INTRODUCTION

The CREB/CREM/ATF family of transcription factors (3, 4, 5, 6, 7, 8, 9, 10, 11) includes a number of similar factors with a basic DNA binding domain that recognizes a highly conserved sequence known as the cyclic AMP response element (CRE),¹ located in the upstream promoter of many cAMP-responsive genes (12, 13, 14, 15, 16, 17, 18). The proteins in this family share other characteristics: a leucine zipper domain for protein dimerization, a phosphorylation site for modulation of functional activity, and a transactivation domain that probably interacts with RNA polymerase II (19).

The cyclic nucleotide response element-binding protein (CREB), identified in studies of somatostatin gene regulation (20), was the first CRE-binding factor to be characterized. Phosphorylation of CREB at serine 133 by phosphokinase A (PKA) or calmodulin kinases, following increases in intracellular cAMP or Ca²⁺ levels, respectively, activates transcription (21, 22). CREB phosphorylation has been shown to be involved in many physiological and pharmacological conditions (23, 24, 25, 26, 27). A search for homologues of CREB led to the discovery of the gene encoding the cyclic nucleotide response modulator (CREM) (9). Through alternative promoter usage, differential splicing and alternative translation, the CREM gene encodes a set of proteins with different DNA binding affinities that either activate or repress transcriptional activity (9, 28, 29, 30). In neuroendocrine cells, alternate transcriptional initiation from an intronic promoter, located between the second glutamine-rich domain and exon  of the CREM gene, produces mRNAs encoding a family of small repressor proteins known as the inducible cAMP early repressor (ICER) (29). Although these proteins were previously localized to PC12 cells and adrenal gland (31), a potential role of ICER in regulation of catecholamine-related gene expression had not been investigated.

Tyrosine hydroxylase (TH) catalyzes the first and rate-limiting step in catecholamine biosynthesis (32). In adult mammals, TH expression is confined to a small number of neuronal groups in the central nervous system and to the neurons of the sympathetic nervous system and neuroendocrine chromaffin cells of the adrenal medulla in the periphery. Levels and activity of the enzyme are finely and differentially regulated in these cells by a large variety of pharmacological and physiological stimuli (33, 34, 35, 36, 37, 38, 39). Modulation of TH activity occurs at the protein level via phosphorylation of the enzyme catalyzed by PKA and possibly other kinases (40), and at the gene expression level via transcriptional changes (14, 41, 42, 43, 44). Transgenic models were developed to identify the minimal promotor regions necessary to drive tissue specificity and stimulus-associated induction in vivo (45, 46, 47). In vitro experiments with reporter gene constructs show that very proximal cis-acting elements play important roles in functional regulation of TH gene transcription (1, 2, 48, 49). Although conflicting data exist, probably related to cell line differences, the CRE, located from 38 to 45 bp upstream of the initiation site, appears to be crucial for both basal and cAMP-induced transcription of the TH gene (1, 2).

To identify members of the CREB/CREM/ATF family that regulate TH transcription, gel mobility shift assays were performed using an oligonucleotide representing the TH CRE (identical in sequence to bp 52 through 32 of the rat TH promoter). Nuclear proteins were obtained from TH-expressing cell lines and rat tissues under control conditions and in response to pharmacological manipulations that increase TH mRNA. We found that ICER is strongly induced in the adrenal gland by reserpine and in PC12 cells by forskolin. Cotransfection experiments confirmed that ICER is indeed capable of modulating TH promoter activity in PC12 cells. These results suggest that, in addition to its previously recognized stimulatory effects on TH transcription, the cAMP signaling cascade can also repress TH expression via an ICER-mediated mechanism.

EXPERIMENTAL PROCEDURES

All procedures were approved by the Institutional Animal Care and Use Committee of the Cornell University Medical College. Male Sprague-Dawley rats (200-300 g), housed 2-3/cage with free access to food and water, under a 12-h light/12-h dark cycle, received subcutaneous injections of reserpine, 10 mg/kg in 20% ascorbic acid, or an equivalent volume of the vehicle, 2-3 h after lights-on. Four hours later, they were decapitated, and the adrenal glands and brains rapidly dissected, frozen in liquid N₂, and stored at 80 °C until extraction of nuclear proteins or RNA.

PC12 cells and CATH.a cells were grown in RPMI 1640 medium, pH 7.2, supplemented with 10% horse serum, 5% fetal calf serum, 50 units/ml penicillin, and 25 mg/ml streptomycin. SK-N-BE(2)C cells were maintained in Dulbecco's modified Eagle's medium, pH 7.5, 10% calf serum, and the above antibiotics. Cells were harvested from 30 min to 8 h after treatment with forskolin (10 µM) and isobutyl methylxanthine (0.5 mM), frozen in liquid nitrogen, and stored as above.

Micropurification of nuclear proteins from either animal tissues or cell line cultures was performed according to Roy et al. (50). Briefly, 4 ml of NE1 (250 mM sucrose, 15 mM Tris-HCl, pH 7.9, 140 mM NaCl, 2 mM EDTA, 0.5 mM EGTA, 0.15 mM spermine, 0.5 mM spermidine, 1 mM dithiothreitol, 0.4 mM phenylmethylsulfonyl fluoride, 25 mM KCl, and 2 mM MgCl₂) were added to tissues (0.2-1.5 mg) or cells (5-10 × 10⁸). Tissues were first homogenized in a Dounce tissue grinder and filtered through cheesecloth. To free nuclei, tissue was further homogenized with added Nonidet P-40® (0.5%), washed twice, and pelleted at 3000 × g. One ``packed cell volume'' of NE2 (NE1 buffer containing 350 mM KCl) was used to resuspend nuclei and, after a 5-min incubation on ice, 20 strokes in the Dounce tissue grinder were performed. The homogenate was centrifuged for 90 min at 4 °C (180,000 × g), and the supernatant was dialyzed for 45 min against 50 mM KCl, 4 mM MgCl₂, 20 mM K₃PO₄ (pH 7.4), 1 mM -mercaptoethanol, and 20% glycerol. Extract protein concentrations were determined with a colorimetric assay (Bio-Rad).

In a 20-µl volume, nuclear proteins (4-15 µg) were incubated at room temperature for 20 min in binding buffer (50 mM Tris, pH 7.5, 500 mM NaCl, 5 mM dithiothreitol, 5 mM EDTA, 20% glycerol) in the presence of poly(dI-dC) (1 µg) with specific double-strand oligonucleotides end labeled with ³²P. The samples were loaded on a 6% acrylamide gel and run at 100 V in a low ionic strength buffer (0.25 × TBE) for 2 h. The gels were dried and autoradiographed at room temperature.

The following oligonucleotide sequences were used. For supershift experiments, the extracts were preincubated for 20 min with appropriate antisera on ice and then with labeled oligonucleotides as above. CREB, CREM, p-CREB, ATF1, and ATF2 were purchased from Upstate Biotechnology Inc. (Lake Placid, NY).

Separation of nuclear proteins (50-150 µg) was carried out by SDS-polyacrylamide gel electrophoresis (8.5%) for 1 h at 20 watts in Tris-glycine buffer (25 mM Tris, 250 mM glycine, and 0.1% SDS) at 22 °C. The proteins were electrotransferred on a supported nitrocellulose membrane (0.45 µm) as described previously (51). The membranes were soaked in phosphate-buffered saline solution and 5% nonfat dry milk for 2 h with gentle agitation. For immunological detection, the membrane was incubated with the primary antibodies, CREB (diluted 1:10,000) and anti-mouse CREM (diluted 1:1,000), for 2 h followed by three washes (10 min each in 5% nonfat dry milk, 0.02% Tween 20®, phosphate-buffered saline) under gentle agitation. Peroxidase-conjugated goat anti-rabbit immunoglobulins (DAKO, Carpinteria, CA) were used as secondary antibodies at a 1:2000 dilution (2 h incubation followed by three wash steps).

The mRNA was extracted from either animal tissues or cells with the Poly(A)Ttract® System1000 (Promega, Madison, WI). mRNA (1-5 µg) were electrophoresed in a 1% agarose gel, transferred to a nylon membrane (Amersham), and hybridized with specific probes labeled by random priming with [-³²P]dCTP. Blots were exposed to Kodak XAR film for 2 days (51).

mRNA was reverse-transcribed with Moloney murine leukemia virus reverse transcriptase in the presence of oligo(dT) and [-³²P]dCTP as radioactive tracer. The cDNA obtained was amplified with specific oligonucleotides designed to prime some of the different CREM isoforms. Some of the fragments amplified by PCR reaction were purified from agarose gels, labeled, and used as probes in the Northern blot analysis. BamHI and EcoRI restriction sites were introduced in the 5 and 3 primers, respectively, to facilitate the following subcloning. The conditions used for amplification are the same as described by Hoeffler et al. (52). Primers are as follows: (B5), ACTTTAGGATCCACTGTGGTACGGCCAAC; (Ia3), GAGCTCGAATTCCCAATTCACACTCTACAGCAG; (Ib3), GTTAAATAGAATTCACTAATCTGTTTTGGG; (A5), GAAGGATCCGGAAGCTGCCCGGGAGTG.

Transfections performed into PC12 cells with the reporter TH2400CAT and the expression PKAc plasmids and the following CAT assay have been already described (1). The expression vector CMV-ICER-Ib was obtained by subcloning the PCR-amplified ICER-Ib cDNA in BamHI and EcoRI sites of the polylinker region of the pcDNAI/AMP plasmid (Invitrogen, San Diego, CA).

RESULTS

As an initial step in defining the role of CRE-binding proteins in TH gene regulation, gel retardation analysis was performed on TH-expressing cell lines and tissues obtained from both peripheral and central nervous system of rats. Gel mobility shift assays, using an oligonucleotide representing the consensus CRE located at 38 to 45 in the rat TH promoter plus some additional flanking sequence, produced distinct retardation patterns that varied among the different extracts (Fig. 1). Nuclear extracts from the adrenal gland and PC12 cells exhibited almost identical binding patterns, which differed from those observed in the extracts from central nervous system or other catecholamine-producing cell lines. Extracts of human neuroblastoma-derived SK-N-BE(2)C cells (53) and rat locus ceruleus-derived CATH.a cells (54) showed unique binding patterns.

Previous studies demonstrated an increase in rat adrenal TH mRNA within 4 h of administration of reserpine (55) and within 4 h of forskolin treatment of PC12 cells (56). We therefore examined gel retardation patterns at 4 h. Nuclear extracts from both forskolin-treated PC12 cells and adrenal glands from reserpine-treated rats displayed dramatic increases in a fast moving protein-DNA complex (Fig. 1). In contrast, neither treatment substantially altered binding patterns in the CATH.a or locus ceruleus extracts (Fig. 1).

To examine the binding specificity of PC12 and adrenal gland extracts, the gel retardation assays were performed in the presence of molar excess of either TH-CRE or TH-AP1 unlabeled nucleotides (Fig. 2, A and B). In extracts from both sources, excess unlabeled CRE prevented the formation of protein-DNA complexes (Fig. 2, A and B). In contrast, even 100-fold molar excesses of the AP-1 oligonucleotide did not efficiently interfere with the binding between the CRE probe and nuclear proteins derived from either PC12 cells (Fig. 2A) or the adrenal gland (data not shown).

Using antibodies against different members of the CREB/CREM/ATF family, supershift assays were performed to make an initial identification of the proteins in the forskolin-induced fast running complex from PC12 cells. The anti-CREM antibody, even at the lowest concentration, supershifted the band in both control (Fig. 3B) and forskolin-treated (Fig. 3, A and B) PC12 preparations. Anti-CREM antibody also supershifted the fast running complex from adrenal gland (Fig. 3B).

Western blot analysis further characterized the antigen recognized by the CREM antibody. As shown in Fig. 4A, six distinct bands, representing antigens with molecular masses ranging from 12 to 43 kDa, interacted with the CREM antibody in extracts from both PC12 cells and adrenal gland. Two small proteins of similar molecular mass, 12 and 13 kDa, are strongly induced by forskolin and reserpine (Fig. 4A). The identification of these proteins as ICER is indicated by the fact that no other CRE-binding polypeptides of the appropriate lengths have been reported and is further supported by the Northern and PCR analysis presented below. As the CREM antiserum used cross-reacts with other members of the CREB/CREM/ATF family, it is not possible to state that all the detected polypeptides derive from the CREM gene.

CREB antiserum also produced a supershift in the protein-DNA binding pattern but did not affect the fast running complex (Fig. 3A). This finding confirms the already reported (1, 2) interaction between CREB and the TH-CRE and rules out the possibility that the protein(s) bound to DNA in the fast running complex are CREB-related. The CREB antibody in Western blot analysis (Fig. 4B) reacted with at least one protein whose molecular mass (43 kDa) matched that previously reported for CREB (21). Note also the increase in the amount of CREB protein following forskolin treatment (Fig. 4B). The inability of phospho-CREB to produce a supershift probably results from phosphatase activity during the nuclear protein extraction, since a supershift occurred with the CREB antibody. Additionally, CREB may already be dephosphorylated at the time analyzed, i.e. 4 h after application of the stimulus. While the ATF-2 antiserum did not affect any protein-DNA binding complex, the anti-ATF-1 antibody supershifted one of the slower moving complexes. This finding was not analyzed further in the present study.

Previous reports identified a small protein, ICER, that is generated by alternative intronic promoter usage from the CREM gene and expressed in PC12 cells and adrenal gland (29, 31). Fig. 5A shows the intron-exon structure of the CREM gene (31). Four possible transcripts can be generated. All ICER transcripts contain 82 intronically derived bases not found in other CREM mRNAs. The four ICER transcripts are generated by either inclusion or exclusion of exon  and alternative splicing of the DNA binding domains Ia and Ib (31). To confirm the presence of the ICER isoforms, we designed specific primers (Fig. 5A) that amplify from the unique 82-base site (B5) to either the Ia site (Ia3) or the Ib site (Ib3). mRNA isolated from forskolin-treated PC12 cells and vehicle- and reserpine-treated adrenal glands were reverse transcribed and subjected to PCR. Two bands, approximately 300 bp long, were amplified from the B5 and Ib3 primers, corresponding to ICER-Ib and ICER--Ib (Fig. 5C). No bands were amplified with the Ia3 primer, regardless of which 5 primer was used (Fig. 5B).

Northern blot analysis confirmed that the pharmacological treatments altered expression of ICER. A 150-bp probe recognizing the Ib domain labeled two bands of approximately 2.4 and 1.5 kb on Northern blots of mRNA obtained from untreated PC12 cells and adrenal glands from control rats (Fig. 6). The Ib domain of the CREM gene is included in all the CREM-derived transcripts that contain a DNA binding domain. The detection of only two isoforms of ICER confirms the PCR results that no full-length CREM transcripts are present either under control conditions or are inducible in PC-12 cells and adrenal glands. These data also support the possibility that 43-kDa, as well as the other, polypeptides detected by the CREM antibody can be CREB isoforms and/or ATF-related proteins. Forskolin and reserpine treatments clearly induced both mRNA bands in PC12 cells and adrenal glands, respectively. In PC12 cells, high levels of mRNA could be detected as soon as 30 min (data not shown) and even 8 h after addition of forskolin. A -actin probe was used to normalize the amount of mRNA loaded on the gel (Fig. 6).

To test directly whether ICER can regulate TH transcription, the reporter plasmid TH2400CAT (1) and the expression plasmid pCMV-ICER-Ib were cotransfected into PC12 cells and CAT activities measured. Cotransfection of the ICER expression construct substantially diminished CAT activity (Fig. 7). ICER-mediated repression of PKA-stimulated transcription of the TH construct was even more dramatic than repression of basal transcription. In control experiments, cotransfection of pUC19 or pcDNAI/AMP (data not shown) did not alter transcription of the TH reporter construct.

DISCUSSION

The present study investigated the role of CRE-binding proteins in TH gene regulation. Nuclear extracts from rat adrenal gland and from the adrenal medulla-derived PC12 cell line (57) produced virtually identical binding patterns in gel mobility shift assays employing a TH-CRE oligonucleotide. The binding patterns of extracts from locus ceruleus differed from patterns produced by adrenal gland, PC12 cells, the catecholamine-producing human neuroblastoma line SK-N-BE(2)C cells (53), and murine locus ceruleus-derived CATH.a cells (54). Thus, relative levels of different CRE-binding transcription factors show a high degree of variability among catecholamine-producing cell types. This observation suggests that regulation of CRE-containing catecholamine genes such as TH and dopamine--hydroxylase (58) occurs by cell-type specific mechanisms that depend, in part, on differential expression of CRE-binding transcription factors. Evaluation of this hypothesis will require more extensive study.

Reserpine and forskolin induce rapid increases in TH levels in adrenal gland and PC12 cells, respectively (33, 37, 43, 56, 59, 60). Abundant data support a direct role for cAMP, PKA, and the TH-CRE as primary mediators of forskolin-induced increases in TH expression (61, 62, 63, 64, 65). In contrast, the molecular events triggering reserpine-activated TH induction are not yet fully understood. Reserpine, which inhibits monoamine uptake in storage vesicles, depletes catecholamine stores. The drop in catecholamine levels triggers a trans-synaptically mediated induction of TH transcription and protein synthesis (34, 37, 59). Although c-Fos mRNA increases after the drug treatment (55, 66) suggesting a role for the TH AP-1 site, additional regulatory mechanisms are likely involved in TH promoter transcription regulation. In fact, increased cAMP levels and induced PKA activity have been reported in rat adrenal chromaffin cells after reserpine treatment (67). Recent reports, underlining the importance of the CRE in TH gene regulation (1, 2), prompted investigation of the possible involvement of CRE-binding proteins in both the forskolin- and reserpine-induced TH gene transcriptional activation.

In gel mobility shift assays, we showed that a fast running CRE-protein complex increased substantially in nuclear extracts from adrenal glands of reserpine-treated rats and forskolin-treated PC12 cells. Supershift experiments indicated that the fast running complex contained one or more members of the CREM family of proteins. Although CREB antiserum produced a supershift of a slow moving complex, indicating that CREB family members are present in the nuclear extract, they do not contribute to the binding activity in the fast running complex. Collectively, Western blot, PCR, and Northern blot analyses demonstrated that both reserpine and forskolin treatments induced two isoforms of the CREM-related ICER protein (ICER-Ib and -ICER-Ib).

Our results confirm the previously reported induction of ICER by forskolin in PC12 cells (31). However, under the present experimental conditions, ICER mRNA levels peaked approximately 4 h after forskolin treatment and remained detectable at 8 h. In the earlier study, maximal levels of ICER mRNA occurred after 2 h (31). These differences probably reflect either variations between the PC12 lines or differing culture conditions. Although the presence of ICER in adrenal gland has been reported (29), to our knowledge this is the first demonstration of robust induction of ICER in adrenal gland by reserpine.

To determine if ICER induction is involved with TH gene transcriptional regulation, cotransfection analyses were performed. The current data clearly demonstrate that ICER represses TH promoter-driven transcription in PC12 cells, with the strongest repression occurring after transcriptional stimulation by PKA. Since PC12 cells are immortalized adrenal medullary cells, the in vitro experiments suggest a role for ICER in TH transcriptional modulation of adrenal medullary cells. The strong ICER induction produced in the adrenal gland after reserpine treatment supported this hypothesis. Further evidence for a role of ICER in TH regulation in vivo is suggested by the dynamics of the TH mRNA response to reserpine. A relatively rapid return of TH mRNA to control levels occurs in adrenal medullary cells, as compared to the response in the locus ceruleus, following reserpine-induced up-regulation of TH mRNA levels (59, 68). Studies, now in progress, will clarify the mechanisms underlying the role of ICER in TH gene modulation in the adrenal gland.

The results of this study suggest that the transcriptional regulation of TH by cAMP-related mechanisms in PC12 cells involves both activating and repressing trans-acting proteins. Similar transcriptional regulation of TH could be simulated in cotransfection experiments performed with SK-N-BE(2)C cells, which do not normally express ICER.² The data indicate that regulation of TH mRNA synthesis occurs in a tissue-specific manner that may be partially determined by a complex balance of activators and repressors. Control could occur by transcription factors competing for the same cis-acting sites. In this scheme CREB and ICER would provide a system for fine modulation of TH gene expression. In fact, CREB, which binds to the TH CRE, is rapidly activated by forskolin (21). Interactions between repressors acting at one site (e.g. ICER binding to the CRE), and activators at other sites (e.g. a Fos-Jun dimer at the AP-1 site) are additional possibilities not addressed by the present work, but represent important directions for future experiments.