Shalini,

use these and see what you get:

http://www.cbs.dtu.dk/services/SignalP/

http://www.cbs.dtu.dk/services/TMHMM/

No Signalpeptide / signalanchor means the likelihood for disulfide bridges goes down. For transmembrane proteins you need to figure out the topology and check whether the Cysteines reside on the cytosolic side or the secretory side.

Be aware that plant proteins can also have a transit peptide for plastid targeting. The redoxpotential of chloroplasts is light dependent, also there are many membranes and thus different topologies / localisations possible.

* Transcription factor = cytosolic/nuclear, hence reduced

* Enzymes with Cys at the catalytic center = reduced (as ox. Cys would render the enzyme inactive)

you can use http://aps.unmc.edu/AP/main.php for prediction (its indirect approach, it will predict AMP with disulfide bonds)

Thanks Neha. But my protein did not work with this software.

Thank you Alison. The software did not pick any disulphide bonds for my protein

(which surely has disulphide bonds). Would you be able to suggest some guidelines for expecting disulphide bonds just by looking at the protein (apart from C-C).

Hello Shalini..first maybe you can identify by yourself what role these disulphide bonds might be playing in your protein(i.e. is your protein a catalytic protein or a structural or a binding protein). The reason a number of disulphide bonds dont get picked up from these servers is that they are overly dependent on evolutionary profiles. However many disulphide bonds particularly allosteric ones will not conform to evolutionary profiles as they have a tendency to form and break based on physiological conditions(even though the C-alpha atoms might be in correct juxtaposition but the oxidation/reduced state of the residues will actually determine if the bonds exist or not). The ones which are more likely to be picked up are catalytic disulphides which would actually be highly conserved. As Alison suggested if you know the 3D structure of the protein/predicted or in the database might be a big help and even in that case one is not sure that in its in vivo state/physiological compartment there is a bond or not. Only functional analysis can give you the definite answer.

I don't know if these programs will work, but they may be worth a shot. http://www.ncbi.nlm.nih.gov/structure

The best approach is to combine all the available information and then make a guess. You can also look at the subcellular localisation of your protein. Cytosolic proteins do generally not have disulfide bridges due to the reducing environment whereas proteins targeted to the secretory pathway / secreted have a much higher propensity for forming disulfide bridges. Are you working with a secret protein? Or can you post the sequence and tell us from which organism it is? And maybe you could let us know why you want to know this, just out of curiousity or out of any practical needs?

It depends on the effort you are willing to spare, and whether you just have a sequence, or you have actually purified the protein. If the latter, you can directly measure its mass by MS; contact the nearest MS facility on how to proceed. If the former, there is software available where you simply feed it the protein sequence, but as others have commented here the results are not very reliable. In the absence of direct structural data, as Markus above points out you simply have to collect all available information and connect the dots. Subcellular localization is important, but you can also do a homology search and comb the literature for disulphide bridiging data on better-known homologues. Sometimes cys residues are involved in catalysis; is your protein an enzyme? Another thing would be to send your protein to a homology modelling server. If a close homologue with a reasonably good structure exists, inspect the resulting model to see if your cystein residues are paired or not, or close enough to pair.

Thanks Alison. The protein I am interested in, does not have a deposited PDB structure. The protein has 9 cys and is approx 500aa long. I can guess the regions where disulphide bonds might occur, just needed a bioinformatic prediction tool to do the same work. But I can definitely check for the structure of the related proteins if that might help with some input.

Shalini,

use these and see what you get:

http://www.cbs.dtu.dk/services/SignalP/

http://www.cbs.dtu.dk/services/TMHMM/

No Signalpeptide / signalanchor means the likelihood for disulfide bridges goes down. For transmembrane proteins you need to figure out the topology and check whether the Cysteines reside on the cytosolic side or the secretory side.

Be aware that plant proteins can also have a transit peptide for plastid targeting. The redoxpotential of chloroplasts is light dependent, also there are many membranes and thus different topologies / localisations possible.

* Transcription factor = cytosolic/nuclear, hence reduced

* Enzymes with Cys at the catalytic center = reduced (as ox. Cys would render the enzyme inactive)

Thank you Arijit. The protein is a membrane bound enzyme. It is true that the functional analysis can only give me the correct answer, but I was just playing around with some bioinformatic prediction tools if they might give me some hint somewhere.

Thanks Casey. I have already tried those, but couldn't get any conclusive answer.

Thank you Markus. It is a human protein, not cytosolic, but Golgi/ER localized. I was asking out of curiosity for practical needs.

Thank you Alejandro for the suggestion. The protein is a membrane bound enzyme. I just have the protein sequence. And a mutation. There is a predicted disulphide bridge at the region of my mutation which might be getting disrupted. I wanted to bioinformatically confirm this prediction with various other tools. Doing functional assay would anyway answer my question, but all I am looking for is a preliminary prediction tool(s) that could give me some valuable inputs.

Markus, thank you for the links. They were helpful. As expected, the protein shows to have a signal peptide with a high value. But I couldn't decipher much from the TMHMM results.

TMHMM results are a bit tricky but if your protein has a clear transmembrane domain you can usually see it. Did you have a look at the graphical results?

PDBsum

Pic server

VADAR

Disulfide by Design

DiANNA : http://clavius.bc.edu/~clotelab/DiANNA/

Dinosolve: http://hpcr.cs.odu.edu/dinosolve

DISULFIND: http://disulfind.dsi.unifi.it/.

GDAP: a web tool for genome-wide protein disulfide bond prediction: http://www.doe-mbi.ucla.edu/Services/GDAP.

Dinsolve is a good option.

Article Dinosolve: A Protein Disulfide Bonding Prediction Server usi...