Research ArticleImmunology

A large Rab GTPase encoded by CRACR2A is a component of subsynaptic vesicles that transmit T cell activation signals

See allHide authors and affiliations

Sci. Signal.  22 Mar 2016:
Vol. 9, Issue 420, pp. ra31
DOI: 10.1126/scisignal.aac9171

Recruiting vesicles to activate T cells

T cell activation by antigens involves the formation of a complex, highly dynamic, yet organized signaling complex at the site of the T cell receptors (TCRs). Srikanth et al. found that the lymphocyte-specific large guanosine triphosphatase of the Rab family CRACR2A-a associated with vesicles near the Golgi in unstimulated mouse and human CD4+ T cells. Upon TCR activation, these vesicles moved to the immunological synapse (the contact region between a T cell and an antigen-presenting cell). The guanine nucleotide exchange factor Vav1 at the TCR complex recruited CRACR2A-a to the complex. Without CRACR2A-a, T cell activation was compromised because of defective calcium and kinase signaling.

Abstract

More than 60 members of the Rab family of guanosine triphosphatases (GTPases) exist in the human genome. Rab GTPases are small proteins that are primarily involved in the formation, trafficking, and fusion of vesicles. We showed that CRACR2A (Ca2+ release–activated Ca2+ channel regulator 2A) encodes a lymphocyte-specific large Rab GTPase that contains multiple functional domains, including EF-hand motifs, a proline-rich domain (PRD), and a Rab GTPase domain with an unconventional prenylation site. Through experiments involving gene silencing in cells and knockout mice, we demonstrated a role for CRACR2A in the activation of the Ca2+ and c-Jun N-terminal kinase signaling pathways in response to T cell receptor (TCR) stimulation. Vesicles containing this Rab GTPase translocated from near the Golgi to the immunological synapse formed between a T cell and a cognate antigen-presenting cell to activate these signaling pathways. The interaction between the PRD of CRACR2A and the guanidine nucleotide exchange factor Vav1 was required for the accumulation of these vesicles at the immunological synapse. Furthermore, we demonstrated that GTP binding and prenylation of CRACR2A were associated with its localization near the Golgi and its stability. Our findings reveal a previously uncharacterized function of a large Rab GTPase and vesicles near the Golgi in TCR signaling. Other GTPases with similar domain architectures may have similar functions in T cells.

INTRODUCTION

The antigen-dependent activation of T cells requires their direct contact with antigen-presenting cells (APCs). The binding of T cell receptors (TCRs) to cognate peptide-bound major histocompatibility complexes on APCs induces clustering of the TCRs and recruitment of lymphocyte-specific protein tyrosine kinase (Lck) and ζ-chain–associated protein kinase of 70 kD (ZAP70). These kinases phosphorylate the signaling adaptor protein Lat, which forms a signalosome containing phospholipase C-γ1 (PLC-γ1) and the guanine nucleotide exchange factor (GEF) and adaptor molecule Vav1 (13). The activity of PLC-γ1 produces the second messenger inositol 1,4,5-trisphosphate (IP3), which binds to the IP3 receptor on the endoplasmic reticulum (ER) and stimulates the depletion of the ER Ca2+ store. By sensing the depletion of Ca2+ from the ER, stromal interaction molecule 1 (STIM1) translocates to the plasma membrane–proximal regions of the ER and activates Orai1, the pore subunit of the Ca2+ release–activated Ca2+ (CRAC) channels (46). Vav1 accumulates at the immunological synapse and recruits small G proteins (guanine nucleotide–binding proteins), such as Rac1 and CDC42 (cell division control protein 42 homolog), to activate the c-Jun N-terminal kinase (JNK) and p38 mitogen-activated protein kinase (MAPK) pathways (7). Activation of both the Ca2+ and MAPK signaling pathways is essential for the differentiation of naïve T cells into T helper cells, and dysregulation of these pathways results in various immune-related diseases in humans and mice (810).

In addition to its localization at the plasma membrane, Lat also exists in subsynaptic vesicles that translocate to the plasma membrane–proximal regions of the immunological synapse after TCR stimulation (1113). Recruitment of this pool of Lat may be important for its phosphorylation. Lat-containing vesicles use a soluble N-ethylmaleimide–sensitive factor attachment protein receptor (SNARE)–dependent trafficking mechanism for their recruitment. The v-SNARE protein VAMP7 guides these vesicles to the plasma membrane potentially by docking to t-SNARE proteins through a mechanism that does not involve membrane fusion (14). These results suggest that components or functional homologs of the molecular machinery used for the trafficking of synaptic vesicles in the neuronal synapse, such as SNAREs and small Rab guanosine triphosphatases (GTPases), play an important role in the trafficking of subsynaptic vesicles in T cells. However, the importance of these subsynaptic vesicles in TCR signaling has been uncovered only recently, and the identities and functions of these subsynaptic vesicles in T cell activation need further investigation.

More than 60 Rab GTPases exist in humans to regulate vesicular trafficking between organelles. Rab GTPases broadly control vesicle budding, uncoating, motility, and fusion through the recruitment of effector molecules, including sorting adaptors, tethering factors, and motors (15, 16). The functions of Rab GTPases, such as membrane association, are regulated by both GTP binding and prenylation (the attachment of isoprenoid lipids) (17). GTP-bound Rab GTPases are retained at the donor membrane to initiate trafficking to the target organelles, whereas guanosine diphosphate (GDP)–bound forms (which are formed after GTP hydrolysis) detach from the membrane and move to the cytoplasm. GDP-bound Rab GTPases are recycled by the exchange of GDP for GTP. The C-terminal prenylation of Rab GTPases is also essential for their association with donor or target membranes. The types of isoprenoid units that are attached depend on the different GTPase families. Ras GTPases are farnesylated by farnesyl transferase, whereas Rac and Rho GTPases are geranylgeranylated by type I geranylgeranyl transferase (GGT). Rab GTPases are also geranylgeranylated; however, this is performed by type II GGT. Statins are useful tools to investigate protein prenylation because they are inhibitors of 3-hydroxyl-3-methyl-glutaryl–coenzyme A (HMG-CoA) reductase, the key, rate-liming enzyme in the cholesterol synthesis pathway. Statins suppress the generation of farnesyl and geranygeranyl pyrophosphate, which are substrates of prenyl transferases, and thus, these drugs inhibit the prenylation of small G proteins (18). Statins, including atorvastatin, are widely prescribed for their cholesterol-lowering effects. Statins also inhibit TCR signaling, and these drugs are used in clinics to suppress autoimmune diseases (1921). These results emphasize the critical role of the prenylation of small G proteins in intracellular signaling for T cell activation.

Although the role of Rab GTPases in membrane trafficking is well studied, little is known about their direct involvement in intracellular signaling. Moreover, our current understanding of the Rab GTPase family is mostly limited to roles of small proteins of 20 to 25 kD in size. Here, we report a previously uncharacterized function of a distinct large Rab GTPase, CRAC channel regulator 2A (CRACR2A) isoform a (CRACR2A-a), which contains multiple functional domains, including the previously identified N-terminal CRAC channel–regulating domain (22), a proline-rich domain (PRD) for protein-protein interactions, and a C-terminal Rab GTPase domain. Our results showed that various domains of this large Rab GTPase contributed to its translocation from the Golgi-proximal area to the immunological synapse in subsynaptic vesicles, which were distinct from those known to contain LAT, to activate the Ca2+-dependent transcription factor nuclear factor of activation T cells (NFAT) and the JNK signaling pathway. Mechanistically, the PRD of CRACR2A-a was important for the recruitment of these vesicles to the immunological synapse through an interaction with Vav1, whereas the GTPase domain regulated the membrane association of CRACR2A-a by GTP binding and prenylation. These observations provide insights into the role of subsynaptic vesicles in mediating the signaling pathways required for T cell activation.

RESULTS

A GTP-GDP switch regulates the association of the large Rab GTPase CRACR2A-a with the Golgi

The human genome database shows two transcriptional isoforms of CRACR2A: CRACR2A isoform a (NM_001144958.1, CRACR2A-a) and isoform c (NM_032680.3, CRACR2A-c); however, the current mouse genome database annotates only one isoform: CRACR2A-c (NM_001033464.3). In a previous study, we showed that CRACR2A-c contains N-terminal EF hands and two coiled-coil domains that are required for CRAC channel function (22). The other isoform, CRACR2A-a, is of particular interest to us because, on the basis of its predicted amino acid sequence, it contains the Orai1 and STIM1-interacting domain (which are shared by CRACR2A-c) together with other potential functional domains (Fig. 1A). First, we validated the presence of CRACR2A proteins in Jurkat cells, a human CD4+ T cell leukemia cell line, by Western blotting analysis (CRACR2A-a, ~90 kD; CRACR2A-c, ~45 kD) (fig. S1A). Next, to confirm the conserved function of the Orai1 and STIM1-interacting domain in CRACR2A-a, we reconstituted its expression in Jurkat cells depleted of both CRACR2A isoforms. Consistent with previous observations, CRACR2A depletion resulted in a profound reduction in store-operated calcium entry (SOCE) (fig. S1, B and C). Exogenous expression of either CRACR2A-a or CRACR2A-c rescued SOCE in CRACR2A-depleted Jurkat cells, validating their conserved function in CRAC channel activation.

Fig. 1 GTP binding regulates the localization of CRACR2A-a at the Golgi and the activation of JNK by CRACR2A-a after TCR stimulation.

(A) Schematic showing the predicted domain structure of human CRACR2A-a. Both CRACR2A-a and CRACR2A-c share EF-hand motifs, coiled-coil domains (CC1 and CC2), and a leucine-rich region (LR), which interact with the Orai1-STIM1 complex to regulate Ca2+ entry. In addition, CRACR2A-a contains a PRD and a predicted Rab GTPase domain with a putative prenylation site at the C terminus. The fragments and mutants of CRACR2A-a used in this study are indicated. (B) Left: Homology modeling of the GTPase domain of CRACR2A-a (yellow) with Rab3a (red). The GTPase domains of CRACR2A-a and Rab3a share 46% sequence identity and 65% sequence similarity (Clustal Omega). Right: MODELLER (41) was used for homology modeling of the CRACR2A-a GTPase domain to a high-resolution structure of a GPPNHP (5′-guanylyl imidodiphosphate)–bound Rab3a (Protein Data Bank ID: 3RAB). A magnified view of the GPPNHP-binding site is also shown. GPPNHP and the side chains of residues important for GTP binding and hydrolysis (Thr559, Gln604, and Asn658) are shown in stick representation. A loop consisting of residues 561 to 570 was removed for clarity. Panels were generated with PyMOL (version 1.5.0.4; Schrödinger, LLC). (C) Left: Intrinsic GTP hydrolysis activity was measured with the indicated FLAG-tagged CRACR2A-a proteins that were immunoprecipitated from transfected human embryonic kidney (HEK) 293 cells. Data are means ± SEM of triplicate samples from a single experiment and are representative of three experiments. Right: That each sample contained similar amounts of proteins was verified by Western blotting (WB) analysis of the immunoprecipitates with an anti-FLAG antibody. WT, wild type; Pi, inorganic phosphate. (D) Left: Jurkat cells transfected with plasmids encoding green fluorescent protein (GFP)–tagged WT and the indicated mutant CRACR2A-a proteins were analyzed by confocal microscopy to examine their subcellular localization. White circles mark the cell periphery. Images are representative of three experiments. Scale bar, 5 μm. Right: Densitometric analysis (in the indicated numbers of cells) of relative fluorescence intensity of GFP in the Golgi compared to that in the cytoplasm. The Golgi was detected by staining for endogenous Rab8a. Data are representative of three experiments. See fig. S4 for images with Rab8a staining. (E) Jurkat cells stably expressing control (Scr) or CRACR2A-specific shRNA were left unstimulated or were stimulated with anti-CD3 antibodies for indicated times and then analyzed by Western blotting with antibodies against the indicated proteins. Western blots are representative of three independent experiments. Right: Densitometric analysis of relative band intensities. Data are means ± SEM of three independent experiments. See fig. S5A for the detection of PLC-γ1, Lat, ERK, and IκB (inhibitor of nuclear factor κB). (F) Jurkat cells transfected with control (Scr) or CRACR2A-specific shRNA (R2A) and CRACR2A-depleted Jurkat cells that were transfected with plasmids encoding shRNA-resistant WT CRACR2A-a (+a), WT CRACR2A-c (+c), or the indicated mutant CRACR2A-a proteins were stimulated with anti-CD3 antibody for 10 min (left) or kept under resting conditions (right). Cell lysates were then analyzed by intracellular flow cytometry with antibodies specific for p-JNK proteins. Data are means ± SD of the percentages of p-JNK–positive cells from three independent experiments. N.S., not significant. *P < 0.05, **P < 0.005, ***P < 0.0005.

Compared to CRACR2A-c, CRACR2A-a contains an additional PRD and a Rab GTPase domain that contains a predicted prenylation site in its C terminus (Fig. 1A). The amino acid sequence of the GTPase domain of CRACR2A-a is similar to that of the Rab GTPases, and three-dimensional homology modeling of this domain to a high-resolution crystal structure of Rab3a (23) showed almost complete overlap (Fig. 1B). The predicted GTPase domain also closely resembles that of other small G proteins, including Rac1 and CDC42, in the overall core fold (fig. S2) (23). The GTPase domain contains the characteristic P-loop, as well as the switch I and switch II regions of small G proteins. The P-loop and the switch I region make contact with the γ-phosphate of GTP through Mg2+ ions, whereas the switch II region plays a major role in the hydrolysis of GTP through its interaction with phosphates and water molecules (23, 24). To confirm the roles of these conserved residues, we generated a mutant CRACR2A-a that involved substitution of an asparagine (N) residue for a threonine (T) residue within the P-loop, which is predicted to cause preferential GDP binding (T559N) (16, 24). The glutamine residue (Q) in the switch II region, which serves as a catalytic residue for intrinsic GTP hydrolysis, was replaced with a leucine residue (Q604L) (2426). The Q604L mutant CRACR2A-a protein was predicted to be in a constitutively GTP-bound form. The conserved, predicted guanidine-binding G4 motif of CRACR2A-a contained an asparagine residue, which was replaced with an isoleucine to generate the N658I mutant. This mutation was predicted to abolish the binding of both GTP and GDP. Wild-type CRACR2A-a hydrolyzed GTP, which validated the presence of a functional GTPase domain in its C terminus (Fig. 1C). The T559N, Q604L, and N658I mutant CRACR2A-a proteins exhibited reduced GTPase activity, possibly due to defects in GTP binding (T559N and N658I) or GTP hydrolysis (Q604L). These results suggest that the GTPase domain of CRACR2A-a, which is similar to those of small G proteins, is functionally active.

Rab GTPases are predominantly involved in vesicle trafficking and often localize in the Golgi and in secretory vesicles. Therefore, we determined the localization of fluorescently tagged CRACR2A-a by costaining Jurkat cells for various Golgi markers, including GM130, Golgin97, and Rab8a. CRACR2A showed the greatest colocalization with Rab8a, which marks the trans-Golgi network (TGN) and associating vesicles (fig. S3A) (27). Endogenous CRACR2A-a also showed substantial colocalization with Rab8a (fig. S3B). Therefore, we concluded that CRACR2A-a was localized primarily at the Golgi-proximal area, and, more specifically, with the membranes of the TGN and associating vesicles in resting T cells. CRACR2A-a was distinct compared to the small GTPases that we tested because of its predominant localization to the Golgi membrane, whereas others, such as Rac1 and CDC42, showed a broad distribution between the Golgi membrane and the cytoplasm (fig. S3C). Next, we examined how the binding of GDP or GTP influenced the intracellular localization of CRACR2A-a. The GDP-bound T559N mutant and the guanidine binding–defective mutant N658I showed a predominantly cytoplasmic localization, whereas the GTP-bound Q604L mutant localized to the Golgi, similar to wild-type CRACR2A-a (Fig. 1D and fig. S4). Together, these data suggest that CRACR2A-a localizes to the Golgi membrane in resting T cells and that GTP binding regulates its localization.

CRACR2A-a plays a role in the activation of the JNK signaling pathway in response to TCR stimulation

To determine whether CRACR2A-a was important for intracellular signaling pathways other than Ca2+ signaling, we examined the extent of phosphorylation of various signaling molecules in control and CRACR2A-deficient T cells upon TCR stimulation. Whereas CRACR2A depletion did not substantially affect the phosphorylation of TCR-proximal signaling molecules, such as PLC-γ1 and Lat, or of downstream signaling molecules, including extracellular signal–regulated kinase (ERK) or p38, we observed a marked reduction in the extent of JNK1/2 phosphorylation in these cells (Fig. 1E and fig. S5A). These results suggest that CRACR2A-a plays a role in selective signaling pathways. To determine whether this function was shared by other CRACR2A isoforms, we examined the effect of reconstitution of cells with different isoforms of CRACR2A on JNK phosphorylation. Only reconstitution with CRACR2A-a rescued the defect in JNK phosphorylation in the CRACR2A-deficient cells (Fig. 1F), suggesting that CRACR2A-a is distinct among CRACR2A proteins in modulating the JNK pathway.

To examine how GTP binding influenced JNK activation by CRACR2A-a, we transfected CRACR2A-deficient Jurkat cells with plasmids encoding CRACR2A-a mutants that were resistant to short inhibitory RNA. In response to TCR stimulation, cells expressing the GTP-bound Q604L mutant showed enhanced JNK phosphorylation compared to that in cells expressing the wild-type protein, whereas cells expressing the GDP-bound T559N mutant did not show any rescue in JNK phosphorylation (Fig. 1F). These results suggest that GTP binding and Golgi localization are required for the activation of JNK by CRACR2A-a. Under resting conditions, only the constitutively active GTP-bound Q604L mutant CRACR2A protein substantially enhanced JNK phosphorylation compared to that in cells expressing the wild-type protein (Fig. 1F).

The JNK pathway plays a broad role in T cell activation, including in the production of cytokines [for example, interleukin-2 (IL-2)] and the expression of T cell activation markers (for example, CD69) through the activation of the AP1 transcription factors (28). To examine the long-term outcomes of JNK signaling, we measured the cell surface abundance of CD69 in CRACR2A-depleted Jurkat cells that expressed short hairpin RNA (shRNA)–resistant wild-type or mutant CRACR2A-a proteins. CRACR2A-depleted Jurkat cells had less cell surface CD69 than did control cells, which was rescued by wild-type CRACR2A-a, but not CRACR2A-c or the T559N mutant CRACR2A-a (fig. S5B), similar to data obtained from experiments examining JNK phosphorylation. The GTP-bound Q604L mutant that mediated enhanced JNK phosphorylation after short-term stimulation of the Jurkat cells through the TCR did not rescue the cell surface expression of CD69 after long-term stimulation; instead, it showed a dominant-negative effect (fig. S5B). This effect may have been caused by the induction of negative feedback on the AP1 signaling pathway because of excessive JNK activation.

In addition to the mutations that affected GTP binding, we also checked the effect on the localization and function of CRACR2A-a of a mutation within the Ca2+-binding EF-hand domain (97DAD99>AAA), which abolishes Ca2+ binding by the CRACR2A-c isoform and enhances its interaction with the Orai1-STIM1 complex (22). However, the extent of localization of the EF-hand mutant of CRACR2A-a to the Golgi-proximal area was similar to that of the wild-type protein, and it rescued JNK phosphorylation when expressed in CRACR2A-depleted cells (fig. S5, C and D). Therefore, the Ca2+-sensing function of CRACR2A-a did not seem to play a role in its subcellular localization or its ability to activate JNK under the conditions tested. Together, these results not only indicate the presence of CRACR2A-a in addition to CRACR2A-c in a human T cell line but also suggest that CRACR2A-a is involved in regulating both the Ca2+ and the JNK signaling pathways, whereas the role of CRACR2A-c appears to be limited to facilitating Ca2+ signaling. Furthermore, these results suggest that the CRACR2A-a–mediated phosphorylation of JNK required CRACR2A-a to be localized to the Golgi membranes, which was dependent on GTP binding.

Loss of CRACR2A impairs both the Ca2+-NFAT and JNK signaling pathways in primary mouse T cells

The current mouse genomic database annotates only one CRACR2A isoform: CRACR2A-c (NM_001033464.3); however, Western blotting analysis of murine tissues indicated the presence of CRACR2A-a, which was particularly abundant in the spleen and lymph nodes (Fig. 2A). To understand the structure of the gene encoding mouse CRACR2A-a, we cloned its complementary DNA (cDNA) with primers homologous to the human cDNA. Amino acid sequence alignments between the human and mouse CRACR2A-a proteins (deduced from the cloned cDNA sequence) showed 80.1% sequence identity and 91.4% sequence similarity (fig. S6A). On the basis of the cDNA sequence information, we generated an intron-exon map encoding murine CRACR2A-a (fig. S6B).

Fig. 2 CRACR2A deficiency impairs activation of the Ca2+-NFAT and JNK signaling pathways in primary T cells.

(A) Lysates of the brain, heart, lung, lymph node (LN), skeletal muscle (Sk M), and spleen (Spl) of C57BL/6J mice were analyzed by Western blotting with antibody (Ab) against CRACR2A-a (R2A-a); β-actin was used as a loading control. Western blots are representative of three experiments. (B) Left: Schematic of the targeting strategy used to delete exons 3 and 4 of mouse CRACR2A. See fig. S6C for the detailed targeting strategy. Right: CD4+ T cells from control (WT, CRACR2Afl/fl; Cre) and knockout (KO, CRACR2Afl/fl; CD4Cre) mice were cultured under nonpolarizing conditions for 4 days and then were analyzed by Western blotting with antibodies against the indicated proteins. (C) WT and CRACR2A-deficient (KO) naïve T cells were subjected to anti-CD3–mediated cross-linking (aCD3 xlink) of the TCR, and intracellular Ca2+ mobilization was then observed. Left: Averaged (± SEM) SOCE responses from control (n = 81) and CRACR2A-deficient (n = 66) CD4+CD25 naïve T cells after stimulation with anti-CD3 antibody and, subsequently, ionomycin (Iono) in the presence of external solution containing 2 mM Ca2+. Right: Averaged (± SEM) SOCE responses from control (n = 73) and CRACR2A-deficient (n = 76) CD4+CD25 naïve T cells after passive store depletion with thapsigargin in the presence of external solution containing 0.5 or 2.0 mM Ca2+ as indicated. Bar graphs show means ± SEM from three independent experiments. (D) NFATc1 and NFATc2 expression in WT and CRACR2A-deficient naïve mouse CD4+ T cells that were left unstimulated or were stimulated with anti-CD3 and anti-CD28 antibodies for the indicated times and then were analyzed by Western blotting with antibodies against NFATc1 and NFATc3. β-actin was used as a loading control. Right: Densitometric analysis of relative band intensities. Data are means ± SEM of three independent experiments. (E) Left: WT and CRACR2A-deficient T cells were stimulated with anti-CD3 antibody for the indicated times and then were analyzed by flow cytometry to detect p-JNK protein. Right: The line graph shows average ± SD of the percentage of T cells that were positive for p-JNK from three independent experiments. *P < 0.05, **P < 0.005, ***P < 0.0005.

We thus designed a targeting vector to delete exons 3 and 4 to abolish the expression of both CRACR2A isoforms, considering their redundant role in Ca2+-NFAT signaling (Fig. 2B and fig. S6C). CRACR2Afl/fl; CD4Cre mice were generated to ablate CRACR2A expression in the T cell lineage, which was validated by Western blotting analysis (Fig. 2B). Consistent with a role for CRACR2A proteins in Ca2+ signaling in Jurkat cells, CRACR2A-deficient naïve CD4+ T cells showed a substantial decrease in the extent of Ca2+ entry after TCR stimulation and in passive store depletion with thapsigargin (Fig. 2C). Expression of the gene encoding NFATc1, but not that encoding NFATc2, is induced upon TCR stimulation in a Ca2+-NFAT pathway–dependent manner (29). Accordingly, we observed the reduced induction of NFATc1 protein production in CRACR2A-deficient naïve T cells upon TCR stimulation, whereas the abundance of NFATc2 remained unchanged, which was suggestive of a role for CRACR2A in long-term regulation of the Ca2+-NFAT pathway (Fig. 2D). Furthermore, we observed a substantial reduction in the extent of JNK phosphorylation in CRACR2A-deficient naïve CD4+ T cells upon TCR stimulation (Fig. 2E). Impaired activation of the Ca2+-NFAT and JNK signaling pathways led to reduced IL-2 production by CRACR2A-deficient cells (fig. S7, A and B), which is consistent with results from the experiments involving CRACR2A-depleted Jurkat cells. These observations from the analyses of CRACR2A-deficient primary CD4+ T cells support our findings from earlier experiments with CRACR2A-depleted Jurkat cells that demonstrated roles for CRACR2A proteins in the Ca2+-NFAT and JNK signaling pathways.

CRACR2A-a translocates from the Golgi-proximal area to the immunological synapse through subsynaptic vesicles that are distinct from those containing Lat

Next, we examined how the Golgi-resident CRACR2A-a plays an active role in TCR signaling. To examine whether CRACR2A-a underwent translocation from the Golgi to the immunological synapse where signaling molecules cluster to activate the JNK pathway, we performed confocal microscopic analyses of T cells stimulated on anti-CD3 antibody–coated coverslips. Upon TCR stimulation with anti-CD3 antibody, endogenous CRACR2A-a was translocated in vesicles from the Golgi-proximal area to the central area of the cell-coverslip contact site (Fig. 3A and movie S1). Furthermore, incubation of Jurkat cells with Raji cells (a human B cell line) loaded with superantigen also resulted in the translocation of CRACR2A-a–containing vesicles to the immunological synapse over a short time period, and the entire Golgi was relocated toward the immunological synapse at longer time points (Fig. 3B). In addition, we identified a role for the PRD of CRACR2A-a in the translocation of these vesicles. We observed numerous CRACR2A-a–containing vesicles in Jurkat cells expressing the ΔPRD mutant CRACR2A; however, these vesicles failed to accumulate at the immunological synapse (Fig. 3B). These data suggest that the PRD region may be important for protein-protein interactions that are essential for the recruitment of CRACR2A-a–containing vesicles to the immunological synapse. The defect in the translocation of the ΔPRD mutant also affected downstream signaling. Compared to CRACR2A-depleted cells expressing wild-type CRACR2A-a, those expressing the ΔPRD mutant did not exhibit substantial rescue of JNK phosphorylation or CD69 expression at the cell surface (Fig. 3C).

Fig. 3 CRACR2A-a translocates from the Golgi to the immunological synapse through subsynaptic vesicles that are distinct from those containing Lat.

(A) Representative confocal images showing localization of endogenous CRACR2A-a in a Jurkat cell allowed to spread on stimulatory anti-CD3 antibody–coated coverslip for 10 min. The Golgi was labeled with WGA-594. Individual confocal sections at indicated depths from the bottom of the cell are shown. White boxes (inset) highlight the regions magnified in the images on the right. Scale bar, 2 μm. Images are representative of 17 cells from three experiments. See movie S1 for three-dimensional (3D) projection images. (B) Analysis of the localization of Vav1-GFP with mCherry-fused WT and ΔPRD mutant of CRACR2A-a at the immunological synapse formed between Jurkat cells and staphylococcal enterotoxin E (SEE)–pulsed Raji cells. Left-most images show the overlap of bright-field and GFP images. Asterisks in the images show the positions of the Raji cells. The top image is of a Jurkat cell incubated with a Raji cell for 10 min, whereas the middle and bottom images are of Jurkat cells incubated with Raji cells for 20 min. White boxes (inset, 1 to 3) highlight the immunological synapse, which is magnified in the numbered images on the right. Scale bar, 5 μm (unless otherwise indicated). Images are representative of 18 cells (for WT) and 12 cells (for the ΔPRD mutant) from three experiments. (C) Jurkat cells transfected with control (Scr) or CRACR2A-specific shRNA (R2A) and CRACR2A-depleted Jurkat cells that were transfected with plasmids encoding shRNA-resistant WT CRACR2A-a (+WT) or the ΔPRD mutant (ΔPRD) CRACR2A-a protein were stimulated with anti-CD3 antibody for 10 min and then were examined by intracellular staining and flow cytometry to detect p-JNK1/2 (left) or were stimulated with anti-CD3 and anti-CD28 antibodies for 18 hours and then were examined by flow cytometry to examine the cell surface expression of CD69 (right). Data are means ± SD from three independent experiments. **P < 0.005. (D) Transfected Jurkat cells expressing Lat-GFP (green) and mCherry (mCh)–CRACR2A-a (red) were allowed to spread on anti-CD3 antibody–coated coverslips, and their localization at the contact site and 1.4 μm away from the contact site were determined by confocal microscopy. Data in images are average Pearson correlation coefficient values ± SEM of 13 cells from three experiments. Scale bar, 5 μm (unless otherwise indicated). (E) Jurkat cells expressing Lat-GFP (green, top) and mCherry-fused CRACR2A-a (red, middle) were imaged by live-cell TIRF microscopy at the contact site between the cells and anti-CD3–coated coverslips. Bottom: Data in merged images show the average Pearson correlation coefficient values ± SEM of 11 cells from four experiments. Scale bar, 5 μm. See movie S2 for time-lapse images of the same cell.

It is generally thought that plasma membrane–resident TCR signaling molecules cluster at the immunological synapse to activate downstream signaling pathways; however, a study found that subsynaptic vesicles also play a role in the recruitment of signaling molecules to the immunological synapse. The most well-studied vesicles are those bearing Lat, translocation of which is regulated by the vesicle docking protein VAMP7 (14). Therefore, we investigated whether the CRACR2A-a–containing vesicles also contained Lat as cargo. However, we could not detect any substantial colocalization between Lat and CRACR2A-a–containing vesicles in Jurkat cells stimulated on coverslips coated with anti-CD3 antibody (Fig. 3D). We also analyzed the accumulation of Lat- and CRACR2A-a–containing vesicles at plasma membrane–proximal regions by total internal reflection fluorescence (TIRF) microscopy. TIRF microscopic imaging of cells expressing Lat-GFP showed Lat molecules prelocalized at the plasma membrane and translocated Lat molecules contained within subsynaptic vesicles (Fig. 3E); however, we did not observe any substantial colocalization between Lat-containing vesicles and CRACR2A-a–containing vesicles (Fig. 3E and movie S2). These results suggest that the CRACR2A-a–containing vesicles are distinct from the subsynaptic vesicles that bear Lat.

The recruitment of CRACR2A-a–containing subsynaptic vesicles to the immunological synapse depends on an interaction with Vav1

To elucidate the mechanism by which the PRD was involved in the recruitment of CRACR2A-a–containing vesicles to the immunological synapse, we investigated possible protein-protein interactions between CRACR2A-a and the Src homology 3 (SH3) domain–containing proximal TCR signaling molecules that have a high affinity for proline-rich sequences. These efforts led to the identification of Vav1 as an interacting partner of the PRD region of CRACR2A (Fig. 4A). Among the Ca2+-binding, coiled-coil, proline-rich, and GTPase domains, the PRD showed the highest binding affinity for Vav1 in pull-down assays with glutathione S-transferase (GST)–fused proteins. Conversely, we also determined the domain within Vav1 that interacted with CRACR2A-a. GST pull-down experiments with lysates of cells expressing either wild-type CRACR2A-a or its ΔPRD mutant upon incubation with a GST-fused Vav1-C fragment (which contains the SH3-SH2-SH3 domains in the C terminus of Vav1) showed that Vav1-C interacted with CRACR2A-a, an interaction that was substantially reduced by truncation of the PRD (Fig. 4B). These results suggest that the PRD of CRACR2A-a binds to Vav1. These results were further validated by immunoprecipitation experiments with endogenous proteins, which showed that the interaction between CRACR2A-a and Vav1 in Jurkat cells was enhanced after TCR stimulation (Fig. 4C).

Fig. 4 The Vav1–CRACR2A-a interaction is required for the accumulation of CRACR2A-a–containing vesicles at the immunological synapse.

(A) Top: Purified recombinant GST fusion proteins containing the indicated functional domains of CRACR2A-a were incubated with the lysates of anti-CD3–stimulated Jurkat cells and then analyzed by Western blotting for endogenous Vav1. Bottom: Ponceau S staining showed similar amounts of recombinant proteins in each lane. Data are representative of four experiments. CBD, Ca2+-binding domain; CCD, coiled-coil domain. (B) Left: Purified recombinant GST–Vav1-C (containing the C terminus of Vav1) was incubated with lysates of Jurkat cells expressing GFP-fused WT or ΔPRD mutant CRACR2A-a and then analyzed by Western blotting. Right: Ponceau S staining showed that similar amounts of recombinant GST fusion proteins were present in each lane. Data are representative of three experiments. (C) Left: Lysates of resting or anti-CD3–stimulated Jurkat cells were subjected to immunoprecipitation (IP) with anti-CRACR2A antibody and were analyzed by Western blotting to detect the indicated proteins. Right: Densitometric analyses of Vav1 band intensity normalized to that of CRACR2A-a. Data are means ± SEM of three independent experiments. IgG, immunoglobulin G. (D) Left: CRACR2A-depleted Jurkat cells coexpressing Vav1-GFP and mCherry-fused WT (top) or ΔPRD mutant (bottom) CRACR2A-a proteins were dropped onto anti-CD3–coated coverslips and subjected to live-cell TIRF microscopy to image the colocalization of Vav1 (green) with the CRACR2A-a proteins (red). Pearson correlation coefficient values (Rr) are shown in individual images. Scale bar, 5 μm. Right: Scatter plot shows mean ± SEM Pearson correlation coefficient values at the indicated times. Cell numbers are in parentheses. The data were fitted with a double-exponential waveform model, and the corresponding 95% confidence interval of the fit (shaded area) was used to analyze the association and dissociation kinetics. See movies S3 and S4 for time-lapse images of Jurkat cells coexpressing Vav1-GFP and mCherry–CRACR2A-a (movie S3) or mCherry-ΔPRD (movie S4). (E) WT Jurkat cells (Control, left) and J.Vav1 cells (right) expressing mCh–CRACR2A-a (red) and ZAP70-GFP (green) were dropped onto anti-CD3 antibody–coated coverslips and analyzed by live-cell TIRF microscopy to visualize vesicular translocation. Scale bar, 5 μm. Images are representative of three experiments. (F) Quantification of the translocation of CRACR2A-a–containing vesicles in the Jurkat (Control) and J.Vav1 cells shown in (E). Left: Line graph shows the mean normalized fluorescence intensity ± SEM of mCherry–CRACR2A-a in Jurkat control cells and J.Vav1 cells. Right: Mean numbers of mCherry–CRACR2A-a–containing vesicles per cell at the contact site at the indicated times. Cell numbers are in parentheses. *P < 0.05, **P < 0.005, ***P < 0.0005.

Many TCR signaling molecules transiently associate and dissociate at the immunological synapse (30). To determine the association and dissociation kinetics of CRACR2A-a and Vav1, and to elucidate the role of Vav1 in the recruitment of CRACR2A-a–containing vesicles to the immunological synapse, we performed real-time TIRF microscopy experiments. We dropped Jurkat cells expressing Vav1-GFP and mCherry-fused wild-type or ΔPRD mutant CRACR2A-a on coverslips coated with anti-CD3 antibody and monitored the recruitment of vesicles to the contact site. Association and dissociation phenomena were analyzed with a double-exponential waveform model to extract the time constants (Fig. 4D). Vesicles containing CRACR2A-a associated with Vav1 at the central region of the contact site at very early times (t1/2 = 33 s; Fig. 4D and movie S3). At later times, Vav1 accumulated in the periphery of the cell, whereas the CRACR2A-a–containing vesicles were localized near the central region of the contact site (Fig. 4D). The association of vesicles containing the ΔPRD mutant with Vav1 had similar kinetics to those of vesicles containing the wild-type protein (t1/2 = 37 s), but the extent of the association was diminished (as indicated by the reduced value of the Pearson correlation coefficient), and their dissociation from Vav1 occurred with faster kinetics compared to those of vesicles containing wild-type CRACR2A (t1/2 = 144 and 312 s for ΔPRD and wild-type CRACR2A-a, respectively; Fig. 4D and movie S4).

In addition to investigating the ΔPRD mutant, we also examined whether the accumulation of the Ca2+ binding–defective 97DAD99>AAA mutant CRACR2A protein in vesicles was affected. Upon stimulation on coverslips coated with anti-CD3 antibody, Jurkat cells expressing the 97DAD mutant CRACR2A-a protein accumulated at the contact sites similar to those expressing the wild-type protein (fig. S8A). Together with previous results from experiments in which JNK phosphorylation was rescued (fig. S5D), these data suggest that a defect in Ca2+ sensing by CRACR2A-a does not alter its translocation in vesicles and its ability to activate JNK under these conditions.

Next, we examined the effect of Vav1 on the accumulation of CRACR2A-a–containing vesicles in cells on anti-CD3 antibody–coated coverslips. Here, we used ZAP70 as a marker for the contact sites. In Vav1-deficient Jurkat cells (known as J.Vav1 cells), the accumulation of CRACR2A-a–containing vesicles at the contact sites was markedly reduced compared to that in control cells (Fig. 4E). Furthermore, the loss of Vav1 substantially decreased the rate of recruitment and the number of CRACR2A-a–containing vesicles at the antibody-coverslip contact sites (Fig. 4F). Collectively, these data from experiments with the ΔPRD mutant and Vav1-deficient cells suggested an essential role for Vav1 in the recruitment of CRACR2A-a–containing subsynaptic vesicles to the immunological synapse. Vav1 acts as a GEF for Rac1 at the immunological synapse; however, through its C-terminal SH3 and SH2 domains, Vav1 also assembles signaling complexes as a scaffold protein independently of its GEF activity (31). It is unlikely that Vav1 acts as a GEF for CRACR2A-a, because CRACR2A-a predominantly exists in a GTP-bound form at the Golgi. Therefore, the function of Vav1 as a signaling adaptor or scaffold protein may be important for the recruitment of CRACR2A-a–containing vesicles from the Golgi-proximal area to the immunological synapse.

In T cells, the CRAC channel components Orai1 and STIM1 cocluster at the immunological synapse upon contact with APCs (32, 33). Because CRACR2A-a also translocated to the immunological synapse and restored SOCE in CRACR2A-depleted Jurkat cells (fig. S1, B and C), we examined its possible colocalization with CRAC channel components. TIRF imaging of Jurkat cells expressing STIM1-YFP (yellow fluorescent protein) and mCherry–CRACR2A-a that were dropped on anti-CD3 antibody–coated coverslips showed substantial colocalization of both proteins (fig. S8B). Furthermore, passive depletion of ER Ca2+ stores by thapsigargin also stimulated the translocation of CRACR2A-a–containing vesicles from the Golgi-proximal area to being in close proximity with STIM1 clusters, albeit to a lesser degree than was caused by TCR stimulation (fig. S8C). These experiments suggest that both passive and active store depletion stimulates the colocalization of CRACR2A-a–containing vesicles with STIM1 to mediate CRAC channel activation.

Dissociation of CRACR2A-a from membranes by deprenylation or GTP hydrolysis stimulates its degradation

Small Rab GTPases associate with the membranes of the Golgi and vesicles depending on whether GTP is bound and on C-terminal prenylation (17). A Prenylation Prediction Suite (PrePS) predicted the attachment of a 20-carbon geranylgeranyl moiety to the C-terminal di-Cys motif of CRACR2A-a with a high confidence score (E = 8.9 × 10−63; Fig. 5A). However, the C-terminal CCG residues of CRACR2A-a do not constitute a conventional prenylation motif (CxC or CC, where x is any amino acid) for Rab GTPases. To determine whether this CCG motif was indeed geranylgeranylated and thus important for the retention of CRACR2A-a in the Golgi, we truncated these three amino acids to generate the ΔC729 mutant CRACR2A-a protein. The deletion of these three residues resulted in the predominantly cytoplasmic localization of the mutant protein, which suggests the possible prenylation-induced association of CRACR2A-a with the plasma membrane (Fig. 5A). Furthermore, the ΔC729 mutant CRACR2A-a protein could not rescue JNK phosphorylation in CRACR2A-depleted Jurkat cells, which suggests a role for prenylation in this function.

Fig. 5 Degradation of CRACR2A-a by atorvastatin, GTP hydrolysis, and TCR stimulation.

(A) Top: Sequence alignment of the prenylation sites in the C termini of CDC42, Rac1, and CRACR2A-a. CRACR2A-a contains di-Cys residues that are potential targets of the type II GGT and positively charged residues (++) that can bind to phospholipids. α5 corresponds to the fifth α-helix in the GTPase domain, and boxes represent positively charged amino acids (++) and prenylation sequences. Middle: Representative confocal images of Jurkat cells expressing GFP-fused WT and mutant CRACR2A-a proteins costained for endogenous Rab8a. White circles mark the cell periphery. Scale bar, 5 μm. Bottom: (Left) Mean relative fluorescence intensity ± SD of GFP in the Golgi compared to that in the cytoplasm in the indicated numbers of cells. (Right) Mean ± SD of the percentages of p-JNK–positive cells among Jurkat cells transfected with control (Scr) or CRACR2A-specific shRNA (R2A) and CRACR2A-depleted Jurkat cells that were transfected with plasmids encoding shRNA-resistant WT CRACR2A-a (+a) or the indicated mutant CRACR2A-a proteins after stimulation with anti-CD3 antibody for 10 min. Images are representative of three experiments. (B) Top: Confocal microscopy images of Jurkat cells expressing GFP-tagged Rac1 and CRACR2A-a that were treated with the indicated concentrations of atorvastatin (Ator) for 16 hours. The atorvastatin target Rac1 was used as a positive control. Scale bar, 5 μm. Bottom: Relative fluorescence intensities of GFP-Rac1 (left) and GFP–CRACR2A-a (right) in the Golgi (identified by WGA-594 staining; see fig. S9 for images with WGA-594 costaining) compared to those in the surrounding areas. Data are means ± SD of three independent experiments. (C) Top: Jurkat cells were treated with the indicated concentrations of atorvastatin for 16 hours and then analyzed by Western blotting with antibodies against the indicated proteins. Bottom: Densitometric analysis of the relative band intensities in treated cells normalized to those in untreated cells. Data are means ± SEM of three independent experiments. (D) Left: HEK 293T cells expressing N-terminally FLAG-tagged WT or the indicated mutant CRACR2A-a cDNAs and GFP from an IRES site were lysed and analyzed by Western blotting to detect the indicated proteins. The Q604L lane was cut out from the same blot (splice point indicated by vertical black line). Right: Densitometric analysis of the relative amounts of the indicated mutant proteins normalized to that of WT CRACR2A-a. Data are means ± SEM of three independent experiments. The y axis contains a break for clear presentation of data. (E) Top: Jurkat cells (left) and naïve mouse CD4+ T cells (right) were stimulated with anti-CD3 antibody for the indicated times, and cellular lysates were then analyzed by Western blotting with antibodies against the indicated proteins. Bottom: Densitometric analysis of the relative amounts of CRACR2A-a normalized to β-actin in Jurkat cells (left) and naïve mouse CD4+ T cells (right). Data are means ± SEM of three independent experiments. **P < 0.005, ***P < 0.0005.

We next examined how treatment with atorvastatin, an inhibitor of prenyl transferases, affected the function of CRACR2A-a and compared it with a known target of statins. Treatment of Jurkat cells with atorvastatin substantially decreased the localization of Rac1, a positive control for atorvastatin treatment, to the Golgi membrane (Fig. 5B and fig. S9A). Under the same conditions, we observed a marked decrease in the extent of the Golgi membrane association of CRACR2A-a. We also noticed a marked decrease in the fluorescence intensity of GFP-tagged CRACR2A-a in atorvastatin-treated cells (Fig. 5B). Consistent with this observation, the abundance of CRACR2A-a protein was decreased in atorvastatin-treated cells, whereas that of Rac1 was not substantially changed (Fig. 5C). These results suggest that the lack of prenylation may induce the degradation of CRACR2A-a.

Because deprenylation by atorvastatin reduced the extent of the association of CRACR2A-a with the Golgi membrane and induced its degradation, we tested the hypothesis that a switch in the binding of GTP for GDP, which is also important for Golgi membrane association, might stimulate the degradation of CRACR2A-a. We transfected HEK 293T cells with plasmids encoding FLAG-tagged wild-type CRACR2A-a, the GTP binding–defective T559N mutant, the guanidine binding–defective N658I mutant, or the GTP hydrolysis–defective Q604L mutant of CRACR2A-a from a single mRNA with GFP, which were separated by an internal ribosomal entry site (IRES). We hypothesized that measurement of the abundance of wild-type or mutant CRACR2A-a protein, when normalized to that of GFP, would give an estimate of the stabilities of the CRACR2A protein because both they and the GFP proteins would be translated from the same mRNA. The GTP binding–defective T559N and N658I mutants were reduced in abundance, whereas the constitutively GTP-bound Q604L mutant protein was increased in abundance compared to that of wild-type CRACR2A-a (Fig. 5D). These results suggest that both prenylation and GTP binding are required for the association of CRACR2A-a with the Golgi membrane and for its stability. To examine the potential physiological effects of this finding, we analyzed the abundance of CRACR2A-a protein in Jurkat cells and mouse CD4+ T cells after TCR stimulation, which would be expected to induce GTP hydrolysis. CRACR2A-a was decreased in abundance in both Jurkat cells and mouse T cells after TCR stimulation (Fig. 5E). Together, these results suggest that TCR stimulation or atorvastatin can cause the cytoplasmic localization of CRACR2A-a and induce its degradation. Furthermore, these data also suggest that CRACR2A-a is a target of atorvastatin and that its degradation may contribute to the anti-inflammatory effects observed by atorvastatin administration (1921).

DISCUSSION

Human CRACR2A encodes two validated transcriptional isoforms: CRACR2A-a and CRACR2A-c. The short isoform CRACR2A-c facilitates CRAC channel function by enhancing the interaction between Orai1 and STIM1 (22). The current mouse genome database shows the presence of only CRACR2A-c. Our analyses validated the presence of the long isoform CRACR2A-a in murine tissues, particularly in mouse lymphoid organs. Both of these isoforms share a conserved function in the Ca2+-NFAT signaling pathway, but only CRACR2A-a plays a role in the activation of the JNK pathway, which emphasizes the importance of the PRD and GTPase domains in this distinctive function. Furthermore, our studies reveal a previously uncharacterized mechanism for a Rab GTPase in TCR signaling, which includes recruitment in subsynaptic vesicles to the immunological synapse and inactivation by protein degradation. The activation and inactivation of CRACR2A-a consists of resting, effector, and termination steps that are controlled by GTP binding and protein degradation (fig. S9B). In resting T cells, CRACR2A-a needed to be bound to GTP and prenylated to be localized at the Golgi. Unlike most small G proteins, CRACR2A-a predominantly exists in a GTP-bound form localized to the Golgi in resting T cells. After TCR stimulation, CRACR2A-a–containing vesicles were recruited to the immunological synapse through an interaction with Vav1, which led to the activation of the downstream JNK pathway.

Passive store depletion also induced the accumulation of CRACR2A-a–containing vesicles in close proximity to the ER–plasma membrane junctions to activate SOCE. The mechanism by which translocation of these vesicles is activated by TCR stimulation or store depletion needs further investigation. Upon TCR stimulation, CRACR2A-a appeared to undergo some sort of posttranslational modification, as evidenced by the detection of a higher molecular mass band in the lysates of stimulated T cells with the anti-CRACR2A antibody (Figs. 4C and 5E). The identity and physiological relevance of this posttranslational modification of CRACR2A-a remain unclear. Inactivation of CRACR2A-a occurred through GTP hydrolysis, which induced its cytoplasmic localization and degradation (termination step). Currently, the composition and fate of the CRACR2A-a–containing vesicles remain unknown because of the lack of additional marker molecules. Small GTPases bind to guanine nucleotide dissociation inhibitors (GDIs) after GTP hydrolysis occurs, which enables their stabilization in the cytoplasm and eventually their reinsertion into the membrane (16, 34). In contrast to small GTPases, CRACR2A-a was not recycled; rather, it was consumed by entering a unidirectional pathway of protein degradation. Our studies also demonstrated that atorvastatin disrupted CRACR2A-a stability by inducing this unstable, cytoplasmic state through deprenylation. The high sensitivity of CRACR2A-a to atorvastatin may be because of a lack of CRACR2A-a–specific GDI molecules.

Small G proteins, including RhoA, Rac1, and CDC42, are important regulators of signaling pathways that control vital functions in T cells, including thymocyte development, cytoskeletal dynamics, gene transcription, and cell cycle progression (35). Small GTPases are present as GDP-bound inactive precursors and are activated by a switch from GDP to GTP to mediate their downstream effects. CRACR2A-a is distinctive because it exists in a GTP- and Golgi membrane–bound form in resting T cells and is degraded after GTP hydrolysis. CRACR2A-a resembles the atypical Rho GTPases, including RhoH, which resides in a constitutively GTP-bound state (36). Hence, it appears that, similar to RhoH, the activity of CRACR2A-a may be regulated by localization.

The detailed mechanism by which CRACR2A-a activates the downstream JNK pathway is unclear. Rac1 and CDC42 play important roles in the stimulation of both p38 and JNK. Activation of the JNK pathway by Rac1 and CDC42 involves their direct interaction with p21-activated kinase (PAK1) or other MAPK kinase kinases (MAPKKKs) to phosphorylate MKK4 and MKK7 (10). It would be interesting to know whether CRACR2A-a acts as a signaling adaptor to directly activate the downstream JNK pathway. Alternatively, it is also possible that CRACR2A-a is involved in the trafficking of vesicles that bear other signaling molecules involved in the activation of JNK as cargo, supporting the conventional role of small Rab GTPases. The v-SNARE protein VAMP7 plays an important role in the recruitment of Lat-containing subsynaptic vesicles to TCR clusters, and this recruitment appears to be important for the phosphorylation of Lat and the formation of the Lat signalosome (14). Further studies are required to determine whether the same cellular machinery is used for the translocation of CRACR2A-a–containing vesicles. The identity and function of CRACR2A-a–containing vesicles is likely to be an exciting area for future investigation.

More than 60 members of the Rab GTPase family have been identified in humans; however, our knowledge of Rab proteins is limited to the scope of the role of small Rab GTPases in membrane trafficking. Here, we elucidated a previously uncharacterized mechanism of a large Rab GTPase in intracellular signaling that is potentially mediated by vesicular trafficking in T cells. Unlike small G proteins, CRACR2A-a contains additional domains, including N-terminal EF hands, a coiled-coil domain, and a PRD. We observed an interaction between CRACR2A-a and Vav1 through the PRD of CRACR2A-a, which was important for the recruitment of CRACR2A-a–containing vesicles to the immunological synapse. These data suggest that domains of CRACR2A-a other than the GTPase domain are important for its distinct functions. These results establish a foundation to understand the function of other large Rab GTPases, such as Rab44 and Rab45, which have domain architectures similar to that of CRACR2A-a (fig. S10). Numerous gene linkage analyses of acute and chronic myeloid leukemia and melanoma identified Rab45 [also known as RASEF (RAS and EF-hand domain containing) or FLJ31614] as a potential tumor suppressor (37, 38); however, the molecular mechanism underlying its link to human diseases is not clearly understood. Therefore, our study of CRACR2A-a may help in uncovering possible roles of these large Rab GTPases in intracellular signaling and vesicle trafficking.

In summary, our study has identified CRACR2A proteins as an intracellular signaling module that bridges two TCR-proximal signaling pathways, Ca2+-NFAT and JNK, to facilitate T cell activation. Whereas the short isoform CRACR2A-c was only involved in the regulation of Ca2+-NFAT signaling, the long isoform (and large Rab GTPase) CRACR2A-a was important for both the Ca2+-NFAT and the JNK signaling pathways. We found that CRACR2A-a was present in subsynaptic vesicles that translocated from the Golgi region to the immunological synapse after ligation of TCRs, which was required for the activation of downstream signaling. Furthermore, the function of CRACR2A-a was regulated by membrane dissociation–induced degradation upon TCR stimulation. Therefore, this study provides a conceptual framework to advance our understanding of the potential roles of large Rab GTPases in intracellular signaling and their pathological mechanisms related to human diseases.

MATERIALS AND METHODS

Chemicals

Fura-2 AM and wheat germ agglutinin (WGA-594) were purchased from Invitrogen. Thapsigargin, phorbol 12-myristate 13-acetate, and ionomycin were purchased from EMD Millipore. Brefeldin A was purchased from eBioscience. Atorvastatin was purchased from Sigma.

Plasmids and cells

Full-length cDNA encoding human CRACR2A-a (National Center for Biotechnology Information Reference Sequence: NM_001144958.1) was cloned from a Jurkat cell cDNA library (see table S1 for primers) into the pMSCV-CITE-eGFP-PGK-Puro vector with an N-terminal FLAG tag. The reverse primer was designed in the putative 3′ untranslated region to sequence the endogenous STOP codon. Our clone contains an additional serine residue at amino acid position 425 within the PRD (codon 5′-AGT-3′). Various mutants of CRACR2A-a were generated by polymerase chain reaction (PCR) amplification and site-directed mutagenesis (see table S1 for primers). cDNA encoding CRACR2A-a was N-terminally tagged with a FLAG tag and subcloned into the lentiviral vector FGllF [a gift from D. S. An, University of California, Los Angeles (UCLA)]. The cDNAs encoding wild-type CRACR2A-1 and its mutants were also subcloned into the plasmids pC1-mCherry and pEGFP-C1 (Clontech) to generate proteins fused N-terminally with mCherry or GFP. HEK 293 and Jurkat E6-1 cell lines were obtained from the American Type Culture Collection. Raji cells (a human B cell lymphoma cell line) were a gift from S. Morrison (UCLA). Vav1-GFP, ZAP70-GFP, and GFP-Rac1 clones were purchased from Addgene. The LAT1-CFP (cyan fluorescent protein) clone and J.Vav1 cells were gifts from L. Samelson (National Institutes of Health). LAT1 cDNA was excised from the LAT1-CFP clone with Eco RI and Bam H1 and subcloned into the plasmid pEGFP-N1.

Western blotting analysis

To analyze tissues by Western blotting, wild-type mice were euthanized and perfused with phosphate-buffered saline (PBS) before their organs were harvested. Tissues were snap-frozen, lysed in radioimmunoprecipitation assay (RIPA) buffer [10 mM tris-Cl (pH 7.5), 1% Triton X-100, 0.1% SDS, 140 mM NaCl, 1 mM EDTA, 0.1% sodium deoxycholate, and protease inhibitor cocktail (Roche)], and centrifuged to remove debris. Samples were resolved by 8 to 10% SDS–polyacrylamide gel electrophoresis (SDS-PAGE). Proteins were transferred to nitrocellulose membranes and subsequently analyzed by Western blotting with relevant antibodies. To detect NFAT, 5 × 106 Jurkat cells were stimulated with plate-bound anti-CD3 (10 μg/ml; OKT3, National Cancer Institute preclinical repository) and soluble anti-CD28 antibody (10 μg/ml) for the times indicated in the figure legends, harvested in PBS, lysed in RIPA buffer, and centrifuged to remove debris. Primary mouse T cells (5 × 106) were stimulated with plate-bound anti-CD3 (10 μg/ml) and soluble anti-CD28 antibody (10 μg/ml) for the times indicated in the figure legends. Lysates were resolved by 10% SDS-PAGE and analyzed by Western blotting to detect NFATc1, NFATc2, and β-actin. To analyze TCR signaling, 5 × 106 Jurkat cells or primary T cells were stimulated with soluble anti-CD3 antibody (10 μg/ml) and cross-linked with anti-mouse or anti-hamster secondary antibody for the times indicated in the figure legends, and cell pellets were lysed in SDS loading dye. Lysates were resolved by 10% SDS-PAGE and analyzed by Western blotting to detect proteins of interest. For atorvastatin treatment, 5 × 106 Jurkat cells were treated with different concentrations of atorvastatin (in dimethyl sulfoxide). Cell pellets were lysed in RIPA buffer and centrifuged to remove debris. Lysates were resolved by 12% SDS-PAGE and analyzed by Western blotting to detect Rac1, CDC42, CRACR2A, and β-actin. The antibodies used were as follows: anti–β-actin (Santa Cruz Biotechnology, clone I-19), anti-FLAG (Sigma), anti-NFATc1 (BD Pharmingen, #556602), anti-NFATc2 (Santa Cruz Biotechnology, #7296), anti–p-ERK (Cell Signaling Technology, #4377), anti-ERK (Cell Signaling Technology, #9102), anti–p-p38 (Cell Signaling Technology, #4511), anti-p38 (Cell Signaling Technology, #9212), anti–p-JNK (Cell Signaling Technology, #9255), anti-JNK (Cell Signaling Technology, #9252), anti–p-IκB (Cell Signaling Technology, #9246), anti–PLC-γ1 (Cell Signaling Technology, #2822), anti–p-PLC-γ1 (Cell Signaling Technology, #2821), anti-Rac1 (Cell Signaling Technology, #2465), anti–p-Lat (Cell Signaling Technology, #3581), anti-Lat (Cell Signaling Technology, #9166), and anti-Vav1 (Cell Signaling Technology, #2502). Polyclonal rabbit antibody to detect CRACR2A was generated with purified human CRACR2A-c protein (Open Biosystems) and used at a concentration of 0.1 μg/ml. Chemiluminescence images were acquired with an Image Reader LAS-3000 liquid-crystal display camera (FujiFilm).

shRNA-mediated knockdown

pLKO.1 plasmids encoding shRNAs for the depletion of CRACR2A were purchased from Open Biosystems (Thermo Fisher). The sequences of the shRNAs are provided in table S1. To generate lentiviruses for transduction, HEK 293T cells were transfected with plasmid(s) encoding the appropriate shRNA and packaging vectors (pMD2.G and psPAX2, purchased from Addgene) with the calcium phosphate transfection method. Culture medium was harvested at 48 and 72 hours after transfection and used to infect Jurkat cells together with Polybrene (8 μg/ml) through the spinfection method. Cells were selected with puromycin (1 μg/ml) 48 hours after infection.

Single-cell Ca2+ imaging, TIRF microscopy, and confocal microscopy

Jurkat cells or primary T cells (1 × 106 cells/ml) were loaded with 1 μM Fura-2 AM for 30 min at 25°C and then attached to poly-l-lysine–coated coverslips. Measurement of the intracellular Ca2+ concentration ([Ca2+]i) was performed essentially as previously described (39). For TIRF analysis, coverslip bottom dishes were coated with anti-CD3 antibody (10 μg/ml, OKT3) at 37°C for 1 hour, washed with PBS, and used for experiments. Cells were resuspended in Ringer’s solution containing 155 mM NaCl, 4.5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 10 mM d-glucose, and 5 mM Na-Hepes (pH 7.4), dropped onto anti-CD3–coated coverslips, and either used for time course imaging or fixed after stimulation for 10 min with 2.5% paraformaldehyde (PFA) at room temperature and used for confocal microscopy analysis. TIRF microscopy was performed with an Olympus IX2 illumination system mounted on an Olympus IX51 inverted microscope as described previously (22). Acquisition and image analysis, including measurement of Pearson correlation coefficient, were performed with SlideBook software (Intelligent Imaging Innovations Inc.), and graphs were plotted with OriginPro8.5 (OriginLab). To quantify TIRF intensity across different cells, individual regions of interest were selected, and data were analyzed as the ratio of fluorescence intensity at each time point (F) to that at the start of the experiment (F0). Vesicle number quantification was performed by automated counting options in ImageJ. For confocal analysis, eGFP (enhanced GFP) and mCherry were excited sequentially on a Zeiss LSM inverted microscope (Axiovert 100 LSM, Carl Zeiss) fitted with a 63× water immersion objective lens (C-Apochromat 63×/1.2 W Corr, Carl Zeiss). Images were acquired and processed for enhancement of brightness and contrast with Pascal5 software (Carl Zeiss).

GTPase activity assay

HEK 293T cells were transfected with empty plasmid or plasmids encoding wild-type or mutant CRACR2A-a proteins. Forty-eight hours after transfection, the cells were harvested, lysed in lysis buffer [20 mM tris-Cl (pH 7.5), 2 mM EDTA, 100 mM NaCl, 10% glycerol, 1.0% Igepal CA-630, and protease inhibitor cocktail], and centrifuged to remove the debris. Cell lysates were then precleared with protein A–Sepharose and subjected to immunoprecipitation overnight with anti-FLAG antibody–conjugated agarose (Sigma). Immunoprecipitated samples were washed in lysis buffer three times and then washed twice more in lysis buffer without EDTA and Igepal CA-630. Immunoprecipitates were directly used for the analysis of GTPase activity with a colorimetric GTPase assay kit (Novus Biologicals), and absorbance was read on a POLARstar Omega plate reader (BMG Labtech).

Measurement of JNK phosphorylation by intracellular staining

Exponentially growing Jurkat cells (5 × 106) were transfected with empty vector or plasmids encoding CRACR2A-c and wild-type or mutant CRACR2A-a proteins by electroporation at 200 V with an ECM 830 electroporator (Harvard Apparatus). Forty-eight hours after transfection, the cells were left untreated or were stimulated with anti-CD3 antibody (10 μg/ml, OKT3) for 10 min. Naïve CD4+ T cells isolated from control and CRACR2Afl/fl-CD4Cre animals were stimulated with anti-CD3 antibody (1 μg/ml, 1452C11) for the times indicated in the figure legends. Cells were fixed with 4% PFA, permeabilized with ice-cold methanol, and stained with an anti–p-SAPK/JNK monoclonal antibody conjugated with Alexa Fluor 647 (Cell Signaling Technology, #9257). Cells were washed twice with PBS and analyzed with a FACSCalibur flow cytometer (Becton Dickinson) and FlowJo software.

Immunological synapse and immunofluorescence analysis

For immunological synapse analysis, Jurkat cells were electroporated with the appropriate plasmids and used for experiments 24 hours later. To visualize Jurkat cells interacting with superantigen-pulsed Raji cells, CMAC (7-amino-4-chloromethylcoumarin) blue dye (Invitrogen)–loaded Raji cells were pulsed with staphylococcal enterotoxin E toxin (1 μg/ml; Toxin Technology) for 45 min in complete medium. After rinsing, the Raji cells were mixed with Jurkat cells at a ratio of 1:1 for 10 or 20 min on poly-d-lysine–coated coverslips at 37°C, fixed with 2.5% PFA, and rinsed with PBS. Immune conjugates were examined on an inverted Zeiss LSM (Axiovert 100 LSM) confocal microscope as described earlier. For immunofluorescence staining of endogenous CRACR2A, Jurkat cells fixed with 2.5% PFA were washed three times with PBS, permeabilized with 0.5% Igepal CA-630, blocked with 2.5% fetal bovine serum, and stained with purified anti-CRACR2A antibody for 1 hour at room temperature. The cells were washed with PBS, incubated with fluorescein isothiocyanate–labeled anti-rabbit antibody for 30 min, mounted, and examined on a Zeiss LSM confocal microscope. Anti-Rab8a antibody (clone 3G1) was purchased from Abnova, anti-GM130 antibody (D6B1) was purchased from Cell Signaling Technology, and anti-Golgin antibody (#A21270) was purchased from Invitrogen.

GST pull-down analysis and immunoprecipitations

Jurkat cells (2 × 107) were harvested in PBS, lysed in lysis buffer, and centrifuged to remove debris before being precleared with glutathione–Sepharose 4B beads. Lysates were incubated with 20 μg of GST or GST-tagged fragments of the Ca2+-binding domain (amino acids 1 to 197), coiled-coil domain (amino acids 198 to 348), PRD (amino acids 349 to 540), or GTPase domain (amino acids 541 to 731) of CRACR2A-a for 18 hours in binding buffer [20 mM tris-HCl (pH 7.5), 100 mM NaCl, 2 mM EDTA, 1.0% Igepal CA-630, protease inhibitors, and 10% glycerol]. Pull-down samples were washed five times with lysis buffer and analyzed by Western blotting. For GST pull-down with the Vav1 C terminus, HEK 293T cells (5 × 106) expressing GFP-tagged wild-type CRACR2A-a or the ΔPRD mutant of CRACR2A were lysed as described earlier and incubated with 20 μg of GST or the GST-tagged C-terminal fragment of Vav1 (amino acids 637 to 870) overnight as described earlier. For immunoprecipitations, Jurkat cells (2 × 107) were lysed in lysis buffer, centrifuged to remove debris, and precleared with protein A–Sepharose. The precleared lysates were incubated with 1 μg of rabbit IgG anti-CRACR2A antibody for 18 hours. Immunoprecipitates were washed five times in lysis buffer and analyzed by Western blotting to detect Vav1 and CRACR2A.

Generation of CRACR2A knockout mice

Targeting of CRACR2A was performed by flanking exons 3 and 4 with LoxP sites by homologous recombination in AB2.2 (129SvEv) embryonic stem (ES) cells. Aberrant splicing of exons 2 to 5 causes a shift in the reading frame, which results in truncation of the protein. Deletion of exons 3 and 4 results in loss of expression of both isoforms of CRACR2A. G418-resistant clones were screened by PCR for homologous recombination at both homology arms. Chimeric mice with floxed CRACR2A alleles were generated by blastocyst injection of heterozygous CRACR2Afl/+ ES cell clones. Founder CRACR2Afl/+ chimeric mice were bred with Flp-deleter mice (Jackson Laboratory) to remove the neomycin resistance gene cassette. CRACR2Afl/fl mice were backcrossed to C57BL/6 mice for at least 10 generations and then bred with CD4Cre mice (Jackson Laboratory) to generate T cell–specific deletion of CRACR2A. All mice were maintained in pathogen-free barrier facilities and used in accordance with protocols approved by the Institutional Animal Care and Use Committee at the UCLA.

T cell purification

T cell purification, activation, and transduction were performed as previously described (40). CD4+ T cells were purified by magnetic sorting from single-cell suspensions generated by mechanical disruption of the spleens and lymph nodes of adult mice according to the manufacturer’s instructions (Invitrogen). For effector T cell differentiation, cells were stimulated with anti-CD3 antibody (1 μg/ml; 1452C11, Bio X Cell) and anti-CD28 antibody (1 μg/ml; Pharmingen) for 48 hours on plates coated with goat anti-hamster secondary antibody for cross-linking the anti-CD3 antibody (0.3 mg/ml; MP Biomedicals).

Statistical analysis

Statistical comparisons were performed with a two-tailed Student’s t test. Normal distribution of data was evaluated with the Shapiro-Wilk test, and equal variance was tested with the F test. For all of the tests, P <0.05 was considered to be statistically significant.

SUPPLEMENTARY MATERIALS

www.sciencesignaling.org/cgi/content/full/9/420/ra31/DC1

Fig. S1. CRACR2A-a plays a role in SOCE in Jurkat cells.

Fig. S2. The C terminus of CRACR2A-a contains a Rab GTPase domain.

Fig. S3. Subcellular localization of CRACR2A-a in comparison with Golgi markers and small GTPases.

Fig. S4. Subcellular localization of wild-type and mutant CRACR2A-a proteins.

Fig. S5. CRACR2A-a is required for the TCR-stimulated phosphorylation of JNK and cell surface expression of CD69.

Fig. S6. Genomic organization of human and murine CRACR2A.

Fig. S7. CRACR2A deficiency or depletion reduces cytokine production by T cells.

Fig. S8. CRACR2A-a colocalizes with STIM1 after store depletion.

Fig. S9. Degradation of CRACR2A-a by atorvastatin.

Fig. S10. Members of the large GTPase family are defined by their similar domain architectures.

Table S1. List of primers and shRNAs used in this study.

Movie S1. Rotation of the 3D projection of endogenous CRACR2A in a Jurkat cell on an anti-CD3 antibody–coated coverslip.

Movie S2. TIRF imaging of Lat and CRACR2A-a at the plasma membrane of a spreading Jurkat cell.

Movie S3. TIRF imaging of Vav1 and CRACR2A-a at the plasma membrane of a spreading Jurkat cell.

Movie S4. TIRF imaging of Vav1 and ΔPRD CRACR2A-a at the plasma membrane of a spreading Jurkat cell.

REFERENCES AND NOTES

Acknowledgments: We thank V. Barr and L. Samelson (NIH) for sharing the Lat plasmid, the J.Vav1 cell line, and protocols for the stimulation of Jurkat cells with anti-CD3 antibody and APCs. We also thank M. Oh-hora (Kyushu University, Japan) and B. Ribalet for helpful suggestions. Funding: This work was supported by NIH grant AI083432 (Y. Gwack) and American Heart Association grant 12SDG12040188 (S.S.). Flow cytometry was performed in the UCLA Jonsson Comprehensive Cancer Center (JCCC) and Center for AIDS Research (CFAR) Flow Cytometry Core Facility, which is supported by JCCC grant P30 CA016042 and CFAR grant 5P30 AI028697. Author contributions: S.S. and Y. Gwack designed the research; S.S. cloned and generated CRACR2Afl/fl animals and performed all the Ca2+ imaging, microscopy, and statin treatment experiments; K.-D.K. performed the in vitro T cell analysis and JNK experiments with help from J.S.W.; Y. Gao performed biochemical experiments to analyze CRACR2A and Vav1 interactions and performed GTPase assays; S.G. performed Western blotting analysis of CRACR2A and NFAT; G.C. helped with statistical analyses; A.P. and J.A. built the structural homology model of the GTPase domain of CRACR2A-a; M.J. helped with the generation of the CRACR2Afl/fl mice; and S.S. and Y. Gwack wrote the manuscript with input from the other authors. Competing interests: The authors declare that they have no competing interests.
View Abstract

Navigate This Article