Hi all,
I posted a topic previously regarding the protein complex I am trying to model. Essentially I am trying to model the complex between an RGS protein and a G-protein alpha subunit (GTP and Mg-bound). In order for the structure of the complex to be consistent with literature, the G-protein alpha subunit must be catalytically competent, meaning GTP and Mg2+ should be bound to certain residues in the G-protein alpha subunit in a manner resembling the transition state in the GTP hydrolysis reaction.
I was told to just run HADDOCK without any unambiguous restraints at first. However, my analysis of the data suggests that there are some issues in terms of the GTP/Mg drifting during the docking simulation. I’ve even seen some structures where the GTP is ejected from the pocket and it just floats in space. The validity of the results depends on the G-protein being in a state that is catalytically competent, so I think that it would be best to introduce unambiguous restraints holding at the least the GTP in place.
I’ve prepared a .tbl file for the unambiguous restraints, but I cannot seem to use it when setting up the simulation.
Are you running using our web server or a local version?
Can you share your PDB file and restraints? Or point to a result page of the server (if using the server)?
I am using the web server at this point.
I have a few result pages. For the most recent run of the simulation, I used ambiguous restraints based on an Alphafold 3 prediction of the complex. I have not analyzed the results from this yet, but I need a comparison with simulations utilizing ambiguous restraints based on the literature.
The simulations I refer to in this post with the drift of the cofactors are found at: ( HADDOCK results page ) and ( HADDOCK results page ). There is one more results page (the very first simulation I ran which included the cofactors) but I cannot put it in this post as I am a new user limited to two links per post.
Ignore the name of the simulation on HADDOCK.
I cannot upload the PDB file on the forum here as I am considered a new user, but I have uploaded them to a file sharing website (Wormhole.app). I can provide a link if necessary.
Here is a direct copy-past of the contents within the restraint file (for unambiguous restraints):
Gln205(NE2)-GTP O18/O3G (gamma-P) (per-rank: 2.801, 2.542, 2.697)
assign (resid 205 and segid A) (resid 356 and name O3G and segid A) 2.680 0.50 0.50
Gly204(N)-GTP O14/O1G (gamma-P) (per-rank: 2.790, 2.779, 2.757)
assign (resid 204 and segid A) (resid 356 and name O1G and segid A) 2.776 0.50 0.50
Ser47(OG)-Mg (MG1) (per-rank: 2.148, 2.141, 2.146)
assign (resid 47 and segid A) (resid 357 and segid A) 2.145 0.50 0.50
Lys46(NZ)-GTP O20/O2B (beta-P) (per-rank: 2.760, 2.744, 2.768)
assign (resid 46 and segid A) (resid 356 and name O2B and segid A) 2.757 0.50 0.50
I know the formatting is not perfect, but that is the file I am referencing. For ambiguous restraints (basically I ran a solvent-accessibility script to filter out residues that aren’t solvent exposed within the experimentally defined interaction regions for homologous proteins), I do have .tbl files, but I have primarily used the interface on the server to set those up.
I am using the Web server. I requested Guru-level access, but I only really need to be able to utilize unambiguous restraints.
The IDs for the simulations in which I did not use any unambiguous restraints are: 725652, 726122, 726401, and 730024.
Took a quick look at the last run you provided.
All looks fine to me. The ligand (GTP) is still in your protein.
Below is an image of the top1 model of the top3 clusters, superimposed on your input model in yellow.
This ligand did not move away. Some small differences are to be expected because of the refinement.
To provide unambiguous restraints you will need expert access if you don’t already have it.
