The root enrichment of bacteria is consistent across different stress-resistant plant species

Bacteria, inhabiting around and in plant roots, confer many beneficial traits to promote plant growth and health. The secretion of root exudates modulates the nutritional state of the rhizosphere and root area, further selecting specific bacteria taxa and shaping the bacteria communities. Many studies of the rhizosphere effects have demonstrated that selection by the plant rhizosphere consistently enriches a set of bacteria taxa, and this is conserved across different plant species. Root selection effects are considered to be stronger than the rhizosphere selection effects, yet studies are limited. Here, we focus on the root selection effects across a group of 11 stress-resistant plant species. We found that the root selection consistently reduced the alpha diversity (represented by total number of observed species, Shannon’s diversity, and phylogenetic diversity) and altered the structure and composition of bacteria communities. Furthermore, root selection tended to enrich for clusters of bacteria genera including Pantoea, Akkermansia, Blautia, Acinetobacter, Burkholderia-Paraburkholderia, Novosphingobium, Massilia, Pseudomonas, Chryseobacterium, and Stenotrophomonas. Our study offers some basic knowledge for understanding the microbial ecology of the plant root, and suggests that several bacteria genera are of interest for future studies.


INTRODUCTION
The plant root and surrounding soil area represents a hotspot for interactions between plants and microbes (Mitter, De Freitas & Germida, 2017;Zhalnina et al., 2018). Two distinct compartments, the rhizosphere and the root or endosphere, are separated based on their physical proximity to the plant and the level of host's influence on the microbial communities assembled from the pool of soil (Fitzpatrick et al., 2018;Yurgel et al., 2018). The rhizosphere compartment is inhabited by microbes that assemble sourrounding the roots, whereas the root compartment is inhabited by the microbes assembled within roots (Fitzpatrick et al., 2018;Yurgel et al., 2018). Both compartments are inhabited by a highly

Plant growth and sampling
Eleven plant species, including Casuarina equisetifolia (CE), Calophyllum inophyllum (CI), Fagraea ceilanica (FC), Guettarda speciosa (GS), Heritiera littoralis (HL), Hernandia sonora (HS), Melaleuca bracteata (MB), Pongamia pinnata (PP), Portulaca pilosa (PPI), Ruellia brittoniana (RB), and Scaevola sericea (SS), were selected in this study based on their ability to thrive under harsh environments, such as high salinity and alkalinity soils and sites with high temperatures and light, and under drought (Tong et al., 2013;Liu et al., 2015;Ren et al., 2017). Plant seeds were sown in pots (24 × 26 cm, diameter × height) filled with field soils (pH = 8.0∼9.0, mean soil organic carbon 0.4%, total nitrogen 0.05%, total phosphorus 0.03%) collected near a greenhouse of Lingye Gardening Limited Compary (Wenchang, China). For each plant species, four seeds were sown in each pot of a total of six pots. After germination, only one seedling was kept growing in each pot, and the remaining plants were grown in the same greenhouse from February to August, 2017. All plant and soil samples were harvested in August 14, 2017. The soils collected from the field, incubated in the pots for the same period to the plant growth, were designated as bulk soil, and the soils, adhered to plant roots, were mechanically brushed and kept as rhizosphere soil. Fine roots, with diameter smaller than two mm, were washed in 1 × TE buffer (Tris-EDTA buffer: 10 mM Tris, 1 mM EDTA, pH = 8.0. Added with 0.1% Triton X-100) for 30s, 75% ethanol for 15s, 2% bleach for 15s, and rinsed three times with sterile water (Agler et al., 2016). The soils and sterile roots were stored under −80 • C overnight before DNA extraction.

DNA extraction and amplicon sequencing
Microbial genomic DNA was extracted from 250 mg soil or 100 mg root with the Powersoil R Kit (MoBio Laboratories Inc., Carlsbad, CA, USA). The primer pair 515F/806R was used to amplify partial sequence of the 16S V4-V5 rRNA region of the bacteria community (Caporaso et al., 2011). PCR reactions were carried out using Phusion High-Fidelity PCR Master Mix (New England Biolabs Inc., Ipswich, MA, USA). The PCR cycle was set with an initial 94 • C for 5 min, followed by 35 cycles of 94 • C for 45s, 56 • C for 30s, and 72 • C for 30s, a final extension of 72 • C for 10 min. The harvested DNA was verified on a 2% agarose electrophoresis gel and prepared for sequencing using the TruSeq R DNA PCR-Free Library Preparation Kit (Illumina, San Diego, CA, USA). Amplicon sequencing was carried out using the Illumina HiSeq2500 platform (Illumina, San Diego, CA, USA).

Bioinformatics analysis
Paired-end reads were merged by FLASH v1.2.7 (Magoč & Salzberg, 2011) and spliced raw tags were quality filtered by QIIME v1.7.0 (http://qiime.org/) to obtain high-quality clean tags (Caporaso et al., 2010). Chimeric sequences were removed by following UCHIME algorithm (Edgar et al., 2011). Then, the effective tags were assigned to Operational Taxonomic Units (OTUs) by Uparse v7.0.1001 (http://drive5.com/uparse/) at 97% similarity level (Edgar, 2013). All singletons were deleted, the remaining sequences of each sample were normalized to the sample with the least number of sequences (32,416). The rarefaction curve showed the number of OTUs/observed species identified in randomly picked sequences, the curve of each sample reached to plateau, implying that deeper sequencing might not result in a significant increase of the number of observed species (Fig. 1A). The taxonomic identity was assigned to bacteria OTUs after searching against the SILVA138 database (Yilmaz et al., 2014).

Statistical analysis
The alpha diversity indices (Observed species, Chao1, Shannon, Simpson, ACE, Good's coverage, and Phylogenetic diversity), summarizing the distribution of species abundances and diversities in a given sample (Thukral, 2017), were calculated in QIIME v1.9.1 and visualized by using R software v4.0.2 (https://www.r-project.org/). For example, the Shannon's diversity index was calculated with Sum (p i log(p i )), the phylogenetic diversity value was calculated as the sum of branch lengthes of all species in a given sample (Faith, 1992;Thukral, 2017). Differences of bacteria community structure among the bulk soils, rhizosphere soils, and roots were assessed by performing PERMANOVA with 999 permutations in the vegan package in R (R Core Team, 2020). A principal components analysis (PCoA) based on bray-curtis dissimilarities was performed using the cmdscale function within vegan package. Function prediction of each OTU was performed by Tax4Fun (Aßhauer et al., 2015). The estimation of fold change of bacteria taxa and predicted functions was performed in DESeq2 (Love, Huber & Anders, 2014). Equality of variances (bartlett.test) and condition of normality (shapiro.test) were tested in R software, ANOVAs followed by Tukey post-hoc pairwise test was used for comparisons among the bulk soils, rhizosphere soils, and root samples, two-tailed t -test was used for comparisons between the rhizosphere soils and the root samples. P values were adjusted using the Benjamini-Hochberg method, and p values smaller than 0.05 were accepted with significance.  Fig. 1D). When examing beta diversity, root samples were also significantly differentiated from soil samples in composition and structure (PCoA 1 = 28.2%, p < 0.01, Fig. 1E), and also in predicted functions (PCoA 1 = 56.3%, p < 0.01, Fig. 1F).

Comparison of rhizosphere and root communities by plant species
In overall analysis, root bacteria communities were composed of different OTUs and overall less diversity compared to the rhizosphere bacteria communities. This phenomenon was consistent across all tested plant species. For each plant species, the observed species, Shannon's diversity, and phylogenetic diversity were typically lower in root samples compared to the rhizosphere samples ( Fig. 2). Specifically, there was a significant reduction in alpha diversity estimates, observed species (

Taxa enrichment and depletion in root compared to rhizosphere soil
To further characterize the differences in bacteria communities between the roots and the rhizosphere soils, the relative abundances of the top 50 bacteria genera and predicted genes were compared. The log fold change of bacteria genera and predicted genes in root, either     enriched or decreased, were plotted by plant species in heatmaps (Fig. 4). From the plot, a clade of bacteria including Candidatus Solibacter, Pseudolabrys, Variibacter, H16 (of the family Desulfurellaceae), Bryobacter, Rhizomicrobium, and Haliangium decreased in the roots of multiple plant species (Fig. 4A). On the contrary, Pantoea, Akkermansia, Blautia, Acinetobacter, Burkholderia-Paraburkholderia, Novosphingobium, Massilia, Pseudomonas, Chryseobacterium, and Stenotrophomonas were enriched in the roots of multiple plant species (Fig. 4A). For example, Pantoea and Pseudomonas were both significantly enriched in the roots of seven out of 11 (63.6%) plant species. The relative abundance of Pantoea increased from 0.6% ± 1.1% in the rhizosphere soil to 12.5% ± 16.6% in the roots, and   the relative abundance of Pseudomonas increased from 1.7% ± 3.7% to 8.3% ± 15.7%. Accordingly, a clade of predicted genes including K02529, K02056, K02029, K18900, K10441, K03406, K03566, and K02030 were also enriched in the roots of different plant species (Fig. 4B). In specific, the enrichment of the genera Pantoea and Pseudomonas in the roots were represented by OTU1 (Pantoea ananatis, Fig. 5A) and OTU3 (Pseudomonas sp., Fig. 5B), respectively. The enrichment of the species complex Burkholderia-Paraburkholderia was represented by four OTUs (OTU2, OTU4, OTU21, and OTU215; Figs. 5C-5F), of which OTU2 was most enriched in the roots (Fig. 5C). In addition, the predicted genes enriched in the roots (Fig. 4B, K02529, K02056, K02029, K18900, K10441, K03406, K03566, and K02030) were catalogued into different cellular pathoways: genes involved in amino acid metabolism (Fig. 6A), biofilm formation (Fig. 6B), carbohydrate metabolism (Fig. 6C), and membrane transport (Fig. 6D) were consistently enriched in the roots.

DISCUSSION
The 11 plant species selected for this study have demonstrated ecological functional traits that make them useful for vegetation restoration of destroyed and degraded ecosystems (Ren et al., 2017;Wang et al., 2019). Using these stress-resistant plant species, we tested d f ** * ns * ** ns *** * ns ns ns * * ns * ** *** * ns *** ns ns ** * * *** ** * ns ns ns ns * ** ns ns * * ns ns ns ns ns ns ns ns * ns ns ns ns ns ns * ns * ns ns ns * ns ns ** ns ns ns  whether root selection of the bacteria community occurs by comparing communities within the bulk soils, the rhizosphere soils, and the roots. The root selection of bacteria from the rhizosphere can be seen as a re-selection process, as the rhizosphere already selects bacteria from the bulk soils (Zhalnina et al., 2018;Schneijderberg et al., 2020). We observed that root selection, leading to the enrichment and depletion patterns of root-selected bacteria, was consistent across 10 of the 11 diverse plant species used in this study. Further, *** *** * ** ** ** ** ns * ns * *** *** * * ns ** *** ns * * * ns * ns *** ns * ns ns ** ns ns * ** ** ** * * ns ns ** ns * Compartment Rhizosphere Root Bulk we identified several bacteria taxa, Pantoea and Pseudomonas, and genes involved in amino acid metabolism, biofilm formation, carbohydrate metabolism, and membrane transport were consistently enriched in the roots. As soil bacteria can be important for improving plant growth and health (Müller et al., 2016;Sánchez-Cañizares et al., 2017), it may be that root bacteria are selected and structured by plant genetic factors and root exudates (Stringlis et al., 2018;Zhalnina et al., 2018) for specific functions. Plants of diverse species can share a core set of microbiota in the rhizosphere and root areas (Fitzpatrick et al., 2018;Yeoh et al., 2017). In specific, bacteria genera, Bacillus, Burkholderia, Pantoea, Pseudomonas, and Rhizobium are more frequently found in the rhizosphere and root areas (Saleem et al., 2018;Cordero, De Freitas & Germida, 2020;Yeoh et al., 2017). In our study, we also found that these genera were among the dominant genera in the rhizosphere and root areas, with highest abundances observed for Pantoea, Pseudomonas, and Burkholderia.
The different distribution of the core species between the rhizosphere and the root implies the effects of the root selection, on the contrary, whether the root-enriched bacteria confer some positive feedbacks back to the root? In the heatmap plot (Fig. 4), bacteria genera Pantoea, Akkermansia, Blautia, Acinetobacter, Burkholderia-Paraburkholderia, Novosphingobium, Massilia, Pseudomonas, Chryseobacterium, and Stenotrophomonas, along with the genera Rhizobium and Methylobacterium were generally enriched in the root (Fig.  4A), suggesting that root selection across diverse plant species is conserved (Cordero, De Freitas & Germida, 2020). To our knowledge, Massilia species are root-colonizing bacteria of many plant species, these species were found to be correlated to plant development and competitive in root colonization (Ofek, Hadar & Minz, 2012), but there function mechanism is less known. Species in the genera Burkholderia, Pantoea, Rhizobium, and Stenotrophomonas have been found with the capabilities of nitrogen fixation (Franche, Lindström & Elmerich, 2009;Ryan et al., 2009;Walterson & Stavrinides, 2015;Lemanceau et al., 2017), which may contribute to their enrichment in plant roots. Besides nitrogen fixation, the species in these genera, along with in the well known plant-associated genus Pseudomonas, harbor versatile skills in improving plant abiotic stress tolerance and suppressing disease occurrence (Lemanceau et al., 2017; see more information in Table  S1). The plant core genus Bacillus, is well known to be benificial to different plant species in growth promotion and disease suppression (Miljaković, Marinković & Balešević-Tubić, 2020), was not enriched in our root samples, maybe because these species mainly inhabit and function in the rhizosphere soils (Miljaković, Marinković & Balešević-Tubić, 2020).
Furthermore, our predicted gene functions imply that the enriched bacteria genera may correlate to four basic bacteria cellular functions: amino acid metabolism, biofilm formation, carbohydrate metabolism, and membrane transport. The prediction results conform to the study of Levy et al. (2017). In their study, they found that the plantassociated bacteria harbor higher numbers of gene clusters involved in carbohydrate metabolism functions, amino acids, and cell membrane functions compared to other bacteria. Taken together, our results show that the root enrichment of specific bacteria genera is a conserved plant trait, and imply that the root-enriched bacteria have more active and closer relationship to plant species. However, what benefits of these enriched bacteria can confer to their host plant still needs more experimental tests.
Bacteria species inhabiting in the root are associated with plant resistance to different stresses, such as soil salinity and drought (Ali, Charles & Glick, 2014;Sandhya et al., 2017). With culture and inoculation dependent methods, the root bacteria are too abundant to isolate. For example, over 300 bacteria genera and over 20,000 OTUs have been detected in our soil and root samples, to isolate and test these bacteria is not near to possible. Our sequencing results indicate that the species of Pantoea ananatis (OTU1), Pseudomonas sp. (OTU3), and Burkholderia-Paraburkholderia (OTU2, OTU4, OTU21, and OTU215) are closely-associated to the plant root and may be selected by the plant root. Results from this study have reduced our needed to culture blindly, but rather we can now target these species for further studies.