Improved genetically-encoded, FlincG-type fluorescent biosensors for neural cGMP imaging

Abstract

Genetically-encoded biosensors are powerful tools for understanding cellular signal transduction mechanisms. In aiming to investigate cGMP signaling in neurones using the EGFP-based fluorescent biosensor, FlincG (fluorescent indicator for cGMP), we encountered weak or non-existent fluorescence after attempted transfection with plasmid DNA, even in HEK293T cells. Adenoviral infection of HEK293T cells with FlincG, however, had previously proved successful. Both constructs were found to harbor a mutation in the EGFP domain and had a tail of 17 amino acids at the C-terminus that differed from the published sequence. These discrepancies were systematically examined, together with mutations found beneficial for the related GCaMP family of Ca(2+) biosensors, in a HEK293T cell line stably expressing both nitric oxide (NO)-activated guanylyl cyclase and phosphodiesterase-5. Restoring the mutated amino acid improved basal fluorescence whereas additional restoration of the correct C-terminal tail resulted in poor cGMP sensing as assessed by superfusion of either 8-bromo-cGMP or NO. Ultimately, two improved FlincGs were identified: one (FlincG2) had the divergent tail and gave moderate basal fluorescence and cGMP response amplitude and the other (FlincG3) had the correct tail, a GCaMP-like mutation in the EGFP region and an N-terminal tag, and was superior in both respects. All variants tested were strongly influenced by pH over the physiological range, in common with other EGFP-based biosensors. Purified FlincG3 protein exhibited a lower cGMP affinity (0.89 μM) than reported for the original FlincG (0.17 μM) but retained rapid kinetics and a 230-fold selectivity over cAMP. Successful expression of FlincG2 or FlincG3 in differentiated N1E-115 neuroblastoma cells and in primary cultures of hippocampal and dorsal root ganglion cells commends them for real-time imaging of cGMP dynamics in neural (and other) cells, and in their subcellular specializations.

Keywords: C-type natriuretic peptide; cGMP; dorsal root ganglion; genetically-encoded biosensor; hippocampus; neuroblastoma; nitric oxide.

Figure 1

Screening of FlincG variants. (A) Schematic diagram illustrating the general FlincG design and specifying the modifications tested. The diagram also defines the nomenclature adopted in text. (B) Brightfield (top) and basal fluorescence (bottom) images of HEK_GC/PDE5 cells 3 days after attempted transfection with identical amounts (0.5 μg/well) of AdV-FlincG plasmid DNA (left) and FGB DNA (right). Scale bar = 100 μm. (C) Sample FACS spectra of suspensions of HEK_GC/PDE5 cells transfected with the indicated FlincG variants, before (black) and after (red) incubation with 8-bromo-cGMP (1 mM). Calibration bars: vertical = frequency; horizontal = log intensity (arbitrary units). Arrows and arrowheads indicate untransfected and transfected components of the control spectra, respectively. (D–F) Results of FACS analysis for groups of FlincG variants; plasmids in panel D all bore a strong Kozak consensus sequence (see Material and Methods). Numbers in parentheses are n-values.

Figure 2

Responsiveness to NO of HEK_GC/PDE5 cells transfected with FlincG variants. (A–E) Representative recordings (from 3 to 4 experiments) of HEK_GC/PDE5 cells transfected with the indicated FlincG variants (grouped as in Figure 1) and exposed to 1 nM NO for 1 min (horizontal bars). Traces are averages of 5–20 cells; vertical bars are ± SEM. Calibrations (C, bottom) apply to all panels. (F) Sample brightfield images (“BF”, left), together with the baseline fluorescence (“Control”, center) and maximal NO-induced change in fluorescence (“Peak”, right), coded into the grayscale shown on the extreme right, of cells transfected with H6-FGA^M (top) and FG^MT (bottom); scale bar (in the top left image) = 50 μm.

Figure 3

Cyclic nucleotide sensitivity of H6-FGA^M and H6-FGB proteins and NO-sensitivity of HEK_GC/PDE5 cells expressing selected FlincG variants. (A,B) Sample excitation (left) and emission (right) spectra of H6-FGB (A) and H6-FGA^M (B) proteins in the absence (black) and presence of 1 μM (red) and 100 μM (blue) cGMP, normalized to the peaks in the absence of cGMP. (C) Cyclic nucleotide concentration-response curves for the two proteins. Points are means of three determinations ± SEM. (D) NO concentration-response curves in HEK_GC/PDE5 cells transfected with the indicated FlincG variants (n = 3–5). (E) Representative recordings of responses to NO contributing to panel D. NO concentrations (specified at bottom) were superfused for the time indicated by the horizontal bars. Traces are means of 11–24 cells; the calibration (top left) applies to all of them. The inset illustrates the heterogeneity in the response profile amongst cells in a given experiment often (but not always) observed at the higher NO concentrations. Also at the higher concentrations, responses were sometimes followed by small, rapid undershoots and small, slow secondary fluorescence increases (or decreases in the case of FG^MT, bottom trace), as was observed previously with adenoviral FlincG-infected HEK_GC/PDE5 cells (Batchelor et al., 2010).

Figure 4

Comparison between H6-FGA^M and FGB as sensors of NO puffs when expressed in HEK_GC/PDE5 cells. (A,B) Representative traces of fluorescence changes to NO puffs of different durations (black traces); the NO applications are quantified from the Texas Red dye intensity (red traces), with the applied durations specified underneath. There was a 5-min interval between applications and each result is from a single cell. Calibration (in B) applies to both sets of data. (C) Mean responses (± SEM) to the different puff durations in four (H6-FGA^M) or seven (FGB) cells in three experiments. (D) Images showing the spread of Texas Red from the puffer pipette (top) within a field of cells (brightfield, middle), two of which are fluorescent (bottom); the arrowed cell is the source of the traces in (B); scale bar (middle) = 50 μm.

Figure 5

Sensitivity of FlincG variants to pH. (A) Responses of HEK_GC/PDE5 cells expressing FGB to 2-min applications of NH₄Cl (horizontal bar). Traces are means from 40 cells; vertical bars are ± SEM. (B) HEK_GC/PDE5 cells expressing FGB were initially stimulated with NO (1 nM; 1 min), after which the combination of CCCP (20 μM) and nigericin (10 μM), which results in an equalization of intracellular and extracellular pH, was superfused at different pH values (as indicated by the stepped horizontal bars below). The trace is the average of 10 cells; vertical bars are ± SEM. (C) Summary pH titration data obtained as in panel (B) on HEK_GC/PDE5 cells expressing different FlincGs or EGFP (n = 3). (D) Sample excitation (Ex) and emission (Em) spectra of purified H6-FGA^M and H6-FGB proteins in the absence (solid lines) or presence (dashed lines) of 100 μM cGMP, at pH 6.5 (red), 7.5 (green) and 8.5 (blue), normalized to the peak at pH 8.5 in the absence of cGMP. (E,F) pH-titration data for the fluorescence (arbitrary units) of the purified FlincG proteins in the absence and presence of 100 μM cGMP (left-hand ordinate), and the dynamic ranges, expressed as ΔF/F₀, at each pH (black squares; right-hand ordinate).

Figure 6

Expression of FGB in differentiated N1E-115 neuroblastoma cells (A–E) and rat hippocampal cell cultures (F,G). (A) Brightfield (left) and fluorescent (right) images of neuroblastoma cells (two transfected, two untransfected); scale bar = 50 μm. (B) Response to increasing clamped NO concentrations applied at the horizontal bars (5 cells). (C) Response to three consecutive 1 nM NO applications followed by 3 nM NO (10 cells). (D) Inhibition of the response to 1 nM NO by ODQ (8 cells). (E) Response to acetylcholine (ACh) in the absence and presence of L-nitroarginine (L-NNA) with application of a high concentration of PAPA/NO at the end (4 cells). Blue lines in B-E are means; gray lines = SEM. (F) Brightfield (left) and fluorescent (right) images of a rat hippocampal cell culture. Cells numbered 1, 2, and 3 are a putative oligodendrocyte, neurone and astrocyte, respectively, based on their morphology. Scale bar = 50 μm. (G) Responses of these 3 cells to superfusion of 8-bromo-cGMP.

Figure 7

Expression of H6-FGA^M in DRG cultures. (A) (a) Brightfield image of a typical culture showing neurones of different sizes and an underlay of flatter, glial-like cells. (b,c) Fluorescence images of cultures after attempted transfection with AdV-FlincG plasmid DNA (b) and H6-FGA^M (c) and exposure to 1 mM 8-bromo-cGMP for 10 min. Scale bar (a) = 50 μm. (B) Time-courses of fluorescent changes in the individual cells arrowed in (Ab) and (Ac) during superfusion of 8-bromo-cGMP. (C) High-power brightfield (a) and fluorescence (b,c) images of a transfected neurone before (b) and after (c) superfusion of 8-bromo-cGMP (1 mM) for 8 min. The images in (b) and (c) were processed identically. Scale bar (a) = 25 μm. (D) Time-course of the changes in fluorescence in the number-coded puncta and cell soma demarcated in red in (c). Relatively large regions of interest were used for the puncta to accommodate their movement during the experiment. This procedure should not have affected the signal amplitude but may have increased the noise. (E–G) Neuronal (E,G) and glial-like (F) cells transfected with H6-FGA^M and exposed to PAPA/NO (E) or PAPA/NO followed 20–30 min later by CNP (F,G). Each panel (E–G) shows pairs of brightfield (a) and fluorescent (b) images; scale bars = 25 μm. Traces (c) are from the regions outlined in red in (b); black horizontal bars = periods of agonist application; in (E) only one punctum (at the top) was analyzed because the others, despite obviously responding, were too motile; calibrations in (Fc) also apply to the traces in (Ec).