MINIREVIEW Roles of base excision repair subpathways in correcting oxidized abasic sites in DNA Jung-Suk Sung 1 and Bruce Demple 2 1 Department of Life Science, Dongguk University, Seoul, South Korea 2 Department of Genetics and Complex Diseases, Harvard School of Public Health, Boston, MA, USA Genetic stability is threatened by the continuous assault on cellular DNA by various reactive species of both endogenous and exogenous origins. The most common types of DNA damage are associated with DNA base alteration. A well-characterized DNA base modification is uracil, which can arise in genomic DNA by misincorporation of dUMP during DNA syn- thesis, or by the spontaneous deamination of cytosine in G : C base pairs to form a premutagenic lesion [1,2]. Reactive oxygen species, the products of normal cellular respiration, also generate a variety of oxidized DNA base damages, including an 8-oxoguanine that is frequently used as a biomarker for oxidative DNA damage [3,4]. Enzymatic methylation of DNA bases, predominantly cytosines, plays an important role in gene regulation, but nonenzymatic alkylation from endogenous sources forms cytotoxic and mutagenic products, such as 3-alkyladenine and O 6 -alkylguanine [5,6]. Metabolic by-products (such as epoxyaldehydes), produced during cellular lipid peroxidation, are Keywords 2-deoxyribonolactone; DNA polymerase beta; DNA–protein crosslinks; FEN1 protein; long-patch BER; oxidized abasic sites; short-patch BER Correspondence B. Demple, Department of Genetics and Complex Diseases, Harvard School of Public Health, Boston, MA 02115, USA Fax: +1 617 432 0377 Tel: +1 617 432 3462 E-mail: bdemple@hsph.harvard.edu (Received 12 December 2005, accepted 6 February 2006) doi:10.1111/j.1742-4658.2006.05192.x Base excision DNA repair (BER) is fundamentally important in handling diverse lesions produced as a result of the intrinsic instability of DNA or by various endogenous and exogenous reactive species. Defects in the BER process have been associated with cancer susceptibility and neurodegenera- tive disorders. BER funnels diverse base lesions into a common intermedi- ate, apurinic ⁄ apyrimidinic (AP) sites. The repair of AP sites is initiated by the major human AP endonuclease, Ape1, or by AP lyase activities associ- ated with some DNA glycosylases. Subsequent steps follow either of two distinct BER subpathways distinguished by repair DNA synthesis of either a single nucleotide (short-patch BER) or multiple nucleotides (long-patch BER). As the major repair mode for regular AP sites, the short-patch BER pathway removes the incised AP lesion, a 5¢-deoxyribose-5-phosphate moi- ety, and replaces a single nucleotide using DNA polymerase (Polb). How- ever, short-patch BER may have difficulty handling some types of lesions, as shown for the C1¢-oxidized abasic residue, 2-deoxyribonolactone (dL). Recent work indicates that dL is processed efficiently by Ape1, but that short-patch BER is derailed by the formation of stable covalent crosslinks between Ape1-incised dL and Polb. The long-patch BER subpathway effectively removes dL and thereby prevents the formation of DNA–protein crosslinks. In coping with dL, the cellular choice of BER subpathway may either completely repair the lesion, or complicate the repair process by forming a protein–DNA crosslink. Abbreviations AP, apurinic ⁄ apyrimidinic; BER, base excision DNA repair; DPC, DNA–protein crosslink; dL, 2-deoxyribonolactone; 5¢-dLp, 5¢-terminal dL-5- phosphate residues; 5¢-dRp, 5¢-deoxyribose-5-phosphate; MEF, mouse embryonic fibroblasts; 5-MF, 5-methylene-2-furanone; PARP-1, poly(ADP-ribose) polymerase; PCNA, proliferating cell nuclear antigen; Polb, DNA polymerase b. 1620 FEBS Journal 273 (2006) 1620–1629 ª 2006 The Authors Journal compilation ª 2006 FEBS reactive to DNA and give rise to covalently modified etheno-adducts involving all four DNA bases [7]. Although the reported endogenous levels of each type of base lesion vary among tissues and with the method of detection, their mutagenic and cytotoxic potential suggests that they must be considered as factors in the induction of cancer and other diseases. Beyond this endogenous burden of DNA damage, exposure of cells to exogenous reactive chemical agents, derived from environmental sources or delivered deliberately as che- motherapeutic drugs, may directly produce further DNA damage or modulate cellular conditions to increase the level of damage indirectly (e.g. by disrupt- ing mitochondrial function). Perhaps the most important cellular defense mechan- ism that evolved to avert the deleterious effects of the most frequent damaged or inappropriate bases in DNA is base excision DNA repair (BER) [8–10]. The initial step of BER involves enzymatic activities that process the N-glycosylic bonds linking the target bases and their deoxyribose sugars. The first such enzyme discovered was bacterial uracil-DNA glycosylase [11]. Subsequently, uracil-DNA glycosylases were found to be widely distributed, and DNA glycosylases acting on other diverse lesions (alkylated, oxidized, or photo- damaged bases, as well as certain undamaged but mispaired bases) have been found and characterized for their biochemical properties and biological roles in BER : mammalian cells contain at least 10 distinct gly- cosylase activities [12,13]. The initial product of a DNA glycosylase is an abasic [apurinic ⁄ apyrimidinic (AP)] site in DNA, which is the central intermediate during BER. AP sites can also arise spontaneously at a substantial rate and are expected to be one of the most frequent lesions in DNA (Fig. 1A). It has been estimated that AP site formation through the sponta- neous hydrolytic loss of purines generates some 10 000 AP sites per day in a mammalian cell [14,15]. Com- bined with the AP sites produced by DNA glycosylas- es, the daily burden of AP sites is probably much higher. One estimate yielded steady-state levels of 50 000–200 000 AP sites per cell in various rat tissues and human liver [16], although that seems likely to be an overestimate [12]. AP sites are dangerous lesions that block normal DNA replication, with cytotoxic and mutagenic consequences [17]. Oxidative damage to DNA, mediated by free radi- cals and reactive oxygen species, produces structurally distinct abasic sites, known as oxidized abasic sites. Oxidized abasic sites include lesions at DNA strand breaks, such as 3¢-phosphoglycolate esters and abasic residues in an uninterrupted phosphodiester backbone. These types of DNA lesions are formed by the action of various physical and chemical agents, including UV and c-irradiation, heterocyclic N-oxides of the tirapazamine family, organometallic oxidants and the anticancer antibiotics (such as neocarzinostatin) of the ene-diyne family [18–21]. The formation of oxid- ized AP sites is initiated by the reaction of free radicals with the deoxyribose sugar components of DNA and subsequent chemical rearrangements that are modula- ted by the presence of molecular oxygen [22,23]. The earliest identified X-ray damage in DNA was a C1¢- oxidized abasic lesion, 2-deoxyribonolactone (dL) [24], which is generated by initial hydrogen abstraction from the deoxyribose C1¢ carbon, followed by O 2 addi- tion and base loss (Fig. 1B). Successive b- and d-elimi- nations of dL residues yields a strand break with 3¢- and 5¢-phosphate ends and liberates 5-methylene-2- furanone (5-MF) (Fig. 1B). 5-MF has been employed as a characteristic product of dL in its detection in DNA [25,26]. As determined by comparing the release of 5-MF with concomitant DNA breakage, dL lesions may account for up to 72% of the total sugar damage in the irradiated DNA in vitro [25]. Comparison of the rate of spontaneous strand scission at dL sites to the regular (aldehyde) AP sites shows that cleavage at dL sites is 12- to 55-fold faster than at AP sites [27]. How- ever, the immediate breakage of DNA at the dL lesion would not be expected under physiological conditions. OPO 3 O N O OH AP site Spontaneous Base Loss Removal of Bases by DNA Glycosylases OPO 3 OPO 3 OPO 3 OPO 3 A C1´-Oxidation O O 2-Deoxyribonolactone O OPO 3 OPO 3 OPO 3 OPO 3 OPO 3 + O O 5-methylene-2-furanone N 2- 2- + B Fig. 1. Abasic DNA damage. Formation of a regular abasic apurinic ⁄ apyrimidinic (AP) site (A) and an oxidized abasic site, 2-deoxyribono- lactone (dL) (B). J S. Sung and B. Demple BER of oxidized abasic sites FEBS Journal 273 (2006) 1620–1629 ª 2006 The Authors Journal compilation ª 2006 FEBS 1621 The half-life of dL for spontaneous cleavage under simulated physiological conditions was estimated to be 32–54 h in duplex DNA [28]. Recent understandings of the chemical properties of dL indicates that these lesions are probably subjected to cellular DNA repair or translesion DNA synthesis, rather than directly con- tributing to the formation of DNA strand scission. Short- and long-patch BER in mammalian cells A simplified version of BER for AP sites can be des- cribed as follows: (a) enzymatic incision of the AP site; (b) excision of the cleaved AP site at the single-strand break; (c) repair DNA synthesis; (d) ligation of the nick in DNA. In mammalian cells, the major AP endo- nuclease, Ape1 (also called Apex, HAP1, or Ref-1), hydrolyzes the 5¢ phosphodiester bond of the AP site to generate a DNA repair intermediate that contains a single strand break with 3¢-hydroxyl and 5¢-deoxy- ribose-5-phosphate (5¢-dRp) termini [29,30]. Further repair is achieved through at least two distinct BER subpathways that involve different subsets of enzymes, and which result in the replacement of one nucleotide (short-patch BER), or two or more nucleotides (long- patch BER) (Fig. 2). In mammalian short-patch BER, the major 5¢-dRp excision is attributable to DNA polymerase b (Polb). The dRp excision involves a lyase activity in the Polb 8 kDa N-terminal domain acting through a covalent, Schiff base intermediate [31,32] (Fig. 3A). Single- nucleotide gap-filling DNA synthesis is associated with the DNA polymerase activity of Polb, which therefore plays dual roles in short-patch BER. In an earlier study, the simplest form of short-patch BER of uracil was reconstituted in vitro by using purified human pro- teins, including Ung, Ape1, Polb and DNA ligase III [33]. Similar in vitro reconstitution experiments for the repair of other base lesions or the AP site also sugges- ted essential roles of Polb in the short-patch BER pathway [34–36]. Involvement of Polb in the short- patch BER of various types of DNA lesions has been demonstrated by using cell extracts from wild-type and Polb null mouse embryonic fibroblasts (MEF) cells [37–40]. Some short-patch BER is still observed with Polb-deficient cell extracts, however, which suggests 5'-P 3'-blocking group 3'-OH 5'-dRP 3'-OH 5'-P 3'-OH 5'-P 3'-OH 5'-P FEN1-PCNA Polβ and/or Polδ/ε-PCNA Base damage AP site Monofunctional DNA glycosylase Bifunctional DNA glycosylase Ape1 Ape1 / PNK Polβ Polβ LIG1 LIGIII-XRCC1 (1 nt patch) (≥2 nt patch) Fig. 2. Short- and long-patch base excision DNA repair (BER) path- ways. The steps involved in both pathways are discussed in the text. O OPO 3 OPO 3 2- OH H 2 N OH OPO 3 H HN + β-elimination A K72 5'-dRP lyase K72 Polβ OPO 3 2- OPO 3 2- Polβ O OPO 3 O OH OPO 3 O HN B H 2 N K72 K72 OPO 3 2- OPO 3 2- O OPO 3 O H 2 N OH OPO 3 O HN AP Lyase C OPO 3 2- OPO 3 2- Polβ Polβ 5'-dRP lyase Fig. 3. Excision of an abasic apurinic ⁄ apyrimidinic (AP) site and formation of a 2-deoxyribonolactone (dL)-mediated DNA–protein crosslink. (A) Repair of a 5¢ incised AP site, a 5¢-deoxyribose-5-phos- phate residue (5¢-dRp) by the dRp lyase activity of DNA polymerase b (Polb). (B) Covalent trapping of Polb by a 5¢ incised dL residue through the dRp lyase active site of the enzyme. (C) Covalent trap- ping of a glycosylase-AP lyase by an uncleaved dL residue. BER of oxidized abasic sites J S. Sung and B. Demple 1622 FEBS Journal 273 (2006) 1620–1629 ª 2006 The Authors Journal compilation ª 2006 FEBS that there is functional redundancy at the level of DNA polymerases to provide cells with backup sys- tems [41–43]. Despite this possibility, Polb is encoded by an essential gene, the deletion of which causes embryonic lethality in mice [44]. Polb-deficient MEFs exhibit hypersensitivity to DNA alkylating agents that require BER [44]. Somewhat surprisingly, near-normal resistance could be restored in MEFs by providing only the N-terminal dRp lyase domain of Polb [45], which suggests greater functional redundancy for BER repair polymerase activities than for dRp excision. The long-patch BER pathway involves strand dis- placement repair synthesis of at least two nucleotides, with excision of the 5¢ -dRp residue as part of a flap oligonucleotide cleaved by the FEN1 nuclease [34,46]. The identity of the polymerases involved in the long- patch BER pathway is not yet fully understood. It has been suggested that Polb may be responsible for the initiation of strand displacement synthesis [40,47]. In addition, the involvement of other DNA polymerases, such as Pold and Pole, in long-patch BER has been suggested [43,48,49]. A reconstituted enzyme system was developed for long-patch BER of a reduced AP site utilizing purified Ape1, Polb, Pold, proliferating cell nuclear antigen (PCNA), FEN1 and DNA ligase I, where Pold substituted for Polb when PCNA was pre- sent in the reaction [34]. PCNA-dependent long-patch BER was also demonstrated in extracts of Polb-defici- ent MEF cells, but it appeared to be heavily dependent on the use of circular DNA substrates [38,41]. During the PCNA-independent long-patch BER mode, Polb may be the major DNA polymerase in strand displace- ment DNA synthesis [40]. However, comparative ana- lysis of BER in wild-type and Polb null cell extracts showed the occurrence of long-patch BER, even in the absence of Polb, suggesting that various DNA poly- merases provide functional redundancy in long-patch BER DNA synthesis [38,41]. Various interactions among BER proteins may alter the choice of BER subpathways. Ape1, when bound to DNA, interacts with Polb, which also physically inter- acts with the scaffold protein, XRCC1 [33,50,51]. Poly(ADP-ribose) polymerase (PARP-1), the enzyme that immediately binds to the incised AP site and undergoes self-ADP-ribosylation, interacts with XRCC1 and Polb and affects BER [51,52]. The involvement of PARP-1 can increase the overall BER rate, especially by enhancing short-patch BER, by ant- agonizing the action of Polb, producing a complete block of long patch BER strand-displacement DNA synthesis [53]. Long-patch BER reactions are also well co-ordinated through protein–protein interactions between PCNA and various BER enzymes, including Polb, Pold ⁄ e, FEN1 and DNA ligase I [9,54–56]. When such interactions are disrupted by p21-derived peptide that binds specifically to PCNA, the mode of AP site repair was skewed towards short-patch BER, but only in the presence of Polb [41,57]. Recently, adenomatous polyposis coli, the tumor suppressor protein, has been implicated in preventing Polb-mediated strand dis- placement synthesis by masking the domain of Polb that interacts with PCNA, thereby decreasing long- patch BER, but not short-patch BER [58]. An additional variation of BER has been suggested, as some bifunctional DNA glycosylases are associated with AP lyase activity that can carry out the cleavage of AP sites by b-elimination. These reactions generate 3¢ termini that are blocked by the lyase product, which must be removed by an enzyme, such as Ape1, to allow repair DNA synthesis (Fig. 2). In this path- way, the 3¢ repair diesterase activity of Ape1 plays an important role [59], as it also does in the excision of 3¢ phosphoglycolate esters generated by ionizing radiation or chemical oxidation [29,60]. More recently, human polynucleotide kinase has been implicated in the repair of 3¢ phosphate damage, and its interaction with other BER proteins, including XRCC1, Polb and DNA ligase III, has been shown [61]. In general, long-patch BER has been considered to be a minor pathway relative to the predominant short- patch BER. However, several in vitro and in vivo stud- ies suggest a significant contribution of the long-patch BER mode in some circumstances, particularly in the repair of regular AP sites or of the damaged base lesions that become AP sites by the action of mono- functional DNA glycosylases [39,41,62,63]. As meas- ured by an in vivo assay using a plasmid containing a single AP site in the stop codon of the gene encoding enhanced green fluorescent protein, > 80% of the repair accompanying the reversion of the stop codon occurred by long-patch BER [63]. This result is consis- tent with a previous observation that 70–80% of uracil-initiated BER was mediated by long-patch BER, when examined by utilizing a circular DNA substrate and cell-free extracts of MEF cells [41]. The detailed mechanism that governs the selection between the short- or long-patch BER modes remains a major unknown. Previously, it has been suggested that it is the nature of the DNA lesion that determines the type of DNA glycosylase (monofunctional versus glycosylase lyase), which, in turn, determines the selection of the repair pathway [39]. BER, initiated by bifunctional DNA glycosylases with associated AP lyase activity, is mainly mediated by the short-patch pathway because the resulting BER intermediate, containing a single nucleotide gap bracketed by a J S. Sung and B. Demple BER of oxidized abasic sites FEBS Journal 273 (2006) 1620–1629 ª 2006 The Authors Journal compilation ª 2006 FEBS 1623 3¢-hydroxyl and a 5¢-phosphate, can be readily filled in by Polb. In contrast, DNA repair, involving a mono- functional DNA glycosylase that generates an AP site, may involve both the short- and long-patch BER path- ways. In this model, the removal of 5¢-dRp, which appears to be the late-limiting step in short-patch BER [64], may be critical in determining the mode of BER. DNA–protein crosslink formation in the short-patch BER of dL Chemical methods for the specific generation of dL lesions within DNA oligonucleotides have been inde- pendently developed by several laboratories [65–67]. All of these methods involve the photolysis of a stable precursor and its conversion to dL at a defined site in synthetic DNA oligonucleotides. These approaches facilitated the study of the biological fate of this key oxidative deoxyribose damage in DNA. Initial investi- gation of dL repair by Escherichia coli endonuclease III, a bifunctional DNA glycosylase associated with AP lyase activity, revealed the formation of a sta- ble DNA–protein crosslink (DPC) with dL, which was dependent on the lyase active-site lysine residue involved in b-elimination [19]. Bifunctional DNA gly- cosylase ⁄ AP lyase enzymes (hOGG1 and hNth1) found in human cells, can also crosslink to dL [68]. On the other hand, the E. coli AP endonucleases exonuclease III and endonuclease IV can efficiently incise dL resi- dues [68,69]. Consistent with these observations, the dL-induced mutation frequency measured in vivo was 32-fold elevated in AP endonuclease-deficient E. coli compared with wild-type bacteria [70]. Human Ape1 protein also incises dL residues rather efficiently, with a turnover rate (2.3 s )1 ) essentially identical to that of regular AP sites (2.4 s )1 ), and only a modest K m difference (98 nm for dL versus 21 nm for AP) [69]. Considering the abundance of Ape1 in most mammalian cell types, the most probable fate of dL residues in vivo would be cleavage on the 5¢ side to yield strand breaks with 5¢-terminal dL-5-phosphate residues (5¢-dLp). The equivalent 5¢-dRp residue is effectively processed by the dRp lyase activity of Polb during short-patch BER. However, reactions of puri- fied Polb with DNA oligonucleotide substrates con- taining Ape1-cleaved 5¢-dLp residues led to the spontaneous formation of covalent crosslinks between the DNA and the polymerase [71]. The formation of such DPCs was shown to be dependent on the dRp lyase active site Lys72 of Polb [71], suggesting that the respective lysine side chains are involved in nucleophi- lic attack on the carbonyl carbon of dL, resulting in the formation of a stable amide bond (Fig. 3B). It has been shown that bacterial nucleotide excision repair can incise DNA containing an AP lyase (or peptide) covalently cross-linked by chemical reduction in an unbroken DNA [72,73]. Unlike the dL-mediated DPC formed with an AP lyase on an unbroken DNA, the DPC formation by Polb trapping to dL occurs at the DNA strand break generated by Ape1 (compare Fig. 3B with Fig. 3C). Whether a DPC located at a DNA strand break can be handled by nucleotide exci- sion repair remains to be addressed. In an effort to determine the biological significance of such crosslink formation, a cell-free extract system was utilized to react with oligonucleotide DNA con- taining a site-specific dL residue [57]. Under nonrepair conditions (no added dNTPs or Mg 2+ ), the most pre- dominant DPC species was found to contain Polb, because this species was not observed in the reactions with extracts of Polb null mouse cells. As the dRp lyase activity of Polb constitutes the major activity for removing 5¢-dRp residues in mammalian cells [32,44], the results indicate that DPC formation, specific to the 5¢-dLp lesion, occurs mainly through the abortive attempt of the dRp lyase activity of Polb to remove this incised dL lesion. Polb displays strong affinity for 5¢-dRp residues at the incised AP site, while Ape1 recruits Polb to the incised AP site and stimulates its dRp lyase activity [50,74]. Thus, this enzyme–substrate specificity may promote the interaction of Polb with a 5¢-dLp lesion at a DNA nick, thereby increasing the rate of Polb-specific DPC formation. On the other hand, it has been recently verified that dRp lyase activ- ity lags behind the polymerase activity in the dual functions of Pol b, while Ape1 suppresses the poly- merase activity [75]. In this scenario, Ape1 may modu- late Polb to pause prior to acting at the 5¢-dLp, possibly suppressing an abortive attempt to excise the lesion. Whether interactions between Ape1 and Polb, or the involvement of other factors, stimulates or inhibits the covalent trapping of Polb to the 5¢-dLp residue, must await further analysis. Use of long-patch BER in the repair of dL The major difference found in the sequential enzymatic steps between short-patch and long-patch BER is the removal of the incised abasic residue (5¢-dRp). While the dRp lyase activity of Polb participates in the processing of this residue, an attempt to remove the 5¢-dLp residue by Polb using the same mechanism results in trapping of the repair enzyme at the lesion. In the alternative long-patch BER pathway, removal of the 5¢-dRp moiety is independent of the Polb dRp BER of oxidized abasic sites J S. Sung and B. Demple 1624 FEBS Journal 273 (2006) 1620–1629 ª 2006 The Authors Journal compilation ª 2006 FEBS lyase activity and is mediated mainly by strand dis- placement DNA synthesis followed by FEN-1 excision. Therefore, it is not unreasonable to expect that the Ape1-incised dL residue may be repaired by the long- patch BER pathway. Reconstitution of dL-mediated BER conducted with partial components of long-patch BER, including Ape1, Polb and FEN-1, revealed that the formation of dL-mediated DPC was dependent on both Ape1 (for cleavage) and Polb, but that the amount of this DPC product was markedly decreased in reactions including FEN-1 and dNTPs (Fig. 4A). Repair DNA synthesis, displacing the 5¢-dLp residue by Polb alone, did not block the DPC formation, indicating that removal of dL-containing DNA fragment by FEN1 plays a key role in preventing crosslinking with the DNA substrate (Fig. 4B). This result suggests that sequential enzymatic activities in long-patch BER can effectively process the lesion and avoid dL-mediated DPC formation. This hypothesis was further supported by the demonstration of efficient processing of a 5¢-dLp flap oligonucleotide by FEN-1 [57], consistent with previous observations showing that the enzyme tolerates a variety of small modifications of the flap 5¢ terminus [76]. Investigation of dL-mediated long-patch BER was performed by util- izing circular DNA with a defined dL residue, incuba- ted with whole-cell extracts [57]. The repair of dL was detected in both wild-type and Polb-null MEF cell extracts, with concomitant reduction of subsequent crosslinking activity. Analysis of the patch size distribu- tion associated with BER of site-specific lesions showed that the single-nucleotide replacement was the predom- inant repair patch (35% of the total) for a regular AP site in the Polb-proficient cell extract, but this event A B X=dL *= 32 P dL 5'-dL Polβ 1 2 3 4 C * * * * * Fig. 4. In vitro reconstituted long-patch base excision DNA repair (BER) mediates the repair of 2-deoxyribonolactone (dL) and inhibits the formation of a dL-mediated DNA–protein crosslink (DPC). (A) A duplex 3¢ 32 P-labeled DNA substrate, containing a site-specific dL, was incu- bated with different combinations of Ape1, DNA polymerase b (Polb) and FEN1 in the presence or absence of a dNTP mix excluding dTTP. After the incubation, one-half of each reaction mixture was analyzed on a DNA sequencing gel. Ape1 converted the majority of the DNA sub- strate to the DNA cleavage product, while additional treatments with Polb and FEN1 mediated further processing of the DNA only in the presence of dNTPs. The generation of the 11-mer is consistent with strand displacement DNA synthesis of seven nucleotides by the poly- merase, followed by removal of the displaced DNA flap by FEN1. (B) The remainder of each reaction mixture was analyzed by SDS ⁄ PAGE. The dL-mediated DPCs with Polb are observed with mobilities slower than those of Polb and the free DNA. The generation of DPC was markedly reduced when the reaction allowed the combined action of repair synthesis by Polb and flap excision by FEN1. (C) Schemes for the Ape1 incision of DNA at the 5¢ side of the dL lesion (1), the strand displacement DNA synthesis of seven nucleotides by the DNA poly- merase activity of Polb (2), removal of the 5¢-dLp-containing flap by FEN1, resulting in a nick on DNA (3), and DPC formation via an abortive attempt to remove the 5¢-dLp residue by the dRp lyase activity of Polb (4). The combined processes of (2) and (3) mediate removal of the dL-containing oligonucleotide fragment from the DNA substrate and prevent DPC formation with Polb (4). Adapted from a previous publicat- ion [57]. J S. Sung and B. Demple BER of oxidized abasic sites FEBS Journal 273 (2006) 1620–1629 ª 2006 The Authors Journal compilation ª 2006 FEBS 1625 was significantly reduced (< 10% of the total) for repair of the dL substrate. Instead, repair patches of two or more nucleotides were the predominant mode for dL with both Polb-proficient and -deficient cell extracts. It was also confirmed that only the long-patch BER mode was mostly associated with the complete repair process, including the final DNA ligation step [57]. Therefore, at least in mammalian cell extracts, dL appears to be resistant to repair by short-patch BER, but effectively and exclusively repaired by long-patch BER, thereby preventing the formation of deleterious DPC adducts in DNA. Concluding remarks In spite of numerous efforts in defining the biological and biochemical mechanisms involved in BER, the cel- lular choice of the specific BER mode remains an intriguing question. A similar diversity in BER modes is also found in E. coli [77–79], which indicates that multiple subpathways of BER are favored by evolution for defending against various types of nonbulky dam- age lesions in the genetic material. Our studies of dL-mediated BER provide at least one clear rationale for the evolution of long-patch BER to handle a naturally occurring lesion. While dL residues present serious problems for cells by mediating stable DPC formation with Polb, particularly in the course of the short-patch BER pathway, it appears that the operat- ion of the long-patch BER pathway substantially avoids this detrimental consequence. However, under conditions of extensive oxidative stress, it seems poss- ible that long-patch BER components may become limiting because of their participation in the repair of many other lesions, with the attendant hazards if short-patch BER increasingly attempts to handle dL lesions. On the other hand, the induction of proteins that could modulate the subpathways of BER, as shown with p21, may alter the outcome of BER oper- ating on dL [57]. In such circumstances, Ape1-incised dL residues could remain in the DNA for longer peri- ods, increasing the opportunity for DPC formation. Further studies of dL will provide more understanding the BER switching mechanism that governs the short- versus long-patch BER distribution under varying circumstances of damage load and repair enzyme avail- ability. Acknowledgements Work in B. Demple’s laboratory was supported by NIH grants GM40000 and CA71993. J. S. Sung was partly supported by Dongguk University Research Fund. We are grateful to our colleagues, especially Dr M. S. DeMott, for helpful discussions. References 1 Duncan BK & Miller JH (1980) Mutagenic deamination of cytosine residues in DNA. Nature 278, 560–561. 2 Mosbaugh DW & Bennett SE (1994) Uracil-excision DNA repair. Prog Nucleic Acid Res Mol Biol 48, 315– 370. 3 Escodd (European Standards Committee on Oxidative DNA Damage) (2003) Measurement of DNA oxidation in human cells by chromatographic and enzymic meth- ods. Free Radic Biol Med 34, 1089–1099. 4 Escodd (European Standards Committee on Oxidative DNA Damage) (2002) Comparative analysis of baseline 8-oxo-7,8-dihydroguanine in mammalian cell DNA, by different methods in different laboratories: an approach to consensus. Carcinogenesis 23, 2129–2133. 5 Schmutte C, Yang AS, Nguyen TT, Beart RW & Jones PA (1996) Mechanisms for the involvement of DNA methylation in colon carcinogenesis. Cancer Res 56, 2375–2381. 6 Rydberg B & Lindahl T (1982) Nonenzymatic methyla- tion of DNA by the intracellular methyl group donor S-adenosyl-L-methionine is a potentially mutagenic reaction. EMBO J 1 , 211–216. 7 Gros L, Ishchenko AA & Saparbaev M (2003) Enzy- mology of repair of etheno-adducts. Mutat Res 531, 219–229. 8 Fortini P, Pascucci B, Parlanti E, D’Errico M, Simonelli V & Dogliotti E (2003) The base excision repair: mechanisms and its relevance for cancer susceptibility. Biochimie 85, 1053–1071. 9 Fan J, Wilson DM III (2005) Protein–protein interac- tions and posttranslational modifications in mammalian base excision repair. Free Radic Biol Med 38, 1121– 1138. 10 Krokan HE, Nilsen H, Skorpen F, Otterlei M & Slup- phaug G (2000) Base excision repair of DNA in mam- malian cells. FEBS Lett 476, 73–77. 11 Lindahl T (1974) An N-glycosidase from Escherichia coli that releases free uracil from DNA containing deaminated cytosine residues. Proc Natl Acad Sci USA 71, 3283–3286. 12 Barnes DE & Lindahl T (2004) Repair and genetic con- sequences of endogenous DNA base damage in mam- malian cells. Annu Rev Genet 38, 445–476. 13 Krokan HE, Standal R & Slupphaug G (1997) DNA glycosylases in the base excision repair of DNA. Bio- chem J 325, 1–16. 14 Lindahl T & Andersson A (1972) Rate of chain break- age at apurinic sites in double-stranded deoxyribonu- cleic acid. Biochemistry 11, 3618–3623. BER of oxidized abasic sites J S. Sung and B. Demple 1626 FEBS Journal 273 (2006) 1620–1629 ª 2006 The Authors Journal compilation ª 2006 FEBS 15 Lindahl T & Nyberg B (1972) Rate of depurination of native deoxyribonucleic acid. Biochemistry 11, 3618– 3623. 16 Nakamura J & Swenberg JA (1999) Endogenous apurinic ⁄ apyrimidinic sites in genomic DNA of mam- malian tissues. Cancer Res 59, 2522–2526. 17 Loeb LA & Preston BD (1986) Mutagenesis by apurinic ⁄ apyrimidinic sites. Annu Rev Genet 20, 201–230. 18 Kappen LS & Goldberg IH (1992) Neocarzinostatin acts as a sensitive probe of DNA microheterogeneity: switching of chemistry from C-1¢ to C-4¢ by a G : T mismatch 5¢-to the site of DNA damage. Proc Natl Acad Sci USA 89, 6706–6710. 19 Hashimoto M, Greenberg MM, Kow YW, Hwang JT & Cunningham RP (2001) The 2-deoxyribonolactone lesion produced in DNA by neocarzinostatin and other damaging agents forms cross-links with the base-exci- sion repair enzyme endonuclease III. J Am Chem Soc 123, 3161–3162. 20 Pratviel G, Pitie M, Bernadou J & Meunier B (1991) Mechanism of DNA cleavage by cationic manganese porphyrins: hydroxylations at the 1¢-carbon and 5¢-car- bon atoms of deoxyriboses as initial damages. Nucleic Acids Res 19, 6283–6288. 21 Sigman DS, Mazumder A & Perrin DM (1993) Chemi- cal nucleases. Chem Rev 93, 2295–2316. 22 Demple B & DeMott MS (2002) Dynamics and diver- sions in base excision DNA repair of oxidized abasic lesions. Oncogene 21, 8926–8934. 23 Dedon PC & Goldberg IH (1992) Free-radical mechan- isms involved in the formation of sequence-dependent bistranded DNA lesions by the antitumor antibiotics bleomycin, neocarzinostatin, and calicheamicin. Chem Res Toxicol 5, 311–332. 24 Von Sonntag C (1987) The Chemical Basis of Radiation Biology. Taylor & Francis, London. 25 Roginskaya M, Bernhard WA, Marion RT & Razska- zovskiy Y (2005) The release of 5-methylene-2-furanone from irradiated DNA catalyzed by cationic polyamines and divalent metal cations. Radiat Res 163, 85–89. 26 Goyne TE & Sigman DS (1987) Nuclease activity of 1,10-phenanthroline-copper ion. Chemistry of deoxy- ribose oxidation. J Am Chem Soc 109, 2846–2848. 27 Roupioz Y, Lhomme J & Kotera M (2002) Chemistry of the 2-deoxyribonolactone lesion in oligonucleotides: cleavage kinetics and products analysis. J Am Chem Soc 124, 9129–9135. 28 Zheng Y & Sheppard TL (2004) Half-life and DNA strand scission products of 2-deoxyribonolactone oxida- tive DNA damage lesions. Chem Res Toxicol 17, 197– 207. 29 Chen DS, Herman T & Demple B (1991) Two distinct human DNA diesterases that hydrolyze 3¢-blocking de- oxyribose fragments from oxidized DNA. Nucleic Acids Res 19, 5907–5914. 30 Demple B, Herman T & Chen DS (1991) Cloning and expression of APE, the cDNA encoding the major human apurinic endonuclease: definition of a family of DNA repair enzymes. Proc Natl Acad Sci USA 88, 11450–11454. 31 Piersen CE, Prasad R, Wilson SH & Lloyd RS (1996) Evidence for an imino intermediate in the DNA poly- merase beta deoxyribose phosphate excision reaction. J Biol Chem 271, 17811–17815. 32 Matsumoto Y & Kim K (1995) Excision of deoxyribose phosphate residues by DNA polymerase beta during DNA repair. Science 269, 699–702. 33 Kubota Y, Nash RA, Klungland A, Schar P, Barnes DE & Lindahl T (1996) Reconstitution of DNA base excision-repair with purified human proteins: interaction between DNA polymerase beta and the XRCC1 pro- tein. EMBO J 15, 6662–6670. 34 Klungland A & Lindahl T (1997) Second pathway for completion of human DNA base excision-repair: recon- stitution with purified proteins and requirement for DNase IV (FEN1). EMBO J 16, 3341–3348. 35 Pascucci B, Maga G, Hubscher U, Bjoras M, Seeberg E, Hickson ID, Villani G, Giordano C, Cellai L & Dogliotti E (2002) Reconstitution of the base excision repair pathway for 7,8-dihydro-8-oxoguanine with puri- fied human proteins. Nucleic Acids Res 30, 2124–2130. 36 Dianova II, Sleeth KM, Allinson SL, Parsons JL, Bres- lin C, Caldecott KW & Dianov GL (2004) XRCC1– DNA polymerase beta interaction is required for effi- cient base excision repair. Nucleic Acids Res 32, 2550– 2555. 37 Braithwaite EK, Prasad R, Shock DD, Hou EW, Beard WA & Wilson SH (2005) Role of DNA polymerase beta in the excision step of long patch mammalian base exci- sion repair. J Biol Chem 280, 18469–18475. 38 Fortini P, Pascucci B, Parlanti E, Sobol RW, Wilson SH & Dogliotti E (1998) Different DNA polymerases are involved in the short- and long-patch base excision repair in mammalian cells. Biochemistry 37, 3575–3580. 39 Fortini P, Parlanti E, Sidorkina OM, Laval J & Dogliotti E (1999) The type of DNA glycosylase deter- mines the base excision repair pathway in mammalian cells. J Biol Chem 274, 15230–15236. 40 Dianov GL, Prasad R, Wilson SH & Bohr VA (1999) Role of DNA polymerase beta in the excision step of long patch mammalian base excision repair. J Biol Chem 274, 13741–13743. 41 Bennett SE, Sung JS & Mosbaugh DW (2001) Fidelity of uracil-initiated base excision DNA repair in DNA polymerase a-proficient and -deficient mouse embryonic fibroblast cell extracts. J Biol Chem 276, 42588–42600. 42 Dianov GL, Bischoff C, Piotrowski J & Bohr VA (1998) Repair pathways for processing of 8-oxoguanine in DNA by mammalian cell extracts. J Biol Chem 273, 33811–33816. J S. Sung and B. Demple BER of oxidized abasic sites FEBS Journal 273 (2006) 1620–1629 ª 2006 The Authors Journal compilation ª 2006 FEBS 1627 43 Stucki M, Pascucci B, Parlanti E, Fortini P, Wilson SH, Hubscher U & Dogliotti E (1998) Different DNA poly- merases are involved in the short- and long-patch base excision repair in mammalian cells. Biochemistry 37, 3575–3580. 44 Sobol RW, Horton JK, Kuhn R, Gu H, Singhal RK, Prasad R, Rajewsky K & Wilson SH (1996) Require- ment of mammalian DNA polymerase-beta in base-exci- sion repair. Nature 379, 183–186. 45 Sobol RW, Prasad R, Evenski A, Baker A, Yang XP, Horton JK & Wilson SH (2000) The lyase activity of the DNA repair protein beta-polymerase protects from DNA-damage-induced cytotoxicity. Nature 405 , 807– 810. 46 Prasad R, Dianov GL, Bohr VA & Wilson SH (2000) FEN1 stimulation of DNA polymerase beta mediates an excision step in mammalian long patch base excision repair. J Biol Chem 275, 4460–4466. 47 Dianov GL, Souza-Pinto N, Nyaga SG, Thybo T, Stevnsner T & Bohr VA (2001) Base excision repair in nuclear and mitochondrial DNA. Prog Nucleic Acid Res Mol Biol 68, 285–297. 48 Dogliotti E, Fortini P, Pascucci B & Parlanti E (2001) The mechanism of switching among multiple BER path- ways. Prog Nucleic Acid Res Mol Biol 68, 3–27. 49 Stucki M, Pascucci B, Parlanti E, Fortini P, Wilson SH, Hubscher U & Dogliotti E (1998) Mammalian base excision repair by DNA polymerases delta and epsilon. Oncogene 17, 835–843. 50 Bennett RA, Wilson DM, 3rd, Wong D & Demple B (1997) Interaction of human apurinic endonuclease and DNA polymerase beta in the base excision repair path- way. Proc Natl Acad Sci USA 94, 7166–7169. 51 Caldecott KW, Aoufouchi S, Johnson P & Shall S (1996) XRCC1 polypeptide interacts with DNA polymerase beta and possibly poly (ADP-ribose) polymerase, and DNA ligase III is a novel molecular dnick-sensorT in vitro. Nucleic Acids Res 24, 4387–4394. 52 Dantzer F, de La Rubia G, Menissier-De Murcia J, Hostomsky Z, de Murcia G & Schreiber V (2000) Base excision repair is impaired in mammalian cells lacking Poly (ADP-ribose) polymerase-1. Biochemistry 39, 7559– 7569. 53 Sukhanova MV, Khodyreva SN, Lebedeva NA, Prasad R, Wilson SH & Lavrik OI (2005) Human base excision repair enzymes apurinic ⁄ apyrimidinic endonuclease1 (APE1), DNA polymerase beta and poly (ADP-ribose) polymerase 1: interplay between strand-displacement DNA synthesis and proofreading exonuclease activity. Nucleic Acids Res 33, 1222–1229. 54 Dianova II, Bohr VA & Dianov GL (2001) Interaction of human AP endonuclease 1 with flap endonuclease 1 and proliferating cell nuclear antigen involved in long- patch base excision repair. Biochemistry 40, 12639– 12644. 55 Levin DS, McKenna AE, Motycka TA, Matsumoto Y & Tomkinson AE (2000) Interaction between PCNA and DNA ligase I is critical for joining of Okazaki frag- ments and long-patch base-excision repair. Curr Biol 10, 919–922. 56 Kedar PS, Kim SJ, Robertson A, Hou E, Prasad R, Horton JK & Wilson SH (2002) Direct interaction between mammalian DNA polymerase beta and prolif- erating cell nuclear antigen. J Biol Chem 277, 31115– 31123. 57 Sung JS, Demott MS & Demple B (2005) Long-patch base excision DNA repair of 2-deoxyribonolactone prevents the formation of DNA-protein cross-links with DNA polymerase beta. J Biol Chem 280, 39095–39103. 58 Narayan S, Jaiswal AS & Balusu R (2005) Tumor sup- pressor APC blocks DNA polymerase a-dependent strand displacement synthesis during long patch but not short patch base excision repair and increases sensitivity to methylmethane sulfonate. J Biol Chem 280, 6942– 6949. 59 Demple B & Harrison L (1994) Repair of oxidative damage to DNA: enzymology and biology. Annu Rev Biochem 63, 915–948. 60 Izumi T, Hazra TK, Boldogh I, Tomkinson AE, Park MS, Ikeda S & Mitra S (2000) Requirement for human AP endonuclease 1 for repair of 3¢-blocking damage at DNA single-strand breaks induced by reactive oxygen species. Carcinogenesis 21, 1329–1334. 61 Whitehouse CJ, Taylor RM, Thistlethwaite A, Zhang H, Karimi-Busheri F, Lasko DD, Weinfeld M & Calde- cott KW (2001) XRCC1 stimulates human polynucleo- tide kinase activity at damaged DNA termini and accelerates DNA single-strand break repair. Cell 104, 107–117. 62 Horton JK, Prasad R, Hou E & Wilson SH (2000) Pro- tection against methylation-induced cytotoxicity by DNA polymerase beta-dependent long patch base exci- sion repair. J Biol Chem 275, 2211–2218. 63 Sattler U, Frit P, Salles B & Calsou P (2003) Long- patch DNA repair synthesis during base excision repair in mammalian cells. EMBO Rep 4, 363–367. 64 Srivastava DK, Berg BJ, Prasad R, Molina JT, Beard WA, Tomkinson AE & Wilson SH (1998) Mammalian abasic site base excision repair. Identification of the reaction sequence and rate-determining steps. J Biol Chem 273, 21203–21209. 65 Kotera M, Roupioz Y, Defrancq E, Bourdat AG, Garcia J, Coulombeau C & Lhomme J (2000) The 7-nitroindole nucleoside as a photochemical precursor of 2¢-deoxyribonolactone: access to DNA fragments containing this oxidative abasic lesion. Chemistry 6, 4163–4169. 66 Lenox HJ, McCoy CP & Sheppard TL (2001) Site-speci- fic generation of deoxyribonolactone lesions in DNA oligonucleotides. Org Lett 3, 2415–2418. BER of oxidized abasic sites J S. Sung and B. Demple 1628 FEBS Journal 273 (2006) 1620–1629 ª 2006 The Authors Journal compilation ª 2006 FEBS 67 Tronche C, Goodman BK & Greenberg MM (1999) DNA damage induced via independent generation of the radical resulting from formal hydrogen atom abstraction from the C1¢-position of a nucleotide. Chem Biol 5, 263–271. 68 Faure V, Saparbaev M, Dumy P & Constant JF (2005) Action of multiple base excision repair enzymes on the 2¢-deoxyribonolactone. Biochem Biophys Res Commun 328, 1188–1195. 69 Xu YJ, DeMott MS, Hwang JT, Greenberg MM & Demple B (2003) Action of human apurinic endonu- clease (Ape1) on C1¢-oxidized deoxyribose damage in DNA. DNA Repair 2, 175–185. 70 Faure V, Constant JF, Dumy P & Saparbaev M (2004) 2¢-deoxyribonolactone lesion produces G fi A transi- tions in Escherichia coli. Nucleic Acids Res 32, 2937–2946. 71 DeMott MS, Beyret E, Wong D, Bales BC, Hwang JT, Greenberg MM & Demple B (2002) Covalent trapping of human DNA polymerase a by the oxidative DNA lesion 2-deoxyribonolactone. J Biol Chem 277, 7637– 7640. 72 Minko IG, Kurtz AJ, Croteau DL, Van Houten B, Harris TM & Lloyd RS (2005) Initiation of repair of DNA-polypeptide cross-links by the UvrABC nuclease. Biochemistry 44, 3000–3009. 73 Minko IG, Zou Y & Lloyd RS (2002) Incision of DNA-protein crosslinks by UvrABC nuclease suggests a potential repair pathway involving nucleotide excision repair. Proc Natl Acad Sci USA 99, 1905–1909. 74 Feng JA, Crasto CJ & Matsumoto Y (1998) Deoxyri- bose phosphate excision by the N-terminal domain of the polymerase a: the mechanism revisited. Biochemistry 37, 9605–9611. 75 Wong D & Demple B (2004) Modulation of the 5¢-deo- xyribose-5-phosphate lyase and DNA synthesis activities of mammalian DNA polymerase a by apurinic ⁄ apyri- midinic endonuclease 1. J Biol Chem 279, 25268–25275. 76 Bornarth CJ, Ranalli TA, Henricksen LA, Wahl AF & Bambara RA (1999) Effect of flap modifications on human FEN1 cleavage. Biochemistry 38, 13347–13354. 77 Sung JS & Mosbaugh DW (2003) Escherichia coli ura- cil- and ethenocytosine-initiated base excision DNA repair: rate-limiting step and patch size distribution. Biochemistry 42, 4613–4625. 78 Sandigursky M, Freyer GA & Franklin WA (1998) The post-incision steps of the DNA base excision repair pathway in Escherichia coli: studies with a closed circu- lar DNA substrate containing a single U : G base pair. Nucleic Acids Res 26, 1282–1287. 79 Dianov G & Lindahl T (1994) Reconstitution of the DNA base excision-repair pathway. Curr Biol 4, 1069– 1076. J S. Sung and B. Demple BER of oxidized abasic sites FEBS Journal 273 (2006) 1620–1629 ª 2006 The Authors Journal compilation ª 2006 FEBS 1629 . MINIREVIEW Roles of base excision repair subpathways in correcting oxidized abasic sites in DNA Jung-Suk Sung 1 and Bruce Demple 2 1 Department of Life. process by forming a protein DNA crosslink. Abbreviations AP, apurinic ⁄ apyrimidinic; BER, base excision DNA repair; DPC, DNA protein crosslink; dL, 2-deoxyribonolactone;