Nuclear import of the BTB/POZ transcriptional regulator Kaiso

Proteins of the BTB/POZ (broad complex, Tramtrack, Bric à brac/pox viruses and zinc fingers) transcription factor family regulate a range of biological processes in Drosophila and vertebrates (for reviews, see Albagli et al., 1995; Bardwell and Treisman, 1994). The earliest identified members in Drosophila have essential roles in development. For example, Bric à Brac controls patterning along the proximal-distal axis of the leg and antennae (Godt et al., 1993), and Abrupt modulates coordinated movement by directing the specificity of neuromuscular connections (Hu et al., 1995). In vertebrates, most BTB/POZ proteins (e.g. HIC-1, Bcl-6, PLZF, ZF5, MIZ-1) function as transcriptional regulators with putative oncogenic or tumor suppressor roles (Chang et al., 1996; David et al., 1998; Kaplan and Calame, 1997; Staller et al., 2001; Wales et al., 1995). This is best exemplified by B-cell-lymphoma 6 (Bcl-6), which directly promotes B-cell lymphomas and functions as a transcriptional repressor of various target genes including cyclin D1 and cyclin D2 (Dent et al., 2002; Kusam et al., 2004). In contrast to Bcl-6, no target genes have yet been identified for the putative tumor suppressor Hypermethylated-in-cancer 1 (HIC-1) but recent studies report that it is a sequence-specific transcriptional repressor in the context of artificial promoters (Pinte et al., 2004). Nevertheless, the mechanism of action of HIC-1 in human tumors includes its modulation by hypermethylation events and p53 expression, which render the HIC-1 gene transcriptionally inactive (Wales et al., 1995). Interestingly, whereas most BTB/POZ proteins function primarily as transcriptional repressors (e.g. Bcl-6 and PLZF) (Chang et al., 1996; David et al., 1998), others function as both activators and repressors (e.g. ZF5 and MIZ-1) (Kaplan and Calame, 1997; Bowen et al., 2002; Wu et al., 2003).

Kaiso is a ubiquitously expressed BTB/POZ transcription factor that was originally identified in a yeast-two-hybrid screen for binding partners of the multifunctional Armadillo-repeat containing protein p120-catenin (hereafter p120^ctn) (Daniel and Reynolds, 1999). Like other BTB/POZ proteins, Kaiso has the characteristic N-terminal protein-protein interaction POZ domain and it is through this domain that Kaiso homodimerizes and heterodimerizes with other BTB/POZ proteins (Daniel and Reynolds, 1999) (our unpublished data). In addition, Kaiso contains three C-terminal Kruppel-like C₂H₂ zinc fingers that mediate DNA binding (Daniel and Reynolds, 1999; Daniel et al., 2002). Structure-function analyses further revealed that only zinc fingers 2 and 3 mediate Kaiso DNA binding (Daniel et al., 2002). However, unlike other POZ transcription factors identified to date, Kaiso is a bimodal DNA-binding transcriptional repressor; Kaiso recognizes and binds a sequence-specific consensus site TCCTGCnA, where `n' is any nucleotide (Daniel et al., 2002), or methylated CpG dinucleotides in various promoters. These promoters include that of the tumorigenesis-associated genes matrilysin (our unpublished data), metastasin (Prokhortchouk et al., 2001) and MTA2 (Yoon et al., 2003), as well as the acetylcholine-receptor-clustering gene rapsyn (Rodova et al., 2004).

The identification of Kaiso as a p120^ctn-specific binding partner was somewhat unexpected considering that, at steady state, endogenous Kaiso localizes predominantly to the nucleus (Daniel et al., 2001; Daniel and Reynolds, 1999), whereas p120^ctn localizes both to cell-cell contacts, where it modulates E-cadherin activity (Ohkubo and Ozawa, 1999; Thoreson et al., 2000), and to the cytosol, where it regulates the activity of Rho GTPases (Anastasiadis et al., 2000; Franz and Ridley, 2004; Grosheva et al., 2001; Noren et al., 2000; Yanagisawa et al., 2004). However, we and others (van Hengel et al., 1999; Roczniak-Ferguson and Reynolds, 2003) had observed nuclear p120^ctn in some cell types and this led to our identification and characterization of a functional p120^ctn nuclear localization signal (NLS) (Kelly et al., 2004). Interestingly, the Kaisop120^ctn interaction maps to a region encompassing the Kaiso zinc-finger domain, and this is consistent with our recent demonstration that p120^ctn modulates Kaiso transcriptional activity by inhibiting Kaiso DNA binding (Daniel et al., 2002; Kelly et al., 2004). This p120^ctn-mediated inhibition of Kaiso function was contingent upon p120^ctn nuclear translocation and not the sequestering of Kaiso in the cytosol, as had been plausible (Kelly et al., 2004). More importantly, Kaiso-mediated transcriptional repression was effectively relieved in the presence of ectopic wild-type p120^ctn but not by ectopic p120^ctn harboring an NLS-inactivating double point mutation (Kelly et al., 2004). Nuclear p120^ctn thus appears to play a novel role as a regulator of Kaiso-mediated transcriptional repression.

Despite uncontested nuclear roles for Kaiso and other BTB/POZ proteins in transcriptional regulation, the precise mechanisms of nuclear import for these proteins remain unknown. Previous studies of Bcl-6 have shown that the BTB/POZ domain is dispensable for Bcl-6 nuclear localization, although this domain is required for targeting Bcl-6 to subnuclear dots (Dhordain et al., 1995). To definitively determine the mechanism of nuclear import for Kaiso, we used an established deletion mutagenesis approach (Sorg and Stamminger, 1999) and identified a highly basic ten-amino-acid NLS, upstream of the Kaiso zinc-finger domain. Fusion of this putative NLS to a heterologous β-galactosidase (β-gal) and green fluorescent protein (GFP) fusion protein efficiently directs the nuclear targeting of the fusion protein, thereby demonstrating functionality of this NLS. Minimal mutation of the identified NLS in the context of full-length Kaiso strongly inhibits Kaiso nuclear localization and ability to repress transcription of a promoter-reporter with four copies of the Kaiso binding site (KBS) controlling luciferase expression in artificial promoter assays. Our data provide the first reported mechanism of nuclear import for a BTB/POZ transcription factor and might help to elucidate how other POZ proteins traffic to the nucleus to regulate the expression of genes linked to development and tumor progression.

In this study, we used the human cervical carcinoma HeLa cell line and the mouse fibroblast cell line NIH 3T3. Both cell lines were grown at 37°C under 5% CO₂ in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 4 mM L-glutamine, 100 U ml^-1 penicillin, 100 μg ml^-1 streptomycin and 0.5 μg ml^-1 fungizone. Where indicated, cells were treated with 50 nM leptomycin B (LC Laboratories, Woburn, MA) for 18 hours.

To generate the various Kaiso deletion-mutant constructs, each cDNA was amplified using the polymerase chain reaction (PCR) from pBluescript-Kaiso (full-length murine). Each primer was designed to incorporate a 5′ SacII site and a 3′ XbaI site (underlined) into the PCR product for subsequent restriction digestion and ligation into the pHM829 plasmid to create an N-terminal fusion to β-gal and a C-terminal fusion to GFP (Sorg and Stamminger, 1999). Primers used were as follows: Kaiso-1 [nucleotides (nt) 256-738] using PS-1, 5′-GCATATCCGCGGATGGAGAGTAGAAAACTG-3′, and PS-2, 5′-GCATATTCTAGAAGCCCCATTATCTTGGGC-3′; Kaiso-2 (nt 687-1128) using PS-3, 5′-GCATATCCGCGGAAAAGCCTTGACATGGAA-3′, and PS-4, 5′-GCATATTCTAGATGGTGCAGTAAGTGGAGC-3′; Kaiso-3 (nt 1111-1575) using PS-5, 5′-GCATATCCGCGGGCTCCACTTACTGCACCA-3′, and PS-6, 5′-GCATATTCTAGAGTAAGTATCTTCACCGAT-3′; Kaiso-4 (nt 1558-1998) using PS-7, 5′-GCATATCCGCGGATCGGTGAAGATACTTAC-3′, and PS-8, 5′-GCATATTCTAGAAAGCTTTGAGTCCCCAGA-3′; Kaiso-5 (nt 1981-2271) using PS-9, 5′-GCATATCCGCGGTCTGGGGACTCAAAGCTT-3′, and PS-10, 5′-GCATATTCTAGACCAGTAAGATTCTGGTAT-3′. Nested deletions of the Kaiso-4 construct were generated similarly using the following primers: Kaiso-4A using PS-9 and PS-11, 5′-GCATATTCTAGAGCATACAATGCAAATATA-3′; Kaiso-4B using PS-12, 5′-GCATATCCGCGGAGGGTCTATTATATTTGC-3′, and PS-10. Kaiso-4A was further dissected into Kaiso-4A1 using primers PS-9 and PS-13, 5′-GCATATTCTAGAAGATGTTTTTGGTAGCTC-3′ and Kaiso-4A2 using PS-14, 5′-GCATATCCGCGGAAAACATCTGGCAGTGAG-3′ and PS-11. To evaluate NLS functionality, double-stranded oligonucleotides encoding the NLS and having flanking SacII and XbaI sites were ligated into pHM829 to generate β-gal/NLS/GFP fusion constructs.

pEGFP-C1-Kaiso was generated by PCR amplification of the full-length Kaiso cDNA from pBS-Kaiso (murine) followed by the ligation of this product into the BspEI and EcoRI sites of pEGFP-C1 (Clontech, Palo Alto, CA). Each ligation reaction was performed using T4 DNA ligase (NEB, Beverly, MA) according to the manufacturer's instructions. All oligonucleotides used were generated by the Central Facility of the Institute for Molecular Biology and Biotechnology (MOBIX) (McMaster University, Hamilton, ON, Canada). Each construct was verified by automated sequencing at MOBIX.

NLS-specific point mutations of pBS-Kaiso were generated using the Quikchange XL Site-Directed Mutagenesis Kit (Stratagene, Cedar Creek, TX) according to the manufacturer's protocols. The various mutated Kaiso cDNAs were then subcloned into the pEGFP-C1 plasmid (Clontech, Palo Alto, CA) to generate a panel of Kaiso constructs fused to enhanced GFP (eGFP) at their N-terminus. To generate Kaiso proteins fused C-terminally to the SV40 large T antigen (SV40LTag) NLS, we performed mutagenesis with primers encoding the SV40LTag NLS. We used pEGFP-C1-Kaiso as template and the Kaiso C-terminal region fused to this heterologous SV40LTag NLS was then subcloned into the pcDNA-Kaiso and pEGFP-C1-Kaiso plasmids. The nucleotide sequence of each construct was verified by automated fluorescence sequencing at MOBIX.

For each transfection experiment, cells were plated onto coverslips in six-well dishes at least 12 hours before transfection and incubated overnight (37°C, 5% CO₂), to have an approximate confluence of 60-70% at the time of transfection. Unless otherwise noted, all transfections were performed using the ExGen-500 reagent (MBI Fermentas, Bethesda, MD). For each construct, 2 μg of plasmid DNA was diluted in Opti-MEM serum-free medium (Gibco, Carlsbad, CA) or sterile 150 mM NaCl. 6 μl XGen-500 reagent was then added to the diluted plasmid and the mixture was gently vortexed to mix. The mixture was incubated without agitation at room temperature for 15-20 minutes to allow reagent-DNA complex formation before its addition to cells in supplemented DMEM (10% FBS, 1% penicillin/streptomycin, 4.0 mM L-glutamine). Unless otherwise noted, the cells were incubated with complexes for ∼3 hours (37°C, 5% CO₂), after which time the complexes were removed by aspiration and cells washed with 1× PBS. The cells were then incubated for 24 hours in supplemented DMEM (10% FBS, 1% penicillin/streptomycin, 4.0 mM L-glutamine) before fixation in 4% paraformaldehyde in PBS (pH 7) for 10 minutes at 4°C. Transfection of cells plated on 100 mm dishes was performed similarly, except that 20 μg plasmid and 60 μl Xgen-500 reagent were used.

24 hours after transfection, transfected cells were washed once in PBS, fixed in 4% paraformaldehyde in PBS on ice for 10 minutes and permeabilized for 5 minutes at room temperature with 0.2% Triton X-100 in PBS. Non-specific antibody binding was blocked by incubation with 3% skimmed milk powder in PBS for 10 minutes at room temperature. Cells were then incubated in primary anti-Kaiso polyclonal antibodies (Daniel et al., 2001) diluted 1/200 in 3% milk/PBS for 30 minutes at room temperature, followed by three washes for 5 minutes each with PBS. The cells were then briefly rinsed with 3% milk/PBS before incubation for 30 minutes at room temperature with Alexa-Fluor-594-conjugated (Molecular Probes, Eugene, OR) goat anti-rabbit secondary antibody at a dilution of 1/400 in 3% milk/PBS. The cells were finally washed three times for 5 minutes each with PBS before mounting onto glass slides with Aqua Poly/Mount (Polysciences, Warrington, PA). Cells were then imaged using a Zeiss Axiovert fluorescent microscope.

For all cell imaging (antibody stained, eGFP and GFP), cell samples were prepared and fixed as described above before the coverslips were mounted onto slides using Aqua Poly/Mount anti-fade solution (Polysciences). The coverslips were then sealed with clear nail polish (toluene and formaldehyde free) and visualized using epifluorescence microscopy on the Zeiss Axiovert 200 or the Zeiss LSM510 laser confocal microscope (Carl Zeiss, Thornwood, NY), as indicated.

24 hours before transfection, HeLa cells were plated on six-well dishes to achieve an approximate confluence of 60-70% at the time of transfection. All transfections were performed using XGen-500 transfection reagent (MBI Fermentas) according to the manufacturer's protocol. We used 400 ng reporter DNA (either pGL3 control or a 4×KBS-pGL3 control) and 1.6 μg of effector DNA (Kaiso-pcDNA3, Kaiso-NLSmut-pcDNA3, Kaiso-NLSmut-SVNLS-pcDNA3 or backbone vector alone). Cells were incubated with the transfection mix at 37°C in 5% CO₂ for ∼12 hours, washed twice with 2 ml PBS and incubated for 12 hours at 37°C in 5% CO₂ with fresh supplemented DMEM before luciferase assay analysis. Each well was treated with 350 μl passive lysis buffer (PLB) (Promega, Madison, WI) and the plates kept at room temperature for 20 minutes with vigorous shaking. Lysates were resuspended by pipetting and 20 μl lysate from each well was assayed for luciferase activity on a Lumat LB 9501 Berthold Luminometer (Fisher Scientific, Toronto, ON). Each experiment was performed in triplicate and all data are representative of the mean of at least two independent trials.

Cells were washed twice with 5 ml PBS (pH 7.4) followed by incubation on ice for 5 minutes with lysis buffer containing 0.5% NP-40, 50 mM Tris, 150 mM NaCl, 1 mM phenylmethylsulfonyl fluoride, 5 μg ml^-1 leupeptin, 2 μg ml^-1 aprotinin, 1 mM sodium orthovanadate and 1 mM EDTA. The cells were then harvested from the plates and transferred to a 1.5 ml Eppendorf tube. The lysate was centrifuged at 14,000 rpm for 5 minutes at 4°C and the supernatant transferred to a new tube. Lysates were quantified by Bradford assay and equal amounts of total protein were used for immunoprecipitation with anti-Kaiso monoclonal antibody 6F (Daniel et al., 2001) or with rabbit anti-GFP polyclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA). The immune complexes were then subjected to sodium-dodecyl-sulfate polyacrylamide-gel electrophoresis (SDS-PAGE) as previously described (Daniel and Reynolds, 1999). After SDS-PAGE, proteins were transferred from the gel to a nitrocellulose membrane using the Hoeffer semi-dry transfer apparatus (Amersham/Pharmacia, San Francisco, CA). The membrane was then briefly blocked at room temperature with 3% skimmed milk powder in TBS (pH 7.4) before incubating at 4°C overnight with rabbit anti-Kaiso polyclonal antibody at 1/11,000 dilution or mouse anti-GFP monoclonal antibody (B-2; Santa Cruz Biotechnology, Santa Cruz, CA) at 1/200 dilution in 3% milk/TBS. The primary antibodies were removed by rinsing the membranes five times with water and then once with TBS, for five minutes each. The membranes were then incubated for 2 hours at room temperature with horseradish-peroxidase-conjugated goat anti-rabbit secondary antibody (for Kaiso blots) or donkey anti-mouse secondary antibody (for GFP blots), diluted 1:40,000 in 3% milk/TBS. Membranes were finally rinsed five times with water and once with TBS (pH 7.4) for 5 minutes each, and processed using the enhanced chemiluminescence (ECL) system (Amersham/Pharmacia) according to the manufacturer's protocols.

All T7 promoter-containing constructs were in vitro transcribed and translated (IVTT) with the TnT-coupled Transcription and Translation kit as recommended by the manufacturer (Promega). 5 μl of each IVTT sample was analyzed by 7% SDS-PAGE gel electrophoresis and fluorography. Translated proteins were stored at -20°C.

Equal amounts of indicated GST-fusion proteins immobilized on glutathione-Sepharose 4B beads (Amersham Pharmacia) were incubated with 10 μl of ³⁵S-IVTT products in binding buffer (25 mM HEPES, pH 7.6, 100 mM NaCl, 10% glycerol, 2 μM ZnCl₂, 3 mM β-mercaptoethanol, 0.1% Nonidet P-40, 3.4 μg/ml phenylmethylsulfonyl fluoride and 10 μg/ml leupeptin) to a final volume of 100as incubated for 2 hours at 4°C, followed by the addition of 400 μl of binding buffer. The beads were then centrifuged at 4°C for 2 minutes at 1000 rpm and were washed three times with 500 μl of binding buffer. Proteins were eluted by boiling in SDS loading buffer and then subjected to electrophoresis on a 7% SDS-PAGE gel alongside 2.0 μl of IVTT (representing 20% of input). Gels were then fixed in fixing solution (25% isopropanol, 65% water, 10% acetic acid) for 30 minutes at room temperature, before treatment with Amplify solution (Amersham Pharmacia) for 15-30 minutes at room temperature. The gels were dried and exposed to film (Kodak X-AR) for 1-2 days at -80°C.

Kaiso localizes predominantly to the nucleus of most cells (Daniel et al., 2001; Daniel and Reynolds, 1999), where it is proposed to regulate the expression of genes through sequence-specific or methylation-dependent DNA elements (Daniel et al., 2002; Kelly et al., 2004; Prokhortchouk et al., 2001; Yoon et al., 2003). Because the molecular weight of Kaiso (at ∼100 kDa) surpasses the upper limit for passive diffusion through the nuclear pore (∼50 kDa), we postulated that Kaiso might possess one or more NLSs that facilitate its active nuclear import.

To identify regions of Kaiso that could serve as NLSs, we performed deletion mutagenesis analyses using an established in vivo assay designed specifically for the identification of NLSs (Sorg and Stamminger, 1999). We generated five deletion mutants of murine Kaiso spanning the entire open reading frame (Fig. 1A). These Kaiso fragments were simultaneously fused N-terminally to β-gal (which adds bulk and prevents passive diffusion through the nuclear pore) and C-terminally to GFP (which allows the subcellular localization of the mutant protein to be monitored). Controls for this assay included a β-gal/GFP fusion protein, which remained predominantly cytosolic, and a β-gal/SV40-NLS/GFP fusion protein, which localized exclusively to nuclei over repeated trials (Fig. 1B). Of our five initial Kaiso deletion mutants, only one reproducibly localized to the nuclei of HeLa cells; the remaining mutants localized exclusively to the cytosol (Fig. 1D). Interestingly, this deletion mutant (Kaiso-4, amino acids 432-581) completely encompassed the zinc-finger domain of Kaiso (amino acids 494-572). To map the Kaiso NLS more finely, we further divided the Kaiso-4 mutant into two smaller regions (Fig. 2A) to generate Kaiso-4A and Kaiso-4B. Kaiso-4A localized exclusively to nuclei, whereas Kaiso-4B was reproducibly targeted to cytosolic aggregates (Fig. 2B). We further divided the Kaiso-4A mutant into Kaiso-4A1 and Kaiso-4A2 (Fig. 2C). Using this approach, we delineated the NLS-containing region to 30 amino acid residues upstream of the zinc-finger domain within the Kaiso-4A2 mutant. Within this region, we noted the highly basic sequence PPNKRMKVKH with the potential to serve as an NLS. Moreover, this region of Kaiso, encoded by amino acids 469-478, was highly conserved across diverse species such as Musmusculus, Homo sapiens and Xenopus laevis, which further highlighted its potential to regulate Kaiso nuclear translocation (Fig. 2D).

To test this region directly for NLS functionality, we generated an expression construct in which these ten residues were flanked at their N-terminus with β-gal and at their C-terminus with GFP. This construct was then transfected into HeLa and NIH 3T3 cells, and its subcellular localization monitored by fluorescence microscopy. Quantification of these results and subsequent NLS characterization experiments are summarized in Table 1. As seen in Fig. 2E and Table 1, this ten-amino-acid sequence was sufficient to target the β-gal/GFP fusion protein to 96% of HeLa nuclei. However, a mutant NLS-containing fusion protein, harboring a lysine-to-alanine point mutation at a crucial residue within the NLS (amino acid 472), localized predominantly to the cytosol in 65% of transfected cells, and equally between the nucleus and cytosol in 35% of transfected cells (Fig. 2F). This indicates the necessity of basic residues for Kaiso NLS functionality. To verify that the NLS had targeted the entire heterologous fusion protein (i.e. β-gal/NLS/GFP) to the nucleus and that the C-terminal GFP moiety had not been cleaved, transfected cells were stained with 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside (X-gal) and β-gal enzymatic activity was detected in the nucleus (data not shown). This observation confirmed that the fusion protein in question was intact and that we were indeed detecting NLS-dependent nuclear targeting of entire β-gal/NLS/GFP fusion proteins, as opposed to the passive nuclear diffusion of the GFP moieties.

To confirm that we had not overlooked another NLS in any of our initial five Kaiso mutants, perhaps owing to the presence of an out-competing nuclear export signal (NES), we monitored the sensitivity of each mutant to leptomycin B (LMB), a specific inhibitor of CRM-I-mediated nuclear export (Kudo et al., 1998). As seen in Fig. 3B, none of the five Kaiso deletion mutants were LMB sensitive, indicating that none of these mutants encoded a functional CRM-I-dependent NES that out-competed a weaker NLS. Although we cannot exclude the possibility that our NLS assay overlooked an unconventional NLS, our data indicate that amino acids 469-478 of Kaiso compose a functional NLS that is capable of facilitating the nuclear localization of heterologous proteins (e.g. β-gal/GFP).

It is currently unclear whether Kaiso encodes a functional NES and hence shuttles from the nucleus to the cytosol. Although a cytosolic function for Kaiso has yet to be reported, Kaiso might traffic transiently to the cytosol to perform as-yet-undiscovered functions. In support of this hypothesis, we have identified several cytoskeletal proteins as potential Kaiso binding partners through a yeast-two-hybrid screen (our unpublished data). Nonetheless, the NLS we identified that drives the nuclear import of the Kaiso might be a dominant signal because, in all the cell lines tested, Kaiso localizes almost exclusively to the nucleus (our unpublished data). Hence, should Kaiso possess a NES, it might be difficult to detect owing to Kaiso's predominant nuclear localization.

To verify that the identified NLS was crucial for the nuclear translocation of full-length Kaiso, we generated an eGFP/Kaiso construct encoding a single point mutation in the putative NLS (K472A). We then monitored the localization of this eGFP/Kaiso fusion protein in fixed cells using confocal microscopy. Consistent with studies on endogenous Kaiso (Daniel et al., 2001; Daniel and Reynolds, 1999), wild-type eGFP/Kaiso localized predominantly to 99.5% of HeLa nuclei (Fig. 4A). Similar results were obtained in human fibroblast Va-2 (Fig. 4A), rat fibroblast Rat-1, monkey fibroblast Cos-1 and human colon carcinoma HCT-116 cells (data not shown). By contrast, the K472A Kaiso NLS mutant (eGFP/Kaiso-NLSmut) demonstrated a pronounced localization to the cytosol in 63% of HeLa cells, and equally to the cytosol and nucleus in the remaining 37% of cells. This confirmed the relevance and functionality of the identified NLS. The incomplete nuclear exclusion of eGFP/Kaiso-NLSmut that we observed could be attributed to the relatively conservative single point mutation (K472A) within the NLS of this Kaiso construct; this mutant might retain some capacity to interact with import factors in vivo.

To verify that our eGFP/Kaiso and eGFP/Kaiso-NLSmut constructs encoded full-length Kaiso, we transiently transfected HeLa cells with these constructs and stained transfected cells with our Kaiso rabbit polyclonal antibodies (Fig. 4B), and performed immunoprecipitation and immunoblot analysis (Fig. 4C). As seen in Fig. 4B, our Kaiso-specific antibodies (Daniel et al., 2001) recognized the eGFP/Kaiso fusion proteins, thereby confirming that each fusion protein was intact and in frame. This also verified that the eGFP moiety had not been cleaved and that the resulting nuclear fluorescence was not due to passive diffusion of eGFP into the nucleus. Moreover, immunoprecipitation and westernblot analysis with anti-GFP antibodies confirmed the correct molecular weight of the eGFP/Kaiso and eGFP/Kaiso-NLSmut fusion proteins (Fig. 4C). Collectively, these data show that the identified NLS is indeed necessary and crucial for the nuclear translocation of full-length Kaiso.

Our detection of a basic monopartite NLS in Kaiso suggested that Kaiso might bind the conventional nuclear import receptors (e.g. Importin-α) to achieve nuclear entry. This pathway (for a review, see Kau et al., 2004) initiates with the binding of Importin-α to the NLS of a substrate protein. Importin-β then binds Importin-α and this tripartite complex enters the nuclear pore complex and is imported into the nucleus. The subsequent binding of Ran-GTP to Importin-β triggers the dissociation of the complex, and the substrate protein is released into the nucleoplasm. Importin-α is then recycled back to the cytoplasm by the export receptor CAS and is equipped for another round of import.

To ascertain whether Kaiso interacts directly with the nuclear import receptor Importin-α2 (Rch1), we performed glutathione-S-transferase (GST) pulldown and co-immunoprecipitation experiments. As seen in Fig. 5A, wild-type but not NLS-defective Kaiso, bound efficiently to immobilized GST/Importin-α2 but not to immobilized GST alone. To determine whether Kaiso and Importin-α2 form a complex in vivo, we performed coimmunoprecipitation experiments using HCT116 (human colon carcinoma) cells, which express high endogenous levels of both Kaiso and Importin-α2 (Fig. 5B). Using our highly specific Kaiso monoclonal antibodies, Importin-α2 was detected in Kaiso immunoprecipitates but not in p120^ctn or negative-control (12CA5) immunoprecipitates. These data demonstrate the direct association between Kaiso and Importin-α2, and further suggests that the nuclear import of Kaiso, as directed by its basic NLS, uses a classical Importin-α2-dependent import pathway.

We and others have previously demonstrated that Kaiso functions as a sequence-specific and a CpG-methylation-dependent transcriptional repressor (Kelly et al., 2004; Kim et al., 2002; Prokhortchouk et al., 2001; Yoon et al., 2003). Our in vitro studies, however, demonstrated that Kaiso has a higher affinity for the sequence-specific consensus site than for the methylated site (Daniel et al., 2002). To test further the effect of the Kaiso NLS mutant on Kaiso gene regulatory function, we performed artificial promoter assays in HeLa cells using a reporter with four reiterated copies of the KBS (4×KBS) upstream of a luciferase gene. Wild-type and NLS-defective Kaiso were co-transfected with the pGL3-4×KBS-luciferase reporter plasmid and, 24 hours after transfection, luciferase activity was quantified. Consistent with previous reports, endogenous Kaiso repressed transcription from the 4×KBS-luciferase reporter by approximately 1.6 times. This observed repression, in the absence of ectopic effectors, was attributed specifically to Kaiso, because Kaiso-specific small interfering RNAs abolished this effect altogether (data not shown). Overexpression of wild-type Kaiso further repressed transcription fivefold from the artificial 4×KBS promoter (Fig. 6A). This was in contrast to NLS-defective Kaiso, which repressed transcription negligibly despite its expression at similar levels to wild-type Kaiso (Fig. 6B). Fusion of the heterologous SV40-LTAg NLS to the C-terminus of NLS-defective Kaiso fully restored the repressive effect of this mutant effector, demonstrating that the lack of repression by NLS-defective Kaiso was indeed due to its decreased nuclear translocation. Importantly, none of the assayed Kaiso effectors altered luciferase activity of a 4×KBS-luciferase reporter harboring a cytosine-to-adenine mutation in the KBS; this mutation was previously shown to abolish Kaiso-DNA interactions in vitro (Daniel et al., 2002). Collectively, these findings demonstrate the functional relevance of the Kaiso NLS and are consistent with the hypothesis that the nuclear translocation of the transcription factor Kaiso is a necessary prelude to its gene regulatory role.

In summary, we report here the identification and characterization of a basic, monopartite NLS in the BTB/POZ transcription factor Kaiso. Whereas the identified Kaiso NLS was necessary and sufficient to target a heterologous β-gal/GFP fusion protein to nuclei, a single-amino-acid inactivating mutation abolished this effect. This NLS thus appears to be the primary regulator for the nuclear translocation of full-length Kaiso, because a single point mutation of a key positively charged lysine (K472) resulted in a marked redistribution of Kaiso to the cytosol of cells. The approach used in this study to identify the Kaiso NLS might be amenable to the identification of NLSs in other BTB/POZ proteins. Ongoing work is aimed at identifying and characterizing the full scope of genes regulated by Kaiso, in order to more comprehensively to understand the cellular role of this novel BTB/POZ transcription factor in tumorigenesis.

We thank R. Truant for the pHM829 plasmid and his technical expertise. We also thank A. Shaw, P. Sood and C. Spring for experimental assistance. This work was supported by a DOD Breast Cancer IDEA grant (DAMD17-02-1-0479) and a CIHR grant (MOP-42405) to J.M.D. K.F.K. is the recipient of an NSERC D2 Doctoral scholarship. A.A.O. is the recipient of an NSERC Master's scholarship.