E. coli mismatch repair acts downstream of replication fork stalling to stabilize the expanded (GAA.TTC)(n) sequence

Abstract

Expanded triplet repeat sequences are known to cause at least 16 inherited neuromuscular diseases. In addition to short-length changes, expanded triplet repeat tracts frequently undergo large changes, often amounting to hundreds of base-pairs. Such changes might occur when template or primer slipping creates insertion/deletion loops (IDLs), which are normally repaired by the mismatch repair system (MMR). However, in prokaryotes and eukaryotes, MMR promotes large changes in the length of (CTG.CAG)(n) sequences, the motif most commonly associated with human disease. We tested the effect of MMR on instability of the expanded (GAA.TTC)(n) sequence, which causes Friedreich ataxia, by comparing repeat instability in wild-type and MMR-deficient strains of Escherichia coli. As expected, the prevalence of small mutations increased in the MMR-deficient strains. However, the prevalence of large contractions increased in the MMR mutants specifically when GAA was the lagging strand template, the orientation in which replication fork stalling is known to occur. After hydroxyurea-induced stalling, both orientations of replication showed significantly more large contractions in MMR mutants than in the wild-type, suggesting that fork stalling may be responsible for the large contractions. Deficiency of MMR promoted large contractions independently of RecA status, a known determinant of (GAA.TTC)(n) instability. These data suggest that two independent mechanisms act in response to replication stalling to prevent instability of the (GAA.TTC)(n) sequence in E. coli, when GAA serves as the lagging strand template: one that is dependent on RecA-mediated restart of stalled forks, and another that is dependent on MMR-mediated repair of IDLs. While MMR destabilizes the (CTG.CAG)(n) sequence, it is involved in stabilization of the (GAA.TTC)(n) sequence. The role of MMR in triplet repeat instability therefore depends on the repeat sequence and the orientation of replication.

Figure 1

E. coli MMR prevents large contractions (>7 triplets) of the (GAA·TTC)_n sequence specifically when GAA is the lagging strand template. (A,B) Representative agarose gels showing PCR products generated from colonies containing the (GAA·TTC)₇₉ repeat tract in the GAA (A) and TTC (B) orientations in the MMR-proficient (wild-type) strain and three isogenic MMR-deficient mutants, mutS, mutL, and mutH. Short changes in length (<7 triplets) were more common in the mutants than in the wild-type for both GAA-79 and TTC-79 (P<0.01). Arrowheads indicate the position of the full-length repeat. Horizontal lines along the right edge of the gel pictures represent the position of 100-bp, 200-bp, 300-bp, 400-bp and 500-bp, based on the 1 Kb Plus DNA Ladder (Invitrogen). (C) The frequency of large contractions, calculated as the number of large contractions per successful PCR, was significantly enhanced for GAA-79 when propagated in the three MMR-deficient strains relative to the isogenic MMR-proficient strain (P<0.01 for each comparison of mutant vs. wild-type). No such increase in the frequency of large contractions was noted for TTC-79. The moderate increase in large contractions observed for mutH, while statistically significant (P<0.01), is unlikely to be biologically relevant. (D) In contrast, the frequency of large contractions was significantly reduced for CTG-98 when propagated in the three MMR-deficient strains compared to the isogenic MMR-proficient strain (P<0.01 for each mutant). Error bars depict +/- 2SEM.

Figure 2

Replication stalling promotes large contractions of the (GAA·TTC)_n sequence in E. coli MMR mutants. Cultures were either treated with 100 mM hydroxyurea to induce replication fork stalling, or with 0 mM hydroxyurea as a control. As expected, in the absence of hydroxyurea, GAA-79 (A) showed a significant increase in the frequency of large contractions in the MMR-deficient strains mutS, mutL, and mutH compared to the isogenic wild-type, but TTC-79 (B) showed no such increase (P<0.01 for GAA-79; P=1.0 for TTC-79). After the induction of replication fork stalling with hydroxyurea, the loss of MMR led to an increased frequency of large contractions for TTC-79 (P<0.01); however, hydroxyurea treatment did not further increase large contractions of GAA-79 (P>0.1 for treated vs. untreated in each strain). Error bars depict +/- 2SEM.

Figure 3

Loss of MMR promotes large contractions of the (GAA·TTC)_n sequence independently of RecA status. The individual mutation of recA or mutS significantly increased the frequency of large contractions for GAA-79 (P<0.01), but had no effect on TTC-79. The recA mutS double mutant showed a higher frequency of large contractions than either single mutant for both GAA-79 and TTC-79 (P<0.001 in both orientations). Error bars depict +/- 2SEM.

Figure 4

Independent functions of RecA and MMR in maintaining stability of the (GAA·TTC)_n sequence in E. coli. The replication fork is known to stall when GAA serves as the lagging strand template. Efficient RecA-dependent restart of stalled replication forks is required to maintain the length of the repeat tract. In a RecA-deficient state, error-prone replication restart promotes large contractions. Additionally, failure to repair stalled forks may lead to DSB formation, and repair of DSBs within the (GAA·TTC)_n sequence is known to cause large contractions. The role of MMR in correcting IDLs that may form at the stalled fork is likely to be critical for preventing large contractions. In the absence of MMR, large contractions of the (GAA·TTC)_n sequence may result from bypass of unresolved IDLs, resulting in a shorter nascent strand during replication.