Apple microbiome-based improvement of biocontrol agents for postharvest pathogens management

Biocontrol using biocontrol agents (BCA) is a sustainable way to manage
postharvest diseases. Nevertheless, their reliability (stability of efficacy) must still be
improved to reach efficacy of synthetic chemical pesticides. Indeed, when BCA are
applied on the plant surface, they meet contrasted microbial community in which
ecological relationships take place, thus potentially interfering with their efficacy and
stability. Understanding better the role of fruit microbiota composition in the disease
development and in the efficacy of BCA is an emerging research area. This could
allow the isolation of more beneficial strains or the identification of molecules that
modulate host microbiota to favour BCA and to limit disease development. In this
thesis, apple microbiota was deeply characterized and used in situ to improve the
efficacy and stability of Pichia anomala strain K, a BCA against Botrytis cinerea rot.
At first, several protocols to harvest the epiphytic microbiota were tested and
compared. Results showed that a first wash harvested one third of culturable
microorganisms, and that three successive washes harvested around 80% of them.
There were significant differences between protocols, thus inducing potential bias
among studies using different protocols to study the epiphytic microbiota. However,
this bias was offset by pooling the successive washes, that besides increased overall
microbial richness.
Then, the metagenome of apple fruit microbiota was studied by shotgun highthroughput sequencing (HTS) on three symptomless samples of apple fruit Pinova
variety. Results showed a very high fungal (79.0 %) and bacterial (13.8 %) diversity.
The prevalent phyla were Ascomycota for fungi and Bacteroides for bacteria. Genera
of known apple postharvest pathogens and known BCA were identified. The
functional analysis showed important genes potentially involved in biocontrol
activity. On the other hand, the incompleteness of current databases and the
complexity of the analysis limited the complete exploitation of the results and are
blockages of using such technic currently.
Finally, contrasted apple microbiomes were harvested and studied to identify
beneficial strains and molecules to control Botrytis cinerea. Eighteen apple microbial
communities (representing fifteen cultivars and four management practices) were
harvested and characterized by HTS of 16s rRNA gene and Internal Transcript Spacer
(ITS) amplicons. The biocontrol efficacy against B. cinerea of these microbiotas
applied alone or with a BCA was also evaluated and the metabolisation of 190 carbon
sources was studied. Results showed that apple microbiota is diversified and
contrasted at taxonomic level, biocontrol efficacy and carbon source metabolisation.
Certain microbiota significantly raised the strain K efficacy up to 100%. The alpha
diversity of the apple microbiota was not correlated to the biocontrol efficacy. Coclustering of Spearman correlations were used to drive the isolation of promising
beneficial strains (BCA candidates or helper strains supporting BCA efficacy) and
molecules allowing better control of the pathogen. Biocontrol efficacy of the new
strains and molecules was further confirmed in vivo. Results of this work can contribute to increase the reliability of BCA in real
condition.