Edited by: D’Arcy Renee Meyer-Dombard, University of Illinois at Chicago, United States
Reviewed by: Virginia Rich, The Ohio State University, United States; Malak M. Tfaily, University of Arizona, United States
This article was submitted to Extreme Microbiology, a section of the journal Frontiers in Microbiology
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Permafrost is an extreme habitat yet it hosts microbial populations that remain active over millennia. Using permafrost collected from a Pleistocene chronosequence (19 to 33 ka), we hypothesized that the functional genetic potential of microbial communities in permafrost would reflect microbial strategies to metabolize permafrost soluble organic matter (OM)
Permafrost underlies one-quarter of the Northern Hemisphere and contains almost half of the Earth’s soil carbon (C) (
Anaerobic decomposition within intact permafrost occurs through a series of cascading hydrolytic and fermentative steps resulting in production of CO2 or CH4 (
Liquid water can exist in permafrost in nanometer-thin brine channels surrounding soil particles. Multiple factors may affect the amount of liquid water in permafrost, including temperature, pore size distribution, and solute concentration (
Permafrost soils host a diverse microbial community, though diversity is typically reduced relative to active layer soils, pointing to its unique environmental stressors (e.g., reduction in energy availability and salinity) (
Here we examine the relationships between the molecular composition of DOM and the functional potential of microbial communities across a Pleistocene-aged (19 to 33 ka) permafrost chronosequence. We sought to determine if changes in DOM chemistry were consistent with microbial community changes and biochemical adaptations inferred from metagenomic data. With these data, we identified cryogenic and energetic constraints on microbial metabolism of OM. Results from this study will help develop our understanding of biological activity in intact permafrost, as well as dominant microbial and ecosystem responses to permafrost thaw.
Permafrost cores were collected in 2012 from the Fox Permafrost Tunnel, operated by the Cold Regions Research and Engineering Laboratory (CRREL) in Fairbanks, Alaska (64.951°N – 147.621°W), and are the same cores used by
A portion of the metagenomic data presented here was previously published in a paper focusing on microbial survival mechanisms (
To identify and annotate reads from genes potentially involved in polysaccharide processing (CAZymes) we performed a two-step process where (1) potential carbohydrate-utilization reads were identified through a blastx-like comparison to dbCAN using permissive parameters and (2) carbohydrate-active enzyme reads were further identified and annotated through comparison to a custom database of HMM profiles using an
Reads were also annotated through comparison to the Kyoto Encyclopedia of Genes and Genomes (KEGG) database (
In the laboratory, the same cores used for metagenomic analyses were disaggregated and sub-sampled for chemical analyses. One portion was thawed and used for pH, δ13C-DOC, and gravimetric ice content (
Extracted permafrost filtrates were analyzed for anions by ion chromatography using a Metrohm 881 Compact IC pro with detection by suppressed conductivity. Anion determinations were performed using a Metrosep A Supp 7/250 column with 3.6 mM Na2CO3 eluent (0.8 mL/min) at 45°C. A complementary set of determinations were performed targeting organic acid anions using a Metrosep Organic Acids 250/7.8 column with 0.5 mM H2SO4 eluent (0.5 mL/min) at 30°C.
Soluble NH4+, NO3–, and NO2– measurements were carried out using batch-automated spectrophotometry (Aquakem 250, Thermo Fisher Scientific, United States). Ammonium was determined using the Phenate method [United States Environmental Protection Agency (U.S. EPA) method 350.1]. Nitrite was determined by diazotization with sulfanilamide and nitrate was catalytically reduced to nitrite using soluble nitrate reductase in the presence of reduced nicotinamide dinucleotide (NADH; EPA Method 353.1). All samples were analyzed in duplicate. Final DOC, DIN, and SCFA data were corrected for dilution and gravimetric water content.
Water-soluble OM was extracted from the same permafrost filtrates using the solid phase method described by
We calculated the nominal oxidation state of carbon (NOSC) according to (
where
All statistical analyses were performed in R version 3.5.0 (
Chemical characterization of permafrost filtrates from the Fox Tunnel Permafrost Chronosequence, different letters indicate statistical differences across ages, within a category (
pH* | 7.730.32 |
7.750.14 |
7.340.09 |
Ice content** | 39.7210.25 |
121.0514.09 |
51.755.24 |
Chloride | 649.32169.22 |
95.797.82 |
237.0256.24 |
Sulfate | 97.1270.48 |
56.108.65 |
165.4874.65 |
NH4 + | 528.4954.25 |
556.5040.96 |
1815.74136.01 |
NO3 | 0.630.09 |
1.371.17 |
0.480.08 |
NO2 | 0.250.14 |
0.610.10 |
0.260.02 |
δ13C-DOC | −26.20.3 |
−26.90.2 |
−27.80.1 |
DOC*** | 74.7020.10 |
124.5025.87 |
181.1337.15 |
Acetate | 0.910.10 |
7.680.82 |
40.456.89 |
Butyrate | 0.000.00 |
3.040.57 |
30.156.49 |
Formate | 0.380.04 |
0.190.01 |
0.340.04 |
Propionate | 6.001.59 |
0.740.06 |
4.780.85 |
Isovalerate | 1.270.36 |
1.100.32 |
7.311.56 |
Glutarate | 0.030.01 |
0.010.00 |
0.030.01 |
Malate | 0.160.03 |
0.180.04 |
0.260.02 |
%C*** | 3.141.94 |
3.110.26 |
3.200.39 |
%N*** | 0.270.16 |
0.210.06 |
0.270.03 |
C/N*** | 11.811.14 |
15.253.30 |
12.100.98 |
Permafrost DOC concentrations increased across the three sampled ages with %C, %N, and C:N remained remarkably consistent across the chronosequence (
Molecular formulae assigned to peaks detected in DOM samples from across the chronosequence totaled 10,396 (
Molecular composition of DOM shifts across the permafrost chronosequence.
Multiple response permutation procedure (MRPP) analysis of the standardized intensities of assigned formulae indicated distinct groupings of the molecular-level composition of DOM by age, with samples from within an age class more similar to each other than to samples from other ages (
Thermodynamic favorability of DOM decreases with older samples. Nominal oxidation state of carbon (NOSC) in permafrost from across the chronosequence as calculated based on molecular formulae identified by FT-ICR MS in DOM. Different letters indicate statistically significant differences (
Shotgun metagenomic sequencing resulted in an average of 22 Gb of sequence data per sample (
Pathways related to acetate, butanoate, and propanoate metabolism. SCFA-related fermentation pathways are shown in orange, the Wood-Ljungdahl pathway is shown in light blue, and the methanogenesis pathway is shown in dark blue. Pathway components with multiple colors indicate participation in multiple pathways. Bar charts show the relative abundance of KEGG genes in 19, 27, and 33 ka samples, respectively, from left to right on each chart.
Next, we explored the microbial communities’ potential for turnover of OC through identification of genes encoding CAZymes with a focus on sequences for glycoside hydrolases (GH), glycosyltransferases (GT), carbohydrate esterases (CE), polysaccharide lyases (PL), and associated carbohydrate-binding modules (CBM). When normalized to a single copy marker gene (the large subunit ribosomal protein L14, K02874) the abundance of CAZymes significantly decreased with increasing age (
Changes in abundance and diversity of CAZymes with soil age.
The most abundant GH families were GH65 and alpha-amylases from GH families 13 and 15, and the most abundant GT families were 1B, 51, 2_6, 20, and 4 (
Our measurements of the relative abundance of CAZyme domains across the chronosequence identified several trends among gene families affected (
We examined the relationship between the molecular composition of DOM (i.e., FT-ICR MS data as relative abundance of molecular formulae), the functional potential of the associated microorganisms (i.e., KEGG ortholog data, KOs), and the potential for polysaccharide processing (CAZyme data) by performing Mantel correlation tests. Both the KO data (
Ordination of functional gene data with subsequent vector fitting by measured permafrost characteristics indicated that microbial function (
Polysaccharide processing potential and DOC structure differs according to age of samples. Non-metric multidimensional scaling ordination based on Bray-Curtis dissimilarities of
Finally, to investigate the relationship between DOM composition and the potential for carbohydrate processing we ordinated the FT-ICR MS data with vector fitting of the identified CAZymes involved in polysaccharide degradation (
Permafrost contains vast stores of OM and is at risk of increased thaw. Yet the variation in OM chemistry is not well constrained, in part because of our limited understanding of the initial OM entrained in permafrost, but also because of poor recognition that microbial communities in permafrost are themselves altering its chemical nature. Recently, Mackelprang et al. demonstrated that microbial populations adapt and survive in permafrost across the Fox Tunnel chronosequence (
Conceptual diagram of the factors influencing microbial processing of C in permafrost. The dashed blue line indicates the transition from active layer soil to permafrost. The decreasing size of green rectangles represents increasing degree of OM decomposition prior to entrainment in permafrost.
The solid and liquid phases of C in permafrost are largely controlled by composition of soil C present, paleoclimate, and vegetation at the time of permafrost formation. At each age along the chronosequence, paleoclimate and paleovegetation likely differed at the time of formation (
Permafrost C is comprised of a complex organic mixture including plant debris and structural polymers (
To understand potential interactions between DOM and the microbial metagenome we screened for genes encoding SCFA analysis (
Cellulose and hemicellulose are the primary plant polymers entrained into terrestrial anaerobic environments, e.g., permafrost (
Several lines of evidence indicate microorganisms begin scavenging proteinaceous compounds from dead microbial biomass in lieu of macromolecular OM over increasing periods of time since permafrost formation.
As microbes scavenge for available C, products of anaerobic OM hydrolysis can serve as substrates for microbial fermentation: a common microbial process found in Alaskan permafrost metagenomes (
The increasing concentrations of electron donors such as DOC and NH3 with increasing permafrost age are another indication of anaerobic decomposition via fermentation (
Our research indicates both the initial conditions of permafrost soils and microbial activity over time can affect the quantity and chemistry of DOM in permafrost soils. Abiotic factors such as potential mineral dissolution, composition of OM prior to freezing, and other physical processes also play a role in the shaping of permafrost DOM pools (
Knowledge of the factors controlling microbial survival and the molecular composition of OM in Pleistocene-aged permafrost is scarce. Through this work, we show further evidence for microbial survival across millennia and that those microbial populations may be actively modifying permafrost DOM
Finally, further exploration of microbial mechanisms of survival in extreme environments could inform the astrobiological search for life on other planets, as permafrost on Earth may be a representative model of extraterrestrial icy habitats, such as Mars, Enceladus, or Europa (
The datasets generated for this study can be found in the NCBI short read archive accession number
M-CL statistically analyzed all data and wrote the original draft manuscript. RM performed the metagenome analyses. RB and AS performed the CAZyme analyses. RB contributed to the statistical analysis and interpretation of CAZyme data. DP, PZ, and RS performed the FT-ICR MS analyses and data processing. CC and JM performed the permafrost filtrate analyses. MW and JM assisted with field work. MW and RM provided the funding. TD provided access to the field site and oversaw sample coring. All authors reviewed and contributed to the final version of the manuscript.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
We thank Dr. Kimberly Wickland for providing an internal USGS review. We also thank Monica Haw and Sabrina Sevilgen for assistance in sample processing.
The Supplementary Material for this article can be found online at:
Heatmap of CAZymes significantly representing
Non-metric multidimensional scaling (NMDS) plot representing the Bray Curtis dissimilarity of all identified KO’s in the permafrost metagenome from across all sites with subsequent vectors of environmental and geochemical data onto the ordination (
Characteristics of compound classes used to categorize FT-ICR MS molecular formulae. Criteria based on
List of protein families (Pfam ID), with corresponding domain names, of potentially involved in polysaccharide processing (CAZymes) identified in metagenome data. The indicated
FT-ICR MS results representing the distribution of DOC leachate by compound class from the Fox Tunnel Permafrost Chronosequence. Data other than total number of formulae and molecular weight (MW) are the average relative abundance of the three replicates with standard deviation from the mean. Compound classes defined in
Multiple-response permutation procedure (MRPP) tests for differentiation of molecular composition of DOC, KO gene counts, and CAZyme Pfam counts.
Relative abundances of KEGG genes in metabolic pathways described in
Permutational multivariate analysis of variance (PERMANOVA) table exploring factors which may explain variation in either the molecular composition of DOC (FT-ICR MS) or functional potential (KEGG).
CAZyme domains which are significant vectors explaining the FT-ICR MS data.