Fourth-generation epac-based FRET sensors for cAMP feature exceptional brightness, photostability and dynamic range: characterization of dedicated sensors for FLIM, for ratiometry and with high affinity

Abstract

Epac-based FRET sensors have been widely used for the detection of cAMP concentrations in living cells. Originally developed by us as well as others, we have since then reported several important optimizations that make these sensors favourite among many cell biologists. We here report cloning and characterization of our fourth generation of cAMP sensors, which feature outstanding photostability, dynamic range and signal-to-noise ratio. The design is based on mTurquoise2, currently the brightest and most bleaching-resistant donor, and a new acceptor cassette that consists of a tandem of two cp173Venus fluorophores. We also report variants with a single point mutation, Q270E, in the Epac moiety, which decreases the dissociation constant of cAMP from 9.5 to 4 μM, and thus increases the affinity ~ 2.5-fold. Finally, we also prepared and characterized dedicated variants with non-emitting (dark) acceptors for single-wavelength FLIM acquisition that display an exceptional near-doubling of fluorescence lifetime upon saturation of cAMP levels. We believe this generation of cAMP outperforms all other sensors and therefore recommend these sensors for all future studies.

Fig 1. Characterization of novel FRET sensors.

A. Localization of the FRET-sensors: N1E-115 cells transfected with mTurquoise2Δ-Epac(CD, ΔDEP)-^cp173Venus-Venus (Epac-S^H126); mTurquoise2Δ-Epac(CD, ΔDEP, Q270E)-td^cp173Venus (Epac-S^H187); mTurquoise2Δ-Epac(CD, ΔDEP)-td^cp173Dark Venus (Epac-S^H159) and mTurquoise2Δ-Epac(CD, ΔDEP)-^cp174Cit (Epac-S^H147). Images were taken 18 hours after transfection. B. FRET efficiency of constructs with different donors. Emission spectra of U2OS cells expressing constructs with donors Cerulaen3 (Epac-S^H105); mTurquoise (Epac-S^H74) or mTurquoise2 (Epac-S^H126) were acquired while exciting at 436 nm. Spectra are normalized to CFP intensity after correction for expression levels (using Venus brightness, excited at 500 nm as detailed in M&M). C. Spectra of constructs with various acceptors. Shown are spectra from N1E-115 cells expressing Epac with acceptor ^cp173Venus-Venus (Epac-S^H134); td^cp173Venus (Epac-S^H187) or td^cp173Dark Venus (Epac-S^H189), illuminated with a 442 nm laser. Emission spectra were normalized to isosbestic point at 505 nm. D. Calibration curves for normal and high-affinity sensors. Shown are the average of three independent calibrations performed on cell lysates of HEK293T cells expressing the indicated constructs. cAMP was titrated in under continuous stirring. The sensors were excited at 420 +/- 3 nm and emission was measured at 530 +/- 10 nm and 490 +/- 10 nm for YFP and CFP respectively. Ratios were calculated as YFP over CFP and were normalized between baseline 0% and maximum response 100%. E. In-vivo experiment revealing the difference between high- and normal affinity sensors (Epac-S^H126 or Epac-S^H134) expressed in Hek293T cells. Isoproterenol induces a graded increase in cAMP levels in these cells and was added in increasing amounts as indicated. Signals were normalized between baseline 0% and maximum response 100%. Representative experiment out of 3 repeats. F. Typical FRET time-lapse trace in N1E-115 cells expressing Epac-S^H187. After recording a baseline, at t = 90 s PGE1 (5 μM) was added, and at t = 250 s IBMX (100 μM) and Forskolin (25 μM) were added for calibration. G. A FLIM-FRET time-lapse recording from N1E-115 cells expressing Epac-S^H189. A baseline was followed by addition of IBMX (100 μM) and Forskolin (25 μM) after 140 seconds.

Fig 2. Summary of FRET changes of the constructs mentioned in this study.

Mentioned are construct details (column 1–3), unique Lab ID (column 4), % ratio change (expressed as cAMP-induced change from an initial baseline ratio of 1, column 5) with its Standard Error of Mean, FRET efficiency E in rest (column 6) with SEM, E after saturation of the sensor with IBMX and Forskolin with SEM (column 7), absolute change of lifetime in ns with SEM (column 8) and % change in lifetime with SEM (column 9). Column 10 contains the fully descriptive name. As mentioned in the main text, we will from now on for brevity name the sensors "Epac-S^LabID", for example, Epac-S^H96 for mECFPΔ_Epac(CD, ΔDEP)_cp¹⁷³Ven_Ven. In gray is the reference construct Epac-S^H74. Abbreviations: nd, not determined; mCer3, monomeric Cerulean 3; mECFP, monomeric enhanced CFP; mTurq, monomeric Turquoise fluorescent protein; Hi-aff, High-affinity (Q270E mutant); ^cp173V, circular permutation of the fluorescent protein Venus(etc); td ^cp173V (etc), tandem of two ^cp173V fluorophores; ^cp173VV, tandem of cpVenus and Venus; CD, catalytically dead mutant (T781A & F782A amino acids of the Epac1 wild type protein); ΔDEP, deletion of the DEP domain to prevent membrane localization. Note that unlike Venus, Citrine contained the A206K monomerizing mutation. FRET efficiency E was calculated as E = 1-τ_{DonorAcceptor}/τ_Donor using τ_Donor = 4.10 ns for mTurquoise2 and 3.70 ns for mTurquoise1.