Genetics is revolutionizing industrial processes for making everything from antibiotics to plastics. Now, as the Fifteenth Enzyme Engineering Conference heard in Kona, Hawaii, last week, researchers are creating new combinations of genetic sequences and mining them to produce biological catalysts that work more efficiently, more cleanly, and perform tasks never conceived by nature.

Biological catalysts are called 'enzymes'. Enzymes make chemical reactions proceed millions of times more rapidly than if the reacting molecules simply collide. They grab specific molecules and hold them together so they can react to form a new product. For decades, biochemists have used enzymes to produce compounds for agriculture, pharmaceuticals and even washing powders.

Until fairly recently, however, these enzymes were identified by laboriously sifting through protein extracts of organisms known to produce useful compounds. Now scientists are using the latest 'gene splicing' techniques to improve upon nature's repertoire of enzymes.

One approach adopted by Diversa, a California biotech company, is to harvest genes encoding useful enzymes from the environment. Diversa is focusing on microbes – an estimated 99% of which are undiscovered – and is searching for these in deep-sea vents, Arctic ice, translucent stones, and other unusual and extreme environments. Eric Mathur, Diversa's Senior Director of Molecular Diversity, reckons that many of these unknown microorganisms will contain new and useful enzymes.

Diversa's technique takes DNA directly from these bacteria, bypassing the usual step where microbes are cultivated in the laboratory – a step that is costly in terms of genetic diversity. But first they separate the microbes from their host organisms – often exotica like Komodo dragon skin or Costa Rican spider gut.

Next they extract DNA from the sample, cut it up into pieces, and transfer the bits to the commonly used laboratory strain of bacteria, Escherichia coli. In this way, a 'library' is produced containing millions of E. coli cells, each producing different proteins encoded by different bits of the harvested genome. The cells are then screened to find the ones that refer to interesting enzymes. This approach has been highly successful, so far yielding over 1,000 unique bioactive molecules.

But that's just the beginning. The company now plans to improve the enzymes by a process called 'directed evolution'. This technology uses the natural evolutionary process – by which organisms acquire random mutations in their genes, some of which impart an advantage and propagate through successive generations – to improve or change enzymes in a test tube.

Indeed, by introducing mutations into a gene in repeated cycles, re-combining the gene with an un-mutated form to eliminate detrimental mutations, and then screening the libraries of evolved genes for the activity desired, Pim Stemmer, working at another Californian biotech company, Maxygen, has succeeded in making an enzyme 32,000 times more powerful than the original.

Both Maxygen and Diversa have filed patents for this gene shuffling technology, and it has already been used to improve laundry detergents, anti-HIV therapies and industrial catalysts. The procedure now has many variations, including 'family shuffling' where genes from different species are shuffled.