Three-dimensional. A structural biology initiative aims to model the structures of membrane proteins such as this one.

Membrane Proteins in 3D

Whereas people use their eyes and ears to get information, cells rely on proteins that span their outer membranes to scan for chemical signals from the outside world. Now a biotech start-up plans to launch an international consortium to determine the three-dimensional crystal structures of 100 such membrane proteins, many of which represent promising drug targets.

Several "structural genomics" efforts have been launched recently to automate the atomic mapping of proteins, but this is the first to concentrate on membrane proteins. The subjects are a class of proteins called G protein-coupled receptors, which are sensitive to stimuli as varied as hormones and photon-altered pigments. Once these proteins detect a specific signal outside the cell, they let loose a cascade of biochemical messengers that alters the cell's chemistry or gene expression. Scientists would love to know more details, but the receptors are notoriously difficult to work with. Removing them from the cell membrane destroys their normal 3D shape and any hope of understanding what they look like in atomic detail.

The consortium, led by start-up Bio-Xtal in Roubaix, France, plans to orchestrate a concerted effort to find new ways to express, crystallize, and image the proteins. If all goes as planned, starting in April the company will collaborate with four academic labs in France, Germany, and the Netherlands. Bio-Xtal has applied to the European Union for half the estimated 10 million Euro ($9.3 million) cost of the 3-year project, and it expects to raise the rest from pharmaceutical sponsors. Seventeen companies, including Roche, Merck, and Astra Zeneca, have already offered support, says Etienne L'Hermite, Bio-Xtal's manager.

The effort to extend structural genomics to membrane proteins "an excellent idea," says Aled Edwards, a structural biologist at the University of Toronto. But because membrane proteins are difficult to express and crystallize--two necessary steps in determining their structure--the project is certain to face slow going, he says. "Calling something 'genomics' implies automation and high throughput," says Edwards. In this case, "it is a bit of a stretch."

Related site
Bio-Xtal's Web site