Chromogranin A: posttranslational modifications in secretory granules.

0013-7227/91/1281-0174$02.00/0 Endocrinology Copyright © 1991 by The Endocrine Society

Vol. 128, No. 1 Printed in U.S.A.

Chromogranin A: Posttranslational Modifications in Secretory Granules JUAN A. BARBOSA, BRUCE M. GILL, MARWAN A. TAKIYYUDDIN, AND DANIEL T. O'CONNOR Department of Medicine and Center for Molecular Genetics, University of California, and Veterans Administration Medical Center, San Diego, California

ABSTRACT. The primary structure of chromogranin A indicates multiple domains which might be subject to posttranslational modification. We explored chromogranin A's proteolytic cleavage, glycosylation, and possible intermolecular disulfide links, using biochemical and cell biological approaches. Antichromogranin A region-specific immunoblots on chromaffin granules suggested bidirectional endoproteolytic cleavage of chromogranin A; control experiments ruled out artifactual cleavage during granule isolation or lysis. Isolation of chromogranin A-derived peptides by gel filtration chromatography or sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDSPAGE), followed by N-terminal amino acid sequencing, established several cleavage sites, including at least two at dibasic sites. Secretion of chromogranin A from bovine chromaffin cells did not initiate further cleavage, nor did prolonged exposure of secreted chromogranins to the secretory cells. The chromogranin A cleavage pattern was qualitatively similar in other neuroen-

docrine tissues, though cleavage was more complete in adrenal medullary than in anterior pituitary hormone storage vesicles, and N-terminal fragments of 45 and 55 kilodaltons were more prominent in the hypothalamus. A similar cleavage pattern was seen in human pheochromocytoma granules, as judged by chromogranin A region-specific immunoblots, fragment isolation by SDS-PAGE, and microsequencing. The presence of full-length chromogranin A as the core protein of a chromaffin granule soluble proteoglycan was suggested in bovine (but not human) chromaffin granules by glycoprotein staining, chondroitinase ABC digestion, chemical deglycosylation, and region-specific immunoblotting. Human (but not bovine) chromogranin A displayed intermolecular disulfide crosslinks on SDS-PAGE gels and immunoblotting. These results document diverse structural paths that the chromogranin A molecule may take in endocrine secretory cells after its translation. (Endocrinology 128: 174190, 1991)

C

posttranslational processing of chromogranin A in endocrine secretory granules.

HROMOGRANIN A is a 48 kilodalton (kDa) acidic protein (1, 2) originally described in the soluble core of catecholamine storage vesicles (3, 4), but more recently found in a wide variety of peptide hormone and neurotransmitter secretory vesicles (5, 6). Its primary structure (Fig. 1), deduced from its cDNA sequence (1, 2, 7-9), contains 8-10 sets of paired basic residues, suggesting that the molecule may be cleaved by proteases of appropriate specificity (1, 2, 7-10). Indeed, an internal region of chromogranin A is identical in sequence to pancreastatin (9), a peptide capable of suppressing pancreatic /3-cell insulin release (11). Pancreastatin may have even more widespread effects on endocrine and exocrine secretion (12-14). Simon et al. (15) have also suggested that fragments of chromogranin A may modulate chromaffin cell catecholamine secretion. We therefore examined several routes of endogenous

Materials and Methods Proteins Chromaffin granules were prepared from normal bovine or human adrenal medulla, and human pheochromocytoma on sucrose density step gradients (0.3 M/1.6 M) as previously described (6, 16, 17). Dopamine-/J-hydroxylase was removed from chromaffin granule lysates by Concavalin A-Sepharose (Pharmacia, Piscataway, NJ) lectin affinity chromatography (18, 19). Bovine anterior pituitary hormone storage vesicles were isolated as previously described (6). Bovine and human chromogranin A were purified from chromaffin granules as previously described (17, 20). Boiled tissue extracts (21) were obtained by immediately boiling fresh neuroendocrine secretory tissues (bovine adrenal medulla, anterior pituitary, hypothalamus, pancreas, and jejunoileum) for 5 min, followed by homogenization in 0.001 M sodium phosphate, pH 6.5, at 1:10 ratio of tissue-buffer, reboiling for 5 min, and centrifugation at 10,000 X g for 10 min to sediment debris. Chromogranin A terminology was standard (22).

Received September 17,1990. Address correspondence and requests for reprints to: Daniel T. O'Connor, M.D., Nephrology/Hypertension (111H), VA Medical Center, 3350 La Jolla Village Drive, San Diego, CA 92161. Supported by the Department of Veterans Affairs and NIH Grants HL-35018 and HL-43275.

Peptides Synthetic peptide haptens were solid-phase synthesized from protected amino acids (23), cleaved from the resin by HF, 174

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 May 2015. at 16:28 For personal use only. No other uses without permission. . All rights reserved.

CHROMOGRANIN PROCESSING

175

Chromogranin A COOH

H,N FIG. 1. Putative structural domain map of bovine chromogranin A, based on its cDNA-deduced amino acid sequence (1,2).

-18+1

+431 ^ ^ Signal peptide WM Paired basic residues PiiH Caz+ binding homologies K ^ RGD (-arg-gly-asp-) [ ? ^ Psncreastatin homology

lyophilized, resuspended in 0.1% trifluoroacetic acid, and purified by reverse-phase HPLC on a semipreparative 1 x 25-cm C-18 silica column (Supelco), with a linear (0-60%, vol/vol) acetonitrile gradient in 0.1% trifluoroacetic acid. The peptides used were: bovine/human chromogranin A (1, 2, 8) N-terminal 17-mer (LPVNSPMNKGDTEVMKY); bovine chromogranin A (1, 2) C-terminal 16-mer (YELEKVAHQLEELRRG); human chromogranin B (24) N-terminal 16-mer (MPVDNRNHNEGMVTRY). Antisera

proteoglycan (32-34), some vesicle samples were electrophoresed and immunoblotted after predigestion by chondroitinase ABC (Sigma; 0.1 U/25 ng protein, 1 h, 37 C) or chemical deglycosylation (35) by trifluoromethane sulfonic acid (Sigma). Band density quantitation by reflectance densitometry on immunoblots (or transmission densitometry on autoradiographs) was done with a model 1650 scanning densitometer (Bio-Rad, Richmond, CA) using a GS-350 data system software package (Hoefer Scientific Instruments, San Francisco, CA) in an IBM-PC-AT microcomputer. Preparative gel filtration chromatography

Rabbit antisera to intact bovine or human chromogranin A were developed as previously described (6, 16, 20). The antichromogranin A intact molecule antisera did not recognize the nitrocellulose-immobilized N- or C-terminal chromogranin A peptides (see Immunoblotting). Peptide haptens were coupled to the carrier protein keyhole limpet hemocyanin (Calbiochem, La Jolla, CA) by bisdiazobenzidine (Sigma Chemical Company, St. Louis, MO), using the adventitious terminal tyrosine residues (25). Rabbits were immunized in multiple intradermal sites (26) at 1-month intervals with hapten-carrier complex (1 mg peptide) in Freund's adjuvant. Antisera were harvested 2-4 weeks after the third boost (fourth immunization).

To isolate proteolytic fragments of chromogranin A, bovine chromaffin vesicle soluble lysates (CVL) were prepared as described (6,16,17), and dopamine-/8-hydroxylase was removed by affinity chromatography on Concavalin A-Sepharose (18, 19). The remaining chromogranins were concentrated by lyophilization, and gel filtered on a 2.5 x 90-cm column of Sephacryl S-300 (Pharmacia) equilibrated and eluted at 4 C with a high ionic strength, neutral pH, volatile buffer (0.3 M ammonium acetate, pH 6.5), monitoring protein elution by A28oEluted fractions were analyzed by SDS-PAGE. Fractions containing partially purified chromogranin A fragments were pooled, lyophilized, and subjected to up to three additional cycles of gel filtration chromatography, again monitoring eluted fractions by A28o and SDS-PAGE. The resulting peptides were Analytical sodium dodecyl sulfate-polyacrylamide gel electropho- sequenced (1 nmol) on a model 670A sequencer (Applied Biosystems, Foster City, CA) (36). resis (SDS-PAGE) and immunoblotting One-dimensional (27) and two-dimensional (28) SDS-PAGE (10% acrylamide) slab gels were run as previously described (27, 28). Some samples were electrophoresed after disulfide reduction by the sulfhydryl reagent dithiothreitol (DTT; Calbiochem, La Jolla, CA) at 17 mM. SDS-PAGE gels were stained with Comassie brilliant blue (29) to visualize total proteins, or the periodic acid-Schiff stain (30) for glycoproteins. Immunoblotting was performed using an avidin-biotin complex bridge (Vectastain, Vector Laboratories, Burlingame, CA) with electrophoretic transfer to nitrocellulose sheets as outlined by Towbin et al. (31). Primary antisera were used at titers of 1:100 to 1:10,000 (vol/vol). Total proteins were stained by amido black (31). To evaluate the presence of chromogranin A epitopes in the core protein of a putative chromaffin vesicle lysate soluble

Polyvinylidenedifluoride (PVDF) blotting and amino acid sequencing To ascertain the N-terminal amino acid sequences of lower molecular weight proteolytic fragments of chromogranin A, bovine chromaffin vesicle soluble proteins (16) were electrophoresed on SDS-PAGE (10% acrylamide) slab gels (27), transferred electrophoretically to PVDF (Immobilon, Millipore) paper (37), and stained with amido black (31). Bands were excised and sequenced as described by Matsudaira (37), on a model 670A sequencer (Applied Biosystems). Preparative gel electrophoresis To resolve fragments of human chromogranin A, a human pheochromocytoma chromaffin vesicle lysate (17, 19) was sub-



176

jected to Concavalin A-Sepharose chromatography to remove dopamine-/?-hydroxylase (18, 19). The remaining chromogranins (5 mg protein) were separated on a 0.15-mm thick SDSPAGE slab gel (10% acrylamide), and the gel was stained with Coomassie brilliant blue (29). Bands were excised, soaked for 2 h in 0.1% SDS, 0 . 1 M HEPES, pH 7, then crushed, placed in dialysis tubing (Spectrapor, obtained from Fisher Scientific, Tustin, CA), and electroeluted overnight at 75 mA (65 V). Afterwards Vio vol 1 M ammonium acetate, pH 6.5, was added to the eluate, followed by 4 vol -20 C acetone (38). The solution was vortexed and incubated overnight at —20 C. The sample (thus precipitated and stripped of dye) was centrifuged (10,000 X g, 10 min, —5 C), whereupon the supernate was aspirated and the pellet was vacuum dried. Purity of the proteins was verified by re-electrophoresis of 15% of each sample on SDSPAGE gels. Proteins were sequenced on an Applied Biosystems Model 670A sequencer (36). Chromaffin cell culture and secretion Bovine adrenal chromaffin cells were isolated and maintained in primary culture in 24 well (2.5 x 105 cells per well) plastic plates by a previously described modification (39) of the protocol of Livett (40). Secretion was stimulated by the nicotinic cholinergic agonist carbamylcholine chloride (carbachol, 10"4 M) in the presence of 2 mM CaCl2 (39) for 10-80 min at 25 C. For pulse-chase experiments, bovine chromaffin cells (5 X 106 cells per plate) were maintained in 10-cm plastic dishes. Cells were pulse-labeled for 24 h with 0.5 mCi 35S-methionine (translation grade, 1055 Ci/mmol, Amersham, Chicago, IL) in methionine-deficient medium [10% dialyzed fetal bovine serum, 90% methionine-deficient Dulbecco's modified Eagle's medium (DMEM)]. At the conclusion of the pulse, cells were rinsed three times in release medium (0.15 M NaCl, 5 mM KC1, 10 mM Na HEPES, pH 7.0). Labeled, rinsed cells were grown again in 37 C, humidified 95% air/5% CO2 in 10% fetal bovine serum/90% DMEM. At appropriate time points (from 0-160 hours), cells were washed three times in release medium without carbachol, then lysed into 2 ml immunoprecipitation buffer (0.1% NP-40,1 mg/ml BSA, 0.15 M NaCl, 0.01 M Na HEPES, pH 7, 5 mM unlabeled methionine). Twenty-five microliters of preimmune rabbit serum and 250 [A killed Staphylococcus aureus cells (Pansorbin; Calbiochem, La Jolla, CA; at 1:10 (vol/ vol) in immunoprecipitation buffer) were added (41). After vortexing 2 h at 4 C, the solution was centrifuged in a microfuge at 13,000 X g for 5 min at 4 C, and the pellet was discarded. To the supernate was added 25 /il specific rabbit antiserum (see Antisera) and another 250 /A 1:10 (vol/vol) killed S. aureus cells. After vortexing and incubation overnight at 4 C, the sample was microfuged (13,000 X g, 4 C, 5 min), and the pellet was washed three times in 1 ml immunoprecipitation buffer, each time by resuspension and recentrifugation. The final pellet was dissolved in SDS gel sample buffer [2% (wt/vol) SDS, 10% glycerol, 10 mM sodium phosphate, pH 6.5, 0.01% bromphenol blue], heated at 100 C for 5 min, and submitted to SDS-PAGE (10% acrylamide slab gels), followed by Enhance (New England Nuclear Corporation, Boston, MA)-assisted fluorography (41). For pulse-release studies, labeled (33 ^Ci 35S-methionine/well;

Endo• 1991 Vol 128 • No 1

1066 Ci/nmol; for 24 h in methionine-deficient medium), rinsed cells (in 25-mm wells, at 2.5 x 106 cells/cm2) were exposed to the secretagogues 0.5 mM carbamylcholine chloride and 50 mM KC1 (100 mM NaCl), along with 1.8 mM CaCl2 in release medium at 37 C for 30 min. At the conclusion of secretagoguestimulated catecholamine release, the release medium was collected, immunoprecipitated, electrophoresed, and fluorographed as described above. Radioiodination of chromogranin A Human and bovine chromogranin A were radioiodinated with Na 125I (0.3 mCi/5 ng protein) using the chloramine-T method, as described (42). Preparations (2.4 X 105 cpm 125Ichromogranin A) were analyzed by SDS-PAGE followed by autoradiography. Some preparations of 125I-chromogranin A were immunoprecipitated (as described under Chromaffin cell culture and secretion) with antichromogranin A antisera, before SDS-PAGE and autoradiography. Disulfide cross-links in chromogranin A Bovine adrenal and human pheochromocytoma chromaffin vesicle lysates (16, 17) were submitted to SDS-PAGE and immunoblotting with or without pretreatment by the sulfhydryl reagent DTT, final concentration 17 mM. Some vesicle preparations were electrophoresed after prolonged (15 h) incubation of the vesicle lysate at 0 C or 37 C, or after preincubation with ascorbic acid (Sigma; 20 mM, 4 C, 15 h). In one experiment, bovine and human CVLs were subjected to SDS-PAGE with or without preincubation (25 C, 30 min) with either the sulfhydryl reagent DTT (Sigma; 20 mM) or the disulfide cystine (cys-SS-cys; Sigma; 20 mM). Assays Bovine (6, 16) and human (42) chromogranin A were measured by homologous-species, soluble-phase equilibrium RIAs, as previously described (6, 16, 42). Protein was measured by the Coomassie blue dye binding method (43).

Results Proteolytic cleavage of bovine chromogranin A in chromaffin cells Figure 2 (immunoblotting of bovine chromaffin vesicle lysate with antisera directed against the synthetic termini of chromogranin A) indicates that antisera to each terminus detect not only chromogranin A itself but also lower molecular size immunoreactive forms, suggesting that the molecule is processed bidirectionally, and predominantly from the N-terminus inward (yielding predominantly C-terminal immunoreactive fragments). How consistent and reproducible is the pattern of proteolysis of chromogranin A? In Fig. 3, region-specific chromogranin A immunoblots reveal essentially the same apparent processing pattern in vesicles obtained on two occasions separated by 3 yr. Hence, the apparent pattern



177

4.2 1

IEF, pH

B

7.5 I

IEF, pH

4.2 I

130- pg - CgA

75v

s" (A O

17-

FlG. 2. Two dimensional (charge, followed by size) gel electrophoretic separation of bovine CVL (chromaffin vesidelysate) proteins. A, Panel stained with amido black for total protein. B, Panel stained with an antiserum directed against a chromogranin A (CgA) synthetic N-terminus (LPVNSPMNKGDTEVMKY). C, Panel stained with an antiserum directed against a chromogranin A synthetic C-terminus (YELEKVAHQLEELRRG). Pg, Proteoglycan. CgB, chromogranin B. DBH, Dopamine-/8-hydroxylase.

503927-

M •

17-

Sample: bovine CVL, 300ug Stain: amido black

Sample: bovine CVL, 300ug Stain: anti N-terminus # 7 8 4 - 2 , 1:100

4.2

13075- ft* 50s-

39-

2

2717-

Sample: bovine CVL, 300ug Stain: anti C-terminus #239-3, 1:100

is unlikely to be a preparation artifact. Could the proteolysis of chromogranin A have occurred during lysis of chromaffin vesicles in vitro after sucrose gradient isolation? Figure 4 reveals that the apparent processing pattern was the same whether gradient-isolated bovine chromaffin granules were lysed into hypotonic buffer (1 mM Na phosphate, pH 6.5) alone, or buffer supplemented with inhibitors of four major classes of proteases: metal ion-dependent proteases (EDTA), acid proteases (pepstatin) sulfhydryl proteases (parachlomeronuribenzoate), or serine proteases (phenylmethylsulfonylfluoride). Thus, the apparent processing pattern did not arise during granule lysis in vitro.

Could the proteolysis of chromogranin A have arisen during granule isolation from the adrenal medulla by homogenization and centrifugation, before granule lysis? Figure 5 shows immunoblots of a gradient-isolated bovine chromaffin vesicle lysate versus a sample of bovine adrenal medulla boiled upon removal from the animal, to inactivate and precipitate proteases while leaving chromogranin A soluble (6, 21). The patterns are qualitatively similar, suggesting that the proteolysis occurred in granules in vivo, before granule isolation. At what residues of chromogranin A does proteolytic cleavage take place? After removal of dopamine-/?-hydroxylase, a bovine chromaffin vesicle lysate was sub-



178 Amido Black

Stain

AntiCgA N-Terminus

Anti CgA Whole Molecule

Anti CgA C-Terminus

13075503927-

17-

Lane

1 2 3

1 2 3

123

1 2 3

1=Size standards 2=8ovine chromaffin vesicle lysate, 1987 3=8ovine chromaffin vesicle lysate, 1984

FIG. 3. Consistency of the apparent proteolytic processing pattern of chromogranin A in bovine chromaffin vesicle lysates (50 ^g protein) obtained on two occasions separated by 3 yr. Vesicles were lysed into 1 mM Na phosphate, pH 6.5, after which membranes were removed by centrifugation. The antibody developed against the chromogranin A intact molecule did not recognize the nitrocellulose-immobilized terminal peptides; hence, it recognizes interior (nonterminal) epitopes.

jected to multiple cycles of gel filtration (Fig. 6), with the resulting fractions analyzed by N-terminal amino acid sequencing (Fig. 6A) and region-specific anti-chromogranin A immunoblotting (Fig. 6B). Peak I had the amino acid sequence expected for chromogranin A (LPVNSPMNKGDTEVMKXIVE; Fig. 6A) and showed principally intact chromogranin A on immunoblotting (Fig. 7). Peak III, containing a variety of size forms on SDS-PAGE, was concentrated and again gel filtered. Since resulting fractions IV and V were still quite heterogeneous by SDS-PAGE (data not shown), they were separately concentrated and gel filtered again, yielding symmetric, single peaks VI and VII. Peak VI contained two sequences: LPVNSPMNKGDXEXM (corresponding to chromogranin A's intact Nterminus), and HSSYEDELSEVLEKP, that is, chromogranin A residues 79 and following; it had thus been cleaved at the first dibasic site near chromogranin A's N-terminus (KK, residues 77 and 78). On SDS-PAGE and immunoblotting, peak VI consisted mainly of an Mr 60 kDa C-terminal chromogranin A fragment with smaller amounts of an Mr 50 kDa C-terminal fragment and an Mr 45 kDa N-terminal fragment (Fig. 6B). Peak VII also contained two chromogranin A sequences: One cleaved at the first dibasic site preceding chromogranin A residue 79 (HSSYEDELSEVL), and

Endo • 1991 Voll28"Nol

another cleaved at the second dibasic site (KR) preceding chromogranin A residue 116 (DDFKEV). On SDS-PAGE and immunoblotting, Peak VII contained mainly an Mr 50 kDa C-terminal chromogranin A fragment, with smaller amounts of C-terminal Mr 60 kDa, 25 kDa, and 10 kDa fragments, as well as an Mr 45 kDa N-terminal fragment. The gel filtration, microsequencing, and region-specific immunoblot results (Fig. 6) thus indicate bidirectional cleavage of chromogranin A to smaller size fragments. At least two dibasic sites toward the N-terminus are used in the cleavage. At least two other cleavage sites toward the C-terminus are also used, but the data do not establish the precise location or sequence of these Cterminal sites. When bovine chromaffin vesicle proteins were separated by SDS-PAGE, transferred to PVDF sheets, stained with amido black, excised, and microsequenced (Fig. 7), the results also supported the conclusion of bidirectional processing. An Mr 60 kDa band contained both chromogranin A's N-terminal sequence (LPVNSPMNK) and a chromogranin A fragment cleaved at the first dibasic site (KK) preceding residue 79 (HSSXEXE). An Mr 50 kDa fragment had chromogranin A's intact N-terminal sequence (LPVNSPMNK), suggesting that it had been cleaved toward the C-terminus. Heterogeneity of the Mr 60 kDa band (Fig. 7) was also apparent upon inspection of immunoblots in Fig. 6B (lanes 1, 3, and 4) and Fig. 2. In Fig. 2, the SDS Mr 60 kDa region of the gel resolved into a more alkaline Nterminal fragment of chromogranin A (panel B) and a more acidic C-terminal fragment of chromogranin A (panel C). The time course of chromogranin A cleavage was examined by an 35S-methionine pulse-chase experiment in primary cultures of bovine chromaffin cells (Fig. 8). There was an immediate post-pulse (t = 0 h of chase) appearance of the Mr «70 kDa chromogranin A as well as an Mr «60 kDa fragment best visualized with the mid/ intact molecule antibody. Although the intensity of the pulse labeled chromogranin A band declined as a function of chase time (half-life = 79 h), there was not clear sequential conversion of the 70 kDa band to lower size forms of immunoreactivity over the time course of this experiment. In these experiments, only the cells themselves (not the media) were analyzed. Effect of secretion of bovine chromogranin A When chromogranins were released by secretagogue (carbachol)-stimulated exocytosis from cultured chromaffin cells into the medium (Fig. 9A), the act of secretion, in itself, did not result in further processing of chromogranin A, as judged by chromogranin A regionspecific immunoblots on secreted vs. stored chromogranins.


CHROMOGRANIN PROCESSING Arrido

AntiCgA

AntiCgA

AntiCgA

black

N-terrrinus

whole moleaJe

C-ternrinus

1 2 3 4 5 6 7

12 3 4 5 6 7

12 3 4 5 6 7

Stain

Lane

179

12 3 4 5 6 7

1: Size standards

Chromaffin vesicles lysed into buffer plus:

2 All 4 inhibitors 3SJTA 4: Pepstatin 5cPOvB 6cPMSF 7: No inhibitor

FIG. 4. Immunoblots of bovine chromaffin vesicle soluble lysates (25 jug protein) with chromogranin A region-specific antisera. The vesicles, obtained from sucrose density (0.3 M/1.6 M) step gradients, were lysed into either buffer alone (1 mM Na phosphate, pH 6.5), or buffer containing one of several protease inhibitors: EDTA (0.1 mM), pepstatin (0.1 Mg/ml), parachloromercuribenzoate (PCMB) (0.1 mM), or phenylmethylsulfonylfluoride (PMSF) (0.1 mM). After vesicle lysis, vesicle membranes were removed by centrifugation (30,000 x g, 3 h), whereupon the soluble proteins in the supernate were analyzed by SDS-PAGE and immunoblotting.

Stain

Amido black

Anti CgA Anti CgA N-terminus Whole mol

Anti CgA C-terminus

SDS

50-

392717- it '

Lane

123

I 123

123

123

1: Size standards 2: Bovine adrenal CVL 3: Bovine boiled adrenal medulla

FIG. 5. Immunoblots of gradient-isolated bovine chromaffin granule core lysate (25 fig protein) us. soluble proteins (25 /xg) obtained from bovine adrenal medulla boiled upon removal from the animal, to inactivate proteases.

Nor was further processing of chromogranin A detected when 35S-methionine pulse-labeled chromaffin cells released labeled chromogranin A into the medium in response to carbachol (Fig. 9B). In this experiment, the secretagogue enhanced secretion of labeled chromogranin A immunoreactivity to 1.8-fold of basal. What if secretagogue-released chromogranins were allowed to remain in contact with chromaffin cells for prolonged periods? Figure 9C shows that even when released chromogranins were exposed to chromaffin cells for up to 80 min, no apparent further processing (beyond that which had already occurred in chromaffin cells) was noted, as judged by chromogranin A region-specific immunoblotting. Cleavage of bovine chromogranin A in other endocrine tissues Region-specific antichromogranin A immunoblots of soluble core proteins of chromaffin granules versus anterior pituitary hormone storage granules, shown in Fig. 10A, suggested a qualitatively similar pattern of process-



180

Endo • 1991 Voll28«Nol

Panel a Sample: CVL-DBH LPVNSPMNKGOT6VMKXIVE (bCgA N-lerminus)

o10 25

0.20

0.15

0 10

0.05

\l

FIG. 6. Isolation and characterization of chromogranin A fragments from bovine chromaffin vesicles. A, Gel filtration of bovine chromogranins (after removal of dopamine-/3-hydroxylase) on a 2.5 X 90cm column of Sephacryl S-300, equilibrated, and eluted with a neutral pH, high ionic strength volatile buffer (0.3 M ammonium acetate, pH 6.5). Fractions were analyzed by SDS-PAGE, pooled, lyophilized, and submitted to repeated cycles of gel filtration to isolate multiple size forms of chromogranin A, which were then submitted to N-terminal amino acid sequencing. The fractions were also analyzed by immunoblotting (B). DBH, Dopamine-/3-hydroxylase. bCgA, Bovine chromogranin A. B, SDSPAGE immunoblots of pooled gel filtration fractions from panel A (isolation of chromogranin A fragments). The immunoblots were performed with antichromogranin A region-specific antibodies.

,LPVNSPMNKGOXEXM (UCgA N-terminus)

Panel d Sample: \

02

B

03

04 05 06 07 08 09 10 Vo/V,

Anti CgA N-terminus

Stain

Anti CgA Whole molecule

Anti CgA C-ter minus

L 3927-

17-

Larte

1 2 3 4 5 6 7

12 3 4 5 6 7

12 3 4 5 6 7

12 3 4 5 6 7

1=Size standards 2=fiovine chromaffin vesicle lysate 3=Peakl(M r 7Okd) 4=Peakll 5=Peaklll 6=Peak VI (Mr6Okd) 7=Peak VII (Mr5Okd)



PVDF BLOT DBH 13O

75

5-

LPVNSPMNKGDT

x

HSSXEXE

50' 39

LPVNSPMNK LPVNSPMNK

-

27 - !

17 -I

Bovine CVL FlG. 7. Microsequencingof bovine chromaffin granule proteins (25/*g) after SDS-PAGE, electrophoretic transfer to PVDF (polyvinylidene difluoride) sheets, amido black staining, and band excision. The sequence beginning LPV . . . corresponds to chromogranin A's intact Nterminus, while that beginning with HSS . . . has been cleaved at the first dibasic site preceding chromogranin A residues 79 and following. DBH, Dopamine-/3-hydroxylase. CgA, chromogranin A. CgB, chromogranin B.

ing, though chromogranin A was a comparatively minor component of pituitary granule cores. By densitometric scan of Fig. 10A, chromogranin A was more extensively processed to immunoreactive fragments in adrenal than in pituitary vesicles: on staining with the anti-intact molecule antibody (137s-l; 1:1000, vol/vol), the ratio of fragmentsxhromogranin A was 10:1 in adrenal us. 4.2:1 in pituitary; on staining with the anti-C-terminal antibody (239-3; 1:100, vol/vol), the ratio of fragmentsxhromogranin A was 5.8:1 in adrenal us. 2.9:1 in pituitary. By RIA (Fig. 10B, Table 1), chromogranin A immunoreactivity was also only a minor component of pituitary granule cores, accounting for 4.5 ± 0.4% of pituitary granule core protein, us. 46.0 ± 2.2% of chromaffin granule core protein. Figure IOC's region-specific anti-chromogranin A immunoblots compare the relative abundance and apparent route of cleavage of chromogranin A in several bovine neuroendocrine tissues: adrenal medulla, anterior pituitary, pancreas, gut (jejunoileum), and hypothalamus. The tissue samples were boiled after removal from the animals to inactivate any proteases that might yield postmortem (artifactual) cleavage. The adrenal medulla contained more chromogranin A immunoreactivity than other neuroendocrine tissues, consistent with prior RIA results (6, 44). The apparent pattern of processing was qualitatively similar across tissues, but the degree of processing (and the resulting fragments accumulating in

181

the tissues) differed markedly. For example, an SDS Mr 45 kDa chromogranin A N-terminal fragment was relatively more prominent in the hypothalamus and small intestine than in the adrenal medulla or anterior pituitary. An SDS Mr 55 kDa N-terminal chromogranin A fragment was especially prominent in the hypothalamus. Proteolysis of human chromogranin A Figure 11 A, region-specific chromogranin A immunoblots of bovine adrenal as well as three human pheochromocytoma chromaffin granule cores, suggests a similar processing pattern in bovine and human vesicles. When chromogranins from one pheochromocytoma were isolated by preparative SDS-PAGE (Fig. 11B), proteins of Mr 70 kDa and 50 kDa were obtained. The Mr 70 kDa chromogranin A had the expected intact Nterminal sequence (LPVNSPMNKGDTEV), whereas the 50 kDa form had been cleaved at the second dibasic site (KR) preceding residue 119 (EDSKEAEKSGEATD) (8, 45). Glycosylation: a chromogranin A core proteoglycan

Glycoprotein staining of bovine chromaffin granule cores after SDS-PAGE (Fig. 12A) revealed an SDS Mr 80-90 kDa glycoprotein spot (also stained by Coomassie brilliant blue) of slower mobility than chromogranin A. Dopamine-/3 hydroxylase removal by lectin affinity on Concavalin A-Sepharose did not affect the 80-90 kDa glycoprotein spot. The Mr 80-90 kDa spot's abundance varied somewhat from preparation to preparation of chromaffin granules (Fig. 3), and it was present in anterior pituitary as well as adrenal chromaffin granules (Fig. 10A). Inspection of Fig. 2 reveals that the Mr 80-90 kDa immunoreactivity was more acidic than chromogranin A, and was stained by both N-terminal and C-terminal chromogranin A antisera. The Mr 80-90 kDa form was exocytotically coreleased with chromogranin A from chromaffin cells during secretagogue exposure (Fig. 9). The 80-90 kDa chromogranin A immunoreactive spot was selectively abolished by predigestion with chondroitinase ABC (Fig. 12B). The action of chondroitinase ABC was specific—the enzyme decreased the densitometric ratio of the Mr 80-90 kDa spot:chromogranin A from 6.6:1 down to 0.27:1; by contrast, the ratio of chromogranin A immunoreactive fragments: chromogranin A was unchanged (1.9:1) by chondroitinase ABC. This 80-90 kDa form of chromogranin A immunoreactivity was also abolished by chemical deglycosylation with trifluoromethane sulfonic acid (data not shown). The Mr 80-90 kDa form was not seen in human pheochromocytoma chromaffin granules (Fig. 12 C).



182

Endo • 1991 Voll28*Nol

PULSE LABELING Antibody

P

N M C

M

P

' , * = 79/7 c£ 10" C 3

/" = 0.96

O 0>

i§ si 103

Lane Chase time, hours

1 2 0

0

3 4

5 6

7 8

0 0

04080160 04080 160

40

80 120 Chase times, hours

9 10 11 12

P: preimmune serum Nfc anti CgA N-terminus Ivl anti CgA intact molecule C: anti CgA C-terminus

FIG. 8. Time course of chromogranin A cleavage in primary cultures of unstimulated bovine chromaffin cells. Three-day cultures (10 cm plates, « 5 x 106 cells per plate) were pulse-labeled for 24 h with 0.5 mCi 3SS-methionine in methionine-deficient medium, rinsed, and incubated in complete medium. At chase time points as indicated, the medium was removed and the cell pellets were lysed, immunoprecipitated, and subjected to SDS-PAGE, whereupon the gel was fixed with a fluor (Enhance), dried, and fluorographed at —70 C. The chromogranin A band was quantified by transmission densitometry.

Disulfide cross-links in human chromogranin A An apparent chromogranin A disulfide-linked dimer was noted in human pheochromocytoma granules but not bovine adrenal granules (Fig. 13A)—in the absence of DTT, a portion of the chromogranin A immunoreactivity migrated at Mr 140 kDa (see triangle, Fig. 13A), us. 70 kDa for intact monomeric chromogranin A. Staining with a chromogranin B antiserum (Fig. 13A) indicated that chromogranin B was relatively more abundant in human than bovine chromaffin granules, but that the Mr 140 kDa chromogranin A immunoreactivity was not chromogranin B. The Mr 140 kDa chromogranin A dimer was not seen on immunoblots of normal human adrenal chromaffin granules (data not shown). Control studies indicated that the apparent dimerization of human chromogranin A was not the result of ascorbate deficiency—incubation of chromaffin granule soluble cores with 20 mM exogenous ascorbate (15 h, 4 C) before SDS-PAGE did not affect the 140 kDa chromogranin A immunoreactive band (data not shown). Nor did prolonged in vitro incubation of human or bovine granule cores (at 4 C or 37 C, for 15 h) before SDS-

PAGE result in further dimer formation (data not shown). While preincubation with the sulfhydryl reagent DTT (20 mM, 25 C, 30 min) abolished the Mr 140 kDa dimer band, the dimer was not affected by preincubation with the disulfide cystine (cys-S-S-cys; 20 mM, 25 C, 30 min) (data not shown). Some purified (17, 20) human chromogranin A preparations, when radioiodinated and analyzed by SDSPAGE followed by autoradiography (Fig. 13B), displayed apparent multimers of up to five disulfide-crosslinked chromogranin A molecules. Control studies indicated that radioiodination itself did not dimerize either bovine or human chromogranin A (Fig. 13C). Nor did in vitro exposure of chromogranin A to the radioiodination oxidizing agent chloramine T (at 1 mM), followed by the radioiodination reducing agent sodium metabisulfite (at 1 mM) (46) result in dimerization of chromogranin A (data not shown)!

Discussion Cleavage of bovine chromogranin A Region-specific anti-chromogranin A immunoblots suggest that chromogranin A is cleaved to lower molec-


CHROMOGRANIN PROCESSING FIG. 9. Effect of secretion on proteolytic cleavage of chromogranin A. A, Regionspecific chromogranin A immunoblots on secreted vs. stored chromogranins. Bovine adrenal chromaffin cells in primary culture (10 cm plates, «2.5 x 106 cells/cm2) were rinsed three times in release medium, and then stimulated to secrete by the nicotinic cholinergic agonist carbachol, 10~4 M, in the presence of 2 mM CaCl2. Lane 2 is 25 ng of heat stable chromaffin cell proteins (see Fig. 5). Lanes 3 and 4 represent release medium after application of the secretagogue for 60 min. B, Release of labeled chromogranin A from cultured bovine chromaffin cells. Cells were labeled with 36 S-methionine as described in Materials and Methods, then rinsed, and stimulated to secrete over 30 min with 0.5 mM carbachol plus 50 mM KC1 (100 mM NaCl) in the presence of 1.8 mM CaCl2. After release, the medium was immunoprecipitated (by preimmune serum or anti-chromogranin A intact molecule serum) as described in Materials and Methods, and subjected to SDS-PAGE, fluorography, and transmission densitometric scanning. The secretagogues enhanced release of labeled chromogranin A to 1.8-fold of basal. C, Immunoblots of secretagogue-released chromogranin A in the medium as a function of time of exposure to chromaffin cells after release. Secretagogue-stimulated release was achieved by carbachol (0.1 mM) as described in Materials and Methods. Chromogranin A region-specific immunoblots are shown. Lane 2 is 25 /xg bovine adrenal CVL protein. Lanes 3-7 represent 10 n\ release medium.

Amido black

Antl CgA N-terminus

Antl CgA Whole mol.

183

B 35

S-chromogranin A release Control

Stimulated

13075" " *

5012 3 4

1: 2: 3: 4:

12 3 4

Size standards Chromaffln cells Carbachol-releas id proteins, 2Oul Carbachol-releas id protolns, 2ul

3927-

Anti CgA

17Antiserum Preimmune Anti CgA

1: Size standards 2: Chromaffin cells 3: 10'

5: 40' 6: 60' 7: 80'

Medium during 0.1 mM carbachol stimulated release

ular weight fragments within chromaffin granules; that the cleavage is bidirectional; and that the main route of cleavage proceeds from the N-terminus inward. Wohlfarter et al. (47) arrived at a similar conclusion using a different set of chromogranin A synthetic peptide epitopes. Our data also establish that the apparent cleavage pattern is consistent, and occurs before granule isolation or granule lysis. Gel filtration and microsequencing of isolated chromogranin A fragment peptides suggested at least four cleavage sites in bovine chromogranin A, two of which are at dibasic sites towards the N-terminus of chromogranin A (residues 79 and 116). Microsequencing of SDS-PAGE-separated, PVDF-immunobilized chromogranin A fragments also confirmed bidirectional cleavage, and verified cleavage at the first dibasic site (residue 79). Cleavage of chromogranin A at dibasic sites is compatible with the action of endopeptidases such as serine

proteases (48). However, the precise endopeptidases that cleave chromogranin A are as yet unknown. Indeed, other kinds of endopeptidases, such as the calcium-dependent protease KEX2 (49), also recognize dibasic sites. A number of endoproteolytic activities have been described in endocrine hormone storage granules, including serine proteases (48, 50), aspartyl proteases (51), sulfhydryl (thiol) proteases (52), and calcium-dependent proteases (53). Hutton et al. (53) reported a chromogranin Acleaving calcium-dependent endoprotease in insulin storage granules, while Seidah et al. (48) have shown that chromogranin A is a substrate for a pituitary serine protease. More recently, Laslop et al. (54) characterized two chromogranin A-processing enzymatic activities found specifically in chromaffin granules, one a trypsinlike protease, the other a calcium-dependent sulfhydryl protease. In vitro, chromogranin A may also be a substrate for the serine esterase acetylcholinesterase (55). Nakano et al. (56) have identified pancreastatin as a cleavage product of bovine adrenal chromogranin A.



184

Endo• 1991 Voll28«Nol

TABLE 1. Immunoreactive chromogranin A measured by intact, whole molecule RIA in bovine hormone storage vesicles"

Antibody directed against

Tissue

Adrenal medulla Anterior pituitary

n

4 6

Immunoreactive chromogranin A, % of total protein (mean ± SEM) Vesicle soluble core

Vesicle membranes

46.0 ± 2.2 4.5 ± 0.4

4.5 ± 0.9 0.06 ± 0.006

° Chromogranin A (intact, whole molecule-based RIA) was measured in subcellular fractions (16). Total protein was measured by the Goomassie blue dye binding technique (43). The adrenal medullary chromaffin granule results have been tabulated previously (6). 1:100

B

1:100 1:1000

Anterior pituitary tissue fraction, 0.01

1

0.1

1

10

10

100

1000

Chromogranin A, ng/tube

Amido black

AntiCgA N-terminus

Anti CgA whole mol.

AntiCgA C-terminus

At what rate does chromogranin A cleavage proceed? A pulse-chase experiment revealed the appearance of both the SDS Mr 70 kDa chromogranin A and an Mr 60 kDa chromogranin A fragment at the end of the 24-h pulse, but there was not clear ongoing progression of the label from higher Mr to lower Mr forms over the ensuing 160 h of chase time. During this time, the SDS Mr 70 kDa parent chromogranin A band declined in intensity, with an apparent half-life of 79 h. Thus, the rate of endoproteolytic cleavage of chromogranin A in vivo must be quite slow, perhaps on the order of days. Other investigators have found that chromogranin A may remain undegraded for substantial time intervals following its biosynthesis in chromaffin cells (57-60). In contrast, Hutton et al. (61) reported that endoproteolytic cleavage of newly synthesized chromogranin A to an approximately 20 kDa fragment began within 20 min in rat pancreatic islet cells. Effect of secretion of bovine chromogranin A

12 3 4 567

123 45 67

1234 5 67

123 4 5 6 7

1: Size standards 2: Chromaffin vesicle lysate 3: Adrenal medulla 4: Anterior pituitary 5: Hypothalamus 6: Small intestine 7: Pancreas

FIG. 10. Presence and processing of chromogranin A in multiple bovine neuroendocrine tissues. A, Region-specific antichromogranin A immunoblots in hormone storage granule cores from bovine adrenal medulla and bovine anterior pituitary. In each case, 25 fig soluble protein were electrophoresed. Hormone storage granules were prepared and lysed, and the granule membranes removed by centrifugation, as described in Methods. B, Bovine chromogranin A RIA (6,16) parallelism in homogenate and hormone storage granules of bovine anterior pituitary. (C) Antichromogranin A region-specific immunoblots on extracts from several bovine neuroendocrine tissues: adrenal medulla,

The act of secretion in itself did not result in further cleavage of either newly synthesized or previously stored chromogranin A. Even when newly secretagogue-released chromogranins were left exposed to chromaffin cells in situ for up to 80 min, no further cleavage (beyond that which had already occurred in chromaffin granules) could be demonstrated. After exocytotic secretion from chromaffin cells in vivo, chromogranins do not appear immediately in the bloodstream along with epinephrine (39, 62), because they cannot easily transit the adrenal capillary wall. The delay in peak plasma chromogranin A concentration behind that of epinephrine may be 60-90 min (39, 62). Our results suggest, however, that even if released chromogranin A remains in the vicinity of the chromaffin cell for up to 80 min, further cleavage is not apparent. anterior pituitary, hypothalamus, small intestine (jejunoileum), and pancreas. Lane 2 was 50 fig CVL proteins. Tissue extracts (lanes 3-7) were 50 fig of heat stable protein from boiled tissue homogenates.


CHROMOGRANIN PROCESSING Stain

FIG. 11. Proteolytic cleavage of human chromogranin A. A, Region-specific antichromogranin A immunoblots of chromaffin granule soluble cores from three human pheochromocytomas as well as from normal bovine adrenal medulla. Twenty-five micrograms of granule protein were applied per lane, after reduction by dithiothreitol (final concentration, 17 mM). B, Isolation of human chromogranin A (Mr 70 kDa) and an Mr 50 kDa fragment of human chromogranin A by preparative ("initial") SDSPAGE of a pheochromocytoma CVL. The purity of the gel-excised proteins was verified by reelectrophoresis (Final) before microsequencing. The sequence of the Mr 70 kDa protein matches the reported amino terminus of human chromogranin A, whereas, the Mr 50 kDa form possessed the sequence expected upon cleavage at the dibasic site preceding human chromogranin A residue 119 (8, 45).

Amido black

Anti CgA N-terminus

12 3 4 5

Lane

185

1234

Anti CgA whole molecule

Anti CgA C-terminus

123 4 5

12345

5

1=Size standards 2=33ovine CVL 3=3>heo. 1 CVL 4=Pheo.2CVL 5=*>heo. 3 CVL

B

Initial

Final

LPVNSPMNKGDTEV

: EDSKEAEKSGEATD

1 2

3

2

4

5

1 : Prestained size standards 2 : Post stained size standards 3 ; Pheochromocytoma CVL 4 : Mr 70 kd chromogranin {

Electron microscopic immunogold cytochemistry reveals chromogranin A confined to secretory granules of porcine Merkel cells.

Biogenesis of secretory granules.

Autoantibodies to posttranslational modifications in rheumatoid arthritis.

Dynamics of posttranslational modifications of p53.

Protein mobility within secretory granules.

Posttranslational Modifications and the Immunogenicity of Biotherapeutics.

Protein synthesis, posttranslational modifications, and aging.

Mutant TP53 posttranslational modifications: challenges and opportunities.

Proteoglycans in secretory granules of NK cells.

The nature of breast dense core granules: chromogranin reactivity.

Posttranslational modifications of α-tubulin in alzheimer disease.

Spatiotemporal regulation of posttranslational modifications in the DNA damage response.

Histone Posttranslational Modifications of CD4⁺ T Cell in Autoimmune Diseases.

Posttranslational modifications of proopiomelanocortin in vertebrates and their biological significance.

Mast cell secretory granules: armed for battle.

Toxoplasma secretory granules: one population or more?

Arrested exocytosis of atrial secretory granules.

Visualizing posttranslational and epigenetic modifications of endogenous proteins in vivo.

Protein Posttranslational Modifications: Roles in Aging and Age-Related Disease.

Oxidative and other posttranslational modifications in extracellular vesicle biology.

Posttranslational Modifications and Clearing of α-Synuclein Aggregates in Yeast.

Posttranslational modifications of FERREDOXIN-NADP+ OXIDOREDUCTASE in Arabidopsis chloroplasts.

Posttranslational modifications of nerve cytoskeletal proteins in experimental diabetes.

β-catenin signaling by posttranslational modifications.