Genome Mining Hits Pay Dirt

CHEMICAL BIOLOGY

Using genome mining, researchers have predicted and then discovered a previously unknown natural product.

The process of discovering natural products from microbes is usually long and laborious. Genome mining--that is, searching a genome for DNA sequences that encode enzymes involved in the biosynthesis of particular products--could provide a significant shortcut, as demonstrated by Gregory L. Challis and coworkers in the chemistry department at the University of Warwick, Coventry, England. They have used genome mining to identify a previously unknown natural product from the soil-dwelling bacterium Streptomyces coelicolor.

In 2000, Challis' group identified a gene cluster that encodes a new nonribosomal peptide synthetase, which is a modular protein system that synthesizes peptides without ribosomes. They predicted that the cluster produces an unknown iron-chelating peptide, which they called coelichelin. Now, they've actually found the peptide and confirmed its structure (Nat. Chem. Biol., published online Sept. 11, dx.doi.org/ 10.1038/nchembio731).

"In our approach, we use analysis of DNA sequence data to predict structural elements of new natural products and then use this information to design strategies for rapidly identifying, purifying, and structurally characterizing the compounds," Challis says.

They predicted how many and which amino acids the product would contain. The information turned out to be crucial for identifying culture conditions in which the product could be produced by the organism and isolated. Surprisingly, even though the gene cluster encodes a nonribosomal peptide synthetase containing only three modules, which usually means that the product would be a tripeptide, coelichelin is in fact a tetrapeptide with one of the predicted amino acids incorporated twice.

"The fact that it is a tetrapeptide rather than a tripeptide is very important because it has potentially far-reaching implications for nonribosomal peptide synthetase multienzymes, which are involved in the biosynthesis of many important microbial natural products," Challis says. These findings suggest that these systems don't always work the way people assume they do.

Brian O. Bachmann, an assistant professor of chemistry at Vanderbilt University, says this work "takes a chunk out of the luck component of natural products discovery, which traditionally relies on screening a large number of extracts. The work shows that a little foreknowledge goes a long way. With a good idea of the natural product structures that a given biological source may contain, the task of isolating them and solving their final structure is greatly simplified."