Nodulation of Retama monosperma by Ensifer aridi in an Abandonned Lead Mine Soils in Eastern Morocco

Millions tons of lead and zinc wastes from the abandoned Touissit mine are stored in the open air as dikes in the vicinity of the villages in Eastern Morocco and pose a real danger to both the environment and local populations. To prevent the movement of minerals to the nearby villages and limit the damages to the environment and health, we proposed the nitrogen-fixing leguminous shrub Retama monosperma, as a model plant to use for phytostabilization experimentations. This plant species is known by its ability to grow in hard climatic conditions and in heavy metals contaminated soils. The isolation of bacterial strains nodulating R. monosperma in the abandoned mine soils will permit the selection of rhizobia to inoculate young plant seedlings before their use for the phytostabilization of the mine tailings. In this work, 44 bacteria were isolated from the root nodules of R. Monosperma grown in the Touissit abandoned mine. Twenty-four isolates were considered as true rhizobia as they possess a copy of the nodC symbiotic gene and were able to renodulate their original host. The phenotypic characterization showed that all the strains are tolerant in vitro to different concentrations of heavy metals. The analysis of the 16S rRNA sequences of two selected representative strains showed they were related to different strains of Ensifer aridi isolated from different legumes in three continents deserts. The glnII, recA, and gyrB housekeeping genes analysis confirmed the affiliation of the strains to E. aridi. Moreover, the phylogenic analysis of nodA, nodC, and nifH symbiotic genes showed that the strains are more related to E. aridi JNVUTP6 species isolated from Tephrosia purpurea root nodules in the Thar Desert in India. To our knowledge, this is the first report about the isolation of E. aridi from R. monosperma root nodules.


INTRODUCTION
The Touissit lead-zinc district is located in the south of Oujda (Eastern Morocco), straddling the Moroccan-Algerian border. The mining vocation of this region goes back to the beginning of this century, following prospections undertaken since 1906. However, such as other lead mines in Morocco, it was abandoned in 2002 since the price of this lead and zinc felt down at international level (Smouni et al., 2010). The wastes generated by mineral extraction were maintained in tailing ponds without any rehabilitation plan. These tailings are rich in heavy metals such as lead and zinc which contaminates the surrounding agricultural fields (Argane et al., 2014).
Metal trace elements contaminate surface water resources as well as agricultural products in surrounding areas. This scourge threatens the health of living beings and confirms hence the need for a quick solution. Several studies have shown the effectiveness of legumes utilization for the phytostabilization of degraded soils contaminated by different metallic trace elements (Foulkes, 2000;Gerhardt et al., 2009).
The efficiency of bioremediation depends mainly on the type of plant used and its ability to accumulate heavy metals pollutants, and on the bacterial partner and its ability to tolerate the toxicity of metal elements, lack of important nutrients and the abiotic stresses (El Aafi et al., 2012).
Several studies have shown the tolerance and resistance of rhizobia to different metal stresses (Jaishankar et al., 2014). These bacteria can colonize the soils of lead-rich mine sites. This metal is an environmental pollutant that has no known biological role (Muñoz Vallés et al., 2013). Lead is very dangerous to human health by inducing neurological and renal damages, it disturbs the blood pressure, and reproduction especially in humans and long-term exposure to this metallic element may even cause cancer (Naik et al., 2013;Najeeb et al., 2017). This metal also disrupts the biological processes of plants by replacing vital elements for the plant such as zinc, copper, or manganese (Najeeb et al., 2017), it can induce the total inhibition of plant growth by disrupting its vital processes such as the photosynthetic path and chlorophyllian chain, which are disturbed by damages of the membrane lipid layer and therefore the disruption of ion absorption.
Mining sites are known for degraded and poor soils lacking vital elements such as nitrogen and organic matter. The indigenous plants growing in the mining regions may be considered as naturally tolerant and may constitute an ecological solution for the phytostabilization of local mines soils.
Retama monosperma is a shrubby legume, native to North Africa and Southern Spain, with a large distribution in Morocco (Muñoz Vallés et al., 2013). It is known for its ability to fix atmospheric nitrogen in symbiosis with rhizobia, which strengthens its ability to tolerate extreme conditions, thus playing a key role in ecosystem restoration, desertification control and dune stabilization (Allen and Allen, 1981).
Retama species increases system's forage productivity (Rivest et al., 2011) as they facilitate the growth of herbaceous plants growing in its understory, constituting "islands of fertility, " which are points of high biological activity (Pugnaire et al., 1996). Soil structure is hence stabilized by the roots, and by higher litter and organic matter content which create a microhabitat surrounding the R. monosperma canopy and participate to the retention of more moisture in the soil and leaching losses reduction (El-Bana et al., 2002;Flores and Jurado, 2003;Pugnaire et al., 2004;Dellafiore et al., 2008). The vegetation growing under Retama canopy increases its productivity as it benefits from the micro-environmental conditions and the nutrition via litter fall (López-Pintor et al., 2003;Muñoz Vallés et al., 2011). It can also be used for the phytostabilization of degraded soils (Caravaca et al., 2003) as well as soils contaminated by heavy metals and mining sites.
Bradyrhizobium is considered so far as the main genus of symbiotic nitrogen-fixing bacteria associated with Retama species (Rodríguez-Echeverría et al., 2003;Boulila et al., 2009;Guerrouj et al., 2013b). Previous studies of Retama rhizobia have shown that this plant establishes a symbiotic relationship exclusively with species of the genus Bradyrhizobium. Rodríguez-Echeverría et al. (2003) isolated some bacteria from nodules of R. sphaerocarpa in central Spain and affiliated them within the genus Bradyrhizobium. These isolates were classified later as B. canariense (Ruiz-Díez et al., 2008). The R. sphaerocarpa nodulating bacteria isolated by Boulila et al. (2009) constituted a new phylogenetic clade in the genus Bradyrhizobium. Guerrouj et al. (2013b) described B. retamae, the first species of Bradyrhizobium genus isolated from R. monosperma, in the subhumid area of Saidia in Northern Morocco, which was able to nodulate R. sphaerocarpa but not Glycine max.
However, Mahdhi et al. (2008) reported the nodulation of Retama raetam in arid zones in Tunisia by fast growing isolates that were close to Sinorhizobium kostiense, S. meliloti, and Sinorhizobium sp. and some isolates were related to Rhizobium leguminosarum, Rhizobium sullae, and Rhizobium sp.
The objective of the present study was the characterization of the fast growing symbiotic bacteria isolated from R. monosperma grown in soils of the abandoned mining site of Touissit in the arid Eastern area of Morocco. This characterization was based on a molecular analysis of symbiotic and housekeeping genes of the isolates, to determine their genetic diversity.

Isolation of Bacteria and Culture Conditions
Forty-four bacteria were isolated from the root nodules of R. monosperma plants collected from ten young plants grown in lead (Pb) tailings soil. The soils were sampled in the abandoned mining residues of Touissit, in the Eastern area of Morocco, at the frontiers border with Algeria (34 • 28 26 N -001 • 46 19 W) situated at 1148 m of altitude. This area is submitted to a double stress; the climate is arid with rare and random rainfall (90-100 mm/year) and the skeletal soils are neutral and contaminated with lead and zinc (1250-3300 mg/kg).
Nodules were washed under running tap water; then surface sterilized by immersion in 0.1% HgCl 2 for 2 min, and finally washed thoroughly with sterile distilled water. Nodules were placed independently in Petri dishes and crushed in a drop of sterile water with a sterile glass.
The resulting suspension was streaked onto Petri dishes containing yeast extract-mannitol (YEM) medium (Vincent, 1970) supplemented with 0.0025% (w/v) Congo red. After incubation of the plates at 28 • C for 10 days, colony forming units which represented all of the colony types that could be distinguished by microscopic observation of living cells were chosen. After identification, all rhizobial strains used in this study were routinely grown in Tryptone yeast-extract (TY) medium (Beringer, 1974).

Heavy Metals Tolerance Assays
The intrinsic heavy metal resistance of the isolates was determined on TY agar medium. To this purpose, we prepared different media containing TY amended with each of the following heavy metals at the mentioned concentrations (in µg ml −1 ) HgCl 2 (5); ZnCl 2 (500; 1000); Pb-acetate (50, 300, 500 and 1000); MnCl 2 .4H 2 O (250 and 500); MgSO 4 (500 and 1000); BaCl 2 , 2H 2 O (1000), CoSO4 (100); and FeCl 3 (250). As the strains were isolated from mine tailings rich in zinc and lead, we tested different concentrations of these two metals on the growth of the strains. The strains were streaked on the different solid media and the Petri plates were incubated at 28 • C for 7 days and any growth of the strains was considered a positive reaction.

DNA Extraction, PCR Amplifications, and Sequencing
For DNA extraction and PCR amplifications, genomic DNA was isolated from bacterial cells as previously described (Guerrouj et al., 2013a). The quantity of DNA was determined by using a NanoDrop spectrophotometer (NanoDrop ND2000/2000c, Thermo Fisher Scientific, United States).
Rep-PCR (repetitive extragenic palindromic polymerase chain reaction) were performed using primers REP1 R-1 and REP2-I according to de Bruijn (1992), to reduce the number of strains and avoid any duplicates or clonality.
The DNA template was denatured for 1 min at 95 • C, and PCR was carried out for 35 cycles (95 • C for 15 s, 40 • C for 15 s, and 72 • C for 10 s), with a final elongation step at 72 • C for 1 min, PCR was performed using a MyTaq Mix, following manufacturer's specifications (Bioline Reagents Ltd.). PCR products were analyzed by horizontal electrophoresis in 2% agarose (Bioline) gels stained with ethidium bromide (EtBr) to a final concentration of 0.4% (p/w) (usually 2.5 µl of lab stock solution per 100 mL gel) in 1× Tris-acetate-EDTA (TAE) buffer at 70 V for 3 h. The profiles were photographed with ENDURO GDS Gel Documentation System. Comparative analysis of electrophoretic rep-PCR (enterobacterial repetitive intergenic consensus PCR) patterns was performed with GelCompar II software (version 2.5 Applied Maths, Belgium) by using UPGMA (Unweighted Pair Group Method with Arithmetic Averages).
PCR amplifications of 16S rRNA gene fragments were done using the two universal primers fD1 and rD1 previously described (Weisburg et al., 1991). Amplification products were checked by horizontal electrophoresis in 1% agarose (Bioline) gels stained with ethidium bromide (EtBr) to a final concentration of 0.4% (p/w) (usually 2.5 µl of lab stock solution per 100 mL gel) in TAE buffer at 70 V for 1 h, finally the gels were photographed under UV light.
The primer pairs recAf/recAr (Stepkowski et al., 2005), TSglnIIf/TSglnIIr and gyrB343F/gyrB1043R were used for amplification of recA, gyrB, and glnII genes, respectively, as described by Martens et al. (2007Martens et al. ( , 2008. For the amplification of nodA, nodC, and nifH symbiotic genes, we used the primers nodA1F/nodAb1r; nodCFn/nodCi; nifHf/nifHi as described by Martens et al. (2007Martens et al. ( , 2008. The sequences obtained were compared with those from GenBank using the BLASTN program (Altschul et al., 1990). They were aligned using MEGA 7 software (Kumar et al., 2016). Distances calculated according to Kimura's two-parameter model (Kimura, 1980) were used to infer phylogenetic trees with the neighborjoining analysis (Saitou and Nei, 1987) with MEGA 7 software (Kumar et al., 2016). The sequences were compared to those of several Ensifer species validly published using MEGA 7 (Kumar et al., 2016). The strain B. retamae Ro9 T was used as an external reference taxon in the construction of the phylogenetic trees of the 16S rDNA and the housekeeping genes. Accession numbers of the nucleotide sequences of the Ensifer strains used in this study are shown in the figure trees. For DNA sequencing, amplification products were purified using the PCR product purification system of Qiagen, and subjected to cycle sequencing using the same primers as for PCR amplification, with ABI Prism Dye Chemistry, and analyzed with a 3130 × l automatic sequencer at the sequencing facilities of Experimental Station of Zaidín, CSIC, in Granada, Spain for the 16S rDNA, and the other genes were sequenced at the sequencing facilities of National Centre for Scientific and Technical Research (CNRST) in Rabat (Morocco).

RESULTS
To test the use of R. monosperma for the stabilization of the lead mine tailings, the seedlings were grown in different soils of the Touissit abandoned mine area, in oriental Morocco. The lead (Pb) concentrations in these soils are higher than 3 g/kg of soil (Smouni et al., 2010). After 4 months of vigorous growth, the plants were recovered and checked for root nodulation and 44 bacteria were isolated.
The overall diversity of 44 isolates nodulating R. monosperma obtained in this study was firstly determined by REP-PCR DNA fingerprinting (Figure 1). This method allows also the prevention a All the bacteria isolated from the root nodules of R. monosperma. b All the strains that do not possess nodC gene copy and unable to renodulate R. monosperma. c All the isolates that possess a nodC gene copy and were able to renodulate R. monosperma. d All the strains defined by REP-PCR fingerprinting. e The two representative strains selected following their differential level of tolerance to Pb-acetate.

Percentage of tolerant strains
Heavy metals concentrations of duplication of the isolates. The 44 isolates were reduced to 29 different fingerprints, i.e., 29 strains (Table 1). This method is highly discriminating and enables identification of genetic diversity of the strains at the intraspecies level (Laguerre et al., 1997). The rep-PCR assays showed a high genetic diversity, and the isolates clustered in five main clusters; although the clustering obtained by this fingerprinting approach does not necessarily warrant an inherent phylogenetic value.
The strains were authenticated as rhizobia by two methods. The first method consisted in the determination of nodC nodulation gene by PCR. After amplification using nodCFn-nodCi primers, the 930 bp characteristic band was obtained with only 24. The PCR method warrants a fast way to screen the symbiotic bacteria among large number of isolates, from which the nodulation was confirmed by the second method, i.e., the inoculation of the original host. All the isolates that possess nodulation genes were able to form root nodules once re-inoculated on R. monosperma seedlings in axenic conditions, while isolates that did not possess nodC gene were not.
The isolation of 20 non-nodulating bacteria from nodules reveals the presence of high number of endophytes among the isolates but this is very common in Mediterranean wild legumes due to the presence of several endophytes that can co-inhabit nodules with rhizobia (Muresu et al., 2008).

Phylogenetic Analysis of 16S rDNA
From the selected strains we retained two strains on the basis of the rep-PCR groupings, for further analysis.
A phylogenetic tree based on the sequences of the rDNA showed that the two representative strains were 100% related to Ensifer aridi strains, recovered from Asian, African, and American deserts (Le Quéré et al., 2017). Therefore, we considered that all our strains belong to the genus Ensifer (Figure 3).

Housekeeping Genes Phylogeny
The results obtained showed that the gyrB sequences of the RMT3 and RMT46 strains have a similarity of 99.64 and 99.32%, respectively, with E. aridi strain JNVUTP6 (Figure 4). Similar results were obtained with glnII sequences which showed a similarity of 99.62 and 100% with E. aridi strain JNVU TP6 (Figure 5). The recA sequences of the two strains have 99.47% similarity with E. aridi strain JNVU TP6 (Figure 6).
Analysis of the concatenated sequences of the recA, glnII, and gyrB housekeeping genes (Figure 7) showed that the two strains were related to E. aridi strain JNVUTP6 with 98.91-99.61% of similarity.

Phylogenic Analysis of Symbiotic Genes
Bacterial nodulation (nod) genes are involved in the production of Nod factors which act as specific signals triggering nodule formation. The nodA, nodB, and nodC genes are conserved in all nodulating rhizobial strains. They are essential for the synthesis of the lipo-oligosaccharide backbone. The nodA gene codes for a 21.8-kDa acyl-transferase and determines the type of the N-acyl substitution transferred into the oligosaccharide backbone of the Nod factor and plays a critical role in making this distinction (Suominen et al., 2001). Thereby, the function of the NodA protein specifies the Nod factor structure and the host range (Ritsema et al., 1996). NodC is an N-acetylglucosaminyl transferase (chitin synthase) that produces chitin oligosaccharides (Machida et al., 2001). Ascertaining the presence of nodC and nodA genes copy is an important requirement for attributing symbiotic properties to a nodule isolate.
To analyze the symbiotic genes diversity in our strains, we amplified and sequenced the two symbiotic genes nodA, nodC, and nifH, the nitrogen fixation gene, using specific primers. Sequencing of the nodC gene and phylogenetic analysis permitted the clustering of the two representative strains RMT3 and RMT46 with E. aridi strain jnvuTP6 with a similarity of 100 and 99.85%, respectively (Figure 8).
Amplification of the nodA gene produced a band around 640 bp. The Neighbor joining phylogenic tree generated by nodA sequences showed that the two representative strains formed a group with E. aridi strain JNVU TP6 with 99.82% similarity with RMT3 and RMT46 (Figure 9).
The nifH gene sequences analysis showed that the two representative strains have a high similarity of 99.38% with E. aridi JNVUTP6 (Figure 10).      The phylogenic analysis of the three symbiotic genes concatenated sequences confirmed the results of 16S rDNA sequencing and housekeeping genes (gyrB, glnII, and recA). These results showed that the two strains are related to E. aridi strain JNVUTP6 (Figure 11).

DISCUSSION
The Eastern area of Morocco and particularly the region of Touissit are known for their hard weather conditions with low and random rainfall as well as a very cold winter and a hot summer. Furthermore, the studied site is located on the lead mine dikes that have been abandoned since 2002 (Smouni et al., 2010). These tons of heavy metal scrap cannot be removed or cleaned up and phyto-stabilization is the only effective way to limit the damages to the environment. Phytostabilization is generally used for soils fixation to prevent the movement of minerals to the nearby villages and population. On this site, several trials of phytostabilization have already been tried without much success, because of the type of introduced plants which are not adapted to the climatic and metallic stresses (Argane et al., 2014). We preferred the use of nitrogen-fixing leguminous trees for the phytostabilization of these rejects, which are dangerous for humans and the environment. We chose R. monosperma, which is known for its ability to establish nitrogen-fixing symbiosis with rhizobia and also for its ability to grow in harsh environments (Hannane et al., 2014). However, extreme conditions may limit the nitrogen fixation process (Berrada et al., 2012), which urges the need to search for the best symbiotic partners that can tolerate these stresses.
The majority of studies done on Retama rhizobia have shown that they belong to the genus Bradyrhizobium (Hannane et al., 2014). Guerrouj et al. (2013b) Ahnia et al. (2018). However, some years ago, Mahdhi et al. (2008) isolated some fast-growing bacteria from root nodules of R. raetam in South-western Tunisia belonging to different taxa and whose 16S rDNA was affiliated to E. kostiense, E. meliloti, R. giardinii, R. leguminosarum, R. sullae, and other Rhizobium spp.
In this work we isolated 44 bacteria from the nodules of R. monosperma grown in the soils of the lead-zinc abandoned mine of Touissit. We found that only 24 isolates were true rhizobia and were able to renodulate their original host and possess the nodC gene. The 20 isolates were considered as endophytes unable to nodulate R. monosperma seedlings in axenic conditions. The genetic diversity as assessed by REP-PCR showed that the isolates were regrouped in 10 different fingerprints meaning that we have 10 strains (Figure 1).
All the 10 strains were tolerant to different concentrations of the tested heavy metals. Two strains were selected on the basis of their tolerance to 1000 µg/ml of Pb-acetate as representatives and retained for further studies. Strain RMT46 is able to grow on TY supplemented with 1 mg/ml of lead, whereas strain RMT3 grew only on 500 µg/ml of this heavy metal. However, the concentration of 500 µg/ml of lead is still considered as a highly toxic concentration for bacterial growth in soils (Tiku et al., 2016;Asad et al., 2018).
The molecular analysis, using 16S rDNA and the three housekeeping genes recA, glnII, and gyrB analysis as well as the three symbiotic genes nifH, nodA, and nodC, showed that our strains are related to E. aridi. This newly described species has been isolated in African, Asian and American deserts, from root nodules of V. gummifera in soils of Merzouga desert in Morocco (Sakrouhi et al., 2016;Le Quéré et al., 2017), root nodules of Tephrosia purpurea in the Thar desert in India (Tak et al., 2016), and root nodules of Phaseolus filiformis in the Mexican desert (Le Quéré et al., 2017).
We remarked also that B. retamae strain Ro19T nodC and nodA sequences were related to Ensifer species that nodulate members of Acacia genus in Africa and America (de Lajudie et al., 1994;Nick et al., 1999;Toledo et al., 2003 ; Figures 8, 9). However, the nifH sequences were completely different (Figure 10). This would suggest some lateral transfer of the symbiotic island between B. retamae and E. aridi.
The presence of excessive levels of heavy metals in soil significantly changes the floristic composition of sites. The deleterious effects of heavy metals on nodulation and nitrogen fixation may be due to their inhibitory effects on the growth and activity of both symbionts. When present in excess, heavy metals may also delay the nodulation process, limit the uptake of water and nutrients by plants and affect concomitantly the health of plants (Terry, 1981;Karpiscak et al., 2001). However, R. monosperma supports excessive levels of metals as it grows wildly in heavy metals polluted areas (Belabed et al., 2014).
Moreover, the strains of E. aridi isolated from nodules of R. monosperma in this work are tolerant to high concentrations of lead, zinc, copper, and mercury. This selective tolerance is probably due to the pressure exerted by heavy metals on microorganisms (Pereira et al., 2006) and plants. Heavy metals can alter and disrupt growth, morphology, and activities of non-tolerant rhizobia (Ahmad et al., 2012). This ability allows them to grow and fix nitrogen in symbiosis with their host plant in the mining environment of Touissit. The plant thus becomes able to establish, fix the soil, and thus participate to the phytostabilization of this mining site.

CONCLUSION
In this study, we isolated 44 bacteria, from the root nodules of R. monosperma grown in the Touissit lead mine tailings, in the arid Eastern Morocco. We selected 2 strains representing the 10 efficient rhizobial strains for molecular analysis and found they are related to E. aridi, a species that is adapted to the extreme deserts conditions and improves the resilience of legumes in these hostile environments.
We consider that R. monosperma and E. aridi would be very good candidates for the restoration of this mining site. It will thus be possible to propose an inoculum based on these strains to the National Forest authorities (Haut-Commissariat aux Eaux et Forêts et Lutte Contre la Désertification) to be used for the inoculation of R. monosperma plants during reforestation and phytostabilization programs and campaigns.

DATA AVAILABILITY
The raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher.