Bernhard Hauer, BASF's vice president for research, says BASF will evaluate "the whole variety of Diversa's enzymes for our processes and for opportunities to make products." Specifically, BASF is interested in enzymes for use in producing high-value chemicals such as chiral intermediates, amino acids, and vitamins. BASF already has R&D in these areas, he says, but Diversa's collection could turn up better enzymes and save BASF development time and effort.

Degussa also is interested in Diversa's enzymes for applications in specialty chemicals. "We don't want to reinvent the wheel," says Stefan Buchholz, head of Project House Biotechnology, Degussa's center for biocatalysis R&D. "Whenever enzymes are available from major enzyme producers or enzyme discovery companies like Diversa, we will apply them to speed our process development," he tells C&EN. Most likely, Degussa's fine and industrial chemicals division, especially the custom manufacturing business, will profit most from shorter development times, he says.

For Givaudan, the agreement with Diversa is part of the company's strategy of forming alliances to access noncore technologies, according to Robert Eilerman, who is Givaudan's senior vice president for R&D. He says Givaudan is looking for biocatalysts that will help reduce the costs of producing existing flavor compounds, as well as others that will allow access to new flavor components.

"We're finally developing a reputation within the chemical industry that we didn't have to a broad extent in the past," Mark J. Burk, Diversa's vice president for chemical product development, tells C&EN. The chemical industry has taken a while to understand "where biotechnology can have an impact" on its activities, Burk notes. Part of the reason on Diversa's side, he says, is that "we weren't able to speak publicly about partner programs. It was difficult to make people aware of our capabilities without concrete examples."

The situation changed in early 2000 when Diversa went public and raised about $200 million. The cash allowed the company to fund its own development program. The company can freely speak about capabilities coming out of this program, as did Burk in Philadelphia last November at ChiraSource 2001, a conference organized by Catalyst Group Resources, Spring House, Pa.

The talk that he gave, Burk says, "was one of the first times that we've really laid out publicly a platform that we're developing internally." The talk also was an opportunity to explain how Diversa's technology has overcome the limitations that have been associated with biocatalysis over the years.

THROUGH WHOLESALE extraction of environmental DNA, Diversa discovers enzymes from organisms that cannot be cultured in the lab, which comprise about 99.9% of the microorganisms in many environments. And by taking samples from places as different as Yellowstone National Park, Iceland, and Kenya, Diversa has global access to microbial diversity.

The DNA is chopped, gene fragments are put into vectors, and the vectors are inserted into easy-to-culture microorganisms such as Escherichia coli and engineered strains of Streptomyces and Bacillus that express the proteins encoded by the genes. Proteins with enzymatic activity are rapidly identified by various screens. Promising enzymes are further developed by directed evolution through single-site mutations or gene reassembly.

This approach to discovery has yielded more than 2 million unique genomes comprising billions of genes in Diversa's environmental libraries. Diversa "maintains large gene libraries that represent significant biodiversity," Givaudan's Eilerman says. "This, together with its high-throughput-screening capability, makes the firm an ideal partner for identifying the new enzymes we require."

In addition to discovery, evolution, and screening, Diversa also engineers hosts for efficient overexpression of enzymes. Efficient hosts bring down the cost of enzymes, Burk says, noting that the expense of enzymes has restricted use of biocatalysts in the past.

The limited number of commercial enzymes, narrow range of useful transformations that enzymes mediate, lack of reaction generality, enzyme instability, and long development times are other drawbacks that have been associated with biocatalysts. Diversa's combination of discovery and evolution has overcome these limitations, Burk says.

Consider, for example, the nitrilases. Diversa started a program to discover enzymes in this class two years ago, when, according to Burk, only 14 such enzymes were known. Since then, Diversa has discovered about 200 novel and unique nitrilases, he says. And of these, 175 are in bottles, available for screening. That rate of discovery would have been impossible with traditional screening of whole organisms. With so many more enzymes to work with, the range of transformations that are amenable to biocatalysis begins to open up.

"EVEN IF NATURE does not get you to an enzyme with the exact properties you require, you can evolve what you've discovered to your desired end point," Burk says. Evolution can yield a range of enzymes to handle a range of substrates and optimize an enzyme to the required process parameters. It may even enable access to both product enantiomers instead of just one. "Having enzymes available and evolution capabilities facilitates development considerably," Burk says.

At ChiraSource 2001, Burk described what Diversa has achieved with nitrilases. To date, Diversa has nitrilases useful for enantioselective one-carbon homologation of aldehydes, preparation of (R)-mandelic acid derivatives at enantiomeric excess (ee) of up to 99%, and preparation of (S)-phenyllactic acid at 99% ee. "Our work shows how effective nitrilases are in producing enantiomerically pure intermediates," he tells C&EN.

Burk's favorite example is the enantioselective preparation of ethyl (R)-3-hydroxy-4-cyanobutyrate. This b-hydroxy ester is an intermediate in the synthesis of the cholesterol-lowering drug Lipitor (atorvastatin calcium). "People have looked at this conversion ever since Lipitor came out 10 years ago. All the enzymes previously discovered provide the S enantiomer," Burk says. "We have four enzymes that give the R enantiomer, and we're getting between 95 and 99% ee."

A published route to the Lipitor intermediate requires six steps, starting from 2,3-dihydroxychloropropane. Using nitrilases and starting with epichlorohydrin, Diversa proposes a different route that takes only three steps. First, epichlorohydrin is chemically converted to 3-hydroxyglutaronitrile. Next, the symmetric prochiral product is desymmetrized by the action of a nitrilase to an (R)-b-hydroxy acid. And finally, chemical esterification yields the required R intermediate. Thus far, Diversa has performed the nitrilase step on the half-gram scale. "We're scaling to 10 g right now," Burk tells C&EN.

With clear-cut examples demonstrating how Diversa's technology can be used, Burk hopes to raise the awareness and comfort level of chemists with regard to biocatalysis. As one with hard-core organic chemistry credentials, Burk says he himself became comfortable with biocatalysis after seeing it perform, particularly in situations where chemistry struggles: "It's very compelling once you see it work."

Chemical & Engineering News
Copyright © 2002 American Chemical Society