Freeze-Coring Method for Characterization of Microbial Community Structure and Function in Wetland Soils at High Spatial Resolution

Abstract

A simple freeze-coring method was developed to obtain structurally intact cores from wetland soils. A copper tube was inserted into the wetland and filled with ethanol and dry ice to freeze the surrounding soil. Biological structure and function could be analyzed, and labile compounds such as mRNA were recovered.

TEXT

Wetlands are regarded as major natural sources of the greenhouse gas methane (CH₄) (29). In submerged, anoxic soils, CH₄ is produced by methanogenic archaea (methanogens) as a final product of organic matter decay. The gas diffuses into the oxic zone, where aerobic methane-oxidizing bacteria (methanotrophs) convert CH₄ to carbon dioxide (4, 5). Therefore, overall CH₄ emissions into the atmosphere are controlled mostly by these two groups of microorganisms. However, in addition, CH₄ can also be directly transported through the aerenchyma of vascular plants (30).

Various techniques have been used to monitor CH₄ production and consumption in the subsurface. For example, turnover rates can be assessed by analyzing concentrations (6) or isotopic signatures of CH₄ along depth profiles (23, 26). In addition, potential CH₄ production and oxidation rates can be assessed by incubating soil slurries and monitoring the concentrations of CH₄ in the headspace (5, 19, 27). The functional genes pmoA and mcrA, which code for key enzymes in the CH₄ oxidation and production pathways, are commonly used as biomarkers for methanotrophs and methanogens (17, 18). Quantification of the abundance of these genes and their transcripts in environmental samples can serve as a proxy for the presence and in situ activity of microorganisms at different depths (7, 15). However, sampling of wetland soils and analysis of processes at high spatial resolution present a significant challenge due to the soil's spongelike consistency (9). Moreover, pore water sampling is often hampered by the extensive root systems of vascular plants. In a study reported by King et al. (12), a core was taken in a wetland for analysis of the rhizosphere structure. However, the core was frozen only after removal and a certain degree of compaction cannot be excluded. For the recovery and analysis of mRNA, rapid freezing of environmental samples is essential to avoid degradation processes (25).

Freeze-coring methods have been developed mainly for the analysis of lake and river sediments to overcome the problems associated with compaction and water loss (3, 8, 21). Subsamples of frozen cores were used for cell counts, fatty acid extraction, and geochemical analyses (8, 21, 32). These approaches were feasible to achieve high-resolution analysis in the first 15 cm of river sediments (8, 32). Freeze-coring of wetland soils such as fens has not been reported.

In this study, we adapted and optimized a freeze-coring method to recover structurally intact cores from an alpine fen located near Oberaar (46°32′50″N, 8°15′41″E, Canton of Bern, Switzerland) at 2,320 m above sea level. Methane emissions from the coring site was previously quantified using static chambers (15) and found to be 225 ± 19 mg CH₄ m⁻² day⁻¹. Our major objective was to retrieve a frozen core of wetland soil that would enable us (i) to investigate the microbial community and rhizosphere structure at high spatial resolution and (ii) to quantify the abundance of selected functional genes and transcripts (i.e., mRNA). This was achieved by manually inserting a closed-bottom copper tube (50-cm length, 3-cm outside diameter, and 1-mm wall thickness) into the wetland soil (Fig. 1). A funnel was connected to the top of the tube, and a plastic cylinder (10.5 cm high, 19 cm wide) covered with insulating material was placed around the tube in the water to minimize thermal exchange and facilitate efficient freezing even at high ambient air and water temperatures (Fig. 1). The tube was filled with roughly 300 ml of ethanol precooled on dry ice. Subsequently, dry ice pellets were added repeatedly to maintain a temperature of approximately −70°C (13). After about 1 h, a frozen core with a diameter of approximately 15 cm and a length of approximately 45 cm was cut out from the surrounding roots with a sharp knife. The core, including the copper tube in the center, was pulled from the wetland using a metal rod inserted into two predrilled holes near the top of the tube. Subsequently, the extracted core was transported to the laboratory on dry ice and stored at −80°C until further analysis.

Fig 1.

Illustration of the method described in this report: 1, copper tube; 2, funnel; 3, ethanol; 4, dry ice pellets and CO₂ bubbles; 5, insulation material (Styropor cylinder; 3.5 cm high, 19 cm in diameter); 6, plastic cylinder; 7, core after freezing; 8, analyzed section of the core.

The frozen core was dissected in a −20°C chamber with a sterile saw, first vertically along the copper tube and subsequently horizontally at approximately 2-cm intervals. Rims and the tube-facing side of 1 cm of each section were cut off to remove material that was disturbed by insertion of the tube and retrieval of the core (8 in Fig. 1). Selected sections (9 to 11, 19 to 21, 29 to 31, and 39 to 41 cm) were used to determine soil dry weight (15) and porosity (22). The overall dry weight was found to be 13% ± 4%, and the bulk soil porosity was 0.84 ± 0.06. One gram of each section was thawed, dissolved in 25 ml of distilled water, and filtered through a 1-mm sieve. The biomass recovered on the filter was resuspended in 10 ml of water and analyzed for root, leaf, and moss surface area using the automated picture analysis software WinRHIZO PRO (Regent Instrument Inc., Quebec, Canada; see Fig. S1 in the supplemental material) as previously described (11). Subsequently, the recovered plant material was filtered through a 2.7-μm glass fiber filter for dry weight, total carbon (TC) and total nitrogen (TN) determination (2). The 1-mm filtrate was treated in the same way. The largest root surface area was detected at a depth of 29 to 31 cm, where the highest dry weight of the recovered fraction >1 mm was also observed (Table 1).

Table 1.

Characterization of 1 g of the frozen core

Depth (cm)	Fraction >1 mm						Fraction between 2.7 μm and 1 mm
	Surface area (cm2)			Dry wt (g)	TC (%)	TN (%)	Dry wt (g)	TC (%)	TN (%)
	Root	Leaf	Moss
9–11	27 (46.3)a	12 (20.0)	20 (33.8)	0.047	49.3	1.7	0.080	32.2	1.8
19–21	50 (61.0)	13 (15.9)	19 (23.2)	0.048	50.6	1.6	0.038	38.9	1.9
29–31	88 (94.1)	6 (5.9)	0 (0.0)	0.076	46.5	1.0	0.032	35.5	1.9
39–41	46 (82.0)	10 (18.0)	0 (0.0)	0.037	46.4	1.6	0.113	32.7	1.9

^aValues in parentheses are percentages of the total surface area of roots, leaves, and mosses.

Subsamples of the different core sections were used for extraction of total DNA and RNA, followed by reverse transcription of RNA and real-time PCR to quantify pmoA and mcrA DNA and cDNA sequence fragments (for detailed methodology, see the supplemental material). The total DNA concentration was low in the top 2 cm of the soil but then increased significantly. Concentrations were above 10 μg (g soil)⁻¹ between 2 and 11 cm and then dropped continuously to reach levels around 5 μg (g soil)⁻¹ at depths below 25 cm (Fig. 2A). The total RNA concentrations showed a similar pattern (Fig. 2A).

Fig 2.

(A) Concentrations of DNA (♦) and RNA () extracted from subsamples of the frozen core. (B) Abundance of total methanogenic archaea and methanotrophic bacteria based on functional gene mcrA (▼) and pmoA () copy numbers, respectively. (C) Abundance of active methanogenic archaea and methanotrophic bacteria based on the copy numbers of transcripts of the functional genes mcrA (▼) and pmoA (), respectively. The x axis starts at 10⁴ for better visualization. Symbols on the left of the broken line represent concentrations below the detection limit (10³ copy numbers [g soil]⁻¹).

Real-time PCR on the pmoA and mcrA genes was used to quantify the total (DNA) and active (cDNA) methanotrophs and methanogens present in the different core sections. The highest pmoA gene DNA copy numbers were detected in the 2- to 4-cm section, and mcrA was most abundant at 4- to 11-cm depths (Fig. 2B). This spatial overlap of the highest abundance of methanotrophs and methanogens coincides with a previous wetland study in which the highest abundance of both groups was also observed at overlapping depths (15). Detected cDNA copy numbers followed a similar trend in that the highest methanotrophic activity was observed at 2 to 6 cm, while the highest methanogenic activity was located in two distinct regions, at 2 to 18 cm and at 31 to 43 cm (Fig. 2C). Nevertheless, despite the detection of the pmoA and mcrA genes throughout the entire core, transcripts of pmoA and mcrA could be detected only at specific depths, indicating the presence of “hot spots” of activity. In agreement with our results, it has been shown elsewhere that potential CH₄ production is high just below the soil surface, decreases with depth, and increases again in deeper soil layers (5, 19). The low potential CH₄ production observed in the intermediate layers has been suggested to be a consequence of competition with sulfate-reducing bacteria (19). However, detailed analyses would be required to determine whether such competition takes place in our fen.

Community profiling of mcrA and pmoA DNA fragments was performed by terminal restriction fragment length polymorphism (T-RFLP; for methodology, see the supplemental material). Profiles clearly showed that the community structure in the top few centimeters of soil was distinct from the remaining sections (Fig. 3). Type II methanotrophs (T-RF 245) were dominant just below the vegetation surface (0 to 4 cm), while type Ia methanotrophs (T-RFs 78, 350, and 513) dominated in all of the remaining sections (Fig. 3A) (10).

Fig 3.

Relative abundances of T-RFs of functional genes pmoA (A) and mcrA (B) obtained along the soil core. The four most dominant T-RFs are represented. The T-RFs are identified by their lengths in base pairs. Each value on the y axis represents the lower end of a section.

The top few centimeters of the fen are influenced by high fluctuations of temperature and water content. Type II methanotrophs might be better adapted to these changing conditions due to their ability to form cysts as resting stages (31) and therefore outcompete type I methanotrophs in the layer just below the vegetation surface. In addition, it has been suggested that type II methanotrophs are dominant in methanotrophic communities at temperatures higher than 15°C (20), which are reached during summer at the surface of the fen studied here (data not shown). Type Ia methanotrophs, on the other hand, have been reported to be dominant under oxygen-limited and methane-rich conditions (1), found, e.g., in the rhizosphere of rice plants (24) or in the deeper layers of the fen studied here. Furthermore, type Ia methanotrophs have been observed to be the dominant community in cold environments such as arctic (14, 16, 28) and Tibetan (33, 34) wetlands.

T-RFLP analyses of the mcrA clones allowed assignment of the T-RFs detected in the soil core to phylogenetic groups of methanogens (see Table S2 in the supplemental material). Four main T-RFs were observed: T-RF 56 (fen cluster), T-RF 175 (fen cluster), T-RF 192 (Methanosaetaceae), and T-RF 470 (rice cluster I). In the sample taken at a depth 0 to 2 cm, only the T-RFs affiliated with the fen cluster could be detected, while all four main T-RFs were present at all other depths (Fig. 3B).

The applicability of the novel freeze-coring method was tested in two other Swiss alpine fens (see the supplemental material). Cores similar in length and diameter could be extracted from both fens despite considerable differences in soil dry weight from the Oberaar fen. Additionally, in one fen, we tested whether ethanol and dry ice could be replaced with liquid nitrogen as a cooling agent. It was found that large quantities of liquid nitrogen were required since it evaporated too quickly. Moreover, the freezing process was uneven and resulted in a pear-shaped core with a greater radius at depths of 30 to 45 cm.

Our results show that the newly developed freeze-coring method using chilled ethanol and dry ice allows the recovery of soil cores for downstream molecular biological and soil physical analyses of wetland soils. Samples can be analyzed at high spatial resolution down to a depth of at least 45 cm. The abrupt changes in community structure and abundance of methanotrophs and methanogens, in particular within the first few centimeters, would most likely have been overlooked by traditional sampling techniques, where compaction of the soil cannot be avoided. In addition, mRNA extraction from soil, a major challenge in environmental studies, was possible and allowed the detection of distinct hot spots of activity.

Nucleotide sequence accession numbers.

The mcrA sequences obtained in this study have been submitted to the EMBL database under accession numbers HE774269 to HE774298.