THE JOURNAL OF BIOLOGICAL CHEMISTRY D 1991 by The American Society for Biochemistry and Molecular Biology, Inc

Vol. 266, No. 23, Issue of August 15, pp. 15377-15383, 1991 Printed in U.S.A.

Structure of the Angiotensin I-converting Enzyme Gene TWO ALTERNATE PROMOTERS CORRESPOND TO EVOLUTIONARY STEPS OF A DUPLICATED GENE*

(Received for publication, November 1, 1990)

Christine Hubert$, Anne-Marie Houot, Pierre Corvol, and Florent Soubrier From the Institut National de la Santi et de la Recherche Medicale Unit 36, College de France, 3 Rue d’Ulm, 75005 Paris, France

Overlapping genomic clones containing the entire sequence of the human angiotensin I-converting en- zyme (ACE) gene were isolated from a X phage human DNA library. This gene spans 21 kilobases (kb) and comprises 26 exons, ranging in size from 88 to 481 base pairs. Intron-exon boundaries were sequenced and the relative positions of the exons were mapped. The two different mRNAs transcribed from the ACE gene were assigned to their respective exons. The large endothelial type ACE mRNA (4.3 kb long) is tran- scribed from exon 1 to exon 26, excluding exon 13. The 3-kb long testicular ACE mRNA is transcribed from exon 13 to exon 26. Exon 13 encodes for the 67 amino acids of the NH2-terminal region of the testicu- lar ACE, whereas downstream exons encode a se- quence common to both isozymes. The gene duplication suggested by the internal homology of the endothelial ACE mRNA is now confirmed by the presence of two homologous clusters of eight exons (exons 4-11 and exons 17-24) having similar sizes and codon phases at exon-intron boundaries. The presence of two alternate promoters was investigated by ribonuclease protection assays. The different 5’ ends of the two ACE tran- scripts revealed a promoter for the endothelial ACE mRNA in the 5”flanking region of the first exon and a promoter for the testicular ACE mRNA situated in intron 12.

Angiotensin I-converting enzyme (ACE)’ (dipeptidyl car- boxypeptidase I, EC 3.4.15.1) is a peptidyldipeptide hydrolase belonging to the class of zinc metalloproteases. It is, with renin, one of the two major enzymes of the renin-angiotensin system and converts the decapeptide angiotensin I into the active octapeptide angiotensin 11. ACE is also able to hydro- lyze the vasodilator peptide bradykinin. The wide distribution of ACE in many body tissues suggests that, in addition to its major role in the metabolism of vasoactive peptides, ACE is probably involved in the metabolism of other peptides, such as neurotransmitters (Erdos and Skidgel, 1987).

The structures of the two forms of the human enzyme have

* This work was supported by Institut National de la Sante et de la Recherche Medicale and Centre National de la Recherche Scien- tifique and by a grant from Bristol-Myers-Squibb Institute for Med- ical Research. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) M73275 and M73276.

f To whom correspondence should be addressed. ’ The abbreviations used are: ACE, angiotensin I-converting en-

zyme; kb, kilobase; bp, base pair.

been determined by molecular cloning of their respective cDNAs (Soubrier et al., 1988; Lattion et al., 1989; Ehlers et al., 1989). A large form ( M , = 170,000) is present in endothelial cells and in the brush border of epithelial cells. A smaller form (Mr = 90,000) is present in the germinal cells of the testis and is encoded by a shorter mRNA (3 instead of 4.3 kb). The large form of the enzyme has a 2-fold internal homology, and each of the homologous domains bears a pu- tative active site, suggesting gene duplication. The small form contains only one of these domains and has a different amino- terminal region. Although the specific substrate of the fully active ACE in germinal cells is not known, it must be impor- tant as this peculiar form is highly conserved, being found in all the mammalian species studied to date, including humans (El Dorry et at., 1982; Bernstein et aL, 1989; Velletri et ai., 1985; Strittmatter et al., 1985). The expressions of the two forms are also under different hormonal regulations; the endo- thelial enzyme is induced by glucocorticoids, whereas the testicular form is stimulated by androgens (Krulewitz et al., 1984; Velletri et al., 1985).

Southern blot experiments indicated that a single gene encodes for both the endothelial and testicular ACE mRNAs (Soubrier et al., 1988). This prompted us to elucidate which of two possible mechanisms, differential splicing or alternate promoters, are responsible for transcription of the two mRNAs, and to determine the structural basis for their dif- ferent tissue specificities and hormonal regulation.

We report here the structural organization of the human ACE gene and show that it contains two overlapping tran- scriptional units. The sequences of ACE cDNAs in other mammalian species have been used to make intragenic and interspecies sequence comparisons and thus provide insight into the evolutionary features of the ACE gene.

MATERIALS AND METHODS

Screening ojthe Human Genomic DNA Library-A human genomic library, prepared by partial AluI-HaeIII digests of human fetal liver DNA and cloned in X phage charon 4A with EcoRI linkers was kindly provided by Drs. Lawn and Maniatis (Lawn et al., 1978). Two million clones were plated on Escherichia coli strain LE392 and screened by plaque hybridization (Benton and Davis, 1977). Filters were hybrid- ized in duplicate with inserts of phage clones X1922 and X2111 containing nearly the whole endothelial ACE cDNA sequence (Souh- rier et al., 1988). These DNA inserts were labeled with [ w ~ ’ P ] ~ C T P by the random primer labeling method (Feinberg and Vogelstein, 1983) using a commercial kit (Amersham Corp.). Hybridization was carried out at 42 “C in 50% formamide, 5 X SSPE (1 X SSPE is 150 mM NaC1, 10 mM NaHgPO,, 1 mM EDTA, pH 7.4), 5 X Denhardt’s solution, 0.1% sodium dodecyl sulfate, and 100 Fg/ml denatured salmon sperm DNA. Filters were washed at 65 “C in 0.1 SSC (1 X SSC is 150 mM NaC1, 15 mM sodium citrate, pH 7), 0.1% sodium dodecyl sulfate and autoradiographed a t -80 “C with an intensifying screen. Positive clones were isolated by repeated phage purification. Phage DNA was purified by the standard method of cesium density gradient centrifugation (Maniatis et al., 1982).

15377

15378 Human Angiotensin I-converting Enzyme Gene

FIG. 1. Organization of the hu- man ACE gene and restriction map of the genomic clones. Top line, scale in kilobases. Second line, location of the 26 numbered exons (vertical bars). Exon 13 (open bar) is specific to testicular ACE. The cleavage sites of seven restric- tion endonucleases are indicated. Those marked with an asterisk are also present in cDNAs. The position of four overlap- ping X phage inserts (hatched bars) and the plasmid subclones (lines) are indi- cated at the bottom of the figure. Plas- mid numberings correspond to their re- spective lengths.

0 10 2 0 30 k b

E x o n s 2 4 6 8 10 12 11 16 18 20 22 24 26

Restr ic t ion maps I 3 5 7 9 11 13 IS 17 I9 21 23 25

ECoR I

BamH I

Hind Ill I I I 1

I I I I I I

I I I I I

11915

)HA 6 5 )HA 4.1 pHA 1.2 - )HA 3 2 c__

)HA 3.5 _.

pHA43 pHAl - 7 )HA 2 I

c__

)HA 3

"

23.1 - 9.4 - 6.5 - 4.3 -

"". I

2.3 - 2.0 -

FIG. 2. Southern blot analysis of the human ACE gene. Hu- man genomic DNA was digested with the restriction endonucleases indicated on the top. The filter was hybridized with a human ACE cDNA fragment corresponding to positions 3778 to 4024 (see "Ma- terials and Methods"). Hind111 fragments of X phage DNA were used as size markers and are indicated in kilobase pairs.

Restriction Endonuclease Mapping of Genomic DNA Clones-Phage DNAs were digested with one or more restriction endonucleases and the resulting fragments were separated by agarose gel (1%) electro- phoresis, denatured in alkali, and transferred to nylon membrane (Hybond N, Amersham Corp.) as described by Southern (1975). Filters were hybridized with selected oligonucleotides, designed from the nucleotide sequence of the cDNA coding for the endothelial ACE, and labeled with [y-:"PP]ATP at their 5' end by T4 polynucleotide kinase (Maniatis et al., 1982).

Subcloning and DNA Sequencing-Selected restriction fragments of genomic DNA inserts were subcloned into the plasmid vectors Bluescript (Stratagene), pTZ18R, and pTZ19R (Pharmacia LKB Biotechnology Inc.).

Detailed restriction maps of each subclone were prepared by re- striction enzyme digestions and Southern blotting. DNA sequencing was performed by the dideoxy chain termination method (Sanger et al., 1977) with T7 (Sequenase, United States Biochemical) or E. coli DNA polymerase (Klenow fragment, Boehringer Mannheim) and a set of synthetic oligonucleotide primers designed from the ACE cDNA sequence.

Southern Blot Analysis of Human Genomic DNA-High molecular

weight DNA was isolated from human placenta (Maniatis et al., 1982). DNA samples (14 pg) were incubated with restriction endo- nucleases (New England Biolabs). Digested DNAs were electropho- resed through a 0.7% agarose gel in 40 mM Tris, 20 mM sodium acetate, 2 mM EDTA, pH 7.8, transferred to nylon filter and hybrid- ized with n-"2P-labeled FspI-EcoRI restriction fragment of ACE cDNA, corresponding to positions 3778-4024 of the endothelial ACE cDNA. The filter was hybridized and washed as described in the screening section.

Mapping the 5' Ends of the Endothelial and Testicular Transcripts by Ribonuclease Protection Assay-Total RNAs from human endo- thelial cells and human testis were prepared by the method of Chirg- win et al. (1979). Poly(A)+ RNAs were purified on oligo(dT) cellulose. Two restriction fragments were subcloned into plasmid vector Blue- script in order to synthesize both sense and antisense transcripts corresponding to the putative 5' ends of the endothelial and testicular ACE transcripts. A 617-bp ApaI-PuuII fragment bearing 225 bp of the 5' end of the endothelial ACE cDNA sequence and the upstream genomic sequence was subcloned into Bluescript SK (pEH18-16). A 244-bp BanII-Ban11 fragment corresponding to 219 bp of the 5' extremity of the testis ACE cDNA sequence and to 25 bp of the upstream genomic sequence was treated with T4 DNA polymerase and subcloned into Bluescript SK (pTH35-1). The conditions used for the following experiments were as in Zinn et al. (1983) unless otherwise stated. In vitro transcriptions were performed in the pres- ence of [a-:"P]UTP to synthesize labeled antisense RNA. Approxi- mately 5 X 1 0 cpm of the probe was coprecipitated with RNAs. Hybridization was performed at 45 and 35 "C for pTH35-1- and pEH18-16-derived RNAs, respectively, in a hybridization mix for 24 h, and was followed by digestion with ribonucleases TI and A. After phenol-chloroform extraction and ethanol precipitation, RNAs were applied to a 5% acrylamide, 7 M urea sequencing gel. Sequencing reactions were used as size markers together with the undigested labeled antisense RNA.

Sequence Alignment-The 1484-bp upstream from the tsp of the endothelial ACE mRNA and intron 12, which corresponds to the DNA region upstream from the tsp of the ACE testicular transcript, were optimally aligned with the program NUCALN (Wilbur and Lipman, 1983). The same program was used to compare the sequences of the human, mouse, and rabbit ACE genes.

RESULTS

Isolation and General Organization of the Human ACE Gene-Out of 2 million phage clones screened with the human endothelial ACE cDNA probe, 18 phages were selected. The isolated genomic clones were segregated into four overlapping X phages (X1011, X1915, X311, and X621) spanning a 35 kb region of continuous human genomic DNA (Fig. 1). No other

Human Angiotensin I-converting Enzyme Gene 15379

TABLE I Nucleotide sequence of the intron-exon junctions in the human ACE gene

Exon sequences are in uppercase letters; intron sequences are in lowercase. The positions of cDNA exons are relative to the start site of the endothelial ACE mRNA, except for exon 13 which is positioned relative to the testicular ACE mRNA start site. The sizes of the introns were estimated from restriction endonuclease analysis or from nucleotide sequences. The codon phase refers to the position of intron in the codon triplet. 0 intron occurs between codons, I intron occurs after the first nucleotide, and I1 intron occurs after the second nucleotide in the codon.

cDNA Exon no.

(size) endothelial position of

exon

5' splice donor

bp 1 (271) 2 (168) GCAGgtgggc

4 (144) CCAGgtacgc

5 (192) GACGgtgagc

6 (98) CTGGgtaagg

7 (173) GCAGgtaagc

8 (224) 9 (145) 1365-1509

ACGGgtatcg TTCGgtgaga

11 (123) 1609-1731 TCAGgtatta TCCGgtgtgt

13 (228) 1-228 ATAGgtaaag CCAAgtggga

14 (137) 1944-2080 15 (159)

TCTGgtggga

16 (88) 17 (144)

CCAGgtgaga

18 (192) CTGGgtaagg

20 (171) AGGGgtccgc

21 (224) 2935-3158 GACGgtcgag 22 (145) 23 (99)

TCAGgttctg

24 (122) 3403-3524 TCAGgtaacg CTGGgtgagt

25 (189) 3525-3713 26 (481) 3714-4195

TCCGgtaccg

1-271 CCAGgtgggc 272-439

3 (94) 440-533 534-677 678-869 870-967 968-1140 TCAGgttcag

1141-1364

10 (99) 1510-1608

12 (212) 1732-1943

2081-2239 GGAGgtcgtc 2240-2327 2328-2471 AATGgtgagt 2472-2663 2664-2763 2764-2934 TCCGgtacat

19 (100)

3159-3303 3304-3402

TABLE I1 Similarity between exons from the first and the second domains of

the ACE gene Codon phases are explained in Table I.

Exon no, Si.es Nucleotide sequence similarity Codon phase

bP % 4 and 17 144/144 54.8 I GIy/I Gly 5 and 18 192/192 70.3 I Gly/I Gly 6 and 19 98/100 67.3 0 Gln-Gly/II Gly 7 and 20 173/171 80.1 I1 Arg/II Arg 8 and 21 224/224 68.3 I Glu/I Glu 9 and 22 145/145 64.8

10 and 23 99/99 I1 Arg/II Arg

60.6 11 and 24 123/122

I1 Arg/II Arg 73.7 I1 Arg/I Ala

type of hybridizing phage was isolated from this library. Several fragments were detected by Southern blot hybridiza- tion of human genomic DNA with the complete ACE cDNA as probe and were all identified in isolatedphages. In contrast, only a single DNA band was hybridized when a short restric- tion fragment of the cDNA, located within exon 26, was used as probe (Fig. 2). These results confirm that a single gene encodes for ACE in the human genome.

A set of oligonucleotides designed from the coding sequence of the endothelial and testicular ACE cDNAs was used to determine the gene structure of ACE by hybridization with cloned genomic sequences. Intron-exon boundaries were se- quenced (Table I). The DNA was sequenced using ACE cDNA-specific oligonucleotides to prime chain elongation on subclones containing the genomic fragments. Intron sizes

Intron no. (size)

bP

3' splice acceptor Codon phase

1 (-600) tgcccaatagGAGG 0

3 (-700) gttgtctcagATCT I 4 (-450) agcccccagGCTT I 5 (-600) ccttatctagGAGA I 6 (-300) cctcctccagGGCT 0 7 (-700) taggccccagGATC I1 8 (-300) ctcccaacagAAAG I 9 (-250) ccttttccagAACC I1

10 (-300) gtgcctgcagGTAC I1 11 (-350) gcccaccagGAAG 11 12 (299) ggggacggtAGCT I 13 (170) tgtcccacagACCT I 14 (-1150) tgtccctcagCTGC 0 15 (-250) CctcctgcagTACA 0 16 (-1850) cccctaccagATCT I 17 (150) ccctctgtagGCTA I 18 (-1800) tcgcctccagGGAA I 19 (156) ctgcttgcagCTGG I1

21 (-250) tgcagAGCA I 22 (310) cactcgacagGCTG I1 23 (-1900) tctgctccagGTAC I1 24 (-250) gcctccccagCGAC I 25 (150) ttccactcagCTCG I

2 (-900) gactccccagTACA 0

20 (-2000) ctattcctagGATC I1

were determined on genomic subclones by restriction endo- nuclease mapping or DNA sequencing (Table I). Some intron sizes were verified on isolated X phages by DNA amplification using primers of the flanking coding region. The human ACE gene contains 26 exons interrupted by 25 introns and spans approximately 21 kb of DNA (Fig. 1). All the exon-intron junction sequences follow the GT/AG rule of Breathnach and Chambon (1981). The 26 exons range in size between 88 bp (exon 16) and 481 bp (exon 26). The 25 intervening sequences vary considerably in length, ranging from 150 bp (introns 17 and 25) to 2 kb (intron 20) (Table I). The polyadenylation site has been determined for the testicular ACE mRNA (Lat- tion et al., 1989; Ehlers et al., 1989). The sequence surrounding the stop codon in the genomic fragment is identical to the published sequences until the polyadenylation signal which is localized 238 bp downstream from the stop codon. For the endothelial ACE mRNA the polyadenylation site was not reported, other site(s), downstream from the testicular poly- adenylation site, cannot be ruled out.

Endothelial ACE Results from Transcription of a Duplicated Cluster of Exons-Previous molecular cloning of the endothe- lial ACE cDNA revealed a high degree of internal homology between two large domains representing the major part of the cDNA. In the gene, each homologous domain is encoded by a group of eight exons: exons 4-11 for the first domain, and exons 17-24 for the second domain. The symmetrical exons in each group have very similar sequences and sizes, and their codon phases are conserved (Table 11). Exon 8 and exon 21, which contain the consensus sequence of metalloprotease active site (H-E-X-X-H) are 68% similar, while exon 7 and

15380 Human Angiotensin I-converting Enzyme Gene G A T C l 2 3 4 5

c 325

c 219

-RNA n s b o

FIG. 3. Determination of transcription initiation sites by ribonuclease protection mapping. Lower panel: parts A and B, schematic diagram showing the cloned DNA fragment which was used as template for the transcription of antisense RNA used as probe. The restriction enzyme sites of the ACE gene used to isolate the fragments are indicated. Endothelial promoter: the ApaI-PuuII fragment was cloned in the ApaI and EcoRV site of Bluescript SK and the resulting plasmid (pEH18-16) was linearized by BspEI, 14 bases downstream from the ApaI site to synthesize antisense RNA probe. Testicular promoter: the BanII-Ban11 fragment was cloned in the EcoRV site of Bluescript SK and the resulting plasmid (pTH35- 1) was cut by Hind111 site of the polylinker to generate antisense RNA. The total length of the antisense testicular probe (325 bases), used as RNA size marker, is obtained by adding the length of the T3 promoter and of the polylinker to the cloned fragment. Top panel: part A, mapping of the endothelial ACE RNA. On the left, sequencing reactions used as size markers. Lane 1 , undigested testicular promoter antisense RNA probe. Lane 2, undigested endothelial promoter anti- sense probe. Lanes 3-5, RNase-protected antisense probe hybridized with 15 pg of cultured endothelial cell poly(A)’ RNA, 15 pg of tRNA, and sense probe, respectively. The calculated size of the protected fragment (225 b) is indicated on the right. Part B, on the left, sequencing reactions used as size markers. Lane 1, undigested anti- sense testis probe. Lanes 2 and 3, RNase-protected antisense probe hybridized with 10 pg of testis RNA and with sense RNA probe,

exon 20 are 80% similar. In contrast to the exons, the introns in the two halves of the gene differ in size. Four introns in the 3‘ part of the gene (introns 16,18,20,23) are much larger than the homologous introns in the 5’ part of the gene (introns 3, 5, 7, 10).

Endothelial ACE Promoter-The transcription start point (tsp) of the endothelial ACE mRNA was determined using the ribonuclease protection assay (see “Materials and Meth- ods”). The main tsp observed was a purine nucleotide (G) 22 bases upstream from the ATG start codon; this tsp corre- sponds to the 5‘ end of endothelial ACE mRNA that was previously determined by cDNA cloning using a specifically primed cDNA library (Fig. 3A) (Soubrier et al., 1988). The DNA sequence of a 1484-bp region upstream of the tsp was determined, and a search for core promoter elements was carried out (Fig. 4). A typical TATA box (TATAA) was found 37 bases upstream from the tsp, but no consensus CCAAT element was found. Three Spl potential binding sites are present and two of them are in tandem (positions -62, -68, -131) within a GC-rich region, reaching 78% on the 200 nucleotides upstream the start site (Dynan and Tjian, 1985). We also searched for hormone-responsive elements, based on the known in vitro regulation of ACE expression in endothe- lial cells (Krulewitz et al., 1984). The hexanucleotide of the consensus glucocorticoid-responsive element was found at position -1213 (TGTCCT), -935 (TGTCCT), -794 (TGTTCT), and -686 (TGTCCT) (Karin et al., 1984).

The Testicular Transcript Is Derived from the 3‘ Moiety of the ACE Gene-The 5‘ end sequence of the testicular mRNA, absent from the endothelial ACE mRNA, contains a non- translated DNA sequence and encodes for 67 amino acids, including a putative signal peptide. The entire testicular- specific DNA sequence is encoded by exon 13 (228 bp long), which is in the middle part of the gene, and is surrounded by exons which are present in the endothelial transcript (Fig. 1). All the exons following exon 13 (exons 14-26) are present in both the testis mRNA and the endothelial transcript, includ- ing the last exon, which encodes the membrane-anchoring segment. We have studied the transcription start site of the testicular mRNA using the ribonuclease protection assay. Human testis poly(A)+ RNA was hybridized with a labeled anti-sense RNA transcribed from a restriction fragment over- lapping the 3’ part of intron 12 and 5’ part of exon 13. A major transcription start point, at 28 nucleotides upstream from the translation initiation site, was found which corre- sponds to the 5‘ end determined on the cDNA clone obtained by screening a specifically primed cDNA library (Fig. 3B).

Sequence Analysis of Intron 12 as an ACE Testicular Pro- moter-The ribonuclease protection assay showed that the 5’ end of testis ACE mRNA corresponds to the 5’ end of exon 13, supporting the hypothesis that a promoter is located upstream, within intron 12. The complete sequence of intron 12 is shown in Fig. 4. We detected a TATA-like sequence (TATT) 15 bases upstream from the major testis mRNA tsp. A sequence very similar to the consensus sequence of the CAMP-responsive element (TGAGGTCA) is present a t posi- tion -44 from the tsp (Roesler et al., 1988). A sequence homologous to the steroid-responsive element was found at position -190 (TGTCCT).

A computer search for homology between the endothelial promoter and intron 12 was carried out with the program NUCALN (Wilbur and Lipman, 1983). Two regions of the endothelial promoter had sequences similar to intron 12. A region between positions -1238 and -1138 was 54% homol-

respectively. Sizes of the undigested probe (325 b) and of the protected fragment (219 b) are indicated on the right.

Human Angiotensin I-converting Enzyme Gene 15381

CTGCAGGTCAGGAGGGAGGCTGGCAAAGCCGGCCAGGGCCAGGCCTCACCCCAGCTCTCCTCTCCCCCCCCCACTGCCGGCCAGTGTTTAACAAGGCCCTGCCTTCTCCCTCTAGTGCTA-1364

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . . . . . GTAAAGCCCTGAGTGAGGATCGTGTGGGGCTAAGGTCGGTCCTC~CTCTGGGCTTGGCCCAGGCCCCAGGTTCCTC

- 2 0 0 -260 -240

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . GTCAGCTCCTACCAGCTGAGCCCTGGTACCCTGTCCTG-GAGGGCCAGGCAGCCCCCC~GCTCATCAGCACGGCCTGCGAGTGGGGACAGGCATGTCTTTCCCCCAGCATCCTAGAGA-

-200 -180 -160 -140 -120

TTGGGGTCTTGGGAGCAGCTGTCCCTGGCTCCTTCCCTGATCCCACCGCCCAGCCTCACCCCACGGTTCCTCCATTGCCCCACCTCCCACTGCGCCGCCGGGCCTCTGCCAGGGTCAAGG-1014

GGCTTCCCCCCTCTGGCAGCAGACGCCATGGTGCCGAGGTGGCCTCCACAACCGCCCTGTGCGCC~TAGGACAAGACTGTCCTCCCTCCCCCACACTTGTCACTTTGAGGG~CACGTGG -094

............................................................

ATGAGACACGAAAACACAGCGCACTGTGGAGAGACCTCACGTCACTTGGAGCAACCCTCTCAACCTGACCGGCAATTTCTTCATCTGT~TGAGGGGGTTGTTCTCATCTCTGAGGCTTT -774 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ."""""""""

-100

GTGTCGCT~TCAAAGCCTGCTAGCCTCGGGTTCTAGCACTCTGTTGCGATC~TGTGTGATGTTTTCTGCTGAGCGACTGGCAGCCTGTGTCCTCGGGGCGAAAGAGGGCAGGCGCTCCAA -654

AGCTCCTGCGCTCTGTGGCTCCCCCTCCCTCGCAGCCCCAAGCCCCP.CGTGTGCCGCCCGCCCTGAGCCCCTCCACCACCTCCCGGAGGCCCCTGCAAGACACCTAAGGTCCCCGCCTCC -534 """""""""_""""""""""""""""""""""~"""""""""""""""""""""""""""" CTCCTCTCCCCCCCGCCACACCCCTACCCCCGGCAGG~GACGTCCCCGCCCCTCGACCATliCCCTG(i'~CAAGAAGCCGGCCAGCCCCGATCAGCCCCATCCCCGCCGCACGAGCGGCGCC - 4 1 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . """""__"""""""""""""""""""""""""""""""""""" CCCTCCCCTTCCAGAGGAAGCCAGACACAAGGCTCT

- B O -60

TGCGGACAGCTCCTGGGGCCCCGGCCTTCTCACTCCCGAGGCCGG~GGC~C~GG~CCGTCGGCCTGGGAACATCGAGCCCGAGGCCGCTAGGCTCCCAGGCCCCGGCCGAGGCTGCGCGG -294

CTGACCTCACACTGCGGGCTCCGCTCT~GCCCAGCGGACCGTAGCTGCAGG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

-4c -20

CCGCACGGTGGGCAGGCTCGGGTGTTCCGGCAAA~TGCCGCGTCCCCATCTTCAAAACAGAGGACGCCCTTTCTCCAGCTTCCTCTGCGGGAGCCCGACCCAGCCCCATCCCGCCACCCC - 1 7 1

CGGGCTGCACCTCGGCCCCTCCCCGGCCCGCCCCCCTGCCCGCGGCCGGCCAGGAACCTCGGCCCGCGCCCCTCGGGACTTTGGAGCGGAGGAGG~GCGCGGCGGGGCGGGGGCGGG~G -54

TGTCTCGGGTTT~CCCGCAGGGCGGCCGCGGCGCAGGAGAAGGGGCAGA~CCGAGCACCGCGCACCGCGTCATG +25

FIG. 4. Sequence comparison of the 5' upstream regions of endothelial and testicular ACE mRNAs. Optimal alignment of 1484 bp of 5"flanking region of endothelial mRNA (upper line) with the 299 bp of upstream sequence of the testicular mRNA (lower line) were obtained with the program NUCALN (see "Materials and Methods"). The endothelial promoter sequence is identical to the sequence recently published by Shai et al. (1990) after this work was completed and is numbered on the right. The intron 12 sequence is numbered under the lower- line. TATA-like sequences and nucleotide of the transcription start sites are underlined.

ogous to the sequence between positions -216 and -107 of intron 12 and another region between positions -449 and -367 of the endothelial promoter and between positions +1 and -81 of intron 12 was 46% homologous. Using the same alignment a short 20-nucleotide region, separated from the two other regions by large gaps, was 80% similar in the two promoters (positions -874 of the endothelial promoter and -103 of intron 12).

Interspecies and Intragenic Nucleotide Sequence Compari- sons-We calculated the approximate rate of divergence per million years between human, mouse, and rabbit at the ACE gene locus by comparing the nucleotide sequences of exons 7 and exons 20, which are the highest homologous regions between the two domains of the ACE gene (Bernstein et al., 1989; Kumar et al., 1989). The divergence rate for exon 20 was 13.5% between human and mouse, 8.2% between human and rabbit, and 13.9% between mouse and rabbit. The diver- gence rate for exon 7 between human and mouse was 13.9%, showing that there is no significant difference between the divergence rates of the two parts of the gene. This calculation could not be done for the rabbit as the endothelial form of rabbit cDNA has not been described. Assuming that the mammalian radiation occurred 80 million years ago, a 1% divergence rate at the exon 20 ACE gene locus is reached in 8018.2 = 9.8 million years from the human-rabbit comparison, in 80113.9 = 5.7 from mouse/rabbit comparison, and in 801 13.5 = 5.9 million years from human-mouse comparison. Intragenic sequence comparison shows that exon 7 presents a 19.9% divergence rate from its homologous exon 20 in the human ACE gene. The same comparison shows a 24.6%

divergence rate in the mouse ACE gene. It is assumed frcm these calculations that ACE gene duplication occurred I: ? -

tween 118 and 194 million years ago, from human/mouse and humanlrabbit comparisons, respectively.

DISCUSSION

The structural organization of the entire gene coding for the two ACE isozymes was established. Out of 26 exons, the endothelial ACE is encoded by 25 exons and the testicular ACE is constituted by 14 exons whose 13 are common with the endothelial enzyme. Molecular cloning of ACE endothelial and subsequently of testicular cDNA showed that the struc- ture of the gene resulted from gene duplication (Soubrier et al., 1988; Lattion et al., 1989). This result is strengthened and extended to the structure of the ACE gene described in this paper. Not only are the sequences of the two domains consti- tuting the major part of the molecule similar, but the homol- ogous exons encoding the two domains have similar sizes and similar codon phases at the exon-intron junctions.

The gene duplication appears to have occurred very early in evolution. Two types of observations support this hypoth- esis. First, the molecular weight of the circulating enzyme or the molecular structure of the ACE mRNA indicates that the ACE gene is probably duplicated in pig, mouse, rat, and rabbit (Hooper et al., 1987; Velletri et al., 1985; El Dorry et al., 1982). Second, interspecies and intragenic nucleotide sequence com- parisons (see "Results") suggest that ACE gene duplication occurred before the mammalian radiation. A more precise dating of the duplication will be possible once the structures of the ACE genes in non-mammalian vertebrates have been

15382 Human Angiotensin I-converting Enzyme Gene

determined. Enzymes similar to ACE and inhibited by the ACE inhibitor, captopril, have been found in both reptiles and birds (Fregly et al., 1981; Gervitz et al., 1987).

Such duplicated structures occur in other brush-border anchored enzymes, including sucrase-isomaltase and lactase- phlorizin hydrolase (Hunziker et al., 1986; Mantei et al., 1988). In the latter, 4-fold internal homology suggests two cycles of duplication. Each of the two domains in these two enzymes bears an active site having a separate catalytic activity (Schle- gel-Haueter et al., 1972; Sjostrom et al., 1980). The activities of each domain of ACE were studied by in vitro expression in our laboratory (Wei et al., 1991). Both domains possess en- zymatic activity, but they have different catalytic properties. Identification of residues responsible for these differences will be of great value for understanding the catalytic mechanisms of ACE. The amino-terminal catalytic site may well have another substrate specificity, as it does in lactase-phlo-rizin hydrolase and sucrase-isomaltase.

No overall sequence similarity was found between ACE and other metallopeptidases, such as neutral endopeptidase and collagenase. Only short amino acid sequences were found to fit the consensus sequences of the catalytic site of zinc metal- lopeptidases (Soubrier et al., 1988). This sequence H-E-X-X- H, in which the 2 histidines are believed to bind the zinc atom, occurs twice in the ACE molecule. Shuffling of an exon bearing this consensus sequence in metallopeptidase genes is not supported by the structure of ACE, collagenase, and neutral endopeptidase genes, as there is no sequence similarity between exons concerned (Collier et al., 1988; D’Adamio et al., 1989).

A remarkable feature of the ACE gene is the existence of two mRNA species, one corresponding to the duplicated gene and present in several tissues, and the other, corresponding to the 3’ half of the gene, expressed solely in the testis. Besides their different tissue specificity, expression of the two ACE transcripts is regulated by different developmental and hormonal controls. The level of expression is also very differ- ent; the testicular ACE mRNA species is much more abundant in the testis than is the larger form in other tissues, assuming that the mRNAs have similar stabilities (Soubrier et al., 1988). DNA elements responsive to glucocorticoids, which induce ACE secretion by cultured endothelial cells (Krulewitz et al., 1984), were found in the upstream region of the endothelial promoter. In vivo studies in the rat have shown that ACE expression in the testis is positively regulated by androgens and luteinizing hormone (Velletri et al., 1985). The testicular transcript could be formed by alternate splicing of a large ACE primary transcript or by transcription from an internal promoter. Mapping of the 5’ end of the testicular ACE mRNA by RNase protection showed that it starts at the 3’ edge of intron 12, indicating an internal promoter. We also found a TATA-like sequence and short DNA sequences very similar to CAMP and consensus steroid hormone-responsive element in intron 12. We therefore propose that the testicular tran- script is transcribed under the control of an internal promoter and suggest that intron 12 is probably responsible for tissue specificities of expression and hormonal regulation, although we cannot rule out a role of other regions of the gene. Simi- larly, Howard et al. (1990) recently showed that the mouse ACE testicular transcript originates from an internal pro- moter.

We identified two similar (50% identity) sequences, sepa- rated by large gaps, in the two putative promoters of the ACE gene. Their presence could indicate that they are derived from a common ancestral sequence and that they diverged toward

different regulatory and tissue specificities after the duplica- tion.

Both the structure of the testis ACE mRNA, transcribed from the 3’ half of the gene and the location of its promoter in the central region of the gene support the hypothesis that this mRNA arises from the ancestral preduplication form of the gene. Alternative promoters have been described in many other genes (for review see Schibler and Sierra, 1987). In these cases they ensure different developmental regulation and tissue-specific expression, but in none of them does transcrip- tion from the alternative promoter appear to produce an “ancestral” mRNA, anterior to duplication of the whole func- tional gene. Interestingly, rat farnesyl pyrophosphate synthe- tase mRNA also appears to have a testis-specific alternative promoter which produces abundant transcription of an mRNA with a different 5’ end (Teruya et al., 1990).

Transfection experiments, with either the endothelial ACE gene promoter or intron 12 driving transcription of a reporter gene, will be necessary to study the tissue specificities and hormonal regulation of the two alternative promoters.

Acknowledgments-We thank J. Allegrini, A. L. Lattion, T. Sarda, and P. Testut for their help during achievement of this work, N. Braure for typing the manuscript, A. Depardieu for artwork, and J-L. Sikorav for critical reading of the manuscript.

REFERENCES Benton, W. D., and Davis, R. W. (1977) Science 196,180-182 Bernstein, K. E., Martin, B. M., Edwards, A. S., and Bernstein, E. A.

Breathnach, R., and Chambon, P. (1981) Annu. Reu. Biochem. 50,

Chirgwin, J . M., Przybyla, A. E., MacDonald, R. J., and Rutter, W. J. (1979) Biochemistry 18, 5294-5299

Collier, I. E., Smith, J., Kronberger, A., Bauer, E. A., Wilhelm, S. M., Eisen, A. Z., and Golberg, G. I. (1988) J. Biol. Chem. 263, 10711- 10713

D’Adamio, L., Shipp, M. A., Masteller, E. L., and Reinherz, E. L. (1989) Proe. Natl. Acad. Sci. U. S. A. 86, 7103-7107

Dynan, W. S., and Tjian, R. (1985) Nature 316, 774-778 Ehlers, M. R. W., Fox, E. A., Strydom, D. G., and Riordan, J. F.

(1989) Proc. Natl. Acad. Sci. U. S. A. 86, 7741-7745 El Dorry, H. A., Pickett, C. B., MacGregor, J. S., and Soffer, R. L.

(1982) Proc. Natl. Acad. Sci. U. S. A. 79, 4295-4297 Erdos, E. G., and Skidgel, R. A. (1987) Lab. Znuest. 56, 365-368 Feinberg, A. P., and Vogelstein, B. (1983) Anal. Biochem. 132, 6-13 Fregly, M. J., Simpson, C. F., and Palmore, W. P. (1981) Proc. SOC.

Gervitz, R. K., Hiraichi, E., Fichman, M., and Lavras, A. A. C. (1987)

Hooper, N. M., Keen, J., Pappin, D. J. C., and Turner, A. J. (1987)

Howard, T. E., Shai, S. Y., Langford, K. G., Martin, B. M., and

Hunziker, W., Spiess, M., Semenza, G., and Lodish, H. F. (1986) Cell

Karin, M., Haslinger, A., Holtgreve, H., Richards, R. I., Krauter, P.,

Krulewitz, A. H., Baur, W. E., and Fanburg, B. L. (1984) Am. J.

Kumar, R. S., Kusari, J., Roy, S. N., Soffer, R. L., and Sen, G. C. (1989) J. Biol. Chem. 264, 16754-16758

Lattion, A., Soubrier, F., Allegrini, J., Hubert, C., Corvol, P., and Alhenc-Gelas, F. (1989) FEBS Lett. 252,99-104

Lawn, R. M., Fritsch, E. F., Parker, R. C., Blake, G., and Maniatis, T. (1978) Cell 15, 1157-1174

Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY

Mantei, N., Villa, M., Enzler, T., Wacker, H., Boll, W., James, P., Hunziker, W., and Semenza, G . (1988) EMBO J. 7, 2705-2713

Roesler, W. J., Vandenbark, G . R., and Hanson, R. W. (1988) J. Eiol. Chem. 263,9063-9066

(1989) J. Biol. Chem. 264,11945-11951

349-383

Exp. Biol. Med. 166,35-38

Comp. Biochem. Physiol. 86A, 503-507

Bwchem. J. 247,85-93

Bernstein, K. E. (1990) Mol. Cell. Biol. 10, 4294-4302

46,227-234

Westphal, H. M., and Beato, M. (1984) Nature 308, 513-519

Physiol. 247, C163-C168

Human Angiotensin I-converting Enzyme Gene 15383

Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc. Natl. Acad. Southern, E. M. (1975) J. Mol. Biol. 98, 503-517 Sci. U. S. A. 74, 5463-5467 Strittmatter, S. M., Thiele, E. A., DeSouza, T. E., and Snyder, S. H.

Schibler, U., and Sierra, F. (1987) Ann. Reu. Genet. 2 1 , 237-257 (1985) Endocrinology 117 , 1374-1379 Schlegel-Haueter, S., Hore, P., Kerry, K. R., and Semenza, G. (1972) TeruYa, J. H.9 KUtsunair s. y.l D. H., Edwards, p. A.3 and

Clarke, C. F. (1990) Mol. Cell. Biol. 10, 2315-2326 Shai, s., Langford, K. G., ti^, B, M., and ~ ~ ~ ~ ~ ~ ~ i ~ , K, E, (1990) Velletri, P. A., Aquilano, D. R., Bmckwick, E., Tsai-Morris, C. H.,

Dufau, M. L., and Lovenberg, W. (1985) Endocrinology 116,2516-

Sjostrom, H., Norin, O., Christiansen, L., Wacker, H., and Semenza, wei, L., Alhenc-Gelas, F., corvol, p., and clause,, E. (1991) J , ~ i ~ l , 2522

Chem. 266,9002-9008

Biochim. Biophys. Acta 258, 506-519

Biochem. Biophys. Res. Commun. 167 , 1128-1133

G. (1980) J. Biol. Chem. 255, 11332-11338

Tregear, G., and Corvol, P. (1988) Proc. Natl. Acad. Sci. U. S . A . A. 80, 726-730 85,9386-9390 Zinn, K., DiMaio, D., and Maniatis, T. (1983) Cell 34,865-879

Soubrier, F., Alhenc-Gelas, F.7 Hubert, c., Alkrini , J-, John, M., Wilbur, W. J., and Lipman, D. J. (1983) Proc. Natl. Acad. Sci. U. S.