E Kaiso-DNA interaction (Figure 4A). The +69 CMUT1 (mutated one 39 CpG to GG but with intact KBS), +69 CMUT2 (mutated the two 59 CpGs to GGs, with intact KBS and 39 CG), +69 CMUT3 (mutated all three CpG sites to GGs but with intact KBS) and +69 ALLMUT (mutated all three CpGs and the KBS) methylated probes were incubated with GST-Kaiso-DPOZ fusion proteins. GST-Kaiso-DPOZ bound the methylated +69 KBS-mutprobe similarly to that of the +69 CMUT1 probe, but with lower affinity than the wild type probe (Figure 4B, compare lanes 6 9 to 3). Since Kaiso did not bind the +69 CMUT2, +69 CMUT3 or +69 ALLMUT probes (Figure 4B, lanes 10?8), this suggests that the two CpG sites immediately SM5688 custom synthesis upstream of the KBS are necessary for Kaiso binding to the cyclin D1-promoter-derived oligonucleotides and supports our 59-azacytidine ChIP experiment (Figure 3B). Taken together, our data suggest that Kaiso’s binding to the +69 KBS region is methyl-CpG-dependent and not KBS-specific. We further confirmed the specificity of Kaiso binding to the methylated +69 core KBS probe via cold competition assays with excess unlabelled probes (data not shown).Figure 3. Kaiso binds the +69 25331948 core KBS region of the cyclin D1 promoter in vitro and in vivo. (A) EMSA revealed that Kaiso bound the methylated cyclin D1+69 KBS promoter region but not the unmethylated +69 KBS probe. Kaiso also bound weakly to the methylated (but KBS mutated) +69 KBS probe compared to the wild type probe. (B) ChIP of the cyclin D1 promoter in HCT 116 and MCF7 cells revealed that Kaiso specifically associated with the cyclin D1 promoter +69 KBS region. 59-azacytidine treatment of MCF7 cells abolished Kaiso’s association with the cyclin D1 promoter and suggests methyl-CpG-dependent binding of Kaiso to the promoter. doi:10.1371/journal.pone.0050398.gKaiso Represses cyclin D1 via KBS and Me-CpG SitesFigure 4. Kaiso binds the +69 core KBS region of the cyclin D1 promoter in a methyl-CpG-specific IPI-145 manner. (A) Summary of Kaiso binding to wild type and mutated +69 core KBS cyclin D1-derived oligonucleotides. The CpGs (red) and KBS (blue) sites are highlighted and the mutations are underlined. (B) EMSA showed that Kaiso binding to the cyclin D1+69 KBS promoter region requires at least two intact methyl-CpG dinucleotides but not an intact KBS site. doi:10.1371/journal.pone.0050398.gKaiso Represses cyclin D1 via KBS and Me-CpG SitesKaiso Represses Transcription from the cyclin D1 Minimal Promoter in a KBS-specific MannerAfter determining that Kaiso bound the cyclin D1 promoter region with dual-specificity (i.e. via the sequence-specific KBS and via methyl-CpG sites), we next assessed Kaiso’s ability to regulate luciferase expression under control of a minimal cyclin D1 promoter. Transfection of MCF7 cells with the unmethylated cyclin D1 promoter-reporter (21748 CD1), containing two KBSs and multiple CpG sites, resulted in an ,35-fold increase in luciferase reporter activity compared to the pGluc-Basic negative control vector lacking the cyclin D1 promoter region (Figure 5A). Co-transfection of the 21748 CD1 promoter-reporter and a Kaiso expression plasmid abrogated this response and resulted in a dosedependent decrease in luciferase activity (Figure 5A). A similar trend was observed in HCT 116 cells (data not shown). To confirm that transcriptional repression was attributed to Kaiso, we depleted endogenous Kaiso with Kaiso-specific siRNA. Increasing amounts of Kaiso-specific siRNA resulted in dose-dependent derepre.E Kaiso-DNA interaction (Figure 4A). The +69 CMUT1 (mutated one 39 CpG to GG but with intact KBS), +69 CMUT2 (mutated the two 59 CpGs to GGs, with intact KBS and 39 CG), +69 CMUT3 (mutated all three CpG sites to GGs but with intact KBS) and +69 ALLMUT (mutated all three CpGs and the KBS) methylated probes were incubated with GST-Kaiso-DPOZ fusion proteins. GST-Kaiso-DPOZ bound the methylated +69 KBS-mutprobe similarly to that of the +69 CMUT1 probe, but with lower affinity than the wild type probe (Figure 4B, compare lanes 6 9 to 3). Since Kaiso did not bind the +69 CMUT2, +69 CMUT3 or +69 ALLMUT probes (Figure 4B, lanes 10?8), this suggests that the two CpG sites immediately upstream of the KBS are necessary for Kaiso binding to the cyclin D1-promoter-derived oligonucleotides and supports our 59-azacytidine ChIP experiment (Figure 3B). Taken together, our data suggest that Kaiso’s binding to the +69 KBS region is methyl-CpG-dependent and not KBS-specific. We further confirmed the specificity of Kaiso binding to the methylated +69 core KBS probe via cold competition assays with excess unlabelled probes (data not shown).Figure 3. Kaiso binds the +69 25331948 core KBS region of the cyclin D1 promoter in vitro and in vivo. (A) EMSA revealed that Kaiso bound the methylated cyclin D1+69 KBS promoter region but not the unmethylated +69 KBS probe. Kaiso also bound weakly to the methylated (but KBS mutated) +69 KBS probe compared to the wild type probe. (B) ChIP of the cyclin D1 promoter in HCT 116 and MCF7 cells revealed that Kaiso specifically associated with the cyclin D1 promoter +69 KBS region. 59-azacytidine treatment of MCF7 cells abolished Kaiso’s association with the cyclin D1 promoter and suggests methyl-CpG-dependent binding of Kaiso to the promoter. doi:10.1371/journal.pone.0050398.gKaiso Represses cyclin D1 via KBS and Me-CpG SitesFigure 4. Kaiso binds the +69 core KBS region of the cyclin D1 promoter in a methyl-CpG-specific manner. (A) Summary of Kaiso binding to wild type and mutated +69 core KBS cyclin D1-derived oligonucleotides. The CpGs (red) and KBS (blue) sites are highlighted and the mutations are underlined. (B) EMSA showed that Kaiso binding to the cyclin D1+69 KBS promoter region requires at least two intact methyl-CpG dinucleotides but not an intact KBS site. doi:10.1371/journal.pone.0050398.gKaiso Represses cyclin D1 via KBS and Me-CpG SitesKaiso Represses Transcription from the cyclin D1 Minimal Promoter in a KBS-specific MannerAfter determining that Kaiso bound the cyclin D1 promoter region with dual-specificity (i.e. via the sequence-specific KBS and via methyl-CpG sites), we next assessed Kaiso’s ability to regulate luciferase expression under control of a minimal cyclin D1 promoter. Transfection of MCF7 cells with the unmethylated cyclin D1 promoter-reporter (21748 CD1), containing two KBSs and multiple CpG sites, resulted in an ,35-fold increase in luciferase reporter activity compared to the pGluc-Basic negative control vector lacking the cyclin D1 promoter region (Figure 5A). Co-transfection of the 21748 CD1 promoter-reporter and a Kaiso expression plasmid abrogated this response and resulted in a dosedependent decrease in luciferase activity (Figure 5A). A similar trend was observed in HCT 116 cells (data not shown). To confirm that transcriptional repression was attributed to Kaiso, we depleted endogenous Kaiso with Kaiso-specific siRNA. Increasing amounts of Kaiso-specific siRNA resulted in dose-dependent derepre.