Microscopy in combination with fluorescence labeling techniques has enabled us to look at diverse biological processes. By tagging a protein of interest, we learn about its intracellular localization and dynamics. Fluorescence resonance energy transfer (FRET) sensors, which carry two fluorescent tags instead of one, a donor and an acceptor, are generated to obtain information about protein function. While the use of multiple labels is straightforward and common in high content imaging and screening, the use of multiple FRET sensors is more difficult. Here we highlight how FRET can be multiplexed and emphasize the potential of molecular translocation as an underlying principle for the development of protein-protein interaction assays suitable for high-content applications.

Development of FRET Sensors

FRET is a radiationless energy transfer between two fluorophores in close proximity [1]. The maximal distance at which FRET is detectable depends on the fluorophores used. When most typical genetically encoded fluorophores, such as cyan and yellow fluorescent proteins, are employed, that value is about 5 nm. Consequently, the method enables probing molecular interactions as well as intramolecular changes in protein conformation. Intramolecular sensors are the most common kind of genetically encoded tools based on FRET and are frequently generated to monitor enzyme activity and changes in protein conformation in living cells [2].
Although the principal design of genetically encoded sensors is usually straightforward, the performance of FRET reporters is not predictable since it mostly depends on small changes in distance and dipole moment of the fluorophores. Therefore this approach often requires significant optimization [3]. An alternative are reporter molecules based on small molecules produced at the chemistry bench [4]. These reporters often exhibit much larger FRET changes but are usually limited to enzymes that hydrolyze a substrate leading to a complete loss of FRET (fig. 1). Small molecule-based reporters are often used for biochemical analysis, but are also chemically modified to enter cells [5-7].

Furthermore, a larger repertoire of fluorophores is available to produce efficient FRET pairs. The largest obstacle is that the synthesis of membrane-permeant variants of small molecule FRET reporters is mostly a tedious and time demanding process. Therefore, small molecule-based FRET reporters are rarely used in high-content screening.

FRET Sensors and High-content Screening

While imaging of FRET changes of a single FRET sensor is usually straightforward, monitoring of more than one sensor in a single cell is challenging [8]. But imaging multiple FRET sensors in parallel is crucial when small temporal differences of the underlying processes, e.g. activation of different enzymes, are investigated. One way towards this goal is to make spectrally non-overlapping and compatible FRET sensors. For instance, when using genetically encoded FRET sensors, CFP and YFP may be used as one FRET pair, while mOrange and mCherry form the second [9].
More opportunities for using multiple reporters are provided by using two CFP-YFP sensors in single cell imaging experiments despite the fact that they consist of exactly the same fluorophores (fig. 2) [9]. The trick is to guide the sensors to different cellular locations. For instance, a plasma membrane-anchored sensor can be generated and co-expressed with a sensor of e.g. cytoplasmic localization. Information about their activation is then extracted from the images by precisely analyzing signals from the cytoplasm and the plasma membrane. Another sensor that has yet different localization may be added, e.g. a reporter for a transcription factor localized in the nucleus. We predict that with further development of fluorescent proteins suitable for FRET imaging, the quality of sensors will increase and their multiplexing using spectral and spatial resolution will become a standard application for comparative analysis of protein activity and conformational dynamics. Additionally, as FRET sensors can be used to screen small molecule libraries [10], multiplexing sensors will strongly increase the information content of screens and enable comparative evaluation of drug effects.

Translocation as Basis for PPI
Assays in Living Cells

Although we consider FRET an excellent approach for the generation of ratiometric enzyme activity sensors, as a method for studying protein-protein interactions (PPI) it is less robust. As PPI are becoming increasingly interesting as drug targets, novel methods that address them are highly desirable. We would like to emphasize the practicability of translocation-based methods for studying binary interactions and multi-component complexes in living cells. One such method recently developed in our lab exploits the property of a protein called annexin A4 to translocate to the plasma and nuclear membranes in the presence of high intracellular calcium levels (fig. 3) [11]. When a bait protein is fused to the annexin, the protein construct translocates jointly. Fluorescently labeled target proteins or peptides that interact with the bait have to follow. Since as many as four (or more) fluorescent proteins can be distinguished using modern microscopes, the method enables multiplexing and the analysis of protein complexes in living cells. For instance, interaction of three different targets with the bait can simultaneously be investigated, which is a clear advantage over other cell-based methods that mostly permit analysis of binary interactions. This concept could easily be adapted to screening applications where interactions of multiple partners with baits are investigated across different cell types and experimental conditions, following different stimuli, or in presence of small-molecule PPI inhibitors.