An ingenious trick that should reveal the side effects of drugs long before trials begin has been developed by a company in California.
The trick is to attach proteins to their genes. This provides a quick way of revealing which of the body's proteins are likely to be caught in the chemical crossfire from a powerful new drug. "Small-molecule drugs are far from being magic bullets," says Bassil Dahiyat, the president of Xencor in Monrovia, California, which developed the technique. "You're basically eating poison."
Once human proteins have been tethered to the genes that made them, the protein-gene "combos" are exposed en masse to known and experimental drugs. Next, technicians take the proteins that bind to a particular drug and sequence the attached genes.
This is a much quicker and simpler way to identify the proteins than sequencing them directly. It also means that you know immediately which genes code for the affected proteins. Normally, researchers would have to isolate the proteins that bind to targets, identify them and then work backwards to find out which genes made them.
Xencor has already tested the method on oestrogen. As well as finding the hormone's receptor protein, as expected, they fished out a second target, which turned out to be a protein that plays a key role in cell division. This explains the long-standing mystery of why oestrogen affects cell division. "It's been doing this for years, and no one knew why," says Dahiyat.
The new technique allows millions of different proteins to be generated, each one tagged to the gene that made it. So far Xencor has only tried the technique with proteins produced by single human cells, but eventually the combos could allow potential drugs to be screened against all human proteins in one go. Another big plus is that the combos can be made in cultures of human cells. This means that their proteins are identical to those found in people-unlike human proteins made by bacteria.
The key to the process is the method, invented by Min Li at the Johns Hopkins School of Medicine in Baltimore, that links the protein to its gene. The first step is to make copies of all the active genes of a particular cell and then insert each copy into a loop of DNA called a plasmid.
The clever part is that the plasmid also contains a special gene that makes a protein "hook". This hook grabs hold of the gene that coded for it. So when the plasmid is put into a culture of human cells, it produces large quantities of the target protein attached to the hook. This links up with the gene, producing the protein-gene "combo" (see Diagram below).
"I think it could be a very important advance," says Raj Parekh, who pioneered protein identification technology at Oxford GlycoSciences in Abingdon. He cautions that proteins might not always adopt their natural shape if burdened with other enzymes and genes. "But if it pans out, it will be valuable," he says.
Precise details are being kept secret until patents are issued.
Author: Andy Coghlan
New Scientist issue: 11 November 2000
PLEASE MENTION NEW SCIENTIST AS THE SOURCE OF THIS STORY AND, IF PUBLISHING ONLINE, PLEASE CARRY A HYPERLINK TO: http://www.newscientist.com