SEARCH

SEARCH BY CITATION

Keywords:

  • cadmium (Cd);
  • genetic markers;
  • heavy metal;
  • plastid DNA;
  • post-glacial colonization;
  • zinc (Zn)

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  • • 
    Thlaspi caerulescens (Brassicaceae) is a promising plant model with which to study heavy metal hyperaccumulation. Population genetics studies are necessary for a better understanding of its history, which will be useful for further genomic studies on the evolution of heavy metal hyperaccumulation.
  • • 
    The genetic structure of 24 natural Swiss locations was investigated using nuclear and plastid loci. Population genetics parameters were estimated and genetic pools were identified using Bayesian inference on eight putatively neutral nuclear loci. Finally, the effect of cadmium (Cd) and zinc (Zn) soil concentrations on genetic differentiation at loci located in genes putatively involved in heavy metal responses was examined using partial Mantel tests in Jura, western Switzerland.
  • • 
    Four main genetic clusters were recognized based on nuclear and plastid loci, which gave mostly congruent signals. In Jura, genetic differentiation linked to heavy metal concentrations in soil was shown at some candidate loci, particularly for genes encoding metal transporters. This suggests that natural selection limits gene flow between metalliferous and nonmetalliferous locations at such loci.
  • • 
    Strong historical factors explain the present genetic structure of Swiss T. caerulescens populations, which has to be considered in studies testing for relationships between environmental and genetic variations. Linking of genetic differentiation at candidate genes with soil characteristics offers new perspectives in the study of heavy metal hyperaccumulation.

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Plant heavy metal hyperaccumulation is only found in c. 400 vascular plant species (< 0.2% of all angiosperms; Baker et al., 2000; Pollard et al., 2002) and this particular trait is currently being investigated in a few models, such as the cadmium (Cd)/zinc (Zn)/nickel (Ni) hyperaccumulator Thlaspi caerulescens (Alpine pennycress; Brassicaceae; Milner & Kochian, 2008). This species is a facultative metallophyte naturally distributed from northern Spain and Italy to the UK, Scandinavia and Poland (Tutin et al., 1993). The physiological, morphological and genetic characteristics of this taxon and its genetic proximity to Arabidopsis thaliana and Arabidopsis halleri (another heavy metal hyperaccumulator) make it an excellent candidate with which to study the genetic basis of Cd, Ni and Zn hyperaccumulation (Escarréet al., 2000; Assunçao et al., 2003; Peer et al., 2006; Roosens et al., 2008). The adaptive value of heavy metal hyperaccumulation is not fully understood in T. caerulescens but recent studies showed that related metallicolous and nonmetallicoulous populations are genetically differentiated for heavy metal tolerance, suggesting local adaptation to variable heavy metal contents in soil (e.g. Dechamps et al., 2007, 2008; Jiménez-Ambriz et al., 2007).

Thlaspi caerulescens is a biennial herbaceous plant. It is self-compatible and according to Riley (1956) freely cross-compatible within populations, although there is some incompatibility in inter-population crosses. The out-crossing rate of T. caerulescens shows strong inter-population variations (Koch et al., 1998; Dubois et al., 2003; Basic & Besnard, 2006). Pollination is preferentially assured by insects and the relatively heavy, thick-walled pollen is transported only short distances, so that gene migration by pollen is restricted (Riley, 1956). Barochore seed dispersal is also limited but the ripe raceme stems can break and the whole fruiting inflorescence may occasionally be blown over some distances (Riley, 1956).

During the last decade, most of the physiological and genetic studies of T. caerulescens have concerned only two main populations (i.e. ‘Ganges’ and ‘Prayon’), which have contrasting responses to heavy metals (e.g. Lombi et al., 2000). Recent studies have, however, suggested that other European populations may also be of interest in view of their various capacities to tolerate and hyperaccumulate some heavy metals (particularly Cd; Roosens et al., 2003; Basic et al., 2006a; Keller et al., 2006). Enlargement of the panel of T. caerulescens populations in physiological investigations is thus required to take into account the variation in heavy metal tolerance and hyperaccumulation in this taxon, for instance for the detection of polymorphism at genes of interest as well as for quantitative genetic studies (e.g. Assunçao et al., 2006; Deniau et al., 2006).

The post-glacial history of T. caerulescens is still unknown, but such knowledge is likely to be very useful in determining the origins of heavy metal hyperaccumulation in this species (e.g. Pauwels et al., 2005, 2008). Indeed, for studies of genomic local adaptation in natural populations, historical effects have to be clearly identified before one can test for relationships between environmental and genetic variations (e.g. Beaumont & Nichols, 1996; Thornsberry et al., 2001). In a recent study, Basic et al. (2006b) reported that genetic differentiation between Swiss T. caerulescens locations should be related to natural variation in soil contents for Cd. However, the conclusions of this preliminary study were limited by the small plant sample, and the putative occurrence of distinct lineages in Switzerland (and particularly in the Jura Mountains), which was not considered. Further investigations are thus necessary to identify potential historical effects on the genetic structure. In addition, the use of polymorphisms in candidate genes involved in heavy metal hyperaccumulation (such as in metal transporters; e.g. Basic & Besnard, 2006) and the comparison of their repartition to that of other loci considered as ‘neutral’ markers would be particularly relevant to the study of local adaptation to various soil contents of heavy metals in the field.

In the present study, we examined the genetic variation of T. caerulescens locations in western Switzerland where natural soil Zn and Cd concentrations were previously reported to be highly variable, particularly in the Jura Mountains (Basic et al., 2006a,b). A set of markers from both the nuclear and plastid genomes was used to characterize our plant samples and additional populations from France (Ganges, Saillagouse and Pla des Aveillans) and Belgium (Prayon). Our main objectives were (1) to determine the geographic structure of genetic diversity in Swiss locations and to identify divergent lineages, allowing the history of T. caerulescens in western Switzerland to be discussed in the light of the genetic data; and (2) to test for relationships between soil Cd and Zn concentrations and genetic differentiation at candidate gene loci (e.g. genes encoding transporters involved in metal uptake from the soil to the root) in the Jura Mountains, potentially reflecting local adaptation.

Materials and Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Plant sampling

Prospects in Switzerland (from May 2001 to July 2002) showed that Thlaspi caerulescens J. & C. Presl (or Noccaea caerulescens (J. & C. Presl) F. K. Mey) is particularly frequent in the Jura Mountains, whereas it is rare in the Alps and Prealps (Basic, 2006). The distribution of the species on the main Swiss mountains (from c. 900 to 1800 m) is thus very fragmented. Individuals of wild populations were collected from 24 locations in western Switzerland (Table 1): 17 from Jura (coded J1 to J16 and J18), two from Prealps (coded PA1 and PA2) and five from the Western Alps (coded A1 to A5). Between 16 and 25 individuals were randomly sampled in each location for a total of 509 individuals. Individuals were collected at least 1 m apart. A T. caerulescens location was defined as a patch (visually well delimited) in which each individual was isolated from its closer congener by a maximum of 10 m. The sampling of a location was performed on a maximum surface of 300 m2. Soil Cd and Zn concentrations were also previously characterized for each Swiss location using an HNO3 extraction method (Basic et al., 2006a,b; Table 1). Preliminary analyses of genetic data showed that five Swiss individuals (J14-4, J14-5, J14-6, A11-2 and A11-3) displayed a very different genetic pattern with specific alleles on most loci (compared with other Swiss individuals; data not shown). Because these individuals apparently belong to another taxon, they were excluded from the analysis and only 504 Swiss individuals were considered.

Table 1.  List of locations (and their codes) analysed in the present study
SiteLocalitynAltitude (m)Latitude (N)Longitude (E)S[Cd]S[Zn]
  1. The number of individuals analysed (n), altitude and coordinates for each site are given. For Jura locations, the soil concentrations of cadmium (S[Cd]) and zinc (S[Zn]) are indicated (data from Basic et al., 2006a,b). BE, Berner Oberland; NE, Canton de Neuchâtel; VD, Canton de Vaud; VS, Valais.

J1VD, St Cergues20110046°27′20″6°09′46″0.7 ± 0.151.0 ± 5.4
J2VD, Col du Marchairuz23133446°32′29″6°15′21″0.9 ± 0.051.5 ± 4.1
J3VD, Col du Marchairuz21134846°32′43″6°15′04″1.3 ± 0.190.4 ± 9.1
J4VD, Vallée de Joux25111446°36′24″6°12′54″1.8 ± 0.471.6 ± 8.3
J5VD, Col du Mollendruz24121946°38′44″6°21′26″1.0 ± 0.254.1 ± 7.7
J6VD, Dent de Vaulion23133946°40′48″6°21′18″1.5 ± 0.295.9 ± 4.2
J7VD, Le Suchet24122346°45′12″6°26′45″1.8 ± 0.279.7 ± 5.2
J8VD, Ste Croix21145346°50′25″6°31′57″4.8 ± 0.9102.1 ± 12.2
J9VD, Ste Croix23145046°50′10″6°31′25″3.7 ± 0.6119.5 ± 33.3
J10VD, Chasseron21137346°51′18″6°33′29″6.3 ± 1.186.5 ± 11.5
J11VD, Chasseron23134346°52′08″6°33′21″2.1 ± 1.266.2 ± 11.5
J12NE, Creux du Van18140046°55′05″6°43′29″2.0 ± 0.593.0 ± 4.7
J13NE, Mont d’Amin21132047°05′16″6°54′06″7.2 ± 3.8181.9 ± 59.0
J14NE, Mont d’Amin22131247°05′08″6°54′07″2.1 ± 0.8103.6 ± 49.7
J15NE, Mont d’Amin23139347°04′55″6°54′13″2.3 ± 0.392.3 ± 4.1
J16NE, Chasseral20151847°07′57″7°01′57″4.7 ± 0.4159.3 ± 38.3
J18NE, Tête de Ran16141447°03′15″6°51′15″2.6 ± 0.6108.6 ± 19.8
PA1VD, Leysin22148546°20′33″6°59′54″
PA2VD, Leysin21143446°21′19″7°01′12″
A1VS, Vallée de Saas20163246°09′50″7°55′27″
A2VS, Vallée de Saas20186346°05′20″7°57′42″
A3VS, Vallée de Saas19164046°05′44″7°56′53″
A4VS, Ulrichen19142446°30′14″8°19′12″
A5BE, Haslital20122546°39′22″8°18′01″
SGSaillagouse, Pyrenees, France10183442°25′06″2°06′14″
AVPla des Aveillans, Pyrenees, France11171042°32′39″2°02′52″
GLa Papeterie, Ganges, France23 18043°55′52″3°40′02″
PPrayon, Belgium21 16050°35′5°40′

Four additional locations from southern France and Belgium were included in our molecular analyses (Table 1): Saillagouse (SG; 10 individuals), Pla des Aveillans (AV; 11 individuals), Ganges (G; 23 individuals) and Prayon (P; 21 individuals).

Molecular characterization

DNA of each individual was extracted from 100 mg of leaf using the FastDNA kit (Qbiogene, Inc., Carlsbad, CA, USA). Individuals were characterized using the 11 polymorphic loci from Basic & Besnard (2006): three cleaved amplified polymorphic sites (CAPS) at loci TcE2F1, TcIRT2 and TcZNT2, seven microsatellites at loci TcIRT1, TcWRKY, TcAGA, TcCP, Tcup1, Tcup2 and TcbHLH and indels in nonrepeated motifs at locus TcZNT1. Six of these loci are located in genes (i.e. IRT1, IRT2, E2F1, WRKY, ZNT1 and ZNT2) putatively implicated in the heavy metal hyperaccumulation and tolerance responses (Basic & Besnard, 2006). Particularly, IRT and ZNT genes encode metal transporters while WRKY and E2F genes encode transcription factors. Three of the 14 loci developed by Basic & Besnard (2006) were not used: Tcup4 and TcNOD because of a high probability of null alleles, and Tcup3 because of some problems encountered with amplification. Three microsatellite polymorphic markers (Thlc1, Thlc2 and Thlc3) developed by Jiménez-Ambriz et al. (2007) were also used. Because the set of loci was different from that of Basic & Besnard (2006), genetic diversity observed for locations J1, J8, J12, A2, A5, G and P is different from that previously reported (see Results). The PCR protocol and polymorphism detection are described in Basic & Besnard (2006) and Jiménez-Ambriz et al. (2007).

Plastid DNA (ptDNA) markers were developed as follows. Eight individuals per location (224 individuals) were characterized using length variation at locus Bras3 (Parisod & Besnard, 2007), which showed polymorphism in a preliminary study (Basic et al., 2006b). This locus is located in the trnS-trnG intergenic spacer. Three length variants (i.e. 266, 297 and 309 bp) were revealed. The trnS-trnG region was then sequenced using the methodology described by Parisod & Besnard (2007) for eight individuals from distant locations (i.e. G-1, P-3, SG-1, A1-1, A5-3, J1-1, J13-1 and PA1-3). Three indels (explaining the length variation in Bras3) and two substitutions were identified. For each substitution, two new CAPS loci named trnS-G-MboI and trnS-G-TaqI were developed (Table 2). Amplifications were performed in a 20-µl total reaction volume with 1X Promega Go-Taq buffer (containing 1.5 mM MgSO4), 2.5 mM dNTPs, 10 µM of each primer (Table 2), 1 U Promega Go-Taq polymerase and c. 10 ng of genomic DNA. For trnS-G-TaqI, we added 1.5 mM MgSO4 to the PCR mixture (final concentration 3 mM MgSO4). Samples were amplified on a Biometra-T3 thermocycler (Biometra, Göttingen, Germany) with a denaturing step of 94°C for 180 s, 36 cycles of 94°C for 45 s, 45 or 53°C for 45 s and 72°C for 45 s, and a final elongation cycle of 72°C for 600 s. Ten µl of the PCR mixture was then restricted with 2 units of either MboI (trnS-G-MboI) or TaqI (trnS-G-TaqI) according to the manufacturer's recommendations (Fermentas, Vilnius, Lithuania). The absence or presence of restriction was revealed after migration on a 3% agarose gel and staining with ethidium bromide. The 224 individuals were characterized using these two loci. For each location, we then sequenced the trnS-trnG region of each detected haplotype in one individual in order to detect putative new polymorphism. A total of 45 individuals were sequenced for this plastid region. The sequences of the five identified haplotypes (see Results) were deposited in the EMBL databank under accession numbers FM178835 to FM178839.

Table 2.  Characteristics of two polymorphic plastid DNA regions of the trnS-trnG spacer, including primers used for the PCR and specific conditions (i.e. annealing temperature (Ta) and MgCl2 concentration)
LocusPrimersTa (°C)MgCl2 (mM)Allele size (bp)
  1. Single substitution polymorphism was revealed by digestion with either MboI or TaqI, and fragment sizes are given before and after restriction.

trnS-G-MboIForward: GATTCCTATCTAATGATCCAG531.5103 [RIGHTWARDS ARROW] 103
Reverse: ACATTTCCACTACACTAATTAGAT  103 [RIGHTWARDS ARROW] 81 + 22
trnS-G-TaqIForward: ATTATTTATTATTATATTCTGATC45384 [RIGHTWARDS ARROW] 84
Reverse: TATATAAATATATAAATAGCTTAG  84 [RIGHTWARDS ARROW] 60 + 24

Data analyses

Nuclear DNA data  A preliminary study (Basic & Besnard, 2006) suggested that the frequency of null alleles for the nuclear loci used is low in Swiss populations and thus that null alleles should not significantly affect the estimation of F-statistics. Weir & Cockerham (1984) estimators of fixation within locations (FIS), over all locations (FIT) and between locations (FST) were calculated at each locus and for all loci across all locations using the Fstat software (version 2.9.3; Goudet, 1995). Standard errors were obtained by jackknifing over locations. The mean allelic richness per locus (A), gene diversity (HT; Nei, 1987) and fixation index (FIS) were also estimated independently for each location using Fstat. The FIS significance was assessed using 10 000 permutations. To assess for a putative correlation between A or HT and Cd and Zn concentrations in soils, we used Kendall's rank correlation test (Kendall, 1938).

The number of genetically homogenous clusters (K) in the 24 Swiss locations was then determined using a model-based clustering method implemented in the Structure software (Pritchard et al., 2000). Bayesian analysis was run under the admixture model for 1 000 000 generations after a burn-in period of 500 000 generations. Ten iterations were performed for each K value between 1 and 15. The most likely number of clusters was determined using the log probability of data as well as the absolute values of the second-order rate of change of the likelihood distribution divided by the standard deviation of the likelihoods (ΔK) following the recommendations of Evanno et al. (2005). This analysis was performed on only eight nuclear loci excluding the six markers putatively involved in heavy metal responses (i.e. TcIRT1, TcIRT2, TcZNT1, TcZNT2, TcE2F1 and TcWRKY) in order to avoid putative selective effects as a result of variable Cd or Zn contents in soil.

Plastid DNA data  For the ptDNA multilocus profiles (assessed with CAPS and Bras3) detected in T. caerulescens, we estimated the gene diversity (HT) for each location using the Fstat software (version 2.9.3; Goudet, 1995). In addition, for each ptDNA profile, indel and nucleotide polymorphisms were coded as present/absent in a matrix, which was then used to reconstruct a reduced Median network with the Network software (Bandelt et al., 1999).

Test for a correlation between genetic and soil Cd and Zn concentration distances  Soil contents for Cd and Zn were previously reported to be highly variable in the Jura Mountains (Basic et al., 2006a,b; Table 1). Therefore, we tested for a correlation between genetic distances and soil heavy metal concentration differences in this mountain range. Establishing causation between a genetic polymorphism and variation of the phenotype is not feasible in natural conditions, because of many confounding factors. However, if the genetic differentiation observed at a locus between locations encountering different ecological conditions is significantly greater than expected by chance, one can conclude that the gene flow is reduced. This suggests that individuals inheriting genes from parents living in different conditions have a reduced reproductive success and consequently that the gene (or a gene physically linked to it) is under divergent selection.

Genetic differentiation between pairs of the 17 Jura locations was assessed by pairwise FST using Fstat (Goudet, 1995), for all the eight gene loci a priori not involved in heavy metal responses (‘neutral loci’: Tcup1, Tcup2, TcbHLH, TcAGA, TcCP, Thlc1, Thlc2 and Thlc3) and individually for each locus located in genes putatively involved in heavy metal responses (‘candidate gene markers’: IRT1, IRT2, E2F1, WRKY, ZNT1 and ZNT2). Partial Mantel tests (Smouse et al., 1986) were then used to test for the correlation of genetic and soil Cd and Zn concentration distances (based on data described in Basic et al., 2006a,b). For each candidate gene marker, a first linear regression was performed between pairwise FST inferred from the marker only and those calculated from the eight ‘neutral’ markers. Residues of this regression represent the part of the genetic variation at this locus that is not correlated with that of the total genome, approximated by the eight other loci. The correlations between these residues and pairwise Cd and Zn differences were then assessed through further linear regressions. Inter-location genetic structure resulting from historical or stochastic processes, such as isolation by distance or genetic drift, is expected to affect all loci similarly. Locus-specific genetic variation can thus not be attributed to such phenomena, excluding spurious correlations resulting from demographic processes. These calculations were performed using the Fstat software. The significance of Mantel tests per locus was assessed through 20 000 randomizations.

Note that, in this analysis, we did not consider the metal concentrations in plants reported by Basic et al. (2006a,b) because these measurements were not performed in controlled conditions but on individuals directly sampled in the field. A strong variation in plant Cd and Zn concentrations was observed among populations, and these contents moreover greatly depend on the soil Cd and Zn concentrations (Basic et al., 2006a). We thus consider that these values are not informative for our study.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Genetic characterization of locations

The number of alleles and fixation indices within locations (FIS), over all locations (FIT) and between locations (FST) at each locus and over all loci across all locations are presented in Table 3. We found between two and 27 alleles at each locus. Five microsatellite loci displayed more than eight alleles: Tcup1, Thlc1, Thlc2, Thlc3 and TcIRT1. By contrast, a maximum of four alleles was revealed at CAPS loci. At each locus, FIS, FIT, and FST values were significantly different from 0. Over all loci, the nuclear gene diversity and genetic differentiation between locations were very high (HT = 0.535; FST = 0.591 ± 0.024). FST per locus ranged from 0.515 (Thlc3) to 0.804 (TcIRT2).

Table 3.  Number of alleles (n), gene diversity (HT), and mean and standard deviation of fixation indices within locations (FIS), over all locations (FIT) and between locations (FST) at each locus and over all loci across all locations
LocusnHTFISFITFST
  1. a Microsatellite locus; b cleaved amplified polymorphic sites (CAPS) or indel locus. Gene abbreviations: AGA, floral homeotic protein AGAMOUS; bHLH, basic helix-loop-helix transcription factor; CP, coatomer protein; E2F1, E2F transcription factor; IRT1/2, Fe(II) transporters; up, unknown protein; WRKY, WRKY transcription factor; ZNT1/2, Zn transporters.

Tcup1a 90.6450.379 ± 0.0520.709 ± 0.0430.533 ± 0.067
Tcup2a 30.5140.369 ± 0.0990.721 ± 0.0620.558 ± 0.075
TcCPa 50.1060.211 ± 0.0060.799 ± 0.1210.746 ± 0.156
TcbHLHa 40.3350.421 ± 0.1280.753 ± 0.0670.574 ± 0.071
TcAGAa 20.3500.361 ± 0.0960.749 ± 0.0760.609 ± 0.110
Thlc1a 90.6980.384 ± 0.0730.747 ± 0.0400.591 ± 0.056
Thlc2a160.7760.426 ± 0.0750.740 ± 0.0470.547 ± 0.061
Thlc3a270.9090.383 ± 0.0500.701 ± 0.0410.515 ± 0.056
TcWRKYa 50.3780.419 ± 0.0770.790 ± 0.0810.640 ± 0.132
TcE2F1b 20.2620.446 ± 0.1670.783 ± 0.0910.610 ± 0.121
TcZNT1b 40.6310.426 ± 0.0760.805 ± 0.0410.661 ± 0.062
TcZNT2b 20.5000.389 ± 0.0790.738 ± 0.0550.572 ± 0.078
TcIRT1a130.7990.322 ± 0.0670.715 ± 0.0390.580 ± 0.049
TcIRT2b 40.5820.375 ± 0.0930.877 ± 0.0390.804 ± 0.060
Over all loci 0.5350.381 ± 0.0110.747 ± 0.0160.591 ± 0.024

Excluding locations J6, A1 and A2, which are nearly fixed (A < 1.2; HT < 0.03), FIS was estimated to be between 0.22 (J15) and 0.82 (J13; Table 4) among Swiss locations. All these values were highly significantly different from 0, suggesting a high degree of inbreeding. The FIS value was also significant for Prayon and Pla des Aveillans but not for Ganges and Saillagouse. In Switzerland, we also found that the allelic richness (A) was positively correlated to the Cd content in soils (z = 2.3359, P = 0.0195) but not to the Zn content (z = 1.787, P = 0.0739), although a highly significant correlation between Cd and Zn contents was revealed (z = 3.2285, P = 0.0012).

Table 4.  Description of the genetic diversity in the 27 Thlaspi caerulescens locations: for nuclear loci, the mean allelic richness per locus (A), gene diversity (HT) and fixation index (FIS) are given, whereas for plastid DNA (ptDNA), only the gene diversity (HT) is indicated
LocationNuclear DNAptDNA
AHTFISHT
  • nsNot significant; *P < 0.05; **P < 0.01.

  • a

    Excluding individuals J14-4, J14-5, J14-6, A1-12 and A1-13.

J11.820.250.348**0.00
J21.440.080.474**0.00
J31.550.110.368**0.00
J41.970.280.396**0.00
J51.490.160.338**0.00
J61.120.01−0.043ns0.00
J71.990.220.530**0.00
J82.530.310.319**0.68
J92.200.310.503**0.22
J101.760.210.547**0.25
J112.220.310.516**0.25
J122.050.300.298**0.00
J131.910.230.821**0.00
J14a1.820.190.589**0.00
J151.720.210.240**0.00
J161.880.160.464**0.00
J182.480.300.409**0.00
PA11.790.270.379**0.00
PA21.850.210.308**0.25
A1a1.110.03−0.007ns0.00
A21.180.02−0.016ns0.00
A31.610.270.188**0.54
A42.310.450.612**0.00
A51.970.280.222**0.00
SG2.000.170.118ns0.00
AV1.310.150.606**0.00
G2.550.320.087ns0.54
P2.120.290.277*0.00

In clustering analyses (for Swiss locations) using the Structure software, optimal cluster number based on ΔK of Evanno et al. (2005) was 2 (Fig. 1a). However, ΔK also showed a peak at K = 4, a K value at which the Loge(K) curve presented a slope break (Fig. 1a), a criterion important for the identification of the true K (Evanno et al., 2005). This suggests a hierarchical population structure, with two upper-level clusters and four lower-level groups, as shown for other taxa (e.g. Besnard et al., 2007). When the number of clusters was set to 2, a clear differentiation between a first group, composed of locations from South Jura (J1 to J7), Prealps (PA1 and PA2) and the Southern Alps (A1 to A4), and a second group, comprising North Jura locations (J8 to J18) and location A5, was revealed (Fig. 1b). The group composed of South Jura and Prealps locations persisted in the four clusters analysis. However, Alps locations A1 and A2 were not grouped with them and formed another cluster (Fig. 1b). North Jura was split into two clusters, according to the geographic repartition of the locations (Fig. 2).

image

Figure 1. Inference of population structure in western Swiss Thlaspi caerulescens locations based on 14 nuclear loci and using Bayesian simulations (Pritchard et al., 2000). (a) Mean log likelihood (loge(K) ± SD) averaged over the 20 iterations (upper graph) and absolute values of the second-order rate of change of the likelihood distribution divided by the SD of the likelihoods (ΔK in Evanno et al., 2005) (lower graph). (b) The percentage of assignment of each individual to the different clusters averaged over 20 iterations is shown for K = 2 and 4 clusters. Each vertical bar represents an individual.

Download figure to PowerPoint

image

Figure 2. Distribution of the genetic diversity in western Swiss Thlaspi caerulescens locations. (a) For the nuclear genome (nDNA), the mean assignment of individuals for K = 4 clusters (Fig. 1) is given for each location, whereas for the plastid genome (ptDNA), the frequency of each haplotype is indicated for each location (based on eight individuals). (b) The reduced Median network for ptDNA haplotypes is also given. Haplotype 1 was only detected in Ganges.

Download figure to PowerPoint

In some locations (in particular J5, J7, J18 and A3), several individuals were not clearly assigned to a cluster, and in the case of location A4 most individuals were assigned to either cluster 1a or cluster 1b (Figs 1b, 2). Moreover, if we consider the distribution of the genetic clusters in Jura and the Alps (Fig. 2), several of these locations were located in putative contact zones (such as A3, J7 and J18) or correspond to an isolated Alpine location (A4). We thus interpret these assignment failures as evidence of admixture or recent colonization by divergent lineages (e.g. A4). In addition, the gene diversity was relatively high in such patches compared with neighbouring locations (e.g. A3 vs A1 or A2; J18 vs J15 or J14; J7 vs J6; Table 4) probably as a consequence of admixture.

Plastid DNA variation

Five ptDNA haplotypes were detected in the whole sample based on sequences and multilocus profiles (Table 5). The amount of variation detected in T. caerulescens was low as only five characters were revealed based on sequence data (three indels and two nulceotidic substitutions). The more divergent haplotypes (haplotypes 1 and 5) were distinguished on four characters. Four ptDNA haplotypes were detected in the Jura Mountains (haplotypes 1, 2, 3 and 4; Fig. 2). Haplotype 1 was only detected in four (50%) individuals from the Ganges location. Individuals from the Prayon, Saillagouse and Pla des Aveillans locations, as well as the four other individuals from Ganges, were all characterized by haplotype 2. Overall, the ptDNA gene diversity and genetic differentiation between locations were very high (HT = 0.65; FST = 0.801 ± 0.055). A strong genetic structure was observed in Switzerland as the three genetic pools identified in the Jura Mountains using nuclear markers were also clearly supported by the distribution of ptDNA haplotypes (Fig. 2). Locations from South Jura were characterized by haplotype 2, those from Central Jura mainly by haplotypes 3 and 5, and lastly those from North Jura mainly by haplotype 4, with the exception of J16, which was characterized by haplotype 2. Locations A1 and A2 were characterized by haplotype 4, while the admixed location A3 from the Southern Alps displayed a mix of haplotypes 2 and 4. Alpine populations A4 and A5 displayed, respectively, haplotypes 2 and 5, and were thus related, respectively, to populations from South Jura/Southern Alps and Central Jura, as for nuclear markers (Fig. 2). Lastly, in Central Jura (locations J8 to J12), the ptDNA diversity was relatively high compared with other locations in this mountain range (Table 4). Interestingly, this was also linked to a relatively high nuclear genetic diversity in these locations (Table 4).

Table 5.  Plastid DNA multilocus profiles identified using the three plastid DNA (ptDNA) markers
Chlorotype numberBras3trnS-G-MboItrnS-G-TaqI
  1. For each locus, the fragment size is given in bp.*Compared with other alleles detected, allele 266 displays two specific indels based on sequence data.

1266*10384
229710384
330810384
429781 + 2284
529781 + 2260 + 24

Genetic differentiation of candidate gene markers in relation to soil metal concentrations

In the Jura Mountains, the genetic differentiation between pairs of locations (based on ‘neutral loci’) was related neither to Zn nor to Cd soil concentrations (Table 6), confirming that the genetic variance at these loci is neutral in terms of these soil metal concentrations. When this overall genetic differentiation was taken into account, the genetic differentiation at one marker (TcZNT2) showed a significant increase with higher soil Zn differences (Table 6), which remained even after Bonferroni correction for multiple tests. Interestingly, this locus is located in a gene encoding a Zn transporter and a higher frequency of allele TcZNT2-141 was observed in populations with high soil Zn contents (Supporting Information Table S1). In addition, three candidate markers (TcIRT1, TcIRT2 and TcE2F1) also showed a genetic differentiation significantly and positively associated with soil Cd concentration differences (Table 6). Except for marker TcIRT2, other relationships remained significant after Bonferroni corrections. Some alleles at these three loci, particularly TcIRT1-183, TcIRT2-99 and TcE2F1-187, were also found at high frequencies in populations with high soil Cd contents (Table S2). E2F1 is a gene encoding a transcription factor putatively involved in heavy metal responses (N. S. Pence & L. V. Kochian, unpublished data) while the genes IRT1 and IRT2 encode iron (Fe) transporters. These two latter genes are thought to be closely physically linked (as supported by the very highly significant linkage disequilibrium between TcIRT1 and TcIRT2; see Basic & Besnard, 2006), and IRT1 has been reported to be involved in Cd transport (Lombi et al., 2002). These results show that the gene flow observed at these loci is significantly reduced between plants encountering different heavy metal soil concentrations, which suggests divergent selection on particular alleles. Genetic differentiation at the TcZNT1 locus, located in a gene encoding a Zn transpoter with a low affinity for Cd (Lasat et al., 2000), did not display any significant association with Cd or Zn soil concentrations. This either results from a lack of power of the analysis or suggests that the function of ZNT1 is constitutive and does not depend on soil metal concentrations in the investigated Swiss populations.

Table 6.  Results of partial Mantel tests comparing genetic differentiation at each candidate gene marker and soil metal differences
LocusNeutral markersdS[Zn]dS[Cd]R2
  1. The genetic differentiation observed between pairs of locations (in the Jura Mountains) was corrected by the genetic differentiation observed at putatively neutral markers and tested against soil zinc differences (dS[Zn]) and cadmium soil differences (dS[Cd]). For each marker, the best model is presented and slopes are given for each factor included in the multiple linear regression, together with the significance (in bold). The percentage of variance of pairwise genetic differentiation explained by the best combination of genetic differentiation at neutral markers, dS[Zn] and dS[Cd] is indicated (R2).

  2. ns, not significant; *, P < 0.05; **, P < 0.01; ***; P < 0.001.

Neutral markersnsns
TcWRKYnsnsns
TcE2F1nsns0.41***17.21
TcIRT10.28**ns0.26**14.47
TcIRT20.50***ns0.21*29.77
TcZNT1nsnsns
TcZNT20.49***0.26**ns30.74

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Genetic variation in Swiss locations of T. caerulescens

As in previous genetic investigations (Koch et al., 1998; Dubois et al., 2003; Basic & Besnard, 2006; Jiménez-Ambriz et al., 2007), high fixation indexes (FIS values between 0.22 and 0.82; Table 2) were observed in Swiss locations, indicating frequent inbreeding events. Self-fertilization and population subdivisions (Walhund effect) on a small scale may explain these observations. This inbreeding coupled with limited gene flow between locations may also contribute to the relatively high genetic differentiation between locations (mean FST = 0.571) that was observed for the 14 loci (Table 3). The Swiss locations displayed mean allelic richness (A = 1.82 ± 0.39) and genetic diversity (HT = 0.22 ± 0.11) similar to those measured in French and Belgium locations (Table 4).

Central Jura locations (J8–J12) displayed a high genetic diversity compared with other Jura locations (Table 4). A large population size (avoiding genetic erosion) and putative ancient admixture could explain this observation. We also showed that a correlation between allelic richness and soil Cd content existed in our study area. Interestingly, soil Cd concentration is particularly high in Central Jura (Basic et al., 2006a,b), and could be an important factor in the maintenance of genetic diversity. Indeed, Dubois et al. (2003) reported a higher density of T. caerulescens in metalliferous locations than in nonmetalliferous locations, probably reflecting a higher effective population size.

This study also reports, for the first time, ptDNA variation in T. caerulescens. Only five variable sites in the trnS-trnG spacer allowed the identification of five chlorotypes. A study based on the same plastid intergenic spacer revealed a higher level of variation (30 variable sites leading to the identification of 21 haplotypes) in another species of Brassicaceae (Biscutella laevigata; Parisod & Besnard, 2007), with a similar plant sampling in western Switzerland. The lower ptDNA variation in T. caerulescens could be attributed to various factors, such as a lower maternal effective population size than in B. laevigata leading to higher genetic drift, and/or a more recent expansion of T. caerulescens in western Switzerland. Even if the trnS-trnG spacer is not very variable in T. caerulescens, substantial ptDNA diversity was detected in the Swiss locations. We thus recommend its use for further phylogeographic analyses of this species on a larger geographic scale.

On the colonization of T. caerulescens in Switzerland

In the Jura Mountains, three genetic pools were identified based on both nuclear and plastid DNA data (Fig. 2). There was also evidence of some putative contacts leading to exchanges between these genetic pools, in particular in the Suchet area (location J7). Interestingly, a similar pattern of plant distribution in the Jura Mountains was observed in the Anthoxanthum complex (Felber, 1986). The diploid form of Anthoxanthum alpinum was geographically separated from the tetraploid form around the Suchet area, as a result of the specific glacial extensions during the Würm period. Similarly, based on nuclear microsatellites and mitochondrial sequences, Dépraz et al. (2008) also reported three distinct groups in the snail Trochullus villosus from the Swiss Jura Mountains. The distribution of these groups was also very similar to those observed in T. caerulescens, and Dépraz et al. (2008) claimed that present T. villosus populations in the Jura Mountains had originated from at least two Last Glacial Maximum (LGM) refugia. We can thus speculate that the genetic pattern presently observed in T. caerulescens resulted from different recolonization routes and/or different LGM refugia located at the periphery of the Jura Mountains.

A fourth genetic pool (e.g. A1, A2, A3 and A4) was identified in the Southern Alps, probably resulting from colonization from northern Italy. In Alpine populations, which are very isolated and thought to have dispersed more recently than the Jura populations (because the Alps were more heavily glaciated during the Quaternary ice ages; e.g. Schönswetter et al., 2005; Parisod, 2008), the four lineages were observed and there was evidence of admixture in several places, particularly in the Saas and Ulrichen valleys (A3 and A4; Fig. 2). Recent human activities probably helped the species to disperse, as already suspected by Koch et al. (1998) for European locations, and could thus have played an important role in the mixing of different lineages during the colonization of the Alps.

As previously mentioned, T. caerulescens is a self-compatible species and a very high level of inbreeding can be observed (Koch et al., 1998; Dubois et al., 2003; Basic & Besnard, 2006). Moreover, its low dispersion capacities (Riley, 1956) and the isolation resulting from geographic barriers (e.g. mountain summits) are factors that can rapidly lead to the genetic fixation of populations (e.g. J6, A1 and A2). Nevertheless, our data support the theory that recurrent admixtures occurred during the recolonization (particularly in the Southern Alps) and this probably increased the success of T. caerulescens by limiting the negative effects of inbreeding. Thus, the admixed location A3 was particularly successful, with several thousands of flowering individuals in a very large continuous patch (> 5000 m2), compared with locations A1 and A2, which were composed of only a few (< 50) individuals in small patches (< 500 m2; N. Basic, personal observation).

This study has highlighted the importance of historical effects in the spatial organization of genetic diversity in T. caerulescens. Such factors should be taken into account in future studies with the aim of detecting genes under selection in relation to heavy metal soil concentration to avoid the finding of spurious relationships (e.g. Beaumont & Nichols, 1996; Thornsberry et al., 2001; Feder & Mitchell-Olds, 2003). On the basis of our molecular methodologies, a phylogeography should also be reconstructed for the whole distribution area of T. caerulescens in order to identify the different European lineages and to infer the post-glacial recolonization of the species at the continental scale, which has not yet been investigated in depth (Koch et al., 1998).

Soil heavy metal concentration and genetic structure in the Jura Mountains

Study of the genetic basis of heavy metal soil adaptation in T. caerulescens is important as it may facilitate the understanding of this adaptive trait (Antonovics et al., 1971; Milner & Kochian, 2008; Roosens et al., 2008). Unfortunately, the strong historical genetic structure observed in our study could hamper investigations of natural genetic variations. To avoid spurious correlations resulting from historical or demographic processes, we corrected the genetic differentiation observed for candidate gene markers by the expected differentiation based on putatively neutral markers. The significant positive relationship between these corrected genetic distances and soil metal differences shows that gene flow between locations living on metalliferous and nonmetalliferous soils is significantly reduced for some markers. Such locus-specific processes suggest strong divergent selective pressures on loci located near such markers. Thus, our candidate genes (or genes in linkage disequilibrium with them) significantly associated with soil metal contents (encoding metal transporters (IRT or ZNT) or transcription factors (E2F)) are potentially involved in the local adaptation to variable heavy metal contents in the soil. In particular, genetic variation in the proximity of TcIRT1 and TcIRT2, two closely linked loci (Basic & Besnard, 2006), is probably involved in the response to variable soil Cd contents. Interestingly, gene expression of IRT1 (encoding an Fe transporter) was shown to be strongly related to the Cd hyperaccumulation variation, and transcription of this gene was also enhanced by exposure to Cd (Lombi et al., 2002; Plaza et al., 2007). By contrast, IRT2 was not reported to be involved in the Cd response (Lombi et al., 2002; Plaza et al., 2007). Consequently, polymorphism in the promoter of IRT1 could be involved in the local adaptation to variable Cd content in soils, as previously suggested by physiological investigations (Lombi et al., 2002; Plaza et al., 2007). The fact that both TcIRT1 and TcIRT2 are significantly correlated to Cd soil content can be explained by a strong physical linkage between the two markers (Basic & Besnard, 2006) and a ‘hitchhiking’ effect.

Similar investigations on other candidate genes involved in metal transport and homeostasis (such as the HMA4 or MT genes; Milner & Kochian, 2008) or genes encoding unknown proteins and displaying particular transcription patterns in response to heavy metals (e.g. Hammond et al., 2006; Rigola et al., 2006; Van de Mortel et al., 2008) should be performed. New environmental variables as well as plant characteristics (such as the heavy metal tolerance and hyperaccumulation capacities of populations assessed in controlled conditions) should also be considered. A genomic scan approach (with mapped loci) seems to offer great promise for the detection of genes involved in local adaptation to heavy metal soil content, and this approach should be complementary to classical quantitative trait locus (QTL) analysis. Indeed, in such a study, a large panel of populations can be considered and this could reveal new genomic regions involved in the adaptation to heavy metal-polluted soils, which could not be detected in a specific cross (in the QTL approach). In addition, candidate genes located in genomic regions putatively involved in heavy metal responses could be further identified using genomic synteny with A. thaliana, as recently performed for QTLs detected in A. halleri (Pauwels et al., 2008). However, some factors, such as a high inbreeding rate, could lead to strong linkage disequilibrium between physically linked markers (Gupta et al., 2005) and thus to the identification of putatively large genomic blocks under selection in T. caerulescens. This could hamper the identification of genes involved in variations of the phenotype. Nevertheless, the combination of multiple approaches (including quantitative genetics, phylogeography and genomic scans) with classical physiological investigations should be very useful in determining the evolution of heavy metal tolerance and hyperaccumulation in different plant models such as T. caerulescens and A. halleri (see Pauwels et al., 2008).

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

The authors thank L. Degen for help in the laboratory, N. Salamin and G. Evanno for helpful discussions and C. Keller for her contribution to soil analyses and for providing plants from Ganges and Prayon. We are also particularly grateful to S. Dubois (ISEM, Montpellier) for providing primers of three microsatellite markers, and to Ph. Danton and A. Baudière for plant prospection in Southern France.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  • Antonovics J, Bradshaw AD, Turner RG. 1971. Heavy metal tolerance in plants. Advances in Ecological Research 7: 185.
  • Assunçao AGL, Pieper B, Vromans J, Lindhout P, Aarts MGM, Schat H. 2006. Construction of a genetic linkage map of Thlaspi caerulescens and quantitative trait loci analysis of zinc accumulation. New Phytologist 170: 2132.
  • Assunçao AGL, Schat H, Aarts MGM. 2003. Thlaspi caerulescens, an attractive model species to study heavy metal hyperaccumulation in plants. New Phytologist 159: 351360.
  • Baker AJM, McGrath SP, Reeves DR, Smith JAC. 2000. Metal hyperaccumulator plants: a review of the ecology and physiology of a biological resource for phytoremediation of metal-polluted soils. In: Terry N, Banuelos G, eds. Phytoremediation of contaminated soils and water. Boca Raton, FL, USA: CRC Press.
  • Bandelt HJ, Forster P, Röhl A. 1999. Median-joining networks for inferring intraspecific phylogenies. Molecular Biology and Evolution 16: 3748.
  • Basic N. 2006. Genetics of cadmium hyperaccumulation variance in Thlaspi caerulescens natural populations from Switzerland. PhD thesis, Lausanne, Switzerland: University of Lausanne.
  • Basic N, Besnard G. 2006. Gene polymorphisms for elucidating the genetic structure of the heavy-metal hyperaccumulating trait in Thlaspi caerulescens J. & C. Presl and their cross-genera amplification in Brassicaceae. Journal of Plant Research 119: 479487.
  • Basic N, Keller C, Fontanillas P, Vittoz P, Besnard G, Galland N. 2006a. Cadmium hyperaccumulation and reproductive traits in natural Thlaspi caerulescens populations. Plant Biology 8: 6472.
  • Basic N, Salamin N, Keller C, Galland N, Besnard G. 2006b. Genetic differentiation of Thlaspi caerulescens natural populations in relation to their cadmium hyperaccumulation capacity. Biochemical Systematics and Ecology 34: 667677.
  • Beaumont MA, Nichols RA. 1996. Evaluating loci for use in the genetic analysis of population structure. Proceedings of the Royal Society of London, Series B 263: 16191626.
  • Besnard G, Christin PA, Baali-Cherif D, Bouguedoura N, Anthelme F. 2007. Spatial genetic structure in the Laperrine's olive (Olea europaea subsp. laperrinei), a long-living tree from central-Saharan mountains. Heredity 99: 649657.
  • Dechamps C, Lefèbvre C, Noret N, Meerts P. 2007. Reaction norms of life history traits in response to zinc in Thlaspi caerulescens from metalliferous and nonmetalliferous sites. New Phytologist 173: 191198.
  • Dechamps C, Noret N, Mozek R, Escarré J, Lefèbvre C, Gruber W, Meerts P. 2008. Cost of adaptation to a metalliferous environment for Thlaspi caerulescens: a field reciprocal transplantation approach. New Phytologist 177: 167177.
  • Deniau AX, Pieper B, Ten Bookum WM, Lindhout P, Aarts MGM, Schat H. 2006. QTL analysis of cadmium and zinc accumulation in the heavy metal hyperaccumulator Thlaspi caerulescens. Theoretical and Applied Genetics 113: 907920.
  • Dépraz A, Cordellier M, Hausser J, Pfenninger M. 2008. Postglacial recolonization at a snail's pace (Trochulus villosus): confronting competing refugia hypotheses using model selection. Molecular Ecology 17: 24492462.
  • Dubois S, Cheptou PO, Petit C, Meerts P, Poncelet M, Vekemans X, Lefèvre C, Escarré J. 2003. Genetic structure and mating systems of metallicolous and nonmetallicolous populations of Thlaspi caerulescens. New Phytologist 157: 633641.
  • Escarré J, Lefèbvre C, Gruber W, Leblanc M, Lepart J, Rivière Y, Delay B. 2000. Zinc and cadmium hyperaccumulation by Thlaspi caerulescens J. et C. Presl. from metalliferous and nonmetalliferous sites in the Mediterranean area: implications for phytoremediation. New Phytologist 145: 429437.
  • Evanno G, Regnaut S, Goudet J. 2005. Detecting the number of clusters of individuals using the software Structure: a simulation study. Molecular Ecology 14: 26112620.
  • Feder ME, Mitchell-Olds T. 2003. Evolutionary and ecological functional genomics. Nature Reviews Genetics 4: 649655.
  • Felber F. 1986. Distribution des cytodèmes d’Anthoxanthum odoratum L. s. lat. en Suisse. Les relations Alpes – Jura. Botanica Helvetica 96: 145158.
  • Goudet J. 1995. FSTAT (Version 1.2): a computer program to calculate F-statistics. Journal of Heredity 86: 485486.
  • Gupta PK, Rustgi S, Kulwal PL. 2005. Linkage disequilibrium and association studies in higher plants: present status and future prospects. Plant Molecular Biology 57: 461485.
  • Hammond JP, Bowen HC, White PJ, Mills V, Pyke KA, Baker AJM, Whiting SN, May ST, Broadley MR. 2006. A comparison of the Thlaspi caerulescens and Thlaspi arvense shoot transcriptomes. New Phytologist 170: 239260.
  • Jiménez-Ambriz G, Petit C, Bourrié I, Dubois S, Olivieri I, Ronce O. 2007. Life history variation in the heavy metal tolerant plant Thlaspi caerulescens growing in a network of contaminated and noncontaminated sites in southern France: role of gene flow, selection and phenotypic plasticity. New Phytologist 173: 199215.
  • Keller C, Diallo S, Cosio C, Basic N, Galland N. 2006. Cadmium tolerance and hyperaccumulation by Thlaspi caerulescens populations grown in hydroponics are related to plant uptake characteristics in the field. Functional Plant Biology 33: 673684.
  • Kendall MG. 1938. A new measure of rank correlation. Biometrika 30: 9193.
  • 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. Biochemical Systematics and Ecology 26: 823838.
  • Lasat MM, Pence NS, Garvin DF, Ebbs SD, Kochian LV. 2000. Molecular physiology of zinc transport in the Zn hyperaccumulator Thlaspi caerulescens. Journal of Experimental Botany 51: 7179.
  • Lombi E, Tearall KL, Howarth JR, Zhao FJ, Hawkesford MJ, McGrath SP. 2002. Influence of iron status on cadmium and zinc uptake by different ecotypes of the hyperaccumulator Thlaspi caerulescens. Plant Physiology 128: 13591367.
  • Lombi E, Zhao FJ, Dunham SJ, McGrath SP. 2000. Cadmium accumulation in populations of Thlaspi caerulescens and Thlaspi goesingense. New Phytologist 145: 1120.
  • Milner JM, Kochian LV. 2008. Investigating heavy-metal hyperaccumulation using Thlaspi caerulescens as a model system. Annals of Botany (London) 102: 313.
  • Nei M. 1987. Molecular evolutionary genetics. New York, NY, USA: Columbia University Press.
  • Parisod C. 2008. Postglacial recolonisation of plants in the western Alps of Switzerland. Botanica Helvetica 118: 112.
  • Parisod C, Besnard G. 2007. Glacial in situ survival in the Western Alps and polytopic autopolyploidy in Biscutella laevigata L. (Brassicaceae). Molecular Ecology 16: 27552767.
  • Pauwels M, Saumitou-Laprade P, Holl AC, Petit D, Bonnin I. 2005. Multiple origins of metallicolous populations of the pseudometallophyte Arabidopsis halleri (Brassicaceae) in Central Europe: the cpDNA testimony. Molecular Ecology 14: 44034414.
  • Pauwels M, Willems G, Roosens N, Frérot H, Saumitou-Laprade P. 2008. Merging methods in molecular and ecological genetics to study the adaptation of plants to anthropogenic metal-polluted sites: implications for phytoremediation. Molecular Ecology 17: 108119.
  • Peer WA, Mahmoudian M, Freeman JL, Lahner B, Richards EL, Reeves RD, Murphy AS, Salt DE. 2006. Assessment of plants from the Brassicaceae family as genetic models for the study of nickel and zinc hyperaccumulation. New Phytologist 172: 248260.
  • Plaza S, Tearall KL, Zhao FJ, Buchner P, McGrath SP, Hawkesford MJ. 2007. Expression and functional analysis of metal transporter genes in two contrasting ecotypes of the hyperaccumulator Thlaspi caerulescens. Journal of Experimental Botany 58: 17171728.
  • Pollard AJ, Powell KD, Harper FA, Smith JA. 2002. The genetic basis of metal hyperaccumulation in plants. Critical Reviews in Plant Sciences 21: 539566.
  • Pritchard JK, Stephens M, Donnelly P. 2000. Inference of population structure from multilocus genotype data. Genetics 155: 945959.
  • Rigola D, Fiers M, Vurro E, Aarts MGM. 2006. The heavy metal hyperaccumulator Thlaspi caerulescens expresses many species-specific genes, as identified by comparative expressed sequence tag analysis. New Phytologist 170: 753766.
  • Riley R. 1956. The influence of breeding system on the genecology of Thlapsi alpestre L. New Phytologist 55: 319330.
  • Roosens N, Verbruggen N, Meerts P, Ximénez-Embún P, Smith JAC. 2003. Natural variation in cadmium tolerance and its relationship to metal hyperaccumulation for seven populations of Thlaspi caerulescens from western Europe. Plant, Cell & Environment 26: 16571672.
  • Roosens NCJ, Willems G, Saumitou-Laprade P. 2008. Using Arabidopsis to explore zinc tolerance and hyperaccumulation. Trends in Plant Science 13: 208215.
  • Schönswetter P, Stehlik I, Holderegger R, Tribsch A. 2005. Molecular evidence for glacial refugia of mountain plants in the European Alps. Molecular Ecology 14: 35473555.
  • Smouse PE, Long JC, Sokal RR. 1986. Multiple-regression and correlation extensions of the Mantel test of matrix correspondence. Systematic Zoology 35: 627632.
  • Thornsberry JM, Goodman MM, Doebley J, Kresovich S, Nielsen D, Buckler ES IV. 2001. Dwarf8 polymorphisms associate with variation in flowering time. Nature Genetics 28: 286289.
  • Tutin TG, Burges NA, Chater AO, Edmondson JR, Heywood VH, Moore DM, Valentine DM, Walters SM, Webbs DA. 1993. Flora Europaea, Vol. 1, 2nd edn. Cambridge, UK: Cambridge University Press.
  • Van de Mortel JE, Schat H, Moerland PD, Ver Loren van Themaat E, van der Ent S, Blankestijn H, Ghandilyan A, Tsiatsiani S, Aarts MGM. 2008. Expression differences for genes involved in lignin, glutathione and sulphate metabolism in response to cadmium in Arabidopsis thaliana and the related Zn/Cd-hyperaccumulator Thlaspi caerulescens. Plant, Cell & Environment 31: 301324.
  • Weir BS, Cockerham CC. 1984. Estimating F-statistics for the analysis of population structure. Evolution 38: 13581370.

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Table S1 Allele frequencies in the 17 Jura populations for the locus TcZNT2 showing genetic differentiation significantly and positively associated with soil zinc (Zn) concentration differences

Table S2 Allele frequencies in the 17 Jura populations for the three loci showing genetic differentiation significantly and positively associated with soil cadmium (Cd) concentration differences

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office.

FilenameFormatSizeDescription
NPH_2706_sm_TableS1-S2.doc146KSupporting info item