Mouse NIPK interacts with ATF4 and affects its transcriptional activity

Mouse NIPK interacts with ATF4 and affects its transcriptional activity

Available online at www.sciencedirect.com R Experimental Cell Research 286 (2003) 308 –320 www.elsevier.com/locate/yexcr Mouse NIPK interacts with ...

487KB Sizes 0 Downloads 4 Views

Available online at www.sciencedirect.com R

Experimental Cell Research 286 (2003) 308 –320

www.elsevier.com/locate/yexcr

Mouse NIPK interacts with ATF4 and affects its transcriptional activity ¨ rda and To˜nis O ¨ rdb,* Daima O a

Institute of Molecular and Cell Biology, Tartu University, Tartu, Estonia b Estonian Biocentre, Tartu, Estonia Received 30 October 2002, revised version received 9 January 2003

Abstract Neuronal cell death-inducible putative kinase (NIPK) is a protein with an unknown function encoded by a gene activated in neuronal cells in cell death-causing conditions (disruption of calcium homeostasis, trophic factor deprivation). Using the yeast two-hybrid screening of an embryonic mouse cDNA library, we identified activating transcription factor 4 (ATF4) as a protein binding to mouse (m) NIPK. The critical domain for mNIPK-binding resides in a 72 amino acid stretch near the N-terminus of ATF4, covering the second leucine zipper motif and the preceding region. mNIPK expressed as fusion protein with enhanced yellow fluorescence protein (EYFP) is localized predominantly in the nucleus, and the mNIPK–ATF4 complex can be immunoprecipitated from cells cotransfected with epitope-tagged mNIPK and ATF4 constructs. The expression of both mNIPK and ATF4 is upregulated in the neuronal cell line GT1-7 in response to disruption of calcium homeostasis by thapsigargin, but ATF4 is induced more rapidly than mNIPK. The coexpression of mNIPK inhibits ATF4 CRE-dependent transcriptional activation activity in transiently transfected cells. At the same time, ATF4 degradation rate is not increased in the cells coexpressing mNIPK, and ATF4, associated to mNIPK, is able to bind to CRE. Thus, mNIPK is a novel regulator of ATF4 transcriptional activity. © 2003 Elsevier Science (USA). All rights reserved. Keywords: Neuronal cell death; Gene expression; Transcriptional repression; Stress response

Introduction A significant fraction (up to 50%) of many types of neurons die during the development of the vertebrate nervous system, presumably as a result of the limited availability of specific neurotrophic factors produced by the target cells [1,2]. The neurons, which have formed synaptic connections with their target cells and survived the physiological cell-death period, are normally maintained for the entire life span of an individual. In the mature nervous system the inappropriate loss of neurons is associated with many acute and chronic neurological disorders, ranging from stroke to neurodegenerative diseases. While developmental neuronal death is mainly apoptotic, the cells dying via apoptosis or necrosis, or exhibiting mixed features of an apoptotic and necrotic mode of death,

* Corresponding author. Estonian Biocentre, 23 Riia St., Tartu, 51010, Estonia. Fax: ⫹372-7-420286. ¨ rd). E-mail address: [email protected] (T. O

are observed in neuropathological states [3]. Inhibitors of transcription and translation block the death of cultured primary neurons and neuronal cell lines caused by a variety of insults (trophic factor deprivation, disruption of calcium homeostasis, oxidative stress, hypoxia), suggesting that de novo RNA and protein synthesis may be necessary in the process of neuronal death [4 – 8]. A number of genes upregulated in neurons in response to induction of cell death by trophic factor withdrawal or loss of membrane depolarization have been identified, including genes encoding apoptotic proteinase caspase-3, cell-cycle regulators cyclin D1 and B, and components of AP-1 transcription factor c-Jun and c-Fos [9 –12]. Administration of inhibitors of cell cycle progression promotes survival of neuronally differentiated PC12 cells and sympthetic and cerebellar granule neurons deprived of trophic factor, suggesting that the activation of cell cycle signals, which is inappropriate in postmitotic background, is an essential component of the neuronal death program [13,14]. Similarly, microinjection of neutralizing anti-c-Jun antibodies, expression of dominant-negative c-

0014-4827/03/$ – see front matter © 2003 Elsevier Science (USA). All rights reserved. doi:10.1016/S0014-4827(03)00070-3

¨ rd, T. O ¨ rd / Experimental Cell Research 286 (2003) 308 –320 D. O

Jun mutants, or inhibition of c-Jun activator JNK (c-Jun NH2-terminal protein kinase) protects neurons and neuronal PC12 cells from the death induced by nerve growth factor (NGF) removal, indicating that the JNK and c-Jun pathway is an important mediator of neuronal cell death [15–18]. Activation of the JNK-c-Jun pathway induces expression of Fas ligand in neuronal PC12 cells and cerebellar granule neurons deprived of NGF and potassium, respectively, providing a mechanism by which c-Jun can trigger apoptosis [19]. Recent experiments using cDNA array technology reveal that withdrawal of survival factors results in upregulation of a wide variety of genes in cerebellar granule neurons, including not only the components of the intrinsic cell death machinery and proapoptotic signaling pathways, but also the genes which are involved in stress response, cell maturation, mitochondrial functions, and synaptic vesicle release [20]. Compared with the gene expression component of the neuronal death program induced by deprivation of survival factors, less is known about the genes activated in the case of cell death evoked by oxidative stress or perturbed Ca2⫹ homeostasis, thought to be associated with neuropathological states [21–23]. Therefore, we carried out differential cloning of genes upregulated in mouse hypothalamic neuronal cell line GT1-7 [24] during induction of cell death by various stressful stimuli, and isolated a gene activated in response to Ca2⫹ release from endoplasmic reticulum (caused by treatment of the cells with thapsigargin, the inhibitor of endoplasmic reticular Ca2⫹-ATPase) and in response to glutathion depletion (caused by treatment of the cells with buthionine sulfoximine, the inhibitor of ␥-glutamylcysteine synthetase). Nucleotide sequence analysis of the gene revealed that it represents the mouse homolog of rat gene NIPK [25], encoding a protein with an unknown function. We show that mNIPK interacts with ATF4 (also known as CREB2 or C/ATF) [26 –28], a ubiquitously expressed member of the ATF/CREB transcription factor family, and affects its transcriptional activity.

309

GT1-7). To induce cell death, GT1-7 cells maintained in the complete medium were exposed to 0.5 mM L-buthionine[S,R]-sulfoximine for 24 h or 50 nM thapsigargin (both purchased from Sigma) for various times as described earlier [8,29,30]. Molecular cloning of mNIPK cDNA Poly(A)⫹ RNA was isolated by the QuickPrep mRNA purification kit (Pharmacia, Sweden) from GT1-7 cells grown under normal or death-inducing conditions, and double-stranded cDNA pools were prepared by TimeSaver cDNA synthesis kit (Pharmacia). The cDNA pools were used for a subtractive cloning procedure representational difference analysis (RDA) according to published protocol [31], and a 0.5-kb cDNA fragment of an undescribed gene (initially named RDA-54), upregulated in the cells treated with buthionine sulfoximine or thapsigargin, was isolated. The full-length coding region of RDA-54 was synthesized by rapid amplification of cDNA ends, using mouse Marathon-Ready cDNA (Clontech, Palo Alto, CA), and several cDNA clones were sequenced by an automated sequence analyzer (Applied Biosystems Model 377). The results indicate that RDA-54 is a mouse homolog of rat gene NIPK [25] and therefore it is designated as mouse (m) NIPK. mNIPK cDNA accession number in the EMBL Nucleotide Sequence Database is AJ514260. Northern blot analysis Total RNA was isolated from GT1-7 cells using the RNeasy kit (Qiagen), separated on 1.2% agarose gel containing formaldehyde, and transfered to a nylon filter (Amersham, UK). A blot containing poly(A)⫹ RNA isolated from various tissues of adult mouse was purchased from Clontech. The filters were hybridized with [32P]dCTP-labeled mNIPK, mouse ATF4, or human ␤-actin cDNA fragments according to standard protocol [32]. Expression constructs

Materials and methods Mammalian cell culture, transfection, and treatment GT1-7 cells (obtained from Dr. Dale E. Bredesen, Buck Institute for Age Research, Novato, CA) were maintained on poly-L-lysine-coated plates in Dulbecco’s modified Eagle’s medium (DMEM; high glucose formulation) supplemented with 10% fetal calf serum (FCS) and 1% penicillinstreptomycin (PS) in an atmosphere of 5% CO2 at 37°C. Cos-7 cells were cultured in Iscove’s modified Dulbecco’s medium supplemented with 10% FCS/1% PS and CHO cells in F12 medium supplemented with 10% FCS/1% PS. Transfection of the cells was carried out by electroporation, using Gene Pulser II (Bio-Rad, Hercules, CA) at a setting of 950 ␮F and 180 V (for cos-7) or 230 V (for CHO and

To make NIPK-pGBT9, the full-length coding region of mNIPK cDNA was amplified by PCR and inserted into vector pGBT9 cut with SmaI/SalI, generating in-frame fusion of mNIPK to the C-terminus of Ga14 DNA-binding domain (DBD). To make ATF4-pVP16, the full-length coding region of mouse ATF4 [28] originating from I.M.A.G.E. EST clone No. 1514797 (purchased from Research Genetics) was amplified by PCR and inserted into vector pVP16 [33] cut with NotI/EcoRI, behind the in-frame activation domain (AD) of VP16. In the same way, plasmids encoding ATF4 deletion mutants m1 (containing amino acid [aa] residues 90 –349 of ATF4), m2 (aa 125–349), m3 (aa 1– 89), and m4 (aa 53– 89) fused to AD of VP16 in vector pVP16 were constructed. Mammalian expression plasmid E2NIPK-pCG encoding mNIPK with the N-terminal E2

310

¨ rd, T. O ¨ rd / Experimental Cell Research 286 (2003) 308 –320 D. O

epitope tag (peptide GVSSTSSDFRDR) was made by insertion of the full-length coding region of mNIPK cDNA into HindIII/KpnI sites of vector pCG-3F12 [34] (obtained from Dr. Mart Ustav, Tartu University, Estonia). Mammalian expression plasmid HA-ATF4-pCG encoding mouse ATF4 with the N-terminal hemagglutinin (HA) tag was made by insertion of the full-length coding region of ATF4 cDNA into XbaI/BamHI sites of plasmid pCGN [35] (a gift from Dr. Mart Ustav). NIPK-pEYFP was constructed by insertion of the coding region of mNIPK cDNA into Hind III/BamH I sites of pEYFP-N1 (Clontech), generating inframe fusion of EYFP to the C-terminus of mNIPK. The structure of the constructs was verified by automated DNA sequencing. Localization of EYFP-fused mNIPK Cos-7, GT1-7, and CHO cells were transfected with NIPK-pEYFP, pEYFP-N1, and Nuc-pECFP (Clontech) and seeded on microscope slides. At 14 h posttransfection, the cells were washed two times with PBS and fixed with 4% paraformaldehyde for 15 min. After two more washes with PBS, the slides were mounted and examined with a Zeiss Axioplan epifluorescence microscope. Yeast two-hybrid system NIPK-pGBT9 and a mouse embryonic cDNA library in pVP16 [33] (provided by Dr. Mart Saarma, University of Helsinki, Finland) were used to cotransform Saccharomyces cerevisiae GC1943 cells by the lithium acetate/polyethylene glycol method, and the transformants were selected on synthetic dropout medium lacking tryptophan, leucine, and histidine for 6 days. Positive clones were picked, and the plasmids were recovered from the colonies, transfered to Escherichia coli HB101, propagated, and used to cotransform S. cerevisiae GC1943 with mNIPK-pGBT9 and control plasmids (empty pGBT9 and SNF1-pGBT9 encoding S. cerevisiae protein kinase SNF1 fused in-frame to the Cterminus of Ga14 DBD [a generous gift from Dr. Juhan Sedman, Tartu University]). The yeast transformants were assayed for ␤-galactosidase activity on filters. cDNA inserts of the mNIPK-specific positive clones were characterized by automated nucleotide sequencing. Coimmunoprecipitation and immunoblotting Cos-7 were cotransfected with the following plasmid mixes (1 ␮g of each plasmid): HA-ATF4-pCG and E2NIPK-pCG, HA-ATF4-pCG and pCG (empty vector), E2NIPK-pCG and pCG. The cells were collected 32 h later and lysed in buffer L (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 1% Triton X-100, 2 mM EDTA, 2 mM EGTA, 50 mM NaF, 5 mM ␤-glycerophosphate, 1 mM sodium orthovanadate, 10 mM DTT, 3 mM Pefabloc [Roche Molecular Biochemicals]). The lysates were incubated on ice for 30 min and

centrifuged at 13,000g for 10 min. The supernatant was preabsorbed with protein G-Sepharose (Pharmacia) for 2 h at 4°C. After that anti-HA monoclonal antibody (MAb) (5 ␮g) (clone 12CA5, Roche Molecular Biochemicals) or anti-E2 MAb (5 ␮g) (clone 3F12 [34], purchased from Quattromed, Estonia) was added and the mixture rotated for 1 h at 4°C. Immunocomplexes were precipitated with protein G-Sepharose for 2 h at 4°C and washed four times in the buffer L. The Sepharose beads were boiled in SDSPAGE sample buffer, samples were resolved by SDSPAGE and transferred to nitrocellulose membrane. Anti-HA MAb conjugated with peroxidase (Roche Molecular Biochemicals) and anti-E2 MAb conjugated with peroxidase (Quattromed) were used for the detection of HA epitopetagged ATF4 and E2 epitope-tagged mNIPK, respectively. The blots were treated by the enhanced chemiluminescence kit from Amersham. CAT assay CHO or cos-7 cells were cotransfected with 0.5 ␮g of CAT reporter plasmid (pEC[ATFx3]-CAT or pELAM-CAT [36], provided by Dr. Tsonwin Hai, Ohio State University, Columbus, OH) along with various amounts of HA-ATF4pCG and E2-NIPK-pCG. Total amount of the plasmid DNA was kept constant by adding empty pCG as appropriate. A plasmid encoding ␤-galactosidase was used as a control of transfection efficiency. At 40 h after transfection, CAT activity was measured by the conversion of 14C-labeled chloramphenicol to the acylated form [32] and normalized for ␤-galactosidase activity. Pulse-chase labeling Cos-7 cells were cotransfected with 1 ␮g of HA-ATF4pCG along with 1 ␮g of E2-NIPK-pCG or empty pCG. Twenty four hours later the cells were rinsed with methionine-cysteine-free medium once, incubated at 37°C for 30 min, and then supplemented with 140 ␮Ci of 35S-methionine-cysteine labeling mix (PRO-MIX, Amersham) in the same medium. After a 1-h pulse, cells were placed into complete medium containing no label for the chase. Cells were lysed at appropriate time points with 0.4 ml of buffer P (20 mM Tris-HCl [pH 7.5], 500 mM NaCl, 1% Triton X-100, 0.1% SDS, 2 mM EDTA, 3 mM Pefabloc) for 30 min on ice. The cell lysates were centrifuged at 13,000g for 10 min, precleared with protein G-Sepharose for 2 h, and incubated with anti-HA MAb for 10 h at 4°C. Immunocomplexes were collected with protein G-Sepharose, washed four times in buffer P, and electrophoresed in 12.5% polyacrylamide-SDS gel. Sodium salicylate fluorography of the gel was performed according to the published protocol [32]. PhosphorImager SI (Molecular Dynamics) was used for quantifying 35S-labeled ATF4.

¨ rd, T. O ¨ rd / Experimental Cell Research 286 (2003) 308 –320 D. O

311

Fig. 1. Northern blot analysis of mNIPK expression. (A) Cell-death-triggering treatment of GTI-7 cells with buthionine sulfoximine or thapsigargin induces expression of mNIPK. The blot containing total RNA isolated from the cells that had been exposed for 24 h to either 0.5 mM buthionine sulfoximine (lane 2) or 50 nM thapsigargin (lane 3) or left untreated (lane 1) was hybridized to mNIPK (upper panel) or actin probe (as a loading control) (lower panel). (B) Analysis of mNIPK expression in various tissues of adult mouse. The blot containing on each lane 2 ␮g of polyA(⫹) RNA isolated from the tissues indicated (Multiple Tissue Northern Blot, Clontech) was hybridized to mNIPK (upper panel) or actin probe (lower panel).

Electrophoretic mobility-shift assay (EMSA) Cos-7 cells were cotransfected with ␤-galactosidase expression plasmid and pairwise combinations of HA-ATF4pCG, E2-NIPK-pCG, and pCG (3 ␮g of each plasmid). After 22 h the cells were suspended in a buffer containing 20 mM Tris-HCl (pH 7.5), 450 mM NaCl, 0.4 mM EDTA, 1 mM DTT, 25% glycerol, and 3 mM Pefabloc. The cells were lysed by three cycles of freezing-thawing and cellular debris was removed by centrifugation (13,000g for 10 min at 4°C). Synthetic oligonucleotide containing CRE (5⬘-AAGATTGCCTGACGTCAGAGAGCTAG) was labeled with [␥-32P]ATP by T4 polynucleotide kinase and annealed with the complementary strand. Equal amounts (3 ␮l) of the cleared lysates normalized for ␤-galactosidase activity were mixed with 5 ␮l of binding buffer [10 mM Tris-HCl (pH 7.5), 2 mM DTT, 2 mM MgCl2, 15% glycerol, 1 mM EDTA, 5 mg/ml BSA, 3 mM Pefabloc, and 200 ␮g/ml poly(dI-dC)], and 0.2 ng of 32P-labeled double-stranded CRE oligonucleotide (in 2 ␮l of TE) was added. Incubation was carried out at room temperature for 15 min and the mixtures were then separated on a 5% nondenaturing polyacrylamide gel. An oligonucleotide 5⬘-ACAAAGTACCGTTGCCGGTCGAA containing the binding site of E2 protein of bovine papillomavirus type 1 [37] was used as a nonspecific competitor. For the supershift assay, 0.4 ␮g of anti-E2 MAb (Quattromed), anti-HA MAb (Roche Molecular Biochemicals), anti-ATF4 antibody, or anti-C/EBP␤ antibody (both from Santa Cruz Biotechnology) was used. Results Cloning and the characterization of mouse NIPK As reported earlier, the exposure of GT1-7 cells to 50 nM thapsigargin triggers an apoptotic mode of death, resulting

in approximately 50% cell loss by 24 h [8], while the exposure to 0.5 mM buthionine sulfoximine for 24 h evokes a necrotic mode of death, resulting in more than 90% cell loss during the following 20-h period [29,30]. Using the PCR-coupled subtractive cDNA cloning procedure RDA [31], we identified a gene (RDA-54) strongly activated in GT1-7 cells treated with thapsigargin and weakly in the cells treated with buthionine sulfoximine (Fig. 1A). Nucleotide sequencing revealed that RDA-54 is 91% homologous to NIPK (neuronal cell death-inducible putative kinase), a cDNA cloned from the PC6-3 subline of rat pheochromocytoma cell line PC12 [25], and therefore we refer to RDA-54 as mouse (m) NIPK. Similarly to its counterpart in GT1-7, NIPK expression in neuronally differentiated PC6-3 cells and in rat sympathetic and cortical neurons is induced in cell death-causing conditions (NGF withdrawal, exposure to the Ca2⫹ ionophore A23187) [25]. To characterize the expression pattern of mNIPK in the adult mouse, a blot containing RNA samples from various tissues was hybridized to mNIPK probe. About 2 kb mNIPK mRNA is detected in the liver, but not in the heart, brain, spleen, lung, skeletal muscle, kidney, or testis (Fig. 1B). mNIPK cDNA contains an open-reading frame encoding a protein of 354 aa residues with 92% identity to rat NIPK (Fig. 2). Remarkably, the C-terminal part of the NIPK proteins is homologous to protein kinase subdomains IV– XI, and in this region a substantial similarity to c5fw, a canine phosphoprotein with an unknown function [38,39], and Tribbles, a recently discovered blocker of mitosis in Drosophila [40 – 42], is observed. The comparison of the amino acid sequences with the consensus sequence of protein kinases [43] reveals that several key residues, including the invariant asparagine (N171 of protein kinase A-C␣) and nearly invariant histidine in the catalytic loop of protein kinases (in subdomain VIB), are not conserved in the NIPK proteins, c5fw and Tribbles. Also, none of the proteins has

312

¨ rd, T. O ¨ rd / Experimental Cell Research 286 (2003) 308 –320 D. O

Fig. 2. Alignment of the aa sequences of mouse NIPK (mNIPK) (the EMBL database Accession Number AJ514260), rat NIPK (rNIPK) (AB020967), canine phosphoprotein c5fw (X99144), Drosophila Tribbles (AF204688), and yeast Candida tropicalis protein kinase SNF1 (AB024535) (only residues 1–391 are shown). Gaps (⫺) were introduced for better alignment. Residues shared by mNIPK and at least one of the other proteins are shown on solid background. Protein kinase subdomains IV–XI are indicated on the SNF1 sequence. Arrow marks the invariant N residue and asterisk the highly conserved H residue in the catalytic loop of protein kinases that are replaced by R and L residues in the NIPK proteins, respectively. mNIPK aa sequence homology to Tribbles and SNF1 is calculated for the kinase-like domain only.

been reported to possess kinase activity. Thus, mNIPK belongs to an emerging group of kinase-like proteins that may not be functional protein kinases. mNIPK-EYFP has predominantly the nuclear localization To characterize the subcellular localization of mNIPK, we fused the full-length coding region of mNIPK cDNA in-frame with EYFP in vector pEYFP-N1, and transfected into cos-7, GT1-7, and CHO cells. The examination of the cells in the fluorescence microscope 14 h later

reveals that mNIPK-EYFP fusion protein resides predominantly in the nuclei of the transfected cells (Figs. 3A and B and data not shown), similarly to the fluorescent protein fused to the nuclear localization signal of SV40 large T antigen expressed in the cells transfected with pECFP-Nuc (Figs. 3E and F). At the same time, in the cells transfected with the control vector pEYFP-N1, the EYFP is distributed equally throughout the whole cell (Figs. 3C and D), suggesting that mNIPK contains a signal involved in targeting of mNIPK-EYFP to the nuclei.

¨ rd, T. O ¨ rd / Experimental Cell Research 286 (2003) 308 –320 D. O

313

To map the mNIPK-binding domain of ATF4, deletion mutants of ATF4 were made and analyzed by the yeast two-hybrid assay. The fragment of ATF4 extending from the LZII to the C-terminus (aa 90 –349) and the region downstream of the LZII (aa 125–349) revealed no binding to mNIPK (Fig. 4). Similarly, ATF4 fragments upstream of the LZII (aa 1– 89 and 53– 89) did not interact with mNIPK in the two-hybrid system. The results suggest that both the LZII and the region upstream of LZII are critical for mNIPK binding to ATF4, and the C-terminal part of ATF4 is not involved in the interaction. Expression of both mNIPK and ATF4 is upregulated in GT1-7 cells exposed to thapsigargin

Fig. 3. mNIPK-EYFP fusion protein is targeted to the nucleus. Fluorescence (A, C, and E) and phase contrast (B, D, and F) micrographs of cos-7 cells transiently transfected with expression plasmid for mNIPK-EYFP (A and B), EYFP (C and D), or ECFP fused to the nuclear localization signal of SV40 large T antigen (ECFP-Nuc) (E and F). Magnification, ⫻400.

Identification of ATF4 as a protein binding to mNIPK by the yeast two-hybrid system To find proteins interacting with mNIPK, the full-length coding region of mNIPK cDNA was cloned into pGBT9, downstream of the Gal4 DNA-binding domain, and used as a bait to screen a murine embryonic cDNA library cloned in fusion with the VP16 activation domain in vector pVP16 [33]. Out of about one million yeast cells cotransformed with the bait and prey plasmids, 45 His autotroph colonies were isolated, 11 of which were positive in the subsequent X-gal test. Plasmids of these yeast clones were propagated in E. coli, tested again in the yeast two hybrid system with different baits (mNIPK and S. cerevisiae protein kinase SNF1), and characterized by nucleotide sequencing. This analysis revealed that two of the mNIPK-specific positive clones (D7 and D24) encode overlapping fragments of transcription factor ATF4. The clone D7 extends from residue 53 to 155 and the clone D24 from residue 29 to 124 of the mouse ATF4 aa sequence [28]. Thus, both clones cover the N-terminal leucine zipper motif of ATF4 (leucine zipper II [LZII] located in position 90 –125). In addition, D7 contains short aa stretches up- and downstream of the LZII (37 and 30 residues, respectively), and D24 contains a 61-residue fragment upstream of the LZII (Fig. 4). The two-hybrid assay of full-length ATF4 and mNIPK, carried out next, gave a positive result. At the same time, control experiments with the empty bait vector and with the yeast kinase SNF1 used as the bait were negative, thus excluding autoactivation and corroborating the specificity of the interaction between ATF4 and mNIPK.

To study temporal expression patterns of mNIPK and ATF4 in thapsigargin-treated GT1-7 cells and confirm their simultaneous expression, mRNA levels of the genes were analyzed at various time points of the treatment. Both mNIPK and ATF4 are upregulated in response to thapsigargin, but demonstrate different profiles of induction (Fig. 5A). While ATF4 mRNA is detectable by Northern analysis also in untreated GT1-7 cells and its level is increased moderately (2- to 4-fold) starting from the first time point analyzed (the cells exposed to thapsigargin for 1 h), increase of mNIPK mRNA level is delayed for several hours, but is more extensive (the mRNA signal is more than 20-fold stronger in the cells treated for 20 h than in the cells treated for 2 h). Immunobloting of the cellular lysates reveals the fast accumulation of ATF4 protein in thapsigargin-treated GT1-7 cells (Fig. 5B), consistent with the translational con-

Fig. 4. mNIPK interacts with ATF4 in the yeast two-hybrid system. Schematic representation of ATF4 (wt) and its fragments (clones D7 and D24, and deletion mutants m1–m4) examined for binding to mNIPK, and the results of the two-hybrid assay. mNIPK (bait) was fused to the Gal4 DBD in vector pGBT9 and ATF4 and its fragments (prey) were fused to the VP16 AD in vector pVP16. The cells cotransfected with the bait and prey constructs were assayed for the activation of the reporter genes HIS3 and lacZ, indicative of the interaction (⫹, His autotrophy and development of dark blue color in the ␤-gal assay in 2 h; ⫺, His auxotrophy and no color development in the ␤-gal assay in 8 h). N-terminal leucine zipper motif (LZ II), the basic DNA-binding domain (b), and C-terminal leucine zipper motif (LZ) of ATF4 are represented by black, striped, and grey boxes, respectively.

314

¨ rd, T. O ¨ rd / Experimental Cell Research 286 (2003) 308 –320 D. O

Fig. 5. Time course of mNIPK and ATF4 expression in GT1-7 cells exposed to thapsigargin. (A) Northern blot analysis of total RNA isolated from the cells that had been treated with 50 nM thapsigargin for the indicated period of time (lanes 2–5) or left untreated (lane 1). The blot was hybridized to mNIPK (upper panel), ATF4 (middle), or actin probe (lower panel). (B) Immunoblot analysis of ATF4 protein content of the cells that had been treated with 50 nM thapsigargin for the indicated period of time (lanes 2– 6) or left untreated (lane 1). Twenty micrograms of total cellular protein was loaded on each lane and the blot was probed with polyclonal anti-ATF4 antibody.

trol mechanism of ATF4 expression described by Harding et al. [44]. ATF4 binds to mNIPK in transiently transfected mammalian cells In order to find out whether ATF4 associates with mNIPK in mammalian cells, cos-7 cells were transfected with expression plasmids encoding HA epitope-tagged ATF4 and E2 epitope-tagged mNIPK. The cells were lysed 32 h later and immunoprecipitation with anti-E2 or anti-HA MAb was carried out. Western blot with anti-HA MAb reveals that HA-tagged ATF4 is precipitated by E2 MAb in the lysate of cells coexpressing HA-ATF4 and E2-mNIPK, but not in the control without E2-mNIPK (Fig. 6). Similarly, the reciprocal experiment with anti-E2 MAb indicates that E2-mNIPK is specifically coprecipitated with HA-ATF4, and thus confirms the formation of mNIPK-ATF4 complex in mammalian cells.

Fig. 7A, the cotransfection of CHO cells with pEC(ATFx3)CAT and increasing the amounts of HA-ATF4-pCG (0.1 to 4 ␮g) activates the reporter dose dependently, yielding up to 18-fold elevation of CAT activity. Raising the HA-ATF4pCG amount further (to 10 ␮g) does not increase the reporter activity (data not shown), probably due to squelching caused by a very high concentration of ATF4 [45]. mNIPK coexpression at a constant low level (0.1 ␮g of E2-NIPKpCG) decreases the ATF4 transcriptional activation activity at least 50% in comparison to the cells transfected with

mNIPK inhibits the transcriptional activation activity of ATF4 To elucidate the functional significance of the interaction between ATF4 and mNIPK, we studied the effect of mNIPK coexpression on the transcriptional activity of ATF4. CHO or cos-7 cells were cotransfected transiently with ATF4 and mNIPK expression constructs, and with a CRE-driven CAT reporter plasmid (either pEC(ATFx3)-CAT with three tandem ATF sites, or pELAM-CAT containing nucleotides ⫺383 to ⫹80 of the E-selectin promoter [36]). As shown in

Fig. 6. mNIPK interacts with ATF4 in mammalian cells. Cos-7 cells were transiently transfected with 1 ␮g of expression plasmids for the proteins indicated (E2-NIPK-pCG, HA-ATF4-pCG, or empty pCG designated as mock). Whole-cell extracts (WCE) were separated by SDS-PAGE and immunoblotted with anti-HA MAb (lanes 1–3) or anti-E2 MAb (lanes 6 – 8). The complexes immunoprecipitated from the WCE using anti-E2 MAb were subjected to immunoblotting with anti-HA MAb (lanes 4 and 5) and the complexes immunoprecipitated with anti-HA MAb were immunoblotted with anti-E2 MAb (lanes 9 and 10). The bands corresponding to E2-NIPK and HA-ATF4 proteins are indicated.

¨ rd, T. O ¨ rd / Experimental Cell Research 286 (2003) 308 –320 D. O

315

Fig. 7. mNIPK inhibits ATF4 transcriptional activation activity. CHO cells were cotransfected with 0.5 ␮g of pEC(ATFx3)-CAT reporter plasmid driven by three tandem ATF sites and with the indicated amounts of expression constructs for ATF4 (ATF4-pCG) and mNIPK (NIPK-pCG). The relative CAT activity is calculated by defining the reporter activity in the presence of vector (pCG) as 1. The results are the average of three independent experiments and standard deviations are indicated by error bars. (A) The transfection of the cells with increasing amounts of the ATF4 expression plasmid produces the increase of ATF4 transcriptional activation activity which is reduced by the cotransfection of 100 ng of the mNIPK expression plasmid. (B) ATF4 transcriptional activation activity produced by constant amount (1 ␮g) of the expression plasmid is decreased by the cotransfection of the increasing amounts of the mNIPK expression plasmid.

HA-ATF4-pCG alone. Similarly, the titration of ATF4 (1 ␮g of HA-ATF4-pCG) with the increasing amounts of coexpressed mNIPK (3 ng to 1 ␮g of E2-NIPK-pCG) reduces the ATF4 transcriptional activation activity consistently, indicating that mNIPK inhibits ATF4 in a dose-dependent manner and is able to block it totally (Fig. 7B). The results confirming the inhibition of ATF4 transcriptional activation activity by mNIPK were also obtained in the experiments carried out in CHO cells with pELAM-CAT reporter plasmid and in cos-7 cells with pEC(ATFx3)-CAT and pELAM-CAT plasmids. At the same time, we observed no influence of mNIPK coexpression on the activity of an unrelated transactivating factor, protein E2 of bovine papillomavirus type 1, in an E2-binding site-driven CAT reporter assay of transiently transfected CHO cells (data not shown). ATF4 degradation rate is not increased in the cells cotransfected with mNIPK expression construct Tribbles, a Drosophila protein with substantial homology to mNIPK, stimulates turnover of C/EBP transcription factor Slbo that can physically associate to Tribbles [46]. To determine whether mNIPK stimulates ATF4 degradation (which might explain the decrease of the ATF4 transcriptional activity described above), a pulse-chase labeling assay was carried out. Cos-7 cells were transfected with expression plasmids for HA-ATF4 and E2-mNIPK, or the empty vector, and labeled metabolically with 35S-methionine and 35S-cysteine. The labeling period of 1 h was followed by a chase period of 2, 4, or 8 h in medium with an excess of nonradioactive methionine and cysteine. Immunoprecipitation of HA-ATF4 by anti-HA MAb at these

time points indicates that the ATF4 half-life is slightly shorter in the cells transfected with expression plasmid for HA-ATF4 alone compared to the cells cotransfected with expression plasmid for E2-mNIPK (Figs. 8A and B). In agreement with this, the estimation of ATF4 steady-state level by immunoblot in transiently transfected cos-7 cells reveals that ATF4 level is slightly increased by mNIPK coexpression (Fig. 8C). These data suggest that mechanisms other than stimulation of degradation are involved in the repression of ATF4 transcriptional activity by mNIPK. ATF4-mNIPK complex binds to CRE An obvious reason for the inhibition of ATF4 transcriptional activity might be blocking of ATF4 DNA-binding ability by mNIPK. To explore this possibility, we employed EMSA of a 32P-labeled CRE-containing oligonucleotide. The incubation of the oligonucleotide with extract of cos-7 cells cotransfected with HA-ATF4-pCG and E2-NIPK-pCG plasmids produces a prominent slower migrating band that is not seen in control experiments with the empty vector and with neither of the expression constructs alone (Fig. 9, lanes 1, 2, 4, and 10). The band is completely supershifted by anti-ATF4 antibody and anti-E2 MAb, indicating that both ATF4 and mNIPK are present in this complex, and not supershifted by a control anti-C/EBP␤ antibody (Fig. 9, lanes 13, 14, and 15). The complex is competed by excess unlabeled CRE oligonucleotide, but not by an unrelated oligonucleotide, confirming that ATF4-mNIPK binds to CRE specifically (Fig. 9, lanes 11 and 12). No binding of mNIPK alone to CRE oligonucleotide is observed (Fig. 9, lanes 2 and 3). At the same time, the formation of an ATF4-CRE complex is detected by EMSA (extract of the

316

¨ rd, T. O ¨ rd / Experimental Cell Research 286 (2003) 308 –320 D. O

Fig. 8. ATF4 degradation rate is not increased in the cells coexpressing mNIPK in the transient transfection assay. (A) Pulse-chase labeling of cos-7 cells transfected with HA-ATF4-pCG (lanes 1– 4) or cotransfected with HA-ATF4-pCG and E2-NIPK-pCG (lanes 5– 8). The cells were metabolically labeled with 35 S-methionine-cysteine mix for 1 h (lanes 1 and 5) and then incubated in nonradioactive rich medium for 2, 4, or 8 h (lanes 2 and 6, 3 and 7, and 4 and 8, respectively). After that the cells were lysed and subjected to immunoprecipitation with anti-HA MAb. The isolated proteins were separated by SDS-PAGE and the gels were exposed to X-ray film. The bands corresponding to HA-ATF4 and E2-NIPK are indicated; identity of the E2-NIPK was confirmed by immunoblot with anti-E2 MAb (not shown). The experiment was repeated twice with similar results. (B) Graphic representation of the results shown on (A). 35 S-labeled ATF4 was quantified with a PhosphorImager and plotted as the percentage of the signal at the beginning of the chase. (C) Analysis of ATF4 steady-state level in transiently transfected cells. Whole-cell extracts of cos-7 cells cotransfected with HA-ATF4-pCG and ␤-gal-pHM6 (lane 1) or with HA-ATF4-pCG, E2-NIPK-pCG, and ␤-gal-pHM6 (lane 2) were resolved by SDS-PAGE and immunoblotted with anti-HA MAb. HA-tagged ␤-galactosidase bands are shown to confirm equal transfection efficiency.

cells transfected with HA-ATF4-pCG (Fig. 9, lane 4) produces an increased amount of protein-CRE complexes compared to extract of the cells transfected with empty vector (Fig. 9, lane 1)) and by antibody supershift assay (a significant proportion of the complexes produced by extract of the cells transfected with HA-ATF4-pCG is supershifted by anti-HA MAb and anti-ATF4 antibody, but not by anti-E2 MAb used as a control) (Fig. 9, lanes 7, 8, and 9).

Discussion In this study we have characterized mNIPK, a gene activated in neuronal cells in response to stressful stimuli (disruption of calcium homeostasis, inhibition of glutathion synthesis) causing cell death eventually, and showed that mNIPK interacts with transcription factor ATF4. Although originally described as a negative regulator of CRE-dependent transcription [27,47,48], ATF4 is now known to contain a strong transcriptional activation domain [45,49] and can act as a positive regulator of transcription [45,50,51].

Both the N-terminal 113 aa and the C-terminal 87 aa of ATF4 interact with transcriptional coactivator CREB-binding protein (CBP) and general transcription factors TFIIB, TATA-binding protein, and the RAP30 subunit of TFIIF, but the N-terminal region activates transcription much more efficiently than the C-terminal region [45,49]. Consistent with its role as an activator, we observed the stimulatory effect of ATF4 on the expression of a reporter gene driven by promoters with the ATF/CRE site in transient transfection assays. The stimulatory effect was reduced by simultaneous expression of mNIPK in a dose-dependent manner, while ATF4 complexed with mNIPK was still able to bind to CRE. The critical region for mNIPK binding is located near the N-terminus of ATF4, covering a heptad repeat of leucines from position 90 to 125 and a preceding stretch of about 40 aa, and thus overlapping with the transcriptional activation domain of ATF4. Therefore, it seems likely that mNIPK inhibits ATF4 transcriptional activation activity by interfering with recruitment of the transcriptional apparatus. mNIPK might do this either by directly competing with CBP and the general transcription factors for the binding

¨ rd, T. O ¨ rd / Experimental Cell Research 286 (2003) 308 –320 D. O

317

Fig. 9. ATF4-mNIPK complex binds to CRE. EMSA of radiolabeled CRE oligonucleotide and extracts made of cos-7 cells transfected with empty vector (pCG) (lane 1), E2-NIPK-pCG (lanes 2 and 3), HA-ATF4-pCG (lanes 4 –9), or cotransfected with E2-NIPK-pCG and HA-ATF4-pCG (lanes 10 –15). Anti-E2 MAb (lanes 3, 9, and 14), anti-HA MAb (lane 7), anti-ATF4 antibody (lanes 8 and 13), and anti-C/EBP␤ antibody (lane 15) were used for supershift analysis. A 100-fold excess of unlabeled CRE oligonucleotide (lanes 5 and 11) or unrelated oligonucleotide (containing binding site of E2 protein of BPV1) (lanes 6 and 12) was added to the extracts as a specificity control. ATF4-NIPK bound to the CRE oligonucleotide is indicated. Asterisk marks ATF4 and endogenous cos-7 proteins bound to the CRE probe.

site on ATF4 molecule or inducing a change in ATF4 conformation or a posttranslational modification state. The previously described repressor of ATF4, latency-associated nuclear antigen of Kaposi’s sarcoma-associated herpesvirus, inhibits the transactivation activity through a different mechanism, interacting with the bZip structure in the Cterminal part of ATF4 and revealing no binding to ATF4CRE complex in EMSA [52]. Similarly, the interaction between ATF4 and CHOP (also known as Gadd153), a transcription factor which is able to down-regulate ATF4 transcriptional activation activity, is mediated by the Cterminal leucine zipper [53]. Thus, mNIPK is the only repressor currently described interacting with the N-terminal region of ATF4. A major portion of mNIPK (region of 200 aa) reveals homology to subdomains IV–XI of a common catalytic core structure of the protein kinase family [43], but several replacements of the invariant or highly conserved amino acid residues make it unlikely that mNIPK is a functional protein kinase. Similar kinase-like domains with severe deviations from the kinase consensus sequence exist also in Drosophila protein Tribbles involved in the cell cycle regulation [40 – 42], and in mammalian proteins c5fw and c8fw with unknown functions [38,39]. The comparison of the aa sequences suggests that the NIPK proteins, Tribbles, c5fw, and c8fw have first diverged from the SNF1 class of serine/ threonine kinases and then from each other. Biochemical analysis is needed to elucidate whether mNIPK and its relatives have evolved as proteins without a catalytic activity or they represent an unrecognized group of enzymes. ATF4 mRNA is present in a wide variety of embryonic and adult tissues [27,54], but physiological functions of the protein are not well understood [55]. ATF4 has been implicated in memory formation [56] and the targeted disruption

of the gene in mouse has disclosed the essential role of ATF4 in mammalian development [57–59]. A growing body of evidence suggests that ATF4 is involved in regulation of gene expression during various conditions of stress [44,60]. The increase of ATF4 mRNA level, similar to that observed in GT1-7 cells in response to thapsigargin, has been reported in several cell types exposed to stressful stimuli like anoxia [61], calcium ionophore [62], and homocysteine [63]. Moreover, the mammalian eIF2a kinases PERK and GCN2 selectively increase translation of ATF4 mRNA in endoplasmic reticulum-stressed and amino aciddeprived cells [44]. The possible target genes activated by ATF4 in stressed cells include CHOP [44] and HO-1 (heme oxygenase-1) [64] that both have been implicated in regulation of apoptosis [65,66]. While HO-1 has a cytoprotective role in many situations triggering cell death (serum deprivation or exposure of the cells to staurosporine, etoposide, or TNF) [65,67], the induction of CHOP serves as a death signal, sensitizing cells to stress by down-regulating expression of antiapoptotic gene bcl2 [68]. Therefore, mNIPK acting as a repressor of ATF4 may have connections to both anti- and proapoptotic pathways of the cell. mNIPK may also influence ATF4 interaction with apoptosis-related kinase Dlk/Zip kinase [69,70] or with the metabotropic GABAB receptor [71,72], or affect cross-talk between ATF4 and CREB, an important mediator of both neuronal plasticity and survival [73,74]. These issues will be addressed in future experiments.

Acknowledgments We thank Drs. Mart Saarma and Stanley M. Hollenberg for providing mouse embryonic cDNA library and Drs.

318

¨ rd, T. O ¨ rd / Experimental Cell Research 286 (2003) 308 –320 D. O

Tsonwin Hai, Mart Ustav, Dale E. Bredesen, and Juhan Sedman for the gift of plasmids and cell lines. We are grateful to Drs. Mart Ustav and Mart Saarma for critical comments. This work was supported in part by Grant 3254 from the Estonian Science Foundation.

References [1] R.W. Oppenheim, Cell death during development of the nervous system, Annu. Rev. Neurosci. 14 (1991) 453–501. [2] M.C. Raff, B.A. Barres, J.F. Burne, H.S. Coles, Y. Ishizaki, M.D. Jacobson, Programmed cell death and the control of cell survival: lessons from the nervous system, Science 262 (1993) 695–700. [3] M. Ba¨ hr, Live or let die—retinal ganglion cell death and survival during development and in the lesioned adult CNS, Trends Neurosci. 23 (2000) 483– 490. [4] C. Bossenmeyer, R. Chihab, S. Muller, H. Schroeder, J.L. Daval, Hypoxia/reoxygenation induces apoptosis through biphasic induction of protein synthesis in cultured rat brain neurons, Brain Res. 787 (1998) 107–116. [5] F. Dessi, C. Charriaut-Marlangue, Y. Ben-Ari, Anisomycin and cycloheximide protect cerebellar neurons in culture from anoxia, Brain Res. 581 (1992) 323–326. [6] D.P. Martin, R.E. Schmidt, P.S. DiStefano, O.H. Lowry, J.G. Carter, E.M. Johnson Jr., Inhibitors of protein synthesis and RNA synthesis prevent neuronal death caused by nerve growth factor deprivation, J. Cell. Biol. 106 (1988) 829 – 844. [7] S. Tan, M. Wood, P. Maher, Oxidative stress induces a form of programmed cell with characteristics of both apoptosis and necrosis in neuronal cells, J. Neurochem. 71 (1998) 95–105. [8] H. Wei, W. Wei, D.E. Bredesen, D.C. Perry, Bcl-2 protects against apoptosis in neuronal cell line caused by thapsigargin-induced depletion of intracellular calcium stores, J. Neurochem. 70 (1998) 2305– 2314. [9] S. Estus, W.J. Zaks, R.S. Freeman, M. Gruda, R. Bravo, E.M. Johnson Jr., Altered gene expression in neurons during programmed cell death: identification of c-jun as necessary for neuronal apoptosis, J. Cell Biol. 127 (1994) 1717–1727. [10] R.S. Freeman, S. Estus, E.M. Johnson Jr., Analysis of cell cyclerelated gene expression in postmitotic neurons: selective induction of Cyclin D1 during programmed cell death, Neuron 12 (1994) 343– 355. [11] P.W. Mesner, C.L. Epting, J.L. Hegarty, S.H. Green, A timetable of events during programmed cell death induced by trophic factor withdrawal from neuronal PC12 cells, J. Neurosci. 15 (1995) 7357–7366. [12] T.M. Miller, K.L. Moulder, C.M. Knudson, D.J. Creedon, M. Desmukh, S.J. Korsmeyer, E.M. Johnson Jr., Bax deletion further orders the cell death pathway in cerebellar granule cells and suggests a caspase-independent pathway to cell death, J. Cell. Biol. 139 (1997) 205–217. [13] J. Padmanabhan, D.S. Park, L.A. Greene, M.L. Shelanski, Role of cell cycle regulatory proteins in cerebellar granule neuron apoptosis, J. Neurosci. 19 (1999) 8747– 8756. [14] D.S. Park, S.E. Farinelli, L.A. Greene, Inhibitors of cyclin-dependent kinases promote survival of post-mitotic neuronally differentiated PC12 cells and sympathetic neurons, J. Biol. Chem. 271 (1996) 8161– 8169. [15] A. Eilers, J. Whitfield, B. Shah, C. Spadoni, H. Desmond, J. Ham, Direct inhibition of c-Jun N-terminal kinase in sympathetic neurones prevents c-jun promoter activation and NGF withdrawal-induced death, J. Neurochem. 76 (2001) 1439 –1454. [16] J. Ham, C. Babij, J. Whitfield, C.M. Pfarr, D. Lallemand, M. Yaniv, L.L. Rubin, A c-Jun dominant negative mutant protects sympathetic neurons against programmed cell death, Neuron 14 (1995) 927–939.

[17] A.C. Maroney, M.A. Glicksman, A.N. Basma, K.M. Walton, E. Knight Jr., C.A. Murphy, B.A. Bartlett, J.P. Finn, T. Angeles, Y. Matsuda, N.T. Neff, C.A. Dionne, Motoneuron apoptosis is blocked by CEP-1347 (KT 7515), a novel inhibitor of the JNK signaling pathway, J. Neurosci. 18 (1998) 104 –111. [18] Z. Xia, M. Dickens, J. Raingeaud, R.J. Davis, M.E. Greenberg, Opposing effects of ERK and JNK-p38 MAP kinases on apoptosis, Science 270 (1995) 1326 –1331. [19] H. Le-Niculescu, E. Bonfoco, Y. Kasuya, F.-X. Claret, D.R. Green, M. Karin, Withdrawal of survival factors results in activation of the JNK pathway in neuronal cells leading to Fas ligand induction and cell death, Mol. Cell. Biol. 19 (1999) 751–763. [20] L.W. Chiang, J.M. Grenier, L. Ettwiller, L.P. Jenkins, D. Ficenec, J. Marin, F. Jin, P.S. DiStefano, A. Wood, An orchestrated gene expression component of neuronal programmed cell death revealed by cDNA array analysis, Proc. Natl. Acad. Sci. USA 98 (2001) 2814 – 2819. [21] M.P. Mattson, F.M. LaFerla, S.L. Chan, M.A. Leissring, P.N. Shepel, J.D. Geiger, Calcium signalling in the ER: its role in neuronal plasticity and neurodegenerative disorders, Trends Neurosci. 23 (2000) 222–229. [22] W. Paschen, Role of calcium in neuronal cell injury: Which subcellular compartment is involved?, Brain Res. Bull. 53 (2000) 409 – 413. [23] J.B. Schulz, J. Lindenau, J. Seyfried, J. Dichgans, Glutathione, oxidative stress and neurodegeneration, Eur. J. Biochem. 267 (2000) 4904 – 4911. [24] P.L. Mellon, J.J. Windle, P.C. Goldsmith, C.A. Padula, J.L. Roberts, R.I. Weiner, Immortalization of hypothalamic GnRH neurons by genetically targeted tumorigenesis, Neuron 5 (1990) 1–10. [25] K. Mayumi-Matsuda, S. Kojima, H. Suzuki, T. Sakata, Identification of a novel kinase-like gene induced during neuronal cell death, Biochem. Biophys. Res. Commun. 258 (1999) 260 –264. [26] T. Hai, F. Liu, W.J. Coukos, M.R. Green, Transcription factor ATF cDNA clones: an extensive family of leucine zipper proteins able to selectively form DNA-binding heterodimers, Genes Dev. 3 (1989) 2083–2090. [27] B.A. Karpinski, G.D. Morle, J. Huggenvik, M.D. Uhler, J.M. Leiden, Molecular cloning of human CREB-2: an ATF/CREB transcription factor that can negatively regulate transcription from the cAMP response element, Proc. Natl. Acad. Sci. USA 89 (1992) 4820 – 4824. [28] M. Vallejo, D. Ron, C.P. Miller, J.F. Habener, C/ATF, a member of the activating transcription factor family of DNA-binding proteins, dimerizes with CAAT/enhancer-binding proteins and directs their binding to cAMP response elements, Proc. Natl. Acad. Sci. USA 90 (1993) 4679 – 4683. [29] D.J. Kane, T.A. Sarafian, R. Anton, H. Hahn, E.B. Gralla, J.S. ¨ rd, D.E. Bredesen, Bcl-2 inhibition of neural death: Valentine, T. O decreased generation of reactive oxygen species, Science 262 (1993) 1274 –1277. ¨ rd, R. Anton, D.E. Bredesen, Expression of bcl-2 [30] D.J. Kane, T. O inhibits necrotic neural cell death, J. Neurosci. Res. 40 (1995) 269 – 275. [31] M. Hubank, D.G. Schatz, Identifying differences in mRNA expression by representational difference analysis of cDNA, Nucleic Acids Res. 22 (1994) 5640 –5648. [32] J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1989. [33] S.M. Hollenberg, R. Sternglanz, P.F. Cheng, H. Weintraub, Identification of a new family of tissue-specific basic helix-loop-helix proteins with a two-hybrid system, Mol. Cell. Biol. 15 (1995) 3813– 3822. [34] N. Kaldalu, D. Lepik, A. Kristjuhan, M. Ustav, Monitoring and purification of proteins using bovine papillomavirus E2 epitope tags, Biotechniques 28 (2000) 456 – 460.

¨ rd, T. O ¨ rd / Experimental Cell Research 286 (2003) 308 –320 D. O [35] M. Tanaka, W. Herr, Differential transcriptional activation by Oct-1 and Oct-2: interdependent activation domains induce Oct-2 phosphorylation, Cell 60 (1990) 375–386. [36] B.P.C. Chen, G. Liang, J. Whelan, T. Hai, ATF3 and ATF3⌬Zip. Transcriptional repression versus activation by alternatively spliced isoforms, J. Biol. Chem. 269 (1994) 15819 –15826. [37] A. Abroi, R. Kurg, M. Ustav, Transcriptional and replicational activation functions in the bovine papillomavirus type 1 E2 protein are encoded by different structural determinants, J. Virol. 70 (1996) 6169 – 6179. [38] F. Wilkin, V. Savonet, A. Radulescu, J. Petermans, J.E. Dumont, C. Maenhaut, Identification and characterization of novel genes modulated in the thyroid of dogs treated with methimazole and propylthiouracil, J. Biol. Chem. 271 (1996) 28451–28457. [39] F. Wilkin, N. Suarez-Huerta, B. Robaye, J. Peetermans, F. Libert, J.E. Dumont, C. Maenhaut, Characterization of a phosphoprotein whose mRNA is regulated by the mitogen pathways in dog thyroid cells, Eur. J. Biochem. 248 (1997) 660 – 668. [40] J. Grosshans, E. Wieschaus, A genetic link between morphogenesis and cell division during formation of the ventral furrow in Drosophila, Cell 101 (2000) 523–531. [41] J. Mata, S. Curado, A. Ephrussi, P. Rørth, Tribbles coordinates mitosis and morphogenesis in Drosophila by regulating String/ CDC25 proteolysis, Cell 101 (2000) 511–522. [42] T.C. Seher, M. Leptin, Tribbles, a cell-cycle brake that coordinates proliferation and morphogenesis during Drosophila gastrulation, Curr. Biol. 10 (2000) 623– 629. [43] S.K. Hanks, T. Hunter, The eukaryotic protein kinase superfamily: kinase (catalytic) domain structure and classification, FASEB J. 9 (1995) 576 –596. [44] H.P. Harding, I. Novoa, Y. Zhang, H. Zeng, R. Wek, M. Schapira, D. Ron, Regulated translation initiation controls stress-induced gene expression in mammalian cells, Mol. Cell 6 (2000) 1099 – 1108. [45] G. Liang, T. Hai, Characterization of human activating transcription factor 4, a transcriptional activator that interacts with multiple domains of cAMP-responsive element-binding protein (CREB)-binding protein (CBP), J. Biol. Chem. 272 (1997) 24088 – 24095. [46] P. Rørth, K. Szabo, G. Texido, The level of C/EBP protein is critical for cell migration during Drosophila oogenesis and is tightly controlled by regulated degradation, Mol. Cell 6 (2000) 23–30. [47] G. Cibelli, S. Ju¨ ngling, S. Schoch, H.H. Gerdes, G. Thiel, Identification of a functional cAMP response element in the secretogranin II gene, Eur. J. Biochem. 236 (1996) 171–179. [48] S. Ju¨ ngling, G. Cibelli, M. Czardubon, H.-H. Gerdes, G. Thiel, Differential regulation of chromogranin B and synapsin I gene promoter activity by cAMP and cAMP-dependent protein kinase, Eur. J. Biochem. 226 (1994) 925–935. [49] S. Schoch, G. Cibelli, A. Magin, L. Steinmu¨ ller, G. Thiel, Modular structure of cAMP response element binding protein 2 (CREB2), Neurochem. Int. 38 (2001) 601– 608. [50] F. Gachon, A. Peleraux, S. Thebault, J. Dick, I. Lemasson, C. Devaux, J.-M. Mesnard, CREB-2, a cellular CRE-dependent transcription repressor, functions in association with Tax as an activator of the human T-cell leukemia virus type 1 promoter, J. Virol. 72 (1998) 8332– 8337. [51] T.R. Reddy, H. Tang, X. Li, F. Wong-Staal, Functional interaction of the HTLV-1 transactivator Tax with activating transcription factor-4 (ATF4), Oncogene 14 (1997) 2785–2792. [52] C. Lim, H. Sohn, Y. Gwack, J. Choe, Latency-associated nuclear antigen of Kaposi’s sarcoma-associated herpesvirus (human herpesvirus-8) binds ATF4/CREB2 and inhibits its transcriptional activation activity, J. Gen. Virol. 81 (2000) 2645–2652. [53] F. Gachon, G. Gaudray, S. Thebault, J. Basbous, J.A. Koffi, C. Devaux, J.-M. Mesnard, The cAMP response element binding pro-

[54]

[55]

[56]

[57]

[58]

[59]

[60]

[61]

[62]

[63]

[64]

[65]

[66]

[67]

[68]

[69]

[70]

319

tein-2 (CREB-2) can interact with the C/EBP-homologous protein (CHOP), FEBS Lett. 502 (2001) 57– 62. P. Murphy, A.-B. Kolstø, Expression of the bZip transcription factor TCF11 and its potential dimerization partners during development, Mech. Dev. 97 (2000) 141–148. T. Hai, M.G. Hartman, The molecular biology and nomenclature of the activating transcription factor/cAMP responsive element binding family of transcription factors: activating transcription factor proteins and homeostasis, Gene 273 (2001) 1–11. T. Abel, K.C. Martin, D. Bartsch, E.R. Kandel, Memory supressor genes: inhibitory constraints on the storage of long-term memory, Science 279 (1998) 338 –341. T. Hettmann, K. Barton, J.M. Leiden, Microphthalmia due to p53mediated apoptosis of anterior lens epithelial cells in mice lacking the CREB-2 transcription factor, Dev. Biol. 222 (2000) 110 –123. H.C. Masuoka, T.M. Townes, Targeted disruption of the activating transcription factor 4 gene results in severe fetal anemia in mice, Blood 99 (2002) 736 –745. T. Tanaka, T. Tsujimura, K. Takeda, A. Sugihara, A. Maekawa, N. Terada, N. Yoshida, S. Akira, Targeted disruption of ATF4 discloses its essential role in the formation of eye lens fibres, Genes Cells 3 (1998) 801– 810. T.W. Fawcett, J.L. Martindale, K.Z. Guyton, T. Hai, N.J. Holbrook, Complexes containing activating transcription factor (ATF)/cAMPresponsive-element-binding protein (CREB) interact with the CCAAT/enhancer-binding protein (C/EBP)-ATF composite site to regulate Gadd153 expression during the stress response, Biochem. J. 339 (1999) 135–141. S.D. Estes, D.L. Stoler, G.R. Anderson, Normal fibroblasts induce the C/EBP␤ and ATF-4 bZIP transcription factors in response to anoxia, Exp. Cell Res. 220 (1995) 47–54. A. Tsujimoto, H. Nyunoya, T. Morita, T. Sato, K. Shimotohno, Isolation of cDNAs for DNA-binding proteins which specifically bind to a tax-responsive enhancer element in the long terminal repeat of human T-cell leukemia virus type I, J. Virol. 65 (1991) 1420 –1426. K. Kokame, H. Kato, T. Miyata, Homocysteine-respondent genes in vascular endothelial cells identified by differential display analysis. GRP78/BiP and novel genes, J. Biol. Chem. 271 (1996) 29659 – 29665. C.H. He, P. Gong, B. Hu, D. Stewart, M.E. Choi, A.M.K. Choi, J. Alam, Identification of activating transcription factor 4 (ATF4) as an Nrf2 interacting protein: implication for heme oxygenase-1 gene regulation, J. Biol. Chem. 276 (2001) 20858 –20865. M.P. Soares, Y. Lin, J. Anrather, E. Csizmadia, K. Takigami, K. Sato, S.T. Grey, R.B. Colvin, A.M. Choi, K.D. Poss, F.H. Bach, Expression of heme oxygenase-1 can determine cardiac xenograft survival, Nature Med. 4 (1998) 1073–1077. H. Zinszner, M. Kuroda, X. Wang, N. Batchvarova, R.T. Lightfoot, H. Remotti, J.L. Stevens, D. Ron, CHOP is implicated in programmed cell death in response to impaired function of the endoplasmic reticulum, Genes Dev. 12 (1998) 982–995. C.D. Ferris, S.R. Jaffrey, A. Sawa, M. Takahashi, S.D. Brady, R.K. Barrow, S.A. Tysoe, H. Wolosker, D.E. Baranano, S. Dore, K.D. Poss, S.H. Snyder, Haem oxygenase-1 prevents cell death by regulating cellular iron, Nature Cell Biol. 1 (1999) 152–157. K.D. McCullough, J.L. Martindale, L.-O. Klotz, T.-Y. Aw, N.J. Holbrook, Gadd153 sensitizes cells to endoplasmic reticulum stress by down-regulating Bcl2 and perturbing the cellular redox state, Mol. Cell. Biol. 21 (2001) 1249 –1259. T. Kawai, M. Matsumoto, K. Takeda, H. Sanjo, S. Akira, ZIP kinase, a novel serine/threonine kinase which mediates apoptosis, Mol. Cell. Biol. 18 (1998) 1642–1651. G. Page, D. Ko¨ gel, V. Rangnekar, K.H. Scheidtmann, Interaction partners of Dlk/ZIP kinase: co-expression of Dlk/ZIP kinase and

320

¨ rd, T. O ¨ rd / Experimental Cell Research 286 (2003) 308 –320 D. O

Par-4 results in cytoplasmic retention and apoptosis, Oncogene 18 (1999) 7265–7273. [71] R.B. Nehring, H.P.M. Horikawa, O. El Far, M. Kneussel, J.H. Brandsta¨ tter, S. Stamm, E. Wischmeyer, H. Betz, A. Karschin, The metabotropic GABAB receptor directly interacts with the activating transcription factor 4, J. Biol. Chem. 275 (2000) 35185–35191. [72] J.H. White, R.A. McIllhinney, A. Wise, F. Ciruela, W.Y. Chan, P.C. Emson, A. Billinton, F.H. Marshall, The GABAB receptor interacts

directly with the related transcription factors CREB2 and ATFx, Proc. Natl. Acad. Sci. USA 97 (2000) 13967–13972. [73] A. Bonni, A. Brunet, A.E. West, S.R. Datta, M.A. Takasu, M.E. Greenberg, Cell survival promoted by the Ras-MAPK signaling pathway by transcription-dependent and -independent mechanisms, Science 286 (1999) 1358 –1362. [74] A. Riccio, S. Ahn, C.M. Davenport, J.A. Blendy, D.D. Ginty, Mediation by a CREB family transcription factor of NGF-dependent survival of sympathetic neurons, Science 286 (1999) 2358 –2361.