BACKGROUND: Bioinformatic analysis of the genes coding for the chitinase in Pyrococcus furiosus and Thermococcus kodakarensis revealed that most likely a one nucleotide insertion in Pyrococcus caused a frame shift in the chitinase gene. This splits the enzyme into two separate genes, PF1233 and PF1234, in comparison to Thermococcus kodakarensis. Furthermore, our attempts to grow the wild type strain of Pyrococcus furiosus on chitin were negative. From these data we assume that Pyrococcus furiosus is most likely unable to use chitin as a carbon source. The aim of this study was to analyze in vivo if the one nucleotide insertion is responsible for the inability to grow on chitin, using a recently described genetic system for Pyrococcus furiosus. RESULTS: A marker-less genetic system for Pyrococcus furiosus was developed using simvastatin for positive selection and 6-methylpurine for negative selection. Resistance against simvastatin was achieved by overexpression of the hydroxymethylglutaryl coenzyme A reductase gene. For the resistance to 6-methylpurine the hypoxanthine-guanine phosphoribosyltransferase gene was deleted. This system was used to delete the additional nucleotide at position 1006 in PF1234. The resulting chitinase in the mutant strain was a single subunit enzyme and aligns perfectly to the enzyme from Thermococcus kodakarensis. A detailed analysis of the wild type and the mutant using counted cell numbers as well as ATP and acetate production as growth indicators revealed that only the mutant is able to use chitin as a carbon source. An additional mutant strain containing a reduced chitinase version containing just one catalytic and one chitin-binding domain showed diminished growth on chitin in comparison to the mutant containing the single large enzyme. CONCLUSIONS: Wild type Pyrococcus furiosus is most likely unable to grow on chitin in the natural biotope due to a nucleotide insertion which separates the chitinase gene into two ORFs, whereas a genetically engineered strain with the deleted nucleotide is able to grow on chitin. The overall high sequence identity of the two chitinases between P. furiosus and T. kodakarensis indicates that this mutation occurred very recently or there is still some kind of selection pressure for a functional enzyme using programmed +/-1 frameshifting.