Soil Depth Determines the Composition and Diversity of Bacterial and Archaeal Communities in a Poplar Plantation

: Understanding the composition and diversity of soil microorganisms that typically mediate the soil biogeochemical cycle is crucial for estimating greenhouse gas ﬂux and mitigating global changes in plantation forests. Therefore, the objectives of this study were to investigate changes in diversity and relative abundance of bacteria and archaea with soil proﬁles and the potential factors inﬂuencing the vertical di ﬀ erentiation of microbial communities in a poplar plantation. We investigated soil bacterial and archaeal community compositions and diversities by 16S rRNA gene Illumina MiSeq sequencing at di ﬀ erent depths of a poplar plantation forest in Chenwei forest farm, Sihong County, Jiangsu, China. More than 882,422 quality-ﬁltered 16S rRNA gene sequences were obtained from 15 samples, corresponding to 34 classiﬁed phyla and 68 known classes. Ten major bacterial phyla and two archaeal phyla were found. The diversity of bacterial and archaeal communities decreased with depth of the plantation soil. Analysis of variance (ANOVA) of relative abundance of microbial communities exhibited that Nitrospirae, Verrucomicrobia, Latescibacteria, GAL15, SBR1093, and Euryarchaeota had signiﬁcant di ﬀ erences at di ﬀ erent depths. The transition zone of the community composition between the surface and subsurface occurred at 10–20 cm. Overall, our ﬁndings highlighted the importance of depth with regard to the complexity and diversity of microbial community composition in plantation forest soils.


Introduction
Microorganisms exist throughout the soil profile and play vital roles in soil biogeochemical cycling, thereby influencing greenhouse gas emissions from soil and plant growth by altering nutrient availability [1,2]. Soils are considered one of the most diverse microbial habitats because of their extensive physical, chemical, and biological heterogeneity [3]. Our current understanding of the distribution of soil microbial communities and associated processes has largely been restricted to the surface soil [4,5]. Plantation forests provide over 45% of the global industrial round wood production and could have applications in mitigating global changes [6]. Within these forests, many functional microbes mediate the cycle of matter and energy in subsurface horizons. Therefore, the subsurface soil cannot be ignored in studies of greenhouse gas emissions from plantations and the underlying mechanisms.
Environmental conditions, such as temperature and moisture, tend to influence soil physical and chemical characteristics in shallow to deeper regions due to additions, losses, transfers, and

Site Description
The site was located in Chenwei forest farm (33 • 20' N, 118 • 20' E), Sihong County, on the west bank of Hongze Lake in northern Jiangsu, China. The climate belongs to a mid-latitude warm zone with long periods of sunshine. The mean annual temperature is 14.4 • C, and the mean monthly temperature ranges from −7 • C in January to 28 • C in July. The mean annual precipitation is 972.5 mm, mostly occurring from June to August. The soil is a Gleysols [17] with clay loam texture derived from lacustrine sediments.
Poplar plantations were established in March 2007 with 1-year-old seedlings of clone "Nanlin-95", a hybrid of clone I-69 (Populus deltoides Bartr. cv. 'Lux') × clone I-45 (P. euramericana [Dode] Guineir. cv. 'I-45/51'), with an area of 6.7 ha. Disking was performed to a depth of 30 cm to improve soil aeration and moisture movement prior to planting. A randomized block design was used to establish the trials with three replicates that were randomly arranged in 12 plots established at the same topographic position, with each plot measuring approximately 1200-1800 m 2 (50 trees/plot). The average diameter at breast height and tree height were 20.7 cm and 21.2 m, respectively, in 2015. The main understory plant species included Echinochloa crusgalli, Youngia japonica, Geranium wilfordii, Duchesnea indica, and Herba cirsii.

Sampling
In July 2018, five pits were selected by S type in each block (four plots). Soils from five depths (D1: 0-10 cm, D2: 10-20 cm, D3: 20-30 cm, D4: 30-40 cm, D5: 40-50 cm) in each pit were collected using soil drills (d = 2.5 cm) after removing surface litter and herbs. Soil samples of the same horizon were mixed in each sampling plot and were then placed in a thermostatic chamber to analyze microbial communities. At the same time, soil samples were collected using a ring core (diameter = 5.05 cm) from each depth to measure bulk density (BD). Approximately 1 kg soil was collected in resalable plastic bags from corresponding pits to evaluate soil physicochemical characteristics. Soil samples used for microbial detection were frozen at −30 • C for 1 day before DNA extraction. One part of fresh soil samples was reserved at 4 • C, and the other part was air-dried and sieved through 2-mm mesh to measure soil properties after discarding fine roots and stones.

Soil Physicochemical Characteristics
Bulk density was measured by the cutting ring method. Soil pH was assessed with a pH probe (AB15 + Basic, Accumet, San Diego, CA, USA) by mixing soil and water in a 1:2.5 volume ratio. Soil organic carbon (SOC) was determined by the potassium dichromate oxidation-ferrous sulfate titration method. Approximately 1 g dry soil was added to 2 mL catalyst and 5 mL concentrated sulfuric acid for digestion; this sample was used to detect the total nitrogen (TN) content with an automatic continuous flow analyzer (AA3; Bran Luebbe, Norderstedt, Germany) [18]. Approximately 10 g dry soil was extracted with distilled water and potassium sulfate, and the extract was used to measure the concentrations of ammonium nitrogen (NH 4 + -N) and nitrate nitrogen (NO 3 − -N) using an ultraviolet-visible (UV-vis) spectrophotometer (UV-2550; Shimadzu, Tokyo, Japan) [18].

DNA Extraction, Polymerase Chain Reaction (PCR) Amplification, and Illumina MiSeq Sequencing
DNA was extracted from 15 soil samples from five depths (three repetitions for each depth) using an E.Z.N.A. Soil DNA Kit (Omega Bio-tek, Norcross, GA, USA), according to the manufacturer's instructions. The final DNA concentration and purification were determined using a NanoDrop 2000 UV-vis spectrophotometer (Thermo Scientific, Wilmington, DE, USA), and DNA quality was checked by 1% agarose gel electrophoresis. The primer sets of 515FmodF (5 -GTGYCAGCMGCCGCGGTAA-3 ) and 806RmodR (5 -GGACTACNVGGGTWTCTAAT-3 ) were designed to amplify the V4 hypervariable region of the 16S rRNA gene from nearly all bacteria and archaea [19].
Illumina Miseq sequencing yielded 882,422 quality sequences. Raw fastq files were quality-filtered by Trimmomatic and merged by Fast Length Adjustment of Short reads (FLASH, https://ccb.jhu.edu/ software/FLASH/) with the following criteria: (i) the reads were truncated at any site receiving an average quality score <20 over a 50 bp sliding window; (ii) sequences whose overlap being longer than 10 bp were merged in terms of their overlap with mismatch no more than 2 bp; (iii) sequences of each sample were separated due to barcodes (exactly matching) and primers (allowing 2 nucleotide mismatching), and then reads containing ambiguous bases were deleted [20,21]. Operational taxonomic units (OTUs) were clustered with 97% similarity cutoff using UPARSE (version 7.1 http://drive5.com/ uparse/) with a "greedy" algorithm that performs chimera filtering and OTU clustering simultaneously. The taxonomy of each 16S rRNA gene sequence was analyzed by Ribosomal Database Project (RDP) Classifier algorithm (http://rdp.cme.msu.edu/) against Silva (SSU128) 16S rRNA database using confidence threshold of 70% [22]. The raw data were submitted into the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) database (Accession Number: PRJNA541200).

Data Analysis
One-way analysis of variance (ANOVA) of soil physical and chemical characteristics at different depths was performed using SPSS 18.0. Shannon [23], Chao1 [24] and Coverage [25] indexes were used to evaluate alpha diversity, abundance, and coverage of the microbial community, respectively, and formulas were listed as follows: where S obs is the number of OTUs actually observed; p i is the proportion of individuals of the i th OTU divided by the total number of sequences (N).
where: n 1 is the number of OTUs with only one sequence; n 2 is the number of OTUs with only two sequences.
To assess beta diversity patterns, i.e., the changes in microbial community composition at different depths, the principal component analysis (PCA) was used. Venn graphs were used to assess the numbers of shared and unique operational taxonomic units (OTUs) at different depths based on a similarity level of 97% OTUs. The relationships between bacterial and archaeal community compositions and edaphic factors were evaluated using redundancy analysis (RDA) and Pearson correlation matrix.

Soil Physicochemical Characteristics with Depth
Soil physicochemical characteristics varied with depth ( Figure 1). Briefly, BD and soil pH increased with depth. Significant differences in BDs of D1, D2, and D5 were observed. Moreover, soil pH was significantly lower in D1 and D2 than in D5. However, SOC and TN decreased with depth. The SOC differed significantly at each depth, whereas only the first four layers showed significant differences for TN. NO 3 − -N and NH 4 + -N concentrations also showed different patterns. For example, NO 3 − -N concentrations generally decreased with depth in the first four layers, whereas the concentration increased at D5. The concentration of NH 4 + -N reached a maximum at D2 and a minimum at D5.

Alpha and Beta Diversity Patterns
Total effective sequences of different soil depths ranged from 45,874 to 57,609 (Table 1). Shannon index declined with depth, and analysis of the Chao index showed that D2 and D3 had higher species richness than the other samples. Coverage of all samples was more than 99%, demonstrating that the depth of the sequence was sufficient. PCA divided the bacterial and archaeal communities into three categories ( Figure 2): D1, D2, and the group of D3-D5. The contributions of two selected principal components for the differences in bacterial and archaeal community compositions were 26.47% and 12.04%, separately.

Alpha and Beta Diversity Patterns
Total effective sequences of different soil depths ranged from 45,874 to 57,609 (Table 1). Shannon index declined with depth, and analysis of the Chao index showed that D2 and D3 had higher species richness than the other samples. Coverage of all samples was more than 99%, demonstrating that the depth of the sequence was sufficient. PCA divided the bacterial and archaeal communities into three categories ( Figure 2): D1, D2, and the group of D3-D5. The contributions of two selected principal components for the differences in bacterial and archaeal community compositions were 26.47% and 12.04%, separately.

Relationship between Bacterial and Archaeal Community Compositions and Edaphic Factors
BD (P = 0.002), pH (P = 0.012), SOC (P = 0.002), TN (P = 0.003), and NO3 --N (P = 0.03) had significant effects on bacterial and archaeal community compositions ( Figure 5). Furthermore, BD and pH were negatively correlated with other soil chemical characteristics. The Pearson correlation   Table 2). Proteobacteria and Bacteroidetes had significantly negative correlations with BD but significantly positive correlations with TN, SOC, and NO 3 − -N. Nitrospirae and GAL15 had significantly positive correlations with BD and pH, but significantly negative correlations with TN, SOC, and NH 4 + -N. Thaumarchaeota was significantly negatively correlated with BD, but was significantly positively correlated with TN, SOC, and NH 4 + -N. Euryarchaeota was significantly positively correlated with pH, but was significantly negatively correlated with TN, SOC, and NO 3 − -N. matrix showed that the soil physicochemical characteristics had different effects on the bacterial and archaeal communities (Table 2). Proteobacteria and Bacteroidetes had significantly negative correlations with BD but significantly positive correlations with TN, SOC, and NO3 --N. Nitrospirae and GAL15 had significantly positive correlations with BD and pH, but significantly negative correlations with TN, SOC, and NH4 + -N. Thaumarchaeota was significantly negatively correlated with BD, but was significantly positively correlated with TN, SOC, and NH4 + -N. Euryarchaeota was significantly positively correlated with pH, but was significantly negatively correlated with TN, SOC, and NO3 --N.

Discussion
In this study, we found that the diversity of bacterial and archaeal communities decreased with depth in a poplar plantation. Some surface-dwelling microbes showed reduced survival in the subsurface soil owing to the strong ecological filtration function in vertical space [4]. These findings are consistent with previous studies on soil microbial community diversity in response to depth across paddy soils [26], grassland soils [27], forest soils [5,28], and tundra soils [29]. Typically, the transition zone for the microbial community is thought to be 10-25 cm in natural forest soils [4]. We found that the transition zone for the bacterial and archaeal communities between surface and subsurface regions occurred at 10-20 cm at the study site.
Environmental conditions, such as oxygen levels and temperature, can affect the soil microbial community. Changes in soil physicochemical characteristics, such as pH, BD, SOC, TN, and NO 3 − -N, with depth had significant positive (SOC and TN) or negative (BD and pH) correlations with the diversity of the microbial community. Soil pH has been identified as the dominant factor shaping microbial communities for surface soils across the continental scale [30]. Soil pH increased with depth, but microbial diversity was reversed in our study. The discrepancy between our soil profile study and previous studies across the continental scale could be explained by the very narrow pH range (6.6-6.8) observed in our soil samples. Furthermore, changes in the quality of SOC (e.g., labile SOC and resistant SOC) with soil depth could have stronger effects on the community structure than the total quantity of SOC [3]. Shifts in microbial communities between surface and subsurface soils could partially result from differences in the availability of labile SOC, leading to physiologically altered (adapted) organisms capable of utilizing more recalcitrant sources of organic carbon [31]. However, in our study, the quantity of SOC still strongly affected the microbial community structure. We observed 10 major bacterial phyla in this study, in contrast to a previous study in which nine dominant groups were obtained from various soils ranging from agricultural land and grassland to pristine forest [32]. The unique bacterial phyla identified in this study were Nitrospirae and GAL15. The relative abundance of Nitrospirae and GAL15 increased significantly with depth, indicating that these organisms have a selective advantage in deeper soils [33]. Moreover, Nitrospira, belonging to the bacterial phylum Nitrospirae, were demonstrated to the most diverse and abundant nitrite-oxidizing bacteria (NOB) [34]. Latescibacteria and Tectomicrobia are potential new bacterial phyla obtained through metagenomic and single-cell genomics in recent years [35,36]. The relative abundance of Latescibacteria at intermediate depth (20-30 cm) was significantly larger than those at shallow and deeper regions, whereas the relative abundance of Tectomicrobia was more uniform with changes in depth.
The abundance of Proteobacteria and Bacteroidetes decreased with depth in the poplar soil profiles, which is consistent with previous studies [4,8,33]. The decreased abundances of Proteobacteria and Bacteroidetes could be attributed to the general copiotrophic properties and responses to labile sources of SOC at shallow depths [37]. Proteobacteria displayed various types of obligate and facultative chemo-and photoautotrophic CO 2 fixation [38]. Moreover, previous studies found that Alphaproteobacteria and Betaproteobacteria were able to induce nitrogen-fixing nodules [39], and Gammaproteobacteria could carry out ammonia oxidation [40]. As well, methanotrophs were found within Alphaproteobacteria, Betaproteobacteria, and Gammaproteobacteria, and these bacteria were more likely to be detected in soil with pH > 6.0 [41].
Additionally, we found that the relative abundances of Actinobacteria, Acidobacteria, Chloroflexi, and Firmicutes exhibited a mid-soil depth peak, but not significant. However, some studies reported that the relative abundances of Actinobacteria, Acidobacteria, Chloroflexi, and Firmicutes increased with depth [42,43], attributing to their adaptation to resource-limited conditions [8,37,44]. This discrepancy may because of the different soil type and land use type. The relative abundance of Planctomycetes did not show clear changes across profiles [4]. This is consistent with our finding. It was also reported that Planctomycetes derive their energy for growth from the conversion of ammonium and nitrite into dinitrogen gas in the complete absence of oxygen, regulating the nitrogen cycle [45].
As there is no significant change of relative abundance with soil depth, anaerobic ammonium oxidation mediated by the bacteria may not change with soil depth in the poplar plantations if the actual effectiveness in surface and subsurface soils is similar.
The relative abundance of Verrucomicrobia was significantly higher between 10 and 20 cm than other layers in this study, which is shallower than the middle depth peak (20-40 cm) in natural forests [4,8]. Verrucomicrobia may prefer a micro-aerobic environment rather than aerobic and extreme anoxic environments [4]. Previous studies have reported that some members of Verrucomicrobia could oxidize methane and take advantage of methane as the only source of carbon, making them the only known aerobic methanotrophs except for the Proteobacteria [46,47]. Consequently, the subsurface peak in Verrucomicrobia abundance may correspond to the peak in methane oxidation [48].
Although archaea were relatively rare in all layers, some individual members of archaea were abundant. For example, Thaumarchaeota was the dominant taxon of archaea within soil, and its relative abundance decreased with depth in this study. This pattern was consistent with a previous study showing that Thaumarchaeota was the only archaeal group in aerobic shallow soil [49]. Furthermore, Thaumarchaeota has been shown to have ammonia oxidation [50,51] and carbon sequestration activities [52,53], but to exhibit high denitrification potential under hypoxic conditions [53]. The relative abundance of Euryarchaeota increased significantly with depth within the first four layers, which could attribute to their preference for anaerobic environment. Notably, members of Euryarchaeota have been found not only to produce methane [54], but also oxidize methane, fix nitrogen, and reduce nitrates [55][56][57]. Generally, methanogens are thought to be strictly anaerobic Euryarchaeota but have recently been found in fully aerated environments of forests, grasslands, and agricultural soils [58][59][60]. Euryarchaeota have been shown to be ubiquitous and highly abundant in aerated upland soils and to participate in global methane production in upland soil [61]. However, it is difficult to detect the specific methanogens at the class level based on the primer used in this study. We will further identify the functional bacteria and archaea of the poplar plantations with specific functional primers (such as methanogens, methanotrophs, nitrogen fixing bacteria, denitrifying bacteria, etc.), and explore the microbial-driven mechanism of soil greenhouse gas productions and emissions.
Assuming that forest managers or scientists are more interested in greenhouse gas budgets from such managed ecosystems, further work will be required on at least two fronts. First, field measurements on CO 2 , CH 4 , and N 2 O emission rates across the surface between soils and the atmosphere must be conducted and microbial composition should be detected accordingly. However, this study is primarily concerned the change of microbial diversity and composition with soil profile, as the plantation ecosystem sometimes showed as a net CH 4 source. Second, laboratory incubation studies must include the greenhouse gas production and corresponding microbial composition change at certain environmental conditions.

Conclusions
As a vital factor in the construction of environmental gradients, soil depth affects the complexity and diversity of bacterial and archaeal community structure. The vertical spatial heterogeneity of soil drives microbes to seek habitats selectively, resulting in apparent changes in the diversity and relative abundances of bacteria and archaea across soil depth. There were significant correlations between shifts of community compositions of bacteria and archaea and soil physicochemical characteristics with changes in depth. Moreover, we observed some bacterial and archaeal communities in poplar plantation soil profiles, and these microbes may regulate important soil carbon and nitrogen processes. Defining the biogeography of soil bacteria and archaea, particularly for deeper regions, could provide insights into unique and potentially significant processes affecting soil ecosystems. Studies such as our current work have produced useful information on microbial diversity in soil profiles. Further studies are needed to focus on functional bacteria such as methanogenesis, methane-oxidizing bacteria, and nitrogen-related bacteria for identifying related biogeochemical processes in man-made forest ecosystems.