The Insulin Receptor-related Receptor TISSUE EXPRESSION, LIGAND BINDING SPECIFICITY, AND SIGNALING CAPABILITIES*

anti-phosphotyrosine immunoprecipitates

Insulin receptor-related receptor (IRR)' is a member of a family of receptors that includes the insulin receptor (IR) and the insulin-like growth factor I receptor (1)(2)(3)(4). A genomic clone of IRR was initially obtained by Shier and Watt (1) by a low stringency screen of a genomic library using human insulin receptor cDNA as a probe. The nucleotide sequence of the IRR gene predicts that it encodes a novel receptor that is homologous to IR and IGF-I receptor (1). All three receptors have similar predicted structures with a single transmembrane domain, an extracellular domain with a cysteine-rich region, and a cytoplasmic domain with an intrinsic tyrosine kinase activity. IRR, like the other two receptors, was also predicted to be synthesized as a single polypeptide, which is then cleaved to yield an CY and / 3 chain, with the (Y chain being completely extracellular and the p chain having both an extracellular and an intracellular domain. The predicted amino acid sequence of IRR is approximately 60% identical to that of IR and IGF-I receptor with the tyrosine kinase domains being most conserved and the carboxyl tails of the receptors showing the lowest sequence identity (1). The genes for all three receptors map to different chromosomes (5).
To begin to study the physiological role(s) for IRR, it is necessary to determine which tissues express this receptor, what its ligand specificity is, and what types of signals it is capable of mediating. In the present studies we have determined the expression of IRR mRNA in different tissues by Northern analyses and obtained IRR cDNA using PCR. In order to define the ligand specificity of IRR, we have constructed a chimeric receptor with the extracellular domain of IRR fused to the transmembrane and cytoplasmic domains of IR (IRR/IRK). This construct was utilized to study the binding of insulin-like molecules to the chimera. We have also constructed a reciprocal chimeric receptor with the cytoplasmic domain of IR replaced with that of IRR (IR/IRRK). This chimeric receptor was utilized to study the substrate specificity of the tyrosine kinase domain of IRR and to define its signaling capabilities. The present results indicate that IRR is expressed in several tissues, including human kidney, heart, skeletal muscle, pancreas, and liver, does not bind any of the known insulin-like molecules, and is similar to the IR in tyrosine kinase activity and ability to mediate various biological responses.
Construction of cDNA Clones Encoding Chimeric ZRRIIRK or ZR/ ZRRK-To construct IRR/IRK, two cDNA fragments (1.4 kb and 0.8 kb) with an overlapping BgZII restriction site were obtained by PCR using QUICK-CloneTM fetal human kidney cDNA (Clontech Laboratories, Inc.) as a template. The reaction consisted of 25 cycles of denaturation at 94 "C for 1 min, annealing at 55-62 "C for 1 min, and polymerization at 72 "C for 1-3 min. The forward/reverse primers are: 5'-CCCGCCAGGCACCTACCAGTAT-3'/GACGTCACCTTC-CAAGGGGATA and 5'-ACTACTGCCACCGCGGCTTG-3'/GG& TATTCCCAGCATCCTCCTCCTCTG (underline denotes an introduced SspI restriction site). These two fragments were cloned using TA CloningTM system (Invitrogen) and then joined at the BgZII site to generate a 2-kb cDNA fragment. This fragment was then digested with EagI and SspI and ligated into SRa-IRRE2,3/IR (7) partially digested with the same enzymes. The resulting construct encodes the extracellular domain of IRR (residues 7-892) fused to the transmembrane and cytoplasmic domain of IR.
To construct IR/IRRK, a 1040-kb cDNA fragment encoding the cytoplasmic domain of IRR (residues 925-1271) was obtained by PCR using a pool of cDNA generated from a human kidney cDNA library (a gift of Dr. Graeme Bell) as a template. The primers are: 5'-AGATCAGTGCCCTGGACCCCCAT (containing an introduced XbaI site). The PCR products were digested with XbaI and BgZI and subcloned into pECE-IR partially digested with the compatible enzymes. The resulting construct encodes a chimeric receptor with the extracellular and transmembrane domains from IR and the cytoplasmic domain from IRR. DNA sequencing was carried out to confirm the identity of the PCR products and to ensure the correct junctions of the chimeric constructs.
Transient and Stable Expression of the Wild-type IR and Chimeric Receptors-COS-7 cells were maintained and transfected with constructs encoding wild-type or chimeric receptors using the calcium phosphate coprecipitation method as described (8). To establish CHO cell lines stably overexpressing chimeric receptors, cells (5 X 10s/lOO mm plate) were co-transfected with plasmid DNA (10 pg of each chimeric construct plus 50 ng of pSVneo) using calcium phosphate co-precipitation as described above. After 48 h, Geneticin (400 pg/ ml) was added to the culture medium to select for neomycin-resistant cells. Cells expressing high level of IR/IRRK (CHO-IR/IRRK) were identified by '261-insulin binding (see below). Cells overexpressing IRR/IRK (CHO-IRR/IRK) were identified by binding of '2SI-labeled monoclonal antibody to the cytoplasmic domain of IR (17A3) (see below).

GGCAGCCAGATGGGCTGTATGCTTCTGTGAAT-3'/CGAm
In Vivo Activation of Tyrosine Kinase Activities of Chimeric Receptors-Confluent CHO-IRR/IRK or CHO-IR/IRRK cells in 6-well dishes were washed and incubated with 0.5 ml serum-free medium with or without 0.5 mM sodium orthovanadate for 4 h at 37 "C. The CHO-IRR/IRK cells were then incubated with either HEPES-buffered saline, insulin, IGF-I, IGF-11, relaxin, bombyxins I1 and IV (gifts of Dr. Suzuki), molluscan insulin-like peptide (a gift of Dr. Smit), or human growth hormone (all at final M) at 37 "C for 30 min. The cells were washed twice with phosphate-buffered saline, lysed in the lysis buffer as described above, and added to wells precoated with monoclonal antibody 29B4. After an overnight incubation, the wells were washed three times with WGBT, and 20 p1 of the kinase reaction mixture (50 mM HEPES, pH 7.6, 150 mM NaCl, 5 mM MgC12, 5 mM MnC12, 0.1% Triton X-100, 1 mg/ml poly(Glu:Tyr)(4:1), 2 pCi of carrier-free [Y-~'P]ATP) was added to each well, and the incubation was continued at 25 'C for 30 min. The reactions were terminated by spotting 10 pl of the reaction mixture on a Whatman No. 3 MM filter strip. After air drying, the filter strips were soaked in ice-cold 10% trichloroacetic acid containing 10 mM sodium pyrophosphate (5 ml/ 1 X 3-cm strip) for 30 min, boiled in 5% trichloroacetic acid for 5 min, washed twice with 95% ethanol and once with acetone, and the radioactivity was determined by Cerenkov counting. For the activation of CHO-IR/IRRK, the cells were incubated with different concentrations of insulin at 37 "C for 10 min, lysates were prepared and tyrosine kinase activities were determined after immunoimmobilization of the receptor on microtiter wells precoated with monoclonal antibody 83-7 as described above.
Insulin Stimulation of Thymidine Incorporation-CHO, CHO.T, and CHO-IR/IRRK cells were plated in 24-well dishes and grown until 80% confluent. The cells were incubated at 37 "C first with serum-free Ham's F-12 with 20 mM HEPES, pH 7.4 for 36 h, and then with 0.5 ml of fresh serum-free medium (Ham's F-12, 20 mM HEPES, pH 7.4, 1 mg/ml bovine serum albumin) and various concentrations of insulin for 8 h.
[3H]Methylthymidine (0.75 pCi/well) was then added, and the incubation was continued at 37 "C for 45 min. The cells were washed twice with HEPES-buffered saline, 0.3 mM CaClz and lysed with 0.075% SDS. The lysates were precipitated with 10% trichloroacetic acid at 4 "C overnight. The precipitates were collected on glass fiber filter circles, washed twice with 5% trichloroacetic acid, and counted in BioSafe-I1 fluid in a scintillation counter.
Insulin Stimulation of Z-Deoxy[l -'4ClGlucose Uptake-Confluent CHO, CHO.T, and CHO-IR/IRRK cells in 24-well plates were treated with 10 mM deoxyglucose in complete medium at 37 "C for 5 h to lower the basal level of glucose uptake (11). The cells were then washed twice with a buffer containing 140 mM NaC1, 2.7 mM KC1, 0.9 mM CaCl,, 1.47 mM potassium phosphate, 8.06 mM sodium phosphate, pH 7.4, 0.46 mM MgCl,, and 1 mg/ml bovine serum albumin. After incubation at 37 "C for 30 min with the same buffer in the presence or absence of lo-@ M insulin, the cells were incubated with 0.1 mM deoxyglucose and 0.2 pCi/well of 2-deo~y-D-[1,2-~H]glucose for 10 min. Cells were then washed three times with the buffer containing 200 pM phloretin and lysed in 0.075% SDS. Aliquots of the lysates were counted and assayed for protein by the Coomassie dye binding assay (Bio-Rad). In each experiment, the amounts of deoxyglucose taken up by the different cell lines were normalized against their protein concentration as compared with CHO cells.
Endogenous Substrates of IRIIRRK-CHO, CHO.T, and CHO-IR/ IRRK cells were incubated in serum-free medium in the presence or absence of 0.5 mM sodium orthovanadate at 37 "C for 4 h. The cells were then either treated with buffer or lo" M insulin for 10 min.
After three washes with phosphate-buffered saline, the cells were lysed in the lysis buffer described above. The lysates were subjected to 10% SDS-polyacrylamide gel electrophoresis followed by Western transfer to nitrocellulose filters. The filters were incubated with either polyclonal anti-phosphotyrosine antibodies, or a monoclonal antibody directed against the a subunit of IR (3Bll) (a gift of Dr. Kozui Shii). Bound antibody was detected using alkaline phosphatase-conjugated anti-immunglobulin (Promega Biotech).
Phosphatidylinositol (PtdZns) 3-Kinase Assays-CHO, CHO.T, and CHO-IR/IRRK cells were treated with increasing concentrations of insulin and lysed in the lysis buffer. The lysates were subjected to immunoprecipitation with either normal mouse IgG, monoclonal antiphosphotyrosine antibody (PY20) (12) (a gift of Dr. John Glenney), or monoclonal anti-IR antibody (5D9). The PtdIns 3-kinase activity was measured directly in the immunoprecipitates as described previously (13).

Properties of the Insulin Receptor-reluted Receptor
PLCy-CHO, CHO.T, and CHO-IR/IRRK cells were treated with either buffer, M insulin, 0.5 mM sodium orthovanadate, or the combination of insulin plus vanadate. The lysates were immunoprecipitated with either polyclonal anti-p85 antibodies (UBI), monoclonal anti-PLCy (UBI), or monoclonal anti-GAP (B4F8) (Santa Cruz Biotech). Control immunoprecipitation was carried out using either rabbit IgG or mouse IgG to match the specific antibody. The immunoprecipitates were then separated with 10% SDS-polyacrylamide gel electrophoresis followed by Western blotting. The blots were incubated with polyclonal anti-phosphotyrosine antibodies. The control blots with anti-p85, anti-GAP, and anti-PLCy immunoprecipitates were incubated with their corresponding antibodies. The bound antibody was detected as described above.

RESULTS
Tissue Expression of IRR-In order to identify tissues expressing IRR, poly(A)-selected mRNAs from human heart, liver, skeletal muscle, kidney, pancreas, brain, lung, and placenta were probed in a Northern blot with an antisense riboprobe directed against exon 3 of human IRR (Fig. 1). Kidney, heart, skeletal muscle, liver, and pancreas were all found to give a positive signal with a band at 5.8 kb. Similar results were obtained with an antisense riboprobe to exon 2 of human IRR (data not shown). To further test whether IRR was expressed in kidney, PCR was performed on cDNA obtained from reverse-transcribed human fetal kidney mRNA and a pool of cDNA generated from an adult human kidney cDNA library. PCR products of the appropriate size for the primers used were generated from both sources. The cDNA encoding the extracellular domain of IRR was cloned from the PCR products with the fetal cDNA, and the cDNA encoding the intracellular domain of IRR was obtained from the PCR products with the kidney library. The nucleotide sequences of these cloned PCR products were identical to those of IRR (1).
Ligand Binding Specificity of the Extracellular Domain of IRR-In order to identify the ligand recognized by IRR, we constructed a cDNA that encodes a chimeric receptor, IRR/ IRK, which contains the extracellular domain of IRR (residues 7-892) on the backbone of the IR (Fig. 2). We first determined the ability of the expressed chimeric receptor to bind to several '9-labeled ligands. Lysates of COS-7 cells transfected with vector or the cDNAs encoding either wildtype IR or IRR/IRK were incubated in microtiter wells coated with a monoclonal antibody to the cytoplasmic domain of IR (29B4). After washing the wells, either '251-labeled insulin, IGF-I, IGF-11, or a monoclonal antibody to a distinct epitope on the cytoplasmic domain of IR (17A3) was added, and the radioactivity bound by the immobilized receptors was determined. As indicated by lZ5I-17A3 binding, similar amounts of wild-type IR and IRR/IRK were expressed in COS-7 cells and bound to the wells (Fig. 3). However, IRR/IRK did not bind significantly greater amounts of either insulin, IGF-I, or IGF-I1 than the vector control ( Fig. 3), suggesting that IRR is not the receptor for any of these ligands. The control transfectants with the native IR bound, as expected, much greater amounts of insulin than the control transfectants.
To further test the ability of these as well as other ligands to bind to IRR/IRK, we examined the ability of various ligands to stimulate the intrinsic tyrosine kinase activity of the chimeric receptor in CHO cells stably overexpressing it. This cell line was demonstrated to overexpress the processed chimeric receptor by immunoblotting in amounts comparable to the overexpressed human IR in CH0.T cells (data not shown). Insulin, IGF-I, and IGF-I1 (all at 100 nM) were all found incapable of activating the intrinsic tyrosine kinase activity of the chimera (Fig. 4). The only other known member of the vertebrate insulin family, relaxin (14), was also incapable of activating the kinase activity of the chimera (Fig. 4). Several invertebrate insulin-like molecules, including bombyxins I1 and IV (15) and molluscan insulin-like peptide (16), as well as an unrelated hormone, human growth hormone, were similarly found to be incapable of activating the tyrosine kinase activity of IRR/IRK (Fig. 4). To test for the intactness of this chimera, we examined whether vanadate and wheat germ agglutin, activators of the insulin receptor kinase (17, 18), could stimulate its activity. Both of these agents were Properties of the Insulin Receptor-related Receptor 18323 X92), which has a much higher affinity for the IR than native insulin (lo), was also found to bind to IRRE2,3/IR in amounts found to significantly stimulate the kinase activity of the chimera ( Fig. 4 and data not shown).
In our prior studies, we had observed a small but significant amount of labeled insulin binding to a chimeric receptor (IRRE2, 3/IR) that only had the amino-terminal 450 amino acid residues of IRR on the backbone of the insulin receptor. We therefore compared the binding of labeled insulin to the two chimeric receptors, IRRE2,3/IR and IRR/IRK. 1251-Labeled insulin was incubated in the presence or absence of a large excess of unlabeled insulin with similar amounts of the two chimeric receptors (as assessed by antibody binding) captured on microtiter wells. IRR/IRK exhibited no significant increase in specific insulin binding in comparison to that for an untransfected control (Fig. 5). However, IRRE2,3/IR was able to bind significantly more insulin than this control ( approximately 7-8 times the amount specifically bound by the control lysates) (Fig. 5). The specific binding of insulin to IRRE2,3/IR was inhibited by a human specific monoclonal antibody to the insulin receptor (5D9) (Fig. 5). This monoclonal antibody has previously been shown to recognize an epitope present in the chimera IRRE2,3/IR ( 7 ) but does not recognize the endogenous insulin and IGF-I receptors in CHO cells (19). Furthermore, a labeled analog of insulin (called approximately 5-6 times greater than that observed bound to the control lysates. This analog dissociated much more rapidly from IRRE2,3/IR than from native IR, indicating that the chimeric receptor IRRE2,3/IR can bind insulin although with a much weaker affinity than the native IR. Tyrosine Kinase Activity of the Cytoplasmic Domain of IRR-Since the ligand that activates the intrinsic kinase activity of IRR was not available, we constructed a cDNA that encodes a chimera, IR/IRRK, with the extracellular and transmembrane domains of IR and the intracellular domain of IRR (Fig. 2). Stable transfectants of CHO cells that overexpress this chimera were selected. These cells were shown to produce the mature chimera by Western blotting in amounts comparable with CH0.T cells that overexpress native IR (see below). This chimera was also able to bind to insulin with an affinity indistinguishable from that for the wild-type IR (Fig.  6). CHO, CHO.T, and CHO-IR/IRRK cells were treated with different concentrations of insulin and the tyrosine kinase activities of the receptors from these cells towards exogenous substrate (poly(G1u:Tyr) (4:l)) were determined. As shown in Fig. 7 , in vivo insulin treatment resulted in a dose-dependent activation of the tyrosine kinase activity of IR/IRRK, which closely followed the activation observed with native IR and was clearly above that observed in the parental CHO cells. However, the maximum kinase activity for IR/IRRK was about one-tenth that observed for wild-type IR, although approximately the same amount of the two receptors were adsorbed to the wells (Fig. 7 ) . Similar results were obtained with a second independent clone of CHO cells expressing the chimera (Fig. 7 B ) . This lower kinase activity with the chimera was also observed in vitro with two other substrates (poly(Glu,Tyr) (1:l) and histone) (data not shown). The activation of the tyrosine kinase activity of the chimera, like the wild-type IR, could also be accomplished in vitro by autophosphorylation (data not shown). As with the in vivo activated receptor, the in vitro activated chimera had approximately one-tenth the kinase activity of the native IR.
To assess the ability of the IRR kinase domain to phosphorylate endogenous substrates, two lines of CHO-IR/IRRK, CHO, and CH0.T cells were treated with or without insulin in the presence or absence of vanadate, and their lysates were examined for tyrosine phosphorylated proteins by immunoblotting with antibodies to phosphotyrosine (Fig. 8A). In the  were treated with the indicated concentrations of insulin for 10 min at 37 C. Cells were then lysed in the presence of phosphatase inhibitors and receptors were captured on microtiter wells coated with monoclonal antibody against IR. Receptor tyrosine kinase activities were determined as the radioactivity incorporated into the substrate poly(Glu,Tyr). In parallel wells, the amount of receptor captured was assessed by labeled antibody binding. In the experiment shown, 6652, 6800, and 6932 cpm of labeled antibody were bound for IR, IR/IRRK-1, and IR/IRRK-2, respectively. Results shown are representative of three independent experiments. CHO-IR/IRRK cells, insulin alone stimulated the tyrosine phosphorylation of a protein of 85 kDa, a value consistent with that expected for the 6 subunit of this receptor, and one of approximately 170 kDa, a value consistent with that expected of a substrate called IRS-1 (20). In the presence of insulin and vanadate, a large number of tyrosine phosphorylated proteins were observed in these transfected cells that were not observed in the control parental CHO cells. The pattern and amount of tyrosine-phosphorylated proteins were similar to those observed in the CH0.T cells overexpressing native IR with the exception of a more prominent 95-kDa band present in the CH0.T cells. This 95-kDa band presumably represents the / 3 subunit of the IR (the 6 subunit of IR is predicted from its sequence to be larger than that of IRR). A control immunoblot with a monoclonal antibody to the a subunit of the insulin receptor demonstrated that comparable levels of receptor were expressed in the different transfected cells (Fig. 8B).
One of the apparent endogenous substrates for the insulin and IGF-I receptors is the PtdIns 3-kinase in that insulin and IGF-I stimulate an increase in the amount of anti-phosphotyrosine-precipitable PtdIns 3-kinase activity (13,(21)(22)(23). We therefore tested whether insulin could also stimulate an increase in anti-phosphotyrosine-precipitable PtdIns 3-kinase activity in CHO-IR/IRRK cells. CHO, CHO.T, and CHO-IR/ IRRK cells were treated in uiuo with insulin and lysed, and the PtdIns 3-kinase activities were determined in anti-phosphotyrosine immunoprecipitates prepared from these cells. As V a n a d a t e -+ --+ + --+ + --+ + shown in Fig. 9, insulin was able to stimulate the amount of PtdIns 3-kinase activity in anti-phosphotyrosine precipitates from CHO-IR/IRRK cells in a dose-dependent manner. The maximal activity precipitated from these cells was 15-fold above that observed in CHO cells, and it was about 50% of that observed in CH0.T cells.
To determine whether the PtdIns 3-kinase was immunoprecipitated with the anti-phosphotyrosine antibodies because of a stable association with the overexpressed receptors, the cell lysates of insulin-treated cells were precipitated with either control IgG, anti-IR antibody 5D9, or the anti-phosphotyrosine antibody PY20. PtdIns 3-kinase activities in the immunoprecipitates were assayed. In both CHO-IR/IRRK and CH0.T cells, less than 2% of the total PtdIns 3-kinase activity (determined in the PY20 immunoprecipitates) is associated with the receptors (determined in the 5D9 immunoprecipitates) (Fig. 10). Immunoblots demonstrated that similar amounts of receptor were precipitated with the anti-IR and anti-phosphotyrosine antibodies (data not shown).
To determine whether insulin stimulated the tyrosine phosphorylation of the PtdIns 3-kinase, cells were treated with insulin in the presence of vanadate (to maximize the extent of substrate phosphorylation), the lysates were immunoprecipitated with an antibody to one of the subunits of the PtdIns 3-kinase (p85) (24-27) and the precipitates were analyzed by blotting with anti-phosphotyrosine antibodies. No significant increase in the level of tyrosine phosphorylation of the 85-kDa subunit of PtdIns 3-kinase was observed in either CH0.T or CHO-IR/IRRK cells even in the presence of the tyrosine phosphatase inhibitor vanadate (Fig. 11). A control blot that was probed with the anti-p85 antibody demonstrated that this antibody was capable of precipitating p85 from these cells (Fig. 11).
We also compared the ability of IR and IRR to tyrosinephosphorylate two other substrates of tyrosine kinases, phospholipase Cy (PLCy) and the GTPase-activating protein of Ras, called GAP (24). Cells were treated with insulin, vanadate, or both, the lysates were immunoprecipitated with anti-PLCy or anti-GAP, and the precipitated proteins subjected to anti-phosphotyrosine immunoblotting. The controls were conducted by probing the same immunoblots with antibodies to either PLC-y or GAP, respectively. In both CH0.T and CHO-IR/IRRK cells, insulin plus vanadate treatment resulted in a small amount of tyrosine phosphorylation of a 145-kDa protein that comigrated with PLCy (Fig. 12). This increase in PLCy phosphorylation was specifically mediated by the overexpressed receptors, since no increase was observed in the parental CHO cells. An even smaller increase in the tyrosine phosphorylation of GAP p120 was observed in CH0.T and CHO-IR/IRRK cells (Fig. 13). However, insulin stimulated an increase in tyrosine phosphorylation of one of the GAP-associated proteins, p60, in the transfected cells but not in the control cells ( Fig. 13)  Test for association of PtdIns kinase activity with the wild-type IR and chimeric IR/IRRK. CHO cells overexpressing IR or IR/IRRK were treated with 10" M insulin a t 37 "C for 10 min. The cells were then lysed, and the extracts were absorbed with protein G-agarose beads coated with either normal mouse IgG (NZg), a monoclonal antibody against the extracellular region of IR (509), or a monoclonal antibody to phosphotyrosine (PY20). The absorbed PtdIns kinase activity was assayed using PtdIns as a substrate. Results shown are representative of three independent experiments. IRRK cells were treated with or without insulin plus vanadate. The lysates were subjected to immunoprecipitation with either normal rabbit IgG (NZg) or antibodies to p85 (a p85) and the precipitates were analyzed by SDS-gel electrophoresis followed by immunoblotting with anti-phosphotyrosine antibodies. The positive control blot was probed with antibody to p85, and the position corresponding to p85 is indicated. insulin plus vanadate, the lysates were immunoprecipitated with either normal mouse IgG (NZg) or anti-PLCy antibody (a PLCy), and the precipitated proteins were subjected to anti-phosphotyrosine immunoblotting. For the positive control, the blot was probed with anti-PLCy. The protein band corresponding to PLCy is indicated. of IRR-CHO cells stably overexpressing IR/IRRK were also utilized to see if the intracellular domain of IRR was capable of mediating stimulation of thymidine incorporation and glucose uptake. Two independent cell lines of CHO-IRR/IRK exhibited a higher basal level of thymidine incorporation than Properties of the Insulin Receptor-related Receptor I n s u l i n --+ -+ -+ -+ -+ -+ V a n a d a t e ---+ + --+ + --+ + MW (kDa) with the indicated concentrations of insulin for 8 h.
[3H]Methylthymidine incorporation was determined as described. The results shown are the mean zk S.E. for five experiments, each conducted in triplicate.
either CHO or CH0.T cells (Fig. 14). Insulin also stimulated in a dose-dependent fashion thymidine incorporation in the cells overexpressing IR/IRRK (Fig. 14). This stimulation clearly occurred at concentrations lower than that required to stimulate the parental CHO cells (Fig. 14), indicating that the kinase domain of IRR can mediate this biological response. Insulin also stimulated thymidine incorporation to a higher maximal level in the cells overexpressing the chimeric receptor than in either the parental CHO or CH0.T cells (Fig. 14).
However the maximal increase in stimulation in the cells overexpressing IR (3.5-fold, average of five independent experiments) was about the same as that observed in the two cell lines overexpressing IR/IRRK (3.6-and 3.9-fold).
The two lines of CHO-IR/IRRK cells also exhibited enhanced basal uptake of 2-deoxyglucose in comparison to the parental CHO cells (Fig. 15). Two additional independent clones of cells overexpressing IR/IRRK also exhibited this same elevated level of basal deoxyglucose uptake. As previously reported (ll), the cells overexpressing native IR, CHO.T, also exhibited a higher basal level of deoxyglucose uptake than the parental CHO cells although not to the same extent as that observed in the cells overexpressing IR/IRRK (Fig. 15). Maximal insulin stimulation of deoxyglucose uptake was only approximately 20-30% above these elevated basal levels of uptake for the transfected cell lines expressing these different receptors, although the absolute increment in deoxyglucose uptake with insulin treatment is greater than that observed in the parental CHO cells (Fig. 15). This low level of stimulation precluded the performance of quantitative dose response curves. However, it appeared that insulin at concentrations of less than 1 nM could stimulate glucose uptake in the transfected cells without stimulating glucose uptake in the parental CHO cells (data not shown). These results indicate that IR/IRRK, like the IR, can mediate the stimulation of glucose uptake.

DISCUSSION
In the present studies the mRNA for IRR has been found by Northern blot analyses to be expressed in a variety of human tissues including kidney, heart, skeletal muscle, pancreas, and liver (Fig. 1). It is likely that other tissues that have not been tested will also express the IRR mRNA. The expression of IRR mRNA in kidney was further confirmed, since we were able to obtain properly processed cDNA for IRR by PCR from both fetal and adult kidney. Our success in detecting IRR mRNA in some of the same tissues that were previously considered negative (1) may be due to the enhanced sensitivity of the antisense riboprobes used in the Northern analyses in the present work. It is also possible that the different results in the two studies arise from species specific differences, since the prior work was with rat tissues whereas the present studies were performed with human tissues. Of course, additional studies are required to demonstrate expression of the IRR protein in these tissues and to determine which cell types express IRR within these tissues.
In the present work we have utilized chimeric receptors constructed from insulin receptor and IRR to examine the ligand specificity of the extracellular domain of IRR and the signal transduction capacity of the cytoplasmic domain of IRR (Fig. 2). Extensive prior studies have shown that chimeric receptors can be constructed between the different domains of the IR and other molecules and that these two domains can mediate their ligand binding and signaling abilities independently (29-34). Thus it is likely that the extracellular domain of IRR can also fold independently and should have been capable of binding its proper ligand. The inability of the chimeric receptor with the extracellular domain of IRR to either bind labeled insulin, IGF-I or IGF-I1 (Fig. 3)  exhibit ligand-stimulated kinase activity ( Fig. 4) with any of the insulin-like molecules tested (which included all of the known vertebrate insulin-like molecules as well as several invertebrate insulin-like molecules) indicates that the ligand recognized by IRR is either a presently unknown insulin-like molecule or a known molecule that has not been previously considered insulin-like. With the chimeric IRR/IRK and the assay procedures described in this study, it is now possible to screen for the presence of a ligand for IRR in various biological fluids and in the supernatants of various cell lines. The inability of the chimera IRR/IRK to bind significantly any insulin differs from the low level of insulin binding observed with the chimera IRRE2,3/IR, which has had only the amino-terminal450 residues of the IR replaced with those of IRR (Fig. 5). The binding of insulin to the latter chimera was inhibited by insulin as well as by a monoclonal antibody that is specific for the human IR (Fig. 5). The inhibition by this antibody demonstrates that the binding of insulin by this chimera is not due to endogenous insulin or IGF-I receptors, since this antibody does not recognize either of these receptors (19). These results are consistent with the hypothesis that regions of the IR carboxyl to the cysteine-rich region also can contribute to ligand binding (7,35,36).
The tyrosine kinase domain of IRR shares a high degree of sequence identity (-80%) to that of both the IR and IGF-I receptor (1)(2)(3)(4). The ATP binding site and the regulatory tyrosine autophosphorylation residues (corresponding to residues 960, 1158, 1162, and 1163 of IR) are all conserved, suggesting that the IRR tyrosine kinase could mediate signal transduction via mechanisms similar to those for IR and IGF-I receptor (1)(2)(3)(4). However, a much lower level of sequence homology was found in the COOH-terminal region of the @ subunit of IRR (only 19% identical to IR) and the intracellular domain of IRR is 48 residues shorter than that of the insulin receptor (1)(2)(3)(4). Truncations of the @ subunit of the IR have in some studies been found to affect the signaling capabilities of the IR, although such differences were not found in other studies (37-39). Therefore, it was possible that IRR tyrosine kinase could differ from the IR in its signaling capabilities.
Since the ligand for IRR is unidentified, it is impossible to study the signaling abilities of IRR using the native receptor. The strategy we used to overcome this limitation was to construct a chimeric receptor (IR/IRRK) composed of the extracellular and the transmembrane domains of the insulin receptor and the intracellular domain of IRR. This approach allowed us to study the signaling potential of IRR following the stimulation of the chimeric receptor with insulin. Such chimeric receptors have also been utilized to study the signaling capabilities of other orphan receptors (40,41).
As expected, the IR/IRRK chimera is functional when transiently expressed in COS-7 cells and stably expressed in CHO cells. The chimeric receptor was able to bind insulin with an affinity equivalent to that for the wild-type insulin receptor (Fig. 6). In vivo stimulation of CHO-IR/IRRK cells with insulin resulted in a dose-dependent activation of the tyrosine kinase activity of IR/IRRK toward exogenous substrate (poly(G1u:Tyr) (4:l)) (Fig. 7). The in vitro assays of the tyrosine kinase activities were conducted using substrates and conditions optimized for IR, which may partially account for the approximate 10-fold lower kinase activity of IR/IRRK compared to the native IR (Fig. 7). The kinase activities of the two receptors appeared more comparable in the intact cells. By immunoblotting with anti-phosphotyrosine antibodies, insulin was found to stimulate the tyrosine phosphorylation of approximately the same spectrum of proteins to approximately the same levels in cells overexpressing the native IR and IR/IRRK (Fig. 8). The only major difference in the patterns observed were in the phosphorylation of the @ subunits of the receptors, with the IR @ subunit having a higher molecular weight and a greater extent of phosphorylation than that of IRR. Both of these differences can be accounted for by the lack of 48 amino acids in the COOH terminus of the @ subunit of IRR, which contains two tyrosine autophosphorylation sites in the IR (1)(2)(3)(4).
The abilities of IR and IR/IRRK to phosphorylate several specific endogenous proteins were also analyzed. Insulin treatment of CHO-IR/IRRK, as previously shown for cells overexpressing IR (13,21), resulted in a large increase in antiphosphotyrosine-precipitable PtdIns 3-kinase activity (Fig.  9). Most (greater than 98%) of this PtdIns 3-kinase activity did not appear to be associated with the chimeric receptor, since it was not precipitated with antibodies to the receptor (Fig. 10). This low level of association of the kinase domain of IRR with the PtdIns 3-kinase is identical to the results with the native IR (Ref. 13 and Fig. 10) and the receptor for 42). Surprisingly, the immunoprecipitation of the 85-kDa subunit of PtdIns 3-kinase did not reveal any insulinstimulated increase in tyrosine phosphorylation of this subunit by immunoblotting in cells overexpressing either the native IR or IR/IRRK (Fig. 11). These results suggest that this protein does not get tyrosine-phosphorylated by the IR and IRR kinase domains. In a recent study, the plateletderived growth factor receptor tyrosine kinase was also reported not to mediate the tyrosine phosphorylation of this protein (43). An explanation for the ability of anti-phophosphotyrosine antibodies to precipitate the PtdIns 3-kinase in insulin-treated cells may be that this enzyme becomes associated with another tyrosine-phosphorylated protein. One such candidate would be IRS-1, a 170-180-kDa protein that appears to associate with PtdIns 3-kinase after becoming tyrosine-phosphorylated by the IR (20).
Two other identified substrates for tyrosine kinases are PLCy and GAP (24). Insulin treatment of CHO-IR/IRRK and CH0.T in the presence of vanadate resulted in the same low level of tyrosine phosphorylation of PLCy (Fig. 12). These results are consistent with prior studies indicating that PLCy is not a very good substrate for the IR tyrosine kinase (44). Insulin treatment of both of these cells also resulted in a large amount of tyrosine phosphorylation of a GAP-associated protein of 60 kDa (Fig. 13). This protein has previously been found to be tyrosine-phosphorylated in cells treated with platelet-derived growth factor (28). The important finding in the present studies is that the ability of the kinase domains of IR and IRR to phosphorylate various specific endogenous proteins are quite similar.
The CHO cells overexpressing IR/IRRK exhibited a much higher basal level of thymidine incorporation and 2-deoxyglucose uptake than either the parental CHO cells or the CHO cells overexpressing the native IR (Figs. 14 and 15). Insulin was also found to stimulate thymidine incorporation and glucose uptake in CHO-IR/IRRK at concentrations that do not significantly stimulate responses in the parental CHO cells (Fig. 14 and data not shown). These results indicate that the kinase domain of IRR, like that of the IR, can mediate both of these biological responses. Moreover, the absolute increment in thymidine and glucose uptake in cells overexpressing the kinase domain of IRR is greater than in the cells overexpressing the native IR. However, because of the higher basal levels of thymidine incorporation and glucose uptake in the CHO-IR/IRRK, it is difficult to conclude whether the kinase domain of IRR is more potent than that of IR in stimulating these responses.
In conclusion, the mRNA for IRR appears to be expressed in a variety of human tissues including human heart, skeletal muscle, kidney, liver, and pancreas. IRR also does not appear to be the receptor for any known member of the insulin family, either in vertebrates or invertebrates. Finally, the tyrosine kinase of IRR is similar to that of IR in that it: 1) is activated by autophosphorylation, 2) phosphorylates the same spectrum of exogenous and endogenous substrates, 3) does not tightly associate with PtdIns 3-kinase, and 4) mediates both thymidine incorporation and glucose uptake.