First, you need to subclone your insert into a better vector. Then you can try a biotin tag - very strong interaction. This tag would allow you to do IP experiments to find interactions, if that is your goal. What system are you working in?

I agree with Lawrence.

Also, can you show that the protein is expressed at all? Like a SDS PAGE before/after induction?

You could try with TOPO vector, it´s kind expensive but works very well. Another option could be pGEM-T. If you want to study protein-protein interaction PLA system could be useful for you, for ex. http://www.sigmaaldrich.com/life-science/molecular-biology/molecular-biology-products.html?TablePage=112232138. Good luck!

Thanks for all the very useful suggestions. I am working on a viral protein. It's not a recombinant overexpressed protein. We are trying to study the interactions after an infection. I have been trying a pcDNA3.1 cloning by PCR and ideally it should be very simple but it isn't working either so I was thinking about the TOPO vector. I like all these ideas and will start looking into these.

If its possible to set a lysine codon at c-terminus of your protein ( only if it doesn't change the protein conformation or affect it's physiological state) and straight away tag it with biotin followed by immobilization on strept beads. Once you have immobilized protein of interest, you can pulldown the interacting protein partners.

Lack of the signal on western with 3 tags, where the antibodies are against the tag means or no expression at all, or transfer problem. Commercial antibodies against common tags are giving extremely strong signal, are suitable to detect protein at picogram levels. Suspect in your case the transfer problem, but to tell you more I need size of your protein and all the details about your western procedure.

Keep fingers cross - Jan