If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
UWA School of Agriculture and Environment, The University of Western Australia, Perth 6009, Western Australia, AustraliaSchool of Veterinary and Life Sciences, Murdoch University, Murdoch 6150, Western Australia, AustraliaAgriculture and Food Division, Department of Primary Industries and Regional Development, South Perth 6151, Western Australia, Australia
Through clearing and use of fertilizer and legumes, areas of southwestern Australia's unique coastal sand plains can support relatively low-cost dairies. However, the ancient, highly weathered nature of the soils in this region makes the dairies susceptible to a range of threats, including nutrient leaching and erosion. Despite this, Western Australian dairy cows typically produce up to 5,500 L of milk per head annually supported by inorganic nitrogen (N) fertilizer (commonly 50:50 urea and ammonium sulfate) at rates up to <320 kg of N/ha per year. Where hotspots exist (up to 2,000 kg of N/ha per year), total N exceeds pasture requirements. We investigated plant and soil bacteria responses to N fertilizer rates consistent with Australian legislated production practices on dairy farms for pure and mixed swards of white clover (Trifolium repens) and Italian ryegrass (Lolium multiflorum) in a long-term pasture experiment in controlled glasshouse conditions. Although the soil bacterial community structure at phylum level was similar for white clover and Italian ryegrass, relative abundances of specific subgroups of bacteria differed among plant species according to the N fertilizer regimen. Marked increases in relative abundance of some bacterial phyla and subphyla indicated potential inhibition of N cycling, especially for N hotspots in soil. Ammonium concentration in soil was less correlated with dominance of some N-cycling bacterial phyla than was nitrate concentration. Changes in bacterial community structure related to altered nutrient cycling highlight the potential for considering this area of research in policy assessment frameworks related to nutrient loads in dairy soils, especially for N.
Nutrient use efficiency, especially for nitrogen (N), is facilitated by maximizing beneficial microbial processes in dairy soil leading to less nutrient leaching and runoff, lower costs, and better outcomes for farmers (
). Australia's ecological sustainability legislation under which the dairy industry operates is outlined in Australia's Environment Protection and Biodiversity Conservation Act (
). This legislation does not include the effects of microbial interactions associated with nutrient loads from the perspective of whole-ecosystem life cycle or environmental fate (
). Reports on soil quality, including N cycling, from New Zealand, the European Union, the United Kingdom, and the United States suggest that Australia's national sustainability legislation may need review in the light of new research linking production practices to microbial interactions in soil management with outcomes for ecosystems and primary producers (
). The addition of high rates of N fertilizer to dairy pastures has potential to significantly reduce the effectiveness of soil microbial processes associated with efficient nutrient cycling, which may lead to overdose fertilization and downstream effects (
). This may occur where excessive N accumulates and forms hotspots (e.g., urine patches), with total N entering ecosystems of up to 2,000 kg/ha per year (
The contribution of cattle urine and dung to nitrous oxide emissions: Quantification of country specific emission factors and implications for national inventories.
). There may be benefits for pasture soil management specific to pasture plant species and associated bacterial groups (e.g., N fixation facilitated by bacteria within Proteobacteria (
). Dominance or subdominance of particular groups of soil bacteria and their functions are now being categorized quantitatively via “omics” technologies. However, their role in soils cannot be determined by merely considering the dynamics and statistical significance of common gene database sequences or bacterial species due to multilayered biological evolutionary selection (
Direct 16S rRNA-seq from bacterial communities: A PCR-independent approach to simultaneously assess microbial diversity and functional activity potential of each taxon.
). For microbiota inhabiting rhizospheres, selection acts simultaneously on genes, individuals, cells, groups, and the holobiont (the plant host with its extended microbiome;
). It is essential to minimize bias surrounding extrapolation (the typical approach in species-specific gene functional analysis) and focus on increasing data reliability (
) and by limiting functional prediction to already-characterized groups within traditional physiological classes in combination with long-term experiments, extrapolations can almost be eliminated (
). We expected that persistent application of N fertilizer on the soil bacteria under swards of white clover (Trifolium repens) and Italian ryegrass (Lolium multiflorum;
This study examined the effect of N and pasture plant species on soil bacterial diversity and relative abundance at the end of a long-term (8 mo) pasture sward experiment set up in a controlled (20°C) glasshouse at The University of Western Australia (UWA).
We selected soil for this study from a dairy farm located in Cowaramup, Western Australia (33°50′35.5″S, 115°11′38.0″E) as part of a collaboration between Dairy Australia, UWA, and the Department of Primary Industries and Regional Development (
. We sampled the soil from the top 10 cm of the profile and transported it to UWA (at 0°C) on the same day, and potted it on the next day. Before potting, the soil was screened, thoroughly sieved over a 5-mm fraction, and mixed to minimize bacterial spatial microclimate effects or bias and to ensure maximum achievable replication for all treatments. Soil bulk density was also determined (
), as was the ratio of water to soil for watering up to (but not over) approximately 60% field capacity. The soil was placed in 10-cm plastic-lined pots, and the surface was covered with plastic beads to minimize moisture loss (0.6 kg of soil/pot).
There were 3 N treatments, and each was calculated for elemental N using yearly known farmer application rates supplied to the Department of Primary Industries and Regional Development and Dairy Australia specific for dairy district practice and for the highest rate to simulate “hotspot” urine-like effects (
). The 3 treatments were 0, 180, and 912 kg of N/ha per year. The total amount received for each pot was the equivalent dose per week, calculated on daily doses of 0, 0.5, and 2.5 kg of N/ha per day for the 0, 180, and 912 kg of N/ha per year treatments, respectively. The inorganic N fertilizer consisted of a 50:50 mix of urea and ammonium sulfate dissolved in aqueous solution. Nitrogen was applied weekly (in aliquots) and adjusted with additional water during the same fertilization event to maintain pots at approximately 60% field capacity.
The pasture species treatments were (1) Italian ryegrass (Lolium multiflorum), (2) white clover (Trifolium repens), and (3) Italian ryegrass and white clover sown at 50:50 (30 plants per pot;
). There were 3 replicates of each treatment and 2 cycles of plant growth consisting of 2 generations of pasture plants grown to full maturity and harvested at the height of 1 cm above the soil level (harvested shoots were removed).
After the first harvest, the original seed stock seeds were resown into the same soil profile at the same seeding rate without significant soil disturbance after a few weeks. Fertilizer treatments were maintained throughout the 2 cycles. No fertilizer was added between growth cycles. At the final harvest (after 2 × 4-mo periods of continual growth), soil cores (1 cm2) from the top 0 to 5 cm of each pot were taken and placed in separate vials and stored in a freezer (−20°C) for wet molecular chemistry, downstream sequencing, and bioinformatics.
At second harvest, shoots were harvested and dried, and the biomass was determined (shoot dry weight). Soil was oven-dried and analyzed by CSBP Laboratories in Bibra Lake, Western Australia, for (1) Colwell phosphorus (mg/kg) and Colwell potassium (mg/kg; methods 9B and 18A1, respectively;
Genomic DNA was isolated from sample vials using the PowerSoil DNA Total Isolation Kit (Mobio Laboratories Inc., Solana Beach, CA). We optimized the manufacturer's protocol for our soil samples by using a bead-beating lysis procedure and by adding 200-μm-diameter microbeads (0.2 g) to each sample vial (
). The DNA extract quality and quantity were checked using a Nanodrop 1000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA). Ribosomal 16S rRNA amplicons were generated via PCR (
). Sequencing was performed using an Ion PGM OT 200 semiconductor (Thermo Fisher Scientific). Briefly, within R (Lucent Technologies, Murray Hill, NJ) packages and implemented through the Anaconda distribution (Anaconda, Austin, TX), Phyloseq (https://github.com/joey711/phyloseq), and biocLite (http://bioconductor.org/bioclite.r; Bioconductor, Huber et al., 2015), FastQ files were checked for accuracy against the latest available benchmark standards (mid-2018 release). For this process, we used Windows OS (Microsoft Corp., Redmond, WA) running Ubuntu 14.04 LTS (Canonical Ltd., London, UK) through a Virtual Box (Oracle Corporation, Redwood Shores, CA) virtual machine within a Python 3.6.5 reconfigured QIIME pipeline (
). The phyla representing nonclassical taxonomic groups or less than 1% of the entire sample size were also removed, and the final taxonomic data sets were compiled (
Data sets were analyzed as follows. First, a range of statistical ANOVA correlations were performed in GenStat (VSN International Ltd., Hemel Hempstead, UK) for significance based on the least significant differences (LSD). This included each N treatment and plant combination (shoot and soil data). Next, relationships between bacterial community diversity and relative abundance and soil data (eigenvectors) were assessed using canonical correspondence analysis in Microsoft Excel add-in XLSTAT ADA (Addinsoft, Paris, France). Then, pie chart representations were made for major taxonomic phyla across plant associations and N treatments. Next, histograms were made representing significant expanded phylum (class, order, and family) versus plant associations and N treatments. Finally, an assessment was then made in relation to all quantitative data sets with regard to the hypotheses (
Effects of N Treatments on Pasture Plant Species (White Clover and Italian Ryegrass)
Application of N to the plant swards increased (P < 0.001) shoot growth for both white clover and Italian ryegrass when grown alone. The highest level of N application influenced the ratio of white clover to Italian ryegrass in the mixed sward (Table 1). When the equivalent of 180 kg of N/ha per year was applied, the ratio of white clover to Italian ryegrass was reduced from 0.16 to 0.08 due to an increase in biomass of Italian ryegrass, but there was no change in white clover biomass compared with 0 kg of N/ha per year (Table 1). When 912 kg of N/ha per year was applied, white clover was completely outcompeted by Italian ryegrass (Table 1). Soil ammonium was higher under white clover than under Italian ryegrass (Table 1). Nitrate was very low in the soil for all treatments (2 mg/kg of soil). Analysis of variance of soil data indicated that the application of N was highly significant (P < 0.001) in its effects on ammonium (LSD = 4.49 mg/kg), sulfur (LSD = 12.2 mg/kg), potassium (LSD = 18.8 mg/kg), and pH (water; LSD = 0.16; Figure 1).
Table 1Dry weight shoot biomass (g/pot) of white clover and Italian ryegrass and soil ammonium (mg of N/kg of soil; n = 3; ±SE) when grown in dairy soil for 4 mo with 1 control rate and 2 N treatment rates (0, 180, and 912 kg of N/ha per year) relative to the already-prevailing N soil concentrations, plus 5% least significant differences (LSD) for plant growth
Figure 1Canonical correspondence analysis (eigenvector analysis) of soil data versus sequenced 16 S rRNA dominant phyla (italicized) shows the variance that the dominant soil phyla have based on N treatments and plant species associations (including 5% SE bars). The proximity of phylum (open circles) to data vectors (bold) Colwell P (mg/kg), Colwell K (mg/kg), KCl 40 S (mg/kg), organic carbon (%), nitrate N (mg/kg), ammonium N (mg/kg), electrical conductivity (dS/m), and pH (H2O; red squares) indicates increasing correlation to that soil data point. Conversely, decreasing proximity to the data vectors indicates decreasing correlation (based on the soil data). The length of each vector indicates the significance of the relationship that the vector (soil property data) has to the overall community relative abundance and diversity makeup for the total sample size. In this canonical correspondence analysis (CCA), the first 2 axes of the eigenvector plots explain more than 77.1% of the total variance, which represents good sample separation about the (x, y) axis.
Effects of N Application and Pasture Plant Species on Dominant Soil Bacterial Phyla and Physiologically Significant Subgroups (Order and Family)
Italian ryegrass grown alone with the highest N application supported an increase in the relative abundance of Proteobacteria (Figure 2) and α-Proteobacteria (Figure 3) and a slight decrease in the relative abundance of Acidobacteria, whereas the relative abundance of Actinobacteria and Firmicutes remained largely unaffected (Figure 2). For white clover grown alone with the highest N application, there was a decrease in the relative abundance of Proteobacteria and increases in the relative abundance of Firmicutes, Acidobacteria, and Actinobacteria (Figure 2). In soil from the mixed ryegrass and clover pasture sward, the relative abundance of Proteobacteria and Acidobacteria decreased slightly, whereas that of Actinobacteria and Firmicutes increased slightly with increasing N application. The effects of N application and plant sward composition were minimal on bacterial groups that were recorded in the soil at low relative abundance.
Figure 2Treatment-specific relative abundance pie charts for domain: bacteria (phyla). Each pie chart indicates the relative abundances of the most dominant phyla present in each of the specific nitrogen (N) treatment pots based on each plant association: Italian ryegrass (ryegrass), white clover (clover), and mixed (ryegrass + clover). Among all phyla displayed are 4 dominant phyla from all pots based on N treatment (0, 180, and 912 kg of N/ha per year for zero, low, and high, respectively), including Proteobacteria, Acidobacteria, and Firmicutes, with lesser abundant groups including the equally distributed across all treatments “bacteria–other” making up total species diversity and relative abundance within the entire sample set (all plant association pots and all treatments). The number 1 in the first pie chart and at the top of the phyla key is for color identification. The arrows indicate how to correlate pie slices with the phyla (clockwise-downward manner).
Figure 3Diversity and relative abundance within phylum Proteobacteria in response to 3 nitrogen (N) treatments (0, 180, and 912 kg of N/ha per year for zero, low, and high, respectively) and 3 plant treatments [Italian ryegrass (rye), white clover (clover), and the combination of Italian ryegrass and white clover (clover + rye)]. The phyla displayed correspond to the colors shown on the right side of the chart. The order in which the phyla appear for the treatments is consistent.
Within Proteobacteria, application of N increased the relative abundance of Caulobacterales when white clover was grown alone, and the relative abundance of Rhizobiales decreased (Figure 3). Within Firmicutes, N application to white clover reduced the relative abundance of some families of Baciliales and increased the relative abundance Clostridiaceae (Figure 4). Furthermore, there was an increase in the relative abundance of families Veillonellaceae and Peptococcaceae with N application when white clover was grown alone (Figure 4).
Figure 4Diversity and relative abundance within phylum Firmicutes in response to 3 nitrogen (N) treatments (0, 180, and 912 kg of N/ha per year for zero, low, and high, respectively) and 3 plant treatments [Italian ryegrass (rye), white clover (clover), and the combination of Italian ryegrass and white clover (clover + rye)]. The phyla displayed correspond to the colors shown on the right side of the chart. The order in which the phyla appear for the treatments is consistent.
For white clover and Italian ryegrass grown either alone or in combination on dairy farms, 50:50 urea and ammonium sulfate application in accordance with industry standards and legislation is likely to have a marked effect on soil bacterial communities related to nutrient uptake by plants in addition to overall plant biomass effects when N exceeds 180 kg of N/ha per year (
A comparative nitrogen balance and productivity analysis of legume and non-legume supported cropping systems: The potential role of biological nitrogen fixation.
. We also expect that the bacterial community changes are associated with N cycling changes; however, because quantifying N cycling and N cycling efficiency is not as simple as assessing N required for pasture plant biomass yield (
), many of the microbiological responses occurring in these production environments are often overlooked. However, they remain highly beneficial sources for improved management (
In this experiment, the concentration of ammonium under white clover was more significant in its effect on soil bacteria than nitrate (ammonia was less correlated than nitrate) based on eigenvector canonical correspondence analyses (Figure 1), which is consistent with the observations of
). Because Rhizobiales was the most dominant order present across all treatments, it needs to be mentioned that the importance of their presence is consistently underrepresented in linking soil quality to soil function and ecosystem services (although not measured for functionality in our experiment;
There was a marked influence of the higher N application on pasture composition for white clover and Italian ryegrass grown together. Clover may mitigate production loss under increased stocking rate in sandy soils where constant N concentrations range from 180 up to (the maximum legislated) 320 kg of N/ha per year. Application rates above this range might trade off beneficial soil processes (especially around N hotspots) due to the inability of clover to compete with ryegrass (
). Interactions between N fertilizer, plant genotype, and soil bacterial community structure contribute to the effects of N on soil bacterial communities rather than N fertilizer alone.
The increase in the relative abundance of the Caulobacterales (within the β-Proteobacteria) with N application for white clover grown alone but not in combination with Italian ryegrass or Italian ryegrass alone corroborates a previous hypothesis that Caulobacterales may dominate N-fixing environments associated with clover (
). The ability of bacteria within Caulobacterales to regulate and transcribe orthologs (through speciation) to associate with N fixation processes is an essential consideration for dairy soils, especially for efficient use of N (including N fixation), as it may include previously unidentified biological pathways and needs investigation (
in: Brenner D. J Krieg N.R. Staley J.T. Bergey's Manual of Systematic Bacteriology: The Alpha-, Beta-, and Epsilon Proteobacteria. 2nd ed. Vol. 2. Part C. Springer Verlag,
New York, NY2005: 555-556
). The relative abundance of Firmicutes was largely unaffected across all N and plant treatments except within the order Clostridia. Increases in the relative abundance of Veillonellaceae and Peptococcaceae after N application to white clover raise the possibility of resistance to acidified soil environments. The link between the increasing negative correlation between N and soil pH variation and the correlation between the presence of Firmicutes and organic carbon and pH supports this assumption (Figure 1). Veillonellaceae decreased again, indicating a buffering capacity threshold for N applied at a rate of 180 kg of N/ha per year < X < 320 kg of N/ha per year (
Determining the fertilizer phosphorus requirements of intensively grazed dairy pastures in south-western Australia with or without adequate nitrogen fertilizer.
The contribution of cattle urine and dung to nitrous oxide emissions: Quantification of country specific emission factors and implications for national inventories.
Nitrogen fertilizer applied at current levels to dairy pastures can alter the structure of soil bacterial communities as well as pasture plant dominance. We showed a marked effect of weekly applications of N fertilizer on the relative abundance of soil bacteria for swards of white clover (Trifolium repens) and Italian ryegrass (Lolium multiflorum) in a long-term glasshouse experiment. There is evidence to support our hypothesis that sustained and elevated N application alters the dominance of N cycling bacteria in this dairy soil, with potential for reduced N use efficiency, which could lead to increased prevalence of oligotrophic soil conditions (
Climate change induces shifts in abundance and activity pattern of bacteria and archaea catalyzing major transformation steps in nitrogen turnover in a soil from a mid-European beech forest.
). Characterization of bacteria from the orders Rhizobiales and Clostridia, within the α-Proteobacteria and Firmicutes, respectively, may provide a litmus test for oligotrophic soil conditions in environments of poor nutrient binding (
). Further understanding of how N fertilizer use influences the functional aspects of bacterial communities associated with N cycling in dairy soils could assist in the development of better sustainability policies related to N application that takes into account group-level concepts of the soil microbiome and surrounding holobiont to minimize related risk.
ACKNOWLEDGMENTS
We acknowledge the contribution of the following people and industry groups. Dairy Australia (Southbank, Victoria) funded this research in collaboration with the University of Western Australia (Crawley, Western Australia). The Department of Primary Industry and Regional Development in Western Australia (previously Department of Agriculture and Food Western Australia, South Perth, Western Australia) provided logistical support. Zakaria Solaiman, Ian Waite, Richard Alcock, Sasha Jenkins, Michael Smirk, and Darryl Roberts (the University of Western Australia), Jim Cook (University of California, Davis), and Kim Angus (https://www.facebook.com/ConsultKimToodyay, PO Box 379, Toodyay, Western Australia), provided valuable advice and editorial support during this study.
REFERENCES
Abbott L.
Mele P.
Jenkins S.
Svatos K.B.W.
Bougoure D.
Hayden H.
Waite I.
Microbial community in response to nitrogen use in dairy soils.
in: Proc. 5th Australasian Dairy Sci. Symposium: Dairy Production Systems, Australasian Dairy Science Symposium (Tallygaroopna, Victoria). 2012: 193
Determining the fertilizer phosphorus requirements of intensively grazed dairy pastures in south-western Australia with or without adequate nitrogen fertilizer.
The contribution of cattle urine and dung to nitrous oxide emissions: Quantification of country specific emission factors and implications for national inventories.
Climate change induces shifts in abundance and activity pattern of bacteria and archaea catalyzing major transformation steps in nitrogen turnover in a soil from a mid-European beech forest.
A comparative nitrogen balance and productivity analysis of legume and non-legume supported cropping systems: The potential role of biological nitrogen fixation.
Direct 16S rRNA-seq from bacterial communities: A PCR-independent approach to simultaneously assess microbial diversity and functional activity potential of each taxon.
in: Brenner D. J Krieg N.R. Staley J.T. Bergey's Manual of Systematic Bacteriology: The Alpha-, Beta-, and Epsilon Proteobacteria. 2nd ed. Vol. 2. Part C. Springer Verlag,
New York, NY2005: 555-556
30634-4&id=gr1.jpg){kind=link}
30634-4&id=gr2.jpg){kind=link}
30634-4&id=gr3.jpg){kind=link}
30634-4&id=gr4.jpg){kind=link}