Down-regulation of the bacterial protein biosynthesis machinery in response to weeks, years, and decades of soil warming

Abstract

How soil microorganisms respond to global warming is key to infer future soil-climate feedbacks, yet poorly understood. Here, we applied metatranscriptomics to investigate microbial physiological responses to medium-term (8 years) and long-term (>50 years) subarctic grassland soil warming of +6°C. Besides indications for a community-wide up-regulation of centralmetabolic pathways and cell replication, we observed a down-regulation of the bacterial protein biosynthesis machinery in the warmed soils, coinciding with a lower microbial biomass, RNA, and soil substrate content. We conclude that permanently accelerated reaction rates at higher temperatures and reduced substrate concentrations result in cellular reduction of ribosomes, the macromolecular complexes carrying out protein biosynthesis. Later efforts to test this, including a short-term warming experiment (6 weeks, +6°C), further supported our conclusion. Down-regulating the protein biosynthesis machinery liberates energy and matter, allowing soil bacteria to maintain high metabolic activities and cell division rates even after decades of warming.

Fig. 1.. Grassland soil samples and differences in physicochemical and biological properties.

(A) Schematic overview of the ForHot experimental warming sites and analyzed samples (see Materials and Methods for more details). (B) Average x-fold change of soil physicochemical and biological properties in response to warming (see table S1 for absolute values). Values <1 represent measures that are lower in the warmed soils (i.e., higher in A_T, highlighted in blue), and values >1 represent measures that are higher in the warmed soils (highlighted in red). Triangles and dots (left) indicate site and temperature contrast (average x-fold change of MTW-E_T versus MTW-A_T, LTW-E_T versus LTW-A_T, and E_T versus A_T). Dots (right) indicate site and temperature contrast with significant differences in the respectively aligned physicochemical and biological properties (two-way ANOVAs on each individual property, TukeyHSD test to correct for multiple testing; see table S2 for the full statistical analysis and exact P values). (C) Nucleic acid (DNA and RNA) content per unit of microbial C (boxplots) and correlation between microbial C (C_MO) and RNA content per gram dry weight soil (scatterplot). See tables S2 and S3 for the full statistical analysis and exact P values (*P_adj < 0.05).

Fig. 2.. Enzymes and enzyme complexes involved in protein biosynthesis and oxidative phosphorylation.

(A) Relative mRNA read abundances of protein biosynthesis complexes in the MTW-A_T, MTW-E_T, LTW-A_T, and LTW-E_T metatranscriptomes (boxplots) and DGE analysis results on associated KOs (table). (B) Relative mRNA read abundances of complexes involved in membrane-bound electron transport and ATP synthesis (oxidative phosphorylation) in the MTW-A_T, MTW-E_T, LTW-A_T, and LTW-E_T metatranscriptomes (boxplots) and DGE analysis results on associated KOs (table). Schematic representations of the enzymes and enzyme complexes are provided above the boxplots and are based on the KEGG pathway drawings. Membrane-bound complexes are embedded in a lipid bilayer. See table S16 for relative abundances of all individual KOs summarized in the boxplots and details on the DGE analysis.

Fig. 3.. Enzymes and enzyme complexes involved in central carbohydrate metabolism and cell replication and taxonomic distribution of overall expression patterns.

(A) Metabolic pathway map focusing on carbohydrate metabolism, associated reactions, and relative mRNA read abundances assigned to these reactions. Reactions are represented by EC numbers (EC-no.) and derived from KEGG pathway drawings. Color code indicates mean relative expression levels (red, higher relative expression levels in E_T; blue, higher relative expression levels in A_T); intensities additionally reflect the presence of differentially expressed KOs and KOs with an FDR of 0.05 to 0.1. Bold fonts indicate Metabolism subcategories (KEGG3); dashed arrows indicate connections to subcategories for which no details are given; the color code (as above) of the frames around these categories indicates if the whole subcategory shows indications of a temperature-dependent response. See table S17 for details. (B and C) Cell replication. Relative mRNA read abundances of DNA polymerase III subunits, bacterial initiation factors (B) and FtsZ (C) in the metatranscriptomes (boxplots), and DGE analysis results on associated KOs (table; see table S16 for details). Schematic illustrations are provided above the boxplots. (D) Taxonomic distribution of observed overall expression patterns on abundant KEGG3 categories comprising genes encoding key functions in central carbohydrate metabolism with higher expression levels and genes encoding key functions in energy metabolism and protein biosynthesis with lower expression levels in the warmed soils, respectively. Pyv MB, pyruvate metabolism; Gly/Glu, glycolysis/gluconeogenesis; C-fix PWs, C fixation pathways in prokaryotes; Ox.phos, oxidative phosphorylation; RNA pol, RNA polymerase; P., protein. Bold numbers aligned to a KEGG3 category highlight taxonomic ranks in which the taxon-specific expression pattern of more than two-thirds of all included taxa (i.e., all taxa with a mean relative abundance of ≥1‰) reflected the global expression pattern of the respective KEGG3 category (see also fig. S11).

Fig. 4.. Hypothesized reduction of cellular ribosome content as response to warming and in situ and experimental indications.

(A) Schematic representations of microbial cells and cellular constituents under different temperature regimes. The cells vary in the number of ribosomes, ratio between ribosomal and metabolic proteins, amount of membrane-bound complexes, and slightly in size. (B) Schematic overview (continuation of Fig. 1A) of the soil sampling conducted in October 2020. The same geothermal soil temperature gradients were samples as in the metatranscriptomics sampling campaign 2016 (see Materials and Methods for more details). (C) Comparison of in situ RNA content per gram dry weight (DW) soil between the same plots sampled in two different seasons (summer versus autumn) and 4 years apart (2016 versus 2020). (D) Schematic overview and results of a short-term warming experiment conducted with soil sampled from an ambient temperature plot at LTW (see Materials and Methods for more details). After 3 weeks (w) of preincubation of the homogenized soil at ambient temperatures (i.e., mean ambient October soil temperature of 7°C), the first samples (t0, n = 5) were taken, and afterward, the soil was distributed to 10 serum bottles. The aerated bottles were incubated for 6 weeks at 7°C (control, n = 5) and 13°C (warming treatment, +6°C above ambient, n = 5), respectively, and sampled after 1, 3, and 6 weeks of incubation. Line charts depict incubation temperatures, mean total RNA content per gram DW soil, and mean total RNA content per milligram of microbial C (C_MO), and error bars represent SD of the mean. For underlying data and details on the statistical analysis presented in (C) and (D), see tables S22 to S24 (*P < 0.05; ***P < 0.001).