Figure 1.
ZBTB family members interact with MTG16.
A. ZBTB family member alignment based on homologous zinc fingers and identification of amino acids used in yeast-two hybrid experiments (yeast two-hybrid clones). B. Yeast-two hybrid assay for interactions between ZBTB family members and MTG16. C. Mapping of the Kaiso binding site on MTG16. Yeast two-hybrid assays were performed in triplicate.
Figure 2.
Kaiso and MTG16 associate and co-localize.
A. Immunoprecipitation of overexpressed Gal-tagged MTG family members in HCT116 cells. Cell lysates were fractionated and immunoprecipitated with anti-Gal or control IgG and analyzed by immunoblotting with Kaiso antibody. B. Myc-tagged MTG16 constructs used for immunoprecipitation mapping of MTG16 binding site. C. Immunoprecipitation of overexpressed Myc-tagged MTG family members in HCT116 cells. Cell lysates were fractionated and immunoprecipitated with anti-Myc or control IgG and analyzed by immunoblotting with Kaiso antibody. D. K562 cells costained for MTG16 (333, FITC) and Kaiso (polyclonal, ABR, Cy3) were imaged using the Deltavision system, and a correlation coefficient was determined for Kaiso and MTG16 colocalization (r = 0.95±0.04, n = 25 cells).
Figure 3.
MTG family members repress the Kaiso binding site reporter (4×KBS).
A. Composition of the KBS reporter constructs. The 4×KBS reporter contains four consensus Kaiso binding site sequences followed by a luciferase gene. The mutant reporter (4×KBSMT) contains a C→A transversion in the Kaiso binding sites which abolishes Kaiso binding. B. 4×KBS artificial promoter assays in HCT116 cells upon titration of MTG family members (0 ng, 200 ng, 400 ng, 800 ng). The graph shows the fold-change in luciferase activity relative to the control reporter, pGL4-TK hRLUC, after transfection of expression plasmids encoding MTG family members, MTG8, MTG16, and MTGR1. The error bars represent the standard error of the mean for 3 replicate experiments performed in triplicate. C. 4×KBS and 4×KBSMT artificial promoter assays in HCT116 cells after transfection of 500 ng of the indicated MTG family member. The error bars represent the standard error of the mean for 4 replicate experiments performed in triplicate. D. Kaiso knockdown by two independent Kaiso shRNA constructs compared to a scrambled shRNA control (left). 4×KBS artificial promoter assay after knockdown of Kaiso and the transfection of 500 ng of MTG16 (right). E. Methylated pSV40 artificial promoter assay in Mbd2−/− cells upon titration of MTG family members (0 ng, 200 ng, 400 ng, 800 ng). The graph shows the fold-change in luciferase activity relative to the control reporter, pGL4-TK hRLUC, after transfection of expression plasmids encoding MTG family members, MTG8, MTG16, and MTGR1. The error bars represent the standard error of the mean for 3 replicate experiments performed in triplicate. F. Model for Kaiso repression of Kaiso binding site (KBS) and methyl CpGs (mCpG) on target promoters. *P<0.05, **P<0.01, ***P<0.001.
Figure 4.
MTG family members depend on Kaiso to repress the human MMP-7 promoter.
A. The composition of the full-length MMP-7 reporter (HMAT-2.3). The HMAT-2.3 reporter contains two Kaiso and two TCF binding sites followed by a luciferase gene. B. HMAT-2.3 artificial promoter assays in HCT116 cells upon titration of MTG family members (0 ng, 200 ng, 400 ng, 800 ng). The graph shows the fold-change in luciferase activity relative to the control reporter, pGL4-TK hRLUC, after transfection of expression plasmids encoding MTG family members, MTG8, MTG16, and MTGR1. The error bars represent the standard error of the mean for 3 replicate experiments performed in triplicate. C. HMAT-2.3 promoter assay utilizing MTG16 deletion constructs in which the indicated nervy homology region is removed. Error bars represent the standard deviation of triplicate samples. D. HMAT-2.3 artificial promoter assay after knockdown of Kaiso and the transfection of 500 ng of MTG16. Error bars represent the standard deviation of triplicate samples. *P<0.05, ***P<0.001.
Figure 5.
MTG family members repress endogenous MMP-7.
A. Primers for the ChIP assay flank the −537 Kaiso binding site in the MMP-7 promoter. B. ChIP assay performed in HCT116 cells using both polyclonal (333, pMTG) and monoclonal (2D1, mMTG16) antibodies for MTG16. The graph shows fold-enrichment compared to IgG control. C. Knockdown of Kaiso by three independent shRNA constructs compared to a scrambled sh-RNA control. D. Chromatin immunoprecipitation assay in HCT116 cells. The graph shows fold enrichment of the −537 Kaiso binding site after precipitation with the MTG16 2D1 antibody compared to the IgG control. E. MMP-7 mRNA expression after transfection of the indicated MTG family member in HT29 cells. The graph shows the fold-change (ΔΔCt) of MMP-7 mRNA compared to an empty control. Error bars represent the standard error for three replicate experiments performed in triplicate (top). Representative immunoblotting of MMP-7 in HT29 cells after transfection of indicated family members and treatment with Brefeldin-A. Protein quantification was performed with the Odyssey Western blot developer system. **P<0.01, ***P<0.001.
Figure 6.
MTG16 and Kaiso are negatively correlated in human CRC patient samples and MMP-7 is up-regulated.
A) Kaiso expression is significantly up-regulated in adenomas and colorectal cancer patients compared with normal adjacent colon tissues (P<0.001, for all comparisons). B) MMP-7 expression is significantly up-regulated in adenoma and colorectal cancer patients for each stage (P<0.001 for all comparisons). C) MTG16 expression is significantly down-regulated in colorectal cancers versus normal adjacent samples (P<0.001; *for all comparisons except adenoma). Wilcoxon Rank Sum test was used to determine significance for A–C. D) Normalized expression of both MTG16 and Kaiso probes (see Methods) were compared in human colorectal cancer patients. The graph demonstrates a significant inverse correlation for MTG16 versus Kaiso (rho = −0.29, P<0.001, inverse correlation depicted by red line, Spearman correlation coefficient).