Molecular Brain Research 74 Ž1999. 208–216 www.elsevier.comrlocaterbres
Research report
AP-2b represses D 1A dopamine receptor gene transcription in Neuro2a cells Sousuke Takeuchi a , Ichiro Imafuku a , Masaaki Waragai a , Christina Roth b, Ichiro Kanazawa a , Reinhard Buettner b, M. Maral Mouradian c , Hitoshi Okazawa
a,)
a
c
Department of Neurology, Graduate School of Medicine, UniÕersity of Tokyo, 7-3-1, Hongo, Bunkyo-ku, Tokyo 113, Japan b Institute for Pathology, UniÕersity of Regensburg Medical School, D-93042 Regensburg, Germany Genetic Pharmacology Unit, Experimental Therapeutics Branch, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, MD 20892, USA Accepted 5 October 1999
Abstract Expression of the D1A dopamine receptor in brain is restricted to specific neuronal populations. To investigate the mechanism of this selective expression, we localized a silencer upstream of the human D1A gene and identified its binding transcription factor in the D 1A-negative neural cell line Neuro2a. Using deletion CAT analysis, we narrowed this silencer to the region between nucleotides y561 and y532 relative to the CAP site. This 30-bp region, designated D1AS1, contains a sequence homologous to the AP-2 binding site and binds to a factor that also interacts with the AP-2 consensus sequence. In gel supershift assays, this factor is recognized by anti-AP-2b antibody. Co-transfection of Neuro2a cells with an AP-2b expression vector repressed the basal CAT activity of D1A promoter–reporter plasmids in a D1AS1-dependent manner. RT-PCR analysis indicated that, among AP-2 family members, Neuro2a cells express only AP-2b. Furthermore, co-transfection of these cells with decoy oligonucleotides corresponding to the D1AS1 sequence de-repressed the D 1A gene promoter. Unlike in Neuro2a cells, AP-2b could not repress the D1A promoter in the D1A-positive neural cell line, NS20Y. In addition, the expression of AP-2b in different brain regions does not inversely correlate with that of D1A dopamine receptor. These observations taken together indicate that AP-2b is a repressive transcription factor that acts on the D1AS1 silencer of the D1A dopamine receptor gene via some cell-specific mechanismŽs. in Neuro2a. q 1999 Elsevier Science B.V. All rights reserved. Keywords: AP-2b; D1A dopamine receptor; Gene regulation; Silencer; Repressor; Differentiation
1. Introduction The central nervous system includes various neuronal populations each with different morphological and functional features. This diversity is due to different subsets of expressed genes which are regulated mainly at the level of transcription. Thus, analyses of cis-regulatory elements, transcription factors, and transcriptional co-factors affecting expression of neural specific genes are essential for AbbreÕiations: PCR, polymerase chain reaction; CAT, chroramphenicol acetyl transferase; REST, the RE1-silencing transcription factor; NRSF, the neuron specific silencer factor; HEPES, N-w2-Hydroxyethlx X piperazine-N -w2-ethanesulfonic acidx; EDTA, ethylenediaminetetraacetic acid; DTT, dithiothreitol; KCl, potassium chloride; CMV, cytomegalovirus ) Corresponding author. Fax: q81-3-5800-6548; e-mail:
[email protected]
understanding the mechanisms underlying neuronal diversity. Repressive transcription factors are equally important as transactivators in the overall regulation of neural specific genes. However, only a small number of repressive transcription factors regulating neuron-specific genes have been reported thus far. These include RESTrNRSF which restricts type II voltage dependent sodium channel expression to neurons w4x and represses multiple neuron-specific genes in non-neuronal tissues and in undifferentiated neurons w20x. Considering the number of known positive transcription factors, far more negative transcription factors are likely to exist than is currently recognized to suppress maladaptive gene expression in specific types of neurons. The D 1A dopamine receptor is one of five known receptors that mediate central dopaminergic neurotransmission and therefore could be involved in the pharmacology
0169-328Xr99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 9 - 3 2 8 X Ž 9 9 . 0 0 2 9 8 - 3
S. Takeuchi et al.r Molecular Brain Research 74 (1999) 208–216
of Parkinson’s disease or even the psychoses that can be treated with dopamine receptor-blocking drugs. We previously reported that expression of the D 1A gene is suppressed by multiple cis-elements in Neuro2a cells w6,17x. Interestingly, the 5X flanking region of the D 1A gene does not include any sequences homologous to the RESTrNRSF consensus binding site, suggesting that other repressive transcription factors suppress the expression of this gene. In this study, we characterized a functional silencer which potently represses the activity of the D 1A dopamine receptor gene promoter in Neuro2a cells and found that AP-2b is the repressive transcription factor acting on this silencer.
2. Materials and methods 2.1. Plasmid construction Plasmids D1A-1510, D1A-672 and D1A-282 are identical to pCAT-HD1G-A, -B and -C vectors described previously w11x. Plasmids D1A-561, -532, -492 and -461 were constructed by serially deleting the 5X end of the insert in D1A-672 with exonuclease III using the kilo deletion kit ŽTakara.. Nucleotide sequences of these reporter plasmids
209
were confirmed by sequencing and their 5X ends are shown in Fig. 1a. 2.2. CAT assays Neuro2a cells were transfected using the calcium phosphate method as described previously w15,16x. Ten mg effector plasmid and 10-mg reporter plasmid were used to co-transfect 5 = 10 6 cells cultured in 10-cm tissue-grade dishes ŽCorning. with a-MEM medium ŽSigma. supplemented with 10% fetal bovine serum ŽDainippon Pharmaceuticals.. Transfection efficiency was verified by co-transfecting 1 mg pCH110 ŽPharmacia., a eukaryotic vector containing the simian virus early promoter and the Escherichia coli-b-galactosidase Ž LacZ . structural gene. Each experiment was repeated at least four times and inter-assay variation of transfection efficiency was less than 20%. NS20Y cells, generously provided by Dr. Nirenberg ŽNHLBI, Bethesda, MD., were cultured and transfected according to the same protocol. 2.3. Preparation of nuclear extract Neuro2a and NS20Y cells were cultured in a-MEM medium ŽSigma. with 10% fetal bovine serum as described
X
Fig. 1. Ža. Genomic structure of the human D1A dopamine receptor gene and 5 deletion CAT constructs used to localize the silencer. Nucleotide numbering system is relative to the main transcription initiation site. Exon 1 and exon 2 are indicated with black boxes. PRE is the POU-responsive element through which Brn-4 transactivates the D1A gene w3x. Žb. Deletion CAT analysis of the silencer repressing the D1A gene. Transactivation by Brn-4 was observed with reporter plasmids shorter than D1A-532, suggesting that the silencer is located around the sequence y561 to y532. Ten micrograms of reporter plasmid, 10 mg effector plasmid ŽpCMVBrn4., and 1 mg pCH110 were used to co-transfect Neuro2a cells. In lanes designated as pCMVBrn4 Žy., 10 mg pBS-SK was used as effector plasmid to equalize total amount of DNA. Transfection efficiency was normalized to b-galactosidase activity. Fold transactivation was calculated relative to basal CAT activity of each reporter plasmid and the mean of four independent transfections is shown.
210
S. Takeuchi et al.r Molecular Brain Research 74 (1999) 208–216
previously w15,16x. Briefly, 4 = 10 8 cells were harvested by scraping, collected in PBS, then centrifuged for 5 min at 100 = g. The pellet was suspended in 8 = vol. of lysis buffer w20 mM HEPES ŽpH 7.9., 1 mM EDTA ŽpH 8.5., 1 mM DTT, 10% glycerol, 0.5 mM spermidine, 1 mM PMSF, 1 mgrml leupeptin, 1 mgrml pepstatin A, 0.3 mgrml antipain x, to which NP-40 was added at a final concentration of 0.3%. After 5 min incubation on ice, the suspension was centrifuged at 700 = g, 48C, for 10 min. The pellet was resuspended in 1 = vol. of lysis buffer with 1 = vol. of 2 M KCl. The suspension was centrifuged at 100 = k g, 48C, for 30 min. The resultant supernatant was loaded on G-50 Sephadex column ŽPharmacia. and eluted with lysis buffer plus 0.05 M KCl. The fraction with the
Fig. 2 Žcontinued..
highest protein concentration was used for gel mobility shift assays. Mouse brain nuclear extract was prepared from 1 g tissue by homogenizing in 5 ml PBS with glass homogenizer, followed by centrifugation at 100 = g for 5 min. Nuclear extract was prepared from the pellet according to the protocol described above for cultured cells. 2.4. Gel mobility shift assays
Fig. 2. DNA–protein interaction in D1AS1 occurs at the AP2 consensus sequence. Ža. Nucleotide sequence of the human D1A dopamine receptor gene silencer ŽD1AS1.. Consensus sequences for known transcription factors and probes ŽA,B. used for gel mobility shift analysis are indicated. Žb. Gel mobility shift assay with Neuro2a cell nuclear extract. The upper band Žlane 2, arrow. is specifically inhibited by 10=cold oligonucleotide Žlane 3.. A non-specific band is indicated with white arrow head. The band indicated with black arrow head was not observed consistently Žsee Žc., lanes 4 and 6., suggesting that perhaps it represents a protein degradation product. Žc. Gel mobility shift assay showing that the binding factor to the D1AS1 region also interacts with AP-2 consensus sequence. AP-2 consensus sequence probe Žlane 2. and probe A Žlane 3. were shifted similarly with nuclear extract from Neuro2a cells. Specific bands shifted with probe A Žlane 4. were inhibited by the AP2 consensus sequence Žlane 5. but not by the POU consensus sequence Žlane 6..
These were performed as described previously w15,16x. Briefly, 4 ml of nuclear extract, 4 ml of 5 = gel shift buffer w100 mM HEPES ŽpH 7.9., 1 mM EDTA, 50% glycerol, 0.5 mM spermidine, 5 mM DTTx and 20.000 c.p.m. probe were incubated in a final volume of 20 ml at room temperature for 15 min. For super shift assays, 2 ml of anti-AP-2a or anti-AP-2b antibody ŽSanta Cruz Biotechnolgy. was added and incubated for another 15 min. Reaction samples were electrophoresed in 4% polyacrylamide gel containing 10% glycerol. Probe sequences after filling-in cohesive ends with Klenow enzyme ŽTakara, Japan. were: A Ž5X-AAGGAGTGAGTCCCCTCGGGGCGCAGC-3X . and B Ž5X -TCGGGGCGCAGCGGCGCGCG-3X .. AP-2 consensus sequence was 5X-AGGAACTGACCGCCCGCGGCCCGTGTGCAGAG-3X w8x. These oligonucleotides were annealed to complementary oligonucleotides lacking five bases from the 3X-ends and radiolabelled by Klenow enzyme. GST-AP-2b binding domain fusion protein expression vector was described previously w13x. 2.5. Decoy oligonucleotide transfection Pre-annealed phosphorothioate-oligonucleotides derived from the D1AS1 sequence was used as decoy to compete for transcription factor binding with D1AS1. The sequence of the decoy oligonucleotide was 5X-TGAGTCCCCTCGG-
S. Takeuchi et al.r Molecular Brain Research 74 (1999) 208–216
GGCGCAG-3X and of the scrambled oligonucleotide was 5X-GGAGTCGCTCCTCGCGCAGG-3X . Ten microgram CAT reporter plasmid and 3 mg decoy or scrambled oligonucleotide were used to cotransfect 5 = 10 6 Neuro2a cells cultured in 10-cm tissue-grade dishes ŽCorning. with a-MEM medium ŽSigma. supplemented with 10% fetal bovine serum ŽDainippon Pharmaceuticals.. Final concentration of the oligonucleotide was 300 ngrml. The molecular ratio between plasmid and oligonucleotide was 1:400, which was necessary to detect the decoy effect efficiently. Transient transfection was performed with Superfect Transfection Reagent ŽQiagen., activated dendrimer w21x, according to the supplier’s protocol.
211
silencer 1 ŽD1AS1.. D1AS1 does not contain any sequence homologous to the previously known neuron-specific silencer RESTrNRSF w4,20x. 3.2. D1AS1 contains an AP-2 consensus-like sequence To investigate whether the D1AS1 contains consensus sequences for previously reported transcription factors, we
2.6. RT-PCR This was performed as described previously w18x with 0.1 mg mRNA from Neuro2A and HeLa cell lines and the following specific primers for AP-2 family members. For AP-2a , AF: 5X-GAGACGTAAAGCTGCCAA-3X ŽGenebank accession number: X74216, nucleotides 907 to 924. and R: 5X-AGATAGTTCTGCAGGGC-3X ŽX74216 nucleotides 1275 to 1259.. For AP-2b, BF: 5X-GAGAGAAAGGCTAGAAAA-3X ŽX78197, nucleotides 1032 to 1049. and R Žnucleotides 1442 to 1426.. For AP-2g, GAGGTGTTCTCAGAAGA-3X ŽX95693, nucleotides 1290 to 1277. and R Žnucleotides 1290 to 1277.. The same reverse primer was used for all three AP2 family members because it corresponds to a sequence conserved among all three.
3. Results 3.1. Characterization of the potent silencer of the D1 A dopamine receptor gene in Neuro2a cells We had demonstrated previously that the POU transcription factor Brn-4 up-regulates transcription of the D1A dopamine receptor gene by interacting with its first intron in Neuro2a cells. We also found that this Brn-4-induced transactivation is suppressed by an upstream region in the D 1A gene w6,17x. In addition, this suppression is more pronounced with longer flanking up-stream regions suggesting that multiple cis-elements participate in repressing the D 1A gene in Neuro2a cells. To investigate these silencers, we performed co-transfection CAT assays using reporter plasmids having various lengths of the D 1A upstream sequence ŽFig. 1a., all including the POU response elements w17x. We observed a dramatic difference in the transactivation by Brn-4 between reporter plasmids D1A561 and D1A-532 suggesting that a strong silencer element which represses Brn-4-induced D 1A gene expression is located between nucleotides y561 and y532 relative to the cap site ŽFig. 1b.. We designated this region as D 1A
Fig. 3. Identification of the AP2 family member that binds to D1AS1. Ža. RT-PCR analysis of the expression of AP-2 family members in Neuro2a cells. mRNA prepared from HeLa cells was used as positive control. Žb. Gel supershift analysis with probe A and Neuro2a cell nuclear extract in the presence of anti-AP-2b and anti-AP-2a antibodies. The specific band is indicated with a small black arrow and its supershift is indicated with a large black arrow. Non-specific bands are indicated with black and white arrow heads. Žc. Gel mobility shift assay with GST-fusion protein flanked with the AP-2b binding domain showing that AP-2b interacts with this probe. The binding of GST-AP-2b fusion protein to the AP-2 consensus sequence probe is shown as positive control Žlane 3, arrow head.. GST-AP-2b similarly binds to probe A Žlane 6.. In the lanes loaded with GST protein Žlanes 2 and 5., very weak non-specific bands could be seen at positions lower than the specific band.
212
S. Takeuchi et al.r Molecular Brain Research 74 (1999) 208–216
searched D1AS1 with a computer-based homology search for transcription factor binding sites ŽTFSEARCH: http: rr pdap1.trc.rwcp.or.jpr research rdbr TFSEARCH. html.. We found homologous sequences to Adf-1 w5x, AP-4 w10x, and HEN w1–3x elements at the center of D1AS1. This homology search also indicated that an AP-2-consensus-like sequence is located within D1AS1 ŽFig. 2a.. Therefore, we synthesized two oligonucleotide probes and analyzed their binding to Neuro2a nuclear proteins in gel mobility shift assays ŽFig. 2a.. Probe A covers the sequence y561 to y535 including the AP-2-consensus-like sequence but destroys the possible binding sites for Adf-1, AP-4 or HEN. Meanwhile, probe B contains Adf-1, AP-4, and HEN sequences but not the full AP-2-consensus-like sequence ŽFig. 2a.. In gel mobility shift assay with probe A, the upper band ŽFig. 2b, lane 2. was considered specific because it disappeared by excess cold oligonucleotide A ŽFig. 2b, lane 3. but not by the POU consensus sequence ŽFig. 2c, lane 6.. On the other hand, gel shift assay with probe B did not show any specific band ŽFig. 2b, lane 5.. In addition, probe A and the AP-2 consensus sequence probe w8x yielded a similar band shift profile with Neuro2a nuclear extract ŽFig. 2c, lane 2 vs. lane 3., and the band shifted by probe A was inhibited by cold consensus AP-2 competitor ŽFig. 2c, lane 5.. Furthermore, we could not observe the specific band with a mutated AP-2 consensus sequence probe whose AP-2 core sequence palindrome w8x was destroyed Ždata not shown.. These results indicated that the factor
binding to the D1AS1 silencer is not Adf-1, AP-4 or HEN-like factor but rather it interacts with its AP-2 consensus sequence. 3.3. AP-2 b binds to the D1AS1 silencer We next tested whether the factor binding to the AP2 consensus sequence in D1AS1 is indeed a member of the AP-2 transcription factor family. To date, three members of this family, a , b, and g have been molecularly cloned w13,22,23x. RT-PCR analysis showed that only AP-2b is expressed in Neuro2a cells, while the same set of primers detected all three members in HeLa cells ŽFig. 3a.. To confirm that AP-2b is a binding factor to probe A, we performed gel supershift assay using antibodies against AP-2a and AP-2b and found that the protein binding to D1AS1 is supershifted by anti-AP-2b antibody but not by anti-AP-2a antibody ŽFig. 3b.. We also tested whether the DNA binding domain of AP-2b binds to probe A ŽFig. 3c.. GST-fusion protein containing the DNA binding domain of AP-2b could bind to probe A Žlane 6. as strongly as to the AP-2 consensus sequence probe Žlane 3.. GST protein itself showed very weak non-specific bands Žlanes 2 and 5., whereas these bands were not retarded similarly to the specific band of GST-AP-2b. These findings indicated that AP-2b binds to the D1AS1 silencer. These data, however, could not totally exclude the possibility that another factor of a similar gel mobility in
Fig. 4. Ža. Construction of D1A-561 and D1A-532 reporter plasmids. D1A-561 but not D1A-532 contains the D1AS1 silencer. Žb. Transcription assay showing that AP-2b inhibits CAT activity via the D1AS1 silencer. Ten micrograms of reporter and effector plasmids were co-transfected into 5 = 10 6 Neuro2a cells. pCMVAP2b or pBS-SK was used as effector. Fold suppression was calculated relative to the CAT activity of cells transfected with pBS-SK. Mean values of four independent transfections are shown. Transfection efficiency was normalized to b-galactosidase activity. Expression of AP-2b suppresses CAT activity either in the absence Žlane 1, 2. or presence Žlane 5, 6. of Brn-4. However, this suppression was not seen when the D1AS1 silencer was deleted from the reporter plasmid Žlanes 3, 4 and 7, 8.. For lanes 1 to 4, fivefold excess amount of nuclear extract was used for CAT assays to observe suppression of the basal activity.
S. Takeuchi et al.r Molecular Brain Research 74 (1999) 208–216
213
that AP-2b suppresses the basal activity of plasmid D1A561 which has D1AS1 but not that of D1A-532 lacking D1AS1 ŽFig. 4b, lanes 1 to 4.. We also confirmed that AP-2b suppresses transcription in the presence of Brn-4 in a D1AS1-dependent manner ŽFig. 4b, lanes 5 to 8., consistent with the result in Fig. 1. These findings indicated that the suppressive effect of AP-2b is mediated by the D1AS1 silencer and supported the idea that AP-2b negatively regulates transcription of the D 1A dopamine receptor gene via D1AS1. 3.5. D1AS1 decoy oligonucleotide up-regulates D1 A gene promoter actiÕity. To further address the importance of D1AS1 in D 1A gene regulation in cells, we tested whether a D1AS1 decoy oligonucleotide up-regulates transcription of this gene. For transfecting oligonucleotides in this assay, Superfect Transfection Reagent ŽQiagen. was employed because of recent reports suggesting that this method is useful for introducing oligonucleotides into cells without injuring
Fig. 5. Ža. CAT assay with cotransfected decoy or scrambled oligonucleotides and reporter plasmids. Žb. Bar graph showing that decoy oligonucleotide co-transfection increases the basal CAT activity of reporter plasmids containing D1AS1 in Neuro2a cells. Percent transactivation was calculated relative to the CAT activity of cells co-transfected with the scrambled oligonucleotide and reporter plasmid A as 100%. Mean" S.E.M. values of four independent transfections are shown. Transfection efficiency was normalized to b-galactosidase activity.
Neuro2a cells also binds to probe A, since the shifted band could not be completely supershifted by anti-AP-2b antibody ŽFig. 3b, lane 2.. 3.4. AP-2 b represses the D1 A dopamine receptor gene promoter To verify that AP-2b represses basal expression of the D 1A gene via the D1AS1 silencer, we performed co-transfection CAT assays using an AP-2b expression vector. Since the basal level of D 1A CAT reporter plasmids is very low in Neuro2a cells, a fivefold increase in the amount of cellular extract was used in these experiments compared with those demonstrating Brn-4-induced transactivation ŽFig. 1.. Using this modification in the assay, we found
Fig. 6. Ža. Gel mobility shift assay showing that no NS20Y cell nuclear protein binds to probe A or to AP-2 consensus sequence probe. Žb. CAT analysis showing that AP-2b does not efficiently inhibit basal transcriptional activity via the D1AS1 silencer in NS 20 Y cells. AP-2b expression vector and the indicated reporter plasmids were used to co-transfect NS20Y cells. The amount of nuclear extract used in this assay is similar to that in Fig. 4, lanes 1 to 4.
214
S. Takeuchi et al.r Molecular Brain Research 74 (1999) 208–216
them w21x. We synthesized a decoy phosphorothioate Ž S . oligonucleotide corresponding to the AP-2-consensus-like sequence within D1AS1 and a scrambled S-oligonucleotide as negative control. Cotransfection of the decoy oligonucleotide increased the CAT activity of reporter plasmid D1A-1510 fourfold compared with the scrambled oligonucleotide ŽFig. 5.. The effect of the decoy oligonucleotide was smaller but still definite on reporter plasmid D1A-561 containing D1AS1. Up-regulation by the decoy oligonucleotide was not observed with reporter plasmid D1A532 which lacks D1AS1. The non-specific effect of the S-oligonucleotide itself was assessed by comparing cells transfected with the scrambled oligonucleotide and nontransfected cells. The latter showed that transfection with S-oligonucleotide reduces CAT activity by 20–30%. These results indicated that D1AS1 plays an essential role in suppressing D 1A dopamine receptor gene expression in Neuro2a cells. Interestingly, up-regulation by the decoy oligonucleotide was most significant with the reporter plasmid D1A-1510 containing a longer upstream region of the D 1A gene ŽFig. 5b.. The latter observation might suggest that upstream sequences homologous to the AP-2 binding site w11x also function to repress the D 1A gene. 3.6. Expression of AP-2 b and D1 A genes We finally studied a possible relationship between the expression of AP-2b and D 1A genes using a D 1A -positive cell line and mouse brain tissues. In NS20Y cells known to express the D1A dopamine receptor protein w12x, RT-PCR revealed lower AP-2b levels than that in Neuro2a cells Ždata not shown., and gel mobility shift assay failed to detect any specific band with either probe A or with AP-2 consensus probe ŽFig. 6a.. Furthermore, in CAT assay of NS20Y cells co-transfected with AP-2b expression vector
Fig. 7. Gel mobility shift assay showing that the shifted bands corresponding to those in Neuro2a cells were reproduced with nuclear extract prepared from mouse temporal cortex ŽT., striatum ŽSt., and cerebellum ŽCl.. Black and white arrowheads indicate specific and non-specific bands, respectively, similar to those observed in Neuro2a cells.
and D 1A -CAT reporter plasmids, AP-2b-induced suppression of the D 1A gene was extremely weak ŽFig. 6b., indicating that AP-2b cannot suppress transcription efficiently in all neural cell lines. The latter finding suggests that the repressive effects of AP2b require a cell-specific mechanism that is present in Neuro2a cells but not in NS20Y cells. This conclusion is particularly plausible since AP-2b is a rather ubiquitous transcription factor in the brain. To test whether the D1AS1 silencer binds to a nuclear proteinŽs. in brain, we performed gel mobility shift assay with nuclear extracts from various mouse brain regions. Retarded bands were seen in all areas tested, but without a clear difference among various regions that express the D 1A gene at different levels ŽFig. 7..
4. Discussion Members of the AP-2 transcription factor family, three of which have been identified thus far, a , b and g, have generally been known to function as positive factors acting on the common GC-rich enhancer and up-regulating housekeeping genes w9,13,19,22,23x. However, AP-2a has recently been reported to suppress the CrEBPa gene during adipogenesis in 3Y3-L1 adipocytes w7x. In the present investigation, we demonstrated that another member of this family, AP-2b, is implicated in repressing transcription of the D 1A dopamine receptor gene in the D 1A -negative neural cell line, Neuro2a. This is the first demonstration of the repressive effect of AP-2b on gene expression. Furthermore, the present report confirms that AP-2b is a novel member of repressive transcription factors that regulate neuron-specific genes. AP-2b was initially reported as the second member of the AP-2 family which transactivates gene expression via the consensus AP-2 binding site w13x. A C-terminal domain known to mediate homodimerization of the previously cloned AP-2a is highly conserved between these two factors suggesting that AP-2b also forms a homodimer binding to the consensus sequence. Our findings suggested that AP2b interacts with the D1AS1 silencing sequence in the D 1A gene because Ž1. Neuro2a cell nuclear extract shifted D1AS1 and AP-2 consensus probes in a similar manner, Ž2. these cells have the mRNA for only the b member of the AP2 family, and Ž3. GST-AP-2b fusion protein binds to the D1AS1 sequence. Moreover, expression of AP-2b repressed transcription of the D 1A gene promoter in a D1AS1-dependent manner. These findings indicated that AP-2b is a repressive transcription factor of the D1A dopamine receptor gene in Neuro2a cells. AP-2b is expressed rather ubiquitously in the adult brain and its expression does not inversely correlate with that of the D 1A dopamine receptor gene. In addition, AP-2b does not remarkably repress transcription through
S. Takeuchi et al.r Molecular Brain Research 74 (1999) 208–216
the D1AS1 silencer in the D 1A -positive NS20Y cells. The mere binding of AP-2b to the D1AS1 sequence does not explain how it represses D 1A gene expression. Therefore, it is conceivable that cell type-specific cofactorŽs. regulate the function of AP-2b and that the amount of this unidentified factorŽs. is different among various brain regions. It is still possible that an additional proteinŽs. with a molecular weight similar to that of AP-2b binds to the AP-2 consensus sequence and represses D 1A gene expression in Neuro2a cells, because gel mobility shift with anti-AP-2b antibody could not supershift the entire band. Unlike the adult brain, the expression pattern of AP-2b in the developing brain has a mutually exclusive relationship with that of the D1A gene. In day 13.5 and 15.5 embryos, expression of AP-2b is observed in midbrain, spinal cord, cerebellum, and dorsal root ganglion w14x, regions of the nervous system that do not express the D 1A gene. The highest level of D 1A expression, however, is observed in the developing cerebrum which expresses only small amounts of AP-2b. Thus, a negative correlation appears to exist in the expression levels of these two genes. Therefore, AP-2b might be able to regulate the D1A dopamine receptor gene during embryonic development differently from that in adulthood. In addition to the sequence characterized in detail in this study, at least six more AP-2 binding consensus sequences are found within a 1.6-kb 5X upstream region of the human D 1A gene w11x. The latter is an intriguing observation in view of the fact that our previous CAT assay data in Neuro2a cells suggested that transactivation of this gene by Brn-4 is suppressed by multiple repressors w17x. This notion is consistent with our present results showing that an AP-2 decoy oligonucleotide enhances transcription from the D 1A promoter to a greater extent when used with reporter plasmids containing longer D 1A upstream regions ŽFig. 5b.. Consensus sequences for AP-2 binding are distributed widely in the promoters of numerous genes. Unraveling the mechanisms underlying AP-2 mediated transcriptional repression could lead to elucidation of the molecular diversity in the nervous system. Our results also showed that specific inhibition of D1AS1 by a decoy oligonucleotide could de-repress the D 1A gene promoter in Neuro2a cells. This notion raises the potential for regulating D 1A gene expression in vivo by using D1AS1 decoy oligonucleotides. Although such an effect could be different in neurons than in cultured cell lines, these findings, nevertheless, identify a specific molecular target for de-repressing the D 1A gene. We recently reported a novel transcription factor, mGIF, which is induced by GDNF and binds to the SP1 consensus sequence w24x. Co-transfection assays indicated that mGIF also represses transcription. Together with RESTrNRSF and the present study, these findings collectively suggest that many more repressive transcription factors exist that regulate neuron-specific genes. Such factors should play critical roles in generating the com-
215
plexity of the central nervous system composed of various types of neurons expressing different sets of specific genes.
Acknowledgements We thank Prof. Hiroshi Hamada ŽOsaka University, Institute of Molecular Biology. for Brn-4 expression vector and Prof. Marshall Nirenberg ŽNHLBI, NIH. for NS20Y cells. This work was supported by grants to H.O. from The Japanese Ministry of Education, Culture, Science and Sports Ž07558233, 10670574. and from The Japanese Ministry of Health and Welfare Ž6A-2.
References w1x C.G. Begley, S. Lipkowitz, V. Gobel, K.A. Mahon, V. Bertness, A.R. Green, N.M. Gough, I.R. Kirsch, Molecular characterization of NSCL, a gene encoding a helix–loop-helix protein expressed in the developing nervous system, Proc. Natl. Acad. Sci. USA 89 Ž1992. 38–42. w2x L. Brown, R.L. Espinosa, M.M. Le Beau, M.J. Siciliano, R. Baer, HEN1 and HEN2: a subgroup of basic helix–loop-helix genes that are coexpressed in a human neuroblastoma, Proc. Natl. Acad. Sci. USA 89 Ž1992. 8492–8496. w3x L. Brown, R. Baer, HEN1 encodes a 20-kilodalton phosphoprotein that binds an extended E-box motif as a homodimer, Mol. Cel. Biol. 14 Ž1994. 1245–1255. w4x J.A. Chong, J. Tapia-Ramirez, S. Kim, J.J. Toledo-Aral, Y. Zheng, M.C. Boutros, Y.M. Altshuller, M.A. Frohman, S.D. Kraner, A. Mandel, REST: a mammalian silencer protein that restricts sodium channel gene expression to neurons, Cell 80 Ž1995. 949–957. w5x B.P. England, A. Admon, R. Tjian, Cloning of Drosophila transcription factor Adf-1 reveals homology to Myb oncoproteins, Proc. Natl. Acad. Sci. USA 89 Ž1992. 683–687. w6x I. Imafuku, M. Kamei, I. Kanazawa, M.M. Mouradian, H. Okazawa, POU transcription factors differentially regulates the D1A dopamine receptor gene in cultured cells, Biochem. Biophys. Res. Commun. 222 Ž1996. 736–741. w7x M.-S. Jiang, Q.-Q. Tang, J. McLenithan, D. Geiman, W. Schillinglaw, W.J. Henzel, M.D. Lane, Depression of the CrEBPa gene during adipogenesis: identification of AP-2a as a repressor, Proc. Natl. Acad. Sci. USA 95 Ž1998. 3467–3471. w8x W. Lee, A. Haslinge, M. Karin, R. Tjian, Activation of transcription by two factors that bind promoter and enhancer sequences of the human metallothionein gene and SV40, Nature 325 Ž1987. 368–372. w9x L.A. Mcpherson, V.R. Baichwal, R.J. Weigel, Identification of ERF-1 as a member of the AP-2 transcription factor family, Proc. Natl. Acad. Sci. USA 94 Ž1997. 4342–4347. w10x N. Mermod, T.J. Williams, R. Tjian, Enhancer binding factors AP-4 and AP-1 act in concert to activate SV40 late transcription in vitro, Nature 332 Ž1988. 557–561. w11x M.T. Minowa, T. Minowa, F.J. Monsma, D.R. Sibley, M.M. MouraX dian, Characterization of the 5 flanking region of the human D1A dopamine receptor gene, Proc. Natl. Acad. Sci. USA 89 Ž1992. 3045–3049. w12x F.J. Monsma, D.L. Brasaard, D.R. Silby, Identification and characterization of D1 and D 2 dopamine receptors in cultured neuroblastoma and retinoblastoma clonal cell lines, Brain Res. 492 Ž1989. 314–324. w13x M. Moser, A. Imhof, A. Pscherer, R. Bauer, W. Amselgruber, F.
216
w14x
w15x
w16x
w17x
w18x
S. Takeuchi et al.r Molecular Brain Research 74 (1999) 208–216 Sinowatz, F. Hofstadter, R. Schule, R. Buettner, Cloning and characterization of a second AP-2 transcription factor: AP-2b, Development 121 Ž1995. 2779–2788. M. Moser, J. Ruschoff, R. Buettner, Comparative analysis of AP-2a and AP-2b gene expression during murine embryogenesis, Dev. Dyn. 208 Ž1997. 115–124. K. Okamoto, H. Okazawa, A. Okuda, M. Sakai, M. Muramatsu, H. Hamada, A novel octamer binding transcription factor is differentially expressed in mouse embryonic cells, Cell 60 Ž1990. 461–472. H. Okazawa, K. Okamoto, S. Ishino-Kaneko, Y. Takeda, Y. Toyoda, M. Muramatsu, H. Hamada, The oct-3 gene, a gene for an embryonic transcription factor, is controlled by a retinoic acid repressible enhancer, EMBO J. 10 Ž1991. 2997–3005. H. Okazawa, I. Imafuku, M.T. Minowa, I. Kanazawa, H. Hamada, M.M. Mouradian, Regulation of striatal D1A dopamine receptor gene transcription by Brn-4, Proc. Natl. Acad. Sci. USA 93 Ž1996. 11933–11938. H. Okazawa, J. Shimizu, M. Kamei, I. Imafuku, H. Hamada, I. Kanazawa, Bcl-2 inhibits retinoic acid-induced apoptosis during the neural differentiation of embryonal stem cells, J. Cell Biol. 132 Ž1996. 955–968.
w19x M. Oulad-Abedelghani, P. Bouillet, C. Chazaud, P. Dolle, P. Chambon, AP-2.2: a novel AP-2-related transcription factor induced by retinoic acid during differentiation of P19 embryonal carcinoma cells, Exp. Cell Res. 225 Ž1996. 338–347. w20x C.J. Schoenherr, D.J. Anderson, The neuron-specific silencer factor ŽNRSF.: a coordinate repressor of multiple neuron-specific genes, Science 267 Ž1995. 1360–1363. w21x M.X. Tang, C.T. Redermann, F.C. Szoka Jr., In vitro gene delivery by degraded polyamidoamine dendrimers, Bioconjugate Chem. 7 Ž1996. 703–714. w22x T. Williams, A. Admon, B. Luscher, R. Tjian, Cloning and expression of AP-2, a cell-type-specific transcription factor that activates inducible enhancer elements, Genes Dev. 2 Ž1988. 1557–1569. w23x J.A. Williamson, J.M. Bosher, A. Skinner, D. Sheer, T. Williams, H.C. Hurst, Chromosomal mapping of the human and mouse homologues of two new members of the AP-2 family of transcription factors, Genomics 35 Ž1996. 262–264. w24x S. Yajima, C.H. Lammers, S.H. Lee, Y. Hara, K. Mizuno, M.M. Mouradian, Cloning and characterization of murine glial cell-derived neurotrophic factor inducible transcription factor ŽMGIF., J. Neurosci. 7 Ž1997. 8657–8666.
