Gene expression and morphological responses of Lolium perenne L. exposed to cadmium (Cd2+) and mercury (Hg2+)

Soil contamination with heavy metals is a major problem worldwide, due to the increasing impact mainly caused by anthropogenic activities. This research evaluated the phytoremediation capacity of, Lolium perenne for heavy metals such as cadmium (Cd2+) and mercury (Hg2+), and the effects of these metals on morphology, biomass production, and the changes on gene expression. Seeds of L. perenne were exposed to six concentrations of Cd2+ and Hg2+ in the range of 0 to 25 mg L−1, and two mixtures of Cd2+–Hg2. The Non-Observed Effect Level (NOEL) was established with dose response curves and the expression of specific genes was evaluated applying a commercially available quantitative reverse transcription (RT-qPCR) assay. There was no significant effect when exposing the seeds to Hg2+, for Cd2+ the maximum concentration was established in 0.1 mg L−1, and for the two concentrations of mixtures, there was a negative effect. An increase of expression of genes that regulate antioxidant activity and stress was found when the plant was exposed to heavy metals. Given the high tolerance to metals analyzed that was reflected both, the development of the plant and in its molecular response, these results highlight that L. perenne is a plant with phytoremediator potential.

www.nature.com/scientificreports/ ated using liquid nitrogen and then processed using Agilent Plant RNA Isolation Mini Kit (Agilent Technologies, USA), according to the manufacturer's recommendations. Isolated RNA was quantified using Nanodrop and its integrity confirmed by a 1.5% agarose gel stained with GelRed (Biotium, Fremont, USA). The extracted RNA was treated with DNAse (Thermo scientific, Waltham, USA) and each sample was diluted to a concentration of 30 ng/μl for processing. This assay was performed by duplicate for each treatment concentrations and the negative control.
The genes evaluated were selected according to the previously reported sequences and their evidence in their participation in tolerance processes of the plant response to growth in contaminated soils 23 . These genes encode for proteins involved in detoxification routes: CAT gene coding for catalase, pAPX gene for enzymatic ascorbate peroxidase, GST for glutathione S-transferase protein, MTPB-1 coding for metal tolerance protein B. In stress-associated routes: HSP70 encoding for heat shock protein 70 and HSP90 encoding for heat shock protein 90. Those are involved in the carbohydrate metabolism routes: SDHA encodes a major catalytic subunit of succinate-ubiquinone oxidoreductase, and the Lhcb-1 gene encodes chlorophyll-binding protein a/b type 1.
The determination of its differential expression was carried out by quantitative PCR using the Brilliant III Ultra-Fast SYBR Green RT-qPCR Master Mix kit (Agilent, USA). For the PCR reaction 2 μl of RNA was added following the manufacturer's recommended protocol. The reactions were carried out by duplicate (n = 2) on the Mx3000P thermal cycler (Agilent, Santa Clara, USA), and if the obtained C T differed more than 5% the reaction was processed again. The temperature profile was 50 °C for 10 min, denaturation at 95 °C for 3 min, 40 ring cycles and extension of 95 °C for 20 s and 55 °C to 62 °C for other 20 s (Table 1). At the end of each run, the melting curves were produced to verify the specific amplification of the reaction, the temperature profile used for this curve was 95 °C for 1 min, 55 °C for 30 s and 95 °C for 40 s. Also, the specificity of the amplification was verified through 1.5% agarose gel stained with GelRed fluorescent (Biotium, Fremont, USA). The expression was quantified by the delta threshold cycle (C T ) method, using TBP-1 (which encodes for TATA-box-binding protein) as the constitutive gene (Table 1) taking into account a corrected efficiency for all reactions of 1.8 32,33 . Data analysis. For the germination and elongation data, the dose response curve was determined, and the toxicity threshold was calculated, defined as the maximum concentration which there is no significant effect (NOEL). In order to determine the ideality of the curve, the normality of the data was evaluated using the Anderson Darling test and subsequently, a parametric or non-parametric ANOVA was performed depending on the data obtained. Likewise, comparisons were made to assess the differences in germination and elongation between treatments with the Kruskal-Wallis or Dunn´s tests depending on the normality of the data. To evaluate the effect of the interaction between the exposure time and the metal concentration, a general linear model (GLM) was performed, where exposure time and metal concentration were included as factors. The software used for the statistical analysis was GraphPad Prism 5.0 and Minitab 2018.

Results
Preliminary determination of germination optimal time and seed viability. The cumulative germination rate of the L. perenne seeds is shown in Fig. 1. It follows a typical S or sigmoidal curve behavior. Likewise, it was found that the emergence of the radicular structure started on average on the fourth day, and it was stabilized on day seven with a percentage of 75% without statistically significant changes for the following two days of observation (P = 0.405). To determine the optimal time, the last day of exponential growth was consid-  -3′)  TCT TGG CGA TGA TGG GGT TGC  220  61  (3′-5′) GAC TCG AAG AAC GCC CTG GAG www.nature.com/scientificreports/ ered, this because until this day the growth of mature seeds is presented and with the necessary characteristics to continue growing even under stressful conditions. The curve found is similar to the one obtained by Bewley et al. 34 who argues that this type of curve is obtained when the seed sample is not uniform, which causes some seeds to complete germination relatively quickly and others require longer times. These cumulative germination curves are normally sigmoidal, typically occurring when dry, viable seeds begin water absorption and a series of events are triggered that end with successful germination [35][36][37] . The foregoing reflects that, like many of the biological properties, the germination time of the seeds presents approximately a normal distribution. Likewise, that the germination percentage has stabilized at a value less than 100%, and that after three days there have been no increases, implies that only a fraction of the selected sample was able to germinate under imbibition conditions, for which the remaining fraction of approximately 24% corresponded to unviable seeds.
Lolium perenne germination under Cd and Hg exposure. The germination rate was reported daily for nine days, taking two measurements per day (Fig. 2). For the concentrations of Cd 2+ (Fig. 2a) and Hg 2+ (Fig. 2b), a germination process like that of Fig. 1 was presented. In contrast, for the binary mixture (Cd 2+ -Hg 2+ ) a delay in the stabilization time was observed, as well as a decrease in the percentage of germination on the last day (Fig. 2c).
Although germination started on the same day for all treatments, the number of germinated seeds showed variations between the concentrations for the different metals. For the control seeds, a percentage of germination of 29 ± 2.51% was presented after the first day. In the case of seeds treated with each of the concentrations of the different metals, a decrease in the germination percentage was observed. According to the Kruskal Wallis test, a significant difference with a P-value < 0.05 was shown for Cd 2+ , Hg 2+ and Cd 2+ -Hg 2+ in comparison with the control.  This potential of heavy metals to decrease germination is due to the fact that germination depends mainly on seed reserves, especially starch, for the supply of metabolites necessary in processes such as respiration, as well as other anabolic reactions that they can be inhibited by metals 38,39 . Table 2 shows the effect of Cd 2+ , Hg 2+ and Cd 2+ -Hg 2+ treatments on different parameters associated with the germination of L. perenne seeds such as IGP, FGP, CVG, MGT, and TGI through analysis of variance. It was observed that for the individual metals there was only a significant difference in the IGP for Hg 2+ with a concentration of 0.05 mg/L and 5.00 mg/L in which there is an inhibition compared to the control. On the other hand, in the case of Cd 2+ -Hg 2+ , statistically significant changes were presented for all the parameters evaluated. This indicates that for the two-mixture concertation (Cd 2+ -Hg 2+ ), the inhibition occurred both in the initial and final germination percentage and in the emergency kinetics.
Treatments exposed to Cd 2+ did not present significant differences, which may be due that this metal prevents the absorption and movement of water by the seed 40 . However, this does not occur with exposure to Hg 2+ that presented significant differences and the results are similar to other studies 41 . These results may be due to the entry of contaminated solution for the seeds exposed to Hg 2+ and not for those exposed to Cd 2+ , which would indicate the inhibitory results for PGI in the case of Hg 2+ . Instead, in Cd 2+ -Hg 2+ , the most pronounced antagonistic effect was presented, very similar to that reported in other studies that associate it with the interaction between metals resulting from the formation of metal complexes with greater toxicity, or also with the influence it generates, one metal in binary mixtures in the adsorption of the other metal in the solution 38,42 . Effects in the root and stem elongation. Root and stem elongation of L. perenne at different concentrations of Cd 2+ , Hg 2+ and Cd 2+ -Hg 2+ were evaluated after the ninth day on which a percentage greater than 80% of germination was reached for all treatments. Regarding the statistical analyses, it was found that for the cases of Cd 2+ and Cd 2+ -Hg 2+ there is a statistically significant effect (P-value < 0.05) of the concentrations in the elongation of both the root and the stem. For the treatments with Hg 2+ , there was no statistical difference between concentrations, therefore, there was no inhibitory or favorable effect for the seeds treated with this metal (Fig. 3).
Regarding the comparison tests, it was found that the root exposed to Cd 2+ in concentrations of 0.25, 5 and 25 mg/L (q = 5.77, q = 4.28 and q = 8.85) presented a statistically significant difference. For the concentration of 0.25 mg/L a stimulation of 28.7% was found and for the two remaining concentrations it was observed and inhibitory effect. The root exposed to Cd 2+ -Hg 2+ presented statistically significant differences for the concentrations of 0.10 and 0.25 mg/L (q = 9.47 and q = 9.52). For the stem exposed to Cd 2+ the concentration with a statistically significant difference was 0.25 mg/L (q = 6.07) presenting a stimulation of 38.6%, and for the stem exposed to Cd 2+ -Hg 2+ , a statistically significant difference was found for the concentrations of 0.10 and 0.25 mg/L (q = 8.06 and q = 8.39) with an inhibitory effect in both concentrations. Table 2. Germination parameters of L. perenne exposed to Cd 2+ Hg 2+ and Cd 2+ -Hg 2+ . The * represents the levels of concentration statistically significant. www.nature.com/scientificreports/ The behavior observed in the case of Cd 2+ can be based on a hormetic process of dose response, where at low concentrations a stimulus occurs in the plant, and at higher concentrations an adverse effect occurs 43 . These percentages of stimulation found were similar to those reported in multiple toxicology databases 43 . Calabrese and Blain 44 indicated that the magnitude of the stimulatory response is characterized by being moderate, being in most cases between 30 and 60% above the control values.
Accordingly, the maximum level with no observable effect (NOEL) was that corresponding to the concentration of 0.1 mg/L for the early growth of the root and stem in the treatments with Cd 2+ . The importance of NOEL is because at this dose the stimulations in the metabolism of plant organs begin to appear, allowing a favorable effect on elongation to be achieved at higher doses. This stimulation in the metabolism could be characterized by overcompensation to an interruption in the homeostasis of the organism 45 . This overcompensation may be the result of incentive of cell proliferation, which could be related to the ability of Cd 2+ to replace the functional metal Zn 2+ , allowing the binding of multiple transcription factors to the regulatory regions of genes 46 . The negative effects at high concentrations presented in Cd 2+ for this study are very similar to those reported by other authors, because it is considered highly toxic 28,40,42,[47][48][49] , while at low concentrations positive effects similar to other studies were presented 45,46,50,51 . This could explain that plants develop oxidative stress when exposed to heavy metals that can cause cellular damage, due to the accumulation of metal ions that generate disturbances in cellular ionic homeostasis 52 .
Regarding the mixture Cd 2+ -Hg 2+ the inhibition percentages were between 37.8 and 53.3%, presenting the greatest negative effect for the concentration of 0.25 mg/L. In the treatments of both mixtures, an inhibitory response increased with increasing concentration and was greater for the root structure which can be attributed to the mobility of metals from roots to stem and direct contact with the root surface. This shows that in L. perenne as in other plants, heavy metals tend to be retained in the root tissues, therefore, the mixing effects are generally greater in the roots than in the stem 21,53 . On the other hand, in the binary mixture the most pronounced antagonistic effect was presented, which may be a consequence of the interaction between the metals that results in the formation of metal complexes with greater toxicity or also the influence that a metal generates in the binary mixtures in the absorption of the other metal in the solution 38,42 .
In the case of Hg 2+ , the absorption could produce serious damage to plants by affecting chlorophyll synthesis and reducing photosynthesis as a result of the replacement of Mg 2+ by Hg 2+ 54 . However, in the present study it was found that there was no inhibitory or stimulating response for any of the Hg 2+ concentrations 55 . This because the absence of an inhibitory effect has been proven on growth of young tissue, which is little affected by increasing concentrations of Hg 2+ . Similarly, it has been found that young plant tissue, which is still in the growth stage, has the possibility of internally diluting Hg 2+ concentrations by increasing biomass 56,57 . Changes in gene expression of L. perenne. The evaluation of the genetic expression was performed by RT-qPCR, nevertheless, the specificity of the amplification was corroborated by melting curve and agarose electrophoresis. For the relative quantification the constitutive gen selected was TBP-1, in order to corroborate the viability of this selection, the C T variation among the treatments was evaluated by analysis of variance using the general linear model finding not differences neither roots nor stems among the treatments (Fig. 4) (P value > 0.05 for the metals, plant structure, and concentrations), which indicates that this gene is suitable to be used as a gene for reference in relative quantification.  Fig. 5. For the detoxification response, the gen CAT (Fig. 5a) was upregulated in the stems for the three different treatments having the highest relative expression when the plant was contaminated with the Hg 2+ , in contrast to the root which has a moderated expression being even less than the reference gene in some treatments. This response is the result that the CAT activity has been associated with H 2 O 2 removal and antioxidant protection in peroxisomes and organelle glyoxysomes found in the entire plant, the enzyme catalase being one of the first to be expressed when oxidative degradation occurs 58 . In contrast, pAPX (Fig. 5c) was overexpressed in the roots exposed to Cd 2+ at the higher concentration, and for the stems and roots exposed to Hg 2+ but not for Cd 2+ + Hg 2+ . This result is similar to the ones obtain in Alfalfa and maize roots where pAPX responded by increasing activity at a moderate degree of toxicity 59 .
The GST gene (Fig. 5b) was downregulated in roots and stems exposed to the different treatments, especially in the different concentrations with Cd 2+ . However, it was evident that the treatments have an influence on the gene expression in comparison with the negative control. GST enzymes catalyze the conjugation between various xenobiotics with electrophilic centers, such as those produced by the metal ions of Cd 2+ and Hg 2+ , and the nucleophile glutathione (GSH), thus marking the xenobiotic for vacuolar sequestration 60,61 . The resulting GSH conjugates are generally less toxic and more soluble in water than the original xenobiotics 62,63 .
In the category stress-related genes, the higher expression levels were for the HSP70 (Fig. 5e), it was upregulated in stems especially exposed with Cd 2+ -Hg 2+ and Hg 2+ at the higher concentrations, however, the expression in the roots did not show a dose response behavior. Hsp70 proteins are crucial as molecular chaperones that inflict the catalysis necessary to rapidly recruit and release proteins 27,64 . This protein has been found to play an important role in the programmed cell death of the leaves of the species Aponogeton madagascariensis, this in order to achieve the elimination of cells compromised with stress. The Hsp90 (Fig. 5f) was only overexpressed in roots exposed with Cd +2 and the stems exposed with Cd 2+ -Hg 2+ and Hg 2+ . Hsp90 proteins are responsible for safeguarding the integrity of the cell membrane when the plant organism is exposed to heavy metals, resulting in reductions in the germination rate and root length 65 . Finally, for the category related to energy production, Lhcb-1 (Fig. 5h) was only expressed in the stem as expected, and it was upregulated in the three treatments, nonetheless, the response was lower for the Cd 2+ -Hg 2+ . The gens MTPB-1 for detoxification (Fig. 5d) and SDHA (Fig. 5g) energy production category were not express either for the control and the different treatments. Since, just two replicates of each treatment could be made, the statistical comparisons were not performed. However, the changes expression levels observed for the genes CAT , GST, pAPX, HSP70 and HSP90 (Fig. 5) indicated physiological mechanisms used by the plant to repair the damage conferring tolerance to grow in contaminated medium with heavy metals and even enhancing their growth under concentrations that are minimal to inhibitory, being the hormetic growth evident. These results are in accordance with previous works, where it has been found that L. perenne is a phytoremediator plant of this type of metals 21 due to its accumulative capacity, and with the evaluation of these genes it is shown that L. perenne can be used in soils for the removal of contaminants such as Hg and Cd.
The RT-qPCR analysis showed that the relative expression levels in the root and stem of L. perenne for the MTPB-1 and Lhcb-1 genes did not present statistical difference between the different treatments. While for the rest of the genes (CAT, pAPX, GST, HSP70 and HSP90) there was a difference in the relative expression for the concentration factor, however, for the metal exposure factor there is no statistically significant difference. In relation to the changes in the relative expression of each gene with respect to the different concentration levels, the following was observed: for the CAT gene there was a decrease in the root and stem for a concentration of 0.25 mg/L, the GST gene increased in the root for all concentrations and in the stem for a concentration of 0.10 mg/L, pAPX increased in the concentration of 0.10 mg/L in the stem, while HSP90 increased in the concentration of 0.10 mg/L in the root, and finally HSP70 increased in the concentration of 0.25 mg/L in the stem (Table 3).

Conclusions
The results of this study indicated that L. perenne is a plant with a significant phytoremediator capacity for heavy metals as Cd 2+ and Hg 2+ individually, since the effects at the morphological and molecular levels indicating the presence of tolerance mechanisms observed in the germination and elongation of the roots and stems. However, for the treatment Cd 2+ -Hg 2+ , the significant inhibition in the germination and elongation were observed and low levels of gen expression, suggesting that the phytoremediation potential may be affected if metal mixtures are found in the medium. Regarding the morphological response of the individual metals, it was found that germination presented an inhibition response relative to increasing concentration. Nevertheless, the plant grows, and an important germination percentage was obtained. The NOEL concentration for Cd 2+ was 0.10 mg/L, in this concentration the longitude increases in both roots and stems, and the change of the genetic expression was observed. This could indicate this level of exposure is where the molecular tolerance response occurs in which the plant seeks to counteract the stress caused by metals mainly performed by detoxification mechanisms, which would be an explanation for the increase in morphological response at the subsequent concentration of 0.25 mg/L. Even though the selection of concentrations occurred as a result of the dose response curve found for Cd 2+ , the molecular response was similar for this metal and for Hg 2+ . Indicating the detoxication molecular rout for this metal has a different morphological response.
Although the limited number of biological replicas did not allow made statistical comparisons among levels of gen expression, changes were observed, these results being a starting point for further investigation of the molecular mechanisms of phytoremediation plants, and their potential use.