Microbial communities associated with diverse plant and animal species exert a powerful influence on host fitness. Specific microbial taxa and functional genes directly affect normal development, nutrition, and immunity. Targeted rRNA and shotgun metagenomic surveys reveal the conservation of microbiome structure at higher taxonomic and functional levels, consistent with its importance to host physiology. These same investigations reveal a surprising amount of variation in the microbiomes of different members at the population level for humans, mice, and zebrafish. Despite the widespread occurrence of this phenomenon, few systematic studies have attempted to reconcile the puzzling combination of strong host control and high inter-individual variability. This gap motivates my long-term research goal: to understand the interaction between selective and random processes in shaping the structure and function of host-associated microbiota. Our objective is to determine the forces regulating microbiome assembly at multiple levels of genetic variation, using microbial colonization of the leaves of Arabidopsis thaliana, a model organism in plant genetics, as a model system.
My laboratory pursues several lines of investigation into the phyllosphere system, and we have achieved exciting results over the past year. We completed an experiment designed to determine the repeatability of community assembly for both taxonomic structure and functional capacity in the Arabidopsis leaf microbiome (the “phyllosphere”) under conditions of unlimited colonization. In unpublished work, we grew large numbers of Arabidopsis plants in the MBL greenhouse facility. Our goals were three-fold: to characterize the temporal trajectory of phyllosphere community structure, assess plant-to-plant variability in community composition, and quantify the influence of airborne microbes on the phyllosphere community. We developed methods to characterize the airborne microbial community colonizing leaves in conjunction with the phyllosphere community over the full plant life cycle. We use high-throughput sequencing of the 16s rRNA gene to measure taxonomic diversity and shotgun metagenomic sequencing to measure the functional capacity of the phyllosphere microbiome.

Figure 1: Development of phyllosphere community structure on greenhouse-grown Arabidopsis thaliana plants over a 73 day time series. (a) Phyllosphere communities follow a clear trajectory in membership from young plants to mature plants, and are distinct from the airborne inoculating community (PCoA plot using the Jaccard beta-diversity index). Arrows are in the direction of increasing time. Bubbles outlined in blue represent air microbes, and bubbles outlined in green represent plant-associated microbes. Bubble centroids are equal to the mean, and bubble diameter is proportional to the standard error of the mean. (b) The Morisita-Horn abundance-based metric of community composition indicate that plants in close spatial association (sharing the same experimental tray) have more similar microbial abundance profiles. Bubbles outlined in green and orange (Tray 1, green; Tray 2, orange) represent individual plants while bubbles outlined in blue represent averages across replicate air samples. Note the high percentage of variance explained by each axis. Separation of trays from air communities is mainly on the PC1 axis, explaining 32% of total variance. Separation of Tray 1 plant associated microbes from Tray 2 associated microbes is mainly on the PC2 axis, which explains 24% of total variance. Analysis performed using QIIME (Caporaso et al. 2010).

Figure 1: Development of phyllosphere community structure on greenhouse-grown Arabidopsis thaliana plants over a 73 day time series. (a) Phyllosphere communities follow a clear trajectory in membership from young plants to mature plants, and are distinct from the airborne inoculating community (PCoA plot using the Jaccard beta-diversity index). Arrows are in the direction of increasing time. Bubbles outlined in blue represent air microbes, and bubbles outlined in green represent plant-associated microbes. Bubble centroids are equal to the mean, and bubble diameter is proportional to the standard error of the mean. (b) The Morisita-Horn abundance-based metric of community composition indicate that plants in close spatial association (sharing the same experimental tray) have more similar microbial abundance profiles. Bubbles outlined in green and orange (Tray 1, green; Tray 2, orange) represent individual plants while bubbles outlined in blue represent averages across replicate air samples. Note the high percentage of variance explained by each axis. Separation of trays from air communities is mainly on the PC1 axis, explaining 32% of total variance. Separation of Tray 1 plant associated microbes from Tray 2 associated microbes is mainly on the PC2 axis, which explains 24% of total variance. Analysis performed using QIIME (Caporaso et al. 2010).

Our unpublished results clearly demonstrate that the membership of phyllosphere communities converges to a stable set of taxa at approximately 50 days post-germination (Figure 1). Moreover, as time progresses, the set of taxa present in plant-associated communities becomes increasingly divergent from the set of taxa that characterize air communities (Figure 2). While the composition of airborne communities remains remarkably stable over time, plant communities follow a clear temporal trajectory from high similarity to air communities towards a distinct stable end point composition (Figure 1a). The variability in community composition between replicate plants sampled at the same timepoint also decreases as plants mature, while the same is not true for air communities. These results suggest that the host plant exerts a strong selective effect on the colonizing microbes, such that only a subset are able to thrive. Strikingly, strong host selection for the presence of particular microbial phylotypes on mature plants was coupled with high stochastic variation in the abundance of dominant taxa. Despite experimental randomization, plants in close spatial association had more similar microbial communities, suggesting plant-to-plant microbial exchange (Figure 1b). This work is the first to measure host-associated microbial community development over time in conjunction with measurements of environmental inocula.

Figure 2: Taxa significantly associated with late-stage stable phyllosphere communities (green) versus contemporaneous air communities (red), shown in cladogram format. Higher-level taxonomic groupings are indicated adjacent to colored wedges. Analysis performed using LeFSe (Segata et al. 2011).

Figure 2: Taxa significantly associated with late-stage stable phyllosphere communities (green) versus contemporaneous air communities (red), shown in cladogram format. Higher-level taxonomic groupings are indicated adjacent to colored wedges. Analysis performed using LeFSe (Segata et al. 2011).

Print Friendly