|Home | About | Journals | Submit | Contact Us | Français|
The mechanism underlying the transient accumulation of CD4 at the immunological synapse (IS) and its significance for T cell activation are not understood. To investigate these issues, we mutated a serine phosphorylation site (S408) in the cytoplasmic tail of murine CD4. Preventing phosphorylation of S408 did not block CD4 recruitment to the IS; rather, it blocked the ability of CD4 to leave the IS. Surprisingly, enhanced and prolonged CD4 accumulation at the c-SMAC had no functional consequence for T cell activation, cytokine production, or proliferation. PKCθ-deficient T cells also displayed enhanced and prolonged accumulation of wild type CD4 at the IS, indicating that θ is the critical PKC isoform involved in CD4 movement. These findings suggest a model wherein recruitment of CD4 to the IS allows its phosphorylation by PKCθ and subsequent removal from the IS. Thus, an important role for PKCθ in T cell activation involves its recruitment to the IS, where it phosphorylates specific substrates that help to maintain the dynamism of protein turnover at the IS.
Antigen recognition involves the interaction between TCRs and pMHC complexes on antigen-presenting cells. The discovery of the co-receptor function of CD4 and CD8 molecules in T cell development and function was a major advancement in the field (1, 2). Subsequently, many features of co-receptor structure, biology, and biochemistry have been elucidated (3–9), with a key attribute being the ability of CD4 and CD8 molecules to bind Lck (10–12). The formation of the immunological synapse (IS) is highly correlated with T cell activation (13, 14). TCR and pMHC are clustered in the center of the contact area, termed the c-SMAC, which is surrounded by the p-SMAC, comprised predominantly of LFA-1 and ICAM. The IS was originally viewed as a stable structure once it was formed (13, 14), but it is now known to be a much more dynamic structure, depends upon the strength of the TCR:pMHC interaction (15, 16).
Despite numerous studies, the precise role of CD4 in IS formation and function is not resolved. With CD4 being able to bind MHC class II molecules on the APC, and bring Lck into the IS, the initial models had CD4 persisting in the c-SMAC (17). This view changed with live cell imaging of CD4 movement. Krummel and Davis, using a T cell clone and transduced CD4-GFP, found a transient accumulation of CD4 at the IS, where it accumulated at the c-SMAC for a few minutes, after which it moved to the p-SMAC (18). Lck was observed to move in a similar fashion (19), and in conjunction with mathematical modeling, the authors proposed that CD4 enhances signaling by coordinating Lck accumulation at the IS in a zero-order ultrasensitivity reaction (20). Zal and Gascoigne using T cell hybrids and FRET analysis, observed that CD4 was recruited to the IS, but not necessarily in an antigen-specific manner (21). They proposed that it was the antigen-specific recruitment of the TCR and its close proximity to CD4 that initiated T cell activation, with CD4 playing a passive, supportive role instead of an active, dynamic role (21). Thus, it is clear that CD4 movement is a dynamic process at the IS, but how and why CD4 moves in and out of the IS is not known.
In this study, we tested the hypothesis that serine phosphorylation of the cytoplasmic tail of CD4 was critical for its movement at the IS. Initial interest in phosphorylation of the cytoplasmic tail of CD4 involved HIV entry into human T cells. It was shown that mutation of the main phosphorylation site, serine 408, when mutated to an alanine, severely impaired the ability of CD4 to be endocytosed, but did not affect HIV infection (22). A subsequent series of studies have shown that the S408 residue is phosphorylated upon stimulation with either antigen or phorbol esters (23, 24), exposing a di-leucine motif, releasing Lck (25–29), and followed by clathrin-mediated endocytosis of CD4 (30, 31). Despite the extensive analysis, essentially nothing is known about the role of serine phosphorylation of CD4 in its movement at the IS and T cell signaling and function.
We report that, in mouse primary T cells, mutation of serine 408 to an alanine (termed S408A) had no effect on the recruitment of CD4 to the IS, but did lead to a prolonged and centralized CD4 accumulation at the c-SMAC. Surprisingly, this centralized localization of CD4 did not enhance T cell responses to strong or weak agonists. T cells deficient in PKCθ had the same centralized CD4 accumulation as the S408A mutation. These findings support the model that PKCθ and CD4 are both recruited to the IS, PKCθ phosphorylates CD4, leading to the elimination of phospho-CD4 from the IS.
Mice expressing the N3.L2 TCR, which recognize Hb(64–76)/I-Ek, has been described previously (32). N3.L2 CD4-deficient T cells (termed CD4KI) were obtained using radiation chimeras in which bone marrow cells from N3.L2 Rag-deficient, CD4-deficient mice were used to reconstitute host mice expressing the antagonist ligand, I72: CD4KI T cells produced under these conditions were functional (33). All mice were bred and housed under specific pathogen-free conditions in the animal facility at the Washington University School of Medicine, and followed protocols that were approved by the Washington University Animal Studies Committee.
Altered peptide ligands were synthesized, purified, and analyzed as described previously (34). The amino acid sequences for the peptides used in this study are: agonist [Hb(64–76)], GKKVITAFNEGLK; weak agonist (T72), GKKVITAFTEGLK, and antagonist (I72), GKKVITAFIEGLK.
The retroviral construct, GFP-RV, was a kind gift from Ken Murphy. Mouse CD4 cDNA was generated by PCR using the bA-L3T4-BssHII-neo plasmid provided by Christophe Benoist. 1) Creation of CD4 and S408A non-fusion proteins. Mouse CD4 and S408A cDNA was generated by PCR or site-directed mutagenesis (T →G to generate the serine to alanine mutation) and cloned into the Bgl II/Xho I site upstream of the internal ribosome entry site (IRES). 2) Creation of CD4-YFP and S408A-YFP. After removal of the stop codon from CD4 or S408A, enhanced YFP (Clontech) was attached in-frame to the C-terminus via a five-amino acid linker (GGAAS) (21), and cloned into the Bgl II/Xho I site of a retroviral construct in which GFP has been removed. 3) Creation of CD3-ζ-CFP. CD3-ζ was excised from the CD3-ζ-GFP construct (gift from Mark Davis), and cloned in-frame with enhanced CFP (Clontech) at the C-terminus. Both fusion and non-fusion CD4 proteins exhibited identical functional properties in response to PMA when expressed in T cell hybrids (data not shown). All constructs were purified using the Endofree Plasmid Kit (Qiagen) prior to transfection into packaging cell lines.
The Platinum-E packaging cell line (gift from Dr. T. Kitamura) was transfected with 25–30 μg of retroviral construct DNA with Lipofectamine 2000 (Invitrogen), and viral supernatant was collected 48 hours after transfection. To obtain purified CD4KI T cells, splenocytes from CD4KI mice were harvested and depleted of CD4+, CD8+, B220+, MHC II+, CD11b+, and CD11c+ radioresistant cells using MACS Microbeads (Miltenyi Biotec). To obtain primary CD4+ N3.L2 T cells, cells were purified using CD4 MACS Beads (Miltenyi Biotec). For T cell activation, 3–5 × 105 T cells were stimulated with 6.5 × 106 irradiated B6.K splenocytes loaded with 10 μM Hb(64–76). At 20 and 24 hours after activation, retroviral supernatant was added to the T cell cultures and spun for 45 minutes at 1800 rpm at 25°C in the presence of Lipofectamine 2000 (Invitrogen) and 125 U/ml IL-2. At days 6–8 after activation, T cells were sorted for equivalent levels of CD4 either via GFP (for CD4 non-fusion proteins) or YFP (for fusion proteins) using the FACSVantage flow cytometer (BD Biosciences) at the Washington University Department of Pathology and Immunology Cell Sorting Facility. For transduction of T cells with two different retroviral constructs, supernatant from individually packaged PLAT-E cells were collected, mixed together, added to T cells, and treated as described above. CD4-deficient 3.L2 T cell hybrids were generated as described (35). For stable expression of CD4 or S408A, the T cell hybridomas were retrovirally transduced with either CD4 or S408A and sorted for equivalent expression levels.
One × 105 of 3.L2-CD4 or 3.L2-S408A T cell hybridoma cells were incubated in each of triplicate wells of a 96 well plate with 10 ng/ml of PMA in a total volume of 200 nl and incubated at 37°C for times ranging from 15 min to 8 hrs. The cells were fixed, stained for CD4 or S408A expression (GK1.5-biotin, streptavidin-cychrome) and analyzed by FACS. The level of CD4 or S408A expression for each well was determined as a percentage of the MFI of the unstimulated T cell hybridomas.
Confocal and differential interference contrast (DIC) images were acquired using the Zeiss LSM 510 microscope (Zeiss, Thornwood, NY) fitted with a 1.3 NA 40X Fluar objective. Live cell imaging of primary T cells transduced with either CD4-YFP or S408A-YFP were done in temperature-controlled Delta T dishes (Bioptechs) using imaging buffer (1% HSA, 1 mM CaCl2, 2 mM MgCl2 in HEPES Buffered Saline, pH 7.4). CH27 cells loaded with 10 μM Hb(64–76) were attached to surfaces coated with anti-H-2Dd (Biolegend) and used as APCs. No significant signal saturation was observed in any of the images used for analysis. To make three dimensional reconstructions, 15 DIC and YFP images spaced 0.66 μm apart in the z-plane were acquired every 15 seconds. For quantitation of cell couples and determination of accumulation indices, unprocessed images were scored using Image J software (NIH, Bethesda, MD) (36). The accumulation index is a ratio based on the mean fluorescent intensity at the IS divided by the mean fluorescent intensity from the rest of the membrane minus background (21). For two-color experiments, CFP and YFP emission was collected using CFP and YFP emission filters, respectively.
For fixed conjugates, T cells were added to CH27 cells loaded with 10 μM Hb(64–76), allowed to interact for the indicated times, and fixed with 4% paraformaldehyde for 20 minutes. CD4 staining was done using the antibody, GK1.5, followed by goat anti-rat Ig Alexa 546 antibody (Molecular Probes). For staining of intracellular molecules, conjugates were also permeabilized with 0.1% Triton X-100, and stained with either pY394 Lck (15) or Lck (sc-013, Santa Cruz Biotechnologies), followed by donkey anti-rabbit Ig Alexa 555 antibody (Molecular Probes).
Statistical analysis was performed using GraphPad Prism 4.0 software. A Student’s t-test with two-tailed distribution was used to determine statistical significance between any two groups, with a p value < 0.05 being considered statistically significant.
For proliferation assays, 1.0 × 105 purified CD4KI T cells were incubated with 5 × 105 irradiated B6.K splenocytes loaded with increasing amounts of peptide for 48 h, pulsed with 0.4μCi [3H]thymidine, and harvested 18–24 hours later. Where CD4 dependency was assessed, T cells were also incubated with 10μg/ml anti-CD4 (GK1.5) during the course of the assay. For measurement of IL-2 production, one fourth of each well was collected 20 hours after antigen stimulation, transferred to the IL-2 indicator cell line, CTLL-2 for 24 hours, pulsed with 0.4μCi [3H] thymidine for 18–24 hours, and harvested.
T cell hybridoma assays were performed as previously described (35). The activation of T cell hybridomas was assessed by measuring IL-2 production by a quantitative ELISA. Briefly, Immulon-2HB 96 well plates were coated with 100μg/ml of a capture anti-IL-2 monoclonal antibody (JES6-1A12, Biolegend) overnight at 4°C. The plates were washed and blocked (PBS, 0.5% BSA, 0.1% Tween-20). Dilutions of the samples and IL-2 standards (Peprotech) were added to triplicate wells and incubated at 25°C for 2 hrs. The plates were washed, and biotin-anti-IL-2(JES6-5H4, Biolegend) at 0.5μg/ml was added to each well, incubated at 25°C for 1 hr, washed, 100nl/well of streptavidin-HRP (Southern Biotech, 1:10,000) was added, incubated at 25°C for 30 min, and washed. 100μl of 1-Step-Ultra TMB substrate (Thermo Scientific) was added to each well, the reaction was stopped with 2M sulfuric acid after 15 min, and the product was quantified using a Victor3 plate reader (Perkin Elmer). The concentration of IL-2 in each well was calculated from the IL-2 standard curve.
For calcium imaging, T cells were loaded for 20 minutes at 37°C with 1μg/ml Fura-2-AM (Molecular Probes), rested for 20 minutes at 37°C, and washed in Ringer’s solution (150 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 10 mM glucose, 5 mM HEPES, pH 7.4) before the onset of the experiment (37). The T cells were added onto a monolayer of H-2Dd-adhered CH27 cells loaded with antigen, and the ratio of fura-2 emission at 510 nm was determined by dividing the emission intensities at 340/380 nm excitation. Data acquisition and analysis was done using Metamorph 6.0 (Universal Imaging). Experiments were done using a Zeiss Axiovert 200 microscope, equipped with a Xenon light source, 20X Fluar objective (NA 0.75), and maintained in a 37°C environmental control chamber.
For intracellular cytokine staining, CD4KI T cells expressing equivalent levels of either CD4 or S408A were stimulated with irradiated B6.K splenocytes loaded with agonist or APLs for 8 hours. During the last 2 hours of stimulation, 10 μg/ml Brefeldin A was added to the culture. Cells were then fixed with 4% paraformaldehyde, permeabilized with 0.5% saponin, and stained for CD25, IFN-γ, and Cab.
Lck kinase assays were done as previously described (38). For CD4 ip kinase assays, equal number of T cell hybrids (2.5×107) expressing equivalent CD4 levels were lysed in lysis buffer (1% NP40, 10 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM PMSF, 10 μg/ml Leupeptin, 10 μg/ml Aprotinin, 0.2mM Na3VO4, 10mM NaF), and incubated with Protein G-Sepharose beads bound with the anti-CD4 antibody, GK1.5, for 2 hours at 4°C. The beads were then washed 3 times in lysis buffer, followed by 2 washes with kinase buffer (50 mM Hepes pH 7.5, 10 mM MnCl2, 0.1 mM Na3VO4) and then subjected to in vitro kinase assay reactions with 1.2 mM of either the substrate peptide (RRLIEDAHYAARG) or control peptide (RRLIEDAHAAARG) in the presence of 0.2 mM ATP and 10 μCi [γ-32P] ATP for 10 minutes at 30°C. Reactions were stopped by the addition of 10% trichloroacetic acid. Reactions were spotted in triplicate onto P81 paper, washed 5 times with 0.5% phosphoric acid, followed by one wash in acetone. Dry filters were then counted in a scintillation counter. For PKCθ kinase assays, purified PKCθ (Millipore) was added to 10 μg of either CD4 peptide (RRRQAARMSQIKRL), or S408A peptide (RRRQAARMAQIKRL), and the reaction was carried out as described in product literature (Millipore).
For this study, we used the 3.L2 TCR transgenic mouse model, specific for Hb/I-Ek. To be able to assess the role of CD4 in primary mouse T cells, we took advantage of our previous finding, that we could generate functional CD4-deficient N3.L2 T cells in vivo (termed CD4KI) by expression of an altered peptide ligand in the thymus (33). This system provided us with a unique source of CD4-deficient N3.L2 primary T cells, and bypassed the critical role of CD4 in T cell development.
We generated and utilized two types of retroviral constructs: 1) for imaging studies, we made fusion proteins consisting of either CD4 or S408A (serine at position 408 mutated to alanine, Figure 1A) fused in-frame with yellow fluorescent protein (designated CD4-YFP or S408A-YFP, respectively 2) for functional studies, we made non-fusion proteins in which CD4 or S408A was inserted into a GFP-reporter retroviral construct (designated as CD4 or S408A, respectively). Both fusion and non-fusion proteins exhibited identical behavior when expressed in T cell hybrids or primary T cells (data not shown). To confirm the lack of endocytosis of S408A, we transduced CD4− T cell hybridomas expressing the 3.L2 TCR and determined the CD4 cell surface expression after activation (Figure 1B). CD4 surface expression decreased rapidly upon activation, to less than 30% by 2hrs. In marked contrast, S408A levels did not decrease at all during the first 2 hours, or at times as long as 8 hrs. These findings confirm the published phenotype of the S408A mutation (22, 30, 31)}, that it prevents endocytosis of CD4 molecules following activation.
We were able to transduce CD4 into CD4KI T cells, and utilized these cells for the functional studies in a comparison of CD4 and S408A. Despite repeated transductions and FACS sorting, the expression levels were still approximately 10-fold less than normal CD4 levels, consistent with other studies involving retroviral expression of proteins in murine T cells (18, 39). However, for the imaging studies, we were concerned that the levels we could achieve may complicate our interpretations of the findings of the movement of CD4, due to the lower expression levels. Therefore, we developed an alternate system whereby we transduced CD4-YFP or S408A-YFP into primary N3.L2 T cells that expressed endogenous CD4, resulting in only a 10% increase in total CD4 levels (data not shown). This was similar to a previous approach using T cell clones (18) and we observed no effects of the slight increase in CD4 levels on T cell proliferation, IL-2 production, or patterns of cytokine secretion such as IFN-γ, IL-4 or IL-5 (data not shown). Thus, we have created a unique system that permitted us to image CD4 dynamics at the IS in a physiological manner and examine its role in T cell signaling and function in normal primary T cells.
We first examined CD4 movement in T cells expressing either CD4-YFP or S408A-YFP in fixed T cell:APC conjugates 30 minutes after cell contact. We observed very little and transient accumulation of CD4-YFP (Figure 1B, top), consistent with previous findings (17, 18). In marked contrast, we detected high level accumulation of S408A-YFP at the IS, achieving a distinct, centralized accumulation (Figure 1B bottom, Figure 1D, and Supplementary Video 1). We quantitated the degree of accumulation at 30 minutes, and found that S408A-YFP accumulation was significantly more than CD4-YFP (Accumulation Index (AI): 2.45 vs. 1.17, p<0.0001, Figure 1C, left). The majority of T cells expressing S408A-YFP showed strong accumulation (60.8%), whereas most of the T cells expressing CD4-YFP had weak accumulation (69.2%); Figure 1C, right). Based on the persistence on the cell surface of S408A upon T cell activation (Fig. 1B), we propose that the increased accumulation of S408A at the IS is due to its lack of internalization.
To determine the kinetics of the accumulation of CD4-YFP and S408A-YFP during IS formation, we performed live-cell imaging on primary N3.L2 T cells transduced with either constructs. We detected modest CD4-YFP accumulation at the interface within 30–60 seconds after IS formation, which disappeared within 120 seconds (Figure 2A and Supplementary Video 2), consistent with the kinetics observed by Davis and colleagues(18). With S408A-YFP, we observed a much different pattern, with strong accumulation within the first 60 seconds after IS formation, which increased over the next 60 seconds, followed by persistence for the duration of the experiment (Figure 2B and Supplementary Video 3). This finding reveals that the accumulation we observed at 30 minutes (Figure 1B), occurred very rapidly and was maintained for a long period of time. Quantitation of the kinetics of the accumulation in representative cells for CD4-YFP and S408A-YFP, is shown in Figure 2C. During the first two minutes after initial contact, we detected a modest and transient accumulation with CD4-YFP (Figure 2C, left), whereas with S408A-YFP we observed a rapid and strong accumulation (Figure 2C, right).
Most of the S408A-YFP accumulation exhibited a centralized localization pattern (Figure 1B), consistent with it being in the c-SMAC. To confirm this localization, we examined S408A-YFP accumulation in relationship to CD3-ζ-CFP, a known c-SMAC marker (14, 40) As shown in Figure 2D, we found that S408A-YFP co-localized within the same region as CD3-ζ-CFP, indicating that centralized S408A accumulation was occurring in the c-SMAC. These findings show that the S408A mutation did not affect the movement of CD4 into the synapse, but significantly altered the movement of CD4 out of the synapse, leading to a large centralized accumulation in the c-SMAC.
From the established role of CD4 as a co-receptor in coordinating Lck accumulation at the IS (20), a prediction would be that enhanced CD4 accumulation at the IS would lead to enhanced T cell signaling and T cell function. We examined the effect of the S408A mutation on calcium mobilization, a well-established early event in T cell activation (41). We reconstituted CD4KI T cells with either CD4 or S408A constructs expressing GFP as a reporter protein, sorted for equivalent GFP levels, and measured calcium levels over time, using the calcium-sensitive dye fura-2, in response to agonist pulsed APCs. Overall, we found that the calcium levels and patterns were the same between the T cells expressing either CD4 or S408 (Figure 3). We did observe a modest increase in peak calcium levels in T cells expressing S408A compared to CD4 in response to low concentrations (1μM) of the agonist, Hb(64–76) (Figure 3A). These differences in peak calcium levels were not detected at higher concentrations of the agonist (data not shown). It is not clear if these observed increases in peak calcium levels are functionally significant, because the normal peak calcium levels are above the threshold needed for T cell activation.
It was possible, that the effect on the calcium mobilization with enhanced accumulation of S408A would be more evident with weaker ligands. We therefore measured calcium mobilization in response to the weak agonist, T72. Overall, calcium responses was weaker for both CD4 and S408A, compared to Hb(64–76) agonist stimulation (Figure 3B). There was again a modest increase in peak calcium levels exhibited by S408A over CD4 (Figure 3B), but the overall pattern was the same. These results show that centralized S408A accumulation had no effect on the early T cell signaling, as measured by calcium mobilization.
We next determined the effect of centralized S408A accumulation at the IS on the later T cell responses of cytokine production and T cell proliferation. When we measured IFN-γ production 8 hours after antigen stimulation, we did not detect an increase in IFN-γ production in T cells expressing S408A (Figure 4A). No significant differences in T cell responses to the weak agonist or antagonist were also seen (Figure 4A): the differences seen in response to the weak agonist, T72, were within normal variation and not significant (H.K. and P.M.A, unpublished observations). When we extended our analysis to include T cell functional responses 20 or 72 hours after antigen stimulation, we detected the same consistent pattern. At 20 hours, T cells expressing either CD4 or S408A produced similar levels of IL-2 in response to the agonist or weak agonist in a dose-dependent manner (Figure 4B). At 72 hours, T cells expressing S408A proliferated equivalently well compared to T cells expressing CD4, in a dose-dependent manner to either the agonist or weak agonist (Figure 4C). Taken together, our results show that centralized CD4 accumulation at the c-SMAC had no effect on early or late T cell signaling or function.
One possibility for the lack of differential T cell functional responses between T cells expressing CD4 or S408A is that the stable expression levels that we were able to achieve by retroviral transduction of primary T cells were not high enough to reveal a difference. To address this issue, we transduced T cell hybridomas cells expressing the 3.L2 TCR with either CD4 or S408A, and FACS sorted for high levels of expression. In the T cell hybridoma cells we were able to achieve high expression for both CD4 and S408A (Fig 6A). These levels were significantly higher than we were able to achieve in the primary T cell blasts, and were approximately 30% of the normal levels. The functional responses of the 3L2 T cells expressing either CD4 or S408A to Hb(64–76) were then examined, and their dose response curves were indistinguishable from each other (Fig 6B). Thus, despite higher levels of CD4 or S408A expression in the T cell hybridomas, the findings confirm our those with the primary T cell blasts (Fig 5), in that despite the significant accumulation of S408A at the IS, there did not appear to be any significant functional consequence.
The finding that enhanced CD4 accumulation did not lead to enhanced T cell functional responses led us to examine whether Lck activation or localization was impaired in T cells expressing S408A. We first wanted to establish that the S408A mutation did not affect the level of Lck activity associated with CD4. To do this, we immunoprecipitated CD4 or S408A from 3.L2 T cell hybrids expressing identical levels of the respective CD4 molecules, and performed Lck kinase assays. Our results show that CD4-associated Lck kinase activity was the same for CD4 or S408A, indicating that the S408A mutation did not affect basal Lck kinase activity (Figure 5A). We next examined active Lck localization at the IS using an antibody specific for the active form of Lck (42). There were no significant differences between T cells expressing either CD4 or S408A in the localization or levels of active Lck (Figure 5B). We did observe two patterns of active Lck staining: accumulation either at the p-SMAC (Figure 5B) or at both the c-SMAC and p-SMAC (data not shown), but they were both found in the CD4 and S408A expressing cells. Total Lck recruitment was also the same between T cells expressing either CD4 or S408A, ruling out the possibility that the S408A mutation affected total lck recruitment (Figure 5C and 5D). For these experiments, it is important to note that CD4 and S408A were expressed in primary CD4KI T cells, in which we not able to achieve full wild type levels of expression. Because of the low expression levels, we did not visualize any enhanced total Lck accumulation as one might have expected from the expressed S408A. These findings with Lck indicated that the S408A mutation did not affect the activity state or recruitment pattern of Lck.
We next investigated what kinase(s) was involved in the phosphorylation of CD4 in the IS. Previous studies had implied that S408 was phosphorylated by a PKC (31, 43). We focused on PKCθ, based on it’s role in T cell activation, and recruitment to the IS (44, 45). The S408 region of CD4 (Figure 1A) contains potential PKC recognition sites (44). We initially determined if PKCθ could directly phosphorylate S408 using an in vitro kinase reaction, with recombinant PKCθ and either an S408 containing peptide or a control S408A peptide. PKCθ could strongly phosphorylate the S408 peptide but not the control S408A peptide (Figure 6A). To assess the effect of PKCθ on CD4 in primary T cells, we used two independent approaches. We then treated normal N3.L2 T cells with rottlerin, a PKCθ inhibitor (46), and visualized CD4 movement by staining for CD4 in fixed T cell:APC conjugates. We observed strong centralized CD4 accumulation 30 minutes after IS formation (Figure 6B), similar to what we had observed with S408A (Figure 1B). This finding showed that inhibiting PKCθ dramatically affected the movement of CD4 at the IS. Because rottlerin could also inhibit the PKCδ isoform, which has been shown to be involved in T cell activation (47), we then examined CD4 accumulation in PKCθ-deficient N3.L2 T cells. We detected strong centralized CD4 accumulation in N3.L2-PKCθ–deficient T cells, whereas no accumulation was detected with N3.L2 T cells (Figure 6C). The CD4 accumulation in both the N3.L2-PKCθ-deficient and rottlerin-treated T cells was significantly higher compared to normal N3.L2 T cells (Figure 6D, left). The distribution of cells exhibiting centralized CD4 accumulation was also greater in rottlerin-treated and N3.L2-PKCθ-deficient compared to WT N3.L2 T cells (Figure 6D, right). Thus, by either inhibiting or eliminating PKCθ in T cells, we were able to dramatically alter the accumulation of CD4 at the IS, recapitulating a pattern generated by the S408A mutation of CD4. These studies reveal a previously unrecognized link between CD4 and PKCθ, and strongly support a model in which PKCθ is directly involved in the removal of CD4 from the IS by phosphorylating the S408 residue.
In this study, we have established a critical role for the serine phosphorylation of the cytoplasmic tail of the CD4 molecule for its movement at the IS. Mutation of the S408 residue did not prevent CD4 from localizing to the IS upon contact with an APC, indicating that it was not required for the recruitment to the IS. However, phosphorylation of S408 essential for the removal of CD4 from the IS, in that the S408A mutation, resulted in prolonged accumulation of CD4 at the IS. The phosphorylation of S408 was performed exclusively by PKCθ, as evidenced by the accumulation of wt CD4 in T cells either treated with a PKCθ inhibitor, or in PKCθ T deficient cells. PKCθ is the only PKC isoform that is recruited to the IS, and we provide compelling evidence that CD4 is a direct substrate of PKCθ, making it the first PKCθ substrate in the IS to be identified. Thus, our findings support a model in which PKCθ participates in the turnover of proteins in the IS by phosphorylating multiple substrates, including CD4. Our studies also help to clarify the reason for the removal of CD4 from the IS. The S408A mutation and the resulting accumulation of CD4 at the IS, did not have any functional consequences on early or late T cell functions. These findings, eliminate the possibility that maintenance of CD4 in the IS would be deleterious, and it is removed from the synapse to help coordinate the signaling initiated through the TCR. They do support the concept that co-receptor function of CD4-Lck is acting as a catalyst, being critically important only during the initial TCR signaling events. The reason for the rapid removal, within minutes, of CD4 from the IS is not fully resolved, in that its retention does not appear to be detrimental, and sustained contact of several hours between the T cell and APC are required for full T cell activation.
The IS is now known to be a highly dynamic structure. The turnover of proteins in the IS could be involved in the tuning of a strong stimulus and/or the dissolution of the IS, allowing the relocation and reformation of the IS. Sims et al. have shown using a supported planar bilayer system that for naive T cells PKCθ is involved in the breaking of the IS, promoting T cell motility (48). Our studies are consistent with this model, and identify CD4 as one of the substrates of PKCθ in the IS. It is well established that PKCθ localizes to the IS, but its precise localization within the IS seems to be dependent on the activated state or type of T cell used. Sims et al. recently reported that PKCθ was localized to the pSMAC in the supported planar bilayer model in naive T cells (48), which was in contrast to the cSMAC localization reported by Kupfer and colleagues using B cell APCs and T cell clones (14). In CD8+ T cells, PKCθ was found to localize to both the cSMAC and pSMAC (49). CD28 has also been shown to be important for the localization of PKCθ to the cSMAC (50, 51). These differences most likely reflect the different T cells, antigenic strength, and the co-receptor dependence, among other variables. We observed that CD4 localized to the cSMAC, but did not address the localization of PKCθ and where CD4 and PKCθ interacted. In our studies, we used B cells as APC and primary mouse T cell blasts, and the PKCθ could have localized to cSMAC, where it encounter CD4. Alternatively, CD4 could encounter PKCθ in the pSMAC, become phosphorylated, migrate to the cSMAC, and then leave the cSMAC. Future studies involving differentially labeled CD4 and PKCθ would help to resolve this issue. The phosphorylation of CD4 initiates several actions, with Lck being dissociated from CD4 and CD4 being endocytosed. The fate of the dissociated Lck in the IS it not currently known. It could remain in the IS, making CD4 an Lck transport molecule or the dissociated Lck could join the pool of free Lck in the T cell. Lck has been shown to interact directly with PKCθ, which could also provide another mechanism for Lck remaining in the IS (52). In our studies, we did not find any differences in the localization of total Lck or active Lck in T cells expressing either CD4 or S408, but we could not rule out some highly localized Lck effects, that were below our level of detection.
In this study, we found little accumulation of WT CD4 at the IS, similar to results reported by Davis and colleagues, who also reported transient accumulation of CD4 at the IS (18). Our findings were somewhat different from those of Zal et al., who showed strong CD4 accumulation early after IS formation, but not necessarily in an antigen-specific manner (21). Possibilities for these differences could be the T cells used- T cell hybrids versus primary T cell blasts, or the affinity of the TCR for the pMHC. It is still possible that the little CD4 accumulation seen with WT CD4 in our system is reflective of the “passive” nature of CD4 movement on the surface of the T cell, with the recruitment of TCR and its proximity to CD4 being the determinant for T cell activation (21). Our studies did not address the important issue of the activation/differentiation state of the T cells on the role and movement of CD4 and PKCθ. For example, PKCθ has been shown to differentially localize in naive versus activated T cells (48), raising the possibility that its role in the phosphorylation of CD4 may vary depending upon the activation state. The T helper subset (Th1, Th2, Th17) may also influence the role of CD4, as has been
PKCθ is an essential component of T cell activation, and is involved in many different signaling pathways (44, 53, 54). Here, we have made the novel link between PKCθ and CD4, and provide strong evidence that CD4 is a direct substrate of PKCθ. We cannot rule out the possibility that PKCθ is acting indirectly on the CD4 through the action of another kinase. We favor the simplest explanation that PKCθ is directly phosphorylating CD4 on the S408 residue, as evidenced by the demonstration that purified PKCθ can phosphorylate a S408 containing peptide, and that inhibition of PKCθ or genetic elimination of PKCθ, resulted in wt CD4 accumulating at the IS, in a similar manner as the S408A mutation. Clearly, the phosphorylation of CD4 by PKCθ is only one of many activities of PKCθ, given that the PKCθ deficient T cells have a profound defect in signaling (54), and that we found no functional defect in the S408A expressing T cells.
In this study, we were able to investigate the role of S408 phosphorylation in primary mouse T cells because of our ability in vivo to generate T cells in the absence of CD4, using the CD4KI system (33). Because previous studies have shown a pivotal role for S408 in controlling CD4 down-regulation in T cell hybrids and non-transformed human cell lines (22, 31), we assumed that this would be similar in primary T cells, and that mutating S408 would lead to increased CD4 surface expression. Instead, we detected enhanced S408A accumulation at the c-SMAC, but S408A down-regulation was similar to wild-type CD4 in mouse primary T cells (data not shown). This implies a different mechanism of control of CD4 down-regulation in primary T cells compared to T cell hybrids, not necessarily dependent on S408. It is also possible that the defect on CD4 down-regulation is more acute and less severe in primary T cells, visible only by imaging, and not by measuring the overall expression of CD4 on the surface.
In conclusion, our findings in mouse primary T cells support the concept that CD4 co-receptor function operates within a spatiotemporal window at the IS, with its primary purpose to first stabilize TCR-pMHC interaction and second, to deliver Lck to the IS to help initiate T cell signaling. We also uncover a new connection between CD4 and PKCθ, which highlights the importance of phosphorylation by PKCθ in the dynamic nature of protein turnover in the IS.
We thank Arup Chakraborty for helpful discussions and reading the manuscript, Saso Cemerski, Nathan Felix, Celeste Morley, Scott Weber, and Emanuele Giurisato for critically reading the manuscript, Scott Weber for help with the calcium imaging, Jennifer Racz technical assistance, Ken Matsui for help in designing the fusion constructs, Darren Kreamelmeyer for animal husbandry, and Stephen Horvath for peptide synthesis.
The authors have no financial conflicts of interest.
1This work was supported by grants AI24157 and AI071195 from The National Institutes of Health.
Publisher's Disclaimer: “This is an author-produced version of a manuscript accepted for publication in The Journal of Immunology (The JI). The American Association of Immunologists, Inc. (AAI), publisher of The JI, holds the copyright to this manuscript. This version of the manuscript has not yet been copyedited or subjected to editorial proofreading by The JI; hence, it may differ from the final version published in The JI (online and in print). AAI (The JI) is not liable for errors or omissions in this author-produced version of the manuscript or in any version derived from it by the U.S. National Institutes of Health or any other third party. The final, citable version of record can be found at www.jimmunol.org.”