• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Jun 14, 2005; 102(24): 8639–8643.
Published online Jun 2, 2005. doi:  10.1073/pnas.0503415102
PMCID: PMC1150857

Variation in efficiency of DNA mismatch repair at different sites in the yeast genome


Evolutionary studies have suggested that mutation rates vary significantly at different positions in the eukaryotic genome. The mechanism that is responsible for this context-dependence of mutation rates is not understood. We demonstrate experimentally that frameshift mutation rates in yeast microsatellites depend on the genomic context and that this variation primarily reflects the context-dependence of the efficiency of DNA mismatch repair. We measured the stability of a 16.5-repeat polyGT tract by using a reporter gene (URA3-GT) in which the microsatellite was inserted in-frame into the yeast URA3 gene. We constructed 10 isogenic yeast strains with the reporter gene at different locations in the genome. Rates of frameshift mutations that abolished the correct reading frame of this gene were determined by fluctuation analysis. A 16-fold difference was found among these strains. We made mismatch-repair-deficient (msh2) derivatives of six of the strains. Mutation rates were elevated for all of these strains, but the differences in rates among the strains were substantially reduced. The simplest interpretation of this result is that the efficiency of DNA mismatch repair varies in different regions of the genome, perhaps reflecting some aspect of chromosome structure.

Keywords: genetic instability, microsatellite, mutation rate, Saccharomyces cerevisiae

Comparisons of amino acid or base sequences of orthologous genes indicate that different genes evolve at different rates (1). Differences in the rates of accumulation of amino acid changes or nonsynonymous base substitutions are influenced by selective constraints (1). For highly expressed genes, the rate of synonymous base substitutions is affected by GC content and codon bias (2, 3). In Saccharomyces cerevisiae and mammalian cells, the rates of synonymous substitutions also vary by a factor of ≈10, depending on the position of the gene in the genome (46). The interpretation of these observations is unclear. It is possible that the misincorporation rates of the replicative DNA polymerases are different at different positions in the genome. Alternatively, the fidelity of DNA polymerases may be invariant, but the detection and repair of misincorporation events may be context-specific.

Microsatellites are regions of DNA in which a single base or a small number of bases is repeated in tandem. The polyGT sequence is a particularly common microsatellite in many eukaryotes (7). We have developed (8, 9) methods of measuring the rate of microsatellite alterations. In this study, we used this assay to measure the mutation rates of the same polyGT microsatellite placed in 10 different chromosomal contexts in the yeast genome. We show that the microsatellite mutation rates vary by more than an order of magnitude among different genomic positions in yeast strains that have wild-type DNA mismatch repair. We have demonstrated (8) that the mutation rates of microsatellites are greatly elevated in yeast strains with deficient mismatch repair. In this study, we find that microsatellite instability is substantially elevated in all contexts examined; however, the relative differences between sites are reduced. This result demonstrates that different genomic regions have different efficiencies of DNA mismatch repair.

Materials and Methods

Strain Constructions. All strains were derived from the haploid MS71 (α ade5-1 his7-2 trp1-289 ura3-52) by transformation. LS48, which contains a complete deletion of the ura3-52 allele, was generated by transforming MS71 with a PCR fragment resulting from amplifying the plasmid pA632 (containing a gene conferring resistance to hygromycin) (10), with the following primers: URA3-F, 5′-TCTTAACCCAACTGCACAGAACAAAAACCTGCAGGAAACGAAGATAAATCCGTAGCGTGCAGGTCGAC-3′; and URA3-R, 5′-GCTCTAATTTGTGAGTTTAGTATACATGCATTTACTTATAATACAGTTTTATCGATGAATTCGAGCTCG-3′. In the hygromycin-resistant transformants, the HYG gene replaces the URA3 sequence. The strains with insertions of the URA3-GT reporter (which confers wild-type URA3 activity) were constructed by transforming LS48 with PCR fragments generated by amplification of sequences derived from plasmid pMBW1 (see Fig. 1), which contains the URA3-GT fusion gene (9). Each primer was ≈75 bases in length, with the 5′ 45–55 bases representing yeast genome sequences flanking the insertion (Stanford University Saccharomyces Genome Database, http://genome-www.stanford.edu) and the 3′ bases representing sequences of pMBW1 flanking the URA3-GT fusion gene. The sequence at the 3′ end of the primers was either 5′-CTCGAGCGCTGATTAAATTACCCCAG-3′ (located upstream of the fusion gene) or 5′-GGATCCAATAACTGATATAAT-3′ (located downstream). The positions of the insertions in the genome were as follows: 74,422–74,423 on III (ARS306-I::URA3-GT and ARS306-O::URA3-GT), 210,606–210,696 on VI (sup6Δ::URA3-GT), 57,148–57,149 on III (SRO9::URA3-GT), 114,597–114,598 on III (CEN3::URA3-GT), 152,212–152,213 on V (CEN5::URA3-GT), 151,045–151,046 on XII (CEN12::URA3-GT), 946,305–946,395 on IV (sup2Δ::URA3-GT), 4,850–4,851 on III (YCLWomega2::URA3-GT), and 1,170–1,171 on III (ARS300::URA3-GT). The LS48 strain was transformed with the resulting fragments, selecting for Ura+ transformants in galactose-containing medium, and the positions of the insertions were confirmed by PCR. We confirmed that the lengths of the micro-satellite sequences in each strain were identical by measuring the sizes of PCR fragments using capillary electrophoresis with an Applied Biosystems Genetic Analyzer. Strains with a null allele of msh2 (msh2::kanMX) were constructed by transformation using the described PCR primers and plasmid substrate (11).

Fig. 1.
Construction of yeast strains with the URA3-GT reporter gene in different chromosome contexts. The plasmid pMBW1 (9) contains a frameshift reporter gene (URA3-GT). This fusion gene has small segments derived from the yeast LYS2 and HIS4 genes, and most ...

Determination of Microsatellite Mutation Rates. For the assay of microsatellite instability, each strain with the URA3-GT reporter gene was grown on solid rich growth medium with 2% galactose (9). Dilutions of 20–50 individual colonies were plated on SG (synthetic galactose) complete medium (to measure the number of cells per colony) or SG plus 5-fluoroorotic acid (5-FOA; to measure the frequency of Ura- derivatives). Rate estimates and 95% confidence limits for these estimates were calculated by using the method of the median, as described (9). Statistical tests of correlations between rates and other factors (such as GC content) were done by using the instat program.

Measurement of Microsatellite Lengths. DNA was isolated from independent 5-FOAR colonies, and PCR was performed by using primers with the following sequences that flank the microsatellite tract: 5′-CCAACGTGGTCATTTAATGAGC-3′ and 5′ HEX-GCTTGAACTCGTCTAATTTG-3′. The HEX tag (Synthegen, Houston) allows the PCR product to be detected by fluorescence. The sizes of the PCR fragments were measured by using capillary electrophoresis on the Applied Biosystems 310 Genetic Analyzer (for details, see ref. 12).


We constructed 10 isogenic yeast strains with a URA3-GT reporter gene (see Fig. 1) including 16.5 in-frame GT-repeats (33 bp) at different locations in the genome. The locations of these constructs in the different strains were as follows: near the centromeres of three different chromosomes (CEN3::URA3-GT, CEN5::URA3-GT, and CEN12::URA3-GT), near a DNA replication origin ARS306 (ARS306-I::UR A3-GT and ARS306-O::URA3-GT at the same point of insertion but with opposite orientations of URA3-GT), near the telomere of chromosome III [ARS300::URA3-GT (1 kb from the telomere) and YCLWomega2::URA3-GT (5 kb from the telomere)], at the positions of tyrosine-tRNA genes (sup2Δ::URA3-GT and sup6Δ::URA3-GT), and near the middle of the left arm of chromosome III (SRO9::URA3-GT). All insertions were in intergenic regions, with the exception of the SUP2 and SUP6 insertions, in which the reporter gene replaced the coding sequence of the tRNA genes. The exact positions of the insertions are given in Materials and Methods.

Because the URA3-GT reporter has an in-frame microsatellite insertion, the strains were Ura+; such strains are sensitive to 5-FOA. Thus, we could select derivatives of the starting strain that had an alteration in the microsatellite that resulted in loss of the correct reading frame (for example, an insertion or deletion of one repeat) by plating the cells on medium containing 5-FOA. For each of the 10 strains, we measured the frequency of 5-FOA-resistant derivatives in ≈20 independent cultures. These frequencies were converted to rates based on a fluctuation analysis. The rates (×10-5 per division) and 95% confidence limits (shown in parentheses) for each mismatch repair (MMR)- proficient (MMR+) strain were as follows: 1.3 (1–1.9), ARS300::URA3-GT; 1.7 (1.3–2.2), YCLWomega2::URA3-GT; 1.7 (1.4–2.0), CEN12::URA3-GT; 2.4 (2–3.3), sup2Δ::URA3-GT; 3.2 (2.8–3.5), SRO9::URA3-GT; 3.6 (2.9–3.9), ARS306-I::UR A3-GT; 3.9 (2.8–4.5), ARS306-O::URA3-GT; 5.4 (4.4–6.7), CEN5::URA3-GT; 6.9 (5.6–7.8), CEN3::URA3-GT; and 21 (1725), sup6Δ::URA3-GT.

A graphical summary of the rate measurements is shown in Fig. 2 (white bars). The rates varied between 1.3 × 10-5 per generation (ARS300::URA3-GT) to 21 × 10-5 per generation (sup6Δ::URA3-GT), a 16-fold difference. The two strains with telomeric insertions had the lowest rates of mutation, although these rates were not significantly different from the strains with the reporters at CEN12 and SUP2. We chose to examine insertions at the SUP2 and SUP6 loci because Ito-Harashima et al. (5) had shown previously that the SUP2 and SUP6 tRNA genes mutated to become ochre suppressors at 20-fold-different frequencies, SUP6 having the higher frequency. In our strains, in which URA3-GT replaced SUP2 or SUP6, we found that the sup6Δ::URA3-GT strain had a 9-fold higher mutation rate than the sup2Δ::URA3-GT strain.

Fig. 2.
Context-dependent rates of microsatellite instability in wild-type (MMR+) and msh2 mutant (MMR-) yeast strains. White bars show the rates derived from MMR+ cells, and gray bars show the rates for the isogenic MMR- strain; 6 of the 10 strains had an isogenic ...

To determine the nature of the microsatellite alterations, we measured the sizes of the microsatellites in independent 5-FOA-resistant derivatives of each strain by PCR and capillary electrophoresis. As expected from previous studies in yeast (13), we found that most of the 5-FOA-resistant derivatives had a tract that was one repeat larger or smaller than the original 33-bp tract (Table 1). There was a ≈2-fold bias for insertions compared with deletions, as observed previously, although this bias was stronger for one of the strains (CEN12::URA3-GT). In approximately half of the 5-FOA-resistant derivatives of the strain with the ARS300::URA3-GT insertion, the URA3-GT reporter gene was lost. This loss is probably the result of a gene conversion event between the reporter gene at ARS300 with other subtelomeric regions, because the ARS300 sequence is part of the subtelomeric X element that is represented at all of the yeast telomeres (14). This class of events was not included in the calculations of the microsatellite mutation rates shown in Fig. 2.

Table 1.
Spectra of mutations observed in 5-FOA-resistant derivatives of the strains with the URA3-GT reporter gene (in bp)

The Msh2 protein is required for DNA mismatch repair in yeast (15). Yeast strains that lack Msh2p have greatly increased rates of microsatellite mutations (8). We made msh2 derivatives of six of the strains described above (CEN12::URA3-GT, CEN3::UR A3-GT, sup2Δ::UR A3-GT, sup6Δ::UR A3-GT, ARS306-O::URA3-GT, and SRO9::URA3-GT). The rates of microsatellite alterations (×10-3 per division) and 95% confidence limits for each MMR- strain were as follows: 1.3 (1.1–1.4), CEN12::URA3-GT; 1.4 (1.2–1.8), sup2Δ::URA3-GT; 1.4 (1–1.6), SRO9::URA3-GT; 1.1 (0.9–1.5), ARS306-O::URA3-GT; 1.5 (1–1.6), CEN3::URA3-GT; and 2.1 (1.6–2.7), sup6Δ::URA3-GT. These data are shown graphically as gray bars in Fig. 2. Although, as expected, the mutation rates for all of the microsatellites were substantially elevated in the MMR- strains, the differences among the mutation rates at different genomic sites were substantially reduced compared with the differences observed in the MMR+ strains. [Note that the scales on the y axes for the MMR- (right axis) and MMR+ (left axis) strains are different.] These results indicate that the context-dependence of mutation rates observed in wild-type strains reflects primarily the context-dependence of DNA mismatch repair rather than variation in the rates of replication errors.

We also examined the spectra of microsatellite alterations in the MMR- strains (Table 1). In contrast to the spectra observed in the MMR+ strains, in which additions of a single repeat are more common than deletions of a single repeat, most of the MMR- strains had no bias or the opposite bias.


These results provide direct experimental evidence for context-specific differences in the rates of frameshift mutations in microsatellite sequences. The loss or gain of repeats within microsatellites is likely to represent DNA polymerase slippage events (13). By this model (Fig. 3), during replication of a repetitive tract of DNA, the primer and template DNA strands temporarily dissociate and then reassociate out of alignment, leading to looping out of one or more repeats of the duplex (16). If the displaced repeat is on the primer strand, failure to correct this intermediate would result in an addition; if the displaced repeat is on the template strand, the resulting intermediate would result in a deletion.

Fig. 3.
Alterations in the length of a microsatellite by DNA polymerase slippage. Replication of a poly GT/CA tract that is seven repeats in length is shown. A slippage event in which the displaced repeat is on the primer strand, if uncorrected, will result in ...

In wild-type cells, it is thought that most of the resulting mismatches are corrected by the cellular DNA mismatch repair system. In support of this argument, cells that lack DNA mismatch repair have dramatically (>100-fold) elevated rates of microsatellite instability (8). As described above, for the 16.5 repeat polyGT tract in wild-type strains, additions are more common than deletions, whereas in the MMR- strains, approximately equal numbers of the two types of changes are observed. The simplest interpretation of this result, which has also been observed previously (13), is that DNA polymerase slippage results in equal numbers of mismatches in which the displaced repeats are located on the primer and template strands. The bias observed in the wild-type strain in favor of additions over deletions argues that the mismatches with a displaced repeat on the template strand are a better substrate for mismatch repair than those with a displaced repeat on the primer strand (13).

The context-dependence differences in microsatellite stability in wild-type strains could reflect context-dependence of the rate of polymerase slippage events, the efficiency of correction of mismatches by the cellular repair systems, or both factors. If the context-dependence of microsatellite stability primarily reflects context-dependence of mismatch correction, the differences among sites in mutation rates should be lost in cells with mutations that eliminate DNA mismatch repair. We find that the context-dependence of mutation rates is reduced from the 16-fold observed in wild-type strains to 2-fold. Thus, we conclude that much of the context-dependence of mutation rates reflects the relative efficiencies of DNA mismatch repair at the different sites. The comparison of mutation rates in the MMR- and MMR+ strains allows calculation of the efficiency of repair at the various sites to be as follows: 99% (CEN12::URA3-GT), 98% (sup2Δ::URA3-GT), 98% (SRO9::URA3-GT), 96% (ARS306-O::URA3-GT), 95% (CEN3::URA3-GT), and 90% (sup6Δ::URA3-GT). However, it is likely that the efficiency of DNA mismatch repair is not the sole determinant of context-dependent effects on mutation rates, because the differences in microsatellite stability for some strains (for example, CEN12::URA3-GT and sup6Δ::URA3-GT) are still statistically significant, even in the msh2 background. Also, differences in mutation rates that are likely to be related to the site of insertion of the microsatellite have been observed in mismatch-repair-defective mammalian cells (17).

There is an alternative explanation for the observation that the context-dependence of microsatellite instability rates is higher in wild-type strains than in msh2 strains. It is possible that some DNA polymerase slippage events occur at a time in the cell cycle (for example, outside of the S period) in which DNA mismatch repair is not active. This hypothesis would be consistent with our data if the frequency of this type of event is rare relative to DNA polymerase slippage events that occur during the S period (which are efficiently corrected in wild-type, but not msh2, strains) and the frequency of these events varies among different regions of the genome. This hypothesis requires the existence of a class of DNA polymerase slippage events that has not been reported.

DNA mismatch repair in eukaryotes is a complex process involving multiple steps (1820). First, the mismatch is bound by a heterodimer of MutS homologues, either Msh2p/Msh6p (for base–base mismatches and one or two base loops) or Msh2p/Msh3p (for loops of 2–14 bp). This binding triggers an ATP-dependent conformational change that results in the recruitment of a MutL heterodimer (Mlh1p/Pms1p). The resulting ternary complex translocates along the DNA until a DNA nick [possibly bound by proliferating cell nuclear antigen (PCNA)] is encountered. The exonuclease Exo1p is recruited to the complex, resulting in excision of the nicked strand. This process (perhaps requiring multiple rounds of excision) eventually results in removal of the mismatch. The resulting single-stranded gap is filled in by DNA polymerase.

Although the efficiency of mismatch repair could be altered by affecting any of these steps described above, we suggest that the most likely steps to be affected are DNA mismatch recognition and DNA mismatch excision. For example, different regions of the chromosome may allow mismatch recognition to occur relatively efficiently or inefficiently. Marsischky and Kolodner (21) showed that the binding of the Msh2p–Msh6p complex to mismatches was affected by sequence context. Although these effects were local (≤18 bp from the mismatch), these studies do not preclude sequence context effects that can act at longer distances from the mismatch.

Alternatively, in different regions of the genome, the average distance between the mismatch and the nick that directs strand-specific removal could vary. Because the efficiency of mismatch repair is reduced as the distance between the mismatch and the nick is increased (21), this property could affect mutation rates. Pavlov et al. (22) showed that the efficiency of repair of an 8-oxoguanine-A mismatch was higher when the mismatch was on the lagging strand of the replication fork than when it was on the leading strand. This conclusion was based primarily on two observations. First, the reversion rate of a mutation (revertible by a C-to-A transversion) in an ogg1 strain (deficient in the removal of 8-oxoguanine) depended on the orientation of the reporter gene with respect to the replication origin. Second, the rate differences observed for the two orientations were substantially reduced by mutations eliminating mismatch repair. The authors suggested that repair was more efficient on the lagging strand because the mismatches were closer to the nicks directing repair. Because we find that the orientation of the reporter relative to the replication origin does not have a significant effect on the rates of instability (comparison of the rates for AR306-I::URA3-GT and ARS306-O::URA3-GT), our results are not explained by this model unless additional conditions are operative. For example, it is possible that DNA polymerase slippage events occur more frequently on the lagging strand, which could balance the effect of more efficient DNA mismatch repair; however, even if this condition were met, it would not explain the variation in rates of instability of the reporter genes located at different genomic positions in our study.

Although we cannot determine which step of mismatch repair is affected by chromosomal context, we examined several chromosome properties including DNA replication, transcription, base composition, and chromatin structure. There does not appear to be a simple relationship between the efficiency of DNA mismatch repair and DNA replication. We compared the positions of the inserted reporter genes with the positions of active replication origins, as established by Raghuraman et al. (23). The four strains with insertions near active ARS elements (ARS306-I::URA3-GT, ARS306-O::URA3-GT, CEN3::URA3-GT, and CEN12::URA3-GT) had quite different rates of microsatellite instability; therefore, the position of the mismatch with respect to the replication origin does not appear to be important in determining the efficiency of the repair event. The expected times of replication (minutes from the beginning of the S period), based on the data of Raghuraman et al. (23) for the different insertions are as follows: 12 (ARS306-I::URA3-GT and ARS306-O::URA3-GT), 17 (sup6Δ::URA3-GT), 18 (SRO9::URA3-GT), 20 (CEN3::URA3-GT and CEN5::UR A3-GT), 23 (CEN12::UR A3-GT), 36 (sup2Δ::URA3-GT), 39 (YCLWomega2::URA3-GT), and 40 (ARS300::URA3-GT). We found no significant correlation between timing of replication and mutation rates (P = 0.27), although the two insertions with the lowest mutation rates were replicated latest in the S period.

In yeast, high rates of transcription result in elevated rates of mutation (9, 24). Because the URA3-GT reporter is expressed from the same strong promoter (GAL1,10) at all positions of the insertion, transcription is unlikely to be relevant to the effects observed in our experiments. Our results also suggest that similar context-dependent effects on mutations rates occur for both RNA polymerase III- and II-transcribed genes (tRNA genes and URA3-GT, respectively).

It has been argued that chromosomal regions of high G+C content may represent regions with relatively efficient DNA mismatch repair, because mismatch repair in yeast and mammals has a bias leading to increased G+C content (25, 26). Consequently, we looked for a correlation between mutation rates and the G+C content of the sequences in 1-, 5-, and 20-kb “windows” flanking the insertions. No significant correlations were found (P > 0.4 for all comparisons).

We suggest that the context-dependence of DNA mismatch repair efficiency is likely to reflect some aspect of chromatin structure that affects the recognition of the DNA mismatch, although other possibilities are not excluded. It is possible that MMR efficiency is a reflection of general chromatin “openness” (often associated with histone hyperacetylation and hypomethylation). Alternatively, the efficiency of repair could be affected by some particular pattern of histone modifications or the binding of nonhistone proteins. Given the plethora of such modifications and such proteins, it is likely to be a challenge to determine the mechanistic basis of these effects. Regardless of the details of the mechanisms, our results demonstrate that the DNA mismatch repair system operates with different efficiencies in different regions of the yeast genome, and that these differences influence mutation rates. These findings substantiate the hypothesis that differences in rates of evolution among genes result partly from variations in mutation rates in different genomic regions (46).


We thank Drs. T. Vision, R. M. Liskay, J. L. Argueso, K. Wolfe, P. Modrich, R. Kolodner, D. Gordenin, T. Kunkel, and S. Jinks-Robertson for useful discussions. This work was supported by National Institutes of Health Grants CA63264 (to R.A.F.) and GM24110 and GM52319 (to T.D.P.).


Author contributions: J.D.H., J.C.B., T.D.P., and R.A.F. designed research; J.D.H., L.S., and J.C.B. performed research; J.D.H., T.D.P., and R.A.F. analyzed data; J.D.H., T.D.P., and R.A.F. wrote the paper.

Abbreviations: 5-FOA, 5-fluoroorotic acid; MMR, mismatch repair.


1. Nei, M. (1987) Molecular Evolutionary Genetics (Columbia Univ. Press, New York).
2. Sharp, P. A. & Li, W.-H. (1987) Mol. Biol. Evol. 4, 222-230. [PubMed]
3. Ticher, A. & Graur, D. (1989) J. Mol. Evol. 28, 289-298. [PubMed]
4. Wolfe, K. H., Sharp, P. A. & Li, W.-H. (1989) Nature 337, 283-285. [PubMed]
5. Ito-Harashima, S., Hartzog, P. E., Sinha, H. & McCusker, J. H. (2002) Genetics 161, 1395-1410. [PMC free article] [PubMed]
6. Ellegren, H., Smith, N. G. C. & Webster, M. T. (2003) Curr. Opin. Genet. Dev. 13, 562-568. [PubMed]
7. Tautz, D. & Renz, M. (1984) Nucleic Acids Res. 12, 4127-4138. [PMC free article] [PubMed]
8. Strand, M., Prolla, T. A., Liskay, R. M. & Petes, T. D. (1993) Nature 365, 274-276. [PubMed]
9. Wierdl, M., Greene, C. N., Datta, A., Jinks-Robertson, S. & Petes, T. D. (1996) Genetics 143, 713-721. [PMC free article] [PubMed]
10. Goldstein, A. L. & McCusker, J. H. (1999) Yeast 15, 1541-1553. [PubMed]
11. Kearney, H. M., Kirkpatrick, D. R., Gerton, J. L. & Petes, T. D. (2001) Genetics 158, 1457-1476. [PMC free article] [PubMed]
12. Hatch, S. B. & Farber, R. A. (2003) Mut. Res. 545, 117-126. [PubMed]
13. Sia, E. A., Jinks-Robertson, S. & Petes, T. D. (1997) Mut. Res. 383, 61-70. [PubMed]
14. Louis, E. J. (1995) Yeast 11, 1553-1573. [PubMed]
15. Reenan, R. A. G. & Kolodner, R. D. (1992) Genetics 132, 975-985. [PMC free article] [PubMed]
16. Streisinger, G., Okada, Y., Emrich, J., Newton, J., Tsugita, A., Terzaghi, E. & Inouye, M. (1966) Cold Spring Harbor Symp. Quant. Biol. 31, 77-84. [PubMed]
17. Hanford, M. G., Rushton, B. C., Gowen, L. C. & Farber, R. A. (1998) Oncogene 16, 2389-2393. [PubMed]
18. Fishel, R. (1999) Nat. Med. 5, 1239-1241. [PubMed]
19. Dzantiev, L., Constantin, N., Genschel, J., Iyer, R. R., Burgers, P. M. & Modrich, P. (2004) Mol. Cell 15, 31-41. [PubMed]
20. Stojic, L., Brun, R. & Jiricny, J. (2004) DNA Rep. 3, 1091-1101. [PubMed]
21. Marsischky, G. T. & Kolodner, R. (1999) J. Biol. Chem. 274, 26668-26682. [PubMed]
22. Pavlov, Y. I., Mian, I. M. & Kunkel, T. A. (2003) Curr. Biol. 13, 744-748. [PubMed]
23. Raghuraman, M. K., Winzeler, E. A., Collingwood, D., Hunt, S., Wodicka, L., Conway, A., Lockhart, D. J., Davis, R. W., Brewer, B. J. & Fangman, W. L. (2001) Science 294, 115-121. [PubMed]
24. Datta, A. & Jinks-Robertson, S. (1995) Science 268, 1616-1619. [PubMed]
25. Brown, T. C. & Jiricny, J. (1989) Genome 31, 578-583. [PubMed]
26. Birdsell, J. A. (2002) Mol. Biol. Evol. 19, 1181-1197. [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...