Change of rhizospheric bacterial community of the ancient wild tea along elevational gradients in Ailao mountain, China

The rhizospheric microbial community is one of the major environmental factors affecting the distribution and fitness of plants. Ancient wild tea plants are rare genetic resource distributed in Southwest China. In this study, we investigated that rhizospheric bacterial communities of ancient wild tea plants along the elevational gradients (2050, 2200, 2350 and 2500 m) in QianJiaZhai Reserve of Ailao Mountains. According to the Illumina MiSeq sequencing of 16 S rRNA gene amplicons, Proteobacteria, Acidobacteria and Actinobacteria were the dominant phyla with the relative abundance 43.12%, 21.61% and 14.84%, respectively. The Variibacter was the most dominant genus in rhizosphere of ancient wild tea plant. Phylogenetic null modeling analysis suggested that rhizospheric bacterial communities of ancient wild tea plants were more phylogenetically clustered than expected by chance. The bacterial community at 2050 m was unique with the highest alpha diversity, tend to cluster the nearest taxon and simple co-occurrence network structure. The unique bacterial community was correlated to multiple soil factors, and the content soil ammonium nitrogen (NH4+-N) was the key factor affecting the diversity and distribution of bacterial community along the elevational gradients. This study provided the necessary basic information for the protection of ancient tea trees and cultivation of tea plants.

the northern slope of Changbai Mountains was clearly influenced by soil pH value and content of nitrate nitrogen of soil 12 .
Recently, a few studies revealed the microbial community of tea rhizosphere and its interaction with environmental factors. For instance, Pandey and Palni 13 found that species of Bacillus dominated the bacterial population in the established rhizosphere of tea bushes in the Himalayas region. Another report found that the dominant bacteria species of ancient tea tree is Streptomyces 14 . Research of Arafat 15 on the south coast of Fujian province showed that the most dominant phyla in tea fields were Proteobacteria and Acidobacteria, and the bacterial taxa were affected by soil pH reduction during long-term tea cultivation. Besides soil pH, the soil organic carbon and available phosphorus meaningfully affected the rhizospheric microbial communities associated with long-term monoculture tea orchards was indicated by Li 16 . Most of these studies have documented the community assembly process and diversity of rhizospheric microbial communities in the well cultivated tea tree 17 , while, the rhizospheric microbiome studies on the endemic ancient wild plants, especially ancient wild tea community, are still largely unknown.
In our previous study, EST-SSR were used to study genetic diversity of ancient wild tea populations at different elevations in Qianjiazhai Reserve 18 . We found that the genetic characteristics of ancient wild tea were affected by the habitat heterogeneity in different elevation. Therefore, the elevational gradients of QianJiaZhai Reserve serve as ideal study areas 18 , for answering questions on how the rhizobacteria of the ancient wild tea community be affected by various environmental factors. In this study, we revealed the rhizospheric soil of ancient wild tea community along the gradient in QianJiaZhai Reserve of Ailao Mountains, China. We analyzed the physicochemical characteristics of rhizospheric soil, the rhizobacterial diversity and community composition of the ancient wild tea at four different elevations (2050, 2200, 2350 and 2500 m). Our results not only can provide the basic and necessary information to better understanding the ecological adaptation of ancient wild tea trees, but also can give theoretical reference for tea plant cultivation and germplasm resources conservation of wild ancient tea tree under the background of climate change and anthropogenic activities.

Result
Soil physiochemical properties. Elevation with 2050 m had the lowest values of total phosphorus (TP), available phosphorus (AP), soil organic matter (SOM), total carbon (C), total nitrogen (N), carbon/nitrogen ratio (C/N), ammonium nitrogen (NH 4+ -N) and nitrate nitrogen (NO 3-N) compared with the other three higher elevations according to the One-way ANOVA test, and most of those edaphic factors had the highest values at 2500 m elevation except NH 4 + -N and NO 3 -N ( Table 1). The value of NH 4 + -N is lowest at 2050 m (29.83 mg/kg), reaching a maximum at 2200 m (48.95 mg/kg), and then decreasing with elevation increased (Table 1). Similarly, the lowest and highest value of NO 3 -N was observed at 2350 m (14.25 mg/kg) and 2200 m (20.62 mg/kg), respectively ( Table 1). The soil pH was acid in the sampled soil and varied from 4.56 to 4.64, there was no significant difference among the samples selected form different elevations.
The value of TP, AP, SOM, N, C and C/N ratio presented a significantly positive relationship with elevation (P < 0.05), whereas no significant relationships with elevation detected on soil pH, NH 4 + -N, NO 3 -N or soil moisture (Table S1). Both NH 4 + -N and NO 3 -N were significantly correlated with organic nitrogen and carbon (P < 0.05). Moreover, the NH 4 + -N was positively correlated with C/N ratio while the NO 3 -N was positively correlated with TP (P < 0.05), respectively. In addition, the correlation results showed that there was a close relationship between TP and soil pH (P < 0.05) (Table S1).
Based on variance analysis (One-way ANOVA) and Duncan's multiple range test at P < 0.05. We found the Chao1 estimator and phylogenetic diversity (Faith's PD) at the lowest elevation (2050 m) significantly higher than those of other elevations (P < 0.05, Table S2). However, the species richness and diversity indices between 2200 and 2350 m, 2350 and 2500 m were not pronounced (Table S2). Mean pairwise distance (MPD) and mean nearest taxon distance (MNTD) had no significant differences at elevations, while showed highest values at 2200 m site (Fig. 2). Both of the standardized effect sizes of MNTD and MPD (ses.MNTD and ses.MPD) values obtained using the null model were significantly negative. The similar variation of ses.MPD values at elevations is consistent with that of MPD, and the standardized metric showed a variation at elevations that was different from that of MNTD (Fig. 2). The Non-metric multidimensional scaling (NMDS) ordination showed that the members of the 2050 m site were clearly separated from the other clusters, and the difference/variation between 2050 m and 2200 m was the most highly ( Fig. 3).
According to the topological properties of the co-occurrence networks, community complexity of bacteria nadired at the 2050 m, as visible as the lowest number of nodes and edges ( Fig. 4; Table S3). The average degree and the number of nodes and edges were the highest at 2500 m, and the clustering coefficient was the highest at 2200 m. These nodes belong to 17 bacteria phyla, with Actinobacteria, Chloroflexi, Proteobacteria, Acidobacteria were mainly found at sites.

Correlation between bacterial communities and soil physicochemical properties. Pearson cor-
relations showed that phylogenetic diversity (Faith's PD) and the Chao1 estimator were negatively correlated with soil N (P < 0.01), soil C (P < 0.01) and soil NH 4 + -N (P < 0.05) (Table S4). Meanwhile, we used linear regression analysis to study the relationship between the alpha diversity (Faith's PD and the Chao1 estimator) and soil characteristics. Both the Faith's PD and the Chao1 estimator were decreased as the increase of N, C and NH 4 + -N contents (Fig. 5). The stepwise multiple regression analysis showed that the ses.MPD was mainly affected by NH 4 + -N (R 2 = 0.28, F = 7.00, P = 0.02), and the ses.MNTD was related to N and soil pH (R 2 = 0.40, F = 5.69, P = 0.01; Table S5).

Discussion
Plants constantly adapt and change the soil during the process of succession. In this study, the contents of most the soil physiochemical in the rhizosphere of the ancient wild tea are lowest and highest at 2050 m and 2500 m, respectively (Table 1, S1). At 2050 m, compared to other higher elevations, the ancient wild tea community was more susceptible to human activities (grazing, logging, etc.) which might be reduced litter and result in hardening of soil, further lead to a decrease in soil nutrients. However, the lowest and highest contents of NH 4 + -N was observed at 2050 m and 2200 m, respectively, and NO 3 -N also showed the highest value at 2200 m (Table 1)  www.nature.com/scientificreports www.nature.com/scientificreports/ soil pH (4.5-6.0) for tea plants. As previous studies reported, soil acidification might be caused by leaf litter and exudates of tea plants such as organic acids, carbonic acid and polyphenols 21,22 .
In this study, Illumina sequencing analyses revealed that the rhizospheric bacterial communities of the ancient wild tea were largely dominated by Proteobacteria (43.12%), Acidobacteria (21.61%) and Actinobacteria (14.84%) at four elevation sites, and it is similar to the previous study on tea orchard 23 . In our study, the dominance of these three phyla in the rhizosphere of ancient wild tea was further confirmed in the genera level (Fig. 1b,d; S1). However, Arafat 16 showed that Proteobacteria and Bacteroidetes were dominant rhizospheric bacterial phylum in the 30-year-old tea plantation, which might be related to the age and environment conditions of the tea tree community 24,25 . To the best of our knowledge, the Proteobacteria play a major role in the global carbon, nitrogen and sulfur cycling 26 . The phylum Proteobacteria was the most dominant phyla in the soil bacteria obtained from various soils ranging from agricultural land and grassland to pristine forest 27 , and was the main rhizosphere bacteria of rice, Allium tuberosum and Prosopis articulate [28][29][30] . In addition, Actinobacteria was important for decomposition of organic matter to improve the quality of agricultural soil 31 , and it may prefer acidic soil as moderately acidophilic heterotrophic bacteria 32 . Variations of rhizospheric bacterial phyla of the wild ancient tea tree were not significant along the elevation gradient. However, some genera belonging to the dominant phylum, which showed significant variations (P < 0.05) along elevations (Fig. 1f, S1). For instance, the relative abundances of Variibacter increased along the increased elevation from 2050 m to 2350 m while decreased at 2500 m (Fig. 1f). The Variibacter, strictly aerobic, Gram-negative, was isolated from soil of the lava forest, Gotjawal, located in Aewol, Jeju, Korea 33 . It has also been reported as a dominant genus form the soil in karst, saline lake and forests from the region surrounding karst plateau lakes in Guizhou Province 34 . In our study, for the rhizospheric bacterial communities of ancient wild tea trees, the dominant phylum is similar while the dominant genera are different at different elevations in Qianjiazhai Reserve. These unique genera taxa in order to better adapt to the corresponding habitat among elevations.  www.nature.com/scientificreports www.nature.com/scientificreports/ According to the topological properties of the co-occurrence networks, community complexity of bacteria nadired at the 2050 m, as visible as the lowest number of nodes and edges ( Fig. 4; Table S3). The complex interactions within microbial communities may contribute to the poor organic matter and insufficient nutrients in the rhizosphere of tea trees at 2050 m, which have few niches and thus with fewer opportunities for interaction between different species 35,36 . These nodes are mainly bacterial phylum with higher relative abundance, such as Actinobacteria, Chloroflexi, Proteobacteria and Acidobacteria. These results revealed that the dominant bacterial phyla in the ancient tea tree bacterial community may play critical ecological functions relating to the construction of the bacterial community 37 .
Elevation is correlated with the variables that affect the ecosystem and play a key role in influencing microbial diversity. In this study, although we found the significantly highest rhizosphere bacterial diversity (the Chao1 estimator and phylogenetic diversity) in the ancient wild tea at the lowest elevation site 2050 m, and did not found an elevational diversity gradient pattern in QianJiaZhai Reserve. As previously reported in soil bacterial studies, for example, Shen 38 reported that soil bacterial diversity exhibited no pattern among elevational gradient on Changbai Mountain, the research of Debnath 11 on Rhododendron arboreum along Eastern Himalayan Slope in Tawang came to the same conclusion. Our result is inconsistent with the findings that the unique stair-step pattern observed in Gongga Mountain 39 or unimodal pattern in alpine tundra 40 about soil microorganisms, the small saddle pattern of rhizospheric bacteria with Stellera chamaejasme 41 . A likely explanation for this differs was that the sampling sites of different elevational gradient, that is, requires broader and more fine-grained spatial extent during sampling 11,39 . On the side, inconsistent plant categories and methods of microbial community analysis may also influence the survey result 41 . Both of the standardized effect sizes of mean nearest taxon distance and mean pairwise distance (ses.MNTD & ses. MPD) values, obtained using the null model were significantly negative (ses.MNTD & ses.MPD < 0; Fig. 2), indicating that bacterial community assembly process tended to be more phylogenetically clustered than expected by chance, which suggested that environment plays a strong ecological filtering role in the process of community construction 42 . Previous studies have reported similar results 40,43 . Furthermore, the MNTD and ses. MNTD indicated that large differences in the phylogenetic clustering of bacteria between 2050 m www.nature.com/scientificreports www.nature.com/scientificreports/    www.nature.com/scientificreports www.nature.com/scientificreports/ and 2200 m and there were more assemblages that share many closely related species at 2050 m. This might be caused by the different combinations of bacterial taxa or complex elevational habitats.
We found the rhizospheric bacterial diversity was significantly correlated with rhizospheric soil N, C, and NH 4 + -N contents based on linear regression analysis and Pearson correlation analysis. Previously, many studies reported that soil pH is the crucial abiotic factor determining the changes in soil bacterial diversity along elevation [44][45][46] . Nevertheless, the lack of significant correlation between soil pH and bacterial diversity along elevations in this study may be pertaining to the narrow soil pH ranges (4.56-4.64) observed. Intriguingly, the vast majority of previous studies mentioned that soil carbon and nitrogen were two of the abiotic soil factors affecting bacterial diversity along elevations 38,39,45,47 , whereas only a few reported that soil C and N were the most prominent abiotic soil factors rather than soil pH or any other. For example, according to a prior study in alpine tundra, soil total carbon content and C:N ratio were the most prominent factors influencing bacterial diversity patterns 40 . Our findings suggested that rhizosphere bacterial alpha diversity of wild ancient tea among elevation in QianJiaZhai Reserve might be driven by the contents of rhizosphere nitrogen (Total N and NH 4 + -N) and total carbon. The stepwise multiple regression analysis showed that the ses.MPD was mainly affected by NH 4 + -N (R 2 = 0.28, F = 7.00, P = 0.02), which revealed that NH 4 + -N could explain 28% of ses.MPD variation. The MPD, mainly reflects the deep phylogenetic structure in a phylogeny 48 , is typically thought to be more sensitive to tree-wide patterns of phylogenetic clustering or overdispersion than to the structure near the tips 49 . Furthermore, the ses.MNTD was related to N and soil pH (R 2 = 0.40, F = 5.69, P = 0.01; Table S5). As such, both low soil N content and high soil pH might lead to the clustering of nearest taxon at 2050 m. Shen et al. also showed that one of the reasons for the phylogenetic clustering at different elevations might be closely related to sharply low nitrogen contents 40 . In our study, NMDS analyses indicated that the bacterial communities of the 2050 m site were clearly separated from the other clusters (2200, 2350 and 2500 m sites) (Fig. 3). According to the Partial Mantel test, the bacterial community structure at four elevational sites was highest associated with rhizosphere soil NH 4 + -N, followed by soil N, C, and elevation ( Table 2). According to the redundancy analysis (RDA), NH 4 + -N also had the greatest impact on the classification of bacterial phylum and genera levels of the ancient wild tea plants on the elevation gradient (Fig. 6a,b). Thus, rhizospheric soil NH 4 + -N is the predominant abiotic soil factor in the rhizospheric bacterial community structure of ancient wild tea trees among the elevation gradient in QianJiaZhai Reserve. With respect to soil NH 4 + -N, which was one of the prominent abiotic soil factors that affected bacterial communities at 0-5 cm soil depth along the elevation gradients in Tibetan Plateau 50 . In another prior study, a significant correlation between soil NH 4 + -N and bacterial communities in rhizospheres of Rhododendron arboretum along Eastern Himalayan Slope in Tawang was been also found 11 . Besides NH 4 + -N, rhizosphere soil N, C, soil pH and SOM also play important roles in the structure of bacterial communities on elevation gradients as reported in previous studies ( Table 2, Fig. 6a,b). For instance, total nitrogen was one determining factor in distributions of bacterial community at relatively high elevation of Himalayan Tawang region 11 . Both total nitrogen and carbon were significantly related to bacterial community structure along the elevational gradients of 1800-4100 m on Gongga Mountain 39 . In addition, Shen 40 found that soil carbon and nitrogen were the most important factors in structuring bacterial communities across a small-scale elevational gradient in alpine tundra. Taken together, rhizospheric soil NH 4 + -N, soil N, C, soil pH and SOM appeared to have effects on bacterial community structure of ancient wild tea among elevational gradient in QianJiaZhai Reserve, with rhizospheric soil NH 4 + -N being the key influencing soil factor by Partial Mantel test and RDA.
For most bacterial phyla (relative abundance> 1%), only one soil property variable was retained in the stepwise regression equation (Table S6). As such, the results show that in the rhizosphere of ancient tea trees, most bacterial phyla (relative abundance> 1%) changed along elevation are driven by only a few soil properties. For instance, the relative abundances of Proteobacteria were positively correlated with soil pH, while Acidobacteria showed the opposite pattern, although only the mild fluctuations (soil pH ranged from 4.56 to 4.64), which concurred with other reports 39,51,52 . These results suggested that soil pH was considered as an important factor of the rhizosphere bacterial community composition in the ancient wild tea by affecting the dominant phylum (Proteobacteria and Acidobacteria). The relative abundance of Actinobacteria was related to rhizospheric soil NO 3 -N and NH 4 + -N (R 2 = 0.39, F = 5.38, P = 0.02; Table S6). A similar result was reported by previous studies in an acidic forest soil 53 . In addition, both the relative abundances of Nitrospirae and Chloroflexi were negatively associated with the contents of rhizospheric soil NH 4 + -N. Taken together, a higher soil NH 4 + -N content would reduce the abundance of some taxa (e.g., Chloroflexi); however, copiotrophic groups (e.g., Proteobacteria and Actinobacteria) could have high richness. These findings were previously explained by the copiotrophic hypothesis 51 . In summary, the aforementioned result demonstrates that rhizosphere inorganic nitrogen (NH 4 + -N&NO 3 -N) and the soil pH value play a key role in rhizospheric bacterial community phyla of the ancient wild tea in QianJiaZhai Reserve.
In conclusion, the rhizosphere bacterial community of the ancient wild tea tree is unique in diversity, phylogeny and bacterial network structure at 2050 m site, which is attributed to the interaction of multiple soil factors, especially ammonium nitrogen. The NH 4 + -N is a key driving soil factor affecting the rhizosphere bacterial community of wild ancient tea trees on the elevation gradients, as well as the relative abundance of some specific bacterial phyla/genera such as Proteobacteria, Actinobacteria, Chloroflexi, Bryobacter and Acidibacter. Our study is the first time to examine the rhizospheric bacterial community of such ancient wild tea tree (DBH approximately 0.3 m, with probably more than 1000 years old), which is the dominant vegetation species in QianJiaZhai Reserve. Hence, our findings will provide researchers the necessary basic information to better understanding the ecological adaptation of ancient wild tea trees. In future, more research needs to be focused on integrative studies in soil microbiology, to reveal the "rhizosphere effect" of ancient wild tea plants. (2020) 10:9203 | https://doi.org/10.1038/s41598-020-66173-9 www.nature.com/scientificreports www.nature.com/scientificreports/

Materials and Methods
Site selection and soil sampling. The QianJiaZhai Reserve (101°14′E; 24°17′N) is located in Yun-nan Ailao Mountains (Zhen-yuan County), Southwest of China. The annual average temperature here is 10-12 °C, and the precipitation is more than 1500 mm. The frost period is from November to March the next year, and the snow period is from December to February the next year. The terrain in the region is complicated, and there are ancient wild tea trees along the elevational gradients (2100-2500 m). As the dominant species in the region, the diameter at breast height of wild ancient tea was frequently more than 0.3 m (ca. 1000 years old). Besides the wild ancient tea, the subdominant species are Lithocarpus xylocarpus and Manglietia insignis (Table S7) 54 .
We set four sites to cover the entire elevation range of wild ancient tea: 2050 m, 2200 m, 2350 m, and 2500 m (Table S7). Five sampling plots were randomly chosen within each elevation along transects of 150 m length. At each sampling plot, mixed rhizospheric soil samples (according to Debnath 11 ) were taken from three randomly selected tea plants (the diameter at breast height was about 0.3 m) and separated by at least 5 m from the others. In total, 20 samples were collected and placed in labeled sterile polypropylene bags in an icebox and transported to the laboratory for analysis. Upon arrival, one part of soil samples is spread out on clean paper to dry in a ventilated room and storage for subsequent element analysis. Another part stored at −80 °C for assessing the rhizospheric bacterial communities.
Physico-chemical characteristics of soil. All rhizospheric soil subsamples were naturally air-dried at indoor and subsequent analysis was performed after screening with a 2 mm sieve. Methods for measuring soil pH with a pH meter: mixing the soil sample and KCl solution (1:2.5 wt/vol) after vigorous shaking and suspending for 1 h. Soil moisture determined was gravimetrical. For the determination of total carbon and total nitrogen the concentrations, the elemental analyzer (Vario MAX, Elementar, Germany) was used. The NO 3 − -N and NH 4 + -N were measured according to standard methods 55 . Other soil chemicals: available phosphorus 56 , and organic matter content 57 were also determined. Processing of illumina sequencing data. Paired-end reads were merged by the FLASH software 58 .
Sequences obtained by Illumina Miseq sequencing were processed and analyzed with the Quantitative Insights into Microbial Ecology pipeline (QIIME) 59 . Using the USEARCH algorithm to identified and removed Chimeric sequences 60 . Simultaneously, in order to rectify the influence of widely different sequencing depth, we rarefied entire samples to the equally sequencing depth by stochastic chosing subset of 25890 sequences from each soil sample. The sequences obtained across samples were specified as operational taxonomic units (OTUs) at 97% analogy. All representative sequences were classified taxonomically according to Silva database.
Diversity and statistical analysis. The Chao1-richness 61 and Faith's phylogenetic diversity (PD) 62 were calculated by with the Mother software (v.1.30.1) to estimate the bacterial alpha diversity. The function ses.mpd and ses.mntd were used to calculate mean pairwise distance (MPD) and mean nearest taxon distance (MNTD) 49 , respectively. We used a network analysis method to visualize the co-occurrence of community structure, characterize intra-community interactions and identify potential shared niches 37 . Non-metric multidimensional scaling analysis (NMDS) tests of Bray-Curtis similarity distance values at the OTU level was used to observe if the samples at different elevational sites formed unique phylogenetically related clusters. Linear regression analysis and Pearson correlation analyses were employed to determine the relationships between abiotic rhizospheric factors and alpha diversity. Partial Mantel based on Bray-Curtis similarities were used to depict the relationships between bacterial community structure and rhizospheric soil factors with Qiime software, while the redundancy analysis (RDA) were carried to further reveal rhizospheric soil elements contribution degree to bacterial at the four sites using CANOCO version 5. In addition, before the RDA analysis, a detrended correspondence analysis for the specific microbial groups was performed to confirm that the linear ordination method was appropriate for the analyses (gradient lengths<3). Used stepwise multiple regression to find the most important some factors related to MPD, MNTD and the relative abundance of bacteria classified at the phyla and genus levels.

Data availability
Illumina MiSeq amplicon sequences of 16 S rRNA genes of bacteria are available in the NCBI Sequence Read Archive (www.ncbi.nlm.nih.gov/sra) as BioProject PRJNA629666.