Antibiotics already are ineffective against E. coli infection. In fact, they make things worse.

"Normally, antibiotics cause a more severe form of the disease through up-regulation of [shiga] toxin release," Adam Linstedt told BioWorld Today. So antibiotics are actually contraindicated during E. coli infections. And though E. coli's death toll is comparatively small in developed countries, numbering in the dozens, the bug kills more than a million people annually in the developing world, mostly through contaminated drinking water.

In work published in the Jan. 20, 2012, issue of Science, Linstedt and first author Somshuvra Mukhopadhyay, who are at Carnegie Mellon University, showed they could protect cells from the shiga toxin that makes some strains of E. coli so deadly by using a simple chemical element: manganese.

Shiga toxin is made by a phage, or virus, and released by E. coli. Once the bacteria release the toxin in the gut, it attaches to a receptor on human cells in the gut epithelium. From there, it gets transported into the host cell's protein processing network, or Golgi network, the endoplasmic reticulum, and finally the cytoplasm, where it proceeds to kill the host cell.

But that trip itself is a highly unusual one.

"The toxin comes into the cells like many things," Linstedt said. "But most things that come in go to the lysosome and get degraded." The lysosome is a cellular defense mechanism against harmful substances that manage to get across the membrane.

Shiga toxin, however, manages to escape that fate by binding to an intracellular protein, GPP130. That protein itself cycles between other parts of the cell, but avoids the lysosome.

The finding that GPP130 is what keeps shiga toxin out of harm's way in the lysosome "was a pretty serendipitous discovery to begin with," based on a toxicologist colleague's observation that GPP130 binds to manganese, Linstedt said.

In their studies, Linstedt and Mukhopadhyay showed they could interrupt the association between shiga toxin and GPP130 by treating cells, or mice, with manganese.

In both cell culture and in vivo experiments, Mukhopadhyay and Linstedt showed the binding rerouted shiga toxin into the lysosome. The lysosome, in turn, did what the lysosome does best: It chewed up the toxin. Animals treated with manganese prior to being exposed to shiga toxin were protected from what normally would be a lethal dose.

Manganese is cheap and stable, and so it could be a realistic treatment for people in the developing world who need it most. Mukhopadhyay and Linstedt intend both to follow up on the practical implications of their findings and to look at their biological underpinnings in greater detail.

For one thing, they want to make their animal model more realistic.

"In our mouse model, we used pure toxin, which is very lethal, but is not the way that humans get infected," Linstedt said.

He and Mukhopadhyay plan to follow up by testing whether manganese also protects mice infected with shiga-producing E. coli.

They also plan to test whether mice whose shiga toxin is being neutralized could benefit from antibiotic treatment, instead of getting sicker from it.

Another plan is to optimize dosing, and to test how late and how short treatment can be and still have an effect.

In the experiments now published in Science, the team treated mice up to five days before exposing them to the poison. And although there is a time window between the appearance of symptoms and the development of life-threatening symptoms, in real life, that window is not five days. But Linstedt said that "the biology suggests that we can go shorter . . . perhaps to as little as 12 hours."

To better understand the basic cell biology of the effect they have discovered, they also plan to study GPP130 itself in greater detail.

"We don't know the actual function of the protein," Linstedt said, though given how tightly it binds manganese, one possibility is that it regulates its levels within the cell – which need to be kept in a certain window because the manganese is "essential, but too much of it is toxic."

Regardless of what the protein does, however, it is apparently dispensable for a short time period.

"We don't definitively know," Linstedt said. But in his and Mukhopadhyay's experiments, "cells that are missing this protein for hours to days don't seem to have big problems."

A key question about the protein's biological role is just how GPP130 avoids the lysosome, cycling between the endosome and the Golgi network instead.

"That's the real key at the molecular level," Linstedt said. "How does GPP130 get into the endosomal tubules?"