Soil bacteria and archaea change rapidly in the first century of Fennoscandian boreal forest development

Soil bacteria and archaea change rapidly in the first century of Fennoscandian boreal forest development

Soil Biology & Biochemistry 114 (2017) 160e167 Contents lists available at ScienceDirect Soil Biology & Biochemistry journal homepage: www.elsevier...

2MB Sizes 0 Downloads 63 Views

Soil Biology & Biochemistry 114 (2017) 160e167

Contents lists available at ScienceDirect

Soil Biology & Biochemistry journal homepage: www.elsevier.com/locate/soilbio

Soil bacteria and archaea change rapidly in the first century of Fennoscandian boreal forest development € gberg b Stephanie A. Yarwood a, *, Mona N. Ho a b

Environmental Science and Technology Department, University of Maryland, College Park, MD, USA Department of Forest Ecology and Management, Swedish University of Agricultural Sciences, SE-901 83 Umeå, Sweden

a r t i c l e i n f o

a b s t r a c t

Article history: Received 26 April 2017 Received in revised form 8 July 2017 Accepted 11 July 2017

Using a gradient of changing ground age, caused by glacial isostatic adjustment, we compared systems that spanned ages from 25 to 560-years-old. Illumina sequencing was applied to determine archaeal and bacterial composition, investigating how different phylogenetic groups change as ecosystems develop. Bacterial communities dramatically changed during early ecosystem development (p < 0.001), evidenced by significant compositional shifts between 25 and 115 year-old-soils. Although significant differences did occur in the three later aged sites, they did not change as much. This was consistent with vegetation that shifted from meadow (25 year) to alder dominated forest (115 year), to ecosystems containing spruce. Correlation networks revealed that the microbial communities became more interconnected in older age ecosystems with a two-fold increase in network density. Species richness had the opposite trend with a decreased number of species: 781 operational taxonomic units (OTUs) in the youngest ecosystem to 366 OTUs in the oldest ecosystem. The observed shifts in community composition are consistent with other reported ecosystem gradients, but here we show that not only does composition change, but as ecosystems age the network connectivity increases indicating potentially more social interactions among microbes or increasingly stringent plant-microbe-soil interactions. © 2017 Published by Elsevier Ltd.

Keywords: Bacterial community Boreal forest soil Soil nitrogen cycling C and N accumulation Network analysis

1. Introduction Glacial isostatic adjustment along the coast lines of northern Sweden and Finland provide a unique system to study microbial community change with soil development and ecosystem age. The age of new ground rising out of the sea is calculated from the known rate of rebound and the height over the ocean. These environments have previously been studied to understand soil € et al., 2006; Wallander et al., 2009; biogeochemistry (Merila Vincent et al., 2013; Blasko et al., 2015) plant succession (Svensson and Jeglum, 2003) on mineral soils, as well as on peat € et al., 2006; Larmola et al., 2014). In this paper we soils (eg. Merila focus exclusively on mineral soils. As these ecosystem ages and plants colonize, there is an increase in soil organic matter (SOM), a decrease in soil pH, and an increase in organic N and soil C/N ratio. Although total organic N increases, plant N decreases along with net and gross N mineralization (Blasko et al., 2015).

* Corresponding author. Environmental Science and Technology Department, University of Maryland, 1204 HJ Patterson Hall, College Park, MD 20742, USA. E-mail address: [email protected] (S.A. Yarwood). http://dx.doi.org/10.1016/j.soilbio.2017.07.017 0038-0717/© 2017 Published by Elsevier Ltd.

Plant species richness increases within the first 100 years of soil development, but decreases and remains constant between 200 and 600 years (Svensson and Jeglum, 2003). Norway spruce (Picea abies L.) colonizes at approximately ~150 years and remains dominant. Both spruce and later Scots pine (Pinus sylvestris (L.)) rely on ectomycorrhizal (ECM) fungi to supply needed N. There is evidence that ECM and associated microorganisms are N-sinks € gberg et al., 2007a), sequestering N inside mycelium and (Ho €sholm et al., potentially exacerbating N limitation to plants (Na 2013). Microbial N-immobilization could thus be one reason why boreal forests become N limited. Microbial studies reveal linkages € between vegetation and microbial community composition (Merila et al., 2002, 2010). In line with this Blasko et al. (2015) observed an increase in fungal:bacterial ratio with increasing age, and microbial composition of the fungi and bacteria differed in the early development stages according to phospholipid fatty acid (PLFA) analysis. Although plant communities have often been studied along chronosequences, fewer studies have examined microbial communities (Fierer et al., 2010). The microbes most extensively studied are the fungal saprobes (Dighton and White, 2017) and arbuscular and ECM fungi (Johnson et al., 1991; Jumpponen and

€gberg / Soil Biology & Biochemistry 114 (2017) 160e167 S.A. Yarwood, M.N. Ho

Egerton-Warburton, 2005; Chagnon et al., 2013). In one study, fungi appeared to change largely in connection to plant succession, but bacteria were correlated to edaphic changes (Cline and Zak, 2015). Additionally, soil bacterial communities, from a dune system in the Great Lakes region of Michigan, USA, changed rapidly within the first 450 years of soil development but then stabilized and remain similar for the next 500 years (Williams et al., 2013). Archaeal community composition also changed during early succession, for example composition differences were observed in deglaciated soils ranging in age from 2 years to 167 years (Zumsteg et al., 2012). We focused on the bacterial and archaeal communities under vegetation on mineral soils, querying how these populations would respond to increased ecosystem development. The chronosequence allowed us to ask: Does ecosystem development lead to differences in the interactions between individual OTUs? Network analysis, a mathematical approach used to visualize and test for interactions or related activities within a complex community (Faust and Raes, 2012), was used to look for these interactions. For example, cooccurrence networks were used in one study to confirm the cooccurrence of N-fixing bacteria and wood decay fungi (Hoppe et al., 2014) and in another study rhizosphere bacterial communities become increasingly connected as plants matured throughout the growing season (Shi et al., 2016). We used cooccurrence, taxonomic association (Hartmann et al., 2014), and bi-partite networking (Bowen et al., 2013) to test three hypotheses along the uplift chronosequence: 1) There will be rapid changes in the bacterial and archaeal community in the early stages of ecosystem development when soil C and N is accumulating and N cycling rates are higher, but communities will become less variable with time; 2) Bacterial and archaeal community composition will vary with ecosystem age due both to loss and recruitment of OTUs over time; 3) The community will be connected over time due to either increased biotic and environmental filtering or interaction. 2. Materials and methods 2.1. Study location The study sites are located in the Gulf of Bothnia (N 63 440 , E on Bjuren island, a nature reserve just off of Sweden's coast. The rate of isostatic rebound in this area is 8.16 mm per year (Ekman, 1996; Vestøl, 2006). Five ecosystem ages were designated along each of three transects: 1) a meadow that was 25 ± 12 years-old with short grasses and rushes (e.g., Agrostis stolonifera var. Bottnica, Juncus balticus); 2) a 115-year-old site dominated by grey alder (Alnus incana (L.) Moench) (112 ± 13 years) which forms root nodule symbioses with N2-fixing actinobacteria of the genus Frankia (Huss-Danell, 1997); 3) a 150-year-old ecosystem (151 ± 15 years) containing Norway spruce (Picea abies (L.)) Karst forest; 4) a Norway spruce forest that was 215 ± 24 years-old; and 5) a 560-year-old (564 ± 31 years) Norway spruce and Scots pine (Pinus sylvestris (L.)) forest (Visual abstract). The soils along each transect were sandy and silty glacial deposits (FAO) with many boulders. The soils with continuous organic mor-layer with F- and H-horizons (150-, 215-, and 560-year-old ground) are classified as Haplic Podzol (FAO classification). For more detailed descriptions of plant community composition and soil characteristics, see Blasko et al. (2015). Soils were sampled along each of the three transects by marking a temporary sampling plot (10 m  15 m) and removing five composite soil cores from randomly selected locations within each sampling plot. The 25- and 115-year-old plots, which lacked a continuous organic layer, were sampled by removing the five cm of the uppermost soil layer using a spade and a frame (30 cm  30 cm). One to two soil samples were composited to make a single soil sample. The three more developed sites contained 20 350 )

161

organic horizons, and F þ H layers of c. 10 cm depth were sampled by removing three soil cores (10 cm diameter corer) and compositing those cores to make one sample. This was repeated to yield five composite samples per ecosystem age. The five representative soil samples were to be analyzed at each ecosystem age in each transect ending up in three mean values. Soils were collected in August and kept on ice for transport back to the lab in Umeå, Sweden (Blasko et al., 2015). Chemical parameters such as pH, Total C and total N have been reported previously (Blasko et al., 2015). A subsample (1e7 g) of each soil was freeze dried, ground, and stored at 80  C until shipment on dry-ice to College Park, MD. 2.2. DNA extraction, quantification, and sequencing Soil DNA was extracted using the MoBio “Powerlyzer” Powersoil DNA isolation kit (MoBio Laboratories, Carlsbad, CA) according to manufacturer's instructions, with the exception of using a FastPrep24 Instrument (MP Biomedical, Solon, OH) set at 5.5 m/s for 45 s for cell lysis (settings recommended by the manufacturer). Purified DNA was quantified by fluorimetry using a Qubit 2.0 (Life Technologies, Carlsbad, CA). Soil DNA used in Q-PCR was diluted to 2.0 ng mL1 and 2 mL of each dilution was used in triplicate Q-PCR reactions. Genes coding for bacterial 16S rRNA were used to quantify the total bacteria and archaea using the universal primer set 505F-806R (Caporaso et al., 2012). Quantification was based on a plasmid standard containing a cloned fragment of the amplicon from an E. coli pure culture. Thermocycling conditions were 95  C for 5 min, followed by 40 cycles of 95  C for 5 s/55  C for 15 s/72  C for 10 s, and a final melting cycle to produce a melting curve for quality control. All runs had standard efficiencies curves of r2 > 0.99 and efficiencies 97e105%. An additional soil standard curve was constructed using a pool of all soil extracts and then diluted. The efficiency of this curve was 102%. In order to ensure that various inhibitors in soil extracts did not affect quantification, 1 mL of each soil extract was added together to make a single composite sample representing all samples. This composite sample was used to make a dilution series that was run against the plasmid dilution series (Hargreaves et al., 2013). In this case, the efficiency and slope of the plasmid standard and composite soil were within 5% of each other and no correction for any inhibitory effects on the PCR reaction was made. Soil DNA was diluted to 5 ng mL1 for amplicon sequencing of the bacterial and archaeal 16S rRNA. The V4 - V5 region of the 16S rRNA gene was amplified using the same universal primers 505F-806R that was used in Q-PCR (Caporaso et al., 2012). The primers contained Illumina sequencing adapters: 50 TCG TCG GCA GCG TCA GAT GTG TAT AAG AGA CAG e Forward Primer e 30 , and 50 GTC TCG TGG GCT CGG AGA TGT GTA TAA GAG ACA G e Reverse Primer e 30 (16S Metagenomic Sequencing Library Preparation Guide). Samples were amplified in duplicate and prepared for sequencing according to Illumina protocol (16S Metagenomic Sequencing Library Preparation Guide). Following clean-up of the PCR products with AMPure XP beads (Beckman Coulter, Pasadena, CA), 8 base pair (bp) nucleotide indexes were added using a Nextera XT Index Kit (Illumina, San Diego, CA). Samples were then cleaned a second time with AMPure XP beads (Beckman Coulter, Pasadena, CA) and all 75 amplified products were pooled and quantified. Each amplification was diluted to 1 ng mL1 and 2 mL of each individual dilution was combined into a single pooled sample. The pooled sample was sent to the Center for Genome Research and Biocomputing at Oregon State University. A spike of 5% PhiX (Illumina, San Diego, CA) was added to the sample. A paired-end 250 bp v2 MiSeq kit was used for sequencing (Illumina, San Diego, CA). All sequences have been deposited in the NCBI database under the BioProject accession number: PRJNA374731.

€gberg / Soil Biology & Biochemistry 114 (2017) 160e167 S.A. Yarwood, M.N. Ho

162

Fig. 1. Non-metric multidimensional scaling ordination using OTU level (97% similarity) matches to the Greengenes database for all bacterial (A) and archaeal (B) sequences. Each point represents the mean of three transects with SE (n ¼ 3). The bacterial ordination had a final stress ¼ 2.4 and the archaeal final stress ¼ 5.9. Vectors show the correlation of the composition with soil chemical variables, with numbers in parentheses indicating the r2 values.

2.3. Statistical analysis and bioinformatics Univariate statistical analyses were performed in SigmaStat (v4.0, Systat Software Inc., CA, USA). We used one-way ANOVA to test for significant differences between ecosystem ages (no

significant differences were found among transects, p > 0.668, n ¼ 3). When the assumptions of normal distribution or equal variance were not met one-way ANOVA on ranks was applied. This was followed by Tukey's post hoc test and multiple pair-wise comparisons.

Table 1 Quantitative and qualitative parameters for the bacterial and archaeal community across different ecosystem ages. Means ± 1 SE, N ¼ 3. Letters denote significant differences at p < 0.05. Fifteen rows of data were used in the statistical analysis i.e., three mean values per ecosystem age. Gene copy numbers are specific (normalized to per gram soil carbon) to account for differences in soil C concentration across the ages. Soil C concentrations are 0.9, 5.3, 36.2, 46.0, and 45.1% of dry soil for the 25-, 115-, 150-, 215-, and 560years-old ecosystems, respectively. Ecosystem Age

16S gene copy numbers g1 soil C

25 115 150 215 560

7.3 2.1 1.1 0.7 0.7

± ± ± ± ±

0.8 0.8 0.5 0.3 0.2

    

1012a 1012b 1012b 1012b 1012b

Species Richness (97% similarity) 781 564 487 400 366

± ± ± ± ±

32a 37b 43b 52b 50b

Species Evenness 1.1 0.8 0.7 0.7 0.6

± ± ± ± ±

0.3a 0.0a 0.0ab 0.0ac 0.0bc

Shannon's Diversity Index 5.0 4.8 4.2 3.9 3.7

± ± ± ± ±

0.0a 0.1a 0.1b 0.0b 0.1b

Simpson's Diversity Index 0.98 0.99 0.98 0.97 0.95

± ± ± ± ±

0.00a 0.00a 0.00ac 0.00bc 0.01b

€gberg / Soil Biology & Biochemistry 114 (2017) 160e167 S.A. Yarwood, M.N. Ho 163

Fig. 2. Taxonomic association network for A) bacteria and B) archaea of abundant bacterial and archaeal taxa. Each node was found in at least five samples and composes at least 1% of the next highest taxonomic group. The size of nodes represents the average abundance across all ecosystem ages and colors denote correlations to ecosystem ages. Red nodes are positively correlated to ecosystem age, blue nodes are negatively correlated to ecosystem age, and purple nodes represent taxonomic groups where abundance did not correlate to ecosystem age. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

164

€gberg / Soil Biology & Biochemistry 114 (2017) 160e167 S.A. Yarwood, M.N. Ho

Illumina sequences were analyzed by first joining the pairedend sequences using fastq-join function with the ea-tools bioinformatics software package (Aronesty, 2011), which is now included in the Qiime pipeline (Caporaso et al., 2010). Sequences were paired if the forward and reverse sequence had a minimum overlap of 50 bp. The demultiplexed samples were quality checked using split_libraries.py with default settings, except the “truncate_remove” option was enabled in order to remove any joined sequence that was missing a complete reverse primer. Individual sequences were concatenated and OTUs were picked using UCLUST at (98% similarity) (Edgar, 2010), and taxonomy was assigned using the Greengenes database (August 2015 release, greengens.org) (McDonald et al., 2012). Approximately 9.7 million high quality sequences were generated, with an average of 130,000 sequences for each of the 75 samples (minimum ¼ 12,700; maximum ¼ 443,000). All samples were rarefied to 11,000 sequences before multivariate statistical analysis. Samples identified at the OTU level (i.e. 98% similarity) were visualized using both PCoA (Qiime) and non-metric multidimensional scaling (McCune and Grace, 2002). Permanova was used to test for differences between ecosystem age and transect, and community composition was correlated to population size and soil edaphic factors using Pearson correlation biplots (McCune and Grace, 2002). Three different forms of network analysis were used to investigate and visualize the communities. Taxonomic association networks were constructed for the most abundant taxa by removing any group that that was absent in five or more samples or making up less than 1% of the next highest taxonomic group. Pearson correlation coefficients were calculated for group abundance correlated to ecosystem age and used to color code the network. A bipartite network was constructed to visualize the presence and absence of individual OTUs across the different ecosystem ages. In this case, all OTUs at 97% similarity and present in at least two samples were represented by points connecting to one or more of the ecosystem ages. Correlation networks were constructed for each ecosystem age (Faust and Raes, 2012). Nodes represent bacterial and archaeal classes matching at 95% confidence to Greengenes and present in at least two samples within that ecosystem age. Spearman's correlation coefficients (so as not to exclude nonlinear relationships) were calculated for each pair of nodes and any significant positive or negative correlations (p < 0.01) were joined by an edge. The networks were each arranged using Fruchterman-Reingold's algorithm (Fruchterman and Reingold, 1991).

included: Proteobacteria (24%), Planctomycetes (21%), Verrucomicrobia (12%), and Acidobacteria (14%) (Fig. 2A). Only Planctomycetes and Acidobacteria positively correlated to ecosystem age (Fig. 2A). Few phyla decreased with ecosystem age, except Cyanobacteria. The archaea contained few taxonomic groups that had any correlation to ecosystem age. The Crenarchaeota had a weak positive correlation, primarily due to the increase of the putative class MBGA (Marine Benthic Group A) (Fig. 2B). The only group that was found to be negatively correlated to ecosystem age was the candidate order YLA114 (Fig. 2B). The 25-year-old ecosystem contained 224 OTUs that were unique (Fig. 3), including several Cyanobacteria and Bacteroidetes. In comparison, other age classes had few unique OTUs and these were a wide diversity of bacteria and archaea belonging to many different phyla (Fig. 3). There were 413 OTUs present at all five ages. These OTUs included members of Verrucomicrobia, Proteobacteria, Actinobacteria, and Euryarchaeota. This ubiquitous cluster of OTUs included the only matches to the Actinobacteria putatively identified as Frankiaceae. Similar to species richness (Table 1), the number of nodes (representing classes) decreased with ecosystem age in the correlation networks (Fig. 4). In the 25-year-old network, each node correlated to 7.6 other nodes (Table 2), and the strongest correlation was between a node classified as NC10 (Ettwig et al., 2016) and the class Methanomicrobia (r ¼ 0.81) (Fig. 4). In comparison, the average connectedness values were 28.6 for the 150-year-old ecosystem (Table 2). The 560-year-old ecosystem had one highly connected module containing 81% of all the classes. This module contained members of: Acidobacteria, Actinobacteria, Proteobacteria, and Chloroflexi. The correlation networks for 150-, 215-, and 560-year-old ecosystems had similar structure and connectivity (Fig. 4 and Table 2). In contrast to the 25-year-old ecosystem, there were 349 connections with r > 0.81 in the 560-year-old ecosystem. Both network degree and diameter increased with ecosystem age (Table 2). 4. Discussion In support of hypothesis one, bacterial communities changed

3. Results Bacterial communities were significantly different between ages (PerMANOVA, F0.05, 4, 14 ¼ 11.40, p ¼ 0.0002), but archaeal communities did not differ (PerMANOVA, F0.05, 4, 14 ¼ 1.37, p ¼ 0.174). When correlated to community composition, pH had the strongest correlation to both the bacteria (r ¼ 0.96) and archaea (r ¼ 0.76) (Fig. 1). Soil C and N increased with ecosystem age and were also found to significantly correlate to community composition (Fig. 1). There was no significant difference between the three transects (data not shown). The 16S gene copy numbers g1 of dry soil was lower in the 25-year-old soils (6.3 ± 1.5  1010 16S rRNA gene copy numbers) and peaked in the 150-year-old sites (3.7 ± 1.5  1011 16S rRNA gene copy numbers). After adjusting for differences in soil C concentration, however, the specific 16S gene copy numbers (g1 soil C) were the highest in the youngest soils and then declined considerably in the 115 and 150-year-old-soils (Table 1). Species richness tended to continuously decrease with ecosystem age, but was only significantly different in the first 25 years (Table 1). Bacterial phyla that were most abundant across all locations

Fig. 3. A bipartite network for OTUs (97% similarity) matching both bacteria and archaea and present in at least two samples. Each circular node represents an OTU and the large hexagons represent the different ecosystem ages: 25 (yellow), 115 (green), 150 (aqua), 215 (blue), and 560 (violet). Nodes are color coded to show their presence in a single ecosystem age (matching color to year) or multiple ecosystem ages. Red nodes were present in all ecosystem age. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

€gberg / Soil Biology & Biochemistry 114 (2017) 160e167 S.A. Yarwood, M.N. Ho

165

Table 2 Features of correlation networks for each ecosystem age. The number of nodes denotes the number of bacterial and archaeal classes that were present in at least two samples within that ecosystem age. Degree refers to the mean significant correlations between any given node and all other nodes. Path length is the mean number of edges needed to complete a circuit within the network and therefore is one parameter associated with modularity. The diameter of the network is the size of the network calculated for each network using the same clustering algorithm. The density refers to the proportion potential edges in the network compared to actual edges (i.e. significant correlations). Ecosystem Age # of Nodes Degree Path Length Network Diameter Density 25 115 150 215 560

213 193 187 174 169

7.6 9.7 28.6 16.1 18.6

5.5 3.7 3.0 2.7 2.4

14 13 12 9 6

0.049 0.051 0.115 0.093 0.108

more in the early stages of ecosystem development from 25 to 150 years (Fig. 1A) which is consistent with results of other chronosequence studies (Schütte et al., 2009; Williams et al., 2013). After 150 years, bacterial community composition became less variable, corresponding to the colonization by coniferous trees, the decline in N mineralization rate and stabilization of edaphic factors such as soil pH (Blasko et al., 2015). The archaeal community did not differ significantly with age (Fig. 1B). Archaeal communities were varied more than bacteria and, although significant, edaphic factors were less correlated to composition (Fig. 1B), suggesting that archaeal composition is effected by unmeasured metrics. The 16S gene copy numbers decreased from the first ecosystem age and forward (Table 1), in line with Blasko et al. (2015) which reported a significant decrease in specific microbial biomass C and bacteria based on PLFA as ecosystems aged. This is likely due to higher bacterial € gberg et al., 2007b) and is likely real biomass at higher soil pH (Ho despite the potential for DNA to persist in soil over time (Dlott et al., 2015). Additionally, 16S gene copy numbers can vary within genomes (1e10 copies) (Acinas et al., 2004), and generally bacteria that contain more 16S gene copies genome1 are associated with higher growth rates (Klappenbach et al., 2000). This could be expected in environments with higher pH where less resources are needed to maintain homeostasis by Hþ excretion and could be allocated to growth (Sigler and Zeyer, 2004). One study observed that more fast-growing bacteria could be cultured from soils that had been deglaciated for 10 years compared to those deglaciated for 100 years, suggesting that young ecosystems may support organisms that have more copies of the 16S gene (Sigler and Zeyer, 2004). Species richness and diversity decreased during the first 115e150 years of development (Table 1). Decreased diversity appeared to be due to a loss of species as ecosystems age, with little evidence for recruitment of new OTUs in older soils (Fig. 3). In the case of Proteobacteria and Verrucomicrobia, there was no difference in relative abundance with ecosystem age, but abundance did differ at finer taxonomic resolution within these phyla (Fig. 2A). For example, the order Rhizobiales, often regarded as important in Nfixation, increased with the ecosystem age (Fig. 2A). The genera Rhizobium and Bradyrhizobium have been observed in association with ECM fungi (Barbieri et al., 2005) and were found in natural N€gberg et al., 2014). N-fixing rich and N-fertilized boreal forests (Ho plant species were not observed in our study, suggesting that these are free living non- or N-fixing species. Rather than being critical for N-fixation these forest ecotypes may be important in the

Fig. 4. Correlation networks for each ecosystem age were constructed using 15 soil samples, 3 transect sites x 5 cores taken at random intervals within the 10  15m sampling plot. Nodes represent OTUs (97% similarity) matching both bacteria and

archaea that were present in at least two transects within that ecosystem. Nodes that are darker in color are on average more abundant. Any two nodes displaying a significant positive or negative correlation are joined by an edge (p < 0.01).

166

€gberg / Soil Biology & Biochemistry 114 (2017) 160e167 S.A. Yarwood, M.N. Ho

breakdown of cyclic C compounds, a trend that makes logical sense given the increased SOM in later ecosystem ages within our study. Cyanobacteria were negatively correlated to ecosystem age (Fig. 2A). The 25-year-old soil contained 48 cyanobacterial OTUs; in contrast, only 6 of those OTUs were found at all ecosystem ages and no other ecosystem age had cyanobacteria unique to that developmental stage. The Leptolyngbya, an abundant genus of cyanobacteria, produce adhesive extra-cellular polysaccharides and organic acids that can degrade rock (Bellezza et al., 2003), an important function as the ecosystem ages. Another study showed that N-fixation by Nostoc cf. sphaericum in the tissue of feather mosses, Pleurozium schreberi, may represent a significant input of 1e3 kg N per ha and year in ecosystems where light is not limiting N-fixation rates and where plant available N is limiting (Deluca et al., 2002; Lindo et al., 2013; Stuiver et al., 2015). Sequences matching Nostocales were only observed in the exposed 25- and 115-year-old sites but were in low abundance in all samples. The family Isosphaeraceae, a group within the planctomycetes, were highest in relative abundance across all ecosystem ages and were most abundant in both the 215- and 560-year-old ecosystem. Members of this family have been identified in northern peat bogs and forest tundra. They seem to be able to withstand low pH conditions and primarily function in the breakdown of pectin, xylan, and licherin (Kulichevskaya et al., 2016). There were fewer trends in abundance differences among the archaea. MBGA, increased with age, and was observed to increase with elevation and decreased pH in Tibetan soils (Wang et al., 2015). Parvarchaeota YLA114 decreased with age (Fig. 2A), and has been proposed as a new phylum (Rinke et al., 2013), with the only sequenced members found to degrade both proteins and carbohydrates, and to exclusively be chemoheterotrophic (Li et al., 2015). The order YLA114 has been observed in marine systems and associated with oil reservoirs (Shelton et al., 2016). It is possible that these sequences represent organisms found mostly in marine systems that were present in the 25-year-old ecosystem, and decreased as the distance from the sea increased during land-uplift. The correlation networks clearly showed fewer nodes that were more significantly correlated in the 150-, 215-, and 560-year-old ecosystems compared to the two earlier locations, supporting hypothesis three. Many of the OTUs that were significantly correlated in the older ecosystem were also present in the younger ecosystems, but did not correlate as strongly. The strongest correlation in the 25-year-old ecosystem was between NC10 (Ettwig et al., 2016) and Methanomicrobia (r ¼ 0.81). The phylum NC10 (Ettwig et al., 2016), with no pure culture representatives, was observed in enrichment cultures from rice paddy soil to facilitate anaerobic methane oxidation (He et al., 2016). The class Methanomicrobia includes species capable of anaerobic methane oxidation and some members have been observed to exist in higher abundance in tundra soils (Mackelprang et al., 2011). Therefore, the correlation network identified the two organisms likely to interact in soil. Although sequences matching NC10 were detected in all the ecosystem ages, Methanomicrobia was only present in the 25- and 115-year-old ecosystems. The class Anaerolineae, phylum Chloroflexi, had the most connections of any node in the 560-year-old ecosystem. This group is ubiquitous in soil, water, and anaerobic digester, but of particular note they have been reported to degrade phenolic compounds (Fang et al., 2006). A wide range of bacteria were correlated to Anaerolineae, but no clearly identified function interaction was observed. Likely, many positive correlations are the result of biological or environmental filtering, but the correlation network cannot distinguish between these possibilities (Blüthgen, 2010). Sampling these locations in time, either seasonally or across years in future studies may help to clarify these relationships.

5. Conclusion Bacterial and archaeal communities changed in composition and structure as ecosystems aged. Species richness and diversity decreased as a result of loss of OTUs with only a handful of unique OTUs present in the older ecosystem and absent in the two younger ages. Multiple lineages of bacteria and archaea changed in abundance. Increased ecosystem age resulted in increased connectedness of the community up to 150 years, indicating that microbial communities may rapidly change and then stabilize within the first century of ecosystem development.

References Acinas, S.G., Marcelino, L.A., Klepac-Ceraj, V., Polz, M.F., 2004. Divergence and redundancy of 16S rRNA sequences in genomes with multiple rRNA operons. Journal of Bacteriology 186, 2629e2635. http://dx.doi.org/10.1128/ JB.186.9.2629-2635.2004. Aronesty, E., 2011. Command-line Tools for Processing Biological Sequencing Data. ea-utils. Expression Analysis, Durham, NC. http://code.google.com/p/ea-utils. Barbieri, E., Bertini, L., Rossi, I., Ceccaroli, P., Saltarelli, R., Guidi, C., Zambonelli, A., Stocchi, V., 2005. New evidence for bacterial diversity in the ascoma of the ectomycorrhizal fungus Tuber borchii Vittad. FEMS Microbiology Letters 247, 23e35. http://dx.doi.org/10.1016/j.femsle.2005.04.027. Bellezza, S., Paradossi, G., De Philippis, R., Albertano, P., 2003. Leptolyngbya strains from Roman hypogea: cytochemical and physico-chemical characterisation of exopolysaccharides. Journal of Applied Phycology 15, 193e200. € gberg, P., Ho €gberg, M.N., Blasko, R., Holm Bach, L., Yarwood, S.A., Trumbore, S.E., Ho 2015. Shifts in soil microbial community structure, nitrogen cycling and the concomitant declining N availability in ageing primary boreal forest ecosystems. Soil Biology and Biochemistry 91, 200e211. http://dx.doi.org/10.1016/ j.soilbio.2015.08.041. Blüthgen, N., 2010. Why network analysis is often disconnected from community ecology: a critique and an ecologist's guide. Basic and Applied Ecology 11, 185e195. http://dx.doi.org/10.1016/j.baae.2010.01.001. Bowen, J.L., Byrnes, J.E.K., Weisman, D., Colaneri, C., 2013. Functional gene pyrosequencing and network analysis: an approach to examine the response of denitrifying bacteria to increased nitrogen supply in salt marsh sediments. Frontiers in Microbiology 4. http://dx.doi.org/10.3389/fmicb.2013.00342. Caporaso, J.G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman, F.D., Costello, E.K., ~ a, A.G., Goodrich, J.K., Gordon, J.I., Huttley, G.A., Kelley, S.T., Fierer, N., Pen Knights, D., Koenig, J.E., Ley, R.E., Lozupone, C.A., McDonald, D., Muegge, B.D., Pirrung, M., Reeder, J., Sevinsky, J.R., Turnbaugh, P.J., Walters, W.A., Widmann, J., Yatsunenko, T., Zaneveld, J., Knight, R., 2010. QIIME allows analysis of highthroughput community sequencing data. Nature Methods 7, 335e336. http:// dx.doi.org/10.1038/nmeth.f.303. Caporaso, J.G., Lauber, C.L., Walters, W.A., Berg-Lyons, D., Huntley, J., Fierer, N., Owens, S., Betley, J., Fraser, L., Bauer, M., Gormley, N., Gilbert, J.A., Smith, G., Knight, R., 2012. Ultra-high-throughput microbial community analysis on the Illumina HiSeq and MiSeq platforms. ISME Journal-International Society for Microbial Ecology 6, 1621e1624. Chagnon, P.-L., Bradley, R.L., Maherali, H., Klironomos, J.N., 2013. A trait-based framework to understand life history of mycorrhizal fungi. Trends in Plant Science 18, 484e491. http://dx.doi.org/10.1016/j.tplants.2013.05.001. Cline, L.C., Zak, D.R., 2015. Soil microbial communities are shaped by plant-driven changes in resource availability during secondary succession. Ecology 96, 3374e3385. Deluca, T.H., Zackrisson, O., Nilsson, Maria-Charlotte, G., Sellstedt, S.J., 2002. Quantifying nitrogen-fixation in feather moss carpets of boreal forestsn. Nature 419, 917e919. http://dx.doi.org/10.1038/nature01136. Dighton, J., White, J.F., 2017. The Fungal Community: its Organization and Role in the Ecosystem, fourth ed. Dlott, G., Maul, J.E., Buyer, J., Yarwood, S., 2015. Microbial rRNA:rDNA gene ratios may be unexpectedly low due to extracellular DNA preservation in soils. Journal of Microbiological Methods 115, 112e120. http://dx.doi.org/10.1016/ j.mimet.2015.05.027. Edgar, R., 2010. Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26, 2460e2461. Ekman, M., 1996. A consistent map of the postglacial uplift of Fennoscandia. Terra Nova 8, 158e165. Ettwig, K.F., Zhu, B., Speth, D., Keltjens, J.T., Jetten, M.S.M., Kartal, B., 2016. Archaea catalyze iron-dependent anaerobic oxidation of methane. Proceedings of the National Academy of Sciences 113, 12792e12796. http://dx.doi.org/10.1073/ pnas.1609534113. Fang, H.H.P., Liang, D.W., Zhang, T., Liu, Y., 2006. Anaerobic treatment of phenol in wastewater under thermophilic condition. Water Research 40, 427e434. http:// dx.doi.org/10.1016/j.watres.2005.11.025. Faust, K., Raes, J., 2012. Microbial interactions: from networks to models. Nature Reviews Microbiology 10, 538e550. http://dx.doi.org/10.1038/nrmicro2832. Fierer, N., Nemergut, D., Knight, R., Craine, J.M., 2010. Changes through time:

€gberg / Soil Biology & Biochemistry 114 (2017) 160e167 S.A. Yarwood, M.N. Ho integrating microorganisms into the study of succession. Research in Microbiology 161, 635e642. http://dx.doi.org/10.1016/j.resmic.2010.06.002. Fruchterman, T.M., Reingold, E.M., 1991. Graph drawing by force-directed placement. Software: Practice and Experience 21, 1129e1164. Hargreaves, S.K., Roberto, A.A., Hofmockel, K.S., 2013. Reaction- and sample-specific inhibition affect standardization of qPCR assays of soil bacterial communities. Soil Biology and Biochemistry 59, 89e97. http://dx.doi.org/10.1016/ j.soilbio.2013.01.007. Hartmann, M., Niklaus, P.A., Zimmermann, S., Schmutz, S., Kremer, J., Abarenkov, K., Lüscher, P., Widmer, F., Frey, B., 2014. Resistance and resilience of the forest soil microbiome to logging-associated compaction. The ISME Journal 8, 226e244. He, Z., Cai, C., Wang, J., Xu, X., Zheng, P., Jetten, M.S.M., Hu, B., 2016. A novel denitrifying methanotroph of the NC10 phylum and its microcolony. Scientific Reports 6, 32241. http://dx.doi.org/10.1038/srep32241. €gberg, M.N., Chen, Y., Ho €gberg, P., 2007a. Gross nitrogen mineralisation and Ho fungi-to-bacteria ratios are negatively correlated in boreal forests. Biology and Fertility of Soils 44, 363e366. http://dx.doi.org/10.1007/s00374-007-0215-9. €gberg, M.N., Ho € gberg, P., Myrold, D.D., 2007b. Is microbial community compoHo sition in boreal forest soils determined by pH, C-to-N ratio, the trees, or all three? Oecologia 150, 590e601. http://dx.doi.org/10.1007/s00442-006-0562-5. €gberg, M.N., Yarwood, S.A., Myrold, D.D., 2014. Fungal but not bacterial soil Ho communities recover after termination of decadal nitrogen additions to boreal forest. Soil Biology and Biochemistry 72, 35e43. http://dx.doi.org/10.1016/ j.soilbio.2014.01.014. Hoppe, B., Kahl, T., Karasch, P., Wubet, T., Bauhus, J., Buscot, F., Krüger, D., 2014. Network analysis reveals ecological links between N-Fixing bacteria and wood-decaying fungi. PLoS One 9, e88141. http://dx.doi.org/10.1371/ journal.pone.0088141. Huss-Danell, K., 1997. Actinorhizal symbioses and their N2 fixation. New Phytol. 136, 375e405. Johnson, N.C., Zak, D.R., Tilman, D., Pfleger, F.L., 1991. Dynamics of vesiculararbuscular mycorrhizae during old field succession. Oecologia 86, 349e358. Jumpponen, A., Egerton-Warburton, L.M., 2005. Mycorrhizal fungi in successional environments: a community assembly model incorporating host plant, environmental, and biotic filters. Mycology Series 23, 139. Klappenbach, J.A., Dunbar, J.M., Schmidt, T.M., 2000. rRNA Operon copy number reflects ecological strategies of bacteria. Applied and Environmental Microbiology 66, 1328e1333. http://dx.doi.org/10.1128/AEM.66.4.1328-1333.2000. , J.S., Kulichevskaya, I.S., Rijpstra, W.I.C., Dedysh, S.N., Suzina, N.E., Sinninghe Damste Ivanova, A.A., 2016. Paludisphaera borealis gen. nov., sp. nov., a hydrolytic planctomycete from northern wetlands, and proposal of Isosphaeraceae fam. nov. International Journal of Systematic and Evolutionary Microbiology 66, 837e844. http://dx.doi.org/10.1099/ijsem.0.000799. Larmola, T., Leppanen, S.M., Tuittila, E.-S., Aarva, M., Merila, P., Fritze, H., Tiirola, M., 2014. Methanotrophy induces nitrogen fixation during peatland development. Proceedings of the National Academy of Sciences 111, 734e739. http:// dx.doi.org/10.1073/pnas.1314284111. Li, M., Baker, B.J., Anantharaman, K., Jain, S., Breier, J.A., Dick, G.J., 2015. Genomic and transcriptomic evidence for scavenging of diverse organic compounds by widespread deep-sea archaea. Nature Communications 6, 8933. http:// dx.doi.org/10.1038/ncomms9933. Lindo, Z., Nilsson, M.-C., Gundale, M.J., 2013. Bryophyte-cyanobacteria associations as regulators of the northern latitude carbon balance in response to global change. Global Change Biology 19, 2022e2035. http://dx.doi.org/10.1111/ gcb.12175. Mackelprang, R., Waldrop, M.P., DeAngelis, K.M., David, M.M., Chavarria, K.L., Blazewicz, S.J., Rubin, E.M., Jansson, J.K., 2011. Metagenomic analysis of a permafrost microbial community reveals a rapid response to thaw. Nature 480, 368e371. http://dx.doi.org/10.1038/nature10576. McCune, B., Grace, J.B., 2002. Analysis of Ecological Communities, second ed. MjM Software Design, Gleneden Beach, OR. McDonald, D., Price, M., Goodrich, J.K., Nawrocki, E., DeSantis, T.Z., Probst, A.J., Andersen, G.L., Knight, R., Hugenholtz, P., 2012. An improved Greengenes taxonomy with explicit ranks for ecological and evolutionary analyses of bacteria and archaea. ISME Journal-International Society for Microbial Ecology 6,

167

610e618. €, K., Meril€ a, P., Galand, P.E., Fritze, H., Tuittila, E.-S., Kukko-oja, K., Laine, J., Yrj€ ala 2006. Methanogen communities along a primary succession transect of mire ecosystems: methane and methanogens in mire ecosystems. FEMS Microbiology Ecology 55, 221e229. http://dx.doi.org/10.1111/j.1574-6941.2005.00030.x. €msa €, M., Spetz, P., Stark, S., Vierikko, K., Derome, J., Meril€ a, P., Malmivaara-La Fritze, H., 2010. Soil organic matter quality as a link between microbial community structure and vegetation composition along a successional gradient in a boreal forest. Applied Soil Ecology 46, 259e267. http://dx.doi.org/10.1016/ j.apsoil.2010.08.003. €mmer, R., Fritze, H., 2002. Soil microbial activity and community Meril€ a, P., Stro structure along a primary succession transect on the land-uplift coast in western Finland. Soil Biology and Biochemistry 34, 1647e1654. €sholm, T., Ho €gberg, P., Franklin, O., Metcalfe, D., Keel, S.G., Campbell, C., Hurry, V., Na € gberg, M.N., 2013. Are ectomycorrhizal fungi alleviating or Linder, S., Ho aggravating nitrogen limitation of tree growth in boreal forests? New Phytologist 198, 214e221. http://dx.doi.org/10.1111/nph.12139. Rinke, C., Schwientek, P., Sczyrba, A., Ivanova, N.N., Anderson, I.J., Cheng, J.-F., Darling, A., Malfatti, S., Swan, B.K., Gies, E.A., Dodsworth, J.A., Hedlund, B.P., Tsiamis, G., Sievert, S.M., Liu, W.-T., Eisen, J.A., Hallam, S.J., Kyrpides, N.C., Stepanauskas, R., Rubin, E.M., Hugenholtz, P., Woyke, T., 2013. Insights into the phylogeny and coding potential of microbial dark matter. Nature 499, 431e437. http://dx.doi.org/10.1038/nature12352. Schütte, U.M.E., Abdo, Z., Bent, S.J., Williams, C.J., Schneider, G.M., Solheim, B., Forney, L.J., 2009. Bacterial succession in a glacier foreland of the High Arctic. The ISME Journal 3, 1258e1268. http://dx.doi.org/10.1038/ismej.2009.71. Shelton, J.L., Akob, D.M., McIntosh, J.C., Fierer, N., Spear, J.R., Warwick, P.D., McCray, J.E., 2016. Environmental drivers of differences in microbial community structure in crude oil reservoirs across a methanogenic gradient. Frontiers in Microbiology 7. http://dx.doi.org/10.3389/fmicb.2016.01535. Shi, S., Nuccio, E.E., Shi, Z.J., He, Z., Zhou, J., Firestone, M.K., 2016. The interconnected rhizosphere: high network complexity dominates rhizosphere assemblages. Ecology Letters 19, 926e936. http://dx.doi.org/10.1111/ele.12630. Sigler, W.V., Zeyer, J., 2004. Colony-forming analysis of bacterial community succession in deglaciated soils indicates pioneer stress-tolerant opportunists. Microbial Ecology 48, 316e323. http://dx.doi.org/10.1007/s00248-003-0189-6. Stuiver, B.M., Gundale, M.J., Wardle, D.A., Nilsson, M.-C., 2015. Nitrogen fixation rates associated with the feather mosses Pleurozium schreberi and Hylocomium splendens during forest stand development following clear-cutting. Forest Ecology and Management 347, 130e139. http://dx.doi.org/10.1016/ j.foreco.2015.03.017. Svensson, J.S., Jeglum, J.K., 2003. Primary succession pathway of Norway spruce communities on land-uplift seashores. Ecoscience 10, 96e109. Vestøl, O., 2006. Determination of postglacial land uplift in Fennoscandia from leveling, tide-gauges and continuous GPS stations using least squares collocation. Journal of Geodesy 80, 248e258. € bner, G., Persson, P., Schleucher, J., Giesler, R., 2013. Vincent, A.G., Vestergren, J., Gro Soil organic phosphorus transformations in a boreal forest chronosequence. Plant and Soil 367, 149e162. http://dx.doi.org/10.1007/s11104-013-1731-z. €rth, C.-M., Giesler, R., 2009. Increasing abundance of soil fungi is a Wallander, H., Mo driver for 15N enrichment in soil profiles along a chronosequence undergoing isostatic rebound in northern Sweden. Oecologia 160, 87e96. http://dx.doi.org/ 10.1007/s00442-008-1270-0. Wang, J.-T., Cao, P., Hu, H.-W., Li, J., Han, L.-L., Zhang, L.-M., Zheng, Y.-M., He, J.-Z., 2015. Altitudinal distribution patterns of soil bacterial and archaeal communities along Mt. Shegyla on the Tibetan Plateau. Microbial Ecology 69, 135e145. http://dx.doi.org/10.1007/s00248-014-0465-7. Williams, M.A., Jangid, K., Shanmugam, S.G., Whitman, W.B., 2013. Bacterial communities in soil mimic patterns of vegetative succession and ecosystem climax but are resilient to change between seasons. Soil Biology and Biochemistry 57, 749e757. http://dx.doi.org/10.1016/j.soilbio.2012.08.023. € ransson, H., Smittenberg, R.H., Brunner, I., Bernasconi, S.M., Zumsteg, A., Luster, J., Go Zeyer, J., Frey, B., 2012. Bacterial, archaeal and fungal succession in the forefield of a receding glacier. Microbial Ecology 63, 552e564. http://dx.doi.org/10.1007/ s00248-011-9991-8.