The Ikaros family protein Eos associates with C-terminal-binding protein corepressors Jose ´ Perdomo and Merlin Crossley Department of Biochemistry, G08, University of Sydney, NSW, Australia Eos is a zinc finger transcription factor of the Ikaros family. It binds typical GGGAA Ikaros recognition sites in DNA and functions as a transcriptional repressor. Here we show that Eos associates with the corepressor C-terminal-binding protein (CtBP). CtBP has previously been shown to bind Pro-X-Asp-Leu-Ser (PXDLS) motifs in several DNA- binding proteins. We note that Eos contains a related motif PEDLA, and we demonstrate that CtBP can bind this site weakly but that it also contacts additional regions of Eos. Consistent with this finding, mutation of the PEDLA motif does not negate CtBP binding or CtBP-mediated repression by Eos. CtBP has previously been shown to bind to a PXDLS-type motif in Ikaros, and we show that another Ikaros-related protein TRPS1 also contains a PXDLS CtBP contact motif within its repression domain. We conclude that several Ikaros family proteins utilize CtBP corepressors to inhibit gene expression. Keywords: corepressors; gene regulation; Ikaros; repression; transcription. The zinc finger transcription factor Ikaros was originally identified as a DNA-binding protein that recognized a critical regulatory region of the T cell-restricted CD3d gene [1]. Ikaros expression is confined to erythroid and myeloid precursors in the early stages of differentiation and to the lymphoid compartment in the adult [2,3]. The Ikaros gene codes for a protein with six zinc fingers that comply with the Kru ¨ ppel C2-H2 consensus. The N-terminal fingers are involved in sequence-specific DNA binding [4], while the two zinc fingers that form the C-terminal domain mediate homodimerization [5]. Alternative mRNA splicing gener- ates at least eight isoforms (Ik-1 to Ik-8) containing subsets of the N-terminal fingers and all sharing the C-terminal domain. Isoforms containing at least three N-terminal fingers are able to bind to the Ikaros consensus recognition sequence [4]. The subsequent cloning of Aiolos and Helios [6–8], and of Eos and Pegasus [9], revealed the existence of a family of related factors. Ikaros, Aiolos and Helios are all abundantly expressed in the haematopoietic system and are all known or predicted to have roles in lymphoid development, whereas Eos and Pegasus are more broadly expressed, as mRNA is detected in several human tissues [9]. Recently a more distantly related member of the Ikaros family, the tricho-rhino-phalangeal syndrome protein TRPS1, has been described [10,11]. This protein contains the characteristic C-terminal domain consisting of two zinc fingers capable of mediating dimerization, but also contains additional Kru ¨ ppel-like zinc fingers and one GATA-type finger. Figure 1 shows schematic representations of Ikaros, Eos and TRPS1. Studies of murine knockouts have revealed that Ikaros is essential for the regulation of commitment of haematopoi- etic stem cells to the lymphoid lineage. In Ikaros null mice (Ik –/– ), B cells and their precursors are absent, and T cells are undetected in the fetus but develop (abnormally) post partum [12]. Mice expressing an Ikaros protein that lacks the DNA-binding domain (dominant negative DN –/– mutation) display more extreme effects, with a complete absence of T cells and death from severe infections soon after birth [2]. The severity of the DN –/– mutation suggests that this aberrant Ikaros protein, which cannot bind DNA, is still able to dimerize with other Ikaros family proteins and, most likely, interfere with their functions. Aiolos –/– mice show expanded B cell populations and autoimmunity, but are normal in their thymic and splenic T cell develop- ment [13], a phenotype consistent with the predominant expression of Aiolos in B cells. No knockouts have been reported for the other family members. The molecular mechanisms by which members of the Ikaros family recognize DNA and regulate gene expression are under intense investigation [14–18]. Ikaros, Aiolos, Helios and Eos all recognize the consensus Ikaros-binding site GGGAA in vitro and in cellular assays, whereas Pegasus recognizes a distinct binding sequence GNNTGTNG [9]. TRPS1, by virtue of its GATA-type zinc finger, can recognize GATA sites in DNA but may also bind to additional elements through its Kru ¨ ppel-like fingers. In transient assays, all these proteins are able to modestly influence the transcription of reporter genes driven by their cognate sites [4,8,9]. Aiolos and Helios have been reported to function as activators, Ikaros has been implicated in both activation and repression, and Eos, Pegasus and TRPS1 have so far only been implicated in transcriptional repres- sion. Recently attention has focused on the role of Ikaros as a repressor, and interactions have been reported with Sin3 [19] and Mi-2 [20], which are components of deacetylase and chromatin remodelling complexes. Correspondence to M. Crossley, Department of Biochemistry, G08, University of Sydney, NSW, Australia, 2006. Fax: 61 29351 4726, Tel.: 61 29351 2233, E-mail: M.Crossley@biochem.usyd.edu.au Abbreviations: TRPS1, tricho-rhino-phalangeal syndrome protein; CtBP, C-terminal-binding protein; GST, glutathione S-transferase. (Received 8 August 2002, revised 19 September 2002, accepted 15 October 2002) Eur. J. Biochem. 269, 5885–5892 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03313.x Ikaros has also been shown to bind coregulatory proteins of the C-terminal-binding protein (CtBP) family (reviewed in [21]). CtBP was named after it was first purified as a protein that bound to the C-terminus of the Adenovirus E1A protein [22]. CtBPs have been identified in Caenor- habditis, Drosophila, and mammals, and data from Drosophila and mammals have shown that CtBPs can function as transcriptional corepressors in vivo.CtBP recognizes PXDLS motifs found in the repression domains of a wide range of transcription factors. Identified partners include Drosophila Hairy, Snail and Kru ¨ ppel [23,24], BKLF/KLF3 [25], FOG [26], KLF8 [27] and Ikaros [15]. There are two highly homologous mammalian CtBP family members, CtBP1 and CtBP2, encoded by separate genes, but, to date, no differences in the activity of these two proteins have been reported. In this study, we used murine CtBP2 [25], but it is likely that the results will also apply to CtBP1, and therefore, in some instances, we use the term CtBP for simplicity. CtBP has been shown to associate with Ikaros through a PEDLS motif in its N-terminus (Fig. 1A,B) [15]. As Ikaros makes multiple contacts with coregulatory proteins, experiments with full-length Ikaros can be difficult to interpret but, by studying the N-terminal repression domain in isolation, Koipally & Georgopoulos [15] were able to show that CtBP contact was required for the repression activity of this domain. This repression activity is histone deacetylase independent, and the precise mech- anism of repression remains unknown. Here, we show that Eos also interacts with CtBP to repress transcription. CtBP recognizes a PEDLA motif in the C-terminus of Eos, but the interaction does not completely depend on this motif, suggesting that the Eos–CtBP interaction involves multiple surfaces. Consistent with this result, we demonstrate that CtBP recognizes several regions within Eos. In addition, we show that CtBP can bind a PXDLS motif in a previously well-characterized C-terminal repres- sion domain in the TRPS1 protein [11], and we suggest that a CtBP-mediated mechanism may be common in Ikaros-like proteins. EXPERIMENTAL PROCEDURES Plasmids Bait plasmids used in the yeast two-hybrid experiments were generated by fusing the desired regions downstream and in-frame of the Gal4 DNA-binding domain of the yeast expression vector pGBT9 (Clontech). All numbers indicated refer to amino acids in the respective sequences. The bait plasmids were pGBT9.Eos364–400, pGBT9.Eos364– 400.mut, pGBT9.Eos364–400DEtoIk (Eos 372 PED- LADGG 379 changed to Ikaros site PEDLSTTS), pGBT9.TRPS1210-1281, and pGBT9.mCtBP2 [25]. Prey plasmids were constructed by inserting the desired sequences in-frame and downstream of the Gal4 activation domain of the pGAD10 vector (Clontech). Full-length vectors are pGAD10.Eos and pGAD10.mCtBP2, pGAD10.Pegasus and pGAD10.Ikaros2. Other plasmids are pGA- D10.Eos101–531, pGAD10.Eos101–531.mut, pGAD10. Eos101–364, pGAD10.Eos101–331, pGAD10.Eos101– 231, pGAD10.Eos364–518, pGAD10.Eos364–518.mut, pGAD10.Ikaros1–81, pGAD10.Ikaros1–81.mut, pGA D10.TRPS1068–1186, pGAD10.TRPS1068–1186.mut, pGAD10.TRPS1210–1281 and pGAD10.Aiolos447–507. Expression vectors for mammalian systems used were constructed in the parental plasmids pcDNA3 (Invitrogen) and pMT2. The vectors include pcDNA3. Gal4DBD, pcDNA3.mCtBP2 [25], pcDNA3.Eos, pcDNA3.Gal4- DBD.Eos364–518, pcDNA3.Gal4DBD.Eos364–518.mut, pcDNA3.Gal4DBD.Ikaros1–81, pcDNA3.Gal4DBD Ikaros1–81.mut, pcDNA3.Gal4DBD.TRPS1068–1281, pcDNA3.Gal4DBD.TRPS1068–1281.mut, pMT2.mCTBP2, pMT2.FLAG.Eos, pMT2.FLAG.Eos.mut, pMT2.FLA- G.Eos101–389, pMT2.FLAG.Eos101–331, and pMT2. FLAG.Eos101–231. Glutathione S-transferase (GST) fu- sion mCtBP2 protein was produced by inserting mCtBP2 cDNA in-frame with GST in the pGEX-2T vector (Amer- sham Pharmacia Biotech) [25]. The luciferase reporter vector contained five copies of the Gal4 DNA-binding domain upstream of the TK promoter in the vector pGL2- TK (Promega). Yeast two-hybrid and pull-down assays The Clontech yeast two-hybrid system was used according to the manufacturer’s instructions. The prey and bait plasmids used are described above. Recombinant GST- mCtBP2 was produced in Escherichia coli strain BL-21, purified as described [28] and immobilized on glutathione beads. 35 S-labelled Eos and Eos.mut production and the pull-down experiments were carried out as described [9]. Ikaros MPVPEDLSTTS 31 41 Eos GEGPEDLADGG 369 379 TRPS1 NDIPLDLAIKH 1160 1170 A B Ikaros Eos TRPS1 DNA binding Dimerization 517 532 1281 Dimerization GATA finger PXDLS motif Intron/Exon boundary Fig. 1. Schematic representation of Ikaros, Eos and TRPS1. (A) Diagram of Ikaros, Eos and TRPS1. The zinc fingers are represented by unfilled semi-ellipses, the positions of the PXDLS-like motif are indicated by filled rectangles, and the intron/exon boundaries by arrows. Amino-acid numbers are indicated in each case. For Ikaros and Eos, the N-terminal fingers involved in sequence-specific DNA recognition are shown, and the GATA-type finger required for binding of TRPS1 to GATA sites is shown. The C-terminal fingers that are characteristic of the Ikaros family and mediate dimerization are also indicated. (B) The sequences corresponding to the PXDLS- like motif for the three proteins in (A) are shown. Amino-acid numbers are indicated. 5886 J. Perdomo and M. Crossley (Eur. J. Biochem. 269) Ó FEBS 2002 In vitro transcription and translation In vitro transcription and translation of proteins has been described [9]. Western blot and immunoprecipitation Transfected COS cells were washed with cold NaCl/P i and resuspended in 400 lL cold solution A (10 m M Hepes, pH 7.8, 1.5 m M MgCl 2 ,10m M KCl) supplemented before use with 1 m M dithiothreitol, 50 ngÆmL )1 phenyl- methanesulfonyl fluoride, 5 lgÆmL )1 leupeptin and 5 lgÆmL )1 aprotinin. The tubes were incubated on ice for 10 min, vortex-mixed for 10 s, and centrifuged for 10 s at 12 000 g to pellet the nuclei. The nuclei were resuspended in 30–50 lL solution C (20 m M Hepes, pH 7.8, 25% glycerol, 420 m M NaCl, 1.5 m M MgCl 2 ,0.2m M EDTA) supplemen- ted as above, centrifuged for 3 min at 14 000 r.p.m. at 4 °C. The extracts were used immediately or stored at )70 °C. Proteins were separated by SDS/PAGE on 8–10% polyacrylamide gels and transferred on to a Biotrace TM nitrocellulose blotting membrane (Pall Gelman Sciences, Ann Arbor, MI, USA) in a TE series Transphor TM electrophoresis unit (Hoefer), at 50 mA overnight at 4 °C. For Western blotting, the membrane was washed once in 50 m M Tris/HCl, pH 7.5, containing 150 m M NaCl and 0.05% Tween-20 (Tris/NaCl/Tween), then incubated at room temperature in skimmed milk powder solution [5% (w/v) in Tris/NaCl/Tween] for 1 h. The membrane was rinsed in Tris/NaCl/Tween and incubated for 1 h with gentle shaking in 10 mL Tris/NaCl/Tween containing 10 lg primary antibody. After a wash with 4 · 100 mL Tris/ NaCl/Tween, the secondary antibody solution was added and incubation was continued for 1 h. The membrane was washed for 1 h in several changes of Tris/NaCl/Tween. Detection was carried out using the RenaissanceÒ Chemi- luminescence reagent plus (NEN Life Sciences, Boston, MA, USA), and the signal detected on X-ray film (Eastman Kodak Company, Rochester, NY, USA) and developed using Kodak reagents. Covalently linked protein A/G-agarose beads (Boehrin- ger, Mannheim, Germany) and the antibody of interest were prepared as follows. The beads plus antibody were incubated for 1 h at room temperature in 1 mL NaCl/P i at 2 lg antibody/lL wet beads. The bead–antibody complex was washed twice with 10 vol. 0.2 M sodium borate, pH 9.0, and the beads were resuspended in 10 vol. 0.2 M sodium borate, pH 9.0. Solid dimethyl pimelimidate (Sigma) was added to a final concentration of 20 m M , incubated for 30 min at room temperature, and the reaction stopped by washing once in 10 vol. 0.2 M ethanolamine, pH 8.0, and then incubating for 2 h at room temperature in 10 vol. 0.2 M ethanolamine, pH 8.0. Coupled beads were resus- pended in 1 vol. NaCl/P i andstoredat4°C. For immunoprecipitations, nuclear extracts were diluted 1 : 3 in Nonidet P40 buffer (50 m M Tris/HCl, pH 7.4, 150 m M NaCl, 0.5–1.0% Nonidet P40, 1 lgÆmL )1 leupep- tin, 1 lgÆmL )1 aprotinin, 1 m M phenylmethanesulfonyl fluoride). Lysates were precleared with 20 lLproteinA/G beads for 30 min at 4 °C. The cleared lysates were treated with 5–15 lL beads–antibody complex for 1 h at 4 °C with rocking. The beads were pelleted at 14 000 r.p.m. for 10 s, the supernatant discarded, and the beads washed (4 · 1mL cold Nonidet P40 buffer). The proteins retained on the beads were separated by SDS/PAGE and detected by Western blotting. Transfections and luciferase assay NIH-3T3 cells were transfected with 3 lg of the reporter pGL2(Gal4) 5 TK and different amounts (0.5–2 lg) of pcDNA3.Gal4DBD, pcDNA3.mCtBP2, pcDNA3.Gal4 DBD.Eos364–518, pcDNA3.Gal4DBD.Eos364–518.mut, pcDNA3.Gal4DBD.Ikaros1–81, pcDNA3.Gal4DBD Ikaros1–81.mut, pcDNA3.Gal4DBD.TRPS1068–1281 and pcDNA3.Gal4DBD.TRPS1068–1281.mut using the calcium phosphate method [29]. Luciferase activity was measured as described [9]. COS cells were transfected with 2 lg pMT2, pMT2.mCTBP2, pMT2.FLAG.Eos, pMT2. FLAG.Eos.mut, pMT2.FLAG.Eos101–389, pMT2.FLAG. Eos101–331 and pMT2.FLAG.Eos101–231 by the DAE- Dextran method [29], and harvested 48–60 h after transfec- tion. The total amount of transfected DNA was kept constant in all cases by addition of ÔnakedÕ pcDNA3 or pMT2 vectors, as appropriate. RESULTS Eos interacts with CtBP in vitro and in vivo The observation that Eos can function to repress gene expression [9] prompted us to investigate whether it associated with recognized corepressor proteins. We noted that the C-terminal region of Eos contains a sequence, 372 PEDLA 376 , that resembles the accepted consensus CtBP- binding motif PXDLS [21]. The Eos motif differs from the consensus at the final residue, but a previously identified partner, Enhancer of Split md, also has alanine at the fifth position, suggesting that this change would not preclude CtBP binding [23]. We first tested the ability of in vitro transcribed and translated Eos protein to interact with bacterially expressed and purified GST–CtBP. Figure 2A shows that GST–CtBP but not GST alone is able to retain 35 S-Eos. The Eos–CtBP contact was also confirmed using the yeast two-hybrid system. We cotransformed yeast with vectors encoding a Gal4 activation domain–Eos fusion, and a Gal4 DNA-binding domain–CtBP fusion and observed activation of the HIS3 reporter gene as indicated by yeast growth in the absence of histidine (Fig. 2B). Finally, we assessed the interaction using coimmunopre- cipitation experiments. COS cells were transfected with vectors expressing FLAG-tagged Eos and native CtBP. FLAG-Eos was recovered by immunoprecipitation with a FLAG antibody, and the presence of CtBP assessed by Western blotting using anti-CtBP serum. As shown in Fig. 2C, CtBP was present in the recovered material. Taken together these results indicate that Eos and CtBP physically interact. The PEDLA motif of Eos is not the sole determinant of the CtBP interaction In most instances, such as the case of Ikaros [15], deletion or mutation of a single critical PXDLS motif results in abrogation of CtBP contact and functional consequences of CtBP association. We examined whether the binding of Ó FEBS 2002 Eos binds the corepressor CtBP (Eur. J. Biochem. 269) 5887 CtBP to Eos required the PEDLA motif by mutating this motif to AAALA (Eos-mut). Interestingly, we found that this mutation did not eliminate the Eos–CtBP interaction. Figure 2D shows that GST–CtBP is also able to retain radiolabelled Eos-mut. Figure 2E shows that yeast har- bouring an expression vector encoding a Gal4 activation domain–Eos-mut fusion and a Gal4 DNA-binding domain–CtBP fusion grow in the absence of histidine, and Fig. 2F indicates that CtBP can be immunoprecipitated with FLAG-tagged Eos-mut. These results suggest that CtBP does not depend exclusively on the Eos PEDLA motif for interaction and may make additional contacts through other domains within the Eos protein. There are several precedents for CtBPs contacting partners through regions outside recognizable PXDLS motifs [30–32]. To delineate additional Eos domains involved in CtBP recruitment, a series of deletion mutants was constructed (Fig. 3A). The CtBP-interacting properties of these mutants were tested using the yeast two-hybrid system. As seen in Fig. 3A, CtBP was able to associate with both the N-terminal and C-terminal domains of Eos. To determine whether the 372 PEDLA 376 motif in the C-terminus was primarily responsible for binding to this domain, we again mutated this motif to AAALA, but this time in the context of the C-terminal domain Eos364–518. Again mutation of the motif did not significantly affect CtBP binding. We carried out the same experiment in the context of a minimal Eos domain, Eos364–400 and found that this region still bound CtBP. In this construct, however, the PEDLA to AAALA mutation reduced binding. We also made a second mutant replacing the nontypical PEDLA motif of Eos with the recognized PEDLS motif of Ikaros and observed a stronger interaction (Fig. 3A). Taken together these results suggest that the PEDLA motif in Eos is suboptimal and not the major determinant of CtBP binding, but that it and other sites within the N-terminus and C-terminus of Eos contribute to CtBP contact. We also confirmed the presence of the N-terminal CtBP-binding domain using coimmunoprecipitation experiments. The FLAG-tagged Eos constructs shown in Fig. 3B were cotransfected with CtBP into COS cells, and immunoprecipitation experiments were carried out. Figure 3C is a Western blot showing that all three Eos constructs are expressed at comparable levels. Figure 3D shows an immunoprecipitation experiment with FLAG antisera or an irrelevant antibody in the mock lane. Western blotting with a CtBP antiserum shows that CtBP is associated with the immunoprecipitated material in all cases in which the FLAG antiserum was used, although most CtBP was retained by the longest construct. The results were confirmed by the reciprocal experiment, immunoprecipitating with CtBP antiserum and analyzing the material by Western blotting with anti-FLAG serum (Fig. 3E). As can be seen, the three Eos deletion constructs are detected in the material immunoprecipitated by the anti-CtBP serum, and again the longest construct appears to have been retained more efficiently. No Eos fragments were detected when an irrelevant antibody was used in a similar experiment (data not shown). These findings confirm the observation that, in addition to binding the Fig. 2. Eos interacts with CtBP. (A) Purified GST and GST–CtPB were used to assess the interacting activities of in vitro transcribed and translated Eos protein. GST–CtBP but not GST alone retained radiolabelled Eos protein. (B) The interaction was confirmed using the yeast two-hybrid system. Plasmids present in the various yeast derivatives are shown. Growth on this plate lacking histidine, leucine and tryptophan is indicative of a positive interaction. (C) FLAG-tagged Eos and CtBP werecotransfectedintoCOScellsandnuclear extracts used for immunoprecipitations. Lane 1 (input) indicates the migration of CtBP in the extracts, control mock transfected cells (lane 2) and detected (lane 3) CtBP after immunoprecipitation with anti-FLAG serum. IP, Immunoprecipitation. (D) Purified GST and GST–CtPB were used to assess the inter- acting activities of in vitro transcribed and translated Eos.mut protein. GST–CtBP retained radiolabelled mutant os protein. (E) and (F), as for (B) and (C). Mutant Eos was tested in both cases. IP, Immunoprecipitation. 5888 J. Perdomo and M. Crossley (Eur. J. Biochem. 269) Ó FEBS 2002 C-terminal domain of Eos, CtBP can also bind sites within the N-terminal domain of Eos. TRPS1 interacts with CtBP The TRPS1 protein is a multi-(zinc finger) protein that contains two C-terminal fingers highly related to the Ikaros family dimerization domain [10,11]. Little is known of the molecular roles of TRPS1 or its target genes. It is known that TRPS1 is capable of binding typical GATA sites via its GATA-type zinc finger and that TRPS1 can act to repress the expression of GATA-dependent reporter genes [11]. The relevant repression domain has been localized to the C-terminal 119 residues of the protein [11]. We noted that this minimal repression domain contains a potential CtBP contact motif, 1163 PLDLA 1167 . This observation implied that the repression domain might function by recruiting the corepressor CtBP. As the yeast two-hybrid system has proved a very reliable indicator of CtBP contact in all instances previously reported [21], we used this assay to determine whether the repression domain of TRPS1 was capable of interacting with CtBP. We found that there was a strong interaction in yeast (Fig. 4A,B). The amount of yeast growth was comparable to that observed for the isolated CtBP-binding region of Ikaros (Fig. 4A,B, construct 1). Mutation of the putative CtBP-binding motif 1163 PLDLA 1167 to ALAAA abolished the interaction (Fig. 4A, construct 4), suggesting that it was the major determinant of CtBP binding within the repression domain. CtBP-interacting regions of Eos, Ikaros and TRPS1 function as CtBP-dependent repression domains Deletion analysis of Ikaros has indicated that it contains distinct domains that are implicated in activating or repressing transcription [4,5,19,33]. One discrete domain within the N-terminus contains the motif PEDLS and has been shown to contact CtBP and function as a CtBP- dependent repression domain. We investigated whether the PXDLS regions of Eos and TRPS1 also functioned as CtBP-dependent repression domains. The regions were tested as Gal4 DNA-binding domain fusions for their ability to repress transcription in transient transfection experiments in mammalian cells. Gal4Ikaros1–81, Gal4Eos364–518, and Gal4TRPS1068–1281 were transfect- ed individually into NIH-3T3 cells and tested against a luciferase reporter gene driven by five Gal4-binding sites upstream of the TK promoter. Figure 5A shows that Gal4Ikaros1–81 represses the transcription of the reporter gene. A mutation in the PEDLS motif abolished the ability of this domain to repress transcription, consistent with previous findings that these residues are required for CtBP recruitment and repression. We also observed that when submaximal amounts of the Ikaros construct (0.1 lg) were Fig. 3. Eos deletion constructs interact with CtBP in yeast and in COS cells. (A) Schematic representation of Eos constructs tested against CtBP in the yeast two-hybrid system. Numbers indicate the amino acids in the Eos sequence, the filled rectangle indicates the position of the PEDLA motif, and the cross represents mutation of this motif. (Rectangle) Ik indicates mutation to resemble the Ikaros motif PED- LSTT.(+)growthobserved;(–)nogrowth.Forcomparison,the CtBP-interacting region of Ikaros (Ikaros1–81) is also shown. Yeast growth on plates lacking histidine is shown for selected constructs. These plates represent growth after 4 days of 10 lL of undiluted, 1 : 10 and 1 : 100 dilutions of D 600 solutions. (B) Schematic repre- sentation of FLAG-tagged Eos constructs used for immunoprecipi- tations. Numbers indicate the amino acids in the Eos sequence. (C) Western blot of the Eos constructs in (B) expressed in COS cells. (D) Immunoprecipitation showing that the three Eos N-terminal con- structs are able to associate with cotransfected CtBP, lanes 1–3. (E) The reciprocal experiment with anti-CtBP serum precipitating the cotransfected Eos constructs (arrows) shown in (B) and (C). The prominent bands seen above the bands of interest (arrowheads) cor- respond to the heavy chain of the antibodies used because of incom- plete covalent coupling of the antibodies to the agarose beads. IP, Immunoprecipitation. Fig. 4. TRPS1 and Ikaros interact with CtBP through their PXDLS- like motifs. (A) Schematic representation of the fragments tested. Numbers indicate the amino acids of the respective protein. Filled rectangles indicate the position of the PXDLS-like motif, and the cross represents mutation of this motif. (B) Yeast growth in the absence of histidine is indicative of a positive interaction. Ó FEBS 2002 Eos binds the corepressor CtBP (Eur. J. Biochem. 269) 5889 used, cotransfection with a CtBP expression vector (0.5 lg) potentiated repression. These results corroborate previous observations on CtBP-dependent repression by this Ikaros domain [15]. Figure 5B shows a similar experiment on the PEDLA motif-containing domain of Eos. This domain also functions to repress the reporter, but mutation of the PEDLA motif does not abrogate repression. This result is consistent with the protein interaction data showing that this mutation does not prevent contact with CtBP. Again, when low amounts of the Eos construct (0.1 lg) were used, addition of a CtBP expression vector (0.5 lg) potentiated repression, confirming the result that Eos364–518 depends on CtBP, although the PEDLA motif is not essential for its recruitment. Finally, Fig. 5C shows results with the TRPS1 PLDLA-containing domain. This portion of the protein functioned as a potent repression domain and mutation of the PLDLA motif abrogated repression activity, consistent with the protein interaction result that this motif was required for CtBP contact. Again addition of a CtBP expression vector potentiated repression when low amounts of the TRPS1 construct were tested. Overall these results indicate that these three Ikaros family proteins all contain repression domains that are dependent on CtBP. DISCUSSION CtBP has previously been shown to bind to repression domains in a number of transcription factors and other regulatory proteins [21]. The results reported here show that CtBP binds repression domains within three mem- bers of the Ikaros family of transcription factors. In each case, the domains contain a recognizable PXDLS-type motif but the PEDLA motif in Eos is not the sole determinant of CtBP contact. Furthermore the PEDLA motif in Eos appears to be a relatively weak binding site, as its replacement with the well-characterized PEDLS motif of Ikaros substantially increased the association with CtBP. Although the PEDLA motif in Eos supports only weak binding of CtBP, we show that additional regions within the N-terminal and C-terminal regions of Eos also contact CtBP, and, taken together, the results of GST-pulldown, yeast two-hybrid and immunoprecipita- tion experiments demonstrate that Eos and CtBP can stably associate. Although PXDLS motifs have often been shown to be the primary determinant of CtBP binding, there are other examples where they are not essential for corepressor contact, and there are even cases where CtBP partner proteins contain no recognizable PXDLS motif. The zinc finger protein Tramtrack69 contains a PPDLS motif, but this is not required for binding [31], and HDAC5 contains arelatedmotif(PVELR)thatisdispensableforCtBP contact [32]. In the case of HDAC1, no canonical CtBP recognition sequences have been found [30]. We presume that generally the PXDLS motif on the DNA-binding protein slots into a putative PXDLS-accepting pocket within CtBP. In the cases where the PXDLS motif is not required for binding and other contacts are made, we expect that other (nonpocket) regions of CtBP will be involved in the interaction. In this way, it seems likely that proteins such as Eos may be able to bind to CtBP that is already associated with another PXDLS motif-containing protein. For instance, Eos may bind to a CtBP molecule, the pocket of which is already complexed to Ikaros. Indeed, we have previously shown that the conserved C-terminal dimerization domain of Eos can associate with the related domain in Ikaros [9], so Eos may in fact be able to bind Ikaros and CtBP to form a trimeric complex. We have also tested the association of Eos with other Ikaros family members and shown that it can bind the dimeriza- tion domain of TRPS1 (unpublished results), raising the possibility that Eos may also function in a similar manner in a complex with TRPS1 and CtBP. There is good evidence that Ikaros proteins dimerize [5] and possibly form higher-order multimers in vivo [7,16], but the precise mechanisms by which various Ikaros-containing complexes operate remains under investigation. In addition to binding CtBP, Ikaros has been found in T cells as part of two discrete histone deacetylase complexes by virtue of its interaction with the ATPase Mi-2 [20] and the corepressor X 1 81 Eos Gal4DBD Ikaros Gal4DBD 0.5µg 0.1µg 1.0µg 0.1µg 1.0µg 0.5µg 0.1µg 1.0µg 1.0µg X 1068 1281 0.5µg 0.1µg 1.0 µg 0.1µg Gal4DBD TRPS1 A B C 0.1µg CtBP + - - + - 364 518 X CtBP + - - + - 0 5 10 15 0 5 10 15 20 01020304050 1.0 µg CtBP + - - + - Fold repression 1 81 Fold repression Fold repression 1 81 1 81 364 518 364 518 364 518 1068 1281 1068 1281 1068 1281 Fig. 5. CtBP binding domains of Ikaros, Eos and TRPS1 act as potent repression domains. Plasmids encoding Gal4 DNA-binding domain alone or fused to Eos364–518, Ikaros1–81 and TRPS1068–1281 were transfected into NIH-3T3 cells and the luciferase activity determined. The rectangle indicates the location of the PXDLS-like motif, and the cross indicates mutation of this motif. (A) Ikaros1–81 is a repressor only when its PEDLS motif is intact. Addition of a minimal amount does not affect transcription, but repression is seen on cotransfection of CtBP. (B) Eos364–518 represses transcription even when the PEDLA motif has been mutated. This Eos domain is responsive to cotrans- fected CtBP. (C) TRPS1068–1281 is a strong repressor the activity of which is abrogated by mutation of its PLDLA motif. Transfection of a minimal amount still represses transcription; this activity is potentiated on CtBP cotransfection. 5890 J. Perdomo and M. Crossley (Eur. J. Biochem. 269) Ó FEBS 2002 Sin3 [19]. It has also been implicated in silencing gene expression in B cells by targeting genes to inactive centro- meric chromatin [3,16,18]. Recently, Ikaros, Helios, Aiolos and murine Eos were shown to be able to interact with CtBP-interacting protein (CtIP) independently of CtBP association [34]. The Ikaros–CtIP complex was shown to be capable of repressing transcription in the absence of histone deacetylase activity and to perhaps function through a mechanism that depends on interactions with components of the basal transcriptional machinery such as TATA binding protein and transcription factor IIB [34]. Thus, the number of possible mechanisms employed by Ikaros complicates studies on the full-length protein, but experi- ments on isolated domains have established a role for the N- terminal CtBP contact region and shown that repression was trichostatin A independent [15]. This result suggests that the CtBP–Ikaros repression domain complex does not repress gene expression through a conventional HDAC mechanism. In our experiments we also found that the Eos– CtBP complex was not sensitive to trichostatin A, again consistent with a non-HDAC mechanism (unpublished results). The specific mechanisms by which other Ikaros family members influence gene expression are still under scrutiny, and very little is known about the overall biological roles of these proteins. Naturally occurring mutations in the TRPS1 gene lead to faciocranial abnor- malities and skeletal deformations [10], but precise target genes remain to be identified. The finding that CtBP associates with a functional repression domain in TRPS1 confirms former results suggesting that TRPS1 acts as a repressor protein and can counter GATA-mediated gene activation [11]. Figure 5C shows strong repression by TRPS1, which suggests that the proposed TRPS1–CtBP interaction is stronger than that of the other constructs investigated. Ultimately determination of the relevant association constants is likely to clarify this observation. Further work to identify target genes upregulated in the absence of functional TRPS1 may illuminate the molecular causes underlying the observed phenotype. ACKNOWLEDGEMENTS We are grateful to members of the laboratory and Margot Kearns for reading the manuscript. We thank Dr R.A. Shivdasani for his gift of a TRPS encoding plasmid. This work was supported by a grant from the Australian Health Management Group to M.C. J.P. is supported by an Australian Post-graduate Award. REFERENCES 1. Georgopoulos, K., Moore, D.D. & Derfler, B. (1992) Ikaros, an early lymphoid specific transcription factor and a putative med- iator for T cell commitment. Science 258, 808–812. 2. Georgopoulos, K., Bigby, M., Wang, J.H., Molnar, A., Wu, P., Winandy, S. & Sharpe, A. (1994) The Ikaros gene is required for the development of all lymphoid lineages. Cell 79, 143–156. 3. Klug,C.A.,Morrison,S.J.,Masek,M.,Hahm,K.,Smale,S.T.& Weissman, I.L. (1998) Hematopoietic stem cells and lymphoid progenitors express different Ikaros isoforms, and Ikaros is loca- lized to heterochromatin in immature lymphocytes. Proc. Natl. Acad.Sci.USA95, 657–662. 4. Molnar, A. & Georgopoulos, K. (1994) The Ikaros gene encodes a family of functionally diverse zinc finger DNA-binding proteins. Mol. Cell. Biol. 14, 8292–8303. 5. Sun,L.,Liu,A.&Georgopoulos,K.(1996)Zincfinger-mediated protein interactions modulate Ikaros activity, a molecular control of lymphocyte development. EMBO J. 15, 5358–5369. 6. Morgan, B., Sun, L., Avitahl, N., Andrikopoulos, K., Ikeda, T., Gonzales, E., Wu, P., Neben, S. & Georgopoulos, K. (1997) Aiolos, a lymphoid restricted transcription factor that interacts with Ikaros to regulate lymphocyte differentiation. EMBO J. 16, 2004–2013. 7. Hahm, K., Cobb, B.S., McCarty, A.S., Brown, K.E., Klug, C.A., Lee,R.,Akashi,K.,Weissman,I.L.,Fisher,A.G.&Smale,S.T. (1998) Helios, a T cell-restricted Ikaros family member that quantitatively associates with Ikaros at centromeric hetero- chromatin. Genes Dev. 12, 782–796. 8. Kelley, C.M., Ikeda, T., Koipally, J., Avitahl, N., Wu, L., Georgopoulos, K. & Morgan, B.A. (1998) Helios, a novel dimerization partner of Ikaros expressed in the earliest hemato- poietic progenitors. Curr. Biol. 8, 508–515. 9. Perdomo, J., Holmes, M., Chong, B. & Crossley, M. (2000) Eos and Pegasus, two members of the Ikaros family of proteins with distinct DNA binding activities. J. Biol. Chem. 275, 38347– 38354. 10. Momeni,P.,Glockner,G.,Schmidt,O.,vonHoltum,D.,Albr- echt, B., Gillessen-Kaesbach, G., Hennekam, R., Meinecke, P., Zabel, B., Rosenthal, A., Horsthemke, B. & Ludecke, H.J. (2000) Mutations in a new gene, encoding a zinc-finger protein, cause tricho-rhino–phalangeal syndrome type I. Nat. Genet. 24, 71–74. 11. Malik, T.H., Shoichet, S.A., Latham, P., Kroll, T.G., Peters, L.L. & Shivdasani, R.A. (2001) Transcriptional repression and devel- opmental functions of the atypical vertebrate GATA protein TRPS1. EMBO J. 20, 1715–1725. 12. Wang, J.H., Nichogiannopoulou, A., Wu, L., Sun, L., Sharpe, A.H.,Bigby,M.&Georgopoulos,K.(1996)Selectivedefectsin the development of the fetal and adult lymphoid system in mice with an Ikaros null mutation. Immunity 5, 537–549. 13. Wang, J.H., Avitahl, N., Cariappa, A., Friedrich, C., Ikeda, T., Renold, A., Andrikopoulos, K., Liang, L., Pillai, S., Morgan, B.A. & Georgopoulos, K. (1998) Aiolos regulates B cell activation and maturation to effector state. Immunity 9, 543–553. 14. Sabbattini, P., Lundgren, M., Georgiou, A., Chow, C., Warnes, G. & Dillon, N. (2001) Binding of Ikaros to the lambda5 promoter silences transcription through a mechanism that does not require heterochromatin formation. EMBO J. 20, 2812–2822. 15. Koipally, J. & Georgopoulos, K. (2000) Ikaros interactions with CtBP reveal a repression mechanism that is independent of histone deacetylase activity. J. Biol. Chem. 275, 19594–19602. 16. Cobb, B.S., Morales-Alcelay, S., Kleiger, G., Brown, K.E., Fisher, A.G. & Smale, S.T. (2000) Targeting of Ikaros to pericentromeric heterochromatin by direct DNA binding. Genes Dev. 14, 2146– 2160. 17. Romero, F., Martinez, A.C., Camonis, J. & Rebollo, A. (1999) Aiolos transcription factor controls cell death in T cells by regu- lating Bcl-2 expression and its cellular localization. EMBO J. 18, 3419–3430. 18. Brown, K.E., Guest, S.E., Smale, S.T., Hahm, K., Merkensch- lager, M. & Fisher, A.G. (1997) Association of transcriptionally silent genes with Ikaros complexes at centromeric hetero- chromatin. Cell 91, 845–854. 19. Koipally, J., Renold, A., Kim, J. & Georgopoulos, K. (1999) Repression by Ikaros and Aiolos is mediated through histone deacetylase complexes. EMBO J. 18, 3090–3100. 20. Kim, J., Sif, S., Jones, B., Jackson, A., Koipally, J., Heller, E., Winandy,S.,Viel,A.,Sawyer,A.,Ikeda,T.,Kingston,R.& Georgopoulos, K. (1999) Ikaros DNA-binding proteins direct formation of chromatin remodeling complexes in lymphocytes. Immunity 10, 345–355. 21. Turner, J. & Crossley, M. (2001) The CtBP family: enigmatic and enzymatic transcriptional co-repressors. Bioessays 23, 683–690. Ó FEBS 2002 Eos binds the corepressor CtBP (Eur. J. Biochem. 269) 5891 22. Boyd, J.M., Subramanian, T., Schaeper, U., La Regina, M., Bayley, S. & Chinnadurai, G. (1993) A region in the C-terminus of adenovirus2/5E1aproteinisrequiredforassociationwithacel- lular phosphoprotein and important for the negative modulation of T24-ras mediated transformation, tumorigenesis and metasta- sis. EMBO J. 12, 469–478. 23. Poortinga, G., Watanabe, M. & Parkhurst, S.M. (1998) Droso- phila CtBP: a Hairy-interacting protein required for embryonic segmentation and hairy-mediated transcriptional repression. EMBO J. 17, 2067–2078. 24. Nibu, Y., Zhang, H. & Levine, M. (1998) Interaction of short- range repressors with Drosophila CtBP in the embryo. Science 280, 101–104. 25. Turner, J. & Crossley, M. (1998) Cloning and characterization of mCtBP2, a co-repressor that associates with basic Kru ¨ ppel-like factor and other mammalian transcriptional regulators. EMBO J. 17, 5129–5140. 26. Holmes,M.,Turner,J.,Fox,A.,Chisholm,O.,Crossley,M.& Chong, B. (1999) hFOG-2, a novel zinc finger protein, binds the co-repressor mCtBP2 and modulates GATA-mediated activation. J. Biol. Chem. 274, 23491–23498. 27. van Vliet, J., Turner, J. & Crossley, M. (2000) Human Kru ¨ ppel- like factor 8: a CACCC-box binding protein that associates with CtBP and represses transcription. Nucleic Acids Res. 28, 1955– 1962. 28. Smith, D.B. & Johnson, K.S. (1988) Single-step purification of polypeptides expressed in Escherichia coli as fusions with glu- tathione S-transferase. Gene 67, 31–40. 29. Sambrook, J., Fritsch, E.F. & Maniatis, T. (1989) Molecular Cloning: a Laboratory Manual, 2nd edn. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 30. Sundqvist, A., Sollerbrant, K. & Svensson, C. (1998) The carboxy- terminal region of adenovirus E1A activates transcription through targeting of a C-terminal binding protein-histone deacetylase complex. FEBS Letts. 429, 183–188. 31. Wen, Y., Nguyen, D., Li, Y. & Lai, Z.C. (2000) The N-terminal BTB/POZ domain and C-terminal sequences are essential for Tramtrack69 to specify cell fate in the developing Drosophila eye. Genetics 156, 195–203. 32. Zhang, C.L., McKinsey, T.A., Lu, Jr & Olson, E.N. (2001) Association of COOH-terminal-binding protein (CtBP) and MEF2-interacting transcription repressor (MITR) contributes to transcritional repression of the MEF2 transcription factor. J. Biol. Chem. 276, 35–39. 33. Koipally, J. & Georgopoulos, K. (2002) A molecular dissection of the repression circuitry of Ikaros. J. Biol. Chem. 277, 27697–27705. 34. Koipally, J. & Georgopoulos, K. (2002) Ikaros–CtIP interactions do not require C-terminal Binding Protein and participate in deacetylase-independent mode of repression. J. Biol. Chem. 277, 23143–23149. 5892 J. Perdomo and M. Crossley (Eur. J. Biochem. 269) Ó FEBS 2002 . The Ikaros family protein Eos associates with C-terminal-binding protein corepressors Jose ´ Perdomo and Merlin Crossley Department of Biochemistry,. aberrant Ikaros protein, which cannot bind DNA, is still able to dimerize with other Ikaros family proteins and, most likely, interfere with their functions.