SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The expansion of normally polymorphic CTG microsatellites in certain human genes has been identified as the causative mutation of a number of hereditary neurological disorders, including Huntington's disease and myotonic dystrophy. Here, we have investigated the effect of methyl-directed mismatch repair (MMR) on the stability of a (CTG)43 repeat in Escherichia coli over 140 generations and find two opposing effects. In contrast to orientation-dependent repeat instability in wild-type E. coli and yeast, we observed no orientation dependence in MMRE. coli cells and suggest that, for the repeat that we have studied, orientation dependence in wild-type cells is mainly caused by functional mismatch repair genes. Our results imply that slipped structures are generated during replication, causing single triplet expansions and contractions in MMR cells, because they are left unrepaired. On the other hand, we find that the repair of such slipped structures by the MMR system can go awry, resulting in large contractions. We show that these mutS-dependent contractions arise preferentially when the CTG sequence is encoded by the lagging strand. The nature of this orientation dependence argues that the small slipped structures that are recognized by the MMR system are formed primarily on the lagging strand of the replication fork. It also suggests that, in the presence of functional MMR, removal of 3 bp slipped structures causes the formation of larger contractions that are probably the result of secondary structure formation by the CTG sequence. We rationalize the opposing effects of MMR on repeat tract stability with a model that accounts for CTG repeat instability and loss of orientation dependence in MMR cells. Our work resolves a contradiction between opposing claims in the literature of both stabilizing and destabilizing effects of MMR on CTG repeat instability in E. coli.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The expansion of CTG repeat tracts is responsible for several important hereditary neurodegenerative and neuromuscular diseases, including Huntington's disease and myotonic dystrophy (reviewed by Reddy and Housman, 1997). The inheritance of these diseases is characterized by anticipation (earlier age at disease onset and more severe progression of symptoms) resulting from further expansion of the trinucleotide repeat upon parent-to-child transmission (reviewed by La Spada, 1997). Mutational mechanisms by which selected trinucleotide repeats are destabilized are still poorly understood. A contradiction has arisen concerning the role of MMR genes in CTG repeat instability from studies of these sequences in Escherichia coli. Jaworski et al. (1995), who studied long CTG repeats (130 and 180 triplets) in E. coli, reported that the action of MMR resulted in elevated frequencies of large deletions when the CTG-containing strand was the template for leading strand synthesis. This was unexpected, because it is known that the action of MMR is normally antimutagenic. Wells et al. (1998) went on to analyse by subcloning the deletion products generated in MMR+ and MMR cells and found evidence for additional events within individual plasmids recovered from MMR mutants. The identification of these additional events, which included small insertions and deletions of one to eight triplets, was possible because of interruptions in the original repeat tract. The small length changes were not observed in the clones obtained from MMR+ cells, and Wells et al. (1998) proposed that misalignment of tract interruptions provided the targets for MMR. Schumacher et al. (1998) have used an assay that is sensitive to both small ( ± one triplet) and large (> ± one triplet) changes in repeat array length within a population of molecules and concluded that an array of (CTG)64 is globally stabilized by the action of MMR by preventing small and long expansions and contractions. They argued that their results were in contradiction to the destabilizing effect of MMR observed previously by Jaworski et al. (1995). We present evidence here for two opposing effects of MMR acting on one CTG repeat array. Our work supports the conclusion that, under certain circumstances, the action of MMR can cause instability.

We have studied the influence of mutations in genes implicated in MMR (mutS, mutL), secondary structure-dependent repair (sbcCD) and recombination (recA) on the stability of a (CTG)43 repeat that was derived from the highly polymorphic CTG repeat of the murine metallothionein-3 (Mt3) gene (Abbott and Chambers, 1994). In E. coli, the mismatch repair protein, MutS, recognizes and binds up to four mismatched basepairs and translocates the DNA flanking the mismatch into an alpha-shaped heteroduplex loop (Allen et al., 1997). MutL binds to MutS–heteroduplex complexes (Grilley et al., 1989), and Drotschmann et al. (1998) recently reported that binding of MutL increases the stability of the DNA–MutS complex. Arrival of the MutS–MutL complex at the hemi-methylated GATC site activates the MutH-associated endonuclease. MutH incises the unmethylated DNA 5′ to the G of GATC sequences either 3′ or 5′ to the mismatch (reviewed by Modrich, 1991). Subsequent steps include unwinding of the DNA from the resulting nick by DNA helicase II (Au et al., 1992; Dao and Modrich, 1998), exonucleolytic degradation of the error-containing strand (Cooper et al., 1993) and resynthesis by DNA polymerase III holoenzyme (Au et al., 1992).

We have also tested the possibility that the (CTG)43 repeat is a substrate for the structure-dependent endonuclease activity of the SbcCD nuclease complex in E. coli. Mutations in sbcC and sbcD, the two genes in E. coli encoding the nuclease complex, can restore viability to E. coli cells carrying palindromic sequences on their chromosome (reviewed by Leach, 1994). In wild-type cells, SbcCD is known to recognize and cleave hairpins that the palindrome can form on the lagging strand template when it is single stranded during replication (Connelly and Leach, 1996; Connelly et al., 1998). Double-strand breaks so generated by SbcCD are repaired via homologous recombination (Leach et al., 1997). If a secondary structure formed by the pseudopalindromic CTG repeat were recognized and cleaved by SbcCD, then misalignment between the invading strand and the template strand could lead to contraction and/or expansion of the repeat during repair by homologous recombination (Darlow and Leach, 1995).

We have followed length changes within a population of repeat arrays in wild-type E. coli and in several DNA repair-deficient strains over ≈ 140 generations. We find that, in wild-type cells, the (CTG)43 repeat expands and contracts, generating every repeat length between five and 69 triplets. In contrast to orientation-dependent repeat instability in wild-type E. coli and yeast, we observe no orientation dependence in MMR cells and suggest that, for the repeat we have studied, orientation dependence in wild-type cells is caused by functional MMR genes. We propose a model in which repeat destabilization by MMR deficiency is caused mainly by DNA slippage events involving 3 bp that occur independently of repeat orientation and in which orientation-dependent repeat instability in wild-type cells is caused by the generation of contractions via the action of MMR.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Assay

The uninterrupted CTG repeat array that we have analysed consists of 43 triplets, which puts it into the lower range of mutated, unstable repeat arrays found in several human genetic disease genes. We have measured repeat length changes in two plasmids carrying the (CTG)43 repeat in the two opposite orientations (see Experimental procedures) in several DNA repair-defective E. coli strains (mutS, mutL, recA, sbcCD, mutS sbcCD, mutL sbcCD). To minimize variation between experiments, populations of more than 500 transformants were used for the first inoculation instead of single transformants. Multiple samples taken from the same mixture of transformants produced identical banding patterns (data not shown). Cultures were grown for ≈ 140 generations, and plasmid DNA was isolated from these cultures at four different times (after ≈ 30, 40, 90 and 140 generations). The instability of the CTG repeat array in both orientations was then assessed by measuring the percentage of each repeat length in the plasmid population (see Experimental procedures). Our assay allowed us to determine the change in the percentage of each repeat length within the contracted and expanded plasmid population, as well as of plasmids with the original repeat length. Experiments that were carried out in preparation for this study showed that a repeat of 43 triplets allows for extensive expansions, and competition experiments detected no replicative advantage during the establishment of pUC18 with respect to a plasmid with the (CTG)43 repeat array.

Repeat instability in MMR+ strains

In MMR+ cells, the CTG repeat array expands and contracts in both orientations, but the degree of this instability is orientation dependent (Table 1, Fig. 1[link]). In both orientations, contracted plasmids can contain as few as five triplets, whereas expanded plasmids can carry up to 69 triplets (Fig. 2). In orientation A, the number of plasmids containing the original repeat length decreases from 70% at the beginning of the time course to 47% after 140 generations. In orientation B, this decrease is greater, with the percentage of plasmids carrying a repeat of the original length falling from 75% to 35%. The amount of contracted plasmids increases from 8% to 30% in orientation A and from 14% to 44% in orientation B. When the CTG repeat serves as the lagging strand template, 26% of all plasmids in the population have lost more than seven triplets, whereas in the opposite orientation, only 12% of all plasmids fall into this group. Plasmids with deletions of one to seven triplets constitute ≈ 18% of all plasmids in both orientations (Table 2, Fig. 3A[link]). This suggests that the group of plasmids that have lost more than seven triplets accounts for the difference in the amount of full-length repeat arrays between the opposite orientations (Table 2, Fig. 3B[link]).

Table 1. . Instability of the (CTG)43 triplet repeat in MMR+ strains (wild type, recA, sbcCD) and MMR strains (mutS, mutL, mutS sbcCD, mutL sbcCD). Percentage values for original repeat arrays (containing 43 triplets), contracted repeat arrays (containing less than 43 triplets) and expanded repeat arrays (containing more than 43 triplets) were determined after ≈ 140 generations in the presence (+) or absence (−) of 1 mM IPTG to induce transcription (values are the average of at least two independent experiments and are shown with 95% confidence intervals). A and B indicate the orientation of the CTG repeat with respect to the origin of replication (A, CTG on the leading strand; B, CTG on the lagging strand).Thumbnail image of
image

Figure 1. . Analysis of the stability of a (CTG)43 repeat as a function of cell growth. Plasmid DNA was extracted from MMR+E. coli strains (wild type, recA) and an MMRE. coli strain (mutS) after 30, 40, 90 and 140 generations. The percentage of repeat arrays containing the original number of 43 triplets was measured on native 5% Long Ranger gels as described in Experimental procedures. Black lines (▴) indicate results for orientation A, while grey lines (▪) show results for orientation B, in the absence of IPTG. Hatched lines in black (◆) and grey (●) indicate results for orientation A and B, respectively, in the presence of IPTG.

Download figure to PowerPoint

image

Figure 2. . PhosphorImages showing the electrophoretic analysis of [α35S]-dATP EcoRI fragments containing CTG repeats from pDL915 (orientation A) and pDL915R (orientation B). Plasmid DNA was extracted and analysed biochemically from wild-type E. coli and from an E. coli mutS mutant after ≈ 40 and 140 generations. All eight lanes show the original repeat length of 43 triplets as the strongest band and numerous other bands (always separated by 3 bp), representing expanded and contracted repeat arrays. A and B indicate the orientation of the CTG repeat with respect to the origin of plasmid replication (A, CTG on the leading strand; B, CTG on the lagging strand).

Download figure to PowerPoint

Table 2. . Subdivision of contracted repeat arrays isolated from MMR+ and MMR strains after ≈ 140 generations (shown in Table 1) into repeat arrays that have lost either one to seven triplets (< seven triplets) or more than seven triplets (> seven triplets), in the presence (+) or absence (−) of 1 mM IPTG to induce transcription (values are the average of at least two independent experiments and are shown with 95% confidence intervals). A and B indicate the orientation of the CTG repeat with respect to the origin of replication (A, CTG on the leading strand; B, CTG on the lagging strand).Thumbnail image of
image

Figure 3. . Effect of functional MMR on small and large contractions within the (CTG)43 repeat measured over 140 generations. Plasmid DNA was extracted analysed and biochemically from wild-type E. coli and from an E. coli mutS mutant after 30, 40, 90 and 140 generations. For this purpose, percentage values for contracted repeat arrays were subdivided into repeat arrays that have lost either one to seven triplets (A) or more than seven triplets (B). A and B indicate the orientation of the CTG repeat with respect to the origin of plasmid replication (A, CTG on the leading strand; B, CTG on the lagging strand).

Download figure to PowerPoint

Expanded repeat arrays are more frequent in orientation A than in orientation B at all time points analysed. However, over the 100 generations studied, the proportion of expanded plasmids increases from 12% to 21% when the CTG repeat is on the lagging strand template (orientation B) and remains at about 23% if the CAG repeat is on the lagging strand template (orientation A). Although there is no net increase in expansion during the time course, expansion will still have to occur at a level that allows the cell to maintain the number of expanded plasmids that were present at the beginning of the time course while contraction is going on. With our assay, however, we cannot establish whether large expansions and contractions are generated by the accumulation of numerous small changes or whether they occur through single large jumps.

Repeat instability in MMR strains

In MMR cells, repeat instability in both orientations is increased. This increase is caused mainly by single triplet insertions and deletions, which are rarely observed in any of the three MMR+ strains (wild type, recA, sbcCD) that we have tested (Fig. 4). Moreover, and in contrast to what we have found in wild-type cells, the stability of the repeat array in MMR cells is not orientation dependent (Fig. 1). In both orientations, the percentage of full-length repeat arrays falls from 70% to 75% at 0 generations to 23–26% after 140 generations (Table 1). The change in the pattern of repeat contraction and expansion in MMR cells is directly evident in a comparison of gel images obtained from several MMR+ and MMR strains (Fig. 4). Six basepair gaps below and above the original band (43 triplets) appear in the three MMR+ strains tested as a result of the absence of plasmids containing 42 and 44 triplets. In MMR cells (mutS, mutL), however, these gaps are starting to be filled in after 40 generations and have disappeared completely at the end of the time course. The introduction of these single triplet changes has a similar effect on repeat stability in both orientations and, therefore, does not explain the elimination of orientation dependence in MMR strains. In fact, our data reveal a stronger destabilizing effect of MMR deficiency on orientation A than on orientation B. When the CAG repeat serves as the lagging strand template (orientation A), we observe after 140 generations 47% of all plasmids with the original repeat length in wild type and 26% plasmids with the original repeat length in mutS. In the opposite orientation (B), the decrease is only from 35% in wild type to 23% in mutS. This loss of orientation dependence is mainly caused by a significant reduction (two-tailed t-test, < 0.05), relative to wild-type cells, in the frequency of plasmids that have lost more than seven triplets in orientation B (Table 2, Fig. 3B[link]). In effect, the size and frequency of contractions that the repeat is subjected to in mutS mutants are similar in the two orientations. Contractions are probably small and accumulate, generating plasmids with deletions as big as 38 triplets. Our data obtained from wild-type cells (Table 1) suggest that expansion events may be slightly more frequent in orientation A than in orientation B. This bias is retained in MMR bacteria in which we observe an increase in expanded arrays (Table 1). As this increase is similar in the two orientations (e.g. from 23% in wild type to 36% in mutS for orientation A, and from 22% in wild type to 33% in mutS for orientation B), any orientation dependence of expansion events that may exist in wild-type cells is not caused by the action of MMR.

image

Figure 4. . Comparison of PhosphorImages showing electrophoretic analysis of [α35S]-dATP EcoRI fragments of pDL915 and pDL915R isolated from MMR+E. coli strains (wild type, recA, sbcCD) and MMRE. coli strains (mutS, mutL). Plasmid DNA was extracted and analysed biochemically after ≈ 40 and 140 generations (see Experimental procedures). After 140 generations, 6 bp gaps are observed just above and below the original band of 43 triplets in MMR strains, whereas these gaps are not present in repeat populations isolated from MMR strains. A and B indicate the orientation of the CTG repeat with respect to the origin of replication (A, CTG on the leading strand; B, CTG on the lagging strand).

Download figure to PowerPoint

Recombination and structure-dependent DNA repair

Using our assay, we could not detect important effects of recA or sbcCD mutations on repeat instability. In order to determine whether this was caused by binding of MutS and MutL protein to the repeat, which might prevent the structure-dependent nuclease complex SbcCD from recognizing any unusual secondary structure adopted by the repeat array, we tested repeat instability in mutS sbcCD and mutL sbcCD double mutants. Table 1 shows that the levels of repeat instability in mutS sbcCD- and mutL sbcCD-deficient cells are similar to those in mutS and mutL single mutants.

Lack of transcriptional effect on CTG repeat instability

It can be seen from Fig. 1 and from Tables 1 and 2 that no aspect of CTG repeat instability that we have studied here was influenced by the absence or presence of 1 mM IPTG in the culture medium. This lack of a transcriptional effect is in contrast to what has been reported for long plasmid-borne (CTG)175 repeats (Bowater et al., 1997) and will be discussed below.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

In this paper, we have shown that MMR has two opposing effects on the stability of one CTG repeat array. On the one hand, as part of the MMR system, MutS initiates repair of heteroduplexes containing 3 bp heterologies, thus preventing single repeat unit insertions and deletions. This effect is independent of the orientation of the repeat relative to the direction of DNA replication, supporting the findings of Schumacher et al. (1998) that the action of MMR leads to an overall stabilization of both CTG repeat orientations. On the other hand, in agreement with the report by Jaworski et al. (1995), we show that, under certain circumstances, a mutS mutation can also lead to a reduction in the frequency of large contractions. This observation of opposing effects of MMR on small- and large-scale repeat length changes has led us to propose a model in which lagging strand-biased slipped structure formation and slipped structure repair explain orientation-dependent CTG repeat instability in E. coli (Fig. 5). This model is similar to that of Jaworski et al. (1995), but has two significant differences. First, it is the CTG-containing lagging strand template that is the source of MMR-promoted contractions, whereas Jaworski et al. (1995) suggested that it was the CTG-containing leading strand. This discrepancy may be explained by pathways that act in addition to MMR-promoted contraction on long but not on short CTG repeats. Second, we demonstrate that the likely source of recruitment of the MMR complex are single triplet mispairs that we show to be effectively corrected by MMR. We do not support the suggestion by Schumacher et al. (1998) that MutS is recruited to A·A and T·T mismatches within the pseudo-hairpins of the CAG and CTG strands. We argue, in contrast, that a significant route to the formation of CTG-containing hairpins on the lagging strand template is via the action of MMR. We show that heterologies derived from interruptions in the array are not required for the MMR-promoted pathway, as has been shown by Wells et al. (1998), but we do not rule out the possibility that they might play a stimulatory role. The orientation dependence for MMR-promoted deletion events argues against their generation by a duplex melting and slippage pathway, as suggested by Wells et al. (1998), because that pathway would not be expected to show strand preference with respect to the leading and lagging strands of the replication fork.

image

Figure 5. . Model for MMR-promoted contraction of short plasmid-borne CTG triplet repeats in E. coli. We propose that mispairs of 3 bp occur preferentially during lagging strand replication and are recognized by a functional MMR machinery. In MMR cells, 3 bp slipped structures persist and lead to single triplet expansion (A) and contraction (B), whereas in MMR+ cells, a single-stranded region is formed during the excision reaction (C) and, if the sequence in this single-stranded region has the capability of secondary structure formation (D), an MMR-dependent contraction is generated during repair synthesis (E). In the absence of secondary structure formation (F), functional MMR does not promote repeat contraction (G). Bold lines symbolize the triplet repeat sequence, while 2 mm vertical lines indicate methylated GATC sites.

Download figure to PowerPoint

Our results suggest that both MMR and some special feature of the (CTG)43 repeat sequence when present on the lagging strand are involved in CTG repeat contraction. Trinh and Sinden (1991) proposed that the asymmetry of leading and lagging strand replication can bias the occurrence of mutations between the strands. In corroboration of this idea, Rosche et al. (1995) reported that, if DNA slippage could be stabilized by a secondary structure in the lagging strand (e.g. a hairpin), deletion occurred 50-fold more frequently than when this structure formed on the leading strand, supporting a model proposed by Leach (1994) in which slipped misalignment involving DNA secondary structure occurs preferentially in the lagging strand during DNA replication. In further support of this idea, Pinder et al. (1998) demonstrated recently that deletions associated with palindromic DNA sequences occur preferentially during lagging strand replication.

The observation that MMR-promoted contractions are seen only when the CTG repeat is the lagging strand template suggests that slipped structures are formed primarily on the lagging strand of the replication fork, independently of repeat orientation. If slipped structures could form at a significant frequency on the leading strand, we would expect to have seen MMR-promoted deletions when the CTG repeat was the template for leading strand synthesis. Frequent dissociation and reassociation of DNA polymerase to the primer end during lagging strand replication was suggested by Veaute and Fuchs (1993) to explain a higher frequency of misincorporation on the lagging strand than on the leading strand. We suggest that 3 bp slipped structures form primarily on the lagging strand but do so independently of repeat orientation, which explains the loss of orientation dependence in MMR mutants. This would lead to similar frequencies of single-stranded gap formation by the action of MMR in the two repeat orientations. However, MMR-promoted deletions are observed in only one orientation, namely if the CTG repeat serves as the lagging strand template. The lower frequency of MMR-promoted contractions in orientation A can be explained by an incapability of the CAG repeat to adopt a stable enough secondary structure. This is consistent with in vitro experiments that have predicted a higher stability of hairpins formed by the CTG repeat than of those formed by the complementary sequence (Gacy et al., 1995; Mitas et al., 1995; Petruska et al., 1996). The situation may be more complex than this, as the recent study by Gacy and McMurray (1998) indicates that CAG and CTG hairpins have a similar ability to inhibit CAG/CTG duplex formation. They propose that in vivo factors may affect these sequences differently. It has been shown that MutS and MutL can remain associated with heteroduplex excision intermediates (Drotschmann et al., 1998), which may aid structure formation. Viswanathan and Lovett (1998) have recently suggested that frameshift mutations are promoted by the MMR pathway in strains deficient in exonuclease I and exonuclease VII. Incomplete degradation of single-stranded DNA by exonuclease I and exonuclease VII may contribute to an elevated level of frameshift mutations in repetitive DNA.

That we have not detected effects of mutations in recA or sbcCD on the stability of the repeat might indicate that the SbcCD system is oversaturated because of the high copy number of repeat-containing plasmids. A similar lack of effect of recA or sbcD mutations on the stability of TG dinucleotide repeats cloned in the E. coli chromosome has been observed (Morel et al., 1998). In dinucleotide experiments, oversaturation of SbcCD by repeat-containing replicons cannot be invoked to explain the lack of effect. Alternatively, the (CTG)43 repeat may not be long enough to form secondary structures that are recognized and cleaved by the SbcCD nuclease complex. This would be compatible with a role for the SbcCD complex in preventing expansion events of substantially longer CTG repeat arrays, as reported recently by Sarkar et al. (1998).

In contrast to Bowater et al. (1997), who reported a destabilizing effect of transcription on a plasmid-borne (CTG)175 repeat in E. coli, we find that transcription into the shorter (CTG)43 repeat has no effect on its stability. This may be explained by the much longer CTG repeat (175 triplets) studied by Bowater et al. (1997) and/or by differences in the experimental design. That transcription-dependent contractions of (CTG)175 were only seen when the CTG sequence was on the lagging strand of the replication fork (referred to here as orientation B) and after the culture had passed through stationary phase suggested to these authors that transcription into the (CTG)175 repeat influences the transition between stationary and exponential growth phase, conferring a growth advantage to cells harbouring deleted CTG repeats. Such a growth disadvantage was not observed when the CTG repeat served as the leading strand or with either orientation of shorter repeats, e.g. (CTG)98 (CTG)17, and may therefore also not be expected for (CTG)43 (Bowater et al., 1996). In addition, differences in the experimental design, such as co-transformation of a plasmid expressing the lacIQ repressor to ensure repression of transcription in the presence of high-copy-number plasmids (Bowater et al., 1997), may also explain why we, without having included such modifications, did not detect a transcriptional effect on the stability of (CTG)43.

In summary, our results suggest that during in vivo replication of a (CTG)43 repeat, small slippages involving 3 bp, 6 bp and 9 bp happen frequently during lagging strand replication. We have shown that 3 bp slipped structures are readily removed by MMR. However, the reliability of this repair may be jeopardized if the DNA sequence on the lagging strand template can adopt stable secondary structures. This can cause repeat contraction and orientation-dependent repeat instability in MMR+ strains. The nature of this orientation dependence strongly suggests that the small slipped structures are primarily formed on the lagging strand of the replication fork. MMR may be one of several molecular mechanisms that involve the formation of single-stranded DNA and, by doing so, affect the stability of triplet repeats.

Experimental procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Strains, media and growth conditions

E. coli strains were all isogenic derivatives of JM83 (Yanisch-Perron et al., 1985). JM83 was transduced to make JM83 mutS::Tn10 (DL902) and JM83 mutL::Tn10 (DL936) using P1 lysate made from CSH115 and CSH116 respectively (Miller, 1972). DL733 carries a genomic deletion of sbcC and sbcD and was P1 transduced to make JM83 ΔsbcCD mutS::Tn10 (DL905) and JM83 ΔsbcCD mutL::Tn10 (DL927). DL887 is JM83 recA::Cm. E. coli strains were grown at 37°C in Luria–Bertani (LB) medium supplemented with ampicillin (100 μg ml−1).

Plasmid construction

Using the polymerase chain reaction (PCR), we have amplified an uninterrupted (CTG)25 repeat from the murine metallothionein III (Mt3) gene including 88 bp of flanking sequence, as described before (Abbott and Chambers, 1994). This PCR product was ligated to 8 bp EcoRI linkers and inserted into the single EcoRI site of the high-copy-number plasmid pUC18 to form plasmid pDL913. After passage of pDL913 through a mutS-defective E. coli strain, we obtained plasmid pDL915, which contains an expanded CTG repeat array of 43 triplets. The EcoRI fragment in pDL915 was excised and then reinserted into pUC18 in the two opposite orientations to form plasmids pDL915 and pDL915R. A single diagnostic Sau3A restriction site in the flanking sequence of the repeat array was used to determine the orientation of the repeat array relative to the origin of replication. In pDL915, referred to here as orientation A, the CAG repeat serves as the lagging strand template, whereas in pDL915R, referred to here as orientation B, the CTG repeat serves as the lagging strand template.

Transformation and time course

CaCl2-competent cells (100 μl) were transformed (Sambrook et al., 1989) with ≈ 100 ng of plasmid DNA. Plasmid DNA was monomeric, and the same DNA preparation was used in all transformations. Approximately 500 transformants were suspended in 5 ml of LB medium, and 50 μl of this suspension was used to inoculate 5 ml of LB liquid medium (supplemented with 100 μg ml−1 ampicillin). Cultures were grown at 37°C and diluted 100-fold every 24 h for 14 days (≈ 140 generations). Cells were grown in the presence or absence of 1 mM IPTG to induce transcription.

DNA extraction and quantification of repeat length changes

Plasmid DNA was prepared from transformants (> 500 pooled) before the first inoculation (≈ 30 generations), and then after ≈ 40 generations (1 day of cultivation), after 90 generations (7 days of cultivation) and after 140 generations (14 days of cultivation), using the Plasmid Mini Spin kit (Qiagen). EcoRI-digested plasmid DNA was labelled radioactively by incorporation of [α35S]-dATP using Klenow enzyme. Labelled restriction fragments were separated on native 5% Long Ranger gels (Flowgen) at a constant power of 35 W for ≈ 1.5 h. Dried gels were exposed overnight to a phosphor storage screen, and radioactive bands were visualized using a Molecular Dynamics phosphorImager. Signals were quantified using Molecular Dynamics image-quant software. The sum of the signal intensity of all radioactive bands in every lane was set at 100%, and the percentage of each band was determined. Using this biochemical assay, we cannot exclude the possibility that, on occasion, one or both EcoRI sites may have been removed during repeat contraction or expansion, thereby excluding these changes from detection.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The authors are grateful to F. Stahl for helpful comments on the manuscript. We thank members of our laboratory for reading the manuscript and for important discussions. K.H.S. is supported by the Darwin Trust (University of Edinburgh), and D.R.F.L. is supported by a grant from the Medical Research Council (UK).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  • 1
    Abbott, C. & Chambers, D. (1994) Analysis of CAG trinucleotide repeats from mouse cDNA sequences. Ann Hum Genet 58: 8794.
  • 2
    Allen, D.J., Makhov, A., Grilley, M., Taylor, J., Thresher, R., Modrich, P., et al (1997) MutS mediates heteroduplex loop formation by a translocation mechanism. EMBO J 16: 44674476.
  • 3
    Au, K.G., Welsh, K., Modrich, P. (1992) Initiation of methyl-directed mismatch repair. J Biol Chem 267: 1214212148.
  • 4
    Bowater, R.P., Rosche, W.A., Jaworski, A., Sinden, R.R., Wells, R.D. (1996) Relationship between Escherichia coli growth and deletions of CAG·CTG triplet repeats in plasmids. J Mol Biol 264: 8296.
  • 5
    Bowater, R.P., Jaworski, A., Larson, J.E., Parniewski, P., Wells, R.D. (1997) Transcription increases the deletion frequency of long CTG·CAG triplet repeats from plasmids in Escherichia coli. Nucleic Acids Res 25: 28612868.
  • 6
    Connelly, J.C. & Leach, D.R.F. (1996) The sbcC and sbcD genes of Escherichia coli encode a nuclease involved in palindrome inviability and genetic recombination. Genes Cells 1: 285291.
  • 7
    Connelly, J.C., Kirkham, L., DeLeau, E., Leach, D.R.F. (1998) The SbcCD nuclease of Escherichia coli is a structural maintenance of chromosomes (SMC) family protein that cleaves hairpin DNA. Proc Natl Acad Sci USA 95: 79697972.
  • 8
    Cooper, D.L., Lahue, R.S., Modrich, P. (1993) Methyl-directed mismatch repair is bidirectional. J Biol Chem 268: 1182311829.
  • 9
    Dao, V. & Modrich, P. (1998) Mismatch-, MutS-, MutL-, and helicase II-dependent unwinding from the single-strand break of an incised heteroduplex. J Biol Chem 273: 92029207.
  • 10
    Darlow, J.M. & Leach, D.R.F. (1995) The effects of trinucleotide repeats found in human inherited disorders on palindrome inviability in Escherichia coli suggest hairpin folding preferences in vivo. Genetics 141: 825832.
  • 11
    Drotschmann, K., Aronshtam, A., Fritz, H.J., Marinus, M.G. (1998) The Escherichia coli MutL protein stimulates binding of Vsr and MutS to heteroduplex DNA. Nucleic Acids Res 26: 948953.
  • 12
    Gacy, A.M. & McMurray, C.T. (1998) Influence of hairpins on template reannealing at trinucleotide repeat duplexes: a model for slipped DNA. Biochemistry 37: 94269434.
  • 13
    Gacy, G.A., Goellner, G., Juranic, N., Macura, S., McMurray, C.T. (1995) Trinucleotide repeats that expand in human disease form hairpin structures in vitro. Cell 81: 533540.
  • 14
    Grilley, M., Welsh, K.M., Su, S.-S., Modrich, P. (1989) Isolation and characterization of the Escherichia coli mutL gene product. J Biol Chem 264: 10001004.
  • 15
    Jaworski, A., Rosche, W.A., Gellibolian, R., Kang, S., Shimizu, M., Bowater, R.P., et al (1995) Mismatch repair in Escherichia coli enhances instability of (CTG) n triplet repeats from human hereditary diseases. Proc Natl Acad Sci USA 92: 1101911023.
  • 16
    La Spada, A.R. (1997) Trinucleotide repeat instability: genetic features and molecular mechanisms. Brain Pathol 7: 943963.
  • 17
    Leach, D.R. (1994) Long DNA palindromes, cruciform structures, genetic instability and secondary structure repair. Bioessays 16: 893900.
  • 18
    Leach, D.R., Okely, E.A., Pinder, D.J. (1997) Repair by recombination of DNA containing a palindromic sequence. Mol Microbiol 26: 597606.
  • 19
    Miller, J.H. (1972) Experiments in Molecular Genetics. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press.
  • 20
    Mitas, M., Yu, A., Dill, J., Kamp, T.J., Chamber, E.J., Haworth, I.S. (1995) Hairpin properties of single-stranded DNA containing a GC-rich triplet repeat: (CTG)15. Nucleic Acids Res 23: 10501059.
  • 21
    Modrich, P. (1991) Mechanisms and biological effects of mismatch repair. Annu Rev Genet 25: 229253.
  • 22
    Morel, P., Reverdy, C., Michel, B., Ehrlich, S.D., Cassuto, E. (1998) The role of SOS and flap processing in microsatellite instability in Escherichia coli. Proc Natl Acad Sci USA 95: 1000310008.
  • 23
    Petruska, J., Arnheim, N., Goodman, M.F. (1996) Stability of intrastrand hairpin structures formed by the CAG/CTG class of DNA triplet repeats associated with neurological diseases. Nucleic Acids Res 24: 19921998.
  • 24
    Pinder, D.J., Blake, C.E., Lindsey, J.C., Leach, D.R.F. (1998) Replication strand preference for deletions associated with DNA palindromes. Mol Microbiol 28: 719727.
  • 25
    Reddy, P.S. & Housman, D.E. (1997) The complex pathology of trinucleotide repeats. Curr Opin Cell Biol 9: 364372.
  • 26
    Rosche, W.A., Trinh, T.Q., Sinden, R.R. (1995) Differential DNA secondary structure-mediated deletion mutation in the leading and lagging strands. J Bacteriol 177: 43854391.
  • 27
    Sambrook, J., Fritsch, E.F., Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual. New York: Cold Spring Harbor Laboratory Press.
  • 28
    Sarkar, P.S., Chang, H.C., Boudi, F.B., Reddy, S. (1998) CTG repeats show bimodal amplification in E. coli. Cell 95: 531540.
  • 29
    Schumacher, S., Fuchs, R.P., Bichara, M. (1998) Expansion of CTG repeats from human disease genes is dependent upon replication mechanisms in Escherichia coli: the effect of long patch mismatch repair revisited. J Mol Biol 279: 11011110.
  • 30
    Trinh, T.Q. & Sinden, R.R. (1991) Preferential DNA secondary structure mutagenesis in the lagging strand of replication in E. coli. Nature 352: 544547.
  • 31
    Veaute, X. & Fuchs, R.P. (1993) Greater susceptibility to mutations in lagging strand of DNA replication in Escherichia coli than in leading strand. Science 261: 598600.
  • 32
    Viswanathan, M. & Lovett, S.T. (1998) Single-strand DNA-specific exonucleases in Escherichia coli. Roles in repair and mutation avoidance. Genetics 149: 716.
  • 33
    Wells, R.D., Parniewski, P., Pluciennik, A., Bacolla, A., Gellibolian, R., Jaworski, A. (1998) Small slipped register genetic instabilities in Escherichia coli in triplet repeat sequences associated with hereditary neurological diseases. J Biol Chem 273: 1953219541.
  • 34
    Yanisch-Perron, C., Vieira, J., Messing, J. (1985) Improved M13 phage cloning vectors and host strains: nucleotide sequences of the M13mp18 and pUC19 vectors. Gene 33: 103119.