Keys to an antibody's near-perfect fit.

A lock that can be opened with many different keys wouldn't do a very good job of protecting a house. Yet a new study suggests that antibodies--the immune system proteins that tag foreign invaders--begin life as adaptable locks, changing shape to accommodate many different molecular keys. When the antibodies mature, they become like traditional locks, accepting just one key.

In the June 13 Science, researchers at the Lawrence Livermore (Calif.) National Laboratory provide evidence supporting this theory. They determined the three-dimensional structures of a mouse antibody that undergoes nine mutations between its immature and mature forms.

"It's a really elegant demonstration," says Sandra Smith-Gill of the National Cancer Institute in Bethesda, Md. "It gives a potential mechanism for [nonspecific] antibodies to become specific." With this strategy, the immune system may be able to recognize more foreign substances than scientists had previously thought.

Antibody maturation resembles evolution except that it occurs on a much more rapid time scale, says Smith-Gill. A foreign substance that binds to an antibody stimulates the production of more antibodies, many of which contain random mutations. Through a selection process, production of the tightest-binding antibodies flourishes, while cells making the other antibodies shut down.

The Livermore researchers measured the binding of the mouse antibody to hapten, a molecule that stimulates an immune response when attached to a protein. The mature antibody binds hapten 30,000 times better than its immature version.

Martin G. Weigert, a molecular biologist at Princeton University, disagrees with the researchers' interpretation of their findings. Weigert suspects that they have in fact determined the structures of two unrelated antibodies. In the currently accepted theory, each antibody binds to one kind of molecule from the start.

The Livermore researchers are studying other antibodies to determine how general their proposed mechanism is. Some newly produced antibodies might come close to matching their eventual targets and thus wouldn't need as many mutations as the mouse antibody did, says study coauthor Raymond C. Stevens. "Everything depends on how good the initial fit is."

An additional question, adds Smith-Gill, is whether mature antibodies can change shape. Studies in her laboratory suggest that they don't always lock out all but one key.