Neuronal cell lines expressing PC5, but not PC1 or PC2, process Pro-CCK into glycine-extended CCK 12 and 22

Peptides 22 (2001) 1271–1277

Neuronal cell lines expressing PC5, but not PC1 or PC2, process Pro-CCK into glycine-extended CCK 12 and 22 Brian M. Cain*, Daesety Vishnuvardhan, Margery C. Beinfeld Department of Pharmacology and Experimental Therapeutics, Tufts University School of Medicine, Boston, MA 02111, USA Received 31 January 2000; accepted 5 March 2001

Abstract Endocrine tumor cells in culture and in vitro cleavage assays have shown that PC1 and PC2 are capable of processing pro-CCK into smaller, intermediate and final, bioactive forms. Similar studies have shown that PC5 has the ability to process a number of propeptides. Here, we use GT1–7 (mouse hypothalamic) and SK-N-MC and SK-N-SH (human neuroblastoma) tumor cell lines to study the ability of PC5 to process pro-CCK. RT-PCR and Western blot analysis showed that the cells express PC5 mRNA and protein, but not PC1 or PC2. They were engineered to stably overexpress CCK and cell media was analyzed for pro-CCK expression and cleavage of the prohormone. Radioimmunoassays showed that pro-CCK was expressed, but no amidated CCK was detected. Lack of production of amidated CCK may be due to the lack of the appropriate carboxypeptidase and amidating enzymes. Production of glycine-extended CCK processing products was evaluated by treatment of media with carboxypeptidase B followed by analysis with a CCK Gly RIA. Glycine-extended forms of the peptide were found in the media. The predominant forms co-eluted with CCK 12 Gly and CCK 22 Gly on gel filtration chromatography. The results demonstrate that these cell lines which express PC5 and not PC1 or PC2 have the ability to process pro-CCK into intermediate, glycine-extended forms more closely resembling pro-CCK products in intestine than in brain. © 2001 Elsevier Science Inc. All rights reserved. Keywords: Cholecystokinin; Procholecystokinin; Prohormone convertase 5; Mouse hypothalamic neuron; Human neuroblastoma cell; Glycine-extended peptide; Prohormone processing

1. Introduction Cholecystokinin (CCK) is a peptide found in the intestine where it regulates gallbladder contraction and pancreatic enzyme secretion [1]. CCK is also abundant in many brain regions, such as the frontal cortex, hippocampus, and striatum [5], where it can act as a neuromodulator and neurotransmitter. During post-translational processing, pro-CCK undergoes modifications that include tyrosine sulfation, endoproteolytic cleavage, basic residue removal, and c-terminal amidation. Within the secretory vesicles, pro-CCK is cleaved into smaller, intermediate forms by a variety of endoproteases. The major bioactive form of this peptide in brain is CCK 8 and the gut makes CCK 58, CCK 33 and CCK 22. The endoproteolytic enzymes responsible for the pro* Corresponding author. Tel.: ⫹1-617-636-2421; fax: ⫹1-617-6366738. E-mail address: [email protected] (B.M. Cain).

cessing of CCK and the temporal order in which these enzymes cleave the prohormone have not been definitively identified. However, several studies have suggested that the subtilisin/kexin-like enzymes, such as prohormone convertase 1 (also known as PC1/PC3, EC 3.4.21.93) and PC2 (EC 3.4.21.94), appear to be involved in the biosynthesis of CCK in endocrine tumor cell lines [4,31–33]. It has also been demonstrated that recombinant PC1 and PC2 in vitro can process pro-CCK into smaller intermediate forms. However, which enzymes are responsible for the complete processing of pro-CCK in neuronal tissues is still unknown. PC5 (EC 3.4.21), a mammalian kex2-like processing endoprotease, has been identified in rat and mouse endocrine and neuroendocrine tissues [19,21,22,26]. The cDNA structure is known to be homologous with proprotein convertases like PACE4, furin, and other PC enzymes [19,21]. PC5 is encoded by multiple mRNAs that generate two splice variant isoforms referred to as PC5A (915 amino acids) and PC5B (1877 amino acids) [22]. PC5A is smaller than PC5B and when stably transfected into endocrine cells it can be processed into two forms (117 kDa and 65 kDa),

0196-9781/01/$ – see front matter © 2001 Elsevier Science Inc. All rights reserved. PII: S 0 1 9 6 - 9 7 8 1 ( 0 1 ) 0 0 4 5 1 - X

1272

B.M. Cain et al. / Peptides 22 (2001) 1271–1277

sorted into regulated secretory granules, and secreted into the media [22]. Several transfection and processing studies have shown that PC5A can cleave various precursor proteins, such as prorenin [15], proMIS (proMullerian Inhibiting Substance) [20], proRPTP (proReceptor Protein Tyrosine Phosphatase) [7], proneurotensin [2], and proglucagon [6]. Recent localization studies have shown that PC5A mRNA (65 kDa form) is expressed in neuronal cells in numerous rat brain regions, including the cortex, hippocampus, and hypothalamus [10,34]. These are the same areas that express high levels of CCK so that it is a good candidate for the processing of pro-CCK in the CNS. In this study we used neuronal tumor cells in culture that express PC5A (referred to as PC5 in this study), but not PC1 or PC2, to examine whether they were able to process pro-CCK.

2. Method 2.1. Preparation of recombinant Pro-CCK expression vector Recombinant pro-CCK expression plasmid was constructed by subcloning full-length rat pro-CCK cDNA (115 amino acids, Mr 12,826) [9] into the HindIII and EcoR1 sites of a pcDNA3 expression vector (Invitrogen Corporation). 2.2. Maintenance and transfection of tissue culture cell Mouse hypothalamic (GTI-7) neurons were obtained from Dr. Pamela Mellon at the University of California, San Diego. Mouse intestinal tumor (␤-STC-1) and mouse insulinoma (␤TC-3) cell lines were obtained from Dr. Douglas Hannahan at the University of California, San Francisco. Human neuroblastoma (SK-N-SH & SK-N-MC), and human neuroepithelioma (SK-N-MC-IXC) cell lines were obtained from the American Type Culture Collection (ATCC) and were maintained as previously described [3,11,12,17]. The cells were grown in Dulbecco’s minimal essential medium (DMEM, GIBCO) containing 5% calf serum (CS, GIBCO), 5% fetal bovine serum (FBS, GIBCO), 50 ␮g/ml streptomycin sulfate, and 50 units/ml penicillin (Sigma Chemical Co) and were maintained at 37°C with 5% CO2 and 95% air. The cell lines were stably transfected with full-length rat pro-CCK cDNA using a LipofectAMINE PLUS Reagent Kit (GIBCO). Cells were transfected with 5–20 ␮g of proCCK expression plasmid according to the manufacturer’s instruction and were maintained using a selection antibiotic (700 ␮g/ml G418). Transfected cells were harvested and cell extract and media samples were analyzed for CCK expression and processing by RIA.

2.3. RT-PCR of PC1, PC2, and PC5 All of the neuronal cell lines were screened for the presence of PC1, PC2, and PC5 mRNA using reverse transcriptase-polymerase chain reaction (RT-PCR). Mouse PC1 (mPC1), rat PC2 (rPC2), and mouse PC5 (mPC5) oligonucleotide primer sequences were designed from the published sequences of the cDNAs encoding these prohormone convertases [19,21,24,25]. The primer sequences were designed to amplify a 500 bp DNA fragment for PC1, a 900 bp fragment for PC2, and an 800 bp fragment for PC5. It should be noted that part of the PC1, PC2, and PC5 full-length cDNA sequences are highly conserved between higher vertebrate species (mouse, rat, and human). Therefore, mouse and rat sequences were used when designing the primer sequences even though the primers were used to screen both human and mouse neuronal cell lines. Additionally, the hPC5 cDNA sequence was only partially sequenced so its cDNA sequence was not used when designing the PC5 primers. The 5⬘ (start) and 3⬘ (stop) primer sequences were primarily taken from the proregions of each endoprotease because these are the least conserved regions among all of the prohormone convertases. The published mPC1 cDNA sequence allowed for the design and synthesis of a 21-mer 5⬘ primer (5⬘-ACTTGCAAGATACCAGAATGA-3⬘) and a 27-mer 3⬘ primer (5⬘-CCCATTCCCTGAAGCCCAGACAAAGAT-3⬘). Similarly, the published rPC2 and mPC5 cDNA sequences allowed for the synthesis of a 24-mer 5⬘ primer (5⬘-GATCCTCTTTTTACAAAGCAATGG-3⬘) and a 24-mer 3⬘ primer (5⬘-GGTGAGCACTGTCAGATGTTGCAT-3⬘) for PC1 and a 20-mer 5⬘ primer (5⬘-CCCAAGTGGCCAAGTATGTG-3⬘) and a 24-mer 3⬘ primer (5⬘GGCAATGATTCCAGCAGCCATGGG-3⬘) for PC5. The specificity of the primers was confirmed by conducting PCR reactions using the appropriate primers with respective stock (PC1, PC2, or PC5) plasmids. Total RNA extractions were performed on all of the neuronal cell lines (GT1–7, SK-N-SH, SK-N-MC, and SKN-MC-IXC) along with a positive control cell line (␤STC-1, mouse intestinal tumor cell line) using guanidine isothiocyanate [8]. RNA was quantitated spectrophotometrically and the purity was examined with a denaturing RNA gel. Once the purity of the samples was confirmed, cDNA templates for PCR were synthesized from 2.5 ␮g of total RNA from these cells by using Moloney murine leukemia virus (MuLV) reverse transcriptase as described by PerkinElmer (GeneAmp RNA PCR Core Kit). The final volume of the RT reaction mixture was 50 ␮l, which contained 2.5 ␮l 50 ␮M random hexamers, 10 ␮l 25 mM MgCl2, 5 ␮l 10⫻ PCR Buffer II, 5 ␮l 10 mM dATP, 5 ␮l 10 mM dCTP, 5 ␮l 10 mM dGTP, 5 ␮l 10 mM dTTP, 2.5 ␮l 20 U/␮l RNase inhibitor, 2.5 ␮l 50 U/␮l MuLV reverse transcriptase, 2.5 ␮g total RNA (concentration-dependent volume), and diethyl pyrocarbonate (DEPC)-treated water balance. All of

B.M. Cain et al. / Peptides 22 (2001) 1271–1277

the RT reaction mixtures were incubated at 25°C for 10 min, 42°C for 1 h, and then 99°C for 5 min. Following the RT step, the samples were either chilled on ice or stored at –20°C until the cDNA fragments were amplified using PCR technique. For all of the samples, 1/25th of the cDNA template reaction mixture from each RT step was used for PCR. PCR reaction mixtures had a final volume of 100 ␮l and contained 2 ␮l cDNA template, 10 ␮l ThermoPol Buffer, 2 ␮l 10 mM dNTP mix, 2 ␮l 20 ␮M mPC1, rPC2, or mPC5 5⬘ primer, 2 ␮l 20 ␮M mPC1, rPC2, or mPC5 3⬘ primer, 81 ␮l sterile water, and 1 ␮l NEB Vent DNA polymerase. Reactions were carried out in a Thermolyne TempTronic thermal cycler for 25 cycles of denaturation (94°C, 1 min 30 sec), annealing (55°C, 1 min 30 sec), and extension (72°C, 1 min). The PCR samples were examined on a 1% agarose gel. In order to confirm the identity of the PCR products, the RT-PCR samples were subcloned into PCBlunt vector (Invitrogen). Restriction enzyme digests were performed on the purified cDNA samples to confirm the orientation of the PC1, PC2, and PC5 RT-PCR products in the PCBlunt vector and the RT-PCR samples were then sequenced (Tufts University School of Medicine Core Facility, Department of Physiology, Boston, MA). The DNA sequences of the RTPCR products for all of the endoproteases were evaluated and compared with published sequences in the literature using a BLAST 2 Sequence software package. 2.4. Western blot analysis for PC1, PC2, and PC5 In order to confirm the expression of PC1, PC2, and PC5 protein, Western blot analysis was performed on cell extracts of GT1–7, SK-N-SH, SK-N-MC, and SK-N-MC-IXC neuronal cells as previously described [32]. Briefly, polyclonal antibodies recognizing PC1 and PC2 proteins were generously provided by Dr. Iris Lindberg at the Louisiana State Medical Center [23,27]. Additional polyclonal antibodies recognizing the amino- and carboxyl-terminals of the PC5 protein were generously provided by Dr. George Greeley of the University of Texas Medical Branch at Galveston. Tissue culture cells were grown to near confluency and harvested by scraping from dishes with 1.0 ml phosphatebuffered saline (PBS), after which an aliquot was taken for total protein measurement by the Lowry assay [18]. After centrifugation, cells were extracted by resuspending in 1⫻ SDS protein loading buffer (25 mM Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, 5% 2-mercaptoethanol, 0.001% bromophenol blue), sonicated, and boiled for 5 min. Protein samples were fractionated by electrophoresis at 20 –25 mAmps for 1.5 h through a 10% Tris-HCl polyacrylamide gel in protein electrophoresis buffer (0.1% SDS, 0.025 M Tris, 192 mM glycine) using a protein minigel apparatus (BioRad). The fractionated proteins were electroblotted onto nitrocellulose in transfer buffer (25 mM Tris, 192 mM glycine, 20% methanol) at 4°C for 1.5 h at 70 V using the

1273

minigel Trans-Blot apparatus (Bio-Rad). Following the transfer the blot was dipped in gel fix for 1 min (methanol [4]:acetic acid [1]:Milli-Q water [5]), rinsed in Milli-Q water, and stored in TBS with 0.06% Tween overnight at 4°C. The next day the nitrocellulose blot was blocked using 3% Carnation nonfat dried milk in TBS (50 mM Tris, 200 mM NaCl, pH 7.4) with 0.06% Tween containing 0.02% sodium azide (Blotto) for 1 h at RT using a shaker. The blot was then incubated with a 1:1000 dilution of primary PC1 or PC2 antibody or a 1:100 dilution of primary PC5 antibody in 10 ml of 33% Blotto/66% TBS-Tween at RT for 1 h with constant shaking. The membrane was then reblocked in Blotto for 30 min at RT on a shaker. Following the reblock step, the blot was incubated with a donkey antirabbit IgG antibody conjugated to horseradish peroxidase in 10 ml of 33% Blotto/66% TBS-Tween at RT for 1 h with constant shaking. This secondary antibody conjugated to horseradish peroxidase (Pierce) was diluted 1:5000. Visualization by enzyme chemiluminescence using the horseradish peroxidase enzyme was performed according to manufacturer’s protocol from the ECL Western detection kit from Amersham Pharmacia Biotech. 2.5. Sample preparation for radioimmunoassay (RIA) and chromatographic analysis The sample preparation of cell extracts and media samples prior to RIA and chromatographic analysis was performed as previously described [28]. Cell extract and media samples from GT1–7, SK-N-SH, SK-N-MC, and SK-NMC-IXC cells expressing the highest pro-CCK content were collected from multiple dishes and were concentrated with Sep-Pak C18 cartridges (Millipore). The cells from several plates were extracted with 0.1 N HCl and centrifuged at 4°C. The supernatant was concentrated using the Sep-Pak C18 cartridges. Similarly, media samples from the neuronal cells were pooled in 50 ml centrifuge tubes, centrifuged at 4°C, and then concentrated using the Sep-Pak C18 cartridges. Samples were eluted from the cartridges with 90% acetonitrile/0.1% trifluoroacetic acid (TFA) and concentrated to 1.0 ml in a Speedvac. Sep-Pak purified cell extract and media samples (1.0 ml) were then treated with 100 ␮l carboxypeptidase B (CpB; 5 mg/ml; Boehringer Mannheim) at 20°C for 30 min and the reaction was stopped by boiling for 10 min [13]. After the boiling step, all samples were stored at 4°C until separation. 2.6. Low pressure gel filtration column chromatography Pretreated cell extract and media samples (1.0 ml) from the GT1–7, SK-N-SH and SK-N-MC neuronal cell lines were separated using low pressure chromatography as previously described [28]. Briefly, the samples were loaded onto a 35 ⫻ 2.5-cm column of Sepharose GCL-90 (Isco), which was eluted with 50 mM Tris, 100 mM NaCl, pH 7.8,

1274

B.M. Cain et al. / Peptides 22 (2001) 1271–1277

Table 1 CCK RIA antibody and trace Antibody

Tracer

Peptide standard

CCK form detectedb

79

Y10Ma

V9Ma

R5

Gastrin17 (G17)

CCK 8

22

CCK 8 Gly

CCK 8 Gly

pro-CCK, aminoterminal amidated CCK peptides glycine extended CCK peptides

(G13 Gly)

(G13 Gly)

a

The RIA to measure pro-CCK (amino terminal) from cell extract and media samples used a Y10M tracer (YVPVEAVDPM) and a V9M peptide standard (VPVEAVDPM). b Specific antibodies and tracers were used to detect pro-CCK and various CCK peptide forms.

containing 0.1% BSA and 0.05% sodium azide. The column was run at 4°C, 1.0 ml fractions were collected, and aliquots were removed for RIA analysis. In addition, prohormone, glycine-extended and amidated CCK peptide standards of various sizes (pro-CCK, CCK 30, CCK 22 Gly, CCK 22, CCK 12 Gly, and CCK 8 Gly) were run on the low pressure gel filtration column. The CCK 22 standard was generated by treating CCK 30 with endolysine C. 2.7. CCK radioimmunoassays (RIAs) The RIA to measure amidated forms of CCK from cell extract and media samples was performed as previously described [5]. The rabbit polyclonal CCK antibody (R5) is highly specific for amidated CCK (Table 1) and it displays a 0.001% cross reactivity with CCK 8 Gly, gastrin 13 Gly, CCK 8 Gly Arg, and CCK 8 Gly Arg Arg. As tracer, the RIA utilized gastrin 17 (G17) produced by iodination using the chloramine-T method [14]. The iodinated G17 was then separated by Sephadex G10 and further purified with a DEAE-A25 ion exchange column eluted with a gradient of sodium chloride. The pro-CCK RIA [30] utilized antiserum 79 which was generated against Y10M (YVPVEAVDPM) with a carboxyl-terminal multiple antigenic peptide tail. It was used to detect pro-CCK and its amino-terminal propeptide (Table 1). The tracer for the RIA was 125I-labeled Y10M. The CCK Gly RIA [29] utilized CCK 8 Gly antiserum 22 generated from unsulfated CCK 8 Gly conjugated at a 50:1 molar ratio to keyhole limpet hemocyanin with 0.1% glutaraldehyde (Table 1). The tracer was 125I-labeled Gastrin 13 Gly (Gly-Leu-Glu-Glu-Glu-Glu-Glu-Ala-Tyr-Met-AspPhe-Gly) separated by Sephadex G10 and further purified with a DEAE-A25 ion exchange column eluted with a gradient of sodium chloride. With 125I-labeled Gastrin 13 Gly as the tracer, antiserum 22 cross reacts 40% with CCK 8 Gly, 0.4% with sulfated CCK 8 amide, 1.25% with unsulfated CCK 8 Gly Arg, and 1.25% with unsulfated CCK 8 Gly Arg Arg.

Fig. 1. Reverse transcription-PCR of PC1, PC2, and PC5 in mouse hypothalamic neurons and human neuroblastoma cell lines. The primer sequences amplify a 500 bp DNA fragment for PC1, a 900 bp fragment for PC2, and an 800 bp fragment for PC5. A, GT1–7; B, SK-N-SH; C, SK-N-MC; D, SK-N-MC-IXC; E, ␤-STC-1; X, PC1; Y, PC2; Z, PC5.

2.8. Data analysis The RIAs were counted with a Micromedic gamma counter (Micromedic Systems, Huntsville, AL) and calculated using an interfaced Micromedic MACC data reduction system. Duplicate sample results were expressed as mean ⫾ SEM and the actual concentrations of individual samples were reported as pg/ml. Data from the RIA analysis of samples and standards passed through the GCL-90 and HPLC columns was plotted with fraction numbers on the x-axis vs peptide concentrations (pg/ml) on the y-axis.

3. Results 3.1. Expression of pro-CCK in mammalian neuronal cell lines The cell lines were engineered to overexpress pro-CCK so its processing could be studied. The expression of proCCK within the cell extract and media samples was confirmed using an RIA specific for the detection of the prohormone. Several subclones of each neuronal cell line were found to express very high levels of pro-CCK within the cell extract and media samples (data not shown). They were maintained for processing studies. 3.2. PC1, PC2, and PC5 expression in mammalian neuronal cell lines Prior to evaluating the postranslational processing of pro-CCK by PC5, the expression of PC1, PC2, and PC5 was examined by RT-PCR to detect enzyme mRNA and Western blot analysis to detect enzyme expression. Fig. 1 shows RT-PCR results for PC1, PC2, and PC from the neuronal cell lines plus a positive control cell line, ␤STC-1, which

B.M. Cain et al. / Peptides 22 (2001) 1271–1277

Fig. 2. Western blot analysis of PC5 in mouse hypothalamic neurons and human neuroblastoma cell lines. The 65 kDa band is representative of active PC5 and the larger 117 kDa band indicates the presence of the proform of the enzyme. A, ␤-STC-1; B, WE; C, GT1–7; D, SK-N-SH; E, SK-N-MC; F, SK-PN-DW.

has all three enzymes. The GT1–7, SK-N-SH, SK-N-MC, and SK-N-MC-IXC cell lines had no bands for PC1 and PC2, but had an 800 bp band for PC5 (Fig. 1). In order to confirm the identity of the PCR products, the RT-PCR samples were sequenced and compared with published sequences in the literature using a BLAST2 Sequence software package. The mPC1 RT-PCR product sequence exhibited a 99% homology with the published mPC1 sequence, mPC2 was 94% homologous, mPC5 was 97% homologous, and hPC5 was 87% homologous with its published sequence (data not shown). Western blot analysis was performed on cell extracts and the neuronal cell lines were found to possess PC5, but not PC1 or PC2 (Fig. 2). The PC5 protein was identified in GT1–7, SK-N-SH, SK-N-MC, and SK-N-MC-IXC cells using primary polyclonal antibodies recognizing the aminoand carboxyl-terminals of the PC5 protein. The results showed a major 65 kDa band which corresponds to active PC5 and small amounts of a larger 117 kDa band which is the proform of the enzyme [19,21]. Several positive control samples, such as rat whole brain and intestinal tissues, D9 (␤-STC-1 clone), WE (mouse thyroid medullary tumor), SK-PN-DW (human neuroblastoma), GH3 (rat pituitary tumor), NTN (human neuronal precursor cells), NT2 (human neuronal teratocarcinoma), and ␤-STC-1 (mouse intestinal tumor) cells, also yielded the same size PC5 proteins (Fig. 2). Negative controls, such as ␤TC-3 (mouse insulinoma) cell extract and media samples, showed no active PC5 protein bands (data not shown). Western blot analysis was also used to screen the neuronal cell lines for the presence of PC1 and PC2 protein expression. None of the neuronal cell lines (GT1–7, SK-NSH, SK-N-MC, and SK-N-MC-IXC) expressed PC1 or PC2 protein in the cell extract samples. However, PC1 was easily detected in the cell extracts of positive control samples (␤TC-3 and ␤-STC-1) with a larger precursor form with a molecular weight of about 100 kDa and a smaller 87 kDa band for active PC1 (data not shown). Similarly, PC2 was

1275

Fig. 3. Gel filtration chromatography (GCL-90) of media samples from various neuronal cell lines transfected with CCK and treated with CpB. RIA analysis was used to determine pro-CCK immunoreactivity. Elution of the Pro CCK standard is indicated.

also detected in positive control samples (WE, GH3, ␤TC-3 and ␤-STC-1) with a larger 75 kDa band, which is the proform of the enzyme, and a smaller 66 kDa band for active PC2. Previously published PC1 and PC2 Western blots of rodent tissues and cell lines yielded the same sized proteins [16,35]. Additionally, negative control samples for PC1 (WE, GH3, NTN, and NT2) and PC2 (NTN and NT2) did not reveal any protein bands for the pro- or the active forms of the two enzymes (data not shown). 3.3. Gel filtration (GCL-90) chromatography of CCK peptides in neuronal cell lines overexpressing pro-CCK The cleavage of pro-CCK by PC5 was evaluated by separating media and cell lysate samples with low pressure chromatography and by using RIAs specific for the detection of glycine-extended and amidated forms of CCK. ProCCK RIA results showed that the transfected cell lines (GT1–7, SK-N-SH, and SK-N-MC) had high levels of proCCK in the media (Fig. 3 and Table 2A). The cell lysate samples of the cultured neuronal cells also had high proCCK content, but the media samples had 3– 4 fold higher levels than the cell lysates (data not shown). Additionally, the effect of the CpB treatment on pro-CCK levels for most of the samples appeared to be negligible and decreases in the prohormone levels following CpB treatment were attributed to a loss of sample (low recovery) following the SepPak treatment. No amidated CCK was detected in cell lysate or media samples. Moderate amounts of CCK-Gly immunoreactive material was detected in the media of the cell lines (Fig. 4 and Table 2B). Cell lysate samples had measurable CCK-Gly immunoreactive forms of the peptide, but the media had 3– 4 fold higher levels than the lysates. These glycineextended, intermediate forms in all of the cell lines were

1276

B.M. Cain et al. / Peptides 22 (2001) 1271–1277

Table 2a Pro-CCK immunoreactivity from media samples of transfected mammalian neuronal cell lines Neuronal cell linea,b

Pro CCK ⫺ CpB (pg/ml)

Pro CCK ⫹ CpB (pg/ml)

GT1–7 SK-N-SH SK-N-MC

9190.41 8076.15 6077.38

9740.79 9501.73 4259.48

Table 2b CCK Gly immunoreactivity from media samples of transfected mammalian neuronal cell lines Neuronal cell linea,b

CCK Gly ⫺ CpBc (pg/ml)

CCK Gly ⫹ CpBc (pg/ml)

GT1–7 SK-N-SH

465.97 230.95 233.53 534.92 625.07

675.69 591.25 514.49 442.75 536.16

SK-N-MC

(Fraction (Fraction (Fraction (Fraction (Fraction

53) 45) 49) 43) 52)

(Fraction (Fraction (Fraction (Fraction (Fraction

54) 44) 47) 44) 54)

a Each media sample data point was harvested from n ⫽ 6 10 cm2 culture dishes containing 5.0 ml of selective media. b The total volume for each data point was 30.0 ml. The cells were 80 –100% confluent at the time of harvesting. c Fraction numbers correspond to fractions collected from the low pressure gel filtration chromatography column.

small in size and, based on protein standards (pro-CCK, CCK 30, CCK 22 Gly, CCK 22, CCK 12 Gly, and CCK 8 Gly), the forms generated in the neuronal cells co-eluted with CCK 22 Gly and CCK 12 Gly (Fig. 4). In the case of CCK-Gly immunoreactivity, the CpB treatment increased the levels of the glycine-extended intermediate forms in most of the media samples (Table 2B). 4. Discussion Three neuronal cell lines, which were found by RT-PCR and Western blotting to express PC5, but not PC1 or PC2,

were engineered to overexpress pro-CCK. None of these cell lines made much amidated CCK before or after stable transfection. They all made and secreted moderate amounts of glycine-extended CCK peptides that were detected by RIA. When treated with carboxypeptidase B, the amount of these peptides increased, indicating that carboxypeptidase activity in these cells was partially deficient. That these cells were unable to produce amidated CCK peptides from these glycine-extended peptides suggests that the amidating enzyme was either absent or not very active. The major forms of glycine-extended CCK peptides secreted into the media by these cells co-eluted on gel-filtration with synthetic CCK 12 Gly and CCK 22 Gly. This indicates that these cell lines are capable of processing pro-CCK into forms that could be converted into active, amidated forms (CCK 12 and CCK 22 amide) by the action of carboxypeptidase E and PAM. These results support the hypothesis that PC5 acting alone is able to appropriately cleave pro-CCK. The possibility that other undiscovered enzymes in these cells are responsible for these cleavages cannot be excluded, but would require further antisense or inhibitor studies. These results suggest that PC5 may (with the help of PC1 and PC2) participate in the processing of pro-CCK in the brain. PC5 is co-localized with oxytocin in the supraoptic and paraventricular nuclei of the hypothalamus, cell populations which also contain CCK. Information about their co-localization in other brain regions is unavailable. However, the forms of glycine-extended CCK produced by these cells are more reminiscent of pro-CCK processing in intestine than in brain. This is entirely reasonable as PC5 is very abundant in the intestine, much higher than PC1 or PC2, although information of the co-localization of CCK with any of these enzymes in the intestine is also unavailable. Additionally, the role of PC5B, the larger PC5 isoform found in the gut only, regarding the processing of pro-CCK in the intestine has not been investigated. In summary, these results demonstrate that these cell lines which express PC5, but not PC1 or PC2, can appropriately cleave pro-CCK into glycine-extended forms. This result supports a role of PC5 in pro-CCK processing in the brain and intestine.

Acknowledgments

Fig. 4. Gel filtration chromatography (GCL-90) of media samples from various neuronal cell lines transfected with CCK and treated with CpB. RIA analysis was used to determine CCK-Gly immunoreactivity. Elution of CCK 22 and CCK 12 Gly standards are indicated.

This work was supported by NIH grant NS31602. Polyclonal antibodies recognizing PC1 and PC2 proteins were generously provided by Dr. Iris Lindberg from the Louisiana State Medical Center. Polyclonal antibodies recognizing the amino- and carboxyl-terminals of the PC5 protein were generously provided by Dr. George Greeley of the University of Texas Medical Branch at Galveston.

B.M. Cain et al. / Peptides 22 (2001) 1271–1277

References [1] Adler G, Beglinger C, Braun U, Reinshagen M, Koop I, Schafmayer A, Rovati L, Arnold R. Interaction of the cholinergic system and cholecystokinin in the regulation of endogenous and exogenous stimulation of pancreatic secretion in humans. Gastroenterology 1991; 100:537– 43. [2] Barbero P, Rovere C, De Bie I, Seidah NG, Beaudet A, Kitabgi P. PC5-A-mediated processing of pro-neurotensin in early compartments of the regulated secretory pathway of PC5-transfected PC12 cells. J Biol Chem 1998;273:25339 – 46. [3] Beak SA, Heath MM, Small CJ, Morgan DGA, Ghatei MA, Taylor AD, Buckingham JC, Bloom SR, Smith DM. Glucagon-like Peptide-1 stimulates Luteinizing Hormone-Releasing Hormone secretion in a rodent hypothalamic neuronal cell line. J Clin Invest 1998;101: 1334 – 41. [4] Beinfeld M.C. CCK mRNA expression, pro-CCK processing, and regulated secretion of immunoreactive CCK peptides by rat insulinoma (RIN 5F) and mouse pituitary tumor (AtT-20) cells in culture. Neuropeptides 1992;2:213–17. [5] Beinfeld MC, Meyer DK, Eskay RL, Jensen RT, Brownstein MJ. The distribution of cholecystokinin immunoreactivity in the central nervous system of the rat as determined by radioimmunoassay. Brain Res 1981;212:51–7. [6] Blache P, Le-Nguyen D, Boegner-Lemoine C, Cohen-Salal A, Bataille D, Kervran A. Immunological detection of prohormone convertases in two different proglucagon processing cell lines. FEBS Lett 1994;344:65– 8. [7] Campan M, Yoshizumi M, Seidah NG, Lee M, Bianchi C, Haber E. Increased proteolytic processing of protein tyrosine phosphatase mu in confluent vascular endothelial cells: The role of PC5, a member of the subtilisin family. Biochemistry 1996;35:3797–3802. [8] Chomczynski P, Sacchi N. Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal Biochem 1987;162:156 –9. [9] Deschenes RJ, Lorenz LJ, Haun RS, Roos BA, Collier KJ, Dixon JE. Cloning and sequence analysis of a cDNA encoding rat precholecystokinin. Proc Natl Acad Sci 1984;81:726 –30. [10] Dong W, Marcinkiewicz M, Vieau D, Chretien M, Seidah NG, Day R. Distinct mRNA expression of the highly homologous convertases PC5 and PACE4 in the rat brain and pituitary. J Neurosci 1995;15: 1778 –96. [11] Dotsch J, Hanze J, Beste O, Behrendt J, Weber WM, Dittrich K, Rascher W. Suppression of neuropeptide Y1 receptor function in SK-N-MC cells by nitric oxide. Am J Physiol 1997;273:C618 –21. [12] Gridley KE, Green PS, Simpkins JW. Low concentrations of estradiol reduce ␤-amyloid (25–35)-induced toxicity, lipid peroxidation, and glucose utilization in human SK-H-SH neuroblastoma cells. Brain Res 1997;778:158 – 65. [13] Hilsted L, Rehfeld JF. Measurement of precursors for ␣ amidated hormones by radioimmunoassay of glycine extended peptides after trypsin-carboxypeptidase B cleavage. Anal Biochem 1986;152:119 – 26. [14] Hunter WM, Greenwood FC. Preparation of Iodine-131 labeled human growth hormone of high specific activity. Nature 1962;194: 495– 6. [15] Laframboise M, Reudelhuber TL, Jutras I, Brechler V, Seidah NG, Day R, Gross KW, Deschepper CF. Prorenin activation and prohormone convertases in the mouse As4.1 cell line. Kidney Internatl 1997;51:104 –9. [16] Lamango NS, Zhu XR, Lindberg I. Purification and enzymatic characterization of recombinant prohormone convertase 2: Stabilization of activity by 21 kDa 7B2. Arch Biochem Biophys 1996;330:238 –50.

1277

[17] Listwak SJ, Gold PW, Whitfield Jr HJ. The human mineralocorticoid receptor gene promoter: Its structure and expression. J Steroid Biochem Molec Biol 1996;58:495–506. [18] Lowry OH, Rosenbrough NJ, Farr AL, Randall RJ. Protein measurement with the folin phenol reagent. J Biol Chem 1951;193:265–75. [19] Lusson J, Vieau D, Hamelin J, Day R, Chretien M, Seidah NG. cDNA structure of the mouse and rat subtilisin/kexin-like PC5: A candidate proprotein convertase expressed in endocrine and nonendocrine cells. Proc Natl Acad Sci 1993;90:6691–5. [20] Nachtigal MW, Ingraham HA. Bioactivation of Mullerian inhibiting substance during gonadal development by a Kex2-like endoprotease. Proc Natl Acad Sci, USA 1996;93:7711–16. [21] Nakagawa T, Hosaka M, Torii S, Watanabe T, Murakami K, Nakayama K. Identification and functional expression of a new member of the mammalian kex2-like processing endoprotease family: its striking structural similarity to PACE4. J Biochem 1993;113:132–5. [22] Nakagawa T, Murakami K, Nakayama K. Identification of an isoform with an extremely large Cys-rich region of PC6, a Kex2-like processing endoprotease. FEBS Lett 993;327:165–71. [23] Shen F-S, Seidah NG, Lindberg I. Biosynthesis of the prohormone convertase PC 2 in chinese hamster ovary cells and in rat insulinoma cells. J Biol Chem 1993;268:24910 –15. [24] Smeekens SP, Avruch AS, LaMendola J, Chan SJ, Steiner DF. Identification of a cDNA encoding a second putative prohormone convertase related to PC2 in AtT-20 cells and islets of Langerhans. Proc Natl Acad Sci, USA 1991;88:340 – 4. [25] Smeekens SP, Steiner DF. Identification of a human insulinoma cDNA encoding a novel mammalian protein structurally related to the yeast dibasic processing protease KEX2. J Biol Chem 1990;265: 2997–3000. [26] Villeneuve P, Seidah NG, Beaudet A. Immunohistochemical distribution of the prohormone convertase PC5-A in rat brain. J Neurosci 1999;92:641– 654. [27] Vindrola L, Lindberg I. Biosynthesis of the prohormone convertase mPC1 in AtT-20 cells. Mol Endo 1992;6:1088 –94. [28] Wang W, Birch NP, Beinfeld MC. Prohormone convertase 1 (PC1) when expressed with pro cholecystokinin (pro CCK) in L cells performs three endoproteolytic cleavages which are observed in rat brain and in CCK-expressing endocrine cells in cell culture, including the production of glycine and arginine extended CCK 8. Biochem Biophys Res Comm 1998;248:538 – 41. [29] Wang W, Cain BM, Beinfeld MC. Adult carboxypeptidase E-deficient fat/fat mice have a near total depletion of brain CCK 8 accompanied by a massive accumulation of glycine and arginine extended CCK: Identification of CCK 8 Gly as the intermediate precursor of CCK 8 in rodent brain. Endocrine 1998;9:329 –32. [30] Wang W, Yum L, Beinfeld MC. Expression of rat pro cholecystokinin (CCK) in bacteria and in insect cells infected with recombinant baculovirus. Peptides 1997;18:1295–9. [31] Yoon JY, Beinfeld MC. A mouse intestinal tumor cell line, STC-1, expresses CCK, PC1, and PC2 mRNA, processes pro CCK to CCK 8, and displays cAMP regulated release. Endocrine 1994;2:973–7. [32] Yoon JY, Beinfeld MC. Prohormone convertase 1 is necessary for the formation of cholecystokinin 8 in Rin5F and STC-1 cells. J Biol Chem 1997;272:9450 – 6. [33] Yoon JY, Beinfeld MC. Prohormone convertase 2 is necessary for the formation of cholecystokinin-22, but not cholecystokinin-8, in RIN5F and STC-1 cells. Endocrinology 1997;138:3620 –3. [34] Zheng M, Seidah NG, Pintar JE. The developmental expression in the rat CNS and peripheral tissues of proteases PC5 and PACE4 mRNAs: Comparison with other proprotein processing enzymes. Dev Biol 1997;181:268 – 83. [35] Zhou Y, Lindberg I. Purification and characterization of the prohormone convertase PC1 (PC3). J Biol Chem 1993;268:5615–23.