We were trying to purify one protein from chickpea seeds. We used to always get one contaminating band during purification. I crystallized that protein and collected diffraction data. The overall resolution is 1.8A. But now we are unable to purify that contaminating protein so we don't have sequence of the same. Is there any way/tool where I can use the diffraction data and solve the structure without protein sequence?
Since you don't know what your protein is, you can't use molecular replacement for phasing. But you can try to soak heavy metal salts into your crystals (e.g.mercury) for phasing by isomorphous replacement (https://en.wikipedia.org/wiki/Multiple_isomorphous_replacement) or, if you have access to a synchrotron, use Multi-Wavelength anomalous dispersion for phasing.
If the protein is small enough you may be able to use Direct Methods (implemented in SHELXT - http://www.iucr.org/news/research-news/shelxt-integrated-space-group-and-crystal-structure-determination). I recently solved a novel protein structure comprised of nonproteinogenic amino acids with 44 residues in the unit cell. Otherwise SAD, MAD, or MIR are the techniques to try (as mentioned by Annemarie).
Nowadays there are also brute-force methods that try massive molecular replacement attempts with general protein parts or using Rosetta fragments: check Arcimboldo and Ample.
Here are few suggestions from my recent experience.
My lab mate crystallised a protein at 1.9 A which we realised to be a contaminant. I did following things,
1) Did a serial Phaser using all available structural domains (~27000 at 95% sequence identity) from SCOP database. Of these only 2 had LLG>140 and TFZ>6 indicating possible solution.
2) These two structural domains were used to search PDB using DALI server. Top 300 structure < 4A RMSD was used serially to run Phaser. 5 of these had TFZ around 12 indicating right fold.
3) An ensemble of these 5 structures was generated using Ensembler program from Phenix with help of Multiple sequence alignment (from ClustalW). Polyalanine of this ensemble was used as Phasing model and Phaser succeed with TFZ=22.
4) I was able to build residues at the core of the protein (but ~60% remained Poly ala). Using this sequence, the sequence family was identified from Pfam.
5) Currently I am using Multiple Sequence Alignment of top hits from Pfam to guide me through Model building. Current LLG (3400) and TFZ (28) clearly indicate that we are taking the right path to structure determination. R/Rfree stands at 28/37 (I am yet to build loops). I hope I will be able to solve this soon...
Hope it helps you.
If in case you still haven't solved this problem:
We have developed an approach (as explained in my previous response) called MarathonMR to specifically address such contaminants. The science is explained in the following papers (links attached). We have developed a wrapper program (unpublished) based on Phenix and GNU Parallel program. There are two versions of it, 1) Single machine version (with multiple core support) and 2) HPC cluster version with MPI support. Please ping me directly so I can share the scripts with you.
Best,
Kaushik
Article Determination of crystal structures of proteins of unknown i...
Article Structure determination of contaminant proteins using the Ma...
@Kaushik Hatti: Thank you sir, could you please send me the second paper.
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Cell Biology
- Culture Cells
- Cell Line