Use of Solvent Mapping for Characterizing the Binding Site and for Predicting the Inhibition of the Human Ether-á-Go-Go-Related K+ Channel.

Molecular dynamics was used to optimize the droperidol-hERG complex obtained from docking. To accommodate the inhibitor, residues T623, S624, V625, G648, Y652, and F656 did not move significantly during the simulation, while F627 moved significantly. Binding sites in cryo-EM structures and in structures obtained from molecular dynamics simulations were characterized using solvent mapping and Atlas ligands, which were negative images of the binding site, were generated. Atlas ligands were found to be useful for identifying human ether-á-go-go-related potassium channel (hERG) inhibitors by aligning compounds to them or by guiding the docking of compounds in the binding site. A molecular dynamics optimized structure of hERG led to improved predictions using either compound alignment to the Atlas ligand or docking. The structure was also found to be suitable to define a strategy for lowering inhibition based on the proposed binding mode of compounds in the channel.

The influence of calculated physicochemical properties of compounds on their ADMET profiles.

We analyzed the influence of calculated physicochemical properties of more than 20,000 compounds on their P-gp and BCRP mediated efflux, microsomal stability, hERG inhibition, and plasma protein binding. Our goal was to provide guidance for designing compounds with desired pharmacokinetic profiles. Our analysis showed that compounds with ClogP less than 3 and molecular weight less than 400 will have high microsomal stability and low plasma protein binding. Compounds with logD less than 2.2 and/or basic pKa larger than 5.3 are likely to be BCRP substrates and compounds with basic pKa less than 5.2 and/or acidic pKa less than 13.4 are less likely to inhibit hERG. Based on these results, compounds with MW < 400, ClogP < 3, basic pKa < 5.2 and acidic pKa < 13.4 are likely to have good bioavailability and low hERG inhibition.

Human PLD structures enable drug design and characterization of isoenzyme selectivity.

Phospholipase D enzymes (PLDs) are ubiquitous phosphodiesterases that produce phosphatidic acid (PA), a key second messenger and biosynthetic building block. Although an orthologous bacterial Streptomyces sp. strain PMF PLD structure was solved two decades ago, the molecular basis underlying the functions of the human PLD enzymes (hPLD) remained unclear based on this structure due to the low homology between these sequences. Here, we describe the first crystal structures of hPLD1 and hPLD2 catalytic domains and identify novel structural elements and functional differences between the prokaryotic and eukaryotic enzymes. Furthermore, structure-based mutation studies and structures of inhibitor-hPLD complexes allowed us to elucidate the binding modes of dual and isoform-selective inhibitors, highlight key determinants of isoenzyme selectivity and provide a basis for further structure-based drug discovery and functional characterization of this therapeutically important superfamily of enzymes.

The use of fake ligands from computational solvent mapping in ligand and structure-based virtual screening.

AIM: Virtual screening selects compounds that resemble a known modulator or compounds that fit into the binding site of a target protein. Computational solvent mapping defines important chemical features for binding to a target protein. Results/methodology: We have tested the ability to use solvent mapping for generating a 'fake' ligand that is a negative image of the binding site. We used this fake ligand as a query for the program ROCS and to define the search space of the docking programs FRED and HYBRID. CONCLUSION: The fake ligands perform comparably to or better than the ligands from crystal structures across a set of ten targets. Thus, the approach is suitable for guiding virtual screening and hit-to-lead optimization.

Discovery of biaryls as RORγ inverse agonists by using structure-based design.

RORγ plays a critical role in controlling a pro-inflammatory gene expression program in several lymphocyte lineages including T cells, γδ T cells, and innate lymphoid cells. RORγ-mediated inflammation has been linked to susceptibility to Crohn's disease, arthritis, and psoriasis. Thus inverse agonists of RORγ have the potential of modulating inflammation. Our goal was to optimize two RORγ inverse agonists: T0901317 from literature and 1 that we obtained from internal screening. We used information from internal X-ray structures to design two libraries that led to a new biaryl series.

Computational solvent mapping in structure-based drug design.

Over the past two decades, solvent mapping has emerged as a useful tool for identifying hot spots within binding sites on proteins for drug-like molecules and suggesting properties of potential binders. While the experimental technique requires solving multiple crystal structures of a protein in different solvents, computational solvent mapping allows for fast analysis of a protein for potential binding sites and their druggability. Recent advances in genomics, systems biology and interactomics provide a multitude of potential targets for drug development and solvent mapping can provide useful information to help prioritize targets for drug discovery projects. Here, we review various approaches to computational solvent mapping, highlight some key advances and provide our opinion on future directions in the field.

In silico prediction of hERG inhibition.

The voltage-gated potassium channel encoded by hERG carries a delayed rectifying potassium current (IKr) underlying repolarization of the cardiac action potential. Pharmacological blockade of the hERG channel results in slowed repolarization and therefore prolongation of action potential duration and an increase in the QT interval as measured on an electrocardiogram. Those are possible to cause sudden death, leading to the withdrawals of many drugs, which is the reason for hERG screening. Computational in silico prediction models provide a rapid, economic way to screen compounds during early drug discovery. In this review, hERG prediction models are classified as 2D and 3D quantitative structure-activity relationship models, pharmacophore models, classification models, and structure based models (using homology models of hERG).

EC144 is a potent inhibitor of the heat shock protein 90.

Alkyne 40, 5-(2-amino-4-chloro-7-((4-methoxy-3,5-dimethylpyridin-2-yl)methyl)-7H-pyrrolo[2,3-d]pyrimidin-5-yl)-2-methylpent-4-yn-2-ol (EC144), is a second generation inhibitor of heat shock protein 90 (Hsp90) and is substantially more potent in vitro and in vivo than the first generation inhibitor 14 (BIIB021) that completed phase II clinical trials. Alkyne 40 is more potent than 14 in an Hsp90α binding assay (IC(50) = 1.1 vs 5.1 nM) as well as in its ability to degrade Her-2 in MCF-7 cells (EC(50) = 14 vs 38 nM). In a mouse model of gastric tumors (N87), 40 stops tumor growth at 5 mg/kg and causes partial tumor regressions at 10 mg/kg (po, qd × 5). Under the same conditions, 14 stops tumor growth only at 120 mg/kg, and does not induce partial regressions. Thus, alkyne 40 is approximately 20-fold more efficacious than 14 in mice.

Can we use docking and scoring for hit-to-lead optimization?

Docking and scoring is currently one of the tools used for hit finding and hit-to-lead optimization when structural information about the target is known. Docking scores have been found useful for optimizing ligand binding to reproduce experimentally observed binding modes. The question is, can docking and scoring be used reliably for hit-to-lead optimization? To illustrate the challenges of scoring for hit-to-lead optimization, the relationship of docking scores with experimentally determined IC(50) values measured in-house were tested. The influences of the particular target, crystal structure, and the precision of the scoring function on the ability to differentiate between actives and inactives were analyzed by calculating the area under the curve of receiver operator characteristic curves for docking scores. It was found that for the test sets considered, MW and sometimes ClogP were as useful as GlideScores and no significant difference was observed between SP and XP scores for differentiating between actives and inactives. Interpretation by an expert is still required to successfully utilize docking and scoring in hit-to-lead optimization.

Virtual screening for kinase targets.

Kinases have become a major area of drug discovery and structure-based design. Hundreds of 3D structures for more than thirty different kinases are available to the public. High structural and sequence homology within the kinase gene family makes the remaining kinases ideal targets for homology modeling and virtual screening. Somewhat surprisingly, however, the number of publications about virtual screening of kinases is very low. Therefore, rather than reviewing the field of virtual screening for kinases, we attempt here a hybrid approach of presenting what is known and common practice together with new studies on CDK2 and SRC kinase. To illustrate the challenges and pitfalls of virtual screening for kinase targets we focus on the question of how ranking is influenced by the database screened, the docking scheme, the scoring function, the activity of the compounds used for testing, and small changes in the binding pocket. In addition, a case study of finding irreversible inhibitors of ErbB2 through in silico screening is presented.

2,3-Disubstituted quinuclidines as a novel class of dopamine transporter inhibitors.

There is considerable interest in developing dopamine transporter (DAT) inhibitors as potential therapies for the treatment of cocaine abuse. We report herein our pharmacophore-based discovery and molecular modeling-assisted rational design of 2,3-disubstituted quinuclidines as potent DAT inhibitors with a novel chemical scaffold. Through 3-D-database pharmacophore searching, compound 12 was identified as a very weak DAT inhibitor with K(i) values of 7.3 and 8.9 microM in [3H]mazindol binding and in inhibition of dopamine reuptake, respectively. Molecular modeling-assisted rational design and chemical modifications led to identification of potent analogues (-)-29 and 34 with K(i) values of 14 and 32 nM for both compounds in binding affinity and inhibition of dopamine reuptake, respectively. Behavioral pharmacological evaluations in rodents showed that 34 has a profile very different from cocaine. While 34 is substantially more potent than cocaine as a DAT inhibitor, it is approximately four times less potent than cocaine in mimicking the discriminative stimulus properties of cocaine in rat. On the other hand, 34 (3-30 mg/kg) lacks either the locomotor stimulant or stereotypic properties of cocaine in mice. Importantly, 34 blocks locomotor stimulant activity induced by 20 mg/kg cocaine in mice, with an estimated ED(50) of 19 mg/kg. Taken together, our data suggest that 34 represents a class of potent DAT inhibitors with a novel chemical scaffold and a behavioral pharmacological profile different from that of cocaine in rodents. Thus, 34 may serve as a novel lead compound in the ultimate development of therapeutic entities for cocaine abuse and/or addiction.

Pharmacophore-based discovery of substituted pyridines as novel dopamine transporter inhibitors.

Abnormal dopamine signaling in brain has been implicated in several conditions such as cocaine abuse, Parkinson's disease and depression. Potent and selective dopamine transporter inhibitors may be useful as pharmacological tools and therapeutic agents. Simple substituted pyridines were discovered as novel dopamine transporter (DAT) inhibitors through pharmacophore-based 3D-database search. The most potent compound 18 has a K(i) value of 79 nM in inhibition of WIN35,248 binding to dopamine transporter and 255 nM in inhibition of dopamine reuptake, respectively, as potent as cocaine. Preliminary structure-activity relationship studies show that the geometry and the nature of the substituents on the pyridine ring determine the inhibitory activity and selectivity toward the three monoamine transporters. The substituted pyridines described herein represent a class of novel DAT inhibitors with simple chemical structures and their discovery provides additional insights into the binding site of DAT.

Discovery of substituted 3,4-diphenyl-thiazoles as a novel class of monoamine transporter inhibitors through 3-D pharmacophore search using a new pharmacophore model derived from mazindol.

Substituted 3,4-diphenyl-1,3-thiazols were identified as a class of novel and potent monoamine transporter inhibitors through a 3-D pharmacophore search using a new pharmacophore model derived from mazindol. The most potent compound (13) has K(i) values of 24 and 23 nM in binding to dopamine transporter and inhibition of dopamine reuptake, respectively.

Structural basis of RasGRP binding to high-affinity PKC ligands.

The Ras guanyl releasing protein RasGRP belongs to the CDC25 class of guanyl nucleotide exchange factors that regulate Ras-related GTPases. These GTPases serve as switches for the propagation and divergence of signaling pathways. One interesting feature of RasGRP is the presence of a C-terminal C1 domain, which has high homology to the PKC C1 domain and binds to diacylglycerol (DAG) and phorbol esters. RasGRP thus represents a novel, non-kinase phorbol ester receptor. In this paper, we investigate the binding of indolactam(V) (ILV), 7-(n-octyl)-ILV, 8-(1-decynyl)benzolactam(V) (benzolactam), and 7-methoxy-8-(1-decynyl)benzolactam(V) (methoxylated benzolactam) to RasGRP through both experimental binding assays and molecular modeling studies. The binding affinities of these lactams to RasGRP are within the nanomolar range. Homology modeling was used to model the structure of the RasGRP C1 domain (C1-RasGRP), which was subsequently used to model the structures of C1-RasGRP in complex with these ligands and phorbol 13-acetate using a computational docking method. The structural model of C1-RasGRP exhibits a folding pattern that is nearly identical to that of C1b-PKCdelta and is comprised of three antiparallel-strand beta-sheets capped against a C-terminal alpha-helix. Two loops A and B comprising residues 8-12 and 21-27 form a binding pocket that has some positive charge character. The ligands phorbol 13-acetate, benzolactam, and ILV are recognized by C1-RasGRP through a number of hydrogen bonds with loops A and B. In the models of C1-RasGRP in complex with phorbol 13-acetate, benzolactam, and ILV, common hydrogen bonds are formed with two residues Thr12 and Leu21, whereas other hydrogen bond interactions are unique for each ligand. Furthermore, our modeling results suggest that the shallower insertion of ligands into the binding pocket of C1-RasGRP compared to C1b-PKCdelta may be due to the presence of Phe rather than Leu at position 20 in C1-RasGRP. Taken together, our experimental and modeling studies provide us with a better understanding of the structural basis of the binding of PKC ligands to the novel phorbol ester receptor RasGRP.