Research Associate The University of Texas at Austin Austin, Texas
Body of Abstract: Microbiome breeding aims to produce microbial communities that benefit hosts through artificial selection. It leverages transplantable microbiomes that can be differentially perpetuated between hosts, passing on microbiomes with desirable effects whilst excluding undesirable ones. It is therefore a type of host-mediated artificial selection shaping microbiomes, independently of selection on the host. This sort of breeding technique has been successfully used to enhance growth, flowering, and seed production in several model species, such as Arabidopsis and wheat.
We recently developed a microbiome breeding protocol to generate microbiomes conferring salt tolerance to Brachypodium distachyon, a model system for temperate grasses. Through successive rounds of differential propagation, we artifactually selected microbiomes conferring tolerance to either sodium (Na) or aluminum (Al) stress. Microbiomes bred with this protocol increased plant fitness after only 1-3 rounds of selections. After 9 rounds of selection, we tested the effects of the bred microbiomes on plants experiencing salt stresses in a 2×2 cross-factored experiment. Microbiomes selected to confer tolerance to Al stress provided nonspecific benefits to plants, whilst those selected to confer tolerance to Na stress provided specific benefits. Seed production of B. distachyon plants given artificially selected microbiomes was up to 205% greater than plants given unselected microbiomes.
Here, we report the metagenomic analyses of the microbiomes generated in this experiment. We show that: (a) beneficial microbiome effects are transplantable and heritable between the two salt stress; (b) selection history impacts root assemblages more than the stress plants are currently experiencing; (c) microbial co-occurrence networks became increasingly more complex with selection and stress, with the most complex networks occurring when selection history matched the salt stress; (d) ASVs from particular bacterial families showed a disproportional eigenvector centrality in networks, with some being intimately associated with a particular stress, whilst others appear to be shared between the two stresses.