Skip to main content
PMC home page
eLife logoLink to eLife
. 2019 May 20;8:e41711. doi: 10.7554/eLife.41711

Fusion pore regulation by cAMP/Epac2 controls cargo release during insulin exocytosis

Alenka Guček 1, Nikhil R Gandasi 1, Muhmmad Omar-Hmeadi 1, Marit Bakke 2, Stein O Døskeland 2, Anders Tengholm 1, Sebastian Barg 1,✉
Editors: Axel T Brunger3, Vivek Malhotra4
PMCID: PMC6557626  PMID: 31099751

Abstract

Regulated exocytosis establishes a narrow fusion pore as initial aqueous connection to the extracellular space, through which small transmitter molecules such as ATP can exit. Co-release of polypeptides and hormones like insulin requires further expansion of the pore. There is evidence that pore expansion is regulated and can fail in diabetes and neurodegenerative disease. Here, we report that the cAMP-sensor Epac2 (Rap-GEF4) controls fusion pore behavior by acutely recruiting two pore-restricting proteins, amisyn and dynamin-1, to the exocytosis site in insulin-secreting beta-cells. cAMP elevation restricts and slows fusion pore expansion and peptide release, but not when Epac2 is inactivated pharmacologically or in Epac2-/- (Rapgef4-/-) mice. Consistently, overexpression of Epac2 impedes pore expansion. Widely used antidiabetic drugs (GLP-1 receptor agonists and sulfonylureas) activate this pathway and thereby paradoxically restrict hormone release. We conclude that Epac2/cAMP controls fusion pore expansion and thus the balance of hormone and transmitter release during insulin granule exocytosis.

Research organism: Human, Mouse

eLife digest

Insulin is the hormone that signals to the body to take up sugar from the blood. Specialized cells in the pancreas – known as β-cells – release insulin after a meal. Before that, insulin molecules are stored in tiny granules inside the β-cells; these granules must fuse with the cells’ surface membranes to release their contents. The first step in this process creates a narrow pore that allows small molecules, but not the larger insulin molecules, to seep out. The pore then widens to release the insulin. Since the small molecules are known to act locally in the pancreas, it is possible that this “molecular sieve” is biologically important. Yet it is not clear how the pore widens.

One of the problems for people with type 2 diabetes is that they release less insulin into the bloodstream. Two kinds of drugs used to treat these patients work by stimulating β-cells to release their insulin. One way to achieve this is by raising the levels of a small molecule called cAMP, which is well known to help prepare insulin granules for release. The cAMP molecule also seems to slow the widening of the pore, and Gucek et al. have now investigated how this happens at a molecular level.

By observing individual granules of human β-cells using a special microscope, Gucek et al. could watch how different drugs affect pore widening and content release. They also saw that cAMP activated a protein called Epac2, which then recruited two other proteins – amisyn and dynamin – to the small pores. These two proteins together then closed the pore, rather than expanding it to let insulin out. Type 2 diabetes patients sometimes have high levels of amisyn in their β-cells, which could explain why they do not release enough insulin. The microscopy experiments also revealed that two common anti-diabetic drugs activate Epac2 and prevent the pores from widening, thereby counteracting their positive effect on insulin release. The combined effect is likely a shift in the balance between insulin and the locally acting small molecules.

These findings suggest that two common anti-diabetic drugs activate a common mechanism that may lead to unexpected outcomes, possibly even reducing how much insulin the β-cells can release. Future studies in mice and humans will have to investigate these effects in whole organisms.

Introduction

Insulin is secreted from pancreatic β-cells and acts on target tissues such as muscle and liver to regulate blood glucose. Secretion of insulin occurs by regulated exocytosis, whereby secretory granules containing the hormone and other bioactive peptides and small molecules fuse with the plasma membrane. The first aqueous contact between granule lumen and the extracellular space is a narrow fusion pore (upper limit 3 nm; Albillos et al., 1997) that is thought to consist of both lipids and proteins (Bao et al., 2016; Sharma and Lindau, 2018). At this stage, the pore acts as a molecular sieve that allows release of small transmitter molecules such as nucleotides and catecholamines, but traps larger cargo (Obermüller et al., 2005; Barg et al., 2002; MacDonald et al., 2006; Alvarez de Toledo et al., 1993). Electrophysiological experiments have shown that the fusion pore is short-lived and flickers between closed and open states, suggesting that mechanisms exist that stabilize this channel-like structure and restrict pore expansion (MacDonald et al., 2006; Hanna et al., 2009; Breckenridge and Almers, 1987; Lollike et al., 1995). The pore can then expand irreversibly (termed full fusion), which leads to mixing of granule and plasma membrane and release of the bulkier hormone content (Obermüller et al., 2005; Barg et al., 2002; Anantharam et al., 2010). Alternatively, the pore can close indefinitely to allow the granule to be retrieved, apparently intact, into the cell interior (termed kiss-and-run or cavicapture) (Obermüller et al., 2005; MacDonald et al., 2006; Taraska et al., 2003; Tsuboi and Rutter, 2003; Shin et al., 2018). Estimates in β-cells suggest that 20–50% of all exocytosis in β-cells are transient kiss-and-run events that do not lead to insulin release (Obermüller et al., 2005; MacDonald et al., 2006). However, kiss-and-run exocytosis contributes to local signaling within the islet because smaller granule constituents, such as nucleotides, glutamate or GABA, are released even when the fusion pore does not expand. Within the islet, ATP synchronizes β-cells (Hellman et al., 2004), and has both inhibitory (Salehi et al., 2005; Poulsen et al., 1999) and stimulatory (Richards-Williams et al., 2008) effects on insulin secretion. ATP suppresses glucagon release from α-cells (Tudurí et al., 2008), and activates macrophages (Weitz et al., 2018). Interstitial GABA leads to tonic GABA-A receptor activation and α-cell proliferation (Jin et al., 2013; Ben-Othman et al., 2017), and glutamate stimulates glucagon secretion (Cabrera et al., 2008).

Regulation of fusion pore behavior is not understood mechanistically, but several cellular signaling events affect both lifetime and flicker behavior. Pore behavior has been shown to be regulated by cytosolic Ca2+, cAMP, PI(4,5)P2, and activation of protein kinase C (PKC) (MacDonald et al., 2006; Hanna et al., 2009; Alés et al., 1999; Calejo et al., 2013; Scepek et al., 1998) and recent superresolution imaging indicates that elevated Ca2+ and dynamin promote pore closure (Shin et al., 2018; Chiang et al., 2014). Both myosin and the small GTPase dynamin are involved in fusion pore restriction (Jackson et al., 2015; Tsuboi et al., 2004; Graham et al., 2002; Artalejo et al., 1995; Aoki et al., 2010), and assembly of filamentous actin promotes fusion pore expansion (Wen et al., 2016), suggesting a link to endocytosis and the cytoskeleton. In β-cells of type-2 diabetics, upregulation of amysin leads to decreased insulin secretion because fusion pore expansion is impaired (Collins et al., 2016), and the Parkinson’s related protein α-synuclein promotes fusion pore dilation in chromaffin cells and neurons (Logan et al., 2017), thus providing evidence for altered fusion pore behavior in human disease.

Inadequate insulin secretion in type-2 diabetes (T2D) is treated clinically by two main strategies. First, sulfonylureas (e.g. tolbutamide and glibenclamide) close the KATP channel by binding to its regulatory subunit SUR1, which leads to increased electrical activity and Ca2+-influx that triggers insulin secretion (Henquin, 2000). Sulfonylureas are given orally and are first-line treatment for type-2 diabetes in many countries. Second, activation of the receptor for the incretin hormone glucagon-like peptide 1 (GLP-1) raises cytosolic [cAMP] and thereby increases the propensity of insulin granules to undergo exocytosis. Both peptide agonists of the GLP-1 receptor (e.g. exendin-4) and inhibitors of DPP-4 are used clinically for this purpose. The effect of cAMP on exocytosis is mediated by a protein-kinase A (PKA)-dependent pathway, and by Epac2, a guanine nucleotide exchange factor for the Ras-like small GTPase Rap (Kawasaki et al., 1998) that is a direct target for cAMP (Ozaki et al., 2000) and is recruited to insulin granule docking sites (Alenkvist et al., 2017). Epac2 has also been suggested to be activated by sulfonylureas (Zhang et al., 2009), which may underlie some of their effects on insulin secretion.

Here, we have studied fusion pore regulation in pancreatic β-cells, using high-resolution live-cell imaging. We report that activation of Epac2, either through GLP1-R/cAMP signaling or via sulfonylurea, restricts expansion of the insulin granule fusion pore by recruiting dynamin and amisyn to the exocytosis site. Activation of this pathway by two classes of antidiabetic drugs therefore hinders full fusion and insulin release, which is expected to reduce their effectiveness as insulin secretagogues.

Results

cAMP-dependent fusion pore restriction is regulated by Epac but not PKA

To monitor single granule exocytosis, human pancreatic β-cells were infected with adenovirus encoding the granule marker NPY-Venus and imaged by TIRF microscopy. Exocytosis was evoked by local application of a solution containing 75 mM K+, which leads to rapid depolarization and Ca2+ influx. Visually, two phenotypes of granule exocytosis were observed. In the first, termed full fusion, fluorescence of a granule that was stably situated at the plasma membrane suddenly vanished during the stimulation (in most cases within <100 ms; Figure 1a–c, left panels). Since the EGFP label is relatively large (3.7 nm vs 3 nm for insulin monomers), this is interpreted as rapid pore widening that allowed general release of granule cargo. The sudden release of material may suggest that this release coincided with the collapse of the granule into the plasma membrane, but we cannot exclude that at least some granules remained intact (Taraska et al., 2003; Shin et al., 2018; Tsuboi et al., 2004; Huang et al., 2018). In the second type, the rapid loss of the granule marker was preceded by an increase in its fluorescence that could last for several seconds (flash events, Figure 1a–c, right panels). We and others have previously shown (Taraska et al., 2003; Ferraro et al., 2005; Gandasi and Barg, 2014) that this reflects neutralization of the acidic granule lumen and dequenching of the EGFP-label, before the labeled cargo is released. Since this neutralization occurs as the result of proton flux through the fusion pore, the fluorescence timecourse of these events can be used to quantitatively study fusion pore behavior.

Figure 1. cAMP-dependent fusion pore restriction depends on Epac (but not PKA).

(a) Examples of single granule exocytosis in human β-cells expressing NPY-Venus and challenged with 75 mM K+. Full fusion (left) and flash event (right), where sudden loss of the granule label was preceded by a transient fluorescence increase. Arrows indicate moment of fusion pore opening (orange) and content release (blue). (b) Cartoons illustrating the interpretation of events in a. (c) Fluorescence time courses for the events in b. Overlaid (green) are fitted functions used to estimate NPY release time. (d) Fraction of flash events in experiments as in a-c, in cells exposed to the indicated agents; forskolin (fsk, p=0.01 vs ctrl), exendin-4 (Ex4, p=0.02 vs ctrl), ESI-09 (p=3*10−4 vs fsk), S223 (p=0.04 vs ctrl), fsk +S223 (p=0.99 vs fsk), RP-8 (p=0.91 vs fsk) and Rp-8 +S223 (p=0.19 vs ctrl; Kruskal Wallis/Dunn). Number of donors analyzed: 7 (CTR); 5 (fsk); 4 (Ex4); 7 (fsk +ESI09); 6 (S223); 6 (fsk +S223); 7 (RP-8); 4 (Rp-8 +S223). n, number of cells. (e) Cumulative frequency histograms of NPY release times; fsk (p=9*10−7 vs ctrl), Ex4 (p=1*10−6 vs ctrl), ESI-09 (p=2*10−4 vs fsk), S223 (p=4*10−6 vs ctrl),Fsk +S223 (n.s. vs fsk), RP-8 (n.s. vs fsk) and RP-8 +S223 (p=0.016 vs ctrl); Kolmogorov-Smirnov test). Inset shows median NPY release times for 170 (CTR), 197 (fsk), 155 (Ex4), 81 (ESI-09), 240 (S223), 328 (fsk +S223), 277 (RP-8) and 227 (Rp-8 +S223) events. (f) Exocytosis during 40 s of K+-stimulation for control (CTR) and with forskolin (fsk, 2 µM, p=0.002 vs ctrl; Kruskal Wallis/Dunn) or Exendin-4 (Ex4, 10 nM, p=0.005 vs ctrl) or S223 (5 µM, p=0.002 vs ctrl) or RP-8 +S223 (p=0.012 vs ctrl and n.s. vs S223) in the bath solution. Inhibitors of Epac (ESI-09, 10 µM, p=9*10-7 vs fsk) or PKA (RP-8, 100 µM, n.s. vs fsk) or Epac2 activator S223 (n.s. vs fsk), one-way ANOVA with Games-Howell post hoc test) were supplied in addition to forskolin. Flash exocytosis (in color) and full fusions (in white) are shown separately. n, number of cells.

Figure 1.

Figure 1—figure supplement 1. cAMP increases NPY release times in INS1 cells.

Figure 1—figure supplement 1.

(a–b) K+ stimulated exocytosis of NPY-tdmOrange2 granules in INS1 cells is significantly increased in presence of (A) forskolin (fsk; 2 µM; p=0.002) or (B) exendin-4 (Ex4; 10 nM; p=0.03). CTR, n = 13 cells; Fsk, n = 15; Ex4, n = 16. n of preps: 4 (CTR); 5 (+fsk); 2 (+Ex4). (c) Total exocytosis in (A–B) separated for events with flashes (in color) and full fusion events (in white); significance (t-test) is given for flash events. (d) Cumulative frequency histograms of NPY release times in (A–B); note longer NPY release times in presence of fsk (p=0.011) or Ex4 (p=0.018, Kolmogorov-Smirnov test). Inset plots the median NPY release times for 38 (CTR), 119 (fsk) and 111 (Ex4) events. (e) Percentage of flash events (p=0.23 for fsk, p=0.14 for Ex4 vs. control, u-test). n, number of cells.
Open in a new tab

In the following, we will report two parameters that reflect fusion pore behavior, the fraction of exocytosis events with flash phenotype (indicating restricted pores, about 40% in control conditions; Figure 1d), and the duration of the flash, referred to as ‘NPY release times’. The latter was estimated by fitting a discontinuous function to the fluorescence timecourse (see Figure 1c, green lines and Figure 1e), which limits the analysis to granules that eventually released their peptide content. The distribution of the NPY release times followed a mono-exponential function and was on average 0.87 ± 0.12 s (186 granules in 26 cells) in control conditions (Figure 1e). Such events are increased by elevated cAMP (MacDonald et al., 2006; Hanna et al., 2009) and likely other conditions that stabilize the fusion pore. Indeed, when forskolin (2 µM; fsk) was added to the bath solution we observed a twofold increase of exocytosis rate (Figure 1f), a threefold increase of NPY release times (Figure 1e), and a nearly doubled fraction of events with restricted fusion pores (Figure 1d,f). The GLP-1 agonist exendin-4 (10 nM; Ex4) had comparable effects (Figure 1d–f). Effects similar to those observed for human β-cells (Figure 1) were observed in the insulin secreting cell line INS-1 (Figure 1—figure supplement 1).

The effect of fsk on fusion pore behavior was mimicked by the specific Epac2 agonist S223 (Schwede et al., 2015). Incubation with S223-acetomethoxyester (5 µM) increased the fraction of flash events by 60% (Figure 1d), doubled average NPY release times (Figure 1e) and doubled the event frequency (Figure 1f); the effects of fsk and S223 were not additive. In contrast, the Epac-inhibitor ESI-09 decreased the exocytosis rate in the presence of fsk by 80% (Figure 1f), and the average NPY release time and the fraction of flash events were both reduced by 60% (Figure 1d–e). PKA inhibition with Rp8-Br-cAMPS (Gjertsen et al., 1995) decreased neither the fraction of flash events, nor average NPY release times (Figure 1e). The results indicate that Epac rather than PKA is responsible for cAMP-dependent fusion pore regulation. Paradoxically, Epac activation increases the rate of exocytosis but slows the rate of peptide release from individual granules.

Epac2 overexpression restricts fusion pores and prolongs their lifetime

We studied the effect of Epac2 overexpression on fusion pore regulation. INS-1 cells were co-transfected with EGFP-Epac2 and NPY-tdmOrange2 and fluorescence was recorded simultaneously in both color channels. Epac2 overexpression had no effect on the overall exocytosis rate in either absence or presence of fsk (Figure 2a), but increased the rate of flash events (Figure 2b–c), supporting our finding, based on manipulation of the endogenous Epac2 activity, that Epac2 is involved in fusion pore regulation (Figure 1d). NPY release times in cells overexpressing Epac2 increased threefold in the absence of fsk, and were similar to controls in presence of fsk (Figure 2d). This indicates that a high Epac concentration can achieve sufficient activity to affect insulin secretion even at basal cAMP level, likely because cAMP acts in part by increasing the Epac concentration at the plasma membrane (Alenkvist et al., 2017).

Figure 2. Epac2 overexpression prolongs NPY release times.

Figure 2.

Open in a new tab

(a) Cumulative exocytosis in INS-1 cells stimulated with 75 mM K+; gray for control cells, purple for cells expressing Epac2-EGFP (both also expressed NPY-tdmOrange2); fsk indicates forskolin in the bath solution. CTR, n = 13 (4 preps); Epac2, n = 11 (2 preps); CTR +fsk, n = 15 (5 preps); Epac2 +fsk, n = 14 cells (2 preps). (b) Total exocytosis in (a), separated into flash events (color) and full fusion (white). Epac2 expression reduced full fusion events (no fsk p=0.06; with fsk p=0.01, Kruskal Wallis/Dunn). n, number of cells. (c) Fraction of flash events in (a–b). (Kruskal Wallis/Dunn). n, number of cells. (d) NPY release times for conditions in a-c. Epac overexpression increased NPY release times in absence (p=0.014) but not in presence of fsk (p=0.87, Kolmogorov-Smirnov test). Inset shows the NPY release times for 38 (CTR), 27 (Epac2), 119 (CTR +fsk) and 77 (Epac2 +fsk) events.

ATP release is accelerated upon Epac inhibition

To test if cAMP-dependent fusion pore restriction affects release of small transmitter molecules, we quantified nucleotide release kinetics from individual granules using patch clamp electrophysiology. The purinergic receptor cation channel P2X2, tagged with RFP (P2X2-RFP), was expressed in INS-1 cells as an autaptic nucleotide sensor (Obermüller et al., 2005) (Figure 3a). The cells were voltage-clamped in whole-cell mode and exocytosis was elicited by including a solution with elevated free Ca2+ (calculated 600 nM) in the patch electrode. In this configuration, every exocytosis event that co-releases nucleotides causes an inward current spike, similar to those observed by carbon fiber amperometry (Figure 3a–b). Including cAMP in the pipette solution increased the frequency of current spikes by 50%, consistent with accelerated exocytosis. This effect of cAMP was blocked if the Epac inhibitor ESI-09 was present (Figure 3b–c). The current spikes (see Figure 3a, right) reflect nucleotide release kinetics during individual exocytosis events. In the presence of cAMP, but not cAMP + ESI-09, they were markedly widened as indicated by on average 20% longer half-widths (Figure 3d), 30% longer decay constants (τ, Figure 3e), and 40% slower rising phases (25–75% slope, Figure 3f), compared with control. This indicates that nucleotide release is slowed by cAMP, likely because of changed fusion pore kinetics. Since the effect is blocked by ESI-09, we conclude that the cAMP effect probably is mediated by Epac.

Figure 3. Cytosolic cAMP slows ATP release by activating Epac.

Figure 3.

Open in a new tab

(a) Electrophysiological detection of nucleotide release events in INS-1 cells expressing P2X2-RFP. Cartoon of the assay (left) and example current spike (black) with fit and analysis parameters (red; Thalf, tau and slope during 25% to 75% of peak). (b) Representative P2X2 currents for control (black), and with cAMP (green) or with cAMP together with ESI-09 (purple) in the electrode solution. (c) Spike frequency conditions in (b). n of events (on top) and n of cells (on bars); two preps for each condition. (d–f) Cumulative frequency histograms of spike half width (d), decay constant tau (e), and slope of the rising phase (25% and 75% of peak, (f)) for CTR (n = 410 spikes, 14 cells), +cAMP (n = 1240, 14 cells) and +ESI-09 + cAMP (n = 552, 15 cells) with medians in the insets. cAMP increased half-width (p=4.1*10−31 vs ctrl, Kolmogorov-Smirnov test), tau (p=2.7*10−32, Kolmogorov-Smirnov test), and rising slope (p=4.7*10−19, Kolmogorov-Smirnov test); the effects were reversed by ESI-09 (p=3.4*10−21, p=3.6*10−22, and p=1.3*10−9, Kolmogorov-Smirnov test), respectively.

cAMP-dependent fusion pore regulation is absent in Epac2-/- (Rapgef4-/-) β-cells

Since ESI-09 blocks all Epac isoforms (Zhu et al., 2015), we characterized fusion pore behavior in isolated β-cells from Epac2-/- (Rapgef4-/-) mice that lack all splice variants of Epac2 (Kopperud et al., 2017). Cells from WT or Epac2-/- mice were infected with adenovirus encoding the granule marker NPY-tdmOrange2 and challenged with 75 mM K+ (Figure 4a–b). In the absence of forskolin, exocytosis was significantly slower in Epac2-/- cells than WT cells, and the fraction of flash-associated exocytosis events was five-fold lower (Figure 4c–e). This was paralleled by strikingly shorter fusion pore life-times in Epac2-/- cells compared with WT (Figure 4f). The data suggest that Epac2 is partially activated in these conditions, consistent with elevated cAMP levels in mouse β-cells in hyperglycemic conditions (Dyachok et al., 2008). As expected, forskolin increased both exocytosis (Figure 4e) and the fraction of flash events (Figure 4c) of WT cells. In contrast, forskolin failed to accelerate exocytosis in Epac2-/- cells, and the fraction of flash events was similar with or without forskolin (Figure 4c,f–g). We conclude therefore that the effects of cAMP on fusion pore behavior are mediated specifically by Epac2.

Figure 4. Fusion pores expand rapidly in Epac2-/- (Rapgef4-/-) mice.

Figure 4.

Open in a new tab

(a–b) Examples of NPY-tdmOrange2 exocytosis events in β-cells from Epac2-/- mice or from wildtype littermates, stimulated with 75 mM K+ in presence of forskolin. Note absence of a flash in Epac2 ko. (c) Fraction of flash events for experiments in (a–b); differences are significant in absence (p=0.027, Kruskal Wallis/Dunn test) or presence of fsk (p=0.011). Number of mice: 4 (WT); 4 (Epac2 KO); 5 (WT + fsk); 2 (Epac2 KO + fsk). n, number of cells. (d) Cumulative exocytosis for experiments in absence of forskolin (a,c left) for wildtype (black) and Epac2-/- cells (red), differences are n.s. (e) Cumulative exocytosis for experiments in presence of forskolin (b,c right) for wildtype (black) and Epac2-/- cells (red). p=0.003, Kruskal Wallis/Dunn test. (f–g) Cumulative frequency histograms and medians (inset) of NPY release times for exocytotic events in d (no forskolin, 23 events for wt, 22 for Epac2-/-) and E (with forskolin, 50 events for wt, nine for Epac2-/-). Differences in f are significant (p=0.043; Kolmogorov-Smirnov test).

Sulfonylureas delay fusion pore expansion through the same pathway as cAMP

Sulfonylureas have been reported to activate Epac (Zhang et al., 2009), in addition to their classical role that involves the sulfonylurea receptor (SUR). We therefore tested the effect of sulfonylureas on fusion pore behavior. INS-1 cells expressing NPY-tdmOrange2 were tested with three types of sulfonylureas, with different relative membrane permeability (tolbutamide < glibenclamide < gliclazide). In addition, diazoxide (200 µM) was present to prevent electrical activity. Exocytosis was not observed under these conditions, but could be triggered by local application of elevated K+ (75 mM). In the absence of fsk, the sulfonylureas accelerated K+-stimulated exocytosis about twofold over that observed in control (Figure 5b, left), which is consistent with earlier findings that sulfonylureas augment insulin secretion via intracellular targets (Barg et al., 1999). This effect was entirely due to an increase in flash-associated exocytosis events (Figure 5b–c) and the average NPY release time increased accordingly in the presence of sulfonylurea (Figure 5d). Fsk strongly stimulated both flash-associated and full fusion exocytosis in absence of sulfonylurea (Figure 5b–c, middle); under these conditions, sulfonylureas tended to decrease full-fusion exocytosis without effect on the frequency of flash-associated events (Figure 5b, middle). Accordingly, NPY release times were elevated compared with control (no fsk), and only marginally longer than with fsk alone (Figure 5d, right). Similar results were obtained in human β-cells, where glibenclamide increased exocytosis in the absence of fsk (p=0.01, n = 13 cells from four donors) but not in its presence (p=0.80, n = 7 cells from four donors; data not shown). The data indicate that sulfonylureas restrict fusion pore expansion through the same intracellular pathway as cAMP, which may counteract their stimulating effect on exocytosis by preventing or delaying peptide release.

Figure 5. Sulfonylureas cause fusion pore restriction.

(a) Cartoon of the experimental design in (b–d). INS-1 cells expressing NPY-tdmOrange2 were bathed in 10 mM glucose, diazoxide (200 µM) and either 200 µM tolbutamide (tolb), 50 µM glibenclamide (glib) or 50 µM gliclizide (gliz); exocytosis was evoked by acute exposure to 75 mM K+. (b) Exocytosis in absence (left) or presence (right) of fsk (2 µM) for flash events (color) and full fusions (white). Total exocytosis was increased by sulfonylurea in absence of fsk (p=0.15 tolb; p=0.05 glib, p=0.005 gliz, Kruskal Wallis/Dunn test vs ctrl/no fsk), but not in its presence (p=0.23 tolb; p=0.16 glib, p=0.10 glic). Sulfonylurea reduced full fusion events in presence of fsk (p=0.0045 tolb, p=0.00032 glib, 0.022 gliz, t-test). n of preps: 4 (CTR); 3 (tolb); 2 (glib); 3 (gliz); 5 (CTR + fsk); 3 (tolb + fsk); 3 (glib + fsk); 2 (gliz + fsk). n, number of cells. (c) Fraction of flash events for experiments in (b); Kruskal-Wallis/Dunn Test against ctrl/no fsk: p=0.015 tolb, p=0.001 glib, p=0.097 gliz, and against control +fsk: p=0.07 tolb; p=0.002 glib; p=0.14 gliz;); n, number of cells. (d) Cumulative frequency histograms and medians (insets) of NPY release times for (b–c). Differences vs control are significant in the absence of fsk: p=9.1*10−4 tolb, p=0.003 glib, p=0.015 gliz, Kolmogorov-Smirnov test). Insets show NPY release times for 38 (CTR), 74 (tolb), 79 (glib), 95 (gliz) events and inset on the right for 111 (CTR), 104 (tolb), 127 (glib) and 54 (gliz) events in presence of fsk. (e) Cartoon of the experimental design in (f–h). Cells were bathed in 10 mM glucose, 2 µM fsk, 50 µM diazoxide and acutely exposed to sulfonylureas (500 µM tolb, 100 µM glib or 100 µM gliz) during the recording period. (f) Exocytosis in presence of fsk (2 µM) for flash events (color) and full fusions (white). Differences are not significant (p=0.16 Kruskal Wallis test). n, number of cells. (g) Fraction of flash events for experiments in (f). Differences are not significant (p=0.98 Kruskal Wallis test). (h) Cumulative frequency histograms and medians (inset) of NPY release times for (f–g). Inset shows NPY release times for 111 (CTR), 68 (tolb), 34 (glib) and 31 (gliz) events.

Figure 5.

Figure 5—figure supplement 1. Granule pH is unchanged by forskolin or tolbutamide and does not affect pore lifetime.

Figure 5—figure supplement 1.

(a) Image sequences of a single NPY-EGFP-mCherry granule in an INS1 cells, exposed to 10 mM NH4+ (arrow); the green (top) and red color channels (bottom) of this ratiometric pH-probe are shown. (b) Green/red fluorescence ratio as measure of granule pH in controls (black), presence of fsk (green), tolbutamide (tlb, cyan), or both (blue); none of the values is significantly different from control; n, number of cells (20 granules each). (c) The fluorescence ratio for the same granules as in (b), after alkalization with 10 mM NH4+ was similar with fsk (p=0.06), tlb (p=0.06) or tlb + fsk (p=0.91). (d) K+-stimulated exocytosis of a single NPY-EGFP-mCherry granule; green (top) and red fluorescence (bottom) are shown. (e) Green/red ratio of granules as in (d), just prior to exocytosis, and separated for events with (flash) or without flash (FF). n, number of events. (f) As in (e), but for exocytosis events in presence of 10 mM NH4+. Values for flash and FF in e-f were not significantly different. n, number of events.
Figure 5—figure supplement 2. Activation of SUR1 by tolbutamide does not affect fusion pore restriction.

Figure 5—figure supplement 2.

(a) Image sequences of a granule undergoing K+-stimulated exocytosis in an INS-1 cell expressing NPY-tdmOrange2 and GFP-SUR1. (b) Quantification of GFP-SUR1 binding to the granule site (ΔF/S) in presence (green) or absence (black) of tolbutamide. (c) Exocytosis (40 s K+) in cells as in (a), separated for restricted fusion pores (flash events, in color) and full fusion events (in white); the decrease with tolbutamide was significant (p=0.001); n, number of cells. (d) Percentage of flash events in cells expressing EGFP-SUR, with or without tolbutamide.
Open in a new tab

Sulfonylureas also bind to SUR1 in the plasma membrane, which leads to rapid closure of KATP channels, depolarization and exocytosis. We tested the involvement of SUR1 by applying sulfonylureas acutely, which is expected to activate SUR1 in the plasma membrane but not Epac the cytosol (Figure 5e). Reduced diazoxide (50 µM) prevented glucose-dependent exocytosis but still allowed acute stimulation of exocytosis by sulfonylureas. Under these conditions, the fraction of flash-associated exocytosis events (Figure 5f–g) and the NPY release times (Figure 5h) were similar to control (stimulation with elevated K+) for all three sulfonylureas. Taken together, the data suggest that sulfonylureas must enter the cytosol to affect fusion pore behavior, and that this effect is not mediated by the plasma membrane SUR. We excluded the possibility that sulfonylureas affect the fluorescence signal indirectly, by altering granule pH (Figure 5—figure supplement 1). Moreover, an EGFP-tagged SUR1 (EGFP-SUR1) expressed in INS-1 cells did not localize to exocytosis sites or affect fusion pore behavior (Figure 5—figure supplement 2). We therefore conclude that sulfonylureas affect fusion pore behavior through Epac2.

Dynamin and amisyn-controlled restriction of the fusion pore is cAMP-dependent

The proteins dynamin and amisyn have previously been implicated in fusion pore regulation in β-cells (Tsuboi et al., 2004; Collins et al., 2016). To understand how these proteins behave around the release site, we expressed EGFP-tagged dynamin1 (Figure 6a) or mCherry-tagged amisyn (Figure 6b) together with a granule marker in INS-1 cells, and stimulated exocytosis with elevated K+. In the presence of fsk, both of the two fluorescent proteins were recruited to the granule site during membrane fusion (Figure 6c,f, and Figure 6—figure supplement 1). Expression of both proteins was about 2–4 fold compared with endogenous levels (Figure 6—figure supplement 2), and markedly increased the NPY release times (Figure 6d, g) and flash-associated exocytosis events (Figure 6e,h). Addition of the Epac inhibitor ESI09 prevented recruitment of both dynamin1 and amisyn during flash events and reduced flash events and NPY release times below control (Figure 6c–h). In the absence of fsk, expression of the two proteins had no effect on fusion pore behavior, and only amisyn (but not dynamin1) was recruited to the exocytosis site (Figure 6i–n). When Epac was activated with S223 (no fsk), dynamin1 and amisyn were recruited during flash events, and NPY release times and flash events were increased for both proteins (Figure 6i–n). The data suggest that dynamin1 and amisyn are acutely recruited to the exocytosis site, where they participate in cAMP-dependent fusion pore restriction.

Figure 6. Fusion pore regulation by dynamin1 and amisyn is cAMP-dependent.

(a–b) Example image sequence of transient recruitment of dynamin1-GFP (a, lower) or mCherry-amisyn (b, lower) to granules (upper, labeled with NPY-tdmOrange2 or NPY-EGFP) during K+-stimulated exocytosis in presence of forskolin. (c) Average time course (± SEM) of dynamin1-GFP (dyn) fluorescence during 34 flash-type exocytosis events (red) and eight full-fusion type events (black) in presence of forskolin; and nine flash events in presence of fsk + ESI09 (blue); data points represent average of five frames and time is relative to the flash onset in the granule signal. (d) Cumulative frequency histograms and medians (inset, with p for Kolmogorov-Smirnov test) of NPY release times in presence of fsk in cells expressing dynamin1-EGFP (red), dynamin with added ESI09 (blue) or control (black). 119 (CTR), 42 (dyn), 24 (dyn + ESI09) events. n of preps: 5 (C + fsk); 1 (dyn); 2 (dyn + ESI-09). (e) Fraction of flash events in (d). n, number of cells, p for Kruskal-Wallis/Dunns test. (f) Average time course (± SEM) of mCherry-amisyn (amis) fluorescence (red n = 274 flash events; black n = 46 full fusion events) or in presence of fsk + ESI09 (blue; n = 56 flash events). (g) Cumulative frequency histograms and medians (inset, with Kolmogorov-Smirnov test) of NPY release times in cells expressing mCherry-amisyn, amysin with ESI09, or control; fsk was present. 213 (CTR), 320 (amisyn), and 90 (amis +ESI09) events. n of preps: two for each. (h) Fraction of flash events in (g); p for Kruskal-Wallis/Dunn test. n, number of cells. (i) As in c, but without forskolin for control (black), dynamin (red), and dynamin with S223 (green); n = 37 flash events, n = 39 full fusion events for dyn and n = 40 flash events for dyn +S223. (j–k) As in (d–e), but for 38 (ctrl, black), 76 (dynamin1, red) and 55 (Dyn + S223, green) events in the absence of forskolin. n of preps: 4 (C-fsk); 2 (dyn); 2 (dyn + S223). (l) As in f, but without forskolin present; 65 flash events (red) and 73 full fusion events (black) for amisyn, and 154 flash events for amisyn + S223 (green). (m–n) As in (g–h), but for 123 (ctrl, black), 138 (amisyn, red) and 174 (amis + S223, green) events in the absence of forskolin. n of preps: 1 (C-fsk); 2 (amis); 2 (amis + S223).

Figure 6.

Figure 6—figure supplement 1. NPY and amisyn/dynamin1 recruitment profiles at the point of release.

Figure 6—figure supplement 1.

(a) Exocytosis events separated for flashes (left, green) and full fusions (right, gray) for dynamin1-GFP and NPY-tdmOrange2 expressing INS-1 cells in presence of forskolin from Figure 6. n, number of events (b) As in a, but for mCherry-amisyn and NPY EGFP expressing INS-1 cells from Figure 6. n, number of events (c) As in a, but in absence of forskolin. n, number of events (d) As in b, but in absence of forskolin. n, number of events.
Figure 6—figure supplement 2. Quantification of overexpression.

Figure 6—figure supplement 2.

Ins1-cells expressing mCherry-amisyn or dynamin1-GFP were fixated and immunostained using anti-amisyn or anti-dynamin1 and fluorescence was quantified for both labels by TIRFM of single cells. (a) Example images of immunostaining (upper) and mCherry-amisyn (lower). (b) Average fluorescence (cell-background) for immunostaining (white) and mCherry-amisyn (gray). (c) Plot of mCherry-amisyn vs immunostaining fluorescence; each symbol represents one cell. The offset at the y-axis corresponds to cells that only express endogenous amisyn. (d–e) as b-c but for dynamin1.
Open in a new tab

Discussion

cAMP-dependent signaling restricts fusion pore expansion and promotes kiss-and-run exocytosis in β-cells (Hanna et al., 2009) and neuroendocrine cells (Calejo et al., 2013; Machado et al., 2001) (but see Hatakeyama et al., 2006). We show here that the cAMP-mediator Epac2 orchestrates these effects by engaging dynamin and perhaps other endocytosis-related proteins at the release site (Figure 7). Since the fusion pore acts as a molecular sieve, the consequence is that insulin and other peptides remain trapped within the granule, while smaller transmitter molecules with para- or autocrine function are released (Obermüller et al., 2005; MacDonald et al., 2006; Taraska et al., 2003; Leclerc et al., 2004). Incretin signaling and Epac activation therefore delays, or altogether prevents insulin secretion from individual granules, while promoting paracrine intra-islet communication that is based mostly on release of small transmitter molecules.

Figure 7. Summary of fusion pore characteristics.

Figure 7.

Open in a new tab

Fraction of events with restricted fusion pores, NPY release time and exocytosis rate for Epac2 KO (first column), controls (second column) and with Epac2 overexpression (third column) in absence (upper rows) and presence of tolbutamide (bottom rows). Changes in exocytosis are compared to controls without (left half columns) or with (right half columns) forskolin. See Figure 7—source data 1 for details.

Figure 7—source data 1.
elife-41711-fig7-data1.xlsx (12.7KB, xlsx)
DOI: 10.7554/eLife.41711.015

Paradoxically, two clinically important classes of antidiabetic drugs, GLP-1 analogs and sulfonylureas, activate Epac in β-cells and caused restriction of the fusion pore. Sulfonylureas have long been known to stimulate insulin secretion by binding to SUR1, which results in closure of KATP channels and depolarization (Henquin, 2000). The drugs also accelerate PKA-independent granule priming in β-cells, which may involve activation of intracellularly localized SUR1 (Eliasson et al., 2003). Our data indicate that sulfonylureas exert a third mode of action that leads to the restriction of the fusion pore and therefore limits insulin release. Two pieces of evidence suggest that SUR1 is not involved in the latter. First, acute exposure to sulfonylureas had no effect on fusion pore behavior, although it blocks KATP channels (indicating SUR1 activation). Only long-term exposure to sulfonylurea resulted in restricted fusion pores, likely because it allowed the drugs to enter the cytoplasm. Second, we could not detect enrichment of SUR1 at the granule release site, which precludes any direct role of the protein in fusion pore regulation. Sulfonylurea compounds have been shown to allosterically stabilize the cAMP-dependent activation of Epac (Takahashi et al., 2013; Herbst et al., 2011). Our finding that sulfonylurea caused fusion pore restriction in the absence of forskolin indicates that basal cAMP concentrations are sufficient for this effect. Since gliclizide binds the CNB1 domain without activating it (Takahashi et al., 2013) and still restricts the fusion pore, Epac localization at the granule site (Alenkvist et al., 2017) may be enough to regulate the downstream proteins (e.g. dynamin and amisyn). It can further be speculated that the competing stimulatory (via exocytosis) and inhibitor effects (via the fusion pore) of sulfonylureas on insulin secretion, contribute to the reduction in sulfonylurea effectiveness with time of treatment. Long-term treatment with GLP-1 analogs disturbs glucose homeostasis (Abdulreda et al., 2016), and combination therapy of sulfonylurea and DPP4 inhibitors (that elevate cAMP) has been shown to lead to severe hypoglycemia (Yabe and Seino, 2014), an effect that likely depends on Epac (Takahashi et al., 2015).

Epac mediates the PKA-independent stimulation of exocytosis by cAMP (Seino et al., 2009) and our data suggests it may affect both priming and fusion pore restriction. This effect is rapid (Eliasson et al., 2003), suggesting that Epac is preassembled at the site of the secretory machinery. Indeed, Epac concentrates at sites of docked insulin granules (Alenkvist et al., 2017), and forms functionally relevant complexes with the tethering proteins Rim2 and Piccolo (Fujimoto et al., 2002). However, the amount of Epac2 present at individual release sites did not correlate with fusion pore behavior, which may indicate that the protein acts indirectly by activating or recruiting other proteins. Indeed, we show here that recruitment of two other proteins, dynamin and amisyn, depends on cAMP and Epac. Other known targets of Epac are the small GTPases Rap1 and R-Ras, for which Epac is a guanine nucleotide exchange factor (GEF). Rap1 is expressed on insulin granules and affects insulin secretion both directly (Shibasaki et al., 2007), and by promoting intracellular Ca2+-release following phospholipase-C activation (Dzhura et al., 2011). R-Ras is an activator of phosphoinositide 3-kinase (Marte et al., 1997). By altering local phosphoinositide levels, Epac could therefore indirectly affect exocytosis via recruitment of C2-domain proteins such as Munc13 (Kang et al., 2006), and fusion pore behavior by recruitment of the PH-domain containing proteins dynamin and amisyn (Ramachandran and Schmid, 2008; Abbineni et al., 2018).

An unresolved question is whether pore behavior is controlled by mechanisms that promote pore dilation, or that instead prevent it. Dynamin causes vesicle fission during clathrin-dependent endocytosis (Marks et al., 2001), and since dynamin is present at the exocytosis site and required for the kiss-and-run mode (Jackson et al., 2015; Tsuboi et al., 2004; Trexler et al., 2016), it may have a similar role during transient exocytosis. An active scission mechanism is also suggested by the finding that granules loose some of their membrane proteins during transient exocytosis (Tsuboi et al., 2004; Perrais et al., 2004). Capacitance measurements have shown that fusion pores initially flicker with conductances similar to those of large ion channels, before expanding irreversibly (Lollike et al., 1995). This could result from pores that are initially stabilized through unknown protein interactions and that eventually give way to uncontrolled expansion. However, scission mechanisms involving dynamin can act even when the pore has dilated considerably beyond limit of reversible flicker behavior (Shin et al., 2018; Taraska and Almers, 2004; Zhao et al., 2016; Anantharam et al., 2011), and even relatively large granules retain their size during fusion-fission cycles (MacDonald et al., 2006; Lollike et al., 1995). Separate mechanisms may therefore operate, one that prevents pore dilation by actively causing scission, similar to the role of dynamins in endocytosis, and another by shifting the equilibrium between the open and closed states of the initial fusion pore. Curvature-sensitive proteins are particularly attractive for such roles since they could accumulate at the neck of the fused granule; such ring-like assemblies that have indeed been observed for the Ca2+-sensor synaptotagmin (Wang et al., 2001). Active pore dilation has also been proposed to be driven by crowding of SNARE proteins (Wu et al., 2017) and α-synuclein (Logan et al., 2017).

β-cell granules contain a variety of polypeptides (insulin, IAPP, chromogranins) and small molecule transmitter molecules (GABA, nucleotides, 5HT) that have important para- and autocrine functions within the islet (Braun et al., 2012; Caicedo, 2013). Insulin modulates its own release by activating β-cell insulin receptors (Leibiger et al., 2008), stimulates somatostatin release (Vergari et al., 2019), and inhibits glucagon secretion (Ravier and Rutter, 2005). Insulin secretion is also inhibited by IAPP/amylin and chromogranin cleavage products such as pancreastatin (Braun et al., 2012). Of the small transmitters, GABA inhibits glucagon secretion from α-cells (Rorsman et al., 1989) and enhances insulin secretion (Soltani et al., 2011), and tonic GABA signaling is important for the maintenance of β-cell mass (Soltani et al., 2011). Adenine nucleotides cause β-cell depolarization, intracellular Ca2+-release and enhanced insulin secretion (Khan et al., 2014; Jacques-Silva et al., 2010), but also negative effects have been reported (Salehi et al., 2005; Poulsen et al., 1999). Paracrine purinergic effects also coordinate Ca2+ signaling among β-cells (Hellman et al., 2004), stimulate secretion of somatostatin from δ-cells (Bertrand et al., 1990), and target islet vasculature and macrophages as part of the immune system (Weitz et al., 2018). By selectively allowing small molecule release, Epac/cAMP-dependent fusion pore restriction is expected to alter both the timing and the relative volume of peptidergic vs. transmitter signaling. Given that granule priming and islet electrical activity are regulated on a second time scale, even small delays between these signals can be envisioned to affect the ratio of insulin to glucagon secretion. As illustrated by the recent finding of altered fusion pore behavior in type-2 diabetes (Collins et al., 2016), Epac-dependent fusion pore regulation may have profound consequences for islet physiology and glucose metabolism in vivo.

Materials and methods

Key resources table.

Reagent or Resource Designation Source of Reference Identifiers Additional Information
Strain, strain background (Adenovirus) NPY-Venus P Rorsman (Oxford)
Strain, strain background (Adenovirus) NPY-tdmOrange2 this paper See Constructs in Materials and methods
Genetic reagent (Mus musculus) Rapgef4 KO and WT (Kopperud et al., 2017)
Cell line (Rattus norvegicus domesticus) INS-1 Clone 832/12 (Hohmeier et al., 2000) RRID:CVCL_7226 H Mulder (Malmö)
Transfected construct (Mus musculus) EGFP-Epac2 (Idevall-Hagren et al., 2013) 1068
Transfected construct (Homo sapiens) NPY-tdmOrange2 (Gandasi et al., 2015) 1140
Transfected construct (Rattus norvegicus) P2X2-mRFP1 (Obermüller et al., 2005) 1226
Transfected construct (Homo sapiens) NPY EGFP mCherry this paper See Constructs in Materials and methods
Transfected construct (Homo sapiens) Cherry2-amisyn This paper NM_001351940.1; 1286 See Constructs in Materials and methods
Transfected construct (Homo sapiens) dynamin1-GFP W Almers (Portland) 1342
Transfected construct (Homo sapiens) NPY EGFP W Almers (Portland) 1008
Biological sample (Homo sapiens) Human pancreatic islets (Goto et al., 2004) Nordic Network for Clinical Islet Transplantation Uppsala
Antibody Rabbit polyclonal anti-amisyn ab153974 abcam 1/50
Antibody Rabbit monoclonal anti-dynamin1 ab52852 abcam PMID:28171750 1/50
Chemical compound, drug Cell dissociation buffer Thermo Fisher 13150016
Chemical compound, drug Trypsin solution Thermo Fisher 12604–021
Chemical compound, drug Lipofectamine 2000 Thermo Fisher 11668–019
Chemical compound, drug Forskolin; Fsk Sigma-Aldrich F6886
Chemical compound, drug Polylysine Sigma-Aldrich P5899
Chemical compound, drug Exendin-4; Ex4 Anaspec (Fremont CA) AS-24463
Chemical compound, drug Diazoxide Sigma-Aldrich D9035
Chemical compound, drug BSA Sigma-Aldrich F0804
Chemical compound, drug RPMI 1640 SVA 992680
Chemical compound, drug L-Glutamine Hyclone SH30034.01
Chemical compound, drug Tolbutamide; tolb Sigma-Aldrich 64-77-7
Chemical compound, drug Glibenclamide; glib Hoechst
Chemical compound, drug Gliclizide; gliz Sigma-Aldrich 21187-98-4
Chemical compound, drug S223 Biolog B 056–01
Software, algorithm MetaMorph Molecular Devices
Open in a new tab

Cells

Human islets were obtained from the Nordic Network for Clinical Islet Transplantation Uppsala (Goto et al., 2004) under full ethical clearance (Uppsala Regional Ethics Board 2006/348) and with written informed consent. Isolated islets were cultured free-floating in sterile dishes in CMRL 1066 culture medium containing 5.5 mM glucose, 10% fetal calf serum, 2 mM L-glutamine, streptomycin (100 U/ml), and penicillin (100 U/ml) at 37°C in an atmosphere of 5% CO2 up to 2 weeks. Prior to imaging, islets were dispersed into single cells by gentle agitation using Ca2+-free cell dissociation buffer (Thermo Fisher Scientific) supplemented with 10% (v/v) trypsin (0.05% Thermo Fisher Scientific). INS1-cells clone 832/13 (Hohmeier et al., 2000) were maintained in RPMI 1640 (Invitrogen) with 10 mM glucose, 10% fetal bovine serum, streptomycin (100 U/ml), penicillin (100 U/ml), Sodium pyruvate (1 mM), and 2-mercaptoethanol (50 μM). The ins1 832/13 cells were screened by PCR and found negative for mycoplasma.

Mouse islets were isolated from 5 to 12 months old WT and Epac2-/- (Kopperud et al., 2017) (Rapgef4-/-) animals. The Epac2 deletion involves exons 12–13, which include the high-affinity cAMP binding domain present in all Epac2 isoforms, in contrast to previously reported knockout strain (Shibasaki et al., 2007), which only lacks the Epac2A isoform. The mice were anesthetized and the pancreas dissected out and cleared from fat and connective tissue in ice-cold Ca5 solution (in mM 125 NaCl, 5KCl, 1.2 MgCl2, 1.28 CaCl2, 10 HEPES; pH 7.4 with NaOH). Pancreas was injected with Collagenase P (1 mg/ml) and cut into small pieces before mechanical dissociation (7 min at 37°C). BSA was added immediately and islets were washed 3X with ice cold Ca5 buffer with BSA. Islets were dispersed into single cells using Ca2+-free cell dissociation buffer (supplemented with 10% (v/v) trypsin) and gentle agitation. Dispersed cells were sedimented by centrifugation, resuspended in RPMI 1640 medium (containing 5.5 mM glucose, 10% fetal calf serum, 100 U/ml penicillin and 100 U/ml streptomycin).

The cells were plated onto 22 mm polylysine-coated coverslips and were transduced the next day using adenovirus (human and mouse cells) or transfected the same day with plasmids (INS1 cells, using Lipofectamine2000, Invitrogen) encoding the granule markers NPY-Venus, NPY-EGFP or NPY-tdOrange. Imaging proceeded 24–36 hr later.

Constructs

The open-reading frame of human amisyn (NM_001351940.1) was obtained as a synthetic DNA fragment (Eurofins, Germany) and was cloned into pCherry2 C1 (Addgene, plasmid nr 54563) by seamless PCR cloning. The linker between Cherry2 and amisyn translates into the peptide SGLRSRAQASNSAV. The plasmid N1 NPY-EGFP-mCherry coding for NPY-linker(TVPRARDPPVAT)-EGFP-linker(KRSGGSGGSGGS)-mCherry was made by seamless PCR cloning. The correct open-reading frame of both Cherry2-linker-amisyn and NPY-EGFP-mCherry was confirmed by Sanger sequencing (Eurofins, Germany). The NPY-tdOrange2 adeno virus was made using the RAPAd vector system (Cell Biolabs, San Diego). NPY-tdOrange2 (Gandasi et al., 2015) was cloned into the pacAd5 CMVK-NpA Shuttle plasmid (Cell Biolabs). Virus was produced in HEK293 cells and isolated according to the instructions of the manufacturer (Cell Biolabs).

Solutions

Cells were imaged in (mM) 138 NaCl, 5.6 KCl, 1.2 MgCl2, 2.6 CaCl2, 10 D-glucose 5 HEPES (pH 7.4 with NaOH) at 32–34°C. Exocytosis was evoked with high 75 mM K+ (equimolarly replacing Na+), applied by computer-timed local pressure ejection through a pulled glass capillary. For K+-induced exocytosis, spontaneous depolarizations were prevented with 200 µM diazoxide (50 µM for Figure 5e–h). In Figure 5e–h, exocytosis was evoked by sulfonylureas (500 µM tolbutamide, 200 µM glibenclamide or 200 µM gliclizide). For electrophysiology, glucose was reduced to 3 mM, and the electrodes were filled with (mM) 125 CsCl, 10 NaCl, 1.2 MgCl2, 5 EGTA, 4 CaCl2, 3 Mg-ATP, 0.1 cAMP, 10 HEPES (pH 7.15 using CsOH).

Immunocytochemistry

To quantify the overexpression, INS-1 cell were transfected with either Cherry2-amisyn or Dynamin1-GFP, fixed 24 hr later in 3.8% formaldehyde in phosphate-buffered saline (PBS) for 30 min at 25°C and washed in PBS. The cells were permeabilized in 0.2% Triton X-100 in PBS for 5 min and washed in PBS. Blocking was done using 5% FBS in PBS for 1–2 hr at 25°C. Cells were then incubated with a primary antibody (anti-Dynamin1, ab52852 abcam or anti-Amisyn, ab153974 abcam) both diluted 1/50 in 5% FCS in PBS over night at 4°C and washed again in PBS. Incubation with secondary antibody (Alexa Fluor 488 anti-rabbit or Alexa Fluor 555 anti-rabbit, Invitrogen) diluted 1/1000 in 5% FCS in PBS was performed for 1 hr at 25°C and subsequently the cells were washed in PBS.

TIRF microscopy

Human cells were imaged using a lens-type total internal reflection (TIRF) microscope, based on an AxioObserver Z1 with a 100x/1.45 objective (Carl Zeiss). TIRF illumination with a calculated decay constant of ~100 nm was created using two DPSS lasers at 491 and 561 nm (Cobolt, Stockholm, Sweden) that passed through a cleanup filter (zet405/488/561/640x, Chroma) and was controlled with an acousto-optical tunable filter (AA-Opto, France). Excitation and emission light were separated using a beamsplitter (ZT405/488/561/640rpc, Chroma) and the emission light chromatically separated (QuadView, Roper) onto separate areas of an EMCCD camera (QuantEM 512SC, Roper) with a cutoff at 565 nm (565dcxr, Chroma) and emission filters (ET525/50 m and 600/50 m, Chroma). Scaling was 160 nm per pixel.

INS1 and mouse cells were imaged using a custom-built lens-type TIRF microscope based on an AxioObserver D1 microscope and a 100x/1.45 NA objective (Carl Zeiss). Excitation was from two DPSS lasers at 473 nm and 561 nm (Cobolt), controlled with an acousto-optical tunable filter (AOTF, AA-Opto) and using dichroic Di01-R488/561 (Semrock). The emission light was separated onto the two halves of a 16-bit EMCCD camera (Roper Cascade 512B, gain setting at 3800 a.u. throughout) using an image splitter (DualView, Photometrics) with ET525/50 m and 600/50 m emission filters (Chroma). Scaling was 100 nm per pixel for INS-1 experiments and 160 nm for mouse cells. The frame rate was 10 frames*s−1, with 100 ms exposures.

Image analysis

Exocytosis events were identified manually based on the characteristic rapid loss of the granule marker fluorescence (most fluorescence lost within 1–2 frames) in cells which exhibited minimum of 1 event/cell (except mouse cells, where all cells were included). Events were classified as flash events if they exhibited an increase in the fluorescence signal before the rapid loss of the granule fluorescence. The NPY release times were obtained for both types of events by non-linear fitting with a discontinuous function in Origin as described previously (Gandasi et al., 2015). Protein binding to the release site (ΔF/S) was measured as described previously (Gandasi and Barg, 2014).

Electrophysiology

ATP release was measured in INS1 cells expressing RFP-tagged P2X2 receptor (Obermüller et al., 2005). Cells were voltage-clamped in whole-cell mode using an EPC-9 amplifier and PatchMaster software (Heka Elektronik, Lambrecht, Germany) with patch-clamp electrodes pulled from borosilicate glass capillaries that were coated with Sylgard close to the tips, and fire-polished (resistance 2–4 MΩ). The free [Ca2+] was calculated to be 600 nM (WEBMAXC standard) and elicited exocytosis that was detected as P2X2-dependent inward current spikes. Currents were filtered at 1 kHz and sampled at 5 kHz. Spike analysis was performed using automated program for amperometric recordings in IGOR Pro (Segura et al., 2000), with the threshold set at eight times the RMS noise during event-free section of recording.

Statistics

Data are presented as mean ± SEM unless otherwise stated. Statistical significance was tested (unless otherwise stated) and is indicated by asterisks (*p<0.05, **p<0.01, ***p<0.001). The not normally distributed exocytosis rates and ratios of flash events were tested with Kruskal Wallis with post hoc Dunn test and NPY release times were tested with Kolmogorov-Smirnov test.

Acknowledgements

We thank J Saras, P-E Lund, Y Xu and A Thonig (Uppsala University) for expert technical assistance, and D Machado (University of La Laguna) for spike analysis software.

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Contributor Information

Sebastian Barg, Email: sebastian.barg@mcb.uu.se.

Axel T Brunger, Stanford University, United States.

Vivek Malhotra, The Barcelona Institute of Science and Technology, Spain.

Funding Information

This paper was supported by the following grants:

  • Family Ernfors Foundation to Alenka Guček, Anders Tengholm, Sebastian Barg.

  • Uppsala Universitet Olga Jönssons stipend to Alenka Guček.

  • P O Zetterlingsstiftelse to Alenka Guček.

  • European Foundation for the Study of Diabetes to Nikhil R Gandasi, Anders Tengholm, Sebastian Barg.

  • Swedish Society for Medical Research to Nikhil R Gandasi.

  • Novo Nordisk to Nikhil R Gandasi, Anders Tengholm, Sebastian Barg.

  • Norwegian Research Council to Marit Bakke.

  • Helse-Bergen to Marit Bakke.

  • Swedish Research Council 2014-02575 to Anders Tengholm, Sebastian Barg.

  • Diabetes Wellness Network Sweden to Anders Tengholm, Sebastian Barg.

  • Swedish Diabetes Society to Anders Tengholm, Sebastian Barg.

  • Exodiab network to Anders Tengholm, Sebastian Barg.

  • Swedish Research Council 2017-00956 to Anders Tengholm, Sebastian Barg.

  • Swedish Research Council 2018-02871 to Anders Tengholm, Sebastian Barg.

  • Hjärnfonden to Sebastian Barg.

Additional information

Competing interests

No competing interests declared.

Author contributions

Conceptualization, Data curation, Formal analysis, Validation, Investigation, Visualization, Writing—original draft, Writing—review and editing.

Conceptualization, Formal analysis, Investigation, Writing—review and editing.

Methodology, Writing—review and editing.

Resources, Funding acquisition.

Resources, Funding acquisition, Writing—review and editing.

Resources, Funding acquisition, Writing—review and editing.

Conceptualization, Resources, Data curation, Supervision, Funding acquisition, Visualization, Methodology, Project administration, Writing—review and editing.

Ethics

Animal experimentation: This study was performed in strict accordance with European and Swedish legislation, fundamental ethical principles and approved by the Regional Ethics Board Uppsala (license number 31, 1-32).

Additional files

Transparent reporting form
elife-41711-transrepform.pdf (474.4KB, pdf)
DOI: 10.7554/eLife.41711.016

Data availability

Source data file has been provided for Figure 7. All raw data are available on the Dryad Digital Repository (http://dx.doi.org/10.5061/dryad.6b604g8).

The following dataset was generated:

Gucek A, Gandasi NR, Omar-Hmeadi M, Bakke M, Døskeland SO, Tengholm A, Barg S. 2019. Data from: Fusion pore regulation by cAMP/Epac2 controls cargo release during insulin exocytosis. Dryad Digital Repository.

References

  1. Abbineni PS, Axelrod D, Holz RW. Visualization of expanding fusion pores in secretory cells. The Journal of General Physiology. 2018;150:1640–1646. doi: 10.1085/jgp.201812186. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Abdulreda MH, Rodriguez-Diaz R, Caicedo A, Berggren PO. Liraglutide compromises pancreatic β cell function in a humanized mouse model. Cell Metabolism. 2016;23:541–546. doi: 10.1016/j.cmet.2016.01.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Albillos A, Dernick G, Horstmann H, Almers W, Alvarez de Toledo G, Lindau M. The exocytotic event in chromaffin cells revealed by patch amperometry. Nature. 1997;389:509–512. doi: 10.1038/39081. [DOI] [PubMed] [Google Scholar]
  4. Alenkvist I, Gandasi NR, Barg S, Tengholm A. Recruitment of Epac2A to insulin granule docking sites regulates priming for exocytosis. Diabetes. 2017;66:2610–2622. doi: 10.2337/db17-0050. [DOI] [PubMed] [Google Scholar]
  5. Alés E, Tabares L, Poyato JM, Valero V, Lindau M, Alvarez de Toledo G. High calcium concentrations shift the mode of exocytosis to the kiss-and-run mechanism. Nature Cell Biology. 1999;1:40–44. doi: 10.1038/9012. [DOI] [PubMed] [Google Scholar]
  6. Alvarez de Toledo G, Fernández-Chacón R, Fernández JM. Release of secretory products during transient vesicle fusion. Nature. 1993;363:554–558. doi: 10.1038/363554a0. [DOI] [PubMed] [Google Scholar]
  7. Anantharam A, Onoa B, Edwards RH, Holz RW, Axelrod D. Localized topological changes of the plasma membrane upon exocytosis visualized by polarized TIRFM. The Journal of Cell Biology. 2010;188:415–428. doi: 10.1083/jcb.200908010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Anantharam A, Bittner MA, Aikman RL, Stuenkel EL, Schmid SL, Axelrod D, Holz RW. A new role for the dynamin GTPase in the regulation of fusion pore expansion. Molecular Biology of the Cell. 2011;22:1907–1918. doi: 10.1091/mbc.e11-02-0101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Aoki R, Kitaguchi T, Oya M, Yanagihara Y, Sato M, Miyawaki A, Tsuboi T. Duration of fusion pore opening and the amount of hormone released are regulated by myosin II during kiss-and-run exocytosis. Biochemical Journal. 2010;429:497–504. doi: 10.1042/BJ20091839. [DOI] [PubMed] [Google Scholar]
  10. Artalejo CR, Henley JR, McNiven MA, Palfrey HC. Rapid endocytosis coupled to exocytosis in adrenal chromaffin cells involves Ca2+, GTP, and dynamin but not clathrin. PNAS. 1995;92:8328–8332. doi: 10.1073/pnas.92.18.8328. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Bao H, Goldschen-Ohm M, Jeggle P, Chanda B, Edwardson JM, Chapman ER. Exocytotic fusion pores are composed of both lipids and proteins. Nature Structural & Molecular Biology. 2016;23:67–73. doi: 10.1038/nsmb.3141. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Barg S, Renström E, Berggren PO, Bertorello A, Bokvist K, Braun M, Eliasson L, Holmes WE, Köhler M, Rorsman P, Thévenod F. The stimulatory action of tolbutamide on Ca2+-dependent exocytosis in pancreatic beta cells is mediated by a 65-kDa mdr-like P-glycoprotein. PNAS. 1999;96:5539–5544. doi: 10.1073/pnas.96.10.5539. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Barg S, Olofsson CS, Schriever-Abeln J, Wendt A, Gebre-Medhin S, Renström E, Rorsman P. Delay between fusion pore opening and peptide release from large dense-core vesicles in neuroendocrine cells. Neuron. 2002;33:287–299. doi: 10.1016/S0896-6273(02)00563-9. [DOI] [PubMed] [Google Scholar]
  14. Ben-Othman N, Vieira A, Courtney M, Record F, Gjernes E, Avolio F, Hadzic B, Druelle N, Napolitano T, Navarro-Sanz S, Silvano S, Al-Hasani K, Pfeifer A, Lacas-Gervais S, Leuckx G, Marroquí L, Thévenet J, Madsen OD, Eizirik DL, Heimberg H, Kerr-Conte J, Pattou F, Mansouri A, Collombat P. Long-Term GABA administration induces alpha Cell-Mediated Beta-like cell neogenesis. Cell. 2017;168:73–85. doi: 10.1016/j.cell.2016.11.002. [DOI] [PubMed] [Google Scholar]
  15. Bertrand G, Gross R, Ribes G, Loubatières-Mariani MM. P2 purinoceptor agonists stimulate somatostatin secretion from dog pancreas. European Journal of Pharmacology. 1990;182:369–373. doi: 10.1016/0014-2999(90)90296-I. [DOI] [PubMed] [Google Scholar]
  16. Braun M, Ramracheya R, Rorsman P. Autocrine regulation of insulin secretion. Diabetes, Obesity and Metabolism. 2012;14:143–151. doi: 10.1111/j.1463-1326.2012.01642.x. [DOI] [PubMed] [Google Scholar]
  17. Breckenridge LJ, Almers W. Currents through the fusion pore that forms during exocytosis of a secretory vesicle. Nature. 1987;328:814–817. doi: 10.1038/328814a0. [DOI] [PubMed] [Google Scholar]
  18. Cabrera O, Jacques-Silva MC, Speier S, Yang SN, Köhler M, Fachado A, Vieira E, Zierath JR, Kibbey R, Berman DM, Kenyon NS, Ricordi C, Caicedo A, Berggren PO. Glutamate is a positive autocrine signal for glucagon release. Cell Metabolism. 2008;7:545–554. doi: 10.1016/j.cmet.2008.03.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Caicedo A. Paracrine and autocrine interactions in the human islet: more than meets the eye. Seminars in Cell & Developmental Biology. 2013;24:11–21. doi: 10.1016/j.semcdb.2012.09.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Calejo AI, Jorgacevski J, Kucka M, Kreft M, Gonçalves PP, Stojilkovic SS, Zorec R. cAMP-mediated stabilization of fusion pores in cultured rat pituitary lactotrophs. Journal of Neuroscience. 2013;33:8068–8078. doi: 10.1523/JNEUROSCI.5351-12.2013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Chiang HC, Shin W, Zhao WD, Hamid E, Sheng J, Baydyuk M, Wen PJ, Jin A, Momboisse F, Wu LG. Post-fusion structural changes and their roles in exocytosis and endocytosis of dense-core vesicles. Nature Communications. 2014;5:3356. doi: 10.1038/ncomms4356. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Collins SC, Do HW, Hastoy B, Hugill A, Adam J, Chibalina MV, Galvanovskis J, Godazgar M, Lee S, Goldsworthy M, Salehi A, Tarasov AI, Rosengren AH, Cox R, Rorsman P. Increased expression of the diabetes gene SOX4 reduces insulin secretion by impaired fusion pore expansion. Diabetes. 2016;65:1952–1961. doi: 10.2337/db15-1489. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Dyachok O, Idevall-Hagren O, Sågetorp J, Tian G, Wuttke A, Arrieumerlou C, Akusjärvi G, Gylfe E, Tengholm A. Glucose-induced cyclic AMP oscillations regulate pulsatile insulin secretion. Cell Metabolism. 2008;8:26–37. doi: 10.1016/j.cmet.2008.06.003. [DOI] [PubMed] [Google Scholar]
  24. Dzhura I, Chepurny OG, Leech CA, Roe MW, Dzhura E, Xu X, Lu Y, Schwede F, Genieser HG, Smrcka AV, Holz GG. Phospholipase C-ε links Epac2 activation to the potentiation of glucose-stimulated insulin secretion from mouse islets of langerhans. Islets. 2011;3:121–128. doi: 10.4161/isl.3.3.15507. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Eliasson L, Ma X, Renström E, Barg S, Berggren PO, Galvanovskis J, Gromada J, Jing X, Lundquist I, Salehi A, Sewing S, Rorsman P. SUR1 regulates PKA-independent cAMP-induced granule priming in mouse pancreatic B-cells. The Journal of General Physiology. 2003;121:181–197. doi: 10.1085/jgp.20028707. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Ferraro F, Eipper BA, Mains RE. Retrieval and reuse of pituitary secretory granule proteins. Journal of Biological Chemistry. 2005;280:25424–25435. doi: 10.1074/jbc.M414156200. [DOI] [PubMed] [Google Scholar]
  27. Fujimoto K, Shibasaki T, Yokoi N, Kashima Y, Matsumoto M, Sasaki T, Tajima N, Iwanaga T, Seino S. Piccolo, a Ca2+ sensor in pancreatic beta-cells. involvement of cAMP-GEFII.Rim2. piccolo complex in cAMP-dependent exocytosis. The Journal of Biological Chemistry. 2002;277:50497–50502. doi: 10.1074/jbc.M210146200. [DOI] [PubMed] [Google Scholar]
  28. Gandasi NR, Vestö K, Helou M, Yin P, Saras J, Barg S. Survey of red fluorescence proteins as markers for secretory granule exocytosis. PLOS ONE. 2015;10:e0127801. doi: 10.1371/journal.pone.0127801. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Gandasi NR, Barg S. Contact-induced clustering of syntaxin and munc18 docks secretory granules at the exocytosis site. Nature Communications. 2014;5:3914. doi: 10.1038/ncomms4914. [DOI] [PubMed] [Google Scholar]
  30. Gjertsen BT, Mellgren G, Otten A, Maronde E, Genieser HG, Jastorff B, Vintermyr OK, McKnight GS, Døskeland SO. Novel (Rp)-cAMPS analogs as tools for inhibition of cAMP-kinase in cell culture. Basal cAMP-kinase activity modulates interleukin-1 beta action. The Journal of Biological Chemistry. 1995;270:20599–20607. doi: 10.1074/jbc.270.35.20599. [DOI] [PubMed] [Google Scholar]
  31. Goto M, Eich TM, Felldin M, Foss A, Källen R, Salmela K, Tibell A, Tufveson G, Fujimori K, Engkvist M, Korsgren O. Refinement of the automated method for human islet isolation and presentation of a closed system for in vitro islet culture. Transplantation. 2004;78:1367–1375. doi: 10.1097/01.TP.0000140882.53773.DC. [DOI] [PubMed] [Google Scholar]
  32. Graham ME, O'Callaghan DW, McMahon HT, Burgoyne RD. Dynamin-dependent and dynamin-independent processes contribute to the regulation of single vesicle release kinetics and quantal size. PNAS. 2002;99:7124–7129. doi: 10.1073/pnas.102645099. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Hanna ST, Pigeau GM, Galvanovskis J, Clark A, Rorsman P, MacDonald PE. Kiss-and-run exocytosis and fusion pores of secretory vesicles in human beta-cells. Pflügers Archiv - European Journal of Physiology. 2009;457:1343–1350. doi: 10.1007/s00424-008-0588-0. [DOI] [PubMed] [Google Scholar]
  34. Hatakeyama H, Kishimoto T, Nemoto T, Kasai H, Takahashi N. Rapid glucose sensing by protein kinase A for insulin exocytosis in mouse pancreatic islets. The Journal of Physiology. 2006;570:271–282. doi: 10.1113/jphysiol.2005.096560. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Hellman B, Dansk H, Grapengiesser E. Pancreatic beta-cells communicate via intermittent release of ATP. American Journal of Physiology-Endocrinology and Metabolism. 2004;286:E759–E765. doi: 10.1152/ajpendo.00452.2003. [DOI] [PubMed] [Google Scholar]
  36. Henquin JC. Triggering and amplifying pathways of regulation of insulin secretion by glucose. Diabetes. 2000;49:1751–1760. doi: 10.2337/diabetes.49.11.1751. [DOI] [PubMed] [Google Scholar]
  37. Herbst KJ, Coltharp C, Amzel LM, Zhang J. Direct activation of epac by sulfonylurea is isoform selective. Chemistry & Biology. 2011;18:243–251. doi: 10.1016/j.chembiol.2010.12.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Hohmeier HE, Mulder H, Chen G, Henkel-Rieger R, Prentki M, Newgard CB. Isolation of INS-1-derived cell lines with robust ATP-sensitive K+ channel-dependent and -independent glucose-stimulated insulin secretion. Diabetes. 2000;49:424–430. doi: 10.2337/diabetes.49.3.424. [DOI] [PubMed] [Google Scholar]
  39. Huang X, Fan J, Li L, Liu H, Wu R, Wu Y, Wei L, Mao H, Lal A, Xi P, Tang L, Zhang Y, Liu Y, Tan S, Chen L. Fast, long-term, super-resolution imaging with hessian structured illumination microscopy. Nature Biotechnology. 2018;36:451–459. doi: 10.1038/nbt.4115. [DOI] [PubMed] [Google Scholar]
  40. Idevall-Hagren O, Jakobsson I, Xu Y, Tengholm A. Spatial control of Epac2 activity by cAMP and Ca2+-mediated activation of ras in pancreatic β cells. Science Signaling. 2013;6:ra29. doi: 10.1126/scisignal.2003932. [DOI] [PubMed] [Google Scholar]
  41. Jackson J, Papadopulos A, Meunier FA, McCluskey A, Robinson PJ, Keating DJ. Small molecules demonstrate the role of dynamin as a bi-directional regulator of the exocytosis fusion pore and vesicle release. Molecular Psychiatry. 2015;20:810–819. doi: 10.1038/mp.2015.56. [DOI] [PubMed] [Google Scholar]
  42. Jacques-Silva MC, Correa-Medina M, Cabrera O, Rodriguez-Diaz R, Makeeva N, Fachado A, Diez J, Berman DM, Kenyon NS, Ricordi C, Pileggi A, Molano RD, Berggren PO, Caicedo A. ATP-gated P2X3 receptors constitute a positive autocrine signal for insulin release in the human pancreatic beta cell. PNAS. 2010;107:6465–6470. doi: 10.1073/pnas.0908935107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Jin Y, Korol SV, Jin Z, Barg S, Birnir B. In intact islets interstitial GABA activates GABA(A) receptors that generate tonic currents in α-cells. PLOS ONE. 2013;8:e67228. doi: 10.1371/journal.pone.0067228. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Kang L, He Z, Xu P, Fan J, Betz A, Brose N, Xu T. Munc13-1 is required for the sustained release of insulin from pancreatic beta cells. Cell Metabolism. 2006;3:463–468. doi: 10.1016/j.cmet.2006.04.012. [DOI] [PubMed] [Google Scholar]
  45. Kawasaki H, Springett GM, Mochizuki N, Toki S, Nakaya M, Matsuda M, Housman DE, Graybiel AM. A family of cAMP-binding proteins that directly activate Rap1. Science. 1998;282:2275–2279. doi: 10.1126/science.282.5397.2275. [DOI] [PubMed] [Google Scholar]
  46. Khan S, Yan-Do R, Duong E, Wu X, Bautista A, Cheley S, MacDonald PE, Braun M. Autocrine activation of P2Y1 receptors couples ca (2+) influx to ca (2+) release in human pancreatic beta cells. Diabetologia. 2014;57:2535–2545. doi: 10.1007/s00125-014-3368-8. [DOI] [PubMed] [Google Scholar]
  47. Kopperud RK, Rygh CB, Karlsen TV, Krakstad C, Kleppe R, Hoivik EA, Bakke M, Tenstad O, Selheim F, Lidén Å., Madsen L, Pavlin T, Taxt T, Kristiansen K, Curry F-RE, Reed RK, Døskeland SO. Increased microvascular permeability in mice lacking Epac1 (Rapgef3) Acta Physiologica. 2017;219:441–452. doi: 10.1111/apha.12697. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Leclerc I, Woltersdorf WW, da Silva Xavier G, Rowe RL, Cross SE, Korbutt GS, Rajotte RV, Smith R, Rutter GA. Metformin, but not leptin, regulates AMP-activated protein kinase in pancreatic islets: impact on glucose-stimulated insulin secretion. American Journal of Physiology-Endocrinology and Metabolism. 2004;286:E1023–E1031. doi: 10.1152/ajpendo.00532.2003. [DOI] [PubMed] [Google Scholar]
  49. Leibiger IB, Leibiger B, Berggren PO. Insulin signaling in the pancreatic beta-cell. Annual Review of Nutrition. 2008;28:233–251. doi: 10.1146/annurev.nutr.28.061807.155530. [DOI] [PubMed] [Google Scholar]
  50. Logan T, Bendor J, Toupin C, Thorn K, Edwards RH. α-Synuclein promotes dilation of the exocytotic fusion pore. Nature Neuroscience. 2017;20:681–689. doi: 10.1038/nn.4529. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Lollike K, Borregaard N, Lindau M. The exocytotic fusion pore of small granules has a conductance similar to an ion channel. The Journal of Cell Biology. 1995;129:99–104. doi: 10.1083/jcb.129.1.99. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. MacDonald PE, Braun M, Galvanovskis J, Rorsman P. Release of small transmitters through kiss-and-run fusion pores in rat pancreatic beta cells. Cell Metabolism. 2006;4:283–290. doi: 10.1016/j.cmet.2006.08.011. [DOI] [PubMed] [Google Scholar]
  53. Machado JD, Morales A, Gomez JF, Borges R. cAmp modulates exocytotic kinetics and increases quantal size in chromaffin cells. Molecular Pharmacology. 2001;60:514–520. [PubMed] [Google Scholar]
  54. Marks B, Stowell MH, Vallis Y, Mills IG, Gibson A, Hopkins CR, McMahon HT. GTPase activity of dynamin and resulting conformation change are essential for endocytosis. Nature. 2001;410:231–235. doi: 10.1038/35065645. [DOI] [PubMed] [Google Scholar]
  55. Marte BM, Rodriguez-Viciana P, Wennström S, Warne PH, Downward J. R-Ras can activate the phosphoinositide 3-kinase but not the MAP kinase arm of the ras effector pathways. Current Biology. 1997;7:63–71. doi: 10.1016/S0960-9822(06)00028-5. [DOI] [PubMed] [Google Scholar]
  56. Obermüller S, Lindqvist A, Karanauskaite J, Galvanovskis J, Rorsman P, Barg S. Selective nucleotide-release from dense-core granules in insulin-secreting cells. Journal of Cell Science. 2005;118:4271–4282. doi: 10.1242/jcs.02549. [DOI] [PubMed] [Google Scholar]
  57. Ozaki N, Shibasaki T, Kashima Y, Miki T, Takahashi K, Ueno H, Sunaga Y, Yano H, Matsuura Y, Iwanaga T, Takai Y, Seino S. cAMP-GEFII is a direct target of cAMP in regulated exocytosis. Nature Cell Biology. 2000;2:805–811. doi: 10.1038/35041046. [DOI] [PubMed] [Google Scholar]
  58. Perrais D, Kleppe IC, Taraska JW, Almers W. Recapture after exocytosis causes differential retention of protein in granules of bovine chromaffin cells. The Journal of Physiology. 2004;560:413–428. doi: 10.1113/jphysiol.2004.064410. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Poulsen CR, Bokvist K, Olsen HL, Høy M, Capito K, Gilon P, Gromada J. Multiple sites of purinergic control of insulin secretion in mouse pancreatic beta-cells. Diabetes. 1999;48:2171–2181. doi: 10.2337/diabetes.48.11.2171. [DOI] [PubMed] [Google Scholar]
  60. Ramachandran R, Schmid SL. Real-time detection reveals that effectors couple Dynamin's GTP-dependent conformational changes to the membrane. The EMBO Journal. 2008;27:27–37. doi: 10.1038/sj.emboj.7601961. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Ravier MA, Rutter GA. Glucose or insulin, but not zinc ions, inhibit glucagon secretion from mouse pancreatic alpha-cells. Diabetes. 2005;54:1789–1797. doi: 10.2337/diabetes.54.6.1789. [DOI] [PubMed] [Google Scholar]
  62. Richards-Williams C, Contreras JL, Berecek KH, Schwiebert EM. Extracellular ATP and zinc are co-secreted with insulin and activate multiple P2X purinergic receptor channels expressed by islet beta-cells to potentiate insulin secretion. Purinergic Signalling. 2008;4:393–405. doi: 10.1007/s11302-008-9126-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Rorsman P, Berggren PO, Bokvist K, Ericson H, Möhler H, Ostenson CG, Smith PA. Glucose-inhibition of glucagon secretion involves activation of GABAA-receptor chloride channels. Nature. 1989;341:233–236. doi: 10.1038/341233a0. [DOI] [PubMed] [Google Scholar]
  64. Salehi A, Qader SS, Quader SS, Grapengiesser E, Hellman B. Inhibition of purinoceptors amplifies glucose-stimulated insulin release with removal of its pulsatility. Diabetes. 2005;54:2126–2131. doi: 10.2337/diabetes.54.7.2126. [DOI] [PubMed] [Google Scholar]
  65. Scepek S, Coorssen JR, Lindau M. Fusion pore expansion in horse eosinophils is modulated by Ca2+ and protein kinase C via distinct mechanisms. The EMBO Journal. 1998;17:4340–4345. doi: 10.1093/emboj/17.15.4340. [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Schwede F, Bertinetti D, Langerijs CN, Hadders MA, Wienk H, Ellenbroek JH, de Koning EJ, Bos JL, Herberg FW, Genieser HG, Janssen RA, Rehmann H. Structure-guided design of selective Epac1 and Epac2 agonists. PLOS Biology. 2015;13:e1002038. doi: 10.1371/journal.pbio.1002038. [DOI] [PMC free article] [PubMed] [Google Scholar]
  67. Segura F, Brioso MA, Gómez JF, Machado JD, Borges R. Automatic analysis for amperometrical recordings of exocytosis. Journal of Neuroscience Methods. 2000;103:151–156. doi: 10.1016/S0165-0270(00)00309-5. [DOI] [PubMed] [Google Scholar]
  68. Seino S, Takahashi H, Fujimoto W, Shibasaki T. Roles of cAMP signalling in insulin granule exocytosis. Diabetes, Obesity and Metabolism. 2009;11 Suppl 4:180–188. doi: 10.1111/j.1463-1326.2009.01108.x. [DOI] [PubMed] [Google Scholar]
  69. Sharma S, Lindau M. The fusion pore, 60 years after the first cartoon. FEBS Letters. 2018;592:3542–3562. doi: 10.1002/1873-3468.13160. [DOI] [PMC free article] [PubMed] [Google Scholar]
  70. Shibasaki T, Takahashi H, Miki T, Sunaga Y, Matsumura K, Yamanaka M, Zhang C, Tamamoto A, Satoh T, Miyazaki J, Seino S. Essential role of Epac2/Rap1 signaling in regulation of insulin granule dynamics by cAMP. PNAS. 2007;104:19333–19338. doi: 10.1073/pnas.0707054104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  71. Shin W, Ge L, Arpino G, Villarreal SA, Hamid E, Liu H, Zhao WD, Wen PJ, Chiang HC, Wu LG. Visualization of membrane pore in live cells reveals a Dynamic-Pore theory governing fusion and endocytosis. Cell. 2018;173:934–945. doi: 10.1016/j.cell.2018.02.062. [DOI] [PMC free article] [PubMed] [Google Scholar]
  72. Soltani N, Qiu H, Aleksic M, Glinka Y, Zhao F, Liu R, Li Y, Zhang N, Chakrabarti R, Ng T, Jin T, Zhang H, Lu WY, Feng ZP, Prud'homme GJ, Wang Q. GABA exerts protective and regenerative effects on islet beta cells and reverses diabetes. PNAS. 2011;108:11692–11697. doi: 10.1073/pnas.1102715108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  73. Takahashi T, Shibasaki T, Takahashi H, Sugawara K, Ono A, Inoue N, Furuya T, Seino S. Antidiabetic sulfonylureas and cAMP cooperatively activate Epac2A. Science Signaling. 2013;6:ra94. doi: 10.1126/scisignal.2004581. [DOI] [PubMed] [Google Scholar]
  74. Takahashi H, Shibasaki T, Park JH, Hidaka S, Takahashi T, Ono A, Song DK, Seino S. Role of Epac2A/Rap1 signaling in interplay between incretin and sulfonylurea in insulin secretion. Diabetes. 2015;64:1262–1272. doi: 10.2337/db14-0576. [DOI] [PubMed] [Google Scholar]
  75. Taraska JW, Perrais D, Ohara-Imaizumi M, Nagamatsu S, Almers W. Secretory granules are recaptured largely intact after stimulated exocytosis in cultured endocrine cells. PNAS. 2003;100:2070–2075. doi: 10.1073/pnas.0337526100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  76. Taraska JW, Almers W. Bilayers merge even when exocytosis is transient. PNAS. 2004;101:8780–8785. doi: 10.1073/pnas.0401316101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  77. Trexler AJ, Sochacki KA, Taraska JW. Imaging the recruitment and loss of proteins and lipids at single sites of calcium-triggered exocytosis. Molecular Biology of the Cell. 2016;27:2423–2434. doi: 10.1091/mbc.e16-01-0057. [DOI] [PMC free article] [PubMed] [Google Scholar]
  78. Tsuboi T, McMahon HT, Rutter GA. Mechanisms of dense core vesicle recapture following "kiss and run" ("cavicapture") exocytosis in insulin-secreting cells. Journal of Biological Chemistry. 2004;279:47115–47124. doi: 10.1074/jbc.M408179200. [DOI] [PubMed] [Google Scholar]
  79. Tsuboi T, Rutter GA. Insulin secretion by 'kiss-and-run' exocytosis in clonal pancreatic islet beta-cells. Biochemical Society Transactions. 2003;31:833–836. doi: 10.1042/bst0310833. [DOI] [PubMed] [Google Scholar]
  80. Tudurí E, Filiputti E, Carneiro EM, Quesada I. Inhibition of Ca 2+ signaling and glucagon secretion in mouse pancreatic α-cells by extracellular ATP and purinergic receptors. American Journal of Physiology-Endocrinology and Metabolism. 2008;294:E952–E960. doi: 10.1152/ajpendo.00641.2007. [DOI] [PubMed] [Google Scholar]
  81. Vergari E, Knudsen JG, Ramracheya R, Salehi A, Zhang Q, Adam J, Asterholm IW, Benrick A, Briant LJB, Chibalina MV, Gribble FM, Hamilton A, Hastoy B, Reimann F, Rorsman NJG, Spiliotis II, Tarasov A, Wu Y, Ashcroft FM, Rorsman P. Insulin inhibits glucagon release by SGLT2-induced stimulation of somatostatin secretion. Nature Communications. 2019;10:139. doi: 10.1038/s41467-018-08193-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  82. Wang CT, Grishanin R, Earles CA, Chang PY, Martin TF, Chapman ER, Jackson MB. Synaptotagmin modulation of fusion pore kinetics in regulated exocytosis of dense-core vesicles. Science. 2001;294:1111–1115. doi: 10.1126/science.1064002. [DOI] [PubMed] [Google Scholar]
  83. Weitz JR, Makhmutova M, Almaça J, Stertmann J, Aamodt K, Brissova M, Speier S, Rodriguez-Diaz R, Caicedo A. Mouse pancreatic islet macrophages use locally released ATP to monitor beta cell activity. Diabetologia. 2018;61:182–192. doi: 10.1007/s00125-017-4416-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  84. Wen PJ, Grenklo S, Arpino G, Tan X, Liao HS, Heureaux J, Peng SY, Chiang HC, Hamid E, Zhao WD, Shin W, Näreoja T, Evergren E, Jin Y, Karlsson R, Ebert SN, Jin A, Liu AP, Shupliakov O, Wu LG. Actin dynamics provides membrane tension to merge fusing vesicles into the plasma membrane. Nature Communications. 2016;7:12604. doi: 10.1038/ncomms12604. [DOI] [PMC free article] [PubMed] [Google Scholar]
  85. Wu Z, Bello OD, Thiyagarajan S, Auclair SM, Vennekate W, Krishnakumar SS, O'Shaughnessy B, Karatekin E. Dilation of fusion pores by crowding of SNARE proteins. eLife. 2017;6:e22964. doi: 10.7554/eLife.22964. [DOI] [PMC free article] [PubMed] [Google Scholar]
  86. Yabe D, Seino Y. Dipeptidyl peptidase-4 inhibitors and sulfonylureas for type 2 diabetes: friend or foe? Journal of Diabetes Investigation. 2014;5:475–477. doi: 10.1111/jdi.12229. [DOI] [PMC free article] [PubMed] [Google Scholar]
  87. Zhang CL, Katoh M, Shibasaki T, Minami K, Sunaga Y, Takahashi H, Yokoi N, Iwasaki M, Miki T, Seino S. The cAMP sensor Epac2 is a direct target of antidiabetic sulfonylurea drugs. Science. 2009;325:607–610. doi: 10.1126/science.1172256. [DOI] [PubMed] [Google Scholar]
  88. Zhao WD, Hamid E, Shin W, Wen PJ, Krystofiak ES, Villarreal SA, Chiang HC, Kachar B, Wu LG. Hemi-fused structure mediates and controls fusion and fission in live cells. Nature. 2016;534:548–552. doi: 10.1038/nature18598. [DOI] [PMC free article] [PubMed] [Google Scholar]
  89. Zhu Y, Chen H, Boulton S, Mei F, Ye N, Melacini G, Zhou J, Cheng X. Biochemical and pharmacological characterizations of ESI-09 based EPAC inhibitors: defining the ESI-09 "therapeutic window". Scientific Reports. 2015;5:9344. doi: 10.1038/srep09344. [DOI] [PMC free article] [PubMed] [Google Scholar]

Decision letter

Editor: Axel T Brunger1
Reviewed by: Ling-Gang Wu2

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

Thank you for submitting your article "Fusion pore regulation by Epac2/cAMP controls cargo release during insulin exocytosis" for consideration by eLife. Your article has been reviewed by Vivek Malhotra as the Senior Editor, a Reviewing Editor, and three reviewers. The following individuals involved in review of your submission have agreed to reveal their identity: Ling-Gang Wu (Reviewer #1).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

In this work the regulation of fusion pore dynamics in pancreatic β-cells is studied. The authors use human pancreatic β-cells for key first experiments, and subsequently from INS-1 cells and the Epac2 knockout mouse. Although it has been previously reported that Epac2 is directly involved in exocytosis at the exocytotic machinery, the precise mechanisms by which Epac2 regulates insulin exocytosis are poorly understood. The results suggest that cAMP and sulfonylureas restrict the fusion pore opening thereby limiting the release of insulin. Finally, the data further suggest that dynamin and amysin are part in the fusion pore restriction and that recruitment of these proteins to the release site is cAMP-dependent. The work adds new information on the role of Epac2 in fusion pore dynamics regarding the release from β-cells of not only insulin but also many other signaling smaller nucleotides.

There are several concerns as outlined below and we outline experimental data that should be included in the revised manuscript.

Essential revisions:

The heterogeneity among cells from different human donors is wide, and it can be argued whether results will hold if the authors increase the number of donors in their study. Moreover, the data from INS-1cells (rat islet origin and not human) require considering differences between human and rodent β-cells. It is essential that the key experiments in Figure 1 and the dynamin/amysin experiments in Figure 6 are repeated in the human cell line EndoC.

The authors do not really show the link between cAMP and dynamin/amysin. What mechanisms are involved in this process? The authors should address this by perhaps a speculative model in their discussion.

Abstract, the conclusion is wrong: cAMP elevation leads to pore expansion and peptide release, but not when Epac2 is inactivated pharmacologically or in Epac2-/- mice. Conversely, overexpression of Epac2 impedes pore expansion. Based on the data, the conclusion might be: cAMP elevation restricts pore and leads to a slower expansion of the fusion pore and peptide release, but not when Epac2 is inactivated pharmacologically or in Epac2-/- mice. Consistently, overexpression of Epac2 impedes pore expansion.

The complete release of NPY proceeds via full collapse fusion as shown in Figure 1(B) and in the text (Introduction), but according to reference (Shin et al., 2018), the vesicles can maintain the omega-shaped structures even after complete NPY release, and the pores can constrict after the complete release. Full-collapse is not observed for dense-core vesicle fusion in chromaffin cells (Shin et al., 2018). In addition, the fusion pore does not have to be larger than 10-30 nm for complete NPY release, because EGFP size is only 3.7nm as stated. The diagram in Figure 1(B) should be modified accordingly to avoid extrapolation from a pore more than 3.7 nm to full collapse fusion, and, "fusion pore lifetime" should be changed to "NPY release time" or other proper terms both in the legends and the main text. The methods used by the authors do not detect fusion pore lifetime, but only the NPY release time course (see Shin et al., 2018 for imaging of the fusion pore lifetime).

What is the reason that authors restricted exocytosis events to the rapid loss of the granule marker fluorescent (1-2 frames)? What about granule marker loss within 3-10 frames or >10 frame? Please clarify why these events are excluded. On the other hand, in Figure 6a, the release time of the NPY after flash takes more than 2 seconds with multiple frames (obviously >4 frames). This is inconsistent with the author's criteria of exocytosis events with the rapid loss within 1-2 frames. Please show the frame rate of all experiments since the different frame rates may generate different fraction of flash events vs full fusion events.

Rapid NPY spot fluorescence decrease can also be due to vesicle undocking (docked vesicle leaving the docking site), rather than fusion as in Figure 1A. The possibility of undocking should be ruled out by control experiments. A control experiment can be co-expressing pH-sensitive vesicular protein or adding dye in the bath solution to indicate fusion.

In subsection “cAMP-dependent fusion pore restriction is regulated by Epac but not PKA”, the authors state that PKA inhibition with Rp8-Br-cAMPS decreases exocytosis ~40% but not the fraction of flash events which is shown in Figure 1F. Then the percentage of flash events should be >80% in Figure 1D for fsk+RP-8, but the flash percentage is similar as fsk alone with ~60-65%. How is this possible? This needs to be clarified to support the conclusion that Epac rather than PKA is responsible for cAMP-dependent fusion pore regulation.

At the end of Introduction, the authors write that activation of Epac2 via GLP-1 and sulfonylurea will hinder full fusion and that this is "expected to reduce their full fusion as insulin secretagogoues". However, there are no data available in this study (or in published elsewhere) showing that addition of GLP-1 or sulfonylurea will reduce insulin secretion. How do the authors validate the importance of their study? Is glucose stimulated insulin secretion decreased in the cells (Figure 2) overexpressing Epac2? What about Epac2 and insulin secretion in EndoC cells?

The efficiency of the overexpression is not shown. What is the mRNA and protein levels of Epac2 (Figure 2), amysin and dynamin (Figure 6) after overexpression?

The authors have not performed the proper statistics for the data. A simple t-test is not enough when comparing multiple conditions. At least one-way ANOVA with ad hoc-test for multiple comparison.

The authors used Epac2 KO mice established in Kopperud et al. 2017, in which the characteristics of microvascular permeability of these mice was examined. However, the characteristics regarding glucose homeostasis including insulin secretion were not well studied in the paper. Furthermore, other lines of Epac2 KO mice have already been established and analyzed. The authors need to examine at least the insulin secretory profiles in the Epac2 KO mice used in this study in vivo and in vitro, and discuss them by referring to the previous studies using other KO strains.

Figure 2 shows that the percentage of flash events is increased by overexpression of Epac2. However, the actual number of flash events is not changed but full fusion events are decreased (Figure 2B), resulting in a decrease in the total number of fusion events (Figure 2A). In Figure 1, Epac activation by S223 increases the total number of exocytotic events by increasing flash events without changing the number of full fusion events. It seems that the fusion pore-regulating mechanism is different between these two conditions. Any thoughts on this issue?

In Figure 4, the total number of exocytotic events is not increased by fsk in Epac2 KO β-cells. However, it is unlikely that cAMP cannot increase the number of exocytotic events in the absence of Epac2, as PKA signaling remains intact. What is the authors' interpretation of these results?

In Figure 4F and 4G, fsk seems to prolong fusion pore lifetime significantly in both WT and Epac2 KO β-cells, although the authors state that the lifetimes are similar with or without forskolin. This should be modified.

In Figure 5, glib, tolb and gliz show similar effects on fusion pore regulation. However, the gliclazide data are inconsistent with the previous reports in which gliclazide among various sulfonylureas does not activate Epac2 (Zhang et al., 2009, Takahashi et al., 2013, Takahashi et al., 2013). The results should be confirmed using β-cells from Epac2 KO mice.

The authors show in Figure 6 that dynamin and amisyn are translocated to the site of exocytosis in a cAMP-dependent manner, and suggest that Epac protein is preassembled at the site of the secretory machinery. The authors also mention the involvement of Epac2 in the recruitment of dynamin and amisyn to the site of exocytosis, based on the observations of the translocation of these proteins by cAMP stimulation, which do not provide sufficient evidence to draw the authors' conclusion. The authors should show the behaviors of dynamin and amisyn in Epac2 KO β-cells and when Epac is activated pharmacologically.

The significance of the restriction of fusion pore expansion is not clear. Both full fusion and flash events eventually release all contents of the secretory vesicle. Thus, the prolonged fusion pore lifetime means only a few seconds delay in insulin release compared to release of small molecules (or small molecules release a few seconds in advance of the insulin release). Although the authors discuss the roles of these small molecules in autocrine and paracrine signaling in islets, the physiological and pathophysiological relevance of the time difference in the release of insulin and small molecules should be further discussed.

It has previously been shown that there are different modes of insulin exocytosis, based on the dynamics of insulin granules (predocked granules, newcomer granules, etc.) (Shibasaki et al.,2007 and Nagamatsu et al., 2006). Is there any difference in the mode of insulin exocytosis between the present study and the previous reports? The authors might want to discuss the modes of exocytotic events found in the present study in relation to previous studies.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "Fusion pore regulation by cAMP/Epac2 controls cargo release during insulin exocytosis" for further consideration at eLife. Your revised article has been favorably evaluated by Vivek Malhotra (Senior Editor), a Reviewing Editor, and three reviewers.

The manuscript has been improved but there are some remaining issues that need to be addressed before acceptance, as outlined below:

1) Figure 1B. The drawing of full fusion with a wide concave structure may be subject to error, as recent super-resolution STED imaging shows that fusing dense-core vesicles maintain an omega-shape till the STED resolution limit (Shin et al., 2018; Chiang et al., 2014). We suggest to re-draw the concave shape in Figure 1B as an omega shape with a larger pore than the other omega profile at the left side. This does not mean that full-collapse does not occur in β-cells, but the authors' data can only say that the fusion pore is larger than NPY-venus.

2) The reviewer's comment: “Figure 2 shows that the percentage of flash events…”

The authors do not respond to the comment properly. The reviewers asked the authors to discuss the mechanisms of the different findings between Figure 2 and Figure 1. This should be reconsidered.

3) The reviewer's comment: “In Figure 5, Glib, tolb and gliz show similar effects…”

The authors have performed the requested experiments, and suggested that Epac2 is required for the effect of gliz on the fusion pore. The result on gliz is different from the previous reports (Zhang et al., 2009; Takahashi et al., 2013). The authors should discuss reasons for the discrepancy.

4) The reviewer's comment: “The role of Eoac2(cAMP-GEFII) in insulin exocytosis…” The authors responded to the comment, but the modified statement has not been incorporated in the text (Introduction).

5) The reviewer's comment: “The source of reagents…” The source is provided, but the source of the Epac2 agonist S223 is missing.

6) The reviewer's comment: The description "inhibitors of the GLP-1 peptidase"… In the authors' response, "inhibitors of DPP-4 peptidase" should be "inhibitors of DPP-4"(delete"peptidase").

eLife. 2019 May 20;8:e41711. doi: 10.7554/eLife.41711.021

Author response

Summary:

In this work the regulation of fusion pore dynamics in pancreatic β-cells is studied. The authors use human pancreatic β-cells for key first experiments, and subsequently from INS-1 cells and the Epac2 knockout mouse. Although it has been previously reported that Epac2 is directly involved in exocytosis at the exocytotic machinery, the precise mechanisms by which Epac2 regulates insulin exocytosis are poorly understood. The results suggest that cAMP and sulfonylureas restrict the fusion pore opening thereby limiting the release of insulin. Finally, the data further suggest that dynamin and amysin are part in the fusion pore restriction and that recruitment of these proteins to the release site is cAMP-dependent. The work adds new information on the role of Epac2 in fusion pore dynamics regarding the release from β-cells of not only insulin but also many other signaling smaller nucleotides.

There are several concerns as outlined below and we outline experimental data that should be included in the revised manuscript.

Essential revisions:

The heterogeneity among cells from different human donors is wide, and it can be argued whether results will hold if the authors increase the number of donors in their study. Moreover, the data from INS-1 cells (rat islet origin and not human) require considering differences between human and rodent β-cells. It is essential that the key experiments in Figure 1 and the dynamin/amysin experiments in Figure 6 are repeated in the human cell line EndoC.

The EndoC line is not available to us, despite a request to the authors. During revision, we instead added data from four additional human donors, so that each of the conditions in Figure 1 now has been measured with 5-7 donors. We do not see strong donor to donor variation with respect to fusion pore behavior, and even average exocytosis is relatively similar between preps from non-diabetic donors (see Author response image 1 and Gandasi et al., 2018). While there are differences in exocytosis rate between Ins1 and human cells, the effect of forskolin is very similar between the two. Exocytosis roughly doubles, the fraction of flash events shifts from 40 to 60%, and pore lifetimes increase from 0.27 to 0.7s for Ins1 and from 0.3 to 0.9s for human cells. We would also like to point out that Ins1 is a far more established model for insulin exocytosis (at least 1800 papers in PubMed) than EndoC (about 60 papers, of which only 4 studied exocytosis).

Author response image 1. K+-stimulated exocytosis in the indicated conditions.

Author response image 1.

Open in a new tab

In (a) each symbol represents the average of a donor, in (b), each symbol represents one cell.

The authors do not really show the link between cAMP and dynamin/amysin. What mechanisms are involved in this process? The authors should address this by perhaps a speculative model in their discussion.

We now include data demonstrating that pharmacological inactivation of Epac prevents cAMP-dependent recruitment of dynamin and amisyn, and that the specific Epac activator S223 leads to recruitment of these proteins during flash exocytosis even in absence of cAMP raising agents (Figure 6).

We speculate (in the Discussion section) that cAMP/Epac controls the local generation of PIP2, which in turn leads to recruitment of the PH-domain containing proteins dynamin and amisyn:

“Indeed, we show here that recruitment of two other proteins, dynamin and amisyn, depends on cAMP and Epac. Other known targets of Epac are the small GTPases Rap1 and R-Ras, for which Epac is a guanine nucleotide exchange factor (GEF). Rap1 is expressed on insulin granules and affects insulin secretion both directly 65, and by promoting intracellular Ca2+-release following phospholipase-C activation 66. R-Ras is an activator of phosphoinositide 3-kinase 67. By altering local phosphoinositide levels, Epac could therefore indirectly affect exocytosis via recruitment of C2-domain proteins such as Munc13 68, and fusion pore behavior by recruitment of the PH-domain containing proteins dynamin and amisyn 69,70.”

Abstract, the conclusion is wrong: cAMP elevation leads to pore expansion and peptide release, but not when Epac2 is inactivated pharmacologically or in Epac2-/- mice. Conversely, overexpression of Epac2 impedes pore expansion. Based on the data, the conclusion might be: cAMP elevation restricts pore and leads to a slower expansion of the fusion pore and peptide release, but not when Epac2 is inactivated pharmacologically or in Epac2-/- mice. Consistently, overexpression of Epac2 impedes pore expansion.

We thank the reviewer for pointing out this error, which has been corrected as suggested.

The complete release of NPY proceeds via full collapse fusion as shown in Figure 1(B) and in the text (Introduction), but according to reference (Shin et al., 2018), the vesicles can maintain the omega-shaped structures even after complete NPY release, and the pores can constrict after the complete release. Full-collapse is not observed for dense-core vesicle fusion in chromaffin cells (Shin et al., 2018). In addition, the fusion pore does not have to be larger than 10-30 nm for complete NPY release, because EGFP size is only 3.7nm as stated. The diagram in Figure 1(B) should be modified accordingly to avoid extrapolation from a pore more than 3.7 nm to full collapse fusion, and, "fusion pore lifetime" should be changed to "NPY release time" or other proper terms both in the legends and the main text. The methods used by the authors do not detect fusion pore lifetime, but only the NPY release time course (see Shin et al., 2018 for imaging of the fusion pore lifetime).

We agree that long-lived omega-shaped structures cannot be excluded in pancreatic β-cells. This is now acknowledged (Introduction). We have also changed the fusion pore lifetime to NPY release time thorough the figures and text, as suggested.

What is the reason that authors restricted exocytosis events to the rapid loss of the granule marker fluorescent (1-2 frames)? What about granule marker loss within 3-10 frames or >10 frame? Please clarify why these events are excluded. On the other hand, in Figure 6a, the release time of the NPY after flash takes more than 2 seconds with multiple frames (obviously >4 frames). This is inconsistent with the author's criteria of exocytosis events with the rapid loss within 1-2 frames. Please show the frame rate of all experiments since the different frame rates may generate different fraction of flash events vs full fusion events.

Detection of events was done by eye, using the criterion that the majority of the signal is lost suddenly (1-2 frames). This is now stated in the Materials and methods section, where we added the following text:

“The frame rate was 10 frames*s-1, with 100 ms exposures” subsection “TIRF microscopy” and “characteristic rapid loss of the granule marker fluorescence (most fluorescence lost within 1-2 frames)” subsection “Image analysis”.

In rare cases, a small portion of the signal remained at the release site (as in Figure 6A, note that only every fifth frame is shown). This is apparent in the quantification of the NPY-mOrange signal (now shown in Figure 6—figure supplement 1). There are four possible explanations for the remaining fluorescence, (1) binding of the label to the outside of the cell after complete release (Michael, 2006), (2) kiss and run followed by undocking or re-acidification (Obermüller, 2005), (3) stray light from a nearby granule that did not undergo exocytosis, or (4) rapid undocking (see answer to the next question). We would maintain that the effect of all of this on our flash/no flash ratio measurement is minor. Occasionally, pore lifetimes for no-flash events are overestimated by the fitter, but this would not affect the conclusions of the paper.

Rapid NPY spot fluorescence decrease can also be due to vesicle undocking (docked vesicle leaving the docking site), rather than fusion as in Figure 1A. The possibility of undocking should be ruled out by control experiments. A control experiment can be co-expressing pH-sensitive vesicular protein or adding dye in the bath solution to indicate fusion.

Undocking is a rare event, 25-100 times less frequently than K+-stimulated exocytosis (eg Gandasi et al., 2018). However, control experiments were included in two of our previous papers. Accordingly, NPY-EGFP fluorescence disappears rapidly during exocytosis, while loss of the signal during undocking takes several seconds (Gandasi and Barg,2014; Figure S7). Multiplane confocal imaging also indicated that undocking is very slow, compared with exocytosis/content release (Barg, 2002, Figure 3).

In subsection “cAMP-dependent fusion pore restriction is regulated by Epac but not PKA”, the authors state that PKA inhibition with Rp8-Br-cAMPS decreases exocytosis ~40% but not the fraction of flash events which is shown in Figure 1F. Then the percentage of flash events should be >80% in Figure 1D for fsk+RP-8, but the flash percentage is similar as fsk alone with ~60-65%. How is this possible? This needs to be clarified to support the conclusion that Epac rather than PKA is responsible for cAMP-dependent fusion pore regulation.

We respectfully disagree. Assuming that PKA only affects the rate of exocytosis (but not fusion pore behavior), we would expect that the fraction of flash events remains constant in presence of the PKA blocker RP-8. This is in fact what is observed (Figure 1D).

At the end of Introduction, the authors write that activation of Epac2 via GLP-1 and sulfonylurea will hinder full fusion and that this is "expected to reduce their full fusion as insulin secretagogoues". However, there are no data available in this study (or in published elsewhere) showing that addition of GLP-1 or sulfonylurea will reduce insulin secretion. How do the authors validate the importance of their study? Is glucose stimulated insulin secretion decreased in the cells (Figure 2) overexpressing Epac2? What about Epac2 and insulin secretion in EndoC cells?

The statement is based on the data in Figure 5 suggesting that SU’s restrict fusion pore expansion by the same Epac-dependent mechanism as cAMP. In situations where exocytosis is strongly stimulated (in Figure 5 by SU, K+ and ex-4), we observe a relative decrease in exocytosis (and therefore likely also insulin release) rather than an additive effect of SU and fsk. In vivo, SU will primarily stimulate exocytosis by closing KATP channels and depolarization, but it will be worth investigating whether the stimulatory effect of GLP-1 agonists is affect by longterm SU treatment.

However, our study is focused on the molecular mechanism by which cAMP regulates fusion pore behavior, and we feel that in vivo studies to determine the physiological role of such mechanisms are beyond the scope of the paper. However, there is some evidence that fusion pore regulation by amisyn restricts insulin secretion, and may play a role in human type-2 diabetes (Collins, 2016). We also discuss the possibility that a delay between the release of small transmitters and insulin after Epac activation could be physiologically relevant, for example by influencing para/autocrine signaling in the islet. Such effects might be difficult to detect with traditional insulin secretion measurements, in particular since small molecule release and exocytosis are not considered in secretion measurements.

Regarding EndoC cells, please refer to our comment above.

The efficiency of the overexpression is not shown. What is the mRNA and protein levels of Epac2 (Figure 2), amysin and dynamin (Figure 6) after overexpression?

We believe that such quantification should be done on single cells, because the transfection only reaches a (varying) fraction of the cells and expression levels vary from cell to cell. We have therefore immunostained INS1 cells that had been transfected with mCherry-amisyn. In these cells we quantified both mCherry-amisyn and immunostaining (anti-amisyn detected with Alexa 488) in a large number of cells, including non-expressing cells (a). As expected, the Alexa fluorescence is higher in transfected cells (b), and there is a near linear relationship between red and green fluorescence (c); the y-axis offset is caused by non-transfected cells that are labeled green but not red, and represents endogenous amisyn expression. From this, we estimate that we worked with cells that overexpress the mCherry-amisyn 2-4 fold compared with endogenous expression. The data also hint that endogenous amisyn may be suppressed by expression of mCherry-amisyn (the average of non-transfected cells is higher than the offset). Similar experiments were done with dynamin1-GFP, and we observed a similar 2-3 fold increase for overexpressed dynamin 1 (d and e). The range of overexpression is similar to a number of other proteins that we have tested (eg syntaxin, CaV1.2), but unfortunately we were unable to source Epac antibodies of sufficient quality to succeed with such quantifications for Epac2.

The authors have not performed the proper statistics for the data. A simple t-test is not enough when comparing multiple conditions. At least one-way ANOVA with ad hoc-test for multiple comparison.

We thank the reviewer for the suggestion. We now provide one-way ANOVA with posthoc tests (Kruskal Wallis/Dunns) instead of t-tests. The new statistics are described in the methods and legends.

The authors used Epac2 KO mice established in Kopperud et al. 2017, in which the characteristics of microvascular permeability of these mice was examined. However, the characteristics regarding glucose homeostasis including insulin secretion were not well studied in the paper. Furthermore, other lines of Epac2 KO mice have already been established and analyzed. The authors need to examine at least the insulin secretory profiles in the Epac2 KO mice used in this study in vivo and in vitro, and discuss them by referring to the previous studies using other KO strains.

The Epac2 knockout (KO) mouse used in this study differs from the original and most widely employed strain in that the deletion involves exons 12-13, which encode the high-affinity cAMP binding domain present in all Epac2 isoforms. With the previously generated mouse, a deletion was made in the first exon, thereby eliminating the dominating longest isoform Epac2A, but allowing expression of the shorter isoforms Epac2B and 2C. Indeed, there is evidence that Epac2B is present in mouse pancreatic islets, although it has not been clarified in which cell type(s) the protein is expressed Høivik et al., 2013. We therefore think that it is advantageous to use the mouse in which all Epac2 isoforms have been eliminated. We have added a short description in the methods section explaining the difference between the present mouse and the Epac2A-KO mouse.

A general characterization of the Epac2-KO mouse is ongoing but beyond the scope of this manuscript. Available insulin secretion data do not indicate any striking phenotypical differences from the first published Epac2A-KO strain. We have performed intraperitoneal glucose tolerance tests showing that the mice are slightly glucose intolerant at an age of 10-11 months (Author response image 2). This is reminiscent of the glucose intolerance reported for Epac2A-KO following increased metabolic demand induced by e.g. high fat feeding (Song et al., 2013).

Author response image 2. Means ± s.e.m. for blood glucose during intraperitoneal glucose tolerance tests in three 10-11-month-old wild type and Epac2-deficient mice.

Author response image 2.

Open in a new tab

Moreover, our analyses show that glucose-stimulated insulin secretion from isolated islets remains intact (Author response image 3 and 4), whereas amplification of secretion by cAMP-elevating glucagon is impaired (Author response image 4). Although the number of observations are too few for the difference to reach statistical significance, the trend is as expected in light of the function of Epac2 and previously published data from Epac2A-deficient islets. We think that this functional characterization of the mouse would not add much to the present manuscript and think that the data would fit better in a forthcoming manuscript with more extensive phenotypic analyses and functional comparison between Epac2 and Epac1.

Author response image 3. Means ± s.e.m. for insulin from isolated islets stimulated by an elevation of the glucose concentration from 3 to 11 mM.

Author response image 3.

Open in a new tab

N=4 mice for each genotype.

Author response image 4. Means ± s.e.m. for blood for insulin from isolated islets stimulated by an elevation of the glucose concentration from 3 to 20 mM and addition of 10 nM glucagon.

Author response image 4.

Open in a new tab

N=3 mice for each genotype.

Figure 2 shows that the percentage of flash events is increased by overexpression of Epac2. However, the actual number of flash events is not changed but full fusion events are decreased (Figure 2B), resulting in a decrease in the total number of fusion events (Figure 2A). In Figure 1, Epac activation by S223 increases the total number of exocytotic events by increasing flash events without changing the number of full fusion events. It seems that the fusion pore-regulating mechanism is different between these two conditions. Any thoughts on this issue?

We speculate that Epac becomes inhibitory for exocytosis due to the strong activation of pore restriction with both fsk and exogenous Epac, which competes with the cAMP-dependent acceleration of priming/exocytosis. We believe that this is also the reason for decreased exocytosis in presence of both forskolin and sulfonylurea (Figure 5B). Collins et al., (2016, and comment by Barg and Gucek, 2016) came to a similar conclusion regarding the role of amisyn in type 2 diabetes.

In Figure 4, the total number of exocytotic events is not increased by fsk in Epac2 KO β-cells. However, it is unlikely that cAMP cannot increase the number of exocytotic events in the absence of Epac2, as PKA signaling remains intact. What is the authors' interpretation of these results?

We are also surprised by the weak forskolin effect and the explanation is not clear. Since the PKA-dependent effect is slower than the Epac effect on exocytosis (Renström, 1997), it may be speculated that it has not become manifested during the relatively short time course of the experiment. Another possibility is that cAMP is already somewhat elevated in the control situation, since the cells are exposed to 10 mM glucose. Elevated cAMP may lead to PKA activation. It has been demonstrated that the apparent Kd for stimulating granule priming/exocytosis is 5-fold lower for PKA than for Epac (Eliasson, 2003).

In Figure 4F and 4G, fsk seems to prolong fusion pore lifetime significantly in both WT and Epac2 KO β-cells, although the authors state that the lifetimes are similar with or without forskolin. This should be modified.

Thank you for pointing this out. While the ratios of flash events are equally low, the NPY release time is increased in Epac2KO mice after addition of fsk. We have edited the text accordingly:

“In contrast, exocytosis was not accelerated by forskolin in Epac2-/- cells, and the fraction of flash events was similar with or without forskolin (Figure 4C, F-G).” (subsection “cAMP-dependent fusion pore regulation is absent in Epac2-/- (Rapgef4-/-) β-cells”).

In Figure 5, glib, tolb and gliz show similar effects on fusion pore regulation. However, the gliclazide data are inconsistent with the previous reports in which gliclazide among various sulfonylureas does not activate Epac2 (Zhang et al., 2009, Takahashi et al., 2013, Takahashi et al., 2013). The results should be confirmed using β-cells from Epac2 KO mice.

We´ve performed the requested experiment, where we bathed Epac2KO β-cells with gliclizide and monitored exocytosis upon K+ depolarization. We observed no increase in either fraction of flash events (a), exocytosis rate (b) or the NPY release times (c), suggesting that Epac2 is required for the effect of gliclazide on the fusion pore. We had already included similar measurements with tolbutamide in the Epac2 KO model (Figure 7 bottom left), indicating that both SU’s affect the pore via Epac2.

Author response image 5.

Author response image 5.

Open in a new tab

The authors show in Figure 6 that dynamin and amisyn are translocated to the site of exocytosis in a cAMP-dependent manner, and suggest that Epac protein is preassembled at the site of the secretory machinery. The authors also mention the involvement of Epac2 in the recruitment of dynamin and amisyn to the site of exocytosis, based on the observations of the translocation of these proteins by cAMP stimulation, which do not provide sufficient evidence to draw the authors' conclusion. The authors should show the behaviors of dynamin and amisyn in Epac2 KO β-cells and when Epac is activated pharmacologically.

Unfortunately, we are unable to perform the suggested experiments in Epac2KO cells, without first generating and testing an adenoviral expression system for labeled dynamin and amisyn. We have instead performed experiments in INS1-cells, in which we quantified amisyn or dynamin binding in presence of S223 or ESI-09 to pharmacologically modulate Epac function.

As evident from the new Figure 6, pharmacological inactivation with ESI09 prevents binding of either dynamin1 (c) or amisyn (f) to the release site, resulting in decreased NPY release times (d,g) and fractions of flash events (e,h). Similarly, in absence of fsk, pharmacological activation of Epac2 by S223 results in accumulation of proteins at the release site at the point of fusion (i,l) and longer NPY release times (j,m) together with higher fraction of flashes (k,n). We´ve changed the figure legends accordingly (changes are highlighted in yellow). The results strengthen the conclusion that Epac activation leads to accumulation of dynamin and amisyn at the release site.

The significance of the restriction of fusion pore expansion is not clear. Both full fusion and flash events eventually release all contents of the secretory vesicle. Thus, the prolonged fusion pore lifetime means only a few seconds delay in insulin release compared to release of small molecules (or small molecules release a few seconds in advance of the insulin release). Although the authors discuss the roles of these small molecules in autocrine and paracrine signaling in islets, the physiological and pathophysiological relevance of the time difference in the release of insulin and small molecules should be further discussed.

The Discussion section has been in part rewritten to strengthen the focus on physiological and pathophysiological relevance.

It has previously been shown that there are different modes of insulin exocytosis, based on the dynamics of insulin granules (predocked granules, newcomer granules, etc.) (Shibasaki et al.,2007 and Nagamatsu et al., 2006). Is there any difference in the mode of insulin exocytosis between the present study and the previous reports? The authors might want to discuss the modes of exocytotic events found in the present study in relation to previous studies.

While a number of labs have reported crash fusion (all using similar protocols), others have not (Tsuboi and Rutter, 2003; Barg et al., 2002; Michael et al., 2007; Hoppa et al., 2009; Gandasi and Barg, 2014). We recently presented a comprehensive analysis of near-membrane granule behavior in human β-cells, where there was no evidence for crash fusion (Gandasi et al., 2018). The reason why some groups observe crash fusion but others not is unclear. Possible differences include the use of glucose instead of KCl for stimulation (Shibasaki et al., 2007), the duration of the stimulation, and technical issues such as the use of thick cell supports (eg laminin), mistargeting of the granule label to other vesicle types, and label bleaching during long recordings. The latter would give the false appearance of no granule at the site of exocytosis. While this question is interesting and important (and should be settled), we believe that it is not central to the mechanism of fusion pore regulation by Epac and should rather be discussed in a topical review.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "Fusion pore regulation by cAMP/Epac2 controls cargo release during insulin exocytosis" for further consideration at eLife. Your revised article has been favorably evaluated by Vivek Malhotra (Senior Editor), a Reviewing Editor, and three reviewers.

The manuscript has been improved but there are some remaining issues that need to be addressed before acceptance, as outlined below:

1) Figure 1B. The drawing of full fusion with a wide concave structure may be subject to error, as recent super-resolution STED imaging shows that fusing dense-core vesicles maintain an omega-shape till the STED resolution limit (Shin et al., 2018; Chiang et al., 2014). We suggest to re-draw the concave shape in Figure 1B as an omega shape with a larger pore than the other omega profile at the left side. This does not mean that full-collapse does not occur in β-cells, but the authors' data can only say that the fusion pore is larger than NPY-venus.

We thank the reviewer for the suggestion. We have redrawn the final fusion step in Figure 1B as requested.

2) The reviewer's comment: “Figure 2 shows that the percentage of flash events…”

The authors do not respond to the comment properly. The reviewers asked the authors to discuss the mechanisms of the different findings between Figure 2 and Figure 1. This should be reconsidered.

This is probably due to its independent action on priming and fusion pore control. With activation (Figure 1; S223) both the effect of priming and the effect on the fusion pore are strongly present. With Epac overexpression only (Figure 2B), there is no effect on the priming, but only on the fusion pore restriction. As we´ve discussed in the response letter, we speculate that Epac, similarly as amisyn, becomes inhibitory for exocytosis due to the strong restriction of the pore with both fsk and exogenous Epac present. We might be under-detecting the restricted fusion pores that close back (kiss-and-run events), which accounts for the observed exocytosis decrease.

We´ve added a short discussion in the Discussion section:” Epac mediates the PKA-independent stimulation of exocytosis by cAMP 59 and our data suggests it may affect both priming and fusion pore restriction.”

3) The reviewer's comment: “In Figure 5, Glib, tolb and gliz show similar effects…”

The authors have performed the requested experiments, and suggested that Epac2 is required for the effect of gliz on the fusion pore. The result on gliz is different from the previous reports (Zhang et al., 2009; Takahashi et al., 2013). The authors should discuss reasons for the discrepancy.

We thank the reviewer for the comment. The mentioned reports state gliz does not activate Epac2A, but it can still bind (only through 1 binding site G114) to the CNB1 domain (Takahashi et al., 2013). We´ve previously shown Epac2 targeting to the plasma membrane doesn’t depend on CNB1 domain. The clustering at the granule site however does depend on the CNB1 domain (Alenkvist et al., 2017). Since we speculate that Epac works only as a regulator of downstream proteins (e.g. dynamin 1 and amisyn), the localization at the granule site without activation may explain the restriction of the fusion pores we´ve observed in this study. This is further supported with the data in Figure 2, where overexpression of Epac in absence of forskolin results in a strong restriction of the fusion pores.

We´ve added a short discussion in the Discussion section: “Since gliclizide binds the CNB1 domain without activating it 54 and still restricts the fusion pore, Epac localization at the granule site 39 may be enough to regulate the downstream proteins (e.g. dynamin and amisyn).”

4) The reviewer's comment: “The role of Eoac2(cAMP-GEFII) in insulin exocytosis…” The authors responded to the comment, but the modified statement has not been incorporated in the text (Introduction).

We thank the reviewer for noticing the mistake. We´ve modified the text accordingly.

5) The reviewer's comment: “The source of reagents…” The source is provided, but the source of the Epac2 agonist S223 is missing.

We thank the reviewer for the comment. We´ve added the source of S223 (Biolog) to the key resource table.

6) The reviewer's comment: The description "inhibitors of the GLP-1 peptidase"… In the authors' response, "inhibitors of DPP-4 peptidase" should be "inhibitors of DPP-4"(delete"peptidase").

We´ve modified the text accordingly.

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Data Citations

    1. Gucek A, Gandasi NR, Omar-Hmeadi M, Bakke M, Døskeland SO, Tengholm A, Barg S. 2019. Data from: Fusion pore regulation by cAMP/Epac2 controls cargo release during insulin exocytosis. Dryad Digital Repository. [DOI] [PMC free article] [PubMed]

    Supplementary Materials

    Figure 7—source data 1.
    elife-41711-fig7-data1.xlsx (12.7KB, xlsx)
    DOI: 10.7554/eLife.41711.015
    Transparent reporting form
    elife-41711-transrepform.pdf (474.4KB, pdf)
    DOI: 10.7554/eLife.41711.016

    Data Availability Statement

    Source data file has been provided for Figure 7. All raw data are available on the Dryad Digital Repository (http://dx.doi.org/10.5061/dryad.6b604g8).

    The following dataset was generated:

    Gucek A, Gandasi NR, Omar-Hmeadi M, Bakke M, Døskeland SO, Tengholm A, Barg S. 2019. Data from: Fusion pore regulation by cAMP/Epac2 controls cargo release during insulin exocytosis. Dryad Digital Repository.

    Articles from eLife are provided here courtesy of eLife Sciences Publications, Ltd

    RESOURCES