Bacterial Physiological Adaptations to Contrasting Edaphic Conditions Identified Using Landscape Scale Metagenomics

Abstract

Environmental factors relating to soil pH are important regulators of bacterial taxonomic biodiversity, yet it remains unclear if such drivers affect community functional potential. To address this, we applied whole-genome metagenomics to eight geographically distributed soils at opposing ends of a landscape soil pH gradient (where "low-pH" is ~pH 4.3 and "high-pH" is ~pH 8.3) and evaluated functional differences with respect to functionally annotated genes. First, differences in taxonomic and functional diversity between the two pH categories were assessed with respect to alpha diversity (mean sample richness) and gamma diversity (total richness pooled for each pH category). Low-pH soils, also exhibiting higher organic matter and moisture, consistently had lower taxonomic alpha and gamma diversity, but this was not apparent in assessments of functional alpha and gamma diversity. However, coherent changes in the relative abundances of annotated genes between low- and high-pH soils were identified; with strong multivariate clustering of samples according to pH independent of geography. Assessment of indicator genes revealed that the acidic organic-rich soils possessed a greater abundance of cation efflux pumps, C and N direct fixation systems, and fermentation pathways, indicating adaptations to both acidity and anaerobiosis. Conversely, high-pH soils possessed more direct transporter-mediated mechanisms for organic C and N substrate acquisition. These findings highlight the distinctive physiological adaptations required for bacteria to survive in soils of various nutrient availability and edaphic conditions and more generally indicate that bacterial functional versatility with respect to functional gene annotations may not be constrained by taxonomy.IMPORTANCE Over a set of soil samples spanning Britain, the widely reported reductions in bacterial taxonomic richness at low pH were found not to be accompanied by significant reductions in the richness of functional genes. However, consistent changes in the abundance of related functional genes were observed, characteristic of differential ecological and nutrient acquisition strategies between high-pH mineral soils and low-pH organic anaerobic soils. Our assessment at opposing ends of a soil gradient encapsulates the limits of functional diversity in temperate climates and identifies key pathways that may serve as indicators for soil element cycling and C storage processes in other soil systems. To this end, we make available a data set identifying functional indicators of the different soils; as well as raw sequences, which given the geographic scale of our sampling should be of value in future studies assessing novel genetic diversity of a wide range of soil functional attributes.

Keywords: ecophysiology; metagenomics; soil microbiology.

FIG 1

Geographically distributed soils from a range of habitats sampled at opposing ends of a landscape pH gradient. The sampling locations are displayed over a soil pH map of Britain derived from the United Kingdom Soils portal (www.ukso.org).

FIG 2

Within-site and across-site taxonomic and functional richness represented by site-based accumulation curves. Standard deviations are calculated from random permutations of the data, with red lines representing low pH and blue lines high pH. (a) Taxonomic richness determined by 16S rRNA sequencing is higher at high pH, both within individual sites (alpha diversity) and accumulated across sites (gamma diversity). (b) Richness of annotated functional genes following whole-genome sequencing is only marginally lower in low pH at individual sites and converges across all sites.

FIG 3

Abundances of annotated functional genes classified at the broadest level (level 1 subsystem classification), with total reads standardized across samples to 92,442 reads.

FIG 4

(a) Ordination of functional genes (classified at the level of function) using two-dimensional nonmetric multidimensional scaling (NMDS) reveals strong clustering of sites by pH irrespective of geographical sampling origin. (b) Network depicting strong (>0.9) positive correlations between annotated functional genes. For clarity, rare genes with less than 10 reads across all samples were omitted. Indicator functional genes are colored according to pH class following indicator (IndVal) analyses.

FIG 5

(Top) Bar plot showing the frequency (freq) of indicator genes at the broad level 1 classification. (Bottom) Circular plot displaying the identity and abundances of significant indicator genes for low- and high-pH soils. Nodes represent individual functional indicators, although they are labeled with the more descriptive subsystem level 3 classification (i.e., repeated node labels indicate different functional indicators within the same level 3 subsystem classification). Node labels are colored red and blue for particular genes that are significantly more abundant in low- or high-pH soils, respectively. Line plots represent total abundances of the indicators within the rarefied data sets and are filled according to pH (red, low; blue, high). The tree depicts the hierarchical subsystem classification, with level 1 classifications being labeled on the internal nodes.

FIG 6

Schematic summarizing some of the main physiological differences for survival, nutrient acquisitions, and substrate metabolism across the pH gradient, as identified from the indicator analyses. We note that inclusion of a gene in either schematic is based on differences in abundances and does not implicate the presence/absence of a particular pathway across the gradient (refer to the circular plot in Fig. 5).