Compositions of fungal and bacterial guilds in the rhizosphere soil across 52 tree species.

The relative abundance of fungal (upper panel) and bacterial (lower panel) guilds is estimated based on sequencing reads. Compositions of the five most abundant ectomycorrhizal (EcM) and saprotrophic genera are also shown. Abbreviations of EcM tree species are in red font, while arbuscular mycorrhizal (AM) tree species are in black font. Complete scientific names of trees are provided in Table S1.

Distance-based redundancy analysis (dbRDA) to determine the predictor variables that significantly influenced the functional compositions of rhizosphere fungal and bacterial communities.

Predictor variables include elevation, host mycorrhizal type, absorptive root traits, and root-zone soil properties. Results show marginal tests based on the Manhattan distance matrix, where var% indicates the relative contributions of predictor variables to fungal and bacterial guild dissimilarity. The marginal test assess each marginal term analyzed in a model with all other variables.

The relative abundance of AM fungi is estimated by ITS sequencing in this table. Significant statistics (P < 0.05) are indicated in bold.

The gradient of rhizosphere fungal guilds in the root economics space.

Principal component analysis followed by varimax rotation was performed on root traits (grey), root-zone soil properties (green), and relative abundances of different fungal guilds (purple). In this trait space, there is a gradient ranging from ectomycorrhizal (EcM) to saprotrophic (SAP) fungal dominance. Greater relative abundance of plant pathogenic fungi (PATH), arbuscular mycorrhizal fungi (AMF, estimated by ITS sequencing here), and higher root nitrogen concentration (N) is associated with the dominance of SAP fungi. Other root traits, including root diameter (D), specific root length (SRL), and partly root tissue density (RTD), comprise the root morphology gradient. Root-zone soil properties, including root-zone pH (PH), gravimetric water content (SW), total carbon (SC), and nitrogen (SN) concentration, mainly occupy the gradient decoupled from most root traits and fungal guilds. Four representative tree species are highlighted with pictures of absorptive roots presented at the same scale: Quercus variabilis (Quva); Pinus taiwanensis (Pita); Torreya grandis (Togr); Aphananthe aspera (Apas).

The gradient of rhizosphere bacterial guilds in the root economics space.

Principal component analysis followed by varimax rotation was performed on root traits (gray), root-zone soil properties (green), and relative abundances of different bacterial guilds (purple). The variation of most bacterial guilds, including aromatic compound degradation (AR_DEG), chitinolysis (CHITIN), ureolysis (UREOL), nitrogen fixation (N_FIX), nitrification (NITRI), nitrate reduction (N_REDU), methylotrophy (METH), phototrophy (PHOTO), and other aerobic chemoheterotrophy (A_CHEM), is strongly related to root-zone pH (pH), but not gravimetric water content (SW), total carbon (SC), and nitrogen (SN) concentration. The pH gradient is also largely decoupled from the traditional root economics traits, including root diameter (D), specific root length (SRL), root nitrogen concentration (N), and root tissue density (RTD). Two representative tree species are highlighted with pictures of absorptive roots presented at the same scale: Cryptomeria japonica var. sinensis (Crja) and Juglans mandshurica (Juma).

Shifts in rhizosphere fungal and bacterial functional compositions along the root-zone pH gradient.

The relative abundance of different fungal (top panel) and bacterial (middle panel) guilds in the rhizosphere is standardized, plotted along the pH gradient, and fitted with a smoothed curve. The relative abundance of arbuscular mycorrhizal fungi (AMF) is estimated by ITS sequencing in this figure. In the bottom panel, root-zone pH values of the common tree species in our dataset (sample size ≥ 4) are shown (box plots), with the shaded area indicating the density curve of pH across all 336 tree hosts. All three panels share the same pH scale.

Compositions of fungal and bacterial guilds in the root tissue across 52 tree species.

The relative abundance of fungal (upper panel) and bacterial (lower panel) guilds is estimated based on sequencing reads. Compositions of the five most abundant ectomycorrhizal and saprotrophic genera are also shown. Abbreviations of ectomycorrhizal tree species are in red font while arbuscular mycorrhizal tree species are in black font. Complete scientific names of trees are provided in Table S1.

The gradient of fungal guilds (x-axis) in the root economics space.

Principal component analyses followed by varimax rotation were performed based on on root traits (grey), root-zone soil properties (green) and relative abundances of different fungal guilds (purple). The data are analyzed at either the individual (a-c) or the species level (d-f). The fungal communities are sampled from either rhizosphere soil (a-b, d-e) or root tissue (c, f). Within the rhizosphere fungal communities, the relative abundance of AM fungi is estimated using either the sequencing method (a, c, d, f, AMF_S) or the qPCR method (b, e, AMF_Q). Statistics of the PCAs are shown in Tables S4-S6. See Fig. 1 for abbreviations of variables.

The gradient of bacterial guilds (x-axis) in the root economics space.

Principal component analyses followed by varimax rotation were performed based on root traits (grey), root-zone soil properties (green) and relative abundances of different bacterial guilds (purple). The data are analyzed at either the individual (a-b) or the species level (c-d). The bacterial communities are sampled from either rhizosphere soil (a, c) or root tissue (b, d). Statistics of the PCAs are shown in Tables S7-S8. See Fig. 2 for abbreviations of variables.

Relative abundance of different fungal and bacterial guilds along the rhizosphere pH intervals.

Relative abundance of AM fungi is estimated using either the qPCR method (AMF_Q) or the sequencing method (AMF_S). Data are ln(x+0.001) transformed. See Figures 1&2 for abbreviations of fungal and bacterial guilds.

Tree species in this study.

Both arbuscular mycorrhizal (AM) and ectomycorrhizal (EM) tree hosts, either evergreen or deciduous, were selected. Only the first three order roots were used for morphology and tissue chemistry measurements. Values were averaged across individuals of the same species. Root-zone soil properties were measured within ca. 5 cm of root surface. D = root diameter (mm), SRL = specific root length (m g-1), N = root nitrogen concentration (mg g-1), RTD = root tissue density (g cm-3), pH = root-zone soil pH, SW = gravimetric soil water content (%), SC = soil total carbon (mg g-1), SN = soil total nitrogen (mg g-1).

Summary of the root-associated fungal and bacterial communities.

Full names of fungal and bacterial guilds are provided in Figs 1&2. The relative abundance of AM fungi is estimated using either the qPCR method (AMF_Q) or the sequencing method (AMF_S).

Distance-based redundancy analysis (dbRDA) to determine the predictor variables that significantly influenced the functional compositions of fungal and bacterial communities within the roots.

Predictor variables include elevation, host mycorrhizal type, absorptive root traits, and root-zone soil properties. Results show marginal tests based on the Manhattan distance matrix, where var% indicates the relative contributions of predictor variable to fungal and bacterial guild dissimilarity. The relative abundance of AM fungi is estimated by ITS sequencing in this table. Significant statistics (P < 0.05) are indicated in bold.

Statistics of the root economics space incorporating fungal guilds across individual trees.

Shown are the proportion of variance and loadings of the relative abundance of fungal guilds, root traits, and root-zone soil properties in the most significant components of the principal component analyses followed by varimax rotation. The relative abundance of AMF is estimated by ITS sequencing. Loadings between -0.1 and 0.1 are not shown. Full names of fungal guilds, root traits, and soil properties are provided in Figure 1.

Statistics of the fungal-extended root economics space at the species level.

Shown are the proportion of variance and loadings in the most significant components of the principal component analyses followed by varimax rotation. The relative abundance of fungal guilds, root traits and soil prosperities are averaged across individuals of the same species. The relative abundance of AM fungi is estimated by ITS sequencing. Loadings between -0.1 and 0.1 are not shown. Full names of fungal guilds, root traits, and soil properties are provided in Figure 1.

Statistics of the fungal-extended root economics space with qPCR-based relative abundance of AM fungi.

Only the relative abundance of fungal guilds in the rhizosphere were included in the analyses. Shown are the proportion of variance and loadings in the most significant components of the principal component analyses followed by varimax rotation at both individual and species levels. Loadings between -0.1 and 0.1 are not shown. Full names of fungal guilds, root traits, and soil properties are provided in Figure 1.

Statistics of the root economics space incorporating bacterial communities across individual tree hosts.

Shown are the proportion of variance and loadings of the relative abundance of bacterial guilds, root traits, and root-zone soil properties in the most significant four or five components of the principal component analyses followed by varimax rotation. Loadings between -0.1 and 0.1 are not shown. Full names of bacterial guilds, root traits, and soil properties are provided in Figure 2.

Statistics of the bacterial-extended root economics space at the species level.

Shown are the proportion of variance and loadings in the most significant components of the principal component analyses followed by varimax rotation. The relative abundance of bacterial guilds, root traits and root-zone soil prosperities are averaged across individuals of the same species. Loadings between -0.1 and 0.1 are not shown. Full names of fungal guilds, root traits, and soil properties are provided in Figure 2.

Influence of elevation and tree species on rhizosphere pH.

Bold values indicate significant effects at P<0.05.

Principal component analysis on common root traits and root-zone soil pH.

Shown are the eigenvalues, proportion of variance and loadings of the first three principal components, without loadings between -0.1 and 0.1.