Ca2+ tunneling architecture and function are important for secretion

Abstract

Ca2+ tunneling requires both store-operated Ca2+ entry (SOCE) and Ca2+ release from the endoplasmic reticulum (ER). Tunneling expands the SOCE microdomain through Ca2+ uptake by SERCA into the ER lumen where it diffuses and is released via IP3 receptors. In this study, using high-resolution imaging, we outline the spatial remodeling of the tunneling machinery (IP3R1; SERCA; PMCA; and Ano1 as an effector) relative to STIM1 in response to store depletion. We show that these modulators redistribute to distinct subdomains laterally at the plasma membrane (PM) and axially within the cortical ER. To functionally define the role of Ca2+ tunneling, we engineered a Ca2+ tunneling attenuator (CaTAr) that blocks tunneling without affecting Ca2+ release or SOCE. CaTAr inhibits Cl- secretion in sweat gland cells and reduces sweating in vivo in mice, showing that Ca2+ tunneling is important physiologically. Collectively our findings argue that Ca2+ tunneling is a fundamental Ca2+ signaling modality.

Figure 1.

Ca²⁺ tunneling in primary salivary gland cells. (A) Cartoon depiction of Ca²⁺ tunneling. STIM1–Orai1 generates localized Ca²⁺ influx within ERPMCS defining the SOCE microdomain. Ca²⁺ taken into the ER by SERCA diffuses through the cortical ER and is released by IP₃R to activate distal cortical targets. Generated using Biorender. (B) Comparison of SOCE and Ca²⁺ tunneling protocols. The ER Ca²⁺ stores are depleted by the reversible SERCA inhibitor CPA. This is followed by a washout phase to restore SERCA activity. SOCE develops upon the addition of extracellular Ca²⁺. Concurrent exposure to an agonist to produce IP₃ or uncaging of IP₃ activates Ca²⁺ tunneling. (C) Example traces of the cortical Ca²⁺ signals recorded during SOCE, SOCE induced by the irreversible blockade of the SERCA pump by thapsigargin (SOCE-Tg), and during tunneling. Traces are taken from Courjaret et al. (2018). (D) Cytosolic Ca²⁺ responses during SOCE and tunneling in isolated salivary gland cells. Tunneling is induced by the addition of CCh (10 µM) simultaneously to Ca²⁺ reperfusion. The bar chart quantifies the peak Ca²⁺ signal of tunneling versus SOCE (n = 26; unpaired t test, P < 0.0001). (E) Whole-cell Cl⁻ currents in response to SOCE versus tunneling, showing similar profiles to the Ca²⁺ signals; and bar chart summarizing the data (n = 9; unpaired t test, P < 0.0001). (F) Localization at the whole cell level of mCh-STIM1 and GFP-Orai1 in HeLa cells. At rest (Ctr) the confocal plane is in the middle of the cell. After store depletion with thapsigargin (Tg) the image was acquired at the PM optical plane to visualize STIM1-Orai1 co-clusters. (G) Orthogonal sections from confocal z-stacks in control conditions and after store depletion. (H and I) Relative intensities of STIM1 and Orai1 across z-stack profiles as in G before (Ctr) and after store depletion (Tg). (J) Pearson correlation coefficient (PCC) between mCh-STIM and GFP-Orai1 measured in orthogonal slices before (Ctr) and after store depletion (Tg) (n = 13–16; unpaired t test, P < 0.0001).

Figure S1.

Store depletion in primary salivary gland cells and STIM1 Orai1 localization in HeLa cells. (A) The transient application of CPA (30 µM) depletes ER Ca²⁺ stores as indicated by the rise in cytosolic Ca²⁺ (blue) which activates the Cl⁻ current (red). Following the washout of CPA, the application of carbachol (CCh, 10 µM) fails to elicit a response, indicating the effective depletion of the ER stores (n = 4–8). (B and C) Kinetics of intracellular Ca²⁺ elevation and Cl⁻ current development during tunneling compared to SOCE (n = 26 (B) and 9 (C); unpaired t test). (D) Enlarged confocal images show the colocalization of STIM1 and Orai1 at the PM focal plane. (E and F) Plots from a virtual line scan drawn across clusters in the PM plane (E, arrow in D) and in the z axis (F).

Figure 2.

Localization of SERCA2b and IP₃R1. (A) Immunostaining of SERCA2b and STIM1 at the whole cell level before (Control) and after store depletion with Thapsigragin. (B and D) High magnification views of the spatial organization of STIM1 and SERCA2b before (Ctr) and after store depletion (Tg) in the lateral dimension (x/y). Scale bar 500 nm. (C and E) Intensity profiles were obtained from virtual line scans in the x/y plane across ER tubules (C) and a STIM1 cluster (E). (F) Distance between the peak signals of mCh-STIM1 (center of the cluster) and the nearest SERCA2b maximum after store depletion as in panel D detected by immunofluorescence. The distance to the peak of GFP-Orai1 was used as a reference (n = 11–16; unpaired t test, P < 0.0001). (G) Pearson’s correlation coefficient (PCC) obtained before (Ctr) and after store depletion (Tg) between endogenous SERCA2b and endogenous STIM1. Colocalization at rest was measured in the middle of the cell and at the PM plane after store depletion (n = 9–11; unpaired t test, P = 0.0011). (H) Orthogonal section across the PM after store depletion illustrating the relative position and intensity of the STIM1 cluster (red) and SERCA2b (green). The arrows indicate the positions of the line analysis inside the cluster (I) and outside (O). The STIM1 and SERCA2b normalized intensities were measured over the z-axis after store depletion from inside and outside a STIM1 cluster. Scale bar 500 nm. (I) Quantification of the peak-to-peak distance in the z-axis between mCh-STIM1, Orai-GFP, and SERCA2b, inside and outside the STIM1 clusters (n = 13–16; one-way ANOVA, P < 0.0001). (J) 3D reconstruction of the relative localization of SERCA2b around the STIM1 cluster. (K) Top: Immunofluorescence for endogenous STIM1 (red) and IP₃R1 (green) after store depletion with Thapsigargin. A virtual line scan (arrow) across the high-magnification image shows the separation between the SOCE cluster and “licensed” IP₃R1s. Bottom: Immunofluorescence for endogenous IP₃R1s and KRAP at the PM focal plane. The intensity profile measured along the white arrow in the high magnification image indicates the high degree of colocalization of the “licensed” IP₃R1 and KRAP. (L) PCC summaries between STIM1, IP₃R1, and KRAP (n = 8–14; one way ANOVA, P < 0.0001). (M) Histogram of the relative frequency of the nearest neighbor distance (NND) between STIM1 clusters and IP₃R1 or KRAP (Outliers removed using the ROUT method and a Q value of 10%). (N) Violin plot comparing the distribution of NNDs between the center of STIM1 clusters and IP₃R1 or KRAP (n = 398–803; one-way ANOVA, P < 0.0001).

Figure S2.

Localization of SERCA2b and mCh-STIM1. (A) HeLa cells transfected with mCh-STIM1 and stained using a SERCA2b antibody. AiryScan images were taken before (Ctr) and after store depletion (Tg). Control images are acquired at a focal plane located in the middle of the cell and store-depleted images at the PM plane, where the STIM1 clusters localize. At rest, STIM1 and SERCA2b colocalize in the ER cisternae, while after store depletion they separate from each other at the PM focal place. (B) Pearson’s correlation coefficient (PCC) obtained before (Ctr) and after store depletion (Tg) between endogenous SERCA2b and overexpressed mCh-STIM1. Colocalization at rest was measured in the middle of the cell and at the PM plane after store depletion (n = 9–11; unpaired t test). (C) Image and line scan of STIM1 clusters on the side of the cell depicting its isolation from SERCA2b. (D) Cartoon illustrating the localization of STIM1 (red) and SERCA (green) after store depletion. The positions of the line scans in Fig. 2 H are also shown.

Figure S3.

Localization of endogenous IP₃R1, STIM1 and KRAP. (A) Relative localization of STIM1 and IP₃R1 detected by immunofluorescence in HeLa cells at rest. While both proteins share the same intracellular compartment, there is no overlap of the signals at the PM where the IP₃R1 fluorescence reveals the “licensed” receptors. (B) Colocalization at the PM focal plane of IP₃R1 and KRAP in HeLa cells at rest.

Figure S4.

Localization of IP₃R1, STIM1, and KRAP. (A) TIRF images of STIM1, KRAP, and IP₃R1 were detected by immunofluorescence in HeLa cells after store depletion. (B) Bar chart summarizing colocalization (PCC) of IP₃R1 with KRAP but not with STIM1 (n = 16–21; unpaired t test). (C) Airyscan images or “licensed” IP₃R1–GFP and KRAP detected by immunofluorescence at the PM plane (n = 6–12; unpaired t test). (D) Relative intensities were measured along the line indicated by the white arrow in C. (E) Colocalization (PCC) between KRAP and IP₃R1–GFP before (Ctr) and after store depletion (Tg). (F) Airyscan images of IP₃R1–GFP and mCh-STIM1 inside the cell (deep ER, in control conditions) and at the PM after store depletion (CPA/Cortical). (G) Relative intensities measured along the line indicated by the white arrow in F in cells at rest. (H) Colocalization (PCC) of IP₃R1–GFP and mCh-STIM1 before and after store depletion (CPA) (n = 7; paired t test). (I) Relative intensities along the line indicated by the white arrow in F in store depleted cells. (J) Lateral distance between the mCh-STIM clusters and either IP₃R1–GFP or endogenous KRAP after store depletion (n = 6–7; unpaired t test). (K) Example of an orthogonal section through a mCh-STIM1 cluster highlighting the localization of “licensed” IP₃R1–GFP outside the cluster and “Free” IP₃R deeper within the ER away from the STIM1 cluster as illustrated by the intensity plot. The cartoon indicates the distribution of STIM1 (red) after store depletion relative to “Licensed” (L) and “Free” (F) IP₃R1–GFP. The lateral distance between STIM1 clusters and licensed IP₃R1 as measured in panel J is indicated as D. (L) Axial distance between STIM1 clusters after store depletion and KRAP, “Free” IP₃R1–GFP, or “Licensed” IP₃R1–GFP, as indicated. “Free” IP₃R1 are receptors that localized deeper in the cell and do not colocalize with KRAP.

Figure 3.

Localization of PMCA4b and design of a tunneling inhibitor. (A) Distribution of GFP-PMCA4b and mCh-STIM1 in HeLa cells before (Ctr) and after (Tg) store depletion. The control confocal image is taken in the middle of the cell and images after store depletion at the PM plane. (B) Bar chart summarizing the PCC between mCh-STIM1, GFP-Orai1 (used as a maximum reference), and GFP-PMCA4b, measured in the orthogonal plane (n = 10–16; one-way ANOVA, P < 0.0001). (C) Quantification of the Z distance between mCh-STIM1 and GFP-PMCA4b peaks (n = 9–11; unpaired t test, P < 0.0001). (D and E) Intensity profiles along the x/y (D) and z (E) axis of mCh-STIM1 and GFP-PMCA4b across a STIM1 cluster (as indicated in A). (F) Cartoon summarizing the strategy to block cortical SERCA to specifically inhibit tunneling. (G) High magnification Airyscan images of GFP-MAPPER and mCh-STIM1 at rest (Ctr) and after store depletion (CPA). Line scans as indicated on the merged image show the colocalization of MAPPER (green) with the diffuse unclustered STIM1 (red) at rest, and the separation of clustered STIM1 from MAPPER. (H) Orthogonal sections of GFP-MAPPER and mCh-STIM1 before (Ctr) and after store depletion (CPA). The bar chart reports the distance in the z-axis between maximum intensities of GFP-MAPPER and mCh-STIM1 signals before (Ctr) and after store depletion (CPA) (n = 8–15; unpaired t test; P < 0.0001). (I) Cartoon depicting the structure of the Ca²⁺ tunneling attenuators (CaTAr1 and 2) compared to MAPPER. (J) Confocal images at the PM plane of CaTAr1 and mCh-STIM1 before (Ctr) and after (CPA) store depletion. The intensity plots using the white arrows in the high-magnification images report the colocalization of STIM1 and CaTAr1 at rest and their separation after store depletion. (K) Bar chart summarizing the PCC between CaTAr1 and mCh-STIM1 before (Ctr) and after store depletion (CPA) (n = 6; paired t test, P = 0.002). (L) Bar chart summarizing the distance between STIM1 clusters and CaTAr1 or MAPPER (MAP) (n = 6–7; unpaired t test, P = 0.91). Cartoons were generated using Biorender.

Figure S5.

Localization of MAPPER-S relative to STIM1. (A) Cartoon illustrating the spatial organization of the tunneling machinery. (B) Airyscan images of MAPPER–GFP (MAPPER) and mCh-STIM1 following store depletion with CPA in HeLa cells show that they do not colocalize. (C) Violin plots showing the nearest neighbor distance (NND) between the edges of the STIM1 and MAPPER clusters (n = 626–1,585, outliers removed using the ROUT routine, one-way ANOVA). (D) Relative localization of mCh-STIM1 and short MAPPER (MAPPER-S) in control conditions (Ctr) and after store depletion (Tg) at the whole cell level. (E) High magnification images of mCh-STIM1 and MAPPER-S clusters with the corresponding PCC (n = 11, paired t test) and line scan profile measured along the white arrow.

Figure S6.

Localization of MAPPER–GFP, E-Syt2-mCh, and TMEM24-mCh. (A) Localization of MAPPER–GFP, E-Syt2-mCh, and TMEM24-mCh at the whole cell level. (B) Bar chart summarizing the PCC of the three tethers relative to STIM1 (n = 7–23; one-way ANOVA). (C) Confocal images and line scans illustrate the colocalization of E-Syt2-mCh and TMEM24-mCh with STIM1-CFP clusters. The intensity plots are obtained from the lines depicted by the white arrows.

Figure S7.

CaTAr2 inhibits Ca²⁺ tunneling. (A) Localization of CaTAr2 relative to STIM1-CFP before (Ctr) and after store depletion (CPA). Colocalization analysis using either intensity plots along the white line or PCC measurements confirms that CaTAr2 localizes in a similar fashion to CaTAr1 and MAPPER relative to STIM1 clusters (n = 5–10; unpaired t test). (B) Colocalization of MAPPER–GFP and CaTAr2 at the whole cell level. (C) Localization and PCC between CaTAr2 and IP₃R1–GFP. The colocalization of CatAr2 with MAPPER was used as a reference value for the PCC analysis (n = 6–14; unpaired t test). (D) Confocal images of HeLa cells expressing CaTAr2 and loaded with Fluo4-AM. The numbered cells refer to the traces in E and F. (E) Ca²⁺ tunneling traces obtained after store depletion with CPA indicate that cells expressing CaTAr2 have a lower tunneling capacity. The bar chart on the right summarizes the inhibition of Ca²⁺ tunneling by CaTAr2. (F) Ca²⁺ release traces were obtained on the same cells as in E by applying histamine (His, 100 μM) after store refilling. The bar chart on the right summarizes the effect of CaTAr2 on Ca²⁺ release (n = 42–342; unpaired t test).

Figure S8.

Phospholamban and CaTArΔpolyK inhibit SERCA. (A) Localization of phospholamban-GFP (PLN-GFP) and mCh-STIM1 in HeLa cells before (Ctr) and after store depletion (Tg). (B) PLN-GFP transfection followed by loading cells with Calbryte 590. The numbered cells match the traces in C. (C) Ca²⁺ release induced by histamine (His, 100 μM) leads to Ca²⁺ release in cell#1, which does not express PLN-GFP; but not in cell#2, which expresses high levels of PLN-GFP. (D) Bar chart summarizing the inhibition of Ca²⁺ release in PLN-expressing cells. The recordings were performed in the presence of 1 mM lanthanum in the extracellular media to inhibit Ca²⁺ influx and recycling at the PM (n = 38–176; unpaired t test). (E) ER localization of CaTArΔpolyK (green), a CaTAr1 construct lacking the terminal poly-lysine domain that localizes it to ERPMCS. (F) Example of Calbryte intensity images before (Ctr) and after histamine addition (His). (G) Ca²⁺ release traces induced by histamine (His, 100 μM) in the two cells labeled in panel F. (H) Bar chart summarizing the inhibition of Ca²⁺ release in CaTArΔpolyK-expressing cells (n = 20–39; unpaired t test; P < 0.0001).

Figure 4.

CaTAr1 inhibits Ca²⁺ tunneling. (A) HeLa cells expressing GFP-MAPPER and loaded with the Ca²⁺ indicator Calbryte 590. (B) Changes in intracellular Ca²⁺ during a tunneling experiment from an untransfected cell (Ctr) and a cell expressing GFP-MAPPER (MAP). (C) Violin plots of the amplitude of the tunneling signal in Ctr and MAPPER expressing cells (MAP) (n = 439/118; 12 dishes; unpaired t test, P = 0.65). (D) Violin plot of the slope of the initial phase of tunneling in MAPPER expressing cells n = 427/110, unpaired t test, P = 0.0008). (E) Example averaged traces of Ca²⁺ release and SOCE in control (Ctr; no detectable CaTAr1 expression) and CatAr1 expressing cells from the same dish. Histamine-induced (His, 100 μM) Ca²⁺ release from stores was followed by CPA to deplete Ca²⁺ store, a wash period in Ca²⁺-free media to allow for SERCA to be active, and then the addition of Ca²⁺ to activate SOCE (n = 9/5). (F) Violin plots summarizing the levels of Ca²⁺ release in response to histamine in a Ca²⁺-free solution (n = 192/77; 5 dishes; unpaired t test, P = 0.085). (G and H) Enlarged traces from the red rectangle in E and violin plots summarizing the levels of SOCE in cells that did not express CatAr1 (Ctr) and CaTAr1 expressing cells (n = 192/77, unpaired t test, P = 0.10). (I) SOCE amplitude after thapsigargin application (1 µM) on control cells (Ctr) and cells expressing CaTAr1 (n = 125/51, 3 dishes, unpaired t test, P = 0.0175). (J) Quantification of NFAT nuclear translocation in Ctr and CaTAr1 expressing cells (n = 269/109, 5 dishes, unpaired t test, P = 0.58). (K) HeLa cells expressing CaTAr1 and loaded with Calbryte 590. The numbers correspond to the cells in L. (L) Traces showing the Ca²⁺ tunneling transient in response to His+Ca²⁺ following the standard store depletion with CPA and wash (not shown for clarity). A second histamine application after a delay in Ca²⁺-containing media confirms that all cells including those expressing CatAr1 refill their stores. (M and N) Violin plots summarizing the levels of Ca²⁺ tunneling (M, n = 262/83, six dishes, unpaired t test, P < 0.0001) and Ca²⁺ release in response to His following the second application in Ca²⁺-containing media (N, n = 182/56, five dishes, unpaired t test, P = 0.0250).

Figure S9.

CaTAr1 effects on tunneling and localization of ANO1–GFP and mCh-STIM1. (A) Dose-dependent inhibition of tunneling by CaTAr1. The amplitude of the tunneling signal is plotted as a function of the GFP signal normalized from minimum to maximum signal in each dish. The data is sorted in bins of 10%. (B) The relative localization of ANO1–GFP and mCh-STIM1 in NCL cells after store depletion is similar to the endogenous ANO1 channel and suggests the separation of ANO1 from the STIM1 cluster. (C) TIRF imaging of ANO1–GFP and mCh-STIM1 during store depletion with CPA confirmed the separation of the two proteins at the PM. Ctr and CPA images are from the same region of interest.

Figure 5.

Uncaging of IP₃ and histamine-induced clustering. (A) HeLa cells were loaded with caged-IP₃ together with Calbryte 590. Ca²⁺ stores were depleted using CPA and the CPA was removed by a 20 min wash. SOCE was triggered by the re-addition of Ca²⁺. After 20 s a UV flash was triggered to induce IP₃ release (arrow). Average example traces are represented with uncaging of IP₃ (red circles, from 97 cells) and without (blue circles, from 29 cells). (B) Single traces illustrating the change in the Ca²⁺ signal during uncaging (arrow, red trace) compared with SOCE only (blue trace) over an expanded time course around the uncaging pulse. (C) Bar chart summarizing the signal amplitude during SOCE and when tunneling is induced by IP₃ uncaging (n = 274/105, P < 0.0001). (D) Pilot experiment showing inhibition of IP₃-induced tunneling by CatAr1 expression. The relative normalized expression levels of CatAr1 are indicated on the x-axis (n = 66/26, P = 0.0007). (E) Airyscan images of HeLa cells expressing STIM1-Ch and CaTAr1 before and after application of histamine (100 µM). (F) Intensities obtained from virtual line scans in the x/y plane (arrows) across CaTAr1 (green) and STIM1 (red) signals before and after the application of histamine. Both signals are colocalized at rest but the CatAr1 segregates away from STIM1 clusters following histamine application. (F) Bar chart summarizing the colocalization intensity between CaTAr1 and STIM1 before (Ctr) and during histamine application (His) (n = 24, paired-t test, P = 0.0009).

Figure 6.

Tunneling in NCL-SG3 cells. (A) Localization of ANO1 by immunocytochemistry and mCh-STIM1 before (Ctr) and after store depletion (Tg). An intensity plot performed along a line (arrow) crossing a STIM1 cluster illustrates the separation of STIM1 and ANO1 at the PM focal plane. (B–D) Comparative localization at the PM plane of mCh-STIM1 with ANO1–GFP and the endogenous ANO1 protein (ANO1–Ig) following store depletion. As indicated by the PCC and the peak-to-peak distances the two proteins do not colocalize laterally but are localized at the same optical plane (z distance) (n = 8–18; one-way ANOVA, P < 0.0001 for B and C, P = 0.7 for D). (E) Intracellular Ca²⁺ elevation induced by trypsin application in NCL-SG3 cells loaded with Fluo4-AM. The application of the SOCE inhibitor BTP-2 (10 μM) reduces the amplitude and the duration of Ca²⁺ release (n = 379/402; unpaired t test, P < 0.0001). (F) Confocal images of NCL-SG3 cells expressing the Cl⁻ sensor mbYFPQS and mCh-STIM1 after store depletion. The orthogonal section through the cell indicates the PM localization of the chloride sensor. (G) Kymographs were measured during a tunneling event on cells expressing mbYFPQS and mCh-STIM1. The line passes through a SOCE cluster and an adjacent cell appendage labeled by mbYFPQS. The changes in Cl⁻ concentration (upper) are distal from the STIM1 cluster, which indicates the Ca²⁺ entry point. (H) Time course of the amplitude of the Cl⁻ signal induced by Ca²⁺ tunneling from NCL-SG3 cells in the same dish that either do not show any CaTAr2 expression (Ctr) and cells expressing CaTAr2. (I) Violin plots summarizing the amplitude of the Cl⁻ signal 3 min after trypsin and Ca²⁺ addition to stimulate tunneling (n = 32–33; unpaired t test, P = 0.0004).

Figure S10.

CaTAr2 and ANO1–GFP in NCL cells. (A) Airy scan images of the Co-expression of CaTAr2 and ANO1–GFP in NCL cells indicate the partial overlap between the two signals. (B) The colocalization is not influenced by store depletion. (C) Relative intensities along a virtual line scan (white arrows in A) reveal the colocalization of ANO1 and CaTAr2 but also enrichment in ANO1 at the edge of the CaTAr2-rich ER patch.

Figure 7.

Inhibition of sweating by CaTAr1. (A) GFP fluorescence in the mouse paw injected with CaTAr1 (right) compared with the contralateral paw (left) on the same animal. (B and C) Sweating was visualized using the iodine/starch technique in GFP- and in CatAr1-expressing paws. (D) Bar chart summarizing the sweating areas recorded after 15 min in control (GFP) and CaTAr1 expressing paws (n = 13–15; unpaired t test, P = 0.0074). (E) Cartoons illustrating the different phases of a typical agonist-driven Ca²⁺ release signal with (Ctr) or without SOCE. The activation state and localization of the Ca²⁺ effectors are shown for the different phases. (F) To scale illustration of the Ca²⁺ effectors involved in tunneling. For STIM1 only the domains with the atomic structure solved are shown. (G) To scale depiction of 5 Orai1 channels and clusters of 125 SERCAs each.