ZNF652, a Novel Zinc Finger Protein, Interacts with the Putative Breast Tumor Suppressor CBFA2T3 to Repress Transcription

ZNF652, A Novel Zinc Finger Protein, Interacts with the Putative Breast Tumor Suppressor CBFA2T3 to Repress Transcription

Raman Kumar,1 Jantina Manning,1 Hayley E. Spendlove,3 Gabriel Kremmidiotis,4 Ross McKirdy,1 Jaclyn Lee,1 David N. Millband,1 Kelly M. Cheney,1 Martha R. Stampfer,5 Prem P. Dwivedi,2 Howard A. Morris,2 and David F. Callen1 1Breast Cancer Genetics Group, Dame Roma Mitchell Cancer Research Laboratories, Department of Medicine, University of Adelaide and Hanson Institute; 2Endocrine Bone Laboratory, Hanson Institute, Adelaide, South Australia, Australia; 3Department of Laboratory Genetics, Women’s and Children’s Hospital, North Adelaide, South Australia, Australia; 4Bionomics, Ltd., Thebarton, South Australia, Australia; and 5Lawrence Berkeley National Laboratory, Berkeley, California

Abstract gene effector zinc finger proteins may specifically The transcriptional repressor CBFA2T3is a putative interact with one or more of the ETO proteins to generate breast tumor suppressor. To define the role of CBFA2T3, a defined range of transcriptional repressor complexes. we used a segment of this protein as bait in a yeast (Mol Cancer Res 2006;4(9):655–65) two-hybrid screen and identified a novel uncharacterized protein, ZNF652. In general, primary tumors and cancer Introduction cell lines showed lower expression of ZNF652 than Tumor growth, characterized by unchecked cell division, normal tissues. Together with the location of this gene results from both the overexpression of growth-promoting on the long arm of chromosome 17q, a region of frequent oncogenes and the reduced expression of growth-inhibiting loss of heterozygosity in cancer, these results suggest tumor suppressor genes. These genes often encode proteins that In silico a possible role of ZNF652 in tumorigenesis. are components of coactivator and corepressor complexes analysis of this protein revealed that it contains multiple involved in the regulation of genes critical for cell division. classic zinc finger domains that are predicted to bind Identification and functional characterization of oncogenes and DNA. Coimmunoprecipitation assays showed that tumor-suppressor genes continues to be the key to an ZNF652 strongly interacts with CBFA2T3and this understanding of the molecular mechanisms of cancer. Detailed interaction occurs through the COOH-terminal 109 genetic and cytogenetic analyses of breast tumors and breast amino acids of ZNF652. In contrast, there was a weak cancer cell lines have shown that there is frequent loss of interaction of ZNF652 with CBFA2T1 and CBFA2T2, the heterozygosity at 16q24.3 (1, 2), suggesting that this band is the other two members of this ETO family. Transcriptional likely location of one or more tumor-suppressor genes. reporter assays further confirmed the strength Subsequent expression studies of genes located at 16q24.3 and selectivity of the ZNF652-CBFA2T3interaction. identified CBFA2T3 (also called MTG16) as a potential breast The transcriptional repression of growth factor tumor-suppressor gene (3). Molecular and cell biology assays independent-1 (GFI-1), a previously characterized ETO showed CBFA2T3 to have characteristics consistent with a effector zinc finger protein, was shown to be enhanced tumor suppressor because expression was significantly reduced by CBFA2T1, but to a lesser extent by CBFA2T2 and in primary breast tumors and in the breast tumor cell lines CBFA2T3. We therefore suggest that each of the various MDA-MB-468 and MDA-MB-231(4, 5). In addition, ectopic expression of CBFA2T3 in breast cancer cell lines inhibited both the ability to form colonies on plastic and anchorage- independent growth on soft agar (4). CBFA2T3, together with the homologues CBFA2T1 (MTG8, Received 11/27/05; revised 6/22/06; accepted 7/5/06. ETO) and CBFA2T2 (MTGR1), form the small ‘‘ETO’’ family Grant support: National Health and Medical Research Council of Australia grant (6), the terminology referring to the Eight-Twenty-One 207703, and Susan G. Komen Breast Cancer Foundation (D.F. Callen); U.S. Department of Energy under contract no. DE-AC03-76SF00098 (M.R. Stampfer); translocation associated with CBFA2T1. The ETO proteins Cancer Council of South Australia (H.A. Morris); Faculty of Health Sciences and show the highest homology within four NHR domains, Department of Medicine, University of Adelaide; and the Hanson Institute, originally identified in the Drosophila melanogaster protein Institute of Medical and Veterinary Science. The costs of publication of this article were defrayed in part by the payment of Nervy (7). The ETO family members function as transcriptional page charges. This article must therefore be hereby marked advertisement in repressors by forming complexes with the transcriptional accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Requests for reprints: David F. Callen, Breast Cancer Genetics Group, Dame corepressors N-CoR, SMRT, mSin3A, and recruit histone Roma Mitchell Cancer Research Laboratories, Hanson Institute, Institute of Medical deacetylases (HDAC). There are some differences between and Veterinary Science, Frome Road, Adelaide, SA 5000, Australia. Phone: 61-8- the ETO proteins in these associations, e.g., CBFA2T1and 8222-23145; Fax: 61-8-8222-3217. E-mail: [email protected]. Copyright D 2006 American Association for Cancer Research. CBFA2T2, but not CBFA2T3, which associate with mSin3A doi:10.1158/1541-7786.MCR-05-0249 (8, 9). The greater differences occur with HDAC interactions;

CBFA2T1interacts with HDAC1to HDAC3, CBFA2T2 CBFA2T3 in a similar screen where atrophin-176-1438 was used interacts only with HDAC3, whereas CBFA2T3 associates as a bait (24). In addition to atrophin-1, a cDNA encoding the with HDAC1, HDAC2, HDAC3, HDAC6, and HDAC8 (8, 9). 61COOH-terminal amino acids of a previously uncharacterized The ETO proteins possess two atypical conserved zinc-finger protein ZNF652 was identified. motifs (-C-x-x-C-7x-C-x-x-C- and -C-x-x-x-C-7x-H-x-x-x-C-) The 5,428 bp ZNF652 cDNA sequence (National Center that are involved in interaction with proteins but not DNA (10, for Biotechnology Information accession no. NM_014897) 11). The gene specificity of ETO-based repressor complexes is located at chromosome band 17q21.32 encodes a predicted determined by the recruitment of proteins that can directly bind 606-amino-acid protein with the presence of seven zinc finger to the promoters of target genes. The DNA-binding partners of motifs located in the central region of the protein. These the ETO proteins include C2H2 zinc finger transcription factors classic C2H2 zinc finger motifs, comprising two conserved BCL6 (12) and PLZF (13), which interact with CBFA2T1, and cysteine and histidine residues, conform to the consensus growth factor independent-1(GFI-1),which interacts with CX2CX12HX3H sequence and three of the zinc fingers are either CBFA2T1or CBFA2T3 (14).Therefore, a major joined by part or all of a consensus TGEKP linker sequence. mechanism whereby the ETO proteins impart their normal These motifs are common to zinc finger proteins involved in function is by transcriptional repression of diverse classes of DNA binding (25, 26), suggesting that ZNF652 is most likely genes through their interaction with different DNA-binding zinc a DNA-binding protein. finger proteins. In addition to interactions with zinc finger– binding proteins, CBFA2T1and CBFA2T2 have been shown to form complexes with the E-box–binding protein HEB, a basic- Expression of ZNF652 helix-loop-helix transcription factor, and these complexes The variation of ZNF652 expression in different normal mediate the roles of HEB in hematopoiesis (15). tissues and the corresponding matched tumors was determined ¶ CBFA2T1 is the best-characterized member of the ETO by probing a cDNA-profiling array with the 5 738 bp of the family due to its involvement in the t(8;21) translocation with ZNF652 open reading frame (Fig. 1A and B). The hybridization RUNX1 (previously called AML1) that generates a RUNX1- signals were normalized against the expression of the CBFA2T1 gene fusion, the major cause of acute myeloid housekeeping gene ubiquitin. There was a large variation in leukemia (16-18). It has been suggested that the normal the level of ZNF652 expression among different nonmalignant function of RUNX1 is to regulate genes that are critical for human tissues, with the highest average expression in the hematopoiesis (19). The RUNX1-CBFA2T1 fusion product normal breast, vulva, prostate, and pancreas. Compared with disrupts this regulation, thereby promoting progression to normal tissues, cancers of breast (one sample with no change, leukemia (20, 21). In patients with therapy-related acute one with overexpression, and the remaining eight showed 40% myeloid leukemia, RUNX1 can also be involved in a down-regulation), vulva (average 69% down-regulation), translocation, t(16;21)(q24;q22), with CBFA2T3 (22). Targeted prostate (three of four samples showed 34% down-regulation), disruption of Cbfa2t1 in mouse reveals a critical role in gut and pancreas (53% down-regulation) showed reduced levels of development (23), whereas disruption of the mouse Cbfa2t2 ZNF652 expression. When averaged over all samples, the shows a role in maintenance of the secretory cell lineage in the tumors showed a 34% down-regulation in ZNF652 expression small intestine (9). Because the ETO proteins may be compared with their matched normal tissues. functionally redundant, the observed phenotypes of these mice An affinity-purified rabbit anti-ZNF652 polyclonal antibody may reflect those tissues that only express a single member of was generated that specifically detected the endogenous the ETO family. ZNF652 protein (Fig. 2A). This antibody was used for Western To further define the role of CBFA2T3 as a tumor blot analysis of protein lysates from finite lifespan human suppressor, we used a yeast two-hybrid screen to identify mammary epithelial cells, nonmalignant immortalized mam- CBFA2T3-interacting proteins from a breast expression library. mary epithelial cells, and breast cancer cell lines. A specific From this screen, a previously uncharacterized zinc finger band of 85 kDa was detected in all cell lines (Fig. 2B). The protein ZNF652 was identified and shown to be a specific level of ZNF652 expression was similar in nonmalignant DNA-binding partner of CBFA2T3. We show that ZNF652 has immortalized and estrogen receptor–positive breast tumor cell a role in tumorigenesis and that ETO proteins selectively lines, but was reduced in three of four estrogen receptor– interact with the DNA-binding proteins ZNF652 and GFI-1. negative breast tumor cell lines.

Results ZNF652 Interacts Specifically with CBFA2T3 Identification and Analysis of ZNF652 To confirm the interaction of the COOH terminus of A yeast two-hybrid screen of a breast cDNA library was ZNF652 with CBFA2T3 observed in yeast, coimmunoprecipi- used to identify novel proteins that were part of the CBFA2T3 tation assays were done in HEK293T mammalian cells repressor complex. As the full-length CBFA2T3 autoactivated transiently expressing hemagglutinin (HA)-ZNF652 and myc- the reporter genes in this assay, CBFA2T3142-567 (amino acids CBFA2T3. HA-ZNF652 was clearly detected in complexes 142-567 of the CBFA2T3b open reading frame) was used as immunoprecipitated with anti-myc antibody that binds to myc- bait. A number of cDNA clones encoding different candidate CBFA2T3 (Fig. 3A, lane 4 and B, lane 7). This interaction was interacting proteins were identified in this screen. Atrophin-1, specific to CBFA2T3 as coimmunoprecipitation of the related one of the most frequently represented proteins, validated our proteins CBFA2T2 and CBFA2T1showed only a weak screening procedure as it has been reported to interact with interaction with ZNF652 (Fig. 3A, lanes 5 and 6, respectively).

Mol Cancer Res 2006;4(9). September 2006 Downloaded from mcr.aacrjournals.org on September 24, 2021. © 2006 American Association for Cancer Research. ZNF652 Interacts with CBFA2T3 to Repress Transcription 657

FIGURE 1. Expression of ZNF652 is reduced in cancer tissues. A. Cancer Profiling Array II was probed with radiolabeled PCR fragment (nucleotides 1- 738) of ZNF652 or a housekeeping ubiquitin cDNA. Levels of ZNF652 hybridization signals from normal and tumor samples of different tissues relative tothe respective ubiquitin signals were assayed and plotted. B. Part of the cancer-profiling array showing normal (N) and tumor (T) breast, vulva, and liver samples probed with ZNF652 (ZNF) or ubiquitin (UBI) radiolabeled DNA.

Negative controls confirmed the specificity of the interactions This interaction was apparently stabilized in the presence of the (not shown). Repeated attempts to perform reciprocal coimmu- zinc finger region because there was a more intense noprecipitations were unsuccessful, possibly due to masking of immunoprecipitated band observed with ZNF652243-606 (com- the HA epitope in the HA-ZNF652 protein. pare lanes 9 and 11 , Fig. 3B). These observations were also To map the region of ZNF652 that interacts with CBFA2T3, consistent with the finding that in the yeast two-hybrid screen, plasmid constructs expressing various HA-tagged regions of the identified ZNF652 cDNA encoded the COOH-terminal ZNF652 were generated (Fig. 3B, bars on right). Each of these region of the ZNF652 protein. In silico analysis of this COOH constructs was included in a coimmunoprecipitation assay with terminus did not identify any typical motifs associated with CBFA2T3. Only ZNF652, ZNF652243-606, and the COOH- protein binding, although a proline- and histidine-rich region terminal ZNF652498-606 proteins coimmunoprecipitated with was present (Fig. 3C). CBFA2T3 (Fig. 3B, lanes 7, 9, and 11 ). Negative controls confirmed the specificity of the interactions (not shown). These ZNF652 Function Is Dependent on CBFA2T3 results indicate that the COOH-terminal 109-amino-acid region The breast cancer cell line MCF-7 (an estrogen receptor– of ZNF652 (amino acids 498-606) interacts with CBFA2T3. positive breast tumor cell line) expresses comparatively higher

levels of CBFA2T3 message, whereas in SKBR3 (an estrogen more pronounced effect in the presence of CBFA2T3 (i.e., in receptor–negative breast tumor cell line) the levels are very low MCF-7 compared with SKBR3) suggests that both proteins are (ref. 5 and unpublished Western blot data). It was of interest to required to functionally interact to mediate their role as determine if ectopic expression of ZNF652 in these cell lines transcriptional repressors. resulted in different phenotypic effects because cell lines with low levels of CBFA2T3 (e.g., SKBR3) would be predicted to ZNF652 Transcriptional Repression Is Specifically show a reduced effect of ZNF652 ectopic expression compared Enhanced by CBFA2T3 with a cell line expressing higher levels of CBFA2T3 (e.g., To assess the effect of ZNF652 on transcription, dual- MCF-7). Accordingly, the effect of retroviral-mediated expres- luciferase reporter assays were done in Chinese hamster ovary sion of ZNF652 on the colony formation and proliferation (CHO) cells. A reference level of expression was observed in ability of MCF-7 and SKBR3 were compared with the effects cells transfected with the pGL2-GAL4-TK-luciferase plasmid of retroviral-mediated expression of the ETO proteins. Controls (Fig. 5A). Transfection of increasing amounts of plasmid were empty vector and expression of p53 (Fig. 4A). Ectopic expressing GAL4-ZNF652 caused a significant (P < 0.05) expression of CBFA2T3 in SKBR3 cells resulted in a greater exponential decrease in luciferase expression (Fig. 5A). reduction (relative to the vector control) in number of colonies Expression of a known strong repressor of transcription, compared with its expression in MCF-7 cells (Fig. 4A), GAL4-NK10 (27), validated the system as it caused a consistent with results obtained previously (4). However, when significant decrease in luciferase expression. These results the cell lines were transduced with retrovirus expressing indicated that ZNF652 functions as a repressor of transcription. ZNF652, the reduction in colony numbers was greater in Because ZNF652 interacts with CBFA2T3, it was predicted MCF-7 than SKBR3, opposite to the effects of CBFA2T3 that the repressive effect of ZNF652 would be enhanced in the ectopic expression. Similar findings were seen in proliferation presence of CBFA2T3. To test this hypothesis, dual-luciferase assays (Fig. 4B). Compared with the empty vector, ectopic reporter assays were done in CHO cells simultaneously expression of ZNF652 resulted in a minimal reduction in transfected with plasmids expressing GAL4-ZNF652 and proliferation of SKBR3, whereas proliferation was markedly myc-CBFA2T3 (Fig. 5B). Two different amounts of myc- reduced in cells ectopically expressing CBFA2T3. In MCF-7, CBFA2T3–expressing plasmid were used to achieve a dose- the effect was reversed with ectopic expression of CBFA2T3 dependent indication of the effects of this protein on luciferase showing no effect on proliferation compared with empty vector, expression. The results confirmed that GAL4-ZNF652 alone whereas expression of ZNF652 reduced proliferation (Fig. 4B). decreased luciferase reporter gene expression and that increas- For both SKBR3 and MCF-7, ectopic expression of CBFA2T1 ing amounts of myc-CBFA2T3 expression caused a further reduced cell proliferation. Ectopic expression of CBFA2T2 dose-dependent decrease in expression (Fig. 5B). Therefore, reduced MCF-7 proliferation but had no effect on SKBR3. CBFA2T3 further enhances the transcriptional repression Therefore, the finding that ectopic expression of ZNF652 has a caused by ZNF652. To test if this repression was specifically caused by CBFA2T3, or would also be displayed by the other ETO family members, dual-luciferase reporter assays were done using CHO cells simultaneously transfected with plasmids expressing GAL4-ZNF652 and either CBFA2T1or CBFA2T2 (Fig. 5B). For each protein, cells were transfected with 50 or 100 ng plasmid. In the presence of GAL4-ZNF652, increasing amounts of CBFA2T2 did not result in repression of luciferase activity, CBFA2T1. There was a significant decrease in luciferase reporter gene expression caused by simultaneous transfection of constructs expressing ZNF652 and CBFA2T1at 50 ng, and this decrease was not altered by an increasing amount (100 ng) of CBFA2T1 expressing plasmid. Therefore, compared with CBFA2T3, only CBFA2T1weakly enhances the transcriptional repression of ZNF652.

GFI-1 Transcriptional Repression Is Specifically Enhanced by CBFA2T1 FIGURE 2. Affinity-purified rabbit polyclonal anti-ZNF652497-606 anti- Because we showed that ZNF652 repression is enhanced body specifically detects the ZNF652 protein and ZNF652 levels vary selectively by CBFA2T3, we investigated for similar selectivity among different breast cancer cell lines. A. Western blots showing that anti-ZNF652497-606 antibody specifically detect the ZNF652 protein. in the interactions between ETO proteins and another zinc Western blots carrying ZR75-1 nuclear extracts (lanes 1, 3, 5, and 7) finger protein. CBFA2T1and CBFA2T3 have previously been and protein lysates from HEK293T cells transfected with pCMV-HA- ZNF652 (lanes 2, 4, 6, and 8) were probed with various primary (shown shown to interact with GFI-1, a zinc-finger transcription below each panel) and appropriate HRP-conjugated secondary anti- repressor (14). We recently showed that GFI-1 represses 25- bodies. B. Western blot carrying protein from various breast cell lines was hydroxyvitamin D 1a-hydroxylase (CYP27B1) expression in probed with anti-ZNF652497-606 antibody. Protein from pCMV-HA- ZNF652 – transfected HEK293T cells showed a single band at 85 kDa. human prostate cancer cells, and that this repression is mediated h-actin was used as a loading control. through GFI-1–specific, DNA-binding sequences within the

FIGURE 3. ZNF652 interacts strongly and specifically with CBFA2T3 through its COOH terminus. A. Immunoprecipitation of transfected ZNF652 and ETO proteins in HEK293T cells showing specific CBFA2T3 interaction. Constructs expressing myc-tagged ETO and HA-tagged ZNF652 proteins were used. Proteins were coimmunoprecipitated (IP) with anti-myc antibody and Western blotted (WB) with anti-HA or anti- myc antibodies (lanes 4-6). All proteins were detected at predicted sizes in the input (lanes 1-3). Nonspecific antibody coimmunoprecipitations showed no HA- ZNF652 bands (not shown). B. ZNF652 interacts with CBFA2T3 through its COOH terminus. Bar diagram, full-length HA-ZNF652 and four regions of the HA- ZNF652 used in immunoprecipitations. Proteins were coimmunoprecipitated with anti-myc antibody and Western blotted with anti-HA or anti-myc antibodies. All proteins were detected at predicted sizes in the input (lanes 1-6). Only the HA-ZNF652 proteins carrying the COOH-terminal 109 amino acids coimmunoprecipitated with CBFA2T3 (lanes 7, 9, and 11). Nonspecific antibody coimmunoprecipitations showed no HA-ZNF652 bands (not shown). C. ZNF652 contains a proline- and histidine-rich region. The COOH-terminal 109 amino acids of ZNF652 (beginning at amino acid 498) contains a proline-rich region (blocked sequence) and a histidine-rich region (under- lined sequence). M, markers.

CYP27B1promoter (28). Reporter assays were done to quantities of CBFA2T2- or CBFA2T3-expressing constructs. investigate if the ETO proteins differ in their ability to further Western blot analysis on LNCaP cells transfected with equal repress the GFI-1–mediated transcriptional repression from the amounts of either of the three ETO plasmids showed lower CYP27B1promoter. For these assays, three firefly luciferase levels of ectopic CBFA2T1expression than either CBFA2T2 or gene constructs containing CYP27B1promoter sequences were CBFA2T3 (data not shown). Therefore, the specific transcrip- used. The first construct, designated pCYP27B1WT(À1200)- tional repression caused by CBFA2T1is likely to be more Luc, contained a natural GFI-1–binding sequence located at pronounced than was actually observed. À1161/À1138; the second construct, pCYP27B1Enh(À997)- To determine if ectopic expression of GFI-1further Luc, was deleted for this GFI-1–binding sequence; and the enhanced the repression of CBFA2T1, reporter activation of third construct, pCYP27B1mGFI1(À1200)-Luc, contained a pCYP27B1WT(À1200)-Luc was assayed in cells transfected mutated GFI-1–binding sequence. The reporter assays were with plasmid constructs expressing both GFI-1and either of the done in LNCaP cells that express low to medium levels of three ETO proteins (Fig. 6B). Ectopic expression of both GFI-1 endogenous GFI-1(data not shown). Compared with and either of the three ETO proteins resulted in a decrease in the pCYP27B1Enh(À997)-Luc, pCYP27B1WT(À1200)-Luc luciferase activity of 54% (CBFA2T1), 27% (CBFA2T2), and showed repression in luciferase activity resulting from 37% (CBFA2T3). These results suggest that the GFI-1and endogenous GFI-1and possibly ETO proteins. This repression CBFA2T1interaction is functionally more significant than the was further significantly reduced in a dose-dependent manner interactions of GFI-1with the other two members of the ETO by ectopic expression of GFI-1(Fig. 6A). Ectopic expression of family. CBFA2T1also significantly repressed luciferase expression To confirm that CBFA2T1mediates its specific repressive from the pCYP27B1WT(À1200)-Luc construct in a dose- effect through GFI-1bound to specific sequences within the dependent manner. In contrast, there was a minor repression in pCYP27B1promoter, reporter assays were done using the luciferase expression in cells transfected with comparable pCYP27B1mGFI1(À1200)-Luc construct. Reporter activity in

LNCaP cells transfected with plasmid containing CYP27B1 ETO proteins have been shown to localize to the nucleus, promoter with a mutated GFI-1binding site was >2-fold function as transcriptional repressors, and can homodimerize higher than in the cells with the wild-type sequence and this and heterodimerize to generate a diversity of repressor did not change in the presence of GFI-1or the ETO proteins complexes (30, 31). The normal physiologic roles of the ETO (Fig. 6C). These results confirm that the repression of proteins are beginning to emerge, although their functions are transcription by CBFA2T1in cells transfected with likely to be broader than presently defined. CBFA2T1is a pCYP27B1WT(À1200)-Luc is mediated through sequence- regulator of early adipogenesis (32) and targeted disruption of specific binding of the endogenous GFI-1at the CYP27B1 Cbfa2t1 in the mouse shows a role in gut development (23). promoter. The role of CBFA2T2 is obscure although disruption of this gene in mice suggests that Cbfa2t2 maintains the secretory cell Discussion lineage in the small intestine (9). CBFA2T3 has been shown to Studies of CBFA2T1 and CBFA2T3 have centered on their coordinate cellular proliferation and differentiation during involvement in the t(8;21) and t(16;21) translocations that are erythropoiesis (33) and has a role as a human breast tumor associated with acute myeloid leukemia (29). These trans- suppressor (4). It is also clear that as ETO family members do locations generate fusion proteins containing the DNA-binding not have the ability to directly bind DNA, they initiate region of the RUNX1transcriptional activator and the majority repression of their respective target genes by interacting with of the transcriptional repressor proteins CBFA2T1or various DNA-binding proteins and by recruitment of corepres- CBFA2T3. The fusion proteins result in the repression of sors to the transcriptional repressor scaffold. For example, it has genes normally activated by the RUNX1protein. All three ETO been shown that CBFA2T1interacts with PLZF, GFI-1,BCL6, proteins have been shown to be transcriptional repressors in and HEB, and CBFA2T3 interacts with GFI-1to repress the reporter assays, and their in vivo function is likely to be transcription of various genes (12-15). The interaction of ETO mediated by the recruitment of various corepressors, including proteins with such sequence-specific, DNA-binding proteins is N-CoR, mSin3A, and HDACs (4, 8, 18, 20). Although ETO essential for their transcriptional repressor functions. proteins are very similar in their amino acid sequences, and Because CBFA2T3 was shown to be a putative tumor similar regions of the proteins are responsible for their repressor suppressor in breast cancer (4), we screened a normal breast function, the proteins do exhibit some differences in their cDNA library with the 142 to 567 amino acids of CBFA2T3 to ability to interact with these corepressors (8). identify interacting proteins. From this yeast two-hybrid screen,

FIGURE 4. ZNF652 and CBFA2T3 functionally interact in vivo. Breast cancer cell lines MCF-7 and SKBR3 transduced with retroviruses expressing ZNF652 or ETO proteins. A. Colony formation of transduced cell lines. Cells were plated 24 hours following transduction, grown for 2 weeks in G418-supplemented medium, and the number of colonies were scored. Col- umns, number of colonies relative to the control vector. B. Growth curves of transduced cell lines. Growth assays were initiated from cultures after selection in G418-supplemented medium for 6 days following transduction with retroviruses expressing either the ETO proteins, ZNF652, p53, or empty vector.

frequency on the long arm that would support, or refute, ZNF652 as a potential tumor suppressor in breast cancer. The levels of expression of ZNF652 protein in breast cancer cell lines may suggest a down-regulation in estrogen receptor– negative cell lines (Fig. 2B), although additional studies are required to confirm this finding. Based on the presence of seven classic (C2H2) zinc finger domains, together with the linkers with consensus sequence TGEKP joining three of these zinc fingers, ZNF652 is predicted to bind DNA (25, 36) and, therefore, be the DNA-binding component of a CBFA2T3 regulatory complex. The interaction of ZNF652 and CBFA2T3 was confirmed by coimmunoprecipitation in HEK293T mammalian cells (Fig. 3A and B; ref. 34). The interaction of ZNF652 and CBFA2T3 occurs through the 109 amino acids at the COOH terminus of ZNF652 (Fig. 3B). This is consistent with the finding that the initial cDNA clone identified in the yeast two-hybrid screen encoded the 61 COOH-terminal amino acids of the ZNF652 protein. Further- more, this interaction is apparently stabilized by the presence of the zinc finger region. A detailed protein database search does not detect significant homology of this region of ZNF652 to any other known proteins. However, further in silico analysis of this region of ZNF652 revealed proline- and histidine-rich regions (Fig. 3C) that are thought to mediate protein-protein interactions critical for the activation, inactivation, and functioning of some proteins (37, 38). These results suggest FIGURE 5. ZNF652 transcriptional repression is enhanced in the that ZNF652 is a specific DNA-binding partner of the presence of CBFA2T3. A. Increasing expression of ZNF652 caused an CBFA2T3 regulatory complex. exponential decrease in the level of luciferase reporter gene expression. Functional assays in the breast cancer cell lines MCF-7 and Dual-luciferase reporter assays were done to assess the effect of ZNF652 on transcription. CHO cells were transfected with the various combinations SKBR3 provided evidence that in vivo thefunctionofZNF652is of plasmid constructs indicated below each column. One hundred enhanced in the presence of CBFA2T3 (Fig. 4). The effect of nanograms of pGL2-GAL4-TK-luciferase provided a reference level of ZNF652 on reducing both the clonability and cell proliferation of expression. The positive control NK10 caused a strong repression. The amount of DNA in each transfection was supplemented with empty vector these cell lines was greater in MCF-7 compared with SKBR3. to a total of 100 ng. B. ZNF652 transcriptional repression is enhanced These effects inversely correlated to endogenous CBFA2T3 primarily by CBFA2T3. Dual-luciferase reporter assays were done as in A. Treatments included expression of increasing amounts of each of the ETO expression, which is high in MCF-7 and low in SKBR3 (ref. 5 proteins to assess the effect of ZNF652 and ETO family members on and unpublished Western blot results). The expression level of transcription. One hundred fifty nanograms of pGL2-GAL4-TK-luciferase ZNF652 in MCF-7 and SKBR3 is comparable (Fig. 2B). These provided a reference level of luciferase expression. The amount of DNA in each transfection was supplemented with empty vector to a total of 150 ng. studies in breast tumor cells suggest that ZNF652 function is There was no reduction of luciferase expression when each of the ETO largely dependent on the presence of CBFA2T3, and support an proteins was expressed in the presence of the pGL2-GAL4-TK-luciferase in vivo interaction of these two proteins. Ectopic expression of (data not shown). CBFA2T1and CBFA2T2 had less pronounced effects on the growth of SKBR3 but a more significant effect in MCF-7. This the novel protein ZNF652 was identified. A Northern blot may be due to the presence of additional ETO DNA-binding probed with a radiolabeled ZNF652 fragment showed that this partners in MCF-7 cells. gene is ubiquitously expressed in human tissues (34). These Transcriptional reporter assays expressing GAL4-DBD observations were confirmed by the results from a cancer- fused to ZNF652 showed that increasing amounts of ZNF652 profiling array (Fig. 1). Of particular interest was the caused an exponential decrease in the level of luciferase consistent reduction in ZNF652 expression observed among reporter gene expression and indicates that ZNF652 functions different types of tumors, and this was more pronounced in as an independent repressor of transcription (Fig. 5A). the breast, prostate, vulva, and pancreas. It is noteworthy that Similarly, BCL6, PLZF, and GFI-1, the other zinc finger ZNF652 is located on chromosome 17q21.32 at 44.7 Mb proteins that interact with the ETO family of proteins, have also (build 35.1of the Human Genome, National Center for been shown to be independent repressors of transcription Biotechnology Information), 6.3 Mb distal to BRCA1. (12-14). Transcriptional reporter assays also showed that the Chromosome 17 is frequently involved in breast tumor– repressor activity caused by ZNF652 was enhanced in the restricted loss of heterozygosity. A large consortium study of presence of CBFA2T3 (Fig. 5B). A similar scenario was seen in 1,280 breast carcinomas showed loss of heterozygosity studies on transcriptional repression mediated by BCL6 (12) averaged 31% at various 17q markers, including those and PLZF (13), where the presence of CBFA2T1 enhanced flanking the location of ZNF652 (35). However, there was their repression. It is not known whether these zinc finger no compelling evidence of peaks in loss of heterozygosity proteins can function in vivo to repress transcription in the

absence of the scaffold ETO proteins. We suggest that ZNF652 determined the repressor activity of GFI-1from the CYP27B1 is a putative breast cancer tumor suppressor because CBFA2T3 promoter. Coexpression of GFI-1and CBFA2T1resulted in a has this role (4), and the two proteins form a transcriptional higher level of repression than coexpression of GFI-1with repressor complex. In tumors, it is proposed that the normal either CBFA2T2 or CBFA2T3 (Fig. 6A and B). Taken together, transcriptional repressor function of the CBFA2T3 complex can these observations suggest that particular combinations of ETO be circumvented by down-regulation, e.g., by loss of proteins and specific DNA-binding transcription factors interact heterozygosity and methylation (1, 5), of either CBFA2T3 or to achieve maximal levels of target gene repression in vivo. ZNF652 expression. This novel finding implies that the ETO family of proteins Although the zinc finger proteins BCL6 and PLZF have generates a diverse group of complexes, with variation in both been shown to interact with CBFA2T1, and GFI-1 with both the architecture and the recruitment of the gene effector zinc CBFA2T1 and CBFA2T3 (12-14), there has been no evidence finger proteins. We therefore suggest that each of the various to suggest that the different zinc finger proteins specifically gene effector zinc finger proteins specifically interacts with one interact with particular ETO family members. However, or more of the ETO proteins to generate a defined range of coimmunoprecipitations done on transiently expressed proteins transcriptional repressor complexes that control the expression in mammalian cells provide evidence that ZNF652 strongly and of specific set of genes. The genes controlled by the ZNF652- specifically interacts with CBFA2T3, but weakly with CBFA2T3 complex are likely to have a role in the CBFA2T1and CBFA2T2 (Fig. 3A). It was of interest to tumorigenesis of breast and other tissues. determine if this observed selectivity of ZNF652 interactions Results using a luciferase reporter system provide an insight with ETO proteins was also reflected in a functional specificity. into the interactions of ZNF652 that may be occurring in vivo. This was indeed the case because in luciferase reporter assays, However, confirmation of these results using reporter-vector CBFA2T3 augmented the repression of ZNF652 but this was constructs carrying the actual DNA-binding sequence of reduced with CBFA2T1and CBFA2T2. To investigate if the ZNF652 upstream of the luciferase gene will be an important functional selectivity between ETO proteins and their interact- next step. We are currently working toward identifying the ing zinc finger proteins is a more general property, we ZNF652 DNA-binding sequences.

FIGURE 6. CBFA2T1 specifically represses transcription through its interaction with GFI- 1 from the CYP27B1 promoter. A. CBFA2T1 causes specific exponential decrease in the level of luciferase reporter gene expression from the CYP27B1 promoter sequences carrying a GFI-1 binding site. Dual-luciferase reporter assays were done to assess the effect of ETO proteins on GFI-1 – mediated transcription. LNCaP cells were transfected with the various combinations of plasmid constructs as indicated below each column. To transfect comparable amounts of DNA, plasmids were supplemented with empty vector. B. GFI-1 repression was specifically enhanced in the presence of CBFA2T1. LNCaP cells were transfected with pCYP27B1WT(À1200)- Luc and various other constructs as indicated below each column. C. CBFA2T1-3 did not repress transcription from the construct carrying the CYB27B1 promoter with a mutated GFI-1 DNA-binding sequence. T1 to T3, CBFA2T1, CBFA2T2, and CBFA2T3. pCYP27B1WT(À1200) contains the 1,200 bp of the CYP27B1 promoter that in- cludes the consensus GFI-1 – binding site. This binding site is deleted in pCYP27B1Enh- (À997)-Luc and mutated in the pCYP27B1mGFI1(À1200)- Luc construct.

Materials and Methods binding domain in pM (BD Biosciences). To generate pGL2- In silico Analysis GAL4-TK-luciferase construct, 5¶ and 3¶ ends of the XbaI- ZNF652 mRNA sequence was obtained from the National BamHI fragment carrying 5Â GAL4 DNA-binding sequences Center for Biotechnology Information (accession no. from pG5CAT (Promega, Madison, WI) were filled-in using NM_014897).6 Protein domain searches were done using Klenow DNA polymerase and cloned at blunted XhoIsiteof ExPASY-PROSITE,7 and protein homology searches were pGL2-TK-luciferase. Constructs used for GFI-1–mediated done using Basic Alignment Search Tools.6 repression assays were pCYP27B1WT(À1200)-Luc that contained a 1,244 bp region of the CYP27B1 promoter (À1200 ¶ Cell Lines and Antibodies together with 44 bp of CYP27B1 5 untranslated region), À 184B and 48RS are finite lifespan human mammary pCYP27B1Enh( 997)-Luc that was deleted for the region con- epithelial cell strains derived from two different individuals’ taining the GFI-1consensus binding site, and pCYP27B1mGFI reduction mammoplasty tissues, whereas 184A1 is a nonma- (-1200)-Luc that contained mutated GFI-1 binding site (28). The lignant immortally transformed cell line derived from normal construct expressing FLAG-GFI-1was kindly provided by H.L. specimen 184 (39). HEK293T (human embryonic kidney), Grimes (Institute for Cellular Therapeutics, University of Louis- CHO, and all other cell lines were purchased from the ville, Louisville, KY). The Renilla luciferase vector (pRL-TK) American Type Culture Collection (Manassas, VA) and grown was from Promega. For repression assays, the NH2 terminus myc-tagged ETO family members were cloned into pcDNA3.1 in the recommended medium at 37jCin5%CO2. Antibodies used were rat anti-HA (Roche Diagnostics, Indianapolis, IN), (Invitrogen, Carlsberg, CA) at XbaI/EcoRI sites for CBFA2T3b, mouse anti-myc (Santa Cruz Biotechnology, Santa Cruz, CA), ApaI/KpnIsitesforCBFA2T2,andXbaI/HindIII sites for horseradish peroxidase (HRP)–conjugated secondary antibody CBFA2T1. DNA was prepared and validated using standard (rabbit anti-rat-IgG-HRP; DakoCytomation, Carpinteria, CA), molecular biology techniques (40). sheep anti-mouse IgG-HRP, and donkey anti-rabbit IgG-HRP (Amersham Biosciences, Piscataway, CA). A rabbit polyclonal Yeast Two-Hybrid Assay anti-ZNF652 antibody was generated. A ZNF652 DNA Display GREEN basic yeast two-hybrid kit (Display fragment expressing amino acids 497 to 606 was cloned in- Systems Biotech, Copenhagen, Denmark) was used to per- frame into the bacterial expression vector pET14b (Novagen, form a yeast two-hybrid screen. DNA sequence encoding Madison, WI). The 6Â His-ZNF652497-606 protein induced CBFA2T3142-567 was PCR amplified with the two primers (the from BL21pLys bacteria harboring the expression construct reverse primer was designed with a 5¶-end extension to generate was purified on Ni-NTA superflow beads (Qiagen, Hilden, a sequence encoding a FLAG-tag at the COOH terminus) and Germany) and injected into White New Zealand rabbits. The cloned in-frame with the lexA DNA-binding protein in the rabbit polyclonal antibody was affinity-purified using bacteri- pdisplayBAIT vector. Human breast cDNA Hybrid Hunter ally expressed GST-ZNF652497-606 protein conjugated to library (Invitrogen) cloned in pYESTrp2 vector was screened CNBr-activated Sepharose 4B matrix (Amersham Biosciences). using pdisplayBAIT-CBFA2T3142-567 and pdisplayREPORTER vectors transformed into yeast H strain (MATa trp1 his3 ura3 Plasmids leu2 6 LexAops-LEU2). All plasmids were constructed with the complete open reading frame of human genes unless specified otherwise. For coimmu- Cancer Profiling noprecipitation assays, the plasmid construct expressing myc- Cancer Profiling Array II (BD Biosciences) was probed with CBFA2T3b (smaller isoform of CBFA2T3) was as reported a PCR product (nucleotides 1-738) of ZNF652. The probe was earlier (4), and CBFA2T1and CBFA2T2 open reading frames labeled with [a-32P]dCTP using the Megaprime DNA labeling were cloned in-frame into the BglII site of pCMV-myc (BD system (Amersham Biosciences). Hybridization and washing Biosciences, San Jose, CA). To generate constructs expressing conditions were as the manufacturer instructed. The membrane various regions of the ZNF652 protein, the relevant fragments of was stripped and then probed with the radiolabeled housekeep- the cDNA were PCR amplified from the existing construct using ing ubiquitin gene probe supplied with the kit. The hybridiza- primers with sequence extensions that created BamHI restriction tion signals were quantified using a PhosporImager (Molecular sites, and cloned in-frame into the BglII restriction site of pCMV- Dynamics), digitized using ImageJ 1.33u (NIH) software, and HA (BD Biosciences). Constructs generated were as follows: normalized for ubiquitin expression. ZNF652 (complete open reading frame, amino acids 1-606), ZNF6521-246 (amino acids 1-246), ZNF652243-606 (amino acids Retoviral-Mediated Expression Studies 243-606), ZNF652243-491 (amino acids 243-491), and Various relevant constructs in the pLNCX2 vector were used ZNF652498-606 (amino acids 498-606). Constructs with various to generate retrovirus and transduce MCF-7 and SKBR3 cells gene fragments in pLNCX2 were used for retroviral experiments. as previously described (41). Following transduction, cells were For luciferase reporter assays, ZNF652 or amino acids 1to 90 of plated in six-well plates, selected for 2 weeks in G-418– a known mouse zinc finger repressor NK10 protein (4, 27) were containing medium, and the number of colonies were scored. cloned in-frame with the region expressing the GAL4 DNA- Results are presented as the mean of three replicates F SE. Proliferation assays were determined on transduced cells 6 http://www.ncbi.nlm.nih.gov. selected for 6 days in the presence of G418. These cells were 7 http://www.expasy.org/prosite. plated at 10% to 20% confluence in 96-well plates and at

various times assayed by incubating cells with the CellTiter-Glo and expressed as relative luciferase units. Each treatment was Luminescent Cell Viability Assay (Promega) according to the done in triplicate and assays were repeated at least thrice. The instructions of the manufacturer. Results are presented as the data presented is of a representative experiment with FSE of mean relative luminescence units of three replicates. mean of three replicates. For GFI-1reporter assays, LNCaP cells (1.5 Â 105) were seeded into 24-well plates and Coimmunoprecipitations transfections were carried out in triplicate. Each transfection À Approximately 1 Â 106 HEK293T cells were transiently was done with 200 ng pCYP27B1WT( 1200)-Luc, À À transfected with 4 Ag of the each relevant plasmid using pCYP27B1Enh( 997)-Luc, or pCYP27B1mGFI( 1200)-Luc LipofectAMINE 2000 (Invitrogen), according to the instruc- together with varying amounts of constructs expressing FLAG- tions of the manufacturer. After 24 hours, cells were harvested GFI-1and/or either of the three ETO proteins and 50 ng pRL- and lysed in 50 mmol/L Tris-HCl (pH 8), 150 mmol/L NaCl, TK plasmid. Cells were harvested and analyzed as above. and 1% Triton X-100 supplemented with complete protease inhibitor cocktail (Roche Diagnostics). Lysates were sonicated References and cell debris was removed by centrifugation at 16,000 Â g. 1. Cleton-Jansen AM, Callen DF, Seshadri R, et al. Loss of heterozygosity mapping at chromosome arm 16q in 712 breast tumors reveals factors that Supernatants were incubated with 500 ng mouse monoclonal influence delineation of candidate regions. Cancer Res 2001;61:1171 – 7. 9E10 anti-myc antibody (Roche Diagnostics) overnight at 4jC, 2. Callen DF, Crawford J, Derwas C, et al. Defining regions of loss of followed by the addition of 15 AL sheep anti-mouse IgG heterozygosity of 16q in breast cancer cell lines. Cancer Genet Cytogenet 2002; Dynabeads (Invitrogen). Beads were washed with lysis buffer, 133:76 – 82. twice with wash buffer [50 mmol/L Tris-HCl (pH 8.0), 150 3. Powell JA, Gardner AE, Bais AJ, et al. Sequencing, transcript identification, and quantitative gene expression profiling in the breast cancer loss of mmol/L NaCl, 1% NP40, 0.5% sodium deoxycholate, and 0.1% heterozygosity region 16q24.3 reveal three potential tumor-suppressor genes. SDS], and then twice with low salt buffer [20 mmol/L Tris-HCl Genomics 2002;80:303 – 10. (pH 7.5)]. Proteins were eluted with 50 ALof1Â SDS protein- 4. Kochetkova M, McKenzie OL, Bais AJ, et al. CBFA2T3 (MTG16) is a putative breast tumor suppressor gene from the breast cancer loss of loading buffer [0.0625 mol/L Tris-HCl (pH 6.8), 2% SDS, 10% heterozygosity region at 16q24.3. Cancer Res 2002;62:4599 – 604. h j glycerol, and 5% -mercaptoethanol] for 5 minutes at 95 C. 5. Bais AJ, Gardner AE, McKenzie OL, et al. Aberrant CBFA2T3B gene Inputs and coimmunoprecipitates were subjected to SDS-PAGE promoter methylation in breast tumors. Mol Cancer 2004;3:22. and transferred to a nitrocellulose membrane (Hybond C-Extra, 6. Calabi F, Cilli V. CBFA2T1, a gene rearranged in human leukemia, is a Amersham Biosciences). member of a multigene family. Genomics 1998;52:332 – 41. 7. Feinstein PG, Kornfeld K, Hogness DS, Mann RS. Identification of homeotic target genes in Drosophila melanogaster including nervy, a proto-oncogene Western Blotting homologue. Genetics 1995;140:573 – 86. For extracting proteins, cells from different cell lines were 8. Amann JM, Nip J, Strom DK, et al. ETO, a target of t(8;21) in acute leukemia, collected at 80% confluence; lysed in 50 mmol/L Tris-HCl makes distinct contacts with multiple histone deacetylases and binds mSin3A through its oligomerization domain. Mol Cell Biol 2001;21:6470 – 83. (pH 7.5), 250 mmol/L NaCl, 1% Triton X-100, 1 mmol/L 9. Amann JM, Chyla BJ, Ellis TC, et al. Mtgr1is a transcriptional corepressor EDTA, 50 mmol/L NaF, 0.1mmol/L Na 3VO4, 1mmol/L DTT, that is required for maintenance of the secretory cell lineage in the small intestine. and 1Â protease inhibitors (Roche Diagnostics) on ice for Mol Cell Biol 2005;25:9576 – 85. 15 minutes; clarified; and then assayed using BCA protein 10. Kitabayashi I, Ida K, Morohoshi F, et al. The AML1-8 leukemic fusion protein forms a complex with a novel member of the MTG8(ETO/CDR) family, assay reagent kit (Pierce, Rockford, IL). Twenty-five micro- MTGR1. Mol Cell Biol 1998;18:846 – 58. grams protein from each sample were used for Western blotting 11. Lutterbach B, Sun D, Schuetz J, Hiebert SW. The MYND motif is required analysis. For detecting endogenous ZNF652 in different cancer for repression of basal transcription from the multidrug resistance 1promoter by cell lines, affinity-purified rabbit polyclonal anti-ZNF652497-606 the t(8;21) fusion protein. Mol Cell Biol 1998;18:3604 – 11. antibody and donkey anti-rabbit IgG-HRP were used. Equal 12. Chevallier N, Corcoran CM, Lennon C, et al. ETO protein of t(8;21) AML is a corepressor for Bcl-6 B-cell lymphoma oncoprotein. Blood 2004;103:1454 – 63. loading of different proteins was confirmed by probing the 13. Melnick AM, Westendorf JJ, Polinger A, et al. The ETO protein disrupted in same membrane with mouse anti-h-actin antibody (Sigma- t(8;21)-associated acute myeloid leukemia is a corepressor for the promyelocytic Aldrich, St Louis, MO). For analyzing the immunoprecipitates, leukemia zinc finger protein. Mol Cell Biol 2000;20:2075 – 86. Western blots carrying inputs and immunoprecipitated samples 14. McGhee L, Bryan J, Elliott L, et al. Gfi-1 attaches to the nuclear matrix, were first incubated with primary antibody (rat anti-HA or associates with ETO (MTG8) and histone deacetylase proteins, and represses transcription using a TSA-sensitive mechanism. J Cell Biochem 2003;89: mouse anti-myc) followed by appropriate HRP-conjugated 1005 – 18. antibodies. Proteins were visualized by enhanced chemilumi- 15. Zhang J, Kalkum M, Yamamura S, Chait BT, Roeder RG. E protein silencing nescence kit (Amersham Biosciences). by the leukemogenic AML1-ETO fusion protein. Science 2004;305:1286 – 9. 16. Hagemeijer A, Garson OM, Kondo K. Fourth International Workshop on Chromosomes in Leukemia 1982: translocation (8;21)(q22;q22) in acute Luciferase Assays nonlymphocytic leukemia. Cancer Genet Cytogenet 1984;11:284 – 7. Approximately 2 Â 105 CHO cells were transiently trans- 17. Linggi B, Muller-Tidow C, van de Locht L, et al. The t(8;21) fusion protein, fected using LipofectAMINE 2000 with 700 ng pGL2-GAL4- AML1ETO, specifically represses the transcription of the p14(ARF)tumor suppressor in acute myeloid leukemia. Nat Med 2002;8:743 – 50. TK-luciferase, constructs expressing varying amounts of 18. Wang J, Hoshino T, Redner RL, Kajigaya S, Liu JM. ETO, fusion partner in GAL4-ZNF652 and/or myc-CBFA2T1, -T1 or -T3, and 50 ng t(8;21) acute myeloid leukemia, represses transcription by interaction with the pRL-TK plasmid (Promega) as an internal transfection control. human N-CoR/mSin3/HDAC1complex. Proc Natl Acad Sci U S A 1998;95: Cells were harvested after 16 hours, lysed, and assayed using 10860 – 5. 19. Okuda T, van Deursen J, Hiebert SW, Grosveld G, Downing JR. AML1, the the dual-luciferase reporter assay system (Promega). Firefly target of multiple chromosomal translocations in human leukemia, is essential for luciferase values were normalized to Renilla luciferase activity normal fetal liver hematopoiesis. Cell 1996;84:321 – 30.

20. Lutterbach B, Hiebert SW. Role of the transcription factor AML-1in acute 32. Rochford JJ, Semple RK, Laudes M, et al. ETO/MTG8 is an inhibitor of C/ leukemia and hematopoietic differentiation. Gene 2000;245:223 – 35. EBPh activity and a regulator of early adipogenesis. Mol Cell Biol 2004;24: 21. Peterson LF, Zhang DE. The 8;21 translocation in leukemogenesis. 9863 – 72. Oncogene 2004;23:4255 – 62. 33. Goardon N, Lambert JA, Rodriguez P, et al. ETO2 coordinates cellular proliferation and differentiation during erythropoiesis. EMBO J 2006;25: 22. Gamou T, Kitamura E, Hosoda F, et al. The partner gene of AML1 in t(16;21) 357 – 66. myeloid malignancies is a novel member of the MTG8(ETO) family. Blood 1998; 91:4028 – 37. 34. Holland J. KIAA0924 is a putative transcriptional repressor potentially involved in breast cancer. Medical and Pharmaceutical Biotechnology (Honors). 23. Calabi F, Pannell R, Pavloska G. Gene targeting reveals a crucial role for Thesis. Adelaide (Australia): Division of Health Sciences, University of South MTG8 in the gut. Mol Cell Biol 2001;21:5658 – 66. Australia; 2002. 24. Wood JD, Nucifora FC, Jr., Duan K, et al. Atrophin-1, the dentato-rubral and 35. Phelan CM, Borg A, Cuny M, et al. Consortium study on 1280 breast pallido-luysian atrophy gene product, interacts with ETO/MTG8 in the nuclear carcinomas: allelic loss on chromosome 17 targets subregions associated with matrix and represses transcription. J Cell Biol 2000;150:939 – 48. family history and clinical parameters. Cancer Res 1998;58:1004 – 12. 25. Wolfe SA, Nekludova L, Pabo CO. DNA recognition by Cys2His2 zinc 36. Laity JH, Dyson HJ, Wright PE. DNA-induced a-helix capping in conserved finger proteins. Annu Rev Biophys Biomol Struct 2000;29:183 – 212. linker sequences is a determinant of binding affinity in Cys(2)-His(2) zinc fingers. 26. Matthews JM, Sunde M. Zinc fingers—folds for many occasions. IUBMB J Mol Biol 2000;295:719 – 27. Life 2002;54:351– 5. 37. Catling AD, Schaeffer HJ, Reuter CW, Reddy GR, Weber MJ. A proline-rich 27. Lange R, Christoph A, Thiesen HJ, et al. Developmentally regulated mouse sequence unique to MEK1and MEK2 is required for raf binding and regulates gene NK10 encodes a zinc finger repressor protein with differential DNA-binding MEK function. Mol Cell Biol 1995;15:5214 – 25. domains. DNA Cell Biol 1995;14:971 – 81. 38. Kay BK, Williamson MP, Sudol M. The importance of being proline: the 28. Dwivedi PP, Anderson PH, Omdahl JL, et al. Identification of growth factor interaction of proline-rich motifs in signaling proteins with their cognate domains. independent-1 (GFI1) as a repressor of 25-hydroxyvitamin D 1-a hydroxylase FASEB J 2000;14:231 – 41. (CYP27B1) gene expression in human prostate cancer cells. Endocr Relat Cancer 39. Stampfer MR, Yaswen P. Human epithelial cell immortalization as a step in 2005;12:351 – 65. carcinogenesis. Cancer Lett 2003;194:199 – 208. 29. Davis JN, McGhee L, Meyers S. The ETO (MTG8) gene family. Gene 2003; 40. Ausubel FM, Brent R, Kingston RE, et al. Current protocols in molecular 303:1– 10. biology. Cambridge: John Wiley and Sons, Inc.; 2001. 30. Rossetti S, Hoogeveen AT, Sacchi N. The MTG proteins: chromatin 41. Kumar R, Neilsen PM, Crawford J, et al. FBXO31 is the chromosome repression players with a passion for networking. Genomics 2004;84:1– 9. 16q24.3 senescence gene, a candidate breast tumor suppressor, and a component 31. Hug BA, Lazar MA. ETO interacting proteins. Oncogene 2004;23:4270 – 4. of an SCF complex. Cancer Res 2005;65:11304 – 13.

Mol Cancer Res 2006;4(9). September 2006 Downloaded from mcr.aacrjournals.org on September 24, 2021. © 2006 American Association for Cancer Research. ZNF652, A Novel Zinc Finger Protein, Interacts with the Putative Breast Tumor Suppressor CBFA2T3 to Repress Transcription

Raman Kumar, Jantina Manning, Hayley E. Spendlove, et al.

Mol Cancer Res 2006;4:655-665.

Updated version Access the most recent version of this article at: http://mcr.aacrjournals.org/content/4/9/655

Cited articles This article cites 39 articles, 19 of which you can access for free at: http://mcr.aacrjournals.org/content/4/9/655.full#ref-list-1

Citing articles This article has been cited by 5 HighWire-hosted articles. Access the articles at: http://mcr.aacrjournals.org/content/4/9/655.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mcr.aacrjournals.org/content/4/9/655. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.