|Home | About | Journals | Submit | Contact Us | Français|
Rap1-GTP-interacting adaptor molecule (RIAM), an adaptor molecule of the Mig-10/RIAM/Lamellipodin (MRL) family, plays a critical role in actin reorganization and inside-out activation of integrins in lymphocytes and platelets. We investigated the role of RIAM in T cell receptor (TCR)-mediated signaling. Elimination of endogenous RIAM by short hairpin RNA (shRNA) resulted in impaired generation of inositol 1,4,5-trisphosphate (IP3) and mobilization of intracellular Ca2+, whereas phosphorylation of ζ chain–associated protein kinase of 70 kD (ZAP-70) and formation of the linker of activated T cells (LAT) signalosome were unaffected. Knockdown of RIAM also resulted in defective nuclear translocation of the transcription factor nuclear factor of activated T cells (NFAT) and activation of Ras guanine nucleotide-releasing protein 1 (RasGRP)1, which led to the diminished transcription of Il2. These events were associated with the impaired translocation of phosphorylated phospholipase C γ1 (PLC-γ1) to the actin cytoskeleton, which was required for the recruitment of PLC-γ1 to the immediate proximity of its substrate phosphatidylinositol 4,5 bisphosphate [PtdIns(4,5)P2], and were reversed by reconstitution of cells with RIAM. Thus, by regulating the activation of PLC-γ1, RIAM has a central role in the activation of T cells and the transcription of target genes.
Binding of the T cell receptor (TCR) to antigens initiates a cascade of molecular events that results in the phosphorylation of tyrosine residues in various substrates, mobilization of Ca2+, activation of signaling pathways that involve mitogen-activated protein kinases (MAPKs) or stress-activated protein kinases [SAPKs, also known as c-Jun N-terminal kinases (JNKs)], and reorganization of the cytoskeleton. Filaments of cytoskeletal actin have a dynamic role during these events and participate in the initiation of molecular movements on the surface of T cells. Reorganization of the actin cytoskeleton is not only a consequence of but also a requirement for T cell activation because treatment of T cells with cytochalasin D, which destabilizes the actin network, abrogates TCR-mediated transcription of Il2, the gene that encodes the cytokine interleukin 2 (IL-2) (1, 2).
Recruitment and activation of phospholipase C γ1 (PLC-γ1) is a key step in the activation process triggered by the TCR (3). Activated PLC-γ1 hydrolyzes phosphatidylinositol 4,5 bisphosphate [PtdIns(4,5)P2] to generate inositol 1,4,5-trisphosphate (IP3), which stimulates the release of Ca2+ from intracellular stores, and diacylglycerol (DAG), which activates protein kinase C (PKC) and signaling pathways dependent on the guanine nucleotide exchange factor (GEF) Ras guanine nucleotide-releasing protein (RasGRP) (4, 5). The accepted model of PLC-γ1 regulation in T cells postulates that the N-terminal Src homology 2 (SH2) domain of PLC-γ1 is both necessary and sufficient for its recruitment to the TCR complex and its phosphorylation following engagement of the TCR, whereas the C-terminal SH2 and the SH3 domains of PLC-γ1 are dispensable (6–8). All three SH domains of PLC-γ1 are required for efficient phosphorylation and activation of PLC-γ1 in T cells; however, recruitment of PLC-γ1 to the signaling complex alone is not sufficient for its activation (9).
Rap1-GTP-interacting adaptor molecule (RIAM), an effector of the small guanosine triphosphatase (GTPase) Rap1, is a member of the MRL family of adaptor molecules, which also includes lamellipodin (Lpd) and its Caenorhabditis elegans ortholog, Mig-10 (10–12). Each of these proteins contains an N-terminal coiled-coil region, central Ras-association and pleckstrin homology (PH) domains, a proline-rich C-terminal region, multiple FPPPP motifs that interact with the Ena-VASP homology 1 (EVH1) domains of the actin regulatory proteins Ena and vasodilator-stimulated phosphoprotein (VASP), and multiple XPPPP motifs that interact with profilin. RIAM is implicated in inside-out signaling, a process of activation-induced modulation of integrin activation through antigen receptors (or other surface receptors) that leads to integrin-mediated adhesion (13). Specifically, RIAM interacts with Rap1-GTP to promote adhesion through β1 and β2 integrin subunits in T cells and adhesion through the integrin αIIβ3 in platelets (10, 14, 15). Several proteins involved in inside-out signaling are components of TCR signaling pathways and have active roles in mediating TCR signaling (13). Moreover, RIAM is recruited to the contact site between the antigen-presenting cell (APC) and the T cell during activation of the T cell (16). Because of these properties, we sought to examine whether RIAM might have a role in regulating signaling events activated by the TCR.
Here, we report that RIAM directly and constitutively interacts with the SH3 domain of PLC-γ1 and is a regulator of the activity of PLC-γ1. Elimination of endogenous RIAM by short hairpin RNA (shRNA) resulted in the impaired generation of IP3 and mobilization of intracellular Ca2+, and defective nuclear translocation of the transcription factor nuclear factor of activated T cells (NFAT). In addition, activation of Ras was impaired due to the defective activation of the diacylglycerol (DAG)- and Ca2+-dependent GEF RasGRP1. These events were associated with the impaired translocation of phosphorylated PLC-γ1 to the actin cytoskeleton. Thus, by regulating the spatiotemporal distribution of activated PLC-γ1, RIAM plays a central role in the generation and functional outcome of TCR mediated signals.
RIAM is recruited to the plasma membrane at the site of the immunological synapse (IS) during the formation of T cell-APC conjugates (16). Because ZAP-70 is a critical TCR-proximal signaling molecule that localizes at the supramolecular activation complex (SMAC) during formation of the IS, we examined whether RIAM might colocalize with ZAP-70. In experiments with Jurkat T cells incubated with Raji B cells that were loaded with the superantigen SEE, which served as APCs, we determined that RIAM colocalized with ZAP-70 at the T cell-APC contact site (fig. S1).
RIAM contains proline-rich motifs that bind to EVH-1 domain–containing proteins and to profilin (10). We used the iSPOT tool and identified several proline-rich sequences in RIAM that may be binding motifs for SH3 domain– and WW domain–containing proteins. Among others, RIAM also contains sites that are predicted to act as regions for the binding of members of the Src family of kinases. Because Fyn and Lck are the most TCR-proximal Src family kinases and are activated during formation of the IS, we examined whether RIAM might interact with or serve as a substrate of these enzymes. Cos cells were transiently cotransfected with plasmids encoding the kinase-active form of Fyn (Fig. 1A) or Lck (Fig. 1B) together with either empty vector or a plasmid encoding Myc-tagged RIAM. Immunoprecipitations with antibodies against Fyn or Lck followed by analysis of Western blots of these samples with antibodies against phosphotyrosine (pTyr) residues or the Myc tag indicated that RIAM underwent tyrosine phosphorylation and coimmunoprecipitated with Fyn and Lck (Fig. 1, A and B). Because RIAM colocalized with ZAP-70 in the IS (fig. S1), we also examined whether RIAM might be a substrate of ZAP-70. Experiments in cells cotransfected with plasmids encoding RIAM and ZAP-70 indicated that RIAM associated with and was a substrate of ZAP-70 (Fig. 1C).
To determine whether RIAM interacted with Fyn and ZAP-70 in primary cells we immunoprecipitated RIAM from extracts of primary human T cell extracts. RIAM underwent tyrosine phosphorylation after T cell activation (Fig. 1D). RIAM displayed a weak association with Fyn and ZAP-70 prior to stimulation that was increased in strength upon stimulation (Fig.1D), indicating that the interaction of RIAM with these kinases could be detected in primary T cells.
Because RIAM is a substrate of TCR-proximal protein tyrosine kinases, we hypothesized that RIAM might have a role in regulating TCR-mediated signaling events and the functional outcome of T cell activation. To address this, we generated various shRNAs specific for RIAM and established stable cell lines containing either RIAM-specific shRNAs or control shRNA sequences, which we named RIAM-KD and control-KD, respectively (fig. S2). RIAM-KD cells displayed an impaired ability to produce IL-2 in response to stimulation with SEE-loaded APCs or crosslinking of CD3 and CD28 compared to that in control-KD cells (Fig. 2A). Similarly, RIAM-KD cells displayed decreased activity of the Il2 promoter in response to crosslinking of CD3 and CD28 (Fig. 2B). Thus, RIAM is required for the transcription of Il2 in response to crosslinking of CD3 and CD28 and it appears to regulate signaling events upstream of Ras and Ca2+, because responses of RIAM-KD cells to stimulation with PMA and ionomycin, which directly activate Ras and cause mobilization of intracellular Ca2+, respectively, were unaffected by the loss of RIAM (Fig. 2, A and B).
Despite the impairment in the expression of Il2 in RIAM-KD cells, analysis of TCR-proximal signaling events did not show any impairment in the activation of ZAP-70, the phosphorylation of LAT, or the formation of the LAT signalosome, which consists of phosphorylated PLC-γ1, the adaptor protein SH2 domain–containing leukocyte phosphoprotein of 76 kD (SLP-76), and the GEF Vav1 (Fig. 2, C to E) (17).
Because TCR-proximal signaling events, but not production of IL-2, remained unaltered when endogenous RIAM was depleted, we examined signals downstream of the TCR. Our analysis showed that phosphorylation of PLC-γ1 occurred to a similar extent in control-KD and RIAM-KD cells (Fig. 3, A and B). Activation of PLC-γ1 hydrolyzes PtdIns(4,5)P2 to generate the second messengers DAG and IP3. IP3 binds to IP3 receptors and triggers the release of Ca2+ from intracellular stores, such as the endoplasmic reticulum (18). Stimulation of the TCR of RIAM-KD cells resulted in a markedly reduced generation of IP3 compared to that of control-KD cells (Fig. 3C), which suggested that the activation of PLC-γ1 was impaired in RIAM-KD cells. Consistent with the diminished generation of IP3, mobilization of Ca2+ in RIAM-KD cells was also impaired (Fig. 3D and fig. S3A). This was due to impaired IP3-mediated release of Ca2+ from the ER and not to defects in store content or in Ca2+ release–activated calcium (CRAC) channel entry as determined in experiments with the Ca2+ ATPase blocker thapsigargin, (fig. S3B). Consistent with this impaired release of Ca2+, dephosphorylation of NFAT, which is dependent on the Ca2+-mediated activation of the phosphatase calcineurin, was impaired (Fig. 3, E and F), and nuclear translocation of NFAT was abrogated in RIAM-KD cells (Fig. 3, G and H).
The activation of Ras upon stimulation of the TCR is dependent on the complex formed between phosphorylated LAT (pLAT), Grb2, and SOS, and on Ca2+- and DAG (CalDAG)-mediated activation of RasGRP1 (19, 20). The latter event represents the dominant mechanism that regulates the TCR-dependent activation of Ras (20, 21) and requires the activation of PLC-γ1 (22). Consistent with the impaired activation of PLC-γ1, thus leading to defective generation of IP3 and mobilization of Ca2+, the activation of Ras was impaired in RIAM-KD cells (Fig. 4, A and B). Activation of mitogen-activated or extracellular signal–regulated protein kinase kinase 1 (MEK1) and MEK2, and the mitogen-activated proteins kinases (MAPKs) extracellular signal–regulated kinases 1 (ERK1), ERK2, and c-Jun N-terminal kinase (JNK) downstream of Ras was also defective in the absence of RIAM, whereas activation of p38 MAPK was mostly unaffected (Fig. 4, C to G). These results are consistent with a specific role for RIAM in PLC-γ1-mediated events downstream of the TCR and indicate that deletion of RIAM does not result in a generalized defect in TCR-mediated signaling.
Activation of RasGRP1 downstream of PLC-γ1 is associated with the partitioning of RasGRP1 to membrane fractions of cells (19) and leads to the activation of Ras at the Golgi apparatus (22, 23). To further investigate whether the impaired activation of Ras in RIAM-KD cells was due to defective activation of RasGRP1, we employed a previously established subcellular fractionation approach to assess partitioning of RasGRP1 into the membrane fraction (4, 19). Membrane translocation of RasGRP1 after stimulation of the TCR was readily detected in control-KD cells but was abrogated in RIAM-KD cells (Fig. 4, H and I). CalDAG-GEFI, another member of the family of CalDAG-dependent GEFs, mediates the activation of Rap1 (24) and, similarly to RasGRP1, displays partitioning to the membrane fraction upon activation (25). We observed that membrane localization of CalDAG-GEFI was readily detected upon stimulation of the TCR in control-KD cells but not in RIAM-KD cells (Fig. 4, H and J). Consistent with this finding, activation of Rap1 was impaired in RIAM-KD cells compared to that in control-KD cells (Fig. 4, K and L).
The specific role of RIAM as a regulator of signaling downstream of Ras secondary to its role in the activation of PLC-γ1 was confirmed by the ability of RIAM-KD cells to activate Ras and its downstream kinases MEK1, MEK2, ERK1, ERK2, and JNK after their reconstitution with a plasmid encoding a variant RIAM (RIAM-KD-R) that contained seven base mismatches with the RIAM-specific shRNA (Fig. 4, A to F). Reconstitution with the shRNA-resistant RIAM also restored the ability of RasGRP1 and CalDAG-GEFI to partition to the membrane fraction of RIAM-KD cells in response to stimulation (Fig. 4, H to J) and to activate Rap1 (Fig. 4, K and L).
Generation of IP3 requires the appropriate docking and positioning of activated PLC-γ1 (26–28). Because our data were consistent with impaired activation of PLC-γ1 in the absence of RIAM, we examined whether RIAM might interact with PLC-γ1 and regulate its localization after stimulation of T cells. Immunoprecipitation experiments with RIAM-specific antiserum followed by Western blotting analysis with an antibody against PLC-γ1 revealed a constitutive association between RIAM with PLC-γ1 that was enhanced after stimulation of the TCR and the tyrosine phosphorylation of RIAM (Fig. 5A). Quantitative analysis indicated that approximately 15% of the total amount of each protein was constitutively associated with the other and that this fraction was increased to approximately 30% after stimulation of the TCR.
PLC-γ1 contains two SH2 domains and one SH3 domain. Based on our data showing a constitutive association between RIAM and PLC-γ1 that was enhanced after stimulation of the TCR (Fig. 5A), we hypothesized that RIAM might interact with the SH3 domain of PLC-γ1 through its proline-rich regions (PRRs) prior to stimulation of T cells; after undergoing tyrosine phosphorylation, RIAM might also interact with the SH2 domains of PLC-γ1. To test these possibilities, we first investigated whether the interaction between RIAM and PLC-γ1 was direct. We examined a potential interaction between PLC-γ1 and RIAM in an in vitro protein association assay. Glutathione S-transferase (GST) and GST fusion proteins of full-length PLC-γ1, the N-terminal and C-terminal SH2 domains of PLC-γ1 (PLC-γ1-N+C-SH2), and the SH3 domain of PLC-γ1 (fig. S4A) were coupled to glutathione-sepharose and incubated with [35S] methionine–labeled RIAM or luciferase, as a negative control. Full-length PLC-γ1 and the SH3 domain of PLC-γ1 bound to RIAM (Fig. 5B), indicating that RIAM interacts directly with PLC-γ1. The interaction between RIAM and PLC-γ1-N+C-SH2 was similar to that of the GST control (Fig. 5B). These results indicate that RIAM interacted with the SH3 domain of PLC-γ1.
We further examined domain-specific interactions of PLC-γ1 with endogenous RIAM in primary human T lymphocytes in experiments with GST-fusion proteins of PLC-γ1. We detected a constitutive interaction between RIAM with PLC-γ1-SH3 that was enhanced after activation of T cells (Fig. 5C). In contrast, no interaction between RIAM and PLC-γ1-N+C-SH2 was detected after the activation of T cells (Fig. 5C), although RIAM underwent tyrosine phosphorylation in activated T cells (Fig. 5A, middle panel). Consistent with our inability to detect an interaction between RIAM and PLC-γ1-N+C-SH2, we could not identify a sequence related to the consensus binding motif (YLVV) for the N-terminal and C-terminal SH2 domains of PLC-γ1 in RIAM.
To identify the region of RIAM that associates with PLC-γ1, we generated GST fusion proteins of full-length RIAM and two truncation mutants, RIAM-D1, which contains only the RA and PH domains of RIAM, and RIAM-C1, which contains only the C-terminal PRR of RIAM (fig. S4B), and assayed them for their ability to interact with PLC-γ1. Full-length RIAM and RIAM-C1 associated with PLC-γ1, whereas RIAM-D1 did not (Fig. 5D). Thus, the C-terminal PRR of RIAM was required for the interaction between RIAM and PLC-γ1, consistent with our observation that PLC-γ1-SH3 displayed robust interaction with RIAM in vitro and in cells (Fig. 5, B and C). Consistent with the specific functional importance of the various regions of RIAM in regulating the association between RIAM and PLC-γ1 and the activation of signaling downstream of PLC-γ1, RIAM-D1 did not restore the capacity of RIAM-KD cells to activate MEK1, MEK2, ERK1, ERK2, or JNK in response to CD3- and CD28-mediated stimulation of cells (fig. S5).
Next, we examined whether the interaction between RIAM and PLC-γ1 might be altered upon phosphorylation of PLC-γ1. We used a GST-fusion protein containing the C-terminal PRR of RIAM (RIAM-C1), which we identified as the region of RIAM required for its association with PLC-γ1, and performed pull-down assays with extracts of T cells before and after stimulation. The interaction between GST-RIAM-C1 and PLC-γ1 was constitutive and was slightly enhanced after activation of T cells (Fig. 5E). Strikingly, upon activation of the T cells, there was a prominent association of RIAM with phosphorylated PLC-γ1 (pPLC-γ1) (Fig. 5E). Thus, although the interaction between PLC-γ1 and RIAM was constitutive (Fig. 5, C E), PLC-γ1 that was associated with RIAM became phosphorylated, resulting in the association of RIAM with activated PLC-γ1 in stimulated cells.
Our studies showed that RIAM could serve as a substrate of the TCR-proximal tyrosine kinases Fyn, Lck, and ZAP-70 (Fig. 1, A to D). Moreover, the constitutive association of RIAM with PLC-γ1 was enhanced after T cell activation and tyrosine phosphorylation of RIAM (Fig. 5A). To determine whether tyrosine phosphorylation of RIAM might play a role in regulating the interaction between RIAM and PLC-γ1 following the activation of T cells, we immunoprecipitated RIAM from extracts of the Lck-deficient Jurkat T cell line J.CaM1 before and after stimulation. Under these conditions, tyrosine phosphorylation of RIAM was essentially undetectable (Fig. 5F). Moreover, the extent of the interaction between RIAM and PLC-γ1 remained unaltered (Fig. 5F, top panel) in contrast to the increase in the amount of PLC-γ1 that coimmunoprecipitated with RIAM following stimulation of control T cells (Fig. 5A). Taken together, these results indicate that although RIAM interacts with only the SH3 domain of PLC-γ1, phosphorylation of RIAM by Lck upon stimulation of T cells increased the extent of this interaction. Because Lck can also phosphorylate PLC-γ1 (29), the enhanced association of RIAM and PLC-γ1 upon phosphorylation might be related to conformational changes of these molecules, although the nature of their interaction remains unaltered.
Studies in other cell types have shown that PLC-γ1 needs to be targeted to the cytoskeleton to act on its specific substrate PtdIns(4,5)P2, which is localized in adjacent membranes; however, the mechanism of this translocation has not been determined (26, 27). RIAM is recruited to polymerized actin at sites of rapid actin turnover (10, 30). Because our studies indicated that RIAM displayed a constitutive interaction with PLC-γ1 and activation-dependent interaction with pPLC-γ1, we investigated whether RIAM might regulate the localization of pPLC-γ1 to the actin cytoskeleton. Isolation of cytoskeletal fractions from cells before and after stimulation revealed evidence of an enhanced association of pPLC-γ1 with the actin cytoskeleton upon T cell activation and showed that this event was impaired in RIAM-KD cells compared to that in control-KD cells (Fig. 6, A and B). Localization of PLC-γ1 at the actin cytoskeleton was also impaired in RIAM-KD cells (Fig. 6A, C). These events were recovered by reconstitution of RIAM-KD cells with RIAM, but not with the RIAM-D1 mutant protein (Fig. 6, A to C).
To further assess the role of RIAM in regulating the recruitment of pPLC-γ1 to the actin cytoskeleton, we performed confocal microscopy to examine the intracellular localization of pPLC-γ1 upon T cell activation. Control-KD, RIAM-KD, and RIAM-KD-R cells were stimulated with monoclonal antibodies (mAbs) against CD3 and CD28, seeded on coverslips, fixed, and incubated with an antibody against pPLC-γ1 and with phalloidin to visualize polymerized actin (F-actin). In control-KD cells (Fig. 6D, first row), pPLC-γ1 was detected predominantly at the plasma membrane where it colocalized with F-actin (Fig. 6, D and E). In RIAM-KD cells, however, membrane recruitment was substantially impaired and pPLC-γ1 was instead detected predominantly in the cytoplasm (Fig. 6D, second row). In these cells, colocalization of pPLC-γ1 and F-actin was dramatically reduced (Fig. 6, D and E). Quantification of the effect of knockdown of RIAM on the colocalization of pPLC-γ1 and F-actin by assessing the overlap between green and red channel pixels in confocal images indicated a 75% loss in the extent of colocalization across the cell (Fig. 6F). These effects were reversed by the reconstitution of RIAM-KD cells with RIAM (Fig. 6D, third row and Fig. 6, E and F), but not with the RIAM-D1 mutant (Fig. 6D, fourth row and Fig. 6, E and F). Incubation with isotype matched control Ig, followed by FITC-conjugated antibody did not show detectable binding (fig. S6A). These results indicate that RIAM has a critical role in regulating the localization of pPLC-γ1 to the actin cytoskeleton following the activation of T cells.
Our data showed that RIAM-KD cells exhibited impaired generation of IP3 upon stimulation compared to that in control cells (Fig. 3C). Activated PLC-γ1 generates IP3 from PtdIns(4,5,)P2 at the plasma membrane. Because RIAM-KD cells displayed defective translocation of pPLC-γ1 to the cytoskeleton and impaired generation of IP3, we hypothesized that RIAM might function as a docking site to position PLC-γ1 in close proximity to PtdIns(4,5)P2-containing membrane. To determine whether cytoskeleton-associated pPLC-γ1, recruited upon T cell activation, might come in closer proximity to its substrate than would the free pool of PLC-γ1, we examined whether PtdIns(4,5)P2 might exhibit close contact with F-actin. Analysis by confocal microscopy revealed that PtdIns(4,5)P2 was detected at the plasma membrane (Fig. 6G), where it colocalized with F-actin (Fig. 6, G to I). Incubation with FITC-conjugated isotype matched control Ig did not show detectable binding to the cells (fig. S7A). These results indicate that recruitment of pPLC-γ1 to the actin cytoskeleton by RIAM brings this enzyme in immediate proximity to its substrate, PtdIns(4,5)P2, which colocalizes with F-actin.
To determine whether our findings in RIAM-deficient Jurkat T cells had direct biological relevance, we performed experiments in primary T lymphocytes isolated from healthy volunteer donors. Endogenous RIAM in these cells was depleted by RIAM-specific shRNA. Control-KD and RIAM-KD primary T lymphocytes were stimulated with antibodies against CD3 and CD28, seeded on coverslips, fixed, and incubated with an antibody against pPLC-γ1 and with phalloidin to visualize F-actin. Similarly to our observations of Jurkat cells, pPLC-γ1 was predominantly detected at the plasma membrane in control-KD primary T lymphocytes, where it colocalized with F-actin (Fig. 7, A and B). In contrast, membrane recruitment of pPLC-γ1 was almost undetectable in RIAM-KD lymphocytes (Fig. 7, A and B); instead, pPLC-γ1 was predominantly distributed in the cytoplasm (Fig. 7A). Quantification of the colocalization of pPLC-γ1 and F-actin in confocal images by assessment of red and green channel pixel overlap revealed that colocalization of PLC-γ1 and F-actin was 15% in RIAM-KD T lymphocytes compared to 95% in control-KD cells (Fig. 7B, C). Incubation with isotype matched control Ig followed by AMCA-conjugated antibody did not show detectable binding (fig. S6B).
To determine whether active pPLC-γ1 could best access PtdIns(4,5)P2 when is recruited to the cortical actin cytoskeleton underlying the plasma membrane in primary T cells by RIAM, we examined whether PtdIns(4,5)P2 exhibited close contact with F-actin. Confocal microscopy showed that PtdIns(4,5)P2 was detected at the plasma membrane (Fig. 7D) where it colocalized with F-actin (Fig. 7, D to F), indicating that recruitment of pPLC-γ1 to the actin cytoskeleton by RIAM brings the enzyme in immediate proximity to its substrate in primary T cells. Incubation with FITC-conjugates isotype matched control Ig did not show detectable binding to these cells (fig. S7B). Furthermore, we assessed the mobilization of Ca2+ in primary T cells to determine whether these events had functional implications on signaling downstream of PLC-γ1. RIAM-KD primary T lymphocytes, displayed impaired Ca2+ flux upon stimulation compared to that in control-KD T lymphocytes (Fig. 7D).
The actin cytoskeleton participates in organizing and maintaining signaling complexes and pathways initiated from cell surface receptors in T and B lymphocytes (31). Although actin regulatory proteins have a mandatory role in this process, how they participate in T cell activation remains poorly understood. Here, we determined that RIAM, an adaptor molecule that interacts with Ena/VASP cytoskeletal regulators, has a central role in T cell signaling by regulating the activation of PLC-γ1 and the positioning of pPLC-γ1. Our data showed that although the overall abundance of pPLC-γ1 was unaffected by the absence of RIAM, translocation of pPLC-γ1 from the cytosol to the actin cytoskeleton was impaired. The direct physical interaction between RIAM and PLC-γ1, the displacement of pPLC-γ1 from the actin cytoskeleton, and the impaired generation of IP3 in the absence of RIAM strongly suggest that RIAM may function as a docking site for PLC-γ1 to position this enzyme in close proximity to PtdIns(4,5)P2-containing membranes (Fig. 8). In agreement with this hypothesis, our studies showed that PtdIns(4,5)P2 was located in close proximity to the F-actin cytoskeleton in T cells.
The SH3 domain of PLC-γ1 is thought to be either dispensable for its activation or an inhibitor of PLC-γ1 activity (6, 8, 32). This region of PLC-γ1 can bind to various proline-rich domain–containing proteins, including SLP76 (33), SOS (34), Itk (35), and PIKE (36), but the importance of these interactions and their occurrence following engagement of the TCR in vivo remains unclear. In contrast, other work has indicated that all three SH domains of PLC-γ1 are required for the efficient phosphorylation and activation of PLC-γ1 in T cells, and that recruitment of PLC-γ1 to the TCR signaling complex alone is not sufficient to activate PLC-γ1 (9). Despite the vital function of PLC-γ1 in T cells and its ubiquitously pivotal role in signal transduction, the mechanism and the precise intracellular location of PLC-γ1 activation are incompletely understood.
Previous studies in other cell types that used GST fusion proteins have shown that the SH3 domain of PLC-γ1 directs its localization to the actin cytoskeleton (28). Moreover, in fibroblasts and epithelial cells, PLC-γ1 associates with the actin cytoskeleton upon activation (26, 27). In those studies, however, although tyrosine phosphorylation of PLC-γ1 did not appear to determine its subcellular localization, there was a substantial increase in the cytoskeletal association of pPLC-γ1 in response to stimulation (26, 27). Moreover, these studies determined that that activity of PLC-γ1 was almost exclusively detected in the cytoskeletal fraction (26). Our present findings provide potential new insights into the mechanisms underlying those observations. Our results indicate that PLC-γ1 constitutively interacted with the C-terminal PRR of RIAM through its SH3 domain, in a manner independent of the tyrosine phosphorylation of PLC-γ1. However, RIAM-associated PLC-γ1 became phosphorylated upon stimulation of T cells, which resulted in an interaction between RIAM and activated PLC-γ1. Because it translocates to sites of actin polymerization upon cellular stimulation (10, 30), RIAM may mediate the concomitant translocation of pPLC-γ1, resulting in the increased cytoskeletal association of pPLC-γ1 upon stimulation (Fig. 8).
PLC-γ1 is thought to mediate its effects on PtdIns(4,5)P2 at the plasma membrane where, in T cells, it is recruited in a LAT-dependent manner (17). However, PLC-γ1 and its substrate localize in other intracellular structures, including endosomes (37) and phosphoinositide-rich vehicles (38). In RIAM-deficient cells there was no difference in LAT-associated PLC-γ1 compared to that in control cells, suggesting that RIAM regulated the activation of PLC-γ1 in a LAT-independent manner. Consistent with this hypothesis, RIAM-PLC-γ1 complexes did not contain LAT and conversely LAT-PLC-γ1 complexes did not contain RIAM (fig. S8). Further studies will be required to identify the precise subcellular compartment(s) in which RIAM regulates association of pPLC-γ1 with its substrate during T cell stimulation.
Our present observations indicate that by regulating the activation of PLC-γ1 and the release of intracellular Ca2+, RIAM has substantial implications for the activation of the CalDAG-GEFs that activate Ras and Rap1 GTPases (Fig. 8). In T lymphocytes, stimulation of the TCR leads to the activation of Ras by RasGRP1 in the Golgi apparatus in a PLC-γ1-dependent manner (22). Similarly, activation of Rap1 through the TCR or CXCR4, the receptor for the chemokine CXCL12, is mediated by CalDAG-GEFI and is dependent on PLC-γ1 (39). Thus, by regulating the activity of PLC-γ1, RIAM functions as a regulator of the activation of Ras and Rap1. PLC-γ1 is required only for the activation of Ras on the Golgi and not for the activation of Ras at the plasma membrane (40). Thus, by regulating the activation of PLC-γ1 and RasGRP1, RIAM may play a role in compartmentalized Ras-MAPK signaling, a mechanism that has implications for the outcome of T cell activation, including decisions between positive versus negative selection (41) and the induction of productive immunity versus T cell anergy (42, 43). In conclusion, our results reveal an unexpected role for RIAM in T cell responses that involves regulation of the spatiotemporal distribution and activation of PLC-γ1, leading to generation of IP3 and mobilization of Ca2+ after stimulation of the TCR.
The cDNAs for Lck, Fyn, ZAP-70, and Myc-tagged RIAM were inserted in pAXEF or pcDNA1.1/Amp vectors (Invitrogen). COS cells were transfected with the indicated plasmids with the DEAE-Dextran method, and Jurkat T cells were transfected by electroporation. To create plasmids containing RIAM-specific shRNAs, specific oligonucleotides were cloned into the pLL3.7 vector (44). As controls, pLL3.7 vector containing shRNA for mouse lamellipodin were used. These sequences and cell lines have been previously described (10). An alternative RIAM-specific shRNA sequence (SuperArray) cloned into the GeneClip U1 hairpin cloning vector (Promega) was also used to knockdown endogenous RIAM. As control, GeneClip vector containing a scrambled artificial sequence not matching any human, mouse, or rat gene was used.
Jurkat, J.CaM1, and COS cells were cultured routinely in RPMI 1640 media supplemented with 10% fetal calf serum (FCS). RIAM-KD and control-KD cell lines transfected with the pLL3.7 vector have been previously described (10). Stably transfected RIAM-KD and control-KD cell lines were also generated with an alternative RIAM-specific shRNA sequence (SuperArray) cloned into the GeneClip U1 hairpin cloning vector. Multiple polyclonal, oligoclonal, and monoclonal RIAM-KD and control-KD Jurkat T cell lines were generated by puromycin-based selection. All experiments were repeated with multiple RIAM-KD and control-KD cell lines. To reconstitute cells with RIAM, RIAM cDNA that was resistant to knockdown by RIAM-specific shRNA was generated by introducing silent mutations with seven base mismatches with the Quickchange Lighting site-directed mutagenesis kit (Stratagene) and was used to transfect RIAM-KD cells that contained RIAM-specific shRNA in the GeneClip vector. All mutants were confirmed by sequencing analysis. The same approach was undertaken to reconstitute expression of RIAM-D1 in RIAM-KD cells, with cDNA that encoded His-tagged RIAM-D1. Primary T cells were isolated from leukopheresis products of healthy volunteer donors with the MACS Pan T cell Isolation kit II (Miltenyi Biotech). To generate RIAM-KD and control-KD T lymphocytes, cells were transfected with the pLL3.7 vector expressing either RIAM-specific shRNA or control shRNA. For stimulation experiments, Jurkat cell lines and primary T cells were left untreated or were incubated at 37°C with 10 µg/ml ofa mAb against CD3 and 10 µg/ml of a mAb against CD28 (CLB; distributed by Fitzgerald Industries) for the indicated times. Reactions were stopped by the addition of cold phosphate-buffered saline (PBS), followed by two washes with cold PBS.
For in vitro protein interaction assays, full-length PLC-γ1, the PLC-γ1 (N+C)-SH2 fragment, and the PLC-γ1-SH3 fragment inserted into pGEX vectors (45) (kindly provided by Dr. Sudeep George, Vanderbilt University) were expressed in BL21(D3)LysS chemically competent, protease-free Escherichia coli cells and purified on glutathione-Sepharose (Amersham Pharmacia Biotech). 10 µg of GST and GST fusion proteins were incubated for 1 hour at 4°C with 50 µl of glutathione-sepharose in GST-buffer [10% glycerol, 50 mM tris-HCl (pH 7.5), 150 mM NaCl, 2 mM MgCl2]. 10 µl of [35S]methionine-labeled RIAM or luciferase synthesized in vitro (TNT T7 System Promega) were suspended in GST buffer and incubated for 1 hour at 4°C with the indicated GST-fusion proteins coupled on glutathione-sepharose. The samples were analyzed by 10% SDS polyacrylamide gel electrophoresis (SDS-PAGE) and visualized by autoradiography. 1 µl of [35S]methionine-labeled products were analyzed with the pull-down assays as control.
For preparation of lysates, cells were washed in PBS and lysed in lysis buffer containing 50 mM tris-HCl (pH 7.4), 150 mM NaCl, 2 mM MgCl2, 10% glycerol, and 1% NP-40 supplemented with 2 mM Na3VO4, 1 mM phenylmethylsulfonyl fluoride (PMSF) and protease Inhibitor Cocktail (Sigma). Cell lysates (50 µg) were separated by SDS-PAGE and analyzed by Western blotting with the indicated antibodies. For immunoprecipitations, cell lysates (1 mg) were incubated overnight at 4°C with the indicated antibodies and immune complexes were captured by protein A/G agarose beads (Santa Cruz) and eluted after gentle rotation for 2 hours at 4°C. For assessment of the activation of Ras and Rap1 by GST pull-down, GST-Raf1-Ras binding domain (RBD) and GST-RalGDS-RBD fusion proteins were purified on glutathione-sepharose (Amersham Pharmacia Biotech). 50 µg of fusion protein coupled to glutathione-sepharose beads (GammaBind Plus) were incubated with Jurkat cell lysates (1 mg) in lysis buffer that did not contain NP-40. Eluted proteins were analyzed by SDS-PAGE and Western blotting with antibodies against Ras and Rap1. To assess domain-specific interactions of RIAM with PLC-γ1, the cDNAs of full-length RIAM, the fragment RIAM-D1 (150 to 430), which contains only the RA and PH domains of RIAM, and the fragment RIAM-C1 (420 to 665), which contains only the C-terminal PRR of RIAM, were inserted into pGEX vectors, expressed in E. coli BL21(D3)LysS, and purified on glutathione-sepharose (Amersham Pharmacia Biotech). 10 µg of GST and GST fusion proteins were incubated for 1 hour at 4°C with 50 µl of glutathione-sepharose in GST-buffer and then incubated with cell extracts (500 µg/sample) overnight followed by SDS-PAGE and Western blotting analysis with the indicated antibodies. For Western blotting, standard immunoblotting methods were employed. The following antibodies were used: RIAM polyclonal antibody #4612 (for immunoprecipitation), RIAM #5541 (for Western blotting) (10), and phosphotyrosine-specific mouse mAb (clone 4G10) were from Upstate Biotechnology; a mouse mAb against Ras was from Calbiochem; a rabbit mAb against Rap1 and a mouse mAb against CalDAG-GEFI were from Novus Bologicals; a mouse mAb against phosphotyrosine residues (clone 4G10), rabbit antiserum against Lck, a mouse mAb against ZAP-70, antiserum against Vav, and a mouse mAb against LAT were all from Upstate Biotechnology; a rabbit polyclonal antibody against Fyn, a goat polyclonal antibody against SLP-76, a mouse mAb against pERK, a mouse mAb against pJNK, a rabbit polyclonal antibody against RasGRP1, and goat antiserum against actin were from Santa Cruz Biotechnology; a mouse mAb against PLC-γ1 was from Abcam; rabbit polyclonal antibodies against pPLC-γ1 (Tyr783) and pMEK1/2 (Ser217/221) were from Cell Signaling Technology; a rabbit polyclonal antibody against pp38 (Thr180/Tyr182)was from New England Biolabs; a mouse mAB against the His tag was from Quiagen; rabbit antiserum against NFAT1 (S671.A20) has been previously described (46). Stripping and re-incubation of the blots with other antibodies was performed as described (47). Where indicated, densitometric analysis was performed and quantification of the integrated density of each band was assessed with ImageJ 1.41o software. Statistical significance was defined where p was less than 0.05 and was calculated with the InStat software.
To assess protein localization at the IS, 5 × 106 Raji B cells were incubated with or without 2.5 µg/ml of superantigen SEE (Toxin Technologies) in 500 µl of RPMI, 10% FCS for 30 min at 37°C. Cells were washed 3 times in warm RPMI and incubated with the same number of Jurkat T cells that were labeled with green cell tracker (Sigma). Cells were incubated for 10 min at 37°C, seeded onto poly-L-lysine-coated slides (Sigma), fixed with 4% (w/v) paraformaldehyde in PHEM buffer [60 mM PIPES, 25 mM hepes, 10 mM EGTA, 2 mM MgCl2, 120 mM sucrose (pH 7.3)], and permeabilized with 0.1 % (v/v) Triton X-100 in tris-buffered saline (TBS). Cells were incubated with a mAb antibody against ZAP-70 and an antibody against RIAM, followed by incubation with Cy5-conjugated donkey anti mouse immunoglobulin G (IgG) and TexasRed- or Cy5-conjugated donkey anti-rabbit IgG (Jackson Immunolabs) or with rhodamine-conjugated phalloidin for the detection of F-actin. Samples were analyzed by confocal laser scanner microscope (Zeiss LSM 520). Sixty cell conjugates were analyzed for each staining combination. Images were processed for presentation with Adobe Photoshop. To localize NFAT, RIAM-KD and control-KD cells were settled onto poly-L-lysine-coated coverslips both before and after stimulation. Cells were permeabilized and incubated with an affinity-purified rabbit antibody against NFAT (S67.1A20) (46), followed by incubation with donkey anti-rabbit IgG conjugated with Alexa fluor 647 (Invitrogen) and with DAPI to visualize nuclei. Samples were analyzed by fluorescent microscopy with a Nikon E800 microscope equipped with a Spot RT cooled CCD camera at the Imaging Core of MRDDRC, at Children’s Hospital Boston. Forty-five cells were analyzed for each culture condition and the experiment was repeated at least three times. Colocalization of the red (NFAT) and blue (DAPI) channel pixels was estimated by the LSM 510 software (Carl Zeiss MicroImaging). Final processing of images was performed with Adobe Photoshop. To localize pPLC-γ1, unstimulated and stimulated Jurkat T cells prior were permeabilized and incubated with a goat polyclonal antibody against pPLC-γ1 (Tyr783) and then with a fluorescein isothiocyanate (FITC)-conjugated F(ab´)2 fragment of donkey anti-goat IgG (Jackson Immunoresearch Labs). For primary human T cells, an amino-methyl-coumarin-acetate (AMCA)-conjugated F(ab´)2 fragment of donkey anti-goat IgG was used for the detection of pPLC-γ1 and analysis was performed by focusing on GFP-positive cells that contained the pLL3.7 vector, expressing either RIAM-specific shRNA or control shRNA. Subsequently, the blue channel was converted to green by image processing to improve image quality. F-actin was stained by rhodamine-phalloidin. For the localization of phosphatidylinositol 4,5-bisphosphate (PIP2) in Jurkat cells or primary T lymphocytes, a FITC-conjugated mouse mAb against PIP2 (Echelon Biosciences) was used. Samples were analyzed with a Perkin Elmer UltraView VoX spinning-disk confocal microscope at the Imaging Core of MRDDRC, at Children’s Hospital Boston. A minimum of eight 0.5-µM z-slices were acquired for each cell and a z-stack was created. Forty-five cells were analyzed for each culture condition and the experiment was repeated three times. Quantitative analysis of colocalization was prepared by the LSM 510 software (Carl Zeiss MicroImaging). Final processing of images was performed with Adobe Photoshop. Where indicated, densitometric analysis of image sections was prepared with ImageJ 1.41o software.
To assess the production of IL-2, 5 × 105 RIAM-KD or control-KD Jurkat T cells were stimulated with 5 × 105 SEE-loaded Raji cells in 24-well plates. For stimulation with anti-CD3 (100 ng/ml) and anti-CD28 (500 ng/ml) or with PMA (50 ng/ml) and ionomycin (1 nM), separate aliquots of 5 × 105 Jurkat T cells were cultured for 48 hours, and the concentration of IL-2 in the supernatants was determined by enzyme-linked immunsorbant assay (ELISA) according to the manufacturer’s instructions (R&D Systems). To assess the extent of transcription of Il2, Jurkat T cells were transiently transfected with 20 µg of a reporter construct that contained the luciferase gene driven by the 2-kB Il2 promoter-enhancer sequence. Forty hours after transfection, separate aliquots of 5 × 105 cells were cultured for 6 hours either alone, with a combination of antibodies against CD3 and CD28, or with PMA and ionomycin and luciferase activity was measured subsequently as described before (47). The efficiency of transfection was normalized by cotransfection with the plasmid pEF-lacZ and assaying for the presence of β-galactosidase.
To measure Ca2+ mobilization by flow cytometry, RIAM-KD and control-KD cells were loaded with 5 µM Indo-1 AM (Molecular Probes) for 45 minutes at 37°C in the dark. Cells were then washed twice with Ca2+-free PBS containing 1% bovine serum albumin (BSA) and were resuspended in RPMI-10 at a density of 1 × 106 cells/ml. After equilibrating the cells at 37°C for 15 minutes, a base line was established for 30 seconds and subsequently cells were stimulated with an antibody against CD3 (1:8000; IgM clone MEM-92, kindly provided by V. Horejsi, Prague, Czech Republic) and analyzed immediately with an LSR II flow cytometer. Assessment of Ca2+ mobilization in RIAM-KD and control-KD primary T lymphocytes was performed in the same way, with the additional step of selecting for green fluorescent protein (GFP)-containing cells, which expressed pLL3.7 carrying either RIAM-specific or control shRNA. To assess Ca2+ mobilization at the single-cell level, RIAM-KD and control-KD cells were loaded with 1 µM fura2 AM (Molecular Probes) in RPMI medium containing 10% FBS for 30 minutes at 37°C. Cells were washed, resuspended in Ringer solution [5 mM hepes (pH 7.4), 155 mM NaCl, 4.5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 10 mM D-glucose] containing 2 mM Ca2+ and attached to poly-L-lysine-coated coverslips to be mounted into a RC-20 closed bath flow chamber (Warner Instruments). Cells were stimulated by perfusing with antibody against CD3 (1:8000; IgM, clone MEM-92) or 1 µM thapsigargin (Calbiochem). Ratiometric measurements of intracellular Ca2+ were performed with a Zeiss Axiovert 200M inverted microscope, OpenLab imaging software (Improvision Inc.), and Excel spreadsheets with tailored macros for numeric data evaluation. Calibration values (Rmin, Rmax, Sf) were obtained from calibration solutions with defined Ca2+ concentrations, as described previously (48).
Preparation of membrane fractions was performed as previously described (4, 19) with minor modifications. Briefly, 100 × 106 cells per sample were washed twice with 1 ml of cold PBS, resuspended in 1 ml of homogenization buffer [20 mM hepes (pH 7.4), 150 mM NaCl, 2 mM MgCl2, 1 mM PMSF, 2 mM Na3VO4, 1 mM NaF, and 10 µl protease inhibitor cocktail/ml], and were gently disrupted at 4°C by 50 strokes in a dounce homogenizer, followed by 10 cycles through a 22-1½ gauge needle. Unbroken cells and nuclei were removed by brief centrifugation at 350 × g, at 4°C and cell homogenates were separated into membranous (particulate, P100) and cytoplasmic (soluble, S100) fractions by ultracentrifugation at 100,000 × g, at 4°C for 1 hour. Proteins in pellets were resuspended in 50 ul homogenization buffer and after quantification, 40 ug of each sample were analyzed by SDS-PAGE and Western blotting with the appropriate antibodies.
IP3 was measured with the Inositol-1,4,5-Trisphosphate [3H] Radioreceptor Assay Kit (PerkinElmer Life Sciences), according to the manufacturer’s instructions, with 10 × 106 cells per sample.
Cytoskeletal fractions were prepared as previously described (49) with minor modifications. Briefly, 0.5 × 106 cells in a final volume of 100 µl of pre-warmed PBS (37°C) were stimulated for 1, 5, or 15 min with antibodies against CD3 and CD28 (each at 10 µg/ml). The reactions were stopped by the addition of cold PBS, followed by two washes in cold PBS. Cells were resuspended in 20 µl of ice-cold cytoskeletal isolation buffer [1% Triton X-100 in 80 mM PIPES (pH 6.8), 5 mM EGTA, 1 mM MgCl2] followed by immediate centrifugation at 13,000 × g for 1 min at 4°C, resulting in a Triton X-100-insoluble pellet. Protein pellets were directly resuspended in sample buffer and analyzed by SDS-PAGE and Western blotting with the appropriate antibodies.
In vitro assays were compared with the unpaired Student’s t test or with analysis of variance (ANOVA) and densitometry data with ANOVA. A p value of <0.05 was considered significant.
Fig. S1. Colocalization of RIAM with ZAP-70 and F-actin at the IS.
Fig. S9. RIAM-KD and control KD cells display comparable levels of LAT phosphorylation.
Fig. S10. Proteins specific for the plasma membrane (Na/K ATPase), Golgi (p58), and the endoplasmic reticulum (calnexin) were not detected in these cytoskeletal fractions
Fig. S2. Reduced abundance of RIAM in RIAM-KD Jurkat T cell lines.
Fig. S3. (A). Knockdown of RIAM results in impaired mobilization of Ca2+ after TCR-mediated activation. (B). Knockdown of RIAM does not impair mobilization of Ca2+ by tapsigargin.
Fig. S4. The interaction between PLC-γ1 and RIAM requires the C-terminal PRR of RIAM.
Fig. S5. Reconstitution of RIAM-KD cells with RIAM-D1 does not restore their ability to activate MEK, ERK, or JNK.
Fig. S6. Negative controls used in detecting the colocalization of phosphorylated PLC-γ1 and F-actin.
Fig. S7. Negative controls used in detecting colocalization of PIP2 with the actin cytoskeleton.
Fig. S8. RIAM and LAT associate with separate pools of PLC-γ1.