Biogeochemical and historical drivers of microbial community composition and structure in sediments from Mercer Subglacial Lake, West Antarctica

Abstract

Ice streams that flow into Ross Ice Shelf are underlain by water-saturated sediments, a dynamic hydrological system, and subglacial lakes that intermittently discharge water downstream across grounding zones of West Antarctic Ice Sheet (WAIS). A 2.06 m composite sediment profile was recently recovered from Mercer Subglacial Lake, a 15 m deep water cavity beneath a 1087 m thick portion of the Mercer Ice Stream. We examined microbial abundances, used 16S rRNA gene amplicon sequencing to assess community structures, and characterized extracellular polymeric substances (EPS) associated with distinct lithologic units in the sediments. Bacterial and archaeal communities in the surficial sediments are more abundant and diverse, with significantly different compositions from those found deeper in the sediment column. The most abundant taxa are related to chemolithoautotrophs capable of oxidizing reduced nitrogen, sulfur, and iron compounds with oxygen, nitrate, or iron. Concentrations of dissolved methane and total organic carbon together with water content in the sediments are the strongest predictors of taxon and community composition. δ¹³C values for EPS (-25 to -30‰) are consistent with the primary source of carbon for biosynthesis originating from legacy marine organic matter. Comparison of communities to those in lake sediments under an adjacent ice stream (Whillans Subglacial Lake) and near its grounding zone provide seminal evidence for a subglacial metacommunity that is biogeochemically and evolutionarily linked through ice sheet dynamics and the transport of microbes, water, and sediments beneath WAIS.

Fig. 1. Locator map of SLM along the confluence of the Mercer and Whillans ice streams.

The drilling sites for SLM, SLW, and WGZ are marked with red dots. The blue lines represent the predicted water flow paths beneath WAIS [15] and the white line indicates the ice-sheet grounding line [63]. The background imagery is MODIS MOA2009 [61] with ice velocity overlain [62].

Fig. 2. Composite depth profile of analytes from SLM sediment cores.

Physical (a, b), geochemical (c–f), isotopic (i), and microbiological (g, h, and j) data are plotted. The Y-axis represents the composite sediment depth, as described in the Supplementary Information, and the dashed horizontal lines are the boundaries between the stratographic units. The open and closed circles for oxygen (c) represent two separate cores used for measurements, while the open circles for sulfide (e), EPS component (h), and cellular DNA (j) represent values too low to precisely measure and are plotted at the procedural detection limit. The oxygen concentration data are from Michaud et al. [47]; TOC and δ¹³C values are from Venturelli et al. [16].

Fig. 3. Bacterial and archaeal ASVs with a relative abundance >1% of total sequences in at least one of the four sediment units (I to IV) from SLM.

ASVs are listed in alphabetical order and with the percent 16S rRNA gene identity to their nearest taxonomic neighbor and its GenBank accession number. Bubble color indicates ASVs having a nearest neighbor with the capacity to use reduced iron and sulfur compounds (orange), ammonia (green), or C1 compounds (blue) as electron donors. Heat map colors represent the correlation coefficient for the relative abundance of each ASV to environmental variables using Spearman’s correlation test. ** = p ≤ .001; * = 0.05 ≥ p > .001.

Fig. 4. Redundancy analysis of microbial assemblages in the SLM sediment profile.

The first two dimensions of the ordination plot are shown for (a) environmental parameters and (b) ASVs that are explanatory variables of community structure (black arrows). The classification of ASVs is shown in Fig. 3 with the exception of ASV_270, which had a relative abundance of 0.15% of total sequences and classifies within the genus Pseudomonas. The two axes in the RDA model explain 38.15% of the cumulative variance in the community composition. Based on a permutation test of total variance for environmental variables and community composition (p value ≤0.001; ANOVA, 999 permutations), the dispersion of unit I in ordination space is significantly different from units II–IV.

Fig. 5. Comparison of communities and taxa in sediment samples from SLM, SLW, and the marine WGZ.

a Nonmetric multidimensional scaling (NMDS) plot of samples from the communities. Distance was calculated using Bray-Curtis dissimilarly for ordination and Adonis was used for testing statistical significance. Sample depths (at 2 cm intervals) for WGZ are between 0 and 10 cm [20], and for SLW, the median depth is 1, 5, 19, and 35 cm [8]. The SLM unit sample depths are shown in Fig. 2. Ellipses are 50% confidence intervals. b–e Maximum likelihood analysis based on aligned partial 16S rRNA gene sequences from related ASVs with >94% identity to each other. Colored lines highlight lineages found exclusively in marine (red; WGZ samples) versus subglacial (blue; SLW and SLM samples) sediments. The colored squares for each taxon represent the site of origin for the ASV: SLM (black), SLW (blue), and WGZ (red). Phylogenetic analysis was carried out with taxa affiliated with the class b) Anaerolineaceae and genera c) Candidatus Nitrotoga, d Pseudohongiella, and (e) Candidatus Nitrosopumilus. The scale bar represents 0.02 fixed substitutions per nucleotide position. Bootstrap values are shown at the nodes and represent 100 replications.