Rhizosphere bacteria community and functions under typical natural halophyte communities in North China salinized areas

Soil salinity is a serious environmental issue in arid China. Halophytes show extreme salt tolerance and are grow in saline-alkaline environments. There rhizosphere have complex bacterial communities, which mediate a variety of interactions between plants and soil. High-throughput sequencing was used to investigated rhizosphere bacterial community changes under the typical halophyte species in arid China. Three typical halophytes were Leymus chinensis (LC), Puccinellia tenuiflora (PT), Suaeda glauca (SG). The dominant phyla were Proteobacteria, Actinobacteria, Chloroflexi, Gemmatimonadetes, Acidobacteria and Bacteroidetes, Suaeda glauca rhizosphere has stronger enrichment of Nitrospirae and Cyanobacteria. The Ace, Chao and Shannon indices were significantly higher in soils under LC and SG (P<0.05). Functional predictions, based on 16S rRNA gene by PICRUSt, indicated that Energy metabolism, Amino acid metabolism, Carbohydrate metabolism and Fatty acid metabolism are dominant bacterial functions in three halophytes rhizosphere soil. Carbon metabolism, Oxidative phosphorylation, Methane metabolism, Sulfur metabolism and Nitrogen metabolism in SG were significantly higher than that in LC and PT. Regression analysis revealed that rhizosphere soil bacterial community structure is influenced by soil organic matter (SOM) and soil water content (SWC), while soil bacterial community diversity is affected by soil pH. This study contributes to our understanding of the distribution characteristics and metabolic functions under different halophyte rhizosphere bacterial communities, and will provide references for the use of rhizosphere bacteria to regulate the growth of halophytes and ecological restoration of saline soil.


Introduction
Soil salinity is one of the rising environmental issues causing considerable yield losses worldwide especially in arid and semiarid regions [1]. It damages the soil structure, reduces soil quality and limits the growing of crops [2]. Halophytes rhizosphere sampling, the roots were gently shaken to remove the loosely adhering soil. The soil attached to the roots, which is considered to be the rhizosphere soil, was collected by using sterile brushes [23]. Rhizosphere samples of each replicate were put in individual sterile plastic bags and transported to the laboratory immediately. After visible stones and plant residues were removed, rhizosphere soil was sieved (2 mm mesh) and separated into two subsamples, one portion was air-dried for the determination of chemical analysis, and the reminder was stored in a -20˚C refrigerator for molecular analysis.

Soil properties determination
Soil pH was measured with a compound electrode (INESA Scientific PHSJ-3F) using a soil to water ratio of 1:5 [24]. Electrical conductivity (EC) was measured using a soil-water suspension (1: 2.5 soil-water ratio) [25]. Soil water content (SWC) was measured after drying in an oven at 105˚C for 24 h [26]. Soil organic matter (SOM) was determined by oxidizing organic C with potassium dichromate (K 2 Cr 2 O7) [27].

DNA extraction and PCR amplification
Soil DNA was extracted from 0.5 g fresh subsamples using Soil DNA Kit (Omega Bio-Tek Inc., Norcross, GA, USA), according to the manufacturer's protocols. The quality of extracted DNA was assessed by 1% agarose gel electrophoresis and substandard samples were extracted again until all the samples passed the quality control. All extracted DNA samples were stored at -20˚C for further analysis. The V4-V5 hypervariable regions of the soil bacterial 16S rRNA gene were subjected to high-throughput sequencing by Majorbio Pharmaceutical Technology Co., Ltd. (Shanghai, China) using PE300 sequencing platform (Illumina, Inc., San Diego, CA, USA). The V4-V5 bacterial 16S rRNA gene were amplified by PCR using the primers pair 515 F/907R. The PCR program was as follows: denaturation at 95˚C for 30 s; annealing at 5˚C for 30 s; extension at 72˚C for 45 s; 27 cycles; holding at 72˚C for 10 min and storing at 10˚C. In order to guarantee the accuracy of the analysis results, Quantitative Insights into Microbial Ecology (QIIME software; version1.9.0) was used for sequence filtering, and then the chimeric sequence was removed using Mothur software to obtain a high-quality sequence for subsequent analysis [28]. The qualified sequences were clustered into operational taxonomic units (OTUs) with a 97% similarity cut-off using the Usearch program (version 7.0; http:// drive5.com/uparse/). The taxonomy of each OTU was analyzed by RDP classifier algorithm against the Silva (SSU123) 16S rRNA database (SILVA ribosomal RNA database project; Max Planck Institute for Marine Microbiology and Jacobs University, Bremen, Germany) using confidence threshold of 70%.

Data processing and statistical analysis
Venn diagrams were used to compare the similarities between soil bacterial communities at OTU level. According to the species abundance of each sample in the OTU list, soil bacterial richness indices (Chao and Ace) and diversity index (Shannon) were calculated by using Mothur software [29]. Analysis of similarity (ANOSIM) and nonmetric multidimensional scaling (NMDS) based on the unweighted UniFrac distance were used to reveal changes in soil bacterial community structures [30]. The influences of soil properties on soil bacterial community structures were examined by Mantel test and distance-based redundancy analysis (Db-RDA). The influences of soils properties on soil bacterial community structures were examined by Linear Regression. PICRUSt (Phylogenetic Investigation of Communities by Reconstruction of Unobserved States) was performed to predict the abundance of main metabolic function genes [19]. The functional genes were predicted from the Kyoto Encyclopedia of Genes and Genomes (KEGG) catalogue [26]. SPSS version 20.0 (SPSS Inc., Chicago, USA) was utilized to data statistics and analysis of variance to determine the difference between soil properties, bacterial community and bacterial function spectrum. R was used to perform and drawn Venn diagram, NMDS, ANOSIM, db-RDA and Linear Regression analysis [31]. The specific functional differences were performed using Origin version 9.0 (Microcal Software, Inc., Northampton, MA, USA).

Soil properties
The changes of rhizosphere soil properties under different halophytes were shown in Table 2. Significant differences of soil pH were found between rhizosphere samples from the different halophytes (P < 0.05, Tukey's test). PT (Puccinellia tenuiflora (Griseb.) Scribn. et Merr.) had the highest soil pH, whereas LC (Leymus chinensis (Trin.) Tzvel.) had the lowest pH value. Soil EC significantly (P < 0.05) varied from 0.45 to 1.44 (mS�cm -1 ) across the rhizosphere samples, and LC hadthe highest values. There was a significant variation in SOM under different halophytes rhizosphere (P < 0.05). The highest SOM content was found in rhizosphere samples from LC, which was 84.31% higher than that of SG (Suaeda glauca (Bunge) Bunge.). SWC followed the order PT > LC > SG, and there was no significant difference between LC and PT (P > 0.05).

Soil bacterial community composition and diversity
A total of 38 bacterial phyla were observed from rhizosphere samples. Proteobacteria, Actinobacteria, Chloroflexi, Gemmatimonadetes, Acidobacteria, Bacteroidetes, Planctomycetes, and Firmicutes were dominant phyla through entire soil samples ( Fig 1A). The highest relative abundance of Proteobacteria, Chloroflexi and Gemmatimonadetes were observed in PT, which were 27.25%, 13.07% and 13.48%, respectively. However, the lowest relative abundance of Proteobacteria (25.37%), Chloroflexiand (11.16%) and Gemmatimonadetes (8.61%) were found in SG, and the relative abundance of Actinobacteria (19.16%) and Acidobacteria (9.45%) were highest in three rhizosphere samples. The relative abundance of Planctomycetes, Entotheonellaeota and Hydrogenedentes in LC were significantly higher than that in PT and SG (S1a Fig, P < 0.05, Tukey's test). The relative abundance of Nitrospirae, Cyanobacteria, Deinococcus-Thermus, Armatimonadetes and BRC1 in SG were significantly higher than those of LC and PT. To further compare the differences in bacterial community composition between rhizosphere samples under different halophytes, we conducted the relative abundance of bacteria (> 0.02%) at the class level ( Fig 1B). The dominant classes were Actinobacteria, Alphaproteobacteria, Gemmatimonadetes, Gammaproteobacteria and Anaerolineae in LC, PT and SG. The relative abundance of Planctomycetacia and Rhodothermia in LC were significantly higher than that in PT and SG (S1b Fig, P < 0.05). The relative abundance of Anaerolineae, Deltaproteobacteria and NC10 in PT were significantly higher than that in LC and SG. The relative abundance of Chloroflexia, Nitrospira, Oxyphotobacteria, Deinococci and Gitt-GS-136 in SG, which were significantly higher than of LC and PT. Venn diagrams indicated that the sum of total observed OTUs in the rhizosphere samples from three halophytes was 10,154 (S2 Fig), and 1,413 OTUs were shared by all three groups. The numbers of OTUs co-occurred in LC, PT and SG were 1,232, 522 and 931, respectively.
To determine whether bacterial community structure was associated with alteration of different halophytes rhizosphere samples, we profiled the overall structural changes of bacterial community by using NMDS based on unweighted UniFrac dissimilarities (Fig 2A). NMDS ordinations showed rhizosphere samples from three regions are accumulated together respectively, samples in PT and SG separated with LC along NMDS 1. The ANOSIM results showed that the differences in the bacterial community structure of three halophytes rhizosphere samples are greater than the differences within the groups (  Soil bacterial diversity is essential for measuring soil micro-ecological health. Shannon, Ace and Chao indices in PT were significantly lower than that in LC and SG (Fig 3, P < 0.05), but there was no significant difference between PT and SG (P > 0.05).

Correlation between bacterial communities and soil properties
Db-RDA analysis was used to study the influence of soil properties on bacterial community composition at phylum level (Mantel test, R = 0.28, P = 0.01). It was difference that the effects of three rhizosphere soil properties on rhizosphere bacterial communities (Fig 4). It is apparent from LC's chart that the two axes of CAP axis explain 63.92% of total variation (Fig 4A), and soil pH, EC, SOM and SWC in LC were No correlation with the CAP axis (S1 Table). Rhizosphere soil properties in PT and SG had a stronger influence on bacterial communities. Soil EC, pH, SOM and SWC were longer arrows. In Fig 4B, two axes of CAP axis explain 70.34% of total variation and soil EC (P = 0.004), pH (P = 0.022), SOM (P = 0.019) and SWC (P = 0.017) were closely correlated to the CAP axis. As for SG, two axes of CAP axis explain 58.59% of total variation, and soil EC (P = 0.023), pH (P = 0.017) and SOM (P = 0.019) were closely correlated to the CAP axis ( Fig 4C). We explored the influence of pH, EC, SOM and SWC on bacterial community composition through Linear regression analysis (Fig 4). The PCA plots  showed that changes in bacterial communities at phylum levels along PCA1 are closely correlated with SOM and SWC, and indicated that the bacterial community structures are influenced by SOM and SWC. In addition, soil bacterial diversity was significantly affected by pH (Fig 5 and Table 3).

Soil potential bacterial functions
We used PICRUSt to predict the rhizosphere bacterial community function under three halophytes. The classification of KEGG functions at pathway Level1 was performed in S3 Fig. The result showed that the predicted functions mainly the bacterial metabolism at pathway Level1, and the relative abundance of three halophytes was above 78%, and SG was higher than LC and PT. Carbohydrate metabolism, Amino acid metabolism, Energy metabolism and Lipid metabolism were main metabolic functions at pathway Level2 ( Fig 6A). Interestingly, the abundance of these metabolism in SG were significantly higher than that in LC and PT (P < 0.05, Tukey's test). At pathway Level3, Carbon metabolism, Oxidative phosphorylation, Carbon fixation pathways in prokaryotes, Methane metabolism, Sulfur metabolism, Nitrogen metabolism were no significant difference between LC and PT (Fig 6B, P > 0.05), which was significantly lower than those in SG (P < 0.05).

Soil properties
Plants, rhizosphere bacterial communities and soil properties have a complex micro-ecological relationship, any changes may affect the bacterial structure and ecological processes [32]. Our study found that soil properties, such as pH, EC and SOM were changed significant in soil

Soil bacterial community composition and diversity
The rhizosphere microbes play an important role in promoting plant survival under adverse plant communities [5]. Proteobacteria, Actinobacteria, Chloroflexi and Gemmatimonadete were the dominant phyla in this study. These phyla have also been recorded in other halophytes rhizosphere reported that Proteobacteria accounting for over 40% in samples from Glaux maritima and Salicornia europaea [13]. We found that the relative abundance of Proteobacteria was exceed 20% in three halophytes rhizosphere samples (Fig 1A). The highest relative abundance of Proteobacteria was observed in Puccinellia tenuiflora, with an increase of 0.89% and 1.88% compared to Leymus chinensis and Suaeda glauca, respectively. The important role of Actinobacteria secrete antimicrobial pathogens compounds and Chloroflexi decompose organic compounds has been performed in previous studies [43]. In contrast with Puccinellia tenuiflora, the relative abundance of Chloroflexi in Leymus chinensis and Suaeda glauca decreased by 0.82% and 1.91%, respectively. However, the relative abundance of Actinobacteria was no significant difference in three halophytes rhizosphere samples. A possible explanation that Actinobacteria widely present in moderate and high salinity environments [44]. Arthrobacter and Kocuria were belongs to Actinobacteria, which contain cellwall-bound amino acids [45]. This indicated that enriched Proteobacteria, Actinobacteria and Chloroflexi in the rhizosphere help three halophytes to adapt to the salty conditions. Plant species certainly affect the structure of bacterial communities, and select specific microbial populations [46]. Specific selection leads to differences in the rhizosphere bacterial communities of halophytes, and confirmed in our research (S1 Fig). Nitrospirae and Cyanobacteria were significantly enriched in the rhizosphere of Suaeda glauca. According to reports, Nitrospirae and Cyanobacteria play an important role in soil C and N cycle and maintain soil fertility [47]. Currently, plant growth-promoting rhizobacteria have been widely used as biological fertilizers to increase the growth and yield of various crops [48]. These results indicate that rhizosphere bacteria have important assistance for halophytes adapt to saline soil, and the rhizosphere of three halophytes has different enrichment of bacteria. Therefore, we can select different halophytes for planting according to the saline environment, and screen the functional rhizosphere microorganism for development and utilization, which is conducive to the improvement of soil quality and ecological environment. Previous studies suggested that soil microbial communities are directly affected by halophytes [49]. The NMDS analysis confirmed that soil bacterial community structure was significantly different under three halophytes rhizosphere. This may be related to the roots of different halophytes secrete different chemical substances, which affects the enrichment of bacteria and changes the community structure [50]. The decomposition of root exudates and litter will change soil characteristics and promote the establishment of specific rhizosphere bacterial communities [34]. Bacterial community structure was also affected by soil properties, such as soil EC, pH and SOM [51,52]. Hence, Leymus chinensis, Puccinellia tenuiflora and Suaeda glauca select for specific bacterial taxa and control microbiomes, which led to differences of bacterial community structure [53]. The composition and structure of the soil bacterial community are the main internal factors affecting the bacterial diversity. The soil bacterial diversity of halophytes rhizosphere in high-salt soil largely depends on the soil properties and plant species [43,63]. Study has shown that the rhizosphere of plants has specific selectivity for bacteria that colonize the rhizosphere, which change the species richness and homogeneity, leading to differences in alpha diversity [54]. In addition, microbial community diversity also affected by different root exudates [55]. The total number of OTUs and bacterial diversity indices (including Ace, Chao, and Shannon) were the highest in Leymus chinensis. This may be related to the higher SOM content and lower pH in the Leymus chinensis rhizosphere soil, since accumulation of organic matter increases the chances of successful migration of soil bacteria community [24]. These consequences indicate the Leymus chinensis is more conducive to the restoration of soil bacterial diversity in saline-alkali conditions compared with Puccinellia tenuiflora and Suaeda glauca.

Correlation between bacterial communities and soil properties
Complex environmental factors determined of vegetation types in different regions, and the soil microbes were particularly sensitive to changes in the external environment. Previous studies suggested that soil properties are the important factors leading to the changes in soil bacterial communities [56]. We found that the rhizosphere bacterial community structures of three halophytes are also influenced by SOM and SWC, which is consistent with findings conducted by Gupta and Drenovsky [48,57]. This suggested that whether in harsh environmental conditions or in farmland, SOM and SWC are very important for the shaping of soil bacterial community.
It is reported that bacteria biogeography differences are largely controlled by edaphic variables, such as soil pH [58]. Linear regression analysis indicated that Ace, Chao, and Shannon indices negatively correlated with soil pH. This is consistent with prior studies that soil bacterial diversity is affected by pH [59]. The rich bacterial diversity in the rhizosphere soil helps to improve the adaptation of halophytes to saline-alkali stress [60]. Therefore, we can evaluate the saline-alkali stress of halophytes through the rhizosphere bacterial diversity. These findings may help us to understand the interaction between different halophyte rhizosphere bacterial communities and their soil environments.

Soil bacterial community potential functions
Soil bacteria play a vital part in nutrient cycling, maintaining soil fertility, and carbon sequestration through Amino acid metabolism and Carbohydrate metabolism [61]. Chaudhary et al. (2018) reported that the soil microbial functional gene abundance was affected by different halophyte vegetations [62]. The relative abundance of bacterial metabolism was above 78% under three halophytes rhizosphere in this study. Carbohydrate metabolism, Amino acid metabolism, Energy metabolism and Lipid metabolism were main metabolic functions at pathway Level2. This is consistent with the findings of previous studies that that Amino acid metabolism and Carbohydrate metabolism are the main metabolic functions by PICRUSt predictions [25,63]. These results also suggested that Amino acid metabolism and Carbohydrate metabolism are main bacterial functions in soil, not only in halophyte vegetation rhizosphere but also in other soil types [64][65][66]. The C, N and P cycle of halophyte rhizosphere bacteria can be evaluated by the abundance of bacterial functional genes [67]. We observed that Carbon metabolism, Sulfur metabolism and Nitrogen metabolism in Suaeda glauca were significantly higher than that in Leymus chinensis and Puccinellia tenuiflora. This discrepancy may be due to their plant root exudates. One of the manifestations of the rhizosphere effect is that rhizosphere bacteria are strongly affected by plant root exudates, and the metabolites of bacteria will be differences [68]. Pii et al. (2016) showed that plants affected the bacterial function through root exudates as C and N sources [69]. Under the dual influence of rhizosphere bacterial metabolites and root exudates, the soil properties of the halophyte rhizosphere will change [70]. Besides, halophytes improve their ability to adapt to saline-alkali conditions by regulating root exudates and affecting the metabolites of rhizosphere bacteria [38]. We found that SOM and SWC in Leymus chinensis and Puccinellia tenuiflora are significantly higher than Suaeda glauca. According to these data, we can infer that the rhizosphere bacterial community of Suaeda glauca adjusts its own metabolism to the arid and barren saline-alkali environment.

Conclusions
This study has shown that soil properties and rhizosphere bacterial community structure showed significant difference in three halophytes. The highest soil EC and SOM content were found in Leymus chinensis. At the phylum level, Leymus chinensis, Puccinellia tenuiflora and Suaeda glauca enriched Proteobacteria, Actinobacteria and Chloroflexi in the rhizosphere to adapt the salt stress. The highest relative abundance of Proteobacteria was observed in Puccinellia tenuiflora, with an increase of 0.89% and 1.88% compared to Leymus chinensis and Suaeda glauca, respectively. In contrast with Puccinellia tenuiflora, the relative abundance of Chloroflexi in Leymus chinensis and Suaeda glauca decreased by 0.82% and 1.91%, respectively. Suaeda glauca rhizosphere has stronger enrichment of Nitrospirae and Cyanobacteria, and Leymus chinensis has higher rhizosphere bacterial diversity. Regression analysis revealed that SOM and SWC have an important influence on rhizosphere bacterial community structure, and pH significantly affected the diversity of bacterial communities. The research has also shown that the rhizosphere bacterial metabolism in Suaeda glauca were significantly higher than that in Leymus chinensis and Puccinellia tenuiflora, which including Carbon metabolism, Sulfur metabolism and Nitrogen metabolism. The limitation of this study is that the rhizosphere samples from different regions with three halophytes are geographically distant, environment and climatic factors are different, and these factors largely affect the diversity of rhizosphere bacterial community. On the other hand, PICRUSt only preliminarily predicted the function of the halophyte rhizosphere bacteria. In future studies, metagenomics and metabolomics should be used for further verification to better understand the functions of different halophyte rhizosphere bacterial communities. In spite of its limitations, the study certainly adds to our understanding of the diversity and functional characteristics of rhizosphere bacterial communities in the typical halophytic vegetation of northern China.