The E. coli outbreak in Germany is enormous. In case numbers (so far) it falls short of the 1996 outbreak in Sakai City in Japan, but the number of those in Germany going on to develop haemolytic uraemic syndrome (HUS), the main complication, which affects blood cells and kidney function, is far greater than in any previous outbreak – 520 on 2 June – and the proportion of those infected that have gone on to develop HUS is also much greater. Germany usually sees about 65 HUS cases every year. In Sakai City only 106 out of 2764 microbiologically confirmed cases developed HUS. The number of deaths in Germany already exceeds the 17 in central Scotland in 1996.

The German problem is being caused by E. coli type O104:H4, which hasn’t been recorded as causing an outbreak before. It is clearly a very hot strain. At least some of its virulence comes from Shiga toxin 2 (Stx2), which is released by the bacterium and sticks to the surface of cells in various organs, including the glomeruli and tubules in the kidney. The cells take up the toxin, which turns off their ability to make proteins, and the cells commit suicide.

Where this strain has come from is currently a mystery. A comparison of victims and well people in Hamburg indicated that the former ate tomatoes, cucumbers and lettuce much more than the latter. But direct attempts to identify the contaminated food have failed. An E. coli strain capable of causing gastroenteritis and HUS was found last week on Spanish organic cucumbers, and Hamburg’s health minister, Cornelia Prüfer-Storcks, went public with this information. At the time it was a reasonable guess that this would turn out to be the outbreak strain. But further work showed that it was not. This didn’t give the cucumbers an absolutely clean bill of health, but it wasn’t the smoking gun. Whether the cause of the outbreak will be established is far from certain. It has been in progress since the beginning of May. Finding contaminated food will depend on whether any remains uneaten and is available for analysis. In many food poisoning outbreaks this doesn’t happen.

Up to now it has been reasonable to assume that the E. coli came from the rectums of cattle or sheep, the natural home of Stx2 carrying strains, and that manure containing it got onto food. But the genome sequence of the strain shows that it is related to another diarrhoea-causing E. coli, a strain from Africa (enteroaggregative E.coli, or EAEC), which only occurs in humans. The sequencing was done in three days by BGI (formerly the Beijing Genomics Institute) in China.

That a nasty new E. coli strain has appeared isn’t surprising. E. coli O157 only showed itself in the early 1980s. In evolutionary terms it is brand new. And E. coli evolves in real time, by taking up DNA from its brothers and sisters as they huddle together in the bowels of their hosts. We may never know where O104 has come from, though we will probably find out quite soon where it lives in nature. Sampling cowpats is straightforward, at least. As for why it’s so nasty, molecular biology will help to find out, but it may take a while: we’ve been studying O157 for a quarter of a century without a full answer. And there’s no knowing whether O104 will stay around for years like O157 or fade away like SARS. The only certainty is that evolution will continue to take us by surprise.