Figures and data
TFIIIC directly interacts with MYCN.
A. Immunoblots showing levels of FLAG-tagged MYCN (amino acids 2-137) and the six subunits of the TFIIIC complex after a pulldown assay using anti-FLAG affinity columns. Multiple columns labelled “Wash” represent the sequential washings (n = 2).
B. Size exclusion chromatography graph of MYCN (aa 1-137)/1A (black trace) or MYCN alone (red trace). The blue box marks the fractions used for panel C and D (n = 2).
C. Coomassie staining of fractions of the MYCN (aa 1-137)/1A complex (fractions marked with blue box in panel B).
D. Immunoblot of fractions of the MYCN (aa 1-137)/1A complex (fractions marked with blue box in panel B).
E. Growth curve (measured as % confluence) of SH-EP-MYCN-ER cells expressing Dox-inducible shRNA targeting TFIIIC2, TFIIIC3 or TFIIIC5 under the indicated conditions. Data show mean ± standard deviation (SD) (n = 3).
MYCN and TFIIIC antagonize accumulation of non-phosphorylated RNAPII.
A. Browser tracks for non-phosphorylated RNAPII (top) and RNAPII pSer2 (bottom) ChIP-Rx at the indicated gene loci. SH-EP-MYCN-ER cells were treated with Dox (1 µg/ml, 48 h) and/or 4-OHT, respectively. EtOH was used as control.
B. Average density plot of ChIP-Rx signal for non-phosphorylated RNAPII. Data show mean (line) ± standard error of the mean (SEM indicated by the shade) of different gene sets based on an RNA-seq of SH-EP-MYCN-ER cells ± 4-OHT. The y-axis shows the number of spike-in normalized reads and it is centered to the TSS ± 2 kb. N = number of genes in the gene set defined in the methods (n = 2).
C. Density plot of ChIP-Rx signal for RNAPII pSer2 as described for panel B. The signal is centered to the TES ± 2 kb (n = 2).
D. Average bin dot plot showing fold change for RNAPII pSer2 ChIP-Rx reads over TES ± 2 kb and RNA-seq of SH-EP-MYCN-ER for the same genes ± MYCN + TFIIIC5 (blue) or + MYCN ± TFIIIC5 (red). The plot shows 20 bins representing a total of 13,239 and 12,330 genes for ± MYCN + TFIIIC5 and + MYCN ± TFIIIC5 datasets, respectively (n = 3 for RNA-seq, n = 2 pSer2 RNAPII ChIP-Rx).
E. Average bin dot plot for RNA-seq of SH-EP-MYCN-ER showing log2 mRNA expression normalized by control per bin. Cells were treated with 1 µg/ml Dox (“– TFIIIC5”, 48 h) and/or 4-OHT (“+ MYCN”, 4 h) or EtOH as control. Expression was normalized by its control. Each bin represents 150 genes of a total of 14,085 genes. Dotted line marks the relative expression at 0 (n = 3).
F. Density plot of ChIP-Rx signal for TFIIIC5. Data show mean (line) ± SEM (shade) for 14,722 genes. The signal is centered to the TSS ± 2 kb (n = 2).
MYCN is part of three-dimensional promoter hubs.
A. Top: Representative browser track of MYCN three-dimensional chromatin interactions. Height shows the number of paired end tags (PETs) indicating the interaction intensity and the width of the line shows the start and end positions of each anchor. Middle and bottom: Browser tracks showing the number of reads of MYCN and total RNAPII ChIP-Rx, respectively. Unless stated, all experiments were performed in SH-EP-MYCN-ER cells treated with 4-OHT (200 nM, 4 h). The ruler at the bottom shows the genomic coordinates. (n = 3 independent biological replicates for MYCN pLHiChIP; n = 2 for RNAPII ChIP-Rx).
B. Bar chart listing functional annotations of all binary MYCN interactions (N = 4,591; N indicates total number).
C. Boxplots showing relative binding of the indicated proteins (RNAPII, MYCN, TFIIIC5) to promoter regions or expression levels of the corresponding genes (mRNA by RNA-seq; 4sU by 4sU-seq). Red boxes: Genes bound by MYCN and part of MYCN-hubs; Blue boxes: Genes bound by MYCN that are not part of MYCN-hubs. Each pair was normalized to the median of the corresponding “blue” gene set. p-values were obtained by pairwise comparisons using Student’s t-test. (n = 2 for TFIIIC5 and RNAPII ChIP-Rx).
D. Boxplot showing the number of promoters in each cluster, with each red dot representing one cluster.
E. Network reconstruction of the three biggest clusters based on MYCN pLHiChIP interactions. Each anchor is represented by a node (“triangle”) and the lines show interactions between the anchors. The colors are indicating the different functional annotation.
TFIIIC antagonizes MYCN participation in promoter hubs.
A. Representative example of pLHiChIP track for MYCN (red) and TFIIIC5 (blue) interactions (conventions as in Figure 3A) (n = 2).
B. Bar chart listing the total number of functional annotations for all TFIIIC5 binary interactions (N = 3,499).
C. Venn diagram showing the number of interactions shared between MYCN and TFIIIC5. The diagram at the left shows the types of overlaps between connections.
D. Bar chart listing the interactions functional annotations for MYCN anchors not overlapping with TFIIIC5 anchors (“MYCN only”) as well as TFIIIC5 anchors without overlapping MYCN anchors (“TFIIIC5 only”) and their joint anchors.
E. Representative example of MYCN spLHiChIP track for MYCN interactions in the presence (blue) or absence (red) of TFIIIC5 (n = 2).
F. Bar graph showing the fold change of all MYCN spLHiChIP interactions comparing “+ TFIIIC5” and “– TFIIIC5” in SH-EP-MYCN-ER cells expressing a Dox-inducible shRNA targeting TFIIIC5. n1,2 indicates two independent biological replicates.
G. Representative example of TFIIIC5 spLHiChIP track without (blue) or with (red) induction of MYCN for SH-EP-MYCN-ER cells (conventions as in Figure 3A).
H. Bar graph showing the number of TFIIIC5 interactions normalized by the relative binding of TFIIIC5 ChIP-Rx signals for the same coordinates. Coordinates defined as TSS ± 2 kb of 14,722 genes.
TFIIIC is required for promoter association of the exosome and of BRCA1.
A. Boxplots showing the number of proximity ligation assay (PLA) signals between RNAPII and NELFE, PP2A, PNUTS or XRN2. SH-EP-MYCN-ER cells were treated with 1 µg/ml Dox (“– TFIIIC5”, 48 h) and/or 4-OHT (“+ MYCN”). EtOH was used as control. For clarity purposes, 500 cells pooled from different replicates were plotted. p-values were calculated comparing the PLA signal of all cells using unpaired Wilcoxon rank sum test. The grey dotted line indicates the median in the control condition (n = 3).
B. Density plot of CUT&RUN for EXOSC5 binding (N = 14,704 genes) in SH-EP-MYCN-ER cells expressing a Dox-inducible shRNA targeting TFIIIC5 treated with 4-OHT. Data show mean ± SEM (shade).
C. BRCA1 ChIP in SH-EP-MYCN-ER cells expressing a Dox-inducible shRNA targeting TFIIIC5 treated with 4-OHT (4 h). Shown is the mean of technical triplicates of one representative experiment with identical results (n = 2).
Model.
Model summarizing our findings. We propose that complex formation with the TFIIIC complex antagonize the localization of MYCN in promoter hubs and that this enables access of the nuclear exosome and BRCA1 to promoters with paused or stalled RNAPII. Both the exosome and BRCA1 have been implicated in fostering the degradation of nascent RNA at promoters. The precise mechanisms by which MYCN and TFIIIC limit accumulation of non-phosphorylated RNAPII at promoters remain to be determined.
Characterization of MYCN/TFIIIC complexes.
A. SDS-PAGE gel of recombinant purified MYCN (aa 1-137) expressed in E. coli cells.
B. SDS-PAGE gel of recombinant purified 1A subcomplex which was expressed in Spodoptera frugiperda (Sf9) cells.
C. Deconvoluted spectra of native mass spectrometry for the 1A/3x FLAG-MYCN complex. Masses are shown in red, identities of complexes are marked in blue.
D. Box plot of the 30 possible 1A/3x FLAG-MYCN complex molecular weights based on the masses observed using intact mass spectrometry. The red line indicates the mass observed for the complex by native mass spectrometry (204,763 Da).
E. Immunoblot showing levels of TFIIIC2, TFIIIC3 or TFIIIC5 and MYC in SH-EP-MYCN-ER cells expressing doxycycline (Dox)-inducible shRNAs targeting the different TFIIIC subunits (n = 3). Where indicated cells were treated with Dox (1 µg/ml, 48 h) and/or 4-OHT (200 nM, 4 h), respectively. EtOH was used as control. In the following panels addition of 4-OHT is indicated by “+ MYCN” and Dox treatment by “– TFIIIC”.
F. Immunoblot of MYC and MYCN in SH-EP-MYCN-ER cells after induction of MYCN with 4-OHT (200 nM, 4 h). VCL was used as a loading control.
Effects of MYCN and TFIIIC on RNA polymerase II.
A. Metagene plot of ChIP-Rx signal for non-phosphorylated RNAPII. Data show mean (line) ± standard error of the mean (SEM indicated by the shade) of different gene sets based on an RNA-seq of SH-EP-MYCN-ER cells ± 4-OHT.
B.. Metagene plot of ChIP-Rx signal for RNAPII pSer2. Data are presented as described in A.
C. Average density plot of ChIP-Rx signal for non-phosphorylated RNAPII in SH-EP cells treated with 4-OHT. Data show mean (line) ± SEM (indicated by the shade) of different gene sets based on an RNA-seq of SH-EP-MYCN-ER cells ± 4-OHT. The signal is centered to the TSS ± 2 kb (n = 2).
D. Average bin dot plot for RNA-seq of SH-EP-MYCN-ER showing mRNA expression normalized by control per bin. Cells were co-treated with 1 µg/ml Dox (“– TFIIIC3”, 48 h) and/or 4-OHT (“+ MYCN”, 4 h) as EtOH as control. Expression was normalized by its control. Each bin represents 100 genes of a total of 12,091 genes. Dotted line marks the relative expression at 1 (n = 3).
E. Browser tracks for TFIIIC5 ChIP-Rx at the indicated gene loci. SH-EP-MYCN-ER cells were treated with DRB and/or 4-OHT, respectively. EtOH was used as control.
Characterization of HiChip Methods.
A. Diagram showing the workflow of pLHiChIP. The workflow corresponds to the original HiChIP protocol. For a detailed description, see Methods.
B. MYCN and MYC ChIP in SH-EP-MYCN-ER cells before and after activation of MYCN with 4-OHT (200 nM, 24 h). Shown is the mean and points of technical triplicates normalized to an intergenic region (n = 2).
C. (Top) Diagram illustrating the difference between valid and invalid pairs after mapping of spLHiChIP data. Valid pairs harbor non-contiguous elements of chromosomal DNA, whereas invalid pairs harbor contiguous stretches of chromosomal DNA. (Bottom) Bar graph showing percentage of valid and invalid reads in our dataset (right) compared to a published HiC protocol (left).
D. Table showing quality controls for MYCN and TFIIIC5 spLHiChIP compared to the original Oct4 HiChIP.
E. Representative example of pLHiChIP track for MYCN (red) and a pLHiC track (blue) showing the strongly increased PETs number of the MYCN interactions after MYCN-immunoprecipitation relative to input (pLHi-C). Data are superimposed with a browser view of a MYCN ChIP-seq.
F. Table showing interactions involving tRNA genes in MYCN pLHiChIP.
G. Table showing the top three terms, their p-values and the FDR from enrichment analysis for MSigDB C5 collection of all MYCN network.
Three-dimensional interactions of MYCN and TFIIIC.
A. Heatmap showing analysis of sequence motifs characteristic for interacting boxes on both anchors of the interactions. Color reflects the number of interactions.
B. Table summarizing numbers of MYCN interactions in the presence and absence of TFIIIC5 in SH-EP-MYCN-ER cells and the TFIIIC5-dependent change. Data merged of 2 independent biological replicates.
C. Table displaying total numbers of TFIIIC5 interactions in SH-EP-MYCN-ER ± MYCN + DRB.
Effects of TFIIIC on splicing and termination factors.
A. Volcano Plot depicting changes in intron retention upon TFIIIC3 knock-down in the absence (left) and presence (right) of MYCN. Triangle reflects the effect of the splicing error after TFIIIC knock-down.
B. Volcano Plot depicting changes in exon skipping upon TFIIIC3 knock-down in the absence (left) and presence (right) of MYCN. Triangle reflects the effect of the splicing error after TFIIIC knock-down.
C. Controls for proximity ligation assays (PLAs) shown in Figure 5A. Immunofluorescence show specificity of the antibody. Single antibodies PLAs were performed in parallel to each PLA. Mean of signal in the nucleus and cytoplasm was calculated and compared to mean of the signal per nucleus of the PLA.
D. Boxplots showing the number of proximity ligation assay (PLA) signals between RNAPII and TFIIIC5. SH-EP-MYCN-ER cells were treated with 1 µg/ml Dox (“– TFIIIC5”, 48 h) and/or 4-OHT (“+ MYCN”). EtOH was used as control. For clarity purposes, 500 cells pooled from different replicates were plotted. p-values were calculated comparing the PLA signal of all cells using unpaired Wilcoxon rank sum test. The grey dotted line indicates the median in the control condition (n = 3).
E. BRCA1 ChIP in SH-EP-MYCN-ER cells expressing a Dox-inducible shRNA targeting TFIIIC5 treated with 4-OHT. Data show fold change of BRCA1 binding after induction of MYCN in the presence (blue) or absence (red) of TFIIIC5 (n = 2).
