Thyroid hormone activation of transcription is potentiated by activators of cAMP-dependent protein kinase.

We characterized the cross-talk between activators of protein kinase A (PKA) and thyroid hormone (T3) in T3 receptor (TR)-mediated transcription. U937 cells were cotransfected with a plasmid expressing the TR and a reporter plasmid containing a T3 response element (TRE) oriented either as a direct repeat or as a palindrome upstream of the thymidine kinase promoter linked to the chloramphenicol acetyltransferase gene. T3 activated transcription by 10-fold. T3 response was potentiated 2.5-3-fold by activators of PKA, but an activator of protein kinase C or of guanylate kinase was ineffective. In the absence of T3, activators of PKA had no effect on transcription. TR heterodimerization with the retinoid X receptor may facilitate T3/PKA cross-talk because coexpression of the retinoid X receptor potentiated cross-talk. Synergy was not observed in JEG-3, F9, CV-1, HeLa, L929, and HTC cells, indicating that it may require cell-specific factors. Synergy required the DNA- and ligand-binding domains, but not the amino-terminal domain, indicating that T3- and TRE-induced conformational changes on the TR are essential for cross-talk. PKA phosphorylated the TR in vitro, suggesting that, like other nuclear receptors, the TR is a target for PKA. These results imply that PKA cross-talks with T3 at the level of the TRE-bound TR, enhancing its transcriptional activity in a cell-specific manner.

The actions of nuclear receptors are known to be influenced by second messenger signaling systems such as those that affect the activity of adenylate cyclase/3Ј,5Ј-cAMP-dependent protein kinase (PKA) and phospholipase/protein kinase C (PKC) (9 -13). Thus, PKA enhances ligand-dependent transcriptional activation of glucocorticoid, retinoic acid (RAR), estrogen, progesterone, and vitamin D receptors (11, 14 -19). PKA has also been shown to phosphorylate the RAR (16,17) and the progesterone receptor (20). Phosphorylation of nuclear receptors has been directly implicated in several of their properties, including hormone binding, nuclear translocation, DNA binding, dimerization, and transcriptional activation (21). However, studies of mutant receptors have not always provided a clear picture of the role of phosphorylation in receptor function. For example, the RAR is phosphorylated directly by PKA in vivo at serine 369, but mutation of serine 369 to alanine failed to abolish the PKA enhancement of RAR-mediated transcription; this finding suggests that PKA may affect the activity of other proteins involved in RAR-mediated transcription (17). Thus, the issue of whether synergy between nuclear receptors and PKA is due to phosphorylation of receptors and/or other proteins is not resolved.
It is important to examine the mechanisms of cross-talk between nuclear receptors and second messenger pathways given the interplay between these receptors and other proteins, all of which are potential targets for regulation by phosphorylation. In addition to RXRs, the action of the TR on the TRE is inhibited or augmented in a ligand-dependent manner by interactions with corepressors or coactivators, including the nuclear receptor corepressor (N-CoR), the silencing mediator for retinoid and thyroid hormone receptors (SMRT), the TR-interacting protein (TRIP1), and the steroid receptor coactivator-1 (SRC-1) (22)(23)(24)(25)(26)(27). Furthermore, PKA-and/or PKC-regulated transcription factors that act through other DNA elements, such as the cAMP response element-binding protein (CREB) and the activator protein-1 (AP-1) complex, may also interact with the TR to influence TR function (28 -30).
There is evidence that T 3 action and second messenger signaling pathways may be interrelated. T 3  sensitivity to ␤-adrenergic stimuli (1). This issue is of clinical importance because T 3 and ␤-adrenergic agonists may interact to promote thyrotoxicosis. Although some of these cross-talk effects may be mediated by influences other than those on transcription, it is reported that T 3 and cAMP effectors can independently stimulate the rat growth hormone promoter activity in rat pituitary GH4C1 cells (31), and their combined effects are synergistic (31). In fact, we previously demonstrated that synergistic activation of the rat growth hormone promoter in non-pituitary U937 cells requires the pituitary-specific factor Pit-1, the TR, and activators of both PKA and PKC, but this synergy was both T 3 -and TRE-independent (32). Despite these indications for potential cross-talk between T 3 and second messenger systems, to our knowledge, there are no studies investigating whether PKA can alter T 3 regulation of TR/TRE-mediated transcription.
In this study, we examined the influence of activators of PKA, PKC, and guanylate kinase on T 3 activation of a minimal promoter through various TREs. We show that activators of PKA (but not PKC or guanylate kinase) potentiated TR-mediated transcription in a cell-specific manner that was dependent on TR binding both to T 3 and to the TRE. The DBD and LBD of the TR, but not the amino-terminal domain, were essential for synergy. The purified TR was phosphorylated by PKA in vitro, suggesting that TR phosphorylation may be involved in the cross-talk.

MATERIALS AND METHODS
Human promonocyte U937, human choriocarcinoma JEG-3, and mouse embryonal carcinoma F9 cell lines, RPMI 1640 medium, Dulbecco's phosphate-buffered saline, glutamine, and penicillin/streptomycin were obtained from the cell culture facility at University of California (San Francisco). Biobrene was purchased from Applied Biosystems. Plasmid Constructions-A series of reporter genes were constructed by ligating double-stranded oligonucleotides containing a TRE into the polylinker upstream of the Ϫ32/ϩ45 herpes simplex thymidine kinase (TK) promoter linked to chloramphenicol acetyltransferase (CAT)-coding sequences in a pUC19-based vector. DR4-TKCAT contains one or two copies of a direct repeat of the consensus TRE half-site spaced by 4 base pairs (5Ј-agcttcAGGTCAccagAGGTCAgag-3Ј). MHC-TKCAT contains one copy of the DR4-like TRE from the rat ␣-myosin heavy chain (Ϫ151/Ϫ124) (5Ј-agcttggctctggAGGTCAccagAGGACAgcg-3Ј). TREpal-TKCAT contains two copies of the consensus half-site unspaced and arranged as a palindrome (5Ј-atattcAGGTCATGACCTgaatat-3Ј). The reporter plasmids had a deletion of a putative AP-1-like site downstream of CAT-coding sequences previously reported to mediate nuclear receptor action (29,32,33). Expression vectors for CREB (34), v-ErbA (35), T⌬T (29), PKI and mutant PKI (36), and TTT and ⌬TT (37) have been described previously. Plasmids for mammalian expression of hTR␤1, hTR␣1, and hTR␣2 were constructed by subcloning their fulllength cDNAs (38 -40) into a plasmid under the control of the Rous sarcoma virus promoter, pRS (41), or the human metallothionein promoter, pLEN (29).
Cell Culture, Electroporation, and CAT Assays-U937, F9, and JEG-3 cells were maintained and subcultured in RPMI 1640 medium containing 10% newborn calf serum, 2 mM glutamine, 50 units/ml penicillin, and 50 g/ml streptomycin. Transfection procedures were as described previously (42). Briefly, cells were collected and resuspended in Dulbecco's phosphate-buffered saline (0.5 ml/1.5 ϫ 10 7 cells) containing 0.1% dextrose, 10 g/ml Biobrene, and typically 10 g of reporter plasmid and 1 g of expression vector. Cells were electroporated at 300 V and 960 microfarads, transferred to RPMI 1640 medium containing 10% stripped newborn calf serum, and then plated in 6-well plates. After incubation for 24 h at 37°C with ethanol, 10 -100 nM T 3 , and/or 10 M forskolin, the cells were collected, and the pellets were solubilized by the addition of 100 l of 0.25 M Tris-HCl, pH 7.6, containing 0.1% Triton X-100. Cellular lysates were assayed for CAT activity using the liquid scintillation method described previously (43).

Nuclear Extracts, T 3 Binding, and Electrophoretic Gel Mobility Shift
Assays-Nuclear extracts of U937 cells treated for 24 h with forskolin or vehicle (ethanol) were prepared as described (42), and protein concentration was measured by Coomassie Blue staining. TR concentration in nuclear extracts was estimated by [ 125 I]T 3 binding assays (44). Binding of the TR to DNA was studied by electrophoretic gel mobility shift assays as described previously (8,44).
In Vitro Phosphorylation of the Purified TR-Bacterially expressed hTR␤1 and rat TR␣1 LBD were purified to homogeneity (Ͼ98% pure) employing procedures described previously (44,45). hTR␤1 (3 pmol) and rat TR␣1 LBD (340 pmol) were added to a reaction mixture containing 25 mM Tris-HCl, pH 7.5, 5 mM magnesium acetate, 100 M ATP, 0.5 mM dithiothrietol, [␥-32 P]ATP, and 2.5 g of the catalytic subunit of PKA in the absence or presence of 2 g of PKI in a final volume of 20 l. We also produced 32 P-phosphorylated and unlabeled phosphorylated and nonphosphorylated (by omitting PKA from the reaction) TRs for electrophoretic gel mobility shift assays by incubating hTR␤1 (1 pmol, 25% pure) in the presence or absence of [␥-32 P]ATP. After the samples were incubated for 30 min at 30°C, the reaction was terminated by the addition of 40 l of 2 ϫ SDS sample buffer containing 5% ␤-mercaptoethanol. The reaction products were examined by SDS-polyacrylamide gel electrophoresis. The dried gel was then exposed to Kodak X-Omat x-ray film at Ϫ70°C. TPA, and 8-Br-cGMP did not regulate DR4-TKCAT in the absence of T 3 , whereas T 3 alone activated it by 10-fold. The activation of DR4-TKCAT by T 3 was potentiated 2.5-fold by 8-Br-cAMP and 3-fold by forskolin, but was not affected by TPA and 8-Br-cGMP. T 3 , forskolin, or both did not activate the parental plasmid missing DR4, indicating that a TRE is essential for both T 3 activation and synergy with forskolin (see Fig.  5). These results demonstrate that activators of PKA, but not of PKC or guanylate kinase, exert a synergistic effect on the T 3 /TR activation of DR4-TKCAT in U937 cells.

Activators of PKA Potentiate T 3 Transcriptional Activation in
To confirm that the effects of cAMP are mediated by PKA, a plasmid expressing either a native or a mutant inhibitor of PKA (PKI) was cotransfected with DR4-TKCAT and hTR␤1 into U937 cells. Fig. 1B shows that native PKI did not affect T 3 activation, but markedly blunted the potentiation by forskolin. In contrast, mutant PKI, which does not inhibit PKA, did not block the forskolin potentiation of T 3 activation. These studies demonstrate that the potentiation by forskolin is mediated by cAMP activation of PKA.
Dose-response experiments showed that forskolin did not shift the T 3 response curve, indicating that synergy is not due to increased affinity of the TR for T 3 (data not shown). Furthermore, the synergy is not due to stimulation of the TR expression vector by forskolin since the number of TRs was similar in control or forskolin-treated U937 cells transfected with hTR␤1 (70.7 Ϯ 6.9 versus 49.0 Ϯ 19.3 fmol of TR/mg of protein for control and forskolin-treated cells, respectively; n ϭ 3; p ϭ 0.28 by Student's paired t test). Finally, TRs commonly form heterodimers with the RXR, and it is conceivable that T 3 activation and synergy with forskolin may be augmented by cotransfection of RXRs. Table I shows that cotransfecting human RXR␣ with hTR␤1 in U937 cells did not alter T 3 activation, but significantly enhanced the forskolin potentiation of T 3 activation.
Forskolin Potentiation of T 3 Transcriptional Activation Is Cell-specific, but Is Not Due to CREB-Cross-talk between nuclear receptors and PKA has been reported in various cell lines (11, 14 -19). However, it is possible that synergy between T 3 and forskolin might differ among cell lines since the content of factors or substrates necessary for PKA action may be tissuespecific. Therefore, as shown in Fig. 2, we tested other cell types known to be responsive to the cAMP/PKA signaling pathway. Whereas T 3 activated hTR␤1-mediated transcription on DR4 (two copies)-TKCAT by Ϸ40-fold in JEG-3 cells and by Ϸ120-fold in F9 cells, the addition of forskolin did not alter the T 3 stimulation. Identical studies in CV-1 monkey kidney, HeLa human cervical carcinoma, L929 mouse fibroblast, and HTC rat hepatocarcinoma cells also showed T 3 activation, but no synergy between T 3 and forskolin was observed (data not shown). CREB is a major target for PKA phosphorylation that is absent in F9 cells. To determine whether CREB mediates the synergy by T 3 and forskolin in F9 cells, we cotransfected CREB with hTR␤1. Fig. 2 shows that CREB did not induce synergy in F9 cells, indicating that the lack of synergy is not due to CREB absence. Altogether, these results indicate that synergy between forskolin and T 3 is cell-specific.
Forskolin Potentiation of T 3 Transcriptional Activation Is Not TR Isoform-specific, but Requires T 3 -bound TRs-We then tested whether cross-talk between PKA and T 3 also occurs with other TR isoforms. Fig. 3 shows that in addition to hTR␤1, forskolin also potentiated the T 3 activation of the hTR␣1 isoform. In contrast to TR␤1 and TR␣1 isoforms, the TR␣2 and v-ErbA TR variants are not activated by T 3 since they have abnormal LBDs that impair T 3 binding. As expected, forskolin had no effects on cotransfected TR␣2 and v-ErbA, indicating that TR LBD structure has to accommodate T 3 -induced structural changes for synergy to occur.
The DBD, but Not the Amino-terminal Domain, of the TR Is Required for Forskolin Potentiation of T 3 Transactivation-To investigate which TR domains participate in forskolin potentiation of T 3 transactivation, DR4-TKCAT was cotransfected with a plasmid expressing full-length TR␤1 (TTT), with TR␤1 with the amino-terminal domain deleted (⌬TT), or with a TR with the DBD deleted (T⌬T). Stimulation of transcription by T 3 and potentiation by forskolin were observed with TTT and ⌬TT, but not with T⌬T, which lacks the DBD (Fig. 4A). These results demonstrate that the amino-terminal transactivation domain does not participate in synergy, whereas deletion of the TR DBD impairs T 3 stimulation and potentiation by forskolin, indicating that TR binding to the TRE is required for synergy.
To explore further the role of the TR DBD, the cDNA sequences encoding the amino terminus of full-length TR␤1 were fused in frame to those encoding the DBD (amino acids 1-147) of the yeast GAL4 transactivator protein (GALTR expression transcription Shown are the responses to T 3 in the absence or presence of forskolin in U937 cells transfected with the TR alone or with the TR and RXR, as described under "Materials and Methods." Fold activation corresponds to the ratio of CAT activity of cells treated with forskolin, T 3 , or T 3 ϩ forskolin to that of ethanol-treated control cells. Synergy index is a measure of the forskolin potentiation of T 3 activation, as calculated by the ratio of (T 3 ϩ forskolin)-induced CAT activity to that induced by T 3 alone. Data are means Ϯ of S.E. of nine independent experiments.

Activation
Synergy index vector). When GALTR was cotransfected with a reporter containing one copy of the GAL4 response element linked to minimal TKCAT, T 3 stimulated CAT expression by 5-fold, but synergy was absent (Fig. 4B). In contrast, T 3 stimulation and synergy with forskolin were observed when GALTR was cotransfected with DR4-TKCAT (Fig. 4B). These results indicate that direct binding of the TR DBD to TREs is required for synergy to occur, implying that DNA-induced allosteric changes in TR conformation are necessary for cross-talk with PKA.
Forskolin Potentiation of T 3 Transcriptional Activation Occurs with Different TREs Linked to TKCAT-Our results demonstrate that synergy requires binding of the TR to DR4. However, the TR can bind other TREs, adopting diverse conformations depending on the orientation and spacing of TRE half-sites. We examined whether synergy depends on specific TR conformations by cotransfecting hTR␤1 with vari-ous TRE-TKCAT plasmids. Fig. 5 shows that T 3 stimulated transcription from all TREs tested. The plasmid containing two copies of DR4 was stimulated the most by T 3 , producing a Ϸ100-fold increase in transcription as compared with Ϸ10-fold on one copy of DR4 and much greater than observed with two copies of TREpal (Ϸ6-fold). Whereas the magnitude of activation was different with T 3 alone, the potentiation by forskolin was ϳ3-fold in each TRE-TKCAT construct, demonstrating that synergy occurs independent of the diverse conformations adopted by TRs when bound to TREs with half-sites positioned in different orientations and spacing.
We also addressed whether the synergy depends on promoter context by substituting the TK promoter with the enhancerless SV40 early promoter (46) on the DR4-TKCAT plasmid. Fig. 5 shows that forskolin also potentiated the T 3 stimulation of DR4-SV40CAT, demonstrating that the synergy is not dependent on promoter context.
The TR Is Phosphorylated in Vitro by PKA, but Phosphorylation Does Not Alter TR DNA Binding Activity-Although a number of different proteins participating in the T 3 transactivation pathway may be a target for PKA, the most likely candidate is the TR because the synergy requires direct binding of the TR to T 3 and to the TRE, and prior reports showed that PKA phosphorylates the progesterone receptor and RAR, modulating their transcriptional influences (17,18). To determine if hTR␤1 is phosphorylated by PKA in vitro, purified unliganded bacterially expressed hTR␤1 was incubated with PKA and [ 32 P]ATP in the absence or presence of PKI. As shown in the autoradiograph obtained by SDS-polyacrylamide gel electrophoresis in Fig. 6A, PKA phosphorylation of a 98% pure hTR␤1 preparation resulted in two 32 P-labeled bands of 52 and 46 kDa (second lane), which correspond to the molecular size of hTR␤1. These bands were abolished by PKI (first lane), but the addition of T 3 did not change the phosphorylation pattern (data not shown). A similar pattern was obtained by phosphorylating this TR preparation using HeLa cell cytosolic kinases (47). Interestingly, PKA did not phosphorylate the LBD fragment of rat TR␣1 purified to homogeneity (Ͼ99% pure) in the presence or absence of PKI (fifth and sixth lanes), nor the LBD fragment of hTR␤1 (data not shown). In parallel experiments, PKA phosphorylated bovine serum albumin (fourth lane), but not ovalbumin (third lane), demonstrating the specificity of PKA.
TR phosphorylation by HeLa cell cytosolic kinases enhances binding of TRs to DNA (47,48). Thus, we compared binding of nonphosphorylated and PKA-phosphorylated TRs to DR4 using   FIG. 3. Forskolin and T 3 synergy is not TR isoform-specific, but requires an intact LBD. U937 cells were cotransfected with 15 g of DR4-TKCAT and 2.5 g of expression vector for hTR␤1, hTR␣1, hTR␣2, or v-ErbA. Cells were treated with ethanol (control), 10 M forskolin (F), 10 nM T 3 , or both T 3 and forskolin (T 3 ϩ F) for 18 h and then assayed for CAT activity. RSV, Rous sarcoma virus.

FIG. 4. Forskolin and T 3 synergy requires the TR DBD and TR DBD binding to the TRE.
A, U937 cells were cotransfected with 15 g of DR4-TKCAT and 2.5 g of full-length TR␤1 (TTT), 2.5 g of TR␤1 with the amino-terminal domain deleted (⌬TT), or 2.5 g of TR␤1 with the DBD deleted (T⌬T). Cells were treated with ethanol (control), 10 M forskolin (F), 10 nM T 3 , or both T 3 and forskolin (T 3 ϩ F) for 18 h and then assayed for CAT activity. B, U937 cells were cotransfected with 2.5 g of GALTR (full-length TR␤1 fused to the GAL4 DBD) and 10 g of either GAL-TKCAT or DR4-TKCAT. Cells were treated with ethanol (control), 10 M forskolin, 100 nM T 3 , or both T 3 and forskolin for 24 h and then assayed for CAT activity.

FIG. 5. Forskolin and T 3 synergy occurs with multiple distinct
TREs. U937 cells were cotransfected with 2.5 g of hTR␤1 and 10 g of reporter plasmid containing either one copy or two copies of DR4, two copies of TREpal, or one copy of the MHC TRE upstream of minimal TKCAT or enhancerless SV40CAT. Cells were treated with ethanol (control), 10 M forskolin (F), 10 nM T 3 , or both T 3 and forskolin (T 3 ϩ F) for 18 h and then assayed for CAT activity.
gel mobility shift assays. Fig. 6B shows that, in the absence of T 3 , no difference in DNA binding was observed between PKAphosphorylated hTR␤1 and nonphosphorylated hTR␤1 as homodimers (lanes 1 and 2) or heterodimers with the RXR (lanes  3 and 4). The addition of T 3 did not alter heterodimeric binding of phosphorylated and nonphosphorylated TRs (lanes 5 and 6), but disrupted their homodimeric binding (lanes 7 and 8). As a control, Fig. 6C shows that PKA-phosphorylated 32 P-labeled TRs bind to nonradioactive DR4 as homodimers (lane 1), and the addition of T 3 disrupts them (lane 2). These results demonstrate that TRs are phosphorylated by PKA and that the DNA binding properties of PKA-phosphorylated TRs remain similar to those of nonphosphorylated TRs.

DISCUSSION
Although several reports have examined interactions between nuclear receptors and second messenger systems, no studies have addressed the influences of second messenger systems on T 3 /TR/TRE-mediated transcription. This study demonstrates that activators of PKA, but not of PKC or guanylate kinase, potentiate T 3 effects on transcription in U937 cells on a promoter activated by TR binding to the TRE. The effects of PKA activation are synergistic with T 3 because it required a TR bound to a TRE and no effects occurred in the absence of T 3 .
The PKA potentiation of the TR was not dependent on the type of TRE structure since it occurred with direct repeats (DR4 and MHC), a palindrome (TREpal), and an inverted palindrome (F2) (data not shown). However, the presence of a functional TRE is essential for synergy since the parental plasmid lacking a TRE showed no T 3 response or PKA effects. The observation that T⌬T, the TR mutant missing the DBD, does not confer synergy also confirms that TR binding to the TRE is essential. Furthermore, when GALTR was cotransfected with a GAL-TKCAT reporter plasmid, the T 3 response was retained, but the synergy was abolished. However, synergy occurred when GALTR was cotransfected with DR4-TKCAT. These results indicate that the T 3 ⅐TR complex assumes a conformation bound to a GAL-binding site that does not permit interaction with other factors that can occur when it is bound to a TRE, implying that TRE structure can allosterically influence TR function. In addition to TR binding to the TRE, synergy requires T 3 binding to the TR since no synergy was observed using TR isoforms with mutations in the TR LBD that render it incapable of binding T 3 . The finding that deletion of the TR amino terminus does not impair synergy indicates that transactivation properties residing in the amino terminus do not play a role in the TR/PKA cross-talk. Consistent with this finding, synergy occurred with TR␤1 and TR␣1, TR isoforms with a divergent amino terminus. Thus, allosteric changes induced by T 3 and DNA binding on the LBD and DBD are essential for synergy, presumably by facilitating PKA action on the TR and/or TR-associated proteins. In this regard, we found that the RXR, which heterodimerizes with the TR, is phosphorylated by PKA in vitro (data not shown) and significantly enhances the forskolin potentiation of T 3 response, supporting the notion that non-TR factors are involved in the cross-talk between PKA and the TR.
A unique aspect of the T 3 and PKA synergy is its cell specificity, occurring in U937 cells, but not in F9, JEG-3, CV-1, HeLa, L929, and HTC cells (Fig. 2 and data not shown). By contrast, PKA synergy with the RAR and glucocorticoid, estrogen, and progesterone receptors occurs in various cell lines (11, 14 -18). These results indicate that U937 cells possess factors that are activated by PKA, which may be absent in other cell types that are responsive to T 3 . One potential factor is CREB since it is phosphorylated by PKA and is absent in F9 cells (14). However, it is unlikely that the lack of synergism between the TR and PKA in F9 and other cell types is due to the absence of FIG. 6. PKA phosphorylates the TR in vitro, but phosphorylation does not alter TR binding to the TRE. A, purified bacterially expressed hTR␤1 or TR␣1 LBD was incubated with the catalytic subunit of PKA and [␥-32 P]ATP in the absence or presence of PKI. As controls, ovalbumin (5 g) and bovine serum albumin (BSA; 2 g) were also incubated with PKA and [␥-32 P]ATP. The reaction products were then examined by SDS-polyacrylamide gel electrophoresis. B, nonphosphorylated (100 fmol) (lanes 1, 3, 5, and 7) or PKA-phosphorylated (100 fmol) (lanes 2, 4, 6, and 8) hTR␤1 was incubated with 32 P-labeled DR4 (14,000 cpm, Ϸ1.5 fmol) in the absence or presence of reticulocyte lysate in vitro translated human RXR␣ (2 l) or 10 nM T 3 and then separated on a nondenaturing polyacrylamide gel. Binding of the RXR to 32  CREB since cotransfection of CREB with the TR in F9 cells did not lead to synergy. The RXR is also unlikely to play a role in the cell-specific nature of synergy since it is present in cell types in which no synergy was observed (7). These studies suggest that U937-specific factors are necessary for synergy, and thus, these cells may provide a useful model to identify and characterize potential cell-specific factors that participate in the cross-talk between PKA and the TR.
In addition to cell-specific factors, other potential targets for PKA that may participate in synergy that are not cell-specific include the TR. Previous studies indicated that phosphorylation of the TR may lead to increased transcriptional activity since treatment of cells with okadaic acid, a phosphatase inhibitor that augments the phosphorylation status of TRs in cells, stimulates transcription (49,50). We did find that PKA efficiently phosphorylates the full-length TR, but not the TR LBD in vitro. These results, combined with those suggesting that the TRE allosterically influences the TR DBD, which is essential for synergy, are consistent with the notion that the TR DBD may be a target for PKA action. However, there are no consensus sites for PKA phosphorylation in the DBD, and further studies will be required to localize the TR sites phosphorylated by PKA. Other studies with the glucocorticoid receptor (14) suggested that the synergy may be due in part to enhanced DNA binding of the PKA-phosphorylated TR. However, it is unlikely that increased binding of the PKA-phosphorylated TR accounts for synergy since no difference in DNA binding was observed when the phosphorylated TR was compared with the nonphosphorylated TR.
The requirements for TR/PKA synergy also differ from those of other nuclear receptors in that it occurs only in the liganded state, whereas other studies found that PKA activates the unliganded progesterone, vitamin D, and estrogen receptors and RAR, mimicking the transcriptional effects of their cognate ligands (18,19,51). Our TRE-TKCAT reporter plasmids have a deletion of an AP-1-like regulatory site found in the backbone of the pUC plasmid (33); this site in TRE-TKCAT constructs conveyed forskolin activation to unliganded TRs (data not shown). In fact, we previously reported that unliganded TRs synergize with an AP-1 complex at this element (29). Thus, it is possible that activation of unliganded receptors found in some studies may be due to the presence of AP-1-like sites in the reporter plasmids.
We have shown that PKA signaling augments TR transcriptional properties in a cell-specific manner, indicating that such cross-talk depends on factors other than the TR and PKA. Also, we show that allosteric influences on TR structure exerted by T 3 and by the TRE are required for synergy. Whether TR/PKA cross-talk requires phosphorylation of the TR, RXR, or coactivators is not yet resolved. We show that the TR is a target for PKA in vitro, but further studies will be needed to define the sites and role of TR phosphorylation, if any, in the synergy. Our results indicate that TR cross-talk with PKA may explain some tissue-specific actions of thyroid hormones and provide a system to identify cell-specific factors involved in T 3 /TR-mediated transcription.