THE OF No. 8, Issue March 1993 Printed in U.S.A ...€¦ · tially at monobasic sites to SS-28 (40-43%) and antrin (27-29%). Processing, however, was inefficient as suggested by large

THE JOURNAL OF BIOLOGICAL CHEMISTRY @ 1993 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 268, No. 8, Issue of March 15, pp. 6041-6049, 1993 Printed in U.S.A.

Heterologous Processing of Prosomatostatin in Constitutive and Regulated Secretory Pathways PUTATIVE ROLE OF THE ENDOPROTEASES FURIN, PC1, AND PC2*

(Received for publication, September 18,1992)

Aristea S. Galanopoulou$, Gillian Kent, Shahida N. Rabbani, Nabil G. Seidah$, and Yogesh C. Pate17 From the Fraser Laboratories, McGill Uniuersity, Departments of Medicine and Neurology and Neurosurgery, Royal Victoria Hospital and Montreal Neurological Institute, Montreal, Quebec H3A 1Al and the $Clinical Research Institute of Montreal, Montreal, Quebec H2W lR7, Canada

Mammalian prosomatostatin (PSS) is cleaved at a dibasic Arg-Lys site to produce somatostatin-14 (SS- 14) and at monobasic Arg and Lys sites to yield SS-28 and PSS(l-lo, (antrin), respectively. Furin, PC1, and PC2 are three recently discovered mammalian endoproteases localized either to the constitutive (furin) or regulated (PC1, PC2) secretory pathways. In this study we have compared the heterologous processing of PSS in transiently transfected endocrine (AtT-20 pituitary) and nonendocrine (COS-7 monkey kidney, PC12 pheochromocytoma) tumor cells. We have correlated the efficiency of processing of PSS to SS-14, SS-28, and PSS+lo, with (i) secretion through the constitutive or regulated pathways; (ii) endogenous expression of mRNA for furin, PC1, and PC2; and (iii) exogenous expression of PC1 and PC2 in cells that do not contain these enzymes in order to delineate the putative role of these enzymes in mediating PSS cleavage at dibasic and monobasic sites and to localize the proteolytic events to specific compartments of the secretory pathways. COS-7 and PC12 cells expressed only furin, secreted constitutively, and processed PSS preferentially at monobasic sites to SS-28 (40-43%) and antrin (27-29%). Processing, however, was inefficient as suggested by large amounts of unprocessed PSS. In contrast, AtT-20 cells showed regulated secretion, expressed all three endoproteases (with high levels of PCl ) , and processed PSS efficiently to mainly SS-14. P C l , but not PC2, exogenously coexpressed with PSS in COS-7 cells produced significant conversion to SS- 14 but not 89-28. This study shows that PSS is capable of monobasic cleavage in the constitutive secretory pathway. Such processing could be mediated by a furin-like enzyme but is relatively inefficient. PC1 can effect dibasic cleavage of PSS whereas PC2 is without influence on PSS processing at least within the constitutive secretory pathway. Although monobasic and dl- basic processing of PSS in COS-7 cells correlates with furin-like and PC1 activity, respectively, the relative inefficiency of such processing suggests that compartmentalization of proteolytic events in secretory vesi-

* This work was supported by Grant MT-6196 from the Canadian Medical Research Council. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ Recipient of studentship support from the Fonds Pour La For- mation de Chercheurs et l’Aide a la Recherche.

7 To whom correspondence should be addressed Rm. M3-15, Royal Victoria Hospital, 687 Pine Ave. West, Montreal, Quebec H3A 1A1 Canada. Tel.: 514-842-1231, ext. 5042; Fax: 514-849-3681.

cles or other more specific endoproteases may be re- quired.

The maturation of secretory proteins to active peptides involves endoproteolytic cleavage of precursor (prohormone) forms at dibasic or single basic amino acid sites (1, 2). In the case of mammalian prosomatostatin (PSS),’ three such cleavage sites have been identified (3-5). Processing occurs prin- cipally at the COOH-terminal segment of the molecule and generates the two bioactive forms somatostatin-14 (SS-14) and somatostatin-28 (SS-28) (3-9). Cleavage at an Arg-Lys doublet yields SS-14 and a peptide of molecular mass 8 kDa (8-lo), whereas endoproteolysis at a monobasic Arg site generates SS-28 and a 7-kDa peptide (Fig. 1) (8, 9, 11, 12). In addition, a second monobasic cleavage site at a Lys residue at the NH2-terminal segment has been recently identified which generates the decapeptide PSS(l-,O, (antrin) without any known biological activity (4, 11). The amino acid sequences at these three cleavage sites have been totally conserved throughout vertebrate evolution (12-14).

Little is currently known about two of the key steps in- volved in PSS maturation, namely the role of the prohormone- converting enzymes and the subcellular compartmentalization of processing events. Secretory proteins such as SS are syn- thesized as precursors on ribosomes, translocated into the lumen of the endoplasmic reticulum, and transported by bud- ding non-clathrin-coated vesicles through the Golgi stacks to the trans Golgi network (15, 16). Here the protein is sorted via clathrin-coated vesicles into a regulated compartment consisting of secretory granules or into a constitutive, non- regulated pathway through non-clathrin-coated vesicles which exit from the Golgi and migrate to the plasma membrane (15-17). Although processing of endogenous and heterologous prohormones at basic residues can occur in the constitutive pathway (18,19), efficient processing is generally believed to require targeting to the secretory pathway (18,19- 24). Three mammalian proprotein-converting enzymes furin, PC1, and PC2 have recently been identified (25-32). These are subtilisin-like serine endoproteases, of which furin is the only member with a membrane-spanning domain (32). Furin exhibits an ubiquitous distribution in all tissues and cells, subcellular localization to the Golgi compartment, a multibasic cleavage motif, and may subserve a functional role in processing protein precursors channeled through the constitutive pathway (18, 20, 26-28, 32, 33). By contrast, PC1 and

’ The abbreviations used are: PSS, prosomatostatin; DMEM, Dul- becco’s modified Eagle’s medium; FBS, fetal bovine serum; HPLC, high performance liquid chromatography; kb, kilobase(s).

~ ~~

604 1

6042

FIG. 1. Schematic depiction of mammalian PSS, its dibasic and monobasic processing sites, and known cleavage products.

Prosomatostatin Processing by PC1, PCZ, and Furin

PSS 1 13 64 TI-76 92

1 76

I 1 63

0 1 10

PC2 are not membrane inserted and appear to be localized in secretory granules to which they may be selectively targeted (30). They have been shown to be preferentially expressed in neurons and endocrine cells and to process hormone precursors such as proopiomelanocortin and prorenin selectively at Lys-Arg and Arg-Arg type dibasic sites (18, 23, 27-29,34,35). The role of any of these enzymes in processing mammalian PSS has not been previously investigated. In particular, it is not known which enzyme, if any, processes at single basic residues or at the Arg-Lys dibasic motif, the least common of the paired basic amino acid combinations which occurs in PSS.

In the present study, we have compared the heterologous processing of PSS in both endocrine and non-endocrine tumor cells using cDNA transfection experiments. We have correlated processing at dibasic and monobasic sites with (i) endogenous expression of mRNA for furin, PC1, and PC2, (ii) exogenous expression of PC1 and PC2 in cells that do not contain these enzymes, and (iii) secretion through the constitutive or regulated pathways, in an attempt to elucidate the putative role of these enzymes in mediating PSS cleavage to SS-14, SS-28, and PSS(l_lo, and to localize the proteolytic events to specific compartments of the secretory pathways.

. MATERIALS AND METHODS

Reagents Synthetic peptides were obtained as follows: SS-14 (Ayerst Labo-

ratories, Montreal, Quebec), SS-28 (Hukabel Scientific, Montreal, Quebec), SS-28(1.12), [TyrISS-14, and [Tyr]SS-28(1-14) (Bachem Fine Chemicals, Torrance, CA), PSS(l-lo) (R. Benoit, Montreal, Quebec), [Tyrlo]PSS(l_lo) (Biomega, Montreal, Quebec), PCl(%-IW) (I. Lindberg, New Orleans, LA). [Tyr]PCl(%-Oe) was prepared by solid-phase syn- thesis. Acetonitrile (CH&N) and trifluoroacetic acid were purchased from Fisher Scientific (Montreal, Quebec). Pepstatin A, phenylmethylsulfonyl fluoride, heptafluorobutyric acid, and forskolin were obtained from Sigma. Dulbecco's modified Eagle's medium (DMEM) and fetal bovine serum were purchased from GIBCO. Ser-X-tend was obtained from Irvine Scientific (Santa Anna, CA). All other reagents were of analytical grade.

Cell Culture COS-7 monkey kidney epithelial cells were obtained from the

National Research Council Biotechnology Research Laboratories,

65 92

111 65 92

8kDa

7kDa

SS-28

U ss-14 19 92

ss-28 [1-121

Montreal (Dr. D. Banville) and cultured in DMEM supplemented with 5% fetal bovine serum (FBS). PC12 rat pheochromocytoma cells were obtained through courtesy of Dr. J. A. Wagner, Dana Farber Cancer Institute, Boston and cultured in DMEM supplemented with 10% FBS, 5% horse serum, and 0.8% glucose. AtT-20/D16 V mouse anterior pituitary cells were cultured in DMEM with 5% FBS supplemented with Ser-X-tend (36). RIN5f cells were provided by Dr. P. Poussier, McGill University, and grown in DMEM with 5% FBS.

Expression Vectors The pre-PSS cDNA expression vector, pKS5, was provided by Dr.

K. Sevarino, Yale University (33). The construct contains the cytomegalovirus promoter, fused to the cDNA sequence of pre-PSS including the 3' translational termination and polyadenylation signals. Murine PC1 and PC2 (mPC1 and mPC2) and rat furin cDNAs were cloned into similar viral promoter-driven expression vectors (27, 28). These consisted of pRC/cytomegalovirus virus (Invitrogen) for mPCl and mPC2 and pSP72/SV40 (Promega) for rat furin.

Northern Blot Analysis

Total RNA was extracted from cultured cells by the acid guanidine- thiocyanate-phenol chloroform method (37). [a-3ZPP]UTP-Labeled cRNA probes were generated by use of RNA transcription kits (Promega) according to the manufacturer's specifications. pSP72 containing the rat furin insert was linearized with Hind111 and transcribed with T7 RNA polymerase. Plasmids containing mPCl and mPC2 were linearized with BglII and transcribed with SP6 RNA polymerase. Probes of 1244 bases (rat furin), 2520 bases (mPCl), and 2232 bases (mPC2) were generated and were purified by standard techniques prior to use. 20-pg samples of total RNA were fractionated by electrophoresis on 1.5% agarose/formaldehyde gels, transferred to Nytran nylon membranes (Schleicher and Schull) using the Vacu- Gene vacuum blotting system (LKB-Pharmacia), and hybridized to [c~-~*P]UTP-labeled cRNA probes for rat furin, mPC1, and mPC2 at 65 "C for 24 h in 50% formamide, followed by washing in high stringency salt conditions. Autoradiograms were prepared by exposing the membranes to Kodak XAR-5 film at -80 "C for 2-4 days using intensifying screens.

Transfections

Tumor cells were plated in 100-mm diameter plastic Petri dishes and grown as monolayers. On the third to fifth day of culture when they had reached 50-70% confluence, cells were transiently transfected with 20 pg of plasmid DNA combined with 20 pg of salmon sperm DNA by the calcium phosphate precipitation method (in 1 X HBS) followed by glycerol shock (15% glycerol in 1 X HBS) for 3

Prosomatostatin Processing by PCl, PC2, and Furin 6043

min. They were then allowed to grow in normal incubation media for 3 days after which they were prepared for studies of basal and stimulated secretion of immunoreactive SS (SSLI) as follows. After removal of the feeding media, separate groups of culture dishes were incubated with secretion medium (DMEM-l% bovine serum albumin together with the protease inhibitors phenylmethylsulfonyl fluoride and pepstatin (20 pg/ml each) with or without forskolin 20 pM for 4 h. Media were then harvested, centrifuged at 1,000 X g for 6 min to remove detached cells and the supernatant acidified to pH 4.8 with 1 M acetic acid and stored at -20 "C. Cells were extracted by scraping in 1 M acetic acid containing phenylmethylsulfonyl fluoride and pepstatin A (20 pg/ml each) at 0 "C (5). The cell suspension was further extracted by sonication followed by centrifugation at 5,000 X g for 30 min. The supernatant was stored at -20 'C pending analysis of SSLI.

HPLC Pooled acidified secretion media and cell extracts were diluted 1:7

with 0.1% trifluoroacetic acid and concentrated using Waters Sep- Pak C,, cartridges. The adsorbed peptides were eluted with 80% CH3CN, 0.1% trifluoroacetic acid. The eluate was analyzed by reverse- phase HPLC on a Waters HPLC apparatus consisting of two solvent delivery systems (M45 and M6000 pumps), M680 automated gradient controller, a C,, pBondapak reverse-phase column, and a Water's model 481 LC variable wave length spectrophotometer (9, 11). The column was eluted at room temperature (21 "C) at 1 ml/min with 12- 55% CH,CN, 0.2% heptafluorobutyric acid gradient over 150 min. Column effluent was monitored for UV absorbance at 230 nm. Frac- tions were collected into borosilicate glass tubes (12 X 75-mm, Fisher) using an LKB Ultravac 7000 fraction collector, spiked with 10 g1 of 10% BSA, and stored at -20 "C until further use. 100-500-p1 aliquots from each fraction were rotary evaporated and assayed for SS immunoreactive peptides by separate region-specific radioimmunoassays. Where necessary, immunoreactive peaks were further char- acterized by gel permeation HPLC for estimation of molecular mass (11).

Somatostatin Radioimmunoassays Two separate radioimmunoassays directed toward the COOH- and

NHz-terminal regions of pro-SS were employed as follows. Radioimmunoassay for SS-14-like Immunoreactivity (SS-14 LZ)-

SS-14 LI was measured with rabbit antibody R149 directed against the central segment of SS-14, I2'I-Tyr SS-14 radioligand, and SS-14 standards (7, 9, 11). This assay detects SS-14 and molecular forms extended at the amino terminus of SS-14, including SS-28 and pro- SS. The minimum detectable dose was 1 pg of SS-14.

Radioimmunoassay for PSS(l-lo, LZ-PSS,l.lol LI was measured using antibody R203, 'z51-Tyr'o PSS,l.lo) ligand, and PSS(l-lo) standards (11). This assay detects PSS(l.lo) and its COOH terminally extended forms including PSS, PSS(1-76) (8-kDa peptide), and PSS(14Bj (7-kDa peptide). The minimum detectable dose was 2 pg of PSS(l-loj.

Radioimmunoassay for PC1 -like Zmmunoreactiuity-A radioimmunoassay for amino segment PC1-like immunoreactivity (PC1 LI) was established using a rabbit antibody (gift of I. Lindberg, New Orleans, LA) against a carbodiimide conjugate of synthetic PC1(&L-loo) coupled to keyhole limpet hemocyanin, "'I-Tyr PCl(sz-saj radioligand, and P C 1 ( ~ - l ~ ) standards. The minimum detectable dose was 51.9 pg of standard. Serial dilution of samples from cells known to express PC1 (AtT-20 cells or secretion media) exhibited parallelism with the standard curve confirming immunological identity between the synthetic peptide standard and the endogenous enzyme.

Zmmunoblots Immunoblot analysis of PC2 was carried out on samples of boiled

cell extracts. 45 pg of protein was resolved by sodium dodecyl sulfate- polyacrylamide gel electrophoresis and immunoblotted on Immobi- lon-P transfer membranes (Millipore) according to the manufacturer's instructions using a rabbit COOH-terminal PC2 antibody (1:1000 dilution) directed against the last 12 COOH-terminal amino acids of PC2 (gift of C. J. Rhodes, Boston, MA).

RESULTS

Secretion Studies Fig. 2 depicts basal and forskolin-stimulated secretion of

SS-14 LI from transfected COS-7, PC12, and AtT-20 cells.

The total amount of secretion during the 4-h incubation periods is compared with the total content of SS-14 LI in extracted cells. COS-7 cells secreted more than they stored and showed no response to forskolin stimulation. This secretory profile is consistent with the well known properties of these cells as a constitutively secreting tumor cell line (14). PC12 cells also exhibited relatively high basal secretion (-20% of cell content/4 h), unresponsiveness to forskolin challenge, and similarly behaved as a constitutive secretor of SSLI. By contrast, AtT-20 cells released ~ 1 0 % of the total ceIIuIar content under basal conditions, showed a doubling of secretion with forskolin stimulation and thus displayed char- acteristics of a regulated secretory cell line.

Processing of PSS to SS-14 and SS-28 in COS-7 Cells, PC12, and AtT-20 Cells

Fig. 3 illustrates HPLC profiles of SS-14-like immunoreactive forms in cell extracts and media from transfected COS-7 cells. Five peaks were observed of which the first two coeluted with synthetic SS-14 (retention time 55 min), and Ss-28 (retention time 65 min), respectively. These two peaks ac- counted for 2% (SS-14) and 40% (SS-28) of total cellular SS- 14 LI (Fig. 3B). The remaining immunoreactivity was distrib- uted between the three additional peaks with retention times of 75 min (43% of total SS-14 LI), 97 min (10% of SS-14 LI), and 110 min (5% of total SS-14 LI). The peak with retention time of 110 min exhibited PSS~l_lo, immunoreactivity and thus corresponded to the full-length PSS molecule (Mr = 10,400). The peaks with retention times of 75 and 97 min represented amino terminally truncated forms of pro-SS. All five peaks found in cell extracts were also released into the culture medium (Fig. 3 C ) in the following proportions (percent of total SS-14 LI): SS-14 (7%), SS-28 (lo%), PSS (8%), peak with retention time 75 min (53%), peak with retention time 97 min (22%). The profile of secreted immunoreactivity was comparable to that in cell extracts except for ss-28 which appeared to be poorly released.

PSS was processed in PC12 cells in an analogous manner to that observed in COS-7 cells. Both SS-14 and Ss-28 peaks were identified in cell extracts and in media but with a considerably higher proportion of SS-28 compared to SS-14: SS-14/SS-28, 4:40% in cells and 3:50% in medium. The remaining peaks comprised unprocessed PSS (both full-length PSS and amino terminally truncated forms of PSS) (data not shown).

In contrast to COS-7 and PC12 cells, AtT-20 cells not only targeted PSS-derived peptides to the regulated secretory pathway (Fig. 2) but also processed PSS efficiently to both SS-14 and SS-28 (75 and 22% of SS-14 LI, respectively, in cell extracts) (Fig. 4A 1. Unprocessed PSS was represented by the full-length PSS molecule only (3% of total cellular SS-14 LI), with complete absence of the two amino terminally truncated forms of PSS observed in COS-7 and PC12 cells. PSS-transfected AtT-20 cells released very small quantities of SSLI under basal conditions which chromatographed as a single HPLC peak coeluting with Ss-14 (Fig. 4B).

Amino-terminal Processing of PSS in COS-7, PC12, and AtT-20 Cells

In order to characterize PSS processing to PSS(l-lo, through cleavage at the second monobasic site at the amino-terminal segment, HPLC eluates of extracts from transfected COS-7, PC12, and AtT-20 cells were analyzed by radioimmunoassay using the amino-terminal R203 antibody. Data obtained from COS-7 and AtT-20 cells are shown in Fig. 5; amino-terminal processing of PSS in PC12 cells was comparable to that in

6044

n

c, aJ N e II

M e W

500

400

300

200

100

0

Prosomatostatin Processing by PCl, PC2, and Furin

CM CM CM CM CM CM

control forskolin control forskolin control forskolin

COS-7 PC1 2 AtT-20 FIG. 2. Comparison of basal and forskolin-stimulated secretion of 88-14 LI from transfected COS-7, PC12, and AtT-20

cells. Transfected cells were incubated for 4 h with control media or media containing 20 PM forskolin. At the end of the incubation, cell extracts (C) and media (M) were separately assayed for SS-14 LI. Cellular immunoreactivity was unaffected by forskolin treatment of any of the tumor cells. COS-7 and PC12 cells exhibited a high level of basal secretion and poor stimulation with forskolin (constitutive secretion). AtT-20 cells secreted 4 0 % of the cellular SS-14 LI basally and responded to forskolin stimulation (regulated secretory cell line). Mean & S.E. (n = 5). *p < 0.01 uersus control.

SS28 4001 1

PSS 1

6.

FIG. 3. HPLC profiles of SS-14 LI in cell extracts (panel B ) and secretion media (panel C) of COS-7 cells transfected with PSS. A Cls reverse- phase column was eluted with 12%-55% CH&N 0.2% heptafluorobutyric acid at 21 "C over 150 min. Column effluent was analyzed by radioimmunoassay for SS- 14 LI. The upper p a n e l ( A ) shows the elution position of synthetic PSS(l-lo,, SS-14, and SS-28 detected by absorbance at 230 nm. Representative of three experiments.

3

3; d

400-

v,

3 0 0 - S5-14 SS28

2 0 0 -

1 0 0 -

04 1

20 40 80 80 100 120

FRACTION NUMBER

PSS 1

C .

FIG. 4. HPLC profiles of SS-14 LI in cell extracts (panel A) and secretion media (panel B ) from AtT-20 cells transfected with PSS. Column conditions and markers are as for Fig. 3. Representative of four experiments. Note the 10-fold difference in the ordi- nates for cellular and secreted SS-14 LI.

Prosomatostatin Processing by PCl, PC2, and Furin

h

8 W

f

100

n

z 8 1

W

0

SS14 1

SS28 1

6045

A.

PSS 1

20 40 00 00 la0 120

FRACTION NUMBER

$" (1 -1 01 100 1

lo 1 7 11-101

E. I I 1.5kDa 2.5kDa 7kDa 8kDa PSS

FIG. 5. HPLC profiles of PSS(l-lo, LI in cell extracts from COS-7 cells (panel A) and AtT-20 cells (panel B ) transfected with PSS. Column conditions are as for Fig. 3. The molecular mass of the immunoreactive peaks represented as 1.5, 2.8, 7, and 8 kDa was determined by gel permeation chromatography as described (11). Repreaenta- tive of two separate experiments for each cell line.

20 lo0 120

FRACTION NUMBER

COS-7 cells (data not shown). A l-kDa peak coeluting with (PSS(1-76)), and 10.4 kDa (PSS) as previously described (11). PSS(l-lo, was the dominant processed product in all three Of these, the 7- and 2.5-kDa peptides were prominent prod- cells, accounting for 27, 29, and 43% of PSS(l-lo, LI in COS-7 ucts (16 and 14% of total immunoreactivity, respectively) in PC12 and AtT-20 cells, respectively. Several other HPLC COS7 and PC12 cells but comprised minor forms only (<4%) peaks were obtained and were shown to correspond to PSS(l-lo,- in AtT-20 cells. By contrast, the 8-kDa peptide was an abun- like immunoreactive peptides of 1.5, 2.5, 3.5, 7 (PSS(1-64J, 8 dant peak in AtT-20 cells (18% of total immunoreactivity)

6046 Prosomatostatin Processing by PCl, PC2, and Furin

but occurred in very low quantities in COS-7 and PC12 cells (<5%). In addition, the SS-28(1-121 peptide analyzed by specific radioimmunoassay (9) was a major product in AtT-20 cells but occurred in negligible amounts in COS-7 and PC12 cells (data not shown).

Expression of Furin, PC1, and PC2 mRNA in COS-7, PC12, and AtT-20 Cells-To correlate the pattern of processing of PSS in tumor cells with the expression of endoproteases, the relative abundance of mRNA for furin, PC1, and PC2 in the COS-7, PC12, and AtT-20 cells was estimated by Northern analysis (Fig. 6). A rat islet insulin-producing secretory cell line (RIN5f) was included for further comparison. The furin cRNA probe hybridized to a 4.5-kb transcript found in all four tumor cell lines but with a high level of expression in PC12 and RIN cells. In the case of PC1, two transcripts of 5 and 3 kb were observed in AtT-20 (27), and RIN cells but no transcripts were observed in COS-7 or PC12 cells. AtT-20 cells expressed very high levels of PC1 mRNA and a relatively low level in RIN5f cells. Northern blots indicated that PC2 hybridized predominantly to a 2.8-kb transcript expressed in AtT-20 and RIN cells, the same cells that also expressed PC1 mRNA. In contrast to PC1 mRNA, however, the relative abundance of PC2 was greater in RIN5f cells compared to AtT-20 cells.

Coexpression of PC1 or PC2 with PSS in COS-7 Cell$- Since furin was the only one of the three known endoproteases expressed in COS-7 cells, the effect of PC1 and PC2 cDNA on PSS processing was determined in separate cotransfection experiments with expression vectors for PSS and either PC1 or PC2. The level of expression of the two exogenously introduced endoprotease genes was determined by Northern analysis of transfected cells for PC1 and PC2 mRNA. COS-7 cells transfected with mPCl or mPC2 showed strong hybridization signals for PC1 mRNA (3.0 kb) and PC2 mRNA (2.8 kb) (Fig. 7). Immunoblot analysis of cell extracts with PC2 antibody revealed a single protein band of 75-kDa molecular mass in PC2 + PSS-cotransfected cells but not in PSS- or PC1- transfected cells (data not shown). The size of this protein corresponds to that of the precursor form of PC2 (35). The level of expression of PC1 was determined by radioimmunoassay analysis of cell extracts and secretion media and compared with endogenous PC1 expression in AtT-20 cells. PC1- transfected COS-7 cells showed undetectable levels of PC1 LI in cell extracts but secreted 272 f 42 pg/dish during the post- transfection period. In comparison, the same number of AtT- 20 cells stored 546 f 94 pg of PC1 LI/dish and secreted 1720 f 120 pg of PC1 LI. HPLC analysis of cell extracts and media (Fig. 8 and Table I) showed a significant increase in the percent of SS-14 processed in cell extracts (from 3.5 f 2 to 13.7 f 3%) as well as in media (from 7 f 1 to 31 f 5%) as a result of cotransfection with PSS and PC1 cDNAs compared to control (PSS alone). In the case of SS-28, PC1 cotransfec-

FURlN

tion led to a small (but not significant) decrease in the proportion of this peptide processed in cells (from 35 f 5 to 22 f 11%) which was offset by a small increase in the amount of SS-28 secreted. Overall, however, there was no net change in SS-28 processing as a result of PC1 transfection (Table I). In contrast to PC1, cotransfection of COS-7 cells with PC2 and PSS cDNAs did not influence the efficiency nor pattern of processing of PSS to SS-14 and SS-28 compared to control (Fig. 9 and Table I).

DISCUSSION

In the present study, we have established that mammalian PSS exogenously introduced into COS-7 and PC12 cells undergoes endoproteolytic processing a t monobasic sites in the constitutive secretory pathway. Such processing could be mediated by furin or a furin-like enzyme, the only one of the three known endoproteases expressed in these cells. Furin (or a related enzyme) is thus a likely candidate for effecting monobasic cleavage in addition to its known role in processing a t multibasic sites. In contrast to COS-7 and PC12 cells, AtT- 20 cells processed PSS efficiently a t its dibasic cleavage site via the regulated secretory pathway. These cells contained not only furin but additionally expressed high levels of PC1 and low levels of PC2. Coexpression of PC1 with PSS in COS- 7 cells which lack endogenous processing activity a t dibasic sites led to significant dibasic cleavage suggesting a functional role of this enzyme in PSS conversion to SS-14. In similar cotransfection experiments, PC2 was without effect on PSS processing a t either monobasic or dibasic sites.

Our finding of furin mRNA in each of the four cell lines tested is consistent with the prevailing view of a generalized distribution of this protease (18, 20, 27, 32). Likewise, the detection of PC1 and PC2 in AtT-20 and RIN5f cells, both of which possess regulated secretory pathways, is in keeping with previous reports of the distribution of these enzymes as well as with the postulated role of these enzymes in mediating processing via the regulated secretory pathway (18, 23, 24, 28). Our AtT-20 cells expressed predominantly PC1 and low levels of PC2 in agreement with previous reports (23). This contrasts with the absence of PC2 in AtT-20 cells studied by Bloomquist et al. (38). We provide clear evidence of endoproteolytic activity in the constitutive pathway of COS-7 and PC12 cells capable of processing PSS a t its monobasic cleavage sites at Arg64 and LysI3 to release SS-28 and pro-SS(l-lol. The finding of significant amounts of the 7-kDa peptide coprocessed in these cells gives further support for monobasic processing at the Arg64 residue, since cleavage at this locus would yield the two products, SS-28 and 7 kDa (Fig. 1). These cells processed PSS only weakly at the Arg-Lys dibasic site to generate SS-14. The relative absence of the 8-kDa peptide and of SS-28(1-121 further attests to the poor dibasic processing in these cells since maturation of these two products also

PC 1 PC2

4.5Kb- L) 5Kb- .)

3Kb- 0

2.8Kb- m

FIG. 6. Expression of furin, PCl, and PC2 mRNA in tumor cells. 20 pg of total RNA from each of the cell lines indicated were electrophoresed on 1.5% agarose/formaldehyde gels and hybridized to cRNA probes for rat furin, PC1, and PC2. The furin probe hybridizes to a 4.5-kb transcript found in all cells but with a high level of expression in PC12 and RIN5f cells. PC1 hybridizes to two transcripts of 5 and 3 kb observed in AtT-20 and RIN5f cells but not in COS-7 or PC12 cells. PC2 is expressed as a 2.8-kb transcript found also in AtT-20 and RIN5f cells but not in COS-7 and PC12 cells.


requires cleavage at the Arg-Lys site (Fig. 1). Weak processing of anglerfish PSSI to SS-14 in COS-7 cells was also observed by Warren and Shields (39). Preferential processing of PSS a t monobasic cleavage sites in COS-7 and PC12 cells could be effected by furin, the only one of the three known endoproteases expressed in these cells, or by as yet undiscovered endoproteases: The recently reported substrate specificity of furin certainly qualifies this enzyme for processing a t mono-

PC 1 PC2

3 K b -

0.8K b -

2.8Kb -

0.8Kb - - a

PSS - + + - + + PC1 - t - - + - PC2 - - t + "

FIG. 7. Northern blot analysis of total RNA from COS-7 cells cotransfected with cDNAs for PSS and either PC1 or PC2. Separate autoradiograms were obtained with cRNA probes for PC1 (left-hand side lanes) and PC2 (right-hand side lanes). PC1- or PCP-transfected COS-7 cells show single specific mRNA bands of 3 kb (PC1 mRNA) and 2.8 kb (PC2 mRNA), respectively. The additional 0.8 kb band observed in PC1-transfected cells is detectable with both enzyme probes and appears to be nonspecific.

"i 40

h

1 0 FIG. 8. HPLC Drofiles of SS-14 LI W

in cell lysates ( b n e l A ) and secretion media (panel B ) obtained from COS-7 cells cotransfected with expression vectors for PSS and PC1. Column conditions and markers are as for Fig. 3. Representative of three experiments.

basic cleavage motifs (see below) (18,33). These studies have identified the Arg-Xaa-Lys/Arg-Arg (RXK/RR) motif as the consensus sequence for precursor cleavage catalyzed by furin. Using direct enzyme-substrate reaction, a recent report has further narrowed the furin consensus sequence to Arg-Xaa- Xaa-Arg as the minimum requirement for catalysis by this enzyme (40). Such a specificity profile shows considerable overlap with the structural requirements for processing at monobasic cleavage sites proposed by Devi (41). These criteria predict the monobasic cleavage motifs Arg-Leu-Glu-Leu-Gln- Arg and Arg-Gln-Phe-Leu-Gln-Lys at the site of SS-28 and PSS(l-lo, processing, respectively. Except for the presence of Lys rather than Arg at the site of PSS(l-lo, conversion, both these sequences possess the necessary substrate specificity for catalysis by a furin-like enzyme: An interesting byproduct of our processing studies was the discovery of significant quantities of the NH2-terminal 2.5-kDa peptide (Fig. 5). This molecule was previously noted as a consistent cellular and secreted product of in vivo PSS processing in the rat, and its occurrence in both COS-7 and PC12 cells raised the possibility that additional sites of processing may exist in the amino- terminal segment of PSS (11). Such processing could occur at the site of a hitherto unsuspected monobasic cleavage motif Lys-Gln-Glu-Leu-Ala-Lys at PSS(21-26).

Concerning the potential role of PC1 and PC2 in PSS processing, the endogenous expression of these two endoproteases correlated with efficient processing. For instance, although PSS was significantly processed by furin or furin-like enzymes in COS-7 and PC12 cells, such processing was relatively inefficient based on the quantities of unprocessed precursor, whereas in AtT-20 cells, PSS was virtually completely processed. Coexpression of PC1 with PSS in COS-7 cells which lack endogenous processing activity at dibasic sites resulted in significant although inefficient production of SS- 14. PC2 on the other hand did not influence monobasic or

55-28 1

LI PSS 1

S5-14 1 II

i o i I 80

PSS 1

I loo

A .

B.

7 120

FRACTION NUMBER

6048 Prosomatostatin Processing by PCl, PC2, and Furin

TABLE I Cumulative data comparingpercentprocessed SS-14 and SS-28 with unprocessed PSS in cells transfected with PSS alone or cotransfected

with PSS + PC2 and PSS + PC1 Coexpression of PC1 with PSS results in a significant increase in the percent of SS-14 processed in cell extracts and secretion media.

Percent SS-28 processed is unaffected by PC1 cotransfection. PC2 is without effect on the percent of SS-14 or SS-28 processed compared to control. Mean data f S.E. (n = 3).

SS-14 (%) SS-28 (96) Unprocessed PSS (%)

Cells Media Cells Media Cells Media

PSS PSS + PC1

3.5 f 2 7 f l 35 f 5 5 f 2 61.5 +- 4 88 f 3 13.7 f 3

PSS + PC2 3 1 2 4 22.3 f 11 1 1 f 4 64 f 7 58 f 4

4 f l 4 + - 9 30.5 f 7 2 f 0.5 65.5 +- 5 84 f 6

I SS28

FIG. 9. HPLC profiles of SS-14 LI in cell lysates (panel A ) and secretion media (panel B ) from COS-7 cells cotransfected with expression vectors for PSS and PC2. Column conditions and markers are as shown in Fig. 3. Representative of three experiments.

A .

PSS 1

1

2 0 0 - SS14

n 100- I

E \ - 0 , #I

4 0 0

B.

3 0 0 I SS14 SS28 1 1

PSS 1

2ool - " 1 0 0 0 20

dibasic processing of PSS in COS-7 cells. These findings correlate with the pattern of processing of PSS in AtT-20 cells which show high endogenous expression of PC1 and which also process the precursor mainly to SS-14. Several reasons could explain the inefficient processing by transfected PC1 in COS-7 cells compared to AtT-20 cells. A major difference could be in the level of endogenous and exogenous PC1 expression in AtT-20 and COS-7 cells, respectively. For instance, although expression of PC1 in transfected COS-7 cells was confirmed by analysis of both mRNA and protein, the level of protein expression was considerably lower than that of endogenous PC1 in AtT-20 cells. Second, compartmentalization of the processing events in the regulated secretory pathway (AtT-20 cells) or the constitutive pathway (COS-7 cells) could also affect processing efficiency as suggested by the ability of PC1 to cleave prorenin at the Lys-Arg site in the regulated pathway of GH, cells but not in the constitutive pathway of Chinese hamster ovary cells (18, 26, 35). Given the limitations of duplicating the exact intracellular milieu for optimal enzyme-substrate reaction in transfection experiments, the full potency of PC1 for efficient catalysis of PSS

4i M, so loo 120

FRACTION NUMBER

to SS-14 cannot be accurately determined. Our studies provide direct proof that PC1 can serve as a SS-14 convertase but do not exclude the involvement of other perhaps more specific endoproteases.

The mammalian SS gene has evolved from two ancestral genes in fish which encode for separate SS-14 and SS-28 type precursors whose products are expressed in separate subpop- ulations of SS cells (13, 42, 43). This means that monobasic and dibasic processing of the fish PSS molecules to either SS- 14 or SS-28 type peptides occurs in different cells probably by separate proteases, whereas the two events appear to be compartmentalized in the same cell in mammals. Two recently identified anglerfish islet proteases have been assigned specific SS-14 and SS-28 converting roles. The SS-28 convertase is a cathepsin D-like aspartyl protease of M, = 39,000 which is distinct from the kex2/furin/PC family of subtilisin proteases (44). I t also differs markedly in size and specificity from two other putative monobasic processing enzymes, a SS- 28 generating rat intestinal convertase and a frog skin endoprotease which cleaves at a typical RXVRG monobasic motif (45, 46). The anglerfish SS-14-converting enzyme is a cal-


cium-activated neutral endoprotease of M, = 52,000-63,000 which shows amino-terminal sequence homology with mammalian PC2 (47). These findings are at variance with the present observations of the conversion of mammalian PSS to SS-14 by PC1 but not PC2. Given the acidic pH optimum of PC2, one explanation for inactivity of PC2 in COS-7 cells could be the compartmentalization of the transfected enzyme to the neutral pH milieu of the constitutive secretory pathway. This, however, may not be a major limitation since other precursors, e.g. proLHRH can be processed by transfected PC2 to mature forms in the constitutive secretory pathway (48). Another possibility could be the relatively slow zymogen activation of PC2 compared to PC1 (35). It is of interest that GHs cells which possess secretory granules and which express PC2 (but not PC1) mRNA process anglerfish PSSI readily to SS-14 (49), whereas the closely related GH4C1 cells which also express only PC2 mRNA2 cleave rat PSS only weakly to SS- 14 (22). Since efficient proprotein processing is determined not just by the type of protease but by a multiplicity of other factors including convertase processing, substrate specificity, and intracellular environment (35) the present studies based on heterologous processing in constitutively secreting COS-7 cells cannot exclude a physiological role of PC2 in PSS processing. Further experiments correlating PSS processing with the expression or inhibition of PC2 in regulated secretory cells may help to resolve this issue. Unfortunately, the presence of endogenous PC2 in most regulated secretory cell lines has so far precluded any direct attempts to examine PC2 specificity for PSS processing by cotransfection experiments in these cells.

In conclusion, we have established that PSS can be processed preferentially at monobasic sites through the constitutive secretory pathway. Such processing could be mediated by a furin-like enzyme but is relatively inefficient. Efficient processing at both dibasic and monobasic sites requires the presence of additional proteases such as PC1 and PC2 and may depend on targeting of PSS with the enzymes to the regulated secretory pathway. PC1 is capable of effecting dibasic cleavage of PSS whereas PC2 is without effect on PSS processing, at least within the constitutive secretory pathway. Whether additional proteases exist capable of efficient monobasic cleavage of PSS remains to be determined.

Acknowledgments-We are grateful to A. Warszynska for technical assistance and to M. Correia for secretarial help.

REFERENCES 1. Steiner, D. F., Quinn, P. S., Chan, S. J., Marsh, J., and Tager, H. S. (1980)

2. Eipper, B. A,, Mains, R. E., and Herbert, E. (1986) Trends Neurosci. 9 , Ann. N. Y. Acad. Sci. 343 , l -16

3. Brazeau, P., Vale, W., Burgus, R., Ling, N., Butcher, M., Rivier, J., and 463-468

Guillemin, R. (1973) Science 179 , 77-79

* N. G. Seidah, unpublished observations.

4. Pradayrol, L., Jornvall, H., Mutt, V., and Ribet, A. (1980) FEBS Lett. 109 ,

5. Benoit, R., Ling, N., and Esch, F. (1987) Science 238,1126-1129 6. Benoit, R., Bohlen, P., Ling, N., Briskin, A., Esch, F., Brazeau, P., Ying, S.

Y., and Guillemin, R. (1982) Proc. Natl..Acad. Sci. U. S. A. 79,917-921 7. Zingg, H. H., and Patel, Y. C. (1982) J. CBn. Inuest. 70(6), 1101-1109 X. Patel. Y. C.. Zinez. H. H.. Srikant. C. B. (1985) in Somutostatin (Patel Y.

55-58

.. - ~ . .~ ~

C.,and TanneiXaum-G. S.: eds)'pp. 71-88, Plenum Press, New York 9. Patel, Y. C., and O'Neil, W. (1988) J. Biol. Chem. 263,745-751

10. Benoit, R., Bohlen, P., Esch, F., and Ling, N. (1984) Brain Res. 311 , 23-

11. Rabbani, S., and Patel, Y. C. (1990) Endocrinology 126(4), 2054-2061 12. Benoit, R., Esch, F., Bennett, H. P. J., Ling, N., Ravazolla. M., Orci, L.,

and Mufson, E. J. (1990) Metab. Clin. Exp. 39(9) , Suppl. 2,22-25 13. Areos. P.. Tavfor. W. L.. Minth. C. D.. and Dixon. J. E. (1983) J. Biol.

29

14.

16. 15.

17. 18.

19.

20.

21.

22.

23.

24.

25. 26.

27.

28.

29. 30.

31.

32. 33.

34.

35.

36.

37. 38.

40. 39.

42. 41.

43.

44.

&m. 258; 87b-8793 Conlon, J. M. (1990) Prog. Clin. Biol. Res. 342 , 10-15 Lingappa, V. R. (1983) J. Clin. Inuest. 83,739-751 Rothman, J. E., and Orci, L. (1992) Nature 3 5 5 , 409-415 Kelly, R. B. (1985) Science 230,25-32 Hosaka, M., Nagahama, M., Kim, W-S., Toshio, W., Hatsuzawa, K.,

Ikemizu, J., Murakami, K., and Nakayama, K. (1991) J. Bwl. Chem.

Thomas, G., Thorne, B. A,, and Hruby, D. E. (1988) Annu. Rev. Physiol. 266,12127-12130

Bresnahan, P. A,, Leduc, R., Thomas, L., Thorner, J., Gibson, H. L., Brake, 50,323-332

Orci, L., Ravazolla, M., Amherdt, M., Madsen, 0.. Vassali, J.-D., and A. J., Barr, P. J., and Thomas, G. (1990) J. Cell Biol. 111,2851-2859

Sevarino, K. A., Felix, R., Banks, C. M., Low, M. J., Montmigny, M. R., Perrelet, A. (1985) Cell 4 2 , 67-81

Benjannet, S., Rondeau, N., Day, R., Chretien, M., and Seidah, N. G. (1991) Mandel, G., and Goodman, R. H. (1987) J. Biol. Chem. 262,4987-4993

Thomas, L., Leduc, R., Thorne, B. A., Smeekens, S. P., Steiner, D. F., and Proc. Natl. Acad. Sci. U. S. A. 88,3564-3568

Fuller, R. S., Brake, A. J., and Thorner, J. (1989) Science 246,482-486 Thomas, G. (1991) Proc. Natl. Acad. Sci. U. S. A. 88,5297-5301

Hatsuzawa, K., Hosaka, M., Nakagawa, T., Nagase, M., Shoda, A,, Murak- ami, K., and Nakayama, K. (1990) J. Bwl. Chem. 266,22075-22078

Seidah, N. G., Gaspar, L., Mion, P., Marcinkiewicz, M., Mbikay, M., and Chretien, M. (1990) DNA Cell Biol. 9(6), 415-424

Seidah, N. G., Marcinkiewicz, M., Benjannet, S., Gaspar, L., Beaubien, G., Mattei,.M. G., Lazure, C., Mbikay, M., and Chretien, M. (1991) Mol. Endocrmnol. 5(1), 111-122

Smeekens, S. P., and Steiner, D. F. (1990) J. Biol. Chem. 266,2997-3000 Smeekens, S. P., Avruch, A. S., LaMendola, J., Chan, S. J., and Steiner, D.

Hakes, D. J., Birch, N. P., Mezey, A., and Dixon, J. E. (1991) Endocrinology

Barr, P. J. (1991) Cell 6 6 , l - 3 Korner, J., Chun, J., OBryan, L., and Axel, R. (1991) Proc. Natl. Acad. Sci.

Korner. J.. Chun. J.. Harter. D.. and Axel. R. (1991) Proc. Natl. Acad. Sci. U. S. A. 88, 11393-11397

F. (1991) Proc. Natl. Acad. Sci. U. S. A. 88,340-344

129,3053-3063

U. S.'A. '88,683416638

and Seidah, N. G: (1992) J. Bwl. Chem. 267 , 11417-11423

126(2), 948-956

. , . . .

Benjannet, S., Reudelhuber, T., Mercure, C., Rondeau, N., Chretien, M.,

Murthy, K. K., Srlkant, C. B., and Patel, Y. C. (1989) Endocrinology

Chomczynzki, P., and Sacchi, N. (1987) Anal. Biochem. 162, 156-158 Bloomquist, B. T., Eipper, B. A., and Mains, R. E. (1992) Mol. Endocrinol.

Warren, T. G., and Shields, D. (1984) Cell 39,547-555 Molloy, S. S., Bresnahan, P. A., Leppla, S. H., Klimpel, K. R., and Thomas,

Devi, L. (1991) FEBS Lett. 280 189-194 McDonald, J. K., Greiner, F., Bauer, G. E., Elde, R. P., and Noe, B. D.

Sevarino, K. A., Stork, P., Ventimiglia, R., Mandel, G., and Goodman, R.

Mackin, R. B., Noe, 8. D., and Spiess, J. (1991) Endocrinology 129 , 1951-

5,2014-2024

G. (1992) J. Biol. Chem. 267,16396-16402

(1987) J. Histochem. Cytochem. 3 5 , 155-162

H. (1989) Cell 57,11-19

1457 45. Bei;;Te'ld, M. C., Bourdais, J., Morel, A., Kuks, P. F. M., and Cohen, P.

46. Kuks, P. F. M., Creminon, C., Leseney, A.-M., Bourdais, J., Morel, A,, and

47. Mcc$, R. B., Noe, B. D., and Spiess, J. (1991) Endocrinology 129,2263-

(1989) J. Biol. Chem. 264,4460-4465

Cohen P. (1985) J. Biol. Chem. 2 6 4 , 14609-14612

48. Wetsel W. C. Thomas, L., Hayflick J. S., Rivera, H. R., Lautermilch, N.,

Endocrine Soctet , June 24 27,1992, San Antonio, Texas, Abstr. 1606 and ?hornis, G. (1992) Program' of the 74th Annual Meeting of the

49. Stoller, T. J., and Aields, D.jl988) J. Cell Biol. 107 , 2087-2095

ZZbD

Documents

THE OF No. 8, Issue March 1993 Printed in U.S.A ...€¦ · tially at monobasic sites to SS-28 (40-43%) and antrin (27-29%). Processing, however, was inefficient as suggested by large