There are thousands of genes in the human genome, and many of them have no cDNA available, Dan Santi, president and CEO of Kosan Biosciences Inc., told BioWorld Today.

The ability to chemically synthesize those genes from their sequence data, which is of course available courtesy of the Human Genome Project, has multiple applications in academic as well as industrial research. There also are endless genes in other species that are interesting, useful, or both. And finally, making artificial genes that can produce proteins not found in nature is yet another possible application of chemical gene synthesis.

Kosan scientists now have published a new procedure for such gene synthesis in the Nov. 2, 2004, issue of the Proceedings of the National Academy of Sciences. In the paper, Kosan scientists describe the method and demonstrate proof of principle by making a polyketide synthase gene cluster, which was successfully assembled and, in turn, made its product in Escherichia coli. Measuring at more than 31,000 base pairs, that gene cluster is more than four times as long as the longest functional man-made DNA to date, the poliovirus genome.

Cut Out The Middleman? No, Quite The Contrary

Genes, and even whole genomes, have been made from scratch before. In fact, the first chemically synthesized genes were reported in peer-reviewed literature in the 1970s. In 2002, Eckard Wimmer and his group at SUNY Stony Brook reported in Science that they had assembled the full 7,000+ base-pair sequence of the poliovirus genome in vitro, and showed that viruses built from the artificially constructed sequence were infective. And in December, Craig Venter and his colleagues at the Institute for Biological Energy Alternatives reported in PNAS that they had made an entire functioning viral genome of a bacteriophage, a virus that infects bacteria.

Asked to compare Kosan's current paper to the bacteriophage genome synthesis, Santi said that the latter essentially used a "one-pot synthesis." Venter's group took short, commercially available oligonucleotides of about 40 base pairs and directly strung together the 5,400-base pair sequence of the bacteriophage.

"That was a terrific accomplishment," Santi said, "and it's a great technology if you have a biological system to do the purification." In the case of the bacteriophage, purification was comparatively easy. While most of the bacteriophages were "riddled with errors," only the few without errors would be able to infect bacteria.

However, many genes of interest would not have an obvious biological system to select for error-free copies. Because of the lower error rate that is achieved by the synthon intermediates, Kosan's method of gene synthesis can make long DNA sequences that do not require such biological selection.

The strength of the approach of Kosan's scientists is their focus on creating a rapid and automated process for synthesizing genes that does not require biological purification. As chemically sequenced DNA gets longer, its error rate increases near-exponentially, which means that for every bit of genetic gold, there is a lot of dross created as well. So Kosan focused on "making smaller, perfect pieces of DNA, and using novel technology for efficiently putting them together in an automated, parallel-processing fashion."

The scientists created 500 to 800 base-pair intermediates prior to assembling the finished gene. Those intermediates, called synthons, were assembled from commercially available short oligonucleotides of about 40 base pairs in length. The synthon length was empirically determined to be short enough to give a low error frequency.

The next step was to combine synthons into longer DNA sequences. With typical methods, that often is done sequentially, and involves "a lot of manual work. It's a pretty tedious process," Santi said.

Kosan scientists avoided that hard labor by combining the synthons in a procedure known as ligation by selection. In that procedure, each synthon is paired with two marker genes conferring antibiotic resistance. Plasmids containing the two synthons are allowed to combine freely, giving four possible combinations of synthons. However, the correct combination of synthons in the correct order has a unique set of antibiotic-resistance markers, allowing scientists to zap the other three combinations. That process can be repeated over as many cycles as desired, doubling the length of the gene with each cycle.

As for the underlying technology, Kosan has patented much of it. Santi said that "we would certainly license it to interested parties - we are freely sharing it with university people right now."

When asked whether the work represents a principally new ability to manipulate genes, Santi demurred. He noted that "people have been making genes for a very long time. This work increases the scope of what can be done to more genes, and to longer and longer genes.

"It could be done before," he said. "It was just much more difficult."