A private firm in California is nearing commercial production of synthetic insulin by means of genetic engineering, and to that end has intensified its research to begin tests on animals.

But in doing so, Genentech Inc. of south San Francisco has bumped up against federal safety guidelines which require National Institutes of Health approval of experiments using more than 10 liter batches, or about 11 quarts, of genetic material. Genentech will use more.

Genentech's experiments involve a new technology whereby pieces of DNA, the genetic material which determines the hereditary characteristics of all; living cells, are extracted from one organism and spliced together in a test tube with those of another. The so-called recombinant DNA could lead to new life forms: rats that warble, for instance, or plants that produce cheap fuel.

Genentech, which last year found a way to make bacteria produce insulin, is probably the leading American firm involved in work with recombinant DNA.

Its breakthrough last year was heralded as a boon for diabetics, one in 20 of whom are allergic to the animal insulin now in use.

The firm has already contracted with Eli Lilly, the giant pharmaceutical company that controls 80 percent of the insulin market, to develop marketable synthetic insulin as soon as possible.

Genentech's decision to speed up its genetic research emphasizes industry's impatience with the NIH guidelines, which are binding only on researchers receiving federal funds, but which commerical scientists have followed voluntarily until this year.

"Because we've been the leaders in this technology, we've had to deal with the problem of how to get the insulin out to diabetics in this country," Genentech President Bob Swanson said. His firm gets no federal funds.

"We're now working as fast as we can to produce enough for animal tests in the fall. You have to do things on a larger scale to make the tests," he added.

According to the British magazine Nature, Genentech's action sets a precedent for at least four similar small companies involved in genetic research -- three others in this country, including two in Maryland, and one in Luxembourg. Representatives from these companies claimed they have so far stayed within NH guidelines.

These federal standards came about in 1976, when may scientists involved in genetic research voiced fearss that the work might inadvertently result in the creation of a super-resistant new bug, such as deadly virus. But now many of the same scientists are chafing under the constraints, and claim that new evidence indicates that some of them may not be necessary.

"The hazards have been exorbitantly overstated," said DeWitt Stetten, NIH deputy director of science. Stetten said he resigned as chairman of the NIH recombinant DNA advisory committee a year and a half ago, "partly because we were in the process of constructing an elaborate system of precautions for something that wasn't really dangerous in the first place."

He said the dangers of the research are "minuscule and trivial compared to the dangers of all sorts of things we do every day."

Stetten's attitude is mirrored in Congress, which has moved from committee demands for federal regulations binding on industry to silence on the subject this year. But Genentech's move has met with active resistance within California, where several localities have ordinances requiring industry to stay within NIH guidelines.

"Industry's willingness to drop its voluntary compliance with the NIH guidelines . . . portends a trend of noncompliance which, if it continues, will cause the loss of the minimum human health safeguards now provided by the guidelines for recombinant DNA research," wrote California state Sen. Barry Keene and Assemblyman tom Bates in letters of protest to a panel of the Nih.

Swanson insisted that Genentech is still complying with those standards. But the firm's failure to provide details of its insulin experiments at the time of its scale-up in October ended its technical compliance, according the director of the NIH Office of Recombinant DNA Affairs, Willian Gartland.

According to Genentech scientist Mike Ross, the firm's past reluctance to reveal the required information for NIH approval stemmed from its fear that competitors would thereby gain access to its tradesecrets.

"All we've done is to withold proprietary information until they have a way of handling it," Ross said. Gartland said that the problem is now solved, and the committee has "as much authority as the FDA [Food and Drug administration] to protect proprietary information." Consequently, Swanson said he is "putting together the information to give NIH in the fall."

The Recombinant DNA Advisory Committee [Rac], which recommends policy changes to the director of the NIH, has not yet discussed Genentech's scale-up, said RAC Chairwoman Jane Setlow. She added that she did not expect the move to prompt changes in the guidelines.

In the meantime, controversy continues about the possible risks of recombinant DNA research.

In a report of the RAC, Harvard Medical School Professor and Rac member Richard Goldstein cited new research indicating that some of the so-called "safe strains" of recombinant bacteria presently used in genetic research can survive both laboratory disposal treatment and conditions inside human intestings for longer than was originally predicted when NIH standards where established.

But an article in a recent issue of nature claims there is a "built-in safeguard" in the strains that would preclude feared epidemics.