Docking lipids on proteins using HADDOCK

HADDOCK
1

I am using the HADDOCK 2.4 webserver to dock a lipid onto a protein. However, the lipid molecule is considered to be rigid and the tail does not seem to be flexible. I have tried specifying the lipid as a fully flexible moiety and also turning off the “rigidmini” parameter. In the first case, the lipid is still rigid and the tail is in the same orientation as the input. In the latter case, the lipid doesn’t reach the protein at all. What should I do to resolve the issue?

2

Dear adev,

Thanks for your interest in using haddock2.4 for your research.

Regarding the question of flexibility:

  • Have you effectively checked that the RMSD between your initial input structure and the final lipid conformation obtained after the run is exactly 0?
  • Turning off the rigidmini parameter is not a wise idea in this case, at it will basically turn off the rigidbody (it0) docking, but just separating the entities by random rotation and translation, hence leading to no contact.

Also, please note that:

  • Flexibility at it1 and itw stages is based on interface, and only interface atoms are considered as flexible. So if the tail is not in contact, it will not be set as flexible in those stages.
  • As lipid are non-amino-acids/non-nucleotides structures, they are considered as small-molecules in haddock. So, similarly to the docking of small-molecules, you could also try to provide an ensemble of initial conformation of your lipid. For this, please refer to the online tutorial about protein-ligand docking where we generate conformational ensemble using rdkit (HADDOCK2.4 shape-restrained protein-small molecule tutorial – Bonvin Lab).

With the hope that this answer will help you.

3

Dear Victor,

Thank you for your response.

Yes, it is indeed 0.

I will give this a shot. Is there an alternative method to dock the lipid as a flexible molecule. As far as the docking result goes, the lipid fits into the binding pocket very well, however it is still in the exact same conformation as the input file and there is no change to the lipid molecule which tells me that the semi-flexible and flexible refinement steps are not working. I find this surprising because any atom in contact with the receptor within 5A should be subject to the refinement by default (based on HADDOCK documentation). Any suggestions on why this might be happening?

4

Dear Archit,

Indeed, as you have clearly stated it: any atom in contact with the receptor within 5A should be subject to the refinement . Therefore it is quite surprising that your lipid is still considered as rigid.

Can you share the url of your run with me?
If you think this is too private, try to send it via email.