Cell-specific processing of preprosomatostatin in cultured neuroendocrine cells

We have previously found that preprosomatostatin is processed accurately to both somatostatin-14 and somatostatin-28 in pituitary gonadotrophs of transgenic mice. The foreign somatostatin peptides have been shown to enter the regulated secretory pathway of these cells. To determine whether accurate preprosomatostatin processing can occur in any neuroendocrine cell, we introduced preprosomatostatin cDNA expression vectors into several different neuroendocrine cell lines. We found that prosomatostatin was cleaved efficiently to somatostatin-14 and somatostatin-28 in RIN 5F and AtT20 cells, but not in GH4 or PC12 cells. The ability of a particular cell type to process prosomatostatin did not correlate with cellular storage capacity and was independent of the level of biosynthesis of the precursor. These data suggest that prosomatostatin processing requires specific pathways which are present in some neuroendocrine cells, but not in others.     

Tissue-specific posttranslational processing of pre-prosomatostatin encoded by a metallothionein-somatostatin fusion gene in transgenic mice

The somatostatins are neuropeptides of 14 and 28 amino acids that inhibit the release of growth hormone and other hypophyseal and gastrointestinal peptides. These neuropeptides are cleaved posttranslationally from a common precursor, pre-prosomatostatin. We report here the production and processing of pre-prosomatostatin by transgenic mice carrying a metallothionein-somatostatin fusion gene. The most active site of somatostatin production, as determined by hormone concentrations in the tissues, is the anterior pituitary, a tissue that does not normally synthesize somatostatin-like peptides. Anterior pituitary processed pre-prosomatostatin almost exclusively to the two biologically active peptides, somatostatin-14 and somatostatin-28, whereas the liver and kidney synthesized much smaller quantities of predominantly a 6000 dalton somatostatin-like peptide. The growth of the transgenic mice was normal despite high plasma levels of the somatostatin-like peptides. These studies indicate that proteases which cleave prosomatostatin to somatostatin-28 and somatostatin-14 are not specific to tissues that normally express somatostatin.     

Somatostatin 28(1-12) in a somatostatin-secreting human medullary thyroid carcinoma cell line

We have identified a system, the TT human medullary thyroid carcinoma cell line, which we found to contain 31.3 +/- 27.7 ng of somatostatin 28(1-12) immunoreactivity/mg protein. Radioimmunoassay of gel filtration fractions showed that the major form of immunoreactive somatostatin 28(1-12) had a molecular weight of 1,500 daltons. During reversed-phase high pressure liquid chromatography, this 1,500-dalton species coeluted with synthetic somatostatin 28(1-12). Somatostatin 28(1-12) containing forms larger than 7,000 daltons were also observed. Further studies will be required to elucidate the route of processing of prosomatostatin. The fact that the products of prosomatostatin processing in these cells are similar to those in normal tissues indicates that the TT medullary thyroid carcinoma cell line constitutes a useful model for human somatostatin gene expression.     

Characterization of three peptides derived from prosomatostatin [prosomatostatin-(1-63)-, -(65-76)- and -(79-92)-peptides] in a human pancreatic tumour.

By using only reverse-phase h.p.l.c., three fragments of prosomatostatin were isolated from an extract of a human pancreatic neuroendocrine tumour that produced somatostatin, vasoactive intestinal polypeptide and gastrin-releasing peptide. The amino acid composition of the peptides indicated that they represented prosomatostatin-(1-63)-peptide, prosomatostain-(65-76)-peptide and prosomatostatin-(79-92)-peptide (somatostatin-14). The identity of prosomatostatin-(1-63)-peptide was confirmed by characterization of the products of digestion with Armillaria mellea (honey fungus) proteinase. Partial micro-sequencing of prosomatostatin-(1-63)-peptide showed that the Gly24-Ala25 bond of preprosomatostatin was the site of cleavage of the signal peptide. Thus human prosomatostatin is a protein of 92 amino acid residues that is proteolytically cleaved in a pancreatic tumour at the site of a dibasic-residue (arginine-lysine) processing site and at a single-monobasic-residue (arginine) processing site.     

Site-specific mutagenesis identifies amino acid residues critical in prohormone processing.

Peptide hormones are generally synthesized as inactive higher mol. wt precursors. Processing of the prohormone into biologically active peptides by specific proteolytic cleavages occurs most often at pairs of basic amino acids but also at single arginine residues. To study the role of protein secondary structure in this process, we used site-directed mutagenesis to modify the predicted secondary structure around the cleavage sites of human prosomatostatin and monitored the processing of the precursor after introduction of the mutated cDNAs in Neuro2A cells. Amino acid substitutions were introduced that affected the possibility of forming beta-turn structures in the immediate vicinity of the somatostatin-28 (S-28) and somatostatin-14 (S-14) cleavage sites. Infection of Neuro2A cells with a retrovirus carrying a human somatostatin cDNA resulted in the expression of prosomatostatin and its processing into S-28 and S-14, indicating that these cells have the necessary enzymes to process prohormone at both single and paired amino acid residues. Disruption of the different beta-turns had various effects on prosomatostatin processing: substitution of Ala for Pro-5 drastically decreased prosomatostatin processing and replacement of Pro-9 by Ala led to the accumulation of the intermediate maturation product [Arg-2Lys-1]-S-14. In contrast, substitution of Ala for Asn-12, Gly+2 and Cys+3 respectively had only very little effect on the proteolytic processing of prosomatostatin. Our results show that amino acids other than the basic amino acid residues are required to define the cleavage sites for prohormone proteolytic processing and suggest that higher orders of protein structure are involved in substrate recognition by the endoproteases.     

Rat pre-prosomatostatin. Structure and processing by microsomal membranes

The tetradecapeptide hormone somatostatin arises from proteolytic processing of a large precursor, pre-prosomatostatin. Studies of other hormone precursors predict that the NH2 terminus of pre-prosomatostatin comprises a leader, or signal, region which is cleaved during its translation. Such co-translational cleavage would generate prosomatostatin. In these studies, we present the complete sequence of rat pre-prosomatostatin, deduced from the nucleotide sequence of cDNAs derived from a somatostatin-rich medullary thyroid carcinoma. These findings indicate that rat pre-prosomatostatin contains 116 amino acids (12,737 daltons). Cell-free translations of medullary thyroid carcinoma mRNA with dog pancreas microsomal membranes were performed to identify the cleavage point of the leader region from prosomatostatin. Partial microsequencing data indicates that the cleavage occurs between the glycine and alanine at positions 24 and 25 of pre-prosomatostatin. Thus, rat prosomatostatin contains 92 amino acids (10,388 daltons). Comparison of the amino acid sequences of the rat and human pre-prosomatostatins reveals only four amino acid substitutions. In view of the high degree of homology between rat and human pre-prosomatostatin, we expect a similar cleavage site and NH2-terminal structure for human prosomatostatin. The high level of conservation between rodents and humans of the entire pre-prosomatostatin molecule further suggests the possibility of biologic functions of the NH2-terminal portions of prosomatostatin.     

Distribution and molecular forms of peptides containing somatostatin immunodeterminants in extracts from the entire gastrointestinal tract of man and pig

Specimens from human porcine mucosal and muscular tissue from the entire gastrointestinal tract were extracted in acid ethanol, subjected to chromatography and analysed for somatostatin-like immunoreactivity by region-specific radioimmunoassays. The concentration of somatostatin-like immunoreactivity from man and pig ranged from 1.13 +/- 0.37 to 101.15 +/- 33.93 pmol/g wet weight, and from 7.64 to 159.48 +/- 23.79 pmol/g wet weight, respectively. In both species the highest concentrations were found in the jejunum. The immunoreactivity in intestinal mucosal extracts was distributed among four major peaks, two of which were identified by HPLC as somatostatin 1-28 and somatostatin 1-14, respectively. A peak of approx. 10 kDa was resolved by ion exchange plus HPLC into three components, two containing at least part of the somatostatin 1-14 sequence as well as (part of) the somatostatin 1-28(1-14) sequence (but differing in charge), the third containing only the 1-28(1-14) sequence. These peptides probably represent uncleaved and partially cleaved prosomatostatin. The fourth component to be identified by gel filtration was slightly larger than somatostatin 1-14. Extracts from the antrum, the pancreas and from muscular tissues contained almost exclusively somatostatin 1-14, and very little somatostatin 1-28, indicating that the somatostatin precursor is processed differently at these sites. Furthermore, extracts of porcine gastric antrum, analysed for somatostatin 1-28(1-14) immunoreactivity, showed two immunoreactive forms in the mucosa and three major forms in the muscular layers.     

Prosomatostatin is processed in the Golgi apparatus of rat neural cells.

Proteolytic processing of somatostatin precursor produces several peptides including somatostatin-14 (S-14), somatostatin-28 (S-28), and somatostatin-28 (1-12) (S-28(1-12)). The subcellular sites at which these cleavages occur were identified by quantitative evaluation of these products in enriched fractions of the biosynthetic secretory apparatus of rat cortical or hypothalamic cells. Each of the major cellular compartments was obtained by discontinuous gradient centrifugation and was characterized both by specific enzyme markers and electron microscopy. The prosomatostatin-derived fragments were measured by radioimmunoassay after chromatographic separation. Two specific antibodies were used, allowing the identification of either S-28(1-12) or S-14 which results from peptide bond hydrolysis at a monobasic (arginine) and a dibasic (Arg-Lys) cleavage site, respectively. These antibodies also revealed prosomatostatin-derived forms containing at their COOH terminus the corresponding dodeca- and tetradecapeptide sequences. Whereas the reticulum-enriched fractions contained the highest levels of prosomatostatin, the proportion of precursor was significantly lower in the Golgi apparatus. In the latter fraction, other processed forms were also present, i.e. S-14 and S-28(1-12) together with the NH2-terminal domain (1-76) of prosomatostatin (pro-S(1-76). Inhibition of the intracellular transport either by monensin or by preincubation at reduced temperature resulted in an increase of prosomatostatin-derived peptides in the Golgi-enriched fractions. Finally, immunogold labeling using antibodies raised against S-28(1-12) and S-14 epitopes revealed the presence of these forms almost exclusively in the Golgi-enriched fraction mainly at the surface of saccules and vesicles. Together these data demonstrate that in rat neural cells, prosomatostatin proteolytic processing at both monobasic and dibasic sites is initiated at the level of the Golgi apparatus.     

Isolation and characterization of S. cerevisiae mutants defective in somatostatin expression: cloning and functional role of a yeast gene encoding an aspartyl protease in precursor processing at monobasic cleavage sites.

The peptide somatostatin exists as two different molecular species. In addition to the most common form, somatostatin-14, there is also a fourteen amino acid N-terminally extended form of the tetradecapeptide, somatostatin-28. Both peptides are synthesized as larger precursors containing paired basic and monobasic amino acids at their processing sites, which upon cleavage generate either somatostatin-14 or -28, respectively. In some species of fish two distinct, but homologous, precursors (prosomatostatin-I and -II) give rise to somatostatin-14 and -28, respectively. Whereas anglerfish prosomatostatin-II was previously shown to release exclusively somatostatin-28, the yeast Saccharomyces cerevisiae proteolytically matures the homologous prosomatostatin-I precursor to somatostatin-28 and -14 as well as to a lysine-extended form of somatostatin-14. The Kex2 endoprotease appears to be essential for the formation of lysine somatostatin-14 and is involved either directly or indirectly in the release of mature somatostatin-14. The isolation of yeast mutants defective in somatostatin-28 expression (sex mutant) allowed the cloning of a non-essential gene, which encodes an aspartyl protease, whose disruption severely affects the cleavage of mature somatostatin-28 from both somatostatin precursors. We conclude that two distinct endoproteases, which demonstrate some cross specificity in vivo, are involved in the proteolytic maturation of prosomatostatin at mono- and dibasic processing sites in yeast.     

Isolation and characterization of proSS1-32, a peptide derived from the N-terminal region of porcine preprosomatostatin

A peptide derived from the N-terminal region of porcine prosomatostatin, proSS1-32, has been purified to homogeneity from extracts of porcine upper intestine. Amino acid analysis revealed that the peptide consists of 32 residues. The complete primary structure was determined as: A P S D P R L R Q F L Q K S L A A A A G K Q E L A K Y F L A E L. This sequence obviously comprises residues 1-32 of porcine prosomatostatin since it is identical to the corresponding sequence in human preprosomatostatin. The postulated cleavage site in porcine prosomatostatin is a Leu-Leu bond between residues 32 and 33, thus confirming previous studies of the processing of the somatostatin precursor in the rat and transgenic mouse.     

Synthetic partial extension peptides of P-450(SCC) and adrenodoxin precursors: effects on the import of mitochondrial enzyme precursors

Various portions of the extension peptides of P-450(SCC) precursor were chemically synthesized. The effects of these peptides on the import of enzyme precursors into mitochondria were examined. Peptides SEP1-15 and SEP1-20, corresponding to the amino terminal portion of the extension peptides, strongly inhibited the import of P-450(SCC) precursor into mitochondria. These peptides were effective at concentrations above 30 microM, and complete inhibition was observed at 100 microM. SEP1-11, which is shorter than SEP1-15 and SEP1-20, showed very weak inhibition. SEP25-39, which corresponds to the carboxy terminal portion of the extension peptide, did not affect the import of the precursor. The import of P-450(11 beta) and adrenodoxin precursors were also inhibited by SEP1-15. Another peptide, AEP1-14, which corresponds to the amino terminal portion of the extension peptide of adrenodoxin precursor, was also synthesized. The peptide inhibited the import of both adrenodoxin and P-450(SCC) precursors into mitochondria. The import of malate dehydrogenase was also inhibited by SEP1-15 and AEP1-14. The rate of the internalization of the precursor into mitochondria was decreased by the peptides. The amount of the precursor bound to the surface of mitochondria and the processing of adrenodoxin precursor were not affected. The respiratory activities of isolated mitochondria were not influenced by SEP1-15 even at 100 microM. We conclude that the inhibitory activities of the synthetic partial extension peptides on the import of enzyme precursors into mitochondria require the presence of about fifteen amino acid residues in the amino terminal portion of the extension peptides, and the inhibition of the import by the peptides was dependent on the blockage of the internalization of the precursors into mitochondria.     

Glucocorticoids stimulate the production of preprocalcitonin-derived secretory peptides by a rat medullary thyroid carcinoma cell line

A rat medullary thyroid carcinoma cell line, CA-77, has been established as a model system for investigating calcitonin biosynthesis and secretion. Growth of this cell line in serum-free defined medium provided suitable conditions for studying steroid hormone effects on the production of calcitonin and related peptides. After exposure for 5 days to a variety of steroids, only dexamethasone and corticosterone increased cellular content of calcitonin and a second secretory peptide (CCAP) derived from the same mRNA translation product as calcitonin. Glucocorticoids had no effect on cellular somatostatin, another secretory product of these cells. Increasing doses of dexamethasone progressively elevated cellular calcitonin and CCAP, with a maximal effect at 10(-8) M; 10(-9) M and lower doses were ineffective. On a molar basis, corticosterone was approximately 50-fold less potent than the synthetic glucocorticoid. An increase in cellular calcitonin content was observed only after 48 h of glucocorticoid treatment; a maximum increase (13-fold) occurred after 7 days. Glucocorticoids also increased basal calcitonin secretion. Similar effects were observed for cellular and secreted CCAP. Withdrawal of dexamethasone after 4 days of treatment lowered cellular calcitonin toward the level of control cultures. Dexamethasone pretreatment potentiated the acute secretory response to calcium for both calcitonin and CCAP, while no such enhancement was noted for calcium stimulation of somatostatin secretion. We conclude that the glucocorticoids specifically stimulate the production and secretion of calcitonin and CCAP, two secretory peptides derived from preprocalcitonin.     

Structural characterization of peptides derived from prosomatostatins I and II isolated from the pancreatic islets of two species of teleostean fish: the daddy sculpin and the flounder

The primary structures of three peptides from extracts from the pancreatic islets of the daddy sculpin (Cottus scorpius) and three analogous peptides from the islets of the flounder (Platichthys flesus), two species of teleostean fish, have been determined by automated Edman degradation. The structures of the flounder peptides were confirmed by fast-atom bombardment mass spectrometry. The peptides show strong homology to residues (49-60), (63-96) and (98-125) of the predicted sequence of preprosomatostatin II from the anglerfish (Lophius americanus). The amino acid sequences of the peptides suggest that, in the sculpin, prosomatostatin II is cleaved at a dibasic amino acid residue processing site (corresponding to Lys61-Arg62 in anglerfish preprosomatostatin II). The resulting fragments are further cleaved at monobasic residue processing sites (corresponding to Arg48 and Arg97 in anglerfish preprosomatostatin II). In the flounder the same dibasic residue processing site is utilised but cleavage at different monobasic sites takes place (corresponding to Arg50 and Arg97 in anglerfish preprosomatostatin II). A peptide identical to mammalian somatostatin-14 was also isolated from the islets of both species and is presumed to represent a cleavage product of prosomatostatin I.     

The complete primary structure of pilin fromHaemophilus influenzae type b strain Eagan

Adherence ofHaemophilus influenzae type b (Hib) to human oropharyngeal cells is mediated by pili which are proteinaceous filaments that extend outward from the bacterial cell surface. Pili from Hib strain Eagan were purified, and the primary structure of the major subunit, pilin, was determined. Sequencing of overlapping peptides showed the mature protein to be comprised of 196 amino acids and to have an M_r of 21,152. The amino terminal sequence was found to be homologous with the sequence previously reported for Hib strain M43 and also to have significant homology to pilins of other gram-negative pathogenic bacteria. Furthermore, Hib pilin had two cysteinyl residues in the amino terminal portion of the protein which were separated by 40 residues (positions 21 and 61); a motif found in other bacterial pilins. The data show that Hib pilin has structural features common to other bacterial pilins.     

Post-translational processing of preprosomatostatin-II examined using fast atom bombardment mass spectrometry

The products and an intermediate of preprosomatostatin-II processing in the anglerfish islet were purified and subjected to structural analysis. The peptides isolated identify the site of signal cleavage (between Ser-24 and Gln-25). The prohormone is further processed at Arg-97 and, to a lesser extent, at the two adjacent basic amino acid residues Lys-61 and Arg-62. A 28-residue somatostatin is also generated which can be hydroxylated at Lys-23. A proteolytic processing site which would form the 14-residue somatostatin does not appear to be used to a significant degree. Fast atom bombardment mass spectrometry (FABMS) was used to demonstrate that the amino-terminal residues of peptides 25-60, and 25-90 are pyroglutamic acid, a modification which precludes Edman degradation of these peptides. Analysis of the peptides and tryptic peptide maps by FABMS allowed confirmation of the sites of prohormone conversion and indicated that terminal basic residues were removed during processing. Three amino acid residues were also found to differ from the amino acid sequence deduced from the cDNA and were localized to specific regions by FABMS analysis. Residues found to differ from the cDNA (cDNA in parentheses) were: Asp-77 (Thr), Val-78 (Phe), and Gly-90 (Glu). Mass assignments were confirmed by running a single cycle of Edman degradation prior to FABMS. The peptides noted above were also examined by Edman sequence analysis. The sequence of a cDNA clone to preprosomatostatin-II was re-examined in light of the observed differences at the protein level. This study emphasizes the utility of FABMS in prohormone processing studies and in identification of post-translational processing events.     

Identification of new epitopes recognized by human monoclonal antibodies with neutralizing and antibody-dependent cellular cytotoxicity activities specific for human T cell leukemia virus type 1.

We have generated a number of EBV-transformed B cell lines producing human mAb against human T cell leukemia virus type 1 (HTLV-1) from the peripheral blood B lymphocytes obtained from patients with HTLV-1-associated myelopathy/tropical spastic paraparesis. Various synthetic peptides corresponding to antigenic regions of HTLV-1 gag and env proteins were used for the screening of antibodies in ELISA. In our study, four IgG mAb to the gag p19 amino acids 100 to 130, and 5 IgG mAb to the env p46 amino acids 175 to 199 were characterized. An immunofluorescence assay showed that all of these mAb specifically bound to the surface of HTLV-1-bearing cell lines. Among these mAb, one anti-gp46 mAb, designated KE36-11, neutralized the infectivity of HTLV-1 as determined by both the inhibition of HTLV-1-induced syncytium formation and transformation assays in vitro. An antibody-binding assay using overlapping oligopeptides revealed that KE36-11 recognized a new epitope locating between the gp46 amino acid sequence 187-193 (Ala-Pro-Pro-Leu-Leu-Pro-His). Another anti-gp46 mAb, designated KE36-7, showed antibody-dependent cellular cytotoxicity against HTLV-1-bearing cell line. KE36-7 bound strongly to the 10-mer peptide-gp46 187-196, and weakly to peptides containing the gp46 amino acid sequence 191-196 (Leu-Pro-His-Ser-Asn-Leu). These two epitopes, which are associated with HTLV-1 neutralization and antibody-dependent cellular cytotoxicity, are thus the first epitopes identified in human HTLV-1 infection. It is possible that passive immunization of humans with these two human mAb are effective on the protection of HTLV-1 infection in vivo.     

Isolation and characterization of copper-binding sites of human ceruloplasmin

A tryptic cleavage procedure was used to obtain stable copper-containing peptide regions of human ceruloplasmin. Tryptic digestion-derived copper peptides were fractionated by gel filtration chromatography, yielding two fractions, one with an apparent molecular weight of 11000 and the other 1000. The high molecular weight fraction (11K fraction) was found to contain 50% of the copper atoms of the ceruloplasmin molecule and behaved as a single copper-containing component by gel filtration chromatography and by isoelectric focusing. The low molecular weight fraction (1K fraction) was found to be a mixture of three or four copper peptides by isoelectric focusing. Acrylamide gel electrophoresis studies, amino acid composition analysis and terminal amino acid determinations showed the 11K fraction to be a complex composed of at least three peptides arising from different regions of the ceruloplasmin molecule. Two of the peptides of the 11K complex appear to be derived from the 19K fragment of the ceruloplasmin molecule (18); one peptide in the complex appears to correspond to the aspartic acid-rich portion, residues 7-30, and the other to the histidine-rich portion, residues 103-157. Most preparations of ceruloplasmin are reported to consist of three non-covalently linked fragments that have molecular weights of 67K, 50K and 19K. Dwulet and Putnam (20) proposed a model for the sequence structure of ceruloplasmin where the molecule exhibits a three-fold repeat pattern of two alternating structures, here termed X and Y.(ABSTRACT TRUNCATED AT 250 WORDS)     

Fibrous Aggregates of Short Peptides Containing Two Distinct Aromatic Amino Acid Residues

The aim of the study was the assessment of the ability of short peptides to form aggregates under physiological conditions. The dipeptides studied were derived from different aromatic amino acids (heteroaromatic peptides). Tripeptides were obtained from two distinct aromatic amino acids and cysteine or methionine residue in the C‐terminal, N‐terminal, or central position. The ability of the peptides to form fibrous aggregates under physiological conditions was evaluated using three independent methods: the Congo Red assay, the Thioflavin T assay, and microscopic examinations using normal and polarized light. Materials potentially useful for regenerative medicine were selected based on their cytotoxicity to the endothelial cell line EA.hy 926 and physicochemical properties of films formed by peptides. The required parameters of biocompatibility were fulfilled by H?PheCysTrp?OH, H?PheCysTyr?OH, H?PheTyrMet?OH, and H?TrpTyr?OH.     

Expression of porcine cholecystokinin cDNA in a murine neuroendocrine cell line. Proteolytic processing, sulfation, and regulated secretion of cholecystokinin peptides

The cDNA for porcine preprocholecystokinin (pre-pro-CCK) was engineered for expression in mammalian cells under the control of the Rous sarcoma virus-long terminal repeat promoter. This expression construct was transfected into the murine anterior pituitary cell line, AtT-20. A stable cell line (AtT-20/CCK) was derived that expresses CCK mRNA indistinguishable from the CCK mRNA found in pig brain or gut. The AtT-20/CCK cells carry out proteolytic processing and sulfation reactions to generate authentic sulfated CCK8 from pro-CCK. The cells also store and secrete CCK-immunoreactive peptides. This secretion can be stimulated with corticotropin releasing factor, the natural secretagogue for anterior pituitary cells. In contrast, monkey kidney epithelial cells (COS cells), which are transiently transfected to express CCK, predominantly secrete nonsulfated pro-CCK into the medium. These studies show that a murine neuroendocrine cell line contains the complete processing machinery required to generate authentic porcine CCK8. The processing events include simultaneous proteolytic processing at one and two basic amino acid sites and sulfation of tyrosine residues. The cell line thus duplicates exactly the processing patterns found to occur in pig brain cortex.     

Characterization of an endoprotease from rat small intestinal mucosal secretory granules which generates somatostatin-28 from prosomatostatin by cleavage after a single arginine residue

We have extracted, characterized, and partially purified an enzyme from secretory granules from rat small intestinal mucosa which cleaves a synthetic prosomatostatin substrate on the carboxyl side of a single arginine residue. This substrate Leu-Gln-Arg-Ser-Ala-Asn-Ser-NH2 contains the monobasic site at which mammalian prosomatostatin is cleaved in vivo to generate somatostatin-28. This activity was released from the granules by osmotic shock followed by extraction with 500 mM KCl. The enzyme had a molecular weight of about 55,000, a pH optimum of about 7.5, and a Km for the synthetic substrate of 20 microM. It was partially inhibited by diisopropyl fluorophosphate, phenylmethanesulfonyl fluoride, iodoacetate, soybean trypsin inhibitor, and EDTA. It was also very sensitive to aprotinin (complete inhibition at 25 micrograms/ml) but was not inhibited by bestatin, pepstatin, or p-chloromercuribenzoate. This endoprotease was unable to cleave three small trypsin and kallikrein substrates (N alpha-benzoyl-L-arginine ethyl ester, N alpha-benzoyl-DL-arginine p-nitroanilide, and N alpha-benzoyl-L-arginine 7-amido-4-methylcoumarin). It was unable to cleave either the Arg-Asp bond in CCK 12 or the Arg-Glu and Arg-Met bonds of synthetic peptides corresponding to sequences of anglerfish prosomatostatin II situated upstream from the somatostatin-28 domain. These observations together suggest that adjacent amino acids play a role in determining the conformational specificity of the monobasic cleavage. This soluble enzyme was also able to cleave three synthetic substrates containing dibasic residues (Arg-Lys or Lys-Arg) on the carboxyl side of the arginine, although it did so less rapidly than at the monobasic cleavage sites. When incubated with partially purified prosomatostatin from anglerfish pancreas, significant quantities of somatostatin-28 II were produced. All these cleavages were completely blocked by preincubation with aprotinin. Although further work is required to clarify the physiological role of this enzyme, it appears, in view of its catalytic properties, this endoprotease could be involved in the conversion of prosomatostatin to somatostatin-28 in intestine mucosal secretory cells.