Scientists reported yesterday that they had discovered a viral gene that they believed could change human cells to cause immensely diverse biological effects such as cancer and acquired immune deficiency syndrome, or AIDS.

The gene discovery has led the scientists at Harvard and the National Institutes of Health to theorize about a new molecular biological mechanism to explain how a group of HTLV viruses, or human T-cell lymphotropic viruses, can alter an infected cell's machinery, allowing the virus to be produced efficiently and to change the cell's growth properties

The scientists also described structural features of the viral genes that they said offered insights into the function of the viruses that could be important for diagnosis, prevention and therapy of cancers and other diseases.

''It is a major scientific discovery'' in explaining how cancer could be caused by viruses under some circumstances, Dr. William A. Haseltine, the cancer researcher who headed the team, said in an interview. He elaborated on findings that are described in three papers published in the July 27 issue of the journal Science, which was released yesterday.

Need for Testing

Although the research ''gives us hope and a path to follow which we didn't have before,'' such therapies and preventions would take years to develop and would require extensive testing in animals before being tried on humans, Dr. Haseltine said. Moreover, he cautioned, ''It doesn't mean that such developments are possible.'

Another virologist not connected with the reports called the new findings ''very significant'' and also cautioned that the proposed practical benefits were scientifically speculative and would take many years to bring about.

Nevertheless, the findings have led the scientists to devise a new scheme to make vaccines against the human diseases caused by HTLV retroviruses. The scheme involves a concept that scientists call the ''screwed up virus'' approach in which sections of the virus can be changed in such a way as to cripple their ability to cause disease but still stimulate the body to form protective antibodies.

The new concept is important because previous efforts to make vaccines against experimental retroviruses have at best had limited success.

There are three types of HTLV viruses, designated HTLV-1, 2 and 3. Type 1 causes a rare type of leukemia. Type 2 has been linked to another rare type of leukemia but the finding is less certain than type 1 because type 2 has been reported in only one or two patients.

A third type has been implicated as the cause of AIDS and it destroys a blood cell known as the T-4 lymphocyte, leading to potentially lethal infections with microorganisms that ordinarily do not harm unaffected people. The T-4 cell, which is destroyed in AIDS, has been called a master gland of the immune system because it produces hormones that regulate immune functions.

French and U.S. Discoveries

American researchers at the National Cancer Institute call this virus HTLV-3; French researchers at the Pasteur Institute in Paris call it LAV. The researchers have said they believe the viruses are the same.

Scientists know that most viruses cause infection by attaching themselves to cells and then entering to cause damage. However, scientists have been baffled by how some members of one family of viruses can cause cancer, and another viral family member can destroy cells such as in AIDS. Cancer is the result of an uncontrolled growth of cells and AIDS results from the destruction of the T-4 lymphocyte.

In their research at the Dana-Farber Cancer Institute in Boston, Dr. Haseltine's team documented a new phenomenon: the ability of HTLV viruses to reprogram the part of an infected cell's machinery so that the viruses can be produced efficiently

That discovery has led the scientists to begin testing their theory that viral proteins produced by the leukemia- causing types of HTLV turn on genes that make infected cells grow, whereas the AIDS virus either turns off such genes or turns on genes that cause the cells to stop growing.

To begin to understand how the virus changes the infected cell's decoding program, the researchers studied similarities and differences in the molecular structure of the HTLV-1 and HTLV- 2 viruses.

Also, in collaboration with other researchers at Harvard and at the National Cancer Institute, Dr. Haseltine's team determined the structure of an area of the virus known as the surface envelope to produce antigens. These foreign substances lead to the production of protective antibodies.

Role in Infection

The surface envelope plays a key role in allowing the virus to enter human cells to infect them. The surface envelope protein is also regarded as a foreign substance that stimulates the body to produce protective antibodies against the virus.

By testing the ability of the protein contained in the surface envelope or fragments of it to produce antibodies, the scientists theorize they could develop schemes to produce vaccines against the rare types of leukemia and AIDS.

They also theorized that they could create a different type of vaccine based on a virus that was missing a section known as LOR, for long open reading frame. Such viruses ought to be crippled in their ability to cause disease but still should be able to grow slowly and induce protection, the researchers said.

A PAINFULLY THIN young man sits in a hospital wheelchair with his legs and torso oddly twisted - as though he has been flung there. Strapped on his arms are heavily padded splints, which are there to protect him against himself. Without them he is likely to bite and horribly injure his fingers because of an uncontrollable urge to mutilate himself.

Dean La Zar has a good memory, a sense of humor, the normal hopes and fears and even some of the ambitions of a 22-year-old. He also has a rare hereditary disorder called Lesch-Nyhan disease, for which there has never been any cure, effective treatment or much hope.

Self-mutilation is one of the most mysterious hallmarks of the disease. Other bizarre quirks of behavior include the compulsive use of four-letter words, spitting at people and vomiting. The behavior varies from patient to patient and from time to time. Some have tried to poke their own eyes out with pencils or have chewed off their own lips. Terrified of their own compulsion, patients often ask to be restrained.

Many Lesch-Nyhan patients are mentally retarded. All of them are confined to bed or wheelchair. Their bodies produce a huge excess of uric acid, which frequently leads to gout and serious kidney problems. Few have lived beyond their 20's; until recent years, most died in childhood.

There are drugs available that can prolong the life of a patient like Dean La Zar and make him more comfortable, but that is all. ''There isn't anything except physical restraint that will protect them from their behavior.'' says Dr. William Nyhan of the University of California at San Diego, who discovered the disease in the mid-1960's with Michael Lesch, a medical student.

Lesch-Nyhan disease is only one of the more than 3,000 hereditary diseases that are known to be caused by one or another single defective gene. Although most of the diseases are rare, altogether they add up to a large public-health problem, affecting many thousands of people; the exact numbers are unknown.

''The management of most genetic diseases is really quite poor,'' says Dr. Theodore Friedmann, a colleague of Dr. Nyhan's at the University of California at San Diego. ''There are three or four for which treatment is good and that's about all.''

Today, at least a half-dozen research groups - mainly at the medical centers of major universities - are intent on changing that. Lesch-Nyhan disease is one of several genetic diseases for which a revolutionary kind of treatment called gene therapy is being considered. The idea is to transplant copies of a normal gene into some of the patient's cells so that they will do the job the patient's own faulty gene cannot do.

If gene transfer - as scientists often call these experimental techniques - can be done effectively, there are many serious hereditary diseases that might, in principle, be treated. Among them could be cystic fibrosis, once the faulty gene has been identified, and phenylketonuria, a metabolic disorder that causes severe mental retardation in infants. Dr. A. Dusty Miller, a member of a major research group at the Fred Hutchinson Cancer Research Center in Seattle, notes that the human gene for blood factor nine, which is crucial to blood clotting, has been identified and grown in the laboratory. If this can be transplanted successfully, it could cure thousands of patients who suffer from one of the major forms of hemophilia.

Some scientists see gene-therapy research as having important possibilities that go far beyond the treatment of hereditary diseases.

''Most human disease has a genetic component,'' says Dr. Friedmann. ''What's going to come out of this genetic approach to disease is understanding how genes function and then how their expression can be manipulated.'' In the short term, he says, the major impact will be profound understanding of gene function.

Dr. Rainer Storb, a colleague of Dr. Miller's, speculates that gene transfer might come into use as a lifesaving adjunct to the drug treatment of some cancers. Genes might be transplanted, for example, to make the patient's normal cells more resistant to a cell-killing drug that must be used to kill the cancer. Research on dogs and other animals is in progress at the Seattle center to test this idea.

To date, the deceptively straightforward strategy of transplanting good copies of a gene into the patient's cells has been thwarted at many points by serious problems. Doing these transplants requires intimate knowledge of how genes are organized, the chemistry of their regulation in the body and many other details that are still only imperfectly understood.

''There have been unforeseen obstacles and twists all along the way that have held us up,'' says Dr. Stuart Orkin, of Harvard Medical School, a senior member of a collaborative team from Harvard and Massachusetts Institute of Technology's Whitehead Institute, which is one of the leading research groups pursuing gene transfer.

Much progress, however, has been made within the last year or so. Some scientists consider the approval in September of the first national guidelines for gene therapy to be a crucial step in bringing this kind of treatment into practical use. The rules, developed by an advisory panel of the National Institutes of Health, are designed to safeguard the patient's interests and to insure that gene therapy will be medically useful and socially acceptable -that is, it will not be used to ''improve'' the human race, and its effects will be limited to patients alone and would not be inheritable. Proposals must first be approved by local review committees and, finally, by the National Institutes of Health.

However, several of the medical research teams that are most advanced in this field represent collaborations between two or more institutions, and some scientists worry that the need for multiple reviews and approvals may be a serious impediment. Many of the patients most likely to be treated will be already close to death and unable to wait for a protracted process of approval.

Nevertheless, the first proposals for gene therapy under the new guidelines may well be submitted before the end of the year. Just how soon an attempt at human gene therapy will be made is anybody's guess; very possibly it will happen within the next three years.

BEFORE 1953, THE gene itself - the basic unit of heredity - was little more than a philosophical concept. A gene is a genetic blueprint that tells a cell how to make a particular substance and also gives that cell orders for actually doing so. The explosion of recombinant DNA research (known popularly as gene splicing), which began in the early 1970's, has made it possible to identify individual genes and grow many copies of them in genetically engineered bacteria for detailed studies. Today, not only can scientists read the messages of the genes, but they can concoct their own artificial messages - synthetic genes that will function in living cells. Genes have become a known quantity of chemistry and constitute one of the hottest subjects in biological research.

Roughly 100,000 genes make up the human body's full complement of genetic instructions. A gene is essentially a piece of DNA (deoxyribonucleic acid), which consists of a long molecule made up of two twisted chains (the famous double helix) of four different subunits, repeated many times in various combinations. It is the sequence of these chemical subunits, called nucleotides, that constitutes the genetic message.

Probably a total of about six billion nucleotides comprise all the genes on the 23 pairs of chromosomes in a human cell. While substantial genetic variation among individuals creates the diversity of the human population, many passages in the hereditary message do not tolerate change. Some serious diseases result from just one nucleotide error.

By the early 1980's, a dozen or so different human genes that could be useful for gene therapy were identified and grown in the laboratory. Among them was the gene that is faulty in the cells of patients suffering from Lesch-Nyhan disease. Copies of the normal gene have been used to correct cells growing in laboratory cultures that had the metabolic defect characteristic of the cells of Lesch-Nyhan patients.

Foreign genes, including some from humans, were transplanted successfully into animals at least five years ago. More recently, mice have grown to twice their normal size because they grew, from the earliest embryonic stage, with transplanted human growth-hormone genes.

There is a crucial difference, however, between this research in animals and work toward human gene therapy. Because the genes transplanted successfuly into animals generally have been incorporated in the earliest embryonic stage, they are sometimes passed on to the recipient's progeny. Such so-called ''germ-line'' gene transfer is not contemplated in humans and would not be approved by the Government in the forseeable future. Fears that someone might try to ''redesign'' the human race have been a factor in the continuing debate over the ethics of gene splicing. Furthermore, an inheritable gene transplanted into a human might well be loosed beyond recall into the general ''gene pool'' of the human race, whether its effects are good or bad.

The only human gene transplant scientists contemplate today is called somatic-cell gene therapy, because it would modify some of the body cells - somatic cells -to correct a devastating disease, but would not alter the genetic characteristics that a patient could pass on to his or her children. It would be done in children or adults, but not in the embryo. Most doctors and ethicists familiar with the field consider gene transfer of this kind to be an acceptable extension of such modern therapies as bone-marrow transplants.

The amount of animal research that would be needed before a human gene therapy attempt could be made is still a matter of debate. So far, research work has been done mostly on mice. Some of the leaders in gene-transfer research, including Dr. W. French Anderson, of the National Heart, Lung and Blood Institute in Bethesda, Md., believe research in primates should probably be done before attempts are made in humans. In fact, his group has already begun such experiments.

Gene therapy, however, was first tried in humans 15 years ago - a collaboration between an American scientist and physicians in West Germany. The patients were two sisters, 2 and 7 years old, who were clearly doomed to severe mental retardation because of a rare enzyme deficiency. They were deliberately infected with a virus, Shope papilloma virus, that causes warts in rabbits, because of evidence that the virus carried a gene for an enzyme that might restore the children's own enzymatic deficiency. The treatment evidently did no harm, but it also brought no improvement. At the time, the unsuccessful experiment received scant attention.

The next known attempt at gene therapy came in 1980. This time, however, there was a great deal of adverse publicity and the chief scientist who did the work, Dr. Martin J. Cline of the University of California at Los Angeles, was censured by the National Institutes of Health. He also lost important Federal research grants.

Dr. Cline, an expert on diseases of the blood, attempted to treat by gene transplant two victims of thalassemia, a hereditary blood disease that is often painful and seriously deforming. One patient was in Italy, the other in Israel. Neither of the attempts succeeded.

Although Dr. Cline had permission from the Italian and Israeli institutions to perform the gene transplants, he had done them without permission from his own university and therefore violated Federal guidelines governing recombinant DNA research.

At the time, many in the scientific community sharply criticized Dr. Cline's attempts on the grounds that they were premature. But some physicians who had much experience with experimental treatment of patients suffering from grave diseases defended the effort, saying that genetic diseases of the blood were a logical first target of gene therapy and that the desperate state of the patients' health warranted the attempts. Today, much more is known about the genetics and molecular biology of the thalassemias, and many experts suspect the chemistry of these blood diseases is too complex to make them a suitable first target for treatment by gene transfer.

CURRENTLY, THE prime targets for gene transfer are diseases in which doctors could hope for significant improvement in their patients if just a little of the needed gene product was produced in an appropriate tissue so that it would circulate in the blood. Two rare, fatal hereditary diseases of the immune system - each caused by the failure of the body to produce a key enzyme - fit into this category.

The diseases are called ADA and PNP deficiencies, for the enzymes involved in each: adenosine deaminase and purine nucleoside phosphorylase. The normal human genes for the two enzymes have been identified and grown in the laboratory. All that appears to be needed is to deliver the appropriate genetic message in a way that will prompt enough cells to produce the enzyme.

ADA deficiency - which is the focus of research efforts by several teams - is often fatal in infancy. Adenosine deaminase is crucial to the body's natural immune defenses against disease; without it, babies are in mortal danger from infections that would be mild or even unnoticed in a normal person. Some experts estimate that ADA deficiency accounts for roughly 10 percent of babies born with severe and general defects in immunity.

A similarly devastating immune-deficiency disease is PNP deficiency. This condition is so rare that fewer than a dozen cases are known throughout the world.

The current gene-therapy strategy is to transfer copies of the needed gene into cells of the patient's bone marrow in the laboratory, and then drip the marrow sample back into the patient through a vein. Much of the current research is focused on bone marrow for several reasons. It is a tissue that can be removed readily from the patient's body, treated in the laboratory and then returned to the patient. Furthermore, the whole purpose of the gene transplant is to deliver the gene's natural product to tissues where it is needed. Bone-marrow cells give rise to blood and immune-defense cells that circulate throughout the body. This should make them excellent delivery vehicles

For years, it has been routine medical practice to remove samples of bone marrow by sucking them out through hollow needles inserted into accessible bones, such as portions of the upper pelvis. Thus, the problem of getting access to at least one appropriate tissue for gene transplant was solved long ago. But there are many other problems.

The human body has billions of cells. Putting a normal gene for ADA into one, two or 1,000 cells would not even be detectable, let alone have any useful effect on the enzyme deficiency that is the key to the ADA disease. The crucial questions are how to get enough of the marrow cells ''infected'' with the new gene, and how to do it in such a way that the newly acquired genes will produce the enzymes in sufficient quantity.

An important early step toward solving the key delivery problem was made in the late 1970's at Stanford University in the laboratory of Dr. Paul Berg, a Nobel Prize winner and pioneer in recombinant DNA research. Dr. Richard C. Mulligan, a young scientist in the laboratory, together with his colleagues, reasoned that if one wanted to insert a gene into a living cell, the thing to do was use some of the world's pre-eminent natural transplanters of genes -viruses.

In fact, viruses have been described as wandering ''packages'' of genes that drift through the world looking for cells to infect. A virus particle consists of a core of either of the two nucleic acids: DNA, the master chemical of the genes in all living things, or RNA (ribonucleic acid), its companion chemical of heredity. This core is ''packaged'' in an outer coat of protein, which determines what cells the virus can infect. In a conventional virus infection, the virus's nucleic acid enters a cell and subverts its genetic machinery so that the cell gives orders for the construction of a new crop of viruses. The cells are often killed in the process.

To be useful for gene-transfer purposes, a virus would have to be reconstructed so that it carried the gene the scientists wanted to transplant. It would also have to be disarmed so that it would ''infect'' cells without doing any damage and without causing the reproduction of other viruses.

In their early work, Dr. Mulligan and his co-workers made dramatic strides toward the goal. They rebuilt viruses, adding the genes they wanted to transplant. Then they used the remodeled viruses as vehicles to get those genes inside animal cells that were growing in tissue-culture flasks. The scientists were delighted to find that copies of the appropriate genes got into the cells' own genetic machinery and went to work in their new home.

The research team used a much-studied animal cancer virus called SV-40. Unfortunately, it would not serve for gene-transfer use in humans since it is a cancer virus. Furthermore, while the gene was successfully transplanted, viruses went along, too, and produced an infection. There was no easy way of ''disarming'' this virus by rebuilding it.

Now at M.I.T.'s Whitehead Institute, Dr. Mulligan has shifted the focus of his research to another kind of virus and has brought gene transfer far closer to practical reality. He is a member of the collaborative research team at the institute and Harvard. The new research uses another kind of virus, known as retroviruses, and capitalizes on new techniques for redesigning its genetic material. Such rebuilt retroviruses are seen as the most promising vehicles (also known as vectors) for gene transplants. Many retroviruses, like SV-40, are dangerous. Several cause cancer in animals. One, called HTLV-I, has been linked to cancer in humans, and another, HTLV-III, is believed to cause the deadly acquired immunodeficiency syndrome - AIDS.

Retroviruses, however, have the advantage of being small, having only a few genes, and being amenable to drastic manipulations. Dr. Mulligan and his colleagues have used great ingenuity in rebuilding them. The first step is to snip out most of the retrovirus's normal genetic instructions, so that it no longer carries the genetic information to make its protein outer coat or to transform cells to a cancerous state. The retrovirus becomes a carrier of only the genes the scientists want to transplant. These genes can be spliced into the retrovirus's core material by conventional gene-splicing techniques.

Such a remodeled virus can replicate itself with the aid of a less constrained ''helper'' virus that can be added to the cell culture. But how does one then separate the ''helper'' from the disarmed virus that is to serve as a gene carrier?

With Dr. Richard Mann and Dr. David Baltimore, Nobel Prize winner and the director of the Whitehead Institute, and others, Dr. Mulligan removed from a retrovirus a key piece of its nucleic acid that gave instructions for packaging the nucleic-acid core within an outer coat of protein. Without this packaging sequence, viruses could never be assembled and exported from a cell even though all the parts were there. Dr. Mulligan's team showed that a ''helper'' virus denuded of its packaging instruction sequence could help another virus be assembled for export, but it could not itself leave the cell it originally invaded.

Cells containing these helper viruses have been grown to serve as ''packaging cell lines.'' When the stripped-down viruses that are to serve as gene carriers are put into such a cell line, new virus particles are produced and exported. When these new viruses enter cells, they lack the ability to order the manufacture of any new virus particles. They deliver their genes and, in effect, they vanish. The genes of the remodeled viruses are perpetuated in the infected cells, but no new viruses can be made.

The group has been successful with two kinds of retroviruses. One of these is limited largely to infecting mice and rats. The other kind, adapted from viruses discovered in wild mice, is called amphotropic - that is, it is capable of infecting many kinds of cells. These viruses are capable of infecting human cells and therefore may give scientists the delivery vehicle they need for human gene therapy. In research with mice, Dr. Ihor R. Lemischka and Dr. David Williams, colleagues of Dr. Mulligan's, have demonstrated that a transplanted gene can be put into the stem cells of bone marrow - cells that will reproduce to form all the components of blood and the immune defense system. The evidence gives reasonable hope that, in humans, this kind of transplant could deliver the ADA gene into cells so that the marrow can be repopulated and produce enough enzyme to cure the genetic defect.

Other research groups, too, are working toward this goal, but gene-transferring vehicles, or vectors, have not yet been perfected. Sometimes, viruses can recombine their genetic material with that of other related viruses. When this happens, the ostensibly virus-free cell culture includes a few of the undesired ''helper'' viruses in it.

Dr. Miller, of the Seattle team, has been working on another major problem facing all medical scientists involved in gene-therapy research, something they call efficiency of expression. When a cell makes a gene's characteristic product, the process is called gene expression. When the expression is efficient in any human or animal tissue, a lot of the product is made. The problem has been difficult, but some research teams appear to be achieving efficiencies high enough for potential use in treating human patients.

WITH RECENT progress suggesting that gene therapy is coming closer and closer to reality, there has been much speculation about which of the teams most active in the work will do it first, and when. According to Dr. E. Donnall Thomas, associate director for clinical research at the Fred Hutchinson Center for Cancer Research, the more important question is: When will it be done wisely?

Every scientist involved in the research knows that developing a system that will work is only part of the problem. There are some safety questions that probably cannot be answered in advance, and these dictate that the first attempts be made in patients whose diseases are grave.

One of the most perplexing questions has to do with genes called oncogenes that are present in human and animal cells. These genes, sometimes called cancer genes, are believed to help start the process of cancer when they are activated - for example, when another extraneous gene rests nearby and somehow triggers the cancer gene.

Could this happen following a human gene transfer? No one knows. Most experts think this risk would be extremely small, but some suggest risk and benefit would be best balanced by limiting the first round of new gene-transfer attempts in humans to diseases that would otherwise be fatal.

Furthermore, since bone-marrow transplant is an accepted surgical procedure, two questions can be raised: Why not give the patient a transplant of whole bone marrow from a normal donor instead of transplanting a gene into the patient's own marrow? Would it not be a good idea to try marrow transplant first as a test of the efficacy of gene transplant?

Dr. Thomas, a pioneer in bone-marrow transplant, says that these operations are, in fact, a form of gene therapy since the entire blood-forming system of the donor, and all of its genes, is put into the recipient. Bone-marrow transplants have been used in recent years to cure a relatively broad range of diseases that are often fatal, including aplastic anemia, some forms of thalassemia and severe immune diseases like ADA deficiency.

The drawbacks to bone-marrow transplants include the need for a compatible donor whose tissue types match almost perfectly those of the patient and, for some diseases, the need to destroy by drugs or radiation the patient's own bone marrow before the marrow transplant can be done. These requirements make bone-marrow transplant a difficult, life-and-death procedure.

''Lesch-Nyhan syndrome is going to be a very difficult disease to treat by gene transfer in my opinion,'' says Dr. C. Thomas Caskey of Baylor College of Medicine in Waco, Tex., who has made important contributions to the study of gene transfer as a treatment for that disease. The main problem is that the disease affects the brain, and current strategies for transferring genes are not considered likely to deliver the corrective genes to brain cells.

Lesch-Nyhan disease stems from the lack of an enzyme called HPRT, a deficiency that leads to a huge and devastating accumulation of uric acid in the patient's body. Why this also affects the brain is unknown.

An important question is whether transplanting the gene to produce the HPRT enzyme in bone-marrow cells would do anything to improve the effects on brain and behavior that are the worst problem of the Lesch-Nyhan patient. The brain has a natural barrier that keeps out many substances that circulate throughout the rest of the body. This blood-brain barrier might prevent the beneficial effects of that enzyme from reaching the patient's brain, but nobody really knows.

Often, medical scientists find a way out of their dilemma through experiments with animals. However, there is no known animal in which that disease can be reproduced.

IF THE PROMISE OF gene therapy meets some of the hopes of today's research workers, there are many genetic diseases that might be treated. It is not yet clear which one will be attacked first. It is likely to be either ADA, PNP or Lesch-Nyhan disease, but perhaps some new development will bring yet another illness to the fore. In this research, already full of surprises, even the experts are unwilling to make guesses.

Given the fact that Dr. Nyhan identified Lesch-Nyhan disease and that it has been prominent in much of the recent speculation on gene therapy, he was asked recently if he would like to see that disease cured first.

''That's a hard question,'' he replied. ''Really, I'd like to cure them all.''

I HAVE A DRAWER IN MY kitchen that's a jumble of odds and ends. But my junk drawer, cluttered and haphazard though it seems to the uninitiated, has an important function in our house. When something breaks down, we run to that drawer and can usually find a handy way to fix the problem. The old can opener buried under the kabob skewers, for instance, turned out to be the perfect way to reach the ''on'' switch for the coffee grinder when the right way no longer worked. And my daughters are always finding uses for the straws and pom-pons and rubber bands that more sensible mothers would have long since thrown away.

Finding a purpose for useless or even dangerous objects is something scientists like to do, too. That explains the current flurry of investigations into the purpose of one of the most notorious, and most bizarre, life forms on earth, the virus. It turns out, surprisingly, that even this wily villain, implicated in so much human suffering - most recently in the AIDS plague - may actually play an important, productive and positive role in the biosphere.

There may, in fact, be two important ways - besides making us sick - that viruses already have affected us greatly and may continue to do so; and these two ways are directly beneficial. First, they may have helped propel the processes of evolution that brought us this far. Second - and this one is more speculative - they may prove to be a handy method for redirecting evolution toward our own ends, for gradually doing away with inherited genetic diseases.

An emerging theory says that the virus may have set the stage for the genetic variation needed for evolution to occur. And, according to this view, the very qualities the virus used in this process - qualities that also enable it to invade, infect and inhabit a human cell - may be something scientists will learn to exploit. Thus, the formerly feared virus may, lo and behold, present itself to physicians one day as a vehicle for inserting healthy genes into otherwise unhealthy people. It therefore becomes conceivable that medical scientists could prevent the actual expression of an inherited fatal illness lying dormant in a person's genes and normally switched on late in life, such as genetically linked cancers, Huntington's chorea or amyotrophic lateral sclerosis (Lou Gehrig's disease).

All living cells have a complete set of genes arrayed like beads in precise sequences along tiny, spiraling chemical strands called chromosomes. Encoded in the sequence of these genes is the information by which hereditary characteristics are passed from generation to generation. Thus, within these genes - or, to use the chemical term, within a cell's deoxyribonucleic acid (DNA) - is the blueprint for building or rebuilding more cells, that is, for reproduction and regeneration. Within these genes are the secrets of how all succeeding generations of the species may look, react, behave.

Each species has its own characteristic collection of genes, called the genome. The human genome, which is still not wholly mapped out, has 50,000 to 100,000 separate genes; many viruses, on the other hand, have fewer than 10. But those few viral genes have been showing up in surprising ways along the human gene sequence.

How did they get there, and what might their purpose be?

The theory is this: The viral sequences may be helping set up the genetic mutations that propel the process of evolution. Evolution requires that slight changes in the gene - owing to chance or to environmental damage - occur periodically and get passed on to a succeeding generation. If the change is debilitating, the offspring may die off rapidly. But if it is not, the offspring and its offspring may actually thrive, and the species eventually changes, or evolves.

This is where viruses come in. Dr. Malcolm A. Martin, chief of molecular microbiology at the National Institute of Allergy and Infectious Diseases, says that viruses, once they invade cells of the human body - and anyone who has ever had a cold knows they can do that -seem to prompt a kind of chromosomal conga line. The viral genes seem to dance in, cause occasional breaks in the human's strands of chromosomes, and insert themselves into the line. ''The human genome is full of DNA that was rearranged by mechanisms that look like viruses had a hand in it,'' says John M. Coffin, a molecular biologist at the Tufts University School of Medicine.

Viral genes probably got themselves embedded into humans as a result of their particular skills of subterfuge. A virus needs to be wily to overcome its critical handicap: it's not really alive at all, at least not in the general understanding of the term. A virus really is nothing but a floating bag of DNA (or its close cousin, RNA) - a little sack of genetic blueprints for life, not life itself. Outside a cell, viral genes cannot reproduce; they just don't have all the equipment. One particularly insidious type of virus - the retrovirus, which is the type that causes AIDS - must find a compatible host cell, invade it and sabotage it. It tricks the host into using the retrovirus's blueprints instead of its own when it constructs new cells.

In this way, Coffin says, some retroviruses have actually managed to become permanently integrated into a host's germ cells - the cells that develop into sperm and eggs - and to be passed from one host generation to the next. As strong if subtle agents for change, ''viruses seem to have been important forces in evolution,'' he says.

Scientists have established that bits of retrovirus DNA found in humans are similar to bits found in contemporary chimpanzees - and probably in chimpanzees of eons ago. ''The retroviruses we see in humans in 1988 probably got into the germ line tens of millions of years ago as a result of exogenous infection - maybe in an epidemic similar to the AIDS one we're experiencing now,'' says Martin.

In most cases, retroviruses exist silently and innocuously in human or animal DNA, and their presence has not mattered much at all over the relatively short term of a lifetime. But retroviruses are not always benign. In certain strains of mice, for example, retroviruses in the germ cell line have been known spontaneously to start expressing themselves and cause disease, the seeds of which were inborn in that individual mouse. Retroviral murine (mouse) disease most often is leukemia or some other cancer. In humans, retroviruses that are acquired rather than inborn can cause cancer, too - the HTLV-1 virus, for instance, causes adult T-cell lymphoma - as well as the deadly AIDS.

Scientists are now trying to put these scheming parasites to good scientific use, trying to manipulate these master manipulators.

One use for viruses is as a ready-made way to study genetic material. Since a virus's set of genes is quite small, it is relatively easy to observe and understand. But another use is even more intriguing; the virus's basic deceptiveness may eventually present medical science with a method for introducing healthier genes into people with genetic disorders. ''The scientific problems of gene therapy are to get the gene into the cell,'' Coffin says, ''to get the cell back into the body, to get the body to accept it, and to get the body to use that cell instead of the one already there.''

Here is where the virus's skills at sabotage may be most useful. If a virus containing a healthy, functioning gene were to insinuate itself into a human cell and fool the human cell into reproducing the healthy gene instead of its own defective one, this would no doubt be one of the neatest molecular tricks of all time. And according to Coffin, research groups working with animals have successfully used viruses to ''put in desirable traits'' - such as increased growth-hormone production - into mice and pigs. Their experience with animals leads them to believe it is theoretically possible to do something about genetic diseases in human beings.

Still, this road remains fraught with problems, ethical as well as scientific. Who is to say which traits are ''desirable'' in humans? How far should such genetic tinkering be allowed to progress? No one would object to treating potentially fatal genetic illnesses, but what about more normal genetic traits -a tendency, for example, to being fat or nearsighted or short?

The ultimate coup in humanity's struggle to dominate nature would be to twist the virus's chicanery toward our own ends, to use viruses as a way to insert theoretically ''better'' genes in the place of abnormal ones. But our zeal to make each individual as healthy as possible should be tempered with a healthy respect for the power of nature's darkest side. Obviously, questions will be raised about which genes are desirable and which are not, about who will decide these matters and who will be the first to be tampered with. But the questions go even deeper than this. What if inherited diseases fulfill some mysterious function, not for the individual victims, of course, but for the species in general? Will we be hurt if we move too far in the direction of homogeneity? Will we get in trouble for altering our genes? This is not to say that scientists should stop trying to manipulate viruses or correct genetic flaws. But they should do so with an appreciation of how much they still do not know, an appreciation of the ancient Greek notion of hubris, of getting too cocky about man's ability to outwit the gods.

A modified, harmless version of the virus that causes AIDS has been successfully used to deliver gene therapy in laboratory experiments, scientists said today.

In gene therapy, a "vector," like a harmless virus or a virus modified to be rendered harmless in humans, is needed to shuttle the therapeutic gene to its target. But scientists have faced tremendous obstacles in finding an appropriate virus that can be used directly on cells in the human body and not just in cumbersome laboratory procedures.

Scientists at the Salk Institute for Biological Studies in San Diego, and the Whitehead Institute for Biomedical Research in Cambridge, Mass., used a disabled version of H.I.V., the AIDS virus, as a vector in experiments on rats' brains and in laboratory cultures of human cells.

"It is truly a case of turning swords into plowshares," said Dr. Fred Gage of the Salk Institute, a co-author of the research being published on Friday in the journal Science.

Advocates of gene therapy, which has not yet lived up to its promise, believe it could potentially be used to treat thousands of genetic disorders, like as cystic fibrosis or muscular dystrophy as well as cancer and degenerative disorders like Alzheimer's disease.

Although the researchers modified the virus, stripping its power to reproduce and leaving only the parts that allow the virus to integrate itself into the host cells, they recognize that H.I.V. is complicated and not completely understood. So if human experiments are carried out, they said, it would prudent to use cow or monkey immunodeficiency viruses instead of H.I.V.

In the experiments, the modified H.I.V. was able to take the therapeutic repair genes directly to diseased neurons in five female rat brains. One Salk researcher, Dr. Inder M. Verma, said the experiments had proved that "our vectors have the ability to affect a nondividing cell.

That is crucial because most cells in the body are nondividing ones and previous gene therapy vectors worked only on dividing cells. That means cells that were dividing had to be removed from muscles, skin, bone marrow or liver, cultured in the laboratory, inoculated with the vectors and then implanted back into the body.

All that can be skipped with the H.I.V.-derived vector, which can deliver the therapeutic genes directly into nondividing cells in the body, the researchers said.

In a bold but potentially frightening effort to turn one of the world's most virulent killers into a cure, scientists and biotechnology companies are trying to tame the AIDS virus and harness it to treat disease.

The scientists say they have stripped the human immunodeficiency virus of its ability to cause disease, while leaving intact its ability to infect human cells. Such a crippled virus, they say, could be used to deliver genes into human cells for gene therapy.

Several university scientists and biotechnology companies hope to begin clinical trials using the modified H.I.V. viruses to carry genes that they hope can be used to treat diseases such as cancer and hemophilia. At least one attempt will even be made to use the modified H.I.V., the virus that causes AIDS, to treat AIDS itself.

''It would be ironic to cure AIDS with the AIDS virus,'' said Dr. Inder M. Verma, a professor at the Salk Institute for Biological Studies here, who has pioneered the effort to harness H.I.V. for gene therapy. But he added, ''There is a saying that diamond cuts diamond.''

Dr. Verma and others involved in such research say it is virtually inconceivable that anyone treated using the crippled H.I.V. could get AIDS as a result. The gene carrier, which is known as a vector, not only is missing the H.I.V. genes that cause disease but also lacks the ability to replicate and spread in the body, they say.

But some experts note that in rare instances, disabled viruses can recombine with genetic material from other viruses or from the person's own cells to regain the ability to replicate. And even if the actual risks are low, patients are likely to be afraid, and regulators cautious, about injecting patients with even a modified AIDS virus.

''It's a human pathogen that's caused a terrible pandemic, so one needs to be thoughtful about using it, even in a crippled form,'' said Dr. Eric Poeschla, an assistant professor of medicine at the University of California at San Diego.

The Food and Drug Administration, whose approval is required for gene therapy trials, agreed. ''There are a number of scientific questions and safety issues to be addressed before any of that could go forward,'' Dr. Philip Noguchi, the director of the agency's division of cellular and gene therapies, said in an interview. ''We don't quite know what we should be concerned about because the biological understanding has not been developed yet.''

The National Institutes of Health held a meeting of experts last March to explore the use of H.I.V. vectors for gene therapy, which involves inserting genes for a particular function into a patient's cells. But Dr. Noguchi said a wider, more open public hearing would be held before his agency would consider approving the first such trial. ''The public hasn't really weighed in with its own opinion,'' he said

It is the very infectiousness of H.I.V. that makes it attractive for gene therapy, which so far has not lived up to its expectations. People with hemophilia, for instance, have an inherited genetic defect that prevents them from making a crucial protein needed for blood clotting. But if enough of the patient's cells could be provided with the proper gene, the patient could manufacture his own blood clotting factor.

To deliver the genes of interest, scientists generally insert them into debilitated viruses, because viruses spread by delivering their own genetic material into the cells of their target. But gene therapy, which has been tried for about a decade, has in general failed because it has been impossible to deliver enough of the genes, and get them to work long enough, to make enough of the required protein.

In many cases the viral gene carriers are attacked and destroyed by the body's immune system. Some of the viruses used so far allow for only transient production of the protein because they do not incorporate the genes they carry into the chromosomes of the target cells.

One of the most commonly used vectors, derived from mouse leukemia virus, can deliver genes into chromosomes, where, it is hoped, they will operate for a long time. But this vector can do this only when cells are dividing, making it difficult to treat diseases in the brain, liver, heart and other organs in which cells divide rarely if at all. And a promising new vector, based on adeno-associated viruses, appears safe and somewhat effective, but is limited in the size of the genes it can carry. H.I.V., on the other hand, is both cunning at evading the body's immune defenses and can carry large genes. Most important, it is one of a small class of viruses, known as lentiviruses, that can incorporate genes into the chromosomes even of nondividing cells.

''Lentiviral vectors are really a new hope in the field of gene therapy,'' said Dr. Philippe LeBoulch, assistant professor of medicine at Harvard University and chief scientific officer of Genetix Pharmaceuticals Inc., a privately held Massachusetts company planning to use H.I.V.-based vectors in gene therapy.

Dr. Verma's lab has used the vector to incorporate genes into the brain, liver, muscle, bone marrow and retinal cells of rats or mice. The genes seem to function for months.

Still, others say, it is not clear whether the H.I.V. vectors will be efficient enough to make gene therapy work. ''They are still going to have to struggle to get genes into people, even using an H.I.V. vector,'' said Dr. Poeschla of the University of California.

Besides being very infectious, the AIDS virus has been extensively studied, so that scientists know how to modify it. ''We know every nucleotide down to the last base,'' Dr. Verma said.

His laboratory's latest vector removes six of the nine genes in the AIDS virus. One big change that is required is that of the virus's protein coat, which determines which cells it can attach to. The AIDS virus mainly infects certain white blood cells known as T cells. So scientists have replaced the H.I.V. protein coat with that of a cattle virus that seems to be able to infect many different types of human cells.

The only genes taken from the H.I.V. itself to make the vector are those that govern the process by which the viral genetic material is incorporated into the chromosome of the target cell. These genes are put into cells known as packaging cells, which produce the actual vectors.

The vectors themselves -- the particles that would be injected into a patient -- contain the viral proteins needed to incorporate the therapeutic gene into the patient's chromosomes, but they do not contain the genes that direct the creation of those proteins. So after it incorporates the gene into one cell, the viral particle cannot go on to direct the cell to produce more viruses.

Dr. Verma and others point out that the H.I.V. vectors are far more different from a real virus than the crippled viruses that are already used in traditional vaccines for deadly diseases like smallpox and polio.

They also note that the other vectors used for gene therapy, while perhaps not effective, have also not been shown to cause disease in patients or to recombine with other genetic material to become active again, in part because the F.D.A. has strict guidelines for vector purity.

New Developments in Cancer Research

Card 1 of 6

Progress in the field. In recent years, advancements in research have changed the way cancer is treated. Here are some recent updates:

Pancreatic cancer. Scientists are exploring whether the onset of diabetes may be an early warning sign of pancreatic cancer, which is on track to become the second leading cause of cancer-related deaths in the U.S. by 2040.

Chemotherapy. A quiet revolution is underway in the field of cancer treatment: A growing number of patients, especially those with breast and lung cancers, are being spared the dreaded treatment in favor of other options.

Prostate cancer. An experimental treatment that relies on radioactive molecules to seek out tumor cells prolonged life in men with aggressive forms of the disease — the second-leading cause of cancer death among American men.

Leukemia. After receiving a new treatment, called CAR T cell therapy, more than a decade ago, two patients with chronic lymphocytic leukemia saw the blood cancer vanish. Their cases offer hope for those with the disease, and create some new mysteries.

Esophageal cancer. Nivolumab, a drug that unleashes the immune system, was found to extend survival times in patients with the disease who took part in a large clinical trial. Esophageal cancer is the seventh most common cancer in the world.

Dr. Michael Emerman of the Fred Hutchinson Cancer Research Center in Seattle, who had questioned the safety of H.I.V. vectors when Dr. Verma's lab reported two years ago on its first one, said the newer ones have fewer H.I.V. genes and appear to be safer.

''A lot of the concerns I had at the time are being addressed,'' he said. One possible concern, he said, is that people who receive the vectors may test positive for AIDS on a frequently used blood test because they may make antibodies to particular AIDS proteins that remain in the vector.

Still, some scientists say it is best not to play with fire. If a lentivirus is needed because of its ability to infect nondividing cells, then animal viruses could be used instead of H.I.V.

Dr. Poeschla and Dr. Garry P. Nolan, an assistant professor at Stanford University School of Medicine, are independently developing vectors based on the feline immunodeficiency virus, which causes an AIDS-like disease in cats but has not been known to cause disease in people. Oxford Biomedica, a British biotechnology company, and Dr. John C. Olsen of the University of North Carolina are separately developing vectors based on equine infectious anemia virus, which causes a disease in horses.

''It's not so much the real danger,'' Dr. Nolan said. ''It's the psychological issues.'' He said that even some of his own researchers would object to using the AIDS virus. ''I want to do research without looking over my shoulders all the time,'' he said.

Dr. Nolan said there was ''absolutely not'' a chance that someone could get AIDS itself from use of such a vector. But he said there was a very small risk that the vector could recombine with another virus, creating a new virus that could spread in the patient's body or even to someone else.

But proponents of the H.I.V. vectors say that there might be some risk of introducing animal viruses into people as well.

''When you take a virus from one species and move it to another, it's very difficult to predict what will happen,'' said Dr. Didier Trono, professor of genetics and microbiology at the University of Geneva. Dr. Trono, who was previously at the Salk Institute, was a major collaborator with Dr. Verma in developing the lentiviral vector, along with F. H. Gage, who is still at Salk, and Luigi Naldini, now at the University of Turin in Italy.

Even those who insist that the vectors are safe acknowledge that political problems will confine H.I.V. vectors to use in the most serious diseases. ''Clearly one is going to want to start to address the questions in patients with terminal diseases,'' said Dr. Stephen A. Sherwin, president of Cell Genesys, a gene therapy company in Foster City, Calif. The company, of which Dr. Verma is a director, is thinking of trying the H.I.V. vector on cancer.

Genetix also is thinking of inserting a gene that confers resistance to chemotherapy drugs into blood stem cells, the cells in bone marrow that give rise to all blood cells. That would allow patients to undergo stronger chemotherapy without destroying their immune systems or requiring bone marrow transplants. The company has a patented H.I.V. vector, based on work by Dr. Joseph G. Sodroski of Harvard Medical School and the Dana-Farber Cancer Institute

But Dr. Verma wants to use his vector to treat hemophilia and hopes to apply for Federal permission for a clinical trial in a year or so.

Dr. Donald Kohn, director of gene therapy at Childrens Hospital Los Angeles, hopes to apply this year for Federal approval to use the H.I.V. vector to treat AIDS, delivering a gene that interferes with the functioning of the H.I.V. virus. One question is whether this gene will also interfere with the functioning of the vector.

Dr. Kohn, who is also a professor at the University of Southern California School of Medicine, has unsuccessfully tried to use gene therapy for AIDS and other diseases, delivering the genes to blood stem cells.

''We have just not been getting the genes into the bone marrow cells,'' he said. But experiments using the Salk Institute vector on bone marrow cells grown in culture show ''the lentivirus is much, much better,'' he said. The safety of the vector, he added, is not that much of a concern for AIDS patients. ''Certainly, nothing you give them will be worse than what they have,'' he said.

In a bold but potentially frightening effort to turn one of the world's most virulent killers into a cure, scientists and biotechnology companies are trying to tame the AIDS virus and harness it to treat disease.

The scientists say they have stripped the human immunodeficiency virus of its ability to cause disease, while leaving intact its ability to infect human cells. Such a crippled virus, they say, could be used to deliver genes into human cells for gene therapy.

Several university scientists and biotechnology companies hope to begin clinical trials using the modified H.I.V. viruses to carry genes that they hope can be used to treat diseases such as cancer and hemophilia. At least one attempt will even be made to use the modified H.I.V., the virus that causes AIDS, to treat AIDS itself.

''It would be ironic to cure AIDS with the AIDS virus,'' said Dr. Inder M. Verma, a professor at the Salk Institute for Biological Studies here, who has pioneered the effort to harness H.I.V. for gene therapy. But he added, ''There is a saying that diamond cuts diamond.''

Dr. Verma and others involved in such research say it is virtually inconceivable that anyone treated using the crippled H.I.V. could get AIDS as a result. The gene carrier, which is known as a vector, not only is missing the H.I.V. genes that cause disease but also lacks the ability to replicate and spread in the body, they say.

But some experts note that in rare instances, disabled viruses can recombine with genetic material from other viruses or from the person's own cells to regain the ability to replicate. And even if the actual risks are low, patients are likely to be afraid, and regulators cautious, about injecting patients with even a modified AIDS virus.

''It's a human pathogen that's caused a terrible pandemic, so one needs to be thoughtful about using it, even in a crippled form,'' said Dr. Eric Poeschla, an assistant professor of medicine at the University of California at San Diego.

The Food and Drug Administration, whose approval is required for gene therapy trials, agreed. ''There are a number of scientific questions and safety issues to be addressed before any of that could go forward,'' Dr. Philip Noguchi, the director of the agency's division of cellular and gene therapies, said in an interview. ''We don't quite know what we should be concerned about because the biological understanding has not been developed yet.''

The National Institutes of Health held a meeting of experts last March to explore the use of H.I.V. vectors for gene therapy, which involves inserting genes for a particular function into a patient's cells. But Dr. Noguchi said a wider, more open public hearing would be held before his agency would consider approving the first such trial. ''The public hasn't really weighed in with its own opinion,'' he said.

It is the very infectiousness of H.I.V. that makes it attractive for gene therapy, which so far has not lived up to its expectations. People with hemophilia, for instance, have an inherited genetic defect that prevents them from making a crucial protein needed for blood clotting. But if enough of the patient's cells could be provided with the proper gene, the patient could manufacture his own blood clotting factor.

To deliver the genes of interest, scientists generally insert them into debilitated viruses, because viruses spread by delivering their own genetic material into the cells of their target. But gene therapy, which has been tried for about a decade, has in general failed because it has been impossible to deliver enough of the genes, and get them to work long enough, to make enough of the required protein.

In many cases the viral gene carriers are attacked and destroyed by the body's immune system. Some of the viruses used so far allow for only transient production of the protein because they do not incorporate the genes they carry into the chromosomes of the target cells.

One of the most commonly used vectors, derived from mouse leukemia virus, can deliver genes into chromosomes, where, it is hoped, they will operate for a long time. But this vector can do this only when cells are dividing, making it difficult to treat diseases in the brain, liver, heart and other organs in which cells divide rarely if at all. And a promising new vector, based on adeno-associated viruses, appears safe and somewhat effective, but is limited in the size of the genes it can carry. H.I.V., on the other hand, is both cunning at evading the body's immune defenses and can carry large genes. Most important, it is one of a small class of viruses, known as lentiviruses, that can incorporate genes into the chromosomes even of nondividing cells.

''Lentiviral vectors are really a new hope in the field of gene therapy,'' said Dr. Philippe LeBoulch, assistant professor of medicine at Harvard University and chief scientific officer of Genetix Pharmaceuticals Inc., a privately held Massachusetts company planning to use H.I.V.-based vectors in gene therapy.

Dr. Verma's lab has used the vector to incorporate genes into the brain, liver, muscle, bone marrow and retinal cells of rats or mice. The genes seem to function for months.

Still, others say, it is not clear whether the H.I.V. vectors will be efficient enough to make gene therapy work. ''They are still going to have to struggle to get genes into people, even using an H.I.V. vector,'' said Dr. Poeschla of the University of California.

Besides being very infectious, the AIDS virus has been extensively studied, so that scientists know how to modify it. ''We know every nucleotide down to the last base,'' Dr. Verma said.

His laboratory's latest vector removes six of the nine genes in the AIDS virus. One big change that is required is that of the virus's protein coat, which determines which cells it can attach to. The AIDS virus mainly infects certain white blood cells known as T cells. So scientists have replaced the H.I.V. protein coat with that of a cattle virus that seems to be able to infect many different types of human cells.

The only genes taken from the H.I.V. itself to make the vector are those that govern the process by which the viral genetic material is incorporated into the chromosome of the target cell. These genes are put into cells known as packaging cells, which produce the actual vectors.

The vectors themselves -- the particles that would be injected into a patient -- contain the viral proteins needed to incorporate the therapeutic gene into the patient's chromosomes, but they do not contain the genes that direct the creation of those proteins. So after it incorporates the gene into one cell, the viral particle cannot go on to direct the cell to produce more viruses.

Dr. Verma and others point out that the H.I.V. vectors are far more different from a real virus than the crippled viruses that are already used in traditional vaccines for deadly diseases like smallpox and polio.

They also note that the other vectors used for gene therapy, while perhaps not effective, have also not been shown to cause disease in patients or to recombine with other genetic material to become active again, in part because the F.D.A. has strict guidelines for vector purity.

Dr. Michael Emerman of the Fred Hutchinson Cancer Research Center in Seattle, who had questioned the safety of H.I.V. vectors when Dr. Verma's lab reported two years ago on its first one, said the newer ones have fewer H.I.V. genes and appear to be safer.

'A lot of the concerns I had at the time are being addressed,'' he said. One possible concern, he said, is that people who receive the vectors may test positive for AIDS on a frequently used blood test because they may make antibodies to particular AIDS proteins that remain in the vector.

Still, some scientists say it is best not to play with fire. If a lentivirus is needed because of its ability to infect nondividing cells, then animal viruses could be used instead of H.I.V.

Dr. Poeschla and Dr. Garry P. Nolan, an assistant professor at Stanford University School of Medicine, are independently developing vectors based on the feline immunodeficiency virus, which causes an AIDS-like disease in cats but has not been known to cause disease in people. Oxford Biomedica, a British biotechnology company, and Dr. John C. Olsen of the University of North Carolina are separately developing vectors based on equine infectious anemia virus, which causes a disease in horses.

''It's not so much the real danger,'' Dr. Nolan said. ''It's the psychological issues.'' He said that even some of his own researchers would object to using the AIDS virus. ''I want to do research without looking over my shoulders all the time,'' he said.

Dr. Nolan said there was ''absolutely not'' a chance that someone could get AIDS itself from use of such a vector. But he said there was a very small risk that the vector could recombine with another virus, creating a new virus that could spread in the patient's body or even to someone else.

But proponents of the H.I.V. vectors say that there might be some risk of introducing animal viruses into people as well.

''When you take a virus from one species and move it to another, it's very difficult to predict what will happen,'' said Dr. Didier Trono, professor of genetics and microbiology at the University of Geneva. Dr. Trono, who was previously at the Salk Institute, was a major collaborator with Dr. Verma in developing the lentiviral vector, along with F. H. Gage, who is still at Salk, and Luigi Naldini, now at the University of Turin in Italy.

Even those who insist that the vectors are safe acknowledge that political problems will confine H.I.V. vectors to use in the most serious diseases. ''Clearly one is going to want to start to address the questions in patients with terminal diseases,'' said Dr. Stephen A. Sherwin, president of Cell Genesys, a gene therapy company in Foster City, Calif. The company, of which Dr. Verma is a director, is thinking of trying the H.I.V. vector on cancer.

Genetix also is thinking of inserting a gene that confers resistance to chemotherapy drugs into blood stem cells, the cells in bone marrow that give rise to all blood cells. That would allow patients to undergo stronger chemotherapy without destroying their immune systems or requiring bone marrow transplants. The company has a patented H.I.V. vector, based on work by Dr. Joseph G. Sodroski of Harvard Medical School and the Dana-Farber Cancer Institute.

But Dr. Verma wants to use his vector to treat hemophilia and hopes to apply for Federal permission for a clinical trial in a year or so.

Dr. Donald Kohn, director of gene therapy at Childrens Hospital Los Angeles, hopes to apply this year for Federal approval to use the H.I.V. vector to treat AIDS, delivering a gene that interferes with the functioning of the H.I.V. virus. One question is whether this gene will also interfere with the functioning of the vector.

Dr. Kohn, who is also a professor at the University of Southern California School of Medicine, has unsuccessfully tried to use gene therapy for AIDS and other diseases, delivering the genes to blood stem cells.

''We have just not been getting the genes into the bone marrow cells,'' he said. But experiments using the Salk Institute vector on bone marrow cells grown in culture show ''the lentivirus is much, much better,'' he said. The safety of the vector, he added, is not that much of a concern for AIDS patients. ''Certainly, nothing you give them will be worse than what they have,'' he said.