The nucleotides of the Mn-superoxide dismutase (SOD) gene of Bacillus circulans and the Fe-SOD gene of Aerobacter aerogenes were sequenced by PCR. These SOD genes were specifically amplified by using oligonucleotide primers corresponding to the amino-terminal amino acid sequences and the antisense strand primer corresponding to the common amino acid sequence near the carboxyl-terminus among various Mn- and Fe-SODs thus far sequenced. The PCR products amplified from B. circulans and A. aerogenes genes contained a 486-nucleotide sequence encoding 162 amino acids and a 507-nucleotide sequence encoding 169 amino acids, respectively. Each sequence seemed to contain most of the open reading frame encoding the SOD protein when compared with other sequenced SODs. The two amino acid sequences deduced from the nucleotide sequences of PCR products had an identity of 66.1%. However, the SODs from B. circulans and A. aerogenes were immunologically distinct from each other judging from an immunoprecipitation test. The two SODs had high homologies with other bacterial Mn-SODs, especially the highest homology of 75.4% and 66.7%, respectively, with the B. stearothermophilus Mn-SOD. Genomic Southern hybridization suggested that each PCR product of the bacterial genes that were synthesized and sequenced was the product of the sole SOD gene in each bacterium.