Trinucleotide repeat instability via DNA base excision repair

Abstract

Trinucleotide repeat (TNR) instability is the cause of over 40 human neurodegenerative diseases and certain types of cancer. TNR instability can result from DNA replication, repair, recombination, and gene transcription. Emerging evidence indicates that DNA base damage and base excision repair (BER) play an active role in regulating somatic TNR instability. These processes may potentially modulate the onset and progression of TNR-related diseases, given that TNRs are hotspots of DNA base damage that are present in mammalian cells with a high frequency. In this review, we discuss the recent advances in our understanding of the molecular mechanisms underlying BER-mediated TNR instability. We initially discuss the roles of the BER pathway and locations of DNA base lesions in TNRs and their interplay with non-B form DNA structures in governing repeat instability. We then discuss how the coordinated activities of BER enzymes can modulate a balance between the removal and addition of TNRs to regulate somatic TNR instability. We further discuss how this balance can be disrupted by the crosstalk between BER and DNA mismatch repair (MMR) machinery resulting in TNR expansion. Finally, we suggest future directions regarding BER-mediated somatic TNR instability and its association with TNR disease prevention and treatment.

Introduction

Trinucleotide repeats (TNRs) are prone to length instability in the form of both expansions and deletions (1,2). TNR expansions are the cause of more than 40 human neurodegenerative diseases including Huntington’s disease (HD)(CAG/CTG) and Friedreich’s ataxia (FRDA) (GAA/TTC), among others (1). Deletions of CAG repeats in the coding region of the androgen receptor gene are associated with prostate and ovarian cancers (3,4).

Over the past 20 years, substantial progress has been made in the understanding of the mechanisms underlying TNR instability using bacteria, yeast, mammalian cells, and mice as model systems (5). It has been shown that TNR instability is mediated by the formation of non-B form DNA structures during DNA replication, repair, recombination, and gene transcription. DNA repair, in particular, is implicated in facilitating the age-dependent TNR expansions in non-dividing neurons (6). It has been shown that only small sizes of TNR expansions (5-20 repeat units) that usually result from DNA repair can occur in neurons (6). Recent studies have further demonstrated that base excision repair (BER) is the major pathway that mediates the age-dependent TNR instability (7). This is because neurons have a high metabolic rate and exhibit an age-dependent increase of oxidative DNA damage and single-strand DNA (ssDNA) breaks that are subject to the BER pathway (7).

TNRs are rich in Gs and Cs, therefore form hotspots of DNA base lesions. Several studies have pointed to an association among oxidative stress, BER, and TNR instability (7,8). It has been shown that increased oxidative DNA damage is associated with TNR expansions (7,8). Previous studies from our group and others also have shown that oxidized and alkylated DNA bases induced by DNA damaging agents can result in both TNR expansions and deletions in human cells (2,9,10). In addition, the Usdin group has shown that exogenous 8-oxoguanine (8-oxoG) induced by potassium bromate can promote CGG repeat expansion in fragile X syndrome transgenic mice (8). These results suggest a role of DNA base damage and BER in facilitating TNR instability in response to oxidative DNA damage. The notion is further supported by the fact that CAG expansions in HD transgenic mice are dependent on 8-oxoG glycosylase (OGG1) (7), and deficiency of the endonuclease VIII-like 1 (NEIL1) glycosylase decreases both germline and somatic CAG repeat expansions in HD mice (11) indicating that mammalian TNR expansions are mediated by DNA glycosylases that initiate the BER pathway. Furthermore, it is found that deficiency of the polymerase activity of a DNA polymerase β (pol β) mutant, Y265C preferentially promotes CGG repeat deletion in fragile X syndrome transgenic mice (12) further demonstrating a crucial role of pol β-mediated BER in regulating TNR instability.

Nevertheless, the molecular mechanisms underlying BER-mediated TNR instability are not well understood. Here we summarize the results from the studies in the past decade that advance our understanding of the mechanisms that govern BER-mediated somatic TNR instability. These include the roles of the locations of DNA base lesions in TNR tracts, the coordination among BER enzymes and cofactors, and the crosstalk between BER and mismatch repair protein MSH2-MSH3 in regulating TNR instability during BER.

TNRs can form a variety of DNA secondary structures during DNA metabolism. These include small loops, hairpins, triplexes, tetraplexes such as G-quadruplexes, and sticky DNA (13). It is generally accepted that TNR secondary structures generated in the newly synthesized strand lead to repeat expansions, whereas the structures generated in the template strand cause repeat deletions (1). During BER, TNR secondary structures can disrupt the coordination between core BER proteins leading to TNR instability (14). We have found that the locations of a base lesion determine if repeats are expanded or deleted. An abasic lesion located at the 5’-end of a CTG repeat tract can result in the formation of a large hairpin at the downstream strand, whereas a lesion located in the middle of the repeats leads to the formation of small upstream and downstream hairpins, respectively (Figure 1) (9). The difference in the location and size of the secondary structures results in the imbalance in the activity of pol β and FEN1 that either adds or removes TNRs, thereby determining if the repeats are expanded or deleted. We show that a base lesion at the 5’-end of a CTG repeat tract induces repeat expansions as a result of a more efficient synthesis of the repeats by pol β than the removal of the repeats by flap endonuclease 1 (FEN1)(Figure 1) (9). However, a base lesion is located in the middle of the repeat tract results in the formation of small upstream and downstream CTG repeat hairpins. We have found that the upstream CTG repeat hairpin inhibits pol β synthesis of the repeats. In contrast, the downstream hairpin is readily to be converted to a flap that can be efficiently cleaved by FEN1 flap cleavage activity. This results in more repeats removed than synthesized leading to repeat deletion (Figure 1) (9). The results indicate a position effect of a base lesion on TNR instability that is mediated by the imbalance of pol β and FEN1 activities.

Figure 1. The locations of DNA base lesions govern TNR instability through BER.

Oxidative stress leads to the formation of oxidized DNA base lesions, i.e., 8-oxoGs, in a CTG repeat tract. OGG1 removes 8-oxoGs leaving an abasic site that is 5′-incised by APE1. In the scenario where a base lesion occurs at the 5′-end of the repeat tract, Pol β performs multi-nucleotide gap-filling DNA synthesis to pass through the template hairpin, whereas FEN1 cleavage of the repeats is inhibited by the hairpin. This results in repeat expansions (the pathway on the left). On the other hand, a base lesion located in the middle of CTG repeats results in the formation of an upstream hairpin and a downstream hairpin along with two hairpins on the template. The upstream hairpin inhibits pol β DNA synthesis. The downstream hairpin is converted into a flap that is efficiently cleaved by FEN1, leading to repeat deletions (the pathway on the right) (9).

BER-mediated TNR expansions are dependent on pol β multi-nucleotide gap-filling synthesis that adds extra repeats through long patch-BER (LP-BER) (14). The pol β multi-nucleotide gap-filling synthesis on CAG repeats not only fills in the gap but also produces additional repeats, preventing downstream TNRs from reannealing to the template and stabilizing the hairpin. The stable hairpin further prevents FEN1 typical flap cleavage activity and forces FEN1 alternate flap cleavage. This leads to the removal of fewer repeats creating a ligatable nick for DNA ligase I (LIG I) to seal, thereby incorporating the hairpin as repeat expansion (Figure 2).

Figure 2. TNR expansions and deletions through BER.

BER of an abasic site in the CAG repeats is initiated by APE1 incision, leaving an ssDNA break. This subsequently leads to the formation of CAG repeat hairpins and a multi-nucleotide gap on the damaged or the template strand. Pol β performs multi-nucleotide gap-filling synthesis to fill the gap on the damaged strand. FEN1 use its alternate flap cleavage activity to process the 5’-end of a short flap attached to the CAG hairpin, leaving a ligatable nicked hairpin that is sealed by LIG I resulting in CAG repeat expansion (the pathway on the left) (14). In a scenario where a hairpin forms on the template strand, pol β skips over the template hairpin and displaces the downstream repeat into a flap that is cleaved by FEN1 causing repeat deletion (the pathway on the right) (15). The BER cofactor, PCNA can interact with LIG I to stimulate ligation across a small template hairpin (the pathway on the right), thereby promoting repeat deletions (18).

TNRs are deleted during BER through pol β skip-over of a template hairpin or loop structure (15). This allows the displacement of the downstream repeats into a flap that is efficiently cleaved by FEN1. Subsequently, this results in more repeats cleaved than synthesized, thereby causing repeat deletions (15) (Figure 2). TNR deletions can also be induced by a bulky DNA base lesion, 5’,8-cyclodeoxyadenosine located in the middle of the template strand of a CTG repeat tract (16) (Figure 2). The bulky DNA base lesion distorts the backbone of the CAG repeat tract and induces the formation of a small loop structure. This then promotes pol β to skip over the loop structure leading to CTG repeat deletions (16) (Figure 2). Also, the weak DNA synthesis resulting from a pol β polymorphism R137Q facilitates the pol β skip-over of a small template TNR loop structure leading to repeat deletions via BER (17) (Figure 2). On the other hand, the BER cofactor, proliferating cell nuclear antigen (PCNA) can promote TNR deletions by stimulating the activity of LIG I on a nick across a small TNR loop during BER (18) (Figure 2). A recent study from the Freudenreich’s group has shown that a bulky non-B form structure, R-loop can promote CAG repeat deletions through the BER pathway in yeast (19). This suggests that TNR R-loops facilitate the removal of TNRs during BER in R-loops, thereby leading to repeat deletions. However, the molecular mechanisms by which BER promotes TNR deletion in the context of R-loops need to be elucidated.

BER can also attenuate TNR expansion by removing TNR hairpins (20,21) (Figure 3). TNR hairpins form hotspots of oxidative DNA damage, and the guanines located in the hairpin loop of a CAG hairpin are particularly susceptible to the formation of oxidized lesions (22). We have found that an 8-oxoG located in the hairpin loop region can be removed by OGG1, leaving an abasic site that is incised by APE1 at the 5’-end. This converts the hairpin into a “double-flap intermediate” containing an upstream 3’-flap and a downstream 5’-flap (20) (Figure 3). The double-flap intermediate is then resolved by FEN1 cleavage of the downstream 5’-flap in coordination with the cleavage of the upstream 3’-flap by the 3’-5’ flap endonuclease, Mus81/Eme1 complex. This results in complete or partial removal of the hairpin through BER preventing or attenuating TNR expansions (20) (Figure 3). We further demonstrate that the removal of a TNR hairpin during BER also can be facilitated by APE1 3’-5’ exonuclease activity (21) (Figure 3), which progressively removes the TNRs in the upstream strand by capturing the annealed region of the upstream 3’-flap. In addition, APE1 can stimulate LIG I to specifically ligate a nick without a hairpin, thereby preventing TNR expansions (21) (Figure 3). As a sliding clamp, PCNA can promote the removal of TNR hairpin by sliding towards the downstream facilitating the reannealing of the upstream 3’-TNR flap and formation of a downstream 5’-TNR flap that is subsequently removed by FEN1 flap cleavage activity (18) (Figure 3). Thus, BER enzymes and cofactors employ their unique mechanisms to process TNR hairpins preventing and attenuating TNR expansion.

Figure 3. BER attenuates TNR expansion via the removal of a trinucleotide repeat hairpin.

Oxidative stress can induce 8-oxoGs in the loop region of a CAG hairpin. OGG1 removes an 8-oxoG leaving an abasic site that is 5’-incised by APE1, resulting in the formation of a double-flap intermediate with a 3’- and 5’-CAG repeat flap. The 3’-flap can be cleaved by the 3’−5’ endonuclease, Mus81/Eme1 (20) or APE1 3’-5’ exonuclease (21), whereas the 5’-flap is cleaved by FEN1. Pol β fills in a gap, and LIG I seals a nick leading to the removal of the TNR hairpin (the pathway on the left). PCNA can be loaded from the upstream of the double-flap intermediate and slides toward the downstream, pushing the 3’-flap to reanneal to the template strand. PCNA then interacts with FEN1 and stimulates its flap cleavage activity, thereby promoting the removal of the hairpin and prevention of TNR expansions (the pathway on the right) (18).

Recent studies from our group have shown that pol β deoxyribose phosphate (dRP) lyase also plays a crucial role in preventing TNR instability. We have discovered that an oxidized sugar, 5'-(2-phosphoryl-1,4-dioxobutane)(DOB) at the double-flap intermediate generated during BER in a CAG repeat hairpin can crosslink with the dRP lyase domain of pol β and LIG I (23) inhibiting the pol β DNA synthesis activity and nick ligation (23). This then results in the accumulation of strand break intermediates in the repeat tract that can be subject to double-strand break repair, promoting large repeat instability (23). Furthermore, we have found that the dRP lyase activity of pol β is essential in preventing TNR deletion (24). The deficiency of pol β dRP lyase activity locks the enzyme to the dRP group tethering the polymerase to the template strand, thereby forcing pol β to perform a processive synthesis of TNRs promoting the hairpin/loop bypassing and repeat deletion (24). Our findings suggest that pol β dRP lyase activity facilitates the interaction between pol β dRP lyase domain and the dRP group that subsequently inhibits pol β processive DNA synthesis and DNA slippage, thereby preventing repeat instability.

TNR instability can also be regulated through the crosstalk between BER and MMR. We have found that pol β can crosstalk with MSH2-MSH3 protein complex by physically interacting with mismatch repair protein on TNR tracts (25). Pol β alone preferentially skips over a template hairpin or loop leading to TNR deletions. However, the crosstalk between pol β and MSH2-MSH3 allows the mismatch protein complex to promote pol β to synthesize through a template hairpin or loop generating a downstream 5’-flap precursor intermediate for repeat expansion (25). This demonstrates that the crosstalk between pol β and MSH2-MSH3 shifts the outcome from repeat deletions to expansions during BER. Thus, the crosstalk between BER and MMR forms “a novel hybrid pathway” that promotes TNR expansion (25) (Figure 4). Our findings indicate the crosstalk between BER and other DNA repair pathways can readily alter the balance of TNR synthesis and removal, thereby the outcome of TNR instability.

Figure 4. Crosstalk between BER and MMR promotes TNR expansions and suppresses TNR deletions.

Oxidative stress induces an 8-oxoG in a TNR tract. OGG1 removes the 8-oxoG and leaves an abasic site that is subsequently 5’-incised by APE1. In the absence of MSH2-MSH3, pol β skips over a template loop structure, leading to repeat deletions (the pathway on the left). In the presence of MSH2-MSH3, MSH2-MSH3 interacts with pol β and guide the polymerase to copy through the repeats within the loop, generating a downstream loop with a short flap that is cleaved by FEN1 alternate flap cleavage. This promotes the synthesis of repeats and results in the incorporation of the loop leading to repeat expansions (the pathway on the right) (25).

Concluding Remarks and Future Directions

The studies in the past decade on the molecular mechanisms underlying BER-mediated TNR instability indicate that a balance between the addition and removal of TNRs plays a crucial role in governing repeat instability. We demonstrate that the balance can be readily modulated by the locations of DNA base lesions in TNR tracts, the formation of hairpins and loops on different DNA strands and their stability, the coordination among BER enzymes and cofactors, and crosstalk between BER and MMR, thereby shifting the outcome of TNR instability (Figure 5). However, our current understanding of the molecular mechanisms underlying TNR instability via BER is limited to the roles of key BER enzymes and cofactors and their coordination with mismatch repair proteins. There remain many outstanding questions regarding somatic TNR instability through BER that need to be addressed. For example, how DNA base lesions in TNRs, BER, other DNA repair pathways, DNA methylation, and chromatin structures can interplay to govern TNR instability and how bulky DNA structures such as R-loops can impact BER-mediated TNR instability through an interplay with epigenetic features on TNR tracts. Future efforts should also be dedicated to the development of BER as a target for effective treatment of TNR-related diseases.

Figure 5. The balance between the addition and removal of TNRs during BER determines the outcome of TNR instability.

The outcome of TNR instability is determined by the balance of the synthesis and removal of TNRs during BER. The enzymes or cofactors, such as pol β, MSH3-MSH3, OGG1, and NEIL1 that favor the generation of ssDNA breaks and addition of the repeats promote repeat expansion via BER. In contrast, the enzymes and cofactors, such as FEN1, APE1, Mus81/EME1, and PCNA that promote the removal of the repeats facilitate repeat deletions through BER.