BioWorld International Correspondent

LONDON - Working out which proteins interact with each other - in normal cells, during infections, and during diseases such as cancer or obesity - will provide key information about which molecules are suitable targets for therapeutic agents.

A French biotechnology company now has developed a fast method for screening proteins for their ability to interact with other proteins. The company predicts that its technique will greatly speed up the process of allocating functions to the many unknown proteins being revealed by genome sequencing work.

Pierre Legrain, vice president of Hybrigenics SA, of Paris, told BioWorld International: "For the past 40 years, the challenge for pharmaceutical companies has been to find new targets at which therapies could be directed. Now, with the sequencing of the human genome, we know that the human genome encodes at least 100,000 different proteins, but we have no idea of the function of most of these proteins." Protein-protein interactions provide many clues to the roles these molecules play in the cell, he said.

Legrain and his colleagues at Hybrigenics, together with collaborators at the Institut Pasteur, also in Paris, describe their screening strategy in a paper in the Jan. 11, 2001, Nature titled, "The protein-protein interaction map of Helicobacter pylori."

Identifying which proteins interact with which is a first step, the researchers said, toward understanding their functions and thus eventually developing therapies that target them. Just as importantly, Legrain predicted, will be the development of a new generation of therapies that target the interactions themselves.

"If you know, for example, that protein A and protein B interact to have an effect on cardiovascular metabolism, and that protein A is present in many other cells but that protein B is present only in certain tissues, then targeting the interaction between A and B could allow you to have a very specific effect with your therapy, and one that has few side effects," he said. "In addition, you might be able to modulate the interaction, either making it stronger if it is too weak, or, if it is too strong, just destabilizing it a little."

The French team decided to evaluate its rapid screening technique using the genome of the bacterium Helicobacter pylori. But the researchers are now applying it to many different organisms, including humans.

The technique involves genetically manipulating yeast cells to include separate fragments of the genome of the organism under study. Yeast cells containing different such fragments are then mated (so that haploid cells become diploid). The cells resulting from this process are manipulated in such a way that they will only grow if the fragments of proteins contained within them interact with each other. The automated process of detecting interactions is weighted in order to reduce false negative and false positive results.

Legrain added, "The risk of false negative results is also greatly reduced because we are using fragments of proteins rather than whole proteins."

For the study of H. pylori, which encodes 1,600 proteins altogether, the researchers tested 261 different proteins encoded in the genome of this bacterium against a library of fragments of the entire genome. They discovered 1,200 interactions, connecting about half of the proteins encoded by the genome.

"Almost all the interactions we discovered are new interactions, but then very little is known about H. pylori," Legrain said. "We also compared our results to what is already known about Escherichia coli, which has been studied much more extensively. We looked at databases and tried to find homology between proteins of E. coli and H. pylori, and tried to find out whether what we observed in H. pylori had already been described for the homologous protein in E. coli." In many cases, it had, the researchers discovered.

Legrain added that most of the information on E. coli had been obtained by very time-consuming experiments. "It is a proof of our concept," he said, "that our high-throughput technology could reveal similar types of interaction in H. pylori, compared to what it has taken many laboratories almost 20 years to discover about E. coli."

The Hybrigenics team and their collaborators already have already set up a similar large-scale screening program for human proteins. It will include examination of protein-protein interactions in human cancer cells, as well as studies of adipocytes and cells of the central nervous system. Other programs already are under way for hepatitis C virus, the human immunodeficiency virus, E. coli and yeast itself.

Future work at Hybrigenics will include carrying out studies to validate biologically the interactions identified. "We have done this already for H. pylori," Legrain said, "and for some of the protein interactions observed we have good evidence that they are a biological feature. It is possible to use fragments of the proteins that have been identified as responsible for the interaction as modulators of these interactions." The company also is contemplating the development of a library of compounds to look for modulators of interactions, he added.