Fig 2: THOC5 influences on transcription elongation(A and B) Metaplot of mean Pol II ChIP signals for 626 THOC5-dependent (A) and 213 THOC5 independent (B) genes were generated using deepTools2.21 50-base bins are shown in flanking regions from 1 kb relative to the TSS and from 3 kb relative to the polyA site. Input and IgG were used as negative control. MACS peak caller module of Seqmonk (p value cut-off = 10−5) was applied to identify and quantify peaks at the TSS and gene body. Read density of peaks was normalized to sequencing depth. p: p value.(C) Metaplot of mean TAP-THOC5 ChIP signals for 626 THOC5-dependent and 213 THOC5 independent genes.(D) Outline of DRB inhibition and 4SU labeling times used for DRB/TTchem-seq.(E) DRB/TTchem-seq metagene profiles of protein-encoding genes between 60 and 300 kb from standard chromosomes (1–22, X, Y) with non-overlapping transcriptional units (n = 4591). Lines are computationally fitted splines.(F) Calculation of Pol II transcription wave position based on metagene profiles using linear regression.(G and H) Coverage profiles of DRB/TT-seq results for THOC5-dependent (I) and THOC5 independent (J) genes. Black arrows indicated the delayed elongation rates upon THOC5 depletion. Red arrows indicated accumulations of Pol II molecules short after TSS.(I and J) DRB/TTchem-seq metagene profiles of 626 THOC5-dependent and 213 THOC5 independent genes.(K) Elongation rates of THOC5-dependent and independent genes were plotted.See also Figure S2 and Table S2.

Fig 3: Chromatin-associated THOC5 interacts with CDK12 in slow Pol II cells but not in fast Pol II cells and influences CDK12 recruitment during transcription elongation(A) Fast and slow Pol II mutants were induced by Doxycycline treatment. Cytoplasmic RNAs were isolated and subjected to Nanopore direct RNA-seq.(B) The ratio (proximal/distal) in fast and slow Pol II cells was plotted. Red dot: slow Pol II-dependent proximal cleavage genes, black dots: slow Pol II independent genes, blue dot: slow Pol II-dependent distal cleavage genes.(C) Wiggle plot of GSPT1, OSBPL3, NUDT21, and USP24 gene: Seqmonk was used to quantitate and visualize sequencing data. Peaks in the wiggle plot represent the normalized RNA-seq read coverage. TSS: transcription start site. TES: transcription end site.(D) pNTAP (TAP) and pNTAP carrying THOC5 cDNA (TAP-THOC5) were transfected into fast and slow Pol II cells for 24 h. After cross-linking by adding formaldehyde, protein and DNA were extracted and the chromatin was sheared by sonication. Cell extracts and binding fractions with streptavidin Sepharose were analyzed by GSPT1 (TSS, +2 kb, +43 kb, +45 , and 48 kb) and OSBPL3 (TSS, 1.8 , 80 , 180 , and 185 kb).(E) Chromatin-associated and nucleoplasmic fractions of HEK293 cells subjected to immunoblot for TREX complex components and NXF1. WTAP and histone H3 were used as markers for chromatin-associated and nucleoplasmic fractions.(F) Venn diagram showing the overlap of proteins identified by the chromatin-associated THOC5 interactome analysis in fast and slow Pol II cells. IPA analysis was performed to identify any functional enrichment observed within the proteins within each sub-group. The significance (p values) for the function enrichment was calculated using the right-tailed Fisher exact test.25(G and H) Interaction between chromatin-associated THOC5 and CDK12, CHTOP, and THOC6 in slow Pol II cells (G) was confirmed by endogenous THOC5 co-IP study in slow Pol II cells. Fast Pol II cells were used as a negative control (H). Chromatin-associated fractions were isolated and supplied for THOC5 co-IP assay.(I) Interaction between chromatin-associated THOC5 and CDK12 in THOC6-depleted cells was examined by endogenous THOC5 co-IP study.(J) Interaction between chromatin-associated THOC5 and THOC6 in CDK12-depleted cells was examined by endogenous THOC5 co-IP study.See also Figures S3, S7, Tables S3 and S4.

Fig 4: THOC5 recruits DDX5 and DXX17 during transcription elongation(A) DDX5, DHX15, DDX17, and DDX50 were depleted using siRNA. Knockdown efficiency was confirmed by qRT-PCR. Three independent experiments were performed, and the average signal of qRT-PCR is shown as the mean value +/−SD.(B) Genomic (g)DNAs were isolated from control and DDX-depleted cells. gDNA with/without RNAse H treatment were subjected to Dot Blot analysis using R-loop-specific antibody (S9.6). ssDNA was used as loading control. Dot blot signal was quantified using ImageJ.(C) Representative images and quantification of nuclear RNA–DNA hybrid accumulation in DDX5- or DDX17-depleted HEK293 cells. Immunostaining with S9.6 monoclonal antibody and DAPI in the presence and absence of RNAse H and RNAse III in siCr, siDDX5, and siDDX17 transfected cells. More than 50 cells per condition were counted in each of the three experiments. The median of the S9.6 signal intensity per nucleus is shown (n = 3). Scale bar, 200 μm. (∗∗∗) p < 10−20, Student t test, two-tailed.(D) R-loop unwinding assay in the presence of increasing DDX5 or DDX17.(E) Representative images and quantification of nuclear RNA–DNA hybrid accumulation in THOC5-depleted HEK293 cells transfected with empty vector (Vector) or FLAG-tagged DDX5 (DDX5 o/e) or DDX17 (DDX17 o/e). More than 50 cells per condition were counted in each of the three experiments. The median of the S9.6 signal intensity per nucleus is shown (n = 3). Scale bar, 200 μm. (∗) p < 10−4, Student t test, two-tailed.(F and G) FLAG-tagged DDX5 (F) and DDX17 (G) were overexpressed in control and THOC5-depleted cells. Cells were fixed in 1% PFA and supplied for ChIP assay using FLAG M2 antibody. The recruitment profile of DDX5 and DDX17 to GSPT1, OSBPL3, and C4ORF46 genes in the control and THOC5-depleted cells is shown.(H and I) FLAG-tagged DDX5 (H) and DDX17 (I) were overexpressed in HEK293 cells. Co-IP studies using FLAG M2 antibody were performed to examine the interaction with THOC5, THOC6, and CDK12. Both input and pull-down samples were detected using FLAG-, THOC5-, THOC6-, and CDK12-specific immunoblot.(J) DDX5 and DDX17 were depleted using siRNA. Control and DDX5/17 double knockdown cells were fixed in 1% (v/v) PFA and subjected to CDK12 ChIP assay. The recruitment profiles of CDK12 to GSPT1, OSBPL3, and C4ORF46 gene in control and DDX5/17-depleted cells are shown. Three independent experiments were performed, and the average signal of ChIP is shown as the mean value +/−SD. ∗: p value <0.05.(K) Novel role of the THOC5/THOC6 complex during transcription elongation. When transcription elongation rate is low and R-loops are formed, the THOC5/THOC6 complex recruits CTD kinase CDK12, which phosphorylates the CTD domain of RNA polymerase II at serine 2 and enhances the elongation. At the same time, the THOC5/THOC6 complex also recruits DNA-RNA helicases DDX5 and DDX17 to resolve the R-loops, thus preventing transcription elongation impairment and premature termination.See also Figure S6.

Fig 5: Depletion of THOC6 and CDK12 affects 3′end cleavage of >50% of mRNAs(A and B) HEK293 cells were transfected with siRNA control (siCr) and siTHOC6 or siCDK12 for 3 days. Cytoplasmic polyA + mRNAs were isolated and subjected to direct RNA-seq. The ratio (proximal/distal) in THOC6- (A) and CDK12 (B)-depleted cells was plotted. Red dot: THOC6/CDK12-dependent proximal cleavage genes, black dots: THOC6/CDK12 independent genes, blue dot: THOC6/CDK12-dependent distal cleavage genes.(C) Venn diagram of THOC5-, THOC6-, and CDK12-dependent genes.(D) Wiggle plot of GSPT1 gene: Seqmonk was used to quantitate and visualize sequencing data. Peaks in the wiggle plot represent the normalized RNA-seq read coverage. TSS: transcription start site. TES: transcription end site.(F) FLAG-tagged THOC6 was overexpressed in HEK293 cells. Cells were fixed in 1% (v/v) PFA and supplied for ChIP assay using FLAG M2 mouse monoclonal antibody. Mouse IgG was used as a control. The recruitment profile of THOC6 to GSPT1, OSBPL3, PSMD12, and C4ORF46 genes is shown. Three independent experiments were performed, and the average signal of ChIP is shown as the mean value +/− SD.(G and H) THOC5 were depleted in HEK293 cells by lentivirus (sh), whereas THOC2 or THOC6 were downregulated by siRNAs. Cells were fixed in 1% PFA and CDK2 and Pol II ChIP assay performed. The recruitment profile of CDK12 (G) and Pol II (H) to GSPT1, OSBPL3, PSMD12, and C4ORF46 genes in the presence and absence of THOC5 or THOC6 is shown. Three to five independent experiments were performed, and the average signal of ChIP is shown as the mean value +/−SD. ∗p value <0.05.See also Figure S4.