Fig 1.
T. atroviride exudates inhibit the growth of rhizobacteria.
A) Bacterial cells were incubated for 24 hrs in spent medium (SM) from T. atroviride cultures (0.2X and 0.8X) with supplementation with fresh medium (R2A at 1X; [66]). The control condition is fresh medium alone (Vogel´s Minimal Media; VMM [64]) supplemented with 1X of R2A. B) Phylogenetic distribution of bacteria used in this study using the rpoB gene (scale bar 0.05 indicates 5 nucleotide substitutions per 100 nucleotides). C) Characterization of the growth of rhizosphere bacteria in response to T. atroviride exudates (OD600) after 72 hrs at 0.2X and 0.8X concentrations of VMM supplemented with fresh R2A medium. Error bars show SD. Asterisks indicate statistical differences, determined using Tukey’s multiple comparison test (p<0.001; Table B in S1 Text). Fig 1A was constructed via BioRender (https://www.biorender.com), agreement number BA25LBKUNK.
Fig 2.
RB-TnSeq of rhizosphere bacteria exposed to T. atroviride exudates.
A) Overview of the strategy for RB-TnSeq profiling of rhizosphere bacteria in response to T. atroviride-spent media. Step 1. RB-TnSeq mutant libraries were grown in LB with kanamycin from archived glycerol stocks. 2. To obtain spent medium, T. atroviride was grown in Vogel’s Minimal Medium [64] for 72 hrs, then both control and inoculated media were passed through a 0.22 μm filter to remove fungal material. 3. Each bacterial RB-TnSeq library was grown in the presence of T. atroviride spent media or uninoculated media control. After the bacteria grew for 24 hrs samples were collected for genomic DNA extraction and BarSeq using primers to conserved sequences that flank the unique barcodes. 4. Gene fitness was calculated by comparing the barcode counts in each gene before (Time 0) and after growth in the experimental condition (Condition). B) For each bacterium, a heatmap of the main functions of genes with negative fitness when exposed to T. atroviride exudates (Fitness < -1); detailed predicted functions are provided in S1 Dataset and Fig 2A was constructed via BioRender (https://www.biorender.com), agreement number DV25LBKS04.
Fig 3.
Comparison of bacterial gene log2 fitness in presence and absence of Trichoderma atroviride-spent media.
At least three replicates per condition were analyzed for all strains. A) Herbaspirillum seropedicae SmR1, B) Klebsiella michiganensis M5aI, C) Pseudomonas simiae WCS417, and D) P. putida KT2440. Each point on the graphs represents the fitness value associated with a disrupted gene, with the gray points indicating no significant change in fitness (|t| > 4). The unnamed genes highlighted in green are those with negative fitness in the presence of T. atroviride exudates (Fitness < -1 in spent media), while orange points indicate positive fitness values. Shown in yellow are mutants that were phenotypically rescued in the presence of exudates compared to those growing in uninoculated media. Mutations in named individual genes with negative fitness scores in other colors are noted and their general function indicated in the box. Light orange dots represent mutants of the flagellar genes (>1 log2 fitness score).
Fig 4.
Pseudomonas simiae and P. putida log2 fitness on polymyxin B.
At least three replicates per condition were analyzed for all conditions. A) BarSeq profile of P. simiae WCS417 mutant library in response to polymyxin B (2 μg/mL). B) BarSeq profile of P. putida KT2440 mutant library in response to polymyxin B (2 μg/mL). Green dots represent genes with negative fitness in response to T. atroviride exudates in spent media (SM). C) Venn diagram of genes important for fitness in T. atroviride exudates from SM and genes important for polymyxin B resistance in P. simiae. D) Venn diagram of genes important for fitness on T. atroviride SM and genes important for polymyxin B resistance in P. putida. arnC: UDP phosphate 4-deoxy-4-formamido-L-arabinose transferase (PS417_13790), arnA: UDP-4-amino-4-deoxy-L-arabinose formyltransferase (PS417_13795), arnT: 4-amino-4-deoxy-L-arabinose transferase (PS417_13805), arnD: 4-deoxy-4-formamido-L-arabinose-phospho-UDP deformylase (PS417_13800), ugd: UDP-glucose 6-dehydrogenase (PS417_13820).
Fig 5.
RB-TnSeq validation using single mutants of P. simiae and P. putida in response to T. atroviride exudates.
A) Eight individual mutants were isolated from the RB-TnSeq library of P. simiae; the location of insertional mutations were confirmed by sequencing (Table C in S1 Text). Two independent mutants were tested with similar results; one is shown. The profiles correlated with the results obtained from BarSeq fitness profiles. Using two different insertional lines for the mprF and oprN genes, a growth experiment for 24 hrs was performed with 0.2X spent medium in both P. simiae (B) and P. putida (C). A one-way ANOVA and a Tukey test were performed to determine statistical differences among the different strains in the same treatment (** P < 0.01). Conditions where growth was not observed are labeled NG (no growth).
Fig 6.
The Hog1-MAPK pathway participates in synthesizing secondary metabolites that affect the fitness of mutants in rhizosphere bacteria.
A) Analysis of the differential expression of the 42 predicted biosynthetic gene clusters in the Δtmk3 mutant vs. the WT strain of T. atroviride (FDR < 0.05). The histograms represent the log2 fold-change values of all the genes within each biosynthetic gene cluster (BGC). At least four Non-Ribosomal Peptide Synthetase (NRPS) gene clusters were affected in their expression in the Δtmk3 mutant (mean log2 fold-change < -4). Black asterisks highlight those clusters that on average were repressed or induced more than 4 log2 fold-change. For each BGC, the expression of all the genes within the cluster were evaluated. As the majority of genes within each BGC did not exhibit significant expression changes, the histogram is centered around zero-fold change. B) The fitness profile changes in the four microbial strains evaluated when competing against exudates from the Δtmk3-mutant relative to media alone. Green dots represent genes with negative fitness (log2 Fitness < -1 & t-score <-4), and orange dots represent positive fitness (log2 Fitness > 1 & t-score > 4).
Fig 7.
Bacterial processes important for fitness on exudates from WT T. atroviride and the Δtmk3 mutant.
A) General processes negatively affected in bacterial species by exposure to WT and Δtmk3 exudates. The colors of the heatmap represent the number of genes with a significant negative change in fitness (Fitness < -1). B-D: Heatmap of gene-fitness values of genes rescued by spent-media of WT and Δtmk3 mutant in H. seropedicae (B), K. michiganensis (C), and P. simiae (D) (log2 Fitness in spent media > -1 & log2 Fitness in media < -1.5).
Fig 8.
Hydrophilic Interaction Liquid Chromatography-ESI (Electrospray) QTOF (quadrupole time of flight) MS/MS (tandem mass spectrometry) (HILIC-MS) untargeted metabolomic profile of spent media from wild-type T. atroviride and the Δtmk3 mutant A) Volcano plot showing the differential peaks between the Δtmk3 mutant versus the WT strain. The x-axis shows the value of log2 fold-change (FC) and the y-axis shows the -log10 value of the adjusted p-value (p). The orange dots show molecules with a log2 fold-change greater than 1, and the blue dots show molecules with log2 fold-change lower than -1 (p-adjusted < 0.05). B) Low mass-charge (m/z) molecules that are primarily present in the Δtmk3 mutant. C) Highest m/z molecules produced in the WT strain that were greatly reduced in the Δtmk3 mutant (Peak height m/z). D) Purine quantification showed that adenine and adenosine 5´-monophosphate occurred in exudates from T. atroviride in the same proportion in both WT and the Δtmk3 mutant. In bar graphs B, C, and D, the negative controls (media only), were excluded as values were very close to zero (S5 Dataset).