Regulated exocytosis: a novel, widely expressed system

articles

Regulated exocytosis: a novel, widely expressed system Barbara Borgonovo*, Emanuele Cocucci*‡, Gabriella Racchetti*§, Paola Podini*§, Angela Bachi§¶ and Jacopo Meldolesi*§# *Department of Neuroscience, Vita-Salute San Raffaele University and Excellence Centre in Cell Differentiation, via Olgettina 58, 20132 Milan, Italy ‡Department of Pharmacology, University of Milan, via Vanvitelli 32, 20133 Milan, Italy §IRCCS San Raffaele, Dibit, via Olgettina 58, 20132 Milan, Italy ¶Department of Cell Pathology, Vita-Salute San Raffaele University and Excellence Centre in Cell Differentiation, via Olgettina 58, 20132 Milan, Italy #e-mail: [email protected]

Published online: 25 November 2002; DOI: 10.1038/ncb888

Electrophysiological studies in some secretory and non-secretory cells have identified an extensive form of calciuminduced exocytosis that is rapid (hundreds of milliseconds), insensitive to tetanus toxin and distinct from regulated secretion. We have now identified a marker of the process, desmoyokin-AHNAK, in a clonal derivative of the neuronal cell line, PC12. In resting cells, desmoyokin-AHNAK is localized within the lumen of specific vesicles, but appears on the cell surface during stimulation. Desmoyokin-AHNAK-positive vesicles exist in a variety of cells and tissues and are distinct from the endoplasmic reticulum, Golgi, trans-Golgi, endosomes and lysosomes, and from Glut4 and constitutive secretion vesicles. They seem to be involved in two models of plasmalemma enlargement: differentiation and membrane repair. We therefore propose that these vesicles should be called ‘enlargosomes’.

egulated exocytosis is the process by which, after an appropriate stimulus, the membrane of specific organelles fuses with and is incorporated into the plasma membrane of the cell1. Regulated exocytosis is often associated with regulated secretion, a process typical of neurons, endocrine and exocrine cells by which secretory granules and vesicles discharge their content under precise temporal and spatial requirements (time constants in neurons and neurosecretory cells are 1–2 ms and several seconds for clear and dense vesicles, respectively2). In contrast to neurons, cells that are not specialized for secretion are widely believed to be incompetent for regulated exocytosis. Recently, however, electrophysiological capacitance assays of nonneurosecretory cells (Chinese hamster ovary (CHO) and 3T3 fibroblasts) and of a clone of the PC12 phaeochromocytoma cell line specifically defective in neurosecretion (named PC12-27), revealed prompt (hundreds of milliseconds) and extensive (15–30%) surface increases after elevation of cytoplasmic calcium concentration ([Ca2+]i) induced by flash photolysis of intracellular caged calcium (refs 3–5). On the basis of extensive evidence, these responses can only be due to calcium-regulated exocytosis of small organelles (less than 100 nm in diameter4,5). Moreover, in a fraction of the CHO cells analysed, the fusion of a few larger organelles (diameter up to 1.5 µm, possibly secretory lysosomes6) was also observed3. Interestingly, the responses of PC12-27 cells were completely unaffected by tetanus toxin, a classical blocker of synaptic and neurosecretory vesicle–granule fusion5. From these results one can conclude that a new type of exocytosis, fast and regulated but mechanistically distinct from the classical form, operates in at least three lines of cultured cells. Evidence of this new process in other cell types, its functional significance and the nature of the exocytotic vesicles remained to be discovered. Here we report on studies utilizing a comprehensive cell biological approach in several PC12 cell clones, competent or defective for neurosecretion, and in a variety of other cell types and tissues. The new exocytotic process was found to have a specific protein marker present in a population of vesicles distinct from any other cytoplasmic organelle. In some cell types, in which expression of the

R

marker is absent or low during growth, an accumulation was observed during differentiation.

Results Identification of an exocytotic marker. Our first task was the identification of one or more markers specific for the organelles sustaining the new calcium-induced exocytotic responses revealed by capacitance assays5. This was achieved by raising monoclonal antibodies against proteins appearing at the surface of PC12-27 cells treated with the calcium ionophore ionomycin, identified by differential biotinylation as described in Methods. Among the generated monoclonal antibodies, one (termed Ab) specific for a very-high-molecular-mass protein is expressed by two neurosecretion-defective PC12 clones, PC12-27 and Trk, and by CHO fibroblasts. In contrast, the antigen was virtually absent in wild-type PC12 cells and in two secretion-competent clones, 16 and 38 (ref. 7) (Fig. 1Aa). In two-dimensional (2D) western blots of immunoprecipitates from PC12-27 cell lysates, various moderately acidic Ab-positive spots appeared, which were excised and sequenced by tandem nanoelectrospray mass spectrometry. All sequences obtained (Fig. 1Ac) were found to be included into the long (128 amino acids), numerous (more than 30) and characteristic repeats of a huge (relative molecular mass ∼700,000 (Mr ~700K)) protein expressed in a variety of cells in multiple species, named both desmoyokin and AHNAK (abbreviated here as dA). In humans, the >4,300-amino-acid repeat core of dA is continuous with both a 251-amino-acid amino-terminal domain and a 1,002amino-acid carboxy-terminal domain8. Further studies confirmed the identification of the Ab antigen as dA. First, the purified product (Mr ~200K) of an Escherichia coliexpressed dA construct, z80, including the glutathione S-transferase (GST) tag, several repeats and 80 amino acids of the C-terminal domain9, was intensely Ab-positive (Fig. 1Ab). Second, mass-spectrometric analysis of Ab-positive bands appearing in the gels of GST-containing proteins purified from z80-expressing E. coli revealed in all cases sequences included within dA repeats (not

NATURE CELL BIOLOGY VOL 4 DECEMBER 2002 www.nature.com/naturecellbiology © 2002 Nature Publishing Group

955

articles A Mr(K)

B PC12 Trk

27

16

38 CHO

Mr(K)

600 -

Mr(K)

a z80

250 -

e.v.

- 250

160 105 -

200 -

- 160

75 -

b

- 105

c - 75

a P1

C

b P2

P3

P2 + P3 +S

S

c P2 + P3 LB +S

D

a

b

c

a

b

c

Figure 1 Ab antigen identification and distribution. Aa, Western blots of wildtype PC12 and CHO cells together with four PC12 clones, two (Trk and 27) defective and two (16 and 38) competent for regulated secretion. Notice that no Ab-positive band appears in secretion-competent PC12, whereas in Trk, 27 and CHO a major band with a very large Mr (>600K) is seen. Ab, Strongly Ab-positive western blot band of the z80 protein product (GST fused to a dA fragment) purified from E. coli transformed with a vector including the corresponding cDNA construct. The preparation purified from E. coli transformed with the empty vector (e.v.) is completely negative. Ac, Mass-spectrometric analysis of peptides obtained from Ab-positive spots, derived from immunoprecipitates and separated by 2D PAGE, revealed in all cases sequences from the internal repeats of dA. B–D, Intracellular distribution of dA, in PC12-27 cells, as revealed by subcellular fractionation (B), confocal immunocytochemistry (red) (C) and immunogold electron microscopy (D). When homogenates were separated by differential ultracentrifugation, the dA western blot band was recovered primarily in the P2 and P3 particulate fractions, whereas the soluble fraction (S) was completely negative (Ba). In the P1 fraction, labelling was much less intense, concentrated not only in the dA band but also in a few lowermolecular-mass bands (Ba) generated by the digestion of dA by cytosolic proteases. This labelling was due not to nuclei but to large membrane sheets (plasmalem-

ma) as revealed by P1 flotation experiments (not shown). The luminal localization of dA is shown in Bb and Bc. Exposure of the intact P2 plus P3 combined fraction to the soluble S fraction had no effect on the dA band (Bb, right lane). Permeabilization of the P2 plus P3 fraction and incubation with LB was also ineffective (Bc, left lane). However, when the fraction was previously permeabilized and then exposed to S (Bc, right lane) the Ab-positive band was converted into a ladder. Similar results were obtained by using trypsin instead of the S fraction (not shown). The distribution of dA in the lumen of cytoplasmic particles was confirmed by the immunofluorescence data (C). In PC12-27 cells whose plasma membrane had been permeabilized with streptolysin-O, the Ab signal was negative (Cb), whereas with antibodies against cytosolic proteins (such as CAMKII, Ca) the signal was prominent. Only when the cells pretreated with streptolysin-O were incubated in DB+, with the resulting permeabilization of the intracellular membranes, strong Ab-positive punctae appeared in the cytoplasm, concentrated in the subplasmalemma area (Cc). Scale bars, 10 µm (that in Ca also applies to Cb). Conventional electron microscopy revealed that the subplasmalemma rim of PC12-27 cytoplasm contained numerous small (50–70 nm in diameter) vesicles (arrows in Da), which in LR White sections processed for Ab immunogold labelling often appeared positive (arrows in Db and Dc). Scale bar, 0.2 µm (also applies to Db and Dc).

shown). Third, binding of a polyclonal anti-dA antibody (named KIS), raised against the N-terminal 16-amino-acid polypeptide of the repeats10 to the z80 protein (revealed by enzyme-linked immunosorbent assay (ELISA)) was competed for by Ab (see Supplementary Information, Fig. S1). We conclude that the target sequence of Ab lies in the internal dA repeats, most probably close to their N-terminal sequence. Cellular distribution of dA. The distribution of dA was investigated by a variety of techniques, first in PC12-27 cells (Figs 1–4) and then in CHO, HeLa and other cell types (Fig. 4 and Supplementary Information, Fig. S2). Confocal microscopy showed strong labelling over numerous punctae, preferentially distributed in the outer rim of the cytoplasm. The nucleus was consistently negative. In cell lines such a distribution was evident both in sparsely distributed, rapidly growing cells and in thick monolayers. Under the electron microscope, the Ab immunogold signal was weak owing to the presence of glutaraldehyde (0.2%, necessary to preserve ultrastructure) combined with formaldehyde in the fixative.

However, in 32 ultrathin sections of PC12-27 cells (analysed area 110 µm2), more than 30% of the immunogold particles were seen in the 0.5-µm-thick rim below the plasmalemma (Fig. 1D), corresponding to less than 10% of the total analysed cytoplasm. This rim is rich in small vesicular profiles ∼60 nm in diameter, most of which were labelled (Fig. 1Da–c). The remaining gold particles, scattered over the rest of the cytoplasm (90% of the area), were randomly distributed. We conclude that the vesicles are specifically labelled and therefore most probably correspond to the positive punctae revealed by immunofluorescence. When resting PC12-27 cells, suspended in 0.32 M sucrose supplemented with an antiprotease cocktail, were mildly homogenized in a cell cracker and fractionated by differential centrifugation, dA was recovered largely in the particulate fractions (P2 and P3 (see Methods), enriched in mitochondria and microsomes), whereas the final supernatant (S, containing the cell cytosol) was consistently negative (Fig. 1Ba). With harsher homogenization (with a Dounce-type homogenizer), a large part of the P2 labelling moved

956

NATURE CELL Group BIOLOGY VOL 4 DECEMBER 2002 www.nature.com/naturecellbiology © 2002 Nature Publishing

articles Mr(K)

a

b

- 250 - 160 - 105

c

- 75

f h d

e

g

P1 P2 P3 S

Figure 2 Surface distribution of dA in stimulated PC12-27 cells. In intact PC12-27 cells, surface immunofluorescence labelling with Ab was almost imperceptible at rest (a) and rose considerably after treatment with ionomycin in calciumcontaining medium (b). In ionomycin-treated cells (c), the surface distribution of dA (green) coincided largely with that of the plasmalemma-directed lectin from Lycopersicon esculentum (red). The appearance of dA on the surface in ionomycintreated PC12-27 cells was also supported by surface biotinylation, showing an almost fivefold increase in the surface signal with respect to resting cells (right and left bands in g) with no change in the total dA pool (corresponding bands in f), and by subcellular fractionation, showing a large increase of the signal in the plasma membrane sheets of the P1 fraction (compare the western blots of h with those of resting cells in Fig. 1Ba). When stimulation was not by ionomycin but by photolysis of caged calcium, PC12-27 cells promptly increased in volume and the surface appearance of dA was extensive (d). When analysed by deconvolution, the dA labelling of photolysed cells was found to consist of numerous small dots (e), as expected for an exocytosis-generated signal. Scale bars, 10 µm (that in b also applies to a, and that in e also applies to c and d).

to P3 (data not shown). In the heavy P1 fraction, labelling was much less intense, distributed not only in the high-molecular-mass band but also in a few smaller bands. When P1, resuspended in 1.8 M sucrose covered with a 0.3 M cushion, was spun for 60 min at 100,000g, most dA floated at the interface (data not shown), showing that its localization is not in the nucleus but in large, light structures, namely plasma membrane sheets. Taken together, immunocytochemical and fractionation data consistently indicate the intracellular distribution of the Ab antigen to be cytoplasmic and particulate. The luminal distribution of dA emerged consistently from both immunofluorescence and the analysis of subcellular fractions. First, when the plasma membrane of living cells was permeabilized with streptolysin-O before application of antibodies (a condition in which proteins in, or exposed to, the cytosol are immunolabelled; Fig. 1Ca), no dA was labelled (Fig. 1Cb). Punctate dA labelling appeared only in cells preincubated in detergent-containing dilution buffer (DB+; Fig. 1Cc). Second, in the isolated P2 and P3 fractions resuspended without solubilization in either the S fraction (10 min, 4 °C) or a trypsin solution (10 min, 22 °C), the large dA band remained unaffected (Fig. 1Bb). When the same treatments were applied to the fractions resuspended in lysis buffer (LB), the large band was converted into a ladder, similar to that observed in the P1 fraction (Fig. 1Cc). In contrast, when the two combined particulate fractions were resuspended in sodium carbonate (0.5 M, pH 11), about 40% of the dA was released (data not shown), as expected for a protein lacking transmembrane sequences. The dA-containing vesicles are exocytotic. The results reported here refer to stimulated PC12-27 cells. Luminal proteins of vesicles competent for regulated exocytosis were expected to appear at the surface when cells are appropriately stimulated. In vivo immunolabelling results showed that this is so with dA (Fig. 2). At rest, the surface of PC12-27 cells is almost unlabelled by Ab (Fig. 2a). In contrast, when stimulation with ionomycin induced an increase in [Ca2+]i more than 5 µM, living cells became rapidly (less than 1 min) immunolabelled (Fig. 2b), showing a pattern coinciding with

a lectin surface marker (Fig. 2c). Consistently, a stimulated-versusresting shift was observed in living PC12-27 cells analysed by fluorescence-activated cell sorting (FACS) for Ab binding (see Supplementary Information, Fig. S3D). In contrast, [Ca2+]i increases of less than 1 µM, induced by ionomycin administered to cells either preloaded with bis-(o-aminophenoxy)ethane-N,N,N´,N´-tetra-acetic acid (BAPTA) (30 min incubation with 50 nM BAPTA acetoxymethyl ester) or bathed in a calcium-free-EGTA medium, failed to induce any increase in surface immunolabelling (Supplementary Information, Fig. S2E–H). The ionomycin-induced surface labelling was long-lasting, with no appreciable decrease at 30 min after ionophore application (data not shown). Together with capacitance data5, this result suggests that the recycling of exocytotic vesicle membrane occurs slowly. Attempts to immunolabel patch-clamped PC12-27 cells were unsuccessful. In contrast, experiments of caged-calcium photolysis, the technique employed in capacitance experiments5 to induce fast [Ca2+]i jumps, were quite revealing. Sparse monolayers, previously loaded with nitrophenyl EGTA (NP-EGTA), were flashed with a xenon arc lamp, transferred to ice, processed for Ab immunofluorescence and finally fixed. Single flashed cells showed prompt swelling, as expected for an exocytotic burst. Concomitantly, a dA signal appeared at the surface, a much stronger one than that induced by ionomycin (compare Fig. 2d with Fig. 2b). Non-flashed cells, in which [Ca2+]i remained at the resting level, failed to show any increase in surface signal. When the images of flashed cells were mathematically deconvoluted, the labelling appeared punctate, as expected after exocytosis of membrane-bound vesicles (Fig. 2e; see also Supplementary Information, Fig. S3). The photolysis-induced surface transfer of dA was unaffected by tetanus toxin; however, it was active in PC12-27 cells, as revealed by the cleavage of a target, cellubrevin (data not shown). Exocytotic transfer of dA to the PC12-27 plasma membrane was also confirmed by subcellular fractionation and by the surface biotinylation of dA. For the former, the dA patterns of stimulated and resting cells were similar in P2 and P3 (large bands) and S (no trace of dA) (Fig. 2h). In contrast, the plasma membrane labelling of P1 was much more prominent and was distributed as a ladder of bands resembling that observed after the proteolysis of solubilized P2–P3 (compare Fig. 2h with Fig. 1Bc). The result is not surprising because plasma membrane sheets are not sealed, and surface proteins therefore remain in direct contact with cytosolic proteases. For surface biotinylation of dA, resting and ionomycin-stimulated PC12-27 cells were cooled, biotinylated on the surface and then lysed. Immunoprecipitates were probed with either Ab or streptavidin, to reveal total and surface-exposed dA, respectively. Stimulation induced a large (almost fivefold; Fig. 2g) increase in the surface-exposed band, with no change in the total pool (Fig. 2f). Unique nature of the new exocytotic vesicles. Confocal analysis of PC12-27 cells doubly labelled with Ab together with antibodies or agents specific for various cytoplasmic organelles failed to reveal any consistent overlap. In particular, the distribution of the dApositive punctae in the proximity of the plasmalemma did not superimpose the endoplasmic reticulum network, revealed by calreticulin (Fig. 3a), the Golgi complex and trans-Golgi network stained with mannosidase II and tgn38, respectively (Fig. 3b, c), early and recycling endosomes (EEA1 and transferrin receptor, Fig. 3d, e), late endosomes, multivesicular bodies and lysosomes (lamp1 (ref. 11) and Lucifer yellow; Fig. 3f, g), the insulin-sensitive Glut4 vesicles (insulin-regulated aminopeptidase (IRAP)12; Fig. 3h), and vesicles of constitutive secretion (labelled by the transfected green fluorescent protein (GFP)–chromogranin B chimaera13 that in PC12-27 cells, in contrast to wild-type PC12, is secreted via the constitutive pathway14; Fig. 3i). Finally, dAmarked vesicles exclude DAMP (β-(2,4-dinitroanilino)-3´amino-N-methyldipropylamine) and are therefore not acidic (data not shown).

NATURE CELL BIOLOGY VOL 4 DECEMBER 2002 www.nature.com/naturecellbiology © 2002 Nature Publishing Group

957

articles Similarly, various primary cultures (dendritic cells and neurons; Fig. 4Ca–c) are dA-positive. In various organs of the rat, expression is widespread but not ubiquitous (see Fig. 4A). In adult rat tissues, dA is present in various cell types. In the pancreas, acinar cells are labelled moderately, whereas a minor component of islets, distinct from insulin and glucagon-secreting cells, is strongly labelled (Fig. 4Cd and Supplementary Information, Fig. S2D). In the same cell type, dA expression can change during differentiation. The positivity of human dendritic cells was reinforced by treatment with tumour necrosis factor-α or lipopolysaccharide15 (Fig. 4Ca, b). More interestingly, dA-negative wild-type and clonal PC12 accumulate the marker beginning at 6 h after the application of nerve growth factor (NGF) (Fig. 4Da, b). In these cells, dA-positive vesicles are distinct from other organelles, in particular dense granules (chromogranin B; data not shown) and clear vesicles (synaptophysin; Fig. 4Dc). The distinction between exocytotic organelles of the classical and new systems was confirmed in differentiated wildtype PC12 cells treated with tetanus toxin, stimulated with ionomycin and then surface-labelled for both synaptotagmin I and dA. In fact, the appearance of synaptotagmin I decreased considerably (−40%), whereas that of dA was unchanged, confirming the insensitivity of the new system to toxin (Fig. 4E). In fully differentiated cells of both wild-type PC12 and PC12-27 cells, dA-positive vesicles accumulate in neurites, especially at growth cones (Fig. 4Dd). Exocytosis of new vesicles, although distributed over the whole plasmalemma, therefore seems to be more concentrated at sites where processes of membrane expansion take place. To investigate the possible involvement of dA-positive vesicles in plasmalemma wound healing16,17, we exposed living PC12-27 cells to various injuries (including scraping and micropipette puncturing) followed, after 1 min at 22 °C for resealing, by Ab labelling. The results showed surface dA appearance in most wounded cells (data not shown). After electroporation, which induces many small holes in the plasmalemma, cells were all surface-positive (Fig. 4Fa, b) except for a few cells that were probably dead before immunolabelling which showed intense positivity for the extracellular marker tetramethylrhodamine β-isothiocyanate (TRITC)–dextran. In parallel experiments, resealed PC12-27 cells, doubly labelled for dA and the lysosome marker lamp1, showed strong surface labelling for the first but not for the second antibody (Fig. 4Fc), which remained distributed in punctae throughout the cytoplasm (Fig. 4Fd).

a

b

c

d

e

f

g

h

Discussion

i

Figure 3 The intracellular distribution of dA is unique. The intracellular distribution of dA does not coincide with that of any known cytoplasmic organelles. In this figure the left panels (red) show dA, the middle panels (green) the markers of the organelles specified below, revealed by antibodies against specific markers (a–f, h), a dye (g) and a transfected secretory-protein–GFP construct (i). The right panels show the merged images. a, Endoplasmic reticulum (calreticulin); b, Golgi complex (mannosidase II); c, trans-Golgi network (tgn38); d, early endosomes (EEA1); e, recycling endosomes (transferrin receptor); f, g, late endosomes and lysosomes (lamp1 and Lucifer yellow), respectively; h, Glut4 vesicles (IRAP); I, constitutive secretion (GFP–chromogranin B). Scale bar, 10 µm (applies to all panels).

Expression and functions of the new vesicles. The expression and surface translocation of dA after stimulation occur not only in PC12-27 cells but also in other lines: CHO (Fig. 4Bb–d), Rin, HeLa (Fig. 4A, Ba) and others (see Supplementary Information, Fig. S2). 958

The principal tool of our work has been Ab, a monoclonal antibody against dA, the very-high-molecular-mass protein that appears at the surface of PC12-27 cells after stimulation by ionomycin or, even more, after photolysis of a calcium caged compound. The identification of dA was based on multiple criteria: peptide sequencing; specific Ab labelling of a protein in E. coli transformed with a dA construct; and competition of Ab with KIS10, another antibody against dA. By the use of Ab, dA was shown to be localized within small cytoplasmic vesicles that, after appropriate increases in [Ca2+]i, undergo exocytotic fusion. After cell stimulation, no dA is recovered in the incubation medium (data not shown). Because the protein seems to be devoid of transmembrane sequences, its persistence and the ensuing slow recycling to the cytoplasm is most probably dependent on specific interaction(s) with other vesicle component(s). These results nicely complement previous data obtained by capacitance assays3–5. dA is therefore the first cell biological marker of a regulated exocytosis system that was previously known only on the basis of electrophysiology. The vesicular localization of dA was unexpected because the protein, cloned almost 10 years ago, had been reported at many other intracellular sites: associated with desmosomes and particles in the Golgi area, within the nucleus and/or the cytosol, attached to the plasmalemma and in transit to and from the above compartments8,10,18–20.

NATURE CELL Group BIOLOGY VOL 4 DECEMBER 2002 www.nature.com/naturecellbiology © 2002 Nature Publishing

articles HeLa

Rin

Sr

Br

Cbl

Hrt

Mus

Liver

Pcrs

Kdn

Lung

Spleen

A

B

a

b

c

d

a

b

c

d

a

b

c

C

D

d c Fluorescence (arbitrary units)

– TeNT

E a

+ TeNT

200

100

b

0

dA

STG

F a

b

c

d

Figure 4 Expression, intracellular distribution and exocytosis of dA in cells and tissues; effects of differentiation and plasma membrane lesion. A, Ab western blots of cells and adult rat tissues. Sr, adrenal gland; Br, brain; Cbl, cerebellum; Hrt, heart; Mus, skeletal muscle; Pcrs, pancreas; Kdn, kidney. B, dA immmunolabelling (red) of resting HeLa (a) and CHO (b) cells permeabilized with DB+, and intact CHO cells, surface-labelled before (c) and after (d) stimulation with ionomycin. C, dA immunolabelling (red) of cells permeabilized with DB+. a, b, Human dendritic cells, before (a) and after (b) differentiation by lipopolysaccharide, documented also by the appearance of the specific antigen, B7.1 (green). c, Rat hippocampal neuron in culture. d, Rat pancreas, showing low dA labelling (red) of acinar cells, intense labelling of an unidentified cell type within and around the islet, and negativity of most islet cells positive for the secretory vesicle protein synaptophysin (green). Da–c, DB+-permeabilized wild-type PC12 cells, competent for regulated secretion, immunolabelled with Ab (red) before (a) and after (b, c) NGFinduced differentiation. Notice the appearance in the latter of dA punctae that are, however, distinct from the conventional secretory vesicles because they do not coincide with the punctae positive for their specific markers, synaptophysin (green, c) and chromogranin B (not shown). d, NGF-differentiated PC12-27 cells immunola-

belled for dA after permeabilization with DB+ buffer. E, Wild-type PC12 differentiated with NGF and stimulated with ionomycin: surface labelling with both Ab (red) and anti-synaptotagmin I antibody (green). The cells in a had received no pretreatment, whereas those in b had been pretreated with tetanus toxin (TeNT). c, Quantitative deconvolution analyses, expressed as arbitrary fluorescence units and shown as means ± s.d., performed in a total of 30 pretreated and 30 non-pretreated cells. STG, synaptotagmin. Fa, b, PC12-27 cells electroporated in the presence of TRITCdextran (red), then resealed and surface-immunolabelled with Ab (green). Most cells appear surface-positive for dA, often without co-labelling with TRITC. In contrast, the cells strongly positive for the latter are often dA-negative and might already have been dead before immunolabelling. c, Surface immunolabelling of electroporated PC12-27 cells for dA (green) and the lysosomal membrane marker lamp1 (red). The latter is not evident at the cell surface. d, As in c, except that cells were permeabilized in DB+ before immunolabelling. Lamp1-positive punctae (red) are numerous within the cytoplasm and do not coincide in distribution with dA (green), which is mostly at the surface and in the subplasmalemma rim of the cytoplasm. Scale bars, 10 µm (that in Ba also applies to Bb–d and E; that in Ca also applies to Cb, Dc and F; that in Cc also applies to Cd, Da, Db and Dd).

However, part of the previous data were obtained by cell-free experiments in which dA, isolated from detergent-lysed cells, was incubated with other proteins or DNA21–23, whereas others were

based on the expression of truncated constructs19,22,23. Additional, extensive studies were conducted with the parallel use of polyclonal antibodies raised against dA peptides8,10,18,20. One such antibody,

NATURE CELL BIOLOGY VOL 4 DECEMBER 2002 www.nature.com/naturecellbiology © 2002 Nature Publishing Group

959

articles KIS, when employed for immunofluorescence (as described in Methods), failed to induce any specific staining even in dA-positive cells. In contrast, when KIS was used for western blotting, the results were fully consistent with those with Ab, in terms of both the cell expression and subcellular distribution of dA. In contrast, immunocytochemical results obtained with another antibody, FEN10, were completely inconsistent. In agreement with previous results8,18,20, we found strong labelling of nuclei (never labelled by Ab) accompanied by weak decoration of cytoplasmic punctae distinct from those labelled by Ab. Moreover, labelling with FEN was not restricted to Ab-positive cells but extended to others, apparently Ab- and KIS-negative, such as the undifferentiated wild-type PC12 cells employed in our laboratory. We therefore consider that FEN (but not KIS) is specific for, or at least cross-reacts with, another antigen distinct from dA. The existence of non-classical routes of exocytosis, needed for the surface transfer of important proteins (such as receptors, channels and transporters24–28), and distinct from both regulated secretion and constitutive cycling to and from the plasma membrane, has emerged during the past several years. However, the possibility that dA-specific exocytosis corresponds to one or more of the above processes seems unlikely because of their distinct properties, including slower time constants and smaller cumulative membrane areas. The new system also seems to be distinct from secretory lysosomes, that is, exocytotic organelles similar to the acidic granules of blood leukocytes29, which have been described in some but not all cells6,29. dA vesicles are much smaller than lysosomes, are not acidic and are negative for lysosomal markers. Finally, dA-marked vesicles are distinct from clear and dense vesicles of neurosecretory cells. In fact, within wild-type PC12 cells treated with NGF, the two families of organelles exhibit distinct intracellular localizations, and their exocytosis is differentially affected by tetanus toxin: inhibition of neurosecretion and no change in dA surface translocation, respectively. The co-expression of distinct families of regulated exocytotic organelles might occur in many cell types. Capacitance assays in mast cells and pancreatic acinar cells have revealed exocytosis of small vesicles together with that of typical large granules30,31, whereas in chromaffin cells classical neurosecretion is accompanied by a minor, calcium-dependent exocytotic process, unaffected by clostridial neurotoxins32. Further information about dA-positive vesicles was obtained by dual-labelling experiments with Ab and antibodies specific for other cytoplasmic organelles. Consistent with their exocytotic role was the observation that dA-positive vesicles are concentrated in the cytoplasmic rim below the plasmalemma, an area from which other organelles (endoplasmic reticulum, Golgi and trans-Golgi elements) are, in contrast, largely excluded. With regard to endosomes at various stages of maturation, and also multivesicular bodies and lysosomes, dA vesicles were distinguished both by localization and by pH (dA-marked organelles are not acidic). In addition, Glut4 and constitutive secretion vesicles do not coincide with dA vesicles. We suggest that the novel, rapid exocytosis system is sustained by a new class of molecularly distinct organelles. The spectrum of cells and tissues rich in dA is broad, ranging from the brain to the exocrine and endocrine glands, from viscera to lymphoid organs. However, within adult rat tissues its expression is not ubiquitous and even in single cell types dA levels can be variable. Wild-type PC12 cells, and also antigen-presenting dendritic cells and neuroblastoma10 cells, although devoid of or poor in dA when undifferentiated, begin to accumulate the marker shortly after exposure to the differentiating conditions, and transfer it to the surface when stimulated. Expression of the new system might therefore not be a stable property of specific cell types, but could be acquired (or increased) transiently at some stages. The differentiation-induced expression of dA vesicles, together with the slow post-exocytotic recycling of their membranes, suggests their involvement in cell surface enlargement, a function that in neurite extension has already been assigned to a still-unidentified 960

vesicle system33. In non-nerve cells, membrane additions for surface expansion, although necessary for key processes such as phagocytosis and locomotion34,35, have not yet been widely investigated. Wound healing is a ubiquitous process performed by exocytosed membranes16,36. So far, the involvement of various types of organelle has been claimed: specific large granules in oocytes16,36, secretory lysosomes in NRK cells17, endosomes in 3T3 fibroblasts37 and synaptotagmin-positive vesicles in wild-type PC12 cells38. Other healing organelles are expected to exist6. In PC12-27 cells wounded by various treatments, dA was always seen to appear on the surface, whereas the lysosomal–endosomal marker lamp1 was not detected. Taken together, the present and previous results seem to suggest that healing is due to the exocytosis of not merely one but several types of organelle competent for rapid exocytosis, whose expression can vary in different types of cell. In conclusion, the list of cytoplasmic organelles needs to be expanded to include a new type of vesicle competent for rapid, calcium-dependent exocytosis. This mechanism of surface plasticity seems to operate in a variety of cell types. Although the information, especially that relating to the functional role, is not yet complete, it is clear that fusion of these vesicles induces an enlargement of the plasmalemma. We therefore propose that they be called ‘enlargosomes’.

Methods Materials The enhanced chemiluminescence detection system, protein G– and protein A–Sepharose agarose beads, 2D strips, IPGphor system and pGEX-4T-1 vector were from Amersham Biosciences (Uppsala, Sweden); the bicinchoninic acid protein assay reagent, sulpho-N-hydroxysuccinimide (NHS)-LCbiotin, sulpho-NHS-SS-biotin and streptavidin–agarose beads were from Pierce (Rockford, IL); Lucifer yellow was from Molecular Probes (Leiden, The Netherlands); the TRITC-labelled Lycopersicon esculentum lectin, streptolysin-O and TRITC-labelled dextran were from Sigma (St Louis, MO); and NGF was from Alamone labs (Jerusalem, Israel). The purified tetanus toxin was a gift from C. Montecucco. The CgB–GFP plasmid13 and the pGEX4T-1-z80 vector containing the AHNAK fragment9 were kindly provided by H. H. Gerdes and F. Sköldberg, respectively. The following antibodies were obtained as gifts: rabbit polyclonal anti-AHNAK KIS and FEN peptides from E. Shtivelman; rabbit polyclonal anti-mannosidase II from M. Farquhar; rabbit polyclonal anti-EEA1 from A. D’Arrigo; rabbit polyclonal anti-CaMKII and anti-synaptophysin from F. Valtorta; rabbit polyclonal anti-IRAP from P. Pilch; mouse-monoclonal IgG1 anti-lamp1 from I. Mellman; rabbit polyclonal anti-synaptotagmin I (luminal epitope) from A. Malgaroli; and anti-chromogranin B from P. Rosa. The following antibodies were from commercial sources: rabbit polyclonal anti-calreticulin and mouse monoclonal IgG1 anti-TGN38 from Affinity BioReagents (Golden, CO); mouse monoclonal IgG1 anti-transferrin receptor, goat anti-mouse IgGγ, IgG(H+L) and IgM (all conjugated with alkaline phosphatase) from Zymed (San Francisco, CA); mouse monoclonal anti-human B7.1 from BD PharMingen (San Diego, CA); peroxidase-conjugated goat anti-mouse and goat anti-rabbit antibodies from Bio-Rad (Hercules, CA); fluorescein isothiocyanate-conjugated and rhodamine-conjugated goat anti-mouse and goat anti-rabbit antibodies, and colloidal gold particles (12 nm) coated with antimouse IgG from Jackson Laboratories (West Grove, PA); and goat anti-mouse IgG subclasses from Southern Biotechnology Associated (Birmingham, AL). The composition of lysis buffer (LB) was 50 mM HEPES pH 7.5, 150 mM NaCl, 15 mM MgCl2, 1 mM EGTA, 10% glycerol and 1% Triton X-100, and that of dilution buffer (DB+) was PBS plus 1% BSA, 0.3% Triton X-100 and 20% goat serum. Dilution buffer without Triton is indicated as DB−. Trisbuffered saline (TBS) contained 200 mM NaCl, 0.5% Tween 20 and 50 mM Tris, pH 7.4. Cytomix buffer contained 120 mM KCl, 0.15 mM CaCl2, 10 mM potassium phosphate, 2 mM EGTA, 5 mM MgCl2 and 25 mM HEPES, pH 7.6.

Cell cultures All cells were grown at 37 °C in a humidified 5% CO2 atmosphere, using media supplemented with 2 mM L-glutamine and 100 U ml−1 penicillin, streptomycin (Biowhittaker, Verviers, Belgium). In particular, PC12 wild-type and cell clones were grown in Dulbecco’s modified Eagle’s medium supplemented with 10% horse serum (Euroclone, Wetherby, UK) and 5% fetal clone III serum (Hyclone, Logan, UT); hybridoma cells in Iscove’s MDM medium with 10% fetal clone I serum, 5% macrophage conditioned medium (MCM) and 50 µM β-mercaptoethanol. PC12 differentiation was induced by NGF39. Cultured rat hippocampal neurons and human blood-derived dendritic cells were generously provided by A. Bergamaschi and C. Paolucci (San Raffaele Institute, Milan, Italy), respectively.

Isolation of exocytosed membrane proteins PC12-27 cell suspensions, washed twice in PBS at 4 °C, were resuspended and incubated at 10 °C for 30 min in the presence of 2 mg ml−1 sulpho-NHS-LC-biotin (Pierce). After three washes in chilled PBS containing 1% BSA (Boehringer Mannheim, Germany), they were incubated at 37 °C for 5 min in PBS with 15 mM CaCl2 (PBS+) supplemented with 5 µM ionomycin to stimulate exocytosis, then chilled again on ice and exposed for 30 min to 2 mg ml−1 sulpho-NHS-SS-biotin40. After two washes in PBS supplemented with 1% BSA, and two washes in PBS, cells were solubilized in LB. Nuclei were removed

NATURE CELL Group BIOLOGY VOL 4 DECEMBER 2002 www.nature.com/naturecellbiology © 2002 Nature Publishing

articles by low-speed centrifugation at 4 °C (10 min). Cell lysate supernatants were then incubated for 1 h (gentle mixing) with equal volumes of streptavidin–agarose beads (Pierce) at 4 °C, then washed three times in PBS. The biotinylated plasma membrane proteins bound to the beads were then incubated for 30 min in PBS containing 50 mM β-mercaptoethanol at 4 °C, and then centrifuged. The supernatant, containing proteins that appeared at the surface during the exocytotic event, was recovered and used to immunize mice.

Immunization and hybridoma production Two 4-week-old female Balb-C mice were injected twice (with a 2 wk interval) in the footpad with the antigen pool diluted in PBS and emulsified in Freund’s adjuvant (the first time complete, the second incomplete). At 16 days after the first injection, animals were first bled (to perform preliminary analyses of sera with ELISA and western blots of wild-type PC12 and PC12-27 cells) and then given a third boost of antigen in PBS. The animals were killed 3 d later. Lymphocyte fusion with P3X63Ag8 myeloma cells was performed as recommended41. Fused cells were centrifuged and resuspended in hybridoma complete medium, seeded in 96-well plates (105 cells per well), and selected with HAT (hypoxanthine, aminopterin, thymidine) medium.

Antibody purification and screening PC12-27 cells (104) were seeded on microtitre plates precoated with 10 µg ml−1 poly-(L-lysine) in PBS, stimulated with 5 µM ionomycin in PBS+ for 5 min, then centrifuged and fixed with 4% paraformaldehyde. After blocking for 1 h with PBS plus 1% BSA, the wells were incubated with hybridoma supernatants for 2 h at 37 °C, washed, and incubated with 1 µg ml−1 goat anti-mouse IgG or IgM, both conjugated with alkaline phosphatase. Hydrolysis of p-nitrophenyl phosphate was measured with an ELISA reader filtered at 405 nm. Positive hybridomas were tested three times, then cloned by limiting dilution to obtain monoclonal cultures. Antibodies were purified either on hydroxyapatite columns (Bio-Rad), or by affinity chromatography on protein G cartridges (Amersham Biosciences) in accordance with the manufacturer’s instructions.

Western blotting Washed cells were solubilized in ice-cold LB, then quickly centrifuged at 20,000g for 5 min to sediment nuclei, which were discarded. Adult rat tissues were homogenized in 0.32 M sucrose, 5 mM HEPES pH 7.4 supplemented with a protease inhibitor mixture (Sigma), with a Dounce hand homogenizer. Protein concentration was determined with the bicinchoninic acid assay. Fixed amounts of protein were separated by SDS–polyacrylamide gel electrophoresis, then transferred to nitrocellulose filters that were processed at 22 °C. Filters were first blocked for 1 h with 5% non-fat dry milk in TBS, then incubated for 3 h with the primary antibody diluted in PBS plus 3% BSA, washed five times for 10 min in TBS, incubated for 1 h with the appropriate peroxidase-conjugated secondary antibody (1 µg ml−1), washed five times for 10 min in TBS and once in PBS, and developed photographically by chemiluminescence.

Subcellular fractionation and protein digestion PBS-washed preparations of PC12-27 cells, resting and pretreated with ionomycin (5 µM, 5 min) were suspended at 4 °C in a mixture containing 0.32 M sucrose, 5 mM HEPES pH 7.4, and protease inhibitor cocktail. After gentle homogenization by 20 passages in a cell cracker (from EMBL; clearance 12 µm), the suspensions were centrifuged for 10 min at 1,000g to obtain P1, which was solubilized in LB and cleared by centrifugation at 20,000g for 5 min to sediment nuclear debris. Alternatively, P1 was resuspended in 1.8 M sucrose, transferred to a centrifuge tube and covered with 0.3 M sucrose. After centrifugation at 100,00g for 1 h, the floating band, containing plasma membrane sheets, was recovered at the interface, diluted to 0.3 M sucrose, sedimented by centrifugation and analysed. The supernatants of the first, low-speed centrifugation were further centrifuged in sequence, first at 10,000g for 10 min to obtain the P2 pellet, then at 100,000g for 1 h to obtain the P3 pellet and the final supernatant, S. All fractions were analysed by western blotting. For digestion experiments in vitro, mixtures of the P2 and P3 fractions were first incubated for 10 min with either trypsin (10 µg ml−1, 22 °C) or S (4 °C), with or without LB containing 0.2% Triton X-100 (final concentration), and then analysed as above.

Immunofluorescence and immunogold Cells plated on poly-(L-lysine)-coated coverslips were fixed for 20 min on ice in 4% formaldehyde dissolved in PBS pH 7.4, quenched with 0.1 M glycine, washed in DB+ and transferred to 22 °C. After exposure for 2 h to the primary antibodies in DB+ they were washed extensively in PBS, exposed for 2 h to fluorescein isothiocyanate-conjugated or TRITC-conjugated goat anti-mouse or anti-rabbit secondary antibodies, incubated for 1 h in DB+, washed, mounted, and finally analysed in a laser scanning confocal microscope (Bio-Rad MRC 1024). For plasma membrane permeabilization, cells plated on coverslips were incubated on ice with 0.5 U ml−1 streptolysin-O in PBS, then warmed (5 min at 37 °C) to induce pore formation. After extensive washing with PBS, the monolayers were either incubated with the primary antibody diluted in DB− (30 min on ice) and then fixed and quenched, or directly fixed, quenched and then exposed to the primary antibody in DB+. After being washed and exposed to the second antibody in the same buffers, the monolayers were processed as above. For surface immunolabelling, monolayers of living cells, covered with PBS+ and kept at 22 °C, were treated with ionomycin (5 µM, 5 min) alone or after loading with BAPTA (BAPTA-AM, 50 µM, 15 min). Additional monolayers were processed while covered with calcium-free PBS supplemented with 1 mM EGTA. After transfer to 4 °C they were incubated (30 min on ice) with the primary antibodies diluted in DB−, then washed, fixed in formaldehyde for 20 min, quenched with glycine, exposed to the second antibody in DB+ and processed as above. In parallel experiments, suspensions of PC1227 cells, at rest or treated with ionomycin, as indicated previously, were quickly chilled and incubated at 4 °C with Ab diluted in DB−. After washing, the fluorescent surface staining of the population was evaluated by FACS analysis. For combined surface labelling with Ab and the TRITC-labelled Lycopersicon esculentum lectin, the latter was included in the primary antibody solution. After incubation in ice for 30 min and washing,

the monolayers were fixed and processed as described in the preceding paragraph. For caged-calcium experiments, cells plated on poly-(L-lysine)-coated coverslips and loaded for 45 min with 10 µg ml−1 NP-EGTA-AM compound (Molecular Probes) were flashed with an ultraviolet xenon lamp to induce exocytosis42. Experiments and imaging were carried out in the laborarory of M. Rupnik at the European Neuroscience Institute, Göttingen. Where indicated, the same experiments were performed after a 24-h preincubation of the cells with tetanus toxin (100 nM). The latter conditions were chosen on the basis of the results of preliminary experiments showing in PC12-27 cells a decrease (about −40%) in a protein target of the toxin, cellubrevin, as revealed by western blotting, and in wild-type PC12 cells a decreased appearance to the surface of synaptotagmin I, after 5 min stimulation with either ionomycin or 50 mM KCl, as revealed by immunofluorescence. The surface appearance of both dA and synaptotagmin I was investigated concomitantly by dual surface labelling in cells expressing both classical and new exocytotic systems, namely NGF-differentiated wild-type PC12 cells, treated or not for 24 h with tetanus toxin (100 nM) and stimulated with ionomycin as specified above. Quantification of the results was made after deconvolution analysis (see below). Optical sections taken every 150 nm with a wide-field microscope were deconvoluted with the softWoRx Deconvolve software (Applied Precision) on the Delta Vision System, to remove the blur in fluorescence and to perform three-dimensional reconstruction of the images. For electron microscopy, PC12-27 cells were fixed in 4% formaldehyde plus 0.2% glutaraldehyde and then processed in parallel for conventional Epon embedding, for LR White resin embedding, and for freezing. Ultrathin sections of Epon samples were stained; ultrathin sections in LR White and cryosections were exposed to Ab and then labelled with immunogold43. Controls were ultrathin sections of wild-type PC12 cells processed as above and of PC12-27 cells exposed to immunogold without previous treatment with Ab.

Immunoprecipitation and protein biotinylation Lysates from PC12-27 cells (1 ml in LB), precleared by the addition of 30 µl of protein G–Sepharose beads, were incubated overnight with Ab at 4 °C, with gentle mixing, after which 50 µl of protein G–Sepharose beads was added. After incubation for 2 h at 4 °C, the beads were washed three times in LB and twice in 10 mM Tris-HCl pH 7.4. Bound proteins were analysed by 2D PAGE staining and western blotting. In specific experiments, cell suspensions at rest or after treatment with ionomycin for 5 min were transferred to ice and labelled with cell-impermeant sulpho-NHS-SS-biotin. After cell lysis in LB, the preparations were immunoprecipitated with Ab as mentioned above, and the proteins bound to the beads were analysed by western blotting probed with either Ab, to evaluate the total amount of antigen immunoprecipitated, or horseradish peroxidase-conjugated avidin, to estimate the surface-biotinylated antigen.

Two-dimensional PAGE, protein expression in E. coli and peptide sequencing Two-dimensional PAGE analyses of Ab-immunoprecipitated proteins were performed by the IPGphor method (Amersham Biosciences). After separation in the second dimension, the proteins were either revealed with silver staining or transferred to nitrocellulose filters. The spots recognized by Ab were excised, digested in gel with pepsin, processed for mass-spectrometric analysis and sequenced as described in ref. 44. The BL21 E. coli, transformed with pGEX-4T-1 empty vector or with the vector containing the z80 construct coding for an AHNAK fragment fused to GST, were induced with isopropyl β-D-thiogalactoside as indicated elsewhere9. The GST fusion proteins produced were recovered from crude bacterial lysates by glutathione–Sepharose 4B resin (Amersham Biosciences) in accordance with the manufacturer’s instructions, and labelled in western blots with Ab. Various GST–z80 fragments recognized by Ab were excised from gels and analysed by mass spectrometry44.

Membrane lesions Micropipette puncturing and scrape wounding were performed as described previously17. The protocol for electroporation45 was as follows. PC12-27 cells were suspended in cytomix buffer supplemented with 2 mM ATP and 5 mM glutathione, in the presence of dextran-TRITC (1 mg ml−1). Suspensions were electroshocked at 400 V and 125 µF in a 2-cm cuvette placed in a Bio-Rad Gene Pulser apparatus equipped with a capacitance extender, then transferred to ice. After washing, surface exposure of the antigen was detected in resealed cells by incubation of coverslips first with Ab with or without anti-lamp1 antibody (30 min on ice), followed by washing, centrifugation onto slides, fixation and processing as mentioned above for surface immunolabelling. In parallel experiments the resealed cells were fixed and permeabilized before exposure to Ab and anti-lamp1, to reveal the intracellular distribution of the two antigens. RECEIVED 26 NOVEMBER 2001; REVISED 29 JULY 2002; ACCEPTED 24 OCTOBER 2002; PUBLISHED 25 NOVEMBER 2002.

1. Jahn, R. & Südhof, T. C. Membrane fusion and exocytosis. Annu. Rev. Biochem. 68, 863–911 (1999). 2. Kasai, H. Comparative biology of Ca2+-dependent exocytosis: implications of kinetic diversity for secretion function. Trends Neurosci. 22, 88–93 (1999). 3. Coorssen, J. R., Schmitt, H. & Almers, W. Ca2+ triggers massive exocytosis in Chinese hamster ovary cells. EMBO J. 15, 3787–3791 (1996). 4. Ninomiya, Y., Kishimoto, T., Miyashita, Y. & Kasai, H. Ca2+-dependent exocytotic pathways in Chinese hamster ovary fibroblasts revealed by a caged-Ca2+ compound. J. Biol. Chem. 271, 17751–17754 (1996). 5. Kasai, H. et al. Multiple and diverse exocytosis in wild-type and defective PC12 cells. Proc. Natl Acad. Sci. USA 96, 945–949 (1999). 6. Andrews, N. W. Regulated secretion of conventional lysosomes. Trends Cell. Biol. 10, 316–321 (2000). 7. Clementi, E., Racchetti, G., Zacchetti, D., Panzeri, M. C. & Meldolesi, J. Differential expression of markers and activities in a group of PC12 nerve cell clones. Eur. J. Neurosci. 4, 944–953 (1992). 8. Sussman, J., Stokoe, D., Ossina, N. & Shtivelmann, E. Protein kinase B phosphorylates AHNAK and regulates its subcellular localization. J. Cell Biol. 154, 1019–1030 (2001). 9. Sköldberg, F. et al. Identification of AHNAK as a novel autoantigen in systemic lupus erythematosus. Biochem. Biophys. Res. Commun. 291, 951–958 (2002). 10. Shtivelman, E. & Bishop, J. M. The human gene AHNAK encodes a large phosphoprotein located primarily in the nucleus. J. Cell Biol. 120, 625–630 (1993).

NATURE CELL BIOLOGY VOL 4 DECEMBER 2002 www.nature.com/naturecellbiology © 2002 Nature Publishing Group

961

articles 11. Piper, R. C. & Luzio, J. P. Late endosomes: sorting and partitioning in multivesicular bodies. Traffic 2, 612–621 (2001). 12. Thoidis, G. & Kandrov, K. V. A Glut4-vesicle marker protein, insulin-responsive aminopeptidase, is localized in a novel vesicular compartment in PC12 cells. Traffic 2, 577–587 (2001). 13. Kaether, C., Salm, T., Glombik, M., Almers, W. & Gerdes, H. H. Targeting of green fluorescent protein to neuroendocrine secretory granules: a new tool for real time studies of regulated protein secretion. Eur. J. Cell Biol. 74, 133–142 (1997). 14. Malosio, M. L. et al. Neurosecretory cells without neurosecretion: evidence of an independently regulated trait of the cell phenotype. J. Physiol. (Lond.) 520, 43–52 (1999). 15. Paolucci, P., Rovere, P., De Nadai, C., Manfredi, A. A. & Clementi, E. Nitric oxide inhibits the tumor necrosis factor alpha-regulated endocytosis of human dendritic cells in a cyclic GMP-dependent way. J. Biol. Chem. 275, 19638–19644 (2000). 16. McNeil, P. L. & Terasaki, M. Coping with the inevitable: how cells repair a torn surface membrane. Nature Cell Biol. 3, E124–E129 (2001). 17. Reddy, A., Caler, E. V. & Andrews, N. W. Plasma membrane repair is mediated by Ca2+-regulated exocytosis of lysosomes. Cell 106, 157–169 (2001). 18. Hashimoto, T. et al. Desmoyokin, a 280 kDa keratinocyte plasma membrane-associated protein, is homologous to the protein encoded by human gene AHNAK. J. Cell Sci. 105, 275–286 (1993). 19. Nie, Z., Ning, W., Amagai, M. & Hashimoto, T. C-terminus of desmoyokin-AHNAK protein is responsible for its translocation between the nucleus and cytoplasm. J. Invest. Dermatol. 114, 1044–1049 (2000). 20. Hashimoto, T., Gamou, S., Shimizu, N., Kitajima, Y. & Nishikawa, T. Regulation of translocation of the desmoyokin-AHNAK protein to the plasma membrane in keratinocytes by protein kinase C. Exp. Cell Res. 217, 258–266 (1995). 21. Haase, H. et al. Signalling from β-adrenoreceptor to L-type calcium channel: identification of a novel cardiac protein kinase A target possessing similarities to AHNAK. FASEB J. 13, 2161–2172 (1999). 22. Sekiya, F., Bae, Y. S., Jhon, D. Y., Hwang, S. C. & Rhee, S. G. AHNAK, a protein that binds and activates phospholipase C-γ1 in the presence of arachidonic acid. J. Biol. Chem. 274, 13900–13907 (1999). 23. Gentil, B. J. et al. The giant protein AHNAK is a specific target for the calcium- and zinc-binding S100B protein. J. Biol. Chem. 276, 23253–23261 (2001). 24. Lu, W. et al. Activation of synaptic NMDA receptors induces membrane insertion of new AMPA receptors and LTP in cultured hippocampal neurons. Neuron 29, 243–254 (2001). 25. Ward, D. T., Hammond, T. G. & Harris, H. W. Modulation of vasopressin-elicited water transport by trafficking of aquaporin2-containing vesicles. Annu. Rev. Physiol. 61, 683–697 (1999). 26. Whitehead, J. P. et al. The role of Ca2+ in insulin-stimulated glucose transport in 3T3-L1 cells. J. Biol. Chem. 276, 27816–27824 (2001). 27. Simpson, F., Whitehead, J. P. & James, D. E. GLUT4—at the cross roads between membrane trafficking and signal transduction. Traffic 2, 2–11 (2001). 28. Banerjee, A., Shih, T., Alexander, E. A. & Schwartz, J. H. SNARE proteins regulate H+-ATPase redistribution to the apical membrane in rat renal inner medullary collecting duct cells. J. Biol. Chem. 274, 26518–26522 (1999). 29. Blott, E. J. & Griffiths, G. M. Secretory lysosomes. Nature Rev. Mol. Cell Biol. 3, 122–131 (2002). 30. Almers, W. & Neher, E. Gradual and stepwise changes in the membrane capacitance of rat peritoneal mast cells. J. Physiol. (Lond.) 386, 205–217 (1987). 31. Thomas, P. et al. Two types of exocytosis observed in single, rat pancreatic acinar cells. J. Physiol. (Lond.) (in the press).

962

32. Xu, T., Binz, T., Niemann, H. & Neher, E. Multiple kinetic components of exocytosis distinguished by neurotoxin sensitivity. Nature Neurosci. 1, 192–200 (1998). 33. Coco, S. et al. Subcellular localization of tetanus neurotoxin-insensitive vesicle-associated membrane protein (VAMP)-VAMP7 in neuronal cells: evidence for a novel membrane compartment. J. Neurosci. 19, 9803–9812 (1999). 34. Hackam, D. J. et al. v-SNARE-dependent secretion is required for phagocytosis. Proc. Natl Acad. Sci. USA 95, 11691–11696 (1998). 35. Mellman, I. Quo vadis: polarized membrane recycling in motility and phagocytosis. J. Cell Biol. 49, 529–530 (2000). 36. Steinhardt, R. A., Bi, G. & Alderton, J. M. Cell membrane resealing by vesicular mechanism similar to neurotransmitter release. Science 263, 390–393 (1994). 37. Togo, T., Alderton, J. M., Bi, G. O. & Steinhardt, R. A. The mechanism of facilitated cell membrane resealing. J. Cell Sci. 112, 719–731 (1999). 38. Detrait, E. R. et al. Plasmalemmal repair of severed neurites of PC12 cells requires Ca2+ and synaptotagmin. J. Neurosci. Res. 62, 566–573 (2000). 39. Leoni, C. et al. Neurite extension occurs in the absence of regulated exocytosis in PC12 subclones. Mol. Biol. Cell 10, 2919–2931 (1999). 40. Schmidt, A., Hannah, M. J. & Huttner, W. B. Synaptic-like microvesicles of neuroendocrine cells originate from a novel compartment that is continuous with the plasma membrane and devoid of transferrin receptor. J. Cell Biol. 137, 445–458 (1997). 41. Galfré, G., Howe, S. C., Milstein, C., Butcher, G. W. & Howard, J. C. Antibodies to major histocompatibility antigens produced by hybrid cell lines. Nature 266, 550–552 (1977). 42. Rupnik, M. et al. Rapid regulated dense-core vesicle exocytosis requires the CAPS protein. Proc. Natl Acad. Sci. USA 97, 5627–5632 (2000). 43. Gatti, G., Podini, P. & Meldolesi, J. Overexpression of calsequestrin in L6 myoblasts: formation of endoplasmic reticulum subdomains and their evolution into discrete vacuoles where aggregates of the protein are specifically accumulated. Mol. Biol. Cell 8, 1789–1803 (1997). 44. Wilm, M. et al. Femtomole sequencing of proteins from polyacrylamide gels by nano-electrospray mass spectrometry. Nature 379, 466–469 (1996). 45. Van den Hoff, M. J. B., Moorman, A. F. M. & Lamers, W. H. Electroporation in ‘intracellular’ buffer increases cell survival. Nucleic Acids Res. 20, 2902 (1992).

ACKNOWLEDGEMENTS We thank H. Kasai and his laboratory staff for their pioneering collaboration and advice; E. Shtivelman and M. Rupnik for generous support; M. Matteoli and G.P. Schiavo for advice, A. Lorusso for participating in some experiments; C. Paolucci for FACS analysis, R. Barsacchi for performing the [Ca2+]i measurements; and D. Dunlap for the critical revision of the text. Imaging experiments were performed within Alembic (Advanced Light and Electron Microscopy Bio-Imaging Center), San Raffaele Scientific Institute. This work, supported by grants from Telethon, the European Union (QLG1–02233), CNR (Ag. 2000 and the CNS degenerative diseases), the Cofin Italian University System (2001.053389-033) and the Armenise–Harvard Foundation, was performed in the framework of the Italian MIUR Center of Excellence in Physiopathology of Cell Differentiation. Correspondence and requests for material should be addressed to J.M.

COMPETING FINANCIAL INTERESTS The authors declare that they have no competing financial interests.

NATURE CELL Group BIOLOGY VOL 4 DECEMBER 2002 www.nature.com/naturecellbiology © 2002 Nature Publishing