PARP dependent acetylation of N4-cytidine in RNA appeared in UV-damaged chromatin

Posttranscriptional RNA modications, including the presence of methyl-6-adenosine (m6A), methyl-5-cytosine (m5C), or pseudo-uridine (Ψ), are known for over many years, but their functional properties have not been fully elucidated yet. Similarly, the regulatory role of N4-cytidine (ac4C) acetylation in RNA must be explored. Here, we observed PARP-dependent accumulation of ac4C RNA at UVA-microirradiated chromatin, which appears 2-5 minutes after genome injury, simultaneously with m 6 A RNAs but with distinct kinetics. When m 6 A RNAs disappeared from the lesions, the high level of ac4C RNA was maintained up to 20 minutes after genome injury. Surprisingly, the process of ac4C RNA accumulation at DNA lesions was not accompanied by the recruitment of acetyltransferase NAT10 to UVA-induced DNA lesions. This process was PARP dependent, and data show how epitranscriptomic features can contribute to DNA damage repair.


Introduction
In general, the acetylation process is a well-described cellular mechanism that regulates gene expression, especially when acetylation appears on the level of histones [1]. However, also N4-acetylcytidine (ac4C), a highly conserved RNA nucleobase, also contributes to the regulation of mRNA stability and translation [2]. On the other hand, information about the function of ac4C RNA in the DNA repair process has not been described yet, but speci c histone acetylation was well described at DNA lesions [3]. For instance, we have also observed HDAC1-dependent deacetylation of histone H3 at lysine 9 position at experimentally induced DNA lesions [4], while Meyer et al. (2016) [5] showed that H3K9 acetylation, a proactivation histone mark, prevents H3K9 methylation; and thus, weakens H3K9me2/3-dependent DNA repair processes. To this fact, Dhar et al. (2017) [6] revised that the histone H4 terminal tails recruit proteins involved in DNA damage repair (DDR), including 53BP1, an important factor of non-homologous end joining (NHEJ) repair. A process mediated by H4 acetylation that provides binding sites for bromodomain proteins, such as ZMYND8 and BRD4, is essential for the repair of double-strand breaks (DSBs) [7,8]. In addition, Jacquet et al. (2016) [9] showed that acetylation of histone H2A abrogates a pronounced H2A ubiquitination at the same lysine position. A very fundamental observation is that H2A.Z exchange at DSBs, in cooperation with cytochrome P400, leads to acetylation of H4 at damaged chromatin via the function of histone acetyltransferase, Tip60 [10,11]. Mentioned data showed that histone acetylation regulates not only gene expression but also DNA damage response. In addition to mentioned epigenetic events, here, we test a hypothesis if, and to which extent, ac4C sites in human RNAs play a role in DNA repair processes. We also took into consideration that ac4C RNA is installed in the genome via the function of Nacetyltransferase, NAT10 [12]. The target of the human enzyme NAT10 is preferentially rRNA and tRNA (tRNA-Ser and tRNA-Leu) [13,14]. To this information, Kudrin et al. (2021) [15] newly identi ed NOP58 as an ac4C-binding protein, and importantly sirtuin 7 (SIRT7) as a speci c ac4C deacetylase. So, taken together, NAT10 can be considered as the essential writer of ac4C on RNA, NOP58 as a reader, while SIRT7 seems to be a highly speci c ac4C RNA eraser. Based on the above-mentioned observation, we addressed a question of how ac4C RNAs contribute to DNA repair machinery and if regulatory protein NAT10 contributes to the process of DNA damage repair.
From this view, we additionally tried to reveal in which DNA repair pathways ac4C RNA is involved. We were inspired by the fact that another RNA modi cation, m 6 A, is signi cantly activated at UV-damaged chromatin [16,17]. Notably, there was described a non-canonical m 6 A-mediated pathway, dependent on the METTL3 and METTL14 enzymes or PARP1/2 proteins [18]. Thus, we also addressed the question if ac4C RNA recruitment to DNA lesions is PARP-dependent. In addition, m 6 A-coated mRNA binds to DNA strands to form hybrid DNA-RNA structures that are recognized by m 6 A RNA readers, including YTHDF1 and YTHDC1. This process mediates both DSB repair as well nucleotide excision repair (NER) mechanism [19]. From this view, a question remains if ac4C RNAs work in an identical way as m 6 A RNAs at DNA lesions.

Cell cultivation and Treatment
Mouse embryonic broblasts (MEFs) and human breast cancer cell line MCF7 were maintained in Dulbecco's modi ed Eagle's medium (DMEM) supplemented with 10% fetal bovine serum FCS (Merck, Darmstadt, Germany), penicillin (1 U/ml), and streptomycin (100 µg/ml) at 37°C in a humidi ed atmosphere containing 5% CO2. For the experiment of dependent recruitment of ac4C on cells cycles, we To inhibit RNA polymerase I and poly ADP ribose polymerase (PARP), at 50% con uence, cells were treated by actinomycin D (#A9415, Merck, Germany), nal concentration 0.5 µg/ml, for 2 h, and by Olaparib (#S1060, Selleckchem, Germany), nal concentration 10 µM, for 24 h, before microirradiation experiments [21,22] To verify the speci city of the antibody, we used treatments by Turbo-DNase (#AM2238, ThermoFisher We observed a signi cant effect after RNase A treatment. The level of ac4C RNA was reduced in all cells, especially in the nucleolus. After RNase H1 treatment, the uorescent intensity was only gently reduced in the nucleoplasm. The signi cant changes were also observed when cells were treated with DNase I, which caused a more diffuse distribution of ac4C RNAs in nuclei of studied cells.

Irradiation by UV light
The cells seeded on 35 mm glass-bottom dishes (#D35-20-1-N, Cellvis Mountain View, CA, USA) and at 50% con uence were sensitized with 10 µM BrdU (#11296736001, Merck, Germany) for 16 h before UVA treatment. The cells were irradiated by the UVA lamps (model GESP-15, 15 W, UVA 330-400 nm wavelength with maximum e ciency at 365 nm) or UVC lamps (Philips, Amsterdam, The Netherlands, model TUV 30 W T8, UVC 254 nm wavelength) for 10 min. After UVC irradiation, the cells were xed at multiple intervals (5 min, 20 min, 60 min, and 120 min after irradiation). The lamp distance from the sample was 2 cm for the UVA source and 60 cm for the UVC source [17].
Immuno uorescence and confocal microscopy Local laser microirradiation and laser scanning confocal microscopy For the micro irradiation experiments using UVA lasers (wavelength 355 nm), cells were seeded on 35 mm gridded microscope dishes (#81166, Ibidi, Fitchburg, WI, USA) and at 50% con uence were sensitized with 10 µM BrdU for 16 to 18 h. The cells were maintained under optimal cultivation conditions in an incubation chamber (EMBL) at 37°C with 5% CO2. In the selected cell nuclei, we irradiated only the de ned region of interest (ROI) using a TCS SP5-X confocal microscope system (Leica, Wetzlar, Germany). The microscope settings for induction of local DNA damage were as follows: power laser (355 nm) 25 mW, 512×512 pixel resolution, 400 Hz, bidirectional mode, 48 lines, zoom 8 and 63x oil objective (HCX PL APO, lambda blue) with a numerical aperture (NA) = 1.4 [17]. The maximum observed irradiation duration was 45 minutes, and for the data analysis, we monitored about 100 cell nuclei. The analysis of 3 biological replicates was performed. After the immunostaining procedure, locally micro-irradiated cells were found according to registered coordinates on gridded microscope dishes. We studied the level of the epigenetic marker N4-acetylcytidine in RNA, NAT 10 acetyltransferase, and the presence of γH2AX (phospho S139), which was also used for the optimization of micro irradiation experiments.
Images were acquired by laser scanning confocal microscopy (Leica TCS SP5-X or Leica SP-8 system).
For observation and image acquisition, the following parameters were used: 1024 × 1024 pixels, 400 Hz, bidirectional scanning mode, 8 lines, zoom 5-8, and 63× oil objective (HCX PL APO, lambda blue) with a numerical aperture (NA) = 1.4. LEICA LAS X software was used for image acquisition and uorescence intensity analysis.

Statistical analysis
Fluorescence intensity values were measured by LAS X software and subsequently analyzed in Python 3.
The obtained data were compared statistically using the ANOVA One-Way test, available in GraphPad Prism software, version 9 for Windows (GraphPad Software, San Diego, CA, USA). All cases where p ≤ 0.05 were considered statistically signi cant and are presented in graphs in the results section.

ac4C RNAs recognizes UVA and UVC-induced DNA lesions
Local laser microirradiation showed that ac4C RNAs recruit to microirradiated chromatin immediately after genome injury. The ac4c RNA signal weakens 20-40 min post-irradiation (Fig. 1). In non-irradiated cells, we have observed that ac4C RNA occupies a compartment of nucleoli (Fig. 1), which was additionally veri ed by dual immunolabelling showing in parallel ac4 RNA and brillarin, a main component of nucleoli ( Fig. 2A, B). The brillarin-positive region of nucleoli of non-irradiated cells was dense on ac4C RNA, and inhibitor of RNA polymerase I, actinomycin D, caused the so-called crescent-like morphology of the brillarin-positive region of nucleoli. Also, these crescents colocalized with ac4C RNAs (Fig. 2A). Ac4C RNAs recognized UVA-induced DNA lesions also when the cells were treated by actinomycin D (Fig. 2B).
When the cells were as the whole cell population irradiated by UVC lamp, we found a high density of ac4C RNA in the nucleoplasm of the cells exposed to UVC and analyzed 20 minutes post-irradiation. Cells analyzed 90-120 minutes after irradiation were characterized by ac4C RNA reorganization into wellvisible, tiny foci ( Fig. 3A-E). In such UVC-irradiated cells, we quanti ed ac4C RNA distribution in the whole nuclear content (Fig. 3B). As can be seen in Figures 3A and 3E, in non-irradiated cells, the most signi cant amount of ac4C RNA was concentrated in nucleoli; on the other hand, after UVC irradiation, the highest uorescence intensity we observed in the nucleoplasm (3A-E). The ratio of the relative uorescence intensities (FI) in the nucleus to the rest of FI of ac4C RNA in the nucleus was 2.42 (median) in the control cells. For irradiated samples, at intervals 5 min, 20 min, 50 min, and 120 min after irradiation, the FI range was from 0.44 to 0.94 (on average). The difference in relative intensities was highly statistically signi cant (p≤0.0001) (Fig. 3E). The most marked changes in the nucleoplasm were detected 5 minutes after irradiation, then the amount of ac4C RNA decreased. Differences in absolute uorescence intensity within the nucleus were statistically signi cant in all compared groups (p≤0.0001); (Fig. 3B, 3C). Interestingly, the total FI of uorescently stained ac4C RNA in the nucleoli was signi cant after irradiation, and an increase in ac4C RNA inside the nucleoli was also 120 min after irradiation, but the trend in the nucleoplasm was the opposite (Fig. 3D, 3E). These results suggest that UV radiation increases the acetylation of N4-cytidine in nucleoplasmic RNAs.
Accumulation of ac4C RNA in DNA lesions is PARP dependent We have addressed the question if a pronounced appearance of ac4C at DNA lesions is PARP-dependent.
We treated the cells with a PARP inhibitor, olaparib. In non-irradiated cells, we have observed a high density of ac4 RNA in nucleoli, 10 min after UVA irradiation, a high level of ac4 RNAs was detected in the nucleoplasm, while the cells treated by PARP inhibitor were characterized by the identical distribution pro le fo ac4C RNA as in control non-irradiated cells (Fig. 4A). PARP inhibitor olaparib also prevents the recruitment of ac4C RNA to UVA-microirradiated chromatin; thus, these data show that accumulation of ac4C RNAs to locally-induced DNA is PARP-dependent (Fig. 4B).

Recruitment of ac4C at UVA-damaged chromatin is not dependent on the function of NAT10 acetyltransferase
It is well-known that RNA cytidine acetyltransferase NAT10 is responsible for the installation of N4acetylcytidine (ac4C) on mRNAs, 18S rRNA, and tRNAs [13,14]. It was observed that NAT10 is responsible for the formation of ac4C at position 1842 in 18S rRNA [13]. Based on this information, we analyzed if recruitment of ac4C RNA to DNA lesions is NAT10 depend. In this case, we found that NAT10 is not recruited to microirradiated chromatin; thus, it seems likely that ac4C installation on RNA at UV-induced DNA lesions is not mediated via NAT10 acetyltransferase (Fig. 5).
Here, we also compare ac4C RNA accumulation at microirradiated chromatin with m 6 A RNA appearance at DNA lesions, as we studied recently [17]. We found distinct recruitment kinetics for ac4C RNAs and m 6 A RNAs at DNA lesions. In general, a peak of accumulation is identical for both ac4C RNAs and m 6 A RNAs, but a high level of ac4C RNAs at microirradiated chromatin remains longer, up to 30-45 min postirradiation, in comparison to m 6 A RNAs (Fig. 6).
An increase in ac4C RNA level after UV irradiation is identical in all phases of the interphase.
For such experiments, we used HeLa-Fucci cells expressing RFP-Cdt1 in the G1 phase and GFP-tagged geminin in the G2 phase of the cell cycle [20]. In G1, S, and G2 cell cycle phases, we observed an identical increase in ac4C RNA level when the whole cell population was irradiated by UV light and compared with non-irradiated cells (Fig. 7A, B).

Discussion
Acetylation processes and their regulation via speci c epigenetic writes (acetyltransferases) and erasers (deacetylases) are well-described on histones. Another component of chromatin, which is DNA, is absent of acetylation marks, but in distinct types of RNAs, acetylation of N4-cytidine (ac4C) was described as a unique regulatory factor that originates from epitranscriptome [25]. From the view of nucleic acid biology and chromatin features, ac4C on RNAs is a unique biochemical event. It is known that RNA can be acetylated on cytidine via a speci c RNA acetyltransferase, called NAT10, and ac4C on RNA can be erased via the function of sirtuin 7 [15]. Thus, these important factors of RNA biology we studied in this paper showing NAT10-independent recruitment of ac4C RNA to UV-induced DNA lesions.
It is well-known that RNA can also be methylated on adenosine (m 6 A), which regulates many biological processes, including transcription, translation, RNA stability, and also DNA repair [16,17].  -30]. In our case, we have observed that these hybrid loops are rather involved in a base excision repair (BER) mechanism because the protein of BER, XRCC1, behaves identically at DNA lesions as another RNA modi cation 8-adenosine methylation (m 8 A) that also recognized microirradiated chromatin [31]. Notably, both XRCC1 and m 8 A RNAs were not detected at microirradiated genomic regions when the cells were treated by PARP inhibitor [31]. Based on these results, we have continued with the analysis of RNA modi cations at DNA lesions. We addressed the question of also ac4C RNAs recognize locally micro-irradiated chromatin. Indeed, a high level of ac4C RNA we observed at irradiated genomic regions by UVA laser (Fig. 1). Also, UVC-light causes elevation of ac4C RNA level in the whole cell nuclei (Fig. 3A). Importantly, similarly to m 6 A RNAs and m 8 A RNAs, ac4C RNAs were accumulated to microirradiated genomic regions, and this process was PARP dependent [18, 31] (Fig. 4A, B). A very interesting is observation that the kinetics of ac4C RNAs at the lesions is different from the kinetics of m 6 A RNAs at microirradiated chromatin ( Fig. 6 and [17]). Also, all studied RNA modi cations were recruited to DNA lesions independently of the cell cycle, as we analyzed in HeLa Fucci cellular system ( [31]; Fig. 7 Declarations Figure 1 Recruitment of ac4C RNAs at UVA-damaged chromatin. Local laser microirradiation showed that ac4C RNAs recognized UVA-microirradiated chromatin immediately after local laser irradiation. In later stages of DDR, 25-40 min post-irradiation, the ac4C RNA signal at DNA lesions disappeared. Scale bars showed 5 µm.   PARP-dependent recruitment of ac4C to UVA-damaged chromatin. Local laser microirradiation showed that ac4C RNAs did not accumulate to DNA lesions when the cells were treated by PARP inhibitor. It was observed in both (A) UVA irradiated MEFs as a whole-cell population and (B) microirradiated MEFs by UVA laser. Scale bars showed 5 µm. Figure 5 NAT10-independent recruitment of ac4C RNA to UVA-damaged chromatin. Local laser microirradiation showed that NAT10 acetyltransferase (red) does not recruit to DNA lesions (positive on ac4C RNA; green) induced in MCF7 cells. These cells were studied instead of MEFs due to antibody availability. Scale bars showed 5 µm.