Multivalent engagement of TFIID to nucleosomes

Abstract

The process of eukaryotic transcription initiation involves the assembly of basal transcription factor complexes on the gene promoter. The recruitment of TFIID is an early and important step in this process. Gene promoters contain distinct DNA sequence elements and are marked by the presence of post-translationally modified nucleosomes. The contributions of these individual features for TFIID recruitment remain to be elucidated. Here, we use immobilized reconstituted promoter nucleosomes, conventional biochemistry and quantitative mass spectrometry to investigate the influence of distinct histone modifications and functional DNA-elements on the binding of TFIID. Our data reveal synergistic effects of H3K4me3, H3K14ac and a TATA box sequence on TFIID binding in vitro. Stoichiometry analyses of affinity purified human TFIID identified the presence of a stable dimeric core. Several peripheral TAFs, including those interacting with distinct promoter features, are substoichiometric yet present in substantial amounts. Finally, we find that the TAF3 subunit of TFIID binds to poised promoters in an H3K4me3-dependent manner. Moreover, the PHD-finger of TAF3 is important for rapid induction of target genes. Thus, fine-tuning of TFIID engagement on promoters is driven by synergistic contacts with both DNA-elements and histone modifications, eventually resulting in a high affinity interaction and activation of transcription.

Figure 1. H3K4_Cme3 nucleosomes bind endogenous TFIID and recombinant TAF3.

(A) Pulldown with the indicated biotinylated peptides using streptavidin coated beads incubated with GST-TAF3 PHD protein lysates. Proteins are visualized using Coomassie blue staining. (B) Immunoblot analysis of endogenous TAF5 binding to immobilized recombinant nucleosomes with the indicated MLA modification. Histones are visualized using Coomassie blue staining. (C) Workflow as applied for quantitative analysis of nucleosome interactors. In brief, heavy and light labeled extracts are used for pull-downs with immobilized, differentially modified nucleosomes. Experiments are also performed with a label swap. Eluted proteins are measured using LC-MS/MS. Enriched proteins in both experiments are selected based on box plot statistics. (D) Scatter plot of SILAC ratios for H3K4_Cme3 versus non-modified nucleosome interacting proteins. Upper right corner significant outliers are depicted and labeled based on the box plot analysis.

Figure 2. Presence of H2AZ, H3K9/K14ac and a TATA sequence enhances binding of endogenous TFIID to recombinant nucleosomes.

(A) Analysis of pulldowns with recombinant nucleosomes immobilized on magnetic streptavidin coated beads and GST-TAF3 (PHD). Proteins are visualized by silver stain. (B) Scatter plot of SILAC ratios for H3K4_Cme3/K27_Cme3 versus non-modified nucleosome interacting proteins. In the upper right corner significant outliers are depicted and labeled in grey based on box plots analysis. (C) Immunoblot analysis of recombinant nucleosomes with the indicated antibodies showing the presence of modifications or variants. (D) TFIID binds to H3K4_Cme3 nucleosomes independently of H2A.Z. Immunoblot analysis of eluted proteins using indicated antibodies. (E) TFIID binding is enhanced by histone H3 acetylation. Immunoblot analysis of eluted proteins using indicated antibodies. TAF5 antibody signal is quantified relative to the H3K_C4me3 pulldown. (F) Schematic representation of the nucleosome with the introduced TATA sequence and biotin group indicated. (G) Combination of TATA DNA and H3K14 acetylation on H3K4_Cme3 nucleosomes increases the interaction with TFIID. Immunoblot analysis of eluted proteins using indicated antibodies. TAF5 antibody signal is quantified relative to the H3K_C4me3 pulldown.

Figure 3. Stoichiometry analysis of endogenous TFIID.

(A) Immunoblot analysis of GFP-TAF3, GFP-TAF3 M882A and GFP-TAF5 expression after 24 hours of doxycycline induction with the indicated antibodies (right). Endogenous proteins and GFP-fusions are indicated on the left. (B) Identification of interacting proteins for GFP-TAF5 by volcano plot. The ratio of identified proteins in all fusion lines over wildtype in label-free quantification are plotted against the -log10 of the false discovery rate (FDR) calculated by a permutation-based FDR adapted t-test. In all experiments FDR was set to <0.05 and S₀ = 1.5. Significant outliers are labeled. (C) Bar graphs indicate the stoichiometry of TFIID subunits (indicated at bottom) relative to TAF6. Black dashed line indicates a ratio to the total TAF6 protein. Error bars indicate the standard deviations of the technical triplicate. (D) Identification of interacting proteins for GFP-TAF3 by volcano plot. (E) Bar graphs indicate the stoichiometry of TFIID subunits (indicated at bottom) relative to TAF6 in GFP-TAF3 (E) and GFP-TAF3 M882A (F) purifications.

Figure 4. TAF3 binding to ER stress gene promoters is dependent on the PHD-finger.

(A) Overlay of ENCODE H3K4me3 profiles from seven human cell lines at the GRP78 locus (upper panel). ChIP analysis of GFP-TAF3 and GFP-TAF3 M882A at GRP78 with the indicated primer sets (lower panel). Standard deviations represent technical triplicates and similar results were observed in at least three independent experiments. (B) Similar labeling as in (A) but for the CHOP locus. (C and D) Analysis of mRNA expression levels of GRP78 and CHOP by quantitative RT-PCR. Levels were normalized to β-ACTIN and are presented as change compared to a control DMSO-treated sample. Samples were analyzed 4 hours after tunicamycin treatment. Standard deviations represent two biological duplicates.