Address correspondence and reprint requests to Kwang-Soo Kim, Molecular Neurobiology Laboratory, McLean Hospital, Harvard Medical School, 115 Mill Street, Belmont, MA 02478, USA. E-mail: firstname.lastname@example.org
The retinoic acid-inducible and developmentally regulated transcription factor AP-2 plays an important role during development. In adult mammals, AP-2 is expressed in both neural and non-neural tissues. However, the function of AP-2 in different neuronal phenotypes is poorly understood. In this study, transcriptional regulation of tyrosine hydroxylase (TH) and dopamine β-hydroxylase (DBH) genes by AP-2 was investigated. AP-2 binding sites were identified in the upstream regions of both genes. Electrophoretic mobility shift assays (EMSA) and DNase I footprinting analyses indicate that the AP−2 interaction with these motifs is more prominent in catecholaminergic SK-N-BE(2)C and CATH.a than in non-catecholaminergic HeLa and HepG2 cell lines. Exogenous expression of AP-2 robustly transactivated TH and DBH promoter activities in non-catecholaminergic cell lines. While AP-2 regulates the DBH promoter activity via a single site, transactivation of the TH promoter by AP-2 appears to require multiple sites. In support of this, mutation of multiple AP-2 binding sites but not that of single site diminished the basal promoter activity of the TH gene in cell lines that express TH and abolished transactivation by exogenous AP-2 expression in cell lines that do not express TH. In contrast, mutation of a single AP-2 binding site of the DBH gene completely abolished transactivation by AP-2. Double-label immunohistochemistry showed that AP-2 is coexpressed with TH in noradrenergic and adrenergic neurons in both the central and peripheral nervous systems of adult rodents. Numerous non-catecholaminergic cell groups within the spinal cord, medulla, cerebellum, and pons also express AP-2. The concentration of AP-2 in dorsomedial locations along the neuraxis suggests a regionally specific role for this transcription factor in the regulation of neuronal function. Based on these findings we propose that AP-2 may coregulate TH and DBH gene expression and thus participate in expression/maintenance of neurotransmitter phenotypes in (nor)adrenergic neurons and neuroendocrine cells.
Eukaryotic gene transcription plays an essential regulatory role in mammalian developmental processes such as cell fate determination and phenotypic specification of terminally differentiated cell types. Understanding these molecular processes in the nervous system poses a unique challenge considering the morphological, biochemical, electrophysiological, and pharmacological heterogeneity of cellular phenotypes. Among the various phenotypes of a particular neuron, neurotransmitter identity is a cardinal feature because it determines the nature of the chemical neurotransmission it will mediate, and its specific connectivity with target neurons. Recently, transcriptional control has been demonstrated as a critical mechanism for specification of neurotransmitter phenotypes in a variety of organisms, such as Canorhabditis elegans, Drosophila, and mouse (reviewed in Goridis and Brunet 1999). Catecholaminergic neurons provide a potentially ideal system to study control mechanisms of neurotransmitter phenotypes because subtypes of catecholaminergic neurons are identified by differential expression of biosynthetic enzymes. Thus, dopaminergic neurons in the substantia nigra pars compacta and ventral tegmental area express the first two enzymes in catecholamine biosynthesis, tyrosine hydroxylase (TH) (Nagatsu et al. 1964) and l-aromatic amino acid decarboxylase (Jaeger et al. 1983). Noradrenergic neurons in the locus coeruleus and other specific cell groups within the brainstem (A1, A2, A4, A5, A7) express an additional enzyme, dopamine β-hydroxylase (DBH), that converts dopamine to noradrenaline (Kirshner and Goodall 1957; Friedman and Kaufman 1965). Finally, adrenergic cells in the C1, C2, and C3 cell groups express phenylethanolamine N-methyltransferase (PNMT) (Axelrod 1962) and produce adrenaline. This study is part of a larger effort to characterize the molecular mechanisms that control the spatially and temporally selective expression of the enzymes involved in catecholamine biosynthesis. In particular, we are interested in identifying and characterizing transcription factors which directly control (transactivate or repress) expression of catecholamine-synthesizing genes. Insights gained from investigation of catecholaminergic systems may provide a model for understanding the molecular control of phenotypes in other populations of central nervous system neurons.
Both in vivo transgenic and in vitro cell culture studies have demonstrated that relatively short upstream sequences of the DBH gene were sufficient for driving noradrenergic-specific reporter gene expression (Shaskus et al. 1992; Ishiguro et al. 1993; Hoyle et al. 1994). Taking advantage of this finding, we systematically characterized the human DBH promoter by electrophoretic mobility shift assay (EMSA), DNase I footprinting, deletional and site-directed mutational analyses (Ishiguro et al. 1993; Kim et al. 1994, 1998a; Seo et al. 1996; Yang et al. 1998a, b). Two cis-regulatory elements, the homeodomain-binding site of the composite promoter (domain IV) and domain II, that are exclusively active in noradrenergic cell lines and thus essential for the cell-specific promoter activity of the DBH gene were identified (Seo et al. 1996; Kim et al. 1998a; Yang et al. 1998a). Of note, both cis-regulatory elements were shown to be binding sites for the paired-like homeodomain protein Phox2a (Kim et al. 1998a; Yang et al. 1998a), which is critical for the development of several major noradrenergic cell groups, including the locus coeruleus (Morin et al. 1997).
Transcriptional regulation of the TH gene has also been extensively investigated using both in vitro cell culture and in vivo transgenic mice experiments. Although several groups have indicated that the 5′ upstream promoter of the TH gene can direct cell type-specific gene expression (Harrington et al. 1987; Cambi et al. 1989; Gandelman et al. 1990; Yoon and Chikaraishi 1992), the relative contribution of different cis-regulatory elements (e.g. AP1, dyad/E box, and CRE) to TH transcription appeared to differ significantly among the cell lines used in these studies (Yoon and Chikaraishi 1992; Kim et al. 1993a; Wong et al. 1994; Lazaroff et al. 1995). As a corollary, cis-regulatory element(s) and their cognate protein factors conferring cell specificity to the TH gene are not clearly defined. Recently various signaling molecules have been identified to be critical for development, specification and/or expansion of dopaminergic neurons (Hynes and Rosenthal 1999 and references therein). Among these, the secreted molecules Sonic hedgehog and fibroblast growth factor 8 are particularly interesting because they appear to be both necessary and sufficient for the specification of major dopaminergic neurons (Hynes et al. 1995; Crossley et al. 1996; Ye et al. 1998). However, the transcription factors which act downstream of these inductive signals and directly transactivate TH transcription have not been identified.
In an attempt to identify cell-specific cis-regulatory elements of the TH gene, we recently examined DNA–protein interactions at the 5′ TH promoter using nuclear extracts isolated from either catecholaminergic or non-catecholaminergic cell lines (Yang et al. 1998b). Among multiple footprinted areas identified in this study (domain I–VII), domain VI was the only one that showed better footprinting by nuclear factors isolated from catecholaminergic in comparison with non-catecholaminergic cell lines. This observation suggests that the cognate transcription factor of domain VI may contribute to cell-specific TH gene expression. Domain VI was found to contain a consensus AP-2-binding motif in the middle. The retinoic acid-inducible transcription factor AP-2, also known as AP-2α, is expressed in neural crest and epidermal cell lineages (Williams et al. 1988; Mitchell et al. 1991; Byrne et al. 1994; Meier et al. 1995) and regulates gene transcription by interacting with the palindromic sequence motif, 5′-GCCNNNGGC-3′. Two independent reports of AP-2 gene inactivation demonstrated that AP-2 is critical for early embryogenesis, especially skeletal development (Schorle et al. 1996; Zhang et al. 1996). More recent analysis of chimeric mice composed of both wild-type and AP-2-null cells indicated that AP-2 is required for multiple independent morphogenic processes, including neural tube and eye formation (Nottoli et al. 1998). It is of great interest whether AP-2 also plays an important role in specifying and maintaining the phenotype of terminally differentiated cells including certain neurons. In this report, we show that AP-2 prominently interacts with promoter elements of the TH and DBH genes in noradrenergic cells. Cotransfection analysis further demonstrates that AP-2 can robustly activate both TH and DBH promoter activity in a cell type-specific manner. In contrast to the DBH promoter which contains only a single AP-2 site, our analyses indicate that the TH promoter may include multiple AP-2-responsive/binding sites. In addition, double-label immunohistochemical analysis of adult mouse brain shows that AP-2 is coexpressed with TH in noradrenergic and adrenergic neurons as well as in adrenal chromaffin cells, suggesting that AP-2 may coregulate cell-specific TH and DBH transcription in these cell types. Interestingly, however, AP-2 immunoreactivity was not detected in dopaminergic neurons suggesting that mechanisms controlling TH gene expression in dopaminergic neurons could be different from those in noradrenergic and adrenergic neurons.
Materials and methods
Cell culture and preparation of nuclear extracts
Human neuroblastoma SK-N-BE(2)C, SK-N-BE(2)M17, and mouse central noradrenergic neuron-derived CATH.a cell lines which express both TH and DBH were maintained as described previously (Seo et al. 1996; Kim et al. 1998a). HeLa, HepG2 and C6 glioma cell lines which express neither TH nor DBH were grown in Dulbecco's modified Eagle medium supplemented with 10% fetal calf serum (HyClone, Logan, UT, USA), 100 µg/mL streptomycin, and 100 units/mL penicillin. Nuclear extracts were prepared from the different cell lines as described by Dignam et al. (1983) with final protein concentrations between 10 and 14 µg/µL. Extracts were stored in aliquots at − 70°C and used within 3 months of preparation.
The TH2400CAT and DBH978CAT reporter constructs contain the 2.4 kb upstream sequences of the rat TH gene and the 978 bp upstream sequences of the human DBH gene, respectively, fused to the bacterial chloramphenicol acetyltransferase (CAT) gene (Ishiguro et al. 1993; Kim et al. 1993a). To generate TH230CAT and TH199CAT plasmids, polymerase chain reaction was performed using oligonucleotides 5′-CCCAAGCTTCTACGTCGTGCCTCGGGC-3′ and 5′-CGGTGGTATATCCAGTG-3′, and 5′-CCCAAGCTTGAGGCAGGTGCCTGTGAC-3′ and 5′-CGGTGGTATATCCAGTG-3′, respectively, with TH2400(II)CAT as the template. The TH2400(II)CAT plasmid was derived from TH2400CAT by removing the AatII site residing at the upstream of the multiple cloning sites. The 230 bp and 199 bp fragments were isolated after digesting the PCR products with HindIII and AatII and were subcloned to TH2400(II)CAT that had been digested with HindIII and AatII, resulting in TH230CAT and TH199CAT plasmids, respectively. The upstream and junction regions of these constructs were confirmed by sequence analysis.
Base substitutions at the AP-2 sites were generated in the context of TH2400CAT using the TransformerTM Site-Directed Mutagenesis kit (Clontech, Palo Alto, CA, USA). Oligonucleotides SHM10 (5′-CCCACCCCCGCCTCCCTCATTAACAGCAGGCG TGGAGAGGA-3′), Shm11 (5′-AGCCCCTGCCCTACGTCGTTTA TCGGGCTGAGGGTGATTCA-3′), SHM12 (5′-GTTATAGTTCT AACATGAGTTATTAGGAAATCCAGCATGGT-3′), SHM13 (5′-GGAACATGGCCCATGTCCTTTAAAAGACTTTATGACAGAC ATCC-3′), and SHM14 (5′-CAGAAGGCCTTAGGGAGCTTTA AGGGGCTAGGGTTGGCACC-3′) were used for the mutagensis of AP-2-II, -III, -IV, -V, and -VI, respectively (underlined bases represent mutated sequences). SHM11, SHM12, and SHM13 were used simultaneously to create triple mutation in AP-2-II, -III, and -IV according to the manufacturer's procedure. Similarly, oligonucleotides SHM11, SHM12, SHM13, SHM14, and SHM15 were simultaneously used to create mutations in all five AP-2-II, -III, -IV, -V, and -VI. Constructs with correct mutations were confirmed by direct sequence analysis.
The plasmid pSPRSV-AP-2 (a kind gift from Dr T. Williams, Yale University, New Haven, CT; Williams and Tjian 1991) expresses full length AP-2 protein under the control of a RSV promoter and was used as an effector plasmid in cotransfection assays. An empty vector, pSPRSV-NN, was used as a negative control for cotransfection assays.
Transient transfection and enzyme assays
Transfection was performed by using the calcium phosphate coprecipitation method (Ishiguro et al. 1993; Seo et al. 1996). For the SK-N-BE(2)C and SK-N-BE(2)M17 cell lines, each 60-mm dish was transfected with 2 µg of the reporter construct, 1 µg of pRSV β-gal, varying amounts of the effector plasmid, and pUC19 plasmid to a total of 5 µg DNA. For the other cell lines, twice as much DNA was used for transfection. Plasmids used for transient transfection assays were prepared using Qiagen columns (Qiagen Co., Santa Clarita, CA). To correct for differences in transfection efficiencies among different DNA precipitates, CAT activity was normalized to that of β-galactosidase. CAT and β-galactosidase activities were assayed as previously described (Ishiguro et al. 1993; Seo et al. 1996).
EMSA and DNase I footprinting
Sense and antisense oligonucleotides corresponding to the sequences of AP-2 binding sites of the rat TH gene were synthesized with the following sequences: THAP2-I-S (5′-GTGGGGGACCCCAGAGGGGCTTTGAC-3′) and THAP2-I-A (5′-CGTCAAAGCCCCTCTGGGTCCCCCA-3′), THAP2-II-S (5′-CCCGCCTCCCTCAGGCACAGCAGGC-3′) and THAP2-II-A (5′-CGCCTGCTGTGCCTGAGGGAGGCGG-3′), THAP2-III-S (5′-CCCTACGTCGTGCCTCGGGCTGAGG-3′) and THAP2-III-A (5′-CCCTCAGCCCGAGGCACGACGTAGG-3′), THAP2-IV-S (5′-ACATGAGCCCTTAGGAAATCCAGCA-3′) and THAP2-IV-A (5′-ATGCTGGATTTCCTAAGGGCTCATG-3′), THAP2-V-S (5′-CATGGCCCATGTCCTGGAGGGGACTTT-3′) and THAP2-V-A (5′-TAAAGTCCCCTCCAGGACATGGGCCAT-3′), THAP2-VI-S (5′-AGGGAGCTGCCAGGGGCTAGGGTTG-3′) and THAP2-VI-A (5′-CCAACCCTAGCCCCTGGCAGCTCCC-3′) to represent the AP-2-I, -II, -III, -IV, -V, and -VI sites, respectively. THAP2m-II-S (5′-CCCGCCTCCCTCATTAACAGCAGGC-3′), THAP2m-II-A (5′-CGCCTTGCTGTTAATGAGGGAGGCGG-3′), THAP2m-III-S (5′-CCCTACGTCGTTTATCGGGCTGAGG-3′), THAP2m-III-A (5′-CCCTCAGCCCGATAAACGACGTAGG-3′), THAP2m-IV-S (5′-ACATGAGTTATTAGGAAATCCAGCA-3′), THAP2m-IV-A (5′-ATGCTGGATTTCCTAATAACTCATG-3′), THAP2m-V-S (5′-TGGCCCATGTCCTTTAAAAGACTTT-3′), THAP2m-V-A (5′-TAAAGTCTTTTAAAGGACATGGGCC-3′), THAP2m-VI-S (5′-AGGGAGCTTTAAGGGGCTAGGGTTG-3′) and THAP2m-VI-A (5′-CCAACCCTAGCCCCTTAAAGCTCCC-3′) were also synthesized for mutant sense and antisense oligonucleotides of the THAP-2-I, -II, -III, -IV, -V, and -VI sites, respectively (the underlined bases represent mutated sequences). In addition, oligonucleotides 5′-GATCGAACTGACCGCCCGCGGCCCG-3′ and 5′-ACGGGCCGCGGGCGGTCAGTTCGAT-3′ were synthesized and used here as the consensus AP-2 (Williams and Tjian 1991). A consensus Sp1 binding site and the AP-2 site of the DBH gene were described in previous studies (Seo et al. 1996; Kim et al. 1998a). The sense and antisense oligonucleotides were annealed, gel-purified, and 32P-labeled with T4 DNA kinase and used as probes in electrophoretic mobility shift assays (EMSA). EMSA and antibody coincubation experiments were performed using 30 000–50 000 cpm of labeled probe (approximately 0.05–0.1 ng) and nuclear extracts (10–30 µg) in a final volume of 20 µL of 12.5% glycerol, 12.5 mm HEPES (pH 7.9), 4 mm Tris-HCl (pH 7.9), 60 mm KCl, 1 mm EDTA, and 1 mm DTT with 1 µg of poly(dI-dC) as described previously (Seo et al. 1996; Yang et al. 1998a). Competition binding assays were performed by adding non-radioactive competitor oligonucleotides in a molar excess prior to adding 32P-labeled oligonucleotides. For antibody coincubation experiments, antibody was coincubated with the nuclear extract mix for 30 min on ice prior to adding the radiolabeled probe. Antibodies against AP-2 and SP1 were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA). Recombinant AP-2 proteins were purchased from Promega Corporation (Madison, WI, USA) and used for the EMSA and DNase I footprinting assay. The DNase I footprinting assay was performed using TH 5′ proximal promoter fragments that had been prepared by PCR as previously described (Seo et al. 1996; Yang et al. 1998b). Approximately 30 000 cpm of labeled probe was incubated with 0.5–2.0 footprint units of recombinant AP-2 protein, and was subject to DNase I digestion. The amount of DNase I was adjusted empirically for each reaction to produce an even pattern of partially cleaved DNA fragments. Location of each band was determined by Maxam–Gilbert sequencing reactions of the labeled probes.
TH and AP-2 immunoreactivity (IR) was localized within the central nervous system of adult FVB mice (Jackson Laboratories, Bar Harbor, ME, USA). Adult Sprague–Dawley rats (Harlan Sprague Dawley Inc., Indianapolis, IN, USA) were used for immunohistochemical analysis of the superior cervical ganglia. The animals were cared for in accordance with the National Institutes of Health's Guide for the Care and Use of Laboratory Animals. Mice were deeply anesthetized with ketamine/xylazine and perfused with normal saline followed by 4% paraformaldehyde in 0.1 m phosphate buffer (pH 7.4). Brains, adrenal glands and superior cervical ganglia were stored in 30% sucrose/0.1 m phosphate-buffered saline (PBS) buffer at 4°C and then sectioned in the coronal plane at 20 µm. Two or more series of brain sections were collected onto glass slides. At least one series was stained with cresyl violet for identification of nuclear boundaries. Other series were stained for AP-2 or AP-2/TH. The primary antibody to AP-2 was a polyclonal rabbit IgG that recognizes an 18 amino acid sequence of human AP-2 (Santa Cruz Biotechnology, catalog # sc-184, Santa Cruz, CA, USA), and does not cross-react with AP-2β or AP-2 γ. The primary antibody to TH was a monoclonal generated against TH from PC12 cells and does not react with DBH, phenylalanine hydroxylase, tryptophan hydroxylase, dihydropteridine reductase, sepiapterin reductase, or PNMT on Western blots (Chemicon, catalog # MAB318, Temecula, CA, USA).
Double-labeling for AP-2/TH proceeded as follows: (1) rinsed in 0.02 m phosphate buffer with 0.9% NaCl (PBS) × 3 over 30 min; (2) quenched endogenous peroxidases with 10% methanol and 3% hydrogen peroxide in PBS for 5 min; (3) rinsed in PBS × 3 over 30 min; (4) blocked with 2% non-fat dry milk, 0.3% Triton X-100 (Sigma), and Avidin D (Vector, Burlinghame, CA, USA, Avidin/biotin blocking kit) for 30 min; (5) rabbit anti-AP-2 at 1 : 2500 dilution, 3% goat serum, biotin (Vector, Avidin/biotin blocking kit), and 0.1% Triton X-100 in PBS with 0.05% NaAzide (Sigma, St Louis, MO, USA) overnight; (6) rinsed in PBS × 3 over 30 min; (7) biotinylated goat antirabbit IgG (Vector) at 1 : 500 dilution, 2% goat serum, and 0.1% Triton X-100 for 4 h; (8) rinsed in PBS × 3 over 30 min; (9) ABC Reagent (Vector) and 0.1% Triton X-100 in PBS for 90 min; (10) rinsed in PBS × 3 over 30 min; (11) nickel-intensified diaminobenzidine (Vector) reaction for generation of a blue/black reaction product; (12) rinsed in PBS × 3 over 30 min; (13) mouse anti-TH at 1 : 2000 dilution, 3% horse serum, and 0.1% Triton X-100 in PBS with 0.05% NaAzide overnight; (14) rinsed in PBS × 3 over 30 min; (15) biotinylated horse antimouse IgG (Vector) at 1 : 500 dilution, 2% horse serum, and 0.1% Triton X-100 in PBS for 4 h; (16) rinsed in PBS × 3 over 30 min; (17) ABC Reagent and 0.1% Triton X-100 in PBS for 90 min; (18) rinsed in PBS × 3 over 30 min; (19) diaminobenzidine reaction for generation of a brown reaction product; (20) rinsed in PBS × 3 over 30 min; (21) dehydrated, cleared, and coverslipped. All cases were reviewed for anatomical localization of AP-2-IR and TH-IR.
5′ Proximal promoter regions of both the TH and DBH genes contain a consensus AP-2 motif which interacts with nuclear proteins in a catecholaminergic cell-preferred manner
The multiple protein-binding sites that we have recently identified in both TH and DBH upstream promoter regions are schematically summarized in Fig. 1 (Kim et al. 1998a; Yang et al. 1998b). Domain VI of TH promoter and domain III of DBH promoter include a consensus AP-2-binding motif (5′-GCCNNNGGC-3′; Mitchell et al. 1991; Williams and Tjian 1991); 5′-GCCTCGGGC-3′ for TH and 5′-GCCTGGGGC-3′ for DBH, respectively. Both sites were footprinted by nuclear factors isolated from catecholaminergic cells better than by those from non-catecholaminergic cells (Kim et al. 1998a; Yang et al. 1998b). While the AP-2 transcription factor is shown to bind to domain III of the DBH promoter and transactivate DBH promoter activity (Greco et al. 1995; Kim et al. 1998a), its role in cell-specific DBH transcription had not been defined, and the potential roles of AP-2 in TH transcription and its interaction with the TH promoter were not known. The results below address the possibility that AP-2 may coregulate TH and DBH transcription and thus contribute to the neurotransmitter phenotype of noradrenergic and adrenergic cells. Our analyses of protein/DNA interactions showed that AP-2 is the cognate transcription factor interacting with domain VI of the TH gene: (1) nuclear proteins prepared from catecholaminergic cells form a specific DNA-protein complex which is competed specifically by the known consensus AP-2-binding motif and domain VI oligonucleotide with the same efficiency; (2) coincubation of nuclear proteins with AP-2-specific antibody diminishes complex formation and generates a supershifted band; and (3) the recombinant AP-2 protein binds to domain VI and forms complex(es) with an identical mobility with C1, which are completely supershifted by addition of AP-2-specific antibody (data not shown). Based on these data, we conclude that AP-2 is the (or one of the) cognate protein factors that interact with domain VI of the TH gene. Because domain VI is footprinted better by nuclear proteins prepared from catecholaminergic than non-catecholaminergic cell lines (Yang et al. 1998b), it is likely that the transcription factor(s) interacting with domain VI (e.g. AP-2) may be preferably expressed in catecholaminergic cells. In support of this, nuclear extracts from TH-positive cells (SK-N-BE(2)C and CATH.a) formed specific DNA–protein complex more efficiently compared to those form TH-negative cells (HeLa and C6) (data not shown).
Exogenous expression of AP-2 robustly transactivates both TH and DBH promoter activities in non-catecholaminergic cell lines
To test whether forced expression of AP-2 can activate TH and/or DBH promoter activity, different amounts of AP-2 expression vector were cotransfected with the reporter construct TH2400CAT or DBH978CAT to catecholaminergic (SK-N-BE(2)C) and non-catecholaminergic (HeLa and HepG2) cell lines. In SK-N-BE(2)C cells, cotransfection with AP-2 activated the promoter activities of TH2400CAT and DBH978CAT constructs only marginally; up to 1.6-fold and 2.4-fold, respectively (Fig. 2). These data are consistent with the idea that SK-N-BE(2)C cells express sufficient levels of AP-2, and thus exogenous expression of AP-2 can only modestly transactivate the promoter activities of these constructs. In contrast, cotransfection with the AP-2 expression vector stimulated reporter gene expression driven by TH2400CAT up to 11-fold in HeLa cells in a dose-dependent manner (Fig. 2). Furthermore, in AP-2-non-expressing HepG2 cells, exogenous AP-2 expression increased reporter gene expression driven by TH2400CAT up-to 20-fold (Fig. 2). In a similar fashion, reporter gene expression by DBH978CAT was much more robustly activated by exogenous AP-2 expression in these non-catecholaminergic cell lines, compared to catecholaminergic SK-N-BE(2)C cell line. These results show that AP-2 is able to up-regulate the promoter activities of both TH and DBH genes in a cell preferred manner.
AP-2 regulates the TH promoter activity via multiple binding sites
Having identified domain VI of the TH promoter and domain III of the DBH promoter as the AP-2-binding sites, we next examined the functional role of these cis-elements in regulating TH and DBH promoter activities. Mutation of the domain III AP-2-binding site in the DBH978CAT construct diminished most of the basal transcriptional activity as well as the AP-2-inducibility of the DBH promoter in the SK-N-BE(2)C cell line (Fig. 3a). This result is in good agreement with a previous study by Greco et al. (1995) which demonstrated that mutation of an AP-2 binding site at −136 to −115 bp of the rat DBH promoter, corresponding to domain III of the human DBH promoter (Fig. 1), reduced basal reporter gene expression sevenfold in the neuroblastoma SHSY-5Y cell line. In the HepG2 cell line, the basal reporter gene expression driven by DBH978CAT was less than 10% of that seen with SK-N-BE(2)C thereby showing the cell type-specificity of gene expression (data not shown; also see Yang et al. 1998b). Mutation of domain III did not affect basal DBH promoter activity in the HepG2 cell line, which is explained by the fact that HepG2 cells do not express AP-2 (Williams et al. 1988). The same mutation, however, almost completely eliminated the AP-2-inducibility of the DBH promoter in HepG2 cells (Fig. 3b). This result strongly suggest that domain III is primarily responsible for transactivation of the DBH gene by AP-2.
Domain VI of the TH gene promoter was mutated in the context of TH2400CAT such that it would not bind AP-2 (Figs 1 and 3). In SK-N-BE(2)C cell line, this mutation modestly decreased the TH basal promoter activity in SK-N-BE(2)C (approximately 20%). Furthermore, AP-2-responsiveness of the TH promoter was affected only marginally, if at all, by this mutation in both cell lines (Fig. 3a,b). These findings prompted us to hypothesize that transcriptional activation of the TH gene by AP-2 may require additional cis-elements than domain domain VI. To test this, we first tested various TH-CAT constructs containing different lengths of TH upstream sequences for their transcriptional activation in response to exogenous AP-2 expression (Fig. 4). Deletions starting from the −2.4 kb to the −0.15 kb site progressively diminished AP-2-responsive promoter activation of the TH gene both in HepG2 and HeLa cell lines, strongly suggesting that the TH promoter may contain multiple AP-2-responsive sites. We next performed DNase I footprinting analysis of the upstream promoter region of the TH gene. Remarkably, this footprinting analysis revealed six AP-2-binding sites (termed AP-2-I to VI) within the 780 bp upstream region, four of which showed almost complete protection (Fig. 5a). AP-2-III precisely corresponded to domain VI that had been previously identified in DNase I footprinting analysis using nuclear extracts (Fig. 1; Yang et al. 1998b). All of these sequences contained either consensus AP-2 binding motif (AP-2-III and -VI) or related sequences (AP-2-I, -II, -IV, and -V) at the middle (Fig. 6b).
Oligonucleotides encompassing the individual AP-2 binding domains as well as mutated sequences (Fig. 5b; also see Materials and methods) were radiolabeled and used as probes in EMSA. Each mutant was designed such that at least three base substitutions are introduced to the consensus AP-2 binding motif therein. As shown in Fig. 6, EMSA using the recombinant AP-2 protein showed that specific DNA/protein complexes were formed between each domain and AP-2 protein. Competition assays and antibody supershift assays demonstrated that these complexes are formed by specific interaction of AP-2 with each motif (data not shown). Consistent with this, all mutant oligonucleotides did not generate any detectable complexes with the recombinant AP-2 protein (Fig. 6). Both DNase I footprinting data (Fig. 5) and EMSA (Fig. 6) show that AP-2-I at −65 to −40 bp has weakest affinity for AP-2. All other motifs (AP-2-II to -VI) bound to AP-2 with an affinity comparable to that of a well characterized consensus motif (Fig. 6). We next introduced the same mutations which abolished AP-2 binding in the context of TH2400CAT construct (Fig. 7a). We focused our study on AP-2-II to -VI because they have higher affinity to AP-2. As shown in Fig. 7(b), individual mutation of AP-2-II, -III, -IV, -V, or -VI reduced the basal transcriptional activity by 15–30% in SK-N-BE(2)C and SK-N-BE(2)M17 cell lines, suggesting that each motif may contribute to basal promoter activity of the TH gene in TH-expressing cells. When multiple motifs were mutated simultaneously, the basal reporter expression was more significantly (approximately 80%) diminished in both cell lines. These results strongly suggest that AP-2 may activate TH transcription by interacting with multiple motifs in these TH-expressing cell lines. We also contransfected each mutant construct with AP-2 expression plasmid or empty plasmid into HeLa and HepG2 cell lines (Fig. 7c). While mutation of AP-2-II, -III, or -IV marginally decreased transactivation of TH promoter by AP-2, that of AP-2-V or –VI did not have any significant effect. Remarkably, triple mutation of AP-2-II, -III, and -IV diminished transactivation by AP-2 from 21-fold to 4.1-fold and from eightfold to 1.8-fold in HepG2 and HeLa cells, respectively. Furthermore, mutation of five motifs (AP-2-II to -VI) completely abolished transactivation of the TH promoter by AP-2 in both cell lines. Taken together, in a sharp contrast to transactivation of DBH transcription by AP-2 via a single site, our results strongly suggest that AP-2 may transactivate TH transcription by interacting with multiple sequence motifs at the 5′ upstream region.
Distinct pattern of AP-2 expression in the nervous system
Within the central nervous system, AP-2-IR was present within the spinal cord, medulla, cerebellum, pons, mesencephalon, and dorsal portions of the caudal diencephalon. There was no AP-2-IR within the telencephalon. AP-2-IR was concentrated in the dorsal-medial portions of the spinal cord, medulla, pons, and midbrain. As shown in Fig. 8, the densest concentrations of AP-2-IR nuclei were in the molecular layer of the cerebellum. AP-2-IR was also prominent in the region of the superior colliculi and pretectum (Fig. 8f). AP-2-IR cerebellar Purkinje cells were concentrated in the vermis, flocculus, and paraflocculus. AP-2-IR was sparse within the granular cell layer and, based on nuclear size, probably localized to Golgi type II cells rather than granule cells.
In the peripheral nervous system, AP-2-IR was present in both the adrenal medulla and superior cervical ganglia (data not shown). Our finding of AP-2 within the adrenal medulla is consistent with the results of previous investigators (Ebert et al. 1998). Both the adrenal medulla and superior cervical ganglia also showed TH-IR.
TH-IR cell bodies were detected within all dopaminergic, noradrenergic, and adrenergic cell groups within the central nervous system. There were no AP-2/TH double-labeled cells at known sites of dopaminergic neurons including the substantia nigra pars compacta (Fig. 8g), ventral tegmental area, olfactory bulb (Fig. 8h), periaqueductal gray, and hypothalamus. There were also presumptive dopaminergic cells with TH-but no AP-2-IR near the dorsal motor nucleus of the vagus and nucleus tractus solitarius. Presumptive adrenergic neurons within the C-1, C-2, and C-3 cell groups were double-labeled for AP-2 and TH (Fig. 8c). The vast majority of cells within the locus coeruleus and nearly all cells in other noradrenergic cell groups (A1, A2, A4, A5, A7) showed label for both TH and AP-2. A dense population of TH/AP-2-IR neurons was localized to the area postrema. There was moderate variability in the density of AP-2-IR among noradrenergic and adrenergic neurons. For example, AP-2-IR was more prominent in the A1, C1, and C2 cell groups than in the locus coeruleus (Fig. 8a-d).
An interesting finding was the presence of double-labeled cerebellar Purkinje cells (Fig. 8c). The AP-2/TH-IR Purkinje cells were limited to the vermis and within the vermis were concentrated in its caudal half. These double-labeled Purkinje cells were localized to bands that were less than eight cells wide. All TH-IR Purkinje cells were also AP-2-IR. However, there were many AP-2-IR Purkinje cells that were not TH-IR.
AP-2, also known as AP-2α, is a developmentally regulated cell type-specific transcription factor which controls eukaryotic gene expression by interacting as a dimer with a palindromic recognition sequence 5′-GCCNNNGGC-3′ (Mitchell et al. 1991; Williams and Tjian 1991). Recent AP-2 gene inactivation experiments demonstrated that AP-2 is critical for normal development of structures such as the face, skull, and eyes (Schorle et al. 1996; Zhang et al. 1996; Nottoli et al. 1998). The expression of AP-2 in neural crest cell lineages and neuroectodermal cells (Mitchell et al. 1991; Phillip et al. 1994) suggests that this protein may also control expression of neural-specific genes in terminally differentiated neurons. In this study, the role of AP-2 in transcriptional activation of two catecholamine-synthesizing enzyme genes, TH and DBH, was characterized using several catecholaminergic and non-catecholaminergic cell lines as model systems. Furthermore, detailed analysis of AP-2 expression in the mouse brain showed that it is coexpressed with TH in both noradrenergic and adrenergic neurons.
Identification of AP-2 as a cognate transcription factor interacting with the TH and DBH promoters in a cell-preferred manner
In a previous study, we identified seven protein-binding sites (domain I-VII) in the upstream promoter region of the rat TH gene (Fig. 1; Yang et al. 1998b). Among these protein-binding sites, domain VI is located −224 to −212 bp upstream of the transcription start site and includes a nucleotide sequence, 5′-GCCTCGGGC-3′, that matches perfectly with the consensus AP-2-binding motif (Mitchell et al. 1991; Williams and Tjian 1991). We were interested in identifying the transcription factor(s) that interact with domain VI because both EMSA and DNase I footprint analyses showed that domain VI is more efficiently bound by nuclear proteins prepared from catecholaminergic cell lines such as the human neuroblastoma SK-N-BE(2)C and central noradrenergic neuron-derived CATH.a cell lines, compared to those prepared from non-catecholaminergic HeLa and C6 cell lines (data not shown; Yang et al. 1998b). Our analyses of DNA/protein interaction using EMSA, competition assays and antibody supershift assay demonstrated that AP-2 is the cognate transcription factor interacting with domain VI (data not shown). Likewise, domain III of the DBH promoter, interacting with nuclear proteins in a catecholaminergic cell-preferred manner, was previously shown to be an AP-2-binding site (Kim et al. 1998a). An AP-2-binding site has also been identified in a corresponding position in the rat DBH promoter (Greco et al. 1995). Taken together, these data shows that AP-2 interacts with at least one cis-element of the TH and DBH promoters in a catecholaminergic cell-preferred manner. Interestingly, while the AP-2-binding site of the TH gene (domain VI) binds to the protein factor with a similar affinity to the well characterized AP-2 consensus motif residing in the human metallothioneine-IIA gene (Fig. 6; Williams and Tjian 1991), the AP-2-binding site of the human DBH gene (domain III) has a significantly higher affinity for AP-2 protein than the binding sites of the TH and metallothionein-IIA genes (Kim et al. 1998a). Because all three sites contain the consensus AP-2 motif (5′-GCCNNNGGC-3′), this observation indicates that the surrounding nucleotides can significantly affect the binding affinity for AP-2. Consistent with this, the AP-2 binding site of the rat DBH promoter, 5′-GCCCCAGGG-3′, has one base deviation form the consensus motif but still prominently binds to AP-2 (Greco et al. 1995).
AP-2 transactivates the promoter activities of the TH and DBH genes in a coordinated, but mechanistically different mode
The above analyses of DNA–protein interactions at the AP-2 binding sites prompted us to hypothesize that AP-2 may coregulate transcription of the TH and DBH genes. We sought to address this hypothesis using transient coexpression assays because different cell lines are available which either express both TH and DBH (e.g. SK-N-BE(2)C and CATH.a) or express neither of them (e.g. HeLa and HepG2). Exogenous expression of AP-2 activated both TH and DBH promoter activities robustly (10–20-fold) in non-catecholaminergic HeLa and HepG2 cells, but increased them only modestly (approximately up to twofold) in catecholaminergic SK-N-BE(2)C cell (Fig. 2). These results suggest that AP-2 may coregulate the transcriptional activities of both TH and DBH genes in TH/DBH-expressing noradrenergic and adrenergic neurons and neuroendocrine cells.
Having confirmed transactivation function of AP-2, we next examined if the AP-2-binding sites identified in the TH and DBH promoter, i.e. domain VI and domain III, respectively, are critical for basal and AP-2-induced transcriptional activities. We introduced base substitutions to the AP-2-binding site, which dramatically reduced the binding affinity (Fig. 1), in the context of either TH2400CAT or DBH978CAT construct, and tested the transcriptional activity by transient expression assay (Fig. 3). Our results suggest that domain III of the DBH gene is a cell type-specific promoter element in that its mutation severely diminished the DBH promoter activity in SK-N-BE(2)C cells, but did not affect it at all in HepG2 cells. Furthermore, this mutated promoter no more responded to the exogenous AP-2 expression in HepG2 and HeLa cells (Fig. 3; data not shown). These results strongly suggest that domain III of the DBH promoter is not only critical for basal transcriptional activity in catecholaminergic cells but also for transcriptional induction by exogenous AP-2 expression in non-catecholaminergic cells. Based on these results, we conclude that transcriptional activation of the human DBH gene by AP-2 is primarily mediated through domain III. In contrast, mutation of domain VI of the TH gene affected the TH promoter activity only modestly: (1) basal reporter expression driven by the TH2400CAT in SK-N-BE(2)C cells was reduced by approximately 20% in the mutant construct (Fig. 3); and (2) transactivation by exogenous AP-2 was affected only marginally, if any, in the mutant construct. This observation suggested that the TH promoter may contain additional AP-2-responsive sites. Several lines of evidence support this: (1) shorter TH-CAT constructs containing different lengths of the TH upstream sequence show a progressive decrease of transactivation by exogenous AP-2 expression in both HeLa and HepG2 cell lines (Fig. 4); (2) DNase I footprinting analysis identified six AP-2 binding sites residing in the 5′ upstream region (Fig. 5); (3) EMSA, competition assays, and antibody supershift assays demonstrated that AP-2 protein interacts with these motifs and form specific DNA/protein complexes (Fig. 6; data not shown). Furthermore, site-directed mutational analyses showed that multiple mutation of these motifs significantly diminished basal promoter activity of the TH gene in catecholaminergic cell lines and almost completely abolished transactivation by exogenous AP-2 expression in non-catecholaminergic cell lines (Fig. 7). While our results indicate that domains II–VI primarily mediate transactivation of the TH gene by AP-2, it is possible that additional sequence motifs residing between −2.4 kb and −700 bp may also be involved (Fig. 4). Based on these results, we propose that AP-2 may transactivate the TH promoter activity via multiple binding sites. It is worthwhile to note that the PNMT gene promoter contains several AP-2-binding sites and is transactivated by AP-2 in concert with other transcription factors such as glucocorticoid receptor and Egr-1 (Ebert et al. 1998; Wong et al. 1998). Taken together, it is tempting to speculate that AP-2 may regulate expression of all three catecholamine-specific genes, i.e. TH, DBH, and PNMT, in co-ordinate but mechanistically distinct fashions.
Co-localized expression of AP-2 and TH in (nor)adrenergic neurons but not in dopaminergic neurons of the mouse brain
To fully characterize AP-2 expression in the adult mammalian brain, particularly within catecholaminergic neurons, immunohistochemistry was performed using antibodies against AP-2 and TH in coronal sections of adult mouse brain. AP-2 is expressed in the cerebellum and dorsal-medial portions of the spinal cord, medulla, pons, and midbrain in a cell type-specific manner (Fig. 8). The AP-2 expression pattern in the CNS revealed by our immunohistochemical analysis is in good agreement with several previous studies using in situ hybridization (Mitchell et al. 1991; Chazaud et al. 1996; Moser et al. 1997; Shimada et al. 1999). However, unlike Shimada et al. (1999) we did not detect AP-2 within the hippocampus or other telencephalic structures. The distinct pattern of AP-2 expression in the adult brain suggests that, in addition to its critical roles for early developmental pathways, AP-2 may be important for phenotypic determination and the maintenance of different functional characteristics of neurons by regulating targeted gene expression. Indeed, regulation by AP-2 has been indicated for ever increasing numbers of neural genes, including acetylcholinesterase, choline acetyltransferase, neuron-specific enolase, synapsin II, presenilin-2 and proenkephalin genes (Baskin et al. 1997; Hyman et al. 1989; Ekstrom et al. 1993; Getman et al. 1995; Petersohn et al. 1995; Twyman and Jones 1997; Pennypacker et al. 1998). Based on our results indicating coregulation of TH and DBH transcription by AP-2, we hypothesized that catecholaminergic neurons may express AP-2 and thus support TH and DBH gene expression. Consistent with this hypothesis, adrenergic and noradrenergic neurons, which express both TH and DBH, are double-labeled for TH and AP-2. In the periphery, TH and AP-2 are coexpressed in the adrenal medulla and the superior cervical ganglia (data not shown; Ebert et al. 1998). Thus, AP-2 may regulate TH and DBH gene expression in both central and peripheral catecholaminergic cell types. Interestingly, however, AP-2-IR was not detected in any dopaminergic neurons (e.g. substantial nigra pars compacta, ventral tegmental area, and olfactory bulb) which express TH, but not DBH (and PNMT). This finding indicates that AP-2 may not play a role in TH gene expression within dopaminergic neurons. One possible explanation for this observation is that regulation of the TH gene may differ among different anatomical sites within the nervous system. Consistent with this possibility, a recent TH promoter study using transgenic founder analysis indicated that different regions of the TH upstream sequence are responsible for its expression depending on where it is expressed (Liu et al. 1997). An alternate, and mutually not exclusive, explanation is that related but antigenically distinct forms of AP-2 may be expressed in dopaminergic neurons and control TH gene expression. Two additional AP-2-like transcription factors, i.e. AP-2β and AP-2 γ (also known as AP-2.2), have been isolated and characterized (Moser et al. 1995; Oulad-Abdelghani et al. 1996). All three members of AP-2 family recognize the same target sequence and share highly conserved DNA-binding and dimerization domains (Williams and Tjian 1991; Moser et al. 1995; Oulad-Abdelghani et al. 1996). Therefore, it is possible that different members of the AP-2 family may regulate catecholamine-synthesizing gene expression in different types of catecholaminergic neurons.
Specification and/or maintenance of neurotransmitter phenotypes in catecholaminergic neurons
Systematic analyses of the TH and DBH promoters and identification and characterization of the cognate transcription factors have revealed several important features of the regulatory mechanisms of TH and DBH gene expression. The cAMP-dependent protein kinase pathway, acting through the cAMP-response element (Kim et al. 1993a; Ishiguro et al. 1993; Lazaroff et al. 1995), appears to play an essential role for both basal and cAMP-inducible transcription of the TH and DBH genes (Kim et al. 1993b, 1994). The importance of the cAMP-pathway for TH and DBH gene expression was further confirmed using primary cultured bovine chromaffin (Hwang et al. 1997) and neural crest stem cells (Lo et al. 1999). However, the cAMP-pathway is not cell-specific, and can not explain cell type-specific TH and DBH expression. Data from several laboratories indicate that Phox2a (or its rat homolog Arix) is a key regulator of catecholamine-specific gene expression (Zellmer et al. 1995; Morin et al. 1997; Kim et al. 1998a; Yang et al. 1998a). Furthermore, using neural crest stem cells, Lo et al. (1999) have recently shown that Phox2a (and/or its close relative Phox2b) are necessary for TH and DBH gene induction by bone morphogenetic protein 2 and cAMP. An orphan member of the nuclear receptor superfamily, Nurr1, was isolated and found to be coexpressed with TH in major dopaminergic neurons such as substantia nigra and olfactory bulb (Law et al. 1992; Zetterstrom et al. 1997). Notably, three groups have independently shown that inactivation of the Nurr1 gene resulted in agenesis of midbrain dopaminergic neurons (Zetterstrom et al. 1997; Castillo et al. 1998; Saucedo-Cardenas et al. 1998). In adult hippocampus-derived progenitor cells, Nurr1 was able to activate transcription of the TH gene by binding a response element (Sakurada et al. 1999). In addition, a homeodomain transcription factor Ptx3, selectively expressed in midbrain dopaminergic neurons, is shown to activate the TH promoter activity by interacting with a sequence element residing at −50 to −45 bp (Cazorla et al. 2000). Thus, Nurr1 and Ptx3 may represent the first known cell-specific transcription factors that may directly controls TH gene activation. This study shows that AP-2 coregulates the promoter activities of the TH and DBH genes and thus, may contributes to expression of catecholamine neurotransmitters. In light of an emerging theme that general transcription factors (e.g. CREB and Sp1) and cell type-specific (e.g. Nurr1, Phox2a/Phox2b, and Ptx3) transcription factors control catecholamine gene expression in a concerted manner (Swanson et al. 1997; Kim et al. 1998a; Wong et al. 1998; Lo et al. 1999; Cazorla et al. 2000), it will be important to determine how AP-2 and additional members of this family (e.g. AP-2β and AP-2 γ) functionally interact with other general and cell-type specific transcription factors for determining and/or maintaining catecholamine gene expression.
We would like to thank Dr Williams Trevor for his generous gift of the AP-2-expression vector. This work was supported by NIH grants MH48866 (to KSK) (P50)NS39793 KOSEF Fellowship (to SJH), and NS01593 (to MSL).
1These authors equally contributed to this work.
2Current address: Department of Neuroscience, Ewha Medical School, 70-bunji, Jongro-6-ka, Jongro-ku, Seoul 110–783, Republic of Korea.