Soil samples were obtained in January 2013 from Lo Valdivia solar saltern, a sea-salt production site located in central Chile (34°41′50.16″S, 72°00′44.3″W), a place free of AC contamination from agro-industrial sources. Soil samples were collected at six points around the salt ponds (1–1.5 m from the pond edge), where A. vaginata was profusely present and thus defined as A. vaginata rhizosphere soil samples (Figure 1). Soil samples were collected to a maximum depth of 15 cm and stored in a plastic bag. The soil samples, free of root pieces and small stones, were mixed and homogenized and then passed through a 3 mm mesh to prepare the microcosms and perform chemical characterization. The latter was carried out in the Faculty of Agriculture and Forestry laboratory at the Pontificia Universidad Católica de Chile, following standard protocols (Robertson et al., 1999). A fresh, not sifted soil sample was used to isolate rhizosphere halophilic microorganisms.

Eight sets of nine microcosms each were set up with homogenized rhizosphere soil samples as follows: 100 mL beakers were filled with 10 g of soil, covered with self-sealing Parafilm M (Bemis, WI, United States), and incubated at room temperature (25°C ± 2°C). Three sampling times were established (1, 15, and 30 days of incubation). At each sampling time, three microcosms from each set were sacrificed, 2 g homogenized soil samples from each microcosm were used for DNA extraction, and the remaining for HPLC analysis. Each microcosm was supplemented with 1 mL of a solution containing 2,4-D, 4-HB, Phe, or Tyr at 5- or 20-mM final concentration. In addition, two control microcosms were used: one treatment without any AC addition and another where only 1 mL of sterilized distilled water was added instead of AC.

Five grams of each saline soil microcosm were mixed with 15 mL of methanol in a conical bottom tube of 50 mL and stirred for 24 h at 200 rpm. The mixture was then centrifuged at 10,860 g for 45 min to recover the supernatant. The supernatant solution was filtered sequentially through 0.45 μm and 0.22 μm pore size filters. To determine the corresponding AC content, the final solution was analyzed by HPLC. HPLC was performed at the Universidad de Chile soil science laboratory in an instrument connected with a Waters 1525 HPLC binary pump and a Waters 2996 diode array. Samples were isocratically eluted using an aqueous solution of phosphoric acid (3%) at 1.5 mL min⁻¹ flow and a mobile phase of 60:40 (methanol/water), using a Kromasil column 100–3.5C18, dimension 4.6 × 150 mm. Absorbance was measured at 228, 254, 211, and 274 nm to detect 2,4-D, 4-HB, Phe, and Tyr, respectively. Data were managed, visualized, and statistically analyzed using RStudio software (RStudio Team, 2020). Statistical analysis of aligned-rank transform (Wobbrock et al., 2011) was applied to detect significant differences among compound and time factors. “ARTool” package was used to perform ANOVA of aligned rank-transformed data (Elkin et al., 2021), and the packages “emmeans” (Lenth et al., 2018) and “multcomp” (Hothorn et al., 2008) were used to perform pairwise contrasts.

Molecular analysis of bacterial and archaeal communities was performed using metagenomic DNA isolated from 1 g of soil from each microcosm soil sample using the FastDNA SPIN Kit for Soil (MP Biomedicals, Santa Ana, California, United States) according to the manufacturer’s instructions. Amplification of 16S rRNA gene was carried out by polymerase chain reaction (PCR) using the primer pairs 63F (6-FAM-5′-AGGCCTAACACATGCAAGTC-3′) and 1087R (5′-CTCGTTGCGGGACTTACCCC-3′) for bacteria domain (Singh et al., 2006), and 21F (6-FAM-5′-TTCCGGTTGATCCTGCCGGA-3′) and 1492R (5′-GGTTACCTTGTTACGACTT-3′) for archaea domain (Waddell et al., 2010). Both forward primers were labeled with the fluorophore 6-carboxy-fluorescein (6-FAM) at the 5′ terminal end. The reaction mixture for the PCR amplification included 5 μL 10× PCR buffer (200 mM Tris-HCl, pH 8.4, 500 mM KCl), 3.5 μL of 25 mM MgCl₂, 1 μL each of primer 63F/1087R (0.2 μM) for bacterial sequences and 21F/1492R (0.2 μM) for archaeal sequences, 1 μL of dNTP (0.2 mM), 1 μL of bovine serum albumin (0.5 mg mL⁻¹), 0.4 μL of 1 U Taq DNA polymerase, 2 μL of environmental DNA, in a total reaction volume of 50 μL. Amplification was performed using the Applied Biosystems 2720 Thermocycler (Applied Biosystems, Foster City, CA, United States). Reactions were held at 95°C for 10 min to allow the DNA to denature, followed by 25 cycles at 95°C for 45 s, 56°C for 1 min, and 72°C for 2 min, followed by a final extension at 72°C for 7 min. PCR products were visualized by electrophoresis on a 1% (w/v) agarose gel using GelRed as the DNA intercalating agent. The PCR products were subjected to digestion with the restriction enzymes AluI and MspI at 37°C for three h, followed by 20 min inactivation at 80°C. After digestion, products were purified with 0.1 vol of 3 M sodium acetate and 2.5 vol 100% ethanol and centrifuged at 26,290 g for 30 min at 4°C. After washing with 100 μL of 70% ethanol and centrifuging, DNA was eluted using 10 μL of ultra-pure water. Restriction fragment size analysis was performed through capillary electrophoresis (Macrogen Inc., Korea). Data from profiles of the restriction fragments were read in Peak Scanner version 1.0 Software (Applied Biosystems) using the standard marker 1200 LIZ and exported as a text format archive. Exported data were read and managed using RStudio software (RStudio Team, 2020). Only restriction fragments between 50 and 500 bp were considered for analysis. Fragments comprising less than 1% relative abundance were excluded from the study. The data were standardized by calculating the area of each peak as a percentage of the total area (Osborne et al., 2006). Using the R-package vegan (Oksanen et al., 2020), data were square root transformed, and the Bray–Curtis distance was calculated. The similarity matrix was used to represent the data with an NMDS (Clarke, 1993), and treatment clusters were highlighted using the ordiellipse function. Each compound/concentration cluster was highlighted using the ordispider function of the R package vegan (version 2.5–7) (Oksanen et al., 2020). The ANOSIM test, equivalent to a non-parametric ANOVA, was used to test similarity distances. Ecological diversity was also calculated with the Shannon–Weaver index (H′ = S pi * ln (pi)) and richness index (S) as the total number of OTUs found (Shannon, 1948; Blackwood et al., 2007).

To isolate halophilic microorganisms capable of AC removal, 2 g of saline soil were mixed in a 500 mL flask with 100 mL of the modified HM liquid medium (Ventosa et al., 1982). This medium (HTM) contained 117 g of NaCl, 29.7 g of MgSO₄, 20.7 g MgCl₂, 3 g KCl, 0.43 g CaCl₂, 0.39 g NaBr, and 0.1 g HNaCO₃ with 10 mL of 10× phosphate buffer and 10 mL of the micronutrient mix to make 1 L, adjusted to pH = 7.2. To isolate halophiles, HTM medium salt concentration was adjusted to 4.5% salinity by diluting the medium with distilled water. 2,4-D, 4-HB, Phe, or Tyr were added at a 5 mM concentration, and flasks were incubated for 1 month at 25°C. After that, aliquots of 100 μL of a 1:1000 cultures dilution were plated on an HTM medium containing 1.5% agar and 1 mM of the corresponding AC. The agar plates were incubated for 1 month at room temperature. Microorganisms showing visible growth and potentially degrading AC were selected and transferred to a new selective solid medium HTM (4.5% salinity, equivalent to EC_susp = 43.8, Table 1, based on the equation EC = 9.7381*Salinity + 0.9771; R² = 0.9987) containing the respective compound.

One isolated colony was used as inoculum for liquid culture, using the same diluted HTM medium and the respective compound where each colony was originally cultured. An aliquot of 50 μL was taken when turbidity was observed and plated in a solid medium containing 500 mL of HTM medium plus 5 g tryptone, 2.5% yeast extract, and 1.5% agar. Forty-one isolated colonies were selected for taxonomical identification based on 16S ribosomal gene sequencing. A colony was taken and resuspended in 800 μL of Chelex100 (BioRad, Hercules, CA, United States) for DNA extraction. The mixture was incubated at 80°C for 30 min and immediately chilled on ice for 5 min. Each mix was centrifuged at 26,290 g for 15 min, and the genomic DNA supernatant was recovered. Genomic DNA extracts were amplified by the above PCR program for T-RFLP analysis, using no fluorescent labeled primer pairs 63F and 1087R or 21F and 1492R. The PCR products were sent to sequencing (Macrogen Inc., Korea). The sequences were analyzed using BLAST (NCBI database GenBank). The 16S rRNA gene sequences have been deposited in the GenBank database under accession numbers MZ322696- MZ322737.

To explore the ability of the isolates to grow on the four AC tested, HTM saline medium was prepared at 4.5% salinity, and ACs were added at 0.5, 1.0 or 2.0 mM final concentration. Non-inoculated, sterile cultures were used as controls. To determine the growth rates, cell concentrations were monitored at 600 nm (UV–VIS Spectrophotometer, Shimadzu, Kyoto, Japan) after 1, 3, 7, 14, 30, and 60 days of incubation at 200 rpm and 25°C. To quantify AC removal rates, 1 mL samples of each culture and controls were filtered with a 0.02 μm pore size filter and monitored by UV-spectrophotometry. Absorbance was measured at 228, 254, 211, and 274 nm to quantify 2,4-D 4-HB, Phe, and Tyr concentration, respectively, and removal percentages were estimated using the formula (([control] − [sample])/[control])*100. GraphPad Prism version 6.0.0 for Windows (GraphPad Software, San Diego, California United States, www.graphpad.com) was used to visualize data.

Five halophilic microorganisms were selected to evaluate their ability to protect A. thaliana plants (Col-0) under salinity conditions and AC chemical stress. A. thaliana seeds were obtained from the Arabidopsis Biological Resource Center (ABRC, Columbus, OH, United States). The selected isolates were Halomonas sp. LV-8T, Alkalihalabacillus sp. LV-39H, Thalassobacillus sp. LV-41P, and Marinobacter sp. LV-48T, representing the genera with the most significant number of isolates. The archaeal isolate Haladaptatus sp. LV-51T was also included as the only isolate belonging to this kingdom. A plant growth-promoting bacteria, Paraburkholderia phytofirmans PsJN (Poupin et al., 2013), was included as a positive control by its ability to promote the growth of A. thaliana under saline stress (Pinedo et al., 2015). Halophile isolates were routinely grown in (100 mM NaCl/10 mM CaCl₂) saline R2A liquid medium, and P. phytofirmans PsJN was grown in the same medium without salt added. Cell suspensions from each inoculum were then collected and adjusted to approximately 10⁸ colony-forming units per milliliter (CFU/mL), as determined by plate counting. Seeds were surface sterilized with 50% sodium hypochlorite 100% commercial laundry bleach containing 0.1% Tween 20, rinsed three times with sterile water, and kept at 4°C for 5 days to synchronize germination. Square Petri dishes were prepared with half-strength (Murashige and Skoog 1962) medium (MS½), 0.8% agar. To prepare the inoculated plates, the initial inoculum (10⁸ CFU/mL) was homogenously diluted in MS½ 0.8% agar just before gelling to reach a final concentration of 10⁴ CFU/mL of medium. This final inoculum size has been proven optimal for exploring plant growth promotion features (Poupin et al., 2013), but does not represent true environmental relative abundances. Sterilized and synchronized seeds were sown in the Petri dishes with MS½ medium inoculated or not with one strain. Plates were placed vertically in a growth chamber at 22°C with a photoperiod of 12/12 h (light/dark) for 7 days after transplant to specific conditions according to experimental design.

After 7 days of sowing (7 DAS), 16 A. thaliana seedlings per treatment were transplanted to a new square Petri dish containing the same volume and amount of agar without inoculation. Two treatments per strain were tested, one containing MS½ medium and a concentration of 100 mM NaCl plus 10 mM CaCl₂ and another in which only MS½ medium was used. Three non-inoculated treatments (WO_MO treatments) were also run: one in the presence of AC, the other under saline conditions, and a third one combining both AC and salty conditions. Two plates per treatment with 15 seedlings each were incubated in a growth chamber under the same conditions described above for 14 days. At the end of the incubation period, the plates were photographed using a scanner (Epson, Perfection V600 Photo, Nagano, Japan), the root length was measured, and the number of leaves was counted with the naked eye. Subsequently, by analyzing the images taken on day 0 (7 DAS) and day 14 (21 DAS), plant growth was calculated as the difference in the rosette area between 0 and 14 days of incubation.

The effect of AC on the germination rate was tested by sterilizing and synchronizing 500 seeds and subjecting a fraction of these to different AC concentrations using the same medium and conditions described above. A concentration of 0.1 mM of phenol and 2,4-D, 0.5 mM of 4-HB, and Tyr were determined to be toxic to A. thaliana.

The rosette area was image analyzed using Adobe Photoshop Cs3 software. This software allows measurement of the number of pixels corresponding to the area of the Petri dish used (12 × 12 cm). Since all the images were captured at the same resolution and height, this parameter was used to estimate the rosette area, converting the number of pixels into cm². Rosette area data was obtained from images taken on days 1 and 14.

Allenrolfea vaginata rhizosphere soil possessed a high content of soluble anions and cations (Table 1), leading to 4.5% salinity. The most conspicuous anions presented were chlorides and sulfates, and cations such as Na⁺ and Mg²⁺ were the most abundant in this soil. At the same time, the organic matter was found in the normal range for agricultural soils (Table 1). In addition, the high electrical conductivity of the extract and suspension at neutral pH indicated that this soil might be classified as saline-sodic (Table 1).

Rhizosphere saline soil microcosms were spiked with five and 20 mM of 2,4-D, 4-HB, Phe, or Tyr and monitored on days 1, 15, and 30 using HPLC analysis to quantify soil concentration of each compound and calculate removal percentages (Figure 2). At day one no removal of these compounds was observed, while at day 15, almost all 4-HB and Tyr were removed, at both concentrations. Full removal levels at 5 mM and 20 mM were observed after 30 days for 4-HB (99.8 ± 0.1% and 99.6 ± 0.2%, respectively) and Tyr (98.4 ± 0.3% and 99.1 ± 0.3%, respectively). 2,4-D and Phe removal levels increased in time but were lower than the other two compounds. 2,4-D was the less-removed compound as maximum removal levels were 33.9 ± 7.0% and 14.9 ± 1.1% at 5 and 20 mM, respectively. Phe was significantly removed at 5 mM (93.1 ± 1.5%), but removal levels decreased at 20 mM (76.9 ± 2.3%). Nonparametric statistical analysis of variance of aligned rank transformed data (Wobbrock et al., 2011) confirmed that significant differences in removal were due to time and compound factors interaction at both concentrations 5 mM (F_{6, 24.0} = 114.726, p = 2.1852E−11) and 20 mM (F_{6, 24.0} = 87.062, p = 4.3228E−15). A post-hoc pairwise Tukey test showed that significant differences were associated with the non-removal of any compound at day one and removal differences in time of each compound under both compound concentration treatments (Figure 2).

To assess the effects of AC on bacterial and archaeal communities, T-RFLP profiles of saline soil microcosms were used to calculate ecological indexes of diversity (H) and richness (S). S and H values for bacterial communities were consistently higher than archaeal communities (Table 2). Archaeal indexes showed almost no variation in time, with differences present only in one or two operational taxonomic units (OTU). The two control conditions not amended with AC displayed higher levels of S and H in archaeal communities than treated microcosms in archaeal communities. Compared with these two controls, bacterial communities were rapidly and strongly affected by the exposure to AC, increasing the number of OTU detected on day 1 (Table 2). When Phe was added, S decreased after 15 and 30 days of incubation. A similar pattern was observed with the H index (Table 2). 2,4-D effects on community indexes highly depended on the compound concentration (Table 2). A noticeable increase of S and H was observed at day 15 with 20 mM 2,4-D. Tyr and 4-HB treatments produced similar changes in community index changes.

The effects of these ACs on the changes in the structure of the saline soil rhizosphere bacterial and archaeal communities were further studied. Non-metric multidimensional scaling (NMDS) comparison of the 16S rRNA gene T-RFLP profiles showed a clear grouping of samples according to treatment (Figures 3A,B). Phe and 2,4-D treatments presented higher dispersion in the NMDS ordering. Particularly noticeable, the 5 mM 2,4-D treatment showed a more substantial dispersion, mainly in archaeal communities (Figure 3A). The finding that the compound component is the more significant dissimilarity factor was confirmed by the analysis of similarity (ANOSIM), achieving a global R of 0.7433 (p-value = 0.001) for bacterial communities, and 0.7802 (p-value = 0.001) for archaeal communities. The R statistic parameter indicates the strength of the difference where 1 is the strongest and 0 is the weakest, representing significant differences between treated communities. T-RFLP profiles from non-treatment or water-only treatment were also clearly different (Figure 3). The results indicated that the increase of 2,4-D concentration had a more significant effect on archaeal community composition (Figure 3A open and solid diamonds) than on bacterial community (Figure 3B). This phenomenon of compositional divergence produced by different concentrations was not observed with the other compounds (Figure 3). On the other hand, there were no signs of resilience within the 30 days, as each compound generated compositional changes that, in general, accentuated through time (Figure 3, black symbols). The pairwise analysis showed that two controls (non-treatment & water-only) presented R values of 0.667 and 0.556 for archaea and bacteria, respectively (Table 3), which indicates that water addition had a low effect on microbial community structure changes. These two controls showed different results at 15 and 30 days, as R values were 1 and 0.481 in archaeal and 0.296 and 0 for bacterial communities, respectively (Table 3). This indicates that the bacterial community structural differences between no-treatment and water-only treatment decreased over time, while archaeal community structural differences peaked at day 15. Bacterial communities exposed to AC showed a fast change in composition after each compound was added (compare treatments and controls), and the amended compound mainly explained the observed differences.

Microorganisms from the rhizospheric saline soil were tested for their ability to remove/degrade ACs. Forty-one isolates from A. vaginata rhizosphere saline soil were obtained using a solid saline medium containing one of the four AC as a sole carbon and energy source. 16S rRNA sequencing analysis revealed that the 41 isolates corresponded to unique 16S rRNA sequences having more than 1% of sequence dissimilarity between the closest sequence of the isolated microorganisms (Figure 4). From these 41 isolates, Halomonas genus was the most represented (21 isolates), of which 16 isolates corresponded to the H. zincidurans species and five isolates corresponded to possibly new Halomonas species, closely related to each other (Figure 4). The isolates included also six possible new species of Alkalihalobacillus, five likely new species of Marinobacter, three potential new species of Oceanobacillus, two isolates closely related to Priestia filamentosa, one isolate closely related to Rossellomorea vietnamensis and possible new species of Halobacillus, Thalassobacillus and Haladaptatus (one isolate each), the latter being the only archaeal isolate obtained (Figure 4). It should be noted that a cautious expression about possible new species was used as only part (about 1,000 bp) of the 16S rRNA sequence was obtained. Among the 41 isolates, 22 were isolated using Tyr, nine using 4-HB, five using Phe, and five using 2,4-D as a sole carbon and energy source.

Nine isolates, including the single archaeal isolate, were selected for preliminary study AC removal versatility in liquid cultures, using two or three concentrations (0.5, 1.0, and 2.0 mM) of each of the four compounds at 4.5% salinity. Isolates of Alkalihalobacillus, Marinobacter, Oceanobacillus, Halobacillus, Thalassobacillus, Haladaptatus, and three Halomonas were selected, as they showed faster growth on solid agar plates. Despite an incubation of up to 60 days, none of these nine isolates could grow on 2,4-D at any concentration, including the Halobacillus sp. LV-2D and Oceanobacillus sp. LV-4D, which were initially isolated using 2,4-D as a growth substrate. Moreover, these two isolates were the only ones unable to grow and remove 4-HB (Figures 5E,G). Thalassobacillus sp. LV-41P showed low removal of 4-HB, presenting maximum removal levels of 48 and 17% at 0.5 and 1 mM after 30 days, respectively (Figure 5H). The other six isolates could almost completely remove 4-HB, although with different time courses (Figures 5A–D,F,I). H. zincidurans was the only isolate that performed better at 0.5 mM than 1 mM (Figure 5C).

In contrast to 4-HB, Phe removal was slower (Figure 6), as after 30 days, only Marinobacter sp. LV-48T attained significant removal of 0.5 mM Phe (Figure 6D), whereas complete removal was only observed after 60 days (Figures 6A–H), except for the archaeal isolate Haladaptatus sp. LV-51T (Figure 6I). Significant removal of 1 mM Phe was only detected in two Halomonas isolates (LV-8T and LV-42T) and both Marinobacter sp. LV-48T and Thalassobacillus sp. LV-41P (Figures 6A,C,D,H). As judged by O.D. 600 nm, biomass levels were lower than those observed with 4-HB.

Tyr was the only compound that allowed growth and removal at 2 mM after 30 days (Figures 7A,D–F,H,I), except for isolates H. zincidurans LV-24T, LV-42T (Figures 7B,C), and Oceanobacillus sp. LV-4D (Figure 7G). Biomass levels, as judged by O.D._{600 nm}, were similar to those observed with 4-HB.

To address plant growth promotion and protection from saline stress by A. vaginata rhizosphere microorganisms, the interactions of A. thaliana Col-0 and five selected isolates or P. phytofirmans PsJN, a plant growth promotion rhizobacteria as positive PGPR control, were tested. At 7 DAS, plants were transferred to MS½ medium without salt (control) or with salt (100 mM NaCl/10 mM CaCl₂). Plants were photographed 14 days after the transplant, and root length, number of leaves, and rosette area were measured. The rosette area values showed more significant variations, reflecting a two-dimensional parameter. In all the parameters measured, plants in control conditions (without salt) exhibited significant differences from those under saline stress (Figure 8A), represented by a greater root length (F_{1, 6.205} = 181.138, p < 2.22E−16), number of leaves (F_{1, 6.205} = 174.138, p < 2.22E−16) and rosette area (F_{1, 6.203} = 59.2656; p = 5.9191E−13). Under non-saline conditions, Alkalihalobacillus sp. LV-39H, and Marinobacter sp. LV-48T enhanced root length growth compared to the control PsJN or WO_MO (Figure 8A, left panel). No significant differences were found in the other three isolates relative to the WO_MO control. Curiously, strain PsJN did not promote root growth relative to the WO_MO control. Under saline stress conditions, the halophile strains LV-48T and LV-41P showed a clear root growth-enhancing effect (Figure 8A, left panel).

Under non-saline conditions, Thalassobacillus sp. LV-41P and Marinobacter sp. LV-48T showed promotion in the number of leaves compared to PsJN (Figure 8A, central panel). Still, the rest of the halophiles isolated did not present significant differences from the WO_MO control (Figure 8A, central panel). In contrast, only PsJN increased the number of leaves under saline stress with respect to WO_MO control (Figure 8A, central panel). Some halophiles also increased the rosette area under non-saline conditions, as shown for Alkalihalobacillus sp. LV-39H, Marinobacter sp. LV-48T, and Halomonas sp. LV-8T, compared to the WO_MO control (Figure 8A, right panel). Under this treatment, strain PsJN showed a more significant difference, promoting more growth than the halophiles (Figure 8A, right panel).

To study growth promotion produced by these five selected halophiles in A. thaliana, under saline conditions and in the presence of AC, plants were exposed in vitro to salt stress, and in the presence of single AC. A. thaliana seeds were sown in MS½ medium with or without inoculation of selected halophiles, or P. phytofirmans strain PsJN used as a PGPR strain control. At 7 DAS, plants were transferred to MS½ medium containing salt (100 mM NaCl/10 mM CaCl₂). Plants were photographed 14 days after the transplant, and root length, number of leaves, and rosette area were measured.

Regarding the effect of 4-HB on the growth of A. thaliana (Figure 8B), only two of the three plant growth parameters studied showed significant differences between treatments: the number of leaves (F_{1, 6.94} = 4.9788; p = 1.815E−4) and rosette area (F_{1, 6.98} = 4.2686; p = 7.304E−4). Pairwise comparisons through the Tukey test showed that the number of leaves was significantly greater when PsJN was inoculated than treatments inoculated with Alkalihalobacillus sp. LV-39H, Marinobacter sp. LV-48T, and Halomonas sp. LV-8T. However, none of these strains, including PsJN, showed significant differences in control treatment without microorganisms (WO_MO). However, it was observed that Marinobacter sp. LV-48T positively affected the rosette area (Figure 8B). Although there were other significant differences between treatments, they are inconclusive to ensure a positive or negative effect of microorganisms, as they do not differ significantly from the non-inoculated, negative control.

The 2,4-D presence provoked detrimental effects on A. thaliana. The statistical results showed differences between treatment root length (F_{1, 6.93} = 2.926, p = 1.1167E−2), number of leaves (F_{1, 6.93} = 7.297, p = 2.09E−6), and rosette area (F_{1, 6.94} = 8.1197; p = 4.448E−7). However, pairwise analysis showed differences between the control WO_MO and all treatments, indicating that halophile and PsJN strains may protect A. thaliana exposed to 2,4-D under salt stress. Post-hoc contrast analysis showed a significant increase in rosette area in the presence of Marinobacter sp. LV-48T, compared to the control WO_MO (Figure 8B).

Statistical analysis applied to the data related to the presence of Phe and salt stress on the growth of A. thaliana showed significant differences associated with root length (F_{1, 6.95} = 2.7981, p = 1.499E−2), number of leaves (F_{1, 6.95} = 1.1217, p = 4.296E−7) and rosette area (F_{1, 6.96} = 12.856; p = 1.3E−10). A growth promotion effect on root length under salt and Phe stress was observed in treatments with Marinobacter sp. LV-48T and Thalassobacillus sp. LV-41P compared to the control WO_MO (Figure 8B). However, strain LV-41P negatively affected the root length of A. thaliana. On the other hand, the number of leaves decreased in the presence of Alkalihalobacillus sp. LV-39H or Thalassobacillus sp. LV-41P compared to the results obtained in treatments where Marinobacter sp. LV-48T, and Halomonas sp. LV-8T were present. However, these results were not significantly different from those of the control WO_MO, indicating that both promotion and detrimental effects by those microorganisms are mild effects on plants (Figure 8B). Similar results were obtained when P. phytofirmans PsJN was present, which showed a promoter/protective effect but to a lesser extent than the halophiles Halomonas sp. LV-8T, and Marinobacter sp. LV-48T. Similarly, the rosette area decreased in the presence of Thalassobacillus sp. LV-41P compared to control WO_MO, in A. thaliana exposed to salt stress conditions and Phe (Figure 8B).

The presence of Tyr under saline stress showed significant differences in root length (F_{1, 6.92} = 3.2172, p = 6.4991E−3) between plants inoculated with strain PsJN compared to those inoculated with Alkalihalobacillus sp. LV-39H, and Thalassobacillus sp. LV-41P. Nevertheless, these results do not allow us to conclude whether these halophiles positively or negatively affect root length in A. thaliana because there were no significant differences compared to the control WO_MO (Figure 8B). Moreover, statistical analysis showed differences in leaf numbers of A. thaliana between treatments inoculated with Marinobacter sp. LV-48T and PsJN compared to those inoculated with Alkalihalobacillus sp. LV-39H and Thalassobacillus sp. LV-41P. Differences were observed between the control WO_MO and any of the five halophiles studied. Finally, rosette area significantly decreased in plants inoculated with Thalassobacillus sp. LV-41P, compared to the control WO_MO (Figure 8B).