Coordinated but physically separable interaction with H3K27-demethylase and H3K4-methyltransferase activities are required for T-box protein-mediated activation of developmental gene expression

Abstract

During cellular differentiation, both permissive and repressive epigenetic modifications must be negotiated to create cell-type-specific gene expression patterns. The T-box transcription factor family is important in numerous developmental systems ranging from embryogenesis to the differentiation of adult tissues. By analyzing point mutations in conserved sequences in the T-box DNA-binding domain, we found that two overlapping, but physically separable regions are required for the physical and functional interaction with H3K27-demethylase and H3K4-methyltransferase activities. Importantly, the ability to associate with these histone-modifying complexes is a conserved function for the T-box family. These novel mechanisms for T-box-mediated epigenetic regulation are essential, because point mutations that disrupt these interactions are found in a diverse array of human developmental genetic diseases.

Figure 1.

T-box domains 1 and 2 are defined by amino acid conservation and clustering of human genetic disease mutations. (A,B) Shown are two viewpoints of a model for the T-box domain cocrystalized with DNA. The model is extrapolated from the Tbx3 structure (Coll et al. 2002) for the T-box domain of T-bet. T-box domain 1 is shown in blue and contains the pocket of conserved amino acids that are mutated in several T-box protein-mediated human genetic diseases. (A) Indicated are the locations for mutations from Tbx5-mediated CHD (P178L, Y182C), Tbx3-mediated Ulnar mammary syndrome (L176P, Y182S), Tbx19-mediated ACTH deficiency (H181R, I271T), and Tbx22-mediated cleft palate (P225L, N226F) (Packham et al. 1995; Bamshad et al. 1997, 1999; Reamon-Buettner and Borlak 2004; Andreou et al. 2007; Vallette-Kasic et al. 2007). The amino acid number indicated is the corresponding position in murine T-bet. (A,B) Shown in green is the pocket of conserved amino acids termed T-box domain 2. DNA contact residues are shown in pink, and the DNA is shown in gray.

Figure 2.

T-box genetic disease mutants severely diminish T-bet’s activity at endogenous target genes. EL4 T cells were transfected with either a pcDNA3.1 control, wild-type T-bet, or mutant T-bet constructs that contain substitution mutations in T-box domain 1. The amino acid position as well as the one-letter symbol for the substituted amino acid are indicated in the key for each graph. A cell aliquot was harvested for RNA and qRT–PCR analysis of endogenous gene expression (A,C,E,G) or Western analysis to monitor construct expression levels from the same transfections to ensure equal expression levels for the mutant proteins (B,D,F,H). Experimental quantitation and normalization were performed as indicated in Supplemental Figure 1. Shown are representative experiments.

Figure 3.

T-box domain 1 mutations preferentially affect chromatin-mediated events. EL4 T-cells were transfected with an Ifnγ-promoter-reporter construct, TK-Renilla control vector, and either a pcDNA control, wild-type T-bet, or the T-box domain 1 mutant constructs as indicated. Cell aliquots from the same transfections were harvested either for luciferase analysis (A,C,E) or Western analysis (B,D,F) to monitor construct expression levels. Representative experiments are shown. (A,C,E) Graphical representation of the luciferase data with the Y-axis representing the ratio of the relative luciferase units compared with the relative transfection efficiency control Renilla units. (G) Shown are the quantitations from representative ChIP experiments examining the ability of the T-box domain 1 mutants to induce the H3K4me2 modification at the endogenous Cxcr3 and Ifnγ promoters. EL4 T cells were transfected with either a pcDNA control, wild-type T-bet, or the indicated T-box domain 1 mutant proteins and processed for standard ChIP analysis as indicated in Supplemental Figure 1. The ChIP samples were analyzed by qPCR using Cxcr3, Ifnγ, or Il4 promoter-specific primers.

Figure 4.

T-box domain 2 is required in chromatin-mediated events. EL4 T cells were transfected with either a pcDNA3.1 control, wild-type T-bet, or single or double mutants in T-box domain 2 as indicated. Representative experiments are shown. Transfections were harvested with one cell aliquot used for RNA and qRT–PCR analysis of endogenous gene expression (A,C) and the second cell aliquot subjected to a Western analysis to examine the level of the transfected protein expression construct (B,D) as described for Supplemental Figure 1. (E,F) Shown is an experiment examining the ability of the double alanine mutants in T-box domain 2 to up-regulate the activity of an Ifnγ-promoter-reporter construct as described in Figure 3. (G) Shown are the quantitations from representative ChIP experiments from EL4 T cells transfected with either single or double alanine mutant proteins in T-box domain 2. An antibody to the H3K4me2 modification or a nonspecific IgG control was used. Quantitative analysis of the ChIP samples with either Cxcr3 or Ifnγ promoter-specific primers was performed as described in Supplemental Figure 1.

Figure 5.

The physical interaction between T-bet and RbBp5 requires T-box domain 1, but not T-box domain 2. (A) T-bet specifically recruits RbBp5 to the Ifnγ and Cxcr3 promoters in primary Th1 cells. Shown is a representative ChIP analysis from primary CD4⁺ T cells isolated from either wild-type (black) or T-bet^−/− (gray) mice and skewed in Th1 conditions for 6 d. Antibodies to RbBp5, H3K4me2, or an IgG control were used, and a standardized aliquot of the total input chromatin was also processed as a control. The samples were subjected to a qPCR analysis with primers specific to the Cxcr3, Ifnγ, or Il4 promoters as described in Supplemental Figure 1. (B–F) Whole-cell extracts were prepared and immunoprecipitated with T-bet or V5 epitope tag-specific antibodies as indicated. The immunocomplexes were resolved by SDS-PAGE, and Western analysis with an RbBp5 antibody was performed. As a control for IP efficiency, all blots were stripped and reprobed with a T-bet- or V5-specific antibody (Supplemental Fig. 4). (B) EL4 T-cells were transfected with either a control pcDNA3.1 vector (lanes 1,3) or wild-type T-bet (lanes 2,4). (C) A co-IP experiment was performed from primary CD4⁺ Th1 whole-cell extracts. The extracts were immunoprecipitated with a control V5 (lane 2) or T-bet specific antibody (lane 3). (D) EL4 T-cells were transfected with either a pcDNA3.1 control (lanes 1,5), wild-type T-bet (lanes 2,6), T-bet Y182S (lanes 3,7), or T-bet L176P (lanes 4,8). (E) EL4 T-cells were transfected with either a pcDNA3.1 control (lanes 1,4), wild-type T-bet (lanes 2,5), or T-bet Q266A + R268A (lanes 3,6). (F) EL4 cells were transfected with either pcDNA (lanes 1,4), Brachyury (lanes 2,5), or an alanine mutant in amino acid position 88, corresponding to the conserved tyrosine residue in T-box domain 1 of Brachyury (lanes 3,6). (G) EL4 cells were transfected with either a pcDNA control (lanes 1,4), wild-type T-bet (lanes 2,5), or the T-bet Y182S mutant (lanes 3,6). The transfected samples were immunoprecipitated with a SET7/9 antibody (N-20) followed by Western analysis with a V5 antibody to detect the T-bet-transfected protein.

Figure 6.

T-bet functionally mediates the demethylation of the repressive H3K27me3 modification through T-box domains 1 and 2. (A,B) T-bet is required and sufficient for the reduction of the repressive H3K27me3 modification at the Cxcr3 and Ifnγ promoters. (A) CD4⁺ T cells were harvested from wild-type (black) or T-bet^−/− (gray) mice, skewed in Th1 conditions and then subjected to a standard ChIP with an antibody specific to the H3K27me3 modification or a nonspecific IgG control. The ChIP samples and standardized total input control were analyzed by qPCR analysis with Cxcr3, Ifnγ, or Il4 promoter-specific primers as described in Supplemental Figure 1. (B–D) EL4 T-cells were transfected with either a pcDNA3.1 control (light gray), wild-type T-bet (solid black), T-bet Y182S (medium gray), or T-bet Q266A + R268A (spotted back). The transfections were harvested for a standard ChIP analysis using an antibody specific to the H3K27me3 modification or a nonspecific IgG control antibody. The ChIP samples were analyzed as described in A. (E) T-bet associates with H3K27me3 demethylase activity. EL4 T-cells were transfected with either a pcDNA3.1 control (lane 1) or wild-type T-bet (lane 2). Whole-cell extracts were prepared from the transfected cells, immunoprecipitated with a T-bet-specific antibody, and subjected to a demethylase reaction. The samples were run on a SDS-PAGE gel, and a Western analysis was performed with an H3K27me3-specific antibody. As a control, the blots were then stripped and reprobed with an antibody to H3 or a T-bet-specific antibody. (F) EL4 cells were transfected with either pcDNA (lane 1), the T-box domain 1 mutant T-bet 182S (lane 2), or wild-type T-bet (lane 3) and subjected to an H3K27-demethylase assay as indicated in E. (G) EL4 cells were transfected with either a pcDNA control (lane 1), the T-box domain 2 mutant protein Q266A + R268A (lane 2), or wild-type T-bet (lane 3), and the samples were subjected to an H3K27-demethylase assay as described in E.

Figure 7.

T-bet target genes require JMJD3 expression for activity. (A–D) EL4 cells were transfected with either a pcDNA control or T-bet (A), Brachyury (B), Tbx5 (C), or eomesodermin (D). The samples were immunoprecipitated with either an antibody to T-bet (A) or the V5-epitope tag (B–D). The co-IPs were then subjected to an H3K27-demethylase assay as described in Figure 6. Aliquots of the demethylase reactions were run on separate gels and probed with either an H3K27me3-, H3K27me2-, or H3K27me1-specific antibody. Blots were then stripped and reprobed for either the wild-type protein or H3 as a loading control. The asterisk in B indicates the heavy chain, which runs close to Brachyury. (E) T-bet associates with JMJD3 in primary Th1 cells. Whole-cell extracts from primary CD4⁺ T cells polarized for 3 d in Th1 conditions were immunoprecipitated with two distinct T-bet-specific monoclonal antibodies (4B10 [lane 2] or 39D [lane 3]), a V5 control antibody (lane 1), or a CDK6 control antibody (lane 4). The immunocomplexes were subjected to SDS-PAGE and Western analysis with an antibody specific to the H3K27-demethylase JMJD3. As a control, the blots were stripped and reprobed with a T-bet-specific antibody. (F,G) T-bet-dependent activation of target genes requires JMJD3 expression. EL4 T cells were transfected with either a pcDNA3.1 control vector or wild-type T-bet and an siRNA to GFP or JMJD3. Transfected cells were harvested for RNA and qRT–PCR analysis (F) or Western analysis (G) to examine T-bet expression construct levels. Equal RNA/cDNA quantities were used in the qRT–PCR. A representative experiment is shown. Independent experiments with this and an additional JMJD3 siRNA are shown in Supplemental Figure 6.

Figure 8.

Model for the T-box-mediated recruitment of H3K27-demethylase and H3K4-methyltransferase activities to target promoters. Shown is a multistep model for T-box domain-mediated regulation of epigenetic events at target promoters. The experimental separation of the amino acid requirements for the physical interaction between T-bet and either H3K27-demethylase or H3K4-methyltransferase activities suggests that each step may be subject to context-dependent regulation, providing a high degree of complexity for the potential promoter and/or cell-type-specific epigenetic states established by the T-box family.