Predicting Protein Ligands For A Protein

Question

[pld]

I have a protein whose structure has been resolved and the structure of the complex it forms with a known target is also known. There is suspicion that it is able to bind other proteins. Structure focused analysis is new to me, can anyone make suggestions for some approaches that can be used to hunt for potential targets of this protein?

In other words, can I use the knowledge of the structure for the protein in question and its only known target to predict other binding interactions. This protein, CPXV203, is a viral protein so I would be interested in human targets. This adds a challenge to the analysis since there is only a few structures of CPXV203 known and only one interaction, binding MHC-1.

[Pappu]

Perform a SILAC pull down assay to find out the interacting partners.

[pld]

The goal is to do some bioinformatic leg work before spending the money on pull downs, plus the cost and time of generating an antibody for the protein in question. Computing is a flat fee per month, reagents aren't.

[Pappu]

You need to have a lot of homologs of the protein in order to do bioinformatic analysis which does not seem to be the case.

Answer 1

[Arnaud Ceol]

You can look for known interactors on protein-protein interaction databases. The psicquic-view app is a fine starting point to access the major databases: http://www.ebi.ac.uk/Tools/webservices/psicquic/view/main.xhtml .

Once you have a list of interactions, you can submit them to interactome3D (http://interactome3d.irbbarcelona.org) in order to see if there is already a known structure or if a model can be built.

You can also start directly from interactome3D and browse the precompile network.

[pld]

There's nothing for the protein I'm interested in, CPXV203. I was more interested in approaches that don't require sets of known interactions. In other words, can I use the knowledge of the structure for the protein in question and its only known target to predict other binding interactions. I've updated the question to better reflect the scenario.

Answer 2

[jackuser1979]

If you structure deposited in pdb you can search using pdb code. If your structure still not submitted in pdb, you can search by your sequence in uniprot blast search, ncbi blast search to get uniprot id or just paste your sequence in SAS.

After perfoming SAS search, you can get either get protein name, gene name, NCBI gene id for your protein structure. Then use protein search in PRIMOS resticting to your opted organism. You can also search with pubmed id in interaction search provided in PRIMOS

[pld]

How should the E-value field be interpreted with SAS? The lowest E-value I get is 1.4. The other problem is how to determine which of the proteins there would be an interaction with when it comes to complexes.