The Activation of Exocytotic Sites by the Formation of Phosphatidylinositol 4,5-Bisphosphate Microdomains at Syntaxin Clusters*

Phosphatidylinositol 4,5-bisphosphate (PI(4,5)P2) is a minor component of the lipid bilayer but plays an important role in various cellular functions, including exocytosis and endocytosis. Recently, PI(4,5)P2 was shown to form microdomains in the plasma membrane. In this study, we investigated the relationship between the spatial organization of PI(4,5)P2 microdomains and exocytotic machineries in clonal rat pheochromocytoma PC12 cells. Both PI(4,5)P2 and syntaxin, a soluble N-ethylmaleimide-sensitive factor attachment protein receptor protein essential for exocytosis, exhibited punctate clusters in isolated plasma membranes. The number of PI(4,5)P2 microdomains colocalizing with syntaxin clusters and large dense core vesicles (LDCVs) was decreased after catecholamine release. Alternatively, the expression of type I phosphatidylinositol-4-phosphate 5-kinase (PIP5KI) increased the number of PI(4,5)P2 microdomains at syntaxin clusters with docked LDCVs and enhanced exocytotic activity, possibly by increasing the number of release sites. About half of the PI(4,5)P2 microdomains were not colocalized with Thy-1, a specific marker of lipid rafts, and the colocalization of transfected PIP5KI with syntaxin clusters was observed. These results suggest that the formation of PI(4,5)P2 microdomains at syntaxin clusters with docked LDCVs is essential for Ca2+-dependent exocytosis.

Ca 2ϩ -dependent exocytosis is a cellular mechanism for the secretion of various bioactive substances such as neurotransmitters and hormones. In response to intracellular Ca 2ϩ elevation, fusion of the secretory vesicle membrane with the plasma membrane causes a release of its contents into the extracellular space (1). Recent studies showed that the complex formation of soluble N-ethylmaleimide-sensitive factor attachment protein receptor proteins, syntaxin and 25-kDa synaptosome-associated protein, in the plasma membrane and vesicle-associated membrane protein-2 in the secretory vesicle membrane, are essential for exocytosis (2,3).
Electrophysiological and biochemical studies showed that the Ca 2ϩ -dependent exocytotic process could be dissected into multiple steps, and different sets of proteins are likely to be involved in each step (4 -6). ATP-dependent priming was recognized as an essential step prior to Ca 2ϩ -dependent membrane fusion to acquire Ca 2ϩ sensitivity of the fusion step (7,8). In membrane-permeabilized chromaffin cells, ATP removal produced the same effects as phosphatidylinositol-4,5-bisphosphate (PI(4,5)P 2 ) 1 -degrading or -masking agents (9), and two enzymes involved in phosphoinositide metabolism, phosphatidylinositol (PI) transfer protein and type I phosphatidylinositol-4-phosphate 5-kinase (PIP5KI), were isolated as the cytosolic factors required for the ATP-dependent priming step (10,11). Overexpression of the pleckstrin homology (PH) domain of phospholipase C␦, which binds to PI(4,5)P 2 with a high affinity (12,13), markedly inhibited the exocytosis of large dense core vesicles (LDCVs) in adrenal chromaffin cells, possibly by reducing PI(4,5)P 2 availability in the plasma membrane (14). Alternatively, the overexpression of constitutive active Arf6, an activator of PIP5KI, induced the accumulation of PIP5KI in the endosomal membrane and in turn the depletion of PI(4,5)P 2 in the plasma membrane, resulting in the inhibition of exocytosis (15). All of these results suggest that the generation of PI(4,5)P 2 in the plasma membrane is crucial for the step preceding Ca 2ϩ -dependent membrane fusion.
It has become generally accepted that lipids in the plasma membrane are not homogeneously distributed and form various microdomains. The most well characterized lipid microdomain is a detergent-insoluble, cholesterol-dependent microdomain, termed a lipid raft. The lipid raft might play important roles in a variety of biological functions by accumulating various sets of proteins to facilitate the efficacy of signal transduction, focal regulation of cytoskeleton, and membrane traffic (16 -18). Recently, Laux et al. (19) showed that PI(4,5)P 2 also formed microdomains in the plasma membrane, and at least part of these microdomains was colocalized with the myristoyl-ated alanine-rich C kinase substrate, a protein enriched in the lipid raft, and involved in the regulation of the actin cytoskeleton. Biochemical studies also revealed that about half of cellular PI(4,5)P 2 was recovered in lipid raft-enriched fractions in discontinuous sucrose density gradient fractionation, suggesting that part of PI(4,5)P 2 was concentrated in lipid raft microdomains (20,21). In addition to exocytosis, PI(4,5)P 2 is also essential for diverse cellular functions, including the production of second messengers, endocytosis, and the regulation of actin cytoskeleton. Since these phenomena are likely to occur in a distinct region in the plasma membrane, PI(4,5)P 2 might be localized in various distinct microdomains to participate in different cellular functions (22)(23)(24)(25).
In this study, we investigated the spatial correlation between PI(4,5)P 2 microdomains and exocytotic machineries in the isolated and intact plasma membrane of PC12 cells. We found that PI(4,5)P 2 was accumulated in syntaxin clusters with docked LDCVs, and exocytotic activity was well correlated with the accumulation of PI(4,5)P 2 in the exocytotic machinery.

MATERIALS AND METHODS
Plasmids-A DNA fragment corresponding to the open reading frame of mouse PIP5KI␤ was amplified by PCR from the mouse brain cDNA library using the oligonucleotides 5Ј-ATAGGATCCATGGCGTC-CGCCTCCTCAGGG-3Ј and 5Ј-ATACTCGAGGTGGGTGAACTCTGAC-TCTGC-3Ј. The fragment was digested with BamHI and XhoI and subcloned into a pBluescript vector (Stratagene, La Jolla, CA), named pBS-PIP5KI␤. The single amino acid substitution mutant K178A was constructed using a QuikChange site-directed mutagenesis kit (Stratagene) (pBS-K178A) according to the instruction manual using the sense primer, 5Ј-CTTCAAGATAATGGCCTACAGCCTGTTGATG-3Ј (mutation site underlined). pBS-PIP5KI␤ and pBS-K178A were subcloned into pcDNA3 (Invitrogen) to construct the mammalian expression vectors pcDNA3-PIP5K and pcDNA3-K178A, respectively. To generate the construct of PH-EGFP, the PH domain of phospholipase C␦ was isolated from the PC12 cDNA library by PCR with specific primers (5Ј-CCGATTCGGGCATGGACTCGGGTAGGGAC-3Ј and 5Ј-CCGGA-TCCCTTCAGGAAGTCCTTCAGCTCC-3Ј) and subcloned into pEGFP (Promega, Madison, WI). A DNA fragment corresponding to the human growth hormone (hGH) gene was isolated from the pXGH5 vector and subcloned into pcDNA3.1/myc-His (Invitrogen), known as hGH-myc. A DNA fragment corresponding to murine Arf6, generously provided by J. Rettig (Universitä t des Saarlandes, Hombrug/Saar, Germany), was subcloned into pBluescript, and a point mutation (Q67L) was introduced to generate the constitutive active Arf6 mutant (Arf6(QL)) (26) by the GeneEditor method (Promega) using the sense primer, 5Ј-GATGT-GGGCGGCCTGGACAAGATCCGG-3Ј (mutation site underlined). It was subcloned into pcDNA4-HisMax (Invitrogen) and used as the mammalian expression vector. All constructs were sequenced to confirm the identity with a designed sequence.
Transfection-PC12 cells were plated on polyethyleneimine (PEI)coated 35-mm culture dishes at a density of 2.08 ϫ 10 5 cells/cm 2 or on PEI-coated coverslips at a density of 1.05 ϫ 10 5 cells/cm 2 . After 18 -24 h, cells were transfected with various vectors using Lipofectamine TM 2000 (Invitrogen) in the presence of serum according to the instruction manual.
Imaging of Membrane Sheets-For the preparation of membrane sheets, PC12 cells were plated on PEI-coated coverslips at a density of 1.05 ϫ 10 5 cells/cm 2 and disrupted as previously reported (29) using a sonicator in ice-cold sonication buffer (120 mM potassium glutamate, 20 mM potassium acetate, 10 mM EGTA, 0.5 mM dithiothreitol, 10 mM NaF, 20 mM Na 4 P 2 O 7 , 1 M pepstatin A, 1 g/ml antipain, 1 g/ml leupeptin, 20 mM Hepes, pH 7.2). Membrane sheets were fixed for 30 min at room temperature in 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer (pH 7.2) and washed three times in Ca 2ϩ -, Mg 2ϩ -free Dulbecco's phosphate-buffered saline (CMF-PBS) for 5 min each. They were then blocked with blocking buffer (5% goat serum in CMF-PBS) for 10 min at room temperature, followed by overnight incubation with the primary antibodies diluted in the blocking buffer at 4°C. After washing three times with CMF-PBS for 5 min each, they were incubated with secondary antibodies diluted in the blocking buffer at room temperature for 1.5 h, washed three times with CMF-PBS, and then mounted with Prolong Antifade reagent (Molecular Probes). For neomycin treatment, blocking buffer including 3 mM neomycin (Sigma) was used for blocking and dilution of first antibodies. Membrane sheets were observed using confocal laser microscopy, the Zeiss LSM510 with a ϫ100 1.4 numerical aperture plan achromate objective (Carl Zeiss, Jena GmbH, Germany). Images were captured in multitrack mode. To image living PC12 cells transfected with PH-EGFP, we employed total internal reflection fluorescence microscopy (or evanescent field microscopy) as described previously (30).
Quantitative Image Analysis of Colocalization-To determine the colocalization of signals corresponding to LDCVs with syntaxin or PI(4,5)P 2 , we used a procedure previously reported (31). Circles were centered around the individual spots in the channel of chromogranin B or hGH-myc and then transferred to the other channel(s). If the fluorescence intensity maximum of a corresponding spot in the other channel was within 300 nm of the spot in the channel of chromogranin B or hGH-myc, it was rated as colocalized. To correct for accidental colocalization, the images of syntaxin and PI(4,5)P 2 were superimposed with a mirror image of chromogranin B or hGH-myc, and colocalization was determined according to the methods of Lang et al. (31). Background correction was performed according to the following formula: corrected colocalization ϭ (measured colocalization Ϫ mirror image colocalization)/(1 Ϫ mirror image colocalization/100). All images were analyzed with the program written by us in Visual Cϩϩ (Microsoft, Redmond, WA).
Release Assay-Dopamine release from PC12 cells was determined as described (32). hGH release from PC12 cells was determined essentially as described previously (32,33). Briefly, cells were washed five times with a low K ϩ solution (140 mM NaCl, 4.7 mM KCl, 1.2 mM KH 2 PO 4 , 2.5 mM CaCl 2 , 1.2 mM MgSO 4 , 11 mM glucose, and 15 mM Hepes-NaOH, pH 7.4). After washing, cells were incubated with low K ϩ solution for 2 min, and then the buffer was changed every 2 min with ionomycin-containing low K ϩ solution to stimulate Ca 2ϩ influx or high K ϩ solution, which raised the KCl concentration as mentioned in the figure legends to induce depolarization and decreased NaCl concentration to maintain the osmolarity. At the end of the experiment, cells were sonicated on ice in 1 ml of chilled 1% Nonidet P-40. Secreted hGH and the content of hGH were measured using an hGH enzyme-linked immunosorbent assay kit (Roche Applied Science).
Floatation Analysis-A detergent-insoluble, cholesterol-dependent raft fraction was isolated following a previous report (21). PC12 cells were plated on PEI-coated 100-mm dishes at a density of 1.59 ϫ 10 6 cells/cm 2 . The cells were washed twice with ice-cold CMF-PBS and once with MBS (150 mM NaCl, 2 mM EDTA, 25 mM MES-NaOH, pH 6.5), and four dishes were scraped into 2 ml of chilled lysis buffer (1% Triton X-100, 150 mM NaCl, 2 mM EDTA, MES-NaOH, pH 6.5). Lysate was passed through a 23-gauge needle 10 times and then mixed with an equal volume of 80% sucrose in MBS. Two ml of the mixture were transferred into two centrifuge tubes and overlaid with 6 ml of 30% sucrose in MBS and 4 ml of 5% sucrose in MBS. The gradient was centrifuged for 18 h at 200,000 ϫ g at 4°C and fractionated into 1-ml fractions. The pellet was resuspended in 1 ml of MBS. The amount of cholesterol in each fraction was measured using a Cholesterol E-test kit (WAKO). The fractions were mixed with one-third volume of 3ϫ sample buffer (6% SDS, 30% glycerol, 0.006% bromophenol blue, 375 mM Tris-HCl, pH 7.4) supplemented with 20 mM dithiothreitol and used for immunoblotting.
Immunoblotting-SDS-PAGE was performed on 7.5 or 12.5% polyacrylamide gel equipped with 5% stacking gel. After separation by SDS-PAGE, the proteins were transferred to polyvinylidene difluoride membranes following standard procedures with a semidry transblot-ting apparatus. The membranes were blocked in 5% (w/v) nonfat milk in TBST (150 mM NaCl, 0.05% (w/v) Tween 20, 25 mM Tris-HCl, pH 7.5) for 30 min at room temperature followed by overnight incubation with primary antibodies diluted in 5% nonfat milk in TBST at 4°C. After washing in TBST, membranes were incubated for 1.5 h at room temperature with peroxidase-labeled anti-rabbit IgG or anti-mouse IgG ϩ A ϩ M antibody (Zymed Laboratories Inc.) in TBST containing 5% nonfat milk. After washing, the immunoreactive bands were visualized by SuperSignal TM (Pierce) and a luminescence image analyzer with an electronically cooled charge-coupled device camera system (LAS-1000; Fuji Photo Film Co.).

PI(4,5)P 2 Microdomains in the Plasma Membrane of PC12
Cells-To study the distribution of PI(4,5)P 2 in the plasma membrane of PC12 cells, membrane-sheet preparation was used as the inside-out layer of the plasma membrane with exocytotic activity (29). PC12 cells were grown on coverslips and subjected to brief sonication to disrupt the upper part of cells, resulting in the generation of membrane sheets on the coverslips. We confirmed that membrane sheets consisted of intact lipid bilayers by staining with DiI, a lipid-soluble fluorescent dye, and were devoid of all other organelles except for docked vesicles by scanning electron microscopy (data not shown). The distribution of PI(4,5)P 2 in the membrane sheets was investigated using a specific antibody to PI(4,5)P 2 (AM-212) (28). Membrane sheets were fixed with 4% PFA and then immunostained using the standard protocol with an anti-PI(4,5)P 2 antibody and a fluorescently labeled Fab fragments of a secondary antibody. As shown in Fig. 1A, the anti-PI(4,5)P 2 antibody yielded numerous small clusters that were scattered over the membrane sheets. Although the specificity of anti-PI(4,5)P 2 antibody used in this study (AM-212) has been established in several systems (28,34), we confirmed whether the AM-212 specifically recognized PI(4,5)P 2 in membrane sheet preparation using neomycin, which strongly binds PI(4,5)P 2 with high affinity (14). Immunostaining of membrane sheets using AM-212 and anti-syntaxin antibody in the presence of neomycin demonstrated that neomycin selectively eliminated the signals derived from AM-212, clearly indicating the specificity of AM-212 for PI(4,5)P 2 in membrane-sheet preparation (Fig. 1B). To further ascertain the availability of membrane sheet preparation for studying the distribution of PI(4,5)P 2 in the plasma membrane, we examined the effects of overexpression of constitutive active Arf6 mutant (Arf6(QL)), which was shown to reduce the amount of PI(4,5)P 2 in the plasma membrane (15). Since the efficiency of transfection was not very high, we co-transfected myc-tagged human growth hormone (hGH-myc) to label LDCVs in Arf6(QL)-transfected PC12 cells (35). As shown in Fig. 1C, signals derived from the anti-PI(4,5)P 2 antibody were remarkably decreased in the membrane sheets of PC12 cells transfected with hGH-myc and Arf6(QL). Quantitative analysis revealed that PI(4,5)P 2 immunoreactivity was selectively reduced in Arf6(QL)-transfected membrane sheets (Fig. 1D). These results, consistent with the previous results in intact and cracked PC12 cells (15), indicated that membrane sheet preparation was suitable to investigate the distribution of PI(4,5)P 2 in the plasma membrane.
The punctate pattern of PI(4,5)P 2 immunoreactivity could be detected in PFA-fixed membrane sheets; however, there remained a possibility that lipids, sphingolipid, and glycosyl phosphatidylinositol-anchored proteins formed large aggregates in the plasma membrane due to cross-linking by the antibodies (19,36,37). To exclude such an artificial capping effect, the distribution of PI(4,5)P 2 was further examined by transient transfection of EGFP-tagged PH domain of phospholipase C␦ (PH-EGFP), which selectively binds to PI(4,5)P 2 in 1:1 stoichiometry (38, 39). As shown in Fig. 1E, PH-EGFP also exhibited a punctate pattern similar to that observed by the anti-PI(4,5)P 2 antibody over the membrane sheets of PH-EGFP-expressing PC12 cells. PI(4,5)P 2 clusters existed not in LDCVs but in the plasma membrane, since no significant PI(4,5)P 2 immunoreactivity was detected in the immunoisolated LDCVs (data not shown), corresponding to a previous report (14). Next, we examined whether these PI(4,5)P 2 clusters could be observed in living PC12 cells. PH-EGFP-expressing PC12 cells on the coverslips were observed by evanescent microscopy enabling the illumination of a thin layer immediately adjacent to the coverslips (30,40). Evanescent microscopy FIG. 1. PI(4,5)P 2 is accumulated at specialized sites in the plasma membrane. Membrane sheets were prepared from PC12 cells as described under "Materials and Methods." A, membrane sheets were fixed with 4% PFA and immunostained with anti-PI(4,5)P 2 antibody, AM-212. To minimize the capping effect, fluorescein isothiocyanateconjugated Fab fragment of the antibody was used as a secondary antibody. Fluorescein isothiocyanate signals were detected by confocal laser-scanning microscopy. B, membrane sheets were immunostained using AM-212 and anti-syntaxin antibody in the absence (top) or presence (bottom) of 3 mM neomycin. Alexa488-conjugated anti-rabbit IgG (for syntaxin) and Alexa555-conjugated anti-mouse IgG 3 (for AM-212) were used to visualize the primary antibodies. Images were detected by confocal laser-scanning microscopy and captured in multitrack mode with the same condition. C, membrane sheets were prepared from PC12 cells transiently co-transfected with hGH-myc and Arf6 mutant (Arf6(QL)) or with hGH-myc and an empty vector (mock) and immunostained for PI(4,5)P 2 , hGH-myc, and syntaxin. Alexa488-conjugated anti-mouse IgG 1 (for myc), Alexa555-conjugated anti-mouse IgG 3 (for PI(4,5)P 2 ), and Alexa647-conjugated anti-rabbit IgG (for syntaxin) were used to visualize the primary antibodies. Images were detected and captured as described above. D, quantitative analysis of signal intensity of syntaxin and PI(4,5)P 2 in the membrane sheets of mock-or Arf6(QL)transfected PC12 cells. A 9-m 2 area was arbitrarily selected to estimate the signal intensity of syntaxin and PI(4,5)P 2 in each membrane sheet. Signal intensities in Arf6(QL)-transfected membrane sheets were normalized by those in mock-transfected membrane sheets. Values are the means Ϯ S.E. of nine representative sheets. E, membrane sheets were prepared from PC12 cells transiently expressing PH-EGFP and fixed with 4% PFA. PH-EGFP signals were detected by confocal laser-scanning microscopy. F, PH-EGFP in the living PC12 cells was detected by evanescent microscopy. The arrowheads represent the sites of PH-EGFP accumulation. Bars, 5 m.
revealed that PH-EGFP could be detected as a footprint image and punctate pattern in living PC12 cells (Fig. 1F) despite the high background fluorescence possibly due to a certain amount of cytosolic PH-EGFP. These results clearly demonstrated that PI(4,5)P 2 was concentrated in a specialized microdomain, termed the PI(4,5)P 2 microdomain, in the plasma membrane of PC12 cells.
Distribution of PI(4,5)P 2 Microdomains and Exocytotic Machinery-Several studies have revealed that PI(4,5)P 2 plays an important role in exocytosis, but the spatial relationship between exocytotic sites and PI(4,5)P 2 microdomains is largely unknown. To examine this relationship, we investigated whether PI(4,5)P 2 microdomains were colocalized with exocytotic machinery in the membrane sheets by immunostaining. Since an exocytotic reaction occurred at syntaxin clusters (29,41), membrane sheets were immunostained with specific antibodies against PI(4,5)P 2 and syntaxin. LDCVs were visualized using an antibody against chromogranin B accumulated within LDCVs of PC12 cells (42). As shown in Fig. 2A, a number of LDCVs docked on the membrane sheets were colocalized with PI(4,5)P 2 microdomains and/or syntaxin clusters. Quantitative analysis revealed that 47.6 Ϯ 2.9% of LDCVs were colocalized with syntaxin clusters, whereas 18.0 Ϯ 3.3% and 12.3 Ϯ 2.5% were colocalized with PI(4,5)P 2 microdomains and both PI(4,5)P 2 microdomains and syntaxin clusters, respectively. Significant colocalizations of LDCVs with syntaxin clusters and PI(4,5)P 2 microdomains were still obvious after correcting the overestimation due to the accidental colocalization according to the methods of Lang et al. (31) (Fig. 2B; 32.4 Ϯ 3.0, 7.8 Ϯ 1.9, and 10.3 Ϯ 2.5% for syntaxin clusters, PI(4,5)P 2 microdomains, and PI(4,5)P 2 microdomains and syntaxin clusters, respectively). Although this procedure tends to underestimate the colocalization, we used it throughout this study to completely exclude the possibility of accidental colocalization. These results suggest that syntaxin clusters were not essential for the docking of LDCVs with the plasma membrane, and a significant number of LDCVs docked on syntaxin clusters were colocalized with PI(4,5)P 2 microdomains.
Next, we examined the effect of mPIP5KI␤ expression on Ca 2ϩ -dependent exocytosis by using hGH co-transfection system (35). As shown in Fig. 5A, Ca 2ϩ /ionomycin-evoked hGH release was markedly enhanced by the expression of mPIP5KI␤ (mock, 23.4 Ϯ 0.3%; mPIP5KI␤, 28.7 Ϯ 0.9%). This effect seemed to be specific, since no significant enhancement was achieved by the expression of a kinase activity-deficient mutant of mPIP5KI␤, K178A (43) (25.0 Ϯ 0.3%). Fig. 5B shows the Ca 2ϩ -dependent release of hGH from mPIP5KI␤-expressed or -unexpressed PC12 cells induced by various high K ϩ depolarization. The significant enhancement of hGH release by mPIP5KI␤ expression was observed in all KCl concentrations, and an almost identical extent of activation of around 20% was achieved. These results suggest that the expression of mPIP5KI␤ enhanced Ca 2ϩ -dependent exocytosis not by modulating the Ca 2ϩ sensitivity of exocytotic machinery, but by increasing the sites for exocytosis.
Heterogeneity of PI(4,5)P 2 Microdomains in the Plasma Membrane-The lipid raft is a microdomain in the plasma membrane and involves in many cellular functions (16 -18). It is thus important and interesting to assess the relationship between the lipid raft and PI(4,5)P 2 microdomain. As shown in Fig. 6A, part of the PI(4,5)P 2 microdomains was overlapped with Thy-1 immunoreactivity in isolated membrane sheets, indicating that about half of the PI(4,5)P 2 microdomains is distinct from the lipid raft. In order to elucidate the mechanism of PI(4,5)P 2 microdomain formation, we examined the localization of PIP5KI in the membrane sheets derived from mPIP5KI␤-expressed PC12 cells. Immunoreactive signals of mPIP5KI␤-myc yielded a punctate pattern in membrane sheets, and most PI(4,5)P 2 clusters were overlapped with mPIP5KI␤-myc signals (Fig. 6B). Interestingly, a large proportion of the mPIP5KI␤-myc signal was also colocalized with syntaxin clusters (Fig. 6C, data not shown). To further examine the relationship between the lipid raft and PI(4,5)P 2 microdomain, we performed a lipid raft floatation assay and examined the recovery pattern of PIP5KI. Thy-1 and growth corn-associated protein (GAP-43) was enriched in cholesterol-enriched lipid raft fractions (Fig. 6D, data not shown), consistent with previous reports (19). In striking contrast, a large part of syntaxin was not recovered in the lipid raft fractions but was enriched in the bottom fractions, containing a substantial amount of endogenous PIP5KI. All of these results suggest that the PI(4,5)P 2 microdomains at syntaxin clusters were likely to be formed by the selective recruitment of PIP5KI to syntaxin clusters, and such nonraft PI(4,5)P 2 microdomains were concerned with exocytosis.

DISCUSSION
In this study, we found that 1) PI(4,5)P 2 formed microdomains in the plasma membrane of PC12 cells; 2) part of the PI(4,5)P 2 microdomains was accumulated at syntaxin clusters with docked LDCVs; and 3) the number of PI(4,5)P 2 -syntaxin-LDCV complexes was correlated with the Ca 2ϩ -dependent exocytotic activity of LDCVs. These results suggest that PI(4,5)P 2 microdomain formation at syntaxin clusters was essential for LDCVs to acquire exocytotic activity.
PI(4,5)P 2 Microdomains in the Plasma Membrane-Aikawa and Martin (15) and Holz et al. (14) showed that PI(4,5)P 2 was FIG. 4. The number of PI(4,5)P 2 microdomains and the proportion of LDCVs colocalized with PI(4,5)P 2 microdomains and syntaxin clusters were increased by mPIP5KI␤ expression. A, membrane sheets were prepared from PC12 cells transiently co-transfected with hGH-myc and mPIP5KI␤ or with hGH-myc and an empty vector (mock) and immunostained for PI(4,5)P 2 (top) and myc (middle). Al-exa488-conjugated anti-mouse IgG 1 (for myc) and Alexa555-conjugated anti-mouse IgG 3 (for PI(4,5)P 2 ) were used to visualize the primary antibodies. Images were detected and captured as described in the legend to Fig. 1B. B, membrane sheets derived from PC12 cells transiently cotransfected with hGH-myc and mPIP5KI␤ (filled bars) and with hGH-myc and an empty vector (open bars) were fixed with 4% PFA and immunostained for syntaxin, PI(4,5)P 2 , and myc. Primary antibodies were visualized, and images were detected and captured described in the legend to Fig. 2A. Quantification of the colocalization of hGH-myc signals with PI(4,5)P 2 microdomains and/or syntaxin clusters was performed and represented as described in the legend to enriched in the plasma membranes of PC12 cells and adrenal chromaffin cells by using an anti-PI(4,5)P 2 antibody and the EGFP-tagged PH domain of phospholipase C␦, respectively, but they did not observe the two-dimensional distribution of PI(4,5)P 2 in the plasma membrane. Recently, Laux et al. (19) showed that the anti-PI(4,5)P 2 antibody yielded a punctate pattern in various clonal cells and in the growth cone of neurons by a standard immunocytochemical procedure. Since they observed PI(4,5)P 2 microdomains in whole-cell preparation, it was not clear whether these signals were derived from the plasma membrane or cytosolic organelles. In this study, we clearly showed that the PI(4,5)P 2 microdomain was located in the plasma membrane by using isolated plasma membranes. Membrane sheets were suitable to investigate the spatial distribution of molecules in the plasma membrane because large areas of the plasma membrane were observed in the single plane and exclude the background noises derived from cytosolic factors. Since Ca 2ϩ -dependent exocytosis of the docked vesicle was observed (29), proteins and lipid components essential for exocytosis were preserved in the membrane sheet preparation in the membrane sheet. It is known that the capping formation by antibody treatments sometimes gives false signals in immunocytochemistry (36,37). To minimize such an artifact effect, Laux et al. (19) used glutaraldehyde for the fixation. In this study, we completely excluded the capping artifact by using PH-EGFP for the detection of PI(4,5)P 2 microdomains. Even using such a monovalent marker for PI(4,5)P 2 , punctate signals were still observed in isolated membrane sheets. Furthermore, we could detect the punctate distribution of the EGFP-tagged PH domain in living cells as reported previously (45). All of these results clearly demonstrated that PI(4,5)P 2 was concentrated in specialized microdomains in the plasma membrane of PC12 cells.
Heterogeneity of PI(4,5)P 2 Microdomains in the Plasma Membrane-Growing evidence indicates that lipids can form specialized domains with compositions and physical properties that differ from the average properties of the membrane. Furthermore, there seem to be many kinds of lipid rafts with different protein composition and functions (18). Since PI(4,5)P 2 plays an important role in the regulation of many types of cellular functions (24), there seem to be several kinds of PI(4,5)P 2 microdomains in the plasma membrane. Laux et al. (19) showed that some PI(4,5)P 2 microdomains were colocalized with myristoylated alanine-rich C kinase substrate, a protein accumulated in the lipid raft, and may be involved in the regulation of actin cytoskeleton. In this study, we showed that part of PI(4,5)P 2 microdomains was colocalized with Thy-1 clusters, suggesting that some PI(4,5)P 2 microdomains were formed in the lipid raft. However, we also found that a substantial number of PI(4,5)P 2 microdomains was accumulated not in Thy-1 clusters, but in syntaxin clusters with docked LDCVs. These results were consistent with previous studies demonstrating that only half of cellular PI(4,5)P 2 was recovered in Thy-1-enriched lipid raft fractions with ultracentrifugation through a sucrose gradient (20,21). Thus, it is quite possible that some PI(4,5)P 2 microdomains play important roles in the exocytosis of LDCVs.
The Functional Relevance of PI(4,5)P 2 Microdomains in Exocytotic Activity-Although previous studies suggested that syntaxin clusters form a release site in the plasma membrane, it was not the only determinant of releasable LDCVs, since not all of LDCVs colocalized with syntaxin clusters fused with the plasma membrane in response to Ca 2ϩ in the membrane sheet preparation derived from PC12 cells and living MIN6 cells (29,41). We found in this study that the proportion of LDCVs colocalized with syntaxin clusters was not correlated with exocytotic activity, but that colocalized with PI(4,5)P 2 microdomains and syntaxin clusters was altered in membrane sheets with different exocytotic activity. Furthermore, we found that the exogenously expressed mPIP5KI␤ increased the proportion of LDCVs colocalized with PI(4,5)P 2 microdomains and syntaxin clusters and enhanced LDCVs exocytosis. Since the extent of stimulation of hGH release was much smaller than that of colocalization of LDCVs with PI(4,5)P 2 , PI(4,5)P 2 -binding protein, such as synaptotagmin and CAPS, was likely to became a rate-limiting factor of exocytosis in the presence of excess numbers of PI(4,5)P 2 microdomains. The subtype specificity seemed to exist in the action of PIP5KI, since the expression of PIP5KI␥, which was endogenously expressed in the brain but not in PC12 cells used in this study (46) (data not shown), failed to enhance the releasing activity in PC12 cells (15). All of these results indicate that the formation of PI(4,5)P 2 microdomains in syntaxin clusters is crucial for the release sites to acquire Ca 2ϩ -dependent exocytotic activity.
The Mechanisms for the Formation of PI(4,5)P 2 Microdomains at Release Sites-If the formation of PI(4,5)P 2 microdomains at syntaxin clusters is the determinant for a releasable site of LDCVs, it is quite interesting and important to elucidate the regulatory mechanism for the formation of PI(4,5)P 2 microdomains. The selective localization of PIP5KI to a syntaxin cluster might be a mechanism of the formation of PI(4,5)P 2 microdomains at the release site, since the exogenously expressed mPIP5KI␤ was colocalized with syntaxin clusters as well as both PI(4,5)P 2 microdomains (Fig. 6, B and C). In the lipid raft floatation assay, endogenous PIP5KI in PC12 cells was not recovered in the Thy-1-enriched lipid raft fraction but rather in the syntaxin-enriched fractions (Fig. 6D) (29). Syntaxin could be associated with Munc13-1, which binds to mSec7, a guanine exchange factor of Arf6, and that in turn binds to PIP5KI for the activation (15, 47-50). Thus, it is possible that a multiprotein complex formation of these proteins might be the mechanism of recruitment of PIP5KI to the FIG. 6. Two distinct types of PI(4,5)P 2 microdomains in the membrane sheet of PC12 cells. A, membrane sheets were immunostained for PI(4,5)P 2 and Thy-1, and the images were detected and captured as described in the legend to Fig. 4A. A substantial number of PI(4,5)P 2 microdomains were colocalized with Thy-1 clusters (arrowheads). B and C, the membrane sheets from PC12 cells transiently transfected with myc-tagged mPIP5KI␤ were immunostained for PI(4,5)P 2 and either myc (B) or syntaxin (C) as described in the legend to Fig. 4A. Alexa488-conjugated anti-mouse IgG 1 and Alexa594-conjugated anti-rabbit IgG were used as secondary antibodies for the detection of myc and syntaxin, respectively. A large population of mPIP5KI␤ was colocalized with PI(4,5)P 2 microdomains (B) and syntaxin clusters (C). Bars, 2 m. D, Triton X-100-solubilized homogenate of PC12 cells were subjected to a lipid raft floatation assay as described under "Materials and Methods," and the proteins in each fraction were subjected to immunoblotting using an antibody against Thy-1, GAP43, and PIP5KI (top). The migration positions of the antigen proteins were indicated by arrowheads on the right. The amount of cholesterol in each fraction was determined using a Cholesterol E-test kit (bottom). release site. Interestingly, Munc13-1 is one of the priming factors of catecholamine release from adrenal chromaffin cells (51). On the other hand, we also observed that some myctagged mPIP5KI␤ was not colocalized with syntaxin clusters, suggesting that other proteins are involved in the recruitment of PIP5KI at nonrelease sites. In preliminary studies, we observed that adaptin ␤, one of the adapter proteins involved in endocytosis, was colocalized with PI(4,5)P 2 microdomains. Further studies are necessary to elucidate the mechanisms of PI(4,5)P 2 microdomain formation at various functional sites for the understanding of the regulatory mechanisms of secretion as well as various PI(4,5)P 2 -related cellular functions.