A Human Homolog of Angiotensin-converting Enzyme

A novel human zinc metalloprotease that has considerable homology to human angiotensin-converting enzyme (ACE) (40% identity and 61% similarity) has been identified. This metalloprotease (angiotensin-converting enzyme homolog (ACEH)) contains a single HEXXH zinc-binding domain and conserves other critical residues typical of the ACE family. The predicted protein sequence consists of 805 amino acids, including a potential 17-amino acid N-terminal signal peptide sequence and a putative C-terminal membrane anchor. Expression in Chinese hamster ovary cells of a soluble, truncated form of ACEH, lacking the transmembrane and cytosolic domains, produces a glycoprotein of 120 kDa, which is able to cleave angiotensin I and angiotensin II but not bradykinin or Hip-His-Leu. In the hydrolysis of the angiotensins, ACEH functions exclusively as a carboxypeptidase. ACEH activity is inhibited by EDTA but not by classical ACE inhibitors such as captopril, lisinopril, or enalaprilat. Identification of the genomic sequence of ACEH has shown that the ACEH gene contains 18 exons, of which several have considerable size similarity with the first 17 exons of human ACE. The gene maps to chromosomal location Xp22. Northern blotting analysis has shown that the ACEH mRNA transcript is ∼3.4 kilobase pairs and is most highly expressed in testis, kidney, and heart. This is the first report of a mammalian homolog of ACE and has implications for our understanding of cardiovascular and renal function.

ACE is in cardiovascular homeostasis through cleavage of the C-terminal dipeptide from angiotensin I to produce the potent vasoconstrictor, angiotensin II (2). ACE also inactivates the vasodilator, bradykinin, by the sequential cleavage of two Cterminal dipeptides (3). ACE can also hydrolyze a wide range of other endogenous bioactive peptides (4).
Two forms of mammalian ACE have been identified to date: the two-domain somatic ACE, containing two catalytic sites and a single domain germinal ACE (5)(6)(7). Both enzymes are derived from the same gene through the use of alternative promoters, and it has been suggested that the ACE gene arose from the duplication of an ancestral gene coding for a single domain enzyme (8). Somatic ACE exists as a type I integral membrane protein anchored to the plasma membrane through a transmembrane domain near the C terminus (9). However, it can also be found in plasma and other body fluids as a soluble enzyme lacking the transmembrane and cytosolic domains (10 -12). This form is thought to arise predominantly by posttranslational proteolytic cleavage at the cell surface through the action of ACE secretase (9,13,14).
ACE-like enzymes have also been found in other non-mammalian species. In particular Musca domestica and Drosophila melanogaster have both been shown to contain single domain, ACE-like proteins (AnCE) (15,16), and more recently, a second ACE-like protein, termed ACEr, has been identified in D. melanogaster (17). ACEr and AnCE appear to be alternatively expressed during D. melanogaster pupal development (18), suggesting different roles for the two enzymes.
In this study we have identified a novel, single domain, human zinc metalloprotease cDNA (ACEH) whose predicted amino acid sequence has significant similarity with mammalian ACE. Furthermore, the genomic structure of ACEH indicates a remarkable exon size similarity with the first 17 exons of ACE. Expression of a soluble, truncated form of ACEH, lacking the transmembrane and cytosolic domains, produced a 120-kDa glycosylated protein that hydrolyzed the C-terminal residue from angiotensins I and II. The transcript for this cDNA is highly expressed in heart, kidney, and testis, implying that the translated protein may play a role in the regulation of cardiovascular and renal function, as well as fertility.

EXPERIMENTAL PROCEDURES
Materials-Captopril was a gift from Bristol-Myers Squibb Co. Enalaprilat (MK422) and lisinopril were gifts from Merck. Other chemicals were obtained from Sigma.
cDNA Cloning-A partial cDNA of 2885 nucleotides encoding ACEH was obtained following identification of a zinc metalloprotease with homology to ACE from a proprietary EST data base and subsequent screening of a human lymphoma cDNA library. To establish whether the clone was a full-length cDNA, 5Ј-and 3Ј-RACE were carried out on a SUPERSCRIPT TM human kidney cDNA library (Life Technologies, Inc.). Gene-specific primers deduced from the partial cDNA library * This work was supported by British Heart Foundation Grant PG/ 97192 and by UK Medical Research Council Grant G9824728. 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 U.S.C. Section 1734 solely to indicate this fact.
Computer-aided Sequence Identification and Manipulation-The genomic sequence of ACEH was identified by searching the GenBank TM data base with the cDNA sequence encoding ACEH using a FASTA program on the GCG suite. The GCG suite was also used to determine the percentage identities and similarities between the predicted ACEH protein and existing members of the ACE family using the Bestfit program. Hydropathy plots, obtained using the algorithms of Kyte and Doolittle (19) and Engelman et al. (20), were also performed using the GCG suite.
Northern Blotting Analysis-Multiple tissue mRNA blots were obtained from CLONTECH (Palo Alto, CA). These blots were probed according to the manufacturer's protocol with cDNA fragments encoding ACEH or endothelial ACE (provided by P. Corvol, Paris, France). A 943-nucleotide cDNA fragment of ACEH was obtained by XmnI restriction endonuclease digestion. A 614-nucleotide cDNA fragment of ACE was obtained using a ScaI digest. The cDNA fragments were labeled with 32 P using a random primed DNA labeling kit (Roche Molecular Biochemicals).
Construction of Expression Plasmid pSTMyc-TM7-The truncated cDNA encoding ACEH was isolated from the human lymphoma cDNA library in plasmid pCMVSport2 which contains cDNA inserts between NotI and SalI restriction sites. By using these sites the partial ACEH cDNA was excised from pCMVSport2 and cloned into pBluescript SK(ϩ). This plasmid was designated pBSKACE10. Additional 5Ј sequence obtained from 5Ј-RACE was added to pBSKACE10, using SalI and HindIII to remove the extra sequence from the RACE product in PCRScript and cloning it into the pBSKACE10 construct using the same sites, generating the plasmid pBSKACE10ϩ5Ј. The cDNA with the additional 5Ј sequence was then excised from pBSKACE10ϩ5Ј using NotI and SalI restriction sites and cloned into vector pC1-neo (Promega, Southampton, UK) using the same sites. PCR was then carried out using antisense primer SMYCPR (5ЈCGAGGGCCCGGAAA-CAGGGGGCTGGTTAGGAGG3Ј) (nucleotides 2300 -2323 which also incorporated an ApaI site) and sense primer NeoT7 (5ЈGGCTAGAG-TACTTAATACGACTCACTATAGG-3Ј) (nucleotides 1055-1085 of pC1neo). This gave a truncated cDNA encoding ACEH, which lacks the transmembrane and cytosolic domains. The PCR product was digested with ApaI and XhoI and ligated, in frame, into expression vector pcDNA3.1Myc-His A which was also digested with XhoI and ApaI. This construct, designated pSTMyc-TM7, gave a truncated ACEH cDNA together with an in frame fusion tag encoding the c-Myc epitope and a hexahistidine tag.
Expression of ACEH in CHO Cells-CHO cells were obtained from the European Collection of Cell Cultures (Salisbury, Wiltshire, UK) and were cultured in Ham's F-12 nutrient mix (Life Technologies, Inc.) supplemented with 2 mM L-glutamine, 10% (v/v) fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin, at 37°C, with 5% CO 2 . 24 h prior to transfection cells were seeded at a density of 1 ϫ 10 6 cells per 75-cm 2 flask. For transient transfection, the monolayer was washed twice with OPTI-MEM (Life Technologies, Inc.) before transfection with 3 g of pSTMyc-TM7 plasmid DNA per flask. LipofectAMINE (Life Technologies, Inc.) was used as cationic lipid at a ratio of DNA/lipid, 1:10 (w/w). This was added to the flasks in 2.5 ml of Opti-MEM and incubated for 16 h before the addition of Ham's F-12 nutrient mix containing 10% (v/v) fetal bovine serum. The medium was removed 24 h after the start of transfection; the monolayer rinsed twice with Opti-MEM, and then 5 ml of Opti-MEM was added to each flask. This was incubated for a further 16 h before harvesting of the medium, containing soluble secreted ACEH protein. The media samples containing protein were concentrated using 4-ml Vivaspin columns (Vivascience, Binbrook, Lincoln, UK).
Deglycosylation of ACEH-30 g of total protein containing soluble secreted ACEH from CHO cell media was incubated overnight at 37°C with 1 l of PNGase F (Oxford Glycosystems, Abingdon, Oxford, UK).
Protein and Enzymic Assays-Protein concentrations were determined using the bicinchoninic acid assay (21) with bovine serum albumin as standard. Assays for ACEH activity were carried out in a total volume of 100 l, containing 100 mM Tris-HCl, pH 7.4, 20 g of protein and either 100 M angiotensin I or II or 500 M bradykinin, Hip-Phe, or Hip-His-Leu as substrates. Where appropriate, inhibitors were added to give final concentrations of 10 M lisinopril, 10 M captopril, 10 M enalaprilat, 100 M benzyl succinate, or 10 mM EDTA. Reactions were carried out at 37°C, for 2 h and stopped by heating to 100°C for 5 min followed by centrifugation at 11,600 ϫ g for 10 min. Carboxypeptidase A assays were carried out at room temperature for 30 min, using 0.1 units of enzyme per assay.
HPLC Analysis of Cleavage Products-Peptide hydrolysis products were separated using reverse-phase HPLC (Bondapak C-18 reverse phase column, Waters) with an UV detector set at 214 nm. All separations were carried out at room temperature, with a flow rate of 1.5 ml/min. Mobile phase A consisted of 0.08% (v/v) phosphoric acid and mobile phase B consisted of 40% (v/v) acetonitrile in 0.08% (v/v) phosphoric acid. A linear solvent gradient of 11% B to 100% B over 15 min with 5 min at final conditions, and 8 min re-equilibration was used. The product from angiotensin I cleavage was collected and analyzed by matrix-assisted laser desorption ionization/time of flight mass spectrometry.
SDS-PAGE and Western Blotting-Proteins were separated by SDS-PAGE, using a 10% resolving gel and a 5% stacking gel according to the method of Laemmli (22). Blotting was carried out as described (23) using an anti-Myc-horseradish peroxidase antibody (Invitrogen, Leek, The Netherlands). Bound antibody was detected using the enhanced chemiluminescence system (Amersham Pharmacia Biotech).

RESULTS
cDNA Sequence Analysis of ACEH-A partial cDNA encoding ACEH was originally identified as a zinc metalloprotease with homology to ACE, in an EST data base. Following isolation of this partial clone from a human lymphoma cDNA library, the full cDNA encoding ACEH (Fig. 1) was deduced in combination with 3Ј-and 5Ј-RACE. Sequence analysis of 5Ј-RACE DNA products revealed an appropriate product that consisted of 100 nucleotides of extra 5Ј sequence. This gave rise to a possible new initiating methionine codon 61 amino acids upstream of the original putative ATG. Two products containing extra 3Ј sequence were identified from sequence analysis of 3Ј-RACE DNA products. The sequence differed at the termination codon of the original clone and downstream from there on. Product 132 contained an extra 207 nucleotides of sequence, including 36 extra codons. Product UB4 contained this sequence plus an additional 774 nucleotides of downstream 3Ј sequence, a total of 981 nucleotides of additional sequence. Hence, the cDNA encoding ACEH consists of 3405 nucleotides with 103 nucleotides of 5Ј-untranslated sequence, 2418 nucleotides of open reading frame, and 884 nucleotides of 3Ј-untranslated region (Fig. 1). The ATG codon at nucleotide position 104 has been assigned as the initiating methionine as it is preceded by two in-frame termination codons, TAA at nucleotide 26 and TAG at nucleotide 59. A polyadenylation signal is located 24 nucleotides upstream of the poly(A) tract.
The open reading frame encodes 805 amino acids, including a potential 17-amino acid N-terminal signal sequence, and has a predicted size of 92.4 kDa. There are 7 potential N-glycosylation sites within the protein sequence. Hydropathy analysis has also revealed a hydrophobic region toward the C terminus of ACEH, indicating that the protein is likely to be membranebound. The predicted amino acid sequence exhibits significant homology to existing members of the ACE family with ϳ60% similarity and 40% identity to the N-and C-terminal domains of human ACE and ϳ56% similarity and 36% identity to AnCE and ACEr. In addition to the conserved zinc metalloprotease consensus sequence, HEXXH, at amino acid positions 374 -378, there is also a conserved glutamate residue (Glu-402) predicted to serve as the third zinc ligand (corresponding to Glu-389 and Glu-987 of the N and C domains of ACE, respectively) 24 amino acids downstream from histidine residue 378. Asp-393 and Asp-991 of the N and C domains of ACE, respectively, which have been proposed to function in the positioning of the first histidine ligand (24), are replaced by a glutamate in ACEH, which may fulfill a similar role. Mutagenesis of the Asp-991 to glutamate in ACE reduces, but does not eliminate, activity (24). There are potential casein kinase II and tyrosine kinase phosphorylation sites in residues 787-790 and 775-781, respectively.
Expression of ACE mRNA and ACEH mRNA in Human Tissues-The expression of mRNA encoding ACEH and ACE was examined in human tissues. Multiple tissue Northern blots were probed with a 32 P-labeled fragment of the cDNA encoding either ACEH or ACE. Autoradiography revealed that expression of ACEH was greatest in kidney, testis, and heart, and moderate levels were also detected in colon, small intestine, and ovary ( Fig. 2A). A single mRNA species of ϳ3.4 kb was detected in these tissues, and an additional, less abundant, 5.9-kb species was also detected in kidney and testis. Two mRNA species were detected in tissues probed with endothelial ACE, of ϳ4.3 and 3.5 kb, corresponding to the previously described alternatively spliced variants of endothelial ACE (25). Expression of ACE mRNA appeared to be more widespread than ACEH, being found in colon, small intestine, ovary, testis, prostate, heart, placenta, liver, skeletal muscle, and pancreas, with the highest levels of expression in lung and kidney (Fig.  2B). The Northern blots were also probed with ␤-actin as a control (Fig. 2C).  protein without its putative C-terminal membrane binding domain and in conjunction with a C-terminal Myc-His fusion protein allowed detection of the protein by Western blotting (Fig. 3). When subjected to SDS-PAGE, the expressed secreted Myc-His tagged enzyme migrated with a molecular mass of ϳ120 kDa (Fig. 3, lane 2), indicating that the protein was glycosylated. Deglycosylation of ACEH with PNGase F resulted in the migration of the protein at the predicted molecular mass of ϳ85 kDa (Fig. 3, lane 1). No protein expression was detected in media taken from untransfected CHO cells. Secreted ACEH protein obtained from the medium was also used to identify potential substrates for the enzyme. High performance liquid chromatography (HPLC) was used to analyze the cleavage products. A parallel preparation taken from the medium of untransfected CHO cells was unable to hydrolyze angiotensin I (Fig. 4A). However, in the presence of ACEH, angiotensin I (retention time 9.4 min) was hydrolyzed to give a single product with a retention time of 6.7 min (Fig. 4B). Mass spectrometric analysis of the peptide recovered from the product peak gave an observed M r of 1183.9 which indicated that ACEH was acting as a carboxypeptidase to cleave the C-terminal leucyl residue from angiotensin I, producing angiotensin-(1-9). This activity was completely inhibited by 10 mM EDTA (Fig. 4C), but activity was unaffected by 10 M lisinopril (Fig. 4D), enalaprilat, or captopril (data not shown). In addition to angiotensin I, ACEH was also able to hydrolyze angiotensin II (Fig. 4E) to give products that co-migrated with angiotensin-(1-7) and phenylalanine but was unable to cleave bradykinin (Fig. 4F) or Hip-His-Leu (data not shown). Identical products were obtained by incubation of angiotensins I or II with carboxypeptidase A. The carboxypeptidase A inhibitor, benzylsuccinate, did not inhibit the hydrolysis of angiotensin I by ACEH under conditions that abolished the hydrolysis by carboxypeptidase A. Hydrolysis of bradykinin or Hip-His-Leu, by ACEH, did not occur even following overnight incubation (data not shown). Overnight incubation of ACEH with the typical carboxypeptidase A substrate, Hip-Phe, resulted in approximately 20% hydrolysis.

Expression and Enzymic Activity of an ACEH Construct-To
Genomic Sequence Analysis of the ACEH Gene-Searches of the GenBank TM data base with the cDNA-encoding ACEH revealed the corresponding genomic sequence. This was located in a sequence submitted to the data base by the Human Genome Sequencing Center, Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, TX. The sequence, obtained from Genome Systems Human BAC library, was defined as Homo sapiens Xp22, BAC GS-594A7 and has accession number AC003669. The ACEH gene contains 18 exons, interspersed with 17 introns, spans approximately 40 kb, and is localized to chromosome X, position p22. All the intronexon junction sequences (Table I) follow the GT/AG rule of Breathnach and Chambon (26). Exon sizes range from 59 nucleotides to 981 nucleotides. There is remarkable size similarity between the sizes of exons 1-4 of ACEH and exons 1-4 of ACE and also exons 7-12 of ACEH and 6 -11 of ACE (Fig. 5). Exon 9 of ACEH contains the HEXXH zinc metalloprotease consensus sequence, whereas exon 8 contains the first HEXXH motif in the genomic sequence of ACE. DISCUSSION Following the identification of a novel zinc metalloprotease from an EST data base, we have determined its cDNA sequence, expressed it as a soluble protein, and determined its activity toward potentially important physiological substrates. As a single domain enzyme, ACEH is similar to AnCE and ACEr, the insect members of the ACE family. When compared with the human ACE isoforms, ACEH shares considerable homology which is particularly marked around the HEXXH zinc-binding domain. This sequence (HEMGH) is identical in ACE and ACEH. A conserved glutamic acid residue, 24 amino acids downstream of the HEXXH motif in ACEH, aligns with the critical glutamate necessary for the catalytic activity of ACE (24). This glutamate serves as the third zinc coordinating ligand. ACEH also contains 8 cysteine residues 6 of which are conserved in the N-and C-terminal domains of endothelial ACE and of testicular ACE. ACEH contains 7 potential Nlinked glycosylation sites (compared with 10 and 7 in the Nand C-terminal domains of endothelial ACE, respectively) and is therefore likely to be glycosylated. This is further reinforced by the molecular mass of truncated, expressed ACEH that migrates at ϳ120 kDa compared with the deglycosylated polypeptide that migrates at 85 kDa.
There is a putative transmembrane domain of 22 amino acids near the C terminus followed by a cluster of charged residues that are likely to constitute a stop-transfer sequence. In contrast, the transmembrane sequence of ACE is predicted to be only 17 amino acids, which is a minimal requirement for a membrane-spanning region. Together with the 17-amino acid signal sequence at the N terminus of ACEH, the enzyme has all the features of a type I integral membrane protein, like ACE. We have shown that a soluble form of ACEH, lacking the transmembrane and cytosolic domains, is secreted from CHO cells and that this form is catalytically active. Surprisingly, ACEH appears to be acting specifically as a carboxypeptidase, rather than as a peptidyl dipeptidase, as it is able to cleave exclusively the C-terminal residues from both angiotensin I and angiotensin II. Bradykinin, which has a C-terminal arginyl residue, is not hydrolyzed, suggesting a carboxypeptidase Alike specificity for ACEH. However, ACEH does not have a typical carboxypeptidase A-like zinc-binding motif (27). Several ACE inhibitors (lisinopril, captopril, and enalaprilat) were not able to inhibit the cleavage of angiotensin I by ACEH, although the metal-chelating agent, EDTA, was an effective inhibitor, showing complete inhibition at 10 mM. This reinforces the proposition that ACEH is a metalloprotease, but with a distinct substrate and inhibitor specificity from ACE. It is perhaps not unexpected that the typical ACE inhibitors do not inhibit ACEH as they have been designed to compete with peptides that are hydrolyzed to release C-terminal dipeptides. Hence, positioning of the inhibitors in the active site should not be in the correct conformation to affect the cleavage of a single amino acid from the C terminus of the substrate.
The high expression of ACEH mRNA in heart and kidney is of interest as these organs are important contributors to blood pressure homeostasis. The highest expression of ACEH mRNA, however, is in testis. Testicular ACE is known to play a key role in fertility (28,29), and ACEH may also therefore have reproductive functions. The tissue distribution of ACE mRNA is more widespread than ACEH, with both 3.5-and 4.2-kb species present in most of the tissues examined.
The genomic sequence of ACEH holds many similarities to the structure of the ACE gene. The sizes of many of the exons are identical. There is, however, a discrepancy at exons 5 and 6  of ACEH, which together appear to correspond with exon 5 of ACE, suggesting a fusion of two exons. The HEXXH motif is therefore located in exon 9 of ACEH but exon 8 of the ACE gene.
The ACEH gene is located on the X chromosome (Xp22) which is similar to the location (Xp22.1) of another membrane metalloproteinase, the product of the PEX gene associated with X-linked hypophosphatemic rickets (30). The Pex protein is a member of the neprilysin (NEP) family (31).
Taken together, these findings indicate that ACEH is a metalloprotease that may have a significant role not only in cardiovascular homeostasis but also in fertility. However, before the physiological roles of ACEH can be elucidated, further enzyme characterization is needed to identify selective inhibitors and the key residues that distinguish its activity from that of ACE itself.