A protein is forcefully unfolded before migrating into the GroEL-GroES chamber, where it either folds properly or must be ejected for another round.
RYE/MACMILLAN
The double-ringed GroEL chaperone complex is well-known for capturing partially folded proteins and confining them within its central cavity to facilitate folding. GroEL is especially important to proteins that have a complex folding pattern, such as Rubisco. A prominent model of GroEL function suggests that the chaperone first partially unfolds its substrate, disrupting misfolded and inhibitory conformations, thus giving them a chance to refold properly once inside the chaperone cavity. But the importance of this effect for successful refolding has been unclear.
To explore this question, the authors tagged the two loose ends of partially folded Rubisco for FRET analysis. They found that the two ends of Rubisco were close upon initial binding to the open end of GroEL. But when ATP bound to the complex and triggered a GroEL conformational change, the Rubisco ends separated. The same conformation change allowed the complex to bind GroES, GroEL's smaller partner, thereby causing Rubisco to enter the cavity.
But did this unfolding aid in proper folding? ATP-driven unfolding works too fast to answer that question, so the authors examined the slower, passive unfolding of Rubisco on a single-ringed form of GroEL called SR1. By incubating Rubisco with SR1 for varying lengths of time before adding ATP and GroES, they showed that more time with SR1, and therefore more time to unfold before entering the SR1 cavity, produced a higher proportion of successfully folded Rubisco molecules.
“We think this represents a really clean demonstration that unfolding can directly stimulate productive folding,” Rye says. 
