A banana a day may not keep the doctor away, but a substance originally found in bananas and carefully edited by scientists could someday fight off a wide range of viruses, including those that cause AIDS, hepatitis C, and even flu, an international team of researchers said Thursday.

The new research, published in the US journal Cell, focused on a protein called banana lectin, or BanLec, that "reads" the sugars on the outside of both viruses and cells.

Five years ago, scientists showed the protein can inhibit HIV infection by binding to sugars on the surface of the AIDS virus, and blocking its entry to the body, but it also caused side effects that limited its potential use.

Now, researchers from Germany, Ireland, Canada, Belgium and the United States reported how they created a new form of BanLec that still fights viruses in mice, but doesn't have a property that causes irritation and unwanted inflammation.

They succeeded in peeling apart these two functions by carefully studying the molecule in many ways, and pinpointing the tiny part that triggered side effects.

Then, the team engineered a new version of BanLec, called H84T, by slightly changing the gene that acts as the instruction manual for building it.

The result is a form of BanLec that worked against the viruses that cause AIDS, hepatitis C and influenza in tests in tissue and blood samples, without causing inflammation.

The researchers also showed that H84T BanLec protected mice from getting infected by flu virus.

"What we've done is exciting because there is potential for BanLec to develop into a broad spectrum antiviral agent, something that is not clinically available to physicians and patients right now," David Markovitz, co-senior author of the new paper, said in a statement.

"But it's also exciting to have created it by engineering a lectin molecule for the first time, by understanding and then targeting the structure," said Markovitz, a professor of internal medicine at the University of Michigan Medical School.

Several years of research were still needed before BanLec can be tested in humans, but the researchers hoped that that their team's work can help address the lack of antiviral drugs that work well against many viruses or against viruses that change rapidly, such as influenza.