Compartment and Plant Identity Shape Tree Mycobiome in a Subtropical Forest

Abstract

Deciphering the relationships between microbes and their host plants is critical for a better understanding of microbial diversity maintenance and community stability. Here, we investigated fungal diversity and community assembly in the phyllosphere and rhizosphere of 13 tree species in a subtropical common-garden experiment. The results showed that fungal community structures significantly differed across compartments (leaf, root, and soil) and different tree species. Higher α-diversity was observed in the phyllosphere than in the roots and rhizospheric soil. Fungal community composition (β-diversity) was significantly affected by both compartment and species identity. The fungal community compositions were significantly correlated with soil pH in the roots and the soils as well as with soil nitrate and leaf total phosphorus in the leaves. We found that fungal community assemblies were mainly driven by deterministic processes, regardless of compartments. Moreover, host preference analyses indicated that stronger plant/fungus preferences occurred in leaves than in roots and soils. Our results highlight the differences in tree mycobiome between aboveground and belowground compartments and have important implications for the promotion of biodiversity conservation and management sustainability for the subtropical forest. IMPORTANCE Subtropical mountain forests are widely distributed in Southern China and are characterized by high biodiversity. The interactions between plants and fungi play pivotal roles in biodiversity maintenance and community stability. Nevertheless, knowledge of fungal diversity and of the community assembly patterns of woody plants is scarce. Here, we investigated fungal diversity and community assembly in the phyllosphere and rhizosphere of 13 tree species in a common-garden experiment. We found that both compartment and plant identity influenced fungal diversity, community, and guild compositions, while deterministic processes mainly governed the fungal community assembly, especially in the rhizospheric fungal communities. Our results demonstrate that tree leaves represent stronger host/fungi preferences than do roots and soils. Together, our findings enhance the understanding of the roles of compartment and plant identity in structuring fungal communities as well as promote fungal diversity maintenance in subtropical mountain forest ecosystems.

Keywords: community composition; deterministic process; host preferences; phyllosphere; α-diversity.

FIG 1

Fungal OTU richness (A) and Venn diagrams (B) comparing the OTU memberships among the three compartments of leaf, root, and soil across 13 tree species in a subtropical forest ecosystem.

FIG 2

Comparison of OTU richness (A, B, C) and Shannon index (D, E, F) among 13 plant species for each of the three compartments of leaf, root and soil. Different letters above the boxes indicate significant differences at P < 0.05. Plant species names: CCam, Cinnamomum camphora; CCar, Castanopsis carlesii; CL, Cunninghamia lanceolate; ED, Elaeocarpus decipiens; KB, Koelreuteria bipinnata; LCh, Liriodendron chinense; LCo, Lindera communis; LF, Liquidambar formosana; MM, Michelia macclurei; PM, Pinus massoniana; PS, Photinia serrulate; SM, Sapindus mukorossi; SS, Schima superba.

FIG 3

The NMDS ordination showing the fungal community composition among the three compartments of leaf, root, and soil (A) and the results of a beta dispersion analysis to show the fungal assembly extent (B). R² and P values were obtained from the PERMANOVA (A), and F and P values were obtained from the ANOVA based on the beta dispersion analysis results (B). Bars without shared letters denote significant differences among the three compartments (B).

FIG 4

Non-metric multidimensional scaling (NMDS) ordinations of the fungal community compositions in 13 plant species within each of the three compartments of leaf (A), root (B), and soil (C). Ellipses indicate 95% confidence intervals around the centroids of the different plants. R² and P values represent the results of a PERMANOVA (adonis). Plant species names: CCam, Cinnamomum camphora; CCar, Castanopsis carlesii; CL, Cunninghamia lanceolate; ED, Elaeocarpus decipiens; KB, Koelreuteria bipinnata; LCh, Liriodendron chinense; LCo, Lindera communis; LF, Liquidambar formosana; MM, Michelia macclurei; PM, Pinus massoniana; PS, Photinia serrulate; SM, Sapindus mukorossi; SS, Schima superba.

FIG 5

Preferences observed in plant-leaf (A), plant-root (B), and plant-soil (C) fungal associations. The standardized d’ estimate of preferences for fungal OTUs is shown for each plant species (column). Likewise, the standardized d’ estimate of preferences for plant species is indicated for each of the observed fungal OTUs (row). Each cell in the matrix indicates a two-dimensional preference (2DP) estimate, which measures the extent to which the association of a focal plant-fungal pair was observed more or less frequently than would be expected by chance. Black lines indicate that the two-dimensional preference (2DP) between a fungus and a plant was not available (NA). The P values were adjusted based on the false discovery rate (FDR).

FIG 6

The neutral community model shows that the neutral interpretation has a good fit to fungal community distribution for the leaf (A), root (B), soil (C), and total (D) fungal data sets. The predicated occurrence frequency is shown as a solid blue line, and dashed blue lines represent 95% confidence intervals around the model prediction. Red and green dots indicate the fungal operational taxonomic units that occur less and more frequently than given by the model. R² and m indicate the degree of fitting to the neutral community model and the immigration rate, respectively. The comparison of the mean habitat niche breadth (Bcom) in all taxa among the three compartments (subsets) and the total fungal community (Kruskal-Wallis test) is shown together (E).