I would suggest using AviTag - BirA system as described in the recent Nature Immunology publication by Rudra et al. (PMID: 22922362).
Briefly, you transfect your cells with a bicistronic vector that contains your gene of interest and BirA biotin-protein Ligase. Your protein is fused to an AviTag sequence, which gets specifically biotinylated by BirA inside the cell. Avitag is a short peptide sequence that usually does not interfere with protein function or localization.
Pulldown is highly efficient using streptavidin-beads and almost free of any unspecific background - due to the high affinity between biotin and streptavidin, this method is far more efficient than classical antibody-based pulldowns (but will need some more cloning, if you don´t already have the required vectors).
We use a lot HA tag in xenopus oocyte. The sigma mouse antibody give no background and is very sensitive. The GFP antibody from roche ( a mix of two monoclonal) is similar.