Fig 1.
Co-occupancy of genes and regulatory sequences by Nipped-B, cohesin, and the TBPH and Lark RNA-binding proteins in BG3 cells.
The genome browser tracks show the log2 ChIP-seq enrichment for cohesin (SA, green), Nipped-B (purple), Lark (brown), TBPH (orange) and Ser5P Pol II (transcriptionally-engaged paused and elongating Pol II, blue) at the string (cdc25) gene, and its upstream enhancers (red boxes). The two active string transcription start sites detected by PRO-seq [17] are indicated by pink arrowheads. Genome-wide Pearson correlation coefficients between Nipped-B, SA, TBPH and Lark ChIP-seq enrichment are in the table below the genome browser panel. The ChIP-seq enrichment for SA, Nipped-B and Ser5P Pol II are the average of two independent biological replicates (independent cell cultures, chromatin preparations and immunoprecipitation). The Lark ChIP-seq data is the average of five biological replicates, and the TBPH ChIP-seq is the average of four biological replicates. Each replicate was sequenced to ~10X genome coverage and normalized to input chromatin to calculate sequence enrichment in 250 bp sliding windows positioned 50 bp apart [24]. This procedure gives smoothened enrichment values at data points spaced 50 bp apart. As illustrated by the browser tracks, this method detects occupancy both in broad regions and in narrow peaks. It provides sensitivity equivalent to ChIP-chip for Nipped-B and cohesin [24]. Red lines in each browser track indicate the 95th percentile for enrichment for that protein, and the bars beneath each track show where enrichment is >95th percentile over regions >300 bp in length (>6 consecutive data points).
Fig 2.
TBPH and Lark bind enhancers, PREs, and the same active as Nipped-B and cohesin in BG3 cells.
The top diagram indicates how active promoters, extragenic enhancers, and PREs were defined for the analyses shown in the box and dot plots. Active promoters (light gray, PRO) are defined as 500 bp regions surrounding the transcription start sites (-250 to +250) that have at least 100 sequence reads in the published PRO-seq data [17] (GEO accession GSE42397). Excluding those that fall in heterochromatin, where ChIP-seq shows poor coverage, there are 7,389 active promoters. Enhancers (ENH) are defined as 500 bp elements centered at the peaks of DNase I hypersensitive sites that show H3K4me1 and H3K27ac histone modifications. This predicts 2,353 active enhancers, many of which correspond with known enhancers, such as those drive expression of string (Fig 1). The analysis performed here used all 523 extragenic enhancers in euchromatin positioned at least 500 bp outside of an actively transcribed gene (except non-coding RNAs of unknown function) [17]. Polycomb Response Elements (PREs) are defined as DNase I hypersensitive sites that show H3K4me1 and H3K27me3 histone modifications [18]. This identifies 195 PREs, including all known active PREs in BG3 cells, such as those silencing the vestigial gene (S2 Fig). As controls we also quantified occupancy of 7,389 gene body 500 bp segments starting 500 nt downstream of all active promoters (BOD, dark gray), and 6,892 randomly-positioned 500 bp regions (RAN, blue). The box plots show the log2 of the average ChIP-seq enrichment of all data points (usually 10) falling within each 500 bp region. An example of the method used to calculate the average enrichment at promoters is shown in S4 Fig. Red lines indicate where there is no ChIP-seq enrichment (log2 = 0, enrichment = 1). The PRO-seq box plots show the log2 of the number of sequence reads from both strands that fall within each 500 bp region as a measure of transcription. The dot plots below show the log2 average enrichment values at all active promoters for Nipped-B vs. TBPH, Nipped-B vs. Lark, and TBPH vs. Lark. Each point represents one of the 7,389 active promoters.
Fig 3.
TBPH and Lark depletion alter Nipped-B and cohesin occupancy at gene regulatory sequences in BG3 cells.
The browser tracks show the log2 enrichment for TBPH, Lark, Nipped-B, SA, and Nipped-B and SA after RNAi depletion of TBPH (iTBPH) and Lark (iLark) for the string gene and its upstream enhancers. The string enhancers that are most resistant to decreases in Nipped-B and SA occupancy upon TBPH depletion are marked with asterisks. The tracks shown are the average of two to five independent biological replicates. The boxplots show the distributions of the average log2 enrichment for all active promoters (gray), extragenic enhancers (yellow) and PREs (orange) in mock-treated (Mock) cells, and cells in which TBPH (iTBPH) or Lark (iLark) have been depleted by RNAi for four to five days. Gene regulatory elements are defined as shown in Fig 2. Asterisks indicate distributions that differ significantly from the mock control using a paired t test. The p values are given in the main text. A paired t test was used because each regulatory sequence is matched with itself in the control and depleted cells and the distributions are close to normal. All comparisons indicated as significant are also significant using the non-parametric Wilcoxon signed rank test. Plots of the values for all individual regulatory sequences in mock control versus the depleted cells are shown in S6 Fig.
Fig 4.
Nipped-B facilitates binding of TBPH and Lark to gene regulatory sequences in BG3 cells.
The browser view at the upper left shows log2 ChIP-seq enrichment tracks for Nipped-B in control cells, and for TBPH and Lark in control cells and cells depleted for Nipped-B (iNipped-B) at the string gene and its enhancers. The boxplots at the upper right show the distributions of ChIP-seq enrichment for TBPH and Lark at active promoters (PRO) extragenic enhancers (ENH) and PREs (PRE) in control (Mock) cells and cells depleted for Nipped-B (iNipped-B). Four biological replicates were averaged for the TBPH ChIP-seq and two for Lark ChIP-seq in Nipped-B depleted cells. Distributions that differ significantly from the Mock control by the paired t test are marked with asterisks, and the p values are given in the main text. All marked comparisons are also significant by the Wilcoxon signed rank test. For each regulatory sequence, the log2 enrichment values in Mock control cells (Mock log2 enrichment) are plotted against the enrichment values in cells depleted for Nipped-B (iNipped-B log2 enrichment) in the dot plots.
Fig 5.
Inhibition of transcription initiation with triptolide does not ablate binding of Nipped-B, TBPH and Lark to gene regulatory sequences in BG3 cells.
The boxplots show the average ChIP-seq enrichment of Ser5P Pol II, Nipped-B, TBPH and Lark at active promoters, extragenic enhancers, and PREs in control cells, and cells treated with 10 μM triptolide for 1, 2, and 4 hours. The genome browser tracks at the right show the log2 ChIP-seq enrichment for the same proteins at the string gene and enhancers over the same time course of triptolide treatment.
Fig 6.
Effects of triptolide treatment on Ser5P Pol II, Nipped-B, TBPH and Lark occupancy at active promoters, extragenic enhancers and PREs in BG3 cells.
The log2 ChIP-seq enrichment for Ser5P Pol II, Nipped-B, TBPH and Lark at all individual active promoters, enhancers and PREs in control untreated (Mock) cells is plotted against the enrichment after treatment of cells with 10 μM triptolide for 1, 2 and 4 hours.
Fig 7.
TBPH and Lark preferentially bind RNAs from cohesin-binding genes in vitro using known RNA-binding domains.
Diagrams of the TBPH and Lark protein structures with known sequence domains shown at the top: RRM, RNA Recognition Motif; ZnF, Zinc Finger. The smallest fragments tested that bind RNA in vitro are underlined in red. The bar graphs below the protein diagrams show examples of RNA-binding competition experiments with TBPH (left) and Lark (right). Whether or not the RNA is from a cohesin-binding gene or contains UG repeats is indicated (Y = yes, N = no). The sequences of all short RNAs tested from cohesin-binding and non-binding genes are in S1 Table, along with a summary of their abilities to bind TBPH and Lark derived from multiple experiments. As detailed in the text, soluble His6-SUMO fusion proteins were immobilized on NTA-Zn2+ agarose beads and incubated with equimolar mixtures of the indicated short RNAs. RNAs that were retained after washing were quantified by real-time PCR and binding was defined by enrichment relative to the amount of the CG6310-3 control RNA, which does not bind to either protein. All RNA fragments were tested in two independent binding experiments with freshly made fusion proteins, and each RNA was measured twice in each experiment. Error bars are the standard deviations of all measurements in all experiments. Enrichment of 10-fold or greater is defined as specific binding when recorded in S1 Table. The bottom bar graphs show example RNA-binding experiments with the indicated fragments of TBPH and Lark to map the domains that bind RNA. SDS-PAGE characterization of the immobilized protein fragments and the residues contained within each fragment are shown in S10 Fig. For truncated proteins, RNA enrichment greater than or equal to half the enrichment obtained with the full-length protein in the same experiment was defined as binding. With TBPH, the RRM1-containing region is necessary and sufficient for RNA-binding. For Lark, the zinc finger (ZnF) is required in addition to the RRM-containing domain. It is unknown if one or both of the Lark RRM domains is required to bind RNA.
Fig 8.
TBPH and Lark interact with Nipped-B.
(A) The western blots show the binding of Nipped-B and cohesin (Rad21) in BG3 cell nuclear extracts to NTA-Zn2+ agarose beads with immobilized His6-SUMO-TBPH and -Lark fusion proteins probed with both Nipped-B and Rad21 antibodies. The western on the far right (Input) shows the Nipped-B and Rad21 proteins in the BG3 cell nuclear extracts used for the binding experiments. Nuclear extracts were prepared from control (Mock) cells and cells depleted for Rad21 and Nipped-B by RNAi. The results shown for mock nuclear extract are representative of five independent experiments. The results shown for nuclear extracts depleted for Rad21 and Nipped-B are representative of two technical replicates. S12 Fig shows that native TBPH and Lark co-immunoprecipitate with Nipped-B from nuclear extracts, that Lark and TBPH do not co-precipitate, and that pre-treatment of BG3 nuclear extract with ribonucleases does not prevent binding of Nipped-B to TBPH and Lark beads. (B) Western blot of Nipped-B binding to the indicated TBPH fragments and full-length TBPH to map the protein domains interacting with Nipped-B. The bottom panel is a longer exposure of the same blot to show low levels of Nipped-B binding to some fragments. The fragments are shown in S10 Fig. Most of the Nipped-B binding occurs with the RRM1 domain of TBPH. The blots shown are representative of three independent experiments. (C) Western blot of Nipped-B binding to the indicated fragments of Lark and full-length Lark. The fragments are shown in S10 Fig. The blot shown is representative of three independent experiments.
Fig 9.
Hypothetical models for roles of TBPH and Lark in Nipped-B and cohesin binding to genes and enhancers, enhancer-promoter interactions, and processing of nascent RNA transcripts.
At promoters (top row) we posit that TBPH binds to UG repeats in the first nascent transcripts produced by elongating Pol II (Ser2P / Ser5P) when a gene is initially activated. TBPH then recruits Nipped-B, which interacts with DNA, loads cohesin and recruits the Lark RNA-binding protein. At enhancers (middle row) activator proteins (purple oval) recruit Mediator (large tan circle) and Nipped-B. Nipped-B then loads cohesin and recruits TBPH and Lark, and TBPH stabilizes binding of the complex. The protein complexes at the enhancers and promoters form enhancer-promoter complexes (bottom row) that are stable for hours even in the absence of new transcription initiation. TBPH contributes to their stability, while Lark destabilizes cohesin and Nipped-B binding, particularly to the promoter. P-TEFb in the enhancer-promoter complex phosphorylates paused Pol II and the pausing factors (not depicted), leading to transcriptional elongation (lower right). Some TBPH and Lark present in the enhancer-promoter complex binds to the nascent RNA produced by the elongating polymerase and can facilitate RNA processing. For example, they can regulate intron removal by splicing as depicted in the lower right diagram, in addition to other processes.