Cloning and expression in Escherichia coli of a human cDNA encoding the DNA repair protein N-methylpurine-DNA glycosylase

A 871-base pair cDNA encoding the human N-methylpurine-DNA glycosylase (MPG) was cloned from a HeLa S3 cDNA expression library in a pUC vector by phenotypic screening of MPG-negative (tag^- alkA^-) Escherichia coli cells exposed to methylmethane sulfonate. The active MPG is expressed as a 31-kDa fusion protein. The human cDNA-encoded MPG releases 3-methyladenine, 7-methylguanine, and 3-methylguanine from DNA and thus has a substrate range similar to that of the indigenous enzyme and the E. coli AlkA protein. The cDNA hybridizes with distinct restriction fragments of mammalian DNAs but not with E. coli or yeast DNA. A search in the GenBank data bank failed to show any other cloned DNA with a similar sequence. Although the human protein has 62% sequence homology with the corresponding rat enzyme, only a few amino acid residues are conserved between the human protein and the E. coli and yeast MPGs. However, a conserved glutamine residue in all MPGs that release 3-alkyladenine and an arginine residue in eukaryotic MPGs and E. coli AlkA that act additionally on N-alkylguanines suggest that these residues are involved in recognition of adenine and guanine adducts in DNA, respectively. Although the 1.1-kilobase mRNAs of MPG from human and rodents are similar in size, liver and cultured cells of rat have much lower levels of MPG mRNA than do human and mouse cells. A hamster cell line variant isolated as being resistant to methylmethane sulfonate does not have a higher level of MPG mRNA than the parent cell line.