Mineral Types and Tree Species Determine the Functional and Taxonomic Structures of Forest Soil Bacterial Communities

Abstract

Although minerals represent important soil constituents, their impact on the diversity and structure of soil microbial communities remains poorly documented. In this study, pure mineral particles with various chemistries (i.e., obsidian, apatite, and calcite) were considered. Each mineral type was conditioned in mesh bags and incubated in soil below different tree stands (beech, coppice with standards, and Corsican pine) for 2.5 years to determine the relative impacts of mineralogy and mineral weatherability on the taxonomic and functional diversities of mineral-associated bacterial communities. After this incubation period, the minerals and the surrounding bulk soil were collected to determine mass loss and to perform soil analyses, enzymatic assays, and cultivation-dependent and -independent analyses. Notably, our 16S rRNA gene pyrosequencing analyses revealed that after the 2.5-year incubation period, the mineral-associated bacterial communities strongly differed from those of the surrounding bulk soil for all tree stands considered. When focusing only on minerals, our analyses showed that the bacterial communities associated with calcite, the less recalcitrant mineral type, significantly differed from those that colonized obsidian and apatite minerals. The cultivation-dependent analysis revealed significantly higher abundances of effective mineral-weathering bacteria on the most recalcitrant minerals (i.e., apatite and obsidian). Together, our data showed an enrichment of Betaproteobacteria and effective mineral-weathering bacteria related to the Burkholderia and Collimonas genera on the minerals, suggesting a key role for these taxa in mineral weathering and nutrient cycling in nutrient-poor forest ecosystems.IMPORTANCE Forests are usually developed on nutrient-poor and rocky soils, while nutrient-rich soils have been dedicated to agriculture. In this context, nutrient recycling and nutrient access are key processes in such environments. Deciphering how soil mineralogy influences the diversity, structure, and function of soil bacterial communities in relation to the soil conditions is crucial to better understanding the relative role of the soil bacterial communities in nutrient cycling and plant nutrition in nutrient-poor environments. The present study determined in detail the diversity and structure of bacterial communities associated with different mineral types incubated for 2.5 years in the soil under different tree species using cultivation-dependent and -independent analyses. Our data showed an enrichment of specific bacterial taxa on the minerals, specifically on the most weathered minerals, suggesting that they play key roles in mineral weathering and nutrient cycling in nutrient-poor forest ecosystems.

Keywords: bacterial communities; forest soil; mesh bags; mineral chemistry; mineral weatherability.

FIG 1

Mineral weatherability under soil conditions and in a microcosm experiment. (A and B) Representative scanning electron micrographs from calcite particle observed before (A) and after (B) a 2.5-year incubation in the soil under a beech stand. As no surface alteration was observed for apatite and obsidian, scanning electron micrographs are not presented. (C) Estimation of the calcite mass loss after a 2.5-year incubation under beech stands (Bc), coppice with standards (CwS), and Corsican pine (Cp) stands. Each value is the mean of three (bulk soil) or four (minerals) replicates with standard error of the mean. As no mass loss was measured for apatite and obsidian, data are not presented. (D) Concentration of the Ca released from obsidian, apatite, and calcite particles over time in the microcosm experiment. (E) Relationship between the relative mineral mass loss and the percentage of Ca released for each mineral. This relationship was determined using the data generated in the mineral weatherability experiment. Three independent replicate reactors were done for each mineral type and are presented. (F) Evolution of the pH of the output solution (initial pH = 4) during the abiotic-dissolution assay performed with obsidian, apatite, and calcite. The dotted line represents the initial pH and corresponds to the pH measured in situ at the Breuil-Chenue experimental site.

FIG 2

Metabolic potentials of microbial communities based on Biolog EcoPlate analysis. The relative use of carbon-substrate guilds by microbial communities (amino acids, amines and amides, carboxylic acid, carbohydrates, and polymers) was estimated under beech stands (A), coppice with standards (B), and Corsican pine stands (C) and for each compartment considered (bulk soil, obsidian, apatite, and calcite). For the different substrate guilds, different capital letters (A, B, or C) above the bars indicate significant differences in the same tree stand according to a one-factor ANOVA and Tukey's multiple pairwise comparisons test, while different lowercase letters (x or y) indicate significant differences between the tree stands. Each value is the mean of three (bulk soil) or four (minerals) replicates, and the error bars indicate the standard errors of the means.

FIG 3

Relative distributions of the major bacterial taxa. The relative abundances of the major phyla and class were calculated as the percentage of sequences belonging to a particular lineage of all 16S rRNA gene sequences under beech and Corsican pine stands and under coppice with standards and for each compartment considered (bulk soil, obsidian, apatite, and calcite).

FIG 4

Multivariate analysis of the bacterial communities based on the relative abundances of bacterial phyla and proteobacterial classes estimated by 16S rRNA gene pyrosequencing analysis. Multivariate analysis was conducted separately for each type of tree stand: beech, coppice with standards, and Corsican pine. The same analysis was conducted for each mineral type, excluding the bulk soil samples: obsidian, apatite, and calcite. For legibility, the samples are presented as follows: orange, stand of beech; green, coppice with standards; blue, Corsican pine stand; circles, calcite samples; squares, apatite samples; triangles, obsidian samples; open diamonds, bulk soil samples. Vectors show the direction of maximum change for variables and longer arrows indicate a greater change in relative abundance. The percentages of the total variance explained by the first two axes, PC1 and PC2, are presented on each graph.

FIG 5

Relative distributions and efficacies of bacterial strains capable of mobilizing phosphorus, iron, and calcite. Pie charts represent the distribution of bacterial isolates recovered from bulk soil, obsidian, apatite, and calcite samples under beech stands, coppice with standards, and Corsican pine stands and tested on TCP medium (A), CAS medium (C), and CAL medium (E). Effective bacterial strains for each bioassay are represented in gray, and noneffective strains are represented in white. For each bioassay, the number of strains considered is presented in the center of the pie chart. For the same compartment, this number may vary between the different bioassays, due to the ability of the bacterial strains to grow on the different culture media. Significant differences in frequencies were evaluated by χ² test. Bar plots represent the averaged efficacy of the bacterial strains recovered from bulk soil, obsidian, apatite, and calcite samples under beech stands, coppice with standards, and Corsican pine stands and tested on TCP medium (B), CAS medium (D), and CAL medium (F). Efficacies were obtained by measuring the discoloration zone (in centimeters) around bacterial colonies on the different media. Significant differences in efficacy were evaluated according to a one-factor ANOVA and Tukey's multiple pairwise comparisons test. Each value is the mean of three (bulk soil) or four (minerals) replicates and the error bars indicate the standard error of the mean. For both pie charts and bar plots, different lowercase letters (a or b) indicate significant differences (P < 0.05) in the same tree stand, while different capital letters (X or Y) indicate significant differences (P < 0.05) between the tree stands.

FIG 6

Relationship between taxonomic belonging and functional efficacy. The averaged P, Fe, and calcite solubilization (in centimeters) were estimated for the major bacterial genera on TCP (A), CAS (B), and CAL (C) media, respectively. The number of strains tested for each bacterial genus and in each biotest is presented in parentheses. For the same genus, this number may vary between the different bioassays, due to the ability of the bacterial strains to grow on the different culture media. For each bioassay, different letters (a, b, and c) above the bars indicate significant differences between the bacterial genera (P < 0.05), according to a one-factor ANOVA and Tukey's multiple pairwise comparisons test (P < 0.05). The error bars indicate the standard errors of the means.