(2008) 27, 5706–5716 & 2008 Macmillan Publishers Limited All rights reserved 0950-9232/08 $32.00 www.nature.com/onc ORIGINAL ARTICLE DAPK2 is a novel /KLF6 target involved in their proapoptotic function

A Britschgi1, E Trinh2, M Rizzi1, M Jenal1, A Ress1, A Tobler1,3, MF Fey1,4, K Helin2 and MPTschan 1

1Experimental Oncology/Hematology, Department of Clinical Research, University of Bern, Bern, Switzerland; 2Centre for Epigenetics, Biotech Research and Innovation Centre, University of Copenhagen, Copenhagen, Denmark; 3Hematology Department, Inselspital, Bern University Hospital, Bern, Switzerland and 4Medical Oncology Department, Inselspital, Bern University Hospital, Bern, Switzerland

Death-associated 2 (DAPK2) belongs to a Introduction family of proapoptotic Ca2 þ /calmodulin-regulated serine/ threonine . We recently identified DAPK2 as an The death-associated protein kinase 2 (DAPK2 or enhancing factor during granulocytic differentiation. To DRP-1) functions as a positive mediator of programmed identify transcriptional DAPK2 regulators, we cloned cell death when overexpressed in various cellular back- 2.7 kb ofthe 5 0-flanking region ofthe DAPK2 gene. We grounds. Up to now, four other members of the DAPK found that E2F1 and Kru¨ ppel-like factor 6 (KLF6) family have been identified. DAPK1, the closest homo- strongly activate the DAPK2 promoter. We mapped the log of DAPK2, is suggested to act as a tumor suppressor E2F1 and KLF6 responsive elements to a GC-rich region (Bialik and Kimchi, 2006) and is epigenetically silenced 50 ofexon 1 containing several binding sites forKLF6 and in a variety of human cancers (Katzenellenbogen et al., Sp1 but not for . Moreover, we showed that 1999; Esteller et al., 2001; Maruyama et al., 2001; Voso transcriptional activation ofDAPK2 by E2F1 and KLF6 et al., 2004; Christoph et al., 2006). Similar to DAPK1, is dependent on Sp1 using Sp1/KLF6-deficient insect DAPK2 has been shown to participate in tumor- cells, mithramycin A treatment to block Sp1-binding or necrosis factor-a and FAS--induced cell death Sp1 knockdown cells. Chromatin immunoprecipitation (Inbal et al., 2000; Shohat et al., 2002). Despite the revealed recruitment ofSp1 and to lesser extent that of obvious importance of this death-promoting kinase, a E2F1 and KLF6 to the DAPK2 promoter. Activation detailed analysis of the participating regulative mechan- ofE2F1 in osteosarcoma cells led to an increase of isms is still missing. We have recently shown that endogenous DAPK2 paralleled by cell death. Inhibition of DAPK2 acts as an enhancer of neutrophilic maturation DAPK2 expression resulted in significantly reduced cell in myeloid leukemic cells (Rizzi et al., 2007), under- death upon E2F1 activation. Similarly, KLF6 expression pinning its relevance in terminal cell differentiation. in H1299 cells increased DAPK2 levels accompanied by The molecular mechanisms that are responsible for cell death that is markedly decreased upon DAPK2 regulation of the DAPK2 gene, however, remain knockdown. Moreover, E2F1 and KLF6 show cooperation unidentified. in activating the DAPK2 promoter. In summary, our In the present study, we analysed the transcriptional findings establish DAPK2 as a novel Sp1-dependent target regulation of DAPK2. Based on in silico DNA sequence gene for E2F1 and KLF6 in cell death response. analysis, the promoter region upstream of exon 1 Oncogene (2008) 27, 5706–5716; doi:10.1038/onc.2008.179; contains multiple motifs similar to E2F as well as published online 2 June 2008 KLF6 DNA binding sites. The predicted KLF6 binding sites are part of a GC box proximal to exon 1, and this Keywords: DAPK2; E2F1; KLF6; Sp1; transcription; GC-rich sequence also displays several putative Sp1 cell death binding sites. As often seen for promoters bearing regulative GC-rich sequences (Blake et al., 1990), no consensus TATA-box was found in the DAPK2 Correspondence: Dr MPTschan, Experimental Oncology/Hematology promoter region. (MEM E829), Department of Clinical Research, University of Bern, The E2F family of transcription factors are key Murtenstrasse 35, Bern CH-3010, Switzerland. regulators of involved in DNA replication, cell- E-mail: [email protected] cycle control, and oncogenic transformation. Authors contributions: AB performed the experimental research, E2F1, the founding member of the family, plays a interpreted the data and drafted the article. ET performed and analysed E2F1 activation and ChIPexperiments, MR cloned promoter unique role among the E2Fs by regulating cellular and lentiviral constructs, performed reporter assays. MJ performed growth in both a positive and negative manner. Besides and analysed U2OS expression experiments. AR performed immuno- its well-documented role of stimulating cell-cycle pro- precipitations. AT, MFF and KH instigated the initial concept and gression, it has more recently become apparent that experimental design and revised the drafted article. MPT designed the project, analysed data and gave final approval of the submitted paper. E2F1 has the ability to induce apoptosis. E2F1- Received 30 November 2007; revised 25 April 2008; accepted 1 May mediated cell death occurs in presence or absence of 2008; published online 2 June 2008 functional and is regarded as a fail-safe mechanism DAPK2 is regulated by E2F1 and KLF6 A Britschgi et al 5707 in response to oncogenic hyperproliferation (Putzer, receptor fusion protein in which E2F1 was activated 2007). in response to 4-hydroxytamoxifen (OHT) treatment In addition to the E2F-binding motifs, the DAPK2 (Muller et al., 2001). Activation of E2F1 in U2OS-ER- promoter region contains a GC box proximal to exon 1 E2F1 osteosarcoma cells significantly increased DAPK2 with consensus binding sites for Kru¨ ppel-like factor 6 mRNA levels 3.6-fold (Figure 1a). As a control for (KLF6) and specificity protein 1 (Sp1). Both successful E2F1 induction, we measured expression of are members of the same Kru¨ ppel-like/Sp family of zinc cyclin E1, a well-known E2F1 target gene (Geng et al., finger transcription factors. KLFs regulate differen- 1996) in all experiments. Similarly, E2F1 activation in tiation and development through their roles in human diploid fibroblasts (TIG3) and its ectopic , cell proliferation, apoptosis and expression in H1299 NSCLCs led to a 2.6- and 3.3-fold angiogenesis (Bieker, 2001; Black et al., 2001; Suzuki et al., increase of endogenous DAPK2 mRNA levels (data not 2005). KLF6 is ubiquitously expressed and contains a shown). highly conserved COOH terminus zinc finger DNA- Next, to characterize the mechanism of DAPK2 binding domain that interacts with GC box motifs in transcriptional activation by E2F1, we cloned a target promoters, often containing no or only weak 2.7 kb fragment upstream of exon 1 (TSS ¼ 62125975 TATA-boxes (Turner and Crossley, 1999). KLF6 is of ctg ¼ NC_000015 on 15) containing at functionally inactivated in a variety of human cancers least two potential E2F-binding sites into the pGL3 (Chen et al., 2003; Narla et al., 2003; Reeves et al., luciferase reporter vector. To identify the E2F respon- 2004; Difeo et al., 2006; Yin et al., 2007) and has been sive element, deletion constructs were generated com- shown to induce apoptosis in non-small cell lung cancer prising different sizes based on the prediction of E2F (NSCLC) in a p53-independent manner (Ito et al., binding sites (Figure 1b, left panel) and were then 2004). transfected into H1299 cells together with an E2F1 Sp1 accounts together with Sp3 for most of the expression plasmid. The full-length promoter construct GC-box binding activities in mammalian cells (Suske, 1999). (À2659/ þ 43) containing two predicted binding sites It has largely been recognized as a basal factor and the deletion construct bearing only one potential controlling initiation of transcription in TATA-less close to the core promoter region (À329/ promoters (Kollmar et al., 1994). Sp1 is essential for þ 43) were both activated to a similar extent (Figure 1b, regulating a growing number of and cancer- right panel). Surprisingly, a minimal DAPK2 promoter related genes, as for example (Koutsodontis and construct (À50/ þ 43) containing a GC-rich region Kardassis, 2004), (Parisi et al., 2002), DHFR upstream from exon 1, lacking all putative E2F binding (Park et al., 2003) or MYCN (Kramps et al., 2004). It sites, was still highly activated in response to E2F1 often does so by interacting with other transcription expression. In addition, mutation of the full-length factors at the sites of transcriptional initiation. promoter through deletion of this 93 bp GC-rich region Because DAPK2 exhibits proapoptotic functions in produced a construct (À2659/À50) with basic activity cancer cells, we examined the transcriptional regulation but that was no longer responsive to E2F1 co-transfec- of the DAPK2 gene by the potentially tumor-suppressive tion (Figure 1b, right panel). Further analysis of this E2F1, KLF6 and Sp1 transcription factors. Our results 93 bp promoter region revealed a cluster of six and five indicate that, indeed, E2F1 and KLF6 activate DAPK2 putative binding sites for Sp1 and KLF6, respectively transcription in an Sp1-dependent fashion. We further (Figure 1c). To further investigate the precise mechan- show that knocking down DAPK2 significantly dam- ism of this E2F1 and possibly Sp1-mediated activation pens E2F1 and KLF6-induced cell death in a p53- of the DAPK2 promoter, we performed transactivation dependent as well as -independent pathway. Finally, we experiments in Schneider SL2 cells. These cells lack Sp1 found cooperative activation of the DAPK2 promoter and related activities and thus represent a useful system by E2F1 and KLF6 most likely due to functional to study Sp1-mediated transcriptional responses. Over- interactions of these two proteins. expression of E2F1 alone was not sufficient to induce activity of the full-length promoter (À2659/ þ 43) and the deletion mutant (À50/ þ 43). In contrast, coexpres- sion of E2F1 along with Sp1 restored the E2F1 Results responsiveness of the promoter constructs and resulted in synergistic transactivation, which was not seen when DAPK2 is activated by E2F1 by an Sp1/KLF6 binding using the deletion mutant lacking the GC box sites cluster (Figure 1d). To identify candidate nuclear factors binding to the We next activated E2F1 in U2OS-ER-E2F1 cells and DAPK2 promoter region, DNA sequence alignment then blocked Sp1-binding to the DNA by mithramycin was performed to search for known transcription factors A treatment (Miller et al., 1987). Indeed, we found that using MatInspector 7.0. Two putative transcription E2F1-mediated upregulation of DAPK2 mRNA was factor binding sites for E2F transcription factors were dose-dependently reduced upon mithramycin A treat- found. Given the role of E2F1 in apoptosis, we planned ment (from 6.7- to 5.3-fold and 2.8-fold), whereas cyclin to measure the mRNA levels of the proapoptotic E1 control mRNA was not (Figure 1e). Finally, we DAPK2 after E2F1 activation. To this end we used knocked down Sp1 in U2OS-ER-E2F1 cells using RNA U2OS osteosarcoma cells expressing an E2F1–estrogen interference and then activated E2F1. DAPK2 was 5.4-

Oncogene DAPK2 is regulated by E2F1 and KLF6 A Britschgi et al 5708

Figure 1 Sp1-dependent transcriptional induction of DAPK2 by E2F1. (a) DAPK2 quantitative RT–PCR analysis of mRNA isolated from U2OS osteosarcoma cells expressing ER-E2F1. Cells were incubated with 600 nM 4-hydroxytamoxifen (OHT) for 8 h before mRNA extraction. Expression of the E2F1 target cyclin E1 (CCNE1) was measured as positive control for E2F1 activation. Results are shown relative to non-treated controls. *Mann–Whitney U-test, Po0.05. (b) Left panel: schematic presentation of progressively truncated DAPK2 promoter constructs. Numbering indicates the relative position to exon 1. Two putative E2F binding sites are shown in circles. The black box indicates a GC-rich sequence. Right panel: transactivation assays of H1299 cells transiently transfected with empty or E2F1 expression vectors together with DAPK2 promoter deletion mutants as indicated. Twenty-four hours after transfection, cells were assayed for luciferase activity. The promoter activity is shown as relative light units (RLUs). Results are the means±s.d. of triplicate transfections. (c) Nucleotide sequence of the GC-rich box representing the E2F1 responsive element. Putative KLF6 and Sp1 binding sites are indicated. (d) Transactivation assays of Sp1 and KLF6 negative Drosophila Schneider cells transiently transfected with E2F1, Sp1 and DAPK2 promoter deletion constructs as indicated. Analysis as in (a). (e) Blocking Sp1-binding using mithramycin A. U2OS-ER-E2F1 cells were induced with 600 nM OHT for 12 h, then treated with 50 and 500 nM mithramycin A or dimethylsulphoxide (DMSO) as a control for 8 h more before mRNA extraction. CCNE1 mRNA levels were measured as positive control for E2F1 induction. Quantitative RT–PCR results are shown relative to DMSO-treated controls. (f) shRNA-induced Sp1 knockdown in U2OS-ER-E2F1 cells. Left panels: cells were transduced with either nontargeting control (SHC002) or Sp1-targeting shRNA (M1 and M2). Transduced and puro-selected cells were induced with 600 nM 4-hydroxytamoxifen (OHT) for 12 h and then mRNA was extracted. CCNE1 mRNA levels were measured as positive control for E2F1 induction. Quantitative RT–PCR results are shown relative to uninduced SHC002 controls. *Mann–Whitney U-test, Po0.05. Right panel: Sp1 knockdown efficiency was measured using western blot analysis. Actin expression is shown as loading control.

Oncogene DAPK2 is regulated by E2F1 and KLF6 A Britschgi et al 5709 fold upregulated in control SHC002 cells (Figure 1f, left sequences (Li et al., 2005; Figure 3a, lanes 3–6). Only the panel) but not in either of the Sp1 knockdown cells antiserum against Sp1 resulted in a clear supershift of (Mix1 and Mix2, 1.8 and 1.7-fold, respectively). In the the bound probe (Figure 3a, lane 4, depicted with **), control experiment, induction of cyclin E1 mRNA levels although a reduction of the bound probe was seen with was not significantly affected by inhibition of Sp1. the E2F1 antibody and no change in binding was Sp1 knockdown was confirmed by western blotting observed with the KLF6 antibody. These experiments (Figure 1f, right panel). Taken together, our findings indicate that Sp1 is the predominant factor binding to suggest that E2F1 activates DAPK2 through an Sp1/ the GC box. KLF6 binding site cluster within the proximal promoter Chromatin immunoprecipitation (ChIP) assays were region of the kinase and that this activation is dependent carried out to test whether Sp1, E2F1 and KLF6 bind to on Sp1 expression. the DAPK2 promoter in vivo. Nuclear lysates from Sp1, E2F1 and KLF6 co-transfected HEK293T were used for pull-downs with Sp1, E2F1 and KLF6 antibodies. KLF6 induces DAPK2 by Sp1 Precipitated chromatin was used to amplify a 150 bp Based on our results from the initial promoter screening fragment of the proximal DAPK2 promoter containing experiments and data presented above, we decided to the GC box. Sp1 strongly associated with the GC box in investigate the activation of DAPK2 by the Sp-like the DAPK2 promoter, whereas only weak binding of factor, KLF6. We chose H1299 NSCLCs for KLF6 E2F1 and KLF6 was observed (Figure 3b, upper panel). expression studies, because these cells show low As negative control for the ChIPassay, we used primers endogenous KLF6 levels. Further, ectopic KLF6 that amplify a region 1.6 kb downstream of the GC Box expression has been shown to induce apoptosis in these (Figure 3b, lower panel). To conclude, these binding cells (Ito et al., 2004). Transient transfections of KLF6 assays suggest that all the three transcription factors, into H1299 NSCLCs led to a 2.8-fold upregulation of Sp1, E2F1 and KLF6, are present at the DAPK2 endogenous DAPK2 message compared to the empty promoter. vector control (Figure 2a). KLF6 mRNA was quantified to test for successful transfection. We next performed luciferase assays in H1299 cells co-transfecting KLF6 DAPK2 knockdown attenuates E2F1 and KLF6-induced expression plasmid or empty vector together with cell death (À2659/ þ 43) and (À50/ þ 43) DAPK2 promoter con- Both E2F1 and KLF6 have the potential to induce cell structs. Both DAPK2 promoter constructs were acti- death. Thus, we next investigated whether knocking vated by KLF6 expression. A mutated promoter down the proapoptotic DAPK2 kinase by RNA construct lacking the GC box was not induced upon interference would have any consequences on E2F1- or KLF6 expression indicating that the GC box represents KLF6-induced cell death. Induction of E2F1 in U2OS- the KLF6-responsive element (Figure 2b). To test for a ER-E2F1 cells resulted in 20% cell death in the small possible role of Sp1 in KLF6-induced DAPK2 promoter hairpin (sh)RNA targeting Firefly luciferase (shFluc) activity, Sp1/KLF6-negative Schneider SL2 cells were expressing control cells as measured in XTT assays, co-transfected with Sp1 and KLF6. Despite potential whereas cell death in DAPK2 knockdown cell lines was KLF6 binding sites present in the GC box, KLF6 did only 12 and 11%, respectively (Figure 4a). Similarly, a not activate the DAPK2 promoter in the absence of significant reduction of cell death was seen by staining Sp1 (Figure 2c). In line, inhibition of Sp1 binding by apoptotic cells using Annexin V (Figure 4b) or by mithramycin A treatment (Figure 2d) or knocking down measuring caspase 3 and 7 activities (Figure 4c). Sp1 using RNA interference completely abolished Upregulation of DAPK2 through activation of E2F1 KLF6-mediated upregulation of DAPK2 mRNA in and knockdown efficiency of DAPK2 shRNAs were H1299 cells (Figure 2f, left panel). Sp1 knockdown assessed by western blotting (Figure 4d, upper panel) efficiency was measured by western blotting (Figure 2f, and real-time quantitative reverse transcription-PCR right panel). In both experiments, equal transfection (RQ-PCR), respectively (Supplementary Figure 2a). efficiency was assessed by measuring KLF6 mRNA Western blotting of nuclear extracts showed equal (Supplementary Figure 1a and b). In summary, KLF6- efficiency of E2F1 translocation into the nucleus in all induced activation of DAPK2 is strictly dependent on the three cell lines (Figure 4d, lower panel). Sp1 despite putative KLF6 binding sites in the GC box. For analysis of KLF6-induced cell death, we tran- siently transfected H1299 cells with KLF6 or the empty Sp1, E2F1 and KLF6 binding to the DAPK2 promoter vector as control. H1299 SHC002 control cells displayed The À50 to þ 27 region of the DAPK2 promoter was 23.2% cell death after 48 h, whereas knocking down tested for in vitro binding to E2F1, Sp1 and KLF6 using DAPK2 resulted in a significant reduction of cell death band shift assays. Electrophoretic mobility shift assay of 12.1 and 14.8%, respectively (Figure 4e). These (EMSA) of HEK293T nuclear lysates and annealed findings were supported by staining for early apoptosis g-32PGC box oligonucleotides was performed (Figure 4f) and by measuring caspase 7/8 activities (Figure 3a). Supershift assays of the bound oligonucleo- (Figure 4g). Upregulation of DAPK2 and knockdown tide probe were carried out using antibodies against efficiency were measured by western blotting (Figure 4h, KLF6, Sp1, E2F1 and CBP/p300, an acetylase shown to upper panel) and real-time quantitative (RQ)–PCR be associated with KLF6 enhancing its binding to target (Supplementary Figure 2b). KLF6 protein expression

Oncogene DAPK2 is regulated by E2F1 and KLF6 A Britschgi et al 5710

Figure 2 DAPK2 induction by KLF6 is Sp1 dependent. (a) DAPK2 and KLF6 quantitative RT–PCR analysis from mRNA extracted from H1299 transiently transfected with KLF6 or empty vector. Results are shown as n-fold regulation of empty vector transfected cells. *Mann–Whitney U-test, Po0.05. (b) Transactivation assays of H1299 cells transiently transfected with KLF6 expression or empty vector together with DAPK2 promoter deletion mutants as indicated. The promoter activity is shown as relative light units (RLUs). Results are the means±s.d. of triplicate transfections. (c) Drosophila Schneider cells were transiently transfected with KLF6, Sp1 and DAPK2 promoter deletion constructs as indicated. Analysis as in (b). (d) Blocking Sp1 binding using mithramycin A. H1299 cells were transfected with KLF6 or pcDNA expression vectors, 12 h later treated with 50 and 500 nM of mithramycin A or DMSO for 8 h more before mRNA extraction and quantitative RT–PCR analysis. Results are shown relative to DMSO-treated, pcDNA- transfected controls. (e) shRNA-induced Sp1 knockdown in H1299 cells. Left panels: non-targeting control (SHC002) or Sp1-targeting (M1 and M2) shRNA-expressing H1299 cells were transfected with KLF6 or pcDNA expression vectors, 12 h later mRNA was extracted and quantitative RT–PCR was performed. Results are shown relative to the pcDNA-transfected SHC002 cells. *Mann– Whitney U-test, Po0.05. Right panel: Sp1 knockdown efficiency was measured using western blot analysis. Actin expression is shown as loading control.

Oncogene DAPK2 is regulated by E2F1 and KLF6 A Britschgi et al 5711 and E2F1 western blotting of the electrophoresed IP were performed. As low KLF6 and E2F1 protein levels in crude lysates of REH cells could not be detected after stripping of the membrane, we show two different KLF6 IPs, one probed with anti-KLF6 and the other with anti- E2F1 antibody. C-jun has been shown to interact with KLF6 (Slavin et al., 2004) and served as a positive control for the IP. As shown in Figure 5c, immunoblot- ting with E2F1 antibody identified E2F1 in the anti- KLF6 immunoprecipitates demonstrating that KLF6 and E2F1 can be in the same protein complex.

Discussion

Previously, we found that DAPK2 is important for Figure 3 Sp1, E2F1 and KLF6 bind to the DAPK2 promoter. (a) Gel electrophoretic mobility shift assays using g-32P-labeled, myeloid development and that it is suppressed in acute double-stranded, oligonucleotide probes corresponding to the myeloid leukemic blast cells, suggesting that DAPK2 GC-rich box incubated with nuclear extracts from HEK293T cells, displays tumor suppressor activities (Rizzi et al., 2007). bandshifts are depicted with (*). Protein/DNA complexes were The aim of the present study was to identify transcrip- supershifted (**) with antibodies as indicated. (b) Chromatin immunoprecipitation (ChIP) assays using HEK293T cells transi- tional regulators of this putative tumor suppressor. ently transfected with Sp1, E2F1 and KLF6 expression plasmids Computational analysis of the proximal DAPK2 pro- and antisera against Sp1, E2F1 and KLF6. Anti-Pol II and IgG moter spanning 2.7 kb upstream of the first exon served as positive and negative controls, respectively. PCR was revealed possible E2F and KLF6 binding sites. We performed using primers encompassing the GC box in the DAPK2 found activation of E2F1 increased DAPK2 message in promoter and an unrelated sequence 1.6 kb downstream of the GC box as a negative control. various cancer cell types. The amount of DAPK2 upregulation upon E2F1 activation in U2OS-E2F1-ER cells was identical to inductions seen for other known did not differ between the three cell lines (Figure 4h, E2F targets such as Apaf-1 (Moroni et al., 2001) or the lower panel). Taken together, these results show for the BH3-only proteins PUMA/Noxa (Hershko and Gins- first time that DAPK2 is significantly involved in E2F1- berg, 2004). KLF6 most efficiently induced DAPK2 in and KLF6-induced cell death pathways. H1299 NSCLC cells. This might be attributed to the fact that these cells as well as primary NSCLC samples (Ito et al., 2004) show very low base-level expression of E2F1 and KLF6 cooperate in activating DAPK2 KLF6. To examine a potential cooperation between E2F1 and Luciferase reporter assays using DAPK2 promoter KLF6 in modulating DAPK2 promoter activity, deletion mutants showed that a GC box proximal to Schneider SL2 cells were transfected with each of the exon 1 is the E2F1 and KLF6-responsive element. expression vectors encoding E2F1, KLF6 or both, Reporter assays in an Sp/KLF-negative cellular back- together with the GC-box DAPK2 promoter reporter ground, as well as inhibition of DAPK2 activation by (À50/ þ 43) and the GC-box deletion mutant (À2659/ blocking Sp1 binding by mithramycin A treatment or À50). As we previously showed that E2F1 and KLF6 by knocking down Sp1 using RNA interference further activation of the DAPK2 promoter is Sp1-dependent, an revealed that activation of DAPK2 is dependent on Sp1 expression plasmid was co-transfected in each direct binding of Sp1 to the promoter. Together with reporter assay. Luciferase assays revealed that KLF6 our ChIPresults showing only weak binding of both, and E2F1 activated the GC-box DAPK2 promoter E2F1 and KLF6, to the DAPK2 promoter compared to reporter 12- and 20-fold, respectively. Combined the readily detectable Sp1, we suggest that binding of expression of KLF6 and E2F1 led to a 52-fold activation E2F1 and KLF6 to the DAPK2 promoter is indirect. indicating that these two transcription factors cooperate This might at first sight be surprising, because both in activating the DAPK2 promoter (Figure 5a). In line E2F1 and KLF6 could—based on computational with these data, we found that simultaneous induction analysis—bind directly to the promoter. On the other of E2F1 and transfection of KLF6 resulted in a hand, E2F1 is an Sp1 binding partner and its cell-cycle significantly higher upregulation of DAPK2 mRNA in regulated association with Sp1 has been proposed to be U2OS-ER-E2F1 cells than each treatment alone important for fine tuning the transcription of respective (Figure 5b). target genes (Karlseder et al., 1996; Lin et al., 1996). Finally, we examined a possible interaction of Similarly, Sp1 interaction with its close relative KLF6 is endogenous KLF6 and E2F1 using nuclear extracts essential for activation of GC-rich promoters (Kim from human B-cell precursor leukemia REH cells. These et al., 1998; Botella et al., 2002). extracts were subjected to immunoprecipitation (IP) We now found that Sp1 is needed for inducible using anti-KLF6 and anti-c-jun antibodies, then KLF6 DAPK2 expression by E2F1, a mechanism that is not

Oncogene DAPK2 is regulated by E2F1 and KLF6 A Britschgi et al 5712

Figure 4 DAPK2 knockdown attenuates E2F1 and KLF6-induced cell death. (a–c) U2OS-ER-E2F1 cells expressing small hairpin (sh)RNAs targeting DAPK2 (shDAPK2 D5 and D6) or a control shRNA targeting Firefly luciferase (shFluc) were incubated with or without 600 nM 4-hydroxytamoxifen (OHT) for 48 h. Cell death was measured using XTT assays (a), Annexin V staining and flow cytometry (b) or a luminescence-based assay for caspase 3 and 7 activity (c). Results are shown relative to untreated controls. *Mann– Whitney U-test, Po0.05. (d) Western blot analysis of whole cell lysates (depicted with C) showing efficient knockdown of DAPK2 or of nuclear extractions (depicted with N) showing equal translocation of E2F1 into the nucleus in all three cell lines upon OHT treatment. Actin expression is shown as loading control. (e–g) H1299 cells expressing a non-targeting control (SHC002) or a DAPK2- targeting shRNA (shDAPK2 759 and 1735) were transiently transfected with KLF6. Cell death was measured as described above. Results are shown relative to pcDNA-transfected controls. *Mann–Whitney U-test, Po0.05. (h) Western blot analysis of whole cell lysates showing efficient knock-down of DAPK2 and equal amounts of transfected KLF6 in all three cell lines. Actin expression is shown as loading control.

without precedent; for example, E2F1-mediated induc- E2F1 and KLF6 upon co-transfection. In addition, we tion of the ARF tumor suppressor is regulated by found in vivo binding of E2F1 and KLF6, pointing to a several Sp1 and E2F binding sites (Robertson and possible mechanism on how these transcription factors Jones, 1998; Parisi et al., 2002). Regarding KLF6- cooperate in activating DAPK2. We suggest that the mediated, Sp1-dependent activation of DAPK2, our observed transcriptional activation through E2F1 and data are in line with earlier reports showing physical KLF6 is important for full activation of the DAPK2 interaction of SP1 and KLF6/Zf9 in activation of promoter, adding KLF6 as a further player to E2F1- transforming -b and endoglin (Kim Sp1-dependent transcription of a gene regulating cell et al., 1998; Botella et al., 2002). Although KLF6 and death. Sp1 have highly similar DNA recognition sites, both Nevertheless, we would like to refrain from simply may be present at GC-rich promoters, stabilizing each adding KLF6 to other components of the general other, recruiting additional factors and thus building a transcription machinery, although KLF6-mediated, transcriptional network enabling the cell to respond Sp1-dependent activation of DAPK2 might have flexibly to various signals. In favor of this hypothesis, we important biological effects in specific cell types and show that the DAPK2 promoter and endogenous cellular contexts. For example, KLF6 expression is low DAPK2 message were synergistically activated by in prostate cancer cells and its induction allows

Oncogene DAPK2 is regulated by E2F1 and KLF6 A Britschgi et al 5713

Figure 5 Cooperation of E2F1, KLF6 and Sp1 in stimulating DAPK2 promoter activity. (a) Cooperative activation of the DAPK2 reporter by E2F1 and KLF6. Schneider SL2 cells were transfected with the DAPK2 promoter constructs (À50/ þ 43) and (À2659/À50) together with KLF6, E2F1 or a combination thereof. Sp1 was co-transfected in each assay. Luciferase activities were measured as in Figure 2c and results are given as n-fold regulation to that of empty vector transfected cells. (b) Cooperative activation of endogenous DAPK2 mRNA by E2F1 and KLF6. U2OS-ER-E2F1 cells were transiently transfected with KLF6 expression plasmids and incubated with 4-hydroxytamoxifen (OHT) where indicated. DAPK2 quantitative RT–PCR analysis was performed as described in Figure 1. Results are shown as n-fold regulation of non-induced, non-transfected cells. *Mann–Whitney U-test, Po0.05. (c) E2F1 and KLF6 protein-protein interaction. Immunoprecipitation of endogenous KLF6 from B cell precursor leukemia REH cells pulled down endogenous E2F1. Immunoprecipitated c-jun protein, a known KLF6 interaction partner, was used as a positive control. IPproducts and crude cell lysates from two experiments were subjected to immunoblotting using anti-KLF6 and -E2F1 antibodies. 1Membrane first probed with anti-E2F1 antibody, stripped and reprobed with anti-KLF6 antibody, 2membrane first probed with anti-KLF6 antibody, stripped and reprobed with anti-E2F1 antibody. (d) Transcriptional regulation of the proapoptotic DAPK2 kinase. E2F1 and KLF6 are implicated in cell death responses. We now propose that these transcription factors partially induce cell death by activating DAPK2 transcription in an Sp1-dependent fashion. We further suggest that E2F1 and KLF6 cooperate in activating DAPK2 expression and show for the first time physical interaction of E2F1 and KLF6. Both proteins have been shown to bind to Sp1. Please see text for details.

upregulation of the p21CIP1 cell-cycle inhibitor in a p53- or inhibition of NF-kB (Bell and Ryan, 2003). To evaluate independent fashion (Narla et al., 2001). Furthermore, the causative role of DAPK2 in E2F1-induced cell KLF6 also induces cell death in p53-negative NSCLC death, we knocked down DAPK2 in U2OS-ER-E2F1 H1299 cells without upregulating p21CIP1 (Ito et al., osteosarcoma cells using RNA interference. We found 2004). So far, the cell death mechanism in the latter that DAPK2 knockdown cells are significantly protected study remained largely unknown. But we now show that from E2F1-induced cell death. The levels of cell death KLF6-mediated cell death in NSCLC H1299 cells is reduction in the present study (from 20% in the control markedly reduced upon knocking down DAPK2 and cell line to 12 and 11% in the knockdown cell lines, that thus identify this proapoptotic kinase as an important is 40 and 45% of reduction) were similar to the KLF6 target gene through which p53-independent cell reduction levels seen in inhibition experiments of other death can be accomplished. E2F1 target genes, for example 45% for (Irwin E2F1 has been shown to play a dual role in either et al., 2000), 35% for Smac/DIABLO (Xie et al., 2006), promoting proliferation or cell death depending on the 35% for Noxa (Hershko and Ginsberg, 2004), 60% for cellular context, the E2F1 levels and its activating Apaf-1 (Moroni et al., 2001) or 50% for DIP(Stanelle signals (DeGregori and Johnson, 2006). Its ability to et al., 2005). The present study now adds the proapop- induce apoptosis is unique among the E2F family totic kinase DAPK2 to the list of E2F1-targeted cell members (Denchi and Helin, 2005) and is a part of an death genes and shows its critical role in E2F1-induced antitumor safeguard mechanism. There are multiple cell death. roads to E2F1-induced cell death: (a) p53 dependent In summary, based on our results and reports through transactivation of ARF, PUMA and Noxa or published by others we present the following model of DAPK1, or (b) p53 independent through activation of transcriptional DAPK2 regulation (Figure 5d). We p73, Smac/DIABLO, DIP, Apaf-1 and caspase 7 suggest that Sp1 is indispensable for basic DAPK2

Oncogene DAPK2 is regulated by E2F1 and KLF6 A Britschgi et al 5714 promoter activity and that E2F1 and KLF6 are needed 10 cm dishes with 12 mg of expression plasmids (E2F1, for further regulation. We postulate binding of a KLF6 and CBP/p300, empty vector control pcDNA3.1) tripartite Sp1-KLF6-E2F1 complex to the DAPK2 with Lipofectine according the manufacturer’s protocol promoter GC box, but cannot exclude sequential (Invitrogen). binding of Sp1 and KL6 to the GC box. The observed, Sp1-dependent transactivation by KLF6 and Real-time quantitative reverse transcription-PCR E2F1 could be a result of enhanced binding that leads Total RNA was extracted using the RNeasy Mini Kit and the RNase-Free DNase Set according to the manufacturer’s to more efficient initiation of transcription. Finally, protocol (Qiagen, Hombrechtikon, Switzerland). Total RNA emphasizing the role of DAPK2 in different cell was reverse transcribed using random primers (Roche Diag- death pathways, we showed that DAPK2 is a critical nostics) and M-MLV reverse transcriptase (Promega Corpora- downstream target in E2F1- and KLF6-induced tion). PCR and fluorescence detection were performed using cell death. the ABI PRISM 7700 Sequence Detection System (Applied Biosystems, Rotkreuz, Switzerland). For quantification of DAPK2, KLF6 and cyclin E1 mRNA TaqMan Assays Hs00204888_m1, Materials and methods Hs00810569_m1 and Hs01026536_m1 (Applied Biosystems) were used. HMBS primers and probes have been described Cell culture (Britschgi et al., 2006). n-Fold changes were calculated using H1299 NSCLC, HEK293T and REH B-cell precursor the DDCt method of relative quantification. leukemia cell lines were purchased from the American Type Culture Collection (ATCC–LGC Promochem, Molsheim, France). Tamoxifen-inducible U2OS-ER-E2F1 cells were Immunoprecipitation and western blot analysis generated as described elsewhere (Muller et al., 2001). Cells IP, immunoblotting and nuclear protein extraction have been were maintained in RPMI-1640 or Dulbecco’s modified described previously (Andrews and Faller, 1991; Radziwill Eagle’s medium (DMEM) supplemented with 10% fetal calf et al., 2003). Primary antibodies used were anti-DAPK2 (2323; serum (FCS), 50 U/ml penicillin and 50 mg/ml streptomycin. ProSci, Poway, CA, USA), anti-E2F1 (sc-251 X), anti-KLF6 Early passage Schneider (SL2) cells were kindly provided by (sc-7158 X), anti-Sp1 (PEP 2, sc-59 X) and anti-c-Jun (sc-1694; D. Kojic (Institute of Cell Biology, University of Bern) and Santa Cruz Biotechnology, Heidelberg, Germany). Secondary maintained in Drosophila’s Schneider Insect Medium (Sigma- antibodies used were donkey antirabbit and sheep antimouse Aldrich, Buchs, Switzerland) supplemented with 10% FCS, horseradish peroxidase-conjugated immunoglobulin (IgG) 50 U/ml penicillin and 50 mg/ml streptomycin without any gas (Amersham, Zurich, Switzerland) exchange at 25 1C. Electrophoretic mobility shift assay Single-stranded oligonucleotides (GC box of DAPK2 promoter, DAPK2 promoter constructs 0 0 forward: 5 -CCGCCGCGCCCGCCGCGCCCGCCGCGCC A 2.7 kb genomic fragment 5 of exon 1 was PCR-amplified CGCCGCGCCCGCCGCGCCCGCC-30, reverse: 50-GGCG from genomic DNA of human leukocytes using the GC-RICH GGCGCGGCGGGCGCGGCGGGCGCGGCGGGCGCGG PCR System (Roche Diagnostics, Rotkreuz, Switzerland) and CGGGCGCGGCGG-30) were subjected to band shift assays subloned into pGL3-basic Luciferase vector (Promega Cor- as described (Britschgi et al., 2006). Anti-KLF6 (sc-7158 X), poration, Madison, WI, USA) using standard cloning techni- anti-Sp1 (sc-59 X), anti-E2F1 (sc-251 X) and anti-CBP/p300 ques. Primer sequences are available upon request. (sc-584 X) were used for supershifts.

Lentiviral knockdown constructs Chromatin immunoprecipitation Lentiviral vectors expressing shRNAs are given in Supple- For ChIPassays, HEK293T cells were transfected with 2 mg mentary Table 1. Lentivirus production, titer determination Sp1, E2F1 and KLF6 expression plasmids using Lipofectine and transduction were carried out as described (Tschan et al., (Invitrogen). The assays were carried out according to the 2003; Rizzi et al., 2007). ChIP-IT Express protocol (Active Motif, Rixensart, Belgium) except for increased crosslinking time to enhance the chance of Reporter assays and transient transfections detecting indirectly associated proteins. Following DNA H1299 and Schneider SL2 cells were transfected with purification, PCR was performed using Hot Start Polymerase Lipofectamine 2000 according to the manufacturer’s protocol (Fermentas, Nunningen, Switzerland) and the following (Invitrogen, Carlsbad, CA, USA). Cells were transfected with primers: DAPK2 promoter forward: 50-TCTGGGAGGA 200 ng reporter, 400 ng effectors and 10 ng of phRL-TK or GAGGACTGCA-30, reverse: 50-TACTCACGGCGGGAGG 20 ng phRL-CMV (for SL2) expression plasmid for Renilla CTGA-30; negative control primers amplifying 188 bp region luciferase (Promega Corporation) and lysed 24 h after trans- (genomic contig 15q22, 35127150–35127338) forward: 50- fection. Reporter expression was analysed using the Dual- GGTGGCTATCAACAGAAGAA-30, reverse: 50-ACTATAT Luciferase Reporter Assay System (Promega Corporation). In GTTGGCGTTCTGG-30. all transient KLF6 expression experiments, the p300/CBP acetylase was co-transfected, because it has been described to Cell death assays be critical for full KLF6 activity (Li et al., 2005). The Firefly U2OS-ER-E2F1 cells were synchronized for 24 h, followed by luciferase activity of each sample was normalized to its Renilla E2F1 induction with 600 nM OHT for 24 h. H1299 cells grown luciferase activity and the fold activation was obtained by in 10 cm plates were transfected with 10 mg of pCI-neo KLF6 setting the value of empty vector control as 1.0. expression plasmid or an empty vector control using Lipo- A total of 1.5 Â 106 U2OS-ER-E2F1 cells were plated in fectine (Invitrogen). For XTT assays (Biological Industries, 10 cm dishes, starved for 24 h and then induced for 6 h with Kibbutz Beit Haemek, Israel) cells were seeded in triplicate, 600 nM of 4-hydroxytamoxifen. H1299 cells were transfected in incubated as indicated by the manufacturer and absorbance

Oncogene DAPK2 is regulated by E2F1 and KLF6 A Britschgi et al 5715 was measured at 500 nm. For Annexin V staining, 0.5 Â 106 Acknowledgements cells were washed with cold phosphate buffer solution/10% FCS, resuspended in 70 ml of Annexin V binding buffer, and Expert technical assistance by D Shan and G Arvidsson is prepared according to the manufacturer’s protocol (BioVision, appreciated. We are grateful to Drs SL Friedmann, G Thiel Mountain View, CA, USA) and at least 104 cells per sample and JJ Rossi for providing the KLF6, CBP/p300 and Sp1 were analysed by flow cytometry. Caspase 3/7 activation was expression plasmids as well as an U6 promoter constructs. We measured using Caspase-Glo 3/7 Assay (Promega Corporation). also thank Dr D Kojic for providing the Schneider SL2 cells. This work was supported by grants from the Swiss National Statistical analysis Foundation 3100-067213 (to AT and MFF); the Marlies- Each value reported represents the mean±s.d. of at least four Schwegler Foundation and the Bernese Foundation of Cancer measurements of at least two independent experiments. Research (to MFF and AT); the Werner and Hedy Berger- Nonparametric Mann–-Whitney U-tests were applied using Janser Foundation of Cancer Research (to MFF and MPT); the program JMP4 (SAS, Cary, NC, USA). P-values o0.05 the Bern University Research Foundation (to MPT); and the were considered to be statistically significant. Danish Cancer Society (to KH).

References

Andrews NC, Faller DV. (1991). A rapid micropreparation technique Irwin M, Marin MC, Phillips AC, Seelan RS, Smith DI, Liu W et al. for extraction of DNA-binding proteins from limiting numbers of (2000). Role for the p53 homologue p73 in E2F-1-induced mammalian-cells. Nucleic Acids Res 19: 2499. apoptosis. Nature 407: 645–648. Bell LA, Ryan KM. (2003). Life and death decisions by E2F-1. Cell Ito G, Uchiyama M, Kondo M, Mori S, Usami N, Maeda O et al. Death Differ 11: 137–142. (2004). Kruppel-Like factor 6 is frequently down-regulated and Bialik S, Kimchi A. (2006). The death-associated kinases: structure, induces apoptosis in non-small cell lung cancer cells. Cancer Res 64: function, and beyond. Ann Rev Biochem 75: 189–210. 3838–3843. Bieker JJ. (2001). Kruppel-like factors: three fingers in many pies. Karlseder J, Rotheneder H, Wintersberger E. (1996). Interaction of J Biol Chem 276: 34355–34358. Sp1 with the growth- and cell cycle-regulated Black AR, Black JD, Azizkhan-Clifford J. (2001). Sp1 and Kruppel- E2F. Mol Cell Biol 16: 1659–1667. like factor family of transcription factors in cell growth regulation Katzenellenbogen RA, Baylin SB, Herman JG. (1999). Hypermethyla- and cancer. J Cell Physiol 188: 143–160. tion of the DAP-Kinase CpG island is a common alteration in B-cell Blake MC, Jambou RC, Swick AG, Kahn JW, Azizkhan JC. (1990). malignancies. Blood 93: 4347–4353. Transcriptional initiation is controlled by upstream Gc-box Inter- Kim Y, Ratziu V, Choi SG, Lalazar A, Theiss G, Dang Q et al. (1998). actions in a tataa-less promoter. Mol Cell Biol 10: 6632–6641. Transcriptional activation of transforming growth factor beta 1 and Botella LM, Sanchez-Elsner T, Sanz-Rodriguez F, Kojima S, Shimada its receptors by the Kruppel-like factor Zf9/core promoter-binding J, Guerrero-Esteo M et al. (2002). Transcriptional activation of protein and Sp1—potential mechanisms for autocrine fibrogenesis in endoglin and transforming growth factor-beta signaling components response to injury. J Biol Chem 273: 33750–33758. by cooperative interaction between Sp1 and KLF6: their potential Kollmar R, Sukow KA, Sponagle SK, Farnham PJ. (1994). Start site role in the response to vascular injury. Blood 100: 4001–4010. selection at the tata-less carbamoyl-phosphate synthase (glutamine- Britschgi C, Rizzi M, Grob TJ, Tschan MP, Hugli B, Reddy VA et al. hydrolyzing) aspartate carbamoyltransferase dihydroorotase pro- (2006). Identification of the p53 family-responsive element in the moter. J Biol Chem 269: 2252–2257. promoter region of the hypermethylated in Koutsodontis G, Kardassis D. (2004). Inhibition of p53-mediated cancer 1. Oncogene 25: 2030–2039. transcriptional responses by mithramycin A. Oncogene 23: Chen CS, Hyytinen ER, Sun XD, Helin HJ, Koivisto PA, Frierson HF 9190–9200. et al. (2003). Deletion, mutation, and loss of expression of KLF6 in Kramps C, Strieder V, Sapetschnig A, Suske G, Lutz W. (2004). E2F human prostate cancer. Am J Pathol 162: 1349–1354. and Sp1/Sp3 synergize but are not sufficient to activate the MYCN Christoph F, Kempkensteffen C, Weikert S, Kollermann J, Krause H, gene in neuroblastomas. J Biol Chem 279: 5110–5117. Miller K et al. (2006). Methylation of tumour suppressor genes Li D, Yea S, Dolios G, Martignetti JA, Narla G, Wang R et al. (2005). APAF-1 and DAPK-1 and in vitro effects of demethylating agents in Regulation of Kruppel-like factor 6 tumor suppressor activity by bladder and kidney cancer. Br J Cancer 95: 1701–1707. acetylation. Cancer Res 65: 9216–9225. DeGregori J, Johnson DG. (2006). Distinct and overlapping roles for Lin SY, Black AR, Kostic D, Pajovic S, Hoover CN, Azizkhan JC. E2F family members in transcription, proliferation and apoptosis. (1996). Cell cycle-regulated association of E2F1 and sp1 is related to Curr Mol Med 6: 739–748. their functional interaction. Mol Cell Biol 16: 1668–1675. Denchi EL, Helin K. (2005). E2F1 is crucial for E2F-dependent Maruyama R, Toyooka S, Toyooka KO, Harada K, Virmani AK, apoptosis. EMBO Rep 6: 661–667. Zochbauer-Muller S et al. (2001). Aberrant promoter methylation Difeo A, Narla G, Hirshfeld J, Camacho-Vanegas O, Narla J, Rose SL profile of bladder cancer and its relationship to clinicopathological et al. (2006). Roles of KLF6 and KLF6-SV1 in ovarian cancer features. Cancer Res 61: 8659–8663. progression and intraperitoneal dissemination. Clin Cancer Res 12: Miller DM, Polansky DA, Thomas SD, Ray R, Campbell VW, 3730–3739. Sanchez J et al. (1987). Mithramycin selectively inhibits Esteller M, Corn PG, Baylin SB, Herman JG. (2001). A gene transcription of G-C containing DNA. Am J Med Sci 294: hypermethylation profile of human cancer. Cancer Res 61: 388–394. 3225–3229. Moroni MC, Hickman ES, Lazzerini Denchi E, Caprara G, Colli E, Geng Y, Eaton EN, Picon M, Roberts JM, Lundberg AS, Gifford A Cecconi F et al. (2001). Apaf-1 is a transcriptional target for E2F et al. (1996). Regulation of transcription by E2Fs and and p53. Nat Cell Biol 3: 552–558. . Oncogene 12: 1173–1180. Muller H, Bracken AP, Vernell R, Moroni MC, Christians F, Grassilli Hershko T, Ginsberg D. (2004). Up-regulation of Bcl-2 homology 3 E et al. (2001). E2Fs regulate the expression of genes involved in (BH3)-only proteins by E2F1 mediates apoptosis. J Biol Chem 279: differentiation, development, proliferation, and apoptosis. Genes 8627–8634. Dev 15: 267–285. Inbal B, Shani G, Cohen O, Kissil JL, Kimchi A. (2000). Death- Narla G, Friedman SL, Martignetti JA. (2003). Kruppel cripples associated protein kinase-related protein 1, a novel serine/threonine prostate cancer—KLF6 progress and prospects. Am J Pathol 162: kinase involved in apoptosis. Mol Cell Biol 20: 1044–1054. 1047–1052.

Oncogene DAPK2 is regulated by E2F1 and KLF6 A Britschgi et al 5716 Narla G, Heath KE, Reeves HL, Li D, Giono LE, Kimmelman AC Slavin DA, Koritschoner NP, Prieto CC, Lopez-Diaz FJ, Chatton B, et al. (2001). KLF6, a candidate tumor suppressor gene mutated in Bocco JL. (2004). A new role for the Kruppel-like transcription prostate cancer. Science 294: 2563–2566. factor KLF6 as an inhibitor of c-Jun proto-oncoprotein function. Parisi T, Pollice A, Di Cristofano A, Calabro V, La Mantia G. (2002). Oncogene 23: 8196–8205. Transcriptional regulation of the human tumor suppressor Stanelle J, Tu-Rapp H, Putzer BM. (2005). A novel mitochondrial p14(ARF) by E2F1, , , and Sp1-like factors. Biochem protein DIPmediates E2F1-induced apoptosis independently of Biophys Res Commun 291: 1138–1145. p53. Cell Death Differ 12: 347–357. Park KK, Rue SW, Lee IS, Kim HC, Lee IK, Ahn JD et al. (2003). Suske G. (1999). The Sp-family of transcription factors. Gene 238: Modulation of Sp1-dependent transcription by a cis-acting E2F 291–300. element in dhfr promoter. Biochem Biophys Res Commun 306: 239–243. Suzuki T, Aizawa K, Matsumura T, Nagai R. (2005). Vascular Putzer BM. (2007). E2F1 death pathways as targets for cancer therapy. implications of the Kruppel-like family of transcription factors. J Cell Mol Med 11: 239–251. Arterioscler Thromb Vasc Biol 25: 1135–1141. Radziwill G, Erdmann RA, Margelisch U, Moelling K. (2003). The Tschan MP, Fischer KM, Fung VS, Pirnia F, Borner MM, Fey MF Bcr kinase downregulates ras signaling by phosphorylating AF-6 et al. (2003). Alternative splicing of the human -binding and binding to its PDZ domain. Mol Cell Biol 23: 4663–4672. Myb-like protein (hDMP1) yields a truncated protein isoform Reeves HL, Narla G, Ogunbiyi O, Haq AI, Katz A, Benzeno S et al. that alters macrophage differentiation patterns. J Biol Chem 278: (2004). Kruppel-Like factor 6 (KLF6) is a tumor-suppressor gene 42750–42760. frequently inactivated in colorectal cancer. Gastroenterology 126: Turner J, Crossley M. (1999). Mammalian Kruppel-like transcription 1090–1103. factors: more than just a pretty finger. Trends Biochem Sci 24: Rizzi M, Tschan MP, Britschgi C, Britschgi A, Hugli B, Grob TJ et al. 236–241. (2007). The death-associated protein kinase 2 is up-regulated during Voso MT, Scardocci A, Guidi F, Zini G, Di Mario A, Pagano L et al. normal myeloid differentiation and enhances neutrophil maturation (2004). Aberrant methylation of DAP-kinase in therapy-related in myeloid leukemic cells. J Leukoc Biol 81: 1599–1608. acute myeloid leukemia and myelodysplastic syndromes. Blood 103: Robertson KD, Jones PA. (1998). The human ARF cell cycle 698–700. regulatory gene promoter is a CpG island which can be silenced Xie W, Jiang P, Miao L, Zhao Y, Zhimin Z, Qing L et al. (2006). by DNA methylation and down-regulated by wild-type p53. Mol Novel link between E2F1 and Smac/DIABLO: proapoptotic Smac/ Cell Biol 18: 6457–6473. DIABLO is transcriptionally upregulated by E2F1. Nucleic Acids Shohat G, Shani G, Eisenstein M, Kimchi A. (2002). The DAP-kinase Res 34: 2046–2055. family of proteins: study of a novel group of calcium-regulated Yin D, Komatsu N, Miller CW, Chumakov AM, Marschesky A, death-promoting kinases. Biochim Biophys Acta (BBA)—Prot McKenna R et al. (2007). KLF6: Mutational analysis and effect on Proteom 1600: 45–50. cancer cell proliferation. Int J Oncol 30: 65–72.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Oncogene