Environmental and edaphic drivers of bacterial communities involved in soil N-cycling.
Nitrogen is a primary driver of ecosystem fertility. The movement of nitrogen into, within, and out of soil is influenced by microbiological processes (De Boer and Kowalchuk 2001; Vitousek et al. 2002; Ogunseitan 2005). Gaining an understanding of the ecology of the soil microbiota involved in N transformations and how their distribution and abundance are affected by soil and environmental properties can be used to better understand the biogeochemical cycling of N in ecosystems. Following from this, knowledge on N form and fate has implications for plant production, nitrate accumulation and leaching to groundwater or other aquatic systems, denitrification (N-based fertility loss from soil), and environmental impacts such as greenhouse-gas formation ([N.sub.2]O).
Natural diversity within soil microbial communities is extensive. As such, understanding biogeochemical cycling by targeting specific phylogenetic groups of organisms is often not possible. An altemative approach to the study of microbial communities of functional relevance is to focus on phylogenetically conserved genes encoding for enzymes involved in nutrient transformations (functional genes). These genes can be targeted to provide useful information on the abundance and, in some cases, diversity of soil bacteria involved in N cycling (WaUenstein and Vilgalys 2005). Of the biological N-transformations, [N.sub.2] fixation, nitrification, and denitrification (nitrate reduction to nitrite) are of particular interest. [N.sub.2] fixation ([N.sub.2] to organic N) uses the nitrogenase enzyme, a conserved subunit of which is nifH (e.g. Kloos et al. 2001; Zehr et al. 2003). The rate-limiting step of nitrification (N[H.sup.+.sub.4] to N[H.sub.2]OH) is undertaken by the ammonia mono-oxygenase enzyme. The amoA gene, encoding for a conserved subunit of the ammonia mono-oxygenase enzyme, can be used as a functional target for bacteria involved in nitrification (Rotthauwe et al. 1997; Kowalchuk and Stephen 2001). The step-wise reductions of N via the denitrification process are each mediated by specific enzymes (e.g. Philippot et al. 2002; Philippot and Hallin 2005; Wallenstein and Vilgalys 2005). However the biological potential for the overall process to occur, i.e. the reduction of N[O.sup.-.sub.3] to N[O.sub.2.sup.-] , can be measured by targeting narG, which encodes for a subunit of periplasmic nitrate reductase enzyme (Philippot et al. 2002). Technical capabilities are being developed to quantify functional genes in both natural and managed ecosystems. The most sensitive of these are realtime (quantitative) polymerase chain reaction (q-PCR) and competitive PCR (e.g. Whelan et al. 2003; Lopez-Gutitrrez et al. 2004; Wallenstein and Vilgalys 2005; Wakelin et al. 2007).
Most applications of functional gene technology to soil nutrient cycling have focused on the investigations of single-gene v. single-geochemical transformations. However, a broader appreciation of ecosystem processes associated with soil N cycling can only be attained via simultaneous analysis of several genes covering various components of the N cycle (e.g. Mergel et al. 2001; Rosch et al. 2002; Taroncher-Oldenburg et al. 2003; Wakelin et al. 2007).
This study assessed the effect of environmental and edaphic properties on the biological potential for N transformations to occur in soil. We selectively targeted key genes regulating key components of the N cycle and related their relative size (abundance within the total DNA pool) to edaphic and environmental properties. Agricultural soils across a rainfall gradient were sampled to provide large variability in potential habitat-selective factors with potential to affect biological communities involved in N cycling. Variation in gene abundance with soil depth was also investigated to examine attenuation or enhancement in the potential for biologically mediated N transformations at varying depth. This approach provided a means to investigate the ecology of N-cycling bacteria at a landscape level and explore effects of long-term environmental (especially climatic) and soil properties on their abundance.
[FIGURE 1 OMITTED]
Materials and methods
Soil sampling and characterisation
The rainfall gradient, spanning over 300mm/year, was represented by 14 soil profiles transecting the top of the eastern side of Gulf Saint Vincent (Balaklava) and progressing east across the lower Clare Valley (Clare) and then into the southern Mallee region (Galga) of South Australia (Fig. 1). Average rainfall throughout 1970-2000 (SILO database; www.bom.gov.au/silo/) ranged from 300mm/year at Balaklava to 600mm/year in the Clare region, and then to below 250 mm/year in the Mallee region.
Four or 5 separate soil cores 30cm long were randomly located and collected within a 100m by 100m zone at each sampling location using an auger (3.9 cm diameter). All sample locations were within non-irrigated agricultural fields with variable cropping histories built around pasture and cereal production. The cores were divided into 0-0.10 and 0.10-0.30m depth layers. Additional soil was collected from >0.70 m depth at half of the sampling locations. Soils sampled from replicate cores within individual fields were combined to give a single composite sample for each soil depth layer at each location. At the Auburn and Caliph sites (Fig. 1), separate samples were collected from the upper (CU, KU) and lower (CL, KL) elevations in the field. All samples were collected over a 2-day period in September and were maintained at 4[degrees]C during transport from the field to the laboratory.
Soil water content at sampling varied across the locations and with depth. All composite samples were air-dried in a fan-forced oven at 40[degrees]C to standardise water contents and facilitate mixing and grinding to <2 mm. Soils in the region are regularly exposed to more severe drying regimes and lower soil water contents than those attained during the air-drying process. Therefore, air-drying was unlikely to perturb the microbial population of the soil samples any more than would be expected under typical environmental conditions. Gravimetric water content of the air-dried soils was determined by oven-drying a subsample at 105[degrees]C to constant mass. Total carbon (TC) and total N (TN) were determined by direct combustion using a LECO CNS 2000. Organic carbon (OC) content was determined using a LECO C-144 C analyser (Merry and Spouncer 1988) and soil materials that had been pre-treated with acid to remove carbonate carbon. Gravimetric carbonate contents (CaC[O.sub.3]) were calculated based on the difference between TC and OC. Inorganic N was determined from a 2M KC1 extraction followed by colourimetric determination of N[H.sup.+.sub.4] and N[O.sup.-.sub.3] (Rayment and Higginson 1992) on an AlpKem Flow Solution III (O.I Analytical, OR, USA). Clay, sand, and char-C contents were derived using partial-least-squares prediction of mid infrared spectral data (Janik et al. 1995). A 1:5 sample: water suspension was used to determine pH (White 1969) and electrical conductivity (EC) (Gillman and Sumpter 1986).
Prior to DNA extraction, all soils were incubated for 12 days at a bulk density of 1.4 g/[cm.sup.3], water-filled pore space of 60%, and constant temperature of 28[degrees]C to activate the soil microbial communities. As noted for the air-drying process, such a wetting event is consistent with the environmental conditions experienced in the field. For each composite sample (all combinations of site x depth) 3 replicate 17 g samples of air-dried soil were prepared and incubated. Immediately after incubation, all samples were stored at -20[degrees]C. DNA was extracted from 3 subsamples of 0.75 g taken from each incubated sample. The extractions were completed the using the MoBio UltraClean Soil DNA kit (MoBio Ltd, CA, USA) immediately after removing the samples from the freezer. The supplier's method was modified to include mechanical disruption using a Fastprep Beadbeater at 6.0m/s for 20s (FP120; Qbiogene, USA). DNA was quantified in the extracts from all 315 soil subsamples (14 locations x 2 or 3 depths x 3 incubated samples x 3 subsamples for DNA extraction) using Quant-it PicoGreen dsDNA Reagent (Invitrogen) on a MX3000P realtime PCR machine (Stratagene, USA). Following quantification of total DNA, the replicate DNA extracts from each incubated sample were pooled. This process provided a single pooled DNA extract, with a known concentration of DNA, for each incubated sample. Therefore, for each composite soil sample, a total of 3 replicate pooled DNA extracts were prepared for subsequent quantification of functional genes.
Quantification of bacterial functional genes
The bacterial functional genes present in the pooled DNA extract for each incubated sample were quantified using q-PCR. PCR primers used were: nifH-F and nifH-R for nifH (Rosch et al. 2002); amoA-1f and amoA-2R for amoA (Rotthauwe et al. 1997; Kowalchuk and Stephen 2001); and 1690f and 2650r for narG (Philippot et al. 2002). Amplification was performed in 25-[micro]L reaction mixes using the QuantiTect SYBR Green PCR kit (Qiagen) on a MX3000P q-PCR system. Final reaction concentrations each used 0.8[micro]m of forward and reverse primer, 2[micro]L of undiluted DNA sample, 12.5 [micro]L of 2 x PCR master mix (including Taq DNA polymerase, SYBR, etc.), and water total volume of 25 [micro]L. PCR cycling conditions were as follows: 40-50 cycles of 95[degrees]C for 45 s, primer annealing for 1 min at 60[degrees]C for nifH or 50[degrees]C for amoA or 55[degrees]C for narG, and template extension at 72[degrees]C for 45 s. The reactions were run for 40+ cycles until no further amplification of samples was observed. The threshold cycle (CT) in which all reactions were in exponential phase amplification was calculated by software and did not require manual adjustment. Functional genes can vary in oligonucleotide composition between bacterial species, which may affect melting (dissociation) curve analysis on q-PCR systems. Therefore, to ensure amplification of specific PCR product, reaction mixes were separated by agarose electrophoresis and stained using ethidium bromide.
Standard curves relating calculated DNA amounts and [C.sub.T] values were used to quantify copy numbers of each gene. PCR products from each gene were cloned into the TOPOTA cloning vector (Invitrogen). Plasmid DNA containing a single gene copy was extracted using the MoBio plasmid prep kit and quantified using PicoGreen as before. 10-fold serial dilutions of the DNA from the plasmid were prepared to obtain a standard curve for each gene spanning 8 orders of magnitude. Each standard curve was run in duplicate. Copy numbers for each sample were calculated (using known sequences of the vector and PCR insert) and the plasmid concentrations and expressed in units of copy numbers per [micro]g DNA extracted to define the abundance of the targeted bacterial functional genes present within the DNA associated with the total microbial community.
Statistical design and analysis
Soils were collected from different locations to examine the potential impact of environmental gradients, in particular rainfall, and soil properties on the abundance of bacterial N cycling genes within the soil microbial community. It was not possible to use a replicated field design for this, as the level of variability in environmental and soil properties at individual locations was inadequate. A consistent methodology was used at each location to acquire a single composite sample for each depth layer. For all soil analyses, replicate determinations were completed for each composite sample and results were averaged to give a single value that was taken into the statistical analyses. Total DNA, N functional gene abundance, and soil properties were compared across sites and depths by an unbalanced analysis of variance (ANOVA) using a general linear model in GENSTAT Version 7.2 (VSN International, Hemel Hempstead, UK). For these analyses, the lack of field-level replication meant that the depth x location interaction mean-square term was used as the denominator to derive F-values to test the significance of the main effects of depth and location. This form of main effect testing was conservative since the mean square associated with the depth x location interaction term can only be equal to or larger than the error term that would have been generated had it been possible to perform field-level replication. Data were tested for the assumption of homogeneity of variance; when the assumption failed the data were [log.sub.10] transformed before ANOVA. Where the influence of depth or location was found to be significant (P< 0.05), the least significant difference (1.s.d.) post-hoc test was used to identify significant differences between depths or locations.
Data of soil physicochemical and environmental properties were normalised and pair-wise correlations between variables made. Soils were then characterised via their edaphic and environmental properties using principal components analysis (PCA). Differences between regions were tested using 2-way crossed (region x depth) analysis of similarities (ANOSIM; Clarke and Warwick 2001).
Multivariate data analyses were used to relate the abundance of N-cycling genes to soil physicochemical and environmental properties. Data for each gene (relative copy numbers) were log(x+1) transformed and resemblance matrices constructed (Bray-Curtis). Associations between the abiotic dataset and biotic data were tested (linked) using BIO-ENV procedure (Clarke 1993; Clarke and Ainsworth 1993) with the Spearman rank coefficient. Permutation (999) of variables was used to generate null distributions from which the significance could be tested. The sample statistic (Spearmans Rho, p) was used to indicate the magnitude of the association (i.e. movement from 0=no association). All multivariate data analysis was conducted using Primer6 (Primer-E Ltd, Plymouth, UK) following multivariate statistical routines outlined by Clarke and Warwick (2001). Correlation and covariance analyses conducted to quantify the relationship between the abundance of different genes were completed using Statictica 8 (Statsoft, Tulsa, OK, USA).
Soil pH and EC increased with depth (P<0.001), pH from 7.8 to 9.3 and EC from 211 to 623 [micro]S/cm (Tables 1 and 2). In contrast, N[H.sup.+.sub.4]-N (from 0.87 to 0.15ms N/ks) and total N (from 0.13 to 0.03%) decreased with depth (P<0.001; Table 2). Soil clay, N[H.sup.+.sub.4]-N, OC, % sand, TC, TN, and pH all varied between sites (P< 0.05; Table 2). This large variation in the soil property dataset was explored using PCA. Using both the physicochemical and environmental factors, the soils grouped into 2 sets (Fig. 2), separated across the x-axis, which were shown to be different by ANOSIM (P< 0.001). The soils from the Clare region were different to those from Balaclava and from the Mallee regions (P=0.001), but there was no significant difference between the soils from Balaclava and Mailee (P=0.66). A large component of the variation in physicochemical properties was explained across PC1 (47.7% of the total variation). The 5 most influential factors linked to PC1 were rainfall (-0.378), elevation (-0.359), TN (-0.374), N[H.sup.+.sub.4]-N (-0.362), and OC (-0.362). Separation of samples across PC2 (22% of the total variation) was associated with variations in sand (0.436), pH (-0.407), EC (-0.472), and TC (-0.377) content of the soils. Eigenvectors were overlaid to the PCA plots to aid in interpretation of regional differences in soils properties (Fig. 2). Soils from the Clare region were characterised by higher rainfall (total and growing season), higher N and C contents, lower pH, and a less sandy texture (Fig. 2).
Quantification of N-cycling genes
The abundance, i.e. relative number of gene copies per ng of DNA, of each the bacterial N-cycling genes was [log.sub.10] transformed before statistical analysis (Table 1). Abundance of the amoA gene copies did not significantly vary between sites, but decreased with depth, from 2.6 at 0-0.10m to 1.3 in the 0.30-0.70 m layer (Table 3). In contrast, copies of the nifH and narG genes were not affected by soil depth but varied across sites (P< 0.05; Table 3), respectively, from 1.6 to 3.3 and from 3.1 to 4.8. Whereas no clear trend of the abundance of nifH gene copies with region could be observed (both the highest and lowest abundance of nifH was found in the Mallee), the abundance of narG was consistently lower in the Mallee region than Clare and Balaclava (Table 1).
Relationships of functional gene abundance to environmental and edaphic factors
The abundance of nifH genes correlated with TC (P=0.006; [rho]=0.382; Table 4) and amoA with EC and OC (P=0.003; [rho] = 0.44). Addition of TC to the correlation slightly increased the explanation in amoA gene abundance ([rho]=0.468). The abundance of the narG gene was positively correlated with rainfall across the sites (P=0.002; [rho]=0.417), but was not associated with soil physiochemical properties. With the exception of amoA (described above), addition of further variables to the correlation reduced the overall explanation of the measured variations in gene abundance.
Relationships between functional gene abundances
Correlation analysis of the abundances of each N transformation gene indicated that ntfH was not linked to amoA or narG; however, a positive relationship was found between amoA and narG ([r.sup.2] = 0.65, P < 0.001). Subsequent analysis of covariance found that the slope of the relationship between amoA and narG abundance within microbial DNA changed (P<0.001) with depth but was only significant for the 0-0.10m (slope=57, P < 0.001) and 0.10-0.30 m (slope = 37, P < 0.001) depths.
The abundance of genes involved in the biogeochemical cycling of N varied across locations (combination of soil type and environment) and/or down soil profiles taken from the mid north of South Australia. The soils, sampled across a wide climatic range, varied in environmental properties but also soil physicochemical attributes, providing an experimental system in which multiple factors could test for association to the relative size of bacterial communities involved in 3 key components of soil N cycling: biological nitrogen fixation, nitrification, and denitrification. The approach also allowed for the broader investigation of edaphic or environmental effects at a landscape level. Most research to date has focused on the effects specific treatment factors (e.g. Yeager et al. 2005; He et al. 2007; Wakelin et al. 2007) within localised sites. Recent publications have also shown that long-term environmental drivers can also affect both abundance and diversity of N-cycling bacteria (Horz et al. 2004; Wallenstein et al. 2006).
[FIGURE 2 OMITTED]
In natural ecosystems, inputs of N to soil are generally dependent on the process of biological N fixation. The subsequent production of ammonium (mineralisation), conversion of ammonium to nitrate (nitrification), and loss of nitrate via denitrification should exhibit a dependence on the prerequisite process. Although we have studied agricultural soils, quantifying the abundance of multiple N cycling genes provides the opportunity to examine linkages between these biologically mediated N transformation processes. Whereas there was no relationship between the abundance of nifH and amoA or narG, the abundance of amoA was positively correlated with the abundance of narG. Annual additions of ammonium-based fertilisers to these soils and the potential existence of nitrifying organisms not detected by amoA may account for the absence of a relationship between nifH and amoA or narG. The positive correlation between amoA and narG abundance is consistent with the fact that amoA activity (nitrification) leads to the production of the substrate (nitrate) required by narG activity (denitrification). Changes in this relationship with depth are consistent with decreasing soil OC (reduced supply of ammonium via mineralisation resulting in reduced nitrification potential) and reduced availability of oxygen (enhanced denitrification potential).
Compared with native ecosystems, soil N cycling within dryland agricultural systems can be considered to be highly disturbed. In particular, the management of N, C, and other inputs to the farming systems is likely to vary between sites according to rainfall and soil properties affecting crop yield potential. The Mallee and Balaclava regions of Australia have low and variable rainfall (mean annual rainfall of 300-400mm/year with a range of 208-494mm/year). To manage the risk of cropping in these areas, many growers adopt conservative strategies (Sadras 2002; Sadras and Roget 2004) such as using N fertiliser rates that target average to below-average yields (Sadras 2002). In higher rainfall areas, such as the Clare region (600 mm rainfall/year), crop failure is less likely and growers adopt higher input strategies designed to optimising potential yields (Sadras and Roget 2004). Across the regions sampled, a variety of application strategies for fertiliser N exist which may perturb the links between N cycling genes and the N transformation processes they control. Additionally, legumes (pulse crops and pastures) are key components of the crop rotations used in these regions. The N returned to the soil by symbiotic N fixation, as well as N spared during the legume growth also have the potential to alter (enhance or limit) N transformations. For example, mineralised organic N in the soil surface layers spared from crop uptake due to N fixation would be converted to nitrate via nitrification, move down the soil profile due to rain events, and be susceptible to denitrification at depth.
The process of [N.sub.2] fixation is energy-intensive, requiring large energy inputs (carbon mineralised) per mole of N fixed. Large inputs of C into soils in the form of crop residues provide a key stimulus for non-symbiotic [N.sub.2] fixation (e.g. Roper 1983). In Australian dryland agricultural systems, C inputs (plant production) are largely influenced by growing season rainfall, and this may, therefore, be expected to be linked with nifH. Our results, however, did not show a significant overall association between nifH gene abundance and rainfall and/or OC content of the soils. While nifH abundance differed between sampling locations, no consistent effects were found within the Mallee, Clare, and Balaclava regions. Rather, variation in the biological potential for [N.sub.2] fixation (nifH gene abundance) was associated with total C content of soils (P=0.006; [rho]=0.382). Soil types across the dryland cropping region of southern Australia vary widely, but include a range of sandy Calcarosols (classification according to Isbell 1996). Such soils, with high TC (due to free carbonates) and low OC, have low potential productivities due to their poor ability to retain nutrients (especially C and N) and water and to their poor ability to provide nutrients for crop growth. Accordingly, these soils have some of the lowest estimated potential values for non-symbiotic [N.sub.2] fixation in Australian cropping systems (Gupta et al. 2006). Although a significant amount of the variation in mfH gene abundance was linked with soil C ([rho]=0.382; P=0.006), most of the variation in the abundance of the gene remained unexplained by the edaphic and environmental factors measured.
The process of nitrification has traditionally been thought to be performed solely by autotrophic bacteria, which gain energy by oxidising N[H.sup.+.sub.4] derived from fertilisers or mineralisation of organic material, and satisfy their C requirements through reduction of C[O.sub.2]. Long-term rates of nitrification and the abundance of nitrifying bacteria within the soil microbial community can therefore be affected by factors controlling net availability of N[H.sup.+.sub.4] in soil. In particular, the balance between N mineralisation and immobilisation rates is likely to be a key factor (Murphy et al. 2003). Within a given soil ecosystem, the cycling/turnover of N can be largely independent of rates of external inputs (N fixation or fertiliser addition) (Murphy et al. 2003). As such, soil and environmental factors affecting the size of functional communities involved in [N.sub.2] fixation or denitrification may or may not be linked with nitrifying bacteria. In this study, the abundance of amoA within the microbial DNA did not vary significantly with site but decreased significantly with soil depth (P=0.006), being greatest in the 0-0.10m zone. The variation in amoA gene abundance (across all soils sites and depths) was associated with variation in EC and also soil OC. The primary factor, soil EC, increased with depth (P<0.001) but was not affected by sampling location. Soil EC values are indicative of the concentration of soluble salts in soils (Rhoades 1993). Although increasing EC is known to be inhibitory to microbial processes such as nitrification (Kumar and Wagenet 1985; Kumar et al. 1988), the EC range at which microbial processes are inhibited is higher than those measured here (Inubushi et al. 1999). The observed link between increasing EC and decreasing abundance of amoA may be more attributable to a 'depth' effect than a direct effect of EC on the abundance of nitrifiers within the microbial community. OC decreased with depth (P<0.001) and varied across sites (P<0.05), consistent with the changes noted for amoA abundance. Mineralisation of organic N, held within the pool of soil OC, represents a process responsible for the generation of soil N[H.sup.+.sub.4], the substrate used by ammonia-oxidising bacteria. The link between OC and amoA gene abundance was therefore considered to be reflective of the importance of soil C in N mineralisation, which should occur at greater rates in soil surface horizons.
However, a link between the mineralisation of soil organic N and the activity of autotrophic ammonium-oxidising bacteria can be complicated by heterotrophic nitrification rates, which are greater than expected (Cookson et al. 2006), and more importantly other soil microbial processes which also consume ammonium. For example, it has recently been demonstrated that Archaea can contribute significantly to nitrification in soil (Leininger et al. 2006; Chen et al. 2008; Shen et al. 2008). The PCR method used in this study was specific for bacterial ammonia mono-oxygenase gene. Therefore, our amoA gene abundance results may not be reflective of the total biological potential in soil for nitrification. However, the recent work by Shen et al. (2008) indicates that in an alkaline sandy loam, while Archeal amoA gene copy numbers were much greater than those of bacterial amoA genes, abundance of ammonia-oxidising bacteria had significant correlations with pH and, more importantly, potential nitrification rates, thus adding weight to the relationships identified in this study.
Denitrification results in a loss of N from ecosystems, either as soluble forms leached through soil or as gaseous N compounds. Our results have shown that the abundance of narG varied across sites (P<0.001) and that this variation was linked with rainfall ([rho]=0.417; P=0.002). Soil DNA from the Mallee sites (all with lowest rainfall) had lower abundance of narG than higher rainfall sites. The link between rainfall and narG may either reflect the long-term [O.sub.2] availability (REDOX) or, as described previously, variation in N management according to site yield potential. Regardless, the strong link between average rainfall and soil denitrification is important, as it demonstrates that soil microbial communities are responsive to long-term climatic conditions. This has wide-ranging implications with regard to shifting of climatic conditions (particularly temperature and rainfall) as a result of global climate change. Soil microbial communities that affect the biogeochemical cycling of nutrients in ecosystems will be affected (Horz et al. 2004).
Of the edaphic factors measured, soil pH, in particular, has previously been shown to affect microbial processes such as nitrification (Haynes 1986; Broos et al. 2007b) and denitrification (Stevens et al. 1998; Bolan et al. 2004). Under acidic conditions the availability of OC and N for mineralisation diminishes, resulting in a smaller microbial community and, consequently, a smaller nitrifying and denitrifying component. However, the sites in this study, with the exception of Marrabel, had a neutral to alkaline pH (7-9). The small range in pH across the soils used in this study may not have provided sufficient potential to measure the influence of pH on biological potential for soil N cycling.
The units used for the functional gene quantifications are an important factor to consider when interpreting gene abundance. The data as presented in this study describe the quantity of a given bacterial N transformation gene per unit of microbial community DNA (e.g. Wallenstein and Vilgalys 2005). As such, the data indicate the 'enrichment' of the bacterially derived N transformation genes (potential for biological N transformation) within soil microbial communities. The data acquired were presented in this manner for the following reasons. Firstly, the overarching goal of this work was to investigate how soil edaphic and environmental factors affected functional groups of microorganisms. It was expected that if specific drivers were affecting the potential for biochemical transformations over time, then either a loss or enrichment of the genes associated with the function would be apparent within the DNA of the microbial community (i.e. relative increase in function per unit of community size). Secondly, specific constraints to quantifying the total abundance of particular functional genes in soil exist. It is very unlikely, given the variations noted in edaphic properties across the soils sampled, that DNA extraction efficiencies from all the soils and soil depths would be similar. Examining the enrichment or reduction of specific functional genes expressed as a function of the total microbial DNA present eliminates this issue and allows changes in abundance to be identified, assuming uniform exaction efficiency for the various types of DNA. Thirdly, spatial and temporal variations in the size and activity of soil microbial communities in the field can change in total size over relatively small periods of time and space (Broos et al. 2007a). To address this, multiple soil samples were pooled to give a composite sample for each location and depth layer, a standardised preincubation protocol was used to minimise the effects of sampling time, soil temperature, and soil water status, and replicate DNA extractions from each incubated sample were pooled.
Summary and conclusions
The biological potential for different aspects of N cycling in soils is each independently associated with various edaphic (pH, EC, OC, and depth) or long-term environmental (rainfall) factors. The differential regulation of the 3 biological functions affecting N input (nif H), N cycling (amoA), and N loss (narG) reveals an additional level of complexity in the ecology of biological N transformations. Investigations of biological N transformations in soil must consider the multiple components of the N cycle if a holistic understanding of N-cycling, and the biology and ecology underpinning N transformations, is to be gained.
This work was funded by the Australian Greenhouse Office. B. Roberts, K. L'Anson, P. Schmaal, R. Graetz, J. Faulkner, D. Wormald, and M. Behn kindly provided access to agricultural sites. Drs Matt Colloff and Don Gomez, CSIRO Entomology, kindly reviewed and commented on the manuscript.
Manuscript received 25 May 2008, accepted 27 January 2009
Bolan N, Saggar S, Luo J, Bhandral R, Singh J (2004) Gaseous emissions of nitrogen from grazed pastures: Processes, measurements and modelling, environmental implications, and mitigation. Advances in Agronomy 84, 37-120. doi: 10.1016/S0065-2113(04)84002-1
Broos K, Macdonald LM, Warne MS J, Heemsbergen DA, Barnes MB, Bell M, McLaughlin MJ (2007a) Limitations of soil microbial biomass carbon as an indicator of soil pollution in the field. Soil Biology & Biochemistry 39, 2693-2695. doi: 10.1016/j.soilbio.2007.05.014
Broos K, Warne MS J, Heemsbergen DA, Stevens D, Barnes MB, Correll RL, McLaughlin MJ (2007b) Soil factors controlling the toxicity of Cu and Zn to microbial processes in Australian soils. Environmental Toxicology and Chemistry 26, 583-590. doi: 10.1897/ 06-302R.1
Chen XP, Zhu YG, Xia Y, Shen JP, He JZ (2008) Ammoniaoxidising archaea: important players in paddy rhizosphere soil? Environmental Microbiology 10(8), 1978-1987. doi: 10.111l/j.1462-2920.2008.01613.x
Clarke KR (1993) Non-parametric multivariate analysis of changes in community structure. Australian Journal of Ecology 18, 117-143. doi: 10.111l/j.1442-9993.1993.tb00438.x
Clarke KR, Ainsworth M (1993) A method of linking multivariate community structure to environmental variables. Marine Ecology Progress Series 92, 205-219. doi: 10.3354/meps092205
Clarke KR, Warwick RM (2001) 'Change in marine communities: an approach to statistical analysis and interpretation.' 2nd edn (Primer-E Ltd: Plymouth, UK)
Cookson WR, Muller C, O'Brien PA, Murphy DV, Grierson PF (2006) Nitrogen dynamics in an Australian semiarid grassland soil. Ecology 87, 2047-2057. doi: 10.1890/0012-9658(2006)87[2047:NDIAAS]2.0.CO;2
De Boer W, Kowalchuk GA (2001) Nitrification in acid soils: microorganisms and mechanisms. Soil Biology & Biochemistry 33, 853-866. doi: 10.1016/S0038-0717(00)00247-9
Gillman GP, Sumpter EA (1986) Modification to the compulsive exchange method for measuring exchange characteristics of soils. Australian Journal of Soil Research 24, 61-66. doi: 10.1071/SR9860061
Gupta VVSR, Roper MM, Roget DK (2006) Potential for non-symbiotic [N.sub.2]-fixation in different agroecological zones of southern Australia. Australian Journal of Soil Research 44, 343-354. doi: 10.1071/ SR05122
Haynes R (1986) Nitrification. In 'Mineral nitrogen in the plant-soil system'. (Ed. RJ Haynes) pp. 127-165. (Academic Press: New York)
He J, Shen J, Zhang L, Zhu Y, Zheng Y, Xu M, Di HJ (2007) Quantitative analyses of the abundance and composition of ammonia-oxidizing bacteria and ammonia-oxidizing archaea of a Chinese upland red soil under long-term fertilization practices. Environmental Microbiology 9, 2364-2374. doi: 10.111l/j.1462-2920.2007.01358.x
Horz HP, Barbrook A, Field CB, Bohannan BJM (2004) Ammonia-oxidizing bacteria respond to multifactorial global change. Proceedings of the National Academy of Sciences of the United States of America 101, 15136-15141. doi: 10.1073/pnas.0406616101
Inubushi K, Barahona MA, Yamakawa K (1999) Effects of salts and moisture content on N20 emission and nitrogen dynamics in Yellow soil and Andosol in model experiments. Biology and Fertility of Soils 29, 401-407. doi: 10.1007/s003740050571
Isbell RF (1996) 'The Australian Soil Classification.' (CSIRO Publishing: Collingwood, Vic.)
Janik LJ, Skjemstad JO, Raven MD (1995) Characterisation and analysis of soils using mid-infra red partial least squares. I. Correlations with XRF-determined major element composition. Australian Journal of Soil Research 33, 621-636. doi: 10.1071/SR9950621
Kloos K, Mergel A, Rrsch C, Bothe H (2001) Denitrification within the genus Azospirillum and other associative bacteria. Australian Journal of Plant Physiology 28, 991-998.
Kowalchuk GA, Stephen JR (2001) Ammonia-oxidising bacteria: A model for molecular microbial ecology. Annual Review of Microbiology 55, 485-529. doi: 10.1146/annurev.micro.55.1.485
Kumar V, Wagenet RJ (1985) Salt effects on urea hydrolysis and nitrification during leaching through laboratory soil columns. Plant and Soil 85, 219-227. doi: 10.1007/BF02139626
Kumar V, Yadav DS, Singh M (1988) Effects of urea rates, farmyard manure, CaC[O.sub.3], salinity and alkalinity levels on urea hydrolysis and nitrification in soils. Australian Journal of Soil Research 26, 367-374. doi: 10.1071/SR9880367
Leininger S, Urich T, Sehloter M, Schwark L, Qi J, Nicol GW, Prosser JI, Schuster SC, Schleper C (2006) Archaea predominate among ammonia-oxidizing prokaryotes in soils. Nature 442, 806-809. doi: 10.1038/ nature04983
Lopez-Gutierrez JC, Henry S, Hallet S, Martin-Laurent F, Catroux G, Philippot L (2004) Quantification of a novel group of nitrate-reducing bacteria in the environment by real-time PCR. Journal of Microbiological Methods 57, 399-407. doi: 10.1016/j.mimet.2004. 02.009
Mergel A, Schmitz O, Mallmann T, Bothe H (2001) Relative abundance of denitrifying and dinitrogen-fixing bacteria in layers of a forest soil. FEMS Microbiology Ecology 36, 33-42. doi: 10.1111/j.1574-6941. 2001.th00823.x
Merry RH, Spouncer LR (1988) The measurement of carbon in soils using a microprocessor-controlled resistance furnace. Communications in Soil Science and Plant Analysis 19, 707-720. doi: 10.1080/0010362 8809367969
Murphy DV, Recous S, Stockdale EA, Fillery IRP, Jensen LS, Hatch D J, Goulding KWT (2003) Gross nitrogen fluxes in soil: theory, measurement and application of [sup.15]N pool dilution techniques. Advances in Agronomy 79, 69-118. doi: 10.1016/S0065-2113(02) 79002-0
Ogunseitan O (2005) 'Microbial diversity.' (Blackwell Science Ltd: Oxford, UK)
Philippot L, Hallin S (2005) Finding the missing link between diversity and activity using denitrifying bacteria as a model functional community. Current Opinion in Microbiology 8, 234-239. doi: 10.1016/j.mib. 2005.04.003
Philippot L, Piutti S, Martin-Laurent F, Hallet S, Germon JC (2002) Molecular analysis of the nitrate-reducing community from unplanted and maize-planted soils. Applied and Environmental Microbiology 68, 6121-6128. doi: 10.1128/AEM.68.12.6121-6128.2002
Rayment GE, Higginson FR (1992) Nitrogen. In 'Australian laboratory handbook of soil and water chemical methods'. (Eds GE Rayment, FR Higginson) pp. 38-57. (Inkata Press: Melbourne, Vic.)
Rhoades JD (1993) Electrical conductivity methods for measuring and mapping soil salinity. Advances in Agronomy 49, 201-251. doi: 10.1016/ S0065-2113(08)60795-6
Roper MM (1983) Field measurements of nitrogenase activity in soils amended with wheat straw. Australian Journal of Agricultural Research 34, 725-739. doi: 10.1071/AR9830725
Rrsch C, Mergel A, Bothe H (2002) Biodiversity of denitrifying and dinitrogen-fixing bacteria in an acid forest soil. Applied and Environmental Microbiology 68, 3818-3829. doi: 10.1128/AEM.68. 8.3818-3829.2002
Rotthauwe J, Witzel KP, Liesack W (1997) The ammonia monooxygenase structural gene amoA as a functional marker: molecular fine-scale analysis of natural ammonia-oxidizing populations. Applied and Environmental Microbiology 63, 4704-4712.
Sadras VO (2002) Interaction between rainfall and nitrogen fertilisation of wheat in environments prone to terminal drought. Economic and environmental risk analysis. Field Crops Research 77, 201-215. doi: 10.1016/S0378-4290(02)00083-7
Sadras VO, Roget DK (2004) Production and environmental aspects of cropping intensification in a semiarid environment of southeastern Australia. Agronomy Journal 96, 236-264.
Shen J, Zhang L, Zhu Y, Zhang J, He J (2008) Abundance and composition of ammonia-oxidising bacteria and ammonia-oxidising archaea communities of an alkaline sandy loam. Environmental Microbiology 10(6), 1601-1611. doi: 10.111 l/j.1462-2920.2008.01578.x
Stevens R J, Laughlin R J, Malone JP (1998) Soil pH affects the processes reducing nitrate to nitrous oxide and di-nitrogen. Soil Biology & Biochemistry 30, 1119-1126. doi: 10.1016/S0038-0717(97)00227-7
Taroncher-Oldenburg G, Griner EM, Francis CA, Ward BB (2003) Oligonucleotide microarray for the study of functional gene diversity in the nitrogen cycle in the environment. Applied and Environmental Microbiology 69, 115-1171. doi: 10.1128/AEM.69.2.1159-1171.2003
Vitousek P, Hattenschwiler S, Olander L, Allison S (2002) Nitrogen and nature. Ambio 31, 97-101.
Wakelin SA, Colloff M J, Harvey PR, Marschner P, Gregg AL, Rogers SL (2007) The effects of stubble retention and nitrogen application on soil microbial community structure and functional gene abundance under irrigated maize. FEMS Microbiology Ecology 59, 661-670. doi: 10.1111/j.1574-6941.2006.00235.x
Wallenstein MD, Myrold DD, Firestone M, Voytek M (2006) Environmental controls on denitrifying communities and denitrification rates: Insights from molecular methods. Ecological Applications 16, 2143-2152. doi: 10.1890/1051-0761 (2006)01612143:ECODCA]2.0.CO;2
Wallenstein MD, Vilgalys RJ (2005) Quantitative analyses of nitrogen cycling genes in soils. Pedobiologia 49, 665-672. doi: 10.1016/j. pedobi.2005.05.005
Whelan JA, Russell NB, Whelan MA (2003) A method for the absolute quantification of cDNA using real-time PCR. Journal of Immunological Methods 278, 261-269. doi: 10.1016/S0022-1759(03)00223-0
White RE (1969) On the measurement of soil pH. Journal of the Australian Institute of Agricultural Science 35, 3-14.
Yeager CM, Northup DE, Grow CC, Barns SM, Kuske CR (2005) Changes in nitrogen-fixing and ammonia-oxidizing bacterial communities in soil of a mixed conifer forest after wildfire. Applied and Environmental Microbiology 71, 2713-2722. doi: 10.1128/ AEM. 71.5.2713-2722.2005
Zehr JP, Jenkins BD, Short SM, Steward GF (2003) Nitrogenase gene diversity and microbial community structure: a cross-system comparison. Environmental Microbiology 5, 539-554. doi: 10.1046/ j. 1462-2920.2003.00451 .x
M. S. Forbes (A,B,D), K. Broos (A,C), J. A. Baldock (A), A. L. Gregg (A), and S. A. Wakelin (A)
(A) CSIRO Land and Water, PMB 2, Glen Osmond, SA 5064, Australia.
(B) Department of Environment and Conservation, Locked Bag 104, Bentley, WA 6983, Australia.
(C) Current address: VITO--Flemish Institute for Technological Research, Boeretang 200, 2400 Mol, Belgium.
(D) Corresponding author. Email: firstname.lastname@example.org
Table 1. Physicochemical, environmental, and functional gene data from agricultural soils sampled across the mid-north of South Australia EC, Electrical conductivity; OC, organic C; TC, total C; TN, total N Site Code Lat. (S), Depth Elev. long. (E) (m) (m a.s.l.) Port FF1 34 20.270 0-0.10 13 Wakefield FF2 138 16.030 0.10-0.30 FF3 >0.70 Avon AV1 34 18.772 0-0.10 20 AV2 138 21.853 0.10-0.30 AV3 >0.70 Balaklava BP1 34 09.121 0-0.10 50 BP2 138 21.529 0.10-0.30 BP3 >0.70 Hoyleton DF1 34 06.064 0-0.10 101 DF2 138 29.552 0.10-0.30 Auburn KU1 34 02.075 0-0.10 346 (upper) KU2 138 39.295 0.10-0.30 KU3 >0.70 Aubum KL1 34 02.174 0-0.10 335 (lower) KL2 138 39.292 0.10-0.30 Mintaro MH1 33 57.808 0-0.10 358 MH2 138 43.047 0.10-0.30 Marrabel IRl 34 04.146 0-0.10 439 IR2 138 52.472 0.10-0.30 IR3 >0.70 Black IBl 33 54.814 0-0.10 456 Springs IB2 138 52.334 0.10--0.30 Sedan North SS1 34 26.989 0.10 135 SS2 139 15.537 0.10-0.30 SS3 >0.70 Sedan South SN1 34 33.473 0-0.10 158 SN2 139 15.110 0.10-0.30 Galga G1 34 43.024 0-0.10 50 G2 139 55.024 0.10-0.30 G3 >0.70 Caliph CU1 34 40.443 0-0.10 49 (upper) CU2 139 42.272 0.10-0.30 Caliph CL1 34 39.949 0-0.10 48 (lower) CL2 140 23.904 0.10-0.30 Site Code Rainfall (A) Sand Clay (mm) (%) Port FF1 358 (85) 84.2 14.8 Wakefield FF2 82.7 17.3 FF3 83.4 16.6 Avon AV1 365 (87) 50.1 18.3 AV2 38.8 22.1 AV3 18.9 34.3 Balaklava BP1 366 (90) 33.6 35.0 BP2 30.1 30.3 BP3 31.7 31.7 Hoyleton DF1 400 (94) 67.3 21.5 DF2 31.3 26.1 Auburn KU1 539 (113) 12.7 32.3 (upper) KU2 12.8 35.6 KU3 15.6 25.2 Aubum KL1 539 (113) 17.4 43.1 (lower) KL2 19.2 34.0 Mintaro MH1 537 (116) 42.1 24.1 MH2 38.7 25.5 Marrabel IRl 529 (114) 26.2 29.5 IR2 28.3 38.5 IR3 28.4 24.3 Black IBl 515 (117) 49.7 28.2 Springs IB2 29.8 30.3 Sedan North SS1 338 (102) 30.4 28.2 SS2 36.3 I5.1 SS3 3.2 22.8 Sedan South SN1 365 (105) 27.0 29.5 SN2 7.3 32.1 Galga G1 298 (90) 92.9 7.1 G2 91.0 9.0 G3 75.9 12.4 Caliph CU1 300 (90) 88.6 11.4 (upper) CU2 91.0 9.0 Caliph CL1 300 (90) 59.3 20.3 (lower) CL2 46.3 22.4 Site Code pH EC OC ([micro]S/cm) Port FF1 8.0 135 1.06 Wakefield FF2 8.7 130 0.42 FF3 8.5 121 0.32 Avon AV1 8.3 192 1.49 AV2 8.7 217 0.70 AV3 9.7 690 0.20 Balaklava BP1 8.1 185 1.74 BP2 8.2 208 0.80 BP3 8.9 1240 0.21 Hoyleton DF1 8.4 163 1.04 DF2 8.7 155 0.47 Auburn KU1 7.7 376 2.90 (upper) KU2 7.9 208 2.16 KU3 8.8 243 1.05 Aubum KL1 7.9 192 2.25 (lower) KL2 8.1 192 1.62 Mintaro MH1 7.7 207 2.67 MH2 8.0 149 1.01 Marrabel IRl 5.8 167 1.77 IR2 6.6 154 0.94 IR3 9.2 574 0.24 Black IBl 7.3 340 1.41 Springs IB2 7.6 304 0.66 Sedan North SS1 8.2 164 1.60 SS2 8.6 147 0.63 SS3 9.8 638 0.32 Sedan South SN1 8.4 486 0.95 SN2 8.7 734 0.76 Galga G1 7.9 162 0.46 G2 8.3 111 0.20 G3 9.3 672 0.19 Caliph CU1 7.9 88 0.30 (upper) CU2 8.5 102 0.17 Caliph CL1 8.0 189 1.20 (lower) CL2 8.7 164 0.63 Site Code TC TN N[H.sub.4]-N N[O.sub.3]-N (%) (mg/kg) Port FF1 1.10 0.09 0.4 0.7 Wakefield FF2 0.85 0.03 0.1 0.4 FF3 1.08 0.02 0.1 0.1 Avon AV1 2.25 0.12 0.4 1.8 AV2 3.37 0.06 0.2 3.8 AV3 2.12 0.03 0.1 8.6 Balaklava BP1 2.15 0.16 0.4 0.6 BP2 2.21 0.08 0.2 3.2 BP3 2.29 0.03 0.1 0.6 Hoyleton DF1 2.84 0.09 1.1 1.6 DF2 5.70 0.03 0.1 3.1 Auburn KU1 3.92 0.24 1.8 4.0 (upper) KU2 3.01 0.16 0.9 3.0 KU3 6.89 0.04 0.2 5.5 Aubum KL1 2.69 0.12 1.1 4.1 (lower) KL2 2.37 0.24 0.8 2.5 Mintaro MH1 3.04 0.26 1.5 2.8 MH2 1.05 0.09 0.3 0.6 Marrabel IRl 1.77 0.18 1.5 10.4 IR2 0.94 0.13 1.6 6.4 IR3 0.34 0.08 0.5 2.5 Black IBl 1.56 0.14 1.0 1.7 Springs IB2 1.33 0.09 0.4 1.7 Sedan North SS1 2.15 0.14 0.8 1.3 SS2 4.24 0.06 0.3 0.8 SS3 7.11 0.02 0.1 6.6 Sedan South SN1 1.43 0.09 0.5 6.0 SN2 1.62 0.08 0.3 6.9 Galga G1 0.46 0.04 0.4 0.8 G2 0.40 0.02 0.2 0.6 G3 0.78 0.02 <0.1 2.0 Caliph CU1 0.44 0.03 0.3 0.8 (upper) CU2 0.17 0.01 0.2 0.5 Caliph CL1 1.20 0.09 1.0 4.4 (lower) CL2 1.13 0.05 0.2 0.9 Site Code nifH amoA narG ([log.sub.10] copies/ng DNA) (B) Port FF1 2.04 1.92 3.22 Wakefield FF2 2.71 2.00 3.48 FF3 3.50 1.38 4.47 Avon AV1 3.03 2.57 4.27 AV2 2.25 2.51 4.44 AV3 2.83 1.00 4.85 Balaklava BP1 3.00 2.68 4.41 BP2 2.11 2.18 4.02 BP3 2.27 0.60 3.80 Hoyleton DF1 2.72 2.92 4.73 DF2 1.80 0.00 4.58 Auburn KU1 2.57 3.08 4.72 (upper) KU2 2.07 3.33 4.98 KU3 1.36 0.00 4.58 Aubum KL1 2.77 2.93 4.78 (lower) KL2 2.17 2.74 4.49 Mintaro MH1 2.10 3.01 4.75 MH2 1.67 2.79 4.73 Marrabel IRl 2.72 2.45 4.33 IR2 2.88 2.25 4.35 IR3 3.02 2.54 4.10 Black IBl 2.69 2.15 3.79 Springs IB2 3.01 2.01 4.12 Sedan North SS1 2.15 2.76 3.23 SS2 1.36 2.92 3.30 SS3 1.32 2.03 3.18 Sedan South SN1 3.10 2.61 4.06 SN2 2.21 2.43 3.19 Galga G1 3.00 2.51 3.47 G2 3.39 2.18 3.41 G3 2.92 2.29 3.69 Caliph CU1 3.42 2.17 3.03 (upper) CU2 3.30 2.06 3.11 Caliph CL1 2.81 2.37 3.04 (lower) CL2 0.85 1.83 3.11 (A) Values in parentheses are standard deviation of annual rainfall determined from 1970 to 2000 from SILO database. (B) Gene abundances are the total gene copy numbers divided by the measured DNA concenhation used per reaction to give gene copies/ng of DNA. Table 2. Summary ANOVA results for soil properties and environmental characteristics compared with soil and depth EC, Electrical conductivity; OC, organic C; TC, total C; TN, total N Sand Clay pH Site P <0.001 <0.001 0.006 Max l.s.d. 19.59 10.50 0.881 Depth P 0.026 0.98 <0.001 Max l.s.d. 9.795 -- 0.441 EC OC TC Site P 0.148 <0.001 0.003 Max l.s.d. -- 0.636 2.411 Depth P <0.001 <0.001 0.270 Max l.s.d. 185.0 0.318 -- TN N[H.sup.+.sub.4]-N N[O.sup.-.sub.3]-N Site P 0.008 0.007 0.071 Max l.s.d. 0.088 0.623 -- Depth P <0.001 <0.001 0.532 Max l.s.d. 0.044 0.311 -- Table 3. Summary ANOVA results for soil and depth with abundance of functional genes (log number of copies/ng DNA extracted) nifH amoA narG Site P 0.030 0.694 <0.001 Max l.s.d. 1.54 -- 0.6484 Depth P 0.082 0.006 0.667 Max l.s.d. -- 0.7355 -- Table 4. Environmental and soil factors related to the abundance of key N cycling genes (across all soils and depths) Gene Function Correlated Spearman Significance (B) variables (A) ([rho]) (P) nifH [N.sub.2] fixation TC 0.382 0.006 amoA Nitrification EC, OC 0.440 0.003 nurG Denitrification Rainfall 0.417 0.002 (A) TC, Total C; EC, electrical conductivity; OC, organic C. (B) Significance of p tested against null distribution involving 999 permutation (randomisation) tests on the each of the associations of the functional genes.
|Printer friendly Cite/link Email Feedback|
|Author:||Forbes, M.S.; Broos, K.; Baldock, J.A.; Gregg, A.L.; Wakelin, S.A.|
|Publication:||Australian Journal of Soil Research|
|Date:||Jul 1, 2009|
|Previous Article:||Long-term effect of pastures on soil quality in acid soil of North-East India.|
|Next Article:||Impact of soil conditions on hydrology and water quality for a brown clay in the north-eastern cereal zone of Australia.|