Last month, the president of fast-food chain Taco Bell appealed to his customers in full-page ads in The New York Times and other US newspapers. “You can be confident our food is safe to eat,” his letter declared.

Green menace? E. coli O157 can spread if greens such as spinach are irrigated with tainted water. Credit: J. ERNST/REUTERS

The ads were the fallout from a food-poisoning outbreak traced to the chain's restaurants, which has affected at least 70 people across five states. The culprit was Escherichia coli O157:H7 – the same as in another outbreak this September and October linked to Californian spinach, which infected nearly 200 across the country and killed three.

The outbreaks have thrown the spotlight on a bacterium that is difficult to detect and virtually impossible to treat or eradicate. “We see it more and more and we don't really know what to do about it,” says microbiologist John Fairbrother of the University of Montreal, Canada.

There are thousands of different strains of E. coli, most of which are harmless. But O157 can make a potent toxin and latch onto intestinal cells, giving it the ability to cause kidney failure and even death. The bugs live harmlessly in cows' large intestine and are thought to be ubiquitous in cattle lots. Bacteria shed in faeces contaminate meat in slaughterhouses or find their way onto vegetables grown near animals or irrigated with water contaminated with manure, as is thought to have happened with the tainted spinach. Lettuce is thought to have spread the Taco Bell outbreak.

There are now some promising research leads that might help prevent future outbreaks. At a meeting earlier this year on pathogenic E. coli, veterinary researcher David Smith of the University of Nebraska, Lincoln, and his colleagues reported that a vaccine containing proteins from O157 cut the number of cows shedding bacteria by 60–70%. Canadian company Bioniche Life Sciences, based in Belleville, Ontario, has submitted the vaccine for regulatory approval in Canada, and plans to do so in the United States.

Other groups are turning to viruses, called bacteriophages, that attack the O157 strain. A group led by microbiologist Todd Callaway of the US Department of Agriculture's Food and Feed Safety Research Unit in College Station, Texas, has found that feeding sheep a mixture of bacteriophages cuts the number of pathogenic bacteria in their guts by over 1,000 times.

Cattle farmers may be forced to adopt vaccines or therapies because of pressure from food processors and the threat of lawsuits. But some microbiologists question whether these reductions will protect humans. Just ten cells of O157 are enough to infect a person, compared with hundreds of thousands needed for a Salmonella or cholera infection. “It's a very different standard for foods to meet,” says James Kaper, an E. coli expert at the University of Maryland, Baltimore. He notes that irradiation would rid food of the bacteria but that the public, food industry and food-safety regulators have been reluctant to adopt it.

So researchers are also working on treatments, for example with antibodies that inactivate the toxin. (Antibiotics aren't recommended for E. coli because by the time the infection is diagnosed, the bugs have usually released so much toxin that killing them doesn't help.) But researchers admit that the demand for such drugs is likely to be low.

Perhaps an underestimated problem are other pathogenic E. coli strains, including O26, O111 and O145. These can also cause serious food poisoning but are more likely to go unnoticed, because lab tests are more difficult or not routine. These strains are more common outside the United States; in Italy, for example, most cases of E. coli food poisoning probably go undetected, says Alfredo Caprioli, who directs the E. coli reference lab at the Istituto Superiore di Sanità in Rome. There is intense interest in finding the exact combination of genes that make strains harmful to humans and quick ways to test for them.

Ultimately researchers must find the critical points in the food supply at which intervention can most reduce contamination, says food scientist Don Schaffner of Rutgers University, New Jersey. “Obviously we haven't studied it enough to solve the problem.”