Bacterial diversity across belowground compartments
We observed 11 bacterial phyla across all samples, which represented 94.8% of reads and had abundance estimates of >10% across all plant compartments, treatments and timepoints throughout this experiment (Figure 2a). The soil-only pots had 8 bacterial phyla, which represented 96.9% of reads and had abundance estimates of
Alpha diversity levels across the soils and rhizospheres were both higher than the T. triandra endospheres in the live (permutedANOVA: F(3,75) = 14.26, p <0.001; Figure 3a) and sterilised treatments (permutedANOVA: F(3,63) = 5.824, p = 0.003). Alpha diversity was also higher for soil-only pots (in all treatments) at the beginning of the trial than at harvest (permutedANOVA: F(2,22) = 7.932, p = 0.01), though there were no differences between soils in the sterilised soil-only pots over time (permutedANOVA: F(1,14) = 0.313, p = 0.59).
Sterilisation reduced alpha diversity of all soils at the beginning of the experiment (p<0.001; Table S4; Figure S12a) and these differences persisted until harvest (p<0.001; Table S4; Figure S12). We saw no effect of soil aridity (live p =0.875; sterile p = 0.086; Table S4; Figure S12) or water-stress treatment (live p = 0.312, sterile p =0.840; Table S4; Figure S12) on soil alpha diversity. The soil-only pots also did not vary in alpha diversity between low and high aridity soil conditions or water availability treatment groups (live permutedLMEM: t-value -0.567, p = 0.584, sterile permutedLMEM: t-value 1.159, p = 0.255).
Bacterial communities significantly varied by compartment (i.e., soils, rhizospheres, endospheres) across all treatments (PERMANOVA: F(2,183)= 7.465, R2=0.075, p<0.001; Figure 3b; Figure S13). Sterilisation (p<0.001, p<0.001, and p=0.013, respectively; Figure 4a; Table S5), soil aridity (p<0.001, p<0.001, and p=0.023, respectively; Figure 4a; Table S5), and water stress treatments (p<0.001, p<0.001, and p=0.033, respectively; Figure 4a; Table S5) each affected soil, rhizosphere and endosphere bacterial community composition.
There was no detectable difference between the bacterial community compositions between the soil-only pots to the plant-present pots (PERMANOVA: F(1, 103)= 0.733, R2= 0.007, p=0.755; Figure 4b), but we did observe a difference in communities from the initial sampling to the harvest (PERMANOVA: F(1, 103)= 7.354, R2= 0.066, p<0.001; Figure 4b).
Soil aridity, sterilisation, and water-stress treatments had effects on differential abundance of bacterial phyla across the soils, rhizospheres and endospheres (Figure 1b; Tables S6-S7).
Bacterial alpha diversity in soils, rhizospheres, and endospheres correlated positively and strongly with plant biomass (LMEM: t-value = 10.857, p <0.001; Figure 5). Biomass increased more with bacterial alpha diversity in high aridity soils (Figure 5).