Melanophilin Accelerates Insulin Granule Fusion without Predocking to the Plasma Membrane

Professional secretory cells store bioactive molecules in vesicles in advance and release them in response to extracellular stimuli by promoting fusion of vesicle membranes to the plasma membrane. In such regulated exocytic pathways, secretory granules carrying proteins as cargo must be regenerated at the Golgi apparatus after releasing their contents; in contrast with that, synaptic vesicles containing low-molecular-weight substances can recycle within the presynaptic terminal. Therefore, newly generated granules must cross a peripheral microfilament web, referred to as the actin cortex, before approaching the plasma membrane. However, the molecular mechanism by which granules link to the F-actin network and are processed toward exocytosis remains poorly understood. This process that accumulates granules in the cell periphery is thought to form a reserve pool to sustain regulated secretion after depletion of a readily releasable pool beneath the plasma membrane. Its disturbance could impair the capacity of secretory cells to cope with external changes and stresses and cause diseases such as type 2 diabetes.

We recently showed that exophilin-8 (also known as MyRIP and Slac2-c), one of the Rab27 effectors that play versatile roles in regulated secretory pathways (1), captures granules within the actin cortex via indirect interaction with myosin-VIIa through binding to RIM-BP2 and that this exophilin-8–RIM-BP2–myosin-VIIa complex formation is critical for peripheral accumulation and efficient exocytosis of insulin granules (2). However, another motor protein on actin filaments, myosin-Va, has also been suggested to function as a carrier to capture and/or transport granules to the vicinity of the plasma membrane (3–6), although the molecular mechanism by which myosin-Va functions in granule exocytosis remains unknown. Myosin-Va does not interact with exophilin-8 but binds another Rab27 effector, melanophilin (also known as exophilin-3 and Slac2-a), in pancreatic β-cells (2). Melanophilin retains melanosomes in the periphery of skin melanocytes by directly interacting with both Rab27a on melanosomes and myosin-Va on cortical actin filaments, which makes melanosomes capable of being transferred to neighboring keratinocytes (7–10). Its functional loss leads to clustering of melanosomes near the perikaryotic regions and causes hypopigmentation in both leaden mice and in patients with Griscelli syndrome (11,12). However, other overt abnormalities have not been reported. The current study demonstrates in vivo function of melanophilin in insulin granule exocytosis that bypasses stable predocking to the plasma membrane.

Leaden (C57J/L) and C57BR/cdJ mice were purchased from The Jackson Laboratory. To minimize potential effects of spontaneous mutations occurring after separation of these inbred strains, C57L/J were crossed with C57BR/cdJ, and the resultant heterozygous mice were intercrossed to generate Mlph^ln/Mlph^ln mice, which were used in the current study. Animal experiments were performed according to the rules and regulations of the Animal Care and Experimentation Committees of Gunma University (Gunma, Japan) and the University of Tokyo (Tokyo, Japan). Only male mice and their tissues and cells were phenotypically characterized in this study. Blood glucose levels were determined by a glucose oxidase method using Glutest Pro GT-1660 (Sanwa Kagaku Kenkyujyo). Insulin was measured by an AlphaLISA insulin kit (PerkinElmer). Pancreatic islet isolation, perifusion secretion assays, and morphometric electron microscopic analysis of granule distribution were performed as described previously (13,14).

The sources of antibodies and their concentrations used are listed in Supplementary Table 1. Cells lysate proteins, separated by gel electrophoresis, were transferred onto an Immobilon-P membrane (Millipore) and were visualized by means of enhanced chemiluminescence (GE Healthcare Biosciences). Immunoprecipitation was performed at 4°C by incubation with primary antibody overnight, followed by the addition of protein G-agarose beads (GE Healthcare Bioscience) for 1 h or by direct incubation with anti- hemagglutinin (HA) affinity matrix beads (Roche Diagnostics) or anti-FLAG affinity gel (Sigma-Aldrich) for 1 h. For immunofluorescence, primary β-cells were fixed by 4% paraformaldehyde for 30 min at room temperature and were rehydrated with PBS for 5 min, followed by PBS plus 0.1% Triton X-100 for 30 min. The cells incubated with primary antibody overnight at 4°C, followed by Alexa Fluor 488- or 568-conjugated secondary antibody for 1 h at room temperature, were observed by a confocal laser scanning microscope. Each image is representative of at least three independent experiments.

Mouse melanophilin and syntaxin-4 cDNAs were derived from MIN6 cells. Point and deletion mutants were generated using a standard PCR-based mutagenesis strategy and were verified by DNA sequencing. The sequences of the primers used are listed in Supplementary Table 2. These cDNAs were subcloned into pcDNA3-HA, pcDNA3-FLAG (Invitrogen), pmCherry-C1, pEGFP-C1 (Clontech), pMAL-cR1 (New England Biolabs), pGEX4T-1 (GE Healthcare Bioscience), or pCAG with a One-STrEP-Flag tag as described previously (2,13). Neuropeptide Y (NPY)-mCherry cDNA was generated by subcloning an mCherry cDNA into the pNPY-Venus-N1 vector. To generate recombinant adenoviruses, they were inserted into pENTR-3C (Invitrogen) and were transferred into pAd/CMV by LR Clonase recombination (Invitrogen). To express exogenous protein, HEK293A cells were transfected with the plasmids using Lipofectamine 2000 reagent (Invitrogen), whereas MIN6 cells were infected with adenoviruses.

Human islets (Supplementary Human Islet Checklist) were provided by the Alberta Diabetes Institute IsletCore of the University of Alberta under full ethical clearance (Yokohama City University Ethics Board, B171100025, Yokohama, Kanagawa, Japan; and Human Tissue MTA from the University of Alberta, UA17-DSA-64, Edmonton, Alberta, Canada). Mouse and human islets were dissociated into monolayer cells by incubation with trypsin-EDTA solution and were cultured on poly-l-lysine–coated 35-mm glass base dishes for 2 days. The cells were infected with adenovirus encoding preproinsulin-EGFP (Insulin-EGFP) or NPY-mCherry and were further cultured for 2 days. Total internal reflection fluorescence (TIRF) microscopy (the penetration depth of the evanescent field: 100 nm) was performed as described previously (15,16). The cells were preincubated for 30 min in 2.8 mmol/L glucose-containing Krebs-Ringer bicarbonate buffer (KRBB) at 37°C, and were exposed to 25 mmol/L glucose stimulation for 20 min. Images were acquired at 103-ms intervals. Fusion events with a flash were manually selected and assigned to one of three types: residents, which are visible over 10 s before fusion; visitors, which have become visible within 10 s before fusion; and passengers, which are not visible before fusion.

Human pancreatic islet cells suspended in 1 × 10⁵ cells/200 μL were transfected with 50 nmol/L control On-Target plus nontargeting pool siRNA or SMARTpool siRNA against human melanophilin (79083; GE Dharmacon) using Lipofectamine RNAiMAX reagent (Invitrogen). After being plated on a glass base dish for 72 h, control and melanophilin siRNA-treated cells were infected with 10 multiplicity of infection of mCherry-tagged, nontargeting (ACTACCGTTGTTATAGGTG) and human melanophilin targeting (GCGTTGAAGGGCAAGATTA) shRNA adenoviruses, respectively. Cells intended for TIRF microscopy were coinfected with adenoviruses encoding shRNA and Insulin-EGFP, wherein infected cells were determined by mCherry and EGFP expression.

Two-photon extracellular polar tracer imaging of exocytic events in islets was performed as described previously (17). The external bathing solution contained 10 mmol/L HEPES (pH 7.4), 150 mmol/L NaCl, 5 mmol/L KCl, 1 mmol/L MgCl₂, 2 mmol/L CaCl₂, 2.8 mmol/L glucose, and 0.7 mmol/L sulforhodamine B. Exocytic events triggered with 20 mmol/L glucose were counted in a region of interest with an area of 3,254–4,992 μm² and normalized to an area of 800 μm². The number of sequential exocytic events was quantified as described previously (18).

All quantitative data are expressed as the mean ± SEM. The P values were calculated using the Student t test or a one-way ANOVA with a Dunnett multiple-comparison test.

The data sets and all noncommercially available resources generated and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Because Rab27a is localized on insulin granules and is involved in their exocytosis (19,20), genetic mutation of any Rab27 effector could potentially cause a defect in insulin secretion. In fact, leaden mice with a natural melanophilin mutation showed glucose intolerance compared with their control mice, although body weight and blood glucose levels in a fasting state or after an insulin load did not differ (Fig. 1A). Melanophilin was expressed in wild-type pancreatic islets (Fig. 1B). We were unable to examine the intracellular distribution of endogenous melanophilin due to lack of an antibody sufficiently durable for immunostaining. However, HA-tagged melanophilin expressed at an endogenous level in monolayer pancreatic β-cells was observed to colocalize with insulin granules, especially those at the cell periphery (Fig. 1C and D). Although melanophilin is known to interact with the actin motor protein, myosin-Va, and/or actin itself (7–10,21), its presence or absence did not affect insulin granule accumulation in the cell corners, where F-actin and myosin-Va were enriched (Supplementary Fig. 1). Electron microscopic analyses also revealed no significant changes in the number, density, or distribution of granules, including the number of docked granules having centers residing within 200 nm of the plasma membrane in leaden β-cells (Supplementary Fig. 2). Perifusion assays of leaden islets revealed a significant decrease in insulin secretion in response to glucose or a stronger stimulus, glucose plus forskolin (Fig. 1E).

After confirming that most of the monolayer cells derived from wild-type and leaden pancreatic islets are β-cells by insulin immunostaining (90.1% and 89.7%, respectively, n = 3 each), we compared fusion profiles in living cells under TIRF microscopy that visualizes Insulin-EGFP–labeled granules just beneath the plasma membrane. Fused granules were categorized into three types as described previously (15,16): residents, granules visible before stimulation; visitors, granules visualized during stimulation; and passengers, granules invisible before fusion. Despite the different prefusion behaviors, the average peak fluorescence intensity during fusion was similar among the three types (15) (see also Fig. 2D for the case of human β-cells), suggesting that they all represent single-granule exocytosis. The total number of fusion events, especially in the late phase of glucose stimulation, was markedly decreased in leaden β-cells (Fig. 2A–C), consistent with findings from the perifusion analysis (Fig. 1E). Remarkably, only the passenger type was decreased by half. Although this type of exocytosis without stable predocking to the plasma membrane has consistently been observed in rodent β-cells expressing Insulin-EGFP (15,22) or Insulin-Venus (23) and in human β-cells expressing NPY-EGFP (24–26), its presence has recently been disputed in human β-cells expressing NPY-mCherry (27). However, because the passenger type is visible only in one frame (103 ms) as a flash by granule neutralization during the fusion, it might be difficult to find whether granules are visualized by pH-insensitive mCherry (28). In fact, in contrast to the granules labeled by Insulin-EGFP, those labeled by NPY-mCherry showed no fluorescence intensity peak even during the resident type of fusion in mouse β-cells (Supplementary Video 1 and Supplementary Fig. 3A). Accordingly, they revealed almost no passenger type but similar numbers of other types (Supplementary Fig. 3B). Thus, the discrepancy should not have arisen from a difference in animal species of β-cells but from a difference in the fluorescent proteins used to visualize the granules. We confirmed in human β-cells that the passenger type of exocytosis occurs at a frequency of ∼40% during 20-min glucose stimulation, although the visitor type of exocytosis is hardly seen and that knockdown of human melanophilin selectively decreases passenger-type exocytic events (Fig. 2D and E, Supplementary Videos 2 and 3, and Supplementary Fig. 4). We then expressed mCherry-melanophilin to label passengers in β-cells. However, mCherry-melanophilin showed too weak fluorescence to monitor granule exocytosis if expressed at an endogenous level and aberrant localization along F-actin if overexpressed (Supplementary Fig. 5A). Nevertheless, immunostaining with anti–red fluorescent protein antibody in cells fixed after 20-min glucose stimulation revealed that mCherry-melanophilin expressed at an endogenous level exists at the site of passenger exocytosis, especially where the exocytosis has occurred at later time points, and thus, the fused granules and associated proteins likely remain at the plasma membrane (Supplementary Fig. 5B). Although EGFP-melanophilin exhibited stronger fluorescence and colocalized with NPY-mCherry (Supplementary Fig. 6A), this set of fluorescent proteins cannot be used to detect passenger exocytosis as described. Further, EGFP-melanophilin colocalized with coexpressed Kusabira-Orange 1–granuphilin, another Rab27 effector mediating resident type of exocytosis (29). We confirmed that HA-melanophilin colocalizes with endogenous granuphilin under confocal microscopy (Supplementary Fig. 6B). These findings indicate that a single granule can carry different Rab27 effectors simultaneously and that melanophilin is not specifically located on granules showing passenger exocytosis.

The passenger type might correspond to sequential exocytosis wherein granules fuse selectively with other granules that have already fused with the plasma membrane, although the frequency of sequential exocytosis is minor (2–3%) in mouse pancreatic islets (18) compared with that of passenger exocytosis found in mouse monolayer β-cells (∼40%) (Fig. 2C). To exclude this possibility, we performed two-photon excitation imaging of wild-type and leaden mouse islets. Although the number of exocytic events during the first 5 min of glucose stimulation was not different, the number of exocytic events during the 5–15 min stimulation was markedly reduced in leaden islets (Fig. 3A and B). This time-dependent difference may correspond to current and previous TIRF microscopic findings (15,22,23) (Fig. 2A) that the relative frequency of the passenger type of exocytosis increases in the late phase of glucose-stimulated insulin secretion (GSIS). However, the frequency of sequential exocytosis did not differ between wild-type and leaden islets (3.3 ± 0.6% vs. 3.1 ± 1.6%) (Fig. 3C), indicating that melanophilin is not involved in sequential exocytosis.

To identify the molecular mechanism by which melanophilin regulates granule exocytosis, we first compared the expression levels of proteins known to interact with melanophilin or to function in insulin granule exocytosis between wild-type and leaden islets. We found that syntaxin-4, a member of the soluble N-ethylmaleimide–sensitive factor attachment protein receptors (SNAREs), was expressed at significantly lower levels in leaden islets (Supplementary Fig. 7). Because protein expression levels often decline with loss of an interacting protein and because syntaxin-4 is known to function in insulin granule exocytosis (25,30), we explored the possibility that melanophilin interacts with syntaxin-4. Melanophilin exogenously expressed in the pancreatic β-cell line MIN6 coprecipitated syntaxin-4 as well as previously known melanophilin-interacting proteins, such as Rab27a, myosin-Va, β-actin, and EB1 (Fig. 4A). In contrast, melanophilin did not interact with syntaxin-1a, -2, or -3. We confirmed that melanophilin forms an endogenous complex with syntaxin-4 in pancreatic islets (Fig. 4B).

This newly identified interaction might be mediated indirectly via actin, because both proteins have been reported to bind actin directly (21,31). To investigate this possibility, we first determined the interacting regions in these two proteins. The cytoplasmic region of syntaxin-4 consists of the N-terminal Habc domain and the COOH-terminal H3 domain containing a SNARE motif. When each domain without the transmembrane domain was coexpressed with melanophilin in HEK293A cells, only the H3 domain formed a complex (Fig. 4C). On the other hand, melanophilin comprises the N-terminal Rab27a binding domain (RBD), the central myosin-Va binding domain (MBD), and the C-terminal actin binding domain (ABD). We found that the N-terminal protein covering the RBD and the MBD interacts with syntaxin-4, whereas the C-terminal protein covering the MBD and the ABD does not (Fig. 4D, left). Furthermore, although the MBD domain only (147–400 amino acid residues) or that with additional N-terminal residues (126–400) could not bind syntaxin-4, the MBD with further N-terminal residues (116–400) was able to bind it (Fig. 4D, right). Because melanophilin binds actin through the ABD (401–590 residues) (21) and because syntaxin-4 interacts with actin through the Hab domains (39–112 residues) (31), the regions responsible for the interaction between melanophilin and syntaxin-4 are completely different from the actin-binding region of each protein. Therefore, it is unlikely that the melanophilin–syntaxin-4 complex is indirectly mediated through the actin binding. We confirmed that the N-terminal melanophilin (1–400 residues) bound syntaxin-4 without involvement of other specific proteins in HEK293A cells (Fig. 4E). Furthermore, the interaction between melanophilin (1–146 residues) and syntaxin-4 (1–273 residues) was observed both in MIN6 cell extracts (Fig. 4F) and between bacterially expressed, purified proteins (Fig. 4G and Supplementary Fig. 8).

These findings indicate that residues 116–125 of melanophilin are crucial for the interaction with syntaxin-4. In fact, the N-terminal 1–130 residues, but not the 1–120 residues, could efficiently bind syntaxin-4 (Fig. 4H). The generation of point mutations around this region revealed that the Y122A/H124A/K130A triple mutant exhibits a greatly reduced syntaxin-4–binding activity (Fig. 4I). The 117–122 residues, SLEWYY, of melanophilin correspond to the SGAWFF structural element of rabphilin that interacts with a deep pocket in Rab3a in the crystal structure (32). Furthermore, the Y121 of melanophilin has been shown to make a hydrogen bond with the R90 carbonyl group of Rab27b (33). Therefore, the triple mutant might also affect the interaction with Rab27a. However, the mutant had a Rab27a-binding activity comparable to that of full-length melanophilin (Fig. 4I), which is consistent with previous findings that either the W120A/Y121A mutation (8) or the broader deletion of 111–145 residues (34) of melanophilin does not affect the interaction with Rab27a. Thus, the Y122A/H124A/K130A mutant specifically loses binding activity to syntaxin-4.

To examine whether the interaction with syntaxin-4 is important for melanophilin’s promotion of insulin granule exocytosis, we first confirmed that the Y122A/H124A/130A mutant loses binding activity to endogenous syntaxin-4, but not to Rab27a or myosin-Va, in MIN6 cells (Fig. 5A). In contrast, the E14A and D378A/E380A/E381A/E382A/A467P mutants selectively disrupted the interaction with Rab27a and myosin-Va, respectively, as previously reported (8,21,35) but not that with syntaxin-4. Immunostaining revealed that the Y122A/H124A/130A mutant expressed in leaden β-cells does not colocalize with syntaxin-4 along the plasma membrane but is only distributed to the F-actin–rich cell corners (Fig. 5B). In contrast, the E14A mutant did not colocalize with the granule-resident Rab27a at all, whereas the D378A/E380A/E381A/E382A/A467P mutant did not colocalize with myosin-Va accumulated at the cell corners, but located only along the plasma membrane. We then performed rescue experiments by introducing wild-type or mutant melanophilin in leaden islets to match the level of endogenous melanophilin in wild-type islets (Fig. 5C). The wild-type melanophilin tended to increase the expression level of syntaxin-4 in leaden islets (Fig. 5C, upper right) and restored GSIS to the level found in wild-type islets expressing the control LacZ protein (Fig. 5D). In contrast, the E14A mutant did not restore insulin secretion at all, confirming the importance of melanophilin’s association with the granule membrane via Rab27a. Neither the Y122A/H124A/K130A nor the D378A/E380A/E381A/E382A/A467P mutant exhibited an enhancement, suggesting that the interactions with syntaxin-4 and myosin-Va are also important. To obtain further specific evidence, we performed rescue experiments in monolayer β-cells under TIRF microscopy (Fig. 5E). Leaden β-cells expressing wild-type melanophilin showed a specific increase in the number of passenger type exocytic events in response to glucose stimulation. In contrast, those expressing the Y122A/H124A/K130A or the D378A/E380A/E381A/E382A/A467P mutant did not exhibit such an increase.

To explore the mechanism underlying melanophilin’s acceleration of exocytosis without a significant presence beneath the plasma membrane, we investigated the mode of interaction between melanophilin and syntaxin-4. Syntaxin members are thought to exist in an equilibrium between the open and closed conformations, and only the open form is capable of forming a trans-SNARE complex with other SNARE proteins to mediate a fusion reaction (36,37). The syntaxin-4 mutant L173A/E174A corresponds to the syntaxin-1a mutant L165A/E166A that adopts a constitutively open conformation (38). We found that exogenously expressed melanophilin interacts with wild-type and L173A/E174A syntaxin-4 similarly in HEK293A cells (Fig. 6A), suggesting that it preferentially binds the fusion-competent, open form. This binding mode could account for the passenger type of instant fusion seen after granules approach the plasma membrane. Importantly, the interaction of melanophilin with syntaxin-4 and its SNARE partners, VAMP2 and SNAP25, was induced by glucose stimulation, and this induction disappeared in the simultaneous presence of the Ca²⁺-chelator EGTA, in MIN6 cells (Fig. 6B and C). This finding suggests that the complex forms only after the stimulus-induced, intracellular Ca²⁺ increase and is consistent with its role in mediating the passenger type of exocytosis wherein granules are recruited to the plasma membrane only after stimulation. This stimulus-dependent binding was similarly observed for the L173A/E174A mutant (Fig. 6D), suggesting that Ca²⁺ does not directly alter the conformation of syntaxin-4 but allows melanophilin-positive granules to reach syntaxin-4 on the plasma membrane by breaking down the F-actin network. Consistent with this idea, melanophilin promptly but transiently dissociated from myosin-Va and actin upon glucose stimulation in a Ca²⁺-dependent manner (Fig. 6E and F).

Under TIRF microscopy visualizing secretory granules using pH-sensitive fluorescent protein, we clearly showed that granules exhibit heterogeneous prefusion behaviors in both mouse and human pancreatic β-cells. The majority of the resident type of exocytosis likely corresponds to the mode of fusion preceded by stable docking to the plasma membrane (29), and several molecules, such as granuphilin and RIM2, have been shown to tether and/or dock granules to the plasma membrane (14,39). In contrast, the molecular mechanism of the latter passenger (undocked) type of exocytosis has not been explored well. The previous findings that syntaxin-4, Munc18-3, and synaptotagmin-7 are involved in both types of exocytosis (24–26) indicate that the passenger type also involves such general exocytic machinery components and represents a real exocytic phenomenon. However, the molecular mechanism unique to this type has still been enigmatic. Given that this type of exocytosis becomes dominant in the late phase of GSIS, it should help sustain granule exocytosis in a prolonged time. Rab27 effectors, such as melanophilin and exophilin-8, are good candidates to form such a reserve pool of granules, because they can link Rab27 on the granule membrane and myosin motors within the cell peripheral actin network. We recently showed that silencing of each component of the exophilin-8–RIM-BP2–myosin-VIIa complex markedly decreases the peripheral accumulation and exocytosis of insulin granules (2). In contrast, loss of melanophilin (present study) or silencing of melanophilin-interacting myosin-Va (2) decreases GSIS but does not affect the granule accumulation at the cell periphery. These findings suggest that exophilin-8 first acts to capture granules within a relatively broad area of the actin cortex (2,40), whereas melanophilin and myosin-Va then function to promote granule exocytosis below the plasma membrane. Indeed, melanophilin mediates the passenger type of exocytosis via its interactions with myosin-Va and syntaxin-4a. Myosin-Va unlikely functions as an active motor for granule movement, because granules showing this type of exocytosis must pass the evanescent field (100–200 nm) per 100 ms (one frame of TIRF microscopy) before fusion, the velocity of which is above its motor speed. A faster kinesin motor may work in their recruitment to the plasma membrane. In any case, those granules should not have located in a deep cell interior, but just above the evanescent field, still close to the plasma membrane, before they fuse in response to external stimulation.

Because granules showing the passenger type of exocytosis are stimulus-dependently recruited from the cell interior and mediate instant fusion without pausing beneath the plasma membrane, they must promptly assemble the fusion machinery after stimulation (41,42). The stimulus-dependent dissociation of melanophilin from myosin-Va would allow granules to be released from the F-actin network and to associate with fusion machinery, syntaxin-4, on the plasma membrane. Further, the specific interaction of melanophilin with the fusion-competent, open form of syntaxin-4 and its SNARE partners in cells after stimulation, should enable granules to fuse instantly with the plasma membrane. In this context, it is interesting that syntaxin-4 interacts with the Ca²⁺-activated F-actin–serving protein, gelsolin and that the complex-dissociating action of secretagogues can induce the open form of syntaxin-4 in MIN6 cells (43), which might facilitate the complex formation between melanophilin and syntaxin-4 at actin filament tips adjacent to the plasma membrane (44). Alternatively, SNAREs in native plasma membranes are constitutively active even if not engaged in fusion events, as previously shown (45). Although the interaction of melanophilin with syntaxin-4 is reminiscent of that of another Rab27 effector, granuphilin, with syntaxins-1a, -2 and -3 (13,14,46), granuphilin specifically interacts with the fusion-incompetent, closed-form of syntaxin-1a in the absence of Ca²⁺ (47) and exclusively mediates exocytosis of fusion-reluctant, stably docked granules (14,15,29). Because melanophilin and granuphilin coexist on almost all of the granules in β-cells, the heterogeneous modes of granule exocytosis observed in living cells should not be determined by the existence of specific Rab27 effector per se but may reflect the existence of distinct modes of interaction between specific tethering factors (Rab27 effectors) and fusion machinery (syntaxins).

In summary, melanophilin links granules within the peripheral actin network via interactions with Rab27a and myosin-Va, dissociates them from it by stimulus-induced F-actin dissolution, and then interacts with the open form of syntaxin-4 to induce immediate fusion. Although loss of melanophilin does not abolish the passenger type of exocytosis, that which remains in leaden β-cells may be mediated by other Rab27 effectors, such as exophilin-7 (16) and exophilin-8 (2,40), or may correspond to sequential exocytosis. Parallel and/or redundant exocytic pathways involving predocked and undocked granules would guarantee robustness to the secretory process that produces many indispensable components, including insulin.

Acknowledgments. The authors thank Prof. P.E. MacDonald (Alberta Diabetes Institute IsletCore) for supplying human pancreatic islets, Prof. A. Miyawaki (RIKEN Brain Science Institute) for providing the pNPY-Venus-N1 vector, H. Kobayashi (Gunma University) for providing guinea pig anti-porcine insulin serum, and the members of Laboratory of Molecular Endocrinology and Metabolism, especially Dr. K. Matsunaga for useful suggestion and discussion, T. Nara and T. Ushigome for their colony management of mice, and S. Shigoka and J. Toshima for assistance in preparing the manuscript.

Funding. This work was supported by Japan Society for the Promotion of Science KAKENHI grants JP18K14647 and JP20K15742 to H.W., JP26670133, JP14F04104, and JP16K15211 to T.I., and Japan Science and Technology Agency (JST)-Core Research for Evolutional Science and Technology (CREST) grant JPMJCR1652 to H.K. It was also supported by grants from Uehara Memorial Foundation (to T.I.), Kobayashi International Scholarship Foundation (to T.I.), Novartis Research Grants, Pfizer Academic Contributions, Astellas Research Support, MSD Scholarship Donation, and Sanofi Scholarship Donation to T.I.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. H.W. and K.M. performed experiments and analyzed data. N.T. performed two-photon excitation microscopy and analyzed data. E.K. performed electron microscopy. J.S. and Y.T. provided experimental reagents. H.K. and K.O. analyzed data. T.I. designed experiments and wrote the paper. T.I. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in abstract form at the 55th Annual Meeting of the European Association for the Study of Diabetes, Barcelona, Spain, 16–20 September 2019.