Geology defines microbiome structure and composition in nunataks and valleys of the Sør Rondane Mountains, East Antarctica

Abstract

Understanding the relation between terrestrial microorganisms and edaphic factors in the Antarctic can provide insights into their potential response to environmental changes. Here we examined the composition of bacterial and micro-eukaryotic communities using amplicon sequencing of rRNA genes in 105 soil samples from the Sør Rondane Mountains (East Antarctica), differing in bedrock or substrate type and associated physicochemical conditions. Although the two most widespread taxa (Acidobacteriota and Chlorophyta) were relatively abundant in each sample, multivariate analysis and co-occurrence networks revealed pronounced differences in community structure depending on substrate type. In moraine substrates, Actinomycetota and Cercozoa were the most abundant bacterial and eukaryotic phyla, whereas on gneiss, granite and marble substrates, Cyanobacteriota and Metazoa were the dominant bacterial and eukaryotic taxa. However, at lower taxonomic level, a distinct differentiation was observed within the Cyanobacteriota phylum depending on substrate type, with granite being dominated by the Nostocaceae family and marble by the Chroococcidiopsaceae family. Surprisingly, metazoans were relatively abundant according to the 18S rRNA dataset, even in samples from the most arid sites, such as moraines in Austkampane and Widerøefjellet ("Dry Valley"). Overall, our study shows that different substrate types support distinct microbial communities, and that mineral soil diversity is a major determinant of terrestrial microbial diversity in inland Antarctic nunataks and valleys.

Keywords: Antarctica; bacteria; bedrock; eukaryotes; metabarcoding; microbial ecology; rRNA.

Figure 1

(A) Satellite image of Sør Rondane Mountains and sampled sites. (B) In situ photographs of the four types of soil ecosystems investigated. More information about the samples can be found and in Supplementary Figure 1. Contains modified Copernicus Sentinel-2 data [2020] and Landsat Image Mosaic of Antarctica (LIMA). Colors differ by sampled site. AU, Austkampane; DV, Dry Valley; PB_N, Perlebandet North; PB_S, Perlebandet South; PA4, Pingvinane 4th nunatak; PA6, Pingvinane 6th nunatak; PT, Petrellnuten; UT, Utsteinen; YO, Yûboku-dani Valley.

Figure 2

Relative abundances of the most abundant phyla (> 1% of total abundances) of bacteria per sample, grouped per substrate type. Colors are according to the Phylum (see legend). Sample names indicate the site of origin (see Supplementary Table S1 for details).

Figure 3

Relative abundances of the most abundant phyla (> 1% of total abundances) of eukaryotes per sample, grouped per substrate type. Colors are according to the Division (see legend). Sample names indicate the site of origin (see Supplementary Table S1 for details).

Figure 4

PCoA of 96 and 97 samples of (A) bacteria and (B) eukaryotes. Circles represent gneiss, triangles granite, squares marble and diamonds moraine samples. The percentage explained by the first two axes is given between brackets.

Figure 5

(A) Co-occurrence network analysis based on bacterial and eukaryotic ASVs (dots) across all sampling sites, colored according to the major identified modules in the different substrate types, with positive Spearman correlations > 0.7 plotted as edges. Size of nodes corresponds to the betweenness centrality degree. (B) Phylum-or Division-level taxonomic log-transformed abundance of those ASVs associated with each of the 11 modules, colors represent the abundance of ASVs per module, grey rectangles highlight eukaryotic phyla. (C) Relative abundance (z-score) of ASVs in each module plotted against the most important substrate geochemical variables identified by RF, Spearman correlation and Pearson correlation as the major explanatory variable for that module (p < 0.01, Holm-corrected). Module 5 represented a low number of samples. Modules 2 and 6 correlated well with decreasing TN values, similar to the plotted module 8.