Massively Parallel Measurement of DNA Mismatch Repair Efficiency In Vivo

Abstract

Due to mutagens or errors in DNA synthesis, single base pair mismatches occur frequently in vivo. To ensure accurate propagation of genetic information, cells monitor and repair such DNA defects. The efficiency of this cellular response is known to depend on the type and the sequence context of the mismatch in bacteria as well as in single- and multi-cellular eukaryotes. Although methods to quantify repair propensities of individual mismatches exist, only a small subset of possible mismatches have been examined due to the labor-intensive nature of these assays. Here, we present a high-throughput approach that can quantify the mismatch repair efficiency. We tagged each plasmid with a random barcode sequence. As the descendants of each plasmid will inherit the same unique barcode, the replication products of the ancestor plasmid can be traced. Into this barcoded library, we inserted a DNA library with 150 different mismatches and observed their fate after transformation into Escherichia coli. If a mismatch gets repaired before the first replication, one of the strands is converted into the proper complementary of the other strand. Subsequent replications of such ancestors yield a pure product. On the contrary, an ancestor plasmid evading repair gives rise to mixture of two products that differ at the position of the mismatch. Mismatch repair efficiency can hence be quantified by next generation sequencing of the DNA products. Repair efficiency was above 95% for majority of mismatch types and contexts, which was greatly reduced in ΔmutS and ΔmutL strains, validating the approach. We also identified some poorly repaired mismatches. Our assay is generally applicable to other organisms and can reveal valuable information about the sequence-dependence of mismatch repair efficiency with implications in evolution of genome, codons and regulatory elements.