Current Topics in Medicinal Chemistry - Volume 11, Issue 15, 2011

The pharmaceutical industry has become highly efficient at synthesizing and testing very large numbers of drug-like molecules, typically through the application of automation in high-throughput screening and combinatorial chemistry. This has yielded something of a “data explosion”, as a result of which it can be challenging simply to manage and store the results, let alone analyze them in any detail. However, for drug discovery to be truly effective, it is important to do just that. Transformation of data into information and information into knowledge can help yield the substantial improvements being demanded of the industry. The scale of the problem is further magnified through the inclusion of vast swathes of genomic data that have become available in recent years (and which have also been generated through the application of automated methodologies). Furthermore, whilst the impact in the GPCR arena is so far modest, we can also anticipate that advances in protein crystallography will yield substantial amounts of structural data that should be very valuable. Chemogenomics is the application of computational methods to make sense of the huge amount of data that sits at this interface between genetics, pharmacology, structural biology and medicinal chemistry. In an attempt to introduce one level of simplification, for this special issue we have chosen to focus on chemogenomics as applied to a single protein family, namely the G-Protein-Coupled Receptors (GPCRs). In its truest sense, however, the technique extends to the prediction of activity for all drug-like molecules at all biological targets and there are elements of that in some of the papers that follow. Given the well-known importance of GPCRs as drug targets, it is perhaps natural to focus there, although the methods described have gained significant traction in other areas as well. This is most notable for the protein kinases where the high levels of similarity in the orthosteric site lend them to such analyses, although results in a somewhat unusual focus on selectivity issues. GPCRs provide an interesting test-bed for chemogenomic methods due to the amount of data available, the relatively diverse nature of their binding site(s) and their on-going importance as drug targets through, for example, the exploitation of targets with hitherto limited chemical tractability, orphan receptors, alternative signaling mechanisms (e.g. beta-arrestin or functional selectivity) and targeted polypharmacology profiles. The principles learned should, however, be applied quite readily to other protein families. Gloriam and Garland have looked at how the available crystal structures can be used to define a reference set of amino acid residues that are accessible for ligand binding in the GPCR transmembrane helical bundle. This set has been used to cluster the receptors to reflect ligand-binding preferences without the “evolutionary noise” associated with non-ligand-binding portions of the sequence. However, the authors show that to achieve high-resolution relationships it is necessary to cluster receptors using a further reduced residue set defined on a per ligand (binding site/mode) basis. This style of analysis has been applied to 3 privileged structures and yields both a rationale for the privileged status as well as predictions for the activity profile across the whole of Family A, with clear applications for ligand/drug design. Frimurer and Hogberg have developed an efficient “physicogenetic” protocol for relating GPCRs with respect to the physicochemical nature of binding sites as well as a “site-directed drug discovery” approach to target- and ligand-based drug design. In one study, a small diverse library directed towards GPR44 (aka CRTh2) generated several useful hit series which were further converted into drug-like lead series. They show the value of including the GPR44 receptor homology model in the design rather than just relying on those targets identified as “close” since, whilst there are similarities in the pockets, there are also some differences that give rise to selectivity. In another analysis, scaffold hopping based on ligand and QSAR data results in the identification of novel MCH1R antagonists. Finally, they have been able to rationalize the known similarity between certain CCR2 and 5HT ligands based on the switch of a key charge-charge interaction point from the Asp of TM3 in 5HT receptors to a Glu on TM7 for CCR2 which influences thinking in terms of receptor similarity......

GPCR binding site-directed techniques are rapidly evolving into powerful tools for modern drug discovery. Many of these approaches bridge chemistry and biology, which are inseparable concepts in nature but are often treated as separate worlds in drug discovery and science in general. This review shows with several examples how focusing on the binding site(s) has a clear advantage when it comes to establishing sequence-correlated pharmacological profiles. By organizing and comparing sequence and structural data it is possible to “borrow” SAR from similar targets to increase the speed of lead-finding and, potentially, to produce ligands for previously intractable receptors. Sequence motifs correlated with ligands can be applied in the design of target-specific focused libraries that are both efficient and cost-effective and should provide increased hit-rates over diversity screening. Furthermore, in the optimization phase, the binding motif approach offers the possibility to identify quickly the most likely off-target candidates to be chosen for selectivity screening, as well as potentially characterizing those pockets which may best be exploited for improved selectivity.

An efficient computational method for hit and lead identification is described. The method that incorporate ligand information from physicogenetically related 7TM receptors, i.e. receptors with similar physicochemical features in the ligand binding pockets, have been developed to aid the construction of pharmacophore queries for mining of vendor and in-house databases to produce small focused libraries for a specific GPCR target. The physicogenetically related targets could be complementary to phylogenetically derived receptors and convey more relevance for the structure-based design approaches suitable for GPCR targets associated with no or limited ligand information. The approach is useful not only in identification of hits but also in the hit-to-lead process as constructed homology receptor models, SAR information and pharmacophore features are collectively utilized in the design of proprietary new lead series. This site-directed drug discovery approach of making smaller receptor-specific libraries displays important advantages over conventional HTSbased generation of hits. The methodology has been exemplified with the CRTH2 receptor, which was associated with minimal ligand information, to produce a small diverse library containing several useful hit series which were further converted into drugable lead series. The use of ligand and QSAR information in scaffold hopping was exemplified with MCH1R antagonists, which had been obtained via chemogenomics-enriched design. Finally, an example on how ligand relationships can be used in identifying receptor relationships was given with CCR2 antagonists to highlight the 3D relationships of GPCR targets not directly evident from either phylogenetic or physicogenetic relationships.

G protein-coupled receptors (GPCRs) share a common architecture consisting of seven transmembrane (TM) domains. Various lines of evidence suggest that this fold provides a generic binding pocket within the TM region for hosting agonists, antagonists, and allosteric modulators. Hence, an automated method was developed that allows a fast analysis and comparison of these generic ligand binding pockets across the entire GPCR family by providing the relevant information for all GPCRs in the same format. This methodology compiles amino acids lining the TM binding pocket including parts of the ECL2 loop in a so-called 1D ligand binding pocket vector and translates these 1D vectors in a second step into 3D receptor pharmacophore models. It aims to support various aspects of GPCR drug discovery in the pharmaceutical industry. Applications of pharmacophore similarity analysis of these 1D LPVs include definition of receptor subfamilies, prediction of species differences within subfamilies in regard to in vitro pharmacology and identification of nearest neighbors for GPCRs of interest to generate starting points for GPCR lead identification programs. These aspects of GPCR research are exemplified in the field of melanopsins, trace amine-associated receptors and somatostatin receptor subtype 5. In addition, it is demonstrated how 3D pharmacophore models of the LPVs can support the prediction of amino acids involved in ligand recognition, the understanding of mutational data in a 3D context and the elucidation of binding modes for GPCR ligands and their evaluation. Furthermore, guidance through 3D receptor pharmacophore modeling for the synthesis of subtype-specific GPCR ligands will be reported. Illustrative examples are taken from the GPCR family class C, metabotropic glutamate receptors 1 and 5 and sweet taste receptors, and from the GPCR class A, e.g. nicotinic acid and 5-hydroxytryptamine 5A receptor.

Thematic Analysis™ is a chemogenomic tool which has been developed and used to aid the process of GPCR drug discovery. This review covers the scientific rationale behind the development of this tool and provides examples of the successful application of the chemogenomic method in both hit finding and hit to lead stages of the drug discovery process.

The potential areas of applications of chemogenomic approaches are very large. Thanks to the large amount of knowledge accumulated during years of research, it is now possible to consider the binding of a ligand to a protein in a much larger context. This knowledge combined with the augmentation of computing capabilities allows global approaches to investigate biological and pharmaceutical problems. Classification of proteins, focused libraries, selectivity profiles and elaboration of new ligands for orphan receptors can all be investigated using chemogenomic. G protein-coupled receptors (GPCRs) constitute a large protein family of significant interest in pharmaceutical research. Despite this interest, and excluding the more than 360 nonolfatory proteins, the endogenous ligands of about 100 GPCRs have still not been identified. The main limitation of GPCRs investigation is the lack of 3D structures. The goal of this review is to present different chemogenomic approaches that can be applied to GPCRs. Three types of such approaches are presented: ligand centered, protein centered and protein-ligand centered approaches. For each of them, current limitations and biases are mentioned.

The degree of applicability of chemogenomic approaches to protein families depends on the accuracy and completeness of pharmacological data and the corresponding level of pharmacological similarity observed among their protein members. The recent public domain availability of pharmacological data for thousands of small molecules on 204 G protein- coupled receptors (GPCRs) provides a firm basis for an in-depth cross-pharmacology analysis of this superfamily. The number of protein targets included in the cross-pharmacology profile of the different GPCRs changes significantly upon varying the ligand similarity and binding affinity criteria. However, with the exception of muscarinic receptors, aminergic GPCRs distinguish themselves from the rest of the members in the family by their remarkably high levels of pharmacological similarity among them. Clusters of non-GPCR targets related by cross-pharmacology with particular GPCRs are identified and the implications for unwanted side-effects, as well as for repurposing opportunities, discussed.

Chemogenomic approaches, which link ligand chemistry to bioactivity against targets (and, by extension, to phenotypes) are becoming more and more important due to the increasing number of bioactivity data available both in proprietary databases as well as in the public domain. In this article we review chemogenomics approaches applied in four different domains: Firstly, due to the relationship between protein targets from which an approximate relation between their respective bioactive ligands can be inferred, we investigate the extent to which chemogenomics approaches can be applied to receptor deorphanization. In this case it was found that by using knowledge about active compounds of related proteins, in 93% of all cases enrichment better than random could be obtained. Secondly, we analyze different cheminformatics analysis methods with respect to their behavior in chemogenomics studies, such as subgraph mining and Bayesian models. Thirdly, we illustrate how chemogenomics, in its particular flavor of ‘proteochemometrics’, can be applied to extrapolate bioactivity predictions from given data points to related targets. Finally, we extend the concept of ‘chemogenomics’ approaches, relating ligand chemistry to bioactivity against related targets, into phenotypic space which then falls into the area of ‘chemical genomics’ and ‘chemical genetics’; given that this is very often the desired endpoint of approaches in not only the pharmaceutical industry, but also in academic probe discovery, this is often the endpoint the experimental scientist is most interested in.

Chemogenomics is an emerging interdisciplinary field that lies in the interface of biology, chemistry, and informatics. Most of the currently used drugs are small molecules that interact with proteins. Understanding protein-ligand interaction is therefore central to drug discovery and design. In the subfield of chemogenomics known as proteochemometrics, protein-ligand-interaction models are induced from data matrices that consist of both protein and ligand information along with some experimentally measured variable. The two general aims of this quantitative multi-structureproperty- relationship modeling (QMSPR) approach are to exploit sparse/incomplete information sources and to obtain more general models covering larger parts of the protein-ligand space, than traditional approaches that focuses mainly on specific targets or ligands. The data matrices, usually obtained from multiple sparse/incomplete sources, typically contain series of proteins and ligands together with quantitative information about their interactions. A useful model should ideally be easy to interpret and generalize well to new unseen protein-ligand combinations. Resolving this requires sophisticated machine-learning methods for model induction, combined with adequate validation. This review is intended to provide a guide to methods and data sources suitable for this kind of protein-ligand-interaction modeling. An overview of the modeling process is presented including data collection, protein and ligand descriptor computation, data preprocessing, machine-learning-model induction and validation. Concerns and issues specific for each step in this kind of data-driven modeling will be discussed.

The G-protein coupled receptors - or GPCRs - comprise simultaneously one of the largest and one of the most multi-functional protein families known to modern-day molecular bioscience. From a drug discovery and pharmaceutical industry perspective, the GPCRs constitute one of the most commercially and economically important groups of proteins known. The GPCRs undertake numerous vital metabolic functions and interact with a hugely diverse range of small and large ligands. Many different methodologies have been developed to efficiently and accurately classify the GPCRs. These range from motif-based techniques to machine learning as well as a variety of alignment-free techniques based on the physiochemical properties of sequences. We review here the available methodologies for the classification of GPCRs. Part of this work focuses on how we have tried to build the intrinsically hierarchical nature of sequence relations, implicit within the family, into an adaptive approach to classification. Importantly, we also allude to some of the key innate problems in developing an effective approach to classifying the GPCRs: the lack of sequence similarity between the six classes that comprise the GPCR family and the low sequence similarity to other family members evinced by many newly revealed members of the family.