EGR2 Induces Apoptosis in Various Cancer Cell Lines by Direct Transactivation of BNIP3L and BAK

EGR2 induces apoptosis in various cancer cell lines by direct transactivation of BNIP3L and BAK

Motoko Unoki and Yusuke Nakamura*

Laboratory of Molecular Medicine, Human Genome Center, Institute of Medical Science, The University of Tokyo, 4-6-1 Shiorokanedai, Minato-ku, Tokyo, 108-8639, Japan

EGR2 plays a key role in the PTEN-induced apoptotic Nakamura, 2001). Further biological analysis to deter- pathway. Using adenovirus-mediated gene transfer to 39 mine the effects of these eight gene products on cell cancer cell lines, we found that EGR2 could induce growth revealed that the EGR2 protein was likely to be apoptosis in a large proportion of these lines by altering an important mediator of the PTEN growth-suppressive the permeability of mitochondrial membranes, releasing signaling pathway (Unoki and Nakamura, 2001). cytochrome c and activating caspase-3, -8, and -9. Induced expression of EGR2 using a plasmid vector Analysis by cDNA microarray and subsequent functional was able to suppress cell growth significantly in several studies revealed that EGR2 directly transactivates ex- cancer-derived cell lines, while inhibition of endogenous pression of BNIP3L and BAK. Our results helped to EGR2 expression using antisense oligonucleotides clarify the molecular mechanism of the apoptotic pathway accelerated cell growth. induced by PTEN-EGR2, and suggested that EGR2 may The Egr2/Krox-20 gene was originally identified as a be an excellent target molecule for gene therapy to treat a serum response immediate-early gene encoding a protein variety of cancers. with three Cys2–His2-type zinc-finger motifs (Chavrier Oncogene (2003) 22, 2172–2185. doi:10.1038/sj.onc.1206222 et al., 1988). Egr2/Krox-20 (À/À) mice display disrupted hindbrain segmentation and development, and differ- Keywords: EGR2; BNIP3L; BAK; apoptosis; PTEN entiation of Schwann cells is blocked at an early stage (Topilko et al., 1994). In human, defects of this gene cause congenital hypomyelinating neuropathy and type 1 Charcot–Marie–Tooth syndrome (Warner et al., Introduction 1998). Egr2/Krox-20 is also associated with the onset of myelination in the peripheral nervous system (Zorick Mutations of the PTEN gene have been found in human et al., 1996). cancers arising from various organs including uterine No association between EGR2 and PTEN has been endometrium, breast, ovary, prostate, and brain reported to date, except in a study using Egr2/Krox-20 (Li et al., 1997b; Risinger et al., 1997; Steck et al., as a hindbrain marker in Pten-deficient mouse embryos. 1997; Obata et al., 1998). Although most tumor At embryonicstage E8.5 of micewith homozygous suppressors appear to be involved directly in regulation deletions of Pten, the cephalic region was severely of the cell cycle, PTEN is located in cytoplasm overgrown and abnormal Egr2/Krox-20 expression and regulates cell growth through its function as a occurred on one side of the prospective forebrain phosphatase; that is, it modulates the phosphati- (Suzuki et al., 1998). This observation provided indirect dylinositol-3-kinase pathway by catalyzing degradat- evidence that Egr2/Krox-20 expression was regulated by ion of the phosphatidylinositol(3,4,5)triphosphate gen- PTEN in vivo. erated by phosphatidylinositol-3-kinase. Since this We report here that exogenous expression of EGR2 process is involved in various pathways that transfer using AdCAEGR2 can induce apoptosis significantly signals into the nucleus, it is not easy to determine all the in various cancer cell lines by altering the permeability mediators involved in PTEN’s growth-suppressive and of mitochondrial membranes, releasing cytochrome apoptosis-inducing activities. c and activating caspase-3, -8, and -9. We also describe To identify additional mediators of the PTEN path- evidences that EGR2 directly induces expression of way, we previously carried out cDNA microarray two proapoptoticproteins of the Bcl-2family, BNIP3L analyses (Matsushima-Nishiu et al., 2001) and selected and BAK, which are localized at mitochondria eight candidate genes (Ono et al., 2000; Unoki and (Chittenden et al., 1995; Farrow et al., 1995; Kiefer et al., 1995; Matsushima et al., 1998; Imazu et al., 1999; Shimizu et al., 1999). These results bring to light novel mechanisms of apoptosis induced by *Correspondence: Y Nakamura; E-mail: [email protected] PTEN–EGR2, and indicate that EGR2 is potentially Received 31 July 2002; revised 28 October 2002; accepted 6 November a useful molecule for gene therapy to treat a wide range 2002 of cancers. EGR2 induces apoptosis by transactivation of BNIP3L and BAK M Unoki and Y Nakamura 2173 Results among AdCAEGR2-infected EGR2-sensitive cells was significantly greater than that in AdCALacZ-infected Adenoviral-mediated gene transfer and expression of cells, but this was not the case in the resistant lines EGR2 (Figure 2b). Subsequently, activities of caspase-3, -8, and -9 were We constructed an adenovirus designed to express examined by spectrophotometric detection of the EGR2 (AdCAEGR2) and performed immuno- chromophore p-nitroanilide (pNA) after cleavage from blotting with anti-Egr2 antibody to examine expression the labeled substrate–pNA. Caspase-3 and -9 activities of EGR2 in cancer-derived cells. EGR2 protein was elevated two times higher 24 h after the AdCAEGR2 detected in cells infected with AdCAEGR2 as early as infection. On the other hand, caspase-8 activation 6 h after infection, but did not occur in cells infected occurred much later (Figure 2c). Degradation of with AdCALacZ (Figure 1a). No endogenous EGR2 procaspase-3, -8, and -9, and production of active forms protein was detectable at hour zero in the cells of these caspases were confirmed by Western blotting examined. (Figure 2d). To examine growth-suppressive effects of exogenous EGR2 in various cell lines, we first evaluated the efficiency of adenovirus-mediated gene transfer by Analysis by cDNA microarray using AdCAEGR2 X-Gal (5-bromo-4chloro-3-indolyl-b-d-galactopyrano- To investigate further the molecular mechanism of side) staining of cells infected with AdCALacZ at EGR2-mediated cell death, we analysed expression MOI 100. As summarized in Table 1, 24 h after profiles of SW480 cells using cDNA microarrays consis- infection, the proportions of b-gal-positive cells for all ting of 4608 genes. As shown in Table 2, a number of 39 of the cancer-derived cell lines were very high (more genes were significantly upregulated in cells infected with than 50%). AdCAEGR2, a result confirmed by semiquantitative EGR2 Subsequently, we examined the effect of AdCA RT–PCR (data not shown). The upregulated genes infection on cell death at 25, 50, and 100 MOI. Table 1 included mitochondrial proapoptotic proteins of the shows averages of the proportion of viable cells that Bcl-2 family such as BNIP3L/NIX (Imazu et al., 1999), were infected with AdCALacZ or AdCAEGR2 at 100 BAK (Shimizu et al., 1999), BAD (Yang et al., 1995), and MOI. For 33 of the 39 lines (84.6%), exogenous NGFR (NGF receptor TNF receptor superfamily member expression of EGR2 decreased cell viability at signifi- 16), a TNF receptor (Rabizadeh et al., 1993). To verify cantly high levels (indicated by MTT dye conversion) as whether upregulation of these genes was important for compared to cells infected with AdCALacZ (Figure 1b); the apoptotic event in cancer cells, we infected apoptosis was observed at 48–60 h after AdCAEGR2 AdCAEGR2 into four EGR2-sensitive cell lines, infection. The remaining six cell lines (indicated as SW480, TYK-nu, Ishikawa3-H-12 and U87MG, and sensitivities M and R in Table 1) were relatively resistant two EGR2-resistant cell lines, A172 and LNCap.FGC. to EGR2 exogenous expression; cell death was induced Then semiquantitative RT–PCR experiments were in a very small proportion of those cells. Induction of performed. As shown in Figure 3a, exogenous EGR2 cell death by AdCA infection was not related to EGR2 induced expression of BNIP3L and BAK genes the mutational status of PTEN, p53, APC, BAX,or in all four EGR2-sensitive cell lines, but not in the hMSH2 (Table 1). The relative level of endogenous EGR2-resistant lines. Expression of BAD was induced EGR2 expression in each cell line was examined by in two of the four EGR2-sensitive cell lines, and NGFR quantitative RT–PCR using the average level among 18 was induced in all six of the cell lines examined. These normal adult tissues as a baseline (Table 1). Expression results indicated that BNIP3L and BAK are likely to in almost all cancer cell lines was less than that in play crucial roles in the EGR2-induced cell-death normal tissues. We examined geneticalterations of the pathway. EGR2 promoter region in several of the cell lines whose We also examined expression of some related genes, endogenous EGR2 expression was very low, but somatic BNIP3/NIP3, BAX, BID, and TNFa, that were not mutations were not observed in any of the cells spotted on our microarray slide (Figure 3a). Expression examined (data not shown). of BNIP3, an isologue of BNIP3L (Chen et al., 1999), was induced in two EGR2-sensitive cell lines. Expres- Apoptosis induced by EGR2 through activation of sion of BID, which interacts with BAK (Wei et al., caspase-3, -8, and -9 2000), and of BAX, a major transcriptional target of p53 (Miyashita et al., 1994), was not induced by exogenous To analyse the EGR2-mediated death of cancer cells EGR2. Since caspase-8 activity is induced by EGR2, further, we performed flowcytometry using six of the some have suggested that a death ligand/receptor- cancer cell lines, SW480, TYK-nu, HEC-88, U87MG, mediated pathway might be activated (Kruidering and A172, and LNCap.FGC. Induction of EGR2 expression Evan, 2000). We found that infection with AdCAEGR2 significantly increased the numbers of cells at Sub-G1 in could induce expression of only NGFR (Rabizadeh et al., the four lines that were EGR2-sensitive (SW480, TYK- 1993) and TNFa (Baker and Reddy, 1998) among the nu, HEC-88, and U87MG), but not in EGR2-resistant 11 TNF ligands and 16 TNF receptors we examined A172 and LNCap.FGC (Figure 2a). By TUNEL assay by cDNA microarray and RT–PCR experiments we observed that the population of apoptoticcells (Figure 3a). Although the 11 TNF ligands including

Oncogene EGR2 induces apoptosis by transactivation of BNIP3L and BAK M Unoki and Y Nakamura 2174

Figure 1 Signiﬁcant loss of cancer cell viability by exogenous EGR2. (a) Expression of the EGR2 protein in SW480 cells after infection by AdCAEGR2. Cells were harvested at times 0, 6, 12, 24, 48, and 72 h after virus infection and cellular proteins were immunoblotted with anti-Egr2 antibody. (b) 39 cell lines were cultured in six-well plates (1 Â 105 cells/well) and infected with AdCAEGR2 or AdCALacZ at 25, 50, or 100 MOI. Cell numbers and viability were assessed by MTT dye conversion 48–72 h after infection. Each data point represents the mean of values from three separate experiments; Error bar, s.d.

Oncogene Table 1 Characterization of cell lines examined Relative expression Average of cell viability at 100 MOI(%)c level of endogenous Transfection Time of data Cell line Origin Reported mutations EGR2a efﬁciency (%)b AdCALacZ AdCAEGR2 obtained (h)d Sensitivitye SW480 Colon ca P53, APC, POLD 0.034 100 67.20 3.96 72 S LoVo Colon ca. APC, hMSH2, BAX, 0.095 100 76.49 4.64 48 S TGFbRII LS174T Colon ca. BAX, TGFbRII 0.068 100 38.57 4.95 60 S HCT 116 Colon ca. APC, POLD, BAX, 16.341 100 44.58 18.30 60 S TGFbRII HT-29 Colon ca. p53 0.157 100 63.75 63.44 72 R

HEC-59 Endometrial ca. p53, PTEN 0.127 100 44.19 1.39 48 S HEC-116 Endometrial ca. p53, PTEN 0.034 100 57.30 2.81 55 S HEC-6 Endometrial ca. PTEN 0.023 79.5 25.24 3.77 60 S HEC-108 Endometrial ca. PTEN 0.021 100 29.48 4.45 48 S HEC-88 Endometrial ca. PTEN 0.004 100 76.09 5.44 48 S HEC-1 Endometrial ca. p53, PTEN, hPMS2 0.010 100 46.80 8.24 60 S Sawano Endometrial ca. PTEN 0.069 94.8 53.17 8.54 55 S HEC-151 Endometrial ca. PTEN 0.150 100 20.89 9.96 55 S HOUA-1 Endometrial ca. p53, PTEN 0.009 94.8 49.34 10.75 48 S HHUA Endometrial ca. PTEN, MSH3, MSH6 0.031 89.7 38.73 11.50 55 S Ishikawa3-H-12 Endometrial ca. PTEN 0.083 100 75.15 14.51 55 S HEC-50B Endometrial ca. PTEN 2.451 100 104.95 29.59 72 S KLE Endometrial ca. p53 0.025 97.3 77.37 42.27 72 M

MKN74 Gastricca. RUNX3 (methylation) 0.069 100 82.86 6.79 48 S

U373MG Glioblastoma p53, PTEN 0.063 100 58.00 8.63 48 S DBTRG-05MG Glioblastoma PTEN 0.091 100 88.71 9.96 72 S BAK Nakamura and Y BNIP3L and of Unoki transactivation M by apoptosis induces EGR2 U87MG Glioblastoma PTEN 2.118 100 104.36 22.02 72 S A172 Glioblastoma PTEN 0.007 100 88.82 59.79 72 R

SNU475 Hepatoma p53 0.069 96.3 76.34 3.66 72 S SNU423 Hepatoma p53 28.828 100 46.43 3.89 72 S Alexander (PLC/PRF/5) Hepatoma p53, AXIN1 0.017 100 82.58 5.53 48 S Huh7 Hepatoma p53 0.011 67.3 87.74 16.46 48 S HepG2 Hepatoma CTNNB 0.012 87.2 96.53 25.69 48 S

A549 Lung ca. K-Ras 0.004 100 53.13 3.69 60 S H1299 Lung ca. p53 0.013 100 46.06 5.20 60 S LU99A Lung ca. 0.076 50 65.90 19.31 72 S

OV-1063 Ovarian ca. 0.127 70 42.10 1.75 48 S SW626 Ovarian ca. p53 0.037 100 86.14 5.94 48 S MDAH 2774 Ovarian ca. p53 0.037 100 23.65 7.26 55 S TYK-nu Ovarian ca. p53 0.106 93 88.59 18.69 60 S NIH:OVCAR-3 Ovarian ca. p53 0.010 83 60.89 23.85 48 S SK-OV-3 Ovarian ca. p53 0.038 71 95.98 43.65 72 M

PC-3 Prostate ca. p53, PTEN 11.414 100 87.42 36.94 72 M LNCaP.FGC Prostate ca. p53, PTEN, hMSH2 7.585 100 104.81 66.16 72 R aRelative levels of endogenous EGR2 expression were examined by quantitative RT–PCR using the average for 18 normal adult tissues as a control. Data represent means of triplicate measurements bPercentage of b-gal-positive cells 24 h after infection with AdCALacZ at 100 MOI. cCell viability was estimated by MTT assay compared with uninfected controls (100%). Data represent means of triplicate measurements. dMTT assay was performed at the indicated times. eSensitivity to AdCAEGR2 infection. S: EGR2 sensitive (viable cells were o30% at 100 MOI. M: middle (viable cells were >30% o50% at 100 MOI). R: resistant (viable cells were >50% at 100 MOI) ca.=cancer Oncogene 2175 EGR2 induces apoptosis by transactivation of BNIP3L and BAK M Unoki and Y Nakamura 2176

Figure 2 Induction of apoptosis through activation of caspases by EGR2. (a) Effect of EGR2 expression on distribution of cell-cycle phases in EGR2-sensitive cell lines SW480, TYK-nu, HEC-88, and U87MG and two resistant cell lines (A172 and LNCap.FGC) when cells were infected with AdCAEGR2 or AdCALacZ at 100 MOI. All ﬂoating and attached cells were collected and ﬁxed after 48 h (TYK-nu and HEC-88) or 72 h (SW480, U87MG, A172, and LNCap.FGC), and DNA content was determined by PI staining. Sub-G1 populations and phase distributions are indicated as percentages. (b) Detection of induced apoptosis in SW480, TYK-nu, and HEC-88 cells. A172 and LNCap.FGC cells were served as negative controls. DNA fragmentation was detected by TUNEL assay, and nuclei were counterstained with PI. Apoptotic cells appear green or yellow; viable cells show only PI staining. (c) Caspase-3/7, caspase-8, and caspase-9 activity in TYK-nu and SW480 cells after AdCAEGR2 infection. (d) Procaspase-3, -8 and -9 were decreased after AdCAEGR2 infection in TYK-nu cells, on the other hand, a cleaved, active form(s) of these caspases were increased

Fas ligand (TNFSF6) were reported to be induced by ephrin-A1 (EFNA1) and interferon regulatory factor 1 EGR2 in lymphocytes (Mittelstadt and Ashwell, 1999), (IRF1); two tumor suppressors, exostoses 1 (EXT 1) and the Fas ligand was not induced by EGR2 in our inhibin alpha (INHA); transcription factors such as experiment. Expression of NGFR was induced in all the activating transcription factor 3 (ATF3)andgeneral cell lines described above, whereas TNFa was induced transcription factor IIB (GTF2b); as well as various only in EGR2-sensitive cell lines. Two TNFa receptors, signal transducers including hematopoietic cell-speciﬁc TNFRSF1A and 1B (Baker and Reddy, 1998), were Lyn substrate 1 (HCLS1)andregulator of G-protein expressed stably or induced by EGR2 in the sensitive cell signaling 3 (RGS3); Table 2. lines. Induction was conﬁrmed by Western blotting with respect to both dimer and monomer forms of BNIP3L, Release of mitochondrial cytochrome c and alteration of BAK, and both hyper- (inactive) and hypophosphory- membrane permeability by EGR2 lated (active) forms of BAD in SW480 and TYK-nu cells (Figure 3b). Some proapoptotic proteins of the Bcl-2 family, includ- The cDNA microarray analysis also revealed that ing BNIP3L (Imazu et al., 1999), BAK (Shimizu et al., many other downstream genes were upregulated by 1999), and BAD (Yang et al., 1995), target mitochondria EGR2, for example, cell-cycle-associated genes such as directly and induce loss of membrane potential

Oncogene EGR2 induces apoptosis by transactivation of BNIP3L and BAK M Unoki and Y Nakamura 2177 Table 2 Genes upregulated by EGR2 transfer Fold changea Clone UniGene Gene I.D. # (Hs.) symbol Gene name Category Function Reference 6 h 12 h 24 h 48 h A0355 76366 BAD Bcl2-antagonist of cell Apoptosis Induces apoptosis. Yang et al., Cell, 80: 1.2 3.6 1.2 2.0 death 285–291, 1995. A2436 93213 BAK1 Bcl2-antagonist/killer 1 Apoptosis Induces apoptosis. Shimizu et al., 1.5 2.7 2.2 3.4 Nature, 399: 411–412, 1999. A1243 132955 BNIP3L Bcl2/adenovinis E1B Apoptosis Induces apoptosis. Imazu et al., Onco- 0.8 2.4 4.4 4.2 19 kDa-interacting gene, 18: 4523–4529, protein 3-like 1999. A0146 1827 NGFR nerve growth factor re- Apoptosis Induces apoptosis in Rabizadeh et al., 1.9 5.2 2.2 1.7 ceptor (TNFR super- some cases. Science, 261: family member 16) 345–348, 1993.

A1217 80645 IRF1 Interferon regulatory Cell cycle, Cooperates with p53 Tanaka et al., Nature, 1.7 2.8 2.7 1.4 factor 1 transcription in response to DNA 382: 816–818, 1996. factor damage. A0340 198951 JUNB jun B proto-oncogene Cell cycle, Suppression of cell Passegue and wagner, 3.6 2.9 4.9 9.3 transcription proliferation by EMBO J., 19: factor transcriptional acti- 2969–2979, 2000. vation of p16(INK4A) expression. A3153 1624 EFNA1 ephrin-A1, TNFa- Cell cycle Regulated by p53 Dohn et al., onco- 4.3 3.6 1.5 1.7 induced protein and p73. Its receptor gene, 20: 6503–6515, 4(TNFAIP4) (EphA2) inhibits cell 2001. proliferation by inducing apoptosis. A5577 180952 LO- Dynactin p62 subunit Cell cycle Cell division and in- Karki and Holzbaur, 1.0 1.2 2.4 2.5 C51164 tracellular transport. Curr. Opin. Cell Biol., 11: 45–53. A0042 184510 SFN Stratiﬁn (14-3-3 Cell cycle p53-regulated inhibi- Hermeking et al., 1.3 0.8 2.1 4.5 sigma) tor of G2/M pro- Mol. Cell, 1: 3–11, gression. 1997.

A0235 184161 EXT1 Exostoses (multiple) 1 Tumor sup- Heparan sulfate bio- McCormick et al., 0.8 3.9 5.4 4.8 pressor synthesis. Nat. Genet., 19: 158–161, 1998. A3180 1734 INHA Inhibin, alpha Tumor sup- Inhibin a subunit Matzuk and Bradley, 2.1 10.2 0.9 1.0 pressor was considered as a Semin. Cancer Biol., tumor suppressor 5: 37–45, 1994. based on functional studies employing transgenicmouse models.

A0745 460 ATF3 Activating transcrip- Transcription Represses 72 kDa Yan et al., J. Biol. 1.2 4.8 6.9 11.9 tion factor 3 factor type IV metallopro- Chem., 277: teinase (MMP-2) ex- 10804–10812, 2002. pression. A1853 258561 GTF2B General transcription Transcription RNA polymerase II Ha et al., Nature, 1.0 2.3 2.0 2.4 factor IIB factor transcription initia- 352: 689–695, 1991. tion factor. A1508 157449 LHX1 LIM homeobox Transcription Plays essential roles Shawlot and Behnn- 1.0 5.7 2.7 2.6 protein 1 factor in the formation of ger, Nature, 374: head, kidney, and 425–430, 1995. gonads. A0805 797 NFYA Nuclear transcription Transcription Transcription factor Liberati et al., FEBS 2.9 16.0 2.8 2.9 factor Y, alpha factor binding to CCAAT Lett., 433: 174–178, boxes. 1998. A3595 100221 NR1H2 Nuclear receptor Transcription Steroid/hormone nu- Shinar et al., Gene, 1.0 1.5 2.1 4.7 subfamily 1, groupH, factor clear receptor. 147: 273–276, 1994. member 2 A1873 169832 ZNF42 Zinc-ﬁnger protein 42 Transcription Regulates the CD34 Morris et al., Blood, 1.5 4.3 3.1 4.7 (myeloid-speciﬁc reti- factor promoter. 86: 3640–3647, 1995. noic acid-responsive)

A4270 179735 ARHC Ras homolog gene Signal trans- GTP-binding pro- Maekawa et al., 1.6 7.8 3.3 3.1 family, member c duction tein; regulates reor- Science, 285: ganization of the 895–898, 1999. actin cytoskeleton.

Oncogene EGR2 induces apoptosis by transactivation of BNIP3L and BAK M Unoki and Y Nakamura 2178 Table 2 (continued) Fold changea Clone UniGene Gene I.D. # (Hs.) symbol Gene name Category Function Reference 6 h 12 h 24 h 48 h A2842 182215 ARL3 ADP-ribosylation Signal trans- Low molecular Cavenagh et al., J. 1.3 3.8 1.7 2.0 factor-like 3 duction weight GTP-binding Biol. Chem., 269: proteins belonging to 18937–18942, 1994. the RAS superfamily. A2596 2281 CHGB Chromogranin B Signal trans- One of major Ca2+ Yoo et al., J. Biol. 4.0 16.0 2.9 6.7 (secretogranin 1) duction storage proteins of Chem., 277: the secretory gran- 16011–16021, 2002. ules of neuroendo- crine cells, and may control transcription in the nucleus. A0325 174142 CSF1R Colony stimulating Signal trans- Belongs to the sub- Yarden and Ullrich, 2.7 16.0 1.8 4.6 factor 1 receptor duction family of receptor Ann. Rev. Biochem., tyrosine kinases. 57: 443–478, 1988. A3741 89739 CHRNB1 Cholinergic receptor, Signal trans- Acetylcholine recep- Gomez et al., 1996. 1.2 3.8 2.2 1.8 nicotinic, beta polypep- duction tor. Its defect causes Ann. Neurol., 39: tide 1 (muscle) slow-channel conge- 712–723. nital myasthenic syndrome. A5159 173664 ERBB2 v-erb-B2 avian Signal trans- Overexpression has Casalini et al., J. Biol. 1.5 6.5 2.4 3.2 erythroblastic leukemia duction been associated Chem., 276: viral oncogen either with prolifera- 12449–12453, 2001. homolog 2 tion or differentia- tion and apoptosis under p53 regulation. A2464 113207 GPR30 G protein-coupled Signal trans- Required for estro- Filardo et al., Mol 5.3 16.0 1.6 1.7 receptor 30 duction gen-induced activa- Endocrinol., 14: tion of Erk-1 and 1649–1660, 2000. Erk-2. A5784 261828 GPRK7 G protein-coupled Signal trans- May function in Chen et al., Mol. Vis., 1.5 1.1 4.5 3.2 receptor kinase 7 duction cone cells as a cone 7: 305–313, 2001. opsin kinase and provide the normal photopicvision. A2396 86859 GRB7 Growth factor recep- Signal trans- Associates with Han et al., J. Biol. 1.7 1.9 2.4 4.0 tor-bound protein 7 duction FAK. Chem., 275: 28911–28917, 2000. A4601 14601 HCLS 1 Hematopoietic cell- Signal trans- One of PTEN upre- Unoki and Naka- 2.4 16.0 1.7 2.7 specificLyn substrate 1 duction gulated genes. mura, Oncogene, 20: 4457–4465, 2001. A3463 182577 INPP5B Inositol polyphosphate- Signal trans- Catalyzes the con- Ross et al., J. Biol. 5.6 16.0 2.5 1.3 5-phosphatase 75 Kda duction version of IP3 to Chem., 266: IP2. 20283–20289, 1991. A5027 26944 NRGN Neurogranin (protein Signal trans- One of EGR 1 puta- Svaren et al., J. Biol. 1.0 16.0 1.5 2.3 kinase C substrate, duction tive target genes. Chem., 275: RC3) 38524–38531, 2000. A2388 202097 PCOLCE Procollagen C-endo- Signal trans- Specific connective Kessler et al., Bio- 2.6 16.0 4.9 5.7 peptidase enhancer duction tissue glycoprotein chem. Biophys. Res. that is likely to reg- Commun., 173: ulate procollagen 81–86, 1990. processing in vivo. A4613 154437 PDE2A Phosphodiesterase 2A, Signal trans- Hydrolyzes cAMP Rosman et al., Gene, 1.6 7.7 1.6 2.0 cGMP-stimulated duction and cGMP. 191: 89–95, 1997. A1663 32971 PIK3C3 Phosphoinositide-3- Signal trans- May be associated Volinia et al., EMBO 1.0 3.0 2.3 0.9 kinase, class 3 duction with intracellular J., 14: 3339–3348, protein trafficking 1995. complex. A1518 82294 RGS3 Regulator of G-protein Signal trans- Inhibits Gb, g2-in- Shi et al., J. Biol. 1.6 16.0 1.7 2.9 signalling 3 duction duced inositol phos- Chem., 276: phate production, 24293–24300, 2001. mitogen-activated protein kinase activation, and Akt activation. A2221 12956 TIP-1 Tax interaction Signal trans- Interacts with the Reynaud et al., J. 1.0 1.6 22 3.9 protein 1 duction Rho effector rhote- Biol. Chem., 275: kin and is involved 33962–33968, 2000

Oncogene EGR2 induces apoptosis by transactivation of BNIP3L and BAK M Unoki and Y Nakamura 2179 Table 2 (continued) Fold changea Clone UniGene Gene I.D. # (Hs.) symbol Gene name Category Function Reference 6 h 12 h 24 h 48 h in Rho signaling to the serum response element. A6286 8375 TRAF4 TNF receptor- Signal trans- Interacts with the Krajewska et al., Am. 0.9 5.2 2.1 1.7 associated factor 4 duction cytosolic domain of J. Pathol., 152: the LTBR and 1549–1561, 1998. weakly with the NGFR but not with TNFRSF1A, TNFRSF1B, Fas, or CD40. aFold change represents a ratio of signal intensity (Cy3/Cy5)

Figure 3 Induced expression of proapoptotic members of the Bcl-2 family by EGR2. RT–PCR analyses of mRNAs from SW480, TYK-nu, Ishikawa3-H-12, U87MG, A172, and LNCap.FGC cells at the indicated times after infection with AdCALacZ or AdCAEGR2. The integrity of each RNA template was controlled through amplification of GAPDH. (b) BNIP3L, BAK, and BAD proteins were increased by AdCAEGR2 infection in SW480 and TYK-nu cells and release of cytochrome c; therefore, we examined Direct regulation of BNIP3L and BAK expression by mitochondrial changes during EGR2-induced EGR2 apoptosis. Release of cytochrome c from mitochondria to cytoplasm occurred in three EGR2-sensitive cell We then performed reporter and electromobility-shift lines examined, SW480, Ishikawa3-H-12, and TYK-nu assays (EMSA) to investigate whether EGR2 directly (Figure 4a left: see *). On the other hand, no release transactivates BNIP3L and BAK. We found six possible was observed in the EGR2-resistant lines, A172 EGR2-binding sequences (Chavrier et al., 1990; Nardelli and LNCap.FGC. To confirm fractionation, we et al., 1991; Sham et al., 1993) in the promoter region detected mitochondria using mitochondria-specific (À1toÀ349) of the BNIP3L gene and two in the antibody (Figure 4a right). Subsequently, we promoter region (À18 to À338) of the BAK gene. examined alteration of mitochondrial membrane perme- Figure 5a illustrates the construction of reporter ability and found that AdCAEGR2 infection induced plasmids in which various lengths of the promoter loss of membrane potential in all of the three EGR2- regions of BNIP3L (P1–P6) and BAK (P1–P3) were sensitive cell lines but not in the resistant cell lines subcloned into pGL3-basic vector. Figure 5b sum- (Figure 4b). marizes results of the reporter-gene assay, indicating

Oncogene EGR2 induces apoptosis by transactivation of BNIP3L and BAK M Unoki and Y Nakamura 2180

Figure 4 Mitochondrial change by AdCAEGR2.(a) Detection of cytochrome c release by AdCAEGR2. At 48 h (TYK-nu and Ishikawa3-H-12) or 72 h (SW480, A172, and LNCap.FGC) after AdCALacZ or AdCAEGR2 infection, proteins included in cytosolic and mitochondrial fractions were extracted separately and immunoblotted using anticytochrome c antibody (* shows cytocrome c in cytosol). Immunodetection of mitochondria using antimitochondrial antibody was performed to confirm fractionation. (b) Reduction of mitochondrial membrane permeability (Dcm) by AdCAEGR2, as measured by fluorescence of the cationic lipophilic dye CMXRos with a flow cytometer, 48 h (TYK-nu and Ishikawa3-H-12) or 72 h (SW480, A172, and LNCap.FGC) after infection with AdCAEGR2. A reduction in Dcm is reported as ‘Dcm low’

Oncogene EGR2 induces apoptosis by transactivation of BNIP3L and BAK M Unoki and Y Nakamura 2181

Figure 5 Direct regulation of BNIP3L and BAK expressions by EGR2. (a) Construction of reporter-gene plasmids representing the BNIP3L and BAK, subcloned into pGL3-basic vector. (b) Reporter assays of BNIP3L and BAK promoter, and four copies of the wt- or mutant- EGR2-binding site in SW480 cells infected with AdCALacZ or AdCAEGR2 were performed in triplicate. Error bar, s.d. (Scheffe´ ’s F-test). (c) Probes for EMSA assay. (d) EMSA using nuclear extracts of SW480 cells infected with AdCAEGR2 and 32P- labeled DNA probes containing wt putative EGR2-binding elements present in the BNIP3L or BAK promoters. No complex of EGR2 protein with the probe containing candidate-1 in the BNIP3L promoter was detected (Lane 1), but the probe containing candidate-2 in this promoter did form a complex with EGR2. The probe containing the candidate binding site in the BAK promoter, and EGR2 protein, was detected as shifted bands (Lanes 3 and 7) that were supershifted by the addition of anti-Egr2 antibody (Lanes 6 and 10). Excess amounts of unlabeled wt DNAs or mutant DNAs were used for the competition experiments (Lanes 2, 4, 5, 8, and 9) that BNIP3L and BAK are direct transcriptional targets (wt) construct (wt Â 4) in the reporter-gene assay of EGR2 and that two genomicsegments ( À140 to À148 (Figure 5b), supporting the conclusion that these two and À32 to À40) in the BNIP3L promoter region apoptosis-related genes are direct transcriptional targets and a segment (À263 to À271) in the BAK promoter of EGR2. region are possible candidates for EGR2-binding motifs. Our EMSA experiments using a nuclear extract from SW480 cells infected with AdCAEGR2 and 25-bp double-stranded DNA probes corresponding Discussion to the possible EGR2-binding motifs (Figure 5c) revealed that EGR2 was able to bind to the candidate-2 PTEN suppresses cell growth and in some cases induces sequence (À63 to À71), but not to the candidate-1 apoptosis through its function as a phosphatase, but the sequence (À140 to À148) in the BNIP3L promoter molecular mechanism by which PTEN mediates cell- region. It also bound to the candidate sequence (À263 to cycle arrest or apoptosis after its signal(s) are transferred À271) in the BAK promoter region (Figure 5d). The into the nucleus is not well understood. To clarify this construct (mt Â 4) containing a three-nucleotide sub- molecular process, we performed cDNA microarray stitution within the binding motif showed signiﬁcantly analysis using an adenovirus vector designed to express low transactivation activity compared with the wild-type PTEN and identiﬁed two novel candidates, BPOZ and

Oncogene EGR2 induces apoptosis by transactivation of BNIP3L and BAK M Unoki and Y Nakamura 2182

Figure 6 Schematic model for the EGR2-mediated cell-death pathway

EGR2, which were able to mediate the growth- membrane permeability and releasing cytochrome c suppressive signal of the PTEN pathway (Unoki and (Imazu et al., 1999; Shimizu et al., 1999; Sattler et al., Nakamura, 2001). EGR2 was previously reported as a 1997). Released cytochrome c leads to the formation transcriptional factor regulating myelination in the of ‘apoptosomes’ consisting of cytochrome c, Apaf-1, peripheral nervous system (Topilko et al., 1994); defects and procaspase-9 (Li et al., 1997a). Autoactivation of of this gene cause myelinopathies in humans (Warner procaspase-9 subsequently activates effector caspases, et al, 1998), but no role in carcinogenesis was reported for example, caspase-3, and kills cells. In our experi- previously. ments with EGR2, we detected alteration of membrane To examine a possible biological function of EGR2 in permeability and release of cytochrome c from mito- growth arrest or apoptosis, we constructed adenovirus- chondria, as well as activation of caspase-9 and -3. EGR2 and transfected it to various cancer cell lines. Therefore, we consider that EGR2 induces apoptosis Among the 39 lines examined, EGR2 induced apoptosis via mitochondrial change through transactivation of in all but six. Endogenous expression of EGR2 in most BNIP3L and BAK. of the cancer cell lines, examined by real-time PCR, was We also observed that activation of caspase-8 lower than in normal tissues. We excluded the pos- occurred slightly after the activation of caspase-9 and sibilities of geneticsomaticmutation in EGR2 gene by -3. Caspase-8 mediates death signals generated at the sequencing. cell membrane by death receptors such as TNFa To investigate the molecular mechanism of the receptor and Fas (Ashkenazi and Dixit, 1999). Although EGR2-mediated cell-death pathway, we performed Fas ligand was reported to be induced by EGR2 in functional analyses of other genes that were transacti- lymphocytes (Mittelstadt and Ashwell, 1999), it was not vated by infection with AdCAEGR2. These experiments induced in our experiment using cancer cell lines. The showed that two apoptosis-related genes, BNIP3L and difference might be because of the difference in types of BAK, are direct targets of EGR2 and play important cells used in the two studies. Regarding TNFa and its roles in the EGR2-induced apoptotic pathway. BNIP3L two receptors, TNFRSF1A was constitutively expressed and BAK localize in mitochondria and induce apoptosis in all six of the cancer cell lines examined in our by interaction with Bcl-2/Bcl-xL, altering mitochondrial experiment, while TNFa was induced in the four EGR2-

Oncogene EGR2 induces apoptosis by transactivation of BNIP3L and BAK M Unoki and Y Nakamura 2183 sensitive cell lines infected with AdCAEGR2, but not in promoter were constructed from the control cosmid pAxCA- the two resistant lines. TNFRSF1B was constitutively LacZ (TaKaRa). The efficiency of adenovirus-mediated gene expressed in SW480 and was inducible in the remaining transfer was assessed by a b-gal transduction assay described three EGR2-sensitive cell lines, but not in the two previously (Minaguchi et al., 1999). resistant cell lines. However, since the transactivation of TNFa occurred at a relatively late stage, we suspect that Immunodetection of EGR2, BNIP3L, BAK, and BAD proteins the mitochondria-mediated pathway involving activa- Proteins extracted from SW480 cells infected with 50 MOI of tion of BNIP3L and BAK is the main cell-death AdCALacZ or AdCAEGR2 were separated by 12% SDS– pathway induced by EGR2, and that the death ligand/ PAGE and immunoblotted with rabbit polyclonal antibody to receptor pathway is likely to play some additional role. mouse Egr2 (amino acids 8–95; 98% homologous to human NGFR, a TNF receptor having low affinity for NGF, in this region; PRB-236P; BabCO, Richmond, CA, USA). To was also induced by EGR2. Expression of NGFR can detect BNIP3L, BAK, and BAD, cellular proteins from induce neural cell death (Minaguchi et al., 1999), and SW480 and TYK-nu cells were separated by 15% SDS–PAGE the presence of this receptor enhances TNF-induced and immunoblotted with rabbit polyclonal antibody to human BNIP3L (PC525; Oncogene Research Products, Boston, MA, apoptosis although TNFa is unable to bind to NGFR USA), rabbit polyclonal antibody to human BAK (66026E; (MacEwan, 1996). Since NGFR was induced in all of our BD Biosciences Pharmingen, San Diego, CA, USA), or rabbit cell lines, whether EGR2-sensitive or -resistant, its role polyclonal antibody to human BAD (H-168; Santa Cruz in the EGR2-induced death pathway, if any, remains Biotechnology, Santa Cruz, CA, USA). HRP-conjugated anti- unclear. rabbit IgG (Santa Cruz Biotechnology) served as the On the basis of the results described here, we propose secondary antibody for the ECL Detection System (Amersham the model of the EGR2-mediated cell-death pathway Pharmacia, Piscataway, NJ, USA). shown in Figure 6. PTEN induces expression of EGR2, and then EGR2 directly transactivates expression of Growth assay BNIP3L and BAK. BNIP3L and BAK can release Cell growth was assessed by MTT dye conversion as described cytochrome c from mitochondria, alter mitochondrial previously (Matsushima-Nishiu et al., 2001) with minor membrane permeability, activate caspase-9, and then modifications. Cells were seeded at 1 Â 105 cells/well in six-well activate caspase-3. Finally, activated caspase-3 cleaves culture plates and infected with AdCAEGR2 or AdCALacZ various substrates and induces apoptosis; the TNFa- at MOIs of 25, 50, or 100. After 48–72 h, virus-contain- mediated pathway may enhance this mitochondria- ing medium was removed and MTT (5 mg/ml in PBS) was mediated process. As EGR2 induced cell death added. efficiently in the majority of cancer cell lines we examined, we believe that EGR2 should be an excellent Real-time quantitative PCR target molecule for potential gene therapy. cDNAs were analysed quantitatively for expression of EGR2 by means of a reported fluorogenic5 0-nuclease PCR assay (Holland et al., 1991). Specific primers (Forward: 50- 0 0 Materials and methods TCTTTCCCAATGCCG-3 and Reverse: 5 -GGAGATCCAA CGACCTCTTCTCT-30) and probes (50-TTGATCATGC- 0 Cell lines and culture conditions CATCTCCGGCCACT-3 ) were obtained from Applied Biosystems. Gene-specific PCR products were measured All endometrial cancer, ovarian cancer, and hepatoma cell continuously with an ABI PRISM 7700 Sequence Detection lines, as well as colon cancer lines SW480, LoVo, and System (Applied Biosystems, Foster City, CA, USA) during 40 HCT116, were obtained and maintained as described cycles. GAPDH was used for normalization using predeve- previously (Minaguchi et al., 1999; Satoh et al., 2000; loped TaqMan Assay reagents (4310884E, Applied Biosys- Matsushima-Nishiu et al., 2001; Unoki and Nakamura, tems). 2001). A172, A549, H1299, LS174 T, HT-29, DBTRG- 05MG, U373MG, U87MG, LNCap.FGC, and PC-3 were Flow cytometry and TUNEL assay obtained from the American Type Culture Collection (Man- assas, VA, USA). LU99A and MKN74 were obtained from Cells were plated at a density of 5 Â 105 cells/100-mm dish the Japanese Collection of Research Bioresources (Tokyo, and infected with 100 MOI of AdCAEGR2 or AdCALacZ. Japan). All cell lines were grown in monolayers in appropriate After 48 h (TYK-nu and HEC-88) or 72 h (SW480, U87MG, media supplemented with 10% fetal bovine serum: Eagle’s A172, and LNCap.FGC), cells were harvested and treated minimal essential medium for A172, DBTRG-05MG, with Cycle TEST PLUS (Becton Dickinson, San Jose, CA, U87MG, and U373MG; F12 medium for PC-3; Dulbecco’s USA) according to the manufacturer’s protocol. Flow modified Eagle’s medium for LS174T; and RPMI 1640 for cytometry was performed as described previously (Matsush- A549, LU99A, H1299, MKN74, LNCap.FGC, and HT-29. ima-Nishiu et al., 2001). Apoptotic cells were detected by fluorescence microscopy (TUNEL assay) using the ApopTag- Construction of recombinant adenovirus kit (Intergen, Purchase, NY, USA) according to the manufacturer’s protocol. To construct AdCAEGR2 viral vectors, a 1.4-kb fragment of EGR2 was cloned into the Swal site of cosmid pAxCAwt Detection of caspase-3, -8, and -9 activities by means of a (TaKaRa, Tokyo, Japan). Recombinant adenovirus was colorimetric protease assay and Western blotting constructed, propagated, purified, and titered as described previously (Minaguchi et al., 1999). As a control, AdCALacZ A colorimetric protease assay kit (Medical & Biological viruses encoding the b-gal gene under the control of the CAG Laboratories, Nagoya, Japan) was used to detect activities of

Oncogene EGR2 induces apoptosis by transactivation of BNIP3L and BAK M Unoki and Y Nakamura 2184 caspase-3, -8, and -9. The assay is based on the spectro- and LNCap.FGC) were incubated with 0.1 mm MitoTracker photometric detection of the chromophore pNA after cleavage Red CMXRos (M-7512, Molecular Probes, Eugene, OR, from the labeled substrate–pNA. The substrate for caspase-3 USA) for 30 min at 371C. Flow cytometry was performed on a (and -7) was DEVD (Asp–Glu–Val–Asp), that for caspase-8 Becton Dickinson FACSCalibur instrument and analysed by was IETD (IIe–Glu–Thr–Asp), and that for -9 was LEHD ModFit software (Verity Software House). (Leu–Glu–His–Asp). The light emission of pNA was quanti- fied using a microplate reader (BIORAD, Randolph, MA, USA) at 405 nm. The assay was performed according to the Reporter assay manufacturer’s protocol. For Western blotting, we used mouse monoclonal antibody to human caspase-3 (19; Medical & pGL3-basicvector(Promega, Madison, WI, USA) was used Biological Laboratories), mouse monoclonal antibody to for the luciferase expression assay. Various lengths of the human caspase-8 (5F7; Medical & Biological Laboratories), promoter regions of BNIP3L and BAK (Figure 5a) as well as and mouse monoclonal antibody to human caspase-9 (5B4; four copies of the 25-bp fragment corresponding to the wild Medical & Biological Laboratories) for the detection of pro- form or the mutant form of EGR2-binding site were inserted caspase-3, -8, and -9 proteins, respectively. into the vector’s cloning site. The cells were infected with 50 MOI of AdCAEGR2 or AdCALacZ, 3 h before transfection of cDNA microarray analysis the reporter constructs into SW480. A plasmid vector, pRL- TK (Promega), was cotransfected with each reporter construct SW480 cells growing in monolayers were infected with using FuGENEt6 Transfection Regent (Roche) according to AdCAEGR2 or AdCALacZ at 50 MOI. After 6, 12, 24, and the supplier’s recommendations. After 48 h, the reporter assays 48 h, total RNAs were extracted with Trizol reagent (Gibco, were carried out using the Dual-Luciferase Reporter Assay Rockville, MD, USA), treated with DNase I (Roche, System (Promega) according to the manufacturer’s protocol. Indianapolis, IN, USA) and purified twice by means of spun columns containing oligo (dT)-cellulose (mRNA Purification Kit, Amersham Pharmacia). The isolated mRNAs were amplified, labeled, and hybridized on cDNA-spotted glass Electromobility-shift assay (EMSA) slides as described elsewhere (Ono et al., 2000). SW480 cells infected with AdCAEGR2 were harvested 36 h later, and nuclear extracts were prepared. The extracts were Semiquantitative RT–PCR incubated for 30 min at 41C with 200 pg of sonicated salmon sperm DNA, 10 ng BSA, 6 mm HEPES (pH 7.6), 6% glycerol, Cells were infected with 50 MOI of MCAEGR2 or AdCALacZ 3mm NaCl, 0.45 mm MgCl2,0.06mm EDTA, 2.3 mm DTT, and harvested after 12, 24, and 48 h. Total RNAs were isolated 0.03% NP40, and a protease inhibitor cocktail (Roche) and using Trizol reagent (Gibco) and treated with DNase I. Equal with either polyclonal anti-Egr2 antibody (BabCO) or cold amounts of RNAs were reverse-transcribed and amplified competitor double-stranded oligonucleotides (mutant and wt 1 by the PCR for 20–35 cycles with gene-specific primers at 94 C sequences are shown in Figure 5c). The appropriate 32P-labeled 1 1 for denaturing, 55 C for annealing, and 72 C for extension. double-stranded wt oligonucleotides were added before incuba- As an internal control, amplification of GAPDH was carried tion for 30 min at room temperature. Each sample was out by RT–PCR using specific primers (Forward: 50-TGGGT 0 0 electrophoresed in a native 5% PAGE using 0.5 Â TBE buffer. GTGAACCATGAGAAG-3 and Reverse: 5 -GTGTCGCTG The gels were dried and exposed for autoradiography at À801C. TTGAAGTCAGA-30).

Detection of cytochrome c release

Cells infected with 50 MOI of AdCAEGR2 or AdCALacZ for Acknowledgements 48 h (TYK-nu and Ishikawa3-H-12) or 72 h (SW480, A172, We thank Hiroyuki Kuramoto (Kitasato University, Sagami- and LNCap.FGC) were separated into mitochondria and hara, Japan), Masato Nishida (Kasumigaura National Hospi- other fractions using digitonin (Wako, Osaka, Japan). To tal, Tsuchiura, Japan), Toyomi Satoh (University of Tsukuba, conﬁrm fractionation, mouse monoclonal antibody to human Tsukuba, Japan), and Isamu Ishiwata (Ishiwata Obstetrics and mitochondria (MAB1273; CHEMICON, Temecula, CA) was Gynecologie Hospital, Mito, Japan) for their help in obtaining served as positive control. To detect cytochrome c, fractio- endometrial cancer cell lines, and Yoichi Furukawa (The nated proteins (15 mg) were separated by 15% SDS–PAGE and University of Tokyo, Tokyo, Japan), Takashi Shimokawa immunoblotted with rabbit polyclonal antibody to human (The University of Tokyo) and Yumi Nakajima (Oncotherapy, cytochrome c (sc-7159, Santa Cruz, Biotechnology). Tokyo, Japan) for their helpful advice and discussion. This work was supported in part by Research for the Future Mitochondrial membrane potential Program Grant #00L01402 from the Japan Society for the Cells infected with 50 MOI of AdCAEGR2 or AdCALacZ for Promotion of Science (JSPS), and the Research Fellowship of 48 h (TYK-nu and Ishikawa3-H-12) or 72 h (SW480, A172, JSPS for Young Scientists.

References

Ashkenazi A and Dixit VM. (1999). Curr. Opin. Cell Biol., 11, Chavrier P, Zerial M, Lemaire P, Almendral J, Bravo R and 255–260. Charnay P. (1988). EMBO J., 7, 29–35. Baker SJ and Reddy EP. (1998). Oncogene, 17, 3261–3270. Chen G, Cizeau J, Vande Velde C, Park JH, Bozek G, Bolton Chavrier P, Vesque C, Galliot B, Vigneron M, Dolle P, J, Shi L, Dubik D and Greenberg A. (1999). J. Biol. Chem., Duboule D and Charnay P. (1990). EMBO J., 9, 1209–1218. 274, 7–10.

Oncogene EGR2 induces apoptosis by transactivation of BNIP3L and BAK M Unoki and Y Nakamura 2185 Chittenden T, Harrington EA, O’Connor R, Flemington C, Ono K, Tanaka T, Tsunoda T, Kitahara O, Kihara C, Lutz RJ, Evan GI and Guild BC. (1995). Nature, 374, Okamoto A, Ochiai K, Takagi T and Nakamura Y. (2000). 733–736. Cancer Res., 60, 5007–5011. Farrow SN, White JH, Martinou I, Raven T, Pun KT, Rabizadeh S, Oh J, Zhong LT, Yang J, Bitler CM, Butcher LL Grinham CJ, Martinou JC and Brown R. (1995). Nature, and Bredesen DE. (1993). Science, 261, 345–348. 374, 731–733. Risinger JI, Hayes AK, Berchuck A and Barrett JC. (1997). Holland PM, Abramson RD, Watson R and Gelfand DH. Cancer Res., 57, 4736–4738. (1991). Proc. Natl. Acad. Sci. USA, 88, 7276–7280. Satoh S, Daigo Y, Furukawa Y, Kato T, Miwa N, Nishiwaki Imazu T, Shimizu S, Tagami S, Matsushima M, Nakamura Y, T, Kawasoe T, Ishiguro H, Fujita M, Tokino T, Sasaki Y, Miki T, Okuyama A and Tsujimoto Y. (1999). Oncogene, 18, Imaoka S, Murata M, Shimano T, Yamaoka Y and 4523–4529. Nakamura Y. (2000). Nat. Genet., 24, 245–250. Kiefer MC, Brauer MJ, Powers VC, Wu JJ, Umansky SR, Sattler M, Liang H, Nettesheim D, Meadows RP, Harlan JE, Tomei LD and Barr PJ. (1995). Nature, 374, 736–739. Eberstadt M, Yoon HS, Shuker SB, Chang BS, Minn AJ, Kruidering M and Evan GI. (2000). IUBMB Life, 50, 85–90. Thompson CB and Fesik SW. (1997). Science, 275, 983–986. Li P, Nijhawan D, Budihardjo I, Srinivasula SM, Ahmad M, Sham MH, Vesque C, Nonchev S, Marshall H, Frain M, Alnemri ES and Wang X. (1997a). Cell, 91, 479–489. Gupta RD, Whiting J, Wilkinson D, Charnay P and Li J, Yen C, Liaw D, Podsypanina K, Bose S, Wang SI, PucJ, Krumlauf R. (1993). Cell, 72, 183–196. Miliaresis C, Rodgers L, McCombie R, Bigner SH, Shimizu S, Narita M and Tsujimoto Y. (1999). Nature, 399, Giovanella BC, Ittmann M, Tycko B, Hibshoosh H, Wigler 411–412. MH and Parsons R. (1997b). Science, 275, 1943–1947. Steck PA, Pershouse MA, Jasser SA, Yung WK, Lin H, Ligon MacEwan DJ. (1996). FEBS Lett., 379, 77–81. AH, Langford LA, Baumgard ML, Hattier T, Davis T, Frye Matsushima M, Fujiwara T, Takahashi E, Minaguchi T, C, Hu R, Swedlund B, Teng DH and Tavtigian SV. (1997). Eguchi Y, Tsujimoto Y, Suzumori K and Nakamura Y. Nat. Genet., 15, 356–362. (1998). Genes Chromosomes Cancer, 21, 230–235. Suzuki A, de la Pompa JL, StambolicV, Elia AJ, Sasaki T, del Matsushima-Nishiu M, Unoki M, Ono K, Tsunoda T, Barco Barrantes I, Ho A, Wakeham A, Itie A, Khoo W, Minaguchi T, Kuramoto H, Nishida M, Satoh T, Tanaka Fukumoto M and Mak TW. (1998). Curr. Biol., 8, T and Nakamura Y. (2001). Cancer Res., 61, 3741–3749. 1169–1178. Minaguchi T, Mori T, Kanamori Y, Matsushima M, Topilko P, Schneider-Maunoury S, Levi G, Baron-Van Ever- Yoshikawa H, Taketani Y and Nakamura Y. (1999). Cancer cooren A, Chennouﬁ AB, Seitanidou T, Babinet C and Res., 59, 6063–6067. Charnay P. (1994). Nature, 371, 796–799. Mittelstadt PR and Ashwell JD. (1999). J Biol Chem., 274, Unoki M and Nakamura Y. (2001). Oncogene, 20, 4457–4465. 3222–3227. Warner LE, Mancias P, Butler IJ, McDonald CM, Keppen L, Miyashita T, Krajewski S, Krajewska M, Wang HG, Lin HK, Koob KG and Lupski JR. (1998). Nat. Genet., 18, 382–384. Liebermann DA, Hoffman B and Reed JC. (1994). Wei MC, Lindsten T, Mootha VK, Weiler S, Gross A, Ashiya Oncogene, 9, 1799–1805. M, Thompson CB and Korsmeyer SJ. (2000). Genes Dev., Nardelli J, Gibson TJ, Vesque C and Charnay P. (1991). 14, 2060–2071. Nature, 349, 175–178. Yang E, Zha J, Jockel J, Boise LH, Thompson CB and Obata K, Morland SJ, Watson RH, Hitchcock A, Chenevix- Korsmeyer SJ. (1995). Cell, 80, 285–291. Trench G, Thomas EJ and Campbell IG. (1998). Cancer Zorick TS, Syroid DE, Arroyo E, Scherer SS and Lemke G. Res., 58, 2095–2097. (1996). Mol. Cell Neurosci., 8, 129–145.

Oncogene