T cell antigen receptor ubiquitination is a consequence of receptor-mediated tyrosine kinase activation.

Engagement of the T cell antigen receptor results in both its phosphorylation and its ubiquitination. T cell antigen receptor ubiquitination was evaluated in Jurkat, a well characterized human T leukemia cell line. Treatment of cells with the tyrosine kinase inhibitor herbimycin A resulted in an inhibition of receptor ubiquitination. Consistent with this, pervanadate, which increases cellular tyrosine phosphorylation, enhanced receptor ubiquitination. A requirement for receptor-mediated tyrosine kinase activity for ubiquitination was confirmed in cells lacking the tyrosine kinase p56 and also in cells that are defective in expression of CD45, a tyrosine phosphatase that regulates the activity of p56. The need for tyrosine kinase activation for ubiquitination was not bypassed by directly activating protein kinase C and stimulating endocytosis of receptors. These observations establish ubiquitination of the T cell antigen receptor as a tyrosine kinase-dependent manifestation of transmembrane signaling and suggest a role for tyrosine phosphorylation in the ligand-dependent ubiquitination of mammalian transmembrane receptors.

Engagement of the T cell antigen receptor results in both its phosphorylation and its ubiquitination. T cell antigen receptor ubiquitination was evaluated in Jurkat, a well characterized human T leukemia cell line. Treatment of cells with the tyrosine kinase inhibitor herbimycin A resulted in an inhibition of receptor ubiquitination. Consistent with this, pervanadate, which increases cellular tyrosine phosphorylation, enhanced receptor ubiquitination. A requirement for receptormediated tyrosine kinase activity for ubiquitination was confirmed in cells lacking the tyrosine kinase p56 lck and also in cells that are defective in expression of CD45, a tyrosine phosphatase that regulates the activity of p56 lck . The need for tyrosine kinase activation for ubiquitination was not bypassed by directly activating protein kinase C and stimulating endocytosis of receptors. These observations establish ubiquitination of the T cell antigen receptor as a tyrosine kinase-dependent manifestation of transmembrane signaling and suggest a role for tyrosine phosphorylation in the ligand-dependent ubiquitination of mammalian transmembrane receptors.
For many transmembrane receptors, including the multisubunit TCR, 1 signaling is initiated by ligand-induced aggregation (1,2). The earliest obligate intracellular event following TCR aggregation is the activation of the src-family protein tyrosine kinases, Lck (p56 lck ) and Fyn (p59 fyn ). Lck and/or Fyn phosphorylate TCR subunits resulting in the association of a third tyrosine kinase, ZAP-70 (70-kDa -associated protein), with the TCR and to subsequent activation events (3)(4)(5). CD45, a tyrosine phosphatase that dephosphorylates key regulatory residues on Lck and Fyn, is also implicated in the initiation of TCR-mediated signaling (6).
TCRs consist of six different polypeptides, these include the antigen-recognition element, in most cells an ␣-␤ heterodimer, and a set of invariant signal transducing subunits. The invariant subunits include CD3-␦, -⑀, and -␥ and the structurally distinct TCRsubunit, which exists within the TCR as a disulfide-linked homodimer (4). The minimal signal transducing element of the TCR is the immunoreceptor tyrosine-based ac-tivation motif (ITAM) (7). monomers have three ITAMs, and each CD3 subunit has one. ITAMs include two tyrosine residues 10 or 11 amino acids apart that are potential phosphorylation sites. The subunit is a particularly prominent substrate for tyrosine phosphorylation; up to 5% of subunits are phosphorylated on multiple tyrosines upon TCR engagement (8 -11).
In addition to being a substrate for tyrosine phosphorylation when cross-linked by antibody or mitogen (11), TCRs also are ubiquitinated. The covalent modification of proteins with chains of ubiquitin, a highly conserved 76-amino acid polypeptide, plays a central role in the targeting of abnormal proteins and a number of regulatory cytosolic and nuclear proteins for degradation in the 26 S proteasome (12)(13)(14)(15)(16). Ubiquitination occurs via a multienzyme process involving families of enzymes termed E1-E3. E1 (ubiquitin-activating enzyme) is involved in the ATP-dependent charging of ubiquitin. The high energy thiol-ester bond between E1 and ubiquitin is transferred to an E2 (ubiquitin-conjugating enzyme). E2s either by themselves or in conjunction with E3s (ubiquitin protein ligases), transfer ubiquitin monomers or multiubiquitin chains to target proteins, where isopeptide linkages are formed with lysine residues.
The signals that lead to ubiquitination of most naturally occurring substrates are unknown. The N-end rule established a relationship between the N-terminal amino acid of certain proteins and susceptibility to ubiquitination (14). For the cyclins and c-Jun, specific internal polypeptide sequences have been implicated in targeting for ubiquitination (17,18), and for cyclins as well as IB␣, serine phosphorylation also plays a role in this process (19 -22).
The TCR is distinguished from most ubiquitination substrates by its long half-life and by being ubiquitinated in response to a specific external stimulus. TCR ubiquitination occurs on multiple subunits and on multiple intracellular lysines, with mono-and multiubiquitinated species detectable (23). As with tyrosine phosphorylation, the subunit is the most prominent substrate for this modification, likely due to the nine intracellular lysines in each monomer, compared with two or three for each of the other invariant subunits. Fundamental to understanding TCR ubiquitination is a determination of the events that couple receptor engagement to this modification. Using Jurkat (24), a well characterized human T leukemia cell line, we address the relationship between receptor occupancy and ubiquitination. Our findings establish a relationship between receptor ubiquitination and early signaling events mediated by TCR engagement and the activation of protein tyrosine kinases.

MATERIALS AND METHODS
Cell Lines and Antibodies-Cell lines were maintained in complete medium containing RPMI 1640 (Biofluids) and 8% fetal calf serum (25). The cell line JCaM1.6 (26) was obtained from the American Tissue * 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.
OKT3 is a monoclonal antibody directed against the ⑀ subunit of the CD3 complex of the TCR (29). Monoclonal anti-phosphotyrosine antibody 4G10 was a gift from Brian Druker (University of Oregon Health Sciences Center). Polyclonal rabbit antisera 527, raised against a peptide corresponding to amino acids 131-143 of human , was generated as described (30). Polyclonal anti-Lck (31) was provided by Lawrence Samelson (National Institutes of Health). The monoclonal antibody A2B4 -2 has been previously described (32). Polyclonal rabbit antiubiquitin has been described (23).
Herbimycin A (Calbiochem, San Diego, CA) was dissolved in dimethyl sulfoxide. Pervanadate was formed by a 1:1 mixture of 100 mM sodium orthovanadate (Sigma) and 0.035% H 2 O 2 and added to cells at 1:50. PMA (Sigma) was dissolved in ethanol.
Experimental Procedures-Except where indicated, 3 ϫ 10 7 cells were stimulated in 1 ml of complete medium with saturating concentrations of OKT3 for 10 or 15 min. Following stimulation, cells were rapidly chilled with cold phosphate-buffered saline supplemented with phosphatase inhibitors (10). Cell pellets were lysed at 4°C in lysis buffer that contained 0.5% (v/v) Triton X-100 (Research Products International, Elk Grove Village, IL). Lysis buffer also included 300 mM NaCl, 50 mM Tris, pH 7.6, and protease and phosphatase inhibitors as described (10). After centrifugation at 15,000 ϫ g for 15 min to remove insoluble material, post-nuclear supernatants were immunoprecipitated with antibodies prebound to protein A-Sepharose (Pierce) for at least 1 h at 4°C. Immunoprecipitates were washed five times with 1 ml of wash buffer that was identical to the lysis buffer except that no iodoacetamide was used and the detergent concentration was 0.1% Triton X-100. Samples were heated to 95°C in SDS-PAGE sample buffer containing 3% ␤-mercaptoethanol prior to resolution on 12% SDS-PAGE (immunoprecipitates) or 7% SDS-PAGE (Triton X-100 soluble whole cell lysates). Gels were transferred to either nitrocellulose membranes (Schleicher & Schuell) or to Immobilon-P (Millipore, Bedford, MA). For anti-ubiquitin blotting, after transfer to Immobilon-P, membranes were washed in phosphate-buffered saline, incubated in 0.1 M potassium phosphate, pH 7.0, containing 0.5% glutaraldehyde for 20 min, and then washed an additional 30 min in phosphate-buffered saline prior to incubation with antibody overnight at room temperature followed by washing, incubation with 125 I protein A (ICN, Costa Mesa, CA), and autoradiography as described (23). Immunoblotting with antiand anti-phosphotyrosine was carried out as described (10). Cell surface analysis of receptor expression was as described (10).

TCR Ubiquitination Is Correlated with Tyrosine
Phosphorylation-To determine whether events distal to TCR aggregation are required for ubiquitination, Jurkat cells were incubated at 37°C for 10 min in the absence or the presence of the TCR agonist OKT3 (Fig. 1A, first two lanes). This monoclonal anti-CD3-⑀ antibody stimulates tyrosine phosphorylation of (33). Detergent-soluble lysates were immunoprecipitated with anti-CD3 together with a polyclonal antiantisera. This combination was used because human dissociates from other TCR components in Triton X-100 (34,35). Immunoblotting with anti-phosphotyrosine revealed the anti-CD3-dependent increase in the 21-kDa form of phosphorylated (Fig. 1A, lower panel, compare first two lanes). When evaluated by immunoblotting with anti-ubiquitin, a number of anti-CD3-dependent ubiquitinated TCR species were detected (Fig. 1A, upper panel). The most prominent of these represent di-, tri-, and tetraubiquitinated forms of , which migrate at 32, 40, and 48 kDa, respectively. The 48-kDa species is superimposed on immunoglobulin heavy chain, which is present regardless of stimulation. Mono-ubiquitinated is seen as a discrete 24-kDa band; this species is often not well visualized, due to poor recognition by polyclonal anti-ubiquitin. In addition to , other TCR com-ponents are also ubiquitinated in response to receptor ligation (11) and contribute to the overall increase in density above 40 kDa. To determine whether tyrosine kinase activation is required for ubiquitination, Jurkat cells were treated with herbimycin A. This tyrosine kinase inhibitor enhances the degradation of src family tyrosine kinases such as Lck and Fyn (36). Herbimycin A at 3 M did not affect viability or receptor levels (not shown) but resulted in a marked diminution in antibodyinduced phosphorylation (Fig. 1A, lower panel). When receptor ubiquitination was evaluated, it was similarly inhibited by this agent (Fig. 1A, upper panel). Incubation of cells with 0.3 M herbimycin A, which did not block phosphorylation, had no effect on TCR ubiquitination.
The herbimycin A results suggest a relationship between TCR ubiquitination and TCR-mediated tyrosine kinase activation. To further assess this relationship, the effects of pervanadate on TCR ubiquitination were evaluated. Pervanadate increases global cellular tyrosine phosphorylation by inhibiting protein tyrosine phosphatases and has been shown to mimic represents Triton X-100-soluble whole cell lysates from 1.5 ϫ 10 6 cell equivalents resolved on 7% SDS-PAGE and immunoblotted with anti-phosphotyrosine. In C, immunoprecipitated TCRs were resolved on 12% gels followed by immunoblotting with either anti-ubiquitin or anti-phosphotyrosine as indicated.
TCR-mediated signaling (37)(38)(39)(40). When Jurkat cells were pretreated with pervanadate, a dramatic increase in total cellular tyrosine phosphorylation was observed in whole cell lysates, regardless of TCR engagement with anti-CD3 (Fig. 1B). When TCRs were specifically immunoprecipitated from pervanadatetreated cells, the subunit of the TCR was also found to be tyrosine-phosphorylated (Fig. 1C, lower panel). The level of 21-kDa phosphoseen with pervanadate alone was substantially greater than that achieved with anti-CD3 (OKT3) crosslinking, and together pervanadate and anti-CD3 were synergistic with regard to phosphorylation. When evaluated for ubiquitination (Fig. 1C, upper panel), pervanadate by itself resulted in the appearance of ubiquitinated TCRs, although the level of ubiquitination achieved with pervanadate was consistently less than that seen with anti-CD3 (Fig. 1C, upper panel, and data not shown). As with tyrosine phosphorylation, anti-CD3 and pervanadate together had synergistic effects on ubiquitination. In multiple experiments, concomitant treatment with pervanadate and anti-CD3 resulted in increases in ubiquitination from 2-to 8-fold relative to anti-CD3 alone. In conjunction with the herbimycin A results, these findings suggest that tyrosine kinase activation plays a crucial role in TCR ubiquitination and that receptor engagement facilitates the generation of TCR-ubiquitin conjugates.
The Tyrosine Kinase Lck Is Required for TCR Ubiquitination-To address whether tyrosine kinases specifically involved in TCR-mediated signaling are required for ubiquitination, we next evaluated the Jurkat variant JCaM1.6. This cell line is defective in TCR-mediated signaling due to loss of enzymatically active Lck (26). JCaM1.6 was subcloned (JCaM1.6.22) to achieve surface TCR levels matching wild type Jurkat and evaluated for both TCR phosphorylation and ubiquitination ( Fig. 2A). When compared with Jurkat, antibody-induced phosphorylation was dramatically reduced (Fig. 2A, lower panel), as was receptor ubiquitination (Fig. 2A, upper panel). To establish if loss of Lck is responsible for the lack of ubiquitination in JCaM1.6.22, Lck expression was restored by transfec-tion of JCaM1.6.22 with an Lck-encoding expression vector. The restoration of Lck expression in two independent stable transfectants, JLck.B3 and JLck.H5, was documented by immunoblotting with anti-Lck (Fig. 2B). Re-expression of Lck resulted in reconstitution of phosphorylation (Fig. 2C, lower  panel). Additionally, re-expression of Lck reconstituted anti-CD3-dependent TCR ubiquitination (Fig. 2C, upper panel), thus establishing a requirement for Lck in this process.
In Lck Negative Cells, TCR Engagement and Protein Kinase C Activation Are Not Sufficient for Ubiquitination-Tyrosine kinase activation plays an important role in a number of events distal to TCR cross-linking. One consequence of TCR-mediated tyrosine kinase activation that directly effects the receptor itself is the activation of protein kinase C. Activation of this serine/threonine kinase stimulates the phosphorylation of CD3 subunits on serine residues and results in enhanced receptor endocytosis (41)(42)(43)(44). Treatment of T cells with the protein kinase C-activating phorbol ester, PMA, bypasses TCR-mediated tyrosine kinase activation and stimulates protein kinase C. This results in the serine phosphorylation of CD3 subunits and an intracellular redistribution of receptors, even in the absence of TCR occupancy (41)(42)(43)(44).
As shown (Fig. 3A), although anti-CD3-dependent receptor endocytosis is less pronounced in Lck-deficient JCaM1.6.22 when compared with Jurkat (Fig. 3A, compare a and b with e and f), treatment with PMA bypasses TCR-mediated signaling and results in an intracellular redistribution of receptors comparable with that seen with Jurkat (Fig. 3A, c and g; d and h). Having established that PMA results in receptor redistribution in JCaM1.6.22, the effects of combinations of PMA and anti-CD3 on TCR ubiquitination were evaluated in this cell line. Neither PMA plus anti-CD3 nor PMA alone (Fig. 3B) resulted in any detectable receptor ubiquitination in JCaM1.6. Thus, serine phosphorylation of TCR subunits and receptor redistribution is not sufficient to result in ubiquitination of TCRs. Ionomycin (a calcium ionophore) and PMA, which together activate Jurkat cells in a TCR-independent fashion (45), also failed to restore ubiquitination in JCaM1.6.22, even when receptors were concomitantly cross-linked with anti-CD3 (not shown). Consistent with the requirement for tyrosine kinase activation for ubiquitination, treatment of wild type Jurkat with PMA and/or ionomycin in the absence of TCR engagement resulted in neither TCR ubiquitination nor tyrosine phosphorylation (not shown).
TCR Ubiquitination Requires the Regulatory Tyrosine Phosphatase CD45-In addition to tyrosine kinases, TCR-mediated signaling also requires the tyrosine phosphatase CD45. This TCR-associated transmembrane phosphatase is believed to function in the initiation of signaling by dephosphorylating key regulatory residues on Lck and Fyn (6). CD45-deficient cells are defective in signaling and do not exhibit the normal increase in TCR phosphorylation seen in response to receptor occupancy (27,46,47). In such a Jurkat variant (J45.01), no ligand-induced TCR ubiquitination was detected (Fig. 4A), despite comparable levels of surface TCR (not shown) and total cellular (Fig. 4B, lower panel). As with the Lck negative cells, PMA and ionomycin did not result in ubiquitination (not shown). As expected, when these cells were reconstituted with CD45 (J45.LB3.3) (28), anti-CD3-dependent TCR ubiquitination was easily detected (Fig. 4B, upper panel). This establishes that TCR ubiquitination is not only dependent on tyrosine kinases but also on the presence of a tyrosine phosphatase that regulates kinase activity. DISCUSSION This study establishes that TCR ubiquitination requires intact coupling to tyrosine kinase activation and is not simply the consequence of the recognition of aggregated TCR cytoplasmic domains by E2/E3 enzymes. This requirement is not bypassed by stimulating serine phosphorylation and internalization of engaged receptors. Pervanadate, which results in an acute increase in tyrosine phosphorylation, is sufficient to result in detectable levels of TCR ubiquitination, even in the absence of receptor ligation. Thus, an acute increase in tyrosine phosphorylation, independent of receptor occupancy, appears to be sufficient to result in ubiquitination. However, the finding that the level of ubiquitination seen with pervanadate is consistently less than that found with receptor occupancy suggests that specific signals generated in response to TCR engagement may be important in stimulating a maximal level of TCR ubiquitination. These signals may be a direct manifestation of TCR oligomerization or perhaps reflect a different temporal order or pattern of phosphorylation induced when tyrosine kinases are activated by TCR ligation.
The platelet-derived growth factor receptor and c-kit (the stem cell factor receptor) are tyrosine kinase-containing members of the growth factor family of receptors that are ubiquitinated in response to their cognate ligands (48,49). In the case of platelet-derived growth factor receptor, mutation of autophosphorylation sites correlates with decreased ligand-dependent ubiquitination (50), and for c-kit, the tyrosine kinase inhibitor genestein results in decreased ligand-dependent ubiquitination (49). Several other mammalian transmembrane receptors that either signal by coupling to tyrosine kinase activation (51) or that contain intrinsic tyrosine kinase activity (52) are ubiquitinated in an occupancy-dependent manner. Taken together with our findings, these observations suggest that tyrosine phosphorylation likely plays an important role in ligand-dependent ubiquitination of a number of mammalian transmembrane receptors. One means by which receptor ubiquitination might occur in response to tyrosine kinase activation is by the phosphorylation of E2/E3 enzymes with resultant changes in their associations and/or activities. In fact, in one case, the in vitro tyrosine phosphorylation of an E2 was found to correlate with enhanced ubiquitination (53). Alternatively, receptors might be ubiquitinated by E2/E3 enzymes that contain sites that bind phosphotyrosine, such as is seen with SH2 domains, although no enzymes fitting this description have thus far been identified.
The function of ubiquitination in the biology of transmembrane receptors remains to be elucidated. It has generally been assumed that transmembrane receptors are degraded in lysosomes. However, to be exposed to lysosomal proteases, intracytosolic receptor domains would first need to be engulfed in autophagocytic vesicles. Alternatively, the intracytosolic domains of receptors could be degraded, at least in part, in a ubiquitin-dependent fashion in the 26 S proteasome. This could be viewed as a protective mechanism, insuring the degradation of the signaling domains of activated receptors. Although re-FIG. 3. Protein kinase C activation fails to stimulate ubiquitination. A, cell surface expression of TCRs in Jurkat (a-d) and JCaM 1.6.22 (e-h) cells in response to combinations of anti-CD3 and PMA (100 ng/ml). Cells were incubated with the indicated additions for 30 min at 37°C and then cooled to 4°C. Samples were incubated with anti-CD3 (solid lines) or a nonbinding murine antibody (A2B4 -2) (dotted lines, in upper two panels only), followed by incubation with fluorescein isothiocyanate-labeled goat anti-mouse F(abЈ) 2 (Southern Biotechnology Associates, Birmingham, AL) and analysis by FACScan (Becton Dickinson, Mountain View, CA). B, cells were incubated at 4°C with the indicated additions prior to stimulation at 37°C followed by lysis, immunoprecipitation, SDS-PAGE, and immunoblotting with anti-ubiquitin. sults obtained with the platelet-derived growth factor receptor are suggestive (50), for no transmembrane receptor has a causal relationship between ligand-induced ubiquitination and proteasomal degradation been established. The recent finding of herbimycin A-induced ubiquitination and proteasomal degradation of the insulin-like growth factor receptor, while not addressing the issue of ligand-dependent ubiquitination, demonstrates that transmembrane receptors may, in fact, be degraded by proteasomes (54).
Regardless of the fate of ubiquitinated receptors, the steric effects of ubiquitin moieties branching off of the intracytoplasmic tails of receptors would be expected to impact negatively on intracellular associations with signaling molecules and between receptors. For the TCR, Fyn, ZAP-70, and CD45 are among the TCR-associated proteins that could be affected by ubiquitination. Thus, whether or not ubiquitination is a major factor in ligand-dependent receptor degradation, it is likely that this modification represents a means of modulating the function of activated transmembrane receptors.