Within and between population variation for zinc and nickel accumulation in two species of Thlaspi (Brassicaceae)


Author for correspondence: Stacy I Taylor Tel: +44 (0)1392 263786 Fax: +44 (0)1392 273700 Email: S.I.Taylor@exeter.ac.uk


  • • In this study, the differences in zinc (Zn) and nickel (Ni) hyperaccumulation were investigated between three populations of Thlaspi pindicum together with genetic variation within populations of T. pindicum and Thlaspi alpinum var. sylvium, both serpentine endemics.
  • • Three experiments were conducted under standard conditions in hydroponic assay. Each experiment contained three treatments of metal: 100 µm Zn, 100 µm Ni, and combined 100/100 µm Zn/Ni. Genetic variation within populations was determined using maternal families.
  • • No genetic variation within populations was found for either Zn or Ni hyperaccumulation for both T. pindicum and T. alpinum var. sylvium, but differences were observed for both Zn and Ni hyperaccumulation between populations of T. pindicum. In combined Zn/Ni treatments, Zn inhibited Ni translocation in both species, which is unexpected considering that these species are serpentine endemics and well known Ni hyperaccumulators.
  • • The lack of genetic variation for metal hyperaccumulation is possibly due to inbreeding. Since Zn hyperaccumulation is not manifested in the field, inadvertent uptake of Zn is a plausible hypothesis for its preferential uptake.


Heavy metal hyperaccumulation is an intriguing natural occurrence where plants can accumulate extraordinary quantities of metals into their shoots and leaves (Reeves & Brooks, 1983; Brooks, 1987), while storing lower concentrations in their roots, typically having root : shoot ratios < 1 (Baker, 1981; Macnair, 2003). Plants are generally considered to be hyperaccumulators at leaf concentrations of 1000 mg kg−1 for nickel (Ni), copper (Cu) and cobalt (Co), 10 000 mg kg−1 for Zn and Mn and 100 mg kg−1 for cadmium (Cd) (Baker et al., 1994; Brooks, 1998). There are approx. 400 species known to hyperaccumulate one metal or another, most of which are Ni hyperaccumulators, all of which occur on serpentine (Brooks, 1987). Serpentine floras and hyperaccumulators exist in several countries, such as New Caledonia, New Zealand, Albania, Greece, Italy and Cuba, and in California in the western USA (Brooks, 1987). A typical characteristic of serpentine soils is the abundance of potentially toxic heavy metals, particularly Ni and chromium. High magnesium (Mg) : calcium (Ca) ratios are also distinctive for these substrates and have been shown to be a factor in preadaptation to serpentine from granite (Taylor & Levy, 2002). Selection pressures associated with high amounts of Ni and other metals can result in the evolution of species and ecotypes that can safely accumulate and tolerate these metals (Kruckeberg, 1967).

Both Zn and Ni hyperaccumulation are more commonly studied than other metals, and some serpentine-adapted species are capable of hyperaccumulating both metals (Reeves & Brooks, 1983). Macnair (2003) suggests that metal hyperaccumulation could result from an enhanced efflux from the roots into the xylem. This suggestion corresponds to the results found by Lasat et al. (1998) who discovered that Zn had a five times greater concentration in the xylem of Thlaspi caerulescens compared with that of the nonaccumulating Thlaspi arvense. In comparison, Krämer et al. (1996) found a strong correlation between histidine and Ni accumulation where histidine concentrations were higher in the xylem of the Ni hyperaccumulator, Alyssum lesbiacum, than in the xylem of related nonaccumulators. Nickel and Zn are possibly transported by similar ligands and thus competition for binding sites is likely to be a factor in the relationship between uptake of Zn and Ni when levels of both are high in the soil (Assunção, 2003).

In addition to the interest in the physiological aspects of hyperaccumulation, such as metal uptake and transport (Krämer et al., 1996; Lasat et al., 1998; Persans & Salt, 2000; Assunção et al., 2001), there has been a considerable focus on the prospect of using hyperaccumulators for the phytoextraction of metals from contaminated land (Brooks, 1998; Salt et al., 1998; Escarréet al., 2000). In the exploration of phytoremediation and the potential of hyperaccumulators, it is important to know the extent of genetic variation within and between populations (Pollard et al., 2002).

The genetic variation between populations in their ability to hyperaccumulate Zn, Ni and Cd has been investigated in several studies using T. caerulescens populations from uncontaminated soils, as well as those from serpentine and calamine ones (Pollard & Baker, 1996; Escarréet al., 2000; Pollard et al., 2002; Assunção et al., 2003a). The results of these studies indicated a high level of variation for Zn, Ni and Cd accumulation between populations. A handful of studies have incorporated sib-analyses (Pollard & Baker, 1996; Meerts & van Isacker, 1997; Escarréet al., 2000; Macnair, 2002; Pollard et al., 2002) to study the quantitative genetics of the accumulation trait. All of these studies showed a degree of within-population genetic variation for Zn accumulation as well as variation for Ni (Pollard et al., 2002) and Cd (Escarréet al., 2000) accumulation within populations. Hitherto, this research strategy has been limited to two species, T. caerulescens and Arabidopsis halleri, both bodenvag species, Zn hyperaccumulators found growing on and off metalliferous soils. Thlaspi caerulescens, the focus of most of these studies, is considered to be a model hyperaccumulator (Assunção et al., 2003b). Many other Thlaspi species are also known to accumulate high levels of Ni and/or Zn (Reeves & Brooks, 1983; Reeves, 1992), yet these species have been far less studied despite the encouragement by Reeves & Brooks (1983).

Thlaspi caerulescens can hyperaccumulate Zn on several soil types, Ni on serpentine and Cd on Cd-rich soils, and there appears to be a diversity of transporter systems for different metals (Assunção, 2003; Assunção et al., 2003a). Assunção (2003) postulated that transporters involved with Ni hyperaccumulation are less metal-specific and are also responsible for low-affinity Zn and Cd accumulation in T. caerulescens populations from normal soils, serpentine (Ni-rich) and Cd-rich soils. She also suggested a high-affinity transporter system with high Zn preference in populations from different soil types: calamine (Zn/Pb-rich), nonmetalliferous (normal), Cd-rich, and Ni-rich (serpentine) (Assunção, 2003). Zinc hyperaccumulation has been considered a constitutive trait in T. caerulescens (Pollard et al., 2002; Assunção et al., 2003a). Thus, our goal was to determine whether two serpentine endemic species in Thlaspi, known to hyperaccumulate Ni in the field, have also evolved the Zn hyperaccumulation trait and to study the degree of genetic variation for Ni and Zn accumulation between and within populations.

We aimed to investigate differences between populations for Zn and Ni hyperaccumulation in Thlaspi pindicum, a Greek serpentine endemic (Tutin et al., 1993) and to examine genetic variation and heritability of hyperaccumulation in T. pindicum and Thlaspi alpinum var. sylvium, a species reported growing only on ultramafic substrates, located in Valle d’Aosta, Italy, and along the Swiss border (Reeves & Brooks, 1983; Verger, 1992). Meyer (1973) allocated both taxa to the Noccaea, the same section of the Thlaspi‘genus’ in which he located T. caerulescens. Thlaspi pindicum and T. alpinum var. sylvium are well-known Ni hyperaccumulators (c. 10 000 mg kg−1 Ni for both species), but Zn levels in herbarium specimens have been reported below 200 mg kg−1 for T. pindicum and ranging from 100 mg kg−1 to 3000 mg kg−1 for T. alpinum var. sylvium (Reeves & Brooks, 1983).

Materials and Methods

Field collection sites

Thlaspi pindicum Hausskn. (= T. tymphaea Hausskn.), was collected from The Katara Pass (Katara), Metsovo Lake (Metsovo) and the Peloponnese (Pelopon), Greece. The Katara Pass 1705 m is a well-exposed serpentine area in the Northern Pindus Mountains on E92, approx. 5 km from Metsovo. A large population of T. pindicum was found with Bournmuellera tymphaea on the south, upper slopes. Metsovo Lake is approx. 3 km south-east of the Katara Pass, and a smaller population of T. pindicum was found on the north-eastern banks of the lake, where serpentine rocks were also apparent. T. pindicum was collected in the Peloponnese by Dr Theophanis Constantinidis (Agricultural University of Athens, Greece). This population was found on serpentine along a forest road in the northern part of Mount Gerania (570–600 m) between summits Korifi and Makria Rachi, 10 km from Pissia village. All seeds in the population experiment were collected and pooled from approx. 20 parents while maternal families were collected individually. A large population of Thlaspi alpinum Crantz var. sylvium Gaudin was collected in the Valle d’Valtournenche, Italy, on a rocky slope approx. 3 km (altitude c. 2200 m) north of Chamois village. Maternal families were collected separately from the bulk collection.

In order to assess relative metal concentrations in the field, soils (four samples per site) and shoot material (six samples per site) were collected from the North Pindus Mountain sites and the Valle d’Valtournenche. Soil samples were taken from approx. 10 cm deep and were prepared for acid digestion by sieving with a 2 mm porous screen followed by grinding into a powder using an agate mortar and pestle. Plant and soil samples were acid digested and tested for metal concentrations (see later). Soil and leaf metal concentrations are given in Table 1 and Table 2, respectively.

Table 1. Mean metal concentrations, pH and magnesium (Mg) : calcium (Ca) ratios determined from sieved soil samples collected in the Northern Pindus Mountains, Greece, and Valle d’Valtournenche, Italy
LocationMetal concentrations (mg kg−1)
NiZnCrMgCaMg : CapH
  1. Data are mean ± SEM; n = 4.

Katara Pass 885 ± 3067 ± 2308 ± 1510852 ± 8209203 ± 7151.186.6
Lake Metsovo1076 ± 10057 ± 0.6282 ± 2531007 ± 15004136 ± 2307.496.5
Valtournenche 244 ± 12094 ± 19145 ± 3 6106 ± 4505914 ± 11001.036.4
Table 2. Mean shoot metal content in Thlaspi species collected in the Northern Pindus Mountains, Greece, and Valle d’Valtournenche, Italy
SpeciesLocationMetal concentrations (mg kg−1)
  1. Data are mean ± SEM; n = 6.

Thlaspi pindicum Katara Pass7241 ± 1292 613 ± 42521 ± 6
Thlaspi pindicum Lake Metsovo5977 ± 796 362 ± 87 6 ± 2
Thlaspi alpinum var. sylviumValtournenche1064 ± 1204703 ± 690Trace

Seed germination

In order to obtain seedlings for experimentation, seeds were sown in a sand/Perlite mix in 9 cm pots and placed in an Exeter glasshouse mist unit. At the cotyledon stage, seedlings were transplanted into a hydroponic environment in the controlled temperature room.

Hydroponic assay and growth conditions

Three experiments were performed to test for differences in Zn and Ni accumulation among different populations of T. pindicum and maternal families of T. pindicum and T. alpinum var. sylvium. The hydroponic system in all experiments consisted of 12-l trays capable of holding nine polystyrene strips. Seven holes were made in each strip and polystyrene plugs held seedlings in place. Nutrient solution composition was based on one by Chaney & Bell (1987) consisting of the following nutrients: 10 µm H3BO3, 0.1 mm K2HPO4, 0.2 mm MgSO4·7H2O, 2 µm MnCl2·4H2O, 0.2 µm CuSO4·5H2O, 0.5 mm Ca(NO3)2·4H2O, 0.1 µm MoO3, 0.5 mm KNO3, 10 µm ferric ethylenediamine di-(2 hydroxyphenyl acetate) (FeEDDHA), 0.2 µm ZnSO4·7H2O. Nickel and Zn treatments were added in the form of NiSO4 and ZnSO4. Seedlings at the cotyledon stage were transferred from sand and randomly placed into nonexperimental solution and left to acclimate for 2 wk before being transferred to experimental solution. Plants were grown in experimental solutions for 4 wk for all three experiments. Solutions were aerated continually and changed weekly in each experiment. The pH was recorded as 5.6 and did not vary much within each week before solutions were changed, and thus pH buffering was unnecessary. All hydroponic experiments were conducted in a controlled temperature room with a 10-h day (25°C day/15°C night).

Determination of metal concentrations

In order to determine the concentration of metals in the shoots and roots, individual plant samples were harvested, acid digested and tested for metal concentration on a UNICAM SP9800 Atomic Absorption Spectrophotometer (Thermo Electron Corporation). First, shoots and roots were separated, rinsed with distilled H2O and dried in an oven at 60°C. The dried plant samples were weighed on a Sartorius L 420+ 3-place balance (Sartorius Ltd, Epsom, UK) and digested in Gerhardt block digesters (C. Gerhardt UK Ltd, Brackley, UK) with Analar nitric acid (4 ml per tube). The digested residues were dissolved in 10 ml of 1 m HCl. After atomic absorption spectrophotometry (AAS) calibration with Spectrosol standards (BHD Chemicals Ltd, Poole, UK), the dissolved samples were analysed for metal concentrations. Soils collected from study sites were dried, sieved and ground to powder with a mortar and pestle before weighed and acid digested (same methods as with plants). Soil pH was determined using a Philips PW9421 pH meter and 1 : 1 (wt : v in distilled H2O) ratio.

Experiment 1

In order to test for population variation for Ni and Zn accumulation in T. pindicum, 20 seedlings from each of the three Greek populations were grown in each of three heavy metal-treated solutions: 100 µm Ni, 100 µm Zn and 100 µm Ni + 100 µm Zn (denoted as 100/100 Ni/Zn). Each tray contained 20 randomized replicates per population

Experiment 2

In order to test for genetic variation for Ni and Zn accumulation in T. pindicum, seeds of 20 maternal families were collected from different mother plants at the Katara Pass. After germination, the seedlings were randomized into hydroponic assay as in Experiment 1. Only the 10 families with nine or more germinated seeds could be used out of the 20 families. For seven of the families (numbered: 4, 9, 10, 13, 17, 18 and 19), five replicates were used per treatment (15 plants total/family). Owing to low germination, only three replicates could be used per treatment for families numbered, 2, 6 and 14 (nine plants total/family).

Experiment 3

In order to test for genetic variation for Ni and Zn accumulation in T. alpinum var. sylvium, seeds of 10 maternal families were collected from different mother plants in the Valle d’Valtournenche. Seedlings were placed into a hydroponic system as in Experiments 1 and 2 except that the concentrations of metals used were 50 µm Ni, 50 µm Zn, 50/50 µm Ni/Zn based upon preliminary tolerance studies. Only the eight families with 12 or more plants could be used out of the 10 families. For six of the families (numbered: 1, 2, 3, 4, 9 and 10), six replicates were used per treatment (18 plants total per family). owing to low germination, only three replicates could be used per treatment for families numbered 5 and 6 (nine plants total per family).

Data analysis

Unbalanced General Linear Model anovas (Minitab Inc., 1993) were conducted to compare metal accumulation and root: shoot ratios in Zn and Ni treatments and combined metal treatments for each experiment. In Experiment 1 with T. pindicum populations, both population and treatment were included in the model as fixed anova factors. For Experiments 2 and 3, maternal family was used in the GLM anova model as a random factor and treatment was incorporated as a fixed factor.


Experiment 1

Zinc accumulation in the shoots differed significantly between the three T. pindicum populations (Table 3) (Fig. 1a). In the 100 µm Zn solution, the Katara and Metsovo populations accumulated similar concentrations of Zn, with the Peloponnese population accumulating much less (Fig. 1a). The combined 100/100 µm Zn/Ni treatment had no overall effect on Zn accumulation in the three populations (Table 3) (Fig. 1a). Nickel hyperaccumulation was observed for all three populations and significantly varied between the populations (Table 3) (Fig. 1b). Unlike Zn, Ni accumulation was significantly reduced (Table 3) in the Zn/Ni treatment for all three populations (Fig. 1b).

Table 3. anova table for zinc (Zn) and nickel (Ni) hyperaccumulation and root : shoot ratios between Thlaspi pindicum populations and hydroponic treatments
Sourcedf F P
  1. Hydroponic treatments:100 µm Zn, 100 µm Ni, and 100/100 µm Zn/Ni.

Zn Accumulation
Populations  2 20.08< 0.001
Treatments  1 < 0.01  0.967
Populations × treatments  2  1.12  0.329
Ni Accumulation
Populations  2  6.19  0.003
Treatments  1195.68< 0.001
Populations × treatments  2  0.78  0.463
Zn Ratios
Populations  2  9.66< 0.001
Treatments  1  2.36  0.128
Populations × treatments  2  0.31  0.733
Ni Ratios
Populations  2  2.67  0.074
Treatments  1 10.48  0.002
Populations × treatments  2  0.42  0.661
Error 99  
Figure 1.

Shoot zinc (Zn) accumulation (a), shoot nickel (Ni) accumulation (b), Zn root : shoot ratios (c) and Ni root : shoot ratios (d) (mean ± SEM) in three populations of Thlaspi pindicum. Light bars, 100 µm Zn or Ni; dark bars, 100 µm Zn + 100 µm Ni (n = 17–20; see the Materials and Methods section).

Similar to the results with shoot Zn accumulation, the root : shoot ratios for Zn concentration were significantly different among populations (Table 3), but not between the Zn and Zn/Ni treatments (Table 3) (Fig. 1c). However, only the population from the Katara Pass was able to accumulate more Zn in the shoots than the roots for both treatments (root : shoot ratio < 1) (Fig. 1c). Unlike the Zn root : shoot ratios, Ni root : shoot ratios did not vary significantly among populations (Table 3), but Ni root : shoot ratios were significantly higher (Table 3) in the combined Ni/Zn treatment (Fig. 1d), and were always less than one (Fig. 1d).

In order to verify that differences in metal tolerance were not causing the observed differences in accumulation, the plant biomass of all populations was analysed. Total plant biomass showed little difference between metal treatments and had no effect on the total metal content or metal concentration for any of the populations or experiments reported here (data not shown). Thus, the metal concentrations accumulated were independent of the plant biomass.

Experiment 2

There was no evidence of heritable variation for either Zn or Ni hyperaccumulation in T. pindicum (NS Families, Table 4). Although there was no difference among the maternal families, Zn accumulation was significantly greater in the combination of Zn and Ni than in Zn alone (Table 4) (Fig. 2a) whereas Ni accumulation was significantly lower in the combination of Ni and Zn than in Ni alone (Treatments, Table 4) (Fig. 2b). The interaction between families and treatments for Ni accumulation was close to significance (P = 0.068, Table 4), but nonsignificant for Zn accumulation (P = 0.856, Table 4).

Table 4. anova table for zinc (Zn) and nickel (Ni) hyperaccumulation and root : shoot ratios between Thlaspi pindicum maternal families and hydroponic treatments
Sourcedf F P
  1. Hydroponic treatments: 100 µm Zn, 100 µm Ni, and 100/100 µm Zn/Ni.

Zn Accumulation
Families 9 1.670.229
Treatments 116.680.002
Families × treatments 9 0.520.856
Ni Accumulation
Families 9 0.380.920
Treatments 115.340.003
Families × treatments 9 1.890.068
Zn Ratios
Families 9 7.130.004
Treatments 1 1.680.217
Families × treatments 9 0.260.984
Ni Ratios
Families 9 0.590.780
Treatments 118.160.001
Families × treatments 9 0.630.764
Figure 2.

Shoot zinc (Zn) accumulation (a), shoot nickel (Ni) accumulation (b), Zn root : shoot ratios (c) and Ni root : shoot ratios (d) (mean ± SEM) in 10 families of Thlaspi pindicum from the Katara Pass, Greece. Light bars, 100 µm Zn or Ni; dark bars, 100 µm Zn + 100 µm Ni (n = 9–15; see the Materials and Methods section).

Zinc root : shoot ratios varied significantly among families (Table 4) (Fig. 2c). Despite the variation of Zn accumulation between Zn and combined Zn/Ni treatments (Table 4) (Fig. 2a), no difference between these treatments was observed for Zn root : shoot ratios (Table 4) (Fig. 2c). These ratios were approximately unity, indicating that the roots and shoots accumulated similar concentrations of Zn (Fig. 2c). Unlike Zn ratios, no significant variation was determined for Ni ratios among maternal families (Table 4) (Fig. 2d). However, Ni root : shoot ratios were significantly higher in the combined Ni/Zn treatment than in Ni alone (Table 4) (Fig. 2d).

Experiment 3

As with the T. pindicum maternal families, Zn hyperaccumulation did not vary significantly among T. alpinum var. sylvium maternal families (Table 5) (Fig. 3a), but Zn concentrations were significantly greater in the combined Zn/Ni treatment than in Zn alone (Table 5) (Fig. 3a). Nickel accumulation also did not vary among maternal families (Table 5) (Fig. 3b), and as in the experiments with T. pindicum, Ni concentrations were significantly reduced in the combined treatment compared with Ni alone (Table 5) (Fig. 3b). Similar to results with T. pindicum, the interaction between families and treatments for Ni accumulation was significant (P = 0.037, Table 5), whereas no interaction was observed for Zn accumulation (P = 0.856, Table 5).

Table 5. anova table for zinc (Zn) and nickel (Ni) hyperaccumulation and root : shoot ratios between Thlaspi alpinum var. sylvium maternal families and hydroponic treatments
Sourcedf F P
  1. Hydroponic treatments: 50 µm Zn, 50 µm Ni, and 50/50 µm Zn/Ni.

Zn Accumulation
Families 7 1.69  0.253
Treatments 1 7.42  0.023
Families × treatments 7 1.30  0.258
Ni Accumulation
Families 7 0.41  0.869
Treatments 146.63< 0.001
Families × treatments 7 2.27  0.037
Zn Ratios
Families 7 1.51  0.299
Treatments 1 6.31  0.029
Families × treatments 7 0.80  0.586
Ni Ratios
Families 7 0.85  0.581
Treatments 112.37  0.007
Families × treatments 7 1.80  0.100
Figure 3.

Shoot zinc (Zn) accumulation (a), shoot nickel (Ni) accumulation (b), Zn root : shoot ratios (c), and Ni root : shoot ratios (d) (mean ± SEM) in nine families of Thlaspi alpinum var. sylvium from Valle d’Valtournenche, Italy. Light bars, 100 µm Zn or Ni; dark bars, 100 µm Zn + 100 µm Ni (n = 9–18; see the Materials and Methods section).

Similar to results with metal accumulation, no significant differences (Table 5) were found for either Zn or Ni root : shoot ratios among families (Fig. 3c,d). However, both Zn and Ni root : shoot ratios were significantly higher (Table 5) in the combined Zn/Ni solution than in Zn and Ni alone (Fig. 3c,d).


The purpose of this study was to investigate, under standard conditions, the accumulation of Zn and Ni by two serpentine endemic taxa of Thlaspi, known from field data to accumulate Ni. We have shown that both taxa, T. pindicum and T. alpinum var. sylvium are capable of hyperaccumulating Zn under laboratory conditions although in the field they only accumulate Ni. There have been few other studies of this aspect of the hyperaccumulation phenomenon. Thlaspi oxyceras, a Turkish serpentine endemic and Ni hyperaccumulator, located in section Thlaspiceras (Meyer, 1973), has been shown to accumulate low levels of Zn – less than 1000 mg kg−1 from soils enriched with Zn and Ni (Peer et al., 2003). Species of Thlaspi in section Noccaea are more closely related to one another than they are to species in Thlaspiceras (Mummenhoff et al., 1997; Koch & Mummenhoff, 2001). Although T. oxyceras accumulates only low levels of Zn, other species adapted to serpentine in the Noccaea, such as Thlaspi ochroleucum, Thlaspi graecum, and Thlaspi epiretum have been found to accumulate higher levels of Zn (greater than 1000 mg kg−1, from serpentine) (Reeves & Brooks, 1983). By contrast, both the species studied here exhibited Zn and Ni accumulation patterns typical of the well-studied metal hyperaccumulator, T. caerulescens (Pollard & Baker, 1996; Meerts & van Isacker, 1997; Escarréet al., 2000; Pollard et al., 2002; Assunção, 2003). This species, which is found on both contaminated and uncontaminated soils, and on both calamine and serpentine soils, exhibits both Zn and Ni accumulation in all populations studied to date.

We found significant between-population variation for both Ni and Zn accumulation in T. pindicum. Populations from the Katara Pass, Metsovo Lake and the Peloponnese accumulated large quantities of Ni and Zn into their shoots, but these amounts were different among populations. The Katara Pass and Metsovo populations were superior to the Peloponnese population in that they accumulated the greatest amounts of Zn and Ni in their shoots, but only the Katara population was able to maintain root : shoot ratios below one in both Zn and Ni as well as the metals in combination. By contrast, we found no evidence of within-population genetic variation in two serpentine endemics, T. pindicum and T. alpinum var. sylvium. Ten maternal families of T. pindicum from the Katara Pass population as well as eight families in T. alpinum var. sylvium showed no significant variation for either Zn or Ni hyperaccumulation. Thus, no signs of genetic variation for metal hyperaccumulation and heritability for either metal could be detected. However, an interaction occurred between families and metal treatments for Ni accumulation for both species, indicating that there could be some genetic variation for the relative uptake of the two metals. There was also significant variation in the Zn root : shoot ratios, which could suggest that that there is variation in the mechanisms for metal translocation and transport. Thus, we have shown significant between-population variation in T. pindicum, but not significant within-population variation. This pattern is typical of a self-fertilizing species, but there has been no study of the breeding system of this species. Thlaspi pindicum is a close relative to T. caerulescens, which is primarily self-fertilizing, although the extent of this varies between populations (Dubois et al., 2003).

Our results can be compared with those of studies of the model hyperaccumulating species T. caerulescens, which is closely related to the serpentine endemics T. pindicum and T. alpinum var. sylvium (Meyer, 1973). Between-population variation in metal accumulation has often been found (Pollard & Baker, 1996; Bert et al., 2000; Escarréet al., 2000; Pollard et al., 2002; Assunção et al., 2003a; Roosens et al., 2003). Several genetic studies have also investigated within-population genetic variation for hyperaccumulation in this species (Pollard & Baker, 1996; Meerts & van Isacker, 1997; Escarréet al., 2000; Pollard et al., 2002). All of these studies have shown a small amount of genetic variation and some degree of heritability for Zn accumulation. Pollard et al. (2002) observed genetic variation among maternal families within three of five populations, including a serpentine population. Heritability estimates were determined to be above zero and some close to one, indicating that Zn and Ni accumulation varies in heritability in both metalliferous and nonmetalliferous populations.

In this study we also investigated the interaction between Zn and Ni accumulation when both metals are present in high concentrations. In all experiments, the accumulation of Ni was reduced when Zn was also present, while in two of the experiments the uptake of Zn was enhanced by the presence of Ni. Assunção (2003) also found that Zn was preferred over Ni in a serpentine population of T. caerulescens and she proposed that Ni hyperaccumulation is less metal-specific than Zn hyperaccumulation. Preference and competition between metals has previously been hypothesized for species growing in high concentrations of various metals (Boyd & Martens, 1998). In our study, competition is apparent between Ni and Zn for both species. The competition between the metals probably occurs in the roots since the Zn root : shoot ratios were not affected by Ni hyperaccumulation and the Ni root : shoot ratios were higher in the combined treatment.

If a serpentine endemic shows preferential accumulation of Zn rather than Ni, when it does not hyperaccumulate Zn in nature, then this might suggest that the processes leading to the evolution of Ni hyperaccumulation on serpentine are not solely or possibly even primarily for Ni accumulation. It is possible that Ni accumulation is the character under direct selection, but this has been achieved by acting on a pre-existing Zn accumulation system. Zinc accumulation would then be a residual effect of this character. Conversely, selection might be for an efficient metal uptake system capable of transporting and accumulating several different metals, as seen in T. caerulescens (Escarréet al., 2000; Lombi et al., 2000). Thlaspi caerulescens is known to hyperaccumulate Ni, Zn, Co, and Cd and can also accumulate lead (Pb) and manganese (Mn) to high levels (Baker et al., 1994). Nonetheless, because of the high Ni : Zn ratio in serpentine soil, such as in Table 1, an enhanced transport system with preference for Zn over Ni would perceivably result in higher Ni accumulation than Zn accumulation. This was evident with T. pindicum in nature (Table 2) but not T. alpinum var. sylvium, which accumulated more Zn in nature than Ni (Table 2), but the Ni concentration in the soil in the Valle d’Valtournenche was much less than that found in the Pindus Mountains (Table 1) where T. pindicum was collected.

The fact that these species appear to have the latent ability to hyperaccumulate Zn, though they do not in nature, would suggest that it would be difficult to postulate an adaptive hypothesis for this Zn accumulation phenotype. Adaptive hypotheses for hyperaccumulation have focused on herbivory deterrence, metal tolerance mechanisms, allelopathy and drought resistance (Boyd & Martens, 1992; Krämer et al., 1997; Pollard & Baker, 1997; Huitson & Macnair, 2003). Another hypothesis however, suggested by Boyd & Martens (1992), is that of inadvertent uptake, in which the phenotype is the byproduct of another process. Thus if some species of Thlaspi hyperaccumulate Zn inadvertently whilst foraging for another nutrient, then this would indicate that Zn hyperaccumulation has not evolved as a result of selection for that aspect of the phenomenon. If serpentine endemics preferentially hyperaccumulate Zn over Ni, then this might indicate that the translocation mechanism has a greater affinity for Zn than Ni. This would also suggest that Ni is transported via Zn transporters and thus Zn accumulation may have evolved first. Moreover, it appears that a high Zn affinity transporter system has evolved in species that grow only on Ni-rich substrates and may be present in most of the Zn and Ni accumulating species in Thlaspi. Further study is needed to focus on the inadvertent uptake hypothesis, which would shed light on why Zn hyperaccumulation occurs on nonmetalliferous soils as well as in serpentine endemics that rarely hyperaccumulate Zn in nature.


We thank Dr T. Constantinidis at the Agricultural University of Athens for his advice on field sites for T. pindicum and collecting seed for us in the Peloponnese. We also thank Dr A. Polatschek at the Vienna Natural History Museum for his advice on field sites for T. alpinum var. sylvium. Fieldwork was supported by field-grants from the Genetics Society and Exeter University School of Biosciences.