Seasonal Variation in the Rhizosphere and Non-Rhizosphere Microbial Community Structures and Functions of Camellia yuhsienensis Hu

Camellia yuhsienensis Hu, endemic to China, is a predominant oilseed crop, due to its high yield and pathogen resistance. Past studies have focused on the aboveground parts of C. yuhsienensis, whereas the microbial community of the rhizosphere has not been reported yet. This study is the first time to explore the influence of seasonal variation on the microbial community in the rhizosphere of C. yuhsienensis using high-throughput sequencing. The results showed that the dominant bacteria in the rhizosphere of C. yuhsienensis were Chloroflexi, Proteobacteria, Acidobacteria, Actinobacteria, and Planctomycetes, and the dominant fungi were Ascomycota, Basidiomycota, and Mucoromycota. Seasonal variation has significant effects on the abundance of the bacterial and fungal groups in the rhizosphere. A significant increase in bacterial abundance and diversity in the rhizosphere reflected the root activity of C. yuhsienensis in winter. Over the entire year, there were weak correlations between microorganisms and soil physiochemical properties in the rhizosphere. In this study, we found that the bacterial biomarkers in the rhizosphere were chemoorganotrophic Gram-negative bacteria that grow under aerobic conditions, and fungal biomarkers, such as Trichoderma, Mortierella, and Lecanicillium, exhibited protection against pathogens in the rhizosphere. In the rhizosphere of C. yuhsienensis, the dominant functions of the bacteria included nitrogen metabolism, oxidative phosphorylation, glycine, serine and threonine metabolism, glutathione metabolism, and sulfur metabolism. The dominant fungal functional groups were endophytes and ectomycorrhizal fungi of a symbiotroph trophic type. In conclusion, seasonal variation had a remarkable influence on the microbial communities and functions, which were also significantly different in the rhizosphere and non-rhizosphere of C. yuhsienensis. The rhizosphere of C. yuhsienensis provides suitable conditions with good air permeability that allows beneficial bacteria and fungi to dominate the soil microbial community, which can improve the growth and pathogen resistance of C. yuhsienensis.


Introduction
Microorganisms play a significant role in the decomposition process of soil organic matter, litter, and wood residues [1][2][3]. The microbial community structures partly reflect the nutrient and health conditions of the soil. Some bacterial phyla only proliferate in copiotrophic substances, such as Alphaproteobacteria, Betaproteobacteria, and Bacteroidetes, while others proliferate in oligotrophic

Site Description
The sampling site was located at a 10-year-old C. yuhsienensis forest in Youxian, Hunan, China (N 27 • 02 , E 113 • 20 ). The plant and row spacing were 2 m and 3 m, respectively. The climate is subtropical monsoon with a mean annual total rainfall and temperature of 1410 mm and 17.8 • C, respectively [33]. The soil at the experimental site is a Quaternary red clay with a pH of 4.1-4.5. The total organic matter content is 17-23 g/kg. The total nitrogen content is approximately 330-810 mg/kg. The available phosphorus content is 0.1-10 mg/kg, and the available potassium content is 64-95 mg/kg. The chemical properties mentioned above are displayed in Table S1.

Experimental Design and Sampling
According to the phenological phase of C. yuhsienensis, the soil samples were collected on 23 October 2018 (autumn, C), 19 January 2019 (winter, D), 5 April 2019 (spring, A), and 22 July 2019 (summer, B). Three quadrats of 20 m × 20 m were selected in the C. yuhsienensis plantation. Five trees were chosen according to an "S" type route in the center of each quadrat. The rhizosphere soil samples of C. yuhsienensis were collected, as described by Reference [34] after shoveling the roots from four ordinations approximately 0.5 m away from the trunk. The corresponding non-rhizosphere soil samples at 0-20 cm depth were collected as a control approximately 1 m away from the trunk. The rhizosphere and non-rhizosphere soils in each quadrate were each well blended as a single sample. The soil samples were then transported back to the laboratory on dry ice. After removing debris and roots, the soil samples were well mixed, ground, and sieved (<2 mm). One part of the soil samples was used for evaluating the physiochemical properties after air-drying in the shade, while the remainder was stored at −80 • C for high-throughput sequencing at Genedenovo Biotechnology Co., LTD (Guangzhou, China). The soil temperature at 0-20 cm depth was detected using a Wdsen Electronic temperature recorder (iButton DS1925) during the whole experimental period ( Figure S1).

Illumina Novaseq6000 Sequencing
Amplicons were extracted from 2% agarose gels and purified using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, USA.) according to the manufacturer's instructions and quantified using an ABI StepOnePlus Real-Time PCR System (Life Technologies, Foster City, CA, USA). Purified amplicons were pooled in equimolar ratios and paired-end sequenced (PE250) on an Illumina platform according to the standard protocols. The raw reads were deposited into the NCBI Sequence Read Archive (SRA) database (BioProject: PRJNA646017).

Statistical and Bioinformatics Analysis
Raw reads were further filtered using FASTP according to the rule: Remove reads containing more than 10% unknown nucleotides (N) and reads with less than 50% of bases with quality (Q-value)>20 [45]. Paired-end clean reads were merged as raw tags using FLASH (version 1.2.11) with a minimum overlap of 10 bp and mismatch error rates of 2% [46]. The effective tags were clustered into operational taxonomic units (OTUs) of ≥ 97% similarity using the UPARSE (version 9.2.64) pipeline [47]. The tag sequence with the highest abundance was selected as a representative sequence within each cluster.
A Venn analysis was used to identify unique and common OTUs among the different groups, was performed using the R "VennDiagram" package (version 1.6.16) [48]. The OTU rarefaction curves and rank abundance curves were plotted using the R "ggplot2" package (version 2.2.1) [49]. Alpha diversities of bacteria and fungi (Sobs, Shannon, and Chao 1 index) were calculated in QIIME (version 1.9.1) [50]. The abundance statistics of each taxonomy were visualized using Krona (version 2.6) [51]. Principal co-ordinates analysis (PCoA) and permutational multivariate analysis of variance (MANOVA) (Permanova) based on unweighted unifrac distance were used to evaluate the influence of plant root and seasonal dynamics on bacterial and fungal community structures. A least discriminant analysis (LDA) effect size (LEfSe) taxonomic cladogram was used to identify bacterial biomarkers (LDA > 3.5) and fungal biomarkers (LDA > 3.5) in the different treatment using LEfSe software [52]. Indicator analysis was used to determine the significantly different species (biomarkers) between rhizosphere and non-rhizosphere soil during the entire year using the R "labdsv" package (version 2.0-1) [53]. The Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis of the OTUs was inferred using Tax4Fun (version 1.0) [54]. The functional group (Guild) of the fungi (relative abundance greater than 0.01% in the rhizosphere) was inferred using FUNGuild (version 1.0) [55]. The bacterial function differences between the rhizosphere and non-rhizosphere samples were calculated by Welch's t-test, and the fungal function differences between the rhizosphere and non-rhizosphere were calculated using the Wilcoxon rank test in the R "vegan" package (version 2.5.3) [56].
One-way analysis of variance (ANOVA) was used to determine the significant differences among different treatments during the whole year. Nonparametric tests with a significance level of 0.05 were used to compare the means of bacterial and fungal alpha diversities in different treatments and seasons. Variance partitioning analysis, generated in the R "vegan" package (version 2.5.3) [56], was used to evaluate the contribution of environmental factors to microbial community structure. Pearson's correlation coefficients between environmental factors and microbial species and alpha diversities were calculated by Omicshare tools, a free online platform for data analysis [57].

Quality Control Report of the High-Throughput Sequencing
The tags and OTUs of the bacteria and fungi are presented in Table S2 and Table S3. In order to ensure the reproducibility and validity of the microbial data, the data extraction flat and dilution cures were processed before analysis [58]. According to Figure S2, the dilution cure of the Sobs index indicated that there would be more bacteria and fungi if sequencing were continued, but the plateau of the dilution cure of Shannon's index was reached early, showing that the number of reads was sufficient for this research.

Microbial Composition
According to Figure 1, the dominant bacteria were Chloroflexi, Proteobacteria, Acidobacteria, Actinobacteria, Planctomycetes, Firmicutes, and Bacteroidetes at the phylum level. The dominant fungi were Ascomycota, Basidiomycota, and Mucoromycota at the phylum level. Venn diagrams ( Figure 2) indicated that, with the exception of summer, the number of specific bacterial OTUs in the rhizosphere was higher than in the non-rhizosphere. In winter, the number of specific bacterial OTUs in the non-rhizosphere and the number of common bacterial OTUs between the rhizosphere and non-rhizosphere were much lower than in the other seasons, while the number of specific bacterial OTUs in the rhizosphere was much higher than in the other seasons. In the rhizosphere, the number of specific bacterial OTUs in spring and winter was higher than in summer and autumn, while the opposite was observed in the non-rhizosphere. The number of common bacterial OTUs among each season in the rhizosphere was 934, which was higher than the value of 456 in the non-rhizosphere.

Microbial Composition
According to Figure 1, the dominant bacteria were Chloroflexi, Proteobacteria, Acidobacteria, Actinobacteria, Planctomycetes, Firmicutes, and Bacteroidetes at the phylum level. The dominant fungi were Ascomycota, Basidiomycota, and Mucoromycota at the phylum level. Venn diagrams ( Figure 2) indicated that, with the exception of summer, the number of specific bacterial OTUs in the rhizosphere was higher than in the non-rhizosphere. In winter, the number of specific bacterial OTUs in the non-rhizosphere and the number of common bacterial OTUs between the rhizosphere and nonrhizosphere were much lower than in the other seasons, while the number of specific bacterial OTUs in the rhizosphere was much higher than in the other seasons. In the rhizosphere, the number of specific bacterial OTUs in spring and winter was higher than in summer and autumn, while the opposite was observed in the non-rhizosphere. The number of common bacterial OTUs among each season in the rhizosphere was 934, which was higher than the value of 456 in the non-rhizosphere. As observed with the bacterial OTUs, the number of specific fungal OTUs in the rhizosphere was higher than in the non-rhizosphere, except for in summer. The highest number of common fungal OTUs between the rhizosphere and non-rhizosphere was in summer. In both the rhizosphere and non-rhizosphere, the highest number of specific fungal OTUs was detected in summer, while the   As observed with the bacterial OTUs, the number of specific fungal OTUs in the rhizosphere was higher than in the non-rhizosphere, except for in summer. The highest number of common fungal OTUs between the rhizosphere and non-rhizosphere was in summer. In both the rhizosphere and non-rhizosphere, the highest number of specific fungal OTUs was detected in summer, while the lowest was found in autumn. The number of common fungal OTUs among each season in the rhizosphere was 94, which was higher than the value of 43 in the non-rhizosphere.

Microbial Diversity
According to Table 1, there were no significant differences in bacterial Sobs and Shannon indexes among different seasons in the rhizosphere. The bacterial Sobs, Chao 1, and Shannon indexes in winter were much lower than in the other seasons in the non-rhizosphere. The highest fungal Sobs, Chao 1, and Shannon indexes were detected in summer in both the rhizosphere and non-rhizosphere. There was no significant difference in fungal Sobs and Shannon index between the rhizosphere and non-rhizosphere during the entire year. In autumn, the fungal Chao 1 index in the rhizosphere was remarkably higher than in the non-rhizosphere. PCoA and Permanova ( Figure 3) indicated that seasonality had a great influence on the soil bacterial communities in the non-rhizosphere (p < 0.001), fungal communities in the rhizosphere (p < 0.001), and fungal communities in the non-rhizosphere (p < 0.001). Across the whole year, the bacterial communities in the rhizosphere differed significantly (p = 0.048) from those in the non-rhizosphere ( Figure 3A). However, there was no significant difference (p = 0.073) between fungal communities in the rhizosphere and non-rhizosphere ( Figure 3B).

Microbial Biomarkers in the Rhizosphere
In order to identify biomarkers in the rhizosphere, LEfSe and indicator analysis were applied. As observed with microbial diversity and composition, the biomarkers in the rhizosphere were also significantly affected by seasonal variation. According to Figure S3, the most bacterial biomarkers were detected in winter, while there were no biomarkers in summer in the rhizosphere. As observed with the bacterial biomarkers, the most fungal biomarkers were detected in winter, while the least was detected in summer ( Figure S4).
Across the entire year, the bacterial and fungal biomarkers in the rhizosphere were both much higher than in the non-rhizosphere. In the rhizosphere, the bacterial biomarkers included Chthoniobacter, Rhodanobacter, Reyranella, Aquicella, Subgroup_10, Granulicella, Labrys, Gemmata, GAS113, Sphingomonas, and Pajaroellobacter at the genus level ( Figure 4A), while the fungal biomarkers were Trechispora, Staphylotrichum, Chaetosphaeria, Trichoderma, Mortierella, Trichoglossum, Lecanicillium, Sarcodon, and Leptodontidium at the genus level ( Figure 4B). In order to identify biomarkers in the rhizosphere, LEfSe and indicator analysis were applied. As observed with microbial diversity and composition, the biomarkers in the rhizosphere were also significantly affected by seasonal variation. According to Figure S3, the most bacterial biomarkers were detected in winter, while there were no biomarkers in summer in the rhizosphere. As observed with the bacterial biomarkers, the most fungal biomarkers were detected in winter, while the least was detected in summer ( Figure S4).

Functional Analysis of Microorganisms
The abundances of nitrogen metabolism, oxidative phosphorylation, glycine, serine and threonine metabolism, glutathione metabolism, and sulfur metabolism in the rhizosphere were significantly higher than in the non-rhizosphere ( Figure 5). According to the FUNGuild analysis, endophytes and ectomycorrhizal fungi in the rhizosphere were significantly higher than in the nonrhizosphere ( Figure 6B). Trophic analysis indicated that symbiotrophs were more abundant in the rhizosphere, while pathotrophs were more abundant in the non-rhizosphere ( Figure 6B).

Functional Analysis of Microorganisms
The abundances of nitrogen metabolism, oxidative phosphorylation, glycine, serine and threonine metabolism, glutathione metabolism, and sulfur metabolism in the rhizosphere were significantly higher than in the non-rhizosphere ( Figure 5). According to the FUNGuild analysis, endophytes and ectomycorrhizal fungi in the rhizosphere were significantly higher than in the non-rhizosphere ( Figure 6B). Trophic analysis indicated that symbiotrophs were more abundant in the rhizosphere, while pathotrophs were more abundant in the non-rhizosphere ( Figure 6B).

Interactions Between Environmental Factors and Microbial Communities
According to Table S1, seasonality and rhizosphere had a remarkable influence on soil physiochemical properties. Different soil physiochemical properties had different responses to seasonal variation. Available nutrients, such as AP, AK, AMN, NN, and AHN, varied more frequently than TP, TN, and TOC along with seasonal dynamics. AP was higher in spring and summer than in winter and autumn, while TK, AMN, NN, and SWC were higher in autumn and winter than in spring and summer. There was no significant difference in most soil physiochemical properties between the rhizosphere and non-rhizosphere, except for AMN, AHN, TOC, and N/P. AMN was significantly higher in the non-rhizosphere than in the rhizosphere in summer. AHN was significantly lower in the rhizosphere than in the non-rhizosphere in spring, while the result was reversed in winter. TOC and N/P were both significantly higher in the rhizosphere than in the non-rhizosphere in winter. Figure S5 indicates the environmental factors with significant contributions to the microbial communities in the rhizosphere and non-rhizosphere soil at the genus level. According to Table 2, there were no significant correlations between environmental factors and bacterial diversity in the rhizosphere. According to Table 3, in the rhizosphere, fungal abundance was positively correlated with TP and Tem, but negatively correlated with TK, NN, and N/P. "*" and "**" indicate the significant correlations at 0.05 and 0.01 level, respectively. AP, available phosphorous; TP, total phosphorous; AK, available potassium; TK, total potassium; AMN, ammonium nitrogen; NN, nitrate nitrogen; AHN, alkaline hydrolyzable nitrogen; TN, total nitrogen; TOC, total organic carbon; SWC, soil water content; Tem, soil monthly mean temperature; C/N, ratio of total carbon to total nitrogen; C/P, ratio of total carbon to total phosphorous; N/P, ration of total nitrogen to total phosphorous. n = 12.
In order to elucidate the interactions between environmental factors and microbial species, the biomarkers and the top 10 abundant species at the genus level were selected to conduct correlation analysis with environmental factors (Figures S6 and S7). In both bacteria and fungi, the interactions between microorganisms and environmental factors were strikingly different between the rhizosphere and non-rhizosphere. pH was significantly correlated with most bacteria in the rhizosphere ( Figure S6). Acidothermus, the most abundant bacteria in the rhizosphere, was significantly positively related to Tem both in the rhizosphere and non-rhizosphere. Compared with bacteria, there were more significant correlations between environmental factors and fungi at the genus level ( Figure S7). Most fungi were negatively correlated with Tem, but positively correlated with NN and AHN in the rhizosphere. "*" and "**" indicate the significant correlations at 0.05 and 0.01 level, respectively. AP, available phosphorous; TP, total phosphorous; AK, available potassium; TK, total potassium; AMN, ammonium nitrogen; NN, nitrate nitrogen; AHN, alkaline hydrolyzable nitrogen; TN, total nitrogen; TOC, total organic carbon; SWC, soil water content; Tem, soil monthly mean temperature; C/N, ratio of total carbon to total nitrogen; C/P, ratio of total carbon to total phosphorous; N/P, ration of total nitrogen to total phosphorous. n = 12.

Discussion
Numerous studies have shown that seasonal variation strongly regulates soil physiochemical properties and microorganisms [10,[59][60][61][62]. However, no studies have assessed the seasonal variation in soil properties and microbial communities in the rhizosphere of C. yuhsienensis. In this study, seasonal variation was the predominant influencing factor on soil physiochemical properties. The soil properties among seasons in this study differed greatly from those of Zhang, Cui, Guo, and Xi [30] in C. oleifera forests. This might be because the climate and soil types are markedly different. Many studies have reported that the soil properties in the rhizosphere differ greatly from those in the non-rhizosphere [30,63]. Our result showed that AHN, TOC, and N/P were both significantly higher in the rhizosphere than in the non-rhizosphere in winter. This indicates that the C. yuhsienensis roots maintained higher activity in winter. No significant differences were found in soil pH between rhizosphere and non-rhizosphere-which is not consistent with Zhang et al. [30], who found that the rhizosphere pH of C. oleifera was significantly lower than non-rhizosphere. A previous study showed that in order to acquire more phosphorous, C. oleifera exuded various organic acid, which led to the decline of pH [64]. These results suggested that C. yuhsienensis roots might not be as active as C. oleifera and had a weaker influence on soil physiochemical properties.
Compared with bacteria, fungi were more sensitive to soil physicochemical properties in this study (Tables 2 and 3). The positive correlations between TP and fungal richness (Table 3) in this study are contrary to those studies which showed that phosphorous is the key limiting factor on the establishment of symbioses between plant and mycorrhizal fungi [65,66]. Previous studies about the negative influences of phosphorous on mycorrhizal fungi were under the high content of phosphorous. However, the contents of the AP and TP were low in this study (Table S1), which is not enough to inhibit mycorrhizal fungi. AP was widely considered to be a limiting factor of yield for oil tea in China [67,68]. Otherwise, mycorrhizal fungi were not dominant in the fungal community in this study, which also explained why phosphorous contents were positively correlated with fungal richness.
The dominant bacteria, Chloroflexi, Proteobacteria, Acidobacteria, Actinobacteria, and Planctomycetes, and fungi, Ascomycota and Basidiomycota, in the rhizosphere of C. yuhsienensis were similar to previous studies in C. oleifera rhizosphere or bulk soil [10,30], but a little different from the rhizosphere soil of Camellia sinensis [69]. These results indicated that nutrient utilization patterns of the plant might significantly alter microbial communities. There were significantly different responses between the rhizosphere and non-rhizosphere microbial communities to seasonal variation. During the entire year, the shared OTUs of the four seasons in the rhizosphere were higher than those in the non-rhizosphere for both bacteria and fungi (Figure 2). The PCoA also indicated that the effect of seasonal variation on bacterial community structure was smaller in the rhizosphere (R 2 = 0.328, p = 0.107) than in the non-rhizosphere (R 2 = 0.398, p < 0.001) ( Figure 3A). On account of the root activity of C. yuhsienensis, the bacterial community structure in the rhizosphere differed significantly from that in the non-rhizosphere (R 2 = 0.064, p = 0.048). Particularly in winter, the bacterial OTUs in the rhizosphere did not decrease due to the low temperature, as observed in the non-rhizosphere, but rather the number of exclusive OTUs increased (Figure 2A). The bacterial Sobs, Chao 1, and Shannon indexes were significantly higher in the rhizosphere than in the non-rhizosphere in winter. These results indicated that the roots of C. yuhsienensis might have higher excretions in winter, thus remarkably increasing the activity of bacteria in winter. This result is consistent with past studies that indicated that the roots of C. yuhsienensis are also active in winter [70,71]. The fungal community structures also varied greatly with seasonal variation ( Figure 3B). Across the entire year, fungi in the non-rhizosphere were more sensitive to seasonal variation than those in the rhizosphere (Table 1 and Figure 2B). Notably, there were slight differences between the rhizosphere and non-rhizosphere microbial communities, regardless of microbial community structure, microbial diversity, or microbial OTUs (Figures 1 and 2 and Table 1), and the number of fungal and bacterial biomarkers was lowest in summer in both the rhizosphere and non-rhizosphere ( Figure S3 and Figure S4). In summary, seasonal variation was the predominant effective factor influencing microbial diversity and structure, and the activity of C. yuhsienensis roots could provide a more stable environment for microorganisms, especially for bacteria in winter.
Exploring the rhizosphere microbial function is a great way to understand the interactions between plants and microorganisms. Although there are already some researchers focusing on microbial communities associated with oil tea [10,30], the microbial functions in the oil tea rhizosphere have never been noticed. Indicator analysis showed that the number of biomarkers in the rhizosphere was much higher than in the non-rhizosphere ( Figure 4). All of the bacterial biomarkers in the rhizosphere in this study were chemoorganotrophic Gram-negative bacteria that grow under aerobic conditions [72][73][74][75][76][77][78]. Some species of Rhodanobacter can degrade γ-hexachlorocyclohexane (lindane; γ-HCH) under aerobic conditions, and some species have the ability to antagonize root rot fungi [79] or to denitrify nitrate [80,81]. Most Labrys species have been reported as biodegraders of some harmful organic matter for animals, such as Labrys portucalensis [82,83]. Sphingomonas, a Gram-negative bacteria that was a biomarker in the rhizosphere, grows under aerobic conditions. Sphingomonas species are widely distributed throughout the soil and water and are associated with plant roots [84]. Sphingomonas species are multifunctional. Some exhibit antagonism against the phytopathogenic fungus Verticillium dahliae [85], while some are the pathogenic bacteria causing plant root disease [86], and some are able to degrade refractory pollutants [87]. The biomarkers reflect that the rhizosphere environment of C. yuhsienensis had high air permeability and abundant organic nutrients for these bacteria.
Most fungal biomarkers in the rhizosphere of C. yuhsienensis are beneficial fungi for plants. Trichoderma has been reported several times to protect plants and contain pathogen populations [88,89]. This is the first study to report Trichoderma in the rhizosphere of C. yuhsienensis. Mortierella is a large and diverse genus, and most species of Mortierella are saprobic and abundant in the soil and plant debris [90]. Arachidonic acid (ARA), produced by most Mortierella species, is an important constituent of biological cells [91]. ARA is not only beneficial for animals [92], but can also antagonize plant disease [93].
Lecanicillium species have great potential as biocontrol agents [94]. Most Lecanicillium species are able to parasitize many insects, such as Lecanicillium longisporum against aphids, Lecanicillium muscarium against whiteflies and thrips [95], and Lecanicillium psalliotae against the root-knot nematode Meloidogyne incognita [96]. Meloidogyne incognita is one of the major nematodes of oil tea roots [97].
In general, most biomarkers in the rhizosphere of C. yuhsienensis are beneficial for plants, and thus, we may speculate that the root activity of C. yuhsienensis markedly improved the rhizosphere conditions and increased the beneficial microorganisms. Functional analysis also verified this (Figures 5 and 6).

Conclusions
The dominant bacteria in the rhizosphere of C. yuhsienensis were Chloroflexi, Proteobacteria, Acidobacteria, Actinobacteria, and Planctomycetes, and the dominant fungi were Ascomycota, Basidiomycota, and Mucoromycota. Seasonal variation was the predominant factor influencing soil microbial abundance, diversity, and structure and the soil physiochemical properties in the rhizosphere of C. yuhsienensis. There were weak correlations between microorganisms and soil physiochemical properties in the rhizosphere. Seasonal variation had significant effects on bacterial abundance, but not on diversity and structure, and had significant effects on fungal abundance and structure, but not on diversity in the rhizosphere. Significantly higher bacterial abundance and diversity were observed in the rhizosphere than in the non-rhizosphere, which reflected the high activity of the C. yuhsienensis roots in winter. All of the bacterial biomarkers in the rhizosphere were chemoorganotrophic Gram-negative bacteria that grow under aerobic conditions, and the fungal biomarkers, including Trichoderma, Mortierella, and Lecanicillium, have a great ability to protect against pathogens in the rhizosphere. In the rhizosphere of C. yuhsienensis, the dominant functions of the bacteria were oxidative phosphorylation and metabolism of nitrogen, glycine, serine, threonine, glutathione, and sulfur in the rhizosphere, while the dominant fungal functional groups were endophytes and ectomycorrhizal fungi of the symbiotroph trophic type. The activity of C. yuhsienensis created better conditions with good air permeability and formed a microbial community dominated by beneficial bacteria and fungi. We strongly suggest that future research should pay greater attention to the root vitality and exudation of C. yuhsienensis in different seasons, which might be the key to a better understanding of the relationships between roots and microbial community. This is the first time to study C. yuhsienensis rhizosphere microbial communities during seasonal variations, and the first time to study Camellia species rhizosphere microbial function. Therefore, we strongly suggest that researchers pay more attention to the rhizosphere microorganisms and provide more scientific data to better study and manage Camellia plantations.
Supplementary Materials: The following are available online at http://www.mdpi.com/2076-2607/8/9/1385/s1, Figure S1: Annual variation of soil temperature in Camellia yuhsienensis Hu. forest. Figure S2: Dilution curves of Sobs index of bacteria (A) and fungus (C) and Shannon index of bacteria (B) and fungus (D). AR, BR, CR and DR indicates rhizosphere soil sample in spring, summer, autumn, and winter, respectively. AN, BN, CN and DN indicates non-rhizosphere soil sample in spring, summer, autumn, and winter, respectively. Figure S3: Least discriminant analysis effect size (LEfSe) analysis of bacterial community in spring (A), summer (B), autumn (C) and winter (D). Red parts of circle in each cladogram indicate the biomarkers in rhizosphere soil samples. Green parts of circle in each cladogram indicate the biomarkers in non-rhizosphere soil samples. Yellow solid circles indicate that the abundance of species have no significant difference in rhizosphere and non-rhizosphere. n = 3. Figure S4: LEfSe analysis of fungal community in spring (A), summer (B), autumn (C) and winter (D). Red parts of circle in each cladogram indicate the biomarkers in rhizosphere soil samples. Green parts of circle in each cladogram indicate the biomarkers in non-rhizosphere soil samples. Yellow solid circles indicate that the abundance of species have no significant difference in rhizosphere and non-rhizosphere. n = 3. Figure S5: Contribution of environmental factors on bacterial community at genus level in rhizosphere (A) and non-rhizosphere (B) and fungal community in rhizosphere (C) and non-rhizosphere (D) using Variance Partitioning Analysis (VPA). AP, available phosphorous; TP, total phosphorous; AK, available potassium; TK, total potassium; AMN, ammonium nitrogen; NN, nitrate nitrogen; AHN, alkaline hydrolyzable nitrogen; TN, total nitrogen; TOC, total organic carbon; SWC, soil water content; Tem, soil monthly mean temperature; C/N, ratio of total carbon to total nitrogen; C/P, ratio of total carbon to total phosphorous; N/P, ration of total nitrogen to total phosphorous. n=3. Figure S6: Heatmap of correlations between environmental factors and bacteria at genus level in rhizosphere (A) and non-rhizosphere (B) based on Pearson correlation coefficients with the two-tail test. Red squares indicate positive correlations. Blue squares indicate negative correlations. "*" and "**" indicate the significant correlations at 0.05 and 0.01 level, respectively. AP, available phosphorous; TP, total phosphorous; AK, available potassium; TK, total potassium; AMN, ammonium nitrogen; NN, nitrate nitrogen; AHN, alkaline hydrolyzable nitrogen; TN, total nitrogen; TOC, total organic carbon; SWC, soil water content; Tem, soil monthly mean temperature; C/N, ratio of total carbon to total nitrogen; C/P, ratio of total carbon to total phosphorous; N/P, ration of total nitrogen to total phosphorous. n = 3. Figure S7: Heatmap of correlations between environmental factors and fungus at genus level in rhizosphere (A) and non-rhizosphere (B) based on Pearson correlation coefficients with the two-tail test. Red squares indicate positive correlations. Blue squares indicate negative correlations. "*" and "**" indicate the significant correlations at 0.05 and 0.01 level, respectively. AP, available phosphorous; TP, total phosphorous; AK, available potassium; TK, total potassium; AMN, ammonium nitrogen; NN, nitrate nitrogen; AHN, alkaline hydrolyzable nitrogen; TN, total nitrogen; TOC, total organic carbon; SWC, soil water content; Tem, soil monthly mean temperature; C/N, ratio of total carbon to total nitrogen; C/P, ratio of total carbon to total phosphorous; N/P, ration of total nitrogen to total phosphorous. n = 3. Table S1: Soil physicochemical property of each sample. Table S2: Quantitative statistics of bacterial OTUs and Tags of each sample. Table S3: Quantitative statistics of fungal OTUs and Tags of each sample.