Synaptotagmin-7 phosphorylation mediates GLP-1–dependent potentiation of insulin secretion from β-cells

Significance

The present study shows that glucagon-like peptide-1 (GLP-1) potentiates insulin release by inducing the PKA-mediated phosphorylation of synaptotagmin-7 in pancreatic β-cells, thereby documenting that the hormone GLP-1 acts by directly enhancing Ca²⁺-triggered insulin exocytosis. PKA phosphorylates synaptotagmin-7 at a single serine residue in the linker between the synaptotagmin-7 transmembrane region and its C-terminal C₂ domains that bind Ca²⁺; accordingly, PKA phosphorylation of synaptotagmin-7 enhances insulin secretion without changing the Ca²⁺ dependence of secretion. The study documents that the efficacy of a synaptotagmin can be modulated by phosphorylation and highlights the importance of synaptotagmin-7 in mediating glucose-stimulated insulin secretion. Given the importance of enhancing β-cell function in the treatment of diabetes, the present findings may offer new pathways for developing future therapeutic strategies.

Abstract

Glucose stimulates insulin secretion from β-cells by increasing intracellular Ca²⁺. Ca²⁺ then binds to synaptotagmin-7 as a major Ca²⁺ sensor for exocytosis, triggering secretory granule fusion and insulin secretion. In type-2 diabetes, insulin secretion is impaired; this impairment is ameliorated by glucagon-like peptide-1 (GLP-1) or by GLP-1 receptor agonists, which improve glucose homeostasis. However, the mechanism by which GLP-1 receptor agonists boost insulin secretion remains unclear. Here, we report that GLP-1 stimulates protein kinase A (PKA)-dependent phosphorylation of synaptotagmin-7 at serine-103, which enhances glucose- and Ca²⁺-stimulated insulin secretion and accounts for the improvement of glucose homeostasis by GLP-1. A phospho-mimetic synaptotagmin-7 mutant enhances Ca²⁺-triggered exocytosis, whereas a phospho-inactive synaptotagmin-7 mutant disrupts GLP-1 potentiation of insulin secretion. Our findings thus suggest that synaptotagmin-7 is directly activated by GLP-1 signaling and may serve as a drug target for boosting insulin secretion. Moreover, our data reveal, to our knowledge, the first physiological modulation of Ca²⁺-triggered exocytosis by direct phosphorylation of a synaptotagmin.

Glucose-stimulated insulin secretion (GSIS) from pancreatic β-cells follows a biphasic time course consisting of an initial, transient first phase lasting 5–10 min followed by a slowly developing, sustained second phase (1). Type 2 diabetes (T2D) is associated with partial or complete loss of the first insulin secretion phase and a reduction in the second insulin secretion phase (2, 3). Incretins, especially GLP-1, boost GSIS in T2D patients, thereby improving glucose homeostasis (4). GLP-1 exerts its action by activating GLP-1R, a G-protein–coupled receptor expressed on the surface of β-cells, which leads to an increase of adenylate cyclase activity and production of cAMP. Elevated cAMP levels in β-cells enhance GSIS through PKA-dependent and -independent (mediated by Epac2) mechanisms (5, 6). Mouse models with constitutively increased PKA activity have established PKA’s predominant role in the GLP-1–induced potentiation of β-cell GSIS (7, 8), but the downstream effectors remain unidentified.

Insulin is secreted in response to glucose by regulated exocytosis of insulin-containing secretory granules. Electrical activity leads to opening of plasmalemmal voltage-gated Ca²⁺ channels (VGCCs) on the β-cell plasma membrane; the resulting increase in [Ca²⁺]_i then triggers Ca²⁺-dependent exocytosis (9). Insulin granule exocytosis is mediated by a multiprotein complex composed of soluble SNAP-receptor (SNARE) proteins (SNAP-25, Syntaxin, and synaptobrevin-2) and Sec1/Munc18-like (SM) proteins (Munc18-1) by a process that shares similarities with synaptic vesicle exocytosis in neurons (10). To date, numerous SNARE isoforms have been implicated in GSIS (11, 12), including Syntaxin-1, Syntaxin-4, SNAP-25 or SNAP-23, and synaptobrevin-2/3 (or VAMP2/3), whereas VAMP8, a nonessential SNARE for GSIS, may be involved in the regulation of GLP-1 potentiation of insulin secretion (13).

In addition to SNARE and SM proteins, a Ca²⁺ sensor is required to initiate membrane fusion during exocytosis. Synaptotagmins, expressed mainly in neurons and endocrine cells, share a similar domain structure: a short N-terminal domain, followed by a transmembrane domain, a linker region with variable length, and two tandem Ca²⁺-binding C₂ domains (C₂A and C₂B) at the C terminus (14, 15). Some synaptotagmins bind to phospholipids in a Ca²⁺-dependent manner and have been identified as major Ca²⁺ sensors for regulated exocytosis (14, 16). Synaptotagmin-1, -2, -7, and -9 function as Ca²⁺ sensors for neurotransmitter release, whereas synaptotagmin-1, -7 (Syt7), and -10 regulate hormone secretion and neuropeptide release (9, 17, 18). Specifically, Syt7 regulates insulin granule exocytosis in insulin-secreting cell lines (19, 20). Syt7 is highly expressed in human pancreatic β-cells, and Syt7 KO mice exhibit reduced insulin secretion and consequently impaired glucose tolerance following glucose stimulation (21–23). Collectively, these studies demonstrate that Syt7 is a major Ca²⁺ sensor mediating GSIS in β-cells.

Given that GLP-1 potentiates insulin secretion in a glucose-dependent manner, it is highly likely that its insulinotropic action is exerted distally to the initiation of electrical activity, possibly at the level of Ca²⁺ sensing and membrane fusion. Here we report that Syt7 is a stoichiometric substrate for PKA and functions as a downstream target of PKA activated by GLP-1 signaling. Compared with wild-type mice, Syt7 KO mice showed reduced insulin secretion ex vivo and in vivo in response to treatment with the GLP-1 analog exendin-4 in a manner that depended on Syt7 phosphorylation at serine-103. Our data not only provide a mechanism by which GLP-1 stimulates insulin secretion, but also report the physiological regulation of a synaptotagmin by phosphorylation.

Results

Syt7 Is Phosphorylated at Serine-103 by PKA.

Among multiple forms of Syt7 that are produced by extensive alternative splicing (24), the short form of Syt7 is the predominant form expressed in pancreatic β-cells. To examine whether Syt7 as the major Ca²⁺ sensor for insulin secretion may be a downstream target of GLP-1 signaling, we first tested whether Syt7 phosphorylation could be induced by forskolin, which stimulates adenylate cyclase and elevates cAMP similar to GLP-1 activation. We consistently observed that forskolin treatment of HEK293 cells expressing Myc-tagged Syt7, of insulin-secreting INS-1E cells, and of isolated mouse islets produced a robust increase in the apparent size of Syt7 as analyzed by SDS/PAGE (Fig. 1A), suggesting a cAMP-dependent modification of Syt7. The molecular weight shift occurred within 5 min after forskolin treatment, lasted for the duration of the stimulation, and was reversible upon washout of forskolin (Fig. 1B). These results suggest that Syt7 undergoes a posttranslational modification mediated by forskolin-activated protein kinase A (PKA).

Fig. 1.

Syt7 is phosphorylated at serine-103 by PKA. (A) Forskolin treatment induces an apparent shift in molecular weight. Images show Syt7 immunoblots of transfected HEK293 cells expressing Myc-Syt7, naive INS-1E cells, and naive mouse pancreatic islets that were treated with DMSO or 10 µM forskolin as indicated. (B) Forskolin-induced molecular weight shift of Syt7 is rapid. Syt7 was visualized by immunoblotting in transfected HEK293 cells expressing Myc-Syt7 that were treated with DMSO or 10 µM forskolin for 5, 15, 30, 60, and 120 min. Washout (W.O.) of forskolin after a 30-min incubation was performed by replacing with fresh medium for another 30-min incubation. (C) Immunoblots with a phospho-serine (pSer) antibody of Syt7 immunoprecipitates from transfected HEK293 cells confirm serine phosphorylation as a function of PKA activation. Cells expressing Myc-Syt7 were treated with DMSO or 10 µM forskolin. Myc-Syt7 was immunoprecipitated with Syt7 antibody and probed with Myc and phospho-serine antibody. (D) HEK293 cells expressing Myc-Syt7 were treated with Epac2 activator 8-pCPT-2′-O-Me-cAMP (50 nM), forskolin (10 µM), or forskolin (10 µM) plus PKA inhibitor H-89 (50 µM). (E) Syt7 phosphorylation in INS-1E cells is stimulated by GLP-1. Cells were treated with 3 mM glucose, 10 mM glucose, 20 mM glucose, 20 mM Na–pyruvate, 10 µM forskolin, and 50 nM GLP-1 for 30 min as indicated and analyzed by Syt7 or actin immunoblotting. (F) Syt7 phosphorylation in INS-1E cells is stimulated by the GLP1 receptor agonist exendin-4. Cells were treated with 50 nM 8-pCPT-2′-O-Me-cAMP, 10 μM forskolin, 10 µM forskolin plus 50 µM H-89, 50 nM exendin-4, and 50 nM exendin-4 plus 50 µM H-89 and analyzed by Syt7 or actin immunoblotting. (G) Schematic of the domain structure of Syt7. The transmembrane domain (TM), C₂ domains (C₂A and C₂B), alternative splicing site (triangle), and phosphorylation site (S103, arrow) are indicated. (H) Forskolin-induced molecular weight shift of Syt7 depends on serine-103. HEK293 cells expressing WT and various point mutants of Syt7 were treated with DMSO or 10 µM forskolin. Syt7 was visualized by immunoblotting with Myc antibody. (I) PKA-dependent response of phospho-inactive (S103A) and phospho-mimetic (S103E) mutants of Syt7. HEK293 cells were transfected with Myc-tagged Syt7-WT, Syt7-S103A, or Syt7-S103E and treated with DMSO or 10 µM forskolin, followed by immunoblotting with Myc antibody.

Because elevated cAMP also activates guanine-nucleotide exchange factor 2 (Epac2) in addition to PKA in insulin-secreting cells (5, 6), we directly tested whether the cAMP-dependent molecular weight shift in Syt7 reflects a phosphorylation event and not another type of modification. Toward this end, we immunoprecipitated Syt7 from control and forskolin-treated cells and immunoblotted the immunoprecipitates with anti–phospho-serine antibodies. We observed a robust signal in the forskolin-treated but not mock-treated sample (Fig. 1C), indicating that the modification represents Syt7 phosphorylation on a serine residue.

We next tested the effects of Epac2 activation and PKA inhibition on forskolin-induced Syt7 phosphorylation. Activation of Epac2 by 8-pCPT-2′-O-Me-cAMP failed to induce a molecular weight shift in Syt7. In contrast, the PKA inhibitor H-89 abolished the forskolin-induced Syt7 molecular weight shift (Fig. 1D). Taken together, these results indicate that the observed upper band is a phosphorylated form of Syt7 and that PKA phosphorylates Syt7 on a serine residue.

Considering that Syt7 is a Ca²⁺ sensor for exocytosis during GSIS, we tested whether Syt7 phosphorylation is stimulated by insulin secretagogues. We treated INS-1E cells with metabolic insulin secretagogues, such as high glucose and pyruvate, or with activators of the GLP-1–signaling pathway such as exendin-4 (a long-acting GLP-1R agonist) (25). We observed Syt7 phosphorylation, monitored via the Syt7 molecular weight shift on SDS/PAGE, only in samples treated with exendin-4 whose effects were similar to that of forskolin (Fig. 1 E and F). Consistent with the findings in HEK293 cells, activation of Epac2 failed to induce Syt7 phosphorylation in INS-1E cells, and elevation of glucose to 10 or 20 mM or addition of pyruvate were likewise without effect (Fig. 1 E and F). The finding that Syt7 is phosphorylated only in response to GLP-1/cAMP/PKA–signaling but not to metabolic insulin secretagogues suggests that Syt7 phosphorylation is not required for GSIS as such, but may mediate incretin- (i.e., GLP-1–) induced potentiation of insulin secretion.

Next we examined a series of Syt7 truncation and point mutants and determined the PKA-dependent phosphorylation site to be within the linker region (Fig. S1). To pinpoint the specific target residue, we evaluated by immunoblotting several point mutants of potential phosphorylation sites within the linker region in response to forskolin treatment. Only mutations of serine-103 (S103), but not any of the other mutations, abolished the forskolin-induced molecular weight shift of Syt7 (Fig. 1 G and H), suggesting that S103 is a putative PKA phosphorylation site. Substitutions of S103 to phospho-inactive S103A or phospho-mimetic S103E blocked the forskolin-induced molecular weight shift of Syt7 (Fig. 1I). Strikingly, the S103E substitution on its own induced a molecular weight shift (Fig. 1I). Interestingly, S103 was previously identified as a phosphorylation site by a phospho-proteome analysis and shown to be significantly phosphorylated in brain (26). The sequences surrounding S103 conform well to the consensus sequence of PKA phosphorylation sites, and S103 is highly conserved in mammalian Syt7 proteins (Fig. S2). Viewed together, these results show that Syt7 is phosphorylated on serine-103 in response to cAMP-mediated PKA activation in cultured cells and ex vivo in isolated islets.

Fig. S1.

Identification of the linker sequence as the target region of the forskolin- and cAMP-induced molecular weight shift of Syt7 (related to Fig. 1). (A) Schematics of the Syt7 truncation mutants. Various fragment of Syt7 was fused at the C-terminal with GFP to increase the total molecular weight of each fusion protein. (B) Forskolin or cAMP treatment induces molecular weight shift of Syt7 only in the fusion protein containing the linker region. HEK293 cells expressing various GFP-Syt7 truncation mutants were treated with DMSO, 10 µM forskolin, or 100 µM 8-pCPT-cAMP. Syt7 was visualized by immunoblotting with GFP antibody.

Fig. S2.

PKA-dependent phosphorylation site is evolutionarily conserved in vertebrate Syt7 proteins. Sequence alignment of mouse, rat, and human Syt7 proteins (short splice variant) showed conservation of the S103 phosphorylation site (asterisk). Syt7 sequences were from NCBI database: NP_061271.1 for mouse Syt7, NP_067691.1 for rat Syt7, and NP_004191.2 for human Syt7. Residuces are color coded based on their degree of conservation.

Syt7 Phosphorylation Potentiates Glucose-Stimulated Insulin Secretion.

To examine the effect of Syt7 phosphorylation on GSIS, we expressed Syt7 phospho-inactive (S103A) and phospho-mimetic (S103E) mutants in Syt7 KO islets (Fig. 2A) and measured insulin secretion from these islets under static incubation conditions. Consistent with previous findings (21), KO of Syt7 significantly reduced GSIS (Fig. 2B). Expression of Syt7-WT or Syt7-S103A fully reversed the defective GSIS caused by the Syt7 KO, suggesting that GSIS is not dependent on Syt7 phosphorylation. Importantly, expression of the phospho-mimetic mutant Syt7-S103E significantly enhanced insulin secretion (Fig. 2B).

Fig. 2.

Phosphorylation of Syt7 enhances glucose-stimulated insulin secretion in mouse islets. (A) Syt7 is phosphorylated on serine-103 in murine pancreatic islets. Islets from control or Syt7 KO (S7KO) mice were infected with adenoviruses expressing GFP, Syt7-WT, Syt7-S103A, or Syt7-S103E. Cell extracts from infected islets were prepared after potentiated GSIS assays and analyzed by immunoblotting for Syt7, GFP, and actin. (B) The phospho-mimetic Syt7 S103E mutant stimulates glucose-induced insulin secretion from murine islets. Mouse islets expressing Syt7 forms were prepared as described for A, and insulin secretion was stimulated in perfused islets with 16.7 mM glucose (insulin secretion is shown as percentage of total islet insulin content; n = 22–35 measurements from three independent islet preparations). (C) The phospho-mimetic Syt7 S103E mutant stimulates depolarization-induced insulin granule exocytosis in isolated β-cells. β-Cells were isolated from murine islets prepared as described for A and analyzed by capacitance measurements (Top) Representative traces shown above the stimulation protocol. (Bottom) Summary graphs of the average membrane capacitance increase (ΔCm) as an indicator of insulin granule exocytosis; n = 10–18 cells from three independent islet preparations). Data in B and C are means ± SEM. Statistical significance was assessed by Student’s t test comparing the various Syt7 KO conditions to the control (**P < 0.01; ***P < 0.001).

We next applied high-resolution capacitance measurements to single β-cells to study exocytosis evoked by a train of depolarization pulses under voltage clamp, which mimics β-cell conditions during GSIS (Fig. 2C). Viral infection and expression of various mutant proteins had no effect on cell size, as evidenced by the basal membrane capacitance measurements (Table S1). Consistent with the islet secretion data, exocytosis—as measured by membrane capacitance changes (ΔCm)—was significantly lower in Syt7 KO β-cells than in WT β-cells. Expression of Syt7-WT or Syt7-S103A fully rescued exocytosis in Syt7 KO β-cells. Strikingly, expression of the phospho-mimetic mutant Syt7-S103E not only rescued the Syt7 KO phenotype, but further enhanced β-cell exocytosis above WT levels (Fig. 2C). These experiments suggest that Syt7 in its nonphosphorylated state fulfills the function of a Ca²⁺ sensor for exocytosis during GSIS and that phosphorylation of Syt7 mediates the potentiation of GSIS by cAMP-PKA signaling.

Table S1.

Basal capacitance values of β-cells under various experimental conditions

	Untreated		Exenatide
	Mean ± SEM	No. of cells analyzed	Mean ± SEM	No. of cells analyzed	Statistics
Control	6.24 ± 0.40	8	6.57 ± 0.31	10	NS
S7KO + GFP	6.68 ± 0.41	9	6.16 ± 0.35	10	NS
S7KO + WT	6.86 ± 0.28	8	6.14 ± 0.24	10	NS
S7KO + S103A	6.76 ± 0.35	9	6.56 ± 0.29	11	NS
S7KO + S103E	6.72 ± 0.38	9	6.84 ± 0.56	10	NS

All data are means ± SEMs; numbers denote capacitance (Cm) measured in femtofarad (fF). Statistical comparisons were performed with Student’s t test. No significant differences were observed in any comparison. NS, nonsignificant.

Syt7 Is a Target of GLP-1 Action in Potentiating Insulin Secretion.

To directly examine whether GLP-1– and cAMP-dependent potentiation of GSIS are mediated through phosphorylation of Syt7 at S103, we performed capacitance measurements in Syt7 KO β-cells expressing Syt7-WT, Syt7-S103A, or Syt7-S103E with the addition of either 50 nM exendin-4 in the bath solution (Fig. 3A) or 0.1 mM cAMP in the recording pipette (Fig. 3B). As expected, in the control condition insulin granule exocytosis was dramatically enhanced by exendin-4 or cAMP (compare with controls in Fig. 2C vs. Fig. 3 A and B). Also as expected, the Syt7 KO suppressed β-cell exocytosis even in the presence of exendin-4 or cAMP. Syt7-WT rescued insulin exocytosis in Syt7 KO β-cells to a level similar to that of WT β-cells, whereas expression of the phospho-inactive Syt7-S103A still showed an approximately 50% reduction in insulin granule exocytosis compared with WT β-cells in the presence of either exendin-4 (Fig. 3A) or cAMP (Fig. 3B). In contrast to the findings without treatment in Fig. 2C, expression of the phospho-mimetic Syt7-S103E did not stimulate exocytosis beyond WT control or Syt7-WT rescued β-cells under exendin-4 or cAMP treatments. Thus, exendin-4 and cAMP occluded the effect of stimulating Syt7 phosphorylation (Fig. 3). Together, these findings suggest that the potentiation of insulin granule exocytosis by GLP-1 and cAMP-PKA signaling is mediated by phosphorylation of Syt7.

Fig. 3.

Syt7 phosphorylation mediates the potentiation of insulin-granule exocytosis by the GLP-1 receptor agonist exendin-4 or by cAMP. (A) The phospho-inactive S103A mutant of Syt7 impairs the increased β-cell exocytosis induced by treatment of the β-cells with 50 nM of exendin-4. β-Cells were isolated from WT control or Syt7 KO (S7KO) mice, and Syt7 KO cells were infected with adenoviruses expressing Syt7-WT, Syt7-S103A, or Syt7-S103E. Capacitance changes induced by depolarization pulses were measured (Top) Representative traces shown above the stimulation protocol. (Bottom) Summary graphs of the average membrane capacitance increase (ΔCm) as an indicator of insulin granule exocytosis; n = 10–15 cells from three independent islet preparations). Note that in the presence of exendin-4 the control capacitance increase is twofold higher than in the absence of exendin-4 (Fig. 2C). (B) Same as A, except that experiments were performed in the presence of 0.1 mM cAMP in the recording patch pipette. Data are means ± SEM. Statistical significance was assessed by Student’s t test comparing the various Syt7 KO conditions to the control (***P < 0.001).

GLP-1 Potentiation of Insulin Secretion Is Diminished in the Absence of Syt7.

To examine whether the Syt7 KO alters the insulin response of mice to increased glucose levels even in the presence of GLP-1R activation, we performed a regular oral glucose tolerance test. This experiment confirmed that Syt7 KO mice showed impaired glucose tolerance (Fig. 4A). Because orally administered glucose also stimulates GLP-1 release and Syt7 KO mice secrete less GLP-1 and insulin under glucose stimulation (27), it is unclear whether the oral glucose tolerance test deficiency that we observed is caused by reduced GLP-1 secretion from the gut, loss of insulin secretion, and/or loss of GLP-1 potentiation of insulin secretion in pancreatic islets. To address this issue, we pretreated Syt7 KO and littermate WT mice with exendin-4 and then used the i.p. glucose tolerance test. Syt7 KO mice continued to exhibit defective glucose clearance compared with their WT littermate controls, confirming that Syt7 is essential for a normal physiological response to increased glucose (Fig. 4B). Moreover, even after pretreatment with exendin-4, insulin secretion during the i.p. glucose tolerance test was significantly lower in Syt7 KO than in WT control mice (Fig. 4C), consistent with a critical role for Syt7 in transducing the GLP-1 receptor signal into a physiological increase in insulin secretion.

Fig. 4.

Loss of Syt7 disrupts GLP-1 potentiation of glucose-stimulated insulin secretion. (A) Syt7 KO impairs glucose tolerance in mice. Blood glucose levels during an oral glucose tolerance test measured in littermate WT (Control) and Syt7 KO (S7KO) mice. Glucose excursion was quantified as area under curve (AUC) during the 120-min test (n = 8–9 mice). (B and C) Syt7 KO decreases glucose tolerance and glucose-stimulated insulin secretion in mice treated with the GLP-1 agonist exendin-4. Blood glucose levels (B) and plasma insulin levels (C) during i.p. glucose tolerance tests performed 30 min after pretreatment of the mice with exendin-4 (s.c. injection at 1 μg/kg). Glucose excursion was quantified as AUC during the 120-min test (B) n = 15 mice per group. (C) n = 8–9 mice per group. Data are means ± SEM. Statistical significance was assessed by Student’s t test comparing the Syt7 KO to the control (*P < 0.05; **P < 0.01).

Discussion

GLP-1 analogs are an important class of diabetes therapies with numerous beneficial effects, one of which is their ability to potentiate insulin secretion and restore glucose homeostasis in diabetic patients (25). Insulin secretion is triggered by increases in cytoplasmic Ca²⁺ and coordinately accomplished by SNARE and SM proteins, synaptotagmins, and associated proteins in a process that bears strong resemblance to neurotransmitter release in synapses (9, 10, 28). Syt7 is a major Ca²⁺ sensor of insulin exocytosis and mediates the final response to elevated [Ca²⁺]_i (9, 21, 29). In addition to Syt7, other proteins, including Doc2 and Piccolo, may function as Ca²⁺ sensors in insulin secretion, but probably act in a modulatory role (30–32).

GLP-1 binds to its receptors on β-cells and activates Epac2 and PKA (25). Epac2 activation elevates Ca²⁺ levels by mobilizing Ca²⁺ from internal stores, thereby increasing insulin secretion (6, 33). PKA activation causes protein phosphorylation and is essential for the enhancement of insulin secretion by GLP-1 because pharmacological inhibitors of PKA abrogate the stimulatory effect of GLP-1 on insulin secretion (25). However, the molecular targets of PKA phosphorylation were poorly defined, which was the starting point of the present study. The fact that GLP-1 and PKA rapidly and fully potentiate insulin secretion within 5–15 min indicates that the effects of GLP-1/PKA signaling are not mediated by altered gene expression, but are likely produced by acute phosphorylation of protein components of the secretory machinery. Our study suggests that Syt7 is at least one of these protein components and thus represents a key point of convergence for the regulation of GSIS by GLP-1.

We show that PKA phosphorylates Syt7 at a single site (S103) in all cells tested and that this phosphorylation causes a major shift in electrophoretic mobility of Syt7, allowing us to conclude that its phosphorylation is stoichiometric. Moreover, we demonstrate that replacement of endogenous Syt7 with S103A-mutant Syt7 lacking the PKA-phosphorylation site prevents the potentiation of insulin secretion by GLP-1, whereas replacement of Syt7 with S103E-mutant Syt7 with a “phospho-mimetic” PKA-phosphorylation site substitution constitutively enhances stimulated insulin secretion and occludes GLP-1 action. It is likely that Syt7 is phosphorylated at additional sites where phosphorylation may not cause an electrophoretic shift and, although other kinases likely phosphorylate S103 and thereby other signaling pathways, probably modulate exocytosis via phosphorylating Syt7. Nevertheless, our findings establish at least one signaling pathway whereby phosphorylation of a synaptotagmin as the final mediator of Ca²⁺-triggered exocytosis potentiates exocytosis and demonstrate that this signaling pathway is activated in a physiologically important context.

Numerous studies on regulated exocytosis in different cell types show that phosphorylation of protein components of the secretory machinery is a common mechanism by which different signaling pathways control exocytosis of neurotransmitters, neuropeptides, and hormones. For example, SNAP-25 is phosphorylated at serine-187 by PKC, regulating exocytosis in neuronal cells and insulin-secreting cell lines (34, 35). Phosphorylation of Snapin by PKA may influence vesicle exocytosis in several cell types and also contribute to GLP-1–potentiated GSIS in β-cells (8, 36–38). Several synaptotagmins have been identified as targets of various kinases. Syt1, a neuronal Ca²⁺ sensor that regulates neurotransmitter release (39, 40), is phosphorylated by CaMKII, PKC, and CK2 in vitro (41–43). Syt2, a close homolog of Syt1, binds to and is phosphorylated on its C₂ domains by WNK1 [with no lysine (K) kinase] (44). Syt4 associates with and is phosphorylated by JNK both in vitro and in vivo (45). In addition, Syt9 and Syt12 are phosphorylated by PKC and PKA, respectively (46, 47). However, whether phosphorylation of synaptotagmins affects their role in regulated exocytosis in a physiologically relevant preparation has remained unclear (41, 42, 48). The present study clarifies this issue by demonstrating that phosphorylation of a synaptotagmin Ca²⁺ sensor physiologically modulates Ca²⁺-triggered exocytosis without being essential for the basic function of the synaptotagmin involved.

It has been proposed that cAMP/PKA stimulates insulin secretion by sensitizing the secretory machinery to Ca²⁺. However, β-cell exocytosis mediated by Syt7 with the phospho-mimetic S103E-mutation exhibits a similar Ca²⁺ dependence as exocytosis mediated by WT Syt7 (Fig. S3). This result is not unexpected as the phosphorylation site is far away from the Ca²⁺-binding C₂ domains. It seems likely that Syt7 phosphorylation affects its interaction with as-yet-unidentified target proteins binding to the linker sequence that contains the phosphorylation site. Future studies on Syt7-interacting partners will be required to address the precise molecular pathway by which Syt7 phosphorylation augments insulin secretion.

Fig. S3.

Phosphorylation site mutations do not affect the Ca²⁺ dependence of Syt7-mediated Ca²⁺-triggered exocytosis (related to Figs. 2 and 3). (A) Syt7 WT, S103A, or S103E was expressed in isolated single β-cells from S7KO mice. Membrane capacitance changes induced by a single 500-ms depolarization pulse were measured at different Ca²⁺ concentrations. (B) K_d values were calculated based on the plots in A. Data shown are means ± SEM. No significant difference in Ca²⁺ dependence was detected.

Previous studies have suggested that Rabphilin-3A and Snapin are also involved in the GLP-1– and PKA-dependent potentiation of GSIS (8, 49). It is not clear whether they function in parallel with or upstream of Syt7, but the general function of Rabphilin-3A and Snapin in exocytosis is not as fundamental as that of Syt7, which is an essential Ca²⁺ sensor for exocytosis. Although we favor the model that GLP-1–signaling pathways converge onto Syt7 to potentiate GSIS, more studies will be necessary to clarify the relation of Syt7 to other potential PKA targets in the modulation of GSIS.

In summary, our study describes stoichiometric phosphorylation of Syt7 by PKA in vitro and in vivo at a single site and demonstrates that, in the case of insulin secretion, Syt7 phosphorylation increases the efficacy of Syt7 as a Ca²⁺ sensor of exocytosis. The PKA-dependent phosphorylation of Syt7 likely accounts for much of the GLP-1–induced enhancement of insulin secretion, thus providing a mechanism of action by which GLP-1 agonists improve insulin secretion as therapeutic agents in diabetes.

Materials and Methods

Animal Studies.

Heterozygous Syt7 knockout (S7KO) (50) mice were used for breeding to generate homozygous mutants and littermate controls (21, 50). All animal studies were approved by A*STAR Institutional Animal Care and Use Committee (Protocol #110683 and #120758).

DNA Constructs.

The rat Syt7 cDNA was cloned into pCMV5-Myc vector in-frame with Myc tag at the 3′-end. The phosphorylation site mutations were obtained by using QuikChange II site-directed mutagenesis kit (Stratagene). The primers for Syt7 S103A and S103E mutants are 5′-AAC GGA GCC CCG TTC CGC TGT CTC GGA CCT CGT CAA-3′ and 5′-AAC GGA GCC CCG TTC CGA AGT CTC GGA CCT CGT CAA-3′, respectively. For rescue and overexpression experiments, Myc-tagged Syt7 (WT, S103A, or S103E) DNA fragments were cloned into pShuttle-CMV (Stratagene).

Virus Preparation and Infection.

Adenoviruses carrying Myc-tagged rat Syt7 or its mutants were prepared using the AdEasy Adenoviral Vector System (Stratagene). Forty-eight hours after infection, islet insulin secretion assays were performed, followed by protein extraction, to determine the insulin content.

Physiology Tests.

Oral and i.p. glucose tolerance tests (GTTs) were performed on male mice (12–16 wk of age) after 16 h of fasting essentially as previously described (21). Exendin-4 treatments were performed by administering the GLP-1 analog exendin-4 (exenatide or Byetta, Eli Lilly) to 12- to 16-wk-old male mice by s.c. injection 30 min before GTTs) at 1 μg/kg body weight.

Mouse Islet Isolation and Insulin Measurements.

Islets were isolated by using 1.5–2 mg/mL of collagenase P (Roche) for each mouse. After digestion, islets were cultured overnight at 11 mM glucose in RPMI medium 1640 supplemented with 10% (vol/vol) FBS, 1% (vol/vol) penicillin, and streptomycin. Insulin levels were measured with an insulin ELISA kit and normalized against the total insulin of the islets.

Electrophysiology.

Mouse primary β-cells were dispersed by trypsin digestion from islets and cultured on glass coverslips that were precoated with poly-l-lysine. After overnight incubation in RPMI medium 1640, dispersed β-cells were infected with adenovirus containing coding sequences of Syt7 and various phospho-mutants. Electrophysiology experiments were performed 96 h after viral infection. Membrane currents and capacitance were recorded from single β-cells by using the standard whole-cell patch-clamp technique as described previously (51, 52) with detailed information in SI Materials and Methods. Exocytosis of insulin-containing granules was detected as changes in cell membrane capacitance (C_m), which was estimated by the Lindau–Neher technique implementing the “Sine+DC” feature of the lock-in module (53). All C_m measurements were performed at 28 °C.

Reagents.

All restriction enzymes, T4 ligase, and DNA polymerases were purchased from New England BioLabs, unless indicated otherwise. INS-1E cells were kindly provided by Claes Wollheim, Lund University, Lund, Sweden. Rabbit polyclonal antibody against Syt7, mouse monoclonal antibodies against SNAP25, Syntaxin1 and Munc18 were purchased from Synaptic Systems. Mouse monoclonal antibodies against c-Myc and β-Actin were obtained from Santa Cruz Biotechnology. All commonly used chemicals were purchased from Sigma-Aldrich.

Statistical Analysis.

The data were presented as means ± SEM. Comparison of data were made by using two-tailed Student’s t test. Statistical significance was denoted as *P < 0.05, **P < 0.01, and ***P < 0.001.

For further experimental details, see SI Materials and Methods.

SI Materials and Methods

Cell Culture.

HEK293 cells were cultured in DMEM with 10% (vol/vol) heat-inactivated FBS, 50 U/mL penicillin, and 50 µg/mL streptomycin (GIBCO Cell Culture). INS-1E cells were grown in RPMI medium 1640 containing 11 mM glucose and 2 mM l-glutamine supplemented with 10 mM Hepes (pH 7.5), 10% (vol/vol) heat-inactivated FBS, 1 mM Na–pyruvate, 50 µM β-mercaptoethanol, 50 U/mL penicillin, and 50 µg/mL streptomycin (49). Cells were maintained in a humidified chamber with 5% CO₂ at 37 °C.

Physiology Tests.

Oral and i.p. GTTs were performed on male mice (12–16 wk of age) after 16 h of fasting essentially as previously described (21). For oral glucose tolerance test and i.p. glucose tolerance test, a dose of 2 g/kg body weight of glucose was given by oral gavage or i.p., respectively. Blood glucose levels were measured at the indicated times. At defined time points, blood samples were collected for plasma insulin measurements by using ELISA (Mercodia). Exendin-4 treatments were performed by administering the GLP-1 analog exendin-4 (exenatide or Byetta, Eli Lilly) to 12- to 16-wk-old male mice by s.c. injection 30 min before GTT) at 1 μg/kg body weight.

Electrophysiology.

Mouse primary β-cells were dispersed by trypsin digestion from islets and cultured on glass coverslips that were precoated with poly-l-lysine. After overnight incubation in RPMI medium 1640, dispersed β-cells were infected with adenovirus containing coding sequences of Syt7 and various phospho-mutants. Electrophysiology experiments were performed 96 h after viral infection. Membrane currents and capacitance were recorded from single β-cells by using the standard whole-cell patch-clamp technique as described previously (50, 51) with detailed information in SI Materials and Methods. Exocytosis was triggered by a single 500-ms depolarization pulse or a train of 10, 500-ms depolarizing pulses from −70–0 mV. Pipettes were filled with intracellular solution containing 125 mM potassium glutamate, 10 mM KCl, 10 mM NaCl, 1 mM MgCl₂, 5 mM Hepes, 0.05 mM EGTA, 0.1 mM cAMP, and 4 mM MgATP (pH 7.1). cAMP (0.1 mM) was added into intracellular solution when required. Extracellular solution contained 118 mM NaCl, 20 mM tetraethylammonium chloride, 5.6 mM KCl, 2.6 mM CaCl₂, 1.2 mM MgCl₂, and 5 mM Hepes (pH 7.4). Cells were stimulated at low frequency (<0.05 Hz) to allow full recovery of exocytotic capacity between pulses. Measurements were performed by using EPC10-2 patch clamp amplifier and PatchMaster software (HEKA Eletronik). Exocytosis of insulin-containing granules was detected as changes in cell membrane capacitance (C_m), which was estimated by the Lindau–Neher technique implementing the “Sine+DC” feature of the lock-in module (52). All C_m measurements were performed at 28 °C.

Western Blotting and Immunoprecipitations.

Islets or cell protein lysates were prepared in lysis buffer. Between 20–30 μg protein lysate was separated on 10% SDS/PAGE and transferred onto nitrocellulose membrane by using i-Blot (Invitrogen). Proteins were identified by specific antibodies. For immunoprecipitation assay, ∼500 μg of cell lysate was incubated with primary antibody at 4 °C for 2–3 h, followed with protein A/G agarose beads for another 2 h. Beads were washed three times with lysis buffer and resuspended in SDS sample buffer. Immunoprecipitated fractions, together with input control, were separated by SDS/PAGE and probed with corresponding antibodies.

Reagents.

All restriction enzymes, T4 ligase, DNA polymerases were purchased from New England BioLabs, unless indicated otherwise. INS-1E cells were kindly provided by Claes Wollheim. Rabbit polyclonal antibody against Syt7, mouse monoclonal antibodies against SNAP25, Syntaxin1, and Munc18 were purchased from Synaptic Systems. Mouse monoclonal antibodies against c-Myc and β-Actin were obtained from Santa Cruz Biotechnology. All commonly used chemicals were purchased from Sigma-Aldrich.