Supplemental Results: Detailed analysis of taxa variation among samples
The microbial community structure of bulk soil samples from this study reflected a nutrient rich environment, with a high ratio of Proteobacteria to Acidobacteria. Proteobacteria are favored in nutrient rich soils, especially fertilized agricultural soils (32).
No OTUs were found in all the samples, but 3 OTUs (97% nucleotide ID) were detected in 75% of the samples. The taxa comprising this ‘core’ microbiota were more abundant belowground and included OTUs associated with Bradyrhizobium, with relative abundances of 0.21-0.87% belowground and 0-0.04% aboveground; Steroidobacter, with a relative abundance of 0.29-0.31% in belowground samples and ~0.01% in aboveground sample fractions; and Acidobacteria-6 OTU (order iii1-15), with relative abundances of 0.31-0.43% and 0-0.03% below and aboveground, respectively.Across vineyard, clone, year, and plant developmental stage,bulk soil samples had 17 OTUs in common, while root zone samples had 15 core OTUs and roots had 10. This core microbiota consisted of OTUs belonging to Acidobacteria (Chloracidobacteria and Acidobacteria-6), Actinobacteria (Psychrolactophilus), Gemmatimonadetes and Xanthomonadales (Steroidobacter) in soil and root zone samples, while Rhizobiales (Bradyrhizobium and Devosia) and Xanthomonadales (Sinobacteraceae and Steroidobacter) were found in 100% of root samples. These values increased to 65, 58, and 42 core OTUs in soil, root zone, and root samples, respectively, when considering OTUs present in 95% of the samples. In aboveground samples, only flowers maintained core OTUs across all samples, with 3 OTUs that belonged to Pseudomanaceae (Pseudomonas and Viridiflava) and Erwinia. Every sample type also had unique OTUs (bulk soil=19, root zone=18, roots=18, grapes=30, flowers=1 and leaves=13), which all had low relative abundance (Table S2).Unique OTUs associated with different plant parts included Cryptosporangiaceae (Actinobacteria), Pseudoalteromonas and Aeromonadaceae (Gammaproteobacteria) in roots; Coriobacteria (Actinobacteria), MCG (Crenarchaeota), Phascolarctobacterium (Firmicutes) in grapes; p-2534-18B5 (Bacteroidetes) in flowers; and Marinicella (Gammaproteobacteria), Arctic97B-4 (Verrucomicrobia), Lentisphaerales (Lentisphaerae) and Gemm-6 (Gemmatimonadetes) in leaves (Table S2).
Many taxa were significantly different among vineyards (Table S4A,C) and the supervised learning classification test proved to be powerful to predict the vineyard of origin of soil and root samples (Table S5). In bulk soil, there were members of Acidobacteria (Koribacteraceae, iii1-15 and RB-41), Bacteroidetes (Chitinophagaceae and Cytophagaceae), Proteobacteria (Hyphomicrobiaceeae, Sphingomonadaceae, Sinobacteraceae) and Verrucomicrobia (Chtoniobaceraceae). While Acidobacteria were the major contributor to differences between all vineyards, Koribacteraceae were the major contributor to the differences at Ruttura. Similarly, RB41 (Chloracidobacteria) abundances distinguished Hearn vineyard samples from the rest, and Cytophagaceae was an important contributor for the differentiation of Sherwood vineyard bulk soil samples (Table S4A). Acidobacteria and Cytophagia correlated with pH (Pearson correlation, r=-0.86 and r=0.51, respectively), as it has been reported previously (32, 34, 35). In roots, Sinobacteraceae contributed to differences among vineyards (Table S4C), while Bradyrhizobiaceae and Koribacteraceae contributed to differences atRuttura. Similar trends in the Acidobacteria and Bacteroidetes community structure were evident in bulk soil and root samples, suggesting soil bacteria as the inoculum, ‘seeding’ the roots with endophytic taxa (10). Aboveground samples did not show such vineyard specificity.
1