Response of archaeal communities in the rhizosphere of maize and soybean to elevated atmospheric CO2 concentrations

Abstract

Background: Archaea are important to the carbon and nitrogen cycles, but it remains uncertain how rising atmospheric carbon dioxide concentrations ([CO(2)]) will influence the structure and function of soil archaeal communities.

Methodology/principal findings: We measured abundances of archaeal and bacterial 16S rRNA and amoA genes, phylogenies of archaeal 16S rRNA and amoA genes, concentrations of KCl-extractable soil ammonium and nitrite, and potential ammonia oxidation rates in rhizosphere soil samples from maize and soybean exposed to ambient (∼385 ppm) and elevated (550 ppm) [CO(2)] in a replicated and field-based study. There was no influence of elevated [CO(2)] on copy numbers of archaeal or bacterial 16S rRNA or amoA genes, archaeal community composition, KCl-extractable soil ammonium or nitrite, or potential ammonia oxidation rates for samples from maize, a model C(4) plant. Phylogenetic evidence indicated decreased relative abundance of crenarchaeal sequences in the rhizosphere of soybean, a model leguminous-C(3) plant, at elevated [CO(2)], whereas quantitative PCR data indicated no changes in the absolute abundance of archaea. There were no changes in potential ammonia oxidation rates at elevated [CO(2)] for soybean. Ammonia oxidation rates were lower in the rhizosphere of maize than soybean, likely because of lower soil pH and/or abundance of archaea. KCl-extractable ammonium and nitrite concentrations were lower at elevated than ambient [CO(2)] for soybean.

Conclusion: Plant-driven shifts in soil biogeochemical processes in response to elevated [CO(2)] affected archaeal community composition, but not copy numbers of archaeal genes, in the rhizosphere of soybean. The lack of a treatment effect for maize is consistent with the fact that the photosynthesis and productivity of maize are not stimulated by elevated [CO(2)] in the absence of drought.

Figure 1. Abundance of archaeal and bacterial 16S rRNA and amoA genes from rhizosphere soil samples collected during the a) 2006 and b) 2008 growing seasons based upon quantitative PCR.

The abbreviations are as follows: maize ambient [CO₂] (Ma), maize elevated [CO₂] (Me), soybean ambient [CO₂] (Sa), and soybean elevated [CO₂] (Se). Mean values (+/− one standard deviation) are shown. Letters indicate statistical differences among treatments.

Figure 2. Phylogenetic trees of archaeal 16S rRNA gene sequences from rhizosphere soil samples from soybean grown at ambient and elevated [CO₂].

The colored bars represent the relative abundances of representative sequences from individual plots for each [CO₂] treatment (Table S1). Each colored bar in each phylogenetic tree represents data from a different plot. Starting from the root, the colors for the leaf ranges indicate crenarchaeota group 1.1a (green), crenarchaeota group 1.1c (dark blue), euryarchaeota (tan), and crenarchaeota group 1.1b (light blue). Clusters within crenarchaeota group 1.1b were further divided into arbitrarily named groups as shown (e.g. 1.1b_1). Only branches with bootstrap support >60 are shown, and identical branch lengths are shown.

Figure 3. Principle component analysis of a) archaeal 16S rRNA and b) archaeal amoA gene sequence data from phylogenetic trees (Figs. 2, S3, S5) of samples collected during the 2006 growing season.

PCA was performed in UniFrac using abundance weights. The symbols are as follows: soybean elevated [CO₂] (Se), diamonds; soybean ambient [CO₂] (Sa), triangles; maize elevated [CO₂] (Me) circles; maize ambient [CO₂] (Ma), squares. The mean PCA scores (+/− one standard deviation) are shown for each axis. For a), no plant/[CO₂] combinations differ on PCA2, whereas Se and Sa are statistically different (p<0.05) from each other and Me and Ma on PCA1. There is no difference (p = 0.72) between Me and Ma on PCA1. For b), no plant/[CO₂] combinations differ on PCA1, whereas Se is significantly different from Ma (p<0.05) on PCA2. Se is marginally different from Me (p = 0.08), and Sa from Ma (p = 0.11), on PCA2.

Figure 4. Percentages of archaeal lineages (identified in Figs. 2, S3) for which there were statistical differences among plant/[CO₂] combinations.

The abbreviations are as follows: maize ambient [CO₂] (Ma), maize elevated [CO₂] (Me), soybean ambient [CO₂] (Sa), and soybean elevated [CO₂] (Se). Mean values (+/− one standard deviation) are shown. Letters indicate statistical differences among treatments.

Figure 5. Average KCl-extractable ammonium and nitrite concentrations, pH, and potential rates of ammonia oxidation from rhizosphere soil samples collected during the 2008 growing season.

The abbreviations are as follows: maize ambient [CO₂] (Ma), maize elevated [CO₂] (Me), soybean ambient [CO₂] (Sa), and soybean elevated [CO₂] (Se). Mean values (+/− one standard deviation) are shown. Letters indicate statistical differences among treatments. For potential ammonia oxidation data, closed bars indicate assays to which antibiotics were added and open bars indicate assays to which antibiotics were not added, as described in the text. Statistical comparison of potential ammonia oxidation data is among assays with antibiotics and among assays without antibiotics.