Methanotrophic Communities in Brazilian Ferralsols from Naturally Forested, Afforested, and Agricultural Sites

Abstract

Conversion of forests to farmland permanently lowers atmospheric methane consumption due to unresolved reasons. Alphaproteobacterial methanotrophs were predominant in forested soils and gammaproteobacterial species were predominant in farmland soils of subtropical ferralsols in Brazil. The capability of atmospheric methane consumption was obliterated in farmland soils, suggesting a shift from oligotrophic to copiotrophic species.

Aerobic methanotrophic communities in aerated soils are the largest biological sink for atmospheric methane (CH₄) (3, 4, 10, 17). Forest soil communities exhibit the highest consumption rates on a global scale (4, 26). The conversion to farmland lowers the sink capacity (26, 28, 29), but it can be restored by afforestation (27). Sink reconstitution is faster in tropical regions than in temperate regions (12, 15). The impact of this land use change on atmospheric CH₄-consuming communities has scarcely been addressed (21, 25). The objective of the present study was to gain insight into CH₄ oxidation kinetics and the composition of methanotrophic communities in soils from a naturally forested site, an afforested site, and two farmland sites from a subtropical region in South America (Mata Atlântica, Brazil).

Soils from an area in Brazil to the west of São Paulo and close to Caucaia do Alto (46°55′W to 47°06′W, 23°39′S to 23°47′S) were sampled in April 2005. The soil type at any site was a ferralsol (8a). The natural forest site was not managed for over 100 years (pH 3.6), whereas the reforested site was restored 20 years ago (pH 3.8). The humus layer was removed before sampling the forest soils. A nearby located conventionally farmed acre (pH 6.1) and an organically farmed acre (pH 6.2) were also sampled. Both sites were forests 50 to 100 years ago, and the forest sites were part of a formerly continuously forested region. Soils were sampled in three spatial replicates from the top 10 cm. Samples were pooled, transported within 2 weeks to Germany, and manually homogenized, and aliquots were stored at −20°C for molecular analyses and at 2°C for activity analyses.

Atmospheric CH₄ consumption and kinetic parameters.

Ten grams of sieved soil samples was incubated in 150-ml gastight vials in triplicate at 22°C. The headspace contained air with 1.75 parts per million by volume (ppmv) CH₄ or a specific mixing ratio adjusted by adding pure CH₄ (99.000%; Riessner-Gase, Germany). Headspace mixing ratios were measured over a period of 6 h using gas chromatography with a flame ionization detector (GC-FID) (6). Soil samples from the forest sites had the potential to oxidize atmospheric CH₄, whereas neither of the farmland soil samples did (Table 1). Measured atmospheric CH₄ consumption rates were lower than those obtained from other Brazilian forest soil samples (20 to 33 pmol g [dry weight] of soil⁻¹) (24), probably due to the long transport. Michaelis-Menten kinetics was determined according to a previous study (6). Apparent half-saturation constants [K_m(app)] (Table 1) of the two forest soils were similar to those of other Brazilian forest soils (9). Thus, soils from these two forest sites might have been sinks for atmospheric CH₄. Both of the farmland soil samples had higher maximal flux rates [ν_max(app)] than the forest soil samples and lacked atmospheric CH₄ consumption (Table 1), suggesting a soil indigenous CH₄ source that provided in situ mixing ratios above atmospheric levels.

TABLE 1.

Methane oxidation kinetic parameters of soil samples

Soil sample	Methane oxidation kinetic parametersa
	vatm (pmol g [dry wt] of soil−1 h−1)b	Km(app) (ppm)	vmax(app) (pmol g [dry wt] of soil−1 h−1)
Forest sites
Natural forest	0.03 ± 0.5	89.5 ± 17.8	1.8 ± 0.01
Afforested	0.03 ± 0.3	260.0 ± 42.9	5.9 ± 0.04
Farmland sites
Organic farming	ND	—	81.0 ± 0.52
Conventional farming	ND	—	53.0 ± 0.31

^aErrors shown are standard errors of the means. ND, not detectable, —, not measured.

^bv_atm, velocity at 1.75 ppmv methane mixing ratio.

Analyses of methanotrophic community structure.

From every pooled soil sample, a pmoA gene (encoding the hydroxylase of the particulate CH₄ monooxygenase [pMMO]) library was set up. Phospholipid fatty acid (PLFA) stable isotope probing (SIP) at 0.5% CH₄ was performed with previously unfrozen soil samples (forest sites) and frozen soil samples (farmland sites). DNA was extracted in four replicates by using a bead beating protocol (Fast Spin kit; Bio 101). pmoA genes were amplified from pooled DNA extracts using primers A189 and A682 (11), according to Ricke and coworkers (22). pmoA PCR products were cloned using the pGEM-T vector system II (Promega, Germany). Gene inserts were sequenced with the services of Macrogen Inc. (South Korea). Resulting pmoA sequences were phylogenetically affiliated using ARB (18) and implemented algorithms (neighbor joining; Tree-Puzzle) (see methods in File S2 of the supplemental material) (8, 19). Every library exhibited coverage of more than 90%. Thus, the analyzed number of sequences (Table 2) was sufficient to assess the expected genotypes. Sequences may be found in the EMBL database (accession numbers FN394081 to FN394214). The gene mmoX (encoding the hydroxylase subunit of the soluble MMO) could not be amplified (primers mmoXf945 and mmoXB1401) (1, 15), indicating that Methylocella species were not detectable, as well as type I or type II methanotrophs harboring this enzyme.

TABLE 2.

Frequency of genotypes in pmoA gene libraries

pmoA gene library	Frequency of genotypes (%)a
	Forests		Farmland
	Natural (n = 37)	Afforested (n = 35)	Organic (n = 31)	Conventional (n = 30)
Methylococcaceae (type I/X)
Methylococcus	ND	ND	ND	36.7
Crenothrix polyspora-like	ND	ND	71.0	ND
Methylocystaceae (type II)
Methylocystis	ND	ND	ND	63.3
Beijerinckiaceae related (type II)
USCα	89.2	94.3	6.5	ND
Cluster 5	5.4	2.8	ND	ND
Ambiguous affiliation to pmoA or amoA
Cluster 2	ND	ND	6.5	ND
Cluster WC306-54	5.4	2.8	ND	ND
Ammonia-oxidizing bacteria (Betaproteobacteria)	ND	ND	12.9	ND

^an, total number of phylogenetically analyzed gene inserts per library; ND, not detectable.

Four aliquots (10 g) of sieved fresh soil were mixed with sterile quartz sand to improve labeling efficiency, used to fill gastight 150-ml vials, and incubated with 0.5% ¹³CH₄ at 30°C. The gravimetric water content was adjusted to 30 to 40% by addition of sterile water. Per site and ¹³C treatment, four incubations were run that were individually stopped at 2, 6, 13, or 30 days. CH₄ concentrations were monitored by a GC-FID. Retrieved soil samples were immediately stored at −80°C until PLFAs were extracted and subjected to GC-combustion-isotope ratio mass spectrometry (GC-C-IRMS) (see methods in File S2 of the supplemental material). Absolute ¹³C incorporation, referred to as ng PLFA (see Table S1 and the methods in File S2 of the supplemental material), was converted to δ¹³C values for a better comparison of the samples (see Fig. 2).

FIG. 2.

δ¹³C values of detected PLFA at 2 (black), 6 (light gray), 13 (dark gray), and 30 (white) days from the natural forest (A), reforested (B), organically farmed (C), and conventionally farmed (D) soils. PLFAs are ordered according to their retention times.

Methanotrophic communities in forest and agricultural soils.

pmoA gene libraries from the naturally forested and afforested sites were dominated by genotypes next related to Beijerinckiaceae. In both gene libraries, the genotype USCα was predominant, and both clusters 5 and WC306-54 were detected. WC306-54-related genotypes could not be functionally assigned to the pmoA or amoA gene (hydroxylase of ammonia monooxygenase) due to their intermediate phylogenetic positions (Fig. 1; Table 2). PLFA ¹³C incorporation patterns were similar in soil samples from both forest soils (Fig. 2A and B). PLFAs a15:0, i16:0, 16:1ω5c, 18:1ω9c, and cy19:0 were significantly labeled in both soils (δ¹³C > 5‰) (Fig. 2A). ¹³C incorporation of PLFAs from forest soils was low compared with that from farmland soils, although the same CH₄ concentrations were used (see Table S1 in the supplemental material). This indicates that primarily the same species that were in situ active and detectable by pmoA utilized ¹³CH₄. It cannot be excluded that type I methanotrophs were also present, since a 16:1 PLFA was labeled. Type I methanotrophs were not present in gene libraries, suggesting that the 16:1 PLFA indicated cross-feeding by nonmethanotrophic soil microbes or type II methanotrophs of the genus Methylocystis (5).

FIG. 1.

Phylogenetic affiliation of PmoA/AmoA sequences retrieved from gene libraries using Tree-Puzzle. OA, organically farmed site; CA, conventionally farmed site; AF, afforested site; NF, natural forest. In parentheses, numbers of similar genotypes found in the same library that share ≥87% sequence similarity on amino acid level. Scale bar represents 0.1 change per amino acid. Rectangles, nodes that were confirmed by an alternative treeing method. Further details are given in the methods in File S2 of the supplemental material.

Communities in farmland soils differed from those in soils from both forest sites, as revealed by pmoA (Table 2). In the agricultural soil under conventional farming, genotypes of Methylocystis and Methylococcus spp. were prevalent. A Crenothrix polyspora-like genotype was frequently found in the organically farmed acre soil. Beijerinckiaceae spp. and cluster 2 were less frequently found (Table 2). ¹³CH₄ incorporation in PLFAs was higher in farmland samples than in forest soil samples (Fig. 2), and consistently, maximal CH₄ oxidation rates were higher in farmland soil samples (Table 1). These observations may indicate that methanotrophic species in farmland soils were adapted to mixing ratios above atmospheric levels. Farmland soil samples were frozen for transport before PLFA SIP. This did not affect the composition of methanotrophic community, since PLFA and pmoA data were basically congruent regarding the prevalence of type I or II methanotrophs. PLFAs known from Methylococcaceae (type I/X) (10) and Methylocystis (5) were significantly labeled; type II-specific ones at a later time point indicated that type II methanotrophs were less abundant or active (δ¹³C > 20,000‰) (Fig. 2C and D). At the end of the incubation period, PLFA 18:1ω7c was detected, with a δ¹³C value above 20,000‰, suggesting a background activity of the present Beijerinckiaceae species.

Conclusions.

The conversion of a natural subtropical forest to farmland led to a loss of atmospheric CH₄ consumption and a shift from Beijerinckiaceae species to Methylococcaceae and Methylocystaceae species. The data also suggest a restoration of the original community and the atmospheric CH₄ sink after afforestation. A similar phenomenon was described by a previous study of tropical soils in Thailand (15). In temperate New Zealand soils, type II methanotrophs were active under a pine forest, whereas type I methanotrophs were active in adjacent pastures. Afforested sites consumed more atmospheric CH₄ than the nonforested ones (25, 27). In contrast to these results, afforestation of a boreal grassland led to reduced atmospheric CH₄ uptake and a reduction of the methanotrophic biomass (20).

The current data suggest that soil capability of atmospheric CH₄ consumption resulted from the dominance of few oligotrophic methanotrophic species (for examples, see reference 16). Methylocystis strains can utilize atmospheric CH₄ (2, 7, 13). However, it is unlikely that the detected Methylocystis spp. in agricultural soils utilized atmospheric CH₄, since the required pmoA2 gene was not detected (2). Consequently, detected Methylococcaceae and Methylocystaceae can be regarded as copiotrophic, whereas detected Beijerinckiaceae represented oligotrophs. Increased CH₄ concentration and availability of nitrogen may alter methanotrophic communities (14, 22). Fertilization may also reduce the abundance of atmospheric CH₄ consuming methanotrophs in soils (20). pH values shifted from approximately 3 to 6 after deforestation. Hence, the sink capacity might have been reduced after deforestation, since oligotrophic species were outcompeted by copiotrophic species due to fertilization, higher organic matter turnover in farmlands, and/or increased pH.