Hsp70s have two major activities: they help unfold stable protein aggregates, and they pull proteins through membrane channels. In the Brownian ratchet model for protein import, Hsp70 grabs part of a protein once it is spontaneously unfolded by random thermal fluctuations, and passively prevents it from sliding backward. By contrast, the power stroke model has the Hsp70 using a channel protein as the fulcrum for a lever arm movement that yanks the protein inwards.
The Lausanne group suggested that an emerging protein with a bulky Hsp70 attached would keep bumping against the nearby membrane. But if the whole complex moved further away from the membrane, this bumping no longer happened. The resulting increase in mobility and thus entropy meant a favorable change in free energy—a change that powers Hsp70's pulling force away from the membrane. Similar physics would draw Hsp70-bound protein segments away from the bulk of an aggregate and thus tease apart the tangles.
“We made a chimeric theory,” says Goloubinoff. “Yes, there is active pulling, but there is no lever.” The Brownian ratchet remains partly intact also, as the unfavorable change of free energy forbids Hsp70 from being pushed back toward the membrane.
“The physics behind it is very simple,” says statistical physicist De Los Rios. “It's been known that you can generate forces on polymers when you manipulate the number of conformations that the polymer can exist in.” With the Hsp70 case, he worked out how much work this change could do and found it was compatible with the forces required for protein unfolding and accelerated import.
In theory, this leap in thinking could have been made at any time in the last 20 years. In fact, says De Los Rios, “the first thing I said is, this is a nice application, but I asked who must I cite. But I couldn't find anything.” Goloubinoff suggests a precedent. “When I was doing my Ph.D. we had all the tools of molecular biology,” he says. “But there was only one person who thought of PCR.”