In the latest example of the wiliness of infectious bacteria, Tomoko Kubori and Jorge E. Galán have shown that Salmonella typhimurium exploits its host's protein-degradation machinery in order to reverse the substantial cellular reorganization that allows the bacterium to gain entry (Cell 115, 333–342; 2003).

A common cause of food poisoning in humans, S. typhimurium eases its passage into intestinal cells by injecting various proteins into them, including one called SopE. This protein causes rearrangements in the cell's actin filaments — part of the intracellular 'skeleton' — and ruffling of the cell membrane, events that enable the bacteria to gain entry. The left image here shows the rearranged actin filaments (red), along with internalized bacteria (blue). Shortly afterwards, another injected bacterial protein, SptP, counters these effects, restoring the status quo (right-hand image).

So how does S. typhimurium ensure that these events happen in the right order? One possibility is that it injects SopE first and SptP later. But Kubori and Galán rule out that idea: they show that SptP is present at the same levels as SopE within about 15 minutes of infection — just after the first evidence of actin rearrangement is seen. This also shows that SopE can produce these rearrangements even when SptP is about.

The authors find that levels of SopE subside later in the infection, when the actin reorganization ceases. But SptP is still easily detected. What causes the difference? Again, the authors eliminate the possibility that the proteins are produced to different extents in the infected cell. So they looked at whether differential degradation might be the key. They find that it is. The main cellular machinery for chewing up unwanted or damaged proteins is the proteasome; when a proteasome inhibitor is added to infected cells, the levels of SopE remain high.

Why is SopE destroyed so quickly after infection whereas SptP is not? Kubori and Galán find that this is not due to the presence of other bacterial factors, so it must be determined by something intrinsic to the proteins themselves. Previous work led the authors to suspect that a particular region of each protein, known as the secretion and translocation domain, controls their degradation. To test this, they swapped these portions of the proteins. The result was that SopE was treated like SptP, and vice versa — the actin rearrangements were not reversed in either case, because SptP was degraded quickly, and SopE was not. So what's the difference between these regions of each protein? That's a question for the future.