Large, medically important class of proteins starts to yield its secrets

GPCRs sit in the cell membrane and sense various molecules outside the cell, including odors, hormones, neurotransmitters, and light. After binding these molecules, GPCRs trigger a specific response inside the cell. Many drugs, including allergy and heart medication and drugs for Parkinson's and Huntington's disease, target these proteins.

This year, a paper published January 19 (Liu et al., Science, 335, 1106) was quickly followed by related publications on the crystal structures of a lipid GPCR (Hanson et al., Science, 335, 851, February 17), the kappa opioid receptor (Wu et al., Nature, 485, 327, March 21), and the nociceptin opioid receptor (Thompson et al., Nature, 485, 395, May 17). The most recent publication is on the 1.8 angstrom high-resolution structure of the A2A adenosine receptor (Liu et al., Science, 336, 232, July 13) and is one of the highest resolution structures to date of a human membrane protein. The structure highlights the receptor and ligand as an allosteric machine controlled by sodium, water, cholesterol, and lipids.

These findings were made possible by technologies developed by the NIH Common Fund Joint Center for Innovative Membrane Technologies (JCIMPT) and the biological questions pursued by the GPCR Network, part of the Protein Structure Initiative:Biology at the National Institute for General Medical Sciences (NIGMS) in Bethesda, Maryland (NIGMS PSI:Biology).

Elusive GPCRs

The precise, three-dimensional arrangement of its constituent atoms is in some ways a protein's ultimate secret. Far more than its amino-acid sequence, the 3D structure holds the key to understanding how a protein interacts with its natural partner molecules in the body or with drug molecules.

But membrane protein structures are as hard to determine as they are valuable, and the most important structures, many of them GPCRs, have often been the most elusive. GPCRs are exceedingly flimsy, fragile proteins when not anchored within their native cell membranes. Coaxing them to line up to form crystals, so that their structures can be determined through X-ray crystallography, has been a formidable challenge for decades. For this reason, the NIH Common Fund highlighted membrane protein expression and stabilization technologies as a priority area of innovation and investment in 2004.

The first high-resolution human GPCR structure determined was the β2 adrenergic receptor, published by Stevens' lab and the lab of Stanford's Brian Kobilka in 2007. In the next three years, only four more GPCR structures made it into the literature -- and finding the structures of the several hundred other medically relevant GPCRs in the human proteome seemed a task for future generations.

Now there is much more optimism. "Ray's network is targeting GPCRs in such a way as to get representative structures from the various GPCR structure groupings, and for these we're probably going to need only about 100 structures," said Bryan Roth, a pharmacologist at the University of North Carolina whose lab collaborated with the GPCR Network to find the kappa opioid and nociceptin receptor structures. "The community may need only a few more years to do that, and of course, already we're finding structures for major GPCR drug targets, such as opioid receptors, that are going to have big impacts on future drug development."

The Technical Breakthroughs

What was the game-changing technical breakthrough?

"I'm always asked that question," Stevens said, "and the answer is that there wasn't just one breakthrough, there were about 15 separate developments, each one critically needed in combination with one another, and they came together after a long time. These are the results of 20 years of commitment by several international groups moving the field forward and the critical decision by the NIH Common Fund to focus on membrane protein technology development as a priority area of importance and investment for all of the NIH institutes."

Some of these breakthroughs have improved researchers' ability to produce and purify GPCRs in quantities sufficient for crystallization -- a process akin to uranium enrichment. Other breakthroughs have been aimed at stabilizing GPCRs, whose core structure is made up of seven membrane-bound helical elements. "When you take away the membrane, these helices have the potential to fall apart," Stevens said. "It is possible that human GPCRs evolved to be unstable as part of their natural function to avoid over stimulation or signaling."

Over the past eight years, researchers with funding from the NIH Common Fund have developed and improved three key GPCR stabilization and crystallization techniques: the use of fusion proteins that stabilize the basic GPCR structure and make it more crystallizable without affecting its function (Chun et al., Structure, 20, 967, 2012); the use of drug compounds that bind to a GPCR and hold it in a specific functional conformation (Xu et al., Science, 332, 322, 2011); and the use of a membrane-mimicking matrix of fat and water molecules, called the lipidic cubic phase (LCP), in which GPCRs, cholesterol, and ligands can form crystals more readily than they do in traditional detergent solutions.

Improvements in these techniques have come by automating and expanding the LCP tool set. Stevens and Assistant Professor Vadim Cherezov built on the LCP work of structural biologist Martin D. Caffrey at the Trinity College Dublin, Ehud Landau at the University of Zurich, and Jurg Rosenbusch at the Biozentrum in Basel, Switzerland. "Stevens and Cherezov really advanced the access and more general use of LCP for crystallizing membrane proteins," said Jean Chin, the NIH program director for both the Common Fund JCIMPT and the GPCR Network.

"The group also developed precrystallization screening tools, next-generation crystallization robotics, imaging technologies, and working with staff at GM/CA beamlines at the Advanced Photon Source (APS) to develop technology to scan the LCP in order to locate the diffraction where the crystals are," said Ward Smith, director of the PSI:Biology program which along with the NIH Common Fund has provided funding for the work.

When GPCR crystals have been formed and isolated, they are sent to a special X-ray crystallography facility -- often the Advanced Photon Source at Argonne National Laboratory in Argonne, Illinois. "At Argonne they've been able to get the X-ray beam smaller and smaller over the years," said Stevens, "and the beamline team at GMCA led by Janet Smith, Robert Fischetti, and Nukri Sanishvili have been incredibly creative and responsive to the needs of our research by developing such novel technologies as the rastering approach to find micro-crystals in the small x-ray beam with our samples as just one example."

Making Sense of the Data

Even when the crystal structure of a GPCR has been solved, scientists need to do further analyses to make full use of the structural data and receptor dynamics are central to their function. GPCRs typically function in part by binding to partner "ligand" molecules that are present outside the cell; that binding event may or may not initiate a biochemical signal inside the cell.

At Scripps Research, the laboratory of Kurt Wüthrich, Nobel Laureate and the Cecil H. and Ida M. Green Professor of Structural Biology, applies nuclear magnetic resonance (NMR) techniques to determine how conformational equilibria and intramolecular rate processes in GPCRs contribute to their binding and signaling functions. "My lab has been working for years with Ray's, combining our NMR expertise with their crystallography expertise," Wüthrich said.

In the paper that started the 2012 wave this past January in Science, for example, Wüthrich's group published the NMR-based study of how different drug compounds affect signaling-related shape changes in the beta-2-adrenergic receptor. Such NMR structure studies were pioneered by Wüthrichand one of the recent innovations has been the use of "microcoils" to enable NMR studies of minute quantities of precious protein samples in solution. Wüthrich's lab, in collaboration with the Stevens lab, has also contributed innovative chemistry techniques for solubilizing GPCRs and other membrane proteins, both for crystallographic and NMR studies. "It's not trivial to make a good solution of a membrane protein," he noted.

The Networks

This quiet revolution in GPCR structure-function hunting has expanded because of the NIGMS's Protein Structure Initiative (PSI), which started in 2000. "Our basic idea was to apply some of the technologies that were enabling faster and cheaper gene sequencing to the problem of determining the structure of proteins and other macromolecules," Smith said. The initiative has been a success on the whole, but progress in solving membrane protein structures lagged at first. Smith noted, "When we renewed the effort in 2010, we decided to emphasize the funding of centers that would concentrate on particular membrane protein families."

The PSI program has been the catalyst in forming efficient structure-hunting research networks. The new NIGMS-supported "GPCR Network," centered at Scripps Research, includes: Stevens's lab; Wüthrich's lab; the lab of LCP innovator Vadim Cherezov, an assistant professor of molecular biology; Professor Hugh Rosen, who leads the NIH Molecular Screening Center at Scripps Research; Professor Peter Kuhn, who is focused on x-ray advancements; Scripps Research Assistant Professor Vsevolod Katrich, who is focused on computational biology; and Professor Ruben Abagyan, University of California, San Diego (UCSD), who also focuses on computational approaches for the GPCR Network.

Some of the network's recent and ongoing projects have been initiated by outside labs that have special expertise in the biology of particular GPCRs. In fact, Smith says, the NIGMS uses its funding to encourage these collaborations. In 2009, for example, UCSD Professor Tracy Handel collaborated with Stevens and others in the GPCR Network to find the structure of CXCR4, an immune receptor on lymphocytes with relevance for HIV infection and the mobilizing of immune cells. The collaboration resulted in the publication of the CXCR4 structure in Science in 2010.

"The GPCR Network has amazing infrastructure and talent that allowed us to rapidly learn from their successes and delve into a challenging area of structural biology, which we could not have done without them," said Handel. "It was an incredible opportunity."

Another collaboration originated in 2008 when Roth, an opioid receptor expert, met Stevens at a conference. Over several years they worked on the problem of crystallizing the kappa opioid receptor. Technological developments from Scripps Research, such as the LCP innovations, made a big difference -- as did Roth's suggestion of a stabilizing ligand, an experimental kappa opioid receptor antagonist called JDTic. Roth's lab also had funding from NIH's National Institute on Drug Abuse to automate post-crystallographic analysis.

"We were able to do a biological validation of the structural model in about three weeks, and that sort of thing used to take about a year," said Roth. "The automated protein engineering and crystallization techniques used in Ray's lab also speeded things up considerably."

Excitement About the Future

Currently only a few labs around the world have access to these new and highly automated technologies for structure determination, but that will change soon. "The NIH Common Fund JCIMPT has made a significant effort in outreach to transfer the technology to other groups interested in GPCR and membrane protein structural biology," Handel said. Roth added, "I think we're going to see a lot more GPCR structures being solved as these technologies are disseminated and these collaborative networks expand."

"We're in only the second year of the GPCR Network, but already the output has been tremendous; we're very excited about the future," said Stevens. "With the continued high impact of the investment in technology, we are only limited by outreach and resources in terms of the output and discoveries of the center. Perhaps most exciting is that we can finally start to address scientific molecular-level questions about GPCR evolution and perhaps even start to make links to human evolution and human cognition, the big picture questions that we are interested in pursuing."