甘肃农业大学：食品科学与工程学院（文献讲义）Bacterial community associated with rhizosphere of maize and cowpea in a subsequent cultivation.pdf

Contents lists available at ScienceDirect Applied Soil Ecology journal homepage: www.elsevier.com/locate/apsoil Bacterial community associated with rhizosphere of maize and cowpea in a subsequent cultivation Ademir Sergio Ferreira de Araujoa,⁎ , Ana Roberta Lima Mirandaa , Ricardo Silva Sousaa , Lucas William Mendesb , Jadson Emanuel Lopes Antunesa , Louise Melo de Souza Oliveiraa , Fabio Fernando de Araujoc , Vania Maria Maciel Melod , Marcia do Vale Barreto Figueiredoe a Soil Quality Lab., Agricultural Science Center, Federal University of Piauí, Teresina, PI, Brazil b Laboratory of Molecular Ecology, CENA-USP, Piracicaba, SP, Brazil c University of Sao Paulo West, UNOESTE, Campus II, Presidente Prudente, SP, Brazil d Laboratório de Ecologia Microbiana e Biotecnologia, Federal University of Ceara, Fortaleza, CE, Brazil e Agronomic Institute of Pernambuco, Av. San Martin, Recife, PE, Brazil ARTICLE INFO Keywords: Bulk soil Metagenomics Zea mays Vigna unguiculata Soil microbiome ABSTRACT The rhizosphere is the soil zone influenced by the roots and its characteristics vary according to plant species and their stage of development. These different characteristics can influence the bacterial community inhabiting the rhizosphere niche that, in turn, influences plant growth and health. In this study, the bacterial community dynamics in the rhizosphere of maize and cowpea, in subsequent cultivation, were assessed using the 16S rRNA sequencing. Thus, during maize growth, soils were collected at 45 (flowering) and 75 (senescence) days. Afterward, during cowpea growth, the sampling occurred at 35 (flowering) and 65 (senescence) days. Our results showed differences between developmental stages within the same plant species. For maize, Acidobacteria decreased from flowering to senescence; while for cowpea, Proteobacteria, Armatimonadetes, WPS-2, and OP11 increased from flowering to senescence. Comparing the same developmental stage for both plant species, Proteobacteria, Elusimicrobia, WPS-2 decreased from maize flowering to cowpea flowering; while for the senescence stage, Acidobacteria, Armatimonadetes, and OP11 increased from maize to cowpea. Also, the rhizosphere community dynamic was more complex at the senescence stage compared to the flowering stage and bulk soil for both plants species. The results showed that the structure and diversity of bacterial community vary significantly according to plant species and, in a minor extent, their developmental stage. Also, we showed that the variation in rhizosphere activity during the plant growth could drive the responses of the bacterial community. 1. Introduction The rhizosphere is defined as the narrow region of soil that is directly influenced by plant roots (Chen et al., 2001). In this region, plant roots release exudates, i.e., organic compounds such as proteins and sugars, which stimulate the microbial community in comparison in with the bulk soil (Mougel et al., 2006). In the rhizosphere microbiome, bacterial community establishes positive interactions with roots and, thus, acts on essential functions in agriculture, such as nutrient acquisition, plant growth-promotion and protection against pathogens (Cavaglieri et al., 2009; Mendes et al., 2014, 2018a). The characteristics of rhizosphere vary in function of different plant species (Eisenhauer et al., 2017). Different plant groups, such as legumes and grass, present contrasting rhizosphere traits regarding root exudation and nutrient requirements (Rovira, 1969; Gransee and Wittenmayer, 2000). Legume roots can exude more amino acids and sugar than grassroots (Gransee and Wittenmayer, 2000), while grass requires more nutrients than legumes (Ghosh et al., 2009). Also, plants species within the same taxonomic group share similar characteristics of the rhizosphere, such as root biomass, and the amount and availability of the exudation (Zhou et al., 2017), while plant species belonging to different groups could present distinct rhizosphere traits. Therefore, it is expected that the bacterial community in the rhizosphere could be driven by differences in these traits, which may select different groups of bacteria. Indeed, a previous study has shown a distinct bacterial community in the rhizosphere of Lotus corniculatus https://doi.org/10.1016/j.apsoil.2019.05.019 Received 4 February 2019; Received in revised form 15 May 2019; Accepted 23 May 2019 ⁎ Corresponding author. E-mail address: ademir@ufpi.edu.br (A.S.F. de Araujo). Applied Soil Ecology 143 (2019) 26–34 Available online 06 June 2019 0929-1393/ © 2019 Elsevier B.V. All rights reserved. T

(legume) when compared with Holcus lanatus (graminea) (Ladygina and Hedlund, 2010). Recently, Zhou et al. (2017) have found differences between the bacterial community in soil under ryegrass (grass) and alfalfa (legume). The plant developmental stage is another factor influencing the bacterial community in the rhizosphere (Mougel et al., 2006). During the plant growth, the roots exudation changes according to the stage of development, and the abundance and structure of the rhizospheric microbial community are shaped according to the pattern of root exudation that may vary from young seedling, flowering, to plant senescence (Dunfield and Germida, 2003). Indeed, Houlden et al. (2008) evaluated the response of the microbial community to the different growth stage of pea, wheat and sugar beet and observed changes in the bacterial community following the developmental stage for all the tested plants. Considering the crucial role of the rhizosphere microbiome for plant growth and health, it is essential to elucidate how different plant species at different growth stages select differently the microbial community inhabiting their rhizosphere. Maize (grass) and cowpea (legume) are two important plant species cultivated in the world being used for human and animal feeding and cover crop (Ramirez-Cabral et al., 2017; Boukar et al., 2016). Although these plants are distributed in several agroecosystems, studies about the bacterial community in the rhizosphere of these species are scarce. Mainly for cowpea, there is no information about the structure of the bacterial community in the rhizosphere. Also, it is unclear the effect of the different stage of development in these plants on the soil bacterial community. In order to determine the extent to which different plant species are able to select a distinct rhizospheric microbiome, in this study we hypothesized that (1) maize and cowpea, as different plant species, exert different rhizosphere effect and shape the diversity and structure of bacterial community differently; and (2) the soil bacterial community changes during the flowering and senescence of plants. Thus, this study aimed to evaluate the diversity and structure of the bacterial community in the rhizosphere of maize and cowpea during the flowering and senescence using the 16S rRNA sequencing. 2. Material and methods 2.1. Experimental site and soil sampling The experiment was conducted at the Department of Soil Science from the Federal University of Piauí, Brazil. The soil of the experimental site is classified as Fluvisol soil. The experimental design was completely randomized with three replicates. Each experimental plot presents 20 m2 (12 m2 of usable area). Maize (Zea mays L.) was sowed at a density of five plants m−1 and was grown for 75 days. Afterward, cowpea [Vigna unguiculata (L.) Walp.] was sowed at the density of six plants m−1 and was grown for 65 days. To measure the rhizospheric influence of each plant species on the bacterial community, soil samples adhered to the roots were collected and mixed to form a composite sample in each plot. During maize growth, soils were collected at 45 (flowering, named MF) and 75 (senescence, MS) days. During cowpea growth, the sampling occurred at 35 (flowering, CF) and 65 (senescence, CS) days. Soil samples without plants effects were collected as control (bulk soil). The soil samples were sieved (2-mm) and stored at −20 °C. Soil chemical and physical properties, estimated according to Embrapa (1997), are shown in Table 1. 2.2. DNA extraction and sequencing Total DNA was extracted from 0.5 g (total humid weight) of soil using the PowerLyzer PowerSoil DNA Isolation Kit (MoBIO Laboratories, Carlsbad, CA, USA), according to the manufacturer's instructions. The DNA extraction was performed in triplicate for each soil sample, totalizing 15 samples [(bulk soil + 2 rhizosphere + 2 developmental stage) × 3 replicates]. The V4 region of the 16S rRNA gene was amplified with regionspecific primers (515F/806R) (Caporaso et al., 2011). Each 25 μL Polymerase Chain Reaction (PCR) reaction volume contained the following: 12.25 μL of nuclease-free water (Certified Nuclease-free, Promega, Madison, WI, USA), 5.0 μL of buffer solution 5× (MgCl2 2 Mm), 0.3 mM dNTP's 0.3 μM of each primer (515 YF 40 μM e 806 R 10 μM), 1.0 unit of Platinum Taq polymerase High Fidelity in concentration of 0.5 μL (Invitrogen, Carlsbad, CA, USA), and 2.0 μL (10 ng μL−1 ) of template DNA. Moreover, a control reaction was performed by adding water in place of DNA. The conditions for PCR were as follows: 95 °C for 3 min to denature the DNA, with 35 cycles at 98 °C for 20 s, 55 °C for 20 s, and 72 °C for 30 s, with a final extension of 3 min at 72 °C to ensure complete elongation. After indexing, the PCR products were cleaned up using Agencourt AMPure XP – PCR purification beads (Beckman Coulter, Brea, CA, USA), according to the manufacturer's manual, and quantified using the dsDNA BR assay Kit (Invitrogen, Carlsbad, CA, USA) on a Qubit 2.0 fluorometer (Invitrogen, Carlsbad, CA, USA). Once quantified, an equimolar concentration of each library (50 ng) was pooled into a single tube. After quantification, the molarity of the pool was determined and diluted to 2 nM, denatured, and then diluted to a final concentration of 8.0 pM with a 20% PhiX (Illumina, San Diego, CA, USA) spike for loading into the Illumina MiSeq sequencing machine (Illumina, San Diego, CA, USA). Sequence data were processed using QIIME following the UPARSE standard pipeline according to Brazilian Microbiome Project (Pylro et al., 2014), to produce an OTU table and a set of representative sequences. Briefly, the reads were truncated at 240 bp and quality-filtered using a maximum expected error value of 0.5. Pre-filtered reads were dereplicated, and singletons were removed and filtered for additional chimeras using the RDP_gold database using USEARCH 7.0. These sequences were clustered into OTUs at a 97% similarity cut-off following the UPARSE pipeline. After clustering, the sequences were aligned and taxonomically classified against the Greengenes database (version 13.8). 2.3. Statistical analysis The community structure and its correlation with soil parameters were visualized using Redundancy analysis (RDA). All matrices were initially analyzed using Detrended Correspondent Analysis (DCA) to evaluate the distribution of the data, revealing that the best-fit mathematical model for the data was RDA. Forward selection (FS) and the Monte Carlo permutation test were applied with 1000 random permutations to verify the significance of environmental parameters upon the microbial community structure. RDA plots were generated using Table 1 Chemical and physical properties of soil. BS MF MS CF CS pH (water) 5.6 6.3 6.7 6.2 6.0 EC (dS m−1 ) 0.3 0.5 0.7 0.5 0.5 TOC (g kg−1 ) 5.9 5.9 6.7 5.4 5.6 P (mg kg−1 ) 5.5 5.8 6.0 4.8 4.3 K (mg kg−1 ) 70 89 85 81 78 Ca (mg kg−1 ) 213 317 328 320 296 Mg (mg kg−1 ) 38 61 64 61 58 Na (mg kg−1 ) 69 98 93 85 91 Sand (%) 60 60 60 60 60 Silt (%) 28 28 28 28 28 Clay (%) 12 12 12 12 12 BS – Bulk soil; MF – maize flowering; MS – maize senescence; CF – cowpea flowering; CS – cowpea senescence; TOC – total organic C; EC – electric conductivity. A.S.F. de Araujo, et al. Applied Soil Ecology 143 (2019) 26–34 27

Canoco 4.5 software (Biometrics, Wageningen, the Netherlands). Permutational multivariate analysis of variance (PERMANOVA) was used to test whether sample categories harbored significantly different microbial community structures. Alpha diversity was calculated from a matrix of richness at the genus level using Shannon's index. PERMANOVA and alpha diversity indexes were calculated with the software PAST 3 (Hammer et al., 2001). Venn diagrams were also constructed to verify the proportion of groups exclusive and shared between treatments using the web tool Venny 2.1.0 (Oliveros, 2007). To visualize the differential microbial community composition among treatments, we used the Statistical Analysis of Metagenome Profile software (STAMP) (Parks et al., 2014). The OTU table generated from the 16S profiling was used as input. The comparison was based on P-values calculated using the two-sided Welch's t-test and the correction was made using Benjamini-Hochberg false discovery rate (Benjamini and Hochberg, 1995). To further assess the microbial community dynamics among the samples we conducted co-occurrence network analysis using the Phyton module ‘SparCC’ (Friedman and Alm, 2012). For the calculations, a table of frequency of hits affiliated at genus level was used for the analysis. For each network, SparCC correlations were calculated between microbial taxa and filtered based on the significance at P < 0.01. The nodes in the reconstructed network represent taxa at the genus level, whereas the edges represent significantly positive or negative correlations between nodes. The network graphs were based on a set of measures, such as the number of nodes, number of edges, modularity, number of communities, average node connectivity, average path length, diameter, and cumulative degree distribution. Networks visualization and properties measurements were calculated with the interactive platform Gephi (Bastian et al., 2009). 3. Results 3.1. Bacterial community structure The redundancy analysis (RDA) explained > 50% of the variation in the first two axes of the plot (Fig. 1A), with the samples clustering according to the niche, i.e. bulk soil and rhizosphere (PERMANOVA, F = 4.75, P = 0.034). The RDA analysis also showed that the bacterial community of the bulk soil correlated with higher values of Ca. For rhizosphere samples, the bacterial community associated to maize correlated with higher values of Mg, P, TOC, and CEC, while for cowpea the community was associated with higher values of pH, K, Na, and EC. Also, the results showed a correlation between maize and cowpea rhizosphere in senescence, showing that bacteria that remain during senescence may be similar in both plant species, maybe due to the same rhizospheric soil that could change in different soils. To compare the bacterial community structure considering only rhizosphere samples, we computed a Bray-Curtis similarity matrix and coordinated into two dimensions using NMDS (Fig. 1B). The samples were grouped according to the plant species (PERMANOVA, F = 2.46, P = 0.005), with differences related to the stage of development only for cowpea. 3.2. Bacterial community composition and diversity After quality trimming, the bacterial community profiling using 16S rRNA gene sequencing generated 230,000 reads (an average of 15,000 sequences per sample) that were used for downstream analysis. The most abundant bacterial groups were affiliated to Actinobacteria (28.1% of the sequences), followed by Proteobacteria (21.5%), Firmicutes (11%), Acidobacteria (8.7%), Chloroflexi (8.7%) and Planctomycetes (8.6%) (Supplementary Fig. 1). When samples were compared using STAMP software, the abundance of specific bacterial phyla showed shifts from bulk soil to rhizospheric soils (Fig. 2). Bulk soil showed a higher abundance of Acidobacteria, Proteobacteria, Elusimicrobia, Armatimonadetes, and OP3; on the other hand, in the rhizosphere samples, there was an increase in abundance of the phyla Actinobacteria, Chlamydiae, Firmicutes and TM6 (P < 0.05). Comparing the rhizosphere sample within the same plant species, we found differences between developmental stages (Fig. 2). For maize, Acidobacteria decreased from flowering to senescence; while for cowpea, Proteobacteria, Armatimonadetes, WPS-2, and OP11 increased from flowering to senescence. Comparing the same developmental stage for both plant species, the results showed a decrease in Proteobacteria, Fig. 1. Structure of bacterial communities in bulk soil and rhizosphere of maize and cowpea plants at flowering and senescence developmental stages. (A) Redundancy analysis (RDA) of bacterial communities and soil characteristics. Arrows indicate correlation between environmental parameters and bacterial profile. (B) Non-metric multidimensional scaling ordination (NMDS) of Bray-Curtis similarity matrix of the bacterial profile from rhizosphere samples. The lines between dots represent the minimal spanning tree, which connects all points with minimal total length, based on the similarity index. Dashed lines in the graph indicate significant clusters (PERMANOVA, P < 0.05). MF = Maize flowering; MS = Maize senescence; CF = Cowpea flowering; CS = Cowpea senescence. A.S.F. de Araujo, et al. Applied Soil Ecology 143 (2019) 26–34 28

Elusimicrobia, WPS-2 from maize flowering to cowpea flowering; while for the senescence stage, Acidobacteria, Armatimonadetes, and OP11 increased from maize to cowpea. Interestingly, the phyla DeinococcusThermus and OP1 were exclusively abundant in maize and cowpea rhizosphere, respectively. In a lower taxonomic level, we compared bulk soil with the rhizosphere of maize and cowpea considering both developmental stages. In general, we observed an enrichment of specific families in the rhizosphere samples (Supplementary Fig. 2). Bulk soil presented a higher abundance of the families Isosphaeraceae, Chitinophagaceae, Nitrosomonadaceae among others (Supplementary Fig. 2A and B). In rhizosphere samples, there was an increased number of sequences af- filiated to the families Paenibacillaceae, Bacillaceae, Gaiellaceae, among others. Comparing the rhizosphere community between the plant species, we observed that 15 families were differently abundant. For example, the families Chthoniobacteraceae and 0319-6A21 (Nitrospirales) were more abundant in cowpea, while Oxalobacteraceae, Acetobacteraceae, Coxiellaceae, and Mycobacteriaceae were more abundant in maize rhizosphere (Supplementary Fig. 2C). We also compared the different developmental stages within each plant species and observed that nine and five bacterial families were differently distributed in the rhizosphere of maize and cowpea, respectively (Supplementary Fig. 3). The pattern of richness and diversity measurements revealed a decreased diversity in the rhizosphere of the cowpea plant in comparison with maize and bulk soil (Fig. 3). When the treatments were compared at the genus level using Venn diagrams, we found that bulk soil and rhizosphere samples shared > 50% of the detected genera (Fig. 4A, B). For maize, the proportion of genera exclusively present in the rhizosphere compared to the bulk soil increased from zero in the flowering, to 8.7% in the senescence (Fig. 4A). On the other hand, the proportion of genera exclusively present genera in cowpea decreased from the flowering (3.6%) to the senescence (2.4%) (Fig. 4B). This result reveals that there is a different rhizosphere effect depending on the plant species and the developmental time. When only the rhizosphere samples were compared, we observed that 51.9% of the detected genera are shared among all the samples, while the senescence of maize presented the highest proportion of exclusive genera (12.7%) (Fig. 4C). Fig. 2. Distribution of the most differential bacterial phyla based on 16S rRNA profile of samples from bulk soil and rhizosphere of maize and cowpea plants at flowering and senescence developmental stages. Boxes indicate IQR (75th to 25th of the data). The median value is shown as a line within the box and outliers are represented by dots. Different lower case letters refer to significant differences between treatments within each soil type based on Tukey's test (P < 0.05). MF = Maize flowering; MS = Maize senescence; CF = Cowpea flowering; CS = Cowpea senescence. Fig. 3. Taxonomic (A) richness and (B) diversity based on genus level at 97% similarity of the 16S rRNA gene sequencing. Error bars represent the standard deviation of three independent replicates. Different lower case letters refer to significant differences between treatments based on Tukey's HSD test (P < 0.05). MF = Maize flowering; MS = Maize senescence; CF = Cowpea flowering; CS = Cowpea senescence. A.S.F. de Araujo, et al. Applied Soil Ecology 143 (2019) 26–34 29

3.3. Correlation between the bacterial community and soil properties and community network The correlation of bacterial phyla with soil chemical properties was analyzed by Spearman's rank correlation (Table 2). The soil factors that presented correlation with more number of bacterial phyla were pH (9 phyla in total), followed by P (6), EC (4) and CEC (3). Interestingly, the phylum that presented more correlation with chemical parameters was Gemmatimonadetes (4 factors in total). The co-occurrence network analysis showed that, in general, rhizosphere samples at the senescence stage presented the highest complexity compared to the flowering stage and bulk soil for both plants species (Fig. 5). In all treatments, the number of positive correlations was higher than negatives; however, we observed an increase of negative correlation in the senescence stage for both plants. More speci- fically, CS was the most complex network, with 152 significant correlations (52% positives), high modularity (8.43) and average degree and coefficient of 4.28 and 0.369, respectively. The MS network was the second most complex, with 133 correlations (52% positives), modularity of 5.22 and average degree and coefficient of 4.83 and 0.516, respectively. Also, based on the number of connections and betweenness centrality we identified the key groups in each network. For bulk soil, the keynote was DA101; for MF were Bacillus and Mycobacterium; for MS were Bacillus, Mycobacterium, and Symbiobacterium; for CF was DA101, and for CS were Bacillus and Microlunatus. 4. Discussion Our study assessed the bacterial community in the rhizosphere of maize (grass) and cowpea (legume) as compared with bulk soil during the flowering and senescence period in subsequent cultivation. In a general view, our study found a dominance of Actinobacteria and Proteobacteria that represented > 20% of the sequences. There are several studies that have found a high abundance of Proteobacteria and Actinobacteria in soils (Liebner et al., 2008, Männistö et al., 2013; Araújo et al., 2017). However, we observed a lower abundance of Acidobacteria in comparison with Actinobacteria and Proteobacteria. In this study, the soil presented suitable pH (6.2) and content of nutrients, i.e. improved soil fertility, and these conditions contributed for a decreased abundance of Acidobacteria, which are better adapted to acid soils and oligotrophic environments (Kielak et al., 2016). In contrast, Actinobacteria and Proteobacteria were favored by these soil conditions since they predominate in neutral and nutrient-rich soils (Yang et al., 2017). The results showed changes in the bacterial community in the rhizosphere as compared with bulk soil, as confirmed by RDA. For the structure of the bacterial community in both rhizosphere soils, the NMDS pattern showed segregation between plant species and, in a minor extent, differences between plant developmental stages. This result was expected since the rhizospheric environment influences the bacterial community via root exudation and nutrient input. Thus, Actinobacteria, Chlamydiae, and Firmicutes increased their abundance Fig. 4. Venn diagrams showing the number and proportion of unique and shared genera at 97% similarity between (A) bulk soil and rhizosphere microbiome of maize; (B) bulk soil and rhizosphere microbiome of cowpea; and (C) between rhizosphere samples at the different developmental stage. MF = Maize flowering; MS = Maize senescence; CF = Cowpea flowering; CS = Cowpea senescence. A.S.F. de Araujo, et al. Applied Soil Ecology 143 (2019) 26–34 30