Edited by Edward F. DeLong, University of Hawaii at Manoa, Honolulu, HI, and approved April 16, 2018 (received for review October 30, 2017)
Marine sediment holds the largest organic carbon pool on earth, where microbial transformation of carbon is considered a key process of carbon cycling. Bathyarchaeota are among the most abundant and active groups of microorganisms in marine sediment. It has been suggested that Bathyarchaeota may play a globally important role in the carbon cycling in the marine environment through fermentation of complex organic substances, acetogenesis, and methanogenesis based on metagenome analysis. Here we provide several lines of converging evidence suggesting the bathyarchaeotal group Bathy-8 is able to grow with lignin as an energy source and bicarbonate as a carbon source. Consequently, members of the Bathyarchaeota are probably important, previously unrecognized degraders of lignin.
Members of the archaeal phylum Bathyarchaeota are among the most abundant microorganisms on Earth. Although versatile metabolic capabilities such as acetogenesis, methanogenesis, and fermentation have been suggested for bathyarchaeotal members, no direct confirmation of these metabolic functions has been achieved through growth of Bathyarchaeota in the laboratory. Here we demonstrate, on the basis of gene-copy numbers and probing of archaeal lipids, the growth of Bathyarchaeota subgroup Bathy-8 in enrichments of estuarine sediments with the biopolymer lignin. Other organic substrates (casein, oleic acid, cellulose, and phenol) did not significantly stimulate growth of Bathyarchaeota. Meanwhile, putative bathyarchaeotal tetraether lipids incorporated 13C from 13C-bicarbonate only when added in concert with lignin. Our results are consistent with organoautotrophic growth of a bathyarchaeotal group with lignin as an energy source and bicarbonate as a carbon source and shed light into the cycling of one of Earth’s most abundant biopolymers in anoxic marine sediment.
The members of Bathyarchaeota (formerly referred to as the “Miscellaneous Crenarchaeotal Group”) (1, 2) are estimated to be among the most abundant microorganisms on the planet (3) and are particularly common in marine sediments (1, 4⇓⇓–7). The phylum Bathyarchaeota contains more than 19 subgroups/lineages with low intragroup similarities (3, 5, 8, 9), and its members have been suggested to play a globally important role in the breakdown of organic matter (10) through fermentation (11), acetogenesis (3), and methanogenesis (12). However, due to the lack of pure culture isolates and difficulty of obtaining enrichments of Bathyarchaeota (13), the metabolic properties and capabilities of these abundant and widespread Archaea have so far been inferred mostly from single-cell genomic and metagenomic analyses (2, 3, 11, 12, 14).
The first metabolic insights into Bathyarchaeota came from organic-rich sub-seafloor sediments of the Peru Margin, where these Archaea were inferred to be assimilating sedimentary organic carbon based on the δ13C-isotopic compositions of intact polar lipids (10). The organoheterotrophic physiology of Bathyarchaeota was further supported through analyses of bathyarchaeotal genomes by single-cell genome and metagenome analyses, and genes encoding enzymes for the degradation, transport, and utilization of detrital proteins, aromatic compounds, and plant-derived carbohydrates were identified (2, 11, 14). Peptidase activity in the sediments was measured, and the gene of an extracellular peptidase from Bathyarchaeota was expressed in vitro and characterized, supporting the inferred capacity of Bathyarchaeota to degrade proteins (11, 15). Meanwhile, stable-isotope probing experiments indicated that members of two subgroups assimilated several organic substrates, including acetate, glycine, urea, lipids, and complex mixtures of organic growth substrates, while showing no significant incorporation of carbon from proteins (16). On the other hand, acetogenesis from H2/CO2 has been proposed for some lineages of Bathyarchaeota (3). The reductive acetyl-CoA pathway for carbon fixation has been identified in most of the obtained bathyarchaeotal genomes (3, 12, 14). Among these, members of the bathyarchaeotal subgroups Bathy-3 and Bathy-8 have been suggested to be capable of methanogenesis (12). All the above studies indicate great versatility in the metabolic potentials of Bathyarchaeota, with some lineages possibly being capable of utilizing both organic and inorganic carbon compounds for biosynthesis and energy production. However, no direct proof of carbon or energy sources has so far been obtained for Bathyarchaeota based on laboratory experiments.
Here we report our efforts to enrich Bathyarchaeota from marine sediment in the laboratory by setting up a series of enrichments with diverse organic compound classes, including lipids (oleic acid), proteins (casein), aromatic monomers (phenol), aromatic polymers (lignin), and structural carbohydrates (cellulose). The addition of lignin stimulated the growth of Bathyarchaeota affiliated with the Bathy-8 subgroup. During growth on lignin, incorporation of inorganic carbon (IC) into archaeal lipids was demonstrated. This study presents an in vitro enrichment of Bathyarchaeota and suggests the organoautotrophic metabolism of members of this phylum, which likely plays an important role in the carbon cycle in marine sediments.
Numerous enrichment cultures from Dayangshan estuarine sediments of the northern East China Sea (Fig. S1) were set up by adding diverse organic compound classes, i.e., the long-chain fatty acid oleic acid, the protein casein, the aromatic monomer phenol, the phenolic polymer lignin, and the polymeric carbohydrate cellulose (Materials and Methods). Sediment slurries without the addition of organic compounds were used as controls. Changes in the abundances of Bacteria, Archaea, and Bathyarchaeota were then monitored by qPCR using universal bacterial and archaeal 16S rRNA gene primers and Bathyarchaeota-specific 16S rRNA gene primers. The organic compounds were initially added at concentrations of 50 mg/L and were increased to 500 mg/L after 3.5- and 6-mo incubation.
Treatment responses to substrate additions and controls are shown in Fig. 1. In the control samples, the abundance of Bacteria, Archaea, and Bathyarchaeota showed only small fluctuations over time. Casein initiated the strongest growth stimulation in total Bacteria and Archaea, with an increase in bacterial and archaeal 16S rRNA gene copies of about three and nine times at 6-mo incubation (t6), respectively; bathyarchaeotal gene copies increased only slightly in response to casein. Cellulose stimulated growth of Bacteria after 3.5 mo incubation (t3.5) but had no influence on archaeal or bathyarchaeotal gene copies. Phenol and oleic acid had no obvious influence on the gene copies of Bacteria, Archaea, or Bathyarchaeota. Lignin showed little influence on the growth of Bacteria but significantly stimulated the growth of Archaea, particularly Bathyarchaeota. Total archaeal gene-copy numbers increased by about two and three times at t6 and 11-mo incubation (t11), while those of Bathyarchaeota climbed to more than 10 times at t11 compared with the original sample. Thus, strong growth of Bathyarchaeota was achieved only in response to lignin addition. Lignin degradation was monitored by the decrease in the concentrations of total dissolved phenolic compounds after incubation. There were ∼102 mg/L phenolic compounds in the culture when 500 mg/L lignin was added, which decreased to ∼25 mg/L after 2.5-mo incubation.
The change in abundance of Bacteria, Archaea, and Bathyarchaeota in sediment slurries after incubation with different organic compounds for 3.5, 6, and 11 mo. The abundance (gene copies per gram wet weight of sediment; gene copies⋅gww−1) is quantified based on the respective 16s rRNA genes. The error bars are obtained from replicate incubations.
Since the enrichment of Bathyarchaeota was the aim of this study, and only lignin showed significant stimulation of bathyarchaeotal growth, further detailed experiments were undertaken with the original sample, lignin enrichments, and controls. Archaeal 16S rRNA gene analyses show that Thaumarchaeota (40%) and Bathyarchaeota (33%) (Fig. 2A) were the dominant archaeal groups in the original sample followed by Thermoplasmata (8%) and Parvarchaea (5%). After amendment with lignin, the relative abundance of Bathyarchaeota within the archaeal community increased to over 65% at t11 (Fig. 2A), compared with a strong decrease in control samples (8–23% of total Archaea) (Fig. 2A). The percentage of methanogenic Methanococci reads increased in both lignin cultures and no-substrate controls, whereas the relative abundance of Thaumarchaeota decreased across lignin treatments and no-substrate controls, showing a stronger percent decrease in the former.
Comparison of archaeal communities at the phylum level (A) and of bathyarchaeotal communities at the subgroup level (B) in response to lignin addition over time. t0, t3.5, t6, and t11 indicate samples that were taken after 0, 3.5, 6, and 11 mo, respectively. Only one sample from t0 and one sample from t3.5 were analyzed, compared with four samples from both t6 and t11. Details of the phylogenetic identifications are in Materials and Methods. AAG, Ancient Archaeal Group; DSEG, Deep Sea Euryarcheotic Group; MBGA, Marine Benthic Group A; MBGB, Marine Benthic Group B; MHVG, Marine Hydrothermal Vent Group; Other, unclassified archaea.
The bathyarchaeotal community in the original sample was dominated by the subgroups Bathy-8 (49%), Bathy-6 (24%), and Bathy-12 (10%) (Fig. 2B) based on a phylogenetic tree of bathyarchaeotal 16S rRNA genes (Fig. S2). Incubation with lignin created a strong selection pressure, resulting in Bathy-8 accounting for 80–90% of the bathyarchaeotal sequences at t6 and t11 (Fig. 2B). Within the Bathy-8 subgroup, a single operational taxonomic unit (OTU), OTU3326, became dominant, accounting for 63–73% of all bathyarchaeotal 16S rRNA gene reads (Table S1).
13C-labeled sodium bicarbonate (13C-IC) was provided to enrichments with and without lignin to test the hypothesis that members of the Bathyarchaeota assimilate IC via the reductive acetyl-CoA pathway. In the original sediment, the δ13C values of the archaeal polar lipid derivatives phytane, biphytane (BP)-0, BP-2, and BP-3 were −35.1‰, −26.5‰, −25.8‰, and −21.9‰, respectively (Fig. 3 and Table S2). No matter whether lignin was added or not, with or without 13C-IC, the δ13C value of BP-3 did not change significantly at t6 and t11. However, the δ13C values of phytane, BP-0, and BP-2 showed a different pattern.
The carbon isotopic composition of archaeal polar lipid derivatives (i.e., phytane, BP-0, BP-2, and BP-3) in the lignin treatment (A) and in the control samples without lignin (B) with/without 13C-IC at t6 and t11.
In the 13C-IC–amended control samples, the δ13C value of phytane increased by 698‰ at t6 and by 1,100‰ at t11, whereas the δ13C values of BP-0, BP-2, and BP-3 did not change (Fig. 3 and Table S2). When grown on lignin with 13C-IC, the δ13C values of phytane, BP-0, and BP-2 increased by 885‰, 137‰, and 18‰, respectively, at t6 (Fig. 3 and Table S2) and by 1,260‰, 234‰, and 81‰, respectively at t11.
Acetate concentration and isotopic data are shown in Table 1. Acetate concentrations show significant fluctuations between treatments and replicates at the two time points sampled after 6 mo and 11 mo. After 6 mo, acetate concentrations were in a similar range in lignin treatments and controls. By comparison, after 11 mo, acetate concentrations had increased in all lignin treatments (P < 0.05, Wilcoxon rank sum test), whereas controls showed no clear increase relative to concentrations after 6 mo (P > 0.05). Production of 13C-acetate, indicative of acetogenesis from 13C-IC, was evident in all 13C-IC–amended controls and lignin treatments after 6 mo, as indicated by strong increases in δ13C-acetate. However, after 11 mo, high δ13C-acetate values were detected only in lignin treatments, consistent with sustained acetogenesis in the presence of lignin.
Carbon isotope composition and concentration of acetate in water medium after incubation with/without lignin and13C-bicarbonate
Marine sediments are one of Earth’s largest organic carbon sinks. Microbial transformation of carbon is considered a key process influencing the carbon flow in sediment and ultimately atmospheric oxygen and carbon dioxide concentrations (17). Nevertheless, the controls on microbial organic carbon cycling are not well understood, including the biogeochemical role of Archaea, which are often abundant in marine sediments (18). In this study, we demonstrate that lignin addition significantly stimulates the growth of Bathyarchaeota (Fig. 1), which are a dominant group of Archaea in marine sediments (11). The relative abundance of Bathyarchaeota reached up to 65% of the total archaeal community after 11 mo of incubation, with the subgroup Bathy-8 accounting for most of this increase (Fig. 2). Our work suggests that members of the Bathyarchaeota are potentially important, previously unrecognized degraders of lignin in marine sediments.
Lignin is a class of complex cross-linked phenolic polymers and, after cellulose, constitutes the second-most abundant biopolymer on Earth. Lignin is particularly important in the formation of cell walls in vascular plants, comprising up to 25% of plant biomass (19). Structural polymers, such as cellulose, hemicellulose, and lignin, constitute the bulk of terrestrial organic matter. Terrestrial organic matter is estimated to contribute one-third of the total buried organic carbon in marine sediments (20). This contribution decreases from nearshore to offshore but remains high, up to 15%, even in basin sediments (21). In terms of global organic carbon burial, nearshore sediments in estuaries and shallow shelf environments play a very important role, accounting for ∼45% of total organic carbon burial in marine sediments (17 and 22). Consequently, anaerobic lignin degradation in marine sediments, in particular in nearshore environments, is probably a globally important microbial process. The anaerobic biodegradation of lignin in sediments has been demonstrated previously (23), but the responsible microbes have remained elusive.
Metagenomic sequence analyses have suggested that Bathyarchaeota are capable of utilizing a variety of organic compounds including proteins (11), aromatic compounds (2), and carbohydrates (3, 14). However, bathyarchaeotal 16S rRNA gene abundances increased 10-fold in response to lignin addition in this study and did not grow in response to proteins, cellulose, phenol, or oleic acid. This indicates that, at least at the study site, members of this phylum play an important role in the degradation of lignin.
In the enrichment with lignin, the doubling (or generation) time of Bathyarchaeota was estimated to be about 2–3 mo, which is similar to that of other marine-sediment archaea enriched in the laboratory (24⇓–26). Considering the typically lower carbon and energy availability in natural sediments, the in situ generation times of bathyarchaeotal cells in marine sediments are likely even longer than in our enrichment cultures. These long generation times may contribute to the difficulty of enriching Bathyarchaeota in the laboratory and explain why no pure isolates exist.
Within the Bathyarchaeota, the subgroup Bathy-8 became predominant during growth on lignin, comprising ∼90% of the bathyarchaeotal community at the end of experiments (Fig. 2B). Within the Bathy-8, OTU3326 became dominant (Table S1), suggesting that this member of Bathy-8 was the main lignin degrader in the community. Bathy-8 members are widely distributed in various marine and terrestrial habitats, including seeps, hydrothermal vents, shallow marine sediments, sediments of hypersaline, saline, and freshwater lakes, salt marsh sediments, hot springs, and Arctic peat soils (5). BLAST analyses furthermore show that highly similar 16S rRNA gene sequences (≥97% identical with OTU3326) occur in diverse environments, including deep-sea fan sediments (KX952596), estuarine sediments (JQ245924), submarine springs (JF971130), and terrestrial habitats.
The metagenome was obtained from the lignin enrichment and was analyzed (for details see SI Materials and Methods and SI Text). Two genomic bins belonging to Bathy-8 (Bin-L-1 and Bin-L-2) were assembled from the lignin enrichment (Fig. S3 and Table S3). A complete Wood–Ljungdahl pathway (WL; also called the “reductive acetyl-CoA pathway”) was found in both genomes, indicating the ability of these Bathy-8 members to fix IC (Tables S4 and S5). The WL differs from other carbon-fixation pathways in being linear rather than cyclic, involving fewer reaction steps and a smaller set of enzymes, and being used for both energy conservation and biosynthesis by the same organisms; these traits likely confer energetic advantages under low-energy conditions because less metabolic energy is required for the synthesis of enzymes and energetically costly intermediates (27⇓–29). High energy efficiency may explain why the WL is widespread among anaerobic microorganisms performing low-energy catabolic reactions, such as methanogenesis and acetogenesis, and in dominant subsurface sedimentary microorganisms. Members of the Bathy-8 group were recently suggested to be capable of methanogenesis because a metagenomic bin contained the entire methyl–coenzyme M reductase (MCR) gene cluster (12). No MCR genes were found in Bin-L-1 and Bin-L-2 (Tables S4 and S5). No known genes for lignin degradation were found in Bin-L-1, while genes of catalase-peroxidase and 4-oxalocrotonate tautomerase for putative lignin and aromatic compound degradation, respectively, were found in Bin-L-2 (Tables S4 and S5). No canonical genes for respiration such as dissimilatory sulfate, nitrate, or iron reduction were found in the genomes (3, 14).
IC was incorporated into phytane as well as biphytanes with 0–2 cycloalkyl rings (Fig. 3). The pattern of label incorporation into these compounds suggests that at least two different functional archaeal groups were assimilating IC for lipid biosynthesis in our experiments. The lignin-unrelated incorporation of 13C-IC into phytane is presumably related to the methanogenic class Methanococci, which were the second-most abundant group of Archaea throughout controls and lignin enrichments and which grow using H2 as an energy source and IC as a carbon source (30). Since the incorporation of 13C label into biphytanes was observed only in lignin treatments, which exhibited strong increases only in bathyarchaeotal populations, we infer that the dominant Bathyarchaeota in lignin treatments were responsible for this incorporation. In the White Oak River estuary, which receives high input of terrestrial organic matter, Bathyarchaeota were suggested to be the producers of isoprenoidal tetraether lipids bearing a carbon isotopic signature of autotrophy (31). Bathyarchaeota are therefore a plausible source of 13C-labeled BP-0 and BP-2 in our study. The fact that the majority of recovered bathyarchaeotal genomes from various subgroups contain genes of the WL (3, 14) is in line with the observed 13C-IC assimilation in lignin treatments. Although the enrichment of Bathyarchaeota in lignin amendments strongly suggests an involvement of Bathyarchaeota in lignin degradation, we cannot entirely rule out the possibility that the IC-derived 13C-labeled incorporation into archaeal lipids is due to the selective growth of other archaeal community members whose growth and lipid biosynthesis were also stimulated selectively by the presence of amended lignin. However, no archaeal group other than the Bathy-8 subgroup shows clear growth based on 16S genes and lipid biosynthesis in response to enrichment with lignin (Figs. 2 and 3 and Table S2).
The enrichment of Bathyarchaeota on lignin suggests that members of this phylum could play an important role in the degradation of lignin in anoxic marine sediments. We hypothesize that bathyarchaeotal community members affiliated with Bathy-8 metabolize methoxy-groups of lignin and combine the resulting methyl groups with CO2 to acetyl-CoA via the carbon monoxide dehydrogenase/acetyl-CoA synthase complex (Codh/Acs). Acetyl CoA is then used as a key intermediate for biosynthesis or is converted to acetate via cleavage of the cofactor to produce energy (3, 28). This mode of lignin metabolism is equivalent to that described for acetogenic bacteria (27), of which 88% of tested cultivars can grow by combining methyl groups from lignin phenols with CO2 (32). Acetate production was indeed measured in the cultures (Table 1). Although both controls and lignin treatments show clear signs of acetogenic incorporation of CO2 into acetate, suggesting that acetogenesis is also an important process in unamended samples and perhaps in native sediments, both the higher acetate concentrations and higher 13C-enrichment of acetate in lignin treatments after 11 mo indicate that prolonged stimulation of acetogenic activity and CO2 reduction occurs only in lignin-amended treatments.
In conclusion, the enrichment of the Bathy-8 subgroup and the incorporation of IC into putative bathyarchaeotal lipids in lignin amendments, combined with genomic and geochemical evidence, strongly suggest an organoautotrophic lifestyle for these Bathyarchaeota. Members of this ubiquitous subgroup may play an important, previously unrecognized role in the degradation of terrestrially derived lignin in marine sediment. Such an affinity for lignin is consistent with high abundances of Bathyarchaeota in terrestrially influenced continental margin settings (5, 8, 31). In addition to accounting for a major fraction of organic carbon burial in marine sediments, lignocellulose material is the most abundant renewable energy resource on our planet. Our finding that Bathyarchaeota mediate lignin degradation may thus result in new strategies for the use of plant material during bioenergy production.
The uppermost 10 cm of intertidal sediments were collected from Dayangshan Island (30.592817 N, 122.083493 E) in Hangzhou Bay, northern East China Sea, China (Fig. S1). Samples were stored anoxically on ice in gas-tight bags and were transported to the laboratory within 3 h, where they were kept at 4 °C until further treatment. The samples were dominated by silt sediment, which is mainly supplied by the Yangtze River (33) and is considered to be rich in terrigenous organic matter (34).
About 600 g of sediment was thoroughly mixed with 2 L of anaerobic artificial seawater medium (26) without Na2SO4 in an anaerobic chamber (Type B vinyl; Coy). After centrifugation at 7,610 × g in 500-mL centrifuge bottles, the supernatant was removed, and this process was repeated again to largely eliminate the original porewater. Afterward, the sediment was mixed again with 3 L of anaerobic artificial seawater medium and was dispensed into serum bottles as 100-mL sediment slurries. The following five organic growth substrates were added to different experiments at a dose of 50 mg/L: oleic acid (Sinopharm 30138518), phenol (Sinopharm 100153008), casein (Sinopharm 69006227), alkali lignin (Sigma 45-471003), and cellulose (Sinopharm 68005761). In addition, 13C-NaHCO3 was added to a final concentration of 0.25 µM [the percentage of 13C-NaHCO3 (mol/mol) to total NaHCO3 was 5%] to test for the assimilation of IC in archaeal lipids. The bottles were sealed with butyl rubber stoppers and aluminum crimp seals, and then the headspace was evacuated by gas pump and refilled with 99.99% N2 gas in the anaerobic chamber three times, after which a pressure of 200 kPa of N2 was applied to each vial. Two replicates of each treatment were incubated; meanwhile two slurries were also incubated without the addition of any organic substrates as controls. Slurries were incubated horizontally without shaking in the dark at 20 °C for 3.5 mo (t3.5), after which 20 mL of slurries were sampled using a syringe with needle. The collected samples were centrifuged at 13,800 × g to separate the supernatant and sediments and were stored at −80 °C for DNA and lipid isolation and measurement of chemical parameters.
Afterward, higher concentrations of organic substrates (500 mg/L of each substrate) and 13C-NaHCO3 [final concentration is 0.5 µM; the percentage of 13C-NaHCO3 (mol/mol) to total NaHCO3 was 10%] were added to the remaining cultures, which were then incubated for another 2.5 mo (t6). Then 20 mL of the slurries were collected, more organic substrates (500 mg/L of each substrate) were added again, and these samples were incubated for additional 5 mo. Thus, slurries were obtained at three time points after incubation, i.e., at t3.5, t6, and t11 mo.
Total DNA was extracted from 0.2–0.3 g of wet sediment using a PowerSoil DNA Isolation Kit (Mo Bio) according to the manufacturer’s protocol. The concentration of DNA was measured by using a NanoDrop 2000 spectrophotometer (Thermo Scientific), followed by freezing at −80 °C until processing.
Bacterial, archaeal, and bathyarchaeotal 16S rRNA genes were quantified by qPCR and were amplified using the primers Uni519F/Arch908R, Bac341F/prokaryotic519R, and Bathy-442F/Bathy-644R, respectively. The reaction solution included 10 μL of SYBR Premix Ex Taq (TaKaRa), 0.4 μL of ROX reference dye (50×; TaKaRa), 0.8 μM of each primer, and 1 μL of template DNA. Further details, such as primer sequences and PCR conditions, are shown in Table S6. Clones with archaeal, bacterial, and bathyarchaeotal 16S rRNA genes (7) were used for standard curve construction. The R2 and efficiency of each individual qPCR assay are shown in Table S6. The abundance of each targeted gene in the DNA assemblage was determined in triplicate analyses.
The mean doubling times (T) of Bathyarchaeota were calculated according to the formula: T = (t2 − t1)/(log2X2 − log2X1) or T = (t2 − t1)/[3.322 (logX2 − logX1)], which is generated from the formula X1 × 2(t2 − t1)/T = X2. X1 and X2 are the bathyarchaeotal abundance at incubation time t1 and t2, respectively.
Since bathyarchaeotal growth was observed only in treatments to which lignin had been added as a growth substrate, only the original sample (t0), the lignin cultures IC+lignin (t3.5), IC+lignin (t6) × 2, 13C-IC+lignin (t6) × 2, IC+lignin (t11) × 2, and 13C-IC+lignin (t11) × 2, and the control cultures IC (t3.5), IC (t6) × 2, 13C-IC (t6) × 2, IC (t11) × 2, and 13C-IC (t11) × 2 (a total of 19 samples) were processed for archaeal 16S rRNA gene sequencing. The hypervariable V4 regions of archaeal 16S rRNA genes were amplified from original and enrichment samples using primer sets U519F/Arch806R (Table S6) (35). Each 50-μL reaction solution contained 10× PCR buffer, dNTP (100 μM each), 0.25 μM of each primer, 2.5 U of DNA polymerase (Ex-Taq; TaKaRa), and ∼10 ng of total DNA (measured by a NanoDrop 2000). PCR products were purified using an E.Z.N.A. Gel Extraction Kit (Omega Bio-tek) according to the manufacturer’s instructions.
The 16S rRNA gene amplicons containing unique 8-mer barcodes used for each sample were pooled at equal concentrations and were sequenced on the Illumina MiSeq platform using 2× 250-bp cycles and the 500-cycle MiSeq Reagent Kit v2 (Illumina) according to the manufacturer’s instructions. Raw reads were removed if they contained a 50-bp continuous fragment with an average quality score of less than 30 and/or any ambiguities. Filtered reads were merged together using FLASH version 1.2.6 (36). Further analysis was performed using the QIIME standard pipeline (37). In particular, sequence reads were first filtered to remove low-quality or ambiguous reads, and chimeric sequences were removed using the UCHIME program version 4.2 (38). The sequences were clustered into OTUs using UCLUST (38) with a 97% sequence identity threshold. Taxonomy was assigned using the Greengenes database (version gg_13_5, greengenes.secondgenome.com/). The obtained archaeal sequences from each sample were randomly subsampled to generate a uniform depth of 10,372 sequences. Species richness, species diversity, and coverage were computed using QIIME’s alpha diversity pipeline. Five metrics were calculated: the observed number of species, coverage (through rarefaction analyses), the Chao1 index, the Shannon index, and the Simpson index. The number of archaeal sequences in the above-mentioned 19 samples ranged from 10,425 to 33,848 (Table S7). The number of archaeal OTUs (cutoff, 97%) in these 19 sediment samples ranged from 423 to 869. The coverage calculation indicated that the analyzed sequences covered the diversity of archaeal populations in the investigated samples (Table S7).
From the Illumina MiSeq dataset, OTUs affiliated with Bathyarchaeota which contained more than 100 sequences were selected for phylogenetic analyses, and the closest phylogenetic affiliations were determined by constructing a bathyarchaeotal 16S rRNA gene tree using reference sequences of the 17 bathyarchaeotal subgroups (5) as phylogenetic anchors. For the construction of the phylogenic tree, the sequences were aligned using CLUSTALX 1.83, and the reference sequences were trimmed, retaining 900 bp after alignment. Neighbor-joining phylogenetic trees were constructed from pairwise comparisons with the Kimura 2-parameter distance model using the molecular evolutionary genetics analysis (MEGA) program, version 6 (39). Bootstrap values are based on 1,000 replicates for testing the robustness of the inferred topology. Interactive Tree of Life (iTOL) (itol.embl.de/) was used to modify the phylogenetic trees.
Sequences of Illumina sequencing raw data were submitted to GenBank under accession numbers PRJNA398600 and PRJNA398689.
The total lipid extracts (TLEs) were extracted from lyophilized samples (∼1 g) by a modified Bligh and Dyer method (40). For the samples at t6, a TLE aliquot of 90% was chemically treated to convert di- and tetraether lipids into their hydrocarbon derivatives phytane and several biphytanes with 0-3 rings via ether-cleavage methods (41). For samples that had been incubated for 11 mo, intact polar archaeal lipid fractions were first purified by preparative HPLC with a fraction collector (42) and then were converted to hydrocarbon derivatives according to the above method (41). The δ13C values of archaeal intact polar lipid derivatives were determined on a Finnigan Trace gas chromatograph (Thermo Scientific) connected to a DELTA Plus XP isotope ratio mass spectrometer via a GC combustion III interface (both from Thermo Fisher) according to the method described by Kellermann et al. (43).
The lignin degradation was initially surveyed by measuring the concentrations of phenolic compounds in the supernatant of slurries using the Folin–Ciocalteu method (Sinopharm 73104861) as described previously (44).
Concentrations and carbon isotope composition of acetate in the supernatants of slurries were analyzed by on LC linked to isotope ratio mass spectrometry (LC-IRMS) according to the method described by Heuer et al. (45).
We thank Karl O. Stetter (University of Regensburg) and Hongbo Yu (Huazhong University of Science and Technology) for kind suggestions to improve the manuscript and Jiahua Wang (Shanghai Jiao Tong University) for help with the metagenome analysis. This work was supported financially by State Key Research and Development Project of China Grant 2016YFA0601102, Natural Science Foundation of China Grants 41525011, 91751205, and 91428308, and the Deutsche Forschungsgemeinschaft through the Gottfried Wilhelm Leibniz Program Award Hi 616-14-1 (to K.-U.H.). This study is also a contribution to the Integrated Marine Biosphere Research International Project and the Deep Carbon Observatory.
↵1T.Y. and W.W. contributed equally to this work.
Author contributions: T.Y. and F.W. designed research; T.Y., W.W., and W.L. performed research; T.Y., W.W., M.A.L., K.-U.H., and F.W. analyzed data; and T.Y., W.W., M.A.L., K.-U.H., and F.W. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
Data deposition: The sequences reported in this paper have been deposited in the GenBank database (accession nos. PRJNA398600 and PRJNA398689). The metagenome and bathyarchaeotal genomes reported in this paper have been deposited in the GenBank database (accession no. PRJNA418892).
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1718854115/-/DCSupplemental.
Published under the PNAS license.
