Arsenic metabolism in high altitude modern stromatolites revealed by metagenomic analysis

Abstract

Modern stromatolites thrive only in selected locations in the world. Socompa Lake, located in the Andean plateau at 3570 masl, is one of the numerous extreme Andean microbial ecosystems described over recent years. Extreme environmental conditions include hypersalinity, high UV incidence, and high arsenic content, among others. After Socompa's stromatolite microbial communities were analysed by metagenomic DNA sequencing, taxonomic classification showed dominance of Proteobacteria, Bacteroidetes and Firmicutes, and a remarkably high number of unclassified sequences. A functional analysis indicated that carbon fixation might occur not only by the Calvin-Benson cycle, but also through alternative pathways such as the reverse TCA cycle, and the reductive acetyl-CoA pathway. Deltaproteobacteria were involved both in sulfate reduction and nitrogen fixation. Significant differences were found when comparing the Socompa stromatolite metagenome to the Shark Bay (Australia) smooth mat metagenome: namely, those involving stress related processes, particularly, arsenic resistance. An in-depth analysis revealed a surprisingly diverse metabolism comprising all known types of As resistance and energy generating pathways. While the ars operon was the main mechanism, an important abundance of arsM genes was observed in selected phyla. The data resulting from this work will prove a cornerstone for further studies on this rare microbial community.

Figure 1

(A) Socompa stromatolites. (B) Vertical section. (C) Taxonomic diversity analysed by different methods: MG-RAST results based on assembled data compared to M5NR database, and MG-RAST assignment of RNA reads, updated Qiime analysis on published 16S rRNA data.

Figure 2

Comparative analysis of several microbial communities at taxonomic phylum level (A) and functional level (B). Abundance data normalized by MG-RAST is represented with a color gradient from white (less abundant) to red (most abundant). Comparisons were performed with R vegan package and heatmaps were generated with R gplot package.

Figure 3

Comparison of Shark Bay smooth mats (blue) and Socompa stromatolites (yellow) by SEED subsystems at level 1 (A) and Virulence, defense and disease level 3 (B) using STAMP. Features displayed show significant differences.

Figure 4

Diversity of arsenic metabolism related proteins, based on NCBI CDD Search, MEGAN, and MG-RAST annotation. Efflux pumps include both Acr3 and ArsB.

Figure 5

Maximum likelihood tree for Acr3 proteins, including 62 sequences from Socompa and 21 sequences from NCBI nr database. Metagenomic sequences with the same affiliation in the same branch are collapsed. The three groups of the Acr3 family are marked, and all of them include Socompa sequences.

Figure 6

ArrA subfamily maximum likelihood tree. The arsenite oxidases clade ArxA is shown in light grey and the arsenate reductases in dark grey. The tree is rooted with scaffold_4302_4, which belongs to another DMSO reductases subfamily. Socompa sequences are in bold. Nodes with bootstraps above 0.7 are marked with a filled circle.

Figure 7

Arsenic genes abundances and affiliation. Counts were log-transformed (log₂(x + 1)) and drawn as a heatmap.

Figure 8

Summary of the arsenic metabolism in Socompa stromatolites. Abundances of the genes are shown in black boxes. arxA sequences are counted among the arrA, although they perform As[III] oxidation.