CaRuby-Nano: a novel high affinity calcium probe for dual color imaging

The great demand for long-wavelength and high signal-to-noise Ca2+ indicators has led us to develop CaRuby-Nano, a new functionalizable red calcium indicator with nanomolar affinity for use in cell biology and neuroscience research. In addition, we generated CaRuby-Nano dextran conjugates and an AM-ester variant for bulk loading of tissue. We tested the new indicator using in vitro and in vivo experiments demonstrating the high sensitivity of CaRuby-Nano as well as its power in dual color imaging experiments. DOI: http://dx.doi.org/10.7554/eLife.05808.001


Introduction
In recent years fluorescence imaging has been one of the fastest growing methods in physiology, cell biology and neuroscience, constantly driving the need for improved fluorescent probes (Wilt et al., 2009;Miyawaki, 2011;Looger and Griesbeck, 2012). The dominance of fluorescein and eGFP in the design of such probes has resulted in an overcrowding of the green spectral band. This makes simultaneous imaging of spatially overlapping signals problematic and emphasizes the need for red-shifted probes . Many of the very favorable photophysical properties of fluorescein and eGFP are shared by the X-Rhodamine chromophore, which is finding increasing use in the development of Ca 2+ -indicators (Eberhard and Erne, 1991;Egawa et al., 2011). A major problem of red-shifted fluorophores is that they are significantly more lipophilic than fluorescein-like dyes. This leads to more leakage through cell membranes as well as to intracellular compartmentalization. These effects can be minimized by using the probes as conjugates of inert hydrophilic compounds such as dextrans. This conjugation commonly uses one of the carboxylic groups of the BAPTA moiety, affecting the calcium affinity of BAPTA-based indicators (Tsien, 1980). Consequently, these indicator-dextran conjugates are strongly shifted to lower affinities making them all but useless for sensitive and quantitative [Ca 2+ ] measurements. Importantly, this lower affinity cannot be compensated for by increasing the concentration of the probe. Such a strategy would lead to a disproportionally large increase in the calcium buffering capacity of the indicator (Neher, 2005), resulting in a stronger disruption of cellular signaling than for a low concentration of high affinity indicator (Markram et al., 1998).
We recently introduced a family of red emitting calcium indicators based on X-Rhodamine: Calcium Ruby (CaRuby) (Collot et al., 2012), which bears an azido side arm for click chemistry and the resulting potential for high-yield coupling reactions (Kolb et al., 2001). This side arm efficiently allows conjugation reactions without significant perturbation of the calcium binding affinity . The dissociation constants of CaRubies ranged from 3.4 to 21.6 μM-too high for the reliable detection of small [Ca 2+ ] transients in biological tissue (Yasuda et al., 2004) (see Appendix 1 for details).

Results
To increase the affinity of CaRuby, we modified the structure of the probe ( Figure 1A), focusing on the Ca 2+ chelating BAPTA moiety, as increasing the electron density of BAPTA lowers its K D for calcium (Tsien, 1980). We introduced an oxygen atom on one of the aromatic rings of BAPTA by a S N Ar reaction. This oxygen also serves as a link for the azido side arm, which was repositioned in the new CaRuby variant (Figure 1-figure supplement 1). Additionally, the fluorophore, which is commonly placed para to the nitrogen of the BAPTA, has an affinity-lowering effect due to its electron withdrawing nature and was therefore placed at a meta position in order to reduce its effect on the ligating nitrogen. These modifications resulted in a CaRuby variant with sub-micromolar affinity ('CaRuby-Nano'). In cuvette calibration experiments CaRuby-Nano was found to have a K D of 258 ± 8 nM, with a 50-fold (±2) increase of fluorescence on binding [Ca 2+ ] ( Figure 1B-C) and a maximum quantum yield of 0.45 (Figure 1-figure supplement 2,3). In addition to being suitable for single photon excitation, CaRuby-Nano also exhibits effective two-photon excitation over a large wavelength band ( Figure 1-figure supplement 4).
For verification of the new probe in biological tissue we used conjugates with 1.5 kD and 6 kD dextrans, which were obtained via click chemistry. As expected, this conjugation had only a small eLife digest The movement of calcium ions within cells controls many vital biological processes, ranging from cell growth to muscle contraction and brain activity. These calcium signals are triggered by stimuli, such as nerve impulses, which drive calcium entry into cells or release calcium from internal stores. These changes in calcium levels can span several orders of magnitude, and can be either localized to very small parts of the cell or span the entire cell.
Scientists have developed numerous indicators or 'probes' that can detect even very low levels of calcium. One common method uses proteins that fluoresce when viewed under a fluorescence microscope each time the protein senses increases of calcium. Most of these probes fluoresce green, and so to view a second signal that occurs in the cell at the same time it's easier to use a probe that fluoresces with a different color, such as red. However, the red-shifted probes that are currently available either produce unreliable results because they tend to leak through cell membranes, or are not very sensitive to calcium ions. New types of red-shifted probes are therefore urgently needed.
In 2012, researchers developed a family of red fluorescent probes known as Calcium Ruby (CaRuby for short) that were more versatile than earlier red probes. Now, Collot, Wilms et al.-including several of the researchers involved in the 2012 research-have enhanced the properties of CaRuby by modifying the chemical structure of the probes. This increased the ability of CaRuby to bind calcium ions, making it more sensitive to small calcium changes. Testing the usefulness of the newly developed probes-called CaRuby Nano-in mouse nerve cells revealed the probes are highly sensitive and can even detect the calcium signal resulting from a single nerve impulse.
Collot, Wilms et al. then went on to demonstrate that CaRuby-Nano can be used alongside a green-fluorescing probe to record two signals at the same time. In one experiment, the release of chemical messengers known as neurotransmitters was stimulated, which caused calcium ions to flow into the observed nerve cells. The researchers succeeded in simultaneously detecting a green signal indicating an increase in neurotransmitter levels and a red signal produced by the corresponding release of calcium. Such dual-color imaging was not possible with previous probes. Finally, it was shown that CaRuby-Nano can also be used to produce dual-color images of the brain activity of live mice.
In summary, these results demonstrate that CaRuby-Nano is a highly sensitive and versatile indicator and can be used together with other probes to observe two simultaneous events in cells.
effect on the affinity of the indicator, increasing the K D from 258 nM in the free salt to 295 nM in the 6 kD dextran conjugate (Figure 1-figure supplement 5).
We first tested if CaRuby-Nano performs comparably to commonly used green emitting [Ca 2+ ] probes. For this, Purkinje cells in acute cerebellar brain slices were filled with CaRuby-Nano (1.5 kD dextran) via a patch-clamp microelectrode ( Figure 1D). Climbing fiber stimulation evoked calcium signals were recorded from multiple spines (range: 4 to 14) at acquisition rates between 2.2 and 5.0 kHz ( Figure 1E,F) using random access two-photon microscopy (Otsu et al., 2008). The rising phase time course (0.55 ms ± 0.13 ms; sigmoidal fit; n = 59 spines from 7 cells) was not significantly different from that found for Fluo-5F (0.40 ± 0.09 ms, n = 36 spines from 4 cells, p = 0.37) under the same conditions, suggesting that CaRuby-Nano has binding kinetics comparable to established small molecule Ca 2+ indicators. These fast kinetics point to a high  sensitivity of CaRuby-Nano for small and fast changes in [Ca 2+ ], such as neuronal action potentials (Otsu et al., 2014).
Thus, we next tested the sensitivity of CaRuby-Nano using in vivo patch-clamp recordings from neocortical layer 2/3 pyramidal neurons in anesthetized mice with simultaneous two-photon [Ca 2+ ] imaging (Svoboda et al., 1997) (Figure 2A). We found that using CaRuby-Nano (6 kD dextran) even single spikes resulted in reliable, easily detected fluorescence transients (mean dR/R 0 = 0.52 ± 0.19, n = 6 cells; Figure 2B). For increasing spike numbers the dR/R 0 vs spike number relation quickly turns sublinear and saturates as expected for high affinity indicators ( Figure 2C). Taken together these experiments demonstrate that CaRuby-Nano is a calcium indicator with a signal quality comparable to previously used high-affinity green emitting probes. Importantly, it is well suited for the detection of small [Ca 2+ ] transients, setting it apart from the previous CaRuby versions.
Having verified the suitability of CaRuby-Nano for single cell imaging experiments in vitro and in vivo, we now set out to test CaRuby-Nano for imaging neuronal network activity when applying sensory stimulation. In the past decade calcium population imaging has commonly been performed using bulk loading (Stosiek et al., 2003;Ohki et al., 2005) of calcium indicators in the AM-ester form (Tsien, 1981). We thus synthesized an AM-ester of CaRuby-Nano and used it to load cerebellar neurons in vivo ( Figure 3A). We performed a series of three experiments. In all cases we found labeling identical to that commonly found in experiments using Oregon Green-488 BAPTA-1 AM (OGB-1 AM) to load cerebellar tissue in vivo ( Figure 3B) (Sullivan et al., 2005;Ozden et al., 2009;Schultz et al., 2009). In all experiments fluorescence traces extracted from identified Purkinje cell dendrites ( Figure 3C) showed clear complex spike activity with a good signal-to-noise ratio ( Figure 3D,F). Both spontaneous activity and sensory evoked responses were again comparable to signals detected in experiments using OGB-1 AM (Sullivan et al., 2005;Ozden et al., 2009;Schultz et al., 2009). These results indicate that CaRuby-Nano AM is a powerful addition to the optophysiological toolbox.
To demonstrate the full power of CaRuby-Nano, we made use of the strong two-photon excitation spectral overlap with eGFP to conduct a set of experiments which were not possible previously: simultaneous imaging of glutamate release onto Purkinje cells using iGluSnFR (a single-wavelength extracellular glutamate indicator constructed from the bacterial glutamate sensor Gltl and circularly permutated GFP [Marvin et al., 2013]) and the resulting post-synaptic [Ca 2+ ] increase (using CaRuby-Nano). Visually identified Purkinje cells showing iGluSnFR expression (7-9 days after viral transfection) were filled with CaRuby-Nano via patch-clamp recording (6 kD dextran; Figure 4A). Activation of the glutamatergic climbing fiber input evoked clear fluorescence transients in both color channels ( Figure 4B). Glutamate signals were confined to distinct subsections of the dendritic tree (i.e., limited to sites of synaptic glutamate release), whereas the resulting [Ca 2+ ] transients were global, with similar amplitudes throughout different regions of the dendritic tree ( Figure 4C) (Lev-Ram et al., 1992). The differential spatial distribution of the signals confirms that the two indicators can be spectrally isolated.
To demonstrate that dual color imaging is also possible in vivo we used CaRuby-Nano (6 kD dextran) to report presynaptic activity in anesthetized Kv3.1-eYFP adult mice (Metzger et al., 2002). In the olfactory bulb of these mice, mitral and tufted cells, as well as a population of periglomerular neurons, strongly express eYFP and their somata and processes clearly demarcate the external glomerular boundaries ( Figure 4D). Olfactory sensory neuron (OSN) terminals, labeled with CaRuby-Nano, filled the inner glomerular boundaries ( Figure 4E). In single glomeruli (n = 8 animals) we could record presynaptic calcium responses with excellent signal to noise ratio. Figure 4F shows a typical example in which presynaptic calcium responses were selectively evoked by odor presentation in a subset of glomeruli. These responses adapted strongly at this high odorant concentration, as reported previously (Lecoq et al., 2009). Taken together, these last two experiments demonstrate the potential of two-channel functional imaging-both in vitro and in vivo-with the red emission and high sensitivity  (Metzger et al, 2002). (E) Olfactory sensory neuron glutamatergic terminals, labeled with CaRuby-Nano dextran, clearly filled the inner boundaries of most glomeruli. (F) A 3 s application of 30% isoamyl acetate reliably triggered presynaptic calcium responses in several glomeruli. DOI: 10.7554/eLife.05808.011 of CaRuby-Nano being an ideal match for numerous other indicators emitting in the green-yellow spectral band.

Discussion
We have developed a high-affinity red-emitting calcium indicator. This novel indicator, CaRuby-Nano, has a K D of 295 nM (for the dextran conjugate). This makes CaRuby-Nano only slightly higher affinity than the commonly used green emitting indicator Fluo-4 (335 nM). On calcium binding CaRuby-Nano shows a 50-fold fluorescence increase (vs. 14-fold and 100-fold for OGB-1 and Fluo-4, respectively). The quantum efficiency of calcium bound CaRuby-Nano (0.45) is lower than that of OGB-1 (∼0.7) but significantly higher than that of Fluo-4 (∼0.14). These values classify CaRuby-Nano as an ideal indicator for the quantification of small intracellular [Ca 2+ ] transients (see Appendix 1). Using a range of different experiments we have demonstrated that CaRuby-Nano is well suited for both in vitro and in vivo imaging experiments requiring high sensitivity to [Ca 2+ ] changes. Finally, we show that CaRuby-Nano can be combined with activity indicators emitting in the green-yellow spectral band, to allow multiplexed imaging. The versatility of the probe is further increased by the azido function, which can easily be reduced to an amine group, thus opening the field to functionalization with numerous molecular tools such as antibodies, benzylguanine (SNAP tag) or peptides to facilitate specific sub-cellular targeting.
CaRuby-Nano's sensitivity and potential for spectral multiplexing allows sophisticated experiments such as simultaneously measuring pre-and postsynaptic activity or imaging of different signaling modalities in the same cell, allowing previously elusive questions to be directly addressed.

General chemical methods
All the solvents were of analytical grade. Chemicals were purchased from commercial sources. 1 H-NMR and 13 C-NMR were measured on a Bruker Avance III-300 MHz spectrometer (Bruker Biospin, The Woodlands, TX, USA) with chemical shifts reported in ppm (TMS as internal standard). Mass spectra were measured on a Focus GC/DSQ II spectrometer (ThermoScientific, Waltham, MA, USA) for IC and an API 3000 spectrometer (Applied Biosystems, PE Sciex) for ES. All pH measurements were made with a Mettler Toledo pH meter. Fluorescence spectra were recorded on a JASCO FP-8300 spectrofluorometer (JASCO, Easton, MD, USA). Absorption spectra were determined on a VARIAN CARY 300 Bio UV-Visible spectrophotometer. All measurements were done at a temperature of 25˚C. The purity of the dyes were checked by RP-HPLC C-18, elutant: ACN 0.1% TFA/Water 0.1% TFA, method: 20/80 to 100/ 0 within 20 min then 100/0 for 10 min detection at λ Abs = 254 nm. The apparent dissociation constant for calcium (K D Ca 2+ ) was measured with a calcium calibration buffer kit from Invitrogen (Lifetechnologies, USA). All mass spectra, NMR spectra and chromatograms are included as supplemental data.

Design of CaRuby-Nano from the first generation CaRubies
In order to develop a high affinity CaRuby, three modifications were carried out based on the first generation CaRubies. First, an oxygen atom was introduced on one of the BAPTA's cycles in order to electronically enrich the latter. Then, this oxygen atom served as an anchor to a spacer terminated by an azide function for further functionalizations either by click chemistry or by reducing it into an amine for coupling with for example, a carboxylic acid. Finally, the fluorophore moiety, an extended rhodamine which is positively charged and therefore has an electron withdrawing effect, was moved from the para position of the aniline to the meta position. As expected, these modifications lead to a significant increase of affinity towards calcium, yielding a CaRuby with a dissociation constant of 258 ± 8 nM.

Synthesis of CaRuby-Nano
The synthesis pathway is displayed in Figure 5 along with the compound numbering. The NMR and mass spectra for both intermediate compounds and final products are contained in Supplementary file 1.
To a solution of de 5-fluoro-2-nitrophenol (14.90 g, 94.84 mmol) in DMF (75 ml) were added dibromoethane (40.90 ml, 472.2 mmol, 5 eq) and K 2 CO 3 (26.30 g, 189.7 mmol, 2 eq), the mixture was allowed to stir at 70˚C for 2 hr. The solvents were evaporated and the product was extracted with EtOAc washed with water (three times) and brine (two times). The organic phase was dried over MgSO 4 , filtered and evaporated to reach a volume of 200 ml. The symmetric dinitro compound crystallizes first and was filtered off. The filtrate was then allowed to crystallize to obtain 20.12 g of 1 (80%) as a yellow powder.  To a solution of 1 (19.79 g, 74.96 mmol) in DMF (75 ml) were added 2-nitrophenol (11.46 g, 82.45 mmol, 1.1 eq) and K 2 CO 3 (15.63 g, 112.4 mmol, 1.5 eq), the mixture was allowed to stir overnight at 70˚C. The solvent was evaporated and the product was extracted with DCM, washed with HCl (1 M) and brine (2 times). The organic phase was dried over MgSO 4 , filtered and evaporated to reach a volume of 200 ml. The product crystallized and was filtered to obtain 12.00 g of 2 (50%) as a yellow powder. 1 H-NMR (300 MHz, DMSO-d6): δ 8.01 (dd, J a-b = 9.1 Hz, J a-F = 6.1 Hz, 1H, H a ), 7.86 (dd, J g-f = 8.1 Hz, J g-e = 1.6 Hz, 1H, H g ), 7.67 (ddd, 3 J = 8.5, 7.4, 4 J e-g = 1. To a stirred solution of 2 (5.91 g, 18.34 mmol) in DMSO (53 ml) was added NaOH 20% (11.5 ml) the solution turned yellow and was allowed to stir at room temperature overnight. 50 ml of water and 10 ml HCl (1 M) were then added and the product was extracted three times with EtOAc. The organic phase was washed three times with water before being dried over MgSO 4 , the solution was filtered and evaporated and crystallized in EtOAc to obtain 4.46 g of 3 (76%) as a yellow powder. To a solution of 3 (4.86 g, 15.19 mmol) in DMF (50 ml) were added dibromohexane (11.12 ml, 45.56 mmol, 3 eq) and K 2 CO 3 (3.16 g, 22.78 mmol, 1.5 eq). The mixture was allowed to stir at 70˚C for 12 hr. The solvents were evaporated and the product was extracted with EtOAc washed with water (three times) and brine (two times). The organic phase was dried over MgSO 4 , filtered and evaporated. The crude was purified by column chromatography on silica gel (Cyclohexane/EtOAc: 7/3) to obtain the crude 4 which was crystallized in a mixture of EtOAc and cyclohexane (3/7) to obtain 2.97 g of pure 4 (40%) as a off white powder. Rf To a solution of 4 (5.00 g, 10.35 mmol) in EtOAc (100 ml) and methanol (30 ml) was added Pd/C (1.10 g). The solution was stirred and degassed before H 2 was allowed to bubble in the solution for 5 hr. The solution was then filtered off celite and rinsed with EtOAc under an atmosphere of argon. The solvents were evaporated and the crude was dissolved in acetonitrile (50 ml), to this solution were added, methyl bromoacetate (12.0 ml, 124.2 mmol, 12 eq) and DIEA (23.0 ml, 124.2 mmol, 12 eq) before being warmed up to 80˚C. The solution was allowed to stir overnight at 80˚C. The solvents were evaporated, the product was extracted with DCM and washed with water. The organic phase was dried over MgSO 4 , filtered and evaporated. The crude was purified by column chromatography on silica gel (Cyclohexane/EtOAc: 7/3) to obtain 3.71 g of 5 (50%) as a yellowish syrup containing impurities (between 2 and 3 ppm in 1 H NMR) that could not be removed. Rf = 0.51 (Cyclohexane/EtOAc, 6/4). (d, J = 7.4 Hz, 12H, 4 OMe), 3.36 (t, J = 6.8 Hz, 2H, CH 2 Br), 1.85-1.80 (m, 2H, CH 2 ), 1.71-1.67 (m, 2H, CH 2 ), 1.42 (t, J = 3.6 Hz, 4H, 2 CH 2 ). MS (ES+), calcd for C 32 H 43 BrN 2 O 11 Na [M + Na] + 735.2, found 735.8.

Animals
All procedures were approved by the local ethical review committee and performed under license from the UK Home Office in accordance with the Animals (Scientific Procedures) Act 1986, and in accordance with the Institut National de la Santé et de la Recherche Médicale (INSERM) Animal Care and Use Committee Guidelines and with Centre National de la Recherche Scientifique (CNRS) animal experimentation guidelines and European laws and policies, as applicable.

Electrophysiology and imaging in cerebellum and neocortex
Full frame and linescan two-photon imaging was performed using microscopes optimized for in vitro (Prairie Technologies, now Bruker Nano Surfaces, USA) or in vivo (MOM, Sutter, Novato, CA, USA) experiments. Two photon excitation was provided by a pulsed Ti:Sa laser (MaiTai HP, Spectra-Physics, Santa Clara, CA, USA), tuned to a central wavelength of 890-920 nm. The microscopes were controlled by ScanImage 3.5 and 3.7.1 (Pologruto et al., 2003) (now Vidrio Technologies, Arlington, VA, USA).
Climbing fiber stimulation-evoked transient [Ca 2+ ] changes in Purkinje cell spines were recorded at high acquisition rate (>2 kHz) by two-photon random-access microscopy, a technique based on the use of acousto-optic deflectors (AODs), which enable selective scanning of defined points (Otsu et al., 2008). Purkinje cells were recorded in current-clamp mode, using 2-3 MΩ patch pipettes containing 300 μM CaRuby-Nano dextran. Recordings were obtained by use of a Multiclamp 700B (Molecular Devices). Following the dialysis of CaRuby-Nano, Purkinje cells in slices were imaged under a 25× Leica water immersion objective (HCX IRAPO L 25×/0.95, Leica Microsystems, Wetzlar, Germany). Two-photon excitation was produced by a pulsed Ti:Sa laser (Chameleon Vision Plus, Coherent, Santa Clara, CA, USA) coupled into the transmitted light pathway of the microscope by a dichroic filter (740dcsx, Chroma) and tuned to a central wavelength of 890 nm. A custom-made user interface based on National Instrument cards programmed under Labview (both National Instruments, Austin, TX, USA) was used to operate the AODs and coordinate the scanning protocols and signal acquisition. A multifunction card (NI-PCI-MIO 16 E-4) was used to pass all the triggers necessary to synchronize the imaging and the electrophysiology and to control the piezo-electric device that moves the objective in Z. Fluorescence photons were detected by cooled AsGaP photomultipliers (H7421-40, Hamamatsu, Hamamatsu, Japan) discriminated and counted on a fast digital card.

Virus injection
Young (P19) C57BL6/J mice were anesthetized using isoflurane, an incision was made into the scalp and a small (∼0.5 mm) craniotomy was performed over lobule V of the cerebellar vermis. A wide bore (∼50 μm) micropipette containing viral suspension (AAV1.hSyn.iGluSnFr.WPRE.SV40, University of Pennsylvania Vector Core) was inserted through the craniotomy and carefully lowered 1.0 mm into the brain. Using application of low pressure 400-800 nl viral suspension were slowly injected (10-20 min). After the injection further 5-10 min were waited before retraction of the injection pipette. The scalp was glued and sutured and the mouse left to recover. At least 7 days incubation time were allowed prior to further experiments.

In vivo bulk loading and imaging
Adult C57BL6 mice (6-9 weeks) were anesthetized with isoflurane, supplemented with 1 mg/kg chlorprothixene. A 1.5-2 mm craniotomy was performed over cerebellar lobule V. Care was taken to leave the dura mater intact. CaRuby-Nano-AM was prepared and injected using standard methods (Stosiek et al., 2003;Sullivan et al., 2005). A 50 μg aliquot was dissolved in 20% Pluronic-127 in DMSO (Invitrogen) and then diluted 1:10 in saline (150 mM NaCl, 2.5 mM KCl, 10 mM HEPES, pH 7.4). This solution was filtered and injected into the cerebellum under visual guidance using a patch-pipette and 500-750 mbar pressure for 1-3 min. After injection the preparation was left to incubate for up to 1.5 hr prior to imaging. This helped improve labeling and lower unspecific fluorescence.

Data analysis and statistics
Imaging data were analyzed using ImageJ (http://rsbweb.nih.gov/ij/). Extracted fluorescence traces, linescans and electrophysiological data were analyzed using in house routines programmed in Igor Pro versions 5 or 6.2 (Wavemetrics) and in pClamp 10 (Molecular Devices). Statistical analysis was performed in Matlab (MathWorks, Natick, MA, USA) or Igor Pro (Wavemetrics, Portland, OR, USA). Experimental groups were compared using a t-test and were assumed to be significantly different if the found p-values were <0.05. a single action potential in CA1 pyramidal neurons [Maravall et al., 2000]) whereas CaRuby-Nano would require transients under 100 nM.
Taken together these results show that while under bright conditions low-affinity indicators have the advantage of larger signal amplitudes, under the dimmer conditions often found in biological imaging, lower affinity dyes will often be working at the edge of the detectability. Thus, for quantification of small, localized (i.e., where it is not possible to average many pixels) Ca 2+ signals, high affinity indicators such as OGB-1 and CaRuby-Nano are needed.