Identification of ATF-2 as a Transcriptional Regulator for the Tyrosine Hydroxylase Gene*

Transcriptional regulation of catecholamine-synthesizing genes is important for the determination of neurotransmitters during brain development. We found that three catecholamine-synthesizing genes were transcriptionally up-regulated in cloned PC12D cells overexpressing V-1, a protein that is highly expressed during postnatal brain development (1). To reveal the molecular mechanism to regulate the expression of tyrosine hydroxylase (TH), which is the rate-limiting enzyme for catecholamine biosynthesis, we analyzed the transcription factors responsible for TH induction in the V-1 clonal cells. First, by using reporter constructs, we found that the transcription mediated by cAMP-responsive element (CRE) was selectively enhanced in the V-1 cells, and TH promoter activity was totally dependent on the CRE in the promoter region of the TH gene. Next, immunoblot analyses and a transactivation assay using a GAL4 reporter system revealed that ATF-2, but not cAMP-responsive element-binding protein (CREB), was highly phosphorylated and activated in the V-1 cells, while both CREB and ATF-2 were bound to the TH-CRE. Finally, the enhanced TH promoter activity was competitively attenuated by expression of a plasmid containing the ATF-2 transactivation domain. These data demonstrated that activation of ATF-2 resulted in the increased transcription of the TH gene and suggest that ATF-2 may be deeply involved in the transcriptional regulation of catecholamine-synthesizing genes during neural development.

L-amino acid decarboxylase (AADC), dopamine to norepinephrine by dopamine ␤-hydroxylase (DBH), and norepinephrine to epinephrine by phenylethanolamine N-methyltransferase. The regulation of the gene expression of these enzymes is important for the determination of the expression of neurotransmitters during brain development as well as for brain function under physiological and pathological conditions. TH is the rate-limiting enzyme for catecholamine biosynthesis. The transcriptional regulation of the TH gene has been extensively studied, and many transcription factors were suggested to regulate TH gene expression. CREB, ATF-1, and CREM were shown to recognize the cAMP-responsive element (CRE) located at position Ϫ45/Ϫ38 in the TH promoter (2)(3)(4)(5)(6). CREB was reported to mediate basal and cAMP-induced TH transcription in various cultured cells including PC12 cells by the use of dominant-negative CREB protein and antisense RNA against CREB (7)(8)(9)(10). CREB is activated by phosphorylation on Ser 133 (11,12). The activation of CREB by phosphorylation has been shown to mediate PKA-dependent (7,(13)(14)(15) and -independent (13,14) induction of TH transcription. Functional and physiological roles of ATF-1 and CREM for TH transcription remained unclear. AP-1 transcription factors bound to the TPA-responsive element located at Ϫ205/Ϫ199 in the TH promoter, which also plays a critical role in the transcriptional induction of the TH gene (16). Expression of AP-1 family was induced by several stimuli to enhance TH gene transcription, and overexpression of Fra-2 and c-Fos stimulated TH transcription in PC12 and PC18 cells, respectively (17,18).
Nurr1 and Ptx3/Pitx3, which are expressed in midbrain dopaminergic neuron at embryonic stage, were recently shown to regulate TH gene expression. Nurr1-null mice lacked TH immunoreactivity in midbrain (19), and overexpression of Nurr1 was shown to induce the TH gene expression in neural stem cells (20,21). However, it is unclear whether or not Nurr1 directly acts on its responsive element in the TH promoter region (20,21). Overepxression of Ptx3/Pitx3 activated the TH promoter activity in neuroblastoma cells through its responsive elements (22,23).
Three of the catecholamine-synthesizing enzymes, i.e. TH, AADC, and DBH, were up-regulated in cloned PC12D cells overexpressing V-1 (1). V-1 is an adaptor-like protein, the expression of which is transiently high during postnatal brain development (24,25). Therefore, V-1 potentially participate TH gene expression in postnatal brain development. We were interested in the intracellular events in the V-1-overexpressing PC12D cells, because none of the transcription factors are known to induce these catecholamine-synthesizing genes simultaneously.
In the present study, we found that the transcription mediated by CRE was specifically and highly elevated in the V-1 cells, and that transcription of the TH gene is enhanced by activation of ATF-2, which directly acted on the CRE in the TH promoter. Our results suggest that ATF-2 is involved in the expression of the TH gene during neural development.

MATERIALS AND METHODS
Cell Culture-Two PC12D clones stably and highly expressing V-1, named V1-46 and V1-69, and two vector control clones, termed C-7 and C-9, were established as described previously (1). Parental PC12D cells were cultured in Dulbecco's modified Eagle's medium containing 10% horse serum and 5% fetal bovine serum. The stable clones were cultured in medium containing 280 g/ml G418 (Invitrogen).
Reporter Plasmids-Mouse TH genomic DNA containing 5.5 kb of its 5Ј-flanking region, and exon 1 was originally isolated from a mouse genomic library and cloned into the pBluescript II KS(ϩ) plasmid (mTH5.5-pBS). First, to produce mTHpro0.8-Luc, which contains the region of Ϫ752 bp from the transcriptional start site to the translational start site of the mouse TH gene, we amplified a fragment containing about 1.2 kb of the mouse TH 5Ј-flanking region by PCR using primers GTTCCCTTAGTGAGAGGACAC (forward) and GGAATTCCATGGTG-CAAGCTGGTGGTC (reverse) and mTH5.5-pBS. The region between Ϫ752 bp from the transcriptional start position and the translational start site was ligated to a firefly luciferase reporter plasmid, PGV-B2 (Toyoink, Tokyo). The entire sequence of the inserted TH 5Ј-flanking region was confirmed by DNA sequencing using an ABI PRISM 310 genetic analyzer (Applied Biosystems). Next, to produce mTHpro4.3-Luc, a fragment that corresponded to the region between Ϫ4326 and Ϫ753 bp from the transcriptional start position was isolated by the digestion of mTH5.5-pBS with XhoI and then inserted into the XhoI site of mTHpro0.8-Luc. To produce a CRE mutant of mTHpro0.8-Luc that contained a 4-base deletion in a canonical CRE sequence of the TH promoter region (TGACGTCA), we digested the wild-type mTHpro0.8-Luc with AatII to result in a plasmid cut in the middle of the CRE sequence, blunted the end with T4 exonuclease, and then religated it.
Firefly luciferase reporter plasmids containing a cyclic AMP-, AP-1-, NF-B-, glucocorticoid-, heat shock protein-, or serum-responsive element upstream of a TATA-like promoter (P TAL ) region taken from herpes simplex thymidine kinase promoter were purchased from Clontech. The following plasmids were purchased from Stratagene as parts of PathDetect trans-reporting systems: pFC-MEKK and pFC-PKA are MEKK (amino acids 380 to 672) and PKA (catalytic subunit) expression vectors, respectively; pFA2-CREB and pFA-ATF-2 are expression vectors that consist of CREB (amino acids 1-280) and ATF-2 (amino acids 1-96) transactivation domains fused to the GAL4 DNA binding domain (GAL4 DBD , amino acids 1-147), respectively; pFC-dbd is the control vector for GAL4 DBD only; and pFR-Luc reporter is a reporter plasmid containing five copies of the GAL4-responsive element.
DNA Transfection and Luciferase Assay-Seapansy luciferase vectors, pRL-CMV and pRL-TK (Toyoink), were used as an internal control to normalize for variations in transfection efficiency. Cells were transfected by lipofection using LipofectAMINE 2000 (Invitrogen). One day prior to transfection, the cells were plated on 24-well plates and transfected at ϳ50% confluence. In the experiments shown in Figs. 1-3, the cells were transfected with 0.75 g of the firefly reporter plasmids and 0.05 g of pRL-CMV per well. In experiments using the GAL4 system shown in Fig. 6, cells were transfected with 0.65 g of pFR-Luc, 0.05 g of pFA2-CREB or pFA-ATF-2, and 0.05 g of pRL-TK per well. As a positive control, 0.05 g of a PKA-expression vector (PKA) or a MEKKexpression vector (MEKK) was used for co-transfection of the parental PC12D cells, and otherwise, pBluescript plasmid was used as a carrier DNA. In the experiments shown in Fig. 7, the cells were transfected with 0.25 g of mTHpro4.3-Luc or PGV-C2 (a SV40 promoter luciferase vector, Toyoink), 0.5 g of pFA-ATF2, pFA2-CREB, or pFC2-dbd, and 0.05 g of pRL-CMV per well. At 48 h after transfection, the cells were harvested and assayed for firefly and seapansy luciferase activities by using a PicaGene Dual luciferase assay kit (Toyoink).
Preparation of Cell Lysates-Cells were washed three times, suspended in ice-cold phosphate-buffered saline, and then pelleted in a microcentrifuge at 300 ϫ g for 3 min. For preparation of nuclear ex-tracts for electrophoretic mobility shift assays and immunoblot analysis of phosphorylated proteins, the cell pellet was lysed, and nuclear proteins were extracted as described (26) except that all buffers contained 0.10 volume of a Protease Inhibitors Cocktail (Sigma) substitute for phenylmethylsulfonyl fluoride and contained 0.01 volume of Phosphatase Inhibitor Cocktails I and II (Sigma). Protein concentration of the nuclear extract was determined by the method of Bradford (27), with bovine ␥-globulin used as a standard. For preparation of whole-cell extracts for immunoblot analysis of phosphorylated proteins, the cell pellet was directly lysed in SDS-sample buffer, and the supernatant was collected as the whole-cell extract. The cell lysates were stored at Ϫ80°C in small aliquots until assayed.
Electrophoretic Mobility Shift Assay-Sense and antisense strands of rat TH-CRE oligonucleotides (sense, GAGGGGCTTTGACGTCAGC-CTGG) were annealed and end-labeled with [␥-32 P]ATP (6000 Ci/mmol, PerkinElmer Life Sciences) by use of T4 polynucleotide kinase. DNAprotein binding reactions were performed as described by Nankova et al. (18) with a slight modification. Briefly, the basic binding buffer contained 0.5 g/l bovine serum albumin, 0.1 g/l poly(dI-dC), 1 ng of end-labeled TH-CRE oligonucleotide (ϳ35,000 cpm), and 10 g of nuclear extract. Double-stranded oligonucleotides or antibodies (1 l; Cell Signaling Technology) were preincubated on ice for 1 h prior to the addition of the labeled oligonucleotide, and the reaction was initiated by the addition of the labeled oligonucleotide. The mixture was then incubated for 20 min at room temperature. Electrophoresis was performed as described by Kapatos et al. (28), and radioactivity was visualized by exposing the x-ray film for 12 h at Ϫ80°C.
Immunoblot Analysis-Phosphorylated form-specific and nonspecific antibodies against CREB, ATF-2, and c-Jun were purchased from Cell Signaling Technology. Immunoblotting were performed following the supplier's protocol. The cell lysate was separated by SDS-PAGE and transferred to a polyvinylidene difluoride membrane (Bio-Rad). Proteins were visualized with ECL plus (Amersham Biosciences).
Statistics-Student's t test was used for statistical evaluations. A level of p Ͻ 0.05 was accepted as statistically significant.

Increased Promoter Activity of the TH Gene in Clonal PC12D
Cells Overexpressing V-1-We previously described that TH enzymatic activity, protein level, and mRNA level were all elevated in V-1-overexpressing clones (V1-46 and V1-69) compared with their values for the control (C-7 and C-9) clones (1). To examine the transcription of the TH gene in the V-1 and control clones, we transfected the cloned cells with plasmid constructs containing 4.3 kb of mouse TH 5Ј-flanking region fused to a luciferase reporter gene (mTHpro4.3-Luc). Reporter activity relative to pRL-CMV was significantly increased in the V-1 clones compared with that in the control clones ( Fig. 1), suggesting that the increased level of TH mRNA in the V-1 clones was mainly due to an increased transcriptional rate and that cis-acting DNA elements located within the 4.3 kb of 5Ј-flanking region of the TH gene were required for the overexpression of the TH gene.
CRE-mediated Transcription Was Increased in the V-1 Clones-To explore the transcriptional events changed in the V-1 clones, we measured promoter activities of reporter genes containing cAMP-, AP-1-, NF-B-, glucocorticoid-, heat shock protein-, and serum-responsive elements. In the V-1 clones, we found that the transcriptional activity mediated by CRE was greatly elevated compared with that in the control clones, whereas that mediated by the other elements was unchanged (Fig. 2).
Evaluation of CRE-dependent Expression of the TH Gene in the V-1 Clones-A CRE consensus motif (TCACGTCA) exists in the 5Ј-flanking region of the rat TH gene (29), and the sequence and its location (near the TATA-box) are highly conserved among rat, mouse, and man (30). It was earlier shown that the TH-CRE was essential for both basal and cyclic AMP-induced transcription of the TH gene in various cell lines, including PC12 cells (3, 31) and a subclone of PC12 (29). To evaluate the contribution of the increased activity of CRE-mediated transcription to the increased transcription of the TH gene in the V-1 clonal cells, we made wild-type and CRE-mutagenized TH reporter vectors containing about 0.8 kb of the mouse TH 5Ј-flanking region upstream from the transcriptional start site (mTHpro0.8-Luc; Fig. 3A). The reporter activity of the wildtype mTHpro0.8-Luc was markedly increased in V1-69 cells (Fig. 3B), as had been the case for mTHpro4.3-Luc (Fig. 1). The reporter activity of the CRE mutant was almost the same as that of the promoter-less vector, PGV-B2 in V1-69, C-7, and the parental PC12D cells (Fig. 3B). Results for the other V-1 clone, V1-46, were similar to those for V1-69 (data not shown). These results indicated the importance of the CRE-mediated transcription for the regulation of the TH gene expression in the V-1 cells.
Identification of ATF-2 as a Transcription Factor Binding to TH-CRE in PC12D Cells-To identify transcriptional factors that might be involved in the CRE-mediated transcription of the TH gene, we conducted an electrophoretic mobility shift assay using nuclear extracts of PC12D cells and analyzed binding proteins that made a complex with the TH-CRE (Fig. 4). There were major bands (Bands 1-3) that disappeared in a dose-dependent manner by the addition of excess amounts of cold TH-CRE competitor (Fig. 4, lanes 2-5). However, an excess molar amount of TH-CRE mutant also decreased the amount of Band 3 to the same extent as did the TH-CRE wild type, whereas Bands 1 and 2 remained almost unaffected (Fig. 4, lanes 6 -8), indicating that Bands 1 and 2 represented the protein-DNA complexes specific to the TH-CRE. An anti-CREB antibody supershifted Band 2 to Band B (Fig. 4, lane 13), and an anti-ATF-2 antibody supershifted Band 1 to Band A (Fig. 4, lane 14), indicating that Bands 1 and 2 represented complexes containing ATF-2 and CREB, respectively. Since ATF-2 was reported to bind to CRE as a homodimer or as a heterodimer with c-Jun (32-34), we examined whether c-Jun was contained in the TH-CRE complex. An anti-c-Jun antibody did not supershift Band 1 containing ATF-2 (Fig. 4, lane 15).
We also conducted an electrophoretic mobility shift assay using TH-CRE and nuclear extracts of V-1 and control clones. There was no significant difference in band patterns between V-1 or control clones and the parental PC12D cells (data not shown). These data suggest that proteins binding to the TH-CRE, including CREB and ATF-2, were not dramatically changed in the V-1 clones.
Activation of ATF-2 in the V-1 Clones-CREB and ATF-2 are well characterized among ATF/CRE binding proteins; phosphorylation of CREB on Ser 133 (11,12) or phosphorylation of ATF-2 on Thr 69/71 (35,36) activates the CRE-mediated transcription. Since we identified CREB and ATF-2 as transcription factors binding to the TH-CRE as was shown in Fig. 4, we next examined the phosphorylation and expression of CREB and ATF-2 in the V-1 clones by Western blot analysis using phosphorylated form-specific and nonspecific antibodies against CREB or ATF-2. As positive controls, PC12D cells were treated with forskolin (FSK) or nerve growth factor (NGF) for 15 min (11,37). In whole-cell extracts of FSK-or NGF-stimulated PC12D cells, the Ser 133 -phosphorylated form of CREB was much increased (Fig. 5A, upper panels). Faster running bands that were immunostained with the antibody against CREB phosphorylated on Ser 133 appeared, and the intensity of bands immunostained with anti-CREB antibody was reduced by FSKand NGF-stimulation (Fig. 5A, upper panels), suggesting that CREB was rapidly degraded after its activation. In addition, phosphorylation of ATF-2 was also increased by FSK or NGF stimulation in PC12D cells, while the total amount of ATF-2 protein was unchanged (Fig. 5A, lower panels). In the V-1 clones, although there were no or few changes in the expression and phosphorylation of CREB compared with those in the control clones and the non-stimulated parental PC12D cells  (Fig. 5A, upper panels), the level of the phosphorylated form of ATF-2 (on Thr 71 ) was increased without any significant change in the expression of ATF-2 (Fig. 5A, lower panels).
Previous mutation analysis of Thr 69 and Thr 71 of ATF-2 strongly indicated that the dual phosphorylated form of ATF-2, i.e. the protein phospho-Thr 69 and -Thr 71 , is the transcriptionally active form (35,36). This form of ATF-2 was greatly increased in nuclear extracts of the V-1 clones, compared with that in those of the control clones (Fig. 5B). On the other hand, the amount of ATF-2 was not increased in the nuclear extracts of the V-1 clones (Fig. 5B).
We also used the GAL4 reporter assay to examine expression vectors expressing the transactivation domain of CREB or ATF-2 (CREB TAD , amino acids 1-280 or ATF-2 TAD , amino acids 1-96, respectively) fused to a GAL4-DNA binding domain (GAL4 DBD ). Although the GAL4 DBD -CREB TAD activity in the parental PC12D cells transiently co-transfected with a PKA expression vector was dramatically increased, this activity in the V-1 clones was unchanged compared with that in the control clone and the non-stimulated parental PC12D cells (Fig.  6A). In contrast, the GAL4 DBD -ATF-2 TAD activity in the V-1 clones was greatly increased and was comparable with that in PC12D cells transiently co-transfected with an active MEKK expression vector as a positive control (Fig. 6B). These data demonstrate that, ATF-2, as a TH-CRE-binding protein, was highly phosphorylated and activated in the V-1 clones.

Enhanced Activity of the TH Promoter in the V-1 Clones Was Attenuated by the Expression of the ATF-2 Transactivation
Domain-To examine whether the enhanced activity of the TH gene transcription was due to activation of ATF-2, we adapted the GAL4 DBD -ATF-2 TAD expression vector to the reporter assay for the TH promoter. Overexpression of the GAL4 DBD -ATF-2 TAD protein containing the phosphorylation sites was expected to interfere competitively with the activation of the endogenous ATF-2 protein for the CRE-mediated TH gene transcription. The reporter activity of mTHpro0.8-Luc in the V1-69 clone was greatly decreased by co-transfection with the GAL4 DBD -ATF-2 TAD expression vector compared with that when the control vector that expressed GAL4 DBD only was used for co-transfection, whereas that in the parental PC12D cells was slightly, but not significantly, decreased (Fig. 7A). The increased TH promoter activity in the V-1 clone remained unchanged by cotransfection with the GAL4 DBD -CREB TAD expression vector in contrast to the GAL4 DBD -ATF-2 TAD expression vector (Fig. 7A). In contrast to the TH promoter, the SV40 promoter was not activated in the V-1 clone and not affected by the expression of either GAL4 DBD -ATF-2 TAD or GAL4 DBD -CREB TAD (Fig. 7B). DISCUSSION In the present study, we found that CRE-mediated transcription was highly elevated in V-1-overexpressing clones of the PC12D cell line. We then identified ATF-2 as one of the TH-CRE-binding proteins in PC12D cells, and, for the first time, found that activation of ATF-2 up-regulated the TH gene transcription via CRE.
ATF-2 (32), also called CRE-BP1 (34), is a member of the ATF/CREB family of transcription factors that bind to CRE. The TH-CRE was shown to be recognized by other members of the ATF/CREB family, i.e. CREB, ATF-1, and CREM (2-6). Among these transcription factors, only CREB was shown to be an activator of the TH gene transcription (7)(8)(9)(10). Our data confirmed that CREB is one of the TH-CRE-binding proteins in PC12D cells (Fig. 4). However, our data showed that the enhanced transcription of the TH gene in the V-1 clones was driven by a CREB-independent mechanism, because neither CREB phosphorylation (Fig. 5) nor GAL4 DBD -CREB TAD activity ( Fig. 6) was enhanced.
In contrast to CREB, we showed that ATF-2 was highly phosphorylated on Thr 69 and Thr 71 in the V-1 clones (Fig. 5) and that GAL4 DBD -ATF-2 TAD activity was greatly enhanced (Fig. 6). There was no significant difference between the V-1 and control cells in expression level of ATF-2 determined by immunoblot analysis (Fig. 5) and in binding of ATF-2 to the TH-CRE determined by the gel mobility shift assay (data not shown). Because the GAL4 DBD -ATF-2 TAD protein containing the phosphorylation sites had a dominant-negative effect for the enhanced activity of the TH promoter in the V-1 cells (Fig.  7), our data collectively show that phosphorylation of ATF-2 induced the TH gene transcription in the V-1 clones.
The stress-activated kinases (SAPKs), such as Jun aminoterminal kinase (JNK) (36) and p38/HOG (38), both of which phosphorylate ATF-2, have been identified as stimulators of ATF-2 (35,36,39). However the phosphorylation state or expression of these kinases in the V-1 clones was little changed compared with that of the control or parental PC12D cells, as determined by Western blotting analysis with antibodies against phosphorylated JNK or p38/HOG (data not shown). Although we have not yet identified the kinase that phosphorylates ATF-2 in the V-1 clones, a small up-regulation of these kinases may contribute to the long-lasting ATF-2 activation in the V-1 clones. Alternatively, there is a possibility that the FIG. 5. Phosphorylation of ATF-2 was increased in the V-1 clones. Western blotting was performed using phosphorylated formspecific or nonspecific antibody against CREB or ATF-2. A, whole-cell extracts were prepared from PC12D cells incubated without or with 10 M FSK or 10 ng/ml NGF and from V-1 or control clones. Whole-cell extracts were separated by SDS-PAGE (10% gel) and then analyzed by immunoblotting with antibodies against CREB phosphorylated on Ser 133 and CREB (left and right of upper panels, respectively) or ATF-2 phosphorylated on Thr 71 and ATF-2 (left and right of lower panels, respectively). B, nuclear extracts were prepared from V-1 and control clones and separated by SDS-PAGE (8% gel) and then analyzed by immunoblotting with antibodies against ATF-2 dual phosphorylated on activity of phosphatases active toward ATF-2, which have not been identified, may be decreased in the V-1 clones. It is of importance that the mechanism of the increased phosphorylation of ATF-2 in the V-1 clones be explored.
Our results suggest the involvement of phosphorylated ATF-2 in the enhanced CRE-mediated transcription in the V-1 clones. Although there are some reports showing enhancement of CRE-mediated transcription by phosphorylation of ATF-2 on Thr 69 and Thr 71 (40,41), the mechanism is largely unknown. In the case of CREB, it is well known that phosphorylation of CREB on Ser 133 directly recruits CBP, a transcriptional coactivator (42)(43)(44), and results in the transcriptional activation of the target gene. The transactivation domain of ATF-2, however, was reported not to interact with CBP even after phosphorylation (45). No protein has been identified as one that specifically binds to phosphorylated ATF-2 to elevate CREmediated transcription. The V-1 cells could be a good material to explore the mechanism.
ATF-2 was reported to bind not only to CRE, but also to AP-1 binding motif (33). Using nuclear extracts of striatal neurons, Guo et al. (46) found that ATF-2 bound to a TH-AP-1 oligonucleotide dissociated from the TH-AP-1 after the cells had been stimulated to induce TH expression. Since Band 1 containing ATF-2 was reduced in the presence of a TH-AP-1 competitor, though to a less extent than a TH-CRE competitor (Fig. 6, lanes  9 -11), our data confirmed weak interaction of ATF-2 with the TH-AP-1. However, our data suggest that the TH-AP-1 does not contribute to the enhanced expression of the TH gene by the activated ATF-2 in the V-1 clones, because the activity of the AP-1-mediated transcription was unchanged in the V-1 clones (Fig. 2). NGF, a factor that induces differentiation of the sympathetic neurons in association with the expression of TH (47), preferentially activates the AP-1-mediated transcription in PC12 cells. It was reported that NGF-induced potentiation of the TH promoter activity was completely blocked by mutation of the TH-AP-1, but not by that of the TH-CRE (31), whereas enhanced activity of the TH promoter via the cAMP-PKA pathway was completely blocked by mutation of the TH-CRE (3, 31). The complete loss of the promoter activity in the TH-CRE mutant in the V-1 clones (Fig. 3) also suggests that AP-1mediated transcription is independent of the enhancement of the TH gene transcription in the V-1 clones.
Mouse null mutants of ATF-2 died shortly after birth and displayed symptoms of severe respiratory distress with lungs filled with meconium (48). In the ATF-2-deficient mice, an increased level of TH mRNA was shown in the embryonic brain, and the authors attributed it to hypoxia in the mice (48). Our data, however, may suggest another possibility that it reflects a direct influence of the disappearance of ATF-2 for the TH gene expression.
Although ATF-2 was reported to be required for postnatal neural development (49), activation of ATF-2 was also observed in apoptotic cells, i.e. activation of ATF-2 reduced the survival rate of differentiated PC12 cells (50). There may be some unknown mechanisms of ATF-2 to regulate both apoptosis and differentiation of neural cells. Because the TH gene expression was induced by ATF-2 activation in the V-1 clones, these cells would be a good model for investigating the function of ATF-2 during neural development.
GTP cyclohydrolase I (GCH) is the rate-limiting enzyme for the de novo synthesis of tetrahydrobiopterin, which is an important regulator of TH enzymatic activity and the protein level (51). In the V-1 clonal cells, we recently showed an increased tetrahydrobiopterin content and enhanced expression of GCH (52). In our previous study, we showed increased promoter activity of the GCH gene in the V-1 clones (52) as we did for that of the TH gene in this report. Recently, ϳ150 bp of the 5Ј-promoter region of the GCH gene was identified as the region contributing to basal and cyclic AMP-induced transcriptional activity and containing a non-canonical CRE (28,53), and we showed that this region was sufficient for the increased activity of the GCH promoter in the V-1 clones (52). Hirayama et al. (53) reported that the proximal promoter region could recruit ATF-2. These observations suggest that ATF-2 may coordinately regulate the expression of both TH and GCH genes.
It is interesting that the mRNA levels of AADC and DBH were also up-regulated in the V-1 clones (1), in addition to those of TH and GCH. Since the DBH gene has a noncanonical CRE in its 5Ј-flanking region to mediate cAMP-responsiveness (54); it is quite possible that the increased activity of the CREmediated transcription in the V-1 clones is responsible for the increased expression of the DBH gene. However, with the respect to the elevated expression of the AADC gene, the relationship between it and the CRE-mediated transcription has not yet been reported. Even though the expression of each enzyme might be governed by different transcription factors, any concerted regulatory mechanism(s) would be expected to play key roles during neural development. The present data suggest that ATF-2 may be involved in coordinate expression of the catecholamine-synthesizing enzymes for the development of catecholaminergic neurons.
In contrast to our observation of the enhanced expression of the catecholamine-synthesizing enzymes in the stable transformants of PC12D cells overexpressing V-1, transient transfection of PC12D cells with V-1 plasmids did not enhance the activities of the CRE-and TH promoter-reporting genes (data not shown). Because an increased expression of TH was observed in transgenic mice overexpressing V-1, 2 our data suggest that the enhanced expression of the TH gene may be a consequence of long lasting expression of V-1. Although we have not fully clarified the action of V-1 in the cells, we showed that ATF-2 in the V-1 cells were activated and that activation of ATF-2 enhances TH gene transcription. We are now investigating the general role of ATF-2 in transcription of the TH gene as well as that of other catecholamine-synthesizing genes.