GNN - Genome News Network  
  Home | About | Topics
DNA shuffle leads to better virus
By Bijal P. Trivedi

A company in California has harnessed a technique called DNA shuffling to breed viruses with desirable characteristics. Viruses that specifically infect only certain cells are of tremendous interest to gene therapy researchers who want to use them to deliver therapeutic genes to specific organs or tissues.

(A). The "parent viruses" all have different envelope proteins (B). The gene for the envelope protein is taken from each parent virus and (C) randomly cut into small pieces (D). The gene pieces reassemble creating chimeric genes, which are put back into a virus. The new envelope proteins are mixtures of parent proteins.

The researchers began with a family of six related murine leukemia viruses (MLV) that only infect mice, and they used bits and pieces of DNA from each MLV to create a new virus that could infect Chinese hamster ovary cells.

The ability of a virus to infect a specific cell type depends on the envelope proteins that cover its outer surface. These envelope proteins interact with receptors on the outer surface of mouse cells, which opens up a route of entry to infect the cell. The viral envelope proteins on the MLV cannot interact with receptors on a hamster ovary cell, which keeps the virus out.

In nature, viruses evolve rapidly, randomly changing their envelope proteins, which enables them to infect new organisms. Nai-Wei Soong, the scientist who led the research at Maxygen Inc. in Redwood City, California, reasoned that he could mimic evolution in test tube and create a virus that would specifically infect hamster ovary cells.

Soong took the envelope genes from the six families of MLV, cut them into pieces, and let them mix and randomly recombine. The result was a vast collection containing millions of chimeric envelope genes. Viruses containing these new genes were then tested to see whether they could infect hamster ovary cells.

The search revealed that one dominant virus containing envelope gene parts from three of the six "parental" strains of MLV could infect hamster ovary cells.

Soong emphasizes that this technique will only work with closely related viruses. "You can't take flu and HIV genes, chop them up, and hope they will recombine in any meaningful way," says Soong.

"I'm a big believer in the technology, and I think it's a very valid means of developing a specific virus. The only limitation is finding a way to select and isolate the one virus out of a million that has the characteristics you want," says Mark Kay, of Stanford University.

. . .

Soong, N.-W. et al. Molecular breeding of viruses. Nat Gen 25, 436-439 (August 2000).

Back to GNN Home Page