Coactivators to Ensure Thymocyte Survival t Recruits Steroid Receptor ...

The Journal of Immunology

ROR␥t Recruits Steroid Receptor Coactivators to Ensure Thymocyte Survival Huimin Xie, Maureen S. Sadim,1 and Zuoming Sun2 Thymocytes undergo apoptosis unless a functional TCR is assembled. Steroid receptor coactivators (SRCs) regulate nuclear receptor-mediated transcription by associated histone acetyltransferase activity. However, it has been a challenge to demonstrate the in vivo function of SRCs due to the overlapping functions among different members of SRCs. In this study, we show that recruitment of SRCs is required for thymic-specific retinoic acid-related orphan receptor ␥ (ROR␥)t-regulated thymocyte survival in vivo. An activation function 2 domain, identified at the carboxyl terminus of ROR␥t, is responsible for recruiting SRCs. A mutation in the activation function domain (Y479F) of ROR␥t disrupted the interaction with SRCs and abolished ROR␥tmediated trans-activation but not its ability to inhibit transcription. Transgenes encoding the wild-type ROR␥t, but not the mutant, restored thymocyte survival in ROR␥ null mice. Our results thus clearly demonstrate that ROR␥t recruits SRCs to impose a gene expression pattern required to expand the life span of thymocytes in vivo, which increases the opportunities for assembling a functional TCR. The Journal of Immunology, 2005, 175: 3800 –3809.

T

hymocytes are subject to a selection process critical for the development of fully functional mature T cells. The life span of thymocytes limits the progression of T cell development (1). Antiapoptotic molecules such as Bcl-xL extend the life span of thymocytes to ensure completion of the developmental process. Thymocytes lacking Bcl-xL undergo premature apoptosis (2, 3), whereas overexpression of Bcl-xL increases thymocyte survival and the chance for completing the selection process (1). Previously we, and subsequently others, have shown that retinoic acid-related orphan receptor ␥ (ROR␥),3 a member of the steroid nuclear receptor family, regulates thymocyte survival via up-regulation of Bcl-xL (4, 5). ROR␥ was initially cloned by its sequence homology to retinoid hormone receptor (6). Another isoform, thymus-specific ROR␥ (ROR␥t), was subsequently cloned while screening for the molecules that inhibit activation-induced cell death in a T cell hybridoma (7). ROR␥t is a truncated form of ROR␥ lacking the first 24 aa due to use of a downstream translation start site. Although ROR␥ is expressed in multiple tissues, ROR␥t is expressed exclusively in thymocytes and a population of lymphoid tissue inducers at the embryonic stage (8, 9). Knockout of the ROR␥ gene, disrupting the expression of both ROR␥ and ROR␥t, resulted in massive apoptosis of thymocytes and defective development of lymph nodes (4, 5). However, less is known about the molecular mechanisms responsible for ROR␥t-regulated thymocyte survival. A specific and heritable pattern of gene expression is imposed upon T cell lineage during the development in the thymus (10). Department of Microbiology and Immunology, College of Medicine, University of Illinois, Chicago, IL 60612 Received for publication June 1, 2005. Accepted for publication July 11, 2005. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 Current address: Medicine-Hematology and Oncology, Northwestern University, 710 North Fairbanks, Olson 8370, Chicago, IL 60611. 2 Address correspondence and reprint requests to Dr. Zuoming Sun, Department of Microbiology and Immunology, College of Medicine, University of Illinois at Chicago, 835 South Wolcott (M/C790), Chicago, IL 60612. E-mail address: [email protected]

Abbreviations used in this paper: ROR␥, retinoic acid-related orphan receptor; ROR␥t, thymic-specific ROR␥; SRC, steroid receptor coactivator; AF2, activation function 2; GRIP, glucocorticoid receptor-interacting protein; Tg, transgenic; IRES, internal ribosome entry site; SP, single positive; DP, double positive; TCRin, TCR intermediate; FasL, Fas ligand. 3

Copyright © 2005 by The American Association of Immunologists, Inc.

Modification of histone by acetylation is a critical step in determining the pattern of gene expression by controlling the magnitude of gene activity (11, 12). Steroid receptor coactivators (SRCs) have the intrinsic or associated histone acetyltransferase activity necessary to remodel the chromatin structure by acetylation of histone. Acetylation is believed to relieve the chromatin-mediated transcriptional repression, resulting in active transcription (13). SRCs consist of three proteins with 50 –55% homology, SRC1 (NcoA1), glucocorticoid receptor-interacting protein (GRIP)1 (TIF2), and SRC3 (p/CIP/RAC3/ACTR/AIB1/TRAM1) (14). In vitro transfection analysis demonstrated that the LXXLL motifs, present on the surface of SRCs, make contact with activation function 2 (AF2) domains located at the carboxyl terminus of ligandbinding domains of nuclear receptors (15–18), resulting in the recruitment of SRCs to nuclear receptors (19, 20). However, it has been a challenge to demonstrate the in vivo function of recruitment of SRCs due to overlapping functions among various members (14). Knockout of an individual SRC member only results in a relatively minor phenotype (21–23), whereas double knockout of SRC1 and GRIP1 is lethal (24). Thus, the in vivo function of a given coactivator may not be revealed by a traditional knockout approach. In this study, we show that ROR␥t recruits SRCs via its AF2 domain to stimulate transcription. In addition, by an AF2-domainindependent pathway, ROR␥t is able to inhibit the activity of NFAT. We created a mutation in the AF2 domain of ROR␥t that disrupted its interaction with SRCs, but not its ability to inhibit NFAT activity. ROR␥ null mice were reconstituted with a wildtype ROR␥t or an AF2 mutant incapable of binding to SRCs or a DNA-binding mutant that does not bind to its target DNA sequence. The massive thymocyte apoptosis was reversed by the transgene encoding the wild-type ROR␥t, but not the DNA-binding mutant, which acts as a negative control. Similar to DNA-binding mutant, AF2 failed to restore thymocyte survival, clearly demonstrating that recruitment of coactivators by ROR␥t is required to establish a gene expression pattern critical for thymocyte survival in vivo.

Materials and Methods Plasmids and reagents The plasmids pSG5-HA-GRIP1 and its LXXLL motif mutant were gifts from Dr. Michael Stallcup (Department of Pathology, University of Southern California, Los Angeles, CA). The third motif of GRIP1 was mutated 0022-1767/05/$02.00

The Journal of Immunology to LXXAA by PCR-based mutagenesis. To generate expression plasmids encoding ROR␥t and its mutants, the DNA fragment containing ROR␥t cDNA was cloned in the pEF-Bos vector that contains a promoter of the elongation factor. ROR␥t reporter plasmid was a generous gift from Dr. Anthony Means (Department of Pharmacology and Cancer Biology, Duke University Medical Center, Durham, NC). The following Abs for FACS analyses were purchased from BD Pharmingen: PE-anti-TCR␤ (catalog no. 553172), biotin-anti-CD4 (catalog no. 553649), streptavidin-PE-Cy5 (catalog no. 554062), annexin V-PE (catalog no. 556421), PE-anti-CD8 (catalog no. RM2204), and anti-Bcl-xL. The hamster anti-ROR␥t Ab was previously described (4). Anti-GRIP1 (MS-1140-P1ABX) and anti-SRC1 (MS-1130-P1ABX) were purchased from Neomarkers; actin (SC8432) and p27kip1 (SC528G) Abs were obtained from Santa Cruz Biotechnology.

Cell culture, transient transfection, and reporter assays 293T cells were cultured in DMEM supplemented with 10% FBS, penicillin, and streptomycin. 293T cells (2 ⫻ 105) were plated in 12-well dishes for 20 h. The cells were then transfected with 10 ng of a reporter construct, 100 ng of ROR␥t (wild type or mutants), 200 ng of SRC1 or GRIP1 expression plasmids, or with the indicated expression plasmids using the calcium phosphate method. The cells were fed with fresh medium 16 h after transfection. Then, after 24 h, cells were lysed in 200 ␮l of lysis buffer (137 mM NaCl, 50 mM Tris, and 0.5% Nonidet P-40), and 5 ␮l was used for luciferase assays as described in the protocol (Promega). A SV40-␤galactosidase expression plasmid was cotransfected for internal controls.

Immunoprecipitation and Western blot analysis Thirty-six hours after transfection of the indicated expression plasmids, 293T cells were lysed in 1 ml of immunoprecipitation lysis buffer (150 mM NaCl, 50 mM Tris, 4 mM KCl, 1 mM MgCl2, 1 mM Na3VO4, 10% glycerol, 1% Nonidet P-40, and protease inhibitor). In the case of using thymocytes, 1 ⫻ 107 cells or sorted GFP⫹ thymocytes were lysed. Lysates were then precleared with 50 ␮l of 50% slurry protein A (Zymed Laboratories) for 3 h. The protein A beads were spun down, and the supernatant was incubated with 1 ␮g of monoclonal anti-SRC1 or GRIP1 Ab for 1 h. Twenty microliters of protein A beads was then added to the above mixtures for an additional 20 min. The protein A beads were spun down and washed four times with washing buffer (lysis buffer without protease inhibitor). The protein A pellets were then boiled in 50 ␮l of SDS loading buffer and resolved on a 10% SDS gel. Western blot analyses were performed with the indicated Abs and developed with ECL (Amersham Biosciences).

Generation of transgenic (Tg) mice Three Tg constructs were cloned as shown in Fig. 4a, encoding either wild-type or mutants ROR␥ driven by a CD4 promoter containing CD4 enhancer but not silencer. The NotI DNA fragment was microinjected into fertilized eggs by Transgenic Production Service in UIC. The founder mice were screened by Southern blot and FACS analyses of GFP expression. The Tg mice were then bred to ROR␥ null mice that were backcrossed to C57BL/6 for at least six generations. All of the mice used in this study were housed in a specific pathogen-free facility at the University of Illinois, Chicago.

Flow cytometric analysis and sorting Thymocytes were stained in the dark with the indicated Abs on ice for 20 min, washed with FACS buffer (PBS with 0.02% sodium azide/1% FBS), and analyzed by a BD Biosciences FACSCalibur with CellQuest software. To analyze the PBLs, RBCs were lysed with ACK buffer (0.15 M NH4Cl, 1 mM KHCO3, and 1 mM EDTA) (Cambrex Bioscience) on ice for 5 min. PBLs were then stained with the indicated Abs. For isolation of GFP⫹CD4⫹CD8⫹ cells, thymocytes were stained with both CD8-PE and CD4-PE-Cy5 and sorted by a DakoCytomation MoFlo high-speed sorter. The GFP⫹CD4⫹CD8⫹ cells were sorted with ⬎98% purity.

Apoptosis assays Thymocytes isolated from each genotype of mice were cultured in RPMI 1640 medium supplemented with 10% FBS at 37°C for different times. The dead cells were then detected by annexin V-PE and propidium iodide staining for 20 min as described previously (4).

Results Identification of a conserved AF2 domain that is responsible for recruiting SRCs To investigate whether ROR␥t recruits other proteins, we first performed sequence analyses to identify the conserved domains

3801 known to be important for nuclear receptors. An AF2 domain (LYKELF) was identified at the carboxyl terminus of the ROR␥t ligand-binding domain, and it was conserved in all members of the ROR family (Fig. 1a). ROR␥, which shares the same DNA-and ligand-binding domains with ROR␥t, also contains this AF2 domain. Ligand-binding domains of nuclear receptors have a common structural feature with 12 ␣ helices (H1–H12) (13). To determine whether the putative AF2 domain of ROR␥t is also conserved in a three-dimensional structure in addition to the primary sequence, we developed a computerized model for the ROR␥t ligand-binding domain (Fig. 1b). Similar to other nuclear receptors, the ligand-binding domain of ROR␥t also contains 12 ␣ helices, and the AF2 domain of ROR␥t is located at the H12 in the predicted model. The AF2 domain is a crucial component for recruiting SRC members containing LXXLL motifs (2). To determine whether ROR␥t associates with SRCs, we conducted immunoprecipitation assays. ROR␥t and SRCs were transiently expressed in 293T cells and immunoprecipitations using Abs specifically against either SRC1 or GRIP1 were conducted (Fig. 1, c and d). It is clear that ROR␥t is capable of forming complexes with both SRC1 and GRIP1. To determine whether the AF2 domain is essential for association with SRCs, an AF2 mutant (ROR␥t-AF2) was created by changing the critical tyrosine 497 in the AF2 domain to phenylalanine (Y479F). As a control, we also created a DNA-binding domain mutant (ROR␥t-DBD) by changing the two arginines located at the base of the DNA-binding zinc finger to alanine and glycine (R35R36-AG). As expected, the ROR␥t-DBD failed to bind to the oligonucleotides containing a ROR␥t binding site in band shift analyses (Fig. 1f), but still formed a complex with SRC1 or GRIP1 (Fig. 1, c and d), demonstrating that mutation in the ROR␥t DNA-binding domain did not affect its interaction with SRCs. In contrast to wild-type ROR␥t and ROR␥t-DBD, ROR␥tAF2 was barely detected in the complexes immunoprecipitated by the specific anti-SRC1 or GRIP1 Ab (Fig. 1, c and d), clearly demonstrating that the AF2 domain is required for ROR␥t to recruit SRCs. To determine whether LXXLL motifs of SRCs are required for the interaction with ROR␥t, we obtained a mutant GRIP1 incapable of binding to nuclear receptors such as thyroid hormone receptor due to the mutations in two of its three LXXLL motifs (25). We generated additional mutations in the third LXXLL motif (LLQLL to LLQAA), so that all three LXXLL motifs were mutated to create GRIP1m. In contrast to wild-type GRIP1, this mutant failed to form a complex with ROR␥t in the immunoprecipitation assays (Fig. 1e), suggesting a critical role for LXXLL motifs of SRCs in the interaction with ROR␥t. Similar results were also obtained from the GRIP1 that had two of the three LXXLL motifs mutated (data not shown), suggesting that mutating two LXXLL motifs is sufficient to disrupt interaction with ROR␥t. Altogether these results demonstrated that the ROR␥t-mediated recruitment of SRCs depends on the AF2 and LXXLL motifs. Recruitment of SRCs augments ROR␥t-mediated trans-activation but has no effect on its ability to inhibit NFAT SRCs are key to enhancing nuclear receptor-mediated transcriptional responses (12, 14). We therefore determined whether recruitment of SRCs regulates ROR␥t-mediated transcriptional activity. A luciferase reporter driven by a TK promoter with three upstream ROR␥t binding sites was used to monitor ROR␥t-mediated transcription. Expression of ROR␥t or SRC1 alone only slightly enhanced reporter activity, whereas significant enhancement of the transcriptional activity was observed in the presence of SRC1 in combination with ROR␥t but not ROR␥t-AF2 (Fig. 2a),

3802

CRITICAL ROLE OF SRCs IN ROR␥t-MEDIATED THYMOCYTE SURVIVAL

FIGURE 1. Identification of an AF2 domain of ROR␥t responsible for recruiting SRCs. a, Identification of a conserved AF2 in ROR␥t ligand-binding domain. Conserved sequence (LYKELF) of the AF2 domain among members of the ROR family. b, A model of ROR␥t ligand-binding domain was predicted by a computer based on the homology to the ligand-binding domains of other nuclear receptors. The AF2 domain is located at helix 12 (H12). The critical tyrosine 479 (Y479) is indicated. c, ROR␥t and ROR␥t-DBD but not ROR␥t-AF2 were coimmunoprecipitated with SRC1. SRC1 was transiently expressed in 293T cells along with ROR␥t, ROR␥t-DBD, and ROR␥t-AF2. SRC1 was then immunoprecipitated (IP) with anti-SRC1-specific Ab or isotype control Ab (C). The presence of various ROR␥t in the immunoprecipitated complexes was detected by Western blot analyses probed with anti-ROR␥t Ab. Lower panels are the expression levels of SRC1 and various ROR␥t in lysates (input). d, ROR␥t and ROR␥t-DBD but not ROR␥t-AF2 coimmunoprecipitated with GRIP1. Similar assays as described in c, but GRIP1 instead of SRC1 were expressed in 293T cells with various ROR␥t. e, Wild-type GRIP1 but not the GRIP1 with mutated LXXLL motifs (GRIP1m) coimmunoprecipitated with ROR␥t. ROR␥t was transiently expressed in 293T cells along with wild-type GRIP1 or GRIP1m. GRIP1 was then immunoprecipitated with anti-GRIP1-specific Ab. The presence of ROR␥t in the immunoprecipitated complexes was detected by Western blot analyses probed with anti-ROR␥t Ab. Lower panels are the expression levels of GRIP1 and ROR␥t in lysates (input). f, 32P-labeled oligonucleotides (probe) containing a ROR␥t binding site was band shifted by bacterially expressed ROR␥t but not ROR␥t-DBD. Data shown are representative of at least three independent experiments.

suggesting that SRC1 recruited by the AF2 domain enhances ROR␥t-mediated transcription. The negative control, ROR␥tDBD, failed to stimulate ROR␥t-DBD activity in the presence of SRC1 as expected, since binding to the target DNA response element is required for a transcription factor to mediate transcription (Fig. 2a). Similar results were obtained using another SRC member, GRIP1 (Fig. 2b), suggesting that SRC members regulate ROR␥t-mediated transcription. Furthermore, wild-type GRIP1, but not GRIP1m incapable of binding to ROR␥t, stimulated

ROR␥t activity (Fig. 2c). These results clearly demonstrated that SRCs recruited by the AF2 domain enhance ROR␥t-mediated transcriptional responses. In addition to trans-activation by recruiting SRCs, we found that ROR␥t was able to inhibit the activity of an important T cell transcription factor, NFAT. The activity of a NFAT reporter was stimulated by forced expression of NFAT as expected (Fig. 2d). However, the stimulated NFAT activity was inhibited by ROR␥t in a dose-dependent manner. NFAT is required to stimulate the IL-2


3803

FIGURE 2. Recruitment of SRCs augments ROR␥t-mediated trans-activation but has no effect on its ability to inhibit NFAT. a, SRC1 facilitates wild-type ROR␥t but not ROR␥t-AF2 to stimulate transcription. Expression plasmids for the wild-type ROR␥t or ROR␥t-DBD or ROR␥t-AF2 were transfected to 293T cell alone (striped bars) or in combination with SRC1 expression plasmid (filled bars). ROR␥t-mediated transcriptional activity was monitored using a luciferase reporter driven by a TK promoter with three upstream ROR␥t binding sites (reporter). Lower panels are the expression levels of various ROR␥t and SRC1 detected by Western blot analyses. b, GRIP1 facilitates wild-type ROR␥t but not ROR␥t-AF2 to stimulate transcription. Similar experiments as described in b were performed, but GRIP1 expression plasmid replaced SRC1 expression plasmid. c, Wild-type GRIP1 but not GRIP1 containing mutated LXXLL motifs (GRIP1m) stimulated ROR␥t-mediated transcription. Expression plasmids for wild-type GRIP1 or GRIP1m were transiently expressed in 293T cell alone (䡺) or along with ROR␥t expression plasmid (f). The transcriptional activity was monitored by a ROR␥t reporter. Right panels are the expression levels of ROR␥t, GRIP1, and GRIP1m detected by Western blot analysis. d, Both wild-type ROR␥t and ROR␥t-AF2 inhibit transcription activity of NFAT. NFAT is transiently expressed in 293T cells alone or with increasing amounts of wild-type ROR␥t or ROR␥t-AF2. NFAT-mediated transcription activity was monitored by a NFAT reporter with three upstream NFAT binding sites. NFAT activity is indicated as the fold of stimulation relative to the activity obtained from cells transfected with reporter alone. Lower panels are the expression levels of ROR␥t or ROR␥t-AF2 and NFAT. e, Forced expression of SRC1 does not affect ROR␥t-mediated inhibition of NFAT. Expression plasmid for NFAT was transfected to 293T cells with expression plasmids encoding proteins as indicated. A NFAT reporter was used to monitor NFAT activity. The reporter activity with error bars denoting SD was averaged from at least three independent experiments.

3804


gene during T cell activation (26). Indeed, overexpression of ROR␥t inhibited IL-2 promoter activity induced by TCR stimulation or PMA/ionomycin treatment (data not shown). One of the common mechanisms for nuclear receptor-mediated inhibition of transcription is the sequestration of coactivators such as SRCs (19). However, overexpression of SRC1 or GRIP1 did not relieve ROR␥t-mediated inhibition of NFAT (Fig. 2e, and data not shown). Furthermore, the ROR␥t-AF2 incapable of binding to SRCs inhibited NFAT activity as efficiently as the wild type (Fig. 2d). Recruitment of SRCs is thus required for ROR␥t-mediated trans-activation but not for inhibition of NFAT activity. Generation of Tg mice To determine the in vivo function of ROR␥t-mediated recruitment of SRCs, we established Tg mice expressing ROR␥t (ROR␥tTg), ROR␥t-DBD (ROR␥t-DBDTg), and ROR␥t-AF2 (ROR␥t-AF2Tg). Since ROR␥t is specifically expressed in thymocytes, we used a CD4 promoter that was able to target transgene expression to T cell compartments, including CD4⫹CD8⫹ thymocytes (27). The cDNAs encoding various ROR␥t were cloned downstream of a CD4 promoter along with an internal ribosome entry site (IRES) and an enhanced GFP coding region (Fig. 3a). The IRES permits both ROR␥t and GFP to be expressed from a single bicistronic mRNA. The founder mice were screened with Southern blot analysis with ROR␥t cDNA as a probe (Fig. 3b). The transgenes were successfully integrated into the genome as indicated by the detection of a 3-kb DNA fragment in BamHI-digested genomic DNA. Expression of GFP along with ROR␥t makes it possible to identify the cells that express transgene. We thus analyzed GFP expression in thymocytes using a flow cytometer (Fig. 3c). Among the Tg mice identified positive by Southern blot analyses, seven mice displayed GFP expression when gated on CD3⫹ T cells, whereas GFP was not detected in the mice negative for transgene, suggesting that the transgene was correctly targeted for expression in T cell compartments. Progeny from different lines of Tg mice expressing the same transgene exhibited a similar phenotype, and in this study results from one line are reported. Since not all of the T cells express GFP, presumably due to the variegation effect (28), we were able to compare the GFP⫹ and GFP⫺ T cells obtained from the same Tg mouse as shown later. The GFP expression also allowed us to sort GFP⫹ cells for analyses, especially in the case of ROR␥t-AF2Tg mice that only contain a relatively small portion, ⬃17%, of GFP⫹ cells in thymus (Fig. 3c). The Tg mice were then crossed to the ROR␥ null mice to eventually obtain the ROR␥ null Tg expressing wild-type ROR␥t (ROR␥⫺/ROR␥tTg), DNA-binding mutant (ROR␥⫺/ROR␥t-DBDTg), or AF2 mutant (ROR␥⫺/ROR␥t-AF2Tg). To confirm whether ROR␥t and the two mutants expressed from the transgenes exhibit similar interactions with SRCs as observed in in vitro experiments, we conducted immunoprecipitation assays using thymocytes obtained from different genotypes of mice. AntiSRC1 (Fig. 3d)- or GRIP1 (Fig. 3e)-specific Ab coimmunoprecipitated ROR␥t in thymocytes obtained from wild-type but not mutant mice, confirming that ROR␥t associates with SRCs in vivo. The immunoprecipitation assays were then performed using GFP⫹ cells sorted from ROR␥-deficient mice expressing various transgenes. These GFP⫹ cells expressed ROR␥t only from transgenes, thus excluding the interference from the endogenous ROR␥t. Similar to the results obtained from transiently transfected cells (Fig. 1), ROR␥t and ROR␥t-DBD but not ROR␥t-AF2 coimmunoprecipitated with GRIP1 and SRC1 (Fig. 3, d and e), suggesting that ROR␥t interacts with members of SRCs in vivo in an AF2-dependent manner. In addition, our results also indicated that the trans-

genes encoding the wild-type and mutant ROR␥t were equivalently expressed (lower panels, Fig. 3, d and e). Transgene encoding the ROR␥t, but not ROR␥t-DBD and ROR␥t-AF2, rescued thymocyte survival To determine the effect of the ROR␥t transgene on thymocyte survival, we analyzed apoptosis of the thymocytes. As reported earlier, ⬃80% of the CD4⫹CD8⫹ROR␥⫺/⫺ thymocytes died after 6 h in medium, whereas in wild-type (ROR␥⫹/⫹) or heterozygous (ROR␥⫹/⫺) mice ⬍20% of the thymocytes underwent apoptosis (Fig. 4a). In the ROR␥⫹/⫺ mice, the transgenes encoding either wild-type or mutant ROR␥t had no significant effect on the survival of thymocytes (data not shown), likely due to the presence of endogenous ROR␥t. We next examined the effect of transgenes on the survival of ROR␥⫺/⫺ thymocytes by analyzing the apoptosis of CD4⫹CD8⫹GFP⫹ and CD4⫹CD8⫹GFP⫺ cells sorted from different genotypes of mice. GFP⫹ cells from ROR␥ null mice expressing the wild-type ROR␥t transgene (ROR␥⫺/ROR␥tTg) survived as well as cells from wild-type mice, whereas GFP⫺ cells underwent rapid apoptosis similar to the ROR␥ null mice (Fig. 4b), suggesting that expression of the ROR␥t is sufficient to rescue thymocytes from apoptosis. Correspondingly, the reduced thymic cellularity and the antiapoptotic Bcl-xL both returned almost to the wild-type levels (Fig. 4, e and f). We previously observed that ROR␥⫺/⫺ thymocytes displayed abnormal cell cycle progression because of reduced levels of p27kip1 (4). Wild-type ROR␥tTg prevented down-regulation of p27kip, suggesting that the cell cycle progression was also restored (Fig. 4f). In contrast to wild-type ROR␥tTg, ROR␥t-DBDTg, the negative control, failed to rescue the ROR␥⫺/⫺ thymocytes from apoptosis (Fig. 4c), confirming that ROR␥t, a transcription factor, regulates thymocyte survival by DNA-binding-dependent mechanisms. Similar to ROR␥t-DBDTg, ROR␥t-AF2Tg also failed to restore the survival of ROR␥t⫺/⫺ thymocytes (Fig. 4d). Accordingly, neither ROR␥t-DBDTg nor ROR␥t-AF2Tg restored thymic cellularity (Fig. 4e) and the levels of Bcl-xL and p27kip1 (Fig. 4f). We also compared apoptosis of GFP⫹ and GFP⫺ cells in the same mice. In contrast to the ROR␥⫺/ ROR␥tTg mice (Fig. 4b), apoptosis of the GFP⫺ cells in ROR␥⫺/ ROR␥t-DBDTg and ROR␥⫺/ROR␥t⫺/AF2Tg mice was similar to that of the GFP⫹ cells (Fig. 4, c and d), strongly suggesting that wild-type ROR␥t but not ROR␥t-DBD and ROR␥t-AF2 supported CD4⫹CD8⫹ thymocyte survival. These results clearly demonstrated a critical role of AF2 domain-mediated recruitment of coactivators in vivo. Flow cytometric analyses of surface expression of CD4, CD8, and TCR on ROR␥⫺/⫺ thymocytes revealed a different expression pattern from that of the wild-type, most likely due to the apoptosis (4). Compared with the wild type (Fig. 5a), CD4⫹CD8⫹ thymocytes in ROR␥ null mice skewed to CD4low (Fig. 5b), which were corrected by the wild-type ROR␥tTg (Fig. 5c, upper panel) but not by the ROR␥t-DBDTg (Fig. 5d, upper panel) and the ROR␥tAF2Tg (Fig. 5e, upper panel). In addition, the CD4 single-positive (SP) cell population was significantly reduced in ROR␥ null mice (0.81%; Fig. 5b) compared with the wild-type mice (7.32%; Fig. 5a). Expression of the ROR␥tTg but not the ROR␥t-DBDTg (Fig. 5d, upper panel) and ROR␥t-AF2Tg (Fig. 5e, upper panel) leads to a significant increase in the CD4 SP cells to a level (14.4%; Fig. 5c, upper panel) much higher than that of the wild-type mice, which is consistent with the increased CD3⫹ cell population in peripheral blood (Fig. 3c). The endogenous ROR␥ gene is turned off during the process of differentiation from double-positive (DP) to SP T cells (7, 8). However, the ROR␥t transgene driven by a CD4 promoter was not turned off, as shown by the GFP expression


3805

FIGURE 3. Generation of Tg mice expressing wild-type ROR␥t, ROR␥t-DBD, and ROR␥t-AF2. a, Tg constructs. cDNAs encoding various ROR␥t were cloned between an upstream CD4 promoter and downstream IRES-GFP. NotI is the restriction site used to release the DNA fragment for generating Tg. b, Integration of the transgene into the genome. DNA prepared from the Tg founders was subjected to digestion with restriction enzyme (BamHI) and to Southern blot analyses with ROR␥t cDNA as a probe. A 3-kb band was detected in mice positive (⫹) but not negative (⫺) for transgene. c, Expression of GFP in thymocytes of the Tg mice. Flow cytometric analyses of the expression of CD3 and GFP in wild-type (ROR⫹), ROR␥ null (ROR⫺), or mice that integrate the transgene encoding wild-type ROR␥t (ROR␥tTg), ROR␥t-DBD (ROR␥t-DBDTg), or ROR␥t-AF2 (ROR␥t-AF2Tg). Numbers are the percentage of GFP⫹ cells within the indicated gates. d, Wild-type ROR␥t and ROR␥t-DBD but not ROR␥t-AF2 coimmunoprecipitated with GRIP1 in thymocytes. GRIP1 in thymocyte lysates was immunoprecipitated (IP) by anti-GRIP1-specific Ab or isotype control Ab (C). The IP complexes were then subjected to Western blot analyses with anti-ROR␥t Ab (upper panel). Lower panels indicate the expression levels of ROR␥t and GRIP1 in the lysates used for immunoprecipitation. e, Wild-type ROR␥t and ROR␥t-DBD but not ROR␥t-AF2 coimmunoprecipitated with SRC1 in thymocytes. Similar assays as described in d, but anti-SRC1 instead of GRIP1 Ab was used. Data shown are representative of at least three independent experiments.

in SP T cells both in thymus and peripheral blood (Figs. 3c and 5), which may be responsible for the increased CD4 cells. With the maturation of T cells, expression of surface TCR increases progressively and the SP cells express the highest levels of TCR. Flow cytometric analyses of TCR revealed that overall TCR levels shifted downward in ROR␥ null mice (Fig. 5, f and g). If the thymocytes were grouped into three arbitrary populations according to their surface TCR levels, TCRlow, TCR intermediate

(TCRin), and TCRhigh, the wild-type mice had ⬃30% of TCRlow cells (Fig. 5f), whereas ROR␥t null mice had ⬃60% of TCRlow cells (Fig. 5g). ROR␥tTg reduced the TCRlow population to the wild-type levels in ROR␥t null mice, ⬃30% (Fig. 5h, upper panel), most likely due to the rescued thymocyte survival. Whereas in ROR␥⫺/ROR␥t-DBDTg (Fig. 5i, upper panel) and ROR␥⫺/ROR␥t-AF2Tg (Fig. 5j, upper panel) mice, the TCRlow population remained at the levels similar to that of the ROR␥ null

3806


FIGURE 4. The wild-type ROR␥tTg but not the ROR␥t-DBDTg and ROR␥t-AF2Tg restored the survival of ROR␥⫺/⫺ thymocytes. a– d, Apoptosis assays. Thymocytes obtained from different genotypes of mice were cultured in medium for the indicated times, and then apoptotic cells were detected by annexin V and propidium iodide staining. Percentage of surviving cells was averaged from at least five mice with the error bar denoting SD. a, ROR␥⫺/⫺ thymocytes undergo rapid apoptosis. Apoptosis of the wild-type (ROR⫹) and ROR␥⫺/⫺ thymocytes (ROR⫺) at various times after removal from mice and cultured in medium. b, Wild-type ROR␥tTg restored thymocyte survival. Apoptosis of the GFP⫹ and GFP⫺ thymocytes obtained from the ROR␥ null mice reconstituted with ROR␥tTg (ROR␥⫺/ROR␥tTg). c, ROR␥t-DBDTg failed to restore thymocyte survival. Apoptosis of the GFP⫹ and GFP⫺ thymocytes obtained from the ROR␥ null mice reconstituted with ROR␥t-DBDTg (ROR␥⫺/ROR␥t-DBDTg). d, ROR␥t-AF2Tg failed to restore thymocyte survival. Apoptosis of the GPF⫹ and GFP⫺ thymocytes obtained from the ROR␥ null mice reconstituted with ROR␥t-AF2Tg (ROR␥⫺/ROR␥t-AF2Tg). e, Wild-type ROR␥tTg but not ROR␥t-DBDTg and ROR␥t-AF2Tg restored thymic cellularity of the ROR␥t null mice. Total thymocyte number was averaged from five mice of each genotype. f, Wild-type ROR␥tTg but not ROR␥t-DBDTg and ROR␥t-AF2Tg restored the levels of Bcl-xL and p27kip1 in ROR␥⫺/⫺ thymocytes. Expression of ROR␥t, Bcl-xL, and p27kip1 was detected by Western blot analyses of the GFP⫹ cells sorted from different genotypes of mice. Actin was used as a control for equal loading. Data shown are representative of three independent experiments.

mice. Interestingly, compared with the ROR␥⫺ mice, slightly higher TCR levels were observed in the TCRin population of the ROR␥⫺/ROR␥t-DBDTg and ROR␥⫺/ROR␥t-AF2Tg mice. To further confirm the function of the transgenes, we also compared the GFP⫹ and GFP⫺ cells in the same mouse. In mice expressing the ROR␥tTg, the GFP⫹ cells (Fig. 5, c and h, upper panel) displayed expression patterns of surface CD4, CD8, and TCR similar to that

of the wild-type mice, whereas the GFP⫺ cells (Fig. 5, a and h, lower panel) displayed the patterns similar to that of the ROR␥⫺/⫺ thymocytes. In contrast, both GFP⫹ and GFP⫺ cells obtained from the ROR␥⫺-/ROR␥t-DBDTg (Fig. 5, d and i) and ROR␥t⫺/ROR␥tAF2Tg (Fig. 5, e and j) mice had surface CD4, CD8, and TCR expression patterns similar to that of the ROR␥t null mice, clearly demonstrating that the wild-type ROR␥t but not ROR␥t-DBD and


3807

FIGURE 5. Flow cytometric analyses of the surface expression of CD4 and CD8 on thymocytes obtained from ROR␥⫹ (a), ROR␥⫺ (b), ROR␥⫺/ ROR␥tTg (c), ROR␥⫺/ROR␥t-DBDTg (d), and ROR␥⫺/ROR␥t-AF2Tg (e) when gated on GFP⫹ cells (upper panels) or GFP⫺ cells (lower panels). Numbers are the percentage of CD4 SP cells within the indicated gates. Surface expression of TCR on thymocytes obtained from ROR␥⫹ (f), ROR␥⫺ (g), ROR␥⫺/ROR␥tTg (h), ROR␥⫺/ROR␥t-DBDTg (i), and ROR␥⫺/ROR␥t-AF2Tg (j) when gated on GFP⫹ cells (upper panels) or GFP⫺ cells (lower panels). Numbers are the percentage of TCRhigh, TCRin, and TCRlow cells within the gated area.

ROR␥t-AF2 is capable of restoring the phenotype of ROR␥⫺/⫺ thymocytes to that seen in the wild-type mice. The above results suggest that the transgene was properly targeted and expressed in thymocytes, and the wild-type ROR␥t but not the ROR␥t-DBD and ROR␥t-AF2 functioned to replace the endogenous ROR␥ gene for thymocyte maturation. Furthermore, since expression of the wild-type ROR␥t in T cell compartments was sufficient to restore thymocyte survival, apoptosis of the ROR␥⫺/⫺ thymocytes is most likely due to the absence of ROR␥t expression in thymus but not due to defects in other somatic tissues.

Discussion SRCs are believed to play a critical role in establishing the specific gene expression pattern required for development and differentiation (14). However, it has been a challenge to demonstrate the in vivo function of SRCs due to overlapping functions among its members (14, 22). In vitro transfection analysis of SRC-mediated transcription often indicates overlapping function among SRCs. Such overlapping functions are indicated by the fact that a single nuclear receptor can interact with multiple members of SRCs (14, 29 –31). Therefore, a gene knockout approach is unlikely to be successful in elucidating the entire function of an individual coactivator in vivo due to the functional redundancy. Indeed, mice deficient in either SRC1 or GRIP1 displayed relatively minor defects (14, 21–23), whereas knockout of both SRC1 and GRIP1 is embryonic lethal (24). Both in transfected cells and in thymocytes, ROR␥t was coimmunoprecipitated with SRC1 as well as GRIP1 (Figs. 1 and 3). ROR␥t is thus capable of interacting with more than one member of the SRCs. Moreover, both SRC1 and GRIP1 potentiate ROR␥t-mediated transcriptional responses (Fig. 2). Such redundancy could explain why knockout of SRC1 or GRIP1 did not lead to defects in thymocyte survival (14). In this study, we created a point mutation in the AF2 domain of ROR␥t that disrupted the LXXLL motif-based interactions. ROR␥ null mice dis-

play thymocyte apoptosis and are fertile, which allows us to determine the function of AF2 domain-mediated recruitment of coactivators in an in vivo model. We clearly demonstrated here that ROR␥t-mediated in vivo thymocyte function requires the recruitment of LXXLL motif-containing coactivators. Nuclear receptors, although capable of interacting with multiple members of SRC, exhibit preferential binding to different SRC family members. For example, androgen receptor binds preferentially to GRIP1 over SRC1 (32). In addition, LXXLL motifs within a given coactivator exhibit binding specificity. The first three LXXLL motifs of SRC1 preferentially mediate the binding to estrogen receptor and progesterone receptor, while the fourth LXXLL motif strongly binds the androgen receptor and glucocorticoid receptor. Using peptides containing various LXXLL motifs, Kurebayashi et al. (33) showed that the binding selectivity of LXXLL peptides for ROR␥ differs from those for other receptors such as estrogen receptor, RAR, glucocorticoid receptor, and vitamin D receptor (33). Our results do not exclude the possibility that ROR␥t preferentially binds to one coactivator over the others. We did observe that GRIP1 is more potent than SRC1 in stimulating ROR␥t-mediated transcription. It is thus very possible that members of the SRCs contribute differentially in the regulation of ROR␥t-mediated thymocyte survival in vivo. Thymocytes have developed a mechanism to switch on Bcl-xL specifically at the DP stage to extend their survival. During transition from the double-negative to DP stage, Bcl-xL is significantly up-regulated. Consistently, mice deficient in Bcl-xL exhibit apoptosis specifically at the DP stage (34). Interestingly, Bcl-xL is switched off again during transition from the DP to SP stage. Our previous results demonstrated that in the absence of ROR␥t, the levels of Bcl-xL mRNA and protein failed to be up-regulated in DP cells, which explains the observed apoptosis (4). ROR␥t is thus critical for inducing Bcl-xL as cells differentiate into the DP stage.

3808


Because the AF2 mutant failed to restore Bcl-xL to wild-type levels (Fig. 4f), recruitment of SRCs by ROR␥t is an essential step in switching on the Bcl-xL gene at the DP stage from its relatively inactive state at the earlier double-negative stage. Nuclear receptors are able to regulate cellular function by mechanisms other than DNA-binding-dependent transcriptional regulation (35, 36). For example, Nur77 induces T cell apoptosis by interfering with Bcl-2 function via direct interaction (37). Nur77-mediated apoptosis is thus independent of its DNA-binding activity. Our results with the DNA-binding mutant of ROR␥t exclude such a possibility. Therefore, SRCs recruited by ROR␥t likely modify the chromatin structure at the DP stage, allowing up-regulation of Bcl-xL and consequent thymocyte survival. ROR␥t has been shown to inhibit up-regulation of the Fas ligand (FasL) and IL-2 production and thus protects T cell hybridoma from activation-induced cell death (7). In this study, we demonstrate that ROR␥t is a potent inhibitor of NFAT (38). Because transcriptional activation of both FasL and IL-2 requires NFAT activity (26, 39), our results explain why forced expression of ROR␥t in hybridoma prevents FasL-dependent activation-induced cell death. In addition to regulating FasL, NFAT plays a critical role in multiple T cell functions including apoptosis (26, 40 – 42). NFAT4⫺/⫺ or NFATc1⫺/⫺ mice have reduced thymic cellularity due to increased apoptosis, suggesting a positive role for both NFAT isoforms in thymocyte survival (40 – 42). Surprisingly, thymocytes deficient in both NFATc1 and NFATc2 do not display obvious defects in thymocyte survival (43). A balanced effect among several isoforms of NFAT is thus critical for maintaining optimum thymocyte survival. Although it is not clear whether ROR␥t inhibits NFAT activity in vivo in thymocytes, our results clearly showed that ROR␥t has an ability to inhibit NFAT activity in in vitro transfection analysis. We found that ROR␥t-AF2, which is still capable of inhibiting NFAT activity, failed to support thymocyte survival. Thus, in contrast to AF2 domain-mediated recruitment of coactivators, the ability to inhibit NFAT does not appear to be critical for ROR␥t-regulated thymocyte survival. However, we do not exclude the possibility that both inhibition of NFAT and recruitment of coactivators are required for thymocyte survival. A mutation that specifically disrupts the inhibitory effect on NFAT, but not the interaction with coactivators, will be useful in examining such a possibility. Knockout of the ROR␥ gene disrupted its expression in thymocytes as well as in other somatic tissues (4, 6, 7). It is therefore possible that the observed apoptosis is a secondary effect resulting from the defects in other somatic tissues. We show here that forced expression of the wild-type ROR␥t in the T cell compartments was sufficient to rescue the defects observed in ROR␥⫺/⫺ thymocytes, including apoptosis and down-regulation of Bcl-xL, CD4, p27kip1, and TCR, strongly suggesting that ROR␥t functions autonomously in thymocytes. In addition, ROR␥t-regulated thymocyte survival is likely independent of its function in lymph node genesis, since the transgene encoding the wild-type ROR␥t did not rescue the lymph node genesis in ROR␥ null mice (data not shown). Therefore, expression of ROR␥t in cells other than thymocytes is required for lymph node development. In agreement, Eberl et al. (44) have shown that expression of ROR␥t in fetal lymphoid inducer cells is critical for lymph node genesis. However, the CD4 promoter may not be active in lymphoid inducer cells, because Cre activity could not be detected in lymphoid inducer cells from CD4-Cre Tg mice (9). It is also possible that cells other than lymphoid inducers are required for lymph node genesis. ROR␥t thus regulates thymocyte survival and lymph node development differently. We show here that recruitment of SRCs by ROR␥t is required to regulate thymocyte survival, presumably by modifying gene ex-

pression. ROR␥t target genes are yet to be identified. Bcl-xL is transcriptionally down-regulated in ROR␥ null mice (4, 5). However, ROR␥t does not appear to regulate Bcl-xL transcription directly, because ROR␥t has no effect on a reporter driven by a Bcl-xL promoter (data not shown). Identification of ROR␥t target genes will facilitate understanding of the mechanisms responsible for ROR␥t-regulated functions.

Acknowledgments We thank Dr. Michael Stallcup for providing expression plasmids of wildtype and mutant GRIP1; Dr. Jeff Staudinger for expression plasmids encoding SRC1, c-RIP140, and vp16-c-RIP140; Dr. Anthony Means for ROR␥t reporter; Dr. Dan Littman for CD4 promoter constructs used in generation of Tg mice; and Drs. Prasad Kanteti and Bellur Prabhakar for critically reading this manuscript and helpful discussion.

Disclosures The authors have no financial conflict of interest.

References 1. Guo, J., A. Hawwari, H. Li, Z. Sun, S. K. Mahanta, D. R. Littman, M. S. Krangel, and Y. W. He. 2002. Regulation of the TCR␣ repertoire by the survival window of CD4⫹CD8⫹ thymocytes. Nat. Immunol. 3: 469 – 476. 2. Gronemeyer, H., and V. Laudet. 1995. Transcription factors 3: nuclear receptors. Protein Profile 2: 1173–1308. 3. Motoyama, N., F. Wang, R.. A. Roth, H. Sawa, K. Nakayama, K. Nakyama, I. Negishi, S. Senju, Q. Zhang, S. Fujii, et al. 1995. Massive cell death of immature hematopoietic cells and neurons in Bcl-x-deficient mice. Science 267: 1506 –1510. 4. Sun, Z., D. Unutmaz, Y. R. Zou, M. J. Sunshine, A. Pierani, S. Brenner-Morton, R. E. Mebius, and D. R. Littman. 2000. Requirement for ROR␥ in thymocyte survival and lymphoid organ development. Science 288: 2369 –2373. 5. Kurebayashi, S., E. Ueda, M. Sakaue, D. D. Patel, A. Medvedev, F. Zhang, and A. M. Jetten. 2000. Retinoid-related orphan receptor ␥ (ROR␥) is essential for lymphoid organogenesis and controls apoptosis during thymopoiesis. Proc. Natl. Acad. Sci. USA 97: 10132–10137. 6. Ortiz, M. A., F. J. Piedrafita, M. Pfahl, and R. Maki. 1995. TOR: a new orphan receptor expressed in the thymus that can modulate retinoid and thyroid hormone signals. Mol. Endocrinol. 9: 1679 –1691. 7. He, Y. W., M. L. Deftos, E. W. Ojala, and M. J. Bevan. 1998. ROR␥ t, a novel isoform of an orphan receptor, negatively regulates Fas ligand expression and IL-2 production in T cells. Immunity 9: 797– 806. 8. Eberl, G., S. Marmon, M. J. Sunshine, P. D. Rennert, Y. Choi, and D. R. Littman. 2004. An essential function for the nuclear receptor ROR␥t in the generation of fetal lymphoid tissue inducer cells. Nat. Immunol. 5: 64 –73. 9. Eberl, G., and D. R. Littman. 2004. Thymic origin of intestinal ␣␤ T cells revealed by fate mapping of ROR␥t⫹ cells. Science 305: 248 –251. 10. Ansel, K. M., D. U. Lee, and A. Rao. 2003. An epigenetic view of helper T cell differentiation. Nat. Immunol. 4: 616 – 623. 11. Spiegelman, B. M., and R. Heinrich. 2004. Biological control through regulated transcriptional coactivators. Cell 119: 157–167. 12. Rosenfeld, M. G., and C. K. Glass. 2001. Coregulator codes of transcriptional regulation by nuclear receptors. J. Biol. Chem. 276: 36865–36868. 13. Moras, D., and H. Gronemeyer. 1998. The nuclear receptor ligand-binding domain: structure and function. Curr. Opin. Cell Biol. 10: 384 –391. 14. Xu, J., and Q. Li. 2003. Review of the in vivo functions of the p160 steroid receptor coactivator family. Mol. Endocrinol. 17: 1681–1692. 15. Henttu, P. M., E. Kalkhoven, and M. G. Parker. 1997. AF-2 activity and recruitment of steroid receptor coactivator 1 to the estrogen receptor depend on a lysine residue conserved in nuclear receptors. Mol. Cell. Biol. 17: 1832–1839. 16. Feng, W., R. C. Ribeiro, R. L. Wagner, H. Nguyen, J. W. Apriletti, R. J. Fletterick, J. D. Baxter, P. J. Kushner, and B. L. West. 1998. Hormonedependent coactivator binding to a hydrophobic cleft on nuclear receptors. Science 280: 1747–1749. 17. Nolte, R. T., G. B. Wisely, S. Westin, J. E. Cobb, M. H. Lambert, R. Kurokawa, M. G. Rosenfeld, T. M. Willson, C. K. Glass, and M. V. Milburn. 1998. Ligand binding and co-activator assembly of the peroxisome proliferator-activated receptor-␥. Nature 395: 137–143. 18. Danielian, P. S., R. White, J. A. Lees, and M. G. Parker. 1992. Identification of a conserved region required for hormone dependent transcriptional activation by steroid hormone receptors. EMBO J. 11: 1025–1033. 19. McKenna, N. J., J. Xu, Z. Nawaz, S. Y. Tsai, M. J. Tsai, and B. W. O’Malley. 1999. Nuclear receptor coactivators: multiple enzymes, multiple complexes, multiple functions. J. Steroid Biochem. Mol. Biol. 69: 3–5. 20. Glass, C. K., and M. G. Rosenfeld. 2000. The coregulator exchange in transcriptional functions of nuclear receptors. Genes Dev. 14: 121–141. 21. Xu, J., Y. Qiu, F. J. DeMayo, S. Y. Tsai, M. J. Tsai, and B. W. O’Malley. 1998. Partial hormone resistance in mice with disruption of the steroid receptor coactivator-1 (SRC-1) gene. Science 279: 1922–1925. 22. Xu, J., L. Liao, G. Ning, H. Yoshida-Komiya, C. Deng, and B. W. O’Malley. 2000. The steroid receptor coactivator SRC-3 (p/CIP/RAC3/AIB1/ACTR/


23.

24.

25.

26. 27.

28. 29.

30.

31.

32.

33.

TRAM-1) is required for normal growth, puberty, female reproductive function, and mammary gland development. Proc. Natl. Acad. Sci. USA 97: 6379 – 6384. Gehin, M., M. Mark, C. Dennefeld, A. Dierich, H. Gronemeyer, and P. Chambon. 2002. The function of TIF2/GRIP1 in mouse reproduction is distinct from those of SRC-1 and p/CIP. Mol. Cell. Biol. 22: 5923–5937. Mark, M., H. Yoshida-Komiya, M. Gehin, L. Liao, M. J. Tsai, B. W. O’Malley, P. Chambon, and J. Xu. 2004. Partially redundant functions of SRC-1 and TIF2 in postnatal survival and male reproduction. Proc. Natl. Acad. Sci. USA 101: 4453– 4458. Darimont, B. D., R. L. Wagner, J. W. Apriletti, M. R. Stallcup, P. J. Kushner, J. D. Baxter, R. J. Fletterick, and K. R. Yamamoto. 1998. Structure and specificity of nuclear receptor-coactivator interactions. Genes Dev. 12: 3343–3356. Rao, A., C. Luo, and P. G. Hogan. 1997. Transcription factors of the NFAT family: regulation and function. Annu. Rev. Immunol. 15: 707–747. Sawada, S., J. D. Scarborough, N. Killeen, and D. R. Littman. 1994. A lineagespecific transcriptional silencer regulates CD4 gene expression during T lymphocyte development. Cell 77: 917–929. Adlam, M., and G. Siu. 2003. Hierarchical interactions control CD4184 gene expression during thymocyte development. Immunity 18: 173. Lee, S. K., H. J. Kim, S. Y. Na, T. S. Kim, H. S. Choi, S. Y. Im, and J. W. Lee. 1998. Steroid receptor coactivator-1 coactivates activating protein-1-mediated transactivations through interaction with the c-Jun and c-Fos subunits. J. Biol. Chem. 273: 16651–16654. Na, S. Y., S. K. Lee, S. J. Han, H. S. Choi, S. Y. Im, and J. W. Lee. 1998. Steroid receptor coactivator-1 interacts with the p50 subunit and coactivates nuclear factor ␬B-mediated transactivations. J. Biol. Chem. 273: 10831–10834. Torchia, J., D. W. Rose, J. Inostroza, Y. Kamei, S. Westin, C. K. Glass, and M. G. Rosenfeld. 1997. The transcriptional co-activator p/CIP binds CBP and mediates nuclear-receptor function. Nature 387: 677– 684. Ding, X. F., C. M. Anderson, H. Ma, H. Hong, R. M. Uht, P. J. Kushner, and M. R. Stallcup. 1998. Nuclear receptor-binding sites of coactivators glucocorticoid receptor interacting protein 1 (GRIP1) and steroid receptor coactivator 1 (SRC-1): multiple motifs with different binding specificities. Mol. Endocrinol. 12: 302–313. Kurebayashi, S., T. Nakajima, S. C. Kim, C. Y. Chang, D. P. McDonnell, J. P. Renaud, and A. M. Jetten. 2004. Selective LXXLL peptides antagonize transcriptional activation by the retinoid-related orphan receptor ROR␥. Biochem. Biophys. Res. Commun. 315: 919 –927.

3809 34. Ma, A., J. C. Pena, B. Chang, E. Margosian, L. Davidson, F. W. Alt, and C. B. Thompson. 1995. Bcl-x regulates the survival of double-positive thymocytes. Proc. Natl. Acad. Sci. USA 92: 4763– 4767. 35. Li, H., S. K. Kolluri, J. Gu, M. I. Dawson, X. Cao, P. D. Hobbs, B. Lin, G. Chen, J. Lu, F. Lin, et al. 2000. Cytochrome c release and apoptosis induced by mitochondrial targeting of nuclear orphan receptor TR3. Science 289: 1159 –1164. 36. Reichardt, H. M., K. H. Kaestner, J. Tuckermann, O. Kretz, O. Wessely, R. Bock, P. Gass, W. Schmid, P. Herrlich, P. Angel, and G. Schutz. 1998. DNA binding of the glucocorticoid receptor is not essential for survival. Cell 93: 531–541. 37. Lin, B., S. K. Kolluri, F. Lin, W. Liu, Y. H. Han, X. Cao, M. I. Dawson, J. C. Reed, and X. K. Zhang. 2004. Conversion of Bcl-2 from protector to killer by interaction with nuclear orphan receptor Nur77/TR3. Cell 116: 527–540. 38. Littman, D. R., Z. Sun, D. Unutmaz, M. J. Sunshine, H. T. Petrie, and Y. R. Zou. 1999. Role of the nuclear hormone receptor ROR␥ in transcriptional regulation, thymocyte survival, and lymphoid organogenesis. Cold Spring Harb. Symp. Quant. Biol. 64: 373–381. 39. Latinis, K. M., L. A. Norian, S. L. Eliason, and G. A. Koretzky. 1997. Two NFAT transcription factor binding sites participate in the regulation of CD95 (Fas) ligand expression in activated human T cells. J. Biol. Chem. 272: 31427–31434. 40. Oukka, M., I. C. Ho, F. C. de la Brousse, T. Hoey, M. J. Grusby, and L. H. Glimcher. 1998. The transcription factor NFAT4 is involved in the generation and survival of T cells. Immunity 9: 295–299. 41. Yoshida, H., H. Nishina, H. Takimoto, L. E. Marengere, A. C. Wakeham, D. Bouchard, Y. Y. Kong, T. Ohteki, A. Shahinian, M. Bachmann, et al. 1998. The transcription factor NF-ATc1 regulates lymphocyte proliferation and Th2 cytokine production. Immunity 8: 115–124. 42. Ranger, A. M., M. R. Hodge, E. M. Gravallese, M. Oukka, L. Davidson, F. W. Alt, F. C. de la Brousse, T. Hoey, M. Grusby, and L. H. Glimcher. 1998. Delayed lymphoid repopulation with defects in IL-4-driven responses produced by inactivation of NF-ATc. Immunity 8: 125–134. 43. Peng, S. L., A. J. Gerth, A. M. Ranger, and L. H. Glimcher. 2001. NFATc1 and NFATc2 together control both T and B cell activation and differentiation. Immunity 14: 13–20. 44. Eberl, G., and D. R. Littman. 2003. The role of the nuclear hormone receptor ROR␥t in the development of lymph nodes and Peyer’s patches. Immunol. Rev. 195: 81–90.