Dual roles of TRPV2 in the innate immune response to cytosolic DNA: Arresting dormant and boosting activated STING

TL;DR

TRPV2 controls the STING immune pathway by keeping it inactive when needed and activating it when cytosolic DNA is present.

Contribution

TRPV2 is identified as a dual regulator of STING dormancy and activation in the innate immune response.

Findings

TRPV2 suppresses spontaneous STING activation in the absence of cytosolic DNA.

TRPV2 promotes STING activation by releasing Ca2+ from the endoplasmic reticulum upon cytosolic DNA detection.

TRPV2 regulates type I interferon production and natural killer cell activity via the cGAS/STING pathway.

Abstract

The cGAS/STING-dependent innate immune pathway is central in the cellular response to cytosolic DNA derived from viral infections, genotoxic stress, or mitochondrial defects. While efficient activation of the pathway is crucial for defending against pathogens and cancer, maintaining its dormancy without stimuli is equally important to avoid autoimmunity. However, the precise control of the cGAS/STING pathway remains poorly understood. Here, we report that the ion channel TRPV2 regulates both the dormancy and activation of STING. TRPV2 associates with STING and suppresses spontaneous STING activation in the absence of cytoDNA but dissociates from STING and promotes its activation by releasing Ca2+ from the endoplasmic reticulum in the presence of cytoDNA, which facilitates STING translocation to Golgi. Consequently, TRPV2 governs type I interferon production and natural killer cell-mediated cytotoxicity through the cGAS/STING pathway. The dual roles of TRPV2 provide an elegant mechanism for a balanced innate immune response.

A tightly controlled innate immune response is vital. Cheng et al. discovered the ion channel TRPV2 as a key regulator of STING-mediated innate immune response. Through their physical interaction, TRPV2 keeps STING dormant; activation by cytosolic DNA triggers their dissociation and ER Ca2+ release, initiating STING signaling.

Figures

H](#F3), no obvious differences in iCa^2+^ levels were observed in the two replacement cell lines depleted of endogenous STING, suggesting that TRPV2(S^m^) mutations do not affect the channel activity of TRPV2 in the absence of STING. In contrast, in the presence of endogenous STING, TRPV2(S^m^)-replacement cells exhibited a higher level of iCa^2+^ compared to TRPV2(WT)-replacement cells ([Figure 3I](#F3)). These observations reinforce the notion that the TRPV2-STING association represses TRPV2 channel activity to prevent unwarranted Ca^2+^ release and STING activation ([Figure 3J](#F3)).^[49]

A](#F4)). Importantly, this STING(C^m^/T^m^) phosphorylation requires both TBK1 and TRPV2 (Figures S4B and S4C). Consistent with its deficiency in binding and inhibiting TRPV2, STING(C^m^/T^m^)-replacement cells also displayed a higher level of iCa^2+^ (Figure S4D). cGAMP transfection did not further increase STING(C^m^/T^m^) phosphorylation, due to its deficiency in cGAMP binding ([Figure 4B](#F4)). Together, these results suggest that disrupting TRPV2 association can bypass the requirement of cGAMP binding for STING activation. In accordance with the effects on STING activation, STING(C^m^/T

E](#F5)). Similarly, depletion of STIM1 in TRPV2(S^m^)-replacement cells also increased STING activation ([Figure 5F](#F5)). Consistent with the effects on STING activation, disruption of the association of both TRPV2 and STIM1 resulted in an additive effect on the ER-to-Golgi translocation of STING ([Figures 5G](#F5) and [5H](#F5); S5F and S5G) and the expression of IL1β, IL-6, IFNβ, and CCL2 (Figure S5H and S5I). These data suggest that TRPV2 and STIM1 contribute collaboratively to the ER retention and repression of STING.

D](#F6)). This mechanistic coupling between STING and TRPV2 activation provides an amplification mechanism for rapid Ca^2+^ release and STING activation. It also enables close coordination between the STING-mediated innate immune response and other Ca^2+^-regulated processes in cells. Intriguingly, disruption of the STING-TRPV2 interaction through mutations in these proteins can partially bypass the requirement of cGAMP binding for STING activation, suggesting that a key function of cGAMP in STING activation is to dissociate TRPV2. It will be interesting to determine precisely how the conforma

Full text

INTRODUCTION

The cGAS/STING-mediated innate immune response system senses and responds to both foreign DNA derived from viruses and self-DNA derived from nuclear or mitochondrial genomes in the cytosol.^1-4^ Upon binding to cytosolic DNA, cGAS undergoes activation and then catalyzes the synthesis of 2′,3′ cyclic GMP-AMP (cGAMP) from ATP and GTP.^5-8^ Serving as a second messenger, cGAMP then binds to STING on the ER, causing its Golgi translocation and oligomerization.^9-17^ The subsequent recruitment of the kinase TBK1 to STING oligomers leads to its S172-autophosphorylation and activation.^18-22^ After activation, TBK1 then phosphorylates STING at S366, leading to the recruitment of the transcription factor IRF3 and its subsequent phosphorylation, also by TBK1.^19,23^ Phosphorylated IRF3 then dimerizes and translocates into the nucleus, where it drives the expression of type I interferon (IFN-I) genes.^18,24^ Additionally, TBK1 activates NF-κB to promote the production of proinflammatory cytokines.^25,26^ Efficient activation of this cGAS/STING/TBK1-mediated immune pathway, which is present in virtually all cell types, is crucial for the defense against pathogen infection and cancer.^1,27^ However, unwarranted activation or hyperactivation of the pathway needs to be avoided, which otherwise can cause autoimmune diseases, chronic inflammation, or tissue damage. As an example, loss-of-function mutations in TREX1, a nuclease that degrades cytosolic DNA, cause Aicardi-Goutieres syndrome (AGS), which is characterized by constitutive cGAS/STING activation and type I interferon production.^28-30^ Notably, even partial loss of cGAS or STING largely rescued the autoimmune and premature mortality phenotypes of mouse models of AGS with Trex1 deletion or mutations.^31-34^ Similarly, gain-of-function mutations in STING that cause its constitutive activation are associated with STING-associated vasculopathy with onset in infancy (SAVI), another autoimmune disorder.^35-38^ Consequently, cGAS/STING antagonism is being pursued as a therapeutic strategy for these diseases. On the other hand, recent studies indicate that this innate immune pathway is also required for cancer cell clearance by immunotherapeutic agents such as anti-PD-1 or anti-PD-L1, pointing out cGAS/STING agonism as an attractive approach for enhancing immunotherapy for certain cancers.^39-41^ However, despite its critical importance in both the innate immune system and multiple diseases, it remains poorly understood how the cGAS/STING pathway is properly controlled to ensure a timely and balanced immune response.

In addition to the innate immune response, cGAS/STING has also been shown to be involved in other cellular processes such as senescence and autophagy.^13,42-47^ We recently identified an important function of cGAS/STING in genome protection during DNA replication stress.^48-50^ Our results indicate that in the absence of replication stress, STING binds to and suppresses the ion channel TRPV2 on the ER to prevent Ca^2+^ release. Upon replication stress, the generation of cytoDNA results in cGAS activation and cGAMP production. The subsequent binding of cGAMP to STING causes its dissociation from TRPV2, leading to TRPV2 derepression and Ca^2+^ release. The resultant elevation of intracellular Ca^2+^ (iCa^2+^) then activates CaMKK2 and downstream kinase AMPK. Following activation, AMPK directly phosphorylates the nuclease EXO1 at S746, leading to the binding of the 14-3-3 adaptor proteins and the sequestration of EXO1. As a result, abnormal fork processing is avoided. Disruption of this pathway caused chromosomal instability and reduced cell viability.^48-52^ At the heart of this signaling pathway is the interplay between STING and TRPV2, which controls Ca^2+^ release from the ER. In the absence of cytosolic DNA, STING binds to TRPV2 and inhibits its channel activity to prevent Ca^2+^ release. Upon cGAMP-mediated stimulation, STING dissociates from TRPV2, leading to its derepression and Ca^2+^ release to activate the downstream pathway.^48-50^

The revealed STING-TRPV2 interaction and its regulation of TRPV2 prompted us to ask whether TRPV2 reciprocally regulates STING activation in the innate immune response. Interestingly, TRPV2 has previously been implicated in the innate immune system, with its roles in facilitating phagocytosis in macrophages, contributing to inflammation in conditions of heart failure, and affecting mouse survival after Listeria monocytogenes infection.^53,54^ However, it is unclear whether TRPV2 plays a direct role in the cGAS/STING innate immune response. In this study, we have identified two distinct functions of TRPV2 in controlling the STING-mediated innate immune response. First, by releasing Ca^2+^ from the ER via its channel activity, TRPV2 promotes STING activation and downstream signaling in response to cytosolic DNA. Second, through its physical association with STING, TRPV2 arrests STING on the ER to prevent its spontaneous activation in the absence of cytoDNA. This STING-TRPV2 interaction also restrains STING to prevent its hyperactivation in the presence of cytoDNA. The dual role of TRPV2 in STING regulation ensures robust immune responses, while avoiding unscheduled or excessive reactions.

RESULTS

TRPV2 promotes STING activation in the innate immune response through its channel activity

To determine whether TRPV2 directly regulates the cGAS/STING-mediate innate immune response, we introduced cytosolic DNA through plasmid DNA transfection into HeLa cells—which have an intact cGAS/STING innate immune signaling pathway—and then assessed the effects of TRPV2 depletion on STING activation. STING activation was assessed using TBK1 phosphorylation at T172 (pTBK1) and STING phosphorylation at S366 (pSTING), both of which are functional markers of innate immune signaling.^18,19,23,24^ Remarkably, siRNA-mediated TRPV2 knockdown abrogated both pTBK1 and pSTING, suggesting that TRPV2 is required for innate immune signaling in response to cytosolic DNA (Figures S1A; 1A). To further test this idea, we treated cells with the cGAS activator MnCl_2_,^55-57^ or with the STING activator cGAMP,^10^ to activate the cGAS/STING pathway. TRPV2 knockdown also abolished pTBK1 and pSTING induced by MnCl_2_ or cGAMP, further confirming the role of TRPV2 in innate immune signaling (Figures 1B and 1C). TRPV2 depletion using siRNAs or shRNAs also prevented pTBK1 and pSTING induced by DNA transfection, MnCl_2_, or cGAMP in multiple other cancer or non-transformed cell lines, including HaCaT, MCF10A, THP-1, and Molm13, or in primary human mammary epithelial cells (HMECs), further demonstrating the function of TRPV2 in the cGAS/STING signaling pathway (Figures S1B-S1I). Consistent with our published finding that AMPK is activated in response to TRPV2-mediated Ca^2+^ release following cGAS/STING activation,^48-50^ DNA transfection, MnCl_2_, or cGAMP treatment also caused AMPK activation in a TRPV2-dependent manner (Figures 1A-1C; S1A-S1I).

We next asked whether the channel activity of TRPV2 is required for cGAS/STING-mediated innate immune signaling. To address this question, we overexpressed an mCherry-tagged, dominant-negative mutant of TRPV2 (TRPV2(DN)) that contains two mutations (E559K/E609K) in the pore region to block TRPV2 channel activity.^49,50,58^ Cells expressing mCherry were used as a control (Figure S1J). As shown previously, TRPV2(DN) overexpression blocked iCa^2+^ elevation induced by DNA transfection, MnCl_2_, or cGAMP, as measured with the intracellular Ca^2+^ biosensor GCaMP6s (Figures S1K-S1M).^49,50^ Interestingly, TRPV2(DN) overexpression also abolished pTBK1 and pSTING induced by DNA transfection, MnCl_2_, or cGAMP in HeLa and MCF10A cells, suggesting that the channel activity of TRPV2 is required for innate immune signaling (Figures 1D-1F; S1N-S1Q). In further support of this idea, we found that treating HeLa cells with the TRPV2 channel inhibitor Tranilast also largely abrogated pTBK1 and pSTING under those conditions (Figures S1R-S1T).^49,50^ In addition to exogenous DNA transfection, cytosolic DNA can be induced internally by replication stressors such as hydroxyurea (HU), which in turn can activate TRPV2.^49,50^ We found that TRPV2(DN) expression also inhibited HU-induced pTBK1 and pSTING in the cells depleted of TREX1, which prevents cytoDNA degradation (Figure S1U). Consistent with the requirement of TRPV2 channel activity for STING activation, we found that TRPV2(DN) overexpression prevented cGAMP-induced production of IFNβ, IL-1β, CCL2 and IL-6, whose genes are targets of the STING-mediated immune response (Figures 1G; S1V).^59^ Taken together, these data strongly suggest that TRPV2 is an integral component of the cGAS/STING pathway and that TRPV2 promotes STING activation through its channel activity.

TRPV2 functions specifically in the cGAS/cGAMP-dependent canonical pathway of STING activation

In addition to the cytosolic DNA-elicited, cGAS/cGAMP-dependent canonical pathway, STING can be activated via non-canonical mechanisms that are independent of cGAS/cGAMP.^60^ Dunphy et al. reported that etoposide treatment induced STING activation in a cGAS-independent, but TRAF6-dependent, manner.^61^ Because etoposide causes DNA damage, which can generate cytosolic DNA, we examined the effects of cGAS depletion on STING activation after a short (6 h) or long (24 h) treatment. Consistent with published results, siRNA-mediated knockdown of TRAF6, but not cGAS, abrogated pSTING and pTBK1 after 6 h etoposide treatment (Figures 1H; S1W). However, cGAS depletion largely inhibited pSTING and pTBK1 after 24 h etoposide treatment, presumably because cytosolic DNA was produced after the long treatment (Figure S1X). Remarkably, similar to cGAS depletion, TRPV2(DN) overexpression also did not affect STING activation after short treatment with etoposide but largely abrogated STING activation after long treatment (Figures 1I; S1Y). These observations suggest that TRPV2 promotes STING activation specifically in the cGAS/cGAMP-dependent canonical pathway (Figure 1J).

iCa2+, but not downstream CaMKK2 or AMPK, promotes STING activation by facilitating STING’s ER-to-Golgi translocation

It has been previously shown that cytoDNA or direct cGAS/STING activation induces TRPV2-mediated ER release of Ca^2+^, leading to intracellular Ca^2+^ (iCa^2+^) elevation and downstream CaMKK2/AMPK activation.^48-50^ The requirement of TRPV2 channel activity revealed above for STING activation suggests that iCa^2+^ is important for innate immune signaling. Indeed, in support of this idea, treating HeLa or MCF10A cells with the Ca^2+^ chelator BAPTA-AM blocked pSTING and pTBK1 induced by transfected DNA, MnCl_2_, or cGAMP (Figures 2A-2C; S2A). We next asked whether CaMKK2 and downstream AMPK are also required for STING activation. To address this, we treated CRISPR-Cas9-engineered HeLa cells that are depleted of CaMKK2 or AMPKα with cGAMP or MnCl_2_ and then examined STING activation.^48^ Unlike TRPV2 or Ca^2+^, depletion of CaMKK2 or AMPKα did not affect pTBK1 or pSTING after cGAMP or MnCl_2_ treatment, suggesting that they are dispensable for innate immune signaling after cGAS/STING activation (Figures 2D and 2E; S2B and S2C). These data indicate that iCa^2+^, but not downstream CaMKK2 and AMPK, is required for STING-mediated innate immune signaling in response to cytosolic DNA or direct cGAS/STING activation.

To define the mechanism by which TRPV2-mediated iCa^2+^ elevation promotes STING activation, we examined the ER-to-Golgi translocation of STING, which is believed to be required for its activation upon cGAMP binding.^62^ To assess STING’s ER-to-Golgi translocation, we analyzed its puncta formation in cells and its colocalization with the Golgi marker GM130.^12,63-65^ As shown in Figures 2F and S2D, TRPV2(DN) overexpression inhibited cGAMP-induced STING translocation after cGAMP transfection. Treating cells with the Ca^2+^ chelator BAPTA-AM also inhibited STING translocation induced by cGAMP transfection (Figures 2G; S2E). Unlike STING, no obvious ER-to-Golgi translocation was observed for TRPV2 in cells after cGAMP transfection (Figure S2F). Taken together, these data suggest that TRPV2-mediated Ca^2+^ release facilitates STING activation by promoting its ER-to-Golgi translocation (Figure 2H).

TRPV2 interaction represses STING to prevent spontaneous activation and hyper-activation

We have previously shown that the STING-TRPV2 interaction represses TRPV2 to prevent spurious Ca^2+^ release in the resting state and that their dissociation facilitates Ca^2+^ release in the presence of cytoDNA.^49,50^ Accordingly, replacing endogenous STING with a STING mutant containing F^117^T^118^W^119^/AAS substitutions (hereafter referred to as STING(T^m^) for simplicity), which is deficient in TRPV2 interaction, caused iCa^2+^ elevation in the absence of cytosolic DNA (Figure S3A).^49,50^ This iCa^2+^ elevation was reversed by TRPV2(DN) overexpression (Figure S3B). Intriguingly, STING(T^m^)-replacement cells exhibited a higher basal level of pTBK1 and pSTING compared with STING(WT)-replacement cells (Figure 3A). STING(T^m^)-replacement cells also exhibited a higher level of activation after cGAMP transfection (Figure 3B). These observations suggest the possibility that TRPV2 reciprocally represses STING to prevent spontaneous activation in the absence of immune stimuli or hyperactivation in the presence of stimuli. To further test this idea, we determined whether the spontaneous phosphorylation and hyper-phosphorylation of STING(T^m^) are also mediated by TBK1 and require TRPV2-mediated Ca^2+^ release, like endogenous STING activation induced by cytoDNA. Indeed, siRNA-mediated knockdown of TBK1 or TRPV2 blocked STING(T^m^) phosphorylation in the absence or in the presence of cGAMP (Figures 3C and 3D). Overexpression of TRPV2(DN) in STING(T^m^)-replacement cells also abrogated STING activation with or without cGAMP transfection (Figure S3C). This requirement of TRPV2 channel activity also explains why TRPV2 depletion did not cause STING activation. Consistent with the results for STING activation, we observed at both mRNA and protein levels elevated expression of IFNβ, IL-6, IL-1β, and CCL2 in STING(T^m^)-replacement cells in the absence or in the presence of cGAMP, compared to STING(WT)-replacement cells (Figures S3D and S3E). Interestingly, unlike STING(T^m^), constitutive activation of the STING(N154S), STING(V147L), and STING(R281Q) mutants identified in SAVI patients does not require TRPV2 (Figure S3F). This observation is consistent with the notions that SAVI-associated mutants are activated through mechanisms distinct from that of WT STING and that constitutive STING activation only partially explains the SAVI phenotypes.^66-70^

To further demonstrate that the physical interaction between STING and TRPV2 restrains STING activation, we set out to map the STING-interaction site(s) in TRPV2. By performing coIP experiments to examine the association between STING-Flag and various truncation mutants of TRPV2, we found that the transmembrane (TM) domain of TRPV2 is both necessary and sufficient for STING binding (Figure S3G). Within the TM domain, the transmembrane segment 2 (S2) and the flanking linkers, when deleted, abolished STING interaction (Figure S3H). Replacing the linker regions with G/S residues did not affect STING binding (Figure S3I, see M1 and M11). Through alanine scanning in the S2 region, we found that the L^438^I^439^L^440^ motif is important for STING binding (Figure S3I, see M5). Importantly, substitution of these residues with alanine in full-length TRPV2 (hereafter referred to as TRPV2(S^m^)) at least partially abrogated its association with STING (Figure 3E). The identification of key residues in the TM domain of both STING and TRPV2 required for their association further suggests that these two proteins interact with each other within the ER membrane.^49^

We next determined whether STING undergoes spontaneous activation in TRPV2(S^m^)-replacement cells in the absence of immune stimuli and hyperactivation in the presence of stimuli, similar to the effects observed for the “FTW/AAS” mutations in STING. Indeed, we detected a higher basal level of pSTING and pTBK1 in TRPV2(S^m^)-replacement cells compared to TRPV2(WT)-replacement cells (Figure 3F). After cGAMP transfection, hyperactivation of STING was also observed in TRPV2(S^m^)-replacement cells, compared to TRPV2(WT)-replacement (Figure 3G). Little or no STING activation was detected in TRPV2(DN)-replacement cells in the absence or presence of cGAMP, consistent with the requirement of TRPV2 channel activity for STING activation (Figures 3F and 3G). In agreement with the effects on STING activation, TRPV2(S^m^)-replacement cells also exhibited elevated expression of IFNβ, IL-6, IL-1β, and CCL2 with or without cGAMP transfection (Figures S3J and S3K). These data further support the idea that the TRPV2-STING interaction prevents spontaneous activation and hyperactivation of STING.

However, an alternative explanation for the observed effects of TRPV2(S^m^) is that these mutations enhanced the intrinsic channel activity of TRPV2 (i.e., independent of STING-interaction), which may cause an elevated level of iCa^2+^, resulting in enhanced STING activation. To test this possibility, we measured iCa^2+^ levels in TRPV2(WT)-replacement cells and TRPV2(S^m^)-replacement cells in the absence and in the presence of endogenous STING. As shown in Figure 3H, no obvious differences in iCa^2+^ levels were observed in the two replacement cell lines depleted of endogenous STING, suggesting that TRPV2(S^m^) mutations do not affect the channel activity of TRPV2 in the absence of STING. In contrast, in the presence of endogenous STING, TRPV2(S^m^)-replacement cells exhibited a higher level of iCa^2+^ compared to TRPV2(WT)-replacement cells (Figure 3I). These observations reinforce the notion that the TRPV2-STING association represses TRPV2 channel activity to prevent unwarranted Ca^2+^ release and STING activation (Figure 3J).^49^

Taken together, our results strongly suggest that TRPV2 plays two distinct roles in regulating STING activity. In the absence of cytoDNA, TRPV2 maintains STING in an inactive state through a channel-independent physical interaction, thereby preventing spontaneous activation of the innate immune response. On the other hand, upon cytoDNA stimulation, TRPV2 facilitates STING activation via its channel-dependent function by mediating Ca^2+^ release from the ER. Notably, even in the presence of cytoDNA, the TRPV2-STING interaction serves to restrain excessive STING activation, ensuring a balanced immune response.

Disruption of the TRPV2-STING interaction can bypass the requirement of cGAMP binding in STING activation

Our results described above suggest that in the resting state TRPV2 and STING associate with and repress each other. Upon induction of cytosolic DNA, cGAS-produced cGAMP binds to STING, leading to its dissociation from TRPV2. This dissociation causes TRPV2 derepression and Ca^2+^ release from the ER. The resulting iCa^2+^ elevation then promotes STING translocation and subsequent activation. To further delineate these signaling events, we asked whether disruption of TRPV2-STING association can bypass the requirement of cGAMP binding for STING activation. To address this question, we introduced the R238A and Y240A mutations into STING(FTW/AAS) to generate a STING(RY/AA, FTW/AAS) mutant (hereafter referred to STING(C^m^/T^m^)), which is deficient in both cGAMP and TRPV2 binding (Figure S4A).^10,19^ We then generated a replacement cell line expressing this mutant after depleting endogenous STING. As a control, we also generated a replacement cell line expressing STING(RY/AA) (hereafter referred to as STING(C^m^)), which is deficient in cGAMP binding but proficient in TRPV2 binding (Figure S4A).^49^ Remarkably, STING(C^m^/T^m^) exhibited a higher basal level of activation in the replacement cells in the absence of cytosolic DNA compared with STING(C^m^) (Figure 4A). Importantly, this STING(C^m^/T^m^) phosphorylation requires both TBK1 and TRPV2 (Figures S4B and S4C). Consistent with its deficiency in binding and inhibiting TRPV2, STING(C^m^/T^m^)-replacement cells also displayed a higher level of iCa^2+^ (Figure S4D). cGAMP transfection did not further increase STING(C^m^/T^m^) phosphorylation, due to its deficiency in cGAMP binding (Figure 4B). Together, these results suggest that disrupting TRPV2 association can bypass the requirement of cGAMP binding for STING activation. In accordance with the effects on STING activation, STING(C^m^/T^m^)-replacement cells also exhibited an elevated expression of IFN-β, IL-6, IL-1β, and CCL2 compared to STING(C^m^)-replacement cells (Figures S4E and S4F).

STING activation requires both TRPV2 dissociation and downstream iCa2+ elevation

We next asked whether iCa^2+^ elevation—which occurs downstream of TRPV2-STING dissociation—can bypass the requirement of TRPV2 dissociation in STING activation. To address this question, we treated cells with thapsigargin (which inhibits the ER uptake of Ca^2+^ or with A23187 (which induces extracellular Ca^2+^ influx) to raise iCa^2+^ and then examined TBK1/STING activation.^71,72^ We observed that at concentrations that cause high levels of iCa^2+^, TBK1 and STING were phosphorylated (Figures S4G and S4H), consistent with a previous report.^73^ However, at lower concentrations that induced an iCa^2+^ level that is comparable to that induced by cGAMP transfection, thapsigargin or A23187 treatment failed to induce pTBK1 or pSTING, suggesting that at a physiologically relevant level, iCa^2+^ alone is not sufficient to activate STING (Figures S4G and S4H). Next, we asked whether at a physiologically relevant concentration, iCa^2+^ raised via TRPV2-independent mechanisms can also activate STING when the STING-TRPV2 association is disrupted. To address this, we first depleted TRPV2 using siRNAs and then raised iCa^2+^ to a physiologically relevant level using thapsigargin or A23187. Remarkably, we observed STING activation in TRPV2-depleted cells but not in cells treated with a control siRNA (Figures 4C and 4D), supporting the idea that STING can be activated by iCa^2+^ in the absence of TRPV2 interaction.

To further test this idea, we overexpressed TRPV2(DN) to block Ca^2+^ release by TRPV2 in STING(T^m^)-replacement cells and in STING(WT)-replacement cells. TRPV2(DN) overexpression blocked spontaneous STING(T^m^) activation, consistent with the requirement of TRPV2 channel activity for STING activation (Figures 4E and 4F). Interestingly, raising iCa^2+^ levels in the replacement cells by thapsigargin or A23187 caused the activation of STING(T^m^) but not STING(WT) (Figures 4E and 4F). We also performed a similar experiment in replacement cells expressing STING(C^m^) or STING(C^m^/T^m^), both of which are deficient in cGAMP binding. As shown in Figures 4G and 4H, STING(C^m^/T^m^), but not STING(C^m^), was activated after thapsigargin or A23187 treatment in the presence of TRPV2(DN) overexpression. Taken together, our results strongly suggest that (1) a key function of cGAMP binding to STING is to promote TRPV2 dissociation and subsequent iCa^2+^ elevation, and (2) both TRPV2 dissociation and iCa^2+^ elevation are required for STING activation in the canonical pathway (Figure 4I).

TRPV2-STING association promotes the ER retention of STING

STING translocation from the ER to Golgi is believed to be required for its activation in response to cytosolic DNA. The physical association between STING and TRPV2 and their ER localization raise the possibility that TRPV2 restrains STING activation through its ER retention. In support of this idea, we found that STING(T^m^) exhibited an elevated level of puncta formation and colocalization with the Golgi marker GM130 compared to STING(WT) in replacement cells (Figures 5A; S5A). This elevated STING(T^m^) translocation was abrogated by TRPV2(DN) overexpression (Figures 5B; S5B). cGAMP transfection further enhanced the Golgi translocation of STING(T^m^), which was also inhibited by TRPV2(DN) overexpression (Figures 5B; S5B). STING(C^m^/T^m^), which is deficient in TRPV2 binding, also exhibited a higher level of puncta formation and GM130 colocalization compared to STING(C^m^) in replacement cells (Figures 5C; S5C). We also observed an elevated level of Golgi translocation of STING in TRPV2(S^m^)-replacement cells compared to TRPV2(WT)-replacement cells in the absence or presence of cGAMP (Figures 5D; S5D), consistent with the deficiency of TRPV2(S^m^) in STING binding (Figure 3E). STING translocation was blocked in TRPV2(DN)-replacement cells in the absence or presence of cGAMP (Figure 5D), consistent with the requirement of TRPV2-mediated Ca^2+^ release for translocation (Figure 2F). Together, these results strongly suggest that the TRPV2-STING association promotes the ER retention of STING, providing a mechanistic explanation for its repressive role in STING activation.

TRPV2 and STIM1 collaborate in restraining STING translocation and activation

Interestingly, the role of TRPV2 in STING repression is similar to that of STIM1, which is also an ER resident protein and also suppresses STING through physical association and ER retention.^59^ Best known for its function in store-operated Ca^2+^ entry (SOCE), STIM1 interacts with and activates the Orai1 channel on the plasma membrane to promote Ca^2+^ influx in response to ER Ca^2+^ depletion.^74^ However, unlike TRPV2 whose channel activity is crucial for STING activation (Figure 1), STIM1 is not required for STING activation in response to cytosolic or direct cGAS activation, because STIM1 depletion increased, instead of decreasing, STING activation under those conditions.^59^

To explore the functional relationship between TRPV2 and STIM1 in STING repression, we determined whether disruption of both the STING-TRPV2 interaction and the STING-STIM1 interaction causes additive/synergistic effects on STING activation. To disrupt STING-STIM1 interaction, STIM1 was depleted from cells using siRNAs. To disrupt the STING-TRPV2 interaction, we used STING(T^m^)- or TRPV2(S^m^)-replacement cells (note that TRPV2 depletion is not suited for this purpose because TRPV2 channel activity is required for STING activation). Consistent with published results, we found that siRNA-mediated STIM1 knockdown in HeLa cells caused a low, but detectable, level of STING activation in the absence of cytosolic DNA (Figure S5E). Depletion of STIM1 in STING(T^m^)-replacement cells caused a further increase in the activation of this mutant (Figure 5E). Similarly, depletion of STIM1 in TRPV2(S^m^)-replacement cells also increased STING activation (Figure 5F). Consistent with the effects on STING activation, disruption of the association of both TRPV2 and STIM1 resulted in an additive effect on the ER-to-Golgi translocation of STING (Figures 5G and 5H; S5F and S5G) and the expression of IL1β, IL-6, IFNβ, and CCL2 (Figure S5H and S5I). These data suggest that TRPV2 and STIM1 contribute collaboratively to the ER retention and repression of STING.

TRPV2 controls STING-driven cell killing by NK cells

Due to the critical role of the STING-mediated innate immune pathway in promoting the tumor infiltration and/or the activity of cytotoxic T cells, dendritic cells, and natural killer (NK) cells, STING agonism is being actively pursued as an approach to enhance the efficacy of cancer immunotherapy.^1,75-77^ To further illustrate the significance of STING regulation by TRPV2, we determined whether TRPV2 controls the ability of STING to promote NK cell-mediated cell killing. To this end, we used a co-culture of mNeonGreen-expressing HeLa cells and NK-92MI cells and tracked HeLa cell number over time via live cell fluorescence imaging in an Incucyte system.^78^ As expected, HeLa cells transfected with cGAMP were susceptible to NK-92MI-mediated killing (Figure S6A). In agreement with the requirement of TRPV2 for STING activation, siRNA-mediated knockdown of TRPV2 in HeLa cells prevented their cGAMP-induced killing by NK-92MI cells (Figure 6A). As described above, disruption of the TRPV2-STING interaction leads to spontaneous STING activation without immune stimuli and to hyperactivation with stimuli. Accordingly, STING(T^m^)-replacement HeLa cells exhibited increased susceptibility to NK-92MI killing compared with STING(WT)-replacement cells, both in the absence and presence of cGAMP transfection (Figure 6B). Likewise, TRPV2(S^m^)-replacement HeLa cells were more efficiently killed by NK-92MI cells than TRPV2(WT)-replacement cells or TRPV2(DN)-replacement cells, with or without cGAMP transfection (Figure 6C). These results further underscore TRPV2’s dual roles in regulating STING-mediated innate immunity.

DISCUSSION

This study has established TRPV2 as both a checkpoint and a facilitator of the cGAS/STING-dependent innate immune response. Through its physical interaction with STING, TRPV2 retains STING on the ER, thereby preventing its Golgi translocation and spontaneous activation in the absence of cytosolic DNA. This checkpoint function of TRPV2 is independent of its channel activity. In the presence of cytosolic DNA, the channel activity of TRPV2, which is liberated after STING dissociation, facilitates STING activation. The ER Ca^2+^ release through TRPV2 promotes the Golgi translocation of STING and subsequent phosphorylation by TBK1. Our findings suggest the following model for STING activation in response to cytosolic DNA: the binding of cGAS-produced cGAMP to STING causes conformational changes, resulting in the dissociation between STING and TRPV2. This dissociation causes TRPV2 derepression and the release of Ca^2+^ from the ER, which in turn promotes STING’s Golgi translocation and subsequent phosphorylation by TBK1, leading to type I interferon production and subsequent immune responses such as NK cell activation (Figure 6D). Thus, TRPV2-mediated Ca^2+^ release is an integral part of the canonical cGAS/STING-mediated innate immune response. However, TRPV2 is apparently not involved in the non-canonical, TRAF6-dependent mechanism of STING activation, suggesting that TRPV2 is a functional marker that can distinguish between the canonical and non-canonical pathways of STING activation.

The dual roles of TRPV2 in STING regulation ensure a tightly controlled, well-balanced innate immune response. In the resting state, the association between these two proteins keeps each other in the dormant state on the ER. Upon cytosolic DNA generation, the binding of cGAMP to STING triggers a sequence of discrete events, including STING-TRPV2 dissociation, TRPV2 activation, Ca^2+^ release, and STING translocation and activation (Figure 6D). This mechanistic coupling between STING and TRPV2 activation provides an amplification mechanism for rapid Ca^2+^ release and STING activation. It also enables close coordination between the STING-mediated innate immune response and other Ca^2+^-regulated processes in cells. Intriguingly, disruption of the STING-TRPV2 interaction through mutations in these proteins can partially bypass the requirement of cGAMP binding for STING activation, suggesting that a key function of cGAMP in STING activation is to dissociate TRPV2. It will be interesting to determine precisely how the conformational changes in STING are transmitted from the C-terminal ectodomain where cGAMP binds to the TM domain where TRPV2 apparently interacts.^14,15,49,50^ In addition to STING dissociation, STING activation requires downstream TRPV2-mediated Ca^2+^ release, which facilitates the ER-to-Golgi translocation of “freed” STING. Interestingly, structural studies on mouse STING suggest that STING contains a Ca^2+^-binding site.^79^ However, direct binding of Ca^2+^ to this site is apparently not responsible for its role in STING activation, because mutations at this site did not block, but actually enhanced, STING activation in both mouse cells^79^ and human cells (Figure S6B). Future studies are needed to determine exactly how Ca^2+^ promotes STING translocation and subsequent activation.

In addition to TRPV2, STING translocation and activation is also restrained by STIM1, another ER resident protein that also interacts with STING.^59^ Our results suggest that TRPV2 and STIM1 function in a partially redundant manner in STING repression. However, unlike TRPV2, STIM1 is not required for STING activation. Although STIM1 can also promote Ca^2+^ entry through the Orai1 channel on the plasma membrane, it is apparently not involved in iCa^2+^ elevation induced by cytosolic DNA, because TRPV2 depletion or inhibition completely blocked iCa^2+^ elevation.^49,50^ Recent studies have suggested that STING also acts as a proton channel.^65,80^ It will be interesting to find out if TRPV2 and STIM1 also play a role in regulating the proton channel activity of STING.

Our findings may have important implications for the development of new therapeutics for human diseases. The requirement of TRPV2 channel activity in STING activation suggests that inhibitors of TRPV2 may be effective in treating chronic inflammation and autoimmune diseases, similar to the effects of direct STING antagonists. Interestingly, Tranilast, a selective TRPV2 inhibitor, is an approved drug that shows benefits in allergic and inflammatory conditions, diabetic nephropathy, and ocular diseases whereby the cGAS/STING pathway likely contributes to the pathogenesis.^81^ A recently identified TRPV2 antagonist, Plumbagin, also exhibited anti-inflammatory effects and could ameliorate brain injury induced by middle cerebral artery occlusion/reperfusion in a mouse model.^82^ The effects of Tranilast and Plumbagin in those pathological conditions may in part be attributable to their inhibitory effects on TRPV2 and resulting STING inhibition. On the other hand, STING agonism is being actively explored to boost the effectiveness of cancer immunotherapy, due to the role of the cGAS/STING pathway in mobilizing and activating T cells, dendritic cells, and NK cells.^1,83^ Compounds that disrupt the TRPV2-STING interaction, thereby activating STING, could potentially enhance immune clearance of cancer cells, akin to the effects observed with direct STING agonists.

Limitations of the study

While our work has identified TRPV2’s multifaceted functions in regulating STING signaling, several important questions remain. It remains to be defined precisely how TRPV2/Ca^2+^ facilitates STING translocation and why TRPV2 is dispensable for the activation of certain constitutively active STING mutants. The impact of relative TRPV2 and STING expression in different cell types on their reciprocal regulation also warrants investigation. Furthermore, direct validation in animal models, particularly in the context of viral infections like HSV-1, is crucial to further confirm the physiological significance of the TRPV2-STING interplay.

STAR★METHODS

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Cell lines

Cell lines used in this study include HeLa, HaCaT, HEK293T, MCF10A, THP-1, and MOLM-13, which were authenticated using short tandem repeat (STR) profiling. NK92MI cells and Primary Mammary Epithelial Cells (HMEC) were newly purchased from ATCC and were not independently authenticated in our laboratory. The cell lines and primary cells used do not retain meaningful sex-specific biological characteristics relevant to the cGAS/STING-mediated innate immune response. No sex-based analyses were performed, and therefore, potential sex-specific effects could not be assessed, which represents a limitation of the study. All the cell lines were routinely tested and confirmed to be free of mycoplasma contamination. Detailed descriptions of cell culture conditions are provided in the method details section.

METHOD DETAILS

Plasmids, cell culture, transfection

Polymerase chain reaction (PCR)-amplified human TRPV2 and STING were cloned into the pCDH-CMV-MCS-EF1-Puro (pCDH) or pCW vectors. STING(C^m^) with R^238^Y^240^/AA mutations, STING(T^m^) with F^117^T^118^/W^119^/AAS mutations, STING(T^m^, C^m^) with R^238^Y^240^/AA and F^117^T^118^/W^119^/AAS mutations, STING (Ca^m^) with D^205^E^316^D^320^/SSS mutations, TRPV2(DN) with E^599^E^609^/KK mutations, TRPV2(S^m^) with L^438^I^439^L^440^/AAA mutations, and TRPV2 truncation mutants in pCDH were generated via site-directed mutagenesis or via DNA fragment synthesis followed by Gibson Assembly. STING (SAVI) with N^154^/S, or V^147^/L, or R^281^/Q mutations were cloned into pCW. For shRNA-mediated gene knockdown, shRNA-encoding sequences were cloned into lentivirus plasmid pLKO.1-puro. siRNAs were purchased from Thermo Fisher. siRNA and shRNA sequences are listed in the key resources table. All constructs were verified by sequencing.

HeLa, HEK293T and HaCaT cells were grown in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin and 100 μg/mL streptomycin. Non-transformed MCF10A cells were grown in DMEM/F-12 supplemented with 5% horse serum, 20 ng/mL EGF, 0.5 mg/mL hydrocortisone, 100 μg/mL cholera toxin, 10 μg/mL Insulin, 100 U/ml penicillin and 100 μg/mL streptomycin. MOLM-13 and THP-1 cells were grown in RPMI 1640 supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin and 100 μg/mL streptomycin. 2-mercaptoethanol (Gibco, 21985023) was added to the medium for THP-1 cells to a final concentration of 0.05 mM before use. NK92-MI cells were grown in MyeloCult H5100 medium (StemCell Technologies cat # 05150) supplemented with 1 μM hydrocortisone, 100 U/ml penicillin and 100 μg/mL streptomycin. HMEC were grown in Mammary Epithelial Cell Basal Medium (PCS-600-030) plus the components of the Mammary Epithelial Cell Growth Kit (PCS-600-040) supplemented with 100 U/ml penicillin and 100 μg/mL streptomycin. All cells above were cultured at 37° C with 5% CO_2_ in a humidified incubator.

Plasmid DNA was transfected into cells using TransIT-LT1 (Mirus, HEK293T) or Lipofectamine 3000 (Invitrogen, for HeLa, MCF10A and HaCaT cells) according to the protocols of the manufacturers. Lipofectamine RNAiMAX transfection reagent (Invitrogen) was used for siRNA transfection according to the protocol of the manufacturer.

Generation of stable cell lines of gene expression, knockdown and knockout

shRNA-mediated knockdown or overexpression of genes was done through lentiviral transduction. GCaMP6s-expressing lentiviruses were generated in HEK293T cells by co-transfecting pMDL, pRev, and pMD2.G together with pBOB-GCaMP6s. The packaging plasmids psPAX2 and pMD2.G were used to produce all other lentiviruses used in this study. Forty-eight hours after transfection, virus-containing medium was collected every 24 h for 2 times, and then filtered using Millex-HV Syringe Filter (0.45 μm, MilliporeSigma). Target cells were transduced with filtered viruses in the presence of polybrene (10 μg/mL). Cell lines expressing GCaMP6s were obtained through single cell cloning. Cells infected with shRNA-, STING-, TRPV2-, mCherry-, mNeonGreen-expressing lentiviruses were selected with puromycin (1.5 μg/mL for HeLa and HEK293T, 2 μg/mL for MCF10A, MOLM-13 and THP-1) for 2 days. All stable cell lines were used for experiments at least 7 days post-infection to ensure that no cytosolic DNA from reverse transcribed lentiviral RNA was present.

AMPKα-KO or CaMKK2-KO HeLa cells were generated using the CRISPR-Cas9 method, as previously described.^48^ Briefly, to create AMPKα-KO cell lines (deleted for both AMPKα1 and AMPKα2), pCRISPRv2-sgRNA constructs (which also express Cas9) targeting the human PRKAA1 and PRKAA2 genes were co-transfected into cells. Twenty-four hours after transfection, cells were selected with puromycin (1.5 μg/mL) for 2 days. Single cells were then grown in 96-well plates for amplification. Individual clones were verified by western blot to detect AMPKα expression. The same strategy was used to generate the CaMKK2-KO cell lines.

DNA and cGAMP transfection, drug treatment

To introduce cytosolic DNA through plasmid DNA transfection, empty pCDH vector (2 μg DNA for HeLa, HaCaT, and 4 μg DNA for MCF10A cells) was transfected into 1 × 10^6^ cells/well plated on a 6-well plate using Lipofectamine 3000. cGAMP (5 μg for HeLa and HaCaT cells, 10 μg for MCF10A cells) was transfected into cells using Lipofectamine 3000. Treatment conditions for drugs, including MnCl_2_, Etoposide, HU, Thapsigargin, A23187, Tranilast, and BAPTA-AM, were described in figure legends.

Immunoprecipitation and immunoblotting

For Flag-tagged protein immunoprecipitation, cells were lysed in the lysis buffer (10 mM NaKPO_4_, pH 7.2, 150 mM NaCl, 1% sodium deoxycholate, 1% Triton X-100, protease inhibitor cocktail, and phosphatase inhibitor cocktail) and then sonicated in iced water for 40 s (Gkika, 2015). The cell lysate was then centrifuged at 12,000 rpm for 10 min at 4° C. The supernatant was then incubated with anti-Flag M2 Magnetic Beads (Sigma, M8823) for 2 h at 4° C. After washing 5 times with wash buffer (10 mM NaKPO4, pH 7.2, 150 mM NaCl, 1% sodium deoxycholate, 1% Triton X-100), bead-bound proteins were dissolved in 2X SDS sample buffer and heated (37° C, 30 min for the samples for blotting for TRPV2, and 95° C, 10 min for blotting for all other proteins) before gel loading. For CaMKK2 KO efficiency detection, 2 μg of CaMKK2 antibody was incubated with 20 μL of Protein A-Sepharose beads at room temperature for 1 h with rotation. Antibody-bound beads were then incubated with cell lysates at 4° C overnight and washed 5 times with wash buffer as described above.

Immunoblotting was performed using DyLight 800- and DyLight 680-conjugated secondary antibodies and an Odyssey M Imaging System from LI-COR Biosciences as previously described.^84^ To detect pT172-phosophorylation of AMPKα, cells were quickly rinsed with cold PBS before flash-freezing by floating the dishes on liquid nitrogen. Cells were then lysed in lysis buffer on ice.

Immunofluorescence staining and confocal microscopy

Immunofluorescent staining was performed essentially as previously described.^49^ To detect the ER-to-Golgi translocation of STING, cells cultured on glass-bottomed dishes (Mattek, P35G-1.5-14-C) were washed once with PBS and then fixed with 4% PFA in PBS for 30 min at room temperature (RT). After washing three times with PBS, cells were permeabilized with PBS with 0.2% Triton X-100 for 20 min at RT. Cells were then washed three times with PBS and incubated in blocking buffer (PBS with 5% FBS, 3% BSA and 0.1% Triton) for 1 h at RT. After washing once with PBS with 0.05% Tween 20, cells were incubated overnight at 4° C with primary antibodies (1:1,000 for anti-HA and 1:500 for anti-GM130) in PBS with 1% FBS, 1% BSA and 0.1% Tween 20. After washing three times (each 10 min) with PBS with 0.05% Tween 20, cells were incubated at RT with secondary antibodies (1:500) in PBS with 1% FBS, 1% BSA and 0.1% Tween 20 for 1 h in a covered container. After washing once with PBS with 0.05% Tween 20 for 10 min, cells were incubated with Hoechst (100 ng/mL) in PBS with 0.1% Triton X-100 for 10 min. After this, cells were washed twice for 10 min each with PBS with 0.05% Tween 20 and once for 5 min with PBS. Fluorescence images were captured using a Nikon Ti2-E CREST X-Light V3 spinning disk confocal microscope.

RNA extraction and RT-qPCR

For RNA extraction, cells in a 6-well plate were lysed in 0.5 mL/well of TRIzol Reagent and samples were transferred to microtubes. Pipet the lysate up and down several times to homogenize. After 5 min incubation, 0.1 mL of chloroform per 0.5 mL TRIzol Reagent was added. After 15 s of rigorous shaking and 2–3 min incubation, samples were centrifuged at 12,000g for 15 min at 4° C. The aqueous phase containing the RNA was transferred to a new tube. 0.25 mL of isopropanol was then mixed with the aqueous phase per 0.5 mL of TRIzol Reagent used for lysis. RNA was pelleted by centrifugation for 10 min at 12,000g at 4° C. After removing the supernatant, 0.5 mL of 75% ethanol was added followed by brief vortex. Samples were then centrifuged for 5 min at 7500g at 4° C. After air drying for 5–10 min, the pellet was resuspended in 35 μL of RNase-free water with 0.1 mM EDTA and incubated at 55° C–60° C for 10–15 min. The trace amount of DNA was removed using a Turbo DNase enzyme kit (ThermoFisher, AM2238), according to the manufacturer’s protocol. RNA concentration was measured using a Nanodrop. For reverse transcription, the PrimeScript RT reagent kit (RR037A) was used. 500 ng RNA in 6.5 μL RNase-free H_2_O was incubated at 65° C for 5 min in a PCR machine, followed by the addition of 3.5 μL of a pre-made master mix of reaction components, consisting of 2 μL 5x PrimeScript buffer, 0.5 μL PrimeScript RT enzyme, 0.5 μL Oligo dT primers (50 μM), 0.5 μL and Random 6 mers (100 μM). The reactions were incubated in a PCR machine at 37° C for 15 min, 85° C for 5 s, and then 4° C to hold. The resulting cDNA was diluted 5–8 times in H_2_O, depending on the abundance of the genes of interest. Quantitative PCR was performed on ViiA 7 Real-Time PCR System or Applied Biosystems QuantStudio3 using PowerUp SYBR Green Master Mix (A25742) according to the manufacturers’ protocols. The following procedure was used for PCR amplification: samples were first incubated at 50° C for 2 min and then at 95° C for 2 min, followed by 40 cycles of denaturation (95 ° C, 1 s) and annealing/extension (60° C, 30 s).

ELISA for protein detection

The protein levels of IFN-β, IL-6, IL-1β and CCL2 were measured in HeLa cells using ELISA kits according to the manufacturers’ instructions. For sample preparation, HeLa cells were lysed in the lysis buffer (10 mM NaKPO4, pH 7.2, 150 mM NaCl, 1% sodium deoxycholate, 1% Triton X-100, protease inhibitor cocktail, and phosphatase inhibitor cocktail). After sonication in iced water for 1 min, samples were centrifuged at 12,000 rpm for 10 min at 4° C and the supernatant was used for ELISA to detect IFN-β, IL-6, IL-1β and CCL2 proteins. Signals were captured using a Cytation 5 microplate reader and analyzed using GraphPad Prism.

Analysis of the ER-to-Golgi translocation of STING

To assess STING translocation from the ER to Golgi, STING puncta formation and its colocalization with the Golgi marker GM130 were analyzed. For puncta analysis, immunofluorescence (IF) staining was performed in cells expressing STING-HA, and the percentage of cells containing STING puncta were determined by manually counting number of STING-HA positive cells with puncta vs. those without. For each sample, 1000 cells were scored. STING colocalization with GM130 with quantified using Fiji ImageJ. A Golgi mask was generated by thresholding the GM130 signal, and a background region was selected adjacent to the Golgi. Specific colocalized STING intensity was calculated by subtracting the mean intensity in the mean background intensity from the mean STING intensity within the Golgi mask. All data and statistical analyses were performed using GraphPad Prism.

Live cell imaging

For Ca^2+^ imaging in live cells, genetically encoded calcium indicators GCaMP6s were used to measure Ca^2+^ levels in the cytoplasm, as described before.^49,85^ Fluorescence signals of GCaMP6s were acquired with a 20× objective using a Nikon Ti2-E CREST X-Light V3 spinning disk confocal microscope in a live cell imaging chamber. 250 cells were scored for each sample. Fluorescence signals in individual cells were quantified using Fiji ImageJ after subtracting the background. To analyze TRPV2-STING colocalization, fluorescence signals of TRPV2-mCherry and STING-mNeonGreen expressed in HeLa cells were acquired with a 20× objective using the same microscope.

NK-92MI-mediated cells killing

mNeonGreen-expressing target HeLa cells were generated via lentiviral infection. For cell killing, target HeLa cells were plated in 12-well dishes and untransfected or transfected with cGAMP. Seven hours after transfection, cell medium was changed to the NK-92MI growth medium (MyeloCult H5100) and NK-92-MI cells were added at a target:effector ratio of 1:2. The co-culture was incubated in a Incucyte X5 and mNeonGreen-positive cells were imaged every 30 min and counted using a Cell-by-Cell software package.

QUANTIFICATION AND STATISTICAL ANALYSIS

Statistical significance tests were done with the GraphPad Prism software. The sample size (n), the number of independent replicates for each experiment, and the tests performed are described in the figure legends.

Supplementary Material

1

Supplemental information can be found online at https://doi.org/10.1016/j.celrep.2025.116745.