G9a inhibits MEF2C activity to control sarcomere assembly

In this study, we demonstrate that the lysine methyltransferase G9a inhibits sarcomere organization through regulation of the MEF2C-HDAC5 regulatory axis. Sarcomeres are essential for muscle contractile function. Presently, skeletal muscle disease and dysfunction at the sarcomere level has been associated with mutations of sarcomere proteins. This study provides evidence that G9a represses expression of several sarcomere genes and its over-expression disrupts sarcomere integrity of skeletal muscle cells. G9a inhibits MEF2C transcriptional activity that is essential for expression of sarcomere genes. Through protein interaction assays, we demonstrate that G9a interacts with MEF2C and its co-repressor HDAC5. In the presence of G9a, calcium signaling-dependent phosphorylation and export of HDAC5 to the cytoplasm is blocked which likely results in enhanced MEF2C-HDAC5 association. Activation of calcium signaling or expression of constitutively active CaMK rescues G9a-mediated repression of HDAC5 shuttling as well as sarcomere gene expression. Our results demonstrate a novel epigenetic control of sarcomere assembly and identifies new therapeutic avenues to treat skeletal and cardiac myopathies arising from compromised muscle function.

Consistent with their central role in regulating MEF2 activity, class II HDACs regulate fiber type specific programs that are MEF2-dependent 22,23 . In addition, they mediate the repressive effects on MEF2-dependent structural and contractile genes in vivo 24 . As such, increased expression of class II HDACs in skeletal muscle are associated with congenital myopathies 25 that display disruption of sarcomere assembly and reduced expression of MEF2 target genes such as Lmod3 26 . Another co-repressor that has been implicated in the regulation of muscle structural and sarcomere genes is G9a. G9a, also known as Euchromatic Histone-lysine N-methyltransferase 2 (EHMT2) or KMT1C, contains a C-terminal enzymatic Su(var)3-9, Enhancer of zeste and Trithorax (SET) domain which mediates repressive histone H3 lysine 9 di-methylation marks at euchromatic regions of the genome [27][28][29] . In embryonic stem cells, G9a exists in a hetero-dimer with its paralog, G9a-Like Protein (GLP), and notably, loss of G9a in mice results in embryonic lethality 30 . Interestingly, however, conditional knockout of G9a in postnatal neurons causes de-repression of muscle sarcomere genes, demonstrating its critical role as a repressor of the contractile machinery in vivo 31 . Although implicated in myogenesis 32,33 , whether G9a regulates sarcomere genes in skeletal muscle cells, and the mechanisms by which it does so have not been investigated.
In this study, we demonstrate that G9a represses sarcomeric genes and inhibits sarcomere assembly by impacting the HDAC5-MEF2 circuitry. Down-regulation of G9a expression or pharmacological inhibition of its methyltransferase activity in skeletal myoblasts results in the up-regulation of several MEF2-dependent structural and sarcomeric genes, including myomesins and myozenin. We demonstrate that G9a associates with MEF2C and blocks nuclear export of HDAC5, resulting in repression of MEF2C transcriptional activity. G9a occupancy and histone H3 lysine 9 di-methylation (H3K9me2) marks are apparent at MEF2 sites on sarcomeric gene promoters, and its over-expression results in sarcomere disorganization. Activation of calcium signaling in G9a over-expressing cells restores HDAC5 shuttling and MEF2C-dependent sarcomere gene expression. Together, these studies provide evidence that G9a forms a repressive complex with MEF2C and HDAC5 to epigenetically control sarcomere organization.

Results
Sarcomeric genes are regulated by G9a. Previous studies have shown that G9a regulates MEF2 transcriptional activity 33 which plays an essential role in sarcomere assembly 20,21 . Interestingly, our recent transcriptome analysis of G9a knockdown myoblasts 34 revealed differential expression of several genes that encode sarcomere proteins, including thin actin filament (Actc1), thick myosin filament (Myh3, Myl4), Z-disc components (Actn2, Csrp3, Fbxo32, Myoz2, Unc45b), M-band (Lmod3, Myom2, Myom3), troponin complex (Tnnc2, Tnni2, Tnnt3), titin (Ttn) and melusin (Itgb1bp2) (Fig. 1a). Since MEF2 directly regulates the expression of many sarcomere and structural genes 3,20 , we hypothesized that G9a may repress their expression via inhibition of MEF2 transcriptional activity. As a start point, we analyzed the expression of Myom2, Myom3, Actn2, Myoz2 and Ttn in C2C12 cells transfected with small interfering RNA (siRNA) specific for G9a (siG9a) compared to control cells (siControl) transfected with scrambled siRNA. Consistent with the microarray data 34 , the expression of these genes was up-regulated at both mRNA and protein levels in proliferating siG9a cells compared to control cells (Fig. 1b). Similar results were obtained in siG9a primary myoblasts ( Supplementary Fig. 1a). To determine if methyltransferase activity is important for repression of these genes, C2C12 cells were treated with UNC0638, a small molecule compound that inhibits G9a activity 35 . Upon treatment with UNC0638, expression of sarcomere genes was increased in C2C12 cells and primary myoblasts (Fig. 1c, Supplementary Fig. 1b), whereas global H3K9me2 was reduced, indicating that the methyltransferase activity of G9a is important for repression of sarcomeric genes. Notably, knock-down of GLP did not result in an increase in sarcomeric genes ( Supplementary Fig. 2), suggesting that the increase in sarcomeric gene expression upon UNC0638 treatment is likely due to inhibition of G9a activity. To examine the functional consequence of G9a-mediated regulation of the contractile apparatus, we retrovirally over-expressed G9a (pBABE-G9a) in C2C12 cells (Fig. 1d). The expression of sarcomeric genes was repressed in proliferating G9a over-expressing cells compared to control cells (Fig. 1d). More importantly, upon differentiation, regularly spaced Actn2 staining in the Z-disc, and F-actin in the thin filament, were evident in control cells. In contrast, pBABE-G9a myotubes showed reduced and irregularly spaced Actn2 staining, and F-actin staining did not reveal any discernible structures (Fig. 1e). Together, these results indicate that sarcomeric structure and myofibril formation was perturbed by G9a over-expression.
G9a deposits repressive epigenetic marks at sarcomeric gene promoters. To examine the mechanisms by which G9a regulates sarcomeric gene expression, we first analyzed the endogenous expression profile of G9a and sarcomeric genes during myogenic differentiation. Consistent with previous studies 32-34 , G9a expression decreased at the mRNA and protein level upon myogenic differentiation and was inversely correlated with expression of sarcomere genes and MEF2C (Fig. 2a,b).
G9a primarily represses its target genes through mediation of histone H3 lysine 9 di-methylation (H3K9me2) repression marks. We therefore performed chromatin immunoprecipitation (ChIP) assays for G9a occupancy and H3K9me2 enrichment at the Myom2 and Myoz2 promoters which have been previously characterized as MEF2 targets 20,36 . In control cells, G9a occupancy (Fig. 2c) and its signature H3K9me2 enrichment (Fig. 2d) were apparent at these promoters in undifferentiated myoblasts (D0), and the occupancy was reduced during differentiation (D2). In pBABE-G9a cells, G9a occupancy and H3K9me2 marks were increased both at the undifferentiated state and upon differentiation. Moreover, this repression mark present at both promoters was reduced in UNC0638-treated cells (Fig. 2e) indicating that G9a represses expression of these genes in a methyltransferase activity-dependent manner.
G9a interacts with and inhibits MEF2C activity. While all MEF2 factors are expressed in skeletal muscle, only the knock-out of MEF2C in mice disrupts sarcomere organization 20 . We therefore examined whether G9a-mediated repression of sarcomeric genes was mediated through an impact on MEF2C. Due to high Scientific RepoRts | 6:34163 | DOI: 10.1038/srep34163 The list is from data set GSE70039 34 . (b) C2C12 myoblasts transfected with scrambled siRNA (siControl) or G9a-specific siRNA (siG9a). The mRNA expression of Myom2, Myom3, Actn2, Myoz2 and Ttn was analyzed by qPCR. The expression in control cells was given an arbitrary value of 1. The fold change relative to control cells is shown. Lysates were analyzed for G9a, Myom2 and Actn2 expression by western blot. β -actin was analyzed as a loading control. (c) Undifferentiated C2C12 cells were treated with DMSO or UNC0638 for 48 hr. mRNA expression of sarcomeric genes in the presence or absence of UNC0638 treatment was examined by qPCR. Myom2 and Actn2 protein levels as well as global H3K9me2 levels were analyzed by western blot. β -actin and Histone H3 were used as loading controls. (d) C2C12 cells were transduced with empty retroviral vector (pBABE) or with pBABE-G9a. G9a expression was analyzed by western blot (upper panel). Densitometric analysis of the western blot showed ~4.5 fold higher expression of G9a in pBABE-G9a cells (lower panel). Sarcomeric gene expression in undifferentiated pBABE and pBABE-G9a cells was probed by qPCR. (e) pBABE and pBABE-G9a cells were differentiated for six days. Actn2 (green), a marker of the Z-disc, and F-actin (red) within the thin filament that form part of the sarcomere were analyzed by immunofluorescence. Confocal images were taken at 100X magnification with 3X focus. Scale bar: 5μ m. Error bars indicate mean ± standard error of n = 3. Numbers indicate molecular weight of proteins.
Scientific RepoRts | 6:34163 | DOI: 10.1038/srep34163 transfection efficiency, we first checked whether G9a interacts with MEF2C in HEK293T cells. EGFP-G9a and Flag-MEF2C were expressed individually or together. Immunoprecipitation of MEF2C revealed a clear interaction with G9a that was not altered upon UNC0638 treatment (Fig. 3a). In reciprocal experiments, MEF2C was co-immunoprecipitated with full length G9a as well as G9a lacking the SET domain (G9a ∆ SET) (Fig. 3b), suggesting that the SET domain is not essential for interaction with MEF2C. To confirm their association in myogenic cells, we immunoprecipitated endogenous G9a from myoblasts and verified its association with endogenous MEF2C (Fig. 3c). To identify the domain in MEF2C that interacts with G9a, Flag-MEF2C deletion mutants (Fig. 3d, left panel) were over-expressed together with EGFP-G9a. G9a immunoprecipitated with full-length MEF2C (FL) and with MEF2C lacking the C-terminal transactivation region (N177), but not with a MEF2C deletion mutant lacking the MADS domain (∆ MADS) (Fig. 3d, right panel).
We then examined the impact of G9a on MEF2 transcriptional activity. C3H10T1/2 (10T1/2) fibroblasts were transfected with a synthetic MEF2-responsive reporter (3XMEF2-Luc) that contains three MEF2-binding sites upstream of the luciferase gene 9 . As expected, the reporter was activated in response to MEF2C. G9a inhibited MEF2C activity in a dose-dependent manner in 10T1/2 cells (Fig. 4a) as well as in C2C12 myoblasts (Fig. 4b).
As G9a-mediated regulation of sarcomeric genes involves its methyltransferase activity, we tested if repression of MEF2C transcriptional activity is dependent on its catalytic activity. Treatment of cells with UNC0638 rescued G9a-mediated repression of MEF2C (Fig. 4c). Consistently, repression of MEF2C was significantly lesser in the presence of G9a ∆ SET compared to full length G9a (Fig. 4d). To validate these findings on natural MEF2 target gene promoters, we analyzed Myom1 and Myom2 that have previously been characterized as direct MEF2C targets 20 . In the presence of G9a, MEF2C-mediated activation of both promoters was inhibited (Fig. 4e,f). Moreover, mutation of MEF2 DNA-binding site in both promoters (pGL3-Myom1∆ MEF2 and pGL3-Myom2∆ MEF2 respectively) abrogated the impact of MEF2C and G9a, and UNC0638 treatment de-repressed MEF2C activity. G9a promotes HDAC5 association with MEF2C. To examine the mechanisms by which G9a inhibits MEF2C transcriptional activity, we analyzed MEF2C sub-cellular localization. MEF2C was found solely in the nucleus, and no difference was observed between pBABE control cells and pBABE-G9a cells (Fig. 6a). In addition, no changes in MEF2C occupancy at Myom2 and Myoz2 promoters were apparent (Fig. 6b) indicating G9a does not alter MEF2C localization or its DNA-binding ability.
Class II HDACs are well-known repressors of MEF2 activity 6 . Notably, class II HDACs have been also implicated in regulation of sarcomeric gene expression by repressing MEF2 activity 22,24 . We therefore examined if G9a modulates this regulatory axis. We focused on HDAC5 because HDAC4 is predominantly cytoplasmic in undifferentiated C2C12 cells 37 . To examine whether G9a has an impact on the association of MEF2 and HDAC5, HDAC5 was immunoprecipitated in control and G9a over-expressing cells. As expected, interaction of HDAC5 and MEF2C was apparent in undifferentiated control cells. Interestingly, this association was enhanced in pBABE-G9a cells despite MEF2C and HDAC5 being expressed at the same level as control cells (Fig. 6c). Moreover, HDAC5 occupancy at Myom2 and Myoz2 gene promoters was decreased in siG9a myoblasts (Fig. 6d) and increased in pBABE-G9a myoblasts (Fig. 6e). During myogenic differentiation, in response to calcium signaling, HDAC5 is phosphorylated at serine (S)259 and S498 and shuttled to the cytoplasm 38 . Given the enhanced association of HDAC5 with MEF2C in proliferating G9a over-expressing cells, we wondered whether phosphorylation and subsequent dissociation of HDAC5 from MEF2C was affected. The level of p-HDAC5 (S498) was reduced in G9a over-expressing cells compared to control cells (Fig. 6f). We then examined HDAC5 localization  in fractioned nuclear and cytoplasmic lysates from control and G9a over-expressing cells. In undifferentiated G9a over-expressing cells, nuclear HDAC5 levels were increased compared to control cells. Correspondingly, the levels were reduced in the cytoplasm in differentiated cells (Fig. 6g). Together, these results suggest that G9a may inhibit calcium signaling and thereby regulate the association of MEF2C with HDAC5. To test this directly, we used a calcium-selective indicator Fura-2 AM. Intracellular Ca 2+ levels were lower in undifferentiated pBABE-G9a cells relative to control cells (Fig. 7a). We then examined HDAC5 phosphorylation and the expression of  MEF2-dependent sarcomeric genes in cells treated with ionomycin, which promotes influx of extracellular calcium into cells to trigger calcium signaling. In response to ionomycin, increased HDAC5 phosphorylation at S498 was apparent in G9a over-expressing cells (Fig. 7b). Concomitantly, a rescue of MEF2C transcriptional activity (Fig. 7c) and expression of MEF2C target genes Myom2 and Myoz2 was evident (Fig. 7d). Similarly, when constitutively active CaMKI was expressed in pBABE-G9a cells, a higher level of HDAC5 phosphorylation (Fig. 7e) coincided with increased MEF2C activity (Fig. 7f) and MEF2C target gene expression (Fig. 7g).

Discussion
The results of this study reveal for the first time that G9a epigenetically controls structural elements in skeletal muscle. G9a interacts with the MEF2C-HDAC5 complex in proliferating myoblasts, and mediates repressive H3K9me2 marks at MEF2 sites in sarcomeric gene promoters that lead to disrupted sarcomere organization.
G9a is expressed in proliferating myoblasts and orchestrates repression of skeletal myogenesis at multiple levels. In addition to regulating cell cycle progression 34 and the expression of differentiation specific genes 32,33,39 , G9a represses maturation by inhibiting expression of sarcomere genes in proliferating myoblasts. The expression of sarcomere/structural genes is up-regulated in proliferating G9a knock down cells. Moreover, G9a occupancy and H3K9me2 enrichment are apparent at sarcomere gene promoters in growth conditions. These observations argue that repression of sarcomere genes is not secondary to G9a-mediated block of myogenic differentiation, but rather reflect an active transcriptional control of contractile apparatus. Consistent with this contention, conditional ablation of G9a in the postnatal brain using Camk2a-Cre results in the up regulation of muscle structural and sarcomere genes in neurons 31 validating that G9a represses muscle sarcomere and structural genes in vivo, independent of myogenic differentiation.
Among the various transcription factors that regulate expression of sarcomere genes in skeletal muscle, MEF2 factors are critical 40 . Disruption of MEF2C has been shown to perturb sarcomere morphology and integrity in various species due to reduced expression of M-line and Z-disc genes. G9a does not appear to inhibit MEF2C expression or block MEF2C DNA-binding ability. However, G9a interacts with MEF2C and blocks its transcriptional activity by enhancing its association with HDAC5. Class II HDACs associate with HDAC3/SMRT/NCoR complex 7 , as well as SUV39H1/HP1 to mediate transcriptional repression 8 . G9a has also been shown to interact with class I HDACs 41 , as well as with SUV39H1 and SETDB1 42 . Thus G9a is likely present in multi-protein repressor complexes that assemble to block MEF2 function in proliferating myoblasts. Class II HDACs are known to undergo nucleocytoplasmic shuttling in response to various signals. During myogenic differentiation, CaMK phosphorylates HDAC5 at S259 and S498 that leads to recruitment of 14-3-3 proteins and its nuclear export. In addition, cAMP/PKA also phosphorylates HDAC5 at S280 to block its nuclear export 43 . A reduction in p-HDAC5 S498 is apparent in G9a overexpressing cells. Conversely, activation of calcium signaling or expression of constitutively active CaMK rescues p-HDAC5 S498 levels and sarcomere gene expression in presence of G9a. Together, these data indicate that G9a modulates calcium signaling dependent HDAC5 dissociation from MEF2C, rather than impact PKA-mediated phosphorylation of HDAC5. Consistent with this notion, nuclear HDAC5 is increased in undifferentiated G9a over-expressing cells, along with a decrease in the cytoplasmic pool in differentiated cells. Nuclear retention of HDAC5 in G9a overexpressing cells may account for the increased association of MEF2 and HDAC5.
In addition to sarcomere genes, MEF2 and class II HDACs (HDAC4) regulate many components of the costamere. Moreover, dystrophin, β -and γ -sarcoglycan of the dystrophin-glycoprotein complex (DGC) are also targets of the class II HDAC-MEF2 axis 24,44 . Interestingly, the expression of β and γ sarcoglycans were also up-regulated in G9a knockdown myoblasts (not shown), suggesting that G9a may also directly regulate components of the DGC complex and the costamere genes in skeletal muscle. The impact of G9a in the regulation of muscle contractile machinery may not be mediated exclusively through MEF2C. G9a influences the function of number of transcription factors such as MyoD 32 , HIF1α 45 , NF-kB 46 that play roles in sarcomere and structural gene expression 40 .
In line with our observations, G9a was recently shown to repress Myh6, an isoform of myosin heavy chain, that is expressed in the heart and whose loss of expression causes contractile dysfunction. G9a expression is elevated in human hypertrophic hearts and contributes to hypertrophic cardiomyopathy 47 . It is noteworthy that hypertrophic cardiomyopathy is found in patients with mutations in MYOM1 48 , ACTN2 49 and MYOZ2 50 , and deletion of Myoz2 in mice increases susceptibility to hypertrophic cardiomyopathy upon cardiac stress 51 . On the other hand, G9a expression is decreased in mouse models of dilated cardiomyopathy 52 whereas increased MEF2 activity in the heart causes dilated cardiomyopathy 53,54 . The impact of G9a on the MEF2-class II HDACs regulatory axis may be relevant in cardiac myopathies associated with dysregulation of sarcomere genes. Pharmacological targeting of G9a activity may be a viable therapeutic avenue in skeletal and cardiac pathologies associated with dysregulation of the MEF2-class II HDAC regulatory circuit.

Materials and Methods
Cell culture. C2C12 mouse myoblasts were cultured in growth medium [Dulbecco's Modified Eagle's Medium (DMEM) high glucose supplemented with 20% fetal bovine serum (FBS)]. To induce differentiation, cells were cultured in differentiation medium (DMEM supplemented with 2% horse serum). Primary myoblast cultures were established from C57BL/6 wild type mice as previously described 55 . All experiments were done in accordance to relevant guidelines. The experimental protocols were approved by the National University of Singapore Institutional Animal Care and Use Committee. Purity of myoblasts was checked by Pax7 immunostaining, and cells were used only when Pax7 positivity was > 95%. Primary myoblasts were cultured on collagen-coated plates in primary culture medium [Ham's F-10 supplemented with 20% FBS and 5ng/ml basic fibroblast growth factor (bFGF)]. HEK293T, Phoenix and C3H10T1/2 (10T1/2) cells were grown in DMEM supplemented with 10% FBS. All cells were incubated at 37°C with 5% CO 2 . For inhibition of methyltransferase activity, cells were cultured in growth medium containing 250 nM of UNC0638 35 for 48 hours (hr) before being harvested. For induction of calcium uptake, C2C12 cells were treated with 2 μ M of ionomycin in growth medium for 4 hr.  38 . Mouse Myom1 promoter (− 1035 to + 88) and Myom2 promoter (− 1058 to + 6) were cloned into pGL3-basic vector (Promega) using primers described previously 20 . Amplification of the Myom1 and Myom2 promoters were performed using the Expand High Fidelity PLUS PCR System (Roche), with mouse chromatin from C2C12 cells as template. Amplicons were digested with KpnI and HindIII for Myom1 promoter, or MluI and HindIII for Myom2 promoter (New England Biolabs), for which restriction sites were introduced into the PCR primers, and ligated into the corresponding sites on the pGL3-basic vector to produce pGL3-Myom1 and pGL3-Myom2 reporter constructs. For pGL3-Myom1 ∆ MEF2 and pGL3-Myom2 ∆ MEF2, four residues within each of the two MEF2 binding sites were mutated that were shown to abrogate MEF2 binding 20  RNA isolation and quantitative real-time polymerase chain reaction (qPCR). Total RNA from cells was extracted using TRIzol reagent (Invitrogen) and treated with DNase using the TURBO DNA-free Kit (Life Technologies). For qPCR, 1 μ g of RNA was converted to complementary DNA (cDNA) with the Superscript III First-Strand Synthesis System using oligo-dT primers (Invitrogen) and subsequently treated with RNase H. The cDNA was then probed through qPCR in triplicates using the LightCycler 480 SYBR Green I Master mix (Roche) and 0.2 μ M of primers and detected using the LightCycler 480 I instrument (Roche) with the following PCR parameters: denaturation at 95 °C for 5 min; amplification for 45 cycles of 95°C for 10 s, 60°C for 10 s and 72°C for 10 s. Melting curve analysis was tested to ensure presence of only a single amplicon product for each primer. Threshold cycle (Ct) values were analyzed on default settings. Relative mRNA expression was calculated using 2 −∆Ct method upon normalization to Ct for Gapdh and expressed as fold change relative to the respective control. Primer sequences for G9a and Gapdh are as previously published 32 . Primers for Myom2, Myom3, Myoz2, Actn2 and Ttn were validated and obtained from Primer Bank 58 . Immunoprecipitation (IP) assay, fractionation, and immunoblotting. Protein lysates were harvested from cells in Radioimmunoprecipitation assay (RIPA) lysis buffer [50 mM NaCl, 50 mM Tris-HCl pH 6.8, 1 mM EDTA, 1% NP-40, 0.1% sodium deoxycholate, protease inhibitor cocktail (Roche)]. Histones were isolated using RIPA lysis buffer supplemented with 2% sodium dodecyl sulphate (SDS).
For cytoplasmic and nuclear fractionation, cells were re-suspended in cytoplasmic extraction buffer [10 mM HEPES, pH7.5, 10 mM KCl, 0.1 mM EDTA, 0.5 mM DTT, 0.1% NP-40 and Protease and Phosphatase Inhibitor (Thermo Scientific)] and incubated on ice for 15 minutes. Lysates were centrifuged at 12,000 g for 10 minutes and supernatant was harvested as cytoplasmic extract. The pellets were washed thrice in cytoplasmic extraction buffer without NP-40. After the last wash, the pellet was re-suspended in RIPA lysis buffer and harvested as nuclear extract.
For immunoprecipitation, lysates were incubated with anti-Flag agarose beads (Sigma-Aldrich). To immunoprecipitate endogenous G9a and HDAC5, 1 mg of lysate was incubated with 2 μ g of primary antibody Flag-G9a, or 200 ng of Flag-G9a∆ SET. After 24 hr, cells were washed with PBS and lysed with Passive Lysis Buffer. Readings were taken using the Varioskan Flash Multimode Reader (Thermo Scientific). Where indicated, 24 hr after transfection, cells were treated with 250 nM UNC0638 for 48 hr. For ionomycin treatment, cells were allowed to recover after transfection for 24 hr before being treated for 4 hr with 2 μ M of ionomycin.

Intracellular calcium assay. Undifferentiated pBABE and pBABE-G9a cells cultured in flat-bottom black
96-well plates were loaded with 5 μ M Fura-2 AM (F1201, Life Technologies) in PBS for 30 min at 37 °C, after which excess Fura-2 AM was washed off. Intracellular levels of Ca 2+ was then measured using Varioskan Flash Multimode Reader (Thermo Scientific) with excitation wavelength of 340 nm or 380 nm with 5 nm bandwidth, and detection of emission at 510 nm for both excitation wavelengths at 1 s interval. Levels of intracellular Ca 2+ was calculated as a ratio of emission at 340 nm over emission at 380 nm and shown relative to control.