Diversity of Arbuscular Mycorrhizal Fungi Associated with a Sb Accumulator Plant, Ramie (Boehmeria nivea), in an Active Sb Mining

Abstract

Arbuscular mycorrhizal fungi (AMF) have great potential for assisting heavy metal hyperaccumulators in the remediation of contaminated soils. However, little information is available about the symbiosis of AMF associated with an antimony (Sb) accumulator plant under natural conditions. Therefore, the objective of this study was to investigate the colonization and molecular diversity of AMF associated with the Sb accumulator ramie (Boehmeria nivea) growing in Sb-contaminated soils. Four Sb mine spoils and one adjacent reference area were selected from Xikuangshan in southern China. PCR-DGGE was used to analyze the AMF community composition in ramie roots. Morphological identification was also used to analyze the species in the rhizosphere soil of ramie. Results obtained showed that mycorrhizal symbiosis was established successfully even in the most heavily polluted sites. From the unpolluted site Ref to the highest polluted site T4, the spore numbers and AMF diversity increased at first and then decreased. Colonization increased consistently with the increasing Sb concentrations in the soil. A total of 14 species were identified by morphological analysis. From the total number of species, 4 (29%) belonged to Glomus, 2 (14%) belonged to Acaulospora, 2 (14%) belonged to Funneliformis, 1 (7%) belonged to Claroideoglomus, 1 (7%) belonged to Gigaspora, 1 (7%) belonged to Paraglomus, 1 (7%) belonging to Rhizophagus, 1 (7%) belonging to Sclervocystis, and 1 (7%) belonged to Scutellospora. Some AMF sequences were present even in the most polluted site. Morphological identification and phylogenetic analysis both revealed that most species were affiliated with Glomus, suggesting that Glomus was the dominant genus in this AMF community. This study demonstrated that ramie associated with AMF may have great potential for remediation of Sb-contaminated soils.

Introduction

Antimony (Sb) is the ninth-most mined metal worldwide, and its compounds are listed as priority pollutants by the US Environment Protection Agency [63] and the European Union [14]. Because Sb occurs in the environment mostly as a co-contaminant of more toxic elements such as Pb or As, its biogeochemistry and ecotoxicity have been generally overlooked in research, especially about contaminanted soils [62]. However, soil pollution with Sb caused by mining and manufacturing has become increasingly serious because of the wide industrial use of this metalloid [23,69]. An estimated 12-16 million tons was consumed worldwide; China is the highest producer of Sb, accounting for approximately 84.0% of the world’s share [24, 26]. Sb is hazardous to human health and some Sb compounds are even considered to be potentially carcinogenic [19, 21]. Ecologically, Sb stress may harm plants by affecting their development, biomass, and quality [13, 44]. Owing to this toxicity and current patterns of pollution, there is a pressing need for research about proper environmental restoration practices for use on Sb-contaminated soils.

Traditional treatments for soil remediation such as physical and chemical approaches are often problematic because they are either labor intensive, costly, or environmentally damaging [45]. In contrast, phytoremediation — the use of plants, and hyper-accumulators in particular, to extract pollutants from soils — has been regarded as an effective, non-intrusive, inexpensive, and more socially acceptable approach [5, 7].

A plant species with great potential in this regard is Boehmeria nivea. Known commonly as ramie or ‘‘China grass,” this perennial has been widely cultivated in China for over 5,000 years. Ramie belongs to Urticaceae whose fiber is widely used in light and heavy textile industries. Ramie has been documented as a dominant plant in many different metal mining sites, and it can accumulate a large amount of Sb, Cd, and Hg [33, 43, 68]. It also has a high biomass, reproduces three times per year, and is highly adaptable to unfavorable growth conditions such as drought, unfertile soils, disease, and insect pests [56]. Owing to these advantages, ramie has huge potential as a fiber crop for phytoremediation of metal-polluted soils, mitigating the danger of toxic metals entering into food chains [72].

Arbuscular mycorrhizal fungi (AMF) are important regulators of plant performance in soils contaminated with heavy metal, and increase the resistance of plants to heavy metal toxicity [11, 17]. In addition, it has been demonstrated that AMF can increase the heavy metal translocation factor, biomass, and trace element concentrations of plants [34, 61, 70]. The combined use of plants with AMF has advantages over the use of hyperaccumulators alone, and has been proposed as one of the most promising green remediation techniques [2, 30, 67]. Even so, it is important to note that the effects of AMF colonization on phytoremediation depend on the combination of AMF species, host plants, and heavy metal types [16]. Typically in natural systems, AMF, host plants, and heavy metals would have reached a state of equilibrium over a long period [74]. A solid understanding of colonization and AMF diversity under natural heavy metal stress is indispensable for appropriate mycorrhizo-remediation in heavy-metal-polluted regions.

It stands to reason then that ramie combined with AMF would be a good choice of environmental restoration method for Sb-contaminated soils. However, little information is available about the symbioses and community composition of AMF associated with ramie growing on Sb-contaminated soils under natural conditions. This knowledge gap is an obstacle for the biotechnological application of ramie growth with AMF on Sb-contaminated soils. Our objectives were (i) to determine whether an Sb accumulator plant was able to develop symbiotic associations with arbuscular mycorrhizal fungi at Sb-polluted sites; (ii) if present, to evaluate mycorrhizal colonization and AMF molecular diversity in the roots of ramie and their relationship with Sb contamination; and (iii) to further explore potential indigenous AMF strains for phytoremediation. Our results will lay the foundation for the use of ramie with AMF in remediation of Sb-contaminated soils.

 

Materials and Methods

Sites and Sampling

The study was conducted at Xikuangshan Sb mine (27.7°N, 111.4°E), near Lengshuijiang City, Hunan Province, in the southern part of China. This is the world’s largest Sb mine, to the point that it is well-known as the “World Capital of Sb.”

Sampling was carried out in an area where mining and smelting operations began in 1897. Based on the distribution of slagheaps, four Sb mine spoils were selected: two slagheaps (sites T1 and T2), an ore charge heap (site T3), and a tailing dam with smelting waste and wastewater (site T4). An area of 30 km from the active mining was chosen as an uncontaminated reference site (site Ref). Soil and plant samples were collected from these five areas in September 2013. Three plots of 10 × 10 m were randomly established at each site. Each plot was then divided into four 5 × 5 m sampling subplots, in which the roots, shoots, and corresponding rhizosphere soils from ramie were collected. Samples from the four subplots were pooled and homogenized to form a composite sample.

Measurement of Soil Physicochemical Properties

Soil and plant samples were immediately transported to the laboratory for further analysis. Shoots were washed carefully with deionized water and dried at 60℃ for 48 h, and then were cut into small pieces for analysis of metal contents. Portions of soils were air-dried, ground in a ceramic mortar, and then sieved through a 2 mm mesh for analysis of metal contents. Total acid digestion of soil and plant samples was carried out in a closed microwave digestion device (MarsXpress, CEM, USA) using concentrated HNO3 (65-68%), HCl (36-38%), and HF (>40%) in a ratio 6:3:1 with 0.2 g of sample and 6 ml of acid (190℃, 15 min). After cooling to room temperature, the samples were filtered and diluted to 50 ml with ultrapure water. Total Sb in soils or shoots were digested by this method. Water-soluble Sb concentrations of soil was carried out as previously described [43]. First, 1.6 g of soil sample (air dried) was weighed into a 50 ml tube and mixed with ultrapure water in a liquid solid (L/S) ratio of 10:1 using a rotary shaker (24 h, 100 rpm, 25℃). Next, the samples were centrifuged (4,000 rpm, 15 min) and filtered at 0.45 µm. Total or waterextracted Sb concentrations in soil or shoots were determined by Inductively Coupled Plasma-Mass Spectrometry (ICP-MS, 7500c; Agilent, USA). Organic matter content was determined using K2Cr2O7–H2SO4 [35]. Total and extractable phosphorus (P) were determined using H2SO4–HClO4 digestion and HCl–H2SO4 extraction, respectively [35]. Total and extractable nitrogen (N) were determined using H2SO4–HClO4 digestion and KCl extraction, respectively [35].The biota to soil accumulation factor (BSAF) was calculated as the ratio of the total Sb content in plant shoots (mg/kg) relative to total soil concentration (mg/kg) [43].

Morphological Identification of AMF and Root Colonization

Spores were extracted from soil by wet sieving [20] and sucrose centrifugation and mounted with PVLG and PVLG + Melzer’s reagent for morphological studies and taxonomic identification using species descriptions from the AMF identification manual by [49] and INVAM homepage (http://www.invam.caf.wvu.edu).

The intensity of root colonization by AMF was determined following a previous method [37]. For each sample, after roots were washed from the soil, a 0.5 g fresh weight subsample of randomly selected fine root segments was placed in histocassettes and stored in 70% ethanol for at least 24 h. Roots were placed in 10% KOH and cleared by double autoclaving, rinsed with water, acidified in 5% HCl for 5 min, and then placed in trypan blue stain (40% glycerol, 20% lactic acid, and 0.01% trypan blue) overnight. Roots were then destained (40% glycerol and 20% lactic acid) for at least 24 h and mounted on slides. Using a compound microscope, slides were scanned, and when vertical crossbars intersected a root, mycorrhizal presence or absence was recorded until 100 hits were recorded. There were three replicates for each subsample.

DNA Extraction from Roots

Approximately 20 mg of lyophilized root samples was used for DNA extraction. Root samples were cut into 2 mm fragments and homogenized in liquid nitrogen. Total DNA was extracted using a DNA Extraction Kit following the manufacturer’s protocol (Axygen Biosciences, China). DNA was finally dissolved in 100 ml of elution buffer.

PCR Amplification of 18S rRNA Genes and Denaturing Gradient Gel Electrophoresis Analysis

Nested PCR was used to increase the resolution yield of denaturing gradient gel electrophoresis (DGGE). The first round of amplification was performed using the universal 18S rDNA primers GeoA2 and Geo11, which produced a 1,800 bp product [54]. PCR amplification was performed in a total volume of 25 µl, containing 1 µl of template solution, 10 mM Tris-HCl (pH 8.3), 9.5 µl of ddH2O, 10 pmol of each primer (1 µl), and 12.5 µl of 2× Master Mix (Promega, USA). The amplified reaction was performed using an initial denaturation at 94℃ for 4 min, followed by 30 cycles of 94℃ for 1 min, 54℃ for 1 min, and 72℃ for 7 min, with a final extension phase of 10 min at 72℃.

The amplification product from the first PCR round was diluted 10-fold; 1 µl of the dilution was used as the template for the second round of PCR. The second round primers AM1/NS31-GC were used to amplify the 18S rRNA gene of the AMF, which produced a 580 bp product with the same reaction ingredients as the first PCR [22,57]. The amplified reaction was performed using an initial denaturation at 94℃ for 5 min, followed by 30 cycles of 94℃ for 45 sec, 65℃ for 1 min, and 72℃ for 45 sec, with a final extension phase of 7 min at 72℃.

The products of the second round of PCR were diluted as above, and used as templates for the third round of PCR using the primers NS31-GC and Glol [9]. The amplified reaction was performed using an initial denaturation at 94℃ for 2 min, followed by 30 cycles of 94℃ for 45 sec, 55℃ for 1 min, and 72℃ for 45 sec, with a final extension phase of 7 min at 72℃, and then chilling to 4℃. The amplicons from the nested PCR were verified with agarose gel electrophoresis (1.5% (w/v) agarose, 100 V, 60 min) and ethidium bromide staining to determine their size (approximately 230 bp) and yield in the presence of a pBR322 DNA/Alul Marker.

A 20 µl volume of nested PCR products was used for DGGE analysis. The gels were 1.5 mm thick (20 × 20 cm) and contained 8% (w/v) polyacrylamide (37:1 acrylamide/bis-acrylamide) plus 1× Tris-Acetate-EDTA buffer. A linear gradient from 30% to 50% denaturant was used, where 100% denaturing acrylamide was defined as containing 7 M urea and 40% formamide [41]. DGGE analysis was conducted using a D-Gene system (Bio-Rad Laboratories, USA) at a constant temperature of 60℃. Electrophoresis was run for 10 min at 200 V, after which the voltage was lowered to 150 V for an additional 6 h. Gels were stained using AgNO3 staining solution, and gel images were digitally captured using a ChemiDoc EQ system (Bio-Rad Laboratories, USA).

Sequence Analysis

Prominent, notable DGGE bands were excised from the acrylamide gel and the DNA was eluted using a Poly-Gel DNA Extraction Kit (Omega Scientific, USA). Eluted DNA (1 ml) was reamplified with the primer pair NS31/Glol. PCR products were purified using a PCR Cleanup Kit (Axygen Biosciences, USA), and then sequenced at another facility (Shanghai Sangon Biological Engineering Technology and Services Company, China), using NS31 and Glol as sequencing primers.

Sequences were compared with known sequences using the basic local alignment search tool (BLAST; http://www.ncbi.nlm.nih.gov/BLAST/) [3], and the nearest neighbor AMF sequences were aligned with sample 18S rRNA sequences using ClustalX 1.83. A phylogenetic tree was inferred by the neighbor-joining method using Mega 4.0 software [60].

Data Analysis

DGGE profiles of amplified AMF fragments from different samples were analyzed using Bio-Rad Quantity One 4.4.0 software. The Shannon-Wiener index (H) was used to characterize AMF diversity [36], using the following formula: where H is the Shannon-Wiener index, S is the total number of bands in each sample, and Pi is the relative abundance (i.e., proportion) of the total sample represented by the i-th band.

Root colonization and diversity index data were analyzed using one-way ANOVA following a test for homogeneity of variance. Fisher’s least significant difference test at the 5% confidence level was used to compare the mean differences among sampling sites. Two-tailed Pearson correlation analysis and principal components analysis (PCA) were used to assess the relationships among various soil properties, total Sb concentrations in plants, root colonization, and AMF diversity. All analyses were performed using the software SPSS ver.10.0 (SPSS Inc., USA) and CANOCO 4.5 (Microcomputer Power, USA).

 

Results

Soil Physicochemical Properties

Soil physicochemical properties at the five sampling sites are shown in Table 1. Site T4 possessed the highest Sb concentrations (both total and extractable Sb), and site Ref (the control) had the lowest Sb concentrations (both total and extractable Sb). When compared with the average background level of Sb in soils from Hunan Province [38], the amount of Sb in the four Sb mine spoils were greater by a factor of 599 (T1), 1,015 (T2), 5,622 (T3), and 8,035 (T4). The Sb content of soil from site Ref was much less than in the mine spoils and similar to the background value.

Table 1.Abbreviations: TSb/soil, total Sb concentration in soil; ESb/soil, extractable Sb concentration in soil; TSb/shoot, total Sb concentration in shoot; OM, soil organic matter; TP, total P; EP, extractable P; TN, total N; EN, extractable N. Data are means ± SEM.

Sb Concentrations in Plants and the BSAF

The Sb concentration in shoots of ramie and the BSAF are shown in Table 1 and Fig. 1. Much higher Sb concentrations were detected in ramie shoots growing on the four Sb mine sites, compared with the reference site. The highest Sb concentrations in plants were found in site T4, followed by sites T3, T2, and T1. Plants growing on site T4 possessed the highest values of BSAF. Sites T3 and T2 possessed similar values of BSAF, followed by T1. Site Ref had the lowest values of BSAF.

Fig. 1.Biota to soil accumulation factor of ramie sampled from four Sb mine spoils and an adjacent reference area. Columns marked by the same letter indicate nonsignificant differences at an alpha level of 0.05. Data are means ± SE.

Mycorrhizal Colonization and Morphological Identification of AMF

Mycorrhizal symbionts had successfully established even in the most heavily polluted sites (Fig. 2). The highest mycorrhizal colonization (64%) occurred in site T4, followed by sites T3, T2, and T1. Site Ref had the lowest mycorrhizal colonization (31%) (Fig. 3). Correlation analysis (Table 2) showed that mycorrhizal colonization was significantly positively correlated with BSAF, total Sb concentration in both soils and plants, extractable soil Sb concentration (p < 0.01), and total soil nitrogen concentration (p < 0.05).

Fig. 2.Intraradical hyphae and vesicles in roots of ramie.

Fig. 3.Mycorrhizal colonization of ramie sampled from four Sb mine spoils and an adjacent reference area. Columns marked by the same letter indicate nonsignificant differences at an alpha level of 0.05. Data are means ± SE.

Table 2.**Denotes significant correlation at p < 0.01. *Denotes significant correlation at p < 0.05.

A total of 14 species were identified by morphological analysis at the five sampling sites (Table 3). From the total number of species, 4 (29%) belonged to Glomus, 2 (14%) belonged to Acaulospora, 2 (14%) belonged to Funneliformis, 1 (7%) belonged to Claroideoglomus, 1 (7%) belonged to Gigaspora, 1 (7%) belonged to Paraglomus, 1 (7%) belonged to Rhizophagus, 1 (7%) belonged to Sclervocystis, and 1 (7%) belonged to Scutellospora. All the species were found in Sb mine areas except Funneliformis geosporum. Acaulospora scrobiculata, Glomus aggregatum, Sclervocystis clavispora, Glomus glomerulatum, Funneliformis mosseae, Paraglomus occultum, and Scutellospora aurigloba were the most common species and possessed higher spore density than other species. Site T1 had the highest spore density, followed by sites T2, T3, and T4. In general, a higher spore density occurred in the Sb mine area than in the reference site.

Table 3.‘-’ for non-occurrence, spore densities defined the number of spores per 50 g soil.

DGGE Analysis of AMF in the Roots of Ramie

Nested PCR amplifications were successful. In terms of the AMF communities associated with different root samples, overall on the DGGE gel (Fig. 4) the band number, intensity, and composition differed significantly among sampling sites. Some bands such as 1, 3, 5, 7, and 8 were present exclusively in sites T1 and T2. Band 6 was detected only in site T3. Some bands were also detected in the most heavily polluted site, T4, such as bands 2, 4, 9, 10, and 11. Site Ref contained the fewest number of bands. ShannonWiener indices, based on the proportional intensity of bands, showed that community T2 had the highest diversity, followed by T1, T3, and T4; site Ref had the lowest diversity (Fig. 5).

Fig. 4.DGGE pattern of nested–PCR amplified 18S rDNA fragments of AMF from the roots of ramie in four Sb mine spoils and an adjacent reference area. The labeled bands in the gel correspond to the sequenced clones.

Fig. 5.Shannon-Weiner diversity index of AMF communities generated from the electrophoresis band pattern. Columns marked by the same letter indicate nonsignificant differences at an alpha level of 0.05. Data are means ± SE.

PCA showed that the first two components explained 81% of the variance in the AMF community. PC1 and PC2 explained 49% and 32% of the total variance, respectively (Fig. 6). Mycorrhizal colonization was significantly positively correlated with total Sb concentration in both soils and plants.

Fig. 6.Principal components analysis of Shannon-Weiner index (H), AMF colonization (AC), total Sb concentration in plants (P-TSb), and soil physicochemical properties. Abbreviations: S-TSb, total Sb concentration in soil; S-ESb, extractable Sb concentration in soil; OM, soil organic matter; TP, total P; EP, extractable P; TN, total N; EN, extractable N. Only the means in the same column were compared.

Phylogenetic Analysis

Eleven bands of interest were excised from the DGGE gel for cloning and sequencing. Results of nucleotide accession numbers for our sequences, together with sequences from their closest relatives in the GenBank database, are shown in Table 4. A phylogenetic reconstruction of these cloned sequences, together with sequences from their closest relatives in the GenBank database, is shown in Fig. 7. Eight bands were found to be closely related to Glomus, two bands belonged to Rhizophagus, and one band belonged to Paraglomus.

Table 4.GenBank homologies for sequences from recovered DGGE bands.

Fig. 7.Neighbor-joining phylogenetic tree of partial 18S rRNA gene sequences amplified from ramie root samples. Bootstrap supports above 50% are given on the branches (1,000 replicates).

 

Discussion

AMF form a putative interaction with roots of approximately 80% of the terrestrial plants in nearly all ecosystems [6,8]. However, many hyperaccumulator plants are from non-mycorrhizal families initially [32]. Recently, AMF have been shown to successfully colonize the roots of some hyperaccumulator plant species, and thereby increase metal tolerance and the accumulation of these toxins in plant tissues [18]. For example, arbuscular mycorrhizae have been found in several non-Brassicaceae hyperaccumulator species [4, 46, 71] and metallophytes in the Brassicaceae [47, 50, 65]. In field surveys of plants, ramie has been shown to accumulate the highest amounts of Sb [12, 29, 43]. Here, we also detected quite a high concentration in its shoots (Table 1), which was much higher than the tolerable concentration 5–10 mg/kg in plants [27]. This further illustrates that ramie possesses high potential for application in the phytoremediation of Sb-contaminated soil as a potential Sb-hyperaccumulator, according to the popular standard suggested for hyperaccumulators [51]. However, whether ramie is able to develop symbiotic associations with AM fungi, especially in Sb-polluted soil, is unknown. In this study, we found mycorrhizal colonization and AMF species in the roots of ramie even in the most heavily Sb-polluted sites (Figs. 2 and 4). This indicates that ramie is able to develop symbiotic associations with AMF under natural conditions. Numerous researchers have indicated that AMF are able to alter plant physiology, increase plant resistance to heavy metals, and increase metal uptake [40,48]. An important question to enable the use of AMF in restoration or phytoremediation projects is whether AMF communities at contaminated sites are formed by siteadapted species of fungi. To the best of our knowledge, our finding of symbiosis between ramie and AMF is the first report of such a system involving an accumulator plant growing on Sb-contaminated soils under natural conditions.

From the unpolluted site Ref to the highest polluted site T4, the spore numbers and AMF diversity increased at first and then decreased (Table 3 and Fig. 5). This indicates that relatively moderate Sb stress can stimulate sporulation and AMF diversity, whereas overhigh Sb concentration will show inhibition. Niu et al. [42] found the same changed trend of AMF diversity in a Pb/Zn mine. The rate of mycorrhizal colonization increased consistently with increasing Sb concentration (Fig. 3). The same results have also been observed in plant species in which higher metal concentrations were associated with higher root fungal colonization, such as the As-hyperaccumulator Pteris vittata and the Cd/Pb/Zn hyperaccumulator Thlaspi praecox [1, 47, 66]. Hildebrandt et al. [25] thought higher mycorrhization values at higher contaminated sites can be explained by the adaptation of these plants to high soil heavy metal contents.

Correlation analysis and PCA (Table 2 and Fig. 6) showed that Sb concentration in shoots of ramie was significantly positively correlated with mycorrhizal colonization (p < 0.01). More interesting was that the biota to soil accumulation factor increased from site Ref to T4 (Fig. 1) with the increasing mycorrhizal colonization, and that the most important covariate for BSAF was mycorrhizal colonization (R = 0.846 , p < 0.01). Furthermore, the extractable soil Sb concentration was significantly positively correlated with mycorrhizal colonization (p < 0.01). The primary factor affecting the uptake of Sb by plants is the phytoavailability of Sb in soils [12, 15]. Many studies have demonstrated that AMF can activate heavy metal absorption by changing the rhizosphere physical and chemical environment, and further impact heavy metal uptake by plants [52, 58]. These results suggested that ramie associated with AMF may have great potential for remediation of Sb-contaminated soils. It was worth to further research the inter-relationship between AMF and Sb accumulation in plants. AMF have also been found to facilitate the uptake of arsenate (As), which show similar chemical properties with Sb as the As hyperaccumulator fern Pteris vittata L. [30,34].

Although soil microbes can usually enhance the potential of plants to grow in polluted soils, and in the case of hyperaccumulators may augment their heavy metal uptake, the selection of appropriate microbes for the right plant is essential [39]. Indigenous AMF isolated from polluted sites have attained the ability to survive under such conditions, and hence can act more efficiently relative to the other AMF species. Such species have developed mechanisms over time that can make AMF able to resist or tolerate heavy metal stress; they can accordingly develop more efficient symbiosis with their host plant under such conditions, making the addition of AMF advantageous for phytoremediation and restoration of polluted sites [18,28]. Therefore, analysis of the AMF community structure in the roots of ramie growing on Sb-polluted sites is needed in order to identify those indigenous and tolerant AMF strains. We found that some DGGE bands appeared at lower Sb concentrations and then disappeared at higher concentrations, while some bands were present at much higher Sb concentrations. These unique AMF sequence types are intriguing: bands that were present exclusively in sites T1 and T2 (bands 1, 3, 5, 7, and 8) represent species that can endure only relatively light Sb pollution, whereas the sequence types found in sites T3 and T4 (bands 2, 4, 6, 9, 10, and 11) indicate AMF species that are more tolerant of Sb (Fig. 4). Interactions between ramie and these tolerant AMF may be important for the survival and growth of plants in highly Sb-contaminated soils, such that those particular AMF species may have the best potential for improving phytoremediation efficiency in Sb-contaminated soils. Leung et al. [31] found that four common plants growing on five toxic mining sites had developed different strategies for survival with the aid of indigenous AMF.

Morphological identification and phylogenetic analysis both revealed that most AMF species were affiliated with Glomus (Tables 3 and 4), suggesting that Glomus was the dominant genus in the AMF community associated with ramie in this Sb mining region. Interestingly, many Glomus species or phylotypes have been commonly reported to dominate heavy-metal-polluted areas [59, 75]. Similar to our findings, Yang et al. [73] found that all AMF sequences detected from the roots of Elsholtzia splendens, a Cu-tolerant plant growing at a Cu mine, belonged to Glomus. The consistent occurrence of Glomus in heavy-metal-polluted areas is likely due to the fact that Glomus is the largest AMF group [55], and possesses a strong ability to adapt to stressful conditions [71]. Its high sporulation rates may also be important [10]. Another possible reason why Glomus species are dominant may be the ability of the Glomeraceae to colonize via fragments of mycelium or mycorrhizal root pieces, making them more competitive at infecting roots under metal stress [64]. AMF are highly diverse [53], and plants may become more closely associated with those species that benefit their growth under specific environmental conditions [16]. These may be the reason why many spores were extracted from contaminated soils but were not observed in roots. Meanwhile, some AMF species may be possible to associate with other plants instead of ramie as the complex plant roots in the field. Hence, AMF species identification should be based on spores morphology associated with the root of ramie. Our results suggest that Glomus has more affinity with ramie, especially under Sb stress, and that they are therefore better candidate species for the phytoremediation of Sb-contaminated soils.