Molecular Cloning of a Novel Angiotensin I1 Receptor Isoform Involved in Phosphotyrosine Phosphatase Inhibition*

There are two major isoforms of the angiotensin I1 receptor, type 1 (AT1) and type 2 (AT,). AT2 is distin- guished from AT1 with respect to its ligand selectivity, its insensitivity to non-hydrolyzable GTP analogues, and its as yet unidentified biological functions. In the present study we have expression-cloned AT, cDNAfrom a cDNA library of a rat pheochromocytoma cell line (PC12w). Rat AT, cDNA encodes a 363-amino acid protein that has seven transmembrane domains. AT1 is the closest in homology to AT, but with only a 32% identity of amino acid sequence. Stably expressed in COS-7 cells, the receptor showed selective binding to AT,-specific li-gands PD123319 and CGP42112A but not to the AT1-spe-cific ligand, losartan. Northern blot analysis revealed that the mRNA of rat AT, was expressed not only in PC12w cells but also in the adrenal glands and in the inferior olive of the brain, both of which are known to contain AT, type binding sites. The expressed AT, recep- tor mediated angiotensin 11-induced inhibition of protein tyrosine phosphatase, an action that was dependent on a pertussis toxin-sensitive G-protein-coupled mecha- nism in COS-7 cells. The AT,-specific ligand CGP42112A an agonist rather than antagonist in the inhibition of phosphotyrosine phosphatase. AT, did not cause a decrease in cGMP in PC12w or COS-7 cells expressing AT2 stably. These results indicate that the AT, receptor is structurally and functionally different from AT1 and suggest novel functional roles of the renin-angiotensin system in cross-talk with phosphotyrosine signaling by modulating protein phosphotyrosine levels.

There are two major isoforms of the angiotensin I1 receptor, type 1 (AT1) and type 2 (AT,). AT2 is distinguished from AT1 with respect to its ligand selectivity, its insensitivity to non-hydrolyzable GTP analogues, and its as yet unidentified biological functions. In the present study we have expression-cloned AT, cDNAfrom a cDNA library of a rat pheochromocytoma cell line (PC12w). Rat AT, cDNA encodes a 363-amino acid protein that has seven transmembrane domains. AT1 is the closest in homology to AT, but with only a 32% identity of amino acid sequence. Stably expressed in COS-7 cells, the receptor showed selective binding to AT,-specific ligands PD123319 and CGP42112A but not to the AT1-specific ligand, losartan. Northern blot analysis revealed that the mRNA of rat AT, was expressed not only in PC12w cells but also in the adrenal glands and in the inferior olive of the brain, both of which are known to contain AT, type binding sites. The expressed AT, receptor mediated angiotensin 11-induced inhibition of protein tyrosine phosphatase, an action that was dependent on a pertussis toxin-sensitive G-protein-coupled mechanism in COS-7 cells. The AT,-specific ligand CGP42112A was an agonist rather than antagonist in the inhibition of phosphotyrosine phosphatase. AT, did not cause a decrease in cGMP in PC12w or COS-7 cells expressing AT2 stably. These results indicate that the AT, receptor is structurally and functionally different from AT1 and suggest novel functional roles of the renin-angiotensin system in cross-talk with phosphotyrosine signaling by modulating protein phosphotyrosine levels.
Angl I1 plays an important role in cardiovascular regulation, fluid volume homeostasis neuroendocrine regulation, and cel-*This work was supported by Research Grants HL35323 and HL14192 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
charges. This article must therefore be hereby marked "aduertisement" The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTMIEMBL Data Bank with accession number(s) 016840.  (1)(2)(3). The highly diverse nature of its functions suggests the presence of multiple isoforms of its receptor. Recent pharmacological studies with isoform-specific antagonists uncovered the presence of two major receptor isoforms, AT1 and AT2 (M). Studies with the AT1-specific antagonist losartan (4) and the cloned and expressed AT1 (7-9) indicated that AT1 mediates many of the biological responses hitherto attributed to Ang I1 (1)(2)(3). However, little is known on the structure and the signaling mechanism of the second major receptor isoform, AT2. Its abundance in the mesenchymal tissues of a developing fetus (10, 11), uterus (161, adrenal medulla (41, pheochromocytoma (12), and specific brain regions (13,14) suggests a hitherto unidentified role of the type 2 receptor, AT2, in neuronal and developmental function. Here we report cloning of the AT2 cDNA from a rat pheochromocytoma cell line, deduction of its amino acid sequence, and observation of a novel signaling mechanism, by which Ang I1 inhibits the protein phosphotyrosine phosphatase (PTP) activity in COS-7 cells permanently transfected with a plasmid harboring the AT2 cDNA and in rat pheochromocytoma PC12w cell membranes.

MATERIALS AND METHODS
cDNA libraries were prepared from poly(A)+ RNA (5 pg) isolated from rat PC12w cells. After ligating the cDNA with a BstXI adaptor, size fractionation by ultracentrifugation for 3 h at 100,000 x g in a concentration gradient of potassium acetate (5-20%), and insertion into pCD-NAI (Invitrogen), the library was used for transforming Escherichia coli (MC 1061/p3) by electroporation in a Gene Pulser (Bio-Rad). Plasmid DNA(5 pg) from pools of about 300 independent clones grown in a 20-ml liquid culture were used to transfect lo6 COS-7 cells by electroporation.
The cells were cultured on a plastic Cell DiskTM (Sumitomo) for 48-72 h, then incubated with 0.3 n M lZ5I-CGP42112A in Dulbecco's modified Eagle's medium containing 0.2% BSA for 2 h at ambient temperature and autoradiographed with Kodak X-OMAT x-ray film for 5 days at -70 "C. From one positive pool, a single positive clone was selected by the method described elsewhere (7).
The base sequence of the cDNA insert was determined with a Sac1 and XbaI fragment of the AT2 cDNA after subcloning it into pBluescript Ks(+). Truncated cDNAs prepared with Erase A Base SystemTM (Promega) were sequenced in both sense and antisense directions with Sequenase version 2.0 (U. S. Biochemical Corp.) by the dideoxynucleotide chain termination method (15). COS-7 cells stably expressing the type 2 receptor AT2 were prepared by transfecting it by electroporation with pRC/CMV (Invitrogen) containing a 2.9-kilobase insert derived from the AT2 cDNA by BstXl, followed by selection with 0.3 mg/ml G418 (Life Technologies, Inc.). For the determination of ligand binding, cells were incubated with 0.5 n M '251-sarile in Dulbecco's modified Eagle's medium containing 0.2% BSA (w/v) for 2 h at ambient temperature, washed with ice-cold Hanks' balanced salt solution, and the bound radioactivity was counted. For ligand binding to membrane receptors, membranes were prepared by homogenizing cells in a 20 m M HEPES buffer (pH 7.4) containing 14 m M NaCl, 0.5 m~ MgCl,, 10 pg/ml aprotinin, and 0.2% BSA (buffer A) and incubated with 1 n~ lZ5I-Ang I1 for 2 h with or without 10 PM GTPyS in buffer A at ambient temperature. After filtration over a Whatman GF/C filter, the bound radioactivity was counted. Nonspecific binding was determined with 1 PM unlabeled compounds.
The PTP activity of plasma membranes was determined as release of 32Pi from 32P-phosphorylated RaytideTM (Oncogene Science) in 15 min at 37 "C, according to Streuli et al. (16) after preincubation with or without Ang I1 for 30 min at an ambient temperature in buffer A containing 25 p~ GTP. p-Nitrophenyl phosphate (pNPP) was also used as an alternative substrate for the determination of the PTP activity (10). sine phosphatase; AT,, type 1 angiotensin I1 receptor; AT2, type 2 angiotensin I1 receptor.

A Novel Angiotensin 11 Receptor
Cyclic GMP was determined by radioimmunoassay as described previously (18). Cross-linking of 1251-sarile to the Ang I1 receptor was performed with BS3 as published (19).
For Northern blot analysis, 2 pg of poly(A)+ RNA (1 pg from fetal tissues) were separated in 1% agarose gel. The RNA was transferred to GeneScreenm (DuPont NEN), and the filters were hybridized with a labeled ApaI-XbaI fragment of AT2 cDNA in 80 m~ sodium phosphate buffer (pH 7.0) containing 40% formamide, 5 x SSC, 2 x Denhardt's solution, 0.8% SDS, 10% dextran sulfate, and 100 pg/ml denatured salmon sperm DNA for 20 h a t 42 "C. The filters were washed with 2 x SSC, 1% SDS a t room temperature twice and then shaken at 55 "C with 0.2 x SSC, 0.1% SDS for 30 min twice.
CGP42112A was obtained from Neosystem (Strasbourg, France); losartan was a giR from DuPont-Merck; PD123319 was from Warner-Lambert-Parke Davis; peptides were purchased from Peninsula Laboratories, and BS3 from Pierce Chemical Co. Other chemicals were from Sigma; [a-32PldCTP (3000 Ci/mmol) was purchased from DuPont NEN; '2sI-CGP42112A and 1251-sarile were prepared as described previously

RESULTS AND DISCUSSION
We isolated a cDNA clone for AT2 from a cDNA library (45,000 clones) derived from PC12w cells using the mammalian expression vector pCDNAI. One hundred fifty pools of about 300 independent clones were transfected into COS-7 cells, which had no detectable endogenous Ang I1 binding sites, and the cells were screened for the expression of the AT2 receptor activity by autoradiography using as a probe 1251-labeled CGP42112A, a potent AT2-specific ligand. Fig. la shows the complete sequence of rat AT2 cDNA comprising 2,868 nucleotides. The initiation codon (positions 140-142) was preceded by CAAT (Kozak's rule (2011, and the open reading frame was terminated by an in-frame termination codon (positions 1,228-1,230). The base sequence encoded 363 amino acid residues corresponding to a theoretical molecular weight of 41,303. The COS-7 cells transfected with the AT2 clone produced a ligand binding protein of -80 kDa as determined by SDS-polyacrylamide gel electrophoresis and autoradiography of the receptor cross-linked with 1251-sarile using BS3. The molecular mass of the expressed receptor was reduced to around 40 kDa upon treatment with N-glycosidase (data not shown). The size of the deglycosylated protein is in good agreement with the length of the open reading frame of the cDNA. A hydropathy analysis revealed that there are seven hydrophobic domains without a signal sequence. Notable features of the receptor molecule are: five potential N-glycosylation sites located exclusively in the N-terminal hydrophilic domain, the presence of the sequence Asp141-Arg142-Tyr143 highly conserved among seven-transmembrane domain receptors in the N-terminal region of the second cytosolic loop, and a short third cytosolic loop (Fig. 1). Homology search using GenBank and Swiss Protein data bases revealed that AT2 is a new protein and that AT2 has only 32% amino acid identity with ratATl (7,9) and little homology with the mas oncogene product (21,221, the former being the closest related molecule recorded. This contrasts with the 96% identity between the AT1 subtypes, ATl* and AT~B (23)(24)(25)(26). Thus, AT2 seems to be a novel and discrete member of the seven-transmembrane domain receptor superfamily.
This feature of the cloned AT2 is in good agreement with that reported for the AT2 receptor in PC12w cells (12).
Consistent with the previous binding studies (4,6,13,14,(27)(28)(29), preincubation with GTPyS did not reduce the binding activity of the cloned AT2 to lZ5I-Ang I1 (Fig. 212). A similar observation was reported for the seven-transmembrane domain dopamine D3 receptor (31). Thus, the lack of the effect of GTPyS on the binding activity, though unusual, does not necessarily indicate the absence of G-protein coupling as further discussed below.
No biological functions for the type 2 receptor AT2 have been rigorously established. Although it is generally agreed that AT2 does not elevate cytosolic [Ca2+l as does AT1 (28,29,321, controversy exists as to whether the stimulation of AT2 by Ang I1 results in a reduction in intracellular cGMP and an increase in the PTP activity (29,30,(32)(33)(34)(35)(36). We addressed these questions using COS-7 cells stably transfected with AT2 cDNA and PC12w cells. Ang I1 M) did not alter cGMP in these cells with or without pretreatment with 100 IMI ANF (Table I)  in a concentration-dependent manner (Fig. 3u) with a K, of around 0.1 m, which agrees well with the binding affinity for Ang I1 (Fig. 2b). Bottari and his associates (35)  ATz. Possible isoforms of ATz could explain these discrepancies.
Besides Ang I1 and Ang 111, the ATz-specific ligand CGP42112A M) was found to inhibit the PTP activity by 31.2 f 4.3% (n = lo), 14.9 2 2.8% (n = 81, and 8.1 4.9% (n = 10) in the plasma membranes of PC12w cells, rat fetal skin (16 embryonic days), and COS-7 cells stably transfected with the ATz cDNA, respectively. Thus, CGP42112A, originally thought to be an ATz antagonist, can now be considered as an ATz agonist in agreement with recent reports (36). The inhibition of PTP by Ang I1 was further confirmed with another PTP sub-stratepNPP (Fig. 3c). The inhibition by Ang I1 was abolished by PD123319, an ATz-specific antagonist, but not by Dup 753, an AT,-specific antagonist, indicating that this inhibition is mediated by an ATz isoform of the Ang I1 receptor. Most (80-81%) of the PTP activity was inhibited by a vanadate/molybdate mixture (100 p~ each). The inhibition by Ang I1 was limited to the vanadatelmolybdate-sensitive phosphatase activity. Besides the permanent transfectants (COS-71, a similar inhibition of the PTP activity by Ang I1 was seen in PC12w cells (Fig. 3d). This observation indicates that the results obtained with the permanent transfectants were neither an artifact arising from genetic manipulation nor a peculiarity of COS-7 cells. To determine the mechanism of the inhibition of PTP, we investigated the effect of pertussis toxin. The Ang 11-induced inhibition of the PTP activity was eliminated by the pretreatment of the membrane preparation with pertussis toxin (10 pg/ml) in the presence of NAD (10 PM), as shown by solid burs in Fig. 3b, indicating that ATz regulates the PTP activity through a Gprotein (Go or Gi)-mediated mechanism. Although the pertussis toxin sensitivity and the absence of GTPyS-induced shift to a low affinity form appear contradictory to each other, a similar observation was reported with the dopamine D3 receptor as discussed above (30).
The inhibition of the PTP activity was only partial even at a saturating concentration of Ang 11. It appears that the AT2- heart ( H t ) , kidney (Kid), liver (Liu). b, comparison of poly(A)+ RNA from the skin and tongue of fetal and neonatal rats. c, cornparison of poly(A)* RNA from whole brain of E -, 17-, 19-, and 21-day-old fetal and 2-daysld neonatal rats. Molecular size standards in kilobases are indicated to the left. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a control probe. E denotes embryonic days; d denotes days after birth; and w denotes weeks after birth. mediated inhibition may be limited to a certain specific PTP(s). Thus, the second messenger of this system is a PTP(s) that can be inhibited by a pertussis toxin-sensitive G-protein. Recently activation of PTP by somatostatin (17) and dopamine (37) through their seven-transmembrane domain receptors has been reported. I t appears that both the activation and inhibition of some PTPs are regulated by the receptor-linked G-proteins. The positive and negative regulation of PTP can provide a possible means for cross-talk between signaling pathways mediated by phosphotyrosine and the G-protein-mediated seven-transmembrane domain receptors.
As shown by Northern blots in Fig. 4u, AT2 mRNA is expressed in various tissues such as the adrenal medulla and inferior olive of the brain (Fig. 4u). The specific localization of AT2 mRNA is consistent with in situ binding studies by autoradiography (4,10,13,14). Furthermore, prominent expression of AT2 mRNA in rat fetal mesenchymal tissues, such as skin and tongue (Fig. 4b), is interesting in view of the postulated developmental roles of Ang I1 (10). The fetal brain also expresses AT2 mRNA (Fig. 4c). However, it is not yet clear how the PTP inhibition by AT2 is related to physiological regulation of various tissues.