Enlargeosomes are cytoplasmic organelles discharged by regulated exocytosis, identified by immunofluorescence of their membrane marker, desmoyokin/Ahnak, but never revealed at the ultrastructural level. Among the numerous enlargeosome-positive cells, the richest and most extensively characterized are those of a PC12 clone, PC12-27, defective of classical neurosecretion. By using ultrastructural immunoperoxidase labeling of formaldehyde-fixed, Triton-X-100-permeabilized PC12-27 cells, we have now identified the enlargeosomes as small vesicles scattered in the proximity of, but never docked to, the plasma membrane. Upon stimulation, these vesicles undergo exocytosis [rapid after the Ca2+ ionophore, ionomycin, much slower after either the phorbol ester, phorbol myristate acetate (PMA), or ATP, working through a P2Y receptor], with appearance in the plasma membrane of typical desmoyokin/Ahnak (d/A)-positive, Ω-shaped and open profiles evolving into flat patches. Postexocytic removal of the exocytized d/A-positive membrane occurs by two processes: generation of endocytic vesicles, predominant after ionomycin and ATP 100–500 μM; and shedding of membrane-bound cytoplasmic bodies, predominant after PMA and 1 mM ATP, containing little or no trace of endoplasmic reticulum, Golgi, endo/lysosomes and also of a plasma membrane marker. Depending on the stimulation, therefore, the cell-surface expansion by enlargeosome exocytosis is not always recycled but can induce release of specific membranes, possibly important in the pericellular environment.
Regulated exocytosis, i.e. the fusion of specific cytoplasmic organelles with the plasma membrane occurring in response to adequate stimulation, is widely envisaged as a unique property of secretory cells, necessary for the release of hydrophilic secretion products segregated within vesicles and granules. Regulated exocytosis, however, occurs also in many other processes, often of great physiological importance, that induce changes of the plasma membrane. These processes induce the transient insertion of transporters and receptors and/or the prompt expansion of the cell surface. Although the number of identified non-secretory exocytic systems is already considerable, information about the generation and intracellular traffic of their vesicles is partial or completely absent. Also largely unknown is whether the vesicles involved in the various systems are distinct from each other, or whether groups of them share molecular and mechanistic properties [for a recent review, see Chieregatti and Meldolesi (1)].
The enlargeosome is one of the regulated, non-secretory exocytic organelles, initially discovered by capacitance/patch clamp electrophysiology in PC12-27, a defective clone of the neurosecretory cell line PC12 (2). Subsequent studies in a variety of cells, in which patch clamping was combined with biochemistry and immunofluorescence of the specific, high-molecular weight membrane marker, desmoyokin/Ahnak (d/A), demonstrated that enlargeosomes are distinct from the classical cytoplasmic organelles. Among the properties of enlargeosomes identified so far are their fast, Ca2+-dependent exocytosis, the detergent resistance of their membrane and their postexocytic endocytosis, carried out by neutral vesicles different from both clathrin-dependent endosomes and caveosomes. So far, however, the ultrastructural identification of enlargeosomes as unique cytoplasmic vesicles had not been achieved (3).
The information reported in this paper provides new direct evidence, expanding considerably our knowledge about enlargeosomes and their membrane traffic. By the use of ultrastructural immunoperoxidase [horseradish peroxidase (HRP)], enlargeosomes are identified with a population of small vesicles, scattered primarily in the proximity of the plasma membrane, undergoing exocytic discharge in response not only to Ca2+ rise but also to stimulation with a receptor agonist, ATP, and with phorbol myristate acetate (PMA). The postexocytic removal of the enlargeosome membrane from the cell surface is shown to occur by two processes: not only endocytosis but also shedding of membrane-bound bodies, predominant after PMA and high ATP. Finally, a change in lipid composition of the d/A-rich membranes following enlargeosome exocytosis was suggested by the change of their detergent resistance. So far, properties such as those now revealed for enlargeosomes have not been investigated for other non-secretory organelles competent for regulated exocytosis. The present results might therefore serve also as a reference for future studies on other exocytic systems.
Anti-d/A antibodies immunolabel a population of small cytoplasmic vesicles
In previous immunofluorescence studies of resting PC12-27 and other types of cells, the specific high-molecular weight marker of enlargeosomes, d/A, was shown in small intracellular puncta located primarily in the peripheral cytoplasm (Figure 1C). Upon [Ca2+]i rise induced either by the photolysis of caged Ca2+ compounds or by administration of a Ca2+ ionophore, the d/A immunolabeling was seen to redistribute rapidly to the cell surface as a consequence of exocytosis (3–6). The subsequent attempts to extend these studies to the ultrastructural level by conventional immunogold labeling of ultrathin sections were unsuccessful because neither the high-affinity anti-d/A monoclonal antibody used by Borgonovo et al. (3) nor additional monoclonals and polyclonals raised by us and others against fragments and peptides of d/A [(7,8) and unpublished data), ever yielded specific and reproducible immunolabeling signals.
The attempts were therefore pursued by testing additional approaches and technical refinements. The best results were obtained by a procedure in which cell monolayers, after fixation with either formaldehyde or a formaldehyde–glutaraldehyde mixture, were permeabilized with Triton-X-100 (TX-100) and then immunodecorated by HRP coupled to either the anti-d/A monoclonal antibody of Borgonovo et al. (3) or to another monoclonal antibody isolated recently in our laboratory. Immunodecorated cells were then postfixed in sequence with glutaraldehyde and OsO4 and finally embedded in Epon (TX-100 + Ab-HRP procedure, see Materials and Methods for details). Figure 1A,B,D,G illustrates the ultrastructure of resting PC12-27 cells processed according to the new procedure. Because of the detergent permeabilization carried out before postfixation, the bilayer structure of membranes, including the plasma membrane, as well as other protein-based structures were no longer evident. However, the general organization of the cell including the nucleus was maintained, and a few organelles, such as the mitochondria (Figures 1H and 5D) and the Golgi complex (Figure 1D), although disrupted, remained often recognizable. In contrast, the endoplasmic reticulum, coated vesicles, multivesicular bodies and lysosomes could no longer be recognized in simple thin sections. In terms of HRP immunolabeling, the cell surface exhibited only a few flat or concave HRP-positive patches (arrows in Figure 1A,D,G; compare with Figures 4 and 5), while the rest (over 95%) appeared completely negative. The only intracellular structures showing weak-to-moderate, but clear HRP positivity was a population of vesicles, spherical or slightly ovoid (Figure 1A,B,D–G), labeled along their contour by a ring of small (∼7 nm), moderately dense and strictly aligned beads (high magnification, B). When corrected in view of the thickness of the ultrathin sections (∼50 nm) as recommended by Parsons et al. (9), the average diameter of the clear core of 85 such vesicles, distributed in 45 cell profiles, turned out to be of 75.0 ± 9.1 nm. The average diameter of the whole vesicles, including the bead coat, was of 115.0 ± 11.1 nm.
Within resting PC12-27 cells, the HRP-labeled vesicles were spread mostly in the external layer of the cytoplasm, at 50–200 nm distance from the surface of the cell, surrounded by the actin cytoskeleton, with no obvious preference for basal, lateral or apical areas. Only a few were close, but never clearly docked to the plasma membrane (Figure 1A,D–G). A lower density was present in the deeper regions of the cytoplasm, except the Golgi/trans Golgi network that was negative. This distribution fits with that reported previously by immunofluorescence (see also Figure 1C). The average density (number of vesicles per micrometer of cell profile) observed in 81 cell profiles chosen at random (total analyzed cytoplasmic area: 121.2 μm2) was of 1.1. Based on a three-dimensional reconstruction of the cytoplasm of PC12 cells (unpublished data), we have calculated the average complement of labeled vesicles to be ∼9650/single cell.
The specificity of the vesicle staining by the anti-d/A-HRP antibody was confirmed by control experiments. Figure 1H shows that no HRP labeling was visible when the TX-100 permeabilization was made after, and not before, the exposure to the immuno-HRP, most likely because d/A, a luminal protein, was not reached by its antibody. Also, the cells processed by the TX-100 + Ab-HRP procedure remained negative, however, using a high-affinity monoclonal antibody (of the same class of the anti-d/A) specific for chromograninB, a protein not expressed by PC12-27 cells (10,11) (Figure 1I). Likewise, no HRP labeling was observed when the cells, permeabilized and processed with the anti-d/A-HRP, were not exposed to the 3,3′-diaminobenzidine tetrahydrochloride (DAB)–H2O2 reaction (data not shown).
The vesicles labeled by anti-d/A-HRP are enlargeosomes
In view of the high affinity and specificity of the anti-d/A antibodies employed, the first hypothesis about the nature of the d/A-HRP-positive vesicles was enlargeosomes. However, their peculiar ultrastructure, in particular the occurrence at their surface profile of the ring of small, strictly aligned HRP-positive beads, was strongly suggestive of a coat covering the vesicle membrane. The distinction of the HRP-positive vesicles from coated vesicles needed therefore to be demonstrated.
Figure 2 illustrates various approaches used to tackle the problem. Images of PC12-27 dually labeled for d/A and the large subunit of clathrin at the immunofluorescence and the ultrastructural HRP/nanogold level are shown in (A–D). In the immunofluorescence image, the two signals appear always distinct, not only in the Golgi area, where clathrin is abundant and d/A absent, but also in the subplasmalemma area, where the two types of puncta never coincide (Figure 2A). At the ultrastructural level, the structure of clathrin coats was not preserved in the TX-100-permeabilized cells, no matter whether immunolabeled or not. Therefore, the identification of coated pits and vesicles could be made only by their clathrin immuno-nanogold labeling. As shown in Figure 2B–D, such a labeling, enhanced (C,D) and not (B), was concentrated on small, poorly preserved vesicular structures of various size (50–100 nm in diameter), i.e. often smaller than the classical coated vesicles (arrows in Figure 2B–D). In contrast, the clathrin-nanogold staining of the over 100 d/A-HRP-positive vesicles investigated in dually labeled sections was at the background level. These results confirm that the two types of immunolabeled structures maintain their distinction even when located close to each other.
Additional ultrastructural experiments were carried out in PC12-27 cells with the TX-100 + Ab-HRP technique, using HRP coupled not to the anti-d/A but to antibodies specific for proteins exposed to the luminal surface of other intracellular organelles (Figure 2E–G). The organelles immunolabeled by HRP coupled to antibodies against transferrin receptor (endosomes, E) or lamp1 (late endosomes and/or lysosomes of various size and shape, F,G) appeared surrounded in part (arrowhead in F) or completely (E,G) by rows of small beads similar in appearance to those observed around the vesicles labeled by d/A-HRP.
Finally a group of experiments were carried out to establish the correlation between the appearance of d/A-immunolabelled vesicles with that of enlargeosomes in wild-type (wt) PC12 cells. Immunofluorescence (H) confirmed that in the growing wt PC12 cells, d/A is lacking (3); however, clathrin-positive puncta are abundant. When these cells were processed for electron microscopy by the TX-100 + Ab-HRP procedure, no HRP-positive vesicles could be seen (I; analyzed cytoplasmic area: 107.5 μm2). On the other hand, treatment of wt PC12 with nerve growth factor (NGF) induced the appearance not only of the neuronal-like phenotype but also of enlargeosomes (3), marked by the d/A immunofluorescence, which did not coincide with that of coated vesicles (J). In the same differentiated wt PC12 cells, electron microscopy showed that d/A-HRP-positive vesicles, identical in terms of size and structure to those of PC12-27, accumulate especially in the proximity of the plasma membrane (arrows in K). A few d/A-positive images of exocytized vesicles were occasionally seen at the plasma membrane of differentiated PC12 cells also at rest (arrowhead in K). Additional experiments carried out with cell types lacking or rich of enlargeosomes [freshly plated primary astrocytes on the one hand; HeLa and CHO cells on the other (3)] yielded the same results: lack or abundance of d/A-HRP-positive vesicles, respectively (data not shown).
We conclude that, with the TX-100 + Ab-HRP ultrastructural procedure, the profile of many, and possibly all positive vesicles, no matter what their nature, appears always surrounded by a ring of moderately dense beads. Therefore, information about the nature of the various HRP-positive vesicles cannot be deduced from their appearance but rather from the specificity of the antibodies used for their revelation, from their intracellular distribution and from the temporal coincidence of their appearance with that of the organelles in question. Based on these criteria, the vesicles labeled by the HRP-anti-d/A antibody can be identified as enlargeosomes.
Enlargeosome exocytosis is regulated by multiple signals
In our previous studies, stimulation of enlargeosome exocytosis was triggered only by a single intracellular signal, the rise of [Ca2+]i induced either by photolysis of caged Ca2+ compounds or by application of the Ca2+ ionophore, ionomycin (2–4,6). Whether other intracellular signals are also able to trigger the response had never been investigated. Here, in addition to ionomycin, we studied two other stimulants: ATP, an agonist of purinergic P2 receptors, and PMA, which may act on exocytosis both directly and as an activator of protein kinase C (PKC) (12). Both typical and atypical forms of PKC are known to be expressed by wt PC12 (13,14). However, in PC12-27 cells, these kinases have never been investigated. As far as purinergic receptors, considerable [Ca2+]i responses are induced in PC12-27 by ATP (15). Preliminary experiments (data not shown) with agonists (ATP, UTP and ADP) and blockers (suramin, 100 μM; reactive blue, 100 μM), tentatively identified the phosphatidylinositol 4,5-bis phosphate (PIP2) hydrolysis-coupled P2Y4 and/or P2Y6(16) as the purinergic receptors involved. The possible involvement of P2X receptors, in particular of P2X7(17), has not been investigated yet.
The enlargeosome exocytoses induced in PC12-27 cells by the three stimulants are illustrated in Figure 3. In most experiments, three concentrations of each agent were employed in parallel. These concentrations are indicated here and in the subsequent text and figures as the low, medium and high concentrations: ionomycin 0.5, 2 and 5 μM; ATP 100, 500 and 1000 μM; PMA 3, 30 and 300 nM. Figure 3A–D shows confocal images illustrating the d/A surface appearance induced by 5-min exposure to the medium concentrations of the various agents. Notice that, at this time point, the responses were approximately of the same size. FM1-43 fluorescence results confirmed that 5-min exposures to the three agents induced similar expansions of the plasma membrane (Figure 3E). However, the kinetics of the FM1-43 fluorescence rises were much faster with ionomycin than with ATP and PMA (Figure 3F). Ionomycin and ATP also induced [Ca2+]i rises which resembled in their kinetics the different FM1-43 responses triggered by the two agents, whereas PMA failed to induce any [Ca2+]i changes (Supplementary Information 1). Both the d/A surface appearance (revealed by confocal microscopy) and the FM1-43 responses induced by ionomycin and by the low and medium concentrations of ATP were largely prevented by preloading of the cells with the Ca2+ chelator, 1,2-bis (2-amino-phenoxy) ethane-N,N,N’, N’-tetraacetic acid (BAPTA) (30 min with the precursor, 1,2-bis (2-amino-phenoxy) ethane-N,N,N’,N’-tetraacetic acid acetoxymethyl ester (BAPTA-AM), 30 μM). With high ATP, part of the exocytic responses persisted, while with PMA they were unchanged by BAPTA preloading of the cells (data not shown).
PC12-27 cell monolayers, stimulated with the medium concentrations of ionomycin and ATP, were investigated also at the ultrastructural level by the TX-100 + Ab-HRP immunolabeling procedure. Overall, the results obtained with the two stimulants were similar. With ionomycin, however, the responses were rapid (not shown), and therefore the sequence of membrane traffic events was difficult to follow. We therefore concentrated our attention on the effects of ATP (Figures 4 and 5). After 1-min stimulation, many d/A-positive fluorescent puncta were already visible at the surface of stimulated PC12-27 cells, as shown by the deconvolved image of Figure 4A. At the same time point, some discrete d/A-HRP-positive vesicles [average diameter of the clear lumen in 20 vesicles, corrected as recommended by Parsons et al. (9): 75.0 ± 19.7 nm; arrowheads in Figure 4B,C,E,F] were still revealed by the electron microscope, accompanied, however, by a considerable number of d/A-positive profiles in direct continuity with the plasma membranes, very similar to those visible, although much more rarely, in resting cells: Ω-shaped exocytic profiles (thin arrows in Figure 4B,E) of approximately the same size of the vesicles (corrected average diameter of 18 clear images: 80.9 ± 12.7 nm); larger concave profiles (corrected average diameter of 12 images: 119.8 ± 23.6 nm; Figure 4C,D,G; thin arrow in F) and also flat patches continuous with the unlabeled cell surface, most likely because of full-collapse fusions (thick arrows in Figure 4C,D,G). After 3 (data not shown) and 5 min of ATP stimulation, the distribution of d/A, moved largely toward the cell surface, was still completely distinct from that of clathrin (Figure 5A). In the HRP-labeled cells, the density of discrete d/A-positive vesicles and Ω-shaped figures (arrowhead in Figure 5F) was decreased, whereas that of the open vesicles (arrows) and flat patches (*) was considerably increased (Figure 5B–F). In some cells, flat d/A-positive labeling of the surface extended for several micrometers (‘*’ in Figure 5B,C,E). Interestingly, the layer of small, strictly aligned beads, still clearly evident in the Ω-shaped and concave profiles (see Figures 4 and 5), was much less prominent in the flat surface profiles (inset in Figure 5B), no matter whether of small or large size.
Postexocytic events: Endocytosis
In previous studies, stimulation of enlargeosome exocytosis by both caged Ca2+ compounds and ionomycin was followed by the endocytosis of discrete, d/A-rich and lumenally neutral vesicles, negative for markers of the clathrin-dependent and caveolin-dependent endocytic systems (4). Whether the d/A-positive endocytosis occurs also after PC12-27 stimulation with ATP or PMA had never been investigated. Figure 6 shows immunofluorescence results obtained in PC12-27 cells stimulated for 15 min with the medium concentrations of ionomycin, ATP or PMA in the presence of the anti-d/A antibody, then washed, fixed, permeabilized and decorated for immunofluorescence. Most cells stimulated with ionomycin or ATP appeared decorated with numerous d/A-positive puncta (Figure 6A,B) that surface immunolabeling with the lectin from Bandeiraea simplicifolia demonstrated to be largely intracellular, and thus because of endocytosis [Figure 6E and data not shown; see also Cocucci et al. (4) and Lorusso et al. (6)]. In contrast, in the cells stimulated by PMA d/A, immunolabeling was marginal (Figure 6C). With ionomycin, the antibody uptake increased progressively with the concentration. With ATP, the uptake also increased, however only from the low to the medium concentration. At high ATP, the uptake was low, similar to that observed at all concentrations of PMA (Figure 6D).
Endocytosis induced by the medium concentrations of ATP was investigated also at the electron microscope level (Figure 7A–E). Two approaches were used. In the first, the cells were stimulated for 15 min in the presence of the anti-d/A antibody conjugated to HRP, then fixed in formaldehyde–glutaraldehyde and processed for the DAB reaction without TX-100 permeabilization. After further fixation, first in glutaraldehyde and then in OsO4, the cells were embedded in Epon. In addition to the flat areas and concave profiles of d/A-HRP positivity at the cell surface (arrows in Figure 7A,B; see also Figure 5), the cells exhibited discrete HRP-positive vesicles, often distributed at some distance from the cell surface, resembling enlargeosomes in their membrane organization but of larger size (corrected average diameter of the clear lumen in 12 images: 113.8 ±18.0 nm) (arrowheads in Figure 7A,B). When, in the labeling procedure, anti-d/A-HRP was replaced by anti-chromograninB–HRP, no positive endocytic vesicles were observed (data not shown).
In other experiments, the cells were stimulated for up to 5 min in the presence of FM1-43. After thorough washing and fixation in a glutaraldehyde–formaldehyde mixture, the cells were incubated in the DAB solution, then photoconverted, postfixed in OsO4 and finally embedded in Epon (18). Figure 7C–E vesicles, of variable size, on the average distinctly larger than the enlargeosomes (corrected average diameter of 20 clear lumena: 133.1 ± 24.2 nm), delimited by rings of dense particles analogous to those of the HRP-positive vesicles, distributed near the surface (Figure 7C,D) and scattered in the cytoplasm (Figure 7E). Photoconversion of FM1-43 does not distinguish among the various types of endocytic vesicles. Therefore, those shown in the Figure 7C–E could well be negative for d/A. However, in separate experiments, d/A alone was found to account for a considerable proportion (38%) of the surface-biotinylated protein internalized in 20 min (data not shown). Moreover, vesicles of the size and aspect of those of Figure 7 were very rare in resting cells photoconverted after exposure to FM1-43. The d/A-positive nature of at least many of the vesicles of Figure 7C–E appears therefore likely.
Postexocytic events: Shedding of membrane-bound bodies
The enlargement of the cell surface resulting from the enlargeosome exocytosis was compensated not only by endocytosis but also by shedding of d/A-rich, membrane-bound bodies. The latter, very rare in the pellets obtained by centrifugation of the media bathing resting cells, increased when the cells were stimulated. Specifically, it was still low at all concentrations of ionomycin and became considerable at the medium/high concentrations of PMA and also at the high concentration of ATP (Figure 8A). The results of d/A-rich body release looked like the quasi-mirror images of the d/A endocytosis results of Figure 6D. The choice between the two processes that follow enlargeosome exocytosis appears therefore to depend on the applied stimulation.
To investigate the ultrastructure of the d/A-rich membrane shedding process, PC12-27 cells were studied by the d/A-HRP procedure (no TX-100 permeabilization) while at rest or after 15-min stimulation with the high concentration of ATP. At rest, the few membrane-bound bodies lying over the cell monolayers, 200–500 nm in diameter, were almost all d/A negative (data not shown). After stimulation, d/A-positive surface bodies (diameter 150–250 nm) became much more numerous and prominent, with a significant fraction still in direct continuity with d/A-negative areas of the plasma membrane (arrows in Figure 8B), while others were apparently discrete (Figure 8B,C). The d/A labeling of many of these structures was flat. Others, however, exhibited HRP-positive dots, 10–30 nm in diameter, apparently bulging out of the membrane, spaced from each other of 10–50 nm or more (Figure 8B), which at the moment remain unexplained. The orientation of their d/A across the limiting membrane was the opposite of enlargeosomes, i.e. it was outward, as shown by their immunolabeling without permeabilization (Figure 8B,C).
A crude pellet heavily contaminated by cell fragments and organelles, recovered by centrifugation of the media bathing cells stimulated with ATP for 5 min (Figure 8D), included shed bodies (average diameter 150–250 nm) marked at their external surface by d/A, as shown by their immunolabeling without permeabilization (Figure 8D) and by the complete removal of their d/A by trypsin (Supplementary Information 2). To enrich the fraction of d/A-positive bodies, the media bathing-stimulated monolayers were centrifuged at low speed and the ensuing supernatants were filtered through a syringe top Millex GP 0.22 μm and then centrifuged at high speed. The pellets of the latter centrifugations, containing approximately one-third of the proteins present in the crude pellets together with almost all the d/A, were greatly enriched in bodies positive for the marker protein (Figure 8E,F). Compared with the crude pellets (Table 1), they were enriched also of two other proteins: a SNARE SNAP23 and annexin2, the latter segregated within the bodies (Supplementary Information 2). In contrast, a few proteins (another SNARE, syntaxin4 and annexin7) were low, i.e. significantly less abundant in the enriched than in the crude pellet, while all the other proteins investigated, including the endosome and lysosome markers, transferrin receptor and lamp1; the endoplasmic reticulum and Golgi markers calnexin, calreticulin and 58K; and also the plasma membrane Na+/K+ ATPase, were undetectable (Table 1).
Table 1. Recovery of proteins in shedding bodies released by ATP -stimulated PC12-27 cells
Asterisks indicate for each protein whether the recovery (on a protein basis, defined by the ratio between d/A in enriched and crude pellets) was high = twofold or more; low = not more than 0.5; undetectable, no specific band visible in the western blot of the enriched pellet.
Triton-X-100 resistance of d/A-rich membranes
In previous studies, we had found that, in contrast to the membranes of other organelles, the d/A-rich membranes of resting PC12-27 cells resist solubilization by cold TX-100, as shown by their large recovery in the floating band of a centrifugation gradient together with typical markers of cholesterol-rich membrane rafts, flotillin, caveolin and Thy1. This result suggested the enlargeosome membrane to be also rich of cholesterol and spingomyelin domains (4). Because, upon exocytosis, the d/A marker translocates first to the plasma membrane and then to the membranes of endosomes and shed vesicles, the study of TX-100 resistance was extended to the subcellular fractions where these d/A-rich membranes are recovered. After delicate homogenization of resting PC12-27 cells, enlargeosomes are mostly recovered in two fractions, microsomes and the plasma membrane sheet-enriched fraction. When these fractions were exposed to cold TX-100, d/A largely resisted solubilization (Figure 9A,B), yielding patterns analogous to that of the homogenate (4). When in contrast, the analysis was carried out with fractions isolated from PC12-27 cells prestimulated for 15 min with the medium concentration of ATP, floatation was shown only by the microsomal d/A (Figure 9A), while the d/A of the plasma membrane-enriched fraction was recovered in the solubilized area of the gradient (Figure 9B). Similar patterns were shown by both the endocytized vesicles, revealed by the anti-d/A antibody taken up by the cells during the 15-min stimulation with the medium concentration of ATP, and by the shedding bodies, enriched by filtration through a syringe top Millex GP 0.22 μm of the incubation medium bathing the cells similarly stimulated with ATP (Figure 9C).
Enlargeosomes were discovered initially by patch clamp capacitance assays as non-secretory cytoplasmic organelles discharged rapidly by exocytosis in response to [Ca2+]i rises (2). Immunofluorescence studies, started upon the identification of d/A as a high-molecular weight protein marker peripheral to the luminal surface of the organelle membrane, lead to the recognition of enlargeosomes in a variety of cell types, both of tissues and cell lines (3). Among these cells, the most advantageous for the study of the organelle were those of PC12-27, a clone of the PC12 pheochromocytoma line defective of neurosecretion. PC12-27 are very rich in enlargeosomes, and the regulated discharge of these organelles is rapid and extensive. Moreover, their lack of the classical neurosecretory organelles, synaptic-like and dense-core vesicles, makes the analyses of exocytic responses much simpler than those of wt neurosecretory cells. Finally, in most previous studies, the data on enlargeosomes obtained in the clone were checked in parallel in other types of cells, with analogous results (3,4,6). At this stage, therefore, PC12-27 appears as the best cellular model of enlargeosome investigation, appropriate for the identification of new properties of the organelle, to be searched later in other cell types.
Compared with other recently discovered non-secretory exocytic organelles (1), the information on the intracellular distribution and composition, the discharge and the endocytic recycling of enlargeosomes were already significant. However, all attempts to investigate the ultrastructure of these organelles had been unsuccessful, both because the antigen recognition of the anti-d/A antibodies was affected by glutaraldehyde fixation of the cells and because the signals obtained by classical procedures (immunogold labeling of ultrathin sections obtained from samples frozen or embedded in hydrophylic resins) were too low to be specifically appreciated. The distinction of the enlargeosomes from other vesicles populating the cytoplasm was therefore impossible.
In the present work, we have employed a new procedure in which PC12-27 cells were fixed in formaldehyde and permeabilized with TX-100 before HRP immunolabeling. The latter is a technique of the highest sensitivity. Our procedure, therefore, sacrifices the preservation of the structures, in particular of many types of membranes that are largely solubilized, in favor of the revelation of specific antigens. In PC12-27 cells, a population of vesicles was found to be surrounded by ring of small, strictly aligned beads weakly, but clearly labeled by d/A-HRP. These vesicles are characterized by a number of properties typical of enlargeosomes, established in previous studies by immunofluorescence and other techniques and expanded in the present work. Among these properties are the expression in various cell types known to be enlargeosome positive (in addition to PC12-27, differentiated wt PC12, CHO, HeLa), and lack in enlargeosome-negative cells (growing wt PC12; freshly plated astrocytes); the distribution in the peripheral layer of the cytoplasm; the rapid exocytosis in response to the Ca2+ ionophore, ionomycin, with appearance of typical images (Ω-shaped, curved and flat profiles) in continuity with the plasma membrane. Two additional properties of the vesicles, their TX-100 resistance and their size, deserve a specific mention. The first fits well with the resistance of enlargeosomes revealed by previous biochemical and biophysical results (4). As far as the size, our data confirm that the vesicles are small, as previously anticipated for enlargeosomes based on capacitance patch clamp results (2). However, because of the permeabilization of the membrane, we do not know whether the HRP-induced ring of beads forms at the external surface or within the vesicle. In the first case, the average diameter would be ∼75 nm, in the second ∼115 nm. In each resting PC12-27 cell, the estimated number of d/A-HRP-positive vesicles was on the average of ∼9650. In case of vesicles 75 nm in diameter, this would correspond to a cumulative membrane area of 169.6 μm2; in case of vesicles of 115 nm, to a cumulative membrane area of 400.7 μm2, i.e. the 54.2% and the 128% of the maximal surface expansion measured in PC12-27 cells by capacitance patch clamping after very strong stimulation [19.1% of 1640 μm2 resting surface area, i.e. ∼313 μm2(2)].
On the other hand, the HRP-decorated ring of beads delimiting the vesicles was suggestive of their coverage by a coat. The question was therefore raised as to whether these structures were not enlargeosomes but classical coated vesicles. Specific confocal and electron microscope results have, however, excluded this possibility. In fact, the immunofluorescence puncta and HRP vesicles labeled by the d/A antibody were absent in cells rich in coated vesicles, but lacking enlargeosomes, while in cells expressing both organelles they did not coincide with puncta and vesicles labeled by anti-clathrin antibodies. Moreover, the d/A-positive vesicles underwent exocytosis upon cell stimulation, a property that coated vesicles do not share. When HRP was coupled to other antibodies, against transferrin receptors (endosomes) and lamp1 (late endosomes and lysosomes), the positive structures, although distinguishable for their variable size and shape, appeared delimited by labeled beads, similar in this respect to the vesicles investigated with the anti-d/A antibody. Finally, a search of the literature revealed numerous examples of ‘coat labeling’ of membranes weakly positive for HRP, in classical papers of the sixties and seventies [see, among others, Figures 5 and 6 of Graham and Karnovsky (19) and Figure 11 of Ceccarelli et al. (20)]. The ‘HRP coat’, therefore, appears to be an artifact of the DAB precipitation, the process that accounts for the HRP labeling, independent on the nature of the membrane involved. This conclusion is reinforced by the presence of bead rings also around endosomes labeled by FM1-43 and DAB, that had been exposed neither to antibodies nor to HRP. Based on the high specificity of the anti-d/A antibodies used (3) and on the properties of the d/A-HRP-positive vesicle expression and distribution listed above, the latter can be safely identified as enlargesomes.
Enlargeosome exocytosis is regulated by multiple mechanisms
In our previous studies, only treatments inducing prompt and considerable rises of [Ca2+]i, i.e. photolysis of specific caged compounds and application of ionomycin, were used to trigger enlargeosome exocytoses (2–4,6). The present data demonstrate that processes comparable in size but much slower in rate than those occurring in response to ionomycin are induced in PC12-27 cells by ATP activation of P2Y4 and/or P2Y6 receptors. The latter are coupled to PIP2 hydrolysis and thus to [Ca2+]i rise and diacylglycerol generation (16,21). Because ATP is released by a variety of cells, it may play an important role in the regulation of enlargeosome exocytosis under physiological conditions.
Of the two second messengers released following activation of P2Y4/P2Y6 receptors, diacylglycerol is known to activate a chain of events synergistic with, but independent from, Ca2+. The results obtained with the diacylglycerol analog, PMA, confirm that a Ca2+-independent stimulation of enlargeosome exocytosis does indeed exist. The mechanisms by which PMA stimulates other types of exocytosis are at least two: activation of PKC, with phosphorylation of SNAREs (SNAP23 and SNAP25) and regulatory proteins (Munc18) (22–24), and direct binding to other regulatory proteins (Munc13 and RasGRP1) (25–27). So far, the mechanistic aspects of the enlargeosome exocytosis are unknown. The present results are, however, the first indication that molecular mechanisms analogous to those operative in other exocytic systems may also have a role in the regulation of the enlargeosome exocytosis.
Enlargeosome membrane recycling: Endocytosis
Extensive studies of other exocytic systems have documented the multiplicity and complexity of the postexocytic membrane interactions and recycling (for reviews, see Maxfield and McGraw (28) and Harata et al. (29)). In the case of enlargeosomes, the only process previously shown to follow ionomycin stimulation was the endocytosis of d/A-rich vesicles revealed, at the fluorescence and ultrastructural HRP level, by the specific antibody taken up from the incubation medium (4). Here we have confirmed the endocytic process in PC12-27 cells stimulated by the medium concentration of ATP and documented its occurrence also by a technique, photoconversion of FM1-43, which cannot be affected by any antibody binding to the recycling vesicle. In addition, we have shown that, upon exocytosis, the d/A-rich membrane modifies at least one of its properties, TX-100 resistance.
The cold TX-100 resistance, documented previously by the floatation of the enlargeosome membranes of resting PC12-27 cell homogenates in a density gradient (4), was confirmed in this study by the analysis of microsome and plasma membrane-enriched fractions containing enlargeosomes distributed at some distance from and in close proximity to the cell surface, respectively. However this property, which reveals the enlargeosome membrane abundance of cholesterol and sphingomyelin, was shared neither by the d/A exocytized to the plasma membrane nor by that redistributed to the endosomes (and also to shed bodies, see below) analyzed 15 min after ATP application. These results suggest that, upon exocytic fusion, the cholesterol-rich enlargeosome membrane intermixes with the adjacent plasma membrane, and that the resulting TX-100 solubility is maintained during the ensuing traffic processes. In resting and ionomycin-stimulated cells, d/A is a slow turnover protein (unpublished data), presumably recycled several times upon exocytosis. The postexocytic association of d/A with an endosomal TX-100 soluble lipid environment may therefore be transient. The d/A turnover in cells stimulated with PMA has not been investigated yet.
The most unexpected result of this study was the duality of the postexocytic, d/A-rich membrane traffic process: not only endocytosis, predominant after ionomycin and low/medium ATP, but also shedding of cytoplasmic bodies (referred to elsewhere with various names: shedding vesicles, microvesicles, ectosomes, argosomes, microparticles: for reviews, see Hugel et al. (30) and Ratajczak et al. (31)), predominant after PMA and high ATP. For quite sometime, shedding bodies were studied primarily in tumor cells and believed to play a role in the digestion of the extracellular matrix, and thus in cell motility and generation of metastases (32–35). Recently, however, the interest about these structures has greatly increased also in non-tumoral cell systems (17,36–38). Interestingly, stimulation of shedding by activators of PKC and by ATP, as observed here, also occurs in other cell types (17,35–38).
Membrane specificity, considered a critical property of shedding bodies, is however still incompletely known in many cell systems. In PC12-27 cells, many surface protrusions, newly formed after stimulation with high ATP, appeared to accumulate d/A, revealed by immuno-HRP, at the expenses of the rest of the plasmalemma that remained free of the protein. Moreover, the bodies strongly positive for d/A were almost absent in the medium bathing resting cells and appeared only after ATP stimulation. When the crude pellets from stimulated cell media were partially purified by filtration, only two proteins out of nine investigated by western blotting were enriched together with d/A, i.e. annexin2, which within intact PC12-27 cells is attached to the cytosolic surface of enlargeosomes (6), and the SNARE SNAP23, which may have a role in the enlargeosome membrane fusion. Other proteins, located in various organelles, endosomes, lysosomes, ER, Golgi complex and also in the plasma membrane (Na+/K+ ATPase), were in contrast low or absent. Taken together with previous patch clamp data (4), these results suggest the d/A-rich shedding bodies to be distinct from exosomes. Generation of d/A-positive bodies may therefore be a specific process. In view of their specificity, the d/A-rich bodies might well play one or more peculiar functions, which however are still unknown.
The present results have expanded significantly our knowledge of enlargeosome traffic and of d/A-rich membranes. The organelles are now identified with small cytoplasmic vesicles competent for extensive exocytosis regulated by various mechanisms, Ca2+ dependent and independent, triggered not only by the non-physiological stimuli investigated previously but also by the activation of one, and possibly other receptors and transduction stimuli. The postexocytic drop of TX-100 resistance, from the enlargeosomes to the d/A-rich membrane domains of the cell surface, endosomes and shed bodies, might play a role in the regulation of membrane traffic and membrane shedding. Finally, the duality of the regulated processes by which d/A-rich membrane patches are removed from the cell surface following exocytosis might serve not only to adjust the cell-surface area but also to carry out additional functions, including the release or recycling of specific proteins, that remain to be investigated.
Materials and methods
Cloned PC12-27 and wt PC12 cells, as well as the IgG2a-purified monoclonals, anti-d/A and anti-chromograninB antibodies, were from our laboratory (3). A second anti-d/A antibody (an IgG1) was isolated recently. The following antibodies and reagents were from commercial sources: mouse monoclonal IgG1 anti-transferrin receptor from Zymed (San Francisco, CA, USA); IgG1 anti-G58K from Abcam (Cambridge, UK); IgG1 anti-annexin2 from BD Biosciences Pharmingen (Heidelberg, Germany); IgG1 anti-lamp1 from Calbiochem (Schwalbach, Germany); IgG1 anti-Na+/K+ ATPase from Upstate Cell Signaling (Lake Placid, NY, USA); rabbit polyclonal anti-calreticulin from Affinity Bioreagents (Golden, CA, USA); anti-calnexin from Stressgen Biotechnology (Victoria BC, Canada); anti-annexin7 from Santa Cruz Biotechnology (Santa Cruz, CA, USA); anti-syntaxin4 and anti-SNAP23 from Synaptic Systems (Goettingen, Germany); a mouse monoclonal (IgG1) and a rabbit polyclonal anti-clathrin heavy chain from Affinity Bioreagents and Abcam, respectively; fluorescein isothiocyanate (FITC)-conjugated and tetraethylrhodamine isothiocyanate (TRITC)-conjugated goat anti-mouse antibodies, goat anti-rabbit antibodies and goat anti-mouse IgG subclasses from Southern Biotechnology Ass. (Birmingham, AL, USA); HRP-conjugated goat anti-mouse and anti-rabbit antibodies from Bio-Rad Labs (Hercules, CA, USA). Nanogold and the HQ-Silver Enhancement kit from Nanoprobes (Yaphank, NY, USA). FM1-43 styryl dye was from Molecular Probes (Eugene, OR, USA); ionomycin from Calbiochem (Schwalbach, Germany); the Activated Peroxidase Antibody Labeling kit and the Bicinchoninic acid (BCA) Protein Assay kit from Pierce Biotechnology (Rockford, IL, USA). All other chemicals were from Sigma (St Louis, MO, USA).
The wt PC12 and PC12-27 clones were grown at 37°C in DMEM supplemented with 10% horse serum (Euroclone, Wetherby, UK), 5% fetal clone III serum (Hyclone, Logan, UT, USA), 2 mM l-glutamine, 100 U/mL−1 penicillin and streptomycin (Biowhittaker, Verviers, Belgium) in a humidified 5% CO2 atmosphere. To induce cell differentiation, wt PC12 were treated with NGF, 100 ng/mL, for up to 5 days. The other cell types were cultured as specified by Borgonovo et al. (3).
SDS–PAGE, Western blotting and biochemical assays
Protein concentration of the samples was determined by the BCA assay. Fixed amounts of protein were separated by SDS–PAGE. The gels were stained with Coomassie blue. Other gels were transferred to nitrocellulose filters that were processed at 22°C, first by blockade for 1 h with 5% non-fat dry milk in Tris-buffered saline (TBS), then by incubation for 3 h with the primary antibody diluted in PBS plus 3% BSA, followed by five 10-min washes in TBS and incubation for 1 h with the appropriate HRP-conjugated secondary antibody (1 μg/mL−1). After further washes in TBS and PBS, the filters were developed photographically by chemiluminescence, using the enhanced chemiluminescence Western Blotting Detection reagent (Amersham Biosciences, Little Chalfont, UK). Signals were acquired by the Personal Densitometer SI and Image Quant (Amersham Biosciences).
[Ca2+]i was assayed by fura-2 as described by Borgonovo et al. (3).
Immunofluorescence of d/A was revealed either at the whole-cell level or at the cell surface only, as detailed in Cocucci et al. (4). In short, monolayers of wt PC12 and PC12-27 cells on poly-(l-lysine)-coated coverslips, to be used for all cell labeling, were kept at rest or exposed to the stimulatory agents while bathed at 37°C in Krebs Ringer Hepes (KRH, containing, in mM: 125 NaCl, 5 KCl, 2.4 MgSO4, 1,2KH2PO4, 2 CaCl2, 25 HEPES, 6 glucose, pH 7.4). At the end of the incubation, they were transferred on ice and fixed for 10 min by 4% formaldehyde dissolved in PBS, pH 7.4, then quenched with 0.1 M glycine and washed in PBS containing 0.5% BSA and goat 5% serum, supplemented with 0.3% TX-100, to induce membrane permeabilization. The latter solution was used also to dissolve and apply the primary and secondary antibodies, and for further washes. For surface immunofluorescence, the same procedure was used, except that no TX-100 was added to the employed solutions. Immunolabeled cells were studied using Bio-Rad MRC 1024 and Perkin Elmer Ultraview ERS confocal microscopes. Image deconvolution was carried out in a wide-field microscope of the Delta Vision system as described by Cocucci et al. (4).
Two approaches were used. For confocal microscopy, PC12-27 and wt PC12 monolayers on coverslips were washed in KRH and then incubated at 37°C for 1–5°min in the same medium supplemented with low, medium and high concentrations of ionomycin (0.5, 2, 5 μM), ATP (0.1, 0.5, 1 mM) or PMA (3, 30, 300 nM). Control cells received equal volumes of solvents. The cells were subsequently processed for d/A surface immunofluorescence microscopy, as previously described. For FM styryl dye fluorescence, PC12-27 cells suspensions in KRH were transferred to fluorometer cuvettes and kept under continuous stirring. FM1-43 dye was added to 4 μM final concentration. After at least 5 min, necessary to reach the dye equilibration, stimulatory agents were added, and fluorescence intensity changes were recorded on-line (excitation and emission wavelengths: 479 and 598 nm) using a LS50B fluorometer (Perkin Elmer, Wellesley, MA, USA) (39).
For immunofluorescence, PC12-27 monolayers, grown on either coverslips or petri dishes, were washed and incubated for 15 min at 37°C in KRH in the presence of one of the stimulatory agents together with the anti-d/A antibodies, as described in the legend of Figure 6(4). After washing with cold PBS, cells were fixed, permeabilized and processed for immunofluorescence as described above; to reveal the cell surface, cells were labeled before permeabilization with the lectin from B. simplicifolia BS-I, FITC labeled.
Monolayers of wt PC12 and PC12-27 cells were fixed with glutaraldehyde followed by OsO4 and embedded in Epon by conventional techniques. Most monolayers to be processed for immuno-HRP electron microscopy were prepared, stimulated or not, fixed, permeabilized with 0.3% TX-100 and immunolabeled as for whole-cell immunofluorescence, however, using antibodies (anti-d/A, anti-transferrin receptor and anti-lamp1) that had been conjugated to HRP (by the Pierce labeling kit, see Materials and Methods). A few additional monolayers were fixed with 4% formaldehyde–2% glutaraldehyde and labeled with the new, IgG1 anti-d/A-HRP antibody. After antibody application, the monolayers were fixed with 2% glutaraldehyde. The DAB reaction was carried out according to Ochs and Press (40) except that H2O2 (0.03%) was included in the mixture and the reaction was arrested after 20 min by washing with Tris buffer. Reacted monolayers were washed with cacodylate buffer, postfixed in sequence with 2% glutaraldehyde and 1% OsO4, each for 10 min, dehydrated and embedded in Epon. For dual labeling with HRP-anti-d/A and nanogold coupled to the anti-clathrin heavy clain, the cells, fixed with formaldehyde and permeabilized as described above, were exposed simultaneously to HRP-anti-d/A and nanogold-coupled anti-clathrin heavy-chain antibodies, then fixed in glutaraldehyde and exposed to the DAB reaction as described. In some experiments, the nanogold was enhanced with the HQ-Silver kit of the Nanoprobes. After OsO4 postfixation, the samples were dehydrated and embedded in Epon.
To investigate endocytosis, two techniques were used. In the first, HRP-conjugated anti-d/A antibodies were administered for different times (from 1 to 15 min) to living cell monolayers together with the stimulatory agents. After cold washing, the cells were fixed with 2% glutaraldehyde–4% formaldehyde. The DAB reaction and the monolayer embedding were carried out as described before. For further details, see Cocucci et al. (4). In the second, monolayers were stimulated for up to 5 min in KRH containing FM1-43, 4 μM, then thoroughly washed to remove the dye from the surface membrane, fixed with 2% glutaraldehyde–4% formaldehyde and quenched in 0.1 M glycin. After 10 min incubation in the DAB solution, the monolayers were washed and photoconverted for 15–25 min under an inverted microscope Axiovert 135 TV Zeiss at 480 nm, working with an Uvico Till Photonics Lamp light source. OsO4-fixed monolayers were finally embedded in Epon (18). All ultrastructural samples were studied in a Leo 912 electron microscope.
Suspensions and density gradient of PC12-27 in a mixture of 0.32 M sucrose, 5 mM HEPES, pH 7.4, and a cocktail of protease inhibitors were gently homogenized in a cell cracker (3) and then centrifuged at 1000 × g for 5 min, to separate the low-speed pellet from the supernatant. The first was resuspended in 2 mL of 1.6 M sucrose, transferred to the bottom of a TLS55 tube and covered with two 1.5 mL cushions of 1.5 and 0.32 M sucrose. After centrifugation at 100 000 × g for 60 min, a floating band, enriched in plasma membrane sheets (Na+/K+ ATPase 11-fold higher than in microsomes), was collected at the interface between the 0.3 and 1.5 cushions and diluted with HEPES 5 mM, pH 7.4, to a final 0.32 M sucrose concentration. The low-speed supernatant, on the other hand, was centrifuged at 10 000 × g for 10 min, and the supernatant of this centrifugation was centrifuged again at 100 000 × g for 1 h, to obtain the microsomal fraction. After protein assay, the two fractions were used for Western blotting and to assay the resistance of their membranes to non-ionic detergents.
Isolation of shedding bodies
PC12-27 monolayers grown in petri dishes were washed with KRH supplemented with protease inhibitors and then exposed, for 15 min at 37°C, to the various concentrations of ionomycin, ATP or PMA, or to their solvents, diluted in the same medium. At the end of the incubations, the media were collected and centrifuged at 1000 × g, to discard large cell debris; the supernatants were centrifuged at 100 000 × g for 1 h to obtain the crude pellet. To isolate preparations enriched in d/A-positive shedding bodies, the low-speed supernatants obtained from cells incubated with medium or high ATP were filtered through a top syringe filter, 0.220 μM cut-off (Millex GP, Millipore, Carringwohill, Ireland), and the filtrate processed by high-speed centrifugation. Both the crude and the filtered pellets were either processed by electron microscopy immuno-HRP, as already described, or resuspended to be used for Western blotting (antigens specified in Table 1), or exposed to TX-100 to assess the resistance of their membrane to non-ionic detergents, as described here below.
Membrane resistance to non-ionic detergents
Microsome and plasma membrane-enriched fractions isolated from PC12-27 cells, resting and stimulated with ATP (medium concentration) in plain medium or in the presence of the anti-d/A antibodies, as well as the shedding bodies were carefully resuspended in cold 0.32 M sucrose containing 5 mM HEPES, pH 7.4, and a cocktail of protease inhibitors, supplemented with 0.5% TX-100. After 30 min in ice, the preparations were diluted to a final 1.23 M sucrose concentration, and 2 mL cushions were transferred to the bottom of a floatation gradient, covered by a 2-mL cushion of 1.10 M sucrose and by 0.5 mL of 0.15 M sucrose (4). After 18 h centrifugation at 100 000 × g in a SW 50.1 Beckman rotor, 10 fractions were collected from the gradients and quickly analyzed by western blotting.
We thank Ilaria Prada and Riccardo Bazzotti for their participation in some of the experiments, Evelina Chieregatti for helpful discussions and Francesca Floriani for editorial assistance. This work was supported in part by grants from the Telethon Fondazione (GGGP030234), the European Community (APOPIS-LSHM-CT-2003-503330) and the FIRB 2004 Program of the Italian Ministry of Research.