Sequence Composition and Context Effects on the Generation and Repair of Frameshift Intermediates in Mononucleotide Runs in Saccharomyces cerevisiae

Corresponding author: Sue Jinks-Robertson, Department of Biology, 1510 Clifton Rd., Emory University, Atlanta, GA 30322., jinks{at}biology.emory.edu (E-mail)

Communicating editor: M. LICHTEN

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

DNA polymerase slippage occurs frequently in tracts of a tandemly repeated nucleotide, and such slippage events can be genetically detected as frameshift mutations. In long mononucleotide runs, most frameshift intermediates are repaired by the postreplicative mismatch repair (MMR) machinery, rather than by the exonucleolytic proofreading activity of DNA polymerase. Although mononucleotide runs are hotspots for polymerase slippage events, it is not known whether the composition of a run and the surrounding context affect the frequency of slippage or the efficiency of MMR. To address these issues, 10-nucleotide (10N) runs were inserted into the yeast LYS2 gene to create +1 frameshift alleles. Slippage events within these runs were detected as Lys⁺ revertants. 10G or 10C runs were found to be more unstable than 10A or 10T runs, but neither the frequency of polymerase slippage nor the overall efficiency of MMR was greatly influenced by sequence context. Although complete elimination of MMR activity (msh2 mutants) affected all runs similarly, analyses of reversion rates in msh3 and msh6 mutants revealed distinct specificities of the yeast Msh2p-Msh3p and Msh2p-Msh6p mismatch binding complexes in the repair of frameshift intermediates in different sequence contexts.

TRACTS of simple repetitive DNA are hotspots for frameshift mutations, and this behavior has been attributed to the propensity of DNA polymerase to "slip" in repetitive sequences during DNA replication (STREISINGER et al. 1966 ; STREISINGER and OWEN 1985 ). Slippage is believed to occur by the transient dissociation of the nascent and template strands during DNA synthesis, followed by out-of-register annealing of the two strands and subsequent extension of the misaligned nascent strand. The unpaired nucleotide(s) in slippage intermediates can be repaired either by the exonucleolytic proofreading activity of DNA polymerase or by the postreplicative mismatch repair (MMR) machinery. If a slippage event is not repaired before the next round of DNA replication, it will give rise to a deletion if the unpaired nucleotide is on the template strand or to an insertion if the unpaired nucleotide is on the nascent strand.

In mononucleotide runs, the frequency of DNA polymerase slippage is positively correlated with the length of the run (KROUTIL et al. 1996 ; TRAN et al. 1997 ), but there is little information about how the sequence of a run (run composition) affects polymerase slippage. In the Escherichia coli lacZ gene, loss of a C from an 8C run occurred at a 10-fold higher rate than loss of an A from an 8A run (SAGHER et al. 1999 ). Although this suggests that G or C runs may be inherently more unstable than A or T runs, it should be noted that the runs were not at identical positions within the lacZ gene. Previous studies in yeast have examined slippage within an 18G run (SIA et al. 1997 ) or within poly(A) tracts varying in length from 4 to 14 nucleotides (TRAN et al. 1997 ). Given the different strain backgrounds and the different assay systems used in these studies, however, it has not been possible to deduce the effect that run composition may have on polymerase slippage in yeast.

Although the frequency of DNA polymerase slippage is positively correlated with mononucleotide run length, the efficiency of removing the corresponding frameshift intermediate via DNA polymerase proofreading activity is inversely related to the length of the run (KROUTIL et al. 1996 ; TRAN et al. 1997 ). As run length increases, it is thought that the extrahelical loop formed when the nascent and template strands reanneal out of register can migrate far enough away from the elongating polymerase to escape detection. In contrast to the inverse length dependence exhibited by proofreading, correction of DNA slippage events by the yeast MMR system appears to become more efficient as run length increases (TRAN et al. 1997 ).

The yeast MMR system can be completely inactivated by eliminating Msh2p, the central homolog of the E. coli MutS mismatch-binding protein. Msh2p functions in heterodimeric complexes with the MutS homologs Msh3p and Msh6p (reviewed in CROUSE 1998 ; KOLODNER and MARSISCHKY 1999 ), and both complexes are capable of initiating the repair of +1 and -1 frameshift intermediates (MARSISCHKY et al. 1996 ; GREENE and JINKS-ROBERTSON 1997 ; SIA et al. 1997 ; HARFE and JINKS-ROBERTSON 1999 ). Although examination of frameshift spectra derived for nonrepetitive DNA indicates that the Msh2p-Msh3p and Msh2p-Msh6p complexes have distinct recognition specificities (GREENE and JINKS-ROBERTSON 1997 ; FLORES-ROZAS and KOLODNER 1998 ; HARFE and JINKS-ROBERTSON 1999 ), the recognition specificities of these complexes for frameshift intermediates in mononucleotide runs of different composition have not been examined. For an 18G run, elimination of either Msh3p or Msh6p produced a modest increase in frameshift events (130-fold and 30-fold, respectively), whereas there was a strong, 5000-fold increase when both were eliminated (SIA et al. 1997 ). Such synergism indicates competing repair specificities for the Msh2p-Msh3p and Msh2p-Msh6p mismatch-binding complexes, with one complex being able to compensate for loss of the other complex. In addition to the MutS homologs, the repair of frameshift intermediates also requires the yeast Pms1p and Mlh1p proteins, which form a heterodimeric complex and are homologs of the E. coli MutL protein (PROLLA et al. 1994A , PROLLA et al. 1994B ). Mlh1p also can form heterodimeric complexes with Mlh2p and Mlh3p (WANG et al. 1999 ), but these complexes have only minor roles in correcting frameshift intermediates (FLORES-ROZAS and KOLODNER 1998 ; HARFE et al. 2000 ).

In addition to considering the composition of a mononucleotide run, the sequences flanking a run can potentially affect the frequency of polymerase slippage. This would be the case, for example, if a base misincorporation preceded the slippage event, with slippage restoring proper base pairing of the 3' end of the nascent strand with the template (see BEBENEK and KUNKEL 1990 ). In addition to context effects on polymerase slippage, sequence context might also impact mismatch recognition by the Msh2p-Msh3p or Msh2p-Msh6p complex. For the human MSH2-MSH6 complex, binding to a frameshift intermediate in vitro was influenced by the immediate flanking sequence (MACPHERSON et al. 1998 ). Similarly, in vitro binding of the yeast Msh2p-Msh6p complex to base substitution intermediates was impacted by sequence context (MARSISCHKY and KOLODNER 1999 ). To date, in vivo studies have not addressed the impact of sequence context either on polymerase slippage or on the repair of frameshift intermediates in mononucleotide runs.

To address the role of sequence composition and context on mononucleotide run stability, we constructed yeast strains in which runs of 10 guanines, 10 cytosines, 10 thymines, or 10 adenines (10G, 10C, 10T, or 10A runs, respectively) were engineered at a defined position in the LYS2 gene. The resulting +1 frameshift alleles reverted predominantly by loss of a single nucleotide from the 10N runs, thus allowing tract stability to be assessed by measuring the rate of Lys⁺ revertants. In addition to varying the run composition, the contexts of the runs were altered either by changing the identities of flanking nucleotides or by changing the position of the runs. The impact of mononucleotide run composition and context on the rate of DNA polymerase slippage and the efficiency of MMR were determined by measuring reversion rates in wild-type cells and in cells defective in various components of the yeast MMR machinery.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Media and growth conditions:
Yeast strains were grown nonselectively in YEP medium (1% yeast extract, 2% Bacto-peptone; 2.5% agar for plates) supplemented with either 2% dextrose (YEPD) or 2% glycerol/2% ethanol (YEPGE). Synthetic complete (SC) medium (SHERMAN 1991 ) containing 2% dextrose and lacking the appropriate amino acid was used for selective growth. Lys^- segregants were selected on minimal medium containing -aminoadipate as the primary nitrogen source (CHATTOO et al. 1979 ). All yeast strains were grown at 30°.

Yeast strain constructions:
All strains used were isogenic derivatives of SJR195 (MAT ade2-101 his3200 ura3Nco; HARFE and JINKS-ROBERTSON 1999 ). MSH2, MSH3, MSH6, MLH1, and PMS1 genes were disrupted as described previously (GREENE and JINKS-ROBERTSON 1997 ). 10N runs were introduced into the LYS2 locus using a two-step allele replacement procedure (ROTHSTEIN 1991 ). In the first step, the relevant plasmid was linearized within the lys2 sequence, and cells that had integrated the plasmid were selected as His⁺ transformants. Plasmid excision events subsequently were identified by selecting Lys^- segregants on -aminoadipate medium and then screening for simultaneous acquisition of a His^- phenotype.

Creation of mononucleotide runs in the LYS2 gene:
The 6A tract beginning at nucleotide 664 of the LYS2 coding strand (nucleotides are numbered from the upstream XbaI site) was replaced with 10N mononucleotide runs using the Chameleon mutagenesis system (Stratagene, La Jolla, CA) and plasmid pSR531 (HARFE and JINKS-ROBERTSON 1999 ) as the template. pSR531 is a nonreplicative plasmid containing HIS3 as a selectable marker and an internal fragment of LYS2. The 10N runs at position 664 are referred to as "position 1" mutations and the resulting alleles will be referred to as lys2-10N-1 alleles. The following oligonucleotides were used for the mutagenesis:

• 10C-1 run: 5'-GACGAGCTAGCTGCCCCCCCCCCTTCAAAGTTGCC

• 10G-1 run: 5'-GACGAGCTAGCTCGGGGGGGGGGTTCAAAGTTGCC

• 10T-1 run: 5'-GACGAGCTAGCTGTTTTTTTTTTGCCAAAGTTGCC

• 10A-1 run: 5'-GACGAGCTAGCTGAAAAAAAAAATTCAAAGTTGCC.

Plasmids containing lys2-10N-1 mutations are as follows: pSR607 (lys2-10A-1), pSR608 (lys2-10G-1), pSR609 (lys2-10T-1), and pSR619 (lys2-10C-1). The mutations were introduced into strain SJR195 to create yeast strains SJR938 (lys2-10G-1), SJR939 (lys2-10A-1), SJR980 (lys2-10T-1), and SJR981 (lys2-10C-1).

To place 10N runs at "position 2" (a unique PflMI site located at nucleotide 686), two complementary oligonucleotides containing PflMI compatible ends were annealed and directionally cloned into the PflMI site of pSR531. The resulting alleles are referred to as lys2-10N-2x alleles, with "x" indicating flanking sequence differences between different position 2 alleles (see Fig 1). The following pairs of oligonucleotides were annealed for the mutagenesis:

• 10C-2a run: 5'-ACCCCCCCCCCACGAT and 5'-GTGGGGGGGGGGTATC

  
  

  View larger version (22K): In this window In a new window Download PPT slide   Figure 1. Position of mononucleotide runs within the LYS2 gene and sequence of nucleotides flanking the runs. The LYS2 coding sequence (6A run) replaced by the 10N runs at position 1 is shown at the top. The position 2 10N runs were introduced at the indicated PflMI site (5'-CCANNNNNTGG; underlined) using appropriate oligonucleotides. When annealed, the oligonucleotides yielded duplex molecules with 3' overhangs complementary to those produced by PflMI digestion. In addition, each 10N run was flanked by one additional nt on the 5' side and two additional nt on the 3' side so that the immediately flanking sequences could be altered. Insertion of the 10N runs at site 2 thus was accompanied by an insertion of three additional nt, and a duplication of the GAT sequence of the PflMI site. At the bottom are shown the sequences of the lys2-10N alleles. Lowercase letters denote nt outside of the 10N runs that were altered. "N" denotes G, C, or T, depending on the allele.

• 10G-2a run: 5'-AGGGGGGGGGGACGAT and 5'-GTCCCCCCCCCCTATC

• 10T-2a run: 5'-ATTTTTTTTTTACGAT and 5'-GTAAAAAAAAAATATC

• 10C-2b run: 5'-ACCCCCCCCCCTTGAT and 5'-AAGGGGGGGGGGTATC

• 10G-2b run: 5'-AGGGGGGGGGGTTGAT and 5'-AACCCCCCCCCCTATC

• 10C-2c run: 5'-GCCCCCCCCCCTTGAT and 5'-AAGGGGGGGGGGCATC

• 10G-2c run: 5'-CGGGGGGGGGGTTGAT and 5'-AACCCCCCCCCCGATC.

Plasmids containing lys2-10N-2x mutations are as follows: pSR646 (lys2-10G-2a), pSR647 (lys2-10T-2a), pSR648 (lys2-10C-2a), pSR649 (lys2-10C-2b), pSR650 (lys2-10G-2b), pSR651 (lys2-10C-2c), and pSR652 (lys2-10G-2c). The mutations were introduced into strain SJR195 by two-step replacement to create yeast strains SJR1047 (lys2-10G-2a), SJR1045 (lys2-10T-2a), SJR1046 (lys2-10C-2a), SJR1093 (lys2-10G-2b), SJR1092 (lys2-10C-2b), SJR1095 (lys2-10G-2c), and SJR1094 (lys2-10C-2c). The lys2-10N-2a alleles were used in a previous study to demonstrate MMR activity of the yeast Mlh2 and Mlh3 proteins (HARFE et al. 2000 ).

Mutation rates and spectra:
Reversion rates to lysine prototrophy were determined by the method of the median (LEA and COULSON 1949 ) using data derived from 10–24 cultures of each strain. It should be noted that two independent isolates of each strain were used to generate these data. For the rate measurements, 5 ml of YEPGE medium were inoculated with single colonies from YEPD plates and the cultures were incubated for 2 days on a roller drum. Cells were harvested by centrifugation, washed with sterile H₂O, and resuspended in 1 ml of H₂O. A total of 100 µl aliquots of appropriate dilutions were plated onto SC-lys to identify Lys⁺ revertants and on YEPD to determine viable cell numbers. Lys⁺ colonies were counted 2 days after selective plating. The 95% confidence interval (CI) for each reversion rate was calculated using the 95% CI for the corresponding experimentally determined median (DIXON and MASSEY 1969 ).

To isolate independent Lys⁺ revertants for DNA sequence analysis, YEPGE cultures were grown as above and plated on SC-lys. To ensure independence, only one revertant from each culture was purified for subsequent molecular analysis. Automated DNA sequence analysis of PCR-amplified genomic fragments was performed as described previously (CHEN and JINKS-ROBERTSON 1998 ) using primer 5'-CGCAACAATGGTTACTCT.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We and others have used the LYS2 locus as an assay system for studying frameshift mutagenesis in yeast (GREENE and JINKS-ROBERTSON 1997 ; TRAN et al. 1997 ; HARFE and JINKS-ROBERTSON 1999 ). In general, these assays have monitored the reversion of lys2 alleles that contain net +1 or -1 frameshift mutations. Compensatory frameshift mutations that restore the correct reading frame can, in principle, occur throughout a reversion window defined by stop codons in the alternative, incorrect reading frames. Our previous studies have focused on the characterization of net +1 and -1 frameshift events that revert the lys2A746 and lys2Bgl frameshift alleles, respectively (GREENE and JINKS-ROBERTSON 1997 ; HARFE and JINKS-ROBERTSON 1999 ). These alleles have overlapping reversion windows near the 5' end of the LYS2 coding sequence that correspond to a functionally dispensable portion of Lys2p, thus allowing a wide variety of different types of frameshift events to be identified. To study the generation and repair of frameshift intermediates that occur in repetitive sequence, TRAN et al. 1997 introduced out-of-frame poly(A) runs of varying length into a different, centrally located region of LYS2. This latter analysis demonstrated that most compensatory frameshift events occurred within the run itself when run length was greater than eight nucleotides.

To examine the effects of nucleotide composition on mononucleotide run stability, we engineered 10N runs into the region of LYS2 that corresponds to the reversion window for the lys2A746 and lys2Bgl frameshift alleles. The 10N runs generate net +1 frameshift mutations, which are most easily "reverted" by deletion of a single base pair. Furthermore, the runs are sufficiently long that most, if not all, reversion events should occur in the runs. DNA sequence analysis confirmed this prediction for all runs used (data not shown), thus validating the assumption that Lys⁺ reversion rates accurately reflect the -1 mutation rates in the 10N runs. To examine the impact of sequence context on mononucleotide run stability, runs of varying composition were introduced at two different positions within the reversion window (see Fig 1). First, a 6A run was replaced with 10N runs, resulting in frameshift alleles referred to as "position 1" lys2-10N-1 alleles. Second, 10N runs were introduced 18 bp downstream of the 6A sequence, thus changing the local context of the runs. The resulting frameshift alleles are referred to as "position 2" lys2-10N-2x alleles. Finally, to examine the effects of flanking sequences on run stabilities, the nucleotides immediately flanking the position 2 runs also were varied (e.g., lys2-10N-2a and lys2-10N-2b).

The yeast MMR system corrects the majority of polymerase slippage events that occur at long mononucleotide runs (TRAN et al. 1997 ). To examine the contributions of individual MMR proteins in the correction of frameshift intermediates in the 10N runs, we individually eliminated yeast MutS (Msh2p, Msh3p, and Msh6p) and MutL (Pms1p and Mlh1p) homologs. Reversion rates of the lys2-10N alleles were measured in both wild-type strains and in msh2, msh3, msh6, pms1, and mlh1 mutant strains. All reversion rate data are presented in Table 1.

Reversion rates of lys2 alleles containing 10N runs at position 1:
In a wild-type background, the reversion rates of the lys2-10N-1 alleles differed over an 20-fold range. The lys2-10A-1 and lys2-10T-1 alleles exhibited the lowest reversion rates (0.9 x 10^-8 and 1.4 x 10^-8, respectively) and the lys2-10G-1 allele the highest reversion rate (1.7 x 10^-7). In the msh2, pms1, and mlh1 strains, the reversion rates were elevated 4,500- to 66,000-fold. Although the reversion rates of any given allele were similar in the three mutant backgrounds, the alleles containing the 10G or 10C runs reverted at a 5- to 10-fold higher rate than the alleles containing the 10A or 10T runs.

Individual disruption of either MSH3 or MSH6 increased the reversion rate of each lys2-10N-1 allele examined, but the reversion rates of different alleles varied over a very broad range. In an msh3 strain, for example, the 10G and 10C runs were much more unstable than the 10A and 10T runs, with the reversion rate of the lys2-10G-1 allele being 340-fold higher than that of the lys2-10T-1 allele (1 x 10^-4 vs. 2.9 x 10^-7, respectively). Furthermore, the 10G run was 10-fold more unstable than the 10C run in an msh3 background. In msh6 strains, the 10G and 10C runs were even more dissimilar, with the reversion rates of the respective lys2-10N-1 alleles differing by more than 200-fold (6.6 x 10^-5 and 2.8 x 10^-7 for the lys2-10G-1 and lys2-10C-1 alleles, respectively). The stability of the 10T run was similar to that of the 10C run in the msh6 strains and these runs each were 10-fold more unstable than the 10A run. Simultaneous disruption of MSH3 and MSH6 resulted in a synergistic increase in mutation rates relative to the single disruptions. As expected, the msh3 msh6 double mutants exhibited mutation rates comparable to those seen in the msh2 single mutants (data not shown).

Reversion rates of lys2 alleles containing 10N runs at position 2:
In the position 1 10N stability assays, the DNA sequences flanking the 10N runs were not identical (Fig 1). Specifically, during the construction of the 10G run, the nucleotide immediately preceding the run was changed from G to C to maintain a constant run length while varying the run composition. Similarly, for the 10T run, the TT dinucleotide immediately following the run was changed to GC. To investigate the contribution of flanking sequence to polymerase slippage, three of the runs (10G, 10C, and 10T) were flanked by adenines and placed 18 bp downstream of position 1 (position 2a; see Fig 1). Placing the runs 18 bp away from position 1 allowed us to assay whether the sequence context of the runs affected polymerase slippage and/or MMR efficiency.

In a wild-type background, the lys2-10N-2a runs had stabilities that differed by approximately fourfold, and it should be noted that the lys2-10C-2a allele was slightly more unstable than the lys2-10G-2a, a reverse of the pattern seen at position 1. As at position 1, the reversion rate of the lys2-10T-2a allele was lower than that of the lys2-10G-2a and lys2-10C-2a alleles in the msh2 strains. Although there were subtle differences between the position 1 and position 2a reversion rates for a given 10N run in wild-type strains, sequence context had no significant effect on reversion rates in msh2 mutants (this also was true for the 2b and 2c runs; see below). In stark contrast, some runs of the same composition behaved very differently in an msh3 mutant and/or an msh6 mutant when the context was altered. Most notably, the 10C run was 50-fold more unstable at position 2a than at position 1 in either an msh3 or an msh6 mutant. The reverse pattern was seen for the corresponding 10G alleles in msh3 and msh6 mutants, with the position 1 allele being two- to eightfold more unstable than the position 2a allele. The lys2-10T-1 and lys2-10T-2a alleles had similar reversion rates in the msh3 and msh6 single mutants.

The very different stabilities of the 10G and 10C runs at positions 1 and 2a in the msh3 and msh6 single mutants could be attributed to the general locations of the runs and/or to the identities of the immediately flanking nucleotides. To address these possibilities, we modified the sequences flanking the position 2a 10G and 10C runs so that they were identical to the position 1 flanking sequences (position 2c runs; see Fig 1). This allowed us to compare the stabilities of 10C and 10G runs having identical 5' and 3' flanking sequences but slightly different positions within LYS2. Finally, strains were made in which the 10G and 10C runs at position 2a were modified so that 3' flanks were identical to those at position 1 (TT dinucleotide) but the 5' flanks were different (position 2b runs; see Fig 1).

As noted previously, context appeared to have no effect on run stabilities in msh2 strains and only small effects on the stabilities of the 10G and 10C runs in wild-type strains. In wild-type strains, the lys2-10G-1 and lys2-10G-2c alleles, which have identical flanking sequences, were more unstable than the lys2-10G-2a and lys2-10G-2b alleles. For the 10C runs in wild-type strains, the stabilities of the lys2-10C-1, lys2-10C-2b, and lys2-10C-2c alleles, all of which have a TT dinucleotide immediately 3' to the run, were greater than that of the lys2-10C-2a allele. In the msh3 and msh6 strains, there were very dramatic context-dependent differences in the lys2-10N reversion rates. First, in an msh3 background the lys2-10C-2a allele was 50-fold more unstable than the lys2-10C-1, lys2-10C-2b, or lys2-10C-2c allele, each of which had a comparable stability. As note previously, the 10C runs in contexts 1, 2b, and 2c have in common a 3' flanking TT dinucleotide, which is changed to AC in context 2a. Second, in an msh6 mutant background the lys2-10C-1 allele was 50-fold more stable than the lys2-10C-2a, lys2-10C-2b, or lys2-10C-2c alleles. The stabilities of the lys2-10G alleles also exhibited context effects in the msh3 and msh6 single mutants, but the mutation rate variations generally were not as dramatic as those observed with the lys2-10C alleles.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Mononucleotide runs are found frequently within the coding regions of eukaryotic genes. In the yeast genome, for example, 25% of genes contain a poly(A) or poly(T) tract that is at least eight nucleotides in length (TRAN et al. 1997 ). Because polymerase slippage events occur at an increased rate in mononucleotide runs, it is important to understand how the nucleotide composition and sequence context of a run affect the rate of DNA polymerase slippage. In addition to affecting the rate of polymerase slippage, run composition and context may also affect the efficiency with which the slippage errors are removed by the MMR machinery. To address these issues, we developed an in vivo reversion assay that allows us to measure the rate of -1 frameshift events within defined 10N mononucleotide runs located in the yeast LYS2 gene (lys2-10N alleles). Both the composition of the runs and the sequences flanking the runs were altered, allowing us to examine the effects of these changes on mononucleotide run stability. To facilitate comparisons relevant to the points discussed below, the reversion rates of the lys2-10N alleles in wild-type and in completely MMR-defective (msh2) strains are presented in Fig 2. Fig 3 presents the reversion rates of the lys2-10N alleles in msh2, msh3, and msh6 strains normalized to the corresponding rates obtained in wild-type strains.

View larger version (42K): In this window In a new window Download PPT slide   Figure 2. Reversion rates of lys2-10N alleles in wild-type (open bars) and msh2 (hatched bars) strains. The 95% CI for each rate is indicated.

View larger version (40K): In this window In a new window Download PPT slide   Figure 3. Lys+ reversion rates in msh2, msh3, and msh6 mutants. For each lys2-10N allele, the mutation rates have been normalized to the corresponding rate obtained in a wild-type strain.

As shown in Fig 2, sequence context had only small (at most fourfold) effects on the stabilities of the 10G, 10C, or 10T runs in MMR competent cells. Because these subtle effects were not evident in completely MMR-defective (msh2) cells, it can be concluded that the overall efficiency of the MMR system is influenced by the sequence context of the mutational intermediate. The context of the 10A run was not altered, but we would expect to obtain similar results. The data from the msh2 strains suggest that most of the -1 frameshift mutations resulted from simple slippage within the runs rather than slippage initiated by base misincorporation events adjacent to the runs. In the latter case, one would expect for slippage rates to exhibit context dependence. We consistently observed a higher mutation rate in the 10G and 10C runs than in the 10T and 10A runs in both wild-type and msh2 strains. One interpretation of this observation is that G and C runs are inherently more unstable than A and T runs, which would be consistent with observations made in E. coli (SAGHER et al. 1999 ). Although one might expect that slippage would occur less often in G/C tracts than in A/T tracts because of the number of hydrogen bonds that must be disrupted, a consideration of base stacking interactions suggests that a slippage intermediate in a G/C run should be more stable than one in an A/T run (see SAGHER et al. 1999 ). In addition, slippage intermediates in G or C runs would be expected to be less efficiently proofread than those in A or T runs because of the relative difficulty of disrupting GC base pairs (BESSMAN and REHA-KRANTZ 1977 ; GOODMAN and FYGENSON 1998 ). It should be noted that the exonucleolytic proofreading activities of the yeast replicative polymerases were present in our studies, so the stability differences between the G/C tracts vs. the A/T tracts could reflect different proofreading efficiencies of -1 frameshift intermediates in A/T vs. G/C runs.

The MMR proteins Msh3p and Msh6p interact with Msh2p to form Msh2p-Msh3p and Msh2p-Msh6p mismatch-recognition complexes (KOLODNER and MARSISCHKY 1999 ). These complexes have overlapping recognition specificities for +1 and -1 frameshift intermediates (MARSISCHKY et al. 1996 ; GREENE and JINKS-ROBERTSON 1997 ; SIA et al. 1997 ; HARFE and JINKS-ROBERTSON 1999 ), and these overlaps are manifested as strong synergistic effects on mutation rates when both complexes are eliminated (i.e., an msh3 msh6 double mutant or an msh2 single mutant). It should be noted that the reversion rate in an msh3 mutant reflects the repair capacity of the remaining Msh2p-Msh6p complex, while the reversion rate in an msh6 mutant indicates the repair capacity of the Msh2p-Msh3p complex. The relative abilities of Msh2p-Msh3p and Msh2p-Msh6p complexes to initiate repair of DNA polymerase slippage errors thus can be assessed by comparing reversion rates of a given allele in msh6 and msh3 mutants, respectively. If the reversion rate of a given allele is similar in the msh3 and msh6 single mutants, then one could argue that the Msh2p-Msh3p and Msh2p-Msh6p complexes have similar affinities for the corresponding frameshift intermediate. If the reversion rate of a given lys2-10N allele differs dramatically in the two single mutants, however, then one would conclude that the complexes have different recognition specificities for the corresponding frameshift intermediate. As shown in Fig 3, mutation rates for the lys2-10C and lys2-10G alleles generally were elevated to a much greater extent in the single mutants than were the reversion rates for the lys2-10A and lys2-10T alleles. It thus appears that the Msh2p-Msh3p and Msh2p-Msh6p complexes may compete more effectively for the repair of frameshift intermediates in A and T runs than in G and C runs. There did not seem to be general specificity of a given complex for an A or T vs. a G or C run.

The Msh2p-Msh3p and Msh2p-Msh6p complexes did not exhibit any striking repair biases in terms of run composition, but we found that the context of a run could dramatically affect the repair of frameshift intermediates. This was particularly evident when examining the reversion rates of the lys2-10C alleles in the mutant strains relative to the corresponding wild-type strains (Fig 3). In an msh6 background, for example, the mutation rates of the three lys2-10C-2x alleles were 50-fold higher than that of the lys2-10C-1 allele. Thus the Msh2p-Msh3p complex appears to recognize frameshift intermediates much better when the 10C run is at position 1 (lys2-10C-1 allele) than when the run is at position 2 (lys2-10C-2x alleles). In addition to this position-dependent context effect, an effect of the immediately flanking nucleotide (nt) on the repair efficiency of the Msh2p-Msh6p complex was evident when the stabilities of the 10C runs were examined in an msh3 background. In this case, the lys2-10C-2a allele had a much higher reversion rate than the lys2-10C-1, lys2-10C-2b, or lys2-10C-2c allele. The lys2-10C-2a allele differs from the other three alleles in that it is immediately flanked on the 3' side by an AC dinucleotide instead of a TT dinucleotide. It should be noted that the reversion rate of the lys2-10C-2a allele in an msh3 background was almost as high as in an msh2 background, indicating that the Msh2p-Msh6p complex has very little repair activity against the relevant frameshift intermediate.

Although context effects for the binding of Msh2p-Msh6p to defined mismatches in vitro have been reported (MACPHERSON et al. 1998 ; MARSISCHKY and KOLODNER 1999 ), the results presented here provide the first clear example of strong context effects on the repair efficiencies of the Msh2p-Msh3p and Msh2p-Msh6p complexes in vivo. Thus, the repair of a frameshift intermediate is not simply related to the identity of the extrahelical nucleotide(s), but also can be profoundly impacted by the nucleotides that flank the mutational intermediate, even when the intermediate is embedded in a highly repetitive sequence. Such context effects may not only be determined by the specificities of the Msh2p-Msh3p and Msh2p-Msh6p mismatch-binding complexes, but also could be related to whether the frameshift intermediate arises during leading strand vs. lagging strand DNA synthesis. In the case of the LYS2 gene, a DNA replication fork would be expected to enter from the 5' end (TRAN et al. 1995 ; FREUDENREICH et al. 1997 ), so that the coding strand would be the template for lagging strand synthesis. In addition, the context effects reported here could be related to the identity of the polymerase that generates the frameshift intermediate or even to the transcriptional state of the sequence (MOREY et al. 2000 ). Additional studies will be required to determine whether these additional factors influence MMR efficiency.

	ACKNOWLEDGMENTS

We gratefully acknowledge the contributions of R. McVoy in constructing some of the plasmids used in this work. We thank K. Hill and R. Spell for critically reading the manuscript. This work was funded by a grant from the National Science Foundation.

Manuscript received January 19, 2000; Accepted for publication June 19, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BEBENEK, K. and T. A. KUNKEL, 1990  Frameshift errors initiated by nucleotide misincorporation. Proc. Natl. Acad. Sci. USA 87:4946-4950[Abstract/Free Full Text].

BESSMAN, M. J. and J. REHA-KRANTZ, 1977  Studies on the biochemical basis of spontaneous mutation. J. Mol. Biol. 116:115-123[Medline].

CHATTOO, F. F., F. SHERMAN, D. A. AZUBALIS, T. A. FJELLSTEDT, and D. MEHNERT et al., 1979  Selection of lys2 mutants of the yeast Saccharomyces cerevisiae by the utilization of alpha-aminoadipate. Genetics 93:51-65[Abstract/Free Full Text].

CHEN, W. and S. JINKS-ROBERTSON, 1998  Mismatch repair proteins regulate heteroduplex formation during mitotic recombination in yeast. Mol. Cell. Biol. 18:6525-6537[Abstract/Free Full Text].

CROUSE, G. F., 1998 Mismatch repair systems in Saccharomyces cerevisiae, pp. 411–448 in DNA Damage and Repair, Volume 1: DNA Repair in Prokaryotes and Lower Eukaryotes, edited by J. A. NICKOLOFF and M. F. HOEKSTRA. Humana Press, Totowa, NJ.

DIXON, W. J., and F. J. MASSEY, JR., 1969 Introduction to Statistical Analysis. McGraw-Hill, New York.

FLORES-ROZAS, H. and R. D. KOLODNER, 1998  The Saccharomyces cerevisiae MLH3 gene functions in MSH3-dependent suppression of frameshift mutations. Proc. Natl. Acad. Sci. USA 95:12404-12409[Abstract/Free Full Text].

FREUDENREICH, C. H., J. B. STAVENHAGEN, and V. A. ZAKIAN, 1997  Stability of a CTG/CAG trinucleotide repeat in yeast is dependent on its orientation in the genome. Mol. Cell. Biol. 17:2090-2098[Abstract].

GOODMAN, M. F. and D. K. FYGENSON, 1998  DNA polymerase fidelity: from genetics toward a biochemical understanding. Genetics 148:1475-1482[Abstract/Free Full Text].

GREENE, C. N. and S. JINKS-ROBERTSON, 1997  Frameshift intermediates in homopolymer runs are removed efficiently by yeast mismatch repair proteins. Mol. Cell. Biol. 17:2844-2850[Abstract].

HARFE, B. D. and S. JINKS-ROBERTSON, 1999  Removal of frameshift intermediates by mismatch repair proteins in Saccharomyces cerevisiae.. Mol. Cell. Biol. 19:4766-4773[Abstract/Free Full Text].

HARFE, B. D., B. K. MINESINGER, and S. JINKS-ROBERTSON, 2000  Discrete in vivo roles for the MutL homologs Mlh2p and Mlh3p in the removal of frameshift intermediates in budding yeast. Curr. Biol. 10:145-148[Medline].

KOLODNER, R. D. and G. T. MARSISCHKY, 1999  Eukaryotic DNA mismatch repair. Curr. Opin. Genet. Dev. 9:89-96[Medline].

KROUTIL, L. C., K. REGISTER, K. BEBENEK, and T. A. KUNKEL, 1996  Exonucleolytic proofreading during replication of repetitive DNA. Biochemistry 35:1046-1053[Medline].

LEA, D. E. and C. A. COULSON, 1949  The distribution of the numbers of mutants in bacterial populations. J. Genet. 49:264-285.

MACPHERSON, P., O. HUMBERT, and P. KARRAN, 1998  Frameshift mismatch recognition by the human MutS complex. Mutat. Res. 408:55-66[Medline].

MARSISCHKY, G. T. and R. D. KOLODNER, 1999  Biochemical characterization of the interaction between the Saccharomyces cerevisiae MSH2-MSH6 complex and mispaired bases in DNA. J. Biol. Chem. 274:26668-26682[Abstract/Free Full Text].

MARSISCHKY, G. T., N. FILOSI, M. F. KANE, and R. KOLODNER, 1996  Redundancy of Saccharomyces cerevisiae MSH3 and MSH6 in MSH2-dependent mismatch repair. Genes Dev. 10:407-420[Abstract/Free Full Text].

MOREY, N. J., C. N. GREENE, and S. JINKS-ROBERTSON, 2000  Genetic analysis of transcription-associated mutation in Saccharomyces cerevisiae.. Genetics 154:109-120[Abstract/Free Full Text].

PROLLA, T. A., D.-M. CHRISTIE, and R. M. LISKAY, 1994a  Dual requirement in yeast DNA mismatch repair for MLH1 and PMS1, two homologs of the bacterial mutL gene. Mol. Cell. Biol. 14:407-415[Abstract/Free Full Text].

PROLLA, T. A., Q. PANG, E. ALANI, R. D. KOLODNER, and R. M. LISKAY, 1994b  MLH1, PMS1, and MSH2 interactions during the initiation of DNA mismatch repair in yeast. Science 265:1091-1093[Abstract/Free Full Text].

ROTHSTEIN, R., 1991  Targeting, disruption, replacement, and allele rescue: integrative DNA transformation in yeast. Methods Enzymol. 194:281-301[Medline].

SAGHER, D., A. HSU, and B. STRAUSS, 1999  Stabilization of the intermediate in frameshift mutation. Mutat. Res. 423:73-77[Medline].

SHERMAN, F., 1991  Getting started with yeast. Methods Enzymol. 194:3-20[Medline].

SIA, E. A., C. A. BUTLER, M. DOMINSKA, P. GREENWELL, and T. D. FOX et al., 1997  Microsatellite instability in yeast: dependence on repeat unit size and DNA mismatch repair genes. Mol. Cell. Biol. 17:2851-2858[Abstract].

STREISINGER, G. and J. OWEN, 1985  Mechanisms of spontaneous and induced frameshift mutation in bacteriophage T4. Genetics 109:633-659[Abstract/Free Full Text].

STREISINGER, G., Y. OKADA, J. EMRICH, J. NEWTON, and A. TSUGITA et al., 1966  Frameshift mutations and the genetic code. Cold Spring Harbor Symp. Quant. Biol. 31:77-84[Medline].

TRAN, H. T., N. P. DEGTYAREVA, N. N. KOLOTEVA, A. SUGINO, and H. MASUMOTO et al., 1995  Replication slippage between distant short repeats in Saccharomyces cerevisiae depends on the direction of replication and the RAD50 and RAD52 genes. Mol. Cell. Biol. 15:5607-5617[Abstract].

TRAN, H. T., J. D. KEEN, M. KRICKER, M. A. RESNICK, and D. A. GORDENIN, 1997  Hypermutability of homonucleotide runs in mismatch repair and DNA polymerase proofreading yeast mutants. Mol. Cell. Biol. 17:2859-2865[Abstract].

WANG, T.-F., N. KLECKNER, and N. HUNTER, 1999  Functional specificity of MutL homologs in yeast: evidence for three Mlh1-based heterocomplexes with distinct roles during meiosis in recombination and mismatch correction. Proc. Natl. Acad. Sci. USA 96:13914-13919[Abstract/Free Full Text].

This article has been cited by other articles:

				J. A. Heck, D. Gresham, D. Botstein, and E. Alani Accumulation of Recessive Lethal Mutations in Saccharomyces cerevisiae mlh1 Mismatch Repair Mutants Is Not Associated With Gross Chromosomal Rearrangements Genetics, September 1, 2006; 174(1): 519 - 523. [Abstract] [Full Text] [PDF]

				N. Erdeniz, S. Dudley, R. Gealy, S. Jinks-Robertson, and R. M. Liskay Novel PMS1 Alleles Preferentially Affect the Repair of Primer Strand Loops during DNA Replication Mol. Cell. Biol., November 1, 2005; 25(21): 9221 - 9231. [Abstract] [Full Text] [PDF]

				H. Gragg, B. D. Harfe, and S. Jinks-Robertson Base Composition of Mononucleotide Runs Affects DNA Polymerase Slippage and Removal of Frameshift Intermediates by Mismatch Repair in Saccharomyces cerevisiae Mol. Cell. Biol., December 15, 2002; 22(24): 8756 - 8762. [Abstract] [Full Text] [PDF]

				J. C. Boyer, N. A. Yamada, C. N. Roques, S. B. Hatch, K. Riess, and R. A. Farber Sequence dependent instability of mononucleotide microsatellites in cultured mismatch repair proficient and deficient mammalian cells Hum. Mol. Genet., March 1, 2002; 11(6): 707 - 713. [Abstract] [Full Text] [PDF]

				A. B. Clark, F. Valle, K. Drotschmann, R. K. Gary, and T. A. Kunkel Functional Interaction of Proliferating Cell Nuclear Antigen with MSH2-MSH6 and MSH2-MSH3 Complexes J. Biol. Chem., November 17, 2000; 275(47): 36498 - 36501. [Abstract] [Full Text] [PDF]