Isolation, Characterization and Symbiotic Efficiency of Nitrogen-Fixing and Heavy Metal-Tolerant Bacteria from a Coalmine Wasteland

Areas affected by coal mining can be recovered by revegetation with leguminous plants associated with nitrogen-fixing bacteria. This study addressed the isolation and characterization of native nitrogen-fixing bacteria from coalmine wasteland under different vegetation restoration approaches using Macroptilium atropurpureum (DC) Urb and Vicia sativa L. as trap plants. The bacteria were characterized and identified on the basis of 16S rRNA sequences. Additionally, nitrogen-fixing strains were characterized for tolerance to high heavy metal and low pH levels, as well as for their effect on growth, nodulation, and symbiotic efficiency of M. atropurpureum and V. sativa. Soil samples were taken from the rhizosphere of eight areas, between 6 and 20 years under vegetation restoration, in the coal mining area of Candiota, RS-Brazil. The following properties were evaluated: colony characterization on solid “79” culture medium; pH (3.0-9.0) and heavy metal (Cr, Cd, Zn, Cu, and Ni) tolerance; partial sequencing of 16S rRNA gene; presence of nodA and nifH genes and symbiotic efficiency. A total of 115 isolates, i.e., 77 from M. atropurpureum and 38 from V. sativa, were obtained. The tolerance of these isolates is high for a wide range of pH levels and heavy metal contents, and 18 among them were selected for symbiotic efficiency and 16S rRNA sequencing. Inoculated with M. atropurpureum, the strains UFSM-B53, UFSM-B64, and UFSM-B74 had high symbiotic efficiency. The nitrogen-fixing bacteria were classified in the genera Rhizobium, Bradyrhizobium, and Burkholderia. The results indicate the potential of these native rhizospheric bacterial strains as inoculants and biofertilizers for legume species under pH and heavy metal stress in coal mining degraded areas in Southern Brazil.


INTRODUCTION
Coal provides two-thirds of the non-renewable energy resources in Brazil and reserves are 20 and 75 times higher than those of oil and natural gas, respectively.A total of 847.5 billion tons of coal is enough to maintain the current output for 130 years (Aneel, 2008).In Brazil, the major coal reserves are located in the region of Candiota, in the state of Rio Grande do Sul, with an estimated reserve of one billion tons, which can be extracted by open-cast mining down to a depth of 50 m (Stumpf et al., 2016).During mining and coal processing, a significant amount of sterile solid waste and sulfiderich (pyrite, marcasite, esfarelita, arseno-pyrite, galena, and chalcopyrite) tailings are produced.Upon exposure to oxygen, water, and Thiobacillus ferrooxidans, these materials are oxidized and cause acidification, inducing acid mine drainage (AMD).Due to AMD, dissolute minerals increase the solubility of some metals (As, Cd, Cu, Ni, Pb, Zn, Al, Cr, Mn, Mg etc) (Fungaro and Izidoro, 2006).As a result of this process, the contamination risk of surface and groundwater as well as of the soil increases, and the restoration of these areas become more expensive.
Biological technologies, e.g., bacterial nitrogen fixation, represent efficient and low-cost alternatives for the recovery of mined areas (Valls et al., 2000;Sriprang et al., 2002;Sriprang et al., 2003;Vázquez et al., 2006;Wu et al., 2006).Bioremediation technologies are usually based on the biochemical and genetic capacity of microorganisms to interact and survive under adverse environmental conditions (Vázquez et al., 2006).Therefore, the introduction of plants in association with nitrogen-fixing bacteria constitutes a promising biotechnological tool for the recovery of problematic environments and degraded ecosystems (Chen et al., 2003;Wani et al., 2007;Ferreira et al., 2012).
Numerous studies have demonstrated the potential of nitrogen-fixing microorganisms isolated from copper, zinc, cadmium, lead, and arsenic mining areas to favor plant growth and enhance soil organic matter (Trannin et al., 2001;Matsuda et al., 2002;Carrasco et al., 2005;Zhan and Sun, 2011;Ferreira et al., 2012;Klonowska et al., 2012).Native nitrogenfixing bacteria in soils with high heavy metal contents and low pH are adapted to these stress types.Therefore, the isolation and selection of bacteria under these conditions can provide important information about the best-adapted genotypes for each scenario (Trannin et al., 2001;Ferreira et al., 2012;Klonowska et al., 2012).In addition, legume inoculation with efficient strains, well-adapted to these conditions, is extremely important both from the ecological and economic point of view (Franco and Faria, 1997).
A wide range of prokaryotic microorganisms with broad morphological, physiological, genetic, biochemical, and phylogenetic diversity perform N fixation.This diversity not only ensures the resilience of biological N fixation in ecosystems, but also facilitates it in the most diverse terrestrial habitats (Moreira and Siqueira, 2006;Moreira et al., 2010).Native nitrogen-fixing bacteria of mine-degraded areas are probably better adapted to the local soil and environmental conditions than non-native populations.However, only one study by Brazilian authors described the isolation, characterization, and symbiotic efficiency of nitrogen-fixing bacteria in another coal mining area, in the state of Santa Catarina, with heavy metal contamination and acidic pH (3.5) (Moura et al., 2016).Nevertheless, a high diversity of these bacteria may favor symbiosis with various legume species and maximize biological N fixation in similarly degraded areas.Therefore, the diversity of these nitrogen-fixing bacteria must be investigated, in view of their important role for ecosystem sustainability.
Therefore, the hypothesis of this study was that bacteria isolated from coal mining areas with low pH and high heavy metal contents will also be tolerant to these conditions in vitro, and are efficient as biological N fixers.This study aimed to isolate nitrogenfixing bacteria from coal mining areas under different vegetation restoration procedures to characterize their tolerance to low pH levels, high heavy metal contents, and their symbiotic efficiency.

Isolation and characterization of nitrogen-fixing bacteria
Macroptilium atropurpureum and V. sativa were used as trap plants for the isolation of nitrogen-fixing bacteria.These crops were chosen because of their well-known (2) Soil digested using HNO 3 , according to EPA 3051 method.
(3) Coalmine waste is the material from coal mining, ash constituents, fragmented rocks, sulfur compounds, and low-quality coal.capacity to harbor a high diversity of symbiotic microorganisms (Lima et al., 2009).Plants were grown from April to May 2012 in 110 cm 3 plastic pots, in a greenhouse of the Department of Soil Science, UFSM.Pots were filled with a vermiculite -sand (ratio of 2:1) mixture and autoclaved for 1 h at 121 °C.A completely randomized design with a factorial arrangement (7 soil samples × 2 controls × 7 dilutions) with three replications was used.The controls were fertilized with 210 mg L -1 N and a low concentration of mineral nitrogen (5.25 mg L -1 N).Each trap plant species corresponded to one experiment.
To isolate nitrogen-fixing bacteria, the trap plants were inoculated with 1 mL suspension of soil sample collected from each of the seven areas.The soil suspensions were prepared by mixing 10 g of soil in 90 mL of 0.85 % NaCl solution, stirred for 30 min.In addition to the soil suspensions, two controls without inoculation (one with and the other without mineral N) were tested.All treatments were replicated twice.
To estimate the most probable number (MPN) of rhizobia in the seven areas under study, 1 mL aliquots of soil suspension per sample were serially diluted from 10 -1 to 10 -7 in saline solution (0.85 %) and inoculated on test plants (Table 2).They were categorized as positive or negative, based on the presence or absence of nodules in each dilution, and the MPN Enumeration System (Mpnes) software was used to estimate the MPN (Bennett et al., 1990).
Seeds were surface-sterilized with 70 % ethanol and 1 % sodium hypochlorite, for 1 min per solution, rinsed six times (1 min each) in sterile distilled water, and four inoculated seeds per pot were sown.Every two days, Hoagland and Arnon (1950) nutrient solution (½ strength) was applied, while autoclaved distilled water with a low N concentration (10 mg L -1 ) was provided occasionally in-between when necessary.
Plants were harvested 50 days after emergence (DAE) for the isolation of root nodule bacteria.Five nodules were carefully removed from the roots and used for rhizobial isolation.The nodules were immersed in 95 % ethanol for 30 s, in H 2 O 2 for 1 min, and then rinsed six times with sterile water to remove excess H 2 O 2 .Nodules were then crushed with sterile forceps and spread onto plates containing culture medium 79 (Fred and Waksman, 1928).After obtaining pure isolate colonies, these were transferred to plates containing culture medium 79 and cultured for phenotypic characterization.Cultures were stored at 4 °C.The morphological characteristics of the isolates, including growth rate, determined by the time of appearance of isolated colonies (rapid growth -after 2 to 3 days; intermediate -4 to 5 days; slow -more than 7 to 9 days) and the changes in pH (acid, alkaline, and neutral medium) were also recorded. (1)Coalmine waste is the material from coal mining, ash constituents, fragmented rocks, sulfur compounds, and low quality coal.
Measurements were performed after nine days (OD, 560 nm), and any particular strain considered tolerant only if the OD ≥0.5.Based on the results, strains with high growth at pH 3.0 and 4.0 and at concentrations of 0.86, 3.0, 5.0, 5.0, and 15.0 mg L -1 of Cr, Cd, Zn, Cu, and Ni, respectively, were selected as potential acidic pH and heavy metal-tolerant candidates.

Identification of acidic pH and heavy metal tolerant strains based on 16S rRNA, nodA, and nifH
The strains (with acidic pH and heavy metal tolerance) were identified by the similarity and phylogenetic analysis of the 16S rRNA gene partial sequence and the presence of nodA and nifH genes.All selected isolates were grown in liquid medium 79 at 28 °C for 3 and 7 days to obtain cells in the logarithmic growth phase.The extraction kit Zr Fungal/Bacterial DNA Miniprep ® (Zymo Research Corp, USA) was used for genomic DNA extraction, following the manufacturer's instructions.The 16S rRNA gene was amplified using primer pair 27F and 1492R (Lane, 1991).Polymerase chain reaction (PCR) was performed with PCR buffer 1X, 2.5 mmol L -1 MgCl 2 , 0.2 mmol L -1 dNTP, 0.2 μmol L -1 primers, 0.02 U Taq DNA polymerase (Invitrogen), ultra-pure sterile H 2 O, and 1 µL DNA, for a final volume of 25 mL.Amplification was performed in a Techne ® Endurance TC-312 thermal cycler.Thermal specifications were: one initial denaturation step at 94 °C for 5 min, 30 denaturation cycles at 94 °C for 40 s, annealing at 55 °C for 40 s, extension at 72 °C for 1.5 min, and a final extension at 72 °C for 7 min (Ferreira et al., 2012).The amplified product was resolved on 1 % agarose gel and visualized under UV light to confirm amplification.Subsequently, the amplified material was purified with a PureLink ® Quick Gel Extraction Kit (Invitrogen, São Paulo, Brazil), and sequenced using an ABI PRISM ® equipment-3100 Genetic Analyzer (Applied Biosystems, USA).Gene sequences were analyzed using the Staden Package 2.0.0bsoftware (Staden et al., 2003) and then subjected to BLASTn (NCBI) for comparison with GenBank sequences.Alignment, editing, and phylogenetic analysis were performed using Mega 5.0 (Tamura et al., 2011).The neighbor-joining clustering method (Saitou and Nei, 1987) and Kimura 2-parameter distance model (Kimura, 1980) were used.Reference strains SEMIA 806 (Mesorhizobium sp.) and CIAT 889 (Rhizobium tropici) were also sequenced.The sequences were deposited in the GenBank database .
The nodA gene was PCR-amplified using the oligonucleotide primers nodABurkF and nodABurkR for both α and β-proteobacteria.The nifH gene was amplified with primers nifH-R and nifH-F for β-proteobacteria, and nifH-F and nifH-I for α-proteobacteria.The PCR products were separated on 1 % agarose gel and visualized under UV light using a 100 pb molecular DNA ladder marker (Invitrogen).

Plant nodulation and symbiotic efficiency tests
The symbiotic efficiency of the selected strains was evaluated on M. atropurpureum and V. sativa plants in a greenhouse.Each species was examined in one experiment.Plants Rev Bras Cienc Solo 2018;42:e0170171 were grown under the conditions described previously for the isolation of the strains.For inoculation, the selected strains were grown in liquid medium 79 (Fred and Waksman, 1928) for five days under constant shaking at 28 °C.Each seed was inoculated with 1.0 mL of bacterial suspension (1 × 10 9 cells mL -1 ).The Hoagland and Arnon (1950) nutrient solution (½ strength) in the plant pots was replaced every two days by fresh autoclaved solution, and the volume maintained by adding sterile distilled water.Ten days after germination only one plant per pot was left.
The experiment was arranged in a completely randomized design with three replications.
The treatments consisted of inoculation with selected strains and two uninoculated controls treated with either 210 mg L -1 or 21 mg L -1 of mineral N.
Plants were harvested 50 days after emergence, and the morphological parameters root (RDM) and shoot dry matter (SDM), number of nodules per plant (NNP), nodule dry weight (NDW), and symbiotic efficiency (SE) were determined.The dry matter yield (dry matter of shoots, dry matter of roots, and nodule dry weight) was estimated after the biomass was dried in a forced-air oven (±65 °C) until reaching constant weight.The symbiotic efficiency of each isolate was calculated by the expression SE (%) = [(Shoot dry matter of inoculated treatments × 100) / (Shoot dry matter of control with 210 mg L -1 mineral N)].

Statistical analysis
The data were analyzed statistically with Sisvar, version 5.6 (Ferreira, 2011).All data were transformed when necessary to meet the assumptions of normality and homoscedasticity.Subsequently, the data were analyzed by Anova and the means compared by the Scott-Knott test (Scott and Knott, 1974), considering a nominal level of significance of 5 % probability (p<0.05).The NNP and NDM data were transformed to (x + 0.5) 0.5.

Isolation and characterization of nitrogen-fixing bacteria
The native rhizobia population was calculated by the serial dilution method using M. atropurpureum and V. sativa as trap plants (Table 2).One hundred and fifteen bacterial strains were obtained (77 strains from nodules on M. atropurpureum and 38 from V. sativa) (Figure 1a).Strains were obtained from all areas, but the soil with the longest restoration period (20 years re-vegetated with Acacia mearnsii + Eucalyptus sp.) had the highest number of strains (or isolates).The results indicated that the area with shortest time of revegetation (A2 -6 years) and one area with Eucalyptus sp.(A4 -20 years) had the lowest number of isolates.
The 115 strains were phenotypically grouped, according to the growth rate and ability to change the pH of the culture medium (Figure 1b).Six phenotypic groups were observed: (i) slow growth, medium alkalinization (SAL); (ii) slow growth, no alteration of medium pH (SN); (iii) slow growth, medium acidification (SA); (iv) fast growth, medium alkalinization (FAL); (v) fast growth, no alteration of medium pH (FN); and (vi) fast-growth, medium acidification (FA).Only strains of M. atropurpureum showed slow growth without altering the medium pH (SN), and fast growth with medium alkalinization (FAL).On the other hand, isolates of V. sativa L. presented slow growth with acidifying capacity (SA).

Acidic pH and heavy metal tolerance assay
The results indicated differential growth and pH tolerance among the 115 isolated strains (Figure 2).Of these 115 strains, 98.2 % grew at pH 9.0, 100 % at pH 7.0, 88.6 % at pH 5.0, 72.2 % at pH 4.0, and only 2.6 % at pH 3.0.The isolates from areas A3, A5, A6, and A7 were more tolerant to pH 4.0.One strain from area A2 and two from area A3 maintained growth at pH 3. For the heavy metal tolerance assay, a significant number of bacterial strains grew at higher metal concentrations, although a differential trend was observed (Figure 3).Forty-one strains were tolerant to 0.86 mmol L -1 Cr.No significant tolerance effects were observed for the strains M. plurifarium (BR 3804), R. tropici (CIAT 899 T ), B. cepacia (LMG 1222 T ), and Bradyrhizobium spp.(BR 2001 andBR 2811), which all responded similarly to Cr stress.Rev Bras Cienc Solo 2018;42:e0170171 Twenty-two percent of the strains were tolerant to 3.0 mmol L -1 Cd, including all reference strains.The highest tolerance levels for Cu (5 mmol L -1 ) and Zn (5 mmol L -1 ) concentrations were observed for 8 and 17 % of the evaluated strains, respectively.Ten strains tolerated the highest Ni concentration (15 mmol L -1 ), including B. cepacia (LMG 1222 T ) and Bradyrhizobium sp.(BR 2811).In general, the strains showed the following order of tolerance to the heavy metals tested: Cr > Cd > Zn > Ni > Cu.Isolates from area A3 were more tolerant to all concentrations of metals, followed by isolates from areas A5, A6, and A7, respectively (Figure 3).

Phylogeny of 16S rRNA and phylogenetic identification of nodA and nifH symbiotic genes
The 18 strains (13 from M. atropurpureum and 5 from V. sativa) selected based on pH tolerance and resistance to high heavy metal concentration (Table 3) were identified by 16S rRNA gene sequence analysis (Table 4).The comparison of the 16S rRNA gene sequences revealed that the bacterial strains were phylogenetically related with nitrogen-fixing bacteria, belonging to genera such as Rhizobium, Bradyrhizobium, and Burkholderia (Table 4).All selected strains for acidic pH and heavy metal tolerance amplified DNA fragments related to the nifH (350 bp) and nodA (530 bp) genes (Table 4).

Plant nodulation and symbiotic efficiency tests
The results for the nodulation and symbiotic efficiency of the 18 selected strains with their respective host plants are shown in table 5.The 13 selected strains isolated from M. atropurpureum favored their growth development.Strains UFSM-B53, UFSM-B64, and UFSM-B74 showed similar results for plant growth, with increases of 220, 240 and 250 % in dry matter, respectively, compared to the control without mineral N application.Inoculation with all strains, except UFSM-B32, provided better root development in both crops.The results also indicated the above strains not only for higher dry matter, but also the highest number of nodules, nodule dry weight and highest symbiotic efficiency of the inoculation treatments.For the other strains the results for nodulation potential and symbiotic efficiency were insignificant.Inoculation with UFSM-B11, UFSM-B12, UFSM-B33, UFSM-B34, and UFSM-B65 in V. sativa did not improve plant growth when compared to the treatment with mineral N (Table 5).These strains were able to induce higher nodulation but had a lower symbiotic efficiency than the treatment with mineral N. (1) Based on the most similar sequence found in GenBank.
(3) pH of the culture medium.

DISCUSSION
The data of this study are not conclusive to support the hypothesis that bacteria isolated from coal mining areas with low pH and high heavy metal concentration are also more tolerant to these conditions in vitro and efficient as biological N fixers.There was no relationship between pH level or heavy metal concentration in the soil inhabited by the bacteria and their growth in culture medium at different acidity levels or metal concentrations.In addition, only three isolates stood out as good N fixers.Even in acid or heavy metal-contaminated soils, it is probable that neutrophilic and alkalophilic or heavy metal-sensitive bacteria can survive at microsites with specific environmental conditions (Vos et al., 2013).
Bacteria were isolated from all studied areas.However, higher numbers of bacteria were isolated from soils under plant establishment (A3 -eight years of spontaneous vegetation growth) and from tailing areas (A5 -20 years under Acacia mearnsii + Eucalyptus sp.plantations) (Figure 1a).The high number of isolated strains may be related to the restoration period of the areas.In these environments, soil organic matter (SOM) is the main indicator of the sustainability of a cultivation system (Conceição et al., 2005), mainly in tropical soils (Solomon et al., 2002).An increase in SOM in these areas has a positive impact on soil biodiversity.
In coal-mining areas, all forest cover and the most fertile soil layer is removed to facilitate the access for mineral exploitation.This, in turn, results in the disruption of nutrients and energy flow.Under natural revegetation, N is one of the most limiting nutrients in the soil surface layers (Vitousek and Howarth, 1991).In degraded areas of tropical countries, soil N availability is low.Supplying N in such areas may favor the process of natural succession.Therefore, herbaceous and tree legume species that establish symbiosis with nitrogen-fixing bacteria can be a primary source of N, facilitating plant recolonization and biodiversity enrichment (Siddique et al., 2008).The addition of low C:N plant residues (Nardoto et al., 2008) contributes to soil recovery, as well as to increase biological activity and SOM stabilization (Lavelle, 2000;Resh et al., 2002).
Similarly to the variability identified by the phenotypic characterization, the number of strains varied as a function of the host plant (Figure 1b).When V. sativa was used as trap plant, there was a predominance of fast-growing strains with acidification capacity, which is a typical characteristic of the genus Rhizobium.The number strains with greater phenotypic diversity was highest for M. atropurpureum.This legume is considered promiscuous in its ability to establish symbiosis with nitrogen-fixing bacteria, capable of associating with various genera of both α and β-rhizobia (Lima et al., 2009).
The strains UFSM-B21 (Bradyrhizobium sp.) and UFSM-B33/UFSM-B34 (Burkholderia sp.) had better growth in culture with pH 3.0 (OD560 nm ≥0.5), similar to the pH of the soil from which they were isolated.Bacteria of very acidic sites must have efficient mechanisms to keep the cytoplasmic pH several units higher than the pH of the external medium.The reviews of Baker-Austin and Dopson (2007) and Krulwich et al. (2011) on homeostasis in acidophilic bacteria indicate the following mechanisms: potassium-transporting ATPases, reduction of permeability of cell membranes, outflow of H + out of the cell, expression of higher number of transporters, binding of H + to molecules within the cytoplasm, more efficient DNA and polypeptide repair systems, and degradation of organic compounds decoupled from the electron transport chain.Although these are the most common mechanisms, not all acidophiles have all these mechanisms, because their occurrence is specific to certain bacterial groups.Several studies demonstrated the presence of nitrogen-fixing bacteria in soils contaminated with Cd, Zn, Cu, Pb, and Ni (Matsuda et al., 2002;Melloni et al., 2004;Carrasco et al., 2005;Becerra-Castro et al., 2011;Weyens et al., 2013).However, this is the first report of nitrogen-fixing bacteria isolated from coal mining areas contaminated with heavy metals under extremely acid (pH = 3.5) soil conditions in Brazil.Strains UFSM-B55 (Bradyrhizobium sp.), UFSM-B65, and UFSM-B74 (both Rhizobium sp.) were grown at 15 mmol L -1 Ni, along with B. cepacia (LMG 1222 T ) and Bradyrhizobium sp.(BR 2811).
The ability of some microorganisms to grow under high metal concentrations can result from intrinsic or induced mechanisms, as well as from other environmental factors in the culture medium, such as pH, redox potential, etc., which may reduce the metal toxicity (Zouboulis et al., 2004;Leedjärv et al., 2008;Xiao et al., 2010).Different mechanisms of metal tolerance of these bacteria are well-documented.They vary little between strains isolated from non-contaminated and contaminated environments (Barkay and Schaefer, 2001;Matsuda et al., 2002;Zouboulis et al., 2004;Ferreira et al., 2012).These characteristics explain the survival and persistence ability of certain strains examined in this study, unlike those isolated from environments with low heavy metal contents.
Out of the 13 selected strains obtained from M. atropurpureum, 12 belong to the genus Bradyrhizobium and one to genus Rhizobium (Table 4).The phylogenetic analysis of 10 strains revealed that seven of them belong to subgroup-II, in which the traditional species B. elkanii is also nested.Three belong to subgroup-I, one of which is B. japonicum (Figure 4).However, due to the high similarity of the 16S rRNA gene, further analyses are needed to better characterize these strains.Many studies that isolated nitrogenfixing bacteria from V. sativa have reported the presence of Rhizobium leguminosarum bv.viciae (van Berkum et al., 1995;Laguerre et al., 2003;Moschetti et al., 2005) and Rhizobium sp.(Lei et al., 2008).Here, out of the five selected strains isolated from V. sativa and tolerant to acidic pH and heavy metals, three belong to Rhizobium and two to Burkholderia (Table 4).Phylogenetic analysis (Figure 5) showed that two strains Rev Bras Cienc Solo 2018;42:e0170171 were close to R. leguminosarum and one to R. phaseoli.This is the first report on the presence of Burkholderia on V. sativa.This bacterium, when re-inoculated, could induce nodule formation with a relative efficiency of 50 % (Table 5).The nodulation ability of V. sativa may be due to a lateral gene transfer from the symbiotic genus Rhizobium into Burkholderia, since symbiotic genes are not very different between α and β-rhizobia (Moulin et al., 2002;Chen et al., 2003).
In this regard, strains UFSM-B53 and UFSM-B54 (Bradyrhizobium sp.) and UFSM-B74 (Rhizobium) had the highest nitrogen-fixing efficiency in M. atropurpureum (Table 5).The selection of efficient nitrogen-fixing strains for the reforestation of degraded areas contributes to the adaptation of plant species.These strains can be an alternative inoculant for coal mining areas (tolerant to environmental stress), since M. atropurpureum may fix up to 181 kg ha -1 yr -1 N (Moreira and Siqueira, 2006).
This study describes some efficient nitrogen-fixing strains in acidic pH and heavy metal tolerance found in coal-mined soils in the state of Rio Grande do Sul, Southern Brazil.Even though the strains proved to be symbiotically efficient, tolerant to high acidity and elevated levels of heavy metals, these isolates still need to be tested under these conditions in field experiments.Efficient isolated strains can provide advances towards plant establishment and restoration of coal-mined degraded areas contaminated with high heavy metal levels.Furthermore, these potential strains must be propagated on a commercial basis to optimize their biological viability and activity for field applications.

CONCLUSIONS
Most of the 115 nitrogen-fixing bacteria isolated from the coal mining areas are tolerant to very acidic pH (4.0), with few tolerating extremely acidic pH levels (3.0).
The ability to tolerate high heavy metal concentrations is highly variable among the isolated nitrogen-fixing bacteria, but the following order of tolerance can be established: Cr > Cd> Zn > Ni > Cu.
The 18 selected nitrogen-fixing bacteria belong to the genera Rhizobium, Bradyrhizobium, and Burkholderia; Rhizobium UFSM-B74 and Bradyrhizobium sp.UFSM-B53 and UFSM-B64 on M. atropurpureum have high symbiotic efficiency.For being tolerant to low pH and high heavy metal concentrations as well, they are promising for the production of specific inoculants for areas affected by mining.

Table 2 .
Most Probable Number of nitrogen-fixing bacteria in each

Table 3 .
Values of pH and heavy metal concentration tolerated (mmol L -1 ) by 18 bacterial strains selected for sequencing the 16S rRNA Rev Bras Cienc Solo 2018;42:e0170171

Table 4 .
Origin, characteristics, and identification of bacterial isolated strains obtained from nodules of Macroptilium atropurpureum and Vicia sativa cultivated in the coal-mining region of Candiota, RS-Brazil

Table 5 .
Influence of the tested strains on shoot dry matter, root dry matter, number of nodules, nodule dry matter, and symbiotic efficiency (SE) of Atropurpureum macroptylium and Vicia sativa and two controls, C/N (with mineral nitrogen) and S/N (without mineral N) Same letter indicates no difference by the Scott-Knott test at 5 % probability.
Phylogenetic tree based on 16S rRNA gene sequence similarity of Bradyrhizobium strains.The phylogenetic tree was built using neighbor-joining and Kimura 2-parameters (K2P).Numbers on branches indicate bootstrap values, with a total of 1000 replicates; only values greater than 50 are shown.Bar indicates 2 substitutions per 100 nucleotides.
Figure 5. Phylogenetic tree based on 16S rRNA gene sequence similarity of Rhizobium strains.The phylogenetic tree was built using neighbor-joining and Kimura 2-parameters (K2P).Numbers on branches indicate bootstrap values, with a total of 1000 replicates; only values greater than 50 are shown.Bar indicates 2 substitutions per 100 nucleotides.