Community proteogenomics reveals the systemic impact of phosphorus availability on microbial functions in tropical soil

Abstract

Phosphorus is a scarce nutrient in many tropical ecosystems, yet how soil microbial communities cope with growth-limiting phosphorus deficiency at the gene and protein levels remains unknown. Here, we report a metagenomic and metaproteomic comparison of microbial communities in phosphorus-deficient and phosphorus-rich soils in a 17-year fertilization experiment in a tropical forest. The large-scale proteogenomics analyses provided extensive coverage of many microbial functions and taxa in the complex soil communities. A greater than fourfold increase in the gene abundance of 3-phytase was the strongest response of soil communities to phosphorus deficiency. Phytase catalyses the release of phosphate from phytate, the most recalcitrant phosphorus-containing compound in soil organic matter. Genes and proteins for the degradation of phosphorus-containing nucleic acids and phospholipids, as well as the decomposition of labile carbon and nitrogen, were also enhanced in the phosphorus-deficient soils. In contrast, microbial communities in the phosphorus-rich soils showed increased gene abundances for the degradation of recalcitrant aromatic compounds, transformation of nitrogenous compounds and assimilation of sulfur. Overall, these results demonstrate the adaptive allocation of genes and proteins in soil microbial communities in response to shifting nutrient constraints.

Novel phosphate-solubilizing bacteria enhance soil phosphorus cycling following ecological restoration of land degraded by mining

“…As an ecological and evolutionary factor, P is thought to override all other elements, including C and N [7]. However, while genome-centric metagenomics has recently allowed great advances in understanding the important role of microbes lacking prior genomic characterization in soil C and N cycling [58,59], few metagenome-assembled genomes encoding enzymes responsible for soil P cycling have been reported in the literature [60]. Here, we reconstructed 39 near-complete genomes involved in soil P cycling from 18 metagenomes associated with an ecological restoration project, which enabled us to deduce that certain previously unknown phosphate-solubilizing bacteria were the main drivers of the enhancement in soil P cycling following restoration.…”

“…Second, it is difficult to recover genomes from metagenomic data obtained from soil samples with high microbial diversity [59]. To the best of our knowledge, only two genome bins involved in microbial P solubilization (as indicated by the presence of genes encoding phytases) have been reported in the literature [60]. Despite this, the role of these two genome bins in soil P cycling remains unclear [60].…”

Contrasting patterns of plant and microbial diversity during long‐term ecosystem development

“…However, the soil pH reached only slightly to moderately acidic values and varied relatively little among the older stages of the chronosequence (pH 6.0–6.4), while the old soils are among the lowest phosphorus soils in the world (Turner & Laliberté, ; Yang, Post, Thornton, & Jain, ). Indeed, several bacteria increased markedly in abundance along the Jurien Bay sequence at resin phosphorus concentrations <2 mg P/kg, which is a threshold that drives variation in above‐ and below‐ground plant and microbial communities in lowland tropical forests (Sheldrake et al., ; Turner, Brenes‐Arguedas, & Condit, ; Yao et al., ). A role for phosphorus therefore seems likely, although experimental work is required to isolate this from the overall decline in soil pH along the chronosequence.…”

Spatial and temporal turnover of soil microbial communities is not linked to function in a primary tropical forest

“…Earth system and ecosystem models that incorporate microbial composition (Sulman et al , ), dormancy (Wang et al ), functional genes and traits (Yao et al ), and physiology (Wang et al , Wieder et al ) more accurately predict soil C storage and fluxes than those that do not (Wieder et al ). Although overall model accuracy is still low, incorporating spatial variation in microbial functional groups (Sulman et al ) improves model fit.…”