Environmental Pollution 199 (2015) 73e82

Contents lists available at ScienceDirect

Environmental Pollution journal homepage: www.elsevier.com/locate/envpol

Genotypic variations in the dynamics of metal concentrations in poplar leaves: A field study with a perspective on phytoremediation Mathieu Pottier a, *, 1, Vanesa S. García de la Torre b, Cindy Victor a, Laure C. David a, 2,
bastien Thomine a Michel Chalot c, d, Se ^t 23A, F-91198 Gif Sur Yvette, France CNRS, Institut des Sciences du V
eg
etal, UPR 2355, Saclay Plant Sciences, Avenue de la Terrasse, Ba Instituto de Ciencias Agrarias, Consejo Superior de Investigaciones Científicas, Serrano 115-bis, 28006 Madrid, Spain c Universit
e de Franche-Comt
e, Laboratoire Chrono-Environnement, 4 place Tharradin, BP 71427, 25 211 Montbeliard, France d Universit
e de Lorraine, Facult
e des Sciences & Technologies, 54506 Vandoeuvre-les-Nancy cedex, France a

b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 23 April 2014 Received in revised form 20 October 2014 Accepted 15 January 2015 Available online

Poplar is commonly used for phytoremediation of metal polluted soils. However, the high concentrations of trace elements present in leaves may return to soil upon leaf abscission. To investigate the mechanisms controlling leaf metal content, metal concentrations and expression levels of genes involved in metal transport were monitored at different developmental stages on leaves from different poplar genotypes growing on a contaminated field. Large differences in leaf metal concentrations were observed among genotypes. Whereas Mg was remobilized during senescence, Zn and Cd accumulation continued until leaf abscission in all genotypes. A positive correlation between Natural Resistance Associated Macrophage Protein 1 (NRAMP1) expression levels and Zn bio-concentration factors was observed. Principal component analyses of metal concentrations and gene expression levels clearly discriminated poplar genotypes. This study highlights a general absence of trace element remobilization from poplar leaves despite genotype specificities in the control of leaf metal homeostasis. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Cadmium Zinc Copper HMA NRAMP

1. Introduction Industrial development, urban activities, and intensive use of fertilizers have released trace elements (TE) in the environment leading to persistent pollution (Sanita di Toppi and Gabbrielli, 1999). Phytoextraction uses plants to clean up TE contaminated soils (Kumar et al., 1995; Vangronsveld et al., 2009). The ideal plant for phytoextraction should display a high biomass yield, have deep roots, be able to grow on poor soils and be tolerant to metal excess (Punshon et al., 1996); moreover it should be possible to valorize the contaminated biomass (Pulford and Watson, 2003). Although no plant combines all these criteria, trees, such as poplar, constitute an interesting starting point, which could be improved by breeding

* Corresponding author. E-mail address: [email protected] (M. Pottier). 1
Catholique de Louvain, Institut des Sciences de la Current address: Universite vie, Croix du Sud, 4-15, 1348 Louvain-la-Neuve, Belgium. 2 Current address: INRA-AgroParisTech, Institut Jean-Pierre Bourgin, UMR1318, Saclay Plant Sciences, Route de St-Cyr (RD10), F-78026 Versailles cedex, France. http://dx.doi.org/10.1016/j.envpol.2015.01.010 0269-7491/© 2015 Elsevier Ltd. All rights reserved.

or by bio-engineering. Since about fifteen years, several studies already investigated the use of poplar for phytoremediation (Di Baccio et al., 2003; Dominguez et al., 2008; Laureysens et al., 2004; Migeon et al., 2009; Robinson et al., 2007; Sebastiani et al., 2004). However, to date, little is known about the molecular mechanisms controlling metal uptake in roots, distribution and storage in the trunk, branches and leaves in poplar. Poplar harvestable parts correspond to the wood which can be safely and profitably used for energy production (Chalot et al., 2012). However, the highest metal concentrations are usually observed in leaves, and the lowest in wood (Dominguez et al., 2008; Laureysens et al., 2004; Migeon et al., 2009; Unterbrunner et al., 2007). Thus, the increase of poplar phytoextraction efficiency requires not only to improve metal uptake from soil but also to limit metal accumulation in leaves, which return to soil upon abscission. This study aims at a better understanding of the dynamics of metal concentrations in leaves of deciduous trees and its variation among poplar genotypes. For this purpose metal concentrations in leaves of 14 poplar genotypes growing on a metal contaminated field were monitored from leaf emergence until leaf abscission with

74

M. Pottier et al. / Environmental Pollution 199 (2015) 73e82

a focus on metal remobilization during autumnal leaf senescence. To get insights into the mechanisms controlling leaf metal concentrations, the expression of genes putatively involved in metal homeostasis was monitored. Knowledge on the dynamics of leaf metal concentrations during senescence should provide a better estimate of the impact of metal return to soil due to leaf abscission in limiting phytoextraction by poplar. Deciphering the molecular network controlling leaf metal homeostasis will allow improvement of poplar phytoextraction efficiency using existing genotypic variation or/and genetic engineering. 2. Materials and methods 2.1. Experimental field and poplar genotypes Poplars grew on a field which is part of an approximately 1200 ha area of sandy soil (Pierrelaye-Bessancourt) located in the northwest suburban area of Paris (France). Initially dedicated to vegetable gardening, this soil was irrigated with raw wastewaters from 1899 to 2002 for fertilization purposes. Consequently, TE accumulated into the soil leading to polymetallic pollution characterized by Pb, Cu, Zn and Cd concentrations 10 times higher than in a non-irrigated reference soil (Table S1; Lamy et al., 2006). Poplars were planted in April 2007 at a density of 1000 stems ha
1. Poplar genotypes were chosen based on their ability to grow on the site and in order to cover poplar genetic diversity (several species and hybrids from different sections) within the limits imposed by French regulations (Table S2). Each of the 14 genotypes was growing on an independent plot of 24.5  20 m, each containing 49 trees. Details on genotypes are given in Table S2. The whole area was irrigated about 4e6 times during the summer. Weather in
te
o-France at a point located at formation was provided by Me res, which is located 10 km south-west of the experimental Ache field (Fig. S1). 2.2. Sampling Leaf blades of 14 different genotypes were harvested monthly from April to October (2011) or November (2012), i.e. until leaf abscission. For each genotype, four trees were randomly selected. For each tree, eight leaves located between 1.5 and 2.5 m from the

Fig. 1. Evolution of leaf chlorophyll content monitored by SPAD chlorophyll meter averaged from 14 to 9 poplar genotypes growing on Pierrelaye contaminated field in 2011 (closed symbol) and 2012 (open symbol), respectively. Error bars represent SE (n ¼ 56 in 2011, n ¼ 36 in 2012). Different letters indicate significant differences between harvests according to a KruskaleWallis test (p < 0.05) followed by a Tukey post hoc test.

ground were collected between 10 a.m. and 2 p.m. SPAD (SinglePhoton Avalanche Diode) index value was measured as a proxy to chlorophyll concentration using a chlorophyll meter (SPAD-502 Plus model, Konica Minolta, Japan). For each tree, SPAD values from eight leaves were averaged. An equal leaf blade area was collected for each leaf using a punch (diameter ¼ 14 mm). The eight leaf disks obtained from the same tree were frozen together for subsequent RNA extraction. The remaining part was washed in milliQ-water, wiped and dried at 60  C for three days. In the same way as for RNA extraction, an equal leaf blade area was collected from each harvested leaf. Dry-weight to area ratio and metal concentrations were determined from pools of eight samples coming from the same tree. The harvests corresponding to “mature leaves” and to “senescent leaves” used for q-RT-PCR studies were determined as the first harvest of leaves having reached the plateau of maximal SPAD index (June) and as the first harvest showing a clear decrease of SPAD index after the plateau (September), respectively (Fig. 1). The entry into senescence was confirmed at the molecular level by the observation of an increase in the expression of the marker of leaf senescence, CYSTEINE PROTEASE (CP), in all studied genotypes (Bhalerao et al., 2003; Couturier et al., 2010) (Tables S2 and S3). 2.3. Metal concentrations Dried samples were digested in 2 ml of 70% nitric acid in a DigiBlock ED36 (LabTech, Italy) at 100  C for 1 h, 120  C for 6 h and 80  C for 1 h. After dilution in trace metal-free water, the metal content of the samples was determined by atomic absorption spectrometry using an AA240FS flame spectrometer (Agilent, USA). The elements to be analyzed in leaves were chosen to cover most of the metal contaminants present in the field. Pb was omitted because (1) this element is often precipitated and taken up by plants at very low efficiency, (2) very little is known about the molecular mechanisms of lead accumulation in plants. Metal concentrations were determined relative to dry-weight and area. Metal evolution (ME) was, then, calculated according to the equation: ME ¼ [(metal concentration last harvest/metal concentration harvest with max SPAD index)  100]
100. 2.4. Gene expression Total RNA was extracted using RNeasy Plant Mini Kit (Qiagen, Hilden, Germany) following the manufacturer's instructions. Buffer RLC was supplemented with 10 mg ml
1 PEG 6000 to scavenge secondary metabolites. Genomic DNA contaminations were eliminated by adding RNase-free DNase (Qiagen, Hilden, Germany) on spin columns. RNA quantity was evaluated using a spectrophotometer (NanoDrop ND1000; Labtech, France) and RNA integrity was checked by agarose gel separation following denaturation during 10 min at 80  C. Five micrograms of DNA-free RNA was used for reverse transcription by the SuperScript III First-Strand synthesis kit (Invitrogen, Carlsbad, California, USA). Reverse transcriptions were performed using random hexamers, allowing comparison of expression levels between different genes despite different amplicon localizations on cDNAs. Gene-specific PCR primers listed in Table S4 were designed according to the transcript sequences (size, about 200 bp; melting temperature, 60  C) using OligoPerfect™ Designer (http://tools.lifetechnologies.com). Onesixth of the cDNAs was used as template in 10 ml q-RT-PCR reactions performed on a Roche LightCycler 480 (Roche Applied Science, Indianapolis, Indiana, USA). For each genotype, primer specificity was confirmed by analysis of the melting curves and sequencing of the PCR products. Relative transcript levels were calculated following the standard curve method with normalization to the geometric mean of the transcript amounts of the

M. Pottier et al. / Environmental Pollution 199 (2015) 73e82

reference genes EF-1A, PP2A and UBC10 (Vandesompele et al.,2002; Basa et al., 2009). For each poplar genotype, primer amplification efficiencies were determined using five standards from serial dilutions of a pool including the four biological repetitions used for gene expression. Primer combinations had at least 85% efficiency. Poplar genes encoding NRAMP, HMA, ZIP4 and CP were previously identified (Couturier et al., 2010; Migeon et al., 2010). Poplar NAP (Potri.010G166200) was identified based on protein sequence homology with AtNAP/ANAC029 (At1g69490) from Arabidopsis thaliana. 2.5. Statistical analysis Statistical analyses were performed using R sofware package with ManneWhitney and Kruskal-Walis non-parametric test for pair comparisons and multiple comparisons, respectively. For multiple comparisons, a Tukey post-hoc test was performed when significant differences were detected. Different letters indicate significant differences between samples. Principal components and correlation analyses were performed using IBM SPSS Statistics 20 software (SPSS Inc., Chicago, IL, USA). 3. Results 3.1. Poplar field experimentation yields reproducible results The study was conducted on 14 poplar genotypes growing on the metal contaminated field of Pierrelaye in 2011. In 2012, 9 genotypes displaying contrasted behavior in the dynamics of their leaf metal concentrations in 2011 were analyzed again. In a first step, data from the different genotypes were averaged in order to focus on general patterns for each parameter. SPAD index, which correlates with chlorophyll concentration, was measured monthly and used as a marker to monitor leaf development and leaf senescence (Fig. 1). SPAD index values followed a bell-shaped curve on both years. As expected, light green developing leaves exhibited lower SPAD index values than dark green mature leaves. Moreover, as chlorophyll is degraded during senescence, SPAD index values decreased just before leaf abscission. However, SPAD index values decreased earlier in 2012 than in 2011, indicating an earlier entry in senescence during the second year. The variation of dry-weight per unit area showed a continuous increase with leaf development and a rapid decrease just before leaf abscission (Fig. 2). Although similar profiles were observed on

Fig. 2. Evolution of the leaf dry-weight to leaf area ratio (mg cm
2) averaged from 14 to 9 poplar genotypes growing on Pierrelaye contaminated field in 2011 (closed symbol) and 2012 (open symbol), respectively. Error bars represent SE (n ¼ 56 in 2011, n ¼ 36 in 2012). Different letters indicate significant differences between harvests according to a KruskaleWallis test (p < 0.05) followed by a Tukey post hoc test.

75

both years, a surprising difference of almost 1.5 mg cm
2; was observed between the two years. This indicates that leaf dry-weight by area may vary depending on the year and the environmental conditions. To allow better estimation of metal evolution, the general decrease of dry-weight per unit area during senescence was corrected by expressing metal concentrations relative to area. To allow comparison with other studies (e.g. Lettens et al., 2011), concentrations relative to dry-weight are provided in Table S5. Despite differences in leaf development, concentrations of all studied metals exhibited similar dynamics on both years. This indicates that the results obtained are reproducible and of general relevance (Fig. 3). As previously described, metal concentrations in leaves ranked as follows Mg > Zn > Mn]Fe > Cu > Cd (Dominguez et al., 2008; Laureysens et al., 2004; Lettens et al., 2011; Migeon et al., 2009). Interestingly, similar patterns were observed for different metals. Iron (Fe) and copper (Cu) mean concentrations decreased at the beginning of leaf development and then, remained constant until leaf abscission (Fig. 3). In contrast, manganese (Mn), cadmium (Cd) and zinc (Zn) concentrations steadily increased from leaf emergence until the end of the summer as previously reported (Laureysens et al., 2004; Lettens et al., 2011). This increase damped just before leaf abscission (Fig. 3). Finally, magnesium (Mg) concentrations exhibited a unique profile, characterized by a clear decrease just before leaf abscission after the beginning of the decline in the SPAD index (Figs. 1 and 3).

3.2. Poplar genotypes exhibit different metal concentrations in senescent leaves To investigate genotypic variation of metal concentrations in senescent leaves, Cd and Zn concentrations at the last leaf harvest before abscission (between October 5 and 24, depending on the genotype) were compared among 14 poplar genotypes in 2011 and a subset of 9 in 2012 (Fig. 4, Tables S2 and S6). Large variations in both Cd and Zn concentrations were observed between genotypes. Taking into account both years, Cd and Zn concentrations in senescent leaves were about two times higher in Dorskamp than in I214. Interestingly, leaves of genotypes that displayed highest and lowest Cd concentrations namely Dorskamp and I214 were also those that had the highest and lowest Zn concentrations.

3.3. Leaf metals have different fates during autumnal senescence The variations in metal concentrations in senescent leaves among genotypes observed in Fig. 4 could be a consequence of differences in metal accumulation during leaf development or in metal remobilization during autumnal leaf senescence. To discriminate between these two possibilities, an indicator for the relative variation in leaf metal concentration during leaf senescence, which we called metal evolution (ME), was designed as described in Materials and Methods. Despite considerable variation between years and biological repetitions, three different patterns of ME were clearly distinguishable depending on the metal considered (Fig. 5 and S2, Table S2). Mg consistently displayed negative ME values, meaning that leaf Mg concentration decreased during senescence in all studied genotypes. This is in agreement with the results shown in Fig. 3 indicating Mg remobilization before leaf abscission. A similar pattern was observed for Cu concentrations, though not consistently in all genotypes (Figs. 5 and S2). In contrast, ME values for Mn and Fe were highly variable depending on the year and the genotype. Finally, Cd and Zn consistently displayed positive ME values, indicating that their concentrations continued to increase until leaf abscission in almost all genotypes.

76

M. Pottier et al. / Environmental Pollution 199 (2015) 73e82

Fig. 3. Evolution of average concentration of iron, copper, manganese, cadmium, zinc and magnesium in leaves of 14 and 9 poplar genotypes in 2011 (closed symbol) and 2012 (open symbol), respectively. Metals exhibiting similar patterns are represented with the same symbol. Metal concentrations are expressed relative to area of dry leaf (mg cm
2). Error bars represent SE (n ¼ 56 in 2011, n ¼ 36 in 2012). Different letters indicate significant differences between harvests according to a KruskaleWallis test (p < 0.05) followed by a Tukey post hoc test. Metal concentrations expressed relative to dry-weight are presented in Supplemental Table 5.

3.4. Pearson's correlations highlight relationships between metals and specificities among genotypes This field experimentation generated more than 4600 metal measurements acquired throughout leaf life cycle on 14 poplar genotypes in 2011 and 9 poplar genotypes in 2012. To explore the relationships between metals in this large data set, systematic correlation analyses using Pearson's coefficients were performed (Fig. 6, Tables S2, S7 and S8). Pooling together all genotypes, the same Pearson's correlation diagrams were obtained from samples harvested in 2011 and in 2012 supporting the robustness of the results (Fig. 6A). Significant (p < 0.01) correlations were found between Zn and Cd, Cd and Mn, Mn and Zn, Mg and Cu and Cu and Fe (Fig. 6A, Tables S7 and S8). Pooling together metal concentrations from 9 poplar genotypes studied in both 2011 and 2012, different correlations were observed between seasons: whereas Mn was not correlated to any other metal during spring, it became connected to Zn and Cd in autumn (Fig. 6B and C, Table S9). These results suggest that different mechanisms control leaf metal concentrations at different stages of leaf life. Metal correlations were then, investigated separately for each genotype revealing genotype specific correlation diagrams (Fig. 6D, Table S10). The correlation

between Cd and Zn was consistent across all genotypes, whereas the correlations between Mn and other elements exhibited variability. These results suggested that different mechanisms control leaf metal concentrations in different genotypes. 3.5. Gene expression analyses correlate poplar NRAMP1 expression with Cd and Zn bio-concentrations and poplar ZIP4 expression with Cd bio-concentration in leaves To get insights in the molecular mechanism controlling leaf metal homeostasis, expression levels of selected genes were measured in leaves of seven poplar genotypes displaying contrasting leaf metal accumulation patterns (Table S2). The genes used for this analysis were selected based on functional knowledge on their orthologues and on poplar transcriptomic data. Genes encoding putative Zn2þ and Cd2þ transporters highly expressed or up-regulated in senescent leaves were included. Among the Heavy Metal P-type ATPase (HMA) family, we focused on HMA1 and HMA2, the only poplar members in HMA family clusters I and II, which are generally involved in Zn2þ/Cd2þ/Pb2þ extrusion from the cytosol (Migeon et al., 2010). ZIP4, one member of the ZRT-IRT-like Protein (ZIP) family involved in influx to the cytosol of a broad

M. Pottier et al. / Environmental Pollution 199 (2015) 73e82

77

Fig. 4. Cd (A) and Zn (B) concentrations in senescent leaves from 14 to 9 poplar genotypes growing on the Pierrelaye experimental field just before leaf abscission in 2011 (white) and 2012 (black), respectively. Metal concentrations are expressed relative to area of dry leaf (mg cm
2). Error bars represent SD (n ¼ 4). Different letters indicate significant differences between genotypes according to a KruskaleWallis test (p < 0.05) followed by a Tukey post hoc test. Metal concentrations expressed relative to dry weight are presented in Supplemental Table 6.

range of divalent cations (including Zn2þ and Cd2þ) was selected based on its strong up-regulation during leaf senescence, according to microarray data (Migeon et al., 2010). Genes from the Natural Resistance Associated Macrophage Protein (NRAMP) family, which also transport a broad range of divalent cations (including Zn2þ and Cd2þ), were included in the analysis (Migeon et al., 2010). In addition, poplar NAP, a close homologue of NAC genes which plays an essential role in senescence (Guo and Gan, 2006; Waters et al., 2009) and CP, a marker of leaf senescence were included in the analysis (Bhalerao et al., 2003; Couturier et al., 2010). Besides CP and ZIP4, poplar transcriptomic data indicate up-regulation of €din et al., 2009). NRAMP3.2 and HMA2 in senescent leaves (Sjo The raw gene expression data presented in Fig. S3 show wide differences among genotypes in expression levels for a given gene (e.g. NRAMP3.1). These variations triggered us to test whether differences in gene expression could be correlated to leaf metal concentrations. Pearson's correlations between metal concentrations and expression levels of selected genes did not reveal any strong (Rp > 0.4) and significant (p < 0.01) correlation. However, heterogeneity of pollution in the experimental field may have masked a correlation (Table S1). To circumvent this possible bias, metal concentrations in leaves were normalized to their concentrations in soil in order to determine the metal bio-concentration factors (Table S11). When using Zn, Cd and Cu leaf bio-concentration factors in the analysis, positive correlations significant to p < 0.01 were observed between NRAMP1 expression levels and Zn bioconcentration factors (Rp ¼ 0.706; Table S12, Fig. 7) as well as between NRAMP1 expression levels and Cd bio-concentration factors (Rp ¼ 0.467, Table S12). Moreover, a significant negative correlation was also detected between ZIP4 expression levels and Cd bioconcentration factors (Rp ¼
0.406; p < 0.01). This data suggests that ZIP4 could act to reduce Cd accumulation while NRAMP1 may contribute significantly to Zn and Cd accumulation in leaves.

3.6. Pearson's correlations between gene expression levels highlight two gene clusters with distinct putative functions in leaf metal homeostasis In a second step, correlations between the expression levels of the genes involved in leaf metal homeostasis and leaf senescence were investigated. Since the quantitative PCR experiments generated over 660 expression data points (Table S2, Fig. S3) and no obvious correlation emerged from visual observation of raw data (Fig. S3), systematic correlation analyses were undertaken (Fig. 8). Two correlated gene clusters were identified. The first one was made of a core of genes HMA1, NRAMP2 and NRAMP3.1 exhibiting strongly correlated expression levels (0.743 < Rp < 0.915; p < 0.01). NRAMP1 impinged on this core through weaker correlations with HMA1 and NRAMP2 (Rp > 0.500; p < 0.01) (Fig. 8, Table S13). The analysis also underscored a negative correlation between ZIP4 and NRAMP2 (Rp ¼
0.276; p < 0.05). Finally, a second correlated gene cluster containing CP, NAP, HMA2 and NRAMP3.2 (0.320 < Rp < 0.541; p < 0.05) was identified. 3.7. Principal components analyses (PCA) highlighted distinct patterns in metal concentration and gene expression among poplar genotypes and across seasons A factorial analysis using PCA method was performed combining metal concentrations and expression levels of genes involved in metal transport and senescence in leaves (Table 1, Fig. 9). Independent analyses were performed at two different developmental stages: “mature leaf” stage (June) and “senescent leaf” stage (September). PCA revealed two principal components (PC) that accounted for 42% and 55% of the total variance in June and September, respectively. For each PC, the coefficients of parameters explaining the variance are listed in Table 1. Fig. 9 shows dispersion of genotypes according to the two PC. In June, PC1 discriminated genotypes of the Tacahamaca section (Bakan; Skado; Trichobel) from genotypes

78

M. Pottier et al. / Environmental Pollution 199 (2015) 73e82

discriminated between genotypes (Fig. 9A). NRAMP2, NRAMP3.1 and HMA1 expression levels contributed to PC2 with high Eigenvalues (Table 1), suggesting that expression of these genes differentiates genotypes within a section. In September, the scattering of the genotypes along the two PCs was strongly modified with four out of seven genotypes clustering together. PC1 and PC2 clearly discriminated Dorskamp and Bakan from the other genotypes, respectively. Whereas the same parameters contributed to PC2 in June and September, Mg, Cd, Zn, Mn concentrations and NRAMP1 expression level contributed to PC1 with high Eigenvalues in September (Table 1). In conclusion, PCA discriminated poplar sections and poplar genotypes along different component in June; the dispersion of poplar genotypes was very different between mature leaves (June) and senescent leaves (September). 4. Discussion The present investigation aimed at a better understanding of the dynamics of metal concentrations in poplar leaves in order to identify rate-limiting steps that could affect poplar phytoextraction efficiency. Combining measurements of leaf metal concentrations and expression levels of selected genes in a range of poplar genotypes, common patterns could be identified among metals, distinct behaviors among genotypes could be demonstrated and molecular mechanisms could be proposed. 4.1. High variation in leaf metal accumulation and remobilization among poplar genotypes

Fig. 5. Metal Evolution (ME) during senescence in poplar leaves. For each of the 9 poplar genotypes growing on the Pierrelaye experimental field analyzed, average of two independent repetitions performed on 2011 and 2012, each containing at least 4 independent biological samples is presented. ME was determined according the following equation ME ¼ [(metal concentration last harvest/metal concentration harvest with max SPAD index)  100]
100. Error bars represent SD (n ¼ 4).

of the Aigeiros section (Lena; Dorskamp; Vesten; Soligo). Three parameters, Cd, Cu and Mg concentrations, contributed to PC1 with high Eigenvalues, suggesting distinct behavior between different sections with respect to these metals. Within each section, PC2

The analyses reported here reproducibly identified a decrease of leaf dry-weight to leaf area ratio during the autumnal senescence. This decrease likely reflects nutrient remobilization during senescence (Couturier et al., 2010). Common patterns were observed between metals in poplar leaves. Fe and Cu concentrations decreased at the beginning of leaf development and then, remained constant until leaf abscission. In agreement, previous studies reported that the highest Cu concentrations are observed in young leaves of poplar (Laureysens et al., 2004; Lettens et al., 2011). The decrease in Fe and Cu concentrations may correspond to the dilution of an initial pool of metal during leaf expansion. This could indicate that most of the Fe and Cu required for leaf throughout their development cycle are initially present at the onset of leaf emergence. In contrast, leaf Mg concentrations did not decrease until the end of summer but decreased just before leaf abscission, suggesting an active remobilization of this metal. Finally, Mn, Zn and Cd concentrations continuously increased from leaf emergence to leaf senescence. Large variations in leaf metal concentrations were observed between genotypes just before leaf abscission (Fig. 4). This finding is consistent with previous studies that reported variations in Cd concentrations among poplar genotypes (Zacchini et al., 2009). In our study, Dvina and I214 genotypes displayed the lowest metal concentrations in senescent leaves. In order to know if these variations of metal concentrations in senescent leaves are due to the establishment of metal remobilization during autumnal senescence in some poplar genotypes, a new indicator was designed to monitor metal evolution (ME) during the last stages of leaf life. While Mg and, to a lesser extent, Cu were remobilized in almost all genotypes, concentrations of other metals did not decrease during leaf senescence. Actually, for Cd and Zn, which would need to be remobilized in the context of phytoextraction, further accumulation was most frequently observed during leaf senescence. In the case of poplar growing on a metal contaminated soil, the continuous accumulation of TE in senescent leaves could be a strategy to exclude them from the tree through leaf abscission. Despite

M. Pottier et al. / Environmental Pollution 199 (2015) 73e82

79

Fig. 6. Metal correlation analyses. Pearson's correlation analyses performed on metal concentrations measured throughout leaf development on 14 poplar genotypes in 2011 (A). Similar correlation diagrams were obtained using the 9 poplar genotypes studied in 2012. Analyses were then performed pooling data from 2011 to 2012 on the 9 poplar genotypes studied on both years, splitting spring (B) and autumn (C). Similarly, metal correlations were investigated independently for each of the 9 genotypes, pooling data from 2011 to 2012 (D). Metal concentrations (mg cm
2) were collected from leaves of poplar genotypes growing on the experimental field of Pierrelaye. Only Pearson's correlation coefficients greater than 0.4 are represented. Full and dashed lines represent positive and negative correlations significant at the 0.01 level, respectively.

variability among poplar genotypes, Fe, Mn, Zn and Cd were not remobilized during the autumnal leaf senescence in any genotype. This result is in agreement with the report by Laureysens et al. (2004). Similarly, recent works performed on Prunus, Vitis and Carpinus species did not show any remobilization of micronutrients before leaf abscission in these species (Shi et al., 2011). In contrast, previous studies showed that poplar remobilizes about 80% of leaf nitrogen, which is mainly localized in chloroplasts (Couturier et al., 2010). Mg is also localized in chloroplasts being part of chlorophylls. During senescence, chlorophyll degradation releases Mg which could be remobilized before leaf abscission (Zavaleta-Mancera et al., 1999). Concerning Cu, Migeon et al. also observed that Cu concentration in poplar leaves is decreased during autumnal senescence. Interestingly, in their study, this decrease is associated with an increase in Cu concentration in perennial tissues (Migeon et al., unpublished data), suggesting that Cu is remobilized from senescent leaves for storage in woody organs over winter. Thus, during senescence, N, Mg and Cu have a distinct fate compared to Zn, Fe, Mn, and Cd. Previous work performed in A. thaliana (L) Heynh, showed that about 50% of Zn is remobilized from leaves during the monocarpic senescence concomitant to seed filling (Himelblau and Amasino, 2001). It is likely that the lack of metal remobilization in poplar is in part due to low metal sink strength in perennial tissues. Thus, species undergoing distinct

types of senescence, handle leaf TE very differently. 4.2. Correlations between metal concentrations suggest common transport pathways Correlated patterns were observed between Zn, Mn and Cd concentrations as well as between Cu and Fe concentrations (Figs. 3 and 6). Positive correlations between Cd and Zn foliar concentrations in poplar were previously reported (Lettens et al., 2011; Migeon et al., 2009), indicating similar mechanisms of uptake and/or storage. As Cd and Zn share similar chemical properties, the non-essential element Cd may use Zn uptake and distribution pathways to accumulate inside the plant. Several reports already showed that some Zn transporters are able to transport Cd. For instance, HMA2 and HMA4, which are necessary for Zn supply to shoots, provide a major contribution to Cd translocation in A. thaliana and in the hyperaccumulator Noccaea caerulescens (J. Presl & C. Presl) F.K.Mey (Craciun et al., 2012; Wong and Cobbett, 2009). Besides, OsNRAMP5, which is essential for Mn uptake from soil in rice, provides the major pathway for Cd accumulation (Ishikawa et al., 2012). The broad substrate range of some transporters may thus account for correlations observed between Cd and Mn, Zn and Mn and Cd and Zn (Fig. 6). Other correlations may be explained by binding of different metals to the same ligands, such as citrate or

80

M. Pottier et al. / Environmental Pollution 199 (2015) 73e82

€mer et al., 2007). For instance, according to nicotianamine (NA) (Kra stability constants and mutant analyses, NA should be mainly associated with Fe and Cu (Takahashi et al., 2003), which may contribute to the correlation observed between these metals (Fig. 6). 4.3. Correlations between gene expression and Cd and Zn bioconcentrations in leaves suggest an important role for poplar NRAMP1

Fig. 7. Correlation between Zinc bio-concentration factor and NRAMP1 relative expression. Pearson's correlation analysis between zinc bio-concentration and NRAMP1 expression measured in June and September on 7 poplar genotypes. Significant positive correlation was observed (Rp ¼ 0.728; p < 0.01). Zn bio-concentration factor which takes into account the heterogeneity of the pollution is defined as the ratio between Zn concentration in leaves and local available Zn concentration in soil (NH4NO3 extraction method).

Fig. 8. Gene expression correlation analyses. Pearson's correlation analysis of gene expression measured in June and September in 7 poplar genotypes (except for NAP which could only be measured on 5 poplar genotypes) Solid lines represent significant positive correlations and dashed lines represent significant negative correlations. Thin and thick lines indicate significance at the 0.05 and 0.01 levels, respectively.

Table 1 Rotated component matrix extracted showing the different eigenvalues for each component. Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization.

NRAMP1 NRAMP2 NRAMP3.1 NRAMP3.2 HMA1 HMA2 CP ZIP4 Cu Mn Fe Cd Zn Mg

Components June

Components September

1

2

1

2


0.198 0.175
0.085 0.329 0.055
0.028 0.116
0.489 0.885 0.432 0.345 0.838 0.133 0.895

0.399 0.866 0.869
0.333 0.911 0.019
0.142
0.270 0.072
0.341 0.187 0.063 0.201
0.129

0.782 0.079
0.176 0.340 0.255
0.204
0.270
0.417 0.551 0.828 0.445 0.914 0.896 0.884

0.372 0.975 0.909
0.106 0.934 0.007 0.249
0.326 0.035
0.101 0.025 0.095
0.044 0.007

Among all possible correlations investigated, the strongest were positive correlations between NRAMP1 expression levels and Zn bio-concentration factors as well was between NRAMP1 expression levels and Cd bio-concentration factors. Accordingly, Dorskamp, which accumulates the highest leaf Zn and Cd concentrations, displayed the highest expression of NRAMP1 both in June and September (Fig. S3). Poplar NRAMP1 homologues in A. thaliana, rice and the metal hyperaccumulator N. caerulescens have been shown to participate in Mn and Cd accumulation in cells (Cailliatte et al., 2010; Milner et al., 2014; Sasaki et al., 2012). Thus, it is tempting to speculate that poplar NRAMP1 could be involved in metal uptake in poplar leaf cells albeit with a distinct selectivity in favor of Zn. 4.4. Correlations between gene expression levels identify two gene clusters with distinct putative functions Gene expression correlations revealed strong relationships between HMA1, NRAMP2 and NRAMP3.1. Although the function of poplar HMA1 has not been investigated, previous studies performed on A. thaliana showed that its homologue is localized in the chloroplast envelope and plays a major role in Cu transport into this organelle (Seigneurin-Berny et al., 2006). Besides, expression of PtNRAMP2 rescues the yeast smf2 mutant impaired in Mn transport to mitochondria and Golgi apparatus (Luk and Culotta, 2001; Pottier et al., unpublished data). This suggests a function in metal distribution to intracellular organelles for NRAMP2. A. thaliana homologues of poplar NRAMP3.1, AtNRAMP3 and AtNRAMP4 play a role in the transfer of essential metals from the vacuole to the chloroplasts (Lanquar et al., 2010). Thus, HMA1, NRAMP2 and NRAMP3.1 functions revolve around essential metal transfer to intracellular organelles, including chloroplasts, which could account for their correlated expression levels. Moreover, poplar NRAMP1 is associated to this cluster; this transporter would participate in metal entry into cells prior to their distribution to organelles. Accordingly, these genes are expressed in mature leaves where photosynthesis takes place and are not induced during € din et al., 2009). poplar autumnal senescence (Fig. S3) (Sjo Gene correlations also highlighted relationships between poplar HMA2, NRAMP3.2, NAP and CP (Fig. 8). This second cluster includes both the senescence marker CP (Bhalerao et al., 2003; Couturier et al., 2010) and NAP, a close poplar homologue of AtNAP/ ANAC029, which encodes a regulator of leaf senescence in A. thaliana (Guo and Gan, 2006). In agreement, microarray analysis showed induction of HMA2 and NRAMP3.2 expression during €din et al., 2009). Poplar NRAMP3.2 is homologous to senescence (Sjo AtNRAMP3 and AtNRAMP4, which were shown to mediate metal release from vacuoles in A. thaliana (Lanquar et al., 2005, 2010; Oomen et al., 2009). Interestingly, AtNRAMP3 expression is also induced in leaves during monocarpic senescence (Breeze et al., 2011). Poplar HMA2 is homologous to AtHMA2 and AtHMA4, which are involved in Cd and Zn extrusion from the cytosol to the apoplast in A. thaliana (Hussain et al., 2004; Migeon et al., 2010; Verret et al., 2004). Based on the functions of their homologues, it is tempting to speculate that poplar NRAMP3.2 and HMA2 act in concert to release Zn and Cd from leaf cells during senescence.

M. Pottier et al. / Environmental Pollution 199 (2015) 73e82

81

Fig. 9. Scatter plots along the two principal components obtained from PCA discriminated the poplar genotypes in June (A) and September (B) 2011. Leaf metal concentrations and gene expression levels from 7 poplar genotypes in June and September harvests were used for PCA. June distribution of the poplar genotypes according to the eigenvalues obtained for the different measured parameters (explained variance percent: 21.8% for PC1 and 20.8% for PC2) (A). September distribution of the poplar genotypes according to the eigenvalues obtained for the different measured parameters (explained variance percent: 33.7% for PC1 and 21.35%) (B). The eigenvalues obtained for the different measured parameters are listed in Table 1.

However, this would not be sufficient for remobilization, as Zn and Cd accumulation in poplar leaves continues until leaf abscission.

Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.envpol.2015.01.010.

4.5. Principal component analyses highlight differences in metal homeostasis between poplar genotypes The PCA of June harvest data discriminated poplar genotypes according to poplar sections (Fig. 9A). Moreover, PCA also discriminated among genotypes within a section in agreement with the distinct patterns of correlations between leaf metal concentrations (Fig. 6D). This suggests that, in the course of evolution, the mechanisms by which poplars control leaf metal concentrations across development have diverged. The distinct behaviors observed among poplar genotypes with respect to leaf metal concentration and gene expression could account for the variability in metal tolerance and accumulation previously reported in the poplar genus (Castiglione et al., 2009; Induri et al., 2012; Migeon et al., 2009). Phytoextraction relies on the capacity of plants to tolerate and accumulate high concentrations of metals in harvestable parts. Here, we addressed both the dynamics of leaf metal concentrations in poplar throughout the vegetative season and the molecular mechanisms controlling this process. The results highlight the absence of TE remobilization from leaves during senescence in all poplar genotypes tested, which is likely to limit phytoremediation efficiency. The results also suggested that poplar NRAMP1 may represent an interesting target for marker assisted selection or genetic engineering of poplar leaf TE content.

Acknowledgments The authors are grateful to Dr Sylvain Merlot, Dr Astrid Agorio,
pre
for Van Anh Le Thi, Sara Martins, Rafael Costa and Sylvain De contribution to harvests. This work was supported by a grant from
gion Ile-de-France to MP, CNRS funding to the ST laboratory and Re ANR Blanc International program (ANR-10-INTB-1703-01BIOFILTREE).

References Basa, B., Solti, A., Sarvari, S., Tamas, L., 2009. Housekeeping gene selection in poplar plants under Cd-stress: comparative study for real-time PCR normalisation. Funct. Plant Biol. 36, 1079e1087. €rkbacka, H., Birve, S.J., Bhalerao, R., Keskitalo, J., Sterky, F., Erlandsson, R., Bjo €m, P., Gustafsson, P., Lundeberg, J., Jansson, S., 2003. Gene Karlsson, J., Gardestro expression in autumn leaves. Plant Physiol. 131, 430e442. Breeze, E., Harrison, E., McHattie, S., Hughes, L., Hickman, R., Hill, C., Kiddle, S., Kim, Y.S., Penfold, C.A., Jenkins, D., 2011. High-resolution temporal profiling of transcripts during Arabidopsis leaf senescence reveals a distinct chronology of processes and regulation. Plant Cell Online 23, 873e894. Cailliatte, R., Schikora, A., Briat, J.-F., Mari, S., Curie, C., 2010. High-affinity manganese uptake by the metal transporter NRAMP1 is essential for Arabidopsis growth in low manganese conditions. Plant Cell Online 22, 904e917. Castiglione, S., Todeschini, V., Franchin, C., Torrigiani, P., Gastaldi, D., Cicatelli, A., Rinaudo, C., Berta, G., Biondi, S., Lingua, G., 2009. Clonal differences in survival capacity, copper and zinc accumulation, and correlation with leaf polyamine levels in poplar: a large-scale field trial on heavily polluted soil. Environ. Pollut. 157, 2108e2117. Chalot, M., Blaudez, D., Rogaume, Y., Provent, A.-S., Pascual, C., 2012. Fate of trace elements during the combustion of phytoremediation wood. Environ. Sci. Technol. 46, 13361e13369. Couturier, J., Doidy, J., Guinet, F., Wipf, D., Blaudez, D., Chalot, M., 2010. Glutamine, arginine and the amino acid transporter Pt-CAT11 play important roles during senescence in poplar. Ann. Bot. 105, 1159e1169. Craciun, A.R., Meyer, C.-L., Chen, J., Roosens, N., De Groodt, R., Hilson, P., Verbruggen, N., 2012. Variation in HMA4 gene copy number and expression among Noccaea caerulescens populations presenting different levels of Cd tolerance and accumulation. J. Exp. Bot. 63, 4179e4189. Di Baccio, D., Tognetti, R., Sebastiani, L., Vitagliano, C., 2003. Responses of Populus deltoides  Populus nigra (Populus  euramericana) clone I-214 to high zinc concentrations. New. Phytol. 159, 443e452. Dominguez, M.T., Maranon, T., Murillo, J.M., Schulin, R., Robinson, B.H., 2008. Trace element accumulation in woody plants of the Guadiamar Valley, SW Spain: a large-scale phytomanagement case study. Environ. Pollut. 152, 50e59. Guo, Y., Gan, S., 2006. AtNAP, a NAC family transcription factor, has an important role in leaf senescence. Plant J. 46, 601e612. Himelblau, E., Amasino, R.M., 2001. Nutrients mobilized from leaves of Arabidopsis thaliana during leaf senescence. J. Plant Physiol. 158, 1317e1323. Hussain, D., Haydon, M.J., Wang, Y., Wong, E., Sherson, S.M., Young, J., Camakaris, J., Harper, J.F., Cobbett, C.S., 2004. P-type ATPase heavy metal transporters with roles in essential zinc homeostasis in Arabidopsis. Plant Cell Online 16, 1327e1339.

82

M. Pottier et al. / Environmental Pollution 199 (2015) 73e82

Induri, B.R., Ellis, D.R., Slavov, G.T., Yin, T., Zhang, X., Muchero, W., Tuskan, G.A., DiFazio, S.P., 2012. Identification of quantitative trait loci and candidate genes for cadmium tolerance in Populus. Tree Physiol. 32, 626e638. Ishikawa, S., Ishimaru, Y., Igura, M., Kuramata, M., Abe, T., Senoura, T., Hase, Y., Arao, T., Nishizawa, N.K., Nakanishi, H., 2012. Ion-beam irradiation, gene identification, and marker-assisted breeding in the development of low-cadmium rice. Proc. Natl. Acad. Sci. 109, 19166e19171. Kr€ amer, U., Talke, I.N., Hanikenne, M., 2007. Transition metal transport. FEBS Lett. 581, 2263e2272. Kumar, P.B.A.N., Dushenkov, V., Motto, H., Raskin, I., 1995. Phytoextraction: the use of plants to remove heavy metals from soils. Environ. Sci. Technol. 29, 1232e1238. re, C., Baize, D., 2006. Use of major-and trace-element Lamy, I., Van Oort, F., De correlations to assess metal migration in sandy luvisols irrigated with wastewater. Eur. J. Soil Sci. 57, 731e740. Lanquar, V., Lelievre, F., Bolte, S., Hames, C., Alcon, C., Neumann, D., Vansuyt, G., €der, A., Kra €mer, U., Barbier-Brygoo, H., Thomine, S., 2005. Curie, C., Schro Mobilization of vacuolar iron by AtNRAMP3 and AtNRAMP4 is essential for seed germination on low iron. EMBO J. 24, 4041e4051. vre, F., Barbier-Brygoo, H., Krieger-Liszkay, A., Lanquar, V., Ramos, M.S., Lelie Kramer, U., Thomine, S., 2010. Export of vacuolar manganese by AtNRAMP3 and AtNRAMP4 is required for optimal photosynthesis and growth under manganese deficiency. Plant Physiol. 152, 1986e1999. Laureysens, I., Blust, R., De Temmerman, L., Lemmens, C., Ceulemans, R., 2004. Clonal variation in heavy metal accumulation and biomass production in a poplar coppice culture: I. Seasonal variation in leaf, wood and bark concentrations. Environ. Pollut. 131, 485e494. Lettens, S., Vandecasteele, B., De Vos, B., Vansteenkiste, D., Verschelde, P., 2011. Intra- and inter-annual variation of Cd, Zn, Mn and Cu in foliage of poplars on contaminated soil. Sci. Total Environ. 409, 2306e2316. Luk, E., Culotta, V., 2001. Manganese superoxide dismutase in Saccharomyces cerevisiae acquires its metal co-factor through a pathway involving the Nramp metal transporter, Smf2p. J. Biol. Chem. 276, 47556e47562. Migeon, A., Blaudez, D., Wilkins, O., Montanini, B., Campbell, M., Richaud, P., Thomine, S., Chalot, M., 2010. Genome-wide analysis of plant metal transporters, with an emphasis on poplar. Cell. Mol. Life Sci. 67, 3763e3784. Migeon, A., Richaud, P., Guinet, F., Chalot, M., Blaudez, D., 2009. Metal accumulation by woody species on contaminated sites in the north of France. Water, Air, Soil Pollut. 204, 89e101. Milner, M.J., Mitani-Ueno, N., Yamaji, N., Yokosho, K., Craft, E., Fei, Z., Ebbs, S., Clemencia Zambrano, M., Ma, J.F., Kochian, L.V., 2014. Root and shoot transcriptome analysis of two ecotypes of Noccaea caerulescens uncovers the role of NcNramp1 in Cd hyperaccumulation. Plant J. 78, 398e410. vre, F., Blanchet, S., Richaud, P., Barbier-Brygoo, H., Oomen, R.J.F.J., Wu, J., Lelie Aarts, M.G.M., Thomine, S., 2009. Functional characterization of NRAMP3 and NRAMP4 from the metal hyperaccumulator Thlaspi caerulescens. New. Phytol. 181, 637e650. Pulford, I.D., Watson, C., 2003. Phytoremediation of heavy metal-contaminated land by trees a review. Environ. Int. 29, 529e540. Punshon, T., Dickinson, N., Lepp, N., 1996. The potential of Salix clones for bioremediating metal polluted soil. In: Institute of Chartered Foresters, Great Britain, Borders Regional Group, British Ecological Society, Forest Ecology Group (Eds.), Heavy Metals and Trees, Meeting. Heavy metals and trees, Edinburgh,

pp. 93e104. Robinson, B.H., Green, S.R., Chancerel, B., Mills, T.M., Clothier, B.E., 2007. Poplar for the phytomanagement of boron contaminated sites. Environ. Pollut. 150, 225e233. Sanita di Toppi, L., Gabbrielli, R., 1999. Response to cadmium in higher plants. Environ. Exp. Bot. 41, 105e130. Sasaki, A., Yamaji, N., Yokosho, K., Ma, J.F., 2012. NRAMP5 is a major transporter responsible for manganese and cadmium uptake in rice. Plant Cell Online 24, 2155e2167. Sebastiani, L., Scebba, F., Tognetti, R., 2004. Heavy metal accumulation and growth responses in poplar clones Eridano (Populus deltoides  maximowiczii) and I214 (P. x euramericana) exposed to industrial waste. Environ. Exp. Bot. 52, 79e88. Seigneurin-Berny, D., Gravot, A., Auroy, P., Mazard, C., Kraut, A., Finazzi, G., Grunwald, D., Rappaport, F., Vavasseur, A., Joyard, J., Richaud, P., Rolland, N., 2006. HMA1, a new Cu-ATPase of the chloroplast envelope, is essential for growth under adverse light conditions. J. Biol. Chem. 281, 2882e2892. € mheld, V., 2011. Is iron phloem mobile during senesShi, R., Bassler, R., Zou, C., Ro cence in trees? A reinvestigation of Rissmuller's finding of 1874. Plant Physiol. Biochem. 49, 489e493. € din, A., Street, N.R., Sandberg, G., Gustafsson, P., Jansson, S., 2009. The populus Sjo genome integrative explorer (PopGenIE): a new resource for exploring the Populus genome. New. Phytol. 182, 1013e1025. Takahashi, M., Terada, Y., Nakai, I., Nakanishi, H., Yoshimura, E., Mori, S., Nishizawa, N.K., 2003. Role of nicotianamine in the intracellular delivery of metals and plant reproductive development. Plant Cell Online 15, 1263e1280. Unterbrunner, R., Puschenreiter, M., Sommer, P., Wieshammer, G., Tlustos, P., Zupan, M., Wenzel, W.W., 2007. Heavy metal accumulation in trees growing on contaminated sites in Central Europe. Environ. Pollut. 148, 107e114. Vandesompele, J., De Preter, K., Pattyn, F., Poppe, B., Van Roy, N., De Paepe, A., Speleman, F., 2002. Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol. 3 research0034. Vangronsveld, J., Herzig, R., Weyens, N., Boulet, J., Adriaensen, K., Ruttens, A., Thewys, T., Vassilev, A., Meers, E., Nehnevajova, E., 2009. Phytoremediation of contaminated soils and groundwater: lessons from the field. Environ. Sci. Pollut. Res. 16, 765e794. Verret, F., Gravot, A., Auroy, P., Leonhardt, N., David, P., Nussaume, L., Vavasseur, A., Richaud, P., 2004. Overexpression of AtHMA4 enhances root-to-shoot translocation of zinc and cadmium and plant metal tolerance. FEBS Lett. 576, 306e312. Waters, B.M., Uauy, C., Dubcovsky, J., Grusak, M.A., 2009. Wheat (Triticum aestivum) NAM proteins regulate the translocation of iron, zinc, and nitrogen compounds from vegetative tissues to grain. J. Exp. Bot. 60, 4263e4274. Wong, C.K.E., Cobbett, C.S., 2009. HMA P-type ATPases are the major mechanism for root-to-shoot Cd translocation in Arabidopsis thaliana. New. Phytol. 181, 71e78. Zacchini, M., Pietrini, F., Mugnozza, G.S., Iori, V., Pietrosanti, L., Massacci, A., 2009. Metal tolerance, accumulation and translocation in poplar and willow clones treated with cadmium in hydroponics. Water, Air, Soil Pollut. 197, 23e34. Zavaleta-Mancera, H., Franklin, K., Ougham, H., Thomas, H., Scott, I., 1999. Regreening of senescent Nicotiana leaves. II. Redifferenciation of plastids. J. Exp. Bot. 50, 1683e1689.