Introduction

Protein-protein interactions play a pivotal role in mediating signal transduction pathways and executing many cellular functions. Defining how each protein interacts with its partners under physiological conditions provides insight into the cellular functions of the protein. Several methods have been developed to study protein-protein interactions in living cells. These include mammalian two-hybrid assays (1), protein complementation assays (2,3,4,5,6,7,8,9,10), and fluorescent resonance energy transfer (FRET) (11,12,13,14). We previously took the protein complementation approach and used enhanced yellow fluorescent protein (EYFP) and enhanced cyan fluorescent protein (ECFP) to develop a bimolecular fluorescence complementation (BiFC) assay and a multicolor BiFC assay to directly visualize protein-protein interactions and to study the competition of multiple proteins in living cells (15,16). Due to stronger complementation signal and direct readout, the BiFC assay has been widely accepted and used for the study of protein-protein interaction in living cells ((17,18), and references therein). However, one limitation of EYFP is the sensitivity of chromophore maturation to higher temperatures (19). This requires preincubation of cells at lower temperatures prior to visualizing the BiFC signal (15,16). Therefore, the availability of new fluorescent protein fragments that can be used under physiological conditions will undoubtedly impact the future applications of the BiFC assay. Here we report the identification of several new combinations for BiFC analysis under physiological conditions.

Materials and Methods

BiFC constructs using fragments derived from three newly engineered fluorescent proteins, Venus (20), Citrine (21), and Cerulean (12) were made. Plasmids pBiFC-bJunYN155 and pBiFC-bFosYC155, as described previously (15), were used as templates. cDNAs encoding the N- or C-terminal fragments of Venus, Citrine, and Cerulean were amplified by PCR and subcloned into pBiFC-bJunYN155 and pBiFC-bFosYC155 to replace those encoding YN155 and YC155 ((Figure 1), A and B). To facilitate the detection of their expression, all N-terminal fragments fused to bJun have a FLAG® tag at their N termini, and C-terminal fragments fused to bFos have a hemagglutin (HA) tag at their N termini.

Figure 1.

COS-1 cells were cultured in Dulbecco's modified Eagle medium (Hyclone, Logan, UT, USA) supplemented with 10% fetal bovine serum with penicillin and streptomycin in 12-well plates at 37°C. The indicated amounts of plasmid encoding bJun or bFos fusion proteins were transfected into COS-1 cells using FuGENE®6 (Roche Applied Science, Indianapolis, IN, USA). For ratio analysis, plasmids encoding ECFP or EYFP were cotransfected as an internal control to measure the BiFC efficiency of fragments derived from YFP or CFP variants, respectively. Images of transfected cells 24 h after transfection were captured using a charge-coupled device (CCD) camera mounted on a TE2000-U inverted fluorescence microscope (Nikon, Melville, NY, USA) with JP4 filters (Chroma, Rockingham, VT, USA). The intensity of more than 100 cells was individually quantified using an automated intensity recognition feature of Metamorph II (Universal Imaging, Downingtown, PA, USA). YFP/CFP ratio analysis was similarly performed as reported previously (15). The median of YFP/CFP ratios was used for comparison of the BiFC efficiency among YFP variants and determination of BiFC specificity, since the distribution of the raw data was highly skewed (15). Since 7.8% of the BiFC signal derived from N173 of Venus paired with C155 of ECFP was read in the CFP channel, we corrected this signal crosstalk by multiplying original CFP images with a correction coefficient. Briefly, cells transfected with plasmids encoding N173 of Venus fused to bJun and C155 of ECFP fused to bFos were used as a control to determine the correction coefficient. Fluorescent images were captured with both YFP and CFP filters, and the correction coefficient was calculated as described for FRET (22). The original CFP images in the multicolor BiFC experiments were multiplied by the correction coefficient using the FRET function provided in Metamorph II, and a representative of the corrected CFP images was presented here (see (Figure 3)B).