Exocytosis of secretory granules entails budding from the trans-Golgi network, sorting and maturation of cargo proteins, and trafficking and fusion to the plasma membrane. Rab27a regulates the late steps in this process, such as granule recruitment to the fusion site, whereas Rab2a functions in the early steps, such as granule biogenesis and maturation. Here, we demonstrate that these two small GTPases simultaneously bind to Noc2 (also known as RPH3AL) in a GTP-dependent manner, although Rab2a binds only after Rab27a has bound. In pancreatic β-cells, the ternary Rab2a–Noc2–Rab27a complex specifically localizes on perinuclear immature granules, whereas the binary Noc2–Rab27a complex localizes on peripheral mature granules. In contrast to the wild type, Noc2 mutants defective in binding to Rab2a or Rab27a fail to promote glucose-stimulated insulin secretion. Although knockdown of any component of the ternary complex markedly inhibits insulin secretion, only knockdown of Rab2a or Noc2, and not that of Rab27a, impairs cargo processing from proinsulin to insulin. These results suggest that the dual effector, Noc2, regulates the transition from Rab2a-mediated granule biogenesis to Rab27a-mediated granule exocytosis.
Regulated secretion is a main pathway in the delivery of the bioactive molecules of a cell to the extracellular environment. The pathway comprises coordinated sequential steps, such as secretory vesicle biogenesis, maturation, trafficking and fusion with the plasma membrane. Although the molecular machinery for these individual processes has been characterized, the precise mechanisms connecting each process remain poorly understood. Previous studies have shown that the small GTPase Rab27 (of which there are two isoforms, Rab27a and Rab27b), regulates the late steps of this pathway through its multiple effector proteins (Izumi et al., 2003; Fukuda, 2006). For example, in pancreatic β-cells, Rab27a localizes on insulin granules (Yi et al., 2002) and forms a complex with its effectors, such as granuphilin (also known as Slp4) (Wang et al., 1999; Gomi et al., 2005), exophilin 7 (also known as JFC1, Slp1 and SYTL1) (Wang et al., 2013), exophilin 8 (also known as MyRIP and Slac2c) (Waselle et al., 2003; Mizuno et al., 2011) and Noc2 (also known as RPH3AL) (Kotake et al., 1997; Cheviet et al., 2004), and it regulates a specific step of insulin granule trafficking and exocytosis. Namely, granuphilin and exophilin 7 control exocytosis of granules that are docked or undocked to the plasma membrane, respectively, whereas exophilin 8 retains granules in the cortical actin network for subsequent release. However, it remains unknown at which step or by what mechanism Noc2 functions, despite the finding that Noc2-knockout mice exhibit impaired insulin secretion when under acute stress (Matsumoto et al., 2004). This is partly because Noc2 is a relatively small protein compared with other Rab27 effectors, and it appears to lack functional domains other than the Rab27-binding domain. In the present study, we show that Noc2 binds to another GTPase, Rab2a, in addition to Rab27a. The ternary Rab2a–Noc2–Rab27a complex specifically localizes on immature granules, and interference in the complex formation inhibits cargo processing and granule exocytosis. We present evidence that this new complex regulates the transition from Rab2a-mediated granule biogenesis to Rab27a-mediated granule exocytosis in the regulated secretory pathway.
Rab2a binds to Rab27a via Noc2 in a GTP-dependent manner
To identify Rab27a-interacting proteins more comprehensively, we stably expressed Rab27a in the β-cell line MIN6, that was tagged with MEF (Myc-TEV-FLAG), which consists of Myc and FLAG epitope tags connected by a spacer sequence containing a TEV protease cleavage site, and then performed tandem affinity purification (Ichimura et al., 2005; Matsunaga et al., 2009). The protein bands specific to the Rab27a immunoprecipitate were then analyzed by a liquid chromatography (LC)-tandem mass spectrometry (MS/MS) system (Fig. S1A; Table S1). In addition to the known Rab27-interacting proteins, such as granuphilin, exophilin 9 (also known as Slp5 and SYTL5) (Kuroda et al., 2002), and Noc2, two Rab GTPases, Rab2 (which has an a and b isoform) and Rab18, were also identified. Co-immunoprecipitation experiments showed that Rab2a, but not Rab18, interacts with Rab27a in MIN6 cells (Fig. S1C). Because the Rab2a-immunoprecipitate also contained Noc2, but not granuphilin (Fig. S1C,D), we performed similar tandem purification in MIN6 cells stably expressing MEF–Noc2, and found Rab2a to be a Noc2-interacting protein, as is Rab27a (Fig. S1B; Table S2). Furthermore, Noc2 formed an endogenous complex with Rab2a in both of the β-cell lines, mouse MIN6 and rat INS1 832/13 (Fig. 1A,B). Noc2 specifically bound to the Rab2a Q65L mutant mimicking the GTP-bound form, but not with the S20N mutant mimicking the GDP-bound form (Fig. 1C). As previously reported (Fukuda et al., 2004), Noc2 interacted with the corresponding Rab27a Q78L mutant, but not with the T23N mutant (Fig. S1E). Although Noc2 also forms a complex with Rab3a (Fig. 1B), as previously reported (Haynes et al., 2001; Cheviet et al., 2004), no Rab3a was present in either the Rab2a or Rab27a immunoprecipitate (Fig. 1C; Fig. S1E). Furthermore, neither the established Rab2a effector ICA69 (also known as ICA1) (Buffa et al., 2008) nor its interacting protein PICK1 (Cao et al., 2007) were found in the Noc2 or Rab27a immunoprecipitate, in contrast to in the Rab2a immunoprecipitate (Fig. 1B,C; Fig. S1E). When Noc2 was downregulated by specific short-hairpin RNA (shRNA), the interaction between Rab2a and Rab27a disappeared (Fig. 1D). These findings indicate that the two Rab proteins interact through Noc2, and that Noc2 simultaneously binds to Rab2a and Rab27a. Furthermore, the ternary Rab2a–Noc2–Rab27a complex appears to exist separately from either the Rab2a–ICA69–PICK1 complex or the Noc2–Rab3a complex.
Rab2a binds to Noc2 only in the presence of Rab27a
We next investigated the formation of the ternary complex. Coexpression of Rab2a and Noc2 in HEK293A cells did not lead to a complex formation between the two proteins, although additional expression of Rab27a did (Fig. 2A). Furthermore, the Noc2 E51A/I55A double mutation known to dramatically reduce the Rab27a-binding activity (Fukuda et al., 2004) simultaneously lost its Rab2a-binding activity in MIN6 cells (Fig. 2B). Furthermore, in contrast to in wild-type β-cell lines (Fig. 1), the Rab2a–Noc2 complex was absent in Rab27a-null β-cells derived from ashen mice (Wilson et al., 2000); however, the complex was present after the introduction of wild-type Rab27a into those cells (Fig. 2C). These findings indicate that Rab2a interacts with Noc2 only in the presence of Rab27a, and probably after Rab27a binds to Noc2. To substantiate this conclusion, we simultaneously expressed different amounts of Rab2a and Rab27a in HEK293A cells expressing One-STrEP-Flag (OSF)-tagged Noc2. Noc2 and the binding proteins were pulled down using Strep-Tactin beads and were subjected to SDS-PAGE and Coomassie Brilliant Blue (CBB) staining (Fig. 2D). There were no specific protein bands other than Rab2a, Rab27a and Noc2. Furthermore, the amount of Rab2a bound to Noc2 was proportional to that of Rab27a bound to Noc2. These findings eliminate the possibility that other proteins are involved in the ternary complex formation and that Rab2a and Rab27a competitively interact with Noc2.
Rab2a and Rab27a bind to Noc2 through distinct N-terminal regions
We next determined the Rab2a-interacting domain of Noc2. A series of Noc2 deletion mutants were expressed as bait in MIN6 cells (Fig. 3A). Rab27 effectors, including Noc2, possess a highly conserved N-terminal Rab27-binding domain, named the RBD (Izumi et al., 2003). Indeed, Noc2 mutants devoid of their RBD (residues 41–158) were unable to bind to either Rab27a or Rab2a (Fig. 3B,C). However, residues further towards the N-terminal of the RBD were specifically required for the binding to Rab2a (Fig. 3C,D). In fact, the minimum deletion mutant Δ(11-20) showed a marked decrease in binding activity to Rab2a, but not to Rab27a (Fig. 3E). These findings indicate that the two Rab proteins interact with Noc2 through different N-terminal regions.
Site-directed mutagenesis analysis in the N-terminus region of Noc2 revealed that mutation of threonine 38, but not that of glutamine 12, specifically eliminated the binding activity to Rab2a (Fig. 3E,F). We explored the possibility that phosphorylation at this threonine residue might be involved in the binding to Rab2a, because endogenous Noc2 proteins in β-cells were detected as multiple bands in gels (Figs 1D, 2C). However, the phosphomimetic mutants, T38D and T38E, failed to interact with Rab2a (Fig. 3F). Furthermore, although in vitro phosphatase treatment of the Noc2 immunoprecipitate did increase the gel mobility of Noc2 protein, it failed to influence the interaction with either Rab2a or Rab27a (Fig. S2), suggesting that phosphorylation of Noc2 plays no role in the formation of the ternary complex.
The Rab2a–Noc2–Rab27a complex localizes on immature proinsulin granules
We next investigated the intracellular localization of Rab27a, Rab2a and Noc2 in INS1 832/13 cells. Because available antibodies to Rab2a and Noc2 are not durable for immunostaining, we examined the localization of exogenously expressed proteins with an N-terminal MEF tag. Although Rab27a and Noc2 colocalized with insulin-positive puncta throughout the cytoplasm, Rab2a did not, and was instead restricted to the perinuclear region (Fig. 4A). However, MEF–Rab2a almost completely colocalized with proinsulin-positive puncta in the perinuclear region, although MEF–Rab27a and MEF–Noc2 also colocalized there (Fig. 4B). The insulin- and proinsulin-containing puncta did not colocalize, indicating that the anti-insulin and anti-proinsulin antibodies hardly crossreacted to the other protein under our experimental condition (Fig. S3A). Furthermore, none of the proteins colocalized with the endosome marker, EEA1, the trans-Golgi network (TGN) marker TGN38 (also known as TGOLN2), or the endoplasmic reticulum marker PDI (also known as P4HB) (Fig. S3B). These findings suggest that the Rab2a–Noc2–Rab27a ternary complex is specifically located on immature granules, whereas the Noc2–Rab27a binary complex is located on mature granules.
In contrast to the wild-type Noc2, the E51A/I55A mutant defective in binding to Rab27a neither localized on insulin-positive mature nor proinsulin-positive immature granules (Fig. 5). However, the Noc2 mutants specifically defective in binding to Rab2a, Δ(11-20) or T38A, still localized to both mature and immature granules. These results indicate that the interaction with Rab27a, but not that with Rab2a, determines the localization of Noc2 on secretory granules.
Rab2a–Noc2–Rab27a complex formation is essential for insulin granule exocytosis
To investigate the functional roles of the Rab2a–Noc2–Rab27a complex, we expressed wild-type Noc2 or its mutants defective in binding to these Rab proteins in INS1 832/13 cells, and examined the effects on insulin secretion (Fig. 6A). The cells expressing exogenous wild-type Noc2 at similar level to the endogenous protein showed a 1.5-fold higher glucose-stimulated insulin secretion compared with control cells (Fig. 6B). By contrast, none of the Noc2 mutants, Δ(11-20), T38A, or E51A/I55A, showed such enhancement. The insulin content was not affected by expression of either wild-type or mutant Noc2 (Fig. 6C). These results suggest that the ternary complex formation is instrumental for evoked granule exocytosis.
We next investigated the effects of downregulation of each component of the complex by adenovirus-mediated shRNA expression (Fig. 7A). Surprisingly, Rab2a knockdown simultaneously downregulated Noc2, despite the presence of Rab27a. This is not an off-targeting effect, because independent knockdown using double-stranded small interfering RNA (siRNA) against a different sequence of Rab2a also led to a markedly decreased expression of Noc2 (Fig. S4A). Because Noc2 stably associates with Rab27a on peripheral mature granules without Rab2a (Fig. 4A), and because Noc2 mutants defective in Rab2a binding still locate on perinuclear immature granules (Fig. 5B), the instability of Noc2 in the absence of Rab2a requires more than the loss of the interaction between the two proteins. Rather, the impairment of granule biogenesis upon Rab2a depletion appears to disrupt the membrane-association site that causes nascent Noc2 to be stabilized. Consistent with this idea, knockdown of ICA69, which is thought to function with Rab2a in the early stage of secretory granule exocytosis, such as in granule biogenesis and maturation (Sumakovic et al., 2009; Cao et al., 2013), dislodges Noc2 from immature granules (Fig. S4D).
Depletion of any component of the complex markedly inhibited glucose-stimulated insulin secretion (Fig. 7B; Fig. S4B). Because insulin secretion could be impaired by inhibition of insulin or granule synthesis, we also measured total insulin content in the cells (Fig. 7C; Fig. S4C). There were notable differences among the cells: Rab27a-knockdown markedly increased insulin content, whereas Rab2a knockdown markedly decreased it. Although the effect of ICA69 depletion was similar to that of Rab2a knockdown, Noc2-depletion caused an intermediate effect between those of the Rab27a and Rab2a depletions. Insulin secretion levels normalized to the total insulin content in cells confirmed that there was an inhibition of exocytosis in all the downregulated cells (Fig. 7D), but Rab27a-depletion appeared to primarily affect granule exocytosis because it increased insulin content, whereas Rab2a depletion likely impaired granule biogenesis and/or maturation because it decreased insulin content. The intermediate phenotypes of Noc2-depleted cells suggest that Noc2 plays a regulatory role in the transition between Rab2a- and Rab27a-mediated processes. Because the Rab2a–Noc2–Rab27a complex should specifically localize on immature proinsulin granules (Fig. 4), we also examined proinsulin secretion and content (Fig. 7E–G). Knockdown of Rab2a, ICA69 or Noc2, but not that of Rab27a, significantly increased the amount of proinsulin secreted in the medium, and the relative ratio between proinsulin and insulin levels in the cells. These results suggest that Noc2, as well as Rab2a and ICA69, is involved in granule maturation and cargo processing.
The present findings indicate that Rab2a and Rab27a bind together through the Noc2 protein. The mode of the complex formation between these three proteins appears to be unique. Although some Rab effector proteins are known to interact with different Rab proteins, such as RUFY1 (Yamamoto et al., 2010), nischarin (Kuij et al., 2013) and golgin family proteins (Gillingham and Munro, 2016), they do so separately or sequentially. There is no reported instance in which the same effector molecule simultaneously binds two different Rab proteins. Although the Rab2a-binding region locates towards the N-terminus further from the Rab27-binding domain, the GTP-bound Rab2a cannot bind to the Noc2 mutant that is incapable of binding to Rab27a, yet can bind to wild-type Noc2 in the presence of Rab27a. These findings indicate that the ternary complex forms after the formation of the binary Noc2–Rab27a complex.
The interaction between Rab2a and Rab27a is intriguing because they have been shown to be involved in early and late stages of secretory granule exocytosis, respectively. In Caenorhabditis elegans, a mutation in UNC-108 (the orthologue of Rab2) or in RIC-19 (the orthologue of ICA69) is unable to prevent cargos of secretory granules from inappropriately entering endosomal compartments during granule maturation (Edwards et al., 2009; Sumakovic et al., 2009). Rab2a forms a complex with ICA69 and PICK1, both of which contain a BAR domain, which is known to bind to lipid membranes and to initiate vesicle formation (Peter et al., 2004). Furthermore, mice deficient in these proteins exhibit defects in granule biogenesis and/or maturation (Cao et al., 2013; Holst et al., 2013). On the other hand, Rab27a associates with secretory granules (Yi et al., 2002) and regulates the late steps of granule exocytosis, such as recruitment and/or docking to fusion sites, through its multiple effector proteins (Izumi, 2007, 2011). Consistent with this, we found that knockdown of Rab2a or ICA69 inhibits insulin synthesis and processing, whereas that of Rab27a primarily affects insulin secretion. The intermediate phenotype caused by knockdown of Noc2 suggests that Noc2 plays a regulatory role in connecting the early and late exocytic pathways by forming a complex with both Rab2a and Rab27a.
Consistent with this model, Rab2a forms a complex with Noc2 and Rab27a, separately from that with ICA69 and PICK1. Although Noc2 and Rab27a locate on both perinuclear immature and peripheral mature granules, Rab2a is restricted to immature granules. ICA69 and PICK1 are also known to localize on immature granules and/or granule budding sites at the TGN (Spitzenberger et al., 2003; Cao et al., 2013; Holst et al., 2013). Interestingly, knockdown of ICA69 dislodges Noc2 from immature granules, and knockdown of Rab2a can eliminate Noc2 expression, which suggests that the two proteins function earlier than does Noc2. The differential effects between Rab2a and ICA69 knockdowns likely reflect the presence of other Rab2 effectors. It has recently been reported that the two other Rab2 effectors, RUND-1 and CCCP-1, function in granule biogenesis and maturation in C. elegans (Ailion et al., 2014). RUND-1 binds RIC-19 (a ICA69 ortholog), whereas CCCP-1 does not bind it and forms a different complex with Rab2. Therefore, Rab2 seems to function in parallel pathways through its multiple effectors, and its absence likely induces more profound effects on granule biogenesis than that of its effector ICA69. In fact, Rab2a knockdown has a more severe effect on insulin content in cells than ICA69 knockdown. Furthermore, Rab2a knockdown decreases the amount of proinsulin in cells, as was found in Rab2 (UNC-108)-mutated C. elegans where granule cargos inappropriately enter endosomal compartments for degradation (Edwards et al., 2009; Sumakovic et al., 2009). By contrast, ICA69 knockdown increases the amount of proinsulin in cells, as was found in ICA69-knockout β-cells where the exit of proinsulin from TGN is blocked (Cao et al., 2013). Despite these differences, the absence of Rab2a or ICA69 appears to prevent granules from being generated or becoming mature enough for nascent Noc2 to associate with them. Taken together, these findings suggest that Rab2a may first regulate granule budding and/or proper cargo sorting at the TGN, through an interaction with ICA69 and other effectors, and then regulates granule maturation and cargo processing by forming a complex with Noc2 bound to Rab27a on immature granules. Further research is required to identify the mechanism and timing of Rab2a dissociation from mature granules, although we suspect that the process is unlikely to involve the phosphorylation of Noc2.
In contrast to for the wild type, Noc2 mutants defective in binding to Rab2a or Rab27a fail to enhance glucose-stimulated insulin secretion, indicating that the ternary complex formation is pivotal for evoked granule exocytosis. The phenotype of Noc2-knockout mice, which have impaired insulin secretion under condition of acute stress (Matsumoto et al., 2004), may represent imbalance between granule maturation and exocytosis. The complex may also play a role in the pathogenesis of human type 2 diabetes, because these patients exhibit a disproportionate level of circulating proinsulin (Røder et al., 1998). However, the complex is not specific to β-cells, because we also found it by similar tandem purification and LC-MS/MS analyses in pancreatic α-cell and melanocyte cell lines (our unpublished observations), both of which employ a Rab27a-regulated exocytic system (Bahadoran et al., 2001; Hume et al., 2001; Yu et al., 2007). Therefore, the present ternary complex may play a universally conserved role in secretory cells.
MATERIALS AND METHODS
All cells were cultured in a humidified incubator with 95% air and 5% CO2 at 37°C, MIN6 cells (Miyazaki et al., 1990) were cultured in Dulbecco's modified Eagle's medium (DMEM) containing 15% fetal bovine serum (FBS) supplemented with 1 mM L-glutamine and 50 µM 2-mercapthoethanol. MIN6 cells stably expressing MEF-tagged Rab27a, Rab2a or Noc2 were cultured in a medium containing 0.5 µg ml−1 of puromycin (Invivogen). INS1 832/13 cells (Hohmeier et al., 2000) were cultured in RPMI1640 containing 10% FBS supplemented with 1 mM L-glutamine, 1 mM HEPES, 1 mM sodium pyruvate and 50 µM 2-mercapthoethanol. HEK293A cells (Invitrogen) were cultured in DMEM containing 10% FBS supplemented with 1 mM L-glutamine. Rab27a-null β-cell lines were established from Rab27a-mutated ashen mice (Wilson et al., 2000), by a method similar to that by which granuphilin-null β-cell lines were previously established (Mizuno et al., 2016), and will be described in detail elsewhere. All animal experiments were performed in accordance with the rules and regulations of the Animal Care and Experimentation Committee, Gunma University.
Rabbit polyclonal antibodies against Rab27a and granuphilin were described previously (Yi et al., 2002) and were used at 1:1000 for immunoblotting. Guinea pig anti-porcine insulin serum was a gift from H. Kobayashi (Gunma University) and used at 1:1000 for immunofluorescence staining. Mouse anti-myc 9E10 monoclonal antibody was purified from the ascites fluid of a hybridoma-injected mouse and was used at 1:200 for MEF-tag immunoprecipitation. Commercially purchased antibodies against the following proteins were also used: rabbit polyclonal antibodies toward FLAG (1:2000 for immunoblotting and 1:500 for immunofluorescence staining; F7425, Sigma-Aldrich), hemagglutinin (HA; 1:2000 for immunoblotting; 561, MBL), GFP (1:2000 for immunoblotting; 598, MBL), Rab27a/b (1:2000 for immunoblotting; 18975, IBL), Noc2 (1:2000 for immunoblotting; 15297-1-AP, Proteintech), Rab2a (1:2000 for immunoblotting; 15420-1-AP, Proteintech), ICA69 (1:2000 for immunoblotting; ab81500, Abcam), and PICK1 (1:2000 for immunoblotting; ab3420, Abcam); and mouse monoclonal antibodies toward Rab3 (1:2000 for immunoblotting; 610379, BD Biosciences), EEA1 (1:100 for immunofluorescence staining; 610457, BD Biosciences), TGN38 (1:100 for immunofluorescence staining; 610849, BD Biosciences), PDI (1:100 for immunofluorescence staining; MA3-018, Affinity BioReagents), α-tubulin (1:3000 for immunoblotting; T5168, Sigma-Aldrich) and proinsulin (1:300 for immunofluorescence staining; clone 3A1; ab8301, Abcam).
Mouse Rab27a wild-type and mutant cDNAs were as described previously (Yi et al., 2002). Mouse Rab2a and Noc2 cDNAs were reverse transcribed from RNA of MIN6 cells. Point and deletion mutants of Rab2a and Noc2 were generated using a standard PCR-based mutagenesis strategy, and were verified by DNA sequencing. These cDNAs were subcloned into pENTR-3C (Invitrogen) or pMRX (Saitoh et al., 2003), and an MEF tag or FLAG tag sequence from the pcDNA3-MEF vector (Ichimura et al., 2005) was incorporated. They were also subcloned into the pCAG vector with or without an OSF tag (Morita et al., 2007). Subconfluent HEK293A cells were transfected with the plasmids using Lipofectamine 2000 reagent (Invitrogen). For generation of recombinant adenoviruses, the cDNAs of Rab27a, Rab2a, and Noc2 in pENTR-3C-MEF/FLAG were transferred into a pAd/CMV vector (Invitrogen) by LR Clonase recombination (Invitrogen). Adenoviral production and infection were performed according to the manufacturer's protocol.
Immunoprecipitation and immunoblotting
Cells were lysed in lysis buffer containing 50 mM Tris-HCl pH 7.5, 150 mM NaCl, 10% (w/v) glycerol, 100 mM NaF, 10 mM ethylene glycol tetraacetic acid, 1 mM Na3VO4, 1% Triton X-100, 5 µM ZnCl2, 1 mM phenylmethylsulfonyl fluoride and complete protease inhibitor cocktail (Roche). The lysates were cleared by centrifugation at 15,000 rpm (20630 g) for 15 min at 4°C. The supernatants were subjected to immunoprecipitation with primary antibody, and Protein G-Sepharose 4FF (GE Healthcare Bioscience), anti-FLAG resin (A2220; Sigma-Aldrich) or Strep-Tactin beads (GE Healthcare Bioscience). After being washed five times with TBS (50 mM Tris-HCl pH 7.5 and 150 mM NaCl) containing 0.1% Triton X-100, the immunoprecipitates were subjected to SDS-PAGE, and then transferred to a polyvinylidene difluoride membrane. The membrane was blocked with TBST (TBS plus 0.1% Tween-20) containing 0.5% nonfat dried milk powder, and then incubated overnight at room temperature with primary antibody diluted in Can Get Signal solution I (TOYOBO). It was then washed three times with TBST, was incubated for 1 h at room temperature with a 5000× dilution of horseradish peroxidase-conjugated secondary antibody (GE Healthcare Bioscience) in TBST containing 0.5% nonfat dried milk powder, and was washed five times. Immunoreactive signals were then detected using ECL prime and an LAS-4000 chemiluminescence detection system (GE Healthcare Bioscience).
MEF-tag-based protein purification and mass spectrometry
The purification procedure was similar to that reported previously (Ichimura et al., 2005; Matsunaga et al., 2009), with minor modifications. Briefly, ∼2×108 cells were lysed in 15 ml of lysis buffer. The interacting proteins were immunoprecipitated with anti-Myc antibody, cleaved by TEV protease (12575015; Invitrogen), re-immunoprecipitated with anti-FLAG antibody, and eluted by FLAG peptides. The final eluate was separated by SDS-PAGE and visualized by Oriole fluorescent gel staining (BioRad). Specific bands were excised and digested in gels with trypsin, and the resulting peptide mixtures were analyzed by LC-MS/MS. All MS/MS spectra were searched against the RefSeq protein sequence database at the National Center for Biotechnology Information using Mascot software (Matrix Science).
Immunofluorescence and microscopy
INS1 832/13 cells cultured on coverslips were fixed with 3% paraformaldehyde in phosphate buffered saline (PBS) for 30 min and permeabilized with 0.1% Triton X-100 in PBS for 30 min. The cells were then treated with 50 mM NH4Cl-PBS for 10 min at room temperature and blocked with PBS containing 1% bovine serum albumin (BSA) for 15 min. The coverslips were incubated with primary antibody overnight, washed three times with PBS, and incubated with Alexa Fluor 488- or 568-conjugated secondary antibody (Invitrogen; 1:500 dilution) for 60 min. Samples were washed five times and mounted using SlowFade Gold (Invitrogen). The microscopic images were obtained with a Fluoview FV1000 (Olympus) confocal laser scanning microscope equipped with a 100× oil immersion objective lens (1.40 NA), or with an A1 (Nikon) confocal laser scanning microscope equipped with a 100× oil immersion objective lens (1.49 NA) and NIS elements. The images were adjusted using Adobe Photoshop CS4 software (Adobe Systems).
shRNA-mediated RNA interference
The oligonucleotide sequences used for shRNA interference were as follows: 298–316 bp of rat Rab27a (5′-GACCTGACAAACGAGCAAA-3′), 11–29 bp of rat Noc2 (5′-CCATCTTCAGCAGTGGAAA-3′), 61-83 bp of rat Rab2a (5′-GCTTATTGCTACAGTTTAC-3′), 129–147 bp of rat ICA69 (5′-GGAAGATGAACATGTCGTT-3′), and 647–665 bp of GFP (5′-GCGATCACATGATCTACTT-3′), respectively, followed by a 9-nucleotide non-complementary spacer (TTCAAGAGA) and the reverse complement of the initial 19-nucleotide sequence. These dsDNA oligonucleotides were cloned into the pENTR/U6 vector (Invitrogen) and transferred into a pAd/PL vector (Invitrogen) by LR Clonase recombination. Sub-confluent INS1 832/13 cells in 35-mm dishes were infected with adenovirus, and were transferred to 60-mm dishes 48 h later. The cells were infected again with the virus 24 h later, and were transferred to 12-well dishes after an additional 24 h. Experiments were performed 24 h after the transfer.
siRNA-mediated RNA interference
On-Target plus SMARTpool siRNA against rat Noc2 (catalog no. 171123) and Rab2a (catalog no. 65158), as well as control On-Target plus non-targeting pool siRNA, were purchased from GE Dharmacon. INS1 832/13 cells plated at a density of 2.5×106 in a 6-well dish were grown for 24 h. Suspended cells after trypsinization were transfected twice with siRNAs using Lipofectamine RNAiMAX reagent (Invitrogen), according to the manufacturer's instructions. The second transfection was performed 48 h later, and the cells were analyzed 48 h thereafter.
Insulin and proinsulin secretion assay
INS1 832/13 cells plated on 6- or 12-well plates were cultured in the RPMI medium for 24 h. The cells were incubated in modified Krebs–Ringer bicarbonate buffer (KRB; 120 mM NaCl, 5 mM KCl, 24 mM NaHCO3, 1 mM MgCl2, 2 mM CaCl2, 15 mM HEPES pH 7.4, 0.1% BSA, 2 mM glucose) for 2 h followed by the same buffer or a buffer containing 25 mM glucose for 2 h. Insulin levels were measured with an AlphaLISA immunoassay kit (PerkinElmer), as described previously (Wang et al., 2013). Proinsulin levels were measured by a proinsulin ELISA assay kit (Shibayagi).
Statistical significance was determined using a two-tailed unpaired t-test.
We are grateful to Drs Shoji Yamaoka (Tokyo Medical and Dental University), Eiji Morita (Hirosaki University), Toshio Kitamura (The University of Tokyo) and Christopher Newgard (Duke University) for supplying pMRX-puro vector, pCAG-OSF vector, PLAT-E cells, and INS1 832/13 cells, respectively. We thank Drs Takuji Fujita and Hiroshi Gomi for generation and characterization of the β-cell lines from ashen mice. We also thank S. Shigoka for her assistance in preparing the manuscript.
K.M. designed and performed experiments and wrote the article. M.T. and T. Isobe contributed to the mass analysis. T. Izumi designed experiments and wrote the article.
This work was supported by KAKENHI grants in aid for scientific research from the Japan Society for the Promotion of Science (JSPS) (JP20113005, JP24390068 and JP16K/5211 to T. Izumi, and JP23790354 and JP25860208 to K.M.). It was also supported by an Insulin Study Award grant from Novo Nordisk (to T. Izumi), and from The Uehara Memorial Foundation, Takeda Science Foundation, the Tokyo Biochemical Research Foundation, and the NOVARTIS Foundation (Japan) for the promotion of science (to K.M.).
The authors declare no competing or financial interests.