Our previous results have shown that oxidative stress may reduce the regeneration potential of protoplasts, but only protoplasts that are able to supply extracellularly H₂O₂ can actually divide (C.I. Siminis, A.K. Kanellis, K.A. Roubelakis-Angelakis [1993] Physiol Plant 87: 263–270; C.I. Siminis, A.K. Kanellis, K.A. Roubelakis-Angelakis [1994] Plant Physiol 1105: 1375–1383; A. de Marco, K.A. Roubelakis-Angelakis [1996a] Plant Physiol 110: 137–145; A. de Marco, K.A. Roubelakis-Angelakis [1996b] J Plant Physiol 149: 109–114). In the present study we have attempted to break down the oxidative burst response into the individual active oxygen species (AOS) superoxide (O₂  ^·−) and H₂O₂, and into individual AOS-generating systems during the isolation of regenerating tobacco (Nicotiana tabacum L.) and non-regenerating grape (Vitis vinifera L.) mesophyll protoplasts. Wounding leaf tissue or applying purified cellulase did not elicit AOS production. However, the application of non-purified cellulase during maceration induced a burst of O₂  ^·− and H₂O₂accumulation in tobacco leaf, while in grape significantly lower levels of both AOS accumulated. AOS were also generated when protoplasts isolated with purified cellulase were treated with non-purified cellulase. The response was rapid: after 5 min, AOS began to accumulate in the culture medium, with significant quantitative differences between the two species. In tobacco protoplasts and plasma membrane vesicles, two different AOS synthase activities were revealed, one that showed specificity to NADPH and sensitivity to diphenyleneiodonium (DPI) and was responsible for O₂  ^·−production, and a second NAD(P)H activity that was sensitive to KCN and NaN₃, contributing to the production of both AOS. The first activity probably corresponds to a mammalian-like NADPH oxidase and the second to a NAD(P)H oxidase-peroxidase. In grape, only one AOS-generating activity was detected, which corresponded to a NAD(P)H oxidase-peroxidase responsible for the generation of both AOS.