|Home | About | Journals | Submit | Contact Us | Français|
T cell receptor (TCR) signaling to NF-κB is required for antigen-induced T cell activation. We conducted an expression-cloning screen for modifiers of CARD11, a critical adapter in antigen-receptor signaling, and identified the kinesin-3 family member GAKIN as a CARD11 inhibitor. GAKIN negatively regulates TCR signaling to NF-κB, associates with CARD11 in a signal-dependent manner, and can compete with the required signaling protein, Bcl10, for association. In addition, GAKIN dynamically localizes to the immunological synapse and regulates the redistribution of CARD11 from the central region of the synapse to a distal region. We propose that CARD11 scaffold function and occupancy at the center of the synapse are negatively regulated by GAKIN to tune the output of antigen-receptor signaling.
During the adaptive immune response, antigen receptor signaling in B and T lymphocytes must be finely tuned so that the immune system can recognize foreign pathogens and respond efficiently without harming the host. Signaling pathways that emanate from TCR and BCR complexes activate programs of gene expression that determine whether a lymphocyte proliferates, becomes activated, anergic, or dies as a result of an encounter with a putative antigen (Cancro, 2009; Smith-Garvin et al., 2009). One of the key transcription factors activated by antigen receptor signaling is NF-κB, which is required in both T and B cells for antigen-induced lymphocyte proliferation, survival, and effector functions (Vallabhapurapu and Karin, 2009).
Optimal antigen-induced activation of NF-κB occurs in T cells in response to concurrent TCR engagement and costimulatory receptor ligation. Antigen recognition initiates receptor clustering and reorganization of signaling components at the cell:cell contact between an antigen presenting cell (APC) and a T cell, forming the immunological synapse (IS) (Lin et al., 2005). The mature IS is segregated into discrete zones termed supramolecular activation clusters (SMAC). The TCR and PKCθ are found in the central SMAC (cSMAC), while adhesion receptors and cytoskeletal proteins, including LFA-1 and Talin, localize to the peripheral SMAC (pSMAC) (Monks et al., 1998). The distal SMAC (dSMAC) is the outermost zone and is defined by the presence of the inhibitory receptor, CD45 (Freiberg et al., 2002).
CARD11 (CARMA1) is a multi-domain scaffold protein that is required for TCR signaling to NF-κB and controls the recruitment of other signaling proteins to the IS (Blonska and Lin, 2009). Prior to TCR engagement, CARD11 is kept in an inactive state by an inhibitory domain (ID) that prevents the binding of multiple proteins (McCully and Pomerantz, 2008). TCR engagement causes the phosphorylation of the ID, mediated in part by PKCθ (Matsumoto et al., 2005; Sommer et al., 2005), which neutralizes its inhibitory effect and allows CARD11 to associate with a group of factors that contribute to the activation of the IKK complex, including Bcl10, TAK1, TRAF6, Caspase-8, and IKKγ (McCully and Pomerantz, 2008). Activated IKK phosphorylates IκB, leading to its ubiquitination and degradation, and the appearance of active NF-κB in the nucleus.
CARD11 activity must be tightly regulated to prevent the hyperactivation of downstream pathways that could lead to dysregulated immune responses or the unwarranted lymphocyte proliferation that is associated with certain NF-κB-dependent types of lymphoma (Jost and Ruland, 2007). Cell lines derived from the ABC subtype of Diffuse Large B Cell Lymphoma (DLBCL) have been shown to require CARD11 signaling for their unchecked ability to proliferate in culture (Ngo et al., 2006). In addition, several mutations in CARD11 have been identified in patient samples of DLBCL that endow the protein with a hyperactive ability to activate NF-κB and confer dysregulated growth (Lenz et al., 2008).
To identify components of the antigen receptor signaling pathway that regulate CARD11, we conducted an expression-cloning screen for enhancers and suppressors of CARD11 signaling activity. We describe the identification of the motor protein GAKIN, a guanylate kinase associated kinesin, as a CARD11 inhibitor that attenuates the extent of NF-κB activation following TCR engagement. We show that GAKIN interacts with CARD11 in an inducible manner during signaling, can compete with Bcl10 for association with CARD11, and regulates a previously unrecognized redistribution of CARD11 from the PKCθ-rich center of the IS to a distal region of the IS.
We adapted the expression cloning strategy of Pomerantz et al. (Pomerantz et al., 2002) to identify cellular factors that could modulate CARD11 signaling activity. Pools of cDNAs from a human spleen expression library were screened for the ability to enhance or suppress the CARD11-mediated activation of the NF-κB-responsive luciferase reporter, Igκ2-IFN-LUC, in HEK293T cells. The individual cDNA responsible for each pool’s activity was then purified by sib selection. Using this approach, we identified the motor protein, guanylate kinase associated kinesin (GAKIN) (Hanada et al., 2000), as an inhibitor of CARD11. GAKIN inhibited CARD11-mediated activation of the Igκ2-IFN-LUC reporter in HEK293T cells in a dose-dependent manner (Figure 1A). The effect of GAKIN on CARD11 was specific; GAKIN did not inhibit NF-κB activation by Bcl10, CARD9, CARD10, or CARD14 (Figure S1A-D). Furthermore, overexpression of GAKIN in Jurkat T cells inhibited NF-κB activation mediated by superantigen (SEE) pulsed-Raji B cells (Figure 1B) indicating an ability to influence CARD11-dependent TCR signaling. In contrast, GAKIN did not inhibit the activation of NF-κB by TNFα, which signals through a CARD11-independent pathway (Figure 1B).
To assess whether endogenous GAKIN regulated CARD11-dependent signaling, we reduced GAKIN expression in Jurkat T cells by lentiviral-mediated RNA interference (RNAi). We expressed two short hairpin RNAs, shRNA2b and shRNA5a, which stably reduced endogenous protein levels of GAKIN by >90% and 60%, respectively (Figures 1C and 1D). Each hairpin resulted in the enhanced activation of NF-κB achieved by anti-CD3/anti-CD28 crosslinking, as compared to that observed with the control hairpin, siGFP (Figure 1C), consistent with an inhibitory role for GAKIN in TCR signaling. The effect of the shRNA2b hairpin was reversed by transient expression of a GAKIN cDNA that contained silent mutations in the shRNA2b target sequence, confirming the specificity of the RNAi (Figure 1E). GAKIN knockdown resulted in enhanced TCR-induced IKK activity (Figure S1E-H), consistent with an effect on a CARD11-dependent step in the pathway, and also resulted in the enhanced anti-CD3/anti-CD28-mediated production of IL-2, an endogenous NF-κB-dependent target of TCR signaling (Figure 1F). Importantly, knockdown of GAKIN in primary human CD4+ T cells by >70% also resulted in enhanced IL-2 production following anti-CD3/anti-CD28 stimulation (Figures 1G and 1H). The observed hyperresponsiveness of GAKIN-deficient T cells to TCR triggering demonstrates that GAKIN functions as an inhibitor at endogenous levels in both Jurkat T cells and primary human CD4+ T cells during antigen receptor signaling.
Prior to TCR engagement, the ID represses the ability of CARD11 to associate with multiple signaling proteins (McCully and Pomerantz, 2008; Sommer et al., 2005). TCR signaling results in the phosphorylation of the ID, which neutralizes its inhibitory effect and activates CARD11 scaffold activity (Matsumoto et al., 2005; Sommer et al., 2005). To test whether GAKIN could inhibit CARD11 signaling after the conversion of CARD11 to an active scaffold, we tested the effect of GAKIN on CARD11ΔID. This mutant, in which the ID has been deleted, is hyperactive, capable of recruiting signaling proteins in a signal-independent manner, and thus behaves as a CARD11 in which the ID has been neutralized by TCR signaling (McCully and Pomerantz, 2008; Sommer et al., 2005). GAKIN potently inhibited the activation of NF-κB achieved by CARD11ΔID in Jurkat T cells (Figure 2A), suggesting that GAKIN inhibits the pathway by targeting a step subsequent to ID neutralization.
To test whether GAKIN and CARD11 could physically associate, we first coexpressed GAKIN with either wild-type CARD11 or CARD11ΔID in HEK293T cells, and assessed association by coimmunoprecipitation. We observed a weak association between myc-tagged CARD11 and HA-tagged GAKIN, but an enhanced association of GAKIN-HA with myc-CARD11ΔID (Figure 2B). These results indicated that the ID of CARD11 interferes with GAKIN binding and predicted that the endogenous proteins would associate only during TCR signaling. Indeed, CARD11 associated with GAKIN in Jurkat T cells in a stimulation-dependent manner; the association of CARD11 with GAKIN was apparent after 5 minutes of PMA/ionomycin treatment, with increased association after 15 minutes and a subsequent decrease after 45 minutes (Figure 2C). Coimmunoprecipitation of CARD11 was not observed with the anti-FLAG control antibody or with the GAKIN-specific antibody in the unstimulated sample. This time-course of association of GAKIN with CARD11 was similar to that observed for Bcl10, TRAF6, Caspase-8, and the IKK complex (McCully and Pomerantz, 2008). These results demonstrate that CARD11 and GAKIN can associate at endogenous levels in T cells and that this association is regulated in a signal-inducible manner.
To identify which domains of GAKIN could associate with CARD11, we tested five regions of GAKIN for the ability to coimmunoprecipitate CARD11ΔID in HEK293T cells (Figures 3A and 3B). The motor domain, MAGUK Binding Stalk (MBS), and two regions from the tail (TR1 and TR2) were each independently capable of associating with CARD11ΔID, while the L1-FHA-L2 segment was not. Each region that could associate with CARD11ΔID was also able to inhibit the ability of CARD11ΔID to activate NF-κB when overexpressed in Jurkat T cells (Figure 3C) at comparable levels of expression (Figure 3D).
GAKIN, like other members of the kinesin family, is regulated by intramolecular interactions that keep the protein in a closed, autoinhibited state prior to the binding of activating ligands or cargo (Yamada et al., 2007). Autoinhibition appears to be mediated by intramolecular interactions between the motor and MBS domains, and between the tail and a region containing the MBS (Yamada et al., 2007). We tested whether the disruption of intramolecular interactions within GAKIN would affect the ability of the protein to associate with CARD11. The deletion of the motor domain resulted in an enhanced ability of GAKIN to associate with CARD11ΔID (Figure 3E). The deletion of the MBS also resulted in an enhanced interaction, while the deletion of the FHA and the CAP-Gly domains had no significant effect (Figure 3E). These results suggest that the association of GAKIN with CARD11 is regulated by intramolecular interactions involving the motor and MBS domains.
To identify which CARD11 domains are required for GAKIN association, a panel of CARD11 deletion constructs made in the context of CARD11ΔID (McCully and Pomerantz, 2008) was coexpressed with GAKIN in HEK293T cells and assayed for coimmunoprecipitation. The deletion of either the CARD or Coiled-coil domains reduced the ability of CARD11ΔID to associate with GAKIN (Figure 3F), while other domain deletions had no significant effect. The requirement of the CARD and Coiled-coil domains for GAKIN association is consistent with the fact that the CARD11:GAKIN association is regulated by the ID and that TCR signaling is required for maximal association (Figures 2B and 2C).
After the signal-induced neutralization of the ID, the CARD and Coiled-coil domains are required for the recruitment of several signaling proteins to CARD11, including Bcl10, Caspase-8, IKKγ, TRAF6 and TAK1 (McCully and Pomerantz, 2008). Since GAKIN association also requires the CARD and Coiled-coil domains, we hypothesized that GAKIN might inhibit CARD11 scaffold function by competing with one or more signaling proteins for the association with CARD11. We tested whether the binding of these proteins to CARD11 could be competed by coexpression of ΔMotor-GAKIN, which displayed an enhanced interaction with CARD11ΔID as compared to full-length GAKIN and contained most of the domains sufficient for association with CARD11ΔID (Figure 3E). Intriguingly, ΔMotor-GAKIN could inhibit the ability of Bcl10 to coimmunoprecipitate CARD11ΔID in a dose-dependent manner (Figure 3G). Several other proteins tested, including TAK1, TRAF6, IKKγ and Caspase-8 did not appear to be competed under these conditions (Figure 3H), suggesting that ΔMotor-GAKIN can specifically influence the association of CARD11 with Bcl10. Consistent with an inhibitory effect of GAKIN on CARD11:Bcl10 association, GAKIN knockdown in Jurkat T cells resulted in a modest, but reproducible enhancement of the amount of CARD11 that dynamically associated with Bcl10 during signaling (Figure 3I). Quantitation of the amount of CARD11 associating with Bcl10 in the GAKIN-knockdown cells, as compared to control cells, suggested that GAKIN inhibits formation of at least 26-44% of the Bcl10:CARD11 complex that would otherwise form in the absence of GAKIN after 10 to 15 minutes of PMA/ionomycin stimulation (Figure 3J).
CARD11 has been observed at the IS (Matsumoto et al., 2005; Tanner et al., 2007; Wang et al., 2004) and regulates the recruitment of the IKK complex to the cSMAC (Hara et al., 2004). Since our results indicated that GAKIN inhibited CARD11 and associated with it upon signaling, we hypothesized that GAKIN would be recruited to the synapse. As GAKIN is a kinesin-3 family member and might exhibit dynamic localization reflective of motor activity, we used live cell imaging with eGFP-tagged GAKIN to assess localization with high temporal resolution. We stably coexpressed GAKIN-eGFP and mCherry-α-tubulin in Jurkat T cells and formed conjugates with SEE-loaded Raji B cells. mCherry-α-tubulin served as a marker of the microtubule organizing complex (MTOC), which undergoes a characteristic signal-dependent relocalization to the area of cell:cell contact in T cells engaged in a productive synapse (Smith-Garvin et al., 2009). Prior to conjugate formation, GAKIN-eGFP was present diffusely in the cytoplasm with some concentrated localization at the MTOC and at the uropod opposite the MTOC (Figure 4A, top panel). At early timepoints after first contact between T cell and APC, GAKIN-eGFP was enriched at the area of cell:cell contact, and also colocalized with the MTOC (Figure 4A, t=60s; see Movie S1 for an entire representative time-course). This early enrichment of GAKIN-eGFP at the IS was observed in 77% of conjugates (n=26) and occurred within 37.7±7.4 seconds of initial T cell:APC contact (Figure 4C). At later timepoints, after MTOC relocalization to the IS, GAKIN-eGFP maintained colocalization with the MTOC at the IS but interestingly, also appeared to concentrate at the distal pole in 69.2% of conjugates (n=26) (Figure 4A, t=600s). These results indicate that GAKIN dynamically localizes to the IS during TCR signaling, and that GAKIN is appropriately localized to exert an inhibitory effect on CARD11 after signaling has been initiated.
The motor domain of GAKIN has been shown to mediate plus-end-directed movement of GAKIN along microtubules (Horiguchi et al., 2006; Yamada et al., 2007). To assess the contribution of the motor domain to GAKIN localization, we stably expressed a motor-deleted variant, ΔMotor-GAKIN-eGFP, in Jurkat T cells with mCherry-α-tubulin. ΔMotor-GAKIN-eGFP displayed a very different pattern of localization (Figure 4B), as compared to GAKIN-eGFP. Prior to and throughout IS formation, ΔMotor-GAKIN-eGFP appeared to concentrate and colocalize with the MTOC, and relocalized to the IS simultaneously with the MTOC (see Movie S2 for an entire representative time-course). Unlike wild-type GAKIN-eGFP, ΔMotor-GAKIN-eGFP did not concentrate at the IS at early timepoints prior to MTOC relocalization, and was recruited to the IS 202.5±40.8 seconds after initial T cell:APC contact (Figure 4C). After MTOC relocalization, ΔMotor-GAKIN-eGFP colocalized with the MTOC at the IS, and did not appear at the distal pole (93.3% of conjugates (n=15)). Thus, the motor domain was required for the concentration of GAKIN at the IS prior to MTOC relocalization, and for the concentration of GAKIN at the distal pole after MTOC relocalization to the IS. The absence of the motor domain appeared to confine ΔMotor-GAKIN-eGFP to the MTOC even as the MTOC relocalized to the area of APC:T cell contact.
To investigate a potential effect of GAKIN on CARD11 localization, we first characterized CARD11 localization during TCR signaling using live cell imaging and PKCθ as a marker of the cSMAC. We co-infected Jurkat T cells with viruses expressing CARD11-mCherry and PKCθ-eGFP and formed conjugates with Raji B cells preloaded with SEE. Prior to APC:T cell contact, both CARD11-mCherry and PKCθ-eGFP exhibited a diffuse localization in the cytoplasm (Figure 5A; Pre). Upon conjugation, PKCθ-eGFP concentrated in the center of the IS during the first 60 sec after cell:cell contact, as expected (Figure 5A; t=60s, 80s) and remained localized in this pattern throughout the observed timecourse. CARD11-mCherry displayed an unexpected dynamic pattern of localization within the contact region. Initially, CARD11-mCherry was enriched in the area of cytoplasm juxtaposed to the APC:T cell contact and partially colocalized with PKCθ-eGFP in the center of the contact (Figure 5A; t=60s). Later in the timecourse, however, CARD11-mCherry concentrated at a distal region of the synapse and no longer significantly colocalized with PKCθ-eGFP (Figure 5A; t=80s). To characterize this transition in CARD11 localization, we determined the fluorescence intensity of CARD11-mCherry and PKCθ-eGFP along a line passing through the APC:T cell contact (in the X-Y plane from B to E) at 60 and 80 seconds after initial cell:cell contact for the conjugate displayed in Figure 5A. The peak of PKCθ-eGFP fluorescence occupied the center of the cell:cell contact at both timepoints (Figure 5B). In contrast, the CARD11-mCherry fluorescence was initially evenly distributed across the cell:cell contact at t=60s (Figure 5B, left), but then resolved into two distal peaks of fluorescence at t=80s that did not overlap with the central peak of PKCθ-eGFP fluorescence (Figure 5B, right). The en face view (X-Z plane) of CARD11-mCherry and PKCθ-eGFP fluorescence (Figure 5C), representing the contact region from the perspective of the APC, also demonstrated the transition in CARD11 localization from a broad pattern throughout the contact at t=60s that partially colocalized with PKCθ, to a distributed distal pattern at t=80s that for the most part did not colocalize with PKCθ. These results indicate that early in APC:T cell conjugation, CARD11 populates the area of cell:cell contact, but its occupancy at the center of the IS is reduced shortly thereafter by a redistribution to distal areas that limits the duration of its colocalization with PKCθ, a key kinase that activates CARD11 scaffold function during TCR signaling.
We investigated whether GAKIN might regulate CARD11 localization by analyzing the dynamic localization of CARD11-mCherry in Jurkat T cells in which GAKIN levels were stably reduced by lentiviral mediated expression of the shRNA2b hairpin. These cells were compared to control cells expressing either no hairpin, or the control non-target hairpin (NTsh). We observed that GAKIN deficiency had no apparent effect on the timing of initial CARD11 occupancy at the center of the APC:T cell contact (Figure 6A), suggesting that GAKIN does not regulate the recruitment of CARD11 to the synapse. However, we observed a significant difference between GAKIN-deficient and control cells in the timepoint at which CARD11 redistributed to the distal region of the IS. While CARD11 redistributed at 62.0±7.6 and 58.0±10.8 seconds after first contact in the NTsh and no-hairpin control cells, respectively, this redistribution occurred at 93.4±9.4 seconds after first contact in the sh2b-expressing cells (Figure 6B). For each conjugate, we measured the interval between initial occupancy in the contact center and distal redistribution and found that while duration of CARD11 occupancy at the center was 46.7±10 and 50.0±8.0 seconds in the no-hairpin and NTsh cells, respectively, this interval increased to 80.1±10.4 seconds in the sh2b cells (Figure 6C). Figures 6E and 6F show representative examples of control and GAKIN-deficient T cells, respectively, at the timepoints of initial APC:T cell contact (t=0s) and distal redistribution (t=60s for NTsh; t=141s for sh2b). We also reduced GAKIN expression in PKCθ-mCherry-expressing cells, and importantly, did not observe any difference in the timepoint of PKCθ recruitment to the cSMAC (Figure 6D) or in the extended occupancy of PKCθ there (data not shown). These data demonstrate that GAKIN regulates the duration of CARD11 localization in the contact center but does not alter the recruitment of another important synapse component, PKCθ.
Antigen-induced signaling through the TCR complex triggers the orchestrated, dynamic movement of signaling machinery to defined regions of the IS. However, the mechanisms that couple cellular localization to signaling activity for most pathway components remain poorly understood. In this report we identify the kinesin GAKIN as an inhibitory component of the TCR pathway that regulates the signaling output and localization of the multi-domain scaffold CARD11 to control the extent of TCR-mediated activation of NF-κB.
It has been previously established that after TCR engagement, CARD11 undergoes a signal-dependent transition from an inactive to an active signaling scaffold (Figures 7A and 7B) that requires the PKCθ-mediated phosphorylation of the ID. ID phosphorylation causes it to disengage from intramolecular interactions with the CARD and Coiled coil domains and allows the recruitment of signaling proteins to CARD11 to induce IKK activation (Figure 7B). This transition in CARD11 likely takes place in the central portion of the IS where PKCθ resides during signaling and where the IKK complex is recruited in a CARD11-dependent manner (Hara et al., 2004).
Our data reveal a subsequent step in TCR signaling at which GAKIN acts to limit CARD11 activity (Figure 7C). During this step, GAKIN associates with CARD11 in a manner that can reduce Bcl10 binding and promote the redistribution of CARD11 away from the PKCθ-rich central portion of the synapse. The redistribution of CARD11 to distal regions of the APC:T cell contact limits the interval during which CARD11 and PKCθ co-occupy the center of the synapse, and likely determines the duration of active CARD11 signaling in that region.
Several mechanisms could explain how occupancy of CARD11 in the center of the synapse would determine CARD11 signaling activity. First, PKCθ must phosphorylate the ID on at least two serines to convert CARD11 into an active scaffold. It is possible that the dwell time of CARD11 in the central region of the contact determines the fraction of CARD11 molecules that become fully phosphorylated by PKCθ, especially if these phosphorylation events are reversible by the action of a phosphatase. Second, active signaling may be restricted to the center of the contact by the CARD11-independent PKCθ-mediated activation of other signaling components that must work in concert with CARD11 in the pathway. For example, Shambharkar et al. demonstrated that during TCR signaling, IKK kinase activation requires both the CARD11-dependent ubiquitination of IKKγ and the CARD11-independent phosphorylation of IKKα/β by TAK1, which was shown to be activated and associate with PKCθ during TCR signaling in a CARD11-independent manner (Shambharkar et al., 2007). The dwell time of CARD11 in the center of the synapse may thus determine the extent to which CARD11-dependent signaling events can couple to CARD11-independent steps.
It is possible that GAKIN transports CARD11 as cargo away from the center of the APC: T cell contact (Figure 7C), although our data does not address this directly. IS formation results in MTOC polarization to the contact (Smith-Garvin et al., 2009), which orients microtubules such that they emanate away from the center of the IS with their plus ends distal to the MTOC. GAKIN is capable of plus-end directed motility on microtubules (Horiguchi et al., 2006; Yamada et al., 2007), and the orientation of microtubules at the IS could allow directional, GAKIN-mediated transport of CARD11 to a distal region. It is important to note that in T cells depleted of GAKIN by RNAi, CARD11 still redistributes to the distal region of the contact, albeit with altered kinetics. This delayed redistribution might be due to the action of the residual GAKIN in these cells, or to a distinct mechanism that functions with GAKIN in CARD11 redistribution.
Our data suggests several important roles for the motor domain of GAKIN during TCR signaling. The motor is required for the dynamic localization of GAKIN at the IS at early timepoints (Figure 4), where CARD11-dependent signaling to the IKK complex likely occurs (Hara et al., 2004). During signaling, the motor appears to override determinants in the other domains of GAKIN that appear to specify localization to the MTOC, because deletion of the motor results in confinement of GAKIN to the MTOC. In addition, the motor participates in an autoinhibitory intramolecular interaction with the MBS that can regulate motor activity (Yamada et al., 2007) and that appears to regulate the GAKIN:CARD11 association since deletion of either of these domains enhances the ability of GAKIN to associate with CARD11ΔID (Figure 3E). Since GAKIN associates with CARD11 in a signal-dependent manner, it is likely that TCR signaling influences GAKIN conformation. It will be important to investigate this possibility, and the possibility that TCR signaling regulates a transition from an inactive to an active motor.
Interestingly, in overexpression studies that bypass the constraints imposed by signal-regulated cellular localization and protein conformation (Figure 3B and C), the motor itself is sufficient for GAKIN-mediated CARD11ΔID association and inhibition, but not necessary, since other GAKIN fragments including the MBS, TR1, and TR2, are also capable of associating with and inhibiting CARD11ΔID. We note that these studies do not address whether these domains bind CARD11 directly or indirectly since the association assays were done in the presence of other cellular proteins. Further studies will be required to understand how the multiple domains of GAKIN integrate function prior to and during antigen receptor signaling.
Bcl10 is an obligate component of the TCR pathway (Ruland et al., 2001) and its recruitment to CARD11 is an essential step in TCR signaling (Blonska and Lin, 2009). The dynamic association of GAKIN with CARD11 appears to reduce the fraction of CARD11 molecules that associate with Bcl10 during signaling (Figure 3I and J). Bcl10 is also targeted by other mechanisms implicated in the termination of TCR signaling to NF-κB, including its signal-induced phosphorylation and degradation (Lobry et al., 2007; Scharschmidt et al., 2004; Zeng et al., 2007). The ability of GAKIN to compete with Bcl10 for CARD11 binding likely works in concert with these degradative mechanisms to accomplish the negative feedback that limits Bcl10 activity in TCR signaling.
The action of GAKIN on CARD11 activity represents a previously unrecognized mechanism used by lymphocytes to limit the extent of antigen receptor signaling to NF-κB. GAKIN is a signaling inhibitor that is poised to act after signaling has been initiated to tune signaling output. GAKIN does not act to terminate signaling, or to prevent signaling in the absence of receptor engagement. This is in contrast to other key inhibitory players in the pathway, including the E3 ligase Cbl-b, the deubiquitinases A20 and CYLD, and the kinase CK1α. Cbl-b appears to act upstream of CARD11 in the pathway, by inhibiting the extent of Akt and PKCθ activation following receptor engagement (Qiao et al., 2008). Downstream of CARD11, A20 appears to dampen pathway output after signaling has begun by removing the K63-linked ubiquitin chains on MALT1 that contribute to the association between MALT1 and the IKK complex (Duwel et al., 2009). CYLD has been shown to prevent spontaneous NF-κB activation in T cells in the absence of receptor engagement by inhibiting the ubiquitination and autoactivation of TAK1 (Reiley et al., 2007). CK1α plays both positive and negative roles in this pathway, and appears to attenuate signaling by phosphorylating residues in the CARD11 ID, although the mechanistic outcome of these phosphorylation events remains unclear (Bidere et al., 2009).
Multiple checkpoints appear to be independently required to tune the output of antigen receptor signaling, to prevent the unwarranted expansion and transformation of lymphocytes observed in lymphoma, and to ensure an appropriate adaptive immune response that does not harm the host. The role of GAKIN in TCR signaling may offer opportunities for the directed manipulation of lymphocyte activation and proliferation that might prove useful in the treatment of autoimmunity, immunodeficiencies, or lymphocytic cancers.
Screening was done using pools of 100 cDNAs/pool isolated from a human spleen expression library (Origene Technologies Inc). The primary screen was carried out as previously described (Pomerantz et al., 2002) except that 50 ng pcCARD11 was cotransfected along with 20 ng Igκ2-IFN-LUC, 6 ng of pCSK-LacZ and 300 ng pool cDNA into HEK293T cells by the calcium phosphate method. Pools altering CARD11 activity at least 3-fold were further purified to a single cDNA by sib selection and sequenced for identification. Secondary screens to confirm NF-κB specificity were conducted in a similar fashion except that 50 ng pCMV6-C/EBPδ was used instead of pcCARD11 to activate 20 ng of a mutant NF-κB reporter (MUT-IFN-LUC). Samples were harvested 40 hr after transfection in 100 μl Promega Lysis buffer and lysates were used to measure Luciferase (Promega) and β-gal activity (Roche) as previously described. Results were calculated as the fold luciferase change, normalized to β-gal activity, compared to the empty vector sample.
Jurkat T cells were plated in 6-well plates at 2.5×105 cells/ml and 2 ml/well. In some assays, Fugene-6 (Roche) or LT-1 (Mirus) was used with 3 μg total DNA per manufacturer’s instructions. Other assays used Lipofectamine LTX (Invitrogen) and 2.5 μg total DNA. Transfections included 200 ng pCSK-LacZ and 750-2000ng Igκ2-IFN-LUC. In each experiment, each sample was supplemented with empty parental expression vector to keep the total amount of expression vector constant. Approximately 40 hr after transfection, cells were stimulated in 1ml media alone or with 75 ng/ml TNFα (Sigma T6674), 1 μg/ml each of mouse anti-human CD3 (BD Pharmingen 555329), mouse anti-human CD28 (BD Pharmingen 555725), anti-mouse IgG1 (BD Pharmingen 02231D), or Raji B cells pulsed with superantigen. To stimulate with the Raji B cells, the B cells were first resuspended in fresh media containing 0.4ng/ml Staphylococcal Enterotoxin E (SEE, Toxin Technology) and incubated for 30 min 37°C 5% CO2 prior to being added to Jurkats at a B:T ratio of 8:1. Stimulations were carried out for 4-6 hr before harvesting samples in 150μl Promega Lysis buffer and assayed for Luciferase and β-gal activity. Results are shown as the fold luciferase change, normalized to β-gal activity, compared to the empty vector sample. In Figures 1B, 1E, ,2A,2A, and and3C,3C, the average fold stimulation under each condition was normalized to that observed with the corresponding unstimulated sample and the unstimulated sample value was set to 1.
Approximately 40 hr after transfection of HEK293T cells via calcium phosphate, cells were harvested in 500 μl IP Lysis buffer, incubated 10 min on ice, and debris cleared by centrifugation at 18300 × g 4°C. Lysates were precleared by incubating with 7 μl bed volume Protein G Sepharose (Amersham) twice for 30 min at 4°C with rotation. An aliquot was removed for input analysis and the remaining lysate was incubated with 1 μg IP antibody for 1.5-2hrs 4°C with rotation. Protein G Sepharose (7 μl bed volume) preblocked with 1% human insulin was added to samples and incubated 1 hr 4°C with rotation. The beads were then washed with rotation 4×5 min 4°C with IP Lysis Buffer before being boiled in the presence of SDS Loading Buffer. IPs were analyzed by western blot using anti-myc (Santa Cruz SC-40), anti-HA (Santa Cruz SC-7392), and M2 anti-FLAG (Eastman Kodak IB13026). For the co-IPs in Figure 3G and 3H, the procedure was carried out as described above except that the lysates were incubated with 2 μg rabbit anti-FLAG and 14 μl bed volume Protein G Sepharose.
To assess the GAKIN:CARD11 association, 108 Jurkat T cells/sample were resuspended in media alone or with 50 ng/ml PMA (Sigma) and 1 μM Ionomycin (Sigma) and stimulated for the indicated times. Cells were then plunged in an ice water bath for 10 min then spun for 10 min 423 × g. Cell pellets were resuspended in 1.5 ml IP Lysis Buffer and processed as above except that the lysates were incubated with rotation, overnight at 4°C with 2 μg rabbit anti-FLAG (Sigma F7425) or 10 μl rabbit anti-GAKIN serum (Covance Research Products Inc.) after the preclear step. CARD11 was detected with goat anti-CARD11 (Imgenex IMG-3653) and GAKIN was detected with rabbit anti-GAKIN serum.
To observe conjugates, 1×106 c/ml of Raji B cells were incubated with 2 μg/ml SEE for 1 hr then washed twice before being resuspended in fresh media at a concentration of 0.5-1×106 cells/ml. Jurkat T cells stably expressing the indicated proteins were resuspended in fresh media to 0.25-0.5×106 cells/ml. Approximately 100μl of T cells were added to the center of a glass bottom culture dish (14mm Microwell, No. 0 Coverglass, MatTek, #P35G-014-C), and approximately 100μl SEE-coated Raji B cells were added dropwise directly to the T cells. 4-D imaging was done by collecting 15 × 1.5μm planes every 20-30 seconds for a total of 60 time points. All images were taken on a motorized Marianas Imaging system (Intelligent-Imaging Innovations) in a humidified, 7.5% CO2 chamber maintained at 37°C and were collected and processed using Slidebook 4.2 or 5.0 (Intelligent-Imaging Innovations). An intensity-based segment mask that indicated pixels within the top 40% of the intensity range was used to determine localization of GAKIN-GFP, ΔMotor-GAKIN-GFP, and CARD11-mCherry. The segment mask for PKCθ-eGFP and PKCθ-mCherry was set to include the top 30% of the intensity range. Line intensity analysis was done using ImageJ; all statistical analysis was carried out in Prism and p values were calculated from a two-tailed, unpaired Student’s t test.
Additional Experimental Procedures are detailed in the Supplementary Material.
We thank M. Meffert, J. Yang, R. Tsien, N. Hacohen, C. Drake, and J. Sisk for reagents; S. Lew, E. Eyler, R. Jattani, A. Kim, R. McCully, d. Mackie and M. Meffert for critical reading of the manuscript; and S. Desiderio, H. Ike, H. Kupfer, M. Spitaler, M. Meffert, T. Bruno and D. Getnet for advice and discussions.
This work was supported by RO1AI078980, and PO1AI072677 from the NIH, RSG-06-172-01-LIB from the American Cancer Society, and funds from the Johns Hopkins University Institute for Cell Engineering. R.L. was supported by Ruth L. Kirschstein National Research Service Award F31AG031689. J.L.P is a recipient of a Kimmel Scholar Award from the Sidney Kimmel Foundation for Cancer Research and a Rita Allen Foundation Scholar.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.