Tripartite Motif-Containing Protein 30 Modulates TCR-Activated Proliferation and Effector Functions in CD4+ T Cells

To avoid excessive activation, immune signals are tightly controlled by diverse inhibitory proteins. TRIM30, a tripartite motif (TRIM)-containing protein is one of such inhibitors known to function in macrophages. To define the roles of TRIM30, we generated Trim30 knockout (Trim30 −/−) mice. Trim30 deletion caused no major developmental defects in any organs, nor showed any discernable defect in the activation of macrophages. But, Trim30 −/− mice showed increased CD4/CD8 ratio when aged and Trim30 −/− CD4+ T cells exhibited an abnormal response upon TCR activation, in particular in the absence of a costimulatory signal. Adoptive transfer of wild-type and Trim30 −/− CD4+ T cells together into lymphopenic hosts confirmed higher proliferation of the Trim30 −/− CD4+ T cells in vivo. Despite the enhanced proliferation, Trim30 −/− T cells showed decreased levels of NF-κB activation and IL-2 production compared to wild-type cells. These results indicate a distinct requirement for TRIM30 in modulation of NF-κB activation and cell proliferation induced by TCR stimulation.


Introduction
Lymphocyte proliferation is important for proper response to infection, but uncontrolled proliferation predisposes individuals to lymphoproliferative disorders (LDs) or autoimmune disease [1,2]. Therefore, peripheral T cells exhibit inhibitory mechanisms to control the two major signals of T cell activation: the MHC-TCR interaction and the CD28 costimulation. Several E3 ligases including Protein pellino homolog 1 (PELI1), Gene Related to Anergy In Lymphocytes (GRAIL), Casitas B-lineage lymphoma b (Cbl-b) appear to play important roles in the downregulation of TCR signals [3]. PELI1 negatively regulates c-Rel by mediating ubiquitination of Lys48 during T-cell activation [4]. Casitas Blineage lymphoma b (Cbl-b) negatively regulates GDP/GTP exchange factor Vav1 phosphorylation [5], while GRAIL promotes CD3 receptor ubiquitination, leading to degradation [6]. Knockout of these E3 ligases in mice results in a high susceptibility to autoimmune diseases [4,7,8] in these animals, indicating that E3 ligase-mediated degradation of TCR signaling molecules is a major inhibitor of T cell activation.
In addition to these regulatory function of TRIM30 in macrophages and dendritic cells, high levels of TRIM30 expression are observed in diverse lymphocyte populations [15] and are inversely related to IL-2R expression in T cells [17]; however, whether TRIM30 controls the NF-kB pathway in lymphocytes is not known. To determine the physiological function of TRIM30 and its role in lymphocytes, we generated Trim30 knockout mice. Interestingly, we found no major defects in the activation of macrophages, but Trim30 2/2 CD4 + T cells exhibited enhanced proliferation and cell cycle progression upon TCR stimulation both in vitro and in vivo compared to wild-type cells; however, CD4 + T cell activation mediated by CD3 stimulation was compromised, particularly with regard to the Genomic DNA fragments from Trim30 +/2 and Trim30 2/2 progeny after Southern blotting with NcoI digestion. Wild-type alleles (10 kb) and the targeted alleles (8 kb) are indicated. (C) Genomic DNA isolated from Trim30 +/+ , Trim30 2/2 , and Trim30 +/2 was subjected to PCR. (D) Tissue Trim30 activation of NF-kB signaling. These findings revealed distinct roles for TRIM30 during the activation of naïve T cells upon TCR stimulation.

Results
Abnormal Age-associated CD4/CD8 Ratio in Trim30 2/2 Mice To understand the physiological function of TRIM30, we generated Trim30 knockout mice. The second exon, which contains the start codon, was replaced with a neomycin selection cassette following a stop codon ( Figure 1A), and the targeted construct was germ-line transformed to generate chimeric Trim30 +/2 mice, which were confirmed both by Southern blotting-and PCR-based genotyping ( Figure 1B-C). The intercross of the Trim30 + /2 heterozygotes produced offspring at the expected Mendelian ratio, indicating that the mutants experienced no obvious developmental defects associated with Trim30 deletion. RT-PCR analysis revealed high Trim30 transcript levels in lymphoid organs (spleen, thymus, and lymph node) and bone marrow in contrast to the low levels of Trim30 transcripts in nonhematopoietic tissues ( Figure 1D). The high levels of basal and induced expression of Trim30 in lymphocytes and macrophages were absent in the Trim30 knockout mouse. Immunoblot analysis of various tissues also confirmed the loss of TRIM30 protein expression in the lymph nodes, spleen, and thymus of Trim30 2/2 mice ( Figure 1E).
To validate its suggested role in NF-kB activation in macrophages, Trim30 +/+ and Trim30 2/2 bone marrow derived macrophages (BMDMs) were challenged with LPS or poly I:C then compared for their cytokine responses. The challenge with TLR ligands induced TRIM30 strongly only in wild-type cells, but there was no discernable difference in the expression of the major cytokines (IL-6, TNFa, IL-12 p40, IFNa, IFNb) ( Figure 1F). Besides, Trim30 2/2 mice showed no defect in resistance to LPS-induced septic shock, as well as in the response to Listeria monocytogenes infection ( Figure 1G). Therefore, TRIM30 appears dispensable for most TLR activations in macrophages.
In contrast to the inducible expression of Trim30 in macrophages, the high basal levels observed in lymphoid organs suggest that TRIM30 protein may be involved in the regulation of lymphocytes. To this end, we first assessed TRIM30 expression in T cells. Immunoblot analysis revealed that TRIM30 is highly expressed in both CD4 + T cells and CD8 + T cells purified from wild-type spleens (Figure 2A). TRIM30 is abundant in the naïve T cells, and high levels of TRIM30 were maintained after T cell activation with anti-CD3/CD28 antibodies or PMA/ionomycin costimulation ( Figure 2B). Comparison of T lymphocyte populations in thymus from Trim30 +/+ and Trim30 2/2 mice revealed no differences in the proportions of double-negative (DN), doublepositive (DP), CD4 single-positive (SP), or CD8SP thymocytes ( Figure 2C). In addition, the four DN populations, which were sorted based on the expression of CD25 and CD44, were comparable in Trim30 +/+ and Trim30 2/2 mice ( Figure 2D). Therefore, the development of T cells appeared normal in the Trim30 mutant mice. However, comparison of aged mice revealed significant difference in the ratios of peripheral CD4/CD8 T cells ( Figure 2E). As mice age, the relative ratio between CD4 + and CD8 + T cells gradually decreases [18,19]; however, in aged Trim30 2/2 mice, the CD4/CD8 ratio remained high in the spleen and lymph nodes. Absolute number of CD4 + T cells was slightly increased in Trim30 2/2 . Difference in CD4/CD8 ratio was due to more abundant CD4 + T cells in Trim30 2/2 mice. These finding suggest that long-term peripheral maintenance of T cells is dysregulated in Trim30 2/2 mice.
Hyper-proliferation of Trim30 2/2 CD4 + T cells We further investigated the role of TRIM30 in the response of CD8 + and CD4 + T cells in vitro. We labeled purified Trim30 +/+ and Trim30 2/2 splenic T cells with the division-tracking dye CFSE, and then stimulated the cells for 72 h with anti-CD3 antibody in the presence or absence of costimulation with anti-CD28 antibody. Upon stimulation with anti-CD3 agonist antibody, 74% of the Trim30 2/2 CD4 + T cells proliferated, while only 40% of the CD4 + T cells proliferated in wild type mice ( Figure 3A). The proliferation disparity between Trim30 +/+ and Trim30 2/2 CD4 + T cells decreased upon stronger stimulation in the presence of costimulatory signal. When cells were stimulated with PMA and ionomycin, which bypass TCR signaling, proliferation of Trim30 2/2 cells was comparable to that of Trim30 +/+ cells. These hyper-proliferation defects appeared specific to CD4 + T cells because CD8 + T cells from Trim30 2/2 mice proliferated normally in response to these stimuli ( Figure 3B). Therefore, TRIM30 seemingly represses premature CD4 + T cell proliferation, particularly in the absence of a costimulatory signal for TCR activation.
Subsequently, we determined whether hyper-proliferation of Trim30 2/2 CD4 + T cells is also observed during antigen-specific T cell stimulation. To this end, we isolated splenocytes from Trim30 2/2 mice crossed with OT-II transgenic mice. Consistent with the above results, Trim30 2/2 OT-II CD4 + T cells exhibited 10-20% enhanced proliferation compared with Trim30 +/+ OT-II CD4 + T cells, particularly when stimulated with a low dose of OVA 323-339 peptide ( Figure 3C). We cannot rule out the possibility of that loss of Trim30 might also compromise antigen presenting cells function. Nevertheless, we detected a normal cytokine expression in Trim30 2/2 macrophage, a type of antigen presenting cells ( Figure 1F), and the hyper-proliferation was more clear when only CD4 + T cells were used ( Figure 3A). Overall, these data indicate that TRIM30 negatively regulates CD4 + T cell proliferation following TCR activation in the absence of costimulatory signaling or low levels of TCR-specific antigens.

Increased Cell Cycle Progression in Trim30 Knockout T cells
To assess the role of TRIM30 in CD4 + T cell proliferation, we analyzed the cell cycle progression of Trim30 2/2 CD4 + T cells. mRNA expression from Trim30 +/+ and Trim30 2/2 mice. RT-PCR analysis revealed high Trim30 transcript levels in lymphoid organs (spleen, thymus, and lymph node) and bone marrow in contrast to the low levels of Trim30 transcripts in non-hematopoietic tissues (E) TRIM30 protein expression level in tissues from Trim30 +/+ and Trim30 2/2 mice as determined with immunoblotting using anti-Trim30 antibody. *, non-specific signal. (F) BMDMs were stimulated with LPS (200 ng/ml) or poly(I:C) (5 mg/ml), and Trim30 transcripts were quantified by quantitative RT-PCR. For detection of cytokine expression, Trim30 +/+ and Trim30 2/2 BMDMs were pretreated for 18 hr with LPS (LSP pre) and then restimulated with LPS (LPS re) indicated time or stimulated with poly(I:C) and transcripts for indicated cytokines were quantified by quantitative RT-PCR. Expression was normalized to GAPDH. (G) Survival of mice (n = 14 per group) given i.p injection of LPS (20 mg/kg) (upper panel). Survival of mice (n = 18 per group) given i.p infection of Listeria monocytogenes (2610 6 CFU per mouse) (lower panel). Data are representative results from three independent experiments. Error bars in D, E, F indicate s.d. doi:10.1371/journal.pone.0095805.g001 Figure 2. T cell lymphoid organ developmental progression is normal in Trim30 knockout mice. Immunoblot analysis of Trim30 expression in splenocytes and purified CD4 + and CD8 + T cells that were (A) unstimulated or, b-actin was used as a loading control (B) stimulated with anti-CD3 (2 mg/ml) and anti-CD28 (2 mg/ml) antibodies (CD3/CD28) or with 10 ng/ml of PMA and 500 ng/ml of ionomycin (P/I) for 3 days. GAPDH was used as a loading control. (C) Representative flow cytometric plots for CD4 and CD8 expression in the thymocyte population from Trim30 +/+ and Upon CD3 stimulation for 72 h, both Trim30 +/+ and Trim30 2/2 CD4 + T cells exhibited a sharp increase in BrdU and 7AAD incorporation, representing active DNA synthesis in S phase. However, more Trim30 2/2 (72.8%) than wild-type (49.3%) CD4 + T cells were found in S phase with a concomitant decrease in the corresponding cells in G0 phase ( Figure 4A). To determine whether Trim30 deletion has any effect on cell viability after TCR signaling, early and late apoptosis was analyzed by annexin V and PI staining ( Figure 4B). CD3 stimulation sharply increased cell viability in both Trim30 +/+ and Trim30 2/2 CD4 + T cells and resulted in no difference in the percentage of annexin V + PI 2 cells (early apoptotic cells) or annexin V + PI + cells (late apoptotic cells). Therefore, Trim30 deficiency caused cell cycle hyper-progression into S phase but did not affect CD4 + T cell death.
We next investigated the expression of cell cycle regulators. Immunoblot analysis showed that in Trim30 +/+ CD4 + T cells, retinoblastoma (Rb) protein phosphorylation and p27kip expression were down-and up-regulated, respectively, 48 h after anti-CD3 stimulation ( Figure 4C); however, Trim30 2/2 CD4 + T cells exhibited an earlier response in these expression patterns only 24 h after anti-CD3 stimulation. These results suggest that cell cycle progression in CD4 + T cells is indeed stimulated by the loss of TRIM30.

Higher Proliferation of Trim30 Knockout T cells in Rag1deficient Mice
To confirm the physiological relevance of the enhanced proliferative phenotype of Trim30 2/2 T cells in vitro, we determined whether Trim30 deletion influenced the homeostatic proliferation of T cells in lymphocyte-deficient mice. We purified CD4 + T cells from both Trim30 2/2 and wild-type mice, which expressed congenic markers CD45.2 and CD45.1, respectively, and labeled these cells with CFSE. In vitro-labeled wild type and Trim30 2/2 CD4 + T cells were mixed at a 1:1 ratio and transferred to Rag1 2/2 recipient mice, which lack B and T cells ( Figure 5A). The relative amounts of wild-type control and Trim30 2/2 splenic CD4 + T cells that were distinctly marked with the CD45.1 + and CD45.1 2 , respectively, were determined 4 days after the adoptive transfer. After selection of the adoptively transferred T cells using CD3 and CD4 markers, cell proliferation was analyzed by CFSE dilution (Fig. 5A). The competitive repopulation of adoptively transferred wild-type and Trim30 2/2 T cells revealed that the proliferating populations contained a significantly greater proportion of Trim30 2/2 T cells compared to Trim30 +/+ cells. Thus, twice the number of Trim30 2/2 T cells was detected in the recipient Rag1 2/2 mice after 4 days compared to wild-type T cells (Fig. 5B-C). These findings confirmed the negative function of TRIM30 in lymphopenia-induced T cell proliferation.
Reduced Effector Function of Trim30 2/2 CD4 + T cells The above results indicate that TRIM30 raises the activation threshold required for T cell proliferation. Therefore, in the absence of Trim30, T cells began to proliferate in response to even weak activation signals from the TCR. Because TCR engagement promotes effector function as well as proliferation, we also analyzed the defects caused by Trim30 deletion on T cell effector functions under the same conditions. Contrary to our expectations, Trim30 2/2 CD4 + T cells produced consistently lower levels of IL-2 compared to Trim30 +/+ CD4 + T cells upon stimulation with anti-CD3 antibody ( Figure 6A). In addition, expression of CD25 and CD69, which mark early T cell activation, was also reduced in Trim30 2/2 CD4 + T cells compared to wild-type CD4 + T cells ( Figure 6B); however, these defects were greatly diminished when cells were strongly stimulated with anti-CD3 and anti-CD28 agonists. Therefore, Trim30 deletion enhanced T cell proliferation but resulted in a higher threshold for activation of effector function via TCR activation.
We then investigated the mechanisms underlying this enhanced CD4 + T cell proliferation and decreased effector function in Trim30 2/2 mice. We stimulated CD4 + T cells for 90 min with anti-CD3 antibodies and then examined cell lysates for phosphorylation of classical TCR signaling molecules. Activation of classical TCR signaling pathways, including the ERK, JNK, and p38 kinases, appeared similar in Trim30 2/2 and Trim30 +/+ CD4 + T cells, although phosphorylation of IkB was lower in Trim30 2/2 CD4 + T cells ( Figure 6C). These observations suggest that TRIM30 is required for NF-kB activation following TCR activation with anti-CD3 agonist and that this effect occurs without enhancing proliferation during TCR activation in the absence of costimulatory signals.

Discussion
TRIM30 is the mouse homolog of human TRIM5a protein, which exerts an anti-retroviral function in various cell types; however, to date, the role of TRIM30 in mice has been linked only with innate immune responses in macrophages and dendritic cells [15,16,20]. In this report, we established TRIM30 as a key regulator of TCR-induced signaling in CD4 + T cells. A lack of TRIM30 caused increased proliferation and cell cycle progression upon TCR cross-linking with anti-CD3 agonist. Furthermore, TRIM30 deficiency led to a high CD4/CD8 ratio in aged mice and increased expansion of adoptively transferred CD4 + T cells in lymphopenic hosts. Another interesting feature of Trim30 2/2 CD4 + T cells is the relatively lower effector function upon CD3 ligation, despite a highly proliferative phenotype. Thus, suppression of T cell proliferation concomitant with promotion of the effector function via TRIM30 may be required to prevent abnormal proliferation of T cells in the absence of optimal T cell activation signals, allowing adequate time for cells to properly differentiate into the correct T cell types before full activation.
Trim30 2/2 CD4+ T cells show the enhanced proliferation despite the reduced IL-2 production. We suggest that TRIM30 mediates T cell expansion through IL-2-independent proliferation. In vitro and in vivo studies present the evidence for IL-2independent cell cycle regulatory mechanisms [21,22]. These studies established that IL-2-independent mechanisms are involved in TCR/CD28 mediated cell cycle controlling [23,24]. Reduced p27kip expression and enhanced phosphorylation of Rb in Trim30 2/2 CD4 + T cells suggest this possibility. Trim30 2/2 mice. The average percentages and absolute cell number of CD4 2 CD8 2 double negative (DN), CD4 + CD8 + double positive (DP), and CD4 + or CD8 + single positive (CD4SP, CD8SP) thymocytes from four independent experiments are shown on the right. (D) Representative flow cytometric plots for CD44 and CD25 expression in the identified DN1 (CD44 + CD25 2 ), DN2 (CD44 + CD25 + ), DN3 (CD44 2 CD25 + ), and DN4 (CD44 2 CD25 2 ) T cell populations. DN1-DN4 thymic populations and absolute cell number from four experiments are shown on the right. (E) Ratios of CD4/CD8 on lymph node cells and splenocytes isolated from 10-week-(young) and 12-month-old (aged) Trim30 +/+ and Trim30 knockout mice. For this analysis, at least four young mice or 12 aged mice were analyzed. Absolute cell number of independent experiment are shown on the right. The P-value was calculated using a t-test. *, P,0.05; ***, P,0.0005. Error bars in C, D, and E represent the s.d. doi:10.1371/journal.pone.0095805.g002 Recent studies showed that proliferation and effector function are regulated by different mechanisms [25,26]. For example, the metabolic switch from oxidative phosphorylation to aerobic glycolysis is specifically required for effector function in T cells but not for proliferation [25]. On the other hand, a high multiplicity of TCR ITAM is required for T cell proliferation, whereas a low multiplicity of CD3 ITAM was sufficient to lead to cytokine secretion [26]. Therefore, activation of effector functions and proliferation upon TCR stimulation can be regulated independently.
Although both CD4 + and CD8 + T cells express TRIM30, our results show that the hyper-proliferation caused by Trim30 deletion is only appear in CD4 + T cells, but not in CD8 + T cells. Because purified CD4 + T cells by using CD4 T Cell Isolation Kit contain Treg cells, the possibility of non-autonomous roles of Trim30 in Treg cells still exist. The percentage of Treg cells (CD25 + Foxp3 + ) in splenic CD4 + T cells was 9 to 10%, but there was no difference in their abundance between Trim30 +/+ and Trim30 2/2 mouse (data not shown). It is necessary to follow-up studies addressing the roles of Trim30 in Treg and CD8 + T cells, albeit not to proliferation mechanism.
Notably, the effects of TRIM30 deficiency were apparent only when cells were stimulated with weak T cell activation signals. T cells should be resistant to activation following weak stimulatory signals in order to avoid hyperactivation of immune responses. We propose that TRIM30 regulates the threshold to effector function and proliferation in the presence of incomplete T cell stimulation; although, strong stimuli may override the regulatory activity of TRIM30 and induce TCR-mediated T cell activation.
In contrast to previous reports on Trim30 knockdown or overexpression in monocytes [15,16], we observed reduced NF-kB activation in Trim30 2/2 CD4 + T cells. This discrepancy may result from differences in the experimental conditions used in these studies (cell line-based Trim30 knockdown vs. Trim30 knockout mouse) or from difference in the tissue types (macrophage vs. T cells). Intriguingly, similar differences were also observed in the analysis of TAB2 and TAB3 mutant mice. TAB2 and TAB3, which are involved in NF-kB activation in monocytes (21), were reported as putative TRIM30 targets in knockdown experiment (13); however, TAB2 knockout mice have normal IL-1b production in response to IL-1R signaling [27]. A recent study of TAB2 and TAB3 knockout mice revealed that TAB2 and TAB3 are dispensable for activation of NF-kB and MAPKs in macrophages and MEFs [28]. Instead, TAB2 and TAB3 were shown to have critical role to control B cell activation by specifically regulating MAPKs [28]. Although the exact role of TRIM30 remains to be elucidated, our data showed that TRIM30 is required for optimal effector function of CD4 + T cells via NF-kB activation and concomitantly negatively regulates the proliferation of these T cells.

Ethics Statement
All animal experiments were performed in accordance with the Korean Food and Drug Administration (KFDA) guidelines. Protocols were reviewed and approved by the Institutional Animal

Animal and Infection
All mice were maintained in the specific pathogen-free facility of the YLARC at Yonsei University following the IACUC guidelines. In experiments for survival study, male six-to ten-week-old littermate (22.961.6 g) Trim30 +/+ and Trim30 2/2 mice (n = 14,18 per group) were used. Mice were challenged with LPS (20 mg/kg) or Listeria Monocytogenes (2610 6 CFU per mouse) by intraperitoneal (i.p) route. The experiment was repeated 3,4 times, and data were pooled. The Kaplan-Meier method was used to plot survival curve. Survival and health status of infected mice were monitored daily. Mice reaching the humane endpoint (compound-related severe illness, as indicated by inability to eat and drink or a lack of response when gently prodded with a forceps) were euthanized by cervical dislocation under anesthesia with isoflurane. All mice were sacrificed by same method.

Flow Cytometry Analysis
We isolated mononuclear cells from thymus, spleen, and lymph nodes of Trim30 +/+ and Trim30 2/2 mice. These cells were stained with antibodies to surface antigens for 30 min at 4uC. After several washes, cytofix/cytoperm was added and incubated for 30 min at 4uC. CD4 + T cell apoptosis was assessed using annexin V-FITC and propidium iodide (PI) staining. Samples were loaded on a FACS Calibur, and data analyses were performed with Flow Jo software.

Quantitative Real-time PCR
Total RNA was isolated using the Trizol reagent (Invitrogen) according to the manufacturer's instructions. Then, cDNA was right. The P-values were calculated with a t -test. *, P,0.05. (B) Apoptotic ratio of Trim30 +/+ and Trim30 2/2 CD4 + T cells with (CD3) or without (No) anti-CD3 stimulation. The percentages of cells in early (annexin positive and PI negative) and late (double positive) apoptosis are given along with those of the viable cells (double negative). The average of five independent experiments is presented with error bars on the right. (C) Immunoblot analysis of cell cycle proteins in whole cell lysates of purified Trim30 +/+ and Trim30 2/2 CD4 + T cells following stimulation with the anti-CD3 antibody (2 mg/ml) using the indicated antibodies for the indicated time period. Error bars in A and B denote s.d. doi:10.1371/journal.pone.0095805.g004 synthesized from 2 mg of the isolated RNA using SuperScript II Reverse Transcriptase (Invitrogen) with random hexamers or oligo-dT as the primers. After synthesis, cDNA was diluted 1:10 in triple distilled water (TDW), and 3 ml was used for real-time PCR using a Bio-Rad CFX. Gene expression levels were normalized to GAPDH RNA levels. The following primers were used: Trim30 forward primer (59-TTCCAGAGGAGGAGCAGAAGGTG-39), Trim30 reverse primer (59-TCACAAAGCCAGCAGATGAC-CATC-3), GAPDH forward primer (59-GGCAAATTCAACGG-CACAGTCAAG-39), GAPDH reverse primer (59-TCGCTCCTGGAAGATGGTGATGG-39), IL-6 forward primer (59-TAAGCATATCAGTTTGTGGAC-39), IL-6 reverse primer Figure 6. Effects of Trim30 deficiency on classical TCR signaling. (A) Primary CD4 + T cells from Trim30 +/+ and Trim30 2/2 mice were stimulated with the indicated antibodies for 3 days, and then the supernatants were collected and assayed for IL-2 production by ELISA. Concentration of IL-2 is shown for three independent experiments. Error bars represent s.d. (B) The numbers of CD252 and CD692 expressing CD4 + T cells from Trim30 +/+ (gray area) and Trim30 2/2 (white area) mice 24 h after treatment with anti-CD3 (2 mg/ml) antibody are shown. Mean fluorescence intensity (MFI) values are given. (C) Purified CD4 + T cells from Trim30 +/+ and Trim30 2/2 mouse spleens were stimulated with anti-CD3 antibody (2 mg/ml) and anti-IgG (10 mg/ml) for the times indicated at the top. Lysates from CD4 + T cells were analyzed by immunoblotting using the indicated antibodies. The bar graph shows the averages of the band intensities. Data are representative results of at least three independent experiments. Error bars in A and C denote the s.d. doi:10.1371/journal.pone.0095805.g006