Non-Specific Interactions of Rhizospheric Microbial Communities Support the Establishment of Mimosa acutistipula var. ferrea in an Amazon Rehabilitating Mineland

Mimosa acutistipula var. ferrea (Fabaceae) is endemic to ferruginous tropical rocky outcrops in the eastern Amazon, also known as canga. Canga are often associated with mining activities and are the target of protection and rehabilitation projects. M. acutistipula stands out in this biodiversity hotspot with high growth rates, even in rehabilitating minelands (RMs). However, little is known about the diversity of soil microorganisms interacting with M. acutistipula in canga and RMs. This study analyzed the rhizosphere-associated bacterial and fungal microbial communities associated with M. acutistipula growing in an RM and a native shrub canga. The fungal phylum Ascomycota was the dominant taxa identified in the rhizosphere of the canga (RA: 98.1) and RM (RA: 93.1). The bacterial phyla Proteobacteria (RA: 54.3) and Acidobacteria (RA: 56.2) were the dominant taxa identified in the rhizosphere in the canga and RM, respectively. Beneficial genera such as Bradyrhizobium, Rhodoplanes, and Paraconiothyrium were identified in the rhizosphere of M. acutistipula in both areas. However, the analyses showed that the fungal and bacterial diversity differed between the rhizosphere of the canga and RM, and that the microbial taxa adapted to the canga (i.e., Rasamsonia, Scytalidium, Roseiarcus, and Rhodomicrobium) were lacking in the RM. This influences the microbe-mediated soil processes, affecting long-term rehabilitation success. The results showed that M. acutistipula established non-specific interactions with soil microorganisms, including beneficial taxa such as nitrogen-fixing bacteria, mycorrhizal fungi, and other beneficial endophytes, well known for their importance in plant adaptation and survival. High levels of microbe association and a plant’s ability to recruit a wide range of soil microorganisms help to explain M. acutistipula’s success in rehabilitating minelands.


Introduction
Brazil is one of the largest iron producers and has the second-largest reservoirs in the world [1]. Serra dos Carajás in the state of Pará covers Brazil's main iron ore reservoirs [2]. Here, iron ore extraction occurs in rocky outcrops [3], where a typical and highly diverse vegetation conforms to an ecosystem known as canga [4]. Iron mining directly affects canga, altering biological diversity and ecological services in this ecosystem [5]. Therefore, the screening of native plants from canga with good yield in post-mining areas is mandatory to reduce biodiversity loss and enhance mineland rehabilitation.
Iron mining in Brazil essentially occurs in open-cast mines. These sites are characterized by the removal of the soil surface layer, soil physical, chemical and biological alterations, soil organic matter depletion, nutrient loss, erosion, and soil acidification [6,7]. In this context, the rehabilitation of minelands is necessary to generate a positive diversity offset [8,9]. The revegetation of post-mining areas using native species is the most recommended alternative to reducing biodiversity loss and progressively recovering the plant coverage in biodiversity hotspots such as the Amazon [10,11]. Here, the interaction of native plants with soil-borne microorganisms is essential to contribute to mineland rehabilitation. Specifically, beneficial rhizosphere-inhabiting microorganisms contribute to abiotic/biotic stress tolerance by direct (i.e., production of plant growth-promoting metabolites, phytohormones, and nutrient solubilization) and/or indirect mechanisms (i.e., protection against phytopathogens, immobilization of metals, and changes in gene expression) [12,13]. Therefore, to identify the microbial diversity associated with plants growing in rehabilitated minelands (RMs), evaluating the underground effects of rehabilitation projects in mining areas in canga is necessary.
The soils from canga are composed of ferruginous fragments in which plant establishment and growth are affected by environmental factors such as the low availability of essential nutrients, low water holding capability, and high soil surface temperatures [14,15]. Mimosa acutistipula var. ferrea Barneby (Fabaceae) inhabits environments characterized by a herbaceous shrub physiognomy and can grow after being sown in waste piles in RMs [16]. Symbiosis with soil-borne microorganisms has been described as essential for plants from the Fabaceae family to fit to challenging environments such as RMs [14,17]. Therefore, it is expected that symbiotic microorganisms, as well as other beneficial rhizosphere-associated microorganisms, influence the establishment of native plants in RMs. In the RMs of the eastern Amazon, few studies have analyzed the interaction of native plants with soil microorganisms. Hence, determining the diversity of soil-borne microorganisms associated with the rhizosphere of plants growing in RMs is necessary for identifying changes in the microbial diversity that underline the loss of essential microbe-mediated process that can influence long-term rehabilitation [18]. The aims of this study were: (i) To describe and compare the rhizosphere-associated fungal and bacterial communities of M. acutistipula growing in its native environment (canga) and RMs, and (ii) to identify beneficial soil-borne microorganisms that contribute to plant establishment in RMs.

Sampling
Soil samples were collected in a native shrub canga (hereafter, "canga"; 6 • 00 41.0" S 50 • 17 45.0" W) and in waste piles of an iron mineland in the canga, with a revegetation program started in 2014 (hereafter, "RM"; 6 • 02 32.0" S 50 • 07 04.0" W), in Serra dos Carajás, Pará, Brazil. Canga soils are characterized by the high adsorption of phosphorous by iron and aluminum oxides and the low availability of soluble forms in the soil solution [19,20]. The revegetation in the RM started in 2014 with commercial and native plants, where fertilization with NPK 04-14-08, mulch, organic compost, and AG60 fixer was used (Table S1). Compound samples from rhizospheric and bulk soils were collected from four M. acutistipula plants growing in the canga and RM. The rhizospheric soil was collected by gently shaking the roots, whereas the bulk soil was collected 5 m away from each sampled plant, at a depth of 10 cm and with no presence of other plants. The samples were immediately stored in 50 mL Falcon tubes in ice, transported to the laboratory, and kept at −80 • C until further processing.

DNA Extraction and Sequencing
Total DNA was extracted from 250 mg of soil using a PowerSoil DNA Isolation Kit (QIAGEN, Hilden, Germany), according to the manufacturer's recommendations. The Processes 2021, 9,2079 3 of 14 DNA concentration was determined with a Qubit fluorometer (Thermo Fisher Scientific, Waltham, MA, USA) and the quality was verified in a 1% electrophoresis agarose gel.
The amplicon libraries for the bacteria were prepared according to the Illumina 16S Metagenomic Sequencing Library Preparation Protocol (Illumina, San Diego, CA, USA). For the fungi, the amplicon libraries were prepared with the respective modifications (fungal primers and PCR conditions). For the bacteria, the V3-V4 region of the 16S rRNA gene was amplified using the primer set S-D-Bact-0341-b-S-17-N (5 -TCGTCGGCAGCGTCA GATGTGTATAAGAGACAGCCTACGGGNGGCWGCAG-3 ) and S-D-Bact-0785-a-A-21-N (5 -GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGACTACHVGGGTATCTAATCC-3 ), whereas for the fungi, the ITS region of the 18S rRNA gene was amplified using the primer set fITS7i (5 -TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGGTGARTCATCGAAT CTTTG-3 ) and ITS4i (5 -GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGTCC TCCGCTTATTGATATGC-3 ). The PCR mix contained 12.5 µL of 2x Kapa Hifi HotStart Ready Mix (Sigma-Aldrich, St Louis, MI, USA), 5 µL of each primer (1 µM), and 2.5 µL of DNA. The PCR cycle for bacteria consisted of an initial denaturing of 3 min at 95 • C, followed by 25 cycles of denaturation at 95 • C for 30 s, annealing at 55 • C for 30 s, extension at 72 • C for 30 s, and a final extension at 72 • C for 5 min. The PCR cycle for the fungi consisted of an initial denaturation at 94 • C for 2 min, followed by 35 cycles of denaturation at 94 • C for 30 s, annealing at 56 • C for 1 min, extension at 72 • C for 30 s, and a final extension at 72 • C for 7 min. The concentrations of the PCR fragments were measured in a Qubit fluorometer using a Qubit™ ds DNA HS Assay (Thermo Fisher Scientific, Waltham, MA, USA). The size and quality of the PCR fragments were estimated on an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA) using a DNA 1000 chip. The PCR reactions were purified with an AMPure XP purification kit (Beckman Coulter, Brea, CA, USA), and the libraries were further processed with a Nextera XT kit (Illumina). The libraries were standardized to a concentration of 4 nM and processed following Illumina 16S Metagenomic Sequencing Library Preparation (Illumina). The 16S rRNA and 18S rRNA gene libraries were sequenced in a Miseq-Illumina platform using a MiSeq V3 reagent kit (600 cycles; Illumina) in the human and medical genetics laboratory at Universidade Federal do Pará (Belém, PA, Brazil).

Sequence Analyses
The ITS and 16S sequences from the rhizospheric and bulk soils were analyzed using the PIMBA pipeline (Pipeline for MetaBarcoding Analysis) [21]. The sequences were trimmed and filtered by quality using PRINSEQ v0.20.4, and forward and reverse sequences were merged using PEAR v0.9.19 [22]. Reads were replicated, singletons removed, and the sequences were truncated to 200 for fungi and 240 for bacteria. Chimeras were filtered, and the sequences were grouped into operative taxonomic units (OTUs) using VSEARCH v2.8.2 at a similarity of ≥97%. Taxonomic assignment was performed using the UNITE database and the Ribosomal Database Project for fungi and bacteria [23,24]. The sequences obtained in this study were submitted to the NCBI Sequence Read Archive (http://trace.ncbi.nlm.nih.gov/Traces/sra/, accessed on 13 February 2021) under accession number PRJNA701717.

Functional Analysis of Rhizosphere-Associated Microbial Communities
The variations in the inferred functions of bacterial and fungal communities associated with M. acutistipula were estimated according to Liang et al. [25], with minor modifications. Briefly, the predicted ecological roles of the fungal OTUs were assigned using FUNGuild v1.1 [26], whereas the predicted functions of the bacterial OTUs were assigned using FAPROTAX v1.2.4 [27], both in Python v3.8.2. The data were plotted in R software v3.6.3 (R Core Team 2018; https://www.R-project.org) using the viridis, dplyr, and scales packages.

Statistical Analyses
Specific data analyses and graphs were performed in R software. Graphs for each sample were constructed considering the alpha and beta diversity using the ggplot2 and vegan packages. Principal coordinate analysis (PCoA) was constructed using "weighted UniFrac distances" in the phyloseq package [28]. Statistical significance of the community data was estimated using permutational multivariate analysis of variance (PERMANOVA) via the function "Adonis" (vegan package). Alpha diversity was estimated as the Shannon and Simpson diversity indices using the vegan package. The shared and unique fungal and bacterial OTUs among different samples were calculated and visualized with Venn diagrams using the MicEco package.
Similarly, heatmaps were constructed with the relative abundance (RA) of beneficial OTUs using the heatmaply and phyloseq packages. In addition, a redundancy RDA analysis was constructed to reveal the relationships between beneficial microbiota and soil properties using the vegan, mvpart, and labdsv packages. Clustering analysis of the data was performed using the "hclust" function in the pvclust package. The linear discriminant analysis (LDA) effect size (LEfSe) was performed with the Kruskal-Wallis test [29] and the effect size was estimated with a logarithmic score of >3.0.

Results
The ITS2 sequencing produced a total of 3,158,043 raw reads across 16 input libraries. After quality filtering, 2,324,549 amplicon sequences were considered. The number of fungal OTUs in the samples ranged from 197 to 737 and was, on average, 496 (Table S2). The soils from RMs showed more fungal sequences (1,202,317; comprising 557 ± 189 OTUs) than canga (1,122,232; comprising 434 ± 115 OTUs) (Table S2). Additionally, the rhizospheric soil samples showed more sequences (1,318,674; comprising 556 ± 116 OTUs) than the bulk soil samples (1,005,875; comprising 435 ± 188 OTUs) ( Table S2). The Shannon diversity index showed higher values in the samples from RMs, whereas the Simpson diversity index was similar in all samples (Figure 1a,b). Venn diagrams of the ITS sequences showed 88 core OTUs in the four samples, and soils from canga-B, RM-B, RM-R, and canga-R had 133, 203, 128, and 221 unique OTUs, respectively ( Figure S1).
In canga, the analysis of the fungal sequences showed Ascomycota as the dominant taxa ( Figure 2), whereas Acidobacteria and Proteobacteria were the dominant bacteria ( Figure 3). Glomeromycota was also detected in bulk soils, with an RA higher in bulk soil samples (Table S4). At the genus level, different fungal taxa were more abundant in the rhizospheric (i.e., Mycosphaerella, Scytalidium, and Talaromyces) ( Figure S2) or bulk soils (i.e., Periconia, Curvularia, and Alternaria) ( Figure S2 and Table S4), but most sequences were not assigned to known taxa ( Figure S2). Regarding the bacterial genera, Bradyrhizobium and Rhodoplanes were the most abundant in the rhizospheric and bulk soil samples ( Figure S3 and Table S4).     In the RM, analysis of the fungal sequences identified Ascomycota as the dominant taxa ( Figure 2), whereas Proteobacteria and Acidobacteria were the dominant bacterial taxa (Figure 3). At the genus level, most sequences were not assigned to known taxa ( Figure S2), followed by Fusarium and Coniosporium in the rhizospheric and bulk soils ( Figure S2 and Table S4). Regarding the bacterial sequences, different genera were dominant in the rhizospheric (i.e., Sphingomicrobium and Pseudomonas) and bulk soils (i.e., Serratia and Hyphomicrobium) ( Figure S3 and Table S4).
In the RM, analysis of the fungal sequences identified Ascomycota as the dominant taxa ( Figure 2), whereas Proteobacteria and Acidobacteria were the dominant bacterial taxa (Figure 3). At the genus level, most sequences were not assigned to known taxa (Figure S2), followed by Fusarium and Coniosporium in the rhizospheric and bulk soils ( Figure  S2 and Table S4). Regarding the bacterial sequences, different genera were dominant in the rhizospheric (i.e., Sphingomicrobium and Pseudomonas) and bulk soils (i.e., Serratia and Hyphomicrobium) ( Figure S3 and Table S4). PCoA and cluster analyses of the microbial community structure among the different samples showed that bacterial (PERMANOVA: p = 0.003) and fungal (PERMANOVA: p = 0.001) communities were distinct between both sampling environments, but similar when comparing the bulk and rhizospheric soils at each sampling environment (Figures 4 and  S4). PCoA and cluster analyses of the microbial community structure among the different samples showed that bacterial (PERMANOVA: p = 0.003) and fungal (PERMANOVA: p = 0.001) communities were distinct between both sampling environments, but similar when comparing the bulk and rhizospheric soils at each sampling environment (Figure 4 and Figure S4).
Differences in the microbial communities' composition were estimated by calculating the LEfSe scores at the family and genus levels and were focused in the rhizosphere. A total of 24 distinct fungal taxa were identified in the rhizosphere of M. acutistipula established in a canga or RM. Most of these taxa were related to Ascomycota and Basidiomycota ( Figure 5). A total of 26 preferential bacterial taxa were identified in M. acutistipula growing in a canga or RM, most of which were related to Proteobacteria and Actinobacteria ( Figure 5).  Differences in the microbial communities' composition were estimated by calculating the LEfSe scores at the family and genus levels and were focused in the rhizosphere. A total of 24 distinct fungal taxa were identified in the rhizosphere of M. acutistipula established in a canga or RM. Most of these taxa were related to Ascomycota and Basidiomycota ( Figure 5). A total of 26 preferential bacterial taxa were identified in M. acutistipula growing in a canga or RM, most of which were related to Proteobacteria and Actinobacteria ( Figure 5).   A total of eight trophic modes were predicted in the fungal OTUs, being the categories saprotroph (45.4%) and pathotroph_saprotroph (31.2%) most abundant in the soils from the RM, whereas pathotroph_symbiotroph (24.6%) was more abundant in the soils from the canga (Figure 6a). The category saprotroph_symbiotroph (2.6%) was exclusively identified in the soils from the canga, whereas symbiotroph was identified in the soils from both the canga (1.8%) and RM (1.5%) (Figure 6a). Similarly, 34 predicted functions were identified in the bacterial OTUs. Among them, chemoheterotrophy (7.79% and 10.2%), aerobic_chemoheterotrophy (7.65% and 10.1%), and nitrogen fixation (4.81% and 3.82%) were the most abundant predicted functions in the canga and RM, respectively. In the bulk soils, nitrate_ammonification (0.92% and 0.85%), nitrite_ammonification (0.82% and 0.88%), and chitinolysis (0.84% and 0.87%) were present in the canga and RM, respectively (Figure 6b). The categories dark_oxidation_of_sulfur_compounds (2.78%) and aro-matic_compound_degradation (0.57%) were associated with the soils from the RM, whereas anoxygenic_photoautotrophy_S_oxydizing (1.71%) and dark_iron_oxidation (0.76%) were exclusive to the soils from the canga (Figure 6b). from the RM, whereas pathotroph_symbiotroph (24.6%) was more abundant in the soils from the canga (Figure 6a). The category saprotroph_symbiotroph (2.6%) was exclusively identified in the soils from the canga, whereas symbiotroph was identified in the soils from both the canga (1.8%) and RM (1.5%) (Figure 6a). Similarly, 34 predicted functions were identified in the bacterial OTUs. Among them, chemoheterotrophy (7.79% and 10.2%), aerobic_chemoheterotrophy (7.65% and 10.1%), and nitrogen fixation (4.81% and 3.82%) were the most abundant predicted functions in the canga and RM, respectively. In the bulk soils, nitrate_ammonification (0.92% and 0.85%), nitrite_ammonification (0.82% and 0.88%), and chitinolysis (0.84% and 0.87%) were present in the canga and RM, respectively (Figure 6b). The categories dark_oxidation_of_sulfur_compounds (2.78%) and aro-matic_compound_degradation (0.57%) were associated with the soils from the RM, whereas anoxygenic_photoautotrophy_S_oxydizing (1.71%) and dark_iron_oxidation (0.76%) were exclusive to the soils from the canga (Figure 6b). Heatmaps showing the RA of potential beneficial taxa showed that they were distinct among the canga and RM for both the fungal and bacterial sequences (except for Bradyrhizobium) (Figure 7). The RDA analysis showed that soil characteristics influence the abundance of beneficial genera without a clear tendency among variables ( Figure S5). Heatmaps showing the RA of potential beneficial taxa showed that they were distinct among the canga and RM for both the fungal and bacterial sequences (except for Bradyrhizobium) (Figure 7). The RDA analysis showed that soil characteristics influence the abundance of beneficial genera without a clear tendency among variables ( Figure S5).

Discussion
Rehabilitation of degraded areas affected by mining has become an urgent issue, es-

Discussion
Rehabilitation of degraded areas affected by mining has become an urgent issue, especially in biodiversity hotspots such as the Amazon [30]. Few studies have analyzed the effects of mineland rehabilitation in the interaction between soil microbial populations and plants [31,32]. Therefore, this study described and characterized the rhizosphere-associated fungal and bacterial communities of M. acutistipula growing in a native canga ecosystem and RM, providing novel evidence about the microbial communities interacting with the plant during mineland rehabilitation.
Ascomycota and Proteobacteria were the most abundant fungal and bacterial phyla detected in both sampling ecosystems, namely, a canga and RM (Figures 2 and 3), which agrees with recent studies analyzing the microbial communities associated with plants growing in stressful ecosystems such as post-mining areas [33,34]. Several taxa belonging to such phyla were identified (i.e., Scytalidium, Paraconiothyrium, Talaromyces, Rhizobium, Rhodomicrobium, and Mesorhizobium) interacting with M. acutistipula in both soil substrates ( Figures S2 and S3). However, specific beneficial taxa can also interact with M. acutistipula in a canga (i.e., Rhodomicrobium and Roseiarcus) or RM (i.e., Sphingomonas and Mesorhizobium) (Figure 7 and Table S4). Such microbial taxa are known for their role in essential microbe-mediated processes such as nutrient mineralization, biological control, and stress tolerance [35][36][37]-mechanisms that are key to sustaining growth in RMs.
The obtained diversity indices in the fungi and bacteria were similar between ecosystems ( Figure 1). Nonetheless, the composition of the fungal and bacterial communities was different between the canga and RM (Figure 4). Such changes in microbial diversity in disturbed ecosystems underline a shift in the diversity of dominant microbial communities, as well as the loss of microbial taxa involved in specific processes such as nutrient cycling, biocontrol, and promotion of stress tolerance [38,39]. Despite the changes in the fungal and bacterial taxa (Figures 4 and 5), beneficial microorganisms were detected in the rhizosphere of M. acutistipula growing in the canga and RM (Figure 7). This difference in the composition of microbial communities is influenced by the low presence of plants colonizing RMs, which decreases the selectivity of microbial taxa and competition for resources [40], thus promoting the changes in microbial communities detected in RMs (Figures 2 and 7). As a canga is a severe ecosystem with high insolation rates, shallow and oxidic soils, intense dry and wet seasons, and low water holding capacity [41], the microbial diversity associated with native plants is also submitted to these stressful environmental conditions. Here, the presence of specific beneficial microorganisms in the rhizosphere of M. acutistipula can be related to beneficial functions to fit with these challenging environmental conditions, which is in line with recent studies analyzing the role of rhizospheric soil communities in the establishment of plants growing in metalliferous soils [42,43].
Microbial interactions in the plant rhizosphere are essential for growth in severe environments such as a canga. In the rhizosphere of M. acutistipula growing in both sampling points, bacterial taxa commonly associated with tolerance to metal stress were also identified, including members of the phyla Proteobacteria and Actinobacteria (i.e., Sphingomonas and Actinotalea) (Table S3) [44,45]. One of the main roles of these soil microorganisms is to contribute to the alleviation of metal stress in plants by the production of natural chelators, which decreases the bioavailability of metals [46], such as those detected in both soil substrates (Table S1). Additionally, the presence of nodule-forming bacteria in the rhizosphere of plants adapted to harmful environments is considered fundamental for symbiotic nitrogen fixation and plant growth in metal-polluted areas [47]. Thus, the preferential taxa involved in N mineralization in the rhizosphere of M. acutistipula (i.e., Rhizobium, Mesorhizobium, Bradyrhizobium, Rhodomicrobium, and Roseiarcus) indicate a strong affinity with these microorganisms (Figure 7), which can contribute to the establishment of plants in RMs. These results are in line with those of Oliveira et al. [48], who detected several nodule-forming bacteria in soil samples from RMs after iron mining in the Quadrilátero Ferrífero region, southeast Brazil. Therefore, such bacterial taxa are vital for the adaptation of this species in canga ecosystems, as well as in RMs. Further studies must be performed in order to identify the specificity of the beneficial taxa associated with vital structures such as nodules or healthy roots.
The study results showed several fungal taxa with beneficial roles for plants growing under severe environments, including dark septate endophytes (DSEs), AMFs, and freeliving fungi. Fungal taxa belonging to DSEs (i.e., Myceliophthora, Thielavia, Coniosporium, and Exophiala) were detected with more abundance in soils associated with the RM ( Figure 5 and Table S4). Fungi belonging to DSEs have been reported as common rhizosphereassociated fungi that promote abiotic stress tolerance in plants [42,49] in order to grow in stressful environments such as RMs. Similarly, the fungal phylum Glomeromycota was detected in the rhizosphere of plants growing in the RM (Table S4). This fungal phylum identified in the canga and RM includes AMFs, which can expand the root system, causing a greater part of the soil to be exploited when searching for resources [50]. As a consequence, plants improve their P uptake and are more resistant to abiotic stress [51]. In RMs, the presence of these microorganisms is fundamental due to the low level of organic matter in the soil after mining (Table S1). This is particularly important in oxidic soils such as those in a canga, where the high adsorption rates of P by iron and aluminum oxides limit their bioavailability (Table S1) [20]. Moreover, beneficial fungi belonging to the phyla Ascomycota and Basidiomycota were identified in the rhizosphere of the canga and/or RM (i.e., Paraconiothyrium, Scytalidium, and Talaromyces) (Figure 7), which include genera with known roles in the promotion of stress tolerance, enhancement of nutrient uptake, growth promotion, and biocontrol capabilities [52,53]. Therefore, rhizosphere-inhabiting fungi associated with M. acutistipula play crucial roles in the establishment at initial stages of mineland rehabilitation, but their benefits for M. acutistipula must be confirmed in further analyses.
As demonstrated in this study, M. acutistipula can associate without specificity with microbial taxa inhabiting the soil substrate (Figures 2 and 3). Plants with this behavior play critical roles in the rehabilitation of mining areas, participating in essential ecosystem services such as nutrient cycling, carbon storage, and symbiosis that can contribute to ecosystem rehabilitation [54]. However, other plant species need specific interaction with soil microbes for growth, being this dependence a bottleneck in the life cycle, affecting the long-term rehabilitation of an altered ecosystem [55,56]. Such specific interactions with soil microorganisms further limits plant establishment in the rehabilitated ecosystem, promoting invasive/exotic species colonization or affecting plant/microbe-related ecosystem services [57,58]. Therefore, a comprehensive management of soil substrates and their associated microbial communities can be further addressed in order to establish other native canga species in RMs.
M. acutistipula has been classified as a promising native plant for mineland rehabilitation in the eastern Amazon by their high drought stress tolerance and capability of hyperaccumulation of metals [59]. In fact, this study showed that M. acutistipula interacts with beneficial fungi and bacteria, independent of the sampling environment ( Figure 7). However, despite beneficial soil microbes interacting with the M. acutistipula rhizosphere in RMs, the differences in rhizosphere-associated microorganisms with canga ecosystems hide key underground microbial-related processes that can alter mineland rehabilitation ( Figure 6). The results agree with recent studies analyzing the performance of M. acutistipula in degraded minelands [54,59] and provide information about specific microbial taxa interacting with the M. acutistipula rhizosphere. An additional strategy to enhance mineland rehabilitation is the search for plants evolved in a severe canga ecosystem that have established non-specific interactions with soil microbes and plant them in future RMs in order to protect the ecosystem services associated with native canga ecosystems.

Conclusions
This study showed that non-specific interactions with rhizosphere microorganisms characterize the growth of M. acutistipula in a native canga ecosystem and RM in the eastern Amazon. Here, interaction with beneficial taxa independent of the soil substrate is one of the mechanisms explaining why M. acutistipula is suitable for use in rehabilitation programs. Therefore, a plant's facility to recruit beneficial microorganisms from the soil substrate, especially those able to establish mycorrhizal or nitrogen fixation symbiosis, is key to selecting plants for further revegetation programs. Moreover, the presence of key microbial groups in soil substrates is crucial for the rehabilitation program.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/pr9112079/s1, Figure S1: Venn diagram of soil fungal (A) and bacterial OTUs (B) in rhizospheric (R) or bulk (B) soil substrate samples of Mimosa acutistipula growing in a canga ecosystem (canga) or a rehabilitating mineland (RM) in Serra dos Carajás, eastern Amazon; Figure S2: Relative abundance at the genus level of major fungal 18S rRNA sequences obtained from Mimosa acutistipula growing in a canga ecosystem (canga) or a rehabilitating mineland (RM) in Serra dos Carajás, eastern Amazon; Figure S3: Relative abundance at the genus level of bacterial 16S rRNA sequences obtained from Mimosa acutistipula growing in a canga ecosystem (canga) or a rehabilitating mineland (RM) in Serra dos Carajás, eastern Amazon; Figure S4: Clustering of fungal 18S rRNA (a) and bacterial 16S rRNA (b) sequences in rhizospheric (R) or bulk (B) soil samples of Mimosa acutistipula (M) growing in a canga ecosystem (canga) or a rehabilitating mineland (RM) in Serra dos Carajás, eastern Amazon; Figure S5: Redundancy analysis (RDA) showing the influence of soil characteristics on the preferential beneficial taxa associated with Mimosa acutistipula growing in a canga (canga) (a) or a rehabilitating mineland (RM) (b) in Serra dos Carajás, eastern Amazon; Table S1: Physical and chemical characteristics of soil substrates associated with Mimosa acutistipula growing in a canga or rehabilitating mineland (RM); Table S2: Fungal 18S rRNA sequences obtained in rhizospheric and bulk soil samples from Mimosa acutistipula growing in a canga (canga) or a rehabilitating mineland (RM); Table S3: Bacterial 16S rRNA sequences obtained in rhizospheric and bulk soil samples from Mimosa acutistipula growing in a canga (canga) or rehabilitating mineland (RM); Table S4: Most abundant fungal and bacterial taxa identified in soils associated with Mimosa acutistipula growing in a canga or rehabilitating mineland (RM) in Serra dos Carajás, eastern Amazon.

Data Availability Statement:
The sequences obtained in this study were deposited in the NCBI Sequence Read Archive (https://www.ncbi.nlm.nih.gov/sra/PRJNA701717, accessed on 13 February 2021) under the accession number PRJNA701717.