Different response of perennial ryegrass—Epichloë endophyte symbiota to the elevated concentration of heavy metals in soil

The phenomenon of plant mutualistic symbiosis with microbes may have a positive effect on the improvement of plant tolerance to environmental stresses as well as on the ability of plants to accumulate heavy metal (HM) ions from soil. The influence of Epichloë fungal endophyte (Ascomycota, Clavicipitaceae) on perennial ryegrass (Lolium perenne L.) plants grown in the presence of elevated concentrations of HM ions (Cd2+, Pb2+, and Cu2+) in soil was studied. The presence of Epichloë in the host grass tissues resulted in different accumulation of HM ions in the aboveground parts of the plants. In some cases, endophyte infection positively affected ryegrass ability to accumulate HM ions from soil. In plants with (E +) and without (E −) endophytes, the hormesis effect was induced by the elevated concentration of Cu2+ ions, resulting in better growth and photosynthesis, as examined by measurements of Chl a fluorescence. The obtained results indicate that based on the laboratory evaluation of the efficiency of HM accumulation, we were able to choose the best associations of perennial ryegrass with endophytes for HM phytoremediation.


Introduction
Endophytes can colonize plant tissues and live without inducing any visible symptoms of biotic stress in plants. In general, as a consequence of host plant-microbe interactions, these endophytes produce a range of alkaloids and stimulate the host plant for enhanced synthesis of primary and secondary metabolites, e.g., free sugars, sugar alcohols, proline, glutamic acid, phospholipids, proteins, and polysaccharides (Avila et al. 2012;Bush et al. 1997;Nagabhyru et al. 2013;Porter 1994;Rasmussen et al. 2008, Soto-Bajas et al. 2016. Hao et al. (2010) observed that treatment of suspension cells of Ginkgo biloba with fungal endophytes resulted in the accumulation of flavonoids, increased abscisic acid (ABA) production, and activation of phenylalanine ammonia-lyase (PAL). Also, the root metabolism is altered in response to colonization of the aboveground parts of plants (Strehmel et al. 2016;Slaughter et al. 2018). Altogether, the mutual associations lead to changes in host plant gene expression and improve plant adaptations to environmental stresses, both biotic (e.g., insects, herbivore animals, diseases) and abiotic (e.g., drought) (Bacon et al. 2015;Dupont et al. 2015;Rodriguez et al. 2008;Schardl et al. 2012Schardl et al. , 2013. Inhibition of photosynthesis by heavy metals (HM) has been well documented (Clijsters and Van Assche 1985;Prasad and Strzałka 1999;Singh et al. 2011). HM stress induces a series of biochemical and physiological modifications in plant tissues that display common characteristics with those induced by drought. Membrane damage and altered enzyme activities lead to a wide range of secondary effects that concern practically all the physiological processes (Barceló and Poschenrieder 1990). Photosynthesis is a very sensitive process due to several structural and metabolic disturbances, like direct interactions of HM ions Communicated by: Izabela Pawłowicz with thiol, histidyl-and carboxyl-groups of cell proteins, induction of reactive oxygen species (ROS) formation, and displacement of essential cations in protein active centers (Hall 2002;Hossain et al. 2012;Farid et al. 2013). Some ions such as Hg 2+ , Cu 2+ , Cd 2+ , Ni 2+ , or Zn 2+ may substitute the central Mg 2+ ion in chlorophyll molecules, forming complexes lowering the quantum efficiency of PSII (Van Assche and Clijsters 1990;Sharma and Dietz 2009). These circumstances affect most of the parameters of chlorophyll a (Chl a) fluorescence detected by the JIP test (Żurek et al. 2014). However, it has been demonstrated that endophytes play a key role in host plant adaptation to polluted environments and that they can enhance phytoremediation by mobilizing/degrading or immobilizing contaminants in the soil, promoting plant growth, decreasing phytotoxicity and improving plants' HM ion tolerance (Soleimani et al. 2010;Li et al. 2012a, b;Li et al. 2016).
Species of the fungal genus Epichloë (Ascomycota, Clavicipitaceae) are specialized fungi of cool-season grasses that can grow throughout the aerial parts of their host plants, forming systemic and predominantly asymptomatic associations, resulting in defensive mutualism (Clay 1988;Tadych et al. 2014). The importance of Epichloë endophytes for ecosystems due to modulation of both below-and aboveground ecosystem processes is well recognized and accepted (Saikkonen et al. 2016).
Phytoremediation is increasingly used as a sustainable approach for soil remediation. However, methodology improvement is constantly forced due to the expected increase in phytoremediation efficacy as well as due to economic reasons. Due to complex biological interactions, currently used methods do not always give the demanded results, so further multidirectional studies are needed (Thijs et al. 2017).
The aim of this study was to describe the different reaction of perennial ryegrass-Epichloë endophyte associations to the elevated concentration of lead, cadmium, and copper in soil with further possible application in the phytoremediation process.

Plant collection
Ecotypes of perennial ryegrass (Lolium perenne L.) were collected from 12 localities in Poland in the form of living plants from permanent grasslands in most cases used for cattle feeding. The term "ecotype" refers to a group of plants within a species that is adapted to particular environmental conditions (locality) and therefore exhibiting structural or physiological differences from the other members of the same species. Those areas were located in Podlaskie (POD), Mazowieckie (MAZ), Lubelskie (LUB), and Świętokrzyskie (SWK) regions located on Central European Plain, in Poland. (Fig. 1, Tab. 1, Supp. Tab.1).
From five to ten plants were picked up in each locality, with the distances of 5-10 m from each other, to avoid collecting clones. GPS coordinates of each locality were recorded. Average concentration of Pb 2+ , Cu 2+ , and Cd 2+ Fig. 1 Map of the distribution of ecotype collection sites in regions of Poland: POD (Podlaskie), MAZ (Mazowieckie), LUB (Lubelskie), SWK (Świętokrzyskie). Ecotypes are identified by reference numbers the same across the whole manuscript. Map drawn with the DIVA-GIS ver. 7.1.7 software (http:///www. diva-gis. org) ions in soils for regions of collection sites were given as per Terelak (2007).
Since plants in the described experiment were part of a large collection, their numbering has no ordinal values. Collected ecotypes were further replanted in a spaced nursery, with 0.5 m distances between plants in Radzików,Poland (52.21 N;20.64 E). No additional treatments (fertilization, watering, chemical weed control) were applied.

Endophyte detection
Perennial ryegrass-Epichloë endophyte associations were identified by the rapid staining method according to Saha et al. (1988). Fungal hyphae (E +), appeared as an intercellular, long, and convoluted hyphae parallel to the leaf-sheath axis of the plant cell without forming haustorial structures (Clay and Holah 1999), (Supp. Fig. S1). On the bases of our recent discoveries and work done on perennial ryegrass endophytes (Wiewióra et al. 2015a(Wiewióra et al. , 2015b, fungus forming fungal hyphae inside intercellular spaces were described as belonging to genus Epichloë (Clavicipitaceae). Based on studies with confirmed isolates describing the typical fungal hyphae in the intercellular space of infected tissues from perennial ryegrass, and our own recent studies, we refer to the endophyte found in Polish perennial ryegrass as Epichloë festucae spp. lolii.
Twelve ecotypes with E + plants were selected as material for further studies. Seeds were collected only from E + plants of those ecotypes grown in the nursery and again tested for the presence of the endophyte hyphae using the rose bengal staining method (Saha et al. 1988). Half of the seeds from each ecotype was treated with Tebuconazole (placing the seeds in a liquid suspension), a triazole fungicide to remove the endophyte from seed bulk (E −). Both E + and E − seeds were sown on filter paper and seedlings were transferred to 0.5 l pots filled with mixture (1:2) of sterilized sand and peat. Seedlings were grown in pots for 4 weeks, with frequent watering and without additional fertilization.
The presence/absence of the endophyte hyphae was again confirmed on 3-4 weeks old seedlings by rose bengal staining before microscopic examination of 3 tillers per each plant. For each ecotype, 12 E − and 12 E + plants were vegetatively propagated: half of each set was intended for HM treatment and half remained as a control (no HM). As a result of the final round of vegetative propagation, 24 plants per ecotype E + and the same number per E − were used in the experiment run in fourfold repetitions per 3 plants each. Again, the endophyte status (E + /E −) was checked.
The pot experiment was arranged into a randomized complete block design with 4 blocks, where each ecotype was grown in 4 pots per block (3 plants per pot): two pots with E + plants and two pots with E − plants. From those four pots, two were treated with HM solution (see below) and two were control. Pots in blocks were re-arranged during the experiment to reduce the positional effect and reduce the residual or pot-to-pot variance. Therefore, two factors were used in the aforementioned experiment: the first, endophyte infection (E + and E − plants) and the second, HM treatment.
The experiment was run in a glasshouse, starting from late spring for 16 weeks in total, with the first 7 weeks of HM treatment. Seedlings were planted into pots, and after 3 weeks of growth in the glasshouse, the first watering was applied; then, watering was applied 9 times during the next 36 days of growth. Control pots were watered with distilled water. Intervals between watering usually were 4-5 days. The whole watering brought in total 20 mg of Cd 2+ and 700 mg of both Pb 2+ and Cu 2+ ions in 1 kg −1 of the used substrate. Finally, HM ion concentration in the substrate, as

Analysis of biomass yields, relative chlorophyll contents, and Chl a fluorescence parameters
Biometric phenotyping of the aboveground part of plants was done to determine the rate of plant growth. Three cuts of plants from all experimental pots were done after 1, 2, and 4 months of plant growth in pots since planting, followed by drying at 70 °C for 3 days for determination of dry matter yield. Dry biomass from each pot was collected to determine HM concentration in plants.
Chlorophyll content index (CCI) was measured with CCM200 Plus (PSI, Brno, Czech Republic), on 3 leaves per plant for a total of 24 plants of both forms E + and E − of each ecotype. The single result consisted of five single measurements per leaf.
Chlorophyll a (Chl a) fluorescence was measured using PocketPEA portable fluorimeter (Hansatech Instruments, King's Lynn, Norfolk, UK). Three measurements per plant (3 plants per ecotype per replication per variant) were done. Fluorescence was induced by saturating, red actinic light with energy of 3.500 μmol·m −2 ·s −1 . Measured and calculated parameters were used for the interpretation of endophyteplant interaction in the presence of HM ions (Paunov et al. 2018).

Chemical analysis
Determination of HM concentration in plants and soil were done as described previously (Żurek et al. 2014) by Regional Agrochemical Station in Warsaw (accredited laboratory acc. PN-EN ISO/IEC 17,025:2005). Plant material was washed with tap water and then with deionized water in an ultrasonic washer to remove all soil particles followed by drying at 70 °C for 3 days. Three hundred mg of dried, ground plant material was wet-washed using concentrated nitric acid (Merck) in a microwave system (MDS 2000, CEM, USA).
For determination of total HM ion (Cd 2+ , Pb 2+ , and Cu 2+ ) concentration in soil, extraction of air-dried soil samples was taken at the end of the experiment from each pot, ground to < 0.25 mm and extracted with concentrated perchloric (HClO 4 ) and fluoric (HF) acids. The amount of Cd 2+ , Pb 2+ , and Cu 2+ ions were measured using inductively coupled plasma spectrometry (ICP-AES, Spectro Analytical Instruments GmbH, Kleve, Germany).

Statistical analysis
All calculations were made with STATISTICA® 12 for Windows (StatSoft Inc. 2014). The significance of differences was accepted with a 95% probability. Two-way factorial ANOVA analysis was performed with ecotypes, presence of HM in soil, and endophyte presence in plants applied as main factors. Least significant differences (LSD) were calculated according to the Fisher test. T tests were performed at independent samples mode for HM ion contents in leaves of E + and E − . Principal component analysis (PCA) based on the correlation matrix algorithm was performed for all chlorophyll fluorescence traits measured and calculated for all ecotypes.

Plant collection sites
Most of the soil beneath meadows from which perennial ryegrass plants were derived, were of mineral or organic type, with medium or low soil moisture content, mainly with medium or low-intensity usage as pastures or for cutting (Supp.

Analysis of biomass yields, relative chlorophyll contents, and Chl a fluorescence parameters
Biomass yields were significantly affected by the ecotype and HM treatment throughout the whole experiment whereas the main effect of the endophyte was significant only for the first (after a month) and second cuts (after two months) ( Table 2).
Generally, for plants grown in the presence of HM ions, dry matter yields combined for three cuts were higher (3.1 g/ plants) than for control plants (1.3-1.5 g/plants) irrespective of endophyte presence in plants (Fig. 2, Supp. Figure 2). The yield of plants grown in the presence of HM, despite the presence of endophyte in plants, was 48% higher than control at 1st cut, 342% at 2nd, and 143% at 3rd cut on average. For the whole experiment, the total yield from HM treated plants was 115% higher than that of the control plants. This difference was statistically significant (p = 0.0000; F = 387.26).
Elevated concentrations of the HM in the soil, as well as the provenance of the tested ecotypes, were the main sources of variation for the relative chlorophyll content, expressed as CCI. In contrast, neither endophyte presence nor its interaction with the plant provenance and HM gave a significant effect on the CCI ( Table 2). The CCI in HM treated ecotypes was on average higher than in non-HM treated ones (Fig. 3) and above difference was also significant (p = 0.0000; F = 86.21).
Elevated concentration of the HM in the soil was also the main source of variation of Chl a fluorescence parameters: (Table 3, Supp. Figure 3).
Neither the ecotype nor endophyte status resulted in a significant effect of any of the abovementioned Chl a fluorescence parameters. However, a significant interaction between HM presence in soil and endophyte presence in plants has been calculated for and Area (Table 3, Fig. 4). For the parameters T FM , RC/ ABS, and PI ABS, none of the main sources of variation nor interactions were significant; therefore, they were not listed in Table 3 and Fig. 4.
(1-Vj)/Vj). F values were given and significance of the effects and interactions with prob-ability higher than 99.9% (***) or 95% (**). For T FM , RC/ABS, and PI ABS , F values for none of main effects or their interactions were significant; therefore, mentioned parameters were not listed below

Source of variation
Chl a fluorescence parameters Considering interactions presented in Fig. 4, perennial ryegrass plants, if grown without the addition of HM, exhibited some negative effects of endophyte presence in tissues, as reflected in lower values of F M , F V , and higher for Area. When HM was added to the soil medium, values of the mentioned parameters increased in the presence of endophytes. However, the value of the parameter reflecting the force of light reactions of PS II (F V /F M ) was significantly lower in the presence of HM in soil and endophytes in plant tissues. Therefore, whether E + plants score higher or lower values of mentioned Chl a parameters than E − plants depends on the addition of HM to the soil medium.    (Table 3, Supp. Figure 3). Interestingly, E + plants collected in more northern localities were characterized by a more visible decline of F V /F M and F V /F 0 ratios. And, as in the case of measured parameters, E + ecotype 730 reacted differently, by their slight increase. The ratio of F V /F 0 was ≤ 4.0 in E − plants, whereas in E + plants in 3 cases, the ratio exceeded 4 (ecotypes 45, 87, and 873). Parameter (1-Vj)/Vj, the measure of forward electron transport, seemed to be slightly affected by HM, especially in the leaves of E + plants.

E (+)
The PCA (principal component analysis) run on the bases of Chl a fluorescence parameters have shown the distribution of ecotypes depending on the endophyte presence mostly over the OX axis (first factor) (Fig. 5, Supp. Table 2) which means that most of the measured parameters, significantly correlated with the first factor (F 0 , F V , F M , and Area), influenced such grouping.
Ecotypes with endophytes, grown in soil without HM, were separated on the left side of the graph, as opposed to E + grown with the addition of HM. Negative values of factor 1, which is negatively correlated with F 0 , F V , F M , and Area, were ascribed to increased values of the mentioned Chl a parameters. On the right side of the OX axis, along with decreasing values of Chl a parameter, points representing E + plants grown with the addition of HM were located. This is another presentation of the interaction between HM and endophyte presence.

HM ion content in E + and E − ecotypes
Analysis of variance for the data of HM ion concentration in the plant tissue revealed a statistically significant influence of both: plant provenance and endophyte presence in the host plant as well as their interaction in case of Cd +2 and Cu +2 ion concentration (Table 4).
The highest concentration of HM ions (sum of Pb 2+ , Cd 2+ , and Cu 2+ ) was detected in the leaves of E + variant of ecotype 160 (102 mg·kg −1 ), whereas in the leaves of the E − plants, the concentration of HM was low (44 mg·kg −1 ) ( Table 5). Differences in the particular ion concentration of the abovementioned ecotype were as follows: almost twofold higher concentration of Pb 2+ and Cd 2+ ions and threefold of Cu 2+ in E + plants compared to E − .
The highest concentration of Pb 2+ (43.9 mg•kg −1 ) was detected in the E − plants of ecotype 50 from POD region, Fig. 5 The graphical presentation of PCA analysis results based on measurements of Chl a parameters for ryegrass-Epichloë symbionts, grown with or without HM supplementation. Each data point represents a different ecotype   (Table 5). In a manner similar to relations described above for Pb 2+ concentration, for all three ecotypes with relatively high Cd 2+ concentration in E + plants, the Cd 2+ ion concentration was significantly higher than the concentration values found in E − plants. Average concentration of Cd +2 in plants was similar between regions of ecotype provenance, and it ranged from 9.1 -11.8 for E − plants and 11.8 to 16.2 for E + plants. For three from 12 tested ecotypes there were no significant difference between Cd +2 concentration in E + and E − plants.
High copper concentration was found in aerial parts of E + ecotypes 160, 273, and 873 (47.9, 40.6, and 37.4 mg·kg −1 , respectively). All mentioned values were significantly higher than in leaves of corresponding E − plants. The average high concentration of Cu 2+ ions in E + plants (ca. 30 mg·kg −1 ) was noted for central and southern regions, i.e., LUB and SWK. But the highest concentration of Cu 2+ (34.7 mg·kg −1 ) was noted in E − plants of ecotype 50, from POD region, which was the northern most exposed.

Discussion
There is increasing evidence that interactions of plants and microbes (including endophytes) play a critical role in metal phytoextraction and metal-mining, as they mediate different physicochemical and biological activities to facilitate ecological performances of the host plant (Muehe et al. 2015). The results of our studies revealed considerable variation in terms of the grass-fungus association's ability to cope with elevated concentration of HM ions in the soil. Mentioned 'variation' should be ascribed to the natural variation between host (perennial ryegrass), fungus and to their interaction. Spatial variation of mutualistic interactions between a host organism (grass plant) and infecting fungus (endophyte) through its intensity (endophyte frequency per locality) and production of toxic metabolite, i.e., ergovaline, has been previously described (Żurek et al. 2013, 2017). Plants subjected to increased HM contents in soil were characterized by significantly higher values of the CCI and plant biomass-this could be explained based on soil fertility. The soil used for this experiment contained a low level of Cu 2+ ions (2.4 mg·kg −1 ) and high amounts of soil organic carbon (SOC), 13%. The natural content of Cu 2+ in soil was in the range of 15 to 40 mg·kg −1 in the 0 -20 cm soil horizon and concentration of Cu 2+ below 3.0 mg·kg −1 is usually defined as a deficit for grass species (Olszewska et al. 2008;Wyszkowska et al. 2013). In the presence of high organic matter content in the soil, the Cu 2+ deficit for plants is quite frequent. Moreover, monocotyledonous plants (e.g., grasses) are particularly sensitive to Cu 2+ deficit (Yamasaki et al. 2008). Unfortunately all these three facts together were met together in our experiment, therefore the addition of Cu to soil medium yielded better growth of HM treated plants, which was manifested in higher CCI values. The differences were not statistically significant for ecotypes collected from the southern region, except for 1 ecotype of E + and 2 ecotypes of E − which could be the result of adaptation to naturally occurring conditions of increased HM content in soil (Rodriguez et al. 2008).
Chl a fluorescence detection and parameters analyses (F O , F M , F V , F V /F M , F V /F O , RCB/ABS, Area, (1-Vj)/Vj, P I ) are simple and widely recognized methods to assess the stress influence on plants (Żurek et al. 2014;Kalaji et al. 2016). Among fluorescence parameters measured in our experiment, F O , F M , F V , as well as the F V /F M , F V /F O and (1-Vj)/Vj, were found to be significantly influenced by both HM ions addition and its interaction with endophyte status. As a reaction to stress, F O value mostly increases, which is interpreted as lower efficiency of energy transfer between chlorophyll antennas in PS II, and our data follow reports in the literature (Prasad and Strzałka 1999). Although the increase of F O was detected in the case of the majority of studied ecotypes, the E + ecotypes, compared to E − , were characterized by lower values of this parameter pointing to the positive influence of Epichloë in the host plants, as was shown in studies on host orchard grass as well (Rozpądek et al. 2015). The F M is decreasing in response to stresses due to the fact that not all electron acceptors in PS II can be reduced. Considering results obtained in our experiment, endophyte presence in plant tissues seems to induce stress to a plant, as reflected by a decrease of F M . The F V /F O ratio, also used for the detection of PS II destruction upon stress, can descend from values of 4-5 down to 1. According to the results obtained in this experiment, a slight but significant (< 5% on average) decrease of F V /F O ratio was detected, showing that the stress did not influence the photosynthetic machinery to a large extent (Kalaji and Łoboda 2010).
The parameter F V /F M is one of the most commonly used in the evaluation of plant physiological status on the bases of fluorescence characteristics. For most healthy plants, it oscillates between 0.80 and 0.83. In our experiment, it fluctuated in 0.78 and 0.81 ranges. Interestingly, HM ions induced a statistically important drop down in this parameter in E + ecotypes originated from northern latitudes.
The distribution of points on the PCA graph indicates that the presence of HM in soil increased stress for plants as reflected by the Chl a parameter describing the efficiency of PS II. Points representing the efficiency of PS II in the presence or absence of HM in soil for E + plants were separated over the OX axis. Considering the negative sign of correlation coefficients between factor 1, factor 2, and Chl a parameters, points on the left side of the OX axis (negative values of factor 1) represent the better status of plants than points on the right (positive values of factor 1). In the case of the absence of endophyte in host plants, there is also no clear separation of points representing the efficiency of PS II in the presence or absence of HM in soil.
Increased nutrient content due to endophyte presence was observed by many authors (Soto-Barajas et al. 2016;Malinowski et al. 2004;Zabalgogeazcoa et al. 2006). In contrast, an absence of endophyte effect for total N (Lewis et al. 1996) and Zn concentration (Monnet et al. 2005) was reported with a single perennial ryegrass genotype evaluated. In the current experiment, we have observed the whole range of possible reactions: from E + plants accumulating less HM than E − plants, through no effect, to increased accumulation of one, two, or three HM ions from the soil by E + plants. Detected differences resulted, probably not only from differences in the endophyte activities but also from strong interactions between the fungus and the host plant, which arose as a result of particular conditions in an origin site. In the current research, spatial aggregation of E + plants able to uptake relatively higher amounts of the HM from the soil has been found for Pb 2+ accumulation. Perennial ryegrass ecotypes collected from the SWK region (locations below the latitude 50.84 N) demonstrated the ability for accumulation of relatively higher concentration of Pb 2+ ions in E + plants than those from the other regions. It could be presumed that it is in line with the natural concentration of Pb 2+ in the soils from this region which was concentrated in average of 17.8 mg·kg −1 of soil compared to 9.4-10.2 mg·kg −1 of the soils from other sampling sites in our experiment (Table S1). Hesse et al. (2003Hesse et al. ( , 2004 concluded that plant-endophyte associations are adapted to their native habitats via natural selection. As we have mentioned before, the natural content of the HM, especially Pb 2+ ions, in soil was higher in the SWK region than in other regions. Probably symbiota of this origin used to accumulate more Pb 2+ than those coming from areas of low Pb 2+ concentration. This could be further hypothesized that the whole microbiome of plants that came from soils of high Pb 2+ concentration could be different from soils of low Pb 2+ concentration. The role of the microbiome on plant health and HM tolerance has been recently widely analyzed and discussed (Dongchu et al. 2019;Ikram et al. 2018;Seneviratne et al. 2017).
The presence of HM tolerant endophytes could improve metals uptake and accumulation in hosting plants (Li et al. 2012a, b). Endophyte colonization promoted Cd 2+ ion accumulation in tall fescue (Ren et al. 2011) and also improved Cd 2+ transport from the root to the shoot. Hesse et al. (2003Hesse et al. ( , 2004 have also found higher abundances of infected perennial ryegrass genotypes on dry sites compared to wet sites and this has been confirmed in our previous research (Żurek et al. 2013, 2017). An abundance of endophyte-infected perennial ryegrass plants was significantly and negatively correlated with annual as well as winter precipitation (multiannual averages, 1950-2000) at localities of their origin. Considering the habitat of symbiota origin, for example, Dobrindt et al. (2013) reported higher incidences of Neotyphodium lolii at sites of limestone bedrock. Therefore, conditions at the place of host plant origin (both climatic and soil) may influence its ability to cope with abiotic stress (drought, soil acidity, toxic metals in soil). Differences observed between the host plants appear to depend on the endophyte and the host life histories, as well as on fungal and plant genotypes, abiotic and biotic environmental conditions, and their interactions (Saikkonen et al. 2013). Specific genotypic combinations of both host and endophyte determine the morphology and physiology of endophyte colonized grasses, as well as regulates how selective pressure acts on them (Hill et al. 1996).

Conclusions
Tested associations (fungus + host) exerted wide variation in response to the presence of an elevated concentration of lead, cadmium, and copper in the soil. In some cases, the presence of Epichloë sp. in perennial ryegrass tissues resulted in the increase of accumulation of above mentioned heavy metals in aerial parts of the host plants. Generally, in the presence of endophyte mycelium, an increased accumulation of cadmium and copper was found, but not for lead.
The phyto-beneficial effect of endophytes was strongly dependent on specific host-fungus associations, which in turn could be the effect of the host plants, i.e., ecotype provenance. However, results obtained in the experiment described above are not sufficient to draw conclusions on the relationship between the provenance of symbiota and their ability to accumulate heavy metals from the soil.
To achieve the best result of the phytoremediation of heavy metals, the choice of the most effective perennial ryegrass-Epichloë symbiosis should be based on their laboratory evaluation.

Competing interests The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.