Protein 14-3-3sigma interacts with and favors cytoplasmic subcellular localization of the glucocorticoid receptor, acting as a negative regulator of the glucocorticoid signaling pathway.

The glucocorticoid receptor (GR) alpha interacts with the highly conserved 14-3-3 family proteins. The latter bind phosphorylated serine/threonine residues of "partner" molecules and influence many signal transduction events by altering their subcellular localization and/or protecting them from proteolysis. To examine the physiologic role of 14-3-3 on the glucocorticoid-signaling pathway, we studied the nucleocytoplasmic shuttling and transactivation properties of GRalpha in a cell line replete with or devoid of 14-3-3sigma. We found that endogenous 14-3-3sigma helped localize green fluorescent protein-fused GRalpha in the cytoplasm in the absence of ligand and potentiated its nuclear export after ligand withdrawal. 14-3-3sigma also suppressed the transcriptional activity of GRalpha on a glucocorticoid-responsive promoter. Disruption of the classic nuclear export signal of 14-3-3sigma inactivated its ability to influence the nucleocytoplasmic trafficking and transactivation activity of GRalpha, whereas introduction of a mutation inactivating the binding activity of 14-3-3sigma to some of its partner proteins did not. 14-3-3sigma bound the ligand-binding domain of GRalpha through its COOH-terminal portion, in a partially ligand-dependent fashion, while it did not interact with "ligand-binding domain" of GRbeta at all. These results suggest that 14-3-3sigma functions as a negative regulator in the glucocorticoid signaling pathway, possibly by shifting the subcellular localization/circulation of this receptor toward the cytoplasm through its nuclear export signal. Since 14-3-3 proteins play significant roles in numerous cellular activities, such as cell cycle progression, growth, differentiation, and apoptosis, these actions might indirectly influence the transcriptional activity of GRalpha. Conversely, through its 14-3-3 protein interactions, GRalpha may influence these processes.

The glucocorticoid receptor (GR) 1 belongs to the superfamily of steroid/thyroid/retinoic acid receptor proteins and mediates the diverse and pivotal actions of glucocorticoids in the maintenance of resting and stress homeostasis (1). The human GR consists of two highly homologous isoforms, ␣ and ␤, produced by alternative use of exon 9 ␣ or ␤ of the GR gene. GR␣ represents the classic glucocorticoid receptor, which binds glucocorticoids and mediates almost all known glucocorticoid effects, while GR␤ does not bind ligand, has dominant negative activity upon GR␣ and unclear physiologic role(s) (2). GR␣, in the ligand-free condition, resides primarily in the cytoplasm in the form of a hetero-oligomer with several heat shock proteins (hsps) and related molecules (2). Upon hormone binding, GR␣ undergoes a conformational change, dissociates from the hsps and translocates into the nucleus depending on nuclear translocation signals 1 and 2 (3)(4)(5). Nuclear translocation signal 1 catalyzes rapid transport of the GR through the nuclear pore, employing the importin-mediated pathway, while nuclear translocation signal 2 contributes to a slower traffic via as yet unknown mechanisms (6).
After entering the nucleus, GR␣ binds as a homodimer to specific DNA enhancer elements, the glucocorticoid response element (GREs) in the promoter regions of glucocorticoid target genes (2). Promoter-bound GR␣, via its two transactivational domains, attracts histone acetyltranferase co-activators and chromatin remodeling complexes, which help transmit signals from the activated ligand-bound GR␣ to the transcription initiation complex (7). After dissociating from DNA, GR␣ is exported into the cytoplasm, becoming again fully competent for ligand binding and signal transmission (8).
Several mechanisms regulate the nuclear export of GR␣. The CRM1/exportin and the classic nuclear export signal (NES)mediated nuclear export machineries were postulated to be involved in GR␣ nuclear export, based on evidence that leptomycin B, an inhibitor of these systems, abrogated the rapid nuclear to cytoplasmic translocation and cytoplasmic retention of GR␣; however, no classic NES(s) are evident in the GR␣ molecule (6,9). The calcium-calreticulin-mediated, classic NES-independent nuclear export system, on the other hand, was also reported to be involved in the nuclear export and cytoplasmic retention of GR␣ (10,11). Reassembly of the GR␣ in the heterocomplex with hsps may not be sufficient to relocate GR␣ in the cytoplasm, since such complexes are also observed in the nucleus both before and after withdrawal of the ligand (8). All three main domains of GR␣, i.e. the amino-terminal, DNA-binding (DBD), and ligand-binding (LBD) domains, seem to be involved in nuclear export of this molecule. A serine residue at position 226 of GR␣ located in the amino-terminal domain is necessary for phosphorylation by the c-Jun NH 2terminal kinase to facilitate the nuclear export of the GR␣, while a 67-amino acid region in the DBD is sufficient to support calreticulin-mediated nuclear export (9,12). In addition, we previously reported that removal of the LBD from GR␣ resulted in constitutive localization of this peptide in the nucleus, indicating that the LBD also contributes to nuclear to cytoplasmic translocation of GR␣ (13).
14-3-3 family proteins constitute a highly conserved family present in high abundance in all eukaryotic cells. They consist of nine isotypes from at least 7 distinct genes in vertebrates and regulate important biologic activities by directly binding to and altering the subcellular localization and/or stability of key molecules in several signaling cascades (14 -16). For example, 14-3-3 proteins regulate the apoptosis pathway by binding BAD and affect the intracellular signaling of several growth factors, including insulin, by interacting with important molecules of their cascades, such as Raf-1, insulin receptor substrate 1 (IRS1), and the forkhead transcription factors (17-21). 14-3-3 proteins also influence other signaling events through physical interaction with members of the protein kinase C family proteins Cbl and polyoma middle-T antigen (15). In addition, 14-3-3 proteins play a critical role in the progression/ arrest of the cell cycle by binding to Cdc25C, Wee1, Cyclin B1, and possibly Chk1 (15,16). Binding of 14-3-3 to Cdc25C segregates the latter into the cytoplasm and eliminates its phosphatase activity from the nucleus, thus inhibiting cells from progressing through the G 2 /M check-point (22)(23)(24).
14-3-3 proteins bind the phosphorylated serine or threonine residues of their partner proteins located within a specific amino acid sequence, RSXpSXP, identified as a "high affinity 14-3-3-binding motif" (25). They contain nine ␣-helical structures and form a homo-or heterodimer through their NH 2terminal portions (25)(26)(27). Their central third to fifth ␣-helices create a binding pocket for a phosphorylated serine/threonine residue, and the C-terminal seventh to ninth helices determine the specificity to target peptide motifs (25,26). 14-3-3 proteins contain one classic NES in their ninth helix, which helps localize 14-3-3/partner protein complexes in the cytoplasm (26,28).
Recent research indicated that GR␣ formed complexes with 14-3-3 proteins and Raf-1 (29). Although an early study reported that GR␣ LBD interacted with 14-3-3 in a ligand-dependent fashion in a yeast two-hybrid assay, a subsequent report indicated that GR␣ was associated with 14-3-3 proteins, both in the ligand-free and -bound conditions (29,30). To further investigate the functional contribution of 14-3-3 to the biologic activity of GR␣, we examined the subcellular localization and transactivation properties of GR␣ in a cell line used in its wild type 14-3-3 replete and in its mutant type 14-3-3deficient forms (31). We found that endogenous 14-3-3 helps localize ligand-free GR␣ in the cytoplasm and contributes to nuclear export of GR␣ after withdrawal of ligand. In addition, endogenous 14-3-3 suppresses ligand-activated GR␣-induced transactivation of a glucocorticoid-responsive promoter. These results indicate that 14-3-3 functions as a negative regulator of the glucocorticoid signaling pathway by shifting the subcellular circulation of this receptor toward the cytoplasm.
Cell Cultures and Transfections-Human colon cancer-derived HCT116 wild type (WT) and 14-3-3 knock-out (KO) cells were kindly provided by Dr. B. Vogelstein (Johns Hopkins University, Baltimore, MD) (31). These cells are defective in functional GR␣ (data not shown). They were maintained in McCoy's 5A medium supplemented with 10% fetal bovine serum, 100 units/ml of penicillin, and 1 g/ml of streptomycin. They were transfected using Lipofectin TM with 1 g/well of pF25-hGR␣ and/or 0.3 g/well of 14-3-3-expressing plasmids for the study of GR␣ subcellular localization, as described previously (32). For reporter assays, 0.5 g/well of pRShGR␣ and 0.3 g/well of 14-3-3expressing plasmids together with 1.0 g/well of pMMTV-Luc and 0.3 g/well of pSV40-␤-Gal were used.
Detection of Subcellular Localization of GFP-fused GR␣ and 14-3-3-Cells were plated on 25-mm dishes and were transfected as described above. 24 h after transfection, the medium was replaced with McCoy's 5A medium containing 10% charcoal/dextran-treated fetal bovine serum with antibiotics. 48 h after transfection, the cells were analyzed with an inverted fluorescence microscope (Leica DM IRB, Wetzlar, Germany) as described previously (13,33). 12-Bit black-andwhite images were captured using a digital CCD camera (Hamamatsu Photonics K.K., Hamamatsu, Japan). Image analysis and presentation was performed using the Openlab software (Improvision, Boston, MA). To examine the subcellular distribution of GFP-GR␣, numbers of cells exhibiting five different distribution patterns from complete cytoplasmic distribution (C) to nuclear localization (N) were counted and percentages of each fraction to the total transfected cell number were calculated. For the experiments examining the nuclear export of GFP-GR␣, cells were exposed to 10 Ϫ6 M dexamethasone 48 h after the transfection. After culturing for 1 h, the cells were washed with PBS two times and placed in McCoy's 5A medium containing 10% charcoal/ dextran-treated fetal bovine serum and antibiotics. Eight hours after replacement of the medium, numbers of cells that contained GFP-GR␣ mainly in the cytoplasm (subgroups corresponding to "C" and "N Ͻ C") were counted, and the nuclear export of GR␣ was expressed as percentages of these numbers to the total transfected cell number.
Reporter Assays-Cells were plated on six-well plates and transfected as described above. 24 h after transfection, the cells were exposed to the indicated amounts of dexamethasone. 48 h after transfection, the cell lysis buffer (Promega) was placed in each well, and the resulting cell lysates were harvested. Luciferase and ␤-galactosidase activities were determined as described previously (32). All measurements of reporter gene activity were conducted in triplicate and all experiments were repeated at least three times.
Statistical Analyses-Statistical analysis was carried out by analysis of variance, followed by Student t test with Bonferroni correction for multiple comparisons.

Endogenous 14-3-3 Helps Non-ligand-bound GR␣ Remain in the Cytoplasm and Facilitates the Nuclear Export of GR␣
after Withdrawal of Dexamethasone-To investigate a role of 14-3-3 family proteins on the subcellular localization and transactivation activity of GR␣, we employed wild type versus mutant HCT116 cells, in which both alleles of the 14-3-3 gene were destroyed by homologous recombination (31). We first tested the subcellular localization of GFP-GR␣ in the absence of ligand. GFP-GR␣ was mainly located in the cytoplasm in the HCT116 WT cells, while substantial amounts of GFP-GR␣ were found in the nucleus in the 14-3-3 KO cells (Fig. 1A). To determine this, we examined multiple cells and graded them into five distribution patterns from complete cytoplasmic dis-tribution (C) to complete nuclear localization (N) (Fig. 1B). Using this analysis, HCT116 WT cells had unliganded GFP-GR␣ mainly in the cytoplasm, while KO cells had more in the nucleus, indicating that endogenous 14-3-3 retained GFP-GR␣ in the cytoplasm in the absence of ligand. We also tested nuclear translocation of GFP-GR␣ in WT cells and KO cells in which 14-3-3 was supplemented by the transfected expressing plasmid. GFP-GR␣ entered the nucleus in 10 -20 min in response to 10 Ϫ6 M dexamethasone in both cell types, indicating that the mechanism supporting the nuclear translocation of GR␣ was intact in these cells (data not shown). We next examined whether 14-3-3 plays a role in the nuclear export of GR␣ after withdrawal of ligand. Eight hours after removing 10 Ϫ6 M dexamethasone from the medium, 32% of HCT116 WT cells had GFP-GR␣ in the cytoplasm, while almost all KO cells still retained it in the nucleus (Fig. 1C). When the wild type 14-3-3 was transfected in KO cells, 22% of the cells had GFP-GR␣ in the cytoplasm, indicating that transfected 14-3-3 helped GR␣ relocate into the cytoplasm after removal of dexamethasone. Endogenous 14-3-3 Suppresses the Transcriptional Activity of GR␣-We then examined the transactivation activity of GR␣ stimulated with increasing concentrations of dexamethasone in WT and KO cells (Fig. 2). GR␣ stimulated the MMTV promoter in response to 10 Ϫ6 M dexamethasone by about 80-and 300-fold in WT cells and KO cells, respectively. The dexamethasone titration curve was shifted upward in the latter cells. Transfection of wild type 14-3-3 partially reversed this change in KO cells. The EC 50 (mean Ϯ S.E.: in mM) was 4.01 Ϯ 0.33 and 6.46 Ϯ 1.04 in WT and KO cells, respectively (p Ͼ 0.10), whereas the B max (mean Ϯ S.E.: ϫ 10 Ϫ2 relative luminescence

FIG. 2. Endogenous 14-3-3 suppresses the transactivation activity of GR␣ by decreasing the Bmax but not the EC 50 of the curve.
WT and KO cells were transfected with pRShGR␣ together with pMMTV-Luc and pSV40-␤-Gal. pCDNA4-14-3-3 was also co-transfected in some KO cells. The cells were then exposed to the indicated concentrations of dexamethasone. Each point represents mean Ϯ S.E. values of the luciferase activity normalized for ␤-galactosidase activity. *, p Ͻ 0.01, comparing to WT cells. unit) was 9.11 Ϯ 0.89 and 28.01 Ϯ 1.19, respectively (p Ͻ 0.001). Transfection of 14-3-3 partially reversed this change in KO cells. EC 50 values were similar in transfected and KO cells (p Ͼ 0.30), while the B max in the transfected cells was significantly lower than in KO cells (p Ͻ 0.01). These results indicated that endogenous 14-3-3 functions as a negative regulator of GR␣-induced transactivation.
The C-terminal Half of 14-3-3 Interacts with the GR␣ LBD in a Yeast Two-hybrid Assay-We next examined the interaction of GR␣ and 14-3-3 in a yeast two-hybrid assay (Fig. 3A). Administration of dexamethasone stimulated the LexA-DBD-GR␣LBD-induced, but not LexA-GR␤LBD-induced, ␤-galactosidase activity by about 3-fold in the EGY48 yeast strain. Co-expression of GAL4-AD fusions of the full-length or the COOH-terminal half of 14-3-3 enhanced ␤-galactosidase activity induced by LexA-DBD-GR␣LBD in a partially dexamethasone-dependent fashion, whereas it did not affect the activity of LexA-DBD-GR␤ LBD. These results indicated that GR␣ LBD interacted with the COOH-terminal half of 14-3-3 in a partially ligand-dependent fashion, while GR␤ LBD did not. No increase of ␤-galactosidase activity was observed in the transformed yeast cells, when they were cultured in galactose-deficient medium that did not support the expression of bait proteins (data not shown). This result indicated that expression of GAL4-AD-fused 14-3-3s did not influence basal promoter activity. We obtained similar results using a plasmid expressing the LexA-DBD fusions of the full-length GR␣ (data not shown).
Our results showed that 14-3-3 interacts with GR␣ in the absence of dexamethasone as well as in its presence. In our system, 14-3-3 interacted with the GR␣ LBD in the absence of dexamethasone and the interaction was enhanced in its presence (Fig. 3B). 14-3-3 fragments, which contained the region from 210 to 240 that corresponds to the ninth ␣-helix, supported the binding to GR␣ LBD.

Destruction of 14-3-3 NES Diminishes the Ability of This Protein to Promote Cytoplasmic Retention/Nuclear Export of GR␣ and Suppression of GR␣ Transactivation-14-3-3
proteins may help translocate their partner proteins into the cytoplasm via their classic NES located in their ninth ␣-helix (16,26). Thus, we examined the contribution of 14-3-3 NES on the cytoplasmic retention of GR␣, by constructing a plasmid expressing a 14-3-3 mutant (14-3-3NES Mut), in which the NES was destroyed by clustered mutations, as a fusion with EGFP (26). The EGFP-fused wild type 14-3-3 was mainly located in the cytoplasm, while a small fraction of this fusion protein was also observed in the nucleus (Fig. 4A). The EGFP-14-3-3NES Mut, on the other hand, was distributed more in the nucleus, indicating that the introduced mutations inactivated the NES. Co-expression of 14-3-3NES Mut did not change the distribution of unliganded GFP-GR␣, in contrast to the wild type 14-3-3 (Fig. 4B). We next examined the effect of this mutant on the nuclear export of GFP-GR␣ after withdrawal of dexamethasone. Supplementation of the wild type 14-3-3 brought GFP-GR␣ into the cytoplasm in 23% of KO cells, while expression of 14-3-3NES Mut did not change the nuclear export of this receptor (Fig. 4C). In a functional reporter assay, 14-3-3NES Mut did not suppress GR␣-induced transactivation of the MMTV promoter in a dexamethasone titration curve (Fig. 4D). These results suggest that 14-3-3 retained the ligand-free GFP-GR␣ in the cytoplasm and helped it through its NES to redistribute in the cytoplasm after the withdrawal of dexamethasone. Since the destruction of NES also abolished the suppressive effect of 14-3-3 on GR␣ transactivation, it is possible that 14-3-3 suppressed GR␣-induced transcriptional activity by segregating GR␣ away from the nucleus.
The Phosphopeptide Binding Activity of 14-3-3 May Not Be Necessary for Cytoplasmic Retention of GFP-GR␣ and Suppression of GR␣-dependent Transactivation-Since a previous publication indicated that 14-3-3 forms a complex with GR␣ together with its partner protein Raf-1, we employed a 14-3-3E182K mutant, to examine whether 14-3-3 partner proteins might contribute to the action of 14-3-3 on the activity of GR␣ (29). This mutation corresponds to the replacement of a glutamic acid at 180 by a lysine in Drosophila 14-3-3⑀ that inactivates the binding activity of this protein to Raf-1 because of destruction of the phosphopeptide-binding pocket (26). 14-3-3E182K was distributed mainly in the cytoplasm similarly to the wild type 14-3-3 (Fig. 5A). 14-3-3E182K preserved property of the wild type 14-3-3 on the subcellular distribution and transactivation activity of GR␣ (Fig. 5, B and C), suggesting that association of 14-3-3 to partner proteins, such as Raf-1, may not be necessary for its effect on these GR␣ activities. DISCUSSION GR␣ forms heterocomplexes with 14-3-3 and its partner protein Raf-1 in the cytoplasmic fraction of adrenalectomized rat liver shown by immunoaffinity chromatography and co-immunoprecipitation experiments (29). Although these complexes were observed in the absence of ligand, the interaction of constituent molecules was further strengthened by the presence of the hormone. Association of GR␣ with one of the 14-3-3 proteins, 14-3-3, was also found in a yeast two-hybrid screening, using the GR␣ LBD as a bait (30). In agreement with the above reports, we examined the functional effect of 14-3-3 on the activity of GR␣ by employing wild type versus mutant HCT116 cells, in which both alleles of the 14-3-3 gene were destroyed FIG. 3. Interaction of 14-3-3 and GR isoforms ␣ and ␤ in a yeast two-hybrid assay. A, 14-3-3 interacts with GR␣ LBD but not with GR␤ LBD via its COOH-terminal portion. Plasmids expressing the indicated bait and prey molecules were co-transfected with p8OP-LacZ in EGY48 strain yeast cells, and their binding activity was tested in the absence or presence of 10 Ϫ5 M dexamethasone. B, 14-3-3 interacts with GR␣ LBD via a portion corresponding to its ninth ␣-helix. Plasmids expressing the indicated bait and prey molecules were co-transfected with p8OP-LacZ in EGY48 strain yeast cells, and their binding activity was tested in the absence or presence of 10 Ϫ5 M dexamethasone. by homologous recombination (31). We found that endogenous 14-3-3 helped GFP-GR␣ remain in the cytoplasm in the absence of dexamethasone and supported the nuclear export of GFP-GR␣ after withdrawal of dexamethasone via its NES. 14-3-3, thus, appeared to function as an "attached" partner NES, a finding that might explain the results of previous reports demonstrating that the nuclear export of GR␣ is sensitive to leptomycin B, even though the GR␣ molecule does not contain a classic NES (6,9). We previously reported that deletion of the GR␣ LBD from GR␣ localized this GR␣ fragment in the nucleus (13). Since 14-3-3 interacts with the LBD, a defect in binding of this domain to 14-3-3 may explain its particular subcellular distribution.
We also demonstrated that endogenous 14-3-3 functioned as a negative regulator of GR␣-induced transactivation. This activity correlated with the ability of 14-3-3 to localize unliganded GFP-GR␣ in the cytoplasm, in the experiment employing the 14-3-3 mutants NES Mut and E182K. Therefore, it is likely that 14-3-3 suppresses GR␣-induced transactivation by shifting intracellular circulation of GR␣ toward the cytoplasm, possibly by reducing the chance of ligand-bound GR␣ to interact with GREs, steroid hormone receptor co-activators, and other related specific or general transcription factors in the nucleus. For instance, GR␣ dynamically interacts with GREs in living cells, binding to and dissociating from them in the order of seconds (35,36). In this situation, the drive created by 14-3-3, which shifts GR␣ toward the cytoplasm, may reduce the probability of GR␣ binding to GREs and, hence, may reduce its transcriptional activity. A similar effect was also observed in a recent report, which showed that c-Jun NH 2 -terminal kinase suppressed GR␣-induced transactivation by phosphorylating serine 216 and facilitating its nuclear export (9). A previous report demonstrated that overexpressed 14-3-3 enhanced GR␣-induced transactivation of a synthetic GREcontaining heterologous promoter in African monkey kidneyderived COS7 cells, while we showed that endogenous 14-3-3 FIG. 4. 14-3-3NES Mut loses the ability to promote the cytoplasmic retention and nuclear export of GFP-GR␣, as well as to suppress GR␣-induced transactivation. A, EGFP-14-3-3NES Mut is more located in the nucleus than the wild type EGFP-14-3-3. KO cells were transfected with EGFP-14-3-3 WT-or NES Mut-expressing plasmids and localization of these EGFP-fused proteins were examined in an inverted fluorescence microscope. B, 14-3-3NES Mut does not retain non-ligand-bound GR␣ in the cytoplasm. KO cells were transfected with pF25-hGR␣ and pCDNA3, pCDNA4-14-3-3 or pCDNA4-14-3-3NES Mut and subcellular localization of non-ligandbound GFP-GR␣ was examined as indicated in Fig. 1B. C, supplementation of 14-3-3NES Mut in KO cells does not facilitate the nuclear export of GFP-GR␣ after withdrawal of dexamethasone. Subcellular localization of GFP-GR␣ 8 h after the withdrawal of dexamethasone was examined in over 100 cells, and percentages of cells having GFP-GR␣ mainly in the cytoplasm (categories C and N Ͻ C) to total transfected cells were calculated. Bars represent mean Ϯ S.E. *, p Ͻ 0.01; n.s., not significant, comparing to KO cells. D, 14-3-3NES Mut fails to suppress the GR␣-induced transactivation of the MMTV promoter. KO cells were co-transfected with plasmids expressing 14-3-3 WT or 14-3-3NES Mut together with pRShGR␣, pMMTV-Luc, and pSV40-␤-Gal. The cells were then exposed to the indicated concentrations of dexamethasone. Each point represents mean Ϯ S.E. of luciferase normalized for ␤-galactosidase activity values.
FIG. 5. 14-3-3E182K retains its cytoplasmic retention of GFP-GR␣ activity and suppresses GR␣-induced transactivation. A, 14-3-3E182K is located mainly in the cytoplasm. KO cells were transfected with EGFP-14-3-3WT-or E182K-expressing plasmids, and localization of EGFP-fused proteins was examined in an inverted fluorescence microscope. B, 14-3-3E182K retains non-ligand-bound GFP-GR␣ in the cytoplasm. KO cells were transfected with pF25-hGR␣ and pCDNA3, pCDNA4-14-3-3, or pCDNA4-14-3-3E182K, and the subcellular localization of non-ligand-bound GFP-GR␣ was examined was examined as indicated in Fig. 1B. C, 14-3-3E182K suppresses GR␣induced transactivation of the MMTV promoter. KO cells were cotransfected with plasmids expressing 14-3-3 WT or 14-3-3E182K together with pRShGR␣, pMMTV-Luc, and pSV40-␤-Gal. The cells were then treated with the indicated concentrations of dexamethasone. Each point represents the mean Ϯ S.E. of luciferase normalized for ␤-galactosidase activity values. *, p Ͻ 0.01, comparing either transfected cell culture to KO nontransfected cells; n.s., not significant, comparing the two transfected lines. suppressed GR␣ transactivation in HCT116 cells (30). This discrepancy may result from differences in the experimental systems employed, including the type of cell lines, the different 14-3-3 isoforms and overexpression versus normal expression versus knock-out of 14-3-3. Indeed, overexpression of a protein may sometimes cause artificial effects (37). We examined 14-3-3 "loss of function" by employing the KO cells and showed that the physiologic activity of 14-3-3 is that of a negative regulator of the glucocorticoid signaling pathway.
We demonstrated that the GR␣ LBD interacts with 14-3-3 as well as in a partially ligand-dependent fashion. A previous report indicated that this domain of GR␣ interacted with 14-3-3 in an absolutely ligand-dependent fashion in the same LexA yeast two-hybrid system (30). Differences in the GR␣ fragments employed or yeast strains used in the assay might have led to the different results. Since a recent report also demonstrated partial ligand-dependent interaction between GR␣ and 14-3-3 in a semiquantitative coimmunoprecipitation assay, it is likely that GR␣ and 14-3-3 proteins associate with each other in the absence of ligand. In contrast to the GR␣ LBD, the GR␤ "LBD" did not interact with 14-3-3 at all. Since GR␤ is constitutively located in the nucleus, the inability of GR␤ to interact with 14-3-3 might, to some extent, contribute to its constitutive nuclear localization (13, 38 -40).
Our results employing 14-3-3E182K suggest that the association of known partner proteins with 14-3-3 may not be necessary for this molecule to influence the subcellular localization of GR␣ and the suppression of its transactivation. However, an E180K mutation in Drosophila 14-3-3⑀, which corresponds to the E182K mutation in human 14-3-3, abolishes the interaction of this protein with Raf-1 and BAD, but preserves that with IRS-1, indicating that the E182K mutation in 14-3-3 might not completely exclude the interaction of this 14-3-3 subtype with all partner proteins (26,41). In agreement with the above-indicated evidence, about half of the 14-3-3-partner proteins use a different phosphopeptide-binding motif to interact with 14-3-3, suggesting that a single surface of 14-3-3 in the phosphopeptide-binding pocket, which the E182K mutation destroys, may not support its association generically with all partner proteins. Further experiments are required to address this issue.
In summary, endogenous 14-3-3 functions as a negative regulator of GR␣-induced transactivation, most likely by shifting the subcellular distribution and circulation of GR␣ toward the cytoplasm. These results indicate that change in the intracellular concentration as well as the subcellular distribution of 14-3-3 may contribute to the altered sensitivity of tissues to glucocorticoids seen in several physiologic and pathologic conditions (1). Since 14-3-3 proteins are involved in a broad array of cellular activities, such as cell cycle progression, growth, differentiation, and apoptosis, these activities might indirectly influence the transcriptional activity of GR␣, by changing the availability of 14-3-3 and/or altering partner proteins associated with 14-3-3. On the other hand, the opposite may be true. Ligand-bound GR␣ may influence these cellular processes by segregating and/or influencing 14-3-3 and partner molecules.