A 'Semi-Protected Oligonucleotide Recombination' Assay for DNA Mismatch Repair in vivo Suggests Different Modes of Repair for Lagging Strand Mismatches

Abstract

In Escherichia coli, a DNA mismatch repair (MMR) pathway corrects errors that occur during DNA replication by coordinating the excision and re-synthesis of a long tract of the newly-replicated DNA between an epigenetic signal (a hemi-methylated d(GATC) site or a single-stranded nick) and the replication error after the error is identified by protein MutS. Recent observations suggest that this 'long-patch repair' between these sites is coordinated in the same direction of replication by the replisome. Here, we have developed a new assay that uniquely allows us to introduce targeted 'mismatches' directly into the replication fork via oligonucleotide recombination, examine the directionality of MMR, and quantify the nucleotide-dependence, sequence context-dependence, and strand-dependence of their repair in vivo-something otherwise nearly impossible to achieve. We find that repair of genomic lagging strand mismatches occurs bi-directionally in E. coli and that, while all MutS-recognized mismatches had been thought to be repaired in a consistent manner, the directional bias of repair and the effects of mutations in MutS are dependent on the molecular species of the mismatch. Because oligonucleotide recombination is routinely performed in both prokaryotic and eukaryotic cells, we expect this assay will be broadly applicable for investigating mechanisms of MMR in vivo.

Figure 1.

A ‘semi-protected oligonucleotide recombination’ (SPORE) assay to quantify mismatch repair (MMR) efficiency in vivo in a nucleotide-, sequence-context-, strand-, direction- and chromosomal context-/orientation-dependent manner. (A) (left) In the SPORE assay presented here, a synthetic oligonucleotide (oligo, red) with significant homology to non-template strand (NT) of galactose kinase gene galK is designed to hybridize with the lagging strand during replication (22). (right) In the E. coli strains used, the oligo is designed to target the region surrounding an amber mutation. (B) Example segments of two of the 70-nucleotide-long synthetic oligos used in the SPORE assay. See text for details. Oligos are designed to possess (i) MMR-inactive ‘control’ mismatch designed to correct the amber mutation after the oligo is incorporated into the genome at the replication fork and (ii) a MMR-reactive ‘probe’ mismatch to one side of the control mismatch that introduces a silent mutation. Phosphorothioate bonds (*), which flank the control mismatch, block long-patch repair of the probe mismatch from the opposite end. (C) Simplified protocol of the SPORE assay. See text and Experimental Procedures for details. (D) Quantification of repair efficiencies is obtained by comparing the decrease in the sequencing signal at the probe mutation site relative to that of a SPORE assay using a MMR-deficient (MutS KO) strain, after selecting for the control mutation by ability to metabolize galactose. See also Supplementary Figure S1 for example chromatograms.

Figure 2.

Repair efficiencies obtained by semi-protected oligonucleotide recombination (SPORE) assay for lagging strand repair of T–T (left panel) and G–T (right panel) mismatches. See Text and Experimental Procedures for details, with oligos showing direction of long-patch repair (LPR) allowed below. Beeswarm plot of individual experimental data points overlaid over bar graphs for repair of each strain (see text for details) tested for 5΄-directed long patch repair (using oligo 5΄-XT, where X is G or T; see Figure 1) or 3΄-long patch repair (using oligo 3΄-XT). Error bars are 95% confidence around mean (bar height) repair, and repair efficiency of 0% is defined as the mean repair efficiency of the mismatch repair defective (MutS KO) strain for each oligo. See Supplementary Tables S1–S5. MutS KO, a MMR-defective, MutS knock-out strain; MutS wt, strain with functional MMR pathway; MutS ¹⁵AAYAAL²⁰, a strain with a mutation in the mutS gene which destabilizes the protein (58); MutS D835R, a strain with a mutation in the mutS gene which disrupts the ability of MutS to tetramerize but does not affect its ability to dimerize (27); MutS ¹⁵AAYAAL²⁰ D835R, a strain with a double mutation in the mutS gene.

Figure 3.

Repair efficiencies, with oligos showing direction of long-patch repair (LPR) allowed below as in Figure 2, obtained by semi-protected oligonucleotide recombination (SPORE) assay of G–T mismatches where the mismatched G ([G]) is flanked by either pYrimidines (Y[G]Y) or puRines (R[G]R).

Figure 4.

Asymmetry in the bi-directional repair of lagging strand mismatches. (i) An ‘activated’ MutS dimer detaches from the replisome at a mismatch site and can initiate and direct repair by diffusing either to the 5΄-single-stranded DNA/double-stranded DNA (ssDNA-dsDNA) junction (with MutL), or (ii) by looping the DNA as a tetramer after diffusing to and nicking a 3΄-hemi-methylated d(GATC) site with MutLH. If MutS is unable to tetramerize, repair directed from hemi-methylated d(GATC) sites becomes uncoordinated.

Scheme 1.

Simplified model of repair efficiencies from SPORE assays.