Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Contrasting bacterial communities in two indigenous Chionochloa (Poaceae) grassland soils in New Zealand

Contrasting bacterial communities in two indigenous Chionochloa (Poaceae) grassland soils in New Zealand

  • Jocelyn C. Griffith, 
  • William G. Lee, 
  • David A. Orlovich, 
  • Tina C. Summerfield


The cultivation of grasslands can modify both bacterial community structure and impact on nutrient cycling as well as the productivity and diversity of plant communities. In this study, two pristine New Zealand grassland sites dominated by indigenous tall tussocks (Chionochloa pallens or C. teretifolia) were examined to investigate the extent and predictability of variation of the bacterial community. The contribution of free-living bacteria to biological nitrogen fixation is predicted to be ecologically significant in these soils; therefore, the diazotrophic community was also examined. The C. teretifolia site had N-poor and poorly-drained peaty soils, and the C. pallens had N-rich and well-drained fertile soils. These soils also differ in the proportion of organic carbon (C), Olsen phosphorus (P) and soil pH. The nutrient-rich soils showed increased relative abundances of some copiotrophic bacterial taxa (including members of the Proteobacteria, Bacteroidetes and Firmicutes phyla). Other copiotrophs, Actinobacteria and the oliogotrophic Acidobacteria showed increased relative abundance in nutrient-poor soils. Greater diversity based on 16S rRNA gene sequences and the Tax4Fun prediction of enhanced spore formation associated with nutrient-rich soils could indicate increased resilience of the bacterial community. The two sites had distinct diazotrophic communities with higher diversity in C. teretifolia soils that had less available nitrate and ammonium, potentially indicating increased resilience of the diazotroph community at this site. The C. teretifolia soils had more 16S rRNA gene and nifH copies per g soil than the nutrient rich site. However, the proportion of the bacterial community that was diazotrophic was similar in the two soils. We suggest that edaphic and vegetation factors are contributing to major differences in the composition and diversity of total bacterial and diazotrophic communities at these sites. We predict the differences in the communities at the two sites will result in different responses to environmental change.


Widespread agricultural development of grasslands has made them among the most threatened ecosystems worldwide [13]. The cultivation of grasslands can modify both bacterial community structure and nutrient cycling [46]. Changes in the bacterial community can arise because different bacterial taxa exhibit different sensitivities to soil conditions including pH [7], C:N ratio [8], moisture [9] and nutrient addition [10,11]. The diversity of soil microbial communities can impact on a range of ecosystem processes; for example, fertilizer addition, elevated temperature and/or CO2 have been linked to altered bacterial community composition and community level activity (reviewed in [12]) [13,14].

For a variety of soil types, differences in N availability, drainage and the proportion of organic C have previously been shown to alter both the total bacterial community [15] and the diazotrophic community [16]. For example, C-rich soils have been proposed to favour the growth of copiotrophic bacteria, whereas higher relative abundance of oligotrophs has been observed in soils with lower organic C [17]. In addition, relative abundance of Acidobacteria has been shown to be negatively dependent on pH, while abundances of Actinobacteria and Bacteroidetes were positively dependent on pH [18]. Consistencies in the response of grassland bacterial community composition to nutrient addition have been identified across a broad geographic range; however, local variation in bacterial community structure demonstrates the need for ecosystem and soil-type specific bacterial community analyses [3,17,19].

Approximately 3.6 million hectares (13.4%) of New Zealand’s land is classified as tussock grassland (as of September 2002) [20]. These areas have significant conservation value as they support indigenous biodiversity and provide ecosystem services such as water regulation and erosion control, as well as being culturally significant [20,21]. Indigenous legumes and nitrogen-fixing actinorhizal plants are rare in New Zealand tussock grasslands, and the contribution of free-living nitrogen-fixing bacteria to biological nitrogen fixation is predicted to be ecologically significant [22]. Global estimates of free-living nitrogen fixation in temperate grasslands vary between 0.1–21 kg ha−1 year−1 with the potential to exceed N inputs from symbiotic nitrogen fixation in environments where symbionts are scarce [23]. Based on the copy number of the nifH gene (encoding a subunit of the nitrogenase enzyme that catalyzes nitrogen fixation), the diazotroph community of New Zealand tussock grasslands appears to be sensitive to cultivation [6]. Similarly in grass-scrubland soils in Australia, higher nitrate concentration as a result of frequent livestock grazing was linked to a lower nifH gene abundance, supporting the hypothesis that nitrogen-fixation decreases when N is widely available [24]. Similarly, fertilization with both organic and inorganic N decreased nitrogen fixation rates in forest soils [2527]. However, it is difficult to assess the significance of potential changes in diazotrophs in response to land use without a better knowledge of the diazotrophs in different undisturbed grasslands.

The tall-tussock grasslands of New Zealand are dominated by long-lived perennial snow tussock grasses (bunchgrasses) of the largely endemic genus Chionochloa and the short tussock grasslands are generally comprised of Festuca and Poa [20]. The alpine zone (800–2000 m) in the Murchison Mountains, Fiordland (South Island, NZ) is dominated by Chionochloa [28,29]. Diversification within this group resulted in more than 10 species in the region segregated according to altitude, soil conditions and snow lie [30,31]. The edaphic segregation of species influences temperature cues, flowering intensity, costs of seed production and vegetative growth [32,33]. Different species of Chionochloa exhibit adaptation to different extremes in natural soil fertility, resulting in distributions that are delimited by soil nutrients [30,32]. Soil N-limitation is a key component differentiating nutrient-rich from nutrient-poor sites in this system [33]. It follows that nitrogen-fixing bacteria must play a significant role in providing biologically available N in the N-limited soils, therefore their diversity and abundance is expected to impact on Chionochloa resource allocation and flowering.

This study examined the diversity of the bacterial community in two indigenous grassland soils in the Takahe Valley, Murchison Mountains, at the edaphic extremes occupied by Chionochloa snow tussock species on mineral soil [34]. Chionochloa teretifolia occupies total N-poor and poorly-drained peaty soils, while C. pallens occupies N-rich and well-drained fertile soils. These soils also differ in the proportion of organic carbon (C), Olsen phosphorus (P) and soil pH [33]. High-throughput 16S rRNA gene sequencing enables high resolution taxonomic characterization of soil bacterial communities [3537]. A similar approach, sequencing part of the nifH gene and comparing these sequences to the extensive database of nifH gene sequences [38,39], enables study of nitrogen-fixing communities. Using this approach we undertook a comprehensive examination of the total bacterial and diazotroph community structure in soils from two undisturbed grassland sites. The aims of this study were to compare the bacterial communities, both total bacterial and diazotroph communities, in the two contrasting indigenous grassland soils and to predict differences in the soil microbial function between the two soil types using Tax4Fun and the 16S rRNA gene dataset.

Materials and methods

Site description, sampling and DNA isolation

Soil samples were collected from Takahe Valley, Murchison Mountains, Fiordland National Park, New Zealand (45°14′ S, 167°33′ E) in March 2013. Line transects of approximately 100 m, marked through both the C. pallens and C. teretifolia communities, were located approximately 1260 and 1150 m above sea level, respectively. This work was performed under the Department of Conservation permit 38477-FLO. Both species transects were divided into five equal-sized, non-overlapping plots. Within each plot, debris (including coarse roots, stones and loose litter) were removed from the soil surface and 1 cm of the surface soils were collected from five separate locations. From each of the two Chionochloa transects, five composite soil samples per species (each composite sample being the result of pooling five individual samples) were collected, and coarse roots and stones were removed. Soil samples were stored in air-tight vials in the dark and then transferred to −20°C. DNA was extracted from 200 mg soil from each composite sample using the MoBio PowerSoil DNA Isolation Kit (MoBio Laboratories) as per manufacturer’s instructions.

Library construction and sequencing

To determine the bacterial and diazotrophic community composition of the grassland soils, barcoded amplicon libraries were constructed and sequenced. The 16S rRNA (V3–4) region and nifH genes were amplified using modified primers 0341F/0785R [40] and 19F/407R [41], respectively. For the first amplification step from soil DNA samples, the primers were synthesized with overhang adapters complementary to the Illumina sequencing primers. A subsequent limited-cycle amplification step was then performed to attach 6-bp multiplexing indices and Illumina sequencing adapters (S1 Table).

PCR amplifications for the bacterial 16S rDNA library were carried out in 50 μl reaction volumes containing final concentrations of: 1× High Fidelity PCR buffer, 2.0 mM MgSO4, 200 μM dNTPs, 200 nM each of 0341F and 0785R primers, 0.05% of bovine serum albumin (BSA) and 1 U Platinum Taq DNA Polymerase High Fidelity (Thermo Fisher Scientific). The reactions were held at 94°C for 2 min, followed by 35 cycles of 94°C for 30 s, 55°C for 30 s, 68°C for 40 s and a final extension of 68°C for 2 min. Amplicons were purified using Agencourt AMPure XP beads (Beckman Coulter) and eluted in 30 μl of UltraPure DNAse/RNAse-free distilled water (Thermo Fisher Scientific) and quantified using the Qubit dsDNA High Sensitivity Assay (Thermo Fisher Scientific). Second round PCR was performed with 2 ng of purified amplicons as template using the same reaction concentrations as the first round. The PCR conditions were 95°C for 20 s, followed by 10 cycles of 95°C for 20 s, 55°C for 20 s, 68°C for 20 s and a final extension of 68°C for 40 s. Amplicons were again purified using Agencourt AMPure XP beads and eluted in 30 μl of nuclease-free distilled water.

PCR amplifications for the nifH library were performed using the universal primer set 19F/407R [41], which has been predicted to have a high coverage of nifH sequences based on in silico testing and also amplified from soil DNA [42]. The 50 μl reaction mixtures contained final concentrations of 1× High Fidelity PCR buffer, 2.0 mM MgSO4, 200 μM of each dNTP, 200 nM each of 19F and 407R primers, 0.05% of BSA and 1 U Platinum Taq DNA Polymerase High Fidelity. The reactions were held at 94°C for 2 min, followed by 30 cycles of 94°C for 30 s, 51°C for 30 s, 68°C for 40 s and a final extension of 68°C for 2 min. Amplicon purification, second round of PCR amplification, and the subsequent purification and elution steps were the same as the 16S rDNA library.

Amplicons were run on the Agilent 2100 Bioanalyzer (Agilent Technologies) for size assessment. A paired-end 250-bp sequencing run was performed on the Illumina MiSeq (New Zealand Genomics Limited). Sequences are deposited in the NCBI Sequence Read Archive (SRP053025) under accession numbers SRX2012063–SRX2012072 (16S rRNA) and SRX2014251–SRX2014260 (nifH).

Bioinformatic analyses

Samples were demultiplexed using bcl2fastq Conversion Software (Illumina). Sequence data were processed and quality controls were performed through a combination of the UPARSE, mothur and QIIME pipelines [4345]. Paired-end reads were joined to build the 16S rRNA and nifH gene sequences using the -fastq_mergepairs option of USEARCH v7.0.1090 [45]. Quality control processing was carried out via the UPARSE pipeline using -fastq_filter command, and sequences with more than 0.5 expected errors were removed (-fastq_maxee 0.5) [45]. Putative frameshifts in nifH sequences were detected and corrected with the FrameBot tool ( [46]. Further processing was performed on mothur v.1.34.0 [45] using the command screen.seqs to include only sequences that were between 400–500 bp and 300–400 bp for the 16S rRNA and nifH gene sequences, respectively. Reads were clustered into OTUs using the UPARSE pipeline based on 97% sequence identity. Chimeric sequences were identified and removed using the uchime_ref command [45] against the ChimeraSlayer “Gold” database from the Broad Microbiome Utilities version microbiomeutil-r20110519 [47] for 16S rRNA sequences and using the reference-free de novo chimera detection mode for nifH sequences [45]. Taxonomic assignment for the 16S rRNA sequence data were made in QIIME against the SILVA 119 database using the Ribosomal Database Project (RDP) Classifier [48]. An aligned nifH gene database [49] using the default assignment method in QIIME (UCLUST) was used for nifH gene taxonomic assignment. Sequences belonging to nifH clusters IV and V that code for genes involved in bacteriochlorophyll and chlorophyll biosynthesis [50] and thus are not related to nitrogen fixation were removed. Singleton OTUs, archaea and chloroplast sequences were removed from the 16S rRNA gene dataset. For the ten soil samples from C. teretifolia and C. pallens-dominated grasslands, a total of 4 893 446 and 2 024 357 reads were obtained through high-throughput sequencing of the 16S rRNA and nifH genes, respectively. Following the removal of singletons, low-quality sequences and nifH sequences belonging to clusters IV and V, 434 260 sequences remained for downstream analyses (S2 Table). This resulted in a total of 19 219 16S rDNA- and 547 nifH-based OTUs. Rarefaction analyses at 97% similarity clustering showed that both 16S rRNA and nifH gene libraries reached their plateau (S1 Fig). Subsamples of 198 769 and 9658 random 16S rRNA and nifH gene sequences, respectively, were analyzed from each sample to compare diversity and richness across sample sites. Tax4Fun [51] was used to predict functional profiles from 16S rRNA genes within the R environment [52].

To construct the nifH phylogenetic tree, nifH sequences were first clustered at 97% similarity OTUs containing > 100 sequences and their closest sequence matches were retrieved from GenBank using the NCBI BLASTX search. The nifH sequences were aligned with MUSCLE v3.8.31 under default parameters [45] and visually inspected in Geneious R9 [53]. A phylogenetic tree of the nucleotide sequences was constructed with Bayesian inference using MrBayes v3.2.5 [54,55] with four separate chains for 72.5 × 106 generations. Sampling frequency was set to every 100 generations and the first 25% of samples were discarded (“burnin”). The Bayesian analysis was run using a GTR + I + Γ (lset nst = 6, rates = invgamma) model of substitution as selected by MrModelTest 2.3 [56] and was run until the standard deviation of split frequencies was < 0.01. The tree was displayed on FigTree and heatmaps were generated with the heatmap.2 function of the ‘gplots’ package for R [52].

Phylogenetic trees of the class Ktedonobacteria (Chloroflexi) and the Nitrosomonadaceae family (Betaproteobacteria) were constructed using a similar method to the nifH tree with the exceptions described below. Closest sequences matches were retrieved form GenBank using the BLASTN search. The nucleotide alignments were visually inspected and gaps were excluded for phylogenetic analysis, and the Markov chains were run for a total of 96 and 2 million generations for Ktedonobacteria and Nitrosomonadaceae Bayesian analyses, respectively.

Real-time PCR assay

Quantitative real-time PCR (qPCR) was performed using a MiniOpticon Real-Time PCR System (Bio-Rad). To estimate total bacterial abundance in soils we targeted the 16S rRNA gene using modified Earth Microbiome Project recommended primers 515f and 806rB [57]. qPCR targeting the nifH gene was performed using the nifH-F (5′-AAAGGYGGWATCGGYAARTCCACCAC-3′) and nifH-R (5′- TTGTTSGCSGCRTACATSGCCATCAT-3′) as these primers have been predicted to have a lower affinity for non-nifH sequences from cluster IV [42,58]. Reaction mixtures of 10 μl were composed of 1 μl gDNA template (1 in 10 dilution, ranging from 1–8 ng), 5 μl of 2 × SensiFAST SYBR No-ROX mix (Bioline) and 400 nM of each primer. The cycling conditions for both genes was as follows: 95°C for 3 min and 40 cycles of 95°C for 5 s, 60°C for 10 s, 72°C for 15 s, and data acquisition at 80°C for 15 s. All qPCR assays were carried out in triplicate and included negative (no template) controls. Melt-profile analyses were performed to confirm the absence of primer dimer or non-specific amplification products. Standard curves were performed for each primer pair using 10-fold serial dilutions of linearized plasmids containing target 16S rRNA or nifH gene fragments as the insert. Linearized plasmids were quantified using Qubit dsDNA High Sensitivity Assay Kit and a Qubit 2.0 fluorometer. PCR efficiencies were calculated from standard curves, these were 86% and 91% for 16S rRNA and nifH genes, respectively.

Statistical analysis

One-way analysis of variance (ANOVA) was used to determine whether there was a significant effect of sample site on diversity or richness of 16S rRNA and nifH gene sequences. Alpha and beta diversity were calculated, and pairwise distances using weighted and unweighted UniFrac matrices [59] were used to generate Principal Coordinates Analysis (PCoA) plots [44]. Permutational multivariate analysis of variance (PERMANOVA) [60] tests were performed to test if beta diversity differed significantly among communities in different sites (permutations = 999). The heatmap was generated using the heatmap.2 function of the ‘gplots’ package [61] for R [52]. Only the top 34 most abundant 16S rRNA gene-based families were used to generate the heatmap. Clustering was accomplished using average linkage method in the hclust function.

Results and discussion

Alpha diversity and richness based on 16S rRNA and nifH gene sequences differ between the two sites

Both the 16S rRNA gene and nifH gene sequences showed more variable α-diversity between the two sites than between samples collected from within either site. For the 16S rRNA gene sequences the number of OTUs (6667 ± 453 [mean ± SD] versus 4852 ± 444; ANOVA, P < 0.001) and the Shannon diversity index (9.9 ± 0.2 versus 9.1 ± 0.3; ANOVA, P < 0.001) were higher in C. pallens soils compared to C. teretifolia soils (Fig 1a). Conversely, in the diazotrophic community the number of OTUs (115 ± 20 versus 205 ± 17; ANOVA, P < 0.001) and Shannon diversity index (4.9 ± 0.4 versus 5.9 ± 0.1; ANOVA, P < 0.001) were significantly lower in the C. pallens soils compared to C. teretifolia soils (Fig 1b). As expected the diversity and richness of the diazotroph community was lower than that of the total bacterial community. However, it was unexpected that the higher diversity and richness of the diazotrophic community was at the C. teretifolia site where there was lower diversity and richness of the bacterial community based on 16S rRNA gene sequences.

Fig 1. Boxplots of sample richness and diversity at a genetic distance of 3% in Chionochloa pallens and Chionochloa teretifolia grassland soils based on (a) 16S rRNA and (b) nifH gene sequences.

Number of unique OTUs and Shannon diversity indices were calculated at 97% similarity for 198 769 and 9658 subsampled 16S rRNA and nifH gene sequences, respectively.

Bacterial communities differ in C. pallens and C. teretifolia grassland soils

Principal coordinates analyses (PCoA) based on the unweighted and weighted UniFrac distances of the 16S rRNA gene sequences showed that the C. pallens samples were distinct from those of the C. teretifolia samples (Fig 2a and 2b). PERMANOVA showed that there was a significant difference in the bacterial communities at the two sites (Pseudo-F = 21.2323, P = 0.009 for weighted; Pseudo-F = 3.4357, P = 0.011 for unweighted analyses). Comparison of nifH sequences showed C. pallens soil samples were separated from the C. teretifolia soil samples in the unweighted UniFrac analysis, but the diazotrophic communities from the two sites were not so clearly separated using weighted UniFrac analysis (Fig 2c and 2d). This was confirmed using PERMANOVA (Pseudo-F = 3.9344, P = 0.01 for weighted; Pseudo-F = 5.912, P = 0.009 for unweighted analyses). For both 16S rRNA and nifH sequences, Bray-Curtis dissimilarities for intrasite samples were significantly lower than intersite samples (P < 0.0001 for 16S rRNA and nifH sequences, two-tailed, two-sample t test) (S3 Table).

Fig 2. Principal coordinates analysis plots illustrating beta diversity of total bacterial communities based on 16S rRNA gene sequences (a and b) and nitrogen-fixing community based on nifH gene sequences (c and d) at Chionochloa pallens (CP) and Chionochloa teretifolia (CT) sample sites.

UniFrac distance matrices are generated from OTUs of 97% identity. Unweighted plots (a and c) are based on the presence or absence of OTUs; weighted plots (b and d) are based on the relative abundance of OTUs. Red dots are samples from C. pallens sites and blue dots are from C. teretifolia sites.

Differences in abundance and diversity of bacterial phyla and nifH clusters in the two Chionochloa grassland soils

The dominant phyla in all samples were Acidobacteria, Actinobacteria, Bacteroidetes, Chloroflexi, Planctomycetes, Proteobacteria and Verrucomicrobia (Fig 3). These were previously reported to be the most abundant phyla in a study of 25 grassland soils worldwide [3]. Less abundant phyla (< 1% of all classified sequences; 20 phyla) are shown in S4 Table. In C. teretifolia sites compared to C. pallens sites there was higher ranking of Acidobacteria and Actinobacteria and lower ranking of Proteobacteria, Planctomycetes, Chloroflexi, Bacteroidetes and Verrucomicrobia (S4 Table). Several of the differences between the C. pallens and C. teretifolia sites were consistent with other findings that have reported an increase of copiotrophic (or K-selected) or faster-growing bacterial taxa under increased available ammonium and nitrate [3,15,17]. This included increased abundance of putative copiotrophic taxa, i.e., Alphaproteobacteria, Betaproteobacteria and Bacteroidetes in the C. pallens sites compared to C. teretifolia sites (S4 Table).

Fig 3. Taxonomic assignments of sequences from Chionochloa pallens (CP) and Chionochloa teretifolia (CT) grassland soils based on (a) 16S rRNA and (b) nifH gene sequences.

Only dominant bacterial phyla (over 1%) of 16S rRNA gene sequences are shown. The nifH clusters are shown as previously designated [62].

In addition, the increased relative abundance of Acidobacteria in the C. teretifolia sites was consistent with identification of this phylum as oligotrophic (or r-selected). Acidobacteria have increased abundance in N-poor soils; this has been suggested to be due to their ability to metabolize less fertile and recalcitrant organic C that is largely unavailable to other bacteria [17,19,63]. This distribution of phyla supports the model of a shift from oligotroph to copiotroph communities based on low to high resource availability [15,17]. However, Actinobacteria which have been identified as copiotrophic [3,17], exhibited higher relative abundance in C. teretifolia soils (13.8 ± 3.9%) compared to C. pallens soils (8.1 ± 1.0%) (S4 Table). Many other soil characteristics differ between the two sites, which may contribute to the differences in bacterial community. For example, greater relative abundance of Actinobacteria was found in C. teretifolia soils, which have a higher C:N ratio compared to C. pallens soils [33]. Actinobacterial abundance in soil was previously positively correlated with total microbial biomass and C:N ratio in grassland soils [35,64].

Increased diversity of 16S rRNA genes per genome within certain bacterial taxa that contain high copy numbers of 16S rRNA can overestimate diversity estimates using OTU clustering [65]. In particular, member of the phyla Firmicutes, Gammaproteobacteria and Bacteroidetes have been reported to have diverse multiple copies of 16S rRNA gene per genome [65] and their higher relative abundance in C. pallens soils might attribute to the higher diversity in these soils.

Differences in bacterial families with high relative abundance in the C. pallens and C. teretifolia soils

Among the most abundant families, Acidobacteriaceae Subgroup 1 (Acidobacteria) and three families belonging to Actinobacteria, i.e., Acidothermaceae, Mycobacteriaceae and YNPFFP1, were significantly more abundant in C. teretifolia grassland soils (P < 0.01 for all families; Fig 4). The mean abundance of Acidobacteriaceae Subgroup 1 in the low pH C. teretifolia soils (mean of pH 4.5 [33]) was more than double that in C. pallens soils (mean of pH 5.2; Fig 4). Previously the abundance of many acidobacterial subgroups, including subgroup 1, were known to correlate negatively with soil pH [66]. Within the Acidothermaceae family (Actinobacteria) and Actinobacteridae subclass, the genus Acidothermus (the most abundant genus across all soil samples) was more abundant in C. teretifolia soils (8.4 ± 3.4%) compared to C. pallens soils (1.2 ± 0.3%) (ANOVA, P = 0.003). However, the opposite pattern was observed in 88 soils across North and South America, where the relative abundance of Actinobacteridae had decreased with increasing pH (between pH < 4 and >8) [18], indicating that other soil variables in addition to pH may be contributing to the changes in the relative abundances of Actinobacteria. The most abundant Proteobacterial families were Bradyrhizobiaceae, Acetobacteraceae and Caulobacteraceae, the latter two were relatively more abundant in C. teretifolia soils while Bradyrhizobiaceae, which includes four genera containing nitrogen-fixing bacteria [67], was relatively enriched in C. pallens soil (Fig 4).

Fig 4. Heatmap of most abundant families in each sample based on 16S rRNA gene sequences.

The heatmap indicates relative proportions of most abundant families within each sample. Double dendrograms are based on average linkage hierarchical clustering.

OTUs unique to either the C. pallens or C. teretifolia soils

Comparative analysis of the bacterial community in both grassland soils revealed 1118 and 334 of the 16S rRNA gene OTUs were present in only C. pallens or C. teretifolia soils, respectively. Many (over 40%) of the sequences that were absent from C. pallens soils were classified as Chloroflexi in the class Ktedonobacteria. Phylogenetic analyses of the most abundant (> 100 reads) Ktedonobacteria OTUs identified four clades consisting of 20 OTUs that were only present in C. teretifolia soils, and none of these were closely related to any previously identified Ktedonobacteria strains (S2 Fig; marked with bars). The closest cultured relative of many of these OTUs was Ktedonobacter racemifier, which has been described as an aerobic heterotroph that grows well at moderately acidic (pH 4.8–6.8) and microaerophilic conditions [68,69]. The C. teretifolia soils are often poorly drained [29,33] and thus microaerobic conditions are likely to develop. The unique Ktedonobacter OTUs in the C. teretifolia soils might indicate an increase in closely related new or rare species that might be capable of growth in microaerobic conditions. This is also consistent with a previous report of positive correlation between increased relative abundance and diversity of Chloroflexi that was proposed to indicate recruitment of new species or increase in rare species [70].

In contrast, sequences that were only present in C. pallens soils consisted of approximately 20% Proteobacteria (mostly Betaproteobacteria, family Nitrosomonadaceae), 17% Chloroflexi (unclassified beyond the phylum level) and 9% each of Planctomycetes and Verrucomicrobia (Spartobacteria). The absence of the family Nitrosomonadaceae in more acidic C. teretifolia soils might reflect selection against ammonium-oxidizing bacteria as they prefer a higher pH range (pH 6–9) for growth and nitrification [71,72]. However, phylogenetic analyses of the abundant OTUs belonging to the family Nitrosomonadaceae did not group the OTUs with closely related to known ammonium-oxidizers in the genera Nitrosospira and Nitrosomonas (S3 Fig).

The relative abundance of Planctomycetes was significantly higher in C. pallens grassland soils (S4 Table), which has higher nitrate N levels compared to C. teretifolia soils [33]. Increased abundance and diversity of Planctomycetes has previously been correlated with higher nitrate concentrations as a result of fertilization and cultivation [73,74]. Despite being the most common verrucomicrobial class in soil [75], only one representative from Spartobacteria has been cultured to date [76]. Therefore, the ecological relevance of the increased relative abundance of Spartobacteria in C. pallens grassland soils is not known.

Differences in putative diazotrophic communities between the two grassland soils

Sequences were distributed into nifH clusters I, II and III (Fig 3b) as defined previously: cluster I includes nifH sequences that encode the conventional FeMo nitrogenases found in Cyanobacteria, Alpha-, Beta- and Gammaproteobacteria; cluster II contains sequences belonging to the alternative FeV and FeFe nitrogenases and some archaea; cluster III includes nifH sequences from anaerobic bacteria such as Clostridia and Deltaproteobacteria [62]. In both C. pallens and C. teretifolia soils the majority of nifH gene sequences belonged to cluster I Proteobacteria (61.1 ± 16.6 and 79.9 ± 3.2%, respectively); however, the majority of OTUs did not form clades with sequences from nitrogen-fixing soil genera such as Azospirillum, Bradyrhizobium, Burkholderia, Methylocystis and Methylocella (Fig 5). The largest difference between the sites was an increase in the number of nifH cluster I Proteobacterial OTUs in C. teretifolia soils (ANOVA, P = 0.038; Fig 3b). In contrast, one of the most abundant OTUs (OTU_13), represented 5.8% of nifH sequences and was almost entirely only found in C. pallens soils (Fig 5). This OTU was in a strongly supported group with an uncultured nitrogen-fixing bacterium from a pine forest site (KF847355; [77]). These were contained in a group with sequences representing members of the genera Azospirillum, Mesorhizobium and Sinorhizobium (Fig 5). Overall, the high abundances of nifH sequences related to Azospirillum and Bradyrhizobium (Fig 5) across both grassland soils are in accordance to the dominance of diazotrophs within with these genera in soils [78]. In addition, the most abundant 16S rRNA gene-based OTU appear to belong to Bradyrhizobium (Alphaproteobacteria). Several of the most abundant nifH OTUS are related to Bradyrhizobia (shaded in gray; Fig 5) that share 91–99% sequence identity (BLASTN) may be represented by the same 16S rRNA-based OTU. The nifH genes are frequently more variable than 16S rRNA genes, for example strains with < 3% sequence dissimilarity for the 16S rRNA genes have been reported to have up to 23% dissimilarity in their nifH genes [39].

Fig 5. Bayesian phylogenetic tree inferred from 97% nucleotide similarity clustered nifH gene sequences and their nearest relatives grouped into canonical nifH clusters.

Heatmap illustrating the relative number of sequences for each OTU is shown, with the lower values of the dataset coloured in blue and the highest value in bright red. Numbers associated with nodes are Bayesian posterior probabilities. GenBank accession numbers are indicated next to reference taxa. For clarity purposes only OTUs with 100 sequences and more have been included in this tree.

The nifH cluster I Cyanobacteria showed trends of increased abundance in C. pallens (6.2 ± 3.2%) versus C. teretifolia (2.9 ± 2.8%) sites (Fig 3). In both sites, nifH sequences comprised of the filamentous heterocyst-forming genera Nostoc, Cylindrospermum and Calothrix. One OTU (OTU_389), which was in a well-supported clade with the heterocystous nitrogen-fixer Cylindrospermum stagnale PCC 7417 and OTU_645, was found in all C. pallens sites but was present at low abundance in only two C. teretifolia sites (Fig 5). Higher phosphate concentrations in C. pallens soils [33] could influence nitrogen-fixation activity, and it has been shown that phosphate inoculations increased the activity of nitrogen-fixing cyanobacteria in a salt marsh [79]. Similar cyanobacteria from the genera Nostoc and Nodularia have already been described as important sources of N input in boreal N-limited forest [80] and alpine ecosystems [81]. In fact, non-symbiotic heterocystous bacteria, primarily Nostoc sp., were attributed as the main contributors of biologically-fixed N in a natural temperate grasslands lacking legumes [82]. However, these represented only a small proportion of nifH OTUs in this study.

The nifH cluster III is composed of anaerobic organisms [38], and this cluster has been estimated to constitute around 11% of the nifH pool in soils [49]. There was increased relative abundance of cluster III nifH sequences in C. pallens sites (22.1 ± 10.7%) compared to C. teretifolia sites (14.1 ± 5.0%) (Fig 3), and a large proportion of nifH cluster III sequences was found predominantly in C. pallens samples (Fig 5). This diverse group consisted of at least nine OTUs that were not closely related to any known cultured diazotroph but is closely related to an uncultured soil bacterium isolated from white spruce rhizospheres (DQ776593; Fig 5) [83]. The most abundant single OTU (OTU_6; Fig 5), is in a well-supported group with other OTUs and uncultured bacterium (EU090282) from an acidic peat bog. Sister to this group is a well-supported clade containing Opitutacaea bacterium TAV5 (Fig 5). There were ten nifH OTUs (> 100 sequences) that were found exclusively in C. pallens soils (Fig 5).

Cluster II represented a smaller percentage of sequences in this study (8.8 ± 12.3% in C. pallens and 2.2 ± 5.5% in C. teretifolia sites). This cluster contains genes encoding the alternative nitrogenases with FeV and FeFe metal clusters (encoded by vnfH and anfH genes, respectively) and nifH sequences of some methanogenic archaea [38]. An earlier survey in NZ soils proposed that nitrogen-fixing Clostridia are more abundant in soils with a higher N availability [84]. Consistent with this, one of the most abundant nifH-based OTUs (OTU_19) in cluster II was more abundant in C. pallens soils (Fig 5). However, as this OTU was only distantly related to the genera Diplosphaera, Treponema and Clostridium (Fig 5), it is difficult to ascertain the phylogenetic affiliation of this OTU and other sequences in cluster II.

The nifH sequences belonging to cluster IA, which contains predominantly obligate anaerobes, were found exclusively in C. pallens soils (Fig 3). The single OTU (OTU_901) in cluster IA with > 100 sequences was related to Geobacter sp. and Pelobacter sp. (Deltaproteobacteria) (Fig 5). The Geobacter-Pelobacter group is anaerobic and the ability to fix N2 in nutrient-deficient environments has seen various Geobacter spp. thrive in a wide variety of ecosystems, from temperate forests [77] to extreme environments such as Antarctic zones and alpine glaciers [85,86]. Therefore, it is surprising that these bacteria might be more common in the generally higher nutrient and well-drained C. pallens soils.

Abundance of 16S rRNA and nifH genes in C. pallens and C. teretifolia soils

A greater number of OTUs and higher Shannon diversity index indicated that nifH gene diversity was greater in C. teretifolia soils. However, the result was not significant when we used nifH OTUs that were clustered based on 90% identity, indicating the diversity in C. teretifolia soils resulted from closely related bacteria. To determine whether the abundance of the nitrogen-fixing community differed in these grassland soils, we quantified the abundance of bacterial and archaeal 16S rRNA and nifH genes. Using qPCR the number of copies of the 16S rRNA and nifH genes were calculated per g of soil. Gene copy numbers of bacterial 16S rRNA genes were significantly higher in C. teretifolia soils compared to C. pallens soils (ranging from 4.9 × 106–3.8 × 107 copies g-1, and 3.3 × 107–1.2 × 108 copies g-1 soil for C. pallens and C. teretifolia soils, respectively) (ANOVA, P <0.01). Gene copy numbers of nifH ranged from 8.0 × 104–4.1 × 105 copies g-1 in C. pallens soils and 7.2 × 105–2.0 × 106 copies g-1 soil in C. teretifolia soils (ANOVA, P <0.001). However, the ratio of nifH to 1000 copies of 16S rRNA gene copies was similar in the two soil types, i.e., 15.2 ± 4.3 and 16.6 ± 4.3 in C. pallens and C. teretifolia soils, respectively (S4 Fig). This indicated an increased number of bacterial and archaeal cells per g soil in C. teretifolia as well as an increase in diazotrophs per g soil, but no change in the proportion of diazotrophs in the community between the two soil types.

Microbial biomass carbon and total N were two of the drivers of nifH abundance in managed and remnant (minimally disturbed) soils in Australia [87]. Therefore, it is suggested that soil N availability not only impacts the above-ground vegetation productivity [33], but also the diazotroph community. However, we did not find support for the hypothesis that greater abundance of diazotrophs based on nifH to 16S rRNA gene ratio would be found in soils with higher N availability (C. pallens soils). Agricultural soils with higher nitrate and ammonium levels have been reported to contained lower nifH gene copy numbers g-1 dry soil following fertilization [78]. Ammonium is known to inhibit nitrogenase [88] and thereby could be causing a decrease of diazotrophs in C. pallens soils. Higher N availability might reflect differences in the nifH transcript levels or different N mineralization rate of the bacterial communities.

Predicted functional analysis using Tax4Fun

To assess how variation in edaphic properties may influence the functional profiles of soil bacterial communities in these grassland soils, we used the approach of inferred metagenomics using Tax4Fun from 16S rRNA data [51]. Tax4Fun analysis revealed many KEGG Ortholog (KO) profiles related to sporulation among the 50 most differentially abundant KOs, these were significantly more abundant in C. pallens soils (P < 0.01, all analyses, Mann-Whitney U-test; S5 Table). Higher relative abundances of spore-forming genera, i.e., Clostridium, Bacillus, Paenibacillus and Anoxybacillus, were found in C. pallens soils (data not shown). Dormancy in both spore-forming and non-spore-forming bacteria is thought to provide the opportunity for functionally redundant species to co-exist [89,90]; and therefore, may buffer the impacts of changing environmental conditions and disturbances [91]. This could lead to the maintenance of ecosystem and community functioning in spite of shifts in soil microbial community composition, which can have major implications for fragile ecosystems.


In this study, we found that both the total bacterial community and diazotrophic community were distinct between two pristine grassland soils that are dominated by different but closely related Chionochloa species in New Zealand. Total bacterial diversity was greater in the C. pallens soils, which had higher nutrient availability [33], and this is consistent with previous observations in grassland soils [92]. Bacterial phyla exhibited ecological traits according to the availability of nutrients in soils as previously predicted [15] to an extent. We found that the more acidic, nutrient poor C. teretifolia soils harboured a greater abundance of oligotrophs, i.e., Acidobacteria, compared to copiotrophs. On the other hand, greater relative abundance of Proteobacteria, Bacteroidetes and Firmicutes were found in C. pallens soils; consistent with slightly higher pH and higher N availability soils favouring copiotrophs over oligotrophs. It should be noted many additional factors differ between the two soil types that may contribute to the differences in the bacterial communities. The increased diversity in the C. pallens soils might be expected to result in a higher level of functional redundancy, and hence, more resistance and resilience to disturbances [93]. Increased abundance of genes relating to sporulation was predicted in C. pallens soils; this could also indicate greater functional redundancy in those soils.

Although total bacterial and nitrogen-fixing communities differed between soils, the proportion of the bacterial community that was diazotrophic was similar. In contrast to the total bacterial community, the diversity of the diazotrophic community was greater in the C. teretifolia soils, potentially increasing resilience of the diazotroph community in this soil. This highlights the importance of understanding the ecological role and responses of different communities to a changing environment, particularly in these habitats where alpine plant biodiversity is vulnerable to the potential effects of climate change, with models predicting a 33–50% loss of indigenous alpine species following a 3°C temperature rise [94]. Microbial communities play important roles in nutrient cycling in soils [95] and have a significant impact on the productivity and diversity of plant communities [96]. Elevated temperature altered bacterial community composition and specifically diazotroph community structure in prairie grasslands [97,98]. Furthermore, changes in soil diazotroph community structure can impact on nitrogen fixation rates [99]. In this study, extensive differences were observed in total bacterial and diazotrophic communities between two sites which are geographically close and dominated by grasses belonging to the same genus. We predict these differences will alter the response of the communities to environmental change.

Supporting information

S1 Fig. Rarefaction curves of sub-sampled datasets indicating the observed number of OTUs within the 198 769 random 16S rRNA gene sequences (a and b) and 9658 random nifH gene sequences (c and d).

Clustering was performed at 97% similarity cut-off.


S2 Fig. Bayesian phylogenetic tree inferred from 97% nucleotide similarity clustered 16S rRNA gene sequences from the class Ktedonobacteria (Chloroflexi) and their nearest relatives.

16S rRNA gene of Chloroflexus aggregans was used as an outgroup. Heatmap illustrating the relative number of sequences for each OTU is shown, with the lower values of the dataset coloured in blue and the highest value in bright red. Numbers associated with nodes are Bayesian posterior probabilities. GenBank accession numbers are indicated next to reference taxa. For clarity purposes only OTUs with 100 sequences and more have been included in this tree.


S3 Fig. Bayesian phylogenetic tree inferred from 97% nucleotide similarity clustered 16S rRNA gene sequences from the family Nitrosomonadaceae (Betaproteobacteria) and their nearest relatives.

16S rRNA gene of Nitrosococcus oceani (Gammaproteobacteria) was used as an outgroup. (Numbers associated with nodes are Bayesian posterior probabilities. GenBank accession numbers are indicated next to reference taxa. For clarity purposes only OTUs with 100 sequences and more have been included in this tree.


S4 Fig. Proportional abundances of nifH per 1000 copies of 16S rRNA gene in Chionochloa pallens (CP) and Chionochloa teretifolia (CT) grassland soil samples, as quantified by real-time quantitative PCR.


S1 Table. Nucleotide sequences of primers used in the construction of libraries for Illumina sequencing.

Lowercase letters indicate adapter sequences required for binding to the flow cell, underlined lowercase indicate binding sites for the Illumina sequencing primers, bold uppercase indicate the Illumina TruSeq 6-bp index sequence, and regular uppercase are the 16S rRNA and nifH gene primers 0341F/0781R (Klindworth et al., 2013) and 19F/407R (Ueda et al., 1995), respectively.


S2 Table. Number of sequences before and after initial processing and filtering steps.


S3 Table. Bray-Curtis distances between 16S rRNA and nifH genes-based bacterial community samples from C. pallens and C. teretifolia soils.


S4 Table. Relative abundances of bacterial phyla and proteobacterial classes in Chionochloa pallens and Chionochloa teretifolia grassland soils.

*P < 0.1, **P < 0.05 and ***P < 0.001 denote significant differences between C. pallens and C. teretifolia samples (one-way ANOVA).


S5 Table. Abundance of KEGG orthologs (KO) and selected genes associated with sporulation predicted using Tax4Fun from 16S rRNA gene dataset.



We thank Dr Adrian Monks, Landcare Research, Dunedin for the collection of soil samples.

Author Contributions

  1. Conceptualization: JCG WGL DAO TCS.
  2. Formal analysis: JCG DAO TCS.
  3. Funding acquisition: TCS JCG.
  4. Investigation: JCG DAO TCS.
  5. Methodology: JCG WGL DAO TCS.
  6. Visualization: JCG TCS.
  7. Writing – original draft: JCG TCS.
  8. Writing – review & editing: JCG WGL DAO TCS.


  1. 1. White RP, Murray S, Rohweder M, Prince SD, Thompson KMJ. Grassland Ecosystems. World Resources Institute Washington, DC, USA; 2000.
  2. 2. Blair J, Nippert J, Briggs J. Grassland Ecology. In: Monson RK, editor. Ecology and the Environment. New York, NY: Springer New York; 2014. pp. 389–423.
  3. 3. Leff JW, Jones SE, Prober SM, Barberán A, Borer ET, Firn JL, et al. Consistent responses of soil microbial communities to elevated nutrient inputs in grasslands across the globe. Proc Natl Acad Sci U S A. National Academy of Sciences; 2015;112: 10967–10972. pmid:26283343
  4. 4. Steenwerth KL, Jackson LE, Calderón FJ, Stromberg MR, Scow KM. Soil microbial community composition and land use history in cultivated and grassland ecosystems of coastal California. Soil Biol Biochem. 2002;34: 1599–1611.
  5. 5. Lauber CL, Strickland MS, Bradford MA, Fierer N. The influence of soil properties on the structure of bacterial and fungal communities across land-use types. Soil Biol Biochem. 2008;40: 2407–2415.
  6. 6. Wakelin SA, Barratt BIP, Gerard E, Gregg AL, Brodie EL, Andersen GL, et al. Shifts in the phylogenetic structure and functional capacity of soil microbial communities follow alteration of native tussock grassland ecosystems. Soil Biol Biochem. 2013;57: 675–682.
  7. 7. Zhalnina K, Dias R, de Quadros PD, Davis-Richardson A, Camargo FAO, Clark IM, et al. Soil pH determines microbial diversity and composition in the park grass experiment. Microb Ecol. 2014;69: 395–406. pmid:25395291
  8. 8. Herzog S, Wemheuer F, Wemheuer B, Daniel R. Effects of Fertilization and Sampling Time on Composition and Diversity of Entire and Active Bacterial Communities in German Grassland Soils. PLoS One. 2015;10: e0145575. pmid:26694644
  9. 9. Lauber CL, Ramirez KS, Aanderud Z, Lennon J, Fierer N. Temporal variability in soil microbial communities across land-use types. ISME J. 2013;7: 1641–1650. pmid:23552625
  10. 10. Koyama A, Wallenstein MD, Simpson RT, Moore JC. Soil bacterial community composition altered by increased nutrient availability in Arctic tundra soils. Front Microbiol. 2014;5: pmid:25324836
  11. 11. Pan Y, Cassman N, de Hollander M, Mendes LW, Korevaar H, Geerts RHEM, et al. Impact of long-term N, P, K, and NPK fertilization on the composition and potential functions of the bacterial community in grassland soil. FEMS Microbiol Ecol. 2014;90: 195–205. pmid:25046442
  12. 12. Bier RL, Bernhardt ES, Boot CM, Graham EB, Hall EK, Lennon JT, et al. Linking microbial community structure and microbial processes: an empirical and conceptual overview. FEMS Microbiol Ecol. 2015;91: 1–11. pmid:26371074
  13. 13. Torsvik V, Øvreås L. Microbial diversity and function in soil: from genes to ecosystems. Curr Opin Microbiol. 2002;5: 240–245. pmid:12057676
  14. 14. Wagg C, Bender SF, Widmer F, van der Heijden MGA. Soil biodiversity and soil community composition determine ecosystem multifunctionality. Proc Natl Acad Sci U S A. 2014;111: 5266–5270. pmid:24639507
  15. 15. Fierer N, Bradford MA, Jackson RB. Toward an ecological classification of soil bacteria. Ecology. 2007;88: 1354–1364. pmid:17601128
  16. 16. Poly F, Ranjard L, Nazaret S, Gourbière F, Monrozier LJ. Comparison of nifH gene pools in soils and soil microenvironments with contrasting properties. Appl Environ Microbiol. 2001;67: 2255–2262. pmid:11319109
  17. 17. Fierer N, Lauber CL, Ramirez KS, Zaneveld J, Bradford MA, Knight R. Comparative metagenomic, phylogenetic and physiological analyses of soil microbial communities across nitrogen gradients. ISME J. 2012;6: 1007–1017. pmid:22134642
  18. 18. Lauber CL, Hamady M, Knight R, Fierer N. Pyrosequencing-based assessment of soil pH as a predictor of soil bacterial community structure at the continental scale. Appl Environ Microbiol. 2009;75: 5111–5120. pmid:19502440
  19. 19. Ramirez KS, Craine JM, Fierer N. Consistent effects of nitrogen amendments on soil microbial communities and processes across biomes. Glob Chang Biol. 2012;18: 1918–1927.
  20. 20. Mark AF, McLennan B. The conservation status of New Zealand’s indigenous grasslands. New Zeal J Bot. 2005;43: 235–270.
  21. 21. McAlpine KG, Wotton DM. Conservation and the delivery of ecosystem services: a literature review. Science for Conservation. 2009. p. 81.
  22. 22. O’Connor KF. Nitrogen balances in natural grasslands and extensively-managed grassland systems. N Z J Ecol. 1983;6: 1–18.
  23. 23. Reed SC, Cleveland CC, Townsend AR. Functional ecology of free-living nitrogen fixation: a contemporary perspective. Annu Rev Ecol Evol Syst. 2011;42: 489–512.
  24. 24. Lindsay EA, Colloff MJ, Gibb NL, Wakelin SA. The abundance of microbial functional genes in grassy woodlands is influenced more by soil nutrient enrichment than by recent weed invasion or livestock exclusion. Appl Environ Microbiol. 2010;76: 5547–5555. pmid:20601513
  25. 25. Crews TE, Farrington H, Vitousek PM. Changes in asymbiotic, heterotrophic nitrogen fixation on leaf litter of Metrosideros polymorpha with long-term ecosystem development in Hawaii. Ecosystems. 2000;3: 386–395.
  26. 26. Barron AR, Wurzburger N, Bellenger JP, Wright SJ, Kraepiel AML, Hedin LO. Molybdenum limitation of asymbiotic nitrogen fixation in tropical forest soils. Nat Geosci. 2008;2: 42–45.
  27. 27. Cusack DF, Silver W, McDowell WH. Biological nitrogen fixation in two tropical forests: Ecosystem-level patterns and effects of nitrogen fertilization. Ecosystems. 2009;12: 1299–1315.
  28. 28. Mark AF. Ecology of snow tussocks in the mountain grasslands of New Zealand. Vegetatio. Springer; 1969;18: 289–306.
  29. 29. Connor HE. Chionochloa Zotov (Gramineae) in New Zealand. New Zeal J Bot. 1991;29: 219–282. Available:
  30. 30. Williams PA, Grigg JL, Nes P, O’Connor KF. Vegetation/soil relationships and distribution of selected macroelements within the shoots of tall-tussocks on the Murchison Mountains, Fiordland, New Zealand. New Zeal J Bot. 1976;14: 29–53.
  31. 31. Mark AF, Barratt BIP, Weeks E, Dymond JR. Ecosystem services in New Zealand’s indigenous tussock grasslands: conditions and trends. Manaaki Whenua Press, Landcare Research, Lincoln, NZ; 2013; 1–33.
  32. 32. Lee WG, Fenner M. Mineral nutrient allocation in seeds and shoots of twelve Chionochloa species in relation to soil fertility. J Ecol. 1989;77: 704–716.
  33. 33. Tanentzap AJ, Lee WG, Coomes DA. Soil nutrient supply modulates temperature-induction cues in mast-seeding grasses. Ecology. 2012;93: 462–469. pmid:22624201
  34. 34. Pirie MD, Lloyd KM, Lee WG, Linder HP. Diversification of Chionochloa (Poaceae) and biogeography of the New Zealand Southern Alps. J Biogeogr. 2010;37: 379–392.
  35. 35. Nacke H, Thürmer A, Wollherr A, Will C, Hodac L, Herold N, et al. Pyrosequencing-based assessment of bacterial community structure along different management types in German forest and grassland soils. PLoS One. 2011;6: e17000. pmid:21359220
  36. 36. Coolon JD, Jones KL, Todd TC, Blair JM, Herman MA. Long-term nitrogen amendment alters the diversity and assemblage of soil bacterial communities in tallgrass prairie. PLoS One. 2013;8: e67884. pmid:23840782
  37. 37. Weber CF, Lockhart JS, Charaska E, Aho K, Lohse KA. Bacterial composition of soils in ponderosa pine and mixed conifer forests exposed to different wildfire burn severity. Soil Biol Biochem. 2014;69: 242–250.
  38. 38. Zehr JP, Jenkins BD, Short SM, Steward GF. Nitrogenase gene diversity and microbial community structure: a cross-system comparison. Environ Microbiol. 2003;5: 539–554. Available: pmid:12823187
  39. 39. Gaby JC, Buckley DH. A comprehensive aligned nifH gene database: a multipurpose tool for studies of nitrogen-fixing bacteria. Database. 2014; article ID bau001. 10.1093/database/bau001
  40. 40. Klindworth A, Pruesse E, Schweer T, Peplies J, Quast C, Horn M, et al. Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next-generation sequencing-based diversity studies. Nucleic Acids Res. 2013;41: 1–11.
  41. 41. Ueda T, Suga Y, Yahiro N, Matsuguchi T. Remarkable N2-fixing bacterial diversity detected in rice roots by molecular evolutionary analysis of nifH gene sequences. J Bacteriol. 1995;177: 1414–1417. pmid:7868622
  42. 42. Gaby JC, Buckley DH. A comprehensive evaluation of PCR primers to amplify the nifH gene of nitrogenase. PLoS One. 2012;7: e42149. pmid:22848735
  43. 43. Schloss PD, Westcott SL, Ryabin T, Hall JR, Hartmann M, Hollister EB, et al. Introducing mothur: open-source, platform-independent, community-supported software for describing and comparing microbial communities. Appl Environ Microbiol. 2009;75: 7537–7541. pmid:19801464
  44. 44. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK, et al. QIIME allows analysis of high-throughput community sequencing data. Nat Methods. 2010;7: 335–336. pmid:20383131
  45. 45. Edgar RC. UPARSE: highly accurate OTU sequences from microbial amplicon reads. Nat Methods. 2013;10: 996–998. pmid:23955772
  46. 46. Wang Q, Quensen JFI, Fish JA. Ecological patterns of nifH genes in four terrestrial climatic zones explored with targeted metagenomics using FrameBot, a new informatics tool. MBio. 2013;4: e00592–13. pmid:24045641
  47. 47. Haas BJ, Gevers D, Earl AM, Feldgarden M, Ward D V, Giannoukos G, et al. Chimeric 16S rRNA sequence formation and detection in Sanger and 454-pyrosequenced PCR amplicons. Genome Res. 2011;21: 494–504. pmid:21212162
  48. 48. Wang Q, Garrity GM, Tiedje JM, Cole JR. Naive Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Appl Environ Microbiol. 2007;73: 5261–5267. pmid:17586664
  49. 49. Gaby JC, Buckley DH. A global census of nitrogenase diversity. Environ Microbiol. 2011;13: 1790–1799. pmid:21535343
  50. 50. Raymond J, Siefert JL, Staples CR, Blankenship RE. The Natural History of Nitrogen Fixation. Mol Biol Evol. 2004;21: 541–554. pmid:14694078
  51. 51. Aßhauer KP, Wemheuer B, Daniel R, Meinicke P. Tax4Fun: predicting functional profiles from metagenomic 16S rRNA data. Bioinformatics. 2015;31: 2882–2884. pmid:25957349
  52. 52. R Core Team. R Development Core Team. In: R: A language and environment for statistical computing. 2013.
  53. 53. Kearse M, Moir R, Wilson A, Stones-Havas S, Cheung M, Sturrock S, et al. Geneious Basic: an integrated and extendable desktop software platform for the organization and analysis of sequence data. Bioinformatics. 2012;28: 1647–1649. pmid:22543367
  54. 54. Altekar G, Dwarkadas S, Huelsenbeck JP, Ronquist F. Parallel metropolis coupled Markov chain Monte Carlo for Bayesian phylogenetic inference. Bioinformatics. 2004;20: 407–415. pmid:14960467
  55. 55. Ronquist F, Teslenko M, van der Mark P, Ayres DL, Darling A, Höhna S, et al. MrBayes 3.2: efficient Bayesian phylogenetic inference and model choice across a large model space. Syst Biol. 2012;61: 539–542. pmid:22357727
  56. 56. Nylander JAA. MrModeltest v2. Program distributed by the author. Evol Biol Centre, Uppsala Univ.
  57. 57. Apprill A, McNally S, Parsons R, Webe L. Minor revision to V4 region SSU rRNA 806R gene primer greatly increases detection of SAR11 bacterioplankton. Aquat Microb Ecol. 2015;75: 129–137.
  58. 58. Rösch C, Mergel A, Bothe H. Biodiversity of denitrifying and dinitrogen-fixing bacteria in an acid forest soil. Appl Environ Microbiol. 2002;68: 3818–3829. pmid:12147477
  59. 59. Lozupone C, Hamady M, Knight R. UniFrac—An online tool for comparing microbial community diversity in a phylogenetic context. BMC Bioinformatics. 2006;7: 371. pmid:16893466
  60. 60. Anderson MJ. Permutational multivariate analysis of variance. Department of Statistics, University of Auckland, Auckland. 2005.
  61. 61. Warnes GR, Bolker B, Bonebakker L, Gentleman R, Huber W, Liaw A, et al. gplots: Various R programming tools for plotting data. 2013. R Packag version. 2014;2.
  62. 62. Chien Y-T, Zinder SH. Cloning, functional organization, transcript studies, and phylogenetic analysis of the complete nitrogenase structural genes (nifHDK2) and associated genes in the archaeon Methanosarcina barkeri 227. J Bacteriol. 1996;178: 143–148. pmid:8550408
  63. 63. Ramirez KS, Lauber CL, Knight R, Bradford MA, Fierer N. Consistent effects of nitrogen fertilization on soil bacterial communities in contrasting systems. Ecology. 2010;91: 3414–3463.
  64. 64. Will C, Thürmer A, Wollherr A, Nacke H, Herold N, Schrumpf M, et al. Horizon-specific bacterial community composition of german grassland soils, as revealed by pyrosequencing-based analysis of 16S rRNA genes. Appl Environ Microbiol. 2010;76: 6751–6759. pmid:20729324
  65. 65. Větrovský T, Baldrian P. The Variability of the 16S rRNA gene in bacterial genomes and its consequences for bacterial community analyses. PLoS One. 2013;8: e57923. pmid:23460914
  66. 66. Jones RT, Robeson MS, Lauber CL, Hamady M, Knight R, Fierer N. A comprehensive survey of soil acidobacterial diversity using pyrosequencing and clone library analyses. ISME J. 2009;3: 442–453. pmid:19129864
  67. 67. Rosenberg E, DeLong EF, Stephen L, Erko S, Thompson F, editors. The Family Bradyrhizobiaceae. The Prokaryotes: Alphaproteobacteria and Betaproteobacteria. Berlin Heidelberg: Springer; 2014. pp. 135–154.
  68. 68. Cavaletti L, Monciardini P, Bamonte R, Schumann P, Rohde M, Sosio M, et al. New lineage of filamentous, spore-forming, gram-positive bacteria from soil. Appl Environ Microbiol. 2006;72: 4360–4369. pmid:16751552
  69. 69. Chang Y, Land M, Hauser L, Chertkov O, Del Rio TG, Nolan M, et al. Non-contiguous finished genome sequence and contextual data of the filamentous soil bacterium Ktedonobacter racemifer type strain (SOSP1-21T). Stand Genomic Sci. 2011;5: 97–111. pmid:22180814
  70. 70. Youssef NH, Elshahed MS. Diversity rankings among bacterial lineages in soil. ISME J. 2009;3: 305–313. pmid:18987677
  71. 71. Stienstra AW, Klein Gunnewiek P, Laanbroek HJ. Repression of nitrification in soils under a climax grassland vegetation. FEMS Microbiol Ecol. 1994;14: 45–52.
  72. 72. Kowalchuk GA, Stephen JR. Ammonia-oxidizing bacteria: a model for molecular microbial ecology. Annu Rev Microbiol. 2001;55: 485–529. pmid:11544365
  73. 73. Buckley DH, Schmidt TM. Diversity and dynamics of microbial communities in soils from agro-ecosystems. Environ Microbiol. 2003;5: 441–452. pmid:12755711
  74. 74. Buckley DH, Huangyutitham V, Nelson TA, Rumberger A, Thies JE. Diversity of Planctomycetes in soil in relation to soil history and environmental heterogeneity. Appl Environ Microbiol. 2006;72: 4522–4531. pmid:16820439
  75. 75. Janssen PH. Identifying the dominant soil bacterial taxa in libraries of 16S rRNA and 16S rRNA genes. Appl Environ Microbiol. 2006;72: 1719–1728. pmid:16517615
  76. 76. Sangwan P, Chen X, Hugenholtz P, Janssen PH. Chthoniobacter flavus gen. nov., sp. nov., the first pure-culture representative of subdivision two, Spartobacteria classis nov., of the phylum Verrucomicrobia. Appl Environ Microbiol. 2004;70: 5875–5881. pmid:15466527
  77. 77. Berthrong ST, Yeager CM, Gallegos-Graves L, Steven B, Eichorst SA, Jackson RB, et al. Nitrogen fertilization has a stronger effect on soil nitrogen-fixing bacterial communities than elevated atmospheric CO2. Appl Environ Microbiol. 2014;80: 3103–3112. pmid:24610855
  78. 78. Pereira e Silva MC, Schloter-Hai B, Schloter M, van Elsas JD, Salles JF. Temporal dynamics of abundance and composition of nitrogen-fixing communities across agricultural soils. PLoS One. 2013;8: e74500. pmid:24058578
  79. 79. Liengen T. Environmental factors influencing the nitrogen fixation activity of free-living terrestrial cyanobacteria from a high arctic area, Spitsbergen. Can J Microbiol. 1999;45: 573–581.
  80. 80. Leppänen SM, Salemaa M, Smolander A, Mäkipää R, Tiirola M. Nitrogen fixation and methanotrophy in forest mosses along a N deposition gradient. Environ Exp Bot. 2013;90: 62–69.
  81. 81. Fritz-Sheridan RP. Physiological ecology of nitrogen-fixing blue-green algal crusts in the upper-subalpine life zone. J Phycol. 1988;24: 302–309.
  82. 82. Dubois JD, Kapustka LA. Biological nitrogen influx in an Ohio reflict prairie. Am J Bot. 1983;70: 8–16.
  83. 83. Lamarche J, Hamelin RC. No evidence of an impact on the rhizosphere diazotroph community by the expression of Bacillus thuringiensis Cry1Ab toxin by Bt white spruce. Appl Environ Microbiol. 2007;73: 6577–6583. pmid:17660307
  84. 84. Di Menna ME. The incidence of Azotobacter and of nitrogen-fixing clostridia in some New Zealand soils. New Zeal J Agric Res. 1966;9: 218–226.
  85. 85. Duc L, Noll M, Meier BE, Bürgmann H, Zeyer J. High diversity of diazotrophs in the forefield of a receding alpine glacier. Microb Ecol. 2009;57: 179–190. pmid:18563478
  86. 86. Niederberger TD, Sohm JA, Tirindelli J, Gunderson T, Capone DG, Carpenter EJ, et al. Diverse and highly active diazotrophic assemblages inhabit ephemerally wetted soils of the Antarctic Dry Valleys. FEMS Microbiol Ecol. 2012;82: 376–390. pmid:22500944
  87. 87. Hayden HL, Drake J, Imhof M, Oxley APA, Norng S, Mele PM. The abundance of nitrogen cycle genes amoA and nifH depends on land-uses and soil types in South-Eastern Australia. Soil Biol Biochem. 2010;42: 1774–1783.
  88. 88. Christiansen-Weniger C, van Veen JA. Nitrogen fixation by Azospirillium brasilense in soil and the rhizosphere under controlled environmental conditions. Biol Fertil Soils. 1991;12: 100–106.
  89. 89. Jones SE, Lennon JT. Dormancy contributes to the maintenance of microbial diversity. Proc Natl Acad Sci. 2010;107: 5881–5886. pmid:20231463
  90. 90. Lennon JT, Jones SE. Microbial seed banks: the ecological and evolutionary implications of dormancy. Nat Rev Microbiol. 2011;9: 119–130. pmid:21233850
  91. 91. Allison SD, Martiny JBH. Resistance, resilience, and redundancy in microbial communities. Proc Natl Acad Sci U S A. 2008;105: 11512–9. pmid:18695234
  92. 92. O’Brien SL, Gibbons SM, Owens SM, Hampton-Marcell J, Johnston ER, Jastrow JD, et al. Spatial scale drives patterns in soil bacterial diversity. Environ Microbiol. 2016;18: 2039–2051. pmid:26914164
  93. 93. Griffiths BS, Philippot L. Insights into the resistance and resilience of the soil microbial community. FEMS Microbiol Rev. 2013;37: 112–129. pmid:22568555
  94. 94. Halloy SRP, Mark AF. Climate-change effects on alpine plant biodiversity: a New Zealand perspective on quantifying the threat. Arctic, Antarct Alp Res. 2003;35: 248–254.
  95. 95. Coleman DC, Crossley DA Jr. Fundamentals of soil ecology. San Diego, CA: Academic Press; 1996.
  96. 96. Van der Heijden MGA, Bardgett RD, Van Straalen NM. The unseen majority: Soil microbes as drivers of plant diversity and productivity in terrestrial ecosystems. Ecol Lett. 2008;11: 296–310. pmid:18047587
  97. 97. Luo C, Rodriguez-R LM, Johnston ER, Wu L, Cheng L, Xue K, et al. Soil microbial community responses to a decade of warming as revealed by comparative metagenomics. Appl Environ Microbiol. 2014;80: 1777–1786. pmid:24375144
  98. 98. Penton CR, St Louis D, Pham A, Cole JR, Wu L, Luo Y, et al. Denitrifying and diazotrophic community responses to artificial warming in permafrost and tallgrass prairie soils. Front Microbiol. 2015;6: pmid:26284038
  99. 99. Hsu S-F, Buckley DH. Evidence for the functional significance of diazotroph community structure in soil. ISME J. 2009;3: 124–136. pmid:18769458