Using a new type of genetic screen, researchers at Harvard Medical School have identified 273 proteins that the AIDS virus needs to survive in human cells, opening up new potential targets for drugs.

Their work, published online on Thursday by Science magazine, used RNA interference to screen thousands of protein-making genes; previously, scientists had identified only 36 human proteins that the virus uses to break into cells, hijack their machinery and start reproducing.

"This is just terrific work," said Dr. Robert C. Gallo, director of the Institute of Human Virology at the University of Maryland and a co-discoverer of the virus. "I think it's destined to be one of the top papers in this field for the decade."

Dr. Anthony S. Fauci, director of the National Institute of Allergy and Infectious Diseases and the government's top AIDS expert, called the Harvard team's work "elegant science," but added a caution.

"It remains to be seen if any of these proteins they identified are useful clinically," Dr. Fauci said. "This is hypothesis-generating, not hypothesis-solving. It creates a lot of work - someone has to go down each of these pathways."

The lead author on the paper, Dr. Stephen J. Elledge, is a geneticist, and this is his first work on the human immunodeficiency virus, which causes AIDS. His previous work has been on cancer, Dr. Elledge said, trying to figure out how cells sense when their chromosomes are broken, and this paper was a collaborative effort.

"I can't even grow H.I.V. in my lab," Dr. Elledge said, so he had to use virus grown by Dr. Judy Lieberman, director of the medical school's AIDS division and one of the co-authors.

Dr. Elledge's team used a library of tens of thousands of different short interfering RNAs, bits of genetic code - each of which, when introduced into a cell, knocks out the cell's ability to make a single protein.

Next, about 21,000 samples of cells, each crippled in its ability to produce one protein, were placed in separate wells on laboratory plates and dosed with the virus.

If the virus could not reproduce normally in a given well, it suggested that the missing protein was one of those it needed.

Of the 273 human proteins identified, only 36 had been previously found by other methods.

The virus, which is itself only a short string of genetic material inside a protective capsule, can make only 15 proteins, so it has to adopt human proteins to its own use.

The advantage of targeting human proteins is that the virus would presumably not be able to mutate to avoid drugs that block them, Dr. Elledge said. Right now, virus strains evolve resistance to antiretroviral drugs, which attack the 15 proteins made by the virus itself, like reverse transcriptase and protease. The mutations force AIDS patients to switch drug regimens - not always successfully.

The disadvantage is that blocking human proteins can, obviously, be fatal to humans. But, as Dr. Gallo pointed out, cancer therapy works that way - doctors try to block proteins that feed fast-growing tumor cells without killing too many other fast-growing cells, like those in the bone marrow.

Right now, Dr. Elledge said, only one drug that targets one of the known human proteins, a receptor called CCR5, has been developed, and it has just won approval.

The new screening technology, known as siRNA, is now used in many laboratories, so this work could theoretically have been done elsewhere, or by using older, more laborious methods.

Dr. Elledge said he benefited from working at Harvard, which could afford the expensive robotics and imaging technology needed.

"And I had lots of collaborators and very dedicated people," he said.

To confirm that the newly identified proteins were important to the life cycle of the virus - which Dr. Elledge described as "opaque" - the team ran further tests on three of them.

Many of the proteins identified by the screen are already known to be important to cells in the immune system, which is the port of entry for H.I.V.

Dr. Abraham L. Brass, a co-author, said the screening method undoubtedly missed other proteins the virus needs, "but the majority of the ones we found are highly likely to play a role in H.I.V. propagation."