Formation of the Catecholamine Release-inhibitory Peptide Catestatin from Chromogranin A

The catestatin fragment of chromogranin A is an inhibitor of catecholamine release, but its occurrence in vivo has not yet been verified, nor have its precise cleavage sites been established. Here we found extensive processing of catestatin in chromogranin A, as judged by catestatin radioimmunoassay of size-fractionated chromaffin granules. On mass spectrometry, a major catestatin form was bovine chromogranin A332–364; identity of the peptide was confirmed by diagnostic Met346oxidation. Further analysis revealed two additional forms: bovine chromogranin A333–364 and A343–362. Synthetic longer (chromogranin A332–364) and shorter (chromogranin A344–364) versions of catestatin each inhibited catecholamine release from chromaffin cells, with superior potency for the shorter version (IC50 ∼2.01 versus∼0.35 μm). Radioimmunoassay demonstrated catestatin release from the regulated secretory pathway in chromaffin cells. Human catestatin was cleaved in pheochromocytoma chromaffin granules, with the major form, human chromogranin A340–372, bounded by dibasic sites. We conclude that catestatin is cleaved extensivelyin vivo, and the peptide is released by exocytosis. In chromaffin granules, the major form of catestatin is cleaved at dibasic sites, while smaller carboxyl-terminal forms also occur. Knowledge of cleavage sites of catestatin from chromogranin A may provide a useful starting point in analysis of the relationship between structure and function for this peptide.

lease (9,10), suggesting a novel autocrine feedback mechanism controlling sympathochromaffin exocytosis. Initial studies of catestatin relied on synthetic peptides (10); hence, the existence of the peptide in vivo remains unexplored, as do its precise cleavage sites from chromogranin A.
In this study, we explored processing of the catestatin region of chromogranin A in secretory granules of chromaffin cells and sympathetic nerves, as well as in human pheochromocytomas. We documented catestatin cleavage from chromogranin A and determined the precise endogenous cleavage sites bounding catestatin in bovine and human chromaffin granules by chromatographic separations coupled with amino-terminal amino acid sequencing, immunoprecipitation, and matrix-assisted laser desorption ionization (MALDI) 1 mass spectrometry. We confirmed catestatin's regulated secretion by chromaffin cells. A major form of catestatin was processed at dibasic sites (bovine chromogranin A 332-364 or human chromogranin A 340 -372 ), while a smaller form was also identified (bovine chromogranin A 343-362 ). Both larger and smaller size forms were synthesized; each displayed specific antagonism of nicotinic cholinergicstimulated catecholamine release, while the smaller form had greater potency of inhibition.

MATERIALS AND METHODS
Preparation of Tissue Fractions-All preparative steps on tissues or protein fractions were conducted at 0 -4°C. Bovine adrenal medullary chromaffin granules were prepared by centrifugation on 0.3 M/1.6 M sucrose density step gradients, as described previously (10). After granule hypotonic lysis and centrifugal removal of granule membranes, the soluble proteins and peptides in the supernatant were size-fractionated on a 2.6 ϫ 80-cm Sephacryl S-300 column (Amersham Pharmacia Biotech), eluting with the volatile buffer 0.3 M ammonium acetate, pH 6.5, as described previously (11). The buffer was removed by lyophilization before further studies. In some experiments, bovine chromaffin granule proteins/peptides (200 l containing 8 mg of protein) were size-fractionated on a Superdex 75 HR 10/30 FPLC gel filtration column (10 ϫ 300 mm, 24-ml bed volume; Amersham Pharmacia Biotech), eluting at 1 ml/min with 0.3 M ammonium acetate, 1 mM EDTA, pH 7.0, collecting fractions every 0.5 ml (0.5 min). Eluted fractions were analyzed for protein by on-line absorbance at 280 nm (A 280 ) and then lyophilized (to remove the volatile buffer) and resuspended in the same volume of radioimmunoassay (RIA) buffer (50 mM Tris-HCl, pH 8.3, 0.3% bovine serum albumin, 0.1% Triton X-100; see below). Chromaffin granules from human pheochromocytomas were also prepared by centrifugation on 0.3 M/1.6 M sucrose density step gradients, followed by hypotonic lysis and membrane removal by centrifugation, as described previously (10). Postganglionic sympathetic nerves were obtained by dissection of splenic nerve in samples from the local slaughterhouse. Nerves were placed in ice-cold 0.3 M sucrose at an ϳ10:1 ratio of tissue to buffer and then minced, homogenized, gauze-filtered, and centrifuged at 1000 ϫ g for 10 min (to remove nuclei and debris). Supernatants were then centrifuged at 10,000 ϫ g for 10 h to pellet a crude fraction of neuronal large dense core vesicles (12), which were lysed by resuspension in 10 mM Hepes, pH 7, and freezing/thawing.
Chromogranin A, Synthetic Peptides, and Antibodies-Bovine or human chromogranin A was isolated from chromaffin granule soluble core proteins by affinity chromatography (to remove dopamine ␤-hydroxylase) followed by gel filtration, as described previously (13). Recombinant human chromogranin A was expressed in Escherichia coli by a modification of previously described methods (14), except that a His 6 tag was used for affinity purification on a Ni 2ϩ -nitrilotriacetic acid column (15). Synthetic peptides (20 -100-mol scale) were prepared by the solid phase method, using Fmoc (N-(9-fluorenyl)methoxycarbonyl) protection chemistry. Purification was by C-18 reverse-phase HPLC (RP-HPLC). Authenticity of the resulting peptides was confirmed by mass spectrometry, using either MALDI or electrospray ionization. Polyclonal rabbit antisera recognizing the catestatin region of chromogranin A (Fig. 1), either bovine chromogranin A 344 -364 (RSMRLS-FRARGYGFRGPGLQL) or human chromogranin A 352-372 (SSMKLS-FRARAYGFRGPGPQL), were developed by a modification of protocols previously described for other chromogranin peptides (11,16). The polyclonal antibody recognizing the catestatin region of chromogranin A was further purified on an Amersham Pharmacia Biotech Hi Trap protein A column in 0.02 M sodium phosphate (pH 7.0), eluted with 0.1 M sodium citrate (pH 3), and the pH was then adjusted back to 7.0 with Tris-HCl (pH 8.8). A polyclonal rabbit antiserum recognizing bovine chromogranin A 316 -329 was obtained from Dr. Marie-France Bader (IN-SERM U-338, Strasbourg, France), and a polyclonal rabbit antiserum recognizing human chromogranin A 367-391 was obtained from Dr. Reiner Fischer-Colbrie (University of Innsbruck, Austria).
Immunoprecipitations-Tissue homogenates, granule soluble core lysates, or gel filtration size-separated fractions were applied to Sep-Pak C-18 cartridges (Waters/Millipore), eluted with 30 -40% acetonitrile, lyophilized, resuspended in 500 l of immunoprecipitation buffer containing protease inhibitors (0.1% Triton X-100, 140 mM NaCl, 0.025% sodium azide, 10 mM Tris-HCl, pH 8.0, 1 mM phenylmethylsulfonyl fluoride, 1 M pepstatin, 1 mM EDTA, 1 mM N-ethylmaleimide), and then immunoprecipitated by a modification of the protocol of Wang et al. (17). To minimize nonspecific binding results, samples were first incubated with 25 l of normal (preimmune) rabbit serum with constant rotator mixing at 4°C for 12-18 h. 60 l of protein G Plus/protein A-agarose beads (33% slurry; Calbiochem) were added, and rotational incubation was continued for 3 h, followed by centrifugation for 2 min at ϳ13,000 ϫ g in an Eppendorf 5417 microcentrifuge, after which the pellet was discarded. To the supernatant, 20 l of rabbit anti-bovine catestatin (10) were added, and rotational incubation was continued for another 12-18 h. 60 l of fresh protein G Plus/protein A-agarose beads were added, and rotational incubation continued for another 3 h, after which the beads were collected by centrifugation, washed three times with immunoprecipitation buffer, and then washed twice with 50 mM Tris-HCl, pH 8 (to remove Na ϩ and detergent).
Mass Spectrometric Analyses of Immunoprecipitated Catestatin-Immunoprecipitated catestatin was eluted from the immune complexes with 20 l of trifluoroacetic acid/water/acetonitrile, 1:20:20 (v/v/v) (17). To identify methionine-containing peptides (by oxidation of methionine to methionine sulfoxide, thereby adding the mass of a single oxygen at 16 daltons), 10 l were oxidized by adding sufficient 3% H 2 O 2 to achieve 10 M final H 2 O 2 concentration. 1-2 l were characterized by MALDI mass spectrometry on a Voyager-Elite mass spectrometer with delayed extraction (PerSeptive Biosystems, Framingham, MA). Samples were embedded in an ␣-cyano-4-hydroxycinnamic acid matrix (18) and then irradiated with a nitrogen laser at 337 nm, and the ions produced were accelerated with a deflection potential of 30,000 V. Ions were then differentiated according to their mass/charge ratio (m/z) using a timeof-flight mass analyzer. The mass error of this method is characteristically Յ0.1% (i.e. Յ1000 ppm; Ref. 18).
After RP-HPLC of bovine chromaffin granule peptides (see below), 50 l of each 500-l HPLC fraction were vacuum-dried, resuspended in 100 l of water, adsorbed by vacuum filtration onto a nitrocellulose membrane using a slot-blotting device (Minifold II; Schleicher and Schuell), and immunoblotted using the anti-bovine catestatin primary antibody (at 1:500 (v/v)) and peroxidase-conjugated anti-rabbit secondary antibody (at 1:7000 (v/v)) with an enhanced chemiluminescence (ECL) detection kit (Amersham Pharmacia Biotech).
Densitometry analysis was performed on a Macintosh computer using the public domain NIH Image program (developed at the National Institutes of Health and available on the Internet. RP-HPLC-Gel filtration size-fractionated bovine chromaffin granule peptides were further separated by RP-HPLC using a 25 ϫ 0.5-cm C-18 column, equilibrated in 0.1% trifluoroacetic acid, and eluted with a linear 0 -60% gradient of acetonitrile in 0.1% trifluoroacetic acid, over 60 min, at 1 ml/min. The elution was monitored by A 214 (peptide bond absorbance), and fractions were collected at 0.5-min (0.5-ml) intervals. 50-l aliquots from each fraction were vacuum-dried and subjected to anti-catestatin slot immunoblotting (see above), and slot-blot positive fractions were further subjected to amino-terminal microsequencing (100 l; see below) and MALDI mass spectrometry (1-2 l; see above).
Mass Spectrometric and Sequencing Analyses-Molecular weights from MALDI mass spectra were interpreted, and peptide fragments within the chromogranin A primary structure were assigned by the program PAWS (Protein Analysis WorkSheet, version 8.1.1, for Macintosh; ProteoMetrics; freeware available on the Internet), assigning average isotopic MH ϩ values for chromogranin A peptides (18). Sequencing results were analyzed at each Edman cycle by the algorithm "Hydrosites" (for Macintosh) to deconvolute multiple amino-terminal sequences derived from the same parent molecule at a given cycle (22). Activity of Synthetic Catestatins-Peptides were subjected to a test of activity by inhibition of secretagogue-stimulated norepinephrine release from [ 3 H]norepinephrine-prelabeled PC12 pheochromocytoma cells, over a 30-min secretion period, as described previously (10,23). The stimuli to catecholamine release were either nicotinic cholinergic (60 M nicotine) or membrane depolarization (by 55 mM KCl). In some control experiments, the peptide was immunoadsorbed overnight (4°C, 1:100 antibody titer in secretion buffer) prior to test of secretion.
Bovine Chromaffin Cell Isolation, Culture, and Stimulation of Secretion-Primary cultures of bovine chromaffin cells were prepared as described previously (24). Secretion was stimulated over a 15-min period. The stimuli were either nicotinic cholinergic (100 M nicotine), membrane depolarization (by 55 mM KCl, in the presence or absence of 1 mM Ca 2ϩ ), or vehicle (mock). Secretion media were collected, concentrated on a Millipore Ultrafree 10K filter, and analyzed by radioimmunoassay.
Statistics-Results are expressed as the mean value Ϯ one S.E.

The Catecholamine Release-inhibitory Activity of Catestatin Is Found in Bovine Adrenal Chromaffin Granule Low Molecular Weight Size
Fractions-Chromaffin granule soluble core proteins (chromogranins) were size-fractionated by gel filtration (Fig. 2); fractions were analyzed by SDS-PAGE (10), and low molecular weight peptide-containing fractions 42-52 were pooled for further study. After peptide adsorption and batch elution from a C-18 hydrophobic affinity matrix (SepPak), this fraction inhibited nicotinic cholinergic-stimulated catecholamine release from [ 3 H]norepinephrine-prelabeled PC12 pheochromocytoma cells; 60 M nicotine stimulated 19.1 Ϯ 0.39% release of cellular norepinephrine, but inclusion of as little as 1.25 g/ml of the peptide fraction diminished nicotinicstimulated release to 1.7 Ϯ 0.19% (10).
Determination of Catestatin Cleavage Sites in Bovine Chromaffin Granules by Mass Spectrometry-To determine whether the catestatin region of chromogranin A is discretely excised at particular amino acid residues, we turned to mass spectrometry on chromogranin A fragments formed endogenously in chromaffin granules. The low molecular weight peptide fraction (Fig. 2, fractions 42-52) was immunoprecipitated by a bovine catestatin (chromogranin A 344 -364 ) antibody and subjected to MALDI mass spectrometry (Fig. 3). The initial spectrum revealed a major peak at m/z ϭ 3829 (Fig. 3B), which, within the 431-amino acid bovine chromogranin A primary structure (25,26), corresponds uniquely to chromogranin A 332-364 (LE-GEEEEEEDPDRSMRLSFRARGYGFRGPGLQL; calculated m/z ϭ 3827.2); the 0.047% difference between experimental and calculated m/z values is within the range of acceptable error of the MALDI method (0.1% m/z; Ref. 18). Preimmune serum did not recognize a peak of this mass (Fig. 3A). Since mass spectra of peptides are susceptible to shifting by formation of peptide-cation (such as Na ϩ ) adducts, we repeated the spectrum on granule peptides that had been desalted by adsorption to a C-18 matrix (Fig. 3C); an m/z ϭ 3827 peak remained the principal component, with some diminution in surrounding minor peaks. To confirm that this peak represents catestatin, the granule peptide sample was gently oxidized with H 2 O 2 to convert any methionine residues into methionine sulfoxide, as described by Wang et al. in studies of ␤-amyloid processing by mass spectrometry (17); since bovine catestatin contains a single methionine residue (bovine chromogranin A Met 346 ; Fig. 1), this procedure provides a further specific diagnostic for the identity of the catestatin region. After oxidation, the major peak was now found at m/z ϭ 3844 (Fig. 3D), a mass shift corresponding to the added 16-dalton oxygen atom in Met 346 -sulfoxide.
Isolation and Characterization of Catestatin Fragments: RP-HPLC, Mass Spectrometry, and Amino-terminal Amino Acid Sequencing-In a second approach to determine the boundaries of endogenous catestatin cleavage, the chromaffin granule low molecular weight peptide fraction (gel filtration fractions 42-52; Fig. 2) was first separated by RP-HPLC (Fig. 4, bottom), and fractions were immunoblotted with anti-bovine catestatin. HPLC fraction 27, containing intense catestatin immunoreactivity (Fig. 4, middle), was then analyzed (Fig. 4, top) by both MALDI mass spectrometry and amino-terminal amino acid sequencing. MALDI revealed two peaks, at m/z ϭ 3718 and 2300 (Fig. 4, top). In the catestatin region of bovine chromogranin A (Fig. 1), m/z ϭ 3718 is compatible with chromogranin A 333-364 (EGEEEEEEDPDRSMRLSFRARGYGFRG-PGLQL), while m/z ϭ 2300 is compatible with chromogranin A 343-362 (DRSMRLSFRARGYGFRGPGL). Fraction 27 was amino-terminally sequenced over 10 residue cycles, with the result suggesting two peptides (22) Cleavage of the Catestatin Region of Chromogranin A in Neurons: Mass Spectrometry-Bovine splenic nerve was homogenized, desalted by adsorption to/elution from a C-18 matrix (SepPak), immunoprecipitated by an anti-catestatin (bovine chromogranin A 344 -364 ) antibody, and subjected to MALDI mass spectrometry. MALDI revealed a peak at m/z ϭ 3832 (Fig.   FIG. 2. Size separation of bovine adrenal medullary chromaffin granule soluble core proteins and peptides. After chromaffin granule isolation and lysis, soluble proteins and peptides were fractionated by gel filtration on a 2.6 ϫ 80-cm column of Sephacryl S-300 (Amersham Pharmacia Biotech), equilibrated and eluted with the volatile buffer 0.3 M ammonium acetate, pH 6.5. Low molecular weight (LMW) peptide fractions 42-52, devoid of chromogranin A by SDS-PAGE analysis, were used for further studies. By Coomassie Blue stain of SDS-PAGE, the low molecular weight fraction molecular mass values ranged from Ͻ55 kDa to the dye front. 5, top), which, in the catestatin region of bovine chromogranin A (Fig. 1), is within 0.125% of 3827.2, the calculated MH ϩ of chromogranin A 332-364 (LEGEEEEEEDPDRSMRLSFRAR-GYGFRGPGLQL). After oxidation by H 2 O 2 , the m/z ϭ 3832 peak diminished, while an m/z ϭ 3843 peak became more prominent (Fig. 5, bottom), a mass shift corresponding to the added 16-dalton oxygen atom in Met 346 -sulfoxide in chromogranin A 332-364 (predicted MH ϩ ϭ 3843.2).
Determination of Catestatin Cleavage Sites in Human Chromogranin A: Mass Spectrometry-After anti-catestatin immunoprecipitation of human pheochromocytoma chromaffin granules, m/z values of 3770 -3771 were noted (Fig. 6), corresponding uniquely within the chromogranin A primary structure to chromogranin A 340 -372 (KRLEGQEEEED-NRDSSMKLSFRARAYGFRGPGPQLRR; calculated m/z ϭ 3771.1), which is bounded on either side by dibasic recognition sites for prohormone cleavage (underlined) (Fig. 1). Upon H 2 O 2 oxidation, in each case these m/z 3770 -3771 peaks shifted to m/z ϭ 3787, consistent with the addition of an oxygen (16 daltons) to form Met 354 -sulfoxide. Other peaks in this region did not shift upon oxidation. Experimental and calculated m/z values are well within the 0.1% expected experimental error of MALDI mass determination (18).
Reproduction of Catestatin Activity in Synthetic Peptides Corresponding to Cleavage Sites: Potency and Specificity of Nicotinic Cholinergic Antagonism-To test the activity and specificity of the bovine catestatin peptides predicted by MALDI mass spectrometry and amino-terminal sequencing (Figs. 3 and 4), we synthesized both longer (chromogranin A 332-364 ; LEGEEEEEEDPDRSMRLSFRARGYGFRGPGLQL) and shorter (chromogranin A 344 -364 ; RSMRLSFRARGYGFRG-PGLQL) versions of catestatin as well as whole chromogranin A and evaluated their effects on catecholamine release from PC12 pheochromocytoma cells (Fig. 7).
As previously reported for the shorter version of catestatin (10), the longer version also selectively inhibited catecholamine release evoked by nicotinic cholinergic stimulation but not that evoked by membrane depolarization (Fig. 7A). Both peptides and chromogranin A showed concentration-dependent inhibition of nicotine-induced secretion (Fig. 7B): intact chromogranin A had an IC 50 of ϳ4.2 M, chromogranin A 344 -364 had an IC 50 of ϳ0.35 M, and the chromogranin A 332-364 IC 50 was ϳ2.01 M. At high dose (10 M) each peptide completely blocked nicotinic-stimulated secretion, while whole chromogranin A only partially blocked secretion.

Cleavage within the Catestatin Region of Bovine Chromogranin A: Immunoblots of Chromaffin Granules with Antisera
Directed to Flanking Peptides-Chromaffin granule soluble proteins were separated by SDS-PAGE and then immunoblotted with not only an antibody directed against the catestatin region (bovine chromogranin A 344 -364 ) but also antibodies directed against peptide regions that are bounded by dibasic cleavage sites and lie either directly amino-terminal (bovine chromogranin A 316 -329 ) or directly carboxyl-terminal (bovine chromogranin A 367-391 ) to catestatin (Fig. 9). Each antibody recognized not only intact chromogranin A (at a molecular mass of ϳ60 -70 kDa) but also several lower molecular mass chromogranin A fragments, ranging from ϳ10 to 50 kDa. All of the lower molecular mass chromogranin A fragments are prominently recognized by all three antisera, with the exception of a ϳ19-kDa fragment (Fig. 9, arrow), which is visualized by antichromogranin A 316 -329 and anti-chromogranin A 344 -364 , al- FIG. 5. Cleavage of the catestatin region from chromogranin in bovine sympathetic neurons: splenic sympathetic nerve. Bovine splenic nerve large dense core granule fractions were desalted by adsorption to/elution from a C-18 matrix (SepPak), immunoprecipitated by an anti-catestatin (bovine chromogranin A 344 -364 ) antibody, and subjected to MALDI mass spectrometry. MALDI revealed a peak at m/z ϭ 3832 (top), which, in the catestatin region of bovine chromogranin A (Fig. 1), is within 0.125% of 3827.2, the calculated MH ϩ of chromogranin A 332-364 (LEGEEEEEEDPDRSMRLSFRARGYGFRGPGLQL). After oxidation by H 2 O 2 , the m/z ϭ 3832 peak diminished, while an m/z ϭ 3843 peak became more prominent (bottom), a mass shift corresponding to the added 16-dalton oxygen atom in Met 346 -sulfoxide in chromogranin A 332-364 (predicted MH ϩ ϭ 3843.2). though not by anti-chromogranin A 367-391 . Thus, the ϳ19-kDa fragment probably represents the peptide just amino-terminal to a dibasic cleavage at Arg 365 -Arg 366 in bovine chromogranin A.
When the low molecular weight chromaffin granule peptides (Fig. 2, fractions 42-52) were subjected to catestatin immunoblotting, followed by densitometry of the immunoreactive bands, more than half of the catestatin immunoreactivity was found in fractions of lower molecular mass than chromogranin A: 23.7% in a ϳ34-kDa band and 27.1% in a ϳ15 kDa band.
Chromogranin A Processing to Catestatin in Human Pheochromocytoma Chromaffin Granules: Immunoblot-The antihuman catestatin antibody (rabbit anti-human chromogranin A 352-372 ) recognized intact human chromogranin A, at M r ϳ70 kDa (Fig. 10, lane 1) in pheochromocytoma chromaffin granule immunoblots. All pheochromocytomas studied exhibited at least some processing of the catestatin region. Low molecular weight peptides, migrating near the tracking dye and bearing the catestatin epitope, are readily apparent in two pheochromocytomas (Fig. 10, lanes 4 and 5), and the overall pattern of catestatin processing from human chromogranin A was strikingly similar in two of five pheochromocytomas (Fig. 10, lanes  4 and 5). In three of the five tumors, more than 50% of the catestatin immunoreactivity still resided in the intact chromogranin A parent molecule (Fig. 10, lanes 4 -6). Similar processing was noted on catestatin immunoblots of chromaffin granules from eight additional pheochromocytomas (data not shown).
We also tested the calcium-dependence of catestatin secretion (Fig. 11B); in the presence of extracellular calcium (1 mM), membrane depolarization (by 100 mM KCl) stimulated catestatin release by ϳ40-fold (from 31.5 Ϯ 5.6 to 1260 Ϯ 19.4 nmol/ ml), while in the absence of extracellular calcium this stimulation was abolished (back to 18.7 Ϯ 1.0 nmol/ml). Qualitatively similar results were observed for enkephalin secretion (Fig.  11B); in the presence of extracellular calcium, membrane depolarization (by 100 mM KCl) stimulated enkephalin release by ϳ27-fold (from 546 Ϯ 104 to 15,000 Ϯ 3710 nmol/ml), while in the absence of extracellular calcium this stimulation was virtually abolished (back to 626 Ϯ 430 nmol/ml).
Using this radioimmunoassay, we detected catestatin immunoreactivity in fetal bovine serum (at 2.71 nM) and adult equine serum (at 22.2 nM).

DISCUSSION
The catestatin region of chromogranin A (bovine chromogranin A 344 -364 ) is a potent and specific inhibitor of chromaffin cell catecholamine release when triggered by nicotinic cholinergic stimulation, the physiologic pathway for such cells (10); catestatin also antagonizes nicotinic desensitization of the process (27). In these studies, we documented cleavage of the catestatin region in normal chromaffin granules (Figs. 3, 4, 8, and 9) and sympathetic nerve catecholamine storage vesicles (Fig. 5) as well as chromaffin granules from human pheochromocytoma (Figs. 6 and 10). Indeed, cleavage of the catestatin region occurs at high frequency in chromogranin A: ϳ46% of catestatin immunoreactivity in chromaffin granules was of a lower molecular size form than intact chromogranin A (Fig. 8A, peaks   FIG. 7. Catecholamine release-inhibitory activity of synthetic catestatins: specificity of mechanism and potency of effect. Longer (bovine chromogranin A 332-364 ; LEGEEEEEEDPDRSMRLSFRAR-GYGFRGPGLQL) versus shorter (bovine chromogranin A 344 -364 ; RSMRLSFRARGYGFRGPGLQL) regions of bovine chromogranin A were synthesized, representing longer and shorter cleavage site versions of bovine catestatin identified in these studies (Figs. 3 and 4).
Size separation of chromaffin granule peptides (Fig. 2), followed by immunoprecipitation and MALDI mass spectrometry  (Fig. 3), revealed a major catestatin form (bovine chromogranin A 332-364 ) cleaved, as expected, at dibasic sites: KRLEGEEE-EEEDPDRSMRLSFRARGYGFRGPGLQLRR. Recognition of dibasic recognition sites by prohormone convertases is well described in chromogranin A; we (28) and others (29) have shown that chromogranin A is a substrate in vivo for prohormone convertases 1 and 2 as well as furin. Chromogranin B and secretogranin II are also cleaved by prohormone convertases (30). Other proteases recognizing dibasic sites may also be active in chromaffin granules (31). Since proteases recognizing dibasic sites cleave to the carboxyl-terminal side of such sites (32), the lack of the carboxyl-terminal residues Arg 365 -Arg 366 suggests, in addition, carboxypeptidase B (33, 34) removal of the remaining Arg 365 -Arg 366 residues after initial cleavage by the prohormone convertase(s). Further resolution of chromaffin granule peptides on reverse-phase chromatography (Fig. 4), followed by MALDI mass spectrometry and amino-terminal sequencing, revealed two further forms: a long form (EGEEEEEEDPDRSMRLSFRARGYGFRGPGLQL; chromogranin A 333-364 ) lacking the amino-terminal Leu 332 of the form previously identified (Fig. 3) and a shorter form (DRSMRLSFRARGYGFRGPGL; chromogranin A 343-362 ). Radioimmunoassay confirmed the secretion and physiological relevance of catestatin.
Previous studies established several cleavage sites in the catestatin region of bovine chromogranin A. Within bovine chromaffin granules, Metz-Boutigue et al. (37) found peptides with amino termini at chromogranin A 332 (Leu 332 ; i.e. after dibasic site Lys 330 -Arg 331 ), chromogranin A 351 (Arg 351 ), and chromogranin A 354 (Gly 354 ); upon secretion into the extracellular space, cleavage was also detected before Gly 359 . Sigafoos et al. (38) also detected a fragment with the amino terminus chromogranin A 342 (Pro 342 ). Evidence for cleavage at the dibasic site Arg 365 -Arg 366 in chromogranin A has not heretofore been obtained. Such previous studies used amino-terminal amino acid sequencing but occurred before the widespread availability of protein and peptide mass spectrometry; thus, carboxyl-terminal boundaries of the detected peptides could only be deduced imprecisely.
Studies on the formation of peptides flanking catestatin, i.e. bovine chromogranin A 316 -329 (referred to as WE14 (39)) and bovine chromogranin A 367-392 (referred to as GE25 (40)), also provide evidence of cleavage in the catestatin (bovine chromogranin A 344 -364 ) region. Our region-specific immunoblots of the catestatin region ( Fig. 9) indicate that chromogranin A fragments bearing the catestatin epitope are at least as prominent as fragments bearing the WE14 or GE25 epitopes and document dibasic cleavage at Arg 365 -Arg 366 in bovine chromogranin A.
Since peptide fractions characterized in these experiments came from sucrose density gradient-purified chromaffin granules, artifactual proteolysis is unlikely to be problematic here.
We detected catestatin immunoreactivity in bovine and equine serum (see "Results"). Using chromogranin A radioimmunoassays directed to regions overlapping the catestatin portion of human (chromogranin A 344 -374 ; Ref. 46) or rat (chromogranin A 359 -389 ; Ref. 47) chromogranin A, Yanaihara and co-workers (46 -49) also found catestatin region immunoreactivity in the circulation (46,47) as well as in saliva, where its release was triggered by autonomic stimulation (48,49). Since catestatin administration into the bloodstream exerts profound effects upon blood pressure (50), detection of catestatin in serum has implications for control of the circulation.
Thus, the catestatin fragment of chromogranin A is formed by endogenous proteolytic cleavage in vivo. Such authentic cleavage sites provide a useful starting point in analysis of the relationship between structure and function for this potent and specific catecholamine release-inhibitory peptide (10).