Fully felixble region did not allow movement


My protein is composed of a few domains, five of which (domains 1-5) are shifting between open and closed conformations.
I have a structure of the open conformation and crosslinking data for the closed conformation. Domain 5 has a couple of links to a region that is far away on the open structure but matches the current model for how the closed conformation looks like.

I set up a docking run when Molecule 1 is the rest of the protein + domains1-3, and Molecule 2 is domains 4-5.
Domains 1-3 probably also move but less. To accommodate this movement I defined the residues in between the domain as fully flexible.
However, domains 1-3 did not move at all.

Is that how fully flexible regions should be used?
If not, how should I do this?

An alternative solution would be to define 6 molecules, one for each domain and one for the rest of the protein. Then add unambiguous restraints between the edges of the domains to keep them chained together.

What is the right way to go about this? Defining flexible regions or using multiple molecules?

Thank you,
Gadi Armony

Large conformational changes are difficult to model by any docking method… Don’t expect miracles.

We did work on that topic in the past. Check the following paper which basically describes a divide and conquer approach:

E. Karaca and A.M.J.J. Bonvin
A multi-domain flexible docking approach to deal with large conformational changes in the modeling of biomolecular complexes.
Structure, 19 555-565(2011).

Another option would be to try to sample some possible conformation prior to docking and provide an ensemble to HADDOCK, e.g. using normal modes.

Your “alternative solution” seems like a good way to proceed (related to the above paper).