ORIGINAL RESEARCH published: 14 May 2021 doi: 10.3389/fmicb.2021.639997 Identifying the Specific Root Microbiome of the Hyperaccumulator Noccaea brachypetala Growing in Non-metalliferous Soils Soledad Martos 1* , Sílvia Busoms 1 , Laura Pérez-Martín 1 , Mercè Llugany 1 , Catalina Cabot 2 and Charlotte Poschenrieder 1 1 Plant Physiology Laboratory, Bioscience Faculty, Universitat Autònoma de Barcelona, Bellaterra, Spain, 2 Department of Biology, Universitat de les Illes Balears, Palma, Spain Edited by: Michel Chalot, Noccaea brachypetala is a close relative of Noccaea caerulescens, a model plant Université Bourgogne Franche-Comté, France species used in metal hyperaccumulation studies. In a previous survey in the Catalan Reviewed by: Pyrenees, we found two occidental and two oriental N. brachypetala populations Shuo Jiao, growing on non-metalliferous soils, with accumulated high concentrations of Cd and Northwest A and F University, China Alison Bennett, Zn. Our hypothesis was that the microbiome companion of the plant roots may The Ohio State University, influence the ability of these plants to absorb metals. We performed high-throughput United States sequencing of the bacterial and fungal communities in the rhizosphere soil and Kristian Harris, The Ohio State University, rhizoplane fractions. The rhizobiomes and shoot ionomes of N. brachypetala plants United States, in collaboration with were analyzed along with those from other non-hyperaccumulator Brassicaceae species reviewer AB found at the same sampling locations. The analyses revealed that microbiome richness *Correspondence: Soledad Martos and relative abundance tended to increase in N. brachypetala plants compared to non- soledad.martos@uab.cat hyperaccumulator species, regardless of plant location. We confirmed that the root compartment is a key factor in describing the community composition linked to the Specialty section: This article was submitted to cohabiting Brassicaceae species, and the rhizoplane fraction contained the specific and Microbe and Virus Interactions with rare taxa associated with each species. N. brachypetala plants harbored a similar relative Plants, abundance of fungi compared to the other plant hosts, but there was a notable reduction a section of the journal Frontiers in Microbiology in some specific taxa. Additionally, we observed an enrichment in the hyperaccumulator Received: 11 December 2020 rhizoplane of previously described metal-tolerant bacteria and bacteria involved in Accepted: 06 April 2021 nitrogen cycling. The bacteria involved in the nitrogen cycle could contribute indirectly Published: 14 May 2021 to the hyperaccumulator phenotype by improving soil quality and fertility. Our results Citation: Martos S, Busoms S, indicate that N. brachypetala captures a particular prokaryotic community from the soil. Pérez-Martín L, Llugany M, Cabot C This particular prokaryotic community may benefit the extraction of metal ions and/or and Poschenrieder C (2021) improve plant nutrition. Our research identified satellite groups associated with the root Identifying the Specific Root Microbiome of the Hyperaccumulator niche of a hyperaccumulator plant that may assist in improving biological strategies in Noccaea brachypetala Growing heavy metal remediation. in Non-metalliferous Soils. Front. Microbiol. 12:639997. Keywords: high-throughput sequencing (deep sequencing), Noccaea brachypetala, root-associated microbial doi: 10.3389/fmicb.2021.639997 communities, Zn/Cd hyperaccumulator, non-metalliferous soil Frontiers in Microbiology | www.frontiersin.org 1 May 2021 | Volume 12 | Article 639997 Martos et al. Unravelling the Root-Community of N. brachypetala GRAPHICAL ABSTRACT INTRODUCTION increased resistance to biotic stresses (Poschenrieder et al., 2006; Cappa and Pilon-Smits, 2014). Metal-rich soils are harsh environments for establishing life. The ability to mobilize, chelate, and transport metals is However, metallophytes have evolved different strategies to directly linked to soil metal bioavailability and the genetic survive under these restrictive growth conditions. Among these background of the plant. However, recent research highlights plant species, hyperaccumulators harbor metal concentrations that plant-associated microorganisms can significantly enhance within aerial tissues at levels far exceeding those present in metal availability. The soil microbial community may greatly their native soil. Without showing significant signs of toxicity, contribute to the hyperaccumulator phenotype (Visioli et al., they can achieve 100-fold higher shoot metal concentrations 2015a). Various studies have reported a high diversity of than non-hyperaccumulator plants (Van der Ent et al., 2013). microbes in the rhizosphere and endosphere (Thijs et al., Hyperaccumulation appears to have evolved several times within 2017). Under unfavorable conditions of metalliferous soils, vascular plants, and the Brassicaceae family contains a high rhizosphere microorganisms may increase or decrease metal percentage of the total hyperaccumulating taxa (Kramer, 2010). availability (Farinati et al., 2011). Previous studies have Noccaea caerulescens (formerly Thlaspi caerulescens) belongs to demonstrated both effects in Noccaea with root-bacteria this family and is one of the best-known hyperaccumulators of favoring metal uptake (Whiting et al., 2001), whereas fungal Cd, Ni, and Zn (Assunçao et al., 2003). This genus contains mycorrhiza reduced plant absorption (Vogel-Mikus et al., several species, including Noccaea praecox (Likar et al., 2010), 2006). However, a proper cohort of microbes may increase Noccaea camlikensis (Aksoy et al., 2015), Noccaea brachypetala the potential of hyperaccumulators to extract metals. Various (Martos et al., 2016), Noccaea kovatsii (Misljenovic et al., studies have confirmed that specific bacteria promote the 2020), and Noccaea japonica (Nishida et al., 2020), which have uptake of Ni in N. caerulescens (Aboudrar et al., 2013; also been described as metal hyperaccumulators. Interestingly, Visioli et al., 2014, 2015b). Therefore, rhizosphere-associated metallophytes are not restricted to metalliferous soils, and it has microbes from hyperaccumulators are a potential source of been suggested that Zn/Cd/Pb hyperaccumulation can evolve on genotypes for biotechnological applications in phytoremediation non-metalliferous lands (Besnard et al., 2009). When growing and agriculture (Rajkumar et al., 2012). Characterization of on common soils, the metal concentrations in their shoots the rhizobiome of hyperaccumulators is needed to enable can achieve unusually high levels. The most popular theory the development of these applications. Despite this, a limited suggests the evolutionary advantage of this behavior is an number of hyperaccumulator species have been investigated Frontiers in Microbiology | www.frontiersin.org 2 May 2021 | Volume 12 | Article 639997 Martos et al. Unravelling the Root-Community of N. brachypetala (Martinez-Diz et al., 2019; Macia-Vicente et al., 2020; Thiergart et al., 2020). Several studies have shown that plant exudates can favor specific microbiome communities mutualistically (Bulgarelli et al., 2013; Stringlis et al., 2018). Exudated organic compounds nurture rhizomicrobiota, allowing the root communities to modulate the plant phenotype. A well-reported case revealed an increased metal tolerance in plants associated with a metal-tolerant microbiome (Luo et al., 2014; Visioli et al., 2015a). However, the potential role of microbiota from non-metalliferous soils to the hyperaccumulation phenotype remains unknown. In the last few years, vast progress has been made in elucidating the composition and dynamics of root communities in the Brassicaceae family. However, additional studies in natural conditions are required. We investigated the impact of plant identity, location, and root compartment on the microbial diversity and community structure of the hyperaccumulator N. brachypetala and other Brassicaceae species (non-hyperaccumulators) coinhabiting undisturbed and non- metalliferous soils of the Catalan Pyrenees. MATERIALS AND METHODS Plant and Soil Sampling Plant and soil materials were collected from four sampling sites in the Catalan Pyrenees. Two sites were located in the Pallars region (Mauri and Alòs), and the other two were in the Ripollès region (Freser and Núria). All sites were located in the alpine biome of the Catalan Pyrenees (Figure 1A). Plants from Mauri and Aneu were grouped as Western populations (WPs), and Freser and Nuria plants were grouped as the Eastern populations (EPs). Both locations are separated by more than 150 km. The geographic FIGURE 1 | Experimental design. (A) Climatic map of Catalonia with the coordinates of each sampling site are listed in Table 1. sampling locations of plants and soils indicated by stars. Western populations At each site, N. brachypetala specimens [previously confirmed are composed form locations of Alòs and Mauri, whereas Eastern populations as Cd/Zn hyperaccumulators (Martos et al., 2016)] and other are from Núria and Freser. (B) Schematic drawing of the analyzed compartments of the root. DNA of the microbiome from the rhizoplane and non-hyperaccumulator Brassicaceae species growing beside the rhizosphere was extracted followed by 16S and ITS sequencing of the specimens were also collected. All collectors wore gloves. A list bacteria and fungi, respectively. of sampled species is provided in Table 1. Selected individuals were excavated with their rhizosphere soil without disturbing the plant root system. Plants were immediately transferred into (Alford et al., 2010; Ma et al., 2011; Sessitsch et al., 2013). polyethylene bags to avoid desiccation. From each site, bulk soil Within the genus Noccaea, the rhizosphere bacteria of Noccaea (five replicates per site) exempt from root or plant debris was goesingense were characterized by Idris et al. (2004), and collected at a depth of 10 cm using a 7-cm diameter Edelman drill mycorrhizal colonization in N. praecox has been investigated (Eijkelkamp, Giesbeek, Netherlands). All collected plant and soil (Pongrac et al., 2007). The microbiomes of other Noccaea samples were refrigerated in a portable icebox and immediately hyperaccumulator species remain unknown. transported to the laboratory for further processing. Recent metagenomic studies revealed that soil properties are the major drivers for the differentiations in the assembly of Analysis of Metal Content From Soil and root communities in Arabidopsis thaliana and in some related Plant Samples species, including the Cd-Zn hyperaccumulator, Arabidopsis Field-collected plants were separated into aerial and below halleri (Bulgarelli et al., 2012; Lundberg et al., 2012; Schlaeppi ground tissue. All aerial tissue was carefully washed with distilled et al., 2014; Glynou et al., 2018). Plant genotype research remains water and oven-dried at 60◦ C for 2 days. Dried shoots (three conflicted. Some authors affirm that plant genotype is a key factor replicates/plant species) were finely powdered using a mortar and in microbiome assembly (Burns et al., 2015; Li et al., 2018); others pestle. In parallel, collected bulk soils (five replicates per sampling do not consider this essential (Glynou et al., 2018; Macia-Vicente site) were homogenized through a 2-mm sieve. Approximately et al., 2020). Proximity to the root is also considered a relevant 1 g of sieved bulk soil and 0.1 g of powdered shoots were acid- parameter that increases soil microbiota diversity and structure digested (HNO3 : H2 O2 69%: 30%, 5:2 vol/vol) in a hot-block Frontiers in Microbiology | www.frontiersin.org 3 May 2021 | Volume 12 | Article 639997 Martos et al. Unravelling the Root-Community of N. brachypetala TABLE 1 | Sampled plants for the rhizosphere microbioma sequencing. fraction was suspended in 50 mL of 0.9% NaCl solution. Roots were then placed into sterile 100-mL tubes containing 50 mL Brassicaceae species DNA extracted Location compartment of 0.9% NaCl solution and washed by manually shaking for 2 min (RP1 fraction). Roots were subsequently immersed in 1 Noccaea brachypetala R, RP1, RP2 Alòs (Western 0.9% NaCl solution and 0.01% Tween 80 and then washed by population) 42◦ 450 2 N. brachypetala R, RP1, RP2 manually shaking for 2 min (RP2 fraction). Rhizosphere and 36.79200 N, 1◦ 40 3 N. brachypetala R, RP1, RP2 13.800 E rhizoplane RP1 and RP2 suspensions were then homogenized 4 N. brachypetala R, RP1, RP2 with gentle agitation (300g, OVAN Orbital Midi OM10E 5 N. brachypetala R, RP1 shaker, 25◦ C, 90 min) followed by centrifugation (5,000g, 25◦ C, 6 Capsella bursa-pastoris R, RP1 10 min) in sterile 100-mL tubes to concentrate soil particles 7 Arabis sp. R, RP1 into a pellet. Then, the supernatant was filtered through a 8 Cardamine impatiens R, RP1 1-mm sieve to remove suspended residuals. Fractions were 9 N. brachypetala R, RP1, RP2 Mauri (Western frozen in liquid nitrogen and stored at −80◦ C until required population) 42◦ 340 10 N. brachypetala R, RP1, RP2 for DNA extraction. 44.50800 N, 1◦ 00 11 N. brachypetala R, RP1, RP2 30.52800 E 12 N. brachypetala R, RP1 Library Generation and Sequencing of 13 Arabis sp. R, RP1 14 N. brachypetala R, RP1, RP2 Núria (Eastern Bacterial and Fungal DNA Regions population) 42◦ 210 The total genomic DNA from the rhizosphere and rhizoplane 15 N. brachypetala R, RP1, RP2 52.16400 N, 2◦ 100 compartments (approximately 200 mg) of the N. brachypetala 16 Capsella sp. R, RP1 5.8800 E specimens and the non-hyperaccumulator Brassicaceae species 17 Iberis sp. R, RP1 was extracted using the NucleoSpin Soil kit (Macherey- R 18 N. brachypetala R, RP1, RP2 Freser (Eastern population) 42◦ 220 Nagel GmbH & Co. KG, Düren, Germany), according to 19 Iberis sp. R, RP1 54.62400 N, 2◦ 120 the manufacturer’s instructions. Extracted DNA was purified 20 Arabis sp. R, RP1 56.26800 E using the QIAamp DNA Stool Mini Kit (Qiagen, West Genomic DNA has been extracted from the rhizosphere soil (R) and Sussex, United Kingdom) and then quantified using the rhizoplane (RP1 and RP2). PicofluorTM method (PicoGreen dsDNA Quantitation Reagent, Molecular Probes, Inc., Eugene, OR, United States). Quality digestion system (SC154-54-Well Hot BlockTM , Environmental was measured using Nanodrop 2000 (Thermo Fisher Scientific, Express, SC, United States). The concentrations of the selected DE, United States) and confirmed by the OD A260 /A280 metal elements (Cd, Cu, Pb, and Zn) and additional elements between 1.8 and 2.0. (Ca, Co, Fe, K, Mg, Mn, Mo, Na, P, and S) were determined The prokaryotic 16S ribosomal RNA gene (16S by inductively coupled plasma (ICP) mass spectrometry (Perkin rRNA) and the internal transcribed spacer (ITS) were Elmer Inc., ELAN 6000, MA, United States) or ICP–optical the barcodes used for the metagenomic study of the emission spectrometry (Thermo Jarrell-Ash, model 61E Polyscan, bacteria and fungi, respectively. Microbial populations England) (Bech et al., 2012). Internal references of the certified were obtained by high-throughput sequencing, targeting material BCR 62 Olea europaea and CRM 142 R Light Sandy Soil the V3–V4 region of 16S rRNA and ITS1 of the ITS were included for quality control. region, using the MiSeq Reagent Kit v2 (500 cycle) R Statistical differences among groups in the element content (MiSeq, Illumina, San Diego, CA, United States). of the soils or leaves were assessed using analysis of variance The V3–V4 region was amplified using primers (two-way ANOVA). Normal distribution, homoscedasticity, and 341F (50 -CCTACGGGAGGCAGCAG-30 ) and 806R the independence of errors were checked prior to analyses. Data (50 -GGACTACHVGGGTWTCTAAT-30 ). The ITS1 region was were log-transformed when necessary. Multiple comparisons amplified using primers ITSF (50 -CTTGGTCATTTAGAGGA of group means were analyzed by Tukey honestly significant AGTAA-30 ) and ITS2 (50 -GCTGCGTTCTTCATCGATGC-30 ). difference using JMP (version 13.0, SAS Institute Inc., 1989- R 2019). Quality Control of Sequencing Data Sequence reads of both datasets (16S and ITS) from the MiSeq Sampling of Rhizosphere and Rhizoplane Illumina system were processed using QIIME v.1.9.1 (Caporaso R Compartments et al., 2010). The quality filter of the already demultiplexed The soil adhering to the roots of each sampled specimen sequences was performed at a maximum unacceptable Phred was separated into two fractions: rhizosphere and rhizoplane quality score of Q20. The resulting reads were clustered into (Figure 1B). The sampling method for the rhizosphere and operational taxonomic units (OTUs) using uclust with 97% rhizoplane compartments followed the protocols described by sequence similarity and the subsampling pick open reference Barillot et al. (2013) and Edwards et al. (2015), with minor method (Rideout et al., 2014) at 10% of the subsampled modifications. In brief, the rhizosphere fraction was manually sequences. Seven samples with number of reads fewer than 5,000 removed by vigorously shaking for 2 min until approximately or a percentage of chimeras higher than 70% were removed 1 mm of soil was firmly attached to the roots. The rhizosphere from the analysis (Supplementary Dataset 1). Rarefaction curves Frontiers in Microbiology | www.frontiersin.org 4 May 2021 | Volume 12 | Article 639997 Martos et al. Unravelling the Root-Community of N. brachypetala showed that 90.7% of the sequenced samples achieved a plateau and minimum λ threshold set to 0.01, for all networks. Bacteria zone, indicating that no additional sequencing effort was required and fungi domains were analyzed together, and independent (Supplementary Figure 1). After filtering, a total of 4,752,917 networks were constructed for each root compartment and high-quality sequences were obtained from 43 samples: 2,442,668 plant species. Fundamental OTU networks (keystones) were the from the 16S region (bacteria) and 2,310,249 sequences from highest “betweenness” and “node count” in more than 50% the ITS region (fungi) (Supplementary Dataset 1). The media of the distribution. Each adjacency matrix was then converted range of sequence reads per sample was 56,806 for bacteria and into an igraph object and visualized as a network using the 53,726 for fungi. adj2igraph() and plot network() functions in SpiecEasi. Representative sequences were assigned to taxonomic groups using the bacterial 16S GreenGenes v.13.8 reference database at a 90% confidence threshold and sequence alignment using Differential Abundance Analysis QIIME. The taxonomy of fungi/oomycetes was assigned to full To establish which taxa were associated more commonly with ITS sequences from the UNITE_v2020 database (Nilsson et al., N. brachypetala compared to non-hyperaccumulator species in 2018) using the DADA2 software package (Callahan et al., 2016). each compartment, a differential abundance analysis (DAA) was To reduce artifacts in the dataset, we removed singletons, OTUs performed using the DESeq2 package. This method models the with relative abundances across all samples less than 0.005% observed abundances using negative binomial distribution after (Bokulich et al., 2012), and OTUs assigned as chloroplasts or normalizing data with the corresponding scaling methods to mitochondria. The OTU table was resampled to the minimum account for differences in sampling fractions, recommended for number of sequences per sample (Weiss et al., 2017), leading to small datasets (< 20 samples per group) (Paulson et al., 2013; a total of 3,110 OTUs in the bacterial domain and 1,060 OTUs in Weiss et al., 2017; Lin and Peddada, 2020). Plot visualization was the fungal domain. performed using the ggplot package. Diversity Metrics and Relative Abundance Metabolic and Functional Analysis To determine the metabolic potential of the N. brachypetala Output data from QIIME and DADA2 were merged and used bacterial communities that were differentially abundant, a to obtain an OUT Phyloseq object generated with R (McMurdie predictive functional analysis of bacterial communities was and Holmes, 2012). The α diversity was calculated using the performed using FAPROTAX software (Louca et al., 2016). Shannon index. The Kruskal–Wallis significance test was used Similarly, functional annotation of the N. brachypetala– for all pairwise combinations (between root compartments specific fungal communities was obtained using FUNGuild and plant species). (Nguyen et al., 2016). Based on the OTU relative abundance, a β-diversity assessment was performed using Bray–Curtis metrics. The resulting distance/dissimilarity matrices were subjected to ordination using principal coordinate analysis. Non-metric multidimensional scaling was analyzed based on RESULTS the Bray–Curtis distances between samples using the Vegan package (Oksanen et al., 2013). The significance of each factor Soil and Leaf Mineral Content [root compartment, location, plant species, and extraction The analyses of the metal concentrations catalogued all the soil methodology (RP1, RP2)] and all the edaphological variables samples as non-metalliferous. The bulk soil collected with the (nutrient mineral content) in the microbial composition were different plant species revealed slight differences in the metal assessed by permutational multivariate ANOVA (PERMANOVA) contents according to their location (EP/WP) (Figure 2A and using the Adonis function (R package vegan). Prior to analysis, Supplementary Dataset 2). For example, WPs of N. brachypetala edaphological data were standardized (Z scores) to avoid bias accumulated higher levels of Cd and Zn compared to the caused by the order of magnitude. All factors were independently EPs. However, the principal component analysis (PCA) of the analyzed. Factors that did not condition the microbiome of any soil samples did not differentiate the samples into location analyzed group were not further analyzed. clusters (Figure 2B). The PCA of the ionomic data identified The assessment of the taxonomic composition was performed the hyperaccumulation trait as the key clustering factor in the at the phylum level using staked plots. To detect differences samples of N. brachypetala with the other Brassicaceae species in in bacterial and fungal community composition, we performed independent groups (Figure 2C). The metal content in the leaves ANOVA pairwise comparisons between root compartments of N. brachypetala was 200 times higher for Cd and more than 100 and plant species. times higher for Zn than for the non-hyperaccumulator species (Figure 2A and Supplementary Dataset 2). The data confirmed Co-occurrence Network the Cd and Zn hyperaccumulator ability of N. brachypetala. The OTU phyloseq object was converted into an adjacency matrix Other elements, such as K, Cu, P, and Pb, were also increased in based on covariance by the spiec.easi() function of the SpiecEasi N. brachypetala leaves but at lower levels (from 1.5 to 10 times package (version 1.0.7) and using MB’s neighborhood selection higher levels, Supplementary Dataset 2). Frontiers in Microbiology | www.frontiersin.org 5 May 2021 | Volume 12 | Article 639997 Martos et al. Unravelling the Root-Community of N. brachypetala FIGURE 2 | Soil and leaf nutrient mineral content and factors potentially influencing the microbiome. (A) Bar plots for soil and leaf metal content of hyperaccumulator and non-hyperaccumulator plants from Western (WP in orange) and Eastern populations (EP in yellow). Error bars correspond to the standard error. Different letters indicate statistically significant differences among treatments at P < 0.05. (B) Principal component analysis of the soil mineral content in the Western (orange dots) and Eastern populations (yellow dots). (C) Principal component analysis of the leaf mineral content. Hyperaccumulator plants are clustered in red and non-hyperaccumulator species in gray. (D) Summarized table of the PERMANOVA on the effect of the soil mineral content, root compartment, plant species, and location on the microbiome community (grouping bacteria and fungi). (E) Summarized table of the PERMANOVA on the effects of location and root compartment or plant species on the microbiome community (grouping bacteria and fungi). Root Compartment and Plant Species of location was irrelevant even when plant species and root Condition the Microbial Composition compartments were analyzed separately (Figure 2E). The statistical analysis (PERMANOVA) of the OTUs from the Additionally, we evaluated the effect of the extraction method bacteria and fungi domains revealed that the root compartment on the rhizoplane. The multivariate ANOVA confirmed no and plant species were relevant factors in explaining the statistically significant difference between the extraction methods variability of microbiome data (Figure 2D). However, the weight (RP1 and RP2) (Supplementary Datasets 1,2). Therefore, of the factor “plant location” was not significant. The effect we selected root compartment and plant species as factors Frontiers in Microbiology | www.frontiersin.org 6 May 2021 | Volume 12 | Article 639997 Martos et al. Unravelling the Root-Community of N. brachypetala FIGURE 3 | Factors influencing the bacterial and fungal community and diversity. Non-metric multidimensional scaling visualization of the Bray–Curtis results of the bacterial domain (A) and fungal domain (B). Plant species are differentiated by color (red is the hyperaccumulator; gray is the non-hyperaccumulator), and root compartments by shape (dots represent the rhizoplane; triangles represent the rhizosphere). (C) Summarized table of the PERMANOVA of the Bray–Curtis distances on the effect of root compartment and plant species for the bacterial and fungal domain, separately. (D) α Diversity boxplot of the bacterial domain. (E) α Diversity boxplot of the fungal domain. The Shannon index divided by root compartment and plant genotype. The x axis indicates root compartment, whereas plant genotype is differentiated by color (red is N. brachypetala; gray represents non-hyperaccumulator species). (F) Summarized table of the Kruskal–Wallis analysis from the Shannon index on the effect of root compartment of the bacterial and also the fungal domain. to investigate their effects on microbial diversity, structure, Moreover, rhizoplane diversity in the bacteria community was and composition. increased in N. brachypetala (hyperaccumulator) compared to the non-hyperaccumulator species (P < 0.01, Figure 3F). Fungal diversity did not vary between plant species (P > 0.05, Figure 3F). High Influence of Root Compartment and Plant Species in the Bacterial Community More Complex Network in the Root-associated microbes were divided into bacteria and fungi Rhizoplane of N. brachypetala to investigate factors that could drive the microbial community The interaction patterns of the overall microbiome were analyzed within each domain (Figures 3A–C). The root compartment and revealed that the hyperaccumulator networks were globally and plant species in the bacterial composition possessed larger more interconnected than the non-hyperaccumulator networks proportions of variation (19%, P < 0.01, and 12%, P < 0.05, (Figures 4A–D). N. brachypetala established more connections respectively) (Figures 3A, C). In contrast, the Bray–Curtis among phyla, as indicated by the increased number of total edges analysis indicated that the root compartment was a unique and when compared to the other species (Figure 4E). Interestingly, significant factor describing the fungal composition, whereas the the rhizoplane of the hyperaccumulator displayed the most plant species was not relevant (Figures 3B, D). complex interaction network with more nodes and edges The α-diversity boxplots (using the Shannon index) showed (Figures 4A,E). In all the represented networks, the central higher diversity in the rhizosphere than in the rhizoplane module was composed of bacterial phyla (circles) surrounded by compartment for both microbial groups (Figures 3D,E). fungal phyla (triangles) (Figures 4A–D). Frontiers in Microbiology | www.frontiersin.org 7 May 2021 | Volume 12 | Article 639997 Martos et al. Unravelling the Root-Community of N. brachypetala FIGURE 4 | Co-occurrence network. Potential relationship among the bacterial and fungal OTUs in the rhizoplane (A,B) and the rhizosphere (C,D) of N. brachypetala (Hyp) (A,C), and the non-hyperaccumulator species (Non-hyp) (B,D). (E) Table with the main parameters of the co-occurrence networks. Frontiers in Microbiology | www.frontiersin.org 8 May 2021 | Volume 12 | Article 639997 Martos et al. Unravelling the Root-Community of N. brachypetala FIGURE 5 | Bacterial and fungal community composition and relative abundance. (A) Means and standard errors (SE) of the relative abundances of the bacteria domain at phylum level. Differences and summarized results of an ANOVA pairwise comparison between N. brachypetala and non-hyperaccumulators at each soil compartment were calculated by the log2 ratio. (B) Bacterial relative abundance at the phylum level in a staked plot. The x axis indicates root compartment and plant genotype. (C) Means and standard errors (SE) of the relative abundances at phylum level on the fungi domain. Differences and summarized results of an ANOVA pairwise comparison between N. brachypetala and the non-hyperaccumulators at each soil compartment calculated by the log2 ratio. (D) Fungal relative abundance (at the phylum level) in a staked plot. The x axis indicates the root compartment and plant genotype. The most important OTUs for the overall structure Higher Relative Abundance of Bacteria (keystones) were identified. The N. brachypetala rhizoplane and Fungi in N. brachypetala With Some presented the highest number of keystones. Proteobacteria and Actinobacteria species and Ascomycota and Basidiomycota Exclusive Taxa species may play an important role in plant-microbial feedback To detect differences in the composition of bacterial and as they possess the largest number of keystones in the four fungal communities, the absolute abundance was transformed networks analyzed (Figures 4A–D). into relative values. Data were clustered by phyla and Frontiers in Microbiology | www.frontiersin.org 9 May 2021 | Volume 12 | Article 639997 Martos et al. Unravelling the Root-Community of N. brachypetala FIGURE 6 | Differential abundance analysis (DAA) and metabolic functions of bacterial OTUs in N. brachypetala and non-hyperaccumulator species. Bubble plots of the differential abundant bacterial OTUs on the rhizoplane (A) and rhizosphere (B). The x axis indicates the log2 ratio (hyperaccumulators vs. non-hyperaccumulators). The y axis indicates bacteria at the taxonomical level of order. Bubble colors indicate the associated phylum. (C) The co-occurrence network of the DAA bacterial OTUs in the rhizoplane of N. brachypetala and a dendrogram with the annotated functions. (D) Table of metabolic functions of the DAA bacteria in N. brachypetala in each root compartment. grouped by hyperaccumulator/non-hyperaccumulator and The bacterial domain was dominated by six major phyla: rhizosphere/rhizoplane (Figure 5). The relative abundance Acidobacteria, Actinobacteria, Bacteroidetes, Planctomycocetes, of each sample is shown in Supplementary Figure 3. Proteobacteria, and Verrucomicrobia, comprising 90.98% of Frontiers in Microbiology | www.frontiersin.org 10 May 2021 | Volume 12 | Article 639997 Martos et al. Unravelling the Root-Community of N. brachypetala the total relative abundance (Figures 5A,B). The remaining of bacteria (by order) in N. brachypetala (Figure 6A). identified phyla (22) showed a relative abundance lower than The highest number of differentially abundant bacteria 3% and were considered rare phyla. Confronting the relative belonged to five dominant phyla (Acidobacteria, Bacteroidetes, abundance between plant groups, four phyla (Fibrobacteres, Planctomycetes, Proteobacteria, and Verrucomicrobia). Four Gemmatimodetes, Planctomycetes, and Verrumicrobia) showed rare phyla (Chloroflexi, Fibrobacteres, Gemmatimonadetes, significantly higher relative abundance in the rhizoplane of the and Nitrospirae) were also abundant in the hyperaccumulator hyperaccumulator compared to the non-hyperaccumulators rhizoplane. At the taxonomic level of family, N. brachypetala (Figure 5A). Additionally, two phyla (BRC1 and OP11) were presented 108 families in the rhizoplane and 112 in the absent on the non-hyperaccumulator rhizoplane but were rhizosphere with a higher relative abundance than in the non- present on the hyperaccumulator. Hereafter, we refer to these hyperaccumulator species (Supplementary Dataset 3). The as “exclusive taxa” of N. brachypetala (Figure 5A). Three more number of more abundant families associated with the non- phyla (Acidobacteria, Elusimicrobia, and WS3) were more hyperaccumulator species was in an order of magnitude lower, abundant in the rhizosphere of the hyperaccumulator. In of 20 and 15 in the rhizoplane and rhizosphere, respectively. The contrast, the phyla Fibrobacteres and Proteobacteria exhibited most prevalent taxa appearing repetitively were Candidatus increased abundance in the non-hyperaccumulator rhizosphere Solibacter and the genera Flavobacterium, Koribacter, and rhizoplane, respectively. The phylum Fibrobacteres Opitutus, Planctomyces, Pseudomonas, and Streptomyces was in a higher abundance in the rhizosphere of the non- (Supplementary Dataset 3). hyperaccumulator and, at the same time, had a larger presence in The DA OTUs from N. brachypetala were explored to identify the rhizoplane of the hyperaccumulator. To sum up, six rare and their metabolic functions (Figures 6C,D). Functional analysis of three dominant phyla differed between N. brachypetala and the the hyperaccumulator bacteria identified 22 metabolic functions. non-hyperaccumulators in both root compartments. The most representative functions were those related to the Regarding the fungal domain, the dominant phyla were nitrogen cycle (Figure 6C). The rhizoplane with 20 of 22 obtained Ascomycota (53%) and Basidiomycota (36%) (Figures 5C,D). functions was the most diverse functional root compartment. The remaining identified phyla (10) had a combined relative Rhizoplane bacteria were involved in functions related to abundance of less than 3% (Figure 5C). The relative the nitrogen cycle, whereas the rhizosphere bacteria were abundance of each sample is provided in Supplementary exclusively linked to cyanobacteria and oxygenic photautotrophy Figure 3. The hyperaccumulator harbored two exclusive (Figure 6C). The interaction network in the DA bacteria of fungal phyla (Mucoromycota and Olpidiomycota) in the rhizoplane and the associated functions are provided in the rhizoplane and one (Monoblepharomycota) in the Figure 6D. The observed pattern revealed that interconnected rhizosphere (Figure 5C). The major phyla were in different bacteria were involved in several metabolic functions. Aerobic composition according to the plant host. Ascomycota were nitrite oxidation was assigned to isolated bacteria (black and more abundant in the rhizosphere of the hyperaccumulator, gray dots), with the nitrification and nitrite respiration assigned whereas Basidiomycota had a larger presence in the rhizosphere to non-connected bacteria (blue and green dots). Therefore, the of the non-hyperaccumulators. Conversely, in the non- functions were not diversified within the connected groups. hyperaccumulators, three phyla were more abundant in their rhizoplane (Chytridiomycota, Kickxellomycota, and Monoblepharomycota) and one (Basidiomycota) in the The Fungal Community of rhizosphere. Additionally, six minor phyla were composed differently between the plant species. N. brachypetala Presents a Reduced Generally, the taxa abundance by prevalence revealed Differential Abundance Compared to the that the most abundant bacteria were also the most Non-hyperaccumulator Species prevalent. In contrast, the most abundant species The DAA of the fungi revealed more abundant fungi in the fungal domain exhibited a lower prevalence (by order) in both root compartments of the non- (Supplementary Figure 2). hyperaccumulator species (Figures 7A,B). Some of the limited fungal species in the rhizoplane of N. brachypetala The Bacterial Community of were Candida railenensis, Clarireedia bennettii, Oidiodendron chlamydosporicum, Trechispora byssinella, and Acremonium N. brachypetala Is Globally More Diverse rutilum (Supplementary Dataset 3). With Higher Relative Abundance in the Conversely, we also detected seven fungi OTUs exclusive Rhizoplane Compartment to the N. brachypetala plants (Supplementary Dataset 3). To gain further insight into the taxa distribution, the bacterial To our knowledge, there are no tools available to classify OTUs were subjected to a DAA comparing N. brachypetala fungal communities into metabolic functional groups, but with the non-hyperaccumulator species (Figure 6). The DAA FUNGuild (Nguyen et al., 2016) predicts function at the guild results were presented at the taxonomic level of order, the level. Figure 7C indicates that the four fungal OTUs exhibit lowest level that was visually informative. The bacterial analyses different trophic strategies. Among them, the genus Xerocomus of the rhizoplane showed an important accumulation of dots deserves highlighting because it is a symbiont described as with positive log2 values, indicating a higher abundance an ectomycorrhiza. Frontiers in Microbiology | www.frontiersin.org 11 May 2021 | Volume 12 | Article 639997 Martos et al. Unravelling the Root-Community of N. brachypetala FIGURE 7 | Differential abundance analysis (DAA) and guilds of fungal OTUs in N. brachypetala and non-hyperaccumulator species. Bubble plots of the differential abundant fungal OTUs in the rhizoplane (A) and rhizosphere (B). The x axis indicates the log2 ratios (hyperaccumulators vs. non-hyperaccumulators). The y axis indicates fungi at the taxonomical level of order. Bubble colors indicate the associated phylum. (C) Table of trophic mode and guilds of DAA fungi in N. brachypetala in each root compartment. DISCUSSION using high-throughput sequencing in order to identify novel taxa that could be involved in metal hyperaccumulation in a Microbiota associated with plant roots are considered important cultivation-independent manner. mediators of plant–soil interactions. In the case of metal According to our results, the hyperaccumulator, hyperaccumulator plants, rhizosphere bacteria have been N. brachypetala, exhibited an increased richness in the bacterial associated with positive impacts on plant growth and increased community compared to the companion non-hyperaccumulator metal accumulation (Visioli et al., 2015a). Therefore, the root Brassicaceae species. This hyperaccumulator also presented the microbiome of hyperaccumulators has attracted increased most complex and highly interconnected interaction network. attention for potential biotechnological applications in In our study, the rhizoplane of N. brachypetala harbored a phytoremediation. Despite the interest, there is little information higher relative abundance of bacterial OTUs, with distinct rare on the root–microbiota communities and their associated taxa. Low-abundance taxa have been recognized to provide functions in most hyperaccumulators (Thijs et al., 2017). critical functions in ecosystems and represent a vast functional Additionally, attention has been focused on hyperaccumulators gene pool (Jousset et al., 2017; Compant et al., 2019; Xiong from metalliferous soils, and those inhabiting non-polluted et al., 2020). Therefore, the identification of satellite groups soils remain practically unexplored. Our study aimed to identify associated with the root niche of a hyperaccumulator plant the bacterial and fungal taxa and their functional attributes growing on natural and undisturbed soil is highly interesting. associated with N. brachypetala plants from non-metalliferous For example, Nitrospiraceae and Gemmatimodaceae (which soils. Microbia identification of field samples was performed are more abundant in the rhizoplane of N. brachypetala) Frontiers in Microbiology | www.frontiersin.org 12 May 2021 | Volume 12 | Article 639997 Martos et al. Unravelling the Root-Community of N. brachypetala have been previously associated with the rhizosphere of the than in the hyperaccumulator. However, it cannot be Cd/Zn hyperaccumulator A. halleri (Muehe et al., 2015). The concluded that N. brachypetala was globally deficient in metal(loid)-tolerant phylum Chlamydiae (Schneider et al., fungi, as the relative abundance for all plant species was high. 2017) also increased in the rhizoplane of our hyperaccumulator. Hyperaccumulator plants harbored seven exclusive fungal taxa. Additionally, the rhizosphere of N. brachypetala includes Nonetheless, certain specific fungal biota were absent in the Thermogemmatisporaceae, which was formerly detected in high hyperaccumulator. If we assume that the distribution of the concentrations in the reclaimed soil of an ancient mine (Pershina root-associated microorganisms is not a hazardous process, we et al., 2020), and Moraxellaceae, which includes heavy metal can speculate a potential positive selection in the rhizoplane resistant species (Chen et al., 2015). of N. brachypetala toward certain bacteria with a detrimental Ultimately, there is a substantial variation in the consortia effect on specific fungal groups. Among the exclusive fungi of microbes in the taxonomically related plants that cohabit in the hyperaccumulator, Xerocomus could positively affect the undisturbed soils, but differ in the hyperaccumulator the functioning of N. brachypetala as it is a symbiotroph phenotype. Considering that the cited microbial taxa are annotated as ectomycorrhizal species. Ectomycorrhizal fungi much less abundant in the soil fraction associated with non- can help the hyperaccumulator phenotype through chemical hyperaccumulator plants, our results suggest a selection of metal- immobilization in the cell wall or complexation with organic tolerant bacteria exist in the root fractions of N. brachypetala. exudations (Luo et al., 2014). The Brassicaceae family has a The metal tolerance of microbia colonizing the rhizosphere of high percentage of non-mycorrhizal species (Wang and Qiu, hyperaccumulators in common soils was not the focus of this 2006), but with exceptions such as N. praecox (a Zn, Cd, research. Therefore, it is an interesting topic for future studies on and Pb hyperaccumulator), which, under field conditions, microbial cultivation. is colonized by mycorrhiza during the developmental stages Our results from the DAA identified Candidatus Solibacter that are coincident with low glucosinolate levels (Vogel- (Acidobacteria) as one of the most prevalent bacteria in the Mikus et al., 2005; Pongrac et al., 2008). Our DA analysis rhizoplane of N. brachypetala growing on non-metalliferous soil. revealed that Xerocomus subtomentosus was 100 times more C. Solibacter has been identified in the core of the bacterial abundant in the rhizoplane of N. brachypetala. However, community in soils polluted with heavy metals (Golebiewski the confirmation of functional root colonization requires et al., 2014; Wang et al., 2018). Kaistobacter (which is further analysis. To the best of our knowledge, any of these associated with hyperaccumulators) has been similarly described increased fungal taxa have been previously related to metal- in contaminated soils (Wu et al., 2019). Another abundant polluted soils or hyperaccumulator species. Only A. rutilum genus was Streptomyces, which has been previously affiliated and O. chlamydosporicum have been described as tolerant to with the plant Sedum alfredii in Cd-polluted soil (Hou et al., abiotic stresses such as salinity and acidic pH, respectively (Fritze 2018). We also identified Flavobacterium with a high DAA and Baath, 1993; Georgieva et al., 2012). These taxa deserve ratio, which has been associated with metal accumulation in further investigation for their potential contributions to the different hyperaccumulators (Idris et al., 2004; Luo et al., 2011, hyperaccumulation phenotype. 2017; Zhang et al., 2012). Any of the differentially abundant Our results demonstrated the root compartment (rhizosphere bacteria of N. brachypetala may be assigned as metal solubilizers vs. rhizoplane) and plant host (N. brachypetala vs. non- because the applied software (FAPROTAX) did not include this hyperaccumulators) were the major factors influencing the functional category. microbiome diversity, structure, and composition. Previous A distinctive functional trait of the bacterial communities in studies have revealed that microbial communities are strongly N. brachypetala is their role in the nitrogen cycle. This function influenced by plant roots increasing toward the root (Bulgarelli is not a direct mechanism of metal acquisition or tolerance. et al., 2012; Lundberg et al., 2012; Martinez-Diz et al., 2019; However, increased nitrogen availability can be advantageous for Thiergart et al., 2020). In contrast, the host plant is suggested to plants growing in metal-polluted soils. Nitrate assimilation is be a weaker predictor of root–microbiota composition (Schlaeppi reduced under moderate metal exposure, confirming that metal et al., 2014; Glynou et al., 2018; Macia-Vicente et al., 2020). toxicity alters nitrogen uptake by plants (Wang et al., 2008). However, host influence has mainly been evaluated among Interestingly, co-cropping of legumes with hyperaccumulators genotypes of a specific plant species. Consequently, studies has been reported to improve soil quality and fertility and addressing the role of host identity in harboring different increase metal yield (Saad et al., 2016, 2021). Our results microbial species are required. Recent field research indicates provide a potential specialized rhizosphere bacterial community that host identity in comparing different plant species is a that facilitates N recycling. The plants in our study inhabited strong variable in the determination of microbial assembly non-polluted sites; therefore, bacterial selection should occur (Burns et al., 2015; Borymski et al., 2018; Li et al., 2018; regardless of the soil metal content. To confirm this hypothesis, Wang et al., 2020). it will be necessary to study the microbial community of Previous studies have indicated that the composition of root- N. brachypetala under metal exposure. inhabiting microbiome communities is highly influenced by soil Our results identified bacteria contrasts with the DAA type (Bulgarelli et al., 2012; Lundberg et al., 2012; Schlaeppi ratios for mycobiota between hyperaccumulator and non- et al., 2014) and geographical location (Glynou et al., 2018; hyperaccumulator species. Fungal species were generally Lin et al., 2020). Nonetheless, both factors affect the root- much more enriched in the non-hyperaccumulator species associated microbes in a domain-dependent manner, with a Frontiers in Microbiology | www.frontiersin.org 13 May 2021 | Volume 12 | Article 639997 Martos et al. Unravelling the Root-Community of N. brachypetala stronger effect of soil properties on bacteria and the geographic DATA AVAILABILITY STATEMENT location of fungi (Thiergart et al., 2020). In our research, the study location was not a significant factor in the microbial All sequence data are freely available in the European Nucleotide community variations. Further large-scale research with more Archive database (Project PRJEB41937) and will be available populations from distant locations should be conducted to upon publication. confirm this result. The Pyrenees are the southern limit of N. brachypetala, where the species is rare and in regression (Martos et al., 2016); future studies should consider researching AUTHOR CONTRIBUTIONS Northern populations. Root–microbial communities play a pivotal role in plant CP conceived the study and revised the manuscript. SM designed performance by improving mineral nutrition (Jacoby et al., 2017). the study and wrote the manuscript. SB, ML, and CC performed Bacteria and fungi have been reported to contribute to the use the field sampling work. LP-M and SB analyzed the data. All of major nutrients such as N, P, and S (Hong et al., 2015; authors approved the final manuscript. Almario et al., 2017; Perez-Izquierdo et al., 2019; Zhang et al., 2019). Microbes have also been involved in the cycling of some micronutrients, mainly Fe, Zn, and Mn (Alegria Terrazas et al., FUNDING 2016). In our study, some soil elements (Ca and K) made a minor but significant contribution to the microbial assembly. This study was funded by the Spanish Government grant To the best of our knowledge, the role of these elements in references BFU2016-75176-R and PID2019-104000-R. the composition and structure of root microbes has yet to be investigated. In conclusion, our results comparing non-hyperaccumulator ACKNOWLEDGMENTS and hyperaccumulator species indicate that the root We would like to thank Rosa Padilla for processing the ionome compartment and the plant host are key factors influencing data and Carol Diaz-Gutierrez for her help with microbiota DNA the diversity, structure, and composition of root-associated extraction. We would also like to thank Raquel Egea for the bacteria. The fungal community was mainly determined by the first analysis of the bacterial dataset. We are grateful to Llorenç root niche, but not by the host. Conversely, location was an Saez and Iñigo Granzow-de la Cerda for their botanical advice irrelevant factor in any microbial group variation. We observed during plant sampling. a global reduction in microbial α diversity in the rhizoplane. N. brachypetala was a notable exception, presenting a prominent diversity of bacteria forming a complex interaction network in the closest-root compartment. N. brachypetala harbored a large SUPPLEMENTARY MATERIAL group of differentially abundant bacteria, whereas fungi were The Supplementary Material for this article can be found more dominantly associated with the non-hyperaccumulator online at: https://www.frontiersin.org/articles/10.3389/fmicb. species. Even though both the non-hyperaccumulator and 2021.639997/full#supplementary-material the hyperaccumulator inhabited non-metalliferous soils, we detected enriched levels in the N. brachypetala root Supplementary Figure 1 | Rarefaction curves. compartments of bacteria previously described as metal-tolerant Supplementary Figure 2 | Prevalence for bacteria and fungi. or formerly associated with hyperaccumulators in metal-rich soils. In contrast, the hyperaccumulator was impoverished Supplementary Figure 3 | Relative abundance per sample. in certain fungi. Interestingly, we found that most of the Supplementary Dataset 1 | Sample information and sequencing quality. Stats differentially abundant bacteria in the hyperaccumulator for RP1 and RP2. were involved in the nitrogen cycle. Our results suggest a Supplementary Dataset 2 | Soil and plant leaf ions composition and potential selection of beneficial taxa in the soil fractions statistics. influenced by the roots of N. brachypetala that can favor Supplementary Dataset 3 | List of bacterial and fungal taxa according to N availability to the plants, indirectly contributing to their the differential abundance analysis (DAA) within each root hyperaccumulator phenotype. compartment. REFERENCES Alegria Terrazas, R., Giles, C., Paterson, E., Robertson-Albertyn, S., Cesco, S., and Mimmo, T. (2016). Plant-Microbiota interactions as a driver of the mineral Aboudrar, W., Schwartz, C., Morel, J. L., and Boularbah, A. (2013). Effect of nickel- turnover in the Rhizosphere. Adv. Appl. Microbiol. 95, 1–67 doi: 10.1016/bs. resistant rhizosphere bacteria on the uptake of nickel by the hyperaccumulator aambs.2016.03.001 Noccaea caerulescens under controlled conditions. J. Soils Sediments 13, 501– Alford, E. M., Pilon-Smits, E. A. H., and Paschke, M. W. (2010). Metallophytes— 507. doi: 10.1007/s11368-012-0614-x a view from the rhizosphere. Plant Soil 337, 33–50 doi: 10.1007/s11104-010- Aksoy, A., Leblebici, Z., and Prasad, M. N. V. (2015). Metal-accumulating plants 0482-3 from serpentine habitats of Kizildag, Konya Province, Turkey. Austral. J. Bot. Almario, J., Jeena, G., Wunder, J., Langen, G., Zuccaro, A., and Coupland, G. 63, 372–378 doi: 10.1071/bt14354 (2017). Root-associated fungal microbiota of nonmycorrhizal Arabis alpine and Frontiers in Microbiology | www.frontiersin.org 14 May 2021 | Volume 12 | Article 639997 Martos et al. Unravelling the Root-Community of N. brachypetala its contribution to plant phosphorus nutrition. Proc. Natl. Acad. Sci. U.S.A. 114, Golebiewski, M., Deja-Sikora, E., Cichosz, M., Tretyn, A., and Wrobel, B. E9403–E9412 (2014). 16S rDNA pyrosequencing analysis of bacterial community in heavy Assunçao, A. G. L., Schat, H., and Aarts, M. G. M. (2003). Thlaspi caerulescens, an metals polluted soils. Microb. Ecol. 67, 635–647. doi: 10.1007/s00248-013- attractive model species to study heavy metal hyperaccumulation in plants. New 0344-7 Phytol. 159, 351–360 doi: 10.1046/j.1469-8137.2003.00820.x Glynou, K., Thines, M., and Macia-Vicente, J. G. (2018). Host species identity in Barillot, C. D. C., Sarde, C. O., Bert, V., Tarnaud, E., and Cochet, N. (2013). annual Brassicaceae has a limited effect on the assembly of root-endophytic A standardized method for the sampling of rhizosphere and rhizoplan soil fungal communities. Plant Ecol. Divers. 11, 569–580. doi: 10.1080/17550874. bacteria associated to a herbaceous root system. Ann. Microbiol. 63, 471–476. 2018.1504332 doi: 10.1007/s13213-012-0491-y Hong, Y., Liao, D., Hu, A., Wang, H., Chen, J., and Khan, S. (2015). Diversity Bech, J., Corrales, I., Tume, P., Barcelo, J., Duran, P., Roca, N., et al. (2012). of endophytic and rhizoplane bacterial communities associated with exotic Accumulation of antimony and other potentially toxic elements in plants Spartina alterniflora and native mangrove using Illumina amplicon sequencing. around a former antimony mine located in the Ribes Valley (Eastern Pyrenees). Can. J. Microbiol. 61, 723–733. doi: 10.1139/cjm-2015-0079 J. Geochem. Explor. 113, 100–105 doi: 10.1016/j.gexplo.2011.06.006 Hou, D. D., Lin, Z., Wang, R. Z., Ge, J., Wei, S., Xie, R. H., et al. (2018). Cadmium Besnard, G., Basic, N., Christin, P. A., Savova-Bianchi, D., and Galland, N. exposure-sedum alfredii planting interactions shape the bacterial community in (2009). Thlaspi caerulescens (Brassicaceae) population genetics in western the hyperaccumulator plant rhizosphere. Appl. Environ. Microbiol. 84:e02797- Switzerland: is the genetic structure affected by natural variation of soil heavy 17. doi: 10.1128/AEM.02797-17 metal concentrations? New Phytol. 181, 974–984 doi: 10.1111/j.1469-8137. Idris, R., Trifonova, R., Puschenreiter, M., Wenzel, W. W., and Sessitsch, A. 2008.02706.x (2004). Bacterial communities associated with flowering plants of the ni Bokulich, N. A., Joseph, C. M. L., Allen, G. R., Benson, A., and Mills, D. A. hyperaccumulator Thlaspi goesingense. Appl. Environ. Microbiol. 70, 2667– (2012). Next-generation sequencing reveals significant bacterial diversity 2677 doi: 10.1128/aem.70.5.2667-2677.2004 of botrytized wine. PLoS One 7:e36357 doi: 10.1371/journal.pone.003 Jacoby, R., Peukert, M., Succurro, A., Koprivova, A., and Kopriva, S. (2017). The 6357 Role of soil microorganisms in plant mineral nutrition—current knowledge and Borymski, S., Cycon, M., Beckmann, M., Mur, L. A. J., and Piotrowska-Seget, future directions. Front. Plant Sci. 8:1617. Z. (2018). Plant species and heavy metals affect biodiversity of microbial Jousset, A., Bienhold, C., Chatzinotas, A., Gallien, L., Gobet, A., and Kurm, V. communities associated with metal-tolerant plants in metalliferous soils. Front. (2017). Where less may be more: how the rare biosphere pulls ecosystems Microbiol. 9:1425 strings. ISME J. 11, 853–862. doi: 10.1038/ismej.2016.174 Bulgarelli, D., Rott, M., Schlaeppi, K., Ver Loren van Themaatm, E., and Kramer, U. (2010). Metal hyperaccumulation in plants. Annu. Rev. Plant Biol. 61, Ahmadinejad, N. (2012). Revealing structure and assembly cues for Arabidopsis 517–534. doi: 10.1146/annurev-arplant-042809-112156 root inhabiting bacterial microbiota. Nature 488, 91–95 doi: 10.1038/ Li, Y., Wu, X., Chen, T., Wang, W., Liu, G., and Zhang, W. (2018). Plant phenotypic nature11336 traits eventually shape its microbiota: a common garden test. Front. Microbiol. Bulgarelli, D., Schlaeppi, K., Spaepen, S., Ver Loren van Themaat, E., and Schulze- 9:2479 Lefert, P. (2013). Structure and functions of the bacterial microbiota of plants. Likar, M., Pongrac, P., Vogel-Mikus, K., and Regvar, M. (2010). Molecular diversity Annu. Rev. Plant Biol. 64, 807–838 doi: 10.1146/annurev-arplant-050312- and metal accumulation of different Thlaspi praecox populations from Slovenia. 120106 Plant Soil 330, 195–205. doi: 10.1007/s11104-009-0192-x Burns, J. H., Anacker, B. L., Strauss, S. Y., and Burke, D. J. (2015). Soil microbial Lin, H., and Peddada, S. D. (2020). Analysis of microbial compositions: a review of community variation correlates most strongly with plant species identity, normalization and differential abundance analysis. npj Biofilms Microbiomes? followed by soil chemistry, spatial location and plant genus. AoB Plants 6:60. 7:lv030 Lin, M., Xiong, H., Xiang, X., Zhou, Z., Liang, L., and Mei, Z. (2020). The effect Callahan, B. J., McMurdie, P. J., Rosen, M. J., Han, A. W., Johnson, A. J. A., and of plant geographical location and developmental stage on root-associated Holmes, S. P. (2016). DADA2: high-resolution sample inference from Illumina microbiomes of Gymnadenia conopsea. Front. Microbiol. 11:1257 amplicon data. Nat. Methods 13, 581–583. doi: 10.1038/nmeth.3869 Louca, S., Parfrey, L. W., and Doebeli, M. (2016). Decoupling function and Caporaso, J. G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman, F. D., taxonomy in the global ocean microbiome. Science 353, 1272–1277 doi: 10. and Costello, E. K., et al. (2010). QIIME allows analysis of high-throughput 1126/science.aaf4507 community sequencing data. Nat. Methods 7, 335–336 Lundberg, D. S., Lebeis, S. L., Paredes, S. H., Yourstone, S., Gehring, J., and Malfatti, Cappa, J. J., and Pilon-Smits, E. A. H. (2014). Evolutionary aspects of S. (2012). Defining the core Arabidopsis thaliana root microbiome. Nature 488, elemental hyperaccumulation. Planta 239, 267–275 doi: 10.1007/s00425-013- 86–90. doi: 10.1038/nature11237 1983-0 Luo, J., Tao, Q., Wu, K., Li, J., Qian, J., and Liang, Y. (2017). Structural Chen, B., Ma, X., Liu, G., Xu, X., Pan, F., and Zhang, J. (2015). An endophytic and functional variability in root-associated bacterial microbiomes of Cd/Zn bacterium Acinetobacter calcoaceticus Sasm3-enhanced phytoremediation of hyperaccumulator Sedum alfredii. Appl. Microbiol. Biotechnol. 101, 7961–7976. nitrate–cadmium compound polluted soil by intercropping Sedum alfredii with doi: 10.1007/s00253-017-8469-0 oilseed rape. Environ. Sci. Pollut. Res. 22, 17625–17635 doi: 10.1007/s11356- Luo, S., Chen, L., Chen, J., Xiao, X., Xu, T., and Wan, Y. (2011). Analysis 015-4933-5 and characterization of cultivable heavy metal-resistant bacterial endophytes Compant, S., Samad, A., Faist, H., and Sessitsch, A. (2019). A review on the plant isolated from Cdhyperaccumulator Solanum nigrum L. and their potential microbiome: ecology, functions, and emerging trends in microbial application. use for phytoremediation. Chemosphere 85, 1130–1138. doi: 10.1016/j. J. Adv. Res. 19, 29–37. doi: 10.1016/j.jare.2019.03.004 chemosphere.2011.07.053 Edwards, J., Johnson, C., Santos-Medellín, C., Lurie, E., Podishetty, N. K., and Luo, Z. B., Wu, C., Zhang, C., Li, H., Lipka, U., and Polle, A. (2014). The role of Bhatnagar, S. (2015). Structure, variation, and assembly of the root-associated ectomycorrhizas in heavy metal stress tolerance of host plants. Environ. Exp. microbiomes of rice. Proc. Natl. Acad. Sci. USA 112, E911–E920. Bot. 108, 47–62, doi: 10.1016/j.envexpbot.2013.10.018 Farinati, S., DalCorso, G., Panigati, M., and Furini, A. (2011). Interaction between Ma, Y., Prasad, M. N. V., Rajkumar, M., and Freitas, H. (2011). Plant growth selected bacterial strains and Arabidopsis halleri modulates shoot proteome and promoting rhizobacteria and endophytes accelerate phytoremediation of cadmium and zinc accumulation. J. Exp. Bot. 62, 3433–3447. doi: 10.1093/jxb/ metalliferous soils. Biotechnol. Adv. 29, 248–258. doi: 10.1016/j.biotechadv. err015 2010.12.001 Fritze, H., and Baath, E. (1993). Microfungal species composition and fungal McMurdie, P. J., and Holmes, S. (2012). Phyloseq: a bioconductor biomass in a coniferous forest soil polluted by alkaline deposition. Microb. Ecol. package for handling and analysis of high-throughput 25, 83–92. phylogenetic sequence data. Pac. Symp. Biocomput. 2012, Georgieva, M. L., Lebedeva, M. P., and Bilanenko, E. N. (2012). Mycelial fungi in 235–246 saline soils of the western Transbaikal region. Eurasian Soil Sci. 45, 1159–1168. Macia-Vicente, J. G., Nam, B., and Thines, M. (2020). Root filtering, rather than doi: 10.1134/s1064229312120058 host identity or age, determines the composition of root-associated fungi Frontiers in Microbiology | www.frontiersin.org 15 May 2021 | Volume 12 | Article 639997 Martos et al. Unravelling the Root-Community of N. brachypetala and oomycetes in three naturally co-occurring Brassicaceae. Soil Biol. Bioch. Schlaeppi, K., Dombrowski, N., Garrido-Oter, R., van Themaat, E. V. L., 146:107806. doi: 10.1016/j.soilbio.2020.107806 and Schulze-Lefert, P. (2014). Quantitative divergence of the bacterial root Martinez-Diz, M. P., Andres-Sodupe, M., Bujanda, R., Diaz-Losada, E., Eichmeier, microbiota in Arabidopsis thaliana relatives. PNAS 111, 585–592 doi: 10.1073/ A., and Gramaje, D. (2019). Soil-plant compartments affect fungal microbiome pnas.1321597111 diversity and composition in grapevine. Fungal Ecol. 41, 234–244 doi: 10.1016/ Schneider, A. R., Gommeaux, M., Duclercq, J., Fanin, N., Conreux, A., and j.funeco.2019.07.003 Alahmad, A. (2017). Response of bacterial communities to Pb smelter pollution Martos, S., Gallego, B., Saez, L., Lopez-Alvarado, J., Cabot, C., and Poschenrieder, in contrasting soils. Sci. Total Environ. 605, 436–444 doi: 10.1016/j.scitotenv. C. (2016). Characterization of Zinc and Cadmium Hyperaccumulation in three 2017.06.159 Noccaea (Brassicaceae) populations from non-metalliferous sites in the Eastern Sessitsch, A., Kuffner, M., Kidd, P., Vangronsveld, J., Wenzel, W. W., and Pyrenees. Front. Plant Sci. 7:128 Fallman, K. (2013). The role of plant-associate bacteria in the mobilization and Misljenovic, T., Jovanovic, S., Mihailovic, N., Gajic, B., Tomovic, G., and phytoextraction of trace elements in contaminated soils. Soil Biol. Biochem. 60, Baker, A. J. M. (2020). Natural variation of nickel, zinc and cadmium 182–194 doi: 10.1016/j.soilbio.2013.01.012 (hyper)accumulation in facultative serpentinophytes Noccaea kovatsii and N. Stringlis, I. A., Yu, K., Feussner, K., de Jonge, R., Van Bentum, S., and Van Verk, praecox. Plant Soil 447, 475–495 doi: 10.1007/s11104-019-04402-5 M. C. (2018). MYB72-dependent coumarin exudation shapes root microbiome Muehe, E. M., Weigold, P., Adaktylou, I. J., Planer-Friedrich, B., Kraemer, U., assembly to promote plant health. Proc. Natl. Acad. Sci. USA 115, E5213–E5222 and Kappler, A. (2015). Rhizosphere microbial community composition affects Thiergart, T., Duran, P., Ellis, T., Vannier, N., Garrido-Oter, R., and Kemen, E. cadmium and zinc uptake of the metal-hyperaccumulating plant Arabidopsis (2020). Root microbiota assembly and adaptive differentiation among European halleri. Appl. Environ. Microbiol. 81, 2173–2181 doi: 10.1128/aem.03359-14 Arabidopsis populations. Nat. Ecol. Evol. 4, 122–131 doi: 10.1038/s41559-019- Nguyen, N. H., Song, Z., Bates, S. T., Branco, S., Tedersoo, L., and Menke, J. (2016). 1063-3 FUNGuild: an open annotation tool for parsing fungal community datasets by Thijs, S., Langill, T., and Vangronsveld, J. (2017). The bacterial and fungal ecological guild. Fungal Ecol. 20, 241–248. doi: 10.1016/j.funeco.2015.06.006 microbiota of hyperaccumulator plants: small organisms, large influence. Adv. Nilsson, R. H., Larsson, K. H., Taylor, A. F. S., Bengtsson-Palme, J., Jeppesen, T. S., Bot. Res. 83, 43–86 doi: 10.1016/bs.abr.2016.12.003 and Schigel, D. (2018). The UNITE database for molecular identification of Van der Ent, A., Baker, A. J. M., Reeves, R. D., Pollard, A. J., and Schat, fungi: handling dark taxa and parallel taxonomic classifications. Nucleic Acids H. (2013). Hyperaccumulators of metal and metalloid trace elements: Res. 47, D259–D264 facts and fiction. Plant Soil 362, 319–334 doi: 10.1007/s11104-012- Nishida, S., Tanikawa, R., Ishida, S., Yoshida, J., Mizuno, T., and Nakanishi, 1287-3 H. (2020). Elevated expression of vacuolar nickel transporter gene IREG2 is Visioli, G., D’Egidio, S., Vamerali, T., Mattarozzi, M., and Sanangelantoni, associated with reduced root-to-shoot nickel translocation in Noccaea japonica. A. M. (2014). Culturable endophytic bacteria enhance Ni translocation in the Front. Plant Sci. 11:610 hyperaccumulator Noccaea caerulescens. Chemosphere 117, 538–544 doi: 10. Oksanen, J., Blanchet, F. G., Kindt, R., Legendre, P., Minchin, P. R., and O’hara, R. 1016/j.chemosphere.2014.09.014 (2013). Package ‘vegan’. Community Ecology Package, Version Visioli, G., D’Egidio, S., and Sanangelantoni, A. M. (2015a). The bacterial Paulson, J. N., Pop, M., and Bravo, H. C. (2013). MetagenomeSeq: statistical rhizobiome of hyperaccumulators: future perspectives based on omics analysis analysis for sparse high-throughput sequencing. Bioconductor Package 1:191. and advanced microscopy. Front. Plant Sci. 5:752 Perez-Izquierdo, L., Zabal-Aguirre, M., Gonzalez-Martinez, S. C., Buee, M., Verdu, Visioli, G., Vamerali, T., Mattarozzi, M., Dramis, L., and Sanangelantoni, A. M. M., and Rincon, A. (2019). Plant intraspecific variation modulates nutrient (2015b). Combined endophytic inoculants enhance nickel phytoextraction cycling through its below-ground rhizospheric microbiome. J. Ecol. 107, 1594– from serpentine soil in the hyperaccumulator Noccaea caerulescens. Front. Plant 1605 doi: 10.1111/1365-2745.13202 Sci. 6:638 Pershina, E., Ivanova, E., Kimeklis, A., Zverev, A., Kichko, A., and Aksenova, T. Vogel-Mikus, K., Drobne, D., and Regvar, M. (2005). Zn, Cd and Pb accumulation (2020). Soil microbiomes of reclaimed and abandoned mines of the Yamal and arbuscular mycorrhizal colonisation of pennycress Thlaspi praecox Wulf. region. Polish Polar Res. 41, 95–114 (Brassicaceae) from the vicinity of a lead mine and smelter in Slovenia. Environ. Pongrac, P., Vogel-Mikus, K., Regvar, M., Tolra, R., Poschenrieder, C., and Barcelo, Pollut. 133, 233–242. doi: 10.1016/j.envpol.2004.06.021 J. (2008). Glucosinolate profiles change during the life cycle and mycorrhizal Vogel-Mikus, K., Pongrac, P., Kump, P., Necemer, M., and Regvar, M. (2006). colonization in a Cd/Zn hyperaccumulator Thlaspi praecox (Brassicaceae). Colonisation of a Zn, Cd and Pb hyperaccumulator Thlaspi praecox Wulfen J. Chem. Ecol. 34, 1038–1044 doi: 10.1007/s10886-008-9502-7 with indigenous arbuscular mycorrhizal fungal mixture induces changes in Pongrac, P., Vogel-Mikus, K., Kump, P., Necemer, M., Tolra, R., and heavy metal and nutrient uptake. Environ. Pollut. 139, 362–371 doi: 10.1016/j. Poschenrieder, C. (2007). Changes in elemental uptake and arbuscular envpol.2005.05.005 mycorrhizal colonisation during the life cycle of Thlaspi praecox Wulfen. Wang, H., Zeng, Y., Guo, C., Bao, Y., Lu, G., and Reinfelder, J. R. (2018). Chemosphere 69, 1602–1609 doi: 10.1016/j.chemosphere.2007.05.046 Bacterial, archaeal, and fungal community responses to acid mine drainage- Poschenrieder, C., Tolra, R., and Barcelo, J. (2006). Can metals defend plants laden pollution in a rice paddy soil ecosystem. Sci. Total Environ. 616–617, against biotic stress? Trends Plant Sci. 11, 288–295 doi: 10.1016/j.tplants.2006. 107–116 doi: 10.1016/j.scitotenv.2017.10.224 04.007 Wang, L., Zhou, Q. X., Ding, L. L., and Sun, Y. B. (2008). Effect of cadmium Rajkumar, M., Sandhya, S., Prasad, M. N. V., and Freitas, H. (2012). Perspectives toxicity on nitrogen metabolism in leaves of Solanum nigrum L. as a newly of plant-associated microbes in heavy metal phytoremediation. Biotechnol. Adv. found cadmium hyperaccumulator. J. Hazard. Mater. 154, 818–825. doi: 10. 30, 1562–1574. doi: 10.1016/j.biotechadv.2012.04.011 1016/j.jhazmat.2007.10.097 Rideout, J. R., He, Y., Navas-Molina, J. A., Walters, W. A., Ursell, L. K., Wang, Y. L., Gao, C., Chen, L., Ji, N. N., Wu, B. W., and Lu, and Gibbons, S. M. (2014). Subsampled open-reference clustering creates P. P. (2020). Community assembly of endophytic fungi in consistent, comprehensive OTU definitions and scales to billions of sequences. ectomycorrhizae of Betulaceae plants at a regional scale. Front. Microb. PeerJ 2:e545 doi: 10.7717/peerj.545 10:3105 Saad, R., Echevarria, G., Rodriguez-Garrido, B., Kidd, P., and Benizri, E. (2021). Wang, B., and Qiu, Y. L. (2006). Phylogenetic distribution and evolution of ∗ A two-year field study of nickel-agromining using Odontarrhena chalcidica mycorrhizas in land plants. Mycorrhiza 16, 299–363 doi: 10.1007/s00572-005- co-cropped with a legume on an ultramafic soil: temporal variation in plant 0033-6 biomass, nickel yields and taxonomic and bacterial functional diversity. Plant Weiss, S., Xu, Z. Z., Peddada, S., Amir, A., Bittinger, K., and Gonzalez, A. (2017). Soil. doi: 10.1007/s11104-021-04834-y Normalization and microbial differential abundance strategies depend upon Saad, R., Kobaissi, A., Robin, C., Echevarria, G., and Benizri, E. (2016). Nitrogen data characteristics. Microbiome 5:27 fixation and growth of Lens culinaris as affected by nickel availability: a pre- Whiting, S. N., Souza, M. P., and Terry, N. (2001). Rhizosphere bacteria mobilize requisite for optimization of agromining. Environ. Exp. Bot. 131, 1–9 doi: Zn for Hyperaccumulation by Thlaspi caerulescens. Environ. Sci. Technol. 35, 10.1016/j.envexpbot.2016.06.010 3144–3150 doi: 10.1021/es001938v Frontiers in Microbiology | www.frontiersin.org 16 May 2021 | Volume 12 | Article 639997 Martos et al. Unravelling the Root-Community of N. brachypetala Wu, Z., Gao, G., and Wang, Y. (2019). Effects of soil properties, heavy metals, and Conflict of Interest: The authors declare that the research PBDEs on microbial community of e-waste contaminated soil. Ecotox. Environ. was conducted in the absence of any commercial or financial Saf. 180, 705–714. doi: 10.1016/j.ecoenv.2019.05.027 relationships that could be construed as a potential conflict of Xiong, C., He, J. Z., Singh, B. K., Zhu, Y. G., Wang, J. T., and Li, P. P. (2020). interest. Rare taxa maintain the stability of crop mycobiomes and ecosystem functions. Environ Microbiol. 23, 1907–1924 doi: 10.1111/1462-2920.15262 Copyright © 2021 Martos, Busoms, Pérez-Martín, Llugany, Cabot and Zhang, J., Liu, Y., Zhang, N., Hu, B., Jin, T., and Xu, H. (2019). NRT1.1B is Poschenrieder. This is an open-access article distributed under the terms of associated with root microbiota composition and nitrogen use in field-grown the Creative Commons Attribution License (CC BY). The use, distribution rice. Nat. Biotech. 37, 676–684 doi: 10.1038/s41587-019-0104-4 or reproduction in other forums is permitted, provided the original author(s) Zhang, X., Lin, L., Chen, M., Zhu, Z., Yang, W., and Chen, B. (2012). A non- and the copyright owner(s) are credited and that the original publication pathogenic Fusarium oxysporum strain enhances phytoextraction of heavy in this journal is cited, in accordance with accepted academic practice. No metals by the hyperaccumulator Sedum alfredii Hance. J. Hazard. Mater. use, distribution or reproduction is permitted which does not comply with 229-230, 361–370 doi: 10.1016/j.jhazmat.2012.06.013 these terms. Frontiers in Microbiology | www.frontiersin.org 17 May 2021 | Volume 12 | Article 639997