Oncogene (1998) 16, 2915 ± 2926  1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00 http://www.stockton-press.co.uk/onc Stress-induced immediate-early gene, egr-1, involves activation of p38/JNK1

Cheh Peng Lim, Neeraj Jain and Xinmin Cao

Signal Transduction Laboratory, Institute of Molecular and Cell Biology, National University of Singapore, Singapore 117609

The Ras/Raf/MAP kinase (ERK) pathway is a major expression of immediate-early genes, such as c-fos and signaling pathway induced by growth factors in egr-1, which contain SRE in their promoters (for mammalian cells. Two other types of mammalian MAP review see Treisman, 1995). kinases, JNK (SAPK) and p38 (RK, CSBP), are induced Two new types of mammalian MAP kinases which by environmental stress. Although the immediate-early are activated by environmental stress and pro- gene, egr-1, is induced by growth factors, cytokines, in¯ammatory cytokines were recently identi®ed. The di€erentiation signals and DNA damaging agents, less is c-Jun NH2-terminal kinase (JNK) group of MAP known about its induction by environmental stress and kinases, also known as stress-activated protein kinases the mechanism involved. Here we report that in NIH3T3 (SAPKs), have been shown to be activated by cells, egr-1 is induced by various stress treatments such interleukin-1 (IL-1), tumor necrosis factor (TNF) and as heat shock, sodium arsenite, ultraviolet (U.V.) U.V. radiation (Kyriakis et al., 1994; De rijard et al., radiation, and anisomycin. p38 and JNK1, but not 1994). JNK binds to the amino terminus of c-jun and ERK2, were activated by these stress treatments. phosphorylates it on serine residues 63 and 73 (Hibi et Induction of egr-1 by anisomycin is inhibited by a al., 1993). The third group of MAP kinases, p38, is speci®c inhibitor of p38, SB 203580. We also show that activated in murine cells by endotoxic lipopolysacchar- p38 and JNK1 activated by their upstream kinases ide (LPS) or hyperosmolarity (Han et al., 1994). induce egr-1 promoter activity through activation of the Homologs of p38 MAP kinase in human cells ternary complex factor, Elk-1. The stress treatments also (CSBP) and Xenopus oocytes (Mpk2) have also been lead to an increase in Egr-1 protein phosphorylation and identi®ed (Lee et al., 1994; Rouse et al., 1994) which its DNA binding activity. Together, our data suggest are very similar to the osmosensing MAPK, HOG1, in that induction of egr-1 gene by growth factors and stress yeast. The p38 kinase was also independently puri®ed are mediated through di€erent subgroups of MAP by two groups as reactivating kinase (RK) in sodium kinases which may also di€erentially a€ect egr-1 arsenite-treated PC12 pheochromocytoma cells and function on its target genes. heat-shocked HeLa cells (Rouse et al., 1994), or as a 40 kDa kinase (p40) in IL-1-stimulated KB cells Keywords: egr-1; stress; MAP kinases; p38/JNK1 (Freshney et al., 1994). Although ERKs, JNKs and the p38 kinase family are closely related, they are distinguishable by their speci®c regulatory TXY motif in which X represents glutamic acid in ERKs, proline Introduction in JNKs, and glycine in p38 (De rijard et al., 1994; Payne et al., 1991; Raingeaud et al., 1995). These Activation of receptor tyrosine kinases by binding of kinases are all activated by dual phosphorylation on growth factors to their receptors stimulates prolifera- Thr and Tyr within this motif and each group has both tion and di€erentiation in mammalian cells. The unique and overlapping upstream activators and mitogen-activated protein kinase (MAPK) cascade is downstream substrates (Brunet and Pouysse gur, 1996; considered to be a major signaling pathway which Wang and Ron, 1996). links signals from the cell surface to the nuclear The immediate-early gene egr-1 (also known as events in these processes (for review see Hill and NGFI-A, zif268, TIS8 and krox-24) encodes a Treisman, 1995). This signaling pathway involves transcription factor containing a DNA binding transient formation of ras-GTP and activation of raf domain consisting of three zinc ®ngers (Sukhatme et kinase at the membrane, sequential activation of al., 1988; Milbrandt, 1987; Christy and Nathans, MAPK kinase (MAPKK) and MAP kinase 1989b; Lim et al., 1987; Lemaire et al., 1988). The (MAPK)/extracellular signal-regulated protein kinase egr-1 protein (Egr-1) binds to a speci®c GC-rich (ERK) (Marshall, 1994; Leevers et al., 1994; Stokoe sequence in the promoter region of many genes to et al., 1994). ERKs translocate to the nucleus, regulate the expression of these target genes including phosphorylate and thus activate the transcription growth factors and cytokines (Christy and Nathans, factor p62TCF (Shaw et al., 1989), which forms a 1989a; Cao et al., 1993). The egr-1 gene is rapidly and ternary complex with serum response factor (SRF) transiently induced by growth factors and differentia- and serum response element (SRE), and activates the tion signals (Sukhatme et al., 1988; Milbrandt, 1987) and is functionally implicated in cell proliferation and di€erentiation processes (Hill and Treisman, 1995; Correspondence: X Cao Sukhatme et al., 1988; Sukhatme, 1990; Nguyen et Received 29 September 1997; revised 7 January 1998; accepted 8 al., 1993). The egr-1 gene has been reported to be January 1998 induced by TNF and IL-1 (Cao et al., 1992a), and Stress-induced egr-1 CP Lim et al 2916 serine/threonine protein phosphatase inhibitor, okadaic Results acid (Herschmann et al., 1989; Cao et al., 1992b), which are known stimulators of p38 and JNK Induction of egr-1 mRNA by stress treatments in (Kyriakis et al., 1994; De rijard et al., 1994; Rouse et NIH3T3 cells al., 1994; Freshney et al., 1994; Cano et al., 1994). In addition, egr-1 is also induced by diverse types of To examine whether egr-1 expression can be induced DNA-damaging agents such as U.V. light (Huang et by various stresses, quiescent NIH3T3 cells were al., 1995), ionizing radiation (Hallahan et al., 1991; treated with heat (428C), anisomycin (25 ng/ml), U.V. Datta et al., 1992), and 1-(b-D-arabinofuranosyl) (40 J/m2), and sodium arsenite (500 mM). Sodium cytosine [(ara-C) Kharbanda et al., 1993]. Since these arsenite mimics the e€ects of heat shock by inducing DNA-damaging agents also activate JNK (Kharbanda the expression of several heat shock factors (Johnston et al., 1995a,b; Chen et al., 1996) and p38 (Pandey et et al., 1980) and is an activator of p38 (Rouse et al., al., 1996), it is possible that JNK and p38 are involved 1994). Anisomycin, on the other hand, is reported to in the egr-1 induction by these agents. However, less is be an activator of JNKs (Cano et al., 1994). Total known about the egr-1 induction by environmental RNA was isolated and subjected to Northern blot stress and the mechanisms of induction involved. To analysis. A low level of egr-1 mRNA was detectable in investigate the role of egr-1 in the cellular response to the untreated cells (lane 1 in Figure 1a ± d), whereas a stress, we tested the e€ects of a variety of environ- transient induction of egr-1 mRNA was observed by mental stresses on the egr-1 expression and analysed treatment of these cells with heat shock, anisomycin, or the possible signaling pathways involved. Here, we U.V. The maximum induction of egr-1 transcription report that both the mRNA and protein levels of egr-1 was approximately 33-fold by anisomycin at 30 min are increased after various stress treatments, such as (Figure 1b, lane 3), eightfold by heat shock at 30 min heat shock, sodium arsenite, anisomycin and U.V. (Figure 1a, lane 3) or threefold by U.V. at 60 min radiation in NIH3T3 cells. Our results indicate that (Figure 1c, lane 4), and dropped to the basal level after p38 and JNK, but not ERK2, are involved in the treatment of the cells for 2 h. On the other hand, induction of egr-1 by stress, suggesting that di€erent sodium arsenite treatment strongly induced egr-1 MAPK pathways are involved in egr-1 expression expression at 30 min (eightfold) and sustained up to induced by growth factors and stress factors. 4 h (14-fold) as shown in Figure 1d. Induction of egr-1

ab HEAT SHOCK ANISOMYCIN

0 15 30 60 120 240 (min) 0 15 30 60 120 180 (min)

egr-1 egr-1

GAPDH GAPDH

fold 1 2 8 4 0.5 0.4 fold 1 20.2 33.4 26 0.7 0.2 1 2 3 4 5 6 1 2 3 4 5 6

cd UV ARSENITE

0 15 30 60 120 (min) 0 30 60 120 180 240 (min)

egr-1 egr-1

GAPDH GAPDH

fold 1 1.1 1.8 2.7 0.1 fold 1 8 10.8 13.6 13.8 14 1 2 3 4 5 1 2 3 4 5 6 Figure 1 Northern blot analysis of egr-1 mRNA induced by various stresses. Quiescent NIH3T3 cells were treated with (a) heat 2 shock (428C), (b) Anisomycin (25 ng/ml), (c) U.V. (40 J/m ), or (d) sodium arsenite (500 mM) for the time periods indicated in minutes (min) on top of each panel. Total RNA was isolated and equal amounts (10 mg) were subjected to Northern blot analysis as described in Materials and methods. egr-1 mRNA is shown in the upper panels. The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA was used to normalize the amount of RNA loaded and is shown in the lower panels. The numbers at the bottom of each panel indicate the fold of egr-1 induction as compared to control normalized by GAPDH Stress-induced egr-1 CP Lim et al 2917 by di€erent doses of sodium arsenite was also protein (indicated by arrows) was detected after 60 min examined and the result showed that the egr-1 mRNA of sodium arsenite treatment which reached a peak at was induced by concentrations as low as 5 mM, 90 min, and decreased after 2 h (Figure 2b). The increased dose-dependently up to 100 mM with no migration of Egr-1 protein notably shifted above further increase beyond 100 mM (not shown). To- 80 kDa and appeared as a complex of 82 ± 84 kDa gether, these results indicate that egr-1 mRNA can be after 60 min, reached a maximum at 90 min and induced by heat shock, anisomycin and U.V. decreased after 120 min of arsenite treatment (Figure transiently and by sodium arsenite persistently. 2b, lanes 4 ± 6). This complex represents the hyperpho- sphorylated forms of Egr-1 which is similar to that observed in the okadaic acid-treated cells as reported Induction of Egr-1 protein by stress treatments previously (Cao et al., 1992b). The transient induction To determine whether Egr-1 protein was also induced of Egr-1 protein by sodium arsenite in contrast to the by stress, quiescent NIH3T3 cells were labeled with 35S- sustained induction of mRNA could be due to a partial methionine for 1 h and subsequently treated with inhibition of the protein synthesis by sodium arsenite. sodium arsenite for 1 h. The labeled Egr-1 protein The three bands below the Egr-1 detected in the was immunoprecipitated with an anti-Egr-1 polyclonal Western blot (Figure 2b) are proteins recognized non- antibody (Cao et al., 1990) and the immunoprecipitates speci®cally by Egr-1 polyclonal antibody. Similarly, were resolved by SDS ± polyacrylamide gel electrophor- Egr-1 protein synthesis and hyperphosphorylation were esis (SDS ± PAGE). A fourfold increase in Egr-1 also induced by heat shock at 428C (data not shown). protein synthesis was observed after treatment of the Together, these data indicate that in addition to egr-1 cells with sodium arsenite for 1 h (Figure 2a, lane 2). mRNA and protein, the phosphorylation of Egr-1 Serum was used for comparison (lane 3) since it is protein are also induced in NIH3T3 cells by various known to induce both egr-1 mRNA and Egr-1 protein stress treatments. synthesis (Sukhatme et al., 1988; Cao et al., 1990). The induction of Egr-1 protein synthesis by sodium arsenite Activation of p38 and JNK1 by sodium arsenite and at various time points was also measured by direct U.V., but not by PDGF in NIH3T3 cells Western blot analysis. A signi®cant increase of Egr-1 To identify which MAP kinases are activated speci®cally in response to stress and thus likely to be involved in egr-1 induction, activation of p38, JNK1, a and ERK2 in NIH3T3 cells were investigated. Since ARSENITE SERUM members of the MAP kinase families are known to be 0 60 60 (min) activated by dual phosphorylation on Tyr and Thr (De rijard et al., 1994; Payne et al., 1991; Raingeaud et al., 1995), we examined the activation of these kinases Egr-1 by checking their Tyr-phosphorylation status using a combination of immunoprecipitation and Western blot 1 2 3 (IP/Blot) techniques. Platelet-derived growth factor (PDGF) strongly activates ERKs and was used as a b positive control in our experiments. Quiescent NIH3T3 ARSENITE cells were treated with PDGF, U.V., or sodium arsenite 0 15 30 60 90 120 180 240 (min) for various times, and the cells were lysed, and kDa incubated with either anti-p38, JNK1 or ERK2 — 195 antibody. The immunoprecipitated proteins were separated by SDS ± PAGE, transferred onto a PVDF — 112 membrane, and probed with an anti-phosphotyrosine — 84 antibody, 4G10. The results showed that no activation Egr-1(HP) of p38, JNK1 and ERK2 (lane 1 in Figure 3a ± c, lane — 63 8 in Figure 3a ± b, upper panels) or low level of ERK2 — 52.5 activation (Figure 3c, lane 8, upper panel) was observed in untreated control cells, whereas strong 1 2 3 4 5 6 7 8 activation of p38 (Figure 3a) and JNK1 (Figure 3b) Figure 2 Induction of Egr-1 protein by sodium arsenite. (a) was detected in both sodium arsenite and U.V. treated Immunoprecipitation of Egr-1 protein. Quiescent NIH3T3 cells 35 cells. The activation of p38/JNK1 by U.V. was were labeled with S-methionine (0.25 mCi/ml) for 1 h and observed after treatment of the cells for 15 min, treated with (lane 2) or without (lane 1) sodium arsenite (10 mM), or with 20% fetal calf serum (lane 3) for 60 min. The cell lysates reached maximum at 30 min and decreased at 60 min were immunoprecipitated with an anti-Egr-1 polyclonal antibody (Figure 3a ± b, lanes 5 ± 7). In contrast, p38 activation (Cao et al., 1990). The immunoprecipitates were resolved by 7.5% by sodium arsenite was observed at 30 min (Figure 3a, SDS ± PAGE and the gel was dried and autoradiographed. (b) lane 10), and continued to increase at 60 min (lane 11), Western blot analysis of Egr-1 protein. NIH3T3 cells were and 120 min (lane 12), and sustained up to 3 h (lane untreated (lane 1) or treated with sodium arsenite (50 mM, lanes 2 ± 8) for the time indicated in min. The cells were lysed and equal 13), whereas the activation of JNK1 was detected at amount of protein (22 mg) were loaded onto a 7.5% SDS ± PAGE, 30 min increased at 60 min, and further increased at 2 transferred to a PVDF membrane and blotted with an anti-Egr-1 and 3 h (Figure 3b, lanes 10 ± 13). On the other hand, antibody, 588, raised against its C-terminus. The position of Egr- PDGF did not stimulate p38 or JNK1 (Figure 3a and 1 protein is indicated by arrows and the hyperphosphorylated Egr-1 protein is indicated as Egr-1 (HP). The molecular mass b, lanes 2 ± 4), but strongly stimulated ERK2 (Figure markers are shown on the right 3c, lanes 2 ± 4). U.V. (Figure 3c, lanes 5 ± 7) and Stress-induced egr-1 CP Lim et al 2918 arsenite (lanes 9 ± 13) treatments did not activate ERK2 hyperphosphorylation were also detected and shown in signi®cantly. These blots were stripped and reprobed Figure 3d. with their respective antibodies used for immunopreci- These data demonstrated that in NIH3T3 cells, pitation to show the amount of proteins present in sodium arsenite and U.V. activated p38 and JNK1, but each lane (Figure 3a ± c, bottom panels). not ERK2. Therefore, ERK2 is not involved in the Aliquots of the cell lysates treated with PDGF or induction of egr-1 by these stress treatments. Instead, U.V. were also subjected to direct Western blot activation of p38 and JNK is most likely to be involved analysis to monitor the induction of Egr-1 protein in egr-1 induction by stress. The observation that U.V. synthesis. The induction of the Egr-1 protein and its induces transient activation of p38/JNK and egr-1

a PDGF UV ARSENITE 0 15 30 60 15 30 60 0 15 30 60 120 180 (min)

IP : p38 BLOT: 4G10

IP : p38 BLOT: p38

1 2 3 4 5 6 7 8 9 10 11 12 13

b PDGF UV ARSENITE 0 15 30 60 15 30 60 0 15 30 60 120 180 (min) IP : JNK1 BLOT: 4G10

IP : JNK1 BLOT: JNK1 1 2 3 4 5 6 7 8 9 10 11 12 13

c PDGF UV ARSENITE 0 15 30 60 15 30 60 0 15 30 60 120 180 (min) IP : ERK2 BLOT: 4G10

IP : ERK2 BLOT: ERK2 1 2 3 4 5 6 7 8 9 10 11 12 13

d PDGF UV 0 15 30 60 15 30 60 (min)

BLOT: Egr-1 Egr-1

1 2 3 4 5 6 7 Figure 3 Tyrosine phosphorylation of p38, JNK1 and ERK2. NIH3T3 cells were left untreated (lanes 1 and 8) or treated with 2 PDGF (50 ng/ml, lanes 2 ± 4), U.V. (40 J/m , lanes 5 ± 7), or sodium arsenite (500 mM, lanes 9 ± 13) for time periods in min as indicated on top of each ®gure. Cell lysates were subjected to immunoprecipitation (IP) by appropriate antibodies (anti-p38 in a, anti-JNK1 in b and anti-ERK2 in c). The immunoprecipitates were resolved by 10% SDS ± PAGE, transferred to PVDF membrane and blotted with an anti-P-Tyr antibody (4G10) or other appropriate antibodies as indicated on the left of each ®gure (BLOT). (d). Western blot analysis of egr-1 protein. 30 mg of total cell lysates were separated by 7.5% SDS ± PAGE and subjected to direct Western blot analysis using anti-egr-1 (588) antibody Stress-induced egr-1 CP Lim et al 2919 expression, whereas sodium arsenite induces a sus- egr-1 gene induction by stress using its speci®c tained p38/JNK activation (Figure 3a and b) and egr-1 inhibitor, SB 203580 (Cuenda et al., 1995). Quiescent expression (Figure 1), indicates a good correlation NIH3T3 cells were pre-treated with or without between these two responses and further suggests the SB 203580 (20 mM) for 15 min followed by treatments involvement of p38/JNK in egr-1 induction by stress. of PDGF, U.V., sodium arsenite, or anisomycin for Since sodium arsenite mimics the e€ect of okadaic acid 45 min. Strong induction of egr-1 by anisomycin (Huang et al., 1995), which induces a sustained egr-1 (Figure 5a, lane 9) was inhibited signi®cantly (more expression (Cao et al., 1992b), the mechanism of the than 50%, lane 10) and only slightly inhibited by U.V. sustained egr-1 induction by sodium arsenite may (lanes 5 and 6). In contrast, SB 203580 did not inhibit therefore involve the inhibitory e€ect of serine/ egr-1 induction by PDGF (lanes 3 and 4). Surprisingly, threonine phosphatases on the dephosphorylation of SB 203580 did not inhibit egr-1 expression induced by p38/JNK. sodium arsenite but rather enhanced it further (lanes 7 To con®rm that the tyrosine phosphorylation of p38 and 8). To further analyse which MAP kinases are and JNK1 corresponds to their enzymatic activity, the involved in egr-1 induction by the individual stress p38 and JNK1 kinase activities were analysed. factor, the e€ects of SB 203580 on p38 and ERK2 Transcription factors ATF2 and c-Jun are known to activation were also examined by IP/Blot experiment. be phosphorylated by p38 and JNK1 respectively NIH3T3 cells were treated as described above, and the (De rijard et al., 1995; Hibi et al., 1993). Quiescent cell lysates were incubated with anti-p38 or -ERK2 NIH3T3 cells were treated with U.V., sodium arsenite antibody. The immunoprecipitates were probed with (ARS) or anisomycin (ANI) and the cell lysates were anti-phosphotyrosine antibody, 4G10. The results immunoprecipitated with either the anti-p38 or anti- showed no activation of p38 by PDGF with or JNK1 antibody. Half of the immunoprecipitates were without pre-treatment of SB 203580 (Figure 5b, lanes assayed for phosphorylation of GST-ATF2 (Figure 4a, 3 and 4) as expected. U.V.-, anisomycin- and sodium upper panel) and GST-c-Jun (Figure 4b, upper panel) arsenite-activated p38 (lanes 5, 7 and 9) was inhibited for kinase activity of p38 and JNK1 respectively, by SB 203580 (lanes 6, 8 and 10). Interestingly, whereas the other half was used for examining their SB 203580 a€ects the ERK2 activity by enhancing its tyrosine phosphorylation by IP/Blot (middle panels). tyrosine phosphorylation in cells treated with PDGF Basal level of GST-ATF2 phosphorylation was (Figure 5c, compare lanes 4 to 3), anisomycin (lanes 8 observed which was enhanced by U.V. and anisomy- to 7), and sodium arsenite (lanes 10 to 9). These cin strongly and by sodium arsenite weakly (Figure 4a, enhancements observed in PDGF- and sodium upper panel). Correspondingly, the tyrosine phosphor- arsenite-treated cells in the presence of the inhibitor ylation of p38 was increased strongly by U.V. and SB 203580 may cause the increase of egr-1 mRNA anisomycin treatment, but weakly by arsenite (middle shown in the Northern blot analysis (Figure 5a, lanes 3 panel). Similarly the correlation between JNK1 kinase and 4; lanes 7 and 8). No inhibition of JNK1 by activity and its tyrosine phosphorylation was also SB 203580 was observed (not shown) which is in shown (Figure 4b). agreement with that reported previously (Wang and Ron, 1996). We also tested the e€ect of SB 203580 on heat shock-induced egr-1 expression and the results E€ects of p38 inhibitor SB 203580 on egr-1 induction by were similar to that of sodium arsenite treatment (data various stress treatments not shown). In summary, by using the p38 speci®c Since the stress treatments tested above stimulated p38 inhibitor SB 203580, we further showed that p38 may and JNK1 activity, we evaluated the role of p38 on be mainly involved in egr-1 induction by anisomycin,

ab U UV ARS ANI U UV ARS ANI

GST-ATF2 GST-cjun

IP : p38 IP : JNK 1 BLOT: 4G10 BLOT: p-JNK

IP : p38 IP : JNK1 BLOT: p38 BLOT: JNK1 1 2 3 4 1 2 3 4 Figure 4 Tyrosine phosphorylation correlates with activation of p38 and JNK1. NIH3T3 cells were left untreated (U, lanes 1) or 2 treated with U.V. (60 J/m for 30 min, lanes 2), sodium arsenite (ARS, 500 mM for 60 min, lane 3) or anisomycin (ANI, 25 ng/ml for 45 min, lane 4). Cell lysates were subjected to immunoprecipitation (IP) by appropriate antibodies (anti-p38 in (a) anti-JNK1 in (b)). The immunoprecipitates were divided into two halves. One half was subjected to in vitro kinase assays using either GST-ATF2 (1 ± 96) as a substrate for p38 kinase (a, upper panel) or GST-c-Jun (1 ± 79) as a substrate for JNK1 kinase (b, upper panel). The other half of the immunoprecipitates were resolved by 10% SDS ± PAGE, transferred to PVDF membrane and blotted with an anti- P-Tyr antibody (4G10) or other appropriate antibodies as indicated on the left of each ®gure (BLOT) (a and b, lower panels). p-JNK: antibody against phosphorylated JNK Stress-induced egr-1 CP Lim et al 2920 but only partially involved in egr-1 induction by U.V. et al., 1988). We next investigated whether p38 can since SB 203580 inhibited p38 activation, but only activate egr-1 promoter and whether the e€ect was slightly decreased egr-1 induction by U.V. On the other mediated through the activation of Elk-1 by transient hand, JNK1 may be involved in egr-1 induction by transfection experiments. MAP kinase kinase 3 U.V. (MKK3), has been reported to phosphorylate and activate p38 in mammalian cells (De rijard et al., 1995). The plasmid pEGR-1.1.2-CAT containing p38 stimulates egr-1 promoter activity through activation approximately 1 kb of the egr-1 promoter sequence of Elk-1 upstream of the reporter CAT gene was co-transfected The ternary complex factor TCF binds to SRF and with pcDNA3.p38 expressing p38, either alone or the c-fos SRE upon stimulation by growth factors and together with plasmid pMKK3(E) expressing a activates c-fos gene expression (for review see Treis- constitutively activated mutant of MKK3, into man, 1995). It has been reported that Elk-1, a member NIH3T3 cells in the absence or presence of Elk-1 of the TCF family, is phosphorylated and therefore expressing plasmid, MLV-Elk-1. The transfected cells activated by ERK as well as p38 (Janknecht et al., were harvested and CAT activities were measured. As 1993; Whitmarsh et al., 1997). There are six core shown in Figure 6a, egr-1 promoter activity did not sequences of SRE present in the promoter region of increase when cells were transfected with MLV-Elk-1 egr-1 gene and four SREs at 5' region of egr-1 (lane 2), or pcDNA3.p38 (lane 3) alone, or together promoter are shown to be important for growth factor (lane 4). However, the promoter activity of reporter and serum induction (Tsai-Morris et al., 1988; Christy plasmid was increased threefold by co-transfection of pcDNA3.p38 and pMKK3(E) (lane 5), and further enhanced (®vefold) in the presence of MLV-Elk-1 a (lane 6). Furthermore, as a control, we observed an U PDGF UV ARS ANI increase in egr-1 promoter activity when the cells were co-transfected with plasmids of ERK2 and its SB – + – + – + – + – + upstream kinase, MEK1 (Mansour et al., 1994), but did not see any increase in the cells co-transfected egr-1 with p38 and MEK1 (data not shown). These results suggest that the egr-1 promoter activity was stimu- lated via the activation of p38 by MKK3 speci®cally. To further verify this, the phosphorylation of p38 by GAPDH MKK3 was examined by IP/Blot experiments. We observed an increase in p38 tyrosine phosphorylation 1 2 3 4 5 6 7 8 9 10 only when it was co-transfected with pMKK3(E) in the absence (Figure 6b, upper panel, lane 5) or b presence of Elk-1 (lane 6). U PDGF UV ANI ARS To further investigate the involvement of Elk-1 in SB – + – + – + – + – + egr-1 gene expression regulated by p38, the activation IP : p38 of Elk-1 by p38 was tested in mobility shift DNA BLOT : 4G10 binding assays. NIH3T3 cells were co-transfected with p38 and/or Elk-1 alone or together with MKK3 and IP : p38 BLOT : p38 the whole cell lysates were prepared and mobility shift 1 2 3 4 5 6 7 8 9 10 DNA binding assays were performed using the most 5' SRE (SRE6) in egr-1 promoter as a probe. Low level c of the ternary complex (TC) was detected in cells U PDGF UV ANI ARS transfected with control plasmid (Figure 6c, lane 1) and increased slightly in cells transfected by either SB – + – + – + – + – + Elk-1 alone (lane 2), p38 alone (lane 3), p38 and Elk-1 IP : ERK2 (lane 4), or p38 and MKK3 together (lane 5). The BLOT : 4G10 amount of ternary complex (TC) increased ®vefold IP : ERK2 when the cells were co-transfected with MKK3, p38, BLOT : ERK2 and Elk-1 (lane 6). The formation of ternary complex 1 2 3 4 5 6 7 8 9 10 was also increased (threefold) when cells co-trans- fected with p38 and Elk-1, were stimulated with U.V. Figure 5 E€ects of SB 203580 on the induction of egr-1 and the (lane 7). Serum treatment of the cells transfected with activation of p38 and ERK2 by stress. Quiescent NIH3T3 cells were pre-treated with (+) or without (7) SB 203580 (20 mM) for plasmid expressing ERK2 and Elk-1 showed ternary 15 min followed by treatments of PDGF, U.V., sodium arsenite complex formation with fourfold increase and served (ARS) and anisomycin (ANI) for 45 min or left untreated (U). (a) as a positive control (lane 8). Taken together, these Northern blot analysis of egr-1 mRNA. Total RNA was isolated data indicate that MKK3 activates p38 which and subjected to Northern blot analysis. egr-1 mRNA is shown in the upper panel and GAPDH is shown in the lower panel. (b and speci®cally induces egr-1 expression through the c). E€ect of SB 203580 on (b) p38 and (c) ERK2 phosphorylation activation of Elk-1. by stress. p38 and ERK2 activation was examined by IP/Blot experiment. Cell lysates were immunoprecipitated with (b) anti- p38 or (c) anti-ERK2 antibody and probed with anti- JNK1 also stimulates egr-1 promoter activity phosphotyrosine antibody 4G10 (upper panels). The blots were stripped and re-probed with (b) anti-p38 or (c) anti-ERK2 We also tested whether JNK1 can stimulate egr-1 antibody (lower panels) as described in Figure 3 promoter activity and whether Elk-1 is required. The Stress-induced egr-1 CP Lim et al 2921 reporter plasmid pEGR-1.1.2-CAT, expression plas- the reporter CAT activity. However, the CAT activity mids JNK1 and/or MEKK1, an upstream kinase in increased approximately threefold in cells transfected JNK pathway (Lange-Carter et al., 1993), were co- with MEKK1 and JNK1 in the presence of Elk-1 (lane transfected with or without Elk-1 into NIH3T3 cells. 6), but did not increase in the absence of Elk-1 (lane 5). CAT assays were performed and the results are shown These results suggest that JNK1 activated by MEKK1 in Figure 7. Elk-1 and JNK1, either alone (lanes 2 and also stimulates egr-1 expression through the activation 3 respectively) or together, (lane 4) did not stimulate of Elk-1.

a b MKK3E – – – – + + p38 – – + + + + MKK3E – – – – + + Elk-1 – + – + – + p38 – – + + + + Elk-1 – + – + – + IP : p38 BLOT : 4G10

IP : p38 BLOT : p38 1 2 3 4 5 6

fold 1 0.7 0.4 0.6 2.8 5.1 1 2 3 4 5 6 c UNINDUCED UV S

MKK3E – – – – + + – – p38 – – + + + + + – p42 – – – – – – – + Elk-1 – + – + – + + + FP

TC

Figure 6 p38 activates egr-1 promoter activity through Elk-1. (a) NIH3T3 cells were co-transfected with reporter plasmid pEGR- 1.1.2-CAT and either together with the control plasmid (lane 1), or with expression plasmids MLV-Elk-1 (lane 2), pcDNA3.p38 (lane 3), pcDNA3.p38 and MLV-Elk-1 (lane 4), pMKK3(E) and pcDNA3.p38 (lane 5), or pMKK3(E), pcDNA3.p38 and MLV- Elk-1 (lane 6). Cells were harvested and the CAT activities were analysed as described in Materials and methods. The experiments were repeated three times and similar results were obtained. A representative result is shown. (b) NIH3T3 cells were transfected as described in (a) except in the absence of the reporter gene. Cell lysates were immunoprecipitated (IP) by an anti-p38 antibody and probed with 4G10 (upper panel). The blots were stripped and reprobed with anti-p38 antibody to show the amount of p38 protein in each sample (lower panel). (c). Phosphorylation of Elk- core SRF/SRE 1 by p38 increases the ternary complex formation. NIH3T3 cells were transfected with either the control plasmid (lane 1) or the various expression plasmids indicated on top of the ®gure. ERK2 expression plasmid is designated as p42. The cells were left untreated (lanes 1 ± 6) or treated with U.V. (lane 7) for 30 min or 20% serum (lane 8, designated as S) for 15 min. The cells were harvested, and the whole cell lysates were prepared and subjected to the mobility shift DNA binding assay using 32P-labeled egr-1 SRE6 as a probe as described in Materials and methods. TC: 1 2 3 4 5 6 7 8 9 ternary complex; core SRF/SRE: complex of core SRF and SRE. Free probe is shown in lane 9 (FP) Stress-induced egr-1 CP Lim et al 2922 involved in the egr-1 induction by pro-in¯ammatory E€ect of stress on Egr-1 DNA binding activity cytokines and stress have not been elucidated, although As described in Figure 2b, both the amount and Ras and protein kinase C pathways have been phosphorylation of Egr-1 protein increased after suggested to be involved in egr-1 induction by U.V. treatment of the cells with stresses. Since as a radiation (Huang et al., 1995). Recently, several MAP transcription factor itself, phosphorylation of Egr-1 is kinase subfamilies including p38 (also called RK or important for its function. The e€ect of these stresses CSBP) and JNKs (also called SAPKs) which are on the Egr-1 DNA binding activity was therefore activated by environmental stresses and pro-inflamma- examined. Nuclear extracts from arsenite-treated cells tory cytokines have been cloned and characterized were prepared and subjected to DNA binding mobility (Kyriakis et al., 1994; De rijard et al., 1994; Hibi et al., shift assay as described previously (Cao et al., 1993). In 1993; Han et al., 1994; Lee et al., 1994; Rouse et al., Figure 8a, a dose-dependent increase in Egr-1 DNA 1994; Freshney et al., 1994; Raingeaud et al., 1995). binding was observed in the arsenite-treated cells. The TNF-a and U.V. were shown to activate p38 and JNK time course showed a maximum binding after 2 h and okadaic acid was subsequently shown to activate treatment which was maintained up to 3 h (Figure 8b, JNK (Cano et al., 1994). DNA-damaging agents, such lanes 5 and 6). The data indicate that the induction as ionizing radiation, also activate JNK (Kharbanda et and phosphorylation of the Egr-1 protein by stress al., 1995a,b; Chen et al., 1996) and p38 (Pandey et al., resulted in an increase in the Egr-1 DNA binding 1996). These data suggest that p38/JNK pathways may activity. A similar increase in DNA binding activity be involved in the induction of egr-1 by these reagents. was also observed with heat shock at 428C for 1 h in The goal of our study is to investigate whether egr-1 is NIH3T3 cells (data not shown). Collectively, these one of the nuclear targets of stress signaling pathways results indicate that stress treatments result in an and the possible function of egr-1 in cellular response induction of Egr-1 protein synthesis and its phosphor- to stress. As the ®rst step, we tested the egr-1 induction ylation which lead to an increase in its DNA binding by various stress treatments in NIH3T3 cells (Figures 1 activity. and 2). In parallel, we also examined the activation of the three subtypes of MAP kinases under similar conditions in the same cell lines (Figures 3 and 4). Our Discussion results provide a side by side comparison of egr-1 induction and p38/JNK activation in NIH3T3 cells by We and others have previously reported that egr-1 stress treatments such as heat shock, sodium arsenite, expression can be induced by TNF-a (Cao et al., anisomycin and U.V. radiation. Another cell type, 1992a) and by a serine/threonine protein phosphatase human carcinoma cell A431, was also used to further inhibitor, okadaic acid (Herschmann et al., 1989; Cao test the correlation of egr-1 induction and activation of et al., 1992b). The egr-1 gene has also been reported to p38/JNK. Both induction of egr-1 expression and the be induced by diverse types of DNA-damaging agents activation of p38 and JNK1 by sodium arsenite were including U.V. light (Huang et al., 1995), ionizing observed in a dose-dependent manner (not shown), radiation (Hallahan et al., 1991; Datta et al., 1992), suggesting that the correlation between activation of and ara-C (Kharbanda et al., 1993). However, the p38/JNK and egr-1 expression is not restricted to one induction of egr-1 by environmental stress, such as heat cell type. In addition, the p38 kinase was originally shock, or chemical stress such as sodium arsenite, has identi®ed as a downstream target of TNF, LPS and not been reported and also the signaling pathways hyperosmolarity signaling pathways in pre-B and cells of monocytic origin (Han et al., 1994). Since the egr-1 expression is also stimulated by LPS (not shown), heat MEKK1 – – – – + + shock (Figure 1) and TNF-a (Cao et al., 1992a), the JNK1 – – + + + + activation of p38 by these treatments was also tested. Elk-1 – + – + – + Our results showed that LPS (10 ng/ml), heat shock (428C), and NaCl (0.2 M) induced strong tyrosine phosphorylation of p38 in NIH3T3 and A431 cells, and a weaker phosphorylation of JNK1 was also observed by these treatments (data not shown). These data provide further evidence of a correlation between p38/JNK activation and egr-1 induction by stress. On the other hand, our results showed that phorbol ester PMA, an activator of PKC, strongly stimulated ERK2 phosphorylation, but did not activate p38 and JNK1 (data not shown). Since ERK2 is not activated by the stress treatments we have tested, it is unlikely that fold 1 1.1 0.7 1.1 0.6 2.7 1 2 3 4 5 6 PKC is involved in egr-1 induction in the NIH3T3 cells. On the other hand, c-Abl and phosphatidylino- Figure 7 JNK1 activates egr-1 promoter activity through Elk-1. sitol (PI) 3-kinase have been suggested to be involved (a) NIH3T3 cells were co-transfected with reporter plasmid pEGR-1.1.2-CAT and either together with the control plasmid in the activation of JNK1 and/or p38 (Kharbanda et (lane 1), or the expression plasmids MLV-Elk-1 (lane 2), pSRa- al., 1995a,b; Pandey et al., 1996), but it is not known HA-JNK1 (lane 3), pSRa-HA-JNK1 and MLV-Elk-1 (lane 4), whether these are required in the egr-1 induction by MEKK1 and pSRa-HA-JNK1 (lane 5), or MEKK1, pSRa-HA- stress. JNK1 and MLV-Elk-1 (lane 6). Cells were harvested and the CAT activities were analysed as described in Materials and In the cell lines which we have tested (NIH3T3 and methods A431), the stress treatments which activate p38 can Stress-induced egr-1 CP Lim et al 2923 also activate JNK1 (Figures 3 and 4). Inhibitor of p38, phosphorylated and therefore activated to form SB 203580, inhibited egr-1 induction by anisomycin ternary complex with SRF and SRE by ERKs, JNKs dramatically, but weakly by U.V. (Figure 5a), and p38 in response to a variety of extracellular stimuli indicating that p38 may be the major MAP kinase (Janknecht et al., 1993; Whitmarsh et al., 1995, 1997). involved in egr-1 induction by anisomycin, whereas There are six SRE-like sequences (Tsai-Morris et al., JNK1 may be involved by U.V. These data indicate 1988) and a few TCF binding sites in the egr-1 that although both p38 and JNK1 are usually activated promoter region. Two TCF sites are located at 5' simultaneously by various stress treatments, the ¯anking region of SRE 6 (the most distal SRE to the induction of the immediate-early genes by di€erent transcriptional starting site) one at the 3' ¯anking stress may preferentially be mediated through di€erent region of SRE 6 and one at 3' of SRE 3. No obvious stress-induced MAP kinases. TCF sequence is found around the other SREs. It has c-fos, a prototype immediate-early gene, is rapidly been reported that the ®rst two or three distal SREs induced by various extracellular stimuli including are functional in the ara-C and X-ray response growth factors, cytokines, phorbol ester and stress (Kharbanda et al., 1993; Datta et al., 1992). factors. Serum response element SRE is the major Consistently, we also detected that the ®rst two distal regulatory element located in the c-fos promoter which SREs are important for induction of egr-1 by sodium is recognized by serum response factor SRF and arsenite (not shown). However, it was unclear whether ternary complex factors, a family of ETS-domain Elk-1 is involved in egr-1 induction by growth factors transcription factors (see Treisman, 1995 for review). and stress. Our results obtained from mobility shift Elk-1 is a member of the TCF family which is DNA binding and CAT assays suggest that Elk-1

a b ARSENITE ARSENITE

0 1 10 25 50 100 (µM) 0 30 60 90 120 180 (min)

Egr-1

1 2 3 4 5 6 1 2 3 4 5 6 Figure 8 Sodium arsenite a€ects Egr-1 DNA binding activity. Nuclear extracts were prepared from NIH3T3 cells treated with either (a) di€erent concentrations of sodium arsenite for 1 h or (b)25mM sodium arsenite for various time points as indicated on top of the panels, and subjected to DNA mobility shift assay using a 32P-labeled probe containing Egr-1 binding sequence as described in Materials and methods Stress-induced egr-1 CP Lim et al 2924 positively regulates egr-1 expression induced by stress et al., 1996). In brief, after various treatments, the cells and growth factor through its activation by p38/JNK were washed three times with cold phosphate-bu€ered and ERK2 respectively (Figures 5 ± 7 and not shown). saline (PBS) and lysed in 1.0 ml RIPA bu€er containing However, since there are six SREs present in the egr-1 150 mM NaCl, 50 mM Tris-HCl (pH 7.2), 1% deoxycholic promoter, whether Elk-1 form ternary complex with acid, 1% Triton X-100, 0.25 mM EDTA (pH 8.0) and the protease and phosphatase inhibitors (5 mg/ml leupeptin, each SRE in di€erential stimulation by various factors 5 mg/ml aprotinin, 2 mg/ml pepstatin A, 1 mM phenyl- and whether the binding is equal to each SRE, are not methylsulfonyl ¯uoride (PMSF), 5 mM NaF and 100 mM known. Furthermore, in addition to Elk-1, the other sodium orthovanadate). The cell lysate containing 0.5 ± family member of TCF, Sap-1a, has been reported to 1.0 mg of protein was incubated with the appropriate be phosphorylated by p38 and involved in the c-fos antibody overnight at 48C followed by incubation with expression through SRE in human embryonal kidney protein-A (Sigma) or protein G PLUS/protein A-agarose 293 cells (Janknecht and Hunter, 1997). It is also (Oncogene Science, Cambridge, MA) for 1 ± 3 h. The possible that Sap-1 and/or other unidenti®ed members immunoprecipitates were washed twice with RIPA bu€er of TCF regulate egr-1 expression via its SREs. Answers and twice with cold PBS. The proteins were eluted and to these questions will enhance our understanding of separated by SDS ± PAGE and transferred to polyvinyli- dene di¯uoride (PVDF) membrane. The membrane was the mechanism how signals of mitogens and stress elicit blocked with PBS containing 0.1% Tween-20 and 1% BSA the nuclear responses. before incubating with the appropriate antibodies. The Prokaryotic and eukaryotic cells respond rapidly to bound antibodies were visualized by using the Amersham environmental stress. Severe stress can inhibit DNA, ECL kit (Amersham International, Buckinghamshire, UK). RNA, and protein synthesis and can eventually kill the Also, total cell lysates were prepared by adding equal cells (Van Wijk et al., 1985, 1993). Xia et al. (1995) amount of 26 Laemlli Bu€er to the total cell lysates, reported that JNK and p38 MAP kinase activation are boiled for 5 min, separated by SDS ± PAGE and trans- involved in the induction of apoptosis in PC-12 cells. ferred to PVDF membrane. For Egr-1 (588) immunoblot, Furthermore, persistent activation of JNK1 was the membrane was blocked in PBS containing 5% skim involved in g radiation-induced apoptosis (Chen et milk and 0.1% Tween-20 overnight at 48Cbefore incubating with the egr-1 (588) antibody. al., 1996). Since p38/JNK activities and egr-1 expression are both induced, and that the Egr-1 protein is hyperphosphorylated and the DNA binding DNA transfection and CAT assays activity is enhanced, it will be interesting to see whether Transfection of the plasmids to NIH3T3 cells was egr-1 may play a role in the process of cell death, performed with lipofectamine reagent (GIBCO ± BRL Life which is currently under investigation. Technologies, Gaithersburg, MD) following the manufac- turer's instruction. pMKK3(E) expressing a constitutively activated human MKK3 mutant (De rijard et al., 1994) in Materials and methods which serine 189 and threonine 193 were replaced by glutamic acid and pcDNA3.p38 containing a mouse p38 Cell cultures and reagents cDNA (Han et al., 1994) were obtained from Dr J Han NIH3T3 cells were grown in Dulbecco's modi®ed Eagle's (The Scripps Research Institute, CA, USA). MEKK1 medium (DMEM) with 10% fetal bovine serum (FBS, expression plasmid was provided by Dr R Janknecht Hyclone, Logan, UT). The cells were rendered quiescent by (The Salk Institute, MBVL, CA, USA), and the pSRa- incubating them in DMEM containing 0.5% FBS for 48 h HA-JNK1 (Minden et al., 1994) was obtained from Dr A after the cells reached a con¯uency of about 90%. The Whitmarsh (University of Massachusetts, Worcester, MA, quiescent cells were treated with di€erent reagents and USA). The ERK2 expression plasmid pcDNA3-ERK2 was harvested for preparation of RNA, nuclear extracts or prepared by inserting the ERK2 coding sequence from total cell lysates. Recombinant mouse TNF-a was plasmid pBA4 (Her et al., 1991) into pcDNA3 (Invitrogen, purchased from Genzyme Corp. (Cambridge, MA). San Diego, CA). The Elk-1 expression plasmid MLV-Elk-1 Sodium arsenite was purchased from Sigma Chemical (Marais et al., 1993) was obtained from Dr R Treisman Corporation (St Louis, MO). Anti-phosphotyrosine anti- (Imperial Cancer Research Fund, London, UK). The cells body (4G10) was purchased from Upstate Biotechnology were co-transfected with either 0.5 mg (for p38) or 1 mg(for (Lake Placid, NY). Anti-ERK2 monoclonal antibody was JNK1) of pEGR-1.1.2-CAT, 1 mg of expression plasmids, purchased from Transduction Laboratories (Lexington, and 1 mgofpCMV-b-gal, and the total amount of DNA KY). Goat and rabbit anti-ERK2 polyclonal antibodies, was added to 5 mg per plate with control plasmid pcDNA3. anti-p38, JNK1, p-JNK (Thr-183 and Tyr-185) and Egr-1 The cells were maintained in DMEM with 0.5% FBS after (588) antibodies, and the substrates GST-ATF2 (1 ± 96) 5 h of transfection, and harvested 40 h later. The cell and GST-c-Jun (1 ± 79) for kinase assays, were purchased extracts were prepared for b-galactosidase activity and from Santa Cruz Biotechnology (Santa Cruz, CA). CAT assays. The amount of cell extract used in each CAT assay was normalized with equivalent b-galactosidase activity. Acetylated and nonacetylated forms of Preparation of RNA and Northern analysis [14C]chloramphenicol were separated by thin layer chroma- Total cytoplasmic RNA was isolated by a one-step tography, followed by autoradiography and quantitation guanidinium thiocyanate/phenol/chloroform method using a Visage 2000 Image System (BioImage Products, (Chomczynski and Sacchi, 1987) and the Northern analysis Ann Arbor, MI) and Bio-Rad GS700 Imaging Densit- was performed as described previously (Cao et al., 1992a). ometer. The blots were autoradiographed and quantitated using the Bio-Rad GS700 Imaging Densitometer. Whole cell and nuclear extraction and mobility-shift DNA binding assay Immunoprecipitation and immunoblotting Whole cell extracts were prepared for analysing the Preparation of total cell lysates, immunoprecipitation and activation of Elk-1 in a mobility shift DNA binding assay immunoblot were performed as described previously (Cao as described (Marais et al., 1993) with minor modi®cations. Stress-induced egr-1 CP Lim et al 2925 The core SRF protein was prepared from a plasmid treated with U.V. (60 J/m2 for 30 min), sodium arsenite encoding for GST-core SRF (Gille et al., 1992). The (500 mM for 60 min) or anisomycin (25 ng/ml for 45 min). GST-core SRF was anity puri®ed using glutathione Cells were lysed in 0.5 ml of RIPA bu€er with protease Sepharose beads and cleaved from GST with thrombin inhibitors listed above. The total protein (1.5 mg) was (100 U/ml) overnight at 48C.ThecoreSRFproteinwas immunoprecipitated with either 1.5 mg of anti-p38 or 2 mg washed repeatedly with 50 mM Tris-HCl (pH 7.4) using of anti-JNK1 antibodies overnight at 48C, followed by Centripep 10 (Amicon, Inc, Beverly, MA) and centrifuga- incubation with 30 ± 40 ml of protein G PLUS/protein A- tion at 3000 r.p.m. at 48C for 30 min. The puri®ed core agarose for 1 h. The immunoprecipitates were washed SRF (1.3 ng) was used in each DNA binding reaction. twice with RIPA and once with cold PBS. The beads were Transfected NIH3T3 cells were left in the DMEM media then resuspended in 50 ± 100 ml of PBS and divided into with 0.5% FBS for 40 h before harvesting. All the two halves. One half of the immunoprecipitates were procedures were performed at 48C. The cells were washed boiled in 26Laemlli Bu€er for 5 min and separated by twice with cold PBS, lysed with 75 mlofBu€erE(20mM 10% SDS ± PAGE, transferred to PVDF membrane and HEPES pH 7.9, 5 mM EDTA, 10 mM NaF, 0.1 mg/ml subjected to Western blotting/ECL. The other half of the okadaic acid, 10% glycerol, 1 mM dithiothreitol, 0.4 mM immunoprecipitates were washed with kinase assay bu€er

KCl, 0.4% Triton X-100 and protease inhibitors including (20 mM HEPES, pH 7.3, 20 mM MgCl2,20mM b- 5 mg/ml leupeptin, 5 mg/ml aprotinin, 5 mg/ml pepstatin A, glycerolphosphate, 0.2 mM sodium orthovanadate, 2 mM 1mM benzamidine, 50 mg/ml PMSF) per 60 mm dish. The DTT) and subjected to solid phase kinase assay. The extracts were sonicated using an ice-waterbath sonicator, kinase reactions were performed in 20 or 30 mlofkinase pulsed at 10 s for four times, and subjected to centrifuga- assaybu€erinthepresenceof20mMcold ATP, 5 mCi of tionat14000r.p.m.,48C for 5 min. The supernatant was [g-32P]ATP and 2 mg of substrates (GST-ATF2 for p38 or collected and used for mobility shift DNA binding assay. GST-c-Jun for JNK1 immunoprecipitates) for 15 min at The whole cell extract (25 mg) was mixed with 300 ng poly 308C as described with slight modi®cations (De rijard et (dI-dC. dI-dC), 2 mM spermidine, 1.33 mM EDTA, 1.3 ng al., 1995; Raingeaud et al., 1995; Hibi et al., 1993). The core SRF and 0.16 of Dz Bu€er (25 mM HEPES pH 7.9, reaction was stopped by adding equal amounts of

20% glycerol, 2 mM MgCl2,1mM dithiothreitol, 0.1 mM 26Laemlli Bu€er, boiled for 5 min, separated by 12%

ZnCl2 and 0.2 mM EDTA) and incubated on ice for SDS ± PAGE and transferred to PVDF membrane, 10 min. 32P-labelled Egr-1 SRE6 probe (100 000 c.p.m.) followed by autoradiography. was added to each tube and the ternary complex was allowed to form for 30 min at room temperature. The reaction was stopped by adding 3 ml of loading dye and the samples were loaded onto a 4% 40 : 1 acryla- mide : bisacrylamide gel and run for 2 h and 45 min at 48Cin0.56TBE. The gel was dried followed by autoradiography. For nuclear extraction and mobility Acknowledgements shift DNA binding assay for Egr-1 protein, quiescent cells We are grateful to Drs J Pouysse gur for p42MAPK,MJ were treated with sodium arsenite or heat shock for various Weber for pBA4 plasmid, J Han and SC Lin for plasmids times. The nuclear extracts were prepared and the mobility pMMK3(E) and pcDNA3.p38 and R Treisman for shift DNA bindng assays were carried out as described MLV.Elk-1. We thank Drs A Whitmarsh and RJ Davis previously (Dignam et al., 1983; Cao et al., 1993). for core SRF and pSRa-HA-JNK1 expression plasmids. We are also grateful to Dr R Janknecht for MEKK1 expression plasmid and Drs JC Lee and P Cohen for Kinase assays SB 203580. We want to thank Mr F Leong and Miss SY NIH3T3 cells were quiesced in 0.5% FBS/DMEM for Oh for photography. This work was supported by the 24 h prior to treatment. Cells were left untreated or National Science and Technology Board of Singapore.

References

Brunet A and Pouysse gur J. (1996). Science, 272, 1652 ± Christy BA and Nathans D. (1989b). Proc. Natl. Acad. Sci. 1655. USA, 86, 8737 ± 8741. Cano E, Hazzalin CA and Mahadevan LC. (1994). Mol. Cell. Cuenda A, Rouse J, Doza YN, Meier R, Cohen P, Gallagher Biol., 14, 7352 ± 7362. TF, Young PR and Lee JC. (1995). FEBS Lett., 364, 229 ± Cao X, Guy GR, Sukhatme VP and Tan YH. (1992a). J. Biol. 233. Chem., 267, 1345 ± 1349. Datta R, Rubin E, Sukhatme V, Qureshi S, Hallahan D, Cao X, Koski RA, Gashler A, McKiernan M, Morris CF, Weichselbaum RR and Kufe DW. (1992). Proc. Natl. Ga€ney R, Hay RV and Sukhatme VP. (1990). Mol. Cell. Acad. Sci. USA, 89, 10149 ± 10153. Biol., 10, 1931 ± 1939. De rijard B, Hibi M, Wu IH, Barrett T, Su B, Deng T, Karin Cao X, Mahendran R, Guy GR and Tan YH. (1992b). J. M and Davis RJ. (1994). Cell, 76, 1025 ± 1037. Biol. Chem., 267, 12991 ± 12997. De rijard B, Raingeaud J, Barrett T, Wu IH, Han J, Ulevitch Cao X, Mahendran R, Guy GR and Tan YH. (1993). J. Biol. RJ and Davis RJ. (1995). Science, 267, 682 ± 685. Chem., 268, 16949 ± 16957. Dignam JD, Lebovitz RM and Roeder RG. (1983). Nucleic Cao X, Tay A, Guy GR and Tan YH. (1996). Mol. Cell. Acids Res., 11, 1475 ± 1489. Biol., 16, 1595 ± 1603. Freshney NW, Rawlinson L, Guesdon F, Jones E, Cowley S, Chen YR, Meyer CF and Tan TH. (1996). J. Biol. Chem., Hsuan J and Saklatvala J. (1994). Cell, 78, 1039 ± 1049. 271, 631 ± 634. Gille H, Sharrocks AD and Shaw PE. (1992). Nature, 358, Chomczynski P and Sacchi N. (1987). Anal. Biochem., 162, 414 ± 417. 156 ± 159. Hallahan DE, Sukhatme VP, Sherman ML, Virudachalam S, Christy BA, Lau LF and Nathans D. (1988). Proc. Natl. Kufe D and Weichselbaum RR. (1991). Proc. Natl. Acad. Acad. Sci. USA, 85, 7857 ± 7861. Sci. USA, 88, 2156 ± 2160. Christy BA and Nathans D. (1989a). Mol. Cell. Biol., 9, Han J, Lee JD, Bibbs L and Ulevitch RJ. (1994). Science, 4889 ± 4895. 265, 808 ± 811. Stress-induced egr-1 CP Lim et al 2926 Her JH, Wu J, Rall TB, Sturgill TW and Weber MJ. (1991). Marais R, Wynne J and Treisman R. (1993). Cell, 73, 381 ± Nucleic Acids Res., 19, 3743. 393. Herschman HR, Lim RW, Brankow DW and Fujiki H. Marshall CJ. (1994). Curr. Opin. Genet. Dev., 4, 82 ± 89. (1989). Carcinogenesis, 10, 1495 ± 1498. Milbrandt J. (1987). Science, 238, 797 ± 799. Hibi M, Lin A, Smeal T, Minden A and Karin M. (1993). Minden A, Lin A, McMahon M, Lange-Carter C, De rijard Genes & Dev., 7, 2135 ± 2148. B, Davis RJ, Johnson GL and Karin M. (1994). Science, Hill CS and Treisman R. (1995). Cell, 80, 199 ± 211. 266, 1719 ± 1723. Huang RN, Ho IC, Yih LH and Lee TC. (1995). Environ. Nguyen HQ, Ho€man-Liebermann B and Liebermann DA. Mol. Mutagen., 25, 188 ± 196. (1993). Cell, 72, 197 ± 209. Huang RP and Adamson ED. (1995). Oncogene, 10, 467 ± Pandey P, Raingeaud J, Kaneki M, Weichselbaum R, Davis 475. RJ, Kufe D and Kharbanda S. (1996). J. Biol. Chem., 271, Janknecht R, Ernst WH, Pingoud V and Nordheim A. 23775 ± 23779. (1993). EMBO J., 12, 5097 ± 5104. Payne DM, Rossomando AJ, Martino P, Erickson AK, Her Janknecht R and Hunter T. (1997). EMBO J., 16, 1620 ± JH, Shabanowitz J, Hunt DF, Weber MJ and Sturgill TW. 1627. (1991). EMBO J., 10, 885 ± 892. Johnston D, Oppermann H, Jackson J and Levinson W. Raingeaud J, Gupta S, Rogers JS, Dickens M, Han J, (1980). J. Biol. Chem., 255, 6975 ± 6980. Ulevitch RJ and Davis RJ. (1995). J. Biol. Chem., 270, Kharbanda S, Ren R, Pandey P, Shafman TD, Feller SM, 7420 ± 7426. Weichselbaum RR and Kufe DW. (1995a). Nature, 376, Rouse J, Cohen P, Trigon S, Morange M, Alonso- 785 ± 788. Llamazares A, Zamanillo D, Hunt T and Nebreda AR. Kharbanda S, Saleem A, Rubin E, Sukhatme V, Blenis J and (1994). Cell, 78, 1027 ± 1037. Kufe D. (1993). Biochemistry, 32, 9137 ± 9142. Shaw PE, SchroÈ ter H and Nordheim A. (1989). Cell, 56 563 ± Kharbanda S, Saleem A, Shafman T, Emoto Y, Taneja N, 572. Rubin E, Weichselbaum R, Woodgett J, Avruch J, Stokoe D, Macdonald SG, Cadwallader K, Symons M and Kyriakis J and Kufe D. (1995b). J. Biol. Chem., 270, Hancock JF. (1994). Science, 264, 1463 ± 1467. 18871 ± 18874. Sukhatme VP. (1990). J. Am. Soc. Nephrol., 1, 859 ± 866. Kyriakis JM, Banerjee P, Nikolakaki E, Dai T, Rubie EA, Sukhatme VP, Cao X, Chang LC, Tsai-Morris CH, Ahmad MF, Avruch J and Woodgett JR. (1994). Nature, Stamenkovich D, Ferreira PC, Cohen DR, Edwards SA, 369, 156 ± 160. Shows TB, Curran T, Le Beau MM and Adamson ED. Lange-Carter CA, Pleiman CM, Gardner AM, Blumer KJ (1988). Cell, 53, 37 ± 43. and Johnson GL. (1993). Science, 260, 315 ± 319. Treisman R. (1995). EMBO J., 14, 4905 ± 4913. Lee JC, Laydon JT, McDonnell PC, Gallagher TF, Kumar S, Tsai-Morris CH, Cao X and Sukhatme VP. (1988). Nucleic Green D, McNulty D, Blumenthal MJ, Heys JR, Land- Acids Res., 16, 8835 ± 8846. vatter SW, Strickler JE, McLaughlin MM, Siemens IR, Van Wijk R, Otto AM and Jimenez de Asua L. (1985). Int. J. Fisher SM, Livi GP, White JR, Adams JL and Young PR. Hyperthermia, 1, 147 ± 156. (1994). Nature, 372, 739 ± 746. Van Wijk R, Welters M, Souren JEM, Ovelgonne H and Leevers SJ, Paterson HF and Marshall CJ. (1994). Nature, Wiegant FAC. (1993). J. Cell. Physiol., 155, 265 ± 272. 369, 411 ± 414. Wang XZ and Ron D. (1996). Science, 272, 1347 ± 1349. Lemaire P, Revelant O, Bravo R and Charnay P. (1988). Whitmarsh AJ, Yang SH, Su MSS, Sharrocks AD and Davis Proc. Natl. Acad. Sci. USA, 85, 4691 ± 4695. RJ. (1997). Mol. Cell. Biol., 7, 2360 ± 2371. Lim RW, Varnum BC and Herschman HR. (1987). Whitmarsh AJ, Shore P, Sharrocks AD and Davis RJ. Oncogene, 1, 263 ± 270. (1995). Science, 269, 403 ± 407. Mansour SJ, Matten WT, Hermann AS, Candia JM, Rong S, Xia Z, Dickens M, Raingeaud J, Davis RJ and Greenberg Fukasawa K, Vande Woude GF and Ahn NG. (1994). ME. (1995). Science, 270, 1326 ± 1331. Science, 265, 966 ± 970.