Vertical distribution of bacterial community diversity in the Greater Khingan Mountain permafrost region

Abstract Soil microorganisms are crucial contributors to the function of permafrost ecosystems, as well as the regulation of biogeochemical cycles. However, little is known about the distribution patterns and drivers of high‐latitude permafrost microbial communities subject to climate change and human activities. In this study, the vertical distribution patterns of soil bacterial communities in the Greater Khingan Mountain permafrost region were systematically analyzed via Illumina Miseq high‐throughput sequencing. Bacterial diversity in the active layer was significantly higher than in the permafrost layer. Principal coordinate analysis (PCoA) indicated that the bacterial community structure in the active layer and the permafrost layer was completely separated. Permutational multivariate analysis of variance (PERMANOVA) detected statistically significant differentiation across the different depths. The relative abundance of the dominant phyla Chloroflexi (17.92%–52.79%) and Actinobacteria (6.34%–34.52%) was significantly higher in the permafrost layer than in the active layer, whereas that of Acidobacteria (4.98%–38.82%) exhibited the opposite trend, and the abundance of Proteobacteria (2.49%–22.51%) generally decreased with depth. More importantly, the abundance of bacteria linked to human infectious diseases was significantly higher in the permafrost layer according to Tax4Fun prediction analysis. Redundancy analysis (RDA) showed that ammonium nitrogen (NH4 +‐N), total organic carbon (TOC), and total phosphorus (TP) were major factors affecting the bacterial community composition. Collectively, our findings provide insights into the soil bacterial vertical distribution patterns and major environmental drivers in high‐latitude permafrost regions, which is key to grasping the response of cold region ecosystem processes to global climate changes.

Permutational multivariate analysis of variance (PERMANOVA) detected statistically significant differentiation across the different depths. The relative abundance of the dominant phyla Chloroflexi (17.92%-52.79%) and Actinobacteria (6.34%-34.52%) was significantly higher in the permafrost layer than in the active layer, whereas that of Acidobacteria (4.98%-38.82%) exhibited the opposite trend, and the abundance of Proteobacteria (2.49%-22.51%) generally decreased with depth. More importantly, the abundance of bacteria linked to human infectious diseases was significantly higher in the permafrost layer according to Tax4Fun prediction analysis. Redundancy analysis (RDA) showed that ammonium nitrogen (NH 4 + -N), total organic carbon (TOC), and total phosphorus (TP) were major factors affecting the bacterial community composition. Collectively, our findings provide insights into the soil bacterial vertical distribution patterns and major environmental drivers in high-latitude permafrost regions, which is key to grasping the response of cold region ecosystem processes to global climate changes.

K E Y W O R D S
bacterial diversity, high-throughput sequencing, permafrost, vertical distribution

| INTRODUC TI ON
Permafrost refers to all types of ice-covered rock and soil that remain at temperatures below 0°C for more than 2 years, and these regions are among the most extreme environments on earth (Jansson & Taş, 2014). Permafrost is a crucial component of the cryosphere and plays a vital role in the global climate system (Heffernan et al., 2021). Climate warming is causing widespread permafrost degradation on a global scale, which is mainly manifested by increasing soil temperatures, shrinking permafrost areas, expanding thaw zones, and deepening the active layer thickness (Gao et al., 2021;Kim et al., 2016). The active layer is a transition layer that enables the exchange of water and heat between the permafrost layer and the atmosphere, and therefore constitutes a link between the atmosphere, the biosphere, and the hydrosphere (Steven et al., 2008).
Changes in this transition layer can not only affect the hydrothermal properties of the soil itself and change the material balance within the soil but also have major implications for the ecological processes in entire cold regions (Helbig et al., 2016). The permafrost layer is transformed into the active layer through complex phase change processes, among which changes in temperature, moisture, and organic matter are highly likely to alter the biogeochemical cycles of key elements mediated by microorganisms under global warming conditions (Hultman et al., 2015). Therefore, microbial ecology associated with permafrost has recently garnered increasing attention in the context of global climate change.
Microorganisms in permafrost, particularly bacteria, can adjust to cryoenvironments and play a vital role in the decomposition and mineralization of soil organic matter, the circulation and transformation of soil nutrients (Fry et al., 2021). Permafrost areas in the northern hemisphere account for 24% of the land surface area (Graham et al., 2012). The organic carbon storage in this area accounts for approximately 50% of the global carbon storage, which is equivalent to the sum of the vegetation and atmospheric carbon storage (Gandois et al., 2019). The degradation of permafrost leads to the decomposition of long-term sequestered organic carbon by microorganisms.
In turn, this leads to the further release of greenhouse gases (CO 2 and CH 4 ), which are considered among the most important contributors to global climate change from terrestrial ecosystems (Yergeau et al., 2010). Soil profiles provide heterogeneous habitats for microorganisms, and bacterial distribution is strongly controlled by the bacterial dispersal ability and environmental conditions, which significantly affect bacterial diversity by driving species replacement or regulating differences in the richness of species (Gittel et al., 2014).
The vertical distribution patterns of microbial communities in the unique permafrost soil remain relatively unexplored. Therefore, a deep understanding of the changes in microbial communities in different soil layers and the factors that form these communities is essential to predict the potential microbial processes and functions of permafrost ecosystem in climate change.
The Greater Khingan Mountain is the only cold-temperate coniferous forest area in China, located on the southern edge of the high-latitude permafrost zones of Eurasia, where the permafrost is fragile, thermally unstable, and vulnerable to climate and external environmental changes. The active layer experiences repeated freeze-thaw cycles and is subject to frequent disturbances by environmental factors, whereas the permafrost layer is an extreme environment with low moisture activity and extremely low nutrient conversion rates. Therefore, both of these environments are likely inhabited by different microbial flora (Mackelprang et al., 2011).
Most current studies on soil microbial communities in permafrost regions have been conducted in the circumpolar Arctic, Siberia, Alaska, and the Qinghai-Tibet Plateau (Aksenov et al., 2021;Singh et al., 2017;Tripathi et al., 2018;Wu et al., 2017), whereas much less attention has been given to the permafrost regions of the Greater Khingan Mountain, an area affected by both the effects of climate change and intense human activities, which limits our knowledge of the spatiotemporal variation trends and potential carbon feed- Miseq high-throughput sequencing. Therefore, this study provides a scientific basis for accurately predicting and assessing the response of high-latitude permafrost ecosystems to climate change.

| Study area
The study area is located at Mohe Forest Ecosystem Research Station (53°17′-53°30′ N, 122°06′-122°27′ E) in Heilongjiang Province, China. The region exhibits a cold-temperate continental monsoon climate, with mild and short summers followed by long cold winters. The average annual precipitation is 430 mm, and the average annual temperature is −4.9°C, with extreme minimum temperature as low as −49.5°C. The research area is widely covered with continuous permafrost, and the main soil type is dark brown forest soil (Liu et al., 2020). The active layer thickness is 0.7-3.0 m.

| Soil sampling
Three 20 m × 20 m sample plots were delimited in typical areas of the Larix gmelinii forest with consistent stand conditions. Next, 1.8-m-depth soil columns were drilled in each sample plot using a powered soil sampler (Drill bit was sterilized). These soil columns were divided into nine 20-cm layers, of which B1-B5 constituted the active layer and B6-B9 were the permafrost layer. A total of 27 soil samples were obtained from the three columns. After removing stones, plant roots, and other debris, the samples were thoroughly mixed, placed in a sterilized ziplock bag, and taken back to the laboratory. A portion of the samples was naturally air-dried and passed through a 2 mm sieve to characterize the physicochemical properties of the soil, and the remaining portion of the samples was stored at −80°C for gene sequencing.

| Determination of soil physicochemical properties
Soil pH value was measured using a pH meter (PHS-3E) at a 2.5:1 water-to-soil ratio. The soil samples were oven-dried at 105°C to measure their soil water content (SWC). Soil total organic carbon (TOC) was determined using a total organic carbon analyzer (Multi

| DNA extraction and sequencing
Total genomic DNA was extracted from the soil samples using
Removing chimeras and low-quality reads using VSEARCH OTUs. To reduce the impacts on differing read numbers across samples, the number of sequences of all samples was rarefied to the lowest read number using the R package phyloseq (rarify depth: 41709; McMurdie & Holmes, 2013). Applying a taxon filtering script provided by QIIME 2 was to separate the OTU tables of single microbial taxa, which were then analyzed the relative abundance of each specific taxa. The bacterial community composition was then described by the abundance of the sequences assigned to each taxon. The heatmaps of the bacterial relative abundance among different samples were created using R package gplots at the phylum and genus level classification (Warnes et al., 2020).
Alpha diversity analysis (Good's coverage index, Chao1 index, ACE index, Simpson index, Shannon index, and Phylogenetic diversity index) and principal coordinate analysis (PCoA) based on Bray-Curtis distances were conducted using QIIME 2 and the R package vegan (Dixon, 2003). The diagram of shared and unique OTUs was generated with the R package VennDiagram (Chen & Boutros, 2011). Permutational multivariate analysis of variance (PERMANOVA) was performed to identify significant differences in bacterial community composition across the different depths

| Soil physicochemical properties at different depths
The soil water content (SWC) in the active layer (B1-B5) decreased with increasing depth and tended to increase in the permafrost layer (B6-B9). The SWC (40.90%) was significantly higher at B1 (0-20 cm) than other depths (p < .05; Table 1). The soil pH varied from 5.32 to 6.65 and was highest at B7 (120-140 cm). Total nitrogen (TN), nitrate nitrogen (NO 3 − -N), and ammonium nitrogen (NH 4 + -N) contents decreased with soil depth in the active layer and were markedly higher at B1 than at other depths (p < .05). Except for NO 3 − -N (B6, 100-120 cm), the soil NH 4 + -N and TN contents did not differ significantly in the permafrost layer (p > .05). The total organic carbon (TOC) and total phosphorus (TP) contents increased first, then decreased, and then increased with depth. These parameters were significantly higher at B2 (20-40 cm) compared to other depths (p < .05).

| OTU statistics and bacterial community diversity
The number of common and unique OTUs in the samples was visualized using Venn diagram (Figure 1)
PCoA (calculated on Bray-Curtis) indicated that the cumulative contribution rate of PC1 and PC2 was 71.25%, of which the bacterial community structure in the active layer (B1-B5) and the permafrost layer (B6-B9) was completely separated on the PC1 axis, forming two distinct clusters (Figure 4). In the active layer, the bacterial community of B1 was clearly distinguishable from B2 to B5. The permafrost samples were divided into two groups, B7-B8 samples occurred in one cluster, whereas B6 was relatively close to B9. PERMANOVA analysis indicated that there were significant differences of bacterial community composition across the different depths (R 2 = 0.163, p = .001).
Functional annotation of six types of Level 1 functional groups and 40 types of Level 2 functional groups was conducted using Tax4Fun. Table 3 summarizes the taxa with relative abundance >1%.
ANOVA analyses of the Level 2 metabolic pathways indicated that the abundance of energy metabolism, nucleotide metabolism, and membrane transport was significantly higher in the active layer, whereas that of xenobiotic biodegradation and metabolism was higher in the permafrost layer. Additionally, the abundance of bacteria associated with human infectious diseases was significantly higher in the permafrost layer (except B9) than in the active layer (p < .05).

| DISCUSS ION
In our study, the bacterial community diversity varied significantly between the active layer and permafrost layer. Studies in the Canadian High Arctic (Jansson & Taş, 2014), Alaskan Arctic (Ji et al., 2020), and Siberian (Belov et al., 2020) permafrost confirmed that the soil microbial diversity was highest in the active layer and decreased with depth in the permafrost layer. Compared with the active layer, the permafrost layer exhibits several unique characteristics, including low temperatures, low oxygen levels, and low water availability. As a powerful ecological filter that limits microbial colonization, the permafrost layer reduces microbial diversity (Tripathi et al., 2018). The permafrost layer remains frozen all year round with a small amount of unfrozen water, thereby limiting the bacterial metabolic activity and completely changing the microbial niche. Interestingly, the diversity and community composition of soil bacteria show stratum specificity. Soil depth can alter soil nutrient availability, which affects the microbial community by regulating the physiological activities of microorganisms. Litter and root exudates control soil microbial community composition and diversity by providing substrates and nutrients (Millard & Singh, 2010). In the high-latitude permafrost region of the Greater Khingan Mountain, there was an extremely high amount of litter input on the soil surface derived from the Larix gmelinii forest, and plant roots were concentrated at 0-80 cm. The soil nutrient increased and sustained microbial growth in the active layer, whereas microbial abundance was suppressed by low-quality substrates in the permafrost layer (Aksenov et al., 2021).
In this study, significant spatial variations of soil microbial community composition were observed, and the main bacterial phyla in the Greater Khingan Mountain permafrost region were Chloroflexi, Acidobacteria, Actinobacteria, and Proteobacteria. This result was consistent with previous studies in the Antarctic, Arctic, and Qinghai-Tibetan Plateau permafrost soils (Tytgat et al., 2016;Wilhelm et al., 2011;Wu et al., 2017). Chloroflexi preferentially grows in low nutrient and anaerobic environments, as confirmed by studies in the Greenland and Svalbard permafrost (Ganzert et al., 2014;Xue et al., 2020). Our findings indicated that the relative abundance of Chloroflexi was highest at the transition layer (100-120 cm, above the permafrost interface) and significantly higher in the permafrost layer than in the active layer, which was probably due to the ability of these bacteria to resist low temperatures and limited nutrient availability. Acidobacteria and Proteobacteria play important roles in material metabolism and organic matter decomposition. Due to their preference for nutrient-rich environments, the higher abundances of Acidobacteria and Proteobacteria in the active layer may be due to an increase in litter and root exudate (Naumova et al., 2021). Consistent with this, other studies found that higher levels of TOC and TN increase the abundance of Acidobacteria and Proteobacteria in different soil depths (Eichorst et al., 2018;Frey et al., 2021). Furthermore, our findings indicated that Actinobacteria was predominant in the permafrost layer, and the dominant genera in the permafrost soil (Oryzihumus and Gaiella) also belonged to the Actinobacteria. By efficiently hydrolyzing complex organic compounds such as starch, cellulose, and xylan, Actinobacteria can maintain metabolic activity and cope with nutrient limitation at low temperatures (Chapman et al., 2017). More importantly, our study found that the lowest TOC content and the highest Actinobacteria abundance occurred at 140-160 cm.
Similarly, Fierer et al. (2003) also reported that Actinobacteria was adapted to low-carbon environments. In summary, there was a shift in the dominant soil bacteria from Acidobacteria and Proteobacteria dominance in the active layer to Chloroflexi and Actinobacteria dominance in the permafrost layer. Zhang et al. (2017)   Moreover, NH 4 + -N was the predominant form of inorganic nitrogen in boreal forest soil and was mainly present in an adsorbed state (Kothawala & Moore, 2009). In this study, soil NH 4 + -N content decreased with depth in the active layer due to the freeze-thaw process, which damaged the soil aggregate structure and thus affected the nitrogen attachment capacity, reducing the soil nitrogen fixation effectiveness (Nagano et al., 2018). Differences in organic carbon availability at different soil depths may partly be responsible for variations in bacterial communities between active layers and permafrost layers (Deng et al., 2015). In turn, this results in physiological adaptations of bacteria in the permafrost layer, which enables them to use more complex organic carbon forms (Eilers et al., 2012).
Because of nonlinear relationships between microbial communities and ecosystem properties and cascading influences of changes in these properties, the decrease in microbial community stability with the deepening of the active layer due to intensified permafrost thawing may potentially cause abrupt shifts in ecosystem states (Monteux et al., 2018). Especially, reduced stability of the microbial community in high-latitude permafrost regions may cause increased organic carbon decomposition, which could induce positive warming feedbacks.

| CON CLUS ION
This study characterized the vertical distribution patterns of microbial communities in the Greater Khingan Mountain permafrost soils.
Our results demonstrated that the soil microbial community diversity and composition exhibited significant spatial variations, and the changes were from Acidobacteria and Proteobacteria dominance in the active layer to Chloroflexi and Actinobacteria dominance in the permafrost layer. There was a strongly pronounced preference of Oryzihumus, Rhizobium and Gaiella for deeper permafrost layer.
In contrast, RB41, Candidatus_Solibacter, Nitrospira, H16, freshwa-ter_sediment_metagenome, Bradyrhizobium, and Bryobacter mostly present in the active layer. RDA provided further evidence that soil NH 4 + -N, TOC, and TP predominantly explained the variability of soil bacterial community structures. Therefore, the present study provides important ecological insights into permafrost microbial communities and their drivers, which will be helpful in predicting their response to changes in high-latitude permafrost ecosystems.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
Raw sequences were deposited in the NCBI public database under the accession number of PRJNA818343.