Scaffold-Based Analytics: Enabling Hit-to-Lead Decisions by Visualizing Chemical Series Linked across Large Datasets.

We present a method for visualizing and navigating large screening datasets while also taking into account their activities and properties. Our approach is to annotate the data with all possible scaffolds contained within each molecule. We have developed a Spotfire visualization, coupled to a fuzzy clustering approach based on the scaffold decomposition of the screening deck, used to drive the hit triage process. Progression decisions can be made using aggregate scaffold parameters and data from multiple datasets merged at the scaffold level. This visualization reveals overlaps that help prioritize hits, highlight tractable series, and posit ways to combine aspects of multiple hits. The structure-activity relationship of a large and complex hit is automatically mapped onto all constituent scaffolds making it possible to navigate, via any shared scaffold, to all related hits. This scaffold "walking" helps address bias toward a handful of potent and ligand-efficient molecules at the expense of coverage of chemical space. We consider two scaffold generation methods and explored their similarities and differences both qualitatively and quantitatively. The workflow of a Spotfire visualization used in combination with fuzzy clustering and structure annotation provides an intuitive view of large and diverse screening datasets. This allows teams to effortlessly navigate between structurally related molecules and enriches the population of leads considered and progressed in a manner complementary to established approaches.

Pyridyl aminothiazoles as potent inhibitors of Chk1 with slow dissociation rates.

Pyridyl aminothiazoles comprise a novel class of ATP-competitive Chk1 inhibitors with excellent inhibitory potential. Modification of the core with ethylenediamine amides provides compounds with low picomolar potency and very high residence times. Investigation of binding parameters of such compounds using X-ray crystallography and molecular dynamics simulations revealed multiple hydrogen bonds to the enzyme backbone as well as stabilization of the conserved water molecules network in the hydrophobic binding region.

Selective activation of the M1 muscarinic acetylcholine receptor achieved by allosteric potentiation.

The forebrain cholinergic system promotes higher brain function in part by signaling through the M(1) muscarinic acetylcholine receptor (mAChR). During Alzheimer's disease (AD), these cholinergic neurons degenerate, therefore selectively activating M(1) receptors could improve cognitive function in these patients while avoiding unwanted peripheral responses associated with non-selective muscarinic agonists. We describe here benzyl quinolone carboxylic acid (BQCA), a highly selective allosteric potentiator of the M(1) mAChR. BQCA reduces the concentration of ACh required to activate M(1) up to 129-fold with an inflection point value of 845 nM. No potentiation, agonism, or antagonism activity on other mAChRs is observed up to 100 microM. Furthermore studies in M(1)(-/-) mice demonstrates that BQCA requires M(1) to promote inositol phosphate turnover in primary neurons and to increase c-fos and arc RNA expression and ERK phosphorylation in the brain. Radioligand-binding assays, molecular modeling, and site-directed mutagenesis experiments indicate that BQCA acts at an allosteric site involving residues Y179 and W400. BQCA reverses scopolamine-induced memory deficits in contextual fear conditioning, increases blood flow to the cerebral cortex, and increases wakefulness while reducing delta sleep. In contrast to M(1) allosteric agonists, which do not improve memory in scopolamine-challenged mice in contextual fear conditioning, BQCA induces beta-arrestin recruitment to M(1), suggesting a role for this signal transduction mechanism in the cholinergic modulation of memory. In summary, BQCA exploits an allosteric potentiation mechanism to provide selectivity for the M(1) receptor and represents a promising therapeutic strategy for cognitive disorders.

Small molecule-mediated allosteric activation of the base excision repair enzyme 8-oxoguanine DNA glycosylase and its impact on mitochondrial function.

8-Oxoguanine DNA glycosylase (OGG1) initiates base excision repair of the oxidative DNA damage product 8-oxoguanine. OGG1 is bifunctional; catalyzing glycosyl bond cleavage, followed by phosphodiester backbone incision via a ß-elimination apurinic lyase reaction. The product from the glycosylase reaction, 8-oxoguanine, and its analogues, 8-bromoguanine and 8-aminoguanine, trigger the rate-limiting AP lyase reaction. The precise activation mechanism remains unclear. The product-assisted catalysis hypothesis suggests that 8-oxoguanine and analogues bind at the product recognition (PR) pocket to enhance strand cleavage as catalytic bases. Alternatively, they may allosterically activate OGG1 by binding outside of the PR pocket to induce an active-site conformational change to accelerate apurinic lyase. Herein, steady-state kinetic analyses demonstrated random binding of substrate and activator. 9-Deazaguanine, which can't function as a substrate-competent base, activated OGG1, albeit with a lower Emax value than 8-bromoguanine and 8-aminoguanine. Random compound screening identified small molecules with Emax values similar to 8-bromoguanine. Paraquat-induced mitochondrial dysfunction was attenuated by several small molecule OGG1 activators; benefits included enhanced mitochondrial membrane and DNA integrity, less cytochrome c translocation, ATP preservation, and mitochondrial membrane dynamics. Our results support an allosteric mechanism of OGG1 and not product-assisted catalysis. OGG1 small molecule activators may improve mitochondrial function in oxidative stress-related diseases.

The M1 muscarinic receptor allosteric agonists AC-42 and 1-[1'-(2-methylbenzyl)-1,4'-bipiperidin-4-yl]-1,3-dihydro-2H-benzimidazol-2-one bind to a unique site distinct from the acetylcholine orthosteric site.

Activation of M1 muscarinic receptors occurs through orthosteric and allosteric binding sites. To identify critical residues, site-directed mutagenesis and chimeric receptors were evaluated in functional calcium mobilization assays to compare orthosteric agonists, acetylcholine and xanomeline, M1 allosteric agonists AC-42 (4-n-butyl-1-[4-(2-methylphenyl)-4-oxo-1-butyl]-piperidine hydrogen chloride), TBPB (1-[1'-(2-methylbenzyl)-1,4'-bipiperidin-4-yl]-1,3-dihydro-2H-benzimidazol-2-one), and the clozapine metabolite N-desmethylclozapine. A minimal epitope has been defined for AC-42 that comprises the first 45 amino acids, the third extracellular loop, and seventh transmembrane domain (Mol Pharmacol 61:1297-1302, 2002). Using chimeric M1 and M3 receptor constructs, the AC-42 minimal epitope has been extended to also include transmembrane II. Phe77 was identified as a critical residue for maintenance of AC-42 and TBPB agonist activity. In contrast, the functional activity of N-desmethylclozapine did not require Phe77. To further map the binding site of AC-42, TBPB, and N-desmethylclozapine, point mutations previously reported to affect activities of M1 orthosteric agonists and antagonists were studied. Docking into an M1 receptor homology model revealed that AC-42 and TBPB share a similar binding pocket adjacent to the orthosteric binding site at the opposite face of Trp101. In contrast, the activity of N-desmethylclozapine was generally unaffected by the point mutations studied, and the docking indicated that N-desmethylclozapine bound to a site distinct from AC-42 and TBPB overlapping with the orthosteric site. These results suggest that structurally diverse allosteric agonists AC-42, TBPB, and N-desmethylclozapine may interact with different subsets of residues, supporting the hypothesis that M1 receptor activation can occur through at least three different binding domains.

Algorithms for the automated selection of fragment-like molecules using single-point surface plasmon resonance measurements.

Fragment-based approaches have added to the arsenal of tools used to identify novel developable leads for drug discovery with high ligand efficiencies. A variety of label-free technologies have been developed and used throughout the industry for fragment screening. Using surface plasmon resonance (SPR) as a fragment screening platform is a relatively new approach. The miniaturization and automation of this technology has led to an associated problem: the large volume of raw data often makes it challenging to analyze and integrate the results of SPR data into the workflow of project teams engaged in the discovery process in a timely fashion. As such, several sets of equations were derived and implemented on Merck's intranet to score single sensorgrams to distinguish stable binders from weak or anomalous binders. This set of equations was optimized and validated on simulated data to both capture "fragment-like" behavior from SPR experiments and filter out much of the anomalous behavior commonly observed. It has subsequently been applied successfully to several in-house discovery programs.

Identification of potent, highly constrained CGRP receptor antagonists.

A novel series of potent CGRP receptor antagonists containing a central quinoline ring constraint was identified. The combination of the quinoline constraint with a tricyclic benzimidazolinone left hand fragment produced an analog with picomolar potency (14, CGRP K(i)=23 pM). Further optimization of the tricycle produced a CGRP receptor antagonist that exhibited subnanomolar potency (19, CGRP K(i)=0.52 nM) and displayed a good pharmacokinetic profile in three preclinical species.

Current and Future Roles of Artificial Intelligence in Medicinal Chemistry Synthesis.

Artificial intelligence and machine learning have demonstrated their potential role in predictive chemistry and synthetic planning of small molecules; there are at least a few reports of companies employing in silico synthetic planning into their overall approach to accessing target molecules. A data-driven synthesis planning program is one component being developed and evaluated by the Machine Learning for Pharmaceutical Discovery and Synthesis (MLPDS) consortium, comprising MIT and 13 chemical and pharmaceutical company members. Together, we wrote this perspective to share how we think predictive models can be integrated into medicinal chemistry synthesis workflows, how they are currently used within MLPDS member companies, and the outlook for this field.

Development of thioquinazolinones, allosteric Chk1 kinase inhibitors.

A high throughput screening campaign was designed to identify allosteric inhibitors of Chk1 kinase by testing compounds at high concentration. Activity was then observed at K(m) for ATP and at near-physiological concentrations of ATP. This strategy led to the discovery of a non-ATP competitive thioquinazolinone series which was optimized for potency and stability. An X-ray crystal structure for the complex of our best inhibitor bound to Chk1 was solved, indicating that it binds to an allosteric site approximately 13A from the ATP binding site. Preliminary data is presented for several of these compounds.

Discovery of Pyrazolocarboxamides as Potent and Selective Receptor Interacting Protein 2 (RIP2) Kinase Inhibitors.

Herein we report the discovery of pyrazolocarboxamides as novel, potent, and kinase selective inhibitors of receptor interacting protein 2 kinase (RIP2). Fragment based screening and design principles led to the identification of the inhibitor series, and X-ray crystallography was used to inform key structural changes. Through key substitutions about the N1 and C5 N positions on the pyrazole ring significant kinase selectivity and potency were achieved. Bridged bicyclic pyrazolocarboxamide 11 represents a selective and potent inhibitor of RIP2 and will allow for a more detailed investigation of RIP2 inhibition as a therapeutic target for autoinflammatory disorders.

Fragment-based discovery of 6-azaindazoles as inhibitors of bacterial DNA ligase.

Herein we describe the application of fragment-based drug design to bacterial DNA ligase. X-ray crystallography was used to guide structure-based optimization of a fragment-screening hit to give novel, nanomolar, AMP-competitive inhibitors. The lead compound 13 showed antibacterial activity across a range of pathogens. Data to demonstrate mode of action was provided using a strain of S. aureus, engineered to overexpress DNA ligase.

Comparison of topological, shape, and docking methods in virtual screening.

Virtual screening benchmarking studies were carried out on 11 targets to evaluate the performance of three commonly used approaches: 2D ligand similarity (Daylight, TOPOSIM), 3D ligand similarity (SQW, ROCS), and protein structure-based docking (FLOG, FRED, Glide). Active and decoy compound sets were assembled from both the MDDR and the Merck compound databases. Averaged over multiple targets, ligand-based methods outperformed docking algorithms. This was true for 3D ligand-based methods only when chemical typing was included. Using mean enrichment factor as a performance metric, Glide appears to be the best docking method among the three with FRED a close second. Results for all virtual screening methods are database dependent and can vary greatly for particular targets.

Optimization of a pyrazoloquinolinone class of Chk1 kinase inhibitors.

The development of 2,5-dihydro-4H-pyrazolo[4,3-c]quinolin-4-ones as inhibitors of Chk1 kinase is described. Introduction of a fused ring at the C7/C8 positions of the pyrazoloquinolinone provided an increase in potency while guidance from overlapping inhibitor bound Chk1 X-ray crystal structures contributed to the discovery of a potent and solubilizing propyl amine moiety in compound 52 (Chk1 IC(50)=3.1 nM).

Scoring of KDR kinase inhibitors: using interaction energy as a guide for ranking.

Within a congeneric series of ATP-competitive KDR kinase inhibitors, we determined that the IC(50) values, which span four orders of magnitude, correlated best with the calculated ligand-protein interaction energy using the Merck Molecular Force Field (MMFFs(94)). Using the ligand-protein interaction energy as a guide, we outline a workflow to rank order virtual KDR kinase inhibitors prior to synthesis. When structural information of the target is available, the ability to score molecules a priori can be used to rationally select reagents. Our implementation allows one to select thousands of readily available reagents, enumerate compounds in multiple poses and score molecules in the active site of a protein within a few hours. In our experience, virtual library enumeration is best used when a correlation between computed descriptors/properties and IC(50) or K (i) values has been established.

Development of 6-substituted indolylquinolinones as potent Chek1 kinase inhibitors.

Through a comparison of X-ray co-crystallographic data for 1 and 2 in the Chek1 active site, it was hypothesized that the affinity of the indolylquinolinone series (2) for Chek1 kinase would be improved via C6 substitution into the hydrophobic region I (HI) pocket. An efficient route to 6-bromo-3-indolyl-quinolinone (9) was developed, and this series was rapidly optimized for potency by modification at C6. A general trend was observed among these low nanomolar Chek1 inhibitors that compounds with multiple basic amines, or elevated polar surface area (PSA) exhibited poor cell potency. Minimization of these parameters (basic amines, PSA) resulted in Chek1 inhibitors with improved cell potency, and preliminary pharmacokinetic data are presented for several of these compounds.

3-(Indol-2-yl)indazoles as Chek1 kinase inhibitors: Optimization of potency and selectivity via substitution at C6.

The development of 3-(indol-2-yl)indazoles as inhibitors of Chek1 kinase is described. Introduction of amides and heteroaryl groups at the C6 position of the indazole ring system provided sufficient Chek1 potency and selectivity over Cdk7 to permit escape from DNA damage-induced arrest in a cellular assay. Enzyme potency against Chek1 was optimized by the incorporation of a hydroxymethyl triazole moiety in compound 21 (Chek1 IC(50)=0.30nM) that was shown by X-ray crystallography to displace one of three highly conserved water molecules in the HI region of the ATP-binding cleft.

Probabilistic neural network multiple classifier system for predicting the genotoxicity of quinolone and quinoline derivatives.

Quinolone and quinoline are known to be liver carcinogens in rodents, and a number of their derivatives have been shown to exhibit mutagenicity in the Ames test, using Salmonella typhimurium strain TA 100 in the presence of S9. Both the carcinogenicity and the mutagenicity of quinolone and quinoline derivatives, as determined by SAS, can be attributed to their genotoxicity potential. This potential, which is measured by genotoxicity tests, is a good indication of carcinogenicity and mutagenicity because compounds that are positive in these tests have the potential to be human carcinogens and/or mutagens. In this study, a collection of quinolone and quinoline derivatives' carcinogenicity is determined by qualitatively predicting their genotoxicity potential with predictive PNN (probabilistic neural network) classification models. In addition, a multiple classifier system is also developed to improve the predictability of genotoxicity. Superior results are seen with the multiple classifier system over the individual PNN classification models. With the multiple classifier system, 89.4% of the quinolone derivatives were predicted correctly, and higher predictability is seen with the quinoline derivatives at 92.2% correct. The multiple classifier system not only is able to accurately predict the genotoxicity but also provides an insight about the main determinants of genotoxicity of the quinolone and quinoline derivatives. Thus, the PNN multiple classifier system generated in this study is a beneficial contributor toward predictive toxicology in the design of less carcinogenic bioactive compounds.