Dutch scientists used a computer-controlled artificial gut to show that DNA remains intact for several minutes in the large intestine.

Hub Noteborn, of the State Institute for Quality Control of Agricultural Products in Wageningen, said in New Scientist magazine: "It was a surprise to see that DNA persisted so long in the colon."

DNA survived in the artificial gut

This persistence means it might be possible for genes to jump from genetically-modified (GM) food into bacteria in the gut of farm animals. Previously some scientists have said there was no risk as the modified DNA breaks down too quickly.

If the transferred genes were the antibiotic-resistance genes used as markers in some GM crops fed to livestock, then serious health risks might result. The danger would be that antibiotic-resistant microbes would spread from animals to humans.

The experiments used bacteria genetically modified to contain antibiotic-resistant genes. As these were digested in the artificial gut, half the DNA survived for six minutes.

"This makes it available to transform cells," said Robert Havenaar, of the TNO Nutrition and Food Research Institute in Zeist. His team designed the artificial gut which provides a mechanical model of the stomach and intestines and contains the normal microbes and enzymes.

Greenpeace protest at GM maize in France in October 1998

For the first time the experiments measured the rate of transfer of DNA from GM bacteria to other bacteria normally found in the gut. Only one in 10 million passed on the DNA. But there are usually around a thousand billion gut bacteria, suggesting many would be transformed.

However, the Flavr Savr tomato, engineered by Calgene to resist rot, did not pass on antibiotic resistant genes, although up to 10% of its DNA reached the colon.

The crucial test will be whether GM foods and bacteria which have infected the gut can transfer genes. Havenaar says he will ask the European Union to fund further research.

The new findings show more research is essential, says Derek Burke, former chair of Britain's Advisory Committee on Novel Foods and Processes. "We can only say that the risk is not zero," he says. "Anything that would help put numbers on this, would be useful."