Experiment 1: rhizotron growing systems and sample collection for microbial analysis
A rhizotron system (30 cm long × 18 cm wide × 2 cm thick) with a removable transparent acrylic side was designed for the current experiments. This system was modified from DeAngelis et al (2009) who confirmed its effectiveness to analyse rhizosphere microbiome by keeping healthy plants and avoiding bacterial contamination. Rhizotrons were covered with aluminium foil to simulate dark conditions for root growth, and angled at 45 degrees, with the transparent side as the base, to enable visualisation of roots under the transparent surface (Fig. S1). Five soil types and three chickpea genotypes with 4 replicates were examined in this rhizotron system. Five soils were collected from the cropping regions of South Australia, Australia, and marked as Soil-A (-34.511, 138.697), Soil-B (-34.509, 138.667) and Soil-C (-34.537, 138.690), Soil-D (-34.967, 138.633) and Soil-E (-34.966, 138.631). All the soils used here were from typical Mediterranean type dryland farming systems with 3 to 4 years of cereal crops (wheat or barley) rotated with one season legume (lentil, field pea or faba bean). To limit the excessive presence of Mesorhizobium ciceri (the symbiotic rhizobia of chickpea), soils were chosen where no chickpea was planted in the last decade. The soils differed in chemical and physical properties (Table S1), which enabled the evaluation of stochastic or deterministic processes of host plants in assembling microbiome at different niches. Soil physiochemical properties including total N, C, P, K, Ca, Zn, Mg, Mn and Cu, pH, EC, and texture were analysed using the protocol from Rayment & Lyons (2011), and the results are presented in Table S1. Chickpea cultivars Genesis090, PBA Slasher and Sonali were used. Previous evaluation across multiple environments with diverse photoperiod, seasonal rainfall, and cropping management during chickpea growth (Sadras et al., 2016; Kaloki et al., 2019) have demonstrated that Sonali (genotype EF) required less thermal time to flowering compared with the late flowering genotypes Genesis090 (genotype LF1) and Slasher (genotype LF2). It has been confirmed that an 11-bp deletion in chickpea early flowering 1 (Efl1 ) locus was associated with early flowering performance (Ridge et al., 2017). We analysed the structure ofEfl1 gene by sequencing the whole genomes of the three cultivars in the present study. DNA sequencing and mapping were described in (Li et al., 2017). Briefly, genomic DNA of Genesis090, PBA Slasher, and Sonali was extracted from young leaf tissue using Qiagen DNeasy Plant Mini Kit. Pair-end reads (100 bp) of each genotype were sequenced using Illumina HiSeq 2000 platform. Sequencing reads were filtered, trimmed, and mapped to the reference genome 2.6.3 (http://cicer.info) using SOAP2. Bam files were filtered out for reads with more than 3 mismatches resulting in ~10 x coverage of each genome. Bam files were imported into the software Integrated Genome Viewer (IGV) for visual inspection of the Efl1 gene.
Chickpea seeds were surface-sterilized by soaking in 10% sodium hypochlorite for 10 min, and were then washed with sterile water three times, soaked in 70% ethanol for 20s, and then washed another three times in sterile water. Surface-sterilised seeds were germinated in a petri dish with autoclaved sand and water until the root emerged to a length of 0.5 mm. Two germinated seeds were transplanted to 2 cm depth in each sterilised rhizotron that had been filled with soil at a consistent bulk density of 1.38 Mg m-3. Rhizotrons without plants were used as the bulk soil control.
Plants were grown in a controlled environment room (800 μmol m-2 s-1 light for 12 hours at 20 °C and 12 hours of darkness at 10 °C), and watered to field capacity every two days by weighing using autoclaved sterilised water. The point of observable root tips was marked each day. At 16 days after planting (DAP), roots with rhizosphere soil were harvested. Root segments between the markers of 1 to 2 DAP and 15 to 16 DAP were cut off with approximately 1.0 mm of soil adhering around the root; these were regarded as the basal and apical zones, respectively. Root segments from two plants in each rhizotron were processed for microbial community analysis. Rhizosphere and endosphere samples were isolated from root sections using a method modified from Bulgarelli et al. (2012). The sampled root segments with attached soil were transferred into 15 mL sterile PBS buffer in a 50 mL Falcon tube, which was placed on an orbital shaker for 20 min and rotated at 180 rpm. Roots were collected using sterilised tweezers for further processing. The rest of the soil suspension was centrifuged at 180 rpm for 20 min. After removing the supernatant, the pellet was defined as rhizosphere soil. Roots were washed again on the shaker for 20 min at 180 rpm, and then transferred into a new 50 mL Falcon tube with 40 mL sterile PBS buffer. The roots in the Falcon tube were sonicated for 10 cycles of 30 s at 43 kHz and 120 W (model 160HD, Soniclean, Australia) with a 30 s break. The microbiome extracted from the sonicated root was considered to be from the endosphere. The rhizosphere and endosphere samples were stored in a -80 °C freezer.
The total DNA of the rhizosphere soil, cleaned root tissue, and bulk soil was extracted using a PowerSoil DNA isolation kit (Mo Bio, Carlsbad, CA, USA) based on the manufacturer’s instructions. Afterwards, the V3-V4 region of 16S rRNA genes was amplified using the forward primer 341F (CCTAYGGGRBGCASCAG) and reverse primer 806R (GGACTACNNGGGTATCTAAT). The PCR products were sequenced by Australian Genome Research Facility on an Illumina MiSeq platform with 300 bp paired-end reads following the manufacturer’s instructions.
Quantitative PCR (qPCR) was performed to determine the quantity of bacterial DNA in the samples, using the FemtoTMBacterial DNA Quantification Kit (Zymo Research, USA) following the manufacturer’s instructions. The universal bacterial 16S rRNA gene was amplified using the Femto Bacterial qPCR premix on the CFX connect Real-Time PCR detection system (Bio-Rad, USA), as follows: initial denaturation at 95°C for 10 min, 40 cycles of 95°C for 30 s, 50°C for 30 s and 72°C for 60 s, and final elongation at 72°C for 7 min. Gene copy numbers were estimated using E. coli strain JM109 as an internal standard. Three technical replicates per sample were analysed. Data were calculated as copy number per gram of the soil or root tissue used for DNA extraction.