If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Molecular Identification and Characterization of Novel Membrane-bound Metalloprotease, the Soluble Secreted Form of Which Hydrolyzes a Variety of Vasoactive Peptides*
* This study was supported by grants from the Ministry of Education, Science, Sports and Culture of Japan.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.The nucleotide sequence(s) reported in this paper has been submitted to the GenBank™/EMBL Data Bank with accession number(s) and (SEP and SEPΔ, respectively).
One class of zinc metalloproteases, represented by neutral endopeptidase 24.11 and endothelin-converting enzyme, has been shown to be involved in proteolytic activation or inactivation of many regulatory peptides. Here, we report molecular cloning and characterization of a novel member of this type II membrane-bound metalloprotease family, termed soluble secreted endopeptidase (SEP). Alternative splicing results in the generation of another transcript, SEPΔ, which lacks a 69-base pair nucleotide segment following the transmembrane helix. Both SEP and SEPΔmRNA are detected in all mouse tissues examined. Transfection of an SEP cDNA expression construct resulted in the expression of the membrane-bound form of SEP in the early secretory pathway as well as the soluble secreted form of the enzyme in the culture medium. In contrast, transfection of the SEPΔ cDNA only results in the expression of the membrane-bound form. In vitroenzymological analysis of the recombinant soluble form of SEP demonstrated that it hydrolyzes a variety of vasoactive peptides, including endothelin-1, atrial natriuretic peptide, and angiotensin I. This activity of SEP was inhibited by phosphoramidon and the neutral endopeptidase 24.11 specific inhibitor thiorphan, but it was only partially inhibited by the endothelin-converting enzyme specific inhibitor FR901533. These findings suggest that SEP is a novel metalloprotease that possesses a broad substrate specificity and that it may be involved in the metabolism of biologically active peptides intracellulary as well as extracellularly.
NEP
neutral endopeptidase 24.11
ECE
endothelin-converting enzyme
SEP
soluble secreted endopeptidase
ET
endothelin
ANP
atrial natriuretic peptide
CHO
Chinese hamster ovary
RT
reverse transcription
PCR
polymerase chain reaction
HPLC
high pressure liquid chromatography
PBS
phosphate-buffered saline
ACE
angiotensin-converting enzyme
Endo H
endo-β-N-acetylglucosaminidase H
PNGase F
peptide-N-glycosidase F
A wide variety of biologically active peptide hormones, regulatory peptides, and neuropeptides have been shown to be proteolytically activated or inactivated by members of zinc metalloproteases (
). One such class of zinc metalloproteases, represented by neutral endopeptidase 24.11 (NEP)1and endothelin-converting enzyme (ECE), has recently been highlighted because of their implications in some disease states and should thus provide plausible therapeutic targets for certain diseases (
). In mammals, six members of this metalloprotease family have been identified: NEP; Kell blood group antigen (KELL); ECE-1 and ECE-2; PEX, which has been associated with X-linked hypophosphatemic rickets; and the recently identified “orphan” peptidase XCE. All these members are type II integral membrane proteins containing a highly conserved consensus sequence of a zinc-binding motif, HEXXH (whereX represents any amino acid), in their extracellular C-terminal domain. Despite the apparent structural similarity among the members of this family, a large diversity of physiological functions exists.
NEP, which is especially abundant in kidney and brain, is also expressed in various tissues as an ectoenzyme that can degrade many circulating small peptide mediators, such as enkephalins, atrial natriuretic peptide (ANP), tachykinins, and endothelins (ETs) (
). NEP is also known as the common acute lymphoblastic leukemia antigen, and its presence on leukemic cells has been associated with a better prognosis. Although the physiological substrates of NEP are still unknown, targeted disruption of the NEP gene in mice caused a dramatic sensitivity to endotoxin shock, suggesting that NEP may provide an unexpected protective role against endotoxin shock (
). Moreover, in vivo pharmacological inhibition of NEP has led to a decrease in blood pressure, and NEP-deficient mice were noted to have lower mean blood pressure levels than wild-type littermates, thus indicating that NEP may also play an important role in blood pressure regulation (
ECE, another well characterized member of this metalloprotease family, is involved in the regulation of vascular tone, as well as in the development of some sets of neural crest cells (
). ECE constitutes a potential regulatory site for the production of the active peptide. Two isozymes of ECE, ECE-1 and ECE-2, have been molecularly identified and make up a subfamily within this group of type II membrane-bound metalloproteases (
). Both enzymes have been shown to cleave big ET-1 to produce ET-1 with a similar overall profile of inhibitor sensitivity in vitro as well as in transfected cells. However, ECE-1 and ECE-2 exhibit the following striking differences: (i) ECE-1 cleaves big endothelins in neutral pH, whereas ECE-2 functions in an acidic pH range, (ii) the sensitivity of ECE-1 to phosphoramidon is 250-fold lower as compared with ECE-2, and (iii) ECE-1 is abundantly expressed in endothelial cells and other cell types known to produce mature ET-1, whereas ECE-2 mRNA is detected in neural tissues including the cerebral cortex, the cerebellum, and the adrenal medulla. Targeted disruption of the ECE-1 gene in mice revealed that ECE-1 is the physiologically relevant enzyme needed to produce active ET-1 (
). The physiological function of ECE-2 has not yet been elucidated.
The physiological substrates of the three other mammalian peptidases, KELL, PEX, and XCE, are still unknown. KELL, expressed on human red cells and other cell types, carries the epitopes for the Kell minor blood group antigen (
). Although the Kell blood group antigen is clinically important, its actual protease activity has yet to be described. The PEX gene was identified by positional cloning as a candidate gene for X-linked hypophosphatemic rickets, a dominant disorder characterized by impaired phosphate uptake in the kidney (
). No protease activity for XCE has been detected; hence, its physiological significance is unknown.
Recent gene targeting studies of ECE revealed that ECE-1 is a bona fide activating protease for big ET-1 and big ET-3 at specific developmental stages (
). However, despite the absence of ECE-1 (which resulted in craniofacial and cardiovascular defects), a significant amount of mature ET-1 peptide was still found inECE-1−/− embryos, suggesting that other proteases can activate ET-1. This consideration led us to search for structurally related enzymes of this metalloprotease family. Here, we report the isolation of a novel enzyme, termed soluble secreted endopeptidase (SEP), by degenerate PCR using cDNA prepared fromECE-1−/− embryos as template. The cDNA sequence predicts that SEP is a type II membrane-bound metalloprotease structurally related to NEP, ECE-1, and ECE-2. Transfection of the SEP cDNA into Chinese hamster ovary (CHO) cells resulted in the presence of SEP protein not only in the membrane fraction of the cells but also in the supernatant, suggesting that these cells release soluble forms of the enzyme through proper secretory machinery. Enzymological analysis of the recombinant soluble SEP protein revealed that SEP hydrolyzes a variety of peptides, which are known as substrates of NEP and/or ECE, including big ET-1, ET-1, angiotensin I, ANP, bradykinin, and substance P. This suggest that SEP is likely a novel member of this metalloprotease family and may be involved in the metabolism of biologically active peptides.
EXPERIMENTAL PROCEDURES
Reagents
Synthetic human big ET-1 (1–38), human ET-1, ANP, angiotensin I, bradykinin, and substance P were obtained from American Peptides. Phosphoramidon, thiorphan, 1,10-phenanthroline, and captopril were obtained from Sigma. FR901533 (WS79089B) was a gift from Fujisawa Pharmaceutical Co., Ltd.
cDNA Cloning and Sequencing
Near-termECE-1−/− embryos were kindly provided by Dr. Masashi Yanagisawa (University of Texas Southwestern Medical Center, Dallas, TX). A partial cDNA clone encoding SEP was obtained by reverse transcription (RT)-PCR against wholeECE-1−/− embryo mRNA with degenerate primers based on the highly conserved amino acid sequences of ECE-1, ECE-2, NEP, and PEX cDNAs (see Fig. 2). The PCR contained 60 mm Tris-Cl (pH 8.5), 15 mm ammonium sulfate, 1.5 mm magnesium chloride, 0.25 mm of each dNTP, 7.5 pmol of each degenerate primer, 5′-AT(A/G/C/T)GT(A/G/C/T)TT(C/T)CC(A/G/C/T)GC(A/T)GG-3′ and 5′-T(AG)TC(A/G/C/T)GC(A/G/T)AT(A/G)TT(C/T)TC-3′, 10 ng of first strand cDNA, and 2.5 units of Taq DNA polymerase. The initial five cycles were carried out at an annealing temperature of 37 °C, and then 35 more cycles were carried out at an annealing temperature of 48 °C. The PCR products were separated in a 1% agarose gel, and an approximately 300-base pair region was excised from the gel. The extracted DNA was subcloned into pT7 vector (Novagen) and sequenced. A cDNA library was constructed by using the SuperScript kit (Life Technologies, Inc.) against poly(A)+ RNA from mouse testis. Approximately 1 × 106 plaques from the unamplified library were screened with random primed 32P-labeled RT-PCR product as probe. The 5′ end of the cDNA was cloned by 5′-rapid amplification of cDNA ends (Life Technologies, Inc.) against theECE-1−/− embryo and mouse brain. The first strand cDNA was synthesized with reverse transcriptase by using a specific primer 5′-TCAGGTCCATTCGGTGGTACAGGGC-3′ (corresponding to amino acids 293–301 of SEP). With terminal deoxynucleotidyltransferase, an oligo(dC) anchor was added to the 3′ end of the first strand cDNA. The first round PCR was performed as recommended by the manufacturer with a specific 3′ primer, 5′-GACATCATGCCTTTTCTCCTGGGGG-3′ (corresponding to amino acids 283–291 of SEP) and a 5′ anchor primer. The product was then subjected to the second amplification by using a nested specific 3′ primer, 5′-ACTCCCGGGATGGCATGCCCAAGGT-3′ (corresponding to amino acids 218–226 of SEP). The product of this PCR was subcloned into pT7 vector and subsequently sequenced. For nucleotide sequencing, overlapping restriction fragments of the cDNA were subcloned into the pBluescript plasmid vector (Stratagene), and double strand plasmid DNA sequenced by a model 310 DNA Sequencer (Applied Biosystems). Both strands of cDNA were covered at least twice.
Figure 2Alignment of protein sequences from mouse SEP, NEP, ECE-1, ECE-2, XCE, and PEX. The SEP sequence, isolated by degenerate PCR, was aligned with other members of the metalloprotease family. All sequences used were from the mouse. The zinc-binding motif is marked by a closed box. The sequence used for designing degenerate RT-PCR primers is marked byasterisks.
First strand cDNA synthesis was carried out with 5 μg of total RNA from mouse tissues and oligo(dT)12–18 primers using SuperScript reverse transcriptase II (Life Technologies, Inc.) as recommended by the manufacturer. The PCR contained 20 mm Tris-Cl (pH 8.5), 15 mm ammonium sulfate, 1.5 mm magnesium chloride, a 0.25 mm concentration of each dNTP, a 100 nmconcentration of each amplification primer, 10 ng of first strand cDNA, and 2.5 units of Taq polymerase. The primers, 5′-GGGAGCCATAGTGACTCTGGGTGTC-3′ (corresponding to amino acids 28–36 of SEP) and 5′-GCTATCACACAGCTTGGGGTGGTGC-3′ (corresponding to amino acids 75–83 of SEP) were used for both spliceoforms of mouse SEP. The PCR products were verified by DNA sequencing.
Antibody and Immunoblotting
Antibody directed against SEP was produced by immunizing rabbits with a synthetic peptide, CPRGSPMHPMKRCRIW, corresponding to the C-terminal 16 amino acids of mouse SEP. Rabbits were immunized with keyhole limpet hemocyanin-coupled peptides in complete adjuvant, followed by antisera preparation. Immunoblot analysis was performed with horseradish peroxidase-conjugated anti-rabbit IgG by using the ECL detection kit (Amersham Pharmacia Biotech) as recommended by the manufacturer.
Endoglycosidase Digestion
Partially purified SEP and the membrane preparations from CHO/SEP and CHO/SEPΔ cells were incubated with either 1 unit of endo-β-N-acetylglucosaminidase or peptide-N-glycosidase F (Roche Molecular Biochemicals) in 50 mm sodium phosphate buffer (pH 5.5) at 37 °C for 16 h. Control samples were incubated in parallel without endoglycosidases in the identical buffer at 37 °C for 16 h. The samples were then subjected to immunoblotting.
). Double transfection of prepro-ET-1 and ECE-1a, SEP, or SEPΔ was performed using LipofectAMINE (Life Technologies, Inc.). Twelve hours after transfection, cells were refed with fresh medium. The medium was conditioned for an additional 18 h and was directly subjected to a sandwich-type enzyme immunoassay that showed no cross-reactivity between big ET-1 and ET-1. For in vitro enzymologic characterization, an SEP expression construct was transiently transfected into CHO-K1 cells. The medium was conditioned for 48 h in CHO-SFM II (Life Technologies, Inc.) and was subjected to a wheat germ lectin column (1 × 1 ml HiTrap wheat germ lectin; Amersham Pharmacia Biotech) equilibrated with 20 mm Tris-Cl (pH 7.4) and 0.5 m NaCl. The column was washed and eluted with the same buffer containing 0.5 mN-acetylglucosamine. The active fraction was subsequently subjected to a Centriprep concentrator (Amicon).
Fluorescent Immunocytochemistry
Cells were seeded onto coverslips and cultured for 2 days. Fluorescent immunocytochemistry was performed as described previously (
). Briefly, for intracellular staining, cells were fixed and permeabilized in methanol for 5 min at −20 °C. After washing in phosphate-buffered saline (PBS), PBS containing 10% (v/v) normal goat serum was added. Following incubation for 1 h at 37 °C, the normal goat serum/PBS was replaced with buffer containing polyclonal antibody (1:100) directed against bovine SEP C-terminal peptides. After incubation for 90 min at 37 °C, the cells were washed and then incubated in normal goat serum/PBS containing 7.5 μg/ml fluorescein isothiocyanate-labeled goat anti-rabbit IgG (Zymed Laboratories Inc.). After 45 min at 37 °C, the cells were extensively washed. The coverslips were mounted on microscope slides with 90% (v/v) glycerol, 50 mm Tris-HCl (pH 9.0), and 2.5% (w/v) 1,4-diazadicyclo-[2.2.2]-octane. For cell surface staining, cells were fixed in PBS containing 4% paraformaldehyde for 15 min at room temperature. Following two washes in PBS, the cells were treated with the SEP antibody and fluorescein isothiocyanate-labeled goat anti-rabbit IgG as described above. Three negative control conditions were examined: staining with preimmune serum, with antibody after preabsorption, and omission of the primary antibody. None of these conditions resulted in cell staining.
Measurement of Enzyme Activity
Reaction mixtures for enzyme assay (100 μl) contained 0.1 m MES-NaOH (pH 7.4), 0.5m NaCl, 0.5 μm peptide, and an enzyme fraction. For some experiments, the reactions were preincubated at 37 °C with various protease inhibitors for 15 min prior to the addition of the peptide. The reactions were incubated at 37 °C for 1–12 h in siliconized 0.5-ml microcentrifuge tubes. The enzyme reaction was terminated by adding 1 μl of 10 mm EDTA. The mixture was then injected into a C18 reverse-phase high pressure liquid chromatography (HPLC) column (μRPC C2/C18, ST4.6/100; Amersham Pharmacia Biotech) that was equilibrated with 10% (v/v) acetonitrile and 0.1% (v/v) trifluoroacetic acid. The column was eluted at a flow rate of 1 ml/min with a 10–80% linear gradient of acetonitrile and 0.1% trifluoroacetic acid over 43 min, followed by a 100% acetonitrile for an additional 3 min. Peptides were detected by absorbance at 220 nm.
Mass Spectrometry
Matrix-assisted laser desorption/ionization mass spectra were acquired on Voyager RP delayed extraction mass spectrometer (PerSeptive Biosystems, Inc.). Radiation from a nitrogen laser (Laser Science, Inc.) (337 nm, 3-ns pulse width) was used to desorb ions from the target. All reflector delayed extraction experiments were performed using an extraction grid voltage of 14.5 kV and a pulse delay of 225 ns.
RESULTS
Cloning of SEP
A pair of highly degenerate oligonucleotide primers was designed based on conserved amino acid sequences of known members of the membrane-bound metalloprotease family: ECE-1, ECE-2, NEP, and PEX (see Fig. 2). Subsequent RT-PCR from near-term wholeECE-1−/− mouse embryo RNA yielded cDNA products of the predicted size. We subcloned these cDNA fragments into plasmid vectors and determined the nucleotide sequence. The sequences from randomly picked plasmid clones revealed that the 300-base pair cDNA product was a mixture of two distinct cDNA sequences: most plasmid clones encoded mouse NEP, whereas the nucleotide sequence from one clone predicted a closely related polypeptide sequence to members of this metalloprotease family. We named this novel putative metalloprotease SEP.
Using the cloned SEP RT-PCR product as a probe, we screened a mouse testis cDNA library, because SEP mRNA was most abundantly expressed in testis (see Fig. 3B). We purified and sequenced several positive clones, with the overlapping nucleotide sequences of all these clones confirmed as being identical. The nucleotide sequences of the longest SEP cDNAs had a 5′ ATG triplet codon that was preceded by an in-frame stop codon and followed by a long open reading frame. The encoded amino acid sequence of SEP is shown in Fig.1.
Figure 3Expression of SEP and SEPΔ.A, schematic representation of the SEP and SEPΔ cDNAs. Closed box, predicted transmembrane domains; hatched box, a 69-base pair segment unique to SEP; arrows, PCR primers used to amplify the SEP and SEPΔ cDNAs. B, RT-PCR of mouse tissue RNA. Total RNA from brain, heart, lung, liver, kidney, adrenal, intestine, testis, and wholeECE-1−/− embryo were reverse-transcribed and PCR-amplified using the primers depicted above. An aliquot of each PCR was electrophoresed and visualized with ethidium bromide.
Figure 1Nucleotide and deduced amino acid sequence encoded by the mouse SEP cDNA. The in-frame stop codon that precedes the 5′ ATG codon is marked by an asterisk. Sequences used for designing degenerate RT-PCR primers are marked bydots. An open box designates the predicted transmembrane domain. The sequence that is not present in SEPΔ is underlined. The zinc-binding motif is marked by a closed box. ‡, predictedN-glycosylation sites.
While screening to isolate a full-length SEP cDNA, we performed 5′-rapid amplification of cDNA ends on RNA fromECE-1−/− embryo, mouse brain, and mouse testis using a nested set of specific internal primers to assess the 5′ diversity of SEP mRNA. This yielded two products. The nucleotide sequence of one product was identical to the 5′-end of the full-length SEP cDNA isolated by screening. However, the sequencing results of other clones revealed that they contained cDNA derived from SEP mRNA but lacked a 69-base pair nucleotide segment immediately following the putative transmembrane helix, presumably due to alternative splicing (Figs. 1 and 3A). Based on these findings, we termed this clone SEPΔ.
Structure of SEP and SEPΔ
The SEP cDNA sequence encodes a novel 765-amino acid polypeptide, which shares important structural features with the NEP metalloprotease family (
). (i) The cDNA predicts a type II integral membrane protein with a 17-residue N-terminal cytoplasmic tail, a 21-residue putative transmembrane helix (Fig. 1, open box), and a large (727 residue) extracellular C-terminal part. (ii) The extracellular portion of SEP constitutes the putative catalytic domain and contains a highly conserved consensus sequence (residues 597–605) of a zinc-binding motif, ØXHEØØHØΨ (where Ø and Ψ represent an uncharged and a hydrophobic amino acid, respectively), that is shared by many metalloproteases. (iii) SEP has nine predicted sites forN-glycosylation in the extracellular domain, suggesting that SEP is a highly glycosylated protein, like NEP and ECE. (iv) There are 10 Cys residues in the extracellular domain that are conserved in all the proteins of this metalloprotease family.
A search using the Entrez sequence data base pointed out a significant similarity of the SEP sequence to NEP, ECE-1, ECE-2, XCE, and PEX (Fig.2). The sequence similarity is especially high within the C-terminal one-third of the putative extracellular domain, including the region around the zinc-binding motif. Within this region (amino acids 511–765 of SEP), the identities of mouse SEP with respect to mouse NEP, human ECE-1, bovine ECE-2, and human XCE are 65.1, 47.7, 44.5, and 46.1%, respectively.
Tissue Distribution of SEP and SEPΔmRNA
Northern blot analysis using total RNA from a variety of mouse tissues revealed relatively large amounts of 3.8-kilobase SEP mRNA in testis (data not shown). Small amounts of SEP mRNA were also expressed in the ovary. No signal was observed in other tissues, including the brain, lung, heart, liver, kidney, adrenal gland, and intestine. We then examined the expression of SEP and SEPΔ mRNA in various mouse tissues by RT-PCR using primers to amplify both subisoforms of SEP (Fig.3A). Two fragments corresponding to SEP and SEPΔ (134 and 65 base pairs, respectively) were detected in all tissues examined as well as in theECE-1−/− embryo (Fig. 3B). Although RT-PCR is not strictly quantitative, the data suggest that SEP is the major isoform in testis, whereas SEPΔ is predominantly expressed in other tissues.
Expression of SEP in Eukaryotic Cells
To characterize the properties of cloned SEP and SEPΔ, we generated transfectant cells, CHO/SEP and CHO/SEPΔ, by transiently transfecting expression constructs driven by the SRα viral promoter (
). Immunoblot analysis with an anti-SEP C-terminal peptide antiserum showed that both SEP and SEPΔ proteins are expressed as an approximate 110-kDa protein in the membrane preparation from these cells (Fig. 4). In addition, we detected appreciable amounts of SEP-immunoreactive material with an apparent molecular mass of approximately 126 kDa in culture medium conditioned with CHO/SEP cells, suggesting that these cells release soluble forms of SEP into the culture medium (Fig. 4). In contrast, we did not detect SEP immunoreactivity in the conditioned media of both CHO/SEPΔ and untransfected CHO cells. These observations demonstrate that CHO/SEP cells express the 110-kDa membrane-bound form of SEP as well as the 126-kDa soluble form of SEP in the medium, whereas CHO/SEPΔ cells express only the membrane-bound form.
Figure 4Immunoblot analysis of supernatant and membrane fractions from CHO/SEP, CHO/SEPΔ, and control cells. CHO cells were transiently transfected with an expression construct containing either the SEP or SEPΔ cDNAs or an empty vector. Supernatant (S) and membrane (M) fractions from each cell were separated on a 7.5% SDS-polyacrylamide gel under reduced conditions, blotted, and detected using anti-C-terminal peptide antisera against SEP.
Because cDNA cloning of SEP revealed that it is a highly glycosylated protein, we assumed that the variation in the apparent molecular mass observed on immunoblot analysis was largely due to the presence of sugar moieties. To analyze the sugar side chains of SEP, we examined the sensitivity of both the 110-kDa membrane-bound SEP and the 126-kDa soluble forms of SEP to endo-β-N-acetylglucosaminidase H (Endo H) and peptide-N-glycosidase F (PNGase F) (Fig.5). Proteins containing high mannose sugar moieties, found in the early secretory pathway, including the endoplasmic reticulum and a portion of the Golgi apparatus, are sensitive to both Endo H and PNGase F. In contrast, proteins of which the sugar side chains have been further modified to complex oligosaccharides, which occurs in the Golgi apparatus, are sensitive to PNGase F but resistant to Endo H. Treatment of solubilized membranes from CHO/SEP cells with either Endo H or PNGase F reduced the apparent molecular mass from 110 to 89 kDa, which corresponds to the calculated molecular mass of SEP. These observations suggest that the 110-kDa species observed in the membrane fraction of the cells is the partially glycosylated protein present in the early secretory pathway. In contrast, although PNGase F treatment of the conditioned medium reduced the size from 126 to 89 kDa, Endo H had no effect on these species, indicating that the SEP protein in conditioned medium is resistant to Endo H. These observations suggest that the presence of the SEP protein in the medium is due to its secretion after complete glycosylation during transit through the Golgi apparatus. Taken together, these results suggest that CHO/SEP cells express membrane-bound SEP protein in the membrane of the compartments along the early secretory pathway in the cell and also secrete a soluble form of protein in the culture medium through proper secretory machinery.
Figure 5Deglycosylation of soluble and membrane-bound forms of SEP . The soluble form of SEP in the supernatant as well as the membrane-bound SEP were incubated with either Endo H (H) or PNGase F (F) as described under “Experimental Procedures.” Control samples (−) were incubated in parallel without endoglycosidases using an identical buffer. The samples were then subjected to immunoblotting.
To examine the subcellular localization of the membrane-bound SEP protein, we immunostained both CHO/SEP and CHO/SEPΔ cells with antibodies that recognize the common C-terminal ectodomain of SEP (Fig. 6). Without permeabilization, both cells stained faintly, indicating that these cells expressed little SEP on the cell surface (Fig. 6, B and D). After permeabilization, both cells showed strong intracellular staining (Fig.6, A and C). These findings indicate that the membrane-bound SEP expressed in CHO cells appears to be located inside the cells, presumably in the early secretory pathway, which is compatible with its sensitivity to Endo H.
Figure 6Fluorescence immunocytochemistry of CHO cells transfected with either the SEP or SEPΔcDNAs. CHO cells transfected with either the SEP or SEPΔ expression construct were stained for intracellular (intra.) (A, C, and E) or cell surface (surface) (B, D, and F) staining as described under “Experimental Procedures.” With permeabilization, both CHO/SEP and CHO/SEPΔ cells exhibited strong intracellular staining (A and C). CHO/SEP and CHO/SEPΔcells were not stained without prior permeabilization (B andD). CHO-K1 cells, the parental CHO cell line, exhibited no staining (E and F).
). CHO/SEP and CHO/SEPΔ cells were transiently transfected with a prepro-ET-1 construct, with the amount of mature ET-1 subsequently secreted from these cells into the medium determined by a sandwich-type enzyme immunoassay. As shown in Fig. 7, parental CHO cells transfected with prepro-ET-1 cDNA did not secrete a significant amount of mature ET-1, consistent with the finding that CHO cells do not have detectable ECE activity (
). On the other hand, CHO/ECE-1a cells, which constitutively express bovine ECE-1a, transfected with the prepro-ET-1 construct, produced large amounts of mature ET-1. This is also consistent with our previous finding that the ECE-1 cDNA confers the ability to secrete mature ET-1 to these cells (
). CHO/SEP and CHO/SEPΔ cells transfected with prepro-ET-1 cDNA also produced significant amounts of mature ET-1, indicating that SEP has an ability to cleave big ET-1 to produce mature ET-1. However, much smaller amounts of mature ET-1 were produced by CHO/SEP cells than those by CHO/ECE-1a cells. These observations suggest that SEP may have smaller ECE-like activity than ECE-1. Alternatively, reduced levels of ET-1 may also be due to the possibility that SEP may further degrade the ET-1 after the cleavage at the ECE-1 specific Trp21-Val22 cleavage site in big ET-1.
Figure 7Production of mature ET-1 by CHO cells doubly transfected with prepro-ET-1 and SEP, SEPΔor ECE-1a cDNAs. Doubly transfected cells were cultured for 12 h, and mature ET-1 in the conditioned medium was subsequently determined by enzyme immunoassay as described under “Experimental Procedures.”
Determination of Cleavage Sites of Big ET-1 and ET-1 by Recombinant Soluble SEP
Previously, we have shown that ECE-1 cleaves the Trp21-Val22 bond of big ET-1(1–38) to produce mature ET-1(1–21) and the C-terminal half of big ET-1(22–38) without further cleaving other parts of big ET-1 or ET-1 (
). To examine the cleavage site(s) of big ET-1 and mature ET-1 by recombinant SEP in anin vitro assay, we partially purified the soluble form of SEP from the conditioned medium of CHO/SEP cells. We then incubated relatively large amounts of big ET-1 (10 μm) and ET-1 (1.5 μm) with partially purified SEP for a prolonged period of time (12 h) while directly monitoring the cleavage by reverse-phase HPLC. Product peaks were collected, and the peptides were identified by mass spectrometry. Big ET-1(1–38) was hydrolyzed to a significant degree (42%) by the soluble form of SEP, and HPLC resolved at least four distinct product peaks (Fig.8B). Two peptides products (Fig. 8C, peaks 1 and 4) were coeluted with the standards of big ET-1(22–38) and mature ET-1(1–21), respectively. Analysis using mass spectrometry confirmed that they were big ET-1(22–38) and mature ET-1(1–21), with an m/z value for (M + H)+ of 1810 and 2493, respectively. These findings indicate that the soluble form of SEP can produce ET-1 by cleaving at the specific Trp21-Val22 site of big ET-1. On the other hand, it appears that ET-1 is further digested by the soluble form of SEP. Fig. 8D shows that ET-1 was digested to near completion, with two major product peaks resolved by HPLC under these conditions, indicating that the soluble form of SEP hydrolyzes mature ET-1 at multiple sites. These two major peaks (Fig. 8D, peaks 1 and 2) from mature ET-1 were coeluted with the peaks (Fig. 8C, peaks 2 and 3, respectively) produced from big ET-1 hydolysis. Using molecular masses, one peptide product (Fig. 8, C, peak 3, and D, peak 2) was identified as ET-1(1–16). Another peptide product (Fig. 8, C, peak 2,and D, peak 1) appeared to be the two-chain ET-1(1–16), which is held together by two disulfide bonds between Cys1and Cys15 and between Cys3 and Cys11, presumably produced by cleavage at one site between Ser4 and Glu10 of ET-1(1–16) (
). A parallel preparation of the protein from untransfected CHO cells exhibited no detectable activity for both big ET-1 and ET-1 (data not shown). These activities of the soluble form of SEP were completely inhibited by 100 μm phosphoramidon (Fig. 8, E andF). These results suggest that big ET-1 is initially cleaved at the Trp21-Val22 site by the soluble SEP, resulting in the production of mature ET-1 and that the newly formed ET-1 may be concomitantly degraded by the soluble SEP.
Figure 8SEP hydrolysis of big ET-1 and ET-1.A and B, a 0.1-ml sample of big ET-1 (A) or ET-1 (B) in a reaction buffer (0.1m MES-NaOH, 0.5 m NaCl, pH 7.4) was analyzed by HPLC as described under “Experimental Procedures.” C andD, big ET-1 (C) or ET-1 (D) was incubated with 10 μg of partially purified SEP in a reaction buffer for 12 h at 37 °C. A 0.1-ml aliquot of this reaction was analyzed by HPLC, and the products were identified as described under “Experimental Procedures.” Peaks 1 and 4 inC were identified as big ET-1(22–38) and mature ET-1(1–21), respectively. E and F, HPLC analysis of a reaction treated with phosphoramidon. The reactions were incubated with 100 μm phophoramidon for 15 min prior to the addition of big ET-1 (E) or ET-1 (F). The peaks designated X are unidentified components of the phosphoramidon preparation. Products eluting between 20–80% acetonitrile are shown.
We next assessed the enzymological properties of cloned SEP by using its ET-1 degrading activity. This activity of soluble form of SEP was inhibitedin vitro by 1,10-phenanthroline, the metalloprotease inhibitor phosphoramidon, and the specific NEP inhibitor thiorphan (Table I). Both phosphoramidon and thiorphan inhibited SEP activity in a dose-dependent manner, with apparent IC50 values of about 6 nmand 2 μm, respectively (Fig.9B). The enzyme was partially inhibited by the ECE specific inhibitor FR901533 and was not inhibited by the angiotensin-converting enzyme inhibitor captopril (
). A pH profiling study revealed a neutral optimal pH at 7.4, with a relatively sharp pH dependence (Fig. 9C). These observations indicate that soluble form of SEP represents a novel metalloprotease with a neutral pH optimum and that it has an inhibitor sensitivity profile similar to that of NEP.
Table IProtease inhibitor profile of SEP partially purified from the conditioned medium of CHO/SEP cells
Figure 9In vitro characterization of partially purified SEP .A, functional expression of the SEP cDNA in transfected CHO cells. The partially purified soluble form of SEP was prepared from the supernatant of CHO/SEP cells and untransfected CHO cells, and was assayed for in vitro SEP activity in the absence and presence of 100 μmphosphoramidon. B, concentration-dependent inhibition of the soluble form of SEP by phosphoramidon, thiorphan, and FR901533. C, pH profile of the soluble form of SEP activity partially purified from the supernatant of CHO/SEP cells.
). Substrate hydrolysis was monitored by HPLC, and the results are summarized in TableII. We found that angiotensin I, ANP, bradykinin, and substance P were all digested to completion or near completion by soluble SEP, whereas a parallel preparation of the protein from untransfected CHO cells exhibited no detectable activity. These observations suggest that SEP possesses a broad substrate specificity that is similar to that of NEP.
Table IISEP hydrolysis of biologically active peptides
We have described the cloning and characterization of SEP, a novel soluble secreted metalloprotease that can hydrolyze a variety of vasoactive peptides. Our initial goal was to isolate another endothelin-converting enzyme, and we have demonstrated that SEP can cleave big ET-1 to produce mature ET-1 in vitro as well as in transfected cells. However, we feel that SEP does not qualify as a physiological ECE because it appears to hydrolyze ET-1 more efficiently than it produces ET-1 from big ET-1. Instead, many factors point to SEP sharing higher structural and functional similarities with NEP than with ECEs or other members of this metalloprotease family. First, the sequence identity of SEP with respect to NEP is higher than those of the other members. In particular, SEP and NEP are 42.7% identical to each other in their N-terminal portions (amino acids 1–510 of SEP), whereas they only slightly resemble ECE-1, ECE-2, and XCE in this region. Second, the two arginine residues known to constitute the substrate binding sites in NEP (Arg102 and Arg747 in human NEP) are conserved in SEP (Arg121 and Arg764 in mouse SEP) (
). In contrast, only one of the two arginine residues is conserved in ECE-1 (Arg129 in human ECE-1b) and ECE-2 (Arg162 in bovine ECE-2), although this has been shown to play an insignificant role in the substrate binding of ECE-1 (
). Fourth, both SEP and NEP rapidly degrade big ET-1 and ET-1 at multiple internal cleavage sites, whereas ECE-1 cleaves specifically at the Trp21-Val22 bond of big ET-1 without cleaving other parts of big ET-1 or ET-1 (
). Fifth, the activity of SEP is efficiently inhibited by the specific NEP inhibitor thiorphan but is not completely inhibited by the specific ECE inhibitor FR901533 (
). Finally, both SEP and NEP cleave many small peptides in a highly promiscuous manner. These findings suggest that SEP is not a physiologically relevant endothelin-converting enzyme and that SEP and NEP may constitute a subfamily within this group of metalloproteases.
Although SEP shares several important features with other known members of this metalloprotease family, it still exhibits striking differences from them. Transfection of an expression construct of SEP results in the release of a functional soluble form of the enzyme into the culture medium. This suggests that the soluble form of SEP may act as a circulating endopeptidase in vivo. These observations are in sharp contrast to the fact that both NEP and ECE act as membrane-bound enzymes, and neither releases a soluble form (
). In fact, we are unaware of any member of this metalloprotease family that releases a functional soluble form of the enzyme. The endogenous proteolytic release of an integral membrane-bound ectoenzyme is well documented for angiotensin-converting enzyme (ACE), a member of another metalloprotease family, which plays a critical role in the maintenance of blood pressure in mammals (
). Although ACE exists primarily as a membrane-bound enzyme, a soluble form is present under normal conditions in many body fluids, including blood plasma (
). In mammals, ACE exists as two distinct isoenzymes, namely somatic and testicular ACE. They are derived from a single gene by alternative splicing. Transfection of the full-length cDNA of either the somatic or testicular isoenzyme results not only in the expression of the membrane-bound form of ACE on the cell surface but also in a secreted form as a result of proteolysis by some enzyme(s) (
). However, SEP exhibits the following differences from ACE: (i) only one spliceoform of SEP expresses a soluble form of the enzyme, whereas both spliceoforms of ACE produce soluble forms of the enzyme; and (ii) membrane-bound ACE is expressed on the cell surface as an ectoenzyme, whereas membrane-bound SEP appears to be expressed in the early secretory pathway, including endoplasmic reticulum and a portion of the Golgi apparatus (see below). Thus, although other metalloproteases are known to produce both forms, SEP is clearly a novel molecule. Whether the soluble form of SEP has in vivo activity remains to be tested.
The SEP and SEPΔ polypeptides predicted from the cDNA are identical except for 23 amino acids that are unique to SEP. Therefore, the structural determinants that cause the release of the soluble form of SEP must be embedded within these 23 amino acid residues. Interestingly, a pair of dibasic residues, Lys62-Arg63, are contained in the C-terminal end of this SEP sequence and are the most common processing sites of prohormones and neuroendocrine precursors found in mammals (
). Although we have not determined the exact cleavage site of SEP, it is tempting to speculate that the biologically active soluble form of SEP is proteolytically released from the membrane-bound SEP by a dibasic processing endoprotease(s). Alternatively, SEP may be cleaved at other sites by the membrane protein secretase in a fashion similar to that of some membrane proteins, such as the β-amyloid precursor protein, the tumor necrosis factor-α ligand/receptor, or ACE (
). A detailed mutagenesis study, as well as an N-terminal amino acid sequence of the purified soluble form of SEP, is required to further elucidate this process.
Our in vitro enzymological analyses suggest that both SEP and NEP may share similar enzymological properties. Both enzymes efficiently hydrolyze circulating vasoactive peptides, including ET-1, angiotensin I, ANP, bradykinin, and substance P, resulting in the degradation of these peptides with a similar overall inhibitor profile. However, from the results presented in this paper, we cannot establish whether SEP can degrade these physiological substrates in vivo. Obviously, further studies are required to demonstrate the physiological relevancy of SEP in the degradation of these vasoactive peptides. To further characterize the physiological function of SEP, it is worth noting that SEP mRNA was found to be highly expressed in mouse testis, which is known to express a variety of metalloproteases including NEP, ECE-1, and ACE (
). Although the physiological roles of NEP and ECE-1 in the testis have not yet been fully elucidated, the high expression of NEP and ECE-1, as well as their putative peptide substrates, suggest that they may have functional roles in the reproductive system (
). Likewise, SEP may have some functional roles in the reproductive system by processing some known or, as yet unknown, substrates. In situ hybridization with an SEP probe and immunohistochemical analysis using our SEP antibody in testis should facilitate investigations of the physiological function of this putative enzyme.
We have not yet formally tested whether the intracellular membrane-bound form of SEP, which is sensitive to Endo H, is functionally active. To characterize the enzymological properties of partially glycosylated SEP, we need a purified preparation of membrane-bound SEP. However, a double transfection assay using prepro-ET-1 and SEPΔ cDNAs resulted in the secretion of significant amounts of mature ET-1. Moreover, ECE-1 has been shown to be active after the removal of N-glycosylation, suggesting that N-glycosylation is not essential for ECE-1 enzyme activity (
). These observations suggest that the membrane-bound SEP may function as an intracellular enzyme.
The steady-state subcellular localization of membrane-bound SEP in CHO cells appears to be different from those of the other metalloprotease family members. Our immunocytochemistry results of the CHO transfectants, as well as our findings of SEP sensitivity to Endo H, suggest that a majority of membrane-bound SEP is located in the early secretory pathway. In contrast, NEP is expressed on the cell surface and functions as an ectoenzyme (
), whereas ECE-2 is thought to be located intracellularly in an acidic compartment, based on the findings that ECE-2 is active at an acidic pH and is virtually inactive at a neutral pH (
). The subcellular localization of ECE-1 has been controversial, as it has been reported to be present in intracellular compartments as well as on the cell surface (
). Previously, we reported that ECE-1 spliceoforms, which differ only in their N-terminal cytoplasmic tail, exhibit distinct subcellular localizations and that the cytoplasmic tail of one spliceoform contains novel signals that mediate constitutive lysosomal targeting (
). Likewise, SEP may contain novel signals that mediate its unique intracellular targeting. An extensive site-directed mutagenesis study should allow us to further elucidate the molecular mechanisms of SEP targeting.
Further studies, including the generation and analysis of mice containing a targeted disruption of the SEP gene, are required to determine the physiological substrates of SEP, and hence the physiological functions of this enzyme. Considering the medical potential of NEP and ECE inhibitors, detailed enzymological characterization and structural studies of SEP will be invaluable in the rational design of selective and mixed inhibitors of NEP and ECEs.
Acknowledgments
We thank Dr. David E. Clouthier for critical reading of the manuscript.
51580-0&id=gr2.jpg){kind=link}
51580-0&id=gr3.jpg){kind=link}
51580-0&id=gr1.jpg){kind=link}
51580-0&id=gr4.jpg){kind=link}
51580-0&id=gr5.jpg){kind=link}
51580-0&id=gr6.jpg){kind=link}
51580-0&id=gr7.jpg){kind=link}
51580-0&id=gr8.jpg){kind=link}
51580-0&id=gr9.jpg){kind=link}