Synergistic Signaling by Corticotropin-Releasing Hormone and Leukemia Inhibitory Factor Bridged by Phosphorylated 3′,5′-Cyclic Adenosine Monophosphate Response Element Binding Protein at the Nur Response Element (NurRE)-Signal Transducers and Activators of Transcription (STAT) Element of the Proopiomelanocortin Promoter

Abstract

Leukemia inhibitory factor (LIF) cooperates with CRH at the pituitary level to induce POMC gene transcription, resulting in activation of the pituitary-adrenal axis. However, the underlying molecular mechanisms remain elusive. Here, we show that the NurRE-signal transducers and activators of transcription (STAT) composite element of the POMC promoter was the predominant target of the LIF-CRH synergy. Whereas NurRE or STAT sites alone conferred synergy, the maximal response was found with the NurRE-STAT reporter, suggesting that direct DNA binding of both transcription factors is required for an optimal synergy. During LIF-CRH stimulation, Nur77 and activated STAT1–3 were bound to the composite element, and the binding of each factor was abolished by appropriate mutations. CREB was also detected in this complex in a stimulation-dependent and DNA binding-independent manner. Nur77 and STAT1–3 bound to the NurRE-STAT site were each sufficient for CREB recruitment. Recombinant CREB directly interacted with recombinant Nur77 or STAT1–3. Moreover, CREB-Nur77 interaction was increased by CREB phosphorylation at Ser-133 and the dominant-negative mutant CREB-M1 efficiently inhibited the synergistic LIF-CRH response. This synergism was also inhibited after transfection of CREB-small interfering RNA. We conclude that both CREB phosphorylation at Ser-133 and level of CREB expression are crucial in LIF-CRH synergism where CREB, without direct DNA binding, could improve the stability of Nur77 and STAT1–3 binding to POMC promoter and facilitate the recruitment of coactivators. This novel intrapituitary signaling mechanism may have more general implications in cross talks between cAMP-protein kinase A and Janus kinase-STAT pathways.

THE CRH, SYNTHESIZED in the hypothalamus and released in the pituitary portal system, stimulates the corticotroph cells of the anterior pituitary via the interaction with its receptor. CRH signaling in corticotroph increases both the transcription of proopiomelanocortin (POMC) gene and the secretion of the mature ACTH peptide (1). Stimulation of POMC gene expression by CRH is also observed in pituitary primary cultures and in the murine corticotroph cell line AtT-20. In all cases, CRH increases cAMP and activates protein kinase A (PKA) that, in turn, modulates L-type Ca²⁺ channels and triggers ACTH secretion (2, 3). Although in corticotroph cells CRH stimulates phosphorylation of CREB [cAMP response element (CRE) binding protein] at Ser-133 by PKA, the POMC promoter does not possess a canonical CRE (4).

The regulation of POMC transcription was shown to be dependent from the early induction of expression of Nur77 and Nurr1, two members of the Nur family of orphan nuclear receptors (5, 6). However, the early induction of POMC transcription by CRH and cAMP is rapid and transient and does not require de novo protein synthesis (7). Indeed, it has been recently reported that CRH, via PKA and/or MAPK, regulates the phosphorylation-dephosphorylation state of Nur77, thus increasing its DNA binding and transcriptional activation properties (8, 9). Two Nur targets have been identified in the POMC promoter, a proximal Nur77 binding response element (NBRE) (−70/−63), which binds Nur77 or Nurr1 monomers, and a distal one, the NurRE, constituted of two everted NBRE-related sites separated by six nucleotides (−404/−382). NurRE binds Nur77 homodimers or Nur77/Nurr1 heterodimers and, in the context of the POMC promoter, this site plays a dominant role, as compared with NBRE, in mediating stimulation by CRH (5, 6, 10).

In addition, it has been recently reported that T-pit, a transcription factor belonging to the Tbox family and cooperating with the homeoprotein Pitx1 for cell-specific expression of the POMC gene (11), is also a mediator of signaling by CRH and could cooperate with Nur77 at the level of POMC promoter (12). CRH also induces the protooncogene c-fos, which, in association with junB, binds the activator protein (AP)-1 site at +41/+47 within the first exon of POMC gene (13, 14). However, the induction of AP-1 components does not explain the early transcriptional activation of POMC gene by CRH, which is protein synthesis independent (7). Another site, −173/−160, called POMC CRH-responsive element, located in the central part of the promoter, was shown to confer strong c-fos-independent stimulation of POMC transcription after CRH treatment (15, 16).

Leukemia inhibitory factor (LIF), a pleiotropic cytokine involved in the inflammatory response, also activates the hypothalamo-pituitary-adrenal axis (17, 18). LIF and its receptor are both expressed in corticotroph cells and in AtT-20 cell line (19, 20) where LIF activates the POMC promoter and ACTH secretion through a Janus kinase/signal transducers and activators of transcription 1–3 (Jak/STAT1–3) pathway (20–22). In this response, STAT3, activated by phosphorylation at Tyr-705-YP (tyrosine phosphorylated), plays a predominant role (22, 23). Recently, a functional STAT1–3 low-affinity binding site was identified in the distal region of the POMC promoter (−387/−379); yet a proximal subregion of the promoter, devoid of STAT1–3 binding properties, also mediates the LIF stimulatory effect (23, 24).

A strong synergism between LIF and CRH on POMC promoter activation was recognized early (16, 21, 23), although the underlying molecular mechanisms remain unclear. These previous studies indicated that both the −173/−160 element and the AP-1 site were involved in this synergism (16, 23). Yet, these data ignored the potential role of the distal region of the POMC gene promoter that is crucial in CRH- and LIF-induced activation.

The promoter subregion −414/−293, starting at the NurRE in the distal region and ending after the Pitx1 site of the central region, was recently defined as responsive to LIF (24). Here, we establish that this subregion also mediates LIF-CRH synergy and that the NurRE-STAT composite element, present in this region, is a significant target of the combined stimulation. This element is composed of two NBRE everted sites, the most 3′ one overlapping in part the unique STAT1–3 site. Although a maximal synergistic response requires the presence of both NurRE and STAT1–3 sites, each site alone is sufficient to induce synergy. Binding of Nur77 and STAT1–3 to NurRE-STAT element, which is increased by CRH and LIF treatment, respectively, is maintained during the combined stimulation. Furthermore, the synergistic response requires recruitment of CREB to DNA-bound Nur77 and STAT1–3. Indeed, CREB phosphorylated at Ser-133 and present in this quaternary complex participates, in a DNA binding-independent manner, to the transcriptional activity of the NurRE-STAT site. Finally, the synergistic response to LIF-CRH combined treatment is inhibited by the dominant-negative mutant of CREB, CREB-M1, where Ser-133 is replaced by Ala (25), or by CREB-small interfering RNA (siRNA), indicating the crucial role of the level of CREB phosphorylation and expression.

RESULTS

Synergistic Response of the POMC Promoter to LIF-CRH Treatment Is Targeted at the Composite NurRE-STAT Element

The kinetics of the response of POMC promoter to CRH, LIF, and LIF-CRH stimulation are shown in Fig. 1. The POMC promoter activities all peaked between 4 and 8 h of each stimulation, albeit at different absolute level. The already described approximately 2-fold increased activity after CRH treatment was low but detectable at 2 h and maximal between 4 and 8 h. By contrast, the LIF-induced activity was of greater amplitude and began as early as 1 h after stimulation (not shown). The time course of the synergistic response to the combined LIF-CRH treatment also started early, displaying a pattern similar to, but higher than LIF response. Because the early effects of CRH and LIF alone do not require de novo protein synthesis (7, 21), these experiments suggest that the early synergistic transcriptional response is also independent from protein synthesis.

Fig. 1.

Kinetics of CRH-, LIF-, and LIF-CRH-Induced POMC Promoter Activities AtT-20 cells were transfected with a reporter plasmid bearing the POMC promoter −480/+63 and treated with LIF (1 nm) and/or CRH (50 nm) for the indicated times. Results, expressed as RLU/βgal activities of control (white bars), LIF (gray bars), CRH (hatched bars), and LIF-CRH (black bars), are the means of three independent experiments ± sem.

Basal activities of different POMC promoter regions and subregions have been already reported (6, 24, 26). The minimal POMC promoter region −34/+63, containing the TATA box and an AP-1-like site into the first exon, has been defined as nonresponsive to CRH and LIF treatment, whereas the −414/−293 subregion was sufficient for LIF responsiveness (6, 24). Within the −414/−293 subregion, the LIF response was dependent on STAT1–3 binding to a low-affinity site partially overlapping the NurRE, which is the major target of the CRH response (5, 6). This subregion also contains all the known DNA elements responsible for the corticotroph-specific POMC expression (11). A schematic representation of the POMC promoter with the subregion and sites relevant to this work is depicted in Fig. 2A. The entire −480/+63 POMC promoter was tested after 6 h stimulation with LIF, CRH, or LIF-CRH, and compared with the minimal POMC promoter or to the subregion −414/−293 fused to the minimal promoter. Although the minimal POMC promoter was not responsive to various treatments, the −414/−293 subregion was equally responsive (2-fold) to CRH and LIF. This subregion was still able to mediate a synergistic response after the combined LIF-CRH treatment (∼9-fold) as compared with the entire promoter (∼9-fold) (Fig. 2B). Thus, we surmised that the NurRE-STAT composite site was an important target for LIF-CRH synergy.

Fig. 2.

The NurRE-STAT Site Is a Target of the Synergistic Response to LIF-CRH A, Schematic representation of POMC promoter regions and the −414/−293 fragment containing the NurRE-STAT composite site. B and C, AtT-20 cells, transfected with the indicated reporter plasmids, were treated with LIF (1 nm) and/or CRH (50 mm) for 6 h. B, Minimal POMC promoter alone or fused downstream to the entire promoter or to the −414/−293 fragment. C, 3xNurRE-STAT, 3xNurRE, 3xSTAT_POMC and 3xSIE (see sequences in Materials and Methods) fused to the minimal POMC promoter. The results are expressed as fold induction over the respective control value (no treatment) arbitrarily taken as 1. All the results are the mean of at least three independent experiments ± sem. The insets in B and C represent, in the same order, the basal activities of each reporter.

The synergistic response to LIF-CRH treat-ment was then tested using constructs containing three copies of the NurRE-STAT composite site (3xNurRE-STAT) or of its mutated variants, the NurRE (3xNurRE) and the low affinity STAT1–3_POMC binding site (3x STAT_POMC) fused to the minimal promoter. The 3xNurRE-STAT reporter (Fig. 2C) was stimulated 4- to 5-fold by LIF or CRH alone and approximately 20-fold by the combined treatment. The 3xNurRE reporter, lacking the entire STAT_POMC binding site, was unresponsive to LIF and synergistically responsive to LIF-CRH treatment, albeit at a level lower than that of the 3xNurRE-STAT, indicating that the binding of activated STAT1–3 to specific DNA sequence is not required for the synergistic effect. The 3xSTAT_POMC reporter mediated a response to LIF-CRH that was more than additive (Fig. 2C), being induced approximately 3-fold by LIF, 1.8-fold by CRH and approximately 8-fold by the combined treatment. The induction of 3xSTAT_POMC by CRH can be explained by the presence of the 3′ half part of the NurRE, which is an NBRE-like sequence. We also analyzed the synergistic response with a reporter bearing three copies of the Sis-inducible element (SIE), a high affinity binding site used as reference for STAT1–3 binding (27, 28). The 3xSIE reporter is also devoid of Nur binding sequences and, differently from the NurRE-STAT and STAT_POMC sites, it contains binding elements for a single class of transcription factors. The SIE construct (Fig. 2 C), highly responsive to LIF (∼10-fold) and not to CRH, yet displayed a synergism of approximately 40-fold, which may depend on its high affinity for STAT1–3. In conclusion, these results demonstrate that NurRE or STAT (SIE) binding sites alone could mediate the synergistic effect of LIF-CRH, even though they were not responsive to one of the single treatment (LIF or CRH, respectively). Moreover, the presence of both NurRE and STAT_POMC site seems necessary to obtain a maximal LIF-CRH synergistic effect in the context of the 3xNurRE-STAT reporter.

To evaluate the contribution of the composite NurRE-STAT site to synergistic response, mutations of its sequence were introduced into the −414/−293 reporter (Fig. 3A). Figure 3B shows that the LIF-CRH synergy of the wild-type construct was almost completely lost in the −414/−293 reporter bearing a mutated NurRE-STAT, indicating the critical role played by this motif in the synergistic response of a partially reconstituted POMC promoter. When introduced into the entire promoter, the same mutations abolished the synergistic response by more than 50% (Fig. 3C), demonstrating the major role of the NurRE-STAT element and, as already mentioned, an additional role of the −293/+63 region, in agreement with the proposed role of other targets (16, 23).

Fig. 3.

Inhibition of the LIF-CRH Synergistic Response by Mutation of the NurRE-STAT Site AtT-20 cells, transfected with the indicated reporter plasmids, were treated with LIF (1 nm) and/or CRH (50 mm) for 6 h. Bars indicate the treatments as in Fig. 2. A, Mutations of the NurRE-STAT site; the NurRE is underlined and the STAT site is in bold. B, −414/−293 wild-type and mutated reporters. C, Wild-type and mutated POMC promoter reporters. The results are expressed as fold induction over the respective control value (no treatment) arbitrarily taken as 1. All the results are the mean of at least three independent experiments ± sem.

Transcription Factors Binding to NurRE-STAT Composite Element

To detect transcription factors bound to the NurRE-STAT composite site and involved in the LIF-CRH synergism, we used a pull-down technique, based on incubation of biotinylated oligonucleotides with nuclear extracts, precipitation of complexes with streptavidine-agarose beads, followed by PAGE and Western blotting of eluted proteins. This approach already allowed us to evidence a low affinity interaction of the STAT_POMC site, overlapping in part the NurRE, with STAT1–3 activated proteins from nuclear extracts of LIF-treated AtT-20 cells, an interaction that was undetectable by EMSA studies (24).

The NurRE-STAT biotinylated oligonucleotide was incubated with nuclear extracts derived from control or CRH- or LIF-stimulated cells. Stimulation by CRH (60 min) enhanced Nur77 binding, and stimulation by LIF (20 min)-induced STAT1–3-YP binding at the NurRE-STAT site (Fig. 4A). The use of a control probe unrelated to Nur77 and STAT1–3 sites confirmed the specificity of Nur77 and STAT1–3 binding to the NurRE-STAT site.

Fig. 4.

Binding of Nur77, STAT1–3, and CREB to the NurRE-STAT Site A, Biotinylated NurRE-STAT or Control (C probe) oligonucleotides were incubated with nuclear extracts from AtT-20 cells stimulated or not (C) by CRH (50 nm, for 60 min) or LIF (1 nm, for 20 min). B and C, Biotinylated Control, NurRE-STAT, NurRE-STATmut, NurREmut-STAT oligonucleotides were incubated with nuclear extract (N.E.) from control or CRH (50 nm)-LIF (1 nm)-treated cells during 30 min. *, Mutated site. D, A 10-fold molar excess of unlabeled CRE, NurRE-STAT (NS), or Control oligonucleotides were added as competitor to nuclear extracts incubated with NurRE-STAT biotinylated oligonucleotide. E, Nuclear extracts were incubated with NurRE, CRE and mutated CRE oligonucleotides. F, His-CREB was incubated with CRE, CRE-mut and NurRE-STAT biotinylated oligonucleotides. Bound Nur77, STAT1–3-YP, CREB, and GR proteins were retrieved by streptavidine-agarose beads, eluted by Laemmli buffer and tested by Western blotting with specific antibodies. N.E. and His-CREB represent 1/10 of the input.

To detect the simultaneous binding of Nur77 and STAT1–3 during LIF-CRH synergy, the NurRE-STAT oligonucleotide and two variants, NurRE-STATmut and NurREmut-STAT (see sequences in Materials and Methods), were incubated with nuclear extracts from control or LIF-CRH-stimulated cells (Fig. 4B). Nur77 binding at the NurRE-STAT site was detected during the costimulation, and this binding was abolished only by the mutation of the NurRE site. Similarly, STAT3-YP binding at the NurRE-STAT site was detected during the combined stimulation, and this binding was abolished only by mutation of the STAT site (Fig. 4B). Similar results were obtained when STAT1-YP binding was tested (data not shown). No binding of Nur77 or STAT1–3 to control probe was found. Thus, despite the partial overlap of sites, both activated Nur77 and STAT1–3 appeared bound to NurRE-STAT, even though we did not find an interaction between Nur77 and STAT1–3 in nuclear extracts.

Because it has been shown that the constitutive active form of CREB (CREB-VP16) activated the POMC promoter and the dominant-negative CREB-M1 inhibited this activation (29), we hypothesized that CRH-activated CREB and/or CREB binding protein (CBP) may participate to stabilization and activation of the DNA-proteins complex composed of NurRE-STAT element, Nur77 and STAT1–3. Using the same pull-down technique, the presence of CREB was indeed detected. Figure 4B (bottom) shows that CREB was retained by NurRE-STAT and each mutated oligonucleotide, whereas it was absent when using the control oligonucleotide, which is also devoid of CRE-like sequences. Moreover, the level of bound CREB was always increased over the control level by the combined stimulation. Whether the level of CREB detected in control conditions with the NurRE-STAT probes (Fig. 4B) represented a technical background or its presence into a complex, remains unclear. In similar experiments, the presence of CBP/p300 was not detected, nor the presence of glucocorticoid receptor (GR), although these proteins were detected in nuclear extracts (see Fig. 4C for GR). Indeed, GR, which binds its DNA consensus sequence after high-salt extraction, was present in the nuclear fraction but not retained by the NurRE-STAT probe.

To exclude the fact that the binding of CREB to NurRE-STAT site was due to nonspecific trapping, a 10-fold excess of unlabeled CRE, NurRE-STAT or control oligonucleotides was added to nuclear extracts incubated with biotinylated NurRE-STAT (Fig. 4D). A CRE or NurRE-STAT excess clearly diminished the CREB binding to NurRE-STAT probe. Moreover, as expected, an excess of NurRE-STAT abolished CREB and STAT3-YP binding. By contrast, an excess of unlabeled CRE suppressed only the binding of CREB and not that of STAT3, whereas an excess of control oligonucleotide did not compete for STAT3 and CREB binding, indicating that CREB binding to NurRE-STAT site was indirect. In addition, using control or LIF-CRH-treated nuclear extracts (Fig. 4E), the binding of CREB to NurRE-STAT oligonucleotide was inducible as reported above (Fig. 4B), but lower than to CRE, whereas CREB binding to mutated CRE was almost absent, suggesting again an indirect binding of CREB to NurRE-STAT.

However, when purified recombinant CREB (His-CREB) was tested directly on CRE, mutated CRE and NurRE-STAT probes, it was bound to CRE, whereas almost no binding was detected to NurRE-STAT or mutated CRE (Fig. 4F). This suggested that CREB binding to NurRE-STAT site, only observed with nuclear extracts, was due to protein-protein interaction. Altogether, these results indicate that, after combined stimulation, increased CREB binding to NurRE-STAT probe takes place via DNA-bound Nur and/or STAT and that the integrity of one of the two sites is sufficient to obtain CREB detection. This suggests an indirect and specific recruitment of CREB at the NurRE-STAT site through a protein-protein interaction with Nur77 and activated STAT1–3.

To further substantiate the possibility of a simultaneous binding of Nur77 and Stat1–3 to the NurRE-STAT element, taking in account the overlap of sites, computer modeling of their binding was performed on the basis of existing crystallographic data (30–32). Figure 5A shows two DNA binding domains of Nur77 monomer bound, in the major groove and in opposite sites of the DNA helix, to each half site of the NurRE sequence. The two subunits are pointing the C-ter region to each other, suggesting that Nur77 dimer could be stabilized by DNA contacts. In Fig. 5B, NurRE, STAT1–3 binding site and their overlap are depicted in blue, red, and violet, respectively. Figure 5C shows two STAT3β subunits in contact with the NurRE-STAT sequence, each subunit interacting with a half site on both DNA strands in the major groove. In the overlapping part ^5′TGCCA^3′ (violet), the major contacts of Nur DNA binding domain concern the second and third bases, and a subunit of STAT is in contact with the second and fourth bases. Then, we can infer that the simultaneous binding of dimeric Nur and STAT1–3 takes place only if one of the transcription factors is in contact with a half site.

Fig. 5.

Computer Modeling of Nur77 and STAT3 Binding to NurRE-STAT Site A, Modeling of NurRE-STAT sequence of the POMC promoter with two DNA binding domains of Nur77. The DNA-protein contacts take place in the major groove and each Nur subunit is located in opposite sites of the DNA helix. B, The NurRE-STAT sequence of the POMC promoter : the NuRE site (red); the STAT1–3 site (blue); the overlap of sites (violet). C, Modeling of the NurRE-STAT sequence of the POMC promoter with two STAT3β subunits. Each subunit makes contact with a half-site on both DNA strands in the major groove.

CREB-NUR77 and CREB-STAT1–3 Interactions

The physical interaction between Nur77 and CREB was investigated using pull-down experiments. Glutathione-S-transferase (GST)-Nur77 recombinant protein bound to glutathione resin (GST-Nur77 resin) was incubated with nuclear extracts from control or LIF-CRH-stimulated cells and levels of bound CREB and Ser-133 phoshorylated CREB (P-CREB) were compared with that of nuclear extracts. Figure 6A confirms that, in nuclear extracts, only P-CREB, and not total CREB, increased after LIF-CRH treatment (lanes N.E.), indicating that the increased CREB phosphorylation, due to CRH signaling, is maintained after the combined treatment (16). Almost no binding of CREB or P-CREB to GST alone was found. LIF-CRH treatment increased, in a parallel fashion, CREB and P-CREB binding to GST-Nur, indicating that P-CREB has a better affinity for Nur77 than CREB and supporting its involvement in the synergy. A similar increased binding of P-CREB was also observed with a Nur77 mutant limited to its N-terminal region (not shown). However, the binding of the His-CREB fusion protein to GST-Nur77 resin, as compared with GST resin (Fig. 6B, top), indicated that the two proteins may interact directly, independently from CREB phosphorylation at Ser-133. The mirror experiment (Fig. 6B, bottom) also supported this conclusion because in vitro-translated ³⁵S-labeled Nur77 was bound to His-CREB resin and not to His-Nter-heat shock protein (HSP) 90 resin.

Fig. 6.

Interactions between Nur77/CREB and STAT1–3/CREB A, Nuclear extracts from AtT-20 cells stimulated or not by LIF (1 nm)-CRH (50 nm) during 30min (N.E. = 1/10 of input) were incubated with either GST resin or GST-Nur77 resin. Ser-133 P-CREB in nuclear extracts or bound to resins was detected by Western blotting and the blot was reprobed using an anti-CREB antibody. B (top), Purified His-CREB (total input) was incubated with either GST resin or GST-Nur77 resin. Bound CREB was detected by Western blot analysis. B (bottom), In vitro-translated ³⁵S-Nur77 was incubated with His-CREB resin or His-HSP90-Nter resin. Bound CREB was detected by Western blot analysis and bound ³⁵S-Nur77 by autoradiography. C, Biotinylated SIE oligonucleotide was incubated with nuclear extracts (N.E. = 1/10 of the input) from AtT-20 cells stimulated or not by LIF-CRH. A 10-fold molar excess of unlabeled SIE oligonucleotide was added as competitor. Bound proteins were visualized by Western blot analysis. D, His-CREB resin or His-Hsp90-Nter resin was incubated with nuclear extracts from AtT-20 cells stimulated by LIF-CRH. Bound proteins were detected by Western blot analysis. E, GST-STAT3 resin was incubated with purified His-CREB. GST and GST-STAT3 resins were incubated with purified His-CREB and His-HSP90, respectively, as control of specificity. Bound CREB was detected by Western blotting. F, His-CREB resin was incubated with purified GST-STAT3. As control of specificity, His-CREB and His-HSP90 resins were incubated with purified GST and GST-STAT3, respectively. Bound GST-STAT3 was detected by Western blotting.

To investigate a potential interaction between STAT1–3 and CREB, we used the biotinylated oligonucleotide/streptavidin-agarose system (Fig. 6C). The SIE oligonucleotide was chosen as a model for its high affinity for STAT1–3 and incubated with nuclear extracts from control or LIF-CRH-stimulated cells. As expected, activated STAT1–3 were bound to the SIE probe (Fig. 6C, top) and CREB was also recruited (Fig. 6C, bottom), whereas only background binding of CREB was detected without LIF-CRH stimulation. An excess of unlabeled SIE efficiently competed for STAT1–3 as well as CREB binding, suggesting that CREB and STAT1–3 interaction was not limited to the STAT_POMC site but, taking place also at the SIE site, was a more general feature of these proteins. Because purified His-CREB was unable of direct binding to SIE (data not shown), a direct or mediated binding between CREB and STAT1–3 is suggested. Moreover, Fig. 6D shows that His-CREB resin incubated with nuclear extracts from LIF-CRH-treated cells, specifically retained activated STAT1–3, whereas the unrelated His-Nter-HSP90 resin did not interact with STAT1–3. The low level of STAT1-YP binding to His-HSP90 could be explained by some residual binding property of this chaperone protein fragment (33).

Similarly to CREB and Nur 77, a direct interaction between recombinant STAT3 and recombinant CREB was demonstrated using GST-STAT3 resin and His-CREB, as well as in the converse experiment using His-CREB resin and GST-STAT3 fusion protein. His-CREB was bound to GST-STAT3 resin and not to GST resin (Fig. 6E). Moreover, only His-CREB resin and not His-HSP90 resin retained GST-STAT3 (Fig. 6F). Altogether, these results show that CREB is able to interact separately with Nur 77 and STAT1–3 whether bound or not to DNA.

Blunting the LIF-CRH Synergistic Response by Dominant-Negative CREB-M1 or CREB-siRNA

The preferential interaction of P-CREB with Nur77 (Fig. 6A) suggests a role for CREB phosphorylation in the synergistic response to LIF-CRH. Two dominant-negative forms of CREB have been described, CREB-M1 (25), in which Ser-133 has been mutated to Ala, and A-CREB (34), which is constituted of an acidic amphipathic extension onto the N terminus of the CREB leucine zipper domain and disrupts the binding of CREB to its consensus DNA motif. We tested their action, first, on the activity of a CRE reporter transfected in AtT20 cells stimulated by CRH and, second, on the constitutively active simian virus 40 (SV40) promoter. Both dominant-negative forms of CREB inhibited the response to CRH of the CRE reporter by 75% (Fig. 7A), whereas they were inactive on the SV40 promoter activity, known to be unresponsive to cAMP stimulation (Fig. 7B).

Fig. 7.

Inhibition of LIF-CRH Synergy by Dominant-Negative CREBM1 A, The CRE-CAT reporter was cotransfected in AtT-20 cells stimulated or not by CRH with the empty or CREB-M1 or A-CREB expression vectors. Results are expressed as CAT/βgal activities. B, The SV40-Luc reporter was cotransfected in AtT-20 cells with empty (white bar) or CREB-M1 or A-CREB (gray bar) expression vectors. Results are expressed as RLU/βgal activities. C, The −414/−293 reporter was cotransfected in AtT-20 cells stimulated by LIF-CRH with the empty (white bar) or CREB-M1 or A-CREB (black bar) expression vectors. Results are expressed as percentage of induction. D, The −414/−293 POMC promoter subregion, NurRE-STAT, NurRE or SIE constructs were cotransfected with the empty (white bar) or CREB-M1 (black bar) expression vectors in AtT-20 cells stimulated or not by LIF-CRH. Results are expressed as percentage of induction as compared with 100% induction by LIF-CRH of each reporter + empty vector. The results reported are the mean of three experiments ± sem.

Then, the effect of both dominant-negative CREB constructs was tested on the −414/−293 reporter. The activity of this promoter subregion after LIF-CRH stimulation was inhibited in a dose-dependent manner by CREB-M1, ratios of 3 and 6 between CREB-M1 and reporter construct resulted in a 40 and 70% inhibition, respectively (Fig. 7C), whereas A-CREB (ratio of 6) was poorly efficient. To further confirm the participation of CREB to LIF-CRH synergy, A-CREB and CREB-M1 were tested on −414/−293, 3xNuRE-STAT, 3xNurRE, 3xSTAT_POMC, and 3xSIE constructs. CREB-M1 inhibited the response of all the constructs by 70–85% (Fig. 7D). Inhibition by A-CREB was never above 20–40%, and this mutant was no longer considered (not shown).

The efficient inhibitory effect of CREB-M1, by contrast to that obtained with A-CREB, suggested that phosphorylation of CREB at Ser133 was a key event in synergism, independent of CREB binding to CRE-like sequences.

To further confirm a functional role for CREB in LIF-CRH synergy, we used the siRNA approach. Transfection of a CREB-siRNA blunted the synergistic response to LIF-CRH stimulation of the −414/−293 subregion by approximately 80%, whereas a scrambled version of the sequence had no effect (Fig. 8A). Because transfection efficiency in AtT20 cell line is low, analysis of CREB level was performed after FACS selection of cells transfected with fluorescent siRNA oligonucleotides. The CREB-siRNA reduced of about 50% the protein level, in agreement with the reduced activity of the reporter construct. The scrambled siRNA (scr-siRNA) had no effect and neither siRNA oligonucleotide reduced the abundance of the regulatory subunit of the PKA-R1α (Fig. 8B). We thus concluded that decreased level of CREB phosphorylation or CREB expression both negatively affected the synergistic response.

Fig. 8.

Blunting the LIF-CRH Synergy by CREB-siRNA A, The −414/−293 reporter was cotransfected with scr-siRNA or specific CREB-siRNA in AtT-20 cells treated (black bars) or not (white bars) by LIF-CRH. Results expressed as RLU/βgal are the mean of three experiments ± sem. B, AtT-20 cells transfected with fluorescent scr-siRNA or specific CREB-siRNA were selected by FACS. Extracts of selected cells were probed by Western blotting with anti-CREB and anti-R1α (subunit of PKA) antibodies.

DISCUSSION

A strong interaction between the signaling pathway of proinflammatory cytokines, like LIF, and that of CRH, the major positive regulator of the POMC gene transcription, has been already reported (16, 21, 23). Indeed, CRH and LIF synergize to induce POMC promoter transcriptional activity and ACTH secretion in AtT-20 cells. However, the precise mechanism of this synergy is not fully understood and the proposed targets, on the POMC promoter, do not account for the overall response. The activation of the POMC CRH-responsive element site (−173/−160), that seems to sustain the LIF-CRH synergistic response in the central part of the POMC promoter, does not involve the Jak-STAT pathway, and its role has not been fully demonstrated in the absence of experiments using a precise mutagenesis (16). As to the AP-1 site, located in the first POMC gene exon, its mutation inhibited only by approximately 20% the synergistic effect. Even though the additive effect of LIF and CRH on c-fos and junB synthesis participates to synergism (23), the AP-1 site cannot account for the early transcriptional response, independent from de novo protein synthesis.

It should also be stressed that we and others, during studies on POMC promoter regions responsible for LIF responsiveness, in a STAT1–3 DNA binding-dependent manner, noticed that the unique STAT1–3 binding site of the POMC promoter overlaps in part the NurRE (23, 24), which is the target of Nur subfamily of nuclear receptors and the major determinant of CRH responsiveness (5, 6). We therefore hypothesized that the composite NurRE-STAT site could be the region in the POMC promoter where the cAMP-PKA and Jak-STAT signaling pathways converge. Indeed, the −414/−293 and the 3xNurRE-STAT constructs were sufficient to confer the synergistic effect of LIF-CRH and mutations on the NurRE-STAT site in the −414/−293 subregion and in the entire promoter, blunted the synergy by 90% and 50% respectively. Surprisingly, when the 3xNurRE and 3xSTAT_POMC reporters were tested, we found again a synergistic response, albeit of lower amplitude as compared with 3xNurRE-STAT construct.

In the context of the entire POMC promoter or the subregion −414/−293, such functional results indicate that the NurRE-STAT element mediates the synergistic effect of LIF-CRH treatment and that the presence of both parts of the composite site is required for an optimal synergy. The finding that each individual site was able to confer a suboptimal synergistic response suggests that the binding to DNA of one class of transcription factors (Nur77 or STAT1–3-YP) is sufficient to elicit synergy between CRH and LIF.

The partial overlapping sequences of NurRE and STAT1–3 sites led us to investigate whether both classes of transcription factors make contacts with the cognate DNA motif NurRE-STAT during the synergistic treatment. CRH or LIF alone induced binding of Nur77 and STAT1–3-YP to NurRE-STAT element, respectively. Thus, the partial overlap of NurRE and STAT1–3 binding site might allow simultaneous binding of Nur and STAT during the combined stimulation. This issue is reinforced by the computer modeling presented in Fig. 5, suggesting that both transcription factors can contact the bases of NurRE-STAT if Nur or STAT is bound to a half site. Because the binding of each transcription factor seems not to be mutually exclusive, the LIF-CRH synergy may be a consequence of an interaction between the two classes of transcriptional regulators. Although coimmunoprecipitation experiments did not reveal such interactions, we cannot exclude that they take place in vivo because contacts between a nuclear receptor and STAT proteins have been already described (35).

It is also known that coactivators and adapters are present as components of supramolecular structures assembled on enhancer regions and that some transcription factors may also play the function of coactivator or adapter molecules (36).

Ser-133 phosphorylation of CREB, as seen after CRH stimulation in AtT-20 cell line (16), is known to be crucial for CBP/p300 recruitment and activation of transcription (37). Besides confirming this increased phosphorylation after the combined LIF-CRH treatment, we found that endogenous CREB from nuclear extracts was indirectly bound to NurRE-STAT element, in a stimulation-dependent manner. In these complexes, the presence of CREB requires the binding of Nur77 or STAT1–3 to DNA, indicating a direct or indirect interaction of CREB with each transcription factor. Indeed, we have demonstrated a direct interaction between CREB and Nur77, as well as CREB and STAT 3 recombinant proteins. Moreover, in LIF-CRH-treated nuclear extracts, CREB displays an increased affinity, related to its Ser-133 phosphorylation, toward recombinant Nur77, suggesting a crucial role for P-CREB in this interaction.

Whether the phosphorylation-dephosphorylation state of Nur77 also regulates the interaction with CREB remains to be determined. It is known that dephosphorylation of Ser-350 favors binding of Nur77 to its DNA motifs and hyperphosphorylation of the N-terminal region (activation function 1) regulates its transcriptional activity (38, 39). Indeed, after CRH stimulation, a PKA/MAPK-dependent phosphorylation of Nur77 has been reported as crucial for its activation (8, 9).

Because CREB is an essentially nuclear protein and one of the most obvious roles of STAT1–3 Tyr-phosphorylation is to induce their nuclear translocation, thus Tyr phosphorylation is a prerequisite for interaction with CREB to take place.

When using NurRE-STAT oligonucleotide pull-down, CREB makes contacts with both Nur77 and STAT1–3. When single NurRE or STAT (SIE) site is used, still CREB participates to a ternary complex with each DNA element and the respective transcription factor. In this case, whether the second DNA binding transcription factor (Nur or STAT) plays also a role remains to be determined. Although the simultaneous presence of activated STAT1–3 and CREB is detected with the SIE probe, as with the natural STAT _POMC site, the synergism observed with NurRE alone is not explained by the physical presence of a component of each signaling pathway. It cannot be excluded that tripartite protein complexes composed of DNA-bound Nur77/CREB/STAT or DNA-bound STAT/CREB/Nur77 exist at each separate site. Nevertheless, the synergism may also be explained by cross talk between CRH and LIF signaling, resulting in posttranslational modification of each factors and/or associated coactivators.

The presence of P-CREB within such complexes can improve the stability of activated Nur and STAT binding to NurRE-STAT site and/or facilitates the recruitment of coactivators acting on chromatin structure, stabilizing contacts with the basal transcriptional machinery. The low level of synergy observed with NurRE or STAT_POMC site, each activated alone, suggests that, when the second activated transcription factor (Nur or STAT) lacks the corresponding DNA-binding site, the stabilizing/activating effect of P-CREB is less pronounced, resulting in an attenuated response. However, this situation is reversed when the reporter is constituted of high affinity binding sites, like repetitions of the SIE. Functional and binding results with SIE suggest that this function of CREB in transcriptional activation, independent from its binding to CRE, could be more widely used than previously thought.

The use of dominant-negative forms of CREB, CREB-M1, and A-CREB, the first one being defective for Ser-133 phosphorylation and the second one being able to disrupt CREB binding to CRE (25, 34), confirmed the importance of P-CREB in the mechanism of LIF-CRH synergy and excluded a major role for CREB binding to CRE. Indeed, CREB-M1 inhibited LIF-CRH synergistic effects of all the constructs whereas A-CREB was modestly efficient as dominant negative. It is reasonable to propose that P-CREB plays here a role of coactivator or adapter. On the contrary, the modest effect of A-CREB suggests that the DNA binding property of CREB may be required for the induction of mRNAs of the Nur family of orphan receptor. Indeed, a CRE is present in the promoter of Nurr1 (40), and induction of Nur expression (5, 6) may constitute a secondary mechanism for its increased transcriptional activity, occurring later, after phosphorylation-dephosphorylation events.

In addition to CREB phosphorylation, the level of CREB in LIF-CRH synergy was also critical, because CREB-siRNA specifically inhibited the synergistic response.

Many transcription factors, like members of nuclear receptor or STAT families, can regulate transcription in the absence of DNA binding (36) and we propose here that CREB could also act by such a mechanism. On the POMC promoter, P-CREB may bridge Nur77 and STAT1–3 bound to the NurRE-STAT composite site and mediate the synergistic response to LIF-CRH. Ours results are similar to those described for p53-responsive genes where P-CREB mediates the recruitment of CBP to DNA-bound p53, constituting an alternate mechanism of coactivator recruitment (41). Future work will focus on requirement of CREB at the NurRE-STAT site for the recruitment of CBP/p300 and/or p160 coactivators because Nur77 (9) and STAT3 (42) share them. In conclusion, these results provide further knowledge on the functional analysis of the POMC gene promoter where the NurRE-STAT composite site behaves as a tethering element (43) for CREB. More importantly, they unravel a new molecular action of the transcription factor CREB that can have more general implications in the cross talk between cAMP/PKA and Jak-STAT1–3 signaling pathways.

MATERIALS AND METHODS

Cell Culture

AtT-20/D16v cells were grown in DMEM/nutrient HAM’s F12 mixture (1/1) (Sigma, St. Louis, MO) supplemented with 10% Fetal Clone III (Hyclone, Logan, UT) and 10% Nu Serum (Becton Dickinson, Franklin Lakes, NJ), 2 mm glutamine, and 0.5 mg/ml gentamicin. All cultures were maintained at 37 C in an atmosphere of 5% CO₂.

Plasmids, Transfection, and Luciferase Assay

The −480/−34 and −414/−293 segments of the rat POMC promoter, fused to the minimal promoter −34/+63 and subcloned into the pXP1-luciferase vector were already described (5, 24). Three repetitions of the POMC [NurRE (Nur response element)-STAT] or POMC (NurRE) (5, 6) or POMC (STAT) or SIE (SIEm67) sequences (27, 28) have been introduced in the same reporter in front of the minimal POMC promoter, giving the constructs 3xNurRE-STAT, 3xNurRE, 3xSTAT_POMC and 3xSIE, respectively.

The −414/−293 construct was widely mutated at the NurRE-STAT site giving the −414/−293 mutated reporter (TAGcagcgcccACCTCCgggcctCAGcggGGC) (lower case letters indicate the mutations; see the wild-type sequence in Biotinylated oligonucleotide-streptavidin pull-down). The cDNA of CREB and of the two dominant-negative forms of CREB, CREB-M1 and A-CREB (25, 34), were subcloned in pRSV expression vector to avoid CRE sequences of the original cytomegalovirus promoter constructs. His-CREB construct was obtained by subcloning the CREB cDNA in a modified pET-28 vector.

AtT-20 cells were plated in six-well plates (4 × 10⁵ cells/well) and allowed to adhere for 24 h. Cells were then transfected using Lipofectamine Plus Reagent (Invitrogen Life Technologies, Carlsbad, CA) in the absence of serum. Each sample mix for lipofection contained 500 ng of POMC promoter-luciferase reporter plasmid and 250 ng of pRSV-lacZ plasmid as an internal control. For experiments with dominant-negative forms of CREB, 3 μg of CREB-M1 or A-CREB were added to the usual sample mix. The day after transfection, cells were treated or not with 1 nm recombinant mouse LIF (Sigma) and/or 50 nm rat CRH (Bachem, Bubendorf, Switzerland) for 6 h. Cells were washed with cold PBS and then lysed in Tris/H₃PO₄ 25 mm (pH 7.8), MgCl₂ 10 mm, EDTA 1 mm, Triton 1%, and dithiothreitol 1 mm. The luciferase and β-galactosidase activities were measured as described (24). Each experiment was independently repeated at least three times, with each assay in triplicate. Results are expressed as relative light units (RLU)/βgal activities or fold induction.

Nuclear Extracts

AtT-20 cells were grown to 80% confluence, then serum-deprived during 16 h before treatment with LIF (1 nm) during 20 min, CRH (50 nm) during 60 min or LIF-CRH during 30 min. Cells were harvested in cold PBS and nuclear extracts prepared as described (24).

Western Blotting and Antibodies

Western blotting was realized as described (24). Detection of Tyr-701-phospho-STAT1 (STAT1-YP), STAT1, Tyr-705-phospho-STAT3 (STAT3-YP), STAT3, Ser-133-P-CREB, CREB, Nur77 and R1α-subunit of PKA was carried out with polyclonal anti-STAT1-YP, monoclonal anti-STAT1, monoclonal anti-STAT3-YP, polyclonal anti-STAT3, polyclonal anti-P-CREB, polyclonal anti-CREB, monoclonal 2E1anti-NGFI-B and monoclonal anti-PKA-RIα antibodies [Upstate Biotechnology (Lake Placid, NY) 06-657, Transduction Laboratories (Lexington, KY) G16920, Upstate Biotechnology 05-485, Santa Cruz Biotechnology Inc. (Santa Cruz, CA) H190, Cell Signaling (Beverly, MA) 9191/9192, a gift from J. Milbrandt and Transduction Laboratories 610609, respectively], in blocking buffer for 16 h at 4 C. Secondary antibody conjugated to horseradish peroxidase (Santa Cruz) was incubated with the membrane in blocking buffer for 1 h at room temperature and detection was accomplished using Enhanced Chemiluminescence (ECL) reagent (Amersham Biosciences Europe GmbH, Orsay, France). When reprobed, the membrane was stripped in sodium dodecyl sulfate 0.2%, NaCl 0.1 m, glycine 0.1 m/HCl (pH 2) for 1 h at room temperature and reequilibrated in Western blot buffer.

Biotinylated Oligonucleotide-Streptavidin Pull-Down

Binding of STAT1–3, Nur77, and CREB to the following 3′-biotinylated DNA oligonucleotides was tested using the biotin-streptavidin affinity system:

Control probe: CATCCTCCGCGGATC

SIE (Refs.27 and 28): CATTTCCCGTAAATC;

CRE (Ref.25): GATTCAATGACATCACGGCTGTG;

CRE mut: GATTCAAGgaACATagCGGCTGTG;

POMC (−407/−376) NurRE-STAT: TAGTGATATTTACCT CCAAATGCCAGGAAGGC; POMC(−407/−376) NurRE-STATmut: TAGTGATATTTAC CTCCAAATGCCAGcggGGC;

POMC (−407/−376) NurREmut-STAT: TAGcagcgcccAC CTCCgggTGCCAGGAAGGC.

The STAT1–3 and CRE sites are underlined, and the NurRE site is in boldface.

After annealing, biotinylated oligonucleotides (1 μg) were incubated with precleared nuclear extracts (0.5–1 mg) derived from AtT-20 cells, treated or not as above, and 100 μl streptavidin-agarose (Pierce, Rockford, IL) in a 2 ml of incubation buffer [Tris 10 mm (pH 7.4), NaCl 50 mm, glycerol 5%, EDTA 1 mm, MgCl₂ 5 mm, BSA 1 μg, poly-deoxyinosine-deoxycytosine 20 μg, antipain 5 μg/ml, leupeptin 5 μg/ml, aprotinin 5 μg/ml, vanadate 1 mm, okadaic acid 0.05 μm]. Incubation was carried out on a rotating wheel for 2 h at 4 C. After centrifugation, the pellet containing the streptavidin-agarose was washed four times with washing buffer [Tris 10 mm (pH 7.4), EDTA 1 mm, NaCl 100 mm] and the proteins eluted from the resin with Laemmli sample buffer were resolved on SDS-PAGE and examined by immunoblotting with the respective specific antibodies. The quantitative binding of oligonucleotides to streptavidine-agarose was verified by the analysis of the supernatant in appropriate nondenaturing polyacrylamide gel

GST-Nur77 Fusion Protein Pull-Down

Constructs encoding fusion proteins between GST and Nur77 or STAT3 were already described (6, 44). Purified GST-Nur77 or GST-STAT3 bound to gluthathione-Sepharose beads (50 μl, Amersham Biosciences) was incubated with nuclear extract (0.5 mg) derived from untreated or LIF-CRH-treated AtT-20 cells, or with purified 6xHis tagged CREB. The incubation was carried out in NTEN buffer [Tris 20 mm (pH 7.4), EDTA 1 mm, NaCl 100 mm, Nonidet P-40 0.5%] containing protease and phosphatase inhibitors (phenylmethylsulfonyl fluoride 1 mm, antipain 5 μg/ml, leupeptin 5 μg/ml, aprotinin 5μg/ml, vanadate 1 mm, okadaic acid 0.05 μm) on a rotating wheel for 2 h at 4 C. After centrifugation, the beads were washed three times with NTEN buffer then eluted with Laemmli sample buffer. The eluate was resolved on SDS-PAGE and examined for the presence of P-CREB or CREB proteins by immunoblotting with the respective specific antibodies. Control experiments included bacterial expressed GST alone to estimate nonspecific interactions with GST and gluthatione-Sepharose beads.

Ni-NTA Magnetic Agarose Beads Pull-Down

Purified 6xHis-tagged CREB (His-CREB, 90 pmol) was incubated with 10 μl of Ni-NTA Magnetic Agarose Beads (QIAGEN, Valencia, CA) in interaction buffer [Tris 0.1 m, NaCl 0.15 m, Imidazole 20 mm (pH 7.4); final volume 500 μl] on a rotating wheel during 1 h at 4 C. The beads were immobilized on a magnetic separator, the supernatant removed, and the beads washed with 500μl of interaction buffer. The His-CREB bound to the beads was then incubated with nuclear extract (1.5 μg) derived from LIF-CRH-treated AtT-20 cells or with purified GST-STAT3, in interaction buffer (final volume 500 μl), on a rotating wheel for 90 min at 4 C. The beads were washed as above and the proteins retained were eluted with Laemmli sample buffer. The eluate was resolved on SDS-PAGE and examined for the presence of STAT1-YP and 3-YP or STAT1–33 by immunoblotting with respective specific antibodies. Control experiments were realized with a 6xHis-tagged N-terminal segment (1–221) of HSP90 protein mutated in the ATP binding site (33).

Computer Modeling of NurRE-STAT Interaction with Nur77 and STAT3

Two complexes between NBRE DNA motif and the Nur77 DNA binding domain obtained from the Protein Data Bank (ID: 1CIT) were linked to obtain a double stranded oligonucleotide of the same length as the NurRE site of the POMC promoter. Bases in this oligonucleotide were substituted or added to generate the sequence of the NurRE-STAT. The STAT 3 dimer was manually docked in its binding site accordingly to the crystal structure obtained from the Protein Data Bank (ID: 1BG1).

Transfection of CREB-siRNA

We used CREB-siRNA (5′-CCUUAGUGCAGCUGCCCAAdTdT-3′) or scr-siRNA (5′-UGCCCAAGCACCUUAGUGCdTdT-3′), fluorescein-labeled at 3′ end (Eurogentec, Herstal, Belgium). For studies on gene reporter, AtT-20 cells were plated in 24-well plates (10⁵ cells/well) 24 h before transfection. Per well, 40 pmol of siRNA were cotransfected with 500 ng of −414/−293 subregion reporter constructs and 250 ng of pRSV-LacZ in AtT-20 cells using Lipofectamine 2000 (Invitrogen Life Technologies). Twenty-four hours later, cells were treated or not (C) with LIF (1 nm)-CRH (50 nm) for 6 h, then washed with PBS and lysed in Tris/H₃PO₄ 25 mm (pH 7.8), MgCl₂ 10 mm, EDTA 1 mm, Triton 1%, dithiothreitol 1 mm. The luciferase and β-galactosidase activities were measured as described (24). Each experiment was independently repeated at least three times, with each assay in triplicate. For studies on the level of CREB expression, 5.10⁶ AtT-20 cells were plated on dish (10 cm diameter). CREB-siRNA or scr-siRNA (1200 pmol/dish) was transfected with Lipofectamine 2000 (Invitrogen Life Technologies) 24 h after plating and a second transfection (boost) was realized 24 h after the first one. Cells were then washed extensively with PBS, then treated with cell dissociation buffer (Invitrogen Life Technologies), collected by 500 × g centrifugation and transferred in tubes for sorting on FACS (EPICS Elite, Coulter). Fluorescent cells were collected and treated with lysis buffer [Tris/HCl 50 mm (pH 7.5), NaCl 0.4 m, EDTA 5 mm, NaF 50 mm, okadaic acid 0,05 μm and Roche Complete proteases inhibitors cocktail] during 30 min at 4 C. After centrifugation 15 min at 3200 × g at 4 C, the supernatant was analyzed on SDS-PAGE and CREB revealed by Western blot analysis. Blot was then reprobed with monoclonal anti-PKA-RIα antibody.

Acknowledgments

We thank Y. de Keyzer (Institut Cochin, Paris, France) for helpful discussion and critical reading of the manuscript and J. Bertherat (Institut Cochin, Paris, France) for help with anti-CREB antibodies. We thank J. Drouin (Institut de Recherche Cliniques de Montréal, Montréal, Canada) for kindly providing many POMC promoter and GST-Nur77 constructs, M. R. Montminy (The Salk Institute of Biological Studies, La Jolla, CA) and D. D. Ginty (The John Hopkins University, School of Medicine, Baltimore, MD) for the gift of CREB, CREB-M1, and A-CREB expression vectors.

This work was supported by the Institut National de la Santé et de la Recherche Médicale, Centre National de la Recherche Scientifique and the Association pour la Recherche sur le Cancer (grant to M.G.C.). V.M. was a fellow of Ministère de la Recherche et de la Technologie and the Association pour la Recherche sur le Cancer, and O.L. was a fellow of Société de Secours des Amis des Sciences and the Ligue Contre le Cancer (Indre).

Abbreviations:

 
• A-CREB,

  Dominant-negative form of CREB constituted of N-terminal acidic extension fused to the leucine zipper of CREB;

 
• AP,

  activator protein;

 
• CBP,

  CREB binding protein;

 
• CRE,

  cAMP response element;

 
• CREB,

  CRE binding protein;

 
• GR,

  glucocorticoid receptor;

 
• GST,

  glutathione-S-transferase;

 
• His-CREB,

  purified recombinant CREB;

 
• HSP,

  heat shock protein;

 
• Jak,

  Janus kinase;

 
• LIF,

  leukemia inhibitory factor;

 
• NBRE,

  Nur77 binding response element;

 
• NurRE,

  Nur response element;

 
• P-CREB,

  phoshorylated CREB;

 
• PKA,

  protein kinase A;

 
• POMC,

  proopiomelanocortin;

 
• RLU,

  relative light units;

 
• scr-siRNA,

  scrambled siRNA;

 
• SIE,

  Sis-inducible element;

 
• siRNA,

  small interfering RNA;

 
• STAT,

  signal transducers and activators of transcription;

 
• SV40,

  simian virus 40;

 
• -YP,

  tyrosine phosphorylated.