Environmental Metabolomics of the Tomato Plant Surface Provides Insights on Salmonella enterica Colonization

Abstract

Foodborne illness-causing enteric bacteria are able to colonize plant surfaces without causing infection. We lack an understanding of how epiphytic persistence of enteric bacteria occurs on plants, possibly as an adaptive transit strategy to maximize chances of reentering herbivorous hosts. We used tomato (Solanum lycopersicum) cultivars that have exhibited differential susceptibilities to Salmonella enterica colonization to investigate the influence of plant surface compounds and exudates on enteric bacterial populations. Tomato fruit, shoot, and root exudates collected at different developmental stages supported growth of S. enterica to various degrees in a cultivar- and plant organ-dependent manner. S. enterica growth in fruit exudates of various cultivars correlated with epiphytic growth data (R(2) = 0.504; P = 0.006), providing evidence that plant surface compounds drive bacterial colonization success. Chemical profiling of tomato surface compounds with gas chromatography-time of flight mass spectrometry (GC-TOF-MS) provided valuable information about the metabolic environment on fruit, shoot, and root surfaces. Hierarchical cluster analysis of the data revealed quantitative differences in phytocompounds among cultivars and changes over a developmental course and by plant organ (P < 0.002). Sugars, sugar alcohols, and organic acids were associated with increased S. enterica growth, while fatty acids, including palmitic and oleic acids, were negatively correlated. We demonstrate that the plant surface metabolite landscape has a significant impact on S. enterica growth and colonization efficiency. This environmental metabolomics approach provides an avenue to understand interactions between human pathogens and plants that could lead to strategies to identify or breed crop cultivars for microbiologically safer produce.

Importance: In recent years, fresh produce has emerged as a leading food vehicle for enteric pathogens. Salmonella-contaminated tomatoes represent a recurrent human pathogen-plant commodity pair. We demonstrate that Salmonella can utilize tomato surface compounds and exudates for growth. Surface metabolite profiling revealed that the types and amounts of compounds released to the plant surface differ by cultivar, plant developmental stage, and plant organ. Differences in exudate profiles explain some of the variability in Salmonella colonization susceptibility seen among tomato cultivars. Certain medium- and long-chain fatty acids were associated with restricted Salmonella growth, while sugars, sugar alcohols, and organic acids correlated with larger Salmonella populations. These findings uncover the possibility of selecting crop varieties based on characteristics that impair foodborne pathogen growth for enhanced safety of fresh produce.

FIG 1

Growth of S. enterica in fruit exudates (A and B), seedling (3-week-old) shoot (C and D) and root (E and F) exudates, and flowering plant (6-week-old) shoot (G and H) and root (I and J) exudates. Population densities were measured at 6 hpi (left) and 24 hpi (right). The arrows on the y axes represent the initial cell density at 0 h. Error bars indicate standard errors of the means, bars labeled with the same letter are not significantly different within the same time point measurement by Tukey's HSD test (P < 0.05), and ‡ in panels C through J denotes that the test was not carried out for these tomato cultivars.

FIG 2

Growth curves of S. enterica in fruit exudates of 13 different tomato cultivars predicted by Buchanan's three-phase linear model (IPMP 2013).

FIG 3

Scatter plot of growth of S. enterica in fruit surface compounds and exudates and epiphytic growth data (4) for fruit of 13 cultivars. The line represents the simple regression line fitted to the data.

FIG 4

Proportions of primary and secondary (fatty acids and phenolics) metabolites in exudates of seedlings (3 weeks postgermination), flowering plants (6 weeks postgermination), and fruit. Hierarchical cluster analysis results generated from ranked similarities of metabolite data obtained by GC-TOF-MS are shown in a dendrogram; similarity was determined by using the Bray-Curtis similarity coefficient. The ANOSIM statistic R and the associated P value are given for pairwise comparisons between different plant organs and developmental stages. R will have a value of 0 when all samples within different groups are the same and approaches 1 when replicates within one group are more similar to each other than to replicates from another group.

FIG 5

Hierarchical cluster analysis of tomato shoot, root, and fruit exudate profiles by classification of primary metabolites, including sugars and sugar alcohols (A), amino acids (B), organic acids (C), phenolics (D), and fatty acids (E), generated from ranked similarities of metabolite data obtained by GC-TOF-MS. Similarity was determined by using the Bray-Curtis similarity coefficient; ANOSIM results are attached to each dendrogram and indicate the test statistic R and the associated P value for pairwise comparisons between different plant organs and developmental stages. R will have a value of 0 when all samples within different groups are the same and approaches 1 when replicates within one group are more similar to each other than to replicates from another group.

FIG 6

Fatty acid composition of shoot and root exudates from flowering plants and fruit of various tomato cultivars. Data on the y axis show cumulative peak heights for the quantification ions at their specific retention index and represent abundance in the sample.