Skip to main content
NCBI home page
Ecology and Evolution logoLink to Ecology and Evolution
. 2018 Jul 22;8(16):8217–8230. doi: 10.1002/ece3.4346

Linking microbial co‐occurrences to soil ecological processes across a woodland‐grassland ecotone

Samiran Banerjee 1,2,✉, Peter H Thrall 1, Andrew Bissett 3, Marcel G A van der Heijden 2,4,5, Alan E Richardson 1
PMCID: PMC6145019  PMID: 30250697

Abstract

Ecotones between distinct ecosystems have been the focus of many studies as they offer valuable insights into key drivers of community structure and ecological processes that underpin function. While previous studies have examined a wide range of above‐ground parameters in ecotones, soil microbial communities have received little attention. Here we investigated spatial patterns, composition, and co‐occurrences of archaea, bacteria, and fungi, and their relationships with soil ecological processes across a woodland‐grassland ecotone. Geostatistical kriging and network analysis revealed that the community structure and spatial patterns of soil microbiota varied considerably between three habitat components across the ecotone. Woodland samples had significantly higher diversity of archaea while the grassland samples had significantly higher diversity of bacteria. Microbial co‐occurrences reflected differences in soil properties and ecological processes. While microbial networks were dominated by bacterial nodes, different ecological processes were linked to specific microbial guilds. For example, soil phosphorus and phosphatase activity formed the largest clusters in their respective networks, and two lignolytic enzymes formed joined clusters. Bacterial ammonia oxidizers were dominant over archaeal oxidizers and showed a significant association (p < 0.001) with potential nitrification (PNR), with the PNR subnetwork being dominated by Betaproteobacteria. The top ten keystone taxa comprised six bacterial and four fungal OTUs, with Random Forest Analysis revealing soil carbon and nitrogen as the determinants of the abundance of keystone taxa. Our results highlight the importance of assessing interkingdom associations in soil microbial networks. Overall, this study shows how ecotones can be used as a model to delineate microbial structural patterns and ecological processes across adjoining land‐uses within a landscape.

Keywords: ammonia oxidizers, ecotone, extracellular enzymes, keystone taxa, kriging, microbial networks

1. INTRODUCTION

Ecotones between adjacent ecosystems or biomes that harbor contrasting plant communities represent useful areas for investigation, as they support unique ecological dynamics (Anadón, Sala, & Maestre, 2014; Archer & Predick, 2014). However, recent studies show that ecotones are highly responsive to environmental change and this is especially true for ecotones in the arid and semi‐arid regions such as the ones in Australia (Delgado‐Baquerizo et al., 2014). Grassland‐woodland ecotones around the world are subject to dynamic shifts toward an unstable state, and this has received considerable research attention in recent years (Bradford, Schlaepfer, Lauenroth, & Burke, 2014; Sala & Maestre, 2014). Ecotones encompass interactions occurring between adjoining systems and are useful because the local effects of shifts in vegetation can be explicitly assessed independently of the environmental variability that may occur over broader spatial scales (Gosz, 1993). In addition, such areas can reveal the edge effect between two adjacent habitats (Lacasella, Gratton, & De Felici, 2015; Malmivaara‐Lämsä et al., 2008; Murcia, 1995). Edge effect is the result of the abiotic and biotic interactions between adjoining habitats when the habitats are separated by an abrupt transition (sensu Murcia, 1995).

Previous studies have focused on ecotones to examine impacts on community structure (e.g., species diversity and distribution patterns) as well as a range of ecological processes such as above‐ground biotic interactions, hydrology, fire dynamics, and responses to climate change (Archer & Predick, 2014; Eldridge et al., 2011; Ratajczak, Nippert, Briggs, & Blair, 2014). In contrast, understanding of belowground communities and interactions within the soil microbiota has received less attention (Malmivaara‐Lämsä et al., 2008). Soil microbiota provide a range of important ecosystem services including soil aggregation, organic matter decomposition, nutrient cycling, and mutualistic and pathogenic interactions with plants (Bardgett & van der Putten, 2014; Killham, 1990; Richardson, Barea, McNeill, & Prigent‐Combaret, 2009; Schimel & Schaeffer, 2012; van der Heijden, Bardgett, & Van Straalen, 2008). While patterns across ecotones have been observed for some soil parameters (e.g., moisture, temperature, carbon storage, etc.) and macrobiota in previous studies (Lacasella et al., 2015; Magura, 2017; Schmidt, Jochheim, Kersebaum, & Lischeid, 2017), little information is available on soil microbiome (Malmivaara‐Lämsä et al., 2008).

The soil microbiome comprises a vast diversity and abundance of different microbial groups and complex trophic interactions (Bardgett & van der Putten, 2014; van der Heijden et al., 2008). Microbial co‐occurrence networks can reveal associations among network members and yield insight into microbiome functioning (Bissett, Brown, Siciliano, & Thrall, 2013; Cardona, Weisenhorn, Henry, & Gilbert, 2016; Faust & Raes 2012; Fuhrman, 2009;). For example, patterns of microbial co‐occurrence have been demonstrated for a diverse range of aquatic and terrestrial environments (Banerjee, Baah‐Acheamfour et al., 2016; Barberán, Bates, Casamayor, & Fierer, 2012; De Menezes et al., 2015; Graham et al., 2017; Shi et al., 2016). Previous studies using network analysis have often only assessed bacterial communities and not fungal or archaeal communities (Banerjee, Baah‐Acheamfour et al., 2016; Barberán et al., 2012; Shi et al., 2016; Vick‐Majors, Priscu, & Amaral‐Zettler, 2014). Thus, the roles of these latter groups have been underrepresented in microbial network analyses and only a few studies have investigated associations all three kingdoms (Ma et al., 2016; Steele et al., 2011). Moreover, network analysis provides a statistical tool to identify keystone taxa that play a key role in microbiome structure and functioning (Banerjee, Schlaeppi, & van der Heijden, 2018; Power et al., 1996). A number of studies have used network‐based scores to identify putative keystone taxa in different environments (Hartman et al., 2018; Lupatini et al., 2014; Shi et al., 2016; Vick‐Majors et al., 2014) and linked their abundance to soil nutrient cycling processes (Banerjee, Kirkby et al., 2016; Li, Chen, Zhang, Yin, & Huang, 2017).

A major challenge in ecology is to link microbial co‐occurrences to processes that contribute to soil function. For example, extracellular enzymes are ubiquitous in soil environments and play critical roles in ecosystem functioning through mediation of carbon (C), nitrogen (N), and phosphorus (P) mineralization, thus, facilitating soil organic matter decomposition (Burns, 1982). Soil enzyme activities have often been used as indicators of soil health and microbial function (Allison & Vitousek, 2005; Saiya‐Cork, Sinsabaugh, & Zak, 2002; Sistla & Schimel, 2013). Likewise, ammonia oxidation is important for soil nutrient availability as it is a key step for nitrification in which ammonia is converted to hydroxyl amine and subsequently to nitrite and nitrate (Kowalchuk & Stephen, 2001). The functional gene, amoA, is present in both bacteria and archaea and has been used in many studies to quantify the abundance of ammonia oxidizers in different environments (Di et al., 2009; Jia & Conrad, 2009; Leininger et al., 2006). Spatial patterns of ammonia oxidizers across ecotones can unravel niche differentiation and partitioning among bacteria and archaea based on nutrient availability (Prosser & Nicol, 2012). However, few studies have assessed microbial co‐occurrences in relation to soil nitrification and enzyme activities.

In a previous study, we found similar spatial patterns for a wide range of soil properties and extracellular enzyme activities across two native woodland‐grassland ecotones (Banerjee, Bora, Thrall, & Richardson, 2016). In this study, we further investigated patterns of abundance, diversity, and co‐occurrence for archaeal, bacterial, and fungal communities in one of these ecotones. Our overall hypothesis was that soil microbial properties are different in the transition zone than either of the adjacent woodland or grassland communities. A multifarious approach was used to address the following specific questions: (a) How do the spatial structure, composition, and co‐occurrences of soil archaeal, bacterial, and fungal communities change across a woodland‐grassland ecotone?; (b) How are ammonia oxidizing bacteria and archaea linked to potential nitrification across such ecotones?; (c) Is the composition of microbial networks related to soil properties and ecological processes?; and (d) Which soil properties drive the abundance of microbial keystone taxa across the woodland‐grassland ecotone?

2. MATERIALS AND METHODS

2.1. Study site and soil sampling

The study was conducted at a native woodland adjacent to native grassland within Namadgi National Park (35.6667° S, 148.950° E) in the Australian Capital Territory. The woodland was dominated by Eucalyptus spp. with scattered Acacia dealbata and Acacia implexa in the understorey (Banerjee, Bora et al., 2016). The grassland consisted of a mix of native grasses including Austrodanthonia sp. and Themeda sp. The mean annual rainfall at Namadgi National Park is 777.3 mm with 4.4°C and 17.3°C minimum and maximum mean annual daily temperature (http://www.weatherzone.com.au). The site was dominated by Brown Sodosols; these are typically composed of fine sandy clay loams with 10%–20% clay content (Isbell, 2002). A 50 m length × 20 m width sampling plot was established across the woodland and grassland (i.e., extending 25 m into both the woodland and grassland; Supporting Information Figure S1). A rectangular grid design consisting of 55 nodes was employed, with adjacent nodes separated by a linear or perpendicular distance of 5 m. Soil samples were collected in September 2013 by sampling at each node. Each sample consisted of a composite of 10 individual soil cores (4 cm diameter) collected at 0–10‐cm depth within a 10‐cm radius of each sampling node. The corer was cleaned between nodes and soil samples were placed on ice in a cooler box for transfer to the laboratory. Samples were processed and subsampled on the same day as collection by removing plant materials, homogenizing, and passing through a 2‐mm sieve.

For vegetation type comparisons, the experimental grid was divided into three components: woodland, transition, and grassland (20, 10, and 20 m, respectively; Supporting Information Figure S1). This assignment was based on field observations that the first 20 m (from the plot perimeters) was consistently woodland and visually homogeneous. Similarly, the last 20 m (toward the end of the sampling plot) was consistently grassland and visually homogeneous. The visual assignment of these components was substantiated with data from soil analyses and extracellular enzyme activities (Banerjee, Bora et al., 2016).

2.2. Soil analyses and quantitative PCR

Detailed analyses of soil properties and extracellular enzyme activities of all sample points across the ecotone were described previously by Banerjee, Bora et al. (2016). Briefly, gravimetric soil moisture content, pH, dissolved organic carbon (DOC), dissolved organic nitrogen (DON), ammonium (NH4), nitrate (NO3), percentage of total carbon (%C), and nitrogen (%N), total, inorganic, and organic phosphorus (P) were measured for each soil sample. Activities of extracellular enzymes such as cellulolytic (β‐1,4‐glucosidase and cellobiohydrolase), lignolytic (peroxidase and phenol oxidase), and complex N or P depolymerizing (acid phosphatase and chitinase) were determined and expressed in units of nmol hr−1 g−1. Potential nitrification activity (PNR) was measured according to Hart, Stark, Davidson, and Firestone (1994). Briefly, 10 g of sieved, field‐moist soil was mixed with 100 ml of solution containing 0.2 M monopotassium dihydrogen phosphate, 0.2 M dipotassium hydrogen phosphate, and 50 mM ammonium sulfate. The mixture was shaken at 100 rpm at room temperature for 24 hr. An aliquot of 10 ml was collected and nitrite (NO2), and NO3 were measured by colorimetric assay and spectrophotometer. The 24‐hr time point was selected based on a time course assay, and final values were calculated by comparing with time zero samples.

For all 55 soil samples, DNA was extracted from 0.25 g soil using the PowerSoil DNA isolation kit (MoBio, Carlsbad, CA) following manufacturer's instructions. DNA concentrations were determined by NanoDrop spectrophotometry (NanoDrop, Wilminghton, DE). The abundances of bacterial, archaeal, and fungal genes were determined by quantitative real‐time PCR using the Qiagen QuantiTect™ SYBR® Green PCR Master Mix (Qiagen Inc., Victoria, Australia), a Cavro® Omni Robot (Tecan Group Ltd., Seestrasse, Switzerland), and an ABI 7900 real‐time PCR machine (Applied Biosystems, Foster City, CA). Bacterial 16S rRNA and amoA genes were quantified using 338F and 518R (Lane et al., 1985; Muyzer, de Wall, & Uitterlinden, 1993) and amoA 1F and amoA 2R (Rotthauwe, Witzel, & Liesack, 1997) primers, respectively. Archaeal 16S rRNA and amoA genes were quantified using 771F and 957R (Ochsenreiter, Selezi, Quaiser, Bonch‐Osmolovskaya, & Schleper, 2003) and Arch‐amoA 104F and Arch‐amoA 616R (Alves et al., 2013) primers, respectively. Fungal ITS was quantified using the primer set ITS1F and ITS4 as reported by White, Bruns, Lee, and Taylor (1990) and Gardes and Bruns (1993). Details of thermal cycling conditions and quality assessment are provided in Supporting Information.

2.3. Amplicon sequencing

Sequencing was carried out using an Illumina MiSeq following a previously published protocol (Bissett et al., 2016). For sequencing, samples were randomly selected equally from three zones with six samples in each habitat component (Supporting Information Figure S1). Of the 11 rows in the sampling grid, nine rows (three for each of woodland, transition, and grassland) were selected for selection of samples, and two rows (one between two adjacent components) were selected as a buffer. Amplicons targeting the bacterial 16S rRNA gene, archaeal 16S rRNA gene, and fungal ITS genes were prepared using the 27F–519R (Lane et al., 1985; Muyzer et al., 1993), A2F–519R (DeLong, 1992) and ITS1F–ITS4 (Gardes & Bruns, 1993; White et al., 1990) primer sets, respectively. For all amplicons, Illumina 300 bp paired‐end sequencing was performed at the Australian Genome Research Facility (Melbourne, Australia). For bacterial and archaeal16S rRNA genes, the quality of R1 and R2 reads was determined using FastQC (Andrews, 2010). For fungal ITS, only R1 sequences were used that assessed the ITS1 region (Bissett et al., 2016). Reads were trimmed to remove base pairs from the end of reads after read quality per sample declined (10 and 50 bp for read1 and read2, respectively). Reads were trimmed by as many base pairs as possible while still leaving an overlap for reliable merge of R1 and R2 reads and then merged using FLASH (Magoč & Salzberg, 2011). FASTA format sequences were extracted from FASTQ files, and sequences <400 bp, with homopolymer >8 bp or containing ambiguous bases were removed using the Mothur (V1.38.0) (Schloss et al., 2009). Singletons were removed, and OTUs were defined by clustering at 97% similarity with the “cluster_otus” function in UPARSE (Edgar, 2013). Sequences were then mapped to these OTUs to produce a OTU abundance table using the “usearch_global” function in USEARCH (Edgar, 2010) and classified according to SILVA v102 using the Naïve Bayesian classifier as implemented in Mothur (Wang, Garrity, Tiedje, & Cole, 2007). For fungal ITS1 region sequences, FASTA files were extracted from FASTQ files, and complete ITS1 regions were extracted from R1 reads using ITSx (Bengtsson‐Palme et al., 2013). Partial ITS1 sequences and those not containing ITS1 were discarded. Remaining ITS1 sequences were used for OTU picking and OTU table production using the same methods described above for 16S rRNA genes. OTUs were classified as above against the ITS fungal database UNITE (V6) (Kõljalg et al., 2005). The total number of OTUs in each sample varied from 825 to 1,832 for bacteria, 15–214 for archaea, and 222–538 for fungi (Supporting Information Figure S2).

2.4. Statistical analyses

Alpha diversity indices such as species richness, Pielou's evenness, and Shannon‐Weaver diversity were calculated from bacterial, archaeal, and fungal OTU tables using the vegan package (Oksanen et al., 2017) in R v3.4 (R Development Core Team, 2016). Microbial beta diversity patterns were assessed on square‐root transformed data with principal coordinate analysis (PCoA) using Bray–Curtis dissimilarity matrix in PRIMER‐E (PRIMER‐E, Plymouth, UK). The effect of habitat edge on microbial communities was assessed by performing PERMANOVA with 999 permutations. Spatial variability was determined using geostatistical analyses in GS+ version 10 (Gamma Design Software, Plainwell, MI, USA). Spatial dissimilarity was computed by calculating the isotopic semivariance (Goovaerts, 1998). Semivariograms were calculated with a minimum of 30 sample pairs per lag class (Journel & Huijbregts, 1978). Spatial dependency (SPD) was calculated as SPD = C/(C + C0), where C is the structural variance, C0 is the nugget, and C + C0 is the sill. Values of SPD vary from 0 (no spatial dependence) to 1 (high spatial dependence). Cross‐validation was performed on semivariograms to insure their suitability for kriging. Regression coefficients, standard errors, and r 2 values of cross‐validation plots were checked before kriging. Ordinary kriging was used to interpolate values between sampling points (i.e., unmeasured locations). Ordinary kriging is a spatial interpolation technique that employs the local mean in the estimation and computes values by selecting weights to minimize estimation variance (Isaaks & Srivastava, 1989). Finally, spatial contour maps were generated using GS+ version 10.

2.5. Network analysis

Co‐occurrences between bacterial, archaeal, and fungal communities were assessed by performing network analysis using the maximal information coefficient (MIC) scores in the MINE statistics (Reshef et al., 2011). MIC score reveals positive, negative, and nonlinear associations among OTUs. To minimize pairwise comparisons and manage the false discovery rate (FDR; Benjamini & Hochberg, 1995), network analysis was performed only on OTUs that were present in at least two samples. This resulted in 1,006 bacterial, 105 archaeal, and 697 fungal OTUs. Relationships between microbial OTUs, vegetation types, and soil ecological processes that were significant at an FDR of 5% were then visualized in Cytoscape version 3.4.0 (Shannon et al., 2003). To indicate the most important interactions, only strong positive (r > 0.8), strong negative (r < −0.8), and strong nonlinear (MIC – ρ 2 > 0.8) relationships were shown in the networks. We compared our network against its randomized version using the Barbasi–Albert model available in Randomnetworks plugin in Cytoscape v2.6.1. Structural attributes of the overall network such as clustering coefficient, average degree, degree distribution, and mean shortest path were significantly different from a random network with equal number of nodes and edges. The NetworkAnalyzer tool was then used to calculate network topology parameters. Relationships between microbial co‐occurrences and ecological processes were visualized in Cytoscape. The OTUs with high mean degree (≥199), high closeness centrality (>0.475), and low betweenness centrality (<0.025) scores were considered as keystone taxa (Berry & Widder, 2014). Random Forest Analysis (Breiman, 2001) was used to explore the edaphic drivers of keystone taxa. The analysis was conducted using the rfPermute package version 2.1.5 in R (Archer, 2016). The best predictor was identified based on statistical significance with 999 permutations.

3. RESULTS

3.1. Spatial patterns of microbial abundance and potential nitrification

Ordinary kriging plots revealed that the overall abundance of bacterial 16S rRNA, archaeal 16S rRNA, and fungal ITS genes were significantly (p < 0.05) higher in the woodland samples than in adjacent grassland samples with a distinction across the transition zone (Figure 1 upper panel; Table 1). In general, bacterial abundance was consistently high along the grid with the number of gene copies varying between 107 and 108 per gram of dry soil. Conversely, archaeal (105–107) and fungal (104–107) abundance showed much larger variation in copy numbers. A number of small areas were found near the transition where the abundances of all three microbial kingdoms were high. Soil PNR closely resembled the spatial patterns of bacterial ammonia oxidizers (Figure 1 lower panel). Similar to overall microbial abundance, the abundance of ammonia oxidizers was significantly (p < 0.05) higher in the woodland than the grassland, and a visually distinct transition zone was also noted. Bacterial amoA displayed a significantly (R2 = 0.317; p < 0.001) positive association with PNR across the ecotone (Supporting Information Figure S3). All five gene abundances and PNR displayed high spatial dependency (SPD = 0.631–0.999) and operated between 14 m and 36 m (Supporting Information Table S2). Spherical and Gaussian models showed significant (R 2 = 0.285–0.900; p < 0.01) spatial structure for all properties tested. Bacterial amoA and PNR structured at a smaller spatial range (14.7–23.7 m) than archaeal amoA, which had a spatial range >50 m.

Figure 1.

Figure 1

Open in a new tab

Geostatistical kriging plots showing the spatial patterns of microbial abundance and potential nitrification across the woodland‐grassland ecotone. A 50 m × 20 m grid was established across woodland and grassland in Namadgi National Park, Australia. Ordinary kriging was performed after semivariance analysis and cross‐validation. The dotted lines on kriging maps indicate the boundary between woodland in the upper half of the map and grassland in the lower half

Table 1.

Soil microbial indices and potential nitrification rate across the woodland‐grassland ecotone at the Namadgi National Park, Australia

Soil microbiota and activities Ecotone components
Woodland Transition Grassland
Abundance (Log copies g−1 dry soil)
Bacterial 16S rRNA 8.40 (0.05)a 8.29 (0.08)a 7.84 (0.12)b
Archaeal 16S rRNA 6.91 (0.11)a 6.96 (0.21)a 6.10 (0.16)b
Fungal ITS 7.69 (0.09)a 7.06 (0.16)a 5.39 (0.66)b
Bacterial amoA 4.55 (0.25)a 4.82 (0.20)a 3.54 (0.34)b
Archaeal amoA 5.86 (0.28)a 5.12 (0.20)a 5.51 (0.19)a
Richness
Archaea 42.2 (7.71)a 21.5 (1.91)b 22.1 (1.30)b
Bacteria 1169 (43.10)b 1147 (96.03)b 1395 (21.87)a
Fungi 374 (14.10)a 308 (27.03)b 337 (12.22)ab
Pielou's evenness
Archaea 0.40 (0.04)a 0.09 (0.02)c 0.21 (0.03)b
Bacteria 0.77 (0.01)a 0.80 (0.01)b 0.82 (0.01)b
Fungi 0.61 (0.04)a 0.68 (0.04)a 0.72 (0.03)a
Diversity (Shannon‐Weaver)
Archaea 1.45 (0.15)a 0.27 (0.07)c 0.66 (0.11)b
Bacteria 5.48 (0.08)b 5.67 (0.15)b 6.00 (0.03)a
Fungi 3.57 (0.23)a 3.86 (0.19)a 4.22 (0.21)a
Activity
Nitrification potential (PNR) (μg NO3‐NO2 g−1 dry soil hr−1) 0.13 (0.04)a 0.36 (0.078)b 0.08 (0.01)a
Open in a new tab

Along the grid length (50 m), the first 20 m was woodland, 10 m was transition, and the last 20 m was grassland, resulting (n) in 20, 15, and 20 samples, respectively. For microbial diversity indices, n = 6.

Soil microbial properties were compared between woodland, transition, and grassland by performing one‐way ANOVA with Duncan post hoc test.

Different letters indicate statistical significance at p < 0.05.

3.2. Microbial community structure, taxonomic composition, and diversity

Archaeal, bacterial, and fungal communities displayed a significant habitat effect with the woodland and grassland samples forming distinct clusters and those from the transition zone samples showing a gradient (Figure 2 upper panel). PERMANOVA confirmed such habitat effects for bacterial (Pseudo‐F = 4.79; p < 0.001), archaeal (Pseudo‐F = 4.42; p < 0.001), and fungal (Pseudo‐F = 3.43; p < 0.001) communities. Principal coordinates explained 52%, 81%, and 30% variation in bacterial, archaeal, and fungal communities, respectively. Similar to community structure, taxonomic composition was also influenced by the habitat edge (Figure 2 lower panel). For bacteria, Acidobacteria, Actinobacteria, and Alphaproteobacteria were the dominant members, representing more than 80% of total bacterial abundance across the ecotone. Acidobacteria and Alphaproteobacteria were more abundant in the woodland samples while Betaproteobacteria and Chloroflexi were the most common within the transition zone (p < 0.05). Nitrososphaerales was the predominant archaeal group, especially, in the grassland and transition zone samples where it represented up to 99% of the total archaea. On the other hand, a number of fungal groups showed significant (p < 0.01) differences across the ecotone. For example, Agaricomycetes and Leotiomycetes were 2–3 times more abundant in woodland samples than either the transition zone or grassland samples. Similarly, Saccharomycetes and Dothideomycetes were significantly more abundant (p < 0.05) in the transition zone and Agaricostilbomycetes (p < 0.001) in the grassland. Microbial alpha diversity indices showed contrasting patterns across the ecotone (Table 1). For example, richness, evenness, and diversity of bacteria were significantly (p < 0.05) lower in the woodland samples than the transition zone and grassland. Conversely, archaeal richness, evenness, and diversity were significantly (p < 0.05) higher in the woodland than either the transition zone or the grassland samples. This indicates that the woodland samples harbored a less diverse bacterial community and a highly diverse archaeal community. On the other hand, the relatively N‐ and P‐rich grassland samples supported a diverse bacterial community (Supporting Information Table S1). Fungi also had a significantly higher richness in woodland samples.

Figure 2.

Figure 2

Open in a new tab

Principal coordinate analysis revealing community structure of bacteria, archaea, and fungi in woodland, grassland, and transition zone (upper panel). Stacked bar chart (bottom panel) showing relative abundance of various phyla and classes of bacteria, archaea, and fungi in woodland, grassland, and transition zone

3.3. Microbial co‐occurrences

Network analysis showed distinct co‐occurrences of archaeal, bacterial, and fungal members across the woodland, grassland, and transition zone (Figure 3). The microbial network including the top 1000 MIC scores comprised 324 nodes (260 bacterial, four archaeal, and 60 fungal OTUs; Figure 4). Among the top 10 keystone taxa, six were bacteria and four were fungi (Table 2; Figure 3a). Half of the bacterial keystone taxa belonged to the class Alphaproteobacteria whereas most fungal keystone taxa were members of the Ascomycetes. Two bacterial OTUs belonged to Rhizobiales and Burkholderiales. Random Forest Analysis showed that ammonium, total carbon, and C:N ratio were the major determinants of the abundance of microbial keystone taxa (Figure 3b). The network across the ecotone consisted of 193 nodes of which woodland and grassland were associated with 116 and 74 nodes, and the transition zone connected to two nodes (Figure 3c). The overall network comprised 1767 nodes and had a diameter of 6 and a radius of 4 (Supporting Information Figure S3). Woodland and grassland formed clusters away from the transition zone. While the network was dominated by bacteria, fungal and archaeal nodes were also abundant in the woodland. Overall, microbial and co‐occurrences were considerably different in three habitat components across the ecotone.

Figure 3.

Figure 3

Open in a new tab

(a) Microbial network showing co‐occurrences of bacterial, archaeal, and fungal OTUs. This network of top 1,000 interactions consisted of 324 nodes. Enlarged nodes represent the top ten microbial keystone taxa of which six were bacterial and four fungal. (b) Results of Random Forest Analysis showing the edaphic drivers of microbial keystone taxa. The MSE indicates vector of mean square errors. (c) Microbial co‐occurrences in the woodland, grassland, and transition zone. This network comprised 193 nodes. To indicate the most important interactions, only strong positive (r > 0.8), strong negative (r < −0.8), and strong nonlinear (MIC – ρ 2 > 0.8) relationships were shown in the networks. Oval nodes represent bacterial OTUs, rectangular nodes represent fungal OTUs, and triangular nodes represent archaeal OTUs. Color of the nodes represents different taxonomic groups while green, red, and wavy lines represent positive, negative, and nonlinear relationships, respectively. Only statistically significant (p < 0.05) relationships are shown

Figure 4.

Figure 4

Open in a new tab

Relationship among microbial co‐occurrence, soil properties, and ecological processes. (a) Archaeal, bacterial, and fungal OTUs formed distinct clusters with soil chemical properties. Large clusters such as P consisted of 164 nodes, C:N comprised 76 nodes, and pH consisted of 42 nodes. (b) Clusters of microbial OTUs linked to potential nitrification and extracellular enzyme activities. The cluster of PNR comprised 49 nodes. To indicate the most important interactions, only strong positive (r > 0.8), strong negative (r < −0.8), and strong nonlinear (MIC – ρ 2 > 0.8) relationships were shown in the networks. Oval nodes represent bacterial OTUs, rectangular nodes represent fungal OTUs, and triangular nodes represent archaeal OTUs. Color of the nodes represents different taxonomic groups while green, red, and wavy lines represent positive, negative, and nonlinear relationships, respectively. Only statistically significant (p < 0.05) relationships are shown

Table 2.

Network features and taxonomy of top ten keystone taxa. OTUs with highest degree, highest closeness centrality, and lowest betweenness centrality were selected as the keystone taxa

OTUid Network features Taxonomy
Betweenness centrality Closeness centrality Degree Kingdom Phylum or class Order
Botu781 0.024 0.502 265 Bacteria Acidobacteria Acidobacteriales
Fotu671 0.021 0.510 254 Fungi Eurotiomycetes Chaetothyriales
Fotu695 0.026 0.508 250 Fungi Dothideomycetes Capnodiales
Botu706 0.015 0.499 246 Bacteria Alphaproteobacteria Rhodospirillales
Fotu626 0.013 0.502 241 Fungi Zygomycota Mortierellales
Botu914 0.012 0.501 238 Bacteria Verrucomicrobia Pedosphaerales
Botu257 0.008 0.484 227 Bacteria Alphaproteobacteria Caulobacterales
Botu890 0.015 0.494 220 Bacteria Alphaproteobacteria Rhizobiales
Fotu569 0.011 0.486 210 Fungi Eurotiomycetes Chaetothyriales
Botu81 0.005 0.476 199 Bacteria Betaproteobacteria Burkholderiales
Open in a new tab

3.4. Relationships among microbial co‐occurrences, soil properties, and ecological processes

Microbial co‐abundances were linked to soil properties and ecological processes, with the subnetworks comprising mainly bacterial OTUs (Figure 4; Supporting Information Figure S4). For soil properties, total soil P formed the cluster with maximum nodes followed by C:N ratio and pH (Figure 4a). Mineral N, DON, C:N, and P clusters shared several nodes and were predominantly connected through positive and linear edges. In general, the subnetworks were dominated by Alphaproteobacteria and Actinobacteria OTUs in bacteria. For soil ecological processes, PNR formed a large and distinct cluster away from soil enzymes and predominantly consisted of bacterial nodes but no archaeal nodes, whereas soil enzymes formed individual clusters but were interconnected through shared nodes (Figure 4b). This is especially true for cellobiohydrolase and phosphatase activities that had a large shared guild. β‐glucosidase had the smallest cluster but was connected with both cellobiohydrolase and phosphatase through a fungal node belonging to Dothideomycetes. Similar to soil P content, phosphatase formed the largest clusters. These clusters of soil enzymes were also dominated by bacterial OTUs. Potential nitrification showed consistently significant (p < 0.01) correlations with soil properties and extracellular enzymes (Supporting Information Table S3). Bacterial amoA gene copy number had a strong association with PNR activity in these relatively N‐rich soils, and this was also supported by the fact that bacterial amoA was strongly (p < 0.01) correlated with soil N content. Taken together, microbial co‐occurrences reflected the differences in soil properties and ecological processes.

4. DISCUSSION

4.1. Microbial communities across the ecotone

In this study, we explored the abundance, structure, and co‐occurrences of soil archaea, bacteria, and fungi, and their relationships with relevant soil ecological processes along a woodland‐grassland ecotone. Firstly, using spatial interpolation, we showed how the overall abundance of archaea, bacteria, and fungi changed between woodland and grassland soil samples. The woodland samples had significantly higher microbial abundance than was observed for the grassland samples with a visually distinct transition zone. This higher abundance was observed generally for overall archaea, bacteria, and fungi, and specifically for ammonia oxidizing archaea and bacteria. The overall gene copy numbers of these microbial groups we found in our woodland and grassland soils are comparable to previous studies (Banerjee, Baah‐Acheamfour et al., 2016; Gleeson et al., 2010; Kemnitz, Kolb, & Conrad, 2007; Lauber, Strickland, Bradford, & Fierer, 2008). Secondly, our analyses of α‐ diversity indices showed that the woodland samples had significantly higher diversity of archaea while the grassland samples had significantly higher diversity of bacteria. It should be noted that the grassland soils at our site had significantly higher N and P levels than the woodland soils. Typically, bacteria are more responsive to nutrient‐rich conditions than archaea (Carey, Dove, Beman, Hart, & Aronson, 2016), which indicates their copiotrophic nature (Fierer, Bradford, & Jackson, 2007).

Our results show that the habitat edge between woodland grassland significantly influenced microbial β‐diversity. Microbial communities formed distinct clusters in woodland and grassland samples with the transition zone forming a gradient between those two adjoining systems. Moreover, OTUs belonging to Acidobacteria in bacteria and Agaricomycetes and Leotiomycetes in fungi was significantly higher in the woodland samples than elsewhere. Several members of these oligotrophic microbial groups are involved in wood decomposition, and our results are consistent with previous studies reporting greater abundance of these groups in forest soils (Edwards & Zak, 2011; Jones et al., 2009). Several members of the Agaricomycetes are ectomycorrhizal (van der Heijden, Martin, Selosse, & Sanders, 2015), which may also explain their higher abundance in the woodland soils. Similarly, we found that the number of OTUs belonging to Agaricostilbomycetes was significantly higher in the grassland samples than the woodland samples. Interestingly, a previous study showed a positive relationship between the abundance of OTUs of this group and plant community richness in grassland (LeBlanc, Kinkel, & Kistler, 2014). Overall, microbial diversity and composition were significantly influenced by the habitat edge as revealed across this ecotone.

4.2. Potential nitrification driven by bacterial ammonia oxidizers

Bacterial rather than archaeal ammonia oxidizers drove potential nitrification in the N‐rich soils at this site, and this pattern was consistently shown by multiple analytical approaches employed in this study. For example, ordinary kriging revealed that both bacterial amoA and PNR had visually similar spatial patterns and operated at similar spatial ranges. Consequently, these groups were also positively correlated (p < 0.001) across the ecotone. Network analysis further indicated that the PNR subnetwork was dominated by unclassified members of Betaproteobacteria and not archaea. It should be noted that soils across this ecotone were relatively N rich with average %N, NH4‐N and DON of 0.287% (w/w), 12.1 μg and 78.8 μg per gram of soil, respectively (Banerjee, Bora et al., 2016). While archaeal ammonia oxidizers are important for nitrification and dominant in many ecosystems (Leininger et al., 2006), they are well‐acknowledged for their oligotrophic nature (Erguder, Boon, Wittebolle, Marzorati, & Verstraete, 2009; Hatzenpichler, 2012). On the other hand, bacterial ammonia oxidizers are typically copiotrophic which makes them particularly suited for more nutrient rich soils. Interestingly, archaeal amoA was more abundant than bacterial amoA in both woodland and grassland soils in our study, but despite this, bacterial ammonia oxidizers displayed a stronger correlation with potential nitrification. Previous studies have similarly found that archaeal ammonia oxidizers are less responsive to nitrification in N‐rich soils even when they are relatively more abundant than their bacterial counterparts (Di et al., 2009). The higher responsiveness of ammonia oxidizing bacteria in N‐rich soils was also noted in a recent global meta‐analysis (Carey et al., 2016). The different spatial ranges of bacterial and archaeal ammonia oxidizers in our study indicate a possible niche differentiation of these communities as previously suggested (Prosser & Nicol, 2012). While the majority of Betaproteobacteria nodes in the PNR subnetwork were unclassified members, the association between the PNR subnetwork and Betaproteobacteria members reinforces the importance of this bacterial group for nitrification in N‐rich soils.

4.3. Microbial co‐occurrences across ecotone

We found a similarity between microbial co‐occurrence and spatial patterns. For example, microbial nodes in the woodland, grassland, and transition zone were structured into separate clusters with the woodland habitat having a significantly higher number of nodes. Similarly, kriging showed a significantly higher abundance of all microbial groups in the woodland samples. Importantly, our results illustrate how network complexity, indicated by the number of nodes and edges, changes between two adjoining ecological systems within one landscape and how archaeal, fungal, and bacterial patterns of co‐occurrence are influenced in the transition zone. Previous studies using network analysis have often only assessed bacterial communities and not fungal or archaeal communities (Banerjee, Baah‐Acheamfour et al., 2016; Barberán et al., 2012; Shi et al., 2016; Vick‐Majors et al., 2014). Thus, the roles of these latter groups have been underrepresented in microbial network analyses (Ma et al., 2016; Steele et al., 2011). While the networks were dominated by bacterial nodes, fungal and archaeal nodes were also abundant. Our results highlight the importance of assessing interkingdom associations in soil microbial networks.

4.4. Relationships between microbial co‐occurrences and ecological processes

Linking microbial community composition to function is a central goal in ecology (Graham et al., 2016; Prosser et al., 2007). In this study, soil P and C:N formed large clusters dominated by bacterial nodes and these clusters were also connected with other C and N properties. Similarly, the processes of C, N, and P cycling were also correlated with microbial co‐occurrence. Interestingly, soil P and phosphatase activity formed the largest clusters in their respective networks whereas two lignolytic enzymes (phenol oxidase and peroxidase) formed joined clusters. Extracellular enzymes are involved in the decomposition and mineralization of soil organic matter, which is a “broad” process that involves many steps and operated by functionally and taxonomically diverse generalist microbial groups (Fierer et al., 2007; Schimel & Schaeffer, 2012). On the other hand, ammonia oxidation is a “narrow” process facilitated by specialist groups of bacteria and archaea (Kowalchuk & Stephen, 2001). The distinct cluster of PNR was mainly dominated by Betaproteobacteria, reinforcing the observation that nitrification at this site was driven by ammonia oxidizing bacteria.

4.5. Keystone taxa and determinants

A useful feature of network analysis is that it can identify “hubs” or keystone taxa that have significant influence on the structure and functioning of microbiomes (Newman, 2003). Identifying keystone taxa and the factors that drive their abundance and spatiotemporal distribution is of particular importance in microbial ecology. The concept of keystone taxa was originally proposed some decades ago by ecologist Paine (1966). Keystone taxa have been identified in microbial communities both statistically (Banerjee, Kirkby et al., 2016; Hartman et al., 2018; Lupatini et al., 2014; Shi et al., 2016) and empirically (Curtis et al., 2014; Fisher & Mehta, 2014; Hajishengallis, Darveau, & Curtis, 2012). Berry and Widder (2014) used network‐based scores such as high mean degree, high closeness centrality, and low betweenness centrality to identify keystone taxa with 85% accuracy. Using the method proposed by Berry and Widder (2014), we identified six bacterial and four fungal OTUs in this study as representing the top ten keystone taxa. In a recent study, we also reported that bacterial and fungal keystones were significantly correlated to organic matter decomposition in an agricultural soil (Banerjee, Kirkby et al., 2016). Soil carbon and nitrogen contents likewise emerged as the drivers of keystone taxa that we identified here across the woodland‐grassland ecotone. One limitation of this study is that high‐throughput sequencing of microbial communities was performed on 18 soil samples. These samples were randomly selected equally from the three zones with six samples at each habitat component. Thus, careful consideration was made to obtain representative samples across this woodland‐grassland ecotone. Moreover, the selection of a single ecotone site in this study was based our previous observation that spatial patterns of a range of soil properties and extracellular enzyme activities were similar across two native woodland‐grassland ecotones (including this one) located approximately 150 km apart (Banerjee, Bora et al., 2016).

5. CONCLUSION

Using geostatistics, quantitative PCR, high‐throughput sequencing and network analysis, we demonstrated spatial patterns and co‐occurrences of archaeal, bacterial, and fungal communities across a woodland‐grassland ecotone. The abundance, structure, and taxonomic composition of soil microbial communities were significantly different in the transition zone than the woodland and grassland. Microbial networks predominantly comprised positive interactions that reflected the high C, N, and P levels at this site. Microbial co‐occurrences showed clusters based on habitats, soil properties, and ecological processes. Although microbial networks were dominated by bacterial OTUs, fungal and archaeal members were also abundant, highlighting the importance of interkingdom associations in soil microbial networks. Nitrification was driven by ammonia‐oxidizing bacteria, and this was supported by the dominance of Betaproteobacteria OTUs in the PNR subnetwork. A coherence of spatial patterns and co‐occurrences of microbial communities was thus demonstrated across the ecotone.

CONFLICT OF INTEREST

None declared.

AUTHOR CONTRIBUTIONS

S.B. designed the study and performed analyses. S.B., A.E.R., and P.H.T. conducted the soil sampling and made substantial contributions to the writing. A.B. assisted in data analyses, and M.v.d.H. contributed to the writing of the article.

DATA ACCESSIBILITY

The raw sequence data of bacterial 16S rRNA, archaeal 16S rRNA, and fungal ITS are available under the NCBI BioProject Accession number PRJNA427915 (SRA accession SRP 131862).

Supporting information

 

Click here for additional data file. (7.6MB, png)

 

Click here for additional data file. (866.4KB, png)

 

Click here for additional data file. (145.2KB, jpeg)

 

Click here for additional data file. (1.1MB, png)

 

Click here for additional data file. (29.6KB, docx)

ACKNOWLEDGMENTS

We thank Sudipta Sankar Bora and Shamshul Hoque for their assistance with field work.

Banerjee S, Thrall PH, Bissett A, van der Heijden MGA, Richardson AE. Linking microbial co‐occurrences to soil ecological processes across a woodland‐grassland ecotone. Ecol Evol. 2018;8:8217–8230. 10.1002/ece3.4346

Contributor Information

Samiran Banerjee, Email: banerjee.soilmicro@gmail.com.

Alan E. Richardson, Email: Alan.Richardson@csiro.au

REFERENCES

  1. Allison, S. D. , & Vitousek, P. M. (2005). Responses of extracellular enzymes to simple and complex nutrient inputs. Soil Biology and Biochemistry, 37(5), 937–944. 10.1016/j.soilbio.2004.09.014 [DOI] [Google Scholar]
  2. Alves, R. J. E. , Wanek, W. , Zappe, A. , Richter, A. , Svenning, M. M. , Schleper, C. , & Urich, T. (2013). Nitrification rates in Arctic soils are associated with functionally distinct populations of ammonia‐oxidizing archaea. The ISME Journal, 7(8), 1620–1631. 10.1038/ismej.2013.35 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Anadón, J. D. , Sala, O. E. , & Maestre, F. T. (2014). Climate change will increase savannas at the expense of forests and treeless vegetation in tropical and subtropical Americas. Journal of Ecology, 102(6), 1363–1373. 10.1111/1365-2745.12325 [DOI] [Google Scholar]
  4. Andrews, S. (2010). FastQC a quality control tool for high throughput sequence data. Retrieved from https://www.bioinformatics.babraham.ac.uk/projects/fastqc/
  5. Archer, E. (2016). Package ‘rfPermute’. Retrieved from https://cran.r-project.org/web/packages/rfPermute/index.html
  6. Archer, S. R. , & Predick, K. I. (2014). An ecosystem services perspective on brush management: Research priorities for competing land‐use objectives. Journal of Ecology, 102(6), 1394–1407. 10.1111/1365-2745.12314 [DOI] [Google Scholar]
  7. Banerjee, S. , Baah‐Acheamfour, M. , Carlyle, C. N. , Bissett, A. , Richardson, A. E. , Siddique, T. , … Chang, S. X. (2016). Determinants of bacterial communities in Canadian agroforestry systems. Environmental Microbiology, 18(6), 1805–1816. 10.1111/1462-2920.12986 [DOI] [PubMed] [Google Scholar]
  8. Banerjee, S. , Bora, S. , Thrall, P. H. , & Richardson, A. E. (2016). Soil C and N as causal factors of spatial variation in extracellular enzyme activity across grassland‐woodland ecotones. Applied Soil Ecology, 105, 1–8. 10.1016/j.apsoil.2016.04.002 [DOI] [Google Scholar]
  9. Banerjee, S. , Kirkby, C. A. , Schmutter, D. , Bissett, A. , Kirkegaard, J. A. , & Richardson, A. E. (2016). Network analysis reveals functional redundancy and keystone taxa amongst bacterial and fungal communities during organic matter decomposition in an arable soil. Soil Biology and Biochemistry, 97, 188–198. 10.1016/j.soilbio.2016.03.017 [DOI] [Google Scholar]
  10. Banerjee, S. , Schlaeppi, K. , & van der Heijden, M. G. A. (2018). Keystone taxa as drivers of microbiome structure and functioning. Nature Reviews Microbiology, Ahead of print. 10.1038/s41579-018-0024-1 [DOI] [PubMed] [Google Scholar]
  11. Barberán, A. , Bates, S. T. , Casamayor, E. O. , & Fierer, N. (2012). Using network analysis to explore co‐occurrence patterns in soil microbial communities. The ISME Journal, 6(10), 343–351. 10.1038/ismej.2011.119 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Bardgett, R. D. , & van der Putten, W. H. (2014). Belowground biodiversity and ecosystem functioning. Nature, 515(7528), 505–511. 10.1038/nature13855 [DOI] [PubMed] [Google Scholar]
  13. Bengtsson‐Palme, J. , Ryberg, M. , Hartmann, M. , Branco, S. , Wang, Z. , Godhe, A. , … Nilsson, R. H. (2013). Improved software detection and extraction of ITS1 and ITS2 from ribosomal ITS sequences of fungi and other eukaryotes for analysis of environmental sequencing data. Methods in Ecology and Evolution, 4(10), 914–919. 10.1111/2041-210X.12073 [DOI] [Google Scholar]
  14. Benjamini, Y. , & Hochberg, Y. (1995). Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society B, 57(1), 289–300. 10.2307/2346101 [DOI] [Google Scholar]
  15. Berry, D. , & Widder, S. (2014). Deciphering microbial interactions and detecting keystone species with co‐occurrence networks. Frontiers in Microbiology, 5, 1–14. 10.3389/fmicb.2014.00219 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Bissett, A. , Brown, M. V. , Siciliano, S. D. , & Thrall, P. H. (2013). Microbial community responses to anthropogenically induced environmental change: Towards a systems approach. Ecology Letters, 16(Suppl. 1), 128–139. 10.1111/ele.12109 [DOI] [PubMed] [Google Scholar]
  17. Bissett, A. , Fitzgerald, A. , Meintjes, T. , Mele, P. M. , Reith, F. , Dennis, P. G. , … Young, A. (2016). Introducing BASE: The Biomes of Australian Soil Environments soil microbial diversity database. GigaScience, 5(1), 21 10.1186/s13742-016-0126-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Bradford, J. B. , Schlaepfer, D. R. , Lauenroth, W. K. , & Burke, I. C. (2014). Shifts in plant functional types have time‐dependent and regionally variable impacts on dryland ecosystem water balance. Journal of Ecology, 102(6), 1408–1418. 10.1111/1365-2745.12289 [DOI] [Google Scholar]
  19. Breiman, L. (2001). Randomforest 2001. Machine Learning, 45, 5–32. 10.1017/CBO9781107415324.004 [DOI] [Google Scholar]
  20. Burns, R. G. (1982). Enzyme activity in soil: Location and a possible role in microbial ecology. Soil Biology and Biochemistry, 14(5), 423–427. 10.1016/0038-0717(82)90099-2 [DOI] [Google Scholar]
  21. Cardona, C. , Weisenhorn, P. , Henry, C. , & Gilbert, J. A. (2016). Network‐based metabolic analysis and microbial community modeling. Current Opinion in Microbiology, 31, 124–131. 10.1016/j.mib.2016.03.008 [DOI] [PubMed] [Google Scholar]
  22. Carey, C. J. , Dove, N. C. , Beman, J. M. , Hart, S. C. , & Aronson, E. L. (2016). Meta‐analysis reveals ammonia‐oxidizing bacteria respond more strongly to nitrogen addition than ammonia‐oxidizing archaea. Soil Biology and Biochemistry, 99, 158–166. 10.1016/j.soilbio.2016.05.014 [DOI] [Google Scholar]
  23. Curtis, M. M. , Hu, Z. , Klimko, C. , Narayanan, S. , Deberardinis, R. , & Sperandio, V. (2014). The gut commensal bacteroides thetaiotaomicron exacerbates enteric infection through modification of the metabolic landscape. Cell Host & Microbe, 16(6), 759–769. 10.1016/j.chom.2014.11.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. De Menezes, A. B. , Prendergast‐Miller, M. T. , Richardson, A. E. , Toscas, P. , Farrell, M. , Macdonald, L. M. , … Thrall, P. H. (2015). Network analysis reveals that bacteria and fungi form modules that correlate independently with soil parameters. Environmental Microbiology, 17(8), 2677–2689. 10.1111/1462-2920.12559 [DOI] [PubMed] [Google Scholar]
  25. Delgado‐Baquerizo, M. , Maestre, F. T. , Escolar, C. , Gallardo, A. , Ochoa, V. , Gozalo, B. , & Prado‐Comesaña, A. (2014). Direct and indirect impacts of climate change on microbial and biocrust communities alter the resistance of the N cycle in a semiarid grassland. Journal of Ecology, 102(6), 1592–1605. 10.1111/1365-2745.12303 [DOI] [Google Scholar]
  26. DeLong, E. F. (1992). Archaea in coastal marine environments. Proceedings of the National Academy of Sciences of the United States of America, 89(12), 5685–5689. 10.1073/pnas.89.12.5685 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Di, H. J. , Cameron, K. C. , Shen, J. P. , Winefield, C. S. , O'Callaghan, M. , Bowatte, S. , & He, J. Z. (2009). Nitrification driven by bacteria and not archaea in nitrogen‐rich grassland soils. Nature Geoscience, 2(9), 621–624. https://doi.org/Doi10.1038/Ngeo613 [Google Scholar]
  28. Edgar, R. C. (2010). Search and clustering orders of magnitude faster than BLAST. Bioinformatics, 26(19), 2460–2461. 10.1093/bioinformatics/btq461 [DOI] [PubMed] [Google Scholar]
  29. Edgar, R. C. (2013). UPARSE: Highly accurate OTU sequences from microbial amplicon reads. Nature Methods, 10(10), 996–998. 10.1038/nmeth.2604 [DOI] [PubMed] [Google Scholar]
  30. Edwards, I. P. , & Zak, D. R. (2011). Fungal community composition and function after long‐term exposure of northern forests to elevated atmospheric CO2 and tropospheric O3. Global Change Biology, 17(6), 2184–2195. 10.1111/j.1365-2486.2010.02376.x [DOI] [Google Scholar]
  31. Eldridge, D. J. , Bowker, M. A. , Maestre, F. T. , Roger, E. , Reynolds, J. F. , & Whitford, W. G. (2011). Impacts of shrub encroachment on ecosystem structure and functioning: Towards a global synthesis. Ecology Letters, 14(7), 709–722. 10.1111/j.1461-0248.2011.01630.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Erguder, T. H. , Boon, N. , Wittebolle, L. , Marzorati, M. , & Verstraete, W. (2009). Environmental factors shaping the ecological niches of ammonia‐oxidizing archaea. FEMS Microbiology Reviews, 33(5), 855–869. 10.1111/j.1574-6976.2009.00179.x [DOI] [PubMed] [Google Scholar]
  33. Faust, K. , & Raes, J. (2012). Microbial interactions: from networks to models. Nature Reviews Microbiology, 10(8), 538–550. 10.1038/nrmicro2832 [DOI] [PubMed] [Google Scholar]
  34. Fierer, N. , Bradford, M. A. , & Jackson, R. B. (2007). Toward an ecological classification of soil bacteria. Ecology, 88(6), 1354–1364. 10.1890/05-1839 [DOI] [PubMed] [Google Scholar]
  35. Fisher, C. K. , & Mehta, P. (2014). Identifying keystone species in the human gut microbiome from metagenomic timeseries using sparse linear regression. PLoS ONE, 9(7), 1–10. 10.1371/journal.pone.0102451 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Fuhrman, J. A. (2009). Microbial community structure and its functional implications. Nature, 459(7244), 193–199. 10.1038/nature08058 [DOI] [PubMed] [Google Scholar]
  37. Gardes, M. , & Bruns, T. D. (1993). ITS primers with enhanced specificity for basidiomycetes, application to the identification of mycorrihiza and rusts. Molecular Ecology, 2, 113–118. 10.1111/J.1365-294x.1993.Tb00005.X [DOI] [PubMed] [Google Scholar]
  38. Gleeson, D. B. , Müller, C. , Banerjee, S. , Ma, W. , Siciliano, S. D. , & Murphy, D. V. (2010). Response of ammonia oxidizing archaea and bacteria to changing water filled pore space. Soil Biology and Biochemistry, 42(10), 1888–1891. 10.1016/j.soilbio.2010.06.020 [DOI] [Google Scholar]
  39. Goovaerts, P. (1998). Geostatistical tools for characterizing the spatial variability of microbiological and physico‐chemical soil properties. Biology and Fertility of Soils, 27(4), 315–334. 10.1007/s003740050439 [DOI] [Google Scholar]
  40. Gosz, J. R. (1993). Ecotone hierarchies. Ecological Applications, 3(3), 369–376. [DOI] [PubMed] [Google Scholar]
  41. Graham, E. B. , Crump, A. R. , Resch, C. T. , Fansler, S. , Arntzen, E. , Kennedy, W. , … Stegen, J. C. (2017). Deterministic influences exceed dispersal effects on hydrologically‐connected microbiomes. Environmental Microbiology, 19, 1552–1567. 10.1111/1462-2920.13720 [DOI] [PubMed] [Google Scholar]
  42. Graham, E. B. , Knelman, J. E. , Schindlbacher, A. , Siciliano, S. , Breulmann, M. , Yannarell, A. , … Nemergut, D. R. (2016). Microbes as engines of ecosystem function: When does community structure enhance predictions of ecosystem processes? Frontiers in Microbiology, 7, 1–10. 10.3389/fmicb.2016.00214 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Hajishengallis, G. , Darveau, R. P. , & Curtis, M. A. (2012). The keystone‐pathogen hypothesis. Nature Reviews Microbiology, 10(10), 717–725. 10.1038/nrmicro2873 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Hart, S. C. , Stark, J. M. , Davidson, E. A. , & Firestone, M. K. (1994). Nitrogen mineralization, immobilization, and nitrification In Page A. L., Miller R. H., Keeney D. R. (Eds.), Methods of soil analysis: Part 2—microbiological and biochemical properties (pp. 985–1018). Madison, Wisconsin, USA: Soil Science Society of America. [Google Scholar]
  45. Hartman, K. , van der Heijden, M. G. A. , Wittwer, R. A. , Banerjee, S. , Walser, J.‐C. , & Schlaeppi, K. (2018). Cropping practices manipulate abundance patterns of root and soil microbiome members paving the way to smart farming. Microbiome, 6(1), 14 10.1186/s40168-017-0389-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Hatzenpichler, R. (2012). Diversity, physiology, and niche differentiation of ammonia‐oxidizing archaea. Applied and Environmental Microbiology, 78(21), 7501–7510. 10.1128/AEM.01960-12 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Isaaks, E. H. , & Srivastava, R. M. (1989). An introduction to applied geostatistics. New York, NY: Oxford University Press. [Google Scholar]
  48. Isbell, R. F. (2002). The Australian soil classification. Collingwood, Vic, Australia: Csiro Publishing. [Google Scholar]
  49. Jia, Z. , & Conrad, R. (2009). Bacteria rather than Archaea dominate microbial ammonia oxidation in an agricultural soil. Environmental Microbiology, 11(7), 1658–1671. 10.1111/j.1462-2920.2009.01891.x [DOI] [PubMed] [Google Scholar]
  50. Jones, R. T. , Robeson, M. S. , Lauber, C. L. , Hamady, M. , Knight, R. , & Fierer, N. (2009). A comprehensive survey of soil acidobacterial diversity using pyrosequencing and clone library analyses. The ISME Journal, 3(4), 442–453. 10.1038/ismej.2008.127 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Journel, A. G. , & Huijbregts, C. J. (1978). Mining geostatistics. Caldwell, NJ: Blackburn Press. [Google Scholar]
  52. Kemnitz, D. , Kolb, S. , & Conrad, R. (2007). High abundance of Crenarchaeota in a temperate acidic forest soil. FEMS Microbiology Ecology, 60(3), 442–448. 10.1111/j.1574-6941.2007.00310.x [DOI] [PubMed] [Google Scholar]
  53. Killham, K. (1990). Nitrification in coniferous forest soils. Plant and Soil, 128(1), 31–44. 10.1007/BF00009394 [DOI] [Google Scholar]
  54. Kõljalg, U. , Larsson, K.‐H. , Abarenkov, K. , Nilsson, R. H. , Alexander, I. J. , Eberhardt, U. , … Ursing, B. M. (2005). UNITE: A database providing web‐based methods for the molecular identification of ectomycorrhizal fungi. The New Phytologist, 166(3), 1063–1068. 10.1111/j.1469-8137.2005.01376.x [DOI] [PubMed] [Google Scholar]
  55. Kowalchuk, G. A. , & Stephen, J. R. (2001). Ammonia‐oxidizing bacteria: A model for molecular microbial ecology. Annual Reviews in Microbiology, 55(1), 485–529. 10.1146/annurev.micro.55.1.485 [DOI] [PubMed] [Google Scholar]
  56. Lacasella, F. , Gratton, C. , & De Felici, S. (2015). Asymmetrical responses of forest and ‘“beyond edge”’ arthropod communities across a forest–grassland ecotone. Biodiversity and Conservation, 24, 447–465. 10.1007/s10531-014-0825-0 [DOI] [Google Scholar]
  57. Lane, D. J. , Pace, B. , Olsen, G. J. , Stahl, D. A. , Sogin, M. L. , & Pace, N. R. (1985). Rapid determination of 16S ribosomal RNA sequences for phylogenetic analyses. Proceedings of the National Academy of Sciences of the United States of America, 82(20), 6955–6959. 10.1073/pnas.82.20.6955 [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Lauber, C. L. , Strickland, M. S. , Bradford, M. A. , & Fierer, N. (2008). The influence of soil properties on the structure of bacterial and fungal communities across land‐use types. Soil Biology and Biochemistry, 40(9), 2407–2415. 10.1016/j.soilbio.2008.05.021 [DOI] [Google Scholar]
  59. LeBlanc, N. , Kinkel, L. L. , & Kistler, H. C. (2014). Soil fungal communities respond to grassland plant community richness and soil edaphics. Microbial Ecology, 70(1), 188–195. 10.1007/s00248-014-0531-1 [DOI] [PubMed] [Google Scholar]
  60. Leininger, S. , Urich, T. , Schloter, M. , Schwark, L. , Qi, J. , Nicol, G. W. , … Schleper, C. (2006). Archaea predominate among ammonia‐oxidizing prokaryotes in soils. Nature, 442, 806–809. 10.1038/nature04983 [DOI] [PubMed] [Google Scholar]
  61. Li, F. , Chen, L. , Zhang, J. , Yin, J. , & Huang, S. (2017). Bacterial community structure after long‐term organic and inorganic fertilization reveals important associations between soil nutrients and specific taxa involved in nutrient transformations. Frontiers in Microbiology, 8, 187 10.3389/fmicb.2017.00187 [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Lupatini, M. , Suleiman, A. K. A. , Jacques, R. J. S. , Antoniolli, Z. I. , de Siqueira Ferreira, A. , Kuramae, E. E. , & Roesch, L. F. W. (2014). Network topology reveals high connectance levels and few key microbial genera within soils. Frontiers in Environmental Science, 2, 1–11. 10.3389/fenvs.2014.00010 [DOI] [Google Scholar]
  63. Ma, B. , Wang, H. , Dsouza, M. , Lou, J. , He, Y. , Dai, Z. , … Gilbert, J. A. (2016). Geographic patterns of co‐occurrence network topological features for soil microbiota at continental scale in eastern China. The ISME Journal, 10(8), 1–11. 10.1038/ismej.2015.261 [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. Magoč, T. , & Salzberg, S. L. (2011). FLASH: Fast length adjustment of short reads to improve genome assemblies. Bioinformatics, 27(21), 2957–2963. 10.1093/bioinformatics/btr507 [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Magura, T. (2017). Edge responses are different in edges under natural versus anthropogenic influence: A meta‐analysis using ground beetles. Ecology and Evolution, 7, 1009–1017. 10.1002/ece3.2722 [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Malmivaara‐Lämsä, M. , Hamberg, L. , Haapamäki, E. , Liski, J. , Kotze, D. J. , Lehvävirta, S. , & Fritze, H. (2008). Edge effects and trampling in boreal urban forest fragments ‐ impacts on the soil microbial community. Soil Biology and Biochemistry, 40(7), 1612–1621. 10.1016/j.soilbio.2008.01.013 [DOI] [Google Scholar]
  67. Murcia, C. (1995). Edge effects in fragmented forests: Implications for conservation. Trends in Ecology and Evolution, 10, 58–62. [DOI] [PubMed] [Google Scholar]
  68. Muyzer, G. , de Wall, E. , & Uitterlinden, A. (1993). Profiling of complex microbial populations by denaturing gradient gel electrophoresis analysis of polymerase chain reaction‐amplified genes coding for 16S rRNA. Applied and Environmental Microbiology, 59(3), 695–700. https://doi.org/0099-2240/93/030695-06$02.00/0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  69. Newman, M. E. J. (2003). The structure and function of complex networks. Society for Industrial and Applied Mathematics, 45(2), 167–256. Retrieved from http://arxiv.org/abs/cond-mat/0303516 [Google Scholar]
  70. Ochsenreiter, T. , Selezi, D. , Quaiser, A. , Bonch‐Osmolovskaya, L. , & Schleper, C. (2003). Diversity and abundance of Crenarchaeota in terrestrial habitats studied by 16S RNA surveys and real time PCR. Environmental Microbiology, 5(9), 787–797. 10.1046/j.1462-2920.2003.00476.x [DOI] [PubMed] [Google Scholar]
  71. Oksanen, J. , Blanchet, F. G. , Friendly, M. , Kindt, R. , Legendre, P. , McGlinn, D. , … Wagner, H. (2017). vegan: Community Ecology Package. Retrieved from https://cran.r-project.org/web/packages/vegan/index.html
  72. Paine, R. T. (1966). Food web complexity and species diversity. The American Naturalist, 100(910), 65–75. 10.1086/282400 [DOI] [Google Scholar]
  73. Power, M. E. , Tilman, D. , Estes, J. A. , Menge, B. A. , Bond, W. J. , Mills, L. S. , … Paine, R. T. (1996). Challenges in the quest for keystones. BioScience, 46(8), 609–620. 10.2307/1312990 [DOI] [Google Scholar]
  74. Prosser, J. I. , Bohannan, B. J. M. , Curtis, T. P. , Ellis, R. J. , Firestone, M. K. , Freckleton, R. P. , … Young, J. P. W. (2007). The role of ecological theory in microbial ecology. Nature Reviews. Microbiology, 5(5), 384–392. 10.1038/nrmicro1643 [DOI] [PubMed] [Google Scholar]
  75. Prosser, J. I. , & Nicol, G. W. (2012). Archaeal and bacterial ammonia‐oxidisers in soil: The quest for niche specialisation and differentiation. Trends in Microbiology, 20(11), 523–531. 10.1016/j.tim.2012.08.001 [DOI] [PubMed] [Google Scholar]
  76. R Development Core Team (2016). R Development Core Team, 2016. Retrieved February 20, 2018, from https://www.r-project.org/
  77. Ratajczak, Z. , Nippert, J. B. , Briggs, J. M. , & Blair, J. M. (2014). Fire dynamics distinguish grasslands, shrublands and woodlands as alternative attractors in the central great plains of North America. Journal of Ecology, 102(6), 1374–1385. 10.1111/1365-2745.12311 [DOI] [Google Scholar]
  78. Reshef, D. N. , Reshef, Y. A. , Finucane, H. K. , Grossman, S. R. , McVean, G. , Turnbaugh, P. J. , … Sabeti, P. C. (2011). Detecting novel associations in large data sets. Science, 334(6062), 1518–1524. 10.1126/science.1205438 [DOI] [PMC free article] [PubMed] [Google Scholar]
  79. Richardson, A. E. , Barea, J. M. , McNeill, A. M. , & Prigent‐Combaret, C. (2009). Acquisition of phosphorus and nitrogen in the rhizosphere and plant growth promotion by microorganisms. Plant and Soil, 321(1–2), 305–339. 10.1007/s11104-009-9895-2 [DOI] [Google Scholar]
  80. Rotthauwe, J.‐H. , Witzel, K.‐P. , & 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(12), 4704–4712. 10.1128/AEM.NA [DOI] [PMC free article] [PubMed] [Google Scholar]
  81. Saiya‐Cork, K. , Sinsabaugh, R. , & Zak, D. (2002). The effects of long term nitrogen deposition on extracellular enzyme activity in an Acer saccharum forest soil. Soil Biology and Biochemistry, 34(9), 1309–1315. 10.1016/S0038-0717(02)00074-3 [DOI] [Google Scholar]
  82. Sala, O. E. , & Maestre, F. T. (2014). Grass‐woodland transitions: Determinants and consequences for ecosystem functioning and provisioning of services. Journal of Ecology, 102(6), 1357–1362. 10.1111/1365-2745.12326 [DOI] [Google Scholar]
  83. Schimel, J. P. , & Schaeffer, S. M. (2012). Microbial control over carbon cycling in soil. Frontiers in Microbiology, 3, 1–11. 10.3389/fmicb.2012.00348 [DOI] [PMC free article] [PubMed] [Google Scholar]
  84. Schloss, P. D. , Westcott, S. L. , Ryabin, T. , Hall, J. R. , Hartmann, M. , Hollister, E. B. , … Weber, C. F. (2009). Introducing mothur: Open‐source, platform‐independent, community‐supported software for describing and comparing microbial communities. Applied and Environmental Microbiology, 75(23), 7537–7541. 10.1128/AEM.01541-09 [DOI] [PMC free article] [PubMed] [Google Scholar]
  85. Schmidt, M. , Jochheim, H. , Kersebaum, K. , & Lischeid, G. (2017). Agricultural and Forest Meteorology Gradients of microclimate, carbon and nitrogen in transition zones of fragmented landscapes – a review. Agricultural and Forest Meteorology, 232, 659–671. 10.1016/j.agrformet.2016.10.022 [DOI] [Google Scholar]
  86. Shannon, P. , Markiel, A. , Ozier, O. , Baliga, N. S. , Wang, J. T. , Ramage, D. , … Ideker, T. (2003). Cytoscape: A software environment for integrated models of biomolecular interaction networks. Genome Research, 13, 2498–2504. 10.1101/gr.1239303.metabolite [DOI] [PMC free article] [PubMed] [Google Scholar]
  87. Shi, S. , Nuccio, E. E. , Shi, Z. J. , He, Z. , Zhou, J. , Firestone, M. K. , & Johnson, N. (2016). The interconnected rhizosphere: High network complexity dominates rhizosphere assemblages. Ecology Letters, 19(8), 926–936. 10.1111/ele.12630 [DOI] [PubMed] [Google Scholar]
  88. Sistla, S. A. , & Schimel, J. P. (2013). Seasonal patterns of microbial extracellular enzyme activities in an arctic tundra soil: Identifying direct and indirect effects of long‐term summer warming. Soil Biology and Biochemistry, 66, 119–129. 10.1016/j.soilbio.2013.07.003 [DOI] [Google Scholar]
  89. Steele, J. A. , Countway, P. D. , Xia, L. , Vigil, P. D. , Beman, J. M. , Kim, D. Y. , … Fuhrman, J. A. (2011). Marine bacterial, archaeal and protistan association networks reveal ecological linkages. The ISME Journal, 5(9), 1414–1425. 10.1038/ismej.2011.24 [DOI] [PMC free article] [PubMed] [Google Scholar]
  90. van der Heijden, M. G. A. , Bardgett, R. D. , & Van Straalen, N. M. (2008). The unseen majority: Soil microbes as drivers of plant diversity and productivity in terrestrial ecosystems. Ecology Letters, 11(3), 296–310. 10.1111/j.1461-0248.2007.01139.x [DOI] [PubMed] [Google Scholar]
  91. van der Heijden, M. G. A. , Martin, F. M. , Selosse, M. A. , & Sanders, I. R. (2015). Mycorrhizal ecology and evolution : The past, the present, and the future. New Phytologist, 205, 1406–1423. 10.1111/nph.13288 [DOI] [PubMed] [Google Scholar]
  92. Vick‐Majors, T. J. , Priscu, J. C. , & Amaral‐Zettler, L. A. (2014). Modular community structure suggests metabolic plasticity during the transition to polar night in ice‐covered Antarctic lakes. The ISME Journal, 8(4), 778–789. 10.1038/ismej.2013.190 [DOI] [PMC free article] [PubMed] [Google Scholar]
  93. Wang, Q. , Garrity, G. M. , Tiedje, J. M. , & Cole, J. R. (2007). Naive Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Applied and Environmental Microbiology, 73(16), 5261–5267. 10.1128/AEM.00062-07 [DOI] [PMC free article] [PubMed] [Google Scholar]
  94. weatherzone (n.d.). Retrieved from http://www.weatherzone.com.au/
  95. White, T. J. , Bruns, T. D. , Lee, S. B. , & Taylor, J. W. (1990). PCR protocols: A guide to methods and applications In Innis M. A., Gelfand D. H., & Sninsky J. J. (Eds.), PCR protocols: A guide to methods and applications (pp. 315–322). New York, NY: Academic Press. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

 

Click here for additional data file. (7.6MB, png)

 

Click here for additional data file. (866.4KB, png)

 

Click here for additional data file. (145.2KB, jpeg)

 

Click here for additional data file. (1.1MB, png)

 

Click here for additional data file. (29.6KB, docx)

Data Availability Statement

The raw sequence data of bacterial 16S rRNA, archaeal 16S rRNA, and fungal ITS are available under the NCBI BioProject Accession number PRJNA427915 (SRA accession SRP 131862).

Articles from Ecology and Evolution are provided here courtesy of Wiley

RESOURCES