Sampling of green foxtail leaves
We selected seven sites to collect leaf samples of Setaria viridis (green foxtail) grown in natural conditions according to the distribution of green foxtail in China38. Within each site, six patches, spaced 5–10 km apart, were choses as replicates. Ten neighboring green foxtail plants at the heading stage within each path were selected and pooled as one leaf sample (replicate). In total, there were 42 leaf samples (7 sites × 6 replicates) and 42 bulk soil samples. Samples were stored at -80°C and used for DNA extraction (Fig. 1A).
Soil physicochemical analysis
Bulk soil samples were collected for soil physicochemical analysis, including pH, soil organic carbon (SOC), total phosphorus (TP), total nitrogen (TN), total potassium (TK), total carbon (TC), nitrate nitrogen (NO3--N), ammonium nitrogen (NH4+-N), available phosphorus (AP), and available potassium (AK) (Table S1), following standard analysis methods described by Bao 202039. Climate data, including mean annual temperature (MAT), accumulated temperature (AT), relative humidity (RH), mean annual precipitation (MAP), accumulated sunshine duration (AS) and annual cumulative solar radiation (AR) were sourced from China Meteorological Data Network (http://data.cma.cn/) and are detailed in Table S1.
DNA extraction of leaf samples
To avoid soil particles contamination, all leaf samples were washed once with sterile deionized water, and dried with sterile filter paper. Subsequently, 100 mg of fresh leaf tissue was weighed in a 2 ml tube with two sterile steel balls (diameter = 4 mm). The tube was immersed in liquid nitrogen for 1 min, and then ground for 2 min at 1600 rate with 2 cycles using SPEX Geno/Grinder 2010 for tissue lysis. Leaf DNA was extracted according to the instructions of DNeasy® PowerSoil® Pro Kit (QIAGEN, Germany; Part Number: 47016). The quality of extracted DNA was determined by 1% agarose gel electrophoresis and the DNA concentration and purity were measured using a NanoDrop 1000 spectrophotometer (Thermo Scientific). Leaf DNA samples were stored at -80°C for subsequent sequencing of 5S rDNA, knotted1 (kn1), 16S rRNA genes and internal transcribed spacer region (ITS).
Sanger sequencing of 5S rRNA and knotted1 plant genes
To confirm species identification as green foxtail, specific primers targeting housekeeping genes 5S rRNA (FP: 5’-GGACCTCCTGCGAAG TCCT-3’ and RP: 5’-CCCATCCGTGTACTACTCTC-3’) and kn1 (kn211-402F: 5’-TCAGAACTTTTGGCCGTGGGT-3’ and kn612-402R: 5’-GAGATGGACAGCGAGTTGAGC-3’) genes were amplified and sequenced via Sanger sequencing40. PCR amplification was performed in a 20-µL system, containing 2 µL of 10× Ex Taq Buffer, 1.6 µL of 2.5 mM dNTPs, 1 µL of each primer (5 µM), 0.2 µL of 5-unit Ex Taq Polymerase, 100 ng of template DNA, and 3.7 µL double-distilled H2O. The PCR conditions were as follows: initial denaturation at 95°C for 5 min; followed by 25 cycles of denaturation at 95°C for 30 s, annealing at 56°C for 30 s, and extension at 72°C for 1.5 min; and a final extension step at 72°C for 10 min. PCR products were purified from a 1.5% (w/v) agarose gel after electrophoresis. Further purification and quantification were performed using an AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, USA) and a QuantiFluor™-ST Fluorometer (Promega, Fitchburg, USA) following the manufacturers’ instructions. Purified amplicons were subjected to Sanger sequencing at Majorbio (Shanghai, China). Phylogeny of 5S rDNA and kn1 gene sequences was performed using the command “qiime phylogeny align-to-tree-mafft-fasttree” in QIIME 2 (Quantitative Insights Into Microbial Ecology 2, version 2021.2) and viewed in iTOL41. Sixty 5S rDNA and kn1 housekeeping gene sequences were used in the phylogenetic analysis, including 42 samples obtained and sequenced in this study, 14 samples from different Setaria genera as documented by Zhao et al., (2013)40, 3 established S. viridis cultivars (Q23, Q19, Q4) from He et al., (2023)3 and an outgroup Zea Mays (GenBank: DQ351339).
Whole genome sequencing of green foxtail
The sequencing library was prepared using the NEB Next® Ultra™ DNA Library Prep Kit for Illumina (NEB, USA) following the manufacturer's guidelines, with index codes added to each sample. Genomic DNA was sonicated to a size of 350 bp, and the resulting fragments underwent end-polishing, A-tailing, and ligation with the Illumina full-length adapter. Subsequent PCR amplification was carried out. After purification using the AMPure XP system (Beckman Coulter, Beverly, USA), DNA concentration was measured using the Qubit® 3.0 Fluorometer (Invitrogen, USA). Library size distribution was analyzed with the Agilent 2100 Bioanalyzer, and quantification was done by real-time PCR (> 2 nM). Index-coded samples were clustered using the cBot Cluster Generation System with the Illumina PE Cluster Kit (Illumina, USA). Following cluster generation, the DNA libraries underwent sequencing on the Illumina platform at Novogene, producing 150 bp paired-end reads. Raw data, in the form of short reads in FASTQ format, were obtained through base calling from the original fluorescence image files.
To ensure reliable downstream bioinformatics analysis, quality control measures were applied, including discarding paired reads containing adapter contamination (> 10 nucleotides aligned to the adapter with ≤ 10% mismatches), discarding reads with more than 10% uncertain bases, and discarding reads with over 50% low-quality bases (Phred quality < 5) through using fastp42. Valid sequencing data were then mapped to the Setaria viridis cultivar A10 reference genome (Accession: GCA_005286985) using the Burrows-Wheeler Aligner (BWA) software43, and the original mapping results were stored in BAM format (parameter: mem -t 4 -k 32 -M). Duplicate reads were removed using SAMtools (parameter: rmdup)43 and Genome Analysis Toolkit (GATK v4.1.4)44. Raw SNPs (singlenucleotide polymorphisms) sets were called by SAMtools (parameters: mpileup -m 2 -F 0.002 -d 1000)45. These sets were filtered based on criteria such as a depth of the variate position > 4 and a mapping quality > 2. Biallelic SNPs with a missing frequency < 10% and a minor allele frequency > 0.05 were retained for phylogenetic analysis and correlation analysis. A total of 12,849,803 SNPs were initially obtained, and 3,297,733 SNPs were retained after applying quality filtering with a minimum allele frequency (MAF) threshold of 0.05.
The dist() function in FactoMineR46 was used to generate a dissimilarity distance matrix of SNP with the Euclidean distance method. The optimal number of clusters was determined as 5 through the gapstat() function within the R package factoextra47. For dendrogram construction, we utilized the R package dendextra48. The vegdist() function in the vegan package was used to generate the distance matrices of SNP data (Euclidean) and microbial community composition (bray)49. The mantel() function was used for the mantel test to determine the significance of Pearson’s correlation between the SNP data and the community composition25.
High-throughput sequencing of 16S rRNA gene and internal transcribed spacer (ITS) region
Amplicon libraries for bacteria and fungi were generated using a two-step PCR approach. The hypervariable V5-V7 region of the bacterial 16S rRNA gene was amplified with the primer pair 799F (5’-AACMGGATTAGATACCCKG-3’)/1193R (5’-ACGTCATCCCCACCTTCC-3’), and the ITS1 region was amplified with primers ITS1-F (5’-CTTGGTCATTTAGAGGAAGTAA-3’)/ITS2 (5’-GCTGCGTTCTTCATCGATGC-3’)11, 50. Each 30-µL PCR reaction contained 15 µL of 2 × Phusion® High-Fidelity PCR Master Mix (New England Biolabs), 3 µL of 2 µM forward and reverse primers, 10 µL of DNA template, and was adjusted to 30 µL with ddH2O. PCR conditions were as follows: initial denaturation at 98°C for 1 min; followed by 30 cycles of denaturation at 98°C for 10 s, annealing at 50°C for 30 s, and extension at 72°C for 30 s; and a final extension step at 72°C for 5 min. The PCR products were extracted from a 2% (w/v) agarose gel after electrophoresis, then mixed and purified with GeneJET Gel Extraction Kit (Thermo Scientific) according to the manufacturers’ instructions. Sequencing libraries were generated with the Illumina TruSeq DNA PCR-Free Library Preparation Kit (Illumina, USA) following the manufacturer’s instructions, with added index codes. Library quality was assessed using the Qubit 2.0 Fluorometer (Thermo Scientific) and Agilent Bioanalyzer 2100 system. Finally, libraries were sequenced on an Illumina NovaSeq platform at Novogene Bioinformatics Technology Co. Ltd (Beijing, China), generating 250 bp paired-end reads.
Quality filtering and statistical analysis of sequencing data
The raw amplicon sequencing data were analyzed using QIIME 2 and the plugins51. DADA2 was employed to obtain amplicon sequence variants (ASVs) by removing low-quality reads, primers, denoise, and chimeras52. Bacterial and fungal ASVs were annotated based on the Silva (Silva (arb-silva.de; v13.8) and UNITE (https://unite.ut.ee, v8.3) databases, respectively53, 54. ASVs belonging to Archaea, Unassigned, Eukaryota, Mitochondria and Chloroplast were removed. Subsequently, 1953 bacterial ASVs with 1,476,076 high-quality reads and 474 fungal ASVs with 3,041,190 high-quality reads were obtained from 42 leaf samples, respectively. The number of reads is sufficient to fully capture the bacterial and fungal diversity (Fig. S12). Downstream analyses were performed in R (http://www.r-project.org, v4.0.3). Alpha diversity was analyzed based on Shannon index, Richness and Faith's Phylogenetic Diversity (PD_whole_tree). ASVs table were normalized using the cumulative-sum scaling (CSS) method prior to β-diversity analysis and subsequent analyzes55. The Bray-Curtis distance between samples was calculated with the “vegan” package. Principal coordinates analysis (PCoA) based on Bray-Curtis distance was performed and the significance of site on community dissimilarity was tested based on analysis of similarities (ANOSIM) and permutational multivariate analysis of variance (PERMANOVA or ADONIS)56. Spearman’s rank correlation (r < − 0.7 and r > 0.7, P < 0.01) was used for identifying the collinearity between environmental variables via corr.test function (“psych” package). The correlated variables were grouped into explanatory groups. Within each group, the variable with the highest R2 in univariate analyses (the best explanatory variable) was used in the model (Fig. 1D, Table S2)57. The niche.width() function (“spaa” packages) was used to calculate the niche breadth at the community level (Fig. 3D) and of dominant genera (Fig. 4B). The corr.test function (“psych” package) was also used for the correlation of core genera (Spearman’s rank correlation, P < 0.05; Fig. 4C). Non-parametric statistical test (Kruskal-Wallis test or Wilcoxon test) was used to evaluate the difference of alpha diversity, and taxonomical composition among sites with the “ggpubr” package. P-values were adjusted by using the Benjamini and Hochberg false discovery rate (FDR) test58. The visualization of boxplot, histogram and scatter plot as well as heatmap were based on the “ggplot2” and “pheatmap” packages, respectively. The relative importance of species sorting and dispersal limitation was determined through variation partitioning analysis. Community variation, based on pairwise Bray–Curtis dissimilarity, was separated into environmental and spatial effects59. Environmental and spatial variables were selected using principal coordinates analysis of neighbor matrices and the ordiR2step() function in the “vegan” package in R49, 60, 61. Pure environmental variation without a spatial component ([E|S]) and pure spatial variation without an environmental component ([S|E]) represent the effects of species sorting and dispersal limitation, respectively23, 61. The relative importance of species sorting versus dispersal limitation was calculated as the ratio ([E|S]/[S|E]). Stochastic processes in community assembly were evaluated through a neutral community model62, where the estimated migration rate (m) measures dispersal limitation. Higher m values indicate less dispersal limitation in microbial community assembly62. R2 reflects the fit of the parameter based on nonlinear least-squares fitting63. Phylogenetic bin-based null model analysis was used to calculate community (only for bacteria) assembly processes and the relative importance of individual lineages (bins) in community assembly using the “iCAMP” package64.
Identification and culture of bacterial and yeast isolates from green foxtail leaves
To obtain culturable bacterial and yeast taxa from leaves, leaf homogenate was prepared via grounding 200 mg leaf tissue in 1 ml sterile 10 mM MgSO4 buffer with two 4-mm steel balls. A serial dilution to up to 10− 5 in 100 µl was made, and 100-µl diluted leaf homogenate was spread onto 90-mm petri dishes containing selective media. For the isolationof bateria, R2A medium (1 L: 0.5 g of yeast extract, 0.5 g of peptone, 0.5 g of glucose, 0.5 g of soluble starch, 0.3 g of sodium pyruvate, 0.024 g of MgSO4, 0.3 g of K2HPO4, 0.5 g of casamino acids and 15 g of agar) was used. For the isolation of yeasts, YPD medium (Yeast Extract Peptone Dextrose Medium; 1 L: 20 g of peptone, 10 g of yeast extract, 20 g of glucose and 15 g of agar) with pH 3.5 and 5.5 were used. Petri dishes were incubated for 3–5 days at 25°C. Pure bacterial isolates were growth in LB broth (1 L: 5 g of yeast extract, 10 g of tryptone, and 10 g of NaCl) and stored in 40% at -80°C. Pure yeast isolates were grown in YPD broth for 1–2 days and stored in 15% glycerol (v/v) at -80°C. A total of 164 bacterial isolates and 46 yeast isolates were selected, based on colony morphology, and Sanger sequenced for taxonomical identification. 16S rRNA genes of the bacterial isolates and ITS region of yeast isolates were amplified using the universal primers 27F (5’-AGAGTTTGATCMTGGCTCAG-3’) and 1492R (5’-CGGTTACCTTGTTACGACTT-3’) and ITS1 (5’-GTTTCCGTAGGTGAACTTGC-3’) 65 and ITS4 (5’-TCCTCCGCTTATTGATATGC-3’)66, respectively. For bacterial isolates, a 50-µl PCR mixture system including 25 µl of 2×T5 Super PCR Mix (Colony) (Beijing Tsingke Biotechnology Co., Ltd.), 2 µl of 10-µM forward and reverse primer, 2 µl of culture and 19 µl of ddH2O. The following PCR conditions were used for bacterial 16S rRNA gene amplification: initial denaturation at 98°C for 10 min; 30 cycles of denaturation (10 s at 98°C), annealing (15 s at 56°C), and extension (50 s at 72°C); and a final extension at 72°C for 5 min. Yeast colony rapid detection kit (CAT: RY8001; Nanjing Ruiyuan Biotechnology Co., Ltd.) was used for yeast isolates colony PCR. Briefly, the yeast colony was picked into the PCR tube filled with 3.5 µl of Yeast lysis buffer. After mixing, heat at 98°C for 10–20 minutes to obtain template of PCR. A 50-µl PCR mixture system including 25 µl of 2×yeast PCR mix, 2 µl of 10-µM forward and reverse primer, 1 µl of template and 20 µl of ddH2O. The following PCR conditions were used for bacterial 16S rRNA gene amplification: initial denaturation at 98°C for 10 min; 30 cycles of denaturation (10 s at 98°C), annealing (15 s at 56°C), and extension (50 s at 72°C); and a final extension at 72°C for 5 min. The PCR products were purified with a GeneClean Turbo kit (Qbiogene Inc., Irvine, CA) according to the manufacturer's suggested protocol. The PCR products were used for Sanger sequencing (Beijing Tsingke Biotechnology Co., Ltd.). To identify unique isolates, BOX-PCR was performed using the primers BOX-A1R (5'-CTACGGCAAGGCGACGCTGACG-3')67.
Antagonistic activities against Alternaria alternata on leaves
To verify that SynComs can effectively protect plants against the pathogens Alternaria alternata BXL, three SynComs were designed and included: 1) core bacterial isolates (BacSynCom: (16 Bacillus, 6 Pantoea, and 1 Methylobacterium isolates), 2) core yeast isolates (YeastSynCom: 13 Filobasidium, 8 Vishniacozyma, and 2 Sporobolomyces isolates), and 3) core bacterial and yeast isolates (BacYeastSynCom: all 46 isolates from SynComs 1 and 2). These SynComs were used for the detached leaf assays. Seeds of the foxtail millet cultivar C31 were surface-sterilized by soaking in sodium hypochlorite for 10 minutes and rinsed with sterile water five times. Sterilized seeds were grown to the 2-leaf stage in sterile vermiculite under controlled conditions at 25℃ with a 16 h light/8 h dark photoperiod and 60% relative humidity. The seedlings were then transplanted into pots (210 mm × 160 mm) filled with peat (Kleisman, Germany) and grown to the 5-leaf stage. Leaf sections with 8-cm in length from the 5th leaf were excised and placed a square Petri dish (90 mm × 90 mm) containing 1.5% plant agar with pH 768. Two wounds were punctured in the center of each leaf section with a sterile toothpick. Bacterial and yeast isolates were cultured in LB and YPD broth overnight at 25°C with 180 rpm, respectively. Cells were washed using water three times by centrifugation at 5, 000 rpm for 5 min and then resuspended in 0.025% Tween 20. Isolates with the same OD600nm value were mixed in equal proportions and the final concentration of each isolate was identical in each synthetic community (the concentration of the synthetic community was the concentration of each isolate * isolates numbers). Subsequently, 5-µl suspensions with the SynComs (OD600nm = 0.01 per isolate) were applied on the wounds. The negative group received 5-µl 0.025% Tween 20. After 48 h, 5 µl A. alternata BXL at concentrations of 107 spore/ml was applied at the same spot. To maintain high humidity, 2 ml of sterile water was added to each plate. The plates were sealed and incubated at 25°C under a 16 h light/8 h dark cycle in a growth chamber for 7 days, following the protocol by Perochon and Doohan (2016)68.
Antifungal activity via in vitro dual culture assay
Antagonistic activities the bacterial and yeast strains against the fungal pathogens A. alternata (BXL, DJQ), Epicoccum nigrum (ACCC 37858) and Pyricularia grisea (hlj-2) were assed as previously described by Gouka et al., (2022)69. BXL (isolated from Astragalus membranaceus roots with root rot disease), DJQ (isolated from Nicotiana tabacum leaves with brown spot), ACCC 37858 (purchased from Agricultural Culture Collection of China) and hlj-2 (isolated from S. italica leaf with blast disease) were incubated on PDA plates at 25°C for 9 d, 15 d, 20 d and 10 d, respectively. Briefly, a mycelial plug (diameter of 5 mm) was placed at the center of a 90-mm-petri dish. Four 2.5-µl droplets of microbial solution (OD600nm = 0.5) were spotted symmetrically at four points away from the mycelial plug 2.5 cm, respectively. Four 2.5-µl droplets of LB (for bacterial isolates) and YPD (for yeast isolates) broth were as control group. Petri dishes were incubated at 25°C. Inhibition percentage of mycelium growth was calculated as [(areacontrol-areatreatment)*100/areacontrol].