I'm preparing to start an enzyme redesign project using Rosetta, but the paper I'm supposed to be working off used the older enzdes application, which doesn't seem to be compatible with newer scorefunctions and sampling methods. CoupledMoves seems preferable, and I was hoping to get some recommendations for putting together a ligand docking and design protocol. I have basic familiarity with Rosetta, but since I'm the only person in my lab using it right now, I'd really appreciate any advice to help speed up the process and reduce time spent on trial and error.
Briefly, I'd like to redesign a NADPH-dependent sulfite reductase to accept NADPH analogues. I don't have a bound structure with NADPH, but I can infer the catalytic geometry and cofactor placement from the structure of a related homologue, so I'd imagine that local high-resolution docking should be sufficient. I'll start with NADPH for now and then hopefully move on to more exotic cofactors later. My basic plan is to pre-relax my protein, since the crystal structure isn't great, then do separate or cyclical high-resolution docking and design steps.
My questions are:
- Should I include the roughly-placed NADPH in the pre-relax or omit it and do blind docking?
- Would fixing the catalytic geometry be as simple passing a constraint file? I haven't seen any examples of this being done with CoupledMoves.
- Would it be best to do cycles of docking and design or just get a nicely docked structure/ensemble and then do design? Either way, I expect to manually check to make sure that the expected H-bonds and water-mediated contacts form after the docking and design steps.
- Does anyone have any RosettaScripts (and/or PyRosetta) code that might be helpful as a starting point? I've looked at the commands from the CoupledMoves papers, as well as the template script in the Rosetta docs, but my use case seems more complicated.
- Any other advice? Any recommendations or guidance would be helpful and greatly appreciated.