A glimpse of the paleome in endolithic microbial communities

Abstract

Background: The terrestrial subsurface is home to a significant proportion of the Earth's microbial biomass. Our understanding about terrestrial subsurface microbiomes is almost exclusively derived from groundwater and porous sediments mainly by using 16S rRNA gene surveys. To obtain more insights about biomass of consolidated rocks and the metabolic status of endolithic microbiomes, we investigated interbedded limestone and mudstone from the vadose zone, fractured aquifers, and deep aquitards.

Results: By adapting methods from microbial archaeology and paleogenomics, we could recover sufficient DNA for downstream metagenomic analysis from seven rock specimens independent of porosity, lithology, and depth. Based on the extracted DNA, we estimated between 2.81 and 4.25 × 10⁵ cells × g^-1 rock. Analyzing DNA damage patterns revealed paleome signatures (genetic records of past microbial communities) for three rock specimens, all obtained from the vadose zone. DNA obtained from deep aquitards isolated from surface input was not affected by DNA decay indicating that water saturation and not flow is controlling subsurface microbial survival. Decoding the taxonomy and functional potential of paleome communities revealed increased abundances for sequences affiliated with chemolithoautotrophs and taxa such as Cand. Rokubacteria. We also found a broader metabolic potential in terms of aromatic hydrocarbon breakdown, suggesting a preferred utilization of sedimentary organic matter in the past.

Conclusions: Our study suggests that limestones function as archives for genetic records of past microbial communities including those sensitive to environmental stress at modern times, due to their specific conditions facilitating long-term DNA preservation. Video Abstract.

Keywords: Chemolithotrophy; DNA damage; Endolithic; Metagenomics; Subsurface.

Fig. 1

Pore space characteristics of samples H13-17 (A–D) and H22-8 (E–G) determined by µCT analysis. Moldic pores (up to large mesopores) dominate the packstone. Scale: 0.5 mm. Plug diameter 13 mm (A). Vertical section shows considerable porosity. The dashed line marks the position of C (B). Horizontal section (C). Reconstructed pore space. Colors mark parts of the pore system that are each connected by throats > 26 µm (D). The plug comprises delithified siliceous marlstone (lower part) and delithified calcareous mudstone (upper part). Scale: 0.5 mm. Plug diameter 13 mm (E). Vertical section shows thin fractures (micropores). The dashed line marks the position of G (F). Horizontal section showing fine fractures and rare micro- to small mesopores. The matrix exhibits no pores connected by throats > 26 µm (G)

Fig. 2

Overview of data (pre-)processing. Samples were quantified by fluorometry and quantitative PCR after DNA extraction (upper panel) and library preparation (lower panel) (A). Sequence length histograms were generated after quality control and trimming based on subsampled (n = 1million read pairs) data sets. The gray shading highlights three data sets for which the read length distribution was skewed to the left. Based on taxonomic profiling (see main text), we summarized these three data sets in two groups: (1) and (2) (B). The proportion of quality-controlled and trimmed sequences that could be assigned taxonomically was determined based on database queries with diamond against NCBI nr

Fig. 3

Taxonomic profiling of rock endolithic microbial communities. Principle component analyses were carried out based on phylum-level (A) and family-level (B) taxonomic profiles, prior to (left) and after (right) decontamination. The color coding indicates the sample type. Phylum-level taxonomic profiles were visualized as heatmap (C). (1) and (2) indicate two groups of samples (see main text for details). White and black boxes indicate if the corresponding profile is based on decontaminated data. Ex. and Lib. BLANKS refer to extraction and library blanks, respectively

Fig. 4

DNA damage pattern analysis. Quality-controlled sequence reads were mapped onto assembled contigs (> 1 kbp). The damage pattern analysis was carried out with mapdamage (v.2.2.1) [69]. The plots show the substitution frequency (5pCtoT [5' cytosine to thymine substitutions], 3pGtoA [3' guanine to adenine substitutions]) versus the relative position (from the 5p and 3p end). n, number of contigs > 1 kbp considered for the analysis; cov, mean coverage of the contigs

Fig. 5

Functional profiling of rock endolithic microbial communities. Profiles were generated based on output from humann regrouped into KEGG orthologies. KEGG orthologies were summarized based on pathways and selected functions as described in the methods. Unmapped indicates the proportion of sequences that did not yield any hits against the pre-compiled UniRef databases shipped with humann. logCoPM, log copies per million