Functional Analysis of the Transmembrane Domain and Activation Cleavage of Human Corin

Corin is a cardiac transmembrane serine protease. In cell-based studies, corin converted pro-atrial natriuretic peptide (pro-ANP) to mature ANP, suggesting that corin is potentially the pro-ANP convertase. In this study, we evaluated the importance of the transmembrane domain and activation cleavage in human corin. We showed that a soluble corin that consists of only the extracellular domain was capable of processing recombinant human pro-ANP in cell-based assays. In contrast, a mutation at the conserved activation cleavage site, R801A, abolished the function of corin, demonstrating that the activation cleavage is essential for corin activity. These results allowed us to design, express, and purify a mutant soluble corin, EKsolCorin, that contains an enterokinase recognition sequence at the activation cleavage site. Purified EKsolCorin was activated by enterokinase in a dose-dependent manner. Activated EK-solCorin had hydrolytic activity toward peptide substrates with a preference for Arg and Lys residues in the P-1 position. This activity of EKsolCorin was inhibited by trypsin-like serine protease inhibitors but not inhibitors of chymotrypsin-like, cysteine-, or metallo-proteases. In pro-ANP processing assays, purified active EKsolCorin converted recombinant human pro-ANP to biologically active ANP in a highly sequence-specific manner. The pro-ANP processing activity of EKsolCorin was not inhibited by human plasma. Together, our data indicate that the transmembrane domain is not necessary for the biological activity of corin but may be a mechanism to localize corin at specific sites, whereas the proteolytic cleavage at the activation site is an essential step in controlling the activity of corin.

Corin mRNA and protein are abundantly expressed in the heart (1,2), suggesting that corin might have a role in the cardiovascular system. In cell-based experiments, we showed that recombinant human corin mediated the conversion of proatrial natriuretic peptide (pro-ANP) and pro-brain natriuretic peptide to mature ANP and brain natriuretic peptide (21), both of which are cardiac hormones important in maintaining normal blood pressure and electrolyte homeostasis (22)(23)(24). Corin, however, does not convert pro-C-type natriuretic peptide to mature C-type natriuretic peptide (25), the third member of the natriuretic peptide family (26), which may play a role in angiogenesis and arterial restenosis. The results from these experiments suggest that corin is the pro-ANP/pro-brain natriuretic peptide convertase in the heart. This hypothesis is further supported by additional experiments in which overexpression of an active site mutant corin or transfection of small interfering RNA duplexes directed against the corin gene completely blocked pro-ANP processing in cultured cardiomyocytes (27). To date, however, the reported studies of corin were performed in cell-based experiments, which do not allow elimination of the possibility that other proteins or enzymes might contribute to the observed pro-ANP processing activities. It is important, therefore, to demonstrate directly that corin itself possesses the pro-ANP processing activity by using purified active corin.
To test the hypothesis that purified active corin is able to process pro-ANP, we first examined the importance of the transmembrane domain and activation cleavage in corin for its biological activity. Our results showed that the transmembrane domain is not required for corin to process pro-ANP, but proteolytic cleavage of corin at its conserved activation site is essential. Based on these results, we designed, expressed, and purified a soluble form of human corin and studied its biochemical properties. Our results showed that activated soluble human corin hydrolyzed synthetic peptic substrates and activated human pro-ANP in a highly sequence-specific manner.

EXPERIMENTAL PROCEDURES
Materials-Cell culture medium, G418, anti-V5 antibody, transfection reagent LipofectAMINE 2000, and expression vector pSecTag/FRT/ V5-His-TOPO were purchased from Invitrogen. Fetal bovine serum was from SeraCare Life Sciences, Inc. (Oceanside, CA). Human embryonic kidney (HEK) 293 cells were obtained from the American Type Culture Collection and maintained at the Core Facility at Berlex Biosciences.
* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  Expression Vectors-Expression plasmids encoding human wild-type corin (pcDNACorin), active site mutant corin S985A (pcDNACorinS985A), human wild-type pro-ANP (pcDNAproANP), and mutant pro-ANPs R98A (pcDNAproANPR98A), R101A (pcDNAproANPR101A), and R102A (pcDNAproANPR102A) were described previously (21,27). A plasmid vector encoding corin activation cleavage site mutant R801A (pcDNACorinR801A) was constructed by a PCR-based mutagenesis method (QuikChange sitedirected mutagenesis kit; Stratagene, La Jolla, CA) using pcDNACorin as a template and oligonucleotide primer 5Ј-CCG AAT GAA CAA AGC AAT CCT TGG AGG TCG-3Ј. To construct a plasmid expressing a soluble corin, a cDNA fragment containing nucleotides 463-3219 of human corin cDNA (1) was amplified by PCR and inserted into the expression vector pSEC to yield the plasmid pSECsolCorin. This plasmid encodes a protein, WTsolCorin, consisting of an Ig signal peptide at the N terminus followed by a 919amino acid sequence from the extracellular region of corin (residues 124-1042) and a viral V5 tag at the C terminus (see Fig. 1). To construct a plasmid encoding a soluble corin that can be activated by EK, PCR-based site-directed mutagenesis was performed using plasmid pSECsolCorin as a template. Nucleotides 2482-2496 (CGA ATG AAC AAA AGG) of human corin cDNA were replaced by nucleotides GAC GAT GAC GAT AAG. The resulting plasmid, pSECEKsolCorin, encodes a soluble corin, EKsolCorin, with the amino acid sequence DDDDK replacing the original sequence RMNKR at the conserved activation cleavage site (see Fig. 1). All plasmid constructs were verified by restriction enzyme digestion and direct DNA sequencing.
Cell Culture, Transfection, and Western Analysis of Pro-ANP Processing-HEK 293 cells were cultured in ␣-minimum essential medium supplemented with 10% fetal bovine serum. Transient transfection was performed using LipofectAMINE 2000 according to the manufacturer's instructions. The conditioned medium was collected 12-16 h after transfection and subjected to centrifugation at 15,000 rpm to remove cell debris. The cells were lysed in a buffer containing 100 mM Tris-HCl, pH 7.5, and 1% Triton X-100. To analyze pro-ANP processing, the conditioned medium containing recombinant wild-type or mutant pro-ANPs was incubated with purified soluble corin (1.8 g/ml) at 37°C for 4 h. Recombinant human pro-ANP and its derivatives in the conditioned medium were immunoprecipitated by an anti-V5 antibody. The protein samples were separated by SDS-PAGE and analyzed by Western blotting using a horseradish peroxidase-conjugated anti-V5 antibody. On Western blots under reducing conditions, recombinant human corin zymogen, the corin protease domain, pro-ANP, and ANP migrate as bands of ϳ150, ϳ35, ϳ24, and ϳ7 kDa, respectively.
Expression and Purification of EKsolCorin-HEK 293 cells were co-transfected with the expression vector pSECEKsolCorin and a plasmid expressing the neomycin resistance gene using LipofectAMINE 2000. Stable clones were selected in ␣-minimum essential medium containing 10% fetal bovine serum and 500 g/ml G418 and screened by Western blotting using an anti-V5 antibody for corin protein expression. Positive clones were adapted for growth in serum-free OPTI-MEM I medium containing 500 g/ml G418. The conditioned medium was collected, passed through a 0.2-m filter, and dialyzed against Buffer A (50 mM Tris-HCl, pH 7.5, 300 mM NaCl) using a Dialyze Direct L Module (Qiagen). The medium was loaded onto a 23-ml nickel-nitrilotriacetic acid Superflow column (Qiagen) that was subsequently washed with Buffer A containing 10 mM imidazole and eluted with a 10 -250 mM imidazole linear gradient in Buffer A. The fractions containing soluble corin were identified by Western blotting and combined. The pooled fractions were then diluted 1:3 in Buffer B (20 mM Tris-HCl, pH 8.0) and loaded onto a 5-ml Hi Trap Q Sepharose column (Amersham Biosciences). The column was equilibrated with Buffer B containing 100 mM NaCl, washed with Buffer B containing 200 mM NaCl, and eluted with a 200 -750 mM NaCl linear gradient in Buffer B. The fractions containing the soluble corin, EKsolCorin, were identified by Western blotting and pooled. The purified protein was further characterized by Coomassie Blue staining, analytical size exclusion chromatography, and N-terminal protein sequencing.
Activation of EKsolCorin by Recombinant EK-To activate the recombinant soluble corin, EKsolCorin, 2.5 g of purified EKsolCorin protein was incubated with increasing concentrations of recombinant EK (1-10 units/ml) in 100 l of Activation buffer (100 mM Tris-HCl, pH 7.5, 10 mM CaCl 2 ) at 25°C for 2 h. The samples (2 l) were taken and analyzed by SDS-PAGE under reducing and nonreducing conditions followed by Western blotting using an anti-V5 antibody. For activation of large batches, 1 mg of purified EKsolCorin in 40 ml of activation buffer was incubated with 300 units of recombinant EK at 25°C for 3 h. To remove the recombinant EK, 11 ml of EKapture beads were added to the solution and incubated at room temperature for 15 min. EKapture beads were removed by centrifugation (1,000 rpm, 10 min), and the supernatant was collected and stored at Ϫ20°C until further use. EK was removed from the EKsolCorin preparation to below the limit of detection when analyzed by both SDS-PAGE followed by silver staining and by HPLC-based analytical size exclusion chromatography. As another control, an assay buffer without corin protein underwent the same EK activation and removal procedures.
Enzyme Kinetics-Kinetic constants were determined using a panel of selected synthetic chromogenic substrates. For each assay, which was carried out in 96-well plates, 50 l of substrates (final concentrations ranging from 0.2 to 2 mM in 100 mM Tris-HCl, pH 7.5, 10 mM CaCl 2 ) were mixed with 50 l of activated EKsolCorin (final concentration of 58 nM). The plates were incubated at 37°C and read at 405-nm wavelength over 15 min at 20-s intervals in a Spectra MAX 250 plate reader (Molecular Devices Corp., Sunnyvale, CA). In these experiments, controls included purified EKsolCorin that was not activated by EK and an assay buffer that underwent the same EK treatment and removal procedures. Readings from the controls, which were minimal, were subtracted as the background. In addition, plasmin (for S-2222, S-2251, S-2302, S-2366, S-2403, and S-2444), kallikrein (for S-2266 and S-2288), thrombin (for S-2238), trypsin (for S-2765), and elastase (for S-2484) were used as positive controls under the conditions recommended by the manufacturer. The K m and V max values were determined by Lineweaver-Burk double-reciprocal plot. Each enzymatic assay was carried out in triplicate and repeated at least twice.
Effects of Protease Inhibitors-Effects of protease inhibitors on EK-solCorin were tested in an assay using the chromogenic substrate S-2403. In each experiment, 45 l of activated EKsolCorin (final concentration of 104 nM) was mixed with 5 l of an inhibitor (final concentrations ranging from 0.1 M to 20 mM) and incubated at 37°C for 30 min. To measure the remaining hydrolytic activity of EKsolCorin, 50 l of S-2403 (final concentration of 500 M) was added to the mixture, and the absorbance was measured at 405 nm after 2 h. Each experiment was performed in triplicate and repeated at least twice.
cGMP Assay-To examine the biological activity of corin-processed recombinant ANP, a cGMP assay was performed using an enzyme immunoassay kit (Biotrak; Amersham Biosciences), as described previously (27). In these experiments, synthetic human ANP (Peninsula Laboratories Inc., San Carlos, CA) was used as a standard. Each experimental condition was assayed in triplicate.

Processing of Pro-ANP by Wild Type and a Soluble Corin,
WTsolCorin-To examine the importance of the transmembrane domain of corin for its pro-ANP processing activity, we constructed a plasmid that encodes a soluble form of human corin containing all of the extracellular domains ( Fig. 1). Recombinant wild-type corin, active site mutant corin S985A, and the soluble corin, WTsolCorin, were expressed transiently in HEK 293 cells. Wild-type corin and active site mutant corin S985A were detected by Western analysis in the cell lysate but not in the conditioned medium ( Fig. 2A), consistent with corin being a transmembrane protein. In contrast, WTsolCorin was detected in both the cell lysate and conditioned medium ( Fig.  2A), confirming that the soluble corin was secreted from the cells. We determined the activity of these recombinant corins in pro-ANP processing in co-transfection experiments using a plasmid expressing human pro-ANP together with plasmids expressing either wild-type corin, mutant corin S985A, or WT-solCorin. Pro-ANP and its derivatives in the conditioned medium were analyzed by Western blotting. As shown in Fig. 2B, pro-ANP, but not ANP, was detected in the cell lysate. In the conditioned medium, conversion of pro-ANP to ANP was ob-served when cells were transfected with the pro-ANP expressing plasmid together with plasmids expressing wild-type corin or WTsolCorin. As controls, the cells were co-transfected with the pro-ANP expressing construct and a plasmid expressing either active site mutant corin S985A or a control vector. Without the presence of a plasmid expressing an active corin, no pro-ANP processing was detected, showing that the cells do not contain any detectable endogenous pro-ANP processing activity. The results indicate that the transmembrane domain of corin is not necessary for the pro-ANP processing activity in this cell-based assay.

Processing of Pro-ANP by Wild-type Corin and Activation
Cleavage Site Mutant Corin R801A-The human corin protein contains a conserved activation cleavage sequence Arg-Ile-Leu-Gly-Gly at residues 801-805 (1). To examine the functional importance of the activation cleavage of human corin, we constructed a plasmid expressing a mutant that would have impaired activation (activation cleavage site mutant corin R801A; Fig. 1). Co-transfection experiments were performed using a plasmid expressing pro-ANP together with plasmids expressing either wild-type corin, active site mutant corin S985A, or activation cleavage site mutant corin R801A. Recombinant corin proteins were present in the cell lysate but not in the conditioned medium (Fig. 3A). Recombinant pro-ANP was detected in the cell lysate after transfection with the pro-ANP expressing plasmid (Fig. 3B, left panel). In the conditioned medium, processing of pro-ANP to ANP was observed when cells were co-transfected with a plasmid encoding wild-type corin but not those encoding mutant corins S985A and R801A or a control vector (Fig. 3B, right panel). The results demonstrate that proteolytic cleavage at Arg 801 is required for the pro-ANP processing activity of corin. The fact that wild-type corin was capable of processing pro-ANP in cell-based transfection experiments indicates that some corin molecules must be activated. In the Western analysis (Figs. 2 and 3), however, we were unable to detect the cleaved protease fragment from wildtype corin or mutant corin S985A, which is expected to migrate as a band at ϳ35 kDa under reducing conditions. This would suggest that the number of activated corin molecules is a low fraction of the overall amount of corin.
Effects of Thrombin, Factor Xa, and Kallikrein on Corin Activation-To examine whether plasma-derived serine proteases could activate corin, we examined the effects of thrombin, blood clotting factor Xa, and kallikrein on corin activation. Recombinant human wild-type corin was stably expressed in HEK 293 cells. Purified human plasma thrombin, factor Xa, or kallikrein was added to the cell culture and incubated at 37°C for 1 h. The cell lysates were prepared and analyzed by Western blotting under reducing conditions. The results showed that recombinant human corin was not cleaved in cells treated with either thrombin, factor Xa, or kallikrein (data not shown). We also added thrombin, factor Xa, or kallikrein directly to 293 cell lysates containing recombinant corin and analyzed the The conserved activation cleavage site is indicated by an arrow. The disulfide bond (S -S) that connects two polypeptide chains after the activation cleavage is also shown. WT corin, wild-type corin; corin S985A, a mutant corin in which the active-site Ser is replaced by Ala; corin R801A, a mutant corin in which the cleavage site Arg is replaced by Ala; WTsol-Corin, a soluble corin that consists of a signal peptide sequence derived from human Ig chain followed by the extracellular domains of corin; EKsolCorin, a soluble corin that contains an EK recognition sequence (DDDDK) at the activation cleavage site. corin protein by SDS-PAGE and Western blotting. Again, no activation cleavage of corin was detected in these experiments (data not shown).
Expression and Purification of Soluble EK-activable Corin, EKsolCorin-To produce an active soluble corin for further biochemical studies, we designed a mutant corin, EKsolCorin, in which an EK recognition sequence (DDDDK) (7) was used to replace the activation cleavage sequence of human corin (RMNKR) (1). Stable cell lines expressing EKsolCorin were established. The conditioned medium from these cells was collected, and EKsolCorin was purified by nickel affinity and ion exchange chromatography. SDS-PAGE followed by Coomassie Blue staining and Western blotting using an anti-V5 antibody showed that EKsolCorin migrated as a single band at ϳ150 kDa under reducing conditions and at ϳ145 kDa under nonreducing conditions (Fig. 4). The results were consistent with the calculated mass of ϳ108 kDa for EKsolCorin, which also contained 19 potential N-linked glycosylation sites (1). HPLCbased gel filtration chromatography showed that the purified EKsolCorin had a purity of Ͼ98% (data not shown). The Nterminal sequence of EKsolCorin was confirmed by protein sequencing. The purified protein was quantified by UV spectrometry at 280 nm using an extinction coefficient (1 mg/ml) of 1.45 calculated from the protein sequence.
Activation of EKsolCorin by EK-Purified EKsolCorin protein was activated with increasing concentrations of recombinant EK at 25°C for 2 h. The protein samples were analyzed by SDS-PAGE under reducing and nonreducing conditions followed by Western analysis using an anti-V5 antibody. As shown in Fig. 5, EKsolCorin was activated by recombinant EK in a dose-dependent manner. Under nonreducing conditions, EKsolCorin appeared as a single band of ϳ145 kDa. Under reducing conditions, activated EKsolCorin migrated as two fragments: an N-terminal propeptide (ϳ115 kDa) and a Cterminal protease domain (ϳ35 kDa) (Fig. 1). Because the V5 tag is located at the C terminus, the anti-V5 antibody detected only the C-terminal protease domain once EKsolCorin was activated (Fig. 5).
Enzymatic Properties of Soluble Corin-Substrate specificity and kinetic constants of EKsolCorin for a panel of selected peptide substrates were determined using purified and activated EKsolCorin. As controls, EKsolCorin without EK activation and a buffer that was treated with EK were included. Initial experiments showed that activated EKsolCorin, but not the zymogen form of EKsolCorin or the buffer control, hydro- ( Table I). The results showed that EKsolCorin cleaved peptide substrates with either Arg or Lys at the P-1 position. For example, the K m values were 1.28 Ϯ 0.46 and 3.52 Ϯ 1.07 mM for S-2403 and S-2366, respectively. Pro, Phe, and Gly residues appeared to be preferred at the P-2 position, and a pyro-Glu residue, an analog of small neutral amino acids, seemed to be preferred at the P-3 position. The overall results are consistent with the corin cleavage sequence (Thr-Ala-Pro-Arg2Ser) in human pro-ANP (28).
Effects of Protease Inhibitors-To examine the effects of protease inhibitors on corin, the hydrolysis of the chromogenic substrate S-2403 by activated EKsolCorin was monitored in the presence of various protease inhibitors. As shown in Table  II, the activity of EKsolCorin was inhibited dose-dependently by nonspecific trypsin-like serine protease inhibitors including benzamidine, phenylmethylsulfonyl fluoride, antipain, leupeptin, aprotinin, tosyl-Lys-chloromethylketone, and soybean trypsin inhibitor. In contrast, the activity of EKsolCorin was not inhibited by inhibitors of chymotrypsin-like serine proteases such as chymostatin (100 M) and Tosyl-Phe-chloromethylketone (100 M) or metallo-and cysteine-protease inhibitors such as phosphoramidon (1 mM), EDTA (20 mM), pepstatin (100 M), and bestatin (100 M). These data are consistent with corin being a trypsin-like serine protease based on its protein sequence.
Processing of Pro-ANP by EKsolCorin-To demonstrate the pro-ANP processing activity of purified and activated EKsol-Corin, recombinant human wild-type pro-ANP and mutant pro-ANP R98A, R101A, and R102A were expressed in HEK 293 cells. Conditioned media were collected and incubated with purified EKsolCorin. Processing of pro-ANP was analyzed by Western blotting. As shown in Fig. 6, activated EKsolCorin converted wild-type pro-ANP and mutant pro-ANPs R101A and R102A, but not mutant pro-ANP R98A, to mature peptides. An additional weak band of ϳ20 kDa was also observed in these experiments, which represented a degradation fragment derived from EKsolCorin (data not shown). This fragment is biologically inactive as measured by cell-based cGMP assay (see below). As controls, recombinant EK or the zymogen form of EKsolCorin did not cleave recombinant pro-ANPs. The results are consistent with our previous finding that corin cleaves human pro-ANP specifically at residue Arg 98 but not at the adjacent arginine residues 101 or 102. The results also show that the introduction of an EK recognition sequence did not alter the sequence specificity of corin for its physiological substrate.
The Activity of EKsolCorin-processed Recombinant ANP-The biological function of ANP is mediated through its receptor that has guanylyl cyclase activity. Binding of ANP to its receptor stimulates the guanylyl cyclase activity, leading to production of intracellular cGMP. To determine whether EKsolCorin-processed recombinant ANP is biologically active, a BHK cell-based cGMP assay was performed. As shown in Fig. 7, little cGMP-stimulating activity was detected in the conditioned medium from 293 cells containing either pro-ANP or activated EKsolCorin. The cGMPstimulating activity was significantly increased when activated EKsolCorin was added to the conditioned medium containing wildtype pro-ANP or mutant pro-ANPs R101A and R102A. In contrast, there was no significant increase in the cGMP-stimulating activity when activated EKsolCorin was added to the conditioned medium containing mutant pro-ANP R98A. These data are consistent with the results showing that mutation at Arg 98 in pro-ANP prevented the conversion of pro-ANP to mature ANP (Fig. 6) and demonstrate that EKsolCorin-processed recombinant ANP is biologically active.
Effect of Human Plasma on the Pro-ANP Processing Activity of EKsolCorin-To examine whether there are potential corin inhibitors present in human plasma, we tested the effects of human plasma on EKsolCorin-mediated processing of pro-ANP. The conditioned medium containing recombinant human pro-ANP was incubated with activated EKsolCorin in the presence of increasing concentrations of pooled human plasma. Processing of pro-ANP was analyzed by Western blotting. As shown in Fig. 8, no endogenous pro-ANP convertase activity was detected in the pooled human plasma. Activated EKsol-Corin converted pro-ANP to ANP. When up to 75% human plasma was included in the reactions, we observed a reduction in corin-derived bands, suggesting that human plasma may interfere with the Western blots. Importantly, however, there was no inhibition of the pro-ANP convertase activity, indicating that EKsolCorin remained active in the presence of plasma. Furthermore, no inhibition of corin activity was observed when the experiment was performed in the presence of 5 or 50

DISCUSSION
In this study, we examined the importance of the transmembrane domain and the activation cleavage of human corin for its activity in the processing of pro-ANP. We showed that a soluble corin that consists of only the extracellular domain was capable of processing pro-ANP in cell-based assays, indicating that the transmembrane domain is not necessary for corin activity. In contrast, a mutation at the conserved activation cleavage site, R801A, abolished the function of corin, demonstrating that the one-chain form of corin is not active but needs to be cleaved into the two-chain form for the protease domain to be active. These results led us to design, express, and purify a soluble corin, EKsolCorin, that contains an EK recognition sequence at the conserved activation site. We showed that purified EKsolCorin was activated by EK and that activated EKsolCorin was capable of hydrolyzing peptide substrates with a preference for Arg and Lys residues at the P-1 position. We further showed that activated EKsolCorin converted pro-ANP to biologically active ANP in a highly sequence-specific manner.
As a member of the type II transmembrane serine protease family, corin contains an integral transmembrane domain near its N terminus. The importance of this transmembrane domain in corin had not previously been examined. For many soluble serine proteases such as blood clotting enzymes, binding to the cell surface through either phospholipids or integral membrane co-factors greatly enhances the rate of their catalytic reactions (29 -31). For corin, however, the transmembrane domain appears to be dispensable. In transfection experiments, both membrane-bound wild-type corin and the soluble corin, WTsol-Corin, had similar activities in processing human pro-ANP (Fig. 2). The results are consistent with reports of other type II transmembrane serine proteases including EK (32), matriptases (11,33,34), hepsin (35,36), spinesin (16), human airway trpsin-like protease (17), and polyserase-I (20), showing that soluble enzymes lacking the transmembrane domain can be catalytically active. It appears, therefore, that the main func- Processing of human wild-type and mutant pro-ANPs R98A, R101A, and R102A by activated EKsolCorin was analyzed by SDS-PAGE under reducing conditions followed by Western analysis using an anti-V5 antibody (bottom). As controls, recombinant EK and the zymogen form of EKsolCorin were included in the experiments.
FIG. 7. Stimulation of intracellular cGMP production. The conditioned media containing wild-type pro-ANP (WT) and mutant pro-ANPs R98A (R98A), R101A (R101A), or R102A (R102A) were incubated with activated EKsolCorin (EKsolcorin) at 37°C for 2 h, then added to BHK cells cultured in 96-well plates, and incubated at 37°C for 10 min. As controls, the cells were also treated with the conditioned medium containing only wild-type pro-ANP or activated EKsolCorin. The cells were lysed by the addition of a buffer containing 2% dodecyl trimethylammonium and 50 mM sodium acetate, pH 5.8. The intracellular concentrations of cGMP in BHK cells was measured with the Biotrak enzyme immunoassay kit as described under "Experimental Procedures." The intracellular concentrations of cGMP in untreated BHK cells, which were minimal, were subtracted as the background. Each experimental condition was assayed in triplicate.
FIG. 8. Effect of human plasma on EKsolCorin-mediated pro-ANP processing. Purified EKsolCorin was activated by recombinant EK and added to reactions containing recombinant human pro-ANP in the presence of increasing concentrations of pooled human plasma. As controls, the reactions containing EK or the zymogen form of EKsol-Corin were also included. Processing of pro-ANP was analyzed by SDS-PAGE followed by Western blotting using an anti-V5 antibody. tion of the transmembrane domain in the proteases of this family is either to retain the activity of these enzymes in specific tissues or to localize the activity to specific subcellular sites.
Most trypsin-like proteases are synthesized as a one-chain zymogen. Proteolytic cleavage at a conserved activation site converts the zymogen to an active enzyme (37)(38)(39). Usually, the activation cleavage occurs at a canonical sequence of Arg2Ile-Val-Gly-Gly. In human corin, the predicted cleavage is located at Arg 801 -Ile 802 (1). In our previous cell-based experiments, human wild-type corin was shown to be capable of converting pro-ANP to ANP (21,27). By Western analysis, however, we did not detect the activated form of corin, which should migrate as a two-chain molecule under reducing conditions. This implies that either the single-chain form of corin has sufficient catalytic activity or only a very small fraction of corin was activated in the transfected cells but was undetectable by Western blotting. To test the importance of the activation cleavage in corin, we made a mutant corin R801A, which is expected to abolish the cleavage at Arg 801 -Ile 802 . We found that mutant corin R801A was inactive in processing pro-ANP in cell-based experiments (Fig. 3), indicating that the cleavage at Arg 801 is critical for the activity of corin and that the one-chain zymogen form of corin does not possess detectable activity by the techniques we used in this study. In this regard, corin appears to be different from tissue-type plasminogen activator, which has significant catalytic activity in its single-chain form (40,41). Further studies, however, are required to determine whether single-chain corin may gain the catalytic activity in vivo in the presence of potential co-factors or interacting proteins on the cell surface.
We attempted to activate corin using some soluble plasma proteases but were unsuccessful. At this time, the physiological activator of corin has not yet been identified, making the production of catalytically active corin a challenge. To circumvent this problem, we made a soluble form of corin, EKsolCorin, in which the activation cleavage site was replaced by an EK recognition sequence. We showed that purified EKsolCorin itself was inactive but readily activated by recombinant EK in a dose-dependent manner. In chromogenic substrate assays, activated EKsolCorin exhibited hydrolytic activities with a preference for Arg/Lys residues at the P-1 position, Pro/Phe/Gly at the P-2 position, and pyro-Glu, which is an analog of small neutral amino acids, at the P-3 position. This substrate profile is consistent with the cleavage sequence of Ala-Pro-Arg2Ser in human pro-ANP (28). The observed k cat /K m values of EKsol-Corin for peptide substrates were similar to those reported for other type II transmembrane serine proteases such as human polyserase-I (20) and mouse matriptase (42) but lower than those for human matriptase (10) and EK (43). Importantly, we showed that purified and activated EKsolCorin cleaves human pro-ANP specifically at Arg 98 but not at the adjacent residues Arg 101 or Arg 102 , converting the prohormone to a biologically active peptide (Figs. 6 and 7). Such stringent sequence specificity was observed with wild-type corin in our previous cellbased studies (21,27). The results from this study demonstrate that introduction of the EK cleavage site did not alter the substrate specificity of the soluble corin and that purified active EKsolCorin was indeed capable of processing pro-ANP.
In chromogenic substrate-based assays, we found that the activity of EKsolCorin was inhibited by nonspecific trypsin-like serine protease inhibitors but not inhibitors of chymotrypsinlike, cysteine-, or metallo-proteases. Although the overall results were consistent with corin being a trypsin-like serine protease, inhibition of EKsolCorin by soybean trypsin inhibitor was surprising. In our previous cell-based experiments, we found that the presence of soybean trypsin inhibitor at up to 600 M in culture medium had little effect on the processing of pro-ANP mediated by either recombinant corin in transfected HEK 293 cells (21) or endogenous corin in cultured cardiomyocytes (27). Such different effects of soybean trypsin inhibitor, which has a molecular mass of ϳ20 kDa, on soluble and membrane forms of proteases have also been reported for other type II transmembrane serine proteases. For example, soybean trypsin inhibitor has been shown to inhibit the activity of soluble hepsin in chromogenic substrate assays but have little effect on that of cell-surface hepsin in factor VII activation assays (44,45). It is possible that the presence of the transmembrane domain hinders the access of large molecule protease inhibitors to the protease active site, making these transmembrane proteases more resistant to protease inhibition. Thus, the transmembrane domain of the type II transmembrane serine proteases may serve as a regulatory mechanism in their interactions with cognate inhibitors. At the present time, it is not known whether physiological corin inhibitors exist. In pro-ANP processing assays, we found that human plasma had little effect on the activity of soluble corin (Fig. 8), indicating the absence of any corin inhibitors in human plasma. Together with the results from mutant corin R801A, it is most likely that corin activity is regulated physiologically by the activation cleavage rather than by inhibition of its activity.