Why don't you use ChIP assay, amplifying the plasmid DNA in the last step?

Are the proteins known or unknown? If known, you can do a plasmid prep and protein extraction separately and then add the protein and plasmids together to analyse binding.

why not to try to crash the cells with lysozyme and spin down debris? I would expect pDNA with proteins will stay in liquid phase.

There is no easy way to isolate plasmids without denaturing the associated protein. Its not clear from your question whether you plan to isolate proteins bound to a specific DNA sequence or DNA binding proteins in general? Either case, you can make a cell lysate and pass it over the specific DNA sequence or random DNA fragments immobilized on a solid support. Yes, lac repressor is a very good control. You might also want to use as control a protein that binds DNA without much sequence specificity, eg., histones.

I am not sure but you may try incorporate zinc finger and leucine zipper. I think Timothy is correct and this assay will be easier than others.

You may try to lyze carefully the E.coli cells and isolate plasmid-protein complexes by the sucrose gradient centrifugation, similarly as it was done for the SV40 minichromosomes. The problems is that E. coli cell is much harder to open than eukaryote cell and rearrangement of proteins may occur. Lac repressor is OK as a control.

Thank you all for valuable input! I'm working on a library approach, hence I need to isolate plasmids+plasmid encoded dna binding protein from individual bacterial cells keeping the original interaction between dna-protein. I'll try using lysozyme and possibly a sucrose gradient centrifugation to separate the plasmid-protein complexes from chromosomal dna and other debris.

Thanks again!