SAN DIEGO, Calif. — Scientists are taking the first steps toward creating alternative life forms — organisms that use a genetic code different from the one used by all other creatures on earth.

Such organisms, bacteria to start with, would have novel chemical units in their DNA and synthetic building blocks in their proteins. Scientists hope that such organisms can be used to study biochemical processes in new ways and to produce new medical or electronic materials that cannot now be made by living things.

The research goes well beyond current genetic engineering, which involves reshuffling the ordinary components of DNA or proteins into new combinations or moving DNA from one organism to another. Adding completely new elements to DNA and proteins is essentially rewriting the genetic code, the fundamental language of life. As such, it is likely to raise new ethical and safety issues, though there has been no controversy yet because the work is still 5 to 10 years from any practical use.

"We're not trying to imitate nature; we're trying to supplement nature," said Dr. Floyd E. Romesburg, an assistant professor of chemistry at the Scripps Research Institute here. "We're trying to expand the genetic code."

So far, scientists are nowhere near creating truly novel life forms. They have been able to get only one unnatural protein building block at a time substituted for a natural one. And no one has been able to get unnatural DNA to function in living cells, although progress has been made in test tubes.

Despite life's vast diversity, all creatures — from yeast to humans, from microbes that live in near-boiling water to those that tolerate freezing temperatures — spell out their genetic instructions using the same four DNA chemical units, known as bases, which are represented by the letters A, C, G and T. Different three-letter combinations specify amino acids, which are strung together like beads to make the proteins that carry out most functions in a cell. With rare exceptions, all living things use the same 20 amino acids.

The genetic code, then, is a language of four letters used to make 20 words. Despite the limited vocabulary, those words can be used to make the huge variety of sentences and paragraphs that characterize life.

But what if there could be additional genetic letters and words? That, scientists say, would allow organisms to be even more versatile, just as some languages have sounds or express concepts not found in English.

Dr. David A. Tirrell, chairman of chemistry and chemical engineering at the California Institute of Technology, has gotten bacteria to make a protein with the nonstick properties of Teflon by having the microbes substitute an unnatural amino acid for one of the 20 natural ones. He said such a protein might one day be used to make artificial blood vessels. Teflon now is used to make them.

Dr. Tirrell and others also imagine incorporating fluorescent amino acids into proteins. That would allow proteins to be studied in finer detail. And synthetic DNA units are already being used in at least one genetic test.

Scientists say creatures with a truly different genetic code would essentially be alien life forms. Indeed, one of the aims of the research is to see what kinds of life may be possible outside earth.

"We can't think of any transparent reason that these four bases are used on earth," said Dr. Steven A. Benner, a professor of chemistry at the University of Florida, "and it wouldn't surprise me in the slightest if life on Mars used different letters."

The scientists working on the creation of novel organisms say that for now at least, there is no chance that the microbes will run amok. The bacteria created so far that use an unnatural amino acid have to pick up the synthetic component from the medium in which they grow. If they escaped into the wild, they would die or revert to using a natural amino acid.

Still, safety questions will no doubt be raised. "It's a powerful technology," said Dr. Jonathan King, a professor of molecular biology at the Massachusetts Institute of Technology, "and like all powerful technologies it needs appropriate oversight and regulation." Dr. King said, for instance, that proteins with artificial amino acids might elicit allergic reactions if they were used as drugs or in food.

While proteins rarely contain amino acids beyond the normal 20, it is not hard to come up with more. Chemists can synthesize dozens of amino acids. And some organisms create such amino acids for purposes other than making proteins. But these other amino acids are not incorporated into proteins — except when the cell makes a mistake.

So most efforts to have bacteria use synthetic amino acids have hinged on encouraging such mistakes. When bacteria are fed a diet rich in an unnatural amino acid that closely resembles a natural one, they may evolve to prefer the new amino acid, even to the extent that they cannot live without it.

Dr. Andrew Ellington, a professor of chemistry and biochemistry at the University of Texas, has used such forced evolution, completely substituting an unnatural amino acid for one of the 20 natural ones.

Efforts to expand the genetic code have drawn new attention with the publication of two papers in the journal Science on April 20. Both are from scientists at the Scripps Research Institute.

Dr. August Böck, chairman of the Institute of Genetics and Molecular Biology at the University of Munich in Germany, commented in Science that the two papers pointed to "a new realm of biology, bordering the world of chemistry, which will allow experimenters to explore ideas about completely new proteins that were once inconceivable."

One of the papers presented a variation of the error-causing theme.

Scientists introduced a genetic change that crippled an enzyme involved in correcting errors in protein formation. That allowed an unnatural amino acid to be taken up at 24 percent of the locations in all the bacteria's proteins where the amino acid called valine was supposed to go. The work was led by Dr. Paul Schimmel at Scripps and Dr. Philippe Marlière of Genoscope, a French research institute, and Evologic, a biotech company.

The second Scripps paper took a different approach. Instead of substituting a new amino acid for one of the 20, the scientists introduced a 21st amino acid. And instead of widespread substitution, they put the new amino acid in a specific spot of their choosing. They did this by creating special molecules to deliver this amino acid to the cell's protein-making machinery.

This work was led by Dr. Peter G. Schultz, a chemistry professor who is also director of the Genomics Institute of the Novartis Research Foundation.

Some experts said the work paved the way for introducing more than one new amino acid into bacteria and doing so with a precision previously unobtainable. "I would say it's a major, major advance," said Dr. Uttam RajBhandary, a molecular biologist at M.I.T., who is doing similar work.

But if scientists are going to add new amino acids this way, they have to specify where in the proteins these new amino acids should go. So they must put the genetic code for the new amino acids into the bacteria's DNA at the right spots.

There is only one problem: there is no sequence of DNA letters that encode for amino acids that nature has not encountered before. With four DNA bases, there are 64 possible three-letter combinations, called codons, which can specify an amino acid. But 61 of them are already used for the 20 natural amino acids. (There are duplications; for instance, six different codons specify leucine.)

When the cell encounters one of three remaining codons that do not specify an amino acid, it stops building the protein. Dr. Schultz picked one of those codons as the code for his new amino acid.

But this approach has an obvious problem. What if the bacteria's DNA naturally contains this codon at spots where protein formation really is supposed to stop? If the 21st amino acid were inserted instead at such spots, erroneous proteins would be made that could kill the organisms.

Dr. Schultz said the codon he chose was rarely used by this bacterium, so this would not be a significant problem. Still, there are only three codons that do not already code for an ordinary amino acid, limiting the number of new amino acids that can be introduced this way.

So if scientists want to introduce many new amino acids, new codons will be needed. That is why they are trying to add letters to the genetic alphabet. If DNA consisted of six bases - say, A, C, G, T, X and Y - there could be 216 codons instead of 64.

Such artificial DNA bases have been made by Dr. Benner in Florida, Dr. Romesburg at Scripps and Dr. Eric T. Kool, a chemistry professor at Stanford. Besides fitting into the double helix of DNA, each artificial base must pair with only one artificial counterpart, just as A always pairs with T, and C with G. Such pairing is essential for accurate DNA replication.

Dr. Benner in one case managed to use an artificial DNA base to produce a protein with an unnatural amino acid — but only in a test tube. It has been extremely difficult to use natural enzymes to replicate DNA that contains artificial bases, even in the test tube. And when artificial DNA is introduced into organisms, the organisms invariably die.

But Dr. Kool is confident that he will achieve replication, at least in the test tube. "In 5 to 10 years, we'll have an alien replicating system," he said. Dr. Romesburg of Scripps said he had achieved test-tube replication of DNA containing one extra base that pairs with itself.

Still, while they cannot yet be used in living cells, Dr. Benner's artificial bases are already being used in tests that read DNA sequences. The tests are sold by EraGen Biosciences of Madison, Wis., which calls the technology Aegis, for "an expanded genetic information system."

Besides answering questions about how life could have evolved elsewhere in space, the research might shed light on evolution on earth. Dr. Schimmel at Scripps said there might have been a stage in evolution when the genetic code was not as precise as it is now. His work, in which the protein proofreading enzyme was disabled, was an attempt to recreate that earlier, sloppier stage.

Dr. Schultz wants to subject bacteria with extra synthetic amino acids to stresses like heat or poison to see if they evolve and adapt faster than natural bacteria. "Will those forms of life with a bigger building-block set be superior to the ones who have 20?" he asked.

Dr. Schultz often says that living things have only 20 amino acids because God rested on the seventh day. "If he worked on Sunday," he said, "what would we look like?"