The Phosphorylation State of CD3γ Influences T Cell Responsiveness and Controls T Cell Receptor Cycling*

The T cell receptor (TCR) is internalized following activation of protein kinase C (PKC) via a leucine (Leu)-based motif in CD3γ. Some studies have indicated that the TCR is recycled back to the cell surface following PKC-mediated internalization. The functional state of recycled TCR and the mechanisms involved in the sorting events following PKC-induced internalization are not known. In this study, we demonstrated that following PKC-induced internalization, the TCR is recycled back to the cell surface in a functional state. TCR recycling was dependent on dephosphorylation of CD3γ, probably mediated by the serine/threonine protein phosphatase-2A, but independent on microtubules or actin polymerization. Furthermore, in contrast to ligand-mediated TCR sorting, recycling of the TCR was independent of the tyrosine phosphatase CD45 and the Src tyrosine kinases p56Lckand p59Fyn. Studies of mutated TCR and chimeric CD4-CD3γ molecules demonstrated that CD3γ did not contain a recycling signal in itself. In contrast, the only sorting information in CD3γ was the Leu-based motif that mediated lysosomal sorting of chimeric CD4-CD3γ molecules. Finally, we found a correlation between the phosphorylation state of CD3γ and T cell responsiveness. Based on these observations a physiological role of CD3γ and TCR cycling is proposed.

Receptor internalization has been extensively studied, and a number of the internalization/sorting signals involved in this event have been identified. Thus, both tyrosine-and leucinebased receptor internalization/sorting motifs have been described in several receptors (1). Some receptors are recycled to the cell surface following ligand or protein kinase C (PKC) 1 -induced internalization and some cycle continuously between the plasma membrane and intracellular endosomal compartments. However, despite extensive studies on receptor sorting, it is still not known whether receptor recycling is a signalmediated or a default sorting event (2). Likewise, the regulation of receptor recycling is poorly understood.
The T cell receptor (TCR) constitutes the clonotypic Ti␣␤ heterodimer, the CD3␥⑀ and CD3␦⑀ dimers, and the 2 homodimer (3)(4)(5)(6)(7). The Ti␣␤ heterodimer is responsible for recognition of antigen (8,9), and the associated CD3 and chains (10) are responsible for the signaling events that follow antigen recognition. It has recently been described that the TCR is internalized and sorted to a degradative compartment after antibody or antigen stimulation (11)(12)(13). In contrast, some studies have indicated that the TCR is not degraded following PKC-mediated internalization (14,15). Furthermore, ligandmediated TCR sorting is dependent on tyrosine phosphorylation but independent of PKC and the CD3␥ Leu-based sorting motif (16,17). In contrast, PKC-mediated TCR internalization is dependent on the Leu-based sorting motif in CD3␥ but independent of tyrosine phosphorylation (16,18,19). Thus, two independent pathways for TCR sorting seem to exist. The physiological role of PKC-mediated TCR internalization and recycling is not known.
The aim for this study was to examine the functional state of recycled TCR and the mechanisms involved in TCR sorting following PKC-induced internalization. We find that following PKC-induced internalization, the TCR recycles to the cell surface in a functional state. Recycling was absolutely dependent on dephosphorylation of CD3␥ but was independent upon actin polymerization, microtubules, and tyrosine phosphorylation. Even though the phosphorylation state of CD3␥ controlled TCR recycling, no recycling signal was contained within the cytoplasmic tail of CD3␥. In addition, our results indicated that not only TCR cycling but also T cell responsiveness can be regulated by the phosphorylation state of CD3␥.
Constructs, Transfection, and TCR Down-regulation-All CD3␥ mutations were constructed as described previously (7,18) by the polymerase chain reaction (PCR) using Vent DNA polymerase containing 3Ј 3 5Ј-reading exonuclease activity (New England Biolabs, Beverly, MA) and the plasmid pJ6T3␥-2 (22) as template. Chimeric CD4-CD3␥ constructs were produced as described previously (19) using the plasmid pCD-L3T4.25 (23) as template. The PCR products were digested with XbaI and EcoRI and subcloned into the expression vector pMH-Neo (24). Mutations were confirmed by DNA sequencing using the ABI 310 (Perkin-Elmer). Transfections were performed using the Bio-Rad Gene Pulser at a setting of 270 V and 960 microfarads with 40 g of plasmid per 2 ϫ 10 7 cells. After 3-4 weeks of selection, G418-resistant clones were expanded and maintained in medium without G418. For TCR down-regulation, cells were adjusted to 2 ϫ 10 5 cells per ml of medium (RPMI 1640 ϩ 10% FCS) and incubated at 37°C with different concentrations of PDBu or F101.01 for 45 min. At the indicated time cells were transferred to ice-cold PBS containing 2% FCS and 0.1% NaN 3 and washed twice. The cells treated with PDBu were stained directly with PE-conjugated UCHT1 (Dakopatts A/S). Cells stimulated with F101.01 were stained with F101.01 followed by PE-conjugated F(ab) 2 fragments of goat anti-mouse Ig H ϩ L. Cells were analyzed in a FACSCalibur flow cytometer (Becton Dickinson, Mountain View, CA). Mean fluorescence intensity (MFI) was recorded and used in the calculation of percent antibody binding: (MFI of phorbol ester-treated cells) divided by (MFI of untreated cells) ϫ 100%. For each construct at least three different clones were analyzed.
Receptor Recycling-Cells were pretreated for 60 min with cycloheximide to stop protein synthesis, incubated with 111 nM PDBu for 45 or 60 min, washed three times in RPMI 1640 ϩ 10% FCS to remove the PDBu, and incubated in medium (RPMI 1640 ϩ 10% FCS) with cycloheximide at 37°C. At different time points, aliquots of cells were analyzed for TCR surface expression. The cells were stained with either PE-conjugated anti-CD3⑀ (UCHT1) or PE-conjugated anti-CD4 (L3T4) mAbs and analyzed in a FACSCalibur flow cytometer (Becton Dickinson). MFI was recorded and used in the calculation of percent anti-CD3 or anti-CD4 mAb binding: (MFI of phorbol ester-treated cells) divided with (MFI of untreated cells) ϫ 100%. For each construct at least three different clones were analyzed.
Cell Permeabilization and Total Staining for CD3⑀ and CD4 -Cells were washed in PBS and fixed for 10 min with 1% paraformaldehyde. The cells were permeabilized for 10 min at room temperature with washing buffer (Hepes-buffered PBS, containing 0.1% saponin) and stained with either PE-conjugated anti-CD3⑀ (UCHT-1) or anti-CD4 (L3T4) mAb. MFI was determined by FACS analysis on a FACSCalibur.

RESULTS
The TCR Recycles to the Cell Surface in a Functional State following PKC-mediated Internalization-In contrast to ligand-mediated sorting of the TCR (11), the TCR is not sorted to a degradative compartment following PKC-induced internalization (14). To analyze whether the TCR recycled to the cell surface in Jurkat cells, cells were pretreated with cycloheximide to stop protein synthesis and then incubated with PDBu for 60 min. The cells were subsequently washed to remove PDBu and incubated in medium without PDBu but with cycloheximide. At different times, aliquots of cells were analyzed for TCR surface expression (Fig. 1A). These experiments showed that following PKC-mediated TCR internalization, the TCR was recycled to the cell surface. Approximately 60 -90 min after removal of PDBu, TCR surface expression reached the same level as untreated cells (Fig. 1A). Omitting cycloheximide did not change TCR internalization or recycling significantly (data not shown). As receptor stimulation and internalization have been shown to result in receptor desensitization of some receptors (28,29), we next analyzed whether the internalized and recycled TCRs were desensitized or were able to signal when re-expressed at the cell surface. Jurkat cells were either left untreated, treated with PDBu for 45 min, or treated with PDBu for 45 min and washed to remove the PDBu followed by incubation in medium without PDBu for 120 min to allow TCR recycling. The cells were then analyzed for TCR surface expression as well as for [Ca 2ϩ ] i responses after stimulation with an anti-TCR antibody. Treatment of the cells with PDBu for 45 min resulted in down-regulation of the TCR and an almost abolished [Ca 2ϩ ] i response (Fig. 1, A and B). However, after 120 min of recycling, the responsiveness of the T cell changed from a state of almost non-responsiveness to that of untreated cells (Fig. 1B). To examine whether the lack of responsiveness following PDBu treatment was due to a reduction in cell-surface expression of the TCR or to an unknown effect of PDBu, JGN␥-S123/126V cells were analyzed. In JGN␥-S123/126V cells the CD3␥ phospho-acceptor group Ser-123 and Ser-126 have been mutated to valines. Consequently CD3␥ is not phosphorylated upon activation of PKC, and TCR internalization is almost abolished (18) (Fig. 1A). In correlation with a minimal internalization of the TCR upon PKC activation, TCR signaling in JGN␥-S123/126V cells as measured by [Ca 2ϩ ] i after stimulation with an anti-TCR antibody was only slightly affected by the PDBu treatment (Fig. 1C). The difference in PDBu-induced internalization and TCR signaling following PBDu treatment of JGN␥-S123/126V cells compared with JGN␥-WT cells was not due to a general defect in JGN␥-S123/126V cells. Thus antibody stimulation leads to TCR internalization as well as increased tyrosine phosphorylation in both JGN␥-S123/126V and JGN␥-WT cells (Fig. 1, D and E). These results demonstrated that activation of PKC induced internalization of the TCR with subsequent sorting to a non-degradative compartment from where the TCR recycled back to the cell surface in a functional state. Thus, the TCR is not desensitized following PKC-induced internalization and recycling. Furthermore, the effect of PKC on T cell responsiveness was solely due to the CD3␥-dependent internalization of the TCR.
Recycling of the TCR Requires PP-2A-mediated Dephosphorylation of CD3␥-We have previously shown that PKC-mediated internalization of the TCR requires phosphorylation of CD3␥ S126 (18). To analyze whether CD3␥ was dephosphorylated after PKC-mediated phosphorylation, Jurkat cells were loaded with 32 P and incubated with PDBu for 10 min. The cells were then washed and transferred to medium without PDBu. At different times, aliquots of cells were lysed, and immunoprecipitation was performed with an anti-CD3⑀ mAb. The precipitates were analyzed by SDS-polyacrylamide gel electrophoresis and autoradiography. CD3␥ was phosphorylated upon PKC activation, but 45 min after removal of PDBu, most CD3␥ had been dephosphorylated ( Fig. 2A, upper panel). This showed that phosphorylated CD3␥ is quickly dephosphorylated in vivo, probably by serine/threonine protein phosphatases in agreement with other studies (30). Furthermore, the kinetics of CD3␥ dephosphorylation correlated with that of TCR recycling, as most of the PKC-induced internalized TCRs were re-expressed at the cell surface after 60 -90 min (Fig. 2B). To analyze whether recycling of the TCR subsequent to PKC-mediated TCR internalization was dependent on serine/threonine protein phosphatases, Jurkat cells were treated with PDBu and subsequently with different serine/threonine protein phosphatase inhibitors. After incubation with PDBu and phosphatase inhibitor for 45 min, the cells were washed to remove PDBu and transferred to medium with phosphatase inhibitor. At different times, aliquots of cells were analyzed for TCR surface expression. The inhibitors calyculin A and okadaic acid strongly inhibited TCR recycling, whereas tautomycin only had a minor effect on TCR recycling (Fig. 2B). Calyculin A most markedly inhibited TCR recycling. However, incubation with PDBu and calyculin A for 4 h did not lead to degradation of the TCR (Fig. 2C). Thus, the lack of recycling in the presence of calyculin A was not due to degradation of internalized TCR. To determine whether calyculin A inhibited dephosphorylation of CD3␥, 32 P analysis was performed with or without treatment with calyculin A, as described above. Calyculin A strongly inhibited dephosphorylation of CD3␥ ( Fig. 2A, lower panel), correlating with calyculin A-mediated inhibition of TCR recycling. From these studies, it may be suggested that dephosphorylation of CD3␥ following PKC-mediated internalization of the TCR is required for TCR recycling.
TCR Recycling Is Not Dependent on Microtubules or Actin Polymerization-Actin polymerization has recently been shown to be involved in antigen-mediated T cell activation (31). Furthermore, both actin filaments and microtubules have been shown to be involved in post-endocytic sorting (32). To analyze whether TCR recycling was dependent upon actin polymerization or microtubules, Jurkat cells were treated with PDBu and subsequently with either cytochalasin D, an inhibitor of actin polymerization (32), or nocodazole, an agent that disrupts microtubules (33). After incubation with PDBu and the inhibitor for 45 min, the cells were washed to remove PDBu and transferred to medium with the inhibitor. At different times, aliquots of cells were analyzed for TCR surface expression. TCR recycling was not affected by cytochalasin D, and nocodazole only slightly reduced TCR recycling (Fig. 2D). This indicates that actin polymerization or microtubules are not important for TCR recycling.
Recycling and Src kinases in ligand-mediated internalization of the TCR (16,17). To analyze the TCR recycling event further, we analyzed whether CD45 or the Src tyrosine kinases p56 Lck or p59 Fyn were required for recycling of the TCR following PKCinduced internalization. TCR recycling was examined in the CD45 negative Jurkat T cell line J. 45.01 (34), the p56 Lck negative Jurkat T cell line JCaM 1.6 (35), or in Jurkat T cells treated for 30 min with PP1, an Src kinase inhibitor (36). TCR recycling was analyzed as described above. Lack of CD45, p56 Lck , or the inhibition of Src kinases did not prevent recycling of the TCR to the cell surface (Fig. 3A). Thus, although CD45 and the Src kinases p56 Lck and p59 Fyn were required for efficient ligand-mediated internalization of the TCR (Fig. 3B) (16,17), these proteins were not required for the recycling of the TCR.
A Lysosomal Sorting Signal in CD3␥ Controls TCR Recycling-Since PKC-induced internalization of the TCR required phosphorylation of CD3␥ and TCR recycling required dephosphorylation of CD3␥, we next analyzed whether TCR recycling was mediated by a specific recycling signal in the cytoplasmic tail of CD3␥. We first analyzed whether the residues C-terminal to the Leu-based motif in CD3␥ contained any sorting information involved in the recycling event. CD3␥ was truncated at Pro-133 (CD3␥-tP133). This construct was transfected into the Jurkat CD3␥ negative (JGN) cell line (20). G418resistant clones were isolated that expressed the TCR at the cell surface and were subsequently analyzed as described above. After PKC-mediated internalization the TCR containing the CD3␥-tP133 chain recycled to the cell surface (Fig. 4B) and was not degraded (Fig. 4C). This indicated that the residues C-terminal to the Leu-based motif in CD3␥ did not play a role in the recycling of the TCR. To analyze whether the residues comprising the remaining part of the cytoplasmic tail of CD3␥ (Gln-117-Leu-132) contained a recycling signal, a chimeric CD4-CD3␥ construct in which the amino acids Gln-117-Leu-132 of CD3␥ substituted for the cytoplasmic tail of CD4 (CD4-3 Gln-117-Leu-132 ) was made. In this construct, CD3␥ Asp-127 was changed to an alanine to allow surface expression of the chimera, as we have previously shown that CD4 containing the CD3␥ SDXXXLL motif is spontaneously sorted to the lysosomes  (19). CD4-3 Gln-117-Leu-132 was indeed expressed at the cell surface in JGN cells (data not shown) and was internalized following PKC activation (Fig. 4B). However subsequent to internalization CD4-3 Gln-117-Leu-132 was not recycled to the cell surface ( Fig. 4B) but was degraded (Fig. 4, C and D). Thus, the CD3␥ sequence Gln-117-Leu-132 in the context of CD4-3 Gln-117-Leu-132 did not lead to sorting via a recycling route but instead sorted the chimera to a degradative compartment. In agreement with previous studies CD4-3stop, which lacked the CD3␥ Leu-based motif, was not internalized upon PKC activation (19) demonstrating that the CD3␥ Gln-117-Ala-125 sequence was not responsible for the PKC-induced internalization and lysosomal sorting of the CD4-3 Gln-117-Leu-132 chimera. Taken together, these results indicated that if TCR recycling is a signal-mediated event, the recycling signal is probably not found within the CD3␥ chain.

Role of Protein Serine/Threonine Phosphatases in
Recycling of the TCR-Ligand stimulation of T cells leads to TCR internalization and sorting to the lysosomes (11). In contrast, PDBuinduced activation of PKC leads to sorting to a non-degradative compartment, presumably the early endosomes, from where the TCR is recycled back to the cell surface (Fig. 1). The rate of TCR recycling correlated with that of CD3␥ dephosphorylation indicating that TCR recycling was controlled by the phosphorylation state of CD3␥. Thus, the phospho-acceptor group in CD3␥ plays at least two roles, a role in TCR internalization and a role in TCR recycling. In agreement with previous studies (30,37) we found that the serine/threonine protein phosphatase acting on CD3␥ probably belongs to the phosphatase-2A group as the inhibitors of CD3␥ dephosphorylation and TCR recycling, calyculin A and okadaic acid, primarily inhibit this group of phosphatases (38,39). A similar phosphorylation/ dephosphorylation-dependent cycling behavior has been described for the ␤-adrenergic (40), the C5a anaphylatoxin (41), the neurokinin 1 (42), and the chemokine CXCR4 (43) receptors. Interestingly, like the TCR chain CD3␥, the C5a receptor contains a cytoplasmic di-leucine motif preceded by an N-terminal potential PKC site containing a serine. In fact, this serine has been shown to be phosphorylated following ligand binding or phorbol ester treatment (44), and as for the TCR (Fig. 2B), treatment of cells with okadaic acid delayed the recycling of the C5a receptor to the cell surface. Furthermore, the CXCR4 receptor also contains a di-leucine like sequence (IL) preceded by a potential PKC site which may be involved in PKC-mediated CXCR4 receptor internalization (43). Thus, it seems likely that these receptors share some of the mechanisms that mediate and regulate their intracellular sorting.
TCR Can Be Sorted by Different Independent Sorting Routes-The polymerization of actin or microtubules was not important for TCR recycling or PKC-mediated internalization of the TCR. In agreement with our results, inhibition of actin polymerization did not affect recycling of the transferrin receptor (45). In contrast, actin polymerization has been shown to be required for ligand-mediated receptor internalization and T cell activation (31,45) indicating that the mechanisms involved in PKC-induced internalization/recycling and ligand-mediated T cell activation/TCR internalization are different. In agreement with recent results (16,17) CD45 and the Src tyrosine kinases were required for efficient ligand-induced internalization of the TCR (Fig. 3B), demonstrating that these proteins play an important role in the internalization event that leads to lysosomal sorting of the TCR. In contrast, TCR recycling was not dependent upon CD45 or the Src kinases p56 Lck and p59 Fyn . Likewise, PKC-induced internalization of the TCR is not dependent upon CD45, the Src kinases p56 Lck and p59 Fyn , or tyrosine phosphorylation (Fig. 3) (16). This indicates that the PKC-dependent and the ligand-dependent TCR sorting routes represent two different pathways of receptor sorting. Stimulation of the TCR with ligand activates both tyrosine kinases and PKC but leads to lysosomal sorting of the TCR. This implies that ligand/tyrosine kinase-mediated TCR sorting to lysosomes of the TCR is dominant over PKC-induced TCR sorting. Thus, sorting of TCR via the recycling pathway may only apply to unstimulated TCR and consequently may predominantly occur prior to ligand stimulation. Regarding the sorting of the TCR following PKC-induced internalization, we found that CD3␥ does not contain a cytoplasmic recycling signal. In contrast, it was shown that the Leu-based motif in CD3␥, Ser-126 to Leu-132, possessed the ability to mediate lysosomal sorting of CD4-CD3 Gln-117-Leu-132 (Fig. 4). Why the TCR is recycled and not sorted to the lysosomes following PKC-mediated internalization despite the potential of the CD3␥ Leu-based motif to mediate lysosomal sorting is still not known. Following PKCmediated TCR internalization CD3␥ was dephosphorylated (Fig. 2). It has been suggested that phosphorylation exposes the Leu-based internalization/degradation motif in CD3␥ (19). According to this suggestion, dephosphorylation of CD3␥ might lead to masking of the CD3␥ Leu-based motif which may explain the lack of lysosomal sorting following PKC-induced internalization of the TCR. However, the present results argue against this explanation. Thus, TCRs were not sorted to a degradative compartment despite being located in an intracellular compartment for 4 h in a state in which CD3␥ was most likely phosphorylated due to the presence of the phosphatase inhibitor calyculin A (Fig. 2C). This indicates that irrespective of an exposed lysosomal targeting signal in CD3␥ other factors and/or TCR chains prevent the sorting to the lysosomes. In line with these experiments, other phosphorylation-dependent recycling receptors have also been shown to be retained, but not degraded, in an intracellular compartment in a phosphorylated state (43,44). Thus, in these receptors, as well as in the TCR, phosphorylation leads to internalization and localization in an intracellular compartment but not to degradation.
A Physiological Function of CD3␥ and TCR Cycling-It is not known why the TCR is sorted to the lysosomes following stimulation with antigen. It could be speculated that the TCR is desensitized after antigenic stimulation and that these stimulated/desensitized TCRs influence the activation threshold of the T cell due to competition for the ligand with unstimulated TCRs. In contrast, the present results showed that the PKCinduced cycling event does not lead to TCR desensitization (Fig.  1). Thus, the internalized TCRs are able to signal when reexpressed at the cell surface. PKC-induced internalization of the TCR and the following recycling to the cell surface probably reflects the spontaneous cycling of the TCR (16). Since the signaling ability of the TCR is not affected by the cycling event, the responsiveness of the T cell can be quickly altered by changing the kinetic parameters in the cycling process. T cell activation requires a certain amount of ligand and surfaceexpressed TCR to enable an activation signal that is above the threshold for T cell activation (46,47). Regarding the amount of surface-expressed TCR, our results showed that this amount, and thereby the activation threshold of the T cell, can be controlled solely by the activities of PKC, a phosphatase that is probably PP-2A, and the CD3␥ SDXXXLL motif (Figs. 1 and 2). Concerning the physiological role of CD3␥ and TCR cycling, the mechanisms involved in TCR cycling may be targets for immune regulatory factors, e.g. certain cytokines, which may modulate the activity of PKC or PP-2A and thereby the phosphorylation state of CD3␥. This in turn will affect TCR surface expression and T cell responsiveness as shown in Fig. 1. Indeed several cytokines, some of which are secreted by professional antigen-presenting cells, have been reported to affect the activities of PP-2A or PKC and may thus affect T cell responsiveness toward antigenic stimulation (48 -50). Apart from playing a role in setting the activation threshold of the T cell, it could also be suggested that TCR cycling may ensure that only functional TCR is expressed at the cell surface. Thus, a role for the constitutive cycling of the TCR or other receptors may be to serve as a quality check of the receptors as only fully functional and assembled receptors can proceed through the entire cycling event.
In conclusion, these studies have identified a specific role for the phosphorylation state of CD3␥ in the internalization and recycling of the T cell receptor and in determining the responsiveness of the T cell. Although controlling both internalization and recycling of the TCR, CD3␥ did not contain a recycling signal in its cytoplasmic tail. Thus, whether TCR recycling is a signal-mediated event involving a recycling signal in other TCR chains or is a default process is still unknown. Likewise, the signal mediating ligand-induced lysosomal sorting of the TCR has not yet been identified. We are presently addressing these questions as well as whether the mechanisms involved in TCR cycling are targets for immune regulatory factors.