Introduction

Methane (CH4) is a potent greenhouse gas with a global warming potential 84-fold that of carbon dioxide (CO2) over a 20-year period (IPCC 2014). Natural wetlands are critical CH4 emission sources contributing 23% of the total global CH4 emissions (Conrad 2009). Compared with freshwater wetlands, coastal salt marshes represent a small source of CH4 emissions (Poffenbarger et al. 2011; Livesley and Andrusiak 2012) but they are highly efficient carbon sinks. It is estimated that vegetated coastal habitats, such as marshes and mangroves, may hold up to 70% of all carbon stored in ocean sediments (Mudd et al. 2009; Nellemann et al. 2010); however, most coastal marshes are now threatened by climate change and invasions by alien plants (Ayres et al. 2004; Kirwan et al. 2010).

CH4 emissions are balanced by production and consumption carried out by methanogens and methanotrophs, respectively. Aerobic methanotrophs are responsible for most of the biological CH4 oxidation (Murrell 2010). Phylogenetically, type I methanotrophs are in the families Methylococcaceae and Methylothermaceae of the Gammaproteobacteria, and also in the family Crenotrichaceae (Hirayama et al. 2014; Knief, 2015). Type II methanotrophs in Alphaproteobacteria are divided into type IIa and type IIb methanotrophs, which are affiliated to the families Methylocystaceae and Beijerinckiaceae, respectively (Dumont et al., 2014). In addition to methanotrophs, both Methylocystaceae and Beijerinckiaceae include non-methanotrophic species (Dunfield et al., 2010). Methanotrophs in the family Methylacidiphilaceae of Verrucomicrobia are referred to as type III methanotrophs (Dunfield et al. 2007; Pol et al. 2007; Knief, 2015). Performing the first step of CH4 oxidation is methane monooxygenase (MMO), a key enzyme of methanotrophs that has particulate (pMMO) and soluble methane monooxygenase (sMMO) forms (Hanson and Hanson 1996). The pmoA gene encoding a subunit of the pMMO is commonly used as a marker to study aerobic methanotrophs (Dumont and Murrell 2005). To the best of our knowledge, however, only a few studies have investigated the activity and community composition of aerobic methanotrophs in coastal marsh ecosystems. Type II methanotrophs of Methylocystis are reportedly dominant in a brackish marsh located on a sub-estuary of Chesapeake Bay, USA (Lee et al. 2017); yet it remains unclear whether this is also the common pattern for other coastal marsh ecosystems, let alone for the functionally active methanotrophic populations in these ecosystems.

The Yancheng tidal marsh (in Jiangsu Province, China) is characterized by an intertidal mudflat and dominated by herbaceous vegetation. Two native species, Suaeda salsa and Phragmites australis, and one alien, Spartina alterniflora, currently dominate this marsh ecosystem. S. alterniflora, native to the Atlantic coast of North and South America (Bertness 1985; Adam 1990), was first introduced to China in 1979 for the sediment stabilization of its intertidal marshes (Qin and Zhong 1992). Since then, this species has spread rapidly and is now mainly distributed in the Jiangsu, Shanghai, Zhejiang, and Fujian Provinces of China (Zuo et al. 2012), where its range has increased to 2661.74 hm2 from 1996 to 2010 (Wang and Liu 2014). In the process, S. alterniflora has gradually replaced the habitats of S. salsa and P. australis in the tidal marshes of Jiangsu Province (Zhang et al. 2004). Compared with the native plants, S. alterniflora has a higher plant productivity, which results in rapid accumulation of soil organic carbon (SOC) (Zhang et al. 2010a). This provides an additional substrate for methanogens to produce CH4 that could alter the CH4 emissions of tidal marshes. CH4 cycle-related research in these vegetated habitats of the Yancheng tidal marsh is largely limited to measuring CH4 fluxes (Zhang et al. 2010b; Xu et al. 2014; Yuan et al. 2014; Xiang et al. 2015), CH4 production potential, and to methanogenic community structure analyses (Yuan et al. 2015). The CH4 oxidation potential and active methanotrophic populations after the S. alterniflora invasion in this tidal marsh have not yet been characterized. To this end, we employed DNA stable-isotope probing (DNA-SIP) in this study to identify the active consumers of CH4 in the soils of the four main habitats of the Yancheng tidal marsh.

Materials and methods

Soil sample collection and analyses

Soil samples were collected in the Yancheng National Wetland Reserve of Jiangsu Province, China (33° 36’ N, 120° 36’ E). Currently, there are four main habitat types in the Yancheng tidal marsh from the coast to terrestrial land (Zuo et al. 2012); (1) the bare tidal mudflat (MF), with no vegetation in this habitat; (2) habitat with the invasive species S. alterniflora (SA); (3) habitat with S. salsa (SS) as the dominant vegetation; and (4) habitat with P. australis (PA) as the dominant vegetation. Along the Yancheng coastal vegetation profile, three soil cores from each habitat and thus 12 samples were collected (8 cm in diameter, 0–5 cm in depth) in June 2015. Soil samples were kept at 4 °C using ice bags during transportation. Upon arrival in the laboratory, the main roots in soils were removed and samples were further divided into two subsamples. The subsample used for soil property analysis was air-dried and passed through a 2-mm sieve, while that used for isotope labeling was stored at 4 °C until SIP incubation was initiated within 1 week.

Soil organic carbon (SOC) and total nitrogen (TN) were determined by dichromate oxidation (Bremner and Jenkinson 1960) and Kjeldahl digestion (Bremner 1960), respectively. Soil available phosphorus (AP) was extracted by sodium bicarbonate and measured using the molybdenum antimony anti-colorimetry method. Soil salinity was determined by measurement of the electrical conductivity of a soil-water sample. Soil pH was measured with a water/soil ratio of 2.5 (v/v).

DNA extraction and quantitative PCR

Total genomic DNA of 0.5 g soil sample was extracted using a FastDNA SPIN Kit for Soil (MP Biomedicals). Primer pairs A189f (5’-GGNGACTGGGACTTCTGG) and mb661r (5’-CCGGMGCAACGTCYTTACC) (Costello and Lidstrom 1999) were used to quantify the abundance of methanotroph pmoA genes (Kolb et al. 2003). Quantitative PCR (qPCR) was conducted in 25-μl reaction mixtures containing 12.5 μl of SYBR premix Ex Taq™ (Takara Co), 0.5 μl of 20 μM of each primer and 1 μl of template DNA. The cycling steps were as follows: 94 °C for 6 min, followed by 40 cycles of 94 °C for 25 s, 65.5 °C for 20 s, and 72 °C for 45 s, and then melting curve analysis from 75 °C to 94.8 °C, increment 0.2 °C for 6 s and data collection. The reaction was conducted on a CFX96 Optical Real-Time Detection System (BioRad). Ten-fold serial dilutions of plasmid DNA from 102 to 108 copy numbers μl−1 were used to perform standard curves. Samples were quantified in triplicate. The amplification efficiencies were between 91.8 and 94.5%.

13CH4 labeling and gradient separation

Ten grams of fresh soils from each sample was incubated in 125-ml serum vials capped with butyl stoppers. A total of 10 ml of air in the headspace of each vial was removed using a syringe, after which 10 ml > 99.9% labeled 13CH4 or natural abundance CH4 (referred to here as 12CH4) was injected into the headspace. Soils were then incubated on a shaker (100 rpm) in the dark at 25 °C. CH4 concentration in the headspace of vials was measured daily using a gas chromatograph (Agilent 7890B GC). Once the CH4 concentration in vials started to increase, the incubation was stopped. In total, soil samples were labeled for 6 days.

Soil DNA was extracted immediately after incubation using a FastDNA SPIN Kit for Soil (MP Biomedicals), after which SIP gradient fractionation was performed as previously described (Neufeld et al. 2007a; Deng et al. 2016). Briefly, 2 μg of purified DNA was centrifuged in CsCl gradients with an average density of 1.725 g ml−1. The CsCl gradients were prepared by mixing 4.9 ml of a 1.85 g ml−1 CsCl stock solution with 1 ml of gradient buffer (GB, 100 mM Tris-HCl, pH 8.0; 100 mM KCl; 1 mM EDTA). The CsCl density gradients were added into 6-ml polyallomer ultracentrifuge tubes and then centrifuged at 190,000×g for 44 h at 20 °C in a Hitachi Preparative Centrifuge CP80WX (Hitachi Ultras). The gradient solution was subsequently fractionated from bottom to top into 11 DNA fractions by displacing with sterile water using an NE-1000 single syringe pump (New Era Pump Systems, Inc.). Each fraction contained about 0.5 ml of solution. The buoyant density of each collected fraction was determined using an AR200 digital hand-held refractometer (Reichert), after which DNA was precipitated from CsCl gradient fractions with polyethylene glycol and dissolved in 30 μl of elution buffer.

Quantitative PCR, terminal restriction length polymorphism, and clone analysis of gradient fractions

The A189f-mb661r primers were used to quantify the pmoA gene copies in DNA of labeled soils and in DNA of different gradient fractions (Kolb et al. 2003) using the same qPCR conditions as described above.

T-RFLP of the bacterial 16S rRNA genes was performed from the heavy and light DNA fractions using the primers 27F (5’-AGAGTTTGATCMTGGCTCAG)-FAM and 907r (5’-CCGTCAATTCMTTTRAGTTT) (Lueders et al. 2004). PCR was conducted in a total volume of 50-μl with an initial DNA denaturation step (94 °C, 3 min), followed by 32 cycles of denaturation (94 °C, 30 s), annealing (52 °C, 45 s), and elongation (72 °C, 1 min), and then final extension at 72 °C for 10 min using an Eppendorf PCR instrument. The PCR products were purified using the GenElute PCR Clean-up Kit (Sigma). Next, approximately 100 ng of purified product was digested with the restriction endonuclease MspI (TaKaRa Co) for 1 h at 37 °C. Reactions were stopped by incubating at 65 °C for 15 min, after which the digested products were sent to Sangon Biotech Company (Shanghai, China) for terminal restriction fragments (T-RFs) separation.

To determine which active species the T-RF peaks represented, the same primers (27F and 907r) and PCR program described above for T-RFLP were used to construct the 16S rRNA gene clone libraries of the heavy fractions (ca. 1.74 g ml−1). The PCR products were then purified using a GenElute PCR Clean-up Kit (Sigma), after which the purified products were ligated into the pEASY-T1 vector (TaKaRa Co) and transformed into Escherichia coli TOP 10 competent cells according to the manufacturer’s instructions. On average, 36 clones (32–40) were collected in each sample and were classified into various cOTUs (Figure S3, we named cOTUs in order to distinguish them from OTUs obtained by Miseq sequencing). All clone sequences obtained were deposited in GenBank under accession numbers MF950917–MF951051.

MiSeq sequencing and sequence analysis

To further confirm the bacterial community that had been labeled by CH4 oxidation, barcoded PCR amplification of the 16S rRNA gene from the DNA of the heavy fractions was conducted using the bacterial primers 515F (5’-GTGCCAGCMGCCGCGGTAA) and 907R (5’-CCGTCAATTCMTTTRAGTTT) (Angenent et al. 2005). The library was sequenced on an Illumina MiSeq system using standard procedures of a 2 × 300 cycle combination mode at Meiji Biotechnology Company (Shanghai, China).

Paired-end reads were first merged by USEARCH using the default settings (Edgar 2010). All reads were then subjected to quality control and chimera and singleton filtering based on the recommendations of the UPARSE pipeline (Edgar 2013). On average, 15,912 (7580–29,282) high-quality reads per sample were generated by amplicon sequencing. Species level OTU representatives for 16S rRNA genes were obtained using cluster_otus at 97% sequence identity in USEARCH. Taxonomic identities of represented OTUs were assigned using the Ribosomal Database Project Classifier against the SILVA 123 SSU Ref database (Quast et al. 2013) at a confidence level of 95% within Mothur (Schloss et al. 2009). Sequence datasets are available in the NCBI Sequence Read Archive under study accession number SRP6023620–23.

Statistical analysis

One-way analysis of variance and Tukey’s HSD (Honestly Significant Difference) were conducted to test for significant differences in estimated CH4 consumption rate and soil properties between the four habitats. Canonical Correspondence Analysis (CCA, Anderson and Willis 2003) was conducted to investigate the effects of soil properties on the community composition of active bacteria involved in oxidizing CH4. At first, CCA analysis was conducted for each soil property, including soil salinity, pH, SOC, TN, and AP. An ANOVA-like permutation test function (anova.cca) was run to detect the significance of each soil property contributing to bacterial communities in the labeled DNA. The variables of bacterial community composition only contained OTUs with the relative abundance higher than 0.1%. CCA statistical analyses were performed using the vegan package in R (Oksanen et al. 2016). For all analyses, a P value of < 0.05 was considered to be significant.

Results

Soil characteristics and methanotroph abundance

In the Yancheng tidal marsh, the S. alterniflora-dominated soil had significantly higher soil organic carbon (SOC), total nitrogen (TN), available phosphorus content (AP), and salinity than did the soils of the other three habitats (Table 1). No significant differences of TN and AP were detected among these three marsh soil habitats (i.e., mudflat, P. australis-dominated, and S. salsa-dominated). However, the average SOC in the mudflat soil was 1.90 g kg−1 and significantly lower than that of the other habitats. By contrast, the mudflat’s salinity was significantly higher than that of the P. australis- and S. salsa-dominated soils. The pH values of S. alterniflora-dominated soil were close to neutral and the lowest among these habitats. However, the methanotroph abundance of the sampled soils, as quantified by qPCR, was 106 pmoA gene copies g−1 fresh soil, with no significant differences detected among the four habitats (Table 1).

Table 1 Soil properties of the four studied habitats in the Yancheng tidal marsh, China. All values are means ± standard deviation (n = 3)
Full size table

The DNA-SIP experiments were performed to detect the active aerobic methanotrophs in each habitat sample. After 5 days of incubation, a significantly higher CH4 consumption rate was detected in the mudflat soil (2.67 ± 0.1 μmol g wet weight−1 day−1, Fig. 1), with ~ 7.5% CH4 in the vial headspace decreasing to ~ 5%. Significantly lower estimated consumption rates were detected in the P. australis- (2.18 ± 0.2 μmol g wet weight−1 day−1) and S. alterniflora- (2.03 ± 0.29 μmol g wet weight−1 day−1) dominated soils. The lowest rates were detected in S. salsa-dominated soil (1.6 ± 0.22 μmol g wet weight−1 day−1).

Fig. 1
figure 1

Eatimated CH4 consumption rate of the Yancheng SIP labeling soils. Soils from four marsh habitats (MF: mudflat, PA: Phragmites australis-, SS: Suaeda salsa-, and SA: Spartina alterniflora-dominated) were incubated with 13CH4 or 12CH4. Error bars represent standard deviations (n = 6)

Full size image

DNA was extracted from the soils on the 6th day of incubation. The copy numbers of methanotrophs in soils of the four habitats all increased from 106 pmoA gene copies g−1 fresh soil in the original soils to 107 after the incubation with 13CH4 or 12CH4 (Figure S1). Results of the qPCR targeting the pmoA genes in DNA of different density gradients are expressed as relative abundances. The ‘heavy’ pmoA gene peaks were detected in the heavy fractions (density > 1.73 g ml−1, Fig. 2) of 13CH4-incubated soils. Moreover, those soils incubated with 12CH4 only formed pmoA gene peaks in the light fractions (density < 1.73 g ml−1). These clear separations were detected in all four soils, indicating that their methanotroph pmoA genes were labeled by 13C, irrespective of any differences in their CH4 oxidation rates.

Fig. 2
figure 2

Distributions of the methanotroph pmoA genes, as determined by qPCR of the DNA recovered from CsCl gradient fractions, after 6 days of incubation with 13CH4 or 12CH4. The y-axis indicates the relative abundance at each gradient fraction, with the total quantity detected in a gradient equal to 1. The four panels correspond to the pmoA distribution from incubations of the four types of soils (MF: mudflat, PA: Phragmites australis-, SS: Suaeda salsa-, and SA: Spartina alterniflora-dominated). Densities between 1.73 and 1.8 g ml−1 corresponded to heavy fractions (labeled DNA), and those between 1.66 and 1.73 g ml−1 corresponded to light fractions (unlabeled DNA)

Full size image

Analysis of T-RFLP and the clone libraries of 16S rRNA genes constructed from heavy DNAs

After high-speed centrifugation and gradient separation, the concentration of DNA obtained was too low to amplify the pmoA genes. Hence, the 16S rRNA genes were used for the T-RFLP analysis of each fraction to characterize the bacteria involved in CH4 consumption. In addition, 16S rRNA gene clone libraries were constructed from the heavy fractions (density > 1.74 g ml−1), by using the same primers of T-RFLP to characterize the phylogenetic information of the obtained T-RF peaks. The T-RFs with lengths of 436 bp, 456 bp, and 489–491 bp were dominant in the heavy fractions of the 13CH4-labeled mudflat soil and not so dominant in the light fractions of the 12CH4-incubated soils (Figure S2). Phylogenetic analyses of the clone sequences from the heavy fractions showed that the T-RFs of 456 bp detected in the mudflat soil represented Methylomonas and Methylomarinum species (Figure S3). There were six cOTUs which had T-RFs of 489–491 bp in the mudflat soil: cOTU_2 (with a relative abundance of 15.6%) in the genus Methylotenera; cOTU_20 and cOTU_47 in the marine methylotrophic groups; and cOTU_7, cOTU_10, and cOTU_16 in the order Methylococcales (Figure S3). Those T-RFs with a length of 483 bp, and belonging to cOTU_1 in the order Methylococcales, were predominant in the S. salsa- and P. australis-dominated soils (Figures S2 and S3), in which representative cOTUs of Methylohalobius (7.5%, cOTU_6) and Methylocystis (2.9%, cOTU_46) were also detected respectively. By contrast, no corresponding cOTUs were identified for the T-RF peaks in the heavy fractions of the S. alterniflora-dominated soil (Figures S2 and S3). In this habitat, the most abundant cOTU was cOTU_10, which belonged to the order Methylococcales and had a relative abundance of 5.4% (Figure S3).

MiSeq sequencing of the 13C-enriched 16S rRNA gene targeting the bacteria actively involved in CH4 consumption

For the analyses of the composition of total bacterial community, only the most abundant OTUs were included— i.e., those with a mean relative abundance > 1% in all habitats or > 3% in one habitat (Fig. 3a). Methanotrophs are a subgroup of methylotrophs, and not all methylotrophs are methanotrophs. Besides CH4, methylotrophs can grow on other one-carbon compounds such as methanol (CH3OH). Except for the S. alterniflora-dominated soil, the methylotroph 16S rRNA sequences comprised ~ 50–70% of the heavy fraction in the other three marsh soil habitats. The abundant methylotrophic community from the labeled DNAs obtained via MiSeq sequencing was consistent with that obtained from the clone sequencing (Figs. 3a and S3).

Fig. 3
figure 3

Relative abundance of the major 16S rRNA gene-derived clusters that were retrieved from the heavy fractions (density = ~ 1.74 g ml−1). a MiSeq sequencing of the OTU levels, b MiSeq sequencing of the phylum levels. Error bars represent standard deviations (n = 3)

Full size image
Fig. 4
figure 4

Canonical correspondence analysis (CCA) ordination diagram of the active bacterial communities associated with the soil variables, based on an Illumina MiSeq sequencing of the 12 samples (colored circles) from the four studied habitats (MF: mudflat, PA: Phragmites australis-, SS: Suaeda salsa-, and SA: Spartina alterniflora-dominated). Blue arrows represent the five soil variables considered: soil organic carbon (SOC), total nitrogen (TN), soil available phosphorus (AP), soil salinity, and pH value. The relative abundance of the OTUs was used for the bacterial community composition. OTUs are shown as red ‘+’ symbols

Full size image

The majority of heavy sequences detected in the mudflat soil habitat were methanotroph 16S rRNA sequences and clustered to OTU_2 (Fig. 3a). It had a relative abundance of 24.2% and belongs to the family Methylococcaceae, as does OTU_1011, which was classified as Methylomonas and accounted for 6.3% of the heavy sequences occurring in the mudflat soil. The methylotroph OTUs—namely the OTUs of Methylotenera (OTU_4), Marine_Methylotrophic_Group_1 (OTU_5), and Methylophaga (OTU_7)—were more abundant in mudflat than other marsh soil habitats. The majority of the heavy sequences detected in the P. australis- and S. salsa-dominated soils were also methanotroph sequences and both belonged to family Methylococcaceae (Fig. 3a); however, unlike OTU_2 in the mudflat soil, they clustered to another OTU, named OTU_1. OTU_1 accounted for 62.1% and 44.6% of the labeled sequences in the P. australis- and S. salsa-dominated soils, respectively (Fig. 3a). Meanwhile, 5.3% of the Methylohalobius sequences (OTU_9) were detected in the S. salsa-dominated soil. The proportion of methylotroph sequences detected in S. alterniflora-dominated soil was much lower than that detected in the other soil habitats (Fig. 3a). OTU_1 and OTU_2 were also detected in S. alterniflora-dominated soil, but they accounted for just ~ 1.7% of the heavy fraction, while in this habitat relatively higher proportions of Methylomicrobium (3.5%, OTU_8) and Methylomonas (3.9%, OTU_1011) were detected.

Most of the abundant OTUs represented the methylotrophic community involved in CH4 oxidization (Fig. 3a). Correspondingly, most of the 16S rRNA gene sequences in the heavy fraction belonged to the Proteobacteria at the phylum level, especially Gammaproteobacteria (Fig. 3b). The proportion of non-Proteobacteria sequences was greater in the S. alterniflora-dominated soil than in the other marsh soil habitats, including bacteria belonging to the phyla Acidobacteria, Chloroflexi, and Gemmatimonadetes (Fig. 3b).

Correlations between the soil properties and the active bacterial community

The CCA and the ANOVA-like permutation test (anova.cca) for each soil property (soil salinity, pH, SOC, TN, and AP) aimed to detect the correlation between soil properties and the active bacterial community matrix. All the tested soil properties significantly accounted for the variation found in the sampled bacterial communities (P < 0.05). Among them, however, salinity was the most important explanatory variable: it independently accounted for 29.2% of variation in the active bacterial community composition. The CCA were further conducted with all soil properties and compositions of the active bacterial community included. The CCA model with all the five soil properties considered represented 79.6% of the total variation in the bacterial communities, of which CCA1 accounted for 40.2% and CCA2 for 27.6%. As Fig. 4 shows, the samples from the mudflat (MF) and S. alterniflora (SA) habitats clustered separately, while those of the more similar P. australis (PA) and S. salsa (SS) habitats clustered together, yet remained distinct from the samples of either MF or SA.

Discussion

Spartina alterniflora invasion is known to stimulate CH4 emission and production potential in the Yancheng tidal marsh (Zhang et al. 2010b; Yuan et al. 2014, 2015). CH4 emission is a balance between production and oxidation, which are performed by distinct groups of microorganisms. In the present study, the estimated CH4 oxidation rate of the S. alterniflora soil was higher than that of S. salsa soil and not significantly lower than P. australis soil (Fig. 1). This indicates that, compared with native plant-dominated soil, the greater CH4 emission of S. alterniflora soil was driven by the increased CH4 production and not decreased CH4 oxidation. It is also reasonable that the increased methanogenesis in this soil will have enriched methanotrophs and stimulated CH4 oxidation capacity.

The active community composition in the S. alterniflora soil differed from the other soils and showed a much lower proportion of 16S rRNA sequences from known methanotrophs among the total labeled bacteria. S. alterniflora and P. australis soils consumed the same amount of 13CH4, and pmoA gene copy numbers increased similarly in all soils. Therefore, it is difficult to explain why the relative abundance of labeled methanotrophs was particularly low in the S. alterniflora soil. One possible explanation is that the labeled Acidobacteria, Chloroflexi, and Gemmatimonadetes in this soil are novel methanotrophs, but this seems unlikely. A more plausible explanation is that greater cross-feeding of 13C from methanotrophs to heterotrophs occurred in the S. alterniflora soil. Cross-feeding is poorly understood, but widely observed for in SIP experiments targeting methanotrophs (Beck et al., 2013). A recent study showed that Methylotenera mobilis, a non-methanotrophic methylotroph, could induce excess methanol production in Methylobacter tundripaludum thus facilitating cross-feeding (Krause et al., 2017). It is plausible that the microbial community in the S. alterniflora soil has more extensive cross-feeding partnerships that the other soils, leading to wider labeling of the non-methanotroph populations with the 13CH4.

Most of the methylotroph OTUs from the 16S rRNA gene sequencing were unclassified at the genus level at a 95%-similarity level. Rather, they could be only identified to family or order and lacked close affiliations with the cultivated methylotrophic species. Moreover, these OTUs were abundant in the heavy fractions after the 13CH4 incubation, indicating that they played important roles in CH4 consumption. A dominance of Methylococcaceae OTUs was detected in the SIP-labeled soils in the Yancheng tidal marsh. This family is composed entirely of type I methanotrophs (Bowman et al. 1993; Knief 2015). Hence, it can be concluded that the active and abundant bacteria that consumed CH4 were unclassified type I methanotrophs.

In both the P. australis and S. salsa soils, OTU_1 belonged to an unclassified genus in the family Methylococcaceae that was dominant in the 13C-enriched fractions after the 6-day incubation. The representative sequences of OTU_1 had a 99.5% identity with the uncultured bacterial sequences obtained from mangrove sediment of Sundarban, India (Ghosh et al. 2010). In another red mangrove soil, unclassified Methylococcaceae were also detected as abundant methanotrophs by 13CH4 DNA-SIP via 16S rRNA MiSeq sequencing (Shiau et al. 2018). These findings indicate that these two typical coastal soil ecosystems, salt marshes and mangroves, likely share similar methanotrophic populations. Furthermore, OTU_1 was 96.9% similar to Methylomicrobium sp. XLMV4 (alternative name: Methylobacter sp. XLMV4) that was isolated from oil sands tailings ponds (Saidi-Mehrabad 2014). Because they are closely related, OTU_1 species likely share physiological characteristics with other Methylomicrobium species, which include well-known halotolerant and halophilic methanotrophs (Knief 2015) that are abundant in saline lake sediments (Deng et al. 2017). Soils in the Yancheng tidal marsh had salinity of 5.0‰–23.7‰, which could provide a suitable salinity for halophilic Methylomicrobium species.

OTU_2 was dominant in mudflat soil and also belonged to the family Methylococcaceae. It had a 98% identity with Methylomarinum sp. SSMP-1, which was isolated from a hydrothermal seep field at the northern terminus of the Eastern Pacific Rise (Fradet et al. 2016). The genus Methylomarinum is the cultivated representative of deep-sea cluster 2 (Knief 2015) while the deep-sea clusters 1 to 5 belong to uncultivated type I methanotrophs that have mostly been detected in marine ecosystems by targeting the pmoA gene as a functional marker (Tavormina et al. 2010; Lüke and Frenzel 2011; Knief 2015). Similar environmental conditions occur in the Yancheng marsh soils with its marine sediments, since the Yancheng marsh is continually flushed with seawater by the tide. Thus, we may infer that the different unclassified OTUs in the family Methylococcaceae obtained from the Yancheng marsh may correspond to the different “deep-sea” methanotroph clusters. However, only OTU_2 was related to deep-sea cluster 2, and no other pieces of conclusive evidence were found to prove this inference for the other unclassified OTUs. Nonetheless, further investigations are necessary to reveal the ecophysiological activity and phylogenetic classification of these unclassified methanotrophic populations.

Type II methanotroph Methylocystis species are reportedly abundant and active CH4 consumers in many peatlands of England, China, and North and South America (Chen et al. 2008; Gupta et al. 2012; Kip et al. 2012; Deng et al. 2013, 2016). However, in the Yancheng tidal marsh, we only found one such OTU belonging to Methylocystis, which accounts for 2.9% of the total labeled bacteria in the P. australis soil. The activity of Methylocystis species may be completely inhibited by 1% NaCl (Han et al., 2017). Therefore, even if Methylocystis was abundant in the in-situ soils, they may primarily occur as metabolically inactive spores, doing little to oxidize CH4. Another possibility is that adding high levels of CH4 into the headspace stimulated the growth of type I, rather than type II, methanotrophs during the SIP experiment we conducted.

We also found that the non-CH4-utilizing methylotrophic species Methylotenera and Methylophaga were active when labeled with 13CH4 in soils of the Yancheng tidal marsh. Methylotrophs are a group of bacteria that can grow on a number of different C1 compounds (Lidstrom 2006). Methylotenera species in the family Methylophilaceae are reportedly highly active in freshwater habitats when incubated with 13CH4 or 13CH3OH (Chistoserdova 2011, 2015). Methylophaga species are moderately halophilic methylotrophs in the family Piscirickettsiaceae (Doronina et al. 1997), which also have been detected as active bacteria by the 13CH3OH SIP method in coastal marine environments (Neufeld et al. 2007b). Therefore, in soils of the Yancheng tidal marsh, Methylotenera and Methylophaga may have used the CH3OH produced by methanotrophs and been detected by the 13CH4 SIP experiment.

That the samples formed three major groups, as suggested by the CCA, is indicative of the different active bacterial communities found among the habitats due to distinguishable soil properties. When compared with the mudflat and S. alterniflora habitats, the samples from P. australis and S. salsa had similar soil properties, especially with respect to soil organic carbon and salinity, and clustered together in the CCA plot. It has been reported that S. alterniflora has a competitive advantage over native species in high salinity conditions. As a consequence, this has changed the vegetation pattern by pushing the native communities to the low salinity zones (Tang et al. 2014). Soil salinity was also the primary factor influencing the bacterial community composition in the labeled soils, which was consistent with other bacterial community analyses of sediment in the Florida Coastal Everglades (Ikenaga et al. 2010) and Gironde estuary (France) (Prieur et al. 1987). These results emphasize the important role that salinity plays in shaping the community composition of active bacteria involved in CH4 oxidation.

Conclusion

In this study, we successfully used 13CH4 DNA-SIP approach and identified the microorganisms involved in CH4 oxidation in four soil habitats of the Yancheng tidal marsh. Compared with native vegetation and mudflat habitats, a lower relative abundance of active methanotrophs was detected in the S. alterniflora-dominated habitat, which suggested different ways of carbon cross-feeding. Our results show that in these four marsh habitats different unclassified type I methanotrophs in the family Methylococcaceae were the major active aerobic methanotrophs, thus suggesting an important role for Methylococcaceae in the dynamics of CH4 oxidation in coastal marshes.