Hi all,

I have a question about the best approach to dock two proteins, with the hypothesis that one protein has a binding cleft with a large labile binding cap formed by a loop region. A summary of my question is outlined in the attached diagram.

Essentially, we believe that one protein ('Protein A') has a binding pocket in which another protein ('Protein B') sits. We've solved the crystal structure of 'Protein A' alone, not conjugated with 'Protein B'. The unbound 'Protein A' show a cleft, above which lies a large disordered loop region (~14 residues that show no density in the crystal structure). We hypothesise that this region may constitute the docking site for 'Protein B', with the loop region constituting a 'cap' that goes across the top of 'Protein B' after it is bound within the cleft. We know that 'Protein B' binds to 'Protein A' via biophysical data.

My question is concerning the best approach to test this hypothesis. I intend to attempt local docking with many cycles of perturbations of 'Protein B' (with 3Å translations and 24° rotations). The Rosetta tutorials and guidance on docking of flexible proteins suggest use of an ensemble of the open and closed states of 'Protein A'. However I do not have empirical evidence of 'open' states of 'Protein A', only a hypothetical query that the loop region represents a binding 'cap'. I have manually 'opened' the cleft in Pymol however, using the pivot function on the loops region, so that the cleft is exposed. I've referred to this as the 'open' state. Although the Rosetta protocol suggests flipping between ensembles of 'open' and 'closed' states when binding, I presume this is based on empirical ensembles, not hypothetical states. So my question is: should I dock 'Protein B' to 'Protein A' purely in the 'open' state, and then relax the structure around the bound complex, or should I dock 'Protein B' to 'Protein A' using the 'open' and 'closed' ensembles of 'Protein A' via the Rosetta protocol for docking of flexible proteins?

Any help or insight from the Rosetta/structural modeling/computational biology community would be very much appreciated!

Thanks a lot for your time,

Rob Barringer