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

Authors

  • JUDITH E. VAN DE MORTEL,

    Corresponding author
    1. Laboratory of Genetics, Wageningen University, Arboretumlaan 4, 6703 BD Wageningen, the Netherlands,
    Search for more papers by this author
  • HENK SCHAT,

    1. Ecology and Physiology of Plants, Vrije Universiteit Amsterdam, de Boelelaan 1087, 1081 HV Amsterdam, the Netherlands,
    Search for more papers by this author
  • PERRY D. MOERLAND,

    1. Bioinformatics Laboratory, Department of Clinical Epidemiology, Biostatistics and Bioinformatics, Academic Medical Center, University of Amsterdam, PO Box 22660, 1100 DD Amsterdam, the Netherlands, and
    Search for more papers by this author
  • EMIEL VER LOREN VAN THEMAAT,

    1. Bioinformatics Laboratory, Department of Clinical Epidemiology, Biostatistics and Bioinformatics, Academic Medical Center, University of Amsterdam, PO Box 22660, 1100 DD Amsterdam, the Netherlands, and
    Search for more papers by this author
  • SJOERD VAN DER ENT,

    1. Plant-Microbe Interactions, Institute of Environmental Biology, Faculty of Science, Utrecht University, PO Box 800.84, 3508 TB Utrecht, the Netherlands
    Search for more papers by this author
  • HETTY BLANKESTIJN,

    1. Laboratory of Genetics, Wageningen University, Arboretumlaan 4, 6703 BD Wageningen, the Netherlands,
    Search for more papers by this author
  • ARTAK GHANDILYAN,

    1. Laboratory of Genetics, Wageningen University, Arboretumlaan 4, 6703 BD Wageningen, the Netherlands,
    Search for more papers by this author
  • STYLIANI TSIATSIANI,

    1. Laboratory of Genetics, Wageningen University, Arboretumlaan 4, 6703 BD Wageningen, the Netherlands,
    Search for more papers by this author
  • MARK G.M. AARTS

    1. Laboratory of Genetics, Wageningen University, Arboretumlaan 4, 6703 BD Wageningen, the Netherlands,
    Search for more papers by this author

M. G. M. Aarts. Fax: +31 317 483146; e-mail: mark.aarts@wur.nl

ABSTRACT

Cadmium (Cd) is a widespread, naturally occurring element present in soil, rock, water, plants and animals. Cd is a non-essential element for plants and is toxic at higher concentrations. Transcript profiles of roots of Arabidopsis thaliana (Arabidopsis) and Thlaspi caerulescens plants exposed to Cd and zinc (Zn) are examined, with the main aim to determine the differences in gene expression between the Cd-tolerant Zn-hyperaccumulator T. caerulescens and the Cd-sensitive non-accumulator Arabidopsis. This comparative transcriptional analysis emphasized the role of genes involved in lignin, glutathione and sulphate metabolism. Furthermore the transcription factors MYB72 and bHLH100 were studied for their involvement in metal homeostasis, as they showed an altered expression after exposure to Cd. The Arabidopsis myb72 knockout mutant was more sensitive to excess Zn or iron (Fe) deficiency than wild type, while Arabidopsis transformants overexpressing bHLH100 showed increased tolerance to high Zn and nickel (Ni) compared to wild-type plants, confirming their role in metal homeostasis in Arabidopsis.

INTRODUCTION

Cadmium (Cd) is a widespread, naturally occurring element present in soil, rock, water, plants and animals. It occurs naturally with deposits of zinc (Zn) and phosphorus (P), but unlike these nutrients, it is not essential for life. Instead, Cd is a health-threatening heavy metal pollutant. High Cd concentrations can result in bone disease and kidney damage in mammals, including humans. Cd uptake by humans occurs through food consumption, smoking and environmental exposure associated with Zn smelting, coal burning and the use of mineral phosphate fertilizers (Pinot et al. 2000).

Uptake of Cd by crop plants is the main entry pathway into the human food chain. Although plants generally try to prevent the uptake of Cd, the uptake depends on the availability of Cd to the plant, which largely depends on the Cd concentration in the soil and the pH of the soil. Cd is thought to be transported into the roots via Zn and iron (Fe) transporters belonging to the ZIP family or via Ca channels (Perfus-Barbeoch et al. 2002). Within the cell Cd is bound to sulphur ligands (Salt et al. 1995; Cobbett 2000; Hall 2002; Küpper et al. 2004) like glutathione (GSH), GSH-derived oligopeptides called phytochelatins (PCs) (Clemens 2006), or to organic acid ligands (Lugon-Moulin et al. 2004). The ligand–Cd complexes are most likely actively sequestered into the vacuoles of plant cells (Vögeli-Lange & Wagner 1990; Salt & Rauser 1995), probably through the activity of Cd/H+ antiporters, such as the divalent cation/H+ transporters (CAX family) (Hirschi et al. 1996), or ABC transporters (Bovet, Feller & Martinoia 2005). The further distribution of Cd to the shoots depends on the transport of ligand–Cd complexes through the xylem and the movement of Cd across the different membranes before xylem loading and after xylem unloading. Recently, Chen, Komives & Schroeder (2006) showed that ligand–Cd complexes are also involved in the long-distance shoot-to-root transport. Cd accumulation in plants affects water and nutrient uptake and photosynthesis and it results in growth inhibition, browning of the root tips, leaf chlorosis and finally death (Kahle 1993).

Although Cd is generally toxic to most plants, there are species that can accumulate high amounts of Cd without any sign of toxicity. Species accumulating metals in the above-ground tissues are so-called hyperaccumulators (Baker & Brooks 1989). Over 400 hyperaccumulator species from a wide range of unrelated families have been described. About 75% are Ni hyperaccumulators, 15 are Zn hyperaccumulators and currently only four species are known to hyperaccumulate Cd: Thlaspi caerulescens, Thlaspi praecox, Arabidopsis halleri and Sedum alfredii (Baker et al. 1992; Brooks 1994; Vogel-Mikuš, Drobne & Regvar 2005; Deng et al. 2006). They are mainly, though not exclusively, found on calamine soils contaminated with Pb, Zn or Cd (Meerts & Van Isacker 1997; Schat, Llugany & Bernhard 2000; Bert et al. 2002). Thlaspi caerulescens J. & C. Presl (Brassicaceae) is a self-compatible species, which is closely related to Arabidopsis thaliana (L.) Heynh. (Arabidopsis), with on average 88.5% DNA identity in coding regions (Rigola et al. 2006) and 87% DNA identity in the intergenic transcribed spacer regions (Peer et al. 2003). T. caerulescens accessions La Calamine (LC) and Ganges (GA) are more tolerant to Zn and Cd compared to the accessions Lellingen (LE) and Monte Prinzera (MP) (Assunção et al. 2003). Especially the Southern France T. caerulescens accessions like GA (Lombi et al. 2000) or St. Felix de Pallières (Peer et al. 2003) are strong Cd hyperaccumulators, with the ability to survive accumulation of more than 3000 mg Cd kg−1 in above-ground parts.

The complex network of homeostatic mechanisms that evolved in plants to control the uptake, accumulation, trafficking and detoxification of metals (Clemens 2001) also applies to metal hyperaccumulators. Previously published transcript profiling studies on the response to Zn or Cd in Arabidopsis (Wintz et al. 2003; Kovalchuk et al. 2005; Herbette et al. 2006; Weber, Trampczynska & Clemens 2006) and comparative analysis of Arabidopsis with the Zn and Cd hyperaccumulating species A. halleri and T. caerulescens (Becher et al. 2004; Weber et al. 2004, 2006; van de Mortel et al. 2006) already identified several genes for which expression changes in response to Zn or Cd exposure in these species. The analyses also revealed that the transcriptional regulation of many genes is strikingly different in A. halleri and T. caerulescens compared to A. thaliana.

In this paper we perform a between- and within-species comparison of transcript profiles of Arabidopsis and T. caerulescens plants exposed to different Cd concentrations, aiming to identify genes that are important for the adaptation of T. caerulescens to high Cd exposure. Therefore, we examined the response of roots of Arabidopsis and T. caerulescens to Cd excess and compared the transcript profiles with that caused by alterations in Zn supply.

MATERIALS AND METHODS

Plant material and growth conditions

Arabidopsis thaliana Col-0 (Arabidopsis) and Thlaspi caerulescens J. & C. Presl accession La Calamine were grown as described by van de Mortel et al. (2006). After 3 weeks of growth on a modified half-strength Hoagland's nutrient solution, the T. caerulescens plants were transferred for 7 d to the same modified half-strength Hoagland's nutrient solution (Schat, Vooijs & Kuiper 1996) with a deficient (0 µm), sufficient (100 µm) or excess (1000 µm) ZnSO4 concentration or a high (0.5, 5.0 or 50 µm) CdSO4 concentration. The Arabidopsis plants were transferred to the same nutrient solution with deficient (0 µm), sufficient (2 µm) or excess (25 µm) ZnSO4 concentration or a high (15 or 25 µm) CdSO4. The Cd-containing modified half-strength Hoagland's solutions with high CdSO4 concentration contained a sufficient (2 µm) ZnSO4 concentration. During the first 3 weeks, the nutrient solution was replaced once a week and thereafter twice a week. Germination and plant culture were performed in a climate chamber [20/15 °C day/night; 250 µmoles m−2 s−1 at plant level during 14 h/d (T. caerulescens) or 12 h/d (Arabidopsis); 75% relative humidity (RH)].

Root and shoot metal accumulation assay

Two pools of three plants, grown as described earlier, were used per treatment. After 4 weeks of growth, the plants were harvested after desorbing the root system with ice-cold 5 mm PbNO3 for 30 min. Roots and shoots were dried overnight at 65 °C, wet-ashed in a 4:1 mixture of HNO3 (65%) and HCl (37%) in Teflon bombs at 140 °C for 7 h and analysed for Cd, Zn, Fe and manganese (Mn) using flame atomic absorption spectrometry (Perkin Elmer 1100B, Perkin Elmer Nederland, Nieuwerkerk a/d IJssel, the Netherlands). Metal concentrations in roots and shoots were calculated as µmoles per gram dry weight. For the analysis of Arabidopsis grown at 0 µm CdSO4, 2 µm ZnSO4, we included the data previously reported (van de Mortel et al. 2006), which plants were grown at the same time in the same climate chamber under the same conditions.

Microarray experiment

The microarray experiment was performed as described by van de Mortel et al. (2006) using the Agilent Arabidopsis2 60-mer oligonucleotide microarrays (Agilent Technologies Inc., Palo Alto, CA, USA) representing ~80% of the Arabidopsis transcriptome. Roots of one pot containing three Arabidopsis or three T. caerulescens plants per treatment were pooled. Each pool was considered one biological replicate and two biological replicates were used for the microarray experiment. After hybridization the slides were scanned, analysed and normalized with the Agilent Feature Extraction Software (Agilent Technologies Inc.) using Agilent's standard normalization within each array. The remaining statistical analysis was performed using the Limma package (Smyth 2005a) in R/BioConductor (Gentleman et al. 2004). Between-array quantile normalization was performed on the common reference channel while leaving the log-ratios unchanged (Yang & Thorne 2003). To find differentially expressed genes, we performed a separate channel analysis (Smyth 2005b) using a moderated t-test (Smyth 2004). Within species all Zn and Cd treatments were compared. For the between-species comparison, each probe was tested for differences in the pattern of response to Zn or Cd in Arabidopsis as compared with T. caerulescens. The resulting P values were corrected for multiple testing using the Benjamini – Hochberg false discovery rate adjustment (Benjamini & Hochberg 1995). A gene was considered to be significantly differentially expressed if FDR < 0.05 (controlling the expected false discovery rate to no more than 5%) and the expression difference was ≥2 (within species) or ≥3 (between species) fold. Genes found to be significantly differentially expressed in all comparisons were clustered using Cluster/Treeview (Eisen et al. 1998). Average linkage clustering with uncentred correlation was used within cluster to perform the clustering analysis.

Genomic DNA hybridizations were performed using 1 µg random primed genomic DNA. As a quality control step we performed a dye-swap hybridization. After hybridization the slides were scanned, analysed and normalized with the Agilent Feature Extraction Software (Agilent Technologies Inc.) using Agilent's Linear & Lowess normalization. The features that hybridized with Arabidopsis genomic DNA (both polarities of the dye swap) and not with T. caerulescens genomic DNA were left out from the data set by Spotfire (Spotfire, Somerville, MA, USA) using a ≥3-fold change as cut-off value.

Semi-quantitative RT-PCR

The semi-quantitative RT-PCR was performed as described by van de Mortel et al. (2006). Selected T. caerulescens genomic and cDNA fragments were PCR-amplified using primers designed for the orthologous Arabidopsis gene. PCR products were sequenced and new primers were designed for semi-quantitative RT-PCR to ensure amplification of the correct T. caerulescens gene. The PCR amplification was performed with a cDNA aliquot (2 µL) and gene-specific primers (Supplementary Table S1). Between 25- to 35-cycle PCRs (30 s at 94 °C, 30s at 50 °C, and 60s at 68 °C) were performed in a 50 µL reaction and 50 µL of the reaction was separated on an ethidium bromide stained 1% agarose gel. Gel-image analysis using QuantityOne Software (BioRad, Hercules, CA, USA) was used to quantify the DNA fragment intensities (Supplementary Table S1). The DNA fragment intensities were corrected for background signal and corrected for cDNA quantity using the intensities of Tubulin.

Construction of the bHLH100 overexpression vector

The full-length cDNA of bHLH100 was amplified from Arabidopsis cDNA by PCR using Gateway primers : 5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTCTCA AAATGTGTGCACTTGT-3′ containing the attB1 sequence (underlined) and 5′-GGGGACCACTTTGTACA AGAAAGCTGGTACGAGAGACAAAACAGAAAA-3′ containing the attB2 sequence (underlined). The PCRs were performed with Pfx polymerase (Fermentas, http://www.fermentas.com) at 94 °C for 2 min, followed by 40 cycles 94 °C for 45 s, 50 °C for 45 s and 68 °C for 5 min, and finished by an extension at 68 °C for 5 min. The PCR product was recombined into pDONR201 (Invitrogen, http://www.invitrogen.com) in a 10 µL BP Clonase (Invitrogen) reaction following the manufacturer's instruction. The cDNA fragment was transferred from the pDONR201 construct into the binary overexpression vector pH7WG2 (Karimi, Inze & Depicker 2002), placing it under the control of the double 35S CaMV enhancer, in a 10 µL LR Clonase (Invitrogen) reaction following the manufacturer's instruction.

Plant transformation

The bHLH100 overexpression construct (35S::bHLH100) was used to transform Agrobacterium tumefaciens strain AGL0, and transformants were selected on Luria-Bertani medium containing spectinomycin (100 µg/mL). These A. tumefaciens were used to transform Arabidopsis Columbia-0 (Col) by the standard flower dip method (Clough & Bent 1998). T1 seeds obtained from self-fertilization of the primary transformants were surface-sterilized and sown on half-strength Murashige and Skoog's medium supplemented with hygromycin (20 µg/mL). Hygromycin-resistant plants were transferred to soil, and the T2 seeds resulting from self-fertilization were collected. The T2 seeds were surface-sterilized, plated on the same selection medium and scored for resistance to hygromycin. Transgenic lines that displayed a 3:1 segregation ratio for hygromycin resistance to sensitivity in the T2 generation and that were 100% hygromycin-resistant in the T3 generation were selected for further analysis. All further experiments were performed with T4 seeds.

Screening myb72 knockout mutants

Homozygous knockout Arabidopsis mutants SAIL_713G10 (Sessions et al. 2002), designated myb72-1, and SALK_052993 (Alonso et al. 2003), designated myb72-2, were identified to contain a T-DNA insertion in the MYB72 gene (At1g56160) (http://www.Arabidopsis.org). To confirm insertion and establish homozygous lines, seeds of wild-type A. thaliana Col-0, myb72-1 (SAIL_713G10) and myb72-2 (SALK_052993) were sown in quartz sand. Two-week-old seedlings were transferred to 60 mL pots containing a sand/potting soil mixture that had been autoclaved twice for 20 min with a 24 h interval. Plants were cultivated in a growth chamber with a 9 h day (200 µE m−2 s−1 at 24 °C) and a 15 h night (20 °C) cycle at 70% RH. Plants were supplied with modified half-strength Hoagland nutrient solution (Hoagland & Arnon 1938) once a week, as described (Pieterse et al. 1996). Confirmation of the T-DNA insert in SAIL_713G10 was obtained by PCR on genomic DNA using T-DNA left border primer T1 5′-TTTCATAACCAATCTCGATACACG-3′ and MYB72-specific primers MYB72-F1 5′-CGTTCCATGCTTTTGG GTCT-3′ and MYB72-R1 5′-TCGAGCAGGAACTGG ATCTC-3′ following a procedure that was described previously (Sessions et al. 2002). The SALK_052993 line was checked for the presence of T-DNA using primer T2 5′-GCGTGGACCGCTTGCTGCAACT-3′, and MYB72-specific primers MYB72-F2 5′-TGCTGCGACAAGA ACAAAGT-3′ and MYB72-R2 5′-TCATAGACATAC TTCTCCGACGAAA-3′. The exact insertion sites of the T-DNAs in myb72-1 and myb72-2 were determined by DNA sequencing of the PCR products. Seeds obtained after self-fertilizing of these plants were used to select homozygous insertion lines.

Metal tolerance screening

Seeds were sterilized and sown on the half-strength Murashige and Skoog's medium containing no Zn (0 µm ZnSO4), excess Zn (400 µm ZnSO4), low Fe (0.15 µm FeEDTA), excess Ni (75 µm NiSO4) or Cd (30 µm CdSO4). Seeds grown on half-strength Murashige and Skoog's medium were used as control. For each treatment, five to six replicates were performed and for each line 20 seeds were sown on each replicate plate. After sowing, seeds were vernalized at 4 °C for 3 d. Germination and plant culture were performed in a climate chamber (20/20 °C day/night temperatures; 250 µmoles light m−2 s−1 at plant level during 12 h/d; 75% RH). The tolerance phenotypes were studied on 5-day-old plants by measuring root length.

RESULTS

Experimental design

To analyse the response of Arabidopsis and T. caerulescens to different Cd exposures, we compared the transcript profiles of roots of plants grown under sufficient nutrient supply without Cd with those of plants grown under excess Cd conditions. To minimize variation in the bioavailability of micronutrients, we used a hydroponic culturing system. For excess Zn or Cd conditions, the induction of severe stress to the plants was avoided by exposing them only for 1 week to these conditions. Arabidopsis plants (accession Columbia) were grown on a nutrient solution containing 2 µm ZnSO4, which yields healthy and robust plants with normal seed set even after prolonged cultivation. After 3 weeks, the plants were transferred to fresh solutions for exposure to Zn deficiency (0 µm ZnSO4), excess Zn (25 µm ZnSO4) or excess Cd (15 µm CdSO4, 2 µm ZnSO4). One quarter of the plants remained as a control at sufficient Zn (2 µm ZnSO4). Upon root harvesting, the plants growing on Zn-deficient and Zn-sufficient medium did not show any visible phenotypic differences, while roots of plants growing on excess Zn or excess Cd showed a slight growth inhibition (data not shown). At this stage, plants were not yet flowering.

For T. caerulescens a similar approach was followed. However, in order to properly compare the results between the two species, we aimed at maintaining comparable physiological conditions. T. caerulescens accession La Calamine is Zn- and Cd-tolerant as well as Zn-hyperaccumulating and requires more Zn than Arabidopsis for normal growth. Therefore, a hydroponic solution containing 100 rather than 2 µm ZnSO4 was used as Zn-sufficient condition. To avoid any problems with possible precipitation of Zn or other minerals at very high Zn concentrations, 1000 µm ZnSO4 was used for excess Zn exposure and 0.5 or 50 µm CdSO4 for low-excess and high-excess Cd treatment, respectively. To avoid severe competition of Cd by Zn upon Cd exposure, which occurs in T. caerulescens (Zhao et al. 2002), we used 2 instead of 100 µm ZnSO4 in the Cd excess solutions. When root tissues were harvested after a 1 week exposure, T. caerulescens plants showed no phenotype that could be attributed to Zn deficiency, Zn excess or Cd excess exposure.

Mineral content in Arabidopsis and T. caerulescens

Cd, Zn, Fe and Mn concentrations were determined in root and shoots of hydroponically grown Arabidopsis and T. caerulescens plants grown at low-excess or high-excess Cd. Under lower Cd concentrations (0.5 and 5.0 µm CdSO4), the Cd content in T. caerulescens is similar to Arabidopsis grown under excess Cd supply (Fig. 1a). At high-excess Cd supply, T. caerulescens accumulates more Cd in the roots compared to Arabidopsis, and the concentration in shoots is comparable in both species.

Figure 1.

Cadmium (a), zinc (b), iron (c) and manganese (d) concentrations (µmoles g−1; mean ± SE) in Arabidopsis and Thlaspi caerulescens roots (white bar) and leaves (black bar). T. caerulescens were grown for 3 weeks on nutrient solution containing no cadmium before exposure to the same solution (0 µm CdSO4; Cd0), to low cadmium (0.5 CdSO4; Cd0), medium (5 µm CdSO4; Cd5.0) or to excess cadmium (50 µm CdSO4; Cd50). Arabidopsis were grown for 3 weeks on nutrient solution containing no cadmium before exposure to excess cadmium (15 or 25 µm CdSO4: Cd15/Cd25). We included data on Arabidopsis grown on medium without Cd (0 µm CdSO4; Cd0) from van de Mortel et al. (2006), in which plants were grown in the same experiment.

Because we expected that differences in Cd supply would also affect the concentration of other metals, we also measured the Zn (Fig. 1b), Fe (Fig. 1c) and Mn (Fig. 1d) concentrations in the same material. Comparison of the metal concentration levels between these two species already displayed the typical difference between a metal hyperaccumulator and a metal non-accumulator (Fig. 1b). T. caerulescens clearly contains more Zn in the leaves than in the roots. Arabidopsis accumulates much more Zn in the roots under excess Cd supply than in the absence of Cd, even more than T. caerulescens, and the concentration in shoots is similar for both species. At these very high Zn concentrations in the roots, Arabidopsis is apparently still able to prevent Zn accumulation in the leaves.

Similar to Zn, the Fe concentration in roots of T. caerulescens increases upon the increase of Cd supply (Fig. 1c-I). At excess Cd supply, the root Fe concentrations are similar for both species. Generally the Fe concentrations are much lower in leaves than in roots. The Fe concentration in leaves is similar for the four T. caerulescens treatments and this is only marginally lower in Arabidopsis under 25 µm CdSO4 (Fig. 1c-II). In Arabidopsis leaves, the Fe concentration decreases marginally with increasing Cd supply.

For Mn concentration in roots, the situation is opposite compared to Fe (Fig. 1d). In T. caerulescens, the Mn concentration decreases with increasing Cd supply and in Arabidopsis, the concentration is equal under the two conditions. The Mn concentration in T. caerulescens roots is about fivefold higher under low Cd supply than in Arabidopsis, but at 5.0 and 50 µm CdSO4 there is hardly any difference between the species. The Mn concentration in the leaves is similar for T. caerulescens and Arabidopsis under all conditions tested.

Cd-regulated genes in Arabidopsis

Genes responding to changes in Cd exposure conditions in roots of Arabidopsis were identified by microarray hybridization using Agilent Arabidopsis2 60-mer oligonucleotide microarrays containing 21 500 probes representing over 80% of the Arabidopsis genome. We only considered the hybridization data of probes with an FDR < 0.05 to be significant. Furthermore, only expression differences of ≥2-fold (between any of the four treatments: 0 µm ZnSO4, 2 µm ZnSO4, 25 µm ZnSO4, 15 µm CdSO4) were considered to be relevant, even though lower expression differences were statistically significant at FDR < 0.05. According to these criteria, we identified 30 genes with a response to Cd only. After hierarchical clustering of these Cd-responsive genes (average linkage hierarchical clustering with uncentred correlation; Eisen et al. 1998), two major clusters were distinguished (Supplementary Table S2). Cluster I consists of four genes, which are lower expressed upon exposure to Cd conditions. One of these genes encodes an urophorphyrin III methylase (UPM1), which is involved in porphyrin biosynthesis. The other three genes encode a phosphate-responsive protein, a nodulin-like protein and a MA3 domain-containing protein. Cluster II (Supplementary Table S2) consists of 26 genes, which are higher expressed because of Cd exposure. Genes in this cluster are involved in transport (CAX3, MATE efflux family protein), transcription (MYB107) or stress response (peroxidases). Other genes in this cluster encode pathogenesis-related proteins and genes involved in cell wall biosynthesis.

Cd- and Zn-regulated genes in Arabidopsis

In total, 48 genes are responding to both Zn and Cd in Arabidopsis when comparing Zn deficiency, Zn sufficiency, Zn excess and Cd exposure. After hierarchical clustering of these genes, five major clusters were distinguished (Table 1 + Supplementary Table S3). Cluster I contains 25 genes, which are higher expressed under excess Zn and Cd exposure. This cluster contains FRO2, which encodes a ferric-chelate reductase. Other genes in this cluster are involved in transcription (MYB4, MYB10, MYB72, ORG2, ORG3, bHLH100), lignin biosynthesis (4CL2, CYP71A20, CYP91A2), stress response and general (secondary) metabolism. Cluster II (Table 1 + Supplementary Table S3) contains five genes, which are higher expressed under Zn deficiency, Zn excess and Cd exposure, encoding for two germin-like proteins, the transcription factor WRKY59 and a protein kinase. The ribosomal protein RSU1 is higher expressed under Zn deficient and Cd excess exposure but not under Zn excess exposure. Cluster III (Table 1 + Supplementary Table S3) consists of 14 genes, which show decreased expression upon exposure to excess Zn and Cd. Five genes in this cluster encode members of the 2OG-Fe(II) oxygenase protein family, two encode members of the acyltransferase family, and four genes belong to the cytochrome P450 family (CYP76G1, CYP93D1, CYP712A1 and CYP712A2). This cluster also contains the ZIP2 gene, involved in Zn transport. The two genes in cluster IV (Table 1 + Supplementary Table S3) encode the metal homeostasis-related genes FRD3 (citrate efflux; Puig et al. 2007) and a nicotianamine synthase, NAS4, needed for biosynthesis of the metal-chelating compound nicotianamine (Shojima, Nishizawa & Mori 1989). Both genes show a similar change in expression profile in response to Zn, but in response to Cd the change in expression is different. FRD3 is repressed under Cd exposure, whereas NAS4 is higher expressed under Cd exposure. Cluster V (Supplementary Table S3) consists of two genes, encoding a hypothetical protein (At3g13950) and a flavonol synthase, which are lower expressed under Zn deficiency, Zn excess and Cd exposure compared to Zn sufficiency.

Table 1.  Cadmium- and zinc-responsive genes in Arabidopsis
ClusterNameCodeaPutative functionGO AnnotationbZn0/Zn2cZn0/Zn25cZn2/Zn25cZn2/Cd15cIntensityd
  • a

    AGI gene code (At..).

  • b

    GO annotations according to biological process.

  • c

    Ratio of significant (FDR < 0.05) differential (≥2) expressed genes between two zinc/cadmium exposure conditions. Zn0 = 0 µm ZnSO4, 0 µm CdSO4; Zn2 = 2 µm ZnSO4, 0 µm CdSO4; Zn25 = 25 µm ZnSO4, 0 µm CdSO4; Cd15 = 2 µm ZnSO4, 15 µm CdSO4.

  • d

    Normalized spot intensity at 2 µm ZnSO4. Genes are ordered according to decreasing spot intensity. Significant values are presented in bold (FDR < 0.05).

IFRO2AT1G01580Ferric-chelate reductaseIron chelate transport0.270.020.080.2714 321
MYB4AT4G38620Putative transcription factor (MYB4)Regulation of transcription1.050.450.430.441 530
4CL2AT3G21240Putative 4-coumarate:CoA ligase 2Lignin metabolism1.150.190.170.411 230
ORG1AT5G53450Protein kinase family proteinProtein amino acid phosphorylation0.790.120.160.301 085
bHLH100AT2G41240bHLH proteinRegulation of transcription0.260.010.020.08804
CYP71A20AT4G13310Cytochrome p450 familyLignin metabolism1.440.250.170.23729
ORG3AT3G56980bHLH proteinRegulation of transcription1.030.060.050.10590
ORG2AT3G56970bHLH proteinRegulation of transcription0.370.020.050.10379
MYB10AT3G12820myb family transcription factorRegulation of transcription0.530.050.100.30316
AT2G20030Putative RING zinc finger proteinBiological process unknown0.970.330.340.49279
MYB72AT1G56160myb family transcription factorRegulation of transcription0.420.030.060.27177
AT5G04150Basic helix-loop-helix (bHLH) family proteinRegulation of transcription0.850.160.190.21151
CYP91A2AT4G37430Cytochrome p450 familyLignin metabolism0.870.340.390.32111
IIWRKY59AT2G21900WRKY family transcription factorRegulation of transcription2.951.280.430.26973
RSU1AT3G48130Ribosomal protein L13 homologBiological process unknown4.064.000.990.1655
IIIZIP2AT5G59520Putative zinc transporter ZIP2-likeZinc ion transport1.468.015.473.2810 400
CYP76G1AT3G52970Cytochrome p450 familyLignin metabolism0.7011.5116.3711.461 227
IVFRD3AT3G08040MATE efflux family protein, putativeIron ion homeostasis4.157.061.702.20802
NAS4AT1G56430Nicotianamine synthase, putativeNicotianamine biosynthesis30.5813.080.430.32461

Heterologous microarray hybridization

We used the same Arabidopsis array platform for profiling of T. caerulescens gene expression through heterologous hybridization with labelled T. caerulescens cDNA. From Expressed Sequence Tag (EST) analysis we previously learned that T. caerulescens shares about 85–90% DNA identity in coding regions with Arabidopsis (Rigola et al. 2006), which should be sufficient for proper heterologous hybridization. Additionally, we recently hybridized Agilent Arabidopsis3 oligonucleotide microarrays with T. caerulescens cDNA, which showed reliable heterologous hybridization (van de Mortel et al. 2006). As an additional confirmation, we performed genomic DNA hybridization of T. caerulescens to the Agilent Arabidopsis2 oligonucleotide array, which showed on average a twofold lower signal intensity for T. caerulescens compared to the Arabidopsis signal intensities. Overall, only 94 probes in the Arabidopsis2 oligonucleotide array hybridized with less than threefold lower signal intensity with T. caerulescens genomic DNA. Previously we performed a similar genomic DNA hybridization to the Agilent Arabidopsis3 oligonucleotide array containing more probes (van de Mortel et al. 2006), which showed considerable overlap in both sets of poorly hybridizing probes. Based on both experiments, we excluded a total of 252 poorly hybridizing probes from the analysis.

Cd-regulated genes in T. caerulescens

To identify genes that are regulated by Cd in T. caerulescens, we compared the expression of genes in roots of T. caerulescens plants grown on low- and high-Cd-containing medium with those exposed to different Zn concentrations. Thus, we identified 171 genes that were significantly (FDR < 0.05) differentially expressed (≥2-fold) in response to Cd only (Supplementary Table S4). Only one of these genes (At2g18370) was also found to be differentially expressed in Arabidopsis in response to Cd (Supplementary Table S4). Seven major clusters were identified upon cluster analysis of the Cd-responsive genes (Supplementary Table S4).

Cluster I (Supplementary Table S4) consists of 49 genes, which are repressed by both low and high Cd exposure compared to the Zn-sufficient condition. Many genes in this cluster are involved in (a)biotic stress response, such as genes encoding the pyruvate decarboxylase-like protein family (PDC2), a wound-induced protein and ACD32.1, which encodes a heat shock protein. There are 10 genes with an unknown function in this cluster. Other genes identified in this cluster encode proteins involved in root development and signal transduction (CLE4 and CLE6; Fiers, Ku & Liu 2007), transcriptional regulation (ERF subfamily transcription factor, Zn finger family protein and a bHLH protein), protein metabolism and two genes involved in starch biosynthesis (APL3 and APL4). Cluster II (Supplementary Table S4) consists of 27 genes, which are repressed only by high Cd exposure compared to the sufficient Zn exposure. These genes are involved in Fe storage (AtFER1, AtFER3), transport (the membrane-trafficking protein PATL1, a sugar transporter and an anion exchange family protein), transcription (bHLH protein and Zn finger protein), lignin biosynthesis and protein metabolism. Cluster III consist of 13 genes, which are higher expressed under low Cd exposure compared to high Cd exposure (Supplementary Table S4). These genes are involved in (a)biotic stress response, transcription and transport. Cluster IV (Supplementary Table S4) consists of four genes, which are higher expressed under low Cd exposure compared to sufficient Zn. These genes (AtEXPA17) are involved in cell wall loosening and metabolism. Clusters V, VI and VII (Supplementary Table S4) consist of 11, 23 and 44 genes, respectively, which are higher expressed under high Cd exposure compared to low Cd exposure and sometimes higher expressed under low Cd exposure compared to Zn sufficiency. Within these clusters, several metal homeostasis-related genes are found, such as genes encoding the transporters AHA9 and ZIP8, but also a copper transporter and an oligopeptide transporter. Other genes in this cluster are involved in transcription (HAP5B, a bHLH protein and a myb family transcription factor), (a)biotic stress response and metabolism.

Cd- and Zn-regulated genes in T. caerulescens

In total, 109 genes were significantly (FDR < 0.05) differentially expressed (≥2-fold) in response to both Cd and Zn. Of these genes, three were also found to be differentially expressed in Arabidopsis in response to Cd (bHLH100, ZIP2 and At5g02780; Table 1 + Supplementary Table S3). After hierarchical clustering, seven clusters were identified (Supplementary Table S5). Cluster I (Supplementary Table S5) consists of only two genes, which are higher expressed under Zn deficiency compared to sufficient and excess Zn and low Cd exposure. These genes encode a legume lectin family protein and a NWMU3-2S albumin 3 precursor. The seven genes in cluster II (Supplementary Table S5) are higher expressed under Zn deficiency, Zn excess and high Cd exposure compared to Zn sufficiency and low Cd exposure. The genes in this cluster are involved in (a)biotic stress response (P5CS1), defence response (ATTI2) and senescence. Three genes in this cluster have no known function. In cluster III (Table 2 + Supplementary Table S5), 38 genes are co-expressed. These genes are higher expressed under Zn deficiency and high Cd exposure compared to sufficient and excess Zn exposure. Among these genes are APS3, APK, AKN1, AKN2 and the low-affinity sulphate transport gene SULTR2;1, which are all involved in sulphate assimilation (Takahashi et al. 1997). Also the genes encoding MYB28 and CYP83A1, which are involved in glucosinolate biosynthesis (Hirai et al. 2007) are identified in this cluster. Other genes co-expressed in this cluster are genes involved in (a)biotic stress response (ATGSTF3, ATGSTU17, CYP79C2, CYP81D1, SAL1 and peroxidases), transport [major intrinsic protein NIP6;1, carbohydrate transporter ZIFL, a phosphate transporter, a mitochondrial substrate carrier family protein (At2g17270)], transcription (ATHB5, involved in abscisic acid signalling) and metabolism. Cluster IV (Supplementary Table S5) consists of 11 genes, which are higher expressed under Zn deficiency and low or high Cd exposure compared to sufficient and excess Zn. Two genes in this cluster encode the metal transporters ZIP2 and ZIP9. Other genes in this cluster encode the high-affinity nitrate transporter WR3, a proton-dependent oligopeptide transporter and a wound-inducible protein. The 16 genes in cluster V (Supplementary Table S5) are higher expressed under excess Zn and low Cd exposure compared to Zn sufficiency and Zn deficiency. Of these 16 genes, one is involved in transcription (MYB111), three genes have an unknown function and three genes are involved in abiotic stress response (ATHSP22.0, CYP78A8, LEA14). Cluster VI (Supplementary Table S5) consists of 27 genes, which are higher expressed under Zn sufficiency, Zn excess, and low and high Cd exposure compared to Zn deficiency. The gene encoding bHLH100 is also co-expressed in this cluster and the expression profile is similar to that in Arabidopsis with expression under excess Zn and high Cd exposure. Other genes in this cluster are involved in (a)biotic stress response, cell organization and signal transduction. The genes in cluster VII (Supplementary Table S5) are higher expressed under excess Zn and low Cd exposure. Two of these six genes are involved in protein metabolism and two genes have an unknown function.

Table 2.  Genes higher expressed under zinc deficiency and high cadmium exposure in Thlaspi caerulescens
NameCodeaPutative functionGO AnnotationbZn0/Zn100cZn0/Zn1000cZn100/Zn1000cCd0.5/Cd50cZn0/Cd0.5cZn0/Cd50cZn100/Cd0.5cZn100/Cd50cZn1000/Cd0.5cZn1000/Cd50cIntensity Zn100d
  • a

    AGI gene code (At..).

  • b

    GO annotations according to biological process.

  • c

    Ratio of significant (FDR < 0.05) differential (≥2) expressed genes between two zinc/cadmium exposure conditions. Zn0 = 0 µm ZnSO4, 0 µm CdSO4; Zn100 = 100 µm ZnSO4, 0 µm CdSO4; Zn1000 = 1000 µm ZnSO4, 0 µm CdSO4; Cd0.5 = 2 µm ZnSO4, 0.5 µm CdSO4; Cd50 = 2 µm ZnSO4, 50 µm CdSO4.

  • d

    Normalized spot intensity at 100 µm ZnSO4. Significant values are presented in bold (FDR < 0.05).

 AT1G74100Sulfotransferase family proteinOther biological processes3.052.080.680.732.011.480.660.480.970.7113 733
SUR1AT2G20610Aminotransferase, putativeSignal transduction3.252.050.630.662.291.510.710.461.120.7413 163
AT5G43180Expressed proteinOther biological processes2.171.850.850.801.100.880.510.410.600.4812 528
AT1G18590Sulfotransferase family proteinOther biological processes3.122.260.730.691.871.300.600.420.830.579 570
AT4G33420Peroxidase, putativeResponse to oxidative stress2.612.150.820.681.801.230.690.470.840.579 299
AT1G74090Sulfotransferase family proteinOther biological processes3.722.300.620.781.911.490.510.400.830.659 052
AT2G41480Peroxidase, putativeResponse to oxidative stress2.031.640.810.601.671.010.820.501.020.617 578
CYP83A1AT4G13770Cytochrome p450 familyGlucosinolate catabolic process4.512.910.650.542.781.510.620.340.950.526 314
ATGSTF3AT2G02930Glutathione S-transferase, putativeToxin catabolic process3.691.800.490.592.131.250.580.341.180.705 705
AT1G20160Subtilisin-like serine proteaseProtein metabolism3.892.050.530.582.161.250.560.321.050.614 143
AT4G20070Allantoate AmidohydrolaseProtein metabolism2.181.950.890.842.281.911.040.381.170.393 168
AT4G28940Nucleosidase-relatedOther biological processes2.601.750.680.572.331.330.900.511.330.761 983
ATHB5AT5G65310Homeobox-leucine zipper proteinTranscription2.051.730.840.721.270.910.620.440.730.531 802
AT1G22190AP2 domain transcription factor RAP2, putativeTranscription2.061.380.670.721.140.830.550.400.830.601 771
AKN2AT4G39940Adenylylsulfate kinase 2Sulphate assimilation3.742.360.630.592.631.540.700.411.120.651 757
ZIFLAT5G13750Transporter-like proteinTransport4.142.440.590.622.491.550.600.371.020.641 693
AT4G23920UDPglucose 4-epimerase – like proteinGalactose metabolic process2.172.120.980.741.441.060.660.490.680.501 691
SAL1AT5G639803(2),5-bisphosphate nucleotidaseResponse to abiotic or biotic stimulus3.042.320.760.691.971.350.650.440.850.581 469
CYP79C2AT1G58260Cytochrome p450 familyElectron transport or energy pathways3.992.380.600.482.601.250.650.311.090.531 370
APS3AT4G14680ATP-sulfurylaseSulphate assimilation4.303.580.830.662.691.780.630.410.750.501 191
ATGSTU17AT1G10370Glutathione transferase, putativeToxin catabolic process9.618.350.870.623.862.390.400.250.460.291 174
AT4G11960Expressed proteinOther biological processes2.081.790.860.771.090.840.520.400.610.471 060
AT5G10770Chloroplast nucleoid DNA-binding proteinProtein metabolism2.542.220.870.522.301.210.910.471.040.54953
AT2G308302-oxoglutarate-dependent dioxygenaseOther biological processes2.612.280.880.631.781.130.680.430.780.49908
SULTR2;1AT5G10180Low-affinity sulfate transporterSulphate transport3.342.120.630.512.041.040.610.310.960.49811
AT5G056002OG-Fe(II) oxygenase family proteinOther biological processes2.861.710.600.711.511.080.530.380.890.63696
AT1G76430Phosphate transporter family proteinTransport2.281.990.870.961.161.110.510.490.580.56679
AT5G64700Nodulin MtN21 family proteinOther biological processes3.252.620.810.762.001.520.610.470.760.58631
AKN1AT2G14750Adenylylsulphate kinase 1Sulphate assimilation3.092.280.740.831.781.470.580.480.780.64622
NIP6;1AT1G80760Major intrinsic family proteinTransport2.031.850.910.701.390.980.690.480.750.53545
AT1G52700Phospholipase/carboxylesterase family proteinOther biological processes3.452.800.810.581.901.100.550.320.680.39489
AT3G20015Aspartyl protease family proteinProtein metabolism2.211.640.740.522.121.090.960.491.300.67435
AT3G15650Phospholipase/carboxylesterase family proteinOther biological processes2.582.510.970.582.031.190.790.460.810.47240
AT1G68830STN7 protein kinaseElectron transport or energy pathways2.101.720.820.611.490.920.710.440.870.53237
CYP81D1AT5G36220Cytochrome p450 familyElectron transport or energy pathways8.173.000.370.823.072.500.380.311.020.84233
AT1G11080Serine carboxypeptidase S10 family proteinProtein metabolism2.951.770.600.542.291.230.780.421.290.69232
MYB28AT5G61420myb-related transcription factorTranscription2.172.531.160.511.740.880.800.410.690.35208
AT2G17270Mitochondrial substrate carrier family proteinTransport2.142.171.010.671.470.980.690.460.680.45163

Difference in Cd and Zn response between Arabidopsis and T. caerulescens

To identify genes that may be crucial for the adaptive differences between Arabidopsis and T. caerulescens, we compared the gene expression profiles between the two species for each of the tested physiological conditions. To avoid the large group of genes that are constitutively differentially expressed between both species (van de Mortel et al. 2006), we only focussed on genes that show a different expression profile between the two species. In addition, we only considered probes to be of biological relevance if they differed significantly (FDR < 0.05) with a more than threefold in expression level between T. caerulescens and Arabidopsis. According to these criteria, 186 Zn-regulated genes were found to be significantly differentially expressed between both species (data not shown) and in total 42 Cd-regulated genes (Table 3). Of the latter, 34 genes were differentially expressed between different Zn/Cd exposures in T. caerulescens but constitutively expressed in Arabidopsis. Ten of these genes are involved in transcription. Other classes represent genes encoding for proteins involved in (a)biotic stress response and metabolism. Of the ZIP metal transporter genes, only ZIP1 is higher expressed under high Cd exposure compared to sufficient Zn condition while this was not observed in Arabidopsis. Six of the Cd-regulated genes were found to be differentially expressed between treatments in Arabidopsis but constitutively in T. caerulescens. Of these six genes, two are involved in protein metabolism.

Table 3.  Genes differentially expressed between Thlaspi caerulescens and Arabidopsis in response to cadmium Thumbnail image of

Next to differences in Cd-regulated genes in both species, there is also a group of genes of which the response to both Cd and Zn exposures is different in T. caerulescens compared to Arabidopsis. This group consists of 33 genes (Table 4), of which 17 are differentially expressed in T. caerulescens and 14 are differentially expressed in Arabidopsis. One of these genes is the Zn transporter gene ZIP2, which is higher expressed under Zn sufficiency compared to Zn excess and Cd excess conditions in Arabidopsis while this profile was not found in T. caerulescens. Furthermore, the ABC transporter ATATH13 is regulated in a different way in the two species. This gene is higher expressed under excess Cd in T. caerulescens, while in Arabidopsis this gene is not differentially expressed. Other genes differentially regulated between T. caerulescens and Arabidopsis are the transcription factors bHLH100, CCA1, LHY1 and WRKY59; (a)biotic stress responsive genes encoding the cytochrome P450 genes CYP83A1, CYP712A1, CYP76G1, the trypsin inhibitor ATTI2, and the glutathione transferases ATGSTF11 and ERD9; and genes involved in metabolism. CYP78A8 and FRO4 respond both in Arabidopsis and in T. caerulescens to Zn and Cd exposure, but in a different way. CYP78A8 is higher expressed under Zn-sufficient and Zn excess conditions in T. caerulescens, while this gene is only induced under Zn-deficient conditions in Arabidopsis. FRO4 is higher expressed under low and high Cd in T. caerulescens, whereas this gene is higher expressed under Zn-sufficient conditions in Arabidopsis.

Table 4.  Genes differentially expressed between Thlaspi caerulescens and Arabidopsis in response to cadmium and zinc Thumbnail image of

Semi-quantitative RT-PCR

For confirmation of the microarray expression profiling data, the expression of a small set of differentially expressed genes was determined by semi-quantitative RT-PCR. In the absence of T. caerulescens DNA sequences for designing species-specific PCR primers, orthologous T. caerulescens gene fragments were first amplified by low-stringency PCR using Arabidopsis-specific primers and their identity was confirmed by DNA sequencing. This sequence was used to design species-specific primers annealing to comparable positions in the T. caerulescens and Arabidopsis cDNA sequences, preferably flanking a predicted intron, for semi-quantitative RT-PCR. Expression of the target genes was studied in both root and leaf tissues of plants grown hydroponically at different Zn and Cd supply conditions (Fig. 2a,b). The root expression levels determined by semi-quantitative RT-PCR were comparable to those determined by microarray analysis, confirming the significance of the heterologous microarray hybridization results.

Figure 2.

Comparative semi-quantitative RT-PCR of differentially expressed genes in Arabidopsis (a) and Thlaspi caerulescens (b) in response to different zinc and cadmium exposures. For amplification, species-specific primers were designed at comparable locations in each orthologous gene pair. Roots and leaves were harvested separately after 1 week of exposure of 3-week-old plants to 0 µm ZnSO4, 2 µm ZnSO4, 25 µm ZnSO4, 15 µm CdSO4 and 25 µm CdSO4 for Arabidopsis and 0 µm ZnSO4, 2 µm ZnSO4, 10 µm ZnSO4, 100 µm ZnSO4, 1000 µm ZnSO4, 0.5 µm CdSO4, 5 µm CdSO4 and 50 µm CdSO4 for T. caerulescens. ATH13, ABC transporter, At5g64940; bHLH100, basic helix-loop-helix transcription factor, At2g41240; FRD3, MATE efflux family protein, At3g08040; FRO2, ferric chelate reductase-like, At1g01580; FRO4, ferric chelate reductase-like, At5g23980; NAS4, nicotianamine synthase, At1g56430; ORG1, protein kinase family protein, At5g53450; ORG3, basic helix-loop-helix transcription factor, At3g56980; WR3, wound-responsive gene 3, At5g50200; 2-oxoglutarate-dependent dioxygenase, At1g52820. Tubulin (At1g04820) was used as a control for equal cDNA use.

When considering the expression in leaves, there were some striking differences between Arabidopsis and T. caerulescens that were not observed in roots. FRD3 and NAS4 are more or less constitutively expressed in T. caerulescens leaves, while NAS4 is not expressed in Arabidopsis leaves and expression of FRD3 is limited to Zn deficiency. Furthermore, ORG3 is constitutively expressed in T. caerulescens while this gene is induced under excess Zn and Cd conditions in Arabidopsis leaves. ORG1 is induced under excess Zn and Cd conditions in both roots and leaves in Arabidopsis, but in T. caerulescens this gene is constitutively expressed in the leaves and induced to higher levels under excess Zn and Cd in the roots. The ferric-chelate reductase gene FRO4 is not expressed in leaves of Arabidopsis, while in T. caerulescens this gene is induced under low and high Cd exposure in both tissues. Remarkable is also that the nitrate transporter gene WR3, which is more or less constitutively expressed in Arabidopsis roots and Zn deficient leaves, is hardly expressed in T. caerulescens leaves and repressed in Zn-deficient roots. Finally, the transcription factor gene bHLH100 is differentially expressed when comparing the two species. In Arabidopsis roots and leaves, this gene is highly expressed under stress-inducing Zn excess and Cd exposure conditions, while the expression in T. caerulescens is lower and only detectable under high Cd exposure in both roots and leaves and under sufficient (100 µm ZnSO4), but not high, Zn in roots.

Transgenic Arabidopsis lines overexpressing bHLH100 exhibit increased tolerance to Zn and Ni

The expression of bHLH100 suggests a function related to stress tolerance in Arabidopsis and if this is true, lower expression in T. caerulescens at high Zn and Cd could mean this species experiences less stress at these conditions, in line with the relatively little adverse effects on T. caerulescens plant phenotype at longer exposures to high Zn and Cd. To establish if increased expression indeed enhances the stress tolerance, transgenic A. thaliana plants were generated overexpressing AtbHLH100 under the control of the constitutive Cauliflower Mosaic Virus 35S promoter. Overexpression plants did not show any obvious morphological alterations when they were grown in soil (data not shown). However, when seedlings of homozygous lines were grown for 5 d on vertical plates containing different heavy metals, roots of bHLH100 overexpression plants were longer compared to wild-type roots especially when exposed to excess Zn and excess Ni (Fig. 3), suggesting that there is indeed a positive effect on stress tolerance under these conditions.

Figure 3.

(a) Semi-quantitative RT-PCR analysis of bHLH100 expression in seedlings of homozygous transgenic Arabidopsis thaliana lines transformed with a CaMV 35S::bHLH100 construct. Tubulin amplification was used as control of the total RNA levels. RT-PCR was performed with 25 and 20 cycles, which appears to be in the linear range. (b) Root lengths (mean ± SE) of 5-day-old 35S::bHLH100 seedlings grown on 0.5 × MS (Murashige and Skoog) medium supplement with different Zn, Fe, Ni and Cd levels. Zn− = 0 µm ZnSO4; Zn+ = 400 µm ZnSO4; Fe− = 0.15 µm FeEDTA; Ni+ = 75 µm NiSO4; Cd+ = 30 µm CdSO4. *Significantly different from wild type (WT) at P < 0.05. **Significantly different from WT at P < 0.01. Significance was determined by one-way analysis of variance. Five replicates with 15 to 20 seedlings were measured for each treatment. Numbers indicate plant lines from two independent transformants.

The myb72 loss of function mutant exhibits increased sensitivity to Zn and Fe

MYB72 is another gene that was especially induced by excess Zn and Cd in Arabidopsis. Expression of an orthologue in T. caerulescens could not be established by microarray, nor by low-stringency semi-quantitative RT-PCR (data not shown), either because such gene is not present in T. caerulescens, or its sequence differs too much from the short probes and primers we used. To further examine the function of this transcription factor, the expression was first confirmed by semi-quantitative RT-PCR in Arabidopsis roots and shoots upon exposure to different Zn and Cd concentrations (Fig. 4a). This showed that MYB72 is only expressed in Arabidopsis roots under Zn and mild Cd excess, but not in leaves.

Figure 4.

(a) Semi-quantitative RT-PCR of MYB72 in Arabidopsis. Roots and leaves were harvested separately after 1 week of exposure of 3-week-old plants to 0 µm ZnSO4, 2 µm ZnSO4, 25 µm ZnSO4, 15 µm CdSO4 and 25 µm CdSO4. (b) Genomic DNA PCR amplification indicating homozygosity of the myb72-1 and myb72-2 mutants using forward (F) and reverse (R) primers flanking each T-DNA insert, and one T-DNA specific primer directed outward towards either a forward or reverse primer. Primer numbers (1, 2) indicate whether they are designed for the myb72-1 or myb72-2 allele. (c) Root length (mean ± SE) of 5-day-old myb72 seedlings grown on 0.5 × MS (Murashige and Skoog) medium supplement with different Zn and Fe levels. Zn− = 0 µm ZnSO4; Zn+ = 400 µm ZnSO4; Fe− = 0.15 µm FeEDTA. **Significantly different from wild type (WT) at P < 0.01. Significance was determined by one-way analysis of variance. Six replicates with 15 to 20 seedlings were measured for each treatment.

To verify the predicted T-DNA insertion sites of the selected homozygous myb72 knockout mutants (Fig. 4b), genomic DNA flanking the T-DNA insertion was PCR-amplified according to Sessions et al. (2002) (Fig. 4b). This confirmed the homozygous insertion of a T-DNA in both mutants. Sequence analysis located the T-DNA inserts at 51 and 736 bp downstream of the MYB72 translation start codon in myb72-1 and myb72-2, respectively (data not shown).

Two different myb72 T-DNA insertion lines were examined for any effect on metal tolerance when grown on vertical plates containing different heavy metals (Fig. 4c). Although in general the myb72 mutants had shorter roots than the wild-type plants, this reduction in growth was especially prominent under excess Zn and Fe deficiency conditions, suggesting that this gene is involved in Fe acquisition rather than in conferring general metal stress tolerance. When grown in soil, the myb72 mutants showed no visibly different morphological phenotype compared to wild type (data not shown).

DISCUSSION

Recently we described the gene expression profiles in roots of Arabidopsis and T. caerulescens in response to different Zn exposures, in order to identify genes involved in Zn tolerance and/or accumulation (van de Mortel et al. 2006). In the present study, we specifically investigated the response of Arabidopsis and T. caerulescens to Cd exposure. We anticipated identifying genes that are primarily involved in Cd tolerance. As Cd resembles Zn in chemical interactions, we also expected that genes normally involved in Zn homeostasis would transcriptionally respond to Cd exposure. When comparing gene expression in both species, we assumed that some of the genes differentially expressed between the two species, and especially those that show a difference in response to changes in the external Zn and/or Cd concentration, would be crucial to the adaptive difference between a hyperaccumulator and a non-accumulator.

The 21 500 probes on the Agilent Arabidopsis2 microarray that we used in this study are also present on the Agilent Arabidopsis3 oligonucleotide microarray that we used previously (van de Mortel et al. 2006). Of the 561 genes, which were found to be significantly differentially expressed in Arabidopsis in response to different Zn exposures in the present experiment, 170 genes were also previously found to be significantly differentially expressed in response to Zn (van de Mortel et al. 2006). For T. caerulescens, this is 93 of the 409 genes. When comparing Arabidopsis with T. caerulescens, 118 of the 186 genes we found to be differentially expressed in response to Zn in the present study were also found previously. So, although there is not a full overlap between both experiments, simply because of unavoidable differences in plant growth conditions, tissue sampling and probe hybridization, the two experiments compare well.

Weber et al. (2006) recently published the transcriptome analysis of A. thaliana and the hyperaccumulator A. halleri in response to a short 2 h Cd exposure. They reported many genes involved in stress response, like heat shock proteins and ethylene-responsive element-binding proteins, to be higher expressed in A. thaliana compared to A. halleri. These genes were not found in our analysis of A. thaliana after 1 week of exposure to 15 µm Cd. The reason is probably due to the difference in exposure time. We compared the root transcriptome of Arabidopsis and T. caerulescens plants exposed for 1 week to Cd, which avoids detection of the acute stress response that is expected after short Cd exposures and which allows adaptation of the plants to the new conditions. Knowing that T. caerulescens survives a long-term Cd exposure, whereas Arabidopsis does not, we assumed that a long-term exposure to low concentrations of Cd is more realistic to identify genes involved in Cd tolerance and accumulation than a short exposure to much higher concentrations.

The results of our analysis prompt us to conclude that (1) Cd has little effect on the regulation of the metal homeostasis mechanism in the plant; (2) Cd appears to induce Fe deficiency in Arabidopsis; (3) Cd has an effect on lignin biosynthesis in both species; (4) Cd has an effect on sulphate assimilation in T. caerulescens; and (5) a number of genes that are induced by Cd exposure in Arabidopsis are constitutively highly expressed in T. caerulescens. These conclusions will be discussed in more detail.

Cd effects on metal homeostasis-related genes in T. caerulescens and Arabidopsis

When we tested metal uptake, only Zn accumulation is clearly different between the two species, which is in line with the hyperaccumulating nature of T. caerulescens. At high Cd exposure, Arabidopsis roots are apparently no longer able to maintain a moderate Zn status (normally around 5–10 µmoles Zn g−1 dry weight), which induces strong accumulation of Zn in Arabidopsis roots. In contrast, there is only a very mild increase in the Zn concentration in T. caerulescens roots while the preferential translocation of Zn to the leaves is maintained (Fig. 1b). It would be interesting to see where the Zn is located in Arabidopsis roots at the (sub)cellular level at these high Cd supplies. The Zn concentrations in T. caerulescens leaves at these levels of Zn/Cd supply are not at the hyperaccumulator level, indicating that 2 µm Zn is a rather low Zn supply, not able to boost the full Zn hyperaccumulation trait.

Cd seems to compete only a little, if at all, with Zn translocation to the shoot despite the increasing accumulation of both Zn and Fe in the roots of T. caerulescens at increasing Cd concentrations (Fig. 1b,c-I). The latter is in line with previous observations that both in T. caerulescens and Arabidopsis the Zn and Fe concentrations in roots increase upon higher Zn supply (van de Mortel et al. 2006). These results suggest that the accumulation of Cd in roots is probably due to inadvertent uptake through Zn and/or Fe transporters with low affinity for Cd. This is similar to what has been found for A. halleri (Zhao et al. 2006). Like with increasing Zn supplies, the Mn accumulation in roots of T. caerulescens decreases upon exposure to higher Cd supplies (Fig. 1d) (van de Mortel et al. 2006).

Unlike the Zn response (van de Mortel et al. 2006), analysis of the Cd response indicated relatively few genes that were differentially expressed in Arabidopsis and T. caerulescens, and there was no clear process that is most affected by the Cd exposure. Because the LC accession of T. caerulescens we used is not hyperaccumulating Cd, our analyses again underline previous observations that Zn and Cd uptake are mediated at least in part by different molecular mechanisms (Zhao et al. 2002; Deniau et al. 2006). Cd exposure also has much less effect on Zn and Fe homeostasis genes than was previously observed upon alterations in Zn exposure and that was expected based on the root metal concentrations. In part this can be explained by the fact that Cd is non-essential and that there is no molecular mechanism to prevent deficiency. A complicating factor for Zn and Fe homeostasis is that high-affinity Zn transporters often have low affinity for Fe and the other way around, which necessitates the activity of a whole set of Zn/Fe metal transporters to compensate for the inadvertent uptake of one element and deficiency of the other. Even though Cd probably enters the plant cell through Zn/Fe transporters, the affinity is generally low, which causes only little competition of Cd uptake with Zn/Fe uptake. In addition, Herbette et al. (2006) found that gene expression in response to Cd was more time-regulated than dose-regulated in Arabidopsis, which could explain why fewer genes were differentially expressed in response to long-term Cd exposure.

Still, among the genes differentially expressed in Arabidopsis in response to Cd exposure are some genes known to be involved in Fe homeostasis, such as the FRO2 ferric-chelate reductase gene, a citrate efflux protein FRD3 (Durrett, Gassmann & Rogers 2007) and a nicotianamine synthase, NAS4 (Table 1 + Supplementary Table S3). The ferric chelate-reductase FRO2 is higher expressed both under excess Zn and excess Cd conditions, suggesting that the plants are experiencing Fe deficiency (Colangelo & Guerinot 2004; Mukherjee et al. 2005; Wu et al. 2005; Sarry et al. 2006) although another Fe deficiency marker, the IRT1 Fe uptake transporter, is not induced by Cd. In line with the latter is that the FRD3 expression is negatively affected by Cd exposure. Although this can be caused by a general down-regulation of metabolic processes by Cd exposure (Herbette et al. 2006), FRD3 was recently found to be involved in effluxing the metal chelator citrate into the root vasculature, enhancing root-to-shoot metal transport (Durrett et al. 2007). This suggests that plants are able to recognize Cd uptake through IRT1 (Rogers, Eide & Guerinot 2000) and Cd root-to-shoot movement through FRD3 and respond accordingly. The down-regulation of the Zn uptake transporter ZIP2 (Grotz et al. 1998), as we found, can be because of the same reason. Like citrate, nicotianamine (NA) is a chelator of metals. NA is made by nicotianamine synthases (NAS) for which Arabidopsis contains four genes. Although we cannot exclude that NA also chelates Cd, the predominant expression of NAS4 under Zn deficiency in roots suggests that the induced expression under Cd exposure is related to a modest response to Zn or Fe deficiency at high Cd exposure.

For T. caerulescens, the Cd response is slightly different from Arabidopsis, with more genes differentially expressed in T. caerulescens. Of the 280 genes that were differentially expressed in response to Cd (Supplementary Tables S4 & S5), 109 genes also responded to the Zn exposures (Supplementary Table S5). Only three of these genes were also significantly differentially expressed in response to Cd in Arabidopsis. The little overlap in expression between the two species is in line with the differences in Cd tolerance of both species. Like in Arabidopsis, we also identified metal homeostasis genes among the differentially expressed genes responding to Cd in T. caerulescens, but the expression profile of metal homeostasis-related genes in the Zn/Cd hyperaccumulator T. caerulescens was less affected by Cd exposure than by alterations in Zn exposure. T. caerulescens LC is known to be more tolerant to Cd but contrary to the Cd hyperaccumulator accession GA, it is a poor Cd accumulator (Assunção et al. 2003). Cd tolerance in higher plants is mainly studied in non-tolerant species and there is not much information on the mechanisms of metal tolerance in hypertolerant species (Schat et al. 2002). However, Cd tolerance is likely to be caused by changes in uptake, detoxification, bioavailability, cell wall properties and general stress tolerance.

Although the role of metal transporters may be limited in controlling Cd tolerance, two members of the ZIP gene family were differentially expressed in T. caerulescens in response to Cd. ZIP8 is induced under excess Cd in T. caerulescens (Supplementary Table S4), whereas this gene is only induced under excess Zn conditions in Arabidopsis. ZIP8 was found before to be higher expressed in Arabidopsis after exposure to excess Zn (van de Mortel et al. 2006) and as Zn excess induced more Fe acquisition genes because of Zn–Fe uptake competition, it was suggested that this gene is involved in Fe homeostasis. However, ZIP8 may also be involved in the transport of Mn or another transition metal that is typically transported by ZIPs and for which uptake may suffer from Zn/Cd competition. ZIP1 is another ZIP family relative (Table 3) that is differentially expressed between Arabidopsis and T. caerulescens. ZIP1 is induced by high Cd exposure in T. caerulescens, whereas this gene is not responding to Cd in Arabidopsis. As ZIP1 is also one of the major Zn uptake transporters (Grotz et al. 1998), this suggests that T. caerulescens may be more sensitive to competition of Zn uptake by Cd, especially under relatively low Zn supply of 2 µM Zn, or alternatively that the Zn deficiency signalling is disturbed because of Cd exposure, as was also described for Arabidopsis (Weber et al. 2006) and Schizosaccharomyces pombe (Chen et al. 2003).

Cd effects on lignin biosynthesis in Arabidopsis

The root cell wall is the structure, which is directly exposed to Cd. Genes involved in lignin biosynthesis such as a 4-coumarate: CoA ligase 2 (4CL2), peroxidases [At2g35380 (PER20), At2g18150] and cytochrome P450 (CYP71A20, CYP76G1, CYP91A2), and also the EXPR3 gene (At2g18660) involved in cell expansion, are higher expressed under excess Cd in Arabidopsis roots (Table 1, Supplementary Tables S2 & S3) suggesting that components to strengthen the cell wall are modified to protect the plants from Cd stress. Previously (van de Mortel et al. 2006) we found genes involved in lignin biosynthesis to be constitutively higher expressed in T. caerulescens compared to Arabidopsis roots exposed to Zn, but not in Arabidopsis in response to high Zn. We hypothesized that lignin biosynthesis prevents excess efflux of metals from the vascular cylinder in T. caerulescens by formation of an extra endodermal layer (van de Mortel et al. 2006). Although it is tempting to conclude the same from this study, the induction only in Arabidopsis, which does not have a preferential metal transport to the shoot as is found in T. caerulescens, could also indicate the activation of a Cd-specific detoxification mechanism that limits the entry of toxic metals in Arabidopsis roots, similar to what has been observed for Phragmites australis (Ederli et al. 2004).

Cd effect on sulphate assimilation in T. caerulescens

Genes involved in sulphate assimilation were only found to be differentially expressed in T. caerulescens in response to Cd exposure (Table 2, Supplementary Table S5). Sulphate is transported by high-affinity sulphate transporters from the soil into the root cells. Low-affinity transporters like SULTR2;1 (AST68) are required for translocation of sulphate within the plant. Intracellular sulphate can be metabolized into primary and secondary metabolites. ATP sulphurylase (APS), the first enzyme in sulphate assimilation, catalyses the formation of adenosine 5′-phosphosulphate (APS) from ATP and sulphate. APS is reduced to sulphite by APS reductase. Sulphite is further reduced by a ferredoxin-dependent sulphite reductase to sulphide, which is incorporated by O-acetylserine (thiol) lyase into the amino acid skeleton of O-acetylserine to form cysteine. Cysteine can be directly incorporated into proteins or peptides, such as glutathione (GSH) (Kopriva 2006). Sulphur is present in plant metabolites in the oxidized state as a sulpho-group, which play various roles in plant defence against biotic and abiotic stress, like Cd exposure. The transfer of the sulpho-group, i.e. sulphation, is catalysed by sulphotransferases (SOT) (Klein & Papenbrock 2004). The SOT reaction requires 3′-phosphoadenosyl 5′-phosphosulphate (PAPS), which is synthesized by phosphorylation of APS by APS kinase. Our results show an induction of SULTR2;1 under Zn deficiency and high Cd exposure compared to sufficient and excess Zn (Table 2, Supplementary Table S5). We did not find sulphate assimilation genes induced in Arabidopsis. Nevertheless a recent study by Herbette et al. (2006) reported that SULTR2;1 is induced in response to short-term Cd exposure in Arabidopsis. Most likely the expression in Arabidopsis is needed for a rapid Cd detoxification response and is normalized again after longer exposures. SULTR2;1 is specifically expressed in the central cylinder of roots and in the vascular tissues of leaves in Arabidopsis during sulphate starvation (Takahashi et al. 1997), suggesting that SULTR2;1 makes a significant contribution to the enhancement of the sulphate transport capacity in the vascular tissues of roots. This response could imply the first step of an adaptive process required to ensure an adequate supply in sulphur-containing compounds during Cd exposure. In Arabidopsis, three ATP sulphurylases (APS1, APS2 and APS3) are identified (Murillo & Leustek 1995). In our study APS3 is higher expressed in T. caerulescens under Zn-deficient and high Cd exposure compared to sufficient and excess Zn (Table 2, Supplementary Table S5). APS3 was also identified as being expressed differentially between A. thaliana and A. halleri when exposed to Cd (Weber et al. 2006). Two genes encoding APS kinases (AKN1 and AKN2) are also differentially expressed in T. caerulescens. Both are higher expressed under Zn-deficient and high Cd exposure compared to sufficient and excess Zn (Table 2, Supplementary Table S5). In addition, four T. caerulescens genes encoding glutathione S-transferase (GST) (ATGSTF3, ATGSTF10, ATGSTU17, ATGSTU28) (Table 2, Supplementary Table S5) are higher expressed under Zn-deficient and Cd excess conditions. GST conjugates endogenous and xenobiotic compounds like Cd with glutathione (GSH) and it is suggested that these complexes are transported to the vacuole via a family of GS-X pumps (Rea et al. 1998). GSH is one of the end products of the sulphate assimilation pathway. T. caerulescens clearly shows an activation of the sulphate assimilation pathway upon Cd exposure, as was also observed in other metal-accumulating Thlaspi species (Freeman et al. 2004). Our results suggest the existence of a transcriptional regulation mechanism of the sulphur metabolism pathway involved in response to Cd. The rate-limiting step for GSH biosynthesis is considered to be the availability of reduced sulphur for cysteine synthesis that occurs at the last step of the sulphur reduction pathway (Cobbett 2000). Herbette et al. (2006) suggest the importance of the availability of reduced sulphur as an essential factor for Cd detoxification. In Arabidopsis it was shown that GSH decreases when exposed to Cd (Herbette et al. 2006). The importance of reduced glutathione in PC synthesis and, consequently, in metal detoxification, is well illustrated by the Cd-sensitive cad2-1 mutant of Arabidopsis, which is deficient in the first enzyme of the GSH biosynthesis pathway, γ-glutamylcysteine synthetase (γ-ECS) and as a result hypersensitive to Cd (Cobbett et al. 1998). A GSH decrease in T. caerulescens as a result of Cd exposure may therefore result in an induction of genes involved in sulphur metabolism. In Saccharomyces cerevisiae, exposure of cells to Cd led to a global drop in sulphur-containing protein synthesis and in a redirection of sulphur metabolite fluxes towards the glutathione pathway (Lafaye et al. 2005). Thus, at least in yeast, heavy metal exposure and sulphur homeostasis are clearly linked and the same seems to be the case for Arabidopsis and T. caerulescens. Future reverse-genetic studies will be useful to determine if modified sulphate assimilation, compared to Arabidopsis, is actually required for Cd tolerance in T. caerulescens.

Another interesting observation in this respect was that in T. caerulescens the R2R3-Myb transcription factor MYB28 was found to be higher expressed upon high Cd exposure (Table 2, Supplementary Table S5). Recently MYB28 was identified as a transcription factor gene involved in the regulation of aliphatic glucosinolate production (Hirai et al. 2007). Glucosinolates are sulphur-rich, anionic natural products and degradation of glucosinolates provides an additional sulphur source under stress (Kliebenstein et al. 2001). Glutathione (gamma-glu-cys-gly) is the major reservoir of non-protein reduced sulphur and serves as an important line of defence against reactive oxygen species, xenobiotics and heavy metals. Recently, Herbette et al. (2006) showed a down-regulation of genes involved in the glucosinolate biosynthesis pathway in Arabidopsis after exposure of Cd and they suggested that this diverted the sulphate demand for PC biosynthesis by decreasing the glucosinolate level in the plant. The higher expression of MYB28 in T. caerulescens suggests an increase in glucosinolate level in the plant after exposure to Cd but no other genes involved in glucosinolate biosynthesis were differentially expressed in T. caerulescens. These results suggest an additional function of MYB28 in the hyperaccumulator T. caerulescens that might be linked to sulphate assimilation.

Regulation of Cd response in T. caerulescens and Arabidopsis

In addition to studying the structural genes involved in Cd response, we are also interested in the regulatory genes that control expression of the structural genes. As there is almost no overlap in genes differentially expressed in response to Cd exposure between T. caerulescens and Arabidopsis, it appears as if there is a strong difference in signalling of Cd exposure response in both species.

The OBP3-responsive genes (ORG1, ORG2 and ORG3) are higher expressed under excess Zn and Cd exposure in Arabidopsis (Table 1 + Supplementary Table S3). OBP3 is a DOF (DNA binding with one finger) protein, which plays an important role in plant growth and development (Kang & Singh 2000). ORG2 and ORG3 encode basic helix-loop-helix domain (bHLH) containing transcription factors (Kang et al. 2003), while ORG1 encodes a protein kinase. ORG1, ORG2 and ORG3 have been found to be higher expressed in response to salicylic acid (SA) (Kang et al. 2003). SA is known to have a signalling role in a wide range of oxidative stresses (Rao & Davis 1999; Borsani, Valpuesta & Botella 2001) and because Cd induces oxidative stress in plants, SA may play a role in enhancing Cd tolerance by inducing the antioxidant defence mechanism (Choudhury & Panda 2004). Freeman et al. (2005) showed that elevation of free SA levels in Arabidopsis, both genetically and by exogenous feeding, enhances the specific activity of serine acetyltransferase, leading to elevated glutathione and increased Ni resistance. ORG1, ORG2 and ORG3 were not as differentially expressed in T. caerulescens as in Arabidopsis, and the semi-quantitative RT-PCR showed only for ORG3 an induced expression in the roots when exposed to Cd (Fig. 2b). This is in line with the previous observation by Freeman et al. (2005) that Ni-hyperaccumulating Thlaspi species have a constitutively elevated level of SA compared to Arabidopsis.

Other transcription factors that are higher expressed under Cd exposure in Arabidopsis are MYB4, MYB10, MYB72 and bHLH100 (Table 1). MYB4 is a key regulator of phenylpropanoid pathway gene expression, and is an example of a MYB protein that functions as a transcriptional repressor (Hemm, Herrmann & Chapple 2001). Its role in metal homeostasis or tolerance is not directly clear. MYB10 and MYB72 were previously identified as Fe-regulated transcription factors and their regulation partly depends on FIT1, which encodes a putative bHLH transcription factor (Colangelo & Guerinot 2004). The myb72 mutant was more sensitive to growth under Fe-deficient conditions and Zn excess conditions (Fig. 4b), which also appears to induce a Fe deficiency response in Arabidopsis (van de Mortel et al. 2006). Thus, Arabidopsis exposed to excess Zn and Cd may become somewhat Fe deficient, which also explains the induction of other Fe homeostasis-related genes by Cd exposure.

In both Arabidopsis and T. caerulescens, the bHLH100 transcription factor was higher expressed under high Cd exposure (Table 1, Supplementary Tables S3 & S5, Fig. 2). The 35S::bHLH100 lines showed enhanced tolerance to excess Zn and Ni compared to wild type (Fig. 3b), but not a very clear enhancement of Cd tolerance, meaning that the function of this gene may be related to general abiotic stress response, but remains still largely elusive.

Differences in Cd response between T. caerulescens and Arabidopsis

Probably most interesting for the identification of genes that contribute to the Cd tolerance of T. caerulescens are the genes that are differentially expressed between T. caerulescens and Arabidopsis at comparable Cd and/or Cd/Zn exposures (Tables 3 & 4). This is a very large set of genes though, with most of them apparently not involved in metal homeostasis (van de Mortel et al. 2006). Therefore we only considered genes that showed a different expression profile when comparing Arabidopsis and T. caerulescens and particularly focused on genes responding differentially to Cd. Many of these genes are involved in transcriptional regulation (Tables 3 & 4). CCA1 (Circadian Clock Associated 1) and LHY1 (Late Elongated Hypocotyl), which encode closely related MYB transcription factors regulating circadian rhythms in Arabidopsis (Green & Tobin 2002) and the Early Phytochrome Responsive 1 (EPR1) gene, encoding another MYB transcription factor highly similar to CCA1 and LHY1 (Kuna et al. 2003), are all higher expressed in T. caerulescens in response to Zn deficiency and particularly Cd exposure, while these genes are not responding to Zn or Cd in Arabidopsis. CCA1 was also higher expressed in A. halleri under low Zn supply (Becher et al. 2004) suggesting some specific role of these circadian rhythm controlling factors in metal hyperaccumulators.

Of course it would also be interesting to compare similarities and differences in gene expression between the hyperaccumulator species T. caerulescens and A. halleri. Unfortunately there are only few reports on the analysis of genes differentially expressed in A. halleri when compared to A. thaliana. We find little overlap when comparing T. caerulescens to A. halleri, but we think this is mainly due to differences in experimental set-up, particularly the exposure time to Cd, which was 7 d in our experiments and only 2 h and 3 d, respectively, in experiments by Weber et al. (2006) and Craciun et al. (2006). The 2 h exposure induces a strong expression of stress response genes, as we already discussed. The analysis by Craciun et al. (2006) involved cDNA-AFLP analysis of Cd-tolerant and non-tolerant genotypes from a backcross population between A. halleri and Arabidopsis lyrata ssp. petrea. This identified only 20 differentially expressed transcripts that respond to changes in Cd exposure and is thus a very small set for comparison, which is probably why we found no overlap.

CONCLUSION

In conclusion, the comparative transcriptional analysis of the Cd response of the Zn/Cd-hyperaccumulator T. caerulescens and the non-accumulator Arabidopsis indicates that there are specific responses to Cd exposure in each species and emphasizes the role of genes involved in lignin, glutathione and sulphate metabolism. It also suggests a similar role for many more genes, with unknown function. Genes found to be differentially expressed within this study are expected to be involved in Cd tolerance, rather than Cd hyperaccumulation, because T. caerulescens LC is not a very efficient hyperaccumulator of Cd.

ACKNOWLEDGMENTS

This research was supported by NWO Genomics Grant 050-10-166 (J.E.v.d.M.) and the FP6 European Union Project ‘Peroxisome’ LSHG-CT-2004-512018 (E.V.L.v.T.). We thank Viivi Hassinen (University of Kuopio, Finland) for providing the primer sequences of the T. caerulescens tubulin gene prior to publication; Lisa Gilhuijs-Pederson and Antoine van Kampen (Academic Medical Center, University of Amsterdam, the Netherlands) for their input in the microarray design; and Maarten Koornneef (Laboratory of Genetics, Wageningen University, the Netherlands) for carefully reading the manuscript.

Ancillary