Removal and Ecotoxicity of 2,4-D and MCPA in Microbial Cultures Enriched with Structurally-Similar Plant Secondary Metabolites

: The removal of contaminants from the environment can be enhanced by interactions between structurally-related plant secondary metabolites (PSMs), selected xenobiotics and microorganisms. The aim of this study was to investigate the e ﬀ ect of selected PSMs (ferulic acid—FA; syringic acid—SA) on the removal of structurally-similar phenoxy herbicides (PHs): 2,4-dichlorophenoxyacetic acid (2,4-D) and 2-methyl-4-chlorophenoxyacetic acid (MCPA). The study also examines the biodegradation potential of soil bacteria, based on the occurrence of functional tdfA -like genes, and the ecotoxicity of the samples against two test species: Sinapis alba L. and Lepidium sativum L. The microbial cultures spiked with the PSMs demonstrated higher phenoxy acid removal: 97–100% in the case of 2,4-D and 99%–100% for MCPA. These values ranged from 5% to 100% for control samples not amended with FA or SA. The higher herbicide removal associated with PSM spiking can be attributed to acceleration of the microbial degradation processes. Our ﬁndings showed that the addition of SA particularly stimulated the occurrence of the total number of tfdA genes, with this presence being higher than that observed in the unamended samples. PSM spiking was also found to have a beneﬁcial e ﬀ ect on ecotoxicity mitigation, reﬂected in high (102%) stimulation of root growth by the test species. The present study used an assay based on two dicotyledonous plants ( L. sativum and S. alba ) to conﬁrm the ecotoxic e ﬀ ect of studied PHs and their mitigation as a result of ongoing physicochemical and biological processes. The ecotoxicity of pure SE at the start of the experiment exceeded the limit of 20% for both L. sativum (51%) and S. alba (57%): as the soil used for the SE preparations was collected from an agricultural ﬁeld, it is possible that it had been exposed to contaminants prior to collection. After incubation, it was found that although the % removal of 2,4-D in the studied cases ranged between 83–100%, the ecotoxic e ﬀ ect did not change over time, remaining above the toxicity limit of 20%. Of the two plants used in the assay, L. sativum was more sensitive to the treatment, displaying 100% root growth inhibition in all samples with 2,4-D, which is consistent with our previous study This ecotoxicity might have been caused by the formation of 2,4-D metabolites, as the catabolism of 2,4-D leads to the subsequent formation of 2,4-DCP and 3,5-DCC. These intermediates, formed during the degradation processes, have sometimes been found to be more toxic than the original compound [55]. 2,4-DCP has previously been observed to exhibit a strong cytotoxic e ﬀ ect and exert high a ﬃ nity to the structures of plan cells


Introduction
Two of the most commonly-used herbicides in agriculture and the home/garden market sector are phenoxy herbicides (PHs), 2,4-D (2,4-dichlophenoxyacetic acid) and MCPA (2-methyl-4-chlorophenoxyacetic acid) [1]. These compounds are able to selectively control the growth of dicotyledonous weeds [2]. All PHs are constructed from ring-like structures, with at least one chlorine atom attached to the ring at different positions [3]. Their mode of action is similar to that of phytohormones (auxins), in that they disturb the physiological processes of plants and their growth regulation [4].

Contaminant. Studied Plant Plant Secondary Metabolite Observed Effect Literature
Polychlorinated biphenyls (PCBs) n.a. flavonoid, naringin enhancement of PCBs reduction [15] Trichloroethene (TCE) n.a. cumen enhancement of the biodegradation of TCE by R. gordoniae up to 75% within 24 h [18] Dichlorodiphenyldichloroethylene (DDE) Cucurbita pepo low molecular weight acids (e.g., citric acid) enhancement of p,p'-DDE bioavailability in soil [19] Polycyclic aromatic hydrocarbons (PAH) Apium graveolens linoleic acid enhancement of benzo[α]pyrene removal [20] 2,4-dichlorophenol (2, soil samples taken from sites located under Pinus sylvestris, Quercus robur limonene and α-pinene induction of 2,4-DCP degradation by indigenous soil microbiota [21] n.a.-not analyzed. The present microcosm study examines the effect of the application of two PHs, viz. 2,4-D and MCPA, on the degradation of two structurally-similar PSMs, viz. FA and SA, to confirm whether such similarity influences the rate of removal [16]. The study examines the influence of the PSMs addition on (1) changes in PH concentration over time, (2) the presence of PH-degrading genes (3) and changes in ecotoxicity occurring throughout the experiment, the latter being measured using two dicotyledonous plant species: Lepidium sativum and Sinapis alba.

Microcosm Setup
The study was conducted in bacterial cultures containing liquid Mineral Salt Medium (MSM) (1 g/L of KNO 3 , 0.5 g/L of K 2 HPO 4 and MgSO 4 7 H 2 O, 0.05 g/L of NaCl and CaCl 2 , and 0.01 g/L of FeCl 3 ). MSM was filtered through a microbiological Corning™ (New York, NY, USA) Disposable Vacuum Filter (0.22 µL) and enriched with soil microorganisms derived from the agricultural soil extract (SE) 50%:50%, v/v (see Supplementary Materials: Texts S1 and S2; Table S1).
Prepared MSM or MSM+SE samples were amended with 2,4-D (≥95.0% technical purity, molecular weight 220.04 g/mol, 677 ppm water solubility at 25 • [22]. Samples containing only sterile MSM, and sterile MSM + sterile soil extract (SSE) were used as controls to assess the degree of physicochemical degradation. The samples were incubated in darkness at 25 • C for 24 days.

2,4-D and MCPA Concentration Measurement and Their Ppercentage Removal
Subsamples were collected at the beginning of the experiment and after the 24-day incubation period, and these were used for monitoring changes in 2,4-D and MCPA concentration.
The concentration of 2,4-D was assessed using an enzyme immunoassay (ELISA) RaPID Assay ® (A00082, Guildford, UK) 2,4-D Test Kit according to the manufacturer's instructions [23]. Each analytical batch contained a sample blank, a control sample of known concentration, four calibration standards (0, 1, 10 and 50 µg/L) and the test samples. As the initial concentrations of 2,4-D (0.1 mM and 0.5 mM) were outside the range of the standard curve, the samples were appropriately diluted. The precision was verified by duplicate analyses, and the test reproducibility was measured using coefficients of variation (CVs); these CVs should be lower than 10% for the calibration standards, and lower than 15% for the samples. If the CVs exceeded the above values, the whole procedure was repeated to achieve good quality of the obtained results. The minimum detection level of the kit was 0.7 µg/L. Samples showing a concentration lower than the minimum detection level were considered to be negative.
The concentration of MCPA was determined using a GC-MS TQ 8040 (Shimadzu Corp., Kyoto, Japan) gas chromatograph equipped with split/splitless injector operating in a splitless mode at 250 • C, and a triple quadrupole mass spectrometer (MS) connected to "LabSolutions" software (version 4.45, Shimadzu Corp., Kyoto, Japan) extended with Pesticide Smart Database (version 1.03, Shimadzu Corp., Kyoto, Japan) (see Supplementary Materials: Text S3).
The working standard solutions for the calibration study were prepared by spiking the tested samples with the standard solution in the concentration range of 0.01-100 µg/mL. The linear range for MCPA was studied by replicate analysis of the standard stock solutions. The linear regression value was calculated with the mean peak areas of five replicate injections. The linear regression was in the range of 0.19-100 µg/mL with a coefficient of determination of 0.9996. The coefficient of variability, i.e., the percentage of relative standard deviation (CV%), calculated as the mean value of the concentrations across the linear range, was found to indicate good precision (1.1%). The sensitivity of the method was 62.5 ng/mL, considered in terms of the limit of detection (LOD), calculated from calibration functions. The limit of quantization (LOQ), defined as three times the LOD value, was 0.19 µg/mL. After the basic validation parameters had been determined, the environmental samples were analyzed in nine replications. For each sample, the mean value was calculated together with the relative standard deviation.
The percentage removal of 2,4-D and MCPA was calculated in reference to the initial concentration of herbicide used (0.1/0.5 mM). The percentage removal of studied phenoxy herbicides was calculated according to Equation (1): Equation (1) PR-percentage removal of phenoxy herbicides after 24 days of incubation, C i -initial concentration of phenoxy herbicide used in the microcosm preparation; C f -final concentration of phenoxy herbicide measured after 24 days of incubation.

Molecular Analysis
Subsamples for molecular analysis were collected every six days, i.e., four times during the incubation period. The collected subsamples were examined for the occurrence of bacterial 16S rRNA gene fragments and five functional phenoxy acid degradative genes (tfdA, tfdAα and tfdA Class I, tfdA Class II and tfdA Class III). DNA was extracted from the soil bacteria using a GeneMATRIX Soil DNA Purification Kit EurX ® ( Table S2) were used for the amplification of given bacterial DNA fragments. The conditions for the PCR were optimized for each studied gene. PCR products were checked using 1.5% agarose gel electrophoresis and stained with ethidium bromide.

Ecotoxicity Assessment
Ecotoxicity measurements were performed twice during the experiment: at the beginning and after a 24-day incubation period, using Phytotoxkit Test (Microbiotest Inc., Nazareth, Belgium) [25], a commercial toxicity bioassay. The test compares the degree of inhibition of root length of certain test species after three days of exposure to a control sample with that of an uncontaminated control containing only distilled water. For the purpose of this experiment, the dicotyledons Lepidium sativum (L.) and Sinapis alba (L.) were used as test plants. The samples were classified as non-toxic when the percent effect of root growth inhibition (PE%) was ≥20%, slightly toxic for 20% ≤ PE < 50%, toxic for 50% ≤ PE < 100%, and highly toxic for PE = 100% [26].
The percentage effect on root growth inhibition of the studied soil was calculated according to Equation (2): Equation (2) PE-percentage effect of root growth inhibition, A-plant root length in control soil; B-plant root length in studied soil.  In contrast, lower removal rates were observed for the samples amended with 0.1 mM MCPA: 11-40% removal was observed for the sterile, untreated samples, 53% for those treated with SE and 99% for those treated with MSM + SE + SA. Similar removal rates were observed for samples amended with 0.5 mM MCPA: MCPA removal ranged from 5% to 27% for sterile untreated samples, which increased to 99% for the SE samples and 100% for the SE + SA. Table 3 and Figure 1 show the results of the molecular analysis for studied samples. The bacterial 16S rRNA gene fragment was detected in all studied samples containing SE. In the MSM + SE variant, the tfdAα gene fragment was detected after 6, 12 and 18 days of incubation and the tfdA gene fragment after 18 and 24 days. In the samples amended with 0.1 mM 2,4-D, the tfdA Class III gene was present during the whole experiment; however, the gene detection pattern changed after the addition of FA: the tfdA Class I gene was detected after 6, 12, 18 days and the tfdAα gene was detected after 18 and 24 days. The use of 0.5 mM 2,4-D, i.e., the higher concentration, shifted the frequency and timing of gene fragment detection: in samples with SE, tfdAα was detected only once after 24 days and tfdA Class III was observed after 12 and 18 days of incubation. The genes were found to be less apparent in the MSM + SE + FA variant (Table 4): only tfdA Class I was observed, on one occasion, after 6 days, tfdAα was observed after 6, 18 and 24 days.      Among the variants amended with 0.1 mM MCPA, no functional genes were detected in the SE samples. However, the SA samples demonstrated tfdAα on day 6 of incubation, tfdA Class II on day 6 and day 24, and tfdA Class III on days 6, 12 and 18. In the samples with 0.5 mM MCPA, all tested gene fragments were observed (tfdAα after 6 and 24 days, tfdA Class I after 24 days, Class II after 6 and 24 days and Class III after 6, 12, 18 days). In the MSM + SE + SA variant, tfdAα was present after 12, 18, 24 days, tfdA Class I after 12 and 18 days, tfdA Class II after 12, 18 and 24 days, and tfdA Class III after 18 and 24 days. In comparison to the sample amended with SE, the addition of SA was associated with a twofold increase in gene detection (Table 4). After 12 days, tfdA Class III was detected throughout the whole incubation period in samples amended with MSM + SE + SA. When SA was applied, the detection of the studied genes was found to double for both the tfdAα and the tfdA Class I genes.

Molecular Analysis
In addition, most of the samples treated with 0.1 mM MCPA did not demonstrate the presence of any tfdA-like genes; however, 30% of the 0.1 mM MCPA samples treated with SA were found to contain three of the five tested genes: tfdAα, tfdA Class II and tfdA Class III (Table 4). Greater detection rates were observed in the case of samples treated with 0.5 mM MCPA: the genes were detected in 45% of all samples, i.e., 40% when PSM was not added and 50% when it was (Table 3).
Our analysis confirmed that SA spiking increased the number of detected functional genes and MCPA removal efficiency. Hence, the sample containing MCPA, SE and SA was selected for further analysis of the biodiversity of the microbial populations which were present in samples amended with herbicide and selected PSMs. Analyses of the 16S rRNA gene sequences within the samples revealed widespread enrichment of several sequence variants associated with known microbial degraders of many classes of environmental organic pollutants (Table 5). These degraders include members of the genera Rhodoferax, Achromobacter, Burkolderia and Cupriavidus (see Supplementary Materials: Figures S1 and S2).  Table 6 shows the results of the ecotoxicity assay based on two dicotyledonous species: L. sativum and S. alba; measurements were taken at the beginning of the experiment and after 24 days of incubation. At the beginning of the experiment, the MSM + SE sample demonstrated root growth inhibition of 51% (toxic) for L. sativum and 57% (toxic) for S. alba. All samples treated with 0.1 mM or 0.5 mM 2,4-D displayed 100% inhibition (highly toxic) for L. sativum; however, the 0.1 mM 2,4-D samples displayed 98% inhibition (toxic) for S. alba and the 0.5 mM 2,4-D samples 94% inhibition (toxic). All samples amended with MCPA (0.1 and 0.5 mM) demonstrated 100% (highly toxic) root growth inhibition in both studied plant species. After 24 days, in sterile samples amended with 0.1 mM 2,4-D, the ecotoxicity still remained very high. Samples treated with MSM showed 100% growth inhibition (highly toxic) for L. sativum and 96% (toxic) for S. alba. Likewise, MSM+FA demonstrated 98% inhibition (toxic) for L. sativum and 95% (toxic) for S. alba. The MSM + SSE variant demonstrated 100% (highly toxic) inhibition for L. sativum and 97% (toxic) inhibition for S. alba.

Changes in Ecotoxicity
At the higher initial concentration of 2,4-D (0.5 mM), sterile samples demonstrated 100% (highly-toxic) inhibition for L. sativum and 96% (toxic) for S. alba. Both samples with SE and those with MSM + SE + FA displayed 100% (highly-toxic) inhibition for L. sativum and 97% (toxic) for S. alba.
The samples treated with MCPA also demonstrated high ecotoxicity, with the sterile samples being 100% (highly toxic) for both plants at both PH concentrations. Nevertheless, the application of soil extract ameliorated this highly toxic effect. Among the samples amended with 0.1 mM MCPA, the SE samples showed a stimulation effect of −102% (non-toxic) for L. sativum and −34% (non-toxic) for S. alba, while the SE + SA samples demonstrated −69% (non-toxic) for L. sativum and −21% (non-toxic) for S. alba. For the samples dosed with 0.5 mM MCPA, the PE% values for the MSM + SE samples were −47% (non-toxic) for L. sativum and −38% (non-toxic) for S. alba, while those for the MSM + SE + SA samples were −55% (non-toxic) for L. sativum and −35% (non-toxic) and S. alba.

Discussion
To improve crop yields and mitigate the risk of economic loss caused by weeds, it is often necessary to apply herbicides, e.g., PHs, to soil; however, inappropriate storage practices and usage can result in PHs being dispersed throughout the environment. Following dispersal, their residues can accumulate in indigenous plant and aquatic organisms, where they can adversely affect their metabolism. Although they are initially released into the environment in the form of commercial products containing phenoxy acids salts or esters, these products immediately hydrolyze to their corresponding anionic or neutral forms upon exposure to environmental conditions [27]. Residual forms of active ingredients (2,4-D, MCPA) can be adsorbed by soil particles and transported to terrestrial and water ecosystems (surfaceand groundwater) with runoff and in the soil profile [28]. Hence, there is a need to identify methods that enable effective elimination of these pollutants from the environment; of these, nature-based approaches offer the greatest potential, as these are cheaper and less likely to be harmful to the environment. One such approach is based on the stimulation of biodegradation: a promising, environmentally-friendly and cost-effective method of enhancing naturally-occurring processes of contaminant removal. The present study compares the potential of selected PSMs as stimulators of biodegradation of certain PHs which form the basis of widely-used herbicides: FA against 2,4-D, and SA against MCPA. The method employed a tripartite approach: phenoxy acid removal efficiency was assayed, the bacterial communities involved in the biodegradation processes were identified by bacterial gene analysis, and changes in sample ecotoxicity were determined.

The Influence of Selected PSMs on 2,4-D/MCPA Removal
Various PSMs appear to stimulate the biodegradation of organic pollutants in soil (Table 1). However, as the sterile samples and those amended with SE demonstrated similar efficiency in reducing 2,4-D concentration, it is possible that such degradation is driven mainly by physicochemical processes. Recent research by McMartin et al. [29] found the half-life values of 2,4-D in water samples to range between 30 and 40 hours, which is consistent with our present results obtained in sterile samples ( Table 2). McMartin et al. also reported that the microorganisms required an acclimation period of 18 days before commencing the biodegradation of 2,4-D.
In contrast to our results, it has previously been suggested that samples amended with Pseudomonas cepacia populations displayed significant reductions of 2,4-D in comparison to controls, and that biodegradation of 2,4-D is determined by biotic factors [30]. McLoughlin et al. [21] also note that PSMs such as α-pinen and limonene play a significant role in enhancing the removal of 2,4-DCP (2,4-D metabolite: 2,4-dichlorophenol) from soil matrices. Our present findings indicated that both 2,4-D and FA, and their metabolites, exerted a toxic effect on the bacterial consortia in the amended samples; hence, it is more likely that the removal of 2,4-D was mainly driven by physicochemical processes.
After 24 days of incubation, MCPA removal in the sterile variants ranged from 19% to 40% when applied at 0.1 mM, and from 5% to 27% when applied at 0.5 mM; the addition of soil microbiota enhanced the removal rate to 53% (0.1 mM) and 99% (0.5 mM). The application of both SE and SA boosted the process of MCPA removal to 99% for 0.1 mM MCPA and to 100% for 0.5 mM MCPA. However, regarding the soil amended with 0.5 mM 2,4-D, no significant difference was observed between the sterile samples and those amended with soil extract.
These results indicated that two related compounds, in this case, 2,4-D and MCPA, can behave very differently in the presence of structurally-related PSMs: although the biodegradation of MCPA was enhanced by SA, no such effect was observed on 2,4-D by FA. Two critical factors determining the biostimulation effect of PSM appear to be the selection of structurally-related PHs and PSMs, and the initial concentration of the contaminant.

The Influence of Selected PSMs on the Degradation Potential of SE Bacteria toward 2,4-D/MCPA
Previous studies indicated that the biodegradation processes taking place in the rhizosphere are stimulated solely by PSMs. They can serve as cometabolites and provide the energy needed by microorganisms to carry out biodegradation processes. It is worth noting that PSMs can stimulate the expression of desirable genes present in environmental matrices if they bear a similar structural similarity to certain xenobiotics. Additionally, PSMs can be used as a primary source of carbon for bacterial communities to support their growth and increase their tolerance to higher concentrations of toxic compounds [21,30]. The presence of functional genes indicates that the microorganisms present in environment have the potential to degrade xenobiotics such as PHs.
PH degradative tfdA-like genes belong to the catabolic gene cluster tfdABCDFE, which is thought to be widely distributed in the environment among Proteobacteria. These genes have been extensively studied since the early use of 2,4-D-based herbicides in the 1950s. These tfdA-like genes have been used as markers for the presence of 2,4-D catabolism for two key reasons: they encode the dioxygenase enzyme, which plays a critical role in the initiation of PH biodegradation, and they are characterized by a unique nucleotide sequence and function [12]. Among the 2,4-D degraders, the bacteria harboring tfdA-like genes can be divided into four distinct groups, according to their base nucleotide sequence. The first group, the α-Proteobacteria, harbor the tfdAα gene; this group comprises bacteria closely related to Bradyrhizobium spp. [31]. The remaining three groups, harboring tfdA Class I, II and III, include various bacterial strains belonging to γand β-Proteobacteria [11,32]. Class I was isolated from contaminated sites and is known to be present in βand γ-Proteobacteria such as Cupriavidus pinatubonensis JMP134; in contrast, tfdA Class II is less widely distributed, being found in Burkholderia spp., while tfdA Class III was found in Comamonas acidovorans [33]. The location of the tfdA gene also differs between classes: tfdA Class I and III tfdA are located on self-transmissible plasmids [34], while tfdA Class II genes are located on bacterial chromosomes [12].
Therefore, the presence and type of tfdA-like genes can be used as indicators to identify the biodegradation potential of bacteria present in studied soil. Our findings confirm the presence of bacterial DNA, including the conservative 16S rRNA gene fragment, in all the samples amended with SE. PCR amplification revealed almost all selected functional gene fragments to be abundant in the studied samples; however, significant differences in gene profile were observed between samples amended with 2,4-D and MCPA at the two initial concentrations.
These functional genes were present in the sample spiked with SE (Table 3), which indicates that PH-degrading bacteria were already present in the studied soil. These results are consistent with those of our previous studies, where tfdA Class III and tfdAα genes were detected in control soil [35]; however, although the soil extract was prepared from the same soil material as in the previous studies, only the tfdA and tfdAα genes were detected [35]. It is important to underline that these two experiments were conducted in different media: soil and liquid MSM. The presence of functional genes suggests that the soil used for soil extract preparation might have been exposed to PHs for a longer time prior to collection. This observation is consistent with ecotoxicity results, where samples with the soil extract exhibited ecotoxicity levels as high as 57%.
In samples amended with 2,4-D, the initial concentration had a significant effect on the presence of the studied genes, with tfdA Class III genes being detected later at the higher dose of 2,4-D (0.5 mM), which may be due to the cytotoxic action of 2,4-D. In addition, higher herbicide concentrations were associated with lower numbers of the present functional genes. Following supplementation with a selected PSM (FA), the pattern of genes in the samples changed to tfdA Class I and tfdAα, indicating that FA influenced the structure of the microbial communities. The same amplicons were detected in the sample treated with SE + FA. FA has a strong effect on the microorganisms in the local environment and their degradative activities; however, its presence does not appear to lower the ecotoxicity effect of 2,4-D or its removal. Additionally, the enrichment of the sample with 0.5 mM 2,4-D and FA suppressed detection of the gene, suggesting that the two substances may have a synergistic negative effect on ecotoxicity.
Different results were obtained for samples amended with MCPA than those treated with 2,4-D. No PCR products were observed in samples treated with 0.1 mM of MCPA. This might be due to the presence of bacteria which possess other genes responsible for PH degradation. Previous studies have identified the presence of other RHDO genes in Bradyrhizobium sp., these being the cadA family [31,33]. These amplicons have also been detected in the environment (e.g., activated sludge) [12]. It is important to note that various fungi have also been found to metabolize PHs and to play the sole role in their biodegradation e.g., Streptomyces sp. (isolated from forest soil in Vietnam), Serratia marcescens and Penicillium sp. (isolated from contaminated soil in Brazil) [36]. Therefore, it is possible that the observed absence of bacterial functional genes might be associated with the action of fungi, indeed, the extract used for the experiment would have contained both bacteria and fungi; however, this was not investigated in the present study. Nevertheless, fungal strains are known to produce several non-specific enzymes which take part in the degradation of phenolic compounds [37].
Following the addition of SA, increased numbers of bacterial MCPA-degradative genes were detected. Samples spiked with both MCPA and MSM + SE + SA displayed twice the numbers of detected genes than those amended only with MCPA. Such an increase in tfdA-like gene numbers suggest that SA can enhance the biodegradation of MCPA, a substance with close structural similarity. As noted above, the samples amended with a higher dose of MCPA demonstrated higher numbers of degradative tdfA-like genes; this is also consistent with the results observed for the degradation and mitigation of ecotoxicity of 2,4-D ( Table 6). The presence of the tdfA-like genes in the soil indicates that of microorganisms capable of degrading MCPA.
These findings confirm those of Baelum et al. (2006) [38], who showed that bacteria harboring tfdA Class III genes were active during the degradation of MCPA; interestingly, they also recorded tfdA Class III and tfdAα in control soil, but not in the MCPA spiked soil samples [38], which is consistent with our previous findings [35]. Similar findings were observed by Poll et al. 2010 [32], who compared the abundance of tfdA and tfdAα in control and MCPA-amended samples. This indicates that the soil was probably enriched in a natural substrate enabling the growth of bacteria harboring the tfdAα gene, not necessarily phenoxy acids [31]. The presence of tfdA Class III amplicons in the SA-amended sample indicates that the addition of PSM influences the metabolic properties of the present microbial communities.
The present study used an assay based on two dicotyledonous plants (L. sativum and S. alba) to confirm the ecotoxic effect of studied PHs and their mitigation as a result of ongoing physicochemical and biological processes. The ecotoxicity of pure SE at the start of the experiment exceeded the limit of 20% for both L. sativum (51%) and S. alba (57%): as the soil used for the SE preparations was collected from an agricultural field, it is possible that it had been exposed to contaminants prior to collection. After incubation, it was found that although the % removal of 2,4-D in the studied cases ranged between 83-100%, the ecotoxic effect did not change over time, remaining above the toxicity limit of 20%. Of the two plants used in the assay, L. sativum was more sensitive to the treatment, displaying 100% root growth inhibition in all samples with 2,4-D, which is consistent with our previous study [35]. This ecotoxicity might have been caused by the formation of 2,4-D metabolites, as the catabolism of 2,4-D leads to the subsequent formation of 2,4-DCP and 3,5-DCC. These intermediates, formed during the degradation processes, have sometimes been found to be more toxic than the original compound [55]. 2,4-DCP has previously been observed to exhibit a strong cytotoxic effect and exert high affinity to the structures of plan cells [56].
The ecotoxicity of 2,4-D was not mitigated by addition of 0.1 or 0.5 mM FA. FA can be excreted to the root rhizosphere in various amounts under stress conditions e.g., when plants are exposed to contaminants that disturb their metabolism [57]. FA also has strong antimicrobial properties [58] and can inhibit seed germination, root and shoot growth and cell division, and exert a negative influence on the physiological parameters of the plants [59]. To conclude, FA was not found to exert a positive effect on ecotoxicity mitigation by increasing the rate of 2,4-D biodegradation, probably due to the formation of metabolites of 2,4-D exhibiting higher toxicity than the original compound. This suggests that both 2,4-D and FA had a toxic effect on the proliferation and catabolic activity of bacterial consortia in the amended samples.
The opposite situation was observed for MCPA: the physicochemical and microbial degradation processes not only reduced the ecotoxicity of the samples but even stimulated the root growth of the test plants. This finding suggests that soil microorganisms introduced to the samples play a major role in the process of herbicide detoxification. Furthermore, although both test plants responded similarly to the initial concentration of PH, higher root growth stimulation was observed for L. sativum after incubation. This is consistent with other studies, where S. alba has been found to be more sensitive to MCPA [35].
The above findings indicate that PSM application may have a positive influence on mitigating the ecotoxicity of the samples; however, this influence depends on the selected compounds and concentration of contaminants. In this case, at the higher initial concentration of MCPA used in the study (0.5 mM), higher stimulation of root growth was observed in samples treated with the PSM, i.e., SA (−55%) than those which were not (−47%).

Conclusions
The study presents an interdisciplinary approach to the problem of PH contamination in soil. It focuses not only on the removal rate of the herbicide, but also examines its influence on soil ecotoxicity and the potential for PH degradation using indigenous soil bacteria. The present findings reveal that measurements of removal percentage can be misleading when estimating ecotoxicity and biodegradation potential. In this case, selected PSMs had different effects on 2,4-D and MCPA biodegradation. Two critical factors were identified for the removal of contaminants and mitigation of the ecotoxicity: the choice of structurally-related phenoxy acid and PSM, and the initial concentration of phenoxy acid. Although the application of the chosen PSM (FA), contributed to the depletion of 2,4-D, a high ecotoxic effect was still observed at the end of incubation. In contrast, SA treatment enhanced the biodegradation of MCPA and encouraged the development of beneficial bacteria harboring a wide array of tfdAlike genes.
This study not only identifies changes in the presence of functional, bacterial degradative genes following PSM application, but also indicates that the two chosen xenobiotics, 2,4-D and MCPA, exert different effects on the abundance of degradative potential and ecotoxicity mitigation. The obtained molecular, instrumental and ecotoxicity assessment results demonstrate that the application of PSM can positively influence the removal of structurally-related herbicides; however, the final effect is highly selective and needs further, more elaborate investigation of the molecular mechanisms behind biostimulation processes.
Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4441/11/7/1451/s1, Table S1: Mean and Standard Deviation for Physical and Chemical Properties of Soil; * Significant Differences at α ≤ 0.05 According to the Mann-Whitney U-Test, Table S2: Target Genes, PCR Rimers and Their Optimal Annealing Temperature, Figure S1: The Alignment Analysis of 16S rRNA Gene Fragment (1300-1400 bp) Amplified in Samples Enriched with MCPA and Siringic Acid (Query Sequence) and Nucleotide Sequence of Rhodoferax Saidenbachensis OX00321, Figure S2

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

Abbreviations
The abbreviations with proper definitions used in the study: