SEARCH

SEARCH BY CITATION

Keywords:

  • methane-oxidizing bacteria;
  • microarray;
  • pmoA;
  • salinity;
  • soil

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References
  8. Supporting Information

Despite their large areas and potential importance as methane sinks, the role of methane-oxidizing bacteria (MOB) in native woodland soils is poorly understood. These environments are increasingly being altered by anthropogenic disturbances, which potentially alter ecosystem service provision. Dryland salinity is one such disturbance and is becoming increasingly prevalent in Australian soils. We used microarrays and analysis of soil physicochemical variables to investigate the methane-oxidizing communities of several Australian natural woodland soils affected to varying degrees by dryland salinity. Soils varied in terms of salinity, gravitational water content, NO3-N,SO4-S and Mg, all of which explained to a significant degree MOB community composition. Analysis of the relative abundance and diversity of the MOB communities also revealed significant differences between soils of different salinities. Type II and type Ib methanotrophs dominated the soils and differences in methanotroph communities existed between salinity groups. The low salinity soils possessed less diverse MOB communities, including most conspicuously, the low numbers or absence of type II Methylocystis phylotypes. The differences in MOB communities suggest niche separation of MOB across varying salinities, as has been observed in the closely related ammonia-oxidizing bacteria, and that anthropogenic disturbance, such as dryland salinity, has the potential to alter MOB community and therefore the methane uptake rates in soils in which disturbance occurs.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References
  8. Supporting Information

Methanotrophs [methane-oxidizing bacteria (MOB)] comprise a distinct group of highly specialized bacteria that are able to use methane as their sole carbon and energy source. They have been detected from diverse environments, including soils, sediments, freshwater and marine habitats (Pester et al., 2004; Abell et al., 2009; Moussard et al., 2009; Antony et al., 2010; Semrau et al., 2010) and anthropogenically created habitats such as landfills (Stralis-Pavese et al., 2004; Hery et al., 2007; Gebert et al., 2009). MOB are important in the global carbon cycle as the largest known atmospheric biological methane sink. Methanotrophs use the enzyme methane monooxygenase (MMO) to catalyse the oxidation of methane to methanol. MMO occurs in either the soluble, cytoplasmic form (sMMO), which is only found in some methanotrophs, or the particulate, membrane bound form (pMMO), which is found in all known methanotrophs, except Methylocella spp. (Dedysh et al., 2000) and Methyolferula stellata (Vorobev et al., 2011). The gene pmoA (encoding the α subunit of pMMO) reflects evolutionary relationships between pmoA-containing taxa and has been successfully used to identify methanotrophs in previous ecological studies (McDonald et al., 2008).

MOB are currently classified into two broad taxonomic groups, the type I and type II methanotrophs, based on cell morphology, phenotypic traits including carbon assimilation pathways, 16S rRNA gene phylogeny, PLFA profiles and intracellular membrane system architecture (Hanson & Hanson, 1996; Bowman, 2006; Trotsenko et al., 2008). Type I methanotrophs (Methylococcaceae) belong to the Gammaproteobacteria and are further divided into the type Ia and type Ib (or type X) methanotrophs. The first of these groups includes the genera Methylomonas, Methylobacter, Methylomicobium, Methylosarcina, Methylosphaera, Methylosoma, Crenothrix, Clonthrix and the second includes Methylococcus, Methylocaldum, Methylohalobius, Methylothermus. The type II methanotrophs (Methylocystaceae) belong to the Alphaproteobacteria and comprise the genera Methylosinus, Methylocystis, Methylocapsa and Methylocella. Recent studies have also demonstrated that anaerobic methane oxidation coupled to denitrification in anaerobic freshwater sediments by Candidatus Methylomirabilis oxyfera (Raghoebarsing et al., 2006; Ettwig et al., 2010) and, further, that extremely acidophilic members of the genus Verrucomicrobium are capable of methane oxidation (Op den Camp et al., 2009).

The ecological roles of different methanotrophic taxa are unclear, but there is some evidence that they occupy specific niches, with factors such as oxygen and methane concentrations, nitrogen, pH, carbon and overlying plant community influencing distribution (Bender & Conrad, 1995; Hanson & Hanson, 1996; Bodelier & Laanbroek, 2004; Bussmann et al., 2006; Noll et al., 2008; Abell et al., 2009; Tsutsumi et al., 2009). For example, (Henckel et al., 2001) suggested that type I methanotrophs react more quickly to changing methane availability, (Bull et al., 2000) suggested that type II methanotrophs may be responsible for high-affinity methane oxidation in well-drained soils, and (Tsutsumi et al., 2009) found that understory plant species composition influenced MOB community structure significantly. Low-affinity MOB are adapted to growth at high methane concentrations, while high-affinity MOB are able to oxidize trace amounts of methane, including that at atmospheric concentrations. Low-affinity methane oxidizers are often found in soils where methane production occurs, are responsible primarily for oxidation of methane produced in their environment and are responsible for the majority of the soil methane sink. High-affinity oxidizers are often found in well-aerated soils, oxidize atmospheric methane (that may be produced elsewhere) and are responsible for a lower, but significant, (5–10%) proportion of the soil methane sink.

Methane oxidation is an important process in many environments and has been investigated most intensively in environments, where methane concentrations may be relatively high (e.g. landfill sites, rice fields). The global soil methane sink is estimated to be 30–60 Tg per year (Dunfield, 2007) and, although much less than the atmospheric sink, is important because it represents the current source-sink imbalance in the global methane budget (Bodelier & Laanbroek, 2004; Knief et al., 2005; Dunfield, 2007). The soil methane sink is responsive to anthropogenic disturbances such as land-use change, acidification and nitrogen deposition (Dunfield, 2007). The role of environmental factors in controlling methanotrophic activity has been investigated in agricultural and other managed systems, but is less well understood in natural environments.

Several environmental parameters (including, moisture, temperature, N supply, salinity, diffusivity) have been found to influence soil methane oxidation (Park et al., 2005; Dunfield, 2007). Soil moisture has been shown to influence methane oxidation rates; low water activity reduces biological activity, while moisture high enough to reduce soil oxygenation also reduces methane oxidation activity (Henckel et al., 2001). Methane oxidation has also been shown to increase with increasing temperature. The influence of nutrient supply on methane oxidation is complex. It has been shown to be both inhibited or stimulated by ammonia and N fertilizers, depending upon the limiting nature of N in the system (Bodelier & Laanbroek, 2004). N supply has also been shown to influence the type of MOB present (Graham et al., 1993). Type II methanotrophs are typically capable of N fixation and may therefore be selected by N limiting conditions. High soil salinity increases osmotic stress, decreases water availability and limits methane oxidation (Dalal et al., 2008). Soil diffusivity has also been shown to effect methane oxidation rates by affecting the flux of O2 available to methanotrophs, whereby soils with lower diffusivity and higher compaction supported lower oxidation rates (Gebert et al., 2011).

The role of different MOB phylotypes in well-drained temperate soils is largely unknown, even though these soils are often a methane sink, oxidize methane directly from the atmosphere and are thought to contribute up to 6% of the global methane sink (IPCC, 2007; Reay et al., 2007; Dalal et al., 2008). For example, native woodland in Australia has been shown to be a major sink for methane, taking up methane significantly faster than managed soils (Livesley et al., 2009). Several other previous studies have investigated methane fluxes in Australian soils (Dalal et al., 2008; Livesley et al., 2009) but have concentrated on managed soils and have investigated seasonal and/or management effects on fluxes, rather than the diversity of the organisms responsible. Studies of MOB from other regions do not necessarily relate to Australian soils, which are highly weathered, relatively dry and nutrient poor (Attiwell et al., 1996). Indeed, even though native woodlands were responsible for the fastest uptake of methane of the different soils sampled by (Livesley et al., 2009), uptake rates were still low relative to those measured in native soils worldwide (Livesley et al., 2009).

Soil salinity is a problem in many countries, causing severe land degradation in both agricultural and native soils. High salinity increases osmotic stress and water availability and limits methane oxidation (Dalal et al., 2008). It is likely, then, that land degradation because of factors such as dryland salinity will influence methane oxidation rates and MOB community structure. Although (Dalal et al., 2008) has recently shown that increased soil salinity decreases methane consumption in Australian soils, the effect of soil salinity on the MOB community has not been investigated. To preserve and manage the soil methane sink, it is not only necessary to understand how methane fluxes change with anthropogenic influence, but also to understand how the MOB community responds.

The aim of this study was to investigate the diversity of MOB in Australian native woodland soils of varying salinity. Given the importance of dryland salinity in processes associated with land degradation, and its potential impacts on soil methane sinks, we were specifically interested in exploring the impact of salinity and other edaphic factors on MOB community structure. Microarray-based molecular analyses of soils were conducted to determine MOB community structure, and soil chemical and physical parameters were determined in soils of varying salinity.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References
  8. Supporting Information

Site descriptions

Soils were collected from 10 natural woodland populations of Acacia stenophylla in north central Victoria, Australia (Fig. 1, Table 1). The sites were chosen deliberately to represent a range of soil salinity levels (Table 1) and classified as slight (ECe 2–4 dS m−1), moderate (ECe 4–8 dS m−1) or high (ECe 8–16 dS m−1) salinity (Marcar & Crawford, 2004). Soils of ECe > 4 are classified as saline and are believed to affect plant performance. ECe refers to the measurement of soil conductivity on a saturated soil extract, while EC1 : 5 is the measurement of soil conductivity on a 1 : 5 w/v soil/water suspension. EC1 : 5 is commonly used to estimate ECe and may be converted to ECe for the soil types we investigated by applying a conversion factor of 9 (Marcar & Crawford, 2004). The salinity range investigated therefore covers much of that encountered in saline soils worldwide but does not extend into those soils classified as extremely saline (ECe > 16).

image

Figure 1. Sites sampled. Dark blue (sites 8, 9, 10) = low salinity, green (sites 1, 2, 5) = moderate salinity and light blue (sites 3, 4, 6, 7) = high salinity soils. At each site three replicate soil samples were taken.

Download figure to PowerPoint

Table 1. Soil chemistry for the three salinity groups described
Salinity groupaSoil TypebGWC (%)pH (CaCl2)Corg (%)NO3-N (mg kg−1)SO4-S (mg kg−1)P (mg kg−1)K (Meq per 100 g)Ca (Meq per 100 g)Mg (Meq per 100 g)EC1 : 5 (dS m−1)
  1. n = 9 (slight and moderate, 12 high). Average (SD). Significantly different salinity groups for each chemical parameter are indicated by unique letters.

  2. a

    Numbers refer to site numbers allocated to each site and identifying sites in Fig. 3.

  3. b

    Soil type determined according to the Factual Key (Northcote, 1984).

Slight (8, 9, 10)Sandy Clay8.3 (4.2) a7.7 (0.3)1.6 (0.8)6.5 (8.6) a59.6 (41.3) a20.7 (10.0)2.7 (1.5)24.1 (14.1)4.2 (1.2) a0.3 (0.1) a
Moderate (1, 2, 5)Medium Clay12.3 (1.2) b7.3 (0.6)2.2 (0.3)17.5 (15.0) ab29.0 (22.1) a13.9 (4.7)2.9 (0.3)17.4 (5.6)10.9 (2.6) b0.6 (0.2) b
High (3, 4, 6, 7)Sandy Clay13.3 (3.8) c7.5 (0.5)2.2 (0.8)47.8 (46.9) b104.0 (47.3) b20.5 (12.2)2.8 (0.9)18.0 (6.4)12.3 (2.8) b1.0 (0.2) c

Sample collection

Samples were collected in October 2008. At each site, 10 cores (0–10 cm, 50 mm diameter) were taken from near the base of individual trees in each of three spatially discrete areas (~10 m2) within the site. These 10 cores were then bulked and mixed to give three replicate samples per site, upon which all analyses (molecular, chemical, physical) were conducted separately (i.e. all analysis comprised 30 samples across 10 sites).

Samples were transported to the laboratory on ice and sieved (2 mm). Soil moisture was determined gravimetrically by drying soil overnight at 105 °C, ~15 g was stored at −20 °C for nucleic acid extraction, and the remaining soils were then air-dried. Five hundred grams samples were sent to Incitec Pivot Nutrient Advantage Laboratories (Werribee, Australia) for soil chemistry analyses listed in Table 1.

Soil chemistry

Soil chemical parameters measured were the following: Organic carbon (Corg), nitrate Nitrogen (NO3-N), sulphate (SO4-S), electrical conductivity (EC1 : 5), phosphorous (P), pH, calcium (Ca), sodium (Na), magnesium (Mg), potassium (K) and were determined by nutrient advantage laboratories of Incitec Pivot Ltd (Australia, http://www.incitecpivot.com.au/). For all analyses, air-dried soil was ground to < 2mm. Corg, was determined by colorimetric analysis following the wet oxidation method of (Walkley & Black, 1934) after extraction with H2SO4 and Na2Cr2O7. NO3-N was determined by automated colorimetric analysis following extraction in a 1 : 5 soil/water solution and shaking for 60 min. Soil pH was determined in 1 : 5 soil/water solution, with and without the addition of CaCl2 (0.01 M) and 60-min shaking. EC was determined on a 1 : 5 soil/water (Miller & Curtin, 2008). NaHCO3 extractable P was determined according to (Olsen & Sommers, 1982), and soil extracted in 1 : 20 soil : 0.5 M NaHCO3 (pH8.5) solution for 30 min. Exchangeable cations (Ca, Mg, Na, K) were measured on centrifuged and filtered samples using ICP-AES after extraction in 1 : 10 soil/ammonium acetate (pH 7.0), 30-min shaking. Gravitational water content (%GWC) was determined by water loss after drying at 105 °C.

Nucleic acid extraction

DNA was extracted in duplicate (2 × 0.25 g) from all 30 soils using the Power Soil DNA Isolation kit (MO-BIO) according to the manufacturer's instructions; duplicates were pooled and quantified spectrophotometerically (Nanodrop ND-1000, Thermoscientific).

PCR

Amplification of pmoA and related sequences (Gammaproteobacteria amoA as well as sequences where the physiology of the encoded protein is uncertain) was amplified using the primers pmoA189 and T7-pmoA682, according to the protocol described previously (Bodrossy et al., 2003).

Microarrays

Microarray analysis was performed as described previously (Stralis-Pavese et al., 2004). In brief, PCR products containing the T7 promoter site were used as template for in vitro transcription, during which Cy3-labelled nucleotides were incorporated. The resulting, single-stranded, fluorescently labelled RNA target was chemically fragmented and hybridized to microarrays spotted in house with the latest pmoA probe set (Supporting Information, Table S2), consisting of 17- to 28-nt-long oligonucleotides. Following overnight hybridization and wash, microarrays were scanned and analysed. Median signal intensities were corrected with background signal and normalized to a positive control to enable comparison between hybridizations. For statistical analysis, these normalized signals were subjected to a second normalization, where the hybridization potential of each probe was considered. In this second normalization step, signals were divided by the reference values for each probe; the reference values were the highest signals obtained while hybridizing the microarray with pure targets (Stralis-Pavese et al., 2004).

Statistical analysis

Differences in environmental variables between salinity groups were analysed by one-way anova. General differences between soil salinity groups were visualized using PCA and the full multivariate data set. MOB community structure was analysed using multivariate methods as outlined in (Clarke & Warwick, 2001); a visual representation of the communities was generated using ordination followed by discrimination of the samples using significance tests and finally exploration of explanatory models. MOB communities were visualized using canonical analysis of principle coordinates (CAP) on Bray–Curtis similarities of square-root transformed microarray data constrained by salinity group. Significant differences between salinity groups were tested using anosim. DISTLM multivariate linear regression and distance-based redundancy analysis (dbRDA) were employed to build models of explanatory environmental variables. Co-varying environmental variables (r2 > 0.7, n = 3) were found and only one such variable was run in the model, while results were interpreted such that this single variable represented the behaviour of all variables with which it was found to co-vary at r2 > 0.7. simper was used to identify phylotypes explaining differences between the salinity group communities. Univariate statistical analyses were conducted using Deducer (Fellow, 2011) in R (R Development Core Team, 2011), and multivariate statistics were conducted using primer 6 and permanova+ (http://www.primer-e.com). Alpha = 0.05 for all statistical significance tests.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References
  8. Supporting Information

Soil chemistry

Sites exhibited a degree of within-site variation for all parameters measured, but on the basis of EC1 : 5, the three salinity groups were clearly distinguishable (Fig. 2, Table 1, Table S1). On this basis, the three salinity groups were significantly different from one another for EC1 : 5 and %GWC (P < 0.001). The salinity of the soils was dominated by Na and Cl, and these two components correlated highly with EC1 : 5 and with one another. Given EC1 : 5 comprises the sum of ion equivalents, we removed it from further analyses (along with Cl), retaining Na. It should be noted that any discussion of Na is inclusive of these two collinear variables. %GWC increased with increasing Na. For NO3-N, SO4-S and Mg, there were also significant differences between the groups (P < 0.001), with the slight and high groups being significantly different and the high salinity group also having higher concentrations of each. For these variables, however, the moderate salinity group shared many characteristics with either or both of the high and slight salinity soils and was often not distinguishable from one or both of them in its own right. The remaining parameters showed no significant differences between the salinity groups (P > 0.05).

image

Figure 2. PCA of salinity group soils sites based on physico-chemical variables. Vectors show major chemical parameters separating soils (R2 > 0.5).

Download figure to PowerPoint

Methanotroph community composition

MOB community composition at 10 sites across a salinity gradient was investigated using microarrays targeting the pmoA gene. The array has been demonstrated to detect methane oxidizers commonly associated with high- and low-affinity methane oxidation, including the high-affinity methane oxidizers of both the USC-gamma and USC-alpha (RA14 clusters). Although the array is limited in its detection to known pmoA sequences, the probe set (Table S2) is comprehensive in that it is able to detect all known phylotypes. Further, the primers employed to amplify the pmoA gene in this study have been shown to cover high- and low-affinity methane oxidizers (Holmes et al., 1995). The relative abundances of the different clades targeted by the probe set on the microarray for each site and salinity group are shown in Fig. 3. Differences between sites and salinity groups were evident, and within-site replicate soils gave consistent results. MOB comprised primarily type II and type Ib methanotrophs, although some type Ia phylotypes were also detected. MOB communities were statistically significantly different between the salinity groups (P < 0.005), with the slight salinity group being different to both the moderate and the high groups (P < 0.05, anosim R = 0.5, 0.28 respectively), while the moderate and high groups could not be separated from one another (P > 0.05, anosim R = 0.04). CAP analysis (Fig. 4a) also showed that MOB communities could be differentiated based on salinity group (P < 0.001). CAP leave-one-out allocation further supported the idea that the slight salinity group is better differentiated than the other groups, with 89% of samples correctly assigned to this group, while 44% were correctly assigned to the moderate group and 67% to the high salinity group.

image

Figure 3. Heatmap of microarray data showing MOB community composition across 10 sites (n = 3). Soil color codes refer to slight (blue), moderate (green) and high (purple) salinity categories as listed in Table 1. Site identifier refers to site number (1–10) and replicate number (a–c).

Download figure to PowerPoint

image

Figure 4. (a) CAP ordination of the MOB community at in all soils sampled, constrained by ‘salinity group’. (b) dbRDA plot showing MOB communities. Vectors show the correlation of soil chemistry variables included in the DITLM model with the MOB community.

Download figure to PowerPoint

Environmental variables potentially explaining MOB community composition were explored using DISTLM and dbRDA. Only chemical variables shown to be statistically significantly different between the salinity groups were included in the analysis (Table 1). Marginal tests indicated that SO4-S (12%), Mg (19%), Na (11%) and NO3-N (9%) all explained significant (P < 0.05) amounts of the variation in MOB communities on their own, but that %GWC did not. The most parsimonious model explaining variation in MOB community structure consisted of Mg, Na, SO4-S and %GWC. This combination of variables explained 90% of the fitted variation and 49% of the total variation (Fig. 4b), indicating that the plot represents the model well and, although there is a considerable amount of variation unaccounted for (51%), the model explained a considerable amount of variability among MOB communities. Interestingly, while %GWC did not explain a significant amount of the variation on its own, it does inform the model significantly when combined with the other variables. NO3-N, on the other hand, did explain a significant amount of variation on its own, but this variation was captured in the other parameters, because NO3-N is not included in the best-fit model generated by DISTLM. Of the variables included in the model, Mg appears to be the most important, as it is the only variable whose R2 does not decrease significantly when fitted to the model last.

Phylotypes contributing to differences in MOB community structure between salinity groups were explored using simper (primer 6). The slight salinity group showed higher abundances of JR3-593, USCG 225, NsNv3 63 and NsNv207 than the moderate group. The high and low salinity groups also showed differences with the low salinity soils having higher numbers of JR3-505, while the high salinity soils showed higher abundances of all Methylocystis phylotypes (Mcy probes), Mha-500, 501–953, 501–286, LP20-644 and TUSC409.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References
  8. Supporting Information

Soils are extremely heterogeneous and as expected the physicochemical properties of the soils in this study exhibited a reasonable degree of within-site variation (Fig. 2). Despite this, all salinity classes were clearly distinguishable by salinity, which was dominated by Na and Cl, and moisture %GWC. Soil moisture, measured as %GWC at time of sampling, is known to vary strongly, both temporally and spatially, and does not necessarily reflect long-term trends in soil moisture. All sampling sites used in this study are, however, in the same climatic zone and experience the same amounts of annual rainfall and rainfall patterns. Small-scale (< m) variation in soil moisture was not assessed in this study, indeed the sampling strategy used (10 bulked cores per sampling area, three areas per ‘site’) ensures that %GWC data represent the broader site characteristics, rather than small-scale variation, which almost certainly exists. Dryland salinity is caused by the water table rising and bringing salt to the soil surface with it. Therefore, periods of water saturation are commonly associated with an increase in dryland salinity and %GWC reflects this at the time of sampling. The soils from the slight and high salinity groups were also distinguishable by nitrate, sulphate and magnesium concentrations. This is not unexpected in soils suffering dryland salinity, as increasing water logging brings with it these soluble species. The salinity groups were not able to be distinguished based on pH, Corg, P, K or Ca concentrations. Soil aggregate stability is determined, in part, by the amount of Mg relative to Ca. Soils higher in Mg often have higher bulk density and lower infiltration and drainage, a result of the greater hydrated radius of Mg ions causing them to hold more water and disperse soil particles. pH and Corg are thought to be among the major determinants of microbial community structure (Fierer & Jackson, 2006), and significantly different levels of these parameters between the salinity groups may have confounded our attempt to investigate the specific effects of salinity on MOB diversity (Bissett et al., 2010). However, given we found no differences in these parameters between salinity groups in these soils, it is unlikely that pH or soil Corg plays significant roles in the differences in MOB communities observed.

The largest differences in MOB communities were seen between the slight and high salinity soils, while the moderate salinity soils appeared to be somewhere between them, sharing some characteristics with either one or both of the high and low salinity groups. The slightly saline soils possessed less diverse MOB communities (Fig. 3) but had higher relative abundance of type Ib phylotypes belonging to the JR3 and USCγ groups. The JR3 are uncultivated methanotrophs originally reported as minor members of rich methanotroph communities of grassland soils (Horz et al., 2005). They are closely related to the USCγ methanotrophs, also uncultivated, which were originally reported from neutral to alkaline soils and where they were thought responsible for high-affinity methane oxidation (Knief et al., 2003). These groups have also been reported in desert soils (Angel & Conrad, 2009), where the JR3 group were thought responsible for the bulk of the methane oxidation observed. These phylotypes have typically been reported from well-drained soils, which include the slightly saline soils of this study. Perhaps, the most conspicuous result regarding the lower salinity soils in this study is the absence of type II phylotypes from two of these soils and lower numbers in the third. Representatives of this group, such as Methylocystis and Methylosinus, have been observed in soils, freshwater and brackish sediments (Bussmann et al., 2004; Abell et al., 2009; Moussard et al., 2009). They are very widely distributed and their apparent absence from some of the soils in this study is surprising, but not without precedent. (Doerr et al., 2010) recently investigated the MOB community of soils from different land-uses in Brazil and reported Methylocystaceae from only one soil type (conventional farmland). They suggested that phylotypes observed from this group were not responsible for oxidation of methane at atmospheric levels in those soils, but experienced higher mixing levels. It is possible that the same is true for the soils sampled in our study.

The high salinity soils also showed higher relative abundances of LP20, Mha500 and TUSC phylotypes. The TUSC methane oxidizers are reported from similar environments (Knief et al., 2005; King & Nanba, 2008) and are thought to fulfil similar roles (high-affinity methane oxidation) to the USCγ phylotypes, so it is interesting that the USCγ phytotypes may perform high-affinity oxidation in the slightly saline soils, while the TUSC clade are present in the more highly saline soils. LP20 phylotypes have been reported from water-saturated soils and sediments, including rice fields, below aquifer soils and littoral zones of lakes (Baker et al., 2001; Lipponen et al., 2004; Horz et al., 2005). The high salinity soils in this study are the result of dry salinity caused by water table changes, so it is possible these phylotypes have travelled with the rising water table into the surface soils. Finally, Mha500 phylotypes are related to Methylohalobius spp., a halophilic group isolated from hypersaline lakes (Heyer et al., 2005) and hydrothermal vents (Nercessian et al., 2005). While the soils in this study possessed much lower salt concentrations than the hypersaline lakes reported in (Heyer et al., 2005), they are still highly saline. Water availability in these soils is likely to be very low (low %GWC combined with high salinity) and these conditions may favour halophilic phylotypes.

Several measured environmental variables, Mg, Na, SO4-S, NO3-N, are among those likely to influence MOB community structure and were able to explain a significant proportion of the variation in the MOB communities. All of these variables affect water activity by influencing soil structure and/or creating osmotic stress. Soil structure may influence mass transfer rates and therefore methanotrophic activity (Gebert et al., 2011). (Bussmann et al., 2004) investigated the influence the single ions Ca2+, Cl, K+, Mg2+, Na+, inline image and the total ionic strength on MOB culturing efficiency [determined by most probable number (MPN)] by manipulating their concentrations in culture media. They found that ionic strength did not have an effect, but Mg2, and SO42 did, MPN counts decreasing with increasing Mg2 and SO42. Despite this effect on culturing efficiency in terms of MPN counts, the effect in terms of ‘diversity’ was less pronounced. While we did not investigate MOB numbers, we found clear differences in community structure apparently related to similar variables, indicating that these may be important in structuring soil MOB communities as well as in attempts to culture them. Physiological adaptations of MOB to high salt and alkalinity have been reported previously (Trotsenko & Khmelenina, 2002ab; Trotsenko et al., 2008) and it appears plausible that similar adaptations also enable niche differentiation between MOB within less extreme environmental gradients.

Interestingly, pH and moisture, which are often associated with changes in community structure (Fierer & Jackson, 2006), were not found to be major determinants of MOB community structure in these soils, although they were found to be significant in the model. One reason for this may be the relatively narrow pH and moisture ranges, compared to the salinity ranges, encountered in these soils. Soil environmental variables such as moisture are temporally and spatially heterogeneous, but many inland Australian soils are relatively dry, all soils were from the same catchment, 10 bulked cores per sample were taken and moisture ranged from only 8–12%. The pH range was also small (7.3–7.7). (Bissett et al., 2010) suggested that pH effects on soil bacterial community structure were less evident at neutral pH levels and when data were collected over a narrow pH range. It appears likely, then, that one effect of salinity on MOB community structure in these soils occurs via its effect on water availability, which appears to be more important than moisture per se.

Soil methane oxidation is an important methane sink in the global carbon cycle, and soil environmental conditions clearly affect MOB community structure. Soil salinity affected MOB community composition and relative abundance, suggesting the possibility of salinity-based niche differentiation in the MOB. MOB communities were less diverse at lower salinity sites, which showed fewer type II Methylocystis phylotypes. Dryland salinity is a significant threat to large areas of Australia (Kingwell et al., 2008) and changing soil salinities may therefore affect MOB community structure as factors such as increasing dryland salinity alter existent soil conditions.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References
  8. Supporting Information
FilenameFormatSizeDescription
fem1341-sup-0001-TableS1.docWord document40KTable S1. Soil chemistry variables for individual sites used to derive Table 1 (soil chemistry variables for salinity grouped soils).
fem1341-sup-0002-TableS2.xlsapplication/msexcel59KTable S2. Final oligonucleotide probe set.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.