Expression of interest

Membrane proteins are fickle. They play essential roles in cellular processes, govern communication between the cell and its surroundings and orchestrate the flux of materials across the cell membrane. They're also the target of about half the drugs on the market. Yet because they are notoriously more difficult to work with than soluble proteins, researchers know little about their structures. The human genome encodes an estimated 10,000 membrane proteins, but less than 1% have had their three-dimensional structures determined.

Encouraged by the reams of data produced by recent sequencing efforts, protein scientists now hope that membrane protein structure determination will soon become automated and accelerated. Increases in funding and new technologies may provide enough impetus to launch membrane protein chemistry into lush new territory. Scientists with skills in protein bioinformatics or expression of soluble membrane proteins, and those who can use technology to solve these structures, will be in demand as labs around the world aim to add membrane proteins to international databases over the next five years.

But to do so, they must first identify target proteins and expression vectors able to churn out milligram amounts of easily purified protein. And they'll need to develop solutions that will let techniques such as X-ray crystallography reveal the protein's three-dimensional structure. Then they'll need to use these imaging technologies to take pictures of the proteins, and convert them into three-dimensional images — a slow and expensive process.

In 2000, the US National Institutes of Health launched a pilot project called the Protein Structure Initiative (PSI) designed to develop methods and techniques to solve protein structures more quickly and cheaply. Five years and 1,100 structures later, the PSI is entering phase two and going after tougher structures, says PSI director John Norvell. Labs in Canada, Japan and Europe are following suit, intensifying their efforts to solve the structures of membrane proteins in both eukaryotes, whose cells have a nucleus, and prokaryotes, whose cells do not. Some groups will develop technologies; others will focus on solving structures.

The three-dimensional structures of membrane proteins are difficult to determine.LEWIS ET AL.

But the transition from soluble proteins to membrane proteins is full of challenges. Initially, it will need scientists who can identify targets: those that are representative of large families or that are biologically significant. People skilled in protein bioinformatics will be in high demand and will set the tone for the rest of the work. These techniques are likely to be very different from the ones used for the smaller soluble proteins, says Wayne Hendrickson, director of the New York Consortium on Membrane Protein Structure (NYCOMPS) based at the New York Structural Biology Center.

"If we do the bioinformatics right, one structure for a family will inform us greatly about the structures of all the family," says Hendrickson. Although it hasn't started recruiting officially, NYCOMPS will be building a protein-production facility from scratch, and recruitment for this site will dominate its short-term planning.

Expression remains the most significant bottleneck in the process. G-protein-coupled receptors (GPCRs) are one type of membrane protein that interests researchers, but they are difficult to produce. These molecules make up the largest known protein family and play a key role in how other proteins in the cell interact. They are the target for about 30% of today's pharmaceuticals because of this role and because lots of small molecules bind to them.

"Until we can produce large amounts of pure, properly folded human ion channels or human GPCRs, there will be limited progress on structural studies of them," says Stephen Burley, chief scientific officer of Structural GenomiX in San Diego, California.

Part of the problem is that researchers still need to identify expression systems that can produce the relatively large quantities of the notoriously unstable proteins necessary for imaging. Few believe that a single technique will do the trick. The European Membrane Protein Consortium (E-MeP) is one of many groups that has chosen several expression systems — microbial, mammalian and cell-free — and will systematically determine the best for a given protein. "There will be no panacea," says E-MeP coordinator Roslyn Bill, of Aston University in Birmingham, UK.

E-MeP, a 10.4-million (US$12.7-million) European Commission programme, is a large consortium of labs in Britain, France, Germany, Sweden and the Netherlands. Bill says it hopes to get 20 new structures from the 300 target membrane proteins with biological or industrial interest during its five years of funding. Scientists who can develop new cloning and reproduction methods and are comfortable working with eukaryotic and prokaryotic expression systems will excel in their job search, she says.

Two of the PSI's new bases — the Center for Structures of Membrane Proteins at the University of California, San Francisco, and NYCOMPS — will be dedicated to membrane protein structural biology. Both will develop technologies and will test expression systems and detergents, as well as experimenting with other modifications, such as antibodies, to increase the solubility of proteins for crystallization. The centres will need people with specialist knowledge of detergents and an aptitude in protein purification.

Aled Edwards, executive director of the not-for-profit Structural Genomics Consortium (SGC), has room for people with a flair for protein purification. "We're always recruiting," says Edwards.

The SGC has a three-year mandate to put 350 three-dimensional structures into an unrestricted database (see Nature 435, 547; 2005). It is based at the universities of Toronto and Oxford and the Karolinska Institute in Stockholm, with its membrane protein efforts centred in Britain. Rather than focus on technologies, the SGC will target proteins important in certain diseases including diabetes and cancer.

John Norvell (right) examines new laser beam lines for structure determination.

Joining a consortium can offer younger scientists the opportunity to expand the breadth of their knowledge. "We're trying to train scientists so that they have an appreciation of all the areas that are required to be a good membrane protein structural biologist," says Bill. E-MeP set aside money from its budget for professional development (see 'Fringe benefits').

Those with a knack for crystallization might look to industry. Structural GenomiX specializes in structure-based drug discovery, with a primary focus on oncology products. The company uses high-throughput X-ray crystallography to screen the binding activity of drug candidate molecules to target proteins. In July, the company expanded its collaboration with the Cystic Fibrosis Foundation to begin doing drug discovery against mutant forms of a protein involved in the disease. Structural GenomiX had 20 positions — ranging from bioinformatics to protein expression to crystallization — posted on its website in August. The company will hire about a dozen people to support the work, which is being carried out at the New York Structural Genomics Research Consortium.

E. MOSS

Harren Jhoti, chief scientific officer at Astex Therapeutics in Cambridge, UK, says that finding people with the right mix of biophysics and X-ray crystallography can be difficult. Rather than identify potential drug targets, one of the company's initiatives hinges on the protein cytochrome P450, which is involved in drug metabolism.

Many academic team leaders fear that the best scientists will head to industry for higher salaries. "It is extremely difficult to get good people who want to work in an academic environment," says Larry Miercke, a professional associate at the Membrane Protein Expression Center at the University of California, San Francisco. But the opportunity to publish in high-impact journals and to be involved in the full breadth of research dissuades many from leaving.

The ability to switch gears and make the leap from academia to industry is more of a management challenge than a technological one, says Burley, who gave up an endowed professorship at Rockefeller University in 2002 to join Structural GenomiX and do high-throughput crystallography. "The science, the technology is all the same." Deadlines are tight and finances are limited. But, above all, researchers must be able to produce a result on a given day. New recruits must recognize that being able to work with others in pursuit of a common goal is a prized quality in industry.

Laboratories around the world are increasing their efforts to solve the crystal structures of membrane proteins. And scientists who are willing to tackle these fickle molecules should flourish in both industry and academia because right now their skills are in demand.

Fringe Benefits

Postdoctoral researchers Richard Darby and Mohammed Jamshad are motivated by the challenges they face in their work. But they're also enjoying the fringe benefits that come from joining a large consortium such as the European Membrane Protein Consortium.

"The whole project fitted the bill in terms of what I wanted," says Jamshad, a microbiologist. "The research was going to be challenging and there would be plenty of opportunities for training."

Both Darby and Jamshad travelled to Chalmers University in Gothenburg, Sweden, early in their projects to hone their medium-scale fermentation skills — a method that increases protein yields from Pichia pastoris yeast.

"A large consortium such as this opens so many doors and benefits," says Darby, a molecular biologist. "We've had the opportunity to train in other areas where we do not have the expertise or where we have limited expertise."