Mismatch repair in Xenopus egg extracts is not strand-directed by DNA methylation

The efficiency of Xenopus laevis egg extract to repair T:G and A:C mismatch ed base pairs in unmethylated, hemimethylated and fully methylated heterodu plexes was investigated. Filamentous phage M13mp18 and its derivative M13mp 18/MP-1 (C changed to T inside the sequence dCC*C GGG, at the position 6248 ) were used for heteroduplexes construction. The three origins of mismatche d base-pairs in the eukaryotic DNA are mimicked by in vitro methylation: he mimethylated DNA (me(-)/me(+)) for replication errors; unmethylated (me(-)/ me(-)) and fullymethylated DNA (me(+)/me(+)) for recombination heteroduplex es, and fullymethylated also for locally, spontaneously deaminated 5-methyl cytosine (5meC) to T, generating the exclusively T:G mismatch. The methylat ions were in CpG dinucleotides, mostly characteristic of eukaryotic cells [ 5, 24]. In vitro methylation was done by HpaII methylase which methylate central C of dCCGG sequence in the manner of eukaryotic methylation. The position of mismatched bases was chosen so that correction of mismatched bases in any s trand would create the sequence for one of the "diagnostic" restriction end onucleases, either BstNI or MspI. Correction efficiency was about 10(8) rep air events per egg equivalent. Correction in favor of C:G base pair restoration occurred regardless of the T:G or C:A mispairs, with almost equal efficiency. Repair of T:G to T:A wa s up to 10 times less efficient comparing to C:G, and repair of C:A to T:A was in our experimental system undetectable. No significant difference in r epair efficiency of mismatched bases situated in unmethylated, hemimethylat ed or fullymethylated heteroduplexes indicate methylation-independent repai r of mismatched bases in X, laevis oocite extracts.