Introduction

Methane (CH4), the second largest contributor to radiative forcing after carbon dioxide (CO2), causes approximately 15–20% of the observed global warming (IPCC 2013). Agriculture methane emissions are crucial to the global methane estimates, and these emissions magnitude are highly dependent on land-uses and fertilizer application (Scheer et al. 2008; Liu et al. 2016). Paddy cultivations have long been recognized as the main contributors to anthropogenic CH4 emissions, ranging from 30 to 50 Mt y−1 (Sass et al. 1994), whereas aerobic upland soils are the second greatest global CH4 sinks, accounting for 6–10% of the total atmospheric CH4 uptake or about 30 Tg y−1 CH4 (IPCC 2013). Therefore, the expected land-use conversion from paddies to uplands could possibly mitigate the global warming potentials by reducing CH4 emissions.

It is urgently demanded to understand the extent to which land-use conversion from paddy fields to upland plots contribute to greenhouse gases budgets, but CH4 fluxes record from continuous measurements of paddy fields and upland plots are limited. In the recent decades, results from field measurements have demonstrated that converting from paddies to uplands would change soil from emission source to sink of atmospheric CH4 (Nishimura et al. 2008; Liu et al. 2016). The dynamics of CH4 in paddy fields significantly differ from those in upland plots in which aerobic decomposition process is dominant (Nishimura et al. 2008; Scheer et al. 2008). However, there have been far fewer field investigations on the underlying mechanisms of decreases in CH4 fluxes induced by converting paddies to uplands. Meanwhile, fertilization is the management practice most frequently used for paddy fields and upland plots (Wang et al. 2015; Fan et al. 2016). While it has been reported that fertilization had controversial effects on soil CH4 fluxes in different land-uses, ranging from positive (Wang et al. 2015), negative (Liu et al. 2016; Zhou et al. 2016), and no significant effects (Fan et al. 2016). Thus, adequate estimates of soil CH4 fluxes, coupling with the changes in land-uses and fertilization, would be more helpful for the global greenhouse gases emissions budget.

Generally, CH4 fluxes from soils are determined by the counterbalance of methane production and subsequent oxidation by methanogenic and methanotrophic microorganisms, respectively (Hanson and Hanson 1996; Watanabe et al. 2007). The mcrA gene commonly encodes the K-subunit of methyl coenzyme M reductase (MCR), which is the key catabolic enzyme of methanogens (Freitag and Prosser 2009). The pmoA gene, commonly recognized as a phylogenetic marker of methanotrophs in ecological studies, encoding the membrane-bound subunit of the particulate methane monooxygenase (MMO) (Henckel et al. 1999). Moreover, it is widely accepted that the particulate methane monooxygenase (pmoA) gene provides a molecular indicator of methane oxidation potential (Scheer et al. 2008; Singh et al. 2009). Therefore, the abundance and community compositions of methanogens and methanotrophs could be respectively characterized by the mcrA and pmoA genes. Within the last decade, studies on the abundance and community compositions of the methanogens and mehanotrophs under different land-uses have become more common (Bao et al. 2016; Wang et al. 2015). Nevertheless, the changes of the methanogens and methanotrophs induced by land-use conversion from paddies to orchards are scarce. In addtion, researches on the linkage between the relating microorganisms and in situ soil CH4 fluxes remains unclear. Therefore, it is essential to explore the relationships between methanogens, methanotrophs, and soil CH4 fluxes to clarify the underlying mechanism driving the soil-atmosphere exchange of CH4.

China is one of the most important rice-producing countries in the world and the annual CH4 emissions from Chinese rice cultivations are about 7.4 Tg CH4 Y−1 (Yan et al. 2009). The hilly red soil region is characteristic of typical agricultural soil and covers an area of almost 11.8% of the country land surface, producing 80% of the rice, and supporting 22.5% of China’s population (Liu et al. 2016). Over the past few years, this region has undergone great changes in land-uses, in particular the land-use conversion from paddies to uplands for cultivating fruits and economic orchards. These changes in land-uses and fertilization might cause marked variations in the soil CH4 fluxes owing to the alteration in the microbial activity. However, little information is available on the exhaustive assessments of the responses of soil methanogens, and methanotrophs, as well as in situ CH4 fluxes to the land-use conversion and fertilization in the hilly red soil region of southern China. Therefore, the present study aimed to (1) evaluate the effects of land-use conversion and fertilization on soil CH4 fluxes; (2) examine the variations in the abundance and communities of methanogenic archaea and methanotrophic bacteria, with the mcrA and pmoA gens, respectively; and (3) explore the potential linkages between soil CH4 fluxes and the abundance and communities of methanogens and methanotrophs.

Materials and methods

Site description and experimental design

The experiment site was conducted at Qianyanzhou Ecological Research Station of Chinese Academy of Science (CAS) (26° 44′ 46″ N, 115° 04′ 05″ E) in Taihe city, Jiangxi province, southern China. According to the meteorological data from 1989 to 2010, this area was characteristic of a subtropical monsoon climate, with an average annual air temperature of 18.0 °C and precipitation of 1509.0 mm. The soil is typically red soil found in southern China and classified as Cambosols in the Ultisols soil classification.

In the present experiment, the two most common land-uses were selected, namely rice paddy (Oryza sativa L.) and citrus orchard (Citrus reticulata L.). The paddy fields had been continuously cultivated for approximately 10 years and a portion of the fields has been converted to orchard plots in June 2012. Meanwhile, in order to evaluate fertilization effects, fertilization and non-fertilization treatments were established under the two land-uses. Therefore, four treatments with four replicates were employed, referred as citrus orchard with fertilization (OF), citrus orchard without fertilization (ONF), rice paddy with fertilization (PF), and rice paddy without fertilization (PNF). The planting density was 235,000 ha−1 and the spacing was 0.25 × 0.17 m in the paddy plots, whereas the converted orchards were planted at a planting density of 833 plants ha−1 with a spacing of 3 × 4 m. Under the two land-uses, fertilizer was evenly sprinkled over the soil surface. The details about the management in the orchards and paddies were described in our early studies (Liu et al. 2016). These fertilization systems, fertilizer rates, and plant population were chosen according to common experiences from local farmers. The detailed description about the intensive cultivations and fertilization managements during our observing period are given in Table 1.

Table 1 Description of management practices used for the paddy and orchard
Full size table

Field soil CH4 fluxes measurements

The exchanges of CH4 fluxes between soil and atmosphere from Aug 2013 to Jul 2014 were measured using static chamber and gas chromatography technology. The chambers were made of stainless steel materials and comprise two separate sections: a cylindrical base frame and a chamber. The base frames (a bottom area of 0.12 m2) without covering the vegetation were installed at depth of 0.05 m in all plots several days before the initial sampling and kept in the places marked. The chambers had a bottom area of 0.12 m2 and heights of 0.39 m (used for orchards) and 0.69 m (used for paddies), respectively. During the gas sampling, chambers were temporarily mounted onto the base frames. All gases were sampled in the 09:00–11:00 a.m. The gas samples were collected daily for 4 days after fertilization and once or twice per week for the remainder of the time. The gases were taken with 100 ml plastic syringes at 0, 10, 20, 30, and 40 min after the chamber was closed. Within 24 h after gas collection, the CH4 concentrations of the gas samples were measured using gas chromatography (Agilent 7890A, Santa Clara, California, USA). Then, the soil CH4 fluxes were computed on the basis of the rate of variations of CH4 concentrations in the chambers, and calculated as the slope of linear or nonlinear regression between concentrations and time. During gas sampling, the temperature of the chamber was recorded using a portable thermometer and then used to calibrate the actual gases concentrations.

Soil sampling and chemical analysis

During Aug 2013 and Jul 2014, soil samples from the top layer (0–10 cm) were sampled once per month or once in every 2 months. In each plot, five soil cores (diameter = 3 cm) were randomly collected and combined together. Then, each bulk soil sample was removed with stones and other materials. After that, the composite soil samples were trisected: the first section was air-dried to analyze soil organic carbon (SOC), total nitrogen (TN), C:N ratio, and pH; the second section was restored at −20 °C for analysis of soil ammonia (NH4 +-N) and dissolved carbon (DOC) contents, and the remaining section was freeze-dried and restored at −80 °C for molecular analysis.

Daily air temperature and precipitation were recorded by the climate station located in this region. Soil temperature and moisture were measured using an automatic data recording system via wireless technology. The pH was measured by a pH meter. Soil SOC, TN, and C:N ratio were determined by an element analyzer (Elementar, Hanau, Germany). For the DOC contents, 20 g of soil was extracted by adding 0.1 L distilled water; after that, the admixture was shaken for 1 h, and filtered, and assessed with a TOC-TN analyzer (Elementar, Laurel, NJ). Soil NH4 +-N contents were analyzed with a flow-injection auto analyzer (Seal AA3, Norderstedt, Germany).

DNA extraction and quantitative real-time PCR

Soil DNA was extracted from freeze-dried bulk soil (0.3 g) using the Mobio Powersoil™ DNA isolation kit (MoBio Laboratories Inc., Carlsbad, CA, USA). The concentration and quality of the extraction was checked by a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA). All the soil DNA extracts were diluted with nuclease-free DI water (DNA:water = 1:5) to reduce potential PCR inhibition and preserved at −80 °C.

Quantitative real-time PCR (qPCR) was applied to quantified the copy numbers of methanogenic mcrA gene (primer set ME1 (GCMATGCARATHGGWATGTC)/ME2 (TCATKGCRTAGTTDGGRTAGT)) (Hales et al. 1996) and methanotrophic pmoA gene (primer set A189F (GGN GAC TGG GAC TTC TGG)/mb661 (CCG GMG CAA CGT CYT TAC C)) (McDonald et al. 2008). The copy numbers were determined by qPCR using an iCycler iQ2 Thermocycler (Bio-Rad Laboratories, Hercules, CA, USA). All the DNA samples were run in triplicates in a final volume of 25 μl reaction system containing 12.5 μl of SYBR® Premix Ex Taq™ (TaKaRa Biotechnology, Dalian, China), 0.5 μl of each primer (10 μM), 2 μl of the diluted DNA (1–10 ng) and 9.5 μl of nuclease-free DI water. In order to generate a standard curve, individual clones with accurate inserts were cultured in Luria–Bertani medium and the plasmid DNA was extracted, purified, and quantified. Plasmid DNA was prepared in a 10-fold dilution series to yield a standard curve covering six orders of magnitude (102 to 108 copies) per assay. The qPCR amplification efficiencies were in the range of 95–100%, and the R 2 was 0.995–0.998. Besides, dissociation curves and gel electrophoresis were employed to confirm whether the amplified DNA fragments were of correct size. The final copies of the methanogenic archaea and methanotrophic bacteria were determined by calculation against the plasmid DNA copies.

Terminal restriction fragment length polymorphism (TRFLP), cloning, and sequence analysis

The methanogen and methanotroph community compositions were determined using soil samples collected in Jul 2014. The PCR amplification was carried out with the same primer pairs as those used for qPCR assays, with the forward primers being labeled with 6-carboxyfluorescein. Then, the PCR products obtained were purified by the PCR purification kit (Promega, San Luis Obispo, USA). Subsequently, the fluorescent-labeled PCR products were digested using the restriction enzymes Eco311 and MspI (Takara Biotechnology, Dalian, China), respectively, for the mcrA and pmoA gene products at 37 °C for 3 h, followed by inactivation at 65 °C for 20 min. The digests were purified by ethanol precipitation and re-suspended in 20 μl of water. Then, 10 μl of the digest were mixed with 15 μl of HiDiformamide and 0.3 μl of internal DNA fragment length standard (Applied Biosystems). The terminal restriction fragments (T-RFs) were calculated with the GeneMarker software package (version 2.2.0). The relative high abundance T-RFs (>1%), which fell in the range of 0–500 bp, were employed for analysis.

Two clone libraries were acquired using the mcrA and pmoA gene fragments retrieved from the PNF and ONF treatments, respectively. The primers used for PCR amplification were the same as those employed earlier but not fluorescent labeled. The purified PCR products were ligated with pGEM-T Easy Vector (Promega, Madison, WI, USA) and transformed into Escherichia coli JM109 cells (Takara Biotechnology, Dalian, China). Subsequently, approximately 50 clones were chosen randomly and sequenced with an ABI 3730 sequencer (Applied Biosystems). Then high-quality sequences were compared to the genes in the NCBI database and examined for chimeras manually, and sequences with alignment splits were removed. The sequences were aligned and two phylogenetic trees were acquired with neighbor-joining tree analysis (1000 bootstrap replicates) by MEGA 6 software (Tamura et al., 2013). To match every clone with a particular T-RF, the sequences acquired were run in REMA software to determine the size of the T-RF (Szubert et al. 2007). The whole sequences acquired in our work were submitted to the GenBank database with the following accession numbers: KX130584–KX130599 for the mcrA gene, and KX077180–KX077192 and KX130600 for the pmoA gene.

Statistical analyses

Repeated measures ANOVA were applied to examine the differences in soil properties, abundance of mcrA and pmoA genes, and CH4 fluxes among the four treatments. Repeated measures ANOVA were also conducted to examine the effects of land-use conversion, fertilization as well as their interactions on soil properties, abundance of mcrA and pmoA genes, and CH4 fluxes. Regression and multiple regression analyses were employed to examine the relationships among CH4 fluxes, abundance of mcrA and pmoA genes, and soil properties. All the statistical analyses were performed using the SPSS software (SPSS Inc., Chicago, IL, USA). Redundancy analysis (RDA) of the T-RFLP profiles was performed to summarize the variations in the methanogen and methanotroph communities that could definitely be explained by the soil properties using CANOCO software (Microcomputer Power, Ithaca, NY, USA).

Results

Soil properties

The soil moisture of the paddy plots was greater than that of the orchards, while the reverse patterns were observed for soil temperature (Table 2). Both land-use conversion and fertilization distinctly reduced soil pH (P < 0.001), with the lowest pH being detected in the OF treatment (Table 2). Moreover, land-use conversion from paddies to orchards distinctly increased the SOC contents, TN contents, and C:N ratio (Table 2, P < 0.05). Fertilization did not affect soil moisture, soil temperature, SOC, and C:N ratio under the orchards and paddies (Table 2, P > 0.05). Soil TN contents were affected by fertilization (Table 2, P < 0.05). The soil DOC contents were significantly lower in the orchard plots when compared with those in the paddy fields (Table 2). Fertilization significantly affect the soil DOC contents (Table 2, P < 0.05). In addition, fertilization significantly increased soil NH4 +-N contents, when compared with unfertilized plots (Table 2, P < 0.05).

Table 2 Effects of land-use conversion and fertilization on soil properties
Full size table

Environmental conditions and soil CH4 fluxes

The annual precipitation was 1424.8 mm, which is below than the long-term average (1509.0 mm). Rainfall between March and June accounted for 57% of the annual precipitation, with the largest and smallest monthly rainfall occurring in May and February, respectively (Fig. 1a). The average annual air temperature was 19.3 °C, with maximum and minimum daily air temperature being 31.9 °C (August 2013) and 0.02 °C (February 2014), respectively (Fig. 1a).

Fig. 1
figure 1

Seasonal variation of the daily air temperature and precipitation (a), and soil CH4 fluxes in the orchard plots (b) and paddy fields (c) during the observation period. The values are means with standard errors. Black arrows indicate the fertilizer applied to OF; gray arrows indicate fertilizer applied to PF. Horizontal bars indicated the period of flooding in the paddy plots. Treatments are OF, orchard with fertilization; ONF, orchard without fertilization; PF, paddy with fertilization; PNF, paddy without fertilization

Full size image

The amounts and patterns of CH4 fluxes were significantly different between the two land-uses (Fig. 1b, c). The CH4 emission fluxes from the conventional paddy plots exhibited distinct seasonal variations, peaked at almost 20 days after flooding and declined and became stabilized on the day before harvest (Fig. 1c). In the converted orchard plots, most of the CH4 exchange fluxes were negative and higher amounts of CH4 were oxidized in summer when the soil moisture was moderate and temperature was higher (Fig. 1a, b). The average CH4 fluxes for the paddy fields were significantly greater than those for the orchard plots during the observation period (Table 3, P < 0.05). However, the average soil CH4 flux between unfertilized and fertilized treatment were not significantly different under the orchard plots and paddy field (Table 3, P > 0.05).

Table 3 Effects of land-use conversion and fertilization on mean soil CH4 fluxes, mcrA, and pmoA abundance
Full size table

Abundance of methanogens and methanotrophs

The mcrA copy numbers ranged from 2.55 × 106 to 2.53 × 107 and from 2.26 × 106 to 6.40 × 107 copies g−1 dry soil, while the pmoA copy numbers ranged from 1.42 × 105 to 4.95 × 105 and from 4.81 × 105 to 4.76 × 106 copies g−1 dry soil in the orchard and paddy soils, respectively (Fig. 2a, b). As for the paddy fields, the mcrA copy numbers were highest in June 2014, whereas maximum pmoA copy numbers were observed in November 2013 (Fig. 2a, b). In the orchard plots, the highest mcrA and pmoA abundance were noted in May 2014 and June 2014, respectively (Fig. 2a, b). Furthermore, the mcrA and pmoA abundance was statistically different between the land-uses, with lower mean mcrA and pmoA abundance detected in orchard plots, irrespective of the fertilization treatment (Table 3). However, although the fertilization treatment showed lower mcrA and pmoA abundance when compared with the non-fertilization treatment in both orchard and paddy plots, the difference between them was not statistically significant (Table 3).

Fig. 2
figure 2

Seasonal variation of mcrA (a), and pmoA abundance (b) under the four treatments. Values are means with standard errors. Black arrows indicate fertilizer applied to OF; gray arrows indicate fertilizer applied to PF; the arrows in (a) applied to (b). Treatments are OF, orchard with fertilization; ONF, orchard without fertilization; PF, paddy with fertilization; PNF, paddy without fertilization

Full size image

Table 4 summarizes the relationships among soil mcrA and pmoA abundance and soil properties. For the orchard plots, the coefficients of the best analysis (i.e., CH4 fluxes and SOC contents) were used in multiple regression study against pmoA abundance, and an R 2 of 0.57 was obtained. However, no significant correlation between mcrA abundance and soil properties was observed. For the paddy plots, the best multiple regression result for the pmoA abundance including soil DOC and CH4 fluxes, with an R 2 of 0.53, whereas the best fit of soil properties against mcrA abundance including soil moisture, pH, and soil temperature, with an R 2 of 0.86.

Table 4 Multiple regressions among CH4 fluxes, copy numbers of mcrA gene and pmoA gene, and soil properties
Full size table

Community structure of methanogens and methanotrophs

In total, six T-RFs were considered to be main peaks in the T-RFLP spectrum of mcrA gene (Fig. 3a). The clone sequences were analyzed to associate the T-RFs with the respective methanogens, which suggested that five T-RFs could be specified to a single lineage, i.e., T-RFs of 97, 292, and 139 bp to Methanosarcina and T-RFs of 195 and 280 bp to Methanocella (Fig. S1). However, T-RF of 329 bp was characteristic of Methanosarcina, Methanoregula, and Methanosaeta (Fig. S1). In the orchard plots, members of Methanosarcina and Methanocella were observed (Fig. 3a), whereas in the paddy fields, members of Methanosarcina, Methanocella, Methanoregula, and Methanosaeta were detected (Fig. 3a). Moreover, the community compositions of methanogens were almost similar between fertilization and non-fertilization treatment under the two land-uses (Fig. 3a).

Fig. 3
figure 3

Community structures of methanogens and methanotrophs based on T-RFLP analysis targeting mcrA (a) and pmoA (b) genes in the four treatments. Values are means with standard errors. Treatments are OF, orchard with fertilization; ONF, orchard without fertilization; PF, paddy with fertilization; PNF, paddy without fertilization

Full size image

Seven T-RFs were considered to be main peaks in the T-RFLP profiles for the pmoA gene (Fig. 3b). Four T-RFs were specified to a single lineage, such as T-RF of 130 bp to Methylococcus, T-RF of 13 bp to Methylobacter, and T-RFs of 401 and 53 bp to Methylocystis, whereas T-RFs of 19 and 295 bp were assigned to type II methanotrophs (Methylocystis and Methylosinus) (Fig. S2). However, T-RF of 76 bp could not be characteristic of any sequences (Fig. S2). In the four treatments, the T-RF of 295 bp showed the highest relative abundance (Fig. 3b). Furthermore, land-use conversion from paddies to orchards reduced type I methanotrophs (Methylococcus and Methylobacter), but increased type II methanotrophs (Methylocystis and Methylosinus) (P < 0.05). Moreover, fertilization significantly affected T-RF of 13 bp under the orchard plots (P < 0.05).

The RDA analysis was conducted by combing compositions of soil methanogen and methanotroph to show the relationships among soil properties and microbial compositions (Fig. 4). The results showed that the first (RD1) and second (RD2) axes explained 89.8 and 2.3% of the total variations of the community structures of methanogens and methanotrophs, respectively (Fig. 4) Furthermore, the results indicated that the orchard plots were best characterized by soil temperature, SOC, TN, C:N ratio, and Methylocystis and Methanosarcina, whereas the paddy fields were characterized by soil CH4 fluxes, DOC contents, pH, soil moisture, and Methanoregula, Methanosarcina, Methylobacter, Methanosaeta, Methylosystis. The different cluster of symbols in Fig. 4 indicates that land-use conversion but not of fertilization significantly induced the variation of microbial communities.

Fig. 4
figure 4

RDA showing the relationships among land-uses (symbols), soil properties (black vectors), and relative abundance of methanogens (gray vectors) and methanotrophs (light gray vectors). The data are shown for four different land uses, namely, orchard with fertilization (circles); orchard without fertilization (square); paddy with fertilization (triangles); and paddy without fertilization (stars). SOC soil organic carbon contents, DOC soil dissolved carbon contents, SM soil moisture, ST soil temperature, TN total nitrogen contents, C/N C:N ratio

Full size image

Correlations between CH4 fluxes and various variables

In the orchard plots, soil CH4 fluxes could not be fitted to mcrA abundance (Fig. 5a), but were negatively correlated with pmoA abundance (Fig. 5b). Meanwhile, the relationships between soil CH4 fluxes and mcrA:pmoA ratio well fitted the quadratic equations (Fig. 5c). In the paddy plots, the relationship between CH4 fluxes and mcrA abundance and mcrA:pmoA ratio fitted the linear equations (Fig. 5d, f). The multiple regression analysis confirmed the correlations between CH4 uptakes and pmoA abundance in the orchards (Table 4). Meanwhile, the DOC contents and mcrA abundance were the main factors controlling soil CH4 fluxes from the paddies (Table 4).

Fig. 5
figure 5

Relationships between soil CH4 fluxes and mcrA abundance, pmoA abundance, and mcrA:pmoA ratio in orchard plots (a–c) and paddy fields (d–f)

Full size image

Discussion

Soil CH4 fluxes

The seasonal variations of soil CH4 flux in this study exhibited similar patterns to those documented in the paddy fields and upland plots (Eusufzai et al. 2010; Wang et al. 2015). The paddy fields showed higher CH4 fluxes than those from orchard plots, especially in the flooding period (Fig. 1b, c), which was consistent with the argument that drainage would bring about the disappearance of CH4 emissions but appearance of CH4 uptakes (Jäckel et al. 2001; Nishimura et al. 2008). It is noteworthy that several reasons might contribute to the reduced CH4 emissions induced by such land-use conversion. First, CH4 emissions could have probably been caused by anaerobic soil conditions induced by flooding depth on soil surface (data are not shown). It has been reported that longer period of flooding in paddy fields significantly leads to anaerobic soil conditions, in contrast to the aerobic soil condition in upland orchard fields (Inubushi et al. 2003; Nishimura et al. 2008). Second, soil DOC is an important parameter of CH4 emissions from paddy soils because the available carbon substrates supply energy sources and consequently enhance methanogenic activity (Zou et al. 2005; Liu et al. 2015b). According to the common experiences, some crop residue was typically incorporated in the paddies, which would result in more soil DOC contents (Table 2). Lastly, land-use conversion could markedly influence the abundance and community structures of methanogens and methanotrophs, thus determining the soil CH4 fluxes (Jäckel et al. 2001; Liu et al. 2015a). From the microbiological point, CH4 fluxes are largely determined by methanogenic archaea and methanotrophic bacteria. Therefore, changes in the CH4 fluxes induced by land-use conversion or fertilization would be inherently caused by the responses of these two functional groups. In the following paragraphs, we will illuminate how these factors affect the soil CH4 fluxes.

Although fertilization could reduce CH4 fluxes in some cases (Liu et al. 2016; Zhou et al. 2016), it might increase or fail to affect soil CH4 fluxes depending on the fertilizer rate (Fan et al. 2016; Reay and Nedwell 2004; Wang et al. 2015). As it is well known that various factors lead to the controversial response of soil CH4 fluxes induced by fertilization. On the one hand, an increase in denitrification intermediates (N2O, NO, NO2 −) induced by fertilization can inhibit methanogens (Bao et al. 2016). On the other hand, NH4 + could suppress CH4 oxidation by binding to the particulate methane monooxygenase and reacting with it (Schimel 2000). Nevertheless, nitrogen fertilizers were shown to be advantageous for the growth of methanogens and some methane-oxidizing bacteria (Bodelier et al. 2000; Lu and Conrad 2005). In this study, fertilization did not significantly affect the abundance of mcrA and pmoA gene (Table 3), which might be explained by that the inhibiting and stimulating effects of fertilization on the methanogens and methanotrophs were almost equal. Similarly, Dan et al. (2001) proposed that fertilization stimulated both CH4 production and CH4 oxidation, resulting in no significant effect on CH4 fluxes.

Abundance and community compositions of methanogens and methanotrophs

In the present study, the abundance of methanogens was significantly lower in the orchard plots than that in the paddy plots (Table 3), which means that land-use conversion from paddies to orchards decreased the abundance of methanogenic archaea. Recently, Liu et al. (2015a) also observed significant decreases in the abundance of methanogens as a result of a relatively lengthy duration of upland conditions in the field study. However, methanogens have been shown to survive well without any distinctive changes in the copy numbers under upland crop cultivation for 7 months (Ma and Lu 2011). These findings together confirmed our previous results that soils were CH4 emissions sources in the early days after land-use conversion from paddies to orchards, and became CH4 uptakes in the second conversion stage (Liu et al. 2016). Nevertheless, the results on the existence of methanogenic archaea in upland plots are highly controversial (Angel et al. 2012; Hu et al. 2013; Lammel et al. 2015). In our results, methanogenic archaea were detected in the newly converted orchard plots (1 year). A possible explanation might be that the presence of detoxification enzymes, such as superoxide dismutase and catalase in some members of methanogens, defended them from harmful products derived from oxygen (Brioukhanov et al. 2000). Therefore, a long-term study on the abundance of methanogenic archaea in an age-sequence of converted orchards should be conducted to clearly show how long they would be survived after land-use conversion from paddies to orchards.

The abundance of methanotrophs in the orchard plots was expected to be higher than that in the paddy fields, as that methanotrophs were considered to be more common in upland fields (Singh et al. 2009; Wang et al. 2015; Lammel et al. 2015). Surprisingly, the abundance of methanotrophs was decreased after land-use conversion from paddies to orchards (Table 3), demonstrating that the abundance of methanotrophs was actively increased in the paddy plots. These results of no significant increases in the methanotrophs abundance in orchard plots were in well agreement with those stated by Jaatinen et al. (2004) who observed higher CH4 oxidizing potential in a Finnish mire than in a well-drained soil. Similarly, Mohanty et al. (2016) also proved that methanotrophs were generally activated at high CH4 concentrations in the culture experiment. Moreover, the significant positive correlations between pmoA abundance and CH4 fluxes in paddy plots confirmed the stimulation of pmoA by CH4 concentrations in the paddy soils (Table 4). Furthermore, although the abundance of mcrA and pmoA differed with time, there was no clear response to fertilization at subsequent “pre-fertilization” date (Fig. 2a, b), because our sampling time was not at each “fertilization” date. These variations across time suggested that temporal dynamics were more important in structuring the microbial processes in the study site.

Our study showed that Methanosarcina and Methanocella were observed in the orchard plots and paddy fields, whereas Methanosaeta and Methanoregula were just detected in the paddy fields (Fig. 3a). Those results suggested that land-use conversion from paddies to orchards resulted in a rather poor diversity of methanogens, which were in well agreement with the statement of former results (Liu et al. 2015a; Scavino et al. 2013). Generally, Methanosarcina and Methanocella were the most abundant methanogens in paddy fields, probably due to their tolerance to oxygen (Angel et al. 2011). However, methanogens belonging to the family Methanosarcina were also detected in the orchard plots owing to their ability to utilize a range of different precursors for methanogenesis (Jetten et al. 1990). Methanocella were known for their ability to withstand toxic oxygen populations (Erkel et al. 2006), which made them better adapted to both paddy and orchard soils (Fig. 3a). In addition, it has been reported that a relatively shorter period of drainage of 7 months may be less harmful to the methanogen community structure (Ma and Lu 2011). Fertilization were shown to have no significant effects on the community compositions of methanogens, which corresponded well with the results of Ho et al. (2016) who proposed the theory of site history and the “resting” community.

Land-use conversion increased type II methanotrophs (i.e., Methylocystis and Methylosinus), and decreased type I methanotrophs (i.e. Methylobacter and Methylococcus) (Fig. 3b). In a study on three land-use types, the community compositions of methanotrophic bacterial were observed to be strongly affected by land-use changes (Zheng et al. 2010). Former studies on the field study indicated that the methanogen communities did not change in the fallow period of rice paddies (Krüger et al. 2005; Ma and Lu 2011). However, there was no field observation on the methanotrophs community induced by land-use conversion from paddies to orchards, thus more studies are urgently needed. In general, soil properties select particular groups of microorganisms potentially by shaping the ecological niches such as substrates concentration and soil properties of the targeted microorganisms (Fig. 4). In addition, although fertilization was found to decrease the T-RF of 13 bp in the orchard plots, no consistent patterns could be generalized with respect to the effects of fertilization on methanotrophs (Alam and Jia 2012).

Correlations between soil CH4 fluxes and methanogens and methanotrophs

The conventional explanation for soil CH4 fluxes induced by land-use conversion included soil condition and physical properties (Inubushi et al. 2003; Nishimura et al. 2008). In the present study, soil CH4 fluxes were significantly related to the abundance of mcrA and soil DOC contents (Table 4), indicating that CH4 production and carbon substrates were directly related to the CH4 emissions from paddy soils. Meanwhile, mcrA and pmoA genes play an obvious mechanistic role in the actual CH4 fluxes from paddy soils (Lee et al. 2014), as that significant positive relationships between CH4 fluxes and mcrA/pmoA were observed (Fig. 5f). Moreover, it is well known that most of the methanogens could utilize H2/CO2 for CH4 production, while Methanosaeta can utilize acetate (Metje and Frenzel 2007). Furthermore, Methanosaeta have been described as slow-growing microorganisms and have a much lower acetate threshold (Jetten et al. 1990). In general, CH4 productions from acetate are much higher (typically 67%) than those from H2/CO2 (generally 33%) (Metje and Frenzel 2007). Thus, in the present study, the survival of Methanosaeta in the paddy plots might have resulted in more CH4 emissions, which might be an alternative explanation for more CH4 fluxes from paddy soils.

Intuitively, the potential CH4 uptakes in the uplands can be attributed to the stimulation of methanotrophs. Though abundance of methanotrophic bacteria was decreased after land-use conversion (Table 3), the soil CH4 uptakes of orchard plots were negatively correlated with the pmoA abundance (Fig. 5b). The result suggested that pmoA abundance could dominate soil CH4 uptakes in the upland plots, which was in line with former studies (Singh et al. 2009; Wang et al. 2015). Moreover, despite the pmoA abundance decreased (Table 3), the type II methanotrophs increased in the orchard plots (Fig. 3b), which could result in the CH4 uptakes. Singh and Tate (2007) also reported that samples of forest soil, where type II methanotrophs were predominant, were at least 10 times more efficient in oxidizing CH4 than those of pasture soil in which type I methanotrophs were predominant. Thus, effective CH4 oxidation induced by land-use conversion from paddies to orchards could be owing to the significant stimulation of type II methanotrophs. Moreover, it was suggested that type I methanotrophs participated in CH4 consumption only from 50 ppmv methane upward, whereas type II methanotrophs consumed at lower methane concentrations (Knief et al. 2006). Therefore, we could speculate that the methanotrophs of paddy soils mainly oxidized the soil CH4 productions, and those in the orchard soils consumed the atmospheric CH4. In addition, it has been reported that fertilization did not affect CH4 consumptions of the soil where type I methanotrophs predominant, but inhibits CH4 uptakes in the soil predominantly comprising type II methanotrophs (Mohanty et al. 2006). Conversely, in the present study, the soil was clearly dominated by type II methanotrophs, the CH4 consumption was not inhibited by fertilization, as some others reported that a few methanotrophs genera are shown to be able to withstand nitrogen species (Noll et al. 2008).

Conclusion

Taken together, our results suggested that land-use conversion from paddies to orchards change soils from emission sources to sink of atmospheric CH4, whereas fertilization did not significantly affect soil CH4 fluxes under the two land-uses. The conversion of paddies to orchards decreased the abundance of methanogens and methanotrophs, resulted in poor diversity of methanogens community, and increased type II methanotrophs. The higher CH4 emissions from paddy soils could mainly attribute to the increased soil DOC contents, more methanogens, and compositions of methanogenic archaea. The compositions of methanotrophs and soil conditions together drive the ability of oxidizing atmosphere CH4 of orchard soils. Moreover, no distinct changes in methanogens and methanotrophs could help to explain the stable response of CH4 fluxes to fertilization. Overall, we provide a certain extent evidence of microbial mechanisms for the soil CH4 fluxes induced by land-use conversion and fertilization. Moreover, the findings of CH4 fluxes and dynamics of methanogens and methanotrophs following land-use conversion and fertilization could provide new insights into the soil carbon biogeochemical cycles in red soils.