Mitogen-activated protein kinase kinase negatively modulates ciliary neurotrophic factor-activated choline acetyltransferase gene expression Tiffany Mellott 1 , Ignacio Lopez-Coviella 2 , Jan Krzysztof Blusztajn 1,2 and Brygida Berse 1 1 Department of Pathology and Laboratory Medicine and 2 Department of Psychiatry, Boston University School of Medicine; Boston, MA, USA The expression of t he c holine acetyltransferase ( ChAT) enzyme that synthesizes the neurotransmitter acetylcholine (ACh) is upregulated by ciliary neurotrophic factor (CNTF). We s tudied the involvement of the mitogen -acti- vated protein kinase (MAPK) pathway in regulating ChAT expression in a murine septal cell line. Surprisingly, we found that PD98059 and U0126, two structurally dis- tinct inhibitors of MAPK kinase (MEK1), increased both basal and CNTF-induced ACh p roduction. Transien t transfections with ChAT promoter-luciferase reporter construct demonstrated synergy between PD98059 and CNTF at the transcriptional level. Moreover, in cotrans- fection studies, overexpression of constitutively activated MEK1 completely abrogated the CNTF-mediated induc- tion of the reporter. Blocking MEK1 did not signi®cantly alter CNTF-induced Tyr705 phosphorylation of the prin- cipal mediator of the CNTF pathway, the transcription factor Stat3. However, PD98059 inhibited Ser727 phos- phorylation of Stat3, demonstrating that the latter is MEK1-dependent. Taken together, these results i ndicate that activation of the MEK1/MAPK pathway inhibits the CNTF-mediated stimulation o f C hAT exp ression, possibly as a p art of a feedback mechanism. Keywords: acetylcholine; choline acetyltransferase; MAP kinase; c iliary neurotrophic factor; Stat3. Biosynthesis o f the neurotransmitter acetylcholine (ACh) from choline and acetyl coenzyme A is catalyzed by the enzyme choline acetyltransferase (ChAT, acetyl-CoA:cho- line O-acetylotransfe rase, EC 2.3.1.6) [1]. Following syn- thesis, ACh is transported into synaptic vesicles by vesicular acetylcholine transporter (VAChT). ChAT and VAChT are encoded within one gene locus and share regulatory elements for transcription [2]. Recent studies have shown that the two genes are coordinately regulated in response to various extracellular factors, including nerve growth factor (NGF) and ciliary neurotrophic factor (CNTF) [3±10]. Signal transduction pathways involved in this regulation are largely unknown. Quirin-Stricker et al. identi®ed s everal transcription factors involved in transcriptional activation of the human cholinergic locus by NGF and demonstrated that cyclic AMP-dependent protein kinase (PKA) is important fo r this process [11]. Recently, it was shown that regulation of the rat cholinergic gene locus by cyclic AMP is mediated by PKA type II, but not by PKA type I [12]. In a previous study, we demonstrated that NGF, acting through receptor t yrosine kinase TrkA, can activate cholinergic gene e xpression and ACh production in the murine septal cell line SN56T17 [9]. At the same time, NGF downregulates the CNTF-induced increases in ACh pro- duction, pointing to interference between the NGF and CNTF signaling pathways. As the major signaling event evoked by NGF in many cell types, including SN56T17, is the activation of mitogen-activated p rotein kinase (MAPK), we investigated the role of this pathway in regulating the cholinergic phenotype and speci®cally in modulating the CNTF effect. MAPK cascades are present in all eukaryotic cells and are utilized in many signal transduction pathways origi- nating from receptors at the cell s urface [13]. The Ras/Raf/ ERK pathway is activated by growth factors and cyto- kines, including NGF and CNTF [14,15]. The transient formation of Ras-GTP and activation of R af kinase at the cell membrane eventually leads to the phosphorylation of the two members of MAPK family, ERK1 (p44MAPK) and ERK2 (p42 MAPK) b y MAPK kinase 1 (MAPKK1 or MEK1) [14]. Activated ERKs phosphorylate and regulate numerous targets in the cytoplasm and in the nucleus, including other protein kinases, transcription factors, cytoskeletal proteins, g rowth factor r eceptors, and other regulatory proteins [13]. Speci®city of MAPK effects in different cell types and in response to different stimuli is achieved by the action of scaffolding proteins, which bring together the components of a given MAPK pathway [16]. Correspondence t o B. Berse, Department of Pathology and Laboratory Medicine, Boston University School of Medicine, 85 East Newton Street, Room M1006, Boston, MA 02118, USA. Fax: + 1 617 638 5400, Tel.: + 1 617 638 5960, E-mail: berse@bu.edu Abbreviations: ACh, acetylcholine; ChAT, choline acetyltransferase; CNTF, ciliary neurotrophic factor; DMEM, Dulbecco's modi®ed Eagle's medium; ERK, extracellular signal-regulated kinase; JAK, Janus kinase; LIF, leukemia inhibitory factor; MAPK, mitogen-acti- vated protein kinase; MEK, mitogen-activated protein kinase kinase; NGF, nerve growth factor; PKA, cyclic AMP-activated protein kin- ase; STAT, signal transducer and activator of transcription; VAChT, vesicular acetylcholine transporter; VIP, vasoactive intestinal peptide. (Received 6 June 2001, revised 12 November 2001, accepted 30 November 2001) Eur. J. Biochem. 269, 850±858 (2002) Ó FEBS 2002 In this study we employed two distinct noncompetitive MEK1 inhibitors to prevent the phosphorylation of MAPK by MEK1 in SN56T17 cells. We report that inhibiting the MEK1/MAPK pathway increases both basal and CNTF- induced ACh synthesis and ChAT promoter activity in SN56T17 cells. Moreover, introducing constitutively active MEK1 into the c ells by transient transfection abrogates the stimulatory effect of CNTF. Experiments with the JAK/ STAT pathway inhibitor tyrphostin AG490 suggest that CNTF-induced ACh production is JAK/STAT dependent. Blocking MAPK does not signi®cantly in¯uence tyrosine phosphorylation of the transcription factor Stat3 evoked by CNTF, but it reduces its serine phosphorylation. We propose that the MEK1/MAPK cascade negatively regu- lates ACh synthesis, possibly as a part of a feedback mechanism. EXPERIMENTAL PROCEDURES Materials PD98059 was obtained from Calbiochem-Novabiochem (La Jolla, CA, USA). U0126, luciferase reporter vector pGL3-Basic, and Luciferase Assay System were from Promega (Madison, WI, USA). The expression plasmid pUSE-MEK1(activated), as well as the control v ector pUSE were obtained from Upstate Biotechnology (Lak e Placid, NY, USA). CNTF was purchased from R & D Systems (Minneapolis, MN, USA) and Pepro Tech Inc. (Rocky Hill, NJ, USA). Fetal bovine serum, N2 supplement, and LipofectAMINE were obtained from Gibco BRL (Bethesda, MD, USA). Polyclonal anti- ChAT Ig AB144P and AB143 were purchased from Chemicon (Temecula, CA, USA), and all other antibod- ies [anti-(phospho p44/42 MAPK) Ig, anti-Stat3 Ig, anti- (phosphoSer727 Stat3) Ig, anti-(phosphoTyr705 Stat3) Ig] were from New England Biolabs (Beverly, MA, USA). Immobilon P membrane for Western blotting was from Millipore (Bedford, MA, USA). Renaissance reagent for chemiluminescence was obtained from N ew England Nuclear (Boston, MA, USA). Sigma Chemical Co. (St Louis, MO, USA) supplied all other chemicals and reagents. Cell culture and treatments The mouse septal cell line SN56 (generously provided by B.H.Wainer,UniversityofChicago,IL,USA),isa product of the fusion of septal neurons from postnatal day 21 mice with murine neuroblastoma N18TG2. The SN56T17 cell line was obtained by stable transfection of SN56 cells with rat trkA cDNA [9]. Both cell lines were maintained at 37 °C in an atmosphere of 95% air, 5% CO 2 in Dulbec co's modi®ed Eagle's medium (DMEM) containing 1 m M pyruvate and 10% fetal bovine serum. LA-N-2 cells (a generous gift from R. Seeger, University of California, LA, USA) were grown in Leibowitz L-15 medium containing 10% fetal bovine serum at 37 °Cin an atmosphere of 100% air. During treatment periods with test compounds, cells were grown in DMEM containing N2 supplement [17] instead of fetal bovine serum, and the medium was changed every 24 h. Acetylcholine measurements The intracellular ACh content was measured by HP LC as described in Berse et al. [9], except 50 l M neostigmine was used in the physiological salt solution. Western blotting and immunoprecipitation For Western analysis, cell extracts were prepared by adding lysis buffer (50 m M Tris pH 7.5, 150 m M NaCl, 1% Nonidet NP-40, 10% glycerol, 2 m M 4-(2-aminoethyl)- benzenesulfonyl ¯uoride, 1 lgámL )1 leupeptin, 2 lgámL )1 aprotinin, 2 lgámL )1 pepstatin) to a 60-mm culture dish, scraping the cells into a tube, mixing, incubating for 15 min on ice, and brie¯y centrifuging to clear. The extracts were normalized for total protein content and separated by SDS/ PAGE. After blotting, membranes were blocked u sing 5% nonfat milk in a NaCl/P i /Tween solution pH 7.4 (contain- ing 140 m M NaCl, 3 m M KCl, 10 m M Na 2 HPO 4 ,2m M KH 2 PO 4 , 0.15% Tween 20). The primary antibodies were used at the dilutions recommended by the manufacturers. The antibody±antigen complexes were detected with anti- (rabbit IgG) Ig peroxidase conjugates and visualized using the chemiluminescence method and Kodak X-Omat AR ®lm. Band intensities were quanti®ed with a d ensitometric scanner and NIH IMAGE or Molecular Dynamics IMAGE- QUANT NT software. For immunoprecipitation, cell lysates were prepared as described above. Cell extracts were incubated overnight with either polyclonal anti-ChAT Ig AB144P (2 lgper 500 lg of extract) or polyclonal anti-Stat3 Ig (1.5 lgper 500 lg of extract) and protein A Sepharose (3.5 mg per sample) in a rotary mixer at 4 °C. Samples were brie¯y centrifuged and washed three times. The pellet was resus- pended in lysis buffer and proteins were electrophoretically separated by SDS/PAGE. Reporter gene assay The ChAT±luciferase reporter construct has been described previously [18]. Brie¯y, a 4.8-kb XhoI±HindIII portion of the ChAT promoter region was excised from a 6.4-kb clone of mouse genomic DNA (GenBank accession no. AF019045) [19,20], kindly provided by J. Naciff (University of Cincinnati College of Medicine, Cincinnati, OH, USA). The 4.8-kb fragment stretches from position )4817 (relative to the ChAT translation start) to position +46, and contains the proximal N and M promoters of the ChAT gene. It was inserted between the XhoIandHindIII sites upstream of the luciferase coding region of the vector pGL3- Basic. For cotransfection studies, the luciferase construct was used together with the pUSE-based expression plasmid containing rat MEK1 cDNA under the control of the CMV promoter. MEK1 cDNA in this construct carries mutations substituting two regulatory serine residues with aspartic acid, which results in constitutive kinase activity. SN56T17 cells were transfected using LipofectAMINE in serum-free DMEM. Three to four hours after transfection, the medium was replaced with DMEM containing 10% fetal bovine serum and the test compounds. The cells were treated for 48 h (with one medium change) and lucife rase activity was measured with the Luciferase Assay System, according to Ó FEBS 2002 MEK1 pathway inhibits cholinergic gene expression (Eur. J. Biochem. 269) 851 the manufacturer's protocol. Light intensity was measured in the Lumat LB 9501 semiautomatic luminometer (EG & G Berthold, Bad Wildbad, Germany), and expressed in arbitrary units. Statistical analysis One- or two-way analysis of variance, as appropriate, and Tukey's test were performed with the help of SYSTAT statistical software (SPSS Inc., Chicago, IL, USA) using Macintosh computers. RESULTS To address the relationship between MAPK activation and ACh synthesis, we used PD98059, a speci®c MEK1 inhibitor [21]. It has been previously shown that PD98059 decreases ERK1 (p44) and ERK2 (p42) activation (mea- sured as the degree of threonine/tyrosine phosphorylation) in a variety of cell lines following short incubations with the inhibitor [21,22]. First we tested the inhibitor's effectiveness at blocking basal and CNTF-enhanced MAPK phospho- rylation in our model septal cell line SN56T17. We probed Western blots with an antibody that recognizes the ERK1 and ERK2 only when phosphorylated on Tyr and Thr residues. The basal level of phosphorylated ERK1 and ERK2 in SN56T17 cells is easily detectable with this antibody. Incubation with CNTF fo r 15 min increased the level of phosp horylate d ERK1/2 by 40% (Fig. 1A). In similar experiments, we determined that treatment with CNTF for time periods up to 30 min increased ERK1/2 phosphorylation in these cells up to twofold (data not shown). Pretreatment with 10 l M PD98059 for 15 min almost completely obliterated MAPK phosphorylation in both CNTF-treated and untreated cells (Fig. 1A). Our results are in agreement with those of Bartoe & Nathanson, who also observed a high level of basal MAPK activation in SN56 cells that could be further stimulated about twofold by a related cytokine, leukemia inhibitory factor (LIF) [23]. Both basal and LIF-stimulated MAPK activity could be eliminated by PD98059 [23]. We then investigated whether th e downregulation of MAPK phosphorylation was maintained following an extended treatment with PD98059. When SN56T17 cells were cultured in the presence of the inhibitor for 48 h, a decrease of up to 70% in MAPK phosphorylation level was visualized by immunoblotting (Figs 1 B,C). After protracted serum withdrawal in a nonsupplemented medium, the basal MAPK phosphorylation level was greatly reduced and a 30- min treatment with CNTF resulted in an eightfold induction of MAPK phosphorylation, which was almost completely prevented by pretreatment w ith the in hibitor (Fig. 1B). In cells grown in N2-supplemented medium in the presence of CNTF for 48 h, there was no signi®cant difference in MAPK phosphorylation between control and CNTF- treated cells, and PD98059 again decreased MAPK activa- tion by 50±70%. Factors present in serum and insulin in the N2-supplemented medium could be responsible for the residual MAPK activation, since MAPK phosphorylation is completely eliminated in these cells following a prolonged incubation with PD98059 in a nonsupplemented medium (data not shown). In order to investigate the role of MEK1/MAPK pathway in ACh synthesis, we ®rst measured the effect of inhibiting this pathway on the intracellular ACh level. We examined the SN56T17 cells and their parental cell line SN56, as well as the human neuroblastoma cell line LA-N-2. In order to rule out unspeci®c effects of PD98059, we also tested another noncompetitive MEK1 inhibitor, U0126. The two compounds are structurally unrelated, although it remains unclear if they share a common binding site on MEK1 [24]. Both inhibitors are highly speci®c to the MEK1/2/ERK pathway, as they do not affect other MAPK pathways MEKK1-3, MKK4/JNK, or MKK6/p38, or numerous other protein kinases tested [22,24,25]. In each cell line, there was a statistically signi®cant effect of both MEK1 inhibitors on ACh production as compared to con- trols, and both inhibitors were equally effective (Fig. 2A). The effect of the inhibitors on LA-N-2 cells was smaller than on the mouse cell lines, probably because of lower basal MAPK activity in LA-N-2 cells as compare d to SN56 cells (data not shown). A time-course study revealed that 10 l M A C PD PD/CNTF CNTF p44 p42 B Control Control CNTF PD PD/CNTF p44 p42 p44 p42 PD CNTF − + − − + + 0’ 0’15’ 15’ 5’ 5’ Fig. 1. Eect of PD98059 on MAP k inase phosphorylation in SN56T17 cells. (A) The cells were serum-sta rved for 3 h, pretreated with 10 l M PD98059 for 1 5 min, and then treated with 20 ngámL )1 CNTF for the indicated periods of time. Cell lysates (40 lg per lane) were analyzed by Western blotting with an antibody speci®c to doubly ph osphorylated forms of ERK1 (p44) and ERK2 (p42). The results are representative of two experiments. (B) The cells were grown for 48 h either in the presence or in the absence of 10 l M PD98059 (PD), serum-starved for 6 h, and then treated with 20 ngámL )1 CNTF for 30 min Western blotting was performed as d escribed in (A). T he results are representative of two experiments. (C) The cells were grown for 48 h either in the presence or in the absence of 10 l M PD98059 (PD), 20 ngámL )1 CNTF, or both. Western blotting was performed as described in ( A). The results are representative of four experiments. 852 T. Mellott et al. (Eur. J. Biochem. 269) Ó FEBS 2002 PD98059 produced a twofold increase in ACh levels in SN56T17 cells and that the full effect could be observed after 48 h of treatment (Fig. 2B). Following a 48-h treat- ment, PD98059 increased the intracellular levels of ACh in SN56T17 cells in a concentration-dependent and saturable manner ( Fig. 2C), with an EC 50 of 3.7 l M , comparable to the reported IC 50 of 4 l M on MEK1 activity in vitro [21,22]. Thus, these data suggest that the increase in intracellular ACh levels is a result of down-regulating MEK1-mediated MAPK activation. Because our previous studies have shown that CNTF upregulates the cholinergic phenotype and that NGF strongly activates MAPK in SN56T17 cells while interfering with the actions of CNTF [5,9], we examined the effect of PD98059 and U0126 on the ability of CNTF to increase ACh synthesis. The results are presented in Fig. 3A. Consistent with our previous ®ndings, CNTF alone signi- ®cantly increased intracellular ACh level by more than threefold. PD98059 increased ACh level by twofold, while the combined treatment with CNTF and PD98059 increased ACh content by sixfold. Substituting U0126 for PD98059 resulted in a similar effect on CNTF-stimulated ACh production. Similar results were obtained with SN56 and L A-N-2 c ells (data not shown). This i ncrease in intracellular ACh levels was the result of an increased amount of ChAT protein, as demonstrated by immunopre- cipitation followed by Western blotting with two polyclonal anti-ChAT Ig (Fig. 3B). We measured ChAT protein in SN56T17 cells which had been treated w ith PD98059 (10 l M ), CNTF (20 ngámL )1 ), or the combination for 48 h. Both PD98059 and CNTF increased ChAT protein levels, and the combined treatment with PD98059 and CNTF had the greatest effect on the level of C hAT protein (Fig. 2B). In summary, both PD98059 and CNTF are able to increase intracellular ACh levels via increasing ChAT p rotein levels. Furthermore, even though the MAPK signaling is required for CNTF-mediated activation of certain genes i n the CNS [26], in this case, blocking MAPK activation enhances the CNTF effect on ACh p roduction. C Acetylcholine, pmol/mg protein 0510 15 20 Concentration of PD98059, µ M 1200 1000 800 600 0 EC 50 = 3.7 µM B 0 50 100 150 200 0 2 4 7 Time, hours Acetylcholine, % Control A 0 50 100 150 200 250 Control PD U Control PD U Control PD U Acetylcholine, % Control Fig. 2. MEK1 inhibitors increase ACh production in cholinergic cell lines. (A) SN56T17, SN56, and LA-N-2 cells were grown for 48 h in the presence or absence of 10 l M PD98059 (PD) or 10 l M U0126 (U). Intracellular A Ch was measured in cell extrac ts as described in Experimental proce dures. The results are presented as means  SEM o f two experiments, each performed in quadruplicate. Acetylcholine levels in cells treated with PD98059 and U0126 were statistically dierent from control values for each cell line (SN56T17 P < 0.0 01 and P < 0.005, respectively; SN56 P < 0.001 and P < 0.001, respectively; and LA-N-2 P <0.01andP < 0 .01, respectively), and were not statistically dierent from each other (SN6T17 P  0.834; SN56 P  0.765; and LA-N-2 P  0.977); Tukey's test. (B) Time Course: the SN56T17 cells were grown for 24, 48 or 72 h in the presence or absence of the MEK inhibitor PD98059 (10 l M ). The results are presented as means  SEM and are representative of three experiments. (C) Concentration-response curve of PD98059: the SN56T17 cells were treated for 48 h in the presence of various PD98059 concentrations (0, 1, 3, 5, 10, 15, and 20 l M ). The dat a are representative of three experiments. A rectangular hyperbola was ®tted to the data, permitting the estimation of the median eective concentration (EC 50  3.7 l M ). A Control PD CNTF PD/CNTF UU/CNTF 0 200 400 600 800 Acetylcholine, % Control ChAT CNTF PD/CNTF Control PD B Fig. 3. Up-regulation of the cholinergic phenotype by CNTF and MEK inhibitors. The SN56T17 cells were grown for 48 h in the presence or absence of 10 l M PD98059 (PD), 10 l M U0126 (U), 20 ngámL )1 CNTF, or combinations. (A) MEK inhibitors and CNTF increase intracellular ACh levels. Intracellular ACh was measured in cell extracts as described in the Experimental procedures. The results are presented as means  S EM of 12 samples (four experiments, each performed in t riplicate). Two-way analysis of variance revealed a sig- ni®cant eect of PD98059 (P < 0.001), U0126 (P < 0.005), and CNTF (P < 0.001), and a signi®cant interaction between PD98059 and CNTF (P < 0.05) and between U0126 and CNTF (P <0.01). (B) ChAT protein levels are increased following PD98059 and CNTF treatment. Cell lysates (250 lg per sample) were immunoprecipitated with anti-ChAT Ig AB144P, and then analyzed by Western blotting using anti-ChAT Ig AB143. Similar results were obtained in four additional experiments. Ó FEBS 2002 MEK1 pathway inhibits cholinergic gene expression (Eur. J. Biochem. 269) 853 In order to examine w hether the effect of MAPK on the cholinergic phenotype involves changes in gene expression, we employed transient transfection assays to measure the activation of the ChAT promoter by PD98059 and CNTF. We used a ChAT±luciferase construct containing the proximal mouse ChAT promoter region linked to a luciferase reporter gene as described previously [18]. Following transfection with the reporter construct, SN56T17 cells were treated with PD98059, CNTF, or both, for 48 h and assayed for luciferase reporter activity (Fig. 4). Luciferase activity increased  fourfold following CNTF treatment and twofold with PD98059. Substituting U0126 for PD98059 resulted in a similar effect on luciferase activity (data not shown). Moreover, there was a substantial synergistic increase i n luciferase activity (by 20-fold) when both PD98059 and CNTF were employed (Fig. 4B). This result suggested that the MEK1/MAPK pathway interferes with the CNTF pathway and that inactivation of the former is necessary to observe the full CNTF effect on the ChAT promoter. We tested this hypothesis directly in the cotrans- fection experiments using the expression vector pUSE with a cDNA insert producing constitutively active MEK1. When this plasmid was cotransfected into SN56T17 cells together with the luciferase reporter construct, the effect of CNTF on the luciferase expression was completely abro- gated, whereas c otransfection with the empty pUSE vector did not inhibit C NTF-induced luciferase expression (Fig. 4 C). Thus, we conclude that the MEK1 pathway interferes with the CNTF action in cholinergic cells. The m ain signaling pathway utilized by CNTF involves activation of transcription factors of the STAT (signal transducers and activators of transcription) family. Upon CNTF binding, the CNTF receptor/Janus protein tyrosine kinase (Jak) 2 complex preferentially phosphorylates Stat3 at residue Tyr705 [27]. We veri®ed the involvement of the Jak2/Stat3 pathway in regulating ACh synthesis using the Jak2 inhibitor tyrphostin AG490. This compound has been shown to be selective for JAK family kinases [28]. Although AG490 at higher concentrations inhibits Jak3 in lympho- cytes [29], its primary speci®c target in most cell types is the Jak2/Stat3 pathway [28,30,31]. We used the inhibitor at 10 l M , the concentration at which it is speci®c for JAK kinases in vitro and signi®cantly reduces the DNA binding activity of Stat3 in vivo [32]. First, we tested the ability of the inhibitor to block the Jak2/Stat3 pathway in SN56T17 cells. Following a 15-min exposure to 10 l M AG490, CNTF- induced Tyr705 phosphorylation of Stat3 was signi®cantly reduced (Fig. 5A). After the cells were treated for 48 h with the inhibitor, either alone or in combination with CNTF, intracellular ACh levels were measured as described above (Fig. 5 B). Tyrphostin AG490 alone had no effect on basal ACh synthesis; however, it reduced CNTF-stimulated ACh A 13 more ChAT coding exons first ChAT coding exon polyA site ATG (Met) ATG (Met) V1 V2 R promoter ChAT N promoter ChAT M promoter Luciferase reporter 4.8-kb ChAT genomic DNA pGL3Basic VAChT ORF HinDIII XhoI 0 5 10 15 20 PD/CNTF CNTF PD Control B ChAT-Luc + MEK1 ChAT-Luc + Vector ChAT-Luc Luciferase Activity, arbitrary units Luciferase Activity, arbitrary units 00.5 11.5 22.5 3 Control CNTF C Fig. 4. MEK1 interferes with the up-regulation of ChAT promoter activity by CNTF in a reporter gene assay. (A) ChAT±luciferase reporter construct. The diagram illustrates the lo calization of the promoter region used in th e construct with in the cholinergic locus. The positions of the distal R promoter, as we ll as V AChT-spe ci®c V1 and V2 promoters and ChAT-speci®c N and M p romoters are marked with horizontal arrows. ORF, open reading frame. (B) SN 56T17 cells were transfected with the ChAT-luciferase rep orter construct and treated with the compounds indicated for 48 h . Luciferase activity was measured in cell extracts as described in Experimental procedures. The results are presented as means  SEM and are representative of three experiments. Two-way analysis of variance revealed a signi®cant eect of each individual treatment (P < 0.001) and a signi®cant interactio n (P < 0.001). (C) SN56T17 cells were transfected with the ChAT±luciferase reporter construct (ChAT±Luc) alone or in combination with the pUSE vector (Vector) or with the constitutive MEK1 expression construct (MEK1). The cells were treated w ith 20 ngámL )1 CNTF for 48 h; lucifer ase activity was measured as above and normalized to untreated controls. The results are presented as means  SEM and are repr esentative of four ex perim ents. Two-way analysis of variance an d Tukey's test rev ealed a signi®cant e ect of CNTF treatme nt i n cells transfected with the luciferase construct alone (P < 0.005) and in combination with pUSE (P < 0.05). The CNTF eects in those two groups of transfectants were not statistically dierent from each other (P  0.97). There was n o eect of CNTF in cells transfec ted with the luciferase construct together with the MEK1 expression vector (P  0.99). 854 T. Mellott et al. (Eur. J. Biochem. 269) Ó FEBS 2002 synthesis to approximately control levels. Therefore, the results suggest that CNTF-stimulated ACh production is mediated by the activation of Stat3 by Jak2. We then examined how blocking the MEK1/MAPK pathway affects the phosphorylation state of Stat3. SN56T17 cells were cultured in the presence or absence of 10 l M PD98059 and then treated with CNTF for 20 min. Stat3 phosphorylation on Tyr705 and Ser727 was visualized by immunoblotting. Shorter (up to 4 h) pretreatment with PD98059 did not alter the level of Stat3 Tyr705 phospho- rylation evoked by CNTF, while longer (up t o 48 h ) exposure to the inhibitor resulted in an increase in Tyr705 phosphorylation to a variable degree. In some experiments, the level of Tyr705 phosphorylation of Stat3 in PD98059- pretreated cells was higher than in control cells, however, statistical analysis revealed that the increase was not signi®cant (Fig. 6). The CNTF treatment resulted in a twofold to threefold increase in Stat3 Ser727 phosphoryl- ation. Pre-treatment with PD98059 reproducibly reduced Stat3 Ser727 phosphorylation to control levels. MAPK phosphorylation levels in the same cells, in the presence and absence of PD98059 and CNTF, mirrored the Stat3 serine phosphorylation status and were consistent with previous ®ndings (Fig. 1). CNTF increased MAPK phosphorylation by 90%, and this phosphorylation was almost completely abrogated by PD98059. Thus, it appears that, in SN56T17 cells, the MEK1/MAPK pathway mediates CNTF-evoked Stat3 phosphorylation on the Ser727 residue, whereas tyrosine phosphorylation of Stat3 is at least partially MAPK-independent. DISCUSSION We demonstrate here that exposure of cholinergic SN56T17 cells to MEK1 inhibitors, PD98059 and U0126, induces an increase in ChAT gene expression, resulting in elevated ChAT prote in levels and ACh production. The effect of MEK1 inhibitors on ACh synthesis is not an exclusive property of the SN56T17 cell line, as it can also be observed in the parental murine cell line SN56, as well as, albeit to a lesser extent, in the human cholinergic LA-N-2 cells. The data suggest that PD98059 and U0126 up-regula te ACh synthesis by inhibiting MAPK activation. This ®nding is surprising, because the cholinergic phenotype is also known to be upregulated by several stimuli (notably NGF), that utilize the MAPK signaling pathway. In our model cell line, PD98059 increases both basal and CNTF-stimulated expression of the luciferase reporter under t he control of the murine ChAT promoter. The results o f the reporter gene experiments suggest a negative regulatory role for MAPK in cholinergic gene expression. It is worth noting that Espinos et al. observed no effect of PD98059 on the human ChAT promoter activity after 6 h of treatment in human neuroepithelioma CHP126 cells [33]. However, we also noticed no effect of PD98059 on at 6 h (data not shown), suggesting that prolonged inhibition of MEK1 is neede to observe changes in cholinergic gene expression. Differences in the cellular context of the cell lines and/or the exact ChAT promoter r egion, as well as species of origin of the gene used in the studies, may also have contributed to the a pparent discrepancy between our ®ndings and those of Espinos et al. We observed a synergistic effect of the combination of PD98059 and CNTF on ChAT promoter-driven reporter expression (but not on ACh or ChAT activity). The synergy on the gene expression level indicates that the two treatments do not work independently, but affect cross- talking pathways. It appears that the MEK1/MAPK pathway blocks the full CNTF effect on the ChAT promoter. This conclusion is supported by the cotransfec- tion experiments, where we demonstrated that constitutively active MEK1 prevents the induction of the ChAT-luciferase construct by CNTF. This action of MEK1 provides an interpretation of our previous study, where we have shown that while both CNTF and NGF separately enhance the cholinergic phenotype, NGF, paradoxically, interferes with the CNTF effect [9]. We also have found that NGF evokes a rapid and profound increase in ERK1/2 phosphorylation in SN56T17 cells, whereas CNTF-evoked activation of MAPK is delayed and weaker ([9]; this study). Hence, the attenuation of the CNTF e ffect caused by NGF could be A Stat3 P Tyr Stat3 CNTF 0' 20' 20'0' Tyr AG 4 90 − − ++ B Acetylcholine, % Control Control CNTF Tyr AG 490 Tyr/CNTF 0 50 100 150 200 250 300 Fig. 5. Inhibition of CNTF-induced e ects by a Jak2 inhibitor. (A) Tyrphostin AG490, a Jak2 inhibitor, reduces CNTF-induced tyrosine phosphorylation of Stat3. The SN56T17 cells were serum- starved for 3 h, pretreated with 10 l M AG490 for 15 min, and then treated with 20 ngámL )1 CNTFfor20min.Celllysates(500lgper sample) were immunoprecipitated with a Stat3 antibody and then analyzed by Western blotting using an antibody speci®c for phos- phoTyr705 Stat3 or the Stat3 an tibody. (B) Tyrphostin AG490 blo cks the CNTF-induced eect on intracellular ACh. The SN56T17 cells were grown for 48 h in the presence or absence of 10 l M AG490 (TyrAG490), 20 ngámL )1 CNTF, or a combination o f the two. Intracellular ACh was measured in cell extracts as described in Experimental procedures. Th e results are presented as mean s  SEM of six samples (two experiments, each performed in triplicate). Two- way analysis of variance and Tukey's test revealed a signi®cant eect of CNTF (P < 0.001), no eect of AG490 alone (P  0.6359), and a signi®cant dierence between the combination and CNTF alone (P < 0.01). Ó FEBS 2002 MEK1 pathway inhibits cholinergic gene expression (Eur. J. Biochem. 269) 855 explained by the inhibitory action of NGF-activated MEK1/MAPK pathway on ChAT gene expression. Simi- larly, in a recent study, Bartoe & Nathanson observed that inhibiting MAPK in SN56 cells stimulated the expression of LIF-induced vasoactive intestinal peptide (VIP) promoter by twofold [23]. In that study, treatment with PD98059 was performed 1 h before a 24-h stimulation with LIF. That particular treatment strategy was not suf®cient to increase LIF-induced ChAT±luciferase reporter. The authors con- cluded t hat MAPK activation is involved in negatively regulating LIF-mediated VIP indu ction, but not ChAT induction. However, the ChAT±luciferase construct employed in that study contains a smaller fragment of murine ChAT promoter than the one used here; it responds relatively weakly to LIF and CNTF [34] and it does not respond to overexp ression of LIF receptor [23]. Thus, the ChAT promoter region tested by Bartoe & Nathanson may not contain all the sequences needed for the attenuation o f CNTF-mediated gene expression by MAPK. Furthermore, although CNTF and LIF share the transmembrane receptor subunits, possible differences between signaling by these two cytokines cannot be excluded. The presence of potential STAT binding sites in the cholinergic locus promoter region indicates that the effect of CNTF on ACh production in cholinergic cells is likely mediated by the transcription factor Stat3. In SN56 cells and their derivatives, the Stat3 homodimer is the prevalen t DNA binding complex activated by CNTF [35,36]. We demonstrated that Stat3 is rapidly tyrosine-phosphorylated (Figs 5 and 6) and translocated to the nucleus upon CNTF treatment of SN56T17 cells [9] and in primary spinal neurons (B. Berse, unpublished data). As CNTF-induced MAPK activation is delayed relative to the rapid Stat3 response, it could serve as a negative feedback mechanism to downregulate the effects of CNTF mediated by Stat3. The mechanism of this downregulation remains to be investi- gated. This study shows that Stat3 tyrosine phosphorylation is not signi®cantly affected by PD98059, although there appears t o be a limited increase in Stat3 Tyr705 phospho- rylation after prolonged exposure t o the inhibitor. Curiously, i n SN56T17 cells, the phosphorylation of Stat3 on Ser727 in response to CNTF appears to be MAPK- dependent. Recent studies on STATs point to the regulatory role of serine/threonine phosphorylation in addition to tyrosine phosphorylation (for a review see [37]). The Ser727 residue of Stat3 was shown to be a target for multiple kinase pathways including MEK1/MAPK [38±43]. There are con¯icting conclusions about the actual biological conse- quences of serine phosphorylation of STATs. The majority of reports demonstrated that serine phosphorylation can enhance Stat3 DNA binding and/or transcriptional activity [37]. Other studies point to a negative role for Stat3 serine phosphorylation, through both MAPK-dependent and -independent mechanisms [38,43±46]. Although it has been documented that ERKs can directly phosphorylate Stat3 on Ser727, the MAPK-mediated inhibition of the IL-6 signal- ing pathway occurs upstream o f S tat3, as it was unaffected by Ser727 to Ala mutation [47]. It is becoming clear that the relationship between STAT- and MAPK-mediated path- ways is complex and the ®nal outcome depends on the A Stat3 P Tyr Stat3 P MAPK P Ser Stat3 20' 0' 20' − ++ CNTF 0' PD − B Stat3 Ser727 Phosphorylation, % Control 0 50 100 150 200 250 300 Control PD 0' CNTF 20' CNTF MAPK Phosphorylation, % Control 0 50 100 150 200 250 Control PD 0' CNTF 20' CNTF Stat3 Tyr705 Phosphorylation, % Control 1000 1500 2000 2500 3000 3500 0 500 Control PD 0' CNTF 20' CNTF Fig. 6. Reduction of Stat3 Ser727 phosphorylation by PD98059. (A) The cells were grown in the presence or absence of 10 l M PD98059 for 48 h and then were serum-starved for 3 h. The cells were then treated with 20 ngámL )1 CNTF for 20 min Cell lysates (500 lg per sample) were immunoprecipitated with a Stat3 antibody and then analyzed by Weste rn blotting using an antibody speci®c for phosphoSe r727-Stat3, phos- phoTyr705-Stat3, or an antibody that recognizes Stat3 regardless of its phosphorylation state. Western blotting with an antibody speci®c for the phosphorylated forms o f MAP kinase was performed on the crude cell lysates (40 lg per lane). Sim ilar results were ob tained in 2±4 additional experiments. (B) Ban d intensities from th e blots shown in (A) were q uanti®ed by densitometry an d the results are presented as mean s  SEM of 3±5 experiments, each performed in duplicate. PhosphoSer727-Stat3 and phosphoTyr705-Stat3 levels were normalized to the total Stat3 level. In the absence of PD98059, phosphoSer727-Stat3, phosphoTyr705-Stat3 and phosphorylated MAPK levels in cells treated with CNTF were statistically dierent from control values (P <0.01,p <0.001,andp < 0.005, respectively). In CNTF-treated cells, PD98059 had a statistically signi®cant eect on phosphoSer727-Stat3 and phosphorylated MAPK levels (P <0.01andp < 0.005, respectively), but not on phosphoTyr705-Stat3 level (P  0.31). 856 T. Mellott et al. (Eur. J. Biochem. 269) Ó FEBS 2002 speci®c genes being regulated, the cell type, and possibly on the cross-talk with other pathways (e.g. other serine kinases). ACKNOWLEDGEMENTS The authors thank Dr Jorge M. Naci for the gift of the mouse genomic DNA clone containing ChAT proximal promoters. This work was supported by National Institutes of Health Grant AG09525 (to J. K. B), National Science Foundation Grant IBN-9907572 (to J. K. B and B. B.), a grant from the W hitehall Foundation (to B. B.), Alzheimer's Association grant IIRG-00±2073 (to I. L C) and National Institutes of HealthTraining Grant AG00115 (to T. M). REFERENCES 1. Wu, D. & Hersh, L.B. (1994) Choline acetyltransferase: celebrat- ing its ®ftieth year. J. Neurochem. 62, 1653±1663. 2. Eiden, L.E. (1998) The cholinergic gene locus. J. Neuroc hem . 70, 2227±2240. 3. Bejanin, S., Habert, E., Berrard, S., Edwards, J.B., Loeer, J.P. & Mallet, J. (1992) Promoter elements of the rat choline acetyl- transferase gene allowing nerve growth factor inducibility in transfected primary cultured cells. J. Neurochem. 58, 1580±1583. 4. Cervini, R., Berrard, S., Bejanin, S. & Mallet, J. (1994) Regulation by CDF/LIF and retinoic acid of multiple ChAT mRNAs pro- duced from distinct promoters. Neuroreport. 5, 1346±1348. 5. Berse, B. & Blusztajn, J.K. (1995) Coordinated up-regulation of choline acetyltransferase and vesicular acetylcholine transporter gene expression by the retinoic acid receptor alpha, cAMP, and leukemia inhibitory factor/ciliary neurotrophic factor signaling pathways in a murine septal cell line. J. Biol. Chem. 270, 22101± 22104. 6. Misawa, H., Takahashi, R. & Deguchi, T. (1995) Coordinate expression of vesicular acetylcholine transporter and choline acetyltransferase in sympathetic superior cervical neurones. Neuroreport. 6, 965±968. 7. Tian, X., Sun, X. & Suszkiw, J.B. (1996) Developmental age- dependent upregulation of choline acetyltransferase and vesicular acetylcholine transporte r m RNA e xpression in neonatal rat sep- tum by nerve growth factor. Neurosci. Lett. 209, 134±136. 8. Pongrac, J.L. & Rylett, R .J. (1998) Molecular mechanisms regu- lating NGF-mediated enhancement of cholinergic neuronal phe- notype: c-fos trans-activation of the choline acetyltransferase gene. J. Mol. Neurosci. 11, 79±93. 9. Berse, B ., Lopez-Coviella, I. & Blusztajn, J.K. (1999) Activation of TrkA by nerve growth factor upregulates expression of the cho- linergic gene locus but attenuates the response to ciliary neuro- trophic growth factor. Biochem. J. 342, 301±308. 10. Oosawa, H., Fujii, T. & Kawashima, K. (1999) Nerve growth factor increases the synthesis and release of acetylcholine and the expression of vesicular acetylcholine transporter in primary cul- tured rat embryonic septal cells. J. Neurosci. Res. 57, 381±387. 11. Quirin-Stricker, C., Mauvais, C. & Schmitt, M. (1997) Tran- scriptional activation of human choline acetyltransferase by AP2- and NGF-induced factors. Brain Res. Mol. Brain Res. 49, 165±174. 12. Shimojo,M.,Wu,D.&Hersh,L.B.(1998)Thecholinergicgene locus is coordinately regulated by protein kinase A II in PC12 cells. J. Neurochem. 71 , 1118±1126. 13. Robinson, M.J. & Cobb, M.H. (1997) Mitogen-activated protein kinase pathways. Curr. Opin. Cell Biol. 9, 180±186. 14. Marshall, C.J. (1994) MAP kinase kinase kinase, MAP kinase kinase and MAP kinase. Curr. Opin. Genet. Dev. 4, 82±89. 15. Hill, C.S. & Treisman, R. (1995) Transcriptional regulation by extracellular signals: mechanisms and speci®city. Cell. 80, 199±211. 16. Garrington, T.P. & Johnson, G.L. (1999) Organization and regu- lation of mitogen-activated protein kinase signaling pathways. Curr. Opin. Cell Biol. 11, 211±218. 17. Bottenstein, J.E. & Sato, G.H. (1979) Growth of a rat neuro- blastoma cell lin e in serum -free supplem ented medium. Proc. Natl Acad.Sci.USA76, 514±517. 18. Lopez-Coviella, I., Berse, B., Krauss, R., Thies, R.S. & Blusztajn, J.K. (2000) Induction and maintenance of the neuronal cholinergic phenotype in the central nervous system by BMP-9. Science 289, 313±316. 19. Naci, J.M., Misawa, H. & Dedman, J.R. (1997) Molecular characterization of the mouse vesicular ac etylcholine transporter gene. Neuroreport 8, 3467±3473. 20. Naci, J.M., Behbehani, M.M., Misawa, H. & Dedman, J.R. (1999) Identi®cation and transgenic analysis of a murine promoter that targets cholinergic neuron expression. J. Neurochem. 72, 17±28. 21. Dudley,D.T.,Pang,L.,Decker,S.J.,Bridges,A.J.&Saltiel,A.R. (1995) A synthetic inhibitor of the mitogen-activated protein kinase cascade. Proc. Natl Acad. Sci. USA 92, 7686±7689. 22. Alessi, D.R., Cuenda, A., Cohen, P., Dudley, D.T. & Saltiel, A.R. (1995) PD 098059 is a speci®c inhibitor of the activation of mitogen- activated protein kinase kinase in vitro and in vivo. J. Biol. Chem. 270, 27489±27494. 23. Bartoe, J.L. & Nathanson, N.M. (2000) Dierential regulation of leukemia inhibitory facto r-stimulated neuronal gene expression by protein phosphatases SHP-1 and SHP-2 through mitogen-acti- vated protein kinase-dependent and -independent pathways. J. Neurochem. 74 , 2021±2032. 24. Favata, M.F., Horiuchi, K.Y., Manos, E.J., Daulerio, A.J., Stradley, D .A., Feeser, W.S., Van Dyk, D.E., Pitts, W.J., Earl, R.A., Hobbs, F., Copeland, R .A., Magolda, R.L., Scherle, P.A. & Trzaskos, J.M. (1998) Identi®cation of a novel inhibitor of mito- gen-activated protein kinase kinase. J. Biol. Chem . 273, 18623± 18632. 25. New England Biolabs (2000±01) Cell Signaling Technology Cata- log. New England Biolabs, MA, USA. 26. Rajan, P. & McKay, R.D. (1998) Multiple routes to astrocytic dierentiation in the CNS. J. Neurosci. 18, 3620±3629. 27. Turnley, A.M. & Bartlett, P.F. (2000) Cytokines that signal through the leukemia inhibitory factor receptor-beta complex in the nervous system. J. Neurochem. 74, 889±899. 28. Meydan, N., Grunberger, T., Dadi, H., Shahar, M., Arpaia, E., Lapidot,Z.,Leeder,J.S.,Freedman,M.,Cohen,A.,Gazit,A., Levitzki, A. & Roifman, C.M. (1996) Inhibition of acute lym- phoblastic leukaemia by a Jak-2 inhibitor. Nature 379, 645±648. 29. Wang, L.H., Kirken, R.A., Erwin, R.A., Yu, C.R. & Farrar, W.L. (1999) JAK3, STAT, and MAPK signaling pathways as novel molecular targets for the tyrphostin AG-490 regulation of IL -2- mediated T cell response. J. Immunol. 162, 3897±3904. 30. Marrero, M.B., Schieer, B., Li, B., Sun, J., Harp, J.B. & Ling, B.N. (1997) Role of Janus kinase/signal transducer and activator of transcription and mitogen-activated protein kinase cascades in angiotensin II- and platelet-derived growth factor-induced vascu- lar smooth muscle cell proliferation. J. Biol. Chem. 272, 24684± 24690. 31. Abe, J. & Berk, B .C. (1999) Fyn and JAK2 mediate Ras activation by reactive oxygen species. J. Biol. Chem. 274, 21003±21010. 32. Zhang, Y., Turkson, J., Carter-Su, C., Smithgall, T., Levitzki, A., Kraker, A., Krolewski, J.J., Medveczky, P. & Jove, R. (2000) Activation of Stat3 in v-Src-transformed Fibroblasts Requires Cooperation of Jak1 Kinase Activity. J. Biol. Chem. 275, 24935± 24944. 33. Espinos, E., Le Van Thai, A., Pomies, C. & Weber, M.J. (1999) Cooperation b etween phosphorylation and acetylation processes in transcriptional control. Mol. Cell. Biol. 19, 3474±3484. 34. Johnson, J.A. & Nathanson, N.M. (1994) Dierential require- ments for p21ras and protein kinase C in the regulation of Ó FEBS 2002 MEK1 pathway inhibits cholinergic gene expression (Eur. J. Biochem. 269) 857 neuronal gene expression by nerve growth factor and neurokin es. J. Biol. Chem. 269, 18856±18863. 35. Di Marco, A., Demartis, A., Gloaguen, I., Lazzaro, D., Delmastro, P., Ciliberto, G. & Laufer, R . (2000) Leptin receptor-mediated regulation of cholinergic neurotransmitter phenotype in c ells of central n ervous system origin. Eur. J. Bio- chem. 267, 2939±2944. 36. Gloaguen, I., Costa, P., Demartis, A., Lazzaro, D., Di Marco, A., Graziani, R., Paonessa, G., Chen, F., Rosenblum, C.I., Van der Ploeg, L.H., Cortese, R., Ciliberto, G. & Laufer, R. (1997) Ciliary neurotrophic factor corrects o besity and diabetes associated with leptin de®c iency and resistance. Proc. Natl Acad. Sci. USA 94, 6456±6461. 37. Decker, T. & Kovarik, P. (2000) Serine phosphorylation of STATs. Oncogene 19, 2628±2637. 38. Chung, J., Uchida, E., Grammer, T.C. & Blenis, J. (1997) STAT3 serine phosphorylation by ERK-dependent and -independent pathways negatively mod ulates its tyrosine phosphorylation. Mol. Cell Biol. 17, 6508±6516. 39. Wen, Z. & Darnell, J.E. Jr (1997) Mapping of Stat3 serine phos- phorylation to a single residue (727) and evidence that serine phosphorylation has no in¯uence on DNA binding of Stat1 and Stat3. Nucleic Acids Res. 25, 2062±2067. 40. Wen, Z., Zhong, Z. & Darnell, J.E. Jr (1995) Maximal activation of transcription by Stat1 and Stat3 requires both tyrosine and serine phosphorylatio n. Cell 82, 241±250. 41. Ng, J. & Cantrell, D. (1997) STAT3 is a serine kinase target in T lymphocytes. Interleukin 2 and T cell antigen receptor signals converge upon serine 727. J. Biol. Chem. 272, 24542±24549. 42. Su, L., Rickert, R.C. & David, M. (1999) Rapid STAT phos- phorylation via the B cell receptor. Modulatory role of CD19. J. Biol. Chem. 274, 31770±31774. 43. Lim, C.P. & Cao, X. (1999) Serine Phosphorylation and Negative Regulation of Stat3 by JNK. J. Biol. Chem. 274, 31055±31061. 44. Jain, N., Zhang, T., Fong, S.L., Lim, C.P. & Cao, X. (1998) Repression of Stat3 activity by activation of mitogen-activated protein kinase (MAPK). Oncogene 17, 3157±3167. 45. Bhat, G.J., Hunt, R.A. & Baker, K.M. (1998) alpha-Thrombin inhibits signal transducers and activators of transcription 3 sig- naling by interleukin-6, leukemia inhibitory factor, and ciliary neurotrophic factor in CCL39 cells. Arch. Biochem. Biophys. 350, 307±314. 46. Woetmann, A., Nielsen, M., Christensen, S.T., Brockdor, J., Kaltoft, K., Engel, A.M., Skov, S., Brender, C., Geisler, C., Svejgaard, A., Rygaard, J., Leick, V. & Odum, N. (1999) Inhibi- tion of protein phosphatase 2A induces serine/threonine phos- phorylation, subcellular redistribution, and functional inhibition of STAT3. Proc. Natl Acad. Sci. USA 96, 10620±10625. 47. Sengupta, T.K., Talbot, E.S., Scherle , P.A. & Ivashkiv, L.B. (1998) Rapid inhibition of interleukin-6 signaling and Stat3 acti- vation mediated by mitogen-activated protein kinases. Proc. Natl Acad. Sci. USA 95, 11107±11112. 858 T. Mellott et al. (Eur. J. Biochem. 269) Ó FEBS 2002 . Mitogen-activated protein kinase kinase negatively modulates ciliary neurotrophic factor-activated choline acetyltransferase gene expression Tiffany Mellott 1 , Ignacio. Stat3. Biosynthesis o f the neurotransmitter acetylcholine (ACh) from choline and acetyl coenzyme A is catalyzed by the enzyme choline acetyltransferase (ChAT, acetyl-CoA:cho- line O-acetylotransfe. pathway inhibits the CNTF-mediated stimulation o f C hAT exp ression, possibly as a p art of a feedback mechanism. Keywords: acetylcholine; choline acetyltransferase; MAP kinase; c iliary neurotrophic