By Dean A. Haycock

Special to BioWorld Today

Some of the best chemists on the planet have no graduate degrees. They are also very small. The bacterium Saccharopolyspora erythraea, for example, has no trouble synthesizing the large and complex antibiotic erythromycin. Human chemists, by contrast, face a monumental task when they attempt to synthesize such complex molecules.

Instead, over the years, scientists have slowly collected and nurtured mutant S. erythraea that produce erythromycin. Drug developers particularly are interested in the large cyclic core of the antibiotic, 6-deoxyerythronolide B (6dEB). This natural product provides the core structure of lovastatin and other polyketides as well as erythromycin.

"About $15 billion a year worth of polyketides are sold," chemists Chaitan Khosla of Stanford University told BioWorld Today. "There are probably a 100-fold more molecules that have been very interesting to the pharmaceutical and chemical industry but have not been developed simply because nobody knows how to manipulate these molecules."

Efforts to induce bacteria that normally manufacture polyketides to increase their output and to produce modified versions of the natural product have been disappointing. The Actinomyces family of bacteria, which produce polyketides, doesn't perform very well in the fermentation vats biotech companies use to grow and harvest natural products. They grow slowly and they are not easily genetically engineered to produce polyketides. If a more cooperative and genomically friendly bacterium such as E. coli could be persuaded to synthesize polyketides, the challenges faced by medicinal chemists would be eased.

That is what Khosla and his colleagues set out to do. "The idea was if you could transport the biosynthetic machinery from esoteric organisms into E. coli, you would be able to use E. coli as a host to manipulate these pathways. That would make things much more efficient and faster," Khosla recalled.

A 'To-Do' List of Genetic Manipulation

The researchers knew that using genetic engineering approaches to do chemistry would not be easy in this case. "We listened to the experts in the field who told us all the reasons why it couldn't be done," Khosla said. "We basically made a 'to-do' list from that."

Among the items on the to-do list were: 1) Expressing a polyketide synthase (PKS), an enzyme that synthesizes the polyketide core, in E. coli. The PKS they inserted, called DEBS, consists of three large proteins. These had to be correctly joined, folded and modified after synthesis in the new host. 2) Providing E. coli with the building blocks for polyketide synthesis, a task that required additional re-engineering of the bacterium. 3) Synchronizing the production of precursors with the activity of the enzyme so the "timing" and availability of "parts" would be optimized.

It took four or five years to check off all the items on the list. But finally the chemists/genetic engineers succeeded in showing that their heavily modified E. coli did synthesize erythromycin's polyketide core. And it did so "with a specific productivity that compares well with a high-producing mutant of S. erythraea that has been incrementally enhanced over decades for the industrial production of erythromycin."

In the past, researchers have routinely inserted the equivalent of tools or manufacturing machines into E. coli. The authors succeeded in inserting something closer to an entire factory. An account of their success is described in a report in the March 2, 2001, issue of Science, "Biosynthesis of Complex Polyketides in a Metabolically Engineered Strain of E. coli."

E. coli Gets An 'A'

The chemists were surprised at how well the altered E. coli performed.

"Frankly, I could not have predicted it when we started this work, but E. coli is not just an OK catalyst to be able to do this kind of biochemistry. It seems to be a very good catalyst," Khosla said. "From an analytic chemistry perspective, you would have to give E. coli a grade of 'A' for its ability to do something which it has absolutely no past history of doing."

Khosla acknowledged that the results represent a significant step forward, but added, "Fundamentally, making erythromycin in E. coli is a natural extension from making insulin in E. coli.

"Doing this kind of chemistry in E. coli simplifies the discovery end for new polyketides. By manipulating the materials - cheaper, faster, better - one might be able to make the accessibility of new clinical candidates more affordable," Khosla told BioWorld Today.

Originally, the chemists wanted to set up E. coli as a discovery platform. "But it turned out that the catalytic properties were so good that one can realistically contemplate E. coli as a manufacturing platform," Khosla said. "There are lots of molecules out there that are commercially produced that could be produced far cheaper in E. coli given all the expertise that has been developed with regard to E. coli fermentation."

If Khosla and his co-workers can further manipulate E. coli to complete the synthesis of erythromycin and other drugs from the polyketide core, their work could make the development of agrochemical and natural polyketide-based products far less troublesome to pursue.

Khosla is the co-founder of Kosan Biosciences in Hayward, Calif. Several years ago, Kosan received an exclusive license from Stanford University for a "fairly general patent in the area of producing polyketides" that originated from work in Khosla's laboratory. The research described in the current Science paper, however, was completely supported by grants from the National Science Foundation and the National Institutes of Health. n