Indirect reduction of Ralstonia solanacearum via pathogen helper inhibition

Rhizosphere soil sampling

A total of 20 rhizosphere soil samples (20 tomato plants) were collected at the flowering stage from a tomato field located in Qilin town, Jiangsu province, China, 118°57’ E, 32°03’ N, which had been infested by the pathogen Ralstonia solanacearum for more than 15 years [8]. After uprooting plants, excess soil was first gently shaken from the roots, and the remaining soil attached to roots was considered as rhizosphere soil. Each rhizosphere soil sample was then used for bacterial strain isolation.

Isolation and identification of rhizobacteria

Isolation

A total of 640 bacterial strains were isolated from the fresh rhizosphere soil samples, according to a previously established protocol [11]. Briefly, 1 g of each rhizosphere sample was mixed with 9 mL MS buffer solution (50 mM Tris-HCl [pH 7.5], 100 mM NaCl, 10 mM MgSO₄, 0.01% gelatin) in a rotary shaker at 170 rpm min⁻¹ for 30 min at 30 °C. After serial dilution in MS buffer solution, 100-μl volumes of the diluted soil suspensions were plated on 1/10 tryptone soy agar (1/10 TSA, 1.5 g L⁻¹ tryptone, 0.5 g L⁻¹ soytone, 0.5 g L⁻¹ sodium chloride, and 15 g L⁻¹ agar, pH 7.0). After a 48-h incubation at 30 °C in the dark, 32 isolates were randomly picked per rhizosphere soil sample. To avoid potential fungal contamination, only highly diluted samples were used for isolation. The isolates were then re-streaked on TSA plates for colony purification. Approximately 5.5% (35 isolates) of the bacterial isolates failed to grow on the TSA plates for unknown reasons when we re-streaked them and were therefore omitted from the dataset. The final collection thus consisted of 605 bacterial isolates derived from 20 rhizosphere soil samples. All purified isolates were cultured in 100 μl tryptone soy broth (TSB, liquid TSA) in 96-well microtiter plates at 30 °C with shaking (rotary shaker at 170 rpm) for 18 h before freezing and storing at −80 °C in 15% glycerol.

Strain identification

We sequenced the full 16 S rRNA gene to taxonomically identify all 605 rhizobacterial isolates. The 16 S rRNA gene was sequenced via Sanger sequencing of PCR products from glycerol stocks by Shaihai Songon Biotechnology Co., Ltd, Shaihai Station. The PCR system (25 µl) was composed of 1 µl of bacterial cells (overnight culture), 12.5 µl mixture, 1 µl of forward (27 F: 5-AGA GTT TGA TCA TGG CTC AG-3) and reverse primer (1492 R: 5-TAC GGT TAC CTT GTT ACG ACT T-3) each [17] and 9.5 µl of sterilized water. PCR was performed by initially denaturizing at 95 °C for 5 min, cycling 30 times with a 30-s denaturizing step at 94 °C, annealing at 58 °C for 30 s, extension at 72 °C for 1 min 30 s, and a final extension at 72 °C for 10 min. The 16 S rRNA gene sequences were identified using NCBI databases and homologous sequence similarity. A total of 90 bacterial isolates that were identified as Ralstonia solanacearum were removed from further analyses, resulting in 515 remaining isolates.

Direct effect of rhizobacteria on pathogen growth in vitro

We used R. solanacearum strain QL-Rs1115 tagged with the pYC12-mCherry plasmid as a model bacterial pathogen [8, 18]. We first tested the direct effects of the 515 non-R. solanacearum bacterial strains on the growth of R. solanacearum in vitro by using supernatant assays. Briefly, after 48 h of growth in NB (nutrient broth) medium (glucose 10.0 g l⁻¹, tryptone 5.0 g l⁻¹, yeast extract 0.5 g l⁻¹, beef extract 3.0 g l⁻¹, pH 7.0) on a shaker at 170 rpm, 30 °C, all bacterial cultures were filter sterilized to remove living cells (0.22 µm filter). Subsequently, 20 µl of sterile supernatant from each strain’s culture and 2 µl overnight culture of the pathogen (adjusted to OD600 = 0.5 after 12 h growth at 30 °C with shaking) were added into 180 µl of fresh NB medium (5-times diluted, in order to better reflect the effect of the supernatant). Control treatments were inoculated with 20 µl of 5 X diluted NB media instead of the bacterial supernatant. Each treatment was conducted in triplicate. All bacterial cultures were grown for 48 h at 30 °C with shaking (170 rpm) before measuring pathogen density as red mCherry protein fluorescence intensity (excitation: 587 nm, emission: 610 nm) [9, 11] which was linearly related to the CFU of pathogen R. solanacearum (Fig. S1). To test for significance of growth promotion or inhibition, R. solanacearum densities were log₁₀-transformed prior to analyses of variance (ANOVA) and Bonferroni t test to compare mean differences between each rhizobacterial supernatant treatment and the control treatment, with p values less than 0.05 considered statistically significant. The effect on pathogen growth was defined as the percentage of improvement or reduction in pathogen growth by the supernatant compared to the control treatment. When the effect on pathogen growth was positive, i.e., when the supernatants from strains significantly promoted the growth of the pathogen, they were considered as helpers of the pathogen. If the effect on pathogen growth was negative, i.e., when the supernatants from strains significantly inhibited the growth of the pathogen, they were considered as inhibitors of the pathogen.

Assessing strain redundancy among the 515 non-Ralstonia solanacearum bacteria

We assessed possible redundancy among the 515 strains of the non-Ralstonia solanacearum rhizobacteria. To encompass both taxonomic and functional redundancies, we considered the 16 S rRNA gene sequences as well as the direct effect of their supernatant on Ralstonia solanacearum. Self BLAST searches were performed on the full 515 sequence dataset using the makeblastdb and blastn commands from the BLAST command line tool [19]. Sequences showing >99% identity over >95% of the full length of the 16 S rRNA gene were considered as taxonomically redundant. We then compared the direct effects on pathogen growth of the taxonomically redundant strains, and removed those showing the same patterns of interactions (positive, negative or neutral). Accordingly, (see the dataset “Library of rhizobacterial strains” in the supplementary information), 355 of the 515 strains (68.9%) were removed from the original dataset for further analyses.

Phylogenetic tree construction

The 16 S rRNA gene sequences of the 160 non–redundant bacteria were aligned using MUSCLE [20]. Sequences in the alignment were trimmed at both ends to obtain maximum overlap using the MEGA X software, which was also used to construct taxonomic cladograms [21]. We constructed a maximum-likelihood (ML) tree, using a General Time Reversible (GTR) + G + I model, which yielded the best fit to our data set. Bootstrapping was carried out with 100 replicates retaining gaps. A taxonomic cladogram was created using the EVOLVIEW web tool (https://evolgenius.info//evolview-v2/). To show the relationship between phylogeny and the effects of rhizobacteria on pathogen growth, we added taxonomic status (phylum) of each rhizobacterial strain and its effect on pathogen growth as heatmap rings to the outer circle of the tree separately (Fig. 2B).

Effects of rhizobacteria on pathogen helper strains growth in vitro

We then assessed the potential of different rhizosphere isolates to inhibit helper strains. We first selected two model helper strains (Phyllobacterium ifriqiyense LM1 (Pi) and Microbacterium paraoxydans LM2 (Mp)), which showed strong positive effects on pathogen growth both in co-culture and in supernatant assays (Fig. S2). We defined the effect of rhizobacterial strains on the growth of helpers as the indirect effect on R. solanacearum growth. To study these indirect effects, we first chose a subset of 46 rhizobacterial strains representing a gradient of positive, neutral or negative effect on pathogen growth based on supernatant assays (results in x axis of Figs. 3C and 4A, B, C). We then tested the effects of these 46 rhizobacterial strains on the growth of each of the two helper strains using supernatant assays. Briefly, after 48 h growth in NB media, each of the 46 bacterial monocultures was passed through a 0.22 µm filter to remove living cells. Then 20 µl of sterile supernatant from each strain’s culture and 2 µl overnight culture of Pi or Mp (adjusted to OD₆₀₀ = 0.5 after 12 h growth at 30 °C with shaking) were added into 180 µl of fresh NB medium (5-times diluted, in order to better reflect the effect of the supernatant). Control treatments were inoculated with 20 µl of 5× diluted NB media instead of a bacterial supernatant. Each treatment was replicated four times. All bacterial cultures were grown for 24 h at 30 °C with shaking (170 rpm) before measuring helper density as optical density (OD₆₀₀). To test for significance of growth promotion or inhibition, we used analyses of variance (ANOVA) and Bonferroni t test to compare mean differences of helper density between each rhizobacterial supernatant treatment and the control treatment, with p values lower than 0.05 being considered statistically significant. The effect of rhizobacteria on the helpers’ growth (results in y axis of Fig. 3C and x axis of Fig. 4D, E, F) was defined as the percentage of increase or reduction in helper growth by the supernatant compared to the control treatment.

In vitro pathogen growth in the presence of a helper strain and supernatant from rhizobacterial isolates

To disentangle the direct effects from the indirect effects of rhizobacteria on R. solanacearum growth, we compared their relative effects using in vitro triculture assays comprised of R. solanacearum, one of the two helper strains and supernatant of one of the 46 chosen rhizobacterial strains. Briefly, after 48 h of growth in NB media, each of the 46 bacterial monocultures was passed through a 0.22 µm filter to remove living cells. Then, 20 µl of sterile supernatant from each strain’s culture and 2 µl overnight culture of Pi or Mp (densities were adjusted to ~10⁷ cells per ml) were added to 180 µl of fresh NB medium (5-times diluted). Each treatment was replicated four times. At the same time, 2 µl overnight culture of mCherry-tagged R. solanacearum (density was adjusted to ~10⁶ cells per ml) was added to each treatment in 96-well plates at 30 °C with shaking (170 rpm). After 24-h growth, R. solanacearum density (results in y axis of Fig. 4A, D) was measured as the red mCherry protein fluorescence intensity (excitation: 587 nm, emission: 610 nm) with a SpectraMax M5 plate reader.

In vivo pathogen growth and plant disease development in the presence of a helper strain and a rhizobacterial strain

To validate in vitro results, we set up greenhouse experiments where plants were inoculated with a bacterial consortium consisting of R. solanacearum, one of the two helper strains and a test rhizobacterial strain. Tomato seeds (Lycopersicon esculentum, cultivar “Ai hong sheng”) were surface-sterilized by soaking them in 3% NaClO for 5 min and in 70% ethyl alcohol for 1 min before being germinated on water-agar plates for 2 days. Seeds were then sown into seedling trays containing gamma irradiation-sterilized (to avoid potential effects of the resident community) seedling substrate (Huainong, Huaian Soil and Fertilizer Institute). At the three-leaf stage, tomato plants were transplanted to seedling trays containing 200 g of the same seedling substrate as describe above.

To relate our results to practical application conditions, we selected a subset of 12 strains that displayed a range of inhibitions effects on pathogen and helpers (Table S1) out of the 46 rhizobacterial isolates used for the in vitro assays. Each rhizobacterial strain was used in combination with each of the two helper strains and R. solanacearum, resulting in a total of 28 treatments (Table S2), including a water control, R. solanacearum alone, and R. solanacearum with just each of the two helper strains (results in Fig. 3B, C). For each treatment, four replicate seedling trays were used, with each replicate seedling tray containing 4 tomato plants. Three days after transplantation, plants of each treatment were inoculated with one of the two helper strains, alone or in combination with one of the rhizobacterial strains, using the root drenching method at a final concentration of 10⁸ CFU g⁻¹ soil for each bacterial strain [22]. Seven days after inoculation of helper alone or together with rhizobacteria, R. solanacearum was introduced to the roots of all plants at a final concentration of 10⁷ CFU g⁻¹ soil. The positive control treatment with R. solanacearum alone was inoculated only with the pathogen, and the negative control treatment was not inoculated with any bacteria. Tomato plants were maintained under standard greenhouse conditions (i.e., at natural temperature variation ranging from 28 °C to 32 °C, 15/9 h day/night conditions) and watered regularly with sterile water. Seedling trays were rearranged randomly every two days. Forty days after transplantation, plants were destructively harvested. The disease index for each plant was recorded based on a scale ranging from 0 to 4 [23]. Disease severity for each replicate seedling plate was calculated as described by: Disease severity = [∑ (The number of diseased plants in the disease index category × disease index category)/ (Total number of plants used in the experiment × highest disease index category)] ×100% [23, 24]. Simultaneously, we collected rhizosphere soil samples following an established protocol [4]. Briefly, two plants were randomly chosen from each replicate seedling tray to collect rhizosphere soils and further combined to yield one sample, resulting in a total of 112 rhizosphere soil samples for which R. solanacearum population densities were determined.

Quantification of R. solanacearum at the end of the in vivo experiment

We determined R. solanacearum densities using quantitative PCR (qPCR). DNA was extracted from rhizosphere soils using a Power Soil DNA isolation kit (Mo Bio Laboratories) following the manufacturer’s protocol. DNA concentrations were determined by using a NanoDrop 1000 spectrophotometer (Thermo Scientific) and extracted DNA was used for R. solanacearum density measurements using specific primers (forward, 5ʹ-GAA CGC CAA CGG TGC GAA CT-3ʹ; reverse, 5ʹ-GGC GGC CTT CAG GGA GGT C-3ʹ) targeting the fliC gene, which encodes the R. solanacearum flagellum subunit [25]. The qPCR analyses were carried out with a StepOnePlus Real-Time RCR Instrument using SYBR green fluorescent dye detection and three technical replicates as described previously [4].

Statistical analyses

To meet assumptions of normality and homogeneity of variance, R. solanacearum densities measured in vitro and in vivo were log10-transformed. When comparing mean differences between treatments, we used analyses of variance (ANOVA) and the Tukey Test, where p values lower than 0.05 were considered statistically significant. R. solanacearum densities were explained by two quantitative indices, the direct effect of rhizobacteria on R. solanacearum growth (the effect of rhizobacteria on R. solanacearum growth) and the indirect effect of rhizobacteria on R. solanacearum growth (the effect of rhizobacteria on helper strains’ growth). Nonlinear regression analyses (Sigmoidal, Sigmoid, 3 Parameter) were used to analyze the relationship between the direct effect and pathogen density, as well as the relationship between indirect effects and pathogen density in the presence of helper strains in vitro. The relationships between them, and between direct/indirect effects and disease severity in the presence of helper strains in vivo, were analyzed using linear regressions. These analyses were carried out using the R 3.6.3 program (www.r-project.org) and Sigma Plot (V.12.5).

To further consider the growth inhibition of R. solanacearum, and disease suppression, we fitted a linear model to estimate the relative importance of direct effects versus indirect effects on the density of R. solanacearum both in vitro and in vivo, and on disease severity. This model considered the interaction scenario where rhizobacterial strains inhibited both the pathogen and its helpers (see the R script “Model” in the supplementary information). These analyses were performed in R version 3.6.3 [26] in conjunction with the package car, readxl and dplyr, and tidyverse 1.2.1 [27]. Briefly, proportional effects were normalized using a folded cube root transformation as suggested in J.W. Tukey [28] and fitted using a linear model with direct effects, indirect effects, and an interaction between helper strains and indirect effects as fixed factors. Normality of residuals was tested using the Shapiro-Wilk normality test and visual inspection of QQ-plots with standardized residuals. Type-II sum of squares were calculated using the ANOVA function from car 3.0-2 [29]. Subsequent visualization of the model outcome (results in Fig. 5) showed the predicted R. solanacearum densities and disease severity for different values of the inhibition via pathogen (Direct) or helper (Indirect) as estimated from the statistical model. For the Direct effect line, the indirect effect is set to be zero, while for the Indirect effect line, the direct effect is set to be zero.

Source: Ecology - nature.com