HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 15, 13007-13015, April 12, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/15/13007    most recent M111371200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Koipally, J. Articles by Georgopoulos, K. Search for Related Content PubMed PubMed Citation Articles by Koipally, J. Articles by Georgopoulos, K. Social Bookmarking                      What's this?

Unconventional Potentiation of Gene Expression by Ikaros^*

Joseph Koipally, Elizabeth J. Heller, John R. Seavitt, and Katia Georgopoulos^§

From the Cutaneous Biology Research Center, Massachusetts General Hospital and Harvard Medical School, Charlestown, Massachusetts 02129

Received for publication, November 28, 2001, and in revised form, January 11, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Ikaros is essential for the normal development and regulated proliferation of lymphoid cells. In lymphocytes, Ikaros exists as an integral component of chromatin-remodeling complexes, including the Mi-2/nucleosome remodeling and deacetylation complex (NuRD) complex. It is expected that Ikaros, together with these associated activities effects repression, but here we show that they may also potentiate gene expression in cycling cells. Ikaros cannot activate transcription by itself; instead, it enhances the activity of both weak and strong activators. For this role in potentiation, Ikaros requires its DNA binding and dimerization domains. The DNA binding and dimerization properties of Ikaros are also responsible for its targeting to pericentromeric heterochromatin (PC-HC). Significantly, Ikaros mutants with altered specificity for DNA binding that are unable to localize to PC-HC are incapable of stimulating transcription from reporters bearing their cognate sites. Thus, potentiation of gene expression by Ikaros correlates strongly with its ability to localize to PC-HC in combination with the chromatin remodeler Mi-2.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Lineage commitment and differentiation along the hemolymphoid pathway rely heavily on Ikaros, which encodes a number of Krüppel-type zinc finger proteins (1-4). Ikaros contains seven coding exons, four of which can be alternatively utilized to generate a number of isoforms (5, 6). These Ikaros proteins differ in the number of N-terminal zinc fingers that constitute their DNA binding domain. Ikaros isoforms with at least three N-terminal zinc fingers (i.e. Ik-1, Ik-2 and Ik-3) are capable of binding a high affinity Ikaros site that contains the GGGAA core motif, whereas all other isoforms cannot bind this site (5). Nonetheless, all of the Ikaros isoforms share two hunchback-related zinc fingers at their C terminus, which are necessary for dimerization between Ikaros proteins and family members (7). Interactions between Ikaros isoforms that can and cannot bind DNA compromise the ability of the resulting complex to bind DNA (7, 8). This indicates that the non-DNA binding Ikaros isoforms can act as naturally occurring dominant negative factors to regulate the activity of the DNA binding isoforms.

A role for Ikaros is manifested from the earliest steps of the hemopoietic pathway. Lack of Ikaros causes a significant reduction (30-40-fold) in hemopoietic stem cell activity that is made more severe (>100-fold) by the increased expression of its dominant negative isoforms (9). Lineage restriction of multipotent hemopoietic progenitors toward the lymphoid pathways is severely affected in the absence of Ikaros. Mice homozygous for an Ikaros null mutation lack all B, natural killer, and fetal T cells as well as the earliest described lymphoid progenitor (10). Nonetheless, some postnatal T cell precursors are generated in these mice, though they display skewing in their differentiation toward the CD4/TCR lineage (10). The limited number of Ikaros-deficient thymocytes and mature T cells display a T cell receptor-mediated hyperproliferative phenotype in vitro and undergo rapid clonal expansions in vivo (10). Mice homozygous for a mutation that generates only dominant negative Ikaros isoforms display similar but more severe defects (8). In addition to the B and natural killer cell deficiency, they lack all fetal and adult T cells. Mice heterozygous for this mutation have lymphocyte populations that appear normal; however, their T cells display augmented T cell receptor-mediated proliferative responses, and they rapidly develop leukemias and lymphomas (11-13). The more severe phenotype of the dominant negative compared with the null mice suggested the presence of family members whose activity was affected by this mutation. Consistent with this hypothesis, three family members were identified (Aiolos (14), Helios (15, 16), and Eos/Daedalus (17, 18)) that are more restricted in hemopoietic expression relative to Ikaros.

The aforementioned studies identify Ikaros as an important regulator of several steps in the hemolymphoid pathways (2, 3, 4, 19). The mechanisms by which Ikaros operates along this pathway have been the focus of diverse studies, some of which implicate it as a repressor of gene expression for the following reasons. First, in actively cycling primary lymphocytes, Ikaros concentrates in distinctive toroidal structures, which are found in apposition to pericentromeric heterochromatin and a variety of transcriptionally silent genes (20). Second, in late S phase, some of the Ikaros toroids become coincident with clusters of DNA replication origins, which presumably are sites of replicating heterochromatin (12). Third, a biochemical purification of Ikaros from lymphocytes has determined that the majority of Ikaros exists within a stable 2MD complex containing components of the NuRD¹ complex that includes the ATPase, Mi-2, and Class I histone deacetylases that are presumed to play a role in repression (21), while a smaller fraction of Ikaros also interacts with the putative co-repressors Sin3 and C-terminal binding protein (22, 23). Fourth, heterologous fusions of Ikaros to the Gal4 DNA binding domain behave as potent transcriptional repressors (22). Furthermore, recent studies have suggested that Ikaros may be involved in the repression of the 5 and terminal deoxynucleotidyl transferase genes (24, 25).

Ikaros has also been reported to function as an activator of transcription. Ectopic expression of Ikaros with reporters containing engineered Ikaros sites in non-lymphoid cells results in gene activation (1, 5, 7). Gene expression-profiling experiments in Ikaros-deficient hemopoietic precursors further support such a role; expression of the tyrosine kinase receptors flk-2 and c-kit in early hemopoietic progenitors is reduced in the absence of Ikaros (9). Ikaros has also been reported to activate from the enhancer of a mink cell focus-inducing virus (26). Additionally, Ikaros functions as a suppressor of variegation for regulatory elements of the CD8 locus.²

Ikaros' potential to function as a repressor and activator of gene expression may provide a molecular basis for its diverse effects in the hemolymphoid system. To gain further insight into this important problem in lymphocyte biology, we have undertaken a comprehensive analysis of the transcriptional properties of Ikaros. Here we report that Ikaros is indeed capable of enhancing gene expression as a potentiator of bona fide transcriptional activators and not by functioning as a classical activator. Potentiation by Ikaros requires its intact DNA binding and dimerization domains, both of which are necessary for its recruitment into a PC-HC-associated nuclear compartment in cycling cells. We show that there is an unexpected correlation between Ikaros localization to these heterochromatin-associated sites and its ability to activate gene expression. These studies also indicate that the presence of Ikaros in this nuclear compartment provide a "landing pad" for its chromatin remodeling partner Mi-2 and presumably the NuRD complex. Two models are proposed to explain these findings.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Plasmids-- 4X IkBS2tkCAT and tkCAT have been previously described (5). 4XIKAS1tkCAT were constructed by cloning ligated oligonucleotides containing the corresponding sites into the tkCAT reporter by standard cloning methods. 4XSp1E1BCAT, G5E1BCAT, CMV2Flag, CMV2Flag-Ik1, CMV2Flag-Ik1M, BXG1, and BXG1-Ik1 have been described previously (22, 23). CMV2Flag-Ikaros and BXG1-Ikaros DBD and activation domain mutants described in this paper were constructed using the Stratagene mutagenesis kit.

Transfections-- 293T and NIH-3T3 cell lines were maintained in Dulbecco's modified Eagle's medium with 10% fetal bovine serum (Hyclone). Transfections of these cell lines were carried out using the HBS-CaP0₄ method. For repression assays, 1 µg of the Gal4 fusion plasmid, 10 µg of the Gal4 reporter plasmid, and 0.5 µg of the pXGH5 growth hormone control plasmid were used. For activation assays, typically 1.5 µg of the reporter, 0.25 µg of Ikaros expression plasmid, and 0.5 µg of the GH plasmid were used. 24 h after transfection, cells were fed with fresh media, and 18-24 h later cells were harvested and processed for CAT assays as described (22). Growth hormone assays were done as recommended by the manufacturer (Nichols Institute). Transfections were typically performed in duplicate and repeated between three to six times. For interaction experiments, 293T cells were transfected with 1-2 µg of the CMV2Flag vector expressing the Ikaros protein of choice. 24 h later, cells were fed with fresh media, and 48 h later cells were harvested to prepare nuclear or whole cell extracts. For immunofluorescence studies, NIH-3T3 cells were transfected with 9 µg of expression plasmid and 1 µg of GH plasmid using ProVera TransIT-L1 transfection reagent. 36 h after transfection, cells were harvested for fixation and staining as described below.

Immunoprecipitation and Western Analysis-- Whole-cell extracts from 293T cells transfected with the relevant plasmids were prepared as previously described (7) and precleared using protein-G-agarose beads (Roche Molecular Biochemicals). The precleared extracts were incubated with the antibody of interest or the relevant isotype control on ice for 1 h. 30 µl of protein-G beads were then added to the extract, and the extracts were rotated overnight. The beads were collected by centrifugation and washed four times with TS buffer (20 mM Tris pH 7.5, 150 mM NaCl) (7). The beads obtained after this procedure were treated with SDS sample buffer, boiled at 95 °C for 15 min, and loaded on an SDS-polyacrylamide gel along with 8-10% of the cell extract used for the immunoprecipitation. The proteins were transferred to a nitrocellulose membrane, probed with the relevant antibody, and examined by autoradiography with ECL (Amersham Biosciences). Antibodies used were: FLAG M2 (Sigma), Gal4 (Santa Cruz), HDAC2 (Zymed Laboratories Inc.), and anti-Ikaros and Mi-2, which have been previously described (21).

Preparation of Nuclear Extracts, Gel Shift, and Supershift Analysis-- Nuclear extracts were prepared from 293T cells transfected with the Ikaros constructs in either CDM8-flag or CMV-2-flag expression vectors as previously described (23). The relative amount of each Ikaros protein was determined by Western blotting using the Ikaros monoclonal antibody, 8H2, and roughly normalized. DNA mobility shift assays were performed as previously described (7). The IKBS1 oligonucleotide TCAGCTTTTGGGAATACCCTGTTCA, containing a high affinity Ikaros binding site, was used as well as an oligonucleotide designed for the altered site specificity Ikaros constructs f2s1 and f2s6, IKAS1, TCAGCTTTTGGGAGTACCCTGTTCA. Radiolabeled oligonucleotides were prepared by end labeling with [³²P]dATP 6000 Ci/mmol (NEN Life Science Products) and then annealed to their complement. A cold competitor oligonucleotide, without an Ikaros binding site, TCAGCTTTTGAAAATACCCTGTTCA, IKm, was added to all reactions to reduce background. A mix of three Ikaros monoclonal antibodies, 4E9A4, 8H2, IK14, was used for the supershift experiments.

Fluorescent Microscopy-- Cells were transfected as described above, cytospun onto Superfrost Plus slides (Fisher), fixed, permeabilized in phosphate-buffered saline with 2% paraformaldehyde and 0.1% Triton X-100 for 20 min on ice, and washed repeatedly. Slides were blocked in phosphate-buffered saline with 3% bovine serum albumin and 1% normal donkey serum for 1 h at room temperature and stained with a 1:200-dilution of anti-Ikaros antibody 4E9-A4 overnight at 4 °C. Slides were washed and incubated with a 1:200-dilution of fluorescein isothiocyanate-conjugated donkey anti-mouse IgG for 45 min at room temperature. Slides were counterstained with 1 µg/ml Hoechst 33342 (Molecular Probes, Eugene, OR) and mounted with Vectashield (Vector, Burlingame, CA). Control staining was performed with an isotype-matched primary antibody; immunoreagents were obtained from Jackson ImmunoResearch (West Grove, PA). Images were obtained with an Olympus BX50 (Tokyo, Japan) fluorescent microscope with blue and UV filter modules.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Ikaros' Activation Function Does Not Rely on a Previously Identified Activation Domain-- We have previously shown that ectopic expression of Ikaros and its family members can transactivate reporter genes (5, 7). However, the localization of Ikaros in nuclear structures that are in apposition to centromeric heterochromatin and silent genes, coupled with molecular data showing that it can repress transcription, has called into question its function as an activator (12, 20-22).

Transcriptional activators typically contain at least one module, termed an activation domain, which is required to enable their interaction with co-activators and the RNA polymerase II complex. Using a yeast one-hybrid assay, a single bipartite activation domain was identified in Ikaros (7). This domain is present at the very N-terminal region of the last exon of Ikaros and is comprised of a putative  helical region followed by a  sheet, both of which are required for maximal activation in Gal4-tethering assays in mammalian cells. The activation domain is highly conserved between the Ikaros family members, Aiolos and Helios, and to a lesser extent in Eos/Daedalus; heterologous fusions of this domain from all Ikaros family members are capable of similar levels of activation (data not shown).

The functional importance of this domain was examined in the context of the full-length Ikaros protein. Several different mutations were generated in this region and tested for their effect on Ikaros' transactivation function on an Ikaros binding site-driven reporter (Fig. 1A, 4XIKBS2tkCAT). Testing of Ikaros and its mutant variants was performed in NIH-3T3 fibroblasts for the following reasons. First, neither Ikaros nor any of its family members are expressed in this cell line. Second, ectopic expression of Ikaros and family members in these cells recapitulates their nuclear compartmentalization (i.e. localization in heterochromatin-associated toroidal structures) observed in cycling primary lymphocytes (27). Third, these cells are not grossly transformed and are amenable to growth controls.

View larger version (23K): [in this window] [in a new window]   Fig. 1.   The activation domain of Ikaros is not necessary for its ability to activate transcription. A, NIH-3T3 cells were transfected with 1.5 µg of 4XIKBS2tkCAT reporter, 0.25 µg of the indicated Ikaros expression vectors and 0.5 µg of GH internal control plasmid. The encoded Ikaros proteins differ in the presence or absence of the helical (-Ah),  strand (-Ab), or the combination (-Adel) of components of the activation domain. Normalized CAT activity of Ikaros and its activation domain mutants were divided by the normalized CAT activity value of the vector, and this ratio is reported as fold activation. Experiments were repeated in duplicate five times. B, NIH-3T3 cells were transfected with the following reporters: 1.5 µg of E1B TATA CAT, 3X Ikaros E1B CAT, or 4X Sp1 E1B CAT together with 0.25 µg of empty or Ikaros expression vector and 0.5 µg of the GH internal control plasmid. Fold activation was determined as described in A. Experiments were done in duplicate four times. C, NIH-3T3 cells were transfected with 2 µg of the reporter plasmid, G5E1BCAT, 0.5 µg of empty vector, Gal4 DBD or Gal4 DBD fusions to full-length Ikaros (-Ik1), the Ikaros activation domain (-IkAD), Sp1 (-Sp1), Sp1 lacking its C-terminal zinc fingers (-Sp1N), and amino acids 399-499 of CTF-1 (-CTF) in the presence or absence of Ikaros (0.25 µg) and 0.5 µg of GH internal control plasmid. Fold activation was calculated by dividing the normalized CAT activity obtained by the vector or Gal4 fusion in the presence of Ikaros by the normalized activity obtained in the absence of Ikaros. Experiments were done in duplicate thrice.

The first class of activation domain mutants we tested consisted of several proline point mutations in the  helical and  sheet domains of Ikaros. None of these mutants had any effect on the ability of Ikaros to transactivate the reporter (data not shown). The helical (-Ah) and  sheet (-Ab) regions of the activation domain were then respectively deleted in the context of full-length Ik1. These deletion mutations also had no effect on transactivation by Ikaros (Fig. 1A). Finally, an Ikaros mutant with the entire activation domain deleted was tested. Surprisingly, this mutant was fully capable of activation (Fig. 1A). These findings argue that the previously identified transactivation domain is not required for the activation properties of the full-length protein. However, we cannot rule out the possibility that this domain is utilized in a cell- and context-specific manner. Nevertheless, in the context of our experiments, activation occurs through other means. Since no other domain of Ikaros was capable of supporting activation in detailed dissections by one-hybrid assays in mammalian cells,³ Ikaros may not contain a classical activation domain. Alternatively, other Ikaros activation domains, if they exist, may not be able to be identified by such assays.

Ikaros Can Function As a Non-classical Transcriptional Activator-- How might Ikaros turn on gene expression in the absence of a canonical activation domain? Our studies thus far were done using reporters containing the thymidine kinase promoter that contains binding sites for the transcription factors, Sp1 and CTF. Activation by Ikaros on these reporters may result from its cooperation (direct or indirect) with these factors. To examine whether Ikaros could transactivate in the absence of other transcriptional activators, we made use of reporters containing the TATA box from the adenovirus E1B gene (Fig. 1B, E1B TATA). We also introduced three high affinity Ikaros sites or, as a control, four Sp1 sites upstream of the E1B TATA box. These reporters were cotransfected with the Ikaros expression vector into NIH-3T3 cells and assayed for CAT activity. Ikaros did not activate from either the E1B-CAT or the Ikaros binding site-containing variant (Fig. 1B). Thus, Ikaros cannot activate a minimal promoter that consists of only a TATA box. On the other hand, the NFB subunit p65, which can also bind to the Ikaros site, strongly activated the latter (data not shown). In striking contrast, Ikaros strongly transactivated the Sp1 E1B CAT reporter (Fig. 1B, roughly 18-fold), which has not been engineered to contain any Ikaros binding sites in the vicinity of the promoter. In addition, Ikaros was also capable of stimulating transcription of the Sp1 site-containing promoter of the tkCAT reporter even in the absence of any introduced Ikaros sites (data not shown). Two Ikaros family members, Aiolos and Helios, were also capable of similar increases in activity from these reporters (data not shown). The stimulatory effect Ikaros (and family members) has on the activity of basal transcription factors (i.e. Sp1) in the absence of any engineered binding sites can be explained in three ways. (a) Ikaros may activate transcription without binding DNA (indirect recruitment) as has been reported in some instances of activation by BRCA1 (28). For example, it may interact with other DNA-bound factors such as Sp1 to regulate transcription. (b) Ikaros may be titrating co-repressors away from the promoter (squelching), hence causing activation of reporters with or without Ikaros sites. In agreement with this hypothesis, Ikaros has been shown to interact with several co-repressors and can function as a transcriptional repressor (21-23). (c) Ikaros may bind to the backbone of the reporters, which incidentally have several high, medium, and low affinity Ikaros-consensus binding sites, and in this manner trans-activate reporter expression from these sites.

To determine whether Ikaros mediates activation through indirect recruitment, we tested Ikaros' ability to stimulate a diverse series of activators/activation domains as Gal4 fusions. These included Gal4 DBD, Gal4-CTF, Gal4-Ikaros activation domain, Gal4-Sp1 and Gal4-Sp1N (lacking its zinc fingers). The potent repressor, Gal4-Ikaros was also examined. These heterologous proteins were transfected with the reporter 5XGal4 E1B CAT into NIH-3T3 cells in the presence or absence of Ikaros or a dimerization-defective Ikaros mutant. Ikaros expression stimulated all of the activation domains tested, albeit to different levels (Fig. 1C). Even the cryptic activation domain contained in the Gal4 DBD was further enhanced by Ikaros. Only the potent repressor, Gal4-Ik1 (and vector alone) was not activated by Ikaros (Fig. 1C). Among the activators used, only Sp1 has the potential for direct interaction with Ikaros (J. Koipally),³ but this interaction was not required for activation as Sp1N, an Ikaros/Sp1 interaction mutant, was also stimulated by Ikaros (Fig. 1C). In contrast to wild type Ikaros, the dimerization-defective Ikaros protein was unable to potentiate reporter gene activity (data not shown).

Collectively, these data indicate that Ikaros' ability to stimulate transcription does not occur via a classical mechanism; Ikaros is unable to activate by itself, but rather enhances gene expression by bona fide activators. Ikaros can potentiate the functionally diverse activation domains present in Sp1 (glutamine-rich), CTF (proline-rich), and Ikaros (acidic and hydrophobic) proteins. Since none of these domains directly interact with Ikaros, its potentiation effect is not expected to be mediated through recruitment by these proteins. Instead, the enhancement of gene expression may result from co-repressor squelching or from Ikaros binding to sites associated with the reporter.

Generation and Characterization of Ikaros DNA Binding Domain Mutants-- Gene activation through squelching of co-repressors would not require an intact DNA binding domain unlike activation through Ikaros sites. To distinguish between these two possibilities, we had two available choices: we could mutate the numerous Ikaros sites in the vector or we could target mutations in the Ikaros DNA DBD. As multiple mutations in the reporter DNA would interfere with vector propagation, we constructed Ikaros DBD mutants. Some mutations were designed to ablate Ikaros' binding to DNA and others to redirect its specificity for DNA. For the design of these mutations, we relied on previous reports on the structure of the zinc finger motif and the amino acids that contact DNA (29, 30). The Krüppel-type zinc finger motif is comprised of a  strand and an  helix. Amino acids at positions 1, 1, 2, 3, and 6 of the alpha helix make contacts with bases in the major groove of DNA and determine the binding specificity of the protein (29, 30). Mutations were introduced in the four Ikaros N-terminal zinc fingers at these positions, some of which were expected to disrupt their DNA binding specificity (Fig. 2A). Two of the second and third zinc finger mutations, f2w and f3m (Fig. 2A) respectively, had been previously identified in a Jurkat T cell subline and in thymic lymphomas that developed in -irradiated mice (31). These mutations had no effect on protein expression (Fig. 2B).

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Ikaros DNA binding domain mutants. A, a schematic representation of the DBD of Ikaros identifying the key amino acids involved in DNA recognition (in bold). The residues that were mutated for this study are underlined. The amino acids in finger 2 are numbered to aid the reader to follow the description under "Results." The finger-DNA contacts identified are simplistic at best but serve to identify some of the most critical residues involved in this interaction. The DBD mutants are listed below the cartoon of the zinc fingers. B, 293T cells were transfected with mammalian expression vectors encoding each of the mutants listed above. Whole cell extracts were prepared from these transfectants, and a Western analysis was performed to ensure that all the mutants could be expressed.

The DNA binding properties of these Ikaros DBD mutants were examined in vitro using gel shift assays. Wild type and mutant proteins (generated in 293T cells) were tested for their ability to bind a consensus Ikaros binding site or a mutated variant. The Ikaros-related nucleoprotein complex was identified by supershifts using Ikaros and control (Aiolos) antibodies. Neither the wild type nor any of the mutant Ikaros proteins tested bound to the mutant site (Fig. 3A). However, wild type Ikaros and three of the DBD mutants, f1s1, Rbm, and f4s1, bound to the Ikaros cognate site. This is not surprising, since these mutations are in the first (f1s1) or fourth finger (f4s1), which are not directly involved in DNA binding (5), or in a region of the second finger (Rbm) that is outside the presumed  helical motif that contacts the major groove of DNA. The protein-DNA complexes detected upon expression of these proteins were supershifted by the Ikaros but not by the Aiolos antibodies (Fig. 3A). The f3m mutation drastically decreased the DNA binding ability of this protein, which was detectable only upon addition of the Ikaros antibody (Fig. 3A, f3m). None of the other Ikaros DBD mutants bound to the Ikaros cognate site at any significant level (Fig. 3A, f2s1, f2s2, f2s3, f2s4, f2s5, f2s6, f2w, and f3s1). In some instances, a greater amount of the mutant protein was used for DNA binding compared with wild type to definitively determine its binding potential (Fig. 3A, f2s1, f2s2, f2s5, and f2s6). These data extend our previous findings (5) that fingers two and three of the Ikaros proteins make contacts with DNA and thus play a critical role in DNA binding.

View larger version (52K): [in this window] [in a new window]   Fig. 3.   Ikaros-mediated activation requires an intact DNA binding domain. A, Gel-shift analysis of the Ikaros DBD mutants. Nuclear extracts prepared from 293T cells were analyzed by Western analysis (top), and the same amounts of extracts used were incubated with mutant (m) or high affinity Ikaros binding sites, IKBS2, as a probe (w). Supershifts were performed using Ikaros (I) or Aiolos (A) antibodies. The asterisk indicates a background band. B, transcriptional assay of DBD mutants. NIH-3T3 cells were transfected with 1.5 µg of 4XIKBS2tkCAT reporter, 0.25 µg of vector, Ikaros wild type, or DBD mutant expression vectors, and 0.5 µg of a GH internal control plasmid. Extracts from transfectants were assayed for CAT activity. Normalized CAT activity of Ikaros and its DBD mutants were divided by the normalized value of the vector, and this ratio is reported as fold activation. Transfections were repeated at least four times in duplicate. The variation between experiments is indicated.

One concern when introducing mutations is that they may cause the protein to be misfolded, rather than targeting a specific function. This possibility was examined by testing whether the mutations in the Ikaros DBD influenced the ability of the protein to interact with other factors and engage in the previously characterized Ikaros-chromatin remodeling complexes. All of the Ikaros DBD mutants interacted at wild type levels with Mi-2 and histone deacetylases, which are Ikaros' major partners in lymphocytes (Fig. 4A). These mutants were also capable of wild type interactions with Sin3A and C-terminal binding protein as well as with other Ikaros isoforms and family members (data not shown). We have previously shown that Ikaros functions as a strong repressor when tethered to the Gal4 heterologous DNA binding domain (22). All of the Ikaros DBD mutants were capable of wild type levels of repression in this system, regardless of their ability to bind Ikaros sites (Fig. 4B). In summary, we have identified mutations in the DBD of Ikaros that can abolish its DNA binding specifically, without altering interactions with Ikaros' corepressor partners or repression mediated by Ikaros.

View larger version (54K): [in this window] [in a new window]   Fig. 4.   Ikaros DBD mutants are wild type with respect to interaction with co-repressors and transcriptional repression. A, 293T cells were transfected with expression vectors of each of the Ikaros DBD mutants listed, and whole cell extracts were prepared and subject to immunoprecipitation with Ikaros antibodies. Immunoprecipitates were analyzed by Western analysis using antibodies to Mi-2 and HDAC2 to detect if these endogenous repressors could still associate with the mutants. B, NIH-3T3 cells were transfected with 10 µg of the reporter, G5tkCAT (diagrammed above the graph), 1 µg of Gal4-Ikaros DBD mutants, and 0.5 µg of GH internal control plasmid. Extracts prepared from harvested transfectants were assayed for CAT activity, which was then corrected for transfection efficiency using the GH plasmid. Fold repression relates the ratio of the normalized CAT value obtained with Gal4 divided by that obtained by the Gal4-Ikaros DBD mutant.

The DNA Binding Properties of Ikaros Are Essential for Activating Transcription-- The Ikaros DBD mutants were tested for their ability to activate gene expression. Ikaros proteins with mutations in the Ikaros DBD that did not affect binding to DNA activated transcription to a similar level as wild type (Fig. 3B, f1s1, Rbm, and f4s1). The DBD mutant, f3 m, which displayed reduced binding to the Ikaros site, was compromised in its potentiation capacity (Fig. 3B, f3m). This mutant showed significant variability in its strength of activation, which may be a result of its weak DNA binding ability. Finally, all Ikaros DBD mutants that were unable to bind DNA also lost their ability to potentiate gene expression (Fig. 3B, f2s1, f2s2, f2s3, f2s4, f2s5, f2s6, and f3s1). A dimerization-domain Ikaros mutant that indirectly affects DNA binding was also unable to activate the reporter (data not shown). Similar results were obtained using the DBD mutants in combination with reporters engineered with and without Ikaros binding sites in the proximity of the promoter (4XIKBS2tkCAT and 4XSp1E1BCAT reporter) (data not shown).

These studies with the Ikaros DBD mutants show that the ability of Ikaros to potentiate gene expression is dependent on zinc fingers 2 and 3 and an intact dimerization domain, both of which are required for efficient DNA binding. While we cannot exclude the possibility that the DBD mutations are impaired in activation because of a disruption of a protein-protein interaction, this possibility seems less likely since several different DBD mutations as well as the dimerization defective mutant are transcriptionally inactive. It also seems unlikely that the observed increase in gene expression by Ikaros results from the simple removal of co-repressors, i.e. squelching, given the capacity of all DBD mutants to interact with repressors and repress transcription when fused to a heterologous DNA binding domain (Fig. 4, A and B). These data suggest that Ikaros potentiates transcription in a DNA binding-dependent manner.

The DNA Binding Specificity of Ikaros Dictates Its Function As a Potentiator-- The role of the Ikaros DBD in potentiation of gene expression was further examined by generating variants with altered DNA binding specificity. The DBD mutants, f2s1 and f2s6, were constructed, based on previously described zinc finger specificity studies, to bind the sequence GGGAG rather than the Ikaros site, GGGAA. This prediction was tested in in vitro DNA binding analysis (Fig. 5A, Ik1, f2s1, f2s6, f2s2, and vector control). Wild type Ikaros protein but not the altered specificity mutants f2s1 and f2s6 interacted with the Ikaros cognate site (Fig. 5A, GGGAA, IKBS2). The altered specificity DBD mutants, f2s1 and f2s6, but not the wild type Ikaros protein, bound to their predicted altered site (Fig. 5A, GGGAG, IKAS1). Another DBD mutant with a mutation in the second zinc finger, f2s2, bound neither the Ikaros nor the altered specificity site (Fig. 5A). Thus, the DBD mutations f2s1 and f2s6 alter the DNA binding specificity of the protein to GGGAG rather than the GGGAA Ikaros consensus.

View larger version (50K): [in this window] [in a new window]   Fig. 5.   The wild type DNA binding context is critical for Ikaros to activate transcription. A, gel shift analysis of nuclear extracts prepared from 293 T cells transfected with the indicated Ikaros expression vector. Supershifts were performed with Ikaros (I) or Aiolos (A) antibodies. f2s1 and f2s6 are altered specificity mutants that bind the sequence GGGAG unlike Ikaros, which binds GGGAA. f2s2 binds neither sequence. B, transcriptional assays were performed to determine whether the altered specificity mutants could support activation of a thymidine kinase reporter driven by four repeats of the altered specificity binding site. The procedure was essentially that described in Fig. 3B.

We next tested whether the f2s1- and f2s6-altered specificity mutants could transactivate from their recognition sites introduced upstream of the tkCAT reporter. NIH-3T3 cells were co-transfected with the altered specificity reporter, 4XIKAS1tkCAT, and the altered specificity mutants or the wild type Ikaros protein. Surprisingly, the altered specificity mutants f2s1 and f2s6, like f2s2, did not transactivate the reporter even though they were able to bind their altered specificity binding site IKAS1 (Fig. 5A). In a separate experiment, another Ikaros-altered specificity mutant, f3s1, which can bind to the GAGAA sequence, also failed to transactivate from its cognate site introduced upstream of the tkCAT reporter (data not shown). In contrast, the wild type Ikaros protein (Ik1) transactivated these reporters.

Taken together, these data argue not only that DNA binding is important for Ikaros' ability to stimulate transcription, but also that the nature of the Ikaros binding domain and the context of its site may ultimately dictate its function in transcriptional regulation. These findings also provide an important distinction between the activation and repression properties of Ikaros; although they both rely on DNA binding, only the latter can be rescued by a heterologous DNA binding domain.

Ikaros' Ability to Potentiate Gene Expression Correlates with Its Targeting to PC-HC-- How does the nature or context of Ikaros DBD sites dictate potentiation of gene expression? Previous studies have indicated that Ikaros utilizes its DNA binding activity to target the vicinity of pericentromeric heterochromatin, possibly through Ikaros binding sites present in  satellite DNA (27).

We explored the relation, if any, between nuclear localization and the transactivation potential of the Ikaros DBD mutants. Wild type and Ikaros mutants were introduced into NIH3T3 cells, and their localization was determined by immunofluorescence. The vast majority of wild type Ikaros protein is found in toroidal structures that surround the densely staining heterochromatin (Fig. 6A). Mutations in the DBD that do not alter its specificity or affinity for DNA (f1s1, Rbm, and f4s1) did not interfere with the localization of the protein in the heterochromatin-associated toroidal structures (Fig. 6A and data not shown). Mutations that alter but do not completely ablate DNA binding reduced the amount of protein in these structures (Fig. 6A, f3m). In contrast, mutations that ablate DNA binding (f2s1, f2s2, and f2s6) did not allow localization in these toroids; instead the mutant proteins were diffusely distributed throughout the nucleus and in some instances were also found in the cytoplasm (Fig. 6A and data not shown). Thus, all Ikaros DBD mutants that bind DNA and activate transcription localize to heterochromatin-associated nuclear structures in actively cycling NIH-3T3 cells. In contrast, the mutants that do not bind wild type Ikaros sites and do not activate transcription are non-heterochromatic in their localization. These data identify a striking correlation between Ikaros' ability to activate transcription and its localization to heterochromatin-associated environments.

View larger version (27K): [in this window] [in a new window]   Fig. 6.   Ikaros DBD mutants that cannot activate are inefficient in localizing to centromeric heterochromatin. A, NIH-3T3 cells were transfected with the indicated Ikaros DBD mutants (f1s1, f2s2, f2s6, f3m) and analyzed by immunofluorescence microscopy for Ikaros protein. B, NIH-3T3 cells, either untransfected (far left) or transfected with Ik1 expression vector. The transfectants were analyzed by immunofluorescence microscopy for Ikaros and Mi-2, as indicated. Central three images are single- and two-color collections of double-stained cell (overlay in yellow). Separately, cells were counterstained with Hoechst to reveal the densely staining regions of heterochromatin relative to Mi-2 (far right, Hoechst in blue).

We have previously shown that in lymphocytes the majority of Ikaros protein and its family members associate with the ATPase Mi-2 in a lymphoid-specific version of the NuRD complex (21). Upon lymphocyte activation, there is a dramatic redistribution of Ikaros and Mi-2 to heterochromatin-associated toroids; however, in Ikaros null lymphocytes Mi-2 is diffusely distributed, suggesting that Ikaros is responsible for its heterochromatic localization (21). In fibroblasts, which lack Ikaros, Mi-2 is also diffusely distributed throughout the nucleus (Fig. 6B). However, when Ikaros is ectopically expressed in these cells, both Ikaros and Mi-2 proteins localize to pericentric heterochromatin-associated toroids. These data provide unequivocal evidence that Ikaros targets Mi-2 to centromeric sites (Fig. 6B and data not shown). Targeting of Mi-2 by Ikaros to PC-HC may be pertinent to the mechanism that underlies Ikaros' potentiation function.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

It is well established that Ikaros plays a critical role at many distinct steps of the hemolymphoid pathway, but the mechanisms involved remain under investigation. Several lines of evidence support roles for Ikaros in repression and activation. Here, we provide evidence in support of a novel mechanism by which Ikaros can modulate gene expression. We show that Ikaros is not a classical activator but rather a potentiator of gene expression. Unlike its capacity to repress transcription, Ikaros' function as a potentiator can not be directed by a heterologous DBD. Instead, Ikaros relies on a specific DNA binding context that permits its localization to pericentromeric heterochromatin together with its chromatin remodeling partner Mi-2.

In our investigation of the mechanism by which Ikaros achieves activation, we obtained some surprising results. A previously characterized bipartite activation domain of Ikaros (7) was found to be unnecessary for Ikaros' ability to activate transcription in the context of the whole protein. Furthermore, Ikaros did not activate a minimal TATA containing promoter; however, when an activator was present, Ikaros significantly enhanced the ability of the activator to stimulate transcription. Distinct activation domains, including the glutamine-rich Sp1, the proline-rich CTF-1 and the acidic-hydrophobic Ikaros activation domain, were all enhanced in their activation capabilities. This enhancement did not require a direct physical interaction between Ikaros and these activators; however, it did require that Ikaros have an intact DBD.

To further investigate the role of DNA binding in Ikaros' ability to potentiate gene expression, four types of mutants were examined. The first type consisted of mutants that remain wild type for binding to the Ikaros recognition site; the second consisted of those that bind this site at a reduced level; the third consisted of those that bind to altered specificity sites; and the fourth consisted of those without DNA binding activity. The importance of a functional Ikaros DBD for its function as a transcriptional potentiator was established by a series of findings. All the Ikaros DBD mutants that were unable to bind DNA failed to transactivate reporter gene expression. A mutant with reduced binding activity displayed a reduction in transactivation potential. Finally, mutations that did not alter DNA binding had no effect on Ikaros' transactivation potential. This study excludes simple squelching of co-repressors by Ikaros or Ikaros' indirect recruitment by activators, neither of which requires its DNA binding properties as mechanisms for activation by Ikaros.

The strong correlation between the DNA binding and activation function of Ikaros makes it likely that Ikaros' ability to enhance gene expression depends on binding DNA. DNA binding per se, however, is insufficient for Ikaros to activate gene expression. Altered specificity Ikaros mutants were incapable of supporting transcription from reporters modified to contain their recognition sites. The fact that Ikaros-altered specificity mutants cannot activate from their cognate sites indicates that Ikaros exhibits a clear specificity for the site context from which it operates. In sharp contrast to the correlation between Ikaros' DNA binding specificity and activation potential, the protein's repression properties could be delivered to a promoter by a heterologous DBD.

Further insight into the question of binding site context is provided by the ability of Ikaros proteins in cycling cells to redistribute into toroidal structures surrounding pericentric heterochromatin (12, 20). This targeting event requires an intact Ikaros DBD and is presumed to occur through interactions of Ikaros with its cognate binding sites in  satellite repeats of pericentromeric heterochromatin (27). In support of this hypothesis, no sites corresponding to the cognate sites of the altered specificity mutants reported here were present in  satellite DNA. Unexpectedly, the ability of Ikaros to localize to heterochromatin correlates with its ability to potentiate gene expression. Ikaros DBD mutants that were unable to localize into heterochromatin-associated structures were unable to activate gene expression. On the other hand, every activation-competent DBD mutant could localize to heterochromatin. The current data suggests that Ikaros may support transcription when localized to heterochromatin since the majority of Ikaros proteins in the cell line assayed are found in this compartment and because Ikaros enhances activation in these cells.

Based on these findings, we propose two non-mutually exclusive models to account for Ikaros' role in potentiating gene expression in cycling cells (Fig. 7). The first model (Fig. 7A) proposes that activation may result from targeted squelching into PC-HC. In the absence of Ikaros, reporter activity is kept to a basal level because of the distribution of the NuRD complex throughout the nucleus (Fig. 7A, left). Upon expression of Ikaros, the targeted removal of Mi-2 (and presumably of the NuRD complex) into PC-HC (Fig. 6B) stimulates reporter activity by facilitating the function of activators (Fig. 7A, middle). In this vein, Ikaros DBD-altered specificity mutants that cannot be targeted to PC-HC but which still interact with all of Ikaros' described co-repressors are unable to potentiate gene expression. This indicates that "simple squelching" is insufficient to alter the local concentration of co-repressors (Fig. 7A, right) and highlights the potential for targeted squelching into PC-HC in Ikaros' apparent function in gene activation. How is specificity of gene regulation explained in this model? In resting lymphocytes, Ikaros is expressed in a diffuse/punctate nuclear pattern and may be recruited to specific gene targets through its DNA binding domain. Upon lymphocyte activation, when a major fraction of Ikaros protein complex moves into PC-HC, gene targets that it leaves behind acquire the potential for activation.

View larger version (27K): [in this window] [in a new window]   Fig. 7.   Two models to explain how Ikaros functions as a potentiator of gene expression in actively cycling cells. A, potentiation by targeted squelching. The histone deacetylase (HD) activity of the NuRD complex is shown reducing access by activators to their binding sites and thus lowering transcription levels. The presence of Ikaros and its association with Mi-2 stimulates gene activity by reorganizing co-repressors into PC-HC, decreasing their local concentration and activity in the proximity of these genes. Activators are then able to bind more efficiently and drive transcription. B, potentiation by chromatin remodeling in PC-HC. This model suggests that Ikaros potentiates the activity of genes to which it is recruited in non-permissive chromatin environments (PC-HC). Ikaros' association with Mi-2 directs remodeling activity to the proximity of its binding sites, increasing accessibility and thus enabling activators to bind and exert their function.

The second model (Fig. 7B) suggests that activation occurs through the direct action of Ikaros on its target gene in a restrictive chromatin environment like PC-HC. It is likely that the heterochromatin compartment houses genes that have to be very tightly controlled, for instance genes that are regulated with the cell cycle or during development. Even if the gene activators are present, they would have significant difficulty in accessing their cognate sites due to the restrictive heterochromatin environment (Fig. 7B, left). Such genes, while predominantly repressed, would have the opportunity to be expressed when their specific activators and Ikaros are both present in the PC-HC compartment (Fig. 7B, middle). In these situations, Ikaros may function as a "potentiator" by helping remodel, through associated Mi-2, a densely packaged chromatin environment, facilitating activator access and binding. Similarly, one could imagine that once the activator is no longer available, Ikaros through its chromatin modifying activities may "close" this gene and bring about its silencing. Thus, Ikaros within the same nuclear compartment may potentiate disparate events in gene expression depending upon the prevailing conditions. As altered specificity Ikaros mutants can interact with chromatin remodeling activities but cannot localize to PC-HC (Fig. 7B, right), this model argues for Ikaros' potentiation of gene expression being important in restrictive rather than permissive chromatin environments. Three recent reports have described that transcriptionally active genes can be associated with centromeric heterochromatin (32, 33).

In summary, this study provides new insight on Ikaros as a non-classical activator of gene expression, a function that is dependent on its DNA binding domain and which correlates with its ability to be targeted together with its associated chromatin remodeling partner, Mi2, to PC-HC. Further studies with stably integrated target genes will allow us to build upon the foundation provided here and determine which of the proposed mechanisms are in operation on Ikaros' chromosomal targets in lymphocytes.

	ACKNOWLEDGEMENTS

We thank Dr. Taku Naito for help with the design of the figures. We also thank Esther Wong and Drs. Taku Naito, Michael Pazin, and Bruce Morgan for careful consideration and comments on this manuscript.

	FOOTNOTES

^* The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Supported by National Institutes of Health Grant RO1-AI380342-08.

^§ Supported by National Institutes of Health Grant F32-GM20724. To whom correspondence should be addressed: Cutaneous Biology Research Center, Massachusetts General Hospital East, Bldg. 149, 13th St., Charlestown, MA 02129. Tel.: 617-726-4445; Fax: 617-726-4453; E-mail: katia.georgopoulos@cbrc2.mgh.harvard.edu.

Published, JBC Papers in Press, January 17, 2002, DOI 10.1074/jbc.M111371200

² D. Kioussis, personal communication.

³ J. Koipally, unpublished observations.

	ABBREVIATIONS

The abbreviations used are: NuRD, nucleosome remodeling and deacetylation complex; PC-HC, pericentromeric heterochromatin; DBD, DNA binding domain; CAT, chloramphenicol acetyl transferase; Ik, Ikaros; GH, growth hormone; CTF, CCAAT-binding factor.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				S. Loeper, S. L. Asa, and S. Ezzat Ikaros Modulates Cholesterol Uptake: A Link between Tumor Suppression and Differentiation Cancer Res., May 15, 2008; 68(10): 3715 - 3723. [Abstract] [Full Text] [PDF]

				K. L. Kathrein, S. Chari, and S. Winandy Ikaros Directly Represses the Notch Target Gene Hes1 in a Leukemia T Cell Line: IMPLICATIONS FOR CD4 REGULATION J. Biol. Chem., April 18, 2008; 283(16): 10476 - 10484. [Abstract] [Full Text] [PDF]

				R. Sridharan and S. T. Smale Predominant Interaction of Both Ikaros and Helios with the NuRD Complex in Immature Thymocytes J. Biol. Chem., October 12, 2007; 282(41): 30227 - 30238. [Abstract] [Full Text] [PDF]

T. Saether, T. Berge, M. Ledsaak, V. Matre, A. H. Alm-Kristiansen, O. Dahle, F. Aubry, and O. S. Gabrielsen The Chromatin Remodeling Factor Mi-2{alpha} Acts as a Novel Co-activator for Human c-Myb J. Biol. Chem., May 11, 2007; 282(19): 13994 - 14005. [Abstract] [Full Text] [PDF]