Escherichia coli endonuclease III (EC 4.2.99.18) (gene nth) [1] is a DNA repair enzyme that acts both as a DNA N-glycosylase, removing oxidized pyrimidines from DNA, and as an apurinic/apyrimidinic (AP) endonuclease, introducing a single-strand nick at the site from which the damaged base was removed. Endonuclease III is an iron-sulfur protein that binds a single 4Fe-4S cluster. The 4Fe-4S cluster does not seem to be important for catalytic activity, but is probably involved in the proper positioning of the enzyme along the DNA strand [2].

Endonuclease III is evolutionary related to the following proteins:

• Fission yeast endonuclease III homolog (gene nth1) [3].
• Escherichia coli and related protein DNA repair protein mutY, which is an adenine glycosylase. MutY is a larger protein (350 amino acids) than endonuclease III (211 amino acids).
• Micrococcus luteus ultraviolet N-glycosylase/AP lyase which initiates repair at cis-syn pyrimidine dimers.
• ORF10 in plasmid pFV1 of the thermophilic archaebacteria Methanobacterium thermoformicicum [4]. Restriction methylase m.MthTI, which is encoded by this plasmid, generates 5-methylcytosine which is subject to deamination resulting in G-T mismatches. This protein could correct these mismatches.
• Yeast hypothetical protein YAL015c.
• Fission yeast hypothetical protein SpAC26A3.02.
• Caenorhabditis elegans hypothetical protein R10E4.5.
• Methanococcus jannaschii hypothetical protein MJ0613.

The 4Fe-4S cluster is bound by four cysteines which are all located in a 17 amino acid region at the C-terminal end of endonuclease III. A similar region is also present in the central section of mutY and in the C-terminus of ORF10 and of the Micrococcus UV endonuclease. The 4Fe-4S cluster region does not exist in YAL015c.

We developed two signature patterns for these proteins: the first corresponds to the core of the iron-sulfur binding domain, the second corresponds to the best conserved region in the catalytic core of these enzymes.