Genetic ablation of synaptotagmin‐9 alters tomosyn‐1 function to increase insulin secretion from pancreatic β‐cells improving glucose clearance

Stimulus‐coupled insulin secretion from the pancreatic islet β‐cells involves the fusion of insulin granules to the plasma membrane (PM) via SNARE complex formation—a cellular process key for maintaining whole‐body glucose homeostasis. Less is known about the role of endogenous inhibitors of SNARE complexes in insulin secretion. We show that an insulin granule protein synaptotagmin‐9 (Syt9) deletion in mice increased glucose clearance and plasma insulin levels without affecting insulin action compared to the control mice. Upon glucose stimulation, increased biphasic and static insulin secretion were observed from ex vivo islets due to Syt9 loss. Syt9 colocalizes and binds with tomosyn‐1 and the PM syntaxin‐1A (Stx1A); Stx1A is required for forming SNARE complexes. Syt9 knockdown reduced tomosyn‐1 protein abundance via proteasomal degradation and binding of tomosyn‐1 to Stx1A. Furthermore, Stx1A‐SNARE complex formation was increased, implicating Syt9‐tomosyn‐1‐Stx1A complex is inhibitory in insulin secretion. Rescuing tomosyn‐1 blocked the Syt9‐knockdown‐mediated increases in insulin secretion. This shows that the inhibitory effects of Syt9 on insulin secretion are mediated by tomosyn‐1. We report a molecular mechanism by which β‐cells modulate their secretory capacity rendering insulin granules nonfusogenic by forming the Syt9‐tomosyn‐1‐Stx1A complex. Altogether, Syt9 loss in β‐cells decreases tomosyn‐1 protein abundance, increasing the formation of Stx1A‐SNARE complexes, insulin secretion, and glucose clearance. These outcomes differ from the previously published work that identified Syt9 has either a positive or no effect of Syt9 on insulin secretion. Future work using β‐cell‐specific deletion of Syt9 mice is key for establishing the role of Syt9 in insulin secretion.


| INTRODUCTION
Glucose elicits a biphasic insulin secretion response from pancreatic βcells to maintain whole-body glucose homeostasis. 1,2 Increases in the cellular ATP to ADP ratio, which results from glucose metabolism, cause the closure of the ATP-sensitive potassium channels (K ATP ), leading to the depolarization of plasma membrane (PM). [3][4][5] Consequently, increases in the Ca 2+ influx through voltage-gated Ca 2+ channels 6 facilitate the fusion of insulin granules to the PM, causing an early phase insulin secretion. [6][7][8] Additionally, glucose-generated amplifying pathways contribute to a more sustained release of insulin. 9 The molecular mechanisms underlying biphasic insulin secretion are not completely understood, especially those that contribute to βcell dysfunction in insulinresistant obese and type 2 diabetic individuals. [10][11][12][13][14][15][16][17][18][19][20][21][22] The soluble N-ethylmaleimide-sensitive factor attachment receptor protein (SNARE) complex formation is required for insulin granules to undergo fusion with the PM in both phases of glucose-stimulated insulin secretion (GSIS). [23][24][25] Biochemical and structural studies have demonstrated that the SNARE core complex assembly occurs between a vesicle (v)-SNARE protein Vamp2 (vesicle-associated protein), which is present on the insulin granules with the PM-bound target (t)-SNARE proteins syntaxin (Stx) and Snap25 (synaptosome-associated protein of 25 kDa). 26 Aided by the accessory proteins, the SNARE core proteins assemble into a complex that is only partially zipped up and clamped. 27, 28 Upon stimulation, increases in [Ca 2+ ] i cause complete zippering of the SNAREs by engaging enhancer proteins, causing membrane fusion and insulin exocytosis. 29 It is accepted that only a fraction of insulin granules undergo fusion to the PM upon stimulation. Thus, understanding how βcells regulate the availability of Stx to form SNARE complexes is key for the fusion of insulin granules in insulin secretion.
Stx1A is a well-characterized t-SNARE that forms a cognate SNARE core complex with Vamp2 and Snap25 to regulate the fusion of insulin granules in early and sustained phases of insulin secretion. [30][31][32][33][34] Clusters of Stx1A are present in the PM with proximity to insulin granules. 35 Moreover, mice with βcell-specific Stx1A deletion exhibit impaired early and sustained phase insulin secretion. 30,36,37 In the islets of individuals with T2D, reduction in biphasic insulin secretion is correlated with decreases in the abundance of islet Stx1A, 38 implicating a clinically relevant role of Stx1A in βcell function. Therefore, factors modulating the essential function of Stx1A have a major effect in regulating biphasic insulin secretion. Proteins such as Munc13-1 39 and Munc18a 31 facilitate the availability of Stx1A forming SNAREs. In contrast, a relatively less characterized syntaxin binding protein 5 (Stxbp5) also known as tomosyn-1, functions to inhibit Stx1A, decreasing Stx1A-SNAREs (Stx1A-Snap25-Vamp2) formation and causing the inhibition of insulin secretion. 40 Interestingly, the role of factors facilitating the formation of the Stx1A-SNARE complex is well characterized. However, less is known about factors that decrease the ability of Stx1A to facilitate SNARE-mediated fusion of insulin granules in insulin secretion.
Synaptotagmin (Syt) isoforms have diverse roles in different cell types. 41 The primary structure of Syt comprises an N-terminal intra-vesicular transmembrane region, a linker domain, and two Ca 2+ -binding cytoplasmic domains (C2A and C2B) at the C-terminal. 42,43 In the presence of Ca 2+ , Syt isoforms such as Syt1 and Syt7, via C2A and C2B, bind SNAREs and PM anionic phospholipids, accelerating SNARE complexes-mediated fusion of granules with the PM in exocytosis. [44][45][46][47] Conversely, Syt1 also functions to inhibit SNARE fusion complexes. 48 The ablation of Syt1 was found to increase the spontaneous release of granules from neurons in mice and drosophila. [49][50][51][52][53] These studies show that Ca 2+ -binding Syt1 can positively and negatively affect vesicle exocytosis. In βcells, Syt7 is a major Ca 2+ sensor that facilitates insulin exocytosis. [54][55][56] Mice with the Syt7 deletion exhibited impaired insulin secretion and glucose intolerance. 57 However, the characterization of other Syt isoforms in regulating the fusion of insulin granules is not completely elucidated. Syt9 (NCBI accession #NP_001347350, also referred to as Syt5 58 ) is abundantly expressed in βcells, 59 and herein, we used the Syt9 gene deletion mouse model and clonal INS1(832/13) βcells combined with metabolic phenotyping, confocal imaging, and biochemical approaches to determine the molecular mechanism by which Syt9 regulates insulin secretion and whole-body glucose homeostasis. effect of Syt9 on insulin secretion. Future work using βcell-specific deletion of Syt9 mice is key for establishing the role of Syt9 in insulin secretion.

| Expression construct
The pcDNA3-m-tomosyn-1 construct was generously provided by Dr. Alexander Groffen, Virije Universiteit, Netherlands. Moloney murine leukemia virus-based retroviral vector (RVV, 3051) was a gift from Dr. Bill Sugden, University of Wisconsin, Madison. The m-tomosyn-1 cDNA was subcloned to generate an m-tomosyn-1-V5 tagged RVV mammalian expression plasmid. The hGH mammalian expression plasmid was a gift from Dr. Edwin Chapman, University of Wisconsin-Madison.

| Animals
The Syt9 −/− mice were generously given to us by Dr. Edwin Chapman, University of Wisconsin-Madison. These mice were generated as described in previous studies. [60][61][62] All pups were weaned between 3 and 4 weeks of age. Male and female mice had free access to water and a chow diet and were housed in a temperature-controlled room with a 12 h light-dark cycle (6 AM-6 PM). All mice were kept following the University of Alabama at Birmingham Animal Research Program and NIH guidelines for the care and use of laboratory animals.
Additionally, an oral glucose tolerance test and plasma insulin levels were assessed in response to oral glucose gavage (2 g glucose/kg BW of mice) after a 12 h fast. Tail vein blood was collected at different times to determine blood glucose and plasma insulin levels. An insulin tolerance test was performed in response to an insulin dose (0.5 U human insulin/kg BW of mice, Humulin R, Lilly, USA, Cat #0028215-01) administered intraperitoneally to mice after a 6 h fast. Blood glucose levels were determined using the Contour Next blood glucose meter (Ascensia Diabetes Care, Switzerland). The plasma insulin concentration was measured by Ultra-Sensitive Mouse Insulin ELISA Kit (Crystal Chem, USA, Cat #90080).

| Islet isolation and cell culture
Mouse islets were isolated from 8-to 10-week-old Syt9 −/− and Syt9 +/+ mice using a collagenase digestion method and the Ficoll gradient procedure described previously. 63,64 Briefly, the pancreas was inflated by injecting collagenase (0.6 mg/ mL) solution through the common bile duct, followed by removal, digestion steps, and Ficoll gradient to obtain isolated islets. Handpicked islets were cultured overnight in supplemented RPMI 1640 culture medium containing 8 mM glucose for perifusion and static insulin secretion experiments. INS1(832/13) βcells, a gift from Dr. Christopher Newgard, Duke University, NC, were cultured in RPMI 1640 media supplemented with 10% heat-inactivated fetal bovine serum, 1 mM sodium pyruvate, 20 mM HEPES, 2 mM l-glutamine, and 100 units/mL of antimycotic-antibiotic along with 50 μM βmercaptoethanol.

| Hormone secretion
Insulin secretion from mouse islets was performed as described previously. 63 Islets were preincubated in 100 μL/well of KRB buffer containing 2.8 mM glucose. After 45 min, the preincubation buffer was replaced with the KRB incubation buffer containing insulin secretagogues. After 45 min, the incubation buffer was collected to estimate secreted insulin. Furthermore, islets were lysed in acid ethanol to determine cellular insulin content.
Insulin or human growth hormone (hGH) secretion from INS1(832/13) βcells was performed as described previously. 66 Cells were seeded at a density of 100 000 cells/well in a 96-well plate at ~80% confluency. Fortytwo-hour post-transfection, cells were washed and preincubated with 100 μL of the KRB-based buffer (118.41 mM NaCl, 4.69 mM KCl, 1.18 mM MgSO 4 , 1.18 mM KH 2 PO 4 , 25 mM NaHCO 3 , 20 mM HEPES, 2.52 mM CaCl 2 , pH 7.4, and 0.2% BSA) containing 1.5 mM glucose. After 2 h, the preincubation buffer was replaced with the KRB incubation buffer containing insulin secretagogues. After 2 h, an incubation buffer was collected to determine secreted insulin or hGH. Furthermore, cells were lysed in acid ethanol and lysis buffer (100 mM Tris-HCl, 300 mM NaCl, 10 mM NaF, 2 mM Na 3 VO 4 , 2% NP-40, and protease inhibitor cocktail) to determine cellular insulin and hGH content, respectively. Quantification of insulin was performed using in-house ELISA. The percent fractional insulin secretion was calculated as the amount of insulin secreted divided by the total insulin content. HGH was quantified using ELISA (Roche, Cat #11585878001) following instructions provided in the kit.

| Islet perifusion
For the perifusion insulin secretion assay (as described 63 ), a high-capacity perifusion system from Bio Rep® Perifusion was utilized. Approximately 75 islets were sandwiched in a chamber between two layers of Bio-Gel P-2 (Bio-Rad, Cat #1504118) bead solution (200 mg beads/ml in KRB buffer). Throughout the experiment, chambers containing islets and buffers were maintained at 37°C. Islets were perfused at a flow rate of 1 mL/min, and the flow-through containing secreted insulin was collected in a 96-well plate using an automatic fraction collector. The in-house ELISA was used to quantify cellular and secreted insulin.
2.9.2 | Immunoprecipitation of endogenous proteins INS1(832/13) cells combined from two 10 cm dishes or human islets (RRID: SAMN12500521, 2000 IEQ) were lysed in 700 μL lysis buffer (25 mM Tris-HCl, pH 7.5, 50 mM NaCl, 2 mM MgCl 2 , 1 mM Na 2 EDTA, 1 mM EGTA, 1 mM CaCl 2 , 0.5% NP-40, 0.1% SDS, 2.5 mM sodium pyrophosphate, 1 mM Na 3 VO 4 , 1 mM PMSF, and protease inhibitor cocktail tablet) by placing on a rotating shaker for 1 h at 4°C followed by centrifugation at 13,000 rpm for 10 min. The precleared lysates were supplemented with 0.2% BSA and 1 mM CaCl 2 and incubated overnight with Stx1A antibody (2 mg/mL) or IgG isotype control. The next day, 50 mL of equilibrated magnetic beads were added to the reaction mixture and incubated at 4°C for 1 h. The immunoprecipitated complex was washed three times with 1 mL lysis buffer supplemented with 0.2% BSA and 1 mM CaCl 2 and eluted in 80 mL 2.5× Laemmli buffer with 1 mM DTT, followed by heating at 95°C for 5 min. Proteins were separated by 10% or 12% SDS-PAGE gels, transferred to a PVDF membrane, followed by immunoblotting and detection with Clarity Western ECL substrate (Bio-Rad, Cat #1705061).

| Quantification of Ca 2+ in IP buffer
Effective calcium concentration in IP buffer was measured using a colorimetric calcium assay kit (Abcam, Cat #ab102505). Briefly, 50 μL of samples (IP buffer supplemented with/without 2 mM CaCl 2 ), controls, and calcium standards were pipetted in a 96-well plate. Then, 90 μL of the chromogenic reagent was added to each wellcontaining standard, sample, or control. Next, 60 μL of calcium assay buffer was added into each well and mixed properly. Absorbance was taken at OD 575 nM immediately after incubation at room temperature for 10 min. The concentration of calcium was calculated as calcium concentration = (Sa/Sv) × D, Where: Sa = Sample Amount (in μg) from the standard curve; Sv = Sample Volume (μL) added into the wells; D = Sample Dilution factor.

| Quantitative real-time PCR
Total RNA was harvested from mouse islets and INS1(832/13) cells by using QIAGEN RNeasy Plus Kit (Cat #74034) and TRIzol™ reagent (Invitrogen, Cat #15596018), respectively. Following RNA extraction, cDNA was synthesized using a high-capacity cDNA reverse transcription kit (Applied Biosystems, Cat# 4368814). The mRNA abundance (Table S2) was determined by quantitative PCR using Fast Start SYBR Green (Roche, Cat #4673484001). The relative mRNA abundance was estimated by the comparative ΔCT method. β-Actin mRNA was used as a housekeeping control.

| Statistical analysis
Data are represented as means ± SEM. Statistical significance was performed using Student's two-tailed unpaired t-test for independent data. The significance limit was set at p < .05.

| Loss of Syt9 improves glucose clearance associated with hyperinsulinemia
The deletion of Syt9 in Syt9 −/− vs. Syt9 +/+ littermate control mice was confirmed by Western blotting. The protein abundance of Syt9 was undetectable in islets and the brain of Syt9 −/− mice ( Figure 1A). Metabolic phenotyping was performed using male Syt9 −/− and Syt9 +/+ mice. No effect on the body weight was observed regardless of the genotypes ( Figure 1B). We then determined random-fed and fasting plasma insulin and glucose levels in Syt9 −/− and Syt9 +/+ control mice on the standard chow diet. No significant difference in the random-fed (measured at 8 AM) ( Figure 1C) and fasting (after 6 h) ( Figure S1A) blood glucose levels were observed at 6 and 10 weeks in Syt9 −/− mice compared to the Syt9 +/+ mice. Interestingly, the random-fed plasma insulin levels were increased twofold (p < .001) ( Figure 1D) at 6 and 10 weeks in Syt9 −/− mice. A small yet significant (1.3-fold, p < .01) increase in fasting plasma insulin levels was observed at 6 weeks in Syt9 −/− mice, but no change was observed at 10 weeks. ( Figure S1B). These data show that the primary effect of Syt9 loss is in increasing plasma insulin levels in the fed state. To evaluate the effect of Syt9 on insulin-stimulated glucose clearance, an oral glucose tolerance test was performed in Syt9 −/− and Syt9 +/+ mice. The loss of Syt9 led to accelerated glucose clearance ( Figure 1E), reflected by the reduction in the glucose area under the curve (AUC) by 20% (p < .0001) that was observed in Syt9 −/− , compared to the Syt9 +/+ mice ( Figure 1F). An insulin tolerance test was performed to evaluate the effect of Syt9 loss on insulin-stimulated glucose clearance in Syt9 −/− vs. Syt9 +/+ control mice ( Figure 1G). Neither the glucose excursions nor the AUCs for Syt9 −/− and Syt9 +/+ mice after insulin injection significantly differed ( Figure 1G,H). Moreover, the slope measured as Kg 30 assessing the percent glucose clearance per min for the first 30 min after insulin administration was not significantly different compared to Syt9 +/+ mice ( Figure 1I). Together, these data implicate the role of Syt9 in modulating pancreatic βcells function. We also performed metabolic phenotyping in Syt9 −/− and Syt9 +/+ female mice. No pronounced effects on glucose homeostasis were observed; fasting and fed plasma insulin and glucose levels, glucose clearance, insulin-stimulated glucose clearance, and glucose-stimulated plasma insulin levels ( Figure S1C-I) remained unaffected by Syt9-deletion, suggesting a sex-specific role of Syt9 in regulating pancreatic βcell function.

| Loss of Syt9 increases stimulus-coupled insulin secretion
Increases in plasma insulin levels were observed in random-fed Syt9 −/− mice. Thus, we investigated the effect of Syt9 loss on insulin secretion. The Syt9 −/− and Syt9 +/+ mice were subjected to an oral glucose challenge, and plasma insulin levels were determined at different time points (Figure 2A). Glucose increased plasma insulin levels at 5 and 15 min compared to the baseline (t = 0 min) in both groups; however, further increases by ~30% (p < .04) and ~65% (p < .03) were observed in Syt9 −/− mice compared to Syt9 +/+ mice ( Figure 2A). These data suggest that increases in insulin secretion contribute to improved glucose clearance in Syt9-deficient mice.
To gain insights into how Syt9 regulates insulin secretion, we performed dynamic insulin secretion measurements by perifusion 63 using islets isolated from Syt9 −/− and Syt9 +/+ mice ( Figure 2B-F). As expected, high glucose (16.7 mM) elicited increases in early (6-16 min) and sustained (17-50 min) phases of insulin secretion compared to low glucose (2.8 mM) in both groups of mouse islets ( Figure 2B). However, the Syt9 −/− mouse islets exhibited higher increases in early and sustained phase insulin secretion compared to Syt9 +/+ mouse islets ( Figure 2B). These are reflected by the AUCs, which revealed a twofold increase in the early ( Figure 2C, p < .04) and a 50% increase in the sustained phase ( Figure 2E, p < .04) insulin secretion from Syt9 −/− vs. Syt9 +/+ mouse islets in response to high glucose. KCl is known to cause PM depolarization and facilitates insulin release in the early phase. Consistently, KCl (40 mM at 1.5 mM glucose)-induced early-phase insulin secretion was observed from both Syt9 −/− vs. Syt9 +/+ mouse islets ( Figure 2B). However, islets from Syt9 −/− mice exhibited significantly increased insulin secretion compared to Syt9 +/+ mouse islets ( Figure 2B), as reflected by the twofold increase in the AUC for KCl-stimulated insulin secretion profile ( Figure 2E, p < .004). These outcomes show that Syt9 functions as an inhibitor of biphasic insulin secretion.
To confirm that Syt9 directly affects insulin secretion from βcells, a small inhibitory (si) RNA approach was used to achieve the knockdown of Syt9 in clonal INS1(832/13) βcells. Greater than 90% reduction in Syt9 mRNA ( Figure S2A) and protein levels ( Figure 5G) were achieved upon Syt9 knockdown in INS1(832/13) cells. As in islets, basal insulin secretion and βcell insulin content were unaffected in INS1(832/13) cells ( Figure S2B,C). However, increases in insulin secretion in response to the stimulation by KCl (40 mM at 1.5 mM glucose) and high glucose (15 mM) were observed upon Syt9knockdown compared to si-scramble INS1(832/13) cells ( Figure 2J). Similarly, the hGH reporter was transfected in INS1(832/13) cells expressing siSyt9 and si-Scr; hGH secretion was used to measure insulin secretion. High glucose and KCl treatment caused increases in GH secretion compared to 1.5 mM glucose in INS1(8321/3) cells transfected with siScr ( Figure S3A-C). Further increases in the GH secretion were observed upon Syt9 knockdown in response to 15 mM glucose and 40 mM KCl, without affecting insulin secretion at 1.5 mM glucose. In this experiment, hGH alone caused increases in glucose and KCl stimulated insulin secretion without affecting basal insulin secretion ( Figure S3D,E). The following outcomes: increases in plasma insulin levels, biphasic and static ex vivo islets insulin secretion, and insulin secretion from clonal β cells upon glucose stimulation in response to the loss/knockdown of Syt9 demonstrate that Syt9 inhibitory function is directly attributed to the inhibiting insulin secretion from β cells.

Stx1A, a key protein that forms the SNARE complex required for the fusion of insulin granules
Stx1A forms SNARE complexes (Stx1A-Snap25-Vamp2) in insulin exocytosis. 67 Thus, we used confocal imaging to evaluate whether endogenous Syt9 colocalizes with Stx1A. Immunostaining was performed in fixed INS1(832/13) cells. The colocalization of Syt9 (red) with   Stx1A (green) is shown as a merge (yellow) ( Figure 3A). Tomosyn-1 is an endogenous inhibitor of insulin exocytosis that inhibits Stx1A from forming Stx1A-SNARE complexes. 68,69 We also observed a colocalization between tomosyn-1 (red) and Stx1A (green) ( Figure 3B). The colocalization of Syt9 and tomosyn-1 with Stx1A suggests that these proteins potentially form a molecular complex.
To further establish this observation, IP was performed to assess the endogenous interaction between Syt9, Stx1A, and tomosyn-1 in INS1(832/13) cells upon KCl (40 mM at 1.5 mM glucose) stimulation. Total cell extracts were prepared to IP Stx1A-protein complex and were subjected to Western blotting for Stx1A, tomosyn-1, and Syt9 proteins. We observed that Syt9 formed a complex with Stx1A and tomosyn-1 in INS1(832/13) cells ( Figure 3D). Moreover, the KCl stimulation did not alter the interactions between these proteins compared to basal glucose (1.5 mM glucose), suggesting the formation of the Syt9-Stx1A-tomosyn-1 molecular complex endogenously. Furthermore, Syt9-Stx1Atomosyn-1 molecular complex formation was observed in human islets ( Figure 3E). These results show that Syt9, Stx1A, and tomosyn-1 form a novel molecular complex in INS1(832/13) βcells and human islets.
To gain insights into what caused the reduction of tomosyn-1 protein upon Syt9 knockdown, we assessed whether Syt9 regulates tomosyn-1 protein turnover by the proteasomal pathway. INS1(832/13) cells transfected with si-Syt9/si-scramble were treated with 5.0 μM of the proteasomal inhibitor MG132 or DMSO for 5 h. No significant change in tomosyn-1 protein abundance was observed between si-scramble transfected cells treated with MG132 vs. DMSO. However, the reduction in tomosyn-1 protein levels due to si-Syt9-mediated knockdown of Syt9 was rescued by MG132 treatment (Figure 5P), suggesting that the loss of Syt9 subjects tomosyn-1 protein to degradation by proteasomal pathway. These outcomes suggest that Syt9 in a Stx1A-tomosyn-1-Syt9 molecular complex potentially regulates tomosyn-1's protein stability.

Syt9 knockdown is blocked by tomosyn-1
We assessed whether the effect of Syt9 on insulin secretion is mediated by tomosyn-1. Insulin secretion was performed by overexpressing tomosyn-1 in si-Syt9 or si-scramble INS1(832/13) cells treated with low glucose (1.5 mM), high glucose (15 mM), and KCl (40 mM at 1.5 mM glucose). KCl ( Figure 6B) and high glucose ( Figure 6C) treatments increased insulin secretion compared to low glucose ( Figure 6A) (ninefold increase, p < .0001). Moreover, the knockdown of Syt9 caused further increases in insulin secretion in response to the  Figure 6B,C). Overexpression of tomosyn-1 inhibited insulin secretion in response to the stimulation by KCl (third vs. second/ first bar, Figure 6B) and high glucose (third vs. second/ first bar, Figure 6C) by ~40% (p < .001). Interestingly, no increase in the KCl-or high glucose-stimulated insulin secretion was observed upon Syt9 knockdown and the overexpression of tomosyn-1 (sixth vs. third bar in Figure 6B,C). The basal insulin secretion and cellular insulin content were not altered ( Figure 6A, Figure S6). These outcomes demonstrate that the increase in insulin secretion observed upon Syt9 loss/knockdown is mediated by tomosyn-1.

| DISCUSSION
Stx1A is a well-characterized t-SNARE required for SNARE complex-mediated fusion of insulin granules in the early and sustained phases of insulin secretion. [25][26][27] Less is known about how βcells clamp or inhibit Stx1A to regulate the formation of SNARE complexes in exocytosis. This study shows that increased plasma insulin levels (fed or in response to the oral glucose challenge) and glucose clearance without affecting insulin action (Figures 1 and 2) were observed in male Syt9 −/− mice but not in the Syt9 −/− female mice. The sex-dependent differences could be attributed to estrous cyclicity in female mice. Increases in glucose-stimulated insulin secretion (dynamic and static) were observed from ex vivo Syt9 −/− islets ( Figure 2B-H). Furthermore, the knockdown of Syt9 in clonal βcells increased the formation of the Stx1A-SNARE complexes and glucose-stimulated insulin secretion ( Figures 2J and 4A-C). Multiple approaches involving clonal INS1(832/13) βcells, ex vivo Syt9 −/− islets, and Syt9 −/− mice were used in this study to show that the loss/knockdown of Syt9 increased glucose-stimulated insulin secretion from βcells (Figures 1-4) enabled us to conclude that Syt9 has an inhibitory role in insulin secretion. The outcomes of this study identify and characterize the role of Syt9-tomosyn-1-Stx1A inhibitory complex that renders insulin granules nonfusogenic. Investigating how βcells overcome the inhibition/clamp to modulate SNARE-complexes-mediated fusion of insulin granules in insulin secretion could provide insights into several unresolved facets of insulin exocytosis, such as why only a fraction of cellular insulin granules undergo PM fusion  70 aged-insulin-containing granules are not preferred for fusion, 71-74 a mechanism for the loss of early-phase insulin secretion with the onset of impaired glucose tolerance when insulin granule content is not altered, [74][75][76][77][78] and an increase in fusion-incompetent docked granules and reduced fusion competency of granules in T2D. 25,26,[79][80][81][82][83][84][85] Insulin granules from different cellular pools contribute to biphasic GSIS. 86 Pre-docked granules are juxtaposed with the PM. They immediately undergo fusion upon stimulation and account for 50 percent of insulin released in the early phase. 87 In comparison, newcomer granules are present in the cytoplasm away from the PM 28 . Upon stimulation, they undergo fusion by two plausible modes, short-dock or no-dock, contributing to insulin released in both phases. Our results show that Syt9 deletion increased early and sustained phases of GSIS from islets ( Figure 2B,F,G), implicating that Syt9 potentially regulates the fusion of pre-docked and newcomer insulin granules.
The docking of the insulin granules to the PM is a temporal constraint on granules. 88 Moreover, mechanisms underlying long-and short-duration docking of pre-docked and newcomer granules still need to be understood entirely. Newcomer granules are trafficked from within the cells to the PM. Thus, it has been postulated that an inhibitor that functions at the insulin granule-PM interface could regulate the fusion of pre-docked and newcomer with short-dock modes of fusion of insulin granules in insulin exocytosis. Stx1A is required for biphasic insulin secretion and facilitates the fusion of pre-docked and newcomer with short-dock insulin granules. 30 Syt9 (an insulin granule protein) colocalizes and binds with the PM Stx1A (Figures 3 and 4). These outcomes, combined with Syt9's inhibitory function on insulin secretion (Figure 2), suggest that Syt9 potentially functions as an inhibitor (or potentially clamps) of Stx1A, decreasing its availability to form Stx1A-SNARE ( Figure 4) complexes-mediated fusion of pre-docked and newcomer (short-dock) insulin granules. How Syt9 modulates fusion modes of insulin granules involving distinct cognate SNARE complexes is part of the future directions.
Syt isoforms have a role in the clamping of SNARE complexes in exocytosis. It has been proposed that the clamping action is achieved by imposing a significant separation between plasma and granule membranes to form SNARE complexes 89 or locking the SNARE complexes in a fusion-incompetent state. [90][91][92] Insights into Syt functioning are predominately derived from studies elucidating the role of Syt1 in neurotransmitter release. However, little is known about Syt isoforms that regulate SNARE-mediated insulin exocytosis from βcells. Our data provide a mechanism by which Syt9 via tomosyn-1 inhibits Stx1A-SNARE complexes in insulin exocytosis.
Tomosyn-1 is a soluble cytoplasmic protein that inhibits SNARE complex formation. 93 It interacts with Stx1A via a short C-terminal domain, clamping Stx1A and Snap25 in a nonfusogenic complex, blocking the stimulatory effects of Munc13-1 and Munc18-1 in facilitating Stx1A-SNARE assembly. 94,95 Also, tomosyn-1 is known to interact with insulin granules via its N-terminal domains. 96-100 Present within its N-terminal region, the two unstructured loops are required for tomosyn-1 inhibitory function; it has been reported that deleting both N-terminal unstructured loops blocked tomosyn-1's ability to inhibit exocytosis while retaining Stx1A binding. 69 Concurrently, the N-terminal region of tomosyn-1 is essential for binding to Syt isoforms. The interaction between tomosyn-1 and granule protein Syt1 was reported in neurons. 98 Our data identify that an insulin granule protein Syt9 forms a molecular complex with a soluble cytoplasmic protein tomosyn-1 and a PM protein Stx1A in βcells (Figures 3, 4, and 5). In addition, the knockdown of Syt9 or tomosyn-1 increased the formation of Stx1A-SNARE complexes and insulin secretion ( Figure 4C,F). These outcomes support the possibility that the Stx1A-tomosyn-1-Syt9 complex is formed at the PM-insulin granules interface, functioning to attenuate the formation of SNARE complexes for the fusion of insulin granules to the PM in insulin exocytosis. This implies that the Syt9 isoform functions in the clamping of the Stx1A-SNARE complexes via tomosyn-1, supporting that the clamping action of Syt can be achieved by locking the SNARE in a fusion incompetent state. The ability of the Syt9-tomosyn-1-Stx1A complex to decrease the availability of Stx1A to form SNAREs rendering insulin granules transiently or completely nonfusogenic remains to be determined.
The availability of Stx1A to form SNARE complexes increases insulin secretion. The data presented here provide a mechanism by which Syt9 inhibits insulin secretion. Syt9 ablation decreased tomosyn-1 protein (not mRNA) levels and the binding of tomosyn-1 with Stx1A ( Figure 5M,N). Alterations in tomosyn-1 protein abundance affect Stx(s) function proportionally. 60 Similarly, increases in Stx1A-Snap25-Vamp2 SNARE complexes were observed upon tomosyn-1 knockdown ( Figure 4E,F). Rescuing tomosyn-1 protein levels blocked increases in insulin secretion due to Syt9 knockdown from βcells ( Figure 6B,C). Altogether these outcomes suggest that the loss of Syt9 via post-transcriptional mechanism decreases tomosyn-1 inhibitory function to increase Stx1A-SNARE-mediated insulin secretion. It has been shown that tomosyn-1 is modified posttranslationally. Hrd1 E3-polyubiquitin ligase 66,101 and protein kinase A 102 in neurons, and mono-ubiquitination in βcells, 93 regulates the function of tomosyn-1, modulating exocytosis. Based on these observations and our data, we propose that Syt9 stabilizes tomosyn-1 in the Syt9-tomosyn-1-Stx1A to form a nonfusogenic complex attenuating the fusion of insulin granules to the PM. Whether Syt9 has tomosyn-1-independent effects on insulin secretion requires further investigation. Moreover, how nutritional and hormonal signaling pathways regulate Syt9 and tomosyn-1 inhibitory functions in affecting Stx1A-SNARE complexes formation remains part of future studies.
The important question that needs to be addressed is the identification of the Syt isoform in βcells that facilitates the formation of Stx1A-SNARE complexes upon Syt9 loss. It is known that multiple Syt isoforms can be present on the same granules. Syt1 is present on granules containing Syt7 and Syt9 103 in PC12 cells. However, in βcells, insulin granules harbor distinct isoforms of Syts, conferring different functional roles in exocytosis. 103 A recent study reported that Syt7 and Syt9 are present on distinct granules in βcells. 104 Our data show that tomosyn-1 inhibits the Syt9 insulin granules from forming Stx1A-SNARE complexes. Thus, the Syt isoform that functions as a Ca 2+ sensor in facilitating the fusion of Syt9-insulin granules in insulin exocytosis remains to be determined. A possible role of a calcium sensor other than Syt9 that increases insulin secretion from the islets of Syt9 knockout mice remains to be identified. Moreover, further studies are required to assess whether Syt9 insulin granules are fusogenic and how these granules undergo fusion to the plasma membrane.
The limitation of this work is the use of a constitutive Syt9 deletion mouse model. However, it also enabled the identification of Syt9-tomosyn-1-Stx1A as an inhibitory complex that modulates the fusion of insulin granules. A previous study reported that mice with the pancreas-wide Syt9 deletion exhibited no difference in the plasma insulin levels and glucose clearance to intraperitoneal glucose injection compared to the control mice. 105 However, whether male or female mice were used in this study is unclear. If female mice were used, the reported phenotypes in this study are consistent with our data obtained using Syt9 −/− female mice (Figure Appendix S1). However, if male mice were used, there exists a divergence in the reported phenotypes between the two studies. Differences in the reported phenotypes could be due to using different Syt9 deletion mouse models used between the two studies (pancreas vs. constitutive). Moreover, differences in the genetic background, microbiome, and choice of glucose clearance (oral GTT vs. IPGTT) could potentially have contributed to differential outcomes. The oral GTT invokes a response from the gut incretins to release insulin secretion from βcells. Glucagon-like peptide-1 (GLP-1) is known to potentiate glucose-stimulated insulin release by inducing the PKA-mediated phosphorylation of Syt7 in pancreatic βcells. 54 Another study reported that the reduction of Syt9 by RNA interference reduced GH released from INS1E cells co-transfected with hGH reporter in response to the glucose stimulation 106,107 ; secretion of hGH was used as a surrogate for insulin secretion. We repeated this experiment in INS1(832/13) cells. Syt9 knockdown increased hGH secretion from INS1(832/13) cells ( Figure S3A-C). In this experiment, we also observed that hGH increased insulin secretion (Figure S3D,E). This observation was consistent with the published studies showing that GH directly increases insulin secretion from βcells (reviewed 108 ). Collectively, these outcomes suggest that using an hGH reporter as a surrogate for insulin secretion is not ideal. However, hGH as a reporter may still be useful for measuring secretion from other cell types. Our data implicate that Syt9 has an inhibitory role in regulating insulin secretion that is mediated by regulating tomosyn-1 protein abundance. The divergence of reported phenotypes between the published studies and this work suggests that a future study is warranted using βcell-specific Syt9 mice, which will be the next step in further clarifying the role of Syt9 in βcell function during the physiology and obesity pathophysiology.
In conclusion, this work has identified Syt9 as an endogenous inhibitor of Stx1A-SNARE complexes-mediate insulin secretion, consequently improving glucose clearance. Insights into the mechanism demonstrate that Syt9 forms a nonfusogenic molecular complex with tomosyn-1 and Stx1A between the PM and insulin granules, limiting the Stx1A-SNARE complex-mediated fusion of insulin granules to the PM. Further, the inhibitory effects of Syt9 on Stx1A-SNARE complex formation and insulin secretion are mediated by modulating the tomosyn-1 protein abundance and binding ability to Stx1A.