Covalent docking as mimic of protein engineering

HADDOCK
1

Dear all,
I have read several threads in this forum and I’d like to have a comment to my procedure. I have a large transmembrane protein (>2000 residues) and I studying the possibility to add a custom tag peptide (max 22 residues long) to the the N-terminal or C-terminal. Since this peptide will be in the cytoplasmic region, my idea is to:

  1. perform a folding study of the peptide (i.e. iTasser)
  2. perform a docking protein-peptide using HADDOCK by using a unambiguous restraint N-ter/C-ter between protein and peptide. Probably I should reduce the vdw weight in the scoring during the it0/it1 in order to satisfy the restraint.
  3. Relax the system protein and embed it into the membrane to perform additional simulations via MD.

I was trying to speed the process exploiting the CHARMM-GUI pipeline or using GalaxyFill (FALC) but they are not cleaver enough to explore the conformational space.

Any comments or suggestions really appreciated.

Best

1 Like
2

Thanks for sharing this protocol - it looks good to me!

Is this tag peptide structured or disordered?

If its a disordered I’d suggest that you use an ensemble of conformations, hopefully the scoring function would be able to fish out the best conformation that fits the complex (and the restraints).

When you move to the MD you’ll probably need to edit the complex to create a bond, but you probably already know that :wink:

Good luck!

1 Like
3

Hi
thanks. Indeed during the topology I will insert the bond.
Yes - I will use the model ensemble from iTasser - or different runs each per model or same run.
The peptide from iTasser is quite disordered but few residues in helix.
Thanks

4
  1. perform a docking protein-peptide using HADDOCK by using a unambiguous restraint N-ter/C-ter between protein and peptide. Probably I should reduce the vdw weight in the scoring during the it0/it1 in order to satisfy the restraint.

This won’t help to fulfil the short bond restraint as the vdw term will prevent a short distance.
What I would do once you have reasonable solutions, is to make it a single chain (make sure from the start that the peptide numbering does not overlap with your protein), and run it then through the refinement server. It should restore then the proper bond distance.

1 Like
5

thanks Alexandre.
So, how much I can push the restraint? I was thinking something like 1.8 ± 0.5, which is longer than C-N of course.

6

It doesn’t really matter - again you won’t get much closer than the sum of the vdw radii

But I would suggest to add a few more distance restraints to define the planarity / trans configuration of the peptide bond

1 Like