In vitro FRET imaging of cAMP signaling
The PDE4DIP NM_001002811 transcript was cloned into an mCherry-tagged plasmid and a point mutation, resulting in an alanine to threonine substitution at amino acid 123, was induced using site directed mutagenesis and verified by sequencing. Human embryonic kidney (HEK) cell line 293T was transfected with the PDE4DIP mutant and control plasmids using the Lipofectamine transfection kit. Transfection efficiency was verified by percentage m-cherry expression on fluorescent microscopy (>50%). The cells were then co-transfected with the FRET sensor plasmids using the Lipofectamine kit and incubated on 35mm microwell dishes with 14mm glass coverslip for 3 days to achieve 60-70% confluency.
Intracellular cAMP imaging was done using FRET sensors and confocal microscopy at baseline, and at 4 minutes after stimulation with 1uM isoproterenol (cAMP activator). Further imaging was done 4 minutes after stimulation with 25 uM of Forskolin (strong cAMP activator) under live confocal microscopy. A fourth generation Epac-based FRET sensor (mTurquoise2Δ-Epac (CD, ΔDEP, Q270E)-tdcp173Venus (Epac-SH187)) was used as previously described (Klarenbeek, Goedhart, van Batenburg, Groenewald, & Jalink, 2015). This sensor consists of a full-length cAMP-binding Rap-1 activating protein Epac, which is sandwiched between the donor and acceptor fluorescent proteins with larger conformational change, and FRET change compared to sensors with partial Epac. Fourth generation of cAMP sensors have superior photo-stability, dynamic range and signal-to-noise ratios. The design is based on mTurquoise2 as a bleaching-resistant donor, and a tandem of two cp173Venus fluorophores as acceptors.
FRET was quantified using 3 cube ratiometric imaging with analysis in MATLAB using previously developed custom software (A. Kumar et al., 2016). Images were acquired on an inverted Nikon Eclipse Ti widefield microscope equipped with a cooled charged-coupled device Cool SNAP HQ2 camera using a 20x 0.75 NA objective at 37°C. Three sequential images were acquired with the following filter combinations: donor (mTurquoise2) channel with 460/20 (excitation filter-ex), T455lp (dichroic mirror-di) and 500/22 (emission filter-em); FRET channel with 460/20 (ex), T455lp (di) and 435/30 (em); and acceptor (Venus) channel with 492/18 (ex), T515lp (di) and 535/30 (em) filter combinations. For analysis, donor leakage was determined from 293T cells transfected with mTurquoise2 and acceptor cross excitation was obtained from Vinculin-Venus transfected cells. All three FRET images (mTurquoise2, Venus, FRET) were background subtracted and filtered by three-point smoothening. FRET maps and pixel-wise FRET index was calculated as:
FRET index = [FRET channel – x(Donor channel) – y(Acceptor channel)]/[Acceptor channel]
Where x is the leakage co-efficient and y is the cross-excitation fraction. Masks for each cell were generated by thresholding mCherry positive cells using the mTurquoise images. Mean FRET index per cell was calculated for each field under each treatment condition.