Masking of the CD3γ di-Leucine-based Motif by ζ is Required for Efficient T-Cell Receptor Expression

Authors


Dr. Carsten Geisler, c.geisler@immi.ku.dk

Abstract

The T-cell receptor (TCR) is a multimeric receptor composed of the Tiαβ heterodimer and the noncovalently associated CD3γδε and ζ2 chains. All of the TCR chains are required for efficient cell surface expression of the TCR. Previous studies on chimeric molecules containing the di-leucine-based endocytosis motif of the TCR subunit CD3γ have indicated that the ζ chain can mask this motif. In this study, we show that successive truncations of the cytoplasmic tail of ζ led to reduced surface expression levels of completely assembled TCR complexes. The reduced TCR expression levels were caused by an increase in the TCR endocytic rate constant in combination with an unaffected exocytic rate constant. Furthermore, the TCR degradation rate constant was increased in cells with truncated ζ. Introduction of a CD3γ chain with a disrupted di-leucine-based endocytosis motif partially restored TCR expression in cells with truncated ζ chains, indicating that the ζ chain masks the endocytosis motif in CD3γ and thereby stabilizes TCR cell surface expression.

T-cells are activated through the interaction between the T-cell receptor (TCR) and major histocompatibility complex (MHC) molecules associated with antigen-derived peptides on antigen-presenting cells. The TCR is a multimeric receptor composed of the polymorphic Tiαβ heterodimer and the noncovalently associated CD3γδε and ζ2 chains (1–6). The Tiαβ dimer is responsible for recognition of MHC/peptide complexes (7,8), and the CD3 and ζ chains are responsible for signal transduction across the plasma membrane (9–13). Efficient cell surface expression of the TCR requires that all of the TCR chains are produced (7,8,14,15) and is closely regulated by different sorting motifs present in the chains (16–22). Regulation of the TCR expression level at the cell surface is most probably an important event determining the fate of the T-cell. Thus, it has been shown that peripheral tolerance can be maintained by down-regulation of the TCR (23,24). Likewise, the fate of mature T-cells following stimulation is dependent on the TCR cell surface expression level (25–27).

So far, three pathways for TCR endocytosis and thus regulation of TCR cell surface expression levels have been described, all of which are dependent on the di-leucine-based (diL) motif of the CD3γ chain: (i) protein kinase C (PKC)-induced, (ii) ligand-induced, and (iii) constitutive. The most extensively examined pathway is endocytosis caused by phorbol ester-induced PKC activation (19,22,28,29). In nonstimulated T-cells, the CD3γ diL motif is hardly accessible to adaptor proteins involved in the endocytosis process. Upon PKC-mediated phosphorylation of serine 126 in the CD3γ chain, the diL motif becomes fully accessible for binding by the clathrin-binding adaptor protein AP2. Binding of AP2 to the CD3γ diL motif subsequently leads to endocytosis of the TCR. The majority of TCR endocytosed following PKC activation is recycled back to the plasma membrane (19,30,31). In contrast to PKC-induced TCR endocytosis, relatively little is known about the mechanisms underlying ligand-induced endocytosis. It is caused by TCR triggering by MHC/peptide complexes (32–34), super-antigens (35,36) or anti-TCR antibodies (37,38). Earlier studies indicated that ligand-induced TCR endocytosis was independent of the CD3γ diL motif and only dependent on src kinase activity (39,40). However, recently it was shown that both the CD3γ diL motif and PKC activity are required for efficient ligand-induced TCR endocytosis (21,41). Endocytosis following ligand stimulation leads to degradation of the TCR rather than recycling (22).

In addition to PKC- and ligand-induced endocytosis, the TCR is subjected to constitutive endocytosis and recycling (42–45). A recent study demonstrated that constitutive TCR endocytosis is independent of PKC and src kinase activity but dependent on the CD3γ diL motif (46). When this motif is present in fully assembled TCR complexes as well as in CD16/CD3γ chimeras, both of which associate with the ζ2 homodimer, constitutive endocytosis is limited (47). However, when the motif is present in CD4/CD3γ chimeras that do not associate with ζ, constitutive endocytosis takes place at a high rate (31,47). Interestingly, the apparent stabilizing effect of ζ on the TCR complex seems to rely more on the length than on the precise sequence of the cytoplasmic tail of ζ as the green fluorescence protein (GFP) can substitute for the cytoplasmic tail of ζ and still rescue cell surface expression of the TCR (48).

The aim of this study was to further investigate the role of ζ and the interplay between ζ and CD3γ in the regulation of TCR cell surface expression levels. Our results confirm that TCR expression levels are dependent on the length of the cytoplasmic tail of ζ. The reduced TCR cell surface expression levels in cells with truncated ζ were caused by a combination of increased endocytic and unaffected exocytic rate constants of the TCR. Furthermore, the degradation rate constant of the TCR was increased in cells with truncated ζ. Introduction of a CD3γ chain with a disrupted diL motif partially restored TCR expression in cells with truncated ζ chains. These results indicated that the ζ chain masks the diL motif in CD3γ and thereby stabilizes TCR cell surface expression.

Results

TCR expression levels are dependent on the length of the ζ tail

We have previously shown that the activity of the CD3γ diL motif is low in completely assembled TCR complexes and in CD16/CD3γ chimeras that associate with ζ, but high in CD4/CD3γ chimeras that do not associate with ζ (31,47). Although most of these experiments were performed with chimeric molecules and not with intact TCR complexes, the results suggested that ζ might control TCR expression by masking the diL motif in CD3γ.

To analyze whether ζ plays a role in regulating TCR cell surface expression levels, a number of truncated ζ constructs were made (Figure 1A) and separately transfected into the ζ-negative cell line MA5.8 (14). This resulted in the cell lines MA5.8-ζ-164, which expresses full-length ζ, MA5.8-ζ-133, which expresses a truncated ζ chain lacking the 31 C-terminal amino acids, MA5.8-ζ-103, which expresses a truncated ζ chain lacking the 61 C-terminal amino acids, and finally MA5.8-ζ-Δ55-84, which expresses a ζ chain lacking the 61 C-terminal amino acids and residues 55–84. As shown in Figure 1B, truncating the ζ chain with 31, 61 and 91 amino acids led respectively to surface expression levels of 60%, 33% and 24% of the TCR surface expression levels in cells with full-length ζ. The transfections were performed by retroviral infection to achieve high transfection efficiency and thereby eliminate the need for subsequent cloning. Preliminary studies using a GFP-containing retroviral vector demonstrated that more than 60% of the cells became transfected and that the construct was highly expressed in the transfectants. Thus, cloning was not necessary, excluding the possibility that the results were a consequence of varying expression of the constructs due to selective cloning. The transfection experiments were repeated three times with similar results. Additional truncations of ζ led to even lower TCR cell surface expression levels but at the same time destroyed the anti-ζ mAb recognition site, preventing proper determination of the transfected ζ chain (the anti-ζ mAb recognition site is given in bold in Figure 1A). To analyze whether the reduced surface expression could be due to differences in expression levels of the ζ constructs the cells were lysed in Triton X-100 and analyzed by Western blotting for the presence of ζ, CD3ε, and GAPDH as a TCR-independent loading control. Semiquantifications of the blots from the transfectants demonstrated that the highest levels of ζ expression were found in MA5.8-ζ-164 and MA5.8-ζ-133 cells and the lowest level in MA5.8-ζ-Δ55-84 cells (Figures 2A,B). Interestingly, the level of ζ correlated with the level of CD3ε. In nontransfected MA5.8 cells the total level of CD3ε expression was only 13% of the expression level in MA5.8-ζ-164 cells. The highest levels of CD3ε expression were found in MA5.8-ζ-164 and MA5.8-ζ-133 cells and gradually decreased with additional truncations of ζ (Figure 2A,B). The reduced level of CD3ε and ζ in transfectants with truncated ζ could be due either to reduced synthesis rates or increased degradation rates of the chains. Metabolic pulse-chase labeling experiments demonstrated that nontransfected and all of the transfected MA5.8 cells synthesized the Tiαβ and the CD3 chains at the same rate (Figure 3 and data not shown). The synthesis rate of the ζ chain could not be compared between the transfectants due to different numbers of methionines and different molecular weight of the constructs. However, the experiments clearly indicated that the presence of ζ had a stabilizing effect on the CD3ε chain that was dependent on the length of the cytoplasmic tail of ζ. Coimmunoprecipitation experiments using anti-CD3ε mAb indicated that all of the ζ constructs had the ability to associate with the rest of the TCR complex (Figure 2D). Semiquantifications of the gels indicated that approximately the same amount of ζ was coprecipitated with CD3ε in the various transfectants when correcting for the amount of CD3ε in the transfectants. The multiple ζ bands (Figure 2C) in the cell lines expressing truncated ζ most probably was caused by different tertiary structures rather than different secondary structures of the chains, as these bands collapsed into one band when the samples were run under reducing conditions (Figure 2A). Interestingly, most of the different folded forms of ζ associated with CD3ε (data not shown).

Figure 1.

Figure 1.

TCR expression correlates inversely with the length of the ζ tail. A) Schematic presentation of the amino acid sequences of the cytoplasmic tails of ζ-164, -133, -103, and -Δ55-84. The amino acids shown in bold represent the sequence recognized by the anti-ζ mAb 6B10.2.B) TCR cell surface expression in MA5.8 cells transfected with ζ-164, -133, -103 or -Δ55-84 as determined by flow cytometry. Transfectants were stained with PE-conjugated anti-CD3ε mAb (black profiles) or an irrelevant mAb (white profiles). The abscissa gives the fluorescence intensity in a logarithmic scale. The ordinate gives the relative cell number in arbitrary units.

Figure 2.

Figure 2.

The ζ constructs are expressed and associate with CD3ε in MA5.8 cellsA) Western blot analysis of lysates from MA5.8 (lane 1), MA5.8-ζ-164 (lane 2), MA5.8-ζ-133 (lane 3), MA5.8-ζ-103 (lane 4), and MA5.8-ζ-Δ55-84 (lane 5) cells. Cells were lysed and the proteins separated by SDS-PAGE under reducing conditions. The proteins were transferred to nitrocellulose filters and immunoblotted with anti-ζ mAb (upper panel), polyclonal anti-CD3ε Ab (middle panel), and anti-GAPDH mAb (lower panel).B) Relative expression levels of CD3γ and ζ. The intensity of the CD3γ and ζ bands was semiquantified and corrected for differences in loading as represented by the intensity of the GAPDH bands. Subsequently, the intensities were normalized by setting the intensity of MA5.8-ζ-164 derived bands to 100% (left and middle group of the columns). Finally, the normalized expression levels of ζ and CD3ε were compared (right group of columns).C) Western blot analysis of lysates from MA5.8 (lane 1), MA5.8-ζ-164 (lane 2), MA5.8-ζ-133 (lane 3), MA5.8-ζ-103 (lane 4), and MA5.8-ζ-Δ55-84 (lane 5) cells. Cells were lysed and the proteins separated by SDS-PAGE under nonreducing conditions. The proteins were transferred to nitrocellulose filters and immunoblotted with anti-ζ mAb.D) Western blot analysis of ζ coprecipitated with anti-CD3ε mAb from MA5.8 (lane 1), MA5.8-ζ-164 (lane 2), MA5.8-ζ-133 (lane 3), MA5.8-ζ-103 (lane 4), and MA5.8-ζ-Δ55-84 (lane 5) cells. Upper panel: Precipitated proteins were separated by SDS-PAGE under reducing conditions, transferred to nitrocellulose filters and immunoblotted with anti-ζ mAb. Lower panel: Reblotting of the filter with polyclonal anti-CD3ε Ab.

Figure 3.

Figure 3.

Unaffected TCR synthesis but increased TCR degradation in cells with truncated. MA5.8-ζ-164 (lane 1–4) and MA5.8-ζ-103 (lanes 5–8) cells were pulse-labeled for 30 min at 37 °C with [35S]-methionine/cysteine. Cells were either placed on ice (0) or chased for 5, 10, or 20 h at 37 °C in complete medium before solubilization and immunoprecipitation with anti-CD3ε mAb. The immunoprecipitates were analyzed by SDS-PAGE under nonreducing conditions. The positions of the TCR chains are indicated (m, mature; i, immature). The position of the 32 kDa ζ-164 dimer is indicated by an asterisk. The ζ-103 dimer with a molecular weight of 25–18 kDa could not be identified in this type of experiment.

ζ truncations result in increased endocytosis rate constants of the TCR

In nonstimulated T-cells, the TCR constitutively cycles between the plasma membrane and intracellular compartments. The amount of TCR present at the plasma membrane and in the intracellular compartments is controlled by the endocytic rate constant k and the exocytic rate constant k'. At steady state, a certain amount of TCR is present at the plasma membrane and a certain amount in intracellular compartments. Increasing or decreasing either the endo- or the exocytic rate constants alters the distribution of the TCR. Decreased TCR cell surface expression is induced by increasing the endocytic rate constant, by decreasing the exocytic rate constant, or by a combination of the two. To analyze whether truncation of the ζ chain affected the kinetics of the TCR, the endo- and exocytic rate constants in MA5.8-ζ-164 and MA5.8-ζ-103 cells were examined. The MA5.8-ζ-103 cells were chosen as these cells expressed the shortest version of the truncated ζ with no internal deletions. It has been shown that the cytoplasmic tail of ζ is sensitive for intracellular proteases and that truncated forms of ζ can be found as part of fully assembled TCR complexes in some circumstances (49,50). The cells were incubated with PE-conjugated anti-CD3ε mAb for 10, 20, or 30 min at 37 °C. The cells were washed, stripped for membrane-bound mAb, and subsequently analyzed for endocytosed mAb. As previously observed, a plateau of intracellular retained mAb was not obtained using this approach, indicating that the mAb in these experiments dissociated from the TCR and accumulated inside the cells (31,48). The data showed that the endocytic rate constant of the TCR in cells with truncated ζ chain was significantly higher than the endocytic rate constant in cells with full-length ζ (Figure 4A). Using estimates from three independent experiments, the calculated endocytic rate constant k was found to be 0.0076 min−1 and 0.0270 min−1 for MA5.8-ζ-164 and MA5.8-ζ-103 cells, respectively. This gave a mean residence time at the cell surface of approximately 132 min for the TCR in cells with full-length ζ and of approximately 37 min in cells with truncated ζ.

Figure 4.

Figure 4.

Enhanced endocytic and unaffected exocytic rate constants in cells with truncated ζ. A) Endocytosis of the TCR in MA5.8-ζ-164 (triangles) and MA5.8-ζ-103 (diamonds) cells was determined as described in Materials and Methods.B) TCR cell surface expression of MA5.8-ζ-164 cells incubated with anti-TCR, mAb. MA5.8-ζ-164 cells were incubated with the indicated concentrations of biotinylated anti-TCR mAb H57-597 for 60 min at 37 °C. The cells were subsequently washed and incubated with biotinylated H57-597 at a final concentration of 10 μg/mL for 30 min at 4 °C. The cells were washed, incubated with Alexa fluor 488-conjugate mouse anti-biotin mAb, and analyzed by flow cytometry.C) Exocytosis experiments. Experimental [TCR]l(t) values for TCR in MA5.8-ζ-164 cells are given as filled squares. The calculated values of [TCR]l(t) for MA5.8-ζ-164 obtained by inserting the estimated values of the common k′ = 0.065 min−1 and the experimentally obtained values for [TCR]l(t = 0) = 1999 and [TCR]tot = 2167 into equation 2 are given as a line.D) Experimental [TCR]l(t) values for TCR in MA5.8-ζ-103 cells are given as filled squares. The calculated values of [TCR]l(t) for MA5.8-ζ-103 obtained by inserting the estimated values of the common k′ = 0.055 min−1 and the experimentally obtained values for [TCR]l(t = 0) = 588, and [TCR]tot = 764 into equation 2 are given as a line.

It has previously been shown that an increase in the TCR endocytic rate constant caused by activation of PKC did not affect the exocytic rate constant of the receptor (44). To analyze whether truncation of the ζ chain affected TCR exocytosis we next examined the TCR exocytic rates in MA5.8-ζ-164 and MA5.8-ζ-103 cells. Cells were incubated with an excess of PE-conjugated anti-TCRβ mAb at 37 °C for the time indicated. In agreement with previous studies (43) the use of the anti-TCRβ mAb did not induce TCR endocytosis during the experiment (Figure 4B) and therefore did not interfere with the interpretation of the results. To stop TCR trafficking and ensure complete labeling of surface-expressed TCR, the cells were subsequently transferred to 12 °C and incubated for 2 h with PE-conjugated anti-TCRβ. The cells were washed and the MFI of labeled TCR ([TCR]l(t)) determined. In contrast to the experiment using anti-CD3ε mAb, a staining plateau was reached in 60 min when using the anti-TCRβ mAb in accordance with previous studies (44). This indicated that the anti-TCRβ mAb did not dissociate from the TCR during TCR recycling. Thus, the increase in MFI represented unlabeled TCR (equation 1, Materials and Methods) transported from the endosomes to the cell surface during the experiment and thereby represented the size of the intracellular pool of cycling TCR as previously described (44). The experimental data demonstrated that truncation of ζ did not significantly affect TCR exocytosis. A comparable increase in TCR cell surface labeling was observed in MA5.8-ζ-164 (Figure 4C, filled squares) and MA5.8-ζ-103 (Figure 4D, filled squares) cells. In line with these results, we found that the TCR recycled equally well after PKC activation regardless of the length of ζ (data not shown). Furthermore, the data indicated that in contrast to the absolute amounts of cell surface expressed TCR, the absolute amounts of TCR in the intracellular pool of cycling TCR were equal in the transfectants regardless of the length of ζ. To get an estimate of the exocytic rate constant k′ the experimental data was fitted into equation 2. Exocytic rate constants of 0.065 min−1 and 0.055 min−1 gave the best fit for MA5.8-ζ-164 and MA5.8-ζ-103 cells, respectively. The curves obtained by inserting the estimated k′ and the experimental obtained [TCR]l(t = 0) and [TCR]tot values into equation 2(Materials and Methods) are shown as lines in Figure 4(C,D). Thus, it could be concluded that truncation of the ζ chain did not significantly affect the exocytic rate constant. From the experimentally obtained [TCR]l(t = 0) and [TCR]tot from three independent experiments it was found that on average 91% of the total cycling pool of TCR was at the cell surface and 9% was found intracellularly in MA5.8-ζ-164 cells. For MA5.8-ζ-103 cells, 77% of the total cycling pool of TCR was at the cell surface and 23% was found intracellularly. Given the estimated exocytic rate constants k' and the fractions of recycling TCR at the cell surface and intracellularly, the endocytic rate constant k was calculated as

image

where [TCR]if and [TCR]sf denote the intracellular and surface fraction of the recycling pool of TCR, respectively. According to this, the endocytic rate constants for the TCR in MA5.8-ζ-164 and MA5.8-ζ-103 cells were 0.0064 min−1 and 0.0154 min−1, respectively. This gave a mean residence time at the cell surface of approximately 156 min for the TCR in cells with full-length ζ and 65 min in cells with truncated ζ.

ζ truncations result in increased degradation rate constants of cell surface-expressed TCR

Although the expression levels of ζ as determined by Western blotting were comparable in the transfectants when normalized to the expression level of CD3ε, truncation of ζ seemed to result in lower expression of total ζ and CD3ε in parallel (Figure 2A,B). This could be caused by differences in the synthesis or the degradation rates of the chains. To analyze this, pulse-chase experiments were performed with MA5.8-ζ-164 and MA5.8-ζ-103 cells. The cells were metabolically pulse labeled for 30 min and subsequently chased for 0, 5, 10, or 20 h at 37 °C. The cells were solubilized in digitonin lysis buffer and the lysates were immunoprecipitated using anti-CD3ε mAb. The immunoprecipitates were analyzed by SDS-PAGE and labeled proteins were visualized by auto-radiography (Figure 3). This revealed that Tiαβ and the CD3 chains were synthesized at the same rates in cells with truncated ζ (MA5.8-ζ-103) and full-length ζ (MA5.8-ζ-164). The full-length ζ was clearly seen (Figure 3, lanes 1–4) whereas the truncated ζ-103 was not seen in these experiments. This was probably due to decreased labeling of the truncated ζ, the ζ-103 dimer containing only two methionines, whereas the full-length ζ-164 dimer contains eight methionines, and because the ζ-103 dimer splits up into three bands with molecular weights of 18–25 kDa (Figure 2C) that overlap with the CD3 chains. Although the CD3 and Tiαβ chains were synthesized at the same rate, the chase experiments indicated that TCR degrad-ation was increased in cells with truncated ζ as compared to cells with full-length ζ.

Degradation of the TCR as determined by metabolic pulse-chase labeling shows total degradation arising from misfolded/misassembled TCR components as well as fully assembled TCR. In an attempt to separate these degradation pathways, MA5.8-ζ-164 and MA5.8-ζ-103 cells were surface biotinylated and left for 0, 5, 10, or 20 h at 37 °C. We have recently shown that only cell surface-expressed molecules become biotinylated with this method (51); thus only fully assembled, cell surface expressed TCR were studied. The cells were lysed in digitonin lysis buffer and the lysates subjected to immunoprecipitation using anti-CD3ε mAb. The immunoprecipitated proteins were separated by SDS-PAGE and biotinylated proteins were detected using HRP-conjugated streptavidin. As shown in Figure 5, the turnover of cell surface-expressed TCR was increased in cells with truncated ζ as compared to cells with full-length ζ. The wild-type ζ chain was not as strongly labeled as the rest of the TCR chains but could be seen as a weak band (Figure 5A, asterisk). The truncated ζ could not be identified due to the weak labeling and the overlap with the CD3 chains.

Figure 5.

Figure 5.

Increased TCR degrad-ation rate constants in cells with truncated ζ. A) Western blot of surface biotinylated MA5.8-ζ-164 (lanes 1–4) and MA5.8-ζ-103 (lanes 5–8) cells. The cells were surface biotinylated, washed, and placed on ice (0) or at 37 °C for 5, 10, or 20 h. TCR chains were coimmunoprecipitated with anti-CD3ε mAb, separated on SDS-PAGE under nonreducing conditions, and transferred to nitrocellulose filters. Immuno-precipitated proteins were detected by HRP-conjugated streptavidin. The positions of the CD3 chains are indicated. The position of the 32 kDa ζ-164 dimer is indicated by an asterisk. The ζ-103 dimer with a molecular weight of 25–18 kDa could not be identified in this type of experiment.B) CD3 band intensities from three independent experiments were semiquantified using the program QuantityOne, normalized to expression at time zero, and plotted against time (diamonds, MA5.8-ζ-164 cells; squares, MA5.8-ζ-103 cells). The lines represent the calculated values for TCR degradation assuming 1. order kinetics.

We have recently demonstrated that the kinetics of the degradation of the individual chains of cell surface-expressed TCR are identical. Thus, to determine the TCR degradation rate constant, in principle it is sufficient to determine the degradation rate of only one of the TCR chains (51). The intensity of the biotinylated CD3 bands from three independent experiments was quantified and plotted against time as normalized values with the bands at time zero set to 100% (Figure 5B). By assuming first order kinetics for degradation of the CD3 chains, the degradation rate constant of the CD3 chains was estimated to be approximately 2.6 times higher in MA5.8-ζ-103 cells than in MA5.8-ζ-164 cells. Thus, the TCR degradation rate constant kd was 0.00087 min−1 in MA5.8-ζ-164 cells and 0.00229 min−1 in MA5.8-ζ-103 cells. This gave a half-life of the TCR of ∼ 13 h for cells with full-length ζ chain and ∼ 5 h for cells with a truncated ζ chain. Taking into account that MA5.8-ζ-103 cells expressed approximately three times less TCR at the cell surface than MA5.8-ζ-164 cells, this experiment indicated that the same absolute amount of TCR was degraded per time unit in the two cell lines. As the amount of degraded TCR must equal the amount of synthesized TCR during steady state, this supports the conclusion that the synthesis rate of the TCR was identical in the cell lines.

CD3γ with a defective endocytosis motif restores TCR expression in cells with truncated ζ

The results above demonstrated that cells with truncated ζ had reduced TCR expression due to an increased constitutive endocytic rate constant of the TCR in combination with an unaffected exocytic rate constant. From previous studies it has been suggested that the ζ chain masks the diL motif in CD3γ (31,47,48), and a recent study demonstrated that constitutive TCR endocytosis is dependent on the CD3γ diL motif (46). Thus, it seemed plausible that ζ is masking this motif. However, a modified diL motif is also found in CD3δ, and it could be speculated that the increased constitutive TCR endocytosis in cells with truncated ζ was caused by exposure of the CD3δ diL motif. To distinguish between these possibilities, the MA5.8-ζ-103 cells were cotransfected with GFP-tagged CD3γ containing a disrupted diL motif (CD3γ-LLAA-GFP). As a control, cells were transfected with GFP-tagged wild-type CD3γ (CD3γ-WT-GFP). If the increased constitutive TCR endocytosis and thereby the reduced TCR expression levels in cells with truncated ζ chains were due to exposure of the CD3γ diL motif, introduction of a CD3γ with a disrupted diL motif should slow down constitutive TCR endocytosis and degradation and thereby restore TCR expression. However, if the reduced cell surface expression was caused by exposure of the CD3δ diL motif, introduction of CD3γ with a disrupted diL motif should have no effect on the endocytic rate constant and consequently on the TCR expression levels. Cells were transfected, selected for expression of GFP and subsequently analyzed for TCR expression. TCR expression levels were partly restored in MA5.8-ζ-103 cells transfected with CD3γ-LLAA-GFP but unaffected in MA5.8-ζ-103 cells transfected with CD3γ-WT-GFP (Figure 6A). Cell surface biotinylation and immunoprecipitation experiments demonstrated that the CD3γ-WT- and -LLAA-GFP chains only partly replaced the endogenous CD3γ (Figure 6B). Thus, the transfectants expressed a mixture of TCR complexes containing either the endogenous CD3γ or the transfected CD3γ, as each TCR complex only comprises one CD3γ chain. This gave us the opportunity to compare the degradation rates of TCR complexes containing endogenous CD3γ and TCR complexes containing the transfected CD3γ in the same cell. Semiquantification of the endogenous CD3γ and the transfected CD3γ-GFP bands from the biotinylation pulse-chase experiments demonstrated that TCR complexes containing endogenous CD3γ or CD3γ-WT-GFP were degraded with a degradation rate constant of approximately 0.00217 min−1, whereas CD3γ-LLAA-GFP was degraded at a degradation rate constant of approximately 0.00108 min−1 (Figure 6C). Thus, TCR containing CD3γ with a disrupted diL motif were degraded at approximately the same rate in cells with truncated ζ as TCR containing an intact CD3γ in cells with full-length ζ.

Figure 6.

Figure 6.

CD3γ with a distorted diL motif restores and stabilizes TCR expression in cells with truncated ζ. A) MA5.8-ζ-103 cells were cotransfected with CD3γ-WT-GFP or CD3γ-LAA-GFP. Clones were selected for expression of GFP and subsequently examined for expression of TCR by flow cytometry using PE-conjugated anti-CD3ε mAb. Results are given as mean ± SD for five different transfectants. MA5.8-ζ-164 and MA5.8-ζ-103 cells were analyzed in parallel.B) Western blot of surface biotinylated MA5.8-ζ-164 (lanes 1–4) and MA5.8-ζ-103 (lanes 5–8) cells. The cells were surface biotinylated, washed, and placed on ice (0) or at 37 °C for 5, 10, or 20 h. TCR chains were coimmunoprecipitated with anti-CD3ε mAb, separated on SDS-PAGE under nonreducing conditions, and transferred to nitrocellulose filters. Immunoprecipitated proteins were detected by HRP-conjugated streptavidin. The positions of the CD3 chains are indicated. The position of the 58 kDa CD3-GFP chimeras is indicated by the arrow.C) The intensity of the CD3-LLAA-GFP (closed squares), CD3-WT-GFP (open diamonds), and endogenous CD3 (open triangles) bands were semiquantified using the program QuantityOne, normalized to expression at time zero, and plotted against time. The lines represent the best fits for CD3-LLAA-GFP vs. CD3-WT-GFP/endogenous CD3 degradation assuming 1st order kinetics.

This strongly indicated that the full-length ζ masks the CD3γ diL motif.

Discussion

In this study, we found that the length of the cytoplasmic tail of the ζ chain plays a crucial role in the regulation of cell surface expression levels of the TCR. TCR expression levels were inversely correlated with the length of the cytoplasmic tail of ζ. Thus, truncation of 31, 61, and 91 residues of the cytoplasmic tail of ζ led to a decrease in TCR cell surface expression of 40%, 67%, and 76%, respectively, compared to cells expressing full-length ζ. The reduced TCR expression level was caused by an increase in the endocytic rate constant combined with an unaffected exocytic rate constant. In MA5.8-ζ-103 cells expressing truncated ζ lacking the 61 C-terminal residues, the endocytic rate constant was approximately threefold higher than in cells with full-length ζ. Thus, ∼ 2.1% of the TCR on the cell surface were internalized per minute in cells with truncated ζ compared to ∼ 0.7% in cells with full-length ζ (Table 1). Our results are in good agreement with a recent study that found that the ζ chain stabilizes TCR residency at the cell surface (48). We extended the study by D'Oro et al. (48) by demonstrating that the exocytic rate constant is unaffected by the length of ζ, whereas the degradation rate constant of cell surface expressed TCR is increased in cells with truncated ζ. The degradation of cell surface-expressed TCR was increased by approximately 2.6 times, from ∼ 0.09% per minute to ∼ 0.23% per minute in cells with truncated ζ. In contrast to this, synthesis and exocytosis of the TCR seemed to be unaffected by the length of the ζ chain.

Table 1. TCR distribution, rate constants and t1/2 for TCR with full length and truncated ζ
Cells[TCR]sfa[TCR]ifbk (min−1)ct (min)dk′ (min−1)ekd (min−1)ft1/2 (h)g;
  • a

    Fraction of the recycling pool of TCR expressed at the cell surface.

  • b

    Fraction of the recycling pool of TCR inside the cell.

  • c

    Mean endocytic rate constant obtained from the endo- and exocytosis experiments.

  • d

    Mean residence time of the TCR at the cell surface,(1/k).

  • e

    Mean exocytic rate constant obtained from the endo- and exocytosis experiments.

  • f

    Degradation rate constant.

  • gHalf-life of the TCR (ln2/kd).

MA5.8-ζ-1640.910.090.00701430.0710.0008713
MA5.8-ζ-1030.770.230.0212470.0730.002295

Taking into account that only a fraction of the total cycling TCR pool is in the endosomes, the rate constant for transportation of the TCR from the endosomes to the lysosomes, kl, could be calculated as

image

Using this equation and the results summarized in Table 1, similar kl of 0.009 min−1 and 0.008 min−1 were obtained for cells with full length ζ ([0.91/0.09] × 0.00087 min−1) and truncated ζ ([0.77/0.23] × 0.00229 min−1), respectively. As the total cycling pool of TCR was approximately three times smaller in MA5.8-ζ-103 cells with truncated ζ than in cells with wild-type ζ, and as the fraction of recycling TCR found in the endosomes at the same time was 2.6 times higher in cells with truncated ζ, there would seem to be an equivalent absolute amount of TCR in the endosomes and transport to degradation from the endosomes to the lysosomes per minute regardless of the length of ζ. As the amount of degraded TCR equals the amount of newly synthesized TCR during steady state, this observation supported the results obtained from the pulse-chase experiments demonstrating that an equivalent amount of TCR was synthesized in the cell lines. Our results strongly support the hypothesis that the fate of an endocytosed TCR, to be recycled back to the cell surface or to be sorted to a lysosome, is stochastic and not dependent on the length of ζ (48). The results of the present study are summarized in Table 1 and Figure 7.

Figure 7.

Figure 7.

Model for constitutive TCR trafficking in MA5.8-ζ-164 and MA5.8-ζ-103 cells. In MA5.8-ζ-164 cells, 91% of the recycling pool of TCR is found at the plasma membrane (PM) and 9% is found in the endosomes (E). For MA5.8-ζ-103 cells, 77% of the recycling pool of TCR is found at the plasma membrane and 23% is found in the endosomes. The exocytic rate constants k′ describing trafficking from the endosomes to the plasma membrane are identical in the cell lines. Likewise, the rate constants for TCR trafficking from the endosomes to the lysosomes kl is independent of the length of the ζ chain. Only the endocytic rate constant k is affected by the length of ζ. (L, lysosomes; GA, Golgi apparatus).

Endocytosis motifs have been identified in all of the CD3 chains, including the ζ chain, in studies using isolated, often chimeric molecules (17,19,20,31). However, the only verified endocytosis motifs in completely assembled TCR complexes are found in the CD3δ and CD3γ chains. Previous studies have suggested that the ζ chain masks the diL motif present in the CD3γ chain, thereby making it inaccessible to the adaptor protein AP2 involved in the endocytosis process (31,47,48,52). In the present study, we demonstrated that this is most probably the case because introduction of a CD3γ chain with a disrupted diL motif partly restored TCR cell surface expression in cells with truncated ζ, and because TCR complexes containing the mutated CD3γ in combination with the truncated ζ was almost as stable as TCR complexes containing wild-type CD3γ in combination with full-length ζ.

One could speculate as to why the truncated ζ-103 containing 51 amino acids in its cytoplasmic tail did not mask the CD3γ diL motif located only 10–16 amino acids downstream of the transmembrane domain as efficiently as full-length ζ. However, it should be noted that not even the full-length ζ chain completely masks the CD3γ diL motif for AP2 binding. Thus, we have recently shown that in cells with full-length ζ, the CD3γ diL motif is to a small extent accessible for AP2 binding (46). We propose here that masking of the CD3γ diL motif by ζ is a dynamic process in balance with the binding of AP2 to the motif. Truncation of the cytoplasmic tail of ζ shifts this balance in favor of increased AP2 binding and thereby increased constitutive endocytosis.

The diL motif of CD3γ has been proposed to be a degradation motif as well as an endocytosis motif. Hence, when present in CD4/CD3γ chimeras the motif is responsible for rapid endocytosis and the chimeras are subsequently transported to the lysosomes for degradation (47). Therefore, it has been suggested that recycling of the TCR back to the cell surface is dependent on the inaccessibility of the diL motif (53). In the endosomes, an intact full-length ζ would mask the CD3γ diL motif and ensure recycling of the TCR, whereas a truncated ζ would not mask the motif, resulting in sorting of the TCR to the lysosomes. However, the present study demonstrated that this suggestion does not hold true. Thus, truncation of the ζ chain did not affect the recycling abilities of the TCR or the kl.

The importance of the ζ chain for T-cell responsiveness is emphasized by the fact that the ζ expression is down-regulated in a number of diseases such a cancer, leprosy and systemic lupus erythematosus, often correlated with reduced TCR cell surface expression levels (54–58). In many of these cases the loss of ζ has been linked to apoptosis and, indeed, ζ is the target of several caspases (49,50).

In conclusion, the present observations demonstrated that TCR cell surface expression levels correlate inversely with the length of the cytoplasmic tail of ζ. The reduced TCR expression levels in cells with truncated ζ were caused by an increased endocytic rate constant combined with an unaffected exocytic rate constant. Furthermore, the degradation rate constant of the TCR was increased in cells with truncated ζ. TCR expression in cells with truncated ζ was restored by introduction of a CD3γ chain with a disrupted diL motif, indicating that the ζ chain masks this motif and thereby stabilizes TCR cell surface expression.

Materials and Methods

Cells, antibodies, and chemicals

MA5.8 cells, which do not synthesize ζ (14), were cultured in RPMI 1640 medium supplemented with penicillin 2 × 105 U/L (Leo Pharmaceutical Products, Ballerup, Denmark), streptomycin 50 mg/L (Merck, Darmstadt, Germany), and 10% (v/v) fetal bovine serum (FBS) (BioWhittaker Europa, Verviers, Belgium) at 37 °C in 5% CO2. Biotin- and PE-conjugated Armenian hamster anti-mouse TCRβ (H57-597) and Armenian hamster anti-mouse CD3ε (145.2C11) mAb were from Pharmingen (San Diego, CA). Mouse anti-human ζ mAb (6B10.2) was from Santa Cruz Biotechnology (Santa Cruz, CA). Polyclonal rabbit anti-human CD3ε, which cross-reacts with mouse CD3ε, HRP-conjugated rabbit anti-mouse Ig and HRP-conjugated swine anti-rabbit Ig were obtained from DakoCytomation (Glostrup, Denmark). Alexa fluor 488-conjugate mouse anti-biotin mAb was obtained from Molecular Probes (Eugene, OR).

Constructs and transfection

All ζ mutations were constructed by polymerase chain reaction (PCR) using Vent polymerase containing 3′→ 5′ proofreading exonuclease activity (New England Biolabs, Inc., Beverly, MA) and as template pGEM-3Z-ζ.

The following upstream primer was used:

  • • hZeta-UP-Xho: 5′-AGTCAACTCGAGACCACCATGAAG-TGGAAGGCGCTTTTC-3′ introducing a Xho I site just 5′ to the ζ sequence.

The following downstream primers were used:;

  • • hZeta-s164: 5′-TCGGGTGAATTCATCCCCTGGCTGTTA-GCGAGG-3′ (corresponding to bases 577–556 and with introduction of an EcoRI site);
  • • hZeta-s133: 5′-TCGGGTGAATTCTTACCGGCGCTCGCC-TTTCATCC-3′ (corresponding to bases 470–449 with introduction of an EcoRI site and a stop codon corresponding to residue 134);
  • • hZeta-s103: 5′-TCGGGTGAATTCTTACCTTCTCGGCTTT-CCCCCCAT-3′ (corresponding to bases 380–359 with introduction of an EcoRI site and a stop codon corresponding to residue 104).

hZeta-s103Δ55-84 was constructed using LZRS-hZeta-s103 as template and the upstream primers hZeta-UP-Xho and hZeta55–84-C: 5′-CCTGAGAGTGAAGTTCTTGGACAAGAGACGTGGCCGGA-3′ and the downstream primers hZeta55–84-B: 5′-CCCGGCCACGTCTCTTGTCCAAGAACTTCCTCTCAGGAACAA-3′ and LZRS-Brother-Puro-D: 5′- CTGACTAATTGAGATGCATGCTTTGCA-3′.

The PCR products were digested with XhoI and EcoRI and cloned into the retroviral vector LZRS-Brother-Puro. The constructs were transfected into the ecotropic packaging line Phoenix-Eco (Nolan Laboratory, Stanford University, CA) using FuGENE 6 (Roche, Indianapolis, IN). In short, 3 μL of FuGENE 6 was mixed with 1 μg of the relevant vector in a total volume of 100 μL serum free medium. The mixture was added dropwise to Phoenix-Eco cells plated at a density of 5 × 105 cells per well (wells with a diameter of 35 mm) in 2 mL complete medium the day before. Stable transfectants were obtained by selection in puromycin (final concentration 3 μg/mL). For retroviral infections, MA5.8 cells were resuspended in virus-containing supernatant collected from stably transfected Phoenix-Eco cells. Polybrene (hexadimethrine bromide, Sigma-Aldrich, St. Louis, MO) was added to a final concentration of 10 μg/mL. After 48 h of culturing, puromycin was added to a final concentration of 3 μg/mL and puromycin-resistant cells were expanded and subsequently maintained in medium without puromycin.

The hCD3γ-WT and -LLAA constructs were produced as previously described by PCR using pJ6T3 γ-2 as template (19). EcoRI and BamHI restriction sites were added using the upstream primer hgam-Up-EcoRI: 5′-GTCAGAATTCCCACCATGGAACAGGGGAAGGGCCTG-3′ and the downstream primer hgam-Down-BamHI: 5′-GGTGGATCCAAATTCCTCCTCAACTGGTT-3′. The PCR product was digested with EcoRI and BamHI and cloned into the expression vector pEGFP-N1 (Clontech, Palo Alto, CA). The constructs were transfected into MA5.8-ζ-103 cells by electrophoration and G418-resistant cells were expanded and subsequently maintained in medium without G418 as previously described (19). All mutations were confirmed by DNA sequencing using an ABI 310 (Perkin Elmer).

Metabolic labeling, surface biotinylation, immunoprecipitation, Western blot and semiquantification

For metabolic labeling, cells were washed twice in PBS and resuspended at 5 × 106 cells/mL RPMI without methionine/cysteine at 37 °C for 30 min as previously described (15). After starvation, cells were washed once, resuspended at 5 × 107 cells/mL RPMI without methionine/cysteine and pulsed for 30 min with [35S]-methionine/cysteine. The cells were washed once and resuspended in RPMI containing nonradioactive methionine/cysteine at 37 °C. At the time indicated, cell aliquots were removed, washed three times in ice-cold PBS, and lysed in lysis buffer (50 mm Tris-HCl, pH 7.6, 150 mm NaCl, 1 mm MgCl2, 1% digitonin, 5 mm EDTA, 10 μg/mL leupeptin, 10 μg/mL pepstatin, 10 mg/mL pefabloc) for 30 min on ice. The lysates were precleared three times with Protein A-sepharose (PA) beads and immunoprecipitated with anti-mouse CD3ε mAb for 1.5 h followed by incubation with PA beads for another 1.5 h. After precipitation, the beads were washed five times with lysis buffer before elution of the proteins from the beads by boiling for 5 min in sample buffer (50 mm Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, 0.01% bromophenol). The immunoprecipitated proteins were electrophoresed in 8–15% gradient SDS-polyacrylamide gels. The gels were fixed for 15 min and dried. Labeled proteins were visualized by auto-radiography.

For biotinylation, cells were washed twice in PBS and resuspended in PBS containing 0.5 mg/mL Sulfo-NHS-biotin (Pierce, Rockford, IL) for 30 min at 12 °C. The cells were washed twice and resuspended at 3 × 105 cells/mL RPMI. The cells were placed at 37 °C and at the time indicated, aliquots were removed, washed, lysed, and immunoprecipitated as described for metabolic labeling. The samples were electrophoresed in 8–15% SDS-polyacrylamide gels. Following SDS-PAGE the proteins were transferred to nitrocellulose-membranes (Hybond, Amersham Laboratories, Amersham, UK) by semidry blotting. Membranes were blocked in 3% nonfat milk in PBS containing 0.2% Tween 20 for 1 h and detection of biotinylated proteins were performed using a 1 : 2000 dilution of HRP-conjugated streptavidin (Amersham). Bound HRP-conjugated streptavidin was visualized using enhanced chemiluminescence (ECL, Amersham). For semiquantification of the bands, the ECL-developed films were scanned into a computer and the intensity of the bands quantified by the program Quantity One (Bio-Rad, Hercules, CA) as previously described (51). Multiple exposure times were applied to each membrane as Quantity One only accepts films that have not been overexposed.

Measurements of TCR endocytosis rate constants

For measurement of TCR endocytosis rate constants, cells were incubated at a cell density of 1 × 106/mL at 37 °C or 4 °C with PE-conjugated anti-CD3ε mAb (145.2C11). At the time indicated, aliquots of cell suspension were washed in ice-cold PBS containing 2% FCS and 0.1% NaN3 and immediately treated with 300 μL 0.5 m NaCl, 0.5 m acetic acid, pH 2.2 for 10 s. The acid resistant fluorescence of the cells representing endocytosed anti-CD3ε mAb was measured by flow cytometry. The percentage of internalized anti-CD3ε mAb to cell surface bound anti-CD3ε mAb was subsequently calculated using the equation:;

image

where HAR is the mean fluorescence intensity (MFI) of acid-treated cells incubated at 37 °C, CAR is the MFI of acid-treated cells incubated at 4 °C, and CT is the MFI of untreated cells incubated at 4 °C as previously described (31).

Measurements of TCR exocytic rate constants

Measurements of exocytic rate constants were done as described previously (44). Briefly, MA5.8-ζ-164 and -103 cells were resuspended at 4 × 105/mL in complete medium and incubated with a final concentration of 4.8 μg/mL of the PE-conjugated anti-TCRβ mAb (H57-597) for various periods at 37 °C. The cells were washed twice in ice-cold PBS containing 2% FCS and 0.1% NaN3 and incubated with PE-conjugated H57-597 at a final concentration of 17.5 μg/mL for 2 h at 12 °C to ensure complete labeling of the cell-surface expressed TCR. The cells were subsequently washed twice in ice-cold PBS containing 2% FCS and 0.1% NaN3 and analyzed by flow cytometry to determine the MFI of labeled TCR [TCR]l(t). In these experiments, we assumed instant binding of the anti-TCRβ mAb to the surface TCR at 37 °C. The exocytic rate constant was determined by solving;

image(1)

as [TCR]tot = [TCR]ul(t) + [TCR]l(t), where [TCR]tot denotes the total number of cycling TCR represented by the MFI when the system has reached steady state, [TCR]l(t) denotes labeled TCR, and [TCR]ul(t) denotes unlabeled TCR. The solution to equation 1 is given by;

image(2)

where [TCR]l(t = 0) denotes the number of labeled TCR at the beginning of the experiment, i.e. the number of TCR at the cell surface at time 0 represented by the MFI of cells directly incubated with anti-TCRβ mAb for 2 h at 12 °C.

Acknowledgments

We thank Dr. R. D. Klausner for plasmid pGEM-3Z-ζ and Dr. T. Backstrom for plasmid LZRS-Brother-puro. The technical help of Bodil Nielsen is gratefully acknowledged.

This work was supported by The Danish Cancer Society, The Danish Medical Research Council, The Carlsberg Foundation, The Foundation of Vilhelm Pedersen and Wife by recommendation of The Novo Nordisk Foundation, The A.P. Møller Foundation for the Advancement of Medical Sciences, and The Astrid Thaysen Foundation for Basic Medical Sciences. M.v.E., C.M.B., and J.P.H.L. were recipients of Ph.D. scholarships from The University of Copenhagen.

Ancillary