Public Health:
Lifestyle Changes May Prevent Type 2 Diabetes

The Medical Curriculum Goes Digital

Structural Biology:
Cell Adhesion Receptor Caught on Bended Knee

Depression Linked to Hot Spot in Brain

Digital Library Now Open for Business to HMS, HSPH ID Holders

International Health:
Armenise Speakers Advance Understanding of Cancer Biology, Genomics

New Books:
Autumn Bookshelf

Early Treatment of Seizure Patients May Limit Harm

Racial Disparities and Overuse Shown in Cardiac Revascularization

Some Residents Feel Unprepared for Certain Patient Populations

Potential Diabetes Culprit Identified

Forsyth Institute Seeks Past Patients to Promote Children's Oral Health

Multicultural Affairs Reception Honors the Incoming Students

Mass. Health Donates Books to Countway

New Arrivals Welcomed to Longwood

Honors and Advances

News Briefs

Medical Frontiers: Where Art and Science Meet Global Economics

German Students Help Blaze New Pathway in Munich

Front Page


Cell Adhesion Receptor Caught on Bended Knee

The First Crystal Structure to Be Solved for Shape-Shifting Integrin

A team led by M. Amin Arnaout, HMS professor of medicine and director of the Structural Biology Program at Massachusetts General Hospital, has produced the first crystal structure of an integrin, a mediator of cell adhesion and numerous cell processes that plays a role in tumor angiogenesis, osteoporosis, and some infectious diseases. A shape-shifter as well as a molecular behemoth at nearly 2,000 amino acids, the integrin was a particularly difficult beast to capture, and Arnaout's team spent five years pursuing it. Their findings appear in the Sept. 6 Science Express.

cell structure

Crystal structure ties up loose ends. At left is a ribbon drawing of the crystallized extracellular segment of alpha-V beta-3. The alpha-V and beta-3 subunits appear in blue and yellow, respectively. The structure is severely bent at the kneelike "genu" (arrows). Disordered regions are in gray. At right is a computer model of the straightened alpha-V beta-3, which was developed by extension (135 degrees) and rotation (120 degrees) of the bent structure at the genu. The approximate location and shape of three small and disordered domains are shown in gray. Illustration courtesy of M. Amin Arnaout

Integrins are receptors that are expressed ubiquitously on the surface of cells to give them a certain stickiness by linking them to one another and to the extracellular matrix. Each integrin forms only a weak link, so they work in teams, clustering together where needed most. Unlike classical receptors, integrins naturally lie dormant until they are switched on by a signal from the cell. This level of control is especially crucial for blood cells, which need to circulate freely without sticking to the lining of blood vessels. Once activated, integrins can transmit signals inward to the cytoskeleton, helping to integrate the cell's external and internal environments.

The Bull's Eye

Arnaout's team, which included Jian-Ping Xiong, HMS instructor of medicine at MGH, and Thilo Stehle, HMS assistant professor of pediatrics at MGH, was specifically interested in integrin alpha-V beta-3, an attractive drug development target for several potential applications. The integrin is expressed on many cell types and has a role in allowing new blood vessels to form at tumor sites and in allowing the epithelial cells of breast tumors to metastasize. It also serves as an entry point for adenovirus as well as the viruses that cause foot-and-mouth disease and HIV. And it is expressed in the bone-resorbing cells implicated in osteoporosis.

Thilo Stehle, first author Jian-Ping Xiong, M. Amin Arnaout, and David Scott

A team of HMS researchers has for the first time determined the crystal structure of one of the many integrin proteins found on the surface of cells. From left are Thilo Stehle, first author Jian-Ping Xiong, M. Amin Arnaout, and David Scott. Stehle and Xiong are modeling 3-D glasses for seeing their structure in depth. Photo by Steve Gilbert

Some compounds have been developed that bind and block integrins, but these ligand mimics can have the unwanted agonistic effect of triggering the outside-in signaling that activates the cell. Knowing the detailed structure of the molecule might suggest other ways of blocking its function, as well as helping gene therapists using adenovirus vehicles to understand their targets better.

Previous studies had analyzed the protein with rotary shadowing, which provides a rough outline of the molecule's shape. The integrin showed a large mushroomlike head with two legs protruding outward, which seemed to have a degree of flexibility. "When you have proteins that are flexible, it is very difficult to make crystals," said Arnaout. "That frustrated people." The integrin is large and complex, with carbohydrates tacked on through glycosylation and many cysteines paired in disulfide bonds. These complexities make the protein difficult to manipulate or split without changing the conformation.

"One important advantage is having good collaborators," Arnaout said. One of the key ingredients for crystallizing a tricky molecule is to have large quantities of the protein of choice available. Arnaout took advantage of the pharmaceutical industry's interest in integrins by persuading a group at Merck to supply him with alpha-V beta-3, which they were already producing. Even with this resource, coaxing the protein to form the repetitive lattice needed for a crystal structure was a daunting task.

After the team succeeded in crystallizing the protein, it still faced the task of getting a handle on it. When a beam of x-rays passes through a crystal, the resulting diffraction pattern is like a coded message giving a complex set of coordinates. "What is also needed is to get the strongest x-ray source you can lay your hands on because you need as many diffractions as possible, especially for a huge protein," Arnaout said. Again, he was fortunate to have help from collaborators at Argonne National Laboratory, which houses one of the strongest beams in the world.

A Molecular Flip-flop

The crystal, surprisingly, did not assume the shape it did in previous rotary shadowings; each of the two legs was severely bent at a single joint. "When we saw this structure folded, it was the first indication that you had a focal point in the structure—which we called the "genu," because it's really like a knee—that is, highly flexible," Arnaout said. It was known that integrins shape-shift when switching from an inactive to an active state, and this flexible joint offers a possible mechanism for this change.

The crystal structure shows the integrin in its active state. If a bent shape indicates activity, perhaps the binding site is not the only place to target a drug. "Maybe we should block the ability to flex," Arnaout said, which may keep the protein inactive.

Another discovery that the structure yielded was the association between the two subunits, joined at the head like two frontal lobes. The structure of the domains that form the integrin head is strikingly similar to that of a G-protein, which contains an alpha and a beta subunit that dissociate from each other to transmit signals in the cell and then recombine. Arnaout believes alpha-V beta-3 may have a similar mechanism. Although the two subunits are joined more firmly than a G-protein, the link between the two—a residue of one subunit resting within a ring of residues from the other—contains bonds that could be reversed, allowing a similar regulatory mechanism. "G-proteins and integrins switch on and off," Arnaout said. "The striking similarity in the way the alpha and beta subunits of both proteins associate may reflect functional similarities that could now be explored."

Finally, by mapping previous findings from mutagenesis studies to the structure, the team was able to determine the site for ligand binding, at the very top of the head where the two subunits join. Because the binding site is at the interface of the two subunits, it raises further possibilities for ligand-induced changes in shape.

—Courtney Humphries