Understanding cultivar-specificity and soil determinants of the cannabis microbiome

Abstract

Understanding microbial partnerships with the medicinally and economically important crop Cannabis has the potential to affect agricultural practice by improving plant fitness and production yield. Furthermore, Cannabis presents an interesting model to explore plant-microbiome interactions as it produces numerous secondary metabolic compounds. Here we present the first description of the endorhiza-, rhizosphere-, and bulk soil-associated microbiome of five distinct Cannabis cultivars. Bacterial communities of the endorhiza showed significant cultivar-specificity. When controlling cultivar and soil type the microbial community structure was significantly different between plant cultivars, soil types, and between the endorhiza, rhizosphere and soil. The influence of soil type, plant cultivar and sample type differentiation on the microbial community structure provides support for a previously published two-tier selection model, whereby community composition across sample types is determined mainly by soil type, while community structure within endorhiza samples is determined mainly by host cultivar.

Figure 1. PCoA plots of microbial community similarity in first experiment for unweighted analysis (A–B) and weighted analysis (C–D).

Plots for unweighted analysis are based on unweighted UniFrac distance, and demonstrate relationship between sample type (A), strain (B), and the major PC axes (PC 1 = 26.46% variance, PC 2 = 7.36% variance). Plots for weighted analysis are based on weighted UniFrac distances, and demonstrate relationship between sample type (C), strain (D), and the major PC axes (PC 1 = 62.98% variance, PC 2 = 15.43% variance). Abbreviations for strains are denoted by B (Burmese), BK (BooKoo Kush), and D (Sour Diesel).

Figure 2. PCoA plots of microbial community similarity in second experiment for unweighted analysis (A–C) and weighted analysis (D–F).

Plots for unweighted analysis are based on unweighted UniFrac distances, and demonstrate relationship between soil type (A), sample type (B), strain (C), and the major PC axes (PC 1 = 32.06% variance, PC 2 = 11.34% variance, PC 3 = 5.67% variance). Plots for weighted analysis are based on weighted UniFrac distances, and demonstrate relationship between soil type (D), sample type (E), strain (F), and the major PC axes (PC 1 = 34.51% variance, PC 2 = 25.41% variance, PC 3 = 19.31% variance). Note that PC 1 in the unweighted analysis is dominated by variation in soil type (A), but PC 1 in weighted analysis is dominated by strain (F). Grey points (Fig. 2c, 2f) represent bulk soil samples that aren't associated with either strain. Abbreviations for strains are denoted by MW (Mauie Wowie) and WW (White Widow), and abbreviations for soil type are denoted by MB (Mo-Bio soil) and OC (Orange County soil).

Figure 3. PCoA plots of microbial community similarity in pooled experiments for unweighted analysis (A–C) and weighted analysis (D–F).

Plots for unweighted analysis are based on unweighted UniFrac distance, and demonstrate relationship between soil type (A), sample type (B), strain (C), and the major PC axes (PC 1 = 13.27% variance, PC 2 = 10.15% variance, PC 3 = 6.15% variance). Plots for weighted analysis are based on weighted UniFrac distances, and demonstrate relationship between soil type (D) sample type (E), strain (F), and the major PC axes (PC 1 = 37.69% variance, PC 2 = 13.95% variance, PC 3 = 11.07% variance). Abbreviations for strains are denoted by B (Burmese), BK (BooKoo Kush), D (Sour Diesel), MW (Mauie Wowie) and WW (White Widow). Abbreviations for soil type are denoted by MB1 (Mo-Bio soil from the first experiment), MB2 (Mo-Bio soil from the second experiment) and OC (Orange County soil).

Figure 4. Ternary plot of distribution of bacterial taxonomic groups among sample types in the second experiment.

Size of circles proportional to the log of the total abundance, taxonomic groups are all phylum-level, except for Proteobacteria, which is by class.

Figure 5. Box plots of beta-diversity distances between communities for both weighted and unweighted analyses.

Initials (i.e. B vs. C) stand for comparisons of beta-distances for samples within groups (R =  rhizosphere, C =  Cannabis endorhiza, B =  bulk soil).

Figure 6. Box plots of alpha diversity (observed species) for endorhiza, rhizosphere, and bulk soil from two separate soil types in the second eperiment.

MB  =  Mo-Bio soil, OC  =  Orange County soil. Note the significant differences between alpha diversity in the bulk soil and rhizosphere but negligible differences between endorhiza alpha diversity between soil types.