Keynote
Can We Read Genes?

Protein folding is among the most important reactions in all of biology. Major progress was achieved in understanding the mechanism of folding of small proteins by fast search over the energy landscape. Yet, current algorithms cannot model the folding of larger proteins (of more than ~100 aa) due to the enormous number of possible configurations to be searched.

The fast folding of such proteins depends on specific preprogramed constraints that reduce the size of the search. We hypothesize that a small number of specific non-local interactions play key role in the initiation of the folding transition of globular proteins. This is the loop hypothesis, very few loops formed in the ensemble of collapsed disordered polypeptides reduce the dimensions of the search and accelerate the folding transition. In order to test this counter-intuitive hypothesis, we developed experimental strategy based on site specific double and triple labeling of model proteins and time resolved FRET detection of fast kinetics experiments (stopped flow and microfluidic mixing). Studies of two model proteins show fast closure of long loops and folding of helical segments (microseconds) while the rate limiting step is in the seconds time regime. The “transfer-quench” technique was developed in order to resolve the question of conditioning of the folding of pairs of sub-domain elements. The long range goal of our studies is to be able to create a “dictionary” of sequence elements which can encode the fast closure of long loops.

This focus on the pre-transition state ensemble events is an important step towards development of an algorithm that would be able to identify sequence elements that form early non-local (and associated local) contacts, and use that knowledge for simulating the folding transition and prediction of the folded structures of most proteins.