Long-Term Warming Alters Carbohydrate Degradation Potential in Temperate Forest Soils

Abstract

As Earth's climate warms, soil carbon pools and the microbial communities that process them may change, altering the way in which carbon is recycled in soil. In this study, we used a combination of metagenomics and bacterial cultivation to evaluate the hypothesis that experimentally raising soil temperatures by 5°C for 5, 8, or 20 years increased the potential for temperate forest soil microbial communities to degrade carbohydrates. Warming decreased the proportion of carbohydrate-degrading genes in the organic horizon derived from eukaryotes and increased the fraction of genes in the mineral soil associated with Actinobacteria in all studies. Genes associated with carbohydrate degradation increased in the organic horizon after 5 years of warming but had decreased in the organic horizon after warming the soil continuously for 20 years. However, a greater proportion of the 295 bacteria from 6 phyla (10 classes, 14 orders, and 34 families) isolated from heated plots in the 20-year experiment were able to depolymerize cellulose and xylan than bacterial isolates from control soils. Together, these findings indicate that the enrichment of bacteria capable of degrading carbohydrates could be important for accelerated carbon cycling in a warmer world.

Importance: The massive carbon stocks currently held in soils have been built up over millennia, and while numerous lines of evidence indicate that climate change will accelerate the processing of this carbon, it is unclear whether the genetic repertoire of the microbes responsible for this elevated activity will also change. In this study, we showed that bacteria isolated from plots subject to 20 years of 5°C of warming were more likely to depolymerize the plant polymers xylan and cellulose, but that carbohydrate degradation capacity is not uniformly enriched by warming treatment in the metagenomes of soil microbial communities. This study illustrates the utility of combining culture-dependent and culture-independent surveys of microbial communities to improve our understanding of the role changing microbial communities may play in soil carbon cycling under climate change.

FIG 1

Effect of chronic warming on the relative abundance of polysaccharide (polysacc.)-degrading genes in organic (A) and mineral (B) soil. Points denote mean percent difference in relative abundance of CAZymes between heated (H) and control (C) plots as a fraction of annotated reads, and are colored where regression coefficients of a negative binomial model regression differed between warmed and control plots (Benjamini-Hochberg corrected Wald test, P < 0.1 for individual sites, P < 0.05 for all sites together). Symbol size is proportionate to genome-standardized abundance in the control plots. Panels are separated into all sites analyzed jointly to look for an overall warming effect (“all sites”) or separately by site, in increasing order of experiment age.

FIG 2

Effects of experimental warming on the fraction of annotated reads assigned to dominant phyla in organic (A) and mineral (B) soil. Circles are plotted as the percent difference between warmed and control plot values, with the size proportionate to the number of polysaccharide-associated reads in the metagenome assigned to the phylum. Other parameters are as per Fig. 1.

FIG 3

Taxonomic distribution of polysaccharide-degrading genes for which the overall abundance was significantly affected by warming in both the organic horizon (top row) and mineral soil (bottom row); see Fig. 1. Differences in fractions of Pfam reads assigned to a given taxon and function were analyzed using a t test with Benjamini-Hochberg correction. (A) Barre Woods. (B) Prospect Hill. No genes were affected by warming treatment in both horizons at SWaN. “Other” includes all reads identifiable to at least the domain level. ∼, P < 0.1; *, P < 0.05.

FIG 4

Phylogenetic tree of bacterial isolates collected from warmed and control plots or immediately adjacent to experimental plots at Prospect Hill. Branches are colored according to phylum or class (for Proteobacteria). Inner ring of colors denotes whether isolate came from warmed (red) or control (blue) plots or outside the plots (green), while outer rings denote whether the isolate was able to degrade the polymer in a 4-day (CMC and xylan) or 11-day (chitin) assay on solid medium, as shown in the key. Breaks in color ring denote type strains inserted for orientation. Archaea were removed from the tree after building. Positions with a color in the color ring but no data for any substrate failed to grow in the plate-based assay.