The identification of a novel alternatively spliced form of the MBD4 DNA glycosylase.

Methyl-CpG binding protein 4 (MBD4) is a mismatch-specific G:T and G:U DNA glycosylase. During an analysis of MBD4 expression in HeLa cells we noted the presence of an unexpectedly short reverse transcribed product. This cDNA lacked the region encoding the methyl-binding domain and exon 3 of MBD4 but retained the glycosylase domain. Sequence comparison indicates the existence of a previously unreported cryptic splice site in the MBD4 genomic sequence thus illuminating a mechanism whereby a glycosylase acquired a methyl-binding capacity, thus targeting potential mutagenic CpG sites. In vitro assays of this highly purified species, refolded in arginine rich conditions, confirmed that this unique, short version of MBD4 possessed uracil DNA glycosylase but not thymine DNA glycosylase activity. We conclude that the identification of a transcript encoding a short version of MBD4 indicates that MBD4 expression may be more complex than previously reported, and is worthy of further investigation.