ABSTRACT: Bimolecular Fluorescence Complementation (BiFC) assay is a method used to directly visualize protein-protein interaction in vivo using live-cell imaging or fixed cells. This protocol described here is based on our recent paper describing the functional association of human chromatin adaptor and transcription cofactor Brd4 with p53 tumor suppressor protein (Wu et al., 2013). BiFC was first described by Hu et al. (2002) using two non-fluorescent protein fragments of enhanced yellow fluorescent protein (EYFP), which is an Aequorea victoria GFP variant protein, fused respectively to a Rel family protein and a bZIP family transcription factor to investigate interactions between these two family members in living cells. The YFP was later improved by introducing mutations to reduce its sensitivity to pH and chloride ions, thus generating a super-enhanced YFP, named Venus fluorescent protein, without showing diminished fluorescence at 37 °C as typically observed with EYFP (Nagai et al., 2006). The fluorescence signal is regenerated by complementation of two non-fluorescent fragments (e.g., the Venus N-terminal 1-158 amino acid residues, called Venus-N, and its C-terminal 159-239 amino acid residues, named Venus-C; see Figure 1A and Gully et al., 2012; Ding et al., 2006; Kerppola, 2006) that are brought together by interaction between their respective fusion partners (e.g., Venus-N to p53, and Venus-C to the PDID domain of human Brd4; see Figure 1B and 1C). The intensity and cellular location of the regenerated fluorescence signals can be detected by fluorescence microscope. The advantages of the proximity-based BiFC assay are: first, it allows a direct visualization of spatial and temporal interaction between two partner proteins in vivo; second, the fluorescence signal provides a sensitive readout for detecting protein-protein interaction even at a low expression level comparable to that of the endogenous proteins; third, the intensity of the fluorescence signal is proportional to the strength of protein-protein interaction (Morell et al., 2008); and fourth, the BiFC signals are derived from intrinsic protein-protein interaction, rather than from extrinsic fluorophores that may not reflect true protein-protein interaction due to their nonspecific association with cellular macromolecules or subcellular compartments. However, some limitations of BiFC include slow maturation (T1/2 ~ 1 h) of an eventually stable BiFC complex (Hu et al., 2002), making it unsuitable for real-time observation of transient interaction that disappears prior to BiFC detection, and enhanced BiFC background at high expression levels due to fusion-independent association between two non-fluorescent fragments association. BiFC signals generated by in vivo protein-protein interaction can be validated by amino acid mutation introduced at the protein-protein contact surfaces. This imaging technique has been widely used in different cell types and organisms (Kerppola, 2006).