Red fluorescent protein-based cAMP indicator applicable to optogenetics and in vivo imaging

cAMP is a common second messenger that is involved in various physiological processes. To expand the colour palette of available cAMP indicators, we developed a red cAMP indicator named “Pink Flamindo” (Pink Fluorescent cAMP indicator). The fluorescence intensity of Pink Flamindo increases 4.2-fold in the presence of a saturating dose of cAMP, with excitation and emission peaks at 567 nm and 590 nm, respectively. Live-cell imaging revealed that Pink Flamindo is effective for monitoring the spatio-temporal dynamics of intracellular cAMP generated by photoactivated adenylyl cyclase in response to blue light, and in dual-colour imaging studies using a green Ca2+ indicator (G-GECO). Furthermore, we successfully monitored the elevation of cAMP levels in vivo in cerebral cortical astrocytes by two-photon imaging. We propose that Pink Flamindo will facilitate future in vivo, optogenetic studies of cell signalling and cAMP dynamics.


Results and Discussion
Development of a red cAMP indicator Pink Flamindo. Since cAMP functions in concert with various other intracellular signalling molecules, expansion of the colour palette of genetically-encoded FP-based cAMP indicators is a key requirement for studies aimed at delineating the interplay and/or hierarchy between cAMP and these molecules. The green FP-based cAMP indicator, Flamindo2, was originally constructed by fusing the green FP variant, Citrine, and the cAMP binding domain of the mouse exchange protein that is directly activated by cAMP 1 (Epac1, NP_ 01171281, 199-358 aa) to two linker peptides 9 . In this study, we inserted the same cAMP binding domain into a red FP variant, known as mApple, to create a red FP-based cAMP indicator 12 . We then produced numerous variants of the construct with different linker lengths and amino acid substitutions, in order to improve upon the dynamic range of the indicator. To achieve this aim, we followed the same method that we employed when previously developing a green cGMP indicator 13 . Interestingly, in the presence of cAMP, these variants showed increases or decreases of fluorescence intensity depending on the linker length ( Supplementary  Fig. S1). Of note, some variants with similar characteristics and structures of amino acids at the same residue showed similar responses ( Supplementary Fig. S2). We found that one variant elicited a notable change in fluorescence intensity in response to cAMP (Fig. 1), and named this genetically-encoded red FP-based cAMP indicator "Pink Flamindo" (Pink Fluorescent cAMP indicator).

Spectral properties of Pink Flamindo in vitro. To investigate the properties of Pink Flamindo in vitro,
we measured the fluorescence excitation and emission spectra of the purified Pink Flamindo protein. Pink Flamindo elicited an excitation peak at 567 nm and an emission peak at 590 nm (Fig. 2a). The fluorescence intensity of Pink Flamindo increased by 4.2-fold upon exposure to 100 μM cAMP, and the response was reversible ( Supplementary Fig. S3a). Additionally, the fluorescence intensity of Pink Flamindo reached a peak within 5 sec even at low cAMP concentrations, which is comparable to the previous FRET-based indicator using Epac ( Supplementary Fig. S3b) 5 . It is worth mentioning that the fluorescence intensity of Pink Flamindo is around 20-fold (without cAMP)/5-fold (with cAMP) lower than that of a red fluorescent protein derived from same origin, mCherry ( Supplementary Fig. S4). To improve brightness, incubation of Pink Flamindo-expressing cells at lower temperature is required for accelerating chromophore maturation ( Supplementary Fig. S5) 14 . The absorption spectra of 30 μM Pink Flamindo protein after the addition of cAMP decreased at 420 nm, and increased at 565 nm (Fig. 2b). mApple has an absorption peak at ~400 nm, which corresponds to the protonated form of the chromophore, and another peak at ~550 nm, which corresponds to the deprotonated form 15 . Given these data, it seems that Pink Flamindo changes the protonated/deprotonated ratio of its chromophore upon binding to cAMP, which leads to an increase in its fluorescence intensity.
We next examined the dose-response relationship of Pink Flamindo with cAMP or cGMP and calculated the dissociation constants (K d ) (Fig. 2c). According to the Hill equation, the cAMP K d = 7.2 μM and the cGMP and unbound (left) to cAMP. Images were created using structural graphics from mCherry (PDB_4ZIO), which has the same origin of gene as mApple and Epac1 (cAMP-unbound: PDB_2BYV, cAMP-bound: PDB_4MGK).
Both these values are slightly higher than those of Flamindo2 (cAMP K d = 3.2 μM and cGMP K d = 22 μM) 9 . Flamindo2 was generated from Flamindo by modification of only the linker length. As previously discussed, we developed Pink Flamindo by also introducing several point mutations into mEpac1, the linker region and the FP sequence. These mutations drastically improved the dynamic range of the indicator. Although the specificity of Pink Flamindo to cAMP was retained, a subtle reduction in the affinity for cAMP was noted. Because previous studies have developed various FRET-based cAMP indicators with K d between 50 nM and 10 μM 6, 16-19 and successfully monitored physiological cAMP changes, Pink Flamindo would be applicable for live-cell imaging analysis. Meanwhile, previous cAMP indicators with different affinities have shown different kinetics 5 , and thus careful choice of cAMP indicator with optimal affinity and kinetics is essential to elucidate the phenomenon of interest. The Hill coefficients for cAMP and cGMP were 1.01 and 1.07, respectively, suggesting that Pink Flamindo does not exhibit allosteric effects. As has been reported for many other single, FP-based indicators [20][21][22] , Pink Flamindo showed an increase in its fluorescent intensity as a result of pH elevation (Fig. 2d). We would, therefore, encourage careful interpretation or pH correction of the data obtained from imaging experiments using Pink Flamindo. 2 μM for cAMP and 94 μM for cGMP. The peak of FI in the absence of cAMP or cGMP was normalised to 0, and the peak of FI in the presence of 1 mM cAMP was normalised to 1. The data represent the means ± standard deviation (n = 3). (d) Titration curves of Pink Flamindo against pH in the presence (•, solid line) or absence (○, dashed line) of 100 μM cAMP. The peak FI for each pH was normalised to the peak of FI in the presence of cAMP at pH 9.0. The data represent the means ± standard deviation (n = 3).

Utility of Pink Flamindo in live-cell imaging and optogenetic studies. To investigate whether
Pink Flamindo is functional for studies of living cells, we sequentially applied adenylyl cyclase activator, forskolin (Fsk), phosphodiesterase inhibitor, 3-isobutyl 1-methylxanthine (IBMX), and adenylyl cyclase inhibitor 2′,5′-dideoxyadenosine (DDA) to HeLa cells overexpressing Pink Flamindo. The fluorescence intensity of Pink Flamindo first showed a transient increase in the presence of 100 μM Fsk, followed by a long-lasting elevation in response to 200 μM IBMX, and finally a decrease in the presence of 100 μM DDA ( Fig. 3a and Supplementary Movie S1). The application of DDA after exposure to Fsk and IBMX did not completely suppress the increase of fluorescence intensity. This would be ascribed to continuous suppression of phosphodiesterases by IBMX, because DDA only inhibits adenylyl cyclase. When stimulated with Fsk and IBMX, the fluorescence intensity of Pink Flamindo saturated at ~2.5-fold, which is lower than its dynamic range (4.2-fold) obtained from in vitro spectroscopic experiments (Fig. 1b). This difference may be explained by a certain basal level of cAMP in cells in a resting state 23 . Indeed application of 500 μM DDA to resting HeLa cells resulted in a decrease in the fluorescence intensity of Pink Flamindo ( Supplementary Fig. S6a). As described before, purified Pink Flamindo protein has pH sensitivity in in vitro spectroscopic experiments (Fig. 2d). Therefore, we attempted to correct the effect of intracellular pH on the responses of Pink Flamindo by co-expressing the fluorescent pH indicator pHluorin in HeLa cells for future necessity. When we corrected responses of Pink Flamindo by 100 μM Fsk according to the previous study 21 (also see Supplementary Methods), we confirmed that the intracellular pH changes can be cancelled by the correction ( Supplementary Fig. S6b). Thus, we should critically evaluate the effect of pH sensitivity under situations when intracellular pH is likely to change. To study whether Pink Flamindo monitors cAMP elevation induced by physiological stimulation as well as pharmacological stimulation, we imaged mouse MIN6 m9 pancreatic β cells under glucose application 24 . Pancreatic β cells take up glucose and secrete insulin, which is partly mediated by the activation of adenylyl cyclases and cAMP production 25 . When we applied 25 mM glucose to MIN6 m9 cells, we found that the fluorescence intensity of Pink Flamindo showed a long-lasting elevation (Fig. 3b, Supplementary Movie S2). Taken together, these results suggest that Pink Flamindo is successfully used to monitor the intracellular dynamics of cAMP respond in response to both pharmacological and physiological stimulations in living cells.
We next examined whether Pink Flamindo can be used in conjunction with optogenetic tools -namely, photoactivated adenylyl cyclase (PAC), which artificially induces cAMP production following blue-light exposure 11 . When HeLa cells co-expressing Pink Flamindo and photoactivated adenylyl cyclase from Beggiatoa sp. (bPAC) were stimulated with 1.8 μW blue light laser (488 nm), we observed a rapid increase in fluorescence intensity. On the other hand, blue light laser stimulation itself induced little increase in the cells expressing Pink Flamindo only (Fig. 3c, Supplementary Fig. S7, Supplementary Movie S3). The peak amplitude and duration of responses were dependent on the duration of the blue light laser stimulation. These results suggest that we can control the amount and duration of intracellular cAMP production by bPAC activation. Therefore, we propose that monitoring the tightly controlled cAMP levels with Pink Flamindo or other red fluorescent probes is an appropriate approach to evaluate the physiological function of cAMP in living cells.
Dual-colour imaging with Pink Flamindo and Ca 2+ indicator, G-GECO. Ca 2+ and cAMP are important signalling molecules that trigger hormone secretion in endocrine cells, including pancreatic β cells 25 . Here, we simultaneously monitored the dynamics of cAMP and Ca 2+ in MIN6 m9 cells, by co-expressing Pink Flamindo and the green Ca 2+ indicator G-GECO. Previous studies have shown that an increased level of ATP by glucose uptake closes ATP-sensitive K + channels (K ATP channels) in pancreatic β cells, which leads to Ca 2+ influx and insulin secretion 26 . Here, application of 200 μM tolbutamide, which artificially closes K ATP channels, induced an increase in the fluorescence intensity of both Pink Flamindo and G-GECO (Fig. 3d,e, Supplementary Movie S4). The fluorescence intensity of Pink Flamindo showed a single, sharp increase before gradually returning to baseline. Conversely, G-GECO exhibited long-lasting oscillations after reaching a peak, which is consistent with data from a previous observation 27 . We previously showed that extracellular Ca 2+ influx activates Ca 2+ -sensitive adenylyl cyclase in MIN6 cells 8 , and these latest results further support this concept. These data suggest that Pink Flamindo is suitable for dual-colour imaging in combination with green fluorescent indicators.
Application to in vivo two-photon imaging in cerebral cortical astrocytes. We next investigated whether Pink Flamindo could be applied to in vivo brain imaging studies of mice. We used a recombinant adeno-associated virus with the human GFAP promoter to specifically express Pink Flamindo in cerebral cortical astrocytes (Supplementary Fig. S8a). Immunohistochemical analyses showed that the expression of Pink Flamindo was cytosolic and co-localised with the astrocyte maker S100B (Supplementary Fig. S8b). We also found that Pink Flamindo was expressed in the somata, fine processes, and endfeet of astrocytes. Individual Pink Flamindo-expressing astrocytes were visualised by two-photon microscopy through an acutely-prepared cranial window made through the skull of mice that were maintained under urethane anaesthesia (Fig. 4a,b). Topical application of 50 μM Fsk and 500 μM IBMX resulted in an increase of fluorescence intensity of Pink Flamindo in the imaged astrocytes (Fig. 4c,d and e, Supplementary Movie S5). The fluorescence intensity of Pink Flamindo returned to baseline following wash-out of both compounds. The elevation of fluorescence intensity for Pink Flamindo could be repeated with a second application of Fsk and IBMX. The population mean increase in fluorescence intensity exhibited by the astrocyte somata, 10 minutes after the first exposure to Fsk and IBMX, was 55.5 ± 13.3% (n = 5 mice; data represent the means ± SEM). The analysis for the Flamindo2, turn-off type green fluorescent indicator, also showed recovery of fluorescence intensity after wash-out of the compounds, and yielded a signal decrease of 64.5 ± 10.9% (n = 5 mice; data represent the means ± SEM) ( Supplementary Fig. S8c,d). The opposing responses of Pink Flamindo and Flamindo2 support the proper response of these cAMP indicators in vivo. The magnitude of the change in fluorescence was comparable between Pink Flamindo and Flamindo2, with a similar dynamic range (4.2 and 4.0 fold, respectively, p = 0.62) and dissociation constant (7.2 μM and 3.2 μM, respectively). These data are consistent with our previous spectroscopic results obtained in vitro.
ScienTific RepoRts | 7: 7351 | DOI:10.1038/s41598-017-07820-6 Astrocytes are the most abundant glial cells in the cerebral cortex and they express various G protein-coupled receptors (GPCRs) for neuromodulators, such as acetylcholine and noradrenaline. These neuromodulators are associated with brain states, such as vigilance and startle, and influence wide brain areas by volume transmission and activating astrocytic GPCRs 28 . Given these data, we propose that Pink Flamindo will serve as a potent tool to investigate physiological Gs-GPCR or Gi-GPCR activation in astrocytes and other brain cells, especially in combination with multi-colour imaging and optogenetics.

Conclusion
This report describes the generation of the first red FP-based cAMP indicator, named Pink Flamindo. Spectroscopic analyses showed that the fluorescence intensity of Pink Flamindo increases with the cAMP concentration in a dose-response manner. The development of red FP cAMP indicator allowed dual cAMP/Ca 2+ imaging as well as cAMP imaging in conjunction with optogenetic stimulation by blue light. These results demonstrate that Pink Flamindo overcomes the limited utility of FRET-based or green FP-based indicators due to the spectral overlap with other fluorescence indicators or optogenetic molecules. Moreover, we were able to apply Pink Flamindo to in vivo two-photon imaging of cerebral cortical astrocytes. This application expands the utility of this indicator to elucidate the complex interplay and hierarchy between different signalling molecules at a high spatiotemporal resolution in living animals. The use of our novel indicator in optogenetic studies and in vivo imaging will open new avenues for investigation of the complex signalling networks between various molecules in cells, tissues, organs and whole animals.
Plasmid construction. The DNA fragment of a red fluorescent protein variant, mApple, utilized in a red Ca 2+ indicator R-GECO 12 was created by DNA synthesis from Integrated DNA Technologies (Coralville, IA, USA). It was engineered to contain SacII and EcoRI restriction enzyme sites between A150 and V151 and cloned into the pRSET-A vector. The cAMP binding domain of mEpac1 (NM_001171281, 199-358 amino acids) derived from Flamindo2 and was inserted into the SacII/EcoRI sites of mApple. To improve the performance of the indicator, a number of variants were created with the addition or deletion of linker amino acids in the both N-terminus and C-terminus of the cAMP binding domain by PCR. Random, site-directed mutations were also introduced into the construct by PCR with sense and anti-sense primers including NNK and MNN, respectively. The mutant construct that elicited the highest dynamic range was selected and named Pink Flamindo. Pink Flamindo was cloned into the pcDNA3.1 (-) vector for expression in mammalian cells. The humanised photoactivated adenylyl cyclase from Beggiatoa sp. (h_bPAC) in a pGEM-HE vector was obtained from Addgene (#28134, Cambridge, MA, USA), and was subcloned into the BglII/EcoRI sites of the pEGFP-C1 vector. The G-GECO plasmid was also obtained from Addgene (#32447). The Flamindo2 plasmid was developed in our previous study 9 . Protein expression and in vitro spectroscopy. For protein expression, Pink Flamindo or mCherry in the pRSET-A vector was transformed into Escherichia coli JM109 (DE3) cells, and the cells were cultured for 4 days at 20 °C and harvested by centrifugation. The harvested cells were suspended in phosphate-buffered saline (PBS) and lysed by three freeze-thaw cycles and sonication with 40 μg/mL lysozyme. After centrifugation, the supernatants containing the Pink Flamindo protein were collected and purified on an Ni-NTA agarose column (QIAGEN, KJ Venlo, Netherlands) followed by clean up through a PD-10 gel filtration column (GE Healthcare) to remove imidazole and elution in HEPES buffer (150 mM KCl and 50 mM HEPES-KOH [pH 7.4]). To generate pH titration curves, purified Pink Flamindo protein was diluted in HEPES buffer (100 mM HEPES-KOH [pH 5 to 9]). The absorption spectra of purified Pink Flamindo protein were measured using an UV spectrophotometer (UV-1800, Shimadzu, Kyoto, Japan), and the fluorescence spectra were measured using a fluorescence spectrophotometer (F-2500, Hitachi, Tokyo, Japan).
Cell culture and transfection. HeLa and MIN6 m9 cells were cultured in high glucose Dulbecco's modified Eagle's medium (Sigma-Aldrich) supplemented with 4.5 g/L glucose, L-glutamine, sodium pyruvate, 10% (v/v) heat-inactivated foetal bovine serum (Sigma-Aldrich), 100 U/mL penicillin and 100 μg/mL streptomycin (Sigma-Aldrich), at 37 °C in an atmosphere of 5% CO 2 . The media to culture MIN6 m9 cells also contained 50 μM 2-mercaptoethanol 24 . For imaging experiments, the cells were dissociated with trypsin and then plated onto glass coverslips coated with poly-L-lysine (Sigma-Aldrich) placed within 35-mm dishes. The cells were transfected with plasmids using Polyethyleneimine "MAX" (Polysciences, Warrington, PA, USA) or Lipofectamine 2000 Transfection Reagent (Thermo Fisher Scientific, Waltham, MA, USA), 2 days after plating. The media was exchanged 4 h after transfection and the cells were cultured at 32 °C for 2or 3 days until imaging.
Visualisation of intracellular cAMP dynamics. HeLa and MIN6 m9 cells cultured in 35-mm dishes were washed twice and imaged in modified Ringer's buffer (RB: 140 mM NaCl, 3.5 mM KCl, 0.5 mM NaH 2 PO 4 , 0.5 mM MgSO 4 , 1.5 mM CaCl 2 , 10 mM HEPES, 2 mM NaHCO 3 , 5 mM glucose for HeLa and 3 mM glucose for MIN6 m9 cells, respectively). MIN6 m9 cells were cultured in the medium containing 1 g/L glucose for overnight, and incubated in RB for 30 min at 37 °C before imaging. Dishes were mounted on a stage that was heated and maintained at 37 °C and imaging was performed using an inverted microscope (IX-71, Olympus, Tokyo, Japan) equipped with an oil-immersion objective lens (UApo/340, 40×, NA = 1.35, Olympus), an intermediate magnification lens (1.6×), and an EM-CCD camera (Evolve, Photometrics, Tucson, AZ, USA). Images were acquired using a xenon lamp, 545-580 nm excitation filter, 585 nm dichroic mirror and 610 nm emission filter (Olympus). The exposure time of the EM-CCD camera was controlled by MetaMorph software (Molecular Devices, Sunnyvale, CA, USA). Images were acquired every 5 sec for 30 min. Data analysis of the acquired images was performed using MetaMorph software and the fluorescence intensity of the cells was quantified. After subtracting the background, basal fluorescence intensity, normalised to 100%, was calculated as the average fluorescence intensity between 90 and 120 sec after the beginning of image acquisition.
ScienTific RepoRts | 7: 7351 | DOI:10.1038/s41598-017-07820-6 Surgical procedures for virus inoculation. The procedures involving animal care, surgery and sample preparation were approved by the Animal Experimental Committee of RIKEN Brain Science Institute and performed in accordance with the guidelines of the Animal Experimental Committee of RIKEN Brain Science Institute (H27-2-230). The adeno-associated virus (AAV) pAAV-hGFAP-Pink Flamindo and pAAV-hGFAP-Flamindo2 constructs were synthesised by replacing the promoter and coding region of the pAAV-hSyn-EGFP vector (Addgene plasmid #50465, a gift from Bryan Roth) with the human GFAP promoter 29 and Pink Flamindo or Flamindo2 sequence; the latter was inserted between the AgeI and NotI sites. AAV9-hGFAP-Pink Flamindo and AAV-hGFAP-Flamindo2 was purified at a titer of 6.6 × 10 13 vg/mL and 1.1 × 10 14 vg/mL, respectively. The AAV was diluted with PBS for microinjection, to 6.6 × 10 12 vg/mL and 1.1 × 10 13 vg/mL for Pink Flamindo and Flamindo2, respectively. For microinjection, mice were initially anesthetised by intraperitoneal injection containing a ketamine cocktail (ketamine 70 mg/kg, xylazine 10 mg/kg), and then fixed in a stereotaxic frame. The mice were then maintained under an anaesthetic of 1% isoflurane for the duration of the experiment. A small craniotomy was made and a glass micropipette containing AAV was inserted to the primary somatosensory cortex at a 20° angle. Microinjection of 500 nL AAV was made over 5 min using a FemtoJet injector (Eppendorf, Hamburg, Germany) at depths of 350 μm and 500 μm below the surface. In vivo imaging experiments were performed 2 weeks after viral inoculation. Immunohistochemistry for Pink Flamindo and S100B was performed using anti-DsRed (1:1000, Clontech #632496, Shiga, Japan) and anti-S100B (1:1000, #S2532, Sigma-Aldrich) antibodies.
Cranial window preparation and in vivo two-photon imaging. On the day of imaging, mice were anesthetised with urethane by intraperitoneal injection (1.5 g/kg) and the body temperature was maintained at 37 °C with a heating pad (BWT-100A, Bio Research Centre, Nagoya, Japan, or TR-200, Fine Science Tools, Foster City, CA, USA) during the course of the surgery and recordings. The skull was exposed by making an incision to the scalp, and a metal frame was then attached to the skull using a dental acrylic (Fuji LUTE BC, GC Corporation, Tokyo, Japan, and Super Bond C&B, Sun Medical, Shiga, Japan). A craniotomy (3 mm in diameter) was made above the virus injection site and the dura mater was surgically removed. After filling the craniotomy with artificial cerebrospinal fluid (aCSF), ~50% of the craniotomy area was gently sealed with a thin glass coverslip (3 mm × 3 mm, thickness: 0.12 mm, Matsunami Glass, Osaka, Japan), leaving an open area for drug application. The cranial window was fixed with dental cement.
In vivo imaging was performed using a resonant scanner-based two-photon microscope (B-Scope, Thorlabs, Newton, NJ, USA) with a Chameleon Ultra 2 laser (Coherent, wavelength 1040 nm and 960 nm for Pink Flamindo and Flamindo2, respectively) and an objective lens (XLPlan N 25×, NA = 1.05, Olympus). The laser power under the objective lens was ~30 mW. The B-Scope was equipped with a reverse dichroic mirror (ZT405/488/561/680-1100rpc, Chroma, Bellows Falls, VT, USA) and the emission light was separated using a dichroic mirror (FF562-Di03, Semrock, Rochester, NY, USA) with bandpass filters FF03-525/50 and FF01-607/70 (both from Semrock) for the green and red channels, respectively. Twenty consecutive images were acquired at a frame rate of 1 frame/sec, where each image was an internally averaged image of 30 frames within a second. Image acquisition was performed every 2 min using ThorImage software (Thorlabs) and the 20 images were averaged to represent the time point.
For data analysis, an offset value, calculated as the mean of 100 lowest pixel values, was subtracted from each frame. Astrocyte somata were marked manually using ImageJ software (National Institutes of Health, Bethesda, MD, USA), and the relative intensity change (ΔF/F, fold change) was calculated from (F max − F 0 )/F 0 for Pink Flamindo, and from (F 0 − F min )/F min for Flamindo2.