Protein interactions are critical for normal biological processes and molecular pathogenesis. While it is important to study these interactions, there are limited assays that are performed inside the cell, in the native cell environment, where the majority of protein-protein interactions take place. Here we present a method of studying protein interactions intracellularly using one protein of interest fused to a localization-controllable enhanced GFP (EGFP) construct and the other protein of interest fused to the red fluorescent protein, DsRed. Nuclear translocation of the EGFP construct is induced by addition of a ligand, and the difference in nuclear localization between the induced and noninduced states of the DsRed construct provides an indication of the interaction between the two proteins. This assay, the nuclear translocation assay (NTA), is introduced here as broadly applicable for studying protein interactions in the native environment inside cells and is demonstrated using forms of the coiled-coil domain from the breakpoint cluster region (Bcr) protein.

Introduction

Protein-protein interactions are an essential component of almost every cellular process. These interactions regulate signaling events that result in differentiation, proliferation, regulation of gene transcription, repair mechanisms, and inhibition or induction of apoptosis to name only a few (1). Studying these interactions can provide valuable insights into the molecular mechanisms governing these cellular processes. Furthermore, aberrant protein interactions can lead to diseases such as cancer (2), diabetes (3), cystic fibrosis (4), Alzheimer's disease (5), and a number of metabolic disorders (6). Thus, the understanding of protein interactions is also critical in the understanding of disease and the development of therapeutic interventions to treat them. The majority of these interactions between proteins occur inside cells, and intracellular experimental assays for protein-protein interactions are the most relevant. However, currently there are few assays that provide a means to quantify protein-protein interactions in the native cell environment. Resonance energy transfer (RET), including bioluminescence resonance energy transfer (BRET) (7) and Förster (fluorescence) resonance energy transfer (FRET) (8), is one mechanism for detecting intracellular protein interactions but requires a specific combination of the right donor and acceptor, along with the sensitive detection of light emitted at specific wavelengths. Further, the proper orientation of the donor and acceptor dipoles within the Förster distance must be achieved for effective energy transfer. The two-hybrid system (yeast or mammalian) is another assay for intracellular protein interactions (9), but the interaction typically takes place inside the cell nucleus, and this assay requires the transcription/translation of a reporter protein that necessitates the expression of three proteins (or even four if normalizing to a second control protein). CytoTrap is one variation of the two-hybrid assay that allows the interaction to occur in the cytoplasm. The split ubiquitin assay (10) is similar to the two-hybrid system, but uses truncations of ubiquitin in place of transcription factor domains. Here we present a new method for the qualitative analysis of protein-protein interactions that is performed inside cells and is based on an optimized ligand-responsive protein and the simultaneous quantification of fluorescence intensity from enhanced GFP (EGFP) and the red fluorescent protein, DsRed.

The controlled nuclear translocation in this assay occurs through what we have termed a protein switch (PS). The PSs we have developed are chimeric proteins consisting of three components: (i) a nuclear export signal (NES); (ii) a nuclear localization signal (NLS); and (iii) a ligand binding domain (LBD). In the absence of ligand, the NES directs the PS to the cytoplasm. However, when the ligand binds the LBD, it causes a conformational change, exposes the NLS, and the protein is redirected to the nucleus. In our laboratory, various combinations of these three components have been studied to identify the optimal combination that results in the greatest amount of nuclear translocation upon ligand induction (11,12). This nuclear translocation is concentration-dependent and reversible if the ligand is removed (12). One such optimized PS, used herein, consists of the NES from HIV rev, the NLS from MycA8, and the LBD from the rat glucocorticoid receptor containing the C656G mutation, which renders it 10 times more sensitive to the synthetic glucocorticoid dexamethasone (13). This PS provides an easy method for controlling the subcellular shuttling between the cytoplasm and the nucleus.

The ability to control the translocation of a protein into the nucleus has led to a biochemical assay for studying protein-protein interactions that we have termed a nuclear translocation assay (NTA). The general concept of the NTA is depicted in Figure 1. A protein of interest is subcloned and expressed as a fusion protein with the PS and EGFP. As a result of the PS, this fusion protein is responsive to ligand and will translocate into the nucleus upon ligand induction. A second, nonligand-responsive protein tagged with DsRed is co-expressed with the PS construct. If the two proteins interact (Figure 1, top) the second protein will translocate into the nucleus along with the PS. The percentage of DsRed inside the nucleus is quantified before and after the addition of ligand, and the difference (termed percent nuclear increase, or PNI) is indicative of the interaction between the two proteins. However, if there is little or no interaction between the two proteins (Figure 1, bottom), the PS alone moves to the nucleus, and there is no change in the subcellular localization of the second protein. Fluorescent protein tags are used as a means of determining the cytoplasmic and nuclear localization of the two proteins through fluorescence microscopy, and thus the fluorescent signals from the two proteins expressed in the same cell need to be distinguishable via excitation/emission filters. Two such fluorescent proteins are EGFP and DsRed, which exhibit excitation/emission of 480 nm/510 nm and 545 nm/620 nm, respectively.