An INF2-dependent actin-mediated step in Inositol 1,4,5-trisphosphate receptor cluster formation and activity

Summary

Intracellular calcium signaling plays a vital role in regulating various cellular processes including gene regulation, motility, metabolism and cell death. Inositol 1,4,5-trisphosphate receptors (IP3R) on the Endoplasmic Reticulum (ER) are a major cation channel that regulates stimulus-induced calcium release from the ER. While several molecular players regulate activity of IP3R, its regulation by actin filaments were uncharacterized. Here we show that actin filaments polymerized by a specific actin nucleator INF2 facilitates agonist-induced IP3R activity. Our results demonstrate that INF2-mediated actin filaments regulate formation and/or stability of IP3R clusters on the ER that have been previously shown to be hotspots of ER calcium release. Using cell-biological and biochemical techniques we further show that INF2 physically interacts with IP3R isoforms, often at IP3R clusters. While INF2-IP3R interaction is independent of INF2-activity, the ability of INF2 to mediate IP3R clusters is dependent on its actin polymerization activity. Finally, we demonstrate that in addition to its calcium mobilization activity, INF2 on ER specifically regulates IP3R cluster positioning to mediate ER-mitochondrial contacts and facilitate ER to mitochondrial calcium transfer. Overall, these results reveal an actin-dependent step in regulation of IP3R activity both in terms of ER calcium release and modulation of ER-mitochondrial contacts.

eTOC blurb

Zavala et al. show that depletion of the actin assembly factor INF2 reduces agonist-induced IP3R-mediated ER calcium release through reduction of IP3R cluster assembly on the ER in an INF2-activity- and actin-dependent manner. Additionally, ER-bound INF2 regulates IP3R cluster positioning to promote ER-mitochondrial contacts and facilitate ER-mitochondrial calcium transfer.

Introduction

Calcium (Ca²⁺) ions operating as secondary messengers, play a fundamental role in intracellular signaling regulating various processes like cellular and mitochondrial metabolism ^1,2, cell motility ³, apoptosis ⁴ and vesicular trafficking ⁵ through integration of a complex Ca²⁺ signaling toolkit. Though an increase in intracellular Ca²⁺ levels above the resting 50-100 nM either through external cues ⁶ or spontaneous internal cues ^7,8 often serve as catalyst to promote Ca²⁺ signaling events, absence of Ca²⁺ spike can also serve as platforms of regulation as shown in certain specific transcriptional repression events ⁹, quiescence ¹⁰ and autophagy ¹¹. Two principal Ca²⁺ sources that cells rely on for signaling events are either Ca²⁺ influx from the extracellular milieu or intracellular Ca²⁺ released by inositol 1,4,5-trisphosphate receptors (IP3Rs) ¹², ryanodine receptors (RyRs) ¹³, two-pore channels (TPCs) ^14–16 or the mucolipin subfamily of transient receptor potential channels (TRPML) ¹⁷. IP3Rs in the non-excitable cells and neurons act as the major Ca²⁺ channel, efflux from which regulate intracellular Ca²⁺ signaling ¹⁸.

IP3Rs are large-conductance cation channels expressed in most cells that mediate Ca²⁺ release from intracellular stores primarily ER ¹⁹ and the Golgi apparatus ²⁰. IP3Rs are activated by IP3 and Ca²⁺ ion itself ^21,22. Three separate genes encode three different isoforms of IP3Rs (IP3R1, IP3R2 and IP3R3) having heterogenous tissue specificity, distinct activation and deactivation patterns, sensitivity to IP3 and ability to mediate ER to mitochondrial contact (ERMC) and Ca²⁺ transfer ²³.

While IP3Rs have been shown to be arranged in mobile and immobile clusters with Ca²⁺ signaling ¹² and ERMC ²³ occurring at these clusters, little is known about the underlying mechanisms that regulate these IP3R clusters. Among several factors that regulate IP3R channel activity and function¹², actin filaments have also been implicated in IP3R-mediated Ca²⁺ release ²⁴ from the ER, the mechanism of which still remains uncharacterized.

INF2 is an actin assembly factor that exists in two isomeric forms (ER-bound and freely cytosolic). Actin filaments generated by ER-bound INF2 have been implicated in mitochondrial division ²⁵ and ER to mitochondrial Ca²⁺ transfer ²⁶. INF2 is activated by raising intracellular Ca²⁺ levels in a Ca²⁺-calmodulin dependent manner ^27–31 probably relieving the auto-inhibition between its N-terminal region (the DID, Diaphanous Inhibitory Domain) with a C-terminal sequence (the DAD, Diaphanous Autoregulatory Domain) ²⁷ thereby activating it. Upon activation, INF2-caax mediated actin filaments on the ER mediates close contacts between ER and mitochondria, facilitating Ca²⁺ transfer from ER to mitochondria that promotes division of the inner mitochondrial membrane (IMM) kinetically distinct from Drp1-mediated division of outer mitochondrial membrane (OMM) ²⁶. While INF2-mediated actin filaments have been explored more for their role in mitochondrial division, little is known about its role on ER function.

In this study we demonstrate that INF2-mediated actin filaments regulate agonist-induced IP3R-mediated ER Ca²⁺ release. Acute and chronic depletion of INF2 significantly reduced histamine, carbachol and ATP-induced ER Ca²⁺ release measured both at the level of cytosolic Ca²⁺ ([Ca²⁺]_C) and ER Ca²⁺ [Ca²⁺]_ER transients. Both isoforms of INF2 could rescue the Ca²⁺ release in the INF2-depleted cells suggesting that ER localization was dispensable for this function. We further showed that loss of INF2 affects the activity of all the three IP3R isoforms. Through multiple approaches we showed that INF2 interacted with IP3R in an activity-independent manner. We further demonstrate that depletion of INF2 significantly reduced assembly of IP3R clusters which could be rescued only by active INF2. Finally, we showed that ER localization of INF2 was imperative to position IP3R2 clusters close to mitochondria which promoted IP3R2-mediated ERMC and Ca²⁺ transfer to mitochondria.

Results

INF2 depletion decreases IP3R-mediated ER calcium release

To study the role of INF2 polymerized actin on IP3R-mediated ER Ca²⁺ release, we created CRISPR-mediated INF2-KO in HeLa cells. As shown before, INF2-KO cells (Figure S1A) lacked Ca²⁺-mediated “actin burst” (Figure S1B). Since the resting cytosolic calcium ([Ca²⁺]_C) levels showed no difference between the two groups (Figures S1C and S1D), we measured agonist-induced [Ca²⁺]_C transients using the Cyto-R-Geco calcium probe. 100 μM histamine induced a significantly higher [Ca²⁺]_C transient in the control cells (11.3 ± 3.53-fold) compared to the INF2-KO cells (7.8 ± 2.51-fold) (Figures 1A and 1B). Further we used two sub-maximal doses of histamine (1 μM and 0.1 μM) which showed significantly higher [Ca²⁺]_C transients in the control cells compared to the INF2-KO cells in both the concentrations tested (Figures 1A and 1B). Interestingly, 0.1 μM histamine in the control cells yielded a similar [Ca²⁺]_C transient (3.8 ± 2.1-fold) as 1μM histamine stimulation in the INF2-KO cells (3.4 ± 1.8-fold). Next this effect was evaluated upon acute depletion (siRNA-mediated KD) of INF2 (Figure S1A). INF2 KD cells had unchanged resting [Ca²⁺]_C (Figures S1C and S1D) and showed notably lower [Ca²⁺]_C transients compared to the control cells when stimulated by both maximal (100 μM) and sub-maximal (1 μM) doses of histamine (Figures 1C and 1D). Interestingly, a significantly higher number of INF2-depleted cells did not respond to the sub-maximal doses of histamine compared to the control cells suggesting that INF2 depleted cells are insensitive to these sub-maximal histamine stimulations (Figure S1E). Faster recording at 1frame/sec revealed that in addition to a decrease in amplitude, INF2 depleted cells not only took longer time to respond but also displayed a slower [Ca²⁺]_C transient upon 1 μM histamine (Figure 1E–1H). To test whether the effects were specific for histamine stimulations, we stimulated control and INF2-KO cells with 100 μM ATP that uses endogenous purinergic receptors. Similar to histamine stimulations, 100 μM ATP stimulations also elicited significantly lower [Ca²⁺]_C transients in INF2-KO cells (Figures S1F and S1G).We next directly evaluated kinetics of ER calcium ([Ca²⁺]_ER) release in control and INF2 depleted cells. Control and INF2 KO (Figures 1I–1J; 1M–1N) or INF2 KD (Figures 1K–1N) cells were transfected with the high affinity luminal ER-GCamp150 probe and stimulated with either 100 μM histamine (Figures 1I–1L) or 1 μM histamine (Figures 1M–1N). While control cells show about 73% acute decrease in ER Ca²⁺ content upon 100 μM histamine stimulation (Figures 1I–1L), both INF2-KO cells (46.7 ± 26 %) and INF2-KD cells (34.6 ± 24 %) showed significantly lower Ca²⁺ release from ER (Figures 1I–1L). A similar trend was noted for the sub-maximal stimulation where control cells showed 56.1 ± 21% decrease in ER Ca²⁺ upon stimulation, significantly higher than both INF2-KO (11.7 ± 18.9%) and INF2-KD (18.9 ± 18.2 %) cells (Figures 1M–1N). We explored this further using a low-sensitive ER-LAR-Geco Ca²⁺ probe having a Kd of 24 μM. Using this probe the control cells showed 28.9 ± 9.5 % acute [Ca²⁺]_ER release which was significantly stronger than both INF2-KO (1.78 ± 3.5%) and INF2-KD (5.47 ± 4.7%) cells (Figures S1H and S1I). We also measured the abundance of ER Ca²⁺ stores through 2μM Thapsigargin treatments which showed similar [Ca²⁺]_C transients among the control and INF2-depleted (KO and KD) cells signifying that the ER stores were not pre-exhausted upon INF2 depletion (Figure S1C, S1J). Cumulatively, these results confirm that both acute and chronic depletion of INF2 significantly subdue agonist-induced IP3R-mediated ER Ca²⁺ release measured at the level of cytosolic and ER Ca²⁺ transients.

Figure 1. Depletion of INF2 reduces IP3R-mediated calcium release.

A: Traces of Cytosolic calcium transients in HeLa WT and INF2-KO cells transfected with Cyto R-GECO plasmid (Kd=0.5 μM), stimulated with different concentrations of histamine (100 μM, 1 μM and 0.1 μM).

B: Quantification of maximal peak of cytosolic Ca²⁺ release as shown in A, expressed as fold change F/F0. Data from 80-130 cells across 3 - 4 independent experiments. Graphs shows Mean ± S.D. Two-way ANOVA used. **** p <0.0001, *** p=0.0005, * p=0.0133.

C: Traces of Cytosolic calcium transients in HeLa WT and INF2-KD cells transfected with Cyto R-GECO plasmid (Kd=0.5 μM), stimulated with different concentrations of histamine (100 μM and 1 μM).

D: Quantification of maximal peak of cytosolic Ca²⁺ release as shown in C, expressed as fold change F/F0. Data from 90-160 cells across 3 - 4 independent experiments. Graphs shows Mean ± S.D. **** p <0.0001, ** p=0.0034.

E: Traces of Cytosolic calcium transients in HeLa WT, INF2-KD and INF2-KO HeLa cells transfected with Cyto R-GECO plasmid (Kd=0.5 μM), stimulated with 1 μM histamine and recorded at 1 frame/sec.

F: Quantification of the time delay between histamine stimulation onset of cytosolic calcium transients in WT, INF2 KD and INF2 KO HeLa cells imaged in Fig 1E. Data from 58 (WT), 32 (KD) and 48 (KD) cells across 3 independent experiments in dataset shown in 1E. Graph shown Mean ± SD. **** p <0.0001, * p=0.0295.

G: Quantification of slope of cytoplasmic calcium rise, expressed as fold change/sec from Dataset shown in Fig 1E in WT, INF2 KD and INF2 KO HeLa cells. Data from 58 (WT), 32 (KD) and 48 (KD) cells across 3 independent experiments. Graph shown Mean ± SD. **** p <0.0001

H: Quantification of time taken for cytoplasmic calcium spike to reach the max value from Dataset shown in Fig 1E in WT, INF2 KD and INF2 KO HeLa cells. Data from 58 (WT), 32 (KD) and 48 (KD) cells across 3 independent experiments. Graph shown Mean ± SD. **** p <0.0001

I: Traces of ER calcium transients in HeLa WT and INF2-KO cells transfected with ER-GCaMP6-150 plasmid (Kd=150 μM), stimulated with histamine (100 μM).

J: Quantification of ER calcium expressed as % of decrease in 10 sec from traces shown in E. Data from 129 (WT) and 102 (INF2-KO) cells from 3 independent experiments in each case. Graphs shows Mean ± S.D. unpaired Student’s t-test used. **** p<0.0001.

K: Traces of ER calcium transients in HeLa WT and INF2-KD cells transfected with ER-GCaMP6-150 plasmid (Kd=150 μM), stimulated with histamine (100 μM).

L: Quantification of ER calcium expressed as % of decrease in 10 sec from traces shown in G.Data from 74 cells (WT) and 98 cells (INF2-KD) across 3 independent experiments. Unpaired student’s t-test used. **** p<0.0001.

M: Traces of ER calcium transients in HeLa WT, INF2-KO and INF2-KD cells transfected with ER G-CaMP6-150 plasmid (Kd=150 μM), stimulated with1uM histamine treatment.

N: Quantification of ER calcium release from I, expressed as % of decrease in 5 sec. Data from 55 (WT), 60 (INF2-KD) and 37 (INF2-KO) cells from 3 independent experiments. Graphs shows Mean ± S.D. Significant was measured using One-way ANOVA. **** p<0.0001.

See also Figure S1 and S2.

ER localization of INF2 is dispensable for IP3R-mediated calcium release

In cells, INF2 is expressed in two isomeric forms: ER-bound INF2-caax and cytosolic INF2-non-caax. We next assessed whether ER residency of INF2 was necessary for ER Ca²⁺ release. We re-expressed GFP-INF2-caax (Figures 2A and 2B; S1K) and GFP-INF2-noncaax (Figures 2C and 2D; Figure S1K) in HeLa INF2-KO cells and measured [Ca²⁺]_C transients upon 100 μM histamine stimulation. Both INF2-caax (ER) (Figures 2A and 2B) and INF2-noncaax (cyto) (Figures 2C and 2D) significantly rescued histamine-induced [Ca²⁺]_C transients in the INF2-KO cells. Similar rescue of [Ca²⁺]_ER transients were also seen in INF2-KO cells re-expressing either INF2-caax or INF2-noncaax (Figures 2E and 2F), further confirming that ER residency of INF2 was not required to mediate IP3R-mediated ER Ca²⁺ release. Interestingly we noticed that while baseline resting [Ca²⁺]_C in INF2-rescue cells remained unaltered compared to the control cells (Figures S1C and S1D), 2 μM Thapsigargin induced [Ca²⁺]_C transients revealed slightly depleted ER stores (Figures S1C and S1J). However, the mildly depleted ER stores in the INF2-rescue conditions did not alter histamine-induced [Ca²⁺]_C or [Ca²⁺]_ER transients. Though INF2 in cells is mostly auto-inhibited due to the intrinsic DID (N-term)-DAD (C-term) binding and therefore tightly regulated, we still wanted to evaluate whether overexpression of INF2-isoforms positively regulated histamine-induced IP3R-mediated [Ca²⁺]_ER transients. To do this, we transfected either INF2-caax or INF2-non-caax in HeLa WT cells (Figure S1L) and measured histamine-induced [Ca²⁺]_C or [Ca²⁺]_ER transients using low-affinity probes Cyto-LAR-Geco (Kd 12 uM) (Figures 2G and 2H) and ER-LAR-Geco (Kd 24 uM) (Figures 2I and 2J) respectively. Overexpression of INF2 isoforms significantly enhanced both [Ca²⁺]_C or [Ca²⁺]_ER transients compared to the un-transfected control cells but did not alter the resting [Ca²⁺]_C or [Ca²⁺]_ER levels (Figures S1M–S1O). These results suggest that abundance of either of the INF2 isoforms significantly regulate histamine-induced IP3R-mediated [Ca²⁺]_C and [Ca²⁺]_ER transients in HeLa cells.

Figure 2. ER localization of INF2 is dispensable for IP3R-mediated calcium release.

A: Traces of Cytosolic calcium transients in HeLa WT, INF2-KO and INF2-caax rescued cells transfected with Cyto R-GECO plasmid (Kd=0.5 μM), stimulated with 100 μM histamine.

B: Quantification of maximal peak of cytosolic Ca²⁺ release as shown in A, expressed as fold change F/F0. Data from 98 (WT), 118 (INF2-KO) and 191 (IN2-KO + INF2-caax) cells across 3 - 4 independent experiments. One way ANOVA used. **** p<0.0001 * p=0.0447

C: Traces of Cytosolic calcium transients in HeLa WT, INF2-KO and INF2-noncaax rescued cells transfected with Cyto R-GECO plasmid (Kd=0.5 μM), stimulated with 100 μM histamine.

D: Quantification of maximal peak of cytosolic Ca²⁺ release as shown in C, expressed as fold change F/F0. Data from 97 (WT), 61 (INF2-KO) and 109 (IN2-KO + INF2-noncaax) cells across 3 independent experiments. One way ANOVA used. **** p<0.0001 * p=0.0318

E: Traces of ER calcium transients in HeLa WT, INF2-KO and INF2-KO rescue cells with INF2-Caax or INF2–noncaax co-transfected with ER-LAR-GECO (Kd=24 μM), stimulated with 100 μM Histamine treatment.

F: Graph quantifying changes in ER calcium release upon the Histamine treatment, expressed as % of depletion at 5 sec. Data from 93 (WT), 172 (INF2-KO), 184 (INF2-KO + INF2-caax) and 146 (INF2-KO + INF2-noncaax) cells across 3 independent experiments. Graph shows Mean ± S.D. One-way ANOVA used. **** p<0.0001, *** p=0.0004.

G: Traces of cytosolic transients in HeLa WT cells either alone or over-expressing INF2-caax or INF2-noncaax co-transfected with Cyto LAR-GECO (Kd=12 μM) probe, stimulated with 100 μM histamine

H: Quantification of maximal peak of cytosolic calcium from G, expressed as F/F0. Data from 79 (WT), 75 (WT + INF2-caax) and 32 (WT + INF2-noncaax) cells across 3 independent experiments Graphs shows Mean ± S.D. Significant was measured using One-way ANOVA. **** p<0.0001, *p=0.0483.

I: Traces of ER calcium transients in HeLa WT cells either alone or over-expressing INF2-caax or INF2-noncaax co-transfected with ER-LAR-GECO probe (Kd=24 μM), stimulated with after 100 μM histamine.

J: Quantification of ER calcium release from I, expressed as % of decrease in 5 sec. Data from 84 (WT), 42 (WT + INF2-caax) and 60 (WT + INF2-noncaax) cells across 3 independent experiments Graph shows Mean ± S.D. Significant was measured using One-way ANOVA. **** p<0.0001

See also Figure S1 and S2.

Depletion of INF2 affects the function of all IP3R isoforms

Since HeLa cells express all the three isoforms of IP3R (IP3R1, IP3R2 and IP3R3), it is unclear whether INF2 has a functional bias towards any of the IP3R isoforms. To address this, we utilized previously characterized HEK-293 cell lines expressing either IP3R1 (IP3R2/3 DKO) or IP3R2 (IP3R1/3 DKO) or IP3R3 (IP3R1/2 DKO) ³². Acute depletion of INF2 (Figure S2A) significantly reduced carbachol-induced [Ca²⁺]_C transients mediated by either IP3R1 (Figures S2B and S2C) or IP3R2 (Figures S2D and S2E) or IP3R3 (Figures S2F and S2G). Further, we created CRISPR-INF2-KO in either IP3R2 or IP3R3 expressing HEK cell lines and validated them for the absence of INF2 (Figure S2A). Similar to results obtained upon acute depletion, INF2 KO in either IP3R2 (Figures S2D and S2E) or IP3R3 (Figures S2F and S2G) expressing cell lines significantly dampened carbachol induced [Ca²⁺]_C transients. Similar to results obtained using HeLa cells, re-introduction of either of the INF2 isoforms completely rescued both [Ca²⁺]_C (Figures S2H and S3I) and [Ca²⁺]_ER (Figures S2J and S2K) transients in INF2 depleted IP3R2 HEK cells. Collectively these results confirm that loss of INF2 affects [Ca²⁺]_C and [Ca²⁺]_ER transients induced by all IP3R isoforms individually.

INF2 interacts with IP3R-isoforms independent of its actin polymerization activity

We wanted to test whether regulation of IP3R function by INF2 is mediated through a direct interaction or is an effect of a signaling cascade. Proximity Ligation Assay (PLA) against IP3R2/INF2 and IP3R3/INF2 using validated IP3R2 and IP3R3 antibodies (Figure S3A) in HeLa control and INF2-KO cells revealed significant proximity of both IP3R2 (Figures 3A and 3B) and IP3R3 (Figures S3B and S3C) with INF2 in control cells. Re-introduction of INF2 isoforms in INF2-KO cells significantly rescued the density of PLA dots for both IP3R2 (Figures 3A and 3B) and IP3R3 (Figures S3B and S3C). Interestingly there was significantly increased density of PLA dots between IP3R3- INF2-caax compared to IP3R3-INF2-non-caax suggesting that INF2-caax on the ER might preferentially interact better with IP3R3 (Figures S3B and S3C). This phenotype was not relevant for IP3R2-INF2 interactions (Figure 3A and 3B).

Figure 3. INF2 interacts with IP3R-isoforms independent of its actin polymerization activity.

A: Representative fixed-cell images from Proximity Ligation Assay (PLA) against IP3R2 and INF2 in HeLa WT, INF2-KO and INF2-KO rescue cells. WT and INF2-KO cells were transfected with GFP-ERSec61β to label ER (magenta) prior to PLA staining. PLA foci shown in green. Scale: 10 μm.

B: Quantification of PLA density expressed as number of dots per 100 μm² (left), and PLA dots, expressed as number/cell (right). Data from 48 (WT), 20 (INF2-KO), 60 (INF2-KO+INF2-caax) and 44 (INF2-KO+INF2-noncaax) cells from 3 independent experiments Graph shows Mean ± S.D. One-way ANOVA used. **** p<0.0001, *** p=0.0001 (PLA density, left); **** p<0.0001, ** p=0.0012 (PLA dots, right).

C: Illustration of INF2 domains structure, showing the WT and with different mutations (C1270S-caax: cytosolic localization; FFC caax (inactive): ΔDID/ I643A/K792A and I676A caax (partially active), S1099A and S1100A.

References: DID, Diaphanous Inhibitory Domain; FH1, Formin Homology 1; FH2, Formin Homology 1, WH2/DAD, Diaphanous Autoregulatory Domain.

D: Western blot images of candidate proteins from GFP-trap pull-down assay, in HEK IP3R2 expressing INF2-KO cells transfected with GFP-tagged INF2-Caax or INF2-noncaax.

E: Western blot images of candidate proteins from GFP-trap pull-down assay, in HEK IP3R2 expressing INF2-KO cells transfected with GFP-tagged INF2 constructs (WT and mutants).

F: Western blot images of candidate proteins from GFP-trap pull-down assay, in HEK IP3R2 expressing INF2-KO cells transfected with GFP-tagged INF2 constructs (WT and serine-mutants).

G: Workflow of APEX2-labelling biotinylating assay using mass spectroscopy, to identify IP3R1 proximity partners in HEK-IP3R TKO cells transfected with either APEX2-IP3R1 or APEX2-ER (used as control) constructs.

H: Plot of known interactors of IPR31, showing INF2. Fold change was calculated based on enrichment over control (ERM-APEX2). An ‘imputed’ averaged value from all samples was used to calculate fold enrichment. This accounts for the plateau in the Y-axis.

I: Graph showing significant enrichment of proteins involved in regulation of actin cytoskeleton.

J: Representative western blot images from streptavidin pull-downs following APEX2 labelling as shown in assay in panel G in different conditions.

K: Illustration of IP3R domains showing the two different mutations: IP3R1-D2550A pore dead mutant and IP3R1-R265Q, ligand binding motif mutant. References: SD, Suppressor domain; IP3R BD, IP3R binding domain; RD, Regulatory domain; TD, Transmembrane domain; C-TD: Carboxyterminal domain.

L: Western blot images of candidate proteins from GFP-trap pull-down assay performed in HEK IP3R-TKO-INF2-KO cells re-expressed with IP3R1 (WT/mutants) and GFP-INF2-caax.

See also Figure S3 and Data S1.

To further validate this interaction, we carried out pull-down assays using GFP-trap agarose beads. GFP-tagged INF2 isoforms (caax and non-caax) (Figure 3C) were transfected into IP3R2 expressing-INF2-KO HEK cells for 24 hours, lysed and pulled down using GFP-trap beads. Endogenous IP3R2 was detected in the eluates of both INF2-caax and INF2-non-caax (albeit at lower abundance) suggesting an interaction between the two proteins. We also detected β-actin (positive control) and IP3R interactor VDAC in the eluates of both INF2 isoforms (Figure 3D). To test whether this interaction is actin-dependent, cells expressing GFP-INF2 isoforms (caax and non-caax) were treated with 0.5 μM latrunculin A for 15 mins prior to lysis and GFP-trap pull-down. While IP3R2 interactions with INF2-caax was not affected by actin depolymerization, interaction with INF2-non-caax was significantly reduced (Figure S3D). Similar fate was also observed for VDAC (Figure S3D). However, we still detected a significant amount of β-actin in the pull-down eluates which might either be actin monomers being brought down by INF2 isoforms or stable actin filaments unperturbed by latrunculin A treatment. To further test the relevance of actin filaments in IP3R2-INF2 interactions, we used two mutant INF2-caax constructs: (1) N-term DID domain truncated FFC construct bearing two mutations, I643A and K792A, in the FH2 domain that makes it inactive; (2) a full-length INF2-caax bearing I676A mutation that makes it partially active (Figures 3C; S1B). While the FFC-caax mutant did not interact with β-actin, I676A mutant had significantly lower β-actin in the pull-down eluates (Figure 3E). IP3R2 was detected in the eluates of both these constructs at levels similar to those from INF2-caax (WT), INF2-non-caax and a cytosol localized INF2-caax C1270S mutant (Figure 3E). These results further strengthened the fact that IP3R2-INF2 interaction was independent of actin filaments. INF2 was recently shown to be phosphorylated at the C-terminal (S1077 (human)/ S1099 (mouse)) by AMPK that regulates mitochondrial fission ³³. Moreover, it was also shown to interact with another ER integral membrane protein, MAL, through the C-terminal 200 aa ³⁴ suggesting that the C-terminal region might be an important interacting hub for INF2. To identify whether the C-terminal S1099 and 1100 was important for its interaction with IP3R, we created an INF2- caax bearing S1099A and S1100A mutations (Figure 3C). Expressing them in INF2-KO cells and performing the GFP-trap assay revealed that both mutants tested were significantly impaired for its interaction with IP3R2 (Figures 3F; S3E). Thus, it is evident that either the phosphorylation at S1099/1100 or the serine residues per se at the C-terminal of INF2 facilitate its binding with IP3R2. Collectively these results confirm an actin-independent interaction between INF2-caax and endogenous IP3R2.

Taking a more unbiased approach, we performed an APEX2 labelling assay to identify binding partners for IP3R1 in HEK-293 cells. APEX2-IP3R1 and an APEX2-ER construct was transfected into HEK-293-TKO (IP3R1/2/3 KO) cells, incubated with biotin, pulsed with 1 mM H₂O₂ for 1 min, quenched, lysed and pulled-down with streptavidin agarose beads. The precipitated molecules were analyzed using LC-MS/MS (Figure 3G). Several known interactors were identified and satisfyingly endogenous INF2 emerged as a potent interactor of IP3R1 (Figure 3H; Data S1). Additional cytoskeletal proteins identified include Arp2/3 components mDia 1/2, Spectrin, α-actinin and Filamin A (Data S1). In fact, further analysis revealed a significant enrichment of proteins involved in regulation of the actin cytoskeleton (Figure 3I). We validated some of these targets. INF2, IRBIT and α-actinin were validated in immunoblots both in control and latrunculin A treated cells suggesting an actin-filament independent binding of these molecules with IP3R1 (Figure 3J). While a high salt elution weakened the abundance of IRBIT and α-actinin in the eluate, the abundance of INF2 was not affected, suggesting a rather strong interaction with IP3R1 (Figure 3J). Finally, we tested the ability of IP3R1 mutants compromised for IP3-binding (R1-R265Q) or channel function (R1-D2550A) to interact with INF2. WT-IP3R1 and the mutants (Figure 3K) were transfected into IP3R TKO INF2-KO HEK cells along with GFP-INF2-caax. GFP-trap precipitation showed that the IP3-binding and pore forming ability is dispensable for IP3R1 binding with INF2 (Figure 3L). Therefore, we conclude through various approaches that INF2-caax interacts with IP3R isoforms in an actin-filament independent way which also does not require IP3R activity.

INF2 regulates formation of functionally competent IP3R clusters in cells

Since INF2 depletion significantly affected agonist-induced (histamine, ATP and carbachol) IP3R-mediated Ca²⁺ transients, we wanted to decipher the molecular mechanism underlying this phenotype. Possible mechanism(s) include insufficient production of agonist stimulated IP3 molecules, altered abundance of IP3R on the ER or changes in IP3R organization on the ER. To test out the first possibility, we transfected YFP-tagged PH-PLC-δ construct ³⁵ that binds to PIP2 at the plasma membrane. Agonist-induced hydrolysis of PIP2 and IP3 production allows re-localization of the probe to cytosol which we use as a surrogate to measure IP3 production. 100 μM histamine and carbachol treatments of control and INF2-KO HeLa (Figures S4A and S4B) and IP3R2-HEK cells (Figures S4A and S4C) respectively, showed similar kinetics of the probe re-localization indicating that the abundance and kinetics of IP3 production might not be affected by INF2 depletion. Next, we looked at IP3R abundance upon INF2 depletion. While IP3R2 abundance seemed to increase slightly in HeLa INF2-KO cells, abundance of IP3R1 and IP3R3 remained unchanged in the INF2-depleted cells (Figure 4A). Similarly, IP3R2 (Figure 4B) and IP3R3 (Figure S4D) abundance in IP3R2 expressing-INF2-KO HEK cells and IP3R3 expressing-INF2-KO HEK cells respectively were also unchanged compared to their respective control cells. These results further shows that chronic depletion of INF2 does not affect the abundance of IP3R isoforms in either HeLa or HEK cells.

Figure 4. INF2 regulates formation of functionally competent IP3R clusters in cells.

A: Western blots showing the expression of all IP3R isoforms in WT and INF2-KO HeLa cells. β-Actin was used as loading control.

B: Western blots showing expression of IP3R2 in HEK IP3R2 expressing (R1/R3 DKO) WT or INF2-KO.

C: Representative confocal micrographs of HeLa WT, INF2-KO and INF2-KO rescue cells (transfected with GFP-INF2-Caax or GFP-INF2-non-caax), immunoassayed for INF2 (green) and IP3R2 (magenta). IP3R2 clusters after thresholding shown in the right-most panel. Scale: 10 μm.

D: Quantification of IP3R2 clusters, expressed as number of clusters per 100 μm² of cell area (from dataset in C). Data from 32 (WT), 22 (INF2-KO), 23 (INF2-KO+INF2-caax) and 27 (INF2-KO+INF2-noncaax) from 2 independent experiments.

E: Representative confocal micrographs of HEK-IP3R2 expressing cells (WT, INF2-KO and rescues), immunoassayed for INF2 (green) and IP3R2 (magenta). IP3R2 clusters after thresholding shown in the right-most panel. Scale: 10 μm.

F: Quantification of IP3R2 clusters, expressed as number of clusters per 100 μm² of cell area (from data set in E). Data from 47 (WT), 49 (INF2-KO), 32 (INF2-KO+INF2-caax) and 40 (INF2-KO+INF2-non-caax) cells from 2 independent experiments Graphs shows Mean ± S.D. Significant was measured using One-way ANOVA. **** p<0.0001.

G: Representative confocal micrographs fixed-HEK IP3R2 expressing INF2-KO cells transfected with either GFP-INF2-Caax (WT) or GFP-INF2 FFC caax (inactive), immunoassayed for IP3R2 (magenta). Scale: 10 μm.

H: Quantification of IP3R2 clusters, expressed as number of clusters per 100 μm² (from data set in G). Data of 18 cells for (INF2-caax-WT) and 39 cells (FFC caax-inactive) from 2 independent experiments. Graphs show Mean ± S.D. unpaired Student’s t-test used. **** p<0.0001.

I: Cytosolic Ca²⁺ transients induced by 1 μM histamine in Cyto R-GECO transfected HeLa WT (red), INF2-KO (green), INF2-KO + INF2-caax WT (purple), INF2-FFC-caax inactive mutant (orange) or INF2-caax-S1099A binding mutant (yellow).

J: Quantification of maximal peak of Cyto Ca²⁺ from I, expressed as F/F0. Data from 130 (WT), 215 (INF2-KO), 169 (INF2-KO + INF2 caax), 167 (INF2-KO + INF2-FFC-caax) and 196 (INF2-KO + INF2-caax S1099A) cells across 5 independent experiments. Graphs show Mean ± S.D. One-way Anova used. **** p<0.0001, * p= 0.0427 (WT vs. + FFC caax), * p=0.0110 (WT vs. + S1099A).

See also Figure S4.

Next, we performed immunofluorescence staining of IP3R2 in cells. Fixed cell analysis of IP3R2 and IP3R3 in HEK cells showed a faint homogenous staining of endogenous IP3R isoforms (Figure S3A) on the ER with intermittent bright immunoreactive cluster also localized to the ER (Figure S3A). Immunoreactive signals for both IP3R2 and IP3R3 was not detected outside the ER signal (Figure S3A). Surprisingly depletion of INF2 drastically decreased the density of the bright immunoreactive cluster of endogenous IP3R2 in HeLa cells (Figure 4C and 4D) and in IP3R2-expressing HEK cells (Figure 4E and 4F). We further stained for endogenous IP3R3 in HeLa cells and saw a similar reduction in IP3R3 clusters upon depletion of INF2 (Figures S4E and S4F). Further we stained for endogenous IP3R3 and simultaneously performed PLA for IP3R3/ INF2 which revealed that about 63.2 ± 7.9% of the PLA dots are on IP3R3 clusters suggesting that substantial interaction between IP3R3 and INF2 occurred at IP3R3 clusters (Figures S4G and S4H). Since both isoforms of INF2 could significantly rescue agonist induced [Ca²⁺]_C and [Ca²⁺]_ER transients in HeLa and HEK cells, we wanted to see whether re-introduction of INF2 isoforms also rescued IP3R2 clustering. Satisfyingly both INF2 isoforms could completely recover clustering of endogenous IP3R2 in both the cell lines (Figures 4C–4F). While INF2-activity seems dispensable for its ability to bind IP3R, we wanted to test whether activity is required for IP3R clustering and agonist-induced IP3R activity. While re-introduction of INF2-caax (WT) rescued both IP3R2 cluster formation (Figures 4G and 4H) and histamine-induced [Ca²⁺]_C transients in HeLa INF2-KO cells (Figures 4I and 4J), the inactive FFC-caax mutant, which still interacts with IP3R2 (Figure 3E) failed to significantly rescue either of the phenotypes (Figures 4G–4J). Additionally, the INF2-caax S1099A mutant, with reduced IP3R2 binding, was also severely compromised for its ability to rescue histamine-induced [Ca²⁺]_C transients in INF2-KO cells (Figures 4I–4J). Interestingly INF2-noncaax S1099E mutant could significantly restore [Ca²⁺]_C transients in INF2-KO cells similar to WT-INF2-noncaax indicating the importance of phosphorylation at this specific site for agonist-induced ER Ca²⁺ release (Figure S4I and S4J). These results suggest that while INF2-IP3R2 interactions are independent of INF2-activity, IP3R2 cluster formation and agonist-mediated IP3R activity is dependent on INF2-activity. Interestingly depletion of INF2 causes a significant increase in peripheral ER sheets (Figure S4K) which might be a possible cause for IP3R inactivity. However, we rule out INF2-mediated ER structural aberration as a plausible mechanism as this ER phenotype is exclusively rescued by INF2-caax (Figure S4K) while IP3R cluster assembly and activity is rescued by both INF2 isoforms. Collectively these results suggest that INF2 plays a significant role in formation and/or stability of endogenous IP3R clusters without altering its absolute abundance on ER. Our results indicated that both IP3R2 and IP3R3 loses their clustering ability upon depletion of INF2, a phenotype that is completely rescued by re-introduction of INF2. Previous results have shown that Ca²⁺ release preferentially occur close to the immobile IP3R clusters on the ER ³⁶ and we propose that cluster formation might be regulated by INF2.

INF2-caax specifically regulate ER-Mito contacts and ER to mito calcium transfer

Our results show that both INF2-caax on ER and INF2-non-caax in the cytosol can interact with IP3Rs and regulate cluster formation thereby allowing for agonist-induced ER Ca²⁺ release. However, we had previously shown that ER localization of INF2 is indispensable for its role in ER-Mitochondrial Ca²⁺ transport and mitochondrial division ²⁶. Therefore, we reasoned that INF2-caax has unique roles beyond IP3R clustering to mediate ER-mitochondrial tethering and agonist-induced ER to mitochondrial Ca²⁺ transfer. To this end we imaged for GFP-Tom20 and endogenous IP3R2 in INF2-KO cells rescued with either INF2-caax or INF2-non-caax. While in both cases we noticed IP3R2 cluster formation, finer evaluation of the clusters revealed that in the case of INF2-caax re-expression the IP3R2 clusters were more aligned with mitochondria compared to the ones with INF2-non-caax re-expression (Figures S5A and S5B). We therefore wanted to evaluate whether depletion of INF2 had any profound effect on the ability of IP3R2 to mediate ER-Mito contacts (ERMC). We performed PLA using VAP-B (ER) and Tom20 (Mito) as a surrogate for ERMC. While IP3R2 expressing HEK cells showed an increased density of PLA dots compared to IP3R-TKO HEK cells (Figures 5A and 5B), depletion of INF2 from these IP3R2-expressing cells significantly diminished the ability of IP3R2 to create ER-Mito contacts (Figures 5A and 5B). Finer evaluation of these contacts from thin-section TEM images showed that depletion of INF2 in the IP3R2 expressing cells compromised the tight contacts (0-20 nm) between ER and mitochondria (Figures 5C and 5D) and concomitantly caused significant increase in the minimum distance between ER and mitochondria (Figure 5D). Infact chronic depletion of INF2 in HeLa cells (that has all the IP3R isoforms) also drastically reduces close contacts and increases the minimum distance between ER and mitochondria as evaluated from thin-section TEM analysis (Figures S5C–S5E). Re-introduction of INF2-caax completely rescued the density of PLA dots (VAP-B/Tom20) but INF2-non-caax could only partially rescue the phenotype (Figures 5A and 5B). The abundance of VAP-B, Tom20 and PTPIP51 was similar across all the conditions tested (Figure S5F). Further we functionally validated the relevance of INF2-mediated ERMC through recording mitochondrial Ca²⁺ uptake in IP3R2-expressing HEK cells. Depletion of INF2 in IP3R2-HEK cells significantly reduced carbachol-induced mitochondrial Ca²⁺ uptake which was significantly restored by re-introduction of INF2-caax but not INF2-non-caax (Figures 5E and 5F). Interestingly, Store Operated Calcium Entry (SOCE) induced mitochondrial Ca²⁺ uptake was unperturbed in the INF2-KO cells suggesting that INF2 specifically altered ER to mitochondrial Ca²⁺ transfer (Figure 5G). Therefore, it was evident that while ER localization of INF2 was not required for agonist-induced IP3R2 activity, INF2-caax specifically influenced the ability of IP3R2 to mediate ERMC and downstream agonist-induced mitochondrial Ca²⁺ uptake. This further prompted us to test whether INF2 have IP3R-independent role in regulation of ERMC. PLA assays showed a significant reduction in the abundance of PLA dots but not PLA density upon INF2 depletion (KO) in IP3R-TKO HEK cells suggesting that INF2 might have a general role in mediating ERMC (Figures 5H and 5I). To dissect this further we evaluated the extent of ERMC in these cells from thin-section TEM. Interestingly, chronic INF2 in depletion in TKO cells could further significantly reduce close ERMC (10-20 and 20-30 nm) while still preserving the tightest contacts (0-10 nm) in these cells (Figures 5J and 5K). Loss of INF2 in the TKO cells also significantly increased the minimum distance between ER and mitochondria (Figure 5K). Collectively these results further confirm that INF2 plays a pivotal role in mediating ERMC and downstream agonist-induced mitochondrial calcium uptake.

Figure 5. INF2-caax specifically regulates ER-Mito contacts and ER to mito calcium transfer.

A: Representative confocal micrographs from PLA assay against VAP-B (ER) and Tom20 (mitochondria) to assess ERMC, in HEK IP3R-TKO, IP3R2 (R1/R3 DKO) WT / INF2-KO and INF2-KO cells rescued with INF2-caax or INF2-noncaax. TKO, WT and INF2-KO cells were transfected with ER-RFP (magenta) to label ER prior to PLA staining. The PLA dots of VAP-B/Tom20 proximity is shown in green. Scale: 10 μm.

B: Quantification of PLA density expressed as number of dots per 100 μm² (left), and PLA dots, expressed as number/cell (right). Data from 49 cells for TKO, 43 for WT, 59 for INF2-KO, 40 for INF2-KO + INF2-caax and 43 INF2-KO + INF2 noncaax from 3 independent experiments. Graphs show Mean ± S.D. One-way Anova used. **** p<0.0001, *** p=0.0004, * p=0.0370 (density); **** p<0.0001, * p=0.0367 (dots).

C: Representative images from thin-section TEM in WT and INF2-KO HEK-IP3R2 (R1/R3 DKO) cells showing ER and mitochondria in the sections. The arrows represent close contact between ER and Mitochondria in the sections. Scale: 200 nm.

D: Scatter plots showing the percentage of close ER-Mito contacts, expressed as % of mitochondrial perimeter, between 0-10, 10-20 and 20-30 nm (upper); Quantification of the minimal ER-Mito distance in nm (lower). Data of 79 (WT) and 89 (INF2-KO) mitochondria, from 2 independent sets of images. Graphs show Mean ± S.D. unpaired Student’s T-test. * 0.0251 (WT/INF2 KO 0-10 nm), * 0.0106 (WT/INF2 KO 10-20 nm) and * 0.0176 (WT/INF2-KO; minimum distance)

E: Traces of mitochondrial calcium transients measured in HEK IP3R2 (R1/R3 DKO) WT, INF2-KO and INF2-KO rescue (+caax or +non-caax) cells, transfected with Mito GCaMPf6 (Kd=0.2 μM).

F: Quantification of maximal peak of mitochondrial Ca²⁺ uptake in C, expressed as fold change F/F0. Graphs show Mean ± S.D. Significant was measured using One-way Anova. Data from 66 (IP3R-TKO), 92 (IP3R2 expressing HEK), 99 (IP3R2 expressing INF2-KO), 96 (INF2-KO + INF2 caax) and 53 (INF2-KO + INF2-noncaax) across -4 independent experiments. Graphs show Mean ± S.D. Significant was measured using One-way ANOVA. **** p<0.0001.

G: Traces of cytosolic (Cyto-R-GECO) and mitochondrial (MitoGCaMPf6) calcium transients measured simultaneously in WT and INF2-KO cells following 2 μM Thapsigargin treatments in calcium free ECM followed by re-introduction of 1 mM CaCl₂. Traces shown are average value from 27 WT and 23 INF2-KO cells across 2 independent experiments.

H: Confocal micrographs from PLA against VAP-B (ER) and TOM20 (mitochondria) to assess ERMC, in HEK TKO and HEK TKO INF2-KO. Cells were transfected with ER-RFP (magenta) to label ER prior to PLA staining. The interaction of these proteins is shown in green dots. Scale: 10 μm.

I: Quantification of PLA density expressed as number of dots per 100 μm² (left), and PLA dots, expressed as number/cell (right). Data of 33 cells for TKO and 46 TKO INF2-KO, from 3 independent experiments. Graphs show Mean ± S.D. Significant was measured using unpaired Student’s t-test. Graphs show Mean ± S.D. One-way ANOVA used. ** p=0.0054.

J: Representative images from thin-section TEM in IP3R-TKO HEK and IP3R-TKO-INF2-KO HEK cells showing ER and mitochondria in the sections. The arrows represent close contact between ER and Mitochondria in the sections. Scale: 200 nm.

K: Scatter plots showing the percentage of close ER-Mito contacts, expressed as % of mitochondrial perimeter, between 0-10, 10-20 and 20-30 nm (upper); Quantification of the minimal ER-Mito distance in nm (lower). Data of 60 (TKO) and 55 (TKO-INF2-KO) mitochondria, from 2 independent sets of images. Graphs show Mean ± S.D. unpaired Student’s T-test. ** 0.0034 (TKO/TKO-INF2 KO 10-20 nm), ** 0.003 (TKO/TKO-INF2 KO 20-30 nm) and * 0.0165 (TKO/TKO-INF2-KO; minimum distance)

L: Working model for INF2-dependent actin-mediated clustering, positioning and activity of IP3R in cells.

See also Figure S5.

Discussion

Collectively our data shows that INF2-mediated actin filaments play an integral role in agonist-induced IP3R-mediated ER calcium release. INF2-localization on the ER is dispensable for its role in IP3R activity. Our study further shows that depletion of INF2 significantly reduced IP3R clustering which have been shown to be important for regulation of IP3R. Further, both INF2 isoforms can significantly interact with all IP3R isoforms in an activity and actin filament independent manner, however, INF2 activity is imperative for IP3R cluster formation and agonist-mediated firing of IP3Rs. Finally, we show that the ER localization of INF2 influences proper positioning of IP3R2 close to mitochondria, mediating ER-Mitochondrial contacts and promoting ER to mitochondrial calcium transport. Overall, we show that in addition to previously described role of INF2 in mitochondrial division, it plays additional role upstream in the process by regulating (1) IP3R clustering; (2) IP3R-mediated ER calcium release and (3) IP3R positioning leading to ER-Mito contacts (Figure 5L). These finding suggests that INF2-mediated actin filaments might work to provide a scaffold for regulation of IP3R clusters regulating both IP3R activity and IP3R-mediated ER-mitochondrial crosstalk.

IP3Rs in cells are shown to be arranged in clusters ^36,37 and further have been classified into mobile and active immobile clusters ³⁶. Ca^{2+ 38}, an open transition state of IP3R ³⁹, agonist-induction ⁴⁰ and IP3 ⁴¹ have been shown to regulate clustering which is important for IP3R activity. It is, however, difficult to imagine what regulates the formation or lends stability to these clusters to allow Ca²⁺ release. Our results show that depletion of INF2 or its activity significantly reduces abundance of IP3R2 and IP3R3 clusters on ER signifying that actin filaments play an important role in regulating the clusters in the following possible ways: (1) INF2-mediated actin filaments might generate force needed for cluster formation; (2) INF2-mediated actin filaments might act as a scaffold required for cluster stability through either a force dependent or independent mechanism. In fact, actin filaments have been implicated in regulating functional clusters formation of Fcɣ receptors ⁴², CD44 on PM ⁴³, influenza hemagglutinin ⁴⁴ and RNA polymerase II in the nucleus ⁴⁵. While INF2 can polymerize actin, extensive actin bundling which often prevents its cofilin-mediated depolymerization ⁴⁶ might be required to aid in IP3R clustering which need to be studied further. Microtubules have been shown to positively regulate agonist-induced Ca²⁺ spike in epithelial cells ⁴⁷ and a recent study have shown that depletion of microtubule (+) end binding protein EB3 significantly decreased IP3R3 cluster lifetime ⁴⁸ raising speculations that actin and microtubule cytoskeleton might work in sync to regulate cluster assembly. At present it is difficult to tease out whether INF2-mediated actin filaments work upstream of EB3; however, since multiple formin proteins including INF2 have been shown to bind and modulate microtubules dynamics ^49–51, it is possible that actin filaments might also allow for IP3R-EB3 binding to allow for greater stability of the IP3R clusters. Furthermore, transition from mobile to immobile clusters have been shown to be mediated by KRAP-induced tethering of IP3R clusters with actin filaments in the PM at sites of Store operated Ca²⁺ entry ⁵² which is likely to be downstream of INF2-mediated assembly of IP3R clusters. Overall, we think that actin filaments, possibly together with the microtubule network, might act as a scaffold to allow for IP3R cluster assembly and stability to facilitate agonist-induced ER Ca²⁺ release.

Through multiple approaches we show a direct binding of INF2 with IP3R isoforms in cells, however, the exact domains involved in this interaction remains to be worked out. Clearly the N-terminal DID region of INF2, and the actin polymerization activity is dispensable for its binding with IP3R (Figure 3) suggesting that FH1, FH2, DAD or the C-terminal region post DAD domain might be important for the interaction. In fact, we show that INF2-caax S1099A and S1100A mutants are significantly compromised for their ability to bind to IP3R2 further implicating that the C-terminal tail might be an important regulation hub for INF2. However, further analysis is required to (1) identify additional factors that might aid in INF2-IP3R interaction; (2) dissect out whether a double S1099/1100A mutant totally abrogates interaction; (3) whether the amino acid residues per se and not the phosphorylation is what determines the interaction. Unlike INF2, IP3R boasts of an exhaustive list of interactors ⁵³ and through this study we add INF2 to the growing list. Clearly IP3 binding and pore-forming ability of IP3R does not influence its binding with INF2. An important molecule sitting at the speculative interface of this binding might be calmodulin which is known to bind IP3R in an inhibitory way ⁵⁴ but has been shown to activate INF2 ^27–31 in a Ca²⁺-dependent manner. It remains to be seen whether INF2-calmodulin binding allows sequestration of calmodulin from IP3R thereby further facilitating Ca²⁺ release.

In addition to mitochondrial fission, we show that INF2 regulates formation of ERMC. Interestingly our previous reports using U2-OS cells ²⁶ showed an unchanged basal ERMC upon INF2-KO, however both in unstimulated HeLa and HEK-293 cells we see a drastic reduction of close ERMC upon INF2 depletion both at the level of light microscopy and TEM analysis (Figure 5; S5). This might be due to a cell-to-cell variation or method of quantification employed in the two studies. We further show that INF2 depletion significantly dampened the ability of IP3R2 to facilitate ERMC (Figure 5) which is probably due to a reduction in IP3R cluster formation. While all IP3R isoforms have been shown to form clusters close to mitochondria ²³, IP3R2 seemed to have a higher overall contacts with mitochondria ²³. Our results further demonstrate that ER localized INF2-caax specifically regulated positioning of IP3R2 clusters close to mitochondria thereby facilitating not only the formation ERMC but also ER to mitochondrial Ca²⁺ transport. In fact, capture of IP3R clusters at ERMC allows it to stimulate local Ca²⁺ transfer to mitochondria boosting oxidative metabolism ⁵⁵ implicating the importance of INF2-mediated ER bound actin filaments in the process. Since ER shaping proteins like REEP1 ⁵⁶ and RTN1A ⁵⁷ have been shown to regulate ERMC, INF2 mediated regulation of ERMC might be due to alteration of ER structure in INF2 depleted cells which is recovered by ER bound INF2-caax only. Does INF2 work only with IP3Rs to mediate ERMC? Most probably the answer is negative. A significant reduction of ERMC in INF2 depleted IP3R-TKO cells point to the fact that ER-bound INF2 might be a global regulator of ERMC in general. Therefore, in addition to positioning important ERMC molecules at the interface, INF2 on the ER might work with other members on the outer mitochondrial membrane like Spire1 ⁵⁸ or Myosin-19 ⁵⁹ to stabilize ERMC. Further work is necessary to elucidate the specific role of INF2 at ERMC and its interplay with various molecular determinants on ER that regulate ER-mitochondrial crosstalk.

Resource Availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Rajarshi Chakrabarti (Rajarshi.Chakrabarti@jefferson.edu)

Materials availability

All plasmids, strains, and other reagents generated in this study are freely available upon request.

STAR Methods

Experimental model and Study Participant Details

Cell culture

Human cervical cancer cells (HeLa) were obtained from American Type Culture Collection (ATCC), and Human Embryonic Kidney (HEK) cells expressing specific IP3R isoforms were a kind gift of Dr. David Yule ³². Cells were grown in regular DMEM (Gibco 11-995-065) supplemented with 10% fetal bovine serum at 37°C with 5% CO2. Cell lines were tested every 3 months for mycoplasma contamination using Universal Mycoplasma detection kit (ATCC, 30-1012K) or MycoAlert Plus Mycoplasma Detection Ki (Lonza, LT07-701).

Generation of INF2-KO cell lines

HeLa and IP3R isoform-specific HEK cells were genome-edited for deletion of INF2 gene using the CRISPR/Cas9 technology. CRISPR-gRNA for INF2 was as follows: 5’-CACCGCGGAGATACGTGCAACGCCG -3’ that was designed to target a specific genomic region in exon 1. Selected gRNAs were cloned into pSpCas9(BB)-2A-Puro (Addgene #48139)⁶⁰, sequence-verified, and 1 μg of respective plasmids were transfected using Lipofectamine2000. Next day, transfected cells were replated into medium containing 2.5 μg/ml puromycin (Gibco #A1113802) and selected for 24 hours. After 24 hours puromycin containing media along with dead cells were replaced with normal DMEM and the adherent cells were allowed to become confluent. The cells were then diluted to ~0.5 cells per well and seeded in a 96 well to obtain single cell derived clones.Single cell clones were expanded and screened for the absence of INF2 protein expression by western blotting using specific antibodies.

Method details

Plasmids

Constructs for calcium transients: Cyto-R-Geco (Kd= 0.5 μM) and Cyto LAR-Geco (Kd=12 μM) was a gift from Y.M. Usachev (University of Iowa Carver College of Medicine, Iowa City, IA), ER GCaMP6 150 (Kd= 150 μM)⁶¹ was a kind gift from Timothy Ryan (Weil Cornell Medicine, NYC, NY), CMV-ER-LAR-GECO1⁶² was a gift from Robert Campbell (Addgene plasmid # 61244), Mito GCaMPf6 (Kd=0.2 μM) was a kind gift from Gyorgy Hajnoczky (Thomas Jefferson University, Philadelphia, USA). mcherry-Ftractin, GFP-INF2-caax, GFP-INF2-non-caax, GFP-INF2-caaxC1270S, GFP-INF2-caax-FFC, GFP-INF2 I676A, GFP-INF2-caax-S1099A, GFP-INF2-caax-S1100A, GFP-INF2-noncaaxS1100A, GFP-INF2-noncaaxS1100E, eGFP-Tom20, ER-RFP and GFP-Sec61β were a kind gift from Henry Higgs (Geisel School of Medicine, Dartmouth). IP3R1 plasmids encoding the D2550A pore mutant and the R265Q ligand-binding domain mutant were from Suresh Joseph (Thomas Jefferson University) and have been described previously ⁶³. YFP-PH-PLCδ plasmid was a kind gift from Tamas Balla (NIH, Bethesda). ERM-APEX2⁶⁴ was a gift from Alice Ting (Addgene plasmid # 79055). To generate Apex2-IP3R1 construct, a 747bp DNA fragment was synthesized by Twist Biosciences (San Francisco, CA) containing the Apex2 enzyme attached to a portion of the N-terminus of IP3R1 by a linker (GGSGGGSG). The fragment was synthesized with flanking EcoRI/KpnI sites which were used to insert the fragment into the full length IP3R1 plasmid.

For all experiments, the following amounts of DNA were transfected per 6-well (individually or combined for co-transfection): 250 ng of Cyto-R-GECO, Cyto-LAR-GECO, mito-GCaMPf6, YFP-PH-PLCδ, mcherry-Ftractin; 500 ng of ER GCaMP6 150, ER LAR-GECO, ER-RFP, GFP-Sec61β; 1000 ng of GFP-INF2-caax, GFP-INF2-non-caax, GFP-INF2-caax C1270S and GFP-INF2-caax FFC; 800 ng of IP3R1-WT, IP3R1-D2550A, IP3R1-R265Q

Plasmid transfections

Cells were seeded at 4×10⁵ cells per well in a 35 mm dish at ~16h before transfection. Plasmid transfections were made in OPTI-MEM media (Gibco, 31985062) containing 2.5 μl of Lipofectamine 2000 (Invitrogen, 11668) following the manufacturer’s protocol. After 4-6 h of transfection, cells were trypsinized and re-plated onto glass bottomed dishes (MatTek Corporation, P35G-1.5-14-C) at ~;1×10⁵ cells per well (for live cell imaging) or onto coverslips (25 mm) for fixed cell analysis.Cells were imaged or fixed ~16–24 h after transfection.

siRNA induced INF2 knockdown

Cells were seeded at 1×10⁵ cells per well in a 35 mm dish ~ 16 h before transfection. 63 pg of siRNA was added to OPTI-MEM (Gibco, 31985062) with 2 μl RNAimax (Invitrogen, 13778100), incubated at RT for 20 minutes and added to cells contained in 1 ml of pre-warmed DMEM media. 72-96 hours later cells were transfected with respective reporter constructs and prepared for live/fixed cell analysis. INF2-siRNA oligo: 5’-GAGCGGAUGAUCUCCAGGUUGAUCCCU-3′

Drug treatments

To evaluate the agonist-induced calcium release from ER histamine (H7125; Sigma-Aldrich), Carbachol (51-83-2; Thermo Scientific) or ATP (A6559; Sigma-Aldrich) were used at different concentrations. Agonist was perfused manually at the fifth frame (~25 sec) during the imaging and not replaced during the entire protocol (5 min). To evaluate basal cytosolic calcium levels, cells were loaded with 2 μM FURA-2 AM (Thermo Fisher Scientific, F1221). To evaluate the ER calcium stores, 2 μM of Thapsigargin (MilliporeSigma, 5860051MG) was added at the thirtieth frame (~;2 min) during the imaging and continued for another 5 min. For the Latrunculin A (Thermo Fisher Scientific, L12370) treatment, cells were pre-treated in DMEM with 0.5 μM of drug during 15 min, washed with PBS1X, and then the samples were collected following the IP protocol.

Immunofluorescence

For IP3R cluster, ER structure, HeLa and HEK cells were subjected to immunofluorescence staining. Cells (transfected or un-transfected) were plated onto coverslips ~16 h prior to fixation and staining. Cells were then fixed with chilled Methanol (for IP3Rs clusters, for 10 min at −20C) or with 4% PFA prewarmed (rest of IF, for 20 min at room temperature). In case of PFA fixation, cells were washed with PBS, permeabilized with 0.1% Triton X-100 for 1 min and washed with PBS three times. In all conditions cells were blocked with 10% FBS in PBS for ~60 min. Primary antibody was diluted in 1% FBS/PBS and coverslips were incubated on a drop of antibody solution on parafilm in a wet chamber for 1 h. Cells were then washed 6 times with 1X PBS and incubated with appropriate secondary antibody (prepared in 1% FBS/PBS) for 1 hour. Coverslips were again washed 6 times with 1X PBS and mounted on glass slides using ProLong Gold antifade mounting media (Invitrogen #P36930).

Primary antibodies:

anti-INF2 antibody (Proteintech, 20466-1-AP, used at 1:250)

anti-IP3R2 (A-5) antibody (Santa Cruz Biotechnologies, sc-398434, used at 1:100)

anti-IP3R3 antibody (BD Biosciences, 610312, used at 1:100)

anti-Tom 20 (Abcam ab78547, used at 1:400)

Secondary antibodies:

Goat anti-Mouse IgG (H+L) Cross-Adsorbed Secondary Antibody, Alexa Fluor^™ 647 (Invitrogen #A21235) used at 1:1000; Goat anti-Rabbit IgG (H+L) Cross-Adsorbed Secondary Antibody, Alexa Fluor^™ 488 (Invitrogen #A11008) used at 1:1000

Proximity Ligation Assay (PLA)

PLA (Duolink^™ PLA technology, In Situ Starter Kit Mouse/Rabbit, MilliporeSigma, cat# DUO92101) was used to evaluate the interaction between INF2/IP3Rs or VAPB/Tom20. Briefly, cells were seeded onto 12 mm coverslips for 16 hours and then were fixed with chilled methanol 10 min/−20C (for INF2/IP3Rs), or with prewarmed 4% PFA for 20 min at RT (to VAP-B/Tom20). Cells were washed with PBS (X 3 times), and in case of PFA fixation, cells were permeabilized using 0.1% Triton-X-100 for 1 min. The cells were then incubated with Duolink blocking buffer for 1 hour and incubated with primary antibodies (diluted in Duolink antibody diluent) for 2 hours at RT. Cells were then washed with 5% BSA (15 min x 3) and incubated with Duolink secondary antibodies for 1 hr at 37C. Cells were then washed with Duolink wash buffer A (5 min x 3) and incubated with Duolink ligase for 30 min at 37C. Cells were further washed with Duolink wash buffer A (2 min x 3) and incubated with Duolink polymerase for 90 min at 37 C. The cells were finally washed with Duolink wash buffer B (10 min x 3) and mounted on glass slides using prolong gold antifade reagent. Images were acquired within 24-48 hours. For multiplexed PLA and IP3R3 staining, Alexafluor-647 coupled secondary antibody was also added at 1:2000 during the secondary antibody incubations to visualize IP3R3 along with the PLA dots. For VAP-B/Tom20 PLA staining, the Alexa fluor 488 Phalloidin was added in the blocking buffer to mark actin filaments.

Antibodies and stains used:

anti-INF2 (Proteintech, 20466-1-AP, used at 1:250)

anti-IP3R2 (A-5) antibody (Santa Cruz Biotechnologies, sc-398434, used at 1:100)

anti-IP3R3 antibody (BD Biosciences, 610312, used at 1:100)

anti-VAP-B (Proteintech 66191-1-Ig used at 1:400)

and anti-Tom20 (Abcam ab78547, used at 1:400)

Alexa fluor 488 Phalloidin (Invitrogen, A12379, used at 1:500)

Microscopy

Agonist-induced calcium dynamics:

The images were acquired using a Nikon A1 on a Nikon Ti-E base and equipped with an iXon Ultra 888 EMCCD camera. A solid-state 405 smart diode 100-mW laser, solid-state 488 OPSL smart laser 50-mW laser, solid-state 560 OPSL smart laser 50-mW laser, and solid-state 637 OPSL smart laser 140-mW laser were used (objective: 20X, 1.45 NA CFI Plan Apo; Nikon). Images were acquired using Nikon Elements v5.02. Images from Figures 1, 2, 4I, 4J, 5C, 5D, and Fig. S1B, S1H, S1I, S3, S5B and S5C were acquired using this system and quantified using Image J.

Basal calcium estimation:

Epifluorescence Ca²⁺ imaging for SI Appendix, Fig. S1C, S1D,S1E, S2C, S2D and S2E, were acquired using the EM-CCD camera (Hamamatsu) assembled to an Olympus IX81 microscope with Lambda DG4 light source operated by Metamorph (Molecular Devices). Time lapse images were acquired using the UV-optimized Olympus UAPO/340 *40/1.35 N.A. oil immersion objective every 4 s.

Fixed samples:

was performed on LSM 880 equipped with 63×/1.4 NA plan Apochromat oil objective using the Airyscan detectors (Carl Zeiss Microscopy). The Airyscan uses a 32-channel array of GaAsP detectors configured as 0.2 airy units per channel to collect data that is subsequently processed using Zen2 software. Cells were imaged with the 405-nm laser and 450/30 filter for BFP, 488 nm and 525/30 for GFP, 561 nm and 595/25 for RFP and 640 nm laser and 650 LP for Far-red. Images for Figures 3A, 4C, 4E, 4G, 5A 5E, Fig. S1B, S4A, S5E, S5F, S5H, S5J and S6A, were acquired using this system and quantified using Image J.

Calcium measurements

Cells expressing respective calcium sensors either alone or with INF2-rescue plasmids were washed twice with PBS1X and perfused with 0.25 % BSA ECM buffer, calcium-free (ECM: 121 mM NaCl, 5 mM NaHCO₃, 10 mM Na-HEPES, 4.6 mM KCl, 1.2 mM KH₂PO₄, 1.2 mM MgSO₄, 10 mM glucose, pH7.4) and transferred to the microscope equipped with a heated (37C) stage-top. Cells were recorded for 25 seconds at 5s/frame, stimulated with agonists and imaged for another 5 minutes at 5s/frame. Images were quantified using ImageJ where values were expressed as fold change F/F0 where F0 is the mean of resting fluorescence intensity before stimulus addition. For rapid kinetics, cells expressing Cyto-R-Geco were recorded at 1frame/sec for 4 minutes.

To determine the basal cytosolic and ER Ca²⁺ levels, cells were loaded with 2 μM Fura-2 AM in the presence of 0.003% Pluronic acid F-127 and 100 μM sulfinpyrazone for 10 min at room temperature, washed with fresh ECM containing 0.25% BSA, calcium-free, and transferred to the microscope stage (37C). Fluorescence was recorded at 340 and 380 nm excitation, using dual-band dichroic and emission filters at 4s/frame. To evaluate the ER stores, 2 μM of Thapsigargin was added after 2 min of initial recording and further imaged for another 8 min. Values were expressed as F340/F380 ratio.

Actin burst

To measure actin burst, U2OS-INF2-KO cells were co-transfected with GFP-INF2 constructs the day before of imaging. After ~16–24 h of transfections, cells were stimulated with 4 uM ionomycin for 45 seconds, fixed with 4% pre-warmed PFA for 20 min at RT, washed with PBS (X3), permeabilized with 0.1% Triton X-100 for 1 min at RT, washed with PBS (X3) and stained with phalloidin-647 for 30 minutes at RT. The cells were then washed with PBS and mounted on slides to be imaged.

Lysate preparation and Western Blot

To evaluate protein abundance (except IP3R), cells from a 35mm dish were lysed using 1% SDS, collected and heated at 95C for 5 min. To evaluate abundance of IP3R, cells were lysed in RIPA buffer (reconstituted with 1X Protease/Phosphatase inhibitors) and saved at −80C without boiling. To run them on SDS-PAGE, samples were mixed with 2xDB (50mM Tris-HCl, pH 6.8, 2mM EDTA, 20% glycerol, 0.8% SDS, 0.02% Bromophenol Blue, 1M NaCl, 4M urea) and separated in a 4-20% precast gradient gels (all proteins except IP3R) or custom-made 5% Tris-glycine gels (in case of IP3Rs), transferred to a polyvinylidene fluoride membrane (EMD Millipore, IPFL00010). The membrane was blocked with TBS-T (20 mM Tris-HCl, pH 7.6, 136 mM NaCl, 0.1% Tween20) containing 3% BSA (VWR Life Science, VWRV0332) for 1h, then incubated with primary antibody solution at 4°C overnight or 2 hours at RT. After washing with TBS-T, the membrane was incubated with requisite secondary antibody for 1h at room-temperature and developed on a Chemi-doc system using appropriate reagents.

Primary Antibodies used:

anti-INF2 antibody (Proteintech, 20466-1-AP, used at 1:1000)

anti-β-Actin antibody (CST, #3700S, used at 1:5000; SIGMA, A2228 used at 1:10000)

anti-IP3R1 (D53A5) antibody (CST, 8568S, used at 1:1000)

anti-IP3R2 (A-5) antibody (Santa Cruz Biotechnologies, sc-398434, used at 1:500)

anti-IP3R3 antibody (BD Biosciences, 610312, used at 1:2000)

anti-VDAC (SA93-03) antibody (Invitrogen, MA5-41088, used at 1:1000)

anti-VAP-B (Proteintech 66191-1-Ig used at 1:2000 at RT/ 2hr)

anti-Tom20 (Abcam ab78547, used at 1:1000)

anti-PTPIP51 (Proteintech 20641-1-AP used at 1:1500 at RT/ 2 hr)

anti-IRBIT (Cell Signaling Technology #94248 used at 1:1000)

anti-α-actinin (Cell Signaling Technology #6487 used at 1:1000)

Secondary Antibodies used:

anti-Rabbit-HRP antibody (Jackson Immuno Research, 111-035-144, used at 1:10000)

anti-Mouse-HRP (Jackson Immuno Research, 715-035-150, used at 1:5000)

anti-rabbit IgG (H+L) antibody conjugated to IRDye 800CW (LICORBio, 926-32211, used at 1:10000)

anti-Mouse IgG (H+L) antibody conjugated to IRDye 680LT (LICORBio, 926-68020, used at 1:10000)

Immunoprecipitation using GFP-Trap agarose beads

INF2-IP3R2 interactions was measured using the GFP-Trap agarose beads (ChromoTek GFP-Trap^® Agarose). Briefly, IP3R2 (IP3R1/3 DKO)-INF2-KO HEK cells were seeded into a 10 cm dish and were transfected with GFP-tagged-INF2-constructs. After 6h of transfection, media was replaced with fresh DMEM. Next day, media was removed, washed with PBS and lysates was prepare using immunoprecipitation buffer (IPB) (IPB: 140 mM KCl, 50 mM Tris-HCl (pH 7.4), 2 mM MgCl2) with protease and phosphatase inhibitors containing 1% Triton-X100, Lat A (2 μM) and Phalloidin (10 μM). Cells were lysed on a rotator for 1 h at 4C, then centrifugated for 15 min at 15000 rpm at 4C to collect the supernatant. ~10% of the volume was used for the Input, and the rest was added to the conditioned GFP-trap-beads. After 2 h mixing on a rotator at 4C, the beads were collected via centrifugation, washed with IPB (X 3) and reconstituted in sample buffer, heated at 95C for 5min, and evaluated through SDS-PAGE.

For the pull-down with Lat A, transfected cells were pre-treated with 0.5 μM of Latrunculin A for 15 min at 37C and then processed as described above.

For pull-down using IP3R1 constructs, IP3-TKO-INF2-KO HEK cells transiently expressing IP3R1 constructs and GFP-INF2-caax were processed as above.

APEX2-labelling for IP3R1

Proximity biotinylation was carried out using a variation of published protocol ⁶⁵. Briefly, triplicate 100mm dishes of IP3R TKO cells were transfected with 10ug each of APEX2-IP3R1 or ERM-Apex2. After 48h, the cells were preincubated with 0.5mM Biotin phenol for 40min and then treated for 1min with 1mM H₂O₂. The biotinylation reaction was then quenched by washing in PBS containing 1mM of Trolox, sodium ascorbate and sodium azide. Cells were lysed in a buffer containing 150mM NaCl, 20mM Tris Hepes (pH 7.8), 0.5mM EDTA, 1% Triton X-100, 0.1% SDS, protease inhibitors and biotinylation quenching reagents. Biotinylated proteins in the lysates were recovered by incubation for 4h with streptavidin agarose beads. The beads were washed as described in Figure 3G and subjected to overnight digestion with trypsin. Samples were processed for LC-MS/MS at the Wistar proteomic facility. Data presented as Data S1.

Transmission Electron Microscopy

HEK cells (Figures 5C, D, I, J) for TEM analysis were fixed using 2% glutaraldehyde, then with 1% OsO4 and stained with 0.5% uranyl acetate, pelleted in 2% agarose (Sigma–Aldrich, Type IX ultralow gelling temperature), dehydrated in a dilution series of acetone/ water, and embedded in Spurr’s resin (Electron Microscopy Sciences). The sections were examined with a JEOL JEM1010 TEM fitted with a side-mounted AMT XR- 50 5Mpx CCD camera or an FEI Tecnai 12 TEM fitted with a bottom-mounted AMT XR-111 10.5 Mpx CCD camera. Images were taken with or ×3200–4400 (FEI) direct magnification. HeLa WT and INF2-KO cells (Figures S6C–S6E) for electron microscopic examination were trypsinized, washed twice with PBS and fixed with 2.5% glutaraldehyde, 2.0% paraformaldehyde in 0.1M sodium cacodylate buffer, pH7.4, overnight at 4C in a rotator. After subsequent buffer washes, the samples were post-fixed in 2.0% osmium tetroxide for 1 hour at room temperature and rinsed in dH2O prior to en bloc staining with 2% uranyl acetate. After dehydration through a graded ethanol series, the cells were infiltrated and embedded in EMbed-812 (Electron Microscopy Sciences, Fort Washington, PA). Thin sections were stained with uranyl acetate and lead citrate and examined with a JEOL 1010 electron microscope fitted with a Hamamatsu digital camera and AMT Advantage NanoSprint500 software. Interfaces between ER and mitochondria were analyzed using a custom ImageJ script: https://sites.imagej.net/MitoCare/ as described elsewhere ⁶⁶. Briefly, mitochondrial and ER perimeter was outlined and evaluated using the above-mentioned script. The data compiled to read out the percentage of mitochondrial perimeter that has an ER fragment within the 0-10, 10-20, 2030 and 30-40 nm within it.

QUANTIFICATION AND STATISTICAL ANALYSIS

Image analysis

Image J (NIH) was used to analyze and process all images

Quantification of calcium transients:

Cells transfected with the indicated calcium probes as described above were plated on MatTek dishes for live-cell imaging. Cells were imaged using a 20X water objective on a heated stage at a single confocal slice in the medial region, approximately 2 μm above the basal surface, to avoid stress fibers. Mean fluorescence intensity was calculated for each cell using the ImageJ plug-in “Time Series Analyzer V3.” after background subtraction (taken from extra-cellular 2-3 ROI in the field). Fluorescence values for each time point after drug treatment were normalized to the mean baseline fluorescence (first five frames) and plotted as intensity fold change over time in seconds. Average fluorescence intensities for mitochondrial calcium and ER calcium were determined by outlining the respective organelle networks. Data were plotted as intensity–time curves each data point using Microsoft Excel.

Cells loaded with FURA-2AM to determine the basal and ER Ca²⁺ stores, were analyzed using the custom-made Spectralyzer (imaging software). A masking for each cell was created and average intensities for F340 and F380 were collected and plotted as a ratio (F340/F380). An individual extra-cellular ROI was evaluated for both wavelengths to determine the background which was factored in the quantification.

Quantification of IP3R Clusters:

z-projected images were background-subtracted using “rolling ball radius” plugin of Image J and threshold corrected to preserve the brightest cluster of IP3R. Similar parameters were used for all conditions imaged. The thresholded IP3R image was converted to 8 bit and binarized. “Analyze Particles” plugin was used to record the total number of clusters (as dots) and presented as number per 100 μm² of cellular area. All images collected were quantified.

Quantification of PLA assays:

z- projected images were background-subtracted using “rolling ball radius” plugin of Image J with a value of 20.0 and phalloidin (actin) channel was used to get the cellular area. Channel containing the PLA dots were converted to 8-bit and binarized. “Analyze particles” plugin was used to record total number of dots and presented as number of dots per 100 um² of cellular area. All images collected were quantified.

Statistical analysis

All statical analysis and p-values were determined using GraphPad Prism 10. To determine p-values, an unpaired Student’s t-test was performed between two groups of data, comparing full datasets as stated in the Figure legends. To compare across more than two conditions, multiple comparisons (unpaired) and one-way or two-ways ANOVA were performed in GraphPad Prism 10. Data are expressed as Mean±SD.

Supplementary Material

2

Data S1. List of biotinylated peptides following expression of IP3R1-APEX and ERM-APEX constructs in HEK-IP3R-TKO cells. Related to STAR Methods and Figure 3.

A: Sheet showing summary of the differential biotinylation assay

B: Sheet showing all peptides.

C: Sheet showing filtered peptides having differential abundance among the two groups.

D: Sheet showing the list of peptides used for graphing.

E: Sheet showing a list of a few other filtered peptides.

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
INF2	Proteintech	Cat# 20466-1-AP, RRID:AB_10694821
β-Actin	Cell Signaling Technology	Cat# 3700, RRID:AB_2242334
β-Actin	Sigma-Aldrich	Cat# A2228, RRID:AB_476697
IP3R1	Cell Signaling Technology	Cat# 8568, RRID:AB_10890699
IP3R2	Santa Cruz Biotechnologies	Cat# sc-398434, RRID:AB_2924300
IP3R3	BD Biosciences	Cat# 610312, RRID:AB_397704
VDAC	Thermo Fisher Scientific	Cat# MA5-41088, RRID:AB_2898842
VAP-B	Proteintech	Cat# 66191-1-Ig, RRID:AB_2881586
Tom20	Abcam	Cat# ab78547, RRID:AB_2043078
PTPIP51	Proteintech	Cat# 20641-1-AP, RRID:AB_10695187
IRBIT	Cell Signaling Technology	Cat# 94248, RRID:AB_2800225
α-actinin	Cell Signaling Technology	Cat# 6487, RRID:AB_11179206
HRP-conjugated anti- Rabbit	Jackson ImmunoResearch Labs	Cat# 111-035-144, RRID:AB_2307391
HRP-conjugated anti- Mouse	Jackson ImmunoResearch Labs	Cat# 715-035-150, RRID:AB_2340770
IRDye 800CW conjugated-anti-Rabbit IgG (H+L)	LICORbio	Cat# 926-32211, RRID:AB_621843
IRDye 680CW conjugated-anti-Mouse IgG (H+L)	LICORbio	Cat# 926-68020, RRID:AB_10706161
Goat anti-Mouse IgG (H+L) Cross-Adsorbed Secondary Antibody, Alexa Fluor™ 647	Thermo Fisher Scientific, Molecular Probes	Cat# A-21235, RRID:AB_2535804
Bacterial and virus strains
Rosetta 2 competent cells	Millipore-Sigma	Cat#71400
Biological samples
Chemicals, peptides, and recombinant proteins
DMEM	Gibco	Cat #11-995-065
PBS	Thermo Fisher Scientific -Gibco	Cat #10010023
Fetal bovine serum	Thermo Fisher Scientific -Gibco	Cat #A5256701
Puromycin	Gibco	Cat #A1113802
OPTI-MEM media	Gibco	Cat #31985062
Lipofectamine 2000	Invitrogen	Cat #11668
RNAimax	Invitrogen	Cat #13778100
Histamine	Sigma-Aldrich	Cat #H7125
Carbachol	Thermo Fisher Scientific	Cat # 51-83-2
ATP	Sigma-Aldrich	Cat #A6559
FURA-2 AM	Thermo Fisher Scientific	Cat # F1221
Thapsigargin	MilliporeSigma	Cat # 5860051MG
Latrunculin A	Thermo Fisher Scientific	Cat #L12370
Triton x100	Cayman Chemical	Cat #600217
Pluronic acid F-127	Thermo Fisher Scientific	Cat #P3000MP
Sulfinpyrazone	Sigma-Aldrich	Cat #S9509
Ionomycin	Sigma-Aldrich	Cat #I0634
Phalloidin-647	Thermo Fisher Scientific	Cat #A22287
Phalloidin	Thermo Fisher Scientific	Cat #T7471
DMSO	Sigma-Aldrich	Cat #D8418
Protease and Phosphatase Inhibitor	Thermo Fisher Scientific	Cat #PIA32961
BSA	VWR Life Science	VWRV0332
PFA	Electron Microscopy Sciences	Cat #15710
Glutaraldehyde	Electron Microscopy Sciences	Cat #16000
Methanol	Fisher Scientific	Cat #A412-4
Sodium Cacodylate Buffer	Electron Microscopy Sciences	Cat # 11652
Alexa Fluor 488 Phalloidin	Invitrogen	Cat #A12379
ProLong Gold antifade mounting media	Invitrogen	Cat #P36930
PLA (Duolink™ PLA technology, In Situ Starter Kit Mouse/Rabbit	MilliporeSigma	Cat# DUO92101
Polyvinylidene fluoride membrane	EMD Millipore	Cat #IPFL00010
ECL Western Blotting Substrate	Thermo Scientific	Cat #PI32106
Critical commercial assays
GFP-Trap agarose beads	Proteintech	ChromoTek GFP-Trap® Agarose
Universal Mycoplasma detection kit	ATCC	30-1012K
MycoAlert Plus Mycoplasma Detection Ki	Lonza	LT07-701
Deposited data
Source data for all Figures (main and Supplementary) and the Apexdataset file have been deposited in the Mendeley repository and can be accessed using the following link: Chakrabarti, Rajarshi (2026), “INF2 regulates IP3R activity”, Mendeley Data, V2, doi: 10.17632/xtx7tycpvc.2
Experimental models: Cell lines
Human HeLa	ATCC	CCL-2
Human HeLa INF2-KO	This paper	N/A
Human HEK IP3R1 (IP3R2/IP3R3 DKO)	Alzayady, K.J. et al 32	N/A
Human HEK IP3R2 (IP3R1/IP3R3 DKO)	Alzayady, K.J. et al 32	N/A
Human HEK IP3R3 (IP3R1/IP3R2 DKO)	Alzayady, K.J. et al 32	N/A
Human HEK IP3R2 (IP3R1/IP3R3 DKO) INF2-KO	This paper	N/A
Human HEK IP3R3 (IP3R1/IP3R2 DKO) INF2-KO	This paper	N/A
Human HEK TKO	Alzayady, K.J. et al 32	N/A
Human HEK TKO INF2-KO	This paper	N/A
Experimental models: Organisms/strains
N/A	N/A	N/A
Oligonucleotides
CRISPR-gRNA for INF2: 5′-CACCG CGGAGATACGTGCAACGCCG -3′	IDT	N/A
INF2-siRNA oligo: 5′-GAGCGGAUGAUCUCCAGGUUG AUCCCU-3′	IDT	N/A
Recombinant DNA
pSpCas9(BB)-2A-Puro	Ran F.A. et al 60	Addgene plasmid # 48139 ; http://n2t.net/addgene:48139 ; RRID:Addgene_48139
Cyto-R-Geco	Y.M. Usachev (University of Iowa Carver College of Medicine, Iowa City, IA)	N/A
Cyto LAR-Geco	Y.M. Usachev (University of Iowa Carver College of Medicine, Iowa City, IA)	N/A
ER GCaMP6 150	de Juan-Sanz, et al 61	Addgene plasmid # 86918 ; http://n2t.net/addgene:86918 ; RRID:Addgene_86918
CMV-ER-LAR-GECO1	Wu J. et al 62	Addgene plasmid # 61244 ; http://n2t.net/addgene:61244 ; RRID:Addgene_61244
Mito GCaMPf6	Gyorgy Hajnoczky (Thomas Jefferson University, Philadelphia, USA)	N/A
mcherry-Ftractin	Henry Higgs (Geisel School of Medicine at Dartmouth, USA)	N/A
GFP-INF2-caax	Henry Higgs (Geisel School of Medicine at Dartmouth, USA)	N/A
GFP-INF2-non-caax	Henry Higgs (Geisel School of Medicine at Dartmouth, USA)	N/A
GFP-INF2-caaxC1270S	Henry Higgs (Geisel School of Medicine at Dartmouth, USA)	N/A
GFP-INF2-caax-FFC	Henry Higgs (Geisel School of Medicine at Dartmouth, USA)	N/A
GFP-INF2 I676A	Henry Higgs (Geisel School of Medicine at Dartmouth, USA)	N/A
GFP-INF2-caax-S1099A	Henry Higgs (Geisel School of Medicine at Dartmouth, USA)	N/A
GFP-INF2-caax-S1100A	Henry Higgs (Geisel School of Medicine at Dartmouth, USA)	N/A
GFP-INF2-noncaaxS1100A	Henry Higgs (Geisel School of Medicine at Dartmouth, USA)	N/A
GFP-INF2-noncaaxS1100E	Henry Higgs (Geisel School of Medicine at Dartmouth, USA)	N/A
eGFP-Tom20	Chakrabarti, R. et al 26	N/A
ER-RFP	Chakrabarti, R. et al 26	N/A
GFP-Sec61β	Chakrabarti, R. et al 26	N/A
IP3R1 D2550A	Boehning, D et al. 60	N/A
IP3R1 R265Q	Boehning, D et al. 60	N/A
YFP-PH-PLCδ	Tamas Balla (NIH, Bethesda, USA)	N/A
ERM-APEX2	Lam S.S. et al.64	Addgene plasmid # 79055 ; http://n2t.net/addgene:79055 ; RRID:Addgene_79055
Software and algorithms
GraphPad Prism version 10.4.1	GraphPad	https://www.graphpad.com/ ; RRID:SCR_002798
Office Excel	Microsoft	https://www.microsoft.com/en-us ; RRID:SCR_016137
FIJI Image analysis	Image J	https://imagej.net/Fiji; RRID:SCR_003070
Spectralyzer	Custom-made Spectralyzer (imaging software)	N/A
Other
N/A	N/A	N/A

Highlights.

• INF2-mediated actin filaments potentiate IP3R-mediated ER calcium release

• INF2 interacts with IP3R in an activity- and actin-filament-independent manner

• INF2 regulates IP3R cluster formation in actin-filament-dependent manner

• INF2 regulates IP3R cluster positioning to promote ER-mitochondrial calcium transfer

Data and code availability

Data are reported in the figures of the paper and supplemental information. Source data for all Figures (main and Supplementary) and the Apex-dataset file (Data S1) have been deposited in the Mendeley repository and can be accessed using the following link: Chakrabarti, Rajarshi (2026), “INF2 regulates IP3R activity”, Mendeley Data, V2, doi: 10.17632/xtx7tycpvc.2. No original code was reported in the paper. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.