|Home | About | Journals | Submit | Contact Us | Français|
The CD3 ε subunit of the TCR complex contains two defined signaling domains, a proline-rich sequence and an ITAM. We identified a third signaling sequence in CD3 ε, termed the basic-rich stretch (BRS). Herein, we show that the positively charged residues of the BRS enable this region of CD3 ε to complex a subset of acidic phospholipids, including PI(3)P, PI(4)P, PI(5)P, PI(3,4,5)P3, and PI(4,5)P2. Transgenic mice containing mutations of the BRS exhibited varying developmental defects, ranging from reduced thymic cellularity to a complete block in T cell development. Peripheral T cells from BRS-modified mice also exhibited several defects, including decreased TCR surface expression, reduced TCR-mediated signaling responses to agonist peptide-loaded APCs, and delayed CD3 ε localization to the immunological synapse. Overall, these findings demonstrate a functional role for the CD3 ε lipid-binding domain in T cell biology.
T cells recognize both self- and foreign-peptide molecules through their Ag-specific TCRs. Interactions between the αβ subunits of the TCR and peptide/MHC complexes are relayed into intracellular signals through the associated CD3 invariant chains (CD3 γ, δ, ε, and ζ). However, the mechanisms whereby extracellular ligand-binding triggers intracellular signals remain unknown. Several studies have shown that TCR engagement induces a conformational change within the cytoplasmic tail of CD3 ε, exposing a proline-rich sequence (PRS)3 that is subsequently bound by the adaptor protein Nck (1, 2). Yet, the PRS is not required for the conformational changes, as knock-in mice with a deletion of the CD3 ε PRS retain TCR-mediated signaling responses (3). Instead, these knock-in mice have increased cell surface TCR expression on immature thymocytes, and increased TCR signaling processes following low avidity TCR-ligand interactions (3).
In addition to initiating conformational changes in CD3 ε, TCR engagement also induces the phosphorylation of the ITAM motifs, present within each of the CD3 subunits (4). Following their phosphorylation, the ITAMs are complexed by Syk/ZAP-70 family of protein tyrosine kinases. This kinase family, in conjunction with Src-family protein tyrosine kinases, phosphorylates and activates multiple effector and adaptor proteins. Mutation of the CD3 ε ITAM reduces the efficiency of positive selection in low avidity TCR transgenic lines (5). This is consistent with observations that the number of TCR ITAMs available for phosphorylation directly regulates both positive and negative selection processes in the thymus (6, 7). Moreover, a minimal number of functional ITAMs within the TCR complex is needed to prevent autoimmunity, demonstrating that the regulation of ITAM phosphorylations are critical for maintaining effective T cell functions (8).
In addition to its PRS and ITAM, the CD3 ε subunit also contains a basic-rich stretch (BRS) of amino acids within the juxtamembrane portion of its cytoplasmic tail (Fig. 1A) (9). Although all three subdomains can interact with proteins, the BRS has uniquely been found to bind acidic phospholipids (1, 9–12). Recent NMR studies have suggested that such lipid interactions may embed the tyrosine residues of the CD3 ε ITAM in the inner leaflet of the plasma membrane, thereby shielding them from phosphorylation (12). This notion is consistent with in vitro studies demonstrating that phospholipid binding by the CD3 chains can limit the magnitude of their ITAM phosphorylation (12, 13). Based on these findings, a new mechanism for regulating the accessibility of the CD3 ε ITAM before TCR activation has been proposed (14). However, it should be noted that these studies all used individual CD3 ε molecules in the absence of the other TCR/CD3 subunits. Consequently, the importance of the BRS in the context of an intact TCR, as well as its role in T cell development and activation, has yet to be established. Moreover, the complexity of the phospholipids that can be bound by CD3 ε remains unclear. Herein, we report that the cluster of basic amino acids within the BRS selectively complexes a subset of charged phospholipids, including PI(3)P, PI(4)P, PI(5)P, PI(3,4,5)P3, and PI(4,5)P2. Elimination of the BRS in transgenic mice resulted in a statistically significant reduction in thymic cellularity. Furthermore, T cells from these mice had reduced TCR expression and diminished TCR-induced tyrosine phosphorylation of several signaling intermediates. Relocation of the BRS to a membrane-distal section of the CD3 ε cytoplasmic tail (following the ITAM) blocked T cell development at the CD4−CD8− stage of thymopoiesis. Finally, the expression of a CD3 ε-GFP sensor in T cells revealed a role for the BRS in TCR relocalization to the immunological synapse. Taken together, these results indicate that the cytoplasmic tail of the CD3 ε subunit contains a phospholipid-binding motif that has diverse functions in T cells.
The cDNA for murine CD3 ε was reverse transcribed using standard RT-PCR cloning procedures (Invitrogen). Mutations, truncations, and relocation of the BRS of CD3 ε were generated by PCR-based mutagenesis strategies, and confirmed by dsDNA sequencing. The various BRS-modified constructs were subcloned into pcDNA3.1, pEGFP-N1, or the VA-CD2 T cell specific transgenic cassette (15). GST-fusion proteins were generated using pGEX-2TK vectors (GE Biosciences).
For transgenic lines, the VA-CD2 plasmids containing the different CD3 ε constructs were injected into fertilized eggs from C57BL/6 mice by the Transgenic Core Facility at the University of Texas Southwestern Medical Center (UTSWMC). Transgenic founders were identified by Southern blotting and PCR procedures. Transgenic founders (at least five per construct) were back-crossed onto a CD3 ε−/− background (16). HY and 5C.C7 TCR transgenic mice have been described (17, 18). All mice were housed in the Specific Pathogen Free Facility at UTSWMC. Mouse experimentation was performed with Institutional Animal Care and Use Committee approved protocols.
The 145-2C11 hybridoma and HEK293T cells were obtained from American Type Culture Collection (ATCC). The anti-CD28 mAb secreting hybridoma was generously provided by Dr. James Allison. Anti-CD3 ε, -CD3 ζ (6B10.2), -ZAP-70 (1E7.2), and -CD28 mAbs were purified from culture supernatants using standard affinity chromatography procedures. Anti-GST, -ERK, and -TCR β (H57-597) mAbs were from BD Biosciences. Anti-phospho-linker for activation of T cells (LAT) (Y191), anti-phospho-ERK (T202/Y204), and anti-phospho-ZAP (Y319) were from Cell Signaling Technology. Anti-CD3 γ (C-20), anti-CD3 δ (M-20), and anti-CD3 ε (M-20) polyclonal antisera were from Santa Cruz Biotechnology. The various fluorochrome-labeled Abs used for flow cytometry were described elsewhere (19, 20). Peptides were synthesized by the UTSWMC Protein Chemistry Technology Center.
PIP Strips (Echelon Biosciences) were blocked in 3% fatty acid-free BSA prepared in TBST (25 mM Tris, 125 mM NaCl, 0.1% Tween 20 (pH 8.0)). The membranes were then incubated with purified GST-tagged proteins (5 μg/ml) in 1% fatty acid-free BSA/TBST overnight at 4°C. The membranes were subsequently washed and immunoblotted with anti-GST mAbs. Lipids were purchased from Avanti Polar Lipids. Liposomes were prepared by sonicating the lipids in a HEPES buffer (50 mM HEPES (pH 7.0), 100 mM NaCl, 1 mM EDTA) containing 0.5 M sucrose. The liposomes were then diluted in sucrose-free buffer. Three μM of GST or GST-BRS was incubated with 100 μg of liposomes (20/80 ratio of phospholipid to PC) for 10 min at 4°C. The liposomes were pelleted by centrifugation, washed in HEPES buffer, resuspended in SDS-sample buffer, and resolved by SDS-PAGE. For the solid-phase ELISA phospholipid binding assays, phospholipids were resuspended in methanol at 1 mg/ml (20/80 ratio of phospholipid to PC). One-hundred-microliter aliquots were coated onto EIA/RIA 96-well plates (Costar) and air dried. The wells were blocked with 10 mg/ml fatty acid-free BSA in PBS. Biotinylated peptides were added for 2 h at room temperature. After washing, peptide binding was detected with streptavidin-HRP colorimetric reactions (0.07 μg/ml).
CD3 ε sensors were generated by linking full-length CD3 ε, CD3 ε BRS-Substitute, -Truncate, or -Displace to the N terminus of GFP. The sensors were retrovirally transduced into primary, in vitro-primed 5C.C7 TCR transgenic T cells. Sensor-transduced T cells were FACS-sorted for defined sensor expression levels and live, sensor-expressing T cells were imaged in three dimensions over time during restimulation with peptide-loaded APCs. The frequencies of occurrence of specific spatiotemporal patterns of CD3 ε accumulation were determined from the imaging data. The 5C.C7 T cell culture, retroviral transduction, FACS sorting, image acquisition, and analysis procedures, including pattern definitions, are described in detail elsewhere (21).
Single cell suspensions were prepared from the thymus or lymph nodes of the different mice. The cells were analyzed for the expression of various cell surface and/or intracellular proteins by flow cytometry using FACSCalibur cytometers as described previously (20). Analysis of cytometric data was performed using both FloJo (TreeStar) and Cell Quest Pro software (BD Biosciences).
Transfections of HEK293T cells were performed using standard CaPO4/DNA precipitation methods as described (22). Forty-eight hours post transfection, cells were harvested and processed for Western blotting as described below. In certain experiments, the cells were stimulated for 10 min at 37°C with pervanadate (11 μM sodium orthovanadate, 3.8% H2O2).
Primary lymphocytes (107 cells) were stimulated with 10 μg/ml 145-2C11 for the indicated times before lysis. The cells were then washed to remove excess Ab, and lysed in a 1% Triton X-100 containing lysis buffer (pH 7.6) (20 mM Tris-HCl, 150 mM NaCl, 2 mM EDTA, 1 mM NaF, and protease inhibitors). For phospho-CD3 ζ analysis, whole cell lysates were incubated with 4 μg of anti-CD3 ε plus protein A-Sepharose beads at 4°C for 1.5 h. Immunoprecipitates were washed and then processed for Western blot analyses.
For stimulation of HY/BRS Tg/CD3 ε−/− lymphocytes, DC2.4 APCs were left untreated or loaded with 100 nM SMCY peptide for 2 h at 37°C. The APCs were then washed to remove unbound peptide. In brief, 107 lymphocytes were incubated with 3.5 × 106 APCs for 10 min at 37°C. The cells were then pelleted and processed as above for Western blotting.
In vivo stimulations were performed by i.p. injecting 200 μg of anti-CD3 ε in the indicated strains of mice. Seven days postinjection, the thymus was isolated, single cell suspensions were prepared, and aliquots were analyzed by flow cytometry.
Primary thymocytes were harvested from the indicated mice and resuspended in PBS containing 0.5 mg of EZ-Link NHS-LC-Biotin (Pierce) for every 1 ml of sample. The cells were incubated at 4°C for 20 min, washed twice in PBS containing 5% FBS, washed twice with PBS alone, and lysed in the 1% Triton X-100 lysis buffer described above. Immunoprecipitations of CD3 ε were performed as described above. Detection of biotin-labeled proteins was performed using streptavidin-HRP (Pierce).
The cytoplasmic tail of CD3 ε contains a cluster of positively charged residues within its juxtamembrane position (Fig. 1A). This BRS enables CD3 ε to associate with certain phospholipids, including phosphatidylglycerol (11, 12). However, a comprehensive profiling of the phospholipids complexed by the BRS and the role of these interactions in TCR-mediated functions remains unknown. To identify which phospholipids are bound by CD3 ε, membranes embedded with an assortment of lipids (PIP strips) were probed with GST-fusion proteins containing either the full-length cytoplasmic tail of CD3 ε (GST-ε) or the different subdomains of CD3 ε (BRS, PRS, or ITAM). GST-ε and GST-BRS bound to several monophosphorylated lipids, including phosphatidylinositol 3-phosphate (PI(3)P), phosphatidylinositol 4-phosphate (PI(4)P), phosphatidylinositol 5-phosphate (PI(5)P), and phosphatidic acid (PA) (Fig. 1A). Weaker binding occurred between the BRS and phosphatidylinositol 3,4-bisphosphate (PI(3,4)P2), phosphatidylinositol 4,5-bisphosphate (PI(4,5)P2), and phosphoinositol 3,4,5-trisphosphate (PI(3,4,5)P3). The PRS and the ITAM failed to associate with any phospholipids, demonstrating that the BRS can independently mediate lipid interactions.
To determine whether the phospholipid binding involved the cluster of lysine and arginine residues in the BRS, PIP strips were probed with GST-BRS fusion proteins that lacked five positively charged residues by either amino acid substitution or truncation (BRS-Substitute and BRS-Truncate) (Fig. 1A). Both modifications eliminated phospholipid binding, demonstrating the importance of the basic amino acids in these interactions (Fig. 1A).
We next examined whether the BRS could complex phospholipids present in a lipid bilayer. Liposomes containing 100% phosphatidylcholine (PC) or 80% PC and either 20% of PI, PI(4)P, or phosphatidylserine were incubated with GST or GST-BRS. Only the BRS-containing fusion protein precipitated with charged phospholipids, including PI(4)P, PI(3)P, and PI(4,5)P2 (Fig. 1B and supplemental Fig. 1).4 This binding was specific because the BRS did not associate with liposomes containing PC or a mix of PC with phosphatidylserine. To exclude the possibility that the GST portion of the fusion protein influenced lipid interactions, ELISA-based binding assays were performed using a biotinylated peptide consisting of just the BRS. The BRS peptide associated with PI(4)P and PI(4,5)P2 in a dose-dependent manner, while no binding was observed with PC. A control peptide containing the phosphorylated ITAM of CD3 ζ did not bind any phospholipids (Fig. 1C). The affinity of the BRS for PI(4)P and PI(4,5)P2 were similar in this assay, with a calculated Kd in the range of 40 nM (data not shown). Taken together, these experiments suggested that the BRS has a high affinity for particular phospholipids.
The ability of the CD3 ε BRS to complex charged phospholipids suggested an important role for this domain in T cells. To examine the role of the BRS within the context of an intact TCR, we generated CD3 ε transgenic lines containing three distinct modifications of the BRS (Fig. 2A). Two mutants (BRS-Substitute and -Truncate) were designed to eliminate phospholipid binding (Fig. 1A). In the third mutant, the BRS was relocated to the membrane-distal portion of the cytoplasmic tail of CD3 ε, following the ITAM (BRS-Displace). Transgenic mice expressing a wild-type version of CD3 ε were also generated to control for potential effects caused by forced expression of CD3 ε from a transgene (BRS-Wild Type). All the transgenic lines were back-crossed with CD3 ε knockout mice (which retain CD3 γ and δ) to eliminate endogenous protein expression, and several lines were selected for similar levels of transgene-driven CD3 ε expression (supplemental Fig. 2) (5, 16). Thymopoiesis in the BRS-Wild Type transgenic lines was relatively normal (Fig. 2B). The BRS-Substitute and -Truncate transgenic mice also exhibited normal percentages of CD4+CD8+ double positive (DP) cells, with a slight increase in the percentage of single positive (SP) thymocytes (Fig. 2B and supplemental Fig. 3). In contrast, thymocytes from the BRS-Displace mice were arrested at the CD4−CD8− double negative (DN) stage (Fig. 2B). Analysis of Thy1.2+ DN thymocytes in the BRS-Substitute and BRS-Truncate lines did not reveal any statistically significant differences in the percentages of cells progressing to the DN4 (CD25−CD44−) stage relative to BRS-Wild Type controls (supplemental Fig. 4). Conversely, the BRS-Displace transgenic lines had an almost complete block at the DN3 (CD25+CD44−) stage of thymopoiesis (four independent founders) (supplemental Fig. 4).
All the BRS-modified transgenic lines had a statistically significant reduction in thymic cellularity, with the numbers in the BRS-Displace lines resembling CD3 ε-deficient mice (Fig. 2C). Staining these thymocytes for markers of apoptosis revealed a 1.5- to 2-fold increase in the number of apoptotic cells, indicating that the reduced thymic cellularity in these mice may be partially due to an increase in cell death (Fig. 2D). On occasion, a small percentage (20–50%) of the BRS-Displace thymocytes progressed to the DP stage (supplemental Fig. 5; example = mouse 1). However, this did not restore normal thymic cellularity (Fig. 2C).
To more carefully define a role for the BRS in thymocyte selection, the BRS transgenic lines were mated with HY TCR transgenic mice. This transgenic line expresses an αβ TCR that recognizes the male specific SMCY peptide (17, 20). Both positive and negative selection processes were intact and relatively equivalent in the various HY/BRS-Wild Type, -Substitute and -Truncate female and male mice, respectively (supplemental Fig. 6, A and B). Interestingly, the thymic cellularity of female HY/BRS-Substitute and -Truncate mice was normal when compared with the HY/BRS-Wild Type controls (supplemental Fig. 6C). Taken together, these findings suggest that the BRS regulates an early step in T cell development as thymocytes expand in number.
The PRS of CD3 ε has been shown to regulate TCR expression, but only on immature DP thymocytes (3). When comparing thymocytes from the BRS-Wild Type, -Substitute, and -Truncate lines, we noted a significant reduction in the mean fluorescence intensity (MFI) of CD3 ε not only on developing DP thymocytes, but also on more mature CD4+CD8− and CD4−CD8+ SP thymocytes (Fig. 3A). The BRS-Displace lines exhibited the most severe reduction in CD3 ε MFIs, primarily due to the limited thymopoiesis in these mice.
Examination of peripheral T cells revealed that all the BRS-mutant lines maintained their statistically significant decrease in the MFI of CD3 ε after they exited the thymus (Fig. 3B). This was also true for the MFI of surface TCR β (data not shown). Importantly, this reduction in TCR surface expression was not dependent on protein levels, as these mice possessed similar CD3 ε protein expression as BRS-Wild Type controls (supplemental Fig. 2). Consistent with the requirement for the BRS in maintaining normal TCR expression levels, peripheral T cells from the HY/BRS-Substitute and HY/BRS-Truncate double transgenic female mice also exhibited a significant reduction in surface TCR expression (supplemental Fig. 6B).
A comparison of peripheral T cell subpopulations revealed that the BRS-Substitute and -Truncate lines had a slight increase and decrease in the percentage of CD8+ and CD4+ T cell subsets, respectively (Fig. 3C). Conversely, the BRS-Displace transgenic lines possessed very few CD4+ or CD8+ T lymphocytes in their peripheral lymphoid organs (average <5%) (Fig. 3C and supplemental Fig. 5). Taken together, our results suggest that the presence and/or location of the BRS within the cytoplasmic tail of CD3 ε is necessary for maintaining normal TCR expression levels.
Upon T cell/APC contact, the TCR rapidly localizes to a structure at the center of the T cell/APC interface known as the central supramolecular activation cluster (cSMAC) (Fig. 4) (23). In primed 5C.C7 TCR transgenic T cells interacting with peptide-loaded APCs, the cSMAC is enriched in proximal signaling intermediates, such as ZAP-70, LAT, Itk, and PLC γ, as well as in more distal signaling intermediates, including PKC θ, Rac, and Rho (21). Thus, in these T cell/APC couples, the cSMAC is a site of active signaling. To explore whether mutagenesis of the BRS affected CD3 ε relocation to the immunological synapse, we performed live cell imaging studies using GFP-fusion proteins containing the BRS-Wild Type, -Substitute, -Truncate, or -Displace versions of CD3 ε. These constructs were retrovirally transfected into primed 5C.C7 TCR transgenic T cells. The cells were then subjected to immunofluorescent imaging with APCs bearing their cognate ligand to determine the spatiotemporal pattern of the CD3 ε-chains. Upon T cell/APC couple formation (t = 0 s), the BRS-Wild Type CD3 ε-chain exhibited a rapid translocation to the T cell/APC interface in up to 80% of cells analyzed (Fig. 4; Any Accumulation at Interface). The majority of CD3 ε-GFP localized to the center of the interface and persisted there for up to 7 min (Fig. 4 and supplemental Movie 1). All of the BRS mutants showed distinct defects in the recruitment of CD3 ε to the T cell/APC interface, particularly to its center (Fig. 4 and supplemental Movies 2–4). The CD3 ε BRS-Substitute and -Truncate proteins exhibited a 20–40 s delay in central accumulation, as well as a significant reduction in the overall percentage of T cells exhibiting this accumulation pattern (29 and 35%, respectively) (Fig. 4). The BRS-Displace CD3 ε-chain exhibited reduced interface recruitment at all time points. These results established that the BRS contributes to the recruitment of CD3 ε to the T cell/APC interface, particularly to its center where numerous downstream signaling intermediates are located.
Recent findings have shown that the cytoplasmic tail of CD3 ε is latterly sequestered along the inner leaflet of the plasma membrane via phospholipid interactions (12). Based on these findings, a signaling model has been proposed wherein TCR engagement induces a conformational change through CD3 ε, disrupting the CD3 ε-phospholipid interactions at the plasma membrane so that the ITAM of CD3 ε becomes tyrosine phosphorylated (14). In the context of this model, eliminating the positively charged residues of the CD3 ε BRS should enhance the tyrosine phosphorylation of the ITAM because it would be constitutively exposed to Src kinases. To test this hypothesis, we coexpressed the full-length BRS-wild type, -substitute, -truncate, or displace versions of CD3 ε into HEK293T cells with the tyrosine kinase, Lck (Fig. 5A) (24). CD3 ζ was coexpressed with Lck as a positive control (lane 1). Phosphotyrosine immunoblotting revealed that CD3 ζ and all the CD3 ε constructs lacking the basic amino acids (CD3 ε BRS-Substitute and BRS-Truncate) were easily detected as tyrosine phosphorylated proteins (lanes 1, 3, and 4). Yet, the two CD3 ε constructs that retained the BRS (CD3 ε BRS-Wild type and BRS-Displace) were not tyrosine phosphorylated (lanes 2 and 5). When the transfected cells were treated with pervanadate to inhibit intracellular phosphatases, all of the BRS constructs were tyrosine phosphorylated (Fig. 5B). Taken together, these findings fully support the model that electrostatic interactions between basic amino acids in the CD3 ε BRS and acidic phospholipids limit the ITAM phosphorylation, at least until the TCR is engaged (14).
To examine the validity of this model using cells expressing an intact TCR, lymphocytes from the BRS-Wild Type, -Substitute, and -Truncate lines were analyzed for phosphoprotein content both before and after TCR cross-linking. Because the BRS-Displace mice generally lacked T cells, these mice were not included in the study. There was no consistent evidence of hyperphosphorylation of the CD3 ε ITAM either before or after TCR engagement in the BRS-mutant transgenic lines (Fig. 5C). These findings are very different compared with the phosphorylation state of monomeric CD3 ε expressed in HEK cells. Furthermore, similar levels and kinetics of tyrosine phosphorylation of other signaling molecules was also noted when comparing the various BRS-modified lines (CD3 ζ, ZAP-70, and ERK) (supplemental Fig. 7). Consequently, no statistical differences were observed in the up-regulation of activation markers (CD69 and CD25), production of several cytokines, and T cell proliferation following anti-CD3 treatment (data not shown).
Because subtle differences in TCR signaling pathways in T cells containing the BRS-modifications could have been masked by the use of anti-TCR cross-linking mAbs, we next examined BRS-mutant TCR signaling responses to peptide-loaded APCs. Experimentally, peripheral T cells from the HY/BRS double-transgenic females were incubated for 10 min with dendritic cells that were either untreated or had been loaded with the SMCY peptide. Analysis of phospho-ζ and phospho-LAT revealed a marked decrease in the phosphorylation of these proteins in HY/BRS-Substitute and -Truncate cells relative to the HY/BRS-Wild Type controls (Fig. 6). These findings suggest that the BRS may contribute to early TCR signaling events. Analysis of later time points (30 and 60 min) revealed variations in the level of phosphoprotein induction, with the HY/BRS-Substitute and -Truncate lines exhibiting both increases and decreases in the levels of phospho-CD3 ζ in comparison to wild-type controls in different assays (data not shown). Yet, the 10-min time point was consistently reduced in the BRS-modified lines compared with the wild-type controls in these additional experiments. The inconsistent differences noted at the later time points might reflect alterations in TCR internalization, degradation, and/or recycling that might be regulated by the BRS. The defects in early signaling events did not translate into altered T cell responses, such as proliferation (data not shown).
The inability of the BRS-Displace transgenic lines to support early T cell development suggested that these mutants had a severe signaling defect. This could be accounted for by the delayed and extremely poor recruitment of the BRS-Displace CD3 ε protein to the cSMAC. Alternatively, inefficient pre-TCR assembly in the BRS-Displace lines could block the progression of thymocytes from the DN3 to DN4 stage. To examine these possibilities, CD3 ε pairing with the other TCR/CD3 subunits was analyzed in BRS-Displace thymocytes. Immunoprecipitation experiments demonstrated that the BRS-Displace CD3 ε-chain paired with both CD3 γ and CD3 δ in whole cell lysates, as well as with TCR β that was isolated from the cell surface (Fig. 7, A and B). To further verify that the BRS-Displace CD3 ε-chain was present at the cell surface, thymocytes from RAG−/−, BRS-Displace, and CD3 ε knockout mice were biotinylated and lysed. Both CD3 ε and CD3 γ were biotinylated when extracted from RAG-deficient and BRS-Displace mice, but not from CD3 ε−/− mice (Fig. 7C). These results indicate that the BRS-Displace CD3 ε-chain can form a cell surface expressed pre-TCR.
Finally, we analyzed whether the surface expressed CD3 ε-chain within the BRS-Displace lines could transduce intracellular signals. Experimentally, BRS-Displace mice were given intra-peritoneal injections of anti-CD3 ε. Using such procedures, thymocytes from the RAG-deficient mice underwent a 10- to 50-fold expansion and up-regulated both CD4 and CD8 (Fig. 7D) (25). The BRS-Displace mice exhibited no increase in thymic cellularity and no differentiation of developing thymocytes to the DP stage (Fig. 7D). The findings confirmed that the BRS-Displace CD3 ε-chain is signaling deficient.
The cytoplasmic tail of CD3 ε contains three distinct protein interaction domains: a BRS, a PRS, and an ITAM. Previous studies have indicated that the BRS may contribute to T cell functions through interactions with signaling proteins, such as GRK2 and CAST (9, 26). The unstructured clusters of positively charged residues present within the BRS of CD3 ε has also been shown to bind certain phospholipids that are expressed on the inner leaflet of the plasma membrane (12). Herein, we have demonstrated that BRS/phospholipid interactions are highly versatile, with the BRS capable of binding phospholipids present both at the plasma membrane (PI(4)P, PI(4,5)P2, and PI(3,4,5)P3) and on different intra-cellular organelles (PI(3)P, PI(4)P, and PI(5)P) (27).
The cytoplasmic tail of CD3 ε undergoes a ligand-induced conformational change that precedes ITAM phosphorylations (1, 28). However, the mechanism whereby conformational changes are projected through CD3 ε has remained elusive. Recent experiments have demonstrated that lipid binding by the BRS may position the tyrosine residues of the CD3 ε ITAM so they can embed in the plasma membrane (12). This result provides a potential mechanism whereby ITAM phosphorylations are inhibited by phospholipid interactions until TCR triggering releases the cytoplasmic tail of CD3 ε from the plasma membrane (14). Consistent with this model, we found that eliminating the positively charged residues of the BRS significantly augmented the phosphorylation of the CD3 ε ITAM, but only when expressed as a monomer in heterologous cells. Transgenic mice bearing mutations within their BRS exhibited a significant reduction in thymic cellularity, partly due to an increase in thymocyte apoptosis. This finding could relate to the hypothesis that the BRS binds to many distinct phospholipids that are involved in synapse formation, endosomal recycling, and TCR degradation, both before and subsequent to TCR/ligand interactions. Disruption of these pathways could lead to increased cell death noted in the developing thymocytes. In the HY/BRS double transgenic mice, the thymic cellularity was comparable and not affected by the BRS-modifications. This may be due to several factors, including transgene driven overexpression of the HY TCR and/or the low avidity nature of the HY TCR. Thus, pre-TCR signals could be more sensitive to the BRS-modifications than TCR-driven signals. The HY TCR transgene, which forces TCR expression at the CD4−CD8− stage of thymopoiesis, may obviate the need for the pre-TCR in the HY/BRS mice.
Interestingly, in the context of an intact TCR, the CD3 ε subunit was not consistently hyperphosphorylated. Rather, eliminating or relocating the clustered lysine and arginine residues within the BRS primarily served to reduce the cell surface expression of the TCR on both developing and peripheral T cells. Mechanistically, this could result from an inability of the BRS-modified CD3 ε-chains to bind phospholipids expressed on intracellular compartments, such as PI(3)P (early endosomes) or PI(4)P (the Trans-Golgi network) (27). Such disruptions could lead to a failure of the TCR to recycle properly between the plasma membrane and endosomes, or in an inability of the TCR to translocate to the cell surface following ER assembly. This may reflect a role for the BRS in binding select phospholipids to properly localize within a T cell. Consistent with this hypothesis, BRS-modified CD3 ε-chains showed a diminished capacity to properly translocate to the immunological synapse following APC interactions. This phenotype is consistent with the fact that TCR-engagement causes transient increases in the levels of PI(4)P and PI(4,5)P2 at the plasma membrane (29, 30). Such increases could contribute to redistribution and stabilization of the TCR at the cSMAC.
In terms of proximal signaling events, BRS-mutant lymphocytes only exhibited consistent defects in their signaling capacity when stimulated for short time periods with peptide-loaded APCs (10 min). There was much more variability at later time point. Such findings may reflect discrepancies in the ability of the TCR to undergo proper recycling in the BRS-modified lines following ligand binding. Thus, while the TCR signals were weaker at early time points, inefficient TCR turnover, recycling, or degradation at later time points could have contributed to prolonged or equivalent signaling after 30–60 min of stimulation. This could arise from the absence of BRS interactions with selected phospholipids located on endosomal compartments, such as PI(3)P.
Relocating the BRS to a membrane-distal position had the most pronounced effect on T cell development. The simplest interpretation for this effect is that displacement of the BRS prevented the regulated binding of CD3 ε to phospholipids at diverse intracellular locations, thereby significantly interfering with pre-TCR signaling. Alternatively, the topology of the CD3 ε subunit could have been affected by displacement of the BRS. Positively charged residues located downstream of a hydrophobic transmembrane segment regulate the proper orientation of transmembrane proteins as they pass through the endoplasmic reticulum (31). Displacing the basic residues of a transmembrane protein up to 30 amino acids away from its hydrophobic segment can invert the NH2-domain of a protein (31, 32). Although the BRS-Displace CD3 ε-chain associated with CD3 γ, CD3 δ, and TCR β, the topology of CD3 ε-chain could have been inverted, preventing effective pre-TCR signaling. Furthermore, the BRS-Displace construct resulted in the repositioning of the PRS and ITAM, which may have also contributed to a more severe phenotype.
In summary, we have demonstrated that the polybasic cluster of amino acids in the cytoplasmic tail of CD3 ε complexes acidic phospholipids. Interestingly, there are a number of additional ITAM (CD3 ζ and FcεRIγ) and non-ITAM containing transmembrane proteins (CD43, CD44, and ICAM-1/2) that also contain polybasic motifs in their cytoplasmic tails. These findings suggest that the function and/or assembly of many transmembrane receptors could be dynamically regulated by changes in protein/phospholipid interactions.
We thank John Ritter and his staff from the UTSWMC Transgenic and Knock-out Facility. In addition, we appreciate scientific discussions and insights from Drs. Helen Yin, Hans Deisenhofer, Sandra Hayes, and Melanie Cobb. We thank Angela Mobley for assistance with flow cytometry.
1This work was supported in part by grants from the National Institutes of Health T32 AI005284 (to A.B. and L.D.), AI42953, and AI71229 (to N.S.C.v.O.).
3Abbreviations used in this paper: PRS, proline-rich sequence; PI(3)P, phosphatidyl-inositol-3-phosphate; PI(4)P, phosphatidylinositol 4-monophosphate; PI(5)P, phosphatidylinositol 5-phosphate; PI(4,5)P2, phosphatidylinositol 4,5-bisphosphate; BRS, basic-rich stretch; PC, phosphatidylcholine; DN, CD4−CD8− double negative; DP, CD4+CD8+ double positive; SP, CD4+ or CD8+ single positive; LAT, linker for activation of T cell; MFI, mean fluorescent intensity.
4The online version of this article contains supplemental material.
The authors have no financial conflict of interest.