One day last month, he and a reporter spent five minutes searching a Harvard Medical School parking lot for a grimy green Honda Accord. Sinclair had forgotten where he had parked his car. "So much for resveratrol improving memory," he grumbled.

The car retrieved, he drove to Sirtris' headquarters in Cambridge, where he shares an office with Christoph Westphal, the company's chief executive. Westphal disagrees with his colleague that taking resveratrol is a good idea, saying a therapeutic dose cannot be maintained in the bloodstream. He politely conceded Sinclair's position that a lower dose might be effective over the long term.

Sirtris has developed a modified form of resveratrol, called SRT501, that reaches high levels in the bloodstream. It is now being tested in people for safety and its ability to control glucose levels. Westphal plans to gauge the drug's use in treating diabetes and a rare form of dementia caused by defective mitochondria. Sirtris has also developed several other chemicals that activate sirtuins at doses one-thousandth that of resveratrol. The Food and Drug Administration will approve them, if safe and effective, only to treat specific diseases, but it could be inferred that the drugs might thereby extend lifespan. "We believe this is a new therapeutic modality," Westphal said. "We think it can change medical care."

Sirtris has raised $82 million so far. It has a heavyweight group of biotech entrepreneurs on its board and well-known MIT researchers, like Philip Sharp and Robert Langer, on its scientific advisory board. Still, these luminaries could be backing the wrong horse. Across town, that is the view at Sirtris' rival, Elixir Pharmaceuticals.

Elixir has chosen to emphasize leads developed from Kenyon's work on a different set of genes that affect aging, rather than on the sirtuin work of Guarente. "We think the sirtuins are extraordinarily interesting but just don't yet have the proof that these enzymes will be useful in metabolic disease," said William Heiden, Elixir's chief executive.

"It's a proven artifact that resveratrol activates sirtuins," said Peter DiStefano, Elixir's chief scientific officer, referring to Sinclair's 2003 search for such chemicals.

Both Elixir executives argue that the biology of the seven SIRT genes needs to be better worked out before it is worth trying to develop drugs based on them. In their view, it is not even clear if the sirtuins should be activated or inhibited for best effect. Indeed, Elixir has developed several chemicals that inhibit SIRT1's sirtuin.

This has brought about the odd circumstance that Sirtris is trying to activate SIRT1 and Elixir to inhibit it. Can both companies possibly be right? Guarente's consulting agreement with Elixir has expired, and he welcomes the interest that Sirtris is now taking in his work. Both activation and inhibition of SIRT1 could be useful, he says judiciously, if during caloric restriction the gene's activity goes up in some tissues and down in others.

The body's metabolism is governed by such a complex array of genetic circuits that it will be years before the role of the seven SIRTs is fully understood. But if they really embody an ancient mechanism for fortifying the body against disease, then all that is needed is a safe drug that tricks the SIRT genes into thinking feast is famine. The theory is enticing, even if sirtuins and certainty still lie far apart.