Interaction of Janus Kinases JAK-1 and JAK-2 with the Insulin Receptor and the Insulin-Like Growth Factor-1 Receptor*

Insulin and insulin-like growth factor-1 (IGF-1) treatment of cells overexpressing the insulin receptor or the IGF-1 receptor promotes phosphorylation and activation of Janus kinases JAK-1 and JAK-2 butnotofTYK-2.Withinsulin,weobservedmaximalphosphorylationofJAK-1within2min(5.2 6 0.6-fold) and maximal phosphorylation of JAK-2 within 10 min (2.4 6 0.6-fold). In cells incubated with IGF-1, we found maximal phosphorylation of JAK-2 within 2 min (1.9 6 0.2-fold) and of JAK-1 within 5 min (4.5 6 0.4-fold). The JAKs from insulin- or IGF-1-stimulated cells were activated, as shown by their autophosphorylation in vitro . Moreover, they were able to phosphor- ylate in vitro native insulin receptor substrate (IRS)-1 and a fragment of IRS-2 (GST-IRS-2591–786). Comparison of 32 P-peptide maps of IRS-1 phosphorylated in vitro by the insulin receptor vs. JAK-1 showed the occurrence of different phosphopeptides, suggesting that different sites are likely to be phosphorylated by the two kinases. Finally, coprecipitation of receptors and JAK-1 was seen, and phosphorylation of both receptors was found to be necessary for receptor bindingtoJAK-1.TwodomainsofJAK-1areinvolvedintheformationofthecomplexbetweenreceptorandJAK-1, i.e. the N-terminal por- tion containing JH7 and JH6 and the C-terminal domain (JH1 domain). our data together, we conclude that: and IGF-1 lead phosphorylation and activation of JAK-1 and JAK-2 in intact cells; phosphorylation of IRS-I by JAK-1 on sites phosphorylated by the insulin receptor; JAK-1 interacts with phosphorylated and IGF-1 receptors; and the JH7-JH6 and JH1 domains of JAK-1 are responsible for with and IGF-1 IGF-1 receptors (wild-type or D 121) were incubated with GST, GST-F1, and GST-F2 fusion proteins (60 pmol/sample) preadsorbed on glutathione-sepharose. After 4 h at 4 C, the pellets were washed twice. The presence of unphosphorylated receptors was tested by ligand-induced autophosphorylation (as described in Materials and Methods ). Samples were analyzed by SDS-PAGE under reducing conditions, followed by autoradiography. and JH-1 domains of JAK-1 play a role in the interaction with insulin receptors and with IGF-1 receptors. IGF-1 and insulin receptors, phosphorylated with [ g - 32 P] ATP, were incu- bated with GST, GST-F1–1, GST-F1–2, GST-F2–1, and GST-F2–2 fusion proteins preadsorbed on glutathione-sepharose (90 pmol/sam- ple). After 4 h at 4 C, the pellets were washed twice. Samples were analyzed by SDS-PAGE followed by autoradiography.

The cytokine receptors, which do not possess intrinsic tyrosine kinase activity, stimulate the JAK/STAT pathway. Two domains on the cytoplasmic part of the receptors are responsible for the interaction with JAKs. The first domain is a proline-rich motif (box a) and the second one is an acidic motif (box b) (32,33). The association of JAKs to cytokine receptors can be constitutive or enhanced by ligand binding (34,35). Dimerization of the receptors upon ligand binding induces autophosphorylation and transphosphorylation of JAKs. Importantly, activated JAKs phosphorylate the recep-tors, allowing the recruitment of SH2 domain-containing signal transducers and activators of transcription (STATs), which in turn, are phosphorylated by the JAKs. Phosphorylated STATs form homodimeric or heterodimeric complexes and translocate to the nucleus, where they activate transcription of specific genes. Recently, it has been shown that JAKs activated by cytokines such as interleukin (IL)2, IL4, IL7, IL9, IL15, leukemia inhibitory factor, interferon ␣ and ␥, and GH also can lead to phosphorylation of IRS-1 and IRS-2 (36 -42).
In the present study, we have compared the time courses of JAK-1 and JAK-2 phosphorylation in response to insulin and IGF-1. We have followed the activation of these JAKs by measurement of IRS-1 and IRS-2 phosphorylation in vitro. We also have compared the tryptic maps of IRS-1 phosphorylated by the insulin receptor vs. JAK-1. Finally, we have explored which JAK-1 region(s) is (are) required for interaction with insulin and IGF-1 receptors.

Materials and Methods Materials
Triton X-100, N-acetyl-d-glucosamine, BSA, aprotinin, and protein A-sepharose were from Sigma Chemical Co. (St. Louis, MO). Wheat germ agglutinin (WGA) agarose was from Biomaker (Rehovot, Israel). Polyvinylidene fluoride membrane was from Millipore (Bedford, MA). The monoclonal mouse antiphosphotyrosine antibody was purchased from UBI (Lake Placid, NY). The polyclonal rabbit antibodies to JAK1, JAK-2, and TYK-2 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Antibody used to immunoprecipitate insulin receptors is directed to the receptor extracellular region and was described previously (43). Recombinant human IGF-1 was a gift from Lilly Laboratories (Indianapolis, IN). The polyclonal rabbit antibody to IRS-1 was produced in our laboratory (44).

Partial purification of receptors
IGF-1 and insulin receptors were partially purified by chromatography on WGA (45). Briefly, confluent cells were serum deprived overnight in DMEM containing 0.2% (wt/vol) BSA (10 dishes of 15-cm diameter). Cells were washed twice with PBS and solubilized for 90 min at 4 C in 50 mm HEPES, 150 mm NaCl, 1% (vol/vol) Triton X-100, 10% (vol/vol) glycerol, 1 mm EGTA, pH 7.6. The supernatant from an ultracentrifugation step (60 min, 100,000 ϫ g, 4 C) was applied to a WGA column, and the receptors were eluted in fractions of 500 l with 0.3 m N-acetyl-d-glucosamine in 50 mm HEPES, 150 mm NaCl, 0.1% (vol/vol) Triton X-100, 10% (vol/vol) glycerol, pH 7.6. Protease inhibitors were present throughout the procedure (20 m leupeptin, 1.25 mm bacitracin, 100 U/ml aprotinin, and 1 mm phenylmethylsulfonyl fluoride). The presence of IGF-1 or insulin receptors in each fraction was tested by IGF-1 or insulin induced-autophosphorylation followed by SDS-PAGE. Usually, fractions 2 and 3 were found to contain the receptors and were pooled.

Transfection of 293-EBNA cells.
The full-length rat IRS-1 complementary DNA (cDNA), obtained from M. F. White (Boston, MA), was subcloned into pCEP-4 expression vector (Invitrogen) between XhoI and NheI sites. The XbaI site was introduced 330 bp before the stop codon of the IGF-1 cDNA fragment by PCR. The ⌬121 IGF-1 cDNA was subcloned into pcDNAneo expression vector (Invitrogen) between XhoI and XbaI sites.
The human JAK-2 cDNA subcloned into pBluescript-SK was a gift from J. Ihle (Department of Biochemistry, St. Jude Children's Research Hospital, Memphis, TN). Construction of kinase-dead JAK-2 was performed as described elsewhere (46). Site-directed mutagenesis of W 1020 G and Q 1024 A within the carboxyl-terminal protein-tyrosine kinase domain was accomplished by unique-site elimination (CLONTECH) using the following primer: 5Ј-ggaaagcccaatattcggtacgcacctgcatccttgac-3Ј. The kinase-dead JAK-2 cDNA was subcloned into pcDNA3 expression vector (Invitrogen) between ApaI and NotI sites.
Exponentially growing cells were trypsinized, seeded at 6 ϫ 10 6 cells per 15 cm plate, and incubated overnight in 20 ml growth medium. Ten to 20 g plasmid DNA (IRS-1 or ⌬121 IGF-1 receptor) was mixed with 1 ml of 0.25 m CaCl 2 and 1 ml of (N, NЈ-bis 2 hydroxyethyl)-2 amino ethanesulfonic acid and was incubated for 30 min at room temperature. The calcium phosphate-DNA solution was added dropwise to the cells, and the mixture was swirled gently and incubated overnight at 35 C under 3% CO 2 . The cells were then incubated with fresh growth medium for 8 h before starvation in DMEM containing 0.2% (wt/vol) BSA for 14 h (47, 48). The cells were cooled to 4 C, washed with ice-cold PBS (140 mm NaCl, 3 mm KCl, 6 mm Na 2 HPO 4 , 1 mm KH 2 PO 4 , pH 7.4), and lysed with 50 mm HEPES, 150 mm NaCl, 1% (vol/vol) Triton X-100, 100 mm NaF, 0.5 mm phenylmethylsulfonyl fluoride, 100 U/ml aprotinin (pH 7.6) for 30 min at 4 C. The solubilizates were clarified by centrifugation at 15,000 ϫ g for 15 min at 4 C.
The ⌬121 IGF-1 receptor was partially purified on WGA, as described above. IRS-1 and kinase-dead JAK-2 were immunoprecipitated with specific antibodies.

Fusion proteins
The human JAK-1 cDNA subcloned into pBluescript-KS was a gift from J. Ihle. The SmaI site was introduced to each end of the desired cDNA fragment by PCR to allow the in-frame insertion into the pGEX-2T vector (Pharmacia Biotech Inc., Uppsala, Sweden). The portion of mouse IRS-2 cDNA coding for the amino acids 591-786 was subcloned into vector pGEX-3X (Pharmacia). The two constructs were expressed into Escherichia coli BL21 bacteria (Stratagene, La Jolla, CA). Expression of the recombinant protein was induced with 0.1 mm isopropyl-␤-d-thiogalactopyranoside for 3 h. Bacteria were lysed with 20 mm Tris, 1 m NaCl, 0.2 mm EDTA, 0.2 mm EGTA, 1 mg/ml lysozyme, 0.5 mm phenylmethylsulfonyl fluoride, 100 U/ml aprotinin, 20 mm leupeptine (pH 7.4) for 30 min in ice, sonicated three times, and frozen in liquid nitrogen. Lysates containing GST-IRS-2591-786 were submitted to centrifugation (30,000 ϫ g at 4 C), and the supernatant was incubated with glutathionesepharose (Pharmacia) for 1 h at 4 C. The fusion proteins were eluted with 50 mm glutathione, and 100 mm HEPES (pH 8). Concerning the GST-JAK fusion proteins, lysates were centrifuged (30,000 ϫ g at 4 C), and the pellets were treated with 8 m urea for 30 min at 4 C. After another centrifugation (30,000 ϫ g at 4 C), the supernatants were dialyzed against 70 mm NaCl, 1.5 mm KCl, 3 mm Na 2 P 2 O 7 , 0.5 mm KH 2 PO 4 (pH 7.4). Then, the lysates were incubated with glutathione-sepharose (Pharmacia) for 1 h at 4 C. The fusion proteins were eluted with 50 mm glutathione, 100 mm HEPES (pH 8).

Phosphorylation of JAKs in intact cells
Confluent cells growing in 145-mm culture dishes were starved in DMEM containing 0.2% (wt/vol) BSA for 15 h before being treated with insulin (10 Ϫ7 m) or IGF-1 (10 Ϫ7 m) for different times. Then the cells were washed with ice-cold buffer containing 50 mm HEPES, 150 mm NaCl, 10 mm EDTA, 10 mm Na 2 P 2 O 7 , 100 mm NaF, and 2 mm vanadate (pH 7.4). The cells were solubilized for 30 min at 4 C in lysis buffer (50 mm HEPES, 150 mm NaCl, 10 mm EDTA, 10 mm Na 2 P 2 O 7 , 100 mm NaF, 2 mm vanadate, 0.5 mm phenylmethylsulfonyl fluoride, 100 U/ml aprotinin, 20 mm leupeptin, and 1% (vol/vol) Triton X-100, pH 7.4). The supernatants from a centrifugation step (15 min at 15,000 ϫ g at 4 C) were incubated for 4 h at 4 C with antibodies preadsorbed on protein-A sepharose (anti-JAK1, anti-JAK2, anti-TYK2 antibodies at 1 g/sample; nonimmune serum at 3 g/sample). Pellets were washed with lysis buffer, resuspended in Laemmli sample buffer, and separated by SDS/ PAGE (49). Proteins were transferred to a polyvinylidene fluoride membrane. The membrane was blocked with saline buffer (10 mm Tris, 140 mm NaCl, pH 7.4) containing 5% (wt/vol) BSA for 2 h at 22 C and incubated overnight with mouse monoclonal antibody to phosphotyrosine (1 g/ml). The membrane was washed three times with saline buffer containing Tween-20. Rabbit antimouse antibody (1 g/ml) was added for 60 min at 22 C, followed by several washes. The membrane was then incubated with 125 I-protein A (500,000 cpm/ml) for 60 min at 22 C. The membrane was washed and autoradiographed. In some cases, the membrane was stripped for 30 min at 50 C in 62 mm Tris, 100 mm 2-mercaptoethanol, and 2% (wt/vol) SDS, and reprobed with the indicated antibodies.

Measurement of JAK activation
Cells overexpressing IGF-1 receptors were depleted for 15 h before being incubated with 500 m vanadate for 45 min. The cells were then stimulated, or not, with IGF-1 (10 Ϫ7 m) for 10 min. Clarified cell lysates were incubated for 4 h at 4 C with anti-JAKs antibodies (1 g/sample) preadsorbed on protein-A sepharose. The pellets were washed twice with 50 mm HEPES, 150 mm NaCl, 0.1% (vol/vol) Triton X-100, (pH 7.5). Autophosphorylation was measured by addition of phosphorylation buffer (10 mm HEPES, 50 mm NaCl, 5 mm MnCl 2 , 5 mm MgCl 2 , pH 7.5) containing varying ATP concentrations. GST-IRS-2 phosphorylation was measured by addition of GST-IRS-2 (2 g/sample) and phosphorylation buffer containing 60 m ATP for 5 min at room temperature. For IRS-1 immunopurification, lysates of cells overexpressing IRS-1 were subjected to immunoprecipitation with antibodies to IRS-1 (1/50 dilution) preadsorbed on protein A-sepharose. After 3 h at 4 C, the pellets were washed with 50 mm HEPES, 150 mm NaCl, 1% (vol/vol) Triton X-100 (pH 7.4). The pellets containing IRS-1 were mixed with pellets containing the immunopurified JAKs. Phosphorylation buffer containing 60 m ATP was added to each sample for 5 min at room temperature.
Reactions were stopped by addition of Laemmli sample buffer (49). The samples were analyzed by SDS-PAGE using a 7.5% resolving gel, followed by Western blot analysis with antibody to phosphotyrosine.

In vitro phosphorylation of kinase-dead JAK-2 by WGApurified IGF-1 receptors
Antibodies to JAK-2 (1 g/sample) were incubated with protein A-sepharose for 1 h at 4 C. The pellets were washed twice with 50 mm HEPES, 150 mm NaCl, pH 7.6. Lysates from transfected cells were incubated with the antibody to JAK-2 for 2 h. The JAK-2-containing pellets were washed twice with Hepes NaCl Triton buffer containing 1% (vol/vol) Triton X-100. Antibodies to IGF-1 receptor (1/50) were incubated with protein A-sepharose for 1 h at 4 C. The pellets were washed twice with 50 mm HEPES, 150 mm NaCl, pH 7.6. Five microliters of WGA-purified IGF-1 receptor were incubated with the antibody to IGF-1 receptor for 2 h. The IGF-1 receptor-containing pellets were washed twice with Hepes NaCl Triton buffer containing 0.1% (vol/vol) Triton X-100.
The pellets with kinase-dead JAK-2 were mixed with pellets containing the immunopurified IGF-1 receptor. IGF-1 (10 Ϫ7 m) was added for 30 min at 22 C. Phosphorylation was then initiated by adding 60 m [␥-32 P] ATP (2.5 Ci/mmol), 50 mm MgCl 2 , and the reaction was stopped after 15 min by addition of Laemmli sample buffer (49). The samples were analyzed by one-dimensional SDS-PAGE using a 7.5% resolving gel.

Phosphopeptide map analysis of IRS-1
IRS-1 overexpressed in 293 EBNA-cells was isolated by a specific antibody and phosphorylated in vitro either by insulin receptors or by JAK-1 or JAK-2, as described above, for 2 h with [␥-32 P] ATP. After SDS-PAGE, 32 P-labeled IRS-1 was localized by autoradiography. The gel pieces corresponding to the labeled bands were excised and incubated in 50 mm NH 4 HCO 3 (pH 8) for 2 h at 37 C. CaCl 2 at a final concentration of 1 mm and trypsine at a final concentration of 50 g/ml were added for 12 h at 37 C. For each sample, the eluted phosphopeptides were lyophilized, washed in H 2 O, and dissolved in NH 3 (N/1000). Phosphopeptides were separated by two-dimensional analysis on silice thin-layer plates, as described (50). The plates were dried and subjected to autoradiography.

Coprecipitation of insulin and IGF-1 receptors with GST-JAKs
Precipitation of nonphosphorylated receptors. Five microliters of WGA-purified IGF-1 or insulin receptors were incubated with the fusion protein (60 pmol/sample) preadsorbed on glutathione-sepharose, or with antibodies to receptor preadsorbed on protein A-sepharose (1/500). After 4 h at 4 C, the pellets were washed twice with 50 mm HEPES, 150 mm NaCl, 0.1% (vol/vol) Triton X-100, pH 7.4. The presence of IGF-1 or insulin receptors was detected by ligand-induced autophosphorylation. Precipitation of ligand-occupied receptors. Five microliters of WGA-purified insulin receptors, preincubated with insulin (10 Ϫ7 m) for 60 min at 22 C, were incubated either with the fusion protein (60 pmol/sample) preadsorbed on glutathione-sepharose, or with antibodies to receptor preadsorbed on protein A-sepharose (1/500). After 4 h at 4 C, the pellets were washed twice. Presence of insulin receptors was detected by ligandinduced autophosphorylation, as described above.
The samples were analyzed by SDS-PAGE. The gel was dried and autoradiographed.

Insulin-and IGF-1-induced phosphorylation and activation of JAK-1 and JAK-2
We first examined whether stimulation of cells with insulin or IGF-1 could induce tyrosine phosphorylation of JAKs in NIH 3T3 cells overexpressing the respective receptor. To accomplish this, we stimulated cells, immunoprecipitated a particular JAK, then analyzed its tyrosine phosphorylation using antibodies to phosphotyrosine. In all cases, the total level of JAK protein was determined by stripping the membranes and reprobing with antibodies to JAK. As can be seen in Figs. 1 and 2, both insulin and IGF-1 were more efficient at stimulating phosphorylation of JAK-1 (5.2 Ϯ 0.6 and 4.5 Ϯ 0.4-fold, respectively), compared with JAK-2 (2.4 Ϯ 0.6 and 1.9 Ϯ 0.2-fold, respectively). In addition, there were differences in the time course of phosphorylation, depending on the stimulus. In response to IGF-1, JAK-2 was phosphorylated more rapidly (within 2 min) than JAK-1 (5 min). In contrast, insulin stimulated JAK-1 phosphorylation before that of JAK-2 (2 min compared with 10 min). In all cases, phosphorylation of both JAKs was transient and returned to basal levels within 10 min after IGF-1 stimulation and 20 min after insulin stimulation (data not shown).
Next, we wished to determine whether this phosphorylation was linked to stimulation of the JAK kinase activity. To do this, we tested whether JAKs from stimulated cells had the ability to autophosphorylate or phosphorylate exogenous substrates in vitro.
JAKs were immunoprecipitated from cells, either stimulated or not (in the presence of vanadate, to preserve the phosphorylation state). Experiments performed in the presence of the phosphatase inhibitor have shown that insulinor IGF-1-induced phosphorylation of JAKs can be maintained until 20 min (data not shown). Figure 3A shows a Western blot of immunoprecipitated JAK proteins. In these conditions, basal (i.e. from unstimulated cells) tyrosine kinase activity of JAK is observed, which is frequently seen in cell-free assays of tyrosine kinases. When JAK-1 was purified from IGF-1-stimulated cells, strong autophosphorylation of JAK-1 was induced. This result indicates that JAK-1 isolated from IGF-1 treated-cells was activated. We obtained the same result with JAK-2 (data not shown). We then examined whether these activated JAKs also were able to phosphorylate exogenous substrates such as IRS-1 and IRS-2. GST-IRS-2 was incubated with JAKs immunoprecipitated from either IGF-1-stimulated or nonstimulated cells. Phosphorylation of GST-IRS-2 was determined by Western blotting and probing with antibodies to phosphotyrosine. As shown in Fig. 3B, GST-IRS-2 phosphor-ylation by JAK-1 and JAK-2 isolated from untreated cells was observed. This phosphorylation was clearly increased when JAK-1 and JAK-2 were purified from cells stimulated by IGF-1.
We also tested whether JAKs could phosphorylate IRS-1 immunoprecipitated from IRS-1 overexpressing cells. The immunopurified IRS-1 was added to the pellets containing JAKs. Phosphotyrosine content of IRS-1 was determined as described in Materials and Methods. As shown in Fig. 4, IRS-1 was strongly phosphorylated by the JAKs purified from cells incubated with insulin, compared with IRS-1 phosphorylated by JAKs purified from unstimulated cells. We obtained the same result with insulin-activated JAKs (data not shown).
From these observations, we conclude that upon incubation of cells with IGF-1 or insulin, JAK-1 and JAK-2 become activated, allowing phosphorylation of proteins such as IRS-1 and -2 in vitro. The fact that insulin and IGF-1 receptors, as well as JAKs, phosphorylated IRS-1 and -2, raises the issue concerning a similar or distinct role in IRS-stimulated pathways. To approach this question, we compared the IRS-1 phosphorylation pattern induced by insulin receptors vs. that seen with JAKs. To do this, we performed 32 P-peptide maps of IRS-1 phosphorylated in vitro either by insulin receptors or by JAK-1 in the presence of [␥ -32 P] ATP. Comparison of the two 32 P-peptide maps showed marked differences (Fig.  5). Indeed, a particular panel of phosphopeptides was found in IRS-1 phosphorylated by the insulin receptor, whereas a different panel was seen only in IRS-1 phosphorylated by JAK-1. This indicates that phosphorylation of IRS-1 by JAK-1 or by the insulin receptor could involve distinct tyrosine residues.

JAK-2 is not a substrate of IGF-1 receptor
To determine whether JAK-2 could be a direct substrate of IGF-1 receptor, we constructed a kinase-dead JAK-2 and tested whether it is phosphorylated by IGF-1 receptor in vitro. To do this, kinase-dead JAK-2 from lysates of cells overexpressing the protein and WGA-purified IGF-1 receptors were immunoprecipitated with specific antibodies. The pellets containing kinase-dead JAK-2 were mixed with pellets containing the immunopurified IGF-1 receptor. IGF-1 (10 Ϫ7 m) was added, or not, for 30 min at 22 C. The phosphorylation reaction was then initiated by adding 60 m [␥-32 P] ATP (2.5 Ci/mmol), 50 mm MgCl 2 ; and the reaction was stopped after 15 min. Figure 6 shows that tyrosine phosphorylation of kinase-dead JAK-2 was not induced by IGF-1 receptors. This result suggests that there is no direct phosphorylation of JAK by IGF-1 receptors.

JH6-JH7 and JH1 domains of JAK-1 are involved in interaction with the phosphorylated insulin or IGF-1 receptors
To determine whether JAK-1 could interact directly with insulin and IGF-1 receptors, we produced two GST-fusion proteins consisting of the N-terminal and C-terminal halves of JAK-1 (GST-F1 and GST-F2 are described in Fig. 7). We performed precipitation of unphosphorylated and ligandoccupied insulin receptors with GST, GST-F1, and GST-F2 fusion proteins preadsorbed on glutathione-sepharose, or antibodies to receptor preadsorbed on protein A-sepharose. After receptor incubation with the GST-JAK fragments, pellets were washed extensively to remove the nonassociated receptors. To detect the precipitated receptors, we then performed an in vitro phosphorylation assay in the presence of [␥-32 P] ATP. In this case, phosphorylation occurs only if native or ligand-occupied receptors were previously retained, as is the case using specific antibodies to receptor (Fig.  8, left and middle panels).
As shown, native or ligand-occupied receptors did not associate with GST-JAK fusion proteins.
In a second series of experiments, insulin receptors were phosphorylated with [␥-32 P] ATP before being incubated with JAK-1 fusion proteins. As shown in Fig. 8 (right panel), phosphorylated receptor ␤-subunit was detected both in GST-F1 and in GST-F2 pellets. These results indicate that only the phosphorylated form of insulin receptor is recognized by JAK-1.
Moreover, as shown in Fig. 9 (left and middle panels), IGF-1 receptor ␤-subunit was detected in GST-F1 and GST-F2 pellets only when the receptor was previously phosphorylated. The IGF-1 receptor, deleted of its 121 C-terminal amino acids, still coprecipitated with GST-F1 and GST-F2 (Fig. 9, right  panel), indicating that the C-terminal part of the IGF-1 receptor is not involved in interaction with JAK-1.
Taking our results together, we conclude that both insulin and IGF-1 receptors directly coprecipitate with JAK-1, i.e. without involvement of an additional molecule. Further, only phosphorylated insulin and IGF-1 receptors are able to bind to JAK-1. Interestingly, precipitation of the phosphoreceptors with each part of JAK-1 indicates that both the N-terminus and the C-terminus are involved in the interaction.
To identify more precisely the JAK-1 domains responsible for the interaction, we constructed other fusion proteins with distinct parts of JAK-1 (described in Fig. 7). Insulin receptors or IGF-1 receptors, phosphorylated in the presence of [␥-32 P] In C, both samples were mixed before being analyzed. 32 P-labeled IRS-1 was extracted from SDS-PAGE gel and treated with trypsin. 32 P-peptides were then separated by two-dimensional electrophoretic analysis on silica thinlayer plates. The plates were dried and subjected to autoradiography. Exposure was for 6 days. We show a representative experiment, out of three. The directions of electrophoresis and chromatography are indicated. The points of sample origin are indicated by arrows. FIG. 8. Phosphorylated insulin receptors interact directly with JAK-1. Native receptors, ligand-occupied receptors, and receptors phosphorylated with [␥-32 P] ATP were incubated with GST, GST-F1, and GST-F2 fusion proteins (60 pmol/sample) preadsorbed on glutathione-sepharose, or antibodies to receptor preadsorbed on protein A-sepharose (1/500). After 4 h at 4 C, the pellets were washed twice. The presence of native and ligand-occupied receptors was tested by ligand-induced autophosphorylation (as described in Materials and Methods). Samples were analyzed by SDS-PAGE under reducing conditions, followed by autoradiography.
FIG. 9. The IGF-1 receptor C-terminus is not implicated in the interaction with JAK-1. Unphosphorylated IGF-1 receptors, or phosphorylated IGF-1 receptors (wild-type or ⌬121) were incubated with GST, GST-F1, and GST-F2 fusion proteins (60 pmol/sample) preadsorbed on glutathione-sepharose. After 4 h at 4 C, the pellets were washed twice. The presence of unphosphorylated receptors was tested by ligand-induced autophosphorylation (as described in Materials and Methods). Samples were analyzed by SDS-PAGE under reducing conditions, followed by autoradiography.
In summary, our results show that insulin and IGF-1 receptors interact with JAK-1 upon autophosphorylation and that two domains of JAK-1 are involved in this process.

Discussion
In this study, we looked at whether JAKs are phosphorylated and activated in cells overexpressing insulin or IGF-1 receptors. We show that both JAK-1 and JAK-2 were indeed phosphorylated upon exposure of cells to insulin and IGF-1, whereas TYK-2 was not. These results extend previous reports showing that JAK-1 (20) or JAK-2 (21) were tyrosine phosphorylated in response to insulin.
Moreover, we found that insulin treatment causes a more rapid tyrosine phosphorylation of JAK-1, whereas IGF-1 causes a more rapid tyrosine phosphorylation of JAK-2. It is possible that JAK-1 and JAK-2 have different affinities for the insulin receptor and the IGF-1 receptor, and this would affect the time course of JAK recruitment by the receptors.
To determine whether JAKs interact directly with insulin or IGF-1 receptors, we constructed fusion proteins containing different portions of JAK-1. By precipitation assays, we show that JAK-1 does not interact with native and ligandoccupied insulin receptors. In contrast, JAK-1 interacts with the phosphorylated forms of insulin and IGF-1 receptors, indicating that receptor phosphorylation is necessary for the interaction. These data are consistent with reports showing that binding of JAK-1 to the PDGF receptor also requires receptor phosphorylation (26). In constrast, JAKs can interact constitutively with cytokine receptors, such as the PRL receptor, in a phosphorylation-independent way (34,35). Tyrosine kinase receptors lack proline-rich and acidic motifs found in cytokine receptors and shown to be indispensable for interaction with JAK (32,33). The domain(s) on the tyrosine kinase receptors responsible for the interaction has (have) yet to be identified. However, we found that the C-terminal tail (121 amino acids) of the IGF-1 receptor ␤-subunit is not involved in interaction with JAK-1.
Next, we localized the JAK-1 regions required for interaction with insulin and IGF-1 receptors. We found that two domains of JAK-1 are involved in binding to phosphorylated receptors. These are: 1) the JH6-JH7 domains of the N-terminus; and 2) the JH1 kinase domain in the C-terminus. Further, we observed that the GST-F2 fragment of JAK-1 precipitates the insulin (or IGF-1) receptor less efficiently than the GST-F1 fragment, whereas the GST-F2-1 fragment is as efficient as GST-F1-1. This suggests the occurrence of an inhibitory effect of F2-2 on F2 binding to the receptors. However, this also could be caused by slight modification of protein conformation after fusion to the GST moiety. In JAK-2, the JH6-JH7 domains, which are highly conserved between the different members of the JAK family, also are implicated in the interaction with the ␤-subunit of the interferon-␥ receptor (51), the ␤-subunit of granulocyte macrophage colony stimulating growth factor receptor (52), and the GH receptor (53). Concerning the interaction of JAK-2 with the GH and the granulocyte macrophage colony stimulating growth factor receptors, the kinase domain (JH1) is not required. Thus, the role of the JH1 domain in the interaction of JAK-1 with phosphorylated insulin and IGF-1 receptors could be specific for these proteins.
JAK-1 and JAK-2 are not only phosphorylated in response to insulin and IGF-1 but also are activated. However, the mechanism of the latter process is unknown. One hypothesis could be that JAK activation is triggered by direct phosphorylation by insulin or IGF-1 receptor. However, we did not detect tyrosine phosphorylation of a kinase-dead JAK-2 by the receptor. This result suggests that there is no direct phosphorylation of JAK by IGF-1 receptors.
Another explanation could be that JAKs are stimulated by insulin and IGF-1 receptors in a mechanism similar to that of cytokine receptors, lacking tyrosine kinase activity such as erythropoietin receptors (54). Dimerization of the erythropoietin receptors upon ligand binding leads to transphosphorylation and activation of JAKs. Clustering of the kinases, which are constitutively associated with the erythropoietin receptors, is thought to induce this activation (55).
Our results raise the question of the physiological role of JAK activation by insulin and IGF-1 receptors. Quelle et al. (56) show that STATs are potentially immediate substrates of activated JAKs. A recent study has shown that STAT1, STAT3, and STAT5 from fibroblasts overexpressing insulin receptors are phosphorylated and activated after hormone treatment (57). Moreover, the insulin receptor interacts with and phosphorylates STAT5 in vitro and in insulin-sensitive tissues (57,58). Thus, these data indicate that STATs can be directly activated by the insulin receptor. However, we cannot rule out the possibility that the JAKs are implicated in activation of STATs in intact cells.
Further, we show that JAKs activated by insulin or IGF-1 receptors can directly phosphorylate IRS-1 and IRS-2. Phosphorylated IRS-1 and IRS-2 generate activation of several signaling pathways. In trying to understand the role of IRS-1 phosphorylation by JAKs, we compared the 32 P-peptide maps of IRS-1 phosphorylated in vitro by the insulin receptor vs. JAK-1 or JAK-2. The phosphorylation profile of IRS-1 was found to be different, depending on the tyrosine kinase used. These differences suggest that IRS-1 phosphorylation by JAK-1 or by JAK-2 could occur on tyrosine residues distinct from those phosphorylated by insulin receptors. Phosphorylation of IRS-1 by JAKs could lead to generation of interaction sites for SH2 domain-containing proteins, which are not produced by the insulin receptor. Hence recruitment of JAK-specific downstream signaling molecules could occur. A major challenge is to understand how insulin generates its considerable spectrum of cellular responses, some of which are specific for insulin, and concern mainly metabolic effects, whereas others are seen with several growth factors and relate to growth and differentiation. It remains to be shown whether the JAK-specific sites modified on IRS molecules generate branches in the insulin signaling pathways, which are responsible for a specific type of cellular program, i.e. growth/differentiation vs. metabolism.