Appalachian Fruit Research Station, USDA Agricultural Research Service, Kearneysville, WV 25430, USA

Background and objectives
Biological control of post-harvest diseases of fruits has advanced greatly during the past decade. Currently, there are two biological control agents registered, a bacterium, Pseudomonas syringae (BioSave 11, BioSave 110), and a yeast Candida oleophila (Aspire) [1]. Biocontrol systems can be improved by expanding efficacy over a broader range of environmental conditions and against a wider array of pathogens. One such improvement has been the development of antagonist mixtures that are superior to the individual antagonists in biocontrol of blue mould (Penicillium expansum) [2]. The superiority of the mixtures over individual antagonists resulted, most likely, from greater reduction of the realized niche of the pathogen by the mixture. Thus, selection of the antagonists for mixtures can be greatly improved if the niche (wound) and antagonist nutritional profile are well characterized. To maximize the beneficial effect of mixtures, an optimal proportion of the antagonists must also be determined. The objectives of this work were: (1) to isolate microorganisms naturally colonizing exposed apple tissue in an orchard and determine their biocontrol potential against P. ;expansum; (2) to characterize major carbon (C) and nitrogen (N) sources in apple wounds; (3) to determine nutritional profile of the antagonists and the utilization kinetics of C and N sources; and (4) to pair antagonists based on complementary utilization of C and N sources and compare biocontrol potential of these mixtures with the individual antagonists against P. ;expansum on apples.

Materiais and methods
Antagonists were isolated from wounded apples in an orchard. Their biocontrol potential against P. ;expansum on apples after harvest was determined and their nutritional profiles with respect to utilization of major C and N sources occurring in apple fruit were characterized. The antagonists were paired based on the complementary nature of the nutritional profiles and biocontrol potential of the mixtures, and the individual antagonists were determined on apples. Optimal proportions of the antagonists in the mixtures were determined using De Wit replacement series.

Results and discussion
Yeasts were the prevailing microorganisms isolated from the fruit and bacteria were isolated only occasionally. Many of these microorganisms had some antagonistic activity against P. ;expansum. The major C sources in wounds of 'Golden Delicious' apples were fructose, sucrose, glucose and sorbitol, and the major N sources were amino acids asparagine, aspartic acid, glutamic acid, serine and alanine. Microorganisms varied widely in utilization of C and N sources and ranged from those able to utilize all major C and N sources that occur in apple wounds to those utilizing only a single C source. All antagonists utilized all major amino acids found in apple. Kinetics studies of the utilization of C and N sources revealed wide differences among antagonists; however, antagonists that utilized one or only a few C sources were frequently more efficient in utilizing these compounds then antagonists utilizing a broad range of C sources. Most of the antagonist mixtures protected apples against P. ;expansum more effectively than the individual antagonists, resulting in lower severity and incidence of the disease.

This study demonstrated that biological control of post-harvest diseases of fruits could be improved by increasing the nutritional niche overlap between antagonists and the pathogen using a mixture of mutually compatible antagonists with complementary catabolism of nutrients occurring in fruit.

1. Janisiewicz WJ, Jeffers SN, 1997. Crop Protection 16, 629-633.