By David N. Leff

A simple soil bacterium, Streptomyces erythraeus, makes its living by eating rotting plant and animal biomass.

Not so simple is what S. erythraeus gives back in return for these nutrients. The bug synthesizes one of the most complex molecules known to the pharmaceutical industry, namely, erythromycin.

This means it puts together 118 atoms of carbon, hydrogen, nitrogen and oxygen, in 30 enzymatic steps, to synthesize the antibiotic. Biochemists describe erythromycin like this: C₃₇H₆₇NO₁₃.

It took the legendary Harvard biochemist, Robert Woodward (Nobel Prize, 1965), and his students literally years to duplicate the bacterium's feat of synthesizing erythromycin. They completed the task in 1981.

"Erythromycin has limitations from a therapeutic point of view," observed biochemist Chaitan Khosla, of Stanford University, in Palo Alto, Calif. "These shortcomings include a short half-life, acid lability in the stomach and increasing drug resistance by pathogens. Yet improving the antibiotic is not attractive for medicinal chemists," he added, "because of its too-complicated structure."

Khosla told BioWorld Today he has invented, and Stanford has patented, "general technologies that can be used to modify the structure of a natural product, such as erythromycin, by manipulating its genes. The idea," he explained, "is analogous to protein engineering at the genetic level."

Erythromycin is a polyketide, and polyketides are the basis of Khosla's new technology.

Meeting The Heavy-Hitting Polyketides

"They comprise a large and diverse family of natural products," he pointed out, "all structurally very complex. A few thousand polyketides have been discovered to date from soil microorganisms, and what's particularly interesting about them is that they've been an extremely, unusually rich source of drug-like molecules.

"I am told that the current market for polyketides," Khosla observed in passing, "exceeds $5 billion a year. They include major heavy-hitter drugs, not just erythromycin but other major antibiotics, immunosuppressives, anti-inflammatories and veterinary chemicals. The list goes on and on."

The premise underlying his new technology, Khosla said, "was that here Nature has developed very elegant methods to make a lot of these molecules, with a lot of interesting pharmaceutical properties. So," he continued, "if we could steal a chapter out of Nature's book, and manipulate their pathways, we could create our own private library of unnatural natural products — a very valuable resource to look for new drugs."

Khosla is senior author of a paper in today's Science, dated July 18, 1997, which bears the title: "Precursor-directed biosynthesis of erythromycin analogs by an engineered polyketide synthase."

To go S. erythraeus one better and create an improved erythromycin molecule, Khosla recounted, "We used genetic engineering to manipulate the genes that make the erythromycin pathway.

"For starters," he continued, "we had a bug that uses 30 genes to synthesize erythromycin in 30 steps. The bacterium basically plays a hand-me-down game, and takes simple chemical building blocks into the complex erythromycin molecule."

What Khosla and his co-authors did was put a block in the first step of that pathway. "It was a null mutant of the polyketide precursor gene," Khosla explained, "which we engineered by site-directed mutagenesis. This made a change in one of the key amino acids of that enzyme, and converted it into an inactive enzyme.

"So now," he went on, "we had a bug with 29 intact steps in the 30, but can't make erythromycin. Then we 'fed' this bacterium a completely synthetic precursor, which we dropped into its fermentation broth. This got it to go to step 2 or step 3 or step 4, depending upon the kinds of bells and whistles that we built into the molecule.

"It's basically a way to get something completely unnatural into the natural erythromycin pathway."

By now, Khosla's laboratory has constructed, he said, "hundreds of such modified erythromycins, all of them showing promise from a medicinal point of view. Besides being superior antibiotics, they have other biological properties that I can't talk about right now."

Neither can Khosla's commercial partner, Daniel Santi, with whom the Stanford scientist co-founded Kosan Biosciences Inc., of Burlingame, Calif., in January 1995.

Santi, the company's chairman, is professor of biochemistry and pharmacological chemistry at the University of California, San Francisco. He told BioWorld Today: "Chaitan is heavily committed to Kosan as a scientific consultant.

Manipulation Aimed At Making Better Drugs

"What the new Chaitan technology does," Santi continued, "is it enables you to go into any of these polyketide drugs, of which there are about a dozen super-important ones, and make changes that nobody's been able to make before.

"Minimally, what you would get," Santi observed, "would be equal to one of these very profitable derivatized drugs, and hopefully something that's better."

Santi and Khosla call their nature-mimicking structural engineering technology "combinatorial biosynthesis."

"What this involves," Santi explained, "is manipulating in a combinatorial fashion the genes of organisms that are responsible for making organic natural products, many of which are current pharmaceuticals. What we do is change the structure of the organic chemicals to produce new natural products.

"Nobody has been able to manipulate the genes that make small organic molecules," he pointed out. "They manipulate genes that make proteins. We manipulate genes that make proteins that make small organic molecules. So it's a totally unique technology."

The market for erythromycin and its current derivative analogues, Santi said, "is approximately two billion dollars."

"Chaitan's technology," he concluded, "provides us with ways now, for example, of going after microorganisms that are resistant to erythromycin." *