On page 1115 of this issue, Robert Sauer and his colleagues offer an insight into what makes molecular machines work. Their focus is an enzyme called ClpX, which unfolds proteins and prepares them for degradation in the cell.

The group's interest in protein degradation began with the question why do some proteins get degraded but not others? The key, it seems, is that for proteins to be unfolded by ClpX, they need to have recognition tags so that the enzyme can bind to them. ClpX also needs to hydrolyse ATP to drive the unfolding process. “ClpX is a molecular machine and ATP is the fuel that powers it,” Sauer says.

The key to understanding this molecular machine for Sauer and his colleagues Tania Baker and Andreas Martin involved taking apart this machine and observing how its individual components drove protein degradation.

But ClpX is not a simple enzyme — it is made up of six identical building blocks or subunits. Earlier experiments had revealed that ClpX was inactive if none of its subunits could hydrolyse ATP, but there was no simple way to see whether ClpX could unfold proteins if one or more of its subunits was inactive.

Martin reasoned that, to understand the unfolding process, he would need to manipulate the enzyme's subunits one at a time. This proved easier said than done. He first tried to connect ClpX's subunits by making genes that expressed enzymes in which individual subunits were connected by different peptide linker sequences. But nothing worked. “It was very discouraging,” Sauer says. “I was ready to quit, but Andreas persisted. He just refused to give up.”

Robert Sauer (right) with Tania Baker (left) and Andreas Martin.

After six months of frustration, Martin found that deleting a non-essential segment of ClpX allowed him to stitch the subunits together and generate an active enzyme. But this, too, proved to be a lengthy business. Martin began by deleting the non-essential domain and then linked two separate subunits together. He then did some careful chemical engineering to link in the other four subunits to make a six-sided molecule.

To test how this re-engineered machine worked with ATP, Martin then made mutant enzymes in which only some subunits could use ATP as a power source. He found that only one active ClpX subunit was needed to allow protein degradation.

This contradicted the two prevailing models, one in which all six subunits had to bind and hydrolyse ATP simultaneously to trigger degradation, and another in which the six subunits had to hydrolyse ATP in a strict sequence, like the sequential firing of pistons in a car engine. This means that the six building blocks of ClpX provide some sort of redundancy to keep the process of protein degradation going even if one of the subunits becomes damaged. “We were very surprised,” Sauer says. “But that's why we do experiments.”

And more experiments lie ahead, Sauer says. “We want to understand how ATP binding and hydrolysis change the structure of ClpX and allow it to take other proteins apart,” he says. This is unlikely to be a straightforward task, but persistence may, once again, pay off.