Lanthanide‐based FRET biosensors for time‐gated imaging and detection of protein‐protein interactions in live mammalian cells

Ting Chen,Ha Pham,Lawrence Miller
DOI: https://doi.org/10.1096/fasebj.31.1_supplement.767.6
2017-04-01
The FASEB Journal
Abstract:Förster resonance energy transfer (FRET) using fluorescent proteins (FPs) as donors and acceptors is commonly used to microscopically image protein‐protein interactions in living cells. However, overlapping FP spectra lower signal‐to‐noise because the FRET‐sensitized acceptor emission signal is convoluted with donor bleedthrough and directly excited acceptor emission, or crosstalk. Quantitative FRET imaging is even more difficult when the local concentrations of donor‐labeled and acceptor labeled proteins are substantially different. We used the well‐studied interaction between FK binding protein 12 (FKBP12) and the rapamycin binding domain of m‐Tor (FRB) as a model system to show that lanthanide‐based biosensors can overcome the limitations of conventional FP‐FRET. The essential features of this approach include the use of luminescent Tb(III) complexes as FRET donors and the elimination of non‐specific fluorescence background that comes from time‐gated detection of long‐lived (~ms) Tb(III) and Tb(III)‐to‐FP emission signals. We prepared stably transformed NIH 3T3 fibroblast lines that express a single polypeptide consisting of the following elements (N to C): 1) FRB; 2) Escherichia coli dihydrofolate reductase (eDHFR); 3) a linker sequence; 4) GFP; and 5) FKBP12. A Tb(III) complex (TMP‐Lumi4) that selectively binds to eDHFR completes the sensor. The linker comprises alternating repeats of four glutamic acid residues followed by four arginine or lysine residues (ER/K) that form a rigid alpha helix in solution. Stochastic breaks in the helix permit close approach and binding of affinity elements positioned at each end. The effective concentration of each binding element depends on linker length and not solution concentration. Consequently, the fraction of two elements in the closed (high‐FRET) state depends on KD and linker length. We tested constructs with ER/K linker lengths of 10, 20 and 30 nm. Following addition of cell permeable TMP‐Lumi4 and rapamycin to culture medium, two‐color, time‐gated images showed significant, ~3‐fold increases in acceptor denominated FRET ratios relative to baseline (without rapamycin). The high sensor dynamic range permitted robust, time‐gated detection (Z'‐factor > 0.8) of the FKBP12/FRB interaction and its inhibition by ascomycin in cells grown in 96‐well plates. Support or Funding Information NIGMS R01 081030
cell biology,biochemistry & molecular biology,biology
What problem does this paper attempt to address?