Early growth response-1 gene ( Egr-1 ) promoter induction by ionizing radiation in U87 malignant glioma cells in vitro Ralph G. Meyer 1,2 , Jan-Heiner KuÈ pper 2 , Reinhard Kandolf 2 and H. Peter Rodemann 1 1 Section of Radiobiology and Molecular Environmental Research, Department of Radiotherapy, and 2 Department of Molecular Pathology, University of Tu È bingen, Germany The promoter of the early growth response gene (Egr-1) has been described to b e activated by ionizing radiation, and it seems to be c lear that this process involves dierent mitogen activated protein (MAP) kinases, dependent on the speci®c cell type examined. However, early steps leading to activation of the corresponding pathways and thus to overexpression of Egr-1 are not well understood. In this study, deletion m utants of the 5¢ upstream region of the Egr-1 gene were generated which allowed us t o correlate the radiation±induction of the Egr-1 promoter in U87 g lioma cells to ®ve serum response elements. Based on the d ata shown, a possible role o f two cAMP responsive elements for radiation-dependent promoter regulation could be ruled out. O n t he basis of a ctivator/inhibitor studies applying fetal bovine serum, EGF, PD98059, anisomycin, SB203580, forskolin a nd w ortmannin, it couldbedemonstratedthatinU87cellstheERK1/2 and potentially SAPK/JNK, but not the p 38MAPK/ SAPK2, pathway contribute to the radiation-induction of Egr-1 promoter. In addition, it was observed that irradi- ated cells secrete a diusible factor into the culture media which accounts for the radiation-induced promoter upregulation. By blocking growth factor receptor activa- tion with suramin, this eect could be completely abolished. Keywords: Egr-1 promoter; growth factor receptor; glio- blastoma cells; ionizing radiation. The immediate e arly gene Egr-1 (synonyms are NG FI-A, zif268, TIS8 and kro x24) encodes a transcription factor involved in cell growth and differentiation. The DNA binding domain of this protein with its three zinc-®nger motifs allows speci®c binding to GC rich recognition sequences in the promoters of many downstream genes and thus r egulation of their expression. Target genes of the Egr-1 transcription factor are numerous and include growth factors such as platelet derived growth factor A chain (PDGF-A [1]), PDGF-B chain [ 2], b asic ®broblast growth factor (bFGF) [3], cytokines such as TGF-b [4] and other proteins that can be affected. Expression of Egr-1 i tself is t ransiently induced by a variety of extracellular stimuli such as cytokines, growth factors s uch as FGF-2 [5], hypoxia [6], shear s tress on vascular cells caused by blood current [7], tissue injury and physical stress in¯icted by ionizing radiation. Publication of the latter ®nding [8] initiated investigations on how the Egr-1 promoter may be employed in cancer gene therapy approaches, controlling ectopic expression of therapeutic genes with the application of X-rays t o target tissues [9±13]. However, initial s teps of the underlying mechanism of the observed radio-induction of Egr-1 expression are not completely understood, although further signal transduc- tion pathways involved in this gene activation are, at least in part, well characterized. Functional dissection of the Egr-1 promoter sequence [14±16] r evealed in previous studies t hat it contains at least ®ve copies [17] of a characteristic transcription factor binding site designated the serum response element (SRE), which is described to be respon- sible for radioinducibility of the ge ne. Due to its consensus sequence CC(A/T) 6 GG this motif is also known as t he ÔCArGÕ element, and represents a combined recognition site for Elk-1, a member of the Ets family which acts a s a ternary complex f actor (TCF) in concert with other transcription factors, mainly p68/SRF s erum response factor [18]. The promoter is strongly activated upon binding of Elk-1/SRF transcription factor complexes to CArG elements [19]. Complex assembly requires phosphorylation of b oth SRF and Elk-1 by speci®c kinases which are, at least in the case of Elk-1, in turn dependent on prior activation of mitogen activated p rotein kinases (MAPK) downstream of different signal transduction pathways. At least three differently regulated but partly overlapping kinase cascade pathways Correspondence to H. Peter Rodemann, Section of Radiobiology and Molecular Environmental Research, Department of Radiation Oncology, University of Tu È bingen, Ro È ntgenweg 11, D-72076 Tu È bingen, Germany. Fax: + 49 7071 29 5900, Tel.: + 49 7071 29 8596 2, E-mail: hans-peter.rodemann@uni-tuebingen.de Abbreviations: Egr-1, early growth response-1 gene; PDGF, platelet derived growth factor; bFGF, basic ®broblast growth factor; SRE, serum response element; MAPK, mitogen activated protein kinases; TCF, ternary complex factor; SAPK, stress activated protein kinases; Raf, ras-activated factor; E RK, extracellularly regulated kinase; PLC, phospholipase C; FGF, ®broblast growth factor; AP-1, activated protein-1; TRE, thiophorbolester responsive element; EgrBS, Egr-1 binding site; CRE, cAMP responsive element; MCS, multiple cloning site; RSV, Rous sarcoma virus; rhEGF, recombinant human epider- mal growth factor; PtdIns3-kinase, phosphatidylinositol 3-kinase; b-Gal, b-galactosidase; PVDF, poly(vinylidene di¯uoride); ECL, enhanced chemoluminescence; IL, interleukin. De®nition: 1 gray (Gy)  100 rads. (Received 18 July 2001, revised 26 October 2001, accepted 6 November 2001) Eur. J. Biochem. 269, 337±346 (2002) Ó FEBS 2002 converge in their activity on the phosphorylation of E lk-1, including the c ascade leading to the activation of stress activated protein kinases (JNK/SAPK), the ras-activated factor/extracellularly regulated kinase (Raf/ERK) pathway and the activation of the p38 MAPK/SAPK2 pathway. Independently of these pathways, protein kinase C (PKC) is able to phosphorylate E lk-1 directly. PKC is activated by diacylglycerol, which is formed by phospholipase C (PLC) mediated phosphatidylinositol 4,5-diphosphate cleavage. PLC in turn is activated by binding to intracellular d omains of activated g rowth facto r receptors, e.g. the ®broblast growth factor (FGF) receptor. In addition to SREs the 700-bp full length Egr-1 promoter comprises several kinds of other binding sites including a JNK/SAPK dependent activated protein-1/thiophorbolester responsive element (AP-1/TRE) site, binding sites f or Egr-1 itself (EgrBS) and two cAMP responsive elements (CRE). For g ene t herapy purposes a core promoter of 490 bp (nucleotides )42 5 t o + 65 rela tive to the putative transcription start [9,10], w as described to be suf®cient for radioactivation. This includes only the SREs, Sp1 sites a nd the t wo CREs (see Fig. 1). Whether or not CRE sites contribute to radioinducibility of the Egr-1 promoter in normal cells was not clearly demonstrated to date, although t here is evidence that this is the case i n ras- mutated J urkat cells which exhibit impaired MAP kinase pathways [20]. In order to i nvestigate the phenomenon of radiation induction of the human Egr-1 promoter, glioblastoma cells U87 were transiently transfected with expression plasmids containing a reporter gene under the control of wild-type and recombinant versions of the human Egr-1 promoter. This system was used to investigate the Egr-1 promoter regulation under different experimental condi- tions. We show that Egr-1 promoter can be induced by a single dose of 4 Gy. The data presented indicate that radiation induction of the Egr-1 promoter is at least in part mediated by protein factors secreted by U87 cells in response t o radiation exposure. The data are discussed in the context o f the potential use o f Egr-1 pro moter for radiation-induced gene therapy strategies in radiation oncology. MATERIALS AND METHODS Cloning of Egr-1 promoter variants A 780-b p fragment of the human Egr-1 promoter was obtained by double restriction digestion of plasmid pGL/TiS8 [15] with SmaI/HindIII, with overhanging 5¢ ends ®lled up by using T4 DNA polymerase (Roche, Mannheim, Germany). T he promoter fragm ent w as then ligated blunt into pCR-S cript SK+ (Stratagene). The resulting p lasmid pEgr was used for sequencing a nd as a source of all following promoter variants which were generated by conventional plasmid construction methods using t he restriction enzymes depicted in Fig. 1. As an exception pD7egr was generated by site directed mutagenesis of pD5egr (see below). Plasmid pD6egr is similar to pD5egr but contains an additional cluster of SRE/Ets sites in a fragment which was generated by cutting pD3egr with Eco47III/AccIII and polishing the ends of the 207 bp fragment with T4 DNA polymerase. All promoter versions were excised from their host plasmids and cloned into pGFL cut with Sm aI. The resulting plasmids were desig- nated pwtegrGFL, pD1egrGFL, pD2egrGFL, etc. Expression plasmid pGFL is based on pGL3basic (Promega) but the luciferase gene was replaced by an in-frame ®re¯y luciferase/EGFP fusion with the original multiple cloning site (MCS) and the synthetic upstream 5¢ polyadenylation signal which blocks unspeci®c transcrip- tion activation. The presence of EGFP as part of the luciferase gene allows additional comparison of transfection ef®ciencies in parallel cultures. Physical properties of the novel GFL protein were investigated in several sets of plasmid t ransfection experiments which s howed, that in the fusion protein luciferase a ctivity was constantly lowered to  70% of the unfused luciferase gene. Also, EGFP ¯uores- cence was decreased to an estimated  65±70% of free EGFP. Extensive tests, including complete sets of experi- ments described in this study were performed in order to ensure that there was no alteration of promoter behaviour in corresponding lucifer ase/luciferase±EGFP vectors ( data not shown). All cloning steps were controlled and veri®ed by restriction analysis. Fig. 1. The human Egr-1 promoter and its regulatory e lements as cloned into p wtegrGFL. As a key region the s equence f rom nucleotides )136 to )56 is shown in detail, comp rising two serum response elemen ts (SRE) ¯anked by two cAMP responsive elements (CRE). Oth er bi nding sites are located upstream of SRE 1±3, n amely three Sp1 binding s ites, two reco gnition sequences fo r the Egr-1 gene product itself (EgrBS) and an AP1 binding site. 338 R. G. Meyer et al. (Eur. J. Biochem. 269) Ó FEBS 2002 Site-directed mutagenesis Plasmid pD7egrGFL was generated by exchange of ®ve nucleotides in CRE2 (nucleotides )71 to )58) from ACGTC t o CTCAT by site directed mutagenesis u sing a kit (Stratagene). The primer sequence (5 ¢)3¢) w as C CCA TATATGCCATGTCTCATCACGACGGAGGCGG. Successful mutagenesis was con®rmed by sequence analysis. The de®ciency of this altered sequence to bind CREB was previously published [18]. As a positive control a well characterized Rous Sarcoma Virus (RSV) promoter, excised from pAdRSVbgal, was inserted into pGFL, resulting i n pRSVGFL and sequenced. This promoter stems from rous sarcoma virus, strain Schmidt±Ruppin (EMBL database accession no. L29198). Cell culture and transfection procedures/irradiation of transfected cells Human U87 glioma cells were kept in DMEM (Gibco), supplemented with 10% fetal bovine serum (Gibco), under standard cell culture conditions (5% CO 2 ,37°C). FuGene (Boehringer Mannheim) transfections were performed a ccording t o i nstructions of the m anufacturer. Transfection ef®ciencies were determined by transfection of plasmid pCMVb (Clontech) into U87 cells; 48 h after transfection cells were stained for b-galactosidase (b-Gal) activity using the b-Gal staining kit purchased from Invitrogen. Routine luciferase activity measurement procedures were performed according t o the following procedure. Cells were plated at a density of 10 5 cells per 35 mm dish and allowed to attach for 24 h. FuGene transfections were performed by using 900 ng pDxEgrGFL per cell culture dish. pCMVb (100 ng; Clontech) was added to each culture dish for internal control of transfection ef®ciency. N egative controls were performed by transfecting p romoterless p GFL a s w ell as pRSVGFL constructs in parallel. As tested by pilot experiments, cotransfection of pCMVb did not in¯uence basal Egr-1 or RSV promoter activity s igni®cantly. Twenty- four hours after transfection, cells were exposed to a single dose of 4 Gy of ionizing radiation generated by a linear accelerator (Mevatron 6MeV, LINAC) as described else- where [20,21]. Cells were harvested 48 h aft er irradiation and assayed for luciferase activity. In experiments using speci®c effectors of Egr-1 promoter activity these factors/substances were added 6 h prior to cell harvest (i.e. after 42 h). Inhibitor and activator studies Speci®c activators and inhibitors to signal transduction pathways, except for fetal bovine serum, were obtained from Calbiochem and dissolved to the following working solutions. Fetal bovine serum. fetal bovine serum (Gibco Life Technologies) was added to a ®nal concentration of 30% in the culture medium serving as a positive control for ERK1/2 activation. Recombinant human epidermal growth factor (rhE- GF). Stock solutions at a concentration of 100 lgámL )1 in an aqueous solution of 0.3% BSA with 10 m M acetic acid were added to t he culture m edia at a ® nal c oncentration of 100 ngámL )1 . r hEGF was used as a positive control for the immediate early activation of the Egr-1 promoter by activated Ras-dependent pathways. PD98059. This was added as a 50-m M solution in dimethylsulfoxide t o the media making a ®nal co ncentration of 100 l M . This c ompound is a potent inhibitor of MEK1 (IC 50  5±10 l M ) a nd to a lesser extent of MEK2 [ 22,23], two upstream protein kinases which activate ERK1/2. Anisomycin. This was dissolved in dimethylsulfoxide (10 mgámL )1 ) and immediately added to the culture media at a ®nal concentration of 50 l M . Anisomycin (from Streptomyces griseolus) activates p38MAPK/SAPK2 and SAPK/JNK and served as a positive control in this study. It is also an inhibitor of protein expression at the translational level. SB203580. This was added as a 1-m M working solution in dimethylsulfoxide to the cells to a ®nal concentration of 10 l M . It was used to inhibit p38MAPK/SAPK2 activity. Forskolin. This was prepared a s a 10-m M stock solution in dimethylsulfoxide and directly added to the culture medium to m ake a ®nal concentration of 10 l M .Beinga strong and speci®c activator of a denylate cyclase, forsko- lin induces increased levels of cAMP in treated cells. This in turn activates cAMP-dependent protein kinase A (PKA) which phosphorylates CREB and other target proteins. Wortmannin. A1m M stock solution in dimethylsulfoxide was further diluted 1 : 100 in NaCl/P i and ®nally added as a 10-l M working solution with a ®nal c oncentration in the media of 100 n M . As a speci®c inhibitor o f phosphat- idylinositol 3-kinase (PtdIns3-kinase, IC 50  5n M )wort- mannin a llows us to investigate d o wnstream signalling of the P tdIns3 pathway and dependent activation e vents. Suramin. Suramin (Sigma, St Louis, MO, USA) was added as an aqueous 3 0 m M stock to the media t o make a ®nal concentration o f 300 l M . This substance interferes w ith t he recognition of several growth factors by their membrane receptors and, in addition, disrupts the interaction of growth factor receptors with corresponding adenylate cyclase activating G-proteins. For control conditions, cells were treated with equal amounts of the corresponding solvent (e.g. dimethylsulf- oxide). Cell lysis and quanti®cation of reporter gene activities Cell monolayers were scraped off with a rubber policeman in 100 lL of r eporter lysis buffer ( Promega), transferred t o 1.5 mL reaction t ubes, centrifuged (13 000 g for 60 s) in a conventional bench-top centrifuge and kept on ice until measurements of luciferase activity, b-Gal activity and protein content which were performed f rom the same cell aliquots. Protein content was estimated using Bradford's reagent (Biorad), b-Gal act ivity was measured by employing a Ó FEBS 2002 Egr-1 promoter induction by ionizing radiation (Eur. J. Biochem. 269) 339 commercial kit (Invitrogen). Luciferase activity was deter- mined with a kit from Promega and measured in a luminometer (Berthold); m easured values were normalized to speci®c b-Gal activity. Egr-1 promoter stimulation by ionizing radiation or chemical compounds was performed only for the three major promoter mutant plasmids pD1egrGFL, pD5egrGFL and pD7EgrGFL. All transfec- tion experiments were performed in triplicates an d repro- duced at least three times. Immunoblotting and detection of activated protein factors with phospho-speci®c antibodies For Western blot analyses cells were cultured in 60 mm dishes under n ormal cell culture conditions. T reatment with speci®c inhibitors or exposure to a single dose of 4 Gy o f c-irradiation were performed as described above. Thirty minutes after treatment with activators or inhibitors or radiation exposure, cells were washed twice with NaCl/P i , SDS sample buffer was added and cells scraped off the plate and subjected t o denaturing SDS/PAGE. Proteins were subsequently blotted to a poly(vinylidene di¯uoride) (PVDF) membrane and detected using phosphospeci®c or phosphorylation-state-independent antibodies according to the recommendations of th e manufacturer ( Cell Signaling Technology, New England Biolabs). As s econdary antibod- ies, monoclonal horseradish coupled anti-(rabbit I g) Ig or anti-(mouse Ig) Ig was used allowing identi®cation of protein bands using enhanced chemoluminescence (ECL) detection. For detection of site-speci®c phosphorylated target proteins, i.e. ERK1/2 (T202/Y204), SAPK/JNK (T183/Y185), p38MAPK/SAPK2 (T180/Y182), ATF-2 (T71), c-Jun (S73) and CREB (S133) phosphospeci®c rabbit polyclonal Ig w as applied. Phosphorylation-state-indepen- dent polyclonal antibodies were used to detect total Egr-1, ATF-2 and c-Jun protein. RESULTS Basal activity of Egr-1 promoter variants Basal activities of cloned variants o f the Egr-1 promoter were analyzed by measuring the luciferase reporter gene activity. Transfection experiments with pwtegrGFL showed that basal activity of t he wild-type Egr-1 promoter construct was roughly 10±12% of the activity displayed b y a control RSV promoter. As demonstrated in Fig. 2, the relative activity of promoter variant p D1egrGFL ( nucleotides )47 4 t o +12) did not differ signi®cantly from that of the wild-type promoter construct pwtegrGFL (nucleotides )720 to +12). These data indicate that regulatory promoter ele- ments upstream of nucleotide )474 do not contribute signi®cantly to basal promoter activity under the conditions applied. In contrast however, deletion o f nucleotides )259 to )126 which results in promoter construct pD3egrGFL led to a s mall but signi®cant increase in relative basal luciferase activity (110.5  2.3% of control activity, P < 0.05). A similar, but more pronounced, effect was observed with promoter variant pD5egrGFL also lacking nucleotides )259 to )126 but, as compared to pD3egrGFL, additionally missing the sequence )474 to +12. This promoter variant presented a signi®cantly( p < 0 .005) enhanced relative basal activity of 137.1  5.7%. These data indicated, that dele- tion of the CRE1 site and of the putative Sp1 recognition site results in an upre gulation of basal Eg r-1 p romoter a ctivity. Insertion of an additional cluster of two SRE/Ets binding sites ( nucleotides )474 to )265) into pD5egrGFL resulting Fig. 2. Basal activity o f Egr-1 promoter variants in U87 glioma cells. Deletion mutants of the human Egr-1 promoter were generated i n order t o con®ne the r adio-inductio n ee ct of the promoter to speci®c tran scription f actor b inding sites. For comparison of promo ter ac tivity i n a no nind uced state, resul ting p ro moter v ariant s D1egr, D2egr, D3egr, D4egr, D5egr, D6egr and D7egr were cloned into pGFL, carrying an in-frame fusion of EGFP and ®re¯y luciferase (termed GFL) and transfected in to U87 c ells. Luciferase activities under n ormal cell culture conditions were considered as basal promoter activities, w hich were compared t o wild-type Egr-1 p romoter activity. 340 R. G. Meyer et al. (Eur. J. Biochem. 269) Ó FEBS 2002 in the variant pD6egrGFL, led to a further but, as compared to pD5egrGFL, nonsigni®cant increase i n b asal promoter activity (158  9 .1%). Removal of this cluster of SRE/Ets bindings sites, as performed i n p D2egrGFL a nd p D4egrGFL, alway s led to a dramatic decrease of basal promoter performance to 11.1  0.5% (pD2egrGFL) and 10.5  0.3% (pD4egr- GFL) of wild-type promoter activity. Site directed mutagenesis of CRE2 in pD5Egr resulted in promoter variant pD7Egr, which contains only the SRE and Ets binding sites of the wild-type promoter. p D7egr- GFL presented a signi®cantly (P < 0.005) reduced pro- moter activity which was 61.5  2.8% of the wild-type promoter. Immunoblotting of U87 cell lysates followed by detec- tion of activated ERK1/2 with phosphospeci®c ERK1/2 (T202/Y204) a ntibodies demonstrated detectable amounts of activated ERK1/2 already formed under normal cell culture conditions. Relatively high c oncentrations of phos- phorylated CREB (S133) and ATF-1 were detected, whereas there appeared to be no detectable amounts of activated ATF-2 nor SAPK/JNK. Effect of serum, rhEGF and ionizing radiation on promoter activity To analyse the effect of fetal bovine serum and human recombinant EGF on Egr-1 promoter activity the promoter variants D1egr, D5egr and D7egr were used. Increasing the concentration of fetal bovine serum in the culture medium to 30% resulted in an ef®cient induction of all three promoter plasmids tested (pD1egrGFL pD5egrGFL and pD7egrGFL) within a 6-h interval of application. As shown in Fig. 3 t he corresponding luciferase activity increased to 157.3  12.6% (pD1egrGFL), 149.2  13.4% (pD5egr- GFL), or 206.6  33.1% (pD7egrGFL), respectively, as compared to wild-type activity. Treatment of U87 cells carrying these promoter con- structs with 1 0 ngámL )1 rhEGF resulted i n a pronounced stimulation of promoter activities to 205.5  18.8% in pD1egrGFL transfectants, to 151.9  11.4% in pD5egr- GFL transfectants and to 2 11.5  22.4% to pD7egr GFL transfectants as compared to cells transfected with the wild- type promoter construct (Fig. 3). As indicated by time kinetic experiments, increased l evels of ®re¯y luciferase activity due to the exposure of pwtegrGFL-, pD1egrGFL-, pD5egrGFL-, and pD7eg- rGFL-transfected U87 cells to a single dose of 4Gy of ionizing irradiation could be observed to be maximal after 40±48 h post IR (data not shown). As compared to sham-irradiated controls luciferase activities measured 48 h post IR were increased signi®cantly to 133.7  4.3% for pD1egrGFL-, 119.1  5.0% for pD5egrGFL and 138.7  4.2% for pD7egrGFL-transfected U87 cells (all P <<0.005). Protein kinase inhibitor/activator studies As shown in Fig. 3, a 6-h treatment of U87 cells transfected with the three Egr-1 promoter mutants with PD98059 (100 l M ), a speci®c inhibitor of MEK1, decreased relative luciferase activities by about 20±30% as compared to wild-type control levels (pD1egrGFL: 82.6  3.5%; pD5egrGFL: 67.6  4.8%; pD7egrGFL: 72.0  4.7%). Under the same treatment conditions anisomycin, a speci®c activator of p38MAPK/SAPK and SAPK/JNK, did not alter the luciferase activity of pD1egrGFL (relative activity: 103.3  6.8%) and pD7egrGFL (relative activity: 106.5  19.8%) as compared to wild-type controls. How- ever, anisomycin resulte d in a signi®cant (P < 0.005) inhibition of the pD5egrGFL luciferase a ctivity by a bout 26% (relative activity: 74  2.1%). Treatment with SB203580, which i s a potent inhibitor of p38MAPK/SAPK, resulted in a pronounced promoter activation of pD1eg- rGFL (131.7  9.5%, P<0.005) and pD7egrGFL (124.2  12.3%). In contrast to these results a signi®cant decrease in luciferase activity by about 27% could be Fig. 3. D1egr, D5egr and D7egr promoter activities as a function of inhibitors/activators of speci®c signal tra nsduction pathways. U87 cells were transfected w ith reporter gen e constructs containing an i n-frame EGFP±luciferase gene under the control of modi®ed Egr-1 promoters (see F ig. 1) using FuGene transfection reagent (Roche, Mannheim, Germany). Forty-two hours after the start o f t ransfection cells were treated with chemical eectors as indicated. Cells were further incubated for 6 h and assayed for luciferase activity 48 h post transfection. Irradiated cells were exposed to a single dose of 4 Gy of ionizing radiation 16 h after the start o f transfection and luciferase activity was deter mined 48 h postirradiation. Cotransfected plasmid pCMVbGal (Clontech) was used as an internal control of transfection eciency in all samples. Data represent the mean  SE from th ree to six exp eriments performed in triplicates. Ó FEBS 2002 Egr-1 promoter induction by ionizing radiation (Eur. J. Biochem. 269) 341 observedinpD5egrGFL-transfected cells ( relative activity: 73.1  6.6%; P < 0.005) (Fig. 3). Treatment w ith 1 0 l M forskolin, a speci®c activ ator o f adenylate c yclase, did not signi®cantly alter the promoter activity in cells transfected with the luciferase reporter constructs pD1egrGFL and pD7egrGFL, but l ed to a slight, but signi®cant decrease to 84.2  8.9% (P <0.05) of relative activity in cells transfected with pD5egrGFL (Fig. 3 ). For all three promoter variants tested, treatment of the transfected cells with wortmannin, a speci®c inhibitor of PtdIns3 k inase, resulted in a m easurable downregulation, which ranged between 10% and 17% (pD1egrGFL: 84.0  6.8%; pD5egrGFL: 82.9  15.3%; pD7egrGFL: 90.8%  4.1%) within a 6-h i nterval. In order to a nalyse, whether factors produced and secreted in response to radiation exposure may alter the activity levels of the promoter variant construct pD7egrGFL in transfected cells, a cross feeding exp eriment as described in Materials and methods was performed. When culture media from irradiated was f ed to nonirradiated cells transfected with promoter variant pD7egrGFL the luciferase activity was increased to approximately the same level (130.6  4.8%) as in irradiated cells (132  10.2%). Levels of activity in the irradiated cells did not decrease signi®cantly after a ddition of control medium f or up to 6 h, indicating a l onger lasting s ecretion of activator molecules or growth factors. Adding suramin, a potent inhibitor of growth factor receptor±ligand binding, to the culture medium of irradiated cells or nonirradiated cells fed with medium from irradiated cells the stimulatory effect on luciferase activity of promoter variant plasmid pD7egrGFL could c ompletely b e abolished (87.49  10.15%, 0 Gy a nd 84.89  11.87%, 4 Gy, Fig. 4). Western-blot analyses For the interpretation of the regulatory function of the different Egr-1 promoter elements and t he role of speci®c signal transduction pathways in activating the Egr-1 pro- moter with a nd without radiation exposure W estern blot analyses of untransfected cells for different target proteins and their phosporylation status were performed. Therefore, protein extracts from the same set of cells used for the analyses of luciferase activity under different treatment conditions with serum, rhEGF and ionizing radiation as well as inhibitors or activators of speci®c protein kinases were used. Fetal bovine serum treatment. After treatment with 30% fetal bovine s erum increased ERK1/2 phosphorylation w as shown by immunoblot analysis (Fig. 5). This could account for the promoter induction shown in Fig. 3. While there was slight activation of c-Jun ( S73), ATF-2 phosphorylation did not seem to be affected by elevated fetal bovine serum concentration (Fig. 5). RhEGF-treatment. Treatment with 10 ngámL )1 rhEGF resulted in the strongest phosphorylation a nd activation of ERK1/2 (T202/Y204) (Fig. 5 ), re¯ecting t he highest Egr-1 promoter induction as shown for all mutants tested (Fig. 3). Immunoblots incubated with phospho-speci®c antibodies revealed increased levels of phosphorylated CREB and ATF-1 (S133), S APK/JNK ( T183/Y185) as well as upre- gulation at the translational level and activation of ATF-2 (T71) and c-Jun (S73). An increase in p38MAPK/SAPK2 (T180/Y182) activation was not observed. Radiation exposure. After exposure to ionizing radiation, phosphorylated SAPK/JNK was detectable, but to a much lesser degree t han it was observed a fter rhEGF treatment (Fig. 5). Activation of p38MAPK/SAPK2 could not be detected and ERK1/2 phospho rylation was slightly lower than in control samples. A s a result, there were only low amounts of phosphorylated CREB (S133); however, the level of phosphorylated ATF-1 remained unchanged as compared to controls. Furthermore, radiation exposure resulted in high amounts of activated ATF-2 and c-Jun. As it could be demonstrated using speci®c phosphorylation- independent antibodies as controls for ATF-2 (T71) and c-Jun (S73) levels the expression of both proteins was strongly upregulated. PD98059. As illustrated in Fig. 5, PD98059 as an inhibitor of MEK1, clearly prevented activation of MEK1 (IC 50  5±10 l M ). As ERK1 and ERK2 are activated by MEK1, treatment of U87 cells with 100 l M PD98059 speci®cally abolished the phosphorylation of ERK1/2 without affecting t he activity status of ATF-1, ATF-2, SAPK/JNK, and c-Jun. Anisomycin. Exposure of U87 cells to 10 l M anisomycin did not result in marked changes in phosphorylation status of ERK1/2. The amount of phospho-SAPK/JNK (T183/Y185) and especially o f phospho-p38MAPK/SAPK2 (T180/Y182) were strongly increased ( Fig. 5). T his activa- tion coincided with high l evels of phosphorylated CREB, ATF-1 and ATF-2; however, phosphorylation or activation Fig. 4. Suramin inhibits radiation-induced Egr-1 p romoter activation. After transient transfection with plasmid pD7egr, U87 cells were irra- diated (4 Gy) and in cu bated without subsequent exchange of c ulture media (Control). After 42 h, media were removed from irradiated cells and replaced by media from parallel, unirradiated, transfected cells (A2). The un irradiated cultures, in turn, re ceived media from corre- sponding irradiated cultures (A1). The addition of culture media from irradiatedcellsledtoanincreaseinluciferaseactivityintheunirradi- ated cells (A1) up to the l ev el reached by irradiated cells (control cells and A2). Addition of 300 l M suramin abolished the eect (B1, B2). Cells were harvested after a t otal time interval of 48 h and luciferase activities were determined. Data represent the mean  SE from three independent experiments performed at least in triplicates. 342 R. G. Meyer et al. (Eur. J. Biochem. 269) Ó FEBS 2002 of c-Jun (S73) was only marginal. By using a pan-c-Jun antibody, which recognizes c-Jun independently of its phosphorylation status, a partial phosphorylation of the c-Jun protein, presumably in position S63, could be demonstrated. SB203580. As control cells, grown under standard conditions, presented basically no phosphorylated p38MAPK/SAPK2, no alterations of the phosphorylation pro®le was to be expected by treating the cells with the p38MAP-kinase inhibitor SB203580. However, potentially as side-effects which have also been described i n the recent literature, SB203580 caused st rong activation of E RK1/2 (T202/Y204) via activation of Raf1 [24±26] and inhibition of CREB/ATF1 activation as shown in immunoblots at a concentration of 10 l M . In this context it remains to be addressed in more detail whether the profoun d promoter a ctivation of p D1egrGFL and pD7egrGFL shown in Fig. 3 is related to t he Western blotting results. Forskolin. While forskolin had no effect on the phospho- rylation of SAPK/JNK, p38MAPK/SAPK2, ATF-2 or c-Jun, a marked increase in phospho-CREB (S133) levels could be observed ( Fig. 5). A s CREB phosphorylation i s a cAMP-dependent reaction [27], this observation proves functionality of the drug. A s shown in Fig. 5 in agreement with results from o thers [28] forskolin abolis hed ERK1/2 phosphorylation in U87 cells. The extent of inhibition was comparable to that shown above for PD98059. Wortmannin. As a consequence of w ortmannin treatment no activation of SAPK/JNK or p38MAPK/SAPK2 was observable and the amounts of phosphorylated ERK1/2 were slightly diminished. Additionally no effect on ATF-2 phoshorylation or expression could be observed; however, reduced levels of activated CREB, ATF-1 and c-Jun were present after treatment with wortmannin, whereas no phosphorylated c-Jun could be detected. DISCUSSION One of the main aims of this work was to investigate mechanisms of radiation induction of the human Egr-1 promoter and its potential use for radiation induced gene therapy as recently reported [9,29]. While it is ®rmly established, that the Egr-1 promoter is effectively and quickly activated by growth factors, such as EGF, bFGF and other serum compounds [5,30,31], mechanisms leading to induction by radiation are by far less well understood. Most likely both, growth factor and radiation stimuli ultimately lead to the binding of SRF and TCF/Elk-1 to SREs which are recognized as overlapping CArG/Ets binding sites forming the core promoter [14±17]. By the Fig. 5. Immunoblot analyses of activated protein factors in U87 cells using polyclonal phosphospeci®c antibodies. U87 cells were subjected to various treatments with chemical compounds for a time interval of 30 min and then lysed. Cells exposed to 4 Gy of ionizing radiation were incubated for 30 min prior to harvest. Lysates were r esolved by SDS/PAGE and subsequent immunoblotting with polyclonal rabb it antisera directed speci®cally against the phosphorylated proteins i ndicated (phosphorylated amino-acid residues are given in b racket s). Phosphorylation-state-independent rabbit polyclonal antib odies recognized total E gr-1, ATF-2 and c-Jun protein. Ó FEBS 2002 Egr-1 promoter induction by ionizing radiation (Eur. J. Biochem. 269) 343 stepwise removal all other known regulatory e lements from the 5¢ upstream region of the wild-type Egr-1- promoter, such as CREs as well as AP1-, Egr-1- and Sp1- binding sites, we were able to provide additional evidence s upporting the prevailing view that these binding sites are suf®cient to maintain responsiveness of the Egr-1 promoter t o EGF and ionizing radiation. In general, the radiation-dependent Egr-1 promoter upregulation in U 87 glioma cells observed in t he presen t study was weak but signi®cant. A single dose of 4 Gy upregulated the promoter variant pD7egrGFL by a factor of 1.4 (p < 0 .05). Similar data of induction of the Egr-1 promoter by ionizing radiation h ave been reported recently [29]. Using synthetic promoter constructs consisting merely of r epetitive SRE consensus sequences Marples et al.[29] described an upregulation of Egr-1 by a factor of 1.5±2.5 after radiation exposure of U87 cells. Based on literature data [32] it could be expected that stimulation w ith EGF results in t he phosphorylation and thus binding of TCF/Elk-1 to SRE through activation of the MAP kinases ERK1/2 or SAPK/JNK or both. However, in our hands strong activation of p38MAPK/SAPK2 was not observed in U87 cells except after anisomycin treatme nt. Immunoblot analyses showed activation of ERK1/2 after stimulation of U87 c ells with fe tal bovine serum, E GF and SB203580 but not after e xposure to i onizing radiation. This data indicate that ERK1/2 and SAPK/JNK pathways may be activated apart from each other. Both, SAPK/JNK and ERK1/2, are activated upon growth factors binding to their receptors. However, signal tran sduction leading to JNK/SAPK activation i s known to b e mainly triggered by ultraviolet light (UV-C) [34±36], ionizing radiation (reviewed in [ 37]), p roin¯ammatory c ytokines [30,31] a nd DNA damaging agents [37], whereas the Raf/ERK pathway is preferentially activated upo n binding of growth factors t o receptor tyrosine kinases ([18] and many others). While there i s some overlap of both pathways our data based on EGF treatment and protein kinase inhibitor studies suggest a clear preference of the ERK1/2 pathway after binding of EGF to its receptor, whereas ionizing radiation seems t o favour SAPK/JNK activation as d em- onstrated by the pronounced c-Jun activation observed by Western blot a nalyses. As a clear immediate early reaction after exposure to ionizing radiation a strong increase in c-Jun activation and expr ession was apparent. A lthough in our study no evidence has been obtained that Egr-1 immediate early gene induction could be observed 30 min after exposure t o i onizing radiation, Liu et al. [ 38] reported that exposure of serum-depleted, quiescent U87 cells presented a slight induction of Egr-1 2 h after exposure to UVC. Forskolin inhibited ERK1/2 phosphorylation under normal cell c ulture conditions, but had no signi®cant effect on Egr-1 promoter activity i n our experimental setup. As it has also been reported by others [28] as a speci®c side-effect of forskolin this compound ab olished ERK1/2 phosphory- lation in the U87 cells used in the present study. Therefore, this ®nding supports the view, that Egr-1 promoter activity is not strictly dependent on ERK1/2 activity, but may in part be regulated b y the SAPK/JNK [39,40] or the PKC pathway. Furthermore, in the present study we were able to show that binding of EGF to its receptor led to strong Egr-1 expression U87 cells transfected with Egr-1 reporter constructs. On t he other hand, expression of ATF-2 and c-Jun was not signi®cantly affected after EGF treatment o f these cells, but strongly induced after exposure to i onizing radiation. The fact that ATF-2 and c-Jun were phospho- rylated to a great extent due to the radiation exposure, provides good evidence for the activation of an intact SAPK/JNK pathway, as this is the major kinase activating the two transcription factors [35,41 ]. Secretion of bFGF and interleukin-1a (IL-1a) has been reported after U V-irradiation of H eLa or normal ®broblast cells [42]. This phenomenon may help to transmit a radiation-induced signal to nonirradiated cells. Moreover it may establish an obligatory growth factor l oop on the producer cell, both leading to expression of immediate early genes, e.g. c-Jun, via the activation of S APK/JNK [42,43]. Growth factor-mediated early step in radiation-induced cell signaling could be inhibited either by speci®c antibodies directed against b FGF or IL-1a as well as by preincubation with suramin [42,43]. These results although obtained after irradiating cells with UV light are in perfect agreement with the immunoblot analyses and suramin inhibition experi- ments presented in our study indicating an auto- and/or paracrine growth factor dependency of Egr-1 promoter activation by ionizing radiation. However, in contrast to the immediate-early expression observed for c-Jun, r adiation- induction of Egr-1 seems to be a rather slow, long-lasting and potentially SAPK/JNK independent phenomenon which may involve additional, so far unknown, mecha- nisms. Based on data r eported by Woloschak et al.[44]itis very likely, that activation of protein kinase C (PKC) a dds to the phosphorylation of Elk-1 and therefore to the induction of Egr-1 expression in response to radiation exposure. With respect to t hese results and our own data it can be assumed, however, that growth factor r elease from irradiated cells unequivocally contributes to radiation- mediated Egr-1 expression via the ERK1/2 or the PKC pathway. However, how the secretion into the media is triggered still remains an open question a nd needs to b e investigated further. As a second aspect of Egr-1 gene regulation, our data indicate a strong impact of CRE sites on basal promoter activity. T his suggests, that the w ild-type promoter in U87 glioma cells is in part also dependent on the phosphoryla- tion state o f CREB, ATF-1, ATF-2 and c-Jun p roteins. As reported by Van Dam et al. [ 39], activation o f ATF-2 upon genotoxic stress is mediated preferentially by SAPK/JNK, leading to the induction of c-Jun expression. According t o our data, the SAPK/JNK pathway transmits at least part of Egr-1 radiation-induction. The CRE sites m ay thus be involved in the radiation-dependent Egr-1 expression. Indeed, stress induced activation o f Egr-1 gene expression via promoter-CRE sites has b een described for ras-mutated Jurkat cells which present a n impaired Ras/Raf/ERK pathway [20]. However, under the experimental conditions applied in t his investigation, a speci®c radiation-dependent regulation of CRE sites was not observed. As indicated by the deletion of CRE2 in the pD7egrGFL mutant, this element positively regulates Egr-1 promoter activity (see Fig. 2). Inhibition of p38MAPK/SAPK2 dependent activation of these proteins may therefore account for the low pD5egrGFL promoter activity after SB203580 incubation. This promoter construct contains only one CRE2 element. Deletion of CRE1 as performed in 344 R. G. Meyer et al. (Eur. J. Biochem. 269) Ó FEBS 2002 pD5egrGFL revealed its role as a negative r egulatory element, which is in line with previously published w ork [15,46]. The p resence of both, an activating CRE2 and an inhibitory CRE1 in pD1egrGFL m ay therefore neutralize the effects of S B203580 on CREB/ATF1 phosphorylation. Thus, in pD1egrGFL and pD7egrGFL transfected U87 cells SB203580 induces increased luciferase activity merely due to ERK1/2 activation. Taken together in the present s tudy we provide evidence for the ra diation-induction of the Egr-1 promoter and the regulatory signal transduction pathways involved in this activation. While no evidence for activation of the p38MAPK/SAPK2 pathway exist from our data, other pathways involving ERK1/2 and potentially SAPK/JNK seem to be primarily involved in the initial signal from the receptor protein to the Egr-1 p romoter. Most interestingly, however, a bFGF-like factor released by irradiated cells contributes markedly to the radiation-induction of the Egr- 1 promoter via upstream S RE elements. Closer insight into the very early events of Egr-1 gene induction and identi®- cation of the secreted factor will be useful for further successful utilization of the Egr-1 promoter in combined radiation-inducible gen e therapy approaches as well as for the understand ing of stress-dependent regulation of cell growth in general. ACKNOWLEDGEMENTS Vector pAdRSVbgal was a kind gift from Dr Ju È rgen Kleinschmidt, Deutsches Krebsforschungszentrum Heidelbe rg. Thanks also to Kathleen M. Sakamoto, University o f California, Lo s Angeles, fo r permission to use plasmid pGL/TiS8. T his work was sup ported by a grant from t he Deutsche Forsc hungsgemeinschaft (Ro527/3-1,2) and the Dr M ildred Scheel Stiftung (10-1 503-Ku È I). REFERENCES 1. Silverman, E.S., Khachigian, L.M., Lindner, V., Williams, A.J. & Collins, T. (1997) Inducible PDGF A-chain transcription in smooth muscle cells is mediated by Egr-1 displacement of Sp1 and Sp3. Am . J. Physiol. 273, 1415±1426. 2. Khachigian, L., Lindner, L., Williams, A. & Collins, T. (1996) Egr-1induced endothelial gene expression: a common theme in vascular i njury. Science 271, 1427±1431. 3. Biesiada, E., Razandi, M. & Levin, E.R. (1996) Egr-1 activates basic ®broblast growth factor transcription. Mechanistic i mp li- cations for astrocyte p roliferation. J. Biol . C hem. 271 , 1 8576± 18581. 4. Kim, S J., Jean g, K T., Glick, A.B., S porn , M.B . & Roberts, A.B. (1989) Promoter sequences of the human transforming growth f actor-b1 gene respo nsive to transforming growth facto r- b1 autoinduction. J. Biol. Chem. 264, 7041±7045. 5. Santiago, F.S., Lowe, H.C., Day, F.L., Chesterman, C.N. & Khachigian, L .M. (1999) Early growth response factor-1 induc- tion by injury is triggered by release and p aracrine activation by ®broblast g rowth factor-2. Am. J. Pathol. 154, 937±944. 6. Yan, S.F., Lu, J., Zou, Y.S., Soh-Won, J., Coh en, D.M., Buttrick, P.M., Cooper, D.R., Ste inbe rg, S.F., Mackman, N., P insk y, D.J. & Stern, D.M. (1999) Hypoxia-associated induction of early growth response-1 gene e xpression. J . Biol. Che m. 274, 15030± 15040. 7. Schwachtgen, J.L., Houston, P., Campbell, C., Sukhatme, V. & Braddock, M. (1998) Fluid shear stress activation of Egr-1 transcription in cultured human endothelial a nd epithelial cells is mediated via the extracellular signal-related kinase 1/2 mitogen- activated protein kinase pathway. J. Clin. Invest. 101, 2540± 2459. 8. Keyse, S.M. (1993) The induc tion of gene expression in mam- malian cells by radiation. Sem. Cancer Biol. 4, 119±128. 9. Weichselbaum, R.R., Hallahan, D.E., Beckett, M.A., Mauceri, H.J., Lee, H., Sukhatme, V.P. & Kufe, D.W. (1994) Gene therapy targeted by radiation p referentially radiosensitizes tumor cells. Cancer Res. 54, 4266±4269. 10. Hallahan, D.E., Mauceri, H.J., Se ung, L.P., Dun phy, E.J., Wayne, J.D., Hanna, N .N., Toledano, A ., Hellman, S., Kufe, D.W. & Weichselbaum, R.R. (1995) Spatial and temporal control of gene therapy using ionizing radiation. Nat. M ed. 1, 786±791. 11. Joki, T., Nakamura, M. & O hno, T. (1 995) Activation o f the radiosensitive EGR-1 promoter induces expression of the herpes simplex virus thymidine k inase g ene a nd sensitivity of human glioma cells to ganciclovir. Hum. Ge ne Ther. 6, 1507±1513. 12. Manome,Y.,Kunieda,T.,Wen,P.Y.,Koga,T.,Kufe,D.W.& Ohno, T. (1998) Transgene expression in malignant glioma using a replication-defective adenoviral vector containing the Egr-1 pro- moter: activation by ion izing radiation or up take of radioactive iododeoxyuridine. Hum. Gene Ther. 9, 1409±1417. 13. Kim, S.H., Kim, J.H., Kolozsvary, A., Brown, S.L. & Freytag, S.O. (1997) Preferential radiosensitization of 9L glioma cells transduced with HSV-tk gene by acyclovir. J. Neurooncol. 33, 189± 194. 14. Sakamoto, K.M., B ardeleben, C., Y ates, K .E., Rai nes, M .A., Golde, D.W. & Gasson, J.C. (1991) 5¢ upstream sequenc e and genomic structure of the primary response g ene EGR-1 /TIS8. Oncogene 6, 867±875. 15. Aicher, W.K., Dinkel, A., Grimbacher, B., Haas, C.V., Seydlitz-Kurzbach, E., Peter, H.H. & Eibel, H. (1999) S erum response elements activate and cAMP responsive elements inhibit expression of tran scription factor E gr-1 in syn ovial ®broblasts of rheumatoid arthritis p atients. Int. Imm unol. 11, 47±61. 16. Christy, B. & Nathans, D. (1989) Functional serum r esponse elements up stream of the growth factor-inducible gene zif68. Mol. Cell Bi ol. 11, 4889±4895. 17. Schwachtgen, J.L., Campbell, C.J. & Braddock, M. (2000) Full promoter sequence of human early growth response factor-1 (Egr-1): demonstration of a ®fth functional serum response element. DNA Seq. 10 , 429±432. 18. Rolli, M., Kotlyarov, A., Sakamoto, K.M., Gaestel, M. & Nein- inger, A. (1999) Stress-induced stimulation of early growth response ge ne-1 by p38/stress-activated protein k inase 2 is medi- ated by a cAMP-responsive promoter e lement in a MAPKAP kinase 2-indepe ndent manne r. J. Biol. Chem. 274, 1 9559±19564. 19. Marais, R., Wynne, J. & Treisman, R. (1993) The SRF accessory protein E lk-1 contains a growth f actor-regulated transcriptional activation d omain. Cell 73, 381±393. 20. Burger, A., L o È er, H., B amberg, M. & Rodemann, H.P. (1998) Molecular and cellular basis of radiation ®brosis. Int. J. Rad. Biol. 73, 401± 408. 21. Hakenjos, L., B amberg, M. & Rodemann, H.P. ( 2000) TG F- b1-mediated alterations of rat lung ®broblast dierentiation re- sulting in the radiation-induced ®brotic response. Int. J. Rad. Biol. 76, 503± 509. 22. Alessi, D.R., Cuenda, A., Cohen, P., Dudley, D.T. & Saltiel, A.R. (1995) PD098059 is a speci®c inhibitor of the activation of mitogen-activated protein kinase kinase in vit ro and in vivo . J. Biol. Chem. 270, 27489±27494. 23. 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. 24. Hall-Jackson, C.A., Eyers, P.A., Cohen, P., Goedert, M., Boyle, F.T., Hewitt, N., Plant, H. & Hedge, P. (1999) Paradoxi- cal activation of Raf by a novel Raf inhibitor. Chem. Biol. 6, 559±568. Ó FEBS 2002 Egr-1 promoter induction by ionizing radiation (Eur. J. Biochem. 269) 345 25. Hall-Jackson, C.A., Goedert, M., Hedge, P. & Cohen, P. (1999) Eect of S B203580 on the a ctivity of c-Raf in vitro and in vivo . Oncogene 18, 2047±2054. 26. Kalmes,A.,Deou,J.,Clowes,A.W.&Daum,G.(1999)Raf-1is activated by the mitogen-activated protein kinase inhibitor, SB203580. FE BS Lett. 444, 71±74. 27. Gonzalez, G.A. & Montminy, M.R. ( 1989) Cyclic AMP stimu- lates somatostatin gene transcription by phosphorylation of CREB at serine 133. Cell 59, 675 ±680. 28. David, M., Petricoin, E . & Larner, A .C. (1996) Activation o f protein kinase A inhibits inte rferon ind uction of t he Jak/Stat pathway in U266 cells. J. Biol. Chem. 27 1, 4585±4588. 29. Marples, B., Scott, S.D., Hendry, J.H., Embleton, M.J., Lashford, L.S. & Margison, G.P. (2000 ) Development of synthetic pro- moters for radiation-mediated gene t herapy. Gene Ther. 7,511± 517. 30. Cao, X.M., Guy, G.R., Sukhatme, V.P. & Tan, Y.H. (1992) Regulation of the Egr-1 genebytumornecrosisfactorandinter- ferons in primary h uman ®broblasts. J. Biol. Chem. 267, 1345± 1349. 31. Groo, M., Boxer, L.M. & Thiel, G . (2000) Nerve growth factor- and epidermal growth factor-regulated gene transcriptio n in PC12 pheochromocyt oma a nd INS-1 insulinoma ce lls. Eur. J. Cell Biol. 79, 924±935. 32. Whitmarsh, A.J., Shore, P., Sharrocks, A.D. & Davis, R.J. (1995) Integration of MAP kinase signal transduction pathways at the serum r esponse e lement. Science 269, 403±407. 33. Hibi, M., Lin, A., Smeal, T., Minden, A. & Karin, M. (1993) Identi®cation of an oncoprotein- and UV-responsive protein kinase that binds and potentiates the c-Jun activation domain. Genes Dev. 7, 2135±2148. 34. Derijard, B., Hibi, M., Wu, I.H., Barrett, T., S u, B., Deng, T., Karin, M. & Davis, R.J. (1994) JNK1: a protein kinase stimulated by UV light and Ha-Ras that b inds an d phosp horylates the c-Jun activation domain. Cell 76, 1025±1037. 35. Wilhelm, D., van Dam, H., Herr, I., Baumann, B., Herrlich, P. & Angel, P. (1995) Both ATF-2 and c-Jun are phosphorylated by stress-activated protein kinases in response to UV irradiation. Immunobiology 193, 143±148. 36. Verheij, M., Ruiter, G.A., Zerp, S.F., van Blitterswijk, W .J., Fuks, Z., Haimovitz-Friedman, A. & Bartelink, H. (1998) The role of the stress-activated protein kinase (SAPK/JNK) signaling pathway in radiation-induced a poptosis. Rad iother. Oncol. 47, 225±232. 37. VanDam,H.,Wilhelm,D.,Herr,I.,Steen,A.,Herrlich,P.& Angel, P. (1995) ATF-2 is preferentially activated by stress- activated protein kinases to mediate c-Jun induction in response to genotoxic agents. EMBO J. 14 , 1798±1811. 38. Liu, C., Yao, J., M ercola, D. & Adamson, E. (2000) The tran- scription factor Egr-1 directly transactivates the ®bronectin gene and enhances attachment of human glioblastoma cell line U 251. J. Biol. Chem. 275, 2 0315±20323. 39. Cavigelli, M., Dol®, F., Claret, F.X. & Karin, M. (1995) Induction of c-fos e xpression through JNK-mediated T CF/Elk-1 p ho spho- rylation. EMBO J. 14, 5957±5964. 40. Gille, H., Strahl, T. & Shaw, P.E. (1995) Activation of ternary complex factor Elk-1 by stress-activated p rotein kinases. Curr. Biol. 5, 1191±1200. 41.Gupta,S.,Campbell,D.,Derijard,B.&Davis,R.J.(1995) Transcription factor ATF-2 regulation by the JNK signal trans- duction pathway. Science 267 , 389±393. 42. Kra È mer, M., Sachsenmaier, C., H errlich, P. & Rahmsdorf, H.J. (1993) UV irradiation-induced Interleukin-1 and basic ®broblast growth factor synthesis and release mediate part of the UV response. J. Bi ol. Chem. 268, 6734±6741. 43. Sachsenmaier, C., Radler-P ohl, A., Zinck, R., Nordheim, A., Herrlich, P. & Rahmsdorf, H.J. (1994) Involvement of growth factor receptors in the mammalian UVC response. Cell 78,963± 972. 44. Woloschak, G .E., C hang-Liu, C.M. & S hearin-Jones, P. (1990) Regulation of protein kinase C by ionizing radiation. Cancer Res. 50, 3963±3967. 45. Aicher, W., Sakamoto, K., Hack, A . & Eibel, H . (1999) Analysis of fu nctional elements in the hum an Egr-1 gene promoter. Rheumatol. Int. 18 , 207±214. 346 R. G. Meyer et al. (Eur. J. Biochem. 269) Ó FEBS 2002 . Early growth response-1 gene ( Egr-1 ) promoter induction by ionizing radiation in U87 malignant glioma cells in vitro Ralph G. Meyer 1,2 , Jan-Heiner. Molecular Pathology, University of Tu È bingen, Germany The promoter of the early growth response gene (Egr- 1) has been described to b e activated by ionizing radiation, and