Despite catalyzing the same reaction (2 H2O2-->2 H2O+O2) heme-containing monofunctional catalases and bifunctional catalase-peroxidases (KatGs) do not share sequence or structural similarities raising the question of whether or not the reaction pathways are similar or different. The production of dioxygen from hydrogen peroxide by monofunctional catalases has been shown to be a two-step process involving the redox intermediate compound I which oxidizes H2O2 directly to O2. In order to investigate the origin of O2 released in KatG mediated H2O2 degradation we performed a gas chromatography-mass spectrometry investigation of the evolved O2 from a 50:50 mixture of H2(18)O2/H2(16)O2 solution containing KatGs from Mycobacterium tuberculosis and Synechocystis PCC 6803. The GC-MS analysis clearly demonstrated the formation of (18)O2 (m/e = 36) and (16)O2 (m/e = 32) but not (16)O(18)O (m/e = 34) in the pH range 5.6-8.5 implying that O2 is formed by two-electron oxidation without breaking the O-O bond. Also active site variants of Synechocystis KatG with very low catalase but normal or even enhanced peroxidase activity (D152S, H123E, W122F, Y249F and R439A) are shown to oxidize H2O2 by a non-scrambling mechanism. The results are discussed with respect to the catalatic mechanism of KatG.