• Arabidopsis halleri halleri;
  • Arabidopsis halleri ovirensis;
  • hydroponics;
  • metal tolerance evolution;
  • metallicolous and nonmetallicolous populations;
  • model species;
  • pseudometallophyte


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. References

Although current knowledge about the overall distribution of zinc (Zn) tolerance in Arabidopsis halleri populations is scarce, the species is an emerging model for the study of heavy metal tolerance in plants. We attempted to improve this knowledge by testing the Zn tolerance of scattered European metallicolous (M) and nonmetallicolous (NM) populations of A. h. subsp. halleri and A. h. subsp. ovirensis in hydroponic culture. The occurrence of constitutive tolerance was unconditionally established in A. h. halleri and tolerance was extended to the subspecies ovirensis. M populations were the most tolerant but there was a continuous range of variation in tolerance from NM to M populations. Finally, relatively high levels of tolerance were detected in some NM populations, suggesting that enhanced tolerance could be present at high frequency in populations that have not experienced metal exposure. We used our results to argue the evolutionary dynamics and origin of Zn tolerance in A. halleri.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. References

Heavy metal tolerance in plants is commonly defined as the ability to survive on soils which would prove toxic to most living things because they contain (very) high levels of one or more metals (Antonovics et al., 1971; Macnair & Baker, 1994). It was first recognized by Prat (1934) who showed that a copper mine population of Silene dioica (syn. Melandrium silvestre) was able to grow far better in copper-contaminated soils than populations from uncontaminated areas. Since this initial work, tolerances to a wide range of heavy metals have been encountered in many plant species belonging to widely separated families (Antonovics et al., 1971). These taxa have been classified as either absolute (strict or eu-) or facultative (pseudo-) metallophytes, depending on whether they occur on contaminated sites only or on metalliferous as well as nonmetalliferous soils (Lambinon & Auquier, 1964).

Metal tolerance has been considered to be ‘an example of more powerful evolution in action than industrial melanism in moths’ (Antonovics et al., 1971). It has therefore been the focus of many studies and the evolutionary framework of heavy metal tolerance in plants is nowadays relatively well documented. Evolutionary studies have argued that metal tolerance could evolve rapidly following exposure to heavy metal stress (Wu et al., 1975; Al-Hiyaly et al., 1988) and that it could have evolved independently in geographically distant conspecific populations (Westerbergh & Saura, 1992; Schat et al., 1996; Vekemans & Lefèbvre, 1997; Koch et al., 1998; Mengoni et al., 2001; Pauwels et al., 2005). Recent genetics studies have demonstrated that in most cases tolerance is governed by a few major genes (for reviews, see Macnair, 1993; Macnair et al., 2000). Quantitative polymorphism is nevertheless usually observed and implies the existence of hypostatic ‘modifier’ genes that influence the degree of expression of the tolerance phenotype (Schat & Ten Bookum, 1992; Schat et al., 1993; Smith & Macnair, 1998; van Hoof et al., 2001). In particular, modifiers are supposed to be responsible for the continued evolution towards a high level of average tolerance in metallicolous populations, in response to high levels of metal exposure (Schat et al., 1993; Smith & Macnair, 1998).

In recent years, substantial efforts have been expended to gain further insights in the genetic mechanisms of metal tolerance in plants, i.e. identifying candidate genes and characterizing physiological mechanisms. Thus, many studies have involved high throughput genomics, transcriptomics, proteomics and metabolomics technologies. Genomic and post-genomic technologies are necessarily organized around only a few model species (Feder & Mitchell-Olds, 2003) which makes it crucial to search for the appropriate model. Whereas classical models (e.g. Arabidopsis thaliana) were selected by scientists for their particular genetic and development features, the study of ecologically important traits is motivated by additional criteria (Jackson et al., 2002). In the context of heavy metal tolerance in plants, two pseudometallophyte species that both tolerate and hyperaccumulate zinc (Zn) and cadmium (Cd) have recently been proposed: Arabidopsis halleri (L.) (O'Kane & Al-Shehbaz) [syn. Cardaminopsis halleri (L.) Hayek] and Thlaspi caerulescens J. & C. Presl., both Brassicaceae (Assunção et al., 2003b).

Because it is the closest metal tolerant relative of the pre-eminent model system A. thaliana (Koch et al., 2001; Al-Shehbaz & O'Kane, 2002), A. halleri doubtlessly presents major advantages for the study of tolerance in plants. Indeed, the fact that many molecular tools developed in A. thaliana can readily be transferred to A. halleri (Mitchell-Olds, 2001; Lexer & Fay, 2005) offers the unparalleled opportunity of an integrative study of metal tolerance. Topics as diverse as the nature, the number and the regulation of genes conferring metal tolerance, the epistatic relationships between them, their origin and evolution of tolerance genes, the impact of selection on these genes can now be addressed. Thus, A. halleri is already included in ‘–omics’ studies, and its efficacy in identifying candidate genes involved in metal tolerance and hyperaccumulation has been demonstrated (Becher et al., 2004; Dräger et al., 2004; Weber et al., 2004).

However, the best molecular technologies will be of limited interest if they are not sustained by a good understanding of the ecological range of the model species (i.e. the relative abundance of populations on and off polluted sites) as well as a good understanding of the distribution of tolerance abilities in populations (is this distribution related to the local amount of metals in soils?). So far, in contrast to T. caerulescens which has been studied for many decades (Koch et al., 1998; Assunção et al., 2003b), this knowledge remains sparse in A. halleri which was sometimes said to be typically found on metalliferous soils (Dahmani-Muller et al., 2000; Pollard et al., 2002) although numerous NM populations have been mentioned (Bert et al., 2000, 2002). Zn tolerance was assumed to occur throughout the species in A. halleri (Bert et al., 2000), but only five populations located in two regions of northern Europe and belonging to the single subspecies A. h. halleri were studied in controlled conditions. Meanwhile, the species is widely distributed in Europe and at least two other subspecies are recognized (Al-Shehbaz & O'Kane, 2002). Although it is assumed to be constitutive, Zn tolerance is generally expected to have evolved towards increased levels in M populations of A. halleri, in response to high selective pressure caused by high level of metals in soils (Bert et al., 2000). A recent phylogeographic study using cpDNA (Pauwels et al., 2005) showed that A. h. halleri population structure was related to geographic isolation rather than to Zn exposure in northern Europe. This implies that geographically close nonmetallicolous (NM) and metallicolous (M) populations (e.g. from southern and northern Germany) are more genetically similar than distant M populations (e.g. from Poland and northern Germany), and that distant M populations have evolved independently. Consequently, if enhanced tolerance was confirmed to be a general feature of M populations, it should have evolved several times in parallel. Moreover, a study of population genetic structure using five nuclear microsatellites over a 500-m-long transect characterized by a gradient of heavy metal concentrations found no evidence of genetic divergence due to spatial heavy metal heterogeneity in these neutral markers, suggesting that long-distance pollen dispersal and extensive pollen flow could attenuate the local effect of metal exposure on differentiation by neutral markers (van Rossum et al., 2004). Altogether, population studies based on neutral genetic markers suggest that the relation between Zn exposure in the wild and the distribution of inherited Zn tolerance abilities in populations might not be as straightforward as usually thought.

The present study attempted to provide a better description and understanding of the distribution of Zn tolerance in A. halleri populations, in particular in relation to the available Zn content in soils, and discusses the origin and the evolutionary dynamics of Zn tolerance in A. halleri in Europe. To assess accurately the overall level of quantitative polymorphism for Zn tolerance, 28 widely distributed populations of A. h. subsp. halleri, the only subspecies which has been recorded on metalliferous soils in Europe, were sampled. To be representative of the ecological range of the subspecies, populations were collected in both polluted and nonpolluted areas. In addition, three populations of A. h. subsp. ovirensis were sampled in the Carpathians Mountains (Romania). To identify the genetic component of population differentiation for Zn tolerance, the tested populations were grown under uniform environmental conditions using hydroponic culture in controlled growth chambers.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. References


In the summers of 1999 and 2003, respectively 12 and 16 widely distributed populations of A. halleri subsp. halleri were sampled so as to cover the range of the subspecies in continental Europe (Table 1). At each study site, viable mature seeds were collected as maternal progenies from distinct individuals in order to obtain living plants and to perform a tolerance test in hydroponic culture. Soil samples were also collected at each study site in order to estimate the concentrations of extractable Zn. Many results from soil sample analysis have already been published in Bert et al. (2002). Additional samples were analysed using the protocol described in Bert et al. (2002) and all the results are reported in Table 1.

Table 1.   Names and locations of population samples of Arabidopsis halleri. When shared, the names corresponds to those used in Pauwels et al. (2005).
PopulationSubspeciesTypeOriginGPS coordinatesAltitude (m)Ecological backgroundNi est.Extractable Zn in soil (μm) ±  SE (ns)Year of testn
  1. A, Austria; B, Belgium; CH, Switzerland; D, Germany; F, France; I, Italy; PL, Poland; RO, Romania; SK, Slovakia; M, metallicolous populations; NM, nonmetallicolous populations; NMp, nonpolluted populations; ovirensis, A. h. ovirensis populations. Ni est.: approximate population size; ns, number of soil samples; n, sample size for sequential test; n.a., not available.

A05A. h. halleriNMTyrol47°13′77811°22′779 807Garden lawn under conifers<507 ± 3 (3)200420
A08A. h. halleriNMTyrol47°25′17711°51′957 522Meadow 100–500177 ± 123 (4)200420
B01A. h. halleriNMHautes Fâgnes50°29′6306°40′ 160Underwood 500–1000n.a.200016
CH01A. h. halleriNMTessin45°59′23308°50′387 340On a shaded wall, around a private garden>1005 (2)200420
CH03A. h. halleriNMTessin46°10′63208°43′182 290Meadow at a wood skirt, along a roadway<5017 ± 6 (3)200418
D01 (G1*)A. h. halleriNMBohemian forest49°10′6412°09′′88 340Nitrogenous Regen river bank<503 ± 2*200010
D08 (G8*)A. h. halleriNMpHarz51°53′7910°29′04 190Old mine (19th century)<50123 ± 9*200013
D09 (G9*)A. h. halleriNMpHarz51°53′4610°25′16 190Roadside<5052*200015
D11 (G11*)A. h. halleriNMpHarz51°51′2710°21′95 674Underwood<507*200011
D12 (G12*)A. h. halleriMHarz51°51′9110°17′90 673Mine rubble>5001272 ± 1047*200016
D13 (G13*)A. h. halleriMHarz51°55′2210°18′50 237Roadside and lawn 50–500405 ± 207*200021
F01 (Auby†)A. h. halleriMNorth50°25′03°03′  20Wood near a smelter plant>10005260 ± 280†2000/200414/20
I01A. h. halleriNMTrentin46°52′78311°24′308 967Small meadow along a roadway<10013 ± 7 (3)20045
I02A. h. halleriNMTrentin46°49′92011°43′823 840Meadow along a footpath 100–5004 ± 4 (2)200420
I07A. h. halleriNMTrentin46°29′62310°53′3491388Meadow>10008 ± 5 (3)200419
I08A. h. halleriNMTrentin46°29′73110°53′7391363Meadow 500–100017 ± 10 (2)200420
I09A. h. halleriNMTrentin46°43′88211°25′9581530Meadow>500011 ± 8 (2)200420
PL01 (P1*)A. h. halleriNMpSilesia50°14′8018°57′04 200Wood in Katovice suburbs>50022 ± 1*200020
PL04 (P4*)A. h. halleriMSilesia50°29′9818°55′79 269Metallurgical factory>1000481*200010
PL06 (P5*)A. h. halleriMSilesia50°16′9519°01′52 306Old mine>10002490*200010
PL07A. h. halleriMSilesia50°17′01419°29′564 335Meadow on a ZN/Pb spoil n.a.n.a.200420
PL08A. h. halleriNMpSilesia50°06′24320°21′569 190Oak-hornbeam forest n.a.n.a.200420
RO05A. h. halleriNMCarpathians46°43′30523°02′637 990Footpath along a forest skirt, near a brook<503 ± 3 (2)200420
RO06A. h. halleriNMCarpathians46°39′08323°02′3641154Picea abies forest skirt<5014 (1)200420
RO08A. h. halleriNMCarpathians46°50′68822°48′082 900River bank<103 (1)200420
RO09A. h. halleriNMCarpathians46°49′47522°45′635 666Meadow beside a road<509 ± 8 (3)200416
RO14A. h. ovirensisNMFagaras Mountains45°36′13924°37′0642050Alpine lawn 50–500n.a.200420
RO15A. h. ovirensisNMFagaras Mountains45°36′13924°37′0642050Alpine lawn 50–500n.a.200420
RO18A. h. ovirensisNMFagaras Mountains45°36′16324°37′3362230Alpine lawn<5021 ± 25 (2)200418
SK02 (Sl2*)A. h. halleriNMTatras48°46′1721°07′81 690Shady meadow>5001*200015
SK05 (Sl4*)A. h. halleriNMHigh Tatras49°16′9820°09′241027Tatransla javorina (Nature reserve)>5001*2000/200421/8

Most populations (17) were scattered throughout the European mountain ranges (Fig. 1), at middle to high altitudes (mean = 822 m), far from any known metal pollution source, in noncontaminated environments (NM populations, Table 1). They were located in the Tatras and High Tatras mountains (Slovakia), in the Apuseni Mountains (Romania), in the Bohemian Forest (Germany) and on northern and southern slopes of the Alps (Austrian Tyrol; Ticino, Switzerland and Trentino, Italy). Eleven populations (D08–09, D11:13, F01, PL01, PL04, PL06:08, Fig. 1) were sampled outside the mountain ranges, at low to moderate altitudes (mean =314 m), in the three known polluted areas the species has colonized, i.e. in northern France, Silesia (Poland) and Harz (Germany). In these disturbed regions, A. h. halleri populations have been shown to locally cluster into homogeneous genetic assemblages (Pauwels et al., 2005). Since the extension of the A. h. halleri distribution area in initially unsuitable disturbed habitats has to be related to anthropogenic metal pollution via industrial or mining activities (Berton, 1946; Fabiszewski, 1986; Ernst, 1990), all populations were classified as M populations and most of them (D12–13, F01, PL04–06–07) indeed occurred on highly polluted soils (Table 1). However, a detailed study of Zn concentration in the soils of each population revealed that the species also colonized slightly polluted habitats (<300 μg g−1, see Bert et al., 2002) where selection pressure towards enhanced tolerance should be strongly reduced (such as in NM populations, see Table 1). In industrial areas, ‘non-M’ populations (D08-09-11 and PL01–08) may nevertheless differed from true NM ones by their ability to exchange genes with the geographically proximate M populations (van Rossum et al., 2004; Pauwels et al., 2005). Both ecological and genetic considerations (Pauwels et al., 2005) further suggests that M populations might have settled first in these areas and that populations only subsequently colonized slightly polluted habitats. Populations D08-09-11 and PL01–08 were therefore considered to be in a particular category and qualified as NMp (for nonmetallicolous in a polluted area).


Figure 1.  Distribution map of sampled Arabidopsis halleri populations in Europe. Nonmetallicolous, nonmetallicolous in polluted area and metallicolous populations are respectively indicated by white, grey and black circles. Populations of A. h. ovirensis are indicated by white triangles.

Download figure to PowerPoint

In addition to A. h. halleri populations, mature seeds from maternal progenies and soil samples from three populations of A. h. subsp. ovirensis were collected at high altitudes (mean = 2110 m) in the Fagaras Mountains (Southern Carpathians, Romania).

Zinc tolerance experiment

Seeds sampled from each population were first sown in sand filled pots in a greenhouse. After 3 weeks of growth, a maximum of 20 seedlings per population (depending on the germination rate) were transferred into hydroponic conditions in a controlled-growth chamber (temperature 20 °C day : 15 °C night; light 14 h day : 10 h night). Ten-litre containers containing perforated polystyrene trays in which the seedlings were planted into nylon wool were used. In order to avoid local environmental effects, plants were randomly distributed in containers so that each population was represented by at least one individual in each container. Moreover, containers were randomly distributed in the growth chamber and moved around once a week. The composition of the nutrient solution in containers was: 0.2 mm MgSO4, 0.5 mm Ca(NO3)2, 0.5 mm KNO3, 0.1 mm K2HPO4, 0.2 μm CuSO4, 2 μm MnCl2, 10 μm H3BO3, 0.1 μm MoO3, 10 μm FeEDDHA and 1 μm ZnSO4. The nutrient solution was changed once a week. After a 3-week period allowing plants to acclimate to hydroponic conditions, the sequential tolerance test started following the method of Schat and Ten Bookum (Schat & Ten Bookum, 1992). It involves the qualitative measurement of root growth (growth vs. no growth) at sequential concentrations. In our experiment, Zn concentrations varied from 1 μm during the first week to 2000 μm during the ninth and last week of treatment, increasing each week by a constant step of 250 μm. At the start of the sequential test, the roots of all plants were blackened by dipping them in a suspension of activated charcoal, and rinsed in deionized water to eliminate excess of charcoal. Plants were returned to 1 μm Zn for a further week when the presence of any new roots (uncoated) visible beyond the charcoal-coated roots was recorded. Roots were then reblackened and transferred to a fresh test solution with an increased Zn concentration for an additional week. The recording and blackening of new roots were repeated for each plant at each concentration tested. In such a test, the presence of white roots at a given Zn concentration indicates growth. The lowest concentration at which root growth definitively stopped (EC100 in Schat & Ten Bookum, 1992) was determined for each individual. EC100 was interpreted as the lowest concentration that was not tolerated by the individual in question. To check this interpretation, we transplanted ‘nontolerant’ plants to a fresh solution containing a Zn dose corresponding to their respective EC100s. Mortality usually occurred after 2 weeks of exposure at such a constant dose, validating the use of ‘presence of root growth or not’ as an indicator of plant tolerance and, by definition, survival. Plants for which no new roots were observed were removed from the experiment.

Populations sampled in 1999 and 2003 were tested for Zn tolerance separately in time (in spring 2000 and spring 2004, respectively, see Table 1). However, in order to be able to combine survival data from both experiments into a single data set, the same protocol was used, except for the plant density in containers (28 plants per container in 2000, 24 in 2004). Populations F01 (M) and SK05 (NM) were also included in each experiment to allow comparison and to ensure the homogeneity of data sets. In 2000, 12 individuals of A. thaliana were randomly distributed in containers so as to be represented one to three times in each container; in 2004, 20 individuals of A. lyrata petrea were included and similarly randomized. Both of these nonmetallophyte relatives were used as nontolerant control species.

Statistical analysis

At each experimental dose, root growth was encoded for each plant as a binary variable and interpreted as individual survival or mortality. Survival proportions SiX of individuals from population i that did not reach their EC100 were estimated at each dose X for which an end of growth was observed in that population. Note that the number ni of survival proportions SiX was a secondary variable that could have differed between populations. The resulting survival curves [Si = f(X), where ‘X’ is the concentration of Zn in the test solution] were fitted to sigmoidal dose–response curves with a variable slope by nonlinear regression using graphpad prism version 4.00 for Windows (GraphPad Software, San Diego California USA, The equation used for regression was a two-parameter logistic equation established as following: Si = 100/(1 + 10?((log(T50i)−log(X))*Bi)) where the variable ‘Si’ represents the survival proportions of population i, expressed as a percentage, ‘X’ is the concentration of Zn in the test solution, expressed in μm, the parameter ‘T50i’ is the concentration for which 50% of individuals from population i had reached their EC100 (median EC100, SiT50i =50), and the parameter ‘Bi’ is the slope factor describing the steepness of the curve. Estimated T50 values were considered as overall estimations of the tolerance of populations whereas the slope factors B estimated the range of within-population polymorphism for tolerance. For both parameters and for all populations, the standard error (SE) and 95% confidence intervals of best-fit values were estimated by GraphPad prism during curve fitting.

In order to test for differences in average levels of tolerance abilities or within-population polymorphism between groups of A. h. halleri populations either tested separately in time (2000 vs. 2004) or corresponding to distinct ecological categories (NM vs. NMp vs. M), box-plots were drawn for both parameters and for each category and Mann–Whitney tests were performed using Minitab 13.20 (Minitab Inc., State College, PA, USA).

In order to estimate precise confidence intervals for T50 and to be able to perform pairwise comparisons of tolerance abilities of populations, the equation used for regression was therefore rewritten so as to directly estimate best-fit values and SE for log(T50) rather than T50 (Motulsky 1999) [S = 100/(1 + 10?((LOGT50 − log(X))*B, where LOGT50 and B are the estimated parameters]. LOGT50 values were then compared through unplanned comparisons between pairs of populations using the GT2-method for unequal sample size (Sokal & Rohlf, 1981). The method computes comparison intervals for LOGT50 for each i population using the two-tailed studentized maximum modulus mα[k*,υ] distribution as a critical value (Sokal & Rohlf, 1981). Best-fit values from distinct populations were declared significantly different if their intervals did not overlap. For a 5% experiment-wise error rate, k* = (35*(35−1))/2 = 595 and υ = 124, we used the tabled value m.05[561,120] = 4.038 as a conservative approximation to the desired value m.05[595,153].


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. References

The distribution of EC100s is presented in Fig. 2. As expected, A. thaliana and A. l. petrea nontolerant individuals reached their EC100 at first doses of exposure (i.e. 250 and 500 μm) and all A. halleri individuals showed clear tolerance abilities. Expect for a few individuals that also reached their EC100 before 750 μm (2% and 1% of NM and M tested individuals, respectively, belonging to populations B1, D1, CH1 and F1), most of the A. halleri individuals ranged between 750 and 2000 μm of Zn in solution culture. Interestingly, mortality events were continuously distributed over all doses of exposure, in almost each type of population (i.e. NM, M and ovirensis, the only exception was for NMp populations in which no mortality event was observed at 1000 μm).


Figure 2.  Histograms showing frequency distribution of EC100s for each category of population. A. h., Arabidopsis halleri; M, metallicolous; NM, nonmetallicolous; NMp: nonmetallicolous in polluted area.

Download figure to PowerPoint

All survival data were well fitted by dose–response curves (all R2 > 0.93). Estimated values for both T50 and B were precise (relatively low standard errors) and consistent (no aberrant results, Table 2). The number ni of distinct survival proportion available for curve fitting varied between populations (from ni = 3 to ni = 8, Table 2) and was particularly low for the control species A. thaliana and A. l. petrea (ni = 3) (Figs 2 and 3). The ni values of A. halleri populations were weakly related to sample size (R2 = 0.282, P = 0.001). However, differences in ni values did not explain variation in T50 and T50 standard error (R2 = 0.068, P = 0.144 and R2 = 0.005, P = 0.692 respectively). A weak but significant correlation was detected between ni and B (R2 = 0.268, P =0.002), and, to some extent, between ni and the standard error of B (R2 = 0.208, P = 0.007). As a low ni always meant that the survival proportions dropped from 100% to 0% in a few successive concentrations of exposure, such a correlation was expected. Moreover, differences in B and the standard error of B were not related to differences in sample size (R2 = 0.012, P = 0.545 and R2 = 0.006, P = 0.666, respectively).

Table 2.   Best-fit values and standard errors of T50, LOGT50 and B.
PopulationniT50 (±SE)LOGT50 (±SE)B (±SE)d.f.
  1. Nonmetallicolous, nonmetallicolous in polluted area and metallicolous populations are respectively indicated in light grey, dark grey and black.

  2. ni, number of doses for which an end-of-growth event was observed in the population i, corresponds to the number of survival proportions available for survival curves fitting; d.f., degree of freedom of parameters (note that d.f. are equal to ni − 2 rather than ni − 1 because both parameters were estimated at the same time, Motulsky, 1999); SE, standard error.

A056901.9 (±17.71)2.955 (±8.529 × 10−3)−6.818 (±0.765)4
A0861204 (±19.72)3.081 (±7.113 × 10−3)−7.218 (±0.805)4
B0161011 (±52.69)3.005 (±2.262 × 10−2)−4.237 (±0.681)4
CH0171129 (±24.75)3.053 (±9.523 × 10−3)−5.868 (±0.724)5
CH0351099 (±33.02)3.041 (±1.305 × 10−2)−7.306 (±1.468)3
D0161233 (±31.68)3.091 (±1.115 × 10−2)−7.533 (±1.347)4
D0841437 (±21.03)3.157 (±6.355 × 10−3)−5.128 (±0.615)2
D0951335 (±44.04)3.125 (±1.433 × 10−2)−7.31 (±1.365)3
D1161247 (±34.96)3.096 (±1.217 × 10−2)−5.727 (±0.865)4
D1261432 (±35.30)3.156 (±1.07 × 10−2)−7.910 (±1.391)4
D1351384 (±23.06)3.141 (±7.239 × 10−3)−8.029 (±1.047)3
F01–200051503 (±18.02)3.177 (±5.205 × 10−3)−9.909 (±1.083)3
F01–200461569 (±39.06)3.196 (±1.081 × 10−2)−9.13 (±1.948)4
I0141092 (±31.44)3.038 (±1.25 × 10−2)−4.921 (±0.697)2
I026993.6 (±12.97)2.997 (±5.669 × 10−3)−8.986 (±1.036)4
I0761348 (±13.99)3.13 (±4.508 × 10−3)−8.876 (±0.720)4
I0851144 (±27.37)3.058 (±1.039 × 10−2)−7.451 (±1.276)3
I0971007 (±23.59)3.003 (±1.017 × 10−2)−5.516 (±0.629)5
PL0161407 (±16.84)3.148 (±5.196 × 10−3)−9.078 (±0.912)4
PL0451509 (±24.02)3.174 (±1.299 × 10−2)−5.827 (±0.742)3
PL0641648 (±15.13)3.217 (±03.99 × 10−3)−15.58 (±1.778)2
PL0761495 (±38.91)3.175 (±1.13 × 10−2)−8.115 (±1.574)4
PL0871390 (±22.05)3.143 (±06.89 × 10−3)−7.324 (±0.768)5
RO0551262 (±66.1)3.101 (±2.275 × 10−2)−5.886 (±1.607)3
RO0681208 (±42.52)3.082 (±1.528 × 10−2)−6.06 (±0.451)6
RO0861199 (±13.01)3.079 (±4.714 × 10−3)−4.375 (±0.923)4
RO0961377 (±49.84)3.139 (±1.572 × 10−2)−6.796 (±2.316)4
RO1461368 (±65.99)3.136 (±2.096 × 10−2)−5.942 (±1.105)4
RO1561162 (±15.05)3.065 (±5.627 × 10−3)−7.092 (±0.495)4
RO1871170 (±59.53)3.068 (±2.21 × 10−2)−6.045 (±1.379)5
SK0251022 (±51.61)3.01 (±2.192 × 10−2)−5.136 (±0.897)3
SK05–200051408 (±35.23)3.149 (±1.087 × 10−2)−11.16 (±2.746)3
SK05–200451208 (±28.06)3.082 (±1.009 × 10−2)−7.165 (±1.18)3
A. thaliana3245.8 (±0.133)2.389 (±1.492 × 10−4)−33.24 (±1.072)1
A. l. petrea3232.4 (±1.175)2.36 (±2.229 × 10−3)−30.13 (±1.272)1

Figure 3.  Population survival curves obtained from fitting to a sigmoidal dose–response model. A. h.: Arabidopsis halleri.

Download figure to PowerPoint

The T50 and B values obtained in 2000 and 2004 were very similar for the F01 population and slightly higher for SK05 in 2000 than in 2004 (Table 2). The GT2-test further attested that the LOGT50 of both populations did not differ between 2000 and 2004 (see below). Mann–Whitney tests showed that T50 values of NM and M A. halleri populations obtained from distinct data sets did not differ from each other (W = 41, P = 0.790 and W =18, P = 0.561 respectively). Finally, comparisons of B values and ni revealed that they did not differ between 2000 and 2004 (W = 246, P = 0.066 and W = 307.5, P = 0.069 respectively), in particular when M populations (over-represented in 2000 in comparison with 2004) were removed from the data sets (W = 112, P =0.317 and W = 196, P = 0.317 respectively). These statistical similarities between experiments allowed us to analyse both data sets simultaneously. As both replicates from SK05 and F01 were then considered separately, this led to 33 A. halleri populations being tested.

Survival rates of both control species (A. thaliana and A. l. petrea) severely dropped in the lowest exposure doses (Fig. 3). The respective T50 values were low and similar (Table 2). In our experiment, the T50 values obtained for control species should be considered to be indicative of the absence of Zn tolerance. In comparison, survival rates remained high to much higher Zn concentrations for all A. h. halleri populations (Fig. 3) and T50 values were from four to seven times higher than for control species (Table 2). This showed that all A. h. halleri populations had a high average level of Zn tolerance, irrespective of their edaphic origin. However, overall differences were observed between NM, NMp and M populations (Fig. 4a). Comparisons of T50 revealed that tolerance abilities differed significantly between types (Table 3a). Unilateral tests further showed that T50 values were significantly increased from NM populations to M ones, with an intermediate position for NMp populations (results not shown, but see Fig. 4a). Retrospectively, we ensured that such differences remained true in each data set (2000 and 2004) considered separately, at least when NM populations were compared with M ones (W = 11, P = 0.037 and W = 112, P = 0.032 respectively). It appeared that the range of T50 values was higher for NM populations [mean (m) = 1158.1; standard deviation (SD) = 140.6; coefficient of variation (CV) = 0.12] than for M and NMp ones (m = 1503.4; SD = 86.4; CV = 0.06 and m = 1363.2; SD = 74,8; CV = 0.06 respectively). This was confirmed by a Levene test of equality of variances performed using Minitab 13.20 (data not shown).


Figure 4.  Box plots of T50 and B obtained with Minitab 13.20 on the entire data set. Different letters above the box plots indicate a significant difference at the 5% level. M, metallicolous populations; NM, nonmetallicolous populations; NMp, nonmetallicolous populations in polluted area; ovirensis: A. h. ovirensis populations; *outlier (i.e. any data that is distant from the upper or lower quartile by more than 1.5 times the standard error).

Download figure to PowerPoint

Table 3.   Results of the two-tailed Mann–Whitney tests for comparisons between median values of (a) T50 and (b) slope factor B of the different A. h. halleri population types (NM, NMp, M) and the A. h. ovirensis populations.
PopulationsMedian value of T50 (μm)NMNMpM
 NMp1390W = 180, P = 0.008 
 M1495W = 153, P < 10−3W = 59, P = 0.035
 ovirensis1170W = 40, P = 0.514W = 8, P = 0.136W = 6, P = 0.023
 Median value of B   
  1. NM, nonmetallicolous; NMp, nonmetallicolous in polluted area; M, metallicolous; ovirensis, A. halleri subsp. ovirensis; W, Mann–Whitney statistic.

 NMp−9.078W = 84, P = 0.080 
 M−8.115W = 187, P = 0.005W = 48, P = 0.745
 ovirensis−5.887W = 19, P = 0.175W = 8, P = 0.136W = 6, P = 0.023

Comparisons of LOGT50 values confirmed that the control species did not differ from each other and that the Zn tolerance of all A. h. halleri populations tested was considerably higher (Fig. 5). The trend towards enhanced tolerance in M populations was also confirmed, with most M populations statistically differing from seven NM populations (out of 18). In particular, the tolerance of the PL06 population was statistically higher than the tolerance of most NM populations (16 of 18). However, twenty-one A. h. halleri populations (M, NM or NMp) could not be distinguished from each other. This demonstrated that, rather than discrete statistical categories, a continuum existed from the least tolerant population (A05, NM) to the most tolerant one (PL06, M). Finally, the test showed that the more tolerant NM populations (I07 and RO09) differed significantly from the less tolerant ones (from A05 to CH03), thus confirming a higher heterogeneity of tolerance abilities in this group.


Figure 5.  Comparison intervals for LOGT50given by the GT2-method of multiple comparisons. Different letters indicate a significant difference at the 5% level. Best fit values of control species, A. h. halleri and A. h. ovirensis are respectively represented by open circles, closed circles and closed triangles. Nonmetallicolous, nonmetallicolous in polluted area and metallicolous populations are respectively indicated in light grey, dark grey and black.

Download figure to PowerPoint

Comparisons of slope factors in A. h. halleri revealed that M populations had significantly higher slope factors than NM populations (Table 3b). Because of the wide range of values (Fig. 4b), NMp populations did not differ either from M or NM populations (although a unilateral test showed that NMp populations had higher B values than NM ones, W = 84, P = 0.04). This suggested an overall tendency towards a reduction of polymorphism around the median EC100 in populations occurring in polluted areas.

In comparison with A. h. halleri, populations of the A. h. ovirensis subspecies showed moderate to average tolerance abilities (Table 2). Tolerance in these populations was intermediate between NM and NMp A. h. halleri populations (Fig. 4a) from which they did not differ statistically (Table 3a). This was confirmed by the post hoc comparisons of LOGT50 that revealed that A. h. ovirensis populations did not differ from most of the A. h. halleri populations but did from the two extremes (A05 and PL06, see Fig. 5). The levels of polymorphism in the A. h. ovirensis populations tested were more similar to NM populations of A. h. halleri (Fig. 4b) and differed significantly from M populations (Table 3b).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. References

Constitutive zinc tolerance in the A. halleri population in Europe

Our survey confirmed that A. halleri is clearly pseudometallophyte in Europe. Most of its populations occurred in mountain areas, at medium to high altitudes, on soils with low levels of metal content. In contrast, in our sampling, both M and NMp A. h. halleri populations are not distributed throughout the species range but at the northern margin of the species range. However, most if not all individual we tested showed tolerance abilities and each population of A. halleri we tested showed high average levels of Zn tolerance, independently of the edaphic or geographic origin. We thus largely extended the results obtained from two M populations from northern France and three NM populations from Czech Republic by Bert et al. (2000) and strongly established that Zn tolerance is actually constitutive in A. h. subsp. halleri. Nevertheless, in a few populations, in particular in NM populations, survival rates fell slightly in the first exposure doses, suggesting that rare genotypes with low tolerance abilities may occur. Such an early stop in rooting could also be an artefact of the sequential test and would have to be confirmed by measurement of the heritability of the trait. An appropriate screening of within-population polymorphism using a constant Zn concentration (e.g. 250 μm) could be helpful in identifying putative nontolerant or low tolerant genotypes that would prove particularly useful in genetic studies of metal tolerance mechanisms (Macnair et al., 1999; Bert et al., 2003, G. Willems, C. Godé, D.B. Dräger, M. Courbot, N. Verbruggen, P. Saumitou-Laprade, unpublished data).

Zinc tolerance was not only shown to be constitutive in A. h. subsp. halleri but also extended to A. h. subsp. ovirensis. Although no current local heavy metal exposure was detected, we showed Zn tolerance and within-population polymorphism in A. h. ovirensis populations. Average tolerance of A. h. ovirensis populations was quite similar to those of A. h. halleri NM populations, i.e. populations that had definitely not been exposed to high levels of heavy metals in the recent times. Further studies should verify if Zn tolerance is not only constitutive in A. h. subsp. halleri but also species-wide in Europe, i.e. constitutive in both subspecies.

Origin of constitutive zinc tolerance in A. halleri

Constitutive metal tolerance is a rare phenomenon in pseudometallophytes (for a review, see Pollard et al., 2002). The most parsimonious hypothesis for the origin of constitutive tolerance would assume that it has been acquired only once. Given the very likely recent history of M populations, their marginal position and the extreme conditions they encountered on polluted soils, it is highly unlikely that metal tolerance spread from an initial M population to the entire species range (Levin, 2000). Conversely, tolerance is likely to have evolved once, at an early stage of the species history, i.e. much earlier than the recent colonization of man-made habitats (Westerbergh, 1994). Two major hypotheses remain equally possible. First, metal tolerance in A. halleri could have evolved in response to an early exposure of the species to metal, as has be been suggested for metalloendemics (Kruckeberg & Kruckeberg, 1990). In these species, the adaptation to metal-enriched soils (including metal tolerance acquisition) is supposed to have led to reproductive isolation (Kruckeberg, 1986; Macnair & Gardner, 1998) and metal tolerance is consequently constitutive. Such a hypothesis would be difficult to confirm as A. halleri populations have never been recorded as occurring on naturally metal-enriched soils. However, comparative phylogeography including both A. halleri and its close relatives (mainly A. lyrata petrea and A. arenosa), should be able to verify the likelihood of such a scenario. Secondly, Zn tolerance could have evolved in response to the hyperaccumulation trait. Metal hyperaccumulation (accumulation of metal in tissues to very high concentrations) is sometimes assumed to be a defence strategy against herbivores that could have evolved in the absence of metal exposure (Boyd, 1998). Evolution of hyperaccumulation may have provoked a co-evolution of sequestration capacity which, in return could have improved the tolerance abilities of the hyperaccumulator species. However, the relationships between both Zn and Cd tolerance and hyperaccumulation (Macnair et al., 1999; Bert et al., 2003) and the defensive role of Zn/Cd hyperaccumulation are still highly debated both in A. halleri and in T. caerulescens (Assunção et al., 2003a; Frérot et al., 2005; Noret et al., 2005). Clearly, the support of this hypothesis will require a better understanding of the genetic make-up of both characters, involving the identification of candidate gene (W. Willems, C. Godé, D.B. Dräger, M. Courbot, N. Verbruggen, P. Saumitou-Laprade, unpublished data).

Evolution of Zn tolerance in Europe

Apart from constitutive tolerance, we revealed significant differences in Zn tolerance abilities of individuals and populations. We made it clear that an actual quantitative polymorphism exists between individuals and that a continuum exists in A. h. halleri from the least tolerant population to the most tolerant one. Up to now, Zn tolerance genetic architecture in A. h. halleri has been studied through analyses of interspecific crosses with A. lyrata petraea: one genetic analysis assumed a control by only one major gene and modifiers (Macnair et al., 1999) and a recent QTL analysis detected three genomic regions of similar major effect (W. Willems, C. Godé, D.B. Dräger, M. Courbot, N. Verbruggen, P. Saumitou-Laprade, unpublished data). Therefore, the quantitative polymorphism we observed at the within-species level could be explained by allelic variation in genes present in the identified QTL regions or by additional genes with minor effects not detected in interspecific crosses. Hypostatic modifier genes have commonly been evoked to explain quantitative polymorphism in tolerance in several pseudometallophytes (Macnair, 1993), e.g. copper tolerance in Mimulus guttatus (Macnair, 1983; Smith & Macnair, 1998) and Silene vulgaris (Schat & Ten Bookum, 1992; Schat et al., 1993; van Hoof et al., 2001).

The observed quantitative polymorphism is a prerequisite for allowing diversifying selection to adapt populations to local conditions (Latta, 1998). Our results strongly attested that there are significant differences in average Zn tolerance among populations. Although no simple correlation could be established between T50 measured in controlled conditions and Zn concentration in soils of sites in which seeds were collected (data not shown), heavy metal exposure of populations clearly distinguished groups of populations. M populations were shown to be significantly more tolerant. They were also shown to have a reduced level of within-population polymorphism for Zn tolerance. It has to be noted, however, that the sequential test we used could have generated a bias towards an underestimation of variation in the more tolerant populations (Schat & Ten Bookum, 1992).

The analysis of cpDNA population structure in the northern part of the species range has shown that the populations we sampled in man-made polluted areas have been founded from NM ones located in geographically proximate mountains area (Pauwels et al., 2005). In the absence of strong founder effects (Pauwels et al., 2005), the combination of a higher tolerance level and reduced polymorphism in M populations indicates that natural selection could have acted during the colonization of polluted areas (Meerts & van Isacker, 1997). The location of the most tolerant populations (PL06, F01 and PL04) on the most polluted soils further suggests that the strength of selection was related to the local degree of heavy metal contamination of the soil. Thus, although it is constitutive, Zn tolerance appears to evolve secondarily towards enhanced abilities in M populations. According to Pauwels et al. (2005) who concluded that the M population in northern and western Europe had an independent origin, genetic mechanisms that have been locally selected could differ in separately founded M populations (Schat et al., 1996). Testing of such a hypothesis in A. halleri will require QTL mapping for tolerance in recombinant populations from multiple intra-specific crosses between an NM population and independently founded M populations.

Post hoc comparison of LOGT50 values demonstrated that many populations could not be significantly distinguished from each other. Comparison made it clear that, although the distinct types of populations differed according to tolerance abilities, they did not actually constitute discrete groups. In particular, an important quantitative polymorphism was observed between and within NM populations, with significant differences between the less and the more tolerant populations, and some NM populations were not distinguished from M ones. The occurrence of quantitative polymorphism among NM populations of A. h. halleri, i.e. in the absence of strong metal pressure, is more difficult to interpret in an evolutionary framework. The comparatively high level of tolerance detected in some NM populations (e.g. SK05, RO09, I07 and RO05) suggests that genetic mechanisms conferring enhanced tolerance could be present at relatively high frequency in populations that have not experienced metal exposure in recent times. Overall, the polymorphism observed among NM populations was not related either to the Zn content in soils (R2 = 0.002, P = 0.867) or to any available geographic variables. No correlation was found between T50 and either latitude (R2 = 2.10−4, P = 0.953) or altitude (R2 = 0.028, P = 0.505); a weak but significant correlation was found between T50 and longitude (R2 = 0.241, P = 0.038), although this result strongly suffered from the discontinuous distribution of NM populations according to longitude. Retrospectively, the absence of correlation between T50 and altitude in NM populations, which ranges from very low (160 m) to high (1530 m) elevations, attested that altitude could not explain the differences in tolerance abilities between NM and M populations whereas all M populations were at low elevation. Finally, our results also suggest that no selection acts against genetic mechanisms conferring enhanced tolerance in slightly or nonpolluted habitats (Harper et al., 1997b). Obviously, such a hypothesis will require more appropriate experiment (e.g. reciprocal transplantation) to be more rigorously supported.

The features of NMp populations in polluted areas should be understood with respect to the narrow genealogical relationships they share with surrounding M ones (Pauwels et al., 2005). A first assumption is that NMp populations have been founded from M ones and have simply inherited enhanced tolerance. The low metal concentrations in soils and the absence of the directional selection that acts on polluted sites could have secondarily resulted in lower average tolerance levels of NMp population when compared with M ones (Harper et al., 1997a, 1998). Conversely, NMp could have acquired enhanced tolerance from M population through gene flow moving metal tolerance genes. The discussion of both hypotheses will require a population genetic structure analysis at a local scale, ideally replicated in both polluted areas where NMp populations have been mentioned. A combination of both hypotheses is not impossible and would explain the large range of within-population polymorphism observed.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. References

The occurrence of constitutive metal tolerance in a pseudometallophyte species such as A. halleri is not in accordance with the classical model accounting for tolerance evolution in pseudometallophytes (Brooks, 1987). In this model, mainly elaborated for the so-called ‘mine taxa’ whose M populations mostly occur in anthropogenic metal polluted areas and which include most pseudometallophyte species, metal tolerance is expected to have evolved as an adaptive response to recent metal exposure, i.e. in M populations only. In contrast, in A. halleri, Zn metal tolerance could result from longer evolutionary history. As a result, the genetics of metal tolerance in these species could strongly differ from mine taxa. Thus, constitutive tolerance could distinguish A. h. halleri (as well as T. caerulescens, Ingrouille & Smirnoff, 1986; Meerts & van Isacker, 1997) from most pseudometallophytes (Pollard et al., 2002). The status of model species of these species is thus not straightforward and the extrapolation of results obtained from them will have to be done with caution.

Metal tolerance is usually defined as the ability to survive on metaliferrous soils (Antonovics et al., 1971; Macnair & Baker, 1994). Although tolerance tests involving root growth measurement are usually used to infer tolerance abilities of pseudometallophytes (Macnair, 1993), it has to be noted that root growth could only be part of the adaptation to metalliferous soils. Thus, tolerance abilities revealed in NM populations in the present paper do not indicate that all these populations would be able either to lastingly develop on polluted soils or to found M populations. Reciprocal translocations on polluted and nonpolluted sites from a single geographic region using different M and NM accessions characterized in this study are currently in progress in southern Poland. The measurement of life history traits involved in the overall fitness of individuals and including both vegetative growth and reproduction should provide a more integrated view of the capacities of NM and M genotypes to respectively colonize or escape polluted sites.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. References

The authors thank Drs Vlastimil Mikolas, Mihai Puscas, Slavomir Sokol, Maciej Szczepka, Thomes Wilhalm, Patrick de Laguerie, Laurent Amsellem and Profs Andreas Bresinsky, Krystyna Grodzinska and Konrad Pagitz for help in finding populations. They are very grateful to Josep Escarré, Henk Schat, Maria Clauss and two anonymous referees for helpful comments on the manuscript. This work was supported by funding from the Contrat de Plan Etat/Région Nord-Pas de Calais (PRC) for the ‘Arabidopsis’ project, from the European FEDER (contract no. 79/1769), from the Institut Français de Biodiversité (contract N °C. 00N55/0083), and from the INSU-CNRS program ACI ECCO (contract no. 04 2 9 FNS). M. Pauwels was funded by the French Ministry of Research and Technology.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. References
  • Al-Hiyaly, S.A., McNeilly, T. & Bradshaw, A.D. 1988. The effect of zinc contamination from electricity pylons – evolution in a replicated situation. New Phytol. 110: 571580.
  • Al-Shehbaz, I.A. & O'Kane, S.L.J. 2002. Taxonomy and Phylogeny of Arabidopsis (Brassicaceae). In: The Arabidopsis Book (C. R.Somerville & E. M.Meyerowitz, eds), doi/10.1199/tab.0001. American Society of Plant Biologist, Rockville.
  • Antonovics, J., Bradshaw, A.D. & Turner, R.G. 1971. Heavy metal tolerance in plants. Adv. Ecol. Res. 7: 185.
  • Assunção, A.G.L., Ten Bookum, W.M., Nelissen, H.J.M., Vooijs, R., Schat, H. & Ernst, W.H.O. 2003a. A cosegregation analysis of zinc (Zn) accumulation and Zn tolerance in the Zn hyperaccumulator Thlaspi caerulescens. New Phytol. 159: 383390.
  • Assunção, A.G.L., Schat, H. & Aarts, M.G.M. 2003b. Thlaspi caerulescens, an attractive model species to study heavy metal hyperaccumulation in plants. New Phytol. 159: 351360.
  • Becher, M., Talke, I.N., Krall, L. & Kramer, U. 2004. Cross-species microarray transcript profiling reveals high constitutive expression of metal homeostasis genes in shoots of the zinc hyperaccumulator Arabidopsis halleri. Plant J. 37: 251268.
  • Bert, V., Macnair, M.R., De Laguerie, P., Saumitou-Laprade, P. & Petit, D. 2000. Zinc tolerance and accumulation in metallicolous and non-metallicolous populations of Arabidopsis halleri (Brassicaceae). New Phytol. 146: 225233.
  • Bert, V., Bonnin, I., Saumitou-Laprade, P., De Laguerie, P. & Petit, D. 2002. Do Arabidopsis halleri from nonmetallicolous populations accumulate zinc and cadmium more effectively than those from metallicolous populations? New Phytol. 155: 4757.
  • Bert, V., Meerts, P., Saumitou-Laprade, P., Salis, P., Gruber, W. & Verbruggen, N. 2003. Genetic basis of Cd tolerance and hyperaccumulation in Arabidopsis halleri. Plant and Soil 249: 918.
  • Berton, A. 1946. Presentation de plantes Arabis halleri, Armeria elongata, Oenanthe fluviatilis, Galinsoga parviflora discoidea. Bulletin de la Société de Botanique de France 93: 139145.
  • Boyd, R.S. 1998. Hyperaccumulation as a plant defense strategy. In: Plants that Hyperaccumulate Heavy Metals (R. R.Brooks, ed.), pp. 181201. CAB international, Oxford.
  • Brooks, R.R. 1987. Serpentine and its Vegetation: A Multidisciplinary Approach. Dioscorides Press, Portland, Oregon.
  • Dahmani-Muller, H., Van Oort, F., Gelie, B. & Balabane, M. 2000. Strategies of heavy metal uptake by three plant species growing near a metal smelter. Env. Pollution 109: 231238.
  • Dräger, D.B., Desbrosses-Fonrouge, A.-G., Krach, C., Chardonnens, A.N., Meyer, R.C., Saumitou-Laprade, P. & Krämer, U. 2004. Two genes encoding Arabidopsis halleri MTP1 metal transport proteins co-segregate with zinc tolerance and account for high MTP1 transcript levels. Plant J. 39: 425439.
  • Ernst, W.H.O. 1990. Mine vegetation in Europe. In: Heavy Metal Tolerance in Plants: Evolutionary Aspects (A. J.Shaw, ed.), pp. 2137. CRC Press, Boca Raton, Florida.
  • Fabiszewski, J. 1986. Heavy metal tolerance of Cardaminopsis halleri (L.) Hayek populations in the Polish Tatra Mts. Acta Soc. Bot. Pol. 55: 421428.
  • Feder, M.E. & Mitchell-Olds, T. 2003. Evolutionary and ecological functional genomics. Nat. Rev. Genet. 4: 649655.
  • Frérot, H., Lefèbvre, C., Petit, C., Collin, C., Dos Santos, A. & Escarre, J. 2005. Zinc tolerance and hyperaccumulation in F1 and F2 offspring from intra and interecotype crosses of Thlaspi caerulescens. New Phytol. 165: 111119.
  • Harper, F.A., Smith, S.E. & Macnair, M.R. 1997a. Can an increased copper requirement in copper-tolerant Mimulus guttatus explain the cost of tolerance ? I. Vegetative growth. New Phytol. 136: 455467.
  • Harper, F.A., Smith, S.E. & Macnair, M.R. 1997b. Where is the cost in copper tolerance in Mimulus guttatus? Testing the trade-off hypothesis. Funct. Ecol. 11: 764774.
  • Harper, F.A., Smith, S.E. & Macnair, M.R. 1998. Can an increased copper requirement in copper-tolerant Mimulus guttatus explain the cost of tolerance? II. Reproductive phase. New Phytol. 140: 637654.
  • Van Hoof, N.A.L.M., Hassinen, V.H., Hakvoort, H.W.J., Ballintijn, K.F., Schat, H., Verkleij, J.A.C., Ernst, W.H.O., Karenlampi, S.O. & Tervahauta, A.I. 2001. Enhanced copper tolerance in Silene vulgaris (Moench) garcke populations from copper mines is associated with increased transcript levels of a 2b-type metallothionein gene. Plant Physiol. 126: 15191526.
  • Ingrouille, M.J. & Smirnoff, N. 1986. Thlaspi caerulescens J. & C. presl. (T. alpestre) in Britain. New Phytol. 102: 219233.
  • Jackson, R.B., Linder, C.R., Lynch, M., Purugganan, M., Somerville, S. & Thayer, S.S. 2002. Linking molecular insight and ecological research. Tree 17: 409414.
  • Koch, M., Mummenhoff, K. & Hurka, H. 1998. Systematics and evolutionary history of heavy metal tolerant Thlaspi caerulescens in Western Europe: evidence from genetic studies based on isozyme analysis. Biochem. Syst. Ecol. 26: 823838.
  • Koch, M., Haubold, B. & Mitchell-Olds, T. 2001. Molecular systematics of the Brassicaceae: evidence from coding plastidic matK and nuclear Chs sequences. Am. J. Bot. 88: 534544.
  • Kruckeberg, A.R. 1986. An essay: the stimulus of unusual geologies for plant speciation. Syst. Bot. 11: 455463.
  • Kruckeberg, A.R. & Kruckeberg, A.L. 1990. Endemic metallophytes: their taxonomic, genetic, and evolutionary attributes. In: Heavy Metal Tolerance in Plants: Evolutionary Aspects (A. J.Shaw, ed.), pp. 301312. CRC Press, Boca Raton, Florida.
  • Lambinon, J. & Auquier, P. 1964. La flore et la végétation des terrains calaminaires de la Wallonie septentrionale et de la Rhénanie aixoise. Types chorologiques et groupes écologiques. Natura mosana 16: 113130.
  • Latta, R.G. 1998. Differentiation of allelic frequencies at quantitative trait loci affecting locally adaptive traits. Am. Nat. 153: 283292.
  • Levin, D.A. 2000. The Origin, Expansion and Demise of Plant Species. Oxford University Press, Oxford.
  • Lexer, C. & Fay, M.F. 2005. Adaptation to environmental stress: a rare or frequent driver of speciation? J. Evol. Biol. 18: 893900.
  • Macnair, M.R. 1983. The genetic control of copper tolerance in the yellow monkey flower, Mimulus guttatus. Heredity 50: 283293.
  • Macnair, M.R. 1993. Tansley review N °C49: the genetic of metal tolerance in vascular plants. New Phytol. 124: 541559.
  • Macnair, M.R. & Baker, J.M. 1994. Metal-tolerant plant: an evolutionary perspective. In: Plants and the Chemical Elements, Biochemistry, Uptake, Tolerance and Toxicity (M. E.Farago, ed.), pp. 6883. VCH, NY.
  • Macnair, M. & Gardner, M. 1998. The evolution of edaphic endemics. In: Endless Forms, Species and Speciation (D. J.Howard & S. H.Berlocher, eds), pp. 157171. Oxford University Press, New York.
  • Macnair, M.R., Bert, V., Huitson, S.B., Saumitou-Laprade, P. & Petit, D. 1999. Zinc tolerance and hyperaccumulation are genetically independent characters. Proc. R. Soc. Lond. B 266: 21752179.
  • Macnair, M.R., Tilstone, G.H. & Smith, S.E. 2000. The genetics of metal tolerance and accumulation in higher plants. In: Phytoremediation of Contaminated Soil and Water (N.Terry, G.Banuelos & J.Vangronsveld, eds), pp. 235250. CRC Press, Boca Raton.
  • Meerts, P. & Van Isacker, N. 1997. Heavy metal tolerance and accumulation in metallicolous and nonmetalicolous populations of Thlaspi caerulescens from continental Europe. Plant Ecol. 133: 221231.
  • Mengoni, A., Barabesi, C., Gonnelli, C., Galardi, F., Gabbrielli, R. & Bazzicalupo, M. 2001. Genetic diversity of heavy metal-tolerant populations in Silene paradoxa L. (Caryophyllaceae): a chloroplast microsatellite analysis. Mol. Ecol. 10: 19091916.
  • Mitchell-Olds, T. 2001. Arabidopsis thaliana and its wild relatives: a model system for ecology and evolution. Tree 16: 693700.
  • Motulsky, H.J. 1999. Analyzing Data with GraphPad Prism. GraphPad Software Inc., San Diego, CA.
  • Noret, N., Meerts, P., Tolra, R., Poschenrieder, C., Barcelo, J. & Escarre, J. 2005. Palatability of Thlaspi caerulescens for snails: influence of zinc and glucosinolates. New Phytol. 165: 763772.
  • Pauwels, M., Saumitou-Laprade, P., Holl, A.-C., Petit, D. & Bonnin, I. 2005. Multiple origin of metallicolous populations of the pseudometallophyte Arabidopsis halleri (Brassicaceae) in Central Europe: the cpDNA testimony. Mol. Ecol. 14: 44034414.
  • Pollard, A.J., Powell, K.D., Harper, F.A. & Smith, J.A.C. 2002. The genetic basis of metal hyperaccumulation in plants. Cri. Rev. Plant Sci. 21: 539566.
  • Prat, S. 1934. Die Erblichkeit der Resistenz gegen Kupfer. Berichte der Deutschen Botanischen Gesellschaft 52: 6578.
  • Van Rossum, F., Bonnin, I., Fenart, S., Pauwels, M., Petit, D. & Saumitou-Laprade, P. 2004. Spatial genetic structure within a metallicolous population of Arabidopsis halleri, a clonal, self-incompatible and heavy-metal-tolerant species. Mol. Ecol. 13: 29592967.
  • Schat, H. & Ten Bookum, W.M. 1992. Genetic control of copper tolerance in Silene vulgaris. Heredity 68: 219229.
  • Schat, H., Kuiper, E., Ten Bookum, W.M. & Vooijs, R. 1993. A general model for the genetic control of copper tolerance in Silene vulgaris: evidence from crosses between plants from different tolerant populations. Heredity 70: 142147.
  • Schat, H., Vooijs, R. & Kuiper, E. 1996. Identical major gene loci for heavy metal tolerances that have independently evolved in different local populations and subspecies of Silene vulgaris. Evolution 50: 18881895.
  • Smith, S.E. & Macnair, M. 1998. Hypostatic modifiers cause variation in degree of copper tolerance in Mimulus guttatus. Heredity 80: 760768.
  • Sokal, R.R. & Rohlf, F.J. 1981. Biometry, 2nd edn. W. H. Freeman and Company, New York.
  • Vekemans, X. & Lefèbvre, C. 1997. On the evolution of heavy-metal tolerant populations in Armeria maritima: evidence from allozyme variation and reproductive barriers. J. Evol. Biol. 10: 175191.
  • Weber, M., Harada, E., Vess, C., Roepenack-Lahaye, E. & Clemens, S. 2004. Comparative microarray analysis of Arabidopsis thaliana and Arabidopsis halleri roots identifies nicotianamine synthase, a ZIP transporter and other genes as potential metal hyperaccumulation factors. Plant J. 37: 269281.
  • Westerbergh, A. 1994. Serpentine and non-serpentine Silene dioica plants do not differ in nickel tolerance. Plant and Soil 167: 297303.
  • Westerbergh, A. & Saura, A. 1992. The effect of serpentine on the population structure of Silene dioica (Caryophyllaceae). Evolution 46: 15371548.
  • Wu, L., Bradshaw, A.D. & Thurman, D.A. 1975. The potential for evolution of heavy metal tolerance in plants. III. The rapid evolution of copper tolerance in Agrostis stolonifera. Heredity 34: 165187.