What matters most? Assessment of within-canopy factors influencing the needle microbiome of the model conifer, Pinus radiata

Abstract

The assembly and function of the phyllosphere microbiome is important to the overall fitness of plants and, thereby, the ecosystems they inhabit. Presently, model systems for tree phyllosphere microbiome studies are lacking, yet forests resilient to pests, diseases, and climate change are important to support a myriad of ecosystem services impacting from local to global levels. In this study, we extend the development of model microbiome systems for trees species, particularly coniferous gymnosperms, by undertaking a structured approach assessing the phyllosphere microbiome of Pinus radiata. Canopy sampling height was the single most important factor influencing both alpha- and beta-diversity of bacterial and fungal communities (p < 0.005). Bacterial and fungal phyllosphere microbiome richness was lowest in samples from the top of the canopy, subsequently increasing in the middle and then bottom canopy samples. These differences maybe driven by either by (1) exchange of microbiomes with the forest floor and soil with the lower foliage, (2) strong ecological filtering in the upper canopy via environmental exposure (e.g., UV), (3) canopy density, (4) or combinations of factors. Most taxa present in the top canopy were also present lower in tree; as such, sampling strategies focussing on lower canopy sampling should provide good overall phyllosphere microbiome coverage for the tree. The dominant phyllosphere bacteria were Alpha-proteobacteria (Rhizobiales and Sphingomonas) along with Acidobacteria Gp1. However, the P. radiata phyllosphere microbiome samples were fungal dominated. From the top canopy samples, Arthoniomycetes and Dothideomycetes were highly represented, with abundances of Arthoniomycetes then reducing in lower canopy samples whilst abundances of Ascomycota increased. The most abundant fungal taxa were Phaeococcomyces (14.4% of total reads) and Phaeotheca spp. (10.38%). A second-order effect of canopy sampling direction was evident in bacterial community composition (p = 0.01); these directional influences were not evident for fungal communities. However, sterilisation of needles did impact fungal community composition (p = 0.025), indicating potential for community differences in the endosphere versus leaf surface compartments. Needle age was only important in relation to bacterial communities, but was canopy height dependant (interaction p = 0.008). By building an understanding of the primary and secondary factors related to intra-canopy phyllosphere microbiome variation, we provide a sampling framework to either explicitly minimise or capture variation in needle collection to enable ongoing ecological studies targeted at inter-canopy or other experimental levels.

Keywords: Community assembly; Conifer; Microbiome; Model system; Phyllosphere.

Fig. 1

Needle sampling strategy; different tree heights, cardinal directions, needle age, and needle compartment. *note that Pinus radiata typically has 3 needles per fascicle. This graphic is for demonstration only

Fig. 2

P. radiata needle microbiome α and β diversity summary plots at different tree heights. (A) bacterial ASV-based Chao1 ‘species’ richness (p = < 0.001), (B) fungal Chao1 richness (p = < 0.001), (C) bacterial Venn diagram displaying ASVs at three different canopy collection, and (D) the fungal Venn diagram also partitioned into canopy height factors. Ordinations (nMDS) using Class-level phylogenetic classification are presented in (E) for bacterial communities from the bottom and middle canopy samples only, and (F) fungal communities

Fig. 3

Heat trees based on bacterial taxa counts at order level from needles collected from the (A) top, (B) middle, and (C) bottom portions of the P. radiata canopy. The size and colour of the nodes and edges are correlated with the abundance of bacterial taxa in the community

Fig. 4

Heat map of pairwise comparisons of the P. radiata needle bacterial microbiome among different canopy heights from a P. radiata tree. Only significant differences are coloured, determined using a Wilcox rank-sum test followed by a Benjamini-Hochber false discovery rate (FDR) correction. Taxa coloured green are enriched in the part of the tree shown in the row (i.e., none in this case) and those coloured brown are enriched in the part of the tree show in the column

Fig. 5

Heat trees based on ASV counts of fungal taxa at order level from needles collected from the (A) top, (B) middle, and (C) and bottom portions of the Pinus radiata canopy. The size and colour of the nodes and edges are correlated with the abundance of fungal ASVs in the community

Fig. 6

Heat map to show pairwise comparisons of the needle fungal communities in different canopy height sections of a P. radiata tree. Only significant differences are coloured, determined using a Wilcox rank-sum test followed by a Benjamini-Hochber FDR correction for multiple comparisons. Taxa coloured green are enriched in the part of the tree shown in the row and those coloured brown are enriched in the part of the tree show in the column