RyR2/IRBIT regulates insulin gene transcript, insulin content, and secretion in the insulinoma cell line INS-1

The role of ER Ca2+ release via ryanodine receptors (RyR) in pancreatic β-cell function is not well defined. Deletion of RyR2 from the rat insulinoma INS-1 (RyR2KO) enhanced IP3 receptor activity stimulated by 7.5 mM glucose, coincident with reduced levels of the protein IP3 Receptor Binding protein released with Inositol 1,4,5 Trisphosphate (IRBIT). Insulin content, basal (2.5 mM glucose) and 7.5 mM glucose-stimulated insulin secretion were reduced in RyR2KO and IRBITKO cells compared to controls. INS2 mRNA levels were reduced in both RyR2KO and IRBITKO cells, but INS1 mRNA levels were specifically decreased in RyR2KO cells. Nuclear localization of S-adenosylhomocysteinase (AHCY) was increased in RyR2KO and IRBITKO cells. DNA methylation of the INS1 and INS2 gene promotor regions was very low, and not different among RyR2KO, IRBITKO, and controls, but exon 2 of the INS1 and INS2 genes was more extensively methylated in RyR2KO and IRBITKO cells. Exploratory proteomic analysis revealed that deletion of RyR2 or IRBIT resulted in differential regulation of 314 and 137 proteins, respectively, with 41 in common. These results suggest that RyR2 regulates IRBIT levels and activity in INS-1 cells, and together maintain insulin content and secretion, and regulate the proteome, perhaps via DNA methylation.

www.nature.com/scientificreports/ cell Ca 2+ measurements using fura-2 AM confirmed the loss of caffeine sensitivity in RyR2 knock out (RyR2 KO ) cells (Fig. 1c); however, these cells displayed a strong increase in [Ca 2+ ] in in response to the muscarinic agonist carbachol (500 µM) (Fig. 1c), suggesting that IP 3 receptor activity is intact. Analysis of single-cell Ca 2+ imaging experiments showed that the Ca 2+ response (AUC) to caffeine was reduced by > 90% in RyR2 KO cells, compared to control INS-1 cells (Fig. 1d). The absence of RyR2 protein in RyR2 KO cells was confirmed by immunoblotting of microsomal proteins from control and RyR2 KO
Ca 2+ dynamics in RyR2 KO cells. Stimulation of both control INS-1 and RyR2 KO cells with 7.5 mM glucose resulted in periodic [Ca 2+ ] in oscillations (Fig. 2a). The Ca 2+ integral was inhibited in both cell lines by 2 µM nicardipine; however, the total AUC in the absence of nicardipine was greater in RyR2 KO cells compared to control cells (Fig. 2b). The Ca 2+ response to 7.5 mM glucose was inhibited by 1 µM xestospongin C (xesto) in RyR2 KO cells but not in controls (Fig. 2c-e). Xesto also increased the time between peaks in RyR2 KO cells (Fig. 2f), but reduced the time between peaks in control INS-1 cells (Fig. 2f). Thus, IP 3 receptors are major contributors to glucose-stimulated Ca 2+ oscillations in RyR2 KO cells, but not in control INS-1 cells.

Regulation of IRBIT by RyR2.
Given the apparent increase in IP 3 receptor activation in response to glucose that we observed in RyR2 KO cells, we examined the ability of glucose to activate phospholipase C (PLC) in both RyR2 KO and INS-1 cells. Stimulation with 7.5 mM glucose (Fig. 3a) or 500 µM of the muscarinic receptor agonist carbachol (Fig. 3b) stimulated PLC activity above basal levels in both control and RyR2 KO cells. However, stimulated PLC activity was decreased in RyR2 KO cells compared to control INS-1 cells (Fig. 3c). Thus, the increased IP 3 receptor activity observed with glucose stimulation in RyR2 KO cells is unlikely the result of increased PLC activity and greater accumulation of IP 3 . Total cellular phosphatidylinositol bisphosphate (PIP 2 ) levels in fixed, saponin-treated cells was measured using immunocytochemistry. RyR2 KO cells contained slightly more PIP 2 than control INS-1 cells (Fig. 3d). Thus, the reduced PLC activity in RyR2 KO cells is unlikely to be the result of limiting substrate levels. Given these findings, we measured the levels of the protein IRBIT (aka AHCYL1). Using semi-quantitative immunoblotting, we found that IRBIT protein levels normalized to actin, were substantially reduced (~80% reduction) in RyR2 KO cells compared to control cells (Fig. 3e). These data suggest that in the absence of RyR2, IRBIT protein levels/activity are suppressed, allowing hyperactivation of IP 3 Rs.

Characterization of IRBIT KO cells.
To decipher which effects of RyR2 deletion are likely directly due to loss of ER Ca 2+ release via RyR2, and which are likely mediated by dysregulation of IRBIT, we deleted IRBIT from INS-1 cells using CRISPR/cas9 gene editing with gRNAs targeted to exon 6 of the AHCYL1 gene. Genomic DNA sequencing identified clones with expected indels (Fig. 4a). Immunoblots of cell lysates from IRBIT KO cells confirmed the absence of IRBIT protein (Fig. 4b). IRBIT mRNA was reduced in IRBIT KO cells compared to controls, but IRBIT mRNA wasn't reduced in RyR2 KO cells (Fig. 4c). Basal (2.5 mM glucose) [Ca 2+ ] in measured with fura-2 AM was greater in IRBIT KO cells compared to RyR2 KO cells, but was not different from that measured in control INS-1 cells (Fig. 4d). The [Ca 2+ ] in response to 5 mM caffeine in IRBIT KO cells, measured with fura-2 AM, was reduced compared to control INS-1 cells, but was much greater than that observed in RyR2 KO cells (Fig. 4e). Using the ER-targeted Ca 2+ indicator D1ER 22 to measure ER Ca 2+ levels, we found that basal ER [Ca 2+ ] was reduced in IRBIT KO cells compared to both RyR2 KO and control INS-1 cells, and that thapsigargin treatment reduced ER [Ca 2+ ] to levels that were not different across the three cell lines (Fig. 4f). The [Ca 2+ ] in response to 7.5 mM glucose in IRBIT KO cells was inhibited by 1 µM xesto (Fig. 4g, h), and was greater than in control INS-1 cells, but not different from that measured in RyR2 KO cells (Fig. 4i). Thus, abolition of the caffeine response in RyR2 KO cells is the direct result of the deletion of RyR2, but IRBIT is required to maintain the full magnitude of RyR2-mediated Ca 2+ release. In contrast, increased Ca 2+ response to 7.5 mM glucose and block of this response by xesto in RyR2 KO cells is likely the result of reduced IRBIT levels.
Insulin content and secretion in RyR2 KO and IRBIT KO cells. Since [Ca 2+ ] in is a key regulator of insulin secretion, we measured insulin secretion in response to glucose in control INS-1, RyR2 KO , and IRBIT KO cells. We examined glucose-stimulated insulin secretion (GSIS) at 2.5 mM and 7.5 mM glucose in all three cell lines, and examined the contribution of L-type Ca 2+ channels and IP 3 receptors to GSIS. 2 µM nicardipine (L-type channel blocker) completely inhibited 7.5 mM GSIS in all three cell lines, but 1 µM xesto didn't affect 7.5 mM GSIS in any of the cell lines. In each case, nicardipine suppressed GSIS to a level not different from that stimulated by 2.5 mM glucose (Fig. 5a). Insulin secretion at both 2.5 mM and 7.5 mM glucose was reduced in both RyR2 KO and IRBIT KO cells compared to controls (Fig. 5b). Insulin content was reduced ~ 70% in RyR2 KO cells and ~ 40% in IRBIT KO cells compared to control INS-1 cells as measured by insulin assay of ethanol/HCl-extracted cells, normalized to protein (Fig. 5c). ER Ca 2+ release contributes to glucose-dependent activation of extracellularsignal regulated protein kinase (ERK) 1/2 3 , which phosphorylates and activates several transcription factors involved in positive regulation of insulin transcription 23 . 7.5 mM glucose stimulated an ~ threefold increase in pERK 1/2 compared to 2.5 mM glucose in control, RyR2 KO , and IRBIT KO cells (Fig. 5d). Glucagon-like peptide 1 (50 nM) potentiated pERK in the presence of 7.5 mM glucose in all three cell lines. Epidermal growth factor (15 nM) potentiated pERK in the presence of 7.5 mM glucose in control and RyR2 KO cells, but this potentiation was abolished in IRBIT KO cells (Fig. 5d)

Regulation of AHCY localization by RyR2 and IRBIT.
One potentially global effect of IRBIT deletion is dysregulation of methyltransferase activity. IRBIT binds to S-adenosyl homocysteinase (AHCY) 24,25 , the only enzyme known to hydrolyze S-adenosyl homocysteine (SAH) and relieve product inhibition of DNA, RNA, and protein methyltransferases, and regulates nuclear localization of AHCY 25 . We examined the subcellular localization of AHCY in control, RyR2 KO , and IRBIT KO cells using immunocytochemistry. Confocal micrographs www.nature.com/scientificreports/ were taken of fixed cells labeled with a primary AHCY antibody, and an Alexa Fluor 488-conjugated secondary antibody, counterstained with Hoechst 33,342 to define the nuclei (Fig. 6a). Control experiments with primary antibodies omitted resulted in cells with negligible Alex Fluor 488 fluorescence (Fig. 6b). In control INS-1 cells, AHCY is preferentially localized in the nucleus relative to the cytoplasm (4:1 ratio) (Fig. 6a, c). In RyR2 KO cells, AHCY localized to the nucleus relative to the cytoplasm (4:1 ratio), but more AHCY was detected in the nucleus compared to control INS-1 cells. Deletion of IRBIT resulted in a marked depletion of AHCY in the cytoplasm, resulting in a nucleus to cytoplasm ratio of 13:1 (Fig. 6a, c). However, the total amount of AHCY detected in the nucleus of IRBIT KO cells was reduced compared to control INS-1 cells (Fig. 6c). Thus, deletion of RyR2 or IRBIT correlates with either increased nuclear AHCY (RyR2 KO cells), or a sharp decrease in non-nuclear localized AHCY (IRBIT KO cells). Increased accumulation of AHCY in the nucleus could potentially increase DNA methyltransferase activity (Fig. 6d).

Regulation of INS1 and INS2 gene methylation by RyR2 and IRBIT.
We examined the possibility that insulin genes in the knockout cells were differentially methylated. PCR amplification of genomic DNA regions, with or without digestion by a methylation-dependent endonuclease using primers that flank potential methylation sites (Fig. 7a, b), provides a measure of the relative amount of DNA methylation in the amplified region 26 . Comparing PCR amplification at promoter regions upstream of the translation start site of the INS1 (Fig. 7c) and INS2 (Fig. 7d) genes revealed low methylation that was not different between RyR2 KO , IRBIT KO , or control cells. The single CpG site in intron 1 of the INS1 gene (1-UP4) was extensively methylated but not altered by deletion of RyR2 or IRBIT (Fig. 7c). Increased methylation was observed in the proximal portion of exon 2 of INS1 (Fig. 7c). At 1DS3, DNA methylation was increased in IRBIT KO and RyR2 KO cells compared to controls. In the 1DS2 region, increased DNA methylation was only observed in the IRBIT KO cells. DNA methylation in the 1DS2 region is much higher (~ tenfold) compared to the 1-DS1 and 1-DS3 regions. A similar analysis of the INS2 gene (Fig. 7d) showed high methylation at regions downstream of the translation start site compared to upstream regions. An increase in DNA methylation was observed in Exon 2 of the INS2 gene of IRBIT KO and Glyceraldehyde-3-phosphate dehydrogenase mRNA was used as the reference for the data shown. Equivalent results were obtained using phosphoglycerate kinase mRNA as the reference (data not shown). Lines represent mean ± SD.  S1) and downregulation of 155 proteins (Fig. S2). Deletion of IRBIT resulted in increased levels of 75 proteins (Fig. S3) and decreased levels of 62 proteins (Fig. S4). Of these, 24 were more abundant in both RyR2 KO and IRBIT KO cells (Fig. 8a) and 17 were less abundant in both RyR2 KO and IRBIT KO cells (Fig. 8b). Gene ontology analysis for overrepresentation of differentially regulated proteins in specific cellular component, biological process, or molecular function categories (Figs. S5 and S6), revealed that proteins more abundant in RyR2 KO cells were overrepresented in cellular component categories that clustered around synaptic proteins, nuclear proteins, and mitochondrial proteins. Overrepresentation of proteins with increased abundance in RyR2 KO cells in biological process (gene expression, cellular nitrogen compound biosynthetic processes, RNA metabolic processes, chromatin organization) and molecular function (RNA binding, mRNA binding, nucleotide catalytic activity, RNA catalytic activity) categories suggest that RyR2 activity may regulate transcription and translation and/or mRNA processing. Proteins more abundant in IRBIT KO cells were overrepresented in three cellular component categories related to mitochondria (mitochondrial matrix, mitochondrial protein-containing complexes, intracellular organelle lumen). Proteins with   www.nature.com/scientificreports/ decreased abundance in RyR2 KO cells are mainly overrepresented in categories related to organelle structure and function. Small GTPase binding proteins were overrepresented among proteins with both increased and increased abundance in RyR2 KO cells. The smaller group of proteins with decreased abundance in IRBIT KO cells showed modest overrepresentation (< twofold) in two cellular component categories, intracellular anatomical structure and cytoplasm.

Discussion
Deletion of RyR2 resulted in no detectable band upon western blot with a pan RyR antibody and essentially abolished mobilization of ER Ca 2+ by caffeine (Fig. 1), suggesting that RyR2 is the predominant, if not sole, functional RyR in INS-1 cells. This result is consistent with studies showing that RyR2 is the prominent RyR transcript detected in INS-1 cells 5 and human islets 9 . RyR2 also contributes to resting cytoplasmic [Ca 2+ ] (Fig. 1F) and PLC activity (Fig. 3). Our results suggest a complicated interplay between RyR2 and IP 3 receptors. In control cells, xesto had no effect on the Ca 2+ integral in response to glucose, but did shorten the period between Ca 2+ oscillations. In contrast, deletion of RyR2 led to a significant increase in the Ca 2+ integral in response to glucose, and xesto both decreased the Ca 2+ AUC and increased the period between oscillations in RyR2 KO cells upon glucose stimulation. This suggests the decrease in IRBIT detected in RyR2 KO cells is functionally significant, leading to enhanced IP 3 R activation during glucose stimulation. RyR2 deletion markedly reduced both basal and stimulated IP 1 accumulation compared to control INS-1 cells, suggesting that RyR2 plays an important role in supporting PLC activity in INS-1 cells. Store-operated Ca 2+ entry (SOCE) plays a key role in supporting PLC activity in pancreatic β-cells 27 , and RyR2 is implicated in gating SOCE channels in rat vascular smooth muscle 28 . It will be of interest to determine if a similar mechanism can account for the decreased PLC activity observed in RyR2 KO cells.
In addition to IP 3 R, IRBIT also regulates many other proteins 29 . Thus, phenotypes of RyR2 KO cells may be attributed to either loss of RyR2 Ca 2+ release directly, or to downregulation of IRBIT. Caffeine-stimulated Ca 2+ transients were maintained in IRBIT KO cells, indicating that RyR2 function is retained in the absence of IRBIT. However, the response was significantly reduced compared to control INS-1 cells. This is likely the result of a reduced pool of Ca 2+ available for release via RyR2, as the ER Ca 2+ levels in IRBIT KO cells were reduced compared to control and RyR2 KO cells. As expected, deletion of IRBIT led to an increase in the Ca 2+ response to glucose compared to controls cells, the result of enhanced activation of IP 3 receptors during glucose stimulation, as in RyR2 KO cells. It's not clear why IRBIT levels/activity are reduced upon RyR2 deletion. We speculate that RyR2 may play a critical role in the phosphorylation of IRBIT on Ser 68 by a Ca 2+ -dependent kinase, leading to further phosphorylation that's required for IP 3 receptor binding 18 . Further, unphosphorylated IRBIT is susceptible to proteolytic cleavage 30 . Thus, hypo-phosphorylation of IRBIT could account for both the elevated activity of IP 3 receptors, and the reduced IRBIT levels observed in RyR2 KO cells.
Deletion of either RyR2 or IRBIT had marked effects on insulin secretion, cellular insulin content, and insulin gene transcript levels. RyR2 deletion had the greatest effect on basal (2.5 mM) and glucose stimulated insulin secretion, but IRBIT deletion also reduced basal and glucose-stimulated insulin secretion compared to control INS-1 cells. Although elevated Ca 2+ release from IP 3 receptors during glucose stimulation was a hallmark of both RyR2 and IRBIT deletion, xesto didn't inhibit insulin secretion in either control, RyR2 KO , or IRBIT KO cells. Our results show that even when IP 3 receptors are highly active, they don't contribute to glucose-stimulated insulin secretion in INS-1 cells. This is consistent with previous reports that IP 3 receptors are upregulated in the β-cells of type 2 diabetics, but associated with impairment of glucose-stimulated insulin secretion and β-cell dysfunction 31 .
To understand what might account for this reduction in insulin secretion, we examined the insulin content of control INS-1, RyR2 KO , and IRBIT KO cells and found an ~ 70% decrease in insulin content of RyR2 KO cells, and an ~ 40% decrease in insulin content in IRBIT KO cells compared to controls. This decrease in insulin content was accompanied by a similar decrease in INS2 gene transcript levels in both RyR2 KO and IRBIT KO cells. Interestingly, the level of INS1 transcript detected in RyR2 KO cells was reduced compared to controls, but wasn't different in IRBIT KO cells. Although the role of IRBIT as a regulator of IP 3 receptor activation is well studied, it's clear that IRBIT plays many other roles 32 . IRBIT contains a highly conserved, but catalytically inactive, S-adenosyl homocysteine hydrolase (AHCY) domain 33 . By virtue of this domain, IRBIT can bind to and may regulate the activity of catalytically active AHCY 24 as well as its distribution between the cytoplasm and the nucleus 25 . Since AHCY degrades S-adenosyl homocysteine (SAH), a potent inhibitor of DNA methyltransferases 34 , dysregulation of AHCY by loss of IRBIT could play a role in the increased INS1 and INS2 gene methylation that we observed in RyR2 KO and IRBIT KO cells. While deletion of RyR2 slightly increased nuclear AHCY localization, deletion of IRBIT caused a marked shift of AHCY from the cytoplasm into the nucleus, though the total amount of AHCY immunostaining detected in the nuclei of IRBIT KO cells was reduced compared to controls. Increased AHCY activity in the nucleus could disinhibit DNA, RNA, and protein methyltransferases (Fig. 6d). AHCY binds to chromatin near transcription start sites of active genes 35 , suggesting that AHCY regulation of DNA methylation may be spatially specific. Given that AHCY is only active as a homotetramer 36 , incorporation of IRBIT via its AHCY domain into AHCY complexes could diminish AHCY activity. We speculate that, in addition to controlling AHCY nuclear accumulation, IRBIT may also limit AHCY activity. Such a scenario would imply that AHCY in the absence of IRBIT not only accumulates preferentially in the nucleus, but may be catalytically more active.
The shift of AHCY from the cytoplasm to the nucleus in IRBIT KO cells and the corresponding changes in insulin mRNA levels were coincident with an increase in methylation of both the INS1 and INS2 genes downstream of the translation start site in intron 2. The promoter/upstream regions were hypomethylated in both INS1 and INS2, which didn't change upon deletion of either RyR2 or IRBIT, consistent with the finding that hypomethylation of the promoter regions of insulin genes serves as a marker of islet cell identity 37 . Methylation of genes within the first exon (or within ~ 200 bp downstream of the transcription start site) is highly correlated with inhibition of transcription 38  The differential effect of RyR2 or IRBIT deletion on INS1 mRNA levels was also coincident with differential changes in methylation of the INS1 gene in exon 2. While CpG site(s) in exon 2 (1-DS3) are hypermethylated in both RyR2 KO and IRBIT KO cells, a specific CpG site in exon 2 (ATG + 136) in the INS1 gene is hypermethylated in IRBIT KO cells compared to control and RyR2 KO cells. It is currently unclear whether increased methylation within the 1-DS3 region accounts for the decrease in INS1 mRNA in RyR2 KO cells, or if the specific increase in methylation at + 136 is responsible for the maintenance of INS1 mRNA levels in IRBIT KO cells. However, this differential regulation of INS1 mRNA levels may account for the smaller decrease in insulin content in IRBIT KO cells, compared to RyR2 KO cells. It will be of interest to determine if RyR2 or IRBIT deletion results in a more global increase in DNA methylation.
Exploratory proteomics analysis suggests that deletion of RyR2 or IRBIT differentially regulates a partially overlapping set of proteins. It's not clear if this regulation is occurring pre-or post-translationally, but the reduction of IRBIT protein in RyR2 KO cells, with no decrease in mRNA levels, demonstrates that RyR2 activity is capable of regulating protein levels post-transcriptionally. GO analysis revealed that RNA binding/processing proteins are overrepresented in the population of proteins increased in abundance by RyR2 deletion, suggesting altered RNA processing in the absence of RyR2. Mitochondrial proteins are also overrepresented in the population of proteins increased in abundance by RyR2 or IRBIT deletion, perhaps reflecting the proposed role of IRBIT in regulating Ca 2+ flux between the ER and mitochondria 41 . The decrease in ATG5 protein in both RyR2 KO and IRBIT KO cells might also explain the increased levels of some mitochondrial proteins, since deletion of ATG5 (autophagy-related gene 5) in T-lymphocytes differentially regulates mitochondrial protein levels and mitochondrial mass 42 . GTPase binding proteins are overrepresented in populations of both increased and decreased abundance proteins in RyR2 KO cells, suggesting a switch in the complement of modulators of small GTPase proteins, which play critical roles in vesicle trafficking 43 . Finally, proteins more abundant upon RyR2 deletion are overrepresented in several categories related to the nucleus, including four components of transcription repressor complexes-Rcor1 (REST co-repressor 1), Ncor1 (nuclear receptor corepressor 1), Ctbp2 (c-terminal binding protein 2), and Coro2a (coronin 2a). Rcor1, also increased in abundance in IRBIT KO cells, and Ctbp2 are of particular interest since they are part of the RE-1 Silencing Transcription factor (REST) repressor complex 44 , which represses genes critical for β-cell function, but is inactivated during differentiation. Several proteins repressed by REST (Pcsk1, neuroendocrine convertase 1; Chga, chromogranin A; Chgb, secretogranin; Stmn2, stathmin 2) 44 are reduced in abundance in RyR2 KO cells. Thus loss of RyR2 function may permit expression of some components of the REST repressor complex, and repression of a subset of genes critical for β-cell function.
Some of the differentially regulated proteins identified in RyR2 KO and IRBIT KO cells are dysregulated in diabetes and/or play a key role in β-cell function. Among these are PCSK1 45 and ORMDL1/2 (sphingolipid biosynthesis regulator 1/2) 46 which are reduced in both RyR2 KO and IRBIT KO cells, Abat (GABA aminotransferase) 47 , and Kcnj11 (Kir6.2) 48 which are reduced in RyR2 KO cells. The Ca 2+ -dependent adhesion molecule Cdh2 (N-cadhedrin) is also reduced in both RyR2 KO and IRBIT KO cells. Cell adhesion and spreading via N-cadhedrin enhances GSIS 49 . The pancreas-specific deletion of HUWE1 (HECT, UBA, and WWE domain containing E3 ubiquitin ligase 1), which is reduced in both RyR2 KO and IRBIT KO cell, leads to increased β-cell apoptosis and reduced β-cell mass 50 . Anks4b which is, together with its binding partner Ush1c (Harmonin), more abundant in both RyR2 KO and IRBIT KO cells, increases susceptibility to ER stress-induced apoptosis when overexpressed in MIN6 cells 51 . Finally, some key proteins in β-cell function were specifically reduced in IRBIT KO cells, including mTOR (mammalian target of rapamycin) 52 , Creb1 (cAMP response element-binding protein 1) 53 , and IRS2 (insulin receptor substrate 2) 54 .
In summary, deletion of RyR2 or IRBIT both enhanced IP 3 receptor activation during glucose stimulation, and reduced insulin secretion, content, and INS2 mRNA. In addition, the INS1 and INS2 genes were hypermethylated in exon 2 upon RyR2 or IRBIT deletion, coincident with alterations in the nuclear localization of AHCY. One limitation of this study is that, while insulin content is clearly reduced by RyR2 or IRBIT deletion, it's not clear if this reduction accounts for the reduced secretion, or if granule trafficking/exocytosis is also impaired by RyR2 or IRBIT deletion. Furthermore, the relationship between RyR2 activity and IRBIT levels remains unknown. Nevertheless, IRBIT regulation of AHCY localization and, potentially, activity positions it to regulate the activity of protein, RNA, and DNA methyltransferases via modulation of local SAH levels, and regulate the proteome. It will be of interest to determine if pathophysiological perturbation of RyR2 activity or dysregulation of ER Ca 2+ levels in pancreatic β-cells leads to dysregulation of IRBIT activity.
Construction of Cas9 plasmids. gRNA sequences were designed using the crispr.mit.org website. Oligo gRNA oligonucleotides were subcloned into pSpCas9(BB) using BbsI (New England Biolabs). Ligation products were used to transform competent DH5α E. coli, and transformants were selected on Luria broth-agar plates containing 100 µg/mL ampicillin. Plasmid DNA was purified and sequenced (Purdue Genomics Core Facility) to confirm assembly of the desired construct.
Generation and validation of knockout clones. INS-1 cells were transfected with 2 μg RyR2-or IRBITtargeted gRNA/Cas9 plasmid or pEGFP-N1 (selection control) using Lipofectamine 2000 (Invitrogen) per manufacturer's instructions. 72 h post-transfection, cells were selected with 3 μg/mL puromycin until no cells remained in the pEGFP-N1 transfected well. Individual clones from the gRNA/Cas9 transfected cells were then isolated by limiting dilution in 96-well plates (Corning). RPMI-1640 media was changed weekly for 4 weeks, and clones were gradually expanded. Once expanded, clones were plated at 90% confluency in 96-well plates and allowed to incubate overnight at 37 °C, 5% CO 2 . Cells were lysed and genomic DNA was extracted using QuickExtract DNA Extraction Buffer (Lucigen) per the manufacturer's instructions. Extracted genomic DNA from INS-1, RyR2 KO , or IRBIT KO cells was subjected to PCR amplification (Herculase II Fusion DNA Polymerase and 5X Herculase II PCR Buffer and dNTP (Agilent Techologies) using primers flanking the region targeted by the gRNA. Purified amplicons were sequenced at the Purdue University Genomics Core.
Single-cell intracellular Ca 2+ assays. INS-1 and RyR2 KO cells were either plated in a 35 mm tissue culture dish (Corning) containing a poly-D-lysine coated round glass coverslip (for assays using perfusion; Warner Instrument) or plated in a poly-D-lysine coated 4-chambered 35 mm glass bottom tissue culture dish (for assays not using perfusion; Cellvis). Cells were incubated overnight in RPMI-1640 media at 37 °C, 5% CO 2 . For glucose stimulation assays, cells were deprived of glucose for an additional 24 h in low glucose RPMI-1640 media. Cells were washed twice with PBS prior to loading with 3 μM of the Ca 2+ indicator Fura-2 AM (Invitrogen) diluted in a modified Krebs-Ringer buffer [KRBH: 134 mM NaCl, 3.5 mM KCl, 1.2 mM KH 2 PO 4 , 0.5 mM MgSO 4 , 1.5 mM CaCl 2 , 5 mM NaHCO 3 , 10 mM HEPES (pH 7.4)] supplemented with 0.05% fatty acid free BSA at room temperature for 1 h. The KRBH containing Fura-2 AM was then removed, and the cells were washed twice with KRBH, then equilibrated for 30 min at room temperature in KRBH alone or KRBH containing a 2 × concentration of indicated inhibitors. For perfusion assays, the glass coverslip was mounted on a perfusion chamber attached to the stage of an Olympus IX50 inverted microscope equipped with a PlanApo 40 × objective lens (0.95 na) and solutions/stimuli were perfused to the chamber at a constant flow rate (1 mL/min) at room temperature. For assays not using perfusion, the 4-chambered 35 mm dish was mounted on a chamber attached to the stage of the microscope. Cells were stimulated with the indicated stimulus at a 2 × concentration. Cells were alternatively excited at 340/11 nm and 380/20 nm wavelengths using a band pass filter shutter (Sutter Instrument) and changes in intracellular Ca 2+ were measured by recording the ratio of fluorescence intensities at 508/20 nm in time lapse (time interval of 0.6 s) using a Clara CCD camera (Andor Technology). Background subtraction from the raw 340/11 nm and 380/20 nm wavelengths was performed, then isolated single cells were selected as regions of interest (ROI) and the 340/11 nm/380/20 nm ratios for each ROI were measured using MetaMorph image analysis software (Molecular Devices). All single-cell Ca 2+ transients were normalized to their baseline intracellular Ca 2+ level, which was obtained by averaging the 340/11 nm/380/20 nm ratios during the first minute of each experiment when no stimulus was present. Ca 2+ transients are plotted as normalized 340/11 nm/380/20 nm ratios against time. Immunocytochemistry. Cells were plated at 50% confluency in poly-D-lysine coated 4-chamber glass bottom dishes (Cellvis) 16-24 h prior to fixation. The following day cells were washed once with PBS, then fixed in 4% paraformaldehyde in PBS for 10 min at room temperature (RT). Cells were then washed three times with PBS and permeabilized in 0.2% Triton X-100 in PBS for 10 min at RT. Cells were blocked in 3% BSA in PBS for 1 h at RT then incubated in primary antibody (mouse anti-AHCY 1:200 or mouse anti-PIP 2 1:50) overnight at 4 °C. Following overnight incubation cells, were washed then incubated with anti-mouse IgG-κ Fc binding protein CFL 488 diluted 1:2000 in 3% BSA for 1 h at RT. After 3 washes, cells were incubated with 5 µg/mL Hoechst 33,342 in PBS for 10 min at RT. Hoechst 33,342 solution was then removed after 10 min and cells were imaged in PBS by confocal microscopy on a Nikon A1Rsi confocal microscope. Nuclear regions of interest (ROIs) were identified in an automated fashion by Hoechst 33,342 staining, and cytosolic regions were identified by manual placement of ROIs within the cell outside of the nucleus. All ROI analysis was performed using NIS Elements (Nikon). pERK1/2 assay. Cells were plated at 70% confluency in 24-well plates and incubated overnight at 37 °C.
16-24 h prior to assay, cells were incubated in serum-free low glucose RPMI-1640 media supplemented with 0.1% fatty acid-free BSA (FAF-BSA) overnight at 37 °C, 5% CO 2 . Cells were washed once with PBS and preincubated with 200 µL KRBH containing 2.5 mM glucose for 30 min at 37 °C, 5% CO2. After 30 min, KRBH was removed and replaced with either 200 µL KRBH containing the indicated concentrations of stimulants and were stimulated for 10 min at 37 °C, 5% CO2. Supernatants were discarded and cells were lysed 75 µL TBS containing 1% Triton-X100 supplemented with protease and phosphatase inhibitors (20 mM sodium fluoride, 2 mM sodium orthovanadate, 10 mM β-glycerolphosphate, and 10 mM sodium pyrophosphate). 50 μg of each lysate was separated by SDS-PAGE on 10% acrylamide gels at 150 V for 90 min. Proteins were transferred onto PVDF membranes in ice-cold Towbin Buffer, 10% ethanol at 100 V for 1 h. Membranes were blocked in 3% BSA in TBST for 1 h at room temperature and then incubated with anti-ERK (1:2000) and anti-pERK (1:1000) in 3% BSA in TBST overnight at 4 °C. Primary antibody was removed the following day, and membranes were incubated in goat anti-mouse IgG HRP (1:10,000) and goat anti-rabbit IR800 (1:10,000) in 3% BSA in TBST for 1 h at room temperature. Chemiluminescence was detected using standard ECL reagents, and chemiluminescence and fluorescence were imaged on Azure Biosystems Sapphire imager.

LC-MS/MS analysis of proteins.
Cells were lysed using a Barocycler (5 °C, 60 cycles: 50 s at 35,000 psi and 10 s at 1 atmospheric pressure) in 100 mM ammonium bicarbonate. Protein concentration was measured by bicinchoninic acid (BCA) assay (Pierce) and 50 ug of total protein for each sample was precipitated with 4 volumes of cold acetone (-20 °C), and used for sample preparation as described previously 55,56 . Dried and C18cleaned peptides were re-suspended in 96.9% purified water, 3% acetonitrile, and 0.1% formic acid at a 1 µg/µL, and 1µL was used for LC-MS/MS analysis in the Orbitrap Fusion Lumos mass spectrometer (Thermo Fisher Scientific) 55,57 . The LC-MS/MS raw data were processed using MaxQuant (v1.6.3.3) 58 for protein identification and label-free quantitation 59 . MaxQuant results files were merged by matching rows based on the gene names. Proteins marked as "contaminants", "reverse" and "only identified by match between runs" were removed. All LFQ values were then Log 2 transformed for normalization, and samples were grouped based on deletions (control, IRBIT KO or RYR2 KO ). Proteins identified in two replicates in at least one group (control or IRBIT KO ; control or RYR2 KO ) were filtered for subsequent processing. Missing values were imputed using a constant (Zero-fill), and the average Log 2 (LFQ) values for each group were then calculated. Downregulated proteins were considered as proteins with Log 2 (Fold-Change) < -1 and average MS/MS count ratio < 0.5, compared to control. Similarly, upregulated proteins were considered as proteins with Log 2 (Fold-Change) > 1 and average MS/MS count ratio > 2, compared to control.
Gene ontology (GO) analysis. Proteins identified as up-or down-regulated in RyR2 KO or IRBIT KO cells were analyzed for over-representation in specific cellular component, biological process, or molecular function categories using PANTHER version 16 (http:// www. panth erdb. org). 60 Results for RyR KO cells and IRBIT KO upregulated proteins were filtered at FDR < 0.05, > twofold enrichment, and a minimum of 4 proteins per category, to exclude categories with 3 or less differentially regulated proteins. IRBIT KO down-regulated proteins which were filtered at FDR < 0.05 only, since no category had greater than twofold enrichment.

Data availability
All data needed to evaluate the conclusions in the paper are present in the paper or the supplementary materials.