Can you knock it out of your bug of interest?  It seems like that might help to give an idea of what it does.

Dear Dr. Thachangattuthodi ,

I dont have a lot of experience in this, but just my quick thought. Would you be able to look for any homologs of the protein based on amino acid sequence identity and predict the function of this protein based on the similarity with other proteins that share a high sequence identity with the protein of interest?

I hope this help!

Hi Tuan Hoang,

I already mentioned that the protein does not show any structural similarity. Similarly it is more unstructured protein.

Any way thanks for your suggestion.

Anish

But are you looking at tertiary structure (i.e. 3D shape) or primary structure (i.e. amino acid sequence)?

So how did you purify it? How did you follow your purification procedure? Based on what? Blotting or how?

Can you do some transcriptional analysis, to see whether it responds to development/stresses etc.? It won't tell you exact funtion but may shed some light into what processes is it involved in.

What about trying to identify interacting partners? From those you could be able to do something.

Do you at least know, whether is it enzyme (has some cofactors) or the function is rather regulatory/structural etc.?

Hi Tomas Hluska,

i did both the tertiary structure analysis as well as the primary structure analysis and it shows it to be highly disordered by tertiary and no homology by primary structure analysis. I clones the protein sequence and purified it.  I have carried out some interactive partner studies using ITC, BLI and peptide arrays with inteacting proteins but these only tell that then interact and not function. Its not know its any an enzyme or not. due to the interactions we assume it to be regulatory but just trying to get how to go about it in that direction.

Thanks Anish

A biological function is an adaptive biological process. Structures correspond to states not processes; a process always entails more than one structural state. So you can't unambiguously assign a function from looking at a structure (at either the primary or tertiary structural levels.) Only the study of the whole organism, in a context that recapitulates its natural habitat, can divulge biological functions, because evolution acts on organisms not on isolated gene products.

To do this properly, you need to understand how the transcription/translation of the gene correlate with various physiological scenarios. You need to understand what metabolites and biomolecules interact with the gene product, and how those interactions correlate with various physiological scenarios.

Whatever you do, I recommend you run an intrinsic disorder predictor (IUPred is my favorite, linked below) to assess whether you have a globular protein. Most of the tertiary structure prediction tools will give confusing results for a highly disordered protein.

http://iupred.enzim.hu/

Can you knock it out of your bug of interest?  It seems like that might help to give an idea of what it does.

Hi Amanda,

We are in the process of doing the knowout to check what happens. Thanks for your suggestion.

Anish

Dear Anish, my suggestion is using the different protein refolding solutions and buffers (kits) followed by tertiary structure analysis.

Afshin

Dear Anish, one way to do is to use fluorescence microscopy. Tag your protein-of-interest with GFP or any other fluorescent marker and express the recombinant protein in the bugs. The location of the protein within the bug’s cell may suggest something about the protein’s function. The disorder in your protein-of-interest was identified by bioinformatics prediction or experimental studies such as CD? If the molecular size of your protein-of-interest is less than 30 kD, then you may also try NMR.    

If it is an Intrinsically Disordered Protein (http://en.wikipedia.org/wiki/Intrinsically_disordered_proteins) then you still have some structural options. Obviously NMR but you can do a Ensemble Of Models with SAXS without isotropic labelling and much smaller amounts of protein. Some old slides here http://www.embl-hamburg.de/biosaxs/courses/embo2012/slides/eom-tutorial-tria.pdf from Hamburg / Giancarlo Tria. 

IDPs are supposed to be more common in eukaryotes. Is it a bacterial or eukaryotic bug? And if you have recombinant expression or antibodies you can do pull-down assays followed by mass spec. 

Overexpressing the protein in the bug might also give you some good clues. Especially if it's a transcription regulator. You can compare mRNA levels. 

Hi Anish, I assume that you have done BLAST of your protein (or nucleotide) sequence?

Dear Anish,

Please look for distant homologs of your protein, if you get a hit where the template protein is an enzyme your job will be half done. Next you can go for SAXS which will tell you about the most probable shape. If the ligand is known for your protein, SAXS may provide a mechanistic insight.

--

best,

Kalpana

Hi Anish, 

You can do the following to test the biological activity of this protein by:

1. Microinjection and observe the effect. 

2. Classical experiment: You can do RNAi and see what is this protein do. 

3. You can try to identify its receptor. 

Good luck

Non trivial question though. Oh! sequence is known, then you can play for a while. Since the sequence is known, try PROSITE. You may find some motifs. Size of protein, composition of sequence and stretches in sequence and any other information may be helpful. Nature of bug (bacteria) I suppose, gram negative; gram positive; micro-biological data would help in some guessing; type of protein (globular; membrane; fibrous). Function was traditionally determined through trial and error and hence there is a long history behind each protein science.

if you known the sequence of the proteins, then you may :

1. look for homologs of the protein in other organisms. start with protein BLAST and see if any proteins with known function match to your query protein sequentially.

2. if step 1 is of no use, look for remote homologs. try programs such as PSIBLAST, FFAS, HMMER, HHPred. 

3. you can try run secondary structure prediction (tool: JPRED) and tertiary structure prediction (eg. MODELLER, iTASSER etc). see if any known proteins pop up in tertiary structure predictions.