Most proteins must adopt a specific structure to function, but we do not fully understand how proteins reliably and rapidly fold in the cell—a question not addressed by structure prediction algorithms such as AlphaFold. Many proteins begin folding co-translationally on the ribosome, and in doing so attain a higher folding yield than upon refolding in solution. Consistent with this, bioinformatics analyses of codon usage reveal widespread evolutionary selection for local variations in translation rate that may facilitate nascent chain folding. Yet despite the critical importance of co-translational folding, its molecular underpinnings remain elusive, as this kinetically complex, multi-body process cannot be probed by standard biophysical techniques. To address these challenges, we have developed a novel application of hydrogen-deuterium exchange mass spectrometry to map real-time folding at multiple sites along an elongating nascent chain during in vitro translation. Unlike other methods that only monitor stalled nascent chains, our approach probes folding in actively translating proteins, and thus can capture potentially crucial non-equilibrium effects. For example, we show that for the protein HaloTag, folding is delayed until after translation is complete, despite observing that partially translated HaloTag nascent chains are folding competent given enough time to reach equilibrium. We explain how this mechanism may allow HaloTag to avoid aggregated states which plague the protein’s refolding off the ribosome. Meanwhile, a second protein from the TIM barrel family shows starkly contrasting behavior, adopting an N-terminal folding intermediate partly through real-time translation which resembles a refolding intermediate observed in the full-length protein. These studies reveal how protein structural and energetic properties lead to diverse co-translational folding mechanisms, which we can uniquely probe at high resolution. The methodology can be extended to related co-translational phenomena including chaperone activity and disease-linked misfolding.
Bitran et al. (Sun,) studied this question.