A hatlike molecule on the surface of a cell tells the body's immune system whether a virus has invaded.

By PHILIP COHEN
Illustrations by Anne Faust

WHEN ATTACKED BY A VIRUS, a cell doesn't die immediately. Its end is far more insidious. The cell becomes a factory for the enemy, merrily churning out an army of viruses. "You can now think of this cell as a traitor to the body," says Martha Zúñiga, a UCSC professor of biology, tapping a diagram of an infected cell with her knuckle. "It doesn't necessarily want to be a traitor. But it is, and life's tough. You've got to get rid of it."

But getting rid of a cellular Benedict Arnold isn't that easy, says Zúñiga, whose specialty is immunology. Consider the challenge: the body somehow needs to find these infected cells and destroy them. But where does it find clues? Looking for viruses alone won't help because thousands are being pumped into the blood like decoys. Anyway, the real danger is the assembly of new viruses, and that's hidden away inside the cells, away from the immune system's grasp. Yet to completely stop viral infection, the body needs to identify these traitor cells and get rid of them. What's a body to do?

The answer is one that Sigmund Freud would have loved, if he had known a little immunology: the infected cell acts as if it wants to be caught. Its clandestine treachery shows on its surface. Zúñiga studies the molecular mechanisms that guarantee that these traitors betray themselves.

Because the body's immune system is as paranoid as communist-hating U.S. Senator Joseph McCarthy was, cells are constantly tested for their loyalty. The enforcers of the system are killer T cells. A killer nestles up to each cell and probes the cell's surface for what amounts to a password. When the cell is corrupted by virus, the password gets garbled and the killer recognizes the danger. Then the killer T cells show no mercy: they pump the infected cell full of holes and put the factory out of business.

But while a pulp fiction spy may be tipped off to a double agent by a fake accent or a garbled code word, how does the cellular password work? Immunologists have known for years that the password is a clump of proteins called the MHC complex. The complex has three components. Two are proteins from the cell itself: a long or heavy protein chain, and a shorter or light chain. The third component is a tiny piece of viral protein called a peptide. When a cell has all three on its surface, the T cell police annihilate it.

The components can't resist each other. Shortly after the heavy and light chains are produced in the cell, they fold around each other like balloons twisted by a magician into an animal. "To me it looks like a moose head," says Zúñiga. The light chain is the snout, the heavy chain is the neck, ears and antlers.

And Bullwinkle the MHC moose sports the virus peptide like a hat. The T cell border guard, when it extends its suspicious caress, actually pets the moose's head. In an uninfected cell, the killer would feel a peptide from the cell itself--Bullwinkle's normal chapeau. But if instead the T cell feels the heavy chain antlers and a viral peptide cap, it knows it has to act.

But if there is no viral peptide hat, the killer sees no sign of trouble and moves on. Likewise, the killer T cell will ignore a virus that is simply floating in the blood (other parts of the immune system take care of those). Only the combination of the viral and cellular proteins twisted together unleashes the T cell's arsenal.

But Zúñiga wondered, if the light chain isn't part of the code--if the T cell doesn't feel the snout--why is it included in the password? Work from other labs gave her a tip. When other researchers looked at cells that had lost the ability to produce the snout, they found that the whole MHC password never got out to the surface of the cell. So the snout was necessary, but for what? Zúñiga set out to test one possibility. Perhaps, without the light chain, the MHC proteins could not attain the shape necessary for binding the viral peptide.

Every protein is first produced in linear form. "Like the belt of my coat," Zúñiga says, pulling it from the beige hoops of a trenchcoat and forming it into a series of loops. But then "the heavy chain has to fold into the right structure so that it looks like Bullwinkle."

That making of the protein Bullwinkle is directed in part by links called disulphide bonds. These function like knots that a balloon artist adds to confine the balloons to a particular shape. One of Zúñiga's graduate students discovered that the proper formation of those disulphide bonds in the heavy chain requires association of the light chain snout. Furthermore, it is only when Bullwinkle is in one piece that the viral peptide can bind tightly. It turns out that the nose knows how to make a stable complex.

But solving one mystery just opened up another. Zúñiga wondered what stops the empty light and heavy chain duo from going to the cell surface before it acquires a peptide hat. She thinks the answer lies with another protein, called calnexin, that lurks in the compartment of the cell where proteins are assembled. The reader will remember that without the light chain present, the heavy chain did not get released to the cell surface at all. However, when researchers removed part of the calnexin protein, the heavy chain alone did get out.

This discovery of an important role for calnexin suggests another image that Freud might have liked. Calnexin is like the doting mother who won't let the moose leave the cell without its hat in place. In a virally infected cell that means the treachery of the cell is guaranteed to show at the surface because the moose won't leave without the incriminating cargo of the viral cap.

That's the theory. One prediction from it is that there must be a signal that causes calnexin to let go of the MHC complex. Zúñiga says the twisting that goes into making this molecular version of a balloon animal may be the key. Using specific probes to detect different parts of the calnexin and the moose proteins, she has shown that as the password assembles, the MHC bends, exposing some parts of the protein while hiding others. Although she needs to test the idea further, this is consistent with the model that once the viral peptide is securely in place the MHC moose provides no more handhold for calnexin, so calnexin loses its grip, and Bullwinkle sets sail for the surface, the viral peptide worn like a scarlet V cap.

Zúñiga's work has important medical implications. Presenting virus peptides to the T cell guards not only activates the T cells' arsenal, it actually trains them to be better killers for the next traitor they encounter. Modern vaccine developers try to exploit this fact to develop a strong response against viral factories. Knowing more about the development of the MHC password may allow the whole system to be revved up, making the traitors confess even faster. This may be especially important to patients with a depressed immune system.

On the flip side of the coin are patients whose immune systems are overreacting, and destroying healthy tissue. Such patients typically suffer from rheumatoid arthritis, lupus, a type of diabetes and other so-called autoimmune diseases, in which the killer cells are doing their job a little too enthusiastically.

Autoimmune diseases involve T cells that mistake normal, healthy cells for enemies. Understanding better these puzzling and destructive processes may help scientists develop drugs that inhibit the assembly of the MHC complex, and therefore prevent the misunderstood password from reaching the cell's surface, where it will invite unwanted destruction by T cells.

Finding an autoimmune disease treatment will be tricky, says Zúñiga, but the more that is known, the better the chances. With the right drug in hand, she'll be able to stop a moose dead in its tracks.