Cyclic AMP-mediated regulation of transcription factor Lot1 expression in cerebellar granule cells.

Lot1, a zinc finger transcription factor acting as a tumor suppressor gene on tumoral cells, is highly expressed during brain development. In developing rat cerebellum, Lot1 expression is high in cerebellar granule cells (CGC), a neuronal population undergoing postnatal neurogenesis. The time course of Lot1 cerebellar expression closely matches the expression of pituitary adenylate cyclase-activating polypeptide (PACAP) receptors coupled to adenylyl cyclase. The aim of this study was to ascertain whether Lot1 expression is regulated by cAMP-dependent pathways and to identify mechanisms of Lot1 activation in CGC cultures. Our results show that Lot1 expression in CGC is cAMP-dependent, as treatments with either forskolin or PACAP-38 induced an increase in its expression at both the mRNA and protein levels. This effect on Lot1 expression was mimicked by dibutyryl cAMP and suppressed by protein kinase A and MEK inhibitors. In parallel, we found that treatments with forskolin and PACAP-38 in precursor CGC inhibited bromodeoxyuridine incorporation by 25 and 35%, respectively, indicating a negative effect on neuronal precursor proliferation. Luciferase reporter analysis and mutagenesis of the Lot1 promoter region indicated a crucial role of the AP1-binding site (located at -268 bp) in cAMP-induced Lot1 transcription. In addition, cotransfection experiments indicated that the c-Fos/c-Jun heterodimer is responsible for cAMP-dependent Lot1 transcriptional activation. In conclusion, our data demonstrate that, in CGC, Lot1 is under the transcriptional control of cAMP through an AP1 site regulated by the c-Fos/c-Jun heterodimer and suggest that this gene may be an important element of the cAMP-mediated pathway that regulates neuronal proliferation through the protein kinase A-MEK signaling cascade.

Lot1 was initially characterized as a growth suppressor gene based on its decrease in a rat ovarian carcinoma cell line and hence called Lot1 for "lost in transformation" (1,2). Its mouse ortholog, which is highly homologous to the human and rat genes (LOT1 and Lot1, respectively), was independently identified by Spengler et al. (3,4) and designated as Zac1. The anti-proliferative properties of Zac1 were demonstrated by its ability to induce apoptosis and concomitantly to arrest the cell cycle in G 1 in osteosarcoma cell lines (3). In accordance with its anti-proliferative role, ablation of Zac1 gene expression increases cell proliferation (3,5). Lot1 encodes a transcription factor with seven zinc fingers of the Cys 2 -His 2 type. The presence of a specific DNA-binding region for Lot1 makes it a transcriptional regulator, acting either as an activator or a repressor of nuclear receptor activity (6 -8). A splice variant of Lot/ Zac1 was independently identified in humans by Kas et al. (7) and named PLAGL1 based on its homology to PLAG1 protein, encoded by the PLAG1 gene localized on chromosome 8q12 (9). Northern blot analysis of different tissues revealed the highest expression of Zac1 mRNA in the pituitary gland, lower levels of expression in several brain areas, and faint expression in some peripheral organs (3). Anatomical assessment of its distribution during development of mouse brain, carried out by in situ hybridization analysis, revealed differential, age-dependent distribution of Zac1 mRNA in various regions of the central nervous system (10). Recently, we have shown that Lot1 mRNA and protein expression is high in the cerebellum of rat pups in the early postnatal period, when neurogenesis of the most abundant cerebellar neuronal population, the granule cells (12), does take place, but undergoes progressive decreases at later stages (11). Moreover, analysis carried out in cultures of cerebellar granule cells (CGC) 2 demonstrated that Lot1 expression during differentiation of these neurons in some way recapitulates what happens during in vivo development. A relatively high expression was indeed present in freshly plated CGC, with a progressive decrease in activity occurring in parallel with differentiation toward mature neuronal phenotype (11).
The available data on the pattern of expression of Lot1/Zac1 during brain development (10,11) are compatible with this transcription factor playing a role in neurogenesis. Vertebrate neurogenesis is a complex process that includes regulation of neuron production, death, and differentiation. Therefore, identifying factors that are involved in the regulation of Lot1 expression during CGC differentiation could represent an important step toward deciphering a possible relationship between Lot1 expression and neurogenesis. Evidence from both in vivo and inculture studies suggests that several steps of neurogenetic process are crucially linked to an increase in intracellular cAMP (13,14). cAMP is a second messenger that regulates a striking number of physiological processes, including intermediary metabolism, cell proliferation, and neuronal signaling, by altering the basic pattern of gene expression (15)(16)(17). The cAMP cascade can be triggered in neuronal cells by different neurotransmitters/neuromodulators, neurotrophic factors, and other conditions known to influence neurogenesis (18). The neuropeptide pituitary adenylate cyclase-activating polypeptide (PACAP), which is a potent stimulator of adenylyl cyclase (19), has recently emerged as one of the factors able to regulate neurogenesis of prenatal hindbrain and cortical neuroblasts by exerting anti-mitogenic effects (20 -22). PACAP and its receptors are widely expressed in neurons of the embryonic and neonatal brains, exhibiting region-specific and developmentally regu-lated patterns of expression (21,(23)(24)(25). Furthermore, agonist and antagonist studies have demonstrate PACAP regulation of neuroblast proliferation and differentiation in both the peripheral and central nervous systems (20 -22, 26). In this context, the regulatory action of PACAP as a sensor coordinating various neurogenetic signals during development has been studied recently in cultures of CGC precursors with reference to the inhibition of the sonic hedgehog Shh mitogenic signal (27). The molecular mechanisms by which cAMP exerts its effects on neurogenesis are still largely unknown. In view of the evidence reported above, Lot1 may be one of the genes that are triggered by the cAMP cascade and that mediate the cAMP action on neurogenesis. The presence of a high density of PACAP receptors coupled to adenylyl cyclase in the external granule cell layer of the rat cerebellum during early postnatal development (28), in a time period closely matching Lot1 expression (11), seems to be in agreement with this hypothesis. The aim of this study was to ascertain whether the expression of Lot1 occurring during neurogenesis and differentiation of the cerebellar granule cells is induced by cAMP and to identify the molecular mechanisms that underlie its transcription.
Cloning of the Rat Lot1 Promoter and Construction of the Luciferase Reporter Vectors-For luciferase reporter vector construction, the rat Lot1 promoter region from 2099 bases upstream of the Lot1 transcription start site (GenBank TM accession number U72620) to 2 bases downstream (from base 8682293 to base 8684394 of contig NW_043337) was cloned into pBlueScript SK Ϫ (Stratagene) by nested PCR with primers Lot-F1 (5Ј-caaggaaagaaaaccacccc-3Ј) and Lot-R1 (5Ј-aactcccgagcgttctcc-3Ј) for the first reaction and primers Lot-F1nex (5Ј-cggatcctgggattacagatgctcataaa-3Ј) and Lot-R1nex (5Ј-cgggatccgcgagtgaggctggagaa-3Ј) for the second (with the BamHI sites used for cloning underlined). The ϳ2.1-kb insert representing the putative Lot1 promoter was subcloned upstream of the firefly luciferase gene into the pGL2-Basic vector with an EcoRV/SacI cut to generate the pGLot-2099 vector. Plasmids pGLot-1373, pGLot-515, and pGLot-327 were also generated by restriction endonuclease digestion of pGLot-2099. Mutation of the activating protein-1 (AP1)-binding site at Ϫ268 bases from the transcription start site was introduced by PCR. The region upstream of the AP1 site (from Ϫ2099 to Ϫ269 bases) was amplified by PCR with primers Lot-F1nex and LotAPm (5Ј-cggaattcttcggctgtctacgcacagc-3Ј), whereas the region downstream of the AP1 site (from Ϫ262 to ϩ2 bases) was amplified by PCR with primers Lot-R1nex and LotAPv (5Ј-cggaattccgcagcgggtgcgtgg-3Ј, with the EcoRI sites underlined). The EcoRI site overlapping the AP1 site at Ϫ268 bp was used for cloning and to introduce the mutation. The two segments (digested with BamHI and EcoRI) were cloned into pBlueScript to generate the full-length promoter bearing the mutation of the AP1 site at Ϫ268 bp (TGACTCA 3 GGAATTC). The mutated promoter was then subcloned into pGL2-Basic to generate vectors pGLot-2099APmut and pGLot-327APmut. Finally, the region downstream of the AP1 site was subcloned in pGL2-Basic to generate the pGLot-262 vector. All plasmids were sequenced with an ABI PRISM BigDye terminator cycle sequencing kit (Applied Biosystems, Foster City, CA) and run on an ABI PRISM 3730 DNA analyzer (Applied Biosystems).
Treatments-Unless specified otherwise, treatments were performed at 7 days in culture. Cells were stimulated with 10 M forskolin (Sigma), 10 nM PACAP-38 (Sigma), or 1 mM dibutyryl cAMP (Sigma) for the indicated times. The protein synthesis inhibitor cycloheximide (Sigma) was used at 10 g/ml. The phosphatidylinositol 3-kinase inhibitor wortmannin (Calbiochem) was used at 100 nM. The MEK1/2 inhibitor U0126 (Calbiochem) was used at 10 M. The putative protein kinase A (PKA) inhibitor KT5720 (Calbiochem) was used at 10 M. Because the latter inhibitor has been demonstrated recently to be a relatively unspecific inhibitor of several kinases (31), we also used the specific PKA inhibitor H-89 (10 M; Calbiochem) (32). Cells were preincubated with inhibitors for 1 h before stimulation.
Real-time Reverse Transcription-PCR-For quantification of Lot1 expression, total RNA was extracted from CGC with TriReagent (Sigma) according to the manufacturer's instructions. Approximately 5 g of RNA were digested with 10 units of RNase-free DNase (Promega) for 30 min at 37°C. After DNase inactivation, RNA was retrotranscribed with an oligo(dT) 12-18 primer (0.5 M) and 200 units of Moloney murine leukemia virus reverse transcriptase (Invitrogen) according to the manufacturer's instructions. For real-time PCR, the following Lot1, glyceraldehyde-3-phosphate dehydrogenase, and c-fos primers were used: rat Lot, 5Ј-TTAGCTGCGTAGTTGCGTGTTA-3Ј (sense) and 5Ј-CGGGTCCCTGAAAAGAACACA-3Ј (antisense); glyceraldehyde-3-phosphate dehydrogenase, 5Ј-GAACATCATCCCTGCATCCA-3Ј (sense) and 5Ј-CCAGTGAGCTTCCCGTTCA-3Ј (antisense); and c-fos, 5Ј-TGGACCTGTCTGGTTCCTTC-3Ј (sense) and 5Ј-ATG-CACCAGCTCAGTCAGTG-3Ј (antisense). Real-time PCR was performed using a SYBR Premix Ex Taq kit (TaKaRa, Shiga, Japan) according to the manufacturer's instructions in an iCycler iQ real-time PCR detection system (Bio-Rad). Fluorescence was determined at each step of every cycle. Real-time PCR was done under the following universal conditions: 2 min at 50°C, 10 min at 95°C, 50 cycles of denaturation for 15 s at 95°C, and annealing/extension for 1 min at 60°C. Relative quantitation was performed using the comparative threshold cycle method, in which arithmetic formulas are used to obtain the same result as the one yielded by the relative standard curve method. The comparative threshold cycle method can be used only when the target gene and the reference control gene (glyceraldehyde-3-phosphate dehydrogenase) have approximately equal amplification efficiency.
Northern Blotting-The expression of Lot1 in CGC was determined by Northern blotting. Approximately 30 g of total RNA/sample was loaded per lane and fractionated on a formaldehyde-containing 1% agarose gel. Following transfer to Hybond N membrane (Amersham Biosciences), Lot1 mRNA was detected using an ␣-32 P-labeled full-length cDNA probe. Cyclophilin probes were used for quantification as described previously (33). Membranes were washed at a maximum stringency of 0.1ϫ SSC and 0.1% SDS at 65°C.
Transfection-All cells were transfected with polyethyleneimine (25 kDa; Sigma) as described previously (34,35). For luciferase assays, HEK293 cells were plated 24 h before transfection in 24-well plates (10 5 cells/well). Cells were transfected with 1-2 g of plasmid DNA as indicated for 3 h in Dulbecco's modified Eagle's medium without serum. The pTK-␤gal reporter plasmid was cotransfected together with the Lot1-luciferase reporter vectors to normalize for transfection efficiency. For luciferase assays, CGC were plated in poly-D-lysine-coated 24-well plates (5 ϫ 10 5 cells/well) and transfected after 24 h with 1.05-1.55 g of plasmid DNA as indicated for 2 h in complete medium. CGC were cotransfected with the pRL-TK reporter plasmid (containing the Renilla luciferase gene) together with Lot1-luciferase reporter vectors to normalize for transfection efficiency.
Luciferase Assay-Luciferase activity in CGC was measured 36 -48 h after transfection using a Dual-Luciferase assay kit (Promega) according to the manufacturer's instructions. Firefly luciferase activity was normalized for each sample by dividing by the Renilla luciferase activity in the same sample.
Luciferase activity in HEK293 cells was measured as described previously (36) and normalized for ␤-galactosidase activity in the same sample (see below). Luciferase activity was measured with a TD-20/20 luminometer (Promega).
For the preparation of total cell extracts, cells were lysed in lysis buffer (10 mM Tris-HCl, 2% SDS, 10 mM dithiothreitol, and 1% protease and phosphatase inhibitor mixture (Sigma)). For the preparation of nuclear extracts, cells were lysed in low salt buffer (10 mM Hepes, 50 mM NaCl, 1 mM EDTA, 0.1% Nonidet P-40, 1 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride, pH 8) for 10 min at 4°C. After centrifugation, nuclei were extracted with hypertonic salt buffer (20 mM Hepes, 420 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, 10% glycerol, and 1 mM phenylmethylsulfonyl fluoride, pH 8). Cells extracts were immediately processed by Western blotting or kept frozen (Ϫ80°C) until assayed. The protein concentration of samples was estimated by the method of Lowry et al. (38). Equivalent amounts (50 g) of protein/sample were subjected to electrophoresis on an SDS-10% polyacrylamide gel. The gel was then blotted onto a nitrocellulose membrane, and equal loading of protein in each lane was assessed by brief staining of the blot with 0.1% Ponceau S. Blotted membranes were blocked for 1 h in 5% milk in Tris-buffered saline (10 mM Tris-HCl and 150 mM NaCl, pH 8.0) and 0.1% Tween 20 and incubated overnight at 4°C with primary antibodies. Membranes were washed and incubated for 1 h at room temperature with peroxidase-conjugated anti-rabbit IgG (1:1000 dilution; Amersham Biosciences). Specific reactions were revealed with the ECL Western blotting detection reagent (Amersham Biosciences).
Immunofluorescence-For immunofluorescence studies, the following antibodies were used: monoclonal anti-NeuN antibody (Chemicon International, Inc.), anti-neural cell adhesion molecule (NCAM) and anti-glial fibrillary acidic protein (GFAP) antibodies (Sigma), anti-bromodeoxyuridine (BrdUrd) and fluorescein isothiocyanate-conjugated anti-BrdUrd antibodies (Roche Applied Science, Mannheim, Germany), and anti-Lot1 polyclonal antibody (11). Cells (plated on poly-Dlysine-coated coverslips) were fixed for 30 min in 4% paraformaldehyde and 4% sucrose in 120 mM sodium phosphate buffer, pH 7.4, and then rinsed three times with phosphate-buffered saline (PBS). For BrdUrd immunofluorescence, coverslips were treated with 2 N HCl for 30 min at 37°C and extensively washed with PBS. Coverslips were incubated overnight at 4°C with appropriate dilutions of the primary antibody in 1.5% goat serum and 0.1% Triton X-100 in PBS, pH 7.4. Cells were then incubated with Cy3-conjugated goat anti-mouse secondary antibody or with fluorescein isothiocyanate-conjugated goat anti-rabbit secondary antibody (1:200 dilution; Sigma) for 1-2 h at room temperature. After all incubations, specimens were extensively washed with PBS containing 0.1% Triton X-100. Coverslips were mounted on glass slides in PBS containing 70% glycerol and Hoechst 33342 (2 g/ml). Confocal images were taken using a Leica TCS confocal microscope. Phase-contrast and fluorescence images were taken using a Nikon Eclipse TE 2000-S microscope equipped with a Zeiss AxioCam MRm digital camera.
Determination of the Labeling Index-CGC (plated on poly-D-lysinecoated coverslips) were cultured for 18 h, treated with 10 M BrdUrd for an additional 6 h, fixed, and processed for BrdUrd immunofluorescence as described above. The corresponding phase-contrast and fluorescence images (taken from random microscopic fields, 10 -12 for each coverslip) were superimposed and used to determine the labeling index), defined as the percent of BrdUrd-positive cells out of the total number of cells in three independent experiments performed in duplicate. Glial cells, which represent Ͻ5% of the total cell number in CGC cultures (37), were recognized on the basis of phase-contrast morphology and were excluded from scoring.
Cell Proliferation Assay-CGC (plated on poly-D-lysine-coated 96-well plates at 5 ϫ 10 4 cells/well) were cultured for 18 h and then treated with 10 M BrdUrd for an additional 6 h. BrdUrd incorporation was measured in quadruplicate with using a cell proliferation enzymelinked immunosorbent assay kit (Roche Applied Science) according to the manufacturer's instructions.
Statistics-Data are expressed as the means Ϯ S.E., and statistical significance was assessed by t test or one-way analysis of variance (ANOVA), followed by Bonferroni's or Dunnett's post hoc test, where appropriate. Differences were considered to be significant starting from p Ͻ 0.05. Statistical analysis was performed using GraphPad Prism 3.00. Fig. 1A, PACAP induced a large increase in Lot1 mRNA in differentiated CGC (7 days in vitro (DIV)), clearly implicating cAMP in this response. The effect of PACAP on Lot1 induction was indeed mimicked by the cAMP analog dibutyryl cAMP and by forskolin, a potent activator of adenylyl cyclase. Because it has been reported that Lot1 expression correlates with apoptotic cell death in other systems (3), we tested whether Lot1 expression is affected by serum and potassium withdrawal, a well established paradigm of apoptosis induction in CGC (12). Shifting to low potassium (5 mM)/serum-free medium did not alter Lot1 expression (Fig. 1A), indicating that it is not involved in induction of apoptosis in CGC. To study the time course of cAMP-mediated Lot1 up-regulation in CGC, we used forskolin to maximally increase intracellular cAMP and examined the expression of Lot1 at the mRNA level. Northern blot analysis of CGC culture extracts demonstrated that induction of Lot1 mRNA was detectable by 30 min after exposure to forskolin, reached its maximum between 2.5 and 6 h, and remained elevated for at least 12 h (Fig. 1B). By Western blot analysis using a specific anti-Lot1 antibody (11), we showed that induction of Lot1 expression by PACAP and forskolin at the protein level reflected that observed at the mRNA level ( Fig.  1, C and D).

PACAP and Forskolin Up-regulate Lot1 Expression through cAMP and Act as Anti-mitogenic Stimuli in CGC-As shown in
As it was previously demonstrated that Lot1 is a nuclear protein (3), we used immunocytochemistry to verify its subcellular localization in CGC. Cultures at 7 DIV were treated with forskolin or PACAP-38 for 24 h; and thereafter, cells were fixed and processed for immunocytochemistry as described under "Experimental Procedures." The results show that Lot1 was barely visible in the nuclei of untreated cells ( Fig.  2A), whereas forskolin (Fig. 2B) and PACAP (Fig. 2C) induced a generalized increase in expression in the nuclei of granule cells positive for the neuronal marker NCAM. The nuclei of some CGC exhibited a particularly strong up-regulation of nuclear protein expression (Fig. 2, B and C, arrows). Confocal microscopic observation of doubly immunostained cultures showed that Lot1 immunoreactivity was localized to nuclei of granule cells that were also positive for the antibody against the post-mitotic neuronal marker NeuN (Fig. 2D), but not to nuclei of astrocytes labeled by the glial marker GFAP (Fig. 2E).
To explore the signaling cascades involved in Lot1 gene expression, we examined the effect of inhibitors of various protein kinases known to be activated in response to the forskolin-induced cAMP increase. We found that induction of Lot1 mRNA caused by exposure to forskolin was strongly counteracted by co-treatment with the PKA inhibitor H-89, the MEK1/2 inhibitor U0126, or the multiple kinase inhibitor KT5720 (Fig.  3, A and B), but not by exposure to the phosphatidylinositol 3-kinase inhibitor wortmannin (Fig. 3A), which is know to block, at the concentration used, Akt phosphorylation in response to insulin activation in CGC cultures (39). In the presence of the protein synthesis inhibitor cycloheximide, forskolin induction of Lot1 was largely counteracted, implying the requirement of de novo protein synthesis for Lot1 mRNA induction (Fig. 3C).
Recent studies have reported that the cAMP-PKA cascade is one of the key signal transduction pathways involved in neuronal proliferation (13,14). To define the effects of adenylate cyclase activation on CGC proliferation, we analyzed BrdUrd incorporation in cells exposed to PACAP or forskolin for 24 h after cell plating. By Western blotting, we determined that also freshly plated CGC responded to forskolin stimulation with a strong induction of Lot1 expression, similar to the more differentiated cultures after 7 DIV (Fig. 4, inset). Experiments of double labeling with BrdUrd and the neuronal marker NCAM (Fig. 4A) or BrdUrd and GFAP (Fig. 4B) demonstrated that most BrdUrd-positive cells were also labeled by NCAM, indicating that the dividing cells were, at this stage, mainly neuronal precursors. Furthermore, microscopic observation of doubly immunostained cultures showed that Lot1 immunoreactivity did not colocalize with BrdUrd-positive cells after forskolin treatment (Fig. 4, C-E). Calculating the labeling index (labeled neurons/total neurons) based on direct counting of neurons identified by their morphology observed by phase-contrast microscopy, we found that exposure to forskolin or PACAP slowed down the CGC proliferation rate, which was decreased by ϳ60 or 70%, respectively, compared with untreated cells (Fig. 4, F-I). Moreover, we verified whether the PKA or MEK1/2 inhibitor that blocked forskolin-or PACAP-induced Lot1 up-regulation also counteracted the decrease in BrdUrd incorporation in CGC precursors. We found that the reduced CGC proliferation rate caused by exposure to forskolin or PACAP was significantly counteracted by co-treatment with H-89 or U0126 (Fig. 4I), whereas these two inhibitors had, by themselves, no effect on cell proliferation. Taken together, these results clearly correlate the effect of cAMP on neuronal proliferation with Lot1 expression.
Functional Analysis of the Lot1 Promoter-We cloned the 5Ј-flanking region of the Lot1 gene into the pGL2-Basic vector upstream of the luciferase gene to determine whether this region possesses transcriptional activity. The cloned rat Lot1 promoter region appears to be highly similar to the mouse Lot1 promoter (GenBank TM accession number AF314094) as observed by aligning the two sequences (data not shown). To identify, within the Lot1 promoter, the binding region responsible for the forskolin-induced activation, serial deletions were made. Complete or deleted constructs were transfected in CGC as described under "Experimental Procedures," and cells were stimulated with forskolin for 24 h before luciferase measurements. To obtain better transfection efficiency, these experiments were done in cultures transfected 24 h after plating and stimulated with forskolin for 24 h after 1 additional day in vitro. The structures of the different constructs are schematically shown in Fig. 5 together with the results of experiments performed using the different plasmids. Insertion of the ϳ2.1-kb rat Lot1 genomic fragment (nucleotides Ϫ2099 to ϩ2 with respect to the 5Ј-end of the Lot1 sequence in GenBank TM accession number U72620) into pGL2 (designated pGLot-2099) resulted in a strong increase (up to ϳ30-fold) in luciferase activity, additional activation in the presence of PACAP (up to ϳ60-fold) (data not shown), and much stronger activation (up to ϳ160fold) in the presence of forskolin (Fig. 5). A computer program (Mathinspector) was used to identify putative binding sites for transcription factors within this region. An AP1 site (located at Ϫ268 bp from the transcription start site and conserved in the mouse Lot1 promoter) was found. Results from transfection studies based on different truncated constructs indicated that the cAMP-dependent activation domain in the Lot1 promoter is in this region. Indeed, the transcriptional activity stimulated by forskolin was abolished following the 5Ј-truncation of the promoter at nucleotide 262, as shown for the pGLot-262 plasmid, whereas other truncated constructs only marginally affected forskolininduced transcriptional activity (Fig. 5). Accordingly, a truncated construct conserving a mutated form of this site (pGLot-327AP1mut) did not exhibit any significant response to forskolin stimulation (Fig. 5). A similar mutation in the complete construct (pGLot-2099AP1mut) also dramatically decreased the effect of forskolin stimulation (compare with pGLot-2099). A minor effect of forskolin over basal conditions was noticed with this construct, possibly due to the presence in the long upstream genomic fragment of forskolin-responsive site(s) with lesser effect on transcription. These results demonstrate that full activation of Lot1 expression by forskolin requires the intact AP1 site at Ϫ268 bp. Accordingly, a construct in which luciferase expression was driven by tandem repeats of AP1 consensus motifs (pAP1-Luc) (Fig. 5), used as control for the AP1-dependent transcriptional activity induced by forskolin, exhibited strong activation when transfected in CGC. Moreover, these experiments revealed that the promoter activity of the constructs deleted or mutated in the AP1 site (pGLot-262, pGLot-2099APmut, and pGLot-327APmut) caused a decrease in basal transcription, suggesting that this site is also important for unstimulated basal transcription (Fig. 5). . GAPDH, glyceraldehyde-3-phosphate dehydrogenase. C, the dependence of forskolin-induced Lot1 mRNA expression on protein synthesis was evaluated by realtime reverse transcription-PCR from CGC at 7 DIV that had been treated with forskolin (10 M) alone or in combination with the protein synthesis inhibitor cycloheximide (10 g/ml). Data are expressed as the means Ϯ S.E. of three independent experiments. *, p Ͻ 0.001 compared with the control; #, p Ͻ 0.001 compared with forskolin-stimulated samples (Bonferroni's test after ANOVA).

Induction of Lot1 Transcriptional Activity and c-Fos Expression by
Forskolin in CGC-To study in further detail the mechanism of cAMPmediated induction of Lot1 transcriptional activity, reporter gene assays were performed using plasmids containing a tandem repeat of Lot1, CREB/ATF-1, AP1, or NF-B consensus response elements placed upstream of a luciferase cDNA (pLot1-Luc, pCRE-Luc, pAP1-Luc, and pNFB-Luc) or a control plasmid (pTK-Luc). Twelve hours after transfection with the reporter plasmids, cells were cultured for an additional 24 h in the presence or absence of forskolin. Although no changes in luciferase activity were observed upon forskolin treatment in pNFB-Luc-or pTK-Luc-transfected cells, this treatment remarkably induced the transcriptional activity of pLot1-Luc (ϳ5-fold), pCRE-Luc (ϳ25fold), and pAP1-Luc (ϳ3.5-fold) (Fig. 6A), thus showing that, besides CREB and AP1 transcriptional activity, also Lot1 transcriptional activity was induced by forskolin in CGC.
As shown in Fig. 3C, inhibition of the MEK-ERK cascade by the MEK inhibitor U0126 or of the PKA cascade by the PKA inhibitor H-89 abrogated Lot1 induction by forskolin, implying that forskolin induces Lot1 expression through activation of these pathways. In agreement with this conclusion, exposure of CGC to forskolin caused a rapid (30 min) and long lasting (3 h) phosphorylation of MEK1/2 and ERK1/2 (Fig. 6B). The MEK-ERK signaling pathway regulates the activity of AP1 sites by increasing both the synthesis and activation of the ligands of this binding site, the immediate-early genes c-fos and c-jun (40). The protooncogene c-fos is known to play a key role in the control of cell proliferation and differentiation and programmed cell death (41)(42)(43)(44) and to be induced by PACAP (32). We therefore examined the kinetics of expression of c-Fos as well as c-Jun in CGC after forskolin treatment. As shown in Fig. 6B, forskolin induced c-Fos protein expression, which peaked 3 h after forskolin stimulation, but not c-Jun expression. Inhibition of PKA by H-89 prevented forskolin-induced MEK1/2 phosphorylation, with a consequent decrease in ERK1/2 phosphorylation (Fig. 6B), CREB phosphorylation (Fig. 6B), and c-Fos induction (Fig. 6, B and D). These results demonstrate that PKA is required for forskolin-induced transcription of c-Fos in CGC. Forskolin-induced c-Fos expression at both the protein and RNA levels was also significantly decreased by directly interfering with the MEK1/2-ERK1/2 pathway through the specific MEK1/2 inhibitor U0126 (Fig. 6, B-D). Additional proof of the importance of the MEK1/2-ERK1/2 pathway in cAMP-induced c-Fos expression was provided by a reporter gene assay using a plasmid for the AP1 consensus response element (pAP1-Luc); c-Fos transcriptional activity induced by forskolin (Fig. 6A) was strongly reduced by co-treatment with U0126 (data not shown).
Lot1 Transcription Does Not Depend on cAMP and the AP1 Promoter Site in HEK293 Cells-We examined the promoter activity of Lot1 in a non-neuronal cell model, HEK293. Transfection of these cells with the Lot1 reporter vector pGLot-2099 resulted in remarkable levels of luciferase activity. However, the transcriptional activity did not respond to stimulation by forskolin, and the AP1 site located between Ϫ268 and Ϫ262 bp was not essential for transcription (Fig. 7A). Furthermore, c-Fos activity in HEK293 cells was not stimulated but was actually depressed by forskolin, whereas the MEK-ERK cascade was only slowly and moderately activated (Fig. 7B). To study in further detail Lot1 transcriptional activity in HEK293 cells, reporter gene assays were performed using the pAP1-Luc or pCRE-Luc reporter plasmid. Although forskolin treatment remarkably induced the transcriptional activity of pCRE-Luc, no change in luciferase activity was observed in pAP1-Luctransfected cells, thus confirming that AP1 transcriptional activity was not induced by forskolin in HEK293 cells (Fig. 7C). Moreover, the basal level of c-Fos in HEK293 cells is not responsible for the strong Lot1 promoter activation, as cotransfection of pGLot-2099 with the dominant-negative mutant A-Fos did not affect Lot1 transcription, whereas treatment with a PKA or MEK1/2 inhibitor did not decrease basal Lot1 transcription in these cells (data not shown). Taken together, these results converge in demonstrating the lack of effect of cAMP on Lot1 expression in HEK293 cells.
Fos/Jun Family Proteins Control Lot1 Promoter Activity-To determine the mechanisms leading to AP1 site activation on the Lot1 promoter in CGC, we examined the transcriptional effects of Fos and Jun family members in transient transfection assays. We transfected wildtype expression constructs (as indicated in Fig. 8A) or transactivationdeficient c-Fos (A-Fos) together with the minimum reporter construct responsive to increasing levels of cAMP (pGLot-327) into CGC, and the cells were then stimulated or not with forskolin. We found that only concomitant c-Fos/c-Jun overexpression significantly increased pGLot-327 promoter activity under basal and forskolin-stimulated conditions (Fig. 8A). All other conditions tested (overexpression of c-Fos, Fra1, c-Jun, JunB, c-Fos/JunB, or c-Jun/JunB) failed to induce any significant increase in pGLot-327 activity. On the other hand, overexpression of dominant-negative A-Fos significantly decreased pGLot-327 activity upon forskolin treatment. We were next interested in determining whether enforced overexpression of Fos and Jun would activate the Lot1 promoter at the level of the AP1 site. To perform these experiments, we cotransfected Lot1-luciferase reporter plasmids (pGLot-2099, pGLot-2099AP1mut, pGLot-327, pGLot-327AP1mut, or pGLot-262) together with the Fos and Jun expression plasmids. Fig. 8B shows that overexpression of c-Fos and c-Jun approximately doubled Lot1 transcription only when the Lot1-luciferase reporter plasmids contained the intact AP1 site (pGLot-2099 and pGLot-327). Furthermore, c-Fos/c-Jun overexpression no longer stimulated reporter gene activity when the AP1 site was selectively mutated (Fig. 8B). These results indicate that Lot1 up-regulation elicited by cAMP elevation relies upon AP1-mediated transcriptional activity and that the composition of the heterodimer that binds at the AP1 site on the Lot1 promoter is likely c-Fos/c-Jun. We also tested whether Fos/Jun overexpression activates Lot1 in HEK293 cells, in which cAMP does not induce c-Fos and Lot1 expression. These experiments revealed that the effect of overexpression of c-Fos/c-Jun on Lot1 transcription in HEK293 cells was similar to that observed in CGC (Fig. 8B).
We have shown above that, in CGC, treatments that increased cAMP levels promoted CREB phosphorylation (Fig. 6B). To test whether CREB may be a primary factor in controlling cAMP-dependent Lot1 transcriptional activation, we performed cotransfections with the Lot1-luciferase reporter plasmid pGLot-2099 or pGLot-327 (a truncated form of Lot1 lacking any obvious cAMP response element (CRE)-binding site) and the dominant-negative mutant A-CREB (45). Fig. 8C shows that, first of all, the Lot1 promoter constructs possessing (pGLot-2099) or not (pGLot-327) a CRE were stimulated by forskolin in a similar manner, thus suggesting that CREB is not necessary for Lot1 activation. Under our experimental conditions and in the presence of very strong activation of endogenous CREB by forskolin (see Fig. 6, A and B), the dominant-negative mutant A-CREB did not significantly inhibit forskolininduced Lot1 transcription in both constructs (Fig. 8C). As a positive control, a plasmid containing tandem repeats of the CRE (pCRE-Luc) was used. The results shown in Fig. 8C demonstrate that, in this control plasmid, A-CREB was able to strongly inhibit both basal and forskolininduced pCRE-Luc transcription (Fig. 8C).

DISCUSSION
In this study, we have shown, at both the mRNA and protein levels, that expression of the growth suppressor gene Lot1 is up-regulated in CGC after stimulation of cAMP production by PACAP-38 or forskolin. Upper, schematic representation of the cloned Lot1 promoter region (not drawn to scale). The AP1 response element and CRE found by software analysis are indicated. Lower left, schematic representation of the reporter plasmids used in these experiments. The plasmids contained a progressively shorter fragment of the Lot1 promoter or Lot1 promoter site mutants driving the expression of the luciferase gene. Lower right, luciferase reporter analysis of the Lot1 promoter. One day after plating, CGC were transfected with 1 g of the indicated Lot1-luciferase (LUC) reporter plasmids and 0.05 g of Renilla luciferase reporter plasmid (pRL-TK) to normalize for transfection efficiency. Twenty-four hours after transfection, cells were incubated with (black bars) or without (white bars) 10 M forskolin for an additional 24 h. The results are given as -fold increase over pGL2-Basic activity and are the means of four to five experiments done in triplicate. Data were analyzed with Bonferroni's test after ANOVA. **, p Ͻ 0.001, and *, p Ͻ 0.01, difference in the forskolin-stimulated versus basal conditions; ##, p Ͻ 0.01, and #, p Ͻ 0.05, difference in the basal versus pGLot-2099 basal conditions; E, p Ͻ 0.05, difference in the forskolinstimulated versus pGLot-2099 stimulated conditions. A post hoc t test confirmed the results from Bonferroni's test and additionally revealed a difference between pGLot-2099AP1mut-stimulated versus basal conditions. §, p Ͻ 0.05 (t test).
Physiologically, Lot1 induction parallels the negative regulation of CGC precursor proliferation mediated by cAMP, as we never observed its induction in dividing, i.e. BrdUrd-positive, cells. Furthermore, Lot1 induction was not related to apoptotic elimination of mature CGC whose survival was challenged by shifting them to non-depolarizing conditions and serum deprivation. These results implicate Lot1 as one of the elements in the cascade of molecular events triggered by cAMP and favoring neuronal survival and differentiation. We have also demonstrated that, in CGC, cAMP-dependent Lot1 stimulation requires activation of PKA and MEK/ERK as well as induction of the immediate-early gene c-fos. We have identified a putative promoter region within 2.1 kb 5Ј-upstream of the Lot1 transcription start site, and we have characterized a regulatory element that is involved in neuronal expression of the rat Lot1 gene after cAMP stimulation. We have provided evidence that the transcription of this gene depends upon an AP1 site in the proximal promoter regulatory region (268 bp 5Ј-upstream of the transcription start site) through the binding of the c-Fos/c-Jun heterodimer. The cAMP-dependent effects elicited by the exogenous adenylate cyclase activator forskolin were also replicated by an endogenous putative neurotrophic factor, the peptide PACAP, which we have demonstrated here to down-regulate neuronal precursor proliferation in parallel with induction of Lot1 expression. PACAP and its receptors are present in the developing nervous system (46 -48). PACAP belongs to a peptide family that includes secretin, glucagon, growth hormonereleasing factor, and vasoactive intestinal peptide and interacts with target cells via three G-protein-coupled receptors (19). Cerebellar granule neurons express PACAP and its receptors during early neurogenesis in vivo, and the presence of the peptide elicits a survival response in culture (49 -51). Notably, a high density of PACAP receptors linked to adenylyl cyclase activation was found in the external granular layer of the developing cerebellum (28), with a time window temporally matching Lot1 expression in the same layer (11). Activation of the PAC1 receptors by PACAP stimulates adenylyl cyclase and phospholipase C activities in CGC, leading to activation of PKA and protein kinase C (49,50). The signal transduction pathways that mediate the anti-apoptotic effect of PACAP in cerebellar granule neurons have been investigated in several experimental models. Although the cAMP-PKA pathway has been implicated in the survival response (52)(53)(54), the role of phospholipase C and protein kinase C is still controversial (55,56). The MEK-ERK pathway is a major signaling pathway in neural cells that can be regulated by PKA, and recent studies have shown that inhibition of activation of the PKA-dependent MEK-ERK pathway abolishes the protective effect of PACAP on rat CGC (57). Therefore, in addition to neurogenesis, PACAP is important for neuroprotection, acting directly or eliciting release of trophic signals, most likely from glial cells (48,58). Here, we have demonstrated that PACAP increases Lot1 gene expression in cultured rat cerebellar neurons, thus suggesting the involvement of Lot1 in the PACAP-mediated survival-and differentiation-promoting action in neurodevelopment. This relationship is supported by studies demonstrating neurotrophic interactions between Lot1/Zac1 and PACAP (59) as well as by available reports on Lot1/Zac1 expression in some brain areas during development (10,11). It is also important to note that, in this study, increased Lot1 expression induced by PACAP as well as by forskolin clearly did not compromise cell viability, as no signs of apoptosis could be detected in Lot1-positive neurons. Recent findings in animal models of excitotoxic injuries go in the same direction by linking Lot1 expression to neuroprotection and brain plasticity (10, 60, 61). Interestingly, in these models, a relationship between Lot1 expres- The role of cAMP in the maturation of newborn neurons is well accepted (13,14), and both PKA-dependent and PKA-independent pathways have been implicated in cAMP-mediated effects. Our study has shown that, in CGC, cAMP activation of Lot1 is PKA-and MEK/ ERK-dependent, as it was counteracted by specific pathway inhibitors. Notably, the PKA-dependent mechanism implicated in Lot1 activation did not act through the CRE site present in the gene promoter. Indeed, deletion of the CRE-binding site on the cloned promoter did not abro-gate Lot1 activation after forskolin stimulation. Besides CREB, previous studies have shown that the cAMP-PKA pathway also involves the MEK-ERK signaling cascade (62). Our data imply that forskolin-induced Lot1 transcription in CGC is mediated by the PKA and MEK-ERK pathways. In many neuronal cells, increased cAMP can induce neuronal differentiation through activation of the mitogen-activated protein kinase cascade (63,64). We have shown that inhibition of PKA completely blocked MEK first and consequently ERK1/2 activation by forskolin, demonstrating dependence of ERK activation by PKA. Moreover, blockade of forskolin-induced MEK activity with a selective inhibitor (U0126) not only reduced the magnitude and duration of ERK phosphorylation, but also blocked transcriptional activation of the Lot1 gene. These results provide clear evidence that ERK activation is absolutely required for transcriptional activation of the Lot1 gene. Interestingly, prolonged or potentiated ERK activation in some cell types leads to the expression of a differentiated phenotype such as neurite outgrowth in PC12 cells (65,66). One of the targets of activated MEK/ERK in mammalian cells is the transcription factor AP1, which is composed of members of the Fos and Jun family of DNA-binding proteins (67,68). The MEK-ERK signaling pathways regulate AP1 activity in several ways, including increased synthesis and/or activation of c-Fos (40). Our present results clearly suggest a link between the PKA-MEK-ERK pathways and c-Fos activity in the regulation of Lot1 expression through the critical AP1 promoter site disclosed by the reported experiments. To be completely described in mechanistic terms, this link will only require that the step from ERK activation to c-Fos expression is identified.
We isolated, mapped, and functionally characterized the 5Ј-flanking region of the rat Lot1 gene. By both 5Ј-deletion analysis and site mutagenesis of the native Lot1 promoter, we directly demonstrated that the region between Ϫ268 and Ϫ262 bp (containing an AP1-binding site) is necessary to achieve a full transcriptional response to forskolin stimulation. Moreover, our results show that forskolin up-regulated the expression of c-Fos, implicating this immediate-early gene as the transcriptional mediator linking ERK phosphorylation to Lot1 promoter activation. Our data show, however, that overexpression of c-Fos alone was not sufficient to activate the Lot1 promoter, but cotransfection of c-Jun and c-Fos together resulted in significant induction of Lot1 promoter activity. Overexpression of other heterodimers such as c-Fos/ JunB and c-Jun/JunB was not effective. The unique role of the c-Fos/c-Jun heterodimer in activation of the AP1-binding complex on the Lot1 promoter is supported by A-Fos overexpression studies. Putative knockout of the AP1-binding activity by overexpression of A-Fos was effective in blocking activation of Lot1 promoter activity. This dominant interfering mutant has been shown previously to disrupt c-Fos/c-Jun heterodimer binding to DNA in a remarkably specific manner (45).
Our results provide novel information on the functional significance of the up-regulation of Lot1 expression by cAMP stimulation in CGC. First, we have shown that Lot1 is induced in CGC by the anti-mitogenic and differentiative stimulus exerted by cAMP on these neuronal cells. Second, our analysis revealed that the cellular mechanisms regulating Lot1 activation, i.e. the PKA-MEK-ERK cascade and c-Fos induction, are the same that regulate crucial steps of the neurogenetic process (69). Third, the characterization of these mechanisms, together with the previous data on Lot1 expression in brain development, enforces the notion that the gene may play an essential role in neurogenesis, with particular reference to cerebellar development (11). Fourth, regulation of the Lot1 gene, described here for neuronal cells, may be different in other cell types. It has been indeed noted in the Introduction that Lot1 and its homologs act as tumor suppressor genes in cancer cells (3,5). With the present observations, we have shown that the epithelial cell line HEK293 responds to cAMP elevation in a very different way compared FIGURE 8. The c-Fos/c-Jun heterodimer is responsible for cAMP-mediated Lot1 transcriptional activity. A, CGC were cotransfected with 0.5 g of the pGLot-327 reporter plasmid without (Control) or with 0.5 g of the indicated expression vector for AP1 family proteins in the presence of 0.05 g of pRL-TK. DNA quantity for each transfection was equalized with empty vector pcDNA3. Twenty-four hours after transfection, cells were incubated with (black bars) or without (white bars) 10 M forskolin for an additional 24 h. The results are expressed as percent induction over control conditions and are the means Ϯ S.E. of three to four experiments done in triplicate. *, p Ͻ 0.01 versus control conditions; **, p Ͻ 0.001 versus control conditions; #, p Ͻ 0.05 versus control conditions (Dunnett's test after ANOVA). B, shown is the specificity of c-Fos and c-Jun binding to the AP1 element involved in Lot1 promoter activation. CGC were cotransfected with 0.5 g of the indicated Lot1-luciferase reporter plasmids without (control) or with 0.5 g of c-Fos and c-Jun expression plasmids in the presence of 0.05 g of pRL-TK. DNA quantity for each transfection was equalized with empty vector pcDNA3. HEK293 cells were cotransfected with 0.5 g the indicated Lot1-luciferase reporter plasmids without (control) or with 0.5 g of c-Fos and c-Jun expression plasmids in the presence of 0.5 g of pTK-␤gal. DNA quantity for each transfection was equalized with empty vector pcDNA3. The results are given as percent induction over control conditions and are the means Ϯ S.E. of three to four experiments done in triplicate. *, p Ͻ 0.01 versus pGLot-2099APmut, pGLot-327APmut, and pGLot-262 (Dunnett's test after ANOVA). C, CGC were cotransfected with 0.75 g of pGLot-2099, pGLot-327, or pCRE-Luc reporter plasmid without (Control) or with 0.75 g of the expression vector for dominant-negative A-CREB in the presence of 0.05 g of pRL-TK. DNA quantity for each transfection was equalized with empty vector pcDNA3. Twenty-four hours after transfection, cells were incubated 10 M forskolin for an additional 24 h. The results are given as the percent over basal control conditions and are the means Ϯ S.E. of three experiments done in triplicate. *, p Ͻ 0.01 versus control conditions (Dunnett's test after ANOVA).