Inter-species interactions between two bacterial flower commensals and a floral pathogen reduce disease incidence and alter pathogen activity

Abstract

Flowers are colonized by a diverse community of microorganisms that can alter plant health and interact with floral pathogens. Erwinia amylovora is a flower-inhabiting bacterium and a pathogen that infects different plant species, including Malus × domestica (apple). Previously, we showed that the co-inoculation of two bacterial strains, members of the genera Pseudomonas and Pantoea, isolated from apple flowers, reduced disease incidence caused by this floral pathogen. Here, we decipher the ecological interactions between the two flower-associated bacteria and E. amylovora in field experimentation and in vitro co-cultures. The two flower commensal strains did not competitively exclude E. amylovora from the stigma habitat, as both bacteria and the pathogen co-existed on the stigma of apple flowers and in vitro. This suggests that plant protection might be mediated by other mechanisms than competitive niche exclusion. Using a synthetic stigma exudation medium, ternary co-culture of the bacterial strains led to a substantial alteration of gene expression in both the pathogen and the two microbiota members. Importantly, the gene expression profiles for the ternary co-culture were not just additive from binary co-cultures, suggesting that some functions only emerged in multipartite co-culture. Additionally, the ternary co-culture of the strains resulted in a stronger acidification of the growth milieu than mono- or binary co-cultures, pointing to another emergent property of co-inoculation. Our study emphasizes the critical role of emergent properties mediated by inter-species interactions within the plant holobiont and their potential impact on plant health and pathogen behavior.

Importance: Fire blight, caused by Erwinia amylovora, is one of the most important plant diseases of pome fruits. Previous work largely suggested plant microbiota commensals suppressed disease by antagonizing pathogen growth. However, inter-species interactions of multiple flower commensals and their influence on pathogen activity and behavior have not been well studied. Here, we show that co-inoculating two bacterial strains that naturally colonize the apple flowers reduces disease incidence. We further demonstrate that the interactions between these two microbiota commensals and the floral pathogen led to the emergence of new gene expression patterns and a strong alteration of the external pH, factors that may modify the pathogen's behavior. Our findings emphasize the critical role of emergent properties mediated by inter-species interactions between plant microbiota and plant pathogens and their impact on plant health.

Keywords: Erwinia amylovora; Malus domestica; Pantoea; Pseudomonas; co-culture; co-inoculation; emergent property; fire blight; meta-transcriptome.

Fig 1

Co-inoculation of two flower commensal bacteria reduces disease incidence in experimental orchards. (A) The drawing shows the experimental procedure of co-inoculating Pseudomonas and Pantoea on apple flowers “Red Delicious.” Briefly, apple trees were either inoculated with water (control), with Ea, with Ea than treated with streptomycin (Ea:strep), inoculated with Pantoea (Pa) or Pseudomonas (Ps) than treated with Ea (Pa:Ea or Ps:Ea, respectively), or co-inoculated with Ps and Pa then treated with Ea (Pa:Ps:Ea). (B) Boxplots show fire blight disease incidence in the field under the five treatments: Ea, Pa:Ea, Ps:Ea, Pa:Ps:Ea, and Ea:strep. Co-inoculation of the flowers with Pseudomonas and Pantoea significantly reduced fire blight incidence in the experimental field. Each circle corresponds to one tree; the letter indicates significant differences.

Fig 2

The two flower commensal bacteria do not competitively exclude E. amylovora from the stigma. (A) The boxplots depict amsC gene copy number of E. amylovora on the flower stigma. Each circle corresponds to one flower, and letters indicate significant differences between treatments. E. amylovora was detected in all treatments. None of the inoculations significantly reduced the population size of Ea, except the application of streptomycin. (B and C) The boxplots show the relative abundance of Erwinia amplicon sequence variant (ASV), Pantoea ASV, and Pseudomonas ASV revealed by amplicon sequencing of the 16S gene and gene transcript, respectively. Although co-inoculation of Pa and Ps reduced the relative abundance of Ea, the 16S gene (B) and gene transcript (C) of E. amylovora were readily detected in all treatments. (D) Micrographs show confocal microscopy of E. amylovora expressing Green Fluorescence Protein (GFP) (top left), Pseudomonas expressing mCherry (top right), Pantoea expressing mTurquoise (bottom left), and the overlay of the three channels (bottom right) on the stigma of apple flower. Ps and Pa were first inoculated on the stigma, then Ea was applied after 1 day. Microscopy pictures were acquired 2 days after the inoculation of Ea. All three bacterial strains are detected and express fluorescent proteins on the stigma of apple flowers.

Fig 3

In vitro co-culture of the two flower commensal bacteria and the fire blight pathogen. (A) The schematic depicts in vitro mono- and co-culture of E. amylovora (Ea), Pantoea (Pa), and Pseudomonas (Ps) using a synthetic stigma exudation medium. (B) Bar plots show bacterial CFUs that were determined after 6 h post-inoculation using Lysogeny Broth agar with a selective antibiotic (Materials and Methods). The three bacterial species reached different growth rates in mono-culture, but none of the bacterial strains were out-competed in co-culture. (C) The boxplots show the relative growth of each bacterium in co-culture compared to its mono-culture growth condition. A relative growth >1 indicates that the strain grows more abundantly in the co-culture vs mono-culture, whereas a value <1 indicates otherwise. Boxplots labeled Ea, Pa, and Ps show growth variations in the mono-culture of E. amylovora, Pantoea, and Pseudomonas, respectively. Ternary co-culture of the bacteria leads to growth depletion, but none of the strain is out-competed to extinction.

Fig 4

Co-culture of the three bacterial strains leads to new expression patterns. (A, C, and E) depict MA plots (log fold-change versus mean expression) showing DEGs in Ea, Pa, and Ps upon binary and ternary co-culture, respectively. Count data were fitted to negative binomial distribution (DESeq2) and DEGs were determined based on the cutoffs of 1.5 and 0.05 in log₂ fold change expression and Benjamini-Hochberg adjusted P-value, respectively. Red and blue indicate significantly up- and down-regulated genes, and gray indicates no significant change in the expression. Numeric values indicate a total number of genes up- or down-regulated. Ternary co-culture of Ea, Pa, and Ps led to more DEGs in the three strains compared to their binary co-culture. (B, D, and F) show Venn charts indicating the number of DEGs shared between binary and ternary co-cultures in Ea, Pa, and Ps, respectively. Top Venn diagrams in (B, D, and F) correspond to significantly up-regulated genes Ea, Pa, and Ps respectively. Bottom Venn diagrams correspond to significantly down-regulated genes in Ea, Pa, and Ps, respectively. In Ea and Ps, we noted no core genes significantly up- or down-regulated shared between all co-culture growth conditions. Pa showed a small fraction of core genes significantly up- or down-regulated shared between all co-culture conditions.

Fig 5

Functional annotation of differentially regulated genes. (A) shows circle chart that depicts the proportion of the functional annotation of DEGs in Ea (top), Pa (middle), and Ps (bottom). Color code indicates category, and values inside the chart indicate gene counts. More DEGs are related to metabolism, than to information storage and processing, or cellular processes and signaling. (B) Barplots indicate the classes of COGs of the annotated DEGs in the three strains. Color denotes bacterial strain. Ternary co-culture results in the increase of DEGs and annotated COGs classes.

Fig 6

Ternary co-culture leads to strong acidification of the pH. The boxplots show pH values in the control (no bacterial inoculation), the mono- and the co-cultures. Inoculation of any of Ea, Pa, Ps or their combinations significantly alters the pH of the growth milieu. Ternary co-culture of the bacteria shifts the pH to more acidic than any binary co-culture or mono-culture. Color denotes growth condition, and letters show significance in the P-value (P < 0.05, Kruskal-Wallis test followed by Conover’s post hoc test).