You are here

comparative modeling of protein-ligand complex

4 posts / 0 new
Last post
comparative modeling of protein-ligand complex

I have the feeling that this question has been posted again but no matter how much I searched I couldn't find anything relevant.

I am modeling a transmembrane protein in the holo form. The natural substrate of my target protein is adenine, which is a planar molecule, but the ligand in the template is not a substrate of the target. I have created many models without using the ligand, but according to some mutagenesis experiments we carried out an aspartate residue which is involved in substrate binding is located too far to interact with the substrates (that was conclude from docking with flexible sidechains using Vina). This time I have shifted helix 10 which carries that residue in the alignment and want to repeat comparative modeling -ideally- with adenine. The reason I want adenine to be present is to attenuate the bias introduced from the template in the active site conformation of my transporter models. So my question is:

Can minirosetta be executed with a template that contains ligands? If yes, how do I specify the ligand in the alignment?

The reason I rejected RosettaLigand and enzyme design is that they don't allow enough backbone flexibility to shift helix 10 to such a position that interaction of the catalytic asparate with adenine will be possible. I believe such drastic movements are more likely to happen during comparative modeling. Also adenine is a planar residue so there is no need for conformation search. Now one could ask "why don't you create new models with the new alignment and then dock the ligand to a selection of them with RosettaLigand application?". Well, the answer is why should I follow all these steps if I can do the same job in a single step?

Post Situation: 
Mon, 2013-09-09 09:06

From what I understand, there isn't the possibility to use ligands with the current comparative modeling protocols. (There should be a way to do it with the upcoming "hybrid protocol", but to my knowledge that hasn't officially been released yet.)

What you would have to do instead is to "fake" the ligand by adding in addtional constraints. If you add constraints to your shifted helix (e.g. by adding constraints between the aspartate and the other ligand contacting residues) you should be able to bias the resulting structure to one where the active site is appropriately constructed.

Mon, 2013-09-09 10:48


Does the Rosetta v3.5 comparative modeling protocol support ligands?

Fri, 2014-02-21 01:46

No, but the "hybrid protocol" that I've mentioned has been released as "RosettaCM" ( and is available with the recent weekly releases.

Mon, 2014-02-24 09:15