Abstract
The postreplicative mismatch repair (MMR) system is important for removing mutational intermediates that are generated during DNA replication, especially those that arise as a result of DNA polymerase slippage in simple repeats. Here, we use a forward mutation assay to systematically examine the accumulation of frameshift mutations within mononucleotide runs of variable composition in wild-type and MMR-defective yeast strains. These studies demonstrate that (i) DNA polymerase slippage occurs more often in 10-cytosine/10-guanine (10C/10G) runs than in 10-adenine/10-thymine (10A/10T) runs, (ii) the MMR system removes frameshift intermediates in 10A/10T runs more efficiently than in 10C/10G runs, (iii) the MMR system removes −1 frameshift intermediates more efficiently than +1 intermediates in all 10-nucleotide runs, and (iv) the repair specificities of the Msh2p-Msh3p and Msh2p-Msh6p mismatch recognition complexes with respect to 1-nucleotide insertion/deletion loops vary dramatically as a function of run composition. These observations are relevant to issues of genome stability, with both the rates and types of mutations that accumulate in mononucleotide runs being influenced by the primary sequence of the run as well as by the status of the MMR system.
Frameshift mutations in coding sequences are caused by the insertion or deletion of base pairs in non-multiples of three and generally result in complete inactivation of the encoded protein. Tandemly repeated sequences have long been recognized as hot spots for frameshift mutations, and maintaining the stability of such repeats presents a formidable challenge during DNA replication (for a review, see reference 39). The instability of tandem repeats has been attributed to DNA polymerase slippage, which involves a transient dissociation of the template and the 3′ end of the primer strand, followed by an out-of-register annealing within the repeat tract (40, 41). The result is a frameshift intermediate with an extrahelical loop comprised of one or more repeat units and stabilized by correct base pairing between flanking repeats. If not corrected before the next round of DNA replication, such an intermediate will yield a deletion event if the extrahelical loop is on the template strand or an insertion event if the loop is on the primer strand (Fig. 1). Both in vivo and in vitro studies have shown that the frequency of polymerase slippage increases as the number of repeat units in a run increases, with the most frequent types of corresponding frameshift mutations being single-unit deletions or insertions (22, 35, 42).
FIG. 1.
Generation of frameshift mutations by template-primer slippage during DNA replication.
Cells have two sequential means of correcting misaligned frameshift intermediates: 3′-to-5′ exonucleolytic proofreading by DNA polymerase and postreplicative mismatch repair (MMR). Although both mechanisms are effective in removing frameshift intermediates in short mononucleotide runs, the efficiency of proofreading decreases as run length increases (22, 42), so that MMR becomes the predominant mechanism for removing frameshift intermediates in long runs (33, 38, 42). The importance of the MMR system in promoting genetic stability is particularly evident in hereditary nonpolyposis colon cancer, where MMR defects are associated with simple repeat instability and lead to the accumulation of mutations that contribute to tumor formation (reviewed in reference 4).
The most extensively characterized mismatch repair system is the methyl-directed MutHLS system of the bacterium Escherichia coli (for a review, see reference 29). Following binding of a MutS homodimer to a mismatch, a MutL homodimer interacts with MutS as well as with the endonuclease MutH, thereby activating the MutH protein and initiating removal of the newly synthesized strand. Eukaryotes possess multiple homologs of the E. coli MutS and MutL proteins, with the functional form of each being a heterodimer instead of a homodimer (reviewed in reference 16). In the yeast Saccharomyces cerevisiae, there are six MutS homologs (Msh1p to Msh6p) and four MutL homologs (Pms1p and Mlh1p to Mlh3p). Mismatch recognition in nuclear replication intermediates is effected by Msh2p-Msh3p or Msh2p-Msh6p heterodimers, with the former recognizing only insertion/deletion loops (IDLs) and the latter recognizing base-base mismatches as well as small IDLs (21, 27). Of particular relevance to the work reported here, either complex can initiate repair of −1 and +1 frameshift intermediates (15, 17, 27, 35). The MutL homolog Mlh1p can form heterodimers with Pms1p, Mlh2p, or Mlh3p (45), but because most MMR involves the Mlh1p-Pms1p complex (11, 19), only this complex will be considered.
The activity of the yeast MMR system in removing frameshift intermediates in tandem repeats has been studied by using artificial repeat tracts that either disrupt or maintain the reading frame of a reporter gene (e.g., see references 18, 35, 38, and 42). Whereas an in-frame forward mutation system detects any insertion/deletion that alters the reading frame of the reporter, a given out-of-frame reversion assay typically detects either insertions or deletions, but not both. We previously used a LYS2-based reversion assay to examine the effect of sequence composition and context on the stabilities of 10-nucleotide (nt) runs in strains defective in individual MMR components (18). Although this system was limited by its ability to detect only −1 events, it nevertheless revealed distinct sequence-related specificities of the Msh2p-Msh3p and Msh2-Msh6p complexes in removing the corresponding frameshift intermediates. In the present study, we have modified the 10-nt assay system so that the mononucleotide runs are in frame with the LYS2 coding sequence, thus allowing a direct comparison of the rates and distributions of +1 versus −1 frameshift events.
MATERIALS AND METHODS
Media and growth conditions.
Yeast strains were grown nonselectively on YEPD medium (1% yeast extract, 2% Bacto-peptone, 2% dextrose, 2% agar for plates) and selectively on synthetic dextrose complete (SC) medium (34) lacking the appropriate amino acid. Synthetic medium was supplemented with α-aminoadipate (αAA) to select Lys− colonies (5). Growth was at 30°C in all experiments.
Plasmid and strain constructions.
Ten-nucleotide runs were introduced into the wild-type LYS2 gene with plasmid pSR531 (18) as a template in the Chameleon site-directed mutagenesis system (Stratagene, La Jolla, Calif.). A six-adenine (6A) run (coding strand nt 368 to 373 relative to the start codon) was replaced with a 10A (pSR607), 10-thymine (10T) (pSR609), 10-cytosine (10C) (pSR619), or 10-guanine (10G) (pSR608) run by using the primers 5′-GACGAGCTAGCTGAAAAAAAAAATTCAAAGTTGCC, 5′-GACGAGCTAGCTGTTTTTTTTTTGCCAAAGTTGCC, 5′-GACGAGCTAGCTGCCCCCCCCCCTTCAAAGTTGCC, or 5′-GACGAGCTAGCTCGGGGGGGGGGTTCAAAGTTGCC, respectively. A 1.6-kb PflMI-PstI fragment of pSR607 and a 4.1-kb PflMI-PstI fragment of pSR585 (17) were ligated to yield plasmid pSR685, containing both a 10A run and the lys2ΔA746 mutation. It should be noted that the position of the lys2ΔA746 frameshift mutation was originally defined with respect to an XbaI site upstream of the LYS2 coding sequence (see reference 17). With respect to the translation start codon, the mutation is at position 450.
The 10-nt runs were introduced into the chromosomal LYS2 locus by transforming strain SJR922 (MATα ade2-101oc his3Δ200 ura3ΔNco lys2ΔA746) with EcoRV-PflMI fragments of plasmid pSR608, pSR609, or pSR619 or an EcoRV-NruI fragment of pSR685. The addition of 4 bp (conferred by the 10-nt runs) to the resident −1 allele (lys2ΔA746) restored the correct reading frame of the LYS2 gene, thus resulting in a selectable Lys+ phenotype. MSH2, MSH3, MSH6, and MLH1 were disrupted by using AatII-XbaI-digested GC1914 (15), EcoRI-digested pEN33 (7), SacII-EcoRI-digested pSR504 (15), and SacI-BamHI-digested mlh1Δ::URA3 (32), respectively. The pms1Δ allele was introduced by two-step allele replacement with BstXI-digested pSR211 (7). All mutant alleles were confirmed by PCR.
Mutation analyses.
Independent Lys− mutants for sequence analysis were obtained by patching individual colonies from YEPD medium onto αAA medium. The relevant region of the LYS2 gene was PCR amplified from total genomic DNA and sequenced as described previously (18). Rates of mutations resulting in a Lys− phenotype were determined by the method of the median (25). Single colonies grown for 2 days on YEPD medium were used to inoculate 1-ml YEPD cultures, and these were grown overnight on a roller drum. Cells were washed with H2O, and aliquots of the appropriate dilutions were plated on αAA plates to identify Lys− mutants and on YEPD plates to determine viable cell numbers. In order to provide a uniform cell density on the αAA plates and thereby equalize the residual growth of the dilutions used to determine Lys− colony numbers, approximately 5 × 106 cells of a LYS2/LYS2 diploid strain (MATα/MATa can1/CAN1 ade2-101oc/ADE2 HIS3/his3Δ1 his4-619/HIS4 LEU2/leu2-3,112 URA3/ura3-52 SUC2/suc2°) were plated on each αAA plate. Colonies were counted 2 or 7 days after plating on YEPD or αAA medium, respectively.
RESULTS
Previous work by Tran et al. (42) demonstrated that a 9A run within the yeast LYS2 gene is sufficiently “hot” that mutations in the run comprise approximately 10 and 80% of lys2 forward mutations in wild-type and MMR-defective strains, respectively. In order to obtain a complete spectrum of mutations occurring in mononucleotide runs of variable composition, we introduced 10A, 10T, 10C, or 10G runs at a defined position within the chromosomal LYS2 locus in a manner that maintains the correct reading frame of the gene (see Materials and Methods for details). Forward mutations resulting in a Lys− phenotype were selected on medium containing αAA (5), and the rates of Lys− mutants were measured in wild-type, pms1Δ, mlh1Δ, msh2Δ, msh3Δ, and msh6Δ strains containing each of the 10-nt runs. In addition, the types and proportions of mutational events within each 10-nt run were determined by sequencing the region containing the run in independent Lys− mutants. The rate and sequencing data were used to calculate the total mutation rate, the +1 mutation rate and the −1 mutation rate within each of the 10-nt runs in each strain background (Tables 1 to 3).
TABLE 1.
Mutations within 10-nt runs in wild-type strainsa
Run | Strain | Total mutations
|
+1 Frameshifts
|
−1 Frameshifts
|
|||
---|---|---|---|---|---|---|---|
Fraction | Rate (10−8) | Fraction | Rate (10−8) | Fraction | Rate (10−8) | ||
10A | SJR1246 | 22/89 | 2.36 | 13/89 | 1.40 | 9/89 | 0.97 |
10T | SJR1245 | 31/81 | 4.27 | 30/81 | 4.14 | 1/81 | 0.14 |
10C | SJR1244 | 40/48 | 341 | 38/48 | 324 | 2/48 | 17.0 |
10G | SJR1243 | 48/49 | 1,050 | 48/49 | 1,050 | 0/49 | ND |
The fraction of the Lys− mutational events that occurred within each run, as well as the fractions of +1 and −1 frameshift events within each run, is given. The denominator of each fraction corresponds to the number of independent Lys− mutants sequenced. Rates (number of mutations per cell per generation) were calculated by multiplying the total Lys− mutation rate by the fraction of the relevant event. Each rate was derived from 12 cultures. ND, not detectable.
TABLE 3.
Mutations within 10-nt runs in msh3Δ and msh6Δ strainsa
Genotype | Run | Strain | Total
|
+1 Frameshifts
|
−1 Frameshifts
|
|||
---|---|---|---|---|---|---|---|---|
Fraction | Rate (10−6) | Fraction | Rate (10−6) | Fraction | Rate (10−6) | |||
msh3Δ | 10A | SJR1276 | 49/117 | 1.72 (59) | 5/117 | 0.18 | 43/117 | 1.51 |
10T | SJR1277 | 25/110 | 0.67 (15) | 4/110 | 0.11 | 21/110 | 0.56 | |
10C | SJR1247 | 48/51b | 63.4 (19) | 7/51 | 9.25 | 40/51 | 52.8 | |
10G | SJR1248 | 47/47 | 810 (77) | 46/47 | 793 | 1/47 | 17.2 | |
msh6Δ | 10A | SJR1251 | 35/106 | 0.22 (7.4) | 28/106 | 0.18 | 7/106 | 0.04 |
10T | SJR1350 | 35/81 | 4.0 (90) | 15/81 | 1.71 | 20/81 | 2.28 | |
10C | SJR1349 | 42/48 | 103 (30) | 37/48 | 90.3 | 5/48 | 12.2 | |
10G | SJR1351 | 43/43 | 163 (15) | 3/43 | 11.3 | 40/43 | 151 |
The fraction of the Lys− mutational events that occurred within each run, as well as the fractions of +1 and −1 frameshift events within each run, is given. The denominator of each fraction corresponds to the number of independent Lys− mutants sequenced. Rates (number of mutations per cell per generation) were calculated by multiplying the total Lys− mutation rate by the fraction of the relevant event. All rates were derived using data from at least 12 cultures. The value in parentheses following each 10-nt rate corresponds to the fold increase in rate relative to the mutation rate for same 10-nt run in a wild-type strain.
One −2 event in the 10C run of the msh3Δ strain.
Mutation rates in wild-type (MMR-proficient) strains.
Although the 10-nt runs occupy <1% of the 4.2-kb LYS2 coding sequence, from 25% (10A run) to as many as 98% (10G run) of the forward mutation events in wild-type strains occurred within the runs (Table 1). Mutations in 10G and 10C runs occurred at similar rates (1.05 × 10−5 and 3.41 × 10−6, respectively), as did mutations in the 10A and 10T runs (2.36 × 10−8 and 4.27 × 10−8, respectively). The 10C/10G runs, however, accumulated frameshift mutations at an approximately 100-fold-higher rate than did the 10A/10T runs. Within the 10T, 10C, and 10G runs, +1 events accounted for more than 90% of the mutations, while there were approximately equal numbers of +1 and −1 events in the 10A run.
Mutation rates in completely MMR-defective strains.
The rates and distributions of mutational events within each 10-nt run were very similar in the msh2Δ, mlh1Δ, and pms1Δ strains (Table 2). All three genotypes thus will be considered to be completely MMR defective and will not be further distinguished. As in wild-type strains, the 10C/10G runs were less stable than the 10A/10T runs, but the magnitude of the stability difference in MMR-defective strains was not as great as that in the wild-type strains (approximately 10- and 100-fold, respectively). Relative to the respective rates in the wild-type strains, mutation rates in the MMR-defective strains were elevated several hundred-fold in the 10C/10G runs versus several thousand-fold in the 10A/10T runs, suggesting more efficient MMR-directed removal of frameshift intermediates in 10A/10T than in 10C/10G runs. In contrast to the strong bias for +1 frameshift events in the wild-type strains, −1 and +1 events occurred at roughly equivalent rates in the 10C/10G runs, and −1 events occurred at a consistently higher (on average, fourfold higher) rate than +1 events in the 10A/10T runs.
TABLE 2.
Mutations within 10-nt runs in msh2Δ, pms1Δ, and mlh1Δ strainsa
Genotype | Run | Strain | Total
|
+1 Frameshifts
|
−1 Frameshifts
|
|||
---|---|---|---|---|---|---|---|---|
Fraction | Rate (10−5) | Fraction | Rate (10−5) | Fraction | Rate (10−5) | |||
msh2Δ | 10A | SJR1352 | 54/67 | 9.9 (3,300) | 14/67 | 2.6 | 40/67 | 7.3 |
10T | SJR1353 | 49/57 | 37.7 (8,500) | 12/57 | 9.2 | 37/57 | 28.4 | |
10C | SJR1354 | 49/52b | 105 (310) | 24/52 | 51.4 | 22/52 | 47.2 | |
10G | SJR1355 | 50/50 | 126 (120) | 24/50 | 60.5 | 26/50 | 65.5 | |
pms1Δ | 10A | SJR1393 | 33/37 | 11.6 (3,900) | 9/37 | 3.2 | 24/37 | 8.5 |
10T | SJR1394 | 40/45 | 24.0 (5,400) | 7/45 | 4.2 | 33/45 | 19.8 | |
10C | SJR1396 | 22/23 | 170 (500) | 12/23 | 92.8 | 10/23 | 77.3 | |
10G | SJR1395 | 36/38b | 124 (120) | 16/38 | 55.2 | 19/38 | 65.5 | |
mlh1Δ | 10A | SJR1388 | 24/29 | 13 (4,500) | 6/29 | 3.3 | 18/29 | 9.9 |
10T | SJR1389 | 39/41 | 25.6 (5,800) | 5/41 | 3.8 | 34/41 | 22.3 | |
10C | SJR1391 | 36/36b | 184 (540) | 16/36 | 81.9 | 18/36 | 92.2 | |
10G | SJR1390 | 39/39 | 185 (180) | 21/39 | 99.5 | 18/39 | 85.3 |
The fraction of the Lys− mutational events that occurred within each run, as well as the fractions of +1 and −1 frameshift events within each run, is given. The denominator of each fraction corresponds to the number of independent Lys− mutants sequenced. Rates (number of mutations per cell per generation) were calculated by multiplying the total Lys− mutation rate by the fraction of the relevant event. msh2Δ rates were derived from 12 cultures; rates for pms1 Δ and mlh1Δ strains were each derived from six cultures. The value in parentheses following each 10-nt rate corresponds to the fold increase in rate relative to the mutation rate for same 10-nt run in a wild-type strain.
Additional events in 10-nt runs: msh2Δ, two +2 event and one −2 event in 10C; mlh1Δ, two +2 events in 10C; pms1Δ, one +2 event in 10G.
Mutation rates in msh3Δ and msh6Δ strains.
Consistent with functional redundancy between the Msh2p-Msh3p and Msh2p-Msh6 complexes with respect to the repair of 1-nt IDLs, the single-mutant msh3Δ or msh6Δ strains had much lower mutation rates in the 10-nt runs than did the completely MMR-defective strains (Table 3). For most of the runs, however, the mutation rates in the msh3Δ and msh6Δ strains were not equivalent. The mutation rate within the 10A run, for example, was elevated 59-fold in the msh3Δ strain relative to the wild-type strain, but only 7-fold in the msh6Δ strain. The reverse pattern was seen for the 10T run, where mutation rates were elevated 15-fold and 90-fold in the msh3Δ and msh6Δ strains, respectively. Finally, the 10G run was approximately 10-fold more unstable than the 10C run in msh3Δ strains, but the runs had similar stabilities in the msh6Δ strains.
Not only did overall mutation rates for a given 10-nt run differ in msh3Δ and msh6Δ strains, but the distributions of +1 and −1 mutations also differed (Table 3). In the 10A run, for example, there were 8.6-fold more −1 events than +1 events in the msh3Δ strain, but there were 4-fold more +1 events than −1 events in the msh6Δ strain. In the 10T run, a similar bias for −1 frameshift events was seen in the msh3Δ strain, but no significant bias was evident in the msh6Δ strain. A striking reversal in the distributions of frameshift events was seen for the 10C/10G runs in msh3Δ versus msh6Δ strains. In the 10C run, there were 5.7-fold more −1 events than +1 events in the msh3Δ strain, while in the msh6Δ strain, there were 7.4-fold more +1 events than −1 events. The pattern was reversed in the 10G run, where there was a large excess of +1 events in the msh3Δ strain (46 +1 events versus only 1 −1 event), but a large excess of −1 events in the msh6Δ strain (40 −1 events versus only three +1 events).
DISCUSSION
Frameshift mutations in wild-type cells represent the cumulative effects of DNA polymerase slippage, DNA polymerase proofreading, and postreplicative MMR. Frameshift mutations that accumulate in MMR-defective cells directly reflect the fidelity of the polymerization process, while the repair specificity of the MMR machinery can be deduced by comparing frameshift mutations that accumulate in wild-type versus MMR-defective cells. In the present study, a forward mutation assay was used to examine the effect of mononucleotide run composition on the generation and removal of frameshift intermediates in yeast. Specifically, 10-nt runs were placed at a fixed position within the LYS2 gene, and mutation rates and spectra for each run were determined in appropriate strain backgrounds (Tables 1 to 3). The data for the 10-nt runs are summarized in Fig. 2, where +1 and −1 mutation rates in wild-type, in partially MMR-defective (msh3Δ and msh6Δ mutants), and in completely MMR-defective (msh2Δ, pms1Δ, and mlh1Δ mutants) strains are presented.
FIG. 2.
Rates of +1 and −1 frameshift mutations in 10-nt (10N) runs. (A) Wild type. (B) MMR− (MMR-defective strain background). (C) msh3Δ. (D) msh6Δ. Because the proportions of −1 events in wild-type strains were too low for accurate rate determination, the rates shown are those determined by Harfe and Jinks-Robertson (18). The MMR− mutation rates were calculated by pooling data for the msh2Δ, pms1Δ, and mlh1Δ strains. Error bars correspond to 95% confidence intervals (8). Solid and hatched bars correspond to +1 and −1 frameshift rates, respectively.
In the complete absence of MMR, the 10C/10G runs accumulated both +1 and −1 frameshift mutations at higher rates than did the 10A/10T runs (Fig. 2B). This observation not only extends our prior results obtained with a −1 frameshift detection system (18), but also is consistent with results obtained in other organisms. In MMR-defective E. coli strains, for example, 6C/6G runs accumulate both +1 and −1 frameshifts at higher rates than do 6A/6T runs (12), and an 8C run has been reported to accumulate −1 frameshifts more often than an 8A run (33). Likewise, a 17G run is more unstable than a 17A run in human cells defective for hMLH1 (3), and 8G runs are more unstable than 8A runs in human tumor cell lines exhibiting the microsatellite instability phenotype characteristic of MMR-defective cells (47). Two factors may contribute to the generally higher level of DNA polymerase slippage in C/G than in A/T mononucleotide runs. First, base-stacking interactions would be expected to stabilize a frameshift intermediate in a C/G run to a greater extent than one in an A/T run (see reference 33). Second, the greater strength of CG base pairs would be expected to reduce the efficiency of proofreading in C/G runs relative to A/T runs (2, 14). As discussed by Boyer et al. (3), the greater inherent instability of C/G runs may be relevant to the distributions and abundances of these tracts in a variety of organisms. A general bias for A/T versus G/C runs would promote overall genome stability, whereas C/G runs might predominate in situations in which instability is favored, such as at loci encoding genes that undergo phase variation in pathogenic bacteria (30).
While the rates of −1 versus +1 events in the 10C/10G runs were equivalent in MMR-defective cells, −1 frameshifts occurred at higher rates than did +1 frameshifts in the 10A/10T runs (Fig. 2B). A bias for −1 events in A/T mononucleotide runs has been previously reported in MMR-defective yeast (42) and E. coli (12), indicating that this is a general feature of A/T runs. A surplus of slippage events leading to −1 errors in A/T runs also has been observed in in vitro studies (23) and has been attributed to differences in the number of base pairs that must be disrupted to form the slippage intermediate and/or the number of base pairs that potentially can stabilize the intermediate (see also reference 41). The equivalent rates of +1 and −1 events observed here in 10C/10G runs suggest that such equilibrium considerations may apply only to A/T runs. Alternatively, the in vivo bias for −1 frameshifts in the 10A/10T runs could reflect more efficient DNA polymerase proofreading of +1 than of −1 frameshift intermediates. With the 10C/10G runs, there may be little, if any, proofreading because of the greater difficulty in disrupting C/G as opposed to A/T base pairs, resulting in the observed equivalent numbers of +1 and −1 events (see references 2 and 14). Regardless of the precise mechanistic basis, the data presented here demonstrate that run composition not only affects the overall accumulation of DNA polymerase slippage intermediates, but also can influence the types of frameshift intermediates that are generated.
Restoration of full MMR activity to the MMR-defective strains lowered overall mutation rates several thousand-fold in the 10A/10T runs, but only several hundred-fold in the 10C/10G runs, indicating less-efficient MMR-directed removal of frameshift intermediates in C/G than in A/T runs. It should be noted that the reverse pattern has been reported in E. coli, with MMR-directed removal of frameshift intermediates being more efficient in C/G than in A/T runs (12, 33). In wild-type strains, the numbers of −1 frameshift events identified in the 10T, 10C, and 10G runs were too small to allow accurate estimates of the corresponding −1 frameshift rates (see Table 1). We previously measured −1 rates in the same runs by using a reversion assay, however, and these rates are included in Fig. 2A (18). We believe this inclusion is justified because the −1 frameshift rates presented in Fig. 2B agree very well with those measured previously in MMR-defective strains. In addition, the low numbers of −1 events reported here for the wild-type strains are consistent with the earlier rate measurements. Comparison of the −1 frameshift rates in the wild type versus an MMR-defective strain (Fig. 2A and 2B, respectively) indicates that run composition has little, if any, effect on the efficiency of removing the corresponding mutational intermediates. In contrast, mononucleotide run composition does appear to affect the correction efficiency of +1 frameshift intermediates, leading to smaller rate decreases in the 10C/10G runs than in the 10A/10T runs upon restoration of the MMR system. The greater overall instability of 10C/10G runs relative to 10A/10T runs observed here in wild-type yeast strains thus can be attributed to two factors: (1) more DNA polymerase slippage in the 10C/10G runs to generate both +1 and −1 errors and (2) less-efficient MMR-directed removal of +1 frameshift intermediates in the 10C/10G runs.
For each 10-nt run, there were more +1 events than −1 events in wild-type strains, whereas the number of +1 events was less than or equal to that of −1 events in the corresponding MMR-defective strains. Thus, for 10-nt runs of all composition, the yeast MMR system removes −1 frameshift intermediates (extrahelical nucleotide on the template strand) more efficiently than it removes +1 intermediates (extrahelical nucleotide on primer strand). A similar bias for the MMR-associated removal of extrahelical repeats on the template strand has been reported for poly(GT) dinucleotide repeats in yeast (37) and in mammalian cells (43), suggesting this bias may be a general feature of eukaryotic MMR systems. Although it is not obvious how the eukaryotic MMR system might target extrahelical repeats on the template strand for more efficient removal, the bias presumably results from an asymmetry in the mismatch recognition process or in subsequent steps, either of which could result if there is a close physical association of the MMR machinery with the replisome (6, 10, 20, 44).
In frameshift assays, deletion of either MHS3 or MSH6 typically results in a weak mutator phenotype, while deletion of both genes has synergistic effects on mutation rates, producing a phenotype indistinguishable from that of a msh2Δ strain (15, 17, 18, 27, 35). Although such rate data indicate functional overlap between the Msh2p-Msh3p and Msh2p-Msh6p complexes, distinct differences in the correction efficiencies of the complexes with respect to defined frameshift intermediates have emerged when comparing frameshift spectra derived from msh3Δ versus msh6Δ strains (e.g., see references 15 and 17). For very unstable sequences, such as 10-nt runs, it should be noted that even very small differences in the correction efficiencies of the complexes can translate into very large differences in mutation rates in msh3Δ versus msh6Δ strains. An extreme example of a disparity in correction efficiencies of the complexes is evident in the 10-nt run data shown in Fig. 2, where the +1 frameshift rates within the 10G run are statistically the same in completely MMR-defective and msh3Δ strains (compare panels B and C). Because repair in an msh3Δ mutant reflects the activity of the remaining Msh2p-Msh6p complex, we conclude that the Msh2p-Msh6p complex does not, or only very inefficiently, remove +1 frameshift intermediates in the 10G run. In contrast to its inability to remove +1 frameshift intermediates in the 10G run, the Msh2p-Msh6p complex was able to remove 98% of the −1 frameshift intermediates in this run (Table 4). Although not as pronounced, a reciprocal pattern was evident for the 10C run, with the Msh2p-Msh6p complex removing +1 frameshift intermediates more efficiently than −1 intermediates (99 and 92%, respectively). Interestingly, the Msh2p-Msh3p complex had the reverse specificity, removing +1 frameshift intermediates more efficiently than −1 intermediates in the 10G run (99 and 81%, respectively), and −1 frameshift intermediates more efficiently than +1 intermediates in the 10C run (98 and 88%, respectively).
TABLE 4.
Mismatch correction efficiencies of Msh2p-Msh3p and Msh2p-Msh6pa
Run | Frameshift intermediate | Mismatch correction efficiency
|
|
---|---|---|---|
Msh2p-Msh3p | Msh2p-Msh3p | ||
10A | +1 | 0.9938 | 0.9938 |
−1 | 0.9995 | 0.9810 | |
10T | +1 | 0.9702 | 0.9981 |
−1 | 0.9908 | 0.9977 | |
10C | +1 | 0.8765 | 0.9874 |
−1 | 0.9826 | 0.9248 | |
10G | +1 | 0.9855 | NDb |
−1 | 0.8132 | 0.9787 |
The efficiencies of mismatch correction were determined according to mutation rates obtained in completely MMR-defective (Rmsh2, where msh2 refers to the pooled data from the msh2Δ, mlh1Δ, and pms1Δ strains), Msh3p-defective (Rmsh3) and Msh6p-defective strains (Rmsh6). Correction efficiency was calculated by subtracting the mutation rate after repair from the mutation rate before repair and then dividing this difference by the mutation rate before repair (see reference 35). To determine the efficiency of the Msh2p-Msh3p complex, mutation rates in the completely MMR-defective and msh6Δ strains were compared by using the formula (Rmsh2− Rmsh6)/Rmsh2. The efficiency of error correction by the Msh2p-Msh6p complex was determined by using the formula (Rmsh2 − Rmsh3)/Rmsh2.
ND, not detectable.
The 10C and 10G runs differ with respect to the nucleotide on the coding strand of the LYS2 gene, and this nucleotide reversal has two obvious consequences: (i) the sequence of the DNA strand used as a template for transcription is changed, and (ii) the sequence of the leading- versus lagging-strand template during DNA replication is altered. Although transcriptional processes could influence the relative activities of the Msh2p-Msh3p and Msh2p-Msh6p complexes (46), it seems more likely that the mutational patterns observed here with the 10C/10G runs might reflect leading- versus lagging-strand differences in slippage and/or in subsequent repair. For example, an inverse correlation between DNA polymerase processivity and slippage within mononucleotide runs has been reported in vitro, leading to the suggestion that most slippage within tandem repeats is initiated during DNA polymerase dissociation or reassociation. Because lagging-strand synthesis is inherently less processive than leading-strand synthesis, involving both reinitiation and polymerase switching, one would predict that more frameshift errors would be generated during lagging-strand synthesis (reviewed in reference 24). Recent data from E. coli indicate that the slippage frequency in mononucleotide runs may indeed differ during leading- versus lagging-strand synthesis (12). In support of replication-related asymmetries in mutation accumulation, it should be noted that deletion events arise more often during lagging-strand synthesis (36) and that base substitutions arise more often during leading-strand synthesis (9, 31). In addition to differences in leading versus lagging strands in DNA polymerase slippage, there also could be differences in leading versus lagging strands in MMR processes. If eukaryotic MMR occurs in the context of DNA replication, then the asymmetry inherent in the replication process might extend to the complexes that edit replication fidelity. Finally, the pattern of frameshift accumulation within 10C/10G runs might also reflect different efficiencies of the Msh2p-Msh3p and Msh3p-Msh6p mismatch-binding complexes with respect to an extrahelical G versus C. In this regard, it should be noted that the complexes indeed exhibit different activities towards defined 1-nt IDLs in both in vitro binding assays (1, 13, 28) and in vivo transformation assays (26).
With the exception of the apparent inability of the Msh2p-Msh6p complex to repair +1 frameshift intermediates in the 10G run, the individual repair efficiencies of the Msh2p-Msh3p and Msh2p-Msh6p complexes were at least 97 and 80% in the 10A/10T and 10C/10G runs, respectively (Table 4). In noniterated sequence or in tandem repeats with a small number of repeat units, comparable repair efficiencies would maintain low frameshift rates, even in the presence of only one of the two mismatch-binding complexes. For highly unstable repeats, such as the 10-nt runs used here, however, the presence of only one mismatch-binding complex can result in mutation rates that are several orders of magnitude higher than those in a wild-type strain. The necessity for maintaining genome stability in the presence of such highly unstable repeats may explain the need for functionally redundant mismatch-binding complexes that can recognize IDLs. The 10-nt stability data presented here not only underscore the complexity of the eukaryotic MMR system with regard to removing frameshift intermediates, but also should be useful for predicting which genes or sequences are most likely to be targets for mutagenesis in completely or partially MMR-defective cells.
Acknowledgments
We thank Gray Crouse, Rosann Farber, Tom Petes, and members of the S.J.R. laboratory for useful discussions and for comments on the manuscript.
This work was supported by a grant from the National Science Foundation to S.J.R.
REFERENCES
- 1.Acharya, S., T. Wilson, S. Gradia, M. F. Kane, S. Guerrette, G. T. Marsischky, R. Kolodner, and R. Fishel. 1996. hMSH2 forms specific mispair-binding complexes with hMSH3 and hMSH6. Proc. Natl. Acad. Sci. USA 93:13629-13634. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Bessman, M. J., and J. Reha-Krantz. 1977. Studies on the biochemical basis of spontaneous mutation. J. Mol. Biol. 116:115-123. [DOI] [PubMed] [Google Scholar]
- 3.Boyer, J. C., N. A. Yamada, C. N. Roques, S. B. Hatch, K. Riess, and R. A. Farber. 2002. Sequence dependent instability of mononucleotide microsatellites in cultured mismatch repair proficient and deficient mammalian cells. Hum. Mol. Genet. 11:707-713. [DOI] [PubMed] [Google Scholar]
- 4.Buermeyer, A., S. M. Deschenes, S. M. Baker, and R. M. Liskay. 1999. Mammalian DNA mismatch repair. Annu. Rev. Genet. 33:533-564. [DOI] [PubMed] [Google Scholar]
- 5.Chattoo, F. F., F. Sherman, D. A. Azubalis, T. A. Fjellstedt, D. Mehnert, and M. Ogur. 1979. Selection of lys2 mutants of the yeast Saccharomyces cerevisiae by the utilization of α-aminoadipate. Genetics 93:51-65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Clark, A. B., F. Valle, K. Drotschmann, R. K. Gary, and T. A. Kunkel. 2000. Functional interaction of proliferating cell nuclear antigen with MSH2-MSH6 and MSH2-MSH3 complexes. J. Biol. Chem. 275:36498-36501. [DOI] [PubMed] [Google Scholar]
- 7.Datta, A., A. Adjiri, L. New, G. F. Crouse, and S. Jinks-Robertson. 1996. Mitotic crossovers between diverged sequences are regulated by mismatch repair proteins in Saccharomyces cerevisiae. Mol. Cell. Biol. 16:1085-1093. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Dixon, W. J., and F. J. Massey, Jr. 1969. Introduction to statistical analysis, 3rd ed. McGraw-Hill, Inc., New York, N.Y.
- 9.Fijalkowska, I. J., P. Jonczyk, M. M. Tkaczyk, M. Bialoskorska, and R. M. Schaaper. 1998. Unequal fidelity of leading strand and lagging strand DNA replication on the Escherichia coli chromosome. Proc. Natl. Acad. Sci. USA 95:10020-10025. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Flores-Rozas, H., D. Clark, and R. D. Kolodner. 2000. Proliferating cell nuclear antigen and Msh2p-Msh6p interact to form an active mispair recognition complex. Nat. Genet. 26:375-378. [DOI] [PubMed] [Google Scholar]
- 11.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. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Gawel, D., P. Jonczyk, M. Bialoskorska, R. M. Schaaper, and I. J. Fijalkowska. 2002. Asymmetry of frameshift mutagenesis during leading and lagging-strand replication in Escherichia coli. Mutat. Res. 501:129-136. [DOI] [PubMed] [Google Scholar]
- 13.Genschel, J., S. J. Littman, J. T. Drummond, and P. Modrich. 1998. Isolation of MutSβ from human cells and comparison of the mismatch repair specificities of MutSβ and MutSα. J. Biol. Chem. 273:19895-19901. [DOI] [PubMed] [Google Scholar]
- 14.Goodman, M. F., and D. K. Fygenson. 1998. DNA polymerase fidelity: from genetics toward a biochemical understanding. Genetics 148:1475-1482. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.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. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Harfe, B. D., and S. Jinks-Robertson. 2000. DNA mismatch repair and genetic instability. Annu. Rev. Genet. 34:359-399. [DOI] [PubMed] [Google Scholar]
- 17.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. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Harfe, B. D., and S. Jinks-Robertson. 2000. Sequence composition and context effects on the generation and repair of frameshift intermediates in mononucleotide runs in Saccharomyces cerevisiae. Genetics 156:571-578. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.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. [DOI] [PubMed] [Google Scholar]
- 20.Johnson, R. E., G. K. Kovvali, S. N. Guzder, N. S. Amin, C. Holm, Y. Habraken, P. Sung, L. Prakash, and S. Prakash. 1996. Evidence for involvement of yeast proliferating cell nuclear antigen in DNA mismatch repair. J. Biol. Chem. 271:27987-27990. [DOI] [PubMed] [Google Scholar]
- 21.Johnson, R. E., G. K. Kovvali, L. Prakash, and S. Prakash. 1996. Requirement of the yeast MSH3 and MSH6 genes for MSH2-dependent genomic stability. J. Biol. Chem. 271:7285-7288. [DOI] [PubMed] [Google Scholar]
- 22.Kroutil, L. C., K. Register, K. Bebenek, and T. A. Kunkel. 1996. Exonucleolytic proofreading during replication of repetitive DNA. Biochemistry 35:1046-1053. [DOI] [PubMed] [Google Scholar]
- 23.Kunkel, T. A. 1990. Misalignment-mediated DNA synthesis errors. Biochemistry 29:8003-8011. [DOI] [PubMed] [Google Scholar]
- 24.Kunkel, T. A., and K. Bebenek. 2000. DNA replication fidelity. Annu. Rev. Biochem. 69:497-529. [DOI] [PubMed] [Google Scholar]
- 25.Lea, D. E., and C. A. Coulson. 1949. The distribution of the numbers of mutants in bacterial populations. J. Genet. 49:264-285. [DOI] [PubMed] [Google Scholar]
- 26.Luhr, B., J. Scheller, P. Meyer, and W. Kramer. 1998. Analysis of in vivo correction of defined mismatches in the DNA mismatch repair mutants msh2, msh3 and msh6 of Saccharomyces cerevisiae. Mol. Gen. Genet. 257:362-367. [DOI] [PubMed] [Google Scholar]
- 27.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. [DOI] [PubMed] [Google Scholar]
- 28.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. [DOI] [PubMed] [Google Scholar]
- 29.Modrich, P., and R. Lahue. 1996. Mismatch repair in replication fidelity, genetic recombination and cancer biology. Annu. Rev. Biochem. 65:101-133. [DOI] [PubMed] [Google Scholar]
- 30.Moxon, E. R., P. B. Rainey, M. A. Nowak, and R. E. Lenski. 1994. Adaptive evolution of highly mutable loci in pathogenic bacteria. Curr. Genet. 4:24-33. [DOI] [PubMed] [Google Scholar]
- 31.Pavlov, Y., C. S. Newlon, and T. A. Kunkel. 2002. Yeast origins establish a strand bias for replicational mutagenesis. Mol. Cell 10:207-213. [DOI] [PubMed] [Google Scholar]
- 32.Prolla, T. A., D.-M. Christie, and R. M. Liskay. 1994. Dual requirement in yeast DNA mismatch repair for MLH1 and PMS1, two homologs of the bacterial mutL gene. Mol. Cell. Biol. 14:407-415. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Sagher, D., A. Hsu, and B. Strauss. 1999. Stabilization of the intermediate in frameshift mutation. Mutat. Res. 423:73-77. [DOI] [PubMed] [Google Scholar]
- 34.Sherman, F. 1991. Getting started with yeast. Methods Enzymol. 194:3-20. [DOI] [PubMed] [Google Scholar]
- 35.Sia, E. A., R. J. Kokoska, M. Dominska, P. Greenwell, and T. D. Petes. 1997. Microsatellite instability in yeast: dependence on repeat unit size and DNA mismatch repair genes. Mol. Cell. Biol. 17:2851-2858. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Sinden, R. R., V. I. Hashem, and W. A. Rosche. 1999. DNA-directed mutations. Leading and lagging strand specificity. Ann. N. Y. Acad. Sci. 870:173-189. [DOI] [PubMed] [Google Scholar]
- 37.Strand, M., M. C. Earley, G. F. Crouse, and T. D. Petes. 1995. Mutations in the MSH3 gene preferentially lead to deletions within tracts of simple repetitive DNA in Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA 92:10418-10421. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Strand, M., T. A. Prolla, R. M. Liskay, and T. D. Petes. 1993. Destabilization of tracts of simple repetitive DNA in yeast by mutations affecting DNA mismatch repair. Nature 365:274-276. [DOI] [PubMed] [Google Scholar]
- 39.Strauss, B. S. 1999. Frameshift mutation, microsatellites and mismatch repair. Mutat. Res. 437:195-203. [DOI] [PubMed] [Google Scholar]
- 40.Streisinger, G., Y. Okada, J. Emrich, J. Newton, A. Tsugita, E. Terzaghi, and M. Inouye. 1966. Frameshift mutations and the genetic code. Cold Spring Harbor Symp. Quant. Biol. 31:77-84. [DOI] [PubMed] [Google Scholar]
- 41.Streisinger, G., and J. Owen. 1985. Mechanisms of spontaneous and induced frameshift mutation in bacteriophage T4. Genetics 109:633-659. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.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. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Twerdi, C. D., J. C. Boyer, and R. A. Farber. 1999. Relative rates of insertion and deletion mutations in a microsatellite sequence in cultured cells. Proc. Natl. Acad. Sci. USA 96:2875-2879. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Umar, A., A. B. Buermeyer, J. A. Simon, D. C. Thomas, A. B. Clark, R. M. Liskay, and T. A. Kunkel. 1996. Requirement for PCNA in DNA mismatch repair at a step preceding DNA resynthesis. Cell 87:65-73. [DOI] [PubMed] [Google Scholar]
- 45.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. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Wierdl, M., C. N. Greene, A. Datta, S. Jinks-Robertson, and T. D. Petes. 1996. Destabilization of simple repetitive DNA sequences by transcription in yeast. Genetics 143:713-721. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Zhang, L., J. Yu, J. K. Willson, S. D. Markowitz, K. W. Kinzler, and B. Vogelstein. 2001. Short mononucleotide repeat sequence variability in mismatch repair-deficient cancers. Cancer Res. 61:3801-3805. [PubMed] [Google Scholar]