This project dealt with fast dynamics of proteins using single-molecule FRET techniques. We studied several protein systems, as discussed below, and came to the general conclusion that proteins tend to demonstrate conformational dynamics on much faster time scales than their functional cycles, even while these conformational dynamics affect significantly the function. This conclusion was based on the following studies:
- In the protein adenylate kinase we found that domain closing and opening occur on time scales of 15 and 45 microseconds, respectively, while the protein turnover time is 2.5 milliseconds. We argued that this disparity is due to the need of the protein to optimize the relative orientation of its substrates for the reaction. Recent molecular dynamics simulations and further experiments on the AMP-induced substrate inhibition mechanism of the protein confirm this mechanism.
- In ClpB, a disaggregation machine, we studied several different parts and observed fast dynamics in them. In particular, we showed that the switch of the machine, the so-called middle domain, jumps between two states on the sub-millisecond time scale, making it a continuous, analog-like switch. We also found that the N-terminal domain of ClpB inhibits the middle domain through space-occupying interactions, a mechanism we called 'entropic inhibition'. Finally, we found that protein loops lining the central cavity of ClpB, which are termed 'pore loops' and have been shown to be important for protein-substrate translocation, are very dynamic and move up and down on very fast time scales. This motion depends on the presence of substrates and nucleotides. We proposed that a Brownian ratchet mechanism may explain our results, with two time scales involved- a fast time scale of the pore-loop motion and a slow time scale of nucleotide hydrolysis.
- In GroEL, the molecular chaperone, we were able to demonstrate that the dynamics of each subunit can be described using four microstates. Under each solution condition (e.g. with or without nucleotides), the relative population of these four states changes, but they still produce an appropriate description of the overall ensemble.
The above studies strongly suggest that the finding of two time scales is quite universal. We are working now to demonstrate the same phenomenon in other proteins, such as the enzyme phosphoglycerate kinase.
