Proximity or no proximity: That is the question—But the answer is more complex

Abstract

THE significance of molecular interactions has been recognized since the dawn of molecular biology, and the real importance of the vast amount of information hidden in the human genome can only be understood if the interactome is also defined. Conventionally, molecular interactions are investigated by chemical crosslinking and coprecipitation. Although these approaches are widely used, they do not provide quantitative information, suffer from the lack of cell-by-cell resolution, and potential artifacts arising from high protein expression and membrane solubilization in the case of membrane proteins. Cytometry offers several alternative methods to assess protein–protein interactions. Fluorescence resonance energy transfer (FRET) has made its way to become the golden standard in quantitative analysis of protein interactions (1). In FRET, a fluorescent donor dye residing on a protein of interest transfers its excitation energy to a nearby acceptor molecule which labels another protein. The rate of this radiationless interaction decreases with the sixth power of the separation distance making FRET a sensitive tool for molecular interactions. In conventional FRET, the pairwise interaction between two spectroscopically distinct molecules is analyzed, and the principle has recently been applied for the simultaneous investigation of two pairwise interactions (2) and for the investigation of the three-way interaction among three molecules (3,4). Although FRET usually takes place between spectroscopically distinct molecules, the same process can also occur between molecules of the same type. This flavor of the interaction is called homo-FRET, and it has been used for the detection of molecular trimers (5) and even larger clusters (6,7). The combination of total internal reflection microscopy with FRET allowed the high-throughput screening of molecular interactions (8). A multitude of correlation techniques are available for the analysis of protein clustering. The mother of all such approaches is fluorescence correlation spectroscopy (FCS) in which the autocorrelation of fluorescence signals is measured in a femtoliter volume to determine the diffusion constant (9,10). FCS has been applied in confocal microscopy for the accurate and quantitative measurement of association states and concentrations in the intracellular space (11,12) and for the monitoring of dynamic properties of membrane proteins (13,14). Fluorescence cross-correlation spectroscopy (FCCS) has been used for the detection of the association of two fluorescently labeled proteins. Dynamic and transient interactions are more conveniently detected by FCCS than by FRET (15). However useful the aforementioned correlation techniques are, they require dedicated instrumentation. The realization that a temporal confocal image series taken by a conventional confocal microscope contains all the necessary information to calculate the diffusion constant and the association state of the fluorescently labeled substance has lead to the development of raster image correlation spectroscopy (RICS) (16) and number and brightness analysis (N&B) (17). Both of these techniques