*2.5. Endophyte Isolation*

The contaminated site is rich in spontaneous species that have developed on the reported soil, most of which were probably already present as seed banks in the fill soil. The prevalent herbaceous vegetation has the homogeneous characteristics of the surrounding area.

For this reason, endophytic bacteria were isolated and characterized from three of the most commonly represented spontaneous species: *Lotus cornicolatus*, *Sonchus tenerrimus*, and *Bromus sterilis*. The soil around the roots was removed by repeatedly rinsing with tap water. To sterilize the root surface, they were treated with 70% EtOH for 5 min, with NaClO for 2 min and again 70% EtOH for another 5 min. They were then thoroughly rinsed at

least three times with sterile H2O. The roots, finely chopped with a sterile scalpel, were placed in sterile flasks containing TYEG (Trypticase Yeast Extract Glucose) medium and incubated for 16 h at 30 ◦C. Serial dilutions (10-4, 10-6, 10-8) of the obtained suspension were prepared, and 100 μL of each dilution (in triplicate) was spread over R2A Agar (Merck®) plates. Then 100 μL of the third rinsing water was plated to confirm the efficiency of sterilization. Endophyte colonies appeared after 4–5 days.

About twenty phenotypically different colonies for each root type were isolated and pure cultures were used for DNA extraction and taxonomic classification as already described in Franchi et al. [23].

## *2.6. PGPR Characterization*

Endophytes isolated were subjected to a series of in vitro assays to assess their plant growth-promoting potential. The production of auxin indole-3-acetic acid (IAA) was estimated following the method proposed by Shahab et al. [39], siderophore molecules release was determined as described by Milagres et al. [40]. Their ability to solubilize mineral phosphorus (P) was determined by growing the strains in NBRIP (National Botanic Research Institute's Phosphate) according to the protocol developed by Nautiyal [41]. Production of exopolysaccharide (EPS) was estimated using a modified Weaver mineral medium enriched with sucrose [42]. Proteolytic activity was determined as described by Nielsen and Sørensen [43]. The production of ammonia was determined in peptone water (5 g L−<sup>1</sup> peptone and 5% NaCl, pH 7.2), following the method proposed by Kifle and Laing [44].

The isolated strains were also tested for their capacity to form biofilm in vitro, inoculating them in glass tubes with 7 mL of LB (*Luria Broth*) medium. The tubes were incubated at 30 ◦C for 7 days without agitation. The formation of a visible layer (pellicle) at the interface between medium and air indicated a potential capability to produce biofilms. The potential capacity to fix atmospheric nitrogen was evaluated by growing the isolates with a specific nitrogen-free medium (NFb) [45]. The strains that showed at least three growth-promoting properties were selected. With the 12 potentially most promising (Figure 1), a microbial consortium (SG\_1) was prepared in the form of lyophilizate and was then used as an inoculum in the phytoremediation tests.


**Figure 1.** The 12 selected endophytes are listed. In vitro, PGP properties and Genbank Accession number are shown. The + sign indicates the presence of the PGP property, the − sign its absence. Each strain was assigned a score calculated considering the number of PGP properties displayed.

#### *2.7. Next-Generation Ion Torrent Sequencing (NGS)*

An amount of 3 ng of the genomic DNA, obtained by extraction of 500 mg of soil samples and about 200 mg of roots samples through the Fast DNA® Spin Kit for Soil (MP Biomedicals, Irvine, CA, USA) and quantified by Qubit® 2.0 fluorometer (Invitrogen, Waltham, MA, USA), was amplified using the 16S Metagenomics Kit (Thermo Fischer Scientific, Waltham, MA, USA).

The amplification program was set up as follows: 95 ◦C for 10 min, followed by 25 cycles at 95 ◦C per 30 s, 58 ◦C for 30 s, and 72 ◦C for 20 s, a final hold time for 7 min at 72 ◦C and a cooling step at 4 ◦C.

The subsequent purification of the amplicons, the preparation, and the sequencing of the libraries followed the protocols for the Ion OneTouch ™ 2 System, the Ion OneTouch ™ ES, and the Ion PGM™, respectively, as previously described in Conte et al. [46].

The run was based on the workflow Metagenomics 16S w1.1 handling the Database Curated microSEQ®16 S and the reference Library 2013.1. The primers detected both ends to obtain 250 base pairs sequences. Alignment in Torrent SuiteTM Software (version 5.8) was performed using the Torrent Mapping Alignment Program (TMAP). The sequences that occurred only once in the entire dataset were removed, and the representative sequences were defined with a 97% similarity cut-off.

After classifying the Operational Taxonomic Unit (OTU) representative sequences, the output was elaborated to obtain a relative abundance (%) of each OTU in the total amounts of the entire sample.

#### *2.8. Quality Assurance and Quality Control*

Quality assurance and quality control were performed by testing two standard solutions (0.5 and 2 mg <sup>L</sup>−1) for every 10 samples. CRM ERM—CC141 for soil and CRM ERM—CD281 for plants were used as certified reference materials. The values obtained for Pb were 31.8 ± 1.2 mg kg−<sup>1</sup> for CRM ERM—CC 141 and 1.69 ± 0.20 mg kg−<sup>1</sup> for CRM ERM—CD281, in agreemen<sup>t</sup> with the certified values of 32.2 ± 1.4 mg kg−<sup>1</sup> and 1.67 ± 0.11 mg kg−1, respectively.

The detection limit for Pb was 5 μg <sup>L</sup>−1, and the recovery of samples with spikes (5%) ranged from 93% to 101% with a relative standard deviation (RSD) of 1.92% of the mean.
