The Inter-Valley Soil Comparative Survey: the ecology of Dry Valley edaphic microbial communities

Abstract

Recent applications of molecular genetics to edaphic microbial communities of the McMurdo Dry Valleys and elsewhere have rejected a long-held belief that Antarctic soils contain extremely limited microbial diversity. The Inter-Valley Soil Comparative Survey aims to elucidate the factors shaping these unique microbial communities and their biogeography by integrating molecular genetic approaches with biogeochemical analyses. Although the microbial communities of Dry Valley soils may be complex, there is little doubt that the ecosystem's food web is relatively simple, and evidence suggests that physicochemical conditions may have the dominant role in shaping microbial communities. To examine this hypothesis, bacterial communities from representative soil samples collected in four geographically disparate Dry Valleys were analyzed using molecular genetic tools, including pyrosequencing of 16S rRNA gene PCR amplicons. Results show that the four communities are structurally and phylogenetically distinct, and possess significantly different levels of diversity. Strikingly, only 2 of 214 phylotypes were found in all four valleys, challenging a widespread assumption that the microbiota of the Dry Valleys is composed of a few cosmopolitan species. Analysis of soil geochemical properties indicated that salt content, alongside altitude and Cu(2+), was significantly correlated with differences in microbial communities. Our results indicate that the microbial ecology of Dry Valley soils is highly localized and that physicochemical factors potentially have major roles in shaping the microbiology of ice-free areas of Antarctica. These findings hint at links between Dry Valley glacial geomorphology and microbial ecology, and raise previously unrecognized issues related to environmental management of this unique ecosystem.

Figure 1

MDS plot of soil physicochemical profiles. An Euclidean distance matrix of soil physicochemical profiles was calculated using soil physicochemical properties that have been square-root and log (X+1) transformed and normalized. A MDS plot was then generated using the distance matrix to represent the relative distances between individual samples. The samples are represented using the following symbols: Miers Valley (MV), gray triangles (); Beacon Valley (BV), upturned black triangles (▾); Battleship Promontory (BP), gray squares (); Upper Wright Valley (UW), gray diamonds (). Contour lines enclose samples with distances below certain thresholds (0.22, solid line; 0.34, dashed line).

Figure 2

MDS plots of bacterial (a) and cyanobacterial (b) ARISA profiles. Bray–Curtis similarity matrices of bacterial and cyanobacterial ARISA profiles were calculated based on presence/absence of AFLs. MDS plots were then generated using the distance matrices to represent the relative distances between individual samples. The samples are represented using the following symbols: Miers Valley (MV), gray triangles (); Beacon Valley (BV), upturned black triangles (▾); Battleship Promontory (BP), gray squares (); Upper Wright Valley (UW), gray diamonds (). Contour lines enclose samples with similarities above certain thresholds (bacterial ARISA: 0.20, solid line; 0.40, dashed line) (cyanobacterial ARISA: 0.20, solid line).

Figure 3

Phylum-level distribution of bacterial OTUs_0.03. Representative sequences of OTUs_0.03 for each study site were phylogenetically assigned using the Ribosomal Database Project Classifier, and phylum-level assignments with >80% confidence threshold were retrieved and plotted.

Figure 4

Venn diagram for bacterial OTUs_0.03 found in the four Dry Valleys.