UV Irradiation Triggers Transcript Elongation Slow-Down
(A) Schematic of UV/DRB/GRO-seq.
(B) Profile of nascent PPP1R12A RNA reads after DRB-release ± UV irradiation (15 J/m² used throughout this study). Arrows indicate transcription wave-fronts.
(C) As in (B), but meta-gene profile of normalized GRO-seq reads across 8,148 genes.
(D) Position of the GRO-seq transcription wave-front for 333 long genes over time ± UV irradiation. Dashed lines indicate median wave-front positions.
(E) Meta-gene profile of normalized GRO-seq reads −2 kb to +120 kb relative to the TSS, ± UV irradiation followed by 2, 8, and 24 hr recovery. Arrows indicate the height of the promoter proximal peak. Shaded areas indicate gene regions characterized by increased (yellow) or decreased (gray) GRO-seq signal 2 hr after UV exposure, normalizing over time.
(F) Gradual recovery of GRO-seq reads 90–120 kb downstream of the TSS following UV irradiation.
See also .

Splicing Analysis Reveals Frequent UV-Induced Alternative Last Exon Splicing
(A) UV-induced splicing events.
(B) Relative expression of terminal exons associated with UV-induced ALE splicing events. The ratio of proximal to distal terminal exon was calculated for the UV-treated sample and normalized to the control.
(C) Exon expression profiles for HERC4 and INTS6 upon UV irradiation. Red, dashed box indicates exons associated with expression of the short isoform. Schematic illustrations are shown below. Red arrows indicate terminal exons specific to the short isoforms.
(D) qRT-PCR validation of isoform expression. GAPDH normalized data relative to untreated conditions.
(E) Change in pre-mRNA length of ALE short events (left) and long events (right). Box and whisker plots with min/max/median represent pre-mRNA lengths.
(F) GRO-seq signal across HERC4 after UV exposure (boxed inset, area of short isoform). Arrowheads highlight recovery of gene synthesis at the 3′ end after 24 hr.
(G) Box and whisker plots (5–95 percentile with min/max/median indicated), showing relative GRO-seq read density of terminal exons following UV irradiation, normalized to untreated. Data for (D) and others like it in the following figures are mean ± SEM, t test, ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001 and ^∗∗∗∗p < 0.0001.
See also and and , , and .

Bulky DNA Lesions Induce ASCC3 Alternative Last Exon Switching
(A) ASCC3 exon expression profiles, as in C.
(B) qRT-PCR validation of the isoform switch 24 hr after UV irradiation. GAPDH normalized data relative to untreated conditions are averaged, ±SEM.
(C) Expression, determined by qRT-PCR, of the different isoforms of ASCC3 upon exposure to 100 nM camptothecin (CPT), 20 μM cisplatin (CisPt), 0.001% MMS, or 5 Gy ionizing radiation (IR). Untreated conditions (Un) set to 1. GAPDH normalized as in (B).
(D) GRO-seq signal across ASCC3, as in F.
(E) qRT-PCR of nascent pre-mRNA across ASCC3 after UV irradiation, using intron-exon junction primers (averaged, 18S normalized data, shown relative to untreated).

ASCC3 Long Isoform Knockdown Increases Global Transcription after UV Irradiation
(A) ASCC3 isoforms and siRNA-targeting regions. Ninety-three nucleotide sequence not present in the long isoform shown in dark blue.
(B) Scores from the genome-wide RNAi screen () with ASCC3 (pool-1), ASCC2, and ASCC1 siRNA pools highlighted in red. ASCC3 (pool-2) is highlighted in blue.
(C) Representative images of cells transfected with non-targeting (NT) siRNA and the ASCC3 siRNA pool-1 20 hr after UV irradiation. Nascent EU-labeled RNA shown in green and DAPI-stained nuclei in blue. 10× objective image on the left, with region in white box enlarged on the right.
(D) Histogram plots of average EU incorporation following knockdown of ASCC1, 2 and 3 (pool-1) 20 hr after UV irradiation. Black and gray stippled lines demarcate thresholds of lowly and highly transcribing cells, respectively.
(E) EU incorporation after treatment with NT siRNA (red) or individual siRNAs targeting ASCC3 long isoform (light and dark blue), with or without UV irradiation, measured after 2 and 20 hr. Data shown as in (D). Arrowheads highlight reduced proportion of lowly transcribing cells and shift of histogram to the right in cells lacking the long ASCC3 isoform.
(F) The ratio of low to high transcribing cells (cells left of the black line over cells right of the gray line in E), in untreated conditions (white bars) or 20 hr after UV irradiation (gray bars). Data were averaged and normalized relative to UV-treated control cells (set to 1), ±SEM.
See also .

Cells Deficient for the Short ASCC3 Isoform Cannot Recover Transcription after UV Irradiation
(A) As in E, but after knockdown of the short ASCC3 isoform with individual siRNAs (light and dark blue).
(B) As in F, but after knockdown of short isoform. Data are mean ±SEM relative to UV-treated control.
(C) UV-sensitivity measured by colony formation after knockdown with individual siRNAs targeting ASCC3 short isoform. Two-way ANOVA test: NT versus ASCC3 siRNA-1 p = 0.0182; NT versus ASCC3 siRNA-2 p = 0.008.
(D) CRISPR-Cas9-mediated knockout of the unique, terminal exon of the short ASCC3 isoform. Genomic PCR fragments isolated from parental MRC5VA cells (P) and two knockout (KO) clones are shown (red arrows, primers; blue scissors, guide RNAs).
(E) qRT-PCR analysis of short isoform RNA expression in the cell lines from (D), showing averaged GAPDH-normalized data, relative to parental cells.
(F) Transcription recovery after deletion of the short ASCC3 isoform, measured as in F and B.
(G) Histograms showing decreased EU intensity/nucleus in ASCC3 short isoform KO clone-2 cells compared to control cells 20 hr after UV irradiation.
See also .

Numerous UV-Regulated Genes Are Antagonistically Regulated by the Long and Short ASCC3 Isoforms
(A) Hierarchical clustering of genes downregulated in UV-treated ASCC3 short knockout cells, shown alongside expression of the same genes in UV-treated ASCC3 Long knockdown cells, relative to expression in control cells. Blue and red bars indicate minimum and maximum log fold-changes, respectively.
(B) qRT-PCR analysis of IL7R and VEGFC expression in UV-treated short KO cells (light gray; top panel) compared to parental cells and in UV-treated long knockdown cells (dark gray; bottom panel) compared to NT shRNA cells (black) 20 hr after UV, shown as averaged GAPDH-normalized data, relative to untreated controls.
(C) As in (B) but analysis of genes with UV-induced expression.
(D) Rescue of gene expression in UV-treated short KO cells by transfection with siRNA targeting the long isoform. Analysis by qRT-PCR, with GAPDH-normalized data shown relative to UV-treated control cells.
(E) As in (D) but for global nascent transcription, as indicated by the low/high transcription ratio 20 hr after UV irradiation, relative to control cells. C, siNT control; P, parental control.
See also and .

The Short ASCC3 Isoform Is a Chromatin-Associated lncRNA
(A and B) Histogram (A) and low/high transcription ratio plot (B), showing the effect on transcription of expressing different siRNA-resistant RNAs in ASCC3 short isoform knockdown cells. (A) Blue arrows indicate the reduction in lowly transcribing cells and concomitant increase in highly transcribing cells following rescue with ASCC3 short isoform constructs containing the 3′UTR. Data in (B) are relative to UV-treated control cells, mean ± SEM. n.s., not significant; CDS, coding sequence; Mut. CDS, stop-containing CDS mutant.
(C and D) As in (A) and (B) but for short isoform knockout cells.
(E) RNA scope In situ hybridization signals for endogenous ASCC3 long and short isoforms. RNA scope signal (red) was overlaid with DAPI to highlight nuclear localization.
(F) Immunoblot showing localization of RNAPII (RPB1 subunit), hnRNPA1, tubulin, and histone H3 following sub-cellular fractionation. S1, cytoplasmic; S2, soluble nuclear material; P2, chromatin pellet.
(G) Enrichment of the short ASCC3 isoform in the S2 and P2 fractions as determined by qRT-PCR. As control, P2 was analyzed without reverse transcriptase (−RT). Data are relative to untreated S1 fraction, mean ± SEM.
(H) Model showing RNAPII (gray sphere) producing nascent ASCC3 transcript (red), including the alternative last exon (thick blue line). Splicing determines exclusion/inclusion of the ALE and 3′-UTR (boxes on right). The protein-encoding long isoform mRNA and the non-coding short isoform have opposite effects on the DNA damage response and affect each other’s function (indicated by double arrow on right).
See also .

Transcription Wave-Front and Elongation Rates, Related to
(A) Elongation rate (kb/min) calculated for 333 long genes for the indicated time intervals in untreated and UV-treated cells. Data are box and whisker (min to max) with median indicated. Mann-Whitney test, ^∗∗∗∗p < 0.0001. (B) Comparison of the elongation rate for the 10-25 and 25-40 min time intervals in UV-treated cells. (C) Meta gene profile across the region −2 kb to 120 kb relative to the TSS of mean normalized read GRO Seq read density from untreated cells (No UV) and cells treated with UV followed by 2, 5, 8, 10, 12 and 24 hr. (D) Computationally determined wave fronts for 141 genes during transcription recovery, 2-24 hr post-UV. Data are box and whisker (min to max), with median indicated. The median wave front for the 24 hr time point is indicated by the red dashed line. (E). The median wave fronts from B were plotted for samples 2-12 hr after UV treatment. A line of best fit and the corresponding equation is shown. The slope of the line indicates the rate of transcription elongation as determined by individual gene wave front calling. (F) Mathematically determined GRO-Seq transcription wave fronts during the transcription recovery 2-12 hr post-UV. A line of best fit and the corresponding equation is shown. The slope indicates the rate of transcription elongation as determined by mathematical wave front calling.

UV-Induced ALE Short Events Reduce Transcript Length and Correlate with Nascent RNA Synthesis Compared to Background Events, Related to
(A) Diagram showing UV-induced alternative last exon splicing and premature termination. Normal, fast elongation favors the skipping of the red exon along with its associated poly-A signal, leading to preferential production of the long isoform. Slower elongation favors splicing of the red exon leading to utilization of the associated poly-A signal and expression of the short isoform. (B) Illustration depicts the variables used for the equation for relative change in isoform length ((b-a)/b). Plot of relative change in isoform length for ALE short events (blue line) (n = 121) compared to non-UV regulated background events (n = 7736) (red line) shows UV-regulated events are characterized by a statistically significant greater difference in short and long isoform length compared to events that were not effected by UV. Wilcox test, p = 7.195 X10⁻⁷. (C) Smoothed GRO-Seq signal across the CNTLN gene showing increased synthesis spanning the 5′ region gene corresponding to the short isoform (see boxed inset, 8 and 24 our after UV) and reduced signal across the rest of the gene corresponding to the long isoform. Arrows highlight the sustained repression of synthesis at the 3′ end of the gene 24 hr after UV correlating with preferential short isoform expression at this time point. (D and E) GRO-Seq read density mapped to the terminal exon of UV-induced short isoforms (D) and UV-suppressed long isoforms (E) at the indicated time points after UV-irradiation normalized to untreated samples. Data are box and whisker (min to max), with median indicated. (F) No increase of short over long isoforms was observed for ‘background’ ALE events that were unaffected by UV treatment. The ratio of short isoform (TSS proximal) to long isoform (TSS distal) terminal exon expression at the indicated times following UV treatment normalized to untreated is shown for background events not regulated by UV (n = 7736) (non-hits, gray boxes) and for UV-induced ALE short events (n = 121) (hits, black boxes). Data are box and whisker (min to max) with median indicated. Mann Whitney test, ^∗∗p < 0.01, ^∗∗∗∗p < 0.0001.

Gene Ontology Analysis of ALE Short Genes, Related to
Gene ontology of ALE short genes reveals a significant enrichment in genes involved in transcription and transcription regulation.

RNAi Knockdown of ASCC3 Isoforms and CSB and EU Labeling Assay, Related to
(A) EU incorporation for untreated or UV-treated cell populations. Data are average EU intensity/nuclei displayed as histograms. There is a rapid reduction in EU incorporation 1 hr after UV treatment as indicated by a shift of the population to the left and corresponding increase in the percentage of cells below the low transcription threshold (indicated by the red dashed line). EU incorporation gradually recovers over the 48 hr time course. (B) RT-qPCR analysis of ASCC3 long isoform expression in cells stably expressing non-targeting (NT) shRNA and two clones stably expressing ASCC3-targeting shRNA (shASCC3) and short isoform expression 48 hr after transfection with NT and short isoform-targeting siRNA. (C) EU incorporation for shNT and shASCC3 cell lines in untreated (left panel) and UV-treated (15 J/m², 18 hr recovery) (right panel) conditions. Data are average EU intensity/nuclei displayed as a histogram. (D) EU incorporation for UV-treated (15 J/m2, 18 hr recovery) NT and CSB-targeting siRNA transfected cells shows deficient recovery of transcription in CSB knockdown cells. Data are average EU intensity/nuclei displayed as a histogram. (E) The ratio of the proportion of low transcribing cells over high transcribing cells in untreated conditions (white bars) or 18 hr after UV (gray bars) transfected with NT or CSB siRNA. Data are mean -/+ SEM relative to UV-treated NT siRNA. t test, ^∗∗∗∗p < 0.0001.

Deficiency in ASCC3 Short Isoform Does Not Affect ASCC3 Long Isoform Expression and Vice Versa, Related to
(A) ASCC3 short isoform knockout cells that were left untreated or UV treated (15 J/m2) followed by 20 hr recovery were analyzed for long isoform expression by RT-qPCR. Data are normalized to GAPDH and relative to untreated parental samples, mean -/+ SEM. Knockout cells have a slight increase in long isoform expression (∼1.2 fold) in untreated conditions compared to parental controls but equally downregulate long isoform expression in response to UV. (B). Cells stably expressing ASCC3 long isoform-targeting shRNA or NT shRNA control were left untreated or UV treated (15 J/m2) followed by 20 hr recovery and analyzed for short isoform expression by RT-qPCR. Data are GAPDH normalized and relative to untreated shNT cells, mean -/+ SEM C. Expression of ASCC3 protein in ASCC3 short knockout cells and parental control cells in untreated and UV-treated (15 J/m², 24 hr) conditions. Tubulin is shown as a loading control. Expression of ASCC3 protein is unaffected by short isoform knockout.

Rescue of Transcription in UV-Treated ASCC3 Short Isoform Knockout Cells by Knockdown of ASCC2, Related to
ASCC3 short isoform knockout cells (clone 2) and parental cells (WT) were transfected with NT or ASCC2-targeting siRNA and incubated for 48 hr prior to exposure with UV (15 J/m²). Expression of IFIT2, IFI44 and IL7R was analyzed by RT-qPCR 20 hr after UV treatment. Data are GAPDH normalized and relative to UV-treated NT siRNA transfected parental cells, mean -/+ SEM. Note that the NT siRNA transfected samples are the same as those shown in E. t test, ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001, ^∗∗∗∗p < 0.0001.

ASCC3 Short Isoform Transgene Expression and RNA Localization in MRC5VA Cells, Related to
(A) Cells transfected with an ASCC3 short isoform coding sequence (CDS) expressing construct or mock transfected were incubated for 30 hr prior to transfection with NT or ASCC3 short isoform-targeting siRNA. Cells were UV-treated 48 hr after siRNA transfection followed by a 20 hr recovery. Data are average EU intensity/nuclei displayed in a histogram. Expression of ASCC3 short isoform CDS did not rescue the low transcription following ASCC3 short isoform knockdown. (B) Inputs (5%) (lanes 1-4) and M2 Flag immunoprecipitation (IP) (lanes 5-8) from untransfected cells and cells ectopically expressing Flag-tagged ASCC3 transcripts containing the coding sequence alone (CDS), coding sequence with 3′ UTR (CDS +3′UTR) and coding sequence with 3′UTR that contains an in-frame premature stop mutation (Mut. CDS +3′UTR). Immunoblot was probed with ASCC3 short isoform specific antibody. ASCC3 short isoform was not detected in input samples therefore Ponceau S stain is shown as a control for equal protein loading. ^∗ indicates a non-specific band in input samples not present in IP elutions. (C) Inputs (10%) (lanes 1-4) and IP with anti-Flag M2 (lane 5) and ASCC3 N-terminal targeting antibodies (lanes 6-8) was performed on untransfected cells and cells transfected with constructs encoding Flag-tagged ASCC3 short isoform followed by immunoblotting using ASCC3 short and long isoform specific antibodies. Ectopically expressed short isoform was pulled down by both Flag and ASCC3 N-terminal antibodies (lanes 5 and 6) however short isoform protein was not detected in ASCC3 N-terminal antibody pull downs from untransfected cells under both untreated and UV-treated conditions (lanes 7 and 8). IP of ASCC3 long isoform is shown as a positive control for efficient N-terminal antibody pull down. ASCC3 short isoform was not detected in input samples. (D) RNA scope In situ hybridization followed by Fast Red staining using probes targeting the ASCC3 short isoform, in untreated cells and UV-treated cells after 18 hr recovery. The nucleus was counterstained with DAPI. The primarily punctate nuclear localization of ASCC3 was not significantly affected by UV treatment over several independent experiments. (E) RNA scope In situ hybridization in short isoform knockout cells with short and long isoform-targeting probes shows specific loss of short isoform signal but not long isoform signal confirming the short isoform probe signal was indeed specific for the short isoform of ASCC3. (F) LncRNA Malat-1 is enriched in the chromatin-associated P2 fractions as determined by RT-qPCR. The P2 fraction was also analyzed in the absence of reverse transcriptase (-RT) as a control. Data are relative to untreated S1 fraction, mean -/+ SEM.