ABSTRACT
Ombrotrophic peatlands are a recognized global carbon reservoir. Without restoration and peat regrowth, harvested peatlands are dramatically altered, impairing their carbon sink function, with consequences for methane turnover. Previous studies determined the impact of commercial mining on the physicochemical properties of peat and the effects on methane turnover. However, the response of the underlying microbial communities catalyzing methane production and oxidation have so far received little attention. We hypothesize that with the return of Sphagnum spp. postharvest, methane turnover potential and the corresponding microbial communities will converge in a natural and restored peatland. To address our hypothesis, we determined the potential methane production and oxidation rates in natural (as a reference), actively mined, abandoned, and restored peatlands over two consecutive years. In all sites, the methanogenic and methanotrophic population sizes were enumerated using quantitative PCR (qPCR) assays targeting the mcrA and pmoA genes, respectively. Shifts in the community composition were determined using Illumina MiSeq sequencing of the mcrA gene and a pmoA-based terminal restriction fragment length polymorphism (t-RFLP) analysis, complemented by cloning and sequence analysis of the mmoX gene. Peat mining adversely affected methane turnover potential, but the rates recovered in the restored site. The recovery in potential activity was reflected in the methanogenic and methanotrophic abundances. However, the microbial community composition was altered, being more pronounced for the methanotrophs. Overall, we observed a lag between the recovery of the methanogenic/methanotrophic activity and the return of the corresponding microbial communities, suggesting that a longer duration (>15 years) is needed to reverse mining-induced effects on the methane-cycling microbial communities.
IMPORTANCE Ombrotrophic peatlands are a crucial carbon sink, but this environment is also a source of methane, an important greenhouse gas. Methane emission in peatlands is regulated by methane production and oxidation catalyzed by methanogens and methanotrophs, respectively. Methane-cycling microbial communities have been documented in natural peatlands. However, less is known of their response to peat mining and of the recovery of the community after restoration. Mining exerts an adverse impact on potential methane production and oxidation rates and on methanogenic and methanotrophic population abundances. Peat mining also induced a shift in the methane-cycling microbial community composition. Nevertheless, with the return of Sphagnum spp. in the restored site after 15 years, methanogenic and methanotrophic activity and population abundance recovered well. The recovery, however, was not fully reflected in the community composition, suggesting that >15 years are needed to reverse mining-induced effects.
INTRODUCTION
Peatlands play a crucial role in the global carbon budget. Although peatlands are a net carbon sink, northern ombrotrophic peatlands are also a source of methane, a potent greenhouse gas with a 34-fold stronger radiative forcing effect than carbon dioxide on a 20-year time scale (1). Peatlands, along with other wetlands, contribute approximately 23% of the total methane budget of 500 to 600 Tg per annum (2). The amount of methane released from peatlands would be higher if not for the methanotrophs acting as a natural filter at oxic-anoxic interfaces to mitigate methane emission (3–5); it has been estimated that up to 90% of biologically produced methane is consumed before being released into the atmosphere in this environment (6). Net methane released is thus regulated by methane production and oxidation. Methane-cycling processes, along with the carbon sink function of peatlands, can be significantly affected by peat drainage and mining.
Mining alters the physicochemical properties of peat, increasing the pH and amount of humified materials by exposing highly decomposed peat, and it causes nutrient deficiency for the microorganisms by removing inorganic compounds (e.g., P, K, and Na); this, in turn, exerts a response in the methane-cycling microorganisms with consequences for methane production and oxidation (3, 4). Depending on the mining method (block-cut or vacuum-harvested peat) and duration after restoration, these effects can be persistent over prolonged periods (up to 3 decades [4]). The criteria suggested for monitoring peat restoration include surveys of biogeochemical cycles, hydrology, nutritional and chemical status, and microbiological functioning (3, 7). While the response of methane-cycling processes, as well as peat respiration driven by changes in the physicochemical properties and substrate chemistry postharvest and during restoration, have been the focus of previous studies (e.g., references 3, 4, and 8), less is known of the responses of the microorganisms mediating these processes. In particular, the microbial biogeochemistry regulating methane emission during Sphagnum-dominated peat restoration has so far received little attention (9–11).
Ombrotrophic Sphagnum-dominated peatlands are characterized by relatively inhospitable conditions (e.g., the low-pH, nutrient-poor, and antimicrobial properties of Sphagnum [12, 13]). These conditions are in part engineered by Sphagnum spp. and likely favor the proliferation of some methanogens and methanotrophs (13). For instance, methane is primarily produced by hydrogenotrophic methanogenic archaea rather than aceticlastic ones in ombrotrophic peatlands (14–18). These include members of the genera Methanoregulaceae and “Candidatus Methanoflorentaceae,” which are predominant, and, to a lesser extent, Methanosarcinaceae (17, 19). The methanotrophs in natural peatlands have been well documented. Using stable isotope labeling approaches and molecular analyses targeting gene transcripts, the alphaproteobacterial methanotrophs (type II subgroup) have generally been shown to form the vast majority of metabolically active methane oxidizers in many peatlands; gammaproteobacterial methanotrophs (type Ia and Ib subgroups) seem to form a minor fraction of the total active community (12, 20, 21). Ecologically relevant active alphaproteobacterial methanotrophs in peatlands include the Methylocystis-Methylosinus group, as well as members belonging to the family Beijerinckiaceae (Methylocella, Methyloferula, and Methylocapsa), which are phylogenetically distinct from Methylocystis and Methylosinus (12). In particular, Methylocystis and Methylosinus have been suggested to possess ecological traits to survive under unfavorable conditions (22). Moreover, diazotrophic methanotrophs play an important role in peatlands, fixing atmospheric N2 to assimilable nitrogen forms, and they can contribute >33% of new N input in developing peatlands (13, 23, 24). Hence, diazotrophs are also a relevant microbial guild in peat environments. Because of the oligotrophic conditions in ombrotrophic peat, many of these microorganisms are slow-growing and resist isolation in the laboratory. To circumvent cultivation, quantitative and qualitative culture-independent approaches targeting the structural genes of methanogens (i.e., mcrA encoding the α-subunit of the methyl-coenzyme M reductase) and methanotrophs (i.e., pmoA encoding the A subunit of the particulate methane monooxygenase, and mmoX encoding the α-subunit of the soluble methane monooxygenase) have been employed to characterize peat-inhabiting methane-cycling microbial communities (20, 25–27).
Although mining alters the physicochemical properties of peat, with the return of Sphagnum spp. and conditions approximating those of the natural site, we hypothesized that the methanogenic and methanotrophic community compositions will converge, and the corresponding process rates will become more similar in the natural and restored sites than in an actively mined or abandoned site. To address our hypothesis, we determined the potential methane production and oxidation rates in an active peat mine, drained and abandoned mine, and restored mine to elaborate the impact of peat mining on methane turnover. A natural (i.e., undisturbed) peatland dominated by Sphagnum spp. served as a reference for comparison to the other sites. Comparing the natural and abandoned/restored sites thus presents an opportunity to determine the effect of peat mining on methane turnover and the response of the associated microbial communities. We enumerated the structural genes, i.e., the mcrA and pmoA genes, to relate the abundances of the methanogens and methanotrophs to the total methane produced and oxidized, respectively, in the same season over two consecutive years. In addition, we followed the response of the methanogenic and methanotrophic community compositions, respectively, using high-throughput amplicon sequencing of the mcrA gene and a pmoA-based terminal restriction fragment length polymorphism (t-RFLP) complemented by cloning and sequencing of the mmoX gene. Our efforts were aimed to determine the impact of peat mining and restoration on methane turnover and the response of the methane-cycling microorganisms.
RESULTS
The abiotic environment and methane production and oxidation potentials.The peat material generally showed consistent physicochemical properties and methane turnover trends between sites over 2 years. Given that the peatlands were located within the same area, we do not anticipate large temperature or climatic variations between sites that would affect the process rates. Discrepancies in physicochemical parameters and methane turnover rates between sites are thus largely attributable to the impact of peat mining and restoration.
Peat physicochemical properties were largely comparable in 2015 and 2016, but values significantly varied between sites. Peat pH was altered during restoration, showing significantly higher values (P < 0.01; Table 1) in the actively mined (pH ∼3.5) and abandoned (pH ∼4.8) sites compared to the natural and restored sites (pH 2.7 to 3.2). Although more pronounced for total N content, both total N and C content were higher in the abandoned sites, which resulted in a significantly lower C:N ratio in these sites (Table 1). While NOx (total of NO2− and NO3− concentrations) varied between sites, the NH4+ concentration was significantly higher particularly in the actively mined site (Table 1). The PO43− concentration was significantly higher in the restored site in 2015, but because of high variability, the trend was not replicated in 2016 (Table 1). With the exception of NH4+ and PO43− concentrations in the restored site in 2015, the physicochemical properties in the natural and restored sites were comparable. Likewise, the physicochemical properties in the two abandoned sites were not significantly different.
Selected peat physicochemical properties determined in 2015 and 2016a
The potential methane production and oxidation rates were initially determined in the 0- to 5-cm and 5- to 10-cm peat intervals in 2015. With the exception of the methane production rate in the abandoned site since 2004, the methane production and oxidation rates were comparable in these layers (see Fig. S1 in the supplemental material). Therefore, peat from the upper 10 cm was homogenized and incubated as a single layer in 2016. Although values differed, the methane production rate was consistently higher in the restored site sampled in 2015 and 2016. An appreciably higher methane production rate was detected after a lag of 1 day in the restored site in 2015; in 2016, methane production was detected only in this site (Fig. 1). The lower methane production in 2016 may be attributable to the sampling time (June), in conjunction with the beginning of the vegetation growing season (May), as documented before (15). In both the peat material sampled in 2015 and 2016, the methane oxidation rate was significantly higher in the natural and restored sites, while the values in the actively mined and abandoned sites were comparably low (Fig. 1). Methane uptake was immediately detected in the natural and restored sites during incubation, indicating the presence of a readily active population of methanotrophs in these peatlands. A lag of 1 to 2 days was detected in the other peat samples. Sphagnum spp. and peat water exhibited comparable methane oxidation rates in 2015 and 2016, but the values were lower than in the peat material (Fig. S2).
Methane production (A) and uptake rates (B) in peat sampled in 2015 and 2016 (mean ± standard deviation [SD]; n = 4 or 5). Methane production was not detected in the natural peatland, actively mined, and abandoned sites in 2016. These sites were not included in the analysis of variance. Letters indicate the level of significance (P < 0.01) between sites in 2015 (uppercase letters) and 2016 (lowercase letters) in methane production and uptake rates. Note the different scale along the y axis.
mcrA, pmoA, and nifH gene abundances.The mcrA and pmoA genes were enumerated using quantitative PCR (qPCR) assays to be used as proxies for methanogenic and methanotrophic abundances, respectively. Because of the predominance of alphaproteobacterial methanotrophs in peatlands (12), we also targeted the pmoA gene of this group using the type II qPCR assay (Table 2) (28). The mcrA gene abundance was up to four orders of magnitude higher in the natural and restored sites than in the other sites, which showed values comparable to one another (Fig. 2A). The trend was similar, but less pronounced, for the pmoA gene, exhibiting significantly higher gene abundances in the natural, actively mined, and restored sites than in the abandoned sites. However, the pmoA gene abundance in the restored site sampled in 2016 was highly varied, and the values were not significantly different from those of the site that was abandoned since 2004 (Fig. 2B). Although the type II pmoA gene abundances were relatively higher in the peat sampled in 2016, the values were generally comparable across sites, with one exception. In 2015, peat sampled from the restored site harbored a significantly higher abundance of the type II pmoA gene than in the other sites (Fig. 2C). Moreover, the nifH gene was enumerated to monitor the abundance of the diazotrophs during peat restoration. The nifH gene was detected in significantly higher abundance in the natural and restored peatlands than in the abandoned sites (Fig. 2D). Generally, the abundance of the targeted genes showed comparable trends in peat sampled from 2015 and 2016, in that the values were significantly higher in the natural and/or restored peatlands, while the target gene copy numbers were lower in the abandoned sites (Fig. 2).
Primers and PCR thermal profile used in this study
Numerical abundance of the mcrA (A), pmoA (B), type II pmoA (C), and nifH (D) genes in peat sampled in 2015 and 2016 (mean ± SD; n = 6). Two qPCRs were performed for each DNA extract (n = 3), yielding a total of six reactions per target gene, site, and year. The lower detection limit was approximately 103 to 104 copy number of target molecules/g (dw), depending on the qPCR assay. Letters indicate the level of significance (P < 0.01) between sites in 2015 (uppercase letters) and 2016 (lowercase letters) for each target gene.
mcrA and pmoA gene diversity.The methanogenic and methanotrophic community compositions were characterized based on the mcrA and pmoA gene amplicons, respectively. The mcrA gene sequence analysis revealed the presence of 15 operational taxonomic units (OTUs) affiliated with the methanogens; one OTU could not be assigned, but this comprised a minor fraction of the total number of reads in all samples (<0.5%) (Fig. S3). Among the methanogen-affiliated OTUs, the Methanoregula-affiliated OTU was overwhelmingly represented in the natural (∼95% relative abundance) and restored (∼75% relative abundance) sites and comprised ∼45% relative abundance of the total methanogenic community in the actively mined site (Fig. S3). While unclassified Methanomicrobiales- and Methanosarcina-affiliated OTUs were predominant in the abandoned site since 2009 (∼91% relative abundance), these OTUs, along with Methanomassiliicoccus- and Methanolinea-affiliated OTUs, form the majority of sequences detected in the site that was abandoned since 2004 (∼83% relative abundance; Fig. S3). The discrepancy in the methanogenic community composition between sites was visualized as a correspondence analysis using the environmental variables (NH4+, PO43−, and NOx concentrations and pH) as constraints (Fig. 3), with pH being the only parameter significantly (P < 0.01) affecting the community composition in all sites. The methanogenic community composition in the two abandoned sites grouped together and could be clearly distinguished from the community composition in the natural and restored peatlands, which separated along constrained correspondence analysis (CCA) axis 1 (Fig. 3). The community composition in the natural and restored sites clustered closely together but did not overlap. There appears to be a shift in the methanogenic community composition after peat mining and abandonment and the return to a community structure resembling the natural peatland after restoration.
Correspondence analysis showing the response of the methanogenic (A) and methanotrophic (B) community composition to peat abandonment and restoration postharvest. The methanogenic and methanotrophic community compositions were derived from Illumina MiSeq amplicon sequencing of the mcrA gene and a pmoA-based t-RFLP analysis, respectively. The analysis was performed for each DNA extract (n = 3) of the starting material sampled in 2015 per site. The distribution and affiliation of the OTUs and t-RFs for the mcrA and pmoA genes are given, respectively, in Fig. S3. Green triangle, natural peatland; dark-blue diamond, actively mined site; light-blue inverted triangle, abandoned site since 2009; gray square, abandoned site since 2004; red square, restored site.
Although we were able to retrieve pmoA amplicons in the first PCR cycle, reamplification of an aliquot of the first PCR with barcoded primer was unsuccessful, even after optimization steps. Therefore, we characterized the diversity of the pmoA gene using t-RFLP. The assignment of the t-RFs was based on comparison to a comprehensive pmoA clone library (>5,000 clones [29]). After standardization and then normalization of the t-RFLP profiles to the overall signal intensity as described by Lüke et al. (29, 30), the major identifiable t-RFs were t-RF 349 (affiliated with type Ia methanotrophs and pmoA2), t-RF 514 (affiliated with the type Ia methanotroph Methylobacter), t-RF 240 (affiliated with type Ia methanotrophs), t-RF 225 (affiliated with type Ib methanotrophs), t-RF 74 (affiliated with the type Ib methanotrophs Methylocaldum and Methylococcus), and t-RF 245 (affiliated with type II methanotrophs). The t-RFLP analysis revealed a predominance of t-RF 245 affiliated with type II methanotrophs in all sites (40 to 50% of total community composition) and was most pronounced in the restored site and the site abandoned since 2004, comprising more than 80% relative abundance (Fig. S3). Differences in the methanotrophic community compositions were further determined by a correspondence analysis using measured environmental variables (NH4+, PO43−, and NOx concentrations and pH) as constraints, which were not significantly correlated to the community composition (P > 0.01), together explaining only 20.3% of the total variance (Fig. 3). Because of the large variability between and within sites, no clear separation of the community composition was discernible (Fig. 3). However, the community composition in the natural site and the site abandoned since 2009 tended to separate along CCA axis 1 (Fig. 3).
As some methanotrophs possess only the mmoX gene (e.g., Methylocella and Methyloferula [12]), we determined the composition of the mmoX-harboring methanotrophs using cloning and sequence analyses. Besides Methylocella-like and Methyloferula-like methanotrophs, the mmoX gene sequences detected were affiliated with Methylocystis and Methylosinus, together forming the majority of alphaproteobacterial methanotrophs in all sites (Fig. S4). Hence, the t-RFLP and the cloning and sequence analyses of the mmoX gene were in agreement with the qPCR analysis, which showed an abundance of the type II pmoA gene.
DISCUSSION
Methane dynamics and the abiotic environment postharvest.The surface peat layer is a dynamic area where rapid changes in physicochemical conditions occur (31), significantly affecting the methane production and oxidation rates, as well as respiration (8, 17, 32). The methanotrophs localized at the surface peat, as well as in Sphagnum spp., and planktonic methanotrophs in the overlaying water play a crucial role as a methane biofilter, consuming methane before release into the atmosphere (4, 5). However, methane can escape via ebullition, bypassing the methanotrophs. Edaphic factors controlling methane production were found to be significantly correlated with methane oxidation, suggesting that methane oxidation was primarily fueled by substrate availability and was dependent on methanogenesis (3, 4, 17). Therefore, sites producing methane are anticipated to show proportional methane-oxidizing potential. This appeared to be the case in the actively mined, abandoned, and restored sites; in the natural peatland, low methane production did not correspond with the significantly higher methane oxidation rate detected (Fig. 1). The physicochemical properties of the natural peatland and restored site were largely similar (Table 1); hence, disproportional methane production and oxidation activities in the natural site are likely caused by differences in the biotic components, e.g., methanotrophic community composition and/or abundance (see below). Moreover, we cannot exclude seasonal variation affecting the methane production and oxidation rates caused by fluctuations of the water table layer. Nevertheless, we showed that trends in methane turnover potential were at least consistent during the summer over 2 years.
In the actively mined and abandoned sites, methane production was low or was not detected postharvest. Mining and abandonment significantly reduce substrate quality and bioavailability by exposing the humified fraction and already-decomposed peat material (3, 8). Not surprisingly, methane turnover in these sites was significantly lower than in the restored sites, where Sphagnum spp. naturally revegetated more recently than in the natural site, which provides a supply of readily metabolized organic substrate (Fig. 1). Besides affecting the availability of organic substrates, mining also removes inorganic nutrients (e.g., Mg, Na, and P), causing nutrient limitation, which appears to affect the rates of processes dependent on the nutrients (e.g., Na limiting methane production [4, 33]). Substrate and micronutrient deficiencies caused by mining, coupled to the drier condition and hence better aeration (gravimetric water content, 68 to 81%, as opposed to 89 to 91% in the natural and restored sites) in the surface peat after drainage, likely attributed to the significantly lower methane production potential in the actively mined and abandoned sites (Fig. 1). The abandoned sites were also characterized by significantly higher pH and total carbon and nitrogen concentrations (Table 1), with the markedly higher total nitrogen content decreasing the C:N ratio, indicating readily decomposable material (34). However, the use of the C:N ratio as a proxy for decomposability is disputable (35), and the total C and N concentrations determined here also include the recalcitrant fractions not bioavailable to the microorganisms. Therefore, mining alters the physicochemical characteristics of peat, but methane-cycling processes were restored to even higher values than in the natural peatland after 15 years.
Response of the methane-cycling microbial community abundance to mining and restoration.Peat mining constrains the total microbial biomass by reducing the bioavailability of substrate and nutrients (3, 4, 36). In these studies, the total microbial biomass significantly decreased in the actively mined or abandoned sites compared to the natural or restored sites. Although the total microbial biomass could be correlated to the carbon dioxide emission (3, 4), the methods used to quantify microbial biomass (e.g., fumigation-extraction technique) were unable to discriminate between microbial groups catalyzing methane production and oxidation from the total community. Applying qPCR assays specifically targeting the mcrA and pmoA genes, we enumerated the abundances of the methanogens and methanotrophs, respectively. Assuming that one methanogen harbors a single mcrA gene copy (37) and that a methanotroph harbors two pmoA gene copies (38), mining dramatically decreased the methanogenic population size by around an order of magnitude, and to a lesser extent in the methanotrophic population (1 × 106 to 1 × 107 cells/g [dry weight {dw}]) depending on the sampling year; (Fig. 2) compared to the natural peatland. Although the changes in the pmoA gene copy numbers were statistically significant, methanotrophs appeared to be less sensitive to mining and abandonment, with the alphaproteobacterial (type II) methanotrophic population seemingly increased after abandonment in 2015 (Fig. 2). Regardless, both methanogenic and methanotrophic populations recovered well in the restored site, showing comparable or even higher mcrA or pmoA gene copy numbers than in the natural peatland. Similarly, the diazotrophic abundance, indicated by the nifH gene, significantly decreased after mining, but the population recovered in the restored site. The diazotrophs have been recognized as a key microbial group in ombrotrophic peatlands, being an important source of assimilable carbon and nitrogen to sustain Sphagnum growth (23, 39, 40). Hence, recovery of the diazotrophic abundance is a relevant indicator for Sphagnum revegetation. Overall, more pronounced for the methanogens, mining compromised the methanogenic, methanotrophic, and diazotrophic abundances, but population size fully recovered after restoration.
Structural genes appear to correlate well to the activity catalyzed by the respective encoded enzymes, as has been shown for methanogenesis and methane oxidation (41–44), as well as other microbe-mediated processes, e.g., 2-methyl-4-chlorophenoxyacetic acid, an herbicide, degradation (45), and N-cycling processes (46). Determining the gene abundances before and after incubation, we relate the magnitude change in the mcrA and pmoA genes to the total methane produced and oxidized, respectively (Fig. 4). Integrating all replicate from all sites showing activity, a highly positive significant correlation (P < 0.01) was found for the magnitude of change in the mcrA gene abundance and methane production, suggesting a coupling of methanogenic growth and activity (Fig. 4). Although the magnitude change in the total pmoA gene abundance was not significantly correlated (P > 0.01; Fig. 4) to the total methane oxidized during incubation, a significantly positive correlation (P < 0.01) was detected between the magnitude of change in type II pmoA gene abundance and methane oxidized (Fig. 4). The significant correlation between methane oxidation and type II pmoA gene abundance, rather than the total pmoA gene abundance, suggests the active role of type II methanotrophs under the incubation conditions. Alphaproteobacterial methanotrophs belonging to the Methylocystis-Methylosinus group, as well as the genera Methylocella, Methylocapsa, and Methyloferula, seemingly form the predominantly active community in many acidic peat environments (25, 27, 47), although the presence of active gammaproteobacterial methanotrophs cannot be excluded (32, 48).
Correlation between total methane produced (A) and consumed (B and C) and the magnitude change in the mcrA and pmoA gene abundances, respectively, during incubation (7 days). Replicates from all sites showing methane production and oxidation were incorporated into the correlation. The magnitude change in the mcrA and type II pmoA gene abundances was significantly correlated (P < 0.01) to the total methane produced and oxidized, respectively. Note the different scales on the x and y axes.
Response of the methane-cycling microbial community composition to mining and restoration.Consistent with previous studies, we detected a predominance of Methanoregula, a genus represented by hydrogenotrophic methanogens which showed a preference for acidic conditions, in the natural and restored peatlands (11, 17–19). Methanoregula-like methanogens decreased with mining and abandonment, while the relative abundances of Methanosaeta-like and Methanosarcina-like methanogens increased. Methanosaeta and Methanosarcina are closely related and belong to Methanosarcinales, with Methanosaeta comprising aceticlastic methanogens, while members of Methanosarcina (e.g., M. barkeri) show metabolic flexibility, being capable of hydrogenotrophic and aceticlastic methanogenesis (49, 50). The apparent shift from a predominance of hydrogenotrophic to aceticlastic methanogens suggests a physiological change presumably driven by nutritional status, including substrate quality and availability, and vegetation type following peat mining. However, the community compositions of the methanogens in the natural and restored peatlands became similar, with Methanoregula-like methanogens dominating the community (Fig. 3), suggesting that recovery of the methanogenic community composition occurs with Sphagnum revegetation.
The methanotrophic community composition was not statistically distinct between sites, but a higher mean relative abundance of alphaproteobacterial methanotrophs was detected in the restored and abandoned sites since 2004. Despite comparable methane oxidation rates and total methanotroph abundances in the natural and restored sites, peat abandonment and restoration may have structured the methanotrophic community composition over time (Fig. 3 and S3). In contrast, the methanotrophic community composition was found to be similar in a natural peatland and a site undergoing 10 to 12 years of restoration in a forestry-drained peat (9). However, it should be noted that the authors targeted the pmoA gene using a rather specific primer pair (A189f/A621r [51]), limiting the pmoA gene coverage exclusively to alphaproteobacterial methanotrophs. Regardless, as shown by the t-RFLP analysis and supported by the qPCR assays, all sites harbored predominantly alphaproteobacterial methanotrophs, which are thought to be the active population in widespread peat environments (13).
Conclusion.Supporting our hypothesis, the results showed that methane turnover potentials and the associated microbial abundance recovered postharvest with the return of Sphagnum spp., but we found less compelling evidence for the convergence of the methane-cycling microbial community compositions in the natural and restored sites. A shift in the microbial community composition may alter the collective traits of the contemporary community (52–54), with implications for their response to future disturbances. Our findings could be further substantiated by field-based flux measurements in future studies.
MATERIALS AND METHODS
Site description and sampling.The sampling sites are regarded as ombrotrophic peatlands, as water and nutrients are rain-fed, and the sites are isolated from the groundwater (Table 3). The abandoned sites were dammed postharvest since 2004 and 2009, but water naturally drained from these sites, and vegetation shifted toward shrub land/forest (Table 3). The restored site was dammed and remained water-logged since 2000. Like in the natural site, Sphagnum spp. predominate in the restored site. Besides damming to prevent water outflow, no other restoration intervention was introduced. The natural peatland, considered the reference site, was an undisturbed site and was declared a nature reserve since 1962. In addition, peat was also sampled from a drained actively mined site managed by Greenyard Horticulture, Poland. Peat was extracted close to the water table in the actively mined, abandoned, and restored sites using the “block peat” or surface milling method (Table 3).
Description of sampling sitesa
Peat material was collected in mid-August 2015 and mid-June 2016 at the same sampling sites. The atmospheric temperatures recorded at the time of sampling were in the range of 30°C to 35°C and 27°C to 32°C in 2015 and 2016, respectively. Peat material was sampled from the surface peat layer (0 to 10 cm) from four or five randomly selected plots in each site. The plots were spaced >2 m apart and were considered to be independent from one another. Three cores (3.5 cm diameter, 10 cm high) were collected from each plot and composited after separating the upper 0 to 5 cm from the subsequent 5 to 10 cm in 2015. In 2016, the cores were composited without distinguishing the 0- to 5- and 5- to 10-cm layers. During sampling, vegetation was collected from ∼1 m by 1 m grids from which the samples were collected and delivered to the laboratory to be identified. In the natural and restored sites where Sphagnum spp. predominate, Sphagnum spp. and the overlaying water were sampled. Water from an adjacent ditch to the actively mined site was also collected. All samples were transported to the laboratory in Styrofoam boxes containing cooling blocks. Samples were immediately processed upon arrival for incubation (in 2015) or kept in the 4°C fridge for 2 days before incubation (in 2016).
Microcosm setup to determine methane production and uptake rates.In the laboratory, roots, stones, and other debris were removed by hand from the samples before incubation. Each microcosm consisted of 7.5 g (fresh weight) of peat in a 120-ml bottle capped with a butyl rubber stopper. To determine the methane production rate, the microcosm was flushed with N2 for 30 min before incubation. To determine the methane uptake rate, methane was added to the microcosm to achieve a headspace concentration of approximately 1 to 2% (vol/vol). The microcosm was shaken (120 rpm) during incubation at 25°C for 7 days. In parallel, approximately 5 g of the same material was oven-dried at 65°C until the weight remained constant (approximately 2 weeks) to determine the gravimetric water content, which was later used to normalize methane production and uptake rates to per gram (dry weight) of material.
Sphagnum spp. were rinsed with deionized water and left on a paper towel to drain excess water before incubation. Sphagnum spp. (0.25 to 0.30 g fresh weight) were placed in a 120-ml bottle, and the headspace methane was adjusted to approximately 1% (vol/vol). Incubation was performed under oxic conditions on a shaker (120 rpm) next to a window receiving natural light in the laboratory at room temperature (22°C to 24°C). As such, light intensity may be suboptimal during incubation. A portion of the same Sphagnum sp. was oven-dried at 65°C until the weight remained constant (approximately 2 weeks) to determine the gravimetric water content, which was used to normalize methane uptake rate. The peat water fraction (5 ml) was incubated oxically under an initial headspace methane of approximately 1% (vol/vol) in a 120-ml bottle. Incubation was performed at 25°C while shaking (120 rpm) in the dark. In the peat incubation, the headspace methane concentration was monitored daily for 7 days, while in the Sphagnum and peat water incubations, headspace methane was followed over 10 to 12 days. After incubation, samples were collected and stored in the −20°C freezer until further analysis.
Headspace methane measurement and soil nutrient content.Headspace methane content was determined using an Ultra gas chromatograph (GC; Interscience, Breda, the Netherlands) equipped with a flame ionization detector (FID). Methane production and uptake rates were determined linearly from the methane increase and decrease during incubation, respectively. In particular, the rate of methane production was determined after a lag period (1 to 3 days), which may not reflect the in situ production rate. Methane was consumed immediately (<1-day lag period) during oxic incubation. Peat pH was determined in 1 M KCl (1:5). The nutrient (NOx, NH4+, and PO43−) contents in the peat were determined using a Seal QuAAtro segmented flow analyzer (SFA) (Beun-de Ronde B.V., Abcoude, the Netherlands) in 1 M KCl (1:5), as described before (55). NOx is the total of NO2− and NO3−. The total C and N content was determined using the Flash EA 1112 CN analyzer (Thermo Fisher Scientific, Breda, the Netherlands) after processing the peat, as described before (35).
DNA extraction and qPCR assays.Peat was randomly selected from three plots per site (starting material) and after oxic and anoxic incubation of the same starting material for DNA extraction. DNA was extracted using the PowerSoil DNA isolation kit (Mo Bio, Uden, the Netherlands), according to the manufacturer's instructions.
We performed qPCR assays targeting the mcrA, pmoA, and nifH (encoding the H subunit of the nitrogenase) genes as proxies to enumerate the abundance of the methanogens, methanotrophs, and diazotrophs (nitrogen fixers), respectively, in the starting material. In addition, we performed a qPCR assay specifically targeting the pmoA gene of type II methanotrophs. The mcrA, pmoA, and type II pmoA qPCR assays were also applied to the peat after the incubations. The qPCR assays were performed in duplicate per DNA extract, yielding a total of six replicates per target gene, site, year, and time (before and after incubation). The primers, primer concentration, and PCR thermal profiles for each assay are given in Table 2. The qPCRs for all assays consisted of 10 μl of 2× SensiFAST SYBR (Bioline, Alphen aan den Rijn, the Netherlands), 3.5 μl each of forward and reverse primers, 1 μl of bovine serum albumin (5 mg · ml−1; Invitrogen, Breda, the Netherlands), and 2 μl of diluted template DNA, with an exception; in the qPCR assay targeting the nifH gene, 2.5 μl each of forward and reverse primers was added into the reaction. DNase- and RNase-free water (Sigma-Aldrich Chemie NV, Zwijndrecht, the Netherlands) was added to reach a final volume of 20 μl, if needed. In a pilot qPCR run, 10-, 50-, and 100-fold dilutions of template DNA extracts from each site were performed to determine the DNA concentration yielding the optimal copy numbers of the target genes. Subsequently, the DNA extract was diluted 10-fold and 100-fold for the qPCR assays targeting the nifH and type II pmoA genes and the pmoA gene, respectively. The DNA extract was diluted 10-fold (natural site, actively mined site, and site abandoned since 2009), 50-fold (site abandoned since 2004), and 100-fold (restored site) for the qPCR assay targeting the mcrA gene. The addition of bovine serum albumin (BSA) and dilution of template DNA were performed to relieve PCR inhibition. Plasmid DNA isolated from pure cultures was used as a standard for the calibration curve. The qPCR assays were performed using a Rotor-Gene Q real-time PCR cycler (Qiagen, Venlo, the Netherlands). The specificity of the amplicon was verified from the melt curve and further confirmed by 1% gel electrophoresis, showing a single band of the correct size in the pilot qPCR run.
mcrA gene amplification and sequencing.Amplification of the mcrA gene for sequencing, targeted using the mlas/mcrA-rev primer pair, was performed according to the method of Herbold et al. (56), using a two-step barcoding approach. In the first PCR, the target gene was amplified with diagnostic primers synthesized with a 16-bp head sequence (5′-GCTATGCGCGAGCTGC-3′). In the second PCR, the amplicon from the first PCR was reamplified with primers that consist of the 16-bp head sequence and included at the 5′ end, a library-specific 8-bp barcode. Each PCR (total volume, 20 μl in the first PCR and 50 μl in the second PCR) consisted of 1× Taq buffer with 1.8 mM MgCl (Roche Diagnostics, Almere, the Netherlands), 0.2 mM dinucleoside triphosphate (dNTP) mix (Roche Diagnostics), 0.025 U Taq DNA polymerase (Roche Diagnostics), 0.2 mg/ml bovine serum albumin (Invitrogen, Breda, the Netherlands), 1 μM each forward and reverse primers, and 1 μl of template. When needed, DNase- and RNase-free water (Sigma-Aldrich Chemie NV, Zwijndrecht, the Netherlands) was added to achieve the total volume. The PCR thermal profile included an initial denaturation step at 95°C for 5 min, followed by 40 cycles (for the first PCR) or 15 cycles (for the second PCR) of 95°C for 1 min, 60°C (for first PCR) or 53°C (for second PCR) for 1 min, and 72°C for 1 min. The final elongation step was 72°C for 5 min. The first PCR was performed in duplicate, screened by 1% gel electrophoresis, and pooled for use as the template in the second PCR, which used only one primer (5′-BARCODE-HEAD-3′). The amplicon from the second PCR was also screened by 1% gel electrophoresis. The PCR product was purified using the QIAquick purification kit (Qiagen, Venlo, the Netherlands), and the concentration was determined using the Fragment Analyzer (Advanced Analytical, Heidelberg, Germany). Equal amounts of the purified PCR products were pooled (10 ng/μl) prior to sequencing. High-throughput sequencing was performed by LGC Genomics (Berlin, Germany) using Illumina MiSeq version 3 chemistry generating 300-bp paired-end reads.
pmoA-based t-RFLP.Detailed methodology for the pmoA-based t-RFLP has been described elsewhere (57). Briefly, the pmoA gene was amplified from each DNA extract (n = 3) from peat sampled in 2015 with the primer set A189f/Mmb661r. The forward primer A189f was labeled with 6-carboxyfluorescein (FAM). The amplicon was digested using the restriction endonuclease MspI, and the t-RFs were separated using the ABI Prism 310 (Applied Biosystems, Darmstadt, Germany). The GeneScan 3.71 software (Applied Biosystems) was used to determine the length of the t-RFs by comparison with an internal standard (MapMarker 1000; BioVentures, Murfreesboro, TN, USA). In silico comparative gene sequence analysis with an extensive pmoA clone library was performed to bin and assign the t-RFs as described in detail before (29, 30, 57). Unidentifiable t-RFs (i.e., t-RFs 69, 242, and 143) and t-RFs comprising <2% of the total peak area were excluded from the t-RFLP analysis.
Cloning and sequencing of the mmoX gene.The mmoX clone libraries were constructed from the DNA extract of all peat material sampled in 2015 using the primer pair mmoX-206F/mmoX-886R, with the primer concentrations and PCR thermal profile as shown in Table 2, according to Liebner and Svenning (27). The PCR mixture consisted of 25 μl of MasterAmp 2× PCR premix F (Epicentre, Illumina, WI, USA), 5 μl of bovine serum albumin (BSA; 5 mg · ml−1; Invitrogen, Breda, the Netherlands), 0.5 μl of each primer, 0.5 μl of Taq polymerase (Thermo Fisher Scientific, Uden, the Netherlands), 1 μl of template DNA, and 17.5 μl of DNase- and RNase-free water (Sigma-Aldrich Chemie NV, Zwijndrecht, the Netherlands), giving a final volume of 50 μl. The PCR amplicon was verified on 1% agarose gel electrophoresis. Three PCRs were performed for each DNA extract and subsequently pooled per site during the PCR cleanup step using the GenElute gel extraction kit (Sigma-Aldrich Chemie NV). The purified PCR amplicon was cloned into the pGEM-T vector (Promega, Mannheim, Germany) and transformed into E. coli JM109 competent cells (Promega). The transformants were screened (blue-white colonies) and confirmed by colony PCR using the primers T7/M13rev_29 before sequencing. Sequencing was performed at ADIS (Max Planck Institute [MPI] for Plant Breeding, Cologne, Germany). Sequences were assembled using the SeqMan software (DNAStar software package; Lasergene, USA), whereby the vector sequence was trimmed. Sequences were compared and identified against the GenBank database using the NCBI BLASTn function.
Statistical analysis.Significant differences (P < 0.01) in the physicochemical parameters (n = 3) and process rates between sites per year (n = 4 or 5) were determined by analysis of variance (ANOVA) using SigmaPlot version 13.0 (Systat Software, Inc., USA).
The mcrA sequencing reads were assembled using the “make.contigs” command in Mothur version 1.3.3 (58). The assembled contigs were then sorted, and the barcodes and primers were removed. Putative chimeric reads were removed using USEARCH 6.0, the nucleotide sequences were translated into amino acids, and frameshifting reads were corrected using the FRAMEBOT program in the FunGene Pipeline, RDP (59) (http://fungene.cme.msu.edu/FunGenePipeline). The inferred amino acid sequences were aligned and clustered into operational taxonomic units (OTUs), with an 89% identity cutoff value, which has been considered to be the species-level affiliation for the methanogen (60). Representative mcrA nucleotide sequences were assigned to their taxonomic affiliations by BLASTn comparisons to the GenBank nonredundant (nr) database, and the results were then imported to the MEGAN software version 5.11.3 with the lowest common ancestor (LCA) and default parameters. As with the t-RFs of the pmoA gene (after normalization to the overall signal intensity), constrained correspondence analysis (CCA) for the mrcA gene sequences was produced in the R statistics software environment (61) using the package Vegan version 2.3.0 (62). The measured environmental parameters (NH4+, NOx, and PO43− concentrations and pH) were used as constraints in the CCA.
Accession number(s).The mcrA and mmoX gene sequences were deposited at the EMBL European Nucleotide Archive (ENA) under the project accession numbers PRJEB22766 and PRJEB22776, respectively.
ACKNOWLEDGMENTS
We are grateful to Iris Chardon and Marion Meima-Franke for excellent technical assistance. We thank Marcin Harnisz (Greenyard Horticulture, Poland) for his invaluable help during sampling. A.H. was financially supported by BE-Basic grant F03.001 (SURE/SUPPORT).
This is publication no. 6423 of the Netherlands Institute of Ecology.
All authors have seen and approved the final version submitted.
We declare no conflicts of interest.
FOOTNOTES
- Received 6 October 2017.
- Accepted 16 November 2017.
- Accepted manuscript posted online 27 November 2017.
Supplemental material for this article may be found at https://doi.org/10.1128/AEM.02218-17.
- Copyright © 2018 American Society for Microbiology.
