The gene which codes for endonuclease III of Escherichia coli has been sequenced. The nth gene was previously subcloned and defined as the gene which led to overproduction of endonuclease III when present on a multicopy plasmid and which created a deficiency in endonuclease III activity when mutated. The nth gene was sequenced and translated into a predicted polypeptide. The molecular weight (23,546), the amino-terminal amino acid sequence, and the amino acid composition of the polypeptide predicted from the nucleotide sequence are excellent agreement with those same properties determined for the purified protein. Thus, the nth gene is the structural gene for endonuclease III. Inspection of the nucleotide sequence reveals that there is an open reading frame immediately upstream of the nth gene, suggesting that it might be part of an operon. There is a region of dyad symmetry which could form a hairpin stem and loop structure if transcribed into RNA characteristic of a rho-dependent terminator downstream from the nth gene. The nth gene of Escherichia coli has been cloned onto a lambda PL expression vector which yields approximately 300-fold overproduction of endonuclease III. We have purified the enzyme to apparent homogeneity using two chromatographic steps. Our purification scheme allowed the preparation of 117 mg of protein from 190 g of E. coli with a 70% yield. The purified protein has both AP endonuclease activity and DNA N-glycosylase activity. The protein has a Stokes radius of 2.25 nm, a sedimentation coefficient of 2.65 S, a molecular weight of 26,300 in the native state and 27,300 in the denatured state, and a frictional ratio of 1.13.(ABSTRACT TRUNCATED AT 250 WORDS)