Nuclear hormone receptor antagonism with AP-1 by inhibition of the JNK pathway

Abstract

The activity of c-Jun, the major component of the transcription factor AP-1, is potentiated by amino-terminal phosphorylation on serines 63 and 73 (Ser-63/73). This phosphorylation is mediated by the Jun amino-terminal kinase (JNK) and required to recruit the transcriptional coactivator CREB-binding protein (CBP). AP-1 function is antagonized by activated members of the steroid/thyroid hormone receptor superfamily. Recently, a competition for CBP has been proposed as a mechanism for this antagonism. Here we present evidence that hormone-activated nuclear receptors prevent c-Jun phosphorylation on Ser-63/73 and, consequently, AP-1 activation, by blocking the induction of the JNK signaling cascade. Consistently, nuclear receptors also antagonize other JNK-activated transcription factors such as Elk-1 and ATF-2. Interference with the JNK signaling pathway represents a novel mechanism by which nuclear hormone receptors antagonize AP-1. This mechanism is based on the blockade of the AP-1 activation step, which is a requisite to interact with CBP. In addition to acting directly on gene transcription, regulation of the JNK cascade activity constitutes an alternative mode whereby steroids and retinoids may control cell fate and conduct their pharmacological actions as immunosupressive, anti-inflammatory, and antineoplastic agents.

Figure 1

Dex blocks UV-induced AP-1 activation by preventing c-Jun amino-terminal phosphorylation in HeLa cells. (A) Inhibition of UV-induced AP-1 activation by Dex in cells transiently transfected with the −73Col–CAT reporter (3 μg), along with pSG5–GR (0.2 μg), when indicated. (B) Dex inhibits the increase in AP-1 DNA-binding activity induced by UV stimulation. EMSA was performed with nuclear extracts (5 μg/lane) prepared from cells treated with vehicle (open bars) or Dex (solid bars) and harvested at the indicated time points after UV irradiation. Mobility of free and AP-1-complexed probe is indicated. (C) Dex blocks UV induction of c-jun and c-fos gene expression. Same extracts as in B were analyzed by immunoblotting (20 μg of extract/lane) to monitor the accumulation of c-Jun and c-Fos. (D) Dex prevents c-Jun amino-terminal phosphorylation in response to UV irradiation. c-Jun isolated by immunoprecipitation from metabolically-labeled cells (see Materials and Methods) was subjected to tryptic phosphopeptide mapping. y and x spots correspond to phosphorylation of Ser63/73, respectively. The map origin is marked by an arrowhead.

Figure 1

Dex blocks UV-induced AP-1 activation by preventing c-Jun amino-terminal phosphorylation in HeLa cells. (A) Inhibition of UV-induced AP-1 activation by Dex in cells transiently transfected with the −73Col–CAT reporter (3 μg), along with pSG5–GR (0.2 μg), when indicated. (B) Dex inhibits the increase in AP-1 DNA-binding activity induced by UV stimulation. EMSA was performed with nuclear extracts (5 μg/lane) prepared from cells treated with vehicle (open bars) or Dex (solid bars) and harvested at the indicated time points after UV irradiation. Mobility of free and AP-1-complexed probe is indicated. (C) Dex blocks UV induction of c-jun and c-fos gene expression. Same extracts as in B were analyzed by immunoblotting (20 μg of extract/lane) to monitor the accumulation of c-Jun and c-Fos. (D) Dex prevents c-Jun amino-terminal phosphorylation in response to UV irradiation. c-Jun isolated by immunoprecipitation from metabolically-labeled cells (see Materials and Methods) was subjected to tryptic phosphopeptide mapping. y and x spots correspond to phosphorylation of Ser63/73, respectively. The map origin is marked by an arrowhead.

Figure 2

TNF-α-induced AP-1 activation and c-Jun amino-terminal phosphorylation is inhibited by Dex. (A) Inhibition of TNF-α-induced AP-1 activity by Dex in HeLa cells transiently transfected with the −73Col–CAT reporter (3 μg) along with pSG5–GR (0.2 μg), when indicated. (B) Dex prevents c-Jun amino-terminal phosphorylation in response to TNF-α. Western blot analysis of nuclear extracts (20 μg/lane) prepared from HeLa cells treated with vehicle (open bars) or Dex (solid bars) and collected at the indicated time points after TNF-α stimulation. Additionally, cells were treated with either vehicle (−Act. D) or 1 μg/ml of actinomycin D (+Act. D) 15 min before TNF-α addition. Specific antibodies to detect c-Jun phosphorylated on Ser-63 (P-S63–c-Jun) (top) or total c-Jun (bottom) were subsequently used on the same membrane after stripping.

Figure 2

TNF-α-induced AP-1 activation and c-Jun amino-terminal phosphorylation is inhibited by Dex. (A) Inhibition of TNF-α-induced AP-1 activity by Dex in HeLa cells transiently transfected with the −73Col–CAT reporter (3 μg) along with pSG5–GR (0.2 μg), when indicated. (B) Dex prevents c-Jun amino-terminal phosphorylation in response to TNF-α. Western blot analysis of nuclear extracts (20 μg/lane) prepared from HeLa cells treated with vehicle (open bars) or Dex (solid bars) and collected at the indicated time points after TNF-α stimulation. Additionally, cells were treated with either vehicle (−Act. D) or 1 μg/ml of actinomycin D (+Act. D) 15 min before TNF-α addition. Specific antibodies to detect c-Jun phosphorylated on Ser-63 (P-S63–c-Jun) (top) or total c-Jun (bottom) were subsequently used on the same membrane after stripping.

Figure 3

Hormone-activated nuclear receptors block UV-induced c-Jun amino-terminal-mediated transactivation and c-Jun amino-terminal phosphorylation independently of the c-Jun DNA-binding domain. (A) Activity of the 3xGHF-1–Luc reporter (3 μg) transfected in F9 cells along with either pRSV–c-Jun-GHF-1 or pRSV–c-JunA63/73-GHF-1 (0.5 μg) and pSG5–GR or pSG5–TRα-1 (0.2 μg) as indicated. Cells were treated with Dex or T3 and UV irradiated as indicated. (B) Tryptic–phosphopeptide maps of c-Jun–GHF-1 chimeric protein. HeLa cells were cotransfected with pRSV–c-Jun–GHF-1 (10 μg) and pSG5–GR (2 μg). Twelve hours after precipitate removal, cells were split into four plates, serum-starved, and metabolically labeled (see Materials and Methods). UV irradiation was performed 45 min after hormone or vehicle addition, and cells were harvested 20 min later. y, x, t₁, and t₂ spots corresponding to phosphorylation of Ser-63/73 and Thr-91/-93, respectively, are indicated. The arrowhead marks map origin.

Figure 3

Hormone-activated nuclear receptors block UV-induced c-Jun amino-terminal-mediated transactivation and c-Jun amino-terminal phosphorylation independently of the c-Jun DNA-binding domain. (A) Activity of the 3xGHF-1–Luc reporter (3 μg) transfected in F9 cells along with either pRSV–c-Jun-GHF-1 or pRSV–c-JunA63/73-GHF-1 (0.5 μg) and pSG5–GR or pSG5–TRα-1 (0.2 μg) as indicated. Cells were treated with Dex or T3 and UV irradiated as indicated. (B) Tryptic–phosphopeptide maps of c-Jun–GHF-1 chimeric protein. HeLa cells were cotransfected with pRSV–c-Jun–GHF-1 (10 μg) and pSG5–GR (2 μg). Twelve hours after precipitate removal, cells were split into four plates, serum-starved, and metabolically labeled (see Materials and Methods). UV irradiation was performed 45 min after hormone or vehicle addition, and cells were harvested 20 min later. y, x, t₁, and t₂ spots corresponding to phosphorylation of Ser-63/73 and Thr-91/-93, respectively, are indicated. The arrowhead marks map origin.

Figure 4

(A) Hormone-activated nuclear receptors specifically block UV-induced transcriptional activation mediated by c-Jun amino-terminal phosphorylation. Activity of the 5xGal4–Luc reporter (3 μg) transfected in F9 cells along with the pSG424 derivative construct (2 μg) encoding the Gal4–c-Jun, Gal4–v-Jun, or Gal4–c-JunA63/73 fusion proteins and the expression vector pSG5–GR, pSG5–RARα or pSG5–TRα-1 (0.2 μg), as indicated. Cells were subjected to the indicated treatments. (B) Ligand-activated RAR prevents UV-induced c-Jun amino-terminal phosphorylation. F9 cells were transiently transfected with 20 μg of pRSV–c-Jun and 2 μg of pSG5–RARα. After precipitated removal, cells were split into four plates, serum starved overnight and, when relevant, UV irradiated 45 min after RA or vehicle addition. Nuclear extracts were prepared from cells harvested 25 min after UV stimulation and analyzed by Western blot (80 μg of nuclear extract/lane). Immunodetection of c-Jun phosphorylated on Ser-63 (top) and the total amount of c-Jun (bottom) was subsequently performed on the same filter after stripping. (C) Ligand-activated TRα-1 prevents UV-induced c-Jun amino-terminal phosphorylation. Assays were performed as in B. Cells were transiently transfected with 20 μg of pRSV–c-Jun–GHF-1 and 2 μg of pSG5–TRα-1 and treated when indicated with thyroid hormone (T3). The total amount of c-Jun–GHF-1 was detected using an anti-GHF-1 antibody. Immunodetections of c-Jun–GHF-1 phosphorylated on Ser-63 and the total amount of c-Jun–GHF-1 are showed in the top and bottom panels, respectively.

Figure 5

Hormone-activated nuclear receptors block activation of JNK signaling pathway. (A) Cellular content of JNK is not modified by Dex. JNK Western blot of whole cell extracts (20 μg/lane) from HeLa cells harvested at the indicated time points after Dex addition. (B) Effect of nuclear hormone receptor activation on UV-induced activity of JNK. Serum-starved cells from the cell lines listed aside were pretreated with the indicated hormone (+hormone) or with vehicle (−hormone) and UV stimulated. Extracts were prepared (see Materials and Methods) at the indicated time points after irradiation. For each cell line, the top panel shows JNK activity as measured by solid-phase kinase assay; the bottom panel shows the total JNK amount present in each cell extract (20 μg/lane) as measured by Western blot. (C) Quantitative analysis of the effect of nuclear hormone receptor activation on JNK activity after UV stimulation. Kinetics of JNK activation after UV stimulation in the indicated cell lines pretreated (•) or not (○) with the indicated hormone. JNK activity is represented as fold activation over that present in vehicle-treated cells at the 0 time point. Average results from three independent experiments performed as in B are shown. (D) Effect of Dex on in-gel JNK activity. In-gel kinase assay of cell extracts prepared from HeLa cells pretreated with either vehicle (−Dex) or dexamethasone (+Dex) and harvested at the indicated time points after UV stimulation.

Figure 6

Dex-activated GR inhibits activation of JNK targets other than c-Jun. (A) Dex inhibits UV-induced ATF-2-dependent transactivation in F9 cells. Activity of the 3xjunTRE–TK–CAT reporter transiently cotransfected (3 μg) along with pSG5–GR (0.2 μg). Cells were subjected to the indicated treatments. (B) Dex blocks UV-induced transcriptional activation mediated by the ATF-2 amino-terminal transactivation domain. Activity of the 5xGal4–Luc reporter (3 μg) transfected in F9 cells along with the pSG424-Δ9 derivative construct (0.5 μg) encoding the Gal4–ATF-2(19–96) (C2) or Gal4–ATF-2(19–96)A69/71 (C2–T1T2) fusion proteins and the expression vector pSG5–GR (0.2 μg), as indicated. Cells were subjected to the indicated treatments. (C) Dex inhibits Elk-1mediated transcriptional activation induced by UV irradiation. Activity of the SRE-TK–CAT(Δ4490–183) (3 μg) transiently transfected in HeLa cells subjected to the indicated treatments. (D) UV-induced phosphorylation of Elk-1 is inhibited by Dex. EMSA was performed using nuclear extracts (10 μg/lane) prepared from HeLa cells pretreated with either vehicle (−Dex) or Dex (+Dex), UV irradiated, and harvested at the indicated time points after UV stimulation. Mobility of the uncomplexed (Free) and Elk-1-complexed [(Un.) uninduced; (Ind.) induced] probe is indicated.

Figure 7

(A) Pretreatment with Dex does not activate JNK. In-gel kinase assay of whole cell extracts from HeLa cells were incubated with Dex at the indicated time points. Control extracts from HeLa cells pretreated or not with Dex for 45 min and UV irradiated were included. (B) Dex inhibits JNK activation by interfering with a step downstream of MEKK activation. HeLa cells were transiently cotransfected with pcDNA3–HA–JNK (2 μg) and pSG5–GR (2 μg). Increasing amounts of pCEV29–ΔMEKK and pEBG–SEK-1 (1 μg) were also included as indicated. Twelve hours after precipitate removal, each plate was split into two. Cells were serum-starved for 16 hr, treated with Dex (or vehicle) for 45 min, and harvested. Activity of HA–JNK was determined by immune complex kinase assay using an anti-HA antibody (top). Levels of HA–JNK in these cell extracts were analyzed by Western blot using an anti-HA antibody (bottom). (C) MKP-1 levels are not increased by glucocorticoids. Immunodetection of MKP-1 in whole cell extracts (50 μg/lane) from HeLa cells treated with Dex for the indicated period of time.

Figure 8

(A,B) Receptor–dosage dependence of JNK inhibition. COS-7 cells were transiently transfected with a constant amount of pcDNA3–HA–JNK (2 μg) and increasing amounts of pSG5–GR (total DNA brought up to 7 μg with pSG5). Twelve hours after precipitate removal, each plate was split into four, and after allowing attachment, cells were serum starved for 16 hr. In each set, two plates were treated with Dex and the other two with vehicle. After 45 min, cells were UV irradiated or not, as indicated, and harvested 20 min later. HA–JNK activity was determined by immune complex assay using an anti-HA antibody. Autoradiographs are shown in A (top) together with the analysis of HA–JNK protein levels by Western blotting using an anti-HA antibody (bottom). Substrate phosphorylation was quantitated in an Instant Imager and, for each initial pSG5–GR input, the percentage of JNK inhibition by Dex vs. vehicle treatment was determined. Average values of three independent experiments are shown in B. (C) Short Dex pretreatment prevents JNK activation. Solid-phase JNK assay of extracts from HeLa cells stimulated with TNF-α at the indicated time points after Dex addition. Cells were harvested 15 min after cytokine stimulation. (D) Actinomycin D does not block Dex action. Serum-starved HeLa cells were treated with either ethanol (−Act. D) or 1 μg/ml of actinomycin D (+Act. D) 15 min before vehicle (−Dex) or Dex (+Dex) addition and UV stimulated 60 min later. Cells were harvested at the indicated time points after UV irradiation, and JNK activity (top) and protein levels (bottom) were determined by immune complex assay and immunoblotting, respectively, using an anti-JNK antibody. Though similar amounts of cell extracts were analyzed for JNK activity, autoradiographs corresponding to cells untreated with actinomycin D were exposed 24 times longer than those from actinomycin D-treated cells. In contrast, JNK Western blots were equally exposed. (E) GR mutant LS7 is as efficient as the wild-type receptor in JNK signaling pathway interference. COS-7 cells were transiently transfected with pcDNA3–HA–JNK (2 μg) along with pRC/βact–GR–wt or pRC/βact–GR–LS7 (200 ng). After removal of precipitates, cells were split into four plates, serum-starved overnight, and subjected to the indicated treatments. Cells were harvested 25 min after UV stimulation, and HA–JNK activity (top) and protein levels (bottom) were determined by immune complex assay and Western blotting, respectively, using an anti-HA antibody.

Figure 9

Model for the antagonistic action of hormone-activated nuclear receptors on AP-1 activity through the inhibition of the JNK signal transduction pathway. Activation of gene transcription by either nuclear hormone receptors or AP-1 relays on an activation step: hormone (▴) binding for nuclear receptors and c-Jun phosphorylation on Ser-63/73 (▪) for AP-1. According to our data, upon hormone activation nuclear receptors block the induction of the JNK signaling pathway and, in consequence, c-Jun amino-terminal phosphorylation. By this mechanism, steroid/thyroid hormones and RARs may antagonize AP-1 activity and inhibit expression of AP-1 target genes.