The Ajuba LIM Domain Protein Is a Corepressor for SNAG Domain–Mediated Repression and Participates in Nucleocytoplasmic Shuttling

Research Article

Kasirajan Ayyanathan,1,4 Hongzhuang Peng,1 Zhaoyuan Hou,1 William J. Fredericks,1 Rakesh K. Goyal,3 Ellen M. Langer,2 Gregory D. Longmore,2 and Frank J. Rauscher III1

1The Wistar Institute, Philadelphia, Pennsylvania; 2Department of Medicine, Washington University School of Medicine, St.Louis, Missouri; 3Blood and Marrow Transplantation Program, Children’s Hospital of Pittsburgh, Pittsburgh, Pennsylvania; 4Center for Molecular Biology and Biotechnology, Department of Biological Sciences, Florida Atlantic University, Boca Raton, Florida

Abstract with the components of the basal transcription machinery, (b) ATP-dependent remodeling of chromatin, and (c) site-specific The SNAG repression domain is comprised of a highly modifications such as acetylation, phosphorylation, methylation, conserved 21–amino acid sequence, is named for its presence and ubiquitination of the nucleosomal histones, which serve as in the Snail/growth factor independence-1 class of zinc finger biochemical codes in the context of complex chromatin structure transcription factors, and is present in a variety of proto- (2).A widely accepted paradigm is that histone acetylation leads to oncogenic transcription factors and developmental regula- gene activation, whereas histone deacetylation results in repres- tors. The prototype SNAG domain containing oncogene, sion.Importantly, many repressors function through corepressor growth factor independence-1, is responsible for the develop- molecules as exemplified by DR-DRAP-1 (3), WRPW-Groucho (4), ment of T cell thymomas. The SNAIL proteins also encode the REST-Co-REST (5), and KRAB–KRAB-associated protein-1 (KAP-1; SNAG domain and play key roles in epithelial mesenchymal ref.6). differentiation events during development and metastasis. The SNAG repression domain (Snail/GFI-1) is present in the Significantly, these oncogenic functions require a functional vertebrate homologues of the Snail/Slug and in the growth factor SNAG domain. The molecular mechanisms of SNAG domain– independence-1 (GFI-1) ZFPs (7–10).The GFI-1 proto-oncogene mediated transcriptional repression are largely unknown. was cloned by using an insertional mutagenesis strategy wherein Using a yeast two-hybrid strategy, we identified Ajuba, a Moloney murine leukemia virus–infected T cells selected for multiple LIM domain protein that can function as a interleukin 2 independence showed non–random integration of corepressor for the SNAG domain. Ajuba interacts with the the provirus at the GFI-1 locus (11).Remarkably, GFI-1 was SNAG domain in vitro and in vivo, colocalizes with it, and independently cloned as a gene, which could cooperate with myc enhances SNAG-mediated transcriptional repression. Ajuba and pim1 in transgenic models of B- and T-cell lymphoma (12). shuttles between the cytoplasm and the nucleus and may form GFI-1 is a nuclear localized ZFP which recognizes a 12 bp DNA a novel intracellular signaling system. Using an integrated consensus sequence (13, 14).The GFI-1 ZFP encodes the 20–amino reporter gene combined with chromatin immunoprecipita- acid SNAG domain in its NH2 terminus.This segment is sufficient tion, we observed rapid, SNAG-dependent assembly of a for repression, and single amino acid substitutions in the SNAG multiprotein complex that included Ajuba, SNAG, and histone domain abolish repression (10).Consistent with its role as a T cell– modifications consistent with the repressed state. Thus, SNAG tropic oncogene, overexpression of GFI-1 in immortalized T cells domain proteins may bind Ajuba, trapping it in the nucleus allowed them to escape the G1 arrest induced by interleukin 2 where it functions as an adapter or molecular scaffold for the withdrawal and, interestingly, a SNAG domain mutant of GFI-1 assembly of macromolecular repression complexes at target was inactive in this capacity (10).Gene knockout studies have promoters. [Cancer Res 2007;67(19):9097–106] shown that mammalian GFI-1 genes are essential for the development of erythroid and megakaryocytic lineages as well as neutro- Introduction phil differentiation (15–17). Transcription factors are modular in architecture, often encod- The vertebrate homologues of the fly Snail and Slug genes ing sequence-specific DNA binding and functionally separable encode SNAG domains and their roles in development and disease effector domains.The Cys2-His2 (C 2H2) type zinc finger proteins have been extensively studied (18).In Xenopus, the Snail/Slug (ZFP) constitute the largest family of transcription factors in the family has been shown to play essential roles in both mesoderm human proteome (1).The ZFPs can be classified into distinct differentiation and in neural crest induction/migration, and these subfamilies such as SNAG-, KRAB-, BTB/POZ-, and SCAN-ZFPs functions require a competent SNAG domain.Both fibroblast based on these highly conserved functional domains often encoded growth factor and hepatocyte growth factor induce expression of Slug, which in turn, induces epithelial-mesenchymal transitions in their NH2 termini.The mechanisms used by these domains during transcriptional regulation include: (a) direct interaction (19).The E2A-HLF fusion protein has been shown to activate the human Slug gene with resultant interleukin-3–independent survival and transformation of pro–B cells (20).Snail family members also play crucial roles in the development of mesoderm and nervous Requests for reprints: Frank J.Rauscher III, The Wistar Institute, 3601 Spruce system by triggering epithelial-mesenchymal transition.Recent Street, Philadelphia, PA 19104-4268.Phone: 215-898-0995; Fax: 215-898-3929; E-mail: studies have shown that the human and mouse E-cadherin [email protected]. I2007 American Association for Cancer Research. promoters are direct targets for Snail ZFPs and the loss of doi:10.1158/0008-5472.CAN-07-2987 E-cadherin expression is a major contributor to the acquisition of www.aacrjournals.org 9097 Cancer Res 2007; 67: (19). October 1, 2007

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 2007 American Association for Cancer Research. Cancer Research an invasive, highly malignant phenotype during human tumor laser-scanning microscope with the help of the Wistar Institute Cancer progression (21, 22).Thus, it is clear that the SNAG-ZFPs play Center Microscopy Core Facility.Images were captured using QED Imaging important and diverse roles in embryonic development and in software. human disease.Despite these advances in defining the roles of Chromatin immunoprecipitation. SPHL11 or SPHL20 cells were plated Â 6 SNAG domain proteins in biological processes, the molecular at 5 10 cells/150 mm dish, treated continuously with either 500 nmol/L of 4-hydroxytamoxifen (4-OHT; +OHT dishes) or 0.1% ethanol (ÀOHT mechanisms of SNAG-mediated transcriptional repression are still dishes) for 2 days, and then fixed in 1% formaldehyde (EM Biosciences) for largely undefined. 20 min at 37jC.Solubilized and sonicated chromatin was prepared as In this article, we identify Ajuba (23, 24) and LIMD1 (25) proteins previously described (29), and immunoprecipitated with a-PAX3, a-Ajuba, as SNAG domain–interacting proteins and provide strong evidence a-LIMD1, a-AcH3, a-AcH4, or a-H3-MeK9 antibodies.Usually, 10% of the that they function as corepressors for the SNAG domain. clarified chromatin was saved as input.The immunocomplexes were processed as previously described (29).Both the input and the immunoprecipitated DNAs were used in quantitative PCR reactions with the primer Materials and Methods pairs illustrated in Fig.5 A.The PBS1 (5¶ GATCGATAATTCGAGCTACTG 3¶) Cell lines and antibodies. COS-1, NIH3T3, and its stable cell line and PBS2 (5¶ GAGCTCGGTACCCGGGTCG 3¶) primer pair amplify the PAX3 derivatives (SPHL11 and SPHL20) were grown as described (26).The binding sites; the primer pair TKP1 (5¶ GCGCGGTCCCAGGTCCACTT 3¶) a-PAX3 (27) and a-LEXA IgGs (Santa Cruz Biotechnology) are rabbit and and LUC1 (5¶ TCCAGGAACCAGGGCGTATCTCT 3¶) amplify the HSV-TK goat polyclonal antibodies which recognize the PAX3 and LEXA DNA- promoter region (26).The DNA fragments were electrophoresed on 1.5% binding domains, respectively.The a-Ajuba and a-LIMD1 are polyclonal agarose gels and either photographed or subjected to Southern blotting and antibodies raised by immunizing rabbits with 6-His fusion proteins of autoradiography. mouse Ajuba (amino acids 1–216) and mouse LIMD1 (amino acids 1–158) as antigens.These antibodies only recognize their cognate antigens and do not cross-react in the assays tested.The a-MYC tag monoclonal antibody Results (clone 9E10), a-AcH3, a-AcH4 (raised against acetylated histone H3 and Characterization of engineered SNAG repressors. The SNAG H4), and a-H3-MeK9 (raised against trimethylated lysine 9, histone H3) transcription factor superfamily includes Snail, Slug, GFI-1, Scratch, antibodies were purchased from Upstate. and IA-1 ZFPs.In spite of possessing a varied number of C 2H2-type Construction of a SNAG domain minigene and its derivatives. A zinc fingers at their COOH-termini, they all contain the highly synthetic SNAG domain coding sequence derived from GFI-1 was conserved NH -terminal SNAG repression domain (Fig.1 A). constructed by standard overlap-extension PCR.A Kozak consensus was 2 added prior to the initiator methionine of the 20–amino acid SNAG domain Multiple sequence alignment using Clustal W indicates that the f (MPRSFLVKSKKAHSYHQPRS) followed by a 6–amino acid spacer SNAG domain homology extends to 20 amino acids with the (PGPDYS).In-frame EcoRI (5¶) and SalI(3¶) sites were included for cloning NH2-terminal first seven amino acids being strictly conserved into the pSP73 vector.Alanine substitution mutagenesis was done via PCR between the different members (Fig.1 B).Furthermore, the SNAG using appropriate mutant oligonucleotides.The wild-type and mutant domains of GFI-1 and GFI-1B genes also possess a well-conserved SNAG minigenes were subcloned into PAX3-, LEXA-, and pBTM116-LEXA nuclear localization signal (KSKK).To determine if the synthetic containing plasmids to create NH2 terminal fusions. SNAG domain would function as a transferable, modular repression Immunoprecipitation and electrophoretic mobility shift assays. domain, we fused it to the NH2 terminus of the minimal PAX3 Each expression plasmid was verified for stable protein expression in vivo as DNA-binding domain (PAX3; Fig.1 C).We have used PAX3 as a previously described (26).The myc-AJUBA and LIMD1 plasmids have also 35 recipient (26, 27, 30) for the following reasons: PAX3 (a)binds been described (23).Briefly, [ S]-L-methionine–labeled whole cell extracts prepared from transiently transfected COS-1 cells were immunoprecipitated DNA as a monomer and recognizes an extended nondegenerate with a-PAX3 IgG or a-LEXA IgG, and proteins analyzed by SDS-PAGE DNA binding site due to paired and homeodomain-DNA binding and fluorography.In coimmunoprecipitation experiments, subsequent to motifs, (b) is easily detectable using PAX3 antibodies, and (c)is cotransfection of either SNAG-PAX3 or SNAG-LEXA with the MYC-Ajuba transcriptionally neutral when bound to DNA in the absence of an plasmid (1:1 ratio), immunoprecipitation was carried out using ELB buffer effector domain.We used alanine scanning mutagenesis to map (28).Nuclear extracts were prepared using a rapid protocol (27) and used the amino acid sequence requirements for SNAG repression.We in electrophoretic mobility shift assays with a 32P-labeled e5 site (a PAX3- targeted both the NH2 terminal block of seven amino acids, which binding site derived from the engrailed gene) probe. is identical among all SNAG domains and the more COOH-terminal Transcriptional repression assays. To monitor the repression poten- region, which is not as highly conserved (Fig.1 C).Each protein was tials of the chimeric SNAG repressor proteins, 2 Â 105 NIH3T3 cells were comparably expressed (Fig.1 D) and showed comparable DNA transfected with 1 Ag of the expression plasmid along with 0.5 Ag of PAX3- or LEXA-luciferase reporter plasmids, and 0.25 Ag of pCMV-LacZ plasmids, binding activity to the e5 site (Fig.1 E).Expression plasmids using LipofectAMINE (Life Technologies).Whole cell extracts were assayed encoding SNAG-PAX3 fusions were cotransfected with the PAX3- for luciferase activities and then normalized to h-galactosidase values for luciferase reporter plasmid and cell extracts assayed for activity. transfection efficiency (26, 27).Fold repression was determined as the ratio Fusion of the SNAG domain to the NH2-terminus of PAX3 created a of normalized light units in vector versus SNAG-expression plasmid– potent repressor (Fig.1 F).Repression was dose-dependent and was transfected cells.To examine the effect of Ajuba on SNAG-PAX3 or SNAG- strictly contingent on PAX3-binding sites in the reporter plasmid LEXA–mediated repression, cotransfection was done as described above. (data not shown).The mutants varied in their abilities to repress Immunofluorescence and colocalization. Immunocytochemistry was the PAX3-luciferase in reporter assays (Fig.1 F).The FLV-AAA done essentially as previously described (28).NIH3T3 cells were transfected substitution severely reduced repression.However, the S-A and KK- with the indicated expression plasmids.The cells were fixed, permeabilized, AA mutants retained significant repression activity, whereas the and then incubated with a-LEXA (1:100 dilution) or a-MYC (1:1,000 dilution) antibodies.The proteins were detected with FITC-conjugated single substitution F-A severely reduced repression.Thus, a a-rabbit IgG (1:500 dilution) or Texas red–conjugated a-mouse IgG (1:1,000 phenylalanine to alanine substitution in the highly conserved dilution), respectively.Finally, the cells were stained for DNA with Hoechst SNAG domain converts a normally powerful repressor of (2 ng/mL) and mounted on glass slides using Fluoromount G (Southern transcription into a neutral DNA binding protein.Whether the Biotechnology Associates, Inc.). Cells were visualized using a Leica confocal seven–amino acid NH2 terminal segment is sufficient for repression

Cancer Res 2007; 67: (19). October 1, 2007 9098 www.aacrjournals.org

Figure 1. Characteristics of SNAG-zinc finger transcription factors. A, alignment of human SNAG-ZFPs. Presence of a highly conserved NH2-terminal SNAG repression module and a variable COOH-terminal zinc finger arrays are highlighted. The Slug and Scratch ZFPs also contain conserved slug and scratch domains, respectively. B, SNAG domain homology. The SNAG repression domain extends to f20 amino acids, of which the NH2-terminal first seven amino acids are strictly conserved. C, diagrammatic representation of the wild-type and mutant SNAG-PAX3 repressors. D, protein expression. The expression plasmids mentioned in C were transfected into COS-1 cells, 35S-labeled cell extracts were immunoprecipitated with a-PAX3 IgG, and analyzed by SDS-PAGE and fluorography. E, gel shift. Nuclear extracts made from transfected COS-1 cells were incubated with 32P-labeled e5 probe and electrophoresed on a nondenaturing PAGE. Arrows, shifted bands. F, transcriptional repression competence of SNAG-PAX3 repressors. The wild-type and mutant SNAG-PAX3 expression plasmids were transfected into NIH3T3 cells along with PAX3-luciferase reporter and CMV-lacZ plasmids. Values for normalized luciferase activity obtained for each construct (bottom).

www.aacrjournals.org 9099 Cancer Res 2007; 67: (19). October 1, 2007

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 2007 American Association for Cancer Research. Cancer Research remains to be tested.Interestingly, a modular repression domain this spectrum of ‘‘baits’’ formed the important set of reagents as small as four amino acids (WRPW) present in the hairy and with which to identify and distinguish true mediators of SNAG achaete-scute bHLH transcription factors is sufficient to confer repression. repression by recruiting the Groucho corepressor (4).We suggest First, we showed that the pBTM116-SNAG-LEXA constructs that the SNAG domain may function in a similar manner. neither activated nor repressed transcription in yeast (data not A yeast two-hybrid strategy to identify SNAG domain– shown).The wild-type SNAG-LEXA fusion was then cotransformed associated proteins. Because direct protein-protein interactions with a murine embryonic (E9-E11) library constructed in pVP16 might be required for SNAG-mediated repression, we screened and scored for histidine prototrophy and LacZ activity in a for such proteins using yeast two-hybrid methods.The 20–amino standard two-hybrid screen (31).From f200 primary hits that acid SNAG domain was used as bait by fusing the SNAG domain were obtained upon screening, f40 million library clones, we f (wild-type and mutants) to the NH2-terminus of LEXA (Fig.2 A).To isolated 20 that interacted strongly with the wild-type SNAG guard against spatial constraints to protein binding, we included a domain.Several of these were re-screened by reintroduction into 17–amino acid, inert spacer between the SNAG domain and the naı¨ve yeast with a variety of other baits.SNAG domain–associated NH2 terminus of LEXA.The SNAG-LEXA fusions were properly proteins, SNAP13 and SNAP20, interacted very strongly with expressed in mammalian cells (Fig.2 A), exhibited proper binding wild-type SNAG but failed to interact with the SNAG (FLV-AAA) to a LEXA DNA-binding site in gel shift assays (data not shown), mutant, LEXA, or any of the other negative control baits including and repressed the LEXA-luciferase reporter plasmid (Fig.2 A).Thus, KAP-1, Lamin, Rho, and protein kinase C (Fig.2 B).We further

Figure 2. Identification and characterization of SNAPs. A, characteristics of SNAG-LEXA repressors. The wild-type and mutant SNAG-LEXA repressors were tested for stable protein expression, and also monitored for repression of a LEXA-luciferase reporter. B and C, specificity of SNAPs. B, the SNAP13 and SNAP20 clones bind to wild-type SNAG, but not to SNAG (FLV-AAA), or unrelated baits in the mating assay. C, these SNAPs bind to SNAG domains that retain repression activity, but not to either LEXA DNA-binding domain or to mutants that abolish repression. Note that the KK-AA mutant is not stably expressed either in COS-1 (A) or in yeast (data not shown), which accounts for its poor association. D, group 3 LIM domain proteins. The SNAG-interacting regions of SNAP13 and SNAP20 contain two NH2-terminal LIM domains of LIMD1 and Ajuba, respectively. The characteristic Gly-, Pro-rich regions and nuclear export signals (NES) present in these family members are highlighted. E, the signature Cys-His spacing (CX2CX16–23HX2CX2CX2CX16–21CX2–3C/ H/D; C, cysteine; H, histidine; D, aspartic acid; and X, any amino acid), and conserved residues (black bars) characteristic to LIM domains are shown as exemplified for by Ajuba LIM domains. F, phylogenetic tree analysis. Sequence conservation between the LIM domains of the indicated three LIM domain proteins was determined using DNASTAR software.

Cancer Res 2007; 67: (19). October 1, 2007 9100 www.aacrjournals.org

Figure 3. Ajuba functions as a SNAG corepressor. A, Ajuba interacts with repression-competent SNAG domains in vivo. The wild-type and mutant SNAG-LEXA plasmids were either transfected alone or cotransfected along with Ajuba into COS-1 cells. The 35S-labeled cell extracts were immunoprecipitated with a-LEXA or a-MYC antibodies and analyzed. Arrow, presence of coprecipitated Ajuba protein. B, enhancement of SNAG-mediated transcriptional repression by Ajuba. The wild-type and mutant SNAG-PAX3 or SNAG-LEXA plasmids were either transfected alone or cotransfected with MYC-Ajuba. Values for relative luciferase activities obtained after normalization (bottom).

characterized SNAP13 and SNAP20 by introducing them into interaction motifs, which encode a signature Cys2-His1-Cys3-Cys naı¨ve yeast along with each of the pBTM116-SNAG-LEXA baits Zn2+ binding domain as shown for the three individual LIM (Fig.2C ).The interactions were almost completely concordant domains of Ajuba (Fig.2 E).A dendrogram analysis constructed with repression activity; each SNAP interacted with active SNAG from the LIM domains of the three LIM domain proteins revealed repressors (wild-type and S-A mutants) but did not interact with that a distinct phylogenetic relationship exists only between the the mutants (FLV-AAA and F-A), which abolish repression.The LIM domains of Ajuba and LIMD1 (Fig.2 F).The SNAG-interacting exception to this concordance is KK-AA, which is expressed at a LIM domains of Ajuba and LIMD1 (f134 amino acids) fell into much-reduced level in yeast (data not shown).Thus, by these a single group with a similarity index of 62.7, implying that they criteria, these SNAPS are strong candidates for mediators of evolved from a common ancestral gene. SNAG function. Ajuba interacts with SNAG in vivo and functions as a The DNA sequence of the SNAPs showed that the proteins corepressor. To verify that SNAG-LIM domain interactions occur were in-frame with the VP16 activation domain encoded by the in vivo in mammalian cells, we used expression vectors for SNAG- vector.In the primary screen, we obtained two independent PAX3, SNAG-LEXA, and MYC epitope–tagged, full-length Ajuba isolates of SNAP13 and multiple, identical isolates of SNAP20. (MYC-Ajuba).Proper expression of the MYC-Ajuba protein was BLAST searches revealed that SNAP13 and SNAP20 were identical confirmed by immunoprecipitation with a-MYC monoclonal to the two NH2-terminal LIM domains of Ajuba and LIMD1, antibody (data not shown).Next, we cotransfected the MYC-Ajuba respectively.LIM is an acronym of three transcription factors, and wild-type or mutant SNAG-LEXA plasmids into COS-1 cells. Lin11, Isl-1, and Mec-3, in which the motif was first identified (32). The 35S-methionine-labeled cell lysates were immunoprecipitated The LIM domain protein subfamily, which includes Ajuba and with either a-MYC or a-LEXA IgG.Figure 3 A shows that the LIMD1, is illustrated in Fig.2 D.The Ajuba, LIMD1, and three a-MYC detects a 66 kDa MYC-Ajuba protein whereas the a-LEXA other LIM domain proteins such as Zyxin, TRIP6, and lipoma- detects a 25 kDa LEXA protein in the MYC-Ajuba + LEXA vector preferred partner proteins contain characteristic glycine/proline- cotransfected cells.However, in the wild-type SNAG-LEXA + MYC- rich regions and nuclear export signals.They belong to the group Ajuba cotransfected cells, the a-LEXA also detects the MYC-Ajuba 3 LIM domain family, the members of which exhibit nucleocyto- protein.As expected, the a-MYC also detects the wild-type SNAG- plasmic distribution and mediate both cellular and nuclear LEXA protein.Coimmunoprecipitation experiments with mutant signaling events (33).LIM domains are bona fide protein-protein SNAG-LEXA plasmids indicated that the repression-incompetent www.aacrjournals.org 9101 Cancer Res 2007; 67: (19). October 1, 2007

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 2007 American Association for Cancer Research. Cancer Research mutants of the SNAG domain fail to bind the Ajuba, strongly SNAG-LEXA repressed the LEXA-luciferase reporter, and this suggesting that this interaction is functionally relevant (Fig.3 A). repression was enhanced by CMV-Ajuba.As expected, Ajuba also To confirm that the region of Ajuba required for SNAG enhanced SNAG-PAX3–mediated repression of the PAX3 luciferase interaction contained the LIM domain, we used the expression reporter (Fig.3 B).These results suggest that Ajuba mediates SNAG plasmids coding for the NH2- or COOH-terminal segments of Ajuba repression and functions as a SNAG corepressor. (defined as ‘‘Ajuba-preLIM’’ or ‘‘Ajuba-LIM’’; ref.23) in cotransfec- Recruitment and nuclear colocalization of Ajuba and SNAG tion experiments along with the SNAG-PAX3 plasmid.The in NIH3T3 cells. The functional studies prompted us to investigate SNAG-PAX3 protein coprecipitated with the Ajuba protein or the the subcellular localization patterns of the MYC-Ajuba, the wild- Ajuba-LIM domain but not the Ajuba-preLIM domain, suggesting type and mutant SNAG-LEXA proteins.Immunofluorescence that the LIM region of Ajuba is the site of SNAG domain interaction analysis revealed that cells transfected with MYC-Ajuba showed (data not shown).Thus, Ajuba-SNAG-PAX3 complexes are very predominant cytoplasmic staining and those that were transfected efficiently formed in vivo following cotransfection and are stable with SNAG-LEXA (wild-type and mutants) showed intense nuclear to moderate salt, detergent extraction procedures.Importantly, staining (Fig.4 A).However, cells that received both Ajuba and this coprecipitation was not detected when a PAX3 protein that SNAG-LEXA plasmids showed abundant nuclear staining of lacks a SNAG domain was used, confirming that a SNAG-Ajuba Ajuba that was solely dependent on the repression competency interaction is occurring (data not shown). of the SNAG domain (Fig.4 B).A majority of cells in this doubly We did cotransfection experiments to determine if Ajuba could transfected population showed the pattern observed in Fig.4 B. modify the transcriptional repression function of the SNAG These remarkable images strongly corroborate our functional domain.We transfected CMV vectors for SNAG-PAX3 or SNAG- studies described earlier.These results also suggest that the SNAG LEXA constructs into NIH3T3 cells, with or without CMV-Ajuba. domain can recruit and/or retain Ajuba in the nucleus, and are

Figure 4. SNAG domain colocalizes with Ajuba in a repression competence– dependent manner. A, localization patterns of Ajuba and SNAG-LEXA proteins. The MYC-Ajuba, wild-type, and mutant SNAG-LEXA plasmids were transfected individually into NIH3T3 cells and immunostaining was carried out with a-LEXA or a-MYC antibodies. Note that Ajuba remains mainly cytoplasmic whereas the wild-type and mutant SNAG-LEXA proteins are predominantly nuclear. B, colocalization of repression competent SNAG-LEXA proteins with MYC-Ajuba protein. The wild-type and mutant SNAG-LEXA plasmids were cotransfected along with the MYC-Ajuba plasmid into NIH3T3 cells. Binding, interaction in yeast two-hybrid analysis. Repression, transcriptional competence in reporter assays. A representative cell is shown for each transfected population.

Cancer Res 2007; 67: (19). October 1, 2007 9102 www.aacrjournals.org

Figure 5. SNAG-mediated transcriptional repression mechanisms on a chromatinized locus. A, development of a dual plasmid system. The SNAG-PAX3-HBD chimeric repressor protein is constitutively expressed from the CMV promoter. The reporter plasmid contains six repeats of PAX3 DNA-binding motifs and a HSV thymidine kinase (HSV TK) promoter that controls the basal expression of the luciferase gene. This plasmid also contains a zeocinR cassette for drug selection. The primer-pair PBS1 and PBS2amplify the PAX3-binding sites (344 bp), whereas the TKP1 and LUC1 primer-pair amplify the TK promoter region (257 bp). B, C, and D, ChIP assays identified putative components of the SNAG repression pathway. B, 4-OHT–dependent enrichment of the SPHBD at the PAX3 sites. Chromatin prepared from (À) or (+) 4-OHT–treated SPHL11 and SPHL20 cells were immunoprecipitated with a-PAX3 (10 Ag) and analyzed by quantitative PCR using PBS1 and PBS2primer-pairs. Arrow, the PAX3 site fragment (344 bp) amplified. C, Ajuba and LIMD1 proteins were inducibly recruited to the TK promoter. ChIP experiments were done with (À) or (+) 4-OHT–treated SPHL11 cells using a-Ajuba (20 AL) and a-LIMD1 (20 AL), a-MeK9 (5 AL) antisera. The immunoprecipitated DNA was amplified using TKP1 and LUC1 primer-pair. Arrow, the amplified TK promoter fragment (257 bp). D, 4-OHT–dependent promoter hypoacetylation. ChIP experiments were done with (À) or (+) 4-OHT–treated SPHL11 and SPHL 20 cell lines using a-AcH3 IgG (5 AL) or a-AcH4 (5 AL) IgG and the retained DNA was tested by quantitative PCR using TKP1 and LUC1 primer-pair. Arrow, the amplified TK promoter fragment (257 bp).

consistent with our earlier observations that Ajuba is both cyto- potentials.In the presence of 4-OHT, f6-fold repression was plasmic and nuclear.Recent reports suggest that group 3 LIM observed (26); however, in the presence of both 4-OHT and TSA, proteins like Zyxin and Ajuba can shuttle between the cytoplasm this repression was totally relieved in both clones (26).This and the nucleus, and thus, may form a novel intracellular signaling experiment suggests that one of the mechanisms of SNAG domain system.It has been hypothesized that there are nuclear ‘‘anchors’’ repression may involve histone deacetylation.To confirm that the for this class of shuttling LIM proteins (34).We hypothesize and SNAG-Ajuba machinery was recruited to the target gene, we did present evidence that the SNAG domain may be such an anchor for comprehensive ChIP experiments using a battery of antibodies: the Ajuba protein. chromatin-associated proteins were cross-linked to DNA in vivo Molecular components of SNAG repression are recruited to with formaldehyde in mock or 4-OHT–treated SPHL11 and SPHL20 the target promoter. We investigated whether the SNAG domain– cells and immunoprecipitated with a-PAX3, a-Ajuba, a-LIMD1, mediated transcriptional repression involves chromatin modifica- a-AcH3, a-AcH4, and a-H3-MeK9 antibodies.The immunoprecipi- tion by using trichostatin A (TSA) and by performing chromatin tated DNA was analyzed by quantitative PCR, using primer pair immunoprecipitation (ChIP) assays.We generated stable cell lines, targeting the 6xPAX3 binding sites (PBS1 and PBS2), the TK which contained a chromatin-integrated PAX3-luciferase reporter promoter regions (TKP1 and LUC1) or the CD19-TK-LUC zeocinR transgenes and a conditional SNAG repressor SNAG-PAX3-HBD locus (26).Fragments, which bracket the PAX3-binding sites, were wherein the SNAG-PAX3 is fused to the estrogen receptor considerably enriched in the PAX3 immunoprecipitates after hormone-binding domain (SPHBD; Fig.5 A).This system has been 4-OHT treatment, indicating that there was a strong enrichment extensively used by this laboratory (26, 27).We tested multiple cell of the SPHBD protein at the PAX3 sites.This experiment was lines for protein expression and hormone-dependent (4-OHT) controlled by the presence of similar levels of the PAX3 binding site repression of the PAX3-luciferase reporter (26).We chose the fragment in the input chromatin (Fig.5 B).Likewise, we observed a SPHL11 and SPHL20 cell clones to study due to their repression strong 4-OHT–dependent enrichment of the 257 bp TK promoter www.aacrjournals.org 9103 Cancer Res 2007; 67: (19). October 1, 2007

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 2007 American Association for Cancer Research. Cancer Research fragment in a-Ajuba, a-LIMD1, and a-H3-MeK9 chromatin transcription factor Lhx3 binds to the pituitary-specific transcrip- immunoprecipitates (Fig.5 C).Thus, in addition to the DNA- tion factor Pitx-1 and also interacts with several coactivator/ binding component, other components of the SNAG repression adapter proteins including the selective LIM-binding protein (44); complex (i.e., Ajuba, LIMD1, and H3-MeK9) were inducibly and (g) cysteine-rich LIM-only proteins (CRP1 and CRP2) play recruited to the target gene.Based on these results, we conclude important roles in organizing multiprotein complexes, both in the that the SPHBD fusion protein, Ajuba, LIMD1, and H3-MeK9 are cytoplasm, where they participate in cytoskeletal remodeling, and strongly recruited to the chromatin regions comprising the PAX3 in the nucleus, where they strongly facilitate smooth muscle binding site, and the TK promoter region, respectively. differentiation (45).These studies clearly show that nuclear LIM In parallel ChIP experiments, we observed a significant reduction proteins can function as adapter molecules in the formation of in the levels of acetylated histones H3 and H4 at the TK promoter large multiprotein complexes that form on DNA and that influence region following 4-OHT-treatment, indicating that histone hypo- transcription. acetylation occurs upon recruitment of SPHBD protein to the Our colocalization studies show that Ajuba by itself remains upstream PAX3 sites (Fig.5 D).These results are in agreement with totally cytoplasmic, however, in the presence of a repression- the repression of this locus by SPHBD following hormone competent SNAG domain, its localization completely shifted to the treatment.Future ChIP experiments will unravel the identity of nucleus.The SNAG domain may facilitate both the import of Ajuba the enzymes involved in SNAG repression.We expect these experi- into the nucleus and its subsequent retention by serving as a ments to yield clues about the nature of the unique nucleocyto- nuclear anchor in chromatin.Similar nuclear targeting functions plasmic signaling mechanism used by Ajuba and LIMD1 proteins. have been observed for other LIM proteins.For example, ( a) stimulation of the Rho signaling pathway induces translocation of the transcriptional LIM-only coactivator FHL2 to the nucleus (46); Discussion (b) in heart, FHL2 interacts with hNP220, a DNA-binding nuclear Mammalian zinc finger transcription factors play vital roles protein to serve as a molecular adapter in the formation of a during organism development and subsequent homeostasis.In multiprotein complex (47); (c) the LIM protein KyoT2, an order to gain insights into the molecular mechanisms of SNAG- alternatively spliced murine isoform of SLIM1 has been shown to mediated repression that are responsible for the biological func- negatively regulate transcription by interacting with RBP-J DNA- tions of SNAG-ZFPs, we isolated and characterized SNAG-specific binding protein and displacing it from the DNA (48, 49); (d) corepressors.The highly modular nature of SNAG and other repres- interaction of Zyxin (usually found in focal adhesions) with the sion domains such as KRAB and BTB/POZ has greatly aided the human papillomavirus–derived E6 protein leads to its accumula- discovery of their mechanisms of action (6, 35–38).We followed tion in the nucleus where it functions as a transcriptional activator. this approach in our earlier studies to understand the mechanistic Significantly, this interaction requires the three LIM domains details of the KRAB-mediated repression by isolating and charac- present at the COOH terminus of Zyxin (50); and (e) the lipoma- terizing the KAP-1 corepressor (6, 28).In this study, we have iden- preferred partner protein exhibits nucleocytoplasmic distribution tified Ajuba and LIMD1 proteins as downstream players in the and possesses focal adhesion and nuclear targeting capabilities. SNAG repression pathway. It does not contain any consensus nuclear localization signals After the initial identification of LIM domain proteins, Ajuba suggesting that it may be imported into the nucleus via a nuclear and LIMD1, as SNAG-domain interactors in the yeast two-hybrid localization signal containing transport protein (51). assay, we did extensive biochemical analyses to confirm this The colocalization studies also suggested that upon SNAG- interaction.Our striking results from coimmunoprecipitation mediated translocation, Ajuba along with its (unidentified) experiments indicate that the SNAG domain interacts avidly with associated proteins could constitute a macromolecular repression Ajuba in vivo and also that Ajuba can enhance SNAG-mediated complex at the target promoter.Based on our findings from the repression.As expected, we observed these results only with ChIP experiments, we favor the notion that the initial DNA-binding repression-competent SNAG domains.The KAP-1 corepressor, by the SNAG-PAX3 repressor at the PAX3 binding sites leads to the which interacts only with repression-competent KRAB domains, subsequent recruitment of Ajuba, and associated proteins to the also enhance KRAB-mediated repression.Because Ajuba possesses promoter region.Because we also observed significant hypoacety- these important characteristics of corepressor proteins similar to lation of histone tails at the promoter region and reversal of KAP-1, we designate Ajuba as a SNAG corepressor. SNAG-mediated repression in the presence of TSA, we believe that LIM domain interactions have been observed in several classes SNAG repression involves histone deacetylases (HDAC) and that of transcription factors and cofactors.Specific examples include: HDAC may be constituents in the SNAG-holo repression complex. (a) the nuclear LIM-only (LMO) protein interact with GATA-1, HDAC and Sin3A have been found in SNAIL complexes, which GATA-2, and Tal1/Scl transcription factors (39); (b) the LIM repress E-cadherin (52).However, our preliminary data suggests domain–binding protein Ldb1, its LIM-only protein partner LMO2, that arginine methyltransferases also play an important role.5 and the DNA-binding SCL/E12, exist in a multiprotein complex It is well established that HDACs can be found in both the that negatively regulates erythroid function (40); (c) FHL2/DRAL cytoplasm and the nucleus.Ajuba may interact with a cytosolic interacts with the promyelocytic leukemia zinc finger protein HDAC and recruit it to the nucleus, or it may interact with an (a sequence-specific transcriptional repressor) and functions as a HDAC1 existing in the nucleus and recruit it to the repression corepressor by augmenting promyelocytic leukemia zinc finger– complex.The RLIM has been shown to interact with members of mediated transcriptional repression (41); (d) FHL2 associates with the HDAC corepressor complex (53).It would be interesting to Jun and Fos and serves as a coactivator of activator protein by identify the nature of the HDAC involved, and to decipher the stimulating Fos- and Jun-dependent transcription (42); (e) the CLIM forms a protein complex consisting of LIM-HD, LMO, bHLH, and GATA at the target promoters (43); (f ) the LIM-homeodomain 5 Z.Hou et al.,unpublished data.

Cancer Res 2007; 67: (19). October 1, 2007 9104 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 2007 American Association for Cancer Research. Ajuba and LIMD1 Are SNAG Domain–Interacting Proteins hitherto unidentified associated proteins that constitute the Acknowledgments macromolecular repressor complex.This necessitates the isolation Received 8/6/2007; accepted 8/17/2007. and characterization of interacting proteins that are present in Grant support: K.Ayyanathan is supported by the NIH (KO1-CA095620).G.D. both cytoplasmic and nuclear compartments, experiments which Longmore is supported by the American Heart Association (9940116N), the Edward Mallinckrodt, Jr., Foundation, and the NIH (RO1-CA75315). G.D. Longmore is an are ongoing.These findings provide another example whereby a Established Investigator of the American Heart Association.F.J.Rauscher is supported small conserved repression domain, the 21–amino acid (SNAG) in part by NIH grants, Core grant CA10815, CA92088, CA095561, DAMD17-96-1-6141, 17-02-1-0631, the Irving A.Hansen Memorial Foundation, the Susan G.Komen Breast functions by recruiting a large macromolecular complex.These Cancer Foundation, and the Emerald Foundation.Support from the Commonwealth results are particularly relevant to the SNAIL-SNAG proteins which Universal Research Enhancement Program, Pennsylvania Department of Health is are up-regulated during the processes of epithelial-mesenchymal gratefully acknowledged. The costs of publication of this article were defrayed in part by the payment of page transition and metastases in many epithelial cancers.A small charges.This article must therefore be hereby marked advertisement in accordance molecule strategy to block the SNAIL-SNAG-Ajuba interaction may with 18 U.S.C. Section 1734 solely to indicate this fact. We thank S.Hollenberg for the murine embryonic (E9-E11) VP16 fusion library, be a feasible and attractive target to provide either preventive or P.Tsichlis for many helpful discussions, and D.C.Schultz for the LEXA-luciferase therapeutic efficacy during tumor progression. reporter plasmid.

References 16. Saleque S, Cameron S, Orkin SH.The zinc-finger JR, Rauscher FJ III.An engineered PAX3-KRAB tran- proto-oncogene Gfi-1b is essential for development of scriptional repressor inhibits the malignant phenotype 1. Hamilton AT, Huntley S, Kim J, Branscomb E, Stubbs the erythroid and megakaryocytic lineages.Genes Dev of alveolar rhabdomyosarcoma cells harboring the L.Lineage-specific expansion of KRAB zinc-finger 2002;16:301–6. endogenous PAX3-FKHR oncogene.Mol Cell Biol 2000; transcription factor genes: implications for the evolu- 17. Hock H, Hamblen MJ, Rooke HM, et al.Intrinsic 20:5019–31. tion of vertebrate regulatory networks.Cold Spring Harb requirement for zinc finger transcription factor Gfi-1 in 31. Hollenberg SM, Sternglanz R, Cheng PF, Weintraub Symp Quant Biol 2003;68:131–40. neutrophil differentiation.Immunity 2003;18:109–20. H.Identification of a new family of tissue-specific basic 2. Berger SL.The complex language of chromatin 18. Manzanares M, Locascio A, Nieto MA.The increasing helix-loop-helix proteins with a two-hybrid system.Mol regulation during transcription.Nature 2007;447:407–12. complexity of the Snail gene superfamily in metazoan Cell Biol 1995;15:3813–22. 3. Kim S, Na JG, Hampsey M, Reinberg D.The Dr1/ evolution.Trends Genet 2001;17:178–81. 32. Freyd G, Kim SK, Horvitz HR.Novel cysteine-rich DRAP1 heterodimer is a global repressor of transcrip- 19. Mayor R, Morgan R, Sargent MG.Induction of the motif and homeodomain in the product of the tion in vivo.Proc Natl Acad Sci U S A 1997;94:820–5. prospective neural crest of Xenopus.Development 1995; Caenorhabditis elegans cell lineage gene lin-11.Nature 4. Fisher AL, Ohsako S, Caudy M.The WRPW motif of 121:767–77. 1990;344:876–9. the hairy-related basic helix-loop-helix repressor pro- 20. Inukai T, Inoue A, Kurosawa H, et al.SLUG, a ces-1- 33. Kadrmas JL, Beckerle MC.The LIM domain: from the teins acts as a 4-amino-acid transcription repression related zinc finger transcription factor gene with cytoskeleton to the nucleus.Nat Rev Mol Cell Biol 2004; and protein-protein interaction domain.Mol Cell Biol antiapoptotic activity, is a downstream target of the 5:920–31. 1996;16:2670–7. E2A-HLF oncoprotein.Mol Cell 1999;4:343–52. 34. Nix DA, Beckerle MC.Nuclear-cytoplasmic shuttling 5. You A, Tong JK, Grozinger CM, Schreiber SL.CoREST 21. Cano A, Perez-Moreno MA, Rodrigo I, et al.The of the focal contact protein, zyxin: a potential mecha- is an integral component of the CoREST-human histone transcription factor snail controls epithelial-mesenchy- nism for communication between sites of cell adhesion deacetylase complex.Proc Natl Acad Sci U S A 2001;98: mal transitions by repressing E-cadherin expression.Nat and the nucleus.J Cell Biol 1997;138:1139–47. 1454–8. Cell Biol 2000;2:76–83. 35. Peng H, Begg GE, Schultz DC, et al.Reconstitution of 6. Friedman JR, Fredericks WJ, Jensen DE, et al.KAP-1, a 22. Batlle E, Sancho E, Franci C, et al.The transcrip- the KRAB-KAP-1 repressor complex: a model system for novel corepressor for the highly conserved KRAB tion factor snail is a repressor of E-cadherin gene defining the molecular anatomy of RING-B box-coiled- repression domain.Genes Dev 1996;10:2067–78. expression in epithelial tumour cells.Nat Cell Biol coil domain-mediated protein-protein interactions.J 7. Hemavathy K, Guru SC, Harris J, Chen JD, Ip YT. 2000;2:84–9. Mol Biol 2000;295:1139–62. Human Slug is a repressor that localizes to sites of 23. Goyal RK, Lin P, Kanungo J, Payne AS, Muslin AJ, 36. Peng H, Begg GE, Harper SL, Friedman JR, Speicher active transcription.Mol Cell Biol 2000;20:5087–95. Longmore GD.Ajuba, a novel LIM protein, interacts DW, Rauscher FJ III.Biochemical analysis of the 8. Nakakura EK, Watkins DN, Schuebel KE, et al. with Grb2, augments mitogen-activated protein kinase Kruppel-associated box (KRAB) transcriptional repres- Mammalian Scratch: a neural-specific Snail family activity in fibroblasts, and promotes meiotic maturation sion domain.J Biol Chem 2000;275:18000–10. transcriptional repressor.Proc Natl Acad Sci U S A of Xenopus oocytes in a Grb2- and Ras-dependent 37. Peng H, Gibson LC, Capili AD, et al.The structurally 2001;98:4010–5. manner.Mol Cell Biol 1999;19:4379–89. disordered KRAB repression domain is incorporated 9. Nibu Y, Zhang H, Bajor E, Barolo S, Small S, Levine M. 24. Kanungo J, Pratt SJ, Marie H, Longmore GD.Ajuba, a into a protease resistant core upon binding to KAP-1- dCtBP mediates transcriptional repression by Knirps, cytosolic LIM protein, shuttles into the nucleus and RBCC domain.J Mol Biol 2007;370:269–89. Kruppel and Snail in the Drosophila embryo.EMBO J affects embryonal cell proliferation and fate decisions. 38. Li X, Peng H, Schultz DC, Lopez-Guisa JM, Rauscher 1998;17:7009–20. Mol Biol Cell 2000;11:3299–313. FJ III, Marmorstein R.Structure-function studies of the 10. Grimes HL, Chan TO, Zweidler-McKay PA, Tong B, 25. Kiss H, Kedra D, Yang Y, et al.A novel gene BTB/POZ transcriptional repression domain from the Tsichlis PN.The Gfi-1 proto-oncoprotein contains a containing LIM domains (LIMD1) is located within the promyelocytic leukemia zinc finger oncoprotein.Cancer novel transcriptional repressor domain, SNAG, inhibits common eliminated region 1 (C3CER1) in 3p21.3. Hum Res 1999;59:5275–82. G1 arrest induced by interleukin-2 withdrawal.Mol Cell Genet 1999;105:552–9. 39. Wadman IA, Osada H, Grutz GG, et al.The LIM-only Biol 1996;16:6263–72. 26. Ayyanathan K, Lechner MS, Bell P, et al.Regulated protein Lmo2 is a bridging molecule assembling an 11. Gilks CB, Bear SE, Grimes HL, Tsichlis PN.Progres- recruitment of HP1 to a euchromatic gene induces erythroid, DNA-binding complex which includes the sion of interleukin-2 (IL-2)-dependent rat T cell mitotically heritable, epigenetic gene silencing: a TAL1, E47, GATA-1 and Ldb1/NLI proteins.EMBO J lymphoma lines to IL-2-independent growth following mammalian cell culture model of gene variegation. 1997;16:3145–57. activation of a gene (Gfi-1) encoding a novel zinc finger Genes Dev 2003;17:1855–69. 40. Visvader JE, Mao X, Fujiwara Y, Hahm K, Orkin protein.Mol Cell Biol 1993;13:1759–68. 27. Ayyanathan K, Fredericks WJ, Berking C, et al. SH.The LIM-domain binding protein Ldb1 and its 12. Zornig M, Schmidt T, Karsunky H, Grzeschiczek A, Hormone-dependent tumor regression in vivo by an partner LMO2 act as negative regulators of erythroid Moroy T.Zinc finger protein GFI-1 cooperates with myc inducible transcriptional repressor directed at the PAX3- differentiation.Proc Natl Acad Sci U S A 1997;94: and pim-1 in T-cell lymphomagenesis by reducing the FKHR oncogene.Cancer Res 2000;60:5803–14. 13707–12. requirements for IL-2.Oncogene 1996;12:1789–801. 28. Ryan RF, Schultz DC, Ayyanathan K, et al.KAP-1 41. McLoughlin P, Ehler E, Carlile G, Licht JD, Schafer 13. Zweidler-Mckay PA, Grimes HL, Flubacher MM, corepressor protein interacts and colocalizes with BW.The LIM-only protein DRAL/FHL2 interacts with Tsichlis PN.Gfi-1 encodes a nuclear zinc finger protein heterochromatic and euchromatic HP1 proteins: a and is a corepressor for the promyelocytic leukemia zinc that binds DNA and functions as a transcriptional potential role for Kruppel-associated box-zinc finger finger protein.J Biol Chem 2002;277:37045–53. repressor.Mol Cell Biol 1996;16:4024–34. proteins in heterochromatin-mediated gene silencing. 42. Morlon A, Sassone-Corsi P.The LIM-only protein 14. Grimes HL, Gilks CB, Chan TO, Porter S, Tsichlis PN. Mol Cell Biol 1999;19:4366–78. FHL2 is a serum-inducible transcriptional coactivator of The Gfi-1 protooncoprotein represses Bax expression 29. Schultz DC, Ayyanathan K, Negorev D, Maul GG, AP-1.Proc Natl Acad Sci U S A 2003;100:3977–82. and inhibits T-cell death.Proc Natl Acad Sci U S A 1996; Rauscher FJ III.SETDB1: a novel KAP-1-associated 43. Sugihara TM, Bach I, Kioussi C, Rosenfeld MG, 93:14569–73. histone H3, lysine 9-specific methyltransferase that Andersen B.Mouse deformed epidermal autoregula- 15. Karsunky H, Zeng H, Schmidt T, et al.Inflamma- contributes to HP1-mediated silencing of euchromatic tory factor 1 recruits a LIM domain factor, LMO-4, tory reactions and severe neutropenia in mice lacking genes by KRAB zinc-finger proteins.Genes Dev 2002;16: and CLIM coregulators.Proc Natl Acad Sci U S A the transcriptional repressor Gfi1.Nat Genet 2002;30: 919–32. 1998;95:15418–23. 295–300. 30. Fredericks WJ, Ayyanathan K, Herlyn M, Friedman 44. Howard PW, Maurer RA.A point mutation in the LIM www.aacrjournals.org 9105 Cancer Res 2007; 67: (19). October 1, 2007

domain of Lhx3 reduces activation of the glycoprotein 48. Taniguchi Y, Furukawa T, Tun T, Han H, Honjo T.LIM papillomavirus type 6 results in its nuclear transloca- hormone a-subunit promoter.J Biol Chem 2001;276: protein KyoT2 negatively regulates transcription by tion.J Virol 2001;75:11791–802. 19020–6. association with the RBP-J DNA-binding protein.Mol 51. Petit MM, Fradelizi J, Golsteyn RM, et al.LPP, an 45. Chang DF, Belaguli NS, Iyer D, et al.Cysteine-rich Cell Biol 1998;18:644–54. actin cytoskeleton protein related to zyxin, harbors a LIM-only proteins CRP1 and CRP2 are potent smooth 49. Brown S, McGrath MJ, Ooms LM, Gurung R, nuclear export signal and transcriptional activation muscle differentiation cofactors.Dev Cell 2003;4:107–18. Maimone MM, Mitchell CA.Characterization of two capacity.Mol Biol Cell 2000;11:117–29. 46. Muller JM, Metzger E, Greschik H, et al.The isoforms of the skeletal muscle LIM protein 1, SLIM1. 52. Peinado H, Ballestar E, Esteller M, Cano A.Snail transcriptional coactivator FHL2 transmits Rho signals Localization of SLIM1 at focal adhesions and the mediates E-cadherin repression by the recruitment of from the cell membrane into the nucleus.EMBO J 2002; isoform slimmer in the nucleus of myoblasts and the Sin3A/histone deacetylase 1 (HDAC1)/HDAC2 21:736–48. cytoplasm of myotubes suggests distinct roles in the complex.Mol Cell Biol 2004;24:306–19. 47. Ng EK, Chan KK, Wong CH, et al.Interaction of the cytoskeleton and in nuclear-cytoplasmic communica- 53. Bach I, Rodriguez-Esteban C, Carriere C, et al.RLIM heart-specific LIM domain protein, FHL2, with DNA- tion.J Biol Chem 1999;274:27083–91. inhibits functional activity of LIM homeodomain binding nuclear protein, hNP220.J Cell Biochem 2002; 50. Degenhardt YY, Silverstein S.Interaction of zyxin, a transcription factors via recruitment of the histone 84:556–66. focal adhesion protein, with the e6 protein from human deacetylase complex.Nat Genet 1999;22:394–9.

Cancer Res 2007; 67: (19). October 1, 2007 9106 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 2007 American Association for Cancer Research. The Ajuba LIM Domain Protein Is a Corepressor for SNAG Domain−Mediated Repression and Participates in Nucleocytoplasmic Shuttling

Kasirajan Ayyanathan, Hongzhuang Peng, Zhaoyuan Hou, et al.

Cancer Res 2007;67:9097-9106.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/67/19/9097

Cited articles This article cites 53 articles, 35 of which you can access for free at: http://cancerres.aacrjournals.org/content/67/19/9097.full#ref-list-1

Citing articles This article has been cited by 17 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/67/19/9097.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://cancerres.aacrjournals.org/content/67/19/9097. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.