Genomic DNA from ribotype-01 and -17 Clostridium difficile strains was used for amplification of the sequences encoding the carboxy-terminal domain of toxins A (TcdA) and B (TcdB). The deduced C-terminal TcdB ribotype-01 and -17 domains share 99.5% amino acid sequence identity while TcdA ribotype-17 comprises a 607 amino acid deletion compared to TcdA-01. When compared to previously sequenced C difficile toxins, 99.3% amino acid identity was found between TcdA-01 and TcdA from strain VPI10643 and 98.8% identity between TcdA-17 and TcdA from strain F-1470. The obtained sequences were fused in 3' to a sequence encoding a hexahistidine tag and cloned into an Escherichia coli expression vector. The recombinant proteins were expressed in E coli and purified using single-step metal-chelate chromatography. The recombinant carboxy-terminal domain of TcdA-01 was purified from the soluble E coli lysate fraction whereas TcdA-17 and TcdB-17 carboxy-terminal domains were purified from inclusion bodies. At least 40mg of each protein was purified per liter of bacterial culture. The recombinant toxin domains were detected specifically by Western blot and ELISA with antibodies against native C difficile toxins. This study demonstrated that the carboxy-terminal domains of TcdA and TcdB can be produced using an E. coli expression system and easily purified. These recombinant, stable, and non-toxic proteins provide a convenient source for use in the diagnosis of C difficile infections, instead of native toxins, as controls and calibrators in immunoassay kits and to obtain specific monoclonal antibodies. (C) 2003 Elsevier Science (USA). All rights reserved.