I am curious whether it is possible to download the pdb file of a mouse kappa light chain, and change the residues (mainly CDRs) to that of my light chain's in order to obtain a plausible 3D structure? I only have experience in Jmol; are these modifications doable using Jmol? Would I need to re-fold the protein or to account for torsion angles in the backbone? I'd be really grateful if someone could point me in the right direction.
Jmol can do simple mutagenesis (http://chemapps.stolaf.edu/jmol/docs/#mutate) Whether this is enough to produce a good model depends on the particular VL sequence. First, I would download several of the best templates from the pdb and do a 3D-alignment of these VL domains to see how well they agree and highlight (by a different color) the residues that are different between your query sequence - if non-conservative substitutions are all solvent exposed, and you do not have conformationally critical residue types involved (from and to Gly, from and to Pro) and no differences in loop length, you may be able to get a pretty good model, especially if your query sequence belongs to a germline sequence that highly represented in the pdb. However, there quite a number of specific antibody modeling servers available that will do a better job:
Rosie (Rosetta): http://rosie.rosettacommons.org/antibody
PIGS (http://circe.med.uniroma1.it/pigs/)
Kotai antibody builder (http://kotaiab.org)
just add some dummy VH sequence if you only need VL, e.g. EVQLVESGGGLVQPGGSLRLSCAASGFTFSSYAMSWVRQAPGKGLEWVSAISGSGGSTYYADSVKGRFTISRDNSKNTLYLQMNSLRAEDTAVYYCARWGGDGFYAMDYWGQGTLVTVSS
Thank you very much for your high-quality answer; it was exactly what I was looking for. Since I asked the question I have managed to stumble into Swiss-Model, and I have basically done what you recommended first, I compared the sequence to numerous known structures with the highest sequence similarity.
However, I did find PIGS as well and it had given me much better results. Having practiced PyMol I now have excellent visualisation of the light chain I'm looking at. The second problem I'd like to tackle is docking of the antigen, a small molecule, onto the CDRs. I tried downloading Autodock, but neither version worked; altough this really is a subject for another question...
Thank you very much again for your response.
For docking, there is the ligand docking server under Rosie:
http://rosie.rosettacommons.org/ligand_docking
- Badges
- Science topic
- Similar topics
- Biological Science
- Immunology
- Antibodies