Twenty Years of The Magnificent Seven

FROM TWO TO HUNDREDS

Courtesy of John Sholtis

Almost exactly 20 years have passed since the first human GPCR was cloned and characterized by Jeremy Nathans while he was a graduate student at Stanford Medical School.¹ Another two years passed before a team from Merck Research Laboratories, Duke University Medical School, and Howard Hughes Medical Institute, headed by Richard Dixon and Robert Lefkowitz, cloned the β2-adrenergic receptor and showed that it was homologous to rhodopsin.² A frenzy of receptor-gene cloning followed, leading to a new era of molecular pharmacology based on some knowledge of receptor structure – mainly from site-directed mutagenesis studies – in both academic and industrial settings. The terms GPCR and seven-helical receptor quickly became part of the required lexicon for both established and aspiring biologists.

Research groups recorded major advances in understanding the molecular pathophysiology of various human ailments. Perhaps the keynote early advance was the striking elucidation, again by Nathans, of the genetic basis of color blindness and a modern molecular proof of Thomas Young's nearly 200-year-old hypothesis of trichromatic color vision.³⁴ Mutagenesis and biophysical studies of visual pigments also provided a reasonable understanding of the physical chemistry underlying visual spectral tuning.⁵ How do different visual pigments, which absorb light over a wide range of wavelengths from ultraviolet to far red, tune the same basic chromophore cofactor, 11-cis retinal?

But the visual-pigment, or opsin, gene family of typically just a few genes (four in humans) makes up only a small part of the GPCR superfamily. The seminal work of Linda Buck and Richard Axel at Columbia University College of Physicians and Surgeons and Howard Hughes Medical Institute in 1991 – recognized by this year's Nobel Prize in physiology or medicine – suggested that a large number of seven-helical receptors make up the olfactory receptor system in mammals, for example.⁶ GPCRs existed for ligands that were not even endogenous to the organism harboring the receptors.

Taxonomies of hundreds of cloned receptor genes suggested evolutionary relationships among receptors for widely disparate natural ligands: from proteins, to lipids, to nucleotides, to calcium. Soup to nuts, GPCRs could apparently feast on almost anything.

WHO'LL SAVE THE ORPHANS?

Despite the cloning frenzy of the 1990s, few were prepared for the vast number of so-called orphan receptors identified by the Human Genome Project. Orphan receptors – those genes that have the molecular hallmarks of GPCRs, and are expressed, but have no known endogenous ligand – became targets for extensive "deorphaning" programs, mainly in the maturing biotech industry. The true orphans were those receptors without known endogenous ligands or small-molecule pharmacophores. The hope was that many of those receptors might be drug targets for the next generation of block-buster therapeutic agents with billion-dollar-per-year sales. After all, small-molecule GPCR modulators have been estimated to make up more than one-half of all commercial pharmaceuticals, including some former best-sellers such as the β-adrenergic receptor blockers, and more recently the selective serotonin reuptake inhibitors.

Some striking examples of receptor deorphaning have been reported recently. For example, a group at Tularic in South San Francisco showed that the orphan receptor GPR91 was a receptor for the citric acid cycle intermediate, succinate, raising the interesting possibility that small molecules in various metabolic/catabolic pathways might exert distant effects in regulating organ physiology.⁷ The group also showed that another orphan receptor, GPR99, bound α-ketoglutarate. In the case of the succinate receptor, which is expressed preferentially in the kidney, activation seems to be linked to the renin-angiotensin system to elevate blood pressure.

The report from Tularic included a site-directed mutagenesis experiment to probe the active site of the newly deorphaned receptor showing the importance of positively charged amino acid side chains in binding the dicarboxylic acid ligands. The interpretation of the mutagenesis results was aided by a three-dimensional model of the receptor, which was based on the atomic resolution crystal structure of rhodopsin. It was remarkable that a homology model of a GPCR had been reduced to an inset in a multipanel figure.

CALL IN THE CAVALRY

The long-term goal of obtaining an atomic resolution structure of a GPCR had only just been achieved in 2000.⁸ Rhodopsin, the prototypical GPCR, became the first receptor to be crystallized. The rhodopsin structure is now being used to guide work to elucidate the molecular mechanism of receptor photoactivation and is likely to contribute to understanding how agonist ligands activate GPCRs in general.⁹

But given the vast academic-industrial complex working on GPCRs, the questions continuously arise as to why more GPCRs have not been de-orphaned and why more GPCRs have not been crystallized. These two questions dominated a recent issue of the 20-questions series in Nature Reviews Drug Discovery.¹⁰ Here, 20 experts answered 20 questions, and their answers were juxtaposed with background information and perspective. The underlying assumption, of course, is that these questions should be driving ongoing and future research programs.

With respect to deorphaning, it is probably true that efforts have peaked and are on the wane. Optimists suggest that some form of an in silico prediction method might save the day. Yet hopes that bioinformatics might prevail over brute force compound screening were dashed, at least temporarily, by the Tularic report, which mentions in the fourth sentence that GPR91 was expected to be a nucleotide receptor, not a dicarboxylic acid receptor, based on its sequence homology with the purinergic receptor P2Y1. Of course, three-dimensional structure comparisons rather than methods based on sequence homology should improve the chances of predicting correctly that two related receptors might have related ligands, but many more receptor structures will be required to improve any computational method for deorphaning, or rational drug design, for that matter.

So the real bulk of future efforts should be focused on trying to obtain high-resolution structures of additional GPCRs, and not only of the rhodopsin-type family A receptors, but of other classes of GPCRs as well. It is striking that of the 30,000 or so known globular and water-soluble proteins, about 6,000 structures are available. But only 20 or so unique structures of membrane proteins have been reported, that is, only 20 of the more than 20,000 membrane proteins estimated to exist. Complete genome sequences are more available than are atomic-resolution, membrane-protein structures.

Future success in crystallizing GPCRs will require the kind of innovation that was applied to the channel field by Rod MacKinnon and colleagues at the Rockefeller University and Howard Hughes Medical Institute.¹¹ In addition, more emphasis should be placed on molecular biophysical studies of receptors in native membranes using advanced imaging and spectroscopic methods. While drug discovery motors forward, efforts to understand the basic membrane biophysics and structural biology of GPCRs should not take a back seat. And while we may not yet need to call upon Steve McQueen and TM7 to bail us out, we do need to focus on innovative interdisciplinary approaches to achieve success in understanding how 7TM receptors function at the molecular level.

Thomas P. Sakmar is the Richard M. and Isabel P. Furlaud Professor at the Rockefeller University in New York where he served as acting president from 2002 to 2003.

He can be contacted at sakmar@rockefeller.edu.