| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
First published online July 14, 2006; 10.1104/pp.105.076232 Plant Physiology 142:148-167 (2006) © 2006 American Society of Plant Biologists Zinc-Dependent Global Transcriptional Control, Transcriptional Deregulation, and Higher Gene Copy Number for Genes in Metal Homeostasis of the Hyperaccumulator Arabidopsis halleri1,[W]Max Planck Institute of Molecular Plant Physiology, D–14476 Potsdam-Golm, Germany
The metal hyperaccumulator Arabidopsis halleri exhibits naturally selected zinc (Zn) and cadmium (Cd) hypertolerance and accumulates extraordinarily high Zn concentrations in its leaves. With these extreme physiological traits, A. halleri phylogenetically belongs to the sister clade of Arabidopsis thaliana. Using a combination of genome-wide cross species microarray analysis and real-time reverse transcription-PCR, a set of candidate genes is identified for Zn hyperaccumulation, Zn and Cd hypertolerance, and the adjustment of micronutrient homeostasis in A. halleri. Eighteen putative metal homeostasis genes are newly identified to be more highly expressed in A. halleri than in A. thaliana, and 11 previously identified candidate genes are confirmed. The encoded proteins include HMA4, known to contribute to root-shoot transport of Zn in A. thaliana. Expression of either AtHMA4 or AhHMA4 confers cellular Zn and Cd tolerance to yeast (Saccharomyces cerevisiae). Among further newly implicated proteins are IRT3 and ZIP10, which have been proposed to contribute to cytoplasmic Zn influx, and FRD3 required for iron partitioning in A. thaliana. In A. halleri, the presence of more than a single genomic copy is a hallmark of several highly expressed candidate genes with possible roles in metal hyperaccumulation and metal hypertolerance. Both A. halleri and A. thaliana exert tight regulatory control over Zn homeostasis at the transcript level. Zn hyperaccumulation in A. halleri involves enhanced partitioning of Zn from roots into shoots. The transcriptional regulation of marker genes suggests that in the steady state, A. halleri roots, but not the shoots, act as physiologically Zn deficient under conditions of moderate Zn supply.
Metal hyperaccumulation is a rare trait found in approximately 440 plant taxa, 11 of which have been reported to hyperaccumulate zinc (Zn; Reeves and Baker, 2000  ). Individuals of these taxa accumulate very high Zn concentrations of more than 10,000 µg g–1 Zn in leaf dry biomass (Baker and Brooks, 1989). The accumulation of higher metal concentrations in shoots than in roots has been described as a characteristic of hyperaccumulation in both hydroponic and soil culture (Baker et al., 1994; Krämer et al., 1996; Weber et al., 2004). All known metal hyperaccumulator taxa occur primarily on geologically metal-rich or metal-contaminated soils and possess extraordinarily high metal tolerance. The number of Zn and nickel hyperaccumulator taxa is higher in the Brassicaceae family than in any other plant family, and includes a large number of nickel hyperaccumulator species in the section Odontarrhena of the genus Alyssum and the two most commonly studied Zn hyperaccumulator model species Thlaspi caerulescens and Arabidopsis halleri (Reeves and Baker, 2000).
Both subspecies of A. halleri, subsp. gemmifera and subsp. halleri, are hyperaccumulators of Zn, and subsp. halleri is able to grow healthily and to accumulate very high leaf Zn concentrations of between 3,000 and 22,000 µg g–1 dry biomass within a wide dynamic range of external Zn concentrations in the field (Bert et al., 2000
Researchers have begun to address the major differences between the metal homeostasis networks of closely related metal hyperaccumulator and nonaccumulator plants at the molecular level (Salt and Krämer, 2000
The observed high transcript levels of candidate genes in A. halleri and in other hyperaccumulator model plants are based on a modified regulation of transcript levels (Pence et al., 2000
For A. halleri MTP1, which is involved in Zn hypertolerance, the amplification of gene copy number contributes to enhanced MTP1 transcript levels, when compared to A. thaliana or A. lyrata (Dräger et al., 2004
Previously, root and shoot transcript profiles were established in separate studies using A. halleri and A. thaliana grown under different conditions with Affymetrix 8 K chips representing only a limited number of genes (Becher et al., 2004 Here we present genome-wide transcriptional profiles of hydroponically cultivated A. halleri and A. thaliana plants maintained under control conditions or upon a short-term exposure to high Zn concentrations. We report a set of candidate genes for metal hyperaccumulation and tolerance, for the adjustment of micronutrient homeostasis, and for the protection from secondary abiotic stress in A. halleri. For several of the most highly expressed candidate genes, genomic copy number is estimated to be higher in A. halleri than in A. thaliana by DNA gel blot. The detailed analysis of transcriptional regulation of candidate genes by real-time reverse transcription (RT)-PCR is combined with existing and newly generated knowledge on gene product function to develop a model of the alterations in the metal homeostasis network underlying metal hyperaccumulation in A. halleri.
Establishment of Experimental Conditions To compare the effect of external Zn supply on Zn accumulation and partitioning in A. thaliana and A. halleri, plants were cultivated hydroponically and supplied with different subtoxic Zn concentrations for 3 weeks (Fig. 1, A and B ). In A. halleri, the ratio of shoot Zn concentrations to root Zn concentrations was above unity in plants cultivated in a Zn-deficient hydroponic solution containing no added Zn for 3 weeks (Fig. 1C). Continuous supply of Zn concentrations of between 1 and 5 µM led to an increase in the shoot-to-root ratio of Zn concentrations, with maximum ratios approximately tripling the ratio determined under Zn-deficient conditions. By contrast, A. thaliana partitioned into the shoots only a small proportion of the Zn concentration accumulated in the roots (Fig. 1, A and B). This was reflected in shoot-to-root Zn concentration ratios substantially below unity, which decreased further with increasing Zn supply (Fig. 1C). These results indicated that A. halleri accumulates Zn predominantly in the shoot even at very low Zn supply, whereas A. thaliana primarily immobilizes Zn in the roots irrespective of the Zn supply. When A. halleri was cultivated in media containing high Zn concentrations of 30 or 300 µM Zn, shoot-to-root ratios of Zn concentrations were below unity and only about one-quarter of the ratios measured at 0.3 µM Zn. In addition to the dramatic interspecies differences in Zn partitioning between roots and shoots, the data suggest a species-specific modulation of Zn partitioning dependent on the Zn supply.
To select suitable conditions and time points for transcript profiling, transcript levels of three metal homeostasis genes, ZIP4, ZIP9, and NAS2, were determined in preliminary experiments (Fig. 2 ). These genes are known to be Zn responsive at the transcript level in A. thaliana (Grotz et al., 1998 ; Wintz et al., 2003) and they have previously been identified as candidate genes for Zn accumulation and tolerance in A. halleri (Becher et al., 2004; Weber et al., 2004).
Plants of two different Zn steady states were subjected to short-term increases in external Zn supply. In one experiment, plants grown under control (1 µM) Zn conditions were exposed to excess Zn concentrations of 30 µM Zn for A. thaliana and 300 µM Zn for A. halleri, for 2 and 8 h. The choice of concentrations reflects naturally selected Zn tolerance in A. halleri, whereas A. thaliana possesses only basic Zn tolerance (Clemens, 2001 ; Becher et al., 2004). (In preliminary experiments, the chosen concentrations were the maximum Zn concentrations that did not cause a reduction in biomass gain in A. thaliana and A. halleri, respectively, during 4 d of exposure [M. Becher, I.N. Talke, A.N. Chardonnens, and U. Krämer, unpublished data].) In an additional experiment, Zn-deficient plants cultivated in a medium lacking added Zn for 3 weeks were resupplied with a Zn concentration of 5 µM (Fig. 2). In roots, Zn-induced changes in transcript levels were observed as early as 2 h after increasing the Zn concentrations in the hydroponic medium (Fig. 2, top row: ZIP4, ZIP9). In shoots, transcriptional responses were observed 2 or 8 h after the increase in Zn concentrations in the hydroponic medium (Fig. 2, bottom row: ZIP4, ZIP9, and NAS2). Transcripts of ZIP9 encoding a transporter of the ZIP family and NAS2 encoding a nicotianamine synthase were not detectable in shoots of either Arabidopsis species grown in the presence of sufficient (control) Zn, but were strongly induced upon Zn deficiency (Fig. 2, ZIP9, NAS2). This suggested that neither of the two species experienced Zn deficiency under the chosen control conditions.
According to the results from the preliminary experiments, microarrays were hybridized with labeled cRNAs prepared from roots and shoots of A. halleri and A. thaliana grown at 1 µM Zn and after 2 and 8 h of exposure to excess Zn for roots and shoots, respectively (see "Material and Methods"). Under control conditions, higher signal intensities were detected for a total of 628 probe sets in roots, and for 1,739 probe sets in shoots of A. halleri compared to A. thaliana (as outlined in "Materials and Methods;" for visualization see all colored, nongray datapoints in Fig. 3, A and B ; complete gene lists with annotations in Supplemental Tables I and II). We also identified genes that responded at the transcript level to excess Zn in A. halleri (see "Materials and Methods;" Supplemental Tables III and IV). To group the identified genes according to functional classes, a genome-wide annotation list was assembled; this list includes the class of metal homeostasis-related genes, which contained all functionally characterized A. thaliana genes encoding proteins involved in metal transport, chelation, binding, and trafficking and all other members of the respective protein families, and genes encoding members of protein families known to participate in metal homeostasis in other organisms (Supplemental Table XI: class "Metal ion homeostasis;" Fig. 3, see "Material and Methods"). The intersection of the gene list of metal homeostasis-related genes and of the genes that are either more highly expressed (Supplemental Tables I and II) or Zn regulated in A. halleri (Supplemental Tables III and IV) contained a total of 17 genes in the roots and 19 genes in the shoots, with seven genes in common between both roots and shoots (Fig. 3; Table I ). We considered these genes as an initial set of major candidate genes for an involvement in naturally selected metal tolerance and hyperaccumulation in A. halleri.
The metal homeostasis genes newly identified here as candidate genes encode proteins encompassing both cellular functions known to be of primary importance in metal homeostasis: membrane transport of metals and metal binding/chelation. The 12 new candidate genes encoding proteins that are likely to be involved in the transport of Zn or other metals across membranes were Zn-regulated transporter/iron-regulated transporter-like protein 10 (ZIP10; roots, 111-fold), metal tolerance protein 8 (MTP8; shoots, 65-fold), heavy metal-associated domain-containing protein 4 or heavy metal ATPase 4 (HMA4; shoots, 53-fold), ferric reductase defective 3 (FRD3; roots, 45-fold), phosphate transporter 1;4 (PHT1;4; roots, 19-fold), cation/H+ exchanger 18 (CHX18; roots, 6.5-fold), MTP11 (shoots, 4.7-fold), P-type ATPase from Arabidopsis 1 (PAA1/HMA6; shoots, 4.6-fold), iron-regulated transporter 2 (IREG2; roots, 4.5-fold), iron-regulated transporter 3 (IRT3; roots, 4.5-fold), yellow-stripe 1-like protein 6 (YSL6; shoots, 4.4-fold), and ZIP3 (down-regulated 2.7-fold in roots following exposure to excess Zn; Fristedt et al., 1999 ; Curie et al., 2001; Mäser et al., 2001; Rogers and Guerinot, 2002; Blaudez et al., 2003; Jensen et al., 2003; Lahner et al., 2003; Shikanai et al., 2003; Green and Rogers, 2004; Hussain et al., 2004; Shin et al., 2004). A total of five new candidate genes encoding metal-binding proteins or proteins involved in the biosynthesis of metal chelators included ferritin 2 (FER2; shoots, 11-fold), S-adenosyl-Met synthetase 1 (SAMS1; shoots, 10-fold), SAMS2 (shoots, 8.5-fold), FER1 (shoots, 5.1-fold), and SAMS3 (roots and shoots, around 5-fold; Peleman et al., 1989; Petit et al., 2001). The compound S-adenosyl-Met is the substrate for nicotianamine synthase, and the A. halleri nicotianamine synthase genes NAS2 and NAS3 have previously been implicated in Zn tolerance of A. halleri (Becher et al., 2004; Weber et al., 2004; see also Table I).
Two additional, closely related genes encoding the protein disulfide isomerases 1 and 2 were included among the set of candidate genes analyzed here because of their interesting regulation and potential function in metal detoxification (PDI1: shoots, 39-fold higher in A. halleri; PDI2: shoots, 17-fold; Table I; Rensing et al., 1997
Among candidate genes, the microarray data suggested a trend toward decreasing transcript levels in response to excess Zn in roots of both A. thaliana and A. halleri (Table I). Conversely, in A. halleri roots transcript levels of ZIP and NAS genes were increased under Zn deficiency, when compared to control conditions (Table I). In shoots of A. thaliana, transcript levels of several candidate genes were increased in response to excess Zn, namely for PDI and SAMS genes, and for YSL6 (Table I). In A. halleri, transcript levels of these genes were constitutively high and unchanged in response to excess Zn (Table I). Our data suggested that, especially in shoots, A. halleri responds differently to excess Zn than A. thaliana at the transcript level. We investigated whether the species-specific patterns observed in metal-dependent transcriptional regulation of candidate genes are also reflected globally across the entire panel of genes represented on the ATH1 microarray. In roots of A. halleri and A. thaliana, exposure to excess Zn resulted in increased transcript levels for 0.5% and 0.9%, respectively, of all expressed genes, and in decreased transcript levels for only 0.09% and 0.07%, respectively, of all expressed genes (Supplemental Tables III and V, P < 0.1). Thus, the predominant directions of transcriptional responses to excess Zn were identical in roots of A. halleri and A. thaliana, with a global trend toward up-regulation opposing a trend toward down-regulation among our set of candidate genes in both species. In shoots of A. thaliana, among all expressed genes, transcriptional up-regulation was observed in response to excess Zn (0% down, 3.2% up, P < 0.1; Supplemental Table VI). By comparison, in shoots of A. halleri a smaller proportion of expressed genes responded, and transcriptional down-regulation was predominant (0.8% down, 0.2% up, P < 0.1; Supplemental Table IV). This indicated that primarily in the shoots, the two species respond differently to excess Zn also at the whole transcriptome level.
To assess the reliability of the cross species expression differences between A. halleri and A. thaliana detected using the microarrays, we determined transcript levels by an alternative method, quantitative real-time RT-PCR, for a subset of 18 genes (Table II
). Overall, there was good qualitative agreement between the ratios of microarray signals in A. halleri relative to A. thaliana and the ratios of transcript levels as determined by real-time RT-PCR (Table II). The false negatives according to microarray hybridization, ZIP3, and HMA4 (in roots), could be attributed to a high divergence of the A. halleri sequence from the A. thaliana sequence in the region covered by the oligonucleotide probes on the microarray, in combination with a modest magnitude of the A. halleri to A. thaliana ratio of transcript levels. For several other genes, quantitative discrepancies between microarray and real-time RT-PCR data are likely to be attributable to cross hybridization between high transcript levels of a gene in A. halleri and the probes for a closely related member of the same gene family (NAS2 on NAS4 in roots, NAS3 on NAS2 in shoots, NAS2 on NAS3 in shoots under Zn deficiency, and PDI1 on PDI2). It was reported earlier that very low signals (for example, for ZIP10, MTP8, and ZIP9 in A. thaliana) are often quantified imprecisely on the ATH1 chip (Czechowski et al., 2004
Out of all candidate genes identified here, absolute transcript levels were highest for A. halleri HMA4 (see Table II), which encodes a membrane transport protein of the P1B transition metal pump family of the P-type ATPases (Axelsen and Palmgren, 2001
Gene Copy Number of Highly Expressed Candidate Genes in A. halleri
Previously, high transcript levels of the Zn tolerance gene MTP1 were partly attributed to the presence of several MTP1 gene copies in the genome of A. halleri, whereas the genome of the closely related nontolerant species A. thaliana contains only a single MTP1 gene copy (Dräger et al., 2004
Regulation of Candidate Gene Expression Transcriptional Zn responses were further investigated for a subset of 18 candidate genes using real-time RT-PCR (Fig. 6 ). In the roots of A. halleri, short-term exposure to 300 µM Zn for 2 and 8 h resulted in the down-regulation of transcript levels of a group of genes, which encode several ZIP family Zn transporters with likely functions in Zn influx into the cytoplasm (Fig. 6A). Down-regulation of transcript levels of this group of genes was clearly less pronounced in the roots of A. thaliana after short-term exposure to 30 µM Zn. It has to be kept in mind that in roots of A. halleri the Zn-induced decrease in transcript levels occurred from very high transcript levels under control conditions. Consequently, despite a more pronounced down-regulation in A. halleri, final root transcript levels 8 h after supply of excess Zn were generally higher in A. halleri than in A. thaliana (compare Fig. 2).
Different from roots, shoot transcript levels of most ZIP genes were generally as low in A. halleri as in A. thaliana under control conditions (see Fig. 2; Table II). In shoots of A. halleri, there was virtually no transcriptional response to a short-term oversupply of Zn, whereas A. thaliana shoots responded by a marked down-regulation of several genes including primarily ZIP genes and PHT1;4 (Fig. 6B). To test whether the shoots of A. halleri were able to respond transcriptionally to changes in Zn supply, we investigated gene expression under Zn deficiency (Fig. 6). Zn deficiency resulted in a very pronounced increase in shoot transcript levels of genes encoding ZIP family transporters and nicotianamine synthase proteins (Fig. 6B; see Table I). In roots of A. halleri, Zn deficiency resulted in a moderate increase in transcript levels of these genes, when compared to the high transcript levels observed under control conditions (Fig. 6A; for comparison with A. thaliana, see Figs. 2 and 7 ). To analyze the specificity of transcriptional Zn responses, candidate gene regulation in A. halleri was also investigated following exposure to moderate excesses of Cd (30 µM) or copper (Cu; 5 µM) and NaCl (100 mM) for 2 and 8 h (Fig. 6). The results suggested that overall, Zn responses were rather specific. Some of the Zn-responsive genes, primarily NAS and ZIP genes, were also down-regulated in response to excess Cu, as for excess Zn. The transcriptional responses to excess Zn and excess Cd, however, were distinct. Furthermore, there was a second group of genes exhibiting a pattern of regulation that was clearly different from the Zn- and Cu-repressed genes. Expression of these genes was induced in response to exposure to 100 mM NaCl and included MTP8, PHT1;4, FRD3, ZIP6, and OASA2. Transcript levels of subsets of these genes were also increased following exposure to excess Cu or Cd, supporting their responsiveness to abiotic stress. PDI transcript levels showed few changes over the entire range of conditions, with an increase in root PDI2 transcript levels in response to excess Cu in A. halleri and an increase in PDI1 transcript levels in response to Zn in shoots of A. thaliana. Transcript levels of HMA4 appeared to be constitutively very high in the roots of A. halleri across all treatments, and only slightly induced under Zn deficiency in A. halleri shoots (Fig. 6, A and B). A detailed analysis confirmed the extremely high expression of HMA4 in roots and shoots of A. halleri (Fig. 7). Under Zn deficiency, shoot HMA4 transcript levels were slightly increased, and they were reduced to control levels shortly after readdition of Zn to the medium (Fig. 7). We investigated transcript levels of the additional newly identified candidate genes ZIP3 and IRT3, under different conditions of Zn deficiency, resupply, and oversupply (Fig. 7). There were overall similarities in short-term Zn responses of plants maintained under control (1 µM added Zn) and Zn-deficient (0 µM added Zn) conditions (Fig. 7; see also Fig. 2). In response to resupply of 5 µM Zn to Zn-deficient plants, there was a decrease in transcript levels for these Zn-deficiency-induced genes in both A. halleri and A. thaliana. This demonstrated clearly that there is no general defect in the magnitude or time scale of transcriptional Zn responses in A. halleri (Fig. 7; see also Fig. 2: ZIP4 and ZIP9 in controls and at time points 8 and 24 h). In roots of A. halleri, after the initial down-regulation, which was strongest 2 h after resupply of Zn, a general trend of a gradual increase in transcript levels was observed at later time points (Fig. 7).
In a microarray-based cross species comparative transcript analysis in A. thaliana and A. halleri, we identified a set of candidate genes expressed at higher levels in A. halleri than in A. thaliana or regulated by Zn in A. halleri (Fig. 3; Table I; Supplemental Tables I–IV). Microarray data were confirmed by an alternative technique of transcript quantification, real-time RT-PCR, for a representative subset of genes. Among the genes identified using microarrays, combined in silico and real-time RT-PCR analyses revealed a total of 29 genes encoding putative metal homeostasis proteins, which we consider here as an initial set of major candidates for a role in naturally selected metal hyperaccumulation and associated metal hypertolerance in A. halleri. Of these, 18 genes have not previously been implicated in these traits, and 11 confirm earlier findings (Tables I and II; Becher et al., 2004 ; Weber et al., 2004; Mirouze et al., 2006).
We performed DNA gel blots to estimate genomic copy numbers for a subset of 12 candidate genes (Fig. 5; Table III; Supplemental Fig. 1). The obtained data are important for future studies of segregation of candidate genes and for the cloning of genomic sequences from A. halleri. Moreover, the combination of DNA gel blot with steady-state transcript level data suggested the presence of more than a single genomic copy in A. halleri of the genes, for which the highest transcript levels were detected, HMA4, ZIP9, and ZIP3 (Tables II and III). As reported previously for MTP1 (Dräger et al., 2004
Ratios of shoot-to-root Zn concentrations are generally below unity in A. thaliana and other nonaccumulator plants, and typically above unity in A. halleri and other Zn hyperaccumulating plants (Baker et al., 1994
Transcript levels of genes known to be transcriptionally regulated by Zn, such as ZIP1, ZIP3, ZIP4, ZIP9, NAS2, and NAS3 can be used as indicators of the Zn status of A. thaliana plants (Grotz et al., 1998 There was a rapid and strong down-regulation of transcript levels of ZIP3, ZIP4, and IRT3 in roots of A. halleri following exposure to excess Zn or resupply of Zn to roots of Zn-deficient plants grown in a medium lacking added Zn (Figs. 2 and 7). Overall, this suggests that short-term Zn-dependent down-regulation of Zn deficiency-induced transcripts in roots of A. halleri is generally as functional and efficient as in A. thaliana. Only NAS2 and ZIP9 transcript levels were not or only partly Zn responsive in roots of A. halleri, respectively. The Zn-dependent control of transcript abundance of these two genes in A. thaliana appears to be partially or fully inactive in roots, but not in shoots, of A. halleri. The simplest explanation for the high expression of Zn deficiency responsive genes in the roots of A. halleri at normal Zn supplies could be a loss of Zn from the roots via high root-to-shoot Zn transport rates (see Fig. 1). The proteins encoded by two of the candidate genes have been implicated in root-to-shoot metal translocation in A. thaliana: HMA4 and FRD3.
The A. thaliana HMA4 protein is a P1B-type metal pump that localizes to the plasma membrane and mediates cellular metal efflux. Together with the related AtHMA2, AtHMA4 has a role in the transport of Zn from the root to the shoot and in metal detoxification (Hussain et al., 2004
The A. thaliana and A. halleri FRD3 proteins are members of the multidrug and toxin efflux family of membrane transport proteins. Although the transport function of AtFRD3 has not yet been directly established, the available data suggest that AtFRD3 has a role in maintaining iron (Fe) mobility between the root and the shoot. In the frd3 mutant, root Fe deficiency responses are constitutively active, leading to secondary accumulation of a number of metals including manganese (Mn), cobalt, Cu, Zn, and Fe in both roots and shoots (Delhaize, 1996
In addition to FRD3, several other genes found to be highly expressed in A. halleri (Table I) were previously associated not with Zn or Cd homeostasis, but with the homeostasis of other transition metals, among them FER1 (Fe), FER2 (Fe), NRAMP3 (Fe and Mn), IREG2 (Fe), HMA6/PAA1 (Cu), MTP8, and MTP11 (Mn; Thomine et al., 2000
To address the specificity of Zn-dependent regulation of candidate gene transcript levels, we exposed plants to an excess of Cu, Cd, or NaCl in the hydroponic medium. Root ZIP3 and ZIP4 transcript levels and transcript levels of all four NAS genes were suppressed following exposure of A. halleri to excess Cu (Fig. 6). Wintz et al. (2003)
There appear to be similarities between major candidate genes expressed at high levels in A. halleri and in the metal-tolerant Zn and Cd hyperaccumulator model species T. caerulescens. The A. halleri/A. lyrata clade is phylogenetically closest to A. thaliana, with an estimated divergence time of between 3.5 and 5.8 million years ago (Koch et al., 2001
The comprehensive identification of candidate genes in several hyperaccumulator model plants is a prerequisite for their functional, regulatory, and sequence analysis. This will be an important step toward a better understanding of the evolution of naturally selected metal hyperaccumulation and associated metal hypertolerance, and of evolutionary genome dynamics in plants. Furthermore, this will allow a comparison with extreme traits in other Arabidopsis relative model systems, such as sodium (Na) tolerance in Thellungiella halophila (Inan et al., 2004 In summary, we have used cross species transcript profiling to successfully identify on a genome-wide scale a number of candidate genes for naturally selected metal tolerance and hyperaccumulation in A. halleri. By exploring candidate genes at the genome, functional, and regulatory level we have obtained insights into the complex molecular basis of an extreme physiological trait.
Plant Material, Growth Conditions, and Experimental Treatments
Plants of Arabidopsis thaliana (L. Heynhold, accession Columbia [Col]) and Arabidopsis halleri (L. O'Kane and Al-Shehbaz subsp. halleri, accession Langelsheim) were grown from seeds, which were the progeny of a single mother plant in a pool of six individuals grown in an open greenhouse from seeds collected at Langelsheim. Plants were cultivated in hydroponic culture as described in Becher et al. (2004) Short-term high Zn, Cd, Cu, and Na treatments were initiated when plants were 6.5 weeks old. For short-term high Zn supply, the medium was replaced with fresh control medium (1 µM ZnSO4) 8 h before harvest. Then nothing (controls) or 30 µM ZnSO4 and 300 µM ZnSO4 for A. thaliana and A. halleri, respectively, was added to the hydroponic solution 8 or 2 h before harvest. Three biologically independent experiments, each comprising five culture vessels per treatment (15 plants per treatment) were conducted. In experiments with Cd, Cu, and Na treatments of A. halleri, media were exchanged (1 µM ZnSO4; Cd: none; Cu: 0.5 µM CuSO4; Na: 0.1 µM Na2MO4) 3 d before harvest; 30 µM CdCl2, 5 µM CuSO4, or 100 mM NaCl were added as treatments 8 or 2 h before harvest. Two biologically independent experiments were conducted, with at least two culture vessels (at least six plants in total) for each treatment. For Zn deficiency experiments with Zn resupply, the treatment was initiated when plants were 4.5 weeks old and lasted for 3 weeks. For deficiency, ZnSO4 was omitted from the hydroponic solution; the sufficient control solution contained 5 µM ZnSO4. The last change of medium was done 3 d before harvest. For Zn resupply, 5 µM ZnSO4 was added to the hydroponic solution of Zn-deficient plants 24, 8, and 2 h before harvest. Two biologically independent experiments were conducted, each comprising at least three culture vessels (at least nine plants) per treatment. In the experiments conducted to determine root and shoot Zn concentrations after long-term treatment with a range of different Zn supplies, plants were 4.5 weeks old when treatments were initiated. For A. halleri, the hydroponic solution contained 0, 0.3, 1, 5, 10, 30, or 300 µM ZnSO4, for A. thaliana 0, 1, 5, or 10 µM ZnSO4. Upon harvest after 3 weeks of treatment, shoots were briefly rinsed in ultrapure water and blotted dry, and roots from three individuals of the same species and hydroponic culture vessel were desorbed together in 150 mL of an ice-cold solution of 5 mM CaCl2 and 1 mM MES-KOH (pH 5.7) for 20 min, with a replacement of the solution after 5 min, followed by two washes in 150 mL ice-cold water, each for 3 min. Tissues were pooled by culture vessel and dried at 60°C for 3 d. Two biologically independent experiments were conducted, each comprising three culture vessels (nine plants) per treatment.
Dried root and shoot material was homogenized, and for each sample, 35 to 70 mg were processed and elemental contents analyzed by inductively coupled plasma optical emission spectroscopy using an IRIS Advantage Duo ER/S (Thermo Jarrell Ash) as described previously (Becher et al., 2004
For microarray expression profiling, total RNA was extracted with Trizol (Invitrogen Life Technologies) according to the manufacturer's instructions. RNeasy spin columns and the RNase-free DNase set for on-column DNase digest (Qiagen) were used according to the manufacturer's instructions for subsequent RNA purification and digestion of genomic DNA, and as an alternative method of total RNA isolation when transcript levels were determined by real-time RT-PCR only (RNeasy plant mini kit and RNase-free DNase set; Qiagen). Quality and quantity of RNA was checked visually by denaturing gel electrophoresis and by photometric analysis (A260 and A280).
Synthesis of cDNA, cRNA labeling, and the hybridization on the GeneChip Arabidopsis ATH1 genome array were done as recommended by the manufacturer (Affymetrix; manual 701025 rev 1, https://www.affymetrix.com/support/downloads/manuals/expression_s2_manual.pdf). In short, 20 µg of total RNA were used for first-strand cDNA synthesis with 100 pmol T7(dT)24 primers and SuperScript II RT (Invitrogen), second-strand cDNA synthesis was carried out with Escherichia coli DNA ligase, DNA polymerase I, and RNase H (all Invitrogen). Biotin-labeled cRNAs were synthesized by in vitro transcription using the Enzo BioArray HighYield RNA transcript labeling kit (Enzo Diagnostics). Labeled cRNAs were quantified according to Affymetrix instructions, and a corrected amount of 25 µg cRNA were fragmented in fragmentation buffer (200 mM Tris-acetate, pH 8.1, 500 mM KOAc, 150 mM MgOAc). Before hybridization on ATH1 arrays, sample quality was assessed by hybridization on the test3-array (Affymetrix) for examination of 3' to 5' intensity ratios of houskeeping genes. ATH1 GeneChips were hybridized with cRNA from the following experiments: of the short-term Zn supply experiments, two for each treatment, tissue, and species; of the Zn deficiency experiments, one for each tissue from the 5 µM Zn (sufficient) and 0 µM Zn (deficient) treatment, respectively (Dr. F. Wagner, Resource Center and Primary Database, Berlin).
The microarray suite software package (MAS 5.0, Affymetrix) was used to generate probe set signals of the scanned ATH1 arrays. At a target intensity of 100, scaling factors for arrays hybridized with cRNA from A. halleri were on average 2-fold higher than scaling factors for A. thaliana arrays (A. halleri: 0.943 ± 0.196; A. thaliana: 0.424 ± 0.093). The arithmetic means of signal intensities over all probe sets were 146.9 ± 8.3 for A. halleri and 135.4 ± 9.6 for A. thaliana. For A. thaliana an average of 13,907 (61.0%) and 11,646 (51.1%) of the total number of 22,810 probe sets were assigned a "present call" by the MAS 5.0 software in roots and shoots, respectively. On average 9,750 (42.8%) and 9,389 (41.2%) probe sets were assigned a "present call" in roots and shoots, respectively, for A. halleri. Pivot data of each hybridized ATH1 GeneChip containing raw signal values and present, marginal, and absent calls (flags) were imported from MAS 5.0 into GeneSpring GX (v 7.2; Agilent Technologies, http://www.chem.agilent.com/). Within GeneSpring, the following normalization steps were carried out for each chip: first, signals below 0.01 were set to 0.01; second, all signals within a dataset were normalized to the 50th percentile of the measurements of the chip, using only measurements with a raw signal of at least 12.5; third, in a per gene normalization, the signal for each probe set was normalized to the respective signal of the control chip (from the same experiment; chips from A. thaliana for interspecies comparisons, control condition for within-species comparisons). Calculation of mean values for raw and normalized signals, statistical analysis, and data filtering were performed in GeneSpring. For A. halleri versus A. thaliana comparisons, the following filters were applied sequentially to identify genes more highly expressed in A. halleri than in A. thaliana (value filter). First, to avoid the identification of false positives in a cross species evaluation, probe sets were selected that exhibited an at least 4-fold higher mean normalized signal in A. halleri than in A. thaliana. Second, among these probe sets, we selected probe sets for which the mean raw signal value was above 25 in A. halleri. Third, we selected only those probe sets for which in a one sample Student's t test the P value was below 0.05, with false discovery rate control adjusted at 5% (Benjamini and Hochberg, 1995 |
TOP
ABSTRACT
RESULTS
from one experiment representative of two independent biological experiments. Tissues from at least three culture vessels containing three plants each were pooled for each condition. Genes were concluded to be not expressed (n.e.) when CT values were above a threshold of 35 and reaction efficiencies (REs) with the respective cDNA were more than 5% below the mean RE for the respective primer pair (see also "Material and Methods").
