Introduction

The ability of cells to generate transient Ca²⁺ fluctuations in subcellular locales is fundamental for controlling essential cellular processes such as gene expression, cell metabolism or neuronal communication^1,2 Organelles such as mitochondria or the ER can function within cells as sinks or sources of Ca²⁺. They modulate Ca²⁺ dynamics across cellular compartments with temporal and spatial precision, shaping cell physiology^3,4. For ample, Ca²⁺ transfer between the ER and mitochondria accelerates mitochondrial metabolism to fuel energy-intensive tasks⁵, modulates mitochondrial fission-fusion dynamics⁶,7 or influences central cellular processes such as autophagy and stress responses⁸,⁹ On the other hand, mitochondria and ER uptake Ca²⁺ in neurons during neurotransmission, shaping the capacity of neuronal communication^10-14. Not surprisingly, owing to the critical role of ER and mitochondrial Ca²⁺ signaling in both physiological and pathological processes^{3, 15-17}, their investigation has garnered significant attetinon over the past several decades^{2 18}.

For many years, insights into organellar Ca²⁺ functions were primarily obtained using indirect cytosolic Ca²⁺ measurements coupled with organelle manipulations’. However, recent advancements now enable measuring Ca²⁺ dynamics directly within the ER or mitochondria using adapted genetically encoded sensors (GECis)<sup>^{11, 12 19-22}Among previous mitochondria and ER Ca²⁺sensors, GCaMP variants have stood out due to their high dynamic range and sensitivity^11-14 -^19-23 -²⁴ However, for intraorganellar use they have inherent limitations. First, their excitation spectrum overlaps with other commonly used fluorescent proteins, sensors and optogenetic tools2s--²¹, complicating multicolor imaging needed for dissecting complex cellular processes; second, blue light excites autofluorescence in biological tissues²⁸, making it difficult to detect calcium dynamics in small organellar structures with green fluorescent indicators^{12 22} third, imaging with blue light is associated with higher phototoxicity²⁹,³⁰ which may complicate measurements over time. These limitations suggest that red-shifted GECls could provide a reliable imaging alternative for imaging Ca²⁺dynamics in organelles. However, despite recent advances, available red-shifted GECls for ER and mitochondrial Ca²⁺ imaging are not ideal due to either an inappropriately high Ca²⁺ affinity2¹ or a limited dynamic range¹⁹,²⁰ Additionally, the lack of far-red Ca²⁺sensors for these organelles has further limited simultaneous multi-wavelength imaging of organellar and cytosolic Ca²⁺ dynamics, along with other key physiological markers, limiting multi-compartment dissections of Ca²⁺ fluxes and their relevance for cellular physiology in health and disease.

Leveraging HaloTag³¹ as a scaffold for generating novel sensors has provided several novel indicators that enable measuring diverse pieces of biology in different colors, as the color of emission is provided by different membrane-permeable rhodamine dye-ligands that irreversibly bind to the HaloTag. These include sensors for voltage^32,33, ATP34, NAD+, ³⁴, dopamine³⁵, pH³⁶ and protein aggregation³⁷, among others^38,39. However, while several HaloTag-based strategies have also provided robust multicolor cytosolic Ca²⁺ sensors40-4³, suitable versions to expand the palette of organellar Ca²⁺sensors do not exist. In this work, we present a new suite of chemigenetic indicators derived from HaloCaMP, a genetically encoded Ca²⁺ sensor whose fluorescence is provided by membrane-permeable Janelia Fluor rhodamine dye-ligands with a choice of colors appended to the HaloTag Ligand (JF-HTL)⁴⁰. We generated novel organellar HaloCaMP sensors, tailored specifically for Ca²⁺ measurements in ER and mitochondria: ER-HaloCaMP and Mito-HaloCaMP. These sensors combine robust responsiveness with the enhanced brightness and photostability provided by JF-HTLs. Along with these advantages, the indicators feature tunable red and far-red emission, enabling multicolor imaging without the spectral overlap or interference issues commonly associated with green GECis. We compared ER-HaloCaMP and Mito-HaloCaMP to existing red-emitting organellar sensors, and found they responded significantly better (1.6-to 2.3-fold) and were significantly brighter (10-to 50-fold). Physiological responses of ER-HaloCaMP reconstituted with JF₅₈₅-HTL (ER-HaloCaMP₅₈₅) in cells and neurons were indistinguishable from the those of the gold-standard green ER Ca²⁺ sensor ER-GCaMP6-210¹2, but the photostability and brightness-over-background of ER-HaloCaMP₅₈₅ were significantly improved, making it the ideal sensor for ER Ca²⁺ measurements. This work provides a new technological framework to enable novel investigations on cellular Ca²⁺ fluxes, offering new insights into the fundamental role of Ca²⁺ signaling in health and disease.

Results

Developing a bright and highly responsive red ER Ca²⁺ sensor based on HaloCaMP

The ER contains ∼2000 times more Ca²⁺ than the cytosol. This surplus of Ca²⁺ inevitably saturates all existing red cytosolic Ca²⁺ GECis ^44-49 making them unresponsive to changes in ER Ca²⁺. To overcome this issue and image ER Ca²⁺ in red, previous efforts have screened for mutations that lower the affinity of red GECI to match resting ER Ca²⁺ levels. These approaches generated LAR-GECOs, which preserved a good dynamic range yet their affinity for Ca²⁺ was not lowered enough to match ER Ca²⁺ levels, limiting their usability2¹. Alternative efforts generated R-CEPIAer, which presented low affinity but its responsiveness remained ten times lower compared to that of green ER Ca²⁺ GECls¹⁹.

To overcome the limitations of existing red ER Ca²⁺ sensors and expand their spectral range beyond red, we first generated a series of HaloCaMP1a variants harboring mutations in key Ca²⁺-coordinating sites to lower their affinity (Fig. lA; Supp. Fig. SlA; Table 1). We then performed Ca²⁺ titrations and identified a variant that presented both high dynamic range (∼15 ΔF/F₀) and low affinity (EC₅₀ = 86 µM), which we named Low Affinity (LA)-HaloCaMP₅₈₅ (Fig. 1B). The general properties of this low affinity variant did not significantly change when Ca²⁺ titrations were performed at 37 °C, supporting its usability for live cell experiments (Supp. Fig. SlC; EC₅₀ = 89 µM, ∼17 ΔF/F₀). As a control, we confirmed that micromolar/millimolar increases in Ca²⁺ did not cause any increase in fluorescence of HaloTag reconstituted by JF₅₈₅-HTL (HaloTa&₈₅; Fig. 1B). Using purified LA-HaloCaMP₅₈₅, we confirmed that its excitation and emission spectra yielded a peak excitation at 593 nm and emission at 607 nm, confirming its usability as a red sensor (Supp. Fig. SlD).

Table 1.

Table 2.

Figure 1.

We next targeted LA-HaloCaMP to the lumen of the ER by flanking it with a modified calreticulin signal peptide and a KDEL retention motif, as successfully used for ER-GCaMP6¹². This construct, which also was fused to mTagBFP2, was efficiently expressed in the ER of HeLa cells. We confirmed that JF₅₈₅-HTL successfully traversed both the plasma and ER membranes, demonstrating that conjugation between sensor and dye-ligand in the ER is possible (Fig. 1C). We performed in-cell Ca²⁺ titrations in permeabilized HEK cells and found that ER-HaloCaMP presents an affinity of 115 µMin cells (Supp. Fig SlD), slightly lower than purified protein. Such a shift between purified and in-cell sensor properties is expected and has been observed for ER-GCaMP6 variants⁵⁰. Given the higher dynamic range of ER-HaloCaMP and its better suited affinity for ER Ca²⁺ than LAR-GECO and R-CEPIAer in purified protein (Fig. 1D), we next sought to compare its responsiveness against these sensors under common ER Ca²⁺ perturbations, such as thapsigargin-induced ER Ca²⁺ depletion⁵¹. Quantifying responses in both HeLa cells and somata of primary cortical neurons, we observed that although LAR-GECO and R-CEPIAer responded to thapsigargin as previously reported^19,21,52, ER-HaloCaMP₅₈₅ exhibited a significantly larger response in both cell types (Fig. 1E).

We next compared the brightness of ER-HaloCaMP₅₈₅ against previous red ER Ca²⁺ sensors. While the ER is abundant and easily detectable when measured in bulk within a cell, the small ∼60 nm diameter of individual ER tubules can hinder ER Ca²⁺ detection in subcellular locales. This problem becomes apparent when studying neuronal projections, as only a single narrow ER tubule may be present in axons^53,54 or dendritic spines⁵⁴,⁵⁵ We thus reasoned that sensors with increased brightness should facilitate studying ER Ca²⁺ in subcellular compartments in neurons and other cells. We first compared the molecular brightness in the Ca²⁺-saturated state of ER-HaloCaMP₅₈₅ against previous red ER Ca²⁺ sensors. Purified ER-HaloCaMP₅₈₅ reconstituted with JF₅₈₅-HTL exhibited a brightness of 90 mM-¹cm-1, substantially higher than LAR-GECOl (7.2 mM-¹cm-¹) and R-CEPIAer (6.2 mM-¹cm-¹) (Table 1), confirming a nearly 15-fold increase in molecular brightness relative to currently available red GECls. We next expressed LAR-GECO, R-CEPIAer and ER-HaloCaMP₅₈₅ in HeLa cells and primary cortical neurons and using identical illumination and detection conditions we compared their relative in-cell brightness. We found that ER-HaloCaMP₅₈₅ was approximately 100-fold brighter in HeLa cells and 20-fold brighter in somata of primary neurons (Fig. 1F, G), supporting data from purified protein. Importantly, these differences in live cell imaging cannot be ascribed to our optical setup, as it is actually better matched to excite and collect fluorescence from LAR-GECOl and R-CEPIAer (see Methods). This suggests that the enhanced detectability of ER-HaloCaMP₅₈₅ in cells is driven by its intrinsically higher brightness.

Next, we investigated the ability of these sensors to detect ER fluorescence within the single, narrow ER tubules of axons, a task that represents a significant challenge in ER Ca²⁺ imaging. This comparison provided a substantial improvement over previous technologies, as neither LAR-GECO nor R-CEPIAer enabled axonal ER detection in any neuron presenting somatic ER labeling, whereas ER-HaloCaMP₅₈₅ allowed us to identify it clearly in over 50% of the transfected neurons when using identical imaging conditions (Fig. 1G, H; see Methods). These results indicate that ER-HaloCaMP₅₈₅ provides a significant improvement in red ER Ca²⁺imaging by providing significantly better responsiveness and brightness.

ER-HaloCaMP enables ER Ca²⁺ measurements in locales with low ER content

As ER-HaloCaMP₅₈₅ is the only red ER Ca²⁺sensor visible in axons of neurons, we sought to evaluate its responsiveness in this small compartment. When neurons fire action potentials, axonal ER uptakes Ca²⁺ from the cytosol through sarcoplasmic/endoplasmic reticulum Ca²⁺-ATPases (SERCAs)^{10, 12}, ⁵⁶ Using field stimulation (see Methods), we electrically stimulated primary cortical neurons to mimic physiological firing paradigms of neurons in vivo, such as firing at 20 Hz for ls⁵⁷, and quantified axonal ER Ca²⁺ dynamics using ER-HaloCaMP₅₈₅. We observed robust activity-driven ER Ca²⁺ uptake, which was completely silenced when inhibiting SERCA pumps using thapsigargin, confirming the specificity of the measurements (Fig. 2A, B). Since other red ER Ca²⁺ sensors were not visible in axons, we compared the performance of ER-HaloCaMP₅₈₅ with the current gold standard ER Ca²⁺ sensor, the GFP-based ER-GCaMP6-210¹². ER-HaloCaMP₅₈₅ exhibited uptake responses comparable to those of ER-GCaMP6-210, demonstrating its high efficacy in measuring ER Ca²⁺ dynamics within small ER volumes (Fig. 2C, D).

Figure 2.

To examine whether ER-HaloCaMP₅₈₅ could accurately detect physiological ER Ca²⁺release (decrease in signal) and not only uptake (increase in signal), we assessed its responsiveness in neuronal dendrites, another subcellular localization containing limited ER volume. During excitatory neurotransmission, dendrites receive extracellular glutamate inputs from other neurons, which drive Ca²⁺ entry from the extracellular space and cause Ca²⁺ release from dendritic ER²⁴, ^{58, 59}

To study this process, we locally released glutamate near single dendritic spines of neurons in culture using glutamate uncaging^60,61 and examined ER Ca²⁺ responses using ER-HaloCaMP₅₈₅. Our experiments revealed that ER-HaloCaMP₅₈₅ responds to single-glutamate uncaging pulses with distinct ER Ca²⁺ release events and a high signal-to-noise ratio. The relative magnitude of these events was indistinguishable from those obtained with ER-GCaMP6-210 (Fig. 2E-G, see Methods).

However, when comparing the photostability of prolonged measurements of ER Ca²⁺ release, we found that signals of ER-HaloCaMP₅₈₅ robustly recovered to the initial baseline, while ER-GCaMP6-210 showed a loss of fluorescence at the end of the experiment, suggesting reduced photostability (Fig. 2H). We also observed that ER-GCaMP6-210 exhibited a rapid decrease in fluorescence immediately after 488 nm light exposure both in dendrites and axons, which then quickly settled into a new steady-state baseline (Fig. S2A, B). While this phenomenon has been previously observed on ER-GCaMPs^58, its origin remains unclear. It may reflect a rapid photophysical transition, such as reversible photoswitching or photoisomerization, in which an initial exposure to blue light drives a fraction of the GFP-based molecules into a less fluorescent state, quickly reaching equilibrium⁶². Contrary to this behavior, ER-HaloCaMP₅₈₅ maintained a stable and bright baseline signal throughout the imaging period in both axons and dendrites, exhibiting superior photostability compared to ER-GCaMP6-210 (Fig. S2A, B).

ER-HaloCaMP fluorescence is tunable and enables far-red ER Ca²⁺ measurements

We next evaluated the usability of ER-HaloCaMP conjugated with JF₆₃₅-HTL (ER-HaloCaMP₆₃₅) for measuring ER Ca²⁺ levels in the far-red spectrum, which, to our knowledge, has not yet been possible. Conjugation with JF₆₃s-HTL instead of JF₅₈s-HTL led to excitation and emission spectra with a peak excitation at 643 nm and emission at 654 nm, confirming its usability as a far-red sensor (Supp. Fig. S3A). Using purified protein, LA-HaloCaMP₆₃₅ presented a shift in the Ca²⁺ affinity compared to LA-HaloCaMP₅₈₅, increasing it by approximately 2.5-fold (Table 1). ER-HaloCaMP₆₃₅ retained a molecular brightness of 48 mM^-1cm^-1 which remains substantially higher than that of conventional red GECis (∼7-fold increase, Table 1). When imaged in HeLa cells, ER-HaloCaMP₆₃₅ generated a far-red fluorescence pattern consistent with ER labeling (Fig. 3A). Upon thapsigargin addition, fluorescence decreased by about 80%, confirming its usability as a far-red ER Ca²⁺ sensor (Fig. 3B). Given the non-existence of other far-red ER Ca²⁺ sensors, we evaluated ER-HaloCaMP₆₃₅ by comparing its AF/F changes to those of ER-GCaMP6-210 and ER-HaloCaMP₅₈₅. The results showed that the changes detected by ER-HaloCaMP₆₃₅ were approximately in the same range as those observed with ER-GCaMP6-210 or ER-HaloCaMP₅₈₅ (Fig. 3C). This result supports the versatility ofER-HaloCaMP in capturing live ER Ca²⁺ dynamics in different fluorescent colors.

Figure 3.

We then investigated whether ER-HaloCaMP₆₃₅ could be used to visualize and monitor ER Ca²⁺ dynamics fluorescence within the narrow ER tubules of axons. We first confirmed that the high brightness of ER-HaloCaMP₆₃₅ enabled successful detection of axonal ER at rest (Fig. 3D). Next, we electrically stimulated neurons and observed robust axonal ER Ca²⁺ uptake. As control, signals were completely silenced by thapsigargin, confirming that the increases in fluorescence depended on ER SERCA pumps (Fig. 3E). Benchmarking against ER-GCaMP6-210 and ER-HaloCaMP₅₈₅ showed that ER-HaloCaMP₆₃₅ exhibited lower responses, which could be a consequence of its higher affinity (Fig. 3F). Despite this caveat, ER Ca²⁺ uptake was consistently detected in response to electrical stimulation suggesting that resting ER Ca²⁺ concentration is low enough in axons for this indicator to be useful for detecting changes in subcellular regions with low ER abundance.

Developing a responsive and tunable red/far-red mitochondrial Ca²⁺ sensor based on BaloCaMP

Building on the robust performance of ER-HaloCaMP, we extended our strategy to develop mitochondrial Ca²⁺ sensors based on HaloCaMPs. Mitochondrial Ca²⁺ entry occurs dynamically in cells and is essential for cellular physiology, as it adjusts energy production, modulates metabolic pathways and influences cell survival and apoptosis¹⁸. Resting free Ca²⁺ levels in mitochondria are higher than in the cytosol, yet remain in the hundreds of nanomolar rather than micromolar range^11,63-65. Previous studies have shown that low-affinity mitochondrial Ca²⁺ sensors exhibit inadequate responses to capture physiological stimuli in HeLa cells¹⁹ and neurons⁶⁶. In contrast, adapted sensors originally developed for cytosolic Ca²⁺ measurements give robust mitochondrial responses^{11, 66}. We targeted HaloCaMPla⁴⁰ to mitochondria using tandem mitochondrial targeting sequences from cyclooxygenase 8¹¹ and fused it with mTagBFP2 to create Mito-HaloCaMP (Fig. 4A). We expressed Mito-HaloCaMP in HeLa cells and conjugated it with JF₅₈₅-HTL, which generated a bright mitochondrial pattern in HeLa cells that upon histamine addition reported a robust change in mitochondrial Ca²⁺ (Fig. 4B). To assess the efficacy ofMito-HaloCaMP₅₈₅, we compared it against Mito⁴x-jRCaMPlb^11,67 by measuring their respective AF/F₀ responses and brightness under identical conditions. Our results demonstrated that Mito-HaloCaMP₅₈₅ not only exhibited significantly larger AF/F changes but was also significantly brighter, indicating enhanced sensitivity and improved signal-to-noise for monitoring mitochondrial Ca²⁺ dynamics in HeLa cells (Fig. 4B-D).

Figure 4.

In non-excitable cells and neuronal somata, mitochondria are easily detectable and form networks of elongated structures. However, in axons their morphology is significantly reduced and they appear as small rounded structures averaging approximately 1 µm in length, thereby making their detection significantly more challenging. We next used Mito-HaloCaMP₅₈₅ to identify axonal mitochondria and electrically stimulated neurons to fire 20AP at 20 Hz as before (Fig. 4E). During neuronal firing, we observed that Mito-HaloCaMP₅₈₅ exhibited robust mitochondrial Ca²⁺ responses, generating fluorescence changes that were 3.5-fold greater than those obtained with Mito-jRCaMPlb under identical conditions (Fig. 4F). Similarly, when stimulating neurons with lO0AP fired at 100 Hz, Mito-HaloCaMPsss still reported significantly larger changes (Fig. 4F,G). These findings indicate that Mito-HaloCaMP₅₈₅ represents a significant advancement in mitochondrial Ca²⁺ imaging technology, providing a brighter and more responsive tool for red mitochondrial Ca²⁺ measurements.

We next evaluated the usability of Mito-HaloCaMP conjugated with JF₆₃₅-HTL (Mito-HaloCaMP₆₃₅) for measuring mitochondrial Ca²⁺ dynamics in the far-red spectrum, which, to our knowledge, has not yet been possible. We reasoned that the ability of measuring organellar Ca²⁺ handling in the far-red spectrum should enable quantifying Ca²⁺ dynamics in the different subcellular compartments involved in ER-mitochondria Ca²⁺ transfer, enabling the simultaneous detection of Ca²⁺ in three independent subcellular locales. We used histamine as a stimulus to drive ER Ca²⁺ release and simultaneously quantified Ca²⁺ dynamics in ER, cytosol and mitochondria of HeLa cells (Fig. SA). These experiments were performed in the absence of extracellular Ca²⁺ to ensure that all observed signals originate from the ER, thereby allowing precise quantification of how much ER Ca²⁺ is released into the cytosol versus taken up by mitochondria (see Methods). We examined the timing of Ca²⁺ changes in each compartment using the cytosolic peak as a reference point and we found that ER Ca²⁺ release occurred earlier, while mitochondrial Ca²⁺ signals lagged several seconds behind (Fig. 5B, C). These results confirm that Ca²⁺ release from the ER is the initiating event that drives subsequent cytosolic and mitochondrial Ca²⁺ accumulation. Single-cell measurements revealed that the amplitude of both cytosolic and ER Ca²⁺ correlated with the amplitude of mitochondrial Ca²⁺ increases (Fig. 5D; Fig. S3B, C). However, we observed significant variability, finding large differences in ER-mitochondrial Ca²⁺ transfer in between different cells (Fig, SE, top panel; Fig. S3A, B). As a control, we confirmed that differences did not arise from saturation of the sensors during histamine responses, as we used ionomycin at the end of every experiment to obtain the maximal response of each sensor (Fig. S3D-G). To quantify ER-mitochondrial Ca²⁺ coupling, we leveraged our simultaneous quantification of ER, mitochondria and cytosol to quantify in individual cells the distribution of ER-derived Ca²⁺ into mitochondrial and cytosolic compartments (Fig. SE, bottom; see Methods). Our results show that as ER Ca²⁺ release increases, cytosolic Ca²⁺ increases and the coupling to mitochondrial Ca²⁺ entry increases cooperatively. This indicates that ER-mitochondria Ca²⁺ transfer is not a linear process but it is dynamically adjusted with variable ER Ca²⁺ release events. This novel strategy to monitor ER-mitochondria Ca²⁺ transfer provides a robust technical framework to explore the fitness of this process in different cell types and cell states in health and disease.

Figure 5.

Lastly, we examined the capabilities of Mito-HaloCaMP₆₃₅ to detect mitochondrial Ca²⁺ uptake in single mitochondria of the axon during neuronal activity (Fig. SF). We electrically stimulated neurons to fire 20AP at 20 Hz or lOOAP at 100 Hz as before and observed that Mito-HaloCaMP₆₃₅ exhibited robust mitochondrial Ca²⁺ responses (Fig. 5G, H). As no far-red mitochondrial Ca²⁺ sensors currently exist, we compared Mito-HaloCaMP ₆₃₅ to Mito-GCaMP6f, the gold standard in the field. We found that Mito-GCaMP6f responses were indistinguishable from those of Mito-HaloCaMP₆₃₅, demonstrating its high sensitivity (Fig. 5I). We also compared these data to Mito-HaloCaMP ₅₈₅, which performed significantly better than Mito-GCaMP6f (Fig. 5I). These results show that Mito-HaloCaMP enables robust detection of mitochondrial Ca²⁺ changes in red and far-red even in challenging imaging conditions, such as detecting fluorescence changes in single small axonal mitochondria.

ER-HaloCaMP Ca²⁺ measurements in brain tissue of different species

Lastly, to extend the applicability of organellar HaloCaMPs beyond cultured cells, we next assessed the ability of ER-HaloCaMP₅₈₅ to report ER Ca²⁺ dynamics in ex vivo brain preparations from rats and Drosophila. We first expressed ER-HaloCaMP in individual CA3 pyramidal neurons in organotypic rat hippocampal slices (Fig. 6A). After labeling the sensor with JF₅₈₅-HTL, we triggered and quantified somatic ER Ca²⁺ release via DHPG (3,5-dihydroxyphenylglycine), a group I metabotropic glutamate receptor agonist that activates the production of inositol 1,4,5-trisphosphate (IP₃) and the release of Ca²⁺

Figure 6.

from internal ER stores^68,69. Brief application of DHPG led to a transient reduction in neuronal ER Ca²⁺ levels, showing that ER-HaloCaMP₅₈₅ can successfully report ER Ca²⁺ dynamics in thick samples using two-photon microscopy (Fig. 6B). Next, we evaluated ER-HaloCaMP₅₈₅ functionality in Drosophila neurons. Specifically, we targeted sensor expression to Myosuppressin (Ms) neurons within the pars intercerebralis, an evolutionarily conserved neuroendocrine center of the fly brain (Fig. 6C). Following labeling with JF₅₈,-HTL, ER-HaloCaMP fluorescence was robustly detectable within the somata of Ms neurons (Fig. 6D). Treatment with thapsigargin resulted in a clear and progressive decrease in ER Ca²⁺ levels in Ms neurons as reported by the decrease of ER-HaloCaMP₅₈₅ fluorescence (Fig. 6E), while the fluorescence of mTagBFP2 remained unaltered (Fig. S4A-C). These results demonstrate that ER-HaloCaMP can report ER Ca²⁺ dynamics across tissues from evolutionarily distant species, validating its versatility as a tool for studying ER Ca²⁺ signaling in intact tissue.

Discussion

The interplay between cytosolic and organellar Ca²⁺ pools orchestrates a myriad of cellular responses essential for cellular function and survival2.⁷⁰ and alterations in organellar Ca²⁺ signaling are heavily associated with disease^{18, 71}, ⁷⁵ However, much of our understanding of organellar Ca²⁺ fluxes has relied on indirect approaches or on monitoring Ca²⁺ in different compartments in separate experiments due to limited multiplexing capabilities^24,76-78. For example, when studying ER-mitochondria Ca²⁺ transfer, dual-compartment Ca²⁺ imaging comparing cytosol and mitochondria, or cytosol and the ER, have been used to infer the relationship between cytosolic, ER and mitochondrial Ca²⁺; ^{11, 24, 77} While these are highly valuable approaches, our triplex ER-cytosol-mitochondria Ca²⁺ measurements showed there is large cell-to-cell variability, suggesting that quantifying ER-mitochondria Ca²⁺ transfer in different populations of cells using only 2 sensors may complicate interpretations^{15, 18}, ⁷⁹, ⁸⁰ The ability to image ER and mitochondrial Ca²⁺ using the far-red part of the spectrum unlocks distinguishing the relationship between the three compartments in single cells, enabling a precise quantification of ER-mitochondria Ca²⁺ transfer that takes into account the amplitude of cytosolic signals. Triplex imaging also enables the dissection of the temporal sequence of Ca²⁺ fluxes, revealing that during histamine stimulation, Ca²⁺ is first released from the ER, subsequently peaks in the cytosol, and is then taken up by mitochondria. Our experiments also reveal that ER Ca²⁺ release favors ER-mitochondria Ca²⁺ coupling in a cooperative manner. These findings in intact cells support the current model established in isolated mitochondria and permeabilized cells, in which opening of the Mitochondrial Ca²⁺ Uniporter (MCU) is controlled by cooperative Ca²⁺ binding to accessory Mitochondrial Calcium Uptake (MICU) proteins^{81 82} Overall, these experiments reveal the importance of simultaneous single-cell measurements of different compartments when studying organellar Ca²⁺ signaling and reveal the greatpotential of far-red imaging using organellar HaloCaMPs to explore complicated biological questions.

In addition to expanded multiplexing capacities, the development of ER-HaloCAMP and Mito-HaloCaMP presents 2 important advancements. First, they show significantly improved brightness and responsiveness in both ER and mitochondrial Ca²⁺ measurements in the red spectrum: the improved brightness of our organellar sensors likely reflects the higher extinction coefficients and quantum yields of synthetic dyes^{40, 41}., as was shown when cytosolic HaloCaMP1 ₃₅ was compared to other protein-based fluorescent Ca²⁺ sensors, such as GCaMP6s⁴⁰. Enhanced brightness is of particular importance when measuring Ca²⁺ in organelles because quantifying Ca²⁺ dynamics in individual mitochondria^11,13,14 or in narrow ER tubules¹² is a demanding task that only a few sensors are able to achieve. Second, they present improved photostability compared to the current best-in-class sensors: we found that upon blue light illumination, ER-GCaMPs underwent a rapid decrease in fluorescence that reaches stability in a few seconds, as shown previously5⁸. Such photoinactivation behavior has already been reported in some cytosolic GCaMPs, including preliminary versions of the original GCaMP, as well as GCaMP7c and jGCaMP8 variants⁸3-8s. Recent work on jGCaMP8 proposed that such behavior is the consequence of reversible photoswitching, likely caused by cis-trans isomerization of the GFP chromophore, which shifts the sensor to a non-fluorescent state independent ofCa²⁺ binding⁸³. This indicates that ER-GCaMPs likely suffer from the same problem, which distorts signals independently of ER Ca²⁺ changes and confounds the quantification of rapid Ca²⁺ transients at the start of the imaging period. Because ER-HaloCaMP uses an external synthetic fluorophore rather than a circularly permuted GFP, it cannot undergo the cis-trans isomerization that promotes photoswitching. Thus, ER-HaloCaMP shows significantly more stable fluorescence signals than ER-GCaMP while retaining similar Ca²⁺ responsiveness, making it the ideal choice for endoplasmic reticulum Ca²⁺ imaging.

Organellar HaloCaMPs also demonstrate robustness across diverse biological models (cultured cell lines, primary neurons and intact tissues from different species) and are compatible with both one- and two-photon excitation This broad compatibility across model and imaging systems underscores the versatility of the HaloCaMP toolkit for investigating organellar Ca²⁺ signaling in diverse biological contexts and experimental conditions, which will facilitate their use in future research. Overall, we show that ER-HaloCaMP and Mito-HaloCaMP present improved properties that make them the preferred choice in several Ca²⁺ signaling paradigms in cell lines and primary cells. Thus, our expanded palette of organellar Ca²⁺ sensors unlocks new possibilities for the study of organellar Ca²⁺ signaling across fields, which will facilitate our understanding of organellar cell physiology in health and disease.

Material and methods

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Dr. Jaime de Juan-Sanz (jaime.dejuansanz@icm-institute.org).

Resources Table.

Molecular biology for generating the low affinity HaloCaMPs

Synthetic DNA oligonucleotides were purchased from Integrated DNA technologies. QS high fidelity DNA polymerase (New England Biolabs) wu used for all PCR amplifications. Isothermal assembly reactions were performed with a NEBuilder HiFi kit (New England Biolabs). Small scale DNA isolation were performed with QIAprep Spin Miniprep Kit (Qiagen). The pRSET vector backbone was acquired from Life Technologies. Inserts and vector backbones were amplified by PCR amplification. Vector backbones and inserts were assembled by isothennal assembly with 10-30 base pair overlap, and sequence verified by Sanger sequencing (Azenta Life Sciences) or by nanopore filll-plasmid sequencing (Plasmidsaurus).

Protein npnssfon and parifleation

For expression and purification of proteins, T7 express (New England Biolabs) were tnmsformed with pRSET plasmids encoding for the protein of interest The bacteria were grown in auto-induction media using the Studier methodn⁸⁷ with antibiotics at 30 °C for 48 h shaking at 200 rotations per miute (r.p.m.). Cell pellets were collected by centrifugation, lysed in TRIS Buffered Saline (TBS; 19.98 mM Tris, 136 mM NaCl, pH 8.0), with n-Octyl-β -D-thioglucopyranoside (S g L^-1). Aggregations were disrupted by sonication and the lysate cleared by centrifugation. Protein purification was performed on a N-tenninal poly-histidine (His₆,) tag using HisPur Ni-NTA resin (Thennofisher Scientific), acoording to manufacturer’s recornmendatio Purified proteins were buffer exchanged into TBS using Amicon concentration filters (Merck). Protein aliquots were stored at 4 °C.

Calefam titrations in purified protein

To determine the calcium affinity and cooperativity of the low affinity HaloCaMP sensom, calcium titrations were performed in a buffer system made from CaNTA and NTA prepared using the pH titration method descried by Tsien and Pozza⁸⁸ mixed in specific ratios to generate known free calcium concentrations. The free calcium ooncentration was calculated assuming the dissociation oonstant of NTA for Ca²⁺ 67 μ M. HaloCaMP proteins were pre-labeled with a limiting dye-HaloTag ligand (20 pM protein, 10 μM JF₅₈₅-HTL). 2 μL of the pre-labeled protein-dye oonjugate was diluted into 98 μL of pre-mixed solutions of Ca-NTA in black 96-well plates. Fluorescence intensities were read on a plate reader (Tecan Spark 20M). Fluorescence intensity was measured at 26 °C and 37 °C. For HaloCaMPs with JF₅₈₅-HTL excitation was 590 nm and emission 620 nm. For GCaMPs excitation was 484 nm and emission wu S10 nm. All bandwidths were set to 5 nm. Changes in fluorescence in addition of Ca²⁺ were calculated in Microsoft Excel. The fluorescence (y) was plotted against the free calcium ooncentration (x) and a four-parameter dose-response curve (variable slope) using GraphPad Prism vl0 software was fit where (a) is the value of fluorescence at the bottomof the curve, (b) is the value of fluorescence at the top of the cuve, (EC₅₀) is the ooncentration of agonist that givesa response halfwaybetweenbottom and the top, and (h) is the hillor cooperative coefficient.

Sequence of LA-HaloCaMPl

Animals

The rats of either sex used in this study to prepare primary cultures were of the Sprague-Dawley strain, Crl (SD), bred by Charles River Laboratories following the international genetic standard protocol (IGS). All experiments conducted at the Paris Brain Institute strictly followed the guidelines set by the European Directive 2010/63/EU and the French Decree n° 2013-118 for the protection of animals used in scientific research. In the case of organotypic slices, rats were housed and bred at the University Medical Center Hamburg-Eppendorf (UKE). All procedures complied with German animal welfare regulations (Tierschutzgesetz der Bundesrepublik: Deutschland, TierSchG) and conformed to Directive 2010/63/EU. Experimental protocols were approved by the Behorde fiir Justiz und Verbraucherschutz (BN), Lebensmittelsicherheit und Veterinärwesen, Hamburg.

Flies were raised on a standard cornmeal/agar diet (6.65% cornmeal, 7.15% dextrose, 5% yeast, 0.66% agar, supplemented with 2.2% nipagin and 3.4mL/L propionic acid). All experimental flies were kept in incubators at 70% humidity and on a 12h light/dark cycle, at 25°C. Flies were transferred to fresh vials every 2 days, and fly density was kept to a maximum of 20 flies per vial.

Generation of UAS-ER-HaLo-CaMP::BFP transgenic line

The ER-HaloCaMP fragment was PCR amplified using the primers (5’-aagatcctctagaggtacccTTAGAGTTCATCCTTGCC-3’) and (5’-actctgaatagggaattgggATGGGACTGTTGTCTGTTC-3’). The fragment was then cloned in the open pUASTattB vector through site-directed mutagenesis using the kit Q5-site directed mutagenesis (NEB E0554S). The pUASTattB vector was opened with restriction digestion using the restriction enzymes EcoRI-HF (NEB R3101S) and Xhol (NEB R0146S). The resulting pUASTattB-ER-HaloCaMP plasmid was used to establish transgenic lines through cpC-31 integrase mediated recombination (BestGene Inc.). The landing attP site used was VK37 PBac{yellow[+]-attP-3B}VK00037, BDSC Stock 9752.

Primary rat co-culture of postnatal cortical neurons and astrocytes

Primary co-cultures of cortical neurons and astrocytes were obtained as previously described with small modifications⁸⁹. P0 to P2 rats of mixed gender were sacrificed, and their brains were dissected in a cold HBSS-FBS solution (lX HBSS + 20% Fetal Bovine Serum) to isolate the cerebral cortexes. These were then cut into small pieces for digestion and dissociation. The tissue was washed twice with 1X HBSS-FBS and lX HBSS before being incubated in a trypsin-based digestion solution containing DNAse I (Merck, D5025) for 15 minutes at room temperature. Trypsin (Merck, T1005) was neutralized with HBSS-FBS solution, followed by two washes with lX HBSS-FBS and two with lX HBSS. The tissue was then transferred to a dissociation solution (lX HBSS, 5.85mM MgSO₄) and dissociated into single cells by gentle pipetting. The cells were centrifuged at 13,000 rpm for 10 minutes at 4°C, and the pellet was resuspended in lX HBSS solution. After a second centrifugation, the pellet was resuspended in a homemade warmed plating media consisting of MEM (Thermo Fisher Scientific, 51200038) supplemented with 20 mM Glucose (Merck, 08270), 0.1 mg/ml transferrin (Merck, 616420), 1% GlutaMAX (Thermo Fisher Scientific, 35050061), 24 µg/ml insulin (Merck, 16634), and 10% FBS (Thermo Fisher Scientific, 10082147) and 2% N-21 (Bio-techne, AR008). Finally, cells were plated in sterilized cloning cylinders (Merck, C7983; 38,000 cells per cylinder) attached to coverslips (Diameter= 25 mm, Warner Instruments, 640705) that had been pre-coated with 0.1 mg/ml poly-ornithine (Merck, P3655). After 3-4 days, the neuronal media was replaced with a homemade feeding media, similar in composition to the plating media but containing 5% FBS and 2 µM cytosine P-d-arabinofuranoside (Merck, C6645) to inhibit glial growth. The primary co-cultures of cortical neurons and astrocytes were maintained at 37°C in a humidified incubator with 95% air/5% CO2 until the imaging experiments, which were performed from days in vitro (DIV)12 to DIV21.

For cultures used in glutamate uncaging experiments shown in Figure 2E-H, conditions were as follows: hippocampal regions were dissected in ACSF containing (in mM) 124 NaCl, 5 KCl, 1.3 MgSO₄:7H₂O, 1.25 NaH₂PO₄:H₂O, 2 CaCl₂, 26 NaHCO₃, and 11 Glucose (stored at 4°C) and stored in hibernate E buffer (BrainBits LLC, stored at 4 °C). Dissected hippocampi were dissociated using Papain Dissociation System (Worthington Biochemical Corporation, stored at 4 °C) with a modified manufacturer’s protocol. Briefly, hippocampi were digested in papain solution (20 units of papain per ml in 1 mM L-cysteine with 0.5 mM EDTA) supplemented with DNase I (final concentration 95 units per ml) and shook for 30-60 min at 37°C, 900 rpm. Digested tissue was triturated and set for 3 min, following which the supernatant devoid of tissue chunks was collected. The supernatant was centrifuged at 300 ref for 5 min and the pellet was resuspended in resuspension buffer (1 mg of ovomucoid inhibitor, 1 mg of albumin, and 95 units of DNase I per ml in EBSS). The cells were forced to pass through a discontinuous density gradient formed by the resuspension buffer and the Ovomucoid protease inhibitor (10 mg per ml) with bovine serum albumin (10 mg per ml) by centrifuging at 600 rpm for 6 min. The final cell pellet devoid of membrane fragments was resuspended in Neurobasal-A medium (Gibco, stored at 4 °C) supplemented with Glutamax (Gibco, stored at -20 °C) and B27 (Gibco, stored at -20 °C). Cells were plated on poly-D-lysine coated coverslips mounted on MatTek dishes at a density of 30000-50000 cells/cm⁹⁰. Cultures were maintained at 37 °C and 5% CO2 with feeding every 3 days using the same medium until transfection. Transfections were performed 12 days after plating by magnetofection using Combimag (OZ biosciences, stored at 4 °C) and Lipofectamine 2000 (Invitrogen, stored at 4 °C) according to manufacturer’s instructions.

Preparation of Organotypic Hippocampal Slices

Organotypic hippocampal slices were prepared from Wistar rats of both sexes at postnatal days 5-7. Following dissection, the hippocampi were sectioned into 400 µm slices using a tissue chopper and transferred onto a porous membrane (Millicell CM, Millipore) for culture. Slices were maintained at 37 °C in an atmosphere of 5% CO2 in a culture medium consisting of 394 mL Minimal Essential Medium (Sigma M7278), 100 mL heat-inactivated donor horse serum (Sigma Hl 138), 1 mM L-glutamine (Gibco 25030-024), 0.01 mg/mL insulin (Sigma 16634), 1.45 mL 5 M NaCl (Sigma S5150), 2 mM MgSO₄ (Fluka 63126), 1.44 mM CaC1₂ (Fluka 21114), 0.00125% ascorbic acid (Fluka 11140), and 13 mM D-glucose (Fluka 49152). The culture medium was partially replaced (60-70%) twice per week. To express ER-HaloCaMP in CA3 neurons of the cultured slice, plasmids encoding ER-HaloCaMP-BFP were diluted to a final concentration of 10 ng/µL in a K-gluconate-based intracellular solution containing (in mM): 135 K-gluconate, 4 MgC1_2, 4 N -ATP, 0.4 Na-OTP, 10 N -phosphocreatine, 3 ascorbate, and 10 HEPES (pH 7.2). Single-cell electroporation was performed between DIV 6 and DIV 10. During the electroporation procedure, slice cultures were maintained in a pre-warmed, HEPES-bu:ffered solution containing (in mM): 145 NaCl, 10 HEPES, 25 D-glucose, 1 MgC1₂, and 2 CaC1₂ (pH 7.4, sterile filtered). Electroporation was carried out using an ELectroPORATOR (NPI), applying 12 voltage pulses (− 12 V, 0.7 ms) at 50 Hz. Slices were incubated for approx. 24h with JF₅₈₅-HTL in culture medium 5-7 days after electroporation, then rinsed in the pre-warmed HEPES-bu:ffered solution before imaging. We found that #x223C; 40% of ER-HaloCaMP transfected neurons in the organotypic slices were not sufficiently labelled by incubation with JF₅ss-HTL, presenting detectable BFP while no ER-HaloCaMP₅₈₅ fluorescence. This could be the consequence of the limited bioavailability of JFsss-HTL⁴¹, although lack of labeling was not observed when reconstituting ER-HaloCaMP in Drosophila neurons.

Cell line culture and transfection

HeLa and HEK cells were cultured either on poly-omithine coated glass coverslips or wells of a chambered coverslip with 8 wells and a glass bottom (CliniSciences, 80807-90). They were cultured in DMEM (Thermo Fisher Scientific, 10566016) supplemented with 10% fetal bovine serum and kept in an incubator with humidified air containing 5% CO₂ at a temperature of 37°C. HeLa and HEK cells were transfected using Lipofectamine 2000 (Invitrogen) the day before imaging following the manufacturer’s recommendations.

HaloCaMP labelling in cultured cells

JF dye-ligands were obtained from Dr. Lavis (Janelia Research Campus; JF₅₈₅-HTL) or from Promega (HT 1050, Janelia Fluor 635 HaloTag Ligand; JF₆₃₅-HTL). The JF dye-ligands obtained from Promega were incubated following the manufacturers instructions. JF dye-ligands obtained from Dr. Lavis were received in aliquots of 100 nmol and resuspended in 200 µL DMSO to yield a 500 µM stock. JF dye-ligands were then diluted in the cell culture media (either DMEM for cell lines or feeding media for primary neuronal cultures or organotypic slices) at a final concentration of 1µM. The cells are usually labelled after 30 minutes of incubation but for practical purposes, they were incubated overnight with the dye-ligand. Next day, the dye-ligand was washed in the morning before the experiment. To minimize non-specific labeling and reduce background fluorescence, a series of washing steps were performed by replacing the existing media with fresh, pre-warmed culture media, followed by a 10-minute incubation period. This washing step was repeated twice to ensure thorough removal of non-specific signals and to facilitate subsequent fluorescence measurements.

Calcium imaging in cells and primary neurons

Live imaging assays of primary cortical neurons transfected with calcium phosphate method⁹¹ were conducted from DIV12 to DIV21. HEK or HeLa cells imaging assays were performed one day after transfection using lipofectamine 2000 (Thermo Fisher Scientific, 11668019). The experiments utilised a custom-built, laser-illuminated epifluorescence microscope (Zeiss Axio Observer 3) paired with an Andor i.Xon Life camera (model IXON-L-897), cooled to -70°C. Illumination was provided by fiber-coupled lasers at wavelengths of 488 and 561 nm (Coherent OBIS 488 nm and 561 nm), combined using the Coherent Galaxy Beam Combiner. Laser illumination was controlled by a custom Arduino-based circuit that synchronized imaging and illumination. Neuron-astrocyte co-cultures, HEK or HeLa on coverslips were placed in a closed bath imaging chamber for field stimulation (Warner Instruments, RC-21BRFS) and imaged with a 40x Zeiss oil objective Plan-Neofluar with a numerical aperture of 1.30 (WD = 0.21 mm). Neurons were stimulated using 1 ms current pulses between platinum-iridium electrodes in the stimulating chamber (Warner Instruments, PH-2), driven by a stimulus isolator (WPI, MODEL A382) controlled by the Arduino-based circuit. All the experiments were performed at 37°C. Temperature was kept constant using a Dual Channel Temperature Controller (Warner Instruments, TC-344C) that controlled the temperature of the stimulation chamber (Warner Instruments, PH-2) and simultaneously warmed the imaging solutions using an in-line solution heater (Warner Instruments, SHM-6). Imaging solutions were flown at 0.35-0.40 mVmin. Imaging was performed using a Tyrode’s solution composed of (in mM) 119 NaCl, 2.5 KCl, 2 CaC12, 2 MgC12, 20 Glucose together with 10 µM CNQX and 50 µM AP5, buffered to pH 7.4 at 37°C using 25 mM HEPES.

Multiplexed confocal calcium imaging in cell lines and analysis of ER-mitochondria coupling

HeLa cells expressing ER-GCaMP6-210, cytosolic RCaMPlh and Mito-HaloCaMP were cultured in chambered coverslips with 8 wells and a #1.5H glass bottom (lbidi, 80827) and transfected using lipofectamine 2000 following manufacturer’s instructions (Thermo Fisher Scientific, 11668019) to be imaged the next day. Live imaging assays were conducted with an inverted Leica spinning disk equipped with a Yokogawa CSU-Xl module and a Hamamatsu Orea Flash 4.0 sCMOS camera. The objective used was 63x, NA 1.4. The effective sampling rate at each wavelength was approximately 0.83 Hz, given a 1.2s interval between successive images. The samples were illuminated using lasers at 488 nm, 561 nm, and 637 nm. The emission was collected through 525/50, 607/35, and 685/40 filters mounted in an automated rotating filter wheel. Imaging was performed using a Tyrode’s solution composed of (in mM) 119 NaCl, 2.5 KCl, 2 CaC1_2, 2 MgC1_2, 20 Glucose, buffered to pH 7.4 at 37°C using 25 mM HEPES.

Regions of interest (ROls) were drawn over individual cells, covering ER, cytosol and mitochondria. Background-substracted fluorescence time-courses (ΔAF/F₀) were calculated and peak responses were extracted from each channel individually after histamine addition. Cells typically showed multiple peak responses but only the first peak was used for analysis for consistency. A modified Tyrode’s solution containing 4mM Ca²⁺ and 500 µM ionomycin was added at the end of the experiment to confirm histamine responses did not saturate any of the sensors. When plotting triplex correlations, the size of the dots representing single cells ranged from 0.01 to 0.90 in arbitrary size units. Responses > 0.90 are shown at maximal dot size. ER-mitochondria coupling was quantified by calculating, for each cell initial response, the ratio of mitochondrial ΔF/F₀ to the sum of mitochondrial plus cytosolic ΔF/F₀. Data points were grouped into bins of 0.15 ΔF/F₀ along the ER ΔF/F₀ axis and the corresponding averaged coupling factors were plotted against ER ΔF/F₀ to assess cooperativity. To obtain averaged responses, cytosolic, ER and mitochondrial ΔF/F₀ traces were time-aligned by setting the histamine-induced cytosolic peak tot= 0, then averaged across cells. Only cells in which the peak was higher than six times the standard deviation of the baseline were included in this analysis (65 out of 71).

Measurement of dendritic ER-HaloCaMP and ER-GCaMP6-210 upon two-photon glutamate uncaging

For the two-photon glutamate uncaging experiments in Figure 2E-H, imaging was conducted 15-16 days after neuronal cell culture plating in a modified E4 imaging buffer containing: 120 mM NaCl, 3 mM KCl, 10 mM HEPES (buffered to pH 7.4), 4 mM CaC1_2, and 10 mM Glucose, lacking MgC1_2. Imaging was performed using a custom-built inverted spinning disk confocal microscope (3i imaging systems; model CSU-Wl) attached to an Andor iXon Life 888. Image acquisition was controlled by SlideBook 6 software. Images were acquired with a Plan-Apochromat 63x/1.4 NA. Oil objective, M27 with DIC III prism, using a CSU-Wl Dichroic for 488/561 nm excitation with Quad emitter and individual emitters, at back aperture laser powers 2.00 mW (488 nm) and 2.65 mW (561 nm) for spine stimulation measurements. During imaging, the temperature was maintained at 37°C using an Okolab stage top incubator with temperature control.

For ER-HaloCaMP measurements, cytosolic GCaMP6s was co-transfected to confirm effective spine stimulation (data not shown). For ER-GCaMP6-210 measurement, neurons were co-transfected with PSD95-mCherry to identify spines for stimulation. Before glutamate uncaging, neurons were treated with 1 µM TTX (Citrate salt, mM stock made in water, Abeam, ab1200552) and 2 mM 4-Methoxy-7-nitroindolinyl-caged-L-glutamate (MNI caged glutamate, Tocris Bioscience 1490, 100 mM stock made in modified E4 buffer) in modified E4 buffer lacking Mg²⁺ (see above). Glutamate uncaging was performed using a multiphoton-laser 720 nm (Mai TAI HP) and a Pocket cell (ConOptics) for controlling the uncaging pulses. Uncaging protocols of 6 uncaging pulses at 0.25 Hz with 100 ms pulse duration per pixel at 7.2 mW back aperture laser power was used.

To measure ER-HaloCaMP and ER-GCaMP6-210 response, ∼2 µm regions of interest (ROI) were drawn at the base of the stimulated spine. The average intensities were background subtracted using the intensity measured from an adjacent background area. For each successive time point during and after stimulation, the normalized intensity (ΔF/F) was calculated using the equation: ΔF/F = (F-F₀)/F₀, where F₀ is defined as the average fluorescence intensity of the consecutive 5 time points before spine stimulation, and F is defined as the fluorescence intensity measured at the time point of interest.

Ex vivo ER-HaloCaMP Ca²⁺ imaging in the Drosophila brain

Drosophila brain live imaging experiments were carried out as in Silva et al.⁹² with some modifications. Briefly, female flies carrying the Myosuppressin (Ms)-Gal4 driver9³ were crossed with male flies carrying the UAS-ER-HaloCaMP-BFP construct (which is fused to mTagBFP2, as in the case of experiments in cells in culture). Crosses for imaging experiments were raised at 25°C. 1-to 2-day-old adult progeny were used for each recording.

A single fly brain was dissected in haemolymph solution and mounted on a glass coverslip coated with poly-L-lysine (Sigma-Aldrich, P1524). The haemolymph solution contained 130mM NaCl (Sigma, S9625), 5 mM KCl (Sigma, P3911), 2 mM MgC1₂ (Sigma, M9272), 2 mM CaC1₂ (Sigma, C3881), 5 mM D-trehalose (Sigma, T9531), 30 mM sucrose (Sigma, S9378) and 5 mM HEPES-hemisodium salt (Sigma, H7637). Then, a total volume of 500 µL ofhaemolymph solution was added on top of the coverslip. Next, 1 µL of the JF₅₈rHTL 500 µM stock solution was incubated for approximately 40 min prior imaging.

Two-photon imaging was performed on a Leica Stellaris 8 DIVE upright microscope equipped with a × 25, 1.00-NA water-immersion objective. Two-photon excitation of mTagBFP2 and ER-HaloCaMP₅₈₅ was achieved using a Mai Tai eHP DeepSee MP laser tuned to 820 nm. Spectral separation was achieved by tuning detectors for mTagBFP2 (445-460nm) and JF₅₈₅-HTL (580-650nm). Then, 512 × 512 images were acquired at 0.5Hz, and the entire duration of each recording was 1200 s. For Ca²⁺ depletion experiments, Thapsigargin (Sigma, T9033) was diluted in haemolymph solution to prepare a 100 µM stock solution. After 60 s of baseline acquisition, 5 µ1of the Thapsigargin stock solution was added to the 500-µl bath solution on top of the brain, resulting in a final concentration of 1 µM.

Image analysis was performed using a custom-written MATLAB script⁹². Regions of interest (ROis) were manually drawn around individual Myosuppressin neuronal somas. The average intensity values for mTagBFP2 and ER-HaloCaMP₅₈₅ channels over each ROI were calculated over time after background subtraction. The ratio was calculated by dividing ER-HaloCaMP₅₈₅ by mTagBFP2 intensity. Traces from all cell bodies were pooled for analysis.

Image analysis and statistics

We used the ImageJ plugin Time Series Analyzer V3 for imaging analysis of dynamic changes in fluorescence, except for ex vivo experiments in which we used a custom-written MATLAB script⁹². Statistical analysis was carried out using GraphPad Prism v10 for Windows, with specific tests indicated in the Supplementary Table STl. For each dataset, normality was assessed using the Shapiro-Wilk test and then based on the result we chose either parametric or non-parametric tests for further comparisons. All n numbers, as well as the number of independent experiments, are detailed in the Supplementary Table STl. Unless otherwise indicated, all data for this study was acquired from at least 3 independent experiments. We used Python to calculate Spearman correlations and their associated p-values with the SciPy library. Pearson’s correlation coefficients and associated p values were calculated using graphpad, and both were incorporated in a single heatmap using illustrator. For the figures, the mean and the standard error of mean are used. Results of statistical analysis are shown in figures corresponding to the following criteria: *p<0.05, **p<0.01, ***p<0.001, ****p < 0.0001, n.s., not statistically significant.

Supplementary figures

The authors declare no competing interests.

Fig. S1.

Fig. S2.

Fig. S3.

Fig. S4.

Acknowledgements

We would like to thank all members of the de Juan-Sanz laboratory for insightful discussions and comments. We thank Dr. Luke Lavis and the Janelia Open Science Initiative for providing the Janelia Fluor dyes used in this study. Part of this work was carried out in the ICM.Quant core facility (RRID:SCR_026393). We gratefully acknowledge Claire Lovo, Meryem Rezzik and Astou Tangara for setting up multiplexing experiments. Animals were hosted with the help of the PHENOPARC facility at ICM.This work was made possible by the Paris Brain Institute Diane Barriere Chair in Synaptic Bioenergetics awarded to J.d.J.-S., funding from the French national program ‘‘Investissements d’avenir’’ ANR-10-IAIHU-0006 awarded to the ICM and funding by the Richard Mille Fund (Project DBS, 2023-2028). Our additional funding sources are an ERC Starting Grant SynaptoEnergy (European Research Council, ERC-StG-852873), 2019 ATIP-Avenir Grant (CNRS, lnserm), and two grants by the French National Research Agency (ANR) under project numbers ANR-24-CE16-0221 and ANR-22-CE16-0020 awarded to J.d.J.-S. D.H. is funded by an ERC Starting Grant GutSense (European Research Council, ERC-StG-101117267). A.M. is the recipient of a predoctoral fellowship from the French Ministry of Science. J.d.J.-S. is a permanent CNRS researcher and a FENS-Kavli Scholar. V.R. is supported by the Max Planck Society.

Additional information

Author contributions

Conceptualization, J.d.J-S.; methodology, A.M., H.F., R.F., K.Z., B.S, C.R., J.d.J.S.; investigation, A.M., H.F., R.F., V. R., B.S., C.R., K.Z., J.d.J.S.; project administration and supervision, J.d.J-S.; writing - original draft, A.M. and J.d.J-S.; writing - review & editing. A.M., H.F., R.F., K.Z., B.S., C.R., C.G., T.O., D.H., V.R., E. R .S. and J.d.J.S.

Additional files

Supp. Table ST1