I have diffracted a protein crystal with a resolution of 2.2A. While solving the structure using molecular replacement I am not getting density for 2 residues that form a turn. These two residues are important as I have incorporated a mutation here (Gly to Arg). I have already changed the I/sigma value as well as resolution to check if I can get density across the residues. However, there wasn't any change. Now what else can be done with the same dataset so that I can get electron density?
Hi, is this turn/loop on the surface of the protein? It sometimes occurs that such residues are not visible in electron density, due to flexibility. If you are confident about the quality of your diffraction data, you could leave out the 2 residues from the model until the end of refinement and include them again in the final stage.
Tjaard Pijning Yes, the loop is on the surface.
Sorry, but I couldn't understand the second part of your reply.
Remove the 2 residues from your model, and perform the refinement of the rest of the protein model until convergence. If you then see positive electron density for the 2 missing residues, you can try to add them again. If you don't see (enough) positive electron density, they are probably too flexible and it is better to leave them out from the model.
Tjaard Pijning Ok. So do you mean to change the occupancy of the residues to zero and refining further or deleting the residues?
I have even changed the mutated residue to alanine and refined without changing the residue in molecular replacement step. However, in both case, I couldn't find the electron density.
Subsequently, since my mutation is in the loop and my aim is to compare the wild type vs mutant protein structure can I further analyze the structures even if my mutant residues do not hold any electron density.
Hi Shubhashish Chakraborty .
First of all, congrats on the new structure!
I agree with Tjaard Pijning: completely remove the two residues from the model, model and refine everything else (main chain, side-chains, cofactors, ligands, waters, etc), and then come back to this loop. If everything goes well, the phases will improve (and so the electron density), and you might be able to get back some electron density for this small stretch.
Besides this approach, you don't mention what program you're using for refinement. I suggest you try BUSTER-TNT: this program has a special flag ("-L"), which quite often helps to improve density. It's very similar to PHENIX Polder Map, but better (more robust) in my opinion.
All the best, HTH,
Jose
Hi Jose A. Brito I am using refmac and phenix for refinement.
Dear Shubhashish Chakraborty ,
I don't use REFMAC for years now, so I really don't know what new features and improvements the program has to help you on this.
Regarding PHENIX, try the POLDER map under "Maps". This usually helps to "improve" electron density, and might help you to resolve the loop.
BUSTER-TNT is also very useful. If you want, we can discuss off-board how to test this.
Best,
Jose
Jose A. Brito
I mutated all the residues within the loop to glycine and now while building the model/loop I can find electron density for most of the residues. However, there are still some on which I have to work and refine the structure so will surely take your suggestion.
- Badges
- Science topic
- Similar topics
- Analytical Chemistry
- Column