Isolation and characteristics of bovine pituitary secretory granules

Biochimica et BiophysicaActa, 797 (1983) 209-216

209

Elsevier BBA21619

I S O L A T I O N AND C H A R A C T E R I S T I C S O F BOVINE P I T U I T A R Y S E C R E T O R Y G R A N U L E S C. MICHAEL MORIARTY a FRANK DOWD b and MICHELE FONTAINE a a Department of Physiology and Biophysics, University of Nebraska College of Medicine, 42nd Street and Dewey Avenue, Omaha, NE 68105 and b Department of Pharmacology, Creighton University School of Medicine, Omaha, NE 68178 (U.S.A.)

(Received February 23rd, 1983) (Revised manuscript received June 14th, 1983)

Key words: Pituitary hormone release; Growth hormone," Prolactin; Secretory granule isolation," Ca" +; (Bovine anterior pituitary)

Secretory granules containing primarily growth hormone and prolactin were isolated from bovine anterior pituitaries. Marker enzyme analysis and electron microscopy indicated that the secretory granule fraction did not contain measureable amounts of other intracellular organelles. Such isolated granules were resistant to a variety of chemical and physical challenges including variations in osmolarity, ionic strength, EGTA, sonication, boiling, etc. The only treatments that were found to routinely result in granules lysis were alkaline pH and 0.5% SDS. Nonspecific leakage of both groth hormone and prolactin was less than 9% of total hormone pool even after a 60-min incubation. The release of prolactin but not growth hormone could be increased by lowering the free calcium concentration. Conversely, 10 -s M ionophore A23187 caused a decrease in nonspecific hormone leakage. This raises the possibility that a nonexocytosis secretory pathway might be involved in pituitary hormone release. The initial secretory granule fraction was further purified using discontinuous sucrose gradient ultracentrifugation to yield a subfraction highly enriched in prolactin granules. These granules had the same stability characteristics as the original secretory granule fraction. The use of such granules should prove useful in our efforts to understand how calcium regulates cellular secretion.

Introduction That calcium plays a key role in the regulation of pituitary hormone secretion is no longer in doubt. The question that remains is not if, but how. Somehow calcium, whether originating from an intracellular or extracellular source, initiates a series of molecular events within the pituitary cell, the culmination of which is hormone release from secretory granules either by exocytosis or selective solubilization of granule contents followed by transport of the molecules across the limiting cell membrane. The nature of the release process in the Abbreviations: GH, growth hormone; TRF, thyrotropin releasing factor; EGTA, ethyleneglycol bis(fl-aminoethyl ether)N, N'-tetraacetic acid. 0304-4165/83/$03.00 © Elsevier Science Publishers B.V.

pituitary is not altogether clear, although the evidence tends to favor the fusion-fission concept of exocytosis. However, a selective release of molecules from the storage reservoir/in the secretory granules can not be ruled out at this time. It is also possible that both exocytosis and nonexocytosis pathways of secretion might be operative within the cell. The ultimate target for the release signal(s) must be the secretory granule. One such signal could be a local change in calcium concentration in the vicinity of the granule. In view of the local domain concept of intracellular calcium distribution [1], localized reductions in free calcium could occur concomitantly with elevations in free calcium elsewhere in the cytoplasm. A more likely signal however, is not calcium itself but rather another

210 cytosolic constituent whose amount, configuration a n d / o r distribution is altered by calcium in such a way as to initiate release from the secretory granule. It seemed to us that a logical approach to the fundamental process of calcium-dependent secretion might start with a consideration of the isolated pituitary secretory granules, their nature and properties and extend to a consideration of both granule-granule and more importantly, granuleplasma membrane interactions. The purpose of this study was therefore to validate the procedure for granule isolation and purification, to characterize the stability of the secretory granules when incubated in vitro and finally to examine the release of both growth hormone (GH) and prolactin from the isolated granules. Materials and Methods

Fresh bovine pituitaries were obtained from animals of random sex at a local slaughterhouse. The procedure involved cutting a window in the top of the skull, removing the brain and dissecting free the entire pituitary. Usually 3-6 pituitaries (6-15 g) were obtained. All pituitaries were transported at room temperature in an empty covered beaker. Transporting pituitaries cold (either dry or wet) was found to increase the fragility of the isolated secretory granules. The entire procedure, from removal of the first pituitary until arrival in the laboratory, averaged 30 min. Upon arrival in the laboratory, the connective tissue sheath around the pituitary was teased away, the neural lobe removed and the resulting anterior lobes pooled and weighed. Following mincing, a 7% (w/v) solution with 0.32 M sucrose was prepared and the mixture homogenized with 20 strokes of a loose-fitting teflon pestle. Following centrifugation at 1500 × gay for 5 min in an IEC 870 rotor at 22°C, the supernatant was decanted and centrifuged at 10000 × gay for 12 min. The resulting tan and white layers were very carefully separated, the white layer resuspended in 0.32 M sucrose (5 ml/pituitary) and centrifuged as before. Any residual tan material was again carefully removed and the remaining white layer resuspended and centrifuged one more time. The pellet obtained was termed the crude secretory granule fraction.

When further purification was desired a modification of the procedure of Zanini et al. [2] was used. Briefly, this involves resuspending the secretory granule fraction in 1 M KC1/5 mM MgC12/0.5 mM puromycin, pH 7 at a protein concentration of 6 m g / m l and incubating the mixture for 15 min at 4°C followed by 10 min at 37°C to lyse a n d / o r aggregate G H granules and detach polyribosomes. Aliquots of this suspension were layered on a discontinuous sucrose gradient as described in Results and centrifuged for 15 min at 132 000 × gmax in a SW 27 rotor. The bands were collected with a siliconized Pasteur pipette. Turbidity measurements were carried out with a temperature-controlled quartz cuvette and a recording spectrophotometer. All measurements were corrected for the corresponding media in the absence of secretory granules. All measurements of G H and prolactin whether on supernatants or solubilized pellets were made via radioimmunoassays using ovine G H and prolactin material supplied by the N I A M D D Pituitary Hormone Program. Such ovine material is supplied with documentation verifying that it cross-reacts equally with corresponding bovine hormone. The marker enzymes succinic dehydrogenase, glucose-6-phosphatase and ouabain-inhibited (Na+-K+)-ATPase were measured as previously described [3]. Protein was measured using the method of Lowry et al. [4] with bovine serum albumin as the standard. Free calcium concentrations were measured with a calcium-sensitive electrode [5,6]. Samples prepared for electron microscopy were fixed for 1 h in 2% glutaraldehyde followed by 1 h postfixation in 1% OsO 4. Samples were dehydrated by transfer through a series of graded alcohols and embedded in Spurr low viscosity medium. Samples for discontinuous SDS-polyacrylamide gel electrophoresis were run on 1.5 mm thick slab gels at room temperature with a 37o acrylamide stacking gel and a separating gel of a 5-15% acrylamide gradient. A voltage of 50 V per 2 slabs was used until the samples had entered the separating gel when it was raised to 100 V. Gels were fixed and stained overnight in a solution consisting of 250 ml isopropyl alcohol, 100 ml glacial acetic acid, 650 ml, H20 and 0.05% Coomassie

211

Blue R-250. Destaining was done in a solution of 100 ml isopropyl alcohol, 100 ml glacial acetic acid and 800 ml H20. All other solutions and procedures were identical to those described by Laemmli [7].

TABLE I CONCENTRATION OF GH AND PROLACTIN IN THE HOMOGENATE A N D SECRETORY G R A N U L E FRACTION GH Prolactin Prolactin/GH

Results

To characterize the secretory granule fraction and to assess the extent of contamination with other cytoplasmic constituents, a series of experiments were carried out. First, the crude secretory granule fraction was assayed for the presence of enzymes considered to be markers for other organelles. We were unable to detect the presence of succinic dehydrogenase, glucose-6-phosphatase and ouabain-inhibited (Na + + K +)-ATPase in this fraction. A typical micrograph of the crude secretory granule fraction is shown in Fig. 1. The granules are seen to be largely independent and not aggregated and usually spherical, although some irregular shapes are seen. At still higher magnifications (not shown), there is an indication of a limiting membrane around the granule. In agreement with the marker enzyme studies, the extent of visible contamination by other cytoplasmic organelles was minimal. Following establishment of the relative purity of the secretory granule fraction, we determined the amount of both G H and prolactin in the crude secretory granule fraction. The results shown in Table I indicate that the procedure of preparing

Homogenate a 171 249 Crude secretory granule fraction b 806 120

1.46 0.15

n g / m g protein; b 8.3 mg protein/g wet weight,

a

the crude secretory granule fraction preferentially factors G H at the expense of prolactin. Since the usefulness of the isolated secretory granules is to a large extent dependent upon their in vitro stability, a series of experiments was desig100'

90'

80. 70-

~60I<

Z50. UJ CL U)

z4O_.g Ul

~30"

~20 10, )r

•

•

--U

• pH 10.9

lb

Fig. 1. Electron micrographs of crude secretory granule fraction. Original magnifications: left panel, 4500×; right panel, 20000x.

2'o

a'0

4b

5b

eb

r'0

TIME (mini Fig. 2. Effect of in vitro incubation on protein release from isolated pituitary secretory granules. Granules were incubated at 23°C in 20 mM potassium phosphate, pH 6 and at indicated intervals an aliquot was removed, centrifuged and protein content of both supernatant and pellet determined. Amount of protein in supernatant expressed as a percent of total (supernatant + pellet). After 60 rain the pH of the incubation medium was made 10.9 with NaOH resulting in complete solubilization of granule proteins.

212

< 0.50"

/

./

0.40 -

o~ o

~ e

o/

•f

0.30"

X.

0.20 "

0.1o-

pH

Fig. 3. Effect of pH on turbidity (As•o) of secretory granule suspension.

ned to determine the extent of spontaneous granule lysis. Isolated granules were resuspended in 20 mM potassium phosphate, pH 6 (osmolality = 68 mmsMol/1). At selected timed, an aliquot of the granule suspension was removed, centrifuged and the protein content of both the supernatant and pellet determined. Fig. 2 shows the percentage of total protein found in the supernatant as a function of time. The total amount of spontaneous solubilization under these conditions was quite low

"S

y S]

uJ

..1

and is comparable to that found for platelet agranules [8]. When, however, a pulse of NaOH was introduced sufficient to raise the pH to 10.9, complete solubilization resulted. To further assess granule stability, a series of turbidity measurements were carried out in which As0o was monitored as a function of time under a variety of conditions. Alkaline pH was found to abruptly decrease granule turbidity as did exposure to 0.5% SDS. Interestingly, exposure of the isolated secretory granules to the following agents a n d / o r treatments had no effect on the measured As0•: resuspension in distilled water, 300 mM sucrose-buffered at pH 6, 1 mM EGTA, 1 mM Ca 2+ with or without 10 -5 M ionophore A23187, 1 M KC1, 3 min at 100°C, 2 min of sonication and 5 sequential freeze-thaw cycles. The resistance of the isolated granules to such severe handling suggests a remarkably stable structure. Sherline et al. [9] reported the use of 60 min of stirring in 15 × excess water at 4°C to lyse pituitary secretory granules. We found such a procedure did not alter the granule turbidity. The effect of pH on granule stability is shown in Fig. 3. A maximum stability is seen at pH 5.5, although between pH 4.5-6.5 the secretory granules are reasonably stable to changes in pH. To validate that such turbidity measurements could detect lysis a similar set of experiments were done using chromaffin granules prepared from bovine adrenal according to the technique of Carty et al. [10]. After resuspension of the purified granules in 0.32 M sucrose, the observed As00 was time-independent over a 15 min period. Exposure to the 5 mM Tris, pH 6 medium reported by Smith

-2

T A B L E I1

15

30 TIME

45 (rnin.)

60

15

30 TIME

45 (rnin.)

60

Fig. 4. Time-dependent release of G H (left panel) and prolactin (PRL) right panel) from isolated secretory granules. Granules incubated in 20 m M potassium phosphate, p H 6 with indicated addition. All values determined in quadruplicate and represented as % of granular content release (.~ _+S.E.). Note scale change between G H and prolactin release. O O, control; • #, 50 n M TRF; Ill, 10 s ionophore A23187. For prolactin, the presence of A23187 inhibited release ( P < 0.05) at times longer than 1 rain.

E F F E C T OF F R E E C A L C I U M C O N C E N T R A T I O N ON T H E S P O N T A N E O U S RELEASE OF G H A N D PROL A C T I N F R O M ISOLATED SECRETORY G R A N U L E S Values for G H and prolactin represent % released in 30 min at 25°C. [Ca] r (~M)

GH

Prolactin

<1 18 200

7.7 5.7 8.4

18.7 16.1 5.9

213 and Winkler [11] to lyse chromaffin granules caused a prompt and irreversible reduction in As0o (data not shown). As a final and important measure of in vitro granule stability, we measured the amount of spontaneous release of both G H and prolactin as a function of time expressed as a percent of the total G H or prolactin present in the granules. The resuits are shown in Fig. 4. Relatively, more G H is released into the media in agreement with Zanini and Giannattasio [12] suggesting that these granules are somehow more fragile a n d / o r susceptible to damage during the isolation procedure. A question of interest, in view of a report by Dular and LaBella [13], was whether introduction of exogenous agents could specifically, and perhaps selectively, alter the release of G H or prolactin from the isolated granules. To test this, isolated granules were suspended in 20 m M potassium phosphate, p H 6, with a calcium concentration of 0.3 m M and various agents added. At selected times, a sample of the suspension having a k n o w n protein concentration was removed, centrifuged and the supernatant analyzed for G H and prolactin. The results are seen in Fig. 4. There was no effect of thyrotropin releasing factor (TRF) on either prolactin or G H release from the isolated granules. Further, under control, unstimulated conditions there was little time dependent release. For both G H and prolactin, the calcium ionophore A23187 appeared to inhibit release. Since calcium is known to be a key element in the secretory process as well as a constituent of secretory granules we sought to determine whether there was a correlation between the calcium concentration of the medium and the extent of G H

PRL/GH

132,000gmax ) 15mln.

I 2.1N j

Band1

0.24

Be~d2

t.51

Band3

4.20

Band4

10.42

Pelst

3.45

Fig. 5. Results of discontinuous density ultracentrifugation to yield Band 4 enriched in prolactin granules.

94K 68K

4OK 25.7K

IL7K A

B

C

D

E

Fig. 6. SDS-polyacrylaminde gel electrophoresis. Lane A: known molecular weight standards (cytochrome C, 11.7 K (kDa); chymotrypsinogen,25.7 K; aldolase,40 K; bovine serum albumin, 68 K; phosphorylase A, 94K), Lane B: NIAMDD ovine prolactin (arrow indicates primary band), Lane C: NIAMDD ovine GH (arrow indicates primary band), Lane D: crude secretory granule fraction, Lane E: Band 4. and prolactin release from the isolated granules. Three free calcium concentrations (<< 1 /~M, 18 # M and 200 #M) were chosen using E G T A as a calcium buffer for the first two. The data shown in Table II indicate that there was no effect of the different free calcium concentrations on spontaneous G H release whereas there is a noticeable tendency for spontaneous prolactin release to increase as free calcium concentration decreases. Since the granule population in the above studies contained both G H and prolactin, we sought to obtain an enriched population of one type in order to draw more specific conclusions. Using the crude secretory granule preparation as a starting point, we sought to increase the p r o l a c t i n / G H ratio. The basic approach of Zanini et al. [2] was used with a few modifications as described in Methods. Fig. 5 indicates that following density gadient centrifugation, three distinct bands were visible at the following sucrose interfaces: 0.32 M / 1 . 2 M, 1.2 M / 1 . 6 M and 1.6 M/2.1 M. In addition to a pellet, there was also a diffuse band which occupied most of the 2.1 M sucrose space which we labeled as Band 4. When each of these bands was solubilized and assayed, it was found that Band 4 had the highest p r o l a c t i n / G H ratio. Accordingly, it was this fraction which we considered as en-

214 riched in prolactin granules and used for future studies. The recovered material in Band 4 had the same stability to in vitro incubation as did the initial secretory granule fraction. Band 4 material was insensitive to H20, 20 mM potassium phosphate, pH 6 and was also rapidly solubilized at alkaline pH. To assess the nature of the proteins associated with the crude secretory granules preparation, samples were solubilized and prepared for polyacrylamide gel electrophoresis. Fig. 6 shows the results of such an experiment. Lane B is a sample of ovine prolactin as supplied by the National Pituitary Agency. Since purified bovine material suitable for radioimmunoassay is not available and since ovine material crossreacts equally with bovine material, it was used throughout. There is significant contamination by unknown proteins of approx. 45 and 21 kDa. Lane C contains purified ovine growth hormone which appears to run with a molecular weight near 21 kDa and may well be the contaminant in the prolactin standard. The G H also had an unidentified lower molecular weight species. Lanes D and E are the solubilized crude secretory granule fraction and the enriched Band 4 fraction, respectively. Although Lane E was overloaded, it is clear that the p r o l a c t i n / G H has increased markedly in agreement with the radioimmunoassay data (Fig. 5). In both lanes D and E there are additional proteins near 75 kDa. Also, in Band 4 (lane E) there is the approx. 45 kDa protein observed in the prolactin standard. This was only faintly visible in the crude secretory granule fraction further supporting the enhanced p r o l a c t i n / G H ratio in Band 4. Discussion Procedures described above resulted in a secretory granule fraction in which G H and prolactin were the primary hormone species as evidenced by radioimmunoassay and electrophoresis. Further centrifugation through a discontinuous sucrose gradient after exposure to high salt and puromycin resulted in a Band enriched in prolactin. Electron micrographs and marker enzyme analysis indicated no significant contamination with other cellular constituents. We have not yet analyzed the fractions for the presence of other pituitary hormones.

The in vitro stability of the isolated pituitary secretory granules is in contrast to those isolated from other tissues. Pancreatic zymogen granules, while stable in isosmotic sucrose or urea and even in distilled water, are unstable in ionic solutions [14]. Platelet a-granules in addition to being osmotically sensitive, exhibit a Ca-dependent lysis in ionic solutions [8]. Chromaffin granules isolated from the adrenal medulla behave as ideal osmometers in a variety of media [15]. For these and other granule types, a reliable technique to assess granule stability has been the measurement of turbidity. For granules that are osmotically sensitive, such as platelet a-granules and chromaffin granules, there is an excellent correlation between lysis and turbidity. For the case of pituitary secretory granules, the other treatments that resulted in a change in turbidity were 0.5% SDS and alkaline pH. This agrees with the finding of Zanini and Giannattasio [12]. We found that there was a measureable amount of spontaneous G H and prolactin release from the isolated secretory granules. A possible conclusion is that release from the isolated secretory granules may not be associated with a decrease in turbidity as would be expected with swelling or lysis of the granules. There may be selective solubilization of the stored G H and release of monomeric units from the intact granules. This is a variation of the exocytosis vs nonexocytosis concept of cellular secretion and, of course, the two pathways are not necessarily mutually exclusive. Since the pituitary secretory granules are resistant to osmotic stress, it is logical to conclude that the contents are not osmotically active a n d / o r the secretory granule membrane is impermeable to water. Experiments with 3H20 indicate that the isolated granules do have a defined exchangeable water space of 0.8/~l/mg protein (J.S. Garrett and C.M. Moriarty, unpublished data). This is lower than that reported for the osmotically sensitive chromaffin granule [16,17] although the latter depends heavily on medium osmolality. We conclude that the contents of the pituitary secretory granule are not osmotically active. Evidence for a secretory granule membrane comes from the electron micrographs as well as the work of Zanini and co-workers [2,12,18] and Giannattasio et al. [19] in which exposure of isolated

215

prolactin granules to the detergent Lubrol solubilized the granule membrane while leaving the granular core intact. It is of interest that such membrane-less granules retained their configuration and did not spontaneously solubilize suggesting that the membrane is not essential for confining the granule hormonal content. The importance of the granule membrane may well relate to the isolation of the granular contents from the rest of the cytoplasm, receptor recycling a n d / o r membrane-membrane interactions essential for exocytosis. Further support for the existence of a granule membrane comes from the measurement of a membrane potential as well as a pH gradient associated with the isolated granules (Garrett and Moriarty, unpublished data and Ref. 20). In the case of pituitary granules as well as granules isolated from a variety of other tissues [14,21-23], lysis is observed in an alkaline pH. Presumably, the alkaline pH dissociates the normally osmotically inactive granular core. Since intragranular H 2 0 space is already low, an increase in intragranular osmolarity would lead to water entry, swelling and eventually lysis. The fact that normal plasma is slightly alkaline may explain the prompt dissolution of exocytosed granular contents. Several of our findings are in contrast with a report of Dular and LaBella [13]. With TRF, we were unable to show any stimulation of either G H or prolactin at calcium concentrations from 1 ktM to 1 mM. The concentration of TRF, 50 nM was shown to be effective on pituitary cells [6,24]. Such results were found in both spring and winter months, making reported seasonal variations [25] an unlikely explanation. When one considers the vectorial nature of the secretory granule-plasma membrane fusion prior to exocytosis, it is clear that any putative T R F receptors associated with the secretory granule must reside on the inward facing surface of the secretory granule membrane and thus unlikely to recognize T R F in the bathing medium. Even if such putative receptors could recognize and bind TRF, there is no a priori reason to suspect that they are of physiological importance in regulating release of granular contents. Calcium is known to be an intracellular messenger involved in the secretory response of both

G H and prolactin cells [26,27]. Ultimately, either calcium per se, or a subsequent messenger initiated by Ca 2÷ must signal the secretory granule, the repository of performed stored hormone. Our results indicate that altering the free calcium concentration has no effect on G H release whereas a reduction in [Ca]f does increase the release of prolactin. This suggests a significant difference in control mechanisms a n d / o r cytoarchitecture associated with the two granule types. Zanini and Giannattasio [12] have also reported marked differences in the fragility of both granule types. We found no corresponding change in the turbidity of the granule suspension to correlate with the increased prolactin release again raising the possibility of selective partial solubilization of granular contents. In agreement with our results, Lemay et al. [28] found that EGTA had no effect on G H release but solubilized prolactin. Howell and Ewart [29] also found EGTA to have no effect on G H release. However, Zanini and Giannattasio [12] reported no effect of EDTA, a less specific calcium chelator, on prolactin release from isolated granules. The possibility exists that changes in local calcium concentrations in the vicinity of the prolactin granule might be a trigger for hormone release. It is of interest to note that the calcium ionophore A23187 actually decreased prolactin release, in effect stabilizing the granule. Calcium is a known component of a variety of secretory granules [30-32] including the anterior pituitary, although the role such calcium plays in the architecture and stability of the granular constituents is not known. In the case of the adrenal and pancreatic islets, the intracellular granules take up calcium when the cells are stimulated to secrete and such calcium is subsequently inaccessible to EDTA. In the case of the anterior pituitary, we have shown that cells stimulated with either depolarizing levels of potassium or the phosphodiesterase inhibitor theophylline show a decrease in the cytochemically localized calcium found associated with the secretory granules [33]. The nature of this calcium and its possible role in hormone solubilization from the prolactin granules together with the possible role secretory granules might play in cytosolic buffering of calcium is currently under investigation.

216

Acknowledgment The authors appreciate the secretarial assistance of Ruth Cozette and the technical assistance of Kyung Park. This work was supported by NIH grant AM-18328.

References 1 Rose, R. and Loewenstein, W.R. (1975) Science 190, 1204-1206 2 Zanini, A., Giannattasio, G., Nussdorfer, G., Margolis, R.K., Margolis, R.U. and Meldolesi, J. (1980) J. Cell Biol. 86, 260-272 3 Dowd, F. (1980) Arch. Oral Biol. 25, 773-780 4 Lowry, O.H, Rosebrough, N.J., Farr, A.L. and Randall, R.J. (1951) J. Biol. Chem. 193, 265-275 5 Bets, D.M. (1982) Am. J. Physiol. 242, C404-C408 6 Moriarty, C.M. and Leuschen, M.P. (1981) Am. J. Physiol. 240, E705-E711 " 7 Laemmli, U.K. (1970) Nature 227, 680-685 8 VanderMeulen, J. and Grinstein, S. (1982) J. Biol. Chem. 257, 5190-5195 9 Sherline, P., Lee, Y.C. and Jacobs, L.S, (1977) J. Cell Biol. 72, 380-388 10 Carty, S.E,, Johnson, R.G. and Scarpa, A. (1980) Anal. Biochem. 106, 438-445 11 Smith, A.D. and Winkler, H, (1966) Biochem. J. 103, 480-482 12 Zanini, A. and Giannattasio, G. (1974) Advances in Cytopharmacology (Ceccarelli, B., Clemente, F. and Meldolesi, J., eds.), Vol. 2, pp. 329-339, Raven Press, New York 13 Dular, R. and LaBeUa, F, (1977) Life Sci. 21, 1527-1534 14 Burwen, S.J. and Rothman, S.S. (1972) Am, J. Physiol. 222, 1177-1181 15 Winkler, H. and Westhead, E. (1980) Neuroscience 5, 1803-1823

16 Pollard, H.B., Zinder, O., Hoffman, P.G. and Nikodejevic, O. (1976) J. Biol. Chem. 254, 4544-4550 17 Johnson, R.G. and Scarpa, A. (1976) J. Gen. Physiol. 68, 601-603 18 Zanini, A. and Giannattasio, G. (1973) Endocrinology 92, 349-357 19 Giannattasio, G., Zanini, A. and Meldolesi, J. (1975) J. Cell Biol. 64, 246-251 20 Carty, S.E., Johnson, R.G. and Scarpa, A. (1982) J. Biol. Chem. 257, 7269-7273 21 Howell, S.L. (1974) in Advances in Cytipharmacology (Ceccarelli, B., Clemente, F. and Meldolesi, J., eds.), Vol. 2, pp. 319-327, Raven Press, New York 22 Poisner, A.M. and Hong, J.S. (1974) in Advances in Cytopharmacology (Ceccarelli, B., Clemente, F. and Meldolesi, J., eds.), Vol. 2, pp. 303-310, Raven Press, New York 23 Coore, H.G., Hellman, B., Pihl, E. and Taljedal, I.B. (1969) Biochem. J. 11,107-113 24 Ostlund, R.E., Jr., Leung, J.T., Hajek, S.V., Winokur, T. and Melman, M. (1978) Endocrinology 103, 1245-1252 25 Dular, R. and LaBella, F. (1977b) Endocrinol. Res. Commun. 4, 195-203 26 Moriarty, C.M. (1978) Life Sci. 23, 185-194 27 Rubin, R.P. (1982) Calcium and cellular secretion, pp. 1-276, Plenum Press, New York 28 Lemay, A., Labrie, F. and Drouin, D. (1974) Canad. J. Biochem. 52, 327-335 29 Howell, S.L. and Ewart, R.B.L. (1973) J. Cell Sci. 12, 23-35 30 Thorn, N.A., Russell, J.T. and Vilhardt, H. (1975) Ann. N.Y. Acad. Sci. 248, 202-217 31 Clemente, F. and Meldolesi, J. (1975) J. Cell Biol. 65, 88-102 32 Borowitz, J.L., Fura, K. and Weiner, N. (1965) Nature 205, 42-43 33 Leuschen, M.P., Moriarty, C.M., Sampson, H.W. and Piscopo, I. (1981) Cell Tiss. Res. 220, 191-200