Recycling of a secretory granule membrane protein after stimulated secretion

Membrane traffic in mammalian cells occurs via the budding and fusion of transport vesicles. During this process there is a net transfer of membrane from one organelle to another. This membrane flow requires recycling pathways in order for the cell to maintain the size and composition of organelle membranes.

The secretory pathway is expressed constitutively in all nucleated mammalian cells (Burgess and Kelly, 1987). Secretory and membrane proteins enter the secretory path- way by cotranslational translocation at the endoplasmic reticulum and are then transported via the Golgi complex to the plasma membrane by vesicular carriers. In addition to the constitutive secretory pathway, several cell types also possess a regulated pathway (Burgess and Kelly, 1987). In this case secretory proteins are stored after exit from the Golgi complex in secretory granules, which do not fuse with the plasma membrane until the cell receives a specific secretory stimulus. During stimulated secretion the membrane of the secretory granule fuses with the plasma membrane, releasing its contents to the extracellular milieu, and the secretory granule membrane proteins become incorporated into the plasma membrane. They are rapidly internalized, and are thought to recycle to newly forming secretory granules, since their rate of turnover is less than that of the secretory proteins (Winkler, 1977). The secretory granule membrane proteins are cleared from the plasma membrane by clathrin-coated-pit-mediated endocytosis (Phillips, 1987).

Several studies have demonstrated the delivery of externally applied antibodies recognising secretory granule membrane proteins to organelles defined by morphological criteria to be secretory granules (Patzak and Winkler, 1986; Yamashita and Yasuda, 1992) that have been interpreted as demonstrating membrane recycling after stimulated secretion. However, morphological criteria alone, in the absence of additional cytochemical markers, may well be inadequate in the discrimination of secretory granules from lysosomes (Patzak et al., 1987) and so it is possible that these studies only showed endocytosis of secretory granule membrane proteins, rather than functional recycling. In order to demonstrate functional recycling it is necessary to show that the proteins travel to a functional secretory granule, i.e. one that can respond to a further secretory stimulus.

Here the recycling of a secretory granule membrane protein, dopamine-β-hydroxylase (DBH), has been monitored by the delivery of antibodies against DBH to an intracellular compartment that can respond to a further secretory stimulus.

Bovine adrenal glands were collected from the local abattoir and processed as soon as possible after slaughter. Tissue culture and other reagents were purchased from Gibco or Sigma unless stated otherwise.

Primary cultures of bovine adrenal chromaffin cells were prepared using a modification (Cheek et al., 1989) of the method of Greenberg and Zinder (1982) omitting the differential plating step, and cultured on glass coverslips in a 37°C humidified incubator with 6% CO₂. Plating medium consisted of DMEM + 20 mM Hepes, 2% foetal calf serum, 25 μg/ml fluorodeoxyuridine, 50 μg/ml ascorbic acid, 3 μg/ml cytosine arabinoside,100 units/ml penicillin, 100 μg/ml streptomycin, 50 μg/ml gentamycin and 0.25 μg/ml fungizone, and was changed 24-48 hours later to maintenance medium that lacked serum. Cells were used 3-10 days after plating.

Cells were stimulated to secrete at room temperature (20-25°C). Chromaffin cells grown on glass coverslips were washed with Locke’s buffer (154 mM NaCl, 5.6 mM KCl, 5.6 mM glucose, 1.2 mM MgCl₂, 2.2 mM CaCl₂, 5 mM Hepes, pH 7.0), allowed to equilibrate in Mg-Locke’s (154 mM NaCl, 5.6 mM KCl, 5.6 mM glucose, 3.4 mM MgCl₂, 5 mM Hepes, pH 7.0) for 10 minutes and then treated with Ba-Locke’s (154 mM NaCl, 5.6 mM KCl, 5.6 mM glucose, 5 mM BaCl₂, 1.2 mM MgCl₂, 5 mM Hepes, pH 7.0) for 10 minutes; control cells were treated in parallel with Locke’s buffer. When appropriate, anti-DBH serum was added to the stimulation medium at a 1:100 dilution. At least 99% of the cells remained viable throughout the experiment. Cells were either fixed immediately or returned to culture for various times after removal of Ba-Locke’s. Approximately 50% of the cells remaining on the coverslips after processing contained DBH and could respond to barium stimulation.

Cells were processed for immunofluorescence microscopy using a method adapted from Timm et al. (1983). Briefly, cells were fixed for 20 minutes in 3% paraformaldehyde in phosphate-buffered saline (PBS) plus calcium (0.1 mM) and magnesium (0.1 mM), washed with PBS, and aldehyde groups were quenched with 50 mM ammonium chloride in PBS for 10 minutes, and then, if internal staining was required, cells were permeabilized with 0.5% Triton X-100 in PBS for 1 minute. Cells were then washed with PBS and protein sites were blocked for 30-60 minutes with 5% goat serum (from the Scottish Antibody Production Unit, Carluke) in PBS. Primary antibodies, if appropriate, were added at 1:100 to 1:500 dilution in 5% goat serum in PBS for 60 minutes, washed with PBS, reblocked, and the second reagent, rhodamine-labelled goat anti-rabbit IgG 1:500 (Sigma) in 5% goat serum in PBS, was applied for 60 minutes; they were then washed and mounted in Mowiol, left to set overnight at room temperature and observed in a Leitz Ortholux II microscope with a ×63 oil immersion lens and photographed on Ilford HP5+ film. All negatives and prints of internalized anti-DBH were exposed, developed and printed under identical conditions to allow direct comparison of images.

The polyclonal antibody against DBH was provided by D. Apps and J. Phillips and is described by Hunter and Phillips (1989).

For FACS analysis freshly isolated cells were stimulated in suspension.

Cells were isolated as usual, but plated onto bacteriological plastic in order to avoid tight adherence. Cells were scraped from the dish and resuspended in Locke’s buffer. Treatments were performed in suspension using the same methods as described for the immunofluorescence experiments. Cells were stimulated for the first time 2-4 hours post-preparation, allowed to recycle overnight and then scraped and resuspended for the second stimulation. Samples of cells (approximately 10⁶) were removed after the first stimulation, after overnight incubation and after restimulation were washed with ice-cold PBS, incubated with anti-DBH for 10 minutes on ice if appropriate, fixed and processed as for immunofluorescence microscopy, omitting the Triton X-100 step. All cells were stained with a 1:100 dilution of phycoerythrin-conjugated goat anti-rabbit IgG (Sigma) in 5% goat serum in PBS. The samples were resuspended and examined in a FACS analyser (Becton and Dickinson FACSCAN).

In order to study membrane recycling after stimulated secretion primary cultures of bovine adrenal chromaffin cells were used. These cells contain stored catecholamines in their regulated secretory granules, the chromaffin granules, and can respond in culture to various secretory stimuli.

A major protein of the chromaffin granule membrane is dopamine-β-hydroxylase (DBH). DBH is a tetramer of 75 kDa subunits associated with the membrane and the lumen of the chromaffin granule. DBH is a suitable candidate for following the recycling pathway because it is a major component of the granule membrane and its rate of synthesis has been shown to be less than that of the granule content proteins (Winkler, 1977).

Cells were stimulated to secrete using 5 mM barium ions. Barium is one of the most efficient secretagogues, consistently promoting the release of approximately 30% of stored catecholamines from chromaffin cells (Heldman et al., 1989; Hunter and Phillips, 1989). Fig. 1 shows the results of indirect immunofluorescence microscopy using a polyclonal antibody that recognises DBH (Hunter and Phillips, 1989). After barium stimulation and fixation the cell surface was labelled with the polyclonal antibody (Fig. 1E), indicating that chromaffin granules had fused with the plasma membrane to expose DBH at the cell surface. Control cells (Fig. 1A) showed only low levels of surface immuno-reactivity, suggesting that very few chromaffin granules fused with the plasma membrane in the absence of a secretory stimulus, as expected for regulated secretory cells. In this panel one cell had been permeabilized during fixation and showed punctate intracellular labelling, serving as an internal control for the immunofluorescence protocol (Fig. 1A, also seen in Fig. 1G). Inside the cells, both stimulated and control, DBH showed a typical strong punctate staining characteristic of chromaffin granules (Fig. 1B, F). In the experiment shown in Fig. 2 cells were stimulated to secrete in the presence of anti- DBH antibodies and fixed, and then surface and internal label was assessed. The antibody was bound to intact, stimulated cells (Fig. 2E), whereas unstimulated cells bound very little antibody (Fig. 2A). This is in good agreement with the results shown in Fig. 1 and indicates that the polyclonal antibody could recognise native DBH when delivered to the plasma membrane in stimulated, unfixed cells. Since the cells were not fixed before addition of the antibody, the inadvertent permeabilization observed in Fig. 1 was not observed. Permeabilized stimulated or unstimulated cells showed very little additional staining immediately after stimulation (Fig. 2B, F), confirming that the cells were intact when the antibody was applied.

Figs 1 and 2 also show the results of reincubation of stimulated cells for 2 hours after the removal of secretagogue. In Fig. 1 the increase in surface DBH seen at t=0 was rapidly removed during reculture of stimulated cells in the absence of secretagogue (Fig. 1G), suggesting that the DBH delivered to the cell surface during exocytosis had been internalized. Fig. 2 demonstrates that the polyclonal anti- DBH that had been bound to intact cells during barium stimulation was rapidly cleared from the cell surface after removal of the secretory stimulus (Fig. 2G), in agreement with the results for unbound DBH in Fig. 1, and appeared in punctate intracellular structures (Fig. 2H). This suggested that the presence of anti-DBH bound to the DBH at the plasma membrane did not inhibit the internalization of DBH. Control cells, as expected, showed only low levels of immuno-reactivity on the surface or inside the cells (Fig. 2C, D). Identical results were observed after only 30 minutes of reculture (not shown).

Thus DBH alone, or when complexed with anti-DBH, was rapidly internalized after removal of the secretory stimulus. Indeed the increase in amount of punctate label in the permeabilized, stimulated cells before reincubation in Fig. 2 (compare Fig. 2F with E) suggested that endocytosis had already been initiated in the presence of stimulus.

Taken together these results indicated that the polyclonal anti-DBH should be a suitable probe to use for recycling studies.

Cross-linking of surface molecules by antibodies is thought to cause their internalization and transport to the lysosome (Mellman and Plutner, 1984) where they will be degraded. Fig. 3 shows that after overnight incubation the bound and internalized anti-DBH antibody was still easily detectable inside the barium-stimulated chromaffin cells in a punctate distribution (Fig. 3F) and, as expected, there was little surface reactivity (Fig. 3E). This suggests that the internalized antibodies had not been delivered to a degradative compartment such as the lysosome. Control cells showed only low levels of antibody after overnight incubation inside or outside the cells (Fig. 3A, B) in agreement with the results shown in Fig. 2.

These results suggest that the anti-DBH was not transferred to the lysosomes in these cells, but was transferred to an intracellular compartment where it was stable overnight. The anti-DBH thus did not appear to cause cross- linking of the DBH and transport to the lysosome, possibly because DBH is a tetramer, and polyvalency of the anti- body would cross-link subunits of DBH without necessarily cross-linking separate oligomers.

The functional demonstration of membrane recycling requires that internalized granule membrane proteins return to secretory granules that can respond to a further secretory stimulus. Fig. 3 also shows the effects of a second secretory stimulus on cells that had previously bound and internalized anti-DBH antibodies. Restimulation caused the appearance of polyclonal anti-DBH at the cell surface (Fig. 3G). This increase in surface antibody in response to stimulation (compare Fig. 3E with G) strongly suggested that at least some of the antibody-DBH complex internalized 24 hours earlier had been recycled to functional secretory granules. The level of labelling was consistent with the finding that only approximately 30% of the chromaffin granules would be expected to be able to respond to barium stimulation. Control cells, as before, showed little antibody either inside or on the cell surface (Fig. 3C, D).

Thus barium-stimulated cells that had bound and internalized polyclonal anti-DBH antibodies could be stimulated to re-expose antibody at the cell surface after further stimulation. This strongly suggests that the antibody bound to DBH was recycled to an internal compartment that was capable of responding to a further secretory stimulus, most probably the secretory granules.

The immunofluorescence analysis described so far strongly suggests that internalized DBH-anti-DBH complex recycles to a regulated secretory compartment, most probably secretory granules. In order to obtain a quantitative estimate of recycling efficiency, and to confirm the interpretation of the immunofluorescence images, FACS analysis was performed. FACS analysis requires single-cell suspensions. Freshly isolated chromaffin cells were maintained in suspension, or lightly adherent so that stimulation and restimulation could be performed in suspension. Cells were stimulated in the presence or absence of the polyclonal anti- DBH antibody and samples were fixed immediately after stimulation, after overnight incubation or after restimulation. The samples lacking anti-DBH were incubated on ice for 10 minutes in the presence of anti-DBH after the required treatment in order to assess the localization of total DBH.

The results are shown in Fig. 4. Fig. 4A shows that total DBH behaved as expected. It was exposed at the surface after stimulation and after restimulation. The restimulation was just as efficient as the original stimulation, suggesting that the cells were capable of responding to a second secretory stimulation with the same efficiency as to the first. In Fig. 4B antibody was applied during the first stimulation to living intact cells. As in the control condition, DBH was only exposed in stimulated cells. After restimulation approximately 30% of the anti-DBH originally bound to the cells was re-exposed at the plasma membrane. Since cells respond to barium stimulation by releasing approximately 30% of their stored catecholamine, this suggests that the recycling of anti-DBH was very efficient.

The data presented here represent one of the first demon-strations of membrane recycling to a functional regulated secretory compartment after stimulated secretion.

In this study primary cultures of bovine adrenal chromaffin cells were used. In the past it has been extremely difficult to demonstrate the recycling of secretory granule membrane proteins in cell lines that express a regulated secretory pathway (see e.g. Green and Kelly, 1992). This is probably a result of the down-regulation of the pathway due to de-differentiation and to the fact that in growing cells the rate of synthesis of granule components should be sufficient to maintain the small granule population that is present.

Here, a secretory granule membrane protein, DBH, exposed at the plasma membrane during stimulated secretion, was rapidly internalized in the presence or absence of anti-DBH antibodies. Indeed uptake was already under way within the 10 minute stimulation period and was complete within 30 minutes of reincubation. The internalized anti- DBH bound to DBH was stable for prolonged periods inside the chromaffin cells and was present in a compartment that could respond to further secretagogue stimulation. This strongly suggested that the majority of the internalized antibody was recycled to the secretory granules and was not targeted to the lysosomes.

Previous work has shown that certain internalized antibodies bound to plasma membrane or granule membrane proteins are transported to the lysosomes, but in this case they are rapidly degraded (Mellman and Plutner, 1984; Patzak et al., 1987). Since the internalized antibody remained stable overnight it is assumed that in this study little transfer to the lysosome had occurred. The precise itinerary of internalized antibody-protein complexes is likely to depend on several factors, including the valency of the antibody and the extent of cross-linking of the antigen.

Other work has shown that incubation of the cells at 4°C immediately after stimulation inhibited recycling and appeared to stimulate degradation of internalized DBH (S.M. Hurtley, unpublished data). In both cases, internalization was complete within 30 minutes at 37°C. A working hypothesis is that granule membrane proteins must be internalized rapidly after delivery to the plasma membrane and that if they remain at the plasma membrane long enough to diffuse away from one another they cannot recycle, but are transported to the lysosome and degraded. In order to compare these two conditions directly it will be important to examine the sorting between internalized secretory granule membrane proteins and other endocytic tracers.

Recycling as measured by FACS analysis appeared to be very efficient. In response to a 10 minute stimulation with barium, cells secrete approximately 30% of their stored catecholamines. In the FACS analysis approximately 30% of internalized anti-DBH could be stimulated to return to the plasma membrane following restimulation.

Assuming that the DBH-anti-DBH complex returned to a homogeneous population of granules, recycling would appear to be extremely efficient. However, the population of chromaffin granules in cells is known not to be equally responsive to a stimulus, and newly synthesized granules are particularly so (Phillips, 1987). The newly synthesized granules would be most likely to contain recycled DBH and therefore it is possible that this method will overestimate recycling efficiency.

These findings should now be extended to other secretory granule membrane proteins as suitable reagents become available. The mechanism, rate and route of recycling to a functional compartment that can respond to further stimulation is now under investigation. In order to estimate the efficiency of the recycling pathway it will be necessary to go on to use a biochemical assay. Hunter and Phillips (1989) biotinylated the surface of stimulated cells to look for recycling and produced preliminary evidence that recycling occurred in their conditions. These findings will now be extended to look for evidence of return to a functional compartment, and to monitor the efficiency and rate of recycling.

The results presented here show that it is possible to demonstrate recycling of a secretory granule membrane protein, DBH, to a functional regulated secretory compartment. This contrasts with simple measurements of endocytosis of granule membrane proteins from the plasma membrane, which do not constitute an assay for recycling. This approach can now be used to define the kinetics and route of recycling.

The polyclonal anti-DBH was kindly provided by David Apps and John Phillips. Andrew Sanderson provided excellent technical assistance with the FACS analysis. Many thanks to David Apps and James Pryde for critical comments on the manuscript. This work was supported by the Wellcome Trust.