Introduction

Methane (CH4) is the second most important anthropogenic greenhouse gas and is about 28 times more efficient at absorbing infrared radiation than CO2 (Stocker et al. 2013). Methane emissions can be attenuated by aerobic methane-oxidizing bacteria known as methanotrophs. These bacteria have the ability to utilize methane as their source of C and energy. Methanotrophs can act either as a bio-filter in wetland agricultural soils (paddy field) or as a methane sink in well-aerated soils (Bridgham et al. 2013; Fan et al. 2019a, b; Kolb 2009; Wei et al. 2019). Although methane can also be oxidized anaerobically by microbes using nitrite, nitrate, ferric iron, and sulfate as electron acceptors, methanotrophs are ubiquitous in nature and represent the only oxic biogenic sink for the greenhouse gas methane (Fan et al. 2019b; Hanson and Hanson 1996; Shen et al. 2019; Trotsenko and Murrell 2008).

Historically, aerobic methanotrophs have been classified as types I, II, and X, based on physiological, biochemical, and phenotypical characteristics (Veraart et al. 2015). Nowadays, they are preferentially classified based on phylogeny, either belonging to the classes of Gammaproteobacteria (referred to as type I or X methanotrophs), Alphaproteobacteria (referred to as type II methanotrophs), or to the phyla Verrucomicrobia (referred to as type III methanotrophs) and NC10 (Knief 2015; Stein et al. 2012). More specifically, type I methanotrophs belong to the Methylococcaceae (type Ia and type Ib) and Methylothermaceae (type Ic) families, while type II methanotrophs are divided into type IIa (Methylocystaceae) and type IIb (Beijerinckiaceae) (Lüke and Frenzel 2011). Converting CH4 to methanol is the key step in the CH4 oxidation pathway, which is catalyzed by the enzyme methane monooxygenase (Hanson and Hanson 1996). The pmoA gene encodes the β-subunit of the particulate methane monooxygenase enzyme (pMMO) and is the most commonly used functional marker for identifying methanotrophs in environmental samples (Dumont 2014).

Grasslands are one of the largest terrestrial ecosystems in the world, covering approximately 20–40% of the earth’s surface and accounting for up to 40% of the total land area in China (Nan 2005). Grassland soils are one of the largest terrestrial methane sinks with the capacity for methane oxidation (Pachauri et al. 2014). As the most popular management of grasslands, animal grazing may lead to different soil physical conditions (e.g., bulk density and aeration status) and chemical properties (e.g., pH and organic matter content) (Li et al. 2008; Steffens et al. 2008). Soil bulk density and moisture are considered to be critical factors for CH4 oxidation activity in soil (Serrano-Silva et al. 2014). Animal grazing alters methanotroph composition and methane uptake activities by reducing soil water content and soil aeration through herbage removing and tramping, respectively (Leriche et al. 2001; Zhou et al. 2008). Heavy grazing has been reported to significantly reduce annual CH4 uptake, while light-to-moderate grazing has been shown to have either a considerable positive impact on CH4 uptake or to not significantly change CH4 uptake (Chen et al. 2011; Ma et al. 2018). Previous studies either measured the abundance and/or composition of methanotroph in grassland soils or try to link CH4 uptake with the abundance of the pmoA gene. However, mere presence of pmoA genes under in situ soil conditions may not necessarily reflect the functional activity of CH4 oxidation (Nannipieri et al. 2019). To our knowledge, there are little studies that link CH4 oxidation with functional active methanotrophs in grazed grassland soils.

The objective of this study was to obtain information on the presence of aerobic methanotrophs affected by grazing gradients in the field and to identify the active microorganisms responsible for CH4 oxidation in a typical steppe grazed soils. We applied quantitative PCR (qPCR) of the pmoA gene and Miseq sequencing coupled with DNA-based stable isotope probing (SIP) to study the active methanotroph communities from field soils. Based on previous studies, we hypothesized that different levels of grazing would alter the abundance and communities of active methanotrophs and change the methane oxidation activity.

Materials and methods

Setup of the experiment

Four different stocking rates were established in 2005, with zero, three, nine, and fifteen sheep per two hectares, at the Inner Mongolia Grassland Ecosystem Research Station (IMGERS, 43° 37′ N, 116° 43′ E) on the Xilingol steppe of the Xilin River basin. The dominant plant species in the study area is Leymus chinensis with some Stipa grandis and Cleistogenes squarrosa. The four stocking rates were classified as ungrazed control (CK), light grazing (G1), moderate grazing (G2), and heavy grazing (G3). Each year, sheep are in the field from June to September (~ 95 days), in accordance with the local summer grazing season. The different stocking rate treatments were arranged in a randomized block design with three replicates separated by fences, and the plot size for each block was 2 ha. Composite soil samples of each replicate were collected from the upper 10-cm layer from 5 random locations using a 5 cm in diameter soil auger in August 2015 and transported to the laboratory with a cold chain. After passing through a 2-mm sieve, soil samples of each replicate were separated into three subsamples for DNA extraction, physicochemical property analysis, and SIP incubation experiments, respectively. Soil physicochemical properties were determined according to the protocols of the Handbook of Soil Analysis (Pansu and Gautheyrou 2007). The soil type is dark chestnut (calcic Chernozem according to the ISSS Working Group RB (1998)).

Microcosm incubations for SIP and CH4 oxidation potential were conducted in sterile 120-mL glass serum vials in triplicate, containing 10 g (dry weight) of soils, and sealed with butyl rubber stoppers. Each microcosm was injected with either 13CH4 (99 atom % 13C, Sigma-Aldrich, USA) or CH4 (99.5% pure, as control) to a final concentration of 8% (v/v) in the headspace and incubated at 25 °C in dark conditions for 28 days. CH4 oxidation potential was assessed from the zero-order decrease in CH4 concentration in the headspace of the serum vials within 12 h and measured with gas chromatography (Shimadzu GC12-A, Japan) (Kightley et al. 1995). After more than 90% of the CH4 was consumed, the headspace was flushed with pressurized synthetic air (20% O2, 80% N2) for 1 min to maintain oxic conditions and, after that, the labelled or unlabelled CH4 was renewed. The atmosphere of the microcosms was renewed three times with pressurized synthetic air and labelled or unlabelled CH4 during the incubation period. Destructive sampling was performed in triplicate for further soil analysis after incubation of SIP microcosms for 28 days.

Nucleic acid extraction and SIP fractionation

DNA was carefully extracted from 0.5 g of soil using a FastDNA SPIN kit for soil (MP Biomedicals; Solon, OH, USA) according to the manufacturer’s instructions. Negative control without soil was not tested as this study focuses on the active methanotrophs which were labelled with 13C (Vestergaard et al. 2017). The concentrations and quality of DNA were estimated by a Nanodrop® ND-2000 UV-Vis spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA) and gel electrophoresis.

Density gradient centrifugation was performed on bulk DNA extracted from the 13CH4 and CH4 treatments as described by Liu et al. (2019a) and Zhang et al. (2019), with a minor modification in scale: 2.5 μg of DNA were mixed with gradient buffer/CsCl solution in a 5.1-mL Beckman ultracentrifuge tube and DNA was fractionated into 14 equal fractions after centrifuged at 177,000gav for 44 h at 20 °C in a Vti65.2 vertical rotor (Beckman Coulter, Palo Alto, CA, USA) (Jia et al. 2019). The fractionated DNA was purified and dissolved in 30 μL of TE buffer.

Quantitative PCR of the pmoA genes

Quantitative PCR analysis of the pmoA gene in bulk DNA and in each DNA gradient fraction was performed to determine the growth and efficiency of 13C incorporation into the genomic DNA of methanotroph communities, respectively. The primer pair A189f and mb661r (Costello and Lidstrom 1999; Holmes et al. 1995) was used for the qPCR of the pmoA gene as described previously (Zheng et al. 2014). qPCR runs were carried out in a LightCycler® 480II (Roche, Germany). Efficiencies of 89–105% were obtained for all gene amplifications, with R2 values ranging between 0.992 and 0.999.

MiSeq sequencing and phylogenetic analysis

Methanotroph-affiliated 16S rRNA gene was amplified in both 13CH4 and CH4-control microcosms in bulk DNA and in heavy fractions of each gradient with standard PCR conditions and quantified DNA as suggested (Schöler et al. 2017). Methanotroph-affiliated 16S rRNA libraries were constructed and sequenced using an Illumina® MiSeq sequencer (Illumina, San Diego, CA, USA) by Majorbio Bio-pharm Technology Co., Ltd. (Shanghai, China) with a universal 515F-907R primer assay as reported previously (Daebeler et al. 2014; Zheng et al. 2014). Raw fastq files were quality-filtered by Trimmomatic and merged by FLASH with the following criteria: (i) The reads were truncated at any site receiving an average quality score < 20 over a 50-bp sliding window. (ii) Sequences with overlap being longer than 10 bp were merged according to their overlap with mismatch no more than 2 bp. (iii) Sequences of each sample were separated according to barcodes (exactly matching) and primers (allowing 2 nucleotide mismatching), and reads containing ambiguous bases were removed. We obtained a total of 1,809,594 high-quality sequences with an average of 37,940 for each sample. Rarefying may bring out some problems but is still a popular method in the study of microbial ecology (Delgado-Baquerizo et al. 2018). To avoid potential bias caused by sequencing depth, all sequence data were rarefied to 17,454 sequences per sample for the downstream analyses. Operational taxonomic units (OTUs) were clustered with 97% similarity cutoff using UPARSE (version 7.1 http://drive5.com/uparse/) with a novel “greedy” algorithm that performs chimera filtering and OTU clustering simultaneously. The taxonomy of each 16S rRNA gene sequence was analyzed by the RDP Classifier algorithm (http://rdp.cme.msu.edu/) against the Silva (SSU123) 16S rRNA database using a confidence threshold of 97%.

Distribution of the methanotroph communities in situ and in the incubation experiment among different grazing intensities was evaluated by principal component analysis (PCA) in R using the vegan package. Redundancy analysis (RDA) was also performed to identify the abiotic factors (bulk density, total C, Olsen P, and NO3−-N) that are most important in shaping active methanotroph communities in the grazed grassland soils. The neighbor-joining tree was constructed by MEGA 7 with 1000-fold bootstrap support (Kumar et al. 2016).

Statistical analysis and sequencing data deposition

Significant differences of CH4 oxidation potential, pmoA, and methanotroph-affiliated 16S rRNA genes relative abundance among different treatments were assessed by one-way analysis of variance (ANOVA) followed by Tukey’s post hoc test. All analyses were conducted by SPSS version 20 (IBM Co., Armonk, NY, USA).

The reads for the 16S rRNA genes of the in situ and incubated soil samples were deposited in the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) under accession number PRJNA432864.

Results

Soil physicochemical properties

The general physicochemical characteristics for the soils from different stocking rates (no grazing (CK), light grazing (G1), moderate grazing (G2), and heavy grazing (G3)) are summarized in Table 1. Soil pH was significantly lower in the G2 (6.89) soil compared with other grazed soils (G1: 8.12; G3: 7.69) and the ungrazed control (CK: 7.62). Soil bulk density showed a significant positive relationship with increasing stocking rates (r = 0.96, p < 0.001). The contents of total soil C and nutrients, including SOM, total C, total N, Olsen P, and K, were highest in the G1 soils compared with other samples. Grazing led to significantly less exchangeable NH4+-N in the grazed soils of G1, G2, and G3, compared with the ungrazed control.

Table 1 Physicochemical properties of the grassland soils used in this study
Full size table

Abundance and diversity of methanotrophs in situ

The abundance of methanotrophs was estimated in soil samples in situ by qPCR targeting the pmoA gene (Fig. 1a). The results showed that pmoA gene abundance ranged from 7.74 × 105 ± 2.13 × 104 copies g−1 soil in G1 soil to 6.59 × 105 ± 0.95 × 105 copies g−1 soil in G3 soil. The abundance of pmoA gene increased with increasing grazing intensities. The light grazing (G1) had a significantly lower abundance of the pmoA gene compared with the ungrazed control (CK).

Fig. 1
figure 1

Abundance of pmoA gene (a) revealed by quantitative PCR and principal component analysis (PCA) of methanotroph-affiliated 16S rRNA reads (b) by MiSeq amplicon sequencing in grazed grassland soil in situ. Bars represent standard error of triplicate samples. The different letters above the columns indicate a significant difference (p < 0.05) based on the analysis of variance with Tukey’s post hoc test

Full size image

PCA showed that over 67% methanotroph variations were explained by the first two axes, with PC1 and PC2 explaining 48.9% and 17.7% of the total variance, respectively (Fig. 1b).

Methane oxidation potential and abundance of methanotrophs

The CH4 oxidation potential under a headspace of 8% (v/v) CH4 concentration ranged from 1.59 ± 0.07 to 2.25 ± 0.07 μmol g dry weight−1 day−1 for the grazed soils (Fig. 2). No significant differences in the CH4 oxidation potential were observed between 13CH4 labelled and CH4 control microcosms. The CH4 oxidation potential was as low as 0.51 ± 0.05 μmol g dry weight−1 day−1 in the ungrazed control, which was significantly lower than the grazed soils. Compared with the light and intermediate grazing, the heavy grazing significantly decreased the CH4 oxidation potential.

Fig. 2
figure 2

CH4 oxidation potential in different grazing soils with a CH4 concentration of 8% (v/v) in air. Bars represent standard error of triplicate samples. The different letters above the columns indicate a significant difference (p < 0.05) based on the analysis of variance with Tukey’s post hoc test

Full size image

The community size of methane-oxidizing bacteria was not significantly different for the ungrazed control after 28 days of incubation (Fig. 3a). The copy number of pmoA genes increased significantly from 1.38 × 107 ± 3.21 × 105 at day 0 to 3.39 × 107 ± 0.5 × 106 in microcosms of the G3 soil after 28 days of incubation. Significantly higher abundances of the pmoA gene were observed in the G1 and G2 soils, representing 12- and 6-fold increases, respectively (Fig. 3a). Similar results were observed from the relative abundance of methanotrophs reads to total 16S rRNA reads by MiSeq amplicon sequencing (Fig. 3b). Methanotrophs reads showed a significant increase after 28 days of incubation in the grazed soils but not in the ungrazed control. Correlation analysis showed a strong correlation between the potential CH4 oxidation rates and pmoA quantities (r = 0.92, p < 0.01) during incubation, but not with pmoA quantities in situ (r = − 0.517, p = 0.085).

Fig. 3
figure 3

Abundances of pmoA gene by quantitative PCR (a) and relative abundance of methanotrophic reads to total 16S rRNA gene reads by MiSeq amplicon sequencing (b) in soil microcosms over an incubation period of 28 days. The error bars represent the standard errors of triplicate microcosms. The different letters above the columns indicate a significant difference (p < 0.05) based on the analysis of variance with Tukey’s post hoc test

Full size image

Active methanotrophs in soils

The relative proportion of pmoA across CsCl gradients was similar among the three soils despite the differences in the intensity of grazing (Fig. 4a). The maximum relative proportion of pmoA was initially detected in the light fractions (around a buoyant density of 1.723 g mL−1) in the CH4 microcosms but shifted to the heavy fractions (around a buoyant density of 1.745 g mL−1) in the 13CH4 microcosms after the 28-day incubation (Fig. 4a). This indicated that methanotrophs grew by assimilating CH4 during active methane oxidation. By contrast, no significant shifts in the relative proportion of pmoA were observed in the CH4 and 13CH4 incubations of the ungrazed control soil (Fig. 4a). Miseq sequencing of methanotroph-affiliated 16S rRNA gene distributed across the CsCl gradient also indicated the assimilation of CH4 in the three grazed soils but not in the ungrazed control during the active methane oxidation (Fig. 4b).

Fig. 4
figure 4

Quantitative distribution of the pmoA gene based on qPCR across the entire buoyant density gradient of the DNA fractions (a) and percentage distribution of methanotroph-affiliated 16S rRNA reads by MiSeq amplicon sequencing for the heavy DNA (fraction as 4–7) (b) from soil microcosms incubated with 12CH4 or 13CH4 for 28 days. The normalized data are the ratios of the gene copy number in each DNA gradient to the maximum quantities from each treatment

Full size image

Phylogenetic analysis revealed that active methanotrophs in the 13C-DNA from the three grazed soils were most closely affiliated with Methylobacter and Methylosarcina of type Ia and Methylocystis of type II methanotrophs (Fig. 5; Supplementary fig. 1). Methylosarcina was enriched in the heavy fractions in grazed soils G1 and G2 after incubation, while Methylocystis was only enriched in G1 soil (Fig. 5).

Fig. 5
figure 5

Proportional changes of methanotrophic phylotypes in SIP microcosms over an incubation period of 28 days. HF indicated the methanotrophs from heavy fraction of DNA at 28 days

Full size image

Correlating soil properties with active methanotroph communities

Bulk density, total C, and Olsen P in combination with NO3−-N explained the highest percentage of the variance of active methanotroph communities (Fig. 6). The soil bulk density and NO3−-N content were significantly correlated with the first axis (explaining 77.3% of the total variance).

Fig. 6
figure 6

Redundancy analysis (RDA) between soil physicochemical properties (bulk density, total C, Olsen P, and NO3−-N) and active methanotrophs revealed by MiSeq amplicon sequencing of labelled DNA during the incubation in the grazed grassland soils

Full size image

Correlation analysis showed that the soil bulk density was negatively correlated with the increase of pmoA gene abundance (r = − 0.767, p < 0.05) and methane oxidation potential (r = − 0.782, p < 0.05) during the incubation in the grazed grassland soils (Supplementary fig. 3). The ungrazed control was excluded for the correlation analysis because of no detectable growth of methanotrophs and extremely low CH4 oxidation potential in the ungrazed control soil (Figs. 2 and 3).

Discussion

Effect of grazing on methanotroph community in situ

The abundance and community composition of methanotrophic bacteria were evaluated for three field samples all exhibiting various levels of grazing. The abundance of a key gene in the methane oxidation pathway, the pmoA, was used as a proxy for the population of methanotrophs in the steppe. Compared with the ungrazed control, the abundance of pmoA gene significantly decreased in G1 soil, while increased in G3 soil (Fig. 1a). Abell et al. (2009) found that the abundance of the predominant type I methanotrophs was positively affected by long-term cattle grazing in an alpine meadow soil. In contrast, no significant change of methanotroph abundance with grazing was observed in an alpine meadow (Zheng et al. 2012). On the other hand, grazing was reported to impact the composition of the methanotrophic community in a typical grassland, while no impact was apparent in an alpine meadow (Zhou et al. 2008; Zheng et al. 2012). In this study, distribution of methanotrophs among the soils analyzed shown by PCA indicated a significant impact of grazing on the methanotroph communities (Fig. 1b). In addition, we observed a significantly lower abundance of methanotrophs in G1 soil, even though the concentrations of SOM and TN were significantly higher than others (Table 1). Some factors like the availability of N, cross-feeding, and other C sources, apart from the CH4 availability, have also been proposed to regulate the population size of methanotrophs in upland soils (Li et al. 2018; Malghani et al. 2016). In this study, the pmoA abundance was strongly correlated with SOM (r = − 0.758, p < 0.01), TN (r = − 0.584, p < 0.05), and moisture content (r = − 0.611, p < 0.05). The low abundance of methanotrophs in G1 soil was mainly ascribed to the competition between methanotrophs, which represent only a small fraction of the total bacterial community, and heterotrophs, which was markedly stimulated by the higher nutrients in G1 soil. These results indicate that long-term grazing changes not only the soil properties but also the abundance and composition of functional microbes like methanotrophs in situ.

Effect of grazing on CH4 oxidation potential

The CH4 oxidation potential measurements are useful for comparing the relative activities of the methanotroph populations within samples from different environmental conditions (Kightley et al. 1995). The potential CH4 oxidation rates were variable between the grassland samples, ranging from 0.51 ± 0.05 to 2.25 ± 0.07 μmol g−1 d−1 (Fig. 2). While the potential CH4 oxidation rates were within ranges seen in wetlands (0.17 to 80 μmol CH4 g−1 d−1) (Graef et al. 2011) and geothermal environments (1.0–141 μmol CH4 g−1 d−1) (Sharp et al. 2014) with similar elevated CH4 concentrations (> 5% v/v), they were higher than those in upland soils (up to 74.64 nmol CH4 g−1 d−1) reported by Knief et al. (2003), and in grassland soils (2.42 to 21.54 nmol CH4 g−1 d−1) reported by Kou et al. (2017) with lower concentrations of CH4 (< 0.1% v/v). The strong correlation (r = 0.92, p < 0.01) between the potential CH4 oxidation rates and pmoA quantities suggests that there are a constant activity and a consistent abundance of methanotrophs throughout the 28-day incubation period. A similar trend was observed from the relative abundance of methanotroph reads to total 16S rRNA gene reads by MiSeq amplicon sequencing (Fig. 3b), further proving the activity of methanotrophs during incubation. However, the CH4 oxidation potential was as low as 0.51 ± 0.05 μmol g−1 dry weight day−1 in the ungrazed control. In agreeing with this, no growth of methanotrophs was detected in the ungrazed control using qPCR of pmoA genes and relative abundance of methanotroph reads to total 16S rRNA gene reads by MiSeq amplicon sequencing (Fig. 3). Also, methanotrophs were not labelled by 13C-CH4 confirming our expectation of low CH4 oxidation activity in the site and negligible growth of methanotrophs (Fig. 4). It is well known that N fertilizers inhibit CH4 oxidation by ammonia, which competes with CH4 for the methane monooxygenases in methanotrophs (Bédard and Knowles 1989). Even though the affinity of MMO for CH4 is 600–1300-fold higher than its affinity for ammonia, high concentrations of ammonium (40 mg NH4+-N kg−1) are known to substantially inhibit CH4 oxidation (Alam and Jia 2012). Grazing exclusion has been reported to increase plant biomass, root biomass, root exudate, and available soil N (Wang et al. 2018). The concentrations of exchangeable NH4+-N were significantly higher in the ungrazed control than the grazed soils, which could partly explain the low CH4 oxidation in the ungrazed control soil. Ho et al. (2019) suggest that “high-affinity” methanotrophs predominate CH4 oxidation in native upland soils, while canonical methanotrophs predominate in the anthropogenic-impacted upland soils. Another possible explanation for the low CH4 oxidation potential in the ungrazed control could be due to the methanotrophs in this site predominantly composed of the putative “high-affinity” methanotrophs. Moreover, no correlation (r = − 0.517, p = 0.085) between the potential CH4 oxidation rates and pmoA abundance in situ was observed, further proving that the mere presence of pmoA genes under in situ soil conditions may not necessarily reflect the functional activity of CH4 oxidation (Nannipieri et al. 2019).

It is widely accepted that heavy grazing would cause a decline in CH4 oxidation rate (Chen et al. 2011). Previous studies have indicated that an increase in stocking rate induced a reduction in CH4 uptake (Holst et al. 2008; Wang et al. 2012). Simulating the effects of grazing management with the PaSim model, Soussana et al. (2004) have suggested that a decline in the greenhouse gas sink activity of managed steppes occurs with increased stocking intensity. In this study, heavy grazing significantly decreased the potential CH4 oxidation rate in G3 soil compared with light and intermediate grazing in G1 and G2 soils (Fig. 2). Heavy grazing could significantly increase soil bulk density and directly affect the air permeability incurred from sheep trampling (Table 1) (Ball et al. 2012; Pan et al. 2018a). Furthermore, heavy grazing would decrease aboveground plant and litter biomass and consequently increase water stress, potentially inhibiting the activities of methanotrophs indirectly (Chen et al. 2011; Cui et al. 2018; Liu et al. 2007). Heavy grazing has been previously reported to inhibit the growth of ammonia-oxidizing bacteria, reducing the nitrification activity in grazed grassland soils (Pan et al. 2018b). In contrast to these negative impacts on plants, soil properties, and microbes by heavy grazing, light grazing could lead to a greater diversity of plant species, and the dense fibrous rooting systems of plants would benefit soil organic matter formation and soil C sequestration (Reeder and Schuman 2002). N returned in animal excreta and/or modification of N uptake and C exudation by frequently defoliated plants could also promote soil fertility and enhance microbial activities (Le Roux et al. 2003; Luo et al. 2019; Pan et al. 2018a; Zhou et al. 2010; Zhu et al. 2018). This study showed that grazing alters soil functional traits with light and intermediate grazing stimulating the growth and activity of methanotrophs, while heavy grazing significantly decreased the abundance of methanotrophs and the methane oxidation potential.

Active methanotrophs in grazed soils

Overall, the methanotroph communities present before and after incubation were similar (Supplementary fig. 1). Communities of the active 13C-labelled methanotrophs were more diverse in the light and intermediate grazed soils than in the heavily grazed soil (Fig. 5). The active methanotroph community included members of Methylobacter and Methylosarcina (type I methanotrophs) and Methylocystis (type II methanotrophs) (Supplementary fig. 2). Most of the active methanotrophs (> 90%) were closely related to the type I methanotroph, Methylobacter luteus (Fig. 5), a species originally isolated from a sewage (Bowman et al. 1993; Romanovskaia et al. 1978). Methylobacter-related type I methanotrophs have been reported to be responsible for the majority of methane oxidation in a long-term grazing site in Austria and also in six grazed grassland soils across New Zealand (Abell et al. 2009; Di et al. 2010). The mean annual temperature of 0.3 °C and maximum monthly mean temperatures of 19 °C in the studied field favored the growth of Methylobacter-related type I methanotrophs, as Methylobacter species have been reported to prefer cold environments such as the active layer of Arctic permafrost (Liebner et al. 2009), high Arctic wetlands (Graef et al. 2011), lake sediments (He et al. 2012), plateau wetlands (Deng and Dumont 2016), and rice fields from cold regions (Sultana et al. 2019). No active methanotrophs in this study were affiliated with the proposed atmospheric methane-oxidizing lineages USCα and USCγ (Knief et al. 2003). Moreover, Methylosarcina of type I methanotrophs, which have shown a transient ability to oxidize methane at atmospheric levels and also possible support “high-affinity” methane oxidation activity in paddy soil (Cai et al. 2016), were detected during CH4 oxidation in G1 and G2 soils. Recently, canonical methanotrophs have been suggested to predominate CH4 oxidation as high-affinity methane-oxidizers in anthropogenically-impacted upland soils (Ho et al. 2019). Furthermore, Methylocystis of type II methanotrophs were only detected in the G1 soil. The significantly higher content of SOM in the G1 soil might partly explain the result, as some Methylocystis species are known to oxidize and grow on acetate and ethanol in addition to methane (Belova et al. 2011; Im et al. 2011). Another reason for it should be the significantly higher Olsen P in the G1 soil, because the abundance of type II methanotrophs are positively related to phosphorus and adopt a competitor-ruderal lifestyle (Ho et al. 2013). It thus suggests that the diverse active methanotrophs and higher abundance of pmoA genes in the light and intermediate grazed soils enabled the significantly higher CH4 oxidation potential. These results indicate that light and intermediate grazing stimulate the growth and activity of diverse methanotrophs, while heavy grazing significantly decreases the abundance and diversity of active methanotrophs in this typical steppe.

Methanotroph activity

Animal grazing alters soil water and energy balance by reducing vegetation, increasing soil compaction, or reducing soil aeration by trampling and also soil chemical properties (e.g., pH and organic matter content), which would subsequently induce variation of microbial communities and activity (Li et al. 2019; Liu et al. 2019b; Lu et al. 2019; Pan et al. 2018c; Saggar et al. 2004; Steffens et al. 2008; Yu et al. 2018). Numerous studies have estimated the impact of grazing on either methanotroph communities or methanotroph activity (CH4 uptake), even though the abundance and composition of methanotrophs may not necessarily reflect their activity (Abell et al. 2009; Savian et al. 2014; Van den Pol-van Dasselaar et al. 1999; Zheng et al. 2012). In this study, DNA-SIP was used to link the identity and function of methanotrophs in grazed grassland soils. RDA showed that soil bulk density and NO3−-N were significantly correlated with the distribution of active methanotrophs in grazed grassland soils (Fig. 6). Methane diffusion, which is determined by the soil bulk density and moisture, is considered the limiting factor for CH4 oxidation in soil (Serrano-Silva et al. 2014; Walkiewicz et al. 2018). The single quantification of target genes may not necessarily reflect their functional activity (Nannipieri et al. 2019). In this study, no correlation (r = − 0.517, p = 0.085) between the potential CH4 oxidation rates and pmoA abundance in situ was observed. We linked the potential CH4 oxidation rates with the increase of pmoA gene abundance during the incubation. The significant correlation between the soil bulk density and the increases of pmoA gene abundance (r = − 0.767, p < 0.05), and also methane oxidation potential (r = − 0.782, p < 0.05) during the incubation in the grazed grassland soils further indicated the important impact of bulk density on the methanotroph communities and functional activity (Supplementary fig. 3). These results indicate that grazing induced variation of bulk density and soil C and N altering the abundance and communities of active methanotrophs and subsequently changes the CH4 oxidation activity.

SIP studies often require in vitro incubations and only partially reflect conditions in situ. This method may distort the relative abundance of organisms active in a particular process (Chen and Murrell 2010; McDonald et al. 2005). Fairly high CH4 concentrations, which probably do not reflect on in situ methane levels (typically atmospheric methane levels), were selected in our incubations in order to ensure the labelling of active microbial communities during the CH4 oxidation. The often occurring cross-feeding effect for SIP experiments is not a problem in this study because methanotrophs prefer to utilize CH4 as their source of C and energy even with the existence of microbial metabolites or microbial residues from labelled methanotrophs (Bao et al. 2019). No significant changes in the methanotroph communities before and after incubation were found in this study, which indicate that SIP results might largely reflect the functional process of methane oxidation under field conditions (Supplementary fig. 1). The lack of labelled methanotrophs in the ungrazed control could mainly be due to little microbial growth during incubation, as DNA-SIP relies on cell proliferation. The lowest CH4 oxidation potential rate observed in the ungrazed control further supports the negligible activity of methanotrophs, even though the activity might have resulted from the activation of dormant microbial populations rather than by their growth (Ho et al. 2015). It is worth mentioning that phospholipid fatty acid analysis (PLFA)–based SIP could ensure labelling methanotrophs at low methane concentrations, however, which would enable the detection of microbial groups but not on microbial genera or species as DNA-SIP (Ho et al. 2019).

Conclusions

Overall, this study showed that high abundance and diversity of methanotroph communities under in situ soil conditions may not necessarily reflect high functional activity of CH4 oxidation. Light and intermediate grazing stimulated the growth and activity of active methanotrophs, while heavy grazing significantly decreased the abundance and diversity of active methanotrophs. Phylogenetic analysis of the 13C-enriched DNA fractions from the DNA-SIP microcosms revealed that the active methanotrophs were dominated by the genus Methylobacter of type I. This study also showed that soil physicochemical properties, bulk density, and soil C and N are key factors determining the abundance and composition of active methanotrophs and subsequently the CH4 oxidation activity in the long-term grazed grassland soil.