To design iGECInano, we used a Ca²⁺-sensing domain from Twitch-2B (Thestrup et al., 2014) developed from a minimal Ca²⁺-binding domain of troponin C (TnC_m). Unlike Ca²⁺-sensing domains in iGECI and NIR-GECO2G that are based on calmodulin, TnC_m does not require a binding partner peptide and only has two Ca²⁺-binding motifs, called EF-hands. Therefore, indicators with TnC_m should have simpler and likely faster conformational changes associated with Ca²⁺-binding and lower calcium buffering in cells. To build a FRET pair, we chose miRFP720 FP as the FRET-acceptor and miRFP670nano FP (Oliinyk et al., 2019) as the FRET-donor. miRFP670nano is a single domain NIR FP characterized by fast maturation time and a twice smaller size compared to all other currently available NIR FPs. To generate an initial library of NIR GECI variants, we fused miRFP670nano and miRFP720 to TnC_m directly or via randomized linkers of 1–4 amino acid residues in length.

An initial screening of NIR GECI variants was performed similarly to the process previously described (Shemetov et al., 2021) (Figure 1A). In short, we transformed the library of NIR iGECI variants into BL21 (DE3) bacterial strain expressing heme oxygenase, collected the brightest cells using a fluorescence-activated cell sorter (FACS), plated them on Petri dishes, and then transferred the mature colonies onto a nitrocellulose membrane. The colonies were further screened for functional response by measuring the baseline FRET in the presence of 1 mM EGTA and 50 μg/ml ionomycin, then, they were sprayed with a solution containing 100 mM CaCl₂ to measure the FRET change in response to Ca²⁺. The NIR GECI variants that had miRFP670nano at the N-terminus and miRFP720 at the C-terminus demonstrated a higher Ca²⁺-response compared to those with the reversed orientation of the NIR FPs. We selected the NIR GECI clones demonstrating the highest Ca²⁺-response for further characterization in bacterial lysates using multiwell plates.

In bacterial lysates, we recorded the fluorescence spectra in the presence of 1 mM Ca²⁺ or 1 mM EGTA. We identified several NIR GECI variants demonstrating an approximately 60% drop of the donor/FRET ratio in response to 1 mM Ca²⁺ (or an approximately 2.5-fold higher donor/FRET ratio in the Ca²⁺-free state compared to Ca²⁺-loaded states). Bacterial lysates of these variants had about 4–6 times brighter fluorescence in the donor channel compared to the lysates expressing iGECI control (Figure 1B), likely due to a higher protein expression level due to smaller size.

NIR GECI variants exhibiting the highest brightness and largest Ca²⁺ response were further tested in lysates of mammalian cells. We transfected HeLa with plasmid-encoding NIR GECI variants under the CMV promoter. 24 h after transfection, 5 μM BV was added, and 48 h after transfection, cells were lysed and their fluorescence spectra were recorded in the presence of 1 mM Ca²⁺ or 1 mM EGTA. The brightest variant with the largest Ca²⁺ response had a linker between miRFP670nano and TnC_m consisting of two Pro amino acid residues (L1) and a linker between TnC_m and miRFP720 consisting of a single Pro residue (L2) (Figures 1C,D). We named this variant iGECInano. We concluded that, short rigid linkers between NIR FPs and the Ca²⁺-sensing domain were essential for advanced iGECInano properties.

The absorption spectrum of iGECInano had two major absorption maxima at 640 and 700 nm corresponding to the absorption of miRFP670nano donor and miRFP720 acceptor (Figure 2A; Table 1). iGECInano also had a smaller absorption peak at 390 nm corresponding to the Soret band of BV in NIR FPs, attributed to the absorbance of individual pyrrole rings, and a peak at 280 nm attributed to a sum of absorbances of BV and aromatic residues of the apoprotein. The indicator had very low absorption at 450–500 nm, allowing it to combine with GFP-based indicators and opsin-based optogenetic actuators.

Next, we measured the fluorescence emission of iGECInano in the presence of 1 mM CaCl₂ (Ca²⁺-loaded state) or 1 mM EGTA (Ca²⁺-free state) in the lysates of HeLa cells. iGECInano fluorescence spectrum had two peaks at 670 and 718 nm, corresponding to miRFP670nano and miRFP720, respectively (Figure 2B). iGECInano exhibited an approximately 400% change in the FRET-acceptor to FRET-donor fluorescence ratio between the Ca²⁺-free and Ca²⁺-loaded states, mainly attributed to the decrease in miRFP670nano3 fluorescence. A small increase of FRET acceptor fluorescence in response to Ca²⁺ likely resulted from the compensation of the decrease of the donor and increase of the acceptor fluorescence signals.

We then measured the Ca²⁺ affinity of iGECInano (Figure 2C). The obtained Kd was 530 ± 20 nM, with a Hill coefficient of 1.53 ± 0.08. As expected, since TnC_m has only two EF-hands, it demonstrated the relatively simple dependency of the signal on Ca²⁺ concentration compared to iGECI (Shemetov et al., 2021). Interestingly, the Ca²⁺ affinity of iGECInano was somewhat lower than reported for Twitch-2B (200 nM), which contained the same TnC_m domain.

We also studied the pH stability of iGECInano in the presence of 1 mM CaCl₂ or 1 mM EGTA (Figure 2D). We observed minimal fluorescence changes between pH 6.0–10.0 and a decrease of fluorescence in the acidic range with pK_a 4.6 ± 0.1 for Ca²⁺-loaded and 4.8 ± 0.2 for Ca²⁺-free states. This pH dependence was similar to iGECI (Shemetov et al., 2021) and substantially broader than observed for the NIR–GECO series of indicators (Qian et al., 2019).

Next, we compared the effective (cellular) brightness of iGECInano, iGECI, and NIR-GECO2G expressed in transiently transfected HeLa cells under the same CMV promoter. In the absence of an externally added BV, iGECInano was approximately 3-folds dimmer compared to iGECI (Figure 3A). However, in the presence of 5 μM external BV, iGECInano demonstrated about 40% higher brightness compared to iGECI (although not statistically significant, p > 0.05) and about 100% higher brightness compared to NIR-GECO2G (Figure 3B). iGECInano had lower than iGECI effective brightness in the absence of exogenous BV, even though neither miRFP670nano nor miRFP720 required BV supply to brightly fluoresce in mammalian cells.

miRFP670nano (Oliinyk et al., 2019) was reported to be more photostable compared to miRFP670 used in iGECI. However, iGECInano appeared to be somewhat less photostable (photobleaching t_0.5 = 780 ± 60 s) compared to iGECI (t_0.5 = 1750 ± 150 s). Still, it was significantly more photostable than NIR-GECO2G (t_0.5 = 75 ± 15 s) (Figure 3C).

Finally, iGECInano allowed reliable recording of histamine-induced Ca²⁺-oscillations in HeLa cells (Figure 3D). Overall, although iGECInano had a relatively low cellular brightness without exogenous BV, its higher photostability and pH dependence allows it to be competitive with NIR-GECO2G. In experiments on cultured cells or tissues, external BV can be easily supplied to increase the iGECInano signal.

Next, we compared the performance of NES-iGECInano and NES-iGECI (versions of the indicators fused with a nuclear export signal, therefore localized primarily in the cytoplasm and not in the nucleus), as well as NIR-GECO2G, in mouse primary neurons isolated from the cortex. On seventh DIV, the cultured neurons were transduced with AAVs encoding iGECInano, iGECI, or NIR-GECO2G under the CAG promoter and incubated for 7–9 days in the presence of 5 μM BV. All three indicators demonstrated a uniformly distributed fluorescence in the cytoplasm of neurons (Figure 4A). We then stimulated neuronal activity using 1-ms single pulse electric field stimulation while imaging the fluorescence of cell bodies (Figure 4B). The changes in fluorescence induced after electric stimulation showed a peak between −21.7% and −44.9% among cells expressing different sensors, without a significant difference between groups [mean −ΔF/F of 21.7% (11.8%–33.5%), 17.8% (14.3%–21.1%), and 44.9% (27.1%–67.5%) for iGECInano, iGECI, and NIR-GECO2G; p > 0.05 for all paired comparison] (Figure 4C). Values are the mean and (25%–75%) quartiles. To assess the kinetics of the evoked signals in cortical neurons, we first measured the onset of the fluorescence responses. Cells expressing iGECInano and iGECI showed similar time-to-peak values [0.7 s (0.4–0.85 s) vs. 0.6 s (0.35–0.65 s) for iGECInano and iGECI, respectively; p = 0.327], while cells expressing NIR-GECO2G presented a response onset of 1.8 s ([1.4–2.2 s), showing a slower kinetics compared to both, iGECInano (p = 0.029) and iGECI (p = 0.018) (Figure 4D). We then measured the decay of the evoked changes in fluorescence, finding that cells expressing iGECInano showed the fastest kinetics, being ∼2.3 times faster than cells expressing iGECI [2.4 s (1.3–3.1 s) vs. 5.7 s (5.4–7.4 s) for iGECInano and iGECI, respectively; p = 0.002], and ∼2 times faster than cells expressing NIR-GECO2G [2.4 s (1.3–3.1 s) vs. 4.9 s (4.0–6.1 s) for iGECInano and NIR-GECO2G, respectively; p = 0.035] (Figure 4E). Taken together, these data show that iGECInano displays faster kinetics when compared to iGECI and NIR-GECO2G.

The observed faster kinetics probably resulted from a smaller and simpler Ca²⁺-sensing domain of iGECInano, which did not require interdomain interactions like the binding of M13 or RS20 peptides in calmodulin-based indicators. Therefore, iGECInano combines fast response kinetics with favorable photochemical characteristics (Table 1).

miRFP670nano and a truncated version of miRFP720 (deletion of 17 N-terminal amino acids, miRFP720Δ17N) were PCR-amplified from pmiRFP670nano and pmiRFP720-N1 plasmids. A minimal TnC (TnC_m) sensing module was PCR-amplified from Twitch-2B-pcDNA3 (Addgene #49531). A NIR-GECO2G gene was PCR-amplified from pAAV-CAG-NIR-GECO2G (Addgene #159605). A plasmid pUCmini-iCAP-PHP.eB encoding-modified AAV2/9 capsid was kindly provided by Prof. V. Gradinaru (California Institute of Technology, Addgene #103005). A pHelper plasmid was from the AAV-Helper Free System kit (Agilent #240071). Plasmids encoding iGECI were previously constructed in our lab. For the cytoplasm-restricted localization, nuclear export (NES) signal peptide NELALKLAGLDINK was appended to the N-terminus of iGECInano via a short GGGS linker using a primer-extension PCR.

For the bacterial expression of GECI variants, pBAD/HisD vector (Life Technologies/Invitrogen) was used. Mammalian expression plasmids were based on a pEGFP-N1 vector (Clontech) with a standard CMV promoter. For the expression in dissociated neurons, iGECInano or NIR-GECO2G genes were cloned into pAAV-CAG-mRuby (Addgene #99123) in place of mRuby.

DNA fragments encoding miRFP670nano, TnC_m, and miRFP720Δ17N were PCR amplified and ligated into the pEGFP-N1 vector. A set of constructs was created with miRFP670nano3 at the N-terminus and miRFP720 at the C-terminus and vice versa connected directly or with 1–4 a.a. randomized linkers. All variants were cloned into the pBAD/His-D vector and transformed into BL21 AI E. coli host (ThermoFisher Scientific) containing a pWA23h plasmid encoding heme oxygenase (HO) for BV synthesis. Cells were grown overnight in an LB medium containing 0.02% rhamnose and 0.05% arabinose for induction of HO and iGECI syntheses, respectively.

The library of clones was sorted with FACS, using double-positive gating for eliminating non-fluorescent clones resulting from stop-codons and frameshifts. The pre-sorted library was plated on Petri dishes containing 0.02% rhamnose and 0.05% arabinose so that approximately 100 colonies were growing on each dish. Dishes were incubated overnight at 37°C, then for 24 h at 18°C. Colonies were transferred to nitrocellulose membranes and permeabilized by spraying with Ca²⁺-free solution (30 mM MOPS, pH 7.5, 100 mM KCl, 50 μg/ml poly-L-lysine, 50 μg/ml ionomycin). Membranes were incubated for 5 min, and basal fluorescence in the donor (ex. 605 nm, em. 680 nm) and FRET (ex. 605 nm, em. 720 nm) channels were acquired using an IVIS instrument (Perkin Elmer/Caliper Life Sciences). Then, membranes were treated by spraying with high Ca²⁺ solution (30 mM MOPS, pH 7.5, 100 mM KCl, 50 μg/ml poly-L-lysine, 50 μg/ml ionomycin, 100 mM CaCl₂), incubated for 5 min, and the Ca²⁺-loaded state of fluorescence was recorded using the same filter sets. Data were analyzed using ImageJ software. 12–16 dishes or approximately 1,200–1,600 colonies were screened on membranes for each library per a round of screening.

Clones with the best Ca²⁺-loaded/basal fluorescence ratio (typically 50–100 colonies per a round of screening) were transferred to 5 ml LB liquid culture in 24 deep-well plates containing 0.02% rhamnose and incubated for 2 h at 37°C on a rotating shaker. Then, arabinose was added to 0.05%, the culture was incubated for 3 h at 37°C, then for 24 h at 25°C. The bacterial pellets were lysed with B-PER (ThermoFisher Scientific). The cleared lysates were transferred to 96-well plates and divided. One part was loaded with 1 mM CaCl₂ and another one with 2 mM EGTA. Fluorescences of the donor (ex. 605 nm, em. 670 nm) and FRET (ex. 605 nm, em. 720 nm) were acquired using the SpectraMax M2 plate reader (Molecular Devices).

The best performing constructs were subcloned into a mammalian expression vector and evaluated in HeLa cell lysates. HeLa cells were transiently transfected using FuGENE HD (Promega); 24 h after transfection, 5 µM BV was added. 48 h after transfection, the cells were harvested and lysed with M-PER (ThermoFisher Scientific). Lysates were cleared by centrifugation and divided in two samples; one part was loaded with 1 mM CaCl₂ and another one with 2 mM EGTA. Fluorescence spectra were recorded with the FluoroMax-3 spectrofluorometer (Horiba). The best performing clones were subjected to a new round of L1 and L2 evolutions in E. coli and HeLa cells. Three rounds of evolution and screening were performed to identify the final indicator variants.

The iGECI constructs with polyhistidine tags at the N-terminus were expressed in the BL21-AI host (Life Technologies/Invitrogen) containing a pWA23h plasmid. Bacteria were grown in an LB medium supplemented with ampicillin, kanamycin, and 0.02% rhamnose for 2 h at 37°C, followed by the induction of protein expression with 0.05% arabinose for 3 h at 37°C, then 24 h at 25°C on a rotating shaker. The proteins were purified using Ni-NTA agarose (Qiagen).

For absorbance measurements, a Hitachi U-2000 spectrophotometer was used. Fluorescence spectra in the range of 650–740 nm were recorded with the FluoroMax-3 spectrofluorometer.

Ca²⁺ titrations were carried out using EGTA-buffered Ca²⁺ solutions (Calcium Calibration Buffer Kit, Life Technologies). We prepared buffers by mixing a Ca²⁺-EGTA buffer and an EGTA buffer to give free Ca²⁺ concentrations ranging from 0 to 39 μM at 25°C. Fluorescence intensities were plotted against Ca²⁺ concentrations and fitted by a sigmoidal binding function to determine Kd.

Using a series of Hydrion buffers (Micro Essential Laboratory), pH stability was studied in the presence of either 2 mM EGTA or 1 mM CaCl₂. Fluorescence was excited at 620 nm, and the emission was recorded at 640–760 nm. The area under the spectra at different pH values was quantified.

A comparison of the brightness of iGECInano, iGECI, and NIR-GECO2G was performed in HeLa cells transiently transfected with FuGENE HD. For a BV-saturated condition, 5 µM BV was added 24 h after transfection. 48 h after transfection, cells were analyzed with a BD LSRII flow cytometer using a 640 nm excitation laser and a 647 nm long-pass edge emission filter. The NIR fluorescence intensity was normalized to the efficiency of absorption at 640 nm for each indicator. Flow cytometry gating was performed using intact, single cells.

Photobleaching measurements of the indicators in live HeLa cells and dissociated mouse neurons were performed with the ×100, 1.4 NA oil-immersion objective lens (UPlanSApo, Olympus) and a 605/30-nm excitation and a 647 nm long-pass emission filters at a light power density of 14 mW cm⁻² measured at the back aperture of the objective lens (∼8.3 W cm⁻² at the specimen plane) and normalized to the efficiency of absorption at 605 nm for each indicator.

To measure Ca²⁺ transitions evoked by histamine, HeLa cells were transiently transfected with iGECI using FuGENE HD (Promega) and cultured for 48 h. Then, the medium was changed to Live Cell Imaging Solution (Life Technologies, Invitrogen), supplemented with 1 mM CaCl₂, 100 mM KCl and 1 mM D-glucose, and basal fluorescence in the donor (excitation, 605 nm; emission, 667/30 nm) and FRET (excitation, 605 nm; emission, 725/40 nm) channels was recorded. Time-lapse imaging was performed with an Olympus IX81 inverted epifluorescence microscope, equipped with a 200 W xenon lamp (Sutter Instruments) and a ×60, 1.35 NA oil-immersion objective lens (UPlanSApo, Olympus). The microscope was operated with SlideBook v. 6.0.8 software (Intelligent Imaging Innovations). A histamine solution was added to the cells to a final concentration of 100 μM, and the fluorescence was recorded for 10 min. After that, the imaging solution with Ca²⁺ and histamine was replaced by a Ca²⁺-free imaging solution, and the cells were incubated for 5 min to equilibrate Ca²⁺. Then, an imaging solution containing 2 mM EGTA was added to the cells, and fluorescence was recorded for another 5 min. The images were analyzed using ImageJ software.

AAV particles were obtained as described (Challis et al., 2019). Briefly, plasmid DNA for AAV production was purified with the NucleoBond Xtra Maxi EF kit (Macherey-Nagel), and AAV-293T cells (Agilent) were co-transfected with pAAV-CAG-NIR-GECO2G or pAAV-CAG-iGECInano plasmid, AAV capsid plasmid pUCmini-iCAP-PHP.eB, and pHelper using polyethyleneimine (PEI; Santa Cruz). Cell medium was collected 72 h after transfection. The cells and the medium were collected 120 h after transfection and combined with the medium collected at 72 h. The cells were harvested by centrifugation and then lysed with a salt-active nuclease (HL-SAN, ArcticZymes). Polyethylene glycol (PEG, 8%) was added to the medium, and the mixture was incubated for 2 h on ice and then pelleted. The PEG pellet was treated with HL-SAN and combined with the lysed cells. The cell suspension was clarified by centrifugation. The supernatant was applied on an iodixanol gradient and subjected to ultracentrifugation for 2  h and 25 min at 350,000 g. The virus fraction was collected, washed, and enriched at an Amicon Ultra-15 100,000 MWCO centrifuge device. The virus titer was determined by qPCR. An aliquot of the virus was treated with DNase I and proteinase K and then used as a template for qPCR. A pAAV-CaMKII–iGECInano plasmid of known concentration that was digested with EcoRI was used as a reference.

Neurons were isolated from the hippocampi of P0–P1 Swiss Webster mice using a published protocol (Beaudoin et al., 2012) and cultured in a Neurobasal Plus medium with B-27 Plus Supplement (Gibco), additional GlutaMAX (1 mM; Gibco), 100 U ml⁻¹ penicillin, and 100 μg ml⁻¹ streptomycin, on poly (D-lysine) (EMD Millipore)-coated glass coverslips (thickness, 0.13–0.17  mm; diameter, 12 mm; ThermoFisher Scientific) at a density of ∼70,000 cells per coverslip. Half of the medium was exchanged twice per week. For experiments with AAVs, neurons were transduced at DIV7 with 10⁹ viral genomes (vg) per well (in 24-well plates) and recorded at DIV16–DIV18 at 37°C. The Grass S48 stimulator (Grass Instruments) and custom platinum electrodes (0.5-mm diameter) were used for field stimulation (1 ms single pulses, 50 V). The following synaptic transmission inhibitors were applied: 10 μM CNQX (R&D Systems), 10 μM gabazine (Santa Cruz Biotechnology), 10 μM (R)-CPP (Enzo Life Sciences), and 1 μM (S)-MCPG (Cayman Chemicals). A field stimulation using 1 ms single pulses, 50 V, in the presence of synaptic transmission inhibitors was previously described to generate single APs in cultured neurons (Wardill et al., 2013). A 617 nm light-emitting diode (LED) (Mightex Systems) was used for fluorescence excitation. The excitation filter was 620/15 nm, with a 640LP dichroic mirror, and the emission filter was 667/30 nm for iGECI and iGECInano and 720/40 nm for NIR-GECO2G. The frame rate was 10 Hz for NIR-GECO2G and 5 Hz for iGECI and iGECInano. Fluorescence was recorded using an Orca-Flash4.0 LT camera (Hamamatsu), an Olympus IX81 microscope, and a LUCPlanFLN ×20, 0.45 NA air objective lens (Olympus). Light power density at the specimen plane was 1.4 W cm⁻² (six-folds lower than in the photobleaching experiments), and the total duration of imaging was less than 0.5 h. The bath solution contained (in mM) 125 NaCl, 2.5 KCl, 1 MgCl₂, 10 HEPES, 3 CaCl₂, and 30 D-glucose at pH 7.3, 305–307 mOsm. The stimulator, camera, and LEDs were controlled by Master-8 (AMPI) and MatLab R2018b (MathWorks). Each neuronal recording was performed from an individual coverslip. Stimulation evoked changes in the fluorescence were quantified relative to the 10 s baseline before stimulation. Time to the peak was measured as the time between stimulus onset and the peak of the evoked response. Decay time was quantified as the time constant of a fitted single exponential curve. Data analysis was performed using Igor Pro 9 software. A statistical analysis was performed using the Kruskal–Wallis test correcting for the False Discovery Rate. We inform p values (q) corrected for multiple comparisons. We used GraphPad Prism v. 9.3.1 software.