DNA extraction, sequencing and bioinformatics
We prepared DNA extractions from T. triandra root endospheres following methods outlined in Hodgson et al. (2024b). This involved cleaning the exterior of plant root surfaces by sonication at five 30 s on/off burst cycles in 0.02% Silwet L-77 supplemented PBS buffer (pH = 6.5) for 5 min, followed by five 5 min washes in sterilised, distilled water. These methods underwent prior validation described in Hodgson et al. (2024b). To extract microbial DNA from rhizospheres, we followed the protocol from McPherson et al. (2018). Briefly, root samples were washed in 0.02% Silwet L-77 supplemented PBS buffer, vortexed for 45 min and then filtered using 100 µm sieves (Thermo Fisher Scientific, Waltham, USA) and centrifuged at 1000 RPM. Soils samples taken from the plant plots at either the start or end of the trial were stored at -20 ˚C after collection. DNA from soil, rhizosphere and endosphere samples was extracted using the DNeasy PowerLyzer PowerSoil Kit (Qiagen, Hilden, Germany), as per manufacturer’s instructions.
Amplicon libraries of the 16S rRNA V3-4 gene region were generated by the Australian Genome Research Facility (Brisbane, Australia). Sequences were generated using the 300 base pair paired end run of the Illumina NextSeq 2000 platform. The DADA2 bioinformatics pipeline was used to infer amplicon sequence variants (ASVs), with monotonicity enforced during error estimation. Additionally, loess smoothing was applied with specified weights, span, and degree to improve error rate modelling. Qiime2 was used to identity profiles from amplicon sequence data from the SILVA database (v138.1) (Wang et al. 2007; Quast et al. 2013), using a naïve Bayesian classifier (Wang et al. 2007; Callahan et al. 2016; Bokulich et al. 2018). Taxa that were not assigned as Bacteria, unassigned at the Phylum level, and associated to mitochondria or chloroplasts were removed. We were unable to extract and sequence viable DNA concentrations from sterilised low aridity conditions at the beginning of the experiment, possibly due to the sterilising effects of autoclaving on microbiota and their DNA.