•Few studies have investigated plant–plant interactions involving hyperaccumulator plants. Here, we investigated the effect of selenium (Se) hyperaccumulation on neighboring plants.
•Soil and litter Se concentrations were determined around the hyperaccumulators Astragalus bisulcatus and Stanleya pinnata and around the nonhyperaccumulators Medicago sativa and Helianthus pumilus. We also compared surrounding vegetative cover, species composition and Se concentration in two plant species (Artemisia ludoviciana and Symphyotrichum ericoides) growing either close to or far from Se hyperaccumulators. Then, Arabidopsis thaliana germination and growth were compared on soils collected next to the hyperaccumulators and the nonhyperaccumulators.
•Soil collected around hyperaccumulators contained more Se (up to 266 mg Se kg−1) than soil collected around nonhyperaccumulators. Vegetative ground cover was 10% lower around Se hyperaccumulators compared with nonhyperaccumulators. The Se concentration was higher in neighboring species A. ludoviciana and S. ericoides when growing close to, compared with far from, Se hyperaccumulators. A. thaliana showed reduced germination and growth, and higher Se accumulation, when grown on soil collected around Se hyperaccumulators compared with soil collected around nonaccumulators.
•In conclusion, Se hyperaccumulators may increase the surrounding soil Se concentration (phytoenrichment). The enhanced soil Se contents around hyperaccumulators can impair the growth of Se-sensitive plant species, pointing to a possible role of Se hyperaccumulation in elemental allelopathy.
The element selenium (Se) is an essential micronutrient for many organisms, including mammals. In these organisms Se is incorporated into essential selenoproteins, some of which have antioxidant functions and may help prevent a variety of cancers (Burke, 2002; Zhang et al., 2006c). Although Se is essential for the growth of some algae and has been shown to promote the growth of many higher plant species, there is no evidence that it is essential for higher plants (Novoselov et al., 2002; Zhang & Gladyshev, 2010). Selenium is toxic to most organisms at elevated concentrations, largely because of its similarity to sulfur (S), which leads to nonspecific replacement of S by Se in proteins (Stadtman, 1990).
Bioavailable Se in soil occurs primarily in the form of selenate (SeO42−) or selenite (SeO32−) (Kocot et al., 2003). Soil Se concentrations vary, and most soils contain between 0.01 and 2.0 mg kg−1 of Se; however, some seleniferous soils can have (total) Se concentrations of > 10 mg kg−1 (Zhu et al., 2009). Despite their apparent lack of a physiological requirement for Se, higher plants readily take up selenate or selenite and convert it into organic forms via S-assimilation mechanisms. Plants can also volatilize Se, in the forms of dimethylselenide or dimethyldiselenide, which have a pungent odor that helps to identify Se-rich plants (Terry et al., 2000). Some plants even actively accumulate Se to concentrations between 0.1% and 1.5% of DW, typically 100-fold higher than other species growing on the same site (Beath et al., 1939). Such species are called Se hyperaccumulators and are found exclusively on seleniferious soils (Beath et al., 1939). Species with intermediate concentrations of Se (between 100 and 1000 mg of Se kg−1 DW), are called secondary Se accumulators, and non-Se accumulators only have trace concentrations of Se when grown in seleniferous habitats (Terry et al., 2000; Hawrylak-Nowak, 2008). Selenium-hyperaccumulating species, such as Astragalus bisulcatus (Fabaceae) and Stanleya pinnata (Brassicaceae), have the ability to store and tolerate extremely high concentrations of Se because they sequester Se in specialized tissues in the form of methylselenocysteine (MeSeCys) (Neuhierl & Bock, 1996; Freeman et al., 2009). This form of Se is not incorporated into proteins and therefore is relatively nontoxic compared with selenate, the primary form of Se found in soils and in nonhyperaccumulating species (Neuhierl & Bock, 1996; de Souza et al., 1998).
Why do some plants hyperaccumulate the toxic and nonessential element Se? There is substantial evidence for the elemental defense hypothesis: Se can protect plants from a variety of herbivores and pathogens (Hanson et al., 2003, 2004; Freeman et al., 2007, 2009; Galeas et al., 2008). While this sheds some light on the possible functional significance of Se hyperaccumulation, alternative hypotheses have been proposed and may be explored (Boyd & Martens, 1992). One alternative hypothesis is that elemental hyperaccumulation may serve an allelopathic function. If hyperaccumulators concentrate the accumulated element in their surrounding soil, a phenomenon called phytoenrichment (Morris et al., 2006, 2009), this may be toxic to some of the neighboring plant species, resulting in less competition. Previous studies investigating the role of hyperaccumulation serving an allelopathic function have shown mixed results. Nickel (Ni) hyperaccumulators were shown to increase surrounding soil Ni concentration, but did not decrease neighboring plant germination (Zhang et al., 2005, 2007). An increased concentration of zinc (Zn) in media was shown to reduce the germination rates of a variety of species (Bottoms, 2001). However, the Zn concentrations in the media were much higher than the concentrations found in the field around Zn-accumulating plants, and therefore this cannot be considered as representative for elemental allelopathy. Morris et al. (2006) found that soil with elevated Zn concentrations collected from around Zn-rich Acroptilon repens did not decrease the germination rate of several species.
The effect of Se-hyperaccumulator plants on germination, growth and Se accumulation in neighboring plants has yet to be reported. It has been shown that the decomposition of Se hyperaccumulator leaf litter can increase soil Se concentration (Quinn et al., 2010). Furthermore, it can be envisioned that Se is released from hyperaccumulator plant roots as exudates or from root turnover, or leached from germinating seeds, further contributing to elevated concentrations of Se around hyperaccumulator plants, which may have an effect on neighboring plants. The objectives of this study were to determine how Se hyperaccumulators affect the surrounding soil Se concentration and, via soil accumulation, the germination, growth and Se accumulation of surrounding plant species, and, with that, the composition of the local plant community.
Materials and Methods
The field site for this study, Pine Ridge Natural Area, is located in South West Fort Collins, CO, USA (40°32.70N, 105°07.87W). Pine Ridge Natural Area is a seleniferous habitat with sandy loam of Cretaceous shale origin, dominated by forb and grass species. The soil has a pH of 7.6, and 11% soil organic matter. Some diethylene triamine pentaacetic acid (DTPA)-extractable nutrient concentrations were (in mg kg−1) as follows: nitrate, 5.1; phosphorous, 12.4; sulfate, 18.6; calcium (Ca), 291; magnesium (Mg), 129; iron (Fe), 12; manganese (Mn), 5.6; copper (Cu), 5.5; Zn, 3; and potassium (K), 425 (Quinn et al., 2010). At least two species of Se-hyperaccumulating plants –A. bisulcatus (two-grooved milkvetch) and S. pinnata (prince’s plume) – thrive at Pine Ridge Natural Area (Galeas et al., 2008). The populations of A. bisulcatus and S. pinnata at Pine Ridge Natural Area are known to accumulate high concentrations of Se, up to 10 000 mg of Se kg−1 for A. bisulcatus and > 6000 mg of Se kg−1 for S. pinnata (Freeman et al., 2006a; Galeas et al., 2007).
Effect of Se hyperaccumulation on soil Se distribution
To investigate the effect of Se hyperaccumulator plants on the distribution of soil Se concentration, soil samples were collected from around the Se hyperaccumulators A. bisulcatus (Fabaceae) and S. pinnata (Brassicaceae) and the nonhyperaccumulators Medicago sativa (alfalfa, Fabaceae) and Helianthus pumilus (little sunflower, Asteraceae) at Pine Ridge Natural Area. These control species have been used in earlier studies (Galeas et al., 2008) and were chosen because among the species on the site they showed the most similar growth habit (similar size, shape, age and flower color) and (in the case of the Fabaceae) relatedness. A. bisulcatus and M. sativa are 25–40 cm in radius and 50–70 cm tall; S. pinnata and Helianthus pumilus are 15–25 cm in diameter and 40–60 cm tall. A soil sample was collected from the top 2 cm of topsoil from c. seven individuals of each plant species directly next to the stem and at 10, 20 and 50 cm from the stem. In addition, soil samples were collected 0–5 cm from the taproot at depths of 0, 5, 10 and 30 cm. The rooting depth of all plants was at least 30 cm. For soil collection, first the spots were determined and labeled, and litter was removed. Then the soil samples from the different depths and distances were collected using a small shovel and a ruler, and placed in plastic bags. Soil samples were dried from the day of collection for 72 h at 50°C and were then sieved using mesh with 1-mm2 holes, which removed leaf litter material and larger arthropods. Soil samples were analyzed for elemental concentrations as described later in the Materials and Methods. In addition, young mature leaves and lateral roots were collected from each of the seven individuals of each of the four plant species, washed in distilled water, and then dried and analyzed for elemental composition, as also described later in the Materials and Methods.
Effect of Se hyperaccumulation on plant community
To investigate the effect of Se hyperaccumulation on the surrounding plant community, the percentage ground cover and plant species composition was determined around the same A. bisulcatus, S. pinnata, M. sativa and H. pumilus individuals described in the previous section (n =7). Ground cover was determined by placing two 0.1-m2 Daubenmire plots directly adjacent to the plant on opposite sides (East and West of the plant, with the edge of the plot touching the stem of the plant). The percentage ground cover for each plot was estimated as described by Daubenmire (1959) and the number of individuals from each plant species within each plot was counted. The percentage ground cover and species composition was then averaged between the two plots.
Effect of proximity to Se hyperaccumulators on neighboring plant Se concentration
Studies were conducted to determine if proximity to Se-hyperaccumulating plants affects the Se concentration in nonhyperaccumulating plant species. Young, mature leaves were sampled for Se concentration from the species Artemisia ludoviciana (white sage; Asteraceae) and Symphyotrichum ericoides (white heath aster; Asteraceae), either growing in close proximity (< 1 m) to the hyperaccumulator species A. bisulcatus or S. pinnata or far away (> 4 m) from any hyperaccumulator (n =3). Young, mature leaves from A. bisulcatus, S. pinnata, M. sativa and H. pumilus were also collected for elemental analysis. In addition, litter was collected from under the canopy of A. bisulcatus, S. pinnata, M. sativa and H. pumilus, as well as soil from 0 to 2 cm and from 2 to 5 cm depths. The litter was collected without disturbing the underlying soil, and was sieved as an extra precaution. The underlying soil was scraped off with a small shovel and collected in plastic bags. Using a ruler, two consecutive soil layers were collected, from 0–2 cm and 2–5 cm depths.
Effect of soil Se concentration on Arabidopsis thaliana germination and growth
Litter and soil were collected at Pine Ridge Natural Area from around the hyperaccumulators A. bisulcatus and S. pinnata (n =10), and from around the nonhyperaccumulators M. sativa and H. pumilus (n =5), and were analyzed for Se concentration as described later in the Materials and Methods. For each plant sampled (30 plants total), topsoil (0–2 cm), with an equal volume of leaf litter on top (1 cm of soil with 1 cm of litter), was placed in petri dishes and 50 Arabidopsis thaliana seeds were sown in each petri dish (n =3 per plant sampled). The germination rates for A. thaliana were recorded after 14 d. In a second experiment, 25 A. thaliana seeds were sown on soil taken from each plant at 2–5 cm depth, using 10 × 10 cm pots (n =3 for each plant sampled). Five weeks after germination, plants were analyzed for growth by determining biomass DW by harvesting and washing the whole plant, including roots, and drying at 50°C for 72 h. Shoot material from the dried A. thaliana plants collected from each pot was then analyzed for elemental concentrations, as described later in Materials and Methods.
Selenium-dependent A. thaliana germination was investigated using half-strength Murashige and Skoog (MS) basal salts agar medium (Murashige & Skoog, 1962) spiked with different concentrations of Se as sodium selenate. Germination rates were compared 6 d after sowing seeds on media with 0, 2.5, 5, 10, 25, 100, 250 and 1000 ppm Se, using three replicates of 25 seeds each per concentration.
The elemental composition of leaves, seeds, litter and soil, collected as described in previous sections of the Materials and Methods, was determined as described previously (Galeas et al., 2007). In short, the samples were dried at 50°C for 72 h and 100 mg DW of each sample was digested in nitric acid, as described by Zarcinas et al. (1987). Inductively coupled plasma atomic emission spectroscopy (ICP-AES) was used as described by Fassel (1978) to determine each digest’s elemental composition.
The software Jmp-in (version 3.2.6; SAS Institute, Cary, NC, USA) was used for statistical data analysis. A Student’s t-test was used to compare differences between two means. ANOVA, followed by a post-hoc Tukey Kramer test, was used when comparing multiple means. Correlation analysis and linear regression were used to correlate the A. thaliana germination rate with substrate Se concentration.
Effect of Se hyperaccumulators on soil Se distribution
A. bisulcatus and S. pinnata had higher concentrations of Se in leaves and roots than the nonhyperaccumulators, reaching contents exceeding 1000 mg kg−1 DW in both roots and shoots (Fig. 1a). Shoots and roots of the two nonhyperaccumulator species M. sativa and H. pumilus all contained < 120 mg of Se kg−1 (Fig. 1a). Sulfur, an element that is chemically similar to Se, was also present at higher concentrations in both Se-hyperaccumulator species than in nonhyperaccumulators (Table 1). Among the other nutrients tested, Mg and Mn were present at higher concentrations, and Cu and Fe at lower concentrations, in leaves of hyperaccumulators compared with nonhyperaccumulators (Table 1).
Table 1. Leaf and root tissue concentration (mg kg−1) of various nutrients in hyperaccumulator (Astragalus bisulcatus and Stanleya pinnata) and nonhyperaccumulator (Helianthus pumilus and Medicago sativa) species growing in the same seleniferous habitat
Data shown are mean ± SE.
Superscript letters denote statistically different means (ANOVA, α = 0.05).
Surface soil collected from beneath the canopy of hyperaccumulator species (0–20 cm) generally showed higher concentrations of Se than soil from around nonaccumulators (Fig. 1b). The most pronounced difference was found between soil collected 20 cm from A. bisulcatus, the species with the highest Se concentration, and soil collected 20 cm from M. sativa, a nonhyperaccumulator in the same family as A. bisulcatus (Fig. 1b). For A. bisulcatus the soil Se concentration next to the tap root was even higher below the soil surface. Soil collected right next to A. bisulcatus roots, at depths of 5, 10 and 30 cm, contained 71–103 mg Se kg−1, which was significantly more Se than in soil collected next to any of the other plant species, and three- to fivefold higher than on the A. bisulcatus soil surface (Fig. 1c).
Effect of Se hyperaccumulation on the neighboring plant community
To determine if the plant communities around Se hyperaccumulators were different from those around nonhyperaccumulators on the same site, we measured vegetative ground cover and species composition around the Se hyperaccumulators A. bisulcatus and S. pinnata and around the nonhyperaccumulators M. sativa and H. pumilus at Pine Ridge Natural Area. The percentage ground cover was c. 10% lower around hyperaccumulators than around nonaccumulators (Fig. 2a). When single species were compared, only S. pinnata and M. sativa differed significantly, but when combined, the two hyperaccumulators had significantly lower average surrounding vegetative cover than the two nonhyperaccumulators (Fig. 2a).
Average species richness was also slightly lower around the two hyperaccumulators compared with the nonaccumulators, but not significantly (Fig. 2b). Table 2 lists the plant species found in the neighboring vegetation. There were no significant differences between the four species with respect to neighboring species composition, with the exception that Bromus japonicus (field brome) occurred more frequently around H. pumilus than around the other three species, and Agropyron repens (quackgrass) occurred more frequently around A. bisulcatus than around M. sativa (Table 2). Other species that showed an interesting, but nonsignificant, trend were Chenopodium berlandieri (which was found more frequently around hyperaccumulators than around nonaccumulators) and Descurainia sp. (which was found less frequently around hyperaccumulators than around nonaccumulators).
Table 2. Plant species surrounding selenium (Se) hyperaccumulating and non-Se hyperaccumulating plants
Values show average number of plants per m2 ± SE.
90.0 ± 18.6a
52.1 ± 12.5ab
32.8 ± 17.8b
48.6 ± 15.8ab
1.4 ± 1.4
0.7 ± 0.7
2.1 ± 1.5
8.6 ± 7.0
2.1 ± 1.5
3.6 ± 3.6
8.6 ± 5.0
11.4 ± 4.5a
24.3 ± 6.7a
20.0 ± 5.5a
56.4 ± 12.2b
3.6 ± 2.4
7.1 ± 7.1
1.4 ± 1.4
2.9 ± 2.9
1.4 ± 1.4
10.0 ± 5.0
7.1 ± 2.6
4.3 ± 3.3
2.9 ± 2.9
4.3 ± 2.0
3.6 ± 2.1
3.6 ± 3.6
2.8 ± 2.1
0.7 ± 0.7
Needle and thread grass
2.1 ± 2.1
6.4 ± 3.7
6.4 ± 3.6
14.3 ± 4.7
0.7 ± 0.7
2.9 ± 2.1
1.4 ± 0.9
0.7 ± 0.7
2.1 ± 1.5
3.6 ± 0.9
1.4 ± 0.9
The leaf Se concentrations of two species, A. ludoviciana (white sagebrush) and S. ericoides (white heath aster), were higher when the plants were growing next to a hyperaccumulator species (A. bisulcatus or S. pinnata) than when they were growing away from hyperaccumulators; the difference was six- to sevenfold for S. ericoides and two- to sixfold for A. ludovicianum (Fig. 3). Interestingly, the Se concentration in A. ludoviciana actually reached hyperaccumulator contents (> 1000 mg kg−1 DW) when it was growing next to a hyperaccumulator species (1400–4600 mg kg−1 DW), but not when it was growing away from them (700 mg kg−1 DW). To our knowledge neither S. ericoides nor A. ludovicianum has previously been reported to be a hyperaccumulator. Despite their elevated tissue Se contents, neither of the two neighboring species showed any visible signs of Se toxicity, such as chlorosis or stunted growth (results not shown).
Effects of Se hyperaccumulators on germination and growth of Se-sensitive neighboring plants
In order to determine whether Se hyperaccumulator plants may have a negative effect on Se-sensitive neighbors owing to their apparent ability to concentrate Se in their surrounding soil, litter and soil were collected from around A. bisulcatus, S. pinnata, M. sativa and H. pumilus, for controlled growth experiments. Leaves and seeds of the corresponding plants were also collected, and together with litter and soil, analyzed for elemental composition.
The Se concentrations in leaves and seeds of A. bisulcatus and S. pinnata were higher than those in M. sativa and H. pumilus, as expected (Fig. 4a). Hyperaccumulators had a higher Se concentration in seeds compared with leaves, whereas nonhyperaccumulators did not show such a pattern. Similarly to Se, tissue S concentrations were higher in seeds and leaves of hyperaccumulators than in nonaccumulators (Fig. 4b). The hyperaccumulators had a higher S concentration in leaves compared with seeds, and thus showed an opposite partitioning for S compared with Se; the nonhyperaccumulators did not show a difference in the concentration of S between these organs. Leaf Se and S concentrations were not significantly correlated in hyperaccumulators, but were correlated in nonhyperaccumulators. Table 3 shows the concentrations of some other nutrients in the leaves and seeds. Leaf Mg and Mn concentrations were again higher in Se hyperaccumulators than in nonaccumulators (Table 3). As for seeds, Fe concentrations were two- to five times lower in hyperaccumulators than in nonaccumulators (Table 3).
Table 3. Leaf and seed tissue concentrations (mg kg−1 DW) of various plant nutrients in hyperaccumulator (Astragalus bisulcatus and Stanleya pinnata) and nonhyperaccumulator (Helianthus pumilus and Medicago sativa) species growing side by side in a seleniferous habitat (Fort Collins, CO, USA)
Shown are mean and standard error of the mean. Superscript letters denote statistically different means (ANOVA, α = 0.05).
Selenium was present at similar concentrations in freshly collected leaves and decaying plant litter (Fig. 5a,b), with hyperaccumulators – particularly A. bisulcatus – having higher Se concentrations than nonaccumulators. The same trend was found for soil taken from 0 to 2 cm and 2–5 cm depths: soil from around hyperaccumulators, particularly A. bisulcatus, generally had a higher Se concentration than soil from around nonhyperaccumulator species. As shown in Table 4, the hyperaccumulator species showed positive and highly significant correlations between their Se concentration in fresh leaves, leaf litter and surrounding soil. For nonhyperaccumulator species the only significant correlation was found between leaf litter and soil at 2–5 cm depth. In contrast to Se, the S concentrations in litter and soil from around hyperaccumulators were not elevated when compared with nonhyperaccumulators (Supporting Information Fig. S1).
Table 4. Correlation coefficient (R) and statistical significance (P) of a correlation analysis between selenium (Se) concentration in leaf, litter and soil at 0–2 cm and 2–5 cm depths
Soil 0–2 cm
Statistically significant correlations (P < 0.05) are shown in bold. All correlations were positive.
Hyperaccumulator species (Astragalus bisulcatus and Stanleya pinnata)
R =0.61/P <0.01
R =0.20/P =0.39
Soil 0–2 cm
R =0.39/P <0.01
R =0.64/P <0.01
Soil 2–5 cm
R =0.56/P <0.01
R =0.39/P <0.01
Nonhyperaccumulator species (Medicago sativa and Helianthus pumilus)
R =0.11/P = 0.76
R =0.39/P =0.86
Soil 0–2 cm
R =0.50/P = 0.14
R =0.05/P =0.99
Soil 2–5 cm
R =0.07/P = 0.87
R =0.60/P <0.01
All four species
R =0.70/P <0.0001
R =0.31/P =0.09
Soil 0–2 cm
R =0.48/P <0.01
R =0.66/P <0.0001
Soil 2–5 cm
R =0.65/P <0.0001
R =0.50/P <0.0001
The litter and soil taken from around the hyperaccumulators (A. bisulcatus and S. pinnata) and nonaccumulators (M. sativa and H. pumilus) was used for a comparative germination and growth study using Arabidopsis thaliana accession Landsberg erecta (Ler). This is not a native neighbor of Se hyperaccumulators, but was chosen because it was shown earlier to be sensitive to selenate (Zhang et al., 2006a,b). In a first test, the soil from 0 to 2 cm depth was placed in petri dishes and covered with litter collected from around the same plant. A. thaliana seeds (50 per petri dish) were sown on this recreated topsoil-litter medium and the plates were cultivated in the laboratory. The germination rates of A. thaliana were significantly higher on soil and litter collected from around nonhyperaccumulator species compared with soil and litter collected from around hyperaccumulator species (Fig. 6a). There was a significant, negative relationship between average Se concentration in soil and litter and the germination rate of A. thaliana across the four species (Fig. 6b).
A second experiment was carried out to determine plant growth as well as Se accumulation in A. thaliana plants sown on soil collected from around hyperaccumulators or nonaccumulators. The A. thaliana seeds were sown in pots (25 per pot) on soil collected at 2–5 cm depth around each of the four species. Fewer A. thaliana seeds germinated when sown on soil from around hyperaccumulator species compared with nonhyperaccumulator soil (Fig. 7a); there was a negative correlation between soil Se concentration and A. thaliana germination (Fig. 7b). From visual observation, the plants that germinated on soil from around hyperaccumulators appeared to be substantially smaller compared with those growing on soil collected from around nonhyperaccumulators (Fig. 7c). While the average DW per plant was twofold lower for plants growing on A. bisulcatus soil and 30-fold lower for plants growing on S. pinnata soil compared with nonaccumulator soils (Fig. 7d), because of the high standard deviation these averages were not significantly different. The concentrations of Se were higher in A. thaliana growing on soil collected from around the hyperaccumulator species A. bisulcatus compared with plants growing on soil taken next to the nonhyperaccumulator species M. sativa and H. pumilus (Fig. 8a). The elemental concentrations could not be determined for A. thaliana growing on S. pinnata soil because none of the sown plants survived beyond the seedling stage. The concentrations of S and, to a lesser extent, those of Mg and Mn, were also elevated in A. thaliana growing on A. bisulcatus soil compared with soil taken next to nonhyperaccumulator species (Fig. 8).
To further investigate whether the Se concentration in the litter and soil used in these studies (ranging from 100 to 2000 mg kg−1 DW) was sufficient to inhibit A. thaliana germination, a controlled experiment was carried out in which seed germination was determined as a function of selenate concentration in agar medium. The germination rate decreased above a Se concentration of c. 5 mg kg−1 DW (Fig. 9). The 50% germination inhibition point was c. 10 mg kg−1 DW (125 μM sodium selenate), an order of magnitude lower than the lowest Se concentrations in the soil and litter collected from around hyperaccumulators.
In this paper we report that soil around hyperaccumulators in the field is enriched with Se, which results in enhanced Se accumulation in neighboring plants, to concentrations that are potentially phytotoxic. This may suggest a role for Se in elemental allelopathy: hyperaccumulators may use Se to keep Se-sensitive neighbors at a distance. This finding is of significance because it sheds new light on the functional significance of elemental hyperaccumulation. Thus, in addition to the previously reported benefits of elemental hyperaccumulation as a protectant against herbivores and pathogens, it may reduce competition with neighboring plants. All of these benefits may have contributed as selection pressures during the evolution of hyperaccumulation.
Soil and litter around Se hyperaccumulators was enriched in Se by generally two- to threefold for soil and by six- to sevenfold for litter, compared with soil and litter around nonhyperaccumulators growing on the same site. The hyperaccumulator plants contained, on average, 20-fold higher Se concentrations than the nonaccumulators. The higher Se concentrations in the soil around hyperaccumulators may be a result of phytoenrichment (deposition over time by the hyperaccumulator, from litter decomposition or root release), or simply because soil Se distribution is heterogeneous and the hyperaccumulators are indicators of the high-Se patches. Based on our sampling we cannot distinguish between the two. However, there is some circumstantial evidence that hyperaccumulator plants concentrate certain elements in their surrounding soil. The hyperaccumulators were shown to have higher tissue concentrations than nonaccumulators of not only Se but also of S, Mg and Mn. Similarly, in A. thaliana grown on soil collected from around hyperaccumulators, the concentrations of Se, S, Mn and Mg all tended to be elevated compared with A. thaliana growing on nonaccumulator soil. Moreover, the decomposing hyperaccumulator litter collected on top of the surrounding soil was very high in Se (600–2000 mg kg−1 DW), and its decomposition is likely to enrich the soil underneath over time, as was found recently in a litter-decomposition study (Quinn et al., 2010). Both S. pinnata and A. bisulcatus are perennial species, so the degree of phytoenrichment may increase with time.
If hyperaccumulator plants indeed phytoenrich their surrounding soil with Se, this may be caused not only by litter deposition but also by root turnover and exudation. In each of these processes the Se deposition is probably in the form of organic Se because hyperaccumulators accumulate Se as methyl-selenocysteine (Freeman et al., 2006b, 2010). This organic Se deposited by hyperaccumulators may be more readily taken up by neighbors than selenate, the predominant form of bioavailable Se in bulk soil, based on earlier uptake studies with different Se species (Zayed et al., 1998). In addition to enriching total Se in their surrounding soil via deposition of litter and root-released compounds, it is feasible that hyperaccumulators mobilize nonlabile pools of soil Se via special exudates, further increasing the amount of bioavailable Se for neighboring plants. Changing the soil Se into more bioavailable forms could either increase or decrease Se toxicity to neighbors. More bioavailable soil Se will lead to higher Se uptake in neighbors, which may increase toxicity. On the other hand, it can be envisioned that the form of bioavailable Se around hyperaccumulators is less toxic, for example because it is less prone, inside cells, to interfere with sensitive biochemical targets. Neighbors of hyperaccumulators contained elevated amounts of Se compared to when they were growing far away from hyperaccumulators, yet they showed no sign of toxicity. However, the Se in hyperaccumulator-derived soil appeared to be toxic to A. thaliana. It will be interesting for future studies to compare total and bioavailable Se contents in rhizosphere soil of hyperaccumulators and bulk soil, and to compare the forms of Se in neighboring plants with those in the same species growing far away from hyperaccumulators.
If hyperaccumulators affect the Se concentration and/or form of Se in their surrounding soil, and consequently in their neighbors, this may have a positive or a negative effect on those neighbors. Higher or lower concentrations of Se may affect plant physiology as well as the plant’s ecological interactions. Our findings indicate that in the field there may be some negative effects on neighboring species because the percentage ground cover was c. 10% lower around hyperaccumulators. The lower percentage ground cover around hyperaccumulators may mean reduced competition for hyperaccumulators as well as for Se-tolerant neighboring species, and selection against Se-sensitive neighbors. As a model Se-sensitive plant, the A. thaliana accession Ler indeed showed significant reduction in germination and growth on soil collected next to hyperaccumulators. In the field we did not see any evidence of toxicity, such as chlorosis or stunted growth, even though the two plant species tested showed four- to sevenfold higher concentrations of Se when growing next to hyperaccumulators compared to when growing away from hyperaccumulators in the field. As the composition of neighboring species in the field was different around hyperaccumulators and nonaccumulators, and some neighboring species appeared to thrive next to hyperaccumulators, it will be interesting for future studies to study the individual responses of different neighboring species. Particularly interesting for further studies are the potential beneficial ecological effects of Se accumulation in neighbors, such as protection from herbivores or pathogens.
It cannot be excluded that a factor other than Se was responsible for the apparent toxicity of the soil and litter collected around hyperaccumulator plants. For the Brassicaceae species S. pinnata, for instance, it can be envisioned that S-rich glucosinolates in leaf litter may be toxic to neighboring plants. However, litter and soil around hyperaccumulators was not elevated in S, only in Se. Moreover, based on the agar experiment, the Se concentration found in soil and litter around hyperaccumulators is high enough to impair A. thaliana seed germination for this accession. Even if only 10% of the Se in the litter and soil was bioavailable, this would still be expected to affect germination. Also, the toxic effect was found for soil collected from two hyperaccumulator species from different families (Brassicaeae and Fabaceae), and not for soil collected from two nonaccumulators from two different families, making it less likely that the two hyperaccumulators both exuded a toxic compound other than Se while the other two species did not. Also, the concentration of Se in plants growing on hyperaccumulator soil was elevated, both in the field and in the pot experiment, making it more feasible that Se was indeed the toxic compound.
Hyperaccumulators showed preferential Se accumulation in seeds vs leaves, which contrasted their S-partitioning pattern. Non-hyperaccumulators, on the other hand, showed similar Se- and S-partitioning patterns, and accumulated both elements to similar concentrations in leaves and seeds. This suggests that hyperaccumulators may have different transporters for Se and S that are regulated differently, while nonaccumulators cannot distinguish between Se and S. The presence of Se-specific transporters is also suggested by the higher Se : S ratio typically observed for hyperaccumulators (Feist & Parker, 2001; White et al., 2007). Furthermore, in our study, hyperaccumulators generally had higher concentrations of S in roots, leaves and seeds than nonaccumulators. We also observed this in an earlier study (Galeas et al., 2007), as did White et al. (2007). The higher concentrations of S in Se hyperaccumulators suggest that they may have upregulated mechanisms of S uptake and translocation. This was indeed recently found in a transcriptomic and biochemical study that compared the hyperaccumulator S. pinnata with the nonhyperaccumulator Stanleya albescens (Freeman et al., 2010). Another interesting finding was that hyperaccumulators tended to have higher concentrations of Mn and Mg. At this point we do not have an explanation for this phenomenon, but it is feasible that Mn and Mg uptake or translocation is facilitated by elevated plant S concentrations.
This study is the first to provide insight into the effects of Se hyperaccumulators on soil Se distribution and plant–plant ecological interactions. This information is interesting, not only from a fundamental scientific perspective, but also because Se is both a micronutrient and an environmental pollutant, and plants are increasingly used both as Se-fortified foods and for phytoremediation of excess Se (Banuelos & Bradley, 2010). Better insight into the effects of hyperaccumulator plants on Se accumulation and speciation in their neighbors may be useful for the further development of these applications. For instance, if Se hyperaccumulators enhance Se accumulation in neighboring crop species and perhaps also facilitate accumulation of more organic, anticarcinogenic Se in these neighbors, this would be very applicable for the development of efficient co-cropping practices.
Funding for these studies was provided by the National Science Foundation grant # IOS-0817748 to Elizabeth A. H. Pilon-Smits and a graduate fellowship from the Libyan government to Ali El-Mehdawi.