INTRODUCTION

The proto-oncogene c-Cbl was originally identified as a cellular homologue of the transforming protein expressed by the murine Cas NS-1 retrovirus (1). The retroviral form, v-Cbl, was found to be expressed as a gag fusion with the C-terminal 355 amino acids of c-Cbl. v-Cbl contains a putative nuclear localization signal, has an ability to bind DNA, and was shown to reside in both the nucleus and cytoplasm. The full-length c-Cbl protein, however, has been characterized as a predominantly cytoplasmic protein that is particularly abundant in lymphoid cells (2).

c-Cbl has many tyrosine residues, some of which become phosphorylated after the stimulation of lymphocytes through their antigen receptors (3). c-Cbl also contains a proline-rich region suitable for binding SH3-containing proteins. In fact, it has been reported that c-Cbl can bind the adapter molecules Grb-2 and Nck through its proline-rich region (4, 5). Furthermore, in vitro studies demonstrate the potential association of c-Cbl with multiple nonreceptor tyrosine kinases (6, 7). Recent work on the function of c-Cbl in Caenorhabditis elegans has suggested that it can act as a negative regulator of EGF¹ receptor signaling (8). This evidence suggests that the role of c-Cbl is to function as a cytoplasmic intermediate in signal transduction pathways elicited through growth factor receptors. However, the specific signal transduction pathways in which c-Cbl participates and the precise role of c-Cbl in these pathways are unclear.

Numerous cellular responses are elicited after stimulation of the T cell antigen receptor·CD3 complex. These responses include the induction of IL-2 secretion, the up-regulation of receptors for IL-2 and transferrin, and the initiation of cell proliferation (9). Numerous proteins, of which c-Cbl is a prominent one, become tyrosine phosphorylated following T cell receptor activation (3). Thus, c-Cbl may be a component of the signal transduction pathways that lead to cell proliferation or adhesion or to both following T cell receptor activation.

Previous studies from our laboratory and others have shown that, on stimulation, c-Cbl associates in vivo with the p85 subunit of PI 3-kinase (4, 10). In fact, c-Cbl is the predominant tyrosine-phosphorylated protein bound to p85 · p110 on lymphocyte stimulation (10). These results suggest that c-Cbl and PI 3-kinase are elements of the same signal transduction pathway. Interestingly, PI 3-kinase does not appear to be involved in mediating mitogenic responses elicited by costimulation in T cells since this response is insensitive to wortmannin, a fungal toxin that completely blocks the activation of PI 3-kinase following T cell receptor activation (11). PI 3-kinase activity, however, has been implicated in the activation of integrin binding to fibronectin in T cells (12). These results suggest that c-Cbl and PI 3-kinase may function in a signal transduction pathway involved in integrin activation and cell adhesion. PI 3-kinase has also been found to be involved in the down-regulation of platelet-derived growth factor receptors expressed in HepG2 cells (13). A role for c-Cbl·PI 3-kinase complexes in the down-regulation of yet unidentified signaling complexes in lymphoid cells may also exist.

To further understand the properties of c-Cbl and the significance of its interactions with PI 3-kinase, we have further studied the cellular biological and biochemical responses of c-Cbl to stimulation in lymphoid cells. Our results indicate that a specific pool of c-Cbl is recruited from the cytosol to the membrane fraction after activation, where it forms a complex with PI 3-kinase. After activation, c-Cbl becomes associated with a detergent-insoluble fraction, possibly composed of cytoskeletal elements. These data support the possibility that c-Cbl·PI 3-kinase complexes operate at the membrane, perhaps to regulate the changes in cytoskeletal function that occur after T cell activation. Interestingly, the phosphorylation of c-Cbl and the recruitment of c-Cbl·PI 3-kinase complexes were also observed in a mutant cell line (JCaM1.6) that lacks p56^lck, a tyrosine kinase which is critical for mitogenic signal transduction in response to T cell receptor activation (15). These results suggest that the activation of c-Cbl and the formation of c-Cbl·PI 3-kinase complexes occur upstream or independently of mitogenic signal transduction pathways in T cells.

MATERIALS AND METHODS

Rabbit polyclonal antisera to rat PI 3-kinase (No. 06-195) and a monoclonal antibody raised against phosphotyrosine (4G10) were obtained from Upstate Biotechnology, Inc. Unlabeled goat anti-Ig (IgM, IgG, IgA) used for stimulating A20 cells and cross-linking OKT3 was obtained from Cappel. Affinity-purified c-Cbl antisera was purchased from Santa Cruz Biotechnology. Polyclonal antisera against c-Cbl (R2) was a generous gift from Dr. Wallace Langdon. This antisera was generated against a fusion protein comprising the C-terminal 366 amino acids of c-Cbl (amino acids 540-906) (2). The OKT3 and OKT4 hybridomas were purchased from ATCC (CRL8001, CRL8002). Antibody was concentrated from cell supernatants using ABx Bakerbond (J. T. Baker Inc.).

Jurkat T cells and A20 B cells were obtained from ATCC. JCaM1.6 cells were provided by Dr. Arthur Weiss and are described elsewhere (15). All cells were grown in complete RPMI 1640 medium supplemented with 10% fetal calf serum. T cells were aliquoted (2.5 × 10⁷ cells) into Microfuge tubes and stimulated using 50 µg/ml ABx-purified OKT3 supplemented with excess goat anti-mouse IgG. B cells were used at 3-5 × 10⁷ cells and stimulated with 10-50 µg of anti-Ig. At the time points indicated in each experiment, cells were pelleted in a Microfuge for 5 s, placed on ice, and resuspended in cold buffer (see below). Wortmannin was obtained from Sigma and kept aliquoted in dimethyl sulfoxide at 70 °C. It was thawed and diluted in phosphate-buffered saline immediately before addition to the cells.

Cells were suspended in 1 ml of an ice-cold buffer composed of 1% Triton X-100, 20 mM Tris, 150 mM sodium chloride, 1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, 1 mM 1,10-phenanthroline, 1 mM sodium vanadate, and 50 mM sodium fluoride. Lysates were clarified by centrifugation at 10,000 × g for 15 min and precleared by incubation with protein A-Sepharose prior to the addition of specific antibodies. After 120 min of incubation at 4 °C, protein A-Sepharose was added, and incubations were continued for an additional 60 min. Protein A beads were then washed three to five times in a wash buffer composed of 20 mM Tris, 150 mM sodium chloride, 0.2% Triton X-100, and 0.1% SDS, followed by one wash in phosphate-buffered saline before boiling in SDS sample buffer. Immunoprecipitates were resolved by SDS-PAGE, and transferred to nitrocellulose for immunoblotting. Primary antibodies were detected by chemiluminescence (Amersham Corp.).

Hypotonic lysis and fractionation were performed as described in Ref. 18. Briefly, cells were incubated on ice in a hypotonic buffer (10 mM Tris, 0.5 mM MgCl₂, and phosphatase and protease inhibitors as described above) for 10 min and then homogenized with 30 strokes in a Dounce glass homogenizer. Tonicity was restored by the addition of 0.25 volume of (10 mM Tris, 0.5 mM MgCl₂, 0.6 M NaCl, and then nuclei and unbroken cells were pelleted at 500 × g for 5 min. EDTA was added to the supernatant to a final concentration of 0.05 M before spinning at 100,000 × g for 45 min. The resulting supernatant constituted the cytosolic fraction. The pellet was resuspended and solubilized in lysis buffer (300 mM NaCl, 50 mM Tris, 0.5% Triton X-100, and protease and phosphatase inhibitors as described above) and centrifuged at 10,000 × g for 15 min. The supernatant from this centrifugation step was termed the ``membrane fraction,'' and the pellet was termed the ``insoluble fraction.''

RESULTS

We have previously shown that the p85 subunit of PI 3-kinase associates with c-Cbl on antigen receptor stimulation of T and B cells (10). To gain further insight on the potential cellular function of these complexes, we analyzed their subcellular localization in fractionated Jurkat T cells or A20 B cells.

Cells were stimulated for 2 min and fractionated as described under ``Materials and Methods.'' Samples from each fraction (cytoplasm, total membrane fraction, and detergent-insoluble fraction) were separated by SDS-PAGE, transferred to nitrocellulose, and probed with antibodies raised against c-Cbl. In both control and stimulated cells, the majority of the cellular pool of c-Cbl was found in the cytoplasmic fraction (Fig. 1). However, within 2 min of stimulation, a 2.5-fold increase in the concentration of c-Cbl in the membrane fraction was observed. This fraction was defined by being Triton X-100 soluble. A 2.5-3-fold increase in the concentration of c-Cbl in the Triton-insoluble fraction of the membrane was also observed in response to stimulation (Fig. 1, A and B).

The recruitment of c-Cbl to the Triton-insoluble membrane fraction lagged behind the recruitment to the total membrane fraction by approximately 2 min (Fig. 1C). The Triton X-100-resistant pool of c-Cbl was also highly resistant to extraction with sodium deoxycholate and octyl glucoside (data not shown). The time course of recruitment and the insolubility in detergent of c-Cbl present in this fraction suggest that c-Cbl becomes associated with cytoskeletal elements after recruitment to membranes in response to stimulation.

Our finding that a fraction of c-Cbl is recruited to cellular membranes in response to stimulation prompted us to analyze the subcellular localization of c-Cbl·PI 3-kinase complexes, which form in response to stimulation of T and B cells. Jurkat cells were stimulated and fractionated into a cytosolic and a membrane fraction as described above. The membrane fraction was separated from Triton X-100-insoluble materials, and antisera to the C terminus of c-Cbl were then used to immunoadsorb c-Cbl from both cytosol and membrane fractions. Immunoprecipitates were resolved on SDS-PAGE and probed with antibodies against phosphotyrosine (Fig. 2A, top), against c-Cbl (Fig. 2A, middle), or against the p85 subunit of PI 3-kinase (Fig. 2A, bottom). Stimulation induced an increase in tyrosine phosphorylation of c-Cbl immunoprecipitated from either the cytosolic or membrane fractions. However, substantially more c-Cbl protein was immunoprecipitated from the cytosolic fraction than from the membrane fraction (Fig. 2A, middle). The ratio of phosphotyrosine/c-Cbl protein immunoprecipitated from the cytosol increased 2-fold with stimulation, whereas the ratio of phosphorylation/c-Cbl protein within the membrane increased 4-fold as the result of stimulation (based on arbitrary values obtained by densitometric analysis of three independent experiments). Thus, the stoichiometry of tyrosine phosphorylation of c-Cbl was higher in immunoprecipitates from the membrane relative to those from the cytosol.

Analysis of c-Cbl immunoprecipitates with antibodies against p85 revealed the presence of PI 3-kinase in immunoprecipitates from the cytosolic fraction from stimulated but not from unstimulated cells. A substantially larger amount of p85 was detected in c-Cbl immunoprecipitates from the membrane fraction compared with the cytosolic fraction of stimulated cells despite the fact that these immunoprecipitates contained less c-Cbl protein and, overall, less phosphotyrosine in c-Cbl. Densitometric analysis of immunoblots from several different experiments indicated that immunoprecipitates of membrane-associated c-Cbl contained approximately 20-fold more p85 than those from cytosolic c-Cbl after stimulation (Fig. 2B). Thus, although the majority of tyrosine-phosphorylated c-Cbl is found in the cytosol, the vast majority of c-Cbl·p85 complexes appear localized to the membrane fraction after stimulation.

Multiple mechanisms, whereby c-Cbl·p85 complexes form and associate with membrane fractions in response to stimulation, can be hypothesized. To sort out these mechanisms, we have begun to dissect the elements of the signal transduction pathways elicited through T cell receptor activation that lead to the formation of c-Cbl·p85 complexes. For this purpose, we employed a mutant Jurkat clone, JCaM1.6, which lacks significant expression of the Src family tyrosine kinase, p56^lck (15). The lack of p56^lck expression results in a severe decrease in the ability to induce tyrosine phosphorylation of numerous cellular proteins on antigen receptor cross-linking and ultimately results in abrogation of responses to T cell stimulation, such as calcium mobilization and IL-2 production.

Immunoprecipitates of c-Cbl from JCaM1.6 cells were probed with antibodies against phosphotyrosine (Fig. 3A, top), against c-Cbl (Fig. 3A, middle), or against the p85 subunit of PI 3-kinase (Fig. 3A, bottom). Tyrosine phosphorylation of c-Cbl following receptor cross-linking was readily observed and was similar to that observed in wild-type Jurkat T cells (compare Figs. 2A and 3A). In fact, c-Cbl is one of the few proteins that is readily tyrosine-phosphorylated in JCaM1.6 cells after stimulation (not illustrated). As was found in wild-type Jurkat cells, a substantially larger amount of p85 was detected in c-Cbl immunoprecipitates from the membrane fraction when compared with the cytosolic fraction of stimulated cells despite the fact that these immunoprecipitates contained less c-Cbl protein and overall less phosphotyrosine in c-Cbl (Fig. 3A). Approximately 7-fold more p85 was found in immunoprecipitates of c-Cbl from the membrane fraction compared with the cytosolic fraction of stimulated JCaM1.6 cells (Fig. 3B). These results indicate that the formation of c-Cbl·p85 complexes at the membrane and their association with a detergent-insoluble fraction (not illustrated) are triggered through a p56^lck-independent pathway.

We next addressed whether the activity of PI 3-kinase was required for the recruitment of c-Cbl and the formation of c-Cbl·p85 complexes. For these experiments, we employed wortmannin, a fungal toxin that irreversibly binds to the catalytic subunit of PI 3-kinase and that is a potent and specific inhibitor of its lipid and protein kinase activities. Jurkat cells were incubated for 10 min at 37 °C in the presence or absence of 100 nM wortmannin and then stimulated with cross-linking antibodies to CD3. The cells were then fractionated, and fractions were analyzed by SDS-PAGE and immunoblotting. Inhibition of PI 3-kinase activity did not prevent the recruitment of c-Cbl to the membrane fraction (Fig. 4A). Furthermore, no differences in the phosphorylation of c-Cbl on tyrosine residues or on the formation of c-Cbl·p85 complexes were observed (Fig. 4B) in response to wortmannin. These results suggest that the catalytic activity of PI 3-kinase does not influence either the association of c-Cbl with p85 or the association of c-Cbl with the membrane and cytoskeletal fractions of Jurkat T cells.

DISCUSSION

Previous work from our laboratory has shown that c-Cbl is the principal tyrosine-phosphorylated protein associated with the p85 subunit of PI 3-kinase following T and B cell activation (10). Here we show that this association occurs concomitantly with the translocation of c-Cbl to membrane and insoluble/cytoskeletal fractions (Fig. 1). In addition, we show that phosphorylation of c-Cbl and formation of membrane-bound c-Cbl·p85 complexes occur by a mechanism independent of the nonreceptor tyrosine kinase p56^lck.

Others have shown by immunofluorescent methods that Fc receptor stimulation of a monocyte cell line induces a translocation of c-Cbl from a diffuse to a concentrated perinuclear staining (14). These authors have suggested that this staining is consistent with a translocation to the trans Golgi network. However, based on previous work on PI 3-kinase in other cell systems, it is equally likely to be an endosomal/lysosomal membrane compartment with which c-Cbl becomes associated (13).

The mechanism whereby c-Cbl associates with membrane fractions is not known. However, it is interesting to note that c-Cbl was found to be associated with membranes prior to stimulation (Figs. 2A and 3A). Under these conditions, c-Cbl is neither tyrosine phosphorylated to a large extent nor associated with p85. Thus, other regions in c-Cbl, such as the large proline-rich region, might be involved in determining its membrane association (2). This association of c-Cbl with membranes may in turn be modulated by posttranslational modification that alters the structure of the protein, such as tyrosine phosphorylation.

Despite a high level of tyrosine phosphorylation of both cytosolic and membrane-associated c-Cbl, the membrane-derived fraction is associated with p85 to a much greater extent. We have previously shown that the association between p85 and c-Cbl is principally mediated through the SH2 domains of p85 (10). When taken together, these results suggest that the phosphorylation of tyrosine residues on c-Cbl may be heterogeneous, and a critical phosphotyrosine residue is phosphorylated preferentially when c-Cbl is associated with the membrane fraction. Tyrosine 751 on c-Cbl is flanked by amino acids that comprise a consensus site for p85 SH2 binding. This tyrosine residue lies within the region to which a specific antisera against c-Cbl (R2) was generated (2). Interestingly, immunoprecipitation using the R2 antisera results in less coprecipitatation of p85 compared with other anti-cbl antibodies (data not shown), lending support to this site being the target for p85 binding to c-Cbl.

The mechanism whereby the p85 binding sites become phosphorylated in response to stimulation is not known. In this paper, however, we find that phosphorylation of c-Cbl and its association with p85 are independent of p56^lck. This result is significant because the stimulation-induced tyrosine phosphorylation of most proteins is abrogated in p56^lck-deficient cells (15). The fact that c-Cbl is one of a few exceptions suggests that a signal transduction pathway distinct from the p56^lck pathway is involved in the activation of c-Cbl and the formation of c-Cbl·p85 complexes. This finding may be relevant in terms of understanding the physiological significance of the formation of c-Cbl·p85 complexes. Abrogation of p56^lck leads to impaired early and late T cell responses, including induction of IL-2 production and mitogenesis. The finding that p56^lck deficiency does not impair the formation of c-Cbl·p85 complexes indicates that these complexes are insufficient to signal a mitogenic response. The role of these complexes may not be directly related to mitogenic signaling but perhaps may be related to other important responses to stimulation, such as adhesion. The finding that inhibition of PI 3-kinase activity with wortmannin does not inhibit mitogenic signaling in lymphocytes, but does inhibit adhesion (12), is consistent with this hypothesis.

Alternatively, c-Cbl·p85 complexes may operate in pathways that involve down-regulation of yet unidentified membrane proteins. A role for PI 3-kinase activity in the down-regulation of receptor tyrosine kinases in other cell systems has been proposed (13). Interestingly, C. elegans, a homologue of c-Cbl (8), functions as a negative regulator of signals elicited through a homologue of the EGF receptor. Results by others have suggested that c-Cbl acts as a cytosolic adaptor protein for PI 3-kinase recruitment to the EGF receptor (16). It is interesting to speculate on the possibility that c-Cbl·PI 3-kinase complexes in this system may operate as negative regulators by virtue of their ability to down-regulate receptor tyrosine kinases. Disruption of c-Cbl function would be predicted to result in enhanced mitogenesis or transformation. Interestingly, when 17 amino acids in the central region of c-Cbl are deleted, the protein becomes transforming (17). This protein may operate as a dominant negative of endogenous c-Cbl function. Continued work to determine the p85 binding site in c-Cbl as well as the regions of c-Cbl that determine its association with cellular membranes will be necessary to answer these questions.