Poly(ADP-ribose) polymerase-1 (PARP-1) is required in murine cell lines for base excision repair of oxidative DNA damage in absence of DNA polymerase

Oxidative DNA base damage is mainly corrected by the base excision repair (BER) pathway which can be divided into two subpathways depending on the length of the resynthetized patch, either one nucleotide for short patch (SP-BER) or several nucleotides for long patch (LP-BER). The role of proteins in the course of BER processes has been investigated in vitro using purified enzymes and cell free extracts. In this study, we have investigated the repair of 8-oxo-7,8-dihydroguanine (8-oxoG) in vivo using wild-type, Pol b -/- , PARP-1 -/- , and Pol b -/- PARP-1 -/- 3T3 cell lines. We used non replicating plasmids containing a 8-oxoG.C base pair to study the repair of the lesion located in a transcribed sequence (TS) or in a non transcribed sequence (NTS). The results show that 8-oxoG repair in TS is not significantly impaired in cells deficient in Pol b or PARP-1 or both. Whereas 8-oxoG repair in NTS is normal in Pol b -null cells, it is delayed in PARP-1-null cells and greatly impaired in cells deficient in both Pol b and PARP-1. The removal of 8-oxoG and presumably the cleavage at the resulting AP site are not affected in the PARP-1 -/- Pol b -/- cell lines. However, 8-oxoG repair is incomplete yielding plasmid molecules with a nick at the site of the lesion. Therefore, PARP-1 -/- Pol b -/- cell lines cannot perform 5’-dRP removal and/or DNA repair synthesis. Furthermore, the poly(ADP-ribosyl)ation activity of PARP-1 is essential for 8-oxoG repair in a Pol b -/- context, since expression of the catalytically inactive PARP-1 (E988K) mutant does not restore 8-oxoG repair whereas an WT PARP-1 does.


Introduction
Reactive oxygen species, generated either endogenously by cellular metabolism or by exposure to environmental oxidants, induce DNA damages which have been implicated in human pathologies such as cancer, neurodegenerative diseases or ageing (1)(2)(3)(4). An oxidatively damaged guanine 8-oxo-7,8-dihydroguanine (8-oxoG), is an important mutagenic DNA lesion due to its potential to mispair with adenine, thus generating G.C to T.A transversions. The biological significance of 8-oxoG is revealed by the spontaneous mutator phenotype of bacterial and yeast mutants impaired in 8-oxoG repair (5)(6)(7)(8)(9). In all organisms, 8-oxoG is primarily repaired by the base excision repair (BER) pathway which is the major process for the elimination of oxidative base damage, alkylation base damage and apurinic / apyrimidinic (AP) sites (10,11). In mammalian cells, BER of 8-oxoG is initiated by the action of the Ogg1 DNA N-glycosylase, which catalyzes the hydrolysis of the N-glycosyl bond linking damaged bases to the sugarphosphate backbone generating AP sites. Ogg1 is also endowed with an AP lyase activity that can incise the phosphodiester bond immediately 3' of an AP site yielding a 3'-terminal sugar phosphate (3'-dRP) (12)(13)(14)(15)(16)(17)(18)(19). However, in the presence of the major AP endonuclease Ape1, AP sites resulting from the removal of 8-oxoG residues by Ogg1 are primarily processed by Ape1 which catalyzes the hydrolytic cleavage of the phosphodiester bond immediately 5' to the AP site generating a 5'-terminal sugar phosphate (5'-dRP) (20,21). Afterwards, the 5'-dRP is removed by a dRPase activity associated with DNA polymerase β (Polβ) which simultaneously adds one nucleotide. The nick is finally sealed by DNA ligase III associated with the X-ray cross-complementing factor 1 (Xrcc1) (22)(23)(24). The entire process results in the removal of 8-oxoG and its replacement with a guanine and constitutes the Short-Patch Base Excision Repair 4 displaces the 5'-dRP residue generating a 5'-flap structure with a 5'-dRP end. The 5'-flap is removed by the Flap endonuclease 1 (Fen1) and a DNA ligase seals the nick. The role of different DNA polymerases in LP-BER in the wild type cellular context is unclear. Recently, DNA polymerase δor ε-dependent LP-BER have been reconstituted with purified human proteins (32). For efficient repair of a regular AP site, in addition to Ape 1 and DNA polymerase δ, the reaction assay required replication factor C (RF-C), proliferating cell nuclear antigen (PCNA), Fen1 and Ligase I (30). LP-BER is a minor pathway for the repair of 8-oxoG and regular AP sites in wild type cell free extracts (26,33). In contrast, LP-BER is thought to be the major pathway for the repair of reduced or oxidized AP sites (31).
PARP-1 binds with high affinity DNA containing single-strand breaks. Upon binding to DNA strand breaks, PARP-1 catalyses the synthesis of poly(ADP-ribose) from NAD + and covalently modifies several nuclear proteins involved in chromatin architecture (such as histones and lamin B) and in DNA metabolism (such as topoisomerases, DNA polymerases and BER factors). The automodification of PARP-1 induces its dissociation from DNA breaks and inhibition of its catalytic activity. PARP-1 and poly(ADP-ribosyl)ation are proposed to be critical for cellular processes such as DNA repair, transcription or energy depletion-induced cell death during inflammatory injury (see for review (50 ,51)). Evidence for the involvement of PARP-1 in BER was provided by the fact that PARP-1-null mice are hypersenstive to ionizing radiation and alkylating agents (52)(53)(54). Moreover, the physical interaction of PARP-1 with proteins such as by guest on March 24, 2020 http://www.jbc.org/ Downloaded from Polβ and Xrcc1 also points to its role in BER (55,56). Recently, PARP-1 was shown to bind with high affinity to BER intermediates harbouring a 5'-dRP (57). Furthermore, reconstitution of BER using purified proteins shows that PARP-1 stimulates two of the key steps of LP-BER : strand displacement synthesis by Polβ and 5'-flap cleavage by Fen1 (58). In addition repair of AP sites is impaired in cell free extracts of PARP-1-null mice cell lines (46). This study also shows that PARP-1-null Polβ-null cell free extracts present a dramatic decrease in LP-BER when compared to PARP-1-null cells. Therefore, results with purified proteins and cell free extracts point to a role of PARP-1 in LP-BER.
To investigate the role of Polβ and PARP-1 in the course of BER in the cellular context,
Transfectants were selected by growth in medium containing hygromycin at increasing concentration up to 400µg/ml. Single clones were isolated after 15 days, propagated in 12-well plates and analyzed for PARP-1 expression by western blotting.
Western blotting : Cell-free protein extracts were prepared from clonal isolates as previously described (59 (63). Assays for removal kinetics of 8-oxoG were carried out using a procedure previously described (59). Briefly, recovered extrachromosomal DNA was treated or not with 5ng of E.coli Fpg protein (64)

Combination of Pol and PARP-1 deficiencies abrogates 8-oxoG.C repair on NTS but not in TS.
To investigate the role of Polβ and PARP-1 proteins in the repair of 8-oxoG in the cellular context, we used 3T3 cell lines, wild type (WT), PARP-1 -/-, Polβ -/and PARP-1 -/-Polβ -/which have been previously characterized (46). The different 3T3 cells were transfected with non replicative plasmids that contain a single 8-oxoG.C base pair. Two constructs were used to allow repair analysis of 8-oxoG located in the same sequence context but with a different transcriptional status (59,66).  To further investigate of the role of Polβ and PARP-1, we analyzed the processing of 8-oxoG.C in a double deficient PARP-1 -/-Polβ -/cell lines. Figure 1D shows that 8-oxoG.C base pair is efficiently repaired when located on a TS plasmid in both WT and PARP-1 -/-Polβ -/cell lines. In contrast, until eighteen hours after transfection, we do not observe detectable repair of 8-oxoG.C in NTS plasmid, whereas all molecules were repaired in WT cells. Finally, less than 10% of repair may be observed in the doubel knock-out cells 24 h after transfection (Fig. 1D).
These results show that full repair of 8-oxoG.C is greatly impaired in NTS plasmid in cells lacking both Polβ and PARP-1 proteins.

Removal of 8-oxoG in NTS is not affected in Pol -/-PARP-1 -/double knock-out cells.
The absence of repair of 8-oxoG.C in the NTS plasmid 12 hours after transfection in the PARP-1-null Polβ-null cells could be due to an impaired recognition and/or excision of 8-oxoG by the Ogg1 DNA N-glycosylase. Therefore, Ogg1 enzyme activity in crude extracts was assayed using as substrate a 34mer oligonucleotide containing a 8-oxoG.C base pair. Figure 2 shows that a cell free extract of the PARP-1-null Polβ-null 3T3 cells efficiently cleaves the [8-  (9,67). Therefore, after amplification, the 8-oxoG-containing plasmids generate a population that contains one or two NgoMIV restriction sites. In contrast, plasmid without 8-oxoG generates a pure population containing a single NgoMIV restriction site as previously described (68). Our results show that only 1 out of 96 clones tested contained the 8-oxoG.C pair (data not shown). Taken together, these data strongly suggest that the removal of the 8-oxoG lesion in NTS is not affected by the deletion of Polβ and PARP-1 in the cell.

Incomplete repair of 8-oxoG in double Pol -/-PARP-1 -/cells in NTS results from a defect in 5'-dRPase and/or DNA repair synthesis activities.
Repair assays with purified proteins or cell free extracts show that AP sites generated by  Figure 5B shows a strong arrest of polymerization with DNA recovered from PARP-1 -/-Polβ -/cells. In contrast, no polymerization arrest is observed using DNA recovered from WT or Polβ -/- (Fig. 5B). Location of the arrest band was determined by comparing primer extension and plasmid sequencing using the same primer. Sequence analysis indicates that the arrest band corresponds to an incorporation opposite to the original position of 8-oxoG in the plasmid DNA (Fig. 5B). This observation strongly suggest a cleavage of DNA by Ape1 at the site of the lesion resulting in the formation of a 5'dRP residue (Fig. 5A).

Discussion
The identification of proteins involved in the processes of base excision repair (BER) in mammalian cells is subject to intense investigation. Most studies are carried out using cell extracts or purified proteins (22-24,26,30,46,70,71 ,72 ). The role of " nonessential factors " such as PARP-1 is also subject of discussion, (for reviews, see (49,73,74)). In this work, we used an in vivo approach based on the transfection of monomodified plasmid in intact cells to study the repair of 8-oxoG in the cell context. In the last decade, shuttle plasmids have been used to analyze mutagenesis and repair in mammalian cell lines and mutagenic potency of specific lesions (75,76). In every cases, results gave a quite good preview of the process in genomic DNA. This system also allowed to study the repair of a 8-oxoG in either transcribed (TS) or nontranscribed (NTS) condition (59,77).
In this study, we  (66,78). In mouse cells, this pathway is dependent of Csb but independent of Ogg1, Polβ and PARP-1 ((59,77) and our study). In contrast, 8-oxoG repair in NTS requires BER proteins such as Ogg1 and the combination of Polβ and PARP-1 but not NER proteins (our study and (59,66,77)). PARP-1 has been reported to be a negative or positive regulator of transcription, by modifying and/or binding several transcription factors see for review (51). The absence of PARP-1 has no effect on TS, indicating that the function of PARP-1 in transcription is not related to DNA repair but is more likely to participate in the organization of the chromatin architecture (79).
Our study also show that Polβ, primarily involved in SP-BER, is not essential in vivo for 8-oxoG repair in NTS, in agreement with studies using cell free extracts (26,80).