Post‐transcriptional polyadenylation site cleavage maintains 3′‐end processing upon DNA damage

Abstract The recognition of polyadenylation signals (PAS) in eukaryotic pre‐mRNAs is usually coupled to transcription termination, occurring while pre‐mRNA is chromatin‐bound. However, for some pre‐mRNAs, this 3′‐end processing occurs post‐transcriptionally, i.e., through a co‐transcriptional cleavage (CoTC) event downstream of the PAS, leading to chromatin release and subsequent PAS cleavage in the nucleoplasm. While DNA‐damaging agents trigger the shutdown of co‐transcriptional chromatin‐associated 3′‐end processing, specific compensatory mechanisms exist to ensure efficient 3′‐end processing for certain pre‐mRNAs, including those that encode proteins involved in the DNA damage response, such as the tumor suppressor p53. We show that cleavage at the p53 polyadenylation site occurs in part post‐transcriptionally following a co‐transcriptional cleavage event. Cells with an engineered deletion of the p53 CoTC site exhibit impaired p53 3′‐end processing, decreased mRNA and protein levels of p53 and its transcriptional target p21, and altered cell cycle progression upon UV‐induced DNA damage. Using a transcriptome‐wide analysis of PAS cleavage, we identify additional pre‐mRNAs whose PAS cleavage is maintained in response to UV irradiation and occurring post‐transcriptionally. These findings indicate that CoTC‐type cleavage of pre‐mRNAs, followed by PAS cleavage in the nucleoplasm, allows certain pre‐mRNAs to escape 3′‐end processing inhibition in response to UV‐induced DNA damage.


Introduction
During UV-induced DNA damage and other genotoxic stresses, several steps in eukaryotic gene expression are repressed. Although global, this repression is associated with an upregulation of the expression of genes encoding proteins that are essential for the adaptation and response to stress. For instance, despite the inhibition of pre-mRNA 3 0 -end processing observed in UV-treated cells (Kleiman, 1999;Kleiman & Manley, 2001;Kim et al, 2006;Nazeer et al, 2011), pre-mRNA 3 0 -end processing of the pre-mRNA encoding the p53 tumor suppressor (TP53) protein is specifically maintained (Decorsi ere et al, 2011;Newman et al, 2017). This maintenance requires several RNA binding proteins, i.e., the heterogeneous nuclear ribonucleoprotein (hnRNP) F/H family of proteins that bind to an RNA G-quadruplex forming sequence located downstream of the p53 polyadenylation site (Decorsi ere et al, 2011), as well as the DHX36 RNA/DNA helicase (Newman et al, 2017).
The main mechanism of pre-mRNA 3 0 -end processing is cleavage and polyadenylation (CPA), which involves endonucleolytic cleavage of newly synthesized transcripts and the addition of adenosine residues constituting the poly(A) tail to the generated 3 0 -end. CPA is crucial for mRNA stability, transport to the cytoplasm, and translation (Millevoi & Vagner, 2010;Shi & Manley, 2015). This nuclear process involves the recognition of cis-acting elements in the pre-mRNA by a complex machinery comprising more than 80 proteins (Shi et al, 2009). The pre-mRNA sequences serving as the polyadenylation signal (PAS) include a hexameric sequence (most often AAUAAA) located 10-30 nucleotides (nt) upstream of the cleavage site (generally a CA dinucleotide) and a downstream sequence element (DSE; U/GU-rich) located within 30 nt downstream of the cleavage site. Additional sequence elements located either upstream (upstream sequence element; USE) or downstream (auxiliary downstream sequence element; AuxDSE) of the cleavage site modulate the recognition of the PAS.
The cleavage reaction at the PAS (called thereafter PAS cleavage), which precedes the addition of the poly(A) tail, generally occurs in a co-transcriptional manner. PAS recognition is indeed tightly coupled to RNA polymerase II (Pol II) transcription termination (Proudfoot, 2016). Rpb1, the largest subunit of Pol II, contains a carboxy-terminal domain (CTD) that is comprised of heptad repeats (consensus Tyr 1 -Ser 2 -Pro 3 -Thr 4 -Ser 5 -Pro 6 -Ser 7 ) and plays a critical role in coupling pre-mRNA 3 0 -end processing and transcription termination, especially through its phosphorylated Ser 2 (phospho-Ser 2 ) residues (Ahn et al, 2004). Several components of the polyadenylation machinery, including PCF11, which is concentrated at the 3 0 -end of genes, preferentially bind the phospho-Ser 2 CTD (Barilla et al, 2001;Licatalosi et al, 2002;Meinhart & Cramer, 2004). In human cells, PCF11 depletion leads to a transcription termination defect through a decrease in the degradation of the downstream RNA, generated after the PAS cleavage (West et al, 2008). In the Pause-Type model of transcription termination, Pol II pauses at a GC-rich region located a few nucleotides downstream of the PAS, stimulating the PAS cleavage of the pre-mRNA in a co-transcriptional manner, i.e., when the pol II-bound pre-mRNA is on the chromatin (Gromak et al, 2006;Nojima et al, 2013;Cortazar et al, 2019).
Another model of transcription termination has been proposed (Dye & Proudfoot, 2001;West et al, 2008;Nojima et al, 2013). In this Co-Transcriptional Cleavage (CoTC)-type model, the pre-mRNA is released from chromatin to nucleoplasm through a cleavage event at a CoTC site located downstream of the PAS, and the PAS cleavage subsequently occurs in the nucleoplasm. This mechanism has been described in several human genes (Nojima et al, 2013). The CoTCtype termination model has also been observed in Drosophilia where a release of pre-mRNA from transcription sites to the nucleoplasm takes place prior to PAS cleavage (Sikes et al, 2002). The CoTC cleavage occurs a few kilobases downstream of the PAS, generally at an AT-rich sequence called CoTC element (White et al, 2013). Mutations in this element induce an inhibition of pre-mRNA 3 0 -end processing in vitro (Teixeira et al, 2004).
Here, we report that upon UV irradiation, PAS cleavage of the p53 pre-mRNA is independent from the cleavage/termination factor PCF11 and CTD Ser 2 phosphorylation and relies on a downstream CoTC site, thereby allowing 3 0 -end processing of the p53 pre-mRNA to escape repression by DNA damage. We also identified several other pre-mRNAs that exhibit a CoTC-type mechanism of 3 0 -end processing in response to UV-induced DNA damage and that escape repression by DNA damage, like the p53 pre-mRNA.

Results
PCF11 is dispensable for p53 pre-mRNA 3 0 -end processing in UVtreated cells To understand the contribution of the pre-mRNA 3 0 -end processing machinery in the response to UV treatment, we analyzed the abundance of 13 proteins constituting the different sub-complexes involved in 3 0 -end processing by Western blot (Fig 1A). To ascertain that the band observed in each Western blot corresponds to the expected protein, we used published siRNAs targeting each of the corresponding mRNAs (Masamha et al, 2014). The experiments were performed in A549 lung tumor cells irradiated with UV (254 nm; 40 J/m 2 ) and harvested after 16 h of recovery, conditions that we previously used to demonstrate the maintenance of p53 pre-mRNA 3 0 -end processing following UV treatment (Decorsi ere et al, 2011;Newman et al, 2017). Consistent with previously reported data (Kleiman & Manley, 2001), we observed no changes in the levels of both CstF64 and CPSF160 in response to UV. The abundance of the other components of the CPSF, CstF, and CFIm complexes was unchanged ( Fig 1A). By contrast, we detected a significant decrease in the abundance of both PCF11 and CLP1 in UV-treated cells (Fig 1A). The UV-dependent reduction in PCF11 expression was confirmed in another set of experiments using 2 different siRNAs targeting PCF11 ( Fig 1B) and was accompanied by a 5-fold decrease in PCF11 mRNA level ( Fig 1C). These observations suggest that PCF11 might be dispensable for p53 pre-mRNA 3 0 -end processing following UV-induced DNA damage.
To confirm that PCF11 is not required for p53 pre-mRNA 3 0 -end processing following UV, we evaluated the effect of the siRNAmediated depletion of PCF11 on the efficiency of PAS cleavage of the p53 pre-mRNA by real-time quantitative PCR analysis (RT-qPCR; Fig 1D). The TBP pre-mRNA was used as a control since it was previously shown to be inhibited at the level of PAS cleavage efficiency due to UV treatment (Decorsi ere et al, 2011;Newman et al, 2017). According to a previously described approach (Decorsi ere et al, 2011;Newman et al, 2017), we measured the ratio of uncleaved RNA to total RNA (that is the sum of cleaved and uncleaved RNA) in the nuclear pool of RNAs, by qPCR with antisense primers located either downstream or upstream of the PAS cleavage site, respectively ( Fig 1D). In untreated cells, PAS cleavage of both the TBP and p53 pre-mRNAs was inhibited by PCF11 depletion, as revealed by the increased ratio of uncleaved/total RNAs in PCF11-depleted cells (Fig 1E). This is consistent with the fact that this factor is essential for the co-transcriptional, Pol II-coupled PAS cleavage reaction (West et al, 2008). Following UV treatment, while TBP PAS cleavage was still inhibited by PCF11 depletion, p53 PAS cleavage was no longer inhibited ( Fig 1E). This effect was specific to PCF11 since the siRNA-mediated depletion of CstF64, CFIm25, and CPSF160 all led to decreased p53 PAS cleavage in UV-treated cells (Appendix Fig S1). Altogether, these data indicate that PCF11, which exhibits reduced RNA and protein levels in UV-treated cells, is dispensable for p53 (but not TBP) pre-mRNA 3 0 -end processing in UV-treated cells.
Previous reports showed that UV-induced DNA damage induces global changes in Pol II phosphorylation, including Ser 2 phosphorylation (Rockx et al, 2000;Muñoz et al, 2009). Considering the link between PCF11 and the Pol II CTD phospho-Ser 2 (PolII Ser2P), we sought to determine whether inhibition of Ser 2 phosphorylation may mimic the effect of depleting PCF11 on p53 3 0 -end processing following UV. We treated cells with the Ser 2 kinase (CDK9) inhibitor DRB (5,6-Dichlorobenzimidazole 1-b-D-ribofuranoside) and then assessed the efficiency of pre-mRNA 3 0 -end processing. DRB reduced Pol II Ser 2 phosphorylation ( Fig EV1A). DRB, as expected, inhibited both TBP and p53 PAS cleavage in untreated cells (Fig EV1B). In UV-treated cells, DRB inhibited TBP, but not p53 PAS cleavage (Fig EV1C).
PAS cleavage of the p53 pre-mRNA occurs in the nucleoplasm following a CoTC event 3 of 15 nucleoplasm, where it is released upon CoTC cleavage. We therefore analyzed the nuclear distribution of the PAS-uncleaved p53 pre-mRNA by RT-qPCR in chromatin and nucleoplasm fractions. The quality of the fractionation was assessed by Western blot against histone H3 as a chromatin marker and topoisomerase IIa as a nucleoplasm marker (Fig 2A). The GAPDH and WDR13 pre-mRNAs were used as controls as they were previously reported to be PAScleaved co-transcriptionally or post-transcriptionally (following a CoTC event), respectively (Nojima et al, 2013). Accordingly, the relative abundance of PAS-uncleaved pre-mRNA in the nucleoplasm, as compared to the chromatin, was much higher for WDR13 than for GAPDH ( Fig 2B). The nucleoplasm/chromatin ratio of PASuncleaved pre-mRNA of p53 was similar to the one of WDR13, suggesting that the p53 pre-mRNA may be released in the nucleoplasm following a CoTC event, while the TBP pre-mRNA behaved similarly to the GAPDH pre-mRNA ( Fig 2B). An AT-rich sequence that could correspond to a potential CoTC sequence element is found around 1,200 nt downstream of the p53 PAS ( Fig 2C). To map the putative CoTC element, chromatinassociated RNA was reverse transcribed using random primers and the obtained cDNA was amplified by PCR using primers complementary to the 3 0 flanking region of the p53 gene ( Fig 2C). PCR amplification was carried out using a single forward primer (F), located upstream of the p53 PAS, in combination with reverse primers (R1-R6) located at an increasing distance downstream of the p53 PAS. The F/(R1-R6) primer pairs were used to amplify genomic DNA as an amplification control ( Fig 2D). In cDNA derived from chromatin-bound RNA, the F-R1, F-R2, F-R3, and F-R4 primer pairs resulted in the detection of PCR products at the expected size of 229, 558, 852, and 1,000 bp ( Fig 2D; lanes 7-10). Of note, these PCR products precisely correspond to bands obtained with genomic DNA (lanes 1-4). By contrast, the F-R5 and F-R6 primer pairs did not yield detectable PCR products with cDNA samples from chromatin-bound RNA (lanes 11-12) even though PCR products were obtained with the genomic DNA control (lanes 5-6). These observations indicate that the p53 pre-mRNA is cleaved in between approximately 1,000 to 1,400 nt downstream of the PAS, in the region where the AT-rich sequence is located. To ascertain that this cleavage event is not linked to the presence of an alternative PAS, we adopted the same mapping strategy using chromatin-bound pre-mRNA, but reverse transcription was performed with an oligo-dT primer. A cDNA derived from an mRNA transcript was included as a control. No bands were detected with all primer pairs used previously, except for the control (Fig 2E). In addition, nucleoplasmic RNA was reverse transcribed using random primers and the obtained cDNA was amplified by PCR as in Fig 2C. The bands obtained show that the CoTC-cleaved RNA can be detected in the nucleoplasm ( Fig 2F). Altogether, these data indicate that the p53 pre-mRNAs cleaved in the vicinity of the AT-rich region (i) do not contain a poly(A) tail, (ii) are generated through a CoTC-type event in a UV-induced manner, and (iii) can be found in the nucleoplasm before PAS cleavage.
Consistently, using single-molecule fluorescence in these in situ hybridization (smFISH; Fig 3A), we found that p53 pre-mRNA regions downstream of the PAS (probe B) were detected following UV exposure (median number of smFISH spots: no UV = 7; with UV = 6, with n = 1,797 and 2,165 cells, respectively; Fig 3B and Appendix Fig S2). This is not true for GAPDH (probe D), as expected for a pre-mRNA that undergoes efficient co-transcriptional PAS cleavage and no CoTC-type cleavage. Without UV exposure, GAPDH downstream regions targeted by probe D were visible and localized to the transcription sites (46% cells containing 1 or more active sites, n = 8,448 cells; Fig 3B). However, following UV exposure, the downstream regions targeted by probe D were observed less frequently in the nucleus (26% cells containing 1 or more active sites, n = 6,481 cells). It is possible that these GAPDH spots in the nucleus reflect transcription past the termination sequence, which is now understood to be widespread (Vilborg et al, 2015). Notably, we rarely observed more than 3 such sites (< 5% of cells), which is consistent with chromatin-bound transcripts at the site of synthesis in stark contrast to the more abundant p53 nuclear transcripts. Finally, RNA regions upstream of the PAS were detected for both p53 (probe A; median number of smFISH spots: no UV = 18; with UV = 26, with n = 1,797 and 2,165 cells, respectively) and GAPDH (probe C; Fig 3B), as expected for mature mRNAs. Thus, our smFISH data are consistent with our RT-qPCR data on chromatin (Fig 2) indicating that PAS cleavage of p53 pre-mRNA occurs posttranscriptionally.
Of note, we have previously shown that hnRNP H/F (Decorsi ere et al, 2011) and DHX36 (Newman et al, 2017) are involved in the regulation of p53 pre-mRNA 3 0 -end processing following UVinduced DNA damage. Consistent with p53 pre-mRNA 3 0 -end processing mostly occurring in the nucleoplasm, the increased uncleaved/total ratio of p53 pre-mRNA following the depletion of DHX36 or hnRNP H/F was significantly higher in the nucleoplasm than in the chromatin (Appendix Fig S3).
A Western blot of pre-mRNA 3 0 -end processing factors in response to UV treatment (40 J/m 2 ) of A549 cells (n = 3), followed by 16 h of recovery. GAPDH was used as a loading control. B Western blot analysis of PCF11 expression in A549 cells (n = 3) transfected for 48 h with two different siRNA targeting PCF11 prior to exposure to UV. C RT-qPCR measuring relative PCF11 mRNA level (n = 3) in A549 cells in response to UV treatment (40 J/m 2 ). The expression was normalized to RP18S. D Scheme representing the RT-qPCR strategy for assessing pre-mRNA 3 0 -end processing efficiency. Primers for uncleaved pre-mRNAs are located downstream of the polyadenylation site, while primers that detect both processed (cleaved) and unprocessed RNAs (uncleaved) amplify upstream of the polyadenylation site. The ratio of uncleaved/total (cleaved + uncleaved) indicates the processing efficiency, where a greater uncleaved/total ratio represents a reduced processing. E RT-qPCR assay on nuclear RNA for assessing the uncleaved/total ratio of p53 and TATA-binding protein (TBP) pre-mRNAs in A549 cells (n = 3) transfected for 48 h with two different siRNA targeting PCF11 prior to exposure (UV+) or not (UVÀ) to UV irradiation (40 J/m 2 ).
Data information: "n" indicates the number of biological replicates for each experiment. All data are presented as the mean AE s.e.m. P-values were calculated using a two-sided unpaired t-test. Source data are available online for this figure.

of 15
The EMBO Journal 42: e112358 | 2023 Ó 2023 The Authors The CoTC site is implicated in the maintenance of p53 pre-mRNA 3 0 -end processing in response to UV-induced DNA damage In order to determine the importance of the CoTC site in p53 pre-mRNA 3 0 -end processing following UV, the p53 CoTC element was deleted using a CRISPR-based strategy in both A549 lung tumor and A375 melanoma cells (Appendix Fig S4A). We obtained an A549 clone with deletion of the CoTC site in all three TP53 alleles existing in these cells (hereafter called DCoTC) and several A549 and A375 clones with deletion of only a subset of alleles (hereafter called pDCoTC; Appendix Fig S4B). DCoTC, pDCoTC, and WT cells were then tested for p53 pre-mRNA 3 0 -end processing efficiency in response to UV. We observed an increase in the PAS-uncleaved to total ratio for p53 in UV-treated DCoTC but not WT cells ( Fig 4A). Similar results were obtained with pDCoTC A549 (Appendix Fig S5A) and pDCoTC A375 (Appendix Fig S5B) cells. As a control, the deletion of the p53 CoTC region had no effect on the UV-dependent regulation of pre-mRNA 3 0 -end processing for WDR13, GAPDH, and TBP ( Fig 4A and Appendix Fig S5). Thus, the p53 CoTC region is required for the maintenance of p53 pre-mRNA 3 0 -end processing upon UV exposure. We also found a decrease in the nucleoplasm/chromatin ratio of the p53 PAS-uncleaved pre-mRNA in DCoTC cells when compared to WT cells ( Fig 4B). This effect was also observed in pDCoTC A549 (Appendix Fig S6A) and pDCoTC A375 (Appendix Fig S6B) cells and was not observed for the WDR13, GAPDH, and TBP pre-mRNAs ( Fig 4B and Appendix Fig S6). This shows that the p53 CoTC site is required for the release of the PASuncleaved p53 pre-mRNA from chromatin to nucleoplasm in response to UV. Consistently, total p53 mRNA levels were decreased in DCoTC and pDCoTC cells, but not in WT cells, in response to UV (Fig 4C and Appendix Fig S7A). In addition, the UV-dependent increase in p53 protein levels in WT cells was not observed in DCoTC and pDCoTC cells ( Fig 4D and Appendix Fig S7B). Altogether, these data show that the CoTC site of p53 is required to maintain p53 PAS cleavage and promote p53 expression following UV irradiation.
We then assessed the potential consequences of CoTC site deletion on downstream functions of p53. A direct transcriptional target of the p53 protein is the CDKN1A/p21 gene, which encodes an inhibitor of cell cycle progression from G1 to S phase (Jeong et al, 2010;Galanos et al, 2016;Matsuda et al, 2017). UV-induced upregulation of p21 mRNA and p21 protein levels was observed in WT cells but not in DCoTC and pDCoTC cells (Fig 4C and D,and Appendix Fig S7A and B). Analysis of cell cycle distribution by FACS showed no effect of CoTC site deletion in the absence of UV ( Fig 4E, left panels). However, a moderate UV treatment, which had no effect on cell cycle distribution in WT cells, led to a decrease in G0/G1 cells in DCoTC and pDCoTC cells (Fig 4E and Appendix Fig S8). Altogether, these data suggest that deletion of the p53 CoTC site leads to impaired p21 induction and enhanced G1-S phase progression following moderate UV irradiation.
The 3 0 -end processing of several pre-mRNAs undergoing a CoTC cleavage event is maintained in response to UV-induced DNA damage Our finding that the CoTC site of p53 pre-mRNA is required for p53 to escape 3 0 -end processing inhibition by UV prompted us to investigate whether CoTC-dependent cleavage may be linked to UVresistant pre-mRNA 3 0 -end processing in other genes. Toward this aim, we first developed a strategy to analyze in a high-throughput manner the efficiency of pre-mRNA 3 0 -end cleavage by RNAsequencing (RNA-Seq). This strategy is based on the evaluation of the number of reads located in 500 nt-long windows either upstream (total RNA) or downstream of the PAS (PAS-uncleaved RNA; Fig 5A). An increase in the ratio of downstream reads to upstream reads indicates inhibition of 3 0 -end cleavage, leading to read-through transcription (Vilborg et al, 2015). Focusing on 4,208 expressed genes (Dataset EV1) with detectable reads downstream of the PAS and using a cut-off of P < 0.05, this analysis identified 378 pre-mRNAs with UV-repressed 3 0 -end processing and 108 pre-mRNAs with a more efficient 3 0 -end processing in UV-treated compared with untreated cells (UV-resistant 3 0 -end processing; Fig 5B). Examples of the read distribution in the 500 nt-long windows located upstream and downstream of the PAS are illustrated in Fig 5C for the ZRANB2 and the HMGB1 genes. In the case of ZRANB2 that belongs to the UV-repressed 3 0 -end processing group, the absence of reads downstream of the PAS in untreated cells indicates a very efficient 3 0 -end processing activity, while the presence of more reads in this window in UV-treated cells indicates a reduced 3 0 -end processing efficiency following UV treatment. In the case of HMGB1 that belongs to the UV-resistant 3 0 -end processing group, we observed an opposite trend showing that there is an increase in pre-mRNA 3 0 -end processing following UV treatment ( Fig 5C).
We randomly chose 9 candidate genes from each group and validated the RNA-Seq data by RT-qPCR on PAS-uncleaved and total RNA using primers located downstream or upstream of the PAS, respectively ( Fig 6A). These RT-qPCR analyses showed that all candidate pre-mRNAs of the UV-repressed group exhibited 3 0 -end processing inhibition by UV (Fig 6A, red bars). By contrast, all ◀ Figure 2. p53 pre-mRNA 3 0 -end cleavage occurs in the nucleoplasm following a CoTC.
A Western blot analysis to verify the quality of the nuclear fractionation of UV-treated A549 cells (n = 3; 40 J/m 2 ). The topoisomerase II alpha was used as a marker of the nucleoplasm compartment and the histone H3 for the chromatin fraction. B RT-qPCR analysis on RNA extracted from nucleoplasm and chromatin fractions. The ratio of uncleaved pre-mRNA (nucleoplasm/chromatin) was calculated to quantify the level of unprocessed p53, TBP, WDR13 and GAPDH pre-mRNA (n = 3 biological replicates) released in the nucleoplasm compared with the chromatin-bound unprocessed pre-mRNA. WDR13 and GAPDH were included as controls as they have been previously reported to be processed post-and co-transcriptionally, respectively. Data are presented as the mean AE s.e.m. C Scheme representing the strategy to map the location of the CoTC site in the p53 pre-mRNA. Forward primer (F) is located upstream of the PAS (poly(A) site) and reverse primers (R1-6) are located downstream at increasing distances from the PAS. D RT-PCR analysis of p53 3 0 flanking region using primer pairs indicated (black arrows) in the scheme above the data panel (n = 3). Lanes 1-6 correspond to amplified genomic DNA used as a PCR amplification control. Lanes 7-12 correspond to cDNA derived from chromatin-associated RNA reverse transcribed (+RT) using random primer. Lanes 13-18 are negative RT control samples (ÀRT). E RT-PCR analysis of p53 3 0 flanking region using the same primers employed in the data panel above (n = 3). Lane 1 is a control PCR amplification of cDNA derived from the reverse transcription of a control mRNA using oligo (dT). Lanes 2-7 are PCR amplification of reverse transcribed p53 chromatin-associated pre-mRNA using oligo oligo (dT). F RT-PCR analysis of p53 3 0 flanking region (n = 3). cDNAs were derived from nucleoplasmic-associated RNA reverse transcribed using random primer.
Source data are available online for this figure.
6 of 15 The EMBO Journal 42: e112358 | 2023 Ó 2023 The Authors candidate pre-mRNAs of the UV-resistant group were resistant to 3 0end processing inhibition by UV (Fig 6A, blue bars). These include several DDR-related genes, namely XRCC5 and HMGB1. Because our RNA-seq approach is limited by the sensitivity of read detection downstream of the PAS, we also analyzed six genes from the p53 pathway by RT-qPCR. One of them (i.e., NOXA) exhibited 3 0 -end processing inhibition by UV, while four genes escaped repression (i.e., MDM2, PUMA, PTEN, and FAS; Fig 6A, right). Altogether, our RT-qPCR analyses identified 15 pre-mRNAs that were resistant to 3 0 -end processing inhibition by UV, and 12 pre-mRNAs that underwent inhibition. Then, we measured by RT-qPCR the nucleoplasm/chromatin ratio of PAS-uncleaved pre-mRNA. Remarkably, all 15 UV-resistant pre-mRNAs had a high nucleoplasm/chromatin ratio, indicating that PAS cleavage occurs at least in part post-transcriptionally ( Fig 6B). By contrast, none of the 12 UV-repressed pre-mRNAs were abundant in the nucleoplasm ( Fig 6B). These data show that resistance of PAS cleavage to UV correlates with its occurrence in the nucleoplasm, thus extending our findings on p53.
Because we found that the UV resistance and nucleoplasmic occurrence of the p53 pre-mRNA PAS cleavage require a downstream CoTC cleavage site (Fig 3), we then used our PCR-based mapping strategy (Fig 2) to locate putative CoTC elements in eight other UV-resistant pre-mRNAs (Appendix Fig S9A and B). For the eight tested genes, the presence of CoTC cleavage sites was validated by the sharp loss of amplification at a given distance (most often about 2.5 kb) downstream of the PAS (Appendix Fig S9C). As for p53 (Fig 2E), the CoTC-dependent cleavage event for the eight tested genes is not linked to the presence of an alternative PAS (Appendix Fig S9D).
Gene Ontology (GO) analysis of pre-mRNAs with a more efficient 3 0 -end processing in UV-treated compared with untreated cells shows an enrichment of genes involved in the inhibition of apoptosis and processes related to two genotoxic stress-inducing agents, doxorubicin and daunorubicin ( Fig EV2). Since we previously showed that p53 pre-mRNA 3 0 -end processing is maintained upon doxorubicin treatment (Decorsi ere et al, 2011), we studied the involvement of a CoTC-based regulation in doxorubicin-treated cells. We observed an increase in the PAS-uncleaved to total ratio for p53 in UV-treated DCoTC and pDCoTC but not in A549 WT cells (Fig EV3 and Appendix Fig S10). Thus, the p53 CoTC region is required for the maintenance of p53 pre-mRNA 3 0 -end processing upon doxorubicin treatment. In addition, RT-qPCR analyses showed that 11 out of 12 UV-repressed pre-mRNAs also exhibited 3 0 -end processing inhibition by doxorubicin and that all 15 pre-mRNAs of the UVresistant group were also resistant to 3 0 -end processing inhibition by doxorubicin (Appendix Fig S11). Altogether, these results provide evidence for an association between maintained 3 0 -end processing following UV irradiation/doxorubicin treatment and the presence of a CoTC element in several genes including DDR-related genes (XRCC5, PUMA, FAS, MDM2, TP53).

Discussion
Pre-mRNA 3 0 -end processing by PAS cleavage and poly(A) tail addition mostly occurs in a co-transcriptional manner. We show here that PAS cleavage of the p53 pre-mRNA occurs at least in part in a manner that is uncoupled from transcriptional termination. It involves a CoTC sequence that lies about 1.2 kb downstream of the PAS and allows a first 3 0 -end cleavage event, leading to the dissociation of the pre-mRNA from chromatin. This is followed by a second 3 0 -end cleavage event (and polyadenylation) occurring at the PAS of the released RNA in the nucleoplasm.
One of the important factors in the coupling between 3 0 -end processing and transcription termination is PCF11, a 3 0 -end processing factor that mediates transcriptional termination in yeast (Grzechnik et al, 2015;Larochelle et al, 2018), as well as in vertebrates (Kamieniarz-Gdula et al, 2019). PCF11 recruits the yeast Rat1 or human Xrn2 exonucleases to exert a 5 0 -3 0 exonucleolytic degradation on the nascent RNA leading Pol II to terminate transcription (Luo, 2006;West & Proudfoot, 2007;Eaton et al, 2018). We show that pre-mRNA 3 0 -end processing is inhibited in the absence of PCF11 (Fig 1). PCF11 interacts with CLP1 to target the cleavage site and modulates the binding and cleavage efficiency of CFII (Zhang et al, 2021). It also regulates polyadenylation site choice and plays a role in controlling the 3 0 Untranslated Region (3 0 UTR) of transcripts (Ogorodnikov et al, 2018;Wang et al, 2019;Nourse et al, 2020). However, p53 pre-mRNA 3 0 -end processing is independent of PCF11 in UV-treated cells, allowing this pre-mRNA to escape from the decrease in PCF11 levels observed in UV-treated cells (Fig 1). Since PCF11 has a CTD Interacting Domain (CID) and binds preferentially the phosphorylated Ser2 of an elongating RNAP II (Meinhart & Cramer, 2004), our results clearly indicate an uncoupling of 3 0 -end processing and transcriptional termination in the p53 pre-mRNA following UV-induced DNA damage.
We demonstrate that p53 pre-mRNA 3 0 -end processing does not require PCF11 because it is processed at a CoTC sequence (Figs 2  and 3). CRISPR-based deletion of the p53 CoTC leads to inhibition of p53 pre-mRNA 3 0 -end processing, decreased p53 and p21 protein levels, and decreased G0/G1 cells in UV-treated cells (Fig 4). This is consistent with the fact that UV radiation-induced cell cycle arrest is correlated with increase in p53 levels (Latonen et al, 2001) and causes retention of cells at the G2-M phase (C eraline et al, 1998;van Oosten et al, 2000;Pavey et al, 2001;Blackford & Jackson, 2017). CoTC-dependent cleavage therefore acts as a mechanism of escape from UV-induced global inhibition of pre-mRNA 3 0 -end processing. Beyond p53, the presence of CoTC elements in the 3 0 flanking regions of a number of genes, including genes implicated in p53-mediated DNA damage response, was validated (Figs 5 and 6).
We thus propose a model where UV-induced or doxorubicintriggered DNA damage inhibits co-transcriptional PAS cleavage, which is chromatin-bound and PCF11-dependent, but not post-◀ Figure 3. PAS cleavage of p53 pre-mRNA occurs post-transcriptionally.
A Probe design for single-molecule Fluorescence in situ hybridization (smFISH) spanning regions upstream "A" and downstream "B" of p53 PAS, as well as those spanning regions upstream "C" and downstream "D" of GAPDH PAS. B Representative images of smFISH in untreated (ÀUV) or UV-treated (+UV) A549 cells with the indicated probes (A, B, C, and D).

of 15
The EMBO Journal 42: e112358 | 2023 Ó 2023 The Authors transcriptional PAS cleavage that is nucleoplasmic and PCF11independent and occurs following a CoTC-dependent release of the pre-mRNA from chromatin. Our model thus explains the rescue of the 3 0 -end processing of specific mRNAs in a transcriptionuncoupled manner, despite the global inhibition by UV-induced DNA damage of the canonical chromatin-associated pre-mRNA ◀ Figure 4. The CoTC is implicated in the maintenance of 3 0 -end processing of p53 pre-mRNA in response to UV-induced DNA damage.
A RT-qPCR assay on nuclear RNA for assessing the uncleaved/total ratio of p53 pre-mRNA in wild type (WT) and CoTC-deleted (DCoTC) A549 cells (n = 3) treated with or without UV irradiation (40 J/m 2 ). B RT-qPCR analysis on RNA extracted from nucleoplasm and chromatin fractions. The ratio of uncleaved pre-mRNA (nucleoplasm/chromatin) was calculated to quantify the level of unprocessed p53, WDR13, GAPDH, and TBP pre-mRNA (n = 3) released in the nucleoplasm compared with the chromatin-bound unprocessed pre-mRNA in wild type (WT) and CoTC-deleted (DCoTC) A549 cells treated with or without UV irradiation (40 J/m 2 ). C RT-qPCR measuring relative p53 and p21 mRNA levels in wild type (WT) and CoTC-deleted (DCoTC) cells (n = 3) in response to UV treatment (40 J/m 2 ). The expression was normalized to HPRT. D Western blot analysis of p53 and p21 expression in wild type (WT) and CoTC-deleted (DCoTC) A549 cells (n = 3) treated with or without UV irradiation (40 J/m 2 ). E Representative flow-cytometry analyses of the cell cycle (DNA content by Propidium Iodide; PI) in wild type (WT) and CoTC-deleted (DCoTC) A549 cells (n = 3) treated with or without UV irradiation (40 J/m 2 ). Indicated: percent of cells in the G0-G1, S, and G2/M phases.
Data information: "n" indicates the number of biological replicates for each experiment. All data are presented as the mean AE s.e.m. P-values were calculated using a two-sided unpaired t-test.
Source data are available online for this figure. A Scheme representing the adopted strategy to study at genome wide level by RNA-sequencing the regulation of pre-mRNA 3 0 -end processing in response to UV. Nuclear RNA from A549 cells UV-irradiated or -unirradiated was extracted and cDNA library preparation was performed to assess RNA-sequencing. The efficiency of 3 0 -end processing was studied by the quantification of the number of reads located in a window of 500 bp downstream (uncleaved RNA) and upstream (total RNA) of the poly (A) site. The ratio of reads downstream / reads upstream reflects the efficiency of pre-mRNA 3 0 -end processing. B RNA-seq plot representing, for each gene, the log2 fold change (comparing UV-irradiated to nonirradiated cells) of the number of reads downstream of the poly(A) site (uncleaved RNA, y axis), and the log2 fold change of the number of reads upstream of the poly(A) site (total RNA, x axis; n = 3). Regulation events are considered significant if P-value is below 0.05. The 108 genes at bottom right have a decreased uncleaved/total RNA ratio, meaning that PAS cleavage is increased upon UV irradiation. By contrast, the 378 genes at top left have an increased uncleaved/total RNA ratio, meaning that PAS cleavage is decreased upon UV irradiation. C Visualization of reads distribution in a window of 500 nts upstream and downstream of the PAS of HMGB1 and ZRANB2 pre-mRNA. ▸ Figure 6. The 3 0 -end processing of diverse pre-mRNAs undergoing a CoTC cleavage event is maintained in response to UV-induced DNA damage.
A RT-qPCR (uncleaved/total RNA) on nuclear RNA extracted from UV-treated or -untreated A549 cells (n = 3), to assess the regulation of 3 0 -end processing of 20 pre-mRNAs randomly selected from the previous RNA-sequencing data. B RT-qPCR (Released uncleaved pre-mRNA nucleoplasm/chromatin) on RNA extracted from the nucleoplasm and the chromatin fractions of A549 UV-treated cells (n = 3).
Data information: "n" indicates the number of biological replicates for each experiment. All data are presented as the mean AE s.e.m. Source data are available online for this figure.
10 of 15 The EMBO Journal 42: e112358 | 2023 Ó 2023 The Authors . Nucleoplasmic PAS-dependent 3 0 cleavage occurs following a CoTC-dependent release of the pre-mRNA, thereby acting as a compensatory mechanism to maintain the expression of genes involved in the p53 pathway and DNA damage response.

Materials and Methods
Cell culture, siRNA transfections, and UV irradiation A549 cells were cultured in DMEM (Eurobio) containing 10% FCS (Pan Biotech) and L-Glutamine (Eurobio) at 37°C in 5% CO 2 . siRNA reverse transfections were performed in 10 cm with Lipofectamine RNAiMAX (Thermo Scientific) at a final concentration of 20 nM siRNA (Eurogentec or Dharmacon; see Appendix Table S1) as per the manufacturer's instructions in OptiMEM reduced serum media (Thermo Scientific). After 48 h transfection, cells were washed with PBS and irradiated with 40 J/m 2 UV (254 nm; Stratalinker), placed in fresh media, and harvested on ice after 16 h of recovery at 37°C.

CRISPR-mediated deletion of CoTC element
CRISPR sgRNAs were designed for the CoTC element deletion of the p53 gene (Appendix Fig S2). sgRNAs were designed using the online tool http://crispr.mit.edu/. Guide sequences are identified that minimize identical genomic matches or near-matches to reduce the risk of cleavage away from target sites (off-target effects). The guide sequences are constructed such that they consist of 20-mer protospacer sequence upstream of an NGG protospacer adjacent motif (PAM) at the genomic recognition site (Appendix Table S2). Two sgRNA oligos are constructed, each of 24-25 mer oligos and their associated reverse complement including additional nucleotides for cloning and expression purposes.
The two plasmids used are namely, pSpCas9 (BB) plasmid pX458 and pX459, which include GFP and puromycin as selectable markers, respectively.
i First the sequences CACC and AAAC are added before the 20-mer guide sequence and the guide's reverse complement for cloning into pX458/pX459 vectors using BbsI restriction enzyme (Appendix Table S3). ii A G nucleotide is added after the CACC sequence and before the 20-mer if the first position of the 20-mer is not G. sgRNA expression from the U6 promoter of the pX458/pX459 vector is enhanced by the inclusion of a G nucleotide after the CACC sequence. iii A C nucleotide is added at the 3 0 -end of the reverse complement oligo. All resultant oligos are 25-mer oligos. The sgRNA oligo sequences were cloned into the pX458 and pX459 plasmids using a Golden Gate assembly cloning strategy (Appendix Fig S3B). The plasmids were amplified followed by transfection. Selection of transfected cells was carried out in puromycincontaining medium. The cells are incubated for a total of 48-72 h after transfection before harvesting for indel analysis.
Primers were designed surrounding the sgRNA cleavage sites for PCR and screening for CRISPR/Cas9 screening deletion (Appendix Fig S3). gDNA was isolated from control or transfected cells, and PCR is performed to validate the primers and verify the presence of the intended genomic deletion.

Cell fractionation
Cell pellets were resuspended in approximately 3× cell pellet volume of lysis Buffer A (10 mM HEPES pH 7.9, 15 mM MgCl 2 , 10 mM KCl, 0.1% NP40, 1 mM DTT) containing RNAseOut (Thermo Scientific) and incubated on ice for 15 min. Cells were then pelleted at 1,000 g for 5 min at 4°C and the supernatant retained for cytoplasmic RNA. Nuclear pellets were washed in 2 × 1 ml Lysis Buffer A at 1,000 g for 5 min at 4°C, resuspended in 2 × pellet volume with Nuclear Lysis Buffer B (20 mM HEPES pH 7.9, 400 mM NaCl, 1.5 M MgCl 2 , 0.2 mM EDTA, 5 mM DTT) containing RNAseOut and incubated on ice for 30 min. Nuclear debris was pelleted at 10,000 g for 15 min at 4°C, and supernatants were placed in Trizol Reagent for RNA extraction.

Nuclear fractionation
Cell nuclei were suspended in 1× nuclei pellet volume of buffer 1 (20 mM Tris-pH 7.9, 75 mM NaCl, 0.5 mM EDTA, 0.85 mM DTT, 0.125 mM PMSF, 0.1 mg of yeast tRNA/ml, 50% glycerol) and 10× nuclei pellet volume of buffer 2 (20 mM HEPES, 300 mM NaCl, 1 mM DTT, 7.5 mM MgCl2, 0.2 mM EDTA, 1 M Urea, 1% NP-40, 0.1 mg of yeast tRNA/ml). After vigorous agitation for 5 s, nuclei pellet was incubated for 10 min on ice. Chromatin fraction was then sedimented by full-speed centrifugation for 5 min at 4°C. The supernatant corresponding to the nucleoplasmic fraction was transferred to a new tube, adjusted to 0.1% of SDS, and trizol RNA extraction was proceeded. The insoluble fraction corresponding to the chromatin was resuspended in buffer 3 (10 mM Tris-pH 7.5, 10 mM MgCl2, 500 mM NaCl), and 20 U of DNAse was added before 30 min incubation at 37°C. RNA extraction was then proceeded.

RT-qPCR and RT-PCR
cDNA was synthesized using Superscript III (Thermo Scientific). qPCR on cDNA derived from nuclear pre-mRNAs or cytoplasmic mRNAs was performed using 2× Power Sybrgreen Master Mix (Thermo Scientific) and 0.4 lM oligonucleotide primers (Appendix Table S4). The ratio of uncleaved/total RNA was calculated using 2 (total-uncleaved) , and the ratio of released uncleaved RNA nucleoplasm/ chromatin-associated uncleaved pre-mRNA was calculated using 2 (Chromatin-nucleoplasm) . For the RNA-IP, samples were normalized to the input using 2 (Input-IP) . When conducting RT-PCR, cDNA amplification was performed using Go-Taq flexi DNA polymerase (Promega). PCR products were then applied to 1% agarose gel.

Western blot
Cells were harvested on ice and pelleted by centrifugation at 400 g for 5 min at 4°C. Pellets were resuspended in RIPA Buffer containing complete protease inhibitors (EDTA-free) and sonicated (Bioruptor,

Single-molecule fluorescence in situ hybridization (FISH)
A549 cells were cultured on #1.5 cover glasses in 12-well plates.
When cells were at approximately 50% confluency, the cover glass was washed once with PBS. Each cover glass was placed in one well of a 12-well plate that was filled with 200 ll PBS, which just barely covered the top of the cover glass. Half of the cover glasses were irradiated with UVC at 50 J/m 2 . After irradiation, cover glasses were returned to the culture medium and placed in the incubator for 4 h. After 4 h, cells on cover glasses were washed 3 times with HBSS before fixation with 4% PFA in PBS. Fixed samples were washed with 1× PBS and stored in 70% ethanol at 4°C overnight. FISH probes were designed and ordered from Biosearch Stellaris using Quasar 570 and 670 fluorophores. Hybridizations were performed according to the manufacturer's protocol with minor modifications. Hybridized samples were mounted in Prolong Gold with DAPI and allowed to dry overnight.
Imaging of FISH was performed on a custom-built microscope. This microscope comprised an ASI (www.asiimaging.com) Rapid Automated Modular Microscope System (RAMM) base, a Hamamatsu ORCA-Flash4 V2 CMOS camera (https://www.hamamatsu. com/, C11440), Lumencore SpectraX (https://lumencor.com/), an ASI High Speed Filter Wheel (FW-1000), and an ASI MS-2000 Small XY stage. Excitation of DAPI, Quasar 570, and 670 was performed using SpectraX violet, red, and green, respectively. Emission filters specific to these spectra were used. Image acquisition was performed through Micro-Manager. We obtained multiple z stacks at 250 ms exposures, 0.5 lm intervals, spanning 3.5 lm. The maximum intensity projections were performed and used for transcript localization and analysis.
FISH analysis was performed with custom MATLAB software. Briefly, images of cells were segmented into the nucleus and cytoplasm. Spots were localized with custom MATLAB software using an algorithm based on Thompson et al. The software outputs the number of nuclear and cytoplasmic spots per cell, and the distribution of spots per cell (Appendix Fig S11).

Propidium iodide staining
Cells were harvested in ice and washed with PBS. They were fixed in 70% ethanol for 30 min at À20°C. They were washed twice in PBS pelleted by centrifugation at 850 g for 5 min at 4°C. The cells were then resuspended in a solution containing 3.5 mM Tris-HCl pH 7.6 (Thermo Scientific), 10 mM NaCl (Thermo Scientific), 50 lg/ml propidium iodide (Sigma P4170), 0.1% IGEPAL (Thermo Scientific), 20 lg/ml RnaseA (Sigma), and water. The acquisition of stained cells was performed using an LSRII flow cytometer (BD Biosciences). The acquired data were analyzed using FlowJo software.

RNA-sequencing analysis
For RNA-seq, nuclear RNA from UV-irradiated and nonirradiated A549 cells (two biological replicates of each condition) was subjected to DNAse I treatment with TURBO DNase I (ThermoFisher Scientific), quantified, and analyzed using an RNA 2100 Bioanalyzer (Agilent). 500 ng of good quality RNA (RIN > 9) was used for Illumina compatible library preparation using the TruSeq Stranded total RNA protocol allowing to take into account strand information. A first step of ribosomal RNA depletion was performed using the Ribo-Zero Gold kit (Illumina). After fragmentation, cDNA synthesis was performed and resulting fragments were used for dA-tailing followed by ligation of TruSeq indexed adapters. PCR amplification was finally achieved to generate the final barcoded cDNA libraries. Libraries were equimolarly pooled. Sequencing was carried out on a HiSeq instrument (Illumina) to obtain around 40 million raw singleend reads of 100 nucleotides per sample.
Fastq files were generated using bcl2fastq. RNA-seq reads of good quality were trimmed in their 5 0 -and 3 0 -ends with the cutadapt software to remove uninformative nucleotides due to primer sequences. Trimmed reads of 100 bp or more were aligned on the Human genome (hg19) using Tophat2. Only reads with a mapping quality score of 20 or more were retained (samtools) for downstream analysis. Gene coordinates were obtained on the basis of overlapping Refseq transcripts with the same gene symbol. For each gene, two 500 bp regions located upstream and downstream of the PAS at the end of the gene were defined (genes with a downstream region overlapping another gene were discarded), and reads located in the upstream and downstream regions were counted in each sample. A table of counts was built with the featureCounts software (R version 3.4.0). Only genes with at least 10 reads in both regions in either condition were kept for further analysis. In total, 4,208 genes passed all these steps and were used for subsequent analysis. The differential analysis between the UV+ and UVÀ conditions was done using two independent biological replicates per condition. For each gene, the fold regulation of the downstream region (that is the ratio of normalized read counts between conditions) was compared with the fold regulation of the upstream region using a Wald test implemented in DESeq2 (Love et al, 2014).
Gene ontology (GO) analysis of genes was carried out by the functional enrichment analysis tool DAVID.

Statistics
Statistical differences between experimental and control samples were assessed by the unpaired t-test using GraphPad Prism, with significance achieved at P < 0.05.

Data availability
The datasets produced in this study are available in the following databases: Gene Expression Omnibus repository (GEO) under accession number GSE203517 (https://www.ncbi.nlm.nih.gov/geo/ query/acc.cgi?acc=GSE203517). All other data generated or analyzed during this study are included in the manuscript.
Expanded View for this article is available online. EV2 Figure EV3. The p53 CoTC region is required for the maintenance of p53 pre-mRNA 3 0 -end processing upon doxorubicin treatment.
RT-qPCR assay on nuclear RNA for assessing the uncleaved/total ratio of p53 pre-mRNA in wild type (WT) and CoTC-deleted (DCoTC) A549 cells treated with or without doxorubicin (3.5 lM). (n = 3); P-values were calculated using a two-sided unpaired t-test.

Validation of PAS cleavage inhibition of candidate pre-mRNAs in
"n" indicates the number of biological replicates for each experiment. All data are presented as the mean ± s.e.m. P values were calculated using two-sided unpaired t-test.    Appendix Figure S10. Partial deletion of the CoTC element inhibits p53 PAS cleavage in response to doxorubicin. RT-qPCR assay on nuclear RNA for assessing the uncleaved/total ratio of p53 pre-mRNA in wild type (WT) and partial CoTC deleted (pΔCoTC) A549 cells (n=3) treated with or without doxorubicin (3.5 μM). "n" indicates the number of biological replicates for each experiment. All data are presented as the mean ± s.e.m. P values were calculated using two-sided unpaired t-test. Appendix Figure S11. Validation of PAS cleavage inhibition of candidate pre-mRNAs in response to doxorubicin. RT-qPCR (uncleaved/total RNA) on nuclear RNA extracted from doxorubicin-treated or untreated A549 cells (n=3), to assess the regulation of 3'end processing of 20 pre-mRNAs randomly selected from the previous RNA-sequencing data. "n" indicates the number of biological replicates for each experiment. All data are presented as the mean ± s.e.m.