Protease-linked AAA+ proteins ('ATPases associated with various cellular activities' proteins) unfold substrate proteins and deliver them to their associated proteases for degradation. For example, the bacterial hexameric ring-shaped AAA+ machine ClpX catalyses the denaturation and translocation of substrates into the protease ClpP in a process that requires hundreds of cycles of ATP hydrolysis. But how exactly do AAA+ machines work? Martin, Baker and Sauer provide new insights in Nature.

They covalently linked different combinations of active and inactive subunits of ClpX to form hexamers, and studied how these variants affected the ClpP-mediated degradation of a denatured substrate. Their first conclusion was that models in which all six ClpX subunits bind and hydrolyse ATP in a concerted fashion cannot be correct, because a ClpX variant with alternating active and inactive subunits had an activity that was roughly proportional to the number of active subunits.

They then looked at asymmetrical hexamer variants of ClpX that contained three, four or five active subunits positioned consecutively. As before, these variants had ATPase activities and produced protein-degradation activities that were proportional to the number of active subunits. A symmetrical arrangement of active subunits is therefore not required for ClpX function.

Significantly, a ClpX variant containing a single active subunit could also drive ClpP-mediated protein degradation. Conformational changes in just one ClpX subunit that are the result of ATP binding and hydrolysis therefore represent the basic power stroke of ClpX and are sufficient to drive ClpP-dependent protein degradation.

Next, the authors showed that the ClpX variants could power the ClpP-mediated degradation of native substrates, and that ATP hydrolysis in just a single active subunit of a ClpX hexamer could drive substrate unfolding as well as translocation.

To conclude, Martin and colleagues have shown that diverse geometric arrangements of ClpX subunits can support substrate unfolding and translocation into ClpP for degradation. It seems that the power stroke is generated by ATP hydrolysis in a single ClpX subunit, and the results obtained here rule out a concerted model and a strict sequential model for ATP hydrolysis by ClpX. Instead, these data indicate a probabilistic sequence of nucleotide hydrolysis. This mechanism would allow any ClpX subunit that contacts a translocating polypeptide to hydrolyse ATP to drive the substrate into ClpP. It would also prevent substrate stalling if a particular subunit was unsuccessful in binding or hydrolysing ATP. Such a probabilistic mechanism might be especially important for molecular machines that function on diverse substrates.