When the structure of one of the subunit in a heterodimeric protein complex is known, can that phase be used to solve the structure of the other subunit for which no homology exists? In that case what is the phasing methods to use?
Any references?
In principle it can be done. It works pretty well if the known subunit (A) is much larger then the unknown subunit (B).
You should make sure that you have identified all subunits A in the asymmetric unit cell (~determine the copy number) before attempting to trace subunit B.
Next, you don't _really_ need the structure of B, but a somewhat homologous structure would be really helpful. If you have two AB dimer in the ASU, search for 2 copies of A first and then for 2 copies of B using the homologue as search model. You do want to remove all loops and regions/sidechains that are not in your protein from the search model for B.
If those two strategies don't work you should use heavy atom incorporated crystals and MR-SAD in order to solve the structure. Se-Met label is easy to generate in protein expressed from bacteria. If you have indigenous metals in your sample, use those instead. It is much easier to find the correct heavy atoms when you have a partial MR solution.
best of luck,
S.
It's a bit case by case, but if one subunit fits well after molecular replacement and refinement, you should get enough phasing power to solve the other one by (automated or manual) model building. ARP/wARP or Buccaneer may help. This is mostly a matter of trial and error.
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Biology
- Biomolecules
- Proteins