A widely accepted model for repeat instability suggests that deletions result from primer-template misalignment, as discussed already. Large deletions have been suggested to occur by replication slippage across d(CTG)„ hairpins in the lagging template strand when it is single-stranded (as shown in Fig. 2) (Kang et al. 1995a; Freudenreich et al. 1997; Schweitzer and Livingston 1998, 1999; Miret et al. 1998; Sinden 1999; Ireland et al. 2000; Rolfsmeier et al. 2001; Hashem et al. 2002; Panigrahi et al. 2002; Lee and Park 2002; Mar-cadier and Pearson 2003; Bhattacharyya and Lahue 2004; Liu et al. 2004b). However, functional RecA and RecB are required for the high rates of repeat instability in E. coli (Hashem et al. 2004b), and a simple model of replication slippage across a hairpin in the lagging template strand cannot account for the involvement of RecA and RecB. Rather, repeat deletions may result from errors occurring during replication restart following the collapse of the replication fork during synthesis of the repeats (Hashem et al. 2002, 2004b; Kim et al. 2006). At present, the molecular events responsible for replication pausing are uncertain; however, hairpins, slipped mispaired DNA, or other secondary structures may play a role in blocking or pausing replication fork progression in d(CTG)„ ■ d(CAG)„ or d(CGG)„ ■ d(CCG)„ repeats (Usdin and Woodford 1995; Kang et al. 1995b; Sinden 1999; Hartenstine et al. 2000; Kamath-Loeb et al. 2001; Heidenfelder et al. 2003), while triplex DNA formation could participate in replication pausing in d(GAA)„ ■ d(TTC)n repeats (Gacy et al. 1998; Grabczyk and Usdin 2000; Potaman et al. 2004; Krasilnikova and Mirkin 2004). Double-strand breaks are often, but not always, associated with recombination and they can result in repeat deletion when they occur within repeats, as shown in both bacteria and mammalian cells (Marcadier and Pearson 2003; Hebert et al. 2004).
Several pathways are available for restarting a collapsed or paused fork. Here, these are described with respect to repeat deletion as understood for E. coli (Kim et al. 2006), but they also act in eukaryotic cells and may be responsible for spontaneous, as well as drug-induced deletion in d(CTG)„ ■ d(CAG)„ deletion in DM1 lymphoblasts (Hashem et al. 2004a). Replication restart of stalled replication forks requires DNA replication, recombination, and repair proteins (Cox et al. 2000, 2001; Marians 2000; McGlynn and Lloyd 2002). A pathway for the orientation when d(CAG) comprises the leading template strand is shown in Fig. 7. Leading-strand synthesis may be spontaneously paused during synthesis of the repeats, stalled by a short (3-bp) misalignment, or may be blocked by a stable DNA secondary structure in the leading template strand (the pause site is denoted by the asterisk in Fig. 7, step A). Following leading-strand blockage, lagging-strand replication continues (Fig. 7, step B). After fork collapse, the unwinding of stalled forks by RecG or RuvABC in E. coli leads to fork reversal and formation of a Holliday junction (here called a "chicken-foot" structure) through annealing of the leading and lagging nascent strands (Fig. 7, step D). Cleavage of the Holliday junction by RuvABC resolvase generates a duplex DNA (Fig. 7, step E) in which the 5' end can be resected by RecBCD nuclease (Fig. 7, step F). RecA can then initiate recombination and restore the fork (Fig. 7, steps G-J) (Hashem et al. 2004b; Kim et al. 2006). This may be the major pathway for repeat deletion, as mutations in recA and recB can decrease deletion rates by factors of more than 1000 (Hashem et al. 2002, 2004b). The potential for d(CTG) hairpin formation when single-stranded (Fig. 7, steps G-I), and a preference for restart via the RecA- and RecBC-dependent pathway may explain the generally observed bias for deletions in this orientation, as discussed previously (Hashem et al. 2004b; Kim et al. 2006).
The stalled fork may also be rescued by other pathways. One pathway employs an exonuclease to trim the lagging nascent strand (Fig. 7, steps C-K). Alternatively, the stalled fork (Fig. 7, step A) could simply collapse (Fig. 7,
Fig. 7 Replication restart can explain orientation dependence for repeat instability. Replication restart following a block to DNA replication and fork collapse is required to complete duplication of the chromosome to ensure cell viability. Several pathways are available for this process. A major pathway involves fork reversal (step D) and the introduction of a double-strand break (step E), which is repaired by recombination functions within a cell (steps F-I). A high rate of repeat instability in E. coli is dependent on RecA and RecBC, which precludes a simple replication-based model (as shown in Fig. 2) for their participation in repeat instability. The pathways shown may account for a high rate of repeat deletion in E. coli and explain the orientation-dependent greater instability when the d(CTG)„ tract comprises the lagging template strand. These pathways are described in detail in the text
Fig. 7 Replication restart can explain orientation dependence for repeat instability. Replication restart following a block to DNA replication and fork collapse is required to complete duplication of the chromosome to ensure cell viability. Several pathways are available for this process. A major pathway involves fork reversal (step D) and the introduction of a double-strand break (step E), which is repaired by recombination functions within a cell (steps F-I). A high rate of repeat instability in E. coli is dependent on RecA and RecBC, which precludes a simple replication-based model (as shown in Fig. 2) for their participation in repeat instability. The pathways shown may account for a high rate of repeat deletion in E. coli and explain the orientation-dependent greater instability when the d(CTG)„ tract comprises the lagging template strand. These pathways are described in detail in the text steps A-K). During digestion, or following collapse, a hairpin may form in the d(CTG)„ strand (Fig. 7, step K). Reannealing of the leading and lagging template strands would then drive the formation of slipped-strand DNA (Fig. 7, step M). Fork reversal could occur (Fig. 7, step L), which would move the slipped-strand DNA away from the Holliday junction, making it available for DNA repair, as observed in several systems (Oussatcheva et al. 2001; Panigrahi et al. 2005), and leading to changes in repeat length. Resolution of the junction shown in step L would create the double-strand break, similar to molecules shown in step F, but with slipped-strand DNA in one molecule.
Samadashwily et al. (1997) have reported the strength of replication fork pausing in E. coli during lagging-strand synthesis to be in the order d(CGG) > d(CCG) > d(CTG) > d(CAG). Pausing during synthesis of d(CGG) ■ d(CCG) tracts between 14 and 31 repeats was clearly evident; however, replication fork pausing in a d(CAG)70 ■ d(CTG)70 tract was only detected following chloramphenicol treatment to induce plasmid amplification. Moreover, pausing was only detected when d(CTG) comprised the lagging template strand. Biochemical detection of pausing in E. coli and yeast has been interpreted to be initiated by DNA secondary structure formation in the lagging template strand (Samadashwily et al. 1997; Pelletier et al. 2003; Krasilnikova and Mirkin 2004). The pathway shown in Fig. 7 discusses pausing as a leading strand event because DNA secondary structure in the lagging strand may not be expected to permanently block fork progression since lagging-strand replication can start on either side of the structure.
The restart of stalled forks is also important for mammalian cells, and pathways analogous to those discussed for E. coli may be important for instability in human cells. Human cells respond rapidly to DNA damage, including stalled replication forks, UV-light-induced photoproducts, and chemothera-peutic drug lesions, by arresting cells in the S phase (intra-S phase checkpoint) and allowing repair of the damage (Kastan and Bartek 2004). DNA damage caused by various exogenous factors leads to the activation of the DNA damage checkpoint pathways (reviewed in Melo and Toczyski 2002). These pathways are essential for preventing irreversible breakdown of replication forks stalled at the sites of DNA damage (Tercero et al. 2003). Intra-S checkpoints have also been shown to be involved in normal DNA replication (Cha and Kleckner 2002). The S. cerevisiae genome contains about 1500 sites where DNA replication slows, and mutations in the MEC1 gene, a human ataxia telangectasia-related and Rad 3 related (ATR) homologue, accentuate stalling at those sites, resulting in chromosomal breakage (Cha and Kleck-ner 2002). Thus, intra-S checkpoints may stabilize stalled replication forks even in the absence of DNA damage. Consistent with the expectation that fork blockage during replication of repeats, or double-strand breaks generated as a consequence of replication fork restart or DNA repair events, might activate the DNA damage checkpoint response, d(CAG) ■ d(CTG) repeats can activate the DNA damage response in S. cerevisiae (Lahiri et al. 2004). Mutations in the MEC1, RAD9, or RAD53 genes increased the rates of chromosome breakage associated with a (CAG) ■ (CTG) repeat tract. Deficiencies in Mec1, Ddc2, Rad17, Rad24, or Rad53 resulted in an increase in the frequency of repeat deletions. Interestingly, expansions were also increased in cells deficient in Rad24, Rad17, and Rad53. These results suggest that replication or repair events are altered when normal checkpoint controls become compromised.
Was this article helpful?