Heritability of tomato rhizobacteria resistant to Ralstonia solanacearum

Background Ralstonia solanacearum (Rs) is a soilborne phytopathogen that causes bacterial wilt and substantial yield losses in many plants, such as tomatoes. A resistant tomato cultivar can recruit a beneficial microbiome from soil to resist Rs. However, whether this recruitment is inheritable from resistant parent to progeny has not been determined. Results In the present study, we investigated the rhizosphere microbiomes of tomatoes with clear kinship and different resistance against Rs. Resistant tomatoes grown with the additions of natural soil or its extract showed lower disease indexes than those grown in the sterile soil, demonstrating the importance of soil microbiome in resisting Rs. The results of 16S ribosomal RNA gene amplicon sequencing revealed that the resistant cultivars had more robust rhizosphere microbiomes than the susceptible ones. Besides, the resistant progeny HF12 resembled its resistant parent HG64 in the rhizosphere microbiome. The rhizosphere microbiome had functional consistency between HF12 and HG64 as revealed by metagenomics. Based on multi-omics analysis and experimental validation, two rhizobacteria (Sphingomonas sp. Cra20 and Pseudomonas putida KT2440) were enriched in HF12 and HG64 with the ability to offer susceptible tomatoes considerable protection against Rs. Multiple aspects were involved in the protection, including reducing the virulence-related genes of Rs and reshaping the transcriptomes of the susceptible tomatoes. Conclusions We found promising bacteria to suppress the tomato bacterial wilt in sustainable agriculture. And our research provides insights into the heritability of Rs-resistant tomato rhizobacteria, echoing the inheritance of tomato genetic material. Video Abstract Supplementary Information The online version contains supplementary material available at 10.1186/s40168-022-01413-w.


Tomato resistance test 15
To verify the tomato resistance against Rs GMI1000 and the importance of the soil 16 microbiome, we performed a pot experiment in the greenhouse. Tomato seeds were germinated 17 on a plate for 7 d at room temperature, and each seedling was transferred into a pot with about 18 100 g sterile nursery soil or a pot of the same size with about 200 g natural soil (equal volume of 19 field soil and sterile nursery soil). Two to three weeks later, 20 mL pathogen Rs GMI1000 20 suspensions that were diluted to an optical density (OD) of 1.0 at 600 nm using dH2O were 21 poured into each tomato to soak the base of them. To further validate the importance of the soil microbiome, we inoculated natural soil extract back into the sterile nursery soil, then transferred 23 the resistant cultivars HF12 and HG64 into it. The soil extract was prepared according to [1] by 24 mixing soil and water with a ratio of 1:10 and standing for 5 min. The upper suspensions were 25 considered as the soil extract. Each pot was inoculated with 30 mL soil extract, and the control 26 pot was inoculated with tap water. Two to three weeks later, inoculating pathogen Rs GMI1000 27 as described above. Each treatment contained 4-5 tomatoes, and experiments were replicated at 28 least three times. Disease progression was recorded till the disease symptoms were stable using 29 the disease index: 0, no disease symptoms; 1, 1-25% of leaves wilted; 2, 26-50% of leaves wilted; 30 3, 51-75% of leaves wilted; 4, 76-100% of leaves wilted [2]. A test of normality was conducted 31 using shapiro.test() function in R [3], and visualizations were performed using the ggpubr 32 package [4]. Statistical significance was tested by the Wilcoxon test. 33

Rhizosphere soil collection 34
To minimize the influences of the surface microbiome, seeds used for microbiome 35 experiments were surface sterilized as follows: 10% NaClO, 5 min; 70% ethanol, 5 min; wash 36 five times with sterile dH2O. Then the seeds were germinated on the water agar media (1.5%-2% 37 agar per liter of water). About two weeks after transplantation into natural soil, tomatoes with 38 similar growth status were selected for T1 sampling time point rhizosphere soil collection. The 39 collection method was referenced to Xu et al. [5]. After squeezing the potting soil, pull out the 40 plants and gently shake off bulk soil. Then an appropriate number and length of roots were cut 41 with sterilized scissors and put in 20 mL of 1× phosphate buffered saline (PBS: Na2HPO412H2O, centrifuge tube. After vortexing for 30 s, remove roots and centrifuge at 3500 rpm for 5 min to get the resulting rhizosphere soil. In the meantime, we inoculated Rs GMI1000 suspensions into 45 the remaining tomatoes to exert challenge and explore the response of different cultivars. Control 46 groups were inoculated with an equal volume of sterile dH2O. Five and ten days later, 47 rhizosphere samples were collected as described above. Each treatment consisted of nine 48 samples, three of them were added with an equal volume of buffer containing 10 mM MgSO4 49 and 50% glycerol for bacteria isolation, three of them were used for amplicon sequencing, and 50 three of them were used for metagenomic sequencing. All the samples were flash-frozen in liquid 51 nitrogen and stored at -80℃ until further use. 52

Amplicon sequencing 53
A total of 60 samples were subjected to bacterial 16S rRNA gene V3-V4 region amplicon 54 sequencing. The methods are the same as our previous publication [6]. 55

Metagenomic sequencing 56
According to the results of the amplicon sequencing, there was little difference in 57 microbiome between samples collected from the second and third sampling time points. 58 Therefore, a total of 36 samples from T1, T2, and T2C that were representative were subjected to 59 metagenomic sequencing. The methods are the same as those described in our previous 60 publication [6]. 61

Bacteria isolation and taxonomic identification 62
Rhizosphere samples of HF12 collected from T1 and T2C that were added with buffer and 63 not challenged with Rs GMI1000 were used for bacterial isolation. Rhizospheres were thawed at 64 room temperature and vortexed for suspension. To isolate bacteria as diverse as possible, we 65 adopted different methods and culture media: (1) 1 mL of mixed rhizosphere suspensions were added into 25 mL 0.1 × TSB (Tryptone, 1.7 g; Soytone, 0.3 g; Glucose, 0.25 g; NaCl, 0.5 g; 67 K2HPO4, 0.25 g; pH 7.3 ± 0.2) for 4 d enrichment at 30℃ with shaking of 180 rpm. Then 100 μL 68 of serial dilutions were plated on 0.1 × TSB agar media, of which agar and culture medium were 69 sterilized together or separately; (2) The serial diluted rhizosphere suspensions without 70 enrichment culture were plated on 0.1 × TSB, R2A [7], and NA (Peptone, 5 g; Glucose, 10 g; 71 Beef extract, 3 g; Yeast extract 0.5 g) agar plates. The agar and culture medium were sterilized 72 together or separately. All the plates were cultured at 28℃ for 2-6 d. A total of 259 isolates 73 covering different morphologies were selected and streaked for the single colony, and they were 74 MEGA-X software [8], the phylogenetic tree was constructed using the maximum likelihood method, and the bootstrap value was set to 1000. The taxonomic annotations on the tree were 89 done with the online tool iTOL [9]. The phylogenetic tree of the genera with the top 5% relative 90 abundance obtained by the amplicon was visualized with GraPhlAn [10], and the isolated genera 91 and their relative abundance (percentage) were annotated outside. 92

SEM 93
To observe the morphology of #276 that was identified as Sphingomonas sp. Cra20 through 94 16S rRNA gene, we performed scanning electron microscopy. After 36 h culture in R2A liquid 95 medium, cell pellets of #276 were collected by centrifugation at 8000 rpm for 5 min. Cell 96 fragments and culture medium were removed by cleaning three times using 0.1M PBS buffer 97 (Na2HPO412H2O, 35.8 g; KH2PO4, 2.5 g; NaCl, 80 g; KCl, 20 g; 1L H2O; pH 7.2-7.4). The 98 clean cell pellets were fixed in 2.5% glutaraldehyde buffer for 4 h at 4℃. After that, they were 99 centrifuged and cleaned using 0.1M PBS buffer for three times. Add 1 mL of 30%, 50%, 80%, 100 and 100% ethanol successively for gradient dehydration, 10 min each time, and finally centrifuge 101 to collect the bacterial cells for overnight freeze-drying. Observations were performed at the state 102 key laboratory of agricultural microbiology using field emission SEM. 103

Genome sequencing and analysis 104
Bacteria pellets were sent for DNA extraction and whole genome sequencing at Shanghai 105 Personal Biotechnology Co., Ltd. (Shanghai, China). DNA library was constructed using 106 TruSeqTM DNA Sample Prep Kit, following the manufacturer's instructions. Sequencing was 107 performed at Illumina NovaSeq platform using the paired-end, 2×150 bp mode. Adapters were 108 removed using AdapterRemoval [11]. All sequences were quality corrected using SOAPec based 109 on the Kmer frequency, and the Kmer was set to 17. Reads without adapters were de novo assembled into Contigs and scaffolds by A5-MiSeq [12]and SPAdes [13]. After comparing and 111 evaluating the assembly quality, we chose the results from SPAdes and corrected the base quality 112 using Pilon [14]. Prediction of protein-coding genes in bacterial genomes was conducted using 113 GeneMarkS software [15]. Clustered Regularly Interspaced Short Palindromic Repeats, 114 CRISPRs) were predicted using the CRISPR recognition tool (CRT) [16]. Protein-coding KO 115 genes and Pathway annotation were mainly completed by KEGG's KAAS automated annotation 116 system [17], in which the gene set was selected as "For Prokaryotes", the annotation method was 117 "blast", and the discrimination rule of gene KO was selected as a bi-directional best hit (BBH). 118 Taxonomic annotation and genome quality were performed using the online tool MiGA with the 119 NCBI Prok method [18]. Biosynthesis gene clusters (BGCs) were analysed by online tool 120 antiSMASH v6.0.1 with the default parameters [19].