Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

ARNT is a bHLH factor that contains a highly conserved PAS domain (Hoffman et al., 1991) which has been identified in dozens of signal transduction molecules in animals, plants, and prokaryotes (Gu et al., 2000). The formation of ARNT heterodimers involves the HLH and PAS domains (Reisz-Porszasz et al., 1994). In response to exposure to selected polycyclic aromatic hydrocarbons and halogenated aromatic hydrocarbons, ARNT forms a heterodimer with the AhR after the ligand-dependent translocation of AhR to the nucleus (Reyes et al., 1992; Hankinson, 1995). Subcellular fractionation and immunohistochemical studies have shown that the AhR resides in the cytoplasm bound to two molecules of the 90-kDa heat-shock protein (Perdew, 1988) and a single molecule of AhR interacting protein/hepatitus B virus X-associated protein/AhR-activated 9 (AIP/XAP2/ARA9; Carver and Bradfield, 1997; Ma and Whitlock, 1997; Meyer et al., 1998). In contrast to the AhR, ARNT is a constitutively nuclear protein (Hord and Perdew, 1994). After translocation to the nucleus, AhR heterodimerizes with ARNT, and the AhR-ARNT heterodimer binds to XREs, causing transactivation of genes encoding xenobiotic metabolizing enzymes (Hankinson, 1995).

The HIF1 protein is a second bHLH-PAS protein; it is responsible for sensing the chemical environment and heterodimerizing with ARNT (Gu et al., 2000). Through an analogous signal transduction pathway to the AhR, reduced cellular oxygen tension leads to increased stability of HIF1 and heterodimerizes with ARNT. Like the AhR, HIF1 is constitutively cytoplasmic; however, upon stabilization, it translocates to the nucleus (Guillemin and Krasnow, 1997). The HIF1-ARNT heterodimer binds to hypoxia response elements and up-regulates genes involved in adaptation to hypoxia (Wang et al, 1995). The characterization of the HIF1-ARNT heterodimer established the role of ARNT as a central dimerization partner with other bHLH-PAS proteins that are responsible for sensing and adapting to environmental change.

In addition to heterodimerizing with the AhR and HIF1, ARNT heterodimerizes with endothelial PAS domain protein-1 and Sim (Schmidt et al., 1996). Endothelial PAS-1-ARNT heterodimers are involved in response to hypoxia and the Sim-ARNT heterodimer is involved in mammalian central midline development (Yamaki et al., 1996). Recently, an ARNT2 factor has been described that shares 63% identity with ARNT (Drutel et al., 1996) and can also heterodimerize with Sim (Michaud et al., 2000). ARNT2 has been implicated in nervous system development (Drutel et al., 1999) as well as the development of neuroendocrine lineages (Michaud et al., 2000). In addition to heterodimerizing with numerous factors, ARNT is also able to homodimerize; this homodimer can bind to and activate class B E-box-driven reporter constructs (Gupta et al., 1993; Antonsson et al., 1995). Presently, it is known that E-box binding factors are important in the regulation of cell differentiation and proliferation (Jan and Jan, 1993; Kadesch, 1993; Weintraub, 1993; Dorshkind, 1994). Although the role of the ARNT homodimer in developmental processes is unknown, ARNT expression is important for normal development. Mouse embryos lacking ARNT died after 10.5 days of development (Maltepe et al., 1997). Lack of ARNT expression resulted in defective angiogenesis of the yolk sack and branchial arches (Maltepe et al., 1997). The abnormalities were characteristic of the angiogenic abnormalities for mice lacking vascular endothelial growth factor (Carmeliet et al., 1996a; Ferrara et al., 1996) or tissue factor (Carmeliet et al., 1996b).

Phosphorylation of transcription factors is a regulatory mechanism that can control signal transduction in several potential ways. Protein phosphorylation can modulate DNA binding activity, translocation to the nucleus, and protein-protein interactions. A previous investigation revealed the importance of AhR-ARNT phosphorylation status on activation of the heterodimer into to its active DNA-binding form (Pongratz et al., 1991). It was reported that incubating transformed cytosolic AhR with acid phosphatase abolished the ability of the receptor to bind XREs. Currently, it is known that DNA binding by human and mouse AhR-ARNT heterodimers requires phosphorylation of both proteins, whereas formation of AhR-ARNT heterodimers requires phosphorylation of only ARNT (Berghard et al., 1993). Putative AhR phosphorylation sites have been localized to two regions in the C-terminal half of the mouse protein. One putative phosphorylated region is within or adjacent to a DNA-binding repressor domain, which prevents constitutive XRE binding by AhR-ARNT complexes (Dolwick et al., 1993); a second putative phosphorylated region is located within the glutamine-rich C terminus (Mahon and Gasiewicz, 1995). The phosphorylation status of HIF1 has also been shown to modulate HIF1-dependent activity.

Activation of signal transduction by a protein can be controlled at the structural level as a result of the addition or removal of a phosphate group(s) (Roach, 1991). Consequently, an array of phosphorylation sites could represent a distribution of differentially activated proteins that vary in their functional abilities. Previously, Tsai and Perdew (1997) demonstrated that a shift in pI of ARNT toward the basic occurs during dimerization with AhR and/or after AhR-ARNT DNA binding. This shift in pI occurs during dimerization and/or after binding to DNA and was hypothesized to represent a change in the phosphorylation status of ARNT. To date, there has been no report characterizing the functional significance of site-specific phosphorylation sites on ARNT or the influence of TCDD-dependent heterodimerization with the AhR on the phosphorylation pattern of ARNT. The current study has identified an ARNT phosphorylation site in the PAS domain and has examined its influence on AhR-ARNT signaling and ARNT homodimer signaling. Additionally, the phosphorylation pattern of ARNT after heterodimerization with the AhR was examined.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. The ARNT expression vector pSV-Sport1-hARNT was obtained from Christopher Bradfield at the University of Wisconsin (Madison, WI) (Dolwick et al., 1993). The hARNT/474-Flag construct was generated by Jo Tsai (Tsai, 1997) and the pCI/hAhR/Flag was generated with the same method by M. Pray-Grant in our laboratory. The Hepa 1 c4 cell line was obtained from Oliver Hankinson (University of California, Los Angeles, Los Angeles, CA). COS-1 cells were purchased from the American Type Culture Collection (Manassas, VA). FBS was purchased from Hyclone Lab (Logan, UT). Acrylamide, ammonium persulfate, acetonitrile, hydrochloric acid, and 100-µM thin-layer chromatography plates were purchased from Fisher (Pittsburgh, PA). Vent DNA polymerase, restriction endonucleases, and other DNA modifying enzymes were purchased from New England Biolabs (Beverly, MA). [³²P]Orthophosphate (370 MBq/ml; 10 mCi/ml) and goat anti-mouse [¹²⁵I]IgG were purchased from Amersham Pharmacia Biotech (Piscataway, NJ). pcDNA3 vector was purchased from Invitrogen (Carlsbad, CA). The Vector VIP substrate kit for immunoblots was purchased from Vector Laboratories Inc. (Burlingame, CA). Oligonucleotides for polymerase chain amplifications were purchased from Operon (Alameda, CA). LipofectAMINE reagent, opti-MEM, and Escherichia coli DH5 were purchased from Life Technologies (Gaithersburg, MD). Sequelon-AA arylamine disks were purchased from Perseptive Biosystems (Bedford, MA). Nitrocellulose and SDS were purchased from Bio-Rad (Hercules, CA). Tricine, glycine, Tris, and CHAPS were purchased from Research Organics (Cleveland, OH). The luciferase assay system, sequencing grade trypsin, and restriction endonucleases were purchased from Promega (Madison, WI). The bicinchoninic acid protein assay reagents were purchased from Pierce (Rockford, IL). TCDD was obtained from Steven Safe (Texas A & M University, College Station, TX). All other chemicals were purchased from Sigma (St. Louis, MO).

Plasmid Construction. hARNT/Flag cDNA was amplified by polymerase chain reaction (PCR) using pSV-Sport1/hARNT as a template. The forward primer, including the start site (5'- CCCAAGCTTGGGATGGCGGCGACTACTGCCAACCCC-3') and an ARNT/Flag reverse primer (5'- CCGCTCGAGCGGCTACTTGTCATCGTCGTCCTTGTAG- TCTTCTGAAAAGGGGGGAAACAT-3'), were used to amplify and add a Flag sequence to the 3' terminus of the hARNT gene using standard PCR techniques. The resulting PCR product was gel-purified, digested with HindIII and XhoI, and subcloned into the HindIII/XhoI sites of pcDNA3. All plasmids were propagated in Escherichia coli strain DH5. The nucleotide sequence of hARNT/Flag, was confirmed with an automated DNA sequencer at the DNA Core Facility at Pennsylvania State University.

Site-Directed Mutagenesis. Site-directed mutagenesis of hARNT constructs was performed using the Stratagene Quickchange site-directed mutagenesis procedure, according to the manufacturers instructions. The nucleotide sequence of hARNT mutants was confirmed with an automated DNA sequencer at the DNA core facility at the Pennsylvania State University.

Transient Transfection with hARNT and [³²P]Orthophosphate Labeling of Transfected Cells. COS-1 cells were cultured in -MEM supplemented with 10% FBS, 100 IU/ml penicillin, and 0.1 mg/ml streptomycin at 37°C in 95% air/5% CO₂. COS-1 cells (80% confluent) were transiently transfected with pcDNA3/hARNT/474-Flag or pcDNA3/hARNT/Flag using the LipofectAMINE procedure. For each transfection, 9 µg of DNA/30 µl of LipofectAMINE complex was added to cells in a 10-cm² plate with serum-free opti-MEM. After 8 h of incubation with the DNA-LipofectAMINE complexes, the cells were rinsed twice with PBS and then incubated with -MEM supplemented with 10% FBS, 100 IU/ml penicillin, and 0.1 mg/ml streptomycin at 37°C for 24 h. Transfected COS-1 cells were rinsed three times with phosphate-free Eagle's MEM, and then preincubated in phosphate-free Eagle's MEM supplemented with 10% dialyzed FBS, 100 IU/ml penicillin, and 0.1 mg/ml streptomycin at 37°C. After the 1-h preincubation, the media was replaced and cells were incubated with [³²P]orthophosphate (1 mCi/ml) for 4 h. After aspirating the [³²P]orthophosphate media, cells were rinsed with PBS, scraped from the dishes, and lysed with 1% NP-40 in 25 mM MOPS, 2 mM EDTA, 0.02% NaN₃, and 10% glycerol, pH 7.4 (MENG) containing 10 mM sodium pyrophosphate, 20 mM sodium molybdate, 10 mM sodium fluoride, 0.4 mM sodium vanadate, (MENGPI) and a protease inhibitor cocktail (2 mM 4-(2-aminoethyl)benzenesulfonyl fluoride, 2 µM trans-epoxysuccinyl-L-leucylamido(4-guanidino)butane, 2 µM bestatin, 2 µM leupeptin, 2 µM aprotinin, and 2 mM sodium EDTA) over 15 min on ice. The cell homogenate was centrifuged at 105,000g for 1 h and the supernatant was collected for immunoprecipitation.

In Vitro Transformation of AhR with TCDD. AhR can be induced to form a heterodimer with ARNT when treated with TCDD in vitro as well as in vivo. To examine the influence of hAhR heterodimerization on the phosphorylation pattern of hARNT, COS-1 cells (80% confluent) in 10-cm² plates were cotransfected with 4.5 µg of pSV-Sport1/hARNT and 4.5 µg of pCI/hAhR/Flag or, as a reference control, with 9 µg of pcDNA3/hARNT/Flag using the LipofectAMINE procedure. Twenty-four hours after transfection, cells were washed three times with phosphate-free Eagle's MEM, and then preincubated in phosphate-free Eagle's MEM supplemented with 10% dialyzed FBS, 100 IU/ml penicillin, and 0.1 mg/ml streptomycin at 37°C. After the 1-h preincubation, the media were replaced and cells were incubated with [³²P]orthophosphate (1 mCi/ml) for 4 h. During the final 1 h of the [³²P]orthophosphate labeling, COS-1 cells cotransfected with hAhR/Flag and hARNT were treated with 10 nM TCDD (dissolved in DMSO, 1 µg/ml) and cells transfected with ARNT/Flag alone were treated only with carrier (DMSO, 1 mg/ml). After aspirating the [³²P]orthophosphate media, cells were rinsed with PBS, and lysed in MENGPI/1% NP-40 containing 1× protease inhibitor on ice for 15 min with frequent vortexing. To isolate nuclei, the MENGPI/1% NP-40 lysate was centrifuged at 1,000g at 4°C for 15 min. Nuclei were washed three times with MENGPI before extraction with 500 mM NaCl, in MENGPI for 1 h on ice. Lysed nuclei were centrifuged at 105,000g at 4°C for 1 h and the supernatant was taken as the high-salt nuclear extract for coimmunoprecipitation. In the control experiment, the MENGPI/1% NP-40 lysate was centrifuged at 105,000g at 4°C for 1 h and the supernatant was collected for immunoprecipitation.

Immunoprecipitation and Gel Purification of Radiolabeled hARNT. Nuclear extracts form COS-1 cells cotransfected with hAhR/Flag and hARNT were coimmunoprecipitated with 50 µl of anti-Flag M2 mAb affinity gel. Similarly, cytosolic lysates from hARNT/474-Flag and hARNT/Flag-transfected COS-1 cells were immunoprecipitated with 50 µl of anti-Flag M2 mAb affinity gel. The final buffer composition during immunoprecipitation of hARNT/Flag was MENGPI, 1% CHAPS, and 250 mM NaCl, pH 7.4, and 1× protease inhibitor cocktail. The final buffer composition during immunoprecipitation of the hAhR/Flag-hARNT complexes extracted from nuclei was MENGPI, 250 mM NaCl, pH 7.4, and 1× protease inhibitor cocktail. In separate experiments, designed to demonstrate the specificity of the Flag-tagged hARNT proteins for the anti-Flag M2 mAb affinity gel, the gel was preincubated for several hours with 90 nmol of the Flag peptide (DYKDDDK) before the cell lysate was added. Immunoprecipitations were run overnight on ice. The immunoprecipitations were washed twice with 1 ml of MENGPI + 500 mM NaCl. For cytosolic extracts, the gel was then incubated with MENGPI + 500 mM NaCl + 2% NP-40 + 0.5% SDS for 1 h, with rocking on ice, and washed twice with MENGPI before gel purification by SDS-PAGE on an 8% Tricine gel. For nuclear extracts, the gel was incubated for 1 h with MENGPI + 500 mM NaCl, and then washed twice with MENGPI before SDS-PAGE on an 8% Tricine gel. Gels were transferred to nitrocellulose at 15 V for 3 h at 4°C in a Genie Blotting unit (Idea Scientific, Minneapolis, MN).

Phosphopeptide Mapping and Phosphoamino Acid Analysis of hARNT. For phosphopeptide mapping, polyacrylamide gels containing radiolabeled hARNT/Flag were transferred to nitrocellulose. After autoradiography, the hARNT/Flag band was excised from the membrane, transferred to a microcentrifuge tube, washed five times with 1 ml of water, and immediately blocked with 0.5% polyvinylpyrrolidone-360 in 100 mM acetic acid for 30 min at 37°C. The membrane was then washed five times with 1 ml of water and then once with freshly made 0.05 M ammonium bicarbonate. The membrane was resuspended in 200 µl of 0.05 M ammonium bicarbonate and 10 µg of tosylphenylalanine chloromethyl ketone-treated trypsin was added and incubated at 37°C for 16 h. At the end of the digestion, 300 µl of water was added to the digestion, the membrane was vortexed for 30 s, the digestion was centrifuged at 10,000g for 10 min, transferred to a new microcentrifuge tube, and the digestion was dried in a speed-vac. The dried sample was resuspended in 300 µl of electrophoresis buffer I, pH 1.9 (15% acetic acid, 5% formic acid), dried, resuspended in 10 µl of pH 1.9 buffer, and spotted onto a 20- × 20-cm cellulose TLC plate and electrophoresed at 1.2 kV for 35 min with a Hunter Thin Layer Electrophoresis System model HTLE-7000 (CBS Scientific, INC., Del Mar, CA). Plates were then air-dried, subjected to ascending chromatography in the second dimension for 16 h in phosphochromatography buffer (37.5% n-butanol, 25% pyridine, 7.5% acetic acid, and 30% deionized water), air dried, and exposed to film (Boyle et al., 1991). Phosphopeptides were visualized by autoradiography at 80°C using an intensifying screen.

Phosphate Release by Manual Edman Degradation. To determine the number and position of phosphorylated residues of a selected phosphopeptide, manual Edman degradation was performed by the method of Sullivan and Wong (1991). Phosphopeptides were scraped from that TLC plate and eluted from the cellulose as described by Boyle et al. (1991), dried down, dissolved in 25 µl of 50% acetonitrile, and covalently linked to an arylamine-Sequelon disk. The disk was then subjected to Edman degradation by treatment for 10 min at 50°C with 0.5 ml of coupling reagent [methanol/water/triethylamine/phenylisothiocyanate; 7:1:1:1 (v/v/v/v)], followed by six washes with 1 ml of methanol, and reheated at 50°C for 6 min with 0.5 ml of trifluoroacetic acid to cleave the N-terminal residue. The disk was then washed with 1 ml of trifluoroacetic acid and 42.5% phosphoric acid (9:1) and combined with the 0.5 ml of trifluoroacetic acid. The combined disk washes were dried under a stream of nitrogen, 10 ml of scintillation cocktail was added, and then they were quantitated for ³²P. The disk was subsequently washed five times with 1 ml of methanol before the next cycle was started.

Functionality of hARNT Mutants and Protein Expression. Comparison of the ability of hARNT and hARNT mutants to trans-activate XRE-driven luciferase activity was made in ARNT-deficient Hepa-1 c4 cells using the LipofectAMINE procedure. Hepa-1 c4 cells (80% confluent) in six-well plates were cotransfected with 2.5 ng of pSV-Sport1/hARNT, 100 ng of pGudLuc 6.1, 100 ng of pSV--galactosidase, and the total amount of DNA was equalized to 1.5 µg with pSV-Sport1. Control transfections contained 100 ng of pGudLuc 6.1, 100 ng of pSV--galactosidase, and the total amount of DNA was equalized to 1.5 µg with empty pSV-Sport1. Twenty-four hours after transfection, cells were incubated with either DMSO (1 µl/ml) or 10 nM TCDD (1 µl/ml) for 8 h. At the end of the exposure period, cells were lysed with a lysis buffer (containing 25 mM Tris-phosphate, 2 mM dithiothreitol, 2 mM CDTA, 10% glycerol, and 1% Triton X-100) and assayed for luciferase activity using the Promega luciferase assay system. Luciferase activity was measured with a Turner Instruments TD-20e luminometer (Sunnyvale, CA) and is expressed as RLUs. Luciferase activity was normalized against -galactosidase activity and hARNT expression levels.

Statistical Analyses. Values are reported as means ± S.E.M. ANOVA with Duncan's multiple range test were performed to test for differences between treatment groups ( = 0.05). All statistical tests were performed with SAS (ver. 8; SAS Institute, Cary, NC) on an IBM microcomputer.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

hARNT Is a Phosphoprotein. Flag-tagged truncated hARNT containing the N-terminal 474 amino acid residues (hARNT/474-Flag), as well as a Flag-tagged full-length hARNT (hARNT/Flag) (Fig. 1A) in transfected COS-1 cells incubated with [³²P]orthophosphate, demonstrated that hARNT is a phosphoprotein and that anti-Flag M2 affinity gel specifically immunoprecipitates each of these Flag-tagged proteins (Fig. 1B and 1C). Phosphoamino acid analysis was performed to determine whether hARNT is phosphorylated on serine, threonine, and tyrosine residues. This analysis demonstrated that both hARNT/474-Flag and hARNT/Flag are predominantly phosphorylated on serine residues (Fig. 1D).

View larger version (45K): [in this window] [in a new window]   Fig. 1.   hARNT is a phosphoprotein that is predominantly phosphorylated on serine residues. A, Flag-tagged hARNT cDNAs used for [32P]orthophosphate labeling studies. hARNT/474-Flag is comprised of the N-terminal 474 amino acids of hARNT and contains the bHLH and PAS domains. B and C, expression of and phosphorylation of wild-type pcDNA3/hARNT/Flag and pcDNA3/hARNT474/Flag in COS-1 cells. COS-1 cells were transiently transfected with hARNT constructs and labeled for 4 h, with 1 mCi/ml [32P]orthophosphate. COS-1 cytosol was immunoprecipitated with anti-Flag M2 affinity gel, resolved by SDS-PAGE on an 8% Tricine gel, transferred to nitrocellulose, probed with an anti-Flag specific mAb and autoradiographed as described under Experimental Procedures. B and C, lanes 2 and 4, demonstrate immunodetection of the hARNT proteins with the anti-Flag specific mAb antibody and the autoradiograph of the orthophosphate labeled hARNT proteins, respectively. The hARNT/Flag proteins are labeled with an arrowhead. Lanes 1 and 3 demonstrate the specificity of the anti-Flag M2 gel for the Flag-tagged hARNT proteins; the anti-Flag M2 gel was incubated with Flag peptide before the cell lysates were added. D, phosphoamino acid analysis of pcDNA3/hARNT/Flag and pcDNA3/hARNT474/Flag from [32P]orthophosphate labeled (1 mCi/ml) COS-1 cells. SDS-PAGE-purified hARNT was hydrolyzed in 5.7 N HCl at 110°C for 1 h and mixed with phosphoamino standards [phospho-serine (pS); phospho-threonine (pT); and phospho-tyrosine (pY)] and vacuum dried. Hydroxylates were separated electrophoretically in two dimensions on cellulose plates, stained with ninhydrin to visualize standards (shown by outlined circles), and autoradiography was performed.

The Phosphorylation Pattern of hARNT Is Unaltered after hAhR Heterodimerization. To assess the influence of heterodimerization with the AhR on the phosphorylation pattern of ARNT, COS-1 cells were either cotransfected with hAhR/Flag and hARNT and treated with 10 nM TCDD in DMSO or transfected with hARNT/Flag alone and treated only with DMSO. Twenty-four hours after transfection, COS-1 cells were [³²P]orthophosphate-labeled for 4 h. COS-1 cells cotransfected with hAhR/Flag and hARNT were treated with 10 nM TCDD during the final 1 h of [³²P]orthophosphate labeling and the hAhR/Flag-hARNT complex was coimmunoprecipitated with anti-Flag M2 mAb affinity gel from nuclear extracts. In the reference control, COS-1 cells transfected with hARNT/Flag alone were treated with DMSO during the final 1 h of [³²P]orthophosphate labeling and hARNT/Flag was immunoprecipitated with anti-Flag M2 affinity gel from cytosolic lysate. Two-dimensional phosphopeptide maps were then generated to compare the phosphorylation pattern of hARNT heterodimerized with the hAhR or in its unheterodimerized form. Based on visual analysis of two-dimensional phosphopeptide maps, it seems that the phosphorylation pattern of hARNT was unaltered whether it was heterodimerized with the hAhR or not (Fig. 2). However, because a stoichiometric comparison was not made between phosphopeptides on these two maps, differences between the stoichiometry of particular hARNT sites in unknown.

View larger version (52K): [in this window] [in a new window]   Fig. 2.   Heterodimerization with the hAhR does not alter the hARNT phosphorylation pattern. A, two-dimensional phosphopeptide map of hARNT/Flag. COS-1 cells were transiently transfected with 9 µg of pcDNA3/hARNT/Flag in 10-cm2 dishes. Cells were labeled for 4 h with 1 mCi/ml [32P]orthophosphate. Cytosolic lysates were prepared as described under Experimental Procedures and hARNT/Flag was immunoprecipitated with anti-Flag M2 affinity gel. hARNT/Flag was resolved by SDS-PAGE on an 8% Tricine gel, transferred to nitrocellulose, and digested with trypsin. Resulting peptides were resolved in two dimensions by electrophoresis and chromatography on TLC plates, and phosphorylated peptides were visualized by autoradiography. B, two-dimensional phosphopeptide map of hARNT that was coimmunoprecipitated with AhR/Flag from nuclear extract. COS-1 cells were transiently transfected with 4.5 µg of pSV-Sport1/hARNT and 4.5 µg of pCI/hAhR/Flag in 10-cm2 dishes and labeled for 4 h with 4 mCi/ml [32P]orthophosphate and during the final hour of [32P]orthophosphate labeling, 10 nM TCDD was introduced to the culture media. Nuclear extracts were prepared as described under Experimental Procedures and AhR/Flag-ARNT complexes were immunoprecipitated with anti-Flag M2 gel. SDS-PAGE purified hARNT was transferred to nitrocellulose and digested with trypsin. Resulting tryptic peptides were resolved in two dimensions by electrophoresis and chromatography on TLC plates, and the phosphorylated peptides were visualized by autoradiography.

Serine 348 Is Phosphorylated. Solid-phase sequencing of tryptic phosphopeptides isolated from a two-dimensional phosphopeptide map of hARNT/474-Flag (Fig. 3A) revealed that the predominant tryptic phosphopeptides were labeled either three or six amino acid residues away from the N terminus (Figs. 3B-I). The hARNT/474-Flag construct was chosen because it expresses the N-terminal amino acid sequence through the end of the PAS domain. Table 1 summarizes the list of candidate hARNT tryptic peptides that have a serine residue at either three or six amino acid residues from the N terminus. Site-directed mutagenesis was used to change each of these candidate serines to an alanine in the hARNT/474-Flag and hARNT/Flag cDNAs. This screening procedure revealed that serine 348 (S348) is phosphorylated in the hARNT/474-Flag construct (Fig. 4A) and in the full-length hARNT/Flag construct (Fig. 4B). In addition, this screening procedure indicated that not any of the other candidate serine residues in Table 1 are phosphorylated. Figure 4A illustrates the disappearance of a cluster of phosphopeptides on the hARNT/474-Flag S348A map. This cluster of missing phosphopeptides on the hARNT/474-Flag map may represent incomplete digestion products that each contain S348. Alternatively, the cluster may represent different oxidation states of either the cysteine residue and/or methionine residue in the tryptic peptide, which contains S348. This cluster was not apparent in the hARNT/Flag map; rather, the S348 containing peptide migrated as a single peptide evidenced by the disappearance of a single spot (Fig. 4B). Changing S348 to a threonine (S348T) did not restore phosphorylation of the missing peptide (Fig. 4B). Phosphoamino acid analysis of the tryptic phosphopeptide that contains S348 demonstrated that this phosphopeptide is exclusively phosphorylated on serine (Fig. 4C).

View larger version (31K): [in this window] [in a new window]   Fig. 3.   Manual Edman degradation of tryptic phosphopeptides to determine cycles of [32P]orthophosphate release. A, tryptic phosphopeptides were generated from pcDNA3/hARNT/474-Flag- transfected COS-1 cells and were separated by electrophoresis and chromatography on a TLC plate. B to I, individual phosphopeptides were eluted from the cellulose and resuspended in 50% acetonitrile. Phosphopeptides were covalently linked to arylamine disks and subjected to a modification of Edman degradation. The counts of [32P]orthophosphate in the released amino acid were determined after each cycle by Cerenkov counting.

View larger version (72K): [in this window] [in a new window]   Fig. 4.   Serine 348 is a phosphorylation site on the hARNT protein. A, two-dimensional tryptic phosphopeptide map of pcDNA3/hARNT/474-Flag and pcDNA3/hARNT/474-Flag S348A, which contains a serine-to-alanine substitution at position 348. COS-1 cells transiently expressing pcDNA3/hARNT/474-Flag cDNAs were labeled for 4 h with 1 mCi/ml [32P]orthophosphate, and cytosol was immunoprecipitated with anti-Flag M2 affinity gel. SDS-PAGE-purified proteins were transferred to nitrocellulose and digested with trypsin. Resulting peptides were resolved by electrophoresis and chromatography, and the phosphorylated peptides were visualized by autoradiography. The filled arrowhead identifies the phosphopeptides that contain S348 and the open arrowheads identify the location of the missing phosphopeptides. B, two-dimensional tryptic phosphopeptide maps of pcDNA3/hARNT/Flag (S348) and mutant pcDNA3/hARNT/Flag proteins (S348A and S348T). COS-1 cells transiently expressing hARNT/Flag cDNAs were labeled for 4 h with 1 mCi/ml [32P]orthophosphate, and cytosol was immunoprecipitated with anti-Flag M2 affinity gel. SDS-PAGE-purified proteins were transferred to nitrocellulose and digested with trypsin. Resulting peptides were resolved by electrophoresis and chromatography, and the phosphorylated peptides were visualized by autoradiography. The filled arrowhead identifies the phosphopeptide that contains S348 and the open arrowhead identifies the location of the missing phosphopeptide. C, phosphoamino acid analysis of the phosphopeptide representing phosphoserine 348. The phosphopeptide containing phosphoserine 348 in B (marked with a filled arrow) was scraped from the TLC plate and eluted from the cellulose as described under Experimental Procedures. The eluent was vacuum-dried and hydrolyzed in 5.7 N HCl at 110°C for 1 h. The hydroxylate was mixed with phosphoamino standards [phospho-serine (pS); phospho-threonine (pT); and phospho-tyrosine (pY)], separated electrophoretically in two dimensions on cellulose plates, stained with ninhydrin to visualize standards (shown by outlined circles), and autoradiography was performed.

Functionality of Mutant hARNT cDNAs Expressed in Cell Lines. Phosphorylation of a protein causes an increase in negative charge. Consequently, it is possible that the effects of phosphorylation on protein function could be reproduced experimentally by the addition of a negative charge at the phosphorylation site. To test this hypothesis, we constructed hARNT expression vectors in which S348 was changed to either the neutrally charged alanine residue (S348A) or the negatively charged glutamic acid residue (S348E) to assess the functional role of phosphoserine 348. Therefore, a comparison between the activity of S348A and S348E can provide a basis for determining whether a phosphorylation site has a positive or negative functional role. Hepa-1 c4 cells were used for this comparison because they do not express ARNT at the protein level. Figure 5A shows that both S348A and S348E mutated hARNT proteins trans-activated XRE driven reporter activity in Hepa-1 c4 cells, in the presence and absence of TCDD, to an equal level as S348. Figure 5B shows the protein expression level of each of the transfected hARNT constructs in Hepa-1 c4 cells. XRE driven reporter activity was normalized against hARNT expression levels and -galactosidase activity. Western blot analysis confirmed that mock-transfected Hepa-1 c4 cells did not express detectable levels of ARNT and that each of the transfected hARNT constructs express at similar levels when normalized to p50 expression levels.

View larger version (29K): [in this window] [in a new window]   Fig. 5.   Changing serine at amino acid position 348 (S348) to an alanine (S348A) or a glutamic acid (S348E) does not influence XRE driven luciferase activity in Hepa-1 c4 cells. A, Hepa-1 c4 cells were transiently transfected with either pSV-Sport1/hARNT/S348, pSV-Sport1/hARNT/S348A, pSV-Sport1/hARNT/S348E (2.5 ng µg), pGudLuc 6.1 (100 ng), and a -galactosidase expression vector (100 ng) in a 6-well plate. The total amount of DNA transfected per well was 1.5 µg and equalized with pSV-Sport1 expression vector. Twenty-four hours after transfection, cells were treated with 10 nM TCDD for 8 h and cells were lysed to measure luciferase activity; luciferase activity is reported as RLUs normalized against -galactosidase expression. Transfection groups are presented as mean ± S.E.M. (n = 6) and groups with the same letter were not significantly different as determined with ANOVA and Duncan's multiple range test ( = 0.05). B, expression of hARNT in Hepa-1 c4 cells. Hepa-1 c4 cells were transiently transfected with the pSV-Sport1/hARNT cDNA (9 µg per 10-cm2 dish) and whole-cell extracts were prepared 36 h after transfection. The control transfection (mock transfected) received 9 µg of pSV-Sport1 expression vector. SDS-PAGE and immunoblotting were conducted as described under Experimental Procedures. hARNT expression levels were normalized against p50 protein levels and then normalized hARNT levels were used along with -galactosidase activity to adjust XRE-driven luciferase activity reported in A.

View larger version (34K): [in this window] [in a new window]   Fig. 6.   Changing serine at amino acid position 348 (S348) to an alanine (S348A) or a glutamic acid (S348E) does not influence class B E-box-driven luciferase activity Hepa-1 c4 cells and COS-1 cells. Hepa-1 c4 cells and COS-1 cells were transiently transfected with either pSV-Sport1/hARNT/S348, pSV-Sport1/hARNT/S348A, pSV-Sport1/hARNT/S348E cDNAs (1.0 µg), either pMyc3E1bluc or pMyc3E1bluc E-box reporter construct (0.4 µg), and a -galactosidase expression vector (0.1 µg) in a 6-well plate. The total amount of DNA transfected per well was 1.5 µg, and for control groups the amount of DNA was equalized with empty pSV-Sport1 expression vector. The pMyc0E1bLuc and the pMyc3E1bLuc reporter constructs consist of zero and three E-boxes, respectively, located 5' to luc+. Thirty-six hours after transfection, cells were lysed and luciferase activity was measured. Class B E-box-driven luciferase activity is reported as RLUs normalized against -galactosidase expression. Transfection groups are presented as mean ± S.E.M. (n = 6) and groups with the same letter were not significantly different as determined with ANOVA and Duncan's multiple range test ( = 0.05).

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

An earlier investigation in our laboratory characterized the level of charge heterogeneity displayed by mARNT (Tsai and Perdew, 1997). A comparison between mARNT as a monomer and mARNT heterodimerized with the murine AhR, demonstrated that a significant shift in the pI of ARNT occurs just before or after heterodimerization in the nucleus. A movement of the predominant ARNT isoforms toward the basic end characterized this shift in charge heterogeneity. Furthermore, this shift in charge heterogeneity was believed to be related in part to a change in ARNT phosphorylation or an alteration in another type of secondary modification. For example, one type of modification that could influence ARNT charge heterogeneity is glycosylation, which can block phosphorylation of specific sites. Previously, it was shown that the bHLH protein c-Myc can be either phosphorylated or glycosylated at threonine 58, which is located in the trans-activation domain (Gupta et al., 1993; Chou et al., 1995). Additionally, it was determined that phosphorylation of threonine 58 can regulate c-MYC signaling (Gupta et al., 1993). Consequently, it was proposed that reciprocal glycosylation and phosphorylation at threonine 58 could play an important role in the regulation of c-MYC activity. Whether hARNT residues can be glycosylated is unknown. However, the possibility exists that glycosylation could potentially influence ARNT charge heterogeneity, phosphorylation status, and activity. Presently, the significance of the aforementioned shift in ARNT charge heterogeneity after heterodimerization with the AhR is unknown. Despite the shift in ARNT charge heterogeneity after heterodimerization with the AhR, heterodimerization does not seem to influence the general phosphorylation pattern of hARNT (Fig. 2). However, because the stoichiometry of each hARNT phosphorylation site was not assessed, it is unknown whether heterodimerization with the AhR significantly influences the stoichiometry of individual phosphorylation sites.

An earlier study that investigated the functional role of phosphorylation in the interaction between AhR and ARNT demonstrated that phosphorylation of ARNT is required for formation of the AhR-ARNT heterodimer (Berghard et al., 1993). In a later study, it was shown that the HLH and PAS domains of ARNT mediate dimerization (Reisz-Porszasz et al., 1994). Taken together, the results from these two studies merited our use of the truncated hARNT/474-Flag construct to search for a phosphorylation site(s) that may be involved in ARNT heterodimerization and homodimerization. Phosphoamino acid analysis indicated that hARNT/474-Flag is phosphorylated exclusively on serine residues and that hARNT/Flag is phosphorylated almost exclusively on serine residues. Solid-phase sequencing of tryptic phosphopeptides isolated from two-dimensional maps of hARNT/474-Flag identified five candidate serine containing tryptic peptides located in the PAS domain and identified one candidate serine containing peptide located in the distal end of the HLH region (Table 1). By changing each of these candidate serine residues to alanine residues, we were able to identify S348 as a phosphorylation site and verify that the other five candidate sites were not phosphorylated. The functional role of S348 was assessed with two ARNT signaling pathways. The first signaling pathway examined the role of S348 in AhR-ARNT signaling by comparing the ability of S348, S348A, and S348E to trans-activate XRE-driven luciferase activity in an ARNT-deficient cell line. The second signaling pathway examined the role of S348 on ARNT-ARNT homodimer signaling by comparing the ability of S348, S348A, and S348E to trans-activate class B E-box-driven luciferase activity in the ARNT deficient cell line and COS-1 cells. These analyses have demonstrated that phosphorylation of S348 apparently does not influence ARNT heterodimerization or homodimerization and therefore does not influence AhR-hARNT or hARNT homodimer mediated gene expression.

In contrast to the study by Berghard et al. (1993), which compared a phosphorylated form of ARNT with a dephosphorylated form of ARNT through the use of alkaline phosphatase, the results of the present study were generated by examining the role of a single phosphorylation site in a functional domain. Therefore, it is difficult to draw a comparison between the results from the present study with those of Berghard et al. (1993). However, an unpublished result from our laboratory demonstrated that the phosphatase inhibitor okadaic acid increased hARNT-hARNT-mediated E-box activity by 2-fold in hARNT transfected COS-1 cells. This 2-fold increase in E-box activity was accompanied by a greater than 10-fold increase in [³²P]orthophosphate incorporation into hARNT without influencing hARNT expression levels in COS-1 cells. The results from our laboratory, along with the results of Berghard et al. (1993), suggest that overall phosphorylation status, or a cluster of sites, rather than site-specific phosphorylation, may be crucial in the regulation of ARNT-mediated signaling.

This study has shown that although hARNT is a phosphoprotein, the phosphorylation of a PAS-region serine residue does not modulate AhR-ARNT and ARNT-ARNT mediated signaling. In addition, the phosphorylation pattern of hARNT seems to be unaltered whether it is a monomer or dimerized with the hAhR. Although phosphorylation of S348 was shown not to modulate AhR and ARNT-ARNT signaling, the phosphorylation status of S348 may play a functional role with other ARNT dimerization partners in their respective systems. Future investigations will identify phosphorylation sites distal to the PAS domain and determine whether S348 plays a functional role in combination with these other hARNT phosphorylation sites in ARNT-mediated signaling systems.