Leukemia (2005) 19, 806–813 & 2005 Nature Publishing Group All rights reserved 0887-6924/05 $30.00 www.nature.com/leu Cloning and functional characterization of MEF2D/DAZAP1 and DAZAP1/MEF2D fusion created by a variant t(1;19)(q23;p13.3) in acute lymphoblastic leukemia

V Prima1, L Gore2,3, A Caires2, T Boomer2, M Yoshinari4, M Imaizumi5, M Varella-Garcia3,6 and SP Hunger1

1Department of Pediatrics, University of Florida College of Medicine and the University of Florida Shands Cancer Center, Gainesville, FL; 2Department of Pediatrics, Denver, CO, USA; 3University of Colorado Cancer Center, Denver, CO, USA; 4Pediatric Oncology, Tohoku University Hospital, Sendai, Miyagi, Japan; 5Department of Hematology and Oncology, Miyagi Children’s Hospital, Sendai, Miyagi, Japan; and 6Department of Medicine, Denver, CO, USA

We analyzed the TS-2 acute lymphoblastic leukemia (ALL) cell present in the regulatory regions of many active in cells of line that contains a t(1;19)(q23;p13.3) but lacks E2A-PBX1 muscle or neural lineage. MEF2D dimers regulate expression of fusion typically present in leukemias with this translocation. We genes involved in muscle-specific and/or growth factor-related found that the t(1;19) in TS-2 fuses the 19p13 DAZAP1 13–15 (Deleted in Azoospermia-Associated 1) to the 1q23 gene transcription. Murine retroviral insertional mutagenesis MEF2D (Myocyte Enhancer Factor 2D), leading to expression of studies have identified MEF2D as a candidate oncogene reciprocal in-frame DAZAP1/MEF2D and MEF2D/DAZAP1 tran- involved in the pathogenesis of leukemia.16,17 DAZAP1 is an scripts. MEF2D is a member of the MEF2 family of DNA binding RNA-binding protein that may be essential for spermatogen- proteins that activate transcription of genes involved in control esis.18 It is expressed primarily in testis, and to a lower level, in of muscle cell differentiation, and signaling pathways that mediate response to mitogenic signals and survival of neurons thymus. DAZAP1 shows dynamic subcellular distribution during and T-lymphocytes. DAZAP1 is a novel RNA binding protein spermatocyte development, consistent with its presumed role in 18,19 expressed most abundantly in the testis. We demonstrate that mRNA transport or processing. In this report, we show that MEF2D/DAZAP1 binds avidly and specifically to DNA in a in-frame MEF2D/DAZAP1 and DAZAP1/MEF2D fusion tran- manner indistinguishable from that of native MEF2D and is a scripts are expressed in TS-2 and define the DNA-, RNA- substantially more potent transcriptional activator than MEF2D. binding, and transcriptional regulatory properties of the resultant We also show that DAZAP1/MEF2D is a sequence-specific RNA- binding protein. MEF2D has been identified as a candidate chimeric proteins. Our studies identify MEF2D and DAZAP1 oncogene in murine retroviral insertional mutagenesis studies. fusion proteins as components of novel pathways that contribute Our data implicate MEF2D in human cancer and suggest that to human leukemogenesis. MEF2D/DAZAP1 and/or DAZAP1/MEF2D contribute to leukemo- genesis by altering signaling pathways normally regulated by wild-type MEF2D and DAZAP1. Leukemia (2005) 19, 806–813. doi:10.1038/sj.leu.2403684 Materials and methods Published online 3 March 2005 Keywords: leukemogenesis; oncogene; chromosomal translocation ALL cell lines

ALL cell lines were grown in RPMI-1640 and CV-1 cells were grown in DMEM supplemented with 10% fetal calf serum at Introduction 371C in a humidified atmosphere containing 5% CO2. Cell lines utilized included TS-2, RCH-ACV (kindly provided by Dr Ram translocations play a critical role in leukemogen- Seshadri) that has a t(1;19)(q23;p13) and E2A-PBX1 fusion, and 12,20 esis.1 The 19p13.3 gene E2A (also termed TCF3) is the target of the B-precursor ALL cell line REH. several recurrent genomic rearrangements in acute lymphoblas- tic leukemia (ALL).2 The t(1;19)(q23;p13.3) is the second most common translocation in ALL.3,4 In 90–95% of patients with a Fluoresence in situ hybridization (FISH) t(1;19), the molecular consequence is creation of an E2A-PBX1 fusion gene on the der(19) that encodes chimeric E2A-PBX1 Single and multiplex FISH was performed as described proteins with transforming properties.5–10 The remaining 5–10% previously using cosmids provided by Anne Olsen and the of patients have translocations that appear identical by conven- human chromosome 19 mapping group at Lawrence Livermore tional cytogenetics, but do not involve E2A or PBX1.11 Yoshinari National Labs.21,22 et al12 described the TS-2 cell line that was established from a 3- year-old girl with ALL and contains a t(1;19)(q23;p13.3), but lacks E2A or PBX1 abnormalities. Molecular analyses We found that the 1;19 translocation in TS-2 fuses the 1q23 gene MEF2D (Myocyte Enhancer Factor 2D) to DAZAP1 Nucleic acids were extracted from cell lines and lambda phage (Deleted in Azoospermia-Associated Protein 1) located at using commercial reagents (QIAGEN, Valencia, CA, USA; chromosome 19p13.3. MEF2D is a member of MEF2 family of Invitrogen Life Technologies, Carlsbad, CA, USA; Promega, transcription factors that bind specifically to the MEF2 element Madison, WI, USA). Southern blot, Western blot and PCR analyses were performed essentially as described previously.23–27 Correspondence: Dr SP Hunger, Pediatric Hematology-Oncology, A TS-2 genomic library was constructed using the Lambda University of Florida College of Medicine, PO Box 100296, Gaines- DASH II kit (Stratagene, San Diego, CA, USA). Approximately ville, FL 32610-0296, USA; Fax: þ 1 352 392 8725; E-mail: [email protected]fl.edu one million recombinant phage were screened with a DAZAP1 Received 22 October 2004; accepted 28 December 2004; Published cDNA probe kindly provided by Dr Pauline Yen (Torrance, CA, online 3 March 2005 USA); positive clones were purified to homogeneity and the MEF2D fusion proteins in ALL V Prima et al 807 phage inserts were characterized by nucleotide sequence values obtained with the empty expression vector and are the analysis as described previously.26 averages of at least three separate experiments with duplicate transfections normalized to b-galactosidase values. cDNA cloning, expression and reporter constructs In vitro RNA-binding assay First-strand cDNA synthesis was performed on total RNA using a mix of oligo(dT) primers and random hexamers with SuperScript Binding of IVT 35S-labeled proteins to RNA homopolymers First-Strand Synthesis System for RT-PCR (Invitrogen). To clone immobilized on agarose bead (Sigma) was carried out in a total full-length cDNAs, first-strand cDNA was used as PCR template of 0.5 ml of binding buffer (10 mM Tris-HCl, pH 7.4; 2.5 mM with the primer sets DAZAP1 (f/r), MEF2D (f/r) (oligonucleotide MgCl2; 0.5% Triton X-100; 0.1 M NaCl) for 10 min with rotation sequences are listed in Supplementary Table 1), and Expand at 41C. Equal amounts of 35S-labeled proteins were incubated High Fidelity PCR System (Roche Molecular Systems, Alameda, with equal volume of agarose-linked RNA homopolymers as CA, USA). The reverse primers added FLAG tags (underlined) to described.18 To study binding-specificity, 50-fold molar excess 0 the 3 -ends of cDNAs. PCR products were cloned into pCR2.1- of soluble RNA homopolymers were included as competitors for TOPO and then subcloned into pBlue-TOPO (T7 promoter) and protein binding. The beads were briefly pelleted in a micro- pRc/RSV vectors (Invitrogen). Reporter plasmid p4XMEF2-FLuc centrifuge and washed five times with binding buffer prior to containing four MEF2 sites upstream of the c-fos minimal resuspension in 70 ml of SDS-PAGE loading buffer. Bound promoter and luciferase gene was a kind gift from Dr Ron protein was eluted from the nucleic acid by boiling 5 min in Prywes (Columbia University). The plasmid pCMVbgal contains gel loading buffer, resolved on 12% SDS-PAA gel, and the b-galactosidase gene under control of a CMV promoter. visualized by autoradiography.

In vitro translation and immunoblot analysis Results In vitro transcriptions and translations (IVT) were performed using the Promega TNT Quick Coupled Transcription/Transla- Localization of the chromosome 19 genomic breakpoint tion System (1 mg/50ml reaction). Following separation on 12% in TS-2 SDS-PAA gels, protein expression was determined by Western blot analysis using the M2 anti-FLAG primary antibody (Sigma, To localize the TS-2 translocation breakpoint with respect to St Louis, MO, USA) and goat-anti-mouse Ig-HRP secondary E2A, we performed metaphase FISH with TS-2 and RCH-ACV antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA). The cell lines using cosmids containing genomic sequences from chemiluminescent signal was detected using the ECL Plus 19p13 (Supplementary Figure 1). These studies first established detection kit (Amersham, Arlington Heights, IL, USA). that the TS-2 chromosome 19 breakpoint is located within the 1500 kilobases (kb) of DNA telomeric to E2A (Supplementary Table 2). This is shown in Figure 1a and b using the two-color 22 Electrophoretic mobility shift assay (EMSA) split signal E2A FISH assay that we described previously. Additional FISH studies narrowed the interval of interest to an EMSA was performed at 201C with IVT proteins. The double- approximately 400 kb region between E2A and POL2RE stranded MEF2 consensus binding oligonucleotide, 50- (Supplementary Table 2). Detailed FISH studies using a panel GATCGCTCTAAAAATAACCCTGTCG-30 (Santa Cruz Biotech- of 12 cosmids that spanned the region between E2A and nology), was end-labeled with g-32P-ATP using T4 polynucleo- POL2RE revealed that the breakpoint occurred in the region tide kinase (Promega). The DNA binding reaction mixtures in contained in cosmid 21277, which was split in TS-2 and 20 ml contained 1 ng of 32P-labeled probe, 1 mg/ml of poly(dI-dC) hybridized to both the der(1) and the der(19) (Supplementary (Sigma), 10 mM Tris-HCl, pH 7.5; 1 mM DTT; 1 mM EDTA; Table 2 and Figure 1c). To better visualize the chromosome 19 50 mM NaCl; 5% glycerol and 5 ml IVT reaction. For competition breakpoint in TS-2, we developed a more robust two-color analyses, 50-fold molar excess of unlabeled consensus or split signal FISH assay. A pool of four cosmids located mutant (50-GATCGCTGTAAACATAACCCTGTCG-30) MEF2 immediately centromeric to cosmid 21277 and telomeric to double-stranded oligonucleotide was added to the binding cosmid 27377 serves as one probe, while the other probe is assay. Samples were separated in a 0.5 Â TBE 5% PAA gel at composed of a pool of four cosmids located immediately 20 mA constant current, 41C, and gels were dried and telomeric to 21277. Split signals in metaphase (Figure 1d) and autoradiographed. interphase (data not shown) FISH confirmed localization of the TS-2 chromosome 19 breakpoint to the region homologous to cosmid 21277. Cell transfections and reporter gene assays

CV-1 cells were cultured in six-well plates and transfected using DAZAP1 is fused to MEF2D by the t(1;19) in TS-2 Lipofectamine 2000 (Invitrogen). In total, 2 mg of the test expression plasmids (empty pRc/RSV, pRc/RSV-DAZAP1-FLAG; A Genbank search for open reading frames within cosmid pRc/RSV-MEF2D-FLAG; pRc/RSV-DAZAP1/MEF2D-FLAG; pRc/ 21277 identified exons of DAZAP1, which encodes a protein RSV-MEF2D/DAZAP1-FLAG) was cotransfected with 1.5 mgof with novel RNA-binding properties that is expressed most p4xMEF2Fluc and 0.5 mg of pCMVbgal. Luciferase and b- abundantly in the testis and had been mapped previously to galactosidase expressions were determined using commercial 19p13.3.18 Southern blots containing TS-2 and control DNA reagents (Promega). Luciferase values are expressed as were probed with a DAZAP1 cDNA probe. Rearranged bands mean7s.e.m. of fold induction seen compared to luciferase were seen only in TS-2, establishing that the t(1;19) interrupts

Leukemia MEF2D fusion proteins in ALL V Prima et al 808

Figure 1 Localization of TS-2 chromosome 19 breakpoint via FISH. (a) Two-color E2A split signal FISH was performed using a pool of four cosmids located immediately centromeric (green) and telomeric (red) to E2A. FISH with RCH-ACV shows a breakpoint within E2A with a fused signal on the normal 19 and split signals on the der(19) (green) and der(1) (red). (b) Two-color E2A split signal FISH with TS-2 shows fused signals on the normal and der(19) homologs, indicating that the chromosome 19 breakpoint occurs telomeric to both probes. (c) Cosmid 21277 hybridizes to the normal 19, der(1) and der(19) in TS-2 metaphases. (d) Two-color split signal FISH using pools of cosmids located immediately centromeric (green) and telomeric (red) to cosmid 21277 was performed with TS-2 metaphases. One fused red/green and split red and green signals confirm that the chromosome 19 breakpoint occurs within cosmid 21277. For the schematic depiction of FISH experiments see Supplementary Figure 1.

the 19p13.3 gene DAZAP1 in TS-2, with the breakpoint region Wild type and chimeric DAZAP1 and MEF2D in approximately the middle of the gene (Figure 2a). transcripts are expressed in TS-2 We constructed a TS-2 genomic library, screened it with a DAZAP1 cDNA probe, and identified a lambda phage clone that To study expression of DAZAP1, MEF2D and the predicted contained a rearranged DNA fragment of approximately 13 kb. fusion transcripts, we performed RT-PCR on TS-2 and REH, a Nucleotide sequence analysis showed that this lambda clone control B-precursor ALL cell line. Both TS-2 and REH expressed contained a portion of DAZAP1 fused to a region of chromo- DAZAP1 and MEF2D transcripts, but MEF2D/DAZAP1 and some 1 that contained MEF2D, which encodes one of the four DAZAP1/MEF2D were expressed only in TS-2 (Figure 2c). members of the MEF2 family (Figure 2b).15 We then amplified Sequence analysis of PCR products showed that MEF2D/ TS-2 DNA with breakpoint-flanking primers and sequenced the DAZAP1 and DAZAP1/MEF2D fusion transcripts were in-frame genomic fusion sites from both the der(1) (Genbank accession (Supplementary Figure 3). AY681493) and der(19) (accession AY681494) (Supplementary Figure 2). The genomic breakpoints occur in introns of MEF2D and DAZAP1. There is a five (bp) insertion at the site of Structure and synthesis of wild type and chimeric genomic fusion on the der(19) that is not derived from either DAZAP1 and MEF2D proteins germline chromosome 1 or 19. Homologous breakpoints occur on the der(1) chromosome with a deletion of 97 bp and a 21 bp Native DAZAP1 (NP_061832) includes two identified RNA GC-rich insertion. recognition motifs (RRM) specified by amino acids 1–87 and

Leukemia MEF2D fusion proteins in ALL V Prima et al 809

Figure 2 The t(1;19) in TS-2 fuses DAZAP1 to MEF2D. (a) Southern blots of TS-2 and control samples were hybridized with a labeled DAZAP1 cDNA probe. Rearranged bands present in TS-2 establish that the t(1;19) in TS-2 interrupts DAZAP1. TS-2, RCH-ACV, REH: ALL cell lines; NL1, NL2: normal DNA. Arrows marked G or R indicate germline or rearranged DNA fragments. (b) Schematic depiction of the cloned lambda fragment containing TS-2 der(19) genomic DNA. Arrows denote primers used for sequencing (Supplementary Table 1); black boxes represent DAZAP1 exons. Nucleotide sequences of the fusion sites of chimeric chromosomes are shown in Supplementary Figure 2. (c) RT-PCR amplification of DAZAP1, MEF2D, DAZAP1/MEF2D and MEF2D/DAZAP1 transcripts shows that wild-type DAZAP1 and MEF2D transcripts are expressed in both TS-2 and REH, while only TS-2 expresses the fusion DAZAP1/MEF2D and MEF2D/DAZAP1 transcripts. For nucleotide sequences of the fusion transcripts see Supplementary Figure 3.

105–190 (Figure 3a). MEF2D (NP_005911) has conserved N- This may reflect the lesser fidelity of translation in vitro allowing terminal MADS-box (MCM1, Agamous, Deficiens and Serum alternative downstream start codons. The smaller DAZAP1/ response factor28) and MEF2 domains involved in protein MEF2D IVT fragments have the C-terminal FLAG tag but not the dimerization, DNA binding and protein–protein interactions, complete N-terminal RNA-binding domains (see Figure 3e and f). and two transcriptional activation domains (TAD), one which is poorly localized carboxy-terminal to the MEF2 domain, and one within the C-terminus.29 DAZAP1/MEF2D fusion cDNAs In vitro RNA binding of DAZAP1, MEF2D and their (accession AY678451) are predicted to encode a chimeric derivatives protein that contains all of the first DAZAP1 RRM and a truncated portion of the second RRM joined to the carboxy- DAZAP1 displays salt-dependent RNA-binding properties; Tsui terminal portion of MEF2D that includes the second TAD. reported that IVT DAZAP1 could bind to RNA homopolymers Reciprocal MEF2D/DAZAP1 fusion transcripts (accession immobilized on agarose beads, with preferential binding to AY675556) are predicted to encode a chimera that includes poly(U) and poly(G), less avid binding to poly(A), and no the MEF2D MADS-box, MEF2 domain, and the first TAD joined binding to poly(C).18.We performed analogous experiments and to the carboxy terminus of DAZAP1 including a truncated found that IVT DAZAP1 bound strongly to poly(U) and poly(G) portion of RRM 2. at 0.1 M NaCl, whereas DAZAP1/MEF2D bound to the same We assembled full-length wild type and fusion MEF2D and homopolymers to a lesser degree (Figure 3e). Neither protein DAZAP1 cDNAs with added C-terminal FLAG epitope tags for bound to poly(C) agarose (data not shown). The decreased IVT and expression in mammalian cells. The proteins were binding of DAZAP1/MEF2D as compared to native DAZAP1 detected following polyacrylamide gel electrophoresis (PAGE) may be due to the fact that DAZAP1/MEF2D fusion proteins either by anti-FLAG immunoblot or by autoradiography (Figure expressed in TS-2 include an intact first RRM but only 3b–d). For IVT DAZAP1/MEF2D (Figure 3c and d), only the approximately one-half of the second RRM (Figure 3a). Un- highest band on PAGE corresponds to the predicted full-size expectedly, we found that both MEF2D and MEF2D/DAZAP1 product observed following transient transfections (Figure 3b). bound strongly to poly(G), but not to poly(U) or poly(C)

Leukemia MEF2D fusion proteins in ALL V Prima et al 810

Figure 3 Structural features, expression and RNA-binding properties of wild type and chimeric MEF2D and DAZAP1 proteins. (a) Predicted functional domains of DAZAP1, MEF2D, DAZAP1/MEF2D and MEF2D/DAZAP1 proteins. Arrows indicate predicted protein breakpoints. (RRM – RNA recognition motif; MADS – DNA binding and protein dimerization domain; MEF2 – cofactor interactions domain; TAD – transcriptional activation domain). (b) Transient transfection of expression constructs with proteins detected via anti-FLAG Western blot. (c) IVT of proteins with detection via anti-FLAG Western blot. (d) IVT with 35S-labeled proteins detected via autoradiography. Lane 1: DAZAP1; lane 2: MEF2D; lane 3: DAZAP1/MEF2D; lane 4: MEF2D/DAZAP1. (e) In vitro binding to RNA homopolymers immobilized on agarose beads. IVT 35S-labeled proteins were incubated with agarose beads containing bound RNA homopolymers. Beads were collected via centrifugation and washed to remove nonspecific binding. Bound proteins were then visualized via gel electrophoresis and autoradiography. (f) To determine specificity of RNA-binding, experiments were performed in the presence of 50-fold molar excess of the soluble RNA homopolymer competitors. The first lane for each protein represents 1/3 of the IVT amount used in binding experiments.

(Figure 3e and data not shown). To determine if the observed MEF2/DAZAP1 retains DNA-binding properties of wild- RNA-binding was specific; we performed RNA binding experi- type MEF2D ments in the presence of soluble RNA homopolymers to act as competitors to the agarose-linked RNA homopolymers. The We compared DNA-binding properties of wild type and presence of 50-fold molar excess of soluble poly(U) but not chimeric MEF2D and DAZAP1 IVT proteins via EMSA using a poly(C) abolished both DAZAP1 and DAZAP1/MEF2D binding radiolabeled MEF2 consensus binding site oligonucleotide. Both to poly(U)-agarose, indicating that this binding was specific IVT MEF2D and MEF2D/DAZAP1 proteins bound avidly to the (Figure 3f). In contrast, MEF2D RNA binding was nonspecific, consensus MEF2 site, while there was no binding by DAZAP1 or because both soluble poly(G) and poly(U) failed to affect DAZAP1/MEF2D (Figure 4). The observed binding was specific, MEF2D binding to poly(G)-agarose beads (Figure 3f). We because preincubation of MEF2D and MEF2D/DAZAP1 with observed a similar pattern of nonspecific RNA binding by 50-fold molar excess of the unlabeled consensus oligonucleo- MEF2D/DAZAP1 and the hepatic leukemia factor protein24,30 tide markedly reduced binding to the labeled oligonucleotide, (data not shown), consistent with weak RNA-binding properties while there was no effect of pre-incubation with 50-fold molar of DNA-binding motifs found in many transcription factors.31 excess of an unlabeled mutant oligonucleotide.

Leukemia MEF2D fusion proteins in ALL V Prima et al 811

Figure 5 MEF2D/DAZAP1 is a more potent transcriptional activator than MEF2D. Expression constructs and reporter genes were transiently transfected into CV-1 cells. The relative luciferase levels were normalized to the levels of b-galactosidase expression. The average fold inductions7s.e. of the means of three separate experiments are indicated.

Figure 4 DNA binding properties of MEF2D, DAZAP1 and chimeric proteins. EMSA was performed with the indicated IVT terminal differentiation of all muscle lineages during Drosophila unlabeled proteins and a 32P-labeled MEF2 consensus oligonucleo- embryogenesis.36–38 MEF2 proteins also have critical functions tide. Where indicated 50-fold molar excesses of either unlabeled consensus MEF2 oligonucleotide or unlabeled mutant oligonucleotide in other cell types. They are transcriptional effectors of were included as competitors. mitogenic signaling pathways initiated by mitogen-activated protein kinases (MAPKs) including p38 and ERK5 (extracellular signal-related kinase 5)/Big MAPK-1, and also play critical roles in calcium-regulated signaling pathways that control survival of MEF2D/DAZAP1 is a more potent transcriptional neurons and T-cells.15,33,39 The effects of MEF2 proteins on activator than wild-type MEF2D survival can be either pro- or antiapoptotic.39–41 While MEF2 proteins have not previously been implicated in human cancer, To characterize the transcriptional regulatory properties of MEF2D can induce expression of c-jun, a known transforming MEF2D/DAZAP1, we transiently transfected CV-1 cells with oncogene, and has recently been identified in murine retroviral expression constructs and p4xMEF2FLuc, which has four con- mutagenesis studies as a candidate oncogene involved in the sensus MEF2 binding sites located 50 to a luciferase reporter gene. pathogenesis of lymphoid malignancies.16,17,42 MEF2D/DAZAP1 was a substantially more potent transcriptional In contrast to MEF2D, relatively little is known concerning activator than wild-type MEF2D (Figure 5). DAZAP1 or DAZAP1/ DAZAP1, a protein originally isolated by virtue of its interaction MEF2D did not activate transcription of p4xMEF2FLuc. with DAZ in a yeast two-hybrid assay.18 DAZAP1 is an RNA- binding protein expressed primarily in testis that contains two conserved RRMs and a proline-rich C-terminal domain.18,19 Discussion There are dynamic changes in temporal and spatial expression of DAZAP1 during spermatogenesis in the mouse, with We report that the variant t(1;19)(q23;p13) present in TS-2 fuses suggestions that expression may be associated with active MEF2D to DAZAP1, show that TS-2 expresses MEF2D/DAZAP1 transcription.19 DAZAP1 mRNA is also expressed in thymus,18 and DAZAP1/MEF2D chimeric transcripts, and characterize the and we show in this report that the transcript is detected via RT- in vitro RNA-, DNA-binding, and transcriptional regulatory PCR in human B-precursor ALL cell lines. properties of the encoded fusions proteins. Since TS-2 expresses both MEF2D/DAZAP1 and DAZAP1/ There are four vertebrate (MEF2A-D) and one Drosophila MEF2D fusion transcripts, there are several potential mechan- (D-MEF2) MEF2 genes that encode proteins originally characterized isms by which the chimeras might contribute to leukemogen- as muscle-specific transcription factors that bind as homodimers esis. We demonstrate that MEF2D/DAZAP1 is a substantially and heterodimers to an A/T-rich DNA sequence present in target more potent transcriptional activator than wild-type MEF2D in gene promoters and regulate transcription of genes critical for transient transfection assays. MEF2D/DAZAP1 might directly myogenic differentiation (reviewed by McKinsey).15 MEF2 activate transcription of genes critical for lymphocyte growth proteins have highly similar amino-terminal regions that contain and/or survival such as interleukin-2, a known transcriptional conserved elements including the MADS box with an adjacent target of MEF2D in T-cells.43 Alternatively, MEF2D/DAZAP1 MEF2 domain involved in dimerization and DNA binding, and could contribute to leukemogenesis via dysregulated activation more divergent carboxy-terminal domains. Members of this of MAPK-mediated cell proliferation pathways, analogous to protein family interact with a variety of other proteins involved constitutive activation of a growth factor receptor. Since MEF2 in transcriptional regulation including histone deacetylases proteins are subject to modification through complex interac- (HDAC), p300 and Cabin1/Cain.32–35 MEF2 proteins are tions with various transcriptional co-activators and co-repres- essential for cardiac morphogenesis in mice (mef2c) and sors, it is also possible that replacement of the C-terminus of

Leukemia MEF2D fusion proteins in ALL V Prima et al 812 MEF2D with that of DAZAP1 critically alters regulatory Loewenstern Family Foundation and Monfort Family Foundation influences on MEF2D function. to SPH, NCI Cancer Center Core Grant CA 46934, and by the It is also important to consider the possibility that DAZAP1/ grants from the Leukemia Research Foundation and the Cancer MEF2D is critical for leukemogenesis. DAZAP1/MEF2D proteins League of Colorado to LG. expressed in TS-2 retain all of the first and approximately one-half of the second RRM of DAZAP1, fused to a truncated C-terminus of MEF2D. Studies of other RNA-binding proteins show that each Supplementary Information RRM can form a globular domain that, in at least some cases, can bind to RNA independently.44 This is consistent with our data Supplementary Information accompanies the paper on the demonstrating that DAZAP1/MEF2D retains sequence-specific Leukemia website (http://www.nature.com/leu). RNA-binding properties. A number of other genes that encode proteins with RNA binding domains are involved in transloca- References tions in human cancer including EWS, TLS/CHOP, RBM15, and RSI2.45–48 As is the case in TS-2, each of these RNA-binding 1 Rabbitts TH. Chromosomal translocations in human cancer. proteins is fused to a transcription factor with DNA-binding Nature 1994; 372: 143–149. properties, but there is no clear understanding whether, and if so 2 Hunger SP. Chromosomal translocations involving the E2A gene in how, their RNA binding properties contribute to transformation. acute lymphoblastic leukemia: clinical features and molecular We have performed FISH with chromosome 19 cosmids on pathogenesis. Blood 1996; 87: 1211–1224. several primary hyperdiploid childhood ALL cases with a 3 Carroll AJ, Crist WM, Parmley RT, Roper M, Cooper MD, Finley t(1;19)(q23;p13) but lacking E2A-PBX1 fusion, and did not find WH. Pre-B cell leukemia associated with chromosome transloca- tion 1;19. Blood 1984; 63: 721–724. evidence that DAZAP1 was interrupted (AC and SPH, unpub- 4 Michael PM, Levin MD, Garson OM. Translocation 1;19 – a new lished observations). However, samples were limited and of cytogenetic abnormality in acute lymphocytic leukemia. Cancer poor quality making these conclusions tentative. A Genet Cytogenet 1984; 12: 333–341. t(1;19)(q23;p13) has also been reported occasionally in AML, 5 Mellentin JD, Murre C, Donlon TA, McCaw PS, Smith SD, Carroll T-cell ALL, and in some lymphomas.4,49–51 These cases should AJ et al. The gene for enhancer binding proteins E12/E47 lies at be analyzed to determine the potential involvement of DAZAP1 the t(1;19) breakpoint in acute leukemias. Science 1989; 246: 379–382. and/or MEF2D, and their possible role(s) in the oncogenic 6 Nourse J, Mellentin JD, Galili N, Wilkinson J, Stanbridge E, Smith properties of such malignancies. SD et al. Chromosomal translocation t(1;19) results in synthesis of While this manuscript was being prepared for submission, a homeobox fusion mRNA that codes for a potential chimeric Yuki et al52 reported cloning of identical MEF2D/DAZAP1 and transcription factor. Cell 1990; 60: 535–545. DAZAP1/MEF2D fusion transcripts from TS-2. Their indepen- 7 Kamps MP, Murre C, Sun XH, Baltimore D. A new homeobox gene dent studies showed that transiently transfected MEF2D, contributes the DNA binding domain of the t(1;19) translocation protein in pre-B ALL. Cell 1990; 60: 547–555. DAZAP1, MEF2D/DAZAP1 and DAZAP1/MEF2D are localized 8 Hunger SP, Galili N, Carroll AJ, Crist WM, Link MP, Cleary ML. 52 to the nucleus in HeLa cells. MEF2D-DAZAP1 was co- The t(1;19)(q23;p13) results in consistent fusion of E2A and PBX1 immunoprecipitated with wild-type MEF2D from HEK293 cells, coding sequences in acute lymphoblastic leukemias. Blood 1991; suggesting that the wild type and chimeric MEF2D proteins 77: 687–693. could form heterodimers and/or associate with one another in a 9 Kamps MP, Baltimore D. E2A-Pbx1, the t(1;19) translocation higher order protein complex in vivo.52 MEF2D/DAZAP1 co- protein of human pre-B-cell acute lymphocytic leukemia, causes acute myeloid leukemia in mice. Mol Cell Biol 1993; 13: 351–357. localized with HDAC4 in HEK293 cells and could be co- 10 Dedera DA, Waller EK, LeBrun DP, Sen-Majumdar A, Stevens ME, 52 immunoprecipitated. Their data also showed that wild-type Barsh GS et al. Chimeric homeobox gene E2A-PBX1 induces MEF2D, MEF2D/DAZAP1 and DAZAP1/MEF2D, but not wild- proliferation, apoptosis, and malignant lymphomas in transgenic type DAZAP1, could promote soft agar colony formation by mice [published erratum appears in Cell 1993 Nov 19;75(4):826]. HeLa cells.52 When these data are taken together with our data Cell 1993; 74: 833–843. summarized above and the murine retroviral insertional 11 Privitera E, Kamps MP, Hayashi Y, Inaba T, Shapiro LH, Raimondi 16,17 SC et al. Different molecular consequences of the 1;19 chromo- mutagenesis studies, it suggests that native MEF2D has somal translocation in childhood B-cell precursor acute lympho- latent transforming properties that can be unmasked via aberrant blastic leukemia. Blood 1992; 79: 1781–1788. protein expression. MEF2D-DAZAP1 has similar DNA binding 12 Yoshinari M, Imaizumi M, Eguchi M, Ogasawara M, Saito T, properties to MEF2D, but is a more potent transcriptional Suzuki H et al. Establishment of a novel cell line (TS-2) of pre-B activator and also may associate more strongly with other acute lymphoblastic leukemia with a t(1;19) not involving the E2A proteins involved in transcriptional regulation (eg HDAC4). gene. Cancer Genet Cytogenet 1998; 101: 95–102. 13 Breitbart RE, Liang CS, Smoot LB, Laheru DA, Mahdavi V, Nadal- These alterations may confer more potent transforming proper- Ginard B. A fourth human MEF2 transcription factor, hMEF2D, is ties to MEF2D/DAZAP1, which can be further augmented by co- an early marker of the myogenic lineage. Development 1993; 118: expression with the reciprocal DAZAP1/MEF2D chimera. In 1095–1106. future studies, it will be critical to directly test the transforming 14 Martin JF, Miano JM, Hustad CM, Copeland NG, Jenkins NA, properties of native and chimeric MEF2D and DAZAP1 proteins Olson EN. A Mef2 gene that generates a muscle-specific in lymphoid cells, and determine the mechanisms by which they isoform via alternative mRNA splicing. Mol Cell Biol 1994; 14: 1647–1656. contribute to leukemogenesis. 15 McKinsey TA, Zhang CL, Olson EN. MEF2: a calcium-dependent regulator of cell division, differentiation and death. Trends Biochem Sci 2002; 27: 40–47. Acknowledgements 16 Lund AH, Turner G, Trubetskoy A, Verhoeven E, Wientjens E, Hulsman D et al. Genome-wide retroviral insertional tagging of genes involved in cancer in Cdkn2a-deficient mice. Nat Genet We thank Anne Olsen for supplying cosmids and BACs, Pauline 2002; 32: 160–165. Yen for the DAZAP1 cDNA, Ron Prywes for the p4xMEF2FLuc 17 Suzuki T, Shen H, Akagi K, Morse HC, Malley JD, Naiman DQ reporter construct, Amanda Rice and Chuhua Zhong for technical et al. New genes involved in cancer identified by retroviral advice and assistance. This work was supported by grants from the tagging. Nat Genet 2002; 32: 166–174.

Leukemia MEF2D fusion proteins in ALL V Prima et al 813 18 Tsui S, Dai T, Roettger S, Schempp W, Salido EC, Yen PH. 36 Bour BA, O’Brien MA, Lockwood WL, Goldstein ES, Bodmer R, Identification of two novel proteins that interact with germ-cell- Taghert PH et al. Drosophila MEF2, a transcription factor that is specific RNA-binding proteins DAZ and DAZL1. Genomics 2000; essential for myogenesis. Genes Dev 1995; 9: 730–741. 65: 266–273. 37 Lilly B, Zhao B, Ranganayakulu G, Paterson BM, Schulz RA, Olson 19 Vera Y, Dai T, Hikim AP, Lue Y, Salido EC, Swerdloff RS et al. EN. Requirement of MADS domain transcription factor D-MEF2 for Deleted in azoospermia associated protein 1 shuttles between muscle formation in Drosophila. Science 1995; 267: 688–693. nucleus and cytoplasm during normal germ cell maturation. 38 Lin Q, Schwarz J, Bucana C, Olson EN. Control of mouse cardiac J Androl 2002; 23: 622–628. morphogenesis and myogenesis by transcription factor MEF2C. 20 Jack I, Seshadri R, Garson M, Michael P, Callen D, Zola H et al. Science 1997; 276: 1404–1407. RCH-ACV: a lymphoblastic leukemia cell line with chromosome 39 Mao Z, Bonni A, Xia F, Nadal-Vicens M, Greenberg ME. Neuronal translocation 1;19 and trisomy 8. Cancer Genet Cytogenet 1986; activity-dependent cell survival mediated by transcription factor 19: 261–269. MEF2. Science 1999; 286: 785–790. 21 Ashworth LK, Batzer MA, Brandriff B, Branscomb E, de Jong P, 40 Woronicz JD, Lina A, Calnan BJ, Szychowski S, Cheng L, Winoto Garcia E et al. An integrated metric physical map of human A. Regulation of the Nur77 orphan steroid receptor in activation- chromosome 19. Nat Genet 1995; 11: 422–427. induced apoptosis. Mol Cell Biol 1995; 15: 6364–6376. 22 Boomer T, Varella-Garcia M, McGavran L, Meltesen L, Olsen AS, 41 Okamoto S, Li Z, Ju C, Scholzke MN, Mathews E, Cui J et al. Hunger SP. Detection of E2A translocations in leukemias via Dominant-interfering forms of MEF2 generated by caspase fluorescence in situ hybridization. Leukemia 2001; 15: 95–102. cleavage contribute to NMDA-induced neuronal apoptosis. Proc 23 Maloney KW, McGavran L, Murphy JM, Odom LF, Stork L, Wei Q Natl Acad Sci USA 2002; 99: 3974–3979. et al. TEL-AML1 fusion identifies a subset of children with standard 42 Han TH, Prywes R. Regulatory role of MEF2D in serum induction risk acute lymphoblastic leukemia who have an excellent of the c-jun promoter. Mol Cell Biol 1995; 15: 2907–2915. prognosis when treated with therapy that includes a single delayed 43 Pan F, Ye Z, Cheng L, Liu JO. Myocyte enhancer factor 2 mediates intensification. Leukemia 1999; 13: 1708–1712. calcium-dependent transcription of the interleukin-2 gene in T 24 Hunger SP, Ohyashiki K, Toyama K, Cleary ML. Hlf, a novel lymphocytes: a calcium signaling module that is distinct from but hepatic bZIP protein, shows altered DNA-binding properties collaborates with the nuclear factor of activated T cells (NFAT). following fusion to E2A in t(17;19) acute lymphoblastic leukemia. J Biol Chem 2004; 279: 14477–14480. Genes Dev 1992; 6: 1608–1620. 44 Nagai K, Oubridge C, Ito N, Avis J, Evans P. The RNP domain: a 25 Hunger SP, Brown R, Cleary ML. DNA-binding and transcriptional sequence-specific RNA-binding domain involved in processing regulatory properties of hepatic leukemia factor (HLF) and the and transport of RNA. Trends Biochem Sci 1995; 20: 235–240. t(17;19) acute lymphoblastic leukemia chimera E2A-HLF. Mol Cell 45 Delattre O, Zucman J, Plougastel B, Desmaze C, Melot T, Peter M Biol 1994; 14: 5986–5996. et al. Gene fusion with an ETS DNA-binding domain caused by 26 Hunger SP, Li S, Fall MZ, Naumovski L, Cleary ML. The proto- chromosome translocation in human tumours. Nature 1992; 359: oncogene HLF and the related basic leucine zipper protein TEF 162–165. display highly similar DNA-binding and transcriptional regulatory 46 Ron D. TLS-CHOP and the role of RNA-binding proteins in properties. Blood 1996; 87: 4607–4617. oncogenic transformation. Curr Top Microbiol Immunol 1997; 27 McCombie WR, Heiner C, Kelley JM, Fitzgerald MG, Gocayne JD. 220: 131–142. Rapid and reliable fluorescent cycle sequencing of double- 47 Mercher T, Coniat MB, Monni R, Mauchauffe M, Khac FN, Gressin stranded templates. DNA Seq 1992; 2: 289–296. L et al. Involvement of a human gene related to the Drosophila 28 Shore P, Sharrocks AD. The MADS-box family of transcription spen gene in the recurrent t(1;22) translocation of acute factors. Eur J Biochem 1995; 229: 1–13. megakaryocytic leukemia. Proc Natl Acad Sci USA 2001; 98: 29 Black BL, Olson EN. Transcriptional control of muscle develop- 5776–5779. ment by myocyte enhancer factor-2 (MEF2) proteins. Annu Rev 48 Barbouti A, Hoglund M, Johansson B, Lassen C, Nilsson PG, Cell Dev Biol 1998; 14: 167–196. Hagemeijer A et al. A novel gene, MSI2, encoding a putative RNA- 30 Inaba T, Roberts WM, Shapiro LH, Jolly KW, Raimondi SC, Smith binding protein is recurrently rearranged at disease progression of SD et al. Fusion of the leucine zipper gene HLF to the E2A gene in chronic myeloid leukemia and forms a fusion gene with HOXA9 as human acute B-lineage leukemia. Science 1992; 257: 531–534. a result of the cryptic t(7;17)(p15;q23). Cancer Res 2003; 63: 31 Wan L, Kim JK, Pollard VW, Dreyfuss G. Mutational definition of 1202–1206. RNA-binding and protein–protein interaction domains of hetero- 49 Miyamoto K, Tomita N, Ishii A, Nonaka H, Kondo T, Tanaka T geneous nuclear RNP C1. J Biol Chem 2001; 276: 7681–7688. et al. Chromosome abnormalities of leukemia cells in adult 32 Sun L, Youn HD, Loh C, Stolow M, He W, Liu JO. Cabin 1, a patients with T-cell leukemia. J Natl Cancer Inst 1984; 73: negative regulator for signaling in T lymphocytes. 353–362. Immunity 1998; 8: 703–711. 50 Vagner-Capodano AM, Mozziconacci MJ, Zattara-Cannoni H, 33 Youn HD, Sun L, Prywes R, Liu JO. Apoptosis of T cells mediated Guitard AM, Thuret I, Michel G. t(1;19) in a M4-ANLL [letter]. by Ca2+-induced release of the transcription factor MEF2. Science Cancer Genet Cytogenet 1994; 73: 86–88. 1999; 286: 790–793. 51 Wlodarska I, Stul M, De Wolf-Peeters C, Verhoef G, Mecucci C, 34 Youn HD, Chatila TA, Liu JO. Integration of calcineurin and MEF2 Cassiman JJ et al. t(1;19) without detectable E2A rearrangements in signals by the coactivator p300 during T-cell apoptosis. EMBO J two t(14;18)-positive lymphoma/leukemia cases [see comments]. 2000; 19: 4323–4331. Genes Chromosomes Cancer 1994; 10: 171–176. 35 Youn HD, Grozinger CM, Liu JO. Calcium regulates transcrip- 52 Yuki Y, Imoto I, Imaizumi M, Hibi S, Kaneko Y, Amagasa T et al. tional repression of myocyte enhancer factor 2 by histone Identification of a novel fusion gene in a pre-B acute lymphoblastic deacetylase 4. J Biol Chem 2000; 275: 22563–22567. leukemia with t(1;19)(q23;p13). Cancer Sci 2004; 95: 503–507.

Leukemia