Carbon and nitrogen assimilation in deep subseafloor microbial cells

Abstract

Remarkable numbers of microbial cells have been observed in global shallow to deep subseafloor sediments. Accumulating evidence indicates that deep and ancient sediments harbor living microbial life, where the flux of nutrients and energy are extremely low. However, their physiology and energy requirements remain largely unknown. We used stable isotope tracer incubation and nanometer-scale secondary ion MS to investigate the dynamics of carbon and nitrogen assimilation activities in individual microbial cells from 219-m-deep lower Pleistocene (460,000 y old) sediments from the northwestern Pacific off the Shimokita Peninsula of Japan. Sediment samples were incubated in vitro with ¹³C- and/or ¹⁵N-labeled glucose, pyruvate, acetate, bicarbonate, methane, ammonium, and amino acids. Significant incorporation of ¹³C and/or ¹⁵N and growth occurred in response to glucose, pyruvate, and amino acids (∼76% of total cells), whereas acetate and bicarbonate were incorporated without fostering growth. Among those substrates, a maximum substrate assimilation rate was observed at 67 × 10⁻¹⁸ mol/cell per d with bicarbonate. Neither carbon assimilation nor growth was evident in response to methane. The atomic ratios between nitrogen incorporated from ammonium and the total cellular nitrogen consistently exceeded the ratios of carbon, suggesting that subseafloor microbes preferentially require nitrogen assimilation for the recovery in vitro. Our results showed that the most deeply buried subseafloor sedimentary microbes maintain potentials for metabolic activities and that growth is generally limited by energy but not by the availability of C and N compounds.

Numerous studies of microbial ecology have shown that the dynamics of microbial activities in any ecosystem on this planet largely depend on the availability of energy and nutrient substrates in the environment (reviewed by Morita in ref. 1). Microbial life has been identified in various natural environments, including deep and ancient subseafloor sediments down to 1,627 m below the seafloor (mbsf) (2). Estimates of naturally occurring microbial populations have suggested that a significant fraction of the total living biomass on Earth is present in the subseafloor sedimentary biosphere (3–5). Chemical profiles in pore water from sediment drill cores have suggested that the metabolic activities of subseafloor life are generally extremely low (6–8) because of severe limitations in the availability of electron donors and acceptors for cell respiration and growth (9). Radioactive tracer incubation studies have consistently shown that the theoretical mean generation time of subseafloor microbial cells is extremely long, ranging from a few to thousands of years (9, 10). The net activity of the subseafloor sedimentary microbial ecosystem plays an ecologically important role in biogeochemical cycles, such as the degradation of buried organic matter and methane production.

Other studies of subseafloor life in sediment on the continental margins have shown that the microbial biomass positively correlated with the organic concentration in sediment (5, 8). This finding indicates that sedimentary microbial ecosystems on most continental margins consist mainly of organic-fueled heterotrophs with relatively minor autotrophic components (e.g., methanogens) (11). The composition of carbon isotopes (δ¹³C) in both extracted intact polar lipids and cellular bodies assessed by secondary ion MS (SIMS) from core sediments off the coast of Peru revealed constant values similar to the δ¹³C of buried organic matter, even at the sulfate–methane transition zones where ¹³C-depleted biogenic methane is expected to mediate microbial carbon metabolism through anaerobic oxidation of methane (12). Molecular ecological studies based on PCR-amplified 16S rRNA (and its DNA) have revealed the presence of diverse archaea and bacteria from organic-rich subseafloor sediments, of which most sequences are phylogenetically distinct from physiologically known isolates (13–16), and their community compositions are stratified with depth and redox interfaces such as sulfate–methane transition zones (17, 18). Metagenomic analysis of the Peru margin sediment revealed that over 80% of the genetic assemblages were functionally unknown (19), whereas some key functional genes and activities related to biogeochemical carbon cycles from various locations on continental margins have been reported (14, 20–22).

Despite the widespread occurrence of significant populations of microbial communities in the subseafloor sedimentary environment, the physiological nature and nutrient energy requirements for the long-term maintenance of essential life functions have seldom been tested (23, 24). For example, whether the low availability of nutrient and energy substrates constrains the physiological status and growth potential of deeply buried microbial cells is essentially unknown. More generally, differentiating whether these subseafloor microbes are alive, growing, simply surviving, dormant, or dead fossils is important. To address these fundamental questions regarding the ecological physiology of subseafloor sedimentary microbial life, we studied the carbon and nitrogen assimilation in a deeply buried microbial community at the single-cell level using stable isotope (¹³C and/or ¹⁵N) tracers, image-based cell enumeration, catalyzed reporter deposition (CARD) -FISH, and nanometer-scale SIMS (NanoSIMS).

Results and Discussion

Sampling Location and Geochemical Settings.

The sample studied herein represents typical organic-rich deep (sediment depth of 219 m) and ancient (ca. 460,000 y old) sedimentary habitats in the northwestern Pacific Ocean (25). The drilling vessel (DV) Chikyu launched the shakedown cruise CK06-06 at Site C9001 in the northwestern Pacific Ocean (41°10′38.28″ N, 142°12′04.89″ E) off the Shimokita Peninsula of Japan in 2006 (Fig. S1). The cored sediment was diatom-rich hemipelagic clay containing a large amount of methane and organic carbon (1.2% weight of total organic carbon). The pore water from the examined sediment core at 219 mbsf was pH 7.8, and it contained dissolved inorganic carbon (DIC), acetate, and ammonium concentrations of ∼104, 0.23, and 15 mM, respectively. Neither sulfate nor nitrate was detected in the pore water. Detailed geochemical profiles of the squeezed pore waters have been described elsewhere (26). Sediment slurry was aseptically prepared in the microbiology laboratory on DV Chikyu and portioned under anaerobic conditions. Subsamples were collected immediately after core recovery. During subsample processing on board, the core was placed in a laminar flow clean cabinet, and its center was sampled (50 cm³) using a 50-mL tip-cut sterilized syringe. Two volumes of sulfate-free, anaerobic artificial seawater (pH 7.8) were added to prepare a slurry sample under an atmosphere of 0.22-μm filtered nitrogen.

¹³C and ¹⁵N Incorporation Analyzed Using NanoSIMS.

Aliquots were separately supplemented with either one ¹³C-labeled carbon substrate (1 mM glucose, 1 mM acetate, 1 mM pyruvate, and 1 mM bicarbonate or 3 atm of methane in headspace) and a ¹⁵N-labeled nitrogen source (0.1 mM ammonium) or a ¹³C- and ¹⁵N-labeled mixture of 20 aa (1 mM in total). After a 65-d stationary incubation at 10 °C in the dark, the total cell abundance and number of cells incorporating carbon (¹³C) and nitrogen (¹⁵N) were estimated by a fluorescent microscopic image-based cell enumeration technique (27) and NanoSIMS.

Analysis using NanoSIMS showed that nitrogen was assimilated under all incubation conditions and that carbon was assimilated in all except those conditions incubated with ¹³C-methane (Figs. 1 and 2A). The number of cells increased in sediment samples incubated with glucose, pyruvate, or amino acids mixture, suggesting that these compounds were used not only for assimilatory processes and meeting maintenance energy requirements but also for cell growth. Approximately 76%, 3.9%, and 41% of the original microbial population incorporated ¹³C from glucose, pyruvate, and amino acids mixture, respectively (Table 1). This finding suggests that a large fraction of subseafloor microbes are (facultative) heterotrophs with anaerobic glycolysis and fermentation pathways through pyruvate and/or aminolysis. Assuming that the cells incorporating ¹³C accounted for the increase in cell abundance over the course of the incubation, we estimated that pyruvate-incorporating cells divided 6.0 times on average, whereas those cells that incorporated glucose and amino acids divided 3.0 and 1.6 times, respectively (Table 1). These results suggested that pyruvate consumers are less widespread in the subseafloor than glucose and amino acids, but when pyruvate becomes available, small populations of pyruvate consumers may respond and rapidly increase.

Fig. 1.

NanoSIMS images of subseafloor cells that incorporated stable isotope (¹³C and/or ¹⁵N) -labeled substrates. Color gradient indicates ¹³C or ¹⁵N abundance expressed as ¹³C/¹²C or ¹⁵N/¹⁴N for incubations with (A) ¹³C-glucose and ¹⁵N-ammonium, (B) ¹³C-acetate and ¹⁵N-ammonium, (C) ¹³C-pyruvate and ¹⁵N-ammonium, (D) ¹³C-bicarbonate and ¹⁵N-ammonium, (E) ¹³C,¹⁵N amino acids mix, and (F) ¹³C-methane and ¹⁵N-ammonium.

Fig. 2.

Numbers of subseafloor microbial cells that incorporated ¹³C- or ¹⁵N-labeled substrates. (A) Number of total cells (orange bars) and cells that incorporated ¹³C (red bars) and ¹⁵N (green bars). (B) Phylogenetic identification of domain-specific probes for bacteria [EUB338 (I, II, III probes mix)] and archaea (ARC915) by CARD-FISH. Bacterial, archaeal, and unidentified (i.e., did not hybridize with these probes) fractions of cells that incorporated ¹³C and/or ¹⁵N are shown in blue, red, and gray, respectively.

Table 1.

Fraction, growth, and carbon and nitrogen assimilation rates of deep subseafloor microbial cells after 65 d of incubation

Substrate		Number of cells (107 cells/cm3)						Assimilation rate‡		Td-ar (estimated Td from assimilation rate)§
13C	15N	Total	13C incorporated	15N incorporated	Original fraction of 13C-incorporated cells (%)*	Estimated cell division (times)†	Td-noc (estimated Td from cell division in days)†	C (10−18 mol/cell per d)	N (10−18 mol/cell per d)	C base (days)	N base (days)
Glucose	Ammonium	16.1	15.4	15.2	76.2	3.0	22.0	14.6 ± 8.7	12.0 ± 5.4	675.8	165.4
Acetate	Ammonium	2.53	0.31	0.62	12.2	NA	NA	21.7 ± 13.0	20.9 ± 13.2	497.2	252.3
Pyruvate	Ammonium	5.60	2.96	2.96	3.9	6.0	10.8	24.8 ± 11.6	20.5 ± 7.0	424.5	62.6
Bicarbonate	Ammonium	2.52	0.56	0.56	22.1	NA	NA	67.3 ± 39.2	17.7 ± 5.5	192.4	92.7
Amino acids	Amino acids	4.79	3.15	3.15	41.1	1.6	40.8	43.2 ± 27.4	4.6 ± 2.1	210.8	261.8
Methane	Ammonium	2.74	ND	1.19	43.4	NA	NA	NA	17.5 ± 9.4	NA	164.7

NA, not applicable; ND, not detected; T_d, doubling time.

*The original fraction of ¹³C- or ¹⁵N-incorporated cells was calculated from the population that did not incorporate ¹³C or ¹⁵N from the supplemented substrates and subtracted from the negative control.

^†The number of cell divisions and T_d were estimated from the original fraction of ¹³C- or ¹⁵N-incorporated cells.

^‡C and N assimilation rates were calculated by the total cellular C and N contents as 86 fg-C and 20 fg-N, respectively (3).

^§Cell division was estimated to occur when the equivalent amount of C and N to original mass was assimilated.

The numbers of cells did not increase in sediment samples supplemented with acetate, bicarbonate, and methane (Fig. 2A). These findings indicate that these subseafloor microbes cannot retrieve energy required for cell growth from these carbon compounds. However, 12% and 22% of the microbial population incorporated ¹³C from acetate and bicarbonate, respectively, and 24% and 22% incorporated ¹⁵N from ammonium in the samples supplemented with acetate and bicarbonate, respectively (Fig. 2A). These results indicate that subseafloor life retains the potential to assimilate carbon and nitrogen and that at least 22% of the examined microbial population is capable of enzymatic CO₂ incorporation (e.g., phosphoenolpyruvate carboxylase) or autotrophic CO₂ fixation. The estimated intracellular ¹³C incorporation rate from bicarbonate was 67 × 10⁻¹⁸ mol/cell per d, which is approximately two- to threefold higher than the rate of heterotrophic substrates (Table 1), suggesting that most bicarbonate is incorporated through an autotrophic and/or mixotrophic (i.e., CO₂ fixation through heterotrophic energy respiration) carbon assimilation pathway. However, it remains unknown if those cells used potential intracellular energy (i.e., ATP) or uncharacterized energy sources in the incubation slurry for the CO₂ fixation.

We also observed ¹⁵N incorporation from ammonium in the sample supplemented with ¹³C-methane (43% of total cells) (Fig. 2A), although ¹³C incorporation was undetectable. Additional experiments where sulfate, nitrate, manganese, or ferrihydrite were added as potential electron acceptors also resulted in neither cell growth nor ¹³C uptake from methane. Why nitrogen uptake nonetheless proceeded is more difficult to explain. Conceivable explanations are that (i) methane cannot act as an energy source or a carbon source for the deep subseafloor microbial community or (ii) the carbon turnover and/or incorporation rate of cells that mediate anaerobic oxidation of methane may take place on longer timescales than the heterotrophic substrates examined (28, 29). Nevertheless, even if cells retrieved energy from methane or uncharacterized energy sources, the amount of retrieved energy was insufficient to support cell growth during the incubation experiment for 65 d.

Visualization of Active Cells Using CARD-FISH.

We attempted to identify ¹³C- and ¹⁵N-incorporating cells using CARD-FISH with archaea- and bacteria-specific probes (30, 31). Contrary to our expectation, most metabolically active cells did not hybridize with common FISH probes (Fig. 2B). All cells that had incorporated acetate and ∼50% of those cells that had incorporated amino acids hybridized with the bacteria-specific probe, whereas less than 1% hybridized with the archaea-specific probe. Although CARD-FISH can detect even a few ribosomal RNA molecules (30), detection generally relies on the permeation of relatively large horseradish peroxidase (HRP)-labeled probes across the cell membrane. Precise detection of archaeal cells surrounded by a rigid cell wall and/or crystal proteins is reportedly difficult (32). In addition, whether probes accurately match the sequences of uncharacterized deep subseafloor life (33) and whether cells that do not hybridize with conventional probes belong to archaea, bacteria, or other unknown life forms remain unknown.

Growth Response to Nutrient and/or Energy Substrates.

We calculated ¹³C and ¹⁵N incorporation in individual cells (atom percent of cellular C and N) within each sample using NanoSIMS. The amounts of carbon and nitrogen incorporated from glucose, pyruvate, and amino acids were significantly lower than expected from cell division (Fig. 3A). The average doubling time (T_d) estimated from cell numbers also considerably deviated (cell division, T_d-noc = 10.8–40.8 d) from those numbers estimated by the assimilation rates of carbon and nitrogen substrates (T_d-ar = 62.6–675.8 d) (Table 1).

Fig. 3.

Carbon and nitrogen incorporation by subseafloor microbial cells from stable isotope-labeled substrates. (A) Ratios of incorporated carbon (red bars) and nitrogen (green bars) contents to total carbon or nitrogen are expressed by atomic percentages. Error bars show SD of incorporated carbon and nitrogen contents from substrates. Concentrations of ammonium and bicarbonate (DIC) in the original sample determined on board (25, 26) were used to calculate substrate incorporation ratio (atom percent) for ammonium and bicarbonate in single cells. (B) Scatter plot of incorporated carbon and nitrogen in total cellular carbon and nitrogen of individual cells. Each point represents the atomic percentage of glucose (red), acetate (blue), pyruvate (green), bicarbonate (brown), amino acids (purple), or methane (orange) in each cell examined by NanoSIMS.

One explanation is that cells under conditions in situ are limited by the availability of compounds for energy production but not by the availability of compounds for biomass assimilation. After cells have been activated in vitro with labile, energy-rich substrates, such as glucose, pyruvate, and amino acids, and are no longer energy-starved, they will readily incorporate organic compounds present in the environment. This finding indicates that not all carbon compounds used in biomass assimilation can support energy production. Our observation that acetate and bicarbonate were incorporated during incubations without fostering cell growth supports this notion, albeit that possible energy sources for acetate and bicarbonate incorporation remain elusive.

The intracellular ratios of ¹³C to ¹⁵N notably showed that atomic ratios of nitrogen incorporation from ammonium in the total cellular nitrogen consistently exceeded those ratios of carbon (Figs. 3B and 4B). These data show that subseafloor cells preferentially require nitrogen assimilation from ammonia for the better recovery in vitro, which may contradict the presence of enough ammonium (∼15 mM) in the sediment pore water. A conceivable explanation is that, although most subseafloor microbes could assimilate nitrogen from ammonium, they might suppress ammonium uptake in situ to conserve or maintain energy (i.e., although pore water from sediments at 219 mbsf contain 15 mM ammonium, the metabolism required for nitrogen assimilation into amino acids or calbamoyl phosphate synthesis might consume excess energy required for long-term survival). Another possibility is that subseafloor cells have incorporated the high C/N buried organic matter into the biomass rather than ammonium, which may subsequently result in nitrogen-deprived biomass. However, these assumptions need to be carefully shown in future studies by evaluating carbon and nitrogen composition (i.e., C/N ratio) and the physiological nature of subseafloor cells.

Fig. 4.

Localization of intracellular ¹³C and ¹⁵N incorporation in a deep subseafloor microbial community. (A) Example of overlaid ratio images of ¹³C_glucose/¹²C (red) and ¹⁵N_ammonium/¹⁴N (green) visualized using NanoSIMS. (B) Scatter plot of atomic percentages of carbon and nitrogen incorporation in each of numbered areas shown in A.

The incorporation ratios of carbon only exceeded the ratios of nitrogen in the incubation with amino acids to which no ¹⁵N-ammonium was added. Amino acids are released by cell death and subsequent enzymatic degradation. Because microbial populations generally decrease with depth, released amino acids or amino acid derivatives might be directly or indirectly used by conversion or synthesis without consuming excess energy to repair aging proteins. Such a process might be very different to surface microbial ecosystems, where most organic matter generally breaks down to amino acids or even carbon skeletons before biosynthesis, and hence, a metabolic suppression and recycling system might play important roles in maintaining life in the deep subseafloor sedimentary environment.

Intracellular Localization of C and N Uptake Visualized Using NanoSIMS.

Subseafloor microbial communities are phylogenetically diverse (14–16), and images of ¹³C- and ¹⁵N-incorporating cells from NanoSIMS analyses revealed a varied size and morphology of metabolically active microbial cells (Figs. 1 and 4A). Cell growth and physiological states are likely to differ among species (34). The NanoSIMS images showed that ¹³C and ¹⁵N uptake is not uniform but localized within individual cells (Fig. 4). These observations suggest that activated deep subseafloor life cannot smoothly respond to sudden changes of surrounding high nutrient and energy conditions in vitro and that some cellular components might become damaged during long-term survival without sufficient repair because of the absence of a retrievable energy source (24). In addition, our results support previous cultivation efforts on deep subseafloor sedimentary habitat (6, 11, 13, 35–37); most of deep microbes are indeed alive and hence, have potential to be cultivated in the laboratory condition.

Ecological Interpretation and Research Perspectives.

Our data showed that microbial life forms recovered from the 460,000-y-old deep subseafloor sediment are generally alive and maintain metabolic potentials of carbon and nitrogen assimilation and growth. Up to 76% and 22% of total cells in sediment incorporated heterotrophic substrates and CO₂ into their biomass carbon, respectively, which is in good agreement with previous observations that the microbial community in organic-rich subseafloor sediments on continental margins consists mainly of heterotrophs (5, 12). One of the interesting features discovered in this study is that subseafloor cells preferentially incorporate nitrogen to biomass during the in vitro incubation experiment. However, in situ ecological physiology, metabolic functions, and survival strategies in the subseafloor sedimentary ecosystem remain to be shown with more detailed molecular and isotopic studies at the single-cell level. Freshly cored sediments at various energetic settings from shallow to deep subseafloor realms need to be investigated through future scientific ocean drilling (38) to increase our understanding of the biogeochemical consequences and microbiological nature of subseafloor life.

Materials and Methods

Incubation Experiments.

Carbon sources, which were 1 mM ¹³C-labeled (99.9% atom) glucose, acetate, pyruvate, bicarbonate, or amino acids mixture (20 aa in roughly equal molar amounts), were added to sediment slurries or 3 atm methane was added in the headspace; 0.1 mM ¹⁵N-labeled ammonium was added to all samples, except those samples with ¹³C- and ¹⁵N-labeled amino acids to provide a nitrogen source. Approximately one-half of each slurry was fixed with 4% paraformaldehyde after 65 d of incubation at 10 °C and preserved in PBS/ethanol (1:1 vol/vol) at −20 °C until analysis.

Cell Detection and Enumeration.

Up to 50 μL fixed slurry were mixed with 3% NaCl, sonicated at 20 W for 1 min, and filtered through a platinum (Pt) -coated polycarbonate membrane (0.22-μm pore size; Millipore). The membrane retained on the filtration device (Millipore) was immersed in 1 mL 0.1 M HCl for 5 min, washed with 5 mL Tris-EDTA (TE) buffer (10 mM Tris⋅HCl, 1.0 mM EDTA, pH 8.0), and air-dried. Approximately one-quarter of the membrane was cut, placed on a cellulose acetate membrane (ADVANTEC), and placed in SYBR Green 1 staining solution (1:40 vol/vol SYBR Green 1 in TE buffer) for 10 min. The staining solution was removed by vacuum filtration, and the membranes were placed on glass microscope slides and mounted with 3 μL mounting solution (2:1 mixture of VECTASHIELD mounting medium H-1000 and TE buffer). The microbial cells on the membrane were counted using an automated epifluorescent microscope (BX-51; Olympus) with a band-pass filter of 490/20 nm (center wavelength to bandwidth) for excitation and a long-pass filter at 510 nm cutoff (27, 39).

CARD-FISH.

Fixed microbial cells were separated from sediment particles by layering a cushion of 500 μL 50% (wt/v) Nycodenz below the slurry through a needle and centrifuging at 3,000 × g for 10 min (40). The supernatant, including the interface region of slurry and Nycodenz layers, was carefully removed, transferred to a separate vial, and filtered through a Pt-coated polycarbonate membrane to trap microbial cells. We performed CARD-FISH using a standard protocol (30) with slight modifications. HRP-labeled origonucleotide probes [EUB338 (I, II, III) for bacteria and ARC915 for archaea (30, 31)] were used for CARD-FISH, and hybridization was conducted under 5% (bacteria) or 35% (archaea) formamide concentration at 40 °C for 2.5 h. Cells were mounted on the membranes with marked grids using a laser microdissection microscope (LMD6000; Leica Microsystems) in a High Efficiency Particulate Air (HEPA)-filtered clean cabinet (Fig. S2). Details of the protocol are provided in SI Materials and Methods.

NanoSIMS Analysis of Single-Cell Image Acquisition and Data Processing.

Microbial cells that incorporated stable isotope-labeled substrates were analyzed using NanoSIMS 50 (AMETEK Co. Ltd.; CAMECA) at the University of Tokyo in Japan and the Curie Institute in France. Samples on the Pt-coated polycarbonate membrane were presputtered at high beam currents (30 pA/s per μm²) before measurement. The ¹²C, ¹³C, ¹²C¹⁴N, and ¹²C¹⁵N ions were collected and measured in parallel at a mass resolution sufficient to separate ¹³C from ¹²CH and ¹²C¹⁵N from ¹³C¹⁴N. Samples were measured using a 2- to 4-pA Cs⁺ primary beam that was stepped over a 50 × 50-μm field of a 512 × 512-pixel raster with a counting time of 5 ms/pixel. Detailed images of cells were obtained using a 0.5- to 1-pA Cs⁺ primary beam stepped over a 5 × 5- to 20 × 20-μm field of a 512 × 512-pixel raster with a counting time of 1 ms/pixel, and final images were created from amalgamating of 5–10 images from the same analysis area. The incised grid on membranes after CARD-FISH was first recorded with a CCD camera, and specific grid locations were overlaid on fluorescent microscopic images. Recorded images and data were processed using CAMECA WinImage software. Different scans of each image were aligned to correct image drift during acquisition. Final images were created by adding the secondary ion counts of each recorded secondary ion from each pixel over all scans. Intracellular carbon and nitrogen uptake from stable isotope-labeled substrates was calculated by drawing regions of interest on CN images and calculating ¹³C/¹²C and ¹⁵N/¹⁴N ratios (inferred from the ¹²C¹⁵N/¹²C¹⁴N ratio), and data from a blank filter area was used for standardizing multiple analysis data. Concentrations of ammonium and DIC in original sample determined on board (25, 26) were used to calculate substrate incorporation ratio (atom percent) for ammonium, acetate, and bicarbonate in single cells.