The gene coding for the DNA polymerase I from Treponema pallidum, Nichols strain, was cloned and sequenced. Depending on which of the two alternative initiation codons was used, the protein was either 997 or 1015 amino acids long and the predicted protein had a molecular mass of either 112 or 114 kDa. Sequence comparisons with other polA genes showed that all three domains expected in the DNA polymerase I class of enzymes were present in the protein (5'-3' exonuclease, 3'-5' exonuclease and polymerase domains). Additionally, there were four unique insertions of 20-30 amino acids each, not seen in other DNA polymerase I enzymes. Two of the inserts were near the boundary of the two exonuclease domains and the other two interrupted the 3'-5' exonuclease domain which is involved in proofreading. The predicted amino-acid sequence had an exceptionally high content of cysteine (2.4% compared with <0.05% for most other sequenced DNA polymerase I enzymes). The polA gene was further cloned into pProEXHTa for expression and purification. The transformants expressed a protein of 115 kDa. Antibodies raised against synthetic peptide fragments of the putative DNA polymerase I recognised the 115-kda band in Western blot analysis. No DNA synthesis activity could be demonstrated on a primed single-stranded template. Although significant quantities of the protein were produced in the host Escherichia coli carrying the plasmid, it was not capable of complementing a polA(-) mutant in the replication of a polA-dependent plasmid.