We detected the generation of the reactive oxygen species (ROS) superoxide anion ( O·^- ₂) and hydrogen peroxide (H₂O₂) in apple wounds 2 immediately after wounding, and assessed the relationships between (i) timely colonization of apple wounds by biocontrol yeasts, (ii) resistance of these microorganisms to oxidative stress caused by ROS, and (iii) their antagonism against postharvest wound pathogens. We analyzed a model system consisting of two yeasts with higher (Cryptococcus laurentii LS-28) or lower (Rhodotorula glutinis LS-11) antagonistic activity against the postharvest pathogens Botrytis cinerea and Penicillium expansum. LS-28 exhibited faster and greater colonization of wounds than LS-11. In contrast to LS-28, the number of LS-11 cells dropped 1 and 2 h after application, and then increased only later. In vitro, LS-28 was more resistant to ROS-generated oxidative stress. The combined application of biocontrol yeasts and ROS-deactivating enzymes in apple wounds prevented the decrease in number of LS-11 cells mentioned above, and enhanced colonization and antagonistic activity of both biocontrol yeasts against B. cinerea and P. expansum. Polar lipids of LS-11 contained the more unsaturated and oxidizable α-linolenic acid, which was absent in LS-28. Resistance to oxidative stress could be a key mechanism of biocontrol yeasts antagonism against postharvest wound pathogens.

Contamination of apples (Malus domestica) and derived juices with fungicide residues and patulin produced by Penicillium expansum are major issues of food safety. Biocontrol agents represent an alternative or supplement to chemicals for disease control. Our data show that these microbes could also contribute to actively decreasing patulin accumulation in apples. Three biocontrol agents, Rhodotorula glutinis LS11, Cryptococcus laurentii LS28, and Aureobasidium pullulans LS30, were examined for their in vitro growth in the presence of patulin and for their capability to decrease mycotoxin recovery from the medium. Strain LS11 yielded the highest growth rates and the greatest decrease of toxin recoveries. Further, it caused the appearance of two major spots on thin-layer chromatography (TLC) plates, suggesting possible metabolization of the mycotoxin. In vivo, i.e., in the low percentage of LS11-pretreated apples infected by P. expansum, patulin accumulation was significantly lower than in nontreated infected fruits. Yeast cells survived and increased in infected apples and, in a model system emulating decaying apple, resulted in accelerated breakdown of patulin and the production of the same TLC spots as those detected in vitro. These data suggest that biocontrol yeast cells surviving in decaying apples could metabolize patulin and/or negatively affect its accumulation or synthesis. To our knowledge, this is the first report describing the effect of a biocontrol agent on patulin accumulation in vivo.