The first C2 domain of synaptotagmin is required for exocytosis of insulin from pancreatic β-cells: action of synaptotagmin at low micromolar calcium
1997; Springer Nature; Volume: 16; Issue: 19 Linguagem: Inglês
10.1093/emboj/16.19.5837
ISSN1460-2075
AutoresJochen Lang, Mitsunori Fukuda, Hui Zhang, Katsuhiko Mikoshiba, Claes B. Wollheim,
Tópico(s)Lipid Membrane Structure and Behavior
ResumoArticle1 October 1997free access The first C2 domain of synaptotagmin is required for exocytosis of insulin from pancreatic β-cells: action of synaptotagmin at low micromolar calcium Jochen Lang Corresponding Author Jochen Lang Division de Biochimie Clinique, Departement de Médecine Interne, Centre Médical Universitaire, CH-1211 Genève 4, Switzerland Search for more papers by this author Mitsunori Fukuda Mitsunori Fukuda Molecular Neurobiology Laboratory, Tsukuba Life Science Center, Institute of Physical, University of Tokyo, Japan Chemical Research (RIKEN), 3-1-1 Koyadai, Tsukuba, Ibaraki 305 and Department of Molecular Neurobiology, Institute of Medical Science, University of Tokyo, Japan Search for more papers by this author Hui Zhang Hui Zhang Division de Biochimie Clinique, Departement de Médecine Interne, Centre Médical Universitaire, CH-1211 Genève 4, Switzerland Search for more papers by this author Katsuhiko Mikoshiba Katsuhiko Mikoshiba Molecular Neurobiology Laboratory, Tsukuba Life Science Center, Institute of Physical, University of Tokyo, Japan Chemical Research (RIKEN), 3-1-1 Koyadai, Tsukuba, Ibaraki 305 and Department of Molecular Neurobiology, Institute of Medical Science, University of Tokyo, Japan Search for more papers by this author Claes B. Wollheim Claes B. Wollheim Division de Biochimie Clinique, Departement de Médecine Interne, Centre Médical Universitaire, CH-1211 Genève 4, Switzerland Search for more papers by this author Jochen Lang Corresponding Author Jochen Lang Division de Biochimie Clinique, Departement de Médecine Interne, Centre Médical Universitaire, CH-1211 Genève 4, Switzerland Search for more papers by this author Mitsunori Fukuda Mitsunori Fukuda Molecular Neurobiology Laboratory, Tsukuba Life Science Center, Institute of Physical, University of Tokyo, Japan Chemical Research (RIKEN), 3-1-1 Koyadai, Tsukuba, Ibaraki 305 and Department of Molecular Neurobiology, Institute of Medical Science, University of Tokyo, Japan Search for more papers by this author Hui Zhang Hui Zhang Division de Biochimie Clinique, Departement de Médecine Interne, Centre Médical Universitaire, CH-1211 Genève 4, Switzerland Search for more papers by this author Katsuhiko Mikoshiba Katsuhiko Mikoshiba Molecular Neurobiology Laboratory, Tsukuba Life Science Center, Institute of Physical, University of Tokyo, Japan Chemical Research (RIKEN), 3-1-1 Koyadai, Tsukuba, Ibaraki 305 and Department of Molecular Neurobiology, Institute of Medical Science, University of Tokyo, Japan Search for more papers by this author Claes B. Wollheim Claes B. Wollheim Division de Biochimie Clinique, Departement de Médecine Interne, Centre Médical Universitaire, CH-1211 Genève 4, Switzerland Search for more papers by this author Author Information Jochen Lang 1, Mitsunori Fukuda2,3, Hui Zhang1, Katsuhiko Mikoshiba2,3 and Claes B. Wollheim1 1Division de Biochimie Clinique, Departement de Médecine Interne, Centre Médical Universitaire, CH-1211 Genève 4, Switzerland 2Molecular Neurobiology Laboratory, Tsukuba Life Science Center, Institute of Physical, University of Tokyo, Japan 3Chemical Research (RIKEN), 3-1-1 Koyadai, Tsukuba, Ibaraki 305 and Department of Molecular Neurobiology, Institute of Medical Science, University of Tokyo, Japan *Corresponding author. E-mail: [email protected] The EMBO Journal (1997)16:5837-5846https://doi.org/10.1093/emboj/16.19.5837 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions Figures & Info The Ca2+- and phospholipid-binding protein synaptotagmin is involved in neuroexocytosis. Its precise role and Ca2+-affinity in vivo are unclear. We investigated its putative function in insulin secretion which is maximally stimulated by 10 μM cytosolic free Ca2+. The well-characterized synaptotagmin isoforms I and II are present in pancreatic β-cell lines RINm5F, INS-1 and HIT-T15 as shown by Northern and Western blots. Subcellular fractionation and confocal microscopy revealed their presence mainly on insulin-containing secretory granules whereas only minor amounts were found on synaptic vesicle-like microvesicles. Antibodies or Fab-fragments directed against the Ca2+-dependent phospholipid binding site of the first C2 domain of synaptotagmin I or II inhibited Ca2+-stimulated, but not GTPγS-induced exocytosis from streptolysin-O-permeabilized INS-1 and HIT-T15 cells. Transient expression of wild-type synaptotagmin II did not alter exocytosis in HIT-T15 cells. However, mutations in the Ca2+-dependent phospholipid binding site of the first C2 domain (Δ180–183, D231S) again inhibited only Ca2+-, but not GTPγS-evoked exocytosis. In contrast, mutations in the IP4-binding sites of the second C2 domain (Δ325–341; K327,328,332Q) did not alter exocytosis. Synaptotagmin II mutated in both C2 domains (Δ180–183/K327,328,332Q) induced greater inhibition than mutant Δ180–183, suggesting a discrete requirement for the second C2 domain. Thus, synaptotagmin isoforms regulate exocytotic events occurring at low micromolar Ca2+. Introduction The secretion of insulin proceeds by exocytosis, i.e. fusion between secretory granules and plasma membranes with subsequent release of the granule content into the extracellular space (Wollheim et al., 1996). Exposure of β-cells to secretagogues increases the cytosolic Ca2+ to micromolar levels, which in turn triggers insulin secretion (Theler et al., 1992; Ammala et al., 1993). How Ca2+ triggers exocytosis in pancreatic β-cells is still not identified. In neuroexocytosis, a pivotal role for the Ca2+- and phospholipid-binding protein synaptotagmin has been demonstrated (Sudhof, 1995). Currently, 11 isoforms of synaptotagmin are known in rodents (Geppert et al., 1991; Hilbush and Morgan, 1994; Mizuta et al., 1994; Craxton and Goedert, 1995; Li et al., 1995; Kwon et al., 1996; Babity et al., 1997). Antibody and peptide injections in PC12 cells provided early evidence for a role of synaptotagmin in exocytosis (Elferink et al., 1993). The deletion mutation of one of these isoforms, i.e. synaptotagmin I, in transgenic mice or of Drosophila synaptotagmin, led to a major impairment of fast Ca2+-induced neuroexocytosis. Nonetheless, the precise role and Ca2+affinity in vivo of synaptotagmin I are unknown (DiAntonio et al., 1993; Broadie et al., 1994; DiAntonio and Schwarz, 1994; Geppert et al., 1994). Synaptotagmin is an integral membrane protein of synaptic vesicles (Brose et al., 1992) and of neurohormone-containing secretory vesicles (Walch Solimena et al., 1993). It is composed of a short intravesicular N-terminal sequence, a single transmembrane region and a large cytosolic portion (Sudhof, 1995). The main features of the cytosolic part are two Ca2+-binding repeats, the C2 domains, whose basic motif was first identified in protein kinase C and has subsequently been found in several other proteins. They mediate Ca2+-induced attachment to membrane lipids (Brose et al., 1992). Several functions have been assigned to the C2 domains in synaptotagmin. The first C2 domain (C2A) binds to phospholipids at ∼5 μM free Ca2+ (Brose et al., 1992; Li et al., 1995; Fukuda et al., 1996) and to the membrane protein syntaxin 1 in the presence of several hundred micromolar Ca2+ (Li et al., 1995). The latter Ca2+ requirement coincides with that for exocytosis in bipolar neurones (Heidelberger et al., 1994). In addition to Ca2+-sensitive properties, the second C2 domain (C2B) binds inositol-1,3,4,5-tetrakisphosphate (IP4) independent of Ca2+ (Fukuda et al., 1994). The injection of IP4 into a preterminal of the squid giant synapse inhibited neuroexocytosis. This effect can be prevented by concomitant injection of an antibody directed against the C2B domain (Llinas et al., 1994; Fukuda et al., 1995b). Thus, an IP4-binding site in the C2B domain is seemingly required in neuroexocytosis and may mediate attachment to membrane lipids (Fukuda et al., 1995b). Endocrine exocytosis, such as that of insulin in the pancreatic β-cells, is maximally stimulated by 7–10 μM free Ca2+ with an EC50 of 1.6 μM (Vallar et al., 1987, Ullrich et al., 1990, Bokvist et al., 1995; Proks et al., 1996). β-Cells therefore provide a good model to test the action of synaptotagmin domains in situ at low micromolar Ca2+. Although the occurrence of several synaptotagmin isoforms has been described in pancreatic β-cells at the RNA level, their subcellular localization is unresolved (Mizuta et al., 1994; Wheeler et al., 1996). This is of crucial importance as insulin-secreting cells contain not only insulin-containing secretory granules, but also small synaptic vesicle-like microvesicles (SLMVs) containing GABA (Reetz et al., 1991). We have therefore investigated the subcellular distribution of two well-characterized isoforms of synaptotagmin, i.e. synaptotagmin I and II, and characterized their putative function. Our findings demonstrate that synaptotagmin I and II are mainly localized on secretory granules. Domains essential for Ca2+-mediated phospholipid binding in the C2A domain of synaptotagmin I and II are required for endocrine exocytosis of insulins operating at low micromolar Ca2+. Results We first investigated the expression and subcellular distribution of synaptotagmin I and II. RNA blot analysis for synaptotagmin I and II revealed the presence of a 4.8 kb transcript for the synaptotagmin I in rat brain RNA and RNA from the insulin-secreting cells RINm5F, HIT-T15 and INS-1 as well as primary islet cells (Figure 1, upper panel). Similarly, a 7 kb band was detected in blots probed for synaptotagmin II (Figure 1, lower panel) and, in addition, a minor transcript migrating at ∼9 kb was apparent. The abundance of mRNAs for synaptotagmin I and II was less in insulin-secreting cells as compared with rat. Figure 1.RNA blot analysis of synaptotagmin I and II mRNAs in insulin-secreting β-cell lines. 30 μg of total RNA from whole rat brain or indicated cell lines and primary islet cells were denatured and electrophoresed as described in the text. Upper panel, expression of synaptotagmin I RNA; Lower panel, expression of synaptotagmin II RNA. For autoradiography, the nylon membrane was exposed to X-ray film with an intensifying screen at −80°C for 3 days. Arrowheads indicate the position of 18S and 28S RNA. Download figure Download PowerPoint Western blot demonstrated the presence of the corresponding proteins in crude membrane preparations (Figure 2). Monoclonal antibodies directed against the C2A domain (mab 41.1) or against the intravesicular N-terminus of synaptotagmin I (mab 604) both detected a band of ∼60 kDa in the three cell lines. Furthermore, two other antibodies recognizing synaptotagmin I and II (mab 48 and anti-sytI/II) gave positive signals. In contrast, a monoclonal antibody directed against the N-terminus of synaptotagmin II (mab 8G2b) recognized a 60 kDa protein only in RINm5F and INS-1 cells, derived from a rat insulinoma, but not in the hamster HIT-T15 cells. Interspecies variations in the synaptotagmin II sequence can be excluded as mab 8G2b strongly reacted with a 60 kDa band in immunoblots from hamster brain (data not shown). We therefore prepared secretory granule membranes and detergent extracts of crude membranes from HIT-T15 cells. In those preparations synaptotagmin II is clearly detectable. This isoform is therefore also expressed in this cell line, albeit at lower levels than in the other insulin-secreting cells. Figure 2.Immunoblot analysis of synaptotagmin I and II in insulin-secreting β-cell lines. (A) Crude membranes (30 μg) from rat brain (1) or the indicated insulin-secreting cell lines RINm5F (2), HIT-T15 (3) or INS-1 (4) were separated on 12% SDS–PAGE, electrotransferred to PVDF and probed with antibodies against the first C2-domain (C2A) of synaptotagmin I (mab 41.1) or of synaptotagmin I and II (anti-sytI/II), against synaptotagmin I and II (mab 48) or against the intravesicular N-terminus of synaptotagmin I (mab 604) or synaptotagmin II (mab 8G2b). (B) Crude membranes (5 μg) from rat brain (5), 200 μg of Triton X-100 extract from crude HIT-T15 membranes (6) and 20 or 50 μg of enriched HIT-T15 insulin granules (7 and 8) were separated by 12% SDS–PAGE, electrotransferred to PVDF and probed with antibody against synaptotagmin II (mab 8G2b). Download figure Download PowerPoint Insulin-secreting cells contain two types of exocytotic vesicles: small synaptic-like mirovesicles, SLMVs (Reetz et al., 1991) and insulin-containing granules. Exocytotic vesicle proteins may be localized on both of them (Regazzi et al., 1995). Therefore, subcellular fractionation was performed on INS-1 cells using a continuous sucrose gradient to distinguish between the two vesicle types. Using this method, insulin-containing granules migrated at ∼1.4 M sucrose (Figure 3, lower panel) as indicated by the distribution of immunoreactive insulin in fractions 12–15. The distribution of the SLMVs is indicated by the vesicle protein synaptophysin, which is recovered at ∼0.8 M sucrose in fractions 5 and 6 (Figure 3, middle panel). The distribution of synaptotagmin I and II was tested with mab 41.1, directed against synaptotagmin I, and with anti-sytI/II, which recognizes synaptotagmin I and II. As indicated in Figure 3 (upper panel), immunoreactivity for synaptotagmin I and II was mainly concentrated in the fractions 12–15 containing secretory granules. In contrast, only a small amount of immunoreactivity was found in the synaptophysin-reactive fractions 5 and 6 containing SLMVs. In addition, some synaptotagmin immunoreactivity sedimented in a region previously identified as containing plasma (Lang et al., 1995). Figure 3.Subcellular distribution of synaptotagmin I and II in INS-1 cells. INS-1 cells were homogenized by nitrogen cavitation and centrifuged through a continuous sucrose gradient. Lower panel: distribution of protein (▪) and sucrose density (dashed line). Middle panel: distribution of synaptophysin (SVP38, •) and insulin (○) as measured by immunoblots with subsequent densitometry or by RIA. Upper panel: distribution of synaptotagmin I (SYT I; mab 41.1) and synaptotagmin I and II (anti-sytI/II). Download figure Download PowerPoint The subcellular distribution of synaptotagmin in INS-1 cells was further examined by histocytochemistry and confocal microscopy (Figure 4). The upper panel demonstrates that insulin and synaptophysin—markers for secretory granules and synaptic-like microvesicles, respectively—do not co-localize. As already observed in subcellular fractionation, synaptotagmin does stain synaptophysin-positive structures (Figure 4). A different situation was encountered in the co-staining for insulin (green) and synaptotagmin (red) (Figure 4B). Synaptotagmin was found in a granular pattern and outlining the cell surface. Most of the immunoreactivity for insulin co-localizes with synaptotagmin, as indicated by the yellow stain. A similar co-localization between synaptotagmin and insulin, but not with synaptophysin, was observed in HIT-T15 cells (data not shown). Immunocytochemistry on pancreatic islets extended previous observations for synaptotagmin I (Jacobsson et al., 1994) to synaptotagmin II. Immunoreactivities for both isoforms were absent from primary β-cells but were found on non-insulin endocrine islet cells (data not shown). Figure 4.Immunofluorescence of synaptotagmin in insulin-secreting INS-1 cells. INS-1 cells were fixed, permeabilized and subsequently incubated with anti-synaptophysin (1:400), anti-sytI/II (1:10 000) or anti-insulin (1:400). After processing with fluorescent second antibodies, pictures were taken by confocal microscopy at the mid-cellular level. (A) Anti-synaptophysin (green) and anti-insulin (red); (B) anti-insulin (green) and anti-synaptotagmin I/II (red); (C) anti-synaptophysin (green) and anti-synaptotagmin I/II (red). Left panels: immunofluorescence; right panels: phase contrast. Bars equal 5 μm. Download figure Download PowerPoint The above results suggested that synaptotagmin I and II are concentrated on insulin-containing secretory granules. We have therefore studied whether they could play a functional role in exocytosis, the final step of insulin secretion. To this end we initially used two antibodies which bind to defined regions of synaptotagmin. Affinity-purified anti-sytI/II binds to the first C2 domain (C2A) of synaptotagmin and inhibits Ca2+-dependent interaction of synaptotagmin II with phospholipid vesicles (Fukuda et al., 1995a; Mikoshiba et al., 1995). The monoclonal antibody mab 41.1 has been shown to require a short stretch of nine amino acids in the C2A domain of synaptotagmin I for binding (Chapman and Jahn, 1994) and these amino acids are required for Ca2+-dependent interaction with phospholipid vesicles (Chapman and Jahn, 1994; Fukuda et al., 1996). In the upper panel of Figure 5, the isoform specificity of these two antibodies is illustrated from experiments with recombinant fusion proteins. Both antibodies interacted with the first C2 domain (sytI–C2A), but not with the second C2 domain of synaptotagmin I (sytI–C2B). Anti-sytI/II reacted equally well with synaptotagmin II, but none of the two antibodies recognized synaptotagmin III, synaptotagmin IV or GST alone, the tag of the fusion protein. To test the functional activity of the antibodies, INS-1 or HIT-T15 cells were permeabilized with streptolysin-O (SL-O) and subsequently preincubated at 0.1 μM Ca2+ with the indicated concentration of affinity-purified IgG (anti-sytI/II) or Fab-fragments of mab 41.1 (Fab 41.1) as shown in Figure 5 (lower panel). After aspiration of antibody solutions, cells were exposed to basal levels of free Ca2+ (0.1 μM), or maximally stimulatory levels of Ca2+ (10 μM) as previously demonstrated for β-cells and derived cell lines (Vallar et al., 1987; Ullrich et al., 1990; Ammala et al., 1993; Bokvist et al., 1995). Alternatively, cells were stimulated with GTPγS. This stable GTP analogue induces Ca2+-independent exocytosis (Vallar et al., 1987; Jonas et al., 1994; Kiraly-Borri et al., 1996). The difference between insulin release at basal Ca2+ conditions and at stimulatory conditions (10 μM Ca2+, 100 μM GTPγS) in the absence of antibodies was normalized to 100%. In both cell lines, HIT-T15 and INS-1, the Fab-fragment and the affinity-purified IgG inhibited Ca2+-evoked insulin release by ∼60%, whereas the GTPγS-induced release remained unchanged (Figure 5, lower panel). In contrast to the inhibitory action of native Fab 41.1 or anti-sytI/II IgG, preheated preparations were inactive (Table I). Furthermore, control IgG or Fab did not alter insulin release, and a monoclonal antibody against the intravesicular N-terminus, which is located inside the secretory granule, did not affect hormone levels. A possible effect of Fab 41.1 or anti-sytI/II IgG on the cell surface could be excluded as Fab fragments or IgG were inactive in cells permeabilized by Staphylococcus aureus α-toxin (Table I). This toxin produces only small pores in the plasma membrane which do not allow the entry of molecules larger than 0.5–2 kDa. The inefficacy of the specific antibodies in α-toxin-permeabilized cells also excludes any contribution by antibody buffers to the inhibitory effect observed in SL-O-permeabilized cells. Figure 5.Effect of functional anti-synaptotagmin antibodies on insulin release from permeabilized INS-1 and HIT-T15 cells. Upper panel: reactivity of antibodies against recombinant proteins. Recombinant GST (GST), first (sytI-C2A) or second C2 domain (sytI-C2B) or both C2 domains of synaptotagmin II, III and IV (100 ng/lane) were separated by SDS-PAGE, immunoblotted and incubated with mab 41.1 (left part) or anti-sytI/II (right part). Lower panel: insulin-secreting cells (INS-1, ▪, □; HIT-T15, •, ○) were permeabilized with streptolysin-O, preincubated at 0.1 μM Ca2+ with the indicated concentrations of Fab-fragments of mab 41.1 (Fab 41.1) or affinity-purified polyclonal IgG against the C2A domain of synaptotagmin II (anti-sytI/II). Subsequently, supernatants were aspirated and replaced by antibody-free solution containing 0.1 μM, 10 μM Ca2+ or 100 μM GTPγS. After 7 min, supernatants were sampled and prepared for the determination of released insulin by radioimmunoassay. The Ca2+- or GTPγS-induced stimulation of insulin release in the absence of IgG or Fab was normalized to 100%. For fold-increase upon stimulation see Table I. n = 6–21 from at least three separate experiments for each point; *, 2P <0.05. Download figure Download PowerPoint Table 1. Effect of IgG or Fab-fragments on exocytosis in SL-O- or α-toxin-permeabilized cells Preincubation SL-O permeabilization α-toxin permeabilization 0.1 μM Ca2+ GTPγS 10 μM Ca2+ 0.1 μM Ca2+ GTPγS 10 μM Ca2+ Control 100 ± 23 256 ± 45 375 ± 28 100 ± 5 375 ± 12 621 ± 43 Anti-sytI/II 112 ± 7 248 ± 41 231 ± 17* 98 ± 12 405 ± 56 604 ± 56 Anti-sytI/II heated 93 ± 30 n.d. 398 ± 21 – – – IgG 105 ± 12 289 ± 58 345 ± 9 114 ± 17 371 ± 4 658 ± 13 Fab 41.1 97 ± 28 234 ± 66 217 ± 28* 103 ± 11 352 ± 31 599 ± 25 Fab 41.1 heated 114 ± 6 n.d. 405 ± 13 – – – Fab 89 ± 17 265 ± 39 402 ± 5 105 ± 9 368 ± 8 683 ± 38 Mab604.4 132 ± 31 276 ± 7 398 ± 31 121 ± 21 389 ± 24 595 ± 37 2+ Cells were permeabilized with streptolysin-O or with α-toxin and pretreated with IgG (anti-sytIIC2A, mab 604.4, non-specific IgG, 40 μg/ml) or Fab-fragments (Fab 41.1, non-specific Fab, 40 μg/ml) as in Figure 6. Heat inactivation of antibodies was performed for 20 min at 60°C. Cells were subsequently exposed to 0.1 μM Ca, 0.1 μM Ca with 100 μM GTPγS (GTPγS) or 10 μM Ca for 7 min and insulin release determined in the supernatants. Values are expressed as percent of controls (0.1 μM Ca in the absence of IgG or Fab) and means ± SEM are given for n = 6–9 from three separate experiments * , 2P <0.05; n.d., not determined. These results suggest a specific role for the C2A domain of synaptotagmin I in Ca2+-stimulated exocytosis from insulin-secreting cells. As the functional antibody anti-sytI/II reacted with both isoforms—that is, synaptotagmin I and II—we used a separate approach to evaluate further the role of synaptotagmin II in the secretion of insulin. In HIT-T15 cells we transiently expressed wild-type synaptotagmin II or synaptotagmin II mutated in the C2 domains. To determine the subcellular localization of transiently expressed synaptotagmin II, we analysed its distribution by confocal microscopy. Thus, we took advantage of the finding that only low endogenous levels of synaptotagmin II are present in HIT-T15 cells (see also Figure 2). Therefore, a monoclonal antibody could be used that is directed against the N-terminus (Nishiki et al., 1996) which is highly isoform-specific (Li et al., 1995). At the dilution used, the antibody did not react with endogenous synaptotagmin II (data not shown). Transiently expressed synaptotagmin II again co-localized largely with secretory granules as indicated by the co-distribution of insulin (Figure 6). In addition, synaptotagmin immunoreactivity was observed on the cell border (Figure 6) similar to endogenous synaptotagmin in INS-1 cells (see Figure 4). To determine whether this localization corresponds indeed to the plasma membrane, we used double staining with anti-synaptotagmin II and an antibody against syntaxin, a known plasma membrane protein. As syntaxin and synaptotagmin are differentially distributed, the rim-like stain observed with anti-synaptotagmin II corresponds probably not to localization at the plasma membrane, but to a subplasmalemmal region (Figure 6). All mutants used in this study exhibited a similar subcellular distribution (data not shown). To determine the levels of overexpressed protein, cells were co-transfected with a plasmid encoding the green fluorescent protein (GFP). This allows subsequent purification of the small population of the transiently transfected cells (∼8–13% of all cells) by fluorescence-activated cell sorting. As shown in the upper panel of Figure 7, again using a monoclonal antibody specific for synaptotagmin II (Nishiki et al., 1996), wild-type protein and its different mutants were expressed at comparable levels. Figure 6.Localization of transiently expressed synaptotagmin II in HIT-T15 cells. Upper panel: immunostaining of transiently transfected HIT-T15 cells for synaptotagmin II (red) and insulin (green). Lower panel: immunostaining for transiently expressed synaptotagmin II (red) and for syntaxin 1 (green). The monoclonal antibody against synaptotagmin II was used at a dilution (1:800) which did not stain endogenous synaptotagmin II. A similar distribution was found for the different synaptotagmin mutants used in this study (data not shown). Bars equal 5 μm. Download figure Download PowerPoint Figure 7.Effect of transiently expressed mutants of synaptotagmin II on insulin release from intact insulin-secreting HIT-T15 cells. Upper panel: HIT-T15 cells transiently expressing synaptotagmin II wild-type or mutants were purified by fluorescence-activated cell sorting using co-transfected green fluorescent protein. Control cells and fluorescent cells were solubilized in SDS-sample buffer and subjected to Western blot analysis (6×104 cells/lane) using an antibody specific for synaptotagmin II (mab 8G2b). Lower panel: HIT-T15 cells were co-transfected with plasmid encoding for human preproinsulin and control plasmid (pcDNA3, CON) or pCDNA3 containing the inserts for synaptotagmin II wild-type (WT), mutants in the first C2 domain (Δ180–183; D231S), in the second C2 domain (Δ325–341; KQ, where K327,328 and 332 were exchanged for Q) or in both domains (double mutant DM carrying the mutations Δ180–183 and the exchange of K327,328,332 to Q). Secretion experiments were performed 48–72 h later using Krebs–Ringer buffer with 3.4 mM KCl (BASAL) or with 50 mM KCl (50 mM KCl). Subsequently, the amount of human insulin C-peptide in the supernatants was measured reflecting the secretion from co-transfected cells only. n = 12–24 from at least three separate experiments for each point; ☆, 2P <0.05. Download figure Download PowerPoint To study the role of synaptotagmin II in secretion and exocytosis, the GFP-encoding plasmid was replaced by a construct expressing human preproinsulin, and human insulin C-peptide release was measured from hamster HIT-T15 cells (Figure 7, lower panel). This approach reliably reflects release from co-transfected cells as published previously (Lang et al., 1995). Wild-type synaptotagmin II (wt) did not alter basal insulin release from intact cells or hormone release induced by stimulation due to depolarization of membrane potential by KCl leading to Ca2+ entry through voltage-dependent Ca2+ channels (Wollheim and Sharp, 1981; Wollheim et al., 1996). Deletion mutation Δ180–183 and the point mutation D231S in the C2A domain of synaptotagmin II have been shown to inhibit Ca2+-dependent phospholipid binding (Fukuda et al., 1996). Here they significantly reduced stimulated insulin secretion by some 50% and 35%, respectively (Figure 7, lower panel). Two mutations in the C2B domain, i.e. Δ325–341 or the exchange of K326,327,332, to Q (KQ), abolishes the binding of IP4 and other inositol high-polyphosphates to synaptotagmin II (Fukuda et al., 1995a), a process which is thought to mimic binding of the C2B domain of synaptotagmin to membranes (Llinas et al., 1994; Mochida et al., 1997). In our hands these mutants did not alter the secretion of insulin. However, a double mutant (DM) containing Δ180–183 and KQ mutations of the C2A and C2B domains inhibited the release to a much larger extent than any mutation in the C2A domain alone (Figure 7). To study whether the observed effects of synaptotagmin II mutants on secretion occur at the level of exocytosis, or occur indirectly, we also measured their effect in SL-O-permeabilized cells. As shown in Figure 8, under this condition the effects of all tested constructs were comparable with those observed in intact cells. Similar to the effect of the functional antibodies, again only the Ca2+-, but not the GTPγS-stimulated exocytosis, was affected. Figure 8.Effect of transiently expressed mutants of synaptotagmin II on exocytosis of insulin from permeabilized HIT-T15 cells. Cells were transfected with the same constructs as in Figure 6. Cells were permeabilized 48–72 h later with recombinant streptolysin-O. Cells were subsequently exposed for 7 min to 0.1 μM Ca2+, 100 μM GTPγS or 10 μM Ca2+ and human insulin C-peptide measured from the supernatants. n = 10–27 from at least three separate experiments for each point; *, 2P <0.05. Download figure Download PowerPoint Discussion Exocytosis from pancreatic β-cells shares a number of features with exocytosis in neurones and neuroendocrine cells. We and others have previously shown that the release of insulin involves essential components of the general docking and fusion machinery. The process is ATP-dependent (Vallar et al., 1987; Lang et al., 1995) and requires the SNARE proteins synaptobrevin/VAMP (Regazzi et al., 1995, 1996), SNAP-25 (Sadoul et al., 1995), syntaxin (Martin et al., 1995, 1996) and α-SNAP (Kiraly-Borri et al., 1996). Insulin release is maximal at low micromolar free Ca2+ (Vallar et al., 1987; Ullrich et al., 1990; Bokvist et al., 1995; Proks et al., 1996). Therefore, insulin-secreting cells provide a suitable system to test the Ca2+-dependency of putative Ca2+-sensing proteins in vivo. Here, we demonstrate that the neuronal Ca2
Referência(s)