The Type I Interferon Receptor Mediates Tyrosine Phosphorylation of the CrkL Adaptor Protein*

Interferon (IFN) α induces rapid and transient tyrosine phosphorylation of the Src homology 2/Src homology 3 (SH2/SH3)-containing CrkL adaptor protein in a time- and dose-dependent manner. Such phosphorylation is most likely regulated by the Type I interferon receptor (IFNR)-associated Tyk-2 kinase, as suggested by the detection of Type I IFN-dependent tyrosine kinase activity in anti-CrkL immunoprecipitates and the IFNα-dependent association of CrkL with Tyk-2 in intact cells. Two other Type I IFNs, IFNβ and IFNω, also induce tyrosine phosphorylation of CrkL, suggesting that the protein is involved in the signaling pathways of several different Type I IFNs. In the IFNα-sensitive U-266 and Daudi cell lines, CrkL interacts via its N terminus SH3 domain with the guanine exchange factor C3G that regulates activation of Rap-1, a small G-protein that exhibits tumor suppressor activity. Thus, tyrosine phosphorylation of CrkL links the functional Type I IFNR complex to the C3G-Rap-1 signaling cascade that mediates growth inhibitory responses.

function of these kinases regulates multiple downstream pathways, including the Stat (reviewed in Ref. 16) and IRS/phosphatidylinositol 3-kinase (17)(18)(19) signaling cascades. Despite the significant advances in our understanding of the early signaling events elicited by the activated Type I IFNR complex, the relationship between the different IFN␣-activated signaling pathways and specific biological activities of Type I IFNs is not well elucidated.
CrkL is a homologue of the v-crk proto-oncogene (20) and belongs to a group of proteins which also includes c-Crk, Grb-2, and Nck. These proteins function as adapters, linking tyrosinephosphorylated receptors or their substrates to downstream signaling elements. CrkL has one SH2 domain and two SH3 domains (21) that mediate its cellular interactions with other proteins. There is accumulating evidence that CrkL associates with C3G (22)(23)(24)(25), which is a guanine exchange factor for Rap-1 (26), a small G-protein that antagonizes the Ras pathway (27) and has tumor suppressor activity (28,29).
In the current report we provide evidence that CrkL is involved in Type I IFN signaling, as evidenced by its rapid tyrosine phosphorylation in response to treatment of cells with IFN␣, IFN␤, and IFN. This phosphorylation is most likely regulated by the Tyk-2 kinase, as suggested by the IFN␣-dependent association of CrkL with Tyk-2 in vivo. We also demonstrate that in the IFN␣-sensitive U-266 and Daudi cell lines, CrkL interacts with C3G via its N terminus SH3 domain, suggesting that CrkL links the functional Type I IFNR complex to the C3G-Rap-1 growth inhibitory pathway.
Immunoprecipitations and Immunoblotting-Cells were stimulated with 1-2 ϫ 10 4 units/ml of IFN␣, IFN␤, or IFN for the indicated times, and the cells were lysed as described previously (30). In some experiments the cells were serum-starved for 2-4 h prior to IFN treatment. Immunoprecipitations were performed as described previously (30). In some experiments the cell lysates were precleared with nonimmune rabbit immunoglobulin prior to immunoprecipitation. Immunoblotting using the ECL method was performed as described previously (30).
Preparation of Glutathione S-Transferase Fusion Proteins and Binding Studies-The construction of the pGEX-CrkL-N-SH3, pGEX-CrkL-C-SH3, and pGEX-CrkL (full length) vectors has been described elsewhere (31). Production of glutathione S-transferase fusion proteins and binding experiments were performed as described previously (17).
In Vitro Kinase Assays-Cell lysates from IFN-treated cells were immunoprecipitated with the indicated antibodies. Immunoprecipitated proteins on protein G-Sepharose beads were washed six times with phosphorylation lysis buffer containing 0.1% Triton X-100, one time in phosphorylation lysis buffer without Triton X-100, and once in kinase buffer (50 mM Hepes, pH 7.4, 10 mM MgCl 2 ), and were subse-* This work was supported by grant CA73381 from the National Institutes of Health and by a grant from the Hairy Cell Leukemia Foundation (to L. C. P.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
quently resuspended in 25 l of kinase buffer containing 10 -20 Ci of [␥-32 P]ATP. The beads were incubated for 30 min at room temperature, and the reaction was terminated by adding 10 l of 50 mM EDTA. The immunoprecipitates were washed once with kinase buffer, and phosphorylated proteins were resolved by SDS-PAGE. The proteins were subsequently transferred to Immobilon membranes, the membranes were incubated for 2-3 h in 1 M KOH at 70°C to select for tyrosinephosphorylated proteins, and phosphorylated proteins were detected by autoradiography.
Cell Proliferation Assays-Cells were seeded in flat bottom 96-well plates at a concentration of 2.5 ϫ 10 5 cells/ml in the presence or absence of the indicated doses of IFN-Con1 and incubated at 37°C for 7 days. Cell proliferation was assessed using an MTT assay as described previously (32).

RESULTS AND DISCUSSION
We initially performed studies in which the IFN␣-sensitive U-266 and Daudi cell lines were incubated in the presence or absence of IFN␣, the cells were lysed, lysates were immunoprecipitated with an anti-CrkL antibody, and immunoprecipitated proteins were resolved in an SDS-PAGE and immunoblotted with antiphosphotyrosine. A 36-kDa tyrosinephosphorylated protein, corresponding to CrkL, was clearly detectable in anti-CrkL immunoprecipitates from IFN␣treated U-266 or Daudi cells (Fig. 1, A and C). Stripping and reprobing the blots with the anti-CrkL antibody confirmed that equal amounts of the CrkL protein were present prior to and after IFN␣ stimulation (Fig. 1, B and D). Similarly, IFN␣-dependent tyrosine phosphorylation of CrkL was seen in studies using Molt-4, HL-60, and 293T cells. 2 We subsequently determined the kinetics and dose dependence of the IFN␣-induced tyrosine phosphorylation of CrkL. IFN␣ induced tyrosine phosphorylation of the protein within 1 min of treatment of U-266 cells. The signal was present at 5 min and declined to base-line levels at 30 -90 min, suggesting that such phosphorylation is rapid and transient (Fig. 2, A and B). The IFN␣-induced phos-phorylation of the protein was also dose-dependent (Fig. 2C) and exhibited a similar dose-response pattern to the one for Type I IFN-induced growth inhibitory effects in these cells (Fig. 2D).
As several previous studies have shown that differences in the biological activities of distinct Type I IFNs exist (33)(34)(35)(36), we determined the effects of other Type I IFNs on the phosphorylation of the protein. Both IFN␤ and IFN induced rapid tyrosine phosphorylation of CrkL in U-266 cells (Fig. 3), suggesting that the protein is involved in the signaling pathways of several different Type I IFNs.
To determine whether CrkL associates with a Type I IFNdependent tyrosine kinase in intact cells, we performed in vitro kinase assays on anti-CrkL immunoprecipitates from Type I IFN-treated cells. Under the conditions of the in vitro kinase assay, phosphorylation of the CrkL protein was detectable prior to IFN treatment and increased further after Type I IFN stimulation (Fig. 4, A and B), strongly suggesting that CrkL associates with and acts as a substrate for a Type I IFNregulated tyrosine kinase.
The Tyk-2 tyrosine kinase is associated with the Type I IFN receptor (9 -12) and is activated during Type I IFN-binding to regulate tyrosine phosphorylation of downstream signaling elements. When anti-Tyk-2 immunoblots were performed on anti-CrkL immunoprecipitates from Molt-4 cells, we noticed that a minimal amount of Tyk-2 was detected in association with CrkL at base line (Fig. 5A). The amount of Tyk-2 associated with CrkL increased significantly after 1 min of treatment with IFN␣, while the signal diminished at 5 min, suggesting that IFN␣ induces a rapid and transient association of CrkL with Tyk-2 (Fig. 5, A and B). Previous studies have also shown that Tyk-2 is constitutively associated with the c-cbl proto-oncogene product (Cbl) (37), which has two binding sites for the CrkL Src homology 2 domain (38). When the interaction of CrkL with Cbl was studied, we noticed that a small amount of Cbl was associated with CrkL at base line (Fig. 5C). The signal increased significantly after 1 min of treatment of cells with IFN␣ and gradually declined to base-line levels (Fig. 5, C and D). Thus, Previous studies have shown that CrkL associates with C3G (22)(23)(24)(25), which functions as a guanine exchange factor for Rap-1 (26). We sought to determine whether CrkL interacts with C3G in U-266 and also Daudi cells, which exhibit sensitivity to the antiproliferative effects of Type I IFNs (35,36,39). C3G was detectable in association with CrkL prior to and after IFN␣ stimulation in both cell lines (Fig. 6). When GST fusion proteins encoding the full-length CrkL protein or the different SH3 domains of CrkL were used in binding studies, C3G bound to the GST-CrkL (full length) and the GST-CrkL-N-SH3, but not the GST-CrkL-C-SH3 (Fig. 6). Thus, CrkL interacts constitutively with C3G in these cells, and such interaction is likely mediated by its N terminus SH3 domain.
Despite the significant advances in our understanding of the mechanisms of early events in Type I IFN signaling and the identification of two signaling cascades (Stat and IRS pathways) activated by IFN␣, the specific roles that different pathways play in the generation of IFN␣-biological responses are not well elucidated at this time. Recent evidence by studies in mice with targeted disruption of the Stat-1 gene have established that the Stat pathway is essential for the antiviral effects of IFN␣ (40,41). On the other hand, our studies have shown that the function of the IRS pathway is not essential for induction of antiviral responses (49). In addition, our data have demonstrated that the function of neither the Stat nor the IRS pathways alone is sufficient to mediate the antiproliferative effects of IFN␣, suggesting that other signaling elements are required (49).
In the current report we provide evidence that the CrkL adaptor protein is engaged in Type I IFN signaling, providing a link between the Type I IFNR-associated Tyk-2 kinase and the guanine exchange factor C3G. Our data also suggest that the CrkL-Tyk-2 interaction may be mediated by the Tyk-2associated c-cbl proto-oncogene product, a finding consistent with the presence of CrkL-binding sites in Cbl (38). The engagement of such a pathway by the Type I IFNR may prove of importance in the generation of IFN␣-induced antiproliferative effects, as C3G regulates activation of Rap-1 that antagonizes Ras and has tumor suppressor activity. The activation of this pathway appears to play a role in cell growth regulation in various other systems. For instance, a recent study has demonstrated that CrkL is tyrosine-phosphorylated during B cell antigen receptor stimulation and interacts constitutively via its N terminus SH3 domain with C3G in Ramos cells (22), suggesting that its function may be involved in the negative regulation of the Ras pathway during B-cell receptor activation (22). Similarly, another recent report has demonstrated that in anergic T-cells, a signaling cascade involving the CrkL-C3G complex is activated, resulting in Rap1 activation and inhibition of T-cell growth (43).
It is of particular interest that CrkL is the major tyrosinephosphorylated protein in the peripheral blood of patients with chronic myelogenous leukemia (CML) (44 -46) and interacts directly with the abnormal BCR-ABL fusion protein (47,48) that causes the malignant transformation (42). IFN␣ has remarkable clinical activity in the treatment of CML patients and in fact is the treatment of choice for patients not eligible for allogeneic bone marrow transplantation. It is tempting to hypothesize that engagement of CrkL by the Type I IFNR mediates the induction of the antiproliferative effects of IFN␣ in CML cells, and future studies should address this issue.