Structure of a dimeric histidine triad (HIT) superfamily protein from B. anthracis has been solved. Central ten-stranded antiparallel beta sheet of each of two dimers in the asymmetric unit sits sagged on a short V-shaped depression formed by two helices (residues 59-81 in a single helix of every monomer) at the oligomerization interface. Residues 115 through 144 of every HIT protein in the dimer cover the central beta sheet from the top of the depression. The N- and C-terminal helices enclose the external sticking-out edges of the sheet in every dimer. H101, H103 and H105 of the enzyme belong to the three-histidine (H-x-H-x-H-x-x) motif found in other HIT superfamily proteins. The histidines participate in binding of a metal and a nucleotide. The HIT proteins are known to be dinucleoside polyphosphate hydrolases, which require millimolar Mn2+, Mg2+ or Ca2+ for optimal activity. These enzymes are inhibited by Ni2+, Zn2+ and Cd2+ (Huang et al., (1995) Biochem. J. 312: 925). The His101 is a part of the metal-binding site, which consists of additional His50, Cys8 and Cys11. Initially, the site possessed a strong positive difference density peak with coordination number equals four, which disappeared after modeling Zn2+ ion. H103 and H105 are part of the nucleoside/nucleotide-binding area and orient a sulfate ion at the same position, at which the phosphate of GMP in the rabbit HIT protein structure (PDB ID 3RHN) is positioned. Both proteins share 35 % sequence identity and have 1.7 Å rmsd of their C-alpha atoms within 107 residues. The mammalian enzyme lacks the C-terminal part, which bridges two monomers in the bacterial protein. On the contrary, the B. subtilis HIT protein is structurally better aligned with the B. anthracis enzyme (0.8 Å rmsd) having 66 % identity. The subtilis homolog has also zinc bound at the metal binding site.