Impossible or Merely Difficult? Two Grand Challenges from a Biologist’s Perspective

Abstract

Here I propose two grand challenges for medicinal chemists: the deorphanization of orphan GPCRs via in silico methods and the design of multitarget drugs with enhanced safety and efficacy over current medications.

Medicinal chemistry is a unique discipline straddling the interface of several sciences, including synthetic organic and inorganic chemistry, pharmacology, drug metabolism and distribution, toxicology, and systems and computational biology.¹ Ultimately, medicinal chemistry “..is concerned with the invention, discovery, design, identification and preparation of biologically active compounds... (and) the interpretation of their mode of action at the molecular level ...”¹ (emphasis mine). Here I will propose what might be considered to be “grand challenges” for medicinal chemistry based on the foregoing, made feasible from recent technological and conceptual breakthroughs.

in silico

Grand Challenge #1: Deorphanizing orphan GPCRs . G-protein coupled receptors (GPCRs) represent the largest single target class for therapeutic drug discovery in the human genome.² Among the 900 or so GPCRs in the human genome, more than 50% represent so-called “orphan GPCRs” because validated endogenous ligands have yet to be identified. As recently highlighted by the awarding of the 2012 Chemistry Nobel Prize to Kobilka and Lefkowitz for their pioneering work on GPCR structure and function, the “deorphanization” of even a single GPCR is a notable achievement. Although many GPCRs remain orphans, considerable progress has been made in elucidating the structures of more than 70 GPCR–ligand complexes—mainly via X-ray crystallography.³ Stevens and colleagues have estimated that 18% of nonolfactory GPCRs—including many orphan GPCRs—can now be faithfully modeled; and several studies suggest that, once modeled, they might be fruitfully interrogated by computer-assisted docking.³ Given the steady growth in GPCR–ligand structure determination, one can anticipate that, within a decade, sufficient structural coverage and advances in modeling and computational docking of the GPCR-ome will be achieved so that most members could be computationally interrogated.

Here, the grand challenge is to identify the endogenous agonist for an orphan GPCR. At least two conceptual and technological roadblocks hinder this goal: (1) Modeling the active state of the binding pocket and (2) creating an in silico catalogue of putative endogenous ligands.

Can we accurately model the agonist-bound state of a GPCR for which we have neither ligands nor structure? Although adequate coverage of the GPCR-ome from a modeling perspective can be confidently predicted in the near term, most of the structures will probably continue to represent inactive receptor states with binding pockets that differ from those found of the agonist-bound active state. Certainly, many more agonist-bound states will need to be solved so that the research community can begin to appreciate the types of conformational changes within and outside the binding pocket that occur upon agonist binding. Simultaneously, advances in predicting and modeling agonist conformations—from both a ligand and side-chain perspective—of the binding pocket will be necessary. Thus, a significant aspect of this grand challenge will be to create and refine computational technologies so that they preferentially and faithfully extract or, as may often be the case, infer agonist ligands from large libraries of small drug-like molecules when a putative agonist structure is being interrogated.

Can we create an in silico library encompassing all endogenous ligands? GPCRs can be activated by a bewildering array of endogenous ligands including photons, ions, intermediary metabolites, fragrances, tastants, peptides, neurotransmitters, and autacoids.⁴ To identify the potential endogenous ligands for an orphan GPCR, an in silico database including all of these is needed. Current databases such as the KEGG resource (e.g KEGG COMPOUND http://www.genome.jp/kegg/compound/) and PubChem (http://pubchem.ncbi.nlm.nih.gov/) provide annotation for both endogenous and exogenous bioactives (including peptides). Because many GPCR agonists represent intermediary metabolites, initiatives in metabolomics aiming to identify the universe of human metabolites (e.g., the “Human Metabolomics Library”: http://www.metabolibrary.ca/) will provide many useful structures for computationally interrogating GPCR active states. Although there have been substantial advances in both cataloguing and predicting endogenous bioactive peptides, we will need a large database of such peptides—along with predicted conformations and post-translational modifications—to reliably deorphanize peptide GPCRs in silico.

Clearly, deorphanizing even a single orphan GPCR-ome by in silico or physical methods represents a considerable achievement, as exemplified by studies wherein the orphan peptide osteocalcin was demonstrated to regulate male fertility via the orphan GPCR GPRC6A.⁵ Genome-wide deorphanizing GPCRs along with all other druggable targets (see for instance ref (6) for a pertinent non-GPCR deorphanization success) thus remain a grand challenge for medicinal chemists and biologists.

Research in the authors lab is supported by grants from the National Institute of Mental Health (U19MH82441; RO1MH61887), the National Institute of Drug Abuse (RO1DA017204; NIDA EUREKA Award), the NIMH Psychoactive Drug Screening Program Contract # HHSN-271-2008-00025-C, the Michael Hooker Distinguished Chair of Pharmacology, the International Rett Syndrome Foundation, and the Simon Foundation.

The authors declare no competing financial interest.

Funding Statement

National Institutes of Health, United States