Successional Variation in the Soil Microbial Community in Odaesan National Park, Korea

: Succession is deﬁned as variation in ecological communities caused by environmental changes. Environmental succession can be caused by rapid environmental changes, but in many cases, it is slowly caused by climate change or constant low-intensity disturbances. Odaesan National Park is a well-preserved forest located in the Taebaek mountain range in South Korea. The forest in this national park is progressing from a mixed-wood forest to a broad-leaved forest. In this study, microbial community composition was investigated using 454 sequencing of soil samples collected from 13 di ﬀ erent locations in Odaesan National Park. We assessed whether microbial communities are a ﬀ ected by changes in environmental factors such as water content (WC), nutrient availability (total carbon (TC) and total nitrogen (TN)) and pH caused by forest succession. WC, TC, TN and pH signiﬁcantly di ﬀ ered between the successional stages of the forest. The WC, TC and TN of the forest soils tended to increase as succession progressed, while pH tended to decrease. In both successional stages, the bacterial genus Pseudolabrys was the most abundant, followed by Aﬁpia and Bradyrhizobium . In addition, the fungal genus Saitozyma showed the highest abundance in the forest soils. Microbial community composition changed according to forest successional stage and soil properties (WC, TC, TN, and pH). Furthermore, network analysis of both bacterial and fungal taxa revealed strong relationships of the microbial community depending on the soil properties a ﬀ ected by forest succession.


Introduction
Recently, research on microbial diversity has progressed actively due to the development of high-throughput sequencing (HTS) technology, and the accumulation of many research results has increased the understanding of the microbial community and the sustainability of microbial resources for industrial and commercial applications [1][2][3]. Over the last decade, studies on the microbial community have focused on measuring the diversity in a given study area and comparing composition and structure between environments [4][5][6]. In particular, microbial communities of forest soils have been used to investigate the microbial variation caused by vegetation type and altitude and the successional Sustainability 2020, 12, 4795 3 of 13

Description of the Study Sites
This study was conducted in Odaesan National Park, South Korea (37 • 42 -52 N 128 • 28 -46 E). This park was designated in 1975 as the 11th national park of Korea. The major peaks of Mt. Odae include Birobong (1563 m), Dongdaesan (1434 m), Durobong (1422 m), and Sangwangbong (1493 m). The mean annual temperature is 6.5 • C, with a monthly mean minimum temperature of −13.8 • C (December) and monthly mean maximum temperature of 25 • C (July) in 2013. The annual precipitation is 1288.7 mm, which is mainly concentrated in summer. All meteorological data were collected by the Korea Meteorological Administration from the Daegwallyeong station (http: //www.kma.go.kr/weather/climate/data_monthly.jsp).

Soil Sampling
Soil samples were collected at each stage of succession from 13 different points selected based on elevation and the distance between locations ( Figure 1 and Table S1). Samples were collected twice. The first sampling was on 22 April 2013, and the second sampling was on 3 September 2013. Topsoil (~10 cm depth) was collected three times using sterile 50-mL conical tubes in a triangular pattern with a 1-m radius at each location. The three soil samples collected from each location were combined and sieved together using 2-mm sieves and frozen at -80 • C until use.

Description of the Study Sites
This study was conducted in Odaesan National Park, South Korea (37°42′-52′ N 128°28′-46′ E). This park was designated in 1975 as the 11th national park of Korea. The major peaks of Mt. Odae include Birobong (1563 m), Dongdaesan (1434 m), Durobong (1422 m), and Sangwangbong (1493 m). The mean annual temperature is 6.5 °C, with a monthly mean minimum temperature of −13.8 °C (December) and monthly mean maximum temperature of 25 °C (July) in 2013. The annual precipitation is 1288.7 mm, which is mainly concentrated in summer. All meteorological data were collected by the Korea Meteorological Administration from the Daegwallyeong station (http://www.kma.go.kr/weather/climate/data_monthly.jsp).

Soil Sampling
Soil samples were collected at each stage of succession from 13 different points selected based on elevation and the distance between locations ( Figure 1 and Table S1). Samples were collected twice. The first sampling was on 22 April 2013, and the second sampling was on 3 September 2013. Topsoil (~10 cm depth) was collected three times using sterile 50-mL conical tubes in a triangular pattern with a 1-m radius at each location. The three soil samples collected from each location were combined and sieved together using 2-mm sieves and frozen at -80 °C until use.

Analysis of Soil Properties
Each soil sample was dried in an oven at 105 °C for 1 day, and the soil water content (WC) was measured. Soil pH was measured in a 1:5 dilution of soil:distilled water using an Orion 3 Star pH meter equipped with an Orion 8157 BNUMD probe (Thermo Scientific, Beverly, MA, USA). Soil total carbon (TC) and total nitrogen (TN) contents were measured by a Flash EA 1112 elemental analyzer (Thermo Elemental, Waltham, MA, USA). Differences in soil properties between successional stages

Analysis of Soil Properties
Each soil sample was dried in an oven at 105 • C for 1 day, and the soil water content (WC) was measured. Soil pH was measured in a 1:5 dilution of soil:distilled water using an Orion 3 Star pH meter equipped with an Orion 8157 BNUMD probe (Thermo Scientific, Beverly, MA, USA). Soil total carbon (TC) and total nitrogen (TN) contents were measured by a Flash EA 1112 elemental analyzer Sustainability 2020, 12, 4795 4 of 13 (Thermo Elemental, Waltham, MA, USA). Differences in soil properties between successional stages were assessed using the Wilcoxon rank sum test with multiple test correction by the false discovery rate (FDR) of Benjamini and Hochberg [31].

DNA Library Preparation and 454 Pyrosequencing
DNA extraction was performed with 0.25 g of each soil sample using a PowerSoil DNA isolation kit (MoBio, Carlsbad, CA, USA) according to the MV method [32]. A DNA library was prepared, and pyrosequencing was performed by Macrogen Ltd. (Seoul, Korea) according to Jang et al. [33]. The FastStart High-Fidelity PCR System (Roche, Germany) was used with 20 ng of DNA, 1 µL of each forward and reverse primer (10 µM), 0.5 µL of dNTP mix (10 mM each), 2.5 µL of FastStart 10× buffer #2, 0.25 µL of FastStart Hifi Polymerase (5 U µL−1), and molecular biology-grade water in a 25-µL reaction. The 16S rRNA gene (specifically, the V1 to V3 region) and the internal transcribed spacer 2 (ITS2) region were amplified with the primers 27F (5 -AGAGTTTGATCCTGGCTCAG-3 )/534R (5 -ATTACCGCGGCTGCTGG-3 ) [34] and ITS3_KYO2 (5 -A-GATGAAGAACGYAGYRAA-3 )/ITS4 (5 -B-TCCTCCGCTTATTGATATGC-3 ) [35] for bacterial and fungal community analysis, respectively. The thermocycler settings for bacterial amplification were as follows: initial denaturation at 95 • C for 2 min, 30 cycles of 95 • C for 20 s, 56 • C for 30 s, and 72 • C for 60 s and final elongation at 72 • C for 5 min. PCR for fungal amplication was performed with the following conditions: an initial denaturation step of 3 min at 94 • C, followed by 35 cycles at 94 • C for 15 s, 55 • C for 45 s, and 72 • C for 1 min. The reaction was concluded by a final elongation step at 72 • C for 8 min. PCR purification was performed with Agencourt AMPure XP (Beckman Coulter, Miami, FL, USA) according to the manufacturer's manual and sent for sequencing (Macrogen Inc., Seoul, Korea). 454 pyrosequencing was performed using the GS FLX Titanium platform (454 Life Sciences, Branford, CT, USA).

Bioinformatics Analysis
The sequences obtained in this study were processed using QIIME v1.8.0 [36]. The primer, key, and barcode sequences were trimmed from both ends. Low-quality (QV < 25, ≥1 ambiguous base, or >6 homopolymers), short (<200 nt), and chimeric sequences were removed, leaving the sequences for analysis. Then, the sequences were clustered into operational taxonomic units (OTUs) based on a ≥97% similarity threshold and the average linkage method using Usearch 5.2.236. The representative sequence that was most abundant for each OTU was taxonomically assigned using the UNITE v7.2 (16.11.2017) database [37] for fungi and EzBioCloud [38] for bacteria. Rarefaction curves evaluating the sufficient sequence depth were made using QIIME, determined the depth for bacterial (2100) and fungal (5600) sequences. Alpha diversity indices were calculated in QIIME and compared between successional stages using Shapiro-Wilk test of normality, t-test, or Wilcoxon test in R. Ordination analysis of community structure was conducted based on Bray-Curtis dissimilarities for fungi and weighted UniFrac dissimilarities for bacteria. The analysis of similarities (ANOSIM) between the successional stages was performed using 9999 permutations within the sampling times (April and September) in R. Non-metric multidimensional scaling (NMDS) and environmental fitting were conducted with the environmental variables using the vegan package in R [39]. Network analysis was performed to identify the bacterial and fungal co-occurrence patterns across the successional stages. For microbial network analysis, correlations between major genera and environmental factors were tested using Spearman's rank correlation coefficient in R. Co-occurence patterns were tested based on Spearman's correlation. The correlation matrix was calculated according to the abundance of genus pairs, and significant correlations (r ≥ 0.7, padj < 0.05) were retained after multiple-test correction with the FDR of Benjamini and Hochberg [31]. Network plots were drawn using the igraph package [40] in R.

Soil Properties
Forest succession caused marked variability in soil properties between broad-leaved-wood (late stage) and mixed-wood (early stage) forests ( Figure 2). The WC, TC, TN, and pH of Odaesan forest soil exhibited successional shifts ( Figure 2). The effect of forest succession on soil properties was similar to the effects on WC, TC, and TN. The soil property values increased in the broad-leaved forest, and conversely, the pH decreased. The organic matter quality, explained by the average C/N ratio, was higher (more than 15%) in the mixed-wood forest. Forest succession decreased soil acidity by up to 1.6 units compared to that in the mixed-wood forest.
Sustainability 2020, 12, x FOR PEER REVIEW 5 of 13 soil exhibited successional shifts ( Figure 2). The effect of forest succession on soil properties was similar to the effects on WC, TC, and TN. The soil property values increased in the broad-leaved forest, and conversely, the pH decreased. The organic matter quality, explained by the average C/N ratio, was higher (more than 15%) in the mixed-wood forest. Forest succession decreased soil acidity by up to 1.6 units compared to that in the mixed-wood forest.

Microbial Composition
At the taxonomic level, 75,796 16S rRNA gene sequences were obtained from soil samples of both forest types. The number of reads varied from 2123 to 3980 per sample, resulting in a total of 2593 OTUs at 97% similarity. Bacterial community analysis showed that the most abundant phylum was Proteobacteria and accounted 37 and 41% of total abundances in the forest soils of early and late stage of succession, respectively ( Figure 3). The relative abundances of the Pseudolabrys, Afipia, and Bradyrhizobium genera belonging to Alphaproteobacteria were greater than 25% in both types of forest soils at Odaesan National Park ( Figure 3a). The relative abundances of the Pseudolabrys, Acidibacter, Koribacter, and Conexibacter genera increased with the progression of forest succession. In contrast, the abundance of the Mycobacterium and Pedomicrobium genera in the early stage of succession were higher than those in the late stage ( Figure S1).
At the taxonomic level, 201,037 ITS sequences were obtained from 26 samples of both forest types. The number of reads varied from 5638 to 11,604 per sample, resulting in a total of 2448 OTUs at 97% similarity. Fungal community analysis showed that the most abundant phylum was Basidiomycota and accounted 60 and 66% in the forest soils of early and late stage of succession, respectively ( Figure 3). Ascomycota exhibited an average abundance of approximately 15.2%, followed by Zygomycota. The total abundances of these three major phyla accounted for 82 and 85% of the total in early and late stage, respectively. The sum of the relative abundances of the Saitozyma, Solicoccozyma, Trichocladium and Mortierella genera was greater than 65% in both types of forest (Figure 3b). In particular, the Saitozyma genus showed a relative abundance greater than 30%. The relative abundances of the Solicoccozyma, and Gyoerffyella genera increased while the abundances of the Schizopora and Exophiala genera decreased with forest succession.

Microbial Composition
At the taxonomic level, 75,796 16S rRNA gene sequences were obtained from soil samples of both forest types. The number of reads varied from 2123 to 3980 per sample, resulting in a total of 2593 OTUs at 97% similarity. Bacterial community analysis showed that the most abundant phylum was Proteobacteria and accounted 37 and 41% of total abundances in the forest soils of early and late stage of succession, respectively ( Figure 3). The relative abundances of the Pseudolabrys, Afipia, and Bradyrhizobium genera belonging to Alphaproteobacteria were greater than 25% in both types of forest soils at Odaesan National Park (Figure 3a). The relative abundances of the Pseudolabrys, Acidibacter, Koribacter, and Conexibacter genera increased with the progression of forest succession. In contrast, the abundance of the Mycobacterium and Pedomicrobium genera in the early stage of succession were higher than those in the late stage ( Figure S1).
At the taxonomic level, 201,037 ITS sequences were obtained from 26 samples of both forest types. The number of reads varied from 5638 to 11,604 per sample, resulting in a total of 2448 OTUs at 97% similarity. Fungal community analysis showed that the most abundant phylum was Basidiomycota and accounted 60 and 66% in the forest soils of early and late stage of succession, respectively ( Figure 3). Ascomycota exhibited an average abundance of approximately 15.2%, followed by Zygomycota. The total abundances of these three major phyla accounted for 82 and 85% of the total in early and late stage, respectively. The sum of the relative abundances of the Saitozyma, Solicoccozyma, Trichocladium and Mortierella genera was greater than 65% in both types of forest (Figure 3b). In particular, the Saitozyma genus showed a relative abundance greater than 30%. The relative abundances of the Solicoccozyma, and Gyoerffyella genera increased while the abundances of the Schizopora and Exophiala genera decreased with forest succession.

Microbial Diversity and Its Relationships with Environmental Variables
There was no statistically significant difference in the richness (Chao1; p = 0.458), diversity (Shannon index; p = 0.126), or dominance (Simpson index; p = 0.521) of the bacterial community between the stages of forest succession (Table S1). In contrast to that in the bacterial community, the number of observed OTUs in the fungal community was higher (p = 0.045) in the early stage of succession (Table S1). There was no significant difference in richness (p = 0.068), diversity (p = 0.154), or dominance (p = 0.21) for the fungal community. By analyzing the beta diversity of each bacterial community through ordination analysis based on weighted UniFrac dissimilarities, we detected a distinct shift in the bacterial community as succession proceeded by ANOSIM test (R = 0.6143; p = 0.0001), and environmental fitting analysis with NMDS revealed that elevation, WC, TC, TN, and pH were significant factors distinguishing the bacterial communities ( Figure 4a and Table S2). Among the bacterial genera, Edaphobacter and Pedomicrobium were significantly related to pH, while Reyranella was significantly related to TC (Figure 5a).When the beta diversity of the fungal community was analyzed by ANOSIM, the distribution of fungal communities was also clearly dependent on the stage of succession (R = 0.1314; p = 0.023) (Figure 4b). In addition, envfit analysis revealed significant influences of soil properties on the fungal community (Table S2). These results were similar to those for the bacterial community. Among the fungal genera, Mortierella showed significant relationships with elevation and WC, and Gyoerffyella showed a significant relationship with WC (Figure 5b).

Microbial Diversity and Its Relationships with Environmental Variables
There was no statistically significant difference in the richness (Chao1; p = 0.458), diversity (Shannon index; p = 0.126), or dominance (Simpson index; p = 0.521) of the bacterial community between the stages of forest succession (Table S1). In contrast to that in the bacterial community, the number of observed OTUs in the fungal community was higher (p = 0.045) in the early stage of succession (Table S1). There was no significant difference in richness (p = 0.068), diversity (p = 0.154), or dominance (p = 0.21) for the fungal community. By analyzing the beta diversity of each bacterial community through ordination analysis based on weighted UniFrac dissimilarities, we detected a distinct shift in the bacterial community as succession proceeded by ANOSIM test (R = 0.6143; p = 0.0001), and environmental fitting analysis with NMDS revealed that elevation, WC, TC, TN, and pH were significant factors distinguishing the bacterial communities ( Figure 4a and Table S2). Among the bacterial genera, Edaphobacter and Pedomicrobium were significantly related to pH, while Reyranella was significantly related to TC (Figure 5a).When the beta diversity of the fungal community was analyzed by ANOSIM, the distribution of fungal communities was also clearly dependent on the stage of succession (R = 0.1314; p = 0.023) (Figure 4b). In addition, envfit analysis revealed significant influences of soil properties on the fungal community (Table S2). These results were similar to those for the bacterial community. Among the fungal genera, Mortierella showed significant relationships with elevation and WC, and Gyoerffyella showed a significant relationship with WC ( Figure 5b). was significantly related to TC (Figure 5a).When the beta diversity of the fungal community was analyzed by ANOSIM, the distribution of fungal communities was also clearly dependent on the stage of succession (R = 0.1314; p = 0.023) (Figure 4b). In addition, envfit analysis revealed significant influences of soil properties on the fungal community (Table S2). These results were similar to those for the bacterial community. Among the fungal genera, Mortierella showed significant relationships with elevation and WC, and Gyoerffyella showed a significant relationship with WC ( Figure 5b).

Microbial Networks
As shown in Figure 6, the microbial networks were determined based on correlations between microbial genus abundances at each successional stage. In the early successional stage, the network consisted of four groups including hybrid or separate compartments of bacterial and fungal taxa. In particular, Mortierella spp. formed a network with four fungal genera, and Roseiarcus spp. formed a network with three taxa. In the late stage, the microbial network became simpler. A total of four groups of networks were constructed, but each group consisted of only two microbial genera.

Microbial Networks
As shown in Figure 6, the microbial networks were determined based on correlations between microbial genus abundances at each successional stage. In the early successional stage, the network consisted of four groups including hybrid or separate compartments of bacterial and fungal taxa. In particular, Mortierella spp. formed a network with four fungal genera, and Roseiarcus spp. formed a network with three taxa. In the late stage, the microbial network became simpler. A total of four groups of networks were constructed, but each group consisted of only two microbial genera.
As shown in Figure 6, the microbial networks were determined based on correlations between microbial genus abundances at each successional stage. In the early successional stage, the network consisted of four groups including hybrid or separate compartments of bacterial and fungal taxa. In particular, Mortierella spp. formed a network with four fungal genera, and Roseiarcus spp. formed a network with three taxa. In the late stage, the microbial network became simpler. A total of four groups of networks were constructed, but each group consisted of only two microbial genera.

Discussion
The bacterial and fungal DNA extracted from forest soil was analyzed to investigate the differences in the microbial community caused by forest succession in Odaesan National Park. In general, the microbial community responds more quickly to environmental disturbances causing ecological succession than to plants, animals, and insects [41,42]. Because the phylogenetic diversity of microorganisms is high and is closely associated with geochemical properties, it is important to monitor and analyze the microbial community as a sensitive index during forest succession [43]. Therefore, in this study, variation in forest soil and the microbial community according to forest succession was analyzed, and we focused on the relationships between soil properties and the microbial community.
With succession, the vegetation in Odaesan National Park changed from P. densiflora to Q. mongolica and Q. variabilis, exhibiting forest succession from an early stage (mixed-wood forest) to a late stage (broad-leaved forest) [28]. Forest succession from the early stage to the late stage also greatly changed the properties of the forest soil ( Figure 2). First, the average WC of the soil was higher in the late stage forest. This result could be explained by late-stage wood having a higher leaf area index than mixed-wood, which was able to block sunlight and shade the surface of soil resulting in wetter soil conditions, but also be caused by differences in altitude, since late-stage forests were located in higher altitudes [44,45].
Forest succession generally leads to soil improvement and increased TC and TN contents [46]. In this study, the TC and TN concentrations in forest soils were higher in the late stage than in the early stage ( Figure 2). As the forest develops, the microbial community adapts to the substrate and decomposes leaves and branches to increase the nutrient concentration in the soil [47]. When the nitrogen and carbon availability in the soil increases, inorganic nitrogen is absorbed by the plants, and the nitrogen concentration in the leaves also increases so that when the leaves are dropped, they again become a useful substrate for the microorganisms and increase nutrient availability [48,49]. Additionally, as the microbial community develops, the carbon content derived from microbial biomass also increases, thereby increasing the TC content of the soil [48]. However, the alpha diversity of the microbial communities in this study did not show any difference in richness according to successional stage (Table S1). Another reason for the high TC and TN concentrations in late-stage forests may be the high moisture content. Soil moisture affects not only soil structure but also the transport of substances and nutrients and microbial activity [50,51]. In soil with low water solubility, microorganismal cells have a low WC, which can prevent microbial enzymes from binding to the substrate and thus decrease microbial activity [50]. As a result, the degradation of organic matter in the soil is restricted, decreasing Sustainability 2020, 12, 4795 9 of 13 the soil nutrient concentration. In contrast, the higher WC at the late stage may have contributed to the high TC and TN contents by increasing nutrient circulation and the activities of the enzymes produced by microorganisms [52]. On the other hand, the C/N ratio of the forest soil in Odaesan National Park ranged from 9 to 15, and the ratio in the soil samples from the late successional stage was slightly lower than that in the soil samples from the early stage. Generally, if the C/N ratio is less than 15, organic matter is mineralized, and nitrogen can be used effectively [53,54]. The relatively low C/N ratio in late stage forest soils may be related to the type of vegetation. At the early stage of succession, the litter produced contained a lignified and aromatic substrate with a high C/N ratio [55]. Thus, the C/N ratio of the soil was also considered to be high. As forest succession progresses, overall plant biomass increases, and the increased root area causes a higher WC in forest soil and a greater abundance of symbiotic microorganisms [46]. Soils in more developed forests have higher TC and TN concentrations due to the quantitative increase in microbial biomass and the development of aerial parts of trees [46]. Although our results did not show any increase in the richness of the microbial community in the developed forest (Table S1), the microbial community may have contributed to the degradation and circulation of organic matter, thereby increasing the TC and TN concentrations.
The structure and composition of the microbial community differed with succession of the forest from mixed-wood forest to broad-leaved forest (Figures 3 and 4). According to forest successional stage, the microbial communities showed different compositions and structures, which were well described by NMDS analysis, revealing a difference in the community composition of the microbial communities between the two successional stages (Figure 4). These results effectively support the fundamental concept that soil and vegetation properties can induce a variety of microhabitats, allowing various microorganisms to coexist [56]. Forest succession causes a change in vegetation, which in turn causes changes in the soil in which the vegetation grows and alters the microbial community [57]. The plant community, which is altered by succession, exhibits regional differences (root depth, bark, leaf shade area, and tree density) and ultimately has a significant impact on microbial community structure [58]. Soil WC, TC, TN, C/N ratio, and pH changed in response to succession, and the bacterial and fungal taxa that were correlated with soil properties differed in abundance according to successional stage, eventually altering microbial community structure. Interestingly, most taxa that showed different relative abundances as succession progressed exhibited no significant correlations with soil properties. This result indicates that other environmental factors determined the microbial habitat and affected the shift in the microbial community. The most prominent factor was the interaction between trees and the shifted taxa. Among the dominant taxa in each stage, the genera whose abundance was significantly changed by succession were Pseudolabrys and Solicoccozyma, both of which showed a high abundance in the late stage. The genus Pseudolabrys is mainly found in soil and belongs to the Rhizobiales, which fix nitrogen and are symbiotic with plant roots [59,60]. Solicoccozyma is a genus of basidiomycetous yeast that is mainly found in soil and frequently produces thick polysaccharide capsules and low-weight aromatic compounds [60]. Although none of these genus has been reported to have a symbiotic relationship with Quercus spp., they are expected to play a major role in the carbon and nitrogen cycles, and their significantly higher abundance in the late stage can serve as a bioindicator for monitoring microbial succession.
We also investigated microbial networks to estimate the effects of forest succession on the relationships of microorganisms and microbial ecology ( Figure 6). Our findings suggest that Odaesan National Park has a large and tight network, implicating stronger coupling between microbes in driving nutrient cycles. The microbial networks were different between the early and late stages of the forest. The number of taxa constituting the microbial network was higher during the early stage, and the relationships between the microbial genera were complex, which contrasts with the general view that microbial networks become better developed with environmental succession [61]. These results were probably due to the early stage being composed of mixed-wood, which provided various substrates to the microbes and resulted in a larger number of microorganisms participating in organic matter decomposition, nutrient circulation, and complex networks. The simplification of the microbial network might also have occurred because the environmental conditions in the late stage were stabilized and simplified via the succession process. Meanwhile, the relationships between taxon pairs were independent of the relationships between the microorganisms and the soil properties ( Figures 5 and 6). Although soil properties, such as organic matter, pH, and moisture, evidently determine microorganismal growth, microbial abundance and networks can also be regulated via competition, various metabolites, antagonism, niche preferences, and other factors [62,63]. Consequently, various interactions between microorganisms and the nutrient composition of the forest soil determined microbial network composition, resulting in the formation of the microbial network unique to Odaesan National Park.

Conclusions
We investigated the changes in forest soil characteristics and microbial community structure between different successional stages in Odaesan National Park. Among the soil properties, elecation, WC, TC, TN, and pH were significantly different between the successional stages, indicating that the forest was more developed in the late successional stage and that the soil was fertile. The structure and composition of the microbial community clearly differed between the successional stages. Although there was no significant difference in the richness or diversity of the microbial community, the difference in community composition was confirmed by NMDS analysis and the ANOSIM. Changes in these microbial communities were considered to be due to taxonomic groups being correlated with soil properties, including elevation, WC, TC, TN, and pH, which significantly affected the microbial community distribution. In addition, various microbial taxa formed a network that depended on the correlations at each successional stage. The composition of the network was not consistent with the group showing similar tendencies in response to changes in soil properties. The microbial network was affected by not only the correlations with soil properties but also the metabolites produced by various strains and the enzyme profile aiding nutrition acquisition. The results of this study suggest that forest succession has a strong influence on the soil properties, microbial community structure, and microbial networks in Odaesan National Park, which sheds light on microbial ecosystems under successional change.

Conflicts of Interest:
The authors declare no conflicts of interest.