Oncogene (2001) 20, 8116 ± 8124 ã 2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00 www.nature.com/onc

Crosstalk between Myc and activating transcription factor 2 (ATF2): Myc prolongs the half-life and induces phosphorylation of ATF2

Josef Miethe1, Claudia Schwartz1, Katja Wottrich1, Dorit Wenning1 and Karl-Heinz Klempnauer*,1

1Institut fuÈr Biochemie, WestfaÈlische-Wilhelms-UniversitaÈtMuÈnster, Wilhelm-Klemm-Str. 2, D-48149 MuÈnster, Germany

Myc is a key regulator of cell growth, di€erentiation and Myc proteins are nuclear phosphoproteins that display apoptosis, and a€ects cell fate decisions by activating as all the characteristics of transcription factors. The well as by inhibiting the expression of cellular genes. carboxyterminus of Myc functions as a basic region ± Myc is a member of the basic region ± helix ± loop ± helix ± loop ± helix ± leucine zipper (b-HLH-Zip) DNA- helix ± leucine zipper (b-HLH-Zip) class of transcription binding domain, which heterodimerizes with the Max factors, which heterodimerizes with the Max protein and protein (Blackwood and Eisenman, 1991; LuÈ scher and recognizes a consensus Myc binding motif. Stimulation Larsson, 1999; Prendergast and Zi€, 1991) and of gene expression by Myc is thought to be mediated by recognizes a consensus Myc binding site. The amino- direct binding of Myc ± Max heterodimers to speci®c terminus of Myc contains a transcriptional activation target genes. So far, only a few genes have been domain (Kato et al., 1990) which is responsible for the identi®ed as direct binding targets of Myc, raising the activation of promoters containing Myc binding sites possibility that Myc a€ects gene expression also by (Amati et al., 1992; Kretzner et al., 1992). Besides Max indirect mechanisms. In this work we present evidence a number of other cellular proteins have been shown to that v-Myc encoded by the avian retrovirus MC29 interact with di€erent domains of Myc and to stimulates activating transcription factor 2 (ATF2)- modulate its transcriptional properties (Sakamuro and dependent transcription. Analysis of the e€ect of Myc Prendergast, 1999). on ATF2 shows that v-Myc prolongs the half-life of It is widely believed that Myc exerts some of its ATF2 and induces the phosphorylation of N-terminal functions by direct binding to and activation of cellular sites of ATF2 (Thr-69 and Thr-71) which have previously target genes (Cole and McMahon, 1999; Dang, 1999; been identi®ed as the target sites of stress-activated Facchini and Penn, 1998; Grandori and Eisenman, protein kinases and implicated in the regulation of ATF2 1997). However, so far very few genes have unambigu- activity. Taken together, our results suggest that v-Myc ously been shown to be direct Myc-regulated genes. can a€ect gene expression indirectly by modulating the Recent DNA microarray analyses suggest that Myc activity of ATF2. Oncogene (2001) 20, 8116 ± 8124. a€ects the expression of a large number of genes (Coller et al., 2000; Boon et al., 2001; Schuhmacher et Keywords: Myc; ATF2; phosphorylation; protein al., 2001). The molecular mechanisms by which Myc turnover regulates these genes have not been elucidated, however it is reasonable to suspect that Myc a€ects at least some of these genes by indirect mechanisms. Introduction Besides acting as a transcriptional activator Myc also downregulates the expression of certain genes (Claassen The c-myc gene is a key regulator of cell growth and and Hann, 1999; Facchini and Penn, 1998). Some of plays a central role as a switch by which normal cells the genes repressed by Myc are themselves known to decide between alternative fates, such as proliferation, play critical roles during proliferation or di€erentia- di€erentiation and apoptosis (Amati and Land, 1994; tion, suggesting that some of the biological functions of Bouchard et al., 1998; Dang, 1999; Evan and Little- Myc are related to gene repression rather than wood, 1993; Henriksson and LuÈ scher, 1996; Schmidt, activation. A recent example for the inhibitory role 1999). Numerous studies have shown that altered of Myc is the downregulation of the Cdk inhibitor forms of the gene, such as retrovirally transduced (v- p15INK4b (Seoane et al., 2001; Staller et al., 2001). myc) or rearranged myc alleles, disturb the balance We have been studying the mechanisms by which the between these processes, resulting in tumorigenesis. v-Myc protein of the avian retrovirus MC29 regulates cellular gene expression. Recently, we have shown that Myc downregulates the expression of genes which are normally activated during myeloid di€erentiation by *Correspondence: K-H Klempnauer; interfering with the activity of several members of the E-mail: [email protected] Received 26 June 2001; revised 5 September 2001; accepted 13 CCAAT box/enhancer binding protein (C/EBP) family September 2001 (Mink et al., 1996). We show here that Myc can also Crosstalk between Myc and ATF2 J Miethe et al 8117 stimulate the activity of other transcription factors. We Finally, we were interested to know whether the present evidence that Myc increases the activity of the stimulatory e€ect on ATF2-dependent transcription activating transcription factor 2 (ATF2). ATF2 is a was also observed for c-Myc. We therefore directly member of the ATF/CREB protein family of basic compared the e€ects elicited by v-Myc and chicken c- region ± leucine zipper proteins (Hai et al., 1989; Myc. As shown in Figure 2c,d both proteins Maekawa et al., 1989) which play important roles in stimulated ATF2-dependent transcription to the same the cellular stress response (Gupta et al., 1995; extent. Livingston et al., 1995; van Dam et al., 1993). Several genes regulated by ATF2 have been identi®ed, including Myc increases the half-life of Gal4/ATF2 the c-jun, TNFa, TGFb, cyclin A, E-selectin and DNA polymerase b genes (Kim et al., 1992; Narayan et al., As a ®rst step to ®nding an explanation for the 1994; Read et al., 1997; Shimizu et al., 1998; Tsai et al., apparent increase of the activity of Gal4/ATF2, we 1996; van Dam et al., 1993). Although some of these determined the amounts of Gal4/ATF2 expressed in ATF2 target genes have been implicated in the the absence or the presence of Myc. To our surprise we development of neoplasia, the possible contribution of found that the amount of Gal4/ATF2 in the ATF2 to this process so far is largely uncharacterized. transfected cells was substantially increased in the We have begun to address the mechanism by which v- presence of Myc (see left panel of Figure 3a). Figure 3a Myc stimulates ATF2. Our results show that v-Myc also shows that the level of endogenous ATF2 in the prolongs the half-life of ATF2 and causes increased transfected cells was not signi®cantly a€ected by Myc. phosphorylation of ATF2 at sites that have been shown Under the transfection conditions used only a small to be crucial for the regulation of ATF2 activity. Taken percentage of the cells (usually less than 5%) actually together, our results suggest a novel mechanism by which take-up the transfected DNA. The endogenous ATF2, v-Myc might regulate the expression of cellular genes. visible in Figure 3a, is therefore predominantly (95%) derived from non-transfected cells, thus explaining the apparent di€erence in the behavior of endogenous and exogenous ATF2. Results To explain the increase of the amount of Gal4/ATF2 in the presence of Myc we ®rst considered the Myc stimulates transactivation mediated by Gal4/ATF2 possibility that Myc might a€ect the CMV promoter Previous work from our laboratory has shown that which drives the expression of Gal4/ATF2. However, Myc inhibits the activity of C/EBPb and of other C/ in a similar transfection experiment there was no EBP family members, leading to the downregulation of obvious e€ect of Myc on the amount of C/EBPb several C/EBP-dependent genes by Myc (Mink et al., expressed from a CMV promoter-containing expression 1996). In the course of this work we have also studied vector (see right panel of Figure 3a). Furthermore, we the e€ects of Myc on several other transcription also did not observe any e€ect of Myc on the reporter factors, including a Gal4/ATF2 fusion construct. plasmid CMVb, which contains the CMV promoter Surprisingly, we found that Myc has a strong fused to the b-galactosidase gene and was used in all enhancing e€ect on the activity of Gal4/ATF2 (Figure transfections to monitor the transfection eciency. 1a). The stimulatory e€ect of Myc was speci®c for the Thus, we concluded that Myc does not increase the ATF2 part of the fusion protein, as other Gal4 fusion activity of the CMV promoter. To explain the elevated proteins, such as Gal4/VP16 (Figure 1b) or Gal4/c- level of Gal4/ATF2 in the presence of Myc we Myb (Figure 1c) were not a€ected signi®cantly by Myc. therefore considered the alternative possibility that To explore the speci®city of the stimulatory e€ect of Myc might increase the half-life of Gal4/ATF2. To Myc on the activity of Gal4/ATF2 we analysed a series address this possibility we performed pulse ± chase of Myc as well as Gal4/ATF2 deletion mutants. Figure experiments using cells transfected with the Gal4/ 2a shows an analysis of a set of Myc mutants, which ATF2 expression vector alone or in combination with included several deletion mutants lacking amino- or Myc. Transfected cells were labeled for 1 h with 35S- carboxyterminal sequences (Myc-DN, Myc-DPA, Myc- methionine. Subsequently, they were transferred to a D45 ± 83, Myc-D45 ± 97, Myc-DLZ), as well as point growth medium lacking radioactive methionine and mutants of the HLH region (Myc-H1, Myc-H2). We supplemented with 1 mM of unlabeled methionine. The found that none of the mutants had completely lost the cells were then incubated in this medium for di€erent ability to increase the activity of Gal4/ATF2, but that periods of time, after which they were lysed and two of the mutants, the amino-terminal deletion mutant subjected to immunoprecipitation using Gal4-speci®c DPA and the carboxyterminal deletion mutant DLZ, had antibodies. Finally, the immunoprecipitated proteins only approximately 20% of the stimulatory activity of wt were visualized by SDS ± PAGE and autoradiography. Myc. As shown in Figure 2b all Myc mutants were As shown in Figure 3b, the band corresponding to expressed at similar levels, in particular those mutants Gal4/ATF2 decreased much faster in the absence of which showed reduced activity were not expressed at Myc than in its presence (compare left and right panels lower levels. Thus, it appears that at least two regions of at top of Figure 3b). Quantitative analysis (bottom of Myc, an N-terminal region as well as the leucine zipper Figure 3b) also showed that the half-life of Gal4/ATF2 domain contribute to the stimulatory e€ect. was less in the absence than in the presence of Myc.

Oncogene Crosstalk between Myc and ATF2 J Miethe et al 8118

Figure 1 Myc stimulates transactivation mediated by Gal4/ATF2. (a) QT6 cells were transfected with 3 mg of the Gal4-dependent reporter gene pG5E4 ± 38Luc and 0.2 mg of the b-galactosidase plasmid pCMVb. In addition, cells were transfected with 3 mg of the Gal4/ATF2 expression vector pGal4/ATF2 (white bars) or, for control, with the same amount of expression vector pSG424 encoding only the Gal4 DNA-binding domain (black bars). Furthermore, cells were transfected with 1 mg of the v-Myc expression vector pCDNA3wtMyc (+Myc) or the empty expression vector pCDNA3 (7 Myc), as indicated at the bottom. Cells were analysed for luciferase and b-galactosidase activities 24 h after transfection. The columns show the average luciferase activity (in arbitrary units) relative to the activity of the co-transfected pCMVb plasmid. Thin lines indicate standard deviations. (b) Cells were transfected as in a except that the Gal4/ATF2 expression vector was replaced by Gal4/VP16 (black bars, 0.1 mg). (c) Cells were transfected as in a except that the Gal4/ATF2 expression vector was replaced by Gal4/c-mybD3 (black bars, 3 mg)

Based on the initial decrease of the amounts of mediated by stress-activated protein kinases (Fuchs radiolabeled Gal4/ATF2, we estimate the half-life of et al., 1997, 2000; Fuchs and Ronai, 1999). In the protein to be less than 1 h in the absence of Myc particular, it has been reported that phosphorylation and approximately 3 h in its presence. of ATF2 at Thr-69 and Thr-71 by JNK/p38- Since both ATF2 as well as Myc are degraded dependent phosphorylation renders ATF2 resistant to through the ubiquitin ± proteasome pathway (Ciechan- ubiquitination and degradation. These observations over et al., 1991; Firestein and Feuerstein, 1998; suggested the possibility that Myc might increase the Salghetti et al., 1999) we considered the possibility stability of ATF2 by inducing N-terminal phosphor- that Myc and ATF2 might compete with each other for ylation of ATF2. To determine whether Myc a€ects a limiting component of the degradation machinery. If the phosphorylation at Thr-69 and Thr-71 of Gal4/ this was the case we would expect not only Myc to ATF2 we used an antibody that speci®cally recognizes prolong the half-life of Gal4/ATF2 but also Gal4/ ATF2 phosphorylated at these aminoacid residues but ATF2 to delay the degradation of Myc and, hence, to does not react with unphosphorylated ATF2. Cells increase the amount of Myc. Figure 3c shows clearly were transfected with expression vectors for Gal4/ that this was not the case. The amount of ATF2 was ATF2 and Myc or for Gal4/ATF2 alone, followed by increased in the presence of Myc, however, at the same Western blot analysis of cell extracts using antibodies time the amount of Myc was not a€ected, arguing speci®c for ATF2 (Figure 4, left panel) or recognizing against competition of both proteins for the degrada- only phosphorylated ATF2 (Figure 4, right panel). In tion machinery as an explanation for the stabilizing this experiment we adjusted the amounts of cell e€ect of Myc on ATF2. extracts loaded onto the gel such that similar amounts of total Gal4/ATF2 were present in the two di€erent lanes (left panel). As shown in the right panel of Myc induces the phosphorylation of Gal4/ATF2 Figure 4, the amount of Gal4/ATF2 phosphorylated Recent work has linked the rate of degradation of at Thr-69 and Thr-71 was indeed increased in the ATF2 to N-terminal phosphorylation of ATF2 presence of Myc.

Oncogene Crosstalk between Myc and ATF2 J Miethe et al 8119

Figure 2 E€ect of Myc mutants on Gal4/ATF2. (a) QT6 cells were transfected with 3 mg of the Gal4-dependent reporter gene pG5E4 ± 38Luc, 0.2 mg of the b-galactosidase plasmid pCMVb, and 3 mg of expression vectors pGal4/ATF2 or pSG424. In addition, cells were transfected with 1 mg of expression vectors for wild-type v-Myc (Myc-wt) or for the Myc mutants indicated at the bottom. Twenty-four hours after transfection cells were analysed for luciferase and b-galactosidase activities. The columns show the stimulation of the activity of Gal4/ATF2 by wt or mutant v-Myc. The stimulation of Gal4/ATF2 by wt v-Myc was designated as 100. Thin lines indicate standard deviations. (b) Aliquots of total cellular protein of cells transfected as in A were fractionated by SDS ± PAGE and analysed by Western blotting using antibodies against Myc. Transfection eciencies were normalized by analysing the activity of the co-transfected b-galactosidase reporter gene CMVb. The mutant Myc proteins correspond to the main protein band in each lane. Faster migrating bands represent degradation products. (c) QT6 cells were transfected as described in a except that 1 mg of expression vector for v-Myc or c-Myc was used, as indicated at the bottom. Twenty-four hours after transfection cells were analysed for luciferase and b-galactosidase activities. The columns show the stimulation of the activity of Gal4/ATF2 by v- Myc or c-Myc. The stimulation of Gal4/ATF2 by v-Myc was designated as 100. Thin lines indicate standard deviations. (d) Aliquots of total cellular protein of cells transfected as in c were fractionated by SDS ± PAGE and analysed by Western blotting as described in b

To further investigate the role of phosphorylation at Analysis of the amounts of the wild-type or mutant Thr-69 and Thr-71 of ATF2 in mediating the Myc Gal4/ATF2 fusion proteins in the presence or absence e€ect we used a mutant of Gal4/ATF2, in which the of Myc showed that the mutant protein still accumu- threonine residues at positions 69 and 71 of ATF2 have lated in the presence of Myc, although the levels been mutated to alanine. Figure 5a shows that the reached were somewhat lower than in case of the wild- overall transactivation activity of the mutant Gal4/ type protein (Figure 5b). Taken together these results ATF2 was much lower than that of the unmutated suggest that Myc stimulates the transactivation protein, both in the absence or in the presence of Myc. potential of Gal4/ATF2 and causes accumulation of This indicates that the transactivation domain of ATF2 the protein, in part, by mechanisms which are is required for the Myc-dependent stimulation. dependent on phosphorylation at residues Thr-69 and Although Myc also a€ects the mutated protein the Thr-71. However, it is also evident that the phosphor- magnitude of stimulation was substantially less than ylation of these sites does not account for all of the observed in case of the unmutated protein (7.3-fold in e€ects induced by Myc. case of Gal4/ATF2wt versus 3.3-fold in case of Gal4/ Most of the experiments described above were ATF2(Ala69,71). This demonstrates that the stimula- performed by using arti®cial expression vectors tory e€ect of Myc depends, at least in part, on transfected transiently into cells, a procedure which increased phosphorylation of Gal4/ATF2 at Thr-69 can give rise to artifacts due to the relatively high levels and Thr-71. However, phosphorylation at Thr-69 and of expression of the proteins involved. We were Thr-71 does not account for all of the stimulatory therefore interested to see whether an e€ect of v-Myc e€ect of Myc, because transactivation mediated by the on ATF2 could also be observed under more mutant Gal4/ATF2 is still stimulated by Myc, although physiological conditions. To address this question we less eciently. analysed the amount and phosphorylation state of

Oncogene Crosstalk between Myc and ATF2 J Miethe et al 8120

Figure 3 v-Myc expression prolongs the half-life of Gal4/ATF2. (a) QT6 cells were transfected with 1 mg each of expression vectors for Gal4/ATF2 (pCDNA3 ± Gal4/ATF2) or C/EBPb (pCDNA3 ± CCR) and v-Myc (pCDNA3wtMyc), as indicated at the bottom. Transfection eciencies were normalized by co-transfecting 0.5 mg of the b-galactosidase reporter gene CMVb. Twenty- four hours after transfection total cell extracts were fractionated by SDS ± PAGE and analysed by Western blotting using ATF2- speci®c or C/EBPb-speci®c antibodies. Gal4/ATF2 and C/EBPb are marked by arrowheads. Endogenous ATF2 is seen as a minor band migrating slightly ahead of Gal4/ATF2. The band visible at the top of b is unspeci®c. Endogenous C/EBPb is not visible under these conditions. (b) QT6 cells were transfected as in a, except that 24 h after transfection cells were radiolabeled with 35S-methionine for 1 h. Cells were then chased with an excess of unlabeled methionine for 0 h (lanes 1 and 5), 1 h (lanes 2 and 6), 2 h (lanes 3 and 7) and 4 h (lanes 4 and 8). Cell extracts were then immunoprecipitated with Gal4-speci®c antibodies and analysed by SDS ± PAGE and autoradiography (top). The radioactive bands corresponding to Gal4/ATF2 (arrow) were then quantitated by densitometry and the decrease of Gal4/ATF2 in the absence (crosses) or presence (circles) of Myc is shown against the time of chase (bottom). The band intensities at 0 h chase were designated as 100%. Lanes 1 ± 4: Gal4/ATF2 in the absence of Myc; Lanes 4 ± 8: Gal4/ATF2 in the presence of Myc. (c) QT6 cells were transfected with 1 mg each of expression vectors for full-length Gal4/ATF2 (pCDNA3-Gal4/ATF2) and v-Myc (pCDNA3wtMyc), as indicated at the bottom. Twenty- four hours after transfection total cell extracts were fractionated by SDS ± PAGE and analysed by Western blotting using Gal4- or Myc-speci®c antibodies. The bands corresponding to Gal4/ATF2 and Myc are marked by black and white arrows, respectively

endogenous ATF2 in BM2 cells, a myelomonocytic chicken cell line transformed by the v-myb oncogene, and of a v-Myc expressing derivative of the BM2 line, referred to as BM2(MC29)cl.4. These cells were obtained by infecting BM2 cells with the v-myc containing retrovirus MC29 (Symonds et al., 1986) and have been used before to demonstrate the inhibitory e€ect of Myc on the expression of C/EBP Figure 4 Myc causes increased phosphorylation of Gal4/ATF2 regulated genes (Mink et al., 1996). As shown by at Thr 69 and 71. QT6 cells were transfected with expression Western blotting, both the amount of total ATF2 as vectors for full-length Gal4/ATF2 (pCDNA3 ± Gal4/ATF2, 3 mg) well as of phosphorylated ATF2 was increased in the and v-Myc (pCDNA3 ± wtMyc, 1 mg), as indicated at the bottom. Twenty-four hours after transfection total cell extracts were BM2(MC29)cl.4 cells compared to the original BM2 fractionated by SDS ± PAGE and analysed by Western blotting line. By contrast, an unrelated protein, v-Myb, was not using ATF2-speci®c antibodies (left) or antibodies speci®c for a€ected (Figure 6a). ATF2 phosphorylated at Thr-69 and -71 (right). The amounts of In addition to these myeloid cells we analysed the cell extracts loaded on the gel were adjusted such that the same levels of ATF2 and phosphorylated ATF2 in Myc amount of total ATF2 was present in the di€erent lanes. The bands corresponding to Gal4/ATF2 and to endogenous ATF2 are knock-out ®broblasts and in the corresponding par- marked by black and white arrows, respectively ental cell line (Mateyak et al., 1997). As shown in

Oncogene Crosstalk between Myc and ATF2 J Miethe et al 8121

Figure 5 E€ect of v-Myc on the T69,71A mutant of Gal4/ ATF2. (a) QT6 cells were transfected with 3 mg of the Gal4/ dependent reporter gene pG5E4 ± 38Luc, 0.2 mg of the b- galactosidase plasmid pCMVb and 0.1 mg each of expression vectors for wild-type (pCDNA3 ± Gal4/ATF2) or the T69,71A mutant of Gal4/ATF2 (pCDNA3 ± Gal4/ATF2 T69,71A). In addition cells were transfected either with 1 mg of expression vector for v-Myc (pCDNA3wtMyc) or with 1 mg of the empty expression vector (pCDNA3). Twenty-four hours after transfec- tion cells were analysed for luciferase and b-galactosidase activities. The columns show the average luciferase activity (in arbitrary units) relative to the activity of the co-transfected pCMVb plasmid. Thin lines indicate standard deviations. The numbers above the columns give the stimulation of the corresponding Gal4 construct by Myc. (b) QT6 cells were transfected with 1 mg each of expression vectors for wild-type or the T69,71A mutant of Gal4/ATF2 and v-Myc or the corresponding empty vector, as indicated at the bottom. Transfection eciencies were normalized Figure 6 Analysis of endogenous ATF2 in the presence or by cotransfecting 0.5 mg of the b-galactosidase reporter gene absence of v-Myc. (a) Equivalent amounts of nuclear extracts CMVb. Twenty-four hours after transfection total cell extracts from BM2 cells (B) or from BM2(MC29)cl.4 cells (M) were were fractionated by SDS ± PAGE and analysed by Western fractionated by SDS ± PAGE and analysed by Western blotting blotting using ATF2-speci®c antibodies. The bands corresponding using antisera speci®c for ATF2 (left) or for ATF2 phosphory- to Gal4/ATF2 and to endogenous ATF2 are marked by black lated at Thr-69 and Thr-71 (middle) or for v-Myb (right). The and white arrows, respectively black and white arrows mark endogenous ATF2 and v-Myb, respectively. (b) Equivalent amounts of extract from Myc+/+ (wt) or Myc 7/7 (ko) cells were analysed as in a using antisera speci®c for ATF2 (left) or for ATF2 phosphorylated at Thr-69 Figure 6b, the levels of total ATF2, as well as of and Thr-71. The black arrow marks endogenous ATF2 phosphorylated ATF2, were reduced in the cells lacking Myc protein. These results are consistent with our other observations and support the notion that 2001; Schuhmacher et al., 2001). According to the Myc a€ects the amount and the phosphorylation status prevailing view gene activation by Myc is due to direct of ATF2. binding of Myc to its target genes, mediated by the DNA-binding domain located at the C-terminus of Myc (Cole and McMahon, 1999; Dang, 1999; Facchini Discussion and Penn, 1998; Grandori and Eisenman, 1997). Although the expression of many genes is a€ected by Myc is a key regulator of cell fate decisions which Myc so far only few genes whose expression is allow cells to switch between alternative fates such as activated by Myc as a result of direct Myc binding proliferation, di€erentiation or apoptosis. Myc is have been unambiguously identi®ed. It is, therefore, thought to act by activating the expression of certain likely that Myc regulates the expression of at least cellular genes, as well as by downregulating the some of its target genes indirectly. expression of others (Amati and Land, 1994; Bouchard In the work described here we have explored the et al., 1998; Dang, 1999; Evan and Littlewood, 1993; possibility that Myc activates cellular gene expression Henriksson and LuÈ scher, 1996; Schmidt, 1999). Indeed, by such indirect mechanisms. We have found that v- large-scale expression analyses have shown that many Myc stimulates the activity of the transcription factor genes are di€erentially expressed in Myc overexpressing ATF2, suggesting a novel mechanism for gene versus normal cells (Coller et al., 2000; Boon et al., activation by v-Myc. Together with our previous

Oncogene Crosstalk between Myc and ATF2 J Miethe et al 8122 work, which showed that v-Myc inhibits the activity e€ect of Myc) but, as shown in Figure 6, has only a of members of the C/EBP family (Mink et al., 1996), minor e€ect on the Myc-induced accumulation of our data suggest that v-Myc a€ects the expression of ATF2. At ®rst glance, this suggests that the increase cellular genes by modulating the activity of other in the half-life and the stimulation of phosphorylation transcription factors, such as ATF2 or C/EBP. of ATF2 are independent events. However, it is v-Myc has at least two e€ects on ATF2: Firstly, v- possible that Myc induces phosphorylation of ATF2 Myc induces the phosphorylation of N-terminal sites at other as yet unknown sites and that phosphoryla- of ATF2 (Thr-69 and Thr-71) which have previously tion of these sites also contributes to increased been shown to play key roles in the regulation of stability of the protein. In this case, stabilization of ATF2 activity. Phosphorylation at Thr-69 and Thr-71 ATF2 would simply be a consequence of phosphor- by stress-activated kinases stimulates the transactiva- ylation induced by Myc. However, it is also possible tion potential of ATF2 and is crucial for the ability of that induction of phosphorylation and stabilization of ATF2 to mediate gene activation in response to stress ATF2 are independent activities of Myc which both (Gupta et al., 1995; Livingston et al., 1995; van Dam contribute to the increase in ATF2-mediated transac- et al., 1995). In addition, phosphorylation of ATF2 at tivation. Thr-69 and Thr-71 has been implicated in the At present, we have not yet identi®ed the mechan- regulation of the degradation of ATF2 by a isms by which Myc causes stabilization and phosphor- ubiquitin ± proteasome dependent mechanism (Fuchs ylation of ATF2. As discussed above, stabilization et al., 1997, 2000; Fuchs and Ronai, 1999). Very could be simply a consequence of phosphorylation or recently, ATF2 has been shown to possess intrinsic could be induced independently. Since ATF2 is histone acetyltransferase activity which is also in¯u- degraded by the ubiquitin-proteasome pathway, one enced by phosphorylation of these N-terminal sites possibility could be a modulation of the activity of (Kawasaki et al., 2000). Thus, phosphorylation at some component(s) of this pathway. In this respect it Thr-69 and Thr-71 of ATF-2 is a central event for the would be very interesting to know if the stability of regulation of ATF2 activity. The relevance of these other proteins which are degraded by the same phosphorylation sites is further substantiated by our pathway are also a€ected by Myc. A possible scenario own data which implicate the phosphorylation of Thr- to explain the increase in phosphorylation of ATF2 by 69 and Thr-71 in mediating the e€ects of v-Myc on Myc envisions a stimulation of the activity of a ATF2. We can exclude the trivial possibility that the member of the stress-activated protein kinase family phosphorylation of Thr-69 and Thr-71 simply re¯ects by Myc, such as JNK or p38. To address this a stress response, caused by the transfection proce- possibility it will be interesting to directly measure dure. In this case, we would expect control transfec- the activity of these kinases in the presence or absence tions (i.e. transfections lacking the v-Myc expression of Myc. vector) to have the same e€ect on the phosphorylation Finally, it is worth considering potential target genes status of ATF2. whose expression might be a€ected by v-Myc via Secondly, v-Myc increases the half-life of ATF2 ATF2. One interesting potential target in this respect is and, thereby, causes an increase of the ATF2 protein the cyclin A gene. Cyclin A expression is increased in a levels. Our results show that the e€ect of v-Myc on number of experimental cell systems overexpressing the half-life of ATF2 is not simply due to overloading Myc (Daksis et al., 1994; Hoang et al., 1994; Jansen- of the cellular protein degradation machinery caused DuÈ rr et al., 1993). It has been suggested that the by competition of ATF2 and v-Myc, which is itself a stimulation of the cyclin A promoter by Myc is short-lived protein (Ciechanover et al., 1991; Firestein mediated by the so-called CDE- and CHR-elements and Feuerstein, 1998; Salghetti et al., 1999), for some and involves a cyclin E-dependent kinase (Rudolph et limiting component of this machinery. If this was the al., 1996; Zwicker et al., 1995). Interestingly, however, case we would expect also to see an e€ect of ATF2 on the cyclin A promoter also contains a conserved the half-life of v-Myc, which was not observed. At binding site for ATF2 in addition to these elements. present, we do not know whether or not the two Moreover, functional analyses have shown that the e€ects of v-Myc which we have detected ± increase of ATF2 binding site is essential for up-regulation of the half-life and induction of phosphorylation ± represent, cyclin A promoter in growth-stimulated human ®bro- in fact, the two sides of the same coin. As mentioned blasts (Nakamura et al., 1995) and that ATF2 can above, phosphorylation of ATF2 at Thr-69 and Thr- stimulate the expression of the cyclin A gene (Shimizu 71 has been shown to extend the half-life of the et al., 1998). Taken together, these observations raise protein (Ciechanover et al., 1991; Firestein and the possibility that the stimulation of cyclin A Feuerstein, 1998; Salghetti et al., 1999); thus the expression by Myc is mediated, at least in part, by accumulation of ATF2 in the presence of Myc could ATF2. simply be a consequence of the phosphorylation. In conclusion, we have described a novel mechan- Mutation of Thr-69 and Thr-71 to alanine residues ism for gene activation by v-Myc. Our work suggests has a signi®cant e€ect on the stimulation of the that Myc a€ects cellular gene expression not only by ATF2-dependent transactivation by Myc (which direct binding to speci®c target genes but also strongly suggests that phosphorylation of Thr-69 and indirectly by modulating the activity of other tran- Thr-71 is indeed involved in mediating the stimulatory scription factors.

Oncogene Crosstalk between Myc and ATF2 J Miethe et al 8123 24 h after transfection. Preparation of cell extracts, luciferase Materials and methods and b-galactosidase assays were performed as described (Burk et al., 1993). Eukaryotic expression vectors Expression vectors for the Gal4 DNA-binding domain Antibodies and Western blotting (pSG424) and for full-length Gal4/ATF2 (pGal4/ATF2) have been described before (Li and Green, 1996; Liu and Green, Monoclonal antibodies against the Gal4 DNA-binding 1990, 1994; Sadowski and Ptashne, 1989) and were kindly domain (RK5C1) and polyclonal antisera against ATF2 (N- provided by M Green. pCDNA3 ± Gal4/ATF2 is a derivative 96) or phosphorylated ATF2 (F-1) were obtained from Santa of pSGATF-2, obtained by subcloning the Gal4/ATF2 Cruz Biotechnology. Polyclonal Myc-speci®c antiserum raised coding region into expression vector pCDNA3 (Invitrogen). against the C-terminus of v-Myc was kindly provided by K A Gal4/ATF expression vector carrying the T69A and T71A Bister. Polyclonal Myb-speci®c antiserum has been described point mutations (Gupta et al., 1995) was kindly provided by before (Klempnauer et al., 1983). Total cell extract was R Davis. pGal4/VP16 encodes a Gal4/VP16 hybrid protein prepared by lysing cells in SDS sample bu€er and boiling and was obtained from P Chambon. pGal4/c ± MybD3 them for 5 min. Extracts were analysed by SDS ± PAGE and encodes a Gal4/chicken c-Myb fusion protein lacking the Western blotting as described (Mink et al., 1996). Nuclei of C-terminus of c-Myb. pGal4/c-MybD3 was constructed by BM2 and BM2(MC29)cl.4 cells were prepared by washing the fusing the c-myb coding region of pCM107 (Burk et al., 1993) cells twice in hypotonic bu€er (10 mM Tris-HCl, pH 7.8; to the Gal4 DNA binding domain. Most of the expression 5mM KCl; 2 mM MgCl2)at48C followed by incubation for vectors encoding wild-type or mutant v-Myc proteins have 5 min in hypotonic bu€er supplemented with 0.2% NP-40. been described (Mink et al., 1996). Myc ± DN encodes a Myc Nuclei were pelleted by centrifugation, washed twice with protein lacking aminoacids 1 ± 45. The full-length chicken c- hypotonic bu€er and ®nally dissolved by boiling in SDS- Myc expression vector SFCV ± c-Myc was obtained from D sample bu€er. Crouch. To obtain comparable expression levels the chicken c-Myc coding region was excised from this vector and cloned Radiolabeling of cells into pCDNA3. Transfected cells were radiolabeled with 35S-methionine (Amersham, 4800 Ci/mmol) in methionine-free medium at Cell lines 100 mCi/ml for the indicated time period. Before labeling cells QT6 is a line of chemically transformed Japanese quail were pre-incubated in methionine-free medium for 30 min. ®broblasts and was grown in Iscove's modi®ed DMEM Cells were then harvested immediately or were incubated supplemented with 8% (vol/vol) fetal calf serum and 2% further in a medium lacking radioactive methionine and (vol/vol) chicken serum (Moscovici et al., 1977). BM2 is a supplemented with 1 mM of unlabeled methionine. For line of avian myeloblastosis virus transformed chicken immunoprecipitation (all steps on ice) cells were lysed in myeloblasts (Moscovici et al., 1982). BM2(MC29)cl.4 is a lysis bu€er (10 mM Tris-HCl, pH 7.8; 50 mM NaCl; 0.5% derivative of the BM2 line obtained by infecting BM2 cells sodium-deoxycholate; 0.5% NP-40; 0.1% SDS). The lysate MC29 virus (Symonds et al., 1986). BM2 and BM2(MC29) was clari®ed by centrifugation and the appropriate antibodies cells were grown in RPMI 1640 medium supplemented with were added, followed by incubation for 1 ± 2 h. Immunocom- 10% (vol/vol) tryptose phosphate broth, 5% (vol/vol) fetal plexes were collected on staphylococcus aureus protein A calf serum and 5% (vol/vol) chicken serum. Wild-type and conjugated sepharose beads, washed three times with lysis Myc-de®cient Rat1 ®broblasts (Mateyak et al., 1997) were bu€er, once with PBS and bound radioactive proteins were obtained from J Sedivy and were grown in DMEM visualized by SDS ± PAGE and autoradiography. Individual supplemented with 10% fetal calf serum. protein bands were quantitated using a phosphorimage analyzer. Reporter genes, transfections, luciferase and b-galactosidase assays The reporter plasmid pG5E4 ± 38Luc has been described (Kamano et al., 1995). pCMVb was obtained from Clontech. Acknowledgments Transfection of the quail ®broblast cell line QT6 was We thank K Bister, D Crouch, R Davis, D Eick, M Green performed as described (Burk et al., 1993). The amounts of and J Sedivy for plasmids, antibodies and cells. This work DNA used for transfection of cells in a 10 cm tissue culture was supported by a grant from the DFG (KL461/8-1) and dish are indicated in the ®gure legends. Cells were harvested by the Fonds der chemischen Industrie.

References

AmatiB,DaltonS,BrooksMW,LittlewoodTD,EvanGI Bouchard C, Staller P and Eilers M. (1998). Trends Cell. and Land H. (1992). Nature, 359, 423 ± 426. Biol., 8, 202 ± 206. Amati B and Land H. (1994). Curr. Opin. Genet. Dev., 4, Burk O, Mink S, Ringwald M and Klempnauer K-H. (1993). 102 ± 108. EMBO J., 12, 2027 ± 2038. Blackwood EM and Eisenman RN. (1991). Science, 251, Ciechanover A, DiGiuseppe JA, Bercovich B, Orian A, 1211 ± 1217. Richter JD, Schwartz AL and Brodeur GM. (1991). Proc. Boon K, Caron HN, van Asperen R, Valentijn L, Hermus Natl. Acad. Sci. USA, 88, 139 ± 143. MC, van Sluis P, Roobeek I, Weis I, Voute PA, Schwab M Claassen GF and Hann SR. (1999). Oncogene, 18, 2915 ± and Versteeg R. (2001). EMBO J., 20, 1383 ± 1393. 2933.

Oncogene Crosstalk between Myc and ATF2 J Miethe et al 8124 Cole MD and McMahon SB. (1999). Oncogene, 18, 2916 ± Mateyak MK, Obaya AJ, Adachi S and Sedivy JM. (1997). 2924. Cell. Growth Di€er., 10, 1039 ± 1048. Coller HA, Grandori C, Tamayo P, Colbert T, Lander ES, Mink S, Mutschler B, Weiskirchen R, Bister K and Eisenman RN and Golub TR. (2000). Proc. Natl. Acad. Klempnauer K-H. (1996). Proc. Natl. Acad. Sci. USA, Sci. USA, 97, 3260 ± 3265. 93, 6635 ± 6640. Daksis JI, Lu RY, Facchini LM, Marhin WW and Penn LJZ. MoscoviciC,MoscoviciMG,JiminezH,LaiMMC, (1994). Oncogene, 9, 3635 ± 3645. Hayman MJ and Vogt PK. (1977). Cell., 11, 95 ± 103. Dang CV. (1999). Mol. Cell. Biol., 19, 1±11. Moscovici C, Zeller N and Moscovici MG. (1982). Expres- Evan GI and Littlewood TD. (1993). Curr. Opin. Genet. sion of Di€erentiated Function in Cancer Cells. Revoltella Dev., 3, 44 ± 49. RF, Pontieri GM, Basilico C, Rovera G, Gallo RC and Facchini LM and Penn LZ. (1998). FASEB J., 12, 633 ± 651. Subak-Sharpe JH. (eds), Raven Press, NY. pp. 435 ± 449. FiresteinRandFeuersteinN.(1998).J. Biol. Chem., 273, Nakamura T, Okuyama S, Okamoto S, Nakajima T, Sekiya 5892 ± 5902. S and Oda K. (1995). Exp. Cell. Res., 216, 422 ± 430. Fuchs SY, Xie B, Adler V, Fried VA, Davis RJ and Ronai Z. Narayan S, Widen SG, Beard WA and Wilson SH. (1994). J. (1997). J. Biol. Chem., 272, 32163 ± 32168. Biol. Chem., 269, 12755 ± 12763. Fuchs SY and Ronai Z. (1999). Mol. Cell. Biol., 19, 3289 ± Prendergast GC and Zi€ EB. (1991). Science, 251, 186 ± 189. 3298. Read MA, Whitley MZ, Gupta S, Pierce JW, Best J, Davis Fuchs SY, Tappin I and Ronai Z. (2000). J. Biol. Chem., 275, RW and Collins T. (1997). J. Biol. Chem., 272, 2753 ± 12560 ± 12564. 2761. Grandori C and Eisenman RN. (1997). Trends Biochem. RudolphB,Sa€richR,ZwickerJ,HengleinB,MuÈ ller R, Science, 22, 177 ± 181. Ansorge W and Eilers M. (1996). EMBO J., 15, 3065 ± Gupta S, Campbell D, Derijard B and Davis RJ. (1995). 3076. Science, 267, 389 ± 393. Sadowski I and Ptashne M. (1989). Nucleic Acids Res., 17, Hai TW, Liu F, Coukos WJ and Green MR. (1989). Genes 7539. Dev., 3, 2083 ± 2090. Sakamuro D and Prendergast GC. (1999). Oncogene, 18, Henriksson M and LuÈ scher B. (1996). Adv. Cancer Res., 68, 2942 ± 2954. 109 ± 182. Salghetti SE, Kim SY and Tansey WP. (1999). EMBO J., 18, Hoang AT, Cohen KJ, Barrett JF, Bergstrom DA and Dang 717 ± 726. CV. (1994). Proc. Natl. Acad. Sci. USA, 91, 6875 ± 6879. Schmidt EV. (1999). Oncogene, 18, 2988 ± 2996. Jansen-DuÈ rr P, Meichle A, Steiner P, Pagano M, Funke K, Seoane J, Pouponnot C, Staller P, Schader M, Eilers M and Botz J, Wessbecher J, Draetta G and Eilers M. (1993). Massague J. (2001). Nat. Cell. Biol., 3, 400 ± 408. Proc. Natl. Acad. Sci. USA, 90, 3685 ± 3689. ShimizuM,NomuraY,SuzukiH,IchikawaE,TakeuchiA, Kamano H, Burk B, Noben-Trauth K and Klempnauer K-H. Suzuki M, Nakamura T, Nakajima T and Oda K. (1998). (1995). Oncogene, 11, 2575 ± 2582. Exp. Cell. Res., 239, 93 ± 103. Kato GJ, Barrett J, Villa-Garcia M and Dang CV. (1990). Schuhmacher M, Kohlhuber F, Holzel M, Kaiser C, Mol. Cell. Biol., 10, 5914 ± 5920. Burtscher H, Jarsch M, Bornkamm GW, Laux G, Polack Kawasaki H, Schiltz L, Chiu R, Itakura K, Taira K, A, Weidle UH and Eick D. (2001). Nucleic Acids Res., 29, Nakatani Y and Yokoyama KK. (2000). Science, 405, 397 ± 406. 195 ± 200. Staller P, Peukert K, Kiermaier A, Seoane J, Lukas J, Kim SJ, Wagner S, Liu F, O'Reilly MA, Robbins PD and Karsunky H, Moroy T, Bartek J, Massague J, Hanel F and Green MR. (1992). Nature, 358, 331 ± 334. Eilers M. (2001). Nat. Cell. Biol., 3, 392 ± 399. Klempnauer K-H, Ramsay G, Bishop JM, Moscovici MG, Symonds G, Klempnauer K-H, Snyder M, Moscovici G, Moscovici C, McGrath JP and Levinson AD. (1983). Cell, Moscovici C and Bishop JM. (1986). Mol. Cell. Biol., 6, 33, 345 ± 355. 1796 ± 1802. Kretzner L, Blackwood EM and Eisenman RN. (1992). Tsai EY, Jain J, Pesavento PA, Rao A and Goldfeld AE. Nature, 359, 426 ± 429. (1996). Mol. Cell. Biol., 16, 459 ± 467. Li X-Y and Green MR. (1996). Genes Dev., 10, 517 ± 527. van Dam H, Duyndam M, Rottier R, Bosch A, de Vries- Liu F and Green MR. (1990). Cell., 67, 1217 ± 1224. Smits L, Herrlich P, Zantema A, Angel P and van der Eb Liu F and Green MR. (1994). Nature, 368, 520 ± 525. AJ. (1993). EMBO J., 12, 479 ± 487. Livingston C, Patel G and Jones NC. (1995). EMBO J., 14, van Dam H, Wilhelm D, Herr I, Ste€en A, Herrlich P and 1798 ± 1811. Angel P. (1995). EMBO J., 14, 1798 ± 1811. LuÈ scher B and Larsson L-G. (1999). Oncogene, 18, 2955 ± Zwicker J, Lucibello FL, Wolfraim LA, Gross C, Truss M, 2966. Engeland K and MuÈ ller R. (1995). EMBO J., 14, 4514 ± Maekawa T, Sakura H, Kanei-Ishii C, Sudo T, Yoshimora T, 4522. Fijisawa J, Yoshida M and Ishii S. (1989). EMBO J., 8, 2023 ± 2028.

Oncogene