Molecular and Cellular Biochemistry 212: 51–60, 2000. © 2000 Kluwer Academic Publishers. Printed in the Netherlands.

51

Regulation of tyrosine hydroxylase gene transcription by the cAMP-signaling pathway: Involvement of multiple transcription factors Jinkyu Lim, Chunying Yang, Seok Jong Hong and Kwang-Soo Kim Molecular Neurobiology Laboratory, McLean Hospital, Harvard Medical School, Belmont, MA, USA

Abstract The conversion of L-tyrosine to 3,4-dihydroxy-L-phenylalanine by tyrosine hydroxylase (TH) is the first and rate-limiting step in biosynthesis of catecholamine neurotransmitters. TH gene expression is regulated in a cell type-specific and cAMP-dependent manner. Evidence from this laboratory and others indicates that the cAMP response element (CRE), residing at –45 to –38 bp upstream of the transcription initiation site, is essential for both basal and cAMP-inducible transcription of the TH gene. To understand the control mechanisms of TH gene transcription in greater detail, we sought to identify and characterize the transcription factors involved in recognition and activation of the CRE of the TH gene. Remarkably, electrophoretic mobility shift assay and antibody supershift experiments indicated that all three major CRE-binding protein factors, i.e. CREB, ATF1, and CREM, may participate in forming specific DNA/protein complexes with the CRE of the TH gene. To address the transcriptional activation function of individual factors, we replaced the TH CRE with a GAL4-binding site and cotransfected this modified TH promoter-reporter gene with an effector plasmid that encodes GAL4-fused transcription factor. Our results indicate that CREB but not ATF1 can support basal promoter activity while both can robustly induce the promoter activity in response to co-expression of the catalytic subunit of cAMP-dependent protein kinase (PKA). We further show that the coactivator CBP up-regulates PKA-mediated activation of the TH promoter and, if tethered to the TH promoter by a GAL4-fusion, can robustly transactivate the TH promoter even in the absence of PKA. Collectively, our results suggest that multiple CRE-binding factors interact with the CRE and regulate, in conjunction with the coactivator CBP, the transcriptional activity of the TH gene. (Mol Cell Biochem 212: 51–60, 2000) Key words: tyrosine hydroxylase, gene regulation, cAMP response element (CRE), cAMP response element binding protein (CREB), CREB-binding protein (CBP), activator transcription factor (ATF)

Introduction Tyrosine hydroxylase (TH; EC 1.14.16.2.) is the first and ratelimiting enzyme in the biosynthesis of catecholamine (CA) neurotransmitters and is responsible for the conversion of tyrosine to 3,4-dihydroxyphenylalanine (L-DOPA) [1]. Accordingly, TH gene expression may play a crucial role in regulating neurotransmission of CA neurotransmitters. TH is

selectively expressed in CA-synthesizing and secreting cells, including dopaminergic, noradrenergic and adrenergic neurons in the central nervous system and sympathetic ganglia and adrenal chromaffin cells in the periphery. Expression of the TH gene is subject to many physiological stimuli including those linked to the cAMP-signaling pathway (reviewed in [2, 3]). In numerous cAMP-inducible eukaryotic genes, an octamer DNA motif with the nucleotide sequence 5′-

Present address: J. Lim, Department of Animal Science and Biotechnology, College of Agriculture, Kyungbuk National University, Taeku, South Korea; C. Yang, Department of Biochemistry, St. Jude Children’s Research Hospital, 332 North Lauderdale, Memphis, TN 38105, USA Address for offprints: K.-S. Kim, Molecular Neurobiology Laboratory, McLean Hospital, Harvard Medical School, 115 Mill Street, Belmont, MA 02478, USA

52 TGACGTCA-3′, termed cAMP response element (CRE), mediates transcriptional induction by the cAMP-regulated signaling pathway [4, 5]. The 5′ flanking sequence of the rat TH gene contains a consensus CRE motif located at –38 to –45 bp upstream of the transcription initiation site [6]. This laboratory and others have previously demonstrated that the cAMP-regulated signaling pathway, via the CRE, regulates TH gene expression. In THexpressing cell lines such as the human neuroblastoma SKN-BE(2)C and rat PC12 [7, 8] or mouse CATH.a and PATH.a [9], site-directed mutation of the CRE significantly diminished both basal and cAMP-inducible transcription of the rat TH gene. Co-expression of the catalytic subunit of PKA (PKAc) dramatically increased the transcriptional activity of the rat TH gene in a dose-dependent manner, while coexpression of the specific inhibitor (PKI) blocks cAMP-stimulated induction and reduces basal transcriptional activity [10]. Furthermore, TH gene expression was found to be significantly attenuated at the transcriptional level in several PKAdeficient PC12 cell lines [11]. These studies strongly suggest that PKA, via the CRE, mediates cAMP-inducible transcription of the TH gene. The cAMP signaling pathway critically regulates transcription of the dopamine β-hydroxylase gene as well, suggesting that it may co-regulate catecholaminesynthesizing enzyme genes [10, 12]. In support of this, treatment of the specific PKA inhibitor N-[2-(p-bromocinnamylamine) ethyl]-5-isoquinoline sulfonamide (H-89) significantly diminished both basal and cAMP-induced transcription of all catecholamine-specific genes (i.e. TH, DBH, and PNMT) and resulted in a significant decreases of the intracellular levels of norepinephrine and epinephrine in the primary cultured bovine chromaffin cells [13]. Transcriptional regulation by the cAMP-signaling pathway is mediated by a family of transcription factors which bind to the CRE. These transcription factors have a similar DNAbinding structural motif, so-called leucine zipper, and include the CRE-binding protein (CREB), the CRE-modulatory protein (CREM), and ATF-1 (reviewed in [14]). Different protein kinases are known to phosphorylate these transcription factors, which appear to be critical in modulation of their transcriptional activities [14]. For instance, phosphorylation at serine 133 of CREB or at serine 63 of ATF1 induces recruitment of the coactivators CREB-binding protein (CBP) and p300, resulting in a prominent activation of the target genes [15–17]. Recent evidence from different laboratories has shown that these coactivators regulate transcription of a wide range of target genes by bridging the interacting transcription factors to the basal transcriptional machinery including THIIB [15], TBP [18], RNA polymerase II [19–21]. In addition, p300/CBP have recently been shown to modulate chromatin/nucleosome structure by acetylating histones, resulting in facilitation of access of transcription factors to the promoter sequences [22, 23].

To understand TH gene regulation by the cAMP-regulated pathway in greater detail, it is of great interest to determine which transcription factors bind to the CRE and how they interact with additional transcriptional factors. To address these questions, we sought to identify protein factors that interact with the CRE of the rat TH gene using the electrophoretic mobility shift assay and antibody supershift experiments. Surprisingly, using nuclear proteins prepared from TH-expressing SK-N-BE(2)C and Cath.a as well as TH-negative HeLa cell lines, our analyses show that all three major CRE-binding proteins, i.e. CREB, ATF1, and CREM, may interact with the TH CRE. We next tested whether individual transcription factors can stimulate the transcriptional activity of the TH gene in response to PKA in the context of the TH promoter using Gal4-fusion protein assay. We also tested whether the coactivator CBP may regulate the TH promoter activity. Together with previous study showing that an isoform of CREM regulates TH gene transcription [24], this study strongly suggests that multiple transcription factors including CREB and ATF1, in conjunction with the coactivator CBP/p300, may control transcriptional regulation of the TH gene in basal and/or cAMP-induced conditions.

Materials and methods Reporter plasmids The TH2.4CAT reporter gene contains 2.4 kb of the rat TH upstream sequences, fused to the bacterial chloramphenicol acetyltransferase (CAT) gene [11]. The reporter plasmid TH2.4(GAL4)CAT was constructed in two steps. First, the consensus CRE sequence (TGACGTCA) of the TH promoter was changed to a nonfunctional one (TGAAGTCG) by a site-directed mutagenesis procedure as described using an oligonucleotide of the sequence 5′-GCCAGGCTGAAGTCGAAGCCCCT-3′ [8]. Then, a single copy of GAL 4 binding site with the nucleotide sequence (5′-CGGAGGACTGTCCTCCG-3′) was subcloned at the Bgl II site residing at –168 bp from the start site. The sequences of the upstream regions of these reporter genes were verified by DNA sequencing analyses.

Effector plasmids The plasmid pGAL4-CREB encodes the chimeric protein composed of the full length CREB protein fused to the GAL4 DNA binding domain, while the plasmid pGAL4-CREB(SA119) encodes the same chimeric protein with a mutation at the serine119 residue (described as pNEXδ-WT and pNEXδSA119, respectively, in [25]). pGAL4-ATF1 and pGAL4ATF2 likewise encode full-length ATF1 and ATF2 fused to

53 the GAL4 DNA binding domain under the control of the SV40 promoter and enhancer [26]. pGAL4-CBP 1,678–2,441 encodes a chimeric protein of the GAL4 DNA binding domain and the C-terminal portion of the CBP, which encompasses a potential transactivation domain [16]. pRc/RSV-CBP is an expression vector encoding full-length mouse CBP under the control of Rous Sarcoma Virus promoter/enhancer [16]. The expression plasmid for the catalytic subunit of PKA (PKAc) was described before [11].

accommodate increasing amounts of pRc/RSV-CBP. Transfected cells were collected 36 h after transfection, and activities of chloramphenicol acetyl transferase (CAT) and β-galactosidase were determined as described elsewhere [30]. The CAT activity was normalized by the β-galactosidase activity to correct for differences in transfection efficiency among different DNA precipitates.

Electrophoretic mobility shift assay (EMSA) and antibody supershift experiment Cell culture and transient transfection assay Human neuroblastoma SK-N-BE(2)C [27, 28] and HeLa cell lines were maintained in Dulbecco’s modified Eagle’s medium (Life Technologies Inc.) containing 10% heat-inactivated fetal calf serum (Hyclone Laboratory), 100 units/ml of penicillin (Life Technologies Inc.), and 100 mg/ml of streptomycin (Life Technologies Inc.). Cath.a cells [29] were grown in RPMI 1640 medium (Life Technologies Inc.), supplemented with 10% heat-inactivated horse serum (Hyclone Laboratory), and 5% heat-inactivated fetal bovine serum. SKN-BE(2)C cells were transfected using the calcium phosphate co-precipitation method as described [30]. Briefly, cells were grown to 50–60% confluent in a 60 mm dish and were transfected with 1 µg of each reporter construct, 1 µg of pRSVβgal, varying amounts of pGAL4-fusion protein plasmid, and the inert plasmid, pUC19, to a total of 5.5 µg of DNA. When indicated, 0.5 µg of PKAc-expressing plasmid was included based on our previous results [11]. For experiments described in Figs 4 and 5, a total of 13.5 µg DNA was transfected to

Nuclear extracts were made from SK-N-BE(2)C and HeLa cell lines based on a described procedure [31]. Sense and antisense strands of oligonucleotides, which encompass the cAMP response element of the TH gene [7], were annealed and labeled by kination using [γ-32P]ATP. Specific antibodies were preincubated with nuclear extracts on ice for 2 h prior to adding the binding mixture. EMSA and antibody supershift experiments were performed using 30,000–50,000 cpm of labeled probe (approximately 0.05–0.1 ng) and nuclear extracts (20–30 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 [32]. For experiments described in Figs 1 and 2, nuclear extracts and labeled probe were added at the last step, respectively. It was noted that adding the labeled probe at the last step reduced the formation of a nonspecific band (C3). Antibodies used were against CREB (a gift from Dr. M. Montminy), ATF-1, ATF-2, ATF-3 (SantaCruz), c-fos, c-jun (Protooncogene), and CREM (Upstate

Fig. 1. Multiple CRE-proteins interact with the TH CRE. An oligonucleotide encompassing the TH CRE was radiolabeled and incubated with nuclear extracts isolated from the human neuroblastoma cell line. Unbound free probe (F) and three DNA-protein complexes (C1, 2, and 3) are indicated. Each nuclear extract was preincubated with increasing amounts of antibody (lanes 2, 4, 6, 8, 10, 12, 14, 16 and 19, 1 µl of 10–1; lanes 3, 5, 7, 9, 11, 13, 15, 17 and 20, 1 µl of 100; lanes 1 and 18, no antiserum) prior to binding reaction. Antibody against CREB (lanes 2, 3, 19 and 20), ATF1 (lanes 4 and 5), ATF2 (lanes 6 and 7), ATF3 (lanes 8 and 9), ATF4 (lanes 10 and 11), c-fos (lanes 12 and 13), c-jun (lanes 14 and 15), and CREM (lanes 16 and 17) are used.

54

Fig. 2. Antibody supershift experiments of the TH CRE oligonucleotide with nuclear extracts isolated from SK-N-BE(2)C, Cath.a, and HeLa cell lines. Antibody supershift assays are performed using 20–25 µg of different nuclear extracts. Antibodies against CREB (lanes 2, 6 and 10), CREM (lanes 3, 7 and 11), and ATF1 (lanes 4, 8 and 12) are preincubated prior to incubation with the radiolabeled probe. Co-incubation of different nuclear extracts with each antibody resulted in the generation of a supershifted band at a similar location with different intensity.

Biotech). Additional antibody against CREB (SC240; SantaCruz) was also used in antibody supershift experiments. DNAprotein complexes and DNA-protein-antibody complexes were resolved on nondenaturing 6% polyacrylamide gels as described [30, 32].

Results CREB, ATF1 and CREM may participate in forming protein/DNA complexes with the CRE of the TH gene Previous studies from this and other laboratories have demonstrated that cAMP signaling pathway critically controls transcriptional regulation of the TH gene via the CRE site. To identify potential transcription factor(s) that bind to the CRE and control TH gene transcription, the electrophoretic mobility shift assay (EMSA) and antibody supershift experiments were performed. Three DNA-protein complexes (C1, C2, and C3) were formed between the oligonucleotide encompassing the TH CRE and nuclear extracts of SK-NBE(2)C cells in EMSA, suggesting the possibility that more than one protein factor may control the transcriptional activity of the CRE of the TH gene (Fig. 1, lane 1). Antibodies against three CRE-binding proteins (ATF1, CREB, and CREM) and structurally related proteins such as ATF2, ATF3, ATF4, cfos, and c-jun were used to examine components of CRE/ protein complexes. Coincubation of antibody against CREB with nuclear extracts of SK-N-BE(2)C cells mostly diminished formation of C2 and resulted in formation of a super-

shifted band (Fig. 1, lanes 19 and 20). Interestingly, incubation of specific antibody against ATF1 also diminished most of the C2 complex and resulted in formation of a supershifted band at the position of C1 (Fig. 1, lanes 4 and 5). In addition, coincubation with antibody against CREM diminished C1 and C2 and resulted in formation of a supershifted band at higher location (Fig. 1, lane 16). Antibodies against CREM and CREB worked in a concentration-dependent manner; while antibody against CREM affected the formation of DNA-protein complexes at lower concentration, antibody against CREB worked better at higher concentration. Antibody against ATF1 worked equally well at both concentrations used in this assay. Notably, all three antibodies commonly blocked formation of C2. One possible interpretation is that C2 complex represents heterodimer between CREB, CREM, and/or ATF1, and thus can be recognized by these antibodies. An alternate, but mutually nonexclusive explanation is that these antibodies show some degree of cross-reactivity with CREB, CREM, and/or ATF1. Formation of C3 was not affected by any of these antibodies, suggesting that this complex may be formed by nonspecific binding. Consistent with this possibility, formation of C3 was not efficiently blocked by molar excess of cold oligonucleotides compared to those of C1 and C2 in the competition assay (data not shown). Incubation with antibodies against ATF2, ATF3, ATF4, c-fos, or c-jun apparently did not change the pattern of formation of DNA-protein complexes and no supershifted band was generated. Notably, co-incubation of another antibody against CREB (SC-240) neither significantly changed the pattern of DNA/protein complexes nor produced a supershifted band (Fig. 1, lanes 2 and

55 3), indicating that this antibody supershift experiment may result in different patterns depending on the properties of individual antibodies. Collectively, these results indicate that all three major CRE-binding transcription factors (i.e. CREB, ATF1, and CREM) are candidate transcription factors that bind to the CRE and regulate TH gene transcription in vivo. To test whether these CRE-binding factors are similarly expressed in different cell lines, we next performed antibody supershift experiments using nuclear extracts of another TH-expressing Cath.a and non-expressing HeLa cell lines (Fig. 2). In these experiments, we incubated the radiolabeled oligonucleotide probe after all other binding reagents including nuclear extracts and antibody were mixed (See Materials and methods). While the same pattern of three complexes (C1, C2, and C3) were formed, nonspecific complex C3 was reduced and C2 was increased in this procedure (compare lane 1 of Figs 1 and 2). When nuclear extracts of SK-NBE(2)C were co-incubated with specific antibodies against CREB, CREM, and ATF1, supershifted bands were generated at the same positions as in Fig. 1 and formation of C2 was diminished to a lesser degree, compared to Fig. 1. When nuclear extracts of Cath.a cell line were used, formation of C1 and C2 was more robust than those of SK-N-BE(2)C. Coincubation of Cath.a nuclear extracts with specific antibodies similarly diminished formation of C2 and generated supershifted bands with an almost identical pattern to SK-NBE(2)C. When HeLa nuclear extracts were co-incubated with these antibodies, supershifted bands were generated in a similar fashion. However, supershifted band generated by CREMspecific antibody was much weaker in HeLa, indicating that CREM protein may exist in a lower concentration in HeLa compared to SK-N-BE(2)C and Cath.a.

While CREB appears to support both basal and cAMPinducible promoter activity, ATF1 primarily supports cAMP-inducible promoter activity of the TH gene We next sought to investigate whether the transcription factors found to interact with the TH CRE can transactivate the basal and/or cAMP-inducible promoter activity of the TH gene. Because CREM is primarily involved in negative gene regulation [33] and one of its isoforms has been shown to repress the TH promoter activity [24], we have focused on ATF1 and CREB in this study. Based on the previous finding that the CRE is critical for both basal and cAMP-inducible promoter activity of the TH gene [8, 11], we have modified the reporter plasmid (TH2.4CAT) which contains the 2.4 kb upstream promoter fused to the reporter gene by two steps: (1) the CRE was rendered nonfunctional by introducing a double mutation to the CRE, and (2) a GAL4 binding site was inserted at –168 bp using the unique Bgl II site (Fig. 3A). As expected, this modified reporter construct

(TH2.4GAL4CAT) drove CAT activity at a level no higher than the promoter-less plasmid, pBLCAT3 [34]. Thus, this new reporter construct can be used for testing the transactivating function of a certain transcription factor which is fused to the GAL4 DNA-binding domain [35]. Expression plasmids encoding a hybrid protein, consisting of CREB, a mutant form of CREB (CREBm), ATF1, and ATF2 proteins fused to the GAL4 DNA binding domain, were used as the transactivator (Fig. 3A). In pGAL4-CREBm, the serine residue was changed to alanine at amino acid position 119 which is essential for phosphorylation of CREB by PKA. Co-transfection of pGAL4CREB increased the CAT activity driven by TH2.4GAL4CAT reporter construct by approximately 4-fold (Fig. 3B). Coexpression of the coding sequence of GAL4 DNA-binding domain (1–147) only did not affect the CAT activity driven by TH2.4GAL4CAT [35], ascribing the observed increase of CAT activities by pGAL4-CREB to CREB. Interestingly, cotransfection of pGAL4-CREBm increased the reporter gene activity to a similar degree (approximately 3-fold), suggesting that the serine residue at 119 amino acid position may not be critically required for supporting basal transcription of the TH gene. pGAL4-ATF1 increased the promoter activity of the TH2.4GAL4CAT construct only marginally, if any. Thus, CREB and ATF1 seem to exhibit differential ability to support basal transcriptional activity of the TH promoter. In a control experiment, co-transfection of pGAL4-ATF2 containing the full length coding region of ATF2 modestly transactivated the reporter gene activity by 2.5-fold (Fig. 3B). To address whether these transcription factors can activate TH gene transcription in response to cAMP-signaling pathway, we next co-transfected an expression plasmid encoding catalytic subunit of PKA (PKAc) together with the reporter construct and each effector plasmid. When PKAc was co-expressed, GAL4-CREB and GAL4-ATF1 dramatically increased the reporter gene activity by 112 and 158-fold, respectively (Fig. 3C). These results support the idea that CREB and ATF1 participate in PKA-inducible transcription of the TH gene. PKAc-induced transactivation by GAL4-CREB was mostly diminished (from 112 to 6-fold) by introducing a mutation (Ser to Ala) at the 119 position of the CREB protein, demonstrating that phosphorylation at this residue by PKA is crucial for cAMP-induced activation of the TH gene. In addition, PKAc barely affected transactivation function of pGAL4ATF2, showing its specific activation of CREB and ATF1.

Transcriptional cofactor CBP may be an important component of cAMP-inducible transcription of the TH gene and is able to transactivate the promoter activity when tethered to the proximal promoter area To test the role of CBP in cAMP-regulated activation of TH gene transcription, we co-transfected a CBP-expression plas-

56

A

B

C

Fig. 3. ATF1 and CREB transactivate the basal and PKA-mediated promoter activities of the TH gene. (A) Diagram of reporter and transactivator plasmids. TH2.4GAL4CAT plasmid containing a GAL4-binding site instead of a nonfunctional CRE is used as the reporter plasmid. Transcriptional start site is denoted by a bent arrow. Various plasmids encoding GAL4-fused transactivators are schematically shown with the promoters used to express these fusion molecules. (B) SK-N-BE(2)C cells were transiently co-transfected with TH2.4GAL4CAT reporter construct and the transactivator plasmids indicated below. The molar ratio of the reporter plasmid to the transactivator was 1 to 3. Samples were harvested 48–72 h after transfection, and relative CAT activities were determined as described in Materials and methods. To compare the fold transactivation directly, the basal CAT activity driven by reporter construct without the transactivator plasmid was set to 1. The significance of the effects of the transactivator plasmids is indicated as follows: n.s. – not significant; *p < 0.02; **p < 0.01; ***p < 0.001. (C) Co-transfection assay was performed in SK-N-BE(2)C cells with addition of 0.5 µg of expression plasmid of the catalytic subunit of PKA (PKAc), when indicate by +. Bars represent the normalized CAT activities driven by TH2.4GAL4CAT and each transactivator-expression plasmid, relative to those without PKAc. The relative values are presented as mean value ± S.E.M. from 6 independent samples. The significance of the effects of the PKAc-expression plasmid on CAT gene expression in the presence of each transactivator plasmid is indicated as follows: n.s. – not significant; *p < 0.02; **p < 0.01.

57 mid along with the reporter plasmid TH2.4CAT (Fig. 4). In agreement with our previous observation [10], co-expression of PKAc robustly transactivated the intact promoter activity of the TH gene (64-fold) in the absence of pCBP. Co-transfection of pCBP further increased the CAT activity driven by TH2.4CAT by an additional 2-fold (from 62 to 128-fold) in a dose-dependent manner showing that increased availability of CBP can further up-regulate PKA-induced transcription of the TH gene. We next tested whether CBP similarly can increase PKA-induced transactivation function of CREB. As shown in Fig. 5, co-transfection of pCBP further increased the CAT activity driven by pTH2.4GAL4CAT and pGAL4CREB from 98 to 187-fold in a dose-dependent manner. Similarly, pCBP also increased the reporter activity driven by pTH2.4GAL4CAT and pGAL4-ATF1 in a dose-dependent manner (data not shown). Finally, we co-transfected pTH2.4GAL4CAT and pGAL4CBP to the SK-N-(BE)2C cell line (Fig. 6). Strikingly, in the absence of ectopic expression of PKA, the full-length CBP protein fused to GAL4 was able to transactivate the modified TH promoter by 192-fold. This result show that tether-

Fig. 4. CBP enhances the PKA-induced promoter activity of the TH gene. SK-N-BE(2)C cells were transfected with TH2.4CAT, PKAc, and increasing amounts of pRc/RSV-CBP (CBP) as shown. The total amount of DNA was kept constant at 13.5 µg using the inert plasmid pUC19. The bars represent the normalized CAT activities relative to that driven by TH2.4CAT only, and are shown as mean value ± S.E.M. from 6 independent samples. The significance of the effects of the CBP-expressing plasmids, relative to the control, on PKAc-induced expression of the TH-CAT fusion gene is indicated as follows: *p < 0.02; **p < 0.01; ***p < 0.001. The transient transfection experiments were performed at least twice more in triplicate with similar patterns.

ing CBP to the TH promoter by GAL4 fusion alleviates the requirement of PKA activity for transactivation of the promoter activity and thus strongly suggest that CBP is an essential component in PKA-induced activation of TH transcription. In addition, co-expression of PKAc further increased the reporter activity up to 562-fold indicating that the transactivation function of CBP can be further facilitated by its phosphorylation. Alternatively, it is also possible that the promoter activity of the reporter construct is stimulated through other elements such as the AP-1 site.

Discussion In this study, we sought to identify transcription factors involved in the transcriptional regulation of the CRE of the TH gene. Previous studies have shown that the CRE, which resides at –38 to –45 bp upstream of the transcription initiation site, is crucial for not only cAMP-mediated induction but also basal transcription of the TH gene [7, 10, 36]. We demonstrate that all three major CRE-binding proteins, i.e. CREB,

Fig. 5. Expression of CBP is able to enhance the transactivation function of pGAL-CREB in the presence of PKAc. SK-N-BE(2)C cells were transfected with TH2.4GAL4CAT, PKAc, and increasing amount of pRc/RSVCBP (CBP) as shown. Transfection assays were performed as in Fig. 4 with a total 13.5 µg of DNA and the relative CAT activities are presented as mean value ± S.E.M. from 6 independent samples. The significance of the effects of the CBP-expressiong plasmids, relative to the control, on PKAc-induced CAT expression in the presence of pGAL-CREB and TH2.4GAL4CAT is indicated as follows: *p < 0.02; **p < 0.01; ***p < 0.001.

58

Fig. 6. GAL4-CBP can dramatically increase the CAT activity driven by TH2.4GAL4CAT. SK-N-BE(2)C cells were transfected with TH2.4GAL4CAT, PKAc, and GAL4-CBP. The total amount of DNA was kept constant at 5 µg. The bars represent the normalized CAT activities relative to that driven by TH2.4GAL4CAT only, and are shown as mean value ± S.E.M. from 6 independent samples. The significance of the effects of GAL4-CBP and PKAc-expressing plasmids on CAT expression driven by TH2.4GAL4CAT is indicated as follows: *p < 0.02; **p < 0.01. The transient transfection experiments were performed at least twice more in triplicate, resulting in similar patterns.

ATF1, and CREM, may participate in forming complexes with the CRE of the TH gene in TH-expressing human neuroblastoma SK-N-BE(2)C and mouse central catecholaminergic Cath.a cell lines. In contrast, EMSA of the present study did not detect AP-1 protein complex (e.g. c-fos and c-jun) or other related proteins (e.g. ATF2, 3, and 4) in the CRE/protein complexes. Previously, an isoform of the CREM family proteins, the inducible cAMP early repressor (ICER), has been shown to be induced by forskolin stimulation in PC12 cells and by reserpine treatment of rats in adrenal glands [24]. In addition, Liu et al. [37] recently showed that CREB and CREM contributed to formation of complexes with the TH CRE in the olfactory bulb dopaminergic neurons. Thus, our present data agree with these studies and further suggest that all three major CRE-binding transcription factors may contribute to formation of DNA/protein complexes with the CRE and regulation of the TH gene. We also tested whether ATF1 and CREB can support basal and cAMP-inducible transcriptional activity of the TH gene promoter. To this end, based on our previous analysis

of the structure/function relationship of the TH CRE [8], we made the CRE of the TH-reporter construct be nonfunctional by introducing a double mutation and then inserted a GAL4binding site (Fig. 3A). Because this mutation rendered the whole 2.4 kb upstream promoter to be as inactive as the promoter-less construct, we were able to assess the transactivation function of each transcription factor which is fused to the GAL4 DNA-binding domain. These co-transfection assays in SK-N-BE(2)C cells showed that CREB modestly activated the basal TH promoter activity (approximately 4-fold) while ATF1 minimally activated it (approximately 1.6-fold). Interestingly, a mutant form of CREB containing alanine at 119 amino acid position instead of serine similarly activated the basal TH promoter activity (approximately 3-fold), indicating that the basal transactivation function of CREB may not depend on phosphorylation by PKA at this position. It is to be noted that exogenous expression of GAL4-CREB did not restore the promoter activity of the modified reporter construct (TH2.4GAL4CAT) to the level of the wild type promoter activity of TH2.4CAT (Fig. 3; [8]). This apparent discrepancy can be explained by several possibilities. First, in our modified reporter construct, the GAL4-binding site was located at a more distal position (–168 bp) compared to the CRE (– 45 to –38 bp) of the wild type promoter. The transactivator may not work efficiently at this new site compared to the original site. However, our previous work does not support this possibility because the CRE worked as efficient as the wild type when located within a window of approximately 200 bp upstream of the transcription start site [8]. Secondly, GAL4-fusion of CREB may have rendered it a less efficient transactivator compared to CREB, possibly due to its structural changes. Thirdly, CREB may be able to transactivate the TH CRE efficiently only in concert with additional CREbinding proteins. For instance, since CREB and ATF1 can form heterodimers, one possibility is that these heterodimeric forms have more potent transactivation function to the TH promoter. In contrast to their limited ability to support the basal activity of the modified TH promoter, both GAL4-CREB and GAL4-ATF1 were shown to dramatically up-regulate the reporter activity in the presence of PKA (Fig. 3C). Mutation of the serine residue to alanine at 119 amino acid position diminished the PKA-mediated transactivation by CREB from 112 to 6-fold, indicating that this serine residue of CREB is critical for PKA-mediated activation of TH transcription. In addition, our data showed that coexpression of the full length CBP further up-regulated the transactivation function of both GAL4-CREB and GAL4-ATF1 in a dose-dependent manner (Fig. 4; data not shown). Furthermore, co-expression of GAL4-CBP dramatically activated the CAT activity driven by TH2.4GAL4CAT by 192-fold in the absence of exogenous PKA expression (Fig. 6). Recently, transcriptional coactivator CBP/p300 have been shown to critically regulate transcrip-

59 tion by interacting with the basal transcriptional machinery including THIIB [16], TBP [18], RNA polymerase II [19–21] as well as by modulating chromatin/nucleosome structure [22, 23]. Additionally, CBP/p300 have been shown to interact with ever increasing numbers of transcription factors including YY1 [38], MyoD [39], nuclear factor 1 [40], and Phox2a/Arix [41]. This converging evidence suggests that CBP/p300 may coordinate development stage-, second messenger-, and cell typespecific control of a wide range of genes. In summary, we showed that multiple members of CREbinding proteins may interact with the TH CRE in TH-positive cells and are capable of differentially supporting basal and cAMP-inducible promoter activity of the TH gene. Together with a previous study showing that an isoform of CREM, ICER, is induced in certain physiological conditions and down-regulates TH gene transcription [24], this study strongly suggests that multiple transcription factors including CREB and ATF1, in conjunction with the coactivator CBP/p300, may intricately control transcriptional regulation of the TH gene in basal and/or cAMP-induced conditions.

Acknowledgements The authors thank Dr. David Stone at McLean Hospital, Harvard Medical School, for critical reading of the manuscript. We also thank Drs. M. Montminy, R.H. Goodman, M. Green and B. Kaang for CREB-specific antibody, CBP-expressing plasmids, ATF1- and ATF2-expressing plasmids and CREBexpressing plasmids, respectively. This work was supported by NIH grants MH48866.

References 1. 2.

3. 4.

5. 6.

7.

8.

Nagatsu T, Levitt M, Udenfriend S: Tyrosine hydroxylase; the initial step in norepinephrine biosynthesis. J Biol Chem 239: 2910–2917, 1964 Zigmond RE, Schwarzschild MA, Rittenhouse AR: Acute regulation of tyrosine hydroxylase by nerve activity and by neurotransmitters via phosphorylation. Annu Rev Neurosci 12: 415–461, 1989 Kumer SC, Vrana KE: Intricate regulation of tyrosine hydroxylase activity and gene expression. J Neurochem 67: 443–462, 1996 Roesler WJ, Vandenbark GR, Hanson RW: Cyclic AMP and the induction of eukaryotic gene transcription. J Biol Chem 263: 9063–9066, 1988 Goodman R: Regulation of neuropeptide gene expression. Annu Rev Neurosci 13: 111–127, 1990 Lewis EJ, Harrington CA, Chikaraishi DM: Transcriptional regulation of the tyrosine hydroxylase gene by glucocorticoid and cAMP. Proc Natl Acad Sci USA 84: 3550–3554, 1987 Kim KS, Lee MK, Carroll J, Joh TH: Both the basal and inducible transcription of the tyrosine hydroxylase gene are dependent upon a cAMP response element. J Biol Chem 268: 15689–15695, 1993 Tinti C, Yang C, Seo H, Conti B, Kim C, Joh TH, Kim KS: Structure/ function relationship of the cAMP response element in tyrosine hydroxylase gene transcription. J Biol Chem 272: 19158–19164, 1997

9. Lazaroff M, Patankar S, Yoon SO, Chikaraishi DM: The cyclic AMP response element directs tyrosine hydroxylase expression in catecholaminergic central and peripheral nervous system cell lines from transgenic mice. J Biol Chem 270: 21579–21589, 1995 10. Kim KS, Tinti C, Song B, Cubells JF, Joh TH: Cyclic AMP-dependent protein kinase regulates basal and cAMP-stimulated but not phorbol ester-stimulated transcription of the tyrosine hydroxylase gene. J Neurochem 63: 834–842, 1994 11. Kim KS, Park DH, Wessel TC, Song B, Wagner JA, Joh TH: A dual role of the cAMP-dependent protein kinase on tyrosine hydroxylase gene expression. Proc Natl Acad Sci USA 90: 3471–3475, 1993 12. Ishiguro H, Kim KT, Joh TH, Kim KS: Neuron-specific expression of the human dopamine β-hydroxylase gene requires both the cAMP-response element and a silencer region. J Biol Chem 268: 17987–17994, 1993 13. Hwang O, Park SY, Kim KS: Inhibition of PKA with H-89 causes reduction in basal gene expression of the catecholamine enzymes. J Neurochem 68: 2241–2247, 1997 14. Montminy M: Transcriptional regulation by cyclic AMP. Annu Rev Biochem 66: 807–822, 1997 15. Chrivia JC, Kwok RP, Lamb N, Hagiwara M, Montminy MR, Goodman RH: Phosphorylated CREB binds specifically to the nuclear protein CBP. Nature 365: 855–859, 1993 16. Kwok RP, Lundblad JR, Chrivia JC, Richards JP, Bachinger HP, Brennan RG, Roberts SG, Green MR, Goodman RH: Nuclear protein CBP is a coactivator for the transcription factor CREB. Nature 370: 223–226, 1994 17. Lundblad JR, Kwok RP, Laurance ME, Harter ML, Goodman RH: Adenoviral E1A-associated protein p300 as a functional homologue of the transcriptional co-activator CBP. Nature 374: 85–88, 1995 18. Dallas PB, Yaciuk P, Moran E: Characterization of monoclonal antibodies raised against p300: both p300 and CBP are present in intracellular TBP complexes. J Virol 71: 1726–1731, 1997 19. Cho H, Orphanides G, Sun X, Yang XJ, Ogryzko V, Lees E, Nakatani Y, Reinberg D: A human RNA polymerase II complex containing factors that modify chromatin structure. Mol Cell Biol 18: 5355–5363, 1998 20. Kee BL, Arias J, Montminy MR: Adaptor-mediated recruitment of RNA polymerase II to a signal-dependent activator. J Biol Chem 271: 2373–2375, 1996 21. Nakajima T, Uchida C, Anderson SF, Lee CG, Hurwitz J, Parvin JD, Montminy M: RNA helicase A mediates association of CBP with RNA polymerase II. Cell 90 :1107–1112, 1997 22. Ogryzko VV, Schiltz RL, Russanova V, Howard BH, Nakatani Y The transcriptional coactivators p300 and CBP are histone acetyltransferases. Cell 87: 953–959, 1996 23. Wolffe AP, Pruss D: Targeting chromatin disruption: Transcription regulators that acetylate histones. Cell 84: 817–819, 1996 24. Tinti C, Conti B, Cubells J, Kim KS, Baker H, Joh, TH: Expression of ICER in PC12 and rat adrenal gland: A role for tyrosine hydroxylase gene transcription repression. J Biol Chem 271: 25375–25381, 1996 25. Kaang BK, Kandel ER, Grant SG: Activation of cAMP-responsive genes by stimuli that produce long-term facilitation in Aplysia sensory neurons. Neuron 10: 427–435, 1993 26. Liu F, Green MR: A specific member of the ATF transcription factor family can mediate transcription activation by the adenovirus E1a protein. Cell 61:1217–1224, 1990 27. Ross RA, Biedler JL, Spengler BA, Reis DJ: Neurotransmitter-synthesizing enzymes in 14 human neuroblastoma cell lines. Cell Mol Neurobiol 1: 301–311, 1981 28. Ciccarone V, Spengler BA, Meyers MB, Biedler JL, Ross RA: Phenotypic diversification in human neuroblastoma cells: Expression of distinct neural crest lineages. Cancer Res 49: 219–225, 1989

60 29. Suri A, Fung BP, Tischler AS, Chikaraishi DM: Catecholaminergic cell lines from the brain and adrenal glands of tyrosine hydroxylase-SV40 T antigen transgenic mice. J Neurosci 13: 1280–1291, 1993 30. Seo H, Yang CY, Kim H-S, Kim KS: Multiple protein factors interact with the cis-regulatory elements of the proximal promoter in a cellspecific manner and regulate transcription of the dopamine β-hydroxylase gene. J Neurosci 16: 4102–4112, 1996 31. Dignam JD, Lebovitz RM, Roeder RG: Accurate transcription initiation by RNA polymerase II in a soluble extract from isolated mammalian nuclei. Nucleic Acids Res 11: 1475–1489, 1983 32. Kim HS, Seo H, Brunet JF, Kim KS: Noradrenergic-specific transcription of the dopamine β-hydroxylase gene requires synergy of multiple cis-regulatory elements including at least two Phox2a-binding sites. J Neurosci 18: 8247–8260, 1998 33. Foulkes NS, Borrelli E, Sassone-Corsi P: CREM gene: Use of alternative DNA-binding domains generates multiple antagonists of cAMPinduced transcription. Cell 64: 739–749, 1991 34. Luckow B, Schutz G: CAT constructions with multiple unique restriction sites for the functional analysis of eukaryotic promoters and regulatory elements. Nucleic Acids Res 15: 5490, 1987 35. Sadowski I, Ptashne M: A vector for expressing GAL4(1-147) fusion in mammalian cells Nucleic Acids Res 17: 7539, 1989 36. Lazaroff M, Patankar S, Yoon SO, Chikaraishi DM: The cyclic AMP

37.

38. 39.

40.

41.

response element directs tyrosine hydroxylase expression in catecholaminergic central and peripheral nervous system cell lines from transgenic mice. J Biol Chem 270: 21579–21589, 1995 Liu N, Cigola E, Tinti C, Jin BK, Conti B, Volpe BT, Baker H: Unique regulation of immediate early gene and tyrosine hydroxylase expression in the odor-deprived mouseolfactory bulb. J Biol Chem 274: 3042– 3047, 1999 Galvin KM, Shi Y: Multiple mechanisms of transcriptional repression by YY1. Mol Cell Biol 17: 3723–3732, 1997 Sartorelli V, Huang J, Hamamori Y, Kedes L: Molecular mechanisms of myogenic coactivation by p300: Direct interaction with the activation domain of MyoD and with the MADS box of MEF2C. Mol Cell Biol 17: 1010–1026, 1997 Leahy P, Crawford DR, Grossman G, Gronostajski RM, Hanson RW: CREB binding protein coordinates the function of multiple transcription factors including nuclear factor I to regulate phosphoenolpyruvate carboxykinase (GTP) gene transcription. J Biol Chem 274: 8813–8822, 1999 Swanson DJ, Adachi M, Lewis EJ: The homeodomain protein arix promotes protein kinase A-dependent activation of the dopamine betahydroxylase promoter through multiple elements and interaction with the coactivator cAMP-response element-binding protein-binding protein. J Biol Chem 275: 2911–2923, 2000