Protein Kinase C Modulates Aryl Hydrocarbon Receptor Nuclear Translocator Protein-mediated Transactivation Potential in a Dimer Context*

Protein kinase C (PKC)- and protein kinase A (PKA)-mediated modulation of the transactivation potential of human aryl hydrocarbon receptor nuclear translocator (hARNT), a basic helix-loop-helix (bHLH)-PAS transcription factor, and the bHLH-ZIP transcription factors USF-1 (for upstream regulatory factor 1) and c-Myc were examined. An 81 nm dose of the PKC activator phorbol-12-myristate-13-acetate (PMA), shown here to specifically activate PKC in COS-1 cells, or a 1 nm dose of the PKA activator 8-bromoadenosine-3′,5′-cyclic monophosphate (8-Br-cAMP) results in 2.6- and 1.9-fold enhancements, respectively, in hARNT-mediated transactivation of the class B, E-box-driven reporter pMyc3E1bLuc relative to identically transfected, carrier solvent-treated COS-1 cells. In contrast, 81 nm PMA and 1 nm 8-Br-cAMP did not enhance transactivation of pMyc3E1bLuc-driven by USF-1 and c-Myc expression relative to identically transfected, carrier-treated COS-1 cells. Co-transfection of pcDNA3/ARNT-474-Flag, expressing a hARNT carboxyl-terminal transactivation domain deletion, and pMyc3E1bLuc does not result in induction of reporter activity, suggesting PMA’s effects do not involve formation of unknown hARNT-protein heterodimers. Additionally, PMA had no effect on hARNT expression relative to Me2SO-treated cells. Metabolic 32P labeling of hARNT in cells treated with carrier solvent or 81 nm PMA demonstrates that PMA does not increase the overall phosphorylation level of hARNT. These results demonstrate, for the first time, that the transactivation potential of ARNT in a dimer context can be specifically modulated by PKC or PKA stimulation and that the bHLH-PAS and bHLH-ZIP transcription factors are differentially regulated by these pathways in COS-1 cells.

P labeling of hARNT in cells treated with carrier solvent or 81 nM PMA demonstrates that PMA does not increase the overall phosphorylation level of hARNT. These results demonstrate, for the first time, that the transactivation potential of ARNT in a dimer context can be specifically modulated by PKC or PKA stimulation and that the bHLH-PAS and bHLH-ZIP transcription factors are differentially regulated by these pathways in COS-1 cells.
The ARNT 1 protein is a member of the bHLH-PAS family of transcription factors and is the central dimerization partner (Fig. 1) for a variety of PAS family transcription factor dimers capable of binding different core DNA elements (1). Currently ARNT is known to be involved in such diverse cell signaling events as xenobiotic response (AhR⅐ARNT dimers), and hypoxic response (HIF1␣⅐ARNT and EPAS⅐ARNT dimers), as well as pathways involving central midline and central nervous system development (Sim⅐ARNT) (2)(3)(4). Importantly, null ARNT allele mice are not viable past embryonic day 10.5, are defective in angiogenesis, and have impaired development (5), demonstrating that ARNT plays an essential role in normal cellular metabolism, growth, and development. Finally, the potential importance of regulating ARNT's activity in these different bHLH-PAS dimer contexts is further underscored by the apparent involvement of one of ARNT's dimerization partners, Sim, in mammalian central nervous system development and Down's syndrome (6).
ARNT is a phosphoprotein that contains discreet, functionally defined domains. DNA binding and the formation of ARNT heterodimers involves the NH 2 -terminal half of this protein, which contains a basic region, a helix-loop-helix motif, and the PAS domain (7), with these latter two domains being required for optimal dimerization. The COOH-terminal 34 amino acids of ARNT can act as a TAD (8), but in the AhR⅐ARNT and HIF1␣⅐ARNT heterodimer contexts the ARNT TAD is not absolutely required for transactivation of target genes (2,9), but appears to be required for a maximal transcriptional response. These observations suggest that modulation of ARNT's transactivation potential 2 in the variety of different dimer contexts involving this bHLH-PAS protein may be an important mechanism contributing to the combinatorial transcriptional regulation of target genes.
c-Myc and USF-1 both belong to the bHLH-ZIP family of transcription factors and are capable of transactivating class B, E-box (5Ј-CACGTG-3Ј) regulated genes (10 -13). c-Myc is involved in cellular proliferation, differentiation, neoplasia, and apoptosis (14,15), while USF-1 is involved in regulation of the * This work was supported by National Institutes of Health NIEHS Grant ES04869 and ES09272. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Recently it has been demonstrated that the bHLH-PAS protein ARNT is capable of forming homodimers that bind class B E-boxes and, like the bHLH-ZIP proteins c-Myc and USF-1, can transactivate class B, E-box-driven reporter constructs (1,24,25). Like c-Myc, hARNT is predominantly phosphorylated on serine and threonine residues, suggesting serine/threonine kinases such as PKC and PKA may also be involved, directly or indirectly, in the regulation of ARNT's transactivation potential. A study of the potential regulatory role of the PKC and PKA serine/threonine kinase pathways, in the context of the hARNT⅐hARNT homodimer, was undertaken to begin to understand the regulation of hARNT's transactivation potential and ultimately the role of ARNT in the combinatorial control of such diverse biological phenomena as xenobiotic-induced toxicity, hypoxic response, central nervous system development, Down syndrome, and other, as yet, unexplored pathways.
An important benefit of examining the role of PKC-or PKAmediated modulation of ARNT transactivation potential in the ARNT homodimer context is that this model is uncomplicated by the presence of ARNT's heterologous dimerization partners. Additionally, the COS-1 model system described here allows the role of PKC and PKA in modulating the transactivation potential of the bHLH-ZIP transcription factors, c-Myc and USF-1, to be examined and compared with the ability of these kinase pathways to regulate the transactivation potential of the bHLH-PAS transcription factor hARNT.
In the COS-1 model system presented here, comparison of the consequences of 81 nM PMA and 1 nM 8-Br-cAMP treatments on hARNT-, c-Myc-, and USF-1-driven transactivation of the class B, E-box-driven, reporter construct pMyc3E1bLuc revealed that these compounds have differential effects on the transactivation potentials of these transcription factors. In contrast to the bHLH-ZIP proteins USF-1 and c-Myc, the transactivation potential of the bHLH-PAS protein hARNT is shown to be enhanced approximately 2-fold by both PMA and 8-Br-cAMP, presumably through stimulation of the PKC and PKA serine/threonine kinase pathways. Importantly, these results may indicate that the previously reported ability of PKC to enhance hAhR⅐hARNT heterodimer-driven transactivation (26,27) may, at least in part, be due to ARNT's contribution to the transactivation potential of this heterodimer. Taken as a whole, these results demonstrate, for the first time, that the transactivation potential of ARNT in a dimer context can be specifically modulated by PKC or PKA stimulation and that the bHLH-PAS and bHLH-ZIP transcription factors are differentially regulated by these pathways in COS-1 cells.
Plasmids-The pSV-␤-galactosidase control vector (pBGAL) and pCI-neo plasmids were from Promega (Madison, WI). pSV-Sport1 and pcDNA3 were from Life Technologies, Inc. and Invitrogen (Carlsbad, CA), respectively. The pMyc0E1bLuc and pMyc3E1bLuc reporter constructs consist of zero and three E-boxes, respectively, located 5Ј to luc ϩ ; construction of these plasmids (gifts from R. Davis) was as described (10). phuARNT, expressing full-length human ARNT (a gift from C. A. Bradfield), was constructed as described (28). pcDNA3/ARNT-474-Flag, in which hARNT's carboxyl-terminal TAD has been deleted and the truncated peptide is FLAG-tagged, was generated by polymerase chain reaction from a phuARNT template, using the forward primer 5Ј-CCC-AAGCTTCGAGCTCACTAGTCG-3Ј and the reverse primer 5Ј-CCGCT-CGAGCTACTTGTCATCGTCGTCCTTGTAGTCTGTAGGCCGTGGTT-C-3Ј, followed by HindIII/XhoI digestion and subcloning of this fragment into the HindIII/XhoI sites of pcDNA3. psvUSF1, expressing full-length USF1 (a gift from M. Sawadogo) was generated as described (13). as a XhoI fragment into pCI-neo, followed by NheI/XbaI digestion of this vector to liberate a 1400-base pair fragment, which was then subcloned into the NheI and XbaI sites of pMK10 59 (29). All plasmids were propagated in Escherichia coli strain DH5␣ (supE44 lacU169 (80 lacZ M15) hsdR17 recA1 endA1 gyrA96 thi-1 relA1).
Cell Culture, Transient Transfections, and Treatment Time Courses-FBS was obtained from HyClone Laboratories (Logan, UT). Opti-MEM ® was purchased from Life Technologies, Inc. Trypsin/EDTA, PBS, and ␣-MEM were from Sigma. COS-1 cells (American Type Culture Collection, Rockville, MD) were grown in ␣-MEM supplemented with 10% FBS, 100 IU/ml penicillin, and 0.1 mg/ml streptomycin at 37°C in 94% air, 6% CO 2 . COS-1 cells to be transiently transfected were grown to 80% confluence in tissue culture vessels and transfected with plasmid DNAs using LipofectAMINE (Life Technologies, Inc.) as described by the manufacturer. Transfected cells were allowed to recover for 14 h in complete medium prior to treatment with compounds for 6 h. Cell lysates were then prepared and assayed for luciferase and ␤-galactosidase activities.
Luciferase and ␤-Galactosidase Assays-Transfected, treated COS-1 cells were washed twice with PBS prior to lysis. Cell lysates were prepared and assayed for luciferase activity using the Promega ® luciferase assay system (Promega Corp., Madison, WI) as specified by the manufacturer. Light output was measured for 15 s at 25°C using a TD-20e Luminometer (Turner Designs, Inc., Sunnyvale, CA). ␤-Galactosidase assays were performed using the Promega ® ␤-galactosidase assay system (Promega Corp., Madison, WI) and spectrophotometrically monitored at A 420 nm as specified by the manufacturer. Luciferase activity was expressed as RLU per A 420 nm unit. Protein content of lysates used in luciferase and ␤-galactosidase assays was determined using the BCA protein assay (Pierce).
Preparation of Whole Cell Lysates-COS-1 cells were grown to 80% confluence, transfected in 60 ϫ 15-mm tissue culture dishes, washed twice with PBS, and incubated in complete medium for 20 h. Cells were washed twice with PBS and then covered in 0.4 ml of HSTD lysis buffer consisting of 50 mM Tris-HCl (pH 7.5), 500 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS, 0.02% NaN 3 , 1 mM EDTA, 100 g/ml phenylmethylsulfonyl fluoride, and 10ϫ protease inhibitor mixture (Sigma). Cell debris was pelleted by centrifugation for 15 min at 17,350 ϫ g, and protein content was determined with the BCA protein assay (Pierce).
Protein Transfer to PVDF Membranes and Quantitative Assay for the hARNT Protein-Protein samples were separated by Tricine SDS-PAGE (3.96% stacking gel/7.92% running gel) for 2.5 h at 30 V, followed by 18 h at 60 V. After electrophoresis, the proteins were electrophorectically transferred to PVDF membrane (Millipore, Bedford, MA). Transfers were performed at 8 V for 6 h in a Genie electroblot unit (Idea Scientific Co., Minneapolis, MN) in transfer buffer (20 mM Tris, 185 mM glycine, 20% (v/v) methanol). Following protein transfer the membrane was blocked with 3% (w/v) BSA in PBS containing 10 mM Na 2 HPO 4 , 150 mM NaCl, and 0.5% (v/v) Tween 20 for 30 min at 25°C. The blots were rinsed once in blot wash buffer consisting of 0.1% (w/v) BSA in PBS containing 0.5% (v/v) Tween 20. The ARNT-specific mAb 2B10 was generated as described (30) while the FLAG (M2)-specific mAb was from Kodak Imaging Systems (Rochester, NY). Blots were incubated with mAb 2B10 (1 g/ml) or M2 (1 g/ml) for 1 h at room temperature, followed by three 10-min rinses with blot wash buffer. mAb 2B10 probed blots were then incubated with goat anti-mouse HRP-IgG conjugate (1 g/ml), biotinylated goat anti-mouse IgG (1 g/ml), or 125 Ilabeled sheep anti-mouse IgG (1 g/ml) for 1 h at 25°C followed by three 10-min rinses with blot wash buffer. mAb M2-probed blots were incubated with goat anti-mouse HRP-IgG conjugate (1 g/ml) for 1 h at 25°C followed by three 10-min rinses with blot wash buffer. Blots probed with biotinylated goat anti-mouse IgG were then incubated with HRP-streptavidin conjugate (1.0 g/ml) for 1 h at 25°C followed by three 10-min rinses with blot wash buffer. HRP conjugate probed blots were then visualized using the Vector ® VIP substrate kit (Vector Laboratories Inc., Burlingame, CA) for peroxidase staining, as specified by the manufacturer, and dried. Dried, 125 I-labeled sheep anti-mouse IgGprobed blots were visualized using a Bio-Rad GS-363 Molecular Imager ® System phosphorimager and quantified with Molecular Analyst version 1.4 software (Bio-Rad). HRP-probed blots were scanned on an Epson Action Scanner II (Epson America Inc., Torrance, CA), and band quantification was also performed using Molecular Analyst version 1.4 software (Bio-Rad).
Preparation of Cytosolic and Membrane Protein Fractions for PKC Translocation Assays-COS-1 cells were grown to 90% confluence in 175-cm 3 tissue culture flasks and treated for 30 min with Me 2 SO, 81 nM PMA, or 81 nM 4-MeO-PMA at which point the preparation of fractions began immediately. Treated cells were washed with PBS, trypsinized, collected by centrifugation, washed with PBS and pelleted again. Cell pellets were resuspended in 0.5 ml of ice-cold homogenization buffer (5 mM EDTA, 10 mM EGTA, 0.3% ␤-mercaptoethanol, 1 mM phenylmethylsulfonyl fluoride, 10 mM benzamide, and 20 mM Tris-HCl, pH 7.5) and disrupted by sonication (3 cycles of 5 s at 22 Hz with intervals of 15 s) on ice using a Branson Sonifier 250 (Branson Ultrasonics Corp., Danbury, CT). Cell lysates were then centrifuged at 100,000 ϫ g for 45 min at 4°C and the supernatant was collected, flash-frozen in an ethanol/ dry ice bath, and stored at Ϫ80°C until needed. The membrane-containing pellet was resuspended in 0.5 ml of homogenization buffer containing 0.5% Brij 58 by sonication on ice (three cycles of 5 s at 22 Hz with intervals of 15 s). Resuspended pellet mixtures were then incubated on ice and periodically vortexed to extract membrane-associated proteins. After 30 min, resuspended pellet mixtures were centrifuged at 100,000 ϫ g for 45 min at 4°C and the supernatant (membrane fraction protein extract) was collected. Protein content of cytosolic and membrane fraction extracts was determined using the BCA protein assay (Pierce).
PKC Translocation Assay-Examination of ␣-PKC levels in the cytosolic protein and the membrane protein-containing fractions was used to assess the ability of a given compound to activate ␣-PKC. Briefly, 1 g of recombinant human ␣-PKC (PanVera Corp., Madison, WI), 100 g of cytosolic protein, or 100 g of membrane fraction protein from treated cells was subjected to Tricine SDS-PAGE and transferred to PVDF membrane as described earlier. Blots were then probed with an ␣-PKCspecific rabbit polyclonal IgG (1 g/ml; Biomol Research Laboratories Inc., Plymouth Meeting, PA) followed by 125 I-labeled donkey anti-rabbit IgG (1 g/ml) and visualized using the methods described earlier.
Radiolabeling of hARNT-COS-1 cells were transiently transfected with 9 g of phuARNT in 100 ϫ 20-mm dishes, using LipofectAMINE (Life Technologies, Inc.) as described by the manufacturer. Transfected cells were allowed to recover for 14 h in ␣-MEM supplemented with 10% FBS, 100 IU/ml penicillin, and 0.1 mg/ml streptomycin at 37°C in 94% air, 6% CO 2 . Cells were then rinsed two times with phosphate-free Eagle's minimum essential medium containing 8% dialyzed FBS and placed in this medium. 32 PO 4 3Ϫ was then added to a final concentration of 1 mCi/ml, and cells were promptly incubated with either Me 2 SO or 81 nM PMA for 6 h prior to lysis and hARNT-specific immunoprecipitations.
hARNT-specific Immunoprecipitations-Prior to hARNT-specific immunoprecipitation, cells were rinsed twice with PBS and lysed with MENGPI (25 mM MOPS, 2 mM EDTA, 0.02% NaN 3 , 10% glycerol, 10 mM Na 4 P 2 O 7 , 10 mM NaF, 0.4 mM NaVO 3 , pH 7.5) containing 1% Nonidet P-40 and 1ϫ protease inhibitor mixture. Cell debris was then removed from lysates by ultracentrifugation at 42,000 rpm in a 70Ti rotor (Beckman Instruments, Inc., Columbia, MD). 2B10-Protein G-Sepharose sufficient for immunoprecipitation of hARNT from a single 100 ϫ 20-mm dish of cells required that 20 g of mAb 2B10 be adsorbed to 100 l of a 50% slurry of Protein G-Sepharose conjugate (Sigma) for 1 h in 0.5 ml of MENGPI at 4°C, followed by two washes of 0.6 ml each in MENGPI ϩ 250 mM NaCl. The washed resin was pelleted and resuspended in 0.6 ml of MENGPI ϩ 250 mM NaCl ϩ 1% CHAPS ϩ 1ϫ protease inhibitor mixture, followed by addition of an equal volume of cell lysate and incubation overnight at 4°C. This resin suspension was then washed twice with MENGPI ϩ 500 mM NaCl, and the resin was extruded, by centrifugation at 2665 ϫ g at 4°C for 10 min, through a 1-ml cushion of MENGPI containing 1 M sucrose ϩ 1% Nonidet P-40 ϩ 500 mM NaCl. This sucrose extruded resin was then washed twice with MENGPI containing 0.5% SDS, 2% Nonidet P-40, 500 mM NaCl and washed two more times with MENGPI, followed by preparation of samples for Tricine SDS-PAGE.
Quantitative Assay for 32 P Labeling of Immunoprecipitated hARNT-Immunoprecipitated proteins were resolved by Tricine SDS-PAGE and electrophorectically transferred to PVDF membrane as described earlier. 32 P-Radiolabeled hARNT bands were visualized with a Bio-Rad GS-363 Molecular Imager ® System phosphorimager and quantified with Molecular Analyst version 1.4 software (Bio-Rad).
Statistical Analysis of Data-Error bars in all figures represent standard deviations for a given measurement. Statistical analysis of data was performed using SigmaStat ® version 2.0 software (Jandel Corp., San Rafael, CA) for the Windows ® 95 operating system (Microsoft Corp., Redmond, WA). Statistically significant differences between samples in a given experiment were identified using one-way analysis of variance with Tukey's all pairwise multiple comparisons test (␣ ϭ 0.05). Three to four independent measurements composed each sample set, and data presented are representative of at least two separate experiments.

PMA
Enhances hARNT-driven Transactivation of pMyc3E1bLuc-Previously it had been reported that ARNT is capable of forming a homodimer that binds class B E-boxes and can transactivate class B, E-box-driven reporter constructs (1,24,25). COS-1 cells were transiently transfected with either the pMyc0E1bLuc or pMyc3E1bLuc reporter constructs, containing zero and three class B E-boxes 5Ј to luc ϩ , respectively, alone or in the presence of phuARNT to confirm these observations. Fig. 2A demonstrates that transient expression of hARNT in pMyc3E1bLuc-transfected cells results in a dramatic induction of this class B, E-box-driven reporter construct relative to cells transiently co-transfected with phuARNT and pMyc0E1bLuc. These observations are consistent with those of others (24,25) and demonstrate that the COS-1 transfection system is a valid model in which to examine the regulation of hARNT homodimer activity.
In order to examine the role of PKC in the regulation of hARNT homodimer transactivation activity, COS-1 cells were co-transfected with phuARNT and either pMyc0E1bLuc or pMyc3E1bLuc, followed by treatment with PMA, a PKC activator (31). Fig. 2B demonstrates that an 81 nM PMA treatment of phuARNT and pMyc3E1bLuc-transfected COS-1 cells results in a statistically significant 2.6-fold enhancement in hARNT-driven transactivation of the transfected reporter construct relative to identically transfected cells treated with Me 2 SO alone. Additionally, treatment with 3 M Chel., a PKC inhibitor (32), attenuated this "PMA effect." Together these results suggest that the transactivation potential of the class B E-box binding, hARNT homodimer can be regulated by the serine/threonine kinase PKC.
PMA Specifically Enhances hARNT-driven Transactivation of pMyc3E1bLuc and Activates ␣-PKC in COS-1 Cells-Control experiments utilizing 4-O-Me-PMA, a PMA structural analog that is a poor PKC activator (33), were performed to confirm that the PMA effect observed in Fig. 2B  reporter construct relative to identically transfected cells treated with Me 2 SO alone. Importantly, these results largely eliminate the possibility that the ability of PMA to enhance hARNT-driven transactivation of pMyc3E1bLuc in COS-1 cells is a nonspecific effect of PMA treatment.
Activation of PKC results in the translocation of this kinase from the cytosol to cell membranes (34); therefore, in order to demonstrate that PMA specifically activates PKCs in COS-1 cells, the ability of this compound and 4-O-Me-PMA to induce the translocation of ␣-PKC to membranes was examined. Fig.  3B demonstrates that PMA treatment of COS-1 cells results in a pronounced decrease in the cytosolic levels of this kinase relative to cytosolic ␣-PKC levels in Me 2 SO-treated cells; importantly, cytosolic levels of ␣-PKC in 4-Me-O-PMA-treated COS-1 cells are essentially the same as those in Me 2 SO-treated cells. Fig. 3C demonstrates that PMA specifically activates ␣-PKC translocation to the membrane relative to Me 2 SOtreated COS-1 cells; 4-Me-O-PMA, however, does not activate ␣-PKC in COS-1 cells. Together, the results in Fig. 3 demonstrate that PMA can specifically activate PKCs in COS-1 cells and indicate that the ability of PMA to enhance hARNT-driven transactivation is a direct result of PKC activation.
USF-1-driven Transactivation of pMyc3E1bLuc Is Unaffected by PMA-USF-1 is also capable of forming a homodimer that binds class B E-boxes and can transactivate class B, Ebox-driven, reporter constructs (12,13). COS-1 cells were grown and transiently transfected with either the pMyc0E1bLuc or pMyc3E1bLuc reporter constructs, alone or in the presence of psvUSF1, to demonstrate that USF-1 is able to drive transactivation of pMyc3E1bLuc in the COS-1 model system. Fig. 4A demonstrates that transient transfection of COS-1 cells with psvUSF1 dramatically increases transactivation of pMyc3E1bLuc relative to cells transfected with pMyc3E1bLuc alone. It is important to note that a pronounced difference exists between the extent of luc ϩ reporter induction in cells co-transfected with psvUSF1 and pMyc0E1bLuc relative to pMyc3E1bLuc-and psvUSF1-co-transfected COS-1 cells. These observations are consistent with those reported by others (12,13) and demonstrate that the COS-1 transfection system is a valid model in which to examine the regulation of USF-1 homodimer activity.
COS-1 cells were co-transfected with psvUSF1 and either pMyc0E1bLuc or pMyc3E1bLuc to examine the effect of the PKC activator PMA (31) on the regulation of USF-1 homodimer-mediated transactivation. Fig. 4B demonstrates that an 81 nM PMA treatment of psvUSF1-and pMyc3E1bLuctransfected COS-1 cells has no effect on transactivation of the reporter construct relative to identically transfected cells treated with Me 2 SO alone. Furthermore, treatment with a 3 M dose of Chel. has no effect on USF-1 homodimer-driven transactivation of pMyc3E1bLuc. These results suggest that the transactivation potential of the class B, E-box binding USF-1 homodimer is not regulated by the serine/threonine kinase PKC and that the PMA effect observed with the bHLH-PAS protein hARNT, in Fig. 2, is not due to artifactual activation of pMyc3E1bLuc.
c-Myc-driven Transactivation of pMyc3E1bLuc Is Unaffected by PMA-c-Myc has been demonstrated to drive transactivation of class B, E-box-driven reporter constructs such as pMyc3E1bLuc (10,11). COS-1 cells were grown and transiently transfected with either the pMyc0E1bLuc or pMyc3E1bLuc reporter constructs alone or in the presence of pMK10 59 /Myc to confirm these observations in the COS-1 model system. Fig. 5A demonstrates that transient transfection of pMK10 59 /Myc increases transactivation of the class B, E-box-driven reporter construct pMyc3E1bLuc relative to cells transfected with pMyc3E1bLuc alone. Furthermore, transient co-transfection of pMK10 59 /Myc in pMyc3E1bLuc-transfected cells causes induction of this class B, E-box-driven reporter construct relative to cells transiently co-transfected with pMK10 59 /Myc and pMyc0E1bLuc (Fig. 5A). These observations are consistent with those reported previously (10,11) and demonstrate that the COS-1 transfection system is a valid model in which to examine the regulation of c-Myc-mediated transcriptional activation.
The transactivation potential of c-Myc has been demonstrated to be modulated by mutation of Thr-58 and Ser-62, both of which are phosphorylated in vivo (10). MAP kinases can phosphorylate c-Myc Ser-62 in vitro (21) and are also capable of modulating c-Myc phosphorylation and transactivation potential in vivo (22). Importantly, the serine/threonine kinase or kinase pathway responsible for phosphorylation of c-Myc Thr-58 has yet to be delineated. In order to examine the possibility that the PKC pathway may modulate c-Myc-driven transactivation, COS-1 cells were co-transfected with pMK10 59 /Myc and either pMyc0E1bLuc or pMyc3E1bLuc to investigate the effect of the PKC activator PMA (31) on transactivation of these reporter constructs. Fig. 5B shows that an 81 nM PMA treatment of pMK10 59 /Myc-and pMyc3E1bLuc-co-transfected COS-1 cells has no effect on transactivation of this reporter construct relative to identically transfected cells treated with Me 2 SO alone. Additionally, a 3 M dose of the PKC inhibitor Chel. has no effect on c-Myc-driven transactivation of pMyc3E1bLuc. These results suggest that the transactivation potential of c-Myc is not regulated by the serine/threonine kinase PKC and that the PMA effect observed in Fig. 2, with the bHLH-PAS protein hARNT is not due to artifactual activation of pMyc3E1bLuc.

8-Br-cAMP Differentially Affects hARNT-, USF-1-, and c-Myc-driven Transactivation of pMyc3E1bLuc-hARNT 3 and c-
Myc are both phosphorylated on serine and theronine residues (10), while USF-1 has recently been demonstrated to be a phosphoprotein in vivo (23). In light of the observation that PMA only enhances hARNT-mediated transactivation of pMyc3E1bLuc, it appeared that the transactivation potential of c-Myc and USF-1 may be differentially regulated by various serine/threonine kinase pathways. To further examine this possibility, the PKA activator 8-Br-cAMP (35) was used to probe the ability of this serine/threonine kinase pathway to modulate hARNT-, c-Myc-, and USF-1-driven transactivation of pMyc3E1bLuc in vivo. Fig. 6A demonstrates that 1 nM 8-Br-cAMP treatment of phuARNT and pMyc3E1bLuc co-transfected COS-1 cells increases hARNT-driven transactivation of the luc ϩ gene 1.9-fold relative to identically transfected, Me 2 SO-treated cells. However, Fig. 6 (B and C) demonstrates that 1 nM 8-Br-cAMP treatment of COS-1 cells, in which transactivation of pMyc3E1bLuc is driven by psvUSF1 or pMK10 59 / Myc, has no statistically significant effect on the transactivation of the luc ϩ reporter gene relative to identically transfected, Me 2 SO-treated cells. Together these results suggest that the PKA pathway is capable of modulating hARNT-driven transactivation of pMyc3E1bLuc, but that PKA does not modulate the transactivation potential of the bHLH-ZIP transcription factors c-Myc and USF-1.
The PMA Effect Requires the hARNT Transactivation Domain-One mechanism by which PMA may enhance the hARNT-driven transactivation of pMyc3E1bLuc is by inducing a second transcription factor capable of forming unknown hARNT⅐protein heterodimers or other dimer species that bind 3 G. H. Perdew, unpublished data.  letter (a, b, or c) were not significantly different as determined with ANOVA and Tukey's all pairwise multiple comparisons test (␣ ϭ 0.05).
class B E-boxes and efficiently activate the luc ϩ reporter gene, although, at present, no group has presented evidence that might support the existence of such a mechanism. Nevertheless, the ability of pcDNA3/ARNT-474-Flag to drive transactivation of pMyc3E1bLuc and mediate the PMA effect was examined in order to address this possibility. pcDNA3/ ARNT-474-Flag expresses a FLAG-tagged hARNT truncation mutant in which the COOH-terminal TAD has been deleted. Fig. 7A demonstrates that the hARNT TAD is required for hARNT to drive transactivation of pMyc3E1bLuc and consequently is required for the PMA effect to occur. Fig. 7C demonstrates that the inability of pcDNA3/ARNT-474-Flag to drive transactivation of pMyc3E1bLuc is not due to a failure of this construct to express properly or at a high level in COS-1 cells. Furthermore, panel B demonstrates that transfection of phuA-RNT into COS-1 cells results in proper expression of full-length hARNT.
The current paradigm for ARNT's role in a variety of dimer contexts is that this transcription factor is a central dimerization partner (Fig. 1), which provides half-site specificity and makes a non-dominant contribution to the ability of various ARNT⅐protein heterodimers to efficiently transactivate responsive genes. This paradigm holds for both the AhR⅐ARNT heterodimer (2) and has recently been demonstrated to hold for the HIF1␣⅐ARNT heterodimer (9); in both these heterodimer contexts, ARNT's TAD is not essential for the transactivation of dioxin and hypoxia responsive genes. The results in Fig. 7 suggest that a hARNT homodimer is responsible both for driv-ing transactivation of pMyc3E1bLuc and for the occurrence of the PMA effect in the COS-1 model system. Finally, the results in Fig. 7 strongly suggest that the PMA effect does not involve formation of unknown hARNT⅐protein heterodimer species capable of binding class B E-boxes and driving transactivation of pMyc3E1bLuc. Instead these results indicate that the transactivation potential of hARNT in a hARNT⅐hARNT homodimer context is being modulated by the PKC activator PMA.
PMA Treatments Do Not Alter Cellular hARNT Levels-One candidate mechanism by which PMA may stimulate hARNT homodimer-driven transactivation of pMyc3E1bLuc is simply by increasing cellular hARNT levels. To address this possibility, COS-1 cells were transiently transfected with phuARNT, treated with Me 2 SO or 81 nM PMA, and hARNT levels in these whole cell lysates were quantified (Fig. 8). These analyses demonstrate that cellular hARNT levels in phuARNT-transfected cells are unchanged by PMA treatments (Fig. 8), relative to identically transfected, Me 2 SOtreated cells. Additionally, the observation that mock-transfected COS-1 cells express ARNT at very low levels (Fig. 8) means that endogenous ARNT levels in the COS-1 model system are minimal, making this model well suited for examining the in vivo regulation of ARNT. Importantly, these data demonstrate that the enhanced transactivation effect induced by the PKC activator, PMA, does not occur through a mechanism in which PMA induces increased expression of hARNT in phuARNT-transfected cells.
PMA Treatments Do Not Increase hARNT Phosphorylation-PMA may stimulate hARNT homodimer-driven transactivation of pMyc3E1bLuc by a direct mechanism involving in- were treated with Me 2 SO or 81 nM PMA and hARNT was simultaneously metabolically labeled with 32 PO 4 3Ϫ hARNT metabolically labeled with 32 P was then immunoprecipitated to assess whether PMA treatment increased phosphorylation of hARNT relative to Me 2 SO-treated cells. Fig. 9 (A and B) demonstrates that specific HRP staining of Western blotted, immunoprecipitated hARNT provides an assay by which to quantitate hARNT antigen. The mAb 2B10 used to immunoprecipitate hARNT and detect it on these HRP stained Western blots is specific for a single 19-amino acid peptide corresponding to the hARNT carboxyl-terminal TAD (30); consequently, the HRP signals quantitated in Fig. 9 are proportional to the number of hARNT protein molecules immunoprecipitated. Fig. 9C shows Western-blotted, HRP-stained, immunoprecipitated hARNT and quantification of the immunoprecipitated hARNT antigen. Fig. 9D presents a fluorogram and quantitation of 32 P incorporation into the immunoprecipitated, HRP-visualized, hARNT bands seen in Fig. 9C. In Fig. 9E, the amount of 32 P incorporated per hARNT antigen is shown. These results demonstrate that PMA treatment does not increase the phosphorylation of hARNT relative to Me 2 SO-treated phuARNT-transfected COS-1 cells and that the enhanced transactivation effect induced by the PKC activator, PMA, is not due to increased overall phosphorylation of hARNT.

DISCUSSION
This report demonstrates, for the first time, that hARNT's transactivation potential can be modulated in a dimer context by PMA-and 8-Br-cAMP-mediated stimulation of the PKC and PKA serine/threonine kinase pathways. Furthermore, in the COS-1 model system described here, PMA-and 8-Br-cAMPmediated stimulation of the PKC and PKA serine/threonine kinase pathways differentially affect the transactivation potentials of the bHLH-PAS transcription factor hARNT and the bHLH-ZIP transcription factors c-Myc and USF-1. Previous studies demonstrating that ARNT's TAD is nondominant in the AhR⅐ARNT (2) and HIF1␣⅐ARNT (9) heterodimer contexts supported a model in which ARNT's role, in these various heterodimer contexts, is primarily limited to providing DNA binding, half-site specificity in a combinatorial type control mechanism. The data presented here are not incompatible with this paradigm, but suggest that modulation of ARNT's transactivation potential may provide an important additional level of control by which target genes may be transcriptionally regulated by various ARNT-containing heterodimers.
Previous work examining the role of ARNT's transactivation domain in the AhR⅐ARNT heterodimer context supports the idea that ARNT's transactivation potential makes an important contribution to the transcriptional regulation of target genes. Reisz-Porszasz et al. (7) previously generated a plasmid construct encoding the bHLHAB ARNT protein in which the bHLH, PAS A, and PAS B functional domains of ARNT are intact, but the carboxyl-terminal ARNT TAD has been deleted and a small deletion has been made at the amino terminus. Experiments performed by Reisz-Porszasz et al. (7) in the ARNT-deficient Hepa-1-c4 (B13NBiil) cell line demonstrated that this peptide is essentially unimpaired in its ability to dimerize with the AhR and bind DREs. However, AhR⅐bHLHAB ARNT heterodimers were only capable of activating a DRE-driven chloramphenicol acetyltransferase reporter with approximately 50% of the efficiency of the wild-type AhR⅐ARNT heterodimer (7) implying that ARNT's TAD contributes to the transcriptional regulation of AhR⅐ARNT heterodimer-responsive genes. Similarly Ko et al. (2) found that the ability of GAL4DBD-AhR⅐ARNT C to transactivate the GAL4-responsive reporter pGE-CAT in response to TCDD was impaired by roughly 50% relative to the ability of GAL4DBD-AhR⅐ARNT to transactivation of this reporter in response to TCDD. Additionally, Ko et al. (2) examined CYP1A1 mRNA levels in ARNT-deficient BP r c1 cells transiently expressing the ARNT C peptide, in which the ARNT TAD has been deleted. Although Ko et al. did not quantify these Northern blots, it appears that a substantial decrease in CYP1A1 transcription occurs in BP r c1 cells when an AhR⅐ARNT TAD dimer is formed instead of a AhR⅐ARNT wt dimer. Importantly, these apparent decreases in AhR⅐ARNT C heterodimer-mediated transactivation (2) are similar to the observations of Reisz-Porszasz et al. (7). Together these observations lend credence to the notion that modulation of ARNT's transactivation potential may provide an important, addi- l of hARNT IP. B, hARNT antigen calibration plot. hARNT antigen quantified in A was plotted as a function of hARNT IP volume and analyzed using linear regression as described under "Experimental Procedures." C, quantification of hARNT antigen immunoprecipitated from 32 P-radiolabeled, Me 2 SO or 81 nM PMA-treated, phuARNT-transfected COS-1 cells. 75 l of 32 P-hARNT immunoprecipitated from transfected, treated cells was Western-blotted, and probed with mAb 2B10 followed by biotinylated-GAM IgG, HRP-streptavidin, band visualization, and antigen quantification as described under "Experimental Procedures." D, quantification of 32 P-labeled hARNT immunoprecipitated from 32 P-radiolabeled, Me 2 SO or 81 nM PMA-treated, phuARNT-transfected COS-1 cells. 32 P-Labeled hARNT in C was quantified as described under "Experimental Procedures." E, 32 P incorporation per hARNT antigen for Me 2 SO or 81 nM PMA-treated, phuARNT-transfected COS-1 cells. tional level of control for the transcriptional regulation of target genes responsive to various ARNT-containing heterodimers.
The observation that PMA can enhance the transactivation potential of bHLH-PAS proteins is not without precedent, as PMA has previously been reported (26,27) to enhance hAhR⅐hARNT heterodimer-mediated transactivation in mammalian cell lines stably transfected with a DRE-driven luciferase reporter construct. Importantly, this effect was shown not to be artifactual, appeared to be cell line-independent, and did not involve any of the following parameters: alteration of cytosolic hAhR or hARNT levels, TCDDinduced down-regulation of the hAhR, alteration of nuclear levels of the hAhR, or hAhR⅐hARNT heterodimer DRE binding activity (26,27). Significantly, the observation that hARNT levels in phuARNTtransfected COS-1 are essentially unaltered by 81 nM PMA treatment (Fig. 8) is consistent with these previous observations (26). While it is possible that PMA's effects on the hARNT-driven transactivation of pMyc3E1bLuc may involve increased homodimerization of this bHLH-PAS transcription factor, the observation that PMA does not increase hAhR⅐hARNT dimer activity (26,27) suggests PMA's effects on bHLH-PAS transcription factor-driven transactivation do not occur through this mechanism. Importantly, it was concluded that one mechanism by which PMA may enhance hAhR⅐hARNT heterodimer-mediated transactivation is by altering the ability of the hAhR⅐hARNT complex to form a fully functional transcriptional complex, perhaps by altering the recruitment or activity of a specific co-activator associated with these bHLH-PAS transcription factors.
The co-activator CBP/p300 has been shown to interact specifically with ARNT's TAD both in vitro and in vivo, but not with the AhR's TAD (36). Interestingly, infection of cells with adenovirus E1A inhibits the ability of the TCDD-induced AhR⅐ARNT heterodimer to transactivate responsive genes (37). Additionally, no existing data suggest a direct interaction between E1A and ARNT (36). Consequently, it has been postulated that E1A inhibits AhR⅐ARNT heterodimer-mediated transactivation of responsive genes by titrating cellular CBP/ p300 away from ARNT's TAD and into pathways involved in viral replication (36). These observations further suggest that modulation of ARNT's transactivation potential may provide an important, additional level of control for the transcriptional regulation of target genes responsive to various ARNT-containing heterodimers. hARNT 3 and c-Myc are both phosphorylated on serine and threonine residues (10), while USF-1 has recently been demonstrated to be a phosphoprotein in vivo (23). Although MAP kinases can phosphorylate Ser-62 of c-Myc in vitro (21), and are also capable of modulating c-Myc phosphorylation and transactivation potential in vivo (22), the serine/threonine kinase or kinase pathway responsible for phosphorylation of c-Myc Thr-58 has yet to be delineated. The data presented here suggest that the PKC and PKA pathways are not involved in modulating the transactivation potential of the bHLH-ZIP transcription factors USF-1 and c-Myc, but that instead stimulation of these kinases differentially affects the transactivation potential of the bHLH-PAS and bHLH-ZIP transcription factors. These differential effects may eventually prove useful for identifying co-activators or other proteins that contribute to the ability of PMA to enhance the transactivation potential of the bHLH-PAS proteins or genes specifically activated by the hARNT homodimer in vivo.
In summary we report for the first time, that ARNT's transactivation potential can be modulated in a dimer context by PMA-and 8-Br-cAMP-mediated stimulation of the PKC and PKA serine/threonine kinase pathways. Additionally, in the COS-1 model system, PMA and 8-Br-cAMP treatments differentially affect the transactivation potentials of the bHLH-PAS transcription factor hARNT and the bHLH-ZIP transcription factors c-Myc and USF-1. The ability of PMA to enhance hARNT-driven transactivation was specific, and based on the ability of PMA to activate ␣-PKC, a ubiquitous isoform representative of the phorbol ester-responsive cPKC class (38), this activation appears to be a direct result of PKC activation in the COS-1 model system. The PMA-induced, enhanced hARNT homodimer-mediated transactivation of pMyc3E1bLuc reported here does not involve increased, direct, PKC-mediated phosphorylation of hARNT or alteration of cellular hARNT expression. These observations suggest that recruitment of a coactivator specific to the hARNT homodimer may be one possible mechanism by which the PMA effect occurs; alternatively, the PMA effect could be due to increased hARNT homodimer formation. Finally, we postulate that modulation of ARNT's transactivation potential may provide an important, additional level of control for the transcriptional regulation of target genes responsive to various ARNT-containing heterodimers, a conclusion that is consistent, and not incompatible, with the conclusions and findings of others (2,7,36).