- Email: [email protected]
Soluble Human Endothelin-Converting Enzyme-1: Expression, Purification, and Demonstration of Pronounced pH Sensitivity
ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS
Vol. 359, No. 2, November 15, pp. 258 –268, 1998 Article No. BB980913
Soluble Human Endothelin-Converting Enzyme-1: Expression, Purification, and Demonstration of Pronounced pH Sensitivity Kyunghye Ahn,1 Sarah B. Herman, and Douglass C. Fahnoe Department of Biochemistry, Parke–Davis Pharmaceutical Research, Division of Warner–Lambert Company, Ann Arbor, Michigan 48105
Received June 10, 1998, and in revised form August 24, 1998
Endothelin-converting enzyme-1 (ECE-1) is a type II integral membrane protein that belongs to a family of metalloproteases which includes ECE-2, neprilysin (neutral endopeptidase 24.11, EC 3.4.24.11), and Kell blood group protein. ECE-1 cleaves its biologically inactive native substrate, big endothelin-1, to generate a powerful vasoactive 21-amino acid peptide, endothelin-1. ECE-1 consists of a short N-terminal cytoplasmic tail, a transmembrane hydrophobic domain, and a large extracellular domain containing the catalytic site with a conserved Zn-binding motif. We have constructed a secreted, soluble form of ECE-1 (solECE-1) by fusing the cleavable N-terminal signal sequence of human alkaline phosphatase in frame with the entire extracellular domain of ECE-1. Stable transfectant CHO cell lines expressing up to 6.1 mg of solECE-1 per liter culture medium were established and solECE-1 was purified to homogeneity using three chromatographic steps with a 24% yield. SolECE-1 behaves as a dimer of 110-kDa subunits. SolECE-1 has a sharp pH optimum, similar to the native form, ECE-1a, but has a slightly more acidic pH optimum of 6.1– 6.4 than that of 6.7– 6.9 for ECE-1a. At its optimal pH of 6.4, solECE-1 cleaved big ET-1:big ET-2:big ET-3 in a ratio of 8.1:1:1.4, was inhibited by phosphoramidon with an IC50 value of 0.35 6 0.05 mM, had a Km value of 4.65 6 0.78 mM for big ET-1, and had a kcat value of 5.82 6 0.21 min21, all values comparable to those for ECE-1a at its optimal pH of 6.8. Phosphoramidon inhibition of both ECE-1a and solECE-1 is highly pH-dependent. At pH 5.8, phosphoramidon inhibited ECE-1a and solECE-1 with IC50 values of 14 and 33 nM, respectively, which are 49- and 1224-fold more potent than at pH 7.2. SolECE-1 is highly glycosylated, similar to ECE-1a. Deglycosyla1 To whom correspondence should be addressed. Fax: (734) 6221355. E-mail: [email protected].
258
tion of solECE-1 by peptide N-glycosidase F shifted the apparent molecular weight of solECE-1 to approximately 80 kDa and the deglycosylated form(s) of solECE-1 preserved at least 72% of the activity of the glycosylated form. © 1998 Academic Press Key Words: soluble endothelin-converting enzyme-1; endothelin; pH sensitivity; glycosylation
Endothelin-1 (ET-1)2 is a potent vasoactive 21-amino acid peptide which was first discovered in the culture medium of porcine aortic endothelial cells (1). Subsequently, two other closely related 21-amino acid peptides, ET-2 and ET-3, which differ from ET-1 by two and six amino acids, respectively, were identified to be encoded by distinct genes (2). Endothelins are produced from peptide precursors of approximately 200amino acid residues. They are first processed by prohormone-processing enzyme(s) (3) into biologically inactive 38- to 41-amino acid intermediates called big ET-1, -2, and -3. Endothelin-converting enzyme (ECE) then cleaves these big endothelins between Trp21 and Val22/Ile22 to generate endothelins with potent biological activity including vasoconstriction and mitogenesis (1, 4). The development of agents to block either the actions of ET-1 by ET receptor antagonists or the production of ET-1 by ECE inhibitors has been the subject of much 2
Abbreviations used: big ET, big endothelin; ECE, endothelinconverting enzyme; ELISA, enzyme-linked immunosorbant assay; EIA, enzyme immunoassay; ET, endothelin; NEP, neprilysin (neutral endopeptidase 24.11, EC 3.4.24.11); PAGE, polyacrylamide gel electrophoresis; PNGase F, peptide N-glycosidase F; SDS, sodium dodecyl sulfate; solECE-1, soluble form of endothelin-converting enzyme-1; CHO, Chinese hamster ovary; MALDI, matrix-assisted laser desorption ionization; WGA, wheat germ agglutinin. 0003-9861/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
SOLUBLE ENDOTHELIN-CONVERTING ENZYME-1
research interest (5–7). Recent studies using ET inhibitors suggest that endothelins may play important roles in a number of pathological conditions including stroke (8), chronic heart failure (9), and hypertension (10). In addition, disruptions of ET-1 (11), ET-3 (12), and ETB receptor genes (13) have demonstrated important roles for endothelins in the development of neural crest-derived tissues. ECE has been shown to be inhibited by phosphoramidon, a nonselective metalloprotease inhibitor, but it is insensitive to inhibitors of other metalloproteases such as thiorphan (neutral endopeptidase 24.11, NEP) and captopril (angiotensin-converting enzyme) (14 –20). Purification of ECE-1 to homogeneity from rat lung (15) and porcine aortic endothelium (16) rapidly led to cloning of the gene for one of the isoenzymes, ECE-1 (17, 18). Subsequently, cDNA for the human enzyme was also obtained (21, 22). Three different isoforms of ECE-1 have been identified and termed ECE-1a, ECE1b, and ECE-1c. All three isoforms differ only in their N-terminal region and have been shown to be encoded by the same gene through the use of alternative promoters (23–25). Another ECE family member was also cloned and termed ECE-2. ECE-2 was shown to have an overall 59% amino acid identity to ECE-1, to have an acidic pH optimum in contrast to the neutral pH optimum for ECE-1, and to be expressed at a much lower level than ECE-1 in various tissues (26). The biochemistry and molecular pharmacology of ECE have been reviewed (27–29). ECE-1 is a type II integral membrane protein with a short N-terminal cytoplasmic tail (56, 68, and 52 amino acids for human ECE-1a, ECE-1b, and ECE-1c, respectively), a single transmembrane domain (21 amino acids), and a large extracellular domain (681 amino acids) containing the catalytic site with a conserved Znbinding motif, HEXXH (17, 18, 21–25). ECE-1 shows significant similarities, particularly in the C-terminal region, to neprilysin (NEP, neutral endopeptidase 24.11, 3.4.24.11) and the human Kell group protein. Within the C-terminal one-third of the extracellular domain, ECE-1 exhibits 58 and 36% amino acid identity to NEP and Kell, respectively. ECE-1 has 10 sites predicted to be N-glycosylated and a high glycosylation level is evident from the apparent molecular weight of 120 –130 kDa estimated from SDS–PAGE. The predicted molecular weight based on amino acid content alone is 85.6 kDa. ECE-1 was reported to form a disulfide-linked dimer because ECE-1 runs as a 120- to 130-kDa or 250- to 300-kDa protein on SDS–PAGE under reducing or nonreducing conditions, respectively (18, 21). Studies using site-directed mutagenesis have indicated the involvement of Cys412 in a dimer formation (30). Further studies identified Glu651 as the third ligand for Zn. Glu592 and His716 were shown to be
259
essential as two mutants, E592Q and H716Q, had a complete loss of ECE-1 activity (30). Purification of the membrane-bound native form of ECE-1 in sufficient quantity to allow thorough study has been difficult. To achieve homogeneity, ECE-1 reportedly required 6300- and 12,000-fold purification from rat lung (15) and porcine aortic endothelium (16), respectively. It is expected that isolation from mammalian cells expressing the recombinant enzyme will likewise require extensive purification. In addition, the requirement of detergent for solubilization of the native form complicates purification. In the present study, we have engineered human ECE-1 to be produced as a secreted, soluble form by fusing the cleavable N-terminal signal sequence of human alkaline phosphatase with the entire extracellular domain of ECE-1. Similar work has been reported for the expression of a secreted, soluble form of NEP by fusing a signal peptide (pro-opiomelanocortin) with the extracellular domain of NEP. Both mammalian and baculovirus/insect cell expressions of this construct have been shown to produce a fully active soluble NEP (31, 32). We report here establishment of a stable transfectant Chinese hamster ovary (CHO) cell line expressing high levels of soluble ECE-1 (solECE-1) and the first purification of solECE-1 to homogeneity. Studies of solECE-1 were carried out in parallel with the native form, ECE-1a. Our results suggest that all the characteristics of the native membrane-bound form at its optimal pH of 6.7– 6.9 are retained with solECE-1, but at a slightly more acidic pH optimum of 6.1– 6.4. Our data demonstrate that ECE-1, both the native and soluble forms, is highly pH-sensitive. Evidence is also presented that N-glycosylation is not essential for enzyme activity. EXPERIMENTAL PROCEDURES Materials. Human big ET-1 (1–38), big ET-2 (1–39), and big ET-3 (1– 41 amide), ET-1 (1–21), ET-2 (1–21), and ET-3 (1–21) were purchased from Peptide International (Louisville, KY). Phosphoramidon, pepstatin A, phenylmethylsulfonyl fluoride, leupeptin, peptide N-glycosidase F, endo-b-N-acetylglucosaminidase H, and O-glycopeptide endo-D-galactosyl-N-acetyl-a-galactosamino hydrolase were from Boehringer-Mannheim. Thiorphan and 1,10-phenanthroline were from Sigma and Aldrich, respectively. Polyoxyethylene-10-lauryl ether (C12E10) was from Calbiochem. Alkyl Superose (HR5/5) and Superose-6 (HR10/30) fast protein liquid chromatography columns, and molecular mass standards were from Pharmacia LKB. DEAE and wheat germ agglutinin agarose were from Toso-Haas and Seikagaku, respectively. Precast SDS–PAGE gels and protein standards were from Novex. Tissue culture medium, G418, and lipofectamine were purchased from Gibco/BRL. Zeocin was from Invitrogen. A highly purified crystallized preparation of trypsin (Sigma, T7418) was used for all the tissue culture procedures as described (18). ECE-1 assay. The typical reaction mixture (100 ml) contained 100 mM buffer at the indicated pH, 0.1 mM big ET-1 (1–38), 0.015– 0.1% C12E10, 50 mM pepstatin A, 100 mM leupeptin, 200 mM phenylmethlysulfonyl fluoride, and the indicated amount of either ECE-1a
260
AHN, HERMAN, AND FAHNOE
or solECE-1. The solubilized membrane fraction of CHO/human ECE-1a cells was prepared as described (18) and was used as a source of the native enzyme, ECE-1a. It has been shown that the crude homogenate of untransfected CHO cells exhibits no detectable ECE-1 activity (18). Enzymes were diluted with 20 mM Tris–HCl (pH 7.4)/0.1% C12E10. After incubation for 1 h at 37°C, the reaction was stopped by adding EDTA to give a final concentration of 5 mM. This final mixture was then analyzed for the amount of ET-1 by enzyme-linked immunosorbant assay (ELISA) or enzyme immunoassay (EIA). ET measurements. ET-1 was measured by ELISA using the method as instructed by the manufacturer (Amersham) or by EIA. ET-3 was measured by EIA using the method as instructed by the manufacturer (Cayman, Ann Arbor, MI). For the measurement of ET-1 and ET-2 by EIA, the following modifications were made: ET samples were added to the plate and incubated for 30 min at 37°C. After washing the plate, the acetlylcholinesterase-linked antibody was added and incubated for 30 min at 37°C before the addition of Ellman’s reagent. Plasmid construction. A cDNA for a secreted, soluble form of ECE-1 was obtained by fusing a cDNA fragment encoding the entire extracellular domain of human ECE-1a in frame with cDNA encoding human alkaline phosphatase signal sequence. To create a SmaI site at the 59 end of the extracellular domain, a 534-bp cDNA fragment from the beginning of the extracellular domain was amplified from cDNA for human ECE-1a (provided by Dr. M. Yanagisawa) using primers of 59-ATCGATCCCGGGCAGTACCAGACAAGATCCCC-39 and 59-ATAGTCTCTCGAGGGCAAGCC-39. The cDNA fragment encoding the alkaline phosphatase signal sequence was amplified from pPbac (Stratagene) using primers of 59-ACCGATGAATTCGCCACCATGGTGGGGCCCTGC-39 (underlined nucleotides; Kozak sequence, ref 33) and 59-TAGCCACCCGGGGTTCTCCTCCTCAACTGGGAT-39 to create EcoRI and SmaI sites at the 59 and 39 ends, respectively. The alkaline phosphatase–ECE-1 cDNA fusion was created by joining the two DNA fragments using the SmaI sites. The cDNA was then digested with EcoRI and subcloned into the mammalian expression vector, pSG5 (Stratagene, SV40 promotor-driven), to create pSG5-solECE-1. This cDNA encodes alkaline phosphatase (amino acids 1 to 32) and ECE-1a (amino acids 78 to 758). Stable expression of solECE-1 in CHO cells. The pSG5-solECE-1 was cotransfected with pZeoSV (Invitrogen) into CHO-K1 cells using lipofectamine. Zeocin-resistant (0.5 mg/ml) colonies were isolated by trypsinization in a cloning cup. Cell colonies expressing ECE-1 activity were selected and subcloned using the limiting dilution technique. Cell lines which express the highest ECE-1 activity were further selected by seeding the each stable cell line onto 12-well plates. At confluency, serum-free medium (750 ml per well) was added and incubated for 48 h. The culture supernatant was collected and used for ECE-1 assays. Cell culture. CHO-K1 cells were cultured in monolayers in Ham F-12 and Dulbecco’s modified Eagle’s medium (1:1 mixture) supplemented with 10% fetal bovine serum. CHO/ECE-1a cells were grown in the same medium as above including 1 mg/ml G418. CHO/ solECE-1 cells were cultured either in the same medium as CHO-K1 cells or serum-free CHO-SFM II including 0.5 mg/ml Zeocin. All cells were grown in a humidified incubator at 37°C in an atmosphere of 5% CO2/95% air. Antibody and Western blot. Antibody against ECE-1 was provided by Dr. M. Yanagisawa and was produced by immunizing rabbits with a synthetic peptide corresponding to the C-terminal 16 amino acids of bovine ECE-1 as previously described (26). The Western blot analysis was performed with horseradish peroxidase conjugated anti-rabbit IgG by using an ECL detection kit (Amersham). Other methods. SDS–PAGE and staining by Coomassie brilliant blue were performed as described (34). The protein concentration
was determined by the Bradford method (35) with bovine serum albumin as standard. The concentration of ECE-1a from the solubilized membrane fraction of CHO/ECE-1a cells was determined by Western blotting using purified solECE-1 as standard. The stock concentrations of big ET and ET peptides used in the experiments to determine isopeptide substrate selectivity and kinetic studies were determined by quantitative amino acid analysis.
RESULTS
Expression of solECE-1 in CHO Cells When stable transfectant CHO cell lines which express solECE-1 (CHO/solECE-1) were selected for their highest ECE-1 activity by using the method described under Experimental Procedures, the best six CHO/ solECE-1 cell lines expressed ECE-1 activity in a range from 0.49 to 1.11 pmol/min/ml medium (at pH 7.2). This range of ECE-1 activity corresponded to 2.7– 6.1 mg of solECE-1 per liter culture medium when estimated from that of solECE-1 purified to homogeneity (see below). Purification of solECE-1 to Homogeneity SolECE-1 was purified to homogeneity, as judged by SDS–PAGE visualized by Coomassie blue staining (Fig. 1). The purification scheme employed is summarized in Table I. Purification to homogeneity required only a 20-fold overall refinement representing a 24% yield of solECE-1 from the culture medium. The Nterminal amino acid sequence of the purified solECE-1 was I-I-P-V-E-E-E as expected from the cleavage between Gly22 and Ile23 of the human placental alkaline phosphatase signal sequence. Characterization of Purified solECE-1 SolECE-1 behaves as a dimer. An apparent molecular weight of 110 –120 kDa as estimated from SDS– PAGE gives evidence that solECE-1 is highly glycosylated (Fig. 1). The predicted molecular weight of solECE-1 based on amino acid content alone is 78.6 kDa. A matrix-assisted laser desorption ionization (MALDI) mass spectrum of solECE-1 showed (M 1 H)1, (M 1 2H)1, and (2M 1 H)1 ions for the protein consistent with a molecular weight of approximately 107.6 kDa. Broad peak widths of approximately 10 kDa indicated a large degree of heterogeneity in the glycosylation level of solECE-1. Under native conditions, solECE-1 behaves as a dimer as shown by gel filtration chromatography on Superose-6. Comparison of the Superose-6 elution volume of solECE-1 with those of reference proteins indicated an apparent molecular mass for solECE-1 of 232 kDa; the recovery of the activity was 40% (Fig. 2A). The Stokes radius of solECE-1 was estimated to be 52.3 Å (Fig. 2B).
261
SOLUBLE ENDOTHELIN-CONVERTING ENZYME-1
FIG. 1. SDS–PAGE analysis of solECE-1. The SDS–PAGE (4 – 20%) was carried out as described (34). Samples were precipitated with 15% trichloroacetic acid and resuspended in 15 ml of sampleloading buffer. Proteins were visualized by Coomassie blue staining: fraction II (Fr II), 13.9 mg; fraction III (Fr III), 4.5 mg; and fraction IV (Fr IV), 2.4 mg. Fractions are shown in Table I. Marker proteins (M) were myosin (200 kDa), b-galactosidase (116 kDa), phosphorylase b (97 kDa), bovine serum albumin (66 kDa), glutamic dehydrogenase (55 kDa), lactate dehydrogenase (37 kDa), carbonic anhydrase (31 kDa), trypsin inhibitor (22 kDa), and lysozyme (14 kDa).
The pH optimum for solECE-1 is slightly more acidic than that for ECE-1a. To accurately determine the pH optimum for each enzyme, buffers were prepared which varied by approximately 0.2 pH units extending from pH 5.4 to 8.0. As shown in Fig. 3, both ECE-1a and solECE-1 possessed very narrow pH optima. Different buffers (Mes-KOH, KPi, and Hepes-KOH) gave slightly different ECE-1 activities but the sharp pH optimum of similar ranges was observed. The width of the peak at half-maximum activity was approximately 1 pH unit. The pH optimum of 6.7– 6.9 for ECE-1a agreed well with the reported neutral pH optimum for the native enzyme. The optimum pH for solECE-1 was at 6.1– 6.4, slightly more acidic than that for ECE-1a. Substrate specificity. ECE-1a or solECE-1 was incubated with 1 mM big ET-1 (1–38), big ET-2 (1–38), or big ET-3 (1– 41, amide) for up to 4 h at pH values of 6.4 and 6.8. The amount of ET-1, -2, and -3 generated was linear with incubation time under the assay conditions and the results at the respective optimal pH values are shown in Fig. 4. These data show that both enzymes have a strong substrate preference for big ET-1. At its optimal pH 6.4, solECE-1 cleaved big ET-1:big ET-2: big ET-3 in a ratio of 8.1:1:1.4, similar to that of 7.6:1: 1.3 for ECE-1a at its optimal pH (6.8). Each enzyme showed a markedly different substrate selectivity at the pH optimum of the other. SolECE-1 at pH 6.8 and ECE-1a at pH 6.4 cleaved big ET-1:big ET-2:big ET-3 in ratios of 25.9:1:1.7 and 3.1:1:1.4, respectively. Effects of protease inhibitors. The purified solECE-1 was inhibited by EDTA, 1,10-phenanthroline, phos-
TABLE I
Purification of Human solECE-1
Fraction
Total volume (ml)
Total protein (mg)
Total activitya (pmol/min)
Sp act (pmol/min/mg)
Yield (%)
I. Medium II. DEAE III. Wheat germ agglutinin (WGA) IV. Alkyl
1,400 84 32.5 5.1
106.40 12.26 4.52 1.29
957.6 595.8 400.7 232.1
9.00 48.60 88.65 179.92
100 62.2 41.8 24.2
Note. A stable transfectant CHO cell line which expresses a secreted, soluble form of ECE-1 (CHO/solECE-1) was grown under conditions described under Materials and Methods (15 3 600 2cm). At confluency, the cells were washed with phosphate-buffered saline and incubated for 48 h in serum-free CHO-SFM II medium (Gibco/BRL). All subsequent operations were at 0 – 4°C unless otherwise noted. After removal of cell debris by centrifugation (5000g), the supernatant (fraction I, 1400 ml) was diluted twofold with buffer A and was applied to a 100-ml DEAE column (1.6 3 50 cm) equilibrated with buffer A. The column was washed with 1100 ml of the equilibration buffer, and the activity was eluted with a 650-ml gradient (0–250 mM NaCl in buffer A). Fractions of 12 ml were collected and peak fractions were pooled (fraction II, 84 ml). For WGA, 30% of fraction II was loaded onto a 17-ml WGA column (1.6 3 8.5) equilibrated with buffer A containing 150 mM NaCl and the column was washed with 350 ml of the equilibration buffer. The activity was eluted with a 64-ml gradient (0 –50 mg/ml N-acetylglucosamine in buffer A). Two more runs were repeated with 70% of fraction II and peak fractions were pooled from three WGA chromatographic steps (fraction III, 32.5 ml). Fraction III was adjusted to 2.4 M ammonium sulfate. After centrifugation (16,000g, 15 min), 40% of the supernatant was loaded onto a 1-ml Alkyl Superose column (HR5/5) equilibrated with buffer B containing 2 M ammonium sulfate. The column was washed with 20 ml of the equilibration buffer. The activity was eluted with a 15-ml gradient (2-0 M ammonium sulfate in buffer B). Two more runs were repeated with 60% of fraction III and peak fractions from three Alkyl Superose columns were pooled (fraction IV, 5.1 ml). Buffers A and B are Tris–HCl (pH 7.4) at 20 and 50 mM, respectively. a ECE-1 activity was determined at pH 7.2 (Hepes-KOH) as described under Experimental Procedures and therefore is much lower than that at its optimal pH of 6.4 (see Fig. 3).
262
AHN, HERMAN, AND FAHNOE
FIG. 2. Gel filtration (A) and determination of the Stokes radius (B) of solECE-1. 3.0 mg of solECE-1 was applied to a Superose-6 fast protein liquid chromatography column equilibrated with 50 mM Tris–HCl (pH 7.4)/150 mM NaCl. Elution was performed with the equilibration buffer at a flow rate of 0.3 ml/min; 0.38-ml fractions were collected. ECE-1 activity was determined at pH 7.0 as described under Experimental Procedures. Marker proteins were detected by monitoring UV absorbances at 280 nm. Abbreviations are ferritin (ferr, 440 kDa, 61.0 Å); catalase (cat, 232 kDa, 52.2 Å); aldolase (aldo, 158 kDa, 48.1 Å); bovine serum albumin (BSA, 67 kDa, 35.5 Å); ovalbumin (oval, 43 kDa, 30.5 Å); ribonuclease A (RNase A, 13.7 kDa, 16.4 Å).
phoramidon, and PD 069185, a novel highly specific ECE-1 inhibitor (36), but was insensitive to thiorphan and captopril, similar to the native enzyme, ECE-1a (Table II). Phosphoramidon inhibition of ECE-1 is highly pHdependent. Our initial data at neutral pH indicated that phosphoramidon inhibited ECE-1a almost 10-fold better than solECE-1. These results prompted us to further investigate the pH dependence of phosphoramidon inhibition for both ECE-1a and solECE-1. As shown in Fig. 5, phosphoramidon inhibited ECE-1a with IC50 values of 0.014 6 0.002, 0.114 6 0.014, 0.249 6 0.031, and 0.675 6 0.061 mM at pH values of 5.8, 6.4, 6.8, and 7.2, respectively. Corresponding IC50 values with solECE-1 were 0.033 6 0.004, 0.349 6 0.045, 2.12 6 0.43, and 40.4 6 3.7 mM at pH values of 5.8, 6.4, 6.8, and 7.2, respectively. In both cases, the potency of phosphoramidon dramatically decreased from pH 5.8 to 7.2 and this dependence of inhibition upon pH was far more extensive with solECE-1 than with ECE-1a, decreasing 49- or 1225-fold for ECE-1a or solECE-1, respectively, from pH 5.8 to 7.2. At pH 5.8, there was a 2.4-fold potency difference between ECE-1a and solECE-1. This potency difference increased rapidly with increasing pH such that at pH 7.2, there was a 59-fold difference. Nevertheless, when compared at their respective optimal pH values, phos-
phoramidon potency was very similar; an IC50 value of 0.35 mM for solECE-1 at its optimal pH (6.4) was similar to that of 0.25 mM for ECE-1a at its optimal pH (6.8). These potency shifts are not due to changes in substrate Km values because the IC50 values for phosphoramidon, a competitive inhibitor (25), are very close to Ki values under the assay conditions where the substrate concentration is well below the Km values (see below) according to the equation, Ki 5 IC50/(1 1 [S]/Km) (37). In addition, the potency difference between ECE-1a and solECE-1 is not due to the presence/ absence of detergent since the enzyme reactions for both solECE-1 and ECE-1a were carried out in the presence of 0.1% C12E10. Furthermore, the concentration variation of C12E10 from 0.016 to 0.5% did not have any effect on the potency of phosphoramidon for either ECE-1a or solECE-1. Kinetic parameters. The high pH dependence of the phosphoramidon potency as well as that of the hydrolysis ratios for big ET substrates prompted us to measure Km and kcat at several pH values. The velocity versus big ET-1 concentration curves at pH 6.4, 6.8, and 7.1 were fit to Michaelis–Menten kinetics as shown in Fig. 6. The Km and kcat values thus obtained were also pH-dependent as listed in Table III. Under the pH conditions tested, the tightest observed substrate binding was obtained at the respective optimal pH values
SOLUBLE ENDOTHELIN-CONVERTING ENZYME-1
263
FIG. 3. pH dependence of ECE-1 activity. ECE-1a or solECE-1, 10 ng of each, was incubated at 37°C for 1 h in a reaction mixture (100 ml) containing 0.1 mM big ET-1, 0.1% C12E10, 50 mM pepstatin A, 100 mM leupeptin, 200 mM phenylmethlysulfonyl fluoride, and a buffer at 100 mM as follows: pH 5.31–7.16 (Mes-KOH), pH 5.92–7.74 (KPi), and pH 6.47–7.96 (Hepes-KOH). The pH values of all the buffers were measured at 100 mM and at 37°C. The amount of ET-1 generated was analyzed by ELISA.
for ECE-1a and solECE-1. At its optimal pH (6.4), the Km and kcat values of 4.65 mM and 5.82 min21 for solECE-1 were similar to those of 2.70 mM and 6.72 min21 for ECE-1a at its optimal pH 6.8. At the nonoptimal pH of 7.1, Km values were 10.4- and 7.4-fold higher for solECE-1 and ECE-1a, respectively, than those at the respective optimal pH values. N-glycosylation is not essential for enzyme activity. In order to remove glycosylation from solECE-1 as completely as possible, glycosidases were incubated at 37°C with fully denatured solECE-1, which was prepared by boiling it in the presence of SDS. Peptide N-glycosidase F (PNGase F), endo-b-N-acetylglucosaminidase H, and Oglycopeptide endo-D-galactosyl-N-acetyl-a-galactosamino hydrolase were incubated with the denatured solECE-1 for 12 h at 37°C. Western blot analysis showed that the 110 kDa band for solECE-1 was completely shifted to the 80-kDa band(s) after incubation with PNGase F which cleaves all Asp-bound N-glycans. Endo-b-N-acetylglucosamidase H and O-glycopeptide endo-D-galactosylN-acetyl-a-galactosamino hydrolase, which remove Asplinked high mannose oligosaccharides and O-glycosylation, respectively, did not shift the molecular weight of solECE-1 under similar conditions (data not shown). Next, in order to find out if N-glycosylation is required for ECE-1 activity, solECE-1 was incubated under the native conditions with PNGase F at 37°C. The commercial PNGase F (Boehringer-Mannheim) is supplied in a buffer
containing 25 mM EDTA and thus was dialyzed extensively against 100 mM KPi (pH 7.5)/10% glycerol to remove EDTA before use. Both the extent of deglycosylation by Western blot and ECE-1 activity were tested at the indicated time. As shown in Fig. 7A, after 2 h incubation with PNGase F, the 110 kDa band for solECE-1 was completely shifted to the 80 kDa band(s), similar to those observed with fully denatured solECE-1. At 30 min, 1 h, and 2 h, the solECE-1 activity was 91.8, 78.4, and 72.0% of the control activity (without PNGase F), respectively (Fig. 7B). These data show that a substantial fraction of ECE-1 activity still remained after the removal of N-glycosylation. There was no self-digestion of solECE-1 under the deglycosylation conditions as shown in the control (Fig. 7A). DISCUSSION
Detailed data on enzymatic characterization of ECE-1 has been limited mostly due to the difficulty in obtaining homogeneous enzyme in sufficient quantities. The reported literature data on big ET isopeptide specificity, inhibition potency of phosphoramidon, and kinetic constants vary widely presumably because most of the early data were obtained using partially purified enzyme which may have been contaminated with other proteases. However, even the data obtained from ECE-1 which was either purified to homogeneity or the
264
AHN, HERMAN, AND FAHNOE
FIG. 4. Substrate specificity. ECE-1a or solECE-1, 30 ng of each, was incubated at 37°C for the indicated period of time in a reaction mixture (100 ml) containing 1 mM big ET-1, -2, or -3, 0.1% C12E10, 50 mM pepstatin A, 100 mM leupeptin, 200 mM phenylmethlysulfonyl fluoride, and 100 mM Hepes-KOH (pH 6.8) for ECE-1a or 100 mM Mes-KOH (pH 6.4) for solECE-1. The same stock solutions of big ET and assay components were used for both ECE-1a and solECE-1 for a better comparison. The amount of ET-1, -2, and -3 was analyzed by EIA.
recombinant enzyme expressed in mammalian cells vary over a wide range under similar assay conditions. For example, there is a large variation in the reported ability of ECE-1 to process big ET isopeptides. With purified rat lung ECE-1, the relative conversion rates for big ET-1:big ET-2:big ET-3 were reported to be 4:2:1 (15). For ECE-1 purified from porcine aortic endothelium, big ET-1 was a much preferred substrate over big
TABLE II
Effects of Protease Inhibitors at 100 mM ECE-1 activity (% of control) Inhibitor (100 mM)
ECE-1a
solECE-1
Phosphoramidon EDTAa PD 069185 1,10-Phenanthroline Thiorphan Captopril
2.9 29.8 4.3 0.7 52.0 95.8
3.2 14.4 5.7 1.6 70.0 106.9
Note. ECE-1a or solECE-1, 10 ng of each, was incubated in a reaction mixture (100 ml) containing 0.1 mM big ET-1, 0.1% C12E10, 50 mM pepstatin A, 100 mM leupeptin, 200 mM phenylmethlysulfonyl fluoride, and 100 mM Hepes-KOH (pH 6.8) for ECE-1a or 100 mM Mes-KOH (pH 6.4) for solECE-1. The final dimethyl sulfoxide concentration was 1.5%. a At 1 mM.
ET-2 (12.5:1). However, it did not convert big ET-3 (16). Bovine ECE-1 expressed in CHO cells processed all three big ET isopeptides, with the conversion rates big ET-1 . big ET-3 . big ET-2 (18). These discrepancies in the reported relative values of big ET conversion might possibly reflect variations in the specificity of several isotypes of ECE-1 for substrate. However, it has been shown that there is no substrate specificity difference at least between ECE-1a and ECE-1b of either human or rat (23). Likewise, the reported Km values for big ET-1 range widely from 0.2 mM (ECE-1 purified from rat lung) to 23 mM (recombinant human ECE-1 expressed in CHO cells) (15, 21). Vmax values of 3.1 and 410 nmol/min/mg protein were reported for the purified ECE-1 from rat lung and porcine aortic endothelium, respectively (15, 16). The reported IC50 values for phosphoramidon with ECE-1 also vary from 0.35 to 8 mM (15, 21). It is unlikely that the wide variations in the literature data discussed above are due to the species difference since similar big ET isopeptide specificity was observed with the human and rat enzymes (23). Likewise, similar inhibitor potencies and kinetic parameters were obtained with the human and bovine enzymes (K. Ahn, S. Herman, and D. Fahnoe, unpublished results). In the current study, we have constructed a mammalian expression system by fusing a signal peptide in frame with the extracellular domain of ECE-1. This con-
SOLUBLE ENDOTHELIN-CONVERTING ENZYME-1
265
FIG. 5. pH-dependence of phosphoramidon inhibition. ECE-1a or solECE-1, 10 ng of each, was incubated at 37°C for 1 h in a reaction mixture (100 ml) containing 0.1 mM big ET-1, 0.1% C12E10, 50 mM pepstatin A, 100 mM leupeptin, 200 mM phenylmethlysulfonyl fluoride, and a buffer at 100 mM as follows: pH 5.8 and 6.4, Mes-KOH; pH 6.8 and 7.2, Hepes-KOH. The same stock solutions of compounds and assay components were used for both ECE-1a and solECE-1 for a better comparison. Under the assay conditions, the production of ET-1 is linear with an incubation time of several hours. The data were plotted as percentage of inhibition vs inhibitor concentration and fit with the equation, y 5 100/1 1 (x/IC50)z, using KaleidaGraph (Synergy Software, Reading, PA), where IC50 is the inhibitor concentration at 50% inhibition and z is the slope of the inhibition curve.
struct allowed the secretion of a fully active enzyme after transfection into CHO cells and a relatively easy homogeneous purification of solECE-1 in milligram quantities from the culture supernatant. All the studies of solECE-1 were carried out in parallel with ECE-1a for comparison. Accurate concentrations of big ET and ET stocks were determined by quantitative amino acid analysis for the experiments described in this paper. In our laboratory,
we have observed discrepancies between the stated weight indicated by the commercial sources of theses peptides, usually supplied as preweighed lyophilized powders, and the actual weight determined in solution. Furthermore, the peptides may not fully solubilize in the chosen buffer. Inaccuracy in respect to peptide concentrations may have also contributed to the variations in the literature data.
FIG. 6. Km and kcat measurements. ECE-1a (4.7 ng) or solECE-1 (9.6 ng) was incubated at 37°C for 30 min in a reaction mixture (50 ml) containing 0.1% C12E10, 50 mM pepstatin A, 100 mM leupeptin, 200 mM phenylmethlysulfonyl fluoride, big ET-1 (varied between 0.5 and 140 mM), and a buffer at 100 mM as follows: pH 6.4 and 6.8, Mes-KOH; pH 7.1, Hepes-KOH. The same stock solutions of big ET-1 and assay components were used for both ECE-1a and solECE-1 for a better comparison. The data were fit to the equation, v 5 Vmax [S]/([S] 1 Km), by a nonlinear least-squares algorithm.
266
AHN, HERMAN, AND FAHNOE TABLE III
Kinetic Parameters of ECE-1a and solECE-1 at Various pH Values ECE-1a
solECE-1
pH
Km (mM)
kcat (min21)
Km (mM)
kcat (min21)
6.4 6.8 7.1
7.51 6 1.53 2.70 6 0.36 19.87 6 2.72
3.17 6 0.16 6.72 6 0.16 17.61 6 0.82
4.65 6 0.78 8.20 6 0.95 48.31 6 4.10
5.82 6 0.21 17.77 6 0.54 21.23 6 0.82
Note. The Km and kcat values were measured as described in the legend to Fig. 6. The molecular weights of 85.6 and 78.6 kDa were used for ECE-1a and solECE-1, respectively.
It was surprising to observe a more acidic pH optimum for solECE-1 compared to that for ECE-1a. At the optimal pH of 6.8 for ECE-1a, solECE-1 has only 11– 33% of its optimal activity at pH 6.4 (Fig. 3). A search of the literature for other enzymes whose truncated forms show a shift in pH optimum revealed one example, protein tyrosine phosphatase containing two Src homology 2 domains. The truncated protein tyrosine phosphatase domain was shown to be most active at pH 6.1– 6.5, which is shifted from an optimal pH of 5.5.–5.7 for the full-length enzyme (38). The lack of more examples may derive from the fact that few enzymes have similarly sharp pH optima, thus a shift might not be reflected as a clear difference in activity. The shift in pH optimum prompted us to evaluate hydrolysis rates, potency of phosphoramidon inhibi-
tion, and kinetic parameters at various pH values. At its optimal pH of 6.4, solECE-1 had a hydrolysis ratio for big ET isopeptides, an inhibition potency of phosphoramidon, a Km value for big ET-1, and a kcat value all comparable to those for the native membranebound form, ECE-1a, at its optimal pH of 6.8. Additionally, these data were highly pH dependent and the results obtained at a nonoptimal pH of only 0.4 pH unit away were very different from those at their respective optimal pH values for both solECE-1 and ECE-1a. This high degree of pH dependence could explain further the wide variations in the literature data as discussed above. While this manuscript was in preparation, a report on a similar expression of solECE-1 appeared (39). The authors did not report a pH optimum shift for solECE-1 compared to that for the native enzyme. For their study of solECE-1, the culture supernatant was used as a source of solECE-1 rather than a homogeneous enzyme (39). The pH dependence of phosphoramidon inhibition potency was large, and it was especially dramatic for solECE-1 (Fig. 5). This pH dependence is not due to changes in the ionization state of phosphoramidon. Since the pKa values of the ionizing groups in phosphoramidon, carboxyl and phosphonyl groups, are less than 3 and 3.6, respectively (40), there is very little change in the ionization state of phosphoramidon in the pH range used in these experiments. Interestingly, phosphoramidon was also reported to be a more potent inhibitor of NEP at pH 6.5 than at pH 8.5 by 150-fold
FIG. 7. Deglycosylation of solECE-1. The reaction mixture (25 ml) contained 100 mM Tris–HCl (pH 8.5), 0.1% C12E10, 50 mM pepstatin A, 100 mM leupeptin, 200 mM phenylmethlysulfonyl fluoride, and 150 ng solECE-1 either in the presence or in the absence of PNGase F (0.2 units). After incubation at 37°C for the indicated amount of time, the reaction was stopped by freezing in liquid nitrogen and was stored until the entire set for the time course was collected; 5 ml of the reaction mixture was loaded onto SDS–PAGE (4 –20%) for a Western blot analysis (A). 10 ml of the reaction mixture was used to determine the ECE-1 activity at pH 7.2 (Hepes-KOH) as described under Experimental Procedures (B).
SOLUBLE ENDOTHELIN-CONVERTING ENZYME-1
(41). The nM potency of phosphoramidon reported for ECE-2 as a distinctive difference compared to the mM potency for ECE-1 (26) now appears to be simply a reflection of the different optimal pH values for ECE-1 and ECE-2. An IC50 value of 4 nM with phosphoramidon reported for ECE-2 that was obtained at its optimal pH of 5.6 (26) is also in the same range as the IC50 value of 14 nM for ECE-1a measured at pH 5.8 (Fig. 5). In summary, ECE-1, in both the native and soluble forms, is a highly pH-sensitive enzyme, and it is necessary to be very cautious when acquiring and comparing data. Our results suggest that solECE-1 undergoes a general shift in the apparent pKa values for active site residues yielding a shifted pH rate profile compared to that for ECE-1a although the profile retains similar width. The comparable Km, kcat, and Ki values between ECE-1a and solECE-1 at the respective optimal pH values suggest that the ionized residues at the active site of solECE-1 bind to substrate and inhibitor in the same manner as with the native form. Clearly, further work is warranted and is currently underway in our laboratory. Retention of ECE-1 activity after full deglycosylation is very encouraging with respect to the prospects for crystallization of solECE-1. Structural information on ECE-1 obtainable from X-ray crystallography will be invaluable for a better understanding of the enzyme and in designing potent inhibitors. It has been known that glycosylation is not required for activity in many glycosylated enzymes (42, 43), and successful crystallization of glycosylated enzymes has been possible after deglycosylation (44). With homogeneous enzyme available, many other biochemical characterizations of ECE-1 including the testing of other hormones as possible substrates for ECE-1 and defining preferred residues around the cleavage site are now possible.
5. 6. 7. 8.
9.
10.
11.
12.
13.
14. 15. 16. 17. 18. 19. 20. 21.
ACKNOWLEDGMENTS We thank Masashi Yanagisawa and Noriaki Emoto for providing cDNA for ECE-1a and antibody against ECE-1 and for discussions. We thank Tracy Stevenson and Kenneth Greis for obtaining the MALDI mass spectrum and the N-terminal sequencing of purified solECE-1, respectively. We also thank Jill Knapp and Andre Sisneros for assisting with the assays.
REFERENCES 1. Yanagisawa, M., Kurihara, H., Kimura, S., Tomobe, Y., Kobayashi, M., Mitsui, Y., Yazaki, Y., Goto, K., and Masaki, T. (1988) Nature 332, 411– 415. 2. Inoue, A., Yanagisawa, M., Kimura, S., Kasuya, Y., Miyauchi, T., Goto, K., and Masaki, T. (1989) Proc. Natl. Acad. Sci. USA 86, 2863–2867. 3. Seidah, N. G., Day, R., Marcinkiewicz, M., and Chre´tien, M. (1993) Ann. N. Y. Acad. Sci. 680, 135–146. 4. Yanagisawa, M., Inoue, A., Ishikawa, T., Kasuya, Y., Kimura, S., Kumagaye, S.-I, Nakajima, K., Watanabe, T. X., Sakakibara, S.,
22.
23. 24. 25.
26. 27. 28. 29.
267
Goto, K., and Masaki, T. (1988) Proc. Natl. Acad. Sci. USA 85, 6964 – 6967. Cheng, X.-M., Ahn, K., and Haleen, S. J. (1997) Annu. Rep. Med. Chem. 32, 61–70. Webb, M. L., and Meek, T. D. (1997) Med. Res. Rev. 17, 17– 67. Ohlstein, E. H., Elliott, J. D., Feuerstein, G. Z., and Ruffolo, R. R., Jr. (1996) Med. Res. Rev. 16, 365–390. Patel, T. R., Galbraith, S. L., McAuley, M. A., Doherty, A. M., Graham, D. I., and McCulloch, J. (1995) J. Cardiovasc. Pharmacol. 26(Suppl. 3), S412–S415. Kiowski, W., Su¨tsch, G., Hunziker, P., Mu¨ller, P., Kim, J., Oechslin, E., Schmitt, R., Jones, R., and Bertel, O. (1995) Lancet 346, 732–736. Clozel, M., Breu, V., Burri, K., Cassal, J.-M., Fischli, W., Gray, G. A., Hirth, G., Lo¨ffler, B.-M., Mu¨ller, M., Neidhart, W., and Ramuz, H. (1993) Nature 365, 759 –761. Kurihara, Y., Kurihara, H., Suzuki, H., Kodama, T., Maemura, K., Nagai, R., Oda, H., Kuwaki, T., Cao, W.-H., Kamada, N., Jishage, K., Ouchi, Y., Azuma, S., Toyoda, Y., Ishikawa, T., Kumuda, M., and Yazaki, Y. (1994) Nature 368, 703–710. Baynash, A. G., Hosoda, K., Giaid, A., Richardson, J. A., Emoto, N., Hammer, R. E., and Yanagisawa, M. (1994) Cell 79, 1277– 1285. Puffenberger, E. G., Hosoda, K., Washington, S. S., Nakao, K., deWit, D., Yanagisawa, M., and Chakravarti, A. (1994) Cell 79, 1257–1266. Ahn, K., Beningo, K., Olds, G., and Hupe, D. (1992) Proc. Natl. Acad. Sci. USA 89, 8606 – 8610. Takahashi, M., Matsushita, Y., Iijima, Y., and Tanzawa, K. (1993) J. Biol. Chem. 268, 21394 –21398. Ohnaka, K., Takayanagi, R., Nishikawa, M., Haji, M., and Nawata, H. (1993) J. Biol. Chem. 268, 26759 –26766. Shimada, K., Takahashi, M., and Tanzawa, K. (1994) J. Biol. Chem. 269, 18275–18278. Xu, D., Emoto, N., Giaid, A., Slaughter, C., Kaw, S., deWit, D., and Yanagisawa, M. (1994) Cell 78, 473– 485. Ahn, K., Pan, S., Beningo, K., and Hupe, D. (1995) Life Sci. 56, 2331–2341. Ahn, K., Pan, S., Zientek, M., Guy, P., and Sisneros, A. (1996) Biochem. Mol. Biol. Int. 39, 573–580. Schmidt, M., Kro¨ger, B., Jacob, E., Seulberger, H., Subkowski, T., Otter, R., Meyer, T., Schmalzing, G., and Hillen, H. (1994) FEBS Lett. 356, 238 –243. Shimada, K., Matsushita, Y., Wakabayashi, K., Takahashi, M., Matsubara, A., Iijima, Y., and Tanzawa, K. (1995) Biochem. Biophys. Res. Commun. 207, 807– 812. Shimada, K., Takahashi, M., Ikeda, M., and Tanzawa, K. (1995) FEBS Lett. 371, 140 –144. Valdenaire, O., Rohrbacher, E., and Mattei, M.-G. (1995) J. Biol. Chem. 270, 29794 –29798. Schweizer, A., Valdenaire, O., Nelbo¨ck, P., Deuschle, U., Dumas Milne Edwards, J.-B., Stumpf, J. G., and Lo¨ffler, B.-M. (1997) Biochem. J. 328, 871– 877. Emoto, N., and Yanagisawa, M. (1995) J. Biol. Chem. 270, 15262–15268. Opgenorth, T. J., Wu-Wong, J. R., and Shiosaki, K. (1992) FASEB J. 6, 2653–2659. Turner, A. J., and Murphy, L. J. (1996) Biochem. Pharmacol. 51, 91–102. Ahn, K. (1998) Handbook of Proteolytic Enzymes (Barrett, A. J., Rawlings, N. D., and Woessner, J. F., Eds.), pp. 1085–1091, Academic Press, San Diego.
268
AHN, HERMAN, AND FAHNOE
30. Shimada, K., Takahashi, M., Turner, A. J., and Tanzawa, K. (1996) Biochem. J. 315, 863– 867. 31. Lemay, G., Waksman, G., Roques, B. P., Crine, P., and Boileau, G. (1989) J. Biol. Chem. 264, 15620 –15623. 32. Fossiez, F., Lemay, G., Labonte´, N., Parmentier-Lesage, F., Boileau, G., and Crine, P. (1992) Biochem. J. 284, 53–59. 33. Kozak, M. (1987) Nucleic Acids Res. 15, 8125– 8148. 34. Laemmli, U. K. (1970) Nature 227, 680 – 685. 35. Bradford, M. M. (1976) Anal. Biochem. 72, 248 –254. 36. Ahn, K., Sisneros, A. M., Herman, S. B., Pan, S. M., Hupe, D., Lee, C., Nikam S., Cheng, X.-M., Doherty, A. M., Schroeder, R. L., Haleen, S. J., Kaw, S., Emoto, N., and Yanagisawa, M. (1998) Biochem. Biophys. Res. Commun. 243, 184 –190. 37. Cheng, Y.-C., and Prusoff, W. H. (1973) Biochem. Pharmacol. 22, 3099 –3108.
38. Pei, D., Neel, B. G., and Walsh, C. T. (1993) Proc. Natl. Acad. Sci. USA 90, 1092–1096. 39. Korth, P., Egidy, G., Parnot, C., LeMoullec, J.-M., Corvol, P., and Pinet, F. (1997) FEBS Lett. 417, 365–370. 40. Umezawa, S., Tatsuta, K., Izawa, O., and Tsuchiya, T. (1972) Tetrahedron Lett. 1, 97–100. 41. Jeng, A., Ansell, J., and Erion, M. D. (1989) Life Sci. 45, 2109 – 2114. 42. Malissard, M., Borsig, L., Di Marco, S., Gru¨tter, M. G., Kragl, U., Wandrey, C., and Berger, E. G. (1996) Eur. J. Biochem. 239, 340 –348. 43. Yu, X. C., Sturrock, E. D., Wu, Z., Biemann, K., Ehlers, M. R. W., and Riordan, J. F. (1997) J. Biol. Chem. 272, 3511–3519. 44. Kalisz, H. M., Hecht, H.-J., Schomburg, D., and Schmid, R. D. (1990) J. Mol. Biol. 213, 207–209.
Vol. 359, No. 2, November 15, pp. 258 –268, 1998 Article No. BB980913
Soluble Human Endothelin-Converting Enzyme-1: Expression, Purification, and Demonstration of Pronounced pH Sensitivity Kyunghye Ahn,1 Sarah B. Herman, and Douglass C. Fahnoe Department of Biochemistry, Parke–Davis Pharmaceutical Research, Division of Warner–Lambert Company, Ann Arbor, Michigan 48105
Received June 10, 1998, and in revised form August 24, 1998
Endothelin-converting enzyme-1 (ECE-1) is a type II integral membrane protein that belongs to a family of metalloproteases which includes ECE-2, neprilysin (neutral endopeptidase 24.11, EC 3.4.24.11), and Kell blood group protein. ECE-1 cleaves its biologically inactive native substrate, big endothelin-1, to generate a powerful vasoactive 21-amino acid peptide, endothelin-1. ECE-1 consists of a short N-terminal cytoplasmic tail, a transmembrane hydrophobic domain, and a large extracellular domain containing the catalytic site with a conserved Zn-binding motif. We have constructed a secreted, soluble form of ECE-1 (solECE-1) by fusing the cleavable N-terminal signal sequence of human alkaline phosphatase in frame with the entire extracellular domain of ECE-1. Stable transfectant CHO cell lines expressing up to 6.1 mg of solECE-1 per liter culture medium were established and solECE-1 was purified to homogeneity using three chromatographic steps with a 24% yield. SolECE-1 behaves as a dimer of 110-kDa subunits. SolECE-1 has a sharp pH optimum, similar to the native form, ECE-1a, but has a slightly more acidic pH optimum of 6.1– 6.4 than that of 6.7– 6.9 for ECE-1a. At its optimal pH of 6.4, solECE-1 cleaved big ET-1:big ET-2:big ET-3 in a ratio of 8.1:1:1.4, was inhibited by phosphoramidon with an IC50 value of 0.35 6 0.05 mM, had a Km value of 4.65 6 0.78 mM for big ET-1, and had a kcat value of 5.82 6 0.21 min21, all values comparable to those for ECE-1a at its optimal pH of 6.8. Phosphoramidon inhibition of both ECE-1a and solECE-1 is highly pH-dependent. At pH 5.8, phosphoramidon inhibited ECE-1a and solECE-1 with IC50 values of 14 and 33 nM, respectively, which are 49- and 1224-fold more potent than at pH 7.2. SolECE-1 is highly glycosylated, similar to ECE-1a. Deglycosyla1 To whom correspondence should be addressed. Fax: (734) 6221355. E-mail: [email protected].
258
tion of solECE-1 by peptide N-glycosidase F shifted the apparent molecular weight of solECE-1 to approximately 80 kDa and the deglycosylated form(s) of solECE-1 preserved at least 72% of the activity of the glycosylated form. © 1998 Academic Press Key Words: soluble endothelin-converting enzyme-1; endothelin; pH sensitivity; glycosylation
Endothelin-1 (ET-1)2 is a potent vasoactive 21-amino acid peptide which was first discovered in the culture medium of porcine aortic endothelial cells (1). Subsequently, two other closely related 21-amino acid peptides, ET-2 and ET-3, which differ from ET-1 by two and six amino acids, respectively, were identified to be encoded by distinct genes (2). Endothelins are produced from peptide precursors of approximately 200amino acid residues. They are first processed by prohormone-processing enzyme(s) (3) into biologically inactive 38- to 41-amino acid intermediates called big ET-1, -2, and -3. Endothelin-converting enzyme (ECE) then cleaves these big endothelins between Trp21 and Val22/Ile22 to generate endothelins with potent biological activity including vasoconstriction and mitogenesis (1, 4). The development of agents to block either the actions of ET-1 by ET receptor antagonists or the production of ET-1 by ECE inhibitors has been the subject of much 2
Abbreviations used: big ET, big endothelin; ECE, endothelinconverting enzyme; ELISA, enzyme-linked immunosorbant assay; EIA, enzyme immunoassay; ET, endothelin; NEP, neprilysin (neutral endopeptidase 24.11, EC 3.4.24.11); PAGE, polyacrylamide gel electrophoresis; PNGase F, peptide N-glycosidase F; SDS, sodium dodecyl sulfate; solECE-1, soluble form of endothelin-converting enzyme-1; CHO, Chinese hamster ovary; MALDI, matrix-assisted laser desorption ionization; WGA, wheat germ agglutinin. 0003-9861/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
SOLUBLE ENDOTHELIN-CONVERTING ENZYME-1
research interest (5–7). Recent studies using ET inhibitors suggest that endothelins may play important roles in a number of pathological conditions including stroke (8), chronic heart failure (9), and hypertension (10). In addition, disruptions of ET-1 (11), ET-3 (12), and ETB receptor genes (13) have demonstrated important roles for endothelins in the development of neural crest-derived tissues. ECE has been shown to be inhibited by phosphoramidon, a nonselective metalloprotease inhibitor, but it is insensitive to inhibitors of other metalloproteases such as thiorphan (neutral endopeptidase 24.11, NEP) and captopril (angiotensin-converting enzyme) (14 –20). Purification of ECE-1 to homogeneity from rat lung (15) and porcine aortic endothelium (16) rapidly led to cloning of the gene for one of the isoenzymes, ECE-1 (17, 18). Subsequently, cDNA for the human enzyme was also obtained (21, 22). Three different isoforms of ECE-1 have been identified and termed ECE-1a, ECE1b, and ECE-1c. All three isoforms differ only in their N-terminal region and have been shown to be encoded by the same gene through the use of alternative promoters (23–25). Another ECE family member was also cloned and termed ECE-2. ECE-2 was shown to have an overall 59% amino acid identity to ECE-1, to have an acidic pH optimum in contrast to the neutral pH optimum for ECE-1, and to be expressed at a much lower level than ECE-1 in various tissues (26). The biochemistry and molecular pharmacology of ECE have been reviewed (27–29). ECE-1 is a type II integral membrane protein with a short N-terminal cytoplasmic tail (56, 68, and 52 amino acids for human ECE-1a, ECE-1b, and ECE-1c, respectively), a single transmembrane domain (21 amino acids), and a large extracellular domain (681 amino acids) containing the catalytic site with a conserved Znbinding motif, HEXXH (17, 18, 21–25). ECE-1 shows significant similarities, particularly in the C-terminal region, to neprilysin (NEP, neutral endopeptidase 24.11, 3.4.24.11) and the human Kell group protein. Within the C-terminal one-third of the extracellular domain, ECE-1 exhibits 58 and 36% amino acid identity to NEP and Kell, respectively. ECE-1 has 10 sites predicted to be N-glycosylated and a high glycosylation level is evident from the apparent molecular weight of 120 –130 kDa estimated from SDS–PAGE. The predicted molecular weight based on amino acid content alone is 85.6 kDa. ECE-1 was reported to form a disulfide-linked dimer because ECE-1 runs as a 120- to 130-kDa or 250- to 300-kDa protein on SDS–PAGE under reducing or nonreducing conditions, respectively (18, 21). Studies using site-directed mutagenesis have indicated the involvement of Cys412 in a dimer formation (30). Further studies identified Glu651 as the third ligand for Zn. Glu592 and His716 were shown to be
259
essential as two mutants, E592Q and H716Q, had a complete loss of ECE-1 activity (30). Purification of the membrane-bound native form of ECE-1 in sufficient quantity to allow thorough study has been difficult. To achieve homogeneity, ECE-1 reportedly required 6300- and 12,000-fold purification from rat lung (15) and porcine aortic endothelium (16), respectively. It is expected that isolation from mammalian cells expressing the recombinant enzyme will likewise require extensive purification. In addition, the requirement of detergent for solubilization of the native form complicates purification. In the present study, we have engineered human ECE-1 to be produced as a secreted, soluble form by fusing the cleavable N-terminal signal sequence of human alkaline phosphatase with the entire extracellular domain of ECE-1. Similar work has been reported for the expression of a secreted, soluble form of NEP by fusing a signal peptide (pro-opiomelanocortin) with the extracellular domain of NEP. Both mammalian and baculovirus/insect cell expressions of this construct have been shown to produce a fully active soluble NEP (31, 32). We report here establishment of a stable transfectant Chinese hamster ovary (CHO) cell line expressing high levels of soluble ECE-1 (solECE-1) and the first purification of solECE-1 to homogeneity. Studies of solECE-1 were carried out in parallel with the native form, ECE-1a. Our results suggest that all the characteristics of the native membrane-bound form at its optimal pH of 6.7– 6.9 are retained with solECE-1, but at a slightly more acidic pH optimum of 6.1– 6.4. Our data demonstrate that ECE-1, both the native and soluble forms, is highly pH-sensitive. Evidence is also presented that N-glycosylation is not essential for enzyme activity. EXPERIMENTAL PROCEDURES Materials. Human big ET-1 (1–38), big ET-2 (1–39), and big ET-3 (1– 41 amide), ET-1 (1–21), ET-2 (1–21), and ET-3 (1–21) were purchased from Peptide International (Louisville, KY). Phosphoramidon, pepstatin A, phenylmethylsulfonyl fluoride, leupeptin, peptide N-glycosidase F, endo-b-N-acetylglucosaminidase H, and O-glycopeptide endo-D-galactosyl-N-acetyl-a-galactosamino hydrolase were from Boehringer-Mannheim. Thiorphan and 1,10-phenanthroline were from Sigma and Aldrich, respectively. Polyoxyethylene-10-lauryl ether (C12E10) was from Calbiochem. Alkyl Superose (HR5/5) and Superose-6 (HR10/30) fast protein liquid chromatography columns, and molecular mass standards were from Pharmacia LKB. DEAE and wheat germ agglutinin agarose were from Toso-Haas and Seikagaku, respectively. Precast SDS–PAGE gels and protein standards were from Novex. Tissue culture medium, G418, and lipofectamine were purchased from Gibco/BRL. Zeocin was from Invitrogen. A highly purified crystallized preparation of trypsin (Sigma, T7418) was used for all the tissue culture procedures as described (18). ECE-1 assay. The typical reaction mixture (100 ml) contained 100 mM buffer at the indicated pH, 0.1 mM big ET-1 (1–38), 0.015– 0.1% C12E10, 50 mM pepstatin A, 100 mM leupeptin, 200 mM phenylmethlysulfonyl fluoride, and the indicated amount of either ECE-1a
260
AHN, HERMAN, AND FAHNOE
or solECE-1. The solubilized membrane fraction of CHO/human ECE-1a cells was prepared as described (18) and was used as a source of the native enzyme, ECE-1a. It has been shown that the crude homogenate of untransfected CHO cells exhibits no detectable ECE-1 activity (18). Enzymes were diluted with 20 mM Tris–HCl (pH 7.4)/0.1% C12E10. After incubation for 1 h at 37°C, the reaction was stopped by adding EDTA to give a final concentration of 5 mM. This final mixture was then analyzed for the amount of ET-1 by enzyme-linked immunosorbant assay (ELISA) or enzyme immunoassay (EIA). ET measurements. ET-1 was measured by ELISA using the method as instructed by the manufacturer (Amersham) or by EIA. ET-3 was measured by EIA using the method as instructed by the manufacturer (Cayman, Ann Arbor, MI). For the measurement of ET-1 and ET-2 by EIA, the following modifications were made: ET samples were added to the plate and incubated for 30 min at 37°C. After washing the plate, the acetlylcholinesterase-linked antibody was added and incubated for 30 min at 37°C before the addition of Ellman’s reagent. Plasmid construction. A cDNA for a secreted, soluble form of ECE-1 was obtained by fusing a cDNA fragment encoding the entire extracellular domain of human ECE-1a in frame with cDNA encoding human alkaline phosphatase signal sequence. To create a SmaI site at the 59 end of the extracellular domain, a 534-bp cDNA fragment from the beginning of the extracellular domain was amplified from cDNA for human ECE-1a (provided by Dr. M. Yanagisawa) using primers of 59-ATCGATCCCGGGCAGTACCAGACAAGATCCCC-39 and 59-ATAGTCTCTCGAGGGCAAGCC-39. The cDNA fragment encoding the alkaline phosphatase signal sequence was amplified from pPbac (Stratagene) using primers of 59-ACCGATGAATTCGCCACCATGGTGGGGCCCTGC-39 (underlined nucleotides; Kozak sequence, ref 33) and 59-TAGCCACCCGGGGTTCTCCTCCTCAACTGGGAT-39 to create EcoRI and SmaI sites at the 59 and 39 ends, respectively. The alkaline phosphatase–ECE-1 cDNA fusion was created by joining the two DNA fragments using the SmaI sites. The cDNA was then digested with EcoRI and subcloned into the mammalian expression vector, pSG5 (Stratagene, SV40 promotor-driven), to create pSG5-solECE-1. This cDNA encodes alkaline phosphatase (amino acids 1 to 32) and ECE-1a (amino acids 78 to 758). Stable expression of solECE-1 in CHO cells. The pSG5-solECE-1 was cotransfected with pZeoSV (Invitrogen) into CHO-K1 cells using lipofectamine. Zeocin-resistant (0.5 mg/ml) colonies were isolated by trypsinization in a cloning cup. Cell colonies expressing ECE-1 activity were selected and subcloned using the limiting dilution technique. Cell lines which express the highest ECE-1 activity were further selected by seeding the each stable cell line onto 12-well plates. At confluency, serum-free medium (750 ml per well) was added and incubated for 48 h. The culture supernatant was collected and used for ECE-1 assays. Cell culture. CHO-K1 cells were cultured in monolayers in Ham F-12 and Dulbecco’s modified Eagle’s medium (1:1 mixture) supplemented with 10% fetal bovine serum. CHO/ECE-1a cells were grown in the same medium as above including 1 mg/ml G418. CHO/ solECE-1 cells were cultured either in the same medium as CHO-K1 cells or serum-free CHO-SFM II including 0.5 mg/ml Zeocin. All cells were grown in a humidified incubator at 37°C in an atmosphere of 5% CO2/95% air. Antibody and Western blot. Antibody against ECE-1 was provided by Dr. M. Yanagisawa and was produced by immunizing rabbits with a synthetic peptide corresponding to the C-terminal 16 amino acids of bovine ECE-1 as previously described (26). The Western blot analysis was performed with horseradish peroxidase conjugated anti-rabbit IgG by using an ECL detection kit (Amersham). Other methods. SDS–PAGE and staining by Coomassie brilliant blue were performed as described (34). The protein concentration
was determined by the Bradford method (35) with bovine serum albumin as standard. The concentration of ECE-1a from the solubilized membrane fraction of CHO/ECE-1a cells was determined by Western blotting using purified solECE-1 as standard. The stock concentrations of big ET and ET peptides used in the experiments to determine isopeptide substrate selectivity and kinetic studies were determined by quantitative amino acid analysis.
RESULTS
Expression of solECE-1 in CHO Cells When stable transfectant CHO cell lines which express solECE-1 (CHO/solECE-1) were selected for their highest ECE-1 activity by using the method described under Experimental Procedures, the best six CHO/ solECE-1 cell lines expressed ECE-1 activity in a range from 0.49 to 1.11 pmol/min/ml medium (at pH 7.2). This range of ECE-1 activity corresponded to 2.7– 6.1 mg of solECE-1 per liter culture medium when estimated from that of solECE-1 purified to homogeneity (see below). Purification of solECE-1 to Homogeneity SolECE-1 was purified to homogeneity, as judged by SDS–PAGE visualized by Coomassie blue staining (Fig. 1). The purification scheme employed is summarized in Table I. Purification to homogeneity required only a 20-fold overall refinement representing a 24% yield of solECE-1 from the culture medium. The Nterminal amino acid sequence of the purified solECE-1 was I-I-P-V-E-E-E as expected from the cleavage between Gly22 and Ile23 of the human placental alkaline phosphatase signal sequence. Characterization of Purified solECE-1 SolECE-1 behaves as a dimer. An apparent molecular weight of 110 –120 kDa as estimated from SDS– PAGE gives evidence that solECE-1 is highly glycosylated (Fig. 1). The predicted molecular weight of solECE-1 based on amino acid content alone is 78.6 kDa. A matrix-assisted laser desorption ionization (MALDI) mass spectrum of solECE-1 showed (M 1 H)1, (M 1 2H)1, and (2M 1 H)1 ions for the protein consistent with a molecular weight of approximately 107.6 kDa. Broad peak widths of approximately 10 kDa indicated a large degree of heterogeneity in the glycosylation level of solECE-1. Under native conditions, solECE-1 behaves as a dimer as shown by gel filtration chromatography on Superose-6. Comparison of the Superose-6 elution volume of solECE-1 with those of reference proteins indicated an apparent molecular mass for solECE-1 of 232 kDa; the recovery of the activity was 40% (Fig. 2A). The Stokes radius of solECE-1 was estimated to be 52.3 Å (Fig. 2B).
261
SOLUBLE ENDOTHELIN-CONVERTING ENZYME-1
FIG. 1. SDS–PAGE analysis of solECE-1. The SDS–PAGE (4 – 20%) was carried out as described (34). Samples were precipitated with 15% trichloroacetic acid and resuspended in 15 ml of sampleloading buffer. Proteins were visualized by Coomassie blue staining: fraction II (Fr II), 13.9 mg; fraction III (Fr III), 4.5 mg; and fraction IV (Fr IV), 2.4 mg. Fractions are shown in Table I. Marker proteins (M) were myosin (200 kDa), b-galactosidase (116 kDa), phosphorylase b (97 kDa), bovine serum albumin (66 kDa), glutamic dehydrogenase (55 kDa), lactate dehydrogenase (37 kDa), carbonic anhydrase (31 kDa), trypsin inhibitor (22 kDa), and lysozyme (14 kDa).
The pH optimum for solECE-1 is slightly more acidic than that for ECE-1a. To accurately determine the pH optimum for each enzyme, buffers were prepared which varied by approximately 0.2 pH units extending from pH 5.4 to 8.0. As shown in Fig. 3, both ECE-1a and solECE-1 possessed very narrow pH optima. Different buffers (Mes-KOH, KPi, and Hepes-KOH) gave slightly different ECE-1 activities but the sharp pH optimum of similar ranges was observed. The width of the peak at half-maximum activity was approximately 1 pH unit. The pH optimum of 6.7– 6.9 for ECE-1a agreed well with the reported neutral pH optimum for the native enzyme. The optimum pH for solECE-1 was at 6.1– 6.4, slightly more acidic than that for ECE-1a. Substrate specificity. ECE-1a or solECE-1 was incubated with 1 mM big ET-1 (1–38), big ET-2 (1–38), or big ET-3 (1– 41, amide) for up to 4 h at pH values of 6.4 and 6.8. The amount of ET-1, -2, and -3 generated was linear with incubation time under the assay conditions and the results at the respective optimal pH values are shown in Fig. 4. These data show that both enzymes have a strong substrate preference for big ET-1. At its optimal pH 6.4, solECE-1 cleaved big ET-1:big ET-2: big ET-3 in a ratio of 8.1:1:1.4, similar to that of 7.6:1: 1.3 for ECE-1a at its optimal pH (6.8). Each enzyme showed a markedly different substrate selectivity at the pH optimum of the other. SolECE-1 at pH 6.8 and ECE-1a at pH 6.4 cleaved big ET-1:big ET-2:big ET-3 in ratios of 25.9:1:1.7 and 3.1:1:1.4, respectively. Effects of protease inhibitors. The purified solECE-1 was inhibited by EDTA, 1,10-phenanthroline, phos-
TABLE I
Purification of Human solECE-1
Fraction
Total volume (ml)
Total protein (mg)
Total activitya (pmol/min)
Sp act (pmol/min/mg)
Yield (%)
I. Medium II. DEAE III. Wheat germ agglutinin (WGA) IV. Alkyl
1,400 84 32.5 5.1
106.40 12.26 4.52 1.29
957.6 595.8 400.7 232.1
9.00 48.60 88.65 179.92
100 62.2 41.8 24.2
Note. A stable transfectant CHO cell line which expresses a secreted, soluble form of ECE-1 (CHO/solECE-1) was grown under conditions described under Materials and Methods (15 3 600 2cm). At confluency, the cells were washed with phosphate-buffered saline and incubated for 48 h in serum-free CHO-SFM II medium (Gibco/BRL). All subsequent operations were at 0 – 4°C unless otherwise noted. After removal of cell debris by centrifugation (5000g), the supernatant (fraction I, 1400 ml) was diluted twofold with buffer A and was applied to a 100-ml DEAE column (1.6 3 50 cm) equilibrated with buffer A. The column was washed with 1100 ml of the equilibration buffer, and the activity was eluted with a 650-ml gradient (0–250 mM NaCl in buffer A). Fractions of 12 ml were collected and peak fractions were pooled (fraction II, 84 ml). For WGA, 30% of fraction II was loaded onto a 17-ml WGA column (1.6 3 8.5) equilibrated with buffer A containing 150 mM NaCl and the column was washed with 350 ml of the equilibration buffer. The activity was eluted with a 64-ml gradient (0 –50 mg/ml N-acetylglucosamine in buffer A). Two more runs were repeated with 70% of fraction II and peak fractions were pooled from three WGA chromatographic steps (fraction III, 32.5 ml). Fraction III was adjusted to 2.4 M ammonium sulfate. After centrifugation (16,000g, 15 min), 40% of the supernatant was loaded onto a 1-ml Alkyl Superose column (HR5/5) equilibrated with buffer B containing 2 M ammonium sulfate. The column was washed with 20 ml of the equilibration buffer. The activity was eluted with a 15-ml gradient (2-0 M ammonium sulfate in buffer B). Two more runs were repeated with 60% of fraction III and peak fractions from three Alkyl Superose columns were pooled (fraction IV, 5.1 ml). Buffers A and B are Tris–HCl (pH 7.4) at 20 and 50 mM, respectively. a ECE-1 activity was determined at pH 7.2 (Hepes-KOH) as described under Experimental Procedures and therefore is much lower than that at its optimal pH of 6.4 (see Fig. 3).
262
AHN, HERMAN, AND FAHNOE
FIG. 2. Gel filtration (A) and determination of the Stokes radius (B) of solECE-1. 3.0 mg of solECE-1 was applied to a Superose-6 fast protein liquid chromatography column equilibrated with 50 mM Tris–HCl (pH 7.4)/150 mM NaCl. Elution was performed with the equilibration buffer at a flow rate of 0.3 ml/min; 0.38-ml fractions were collected. ECE-1 activity was determined at pH 7.0 as described under Experimental Procedures. Marker proteins were detected by monitoring UV absorbances at 280 nm. Abbreviations are ferritin (ferr, 440 kDa, 61.0 Å); catalase (cat, 232 kDa, 52.2 Å); aldolase (aldo, 158 kDa, 48.1 Å); bovine serum albumin (BSA, 67 kDa, 35.5 Å); ovalbumin (oval, 43 kDa, 30.5 Å); ribonuclease A (RNase A, 13.7 kDa, 16.4 Å).
phoramidon, and PD 069185, a novel highly specific ECE-1 inhibitor (36), but was insensitive to thiorphan and captopril, similar to the native enzyme, ECE-1a (Table II). Phosphoramidon inhibition of ECE-1 is highly pHdependent. Our initial data at neutral pH indicated that phosphoramidon inhibited ECE-1a almost 10-fold better than solECE-1. These results prompted us to further investigate the pH dependence of phosphoramidon inhibition for both ECE-1a and solECE-1. As shown in Fig. 5, phosphoramidon inhibited ECE-1a with IC50 values of 0.014 6 0.002, 0.114 6 0.014, 0.249 6 0.031, and 0.675 6 0.061 mM at pH values of 5.8, 6.4, 6.8, and 7.2, respectively. Corresponding IC50 values with solECE-1 were 0.033 6 0.004, 0.349 6 0.045, 2.12 6 0.43, and 40.4 6 3.7 mM at pH values of 5.8, 6.4, 6.8, and 7.2, respectively. In both cases, the potency of phosphoramidon dramatically decreased from pH 5.8 to 7.2 and this dependence of inhibition upon pH was far more extensive with solECE-1 than with ECE-1a, decreasing 49- or 1225-fold for ECE-1a or solECE-1, respectively, from pH 5.8 to 7.2. At pH 5.8, there was a 2.4-fold potency difference between ECE-1a and solECE-1. This potency difference increased rapidly with increasing pH such that at pH 7.2, there was a 59-fold difference. Nevertheless, when compared at their respective optimal pH values, phos-
phoramidon potency was very similar; an IC50 value of 0.35 mM for solECE-1 at its optimal pH (6.4) was similar to that of 0.25 mM for ECE-1a at its optimal pH (6.8). These potency shifts are not due to changes in substrate Km values because the IC50 values for phosphoramidon, a competitive inhibitor (25), are very close to Ki values under the assay conditions where the substrate concentration is well below the Km values (see below) according to the equation, Ki 5 IC50/(1 1 [S]/Km) (37). In addition, the potency difference between ECE-1a and solECE-1 is not due to the presence/ absence of detergent since the enzyme reactions for both solECE-1 and ECE-1a were carried out in the presence of 0.1% C12E10. Furthermore, the concentration variation of C12E10 from 0.016 to 0.5% did not have any effect on the potency of phosphoramidon for either ECE-1a or solECE-1. Kinetic parameters. The high pH dependence of the phosphoramidon potency as well as that of the hydrolysis ratios for big ET substrates prompted us to measure Km and kcat at several pH values. The velocity versus big ET-1 concentration curves at pH 6.4, 6.8, and 7.1 were fit to Michaelis–Menten kinetics as shown in Fig. 6. The Km and kcat values thus obtained were also pH-dependent as listed in Table III. Under the pH conditions tested, the tightest observed substrate binding was obtained at the respective optimal pH values
SOLUBLE ENDOTHELIN-CONVERTING ENZYME-1
263
FIG. 3. pH dependence of ECE-1 activity. ECE-1a or solECE-1, 10 ng of each, was incubated at 37°C for 1 h in a reaction mixture (100 ml) containing 0.1 mM big ET-1, 0.1% C12E10, 50 mM pepstatin A, 100 mM leupeptin, 200 mM phenylmethlysulfonyl fluoride, and a buffer at 100 mM as follows: pH 5.31–7.16 (Mes-KOH), pH 5.92–7.74 (KPi), and pH 6.47–7.96 (Hepes-KOH). The pH values of all the buffers were measured at 100 mM and at 37°C. The amount of ET-1 generated was analyzed by ELISA.
for ECE-1a and solECE-1. At its optimal pH (6.4), the Km and kcat values of 4.65 mM and 5.82 min21 for solECE-1 were similar to those of 2.70 mM and 6.72 min21 for ECE-1a at its optimal pH 6.8. At the nonoptimal pH of 7.1, Km values were 10.4- and 7.4-fold higher for solECE-1 and ECE-1a, respectively, than those at the respective optimal pH values. N-glycosylation is not essential for enzyme activity. In order to remove glycosylation from solECE-1 as completely as possible, glycosidases were incubated at 37°C with fully denatured solECE-1, which was prepared by boiling it in the presence of SDS. Peptide N-glycosidase F (PNGase F), endo-b-N-acetylglucosaminidase H, and Oglycopeptide endo-D-galactosyl-N-acetyl-a-galactosamino hydrolase were incubated with the denatured solECE-1 for 12 h at 37°C. Western blot analysis showed that the 110 kDa band for solECE-1 was completely shifted to the 80-kDa band(s) after incubation with PNGase F which cleaves all Asp-bound N-glycans. Endo-b-N-acetylglucosamidase H and O-glycopeptide endo-D-galactosylN-acetyl-a-galactosamino hydrolase, which remove Asplinked high mannose oligosaccharides and O-glycosylation, respectively, did not shift the molecular weight of solECE-1 under similar conditions (data not shown). Next, in order to find out if N-glycosylation is required for ECE-1 activity, solECE-1 was incubated under the native conditions with PNGase F at 37°C. The commercial PNGase F (Boehringer-Mannheim) is supplied in a buffer
containing 25 mM EDTA and thus was dialyzed extensively against 100 mM KPi (pH 7.5)/10% glycerol to remove EDTA before use. Both the extent of deglycosylation by Western blot and ECE-1 activity were tested at the indicated time. As shown in Fig. 7A, after 2 h incubation with PNGase F, the 110 kDa band for solECE-1 was completely shifted to the 80 kDa band(s), similar to those observed with fully denatured solECE-1. At 30 min, 1 h, and 2 h, the solECE-1 activity was 91.8, 78.4, and 72.0% of the control activity (without PNGase F), respectively (Fig. 7B). These data show that a substantial fraction of ECE-1 activity still remained after the removal of N-glycosylation. There was no self-digestion of solECE-1 under the deglycosylation conditions as shown in the control (Fig. 7A). DISCUSSION
Detailed data on enzymatic characterization of ECE-1 has been limited mostly due to the difficulty in obtaining homogeneous enzyme in sufficient quantities. The reported literature data on big ET isopeptide specificity, inhibition potency of phosphoramidon, and kinetic constants vary widely presumably because most of the early data were obtained using partially purified enzyme which may have been contaminated with other proteases. However, even the data obtained from ECE-1 which was either purified to homogeneity or the
264
AHN, HERMAN, AND FAHNOE
FIG. 4. Substrate specificity. ECE-1a or solECE-1, 30 ng of each, was incubated at 37°C for the indicated period of time in a reaction mixture (100 ml) containing 1 mM big ET-1, -2, or -3, 0.1% C12E10, 50 mM pepstatin A, 100 mM leupeptin, 200 mM phenylmethlysulfonyl fluoride, and 100 mM Hepes-KOH (pH 6.8) for ECE-1a or 100 mM Mes-KOH (pH 6.4) for solECE-1. The same stock solutions of big ET and assay components were used for both ECE-1a and solECE-1 for a better comparison. The amount of ET-1, -2, and -3 was analyzed by EIA.
recombinant enzyme expressed in mammalian cells vary over a wide range under similar assay conditions. For example, there is a large variation in the reported ability of ECE-1 to process big ET isopeptides. With purified rat lung ECE-1, the relative conversion rates for big ET-1:big ET-2:big ET-3 were reported to be 4:2:1 (15). For ECE-1 purified from porcine aortic endothelium, big ET-1 was a much preferred substrate over big
TABLE II
Effects of Protease Inhibitors at 100 mM ECE-1 activity (% of control) Inhibitor (100 mM)
ECE-1a
solECE-1
Phosphoramidon EDTAa PD 069185 1,10-Phenanthroline Thiorphan Captopril
2.9 29.8 4.3 0.7 52.0 95.8
3.2 14.4 5.7 1.6 70.0 106.9
Note. ECE-1a or solECE-1, 10 ng of each, was incubated in a reaction mixture (100 ml) containing 0.1 mM big ET-1, 0.1% C12E10, 50 mM pepstatin A, 100 mM leupeptin, 200 mM phenylmethlysulfonyl fluoride, and 100 mM Hepes-KOH (pH 6.8) for ECE-1a or 100 mM Mes-KOH (pH 6.4) for solECE-1. The final dimethyl sulfoxide concentration was 1.5%. a At 1 mM.
ET-2 (12.5:1). However, it did not convert big ET-3 (16). Bovine ECE-1 expressed in CHO cells processed all three big ET isopeptides, with the conversion rates big ET-1 . big ET-3 . big ET-2 (18). These discrepancies in the reported relative values of big ET conversion might possibly reflect variations in the specificity of several isotypes of ECE-1 for substrate. However, it has been shown that there is no substrate specificity difference at least between ECE-1a and ECE-1b of either human or rat (23). Likewise, the reported Km values for big ET-1 range widely from 0.2 mM (ECE-1 purified from rat lung) to 23 mM (recombinant human ECE-1 expressed in CHO cells) (15, 21). Vmax values of 3.1 and 410 nmol/min/mg protein were reported for the purified ECE-1 from rat lung and porcine aortic endothelium, respectively (15, 16). The reported IC50 values for phosphoramidon with ECE-1 also vary from 0.35 to 8 mM (15, 21). It is unlikely that the wide variations in the literature data discussed above are due to the species difference since similar big ET isopeptide specificity was observed with the human and rat enzymes (23). Likewise, similar inhibitor potencies and kinetic parameters were obtained with the human and bovine enzymes (K. Ahn, S. Herman, and D. Fahnoe, unpublished results). In the current study, we have constructed a mammalian expression system by fusing a signal peptide in frame with the extracellular domain of ECE-1. This con-
SOLUBLE ENDOTHELIN-CONVERTING ENZYME-1
265
FIG. 5. pH-dependence of phosphoramidon inhibition. ECE-1a or solECE-1, 10 ng of each, was incubated at 37°C for 1 h in a reaction mixture (100 ml) containing 0.1 mM big ET-1, 0.1% C12E10, 50 mM pepstatin A, 100 mM leupeptin, 200 mM phenylmethlysulfonyl fluoride, and a buffer at 100 mM as follows: pH 5.8 and 6.4, Mes-KOH; pH 6.8 and 7.2, Hepes-KOH. The same stock solutions of compounds and assay components were used for both ECE-1a and solECE-1 for a better comparison. Under the assay conditions, the production of ET-1 is linear with an incubation time of several hours. The data were plotted as percentage of inhibition vs inhibitor concentration and fit with the equation, y 5 100/1 1 (x/IC50)z, using KaleidaGraph (Synergy Software, Reading, PA), where IC50 is the inhibitor concentration at 50% inhibition and z is the slope of the inhibition curve.
struct allowed the secretion of a fully active enzyme after transfection into CHO cells and a relatively easy homogeneous purification of solECE-1 in milligram quantities from the culture supernatant. All the studies of solECE-1 were carried out in parallel with ECE-1a for comparison. Accurate concentrations of big ET and ET stocks were determined by quantitative amino acid analysis for the experiments described in this paper. In our laboratory,
we have observed discrepancies between the stated weight indicated by the commercial sources of theses peptides, usually supplied as preweighed lyophilized powders, and the actual weight determined in solution. Furthermore, the peptides may not fully solubilize in the chosen buffer. Inaccuracy in respect to peptide concentrations may have also contributed to the variations in the literature data.
FIG. 6. Km and kcat measurements. ECE-1a (4.7 ng) or solECE-1 (9.6 ng) was incubated at 37°C for 30 min in a reaction mixture (50 ml) containing 0.1% C12E10, 50 mM pepstatin A, 100 mM leupeptin, 200 mM phenylmethlysulfonyl fluoride, big ET-1 (varied between 0.5 and 140 mM), and a buffer at 100 mM as follows: pH 6.4 and 6.8, Mes-KOH; pH 7.1, Hepes-KOH. The same stock solutions of big ET-1 and assay components were used for both ECE-1a and solECE-1 for a better comparison. The data were fit to the equation, v 5 Vmax [S]/([S] 1 Km), by a nonlinear least-squares algorithm.
266
AHN, HERMAN, AND FAHNOE TABLE III
Kinetic Parameters of ECE-1a and solECE-1 at Various pH Values ECE-1a
solECE-1
pH
Km (mM)
kcat (min21)
Km (mM)
kcat (min21)
6.4 6.8 7.1
7.51 6 1.53 2.70 6 0.36 19.87 6 2.72
3.17 6 0.16 6.72 6 0.16 17.61 6 0.82
4.65 6 0.78 8.20 6 0.95 48.31 6 4.10
5.82 6 0.21 17.77 6 0.54 21.23 6 0.82
Note. The Km and kcat values were measured as described in the legend to Fig. 6. The molecular weights of 85.6 and 78.6 kDa were used for ECE-1a and solECE-1, respectively.
It was surprising to observe a more acidic pH optimum for solECE-1 compared to that for ECE-1a. At the optimal pH of 6.8 for ECE-1a, solECE-1 has only 11– 33% of its optimal activity at pH 6.4 (Fig. 3). A search of the literature for other enzymes whose truncated forms show a shift in pH optimum revealed one example, protein tyrosine phosphatase containing two Src homology 2 domains. The truncated protein tyrosine phosphatase domain was shown to be most active at pH 6.1– 6.5, which is shifted from an optimal pH of 5.5.–5.7 for the full-length enzyme (38). The lack of more examples may derive from the fact that few enzymes have similarly sharp pH optima, thus a shift might not be reflected as a clear difference in activity. The shift in pH optimum prompted us to evaluate hydrolysis rates, potency of phosphoramidon inhibi-
tion, and kinetic parameters at various pH values. At its optimal pH of 6.4, solECE-1 had a hydrolysis ratio for big ET isopeptides, an inhibition potency of phosphoramidon, a Km value for big ET-1, and a kcat value all comparable to those for the native membranebound form, ECE-1a, at its optimal pH of 6.8. Additionally, these data were highly pH dependent and the results obtained at a nonoptimal pH of only 0.4 pH unit away were very different from those at their respective optimal pH values for both solECE-1 and ECE-1a. This high degree of pH dependence could explain further the wide variations in the literature data as discussed above. While this manuscript was in preparation, a report on a similar expression of solECE-1 appeared (39). The authors did not report a pH optimum shift for solECE-1 compared to that for the native enzyme. For their study of solECE-1, the culture supernatant was used as a source of solECE-1 rather than a homogeneous enzyme (39). The pH dependence of phosphoramidon inhibition potency was large, and it was especially dramatic for solECE-1 (Fig. 5). This pH dependence is not due to changes in the ionization state of phosphoramidon. Since the pKa values of the ionizing groups in phosphoramidon, carboxyl and phosphonyl groups, are less than 3 and 3.6, respectively (40), there is very little change in the ionization state of phosphoramidon in the pH range used in these experiments. Interestingly, phosphoramidon was also reported to be a more potent inhibitor of NEP at pH 6.5 than at pH 8.5 by 150-fold
FIG. 7. Deglycosylation of solECE-1. The reaction mixture (25 ml) contained 100 mM Tris–HCl (pH 8.5), 0.1% C12E10, 50 mM pepstatin A, 100 mM leupeptin, 200 mM phenylmethlysulfonyl fluoride, and 150 ng solECE-1 either in the presence or in the absence of PNGase F (0.2 units). After incubation at 37°C for the indicated amount of time, the reaction was stopped by freezing in liquid nitrogen and was stored until the entire set for the time course was collected; 5 ml of the reaction mixture was loaded onto SDS–PAGE (4 –20%) for a Western blot analysis (A). 10 ml of the reaction mixture was used to determine the ECE-1 activity at pH 7.2 (Hepes-KOH) as described under Experimental Procedures (B).
SOLUBLE ENDOTHELIN-CONVERTING ENZYME-1
(41). The nM potency of phosphoramidon reported for ECE-2 as a distinctive difference compared to the mM potency for ECE-1 (26) now appears to be simply a reflection of the different optimal pH values for ECE-1 and ECE-2. An IC50 value of 4 nM with phosphoramidon reported for ECE-2 that was obtained at its optimal pH of 5.6 (26) is also in the same range as the IC50 value of 14 nM for ECE-1a measured at pH 5.8 (Fig. 5). In summary, ECE-1, in both the native and soluble forms, is a highly pH-sensitive enzyme, and it is necessary to be very cautious when acquiring and comparing data. Our results suggest that solECE-1 undergoes a general shift in the apparent pKa values for active site residues yielding a shifted pH rate profile compared to that for ECE-1a although the profile retains similar width. The comparable Km, kcat, and Ki values between ECE-1a and solECE-1 at the respective optimal pH values suggest that the ionized residues at the active site of solECE-1 bind to substrate and inhibitor in the same manner as with the native form. Clearly, further work is warranted and is currently underway in our laboratory. Retention of ECE-1 activity after full deglycosylation is very encouraging with respect to the prospects for crystallization of solECE-1. Structural information on ECE-1 obtainable from X-ray crystallography will be invaluable for a better understanding of the enzyme and in designing potent inhibitors. It has been known that glycosylation is not required for activity in many glycosylated enzymes (42, 43), and successful crystallization of glycosylated enzymes has been possible after deglycosylation (44). With homogeneous enzyme available, many other biochemical characterizations of ECE-1 including the testing of other hormones as possible substrates for ECE-1 and defining preferred residues around the cleavage site are now possible.
5. 6. 7. 8.
9.
10.
11.
12.
13.
14. 15. 16. 17. 18. 19. 20. 21.
ACKNOWLEDGMENTS We thank Masashi Yanagisawa and Noriaki Emoto for providing cDNA for ECE-1a and antibody against ECE-1 and for discussions. We thank Tracy Stevenson and Kenneth Greis for obtaining the MALDI mass spectrum and the N-terminal sequencing of purified solECE-1, respectively. We also thank Jill Knapp and Andre Sisneros for assisting with the assays.
REFERENCES 1. Yanagisawa, M., Kurihara, H., Kimura, S., Tomobe, Y., Kobayashi, M., Mitsui, Y., Yazaki, Y., Goto, K., and Masaki, T. (1988) Nature 332, 411– 415. 2. Inoue, A., Yanagisawa, M., Kimura, S., Kasuya, Y., Miyauchi, T., Goto, K., and Masaki, T. (1989) Proc. Natl. Acad. Sci. USA 86, 2863–2867. 3. Seidah, N. G., Day, R., Marcinkiewicz, M., and Chre´tien, M. (1993) Ann. N. Y. Acad. Sci. 680, 135–146. 4. Yanagisawa, M., Inoue, A., Ishikawa, T., Kasuya, Y., Kimura, S., Kumagaye, S.-I, Nakajima, K., Watanabe, T. X., Sakakibara, S.,
22.
23. 24. 25.
26. 27. 28. 29.
267
Goto, K., and Masaki, T. (1988) Proc. Natl. Acad. Sci. USA 85, 6964 – 6967. Cheng, X.-M., Ahn, K., and Haleen, S. J. (1997) Annu. Rep. Med. Chem. 32, 61–70. Webb, M. L., and Meek, T. D. (1997) Med. Res. Rev. 17, 17– 67. Ohlstein, E. H., Elliott, J. D., Feuerstein, G. Z., and Ruffolo, R. R., Jr. (1996) Med. Res. Rev. 16, 365–390. Patel, T. R., Galbraith, S. L., McAuley, M. A., Doherty, A. M., Graham, D. I., and McCulloch, J. (1995) J. Cardiovasc. Pharmacol. 26(Suppl. 3), S412–S415. Kiowski, W., Su¨tsch, G., Hunziker, P., Mu¨ller, P., Kim, J., Oechslin, E., Schmitt, R., Jones, R., and Bertel, O. (1995) Lancet 346, 732–736. Clozel, M., Breu, V., Burri, K., Cassal, J.-M., Fischli, W., Gray, G. A., Hirth, G., Lo¨ffler, B.-M., Mu¨ller, M., Neidhart, W., and Ramuz, H. (1993) Nature 365, 759 –761. Kurihara, Y., Kurihara, H., Suzuki, H., Kodama, T., Maemura, K., Nagai, R., Oda, H., Kuwaki, T., Cao, W.-H., Kamada, N., Jishage, K., Ouchi, Y., Azuma, S., Toyoda, Y., Ishikawa, T., Kumuda, M., and Yazaki, Y. (1994) Nature 368, 703–710. Baynash, A. G., Hosoda, K., Giaid, A., Richardson, J. A., Emoto, N., Hammer, R. E., and Yanagisawa, M. (1994) Cell 79, 1277– 1285. Puffenberger, E. G., Hosoda, K., Washington, S. S., Nakao, K., deWit, D., Yanagisawa, M., and Chakravarti, A. (1994) Cell 79, 1257–1266. Ahn, K., Beningo, K., Olds, G., and Hupe, D. (1992) Proc. Natl. Acad. Sci. USA 89, 8606 – 8610. Takahashi, M., Matsushita, Y., Iijima, Y., and Tanzawa, K. (1993) J. Biol. Chem. 268, 21394 –21398. Ohnaka, K., Takayanagi, R., Nishikawa, M., Haji, M., and Nawata, H. (1993) J. Biol. Chem. 268, 26759 –26766. Shimada, K., Takahashi, M., and Tanzawa, K. (1994) J. Biol. Chem. 269, 18275–18278. Xu, D., Emoto, N., Giaid, A., Slaughter, C., Kaw, S., deWit, D., and Yanagisawa, M. (1994) Cell 78, 473– 485. Ahn, K., Pan, S., Beningo, K., and Hupe, D. (1995) Life Sci. 56, 2331–2341. Ahn, K., Pan, S., Zientek, M., Guy, P., and Sisneros, A. (1996) Biochem. Mol. Biol. Int. 39, 573–580. Schmidt, M., Kro¨ger, B., Jacob, E., Seulberger, H., Subkowski, T., Otter, R., Meyer, T., Schmalzing, G., and Hillen, H. (1994) FEBS Lett. 356, 238 –243. Shimada, K., Matsushita, Y., Wakabayashi, K., Takahashi, M., Matsubara, A., Iijima, Y., and Tanzawa, K. (1995) Biochem. Biophys. Res. Commun. 207, 807– 812. Shimada, K., Takahashi, M., Ikeda, M., and Tanzawa, K. (1995) FEBS Lett. 371, 140 –144. Valdenaire, O., Rohrbacher, E., and Mattei, M.-G. (1995) J. Biol. Chem. 270, 29794 –29798. Schweizer, A., Valdenaire, O., Nelbo¨ck, P., Deuschle, U., Dumas Milne Edwards, J.-B., Stumpf, J. G., and Lo¨ffler, B.-M. (1997) Biochem. J. 328, 871– 877. Emoto, N., and Yanagisawa, M. (1995) J. Biol. Chem. 270, 15262–15268. Opgenorth, T. J., Wu-Wong, J. R., and Shiosaki, K. (1992) FASEB J. 6, 2653–2659. Turner, A. J., and Murphy, L. J. (1996) Biochem. Pharmacol. 51, 91–102. Ahn, K. (1998) Handbook of Proteolytic Enzymes (Barrett, A. J., Rawlings, N. D., and Woessner, J. F., Eds.), pp. 1085–1091, Academic Press, San Diego.
268
AHN, HERMAN, AND FAHNOE
30. Shimada, K., Takahashi, M., Turner, A. J., and Tanzawa, K. (1996) Biochem. J. 315, 863– 867. 31. Lemay, G., Waksman, G., Roques, B. P., Crine, P., and Boileau, G. (1989) J. Biol. Chem. 264, 15620 –15623. 32. Fossiez, F., Lemay, G., Labonte´, N., Parmentier-Lesage, F., Boileau, G., and Crine, P. (1992) Biochem. J. 284, 53–59. 33. Kozak, M. (1987) Nucleic Acids Res. 15, 8125– 8148. 34. Laemmli, U. K. (1970) Nature 227, 680 – 685. 35. Bradford, M. M. (1976) Anal. Biochem. 72, 248 –254. 36. Ahn, K., Sisneros, A. M., Herman, S. B., Pan, S. M., Hupe, D., Lee, C., Nikam S., Cheng, X.-M., Doherty, A. M., Schroeder, R. L., Haleen, S. J., Kaw, S., Emoto, N., and Yanagisawa, M. (1998) Biochem. Biophys. Res. Commun. 243, 184 –190. 37. Cheng, Y.-C., and Prusoff, W. H. (1973) Biochem. Pharmacol. 22, 3099 –3108.
38. Pei, D., Neel, B. G., and Walsh, C. T. (1993) Proc. Natl. Acad. Sci. USA 90, 1092–1096. 39. Korth, P., Egidy, G., Parnot, C., LeMoullec, J.-M., Corvol, P., and Pinet, F. (1997) FEBS Lett. 417, 365–370. 40. Umezawa, S., Tatsuta, K., Izawa, O., and Tsuchiya, T. (1972) Tetrahedron Lett. 1, 97–100. 41. Jeng, A., Ansell, J., and Erion, M. D. (1989) Life Sci. 45, 2109 – 2114. 42. Malissard, M., Borsig, L., Di Marco, S., Gru¨tter, M. G., Kragl, U., Wandrey, C., and Berger, E. G. (1996) Eur. J. Biochem. 239, 340 –348. 43. Yu, X. C., Sturrock, E. D., Wu, Z., Biemann, K., Ehlers, M. R. W., and Riordan, J. F. (1997) J. Biol. Chem. 272, 3511–3519. 44. Kalisz, H. M., Hecht, H.-J., Schomburg, D., and Schmid, R. D. (1990) J. Mol. Biol. 213, 207–209.