Effects of 5′,8′-Cyclo-2′-Deoxypurines on the Base Excision Repair of Clustered DNA Lesions in Nuclear Extracts of the XPC Cell Line

Clustered DNA lesions (CDL) containing 5′,8-cyclo-2′-deoxypurines (cdPus) are an example of extensive abnormalities occurring in the DNA helix and may impede cellular repair processes. The changes in the efficiency of nuclear base excision repair (BER) were investigated using (a) two cell lines, one of the normal skin fibroblasts as a reference (BJ) and the second from Xeroderma pigmentosum patients’ skin (XPC), and (b) synthetic oligonucleotides with single- and double-stranded CDL (containing 5′,8-cyclo-2′-deoxyadenosine (cdA) and the abasic (AP) site at various distances between lesions). The nuclear BER has been observed and the effect of both cdA isomers (5′R and 5′S) presence in the DNA was tested. CdPus affected the repair of the second lesion within the CDL. The BER system more efficiently processed damage in the vicinity of the ScdA isomer and changes located in the 3′-end direction for dsCDL and in the 5′-end direction for ssCDL. The presented study is the very first investigation of the repair processes of the CDL containing cdPu considering cells derived from a Xeroderma pigmentosum patient.


Introduction
Eukaryotic cells have developed a system that ensures genome integrity, called the DNA damage response (DDR) pathway. Given that each human cell can develop approximately 10 2 -10 5 lesions per day, multiple repair pathways are beneficial, e.g., mismatch repair (MMR), base excision repair (BER), nucleotide excision repair (NER), non-homologous repair end joining (NHEJ), or homologous recombination (HR) [1][2][3]. During DDR, cell cycle progression is blocked, and DNA damage sensors activate DNA repair mechanisms specific to the type of detected lesion. Single-stranded DNA breaks or oxidized nucleobases are mainly repaired by BER, large DNA adducts and cross-links are mainly repaired by NER, and minor nucleotide damage such as alkylation is repaired by MMR [4,5].
The harmful factors damaging cellular components include ionizing radiation. The lethal dose of 4 Gy of ionizing radiation can generate about 5000 single-strand breaks (SSBs), 160 DNA double-strand breaks (DSBs), 600 cross-links, or~5000 DNA base damage [6]. Additionally, a low dose of 1 Gy of ionizing radiation can generate the above DNA damage in the form of isolated lesions as well as clustered DNA lesions (CDL) [7]. A single event of DNA irradiation may result in an additional inter-or intra-nucleotide linkage. This leads to the formation of bulky 5 ,8-cyclo-2 -deoxypurines (cdPus) structures in DNA, where the base and sugar moiety of the purine nucleoside in DNA are simultaneously damaged. Therefore, cdPus have been classified as tandem changes which include two adjacent DNA changes located on the same strand [8][9][10]. The production yields of radiation-induced purine 5 ,8-cyclo-2 -deoxyribonucleosides in human monocyte DNA were determined by HPLC-MS/MS of 20 and 4 changes per 10 9 normal nucleosides per Gy for 5 ,8-cyclo-2deoxyguanosine (cdG) and 5 ,8-cyclo-2 -deoxyadenosine (cdA), respectively [11]. Other It has been recently reported that the successful repair of minor lesions (e.g., AP site) within a CDL containing cdPus depends on the distance and the relative position between lesions [21]. Moreover, differences in BER activity were observed between 5 S and 5 R diastereomers and between purines (cdA and cdG). Repair of the dsCDL was tested in the nuclear extracts (NE) of xrs5 cells, as a well-established model for such experiments [1,10,[37][38][39][40].
The presented study analyzed whether the presence of tandem lesions (RcdA or ScdA) and their relative distance from a single lesion (AP site) on the opposite or the same DNA strand affect the repairability of the latter. The main steps of the repair pathway (strand incision, elongation, and rejoining) were tested in NE of XPC cell line (human fibroblasts from XP patient, complementation group C) that carries the NER defect. As a reference system, the BJ cell line (human skin fibroblasts) was used, which represents normal cells with properly functioning DNA repair systems.
The presence of cdPus in DNA can lead to inhibition of gene expression or transcriptional mutagenesis. It is important to study the cdPus' function to better understand its biological significance, its effects on DNA structure, and the key aspects of cell survival, such as the repair of damaged genetic material. It is also important to test whether the obtained observations are confirmed regardless of the cellular model involved in the research.

The Substrate Oligonucleotides
The oligonucleotides used as a substrate in the presented study were prepared on a Geneworld synthesizer (K&A Laborgeraete GbR, Schaafheim, Germany) in the Bioorganic Chemistry Department of the Polish Academy of Science (Lodz, Poland). Nucleotide phosphoroamidites needed for synthesis were purchased from the ChemGenes Corporation (Wilmington, MA, USA) while the phosphoroamidite cdPus' derivatives were synthesized according to the previously described protocol [41]. The crude oligos were HPLC-purified (C-18 column, Phenomenex, Synergi 4 µm Fusion-RP 80 Å, 250 × 4.6 mm) and the concentration was measured (Varian Cary 1.3E spectrophotometer (Varian, Brunn am Gebirge, Austria)). Mass spectrometry analysis was performed in the full-scan negative-ion mode (mass range of 50-2000 m/z, Waters Synapt G2-Si HDMS, quadrupole time of flight hybrid mass spectrometer (Waters, Manchester, UK)). The results are presented in the Supplementary Materials, i.e., the found and calculated masses are reported in Table S1 and the mass spectra in Figure S2. The substrate oligos were stable in the conditions of the experiment as shown in previous studies: melting temperatures were above 70 • C [42] and matrix strands (native and containing cdPu) were not susceptible to the action of repair enzymes present in NE [2,10,40]. Table 1 shows the list of ds-oligonucleotides used in the experiments and their sequences. Each oligo contained 2 -deoxyuridine (dU) residue which has a specific number assigned. The number characterizes the number of base pairs between cdA and dU. For double-stranded lesions ([ds]), when dU was located in the 3 direction in relation to cdA located on the opposite strand, it is described with a negative number; when dU was located in the 5 direction in relation to cdA located on the opposite strand, it is described with a positive number. For single-stranded lesions ([ss]), when dU was located in the 5 direction in relation to cdA located on the same strand, it is described with a negative number; when dU was located in the 3 direction in relation to cdA located on the same strand, it is described with a positive number.

Hybridization of ss-Oligonucleotides
The hybridization of radiolabeled ss-oligos was performed in pure H 2 O with a 1.5-fold excess of the suitable complementary strand (purified and non-radiolabeled, Table 1) at 90 • C for 10 min and was followed by slow cooling to room temperature. The ds-oligos were purified by precipitation (see Section 2.8). The hybridization was verified, and the purity was checked on the 15% native polyacrylamide gel ( Figure S1A).

Preparation of ds-Oligonucleotides with AP Sites
The radiolabeled, dry ds-oligos with dU in their sequence were incubated with uracil-DNA glycosylase (UDG) (5U, New England BioLabs Ipswich, MA, USA), which recognizes and removes dU from DNA. Reactions were performed in 20 µL of the reaction buffer (20 mM Tris-HCl, 1 mM EDTA, 1 mM DTT, pH 8.0 at 25 • C) at 37 • C for 30 min. The ds-oligos containing AP sites were purified by precipitation (see Section 2.8). Additionally, the AP sites formation was verified. The ds-oligos with AP sites were incubated with human apurinic/apyrimidinic endonuclease (APE1) (5U, New England BioLabs Ipswich, MA, USA) in 10 µL of the reaction buffer (50 mM potassium acetate, 20 mM Tris-acetate, 10 mM magnesium acetate, 1 mM DTT, pH 7.9 at 25 • C) at 37 • C for 30 min to generate SSBs visible as shorter DNA fragments with a length of 15-25 nucleotides. The purity and formation of SSBs were checked on 15% denaturing polyacrylamide gel ( Figure S1B).

Cell Cultures
The study was performed using two cell lines-BJ as a reference cell line and XPC as a tested cell line. The BJ, human fibroblasts cell line (CRL-2522, ATCC, Vancouver, CA, USA) was cultured in MEM with Earle's salts and non-essential amino acids supplemented (Gibco) with 10% FBS (Biowest, MO, USA). Originally cultured from neonatal foreskin tissue, the BJ cell line is a recognized model of replicative senescence [43]. It results from the Hayflick limit or exhaustion of the division limit and affects most types of somatic cells in vitro and in vivo. It is caused by the shortening of the telomeres. As the primary cell line, it was adopted in this study as the reference line with properly functioning cellular DNA repair processes.
Cells of the XP patient belonging to the XPC group (GM17420 skin fibroblasts) were purchased from the Coriell Institute for Medical Research (Camden, NJ, USA). Clinical, cellular, and molecular data for GM17420 is available upon request from the Coriell Institute. XPC fibroblasts were cultured in MEM with Earle's salts and non-essential amino acids (Gibco) and supplemented with 15% FBS (Biowest, MO, USA). Compared to normal fibroblasts, XP cells are hypersensitive to UV radiation and/or chemical factors, which results in a significantly increased frequency of chromosome breaks [23]. Moreover, the damage lasts longer, which indicates impaired repair processes. Measurements of the efficiency of cell repair systems showed that XPC cells had 10-20% of the repair synthesis shown by normal cells [44].

Preparation of Nuclear Extracts
Cells from each cell line were harvested in the exponential phase. After pelleting, they were treated with the NE-PER™ Nuclear and Cytoplasmic Extraction Reagents kit (ThermoFisher Scientific, Waltham, MA, USA) according to the protocol provided by the manufacturer. The concentration of nuclear extracts (NE) was determined using the Pierce™ 660 nm Protein Assay colorimetric assay (ThermoFisher Scientific, Waltham, MA, USA). Concentrations were found as follows: BJ-between 5.4-8.5 mg/mL; XPC-between 4.2-6.4 mg/mL. Aliquots of cellular extracts were stored at −80 • C for a maximum period of 6 months.
Each set of reactions was repeated three times for its consistency and reliability. Data were quantified using Quantity One (Bio-Rad, Hercules, CA, USA) where the AP site repair in relation to reaction time was determined. The intensity of the bands was analyzed and disclosed as a % value of the overall intensity of bands within individual lanes on the autoradiogram. The slight differences in the repair activity between batches of NE were taken into account; hence, all results were compared with the control containing AP site as single damage.

General Procedures
Purification of oligonucleotides: 250 µL of ice-cold ethanol was added to samples which were incubated at −80 • C for 30 min and subsequently centrifuged at 4 • C for 30 min (13,000 rpm). Afterward, the samples were dried under reduced pressure at room temperature and dissolved in pure H 2 O.
Native PAGE: reactions were stopped by placing the samples on ice and adding the native loading buffer (40% sucrose, 0.025% bromophenol blue, and 0.025% xylene cyanole). Samples were subjected to electrophoresis on a 15% native polyacrylamide gel in 1× TBE (89 mM Tris-HCl, 89 mM boric acid, 2 mM EDTA).
The PAGE electrophoresis was performed at a constant power of 45 W for 120 min. The obtained results were visualized by autoradiography.

Results
The study examined whether the diastereomeric form of cdA and the distance between lesions located within single-or double-stranded clustered DNA damage (ssCDL or dsCDL) had an impact on the repair processes. Moreover, this work tried to answer the oftenraised question about the influence of the clustered DNA lesions on the DNA repair in Xeroderma Pigmentosum patients. The two cell lines were used-BJ (human skin fibroblasts) and XPC (skin fibroblasts from Xeroderma Pigmentosum patient, complementation group C, see Section 2.5). The experimental model was, as previously described, synthetic double-stranded oligonucleotides containing dU (AP site precursor) and 5 ,8-cyclo-2deoxyadenosine (cdA) (5 S and 5 R isomer) on the same (ssCDL) or the opposite strand (dsCDL) ( Table 1) [10,40,42]. Each step of the substrate ds-oligonucleotide preparation was verified, and the results are presented in Figure S1. As shown previously, both diastereomers of cdA are stable in experimental conditions, and during reactions with NE and chosen glycosylases [1,2,10,40,42].
NE obtained from BJ and XPC cell lines were used to determine the repair efficiency of AP sites located within ssCDL and dsCDL. The goal was to examine the main stages of BER (AP site incision, strand elongation, and strand rejoining) taking place in the nucleus of XPC cells (compared to BJ as a reference cell line).
Control 1 (containing a single AP site, Table 1) was used to confirm the activity of proteins in the NE (Figure 1). In the case of the reference cell line (BJ), the endonucleolytic activity (strand incision) was observed after 1 min, and polymerase activity was detected after 5 min (Table S5). AP site rejoining (strand reconstitution) was observed after 30 min, increasing with time up to 82.98 ± 7.68% after 120 min (Table S4). The XPC cells showed endonucleolytic activity after 1 min, polymerase activity after 15 min, and strand rejoining after 30 min, reaching 32.96 ± 7.65% after 120 min (Table S2).
The presented study analyzed whether the presence of tandem lesions (RcdA or ScdA) and their relative distance of 1-5 nucleobases in 3 -end and 5 -end direction from a single lesion (AP site) on the opposite or the same strand affected the cell's ability to repair the latter. The repair was tested in NE obtained from XPC and BJ cells. Moreover, it was investigated whether the XPC defect will affect the repair processes carried out by BER and if it would be an additional factor (next to cdPus) limiting the repair of small lesions such as AP sites. The presented study analyzed whether the presence of tandem lesions (RcdA or ScdA) and their relative distance of 1-5 nucleobases in 3′-end and 5′-end direction from a single lesion (AP site) on the opposite or the same strand affected the cell's ability to repair the latter. The repair was tested in NE obtained from XPC and BJ cells. Moreover, it was investigated whether the XPC defect will affect the repair processes carried out by BER and if it would be an additional factor (next to cdPus) limiting the repair of small lesions such as AP sites.

The Influence of Single-Stranded CDL on the Nuclear BER Pathway in XPC Cells
The impact of the cdA appearing in the genome on the DNA repair pathways is widely studied [1,2,8,10,40,[45][46][47]. Single-stranded CDL (ssCDL) containing ScdA or RcdA have been shown to impair the activity of specific proteins involved in the BER process, e.g., glycosylases and polymerases [10,[37][38][39]42,48]. This study explored the repair via BER (strand incision, elongation, and rejoining) in the NER-deficient cell line derived from the XP patient.

Strand Incision
The incision of the AP sites resulting from the endonucleolytic activity of NE was observed for all tested substrates. SSBs were observed as bands corresponding to 18-mer for dU−5 and dU−5/+5, and to 28-mer for dU+5 (Figures 2, S3A-S6A).

The Influence of Single-Stranded CDL on the Nuclear BER Pathway in XPC Cells
The impact of the cdA appearing in the genome on the DNA repair pathways is widely studied [1,2,8,10,40,[45][46][47]. Single-stranded CDL (ssCDL) containing ScdA or RcdA have been shown to impair the activity of specific proteins involved in the BER process, e.g., glycosylases and polymerases [10,[37][38][39]42,48]. This study explored the repair via BER (strand incision, elongation, and rejoining) in the NER-deficient cell line derived from the XP patient.

Strand Incision
The incision of the AP sites resulting from the endonucleolytic activity of NE was observed for all tested substrates. SSBs were observed as bands corresponding to 18-mer for dU−5 and dU−5/+5, and to 28-mer for dU+5 (Figures 2 and S3A-S6A).

Strand Elongation
The activity of polymerases was also taken under consideration for both cell lines. Incorporation of up to two nucleotide units (NUs) was observed for XPC suggesting that the LP-BER pathway was activated (Figures 2, S3 and S4, parts A). The same observation was made for BJ, where two NUs were incorporated ( Figures S8 and S9).

Strand Elongation
The activity of polymerases was also taken under consideration for both cell lines. Incorporation of up to two nucleotide units (NUs) was observed for XPC suggesting that the LP-BER pathway was activated (Figures 2, S3 and S4, parts A). The same observation was made for BJ, where two NUs were incorporated ( Figures S8 and S9).

The Influence of Double-Stranded CDL on the Nuclear BER Pathway in XPC Cells
As shown previously for xrs5 cell line, dsCDL containing ScdA or RcdA impact the main stages of the BER pathway, i.e., strand incision, elongation, and reconstitution [40]. In this study, those stages of repair via base excision were further examined in the context of nuclear extracts of the NER-deficient XPC cell line.

Strand Elongation
The efficiency of strand elongation by polymerases was taken under consideration for dsCDL. The activity of SP-BER has been detected (incorporation of 1NU) for most examined substrates (both cell lines). No polymerase activity (no SSB+1 bands) was observed for lesions positioned opposite (position dU0) and +1 to cdA (dU+1) for all tested variants.
The overall trend for the DNA synthesis by polymerases increased in the following order (data compared for 30 min of reaction time): +4 < −1 < −4 (for ScdA and RcdA; XPC, and BJ).

Discussion
The presented study investigated how (5 S) and (5 R) 5 ,8-cyclo-2 -deoxyadenosine affect the BER repair of ssCDL and dsCDL in NE of XPC cell line. XPC cells obtained from a patient with Xeroderma pigmentosum are NER-deficient. Given that cdPus are not a substrate for BER, this study focused on the second lesion (AP site) present in ssCDL or dsCDL and on the repair efficiency demonstrated by the total activity of nuclear BER proteins.
BER repair enzymes that excise isolated lesions are inhibited when the damage is a part of double-or single-stranded clusters. Previously presented observations confirmed that shifting the distribution of lesions by one nucleotide may change the course of repair or stop it [10,40]. An important observation was the fact that the first stages of the BER system worked more efficiently on oxidative damage (AP site) if it was located within ssCDL or dsCDL systems with the ScdA isomer. This is interesting because, with the rather limited efficiency of the NER system in removing cdPus, excision of 5 R isomer was more efficient than 5 S [12,13,49].
Moreover, as expected, the reference BJ cell line showed the same trends and achieved a higher enzymatic activity of nuclear extracts than XPC. The strand incision and reconstitution were more efficient for BJ, while polymerase activity showed similar or higher levels for BJ than for XPC. Our observations indicate that CDL is also a problem for BJ cells, confirming other results showing that clustered DNA damage is more difficult to repair for the cellular machinery than single lesions [50].
The results showed a similar strand incision trend for XPC and xrs5 cell lines; total endonucleolytic efficiency of NE was higher in the presence of ScdA for both ssCDL and dsCDL [40]. Consistent with the earlier observations obtained for line xrs5, lesions located in the 5 -end direction (positive numbers) of RcdA and ScdA were repaired less efficiently than those located in the 3 -end direction (negative numbers) for dsCDL. Polymerase activity was also observed. It was inhibited by the presence of RcdA even up to 28% more than for ScdA (for both, ssCDL and the dsCDL). It also confirms previous results obtained for xrs5 where the polymerase activity towards RcdA was lower by about 8-20% for most of the tested substrates. Additionally, the polymerase activity was higher in the case of the lesions located in the 3 -end direction for dsCDL ScdA/dU−4 (~44%), where a rate of DNA synthesis of several percent was observed for changes in the 5 -end direction. Again, this supported the observations from xrs5 where DNA synthesis efficiency increased towards the 3 -end [40]. However, results were different for ssCDL, where the polymerase activity was higher for the lesions directed towards 5 -end. The highest activity for XPC extracts was observed in ssCDL for ScdA/dU−5 (~87%), which confirms the observations for BER activity towards ssCDL in extracts obtained from xrs5 [10]. Nevertheless, the small differences in the repair for the dsCDL are very interesting. The extracts from all three discussed lines (BJ, XPC, and xrs5) most efficiently repaired the lesion shifted by 1 NU towards 3 -end for the ScdA and then ScdA/dU−4 (dU−1 > dU−4). The minor differences relate to the repairability of the RcdA isomer. In the case of the BJ reference cell line, the repair trend is the same as for the S isomer (dU−1 > dU−4). However, with a reduced repair efficiency in XPC, the best strand reconstruction was observed for the RcdA/dU−4, the repair of which is blocked in the case of xrs5 extracts.
By comparing the results obtained for XPC and previously for xrs5, the common trends emerged: (a) the BER system processes lesions in the vicinity of the ScdA more efficiently than RcdA which probably results from the fact that 5 R isomer distorts DNA helix to a higher degree than 5 S [51] and (b) lesions located in the 3 -end direction for dsCDL and the 5 -end direction for ssCDL are processed better by BER nuclear proteins, which confirms previous studies ( Table 2).

Conclusions
In summary, the presented results indicate that the BER repair system in XPC nuclear extracts can repair the oxidative damage that constitutes CDL along cdPu. The efficiency and degree of repair depend on the mutual distance of individual lesions within the cluster, diastereomeric form of cdA, and the lesion arrangement on DNA strands (single-or doublestranded clustered DNA damage). This may explain why some lesions are more durable and their repair in the cell is delayed.
Moreover, the presented results confirm what was observed for xrs5 nuclear extracts. This fact additionally validates the use of the xrs5 model for the clustered DNA lesion studies. From the technical point of view, it is beneficial, because working with xrs5 cells is cheaper, more convenient, easier, and faster.
Detailed knowledge on how the distribution of damage within the CDL affects the efficiency of repair processes may have further application in radiotherapy, e.g., in determining the therapeutic dose of radiation. Moreover, it may explain why XP develops various clinical phenotypes which are difficult to predict. The accumulation of cdPus observed in XP patients may be the cause of the development and worsening of XP symptoms and/or other neurological diseases [52,53]. CdA diastereomers are repaired diversely; therefore, it is not surprising that the ScdA and RcdA affect BER in a different way. To understand the potential biological consequences of CDL, it is imperative to carefully recognize what happens to CDL in cells with physiological levels of repair enzymes.