Distance-decay relationships partially determine diversity patterns of phyllosphere bacteria on Tamarix trees across the Sonoran Desert [corrected]

Abstract

Dispersal limitation in phyllosphere communities was measured on the leaf surfaces of salt-excreting Tamarix trees, which offer unique, discrete habitats for microbial assemblages. We employed 16S rRNA gene pyrosequencing to measure bacterial community dissimilarity on leaves of spatially dispersed Tamarix specimens in sites with uniform climatic conditions across the Sonoran Desert in the Southwestern United States. Our analyses revealed diverse bacterial communities with four dominant phyla that exhibited differential effects of environmental and geographic variables. Geographical distance was the most important parameter that affected community composition, particularly that of betaproteobacteria, which displayed a statistically significant, distance-decay relationship.

Fig 1

Sampling site locations in the Sonoran Desert, Southwestern United States. The map was generated using the Map Maker application from the National Atlas of the United States website.

Fig 2

Taxonomic assignment of 16S rRNA gene sequences to genera in each of the T. aphylla leaf samples. Only the 22 genera found at a relative abundance greater than 5% are included by name. All other genera are grouped into the gray areas on the right hand side of each of the bars, completing the count to 100% of the number of identified genera. Taxonomic names ending in “_NA” have no assignment of affiliation at the relevant taxonomic level.

Fig 3

Unweighted-pair group method using average linkages (UPGMA) dendrogram depicting Chao pairwise dissimilarities (based on taxonomic affiliations) between a global array of samples. The samples cluster first according to climate (temperate sites on the left, extreme desert sites on the right), and then by location (sites in Israel in the left subclusters, U.S. samples in the right subclusters).

Fig 4

Correlations between environment, geography, and community dissimilarity. Environmentally fitted NMDS plots based on Chao pairwise dissimilarity index for the entire population (left) or for betaproteobacteria only (right). The points indicate the sampling site (as shown in Fig. 1), and the vectors represent the mean direction and strength of correlation of environmental variables. P values noted by the vectors were calculated by permutations of environmental data.

Fig 5

Correlation between geographical distance (A, C), environmental distance (B, D), and Chao pairwise dissimilarity index for the entire population (A, B) or for betaproteobacteria only (C, D). Locally weighted polynomial regression lines (3) are shown. The significance of correlation was tested using partial Mantel tests.