Fluorescent and bioluminescent calcium indicators with tuneable colors and affinities

We introduce a family of bright, rhodamine-based calcium indicators with tuneable affinities and colors. The indicators can be specifically localized to different cellular compartments and are compatible with both fluorescence and bioluminescence readouts through conjugation to HaloTag fusion proteins. Importantly, their increase in fluorescence upon localization enables no-wash live-cell imaging, which greatly facilitates their use in biological assays. Applications as fluorescent indicators in rat hippocampal neurons include the detection of single action potentials and of calcium fluxes in the endoplasmic reticulum (ER). Applications as bioluminescent indicators include the recording of the pharmacological modulation of nuclear calcium in high-throughput-compatible assays. The versatility and remarkable ease of use of these indicators make them powerful tools for bioimaging and bioassays. Graphical abstract

compatible assays. The versatility and remarkable ease of use of these indicators make them powerful tools for bioimaging and bioassays.

Introduction
The second messenger calcium is involved in a plethora of signaling pathways and biochemical processes. 1 The elucidation of its function in cellular processes has become possible largely through the development of calcium indicators. [2][3][4] While early development focused on synthetic calcium indicators, genetically encoded calcium indicators (GECIs) have now become the gold standard. The main reason for this is that GECIs can be genetically targeted to specific cellular populations and subcellular localizations, whereas the cellular uptake of synthetic calcium indicators lacks selectivity and is often inefficient. However, GECIs possess lower brightness, slower response kinetics and a limited color range (especially in the far-red in comparison to synthetic indicators). 5,6 These limitations are of particular concern when highly localized areas, such as micro-and even nanodomains are investigated, and more demanding microscopy techniques are used. [7][8][9] A possibility to combine the brightness, response kinetics and spectral range of synthetic fluorescent indicators with the targetability of GECIs is the use of self-labeling protein tags such as SNAP-tag and HaloTag. 10,11 Self-labeling proteins form a covalent bond to a specific substrate and through this enable precise localization of synthetic molecules to proteins of interest (POI). This approach has been used to create a number of localizable synthetic calcium indicators, e.g. BG3-Indo-1, 12 BOCA-1-BG 13 or RhoCa-Halo 14 and the far-red indicator JF646-BAPTA. 5,15 However, these probes have limited cell permeability and solubility, and furthermore require washing steps to remove unreacted probes, greatly limiting their applicability. 13,14 The use of bright synthetic fluorophores for calcium sensing was enabled developing chemogenetic sensors in which the protein-based calcium-sensing domain Calmodulin (CaM) interacts with an environmentally sensitive dye (e.g. rHCaMP or HaloCaMP). 16,17 However, based on the same calcium sensing domain as most GECIs, they suffer from relatively slow response kinetics. 16 Furthermore, there is currently no localizable synthetic far-red calcium indicator with a suitable calcium affinity for calcium-rich areas like the endoplasmic reticulum (ER) or calcium microdomains. 18,19 Here we present MaPCa dyes, a family of highly permeable calcium indicators with different colors and calcium affinities that can be coupled to HaloTag. As the reaction with HaloTag shifts the fluorescent scaffold of the indicator from a non-fluorescent into a fluorescent configuration, these probes can be used without any washing steps to remove unbound probe.

Design principle, synthesis and in vitro characterization of MaPCa dyes
The design of our calcium indicators is based on the recently introduced MaP dyes, in which the lactone-forming carboxylic acid of a rhodamine is replaced with an amide attached to an electron withdrawing group (e.g. sulfonamides). 20,21 This results in dyes that preferentially exist as a nonfluorescent spirolactam in solution, but shift to an open, fluorescent state upon binding to HaloTag, enabling no-wash imaging with low background. We envisioned designing fluorogenic calcium indicators by attaching a calcium chelator such as BAPTA (1,2-bis(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid) through a benzene sulfonamide to the ortho-carboxylate of rhodamines and a chloroalkane (CA) through a carboxylate at the 6-position of the benzyl-ring (Fig. 1). BAPTA would be thereby positioned in close proximity to the rhodamine core, which is an important factor for effective PET-quenching of the rhodamine by the free chelator. 15,22 Attachment of the CA via the 6-position of the benzyl-ring would enable HaloTag to shift the equilibrium from spirocyclization to the fluorescent, open form, thereby resulting in fluorogenicity. Furthermore, attachment of the CA via the 6-position would ensure a high labeling speed of the resulting HaloTag substrate. 23 We set out to synthesize a set of such indicators based on the high-affinity calcium chelator BAPTA and the low-affinity chelator MOBHA (2-(2′-morpholino-2′-oxoethoxy)-N,Nbis(hydroxycarbonylmethyl) aniline) 24 in combination with commercially available rhodamine-CA substrates TMR-CA, CPY-CA and SiR-CA, covering the spectrum from 550 to 650 nm (Fig. 2). In a first step, a sulfonamine was attached to the previously described BAPTA-ethylester 25 (01) or MOBHA-ethylester (02) via chlorosulfonation followed by amination (03,04). These two intermediates were then coupled to the commercially available rhodamine-CAs TMR-CA, CPY-CA and SiR-CA using activation by chlorosulfonic acid. The indicators were obtained as free acids after saponification with KOH ( Fig. 2a). For in cellulo experiments, acetoxymethyl (AM) esters of the indicators were synthesized by prior transesterification of the chelator (05,06) and subsequent coupling to the fluorophore. The AM-esters serve to mask the carboxylic acids to ensure cellpermeability, but are cleaved inside the cell by endogenous esterases. 26 We named these indicators MaPCa dyes (for Max-Planck-Calcium sensor), with a postfix expressing the absorption maxima in nm (TMR = 558; CPY = 619; SiR = 656) and the subscripts 'high' or 'low' for indicating the calcium affinity range. The AM-esters of the dyes are marked with an additional AM, in contrast to the saponified probes. It should be noted that this short and convergent synthetic scheme should enable the conversion of most rhodamine-CAs into calcium sensors in a single step. MaPCa-619high and MaPCa-656high showed a significant 7-fold and even 120-fold increase upon binding to HaloTag, respectively, in the calcium-bound state. The higher fluorogenicity of MaPCa-656high can be rationalized considering the higher propensity of SiR derivatives to exist in the nonfluorescent spirocyclic form than the corresponding rhodamine and carborhodamine derivatives. 27 The dyes possess a high brightness in the calcium-bound state (quantum yield >40%; extinction coefficient >80'000 M -1 cm -1 ) and display calcium-affinities in a suitable range for cytosolic measurements (KD(Ca 2+ ): 410-580 nM) with turn-ons of around 6-fold upon calcium binding ( HaloTag binding, while MaPCa-619low (28-fold) and MaPCa-656low (208-fold) are highly fluorogenic. The calcium affinities of these dyes are in the range of 220-460 µM and they show a 7 to 11-fold turn-on upon calcium binding (Table 1, Supporting Table 1). The extinction coefficient of MaPCa-656low is significantly lower than those of the other MaPCa-indicators, suggesting that it does not fully convert to the open state. Nevertheless, its brightness of ~15 mM -1 cm -1 is in the same order of magnitude than genetically encoded red-shifted indicators (brightness FR-GECO1c: 9.3 mM -1 cm -1 ). 28  ( Fig. 3e, Supporting Fig. 6). and MaPCa-656high AM (Fig. 4a, Supporting Fig. 7). To test the sensitivity of the high affinity MaPCa indicators, labeled neurons were stimulated with a distinct number of action potentials (APs) using electric field stimulation. 30 All dyes allowed the detection of a single AP with F/F0 values ranging between 3% (MaPCa-558high AM) and 6% (MaPCa-656high AM), while F/F0 of 120% was obtained using MaPCa-656high AM with a 160 AP burst (Fig. 4b, Supporting Fig. 8, The lower calcium affinity of the MaPCalow series allows to report calcium fluctuations in compartments with high basal calcium concentrations such as the ER (Ca 2+ conc.: ~ 500 µM ). 19 Therefore, the MaPCa dyes were target to the ER through rat hippocampal neuron transduction localizing a HaloTag-SNAP-tag fusion in the ER. Co-staining of SNAP-tag confirmed efficient and specific labeling of HaloTag with MaPCalow dyes under no-wash conditions, with the exception of MaPCa-558low AM that required a washing step to reduce background (Supporting Fig. 9). The ER is a calcium store which, upon stimulation, can release calcium into the cytosol. Here, the RyR2 channel plays a crucial role as a calcium-induced calcium-release channel. 31

Bioluminescence as a readout
The MaPCa dyes could potentially also be used for the labeling of H-Luc, a chimera between HaloTag and the furimazine-dependent luciferase NanoLuc. 33 Labeling of H-Luc with rhodamine dyes can result in efficient BRET from NanoLuc to the bound rhodamine, such that emission at both 450 nm and at the emission wavelength of the bound rhodamine can be observed. We hypothesized that labeling H-Luc with MaPCa dyes would lead to the development of bioluminescent calcium indicators with tunable emission wavelengths with up to far-red light emission (Fig. 5a). Existing bioluminescent calcium indicators, such as Orange CAMBI, 34 GLICO, 35 LUCI-GECO1, 36 CeNL 37 or CalfluxVTN 38 rely exclusively on fluorescent proteins that possess emission maxima restricted below 600 nm. We therefore labeled H-Luc with the MaPCa dyes and recorded the emitted light upon addition of furimazine in the absence and presence of calcium. As is already apparent by eye (Fig. 5b), the color of the emitted light dramatically depends on both, the presence of calcium as well as the nature of the MaPCa dye attached to H-Luc (Fig.   5c). The efficiency of BRET is largest for MaPCa-558high attached to H-Luc, as it has the largest spectral overlap with the BRET donor. As the intensity of the emission of the MaPCa dye depends on the concentration of calcium, measuring the ratio of the intensity of emitted light at 450 nm versus the intensity of the light emitted at the emission maximum of the rhodamine dye can thus be used to record changes in calcium concentrations (Fig. 5b, c, Supporting Fig. 11). The maximal change in ratio ranged from 6. MaPCa-558high AM (1.7-fold) and the smallest for MaPCa-656high AM (1.3-fold) (Fig. 5d). In such given experimental conditions, each channel luminescence intensity was integrated in less than 500 ms, allowing changes in calcium concentrations to be followed with good temporal resolution.
The z-factor is a measure for the statistical effect size used to judge the suitability of an assay for high-throughput screening (HTS) approaches. 293-cells expressing H-Luc labeled with MaPCa-558high AM presented a z-value of 0.58 upon ATP/thapsigargin treatment, highlighting the suitability of such bioassay for HTS (z-factors ≥ 0.5 indicate excellent suitability). 39 Finally, lowaffinity bioluminescent calcium indicators could be generated by labeling H-Luc with the MaPCalow indicators, demonstrating the modularity of the approach (Supporting Fig. 12).