Introduction

Understanding the mechanism of heavy metal (HM) homeostasis and transport in metal tolerant and accumulator plants could become vital in formulating a cheap and environmentally friendly strategy for soil decontamination. HM transporters play key roles in the uptake, transport, sequestration, and efflux of metals and are among the most studied homeostatic genes relating to metal tolerance and accumulation abilities (Kim et al. 2004; Papoyan and Kochian 2004; Pittman et al. 2005). In this context, several candidate HM transporters have been cloned and characterized from various metal tolerant and hyperaccumulator plants (Kramer et al. 2007; Mizuno et al. 2005; Roberts et al. 2004). The metal accumulation levels of Brassica juncea do not qualify it as a ‘hyper accumulator’. However, its ability to accumulate and transport a wide variety of HMs like lead, cadmium, chromium, copper, zinc, and uranium (Duquene et al. 2009) to the shoots, in addition to its fast growing and high biomass producing nature, makes it more desirable for phytoremediation applications than ‘model’ hyperaccumulators such as Noccaea (Thlaspi) caerulescens (Chang et al. 2005; Milner and Kochian 2008). B. juncea is an allopolyploid (AABB) of B. rapa (n = 10, AA) and B. nigra (n = 8, BB). Polyploidy is a common phenomenon in the evolution of flowering plants that contributes to the duplication of genes leading to neofunctionization and subfunctionization events. The consequent expansion of desired genes often results in increased adaptive fitness of the polyploid offspring compared to their parents (Dubcovsky and Dvorak 2007; Leitch and Leitch 2008). Moreover, recently it was proposed that gene copy expansion contributes towards high basal level expression as a strategy for metal hyperaccumulation in zinc hyperaccumulator Arabidopsis halleri (Hanikenne et al. 2008). The current information makes it apparent that orthologs of metal homeostatic genes in hyperaccumulator and tolerant plants do not necessarily vary in their functionality, but more likely in their expression levels or expression patterns (Gendre et al. 2007; Oomen et al. 2009). We therefore attempted to search for HM transporters as well as their allelic variants from B. juncea, with the objective of understanding their role in the broad-spectrum metal tolerance and root-to-shoot transport ability in the plant. The transporter families chosen in the study were natural resistance-associated macrophage proteins (NRAMPs) and yellow stripe-like proteins (YSLs) for their known role in subcellular and intracellular transport of metals, respectively.

Genes encoding NRAMPs represent a family of proteins with significant HM homeostatic functions ubiquitously found in bacteria, fungi, animals, and plants (Cellier et al. 1995). In Arabidopsis thaliana, NRAMPs have been shown to transport iron primarily. A. thaliana expresses six NRAMP isoforms, of which distinct roles in metal transport have been detected for AtNRAMP1-4 (Curie et al. 2000). Plants overexpressing AtNRAMP1 and 2 show an increased resistance to iron overload (Curie et al. 2000). AtNRAMP3 and AtNRAMP4 on the other hand are induced by iron starvation and are required for mobilization of vacuolar iron stores during germination of seeds under low iron conditions (Lanquar et al. 2005; Lanquar et al. 2010). Some of these NRAMPs are also involved in transporting metals other than iron albeit with lower affinity (Kaiser et al. 2003; Thomine et al. 2003; Thomine et al. 2000). Only recently AtNRAMP 1, 3, and 4 were implicated in manganese homeostasis (Cailliatte et al. 2010; Lanquar et al. 2010). AtNRAMP4 also has a prominent role in zinc transport (Oomen et al. 2009). AtNRAMP3 and 6 were shown to be involved in homeostasis of nonessential metals like cadmium (Cailliatte et al. 2009; Thomine et al. 2003). NRAMP isoforms from hyperaccumulating plants are relatively less characterized. In a transcriptomics study with zinc hyper accumulator A. halleri, NRAMP3 was found to show high level of expression relative to non-accumulator A. thaliana (Talke et al. 2006). Similar high basal level expression was also reported for TcNRAMP3 and 4, orthologs of AtNRAMP3 and 4 from the zinc, cadmium, and nickel hyperaccumulator Noccaea (Thlaspi) caerulescens (Milner and Kochian 2008). TjNRAMP4, a NRAMP isoform from the Ni hyper accumulator plant Thlaspi japonicum, showed an unusual specificity for nickel in yeast uptake experiments unlike the Arabidopsis and Noccaea (Thlaspi) caerulescens isoforms (Mizuno et al. 2005). The second family used in this study was the YSL family of transporters. These proteins belong to a distinct class of oligopeptide transporter proteins found in archaebacteria, bacteria, fungi, and plants. YSLs constitute a plant-specific clade within these transporters that transport metal-bound nonreactive carrier molecule like secondary amino acids (Lubkowitz 2006). YSL proteins gained importance in plant metal homeostasis through the identification of the maize ZmYS1 gene (Curie et al. 2001) which is implicated in the uptake of Fe phytosiderophore complexes at the root. YSL family of transporters are also involved in root-to-shoot transport of iron bound to the plant-borne chelator nicotinamine (NA) (Curie et al. 2009). Other YSL proteins might be involved in the long distance transport of metals other than iron. OsYSL2, a rice YSL homolog, was shown to specifically transport Fe(II) NA and Mn(II) NA complexes. Consistent with the role of YSL proteins in long distance transport, AtYSL2 is expressed in the root endodermis and pericycle cells facing the meta-xylem tubes and was predicted to control lateral movement of metal through the vasculature and loading of metals into xylem (DiDonato et al. 2004; Schaaf et al. 2004). In graminaceous plants like barley and rice, YSLs are implicated in metal acquisition through phytosiderophore complexes (PS). For example, HvYS1 uses Fe(III)–PS complex as substrate (Harada et al. 2007). OsYSL15, the rice homolog of ZmYS1 and OsYSL18 are involved in transport of Fe(III)–PS complexes in a tissue- or organ-specific manner (Aoyama et al. 2009; Inoue et al. 2009). Apart from iron and manganese, certain YSLs also transport Zn phytosiderophore complex, and their expressions depend on the zinc status (Schaaf et al. 2004; van de Mortel et al. 2006). YSLs vary widely in their in planta distribution, and in the way they respond to different metal ions (Gendre et al. 2007; Koike et al. 2004; Le Jean et al. 2005). Response of NRAMP and YSLs to metal deficient, sufficient or excess conditions depend on the isoforms and also on the plant species (Gendre et al. 2007; Lee et al. 2005; Schaaf et al. 2004; van de Mortel et al. 2006). In hyperaccumulator Noccaea (Thlaspi) caerulescens (J.& C Presl, Ganges), three YSL genes were found to show distinct tissue-specific expression, among which TcYSL3 was found to be a transporter of both Fe(II)NA and Ni(II)NA complex.

Amplified rDNA restriction analysis (ARDRA) is a molecular fingerprinting technique routinely used for identification and characterization of bacterial species from complex mixed populations like soil and body fluids (Stakenborg et al. 2005). The technique utilizes restriction analysis of PCR-amplified 16SrRNA genes from bacterial pools, preferably with four cutter enzymes, for diagnosis of closely related bacterial species. A specific isolate can be distinguished by the different digestion patterns detected on agarose gels. This provides an effective and inexpensive method for identification and distinction of closely related sequences instead of employing large-scale sequencing. In this study, ARDRA was adopted together with degenerate primer-based amplification of gene families from B. juncea to clone and distinguish several family members/allelic variants. It can be a useful method for high throughput cloning of any gene family from a relatively less characterized organism.

We cloned 23 NRAMP and 27 YSL ESTs from allotetraploid B. juncea using this method. Many of these may potentially represent distinct isoforms/allelic variants/alternatively spliced forms. These were classified according to their homology with known A. thaliana and N. caerulescens genes. Sub members belonging to each family varied in their EST abundance and expression, which can have important functional consequence in contributing to tolerance and accumulation properties. B. juncea L. Czern var. 211000 used in our study showed high levels of cadmium and lead tolerance in vitro. Semi-quantitative RT-PCR-based expression analysis of randomly chosen members showed that different members responded differentially to metal treatment, both essential and nonessential, depending on the tissue in question. Interestingly, BjYSL6.1 showed high levels of expression in shoots specifically when plants were treated with cadmium, and BjYSL5.8 showed elevated expression in roots treated with lead. Cloning of full-length genes of chosen candidates indicated that they encoded functional metal transporter. In addition, Brassica-specific changes in the predicted extracellular as well as intracellular loop for the translated proteins indicated their possible functional diversity from Arabidopsis orthologs.

Materials and methods

Plant materials and growth conditions

Brassica juncea PI 211000 seeds were obtained from USDA, ARS, NCPRIS, Regional Plant Introduction station (USA). Seeds were surface sterilized with 1% bavistine (fungicide, Sun Agro Chemical Industries, India) for 5 min followed by 30 s treatment in absolute alcohol and 5 min treatment in 0.2% HgCl2. Seeds were then germinated on germination media (3% sucrose, 1% agar) in a growth chamber under 16 h light and 8 h dark cycles at 24°C and 60% relative humidity. Seven-day-old seedlings were transferred to Murashige and Skoog media (MS) containing 10 mM MES–NaOH pH 6 supplemented with 3% sucrose under control and iron deplete (−Fe) conditions and were maintained for five additional days. The control and −Fe plants were frozen in liquid nitrogen and stored at –80°C until RNA extraction.

RNA isolation and reverse transcriptase reaction

Total RNA was isolated from the frozen tissue of control and −Fe plants using Plant RNA isolation reagent (invitrogen). RNA integrity was checked on a 1.5% formaldehyde agarose gel. RNA was treated with DNase (10 Units/2.5 μg of RNA) to remove genomic DNA contamination before use in reverse transcriptase reaction. Five micrograms of total RNA was used to synthesize first strand cDNA with anchored oligo dT primer using Superscript II reverse transcriptase (Invitrogen, 200 unit/2.5 μg RNA) as per manufacturer’s instruction. The cDNAs were treated with E. coli RNase H (2 Units) and used for PCR amplification. For semi-quantitative RT-PCR (SQ RT-PCR), 300 ng of total RNA was used from root and shoots of plants grown for 5 days under control (C), iron deplete (−Fe), iron excess (Fe, 1 mM), manganese deplete (−Mn), manganese excess (Mn, 1 mM), zinc deplete (−Zn), zinc excess (Zn, 0.2 mM), lead-EDTA (Pb 75 μM, 0.25 mM EDTA), or cadmium (Cd, 10 μM) conditions.

RT-PCR and nested PCR amplification of NRAMP and YSL genes

150 ng of cDNA was used for PCR amplification using 3′ anchor primer (A) and degenerate gene-specific primers (NF1 for NRAMP forward primer 1, YF1 for YSL forward primer 1). This PCR product was used for nested PCR using two degenerate gene-specific forward primers (NF1, NF2 for NRAMPs and YF1, YF2 for YSLs) in combination with two degenerate gene-specific reverse primers (NR1, NR2 for NRAMP and YR1, YR2 for YSL family, respectively) in a 40 μl reaction volume containing 5 μl of 10× PC2 reaction buffer with MgCl2 as described by (Barnes 1994) with 0.25 mM dNTP (15:1) Taq (Sigma): Phusion (Finnzyme) mix, 1 μl of 10 μM of each primer. PCR was performed on a Bio-Rad cycle sequencer. The primer sequences and PCR parameters are shown in Supplementary Table 1.

Cloning and verification of cloned nested product

The PCR products were directly used for cloning in cloneJET™ vector (Fermentas) as per manufacturer’s instruction. The ligated products were transformed in chemically competent E. coli Mach1 cells (invitrogen) and selected on Ampicillin (100 μg/ml) plates. Colonies obtained were screened by colony PCR using the vector-specific (pJET1.2) forward (5′CGACTCACTATAGGGAGAGCGGC3′) and reverse (5′AAGAACATCGATTTTCCATGGCAG3′) primers. The PCR products were run on 1% agarose gel. Positive clones within the expected size ranges (Table 1) were grown in ampicillin-selected cultures, and plasmids were prepared using the Purelink quick Plasmid miniprep kit (Invitrogen). The clones with apparent size differences were directly sequenced, and clones with same size in an agarose gel were used for ARDRA-based molecular fingerprinting.

Table 1 Expected size ranges of the amplified PCR products as per the alignment of all AtNRAMP and AtYSL family members
Full size table

ARDRA-based molecular fingerprinting technique to differentiate family members

Inserts were amplified from plasmids with pJET1.2 forward and reverse primers. PCR products were digested with HhaI (Vivantis) and MspI (Fermentas) and run on a 1% agarose gel along with uncut amplified inserts (Lane U). The digestion patterns were compared visually. Clones showing unique digestion patterns after using both the enzymes were sequenced.

Sequence processing and domain analysis

The sequences were assembled using Contig express (Vector NTI Advance™ 10, Invitrogen). The contigs were manually curated for errors and redundant sequences. Blastn and Blastx (NCBI) were used for homology searches (Altschul et al. 1997); the Blastx results were used to classify clones in different isoforms. The nucleotide sequences of the clones were translated in all reading frames using Vector NTI Advance™ 10 (Invitrogen), and the correct frame was chosen for each clone. Translated protein sequences for BjNRAMPs and BjYSLs were aligned with their closest orthologs with AlignX (Vector NTI Advance™ 10, Invitrogen), and the longest translated partial protein was chosen for representing the domains for each of the isoforms (boxed region). Transmembrane domains (TM) were drawn according to the prediction for their Arabidopsis orthologs by Aramemnon 6.2. (http://aramemnon.botanik.uni-koeln.de/) using a cut off score of 0.6 for each TM domain. Encircled residues are positions of amino acids in the closest Arabidopsis/Noccaea orthologs that have undergone B. juncea-specific changes. These changes (shown at the apex of the circle separated by “/”) represent amino acid variations in all the sub members for an isoform and are shown by seven diagrams. The phylogenetic tree was created with Mega 4.1 (Beta), applying the maximum parsimony method using translated NRAMP and YSL isoforms as input sequences.

Full length cloning of chosen NRAMP and YSL isoforms

3′ RACE strategy was employed to obtain 3′ end of the partial genes (Sambrook and Russell 2001) using gene-specific forward primer and anchord oligo-dT primer combinations. For obtaining 5′ ends of the genes, degenerate forward primers were designed by aligning the closest known ortholog from A. thaliana or N. caerulescens (T. caerulescens) with available Brassica (taxid 3705) ESTs obtained through BlastN analysis. Both for 3′- and 5′-end clonings, a primary PCR was carried out followed by nested PCR. The primers and PCR conditions are represented in Supplementary Table 2.

Semi-quantitative RT-PCR

cDNA obtained from roots and shoots of control (C), metal deplete (−Fe, −Mn, −Zn) and metal excess conditions [Fe (1 mM), Mn (1 mM), Zn (0.2 mM), Pb (75 μM), and Cd (10 μM)] were used for SQ RT-PCR. In SQ RT-PCR for essential metal ions (Fe, Mn, Zn), cDNA was diluted (1:125) and used for PCR. In SQ RT-PCR for non-essential metal ions (Pb, Cd) undiluted cDNA was used. Allele-specific primers were used at 200 nM final concentration, and PCR conditions are described in Supplementary Table 4. A primer was designed for actin 2 (At3g18780) (Schenk et al. 2003) and was used as internal control. SQ RT-PCR experiment was done in triplicate using two independent cDNAs. Representative results are shown in Figs. 6 and 7. Intensities of all bands from the representative data were densitometrically estimated by Image Quanta TL v2005, normalized against the intensity of the actin transcript, and the relative intensities were plotted by Graph pad prism version 4 for better understanding of the data. Representative results for the biological replicates are given in Supplementary Fig. S5.

Root–shoot assay and expression analysis under heavy metal lead and cadmium

Four-day-old B. juncea plants grown in germination media were transferred to MS media supplemented with lead-EDTA [75 μM Pb(NO3)2, 0.25 mM EDTA] or cadmium (10 μM CdSO4) and grown for another 5 days. Root–shoot lengths and leaf area were measured from eight independent plants grown under control, control supplemented with 0.25 mM EDTA or treated conditions and plotted using Graph pad prism version 4.

Results

RT-PCR amplification and cloning of NRAMP and YSL gene family

Gene-specific degenerate primers (Supplementary Table 1) were designed based on conserved regions of A. thaliana NRAMP and YSL genes. RNAs were isolated from plants grown under metal replete (C) and iron deplete (−Fe) conditions in accordance with the expression conditions reported for most A. thaliana NRAMP and YSL genes (Curie et al. 2001; Thomine et al. 2003). The cDNAs obtained from reverse transcription were subjected to PCR amplification using degenerate gene-specific forward and anchor primers [Fig. 1 (a, b for NRAMP; e, f for YSL)], followed by nested PCR using the same forward primers in combination with the degenerate gene-specific reverse primers [Fig. 1 (c, d for NRAMP; g, h for YSL)]. The primers used were designed to amplify orthologs of AtNRAMP2, 3, 4 only, which are phylogenetically distinct from AtNRAMP 1 and 6. For YSLs, the degenerate primers were predicted to amplify Brassica orthologs of all eight Arabidopsis YSLs. Amplicons of desired size range (Table 1) were produced with all primer combinations (Fig. 1e–h) except YF1–YR1 primer set in iron deplete condition (Fig. 1g; −Fe). The nested products were used for cloning. We expected to amplify ESTs of different isoforms as well as their allelic variants with our strategy using degenerate gene-specific primers. Therefore, all colonies obtained in the cloning procedure (40–50 per plate) were screened for inserts of desired size by colony PCR (Table 1). Different clones with distinct but desired sizes were obtained (Supplementary Fig. S1), and clones representing each separate size class were sequenced at once. Another type of colony PCR results showed inserts of the same sizes (Fig. 2). For primer combination NF2–NR1, colony nos. 3, 5, 9, 11, 12, 13, and 17 from control (left panel; −Fe) and colonies 20, 21, 24, 26, and 27 (right panel; C) showed a single size class (Fig. 2a). Similarly in YSLs, for YF1–YR1 primer combinations, colony nos. 4, 5, 6, 12, 13, 15, 18, 19, 20, and 21 showed only one size class (Fig. 2b). These were subjected to molecular fingerprinting analysis with the assumption that they may include different isoforms/allelic variants not separable on agarose gel. Clones having inserts beyond the desired size range were not analyzed further.

Fig. 1
figure 1

Amplification of NRAMP and YSL gene family with gene-specific degenerate primer combinations. Total RNA was extracted from Brassica juncea plant grown under control (C) and iron deplete (−Fe) condition. The RNA was reverse transcribed using an Anchor oligo dT primer, and RT product was used for PCR amplification with NRAMP- and YSL-specific degenerate forward primer and Anchor primer (A) combination followed by nested PCR. Primer sequences and PCR conditions are summarized in Supplementary Table 1. a Products of NRAMP gene-specific degenerate forward 1 (NF1) and Anchor primer (A). b Products of NRAMP gene-specific degenerate forward 2 (NF2) and Anchor primer (A). c Nested products of NRAMP gene-specific degenerate forward 1 (NF1) and gene-specific degenerate reverse 1 and 2 (NR1 and NR2). d Nested products of NRAMP gene-specific degenerate forward 1 (NF2) and gene-specific degenerate reverse 1 and 2 (NR1 and NR2). e Products of YSL gene-specific degenerate forward 1 (YF1) and Anchor primer (A). f Products of YSL gene-specific degenerate forward 2 (YF2) and Anchor primer (A). g Nested products of YSL gene-specific degenerate forward 1 (YF1) and gene-specific degenerate reverse 1 and 2 (YR1 and YR2). h Nested products of YSL gene-specific degenerate forward 1 (YF2) and gene-specific degenerate reverse 1 and 2 (YR1 and YR2). Molecular weight marker (M; Fermentas 1 kb) is shown in the middle lane

Full size image
Fig. 2
figure 2

Representative colony PCR result for NRAMP and YSL clones showing inserts of same size. PCR products obtained by family-specific primers were cloned. All the colonies obtained under ampicillin selection were screened for inserts. The lane numbers in the agarose gel corresponds to the colony numbers. a NRAMPs (NF2–NR1 primer combination) colony nos. 3, 5, 9, 11, 12, 13, 17 (left panel −Fe), 20, 21, 24, 26, and 27 (right panel C) showed one distinct size variant. Colonies 12, 13 (−Fe), 26, and 27 (C) with identical sizes coded for BjNRAMP3.8, BjNRAMP3.4, BjNRAMP4.2 and BjNRAMP3.9, respectively (distinguished through ARDRA analysis). b YSLs (YF1–YR1 primer combination) colony nos. 4, 6; 5, 12; 13, 15; 18 19, 20, 21 (C) showed one distinct size variant. Colonies 4, 5, 12, 13, 18, 19, 20, and 21 with identical sizes coded for BjYSL5.8, BjYSL8.1, BjYSL5.11, BjYSL5.1, BjYSL6.1, BjYSL2.1, BjYSL6.4, and BjYSL5.12, respectively (distinguished through ARDRA analysis)

Full size image

ARDRA-based restriction analysis of clones of same size

Brassica juncea is an allopolyploid plant with a 6.8 times larger genome than that of A. thaliana (http://www.brassica.info/info/reference/genome-sizes.php, http://data.kew.org/cvalues/). Its genome has several blocks that have undergone duplication (Panjabi et al. 2008) during evolution relative to Arabidopsis. Brassica therefore can have larger number of isoforms and/or allelic variants for a single gene compared to Arabidopsis. Clones showing identical insert size in colony PCR analysis were subjected to molecular fingerprinting based on ARDRA to distinguish between isoforms/allelic variants. The inserts were amplified using vector-specific primers followed by restriction analysis with two frequent cutters HhaI and MspI. Sequences with more than 90% identities at nucleotide level could be distinguished by this fingerprinting technique.

A representative restriction analysis of NRAMP (NF2–NR1) and YSL clones (YF1–YR1) of same size produced four (Fig. 3a) and eight (Fig. 3b) different unique patterns, respectively. The upper and lower panels in Fig. 3 show gels for HhaI and MspI digestion. NRAMP and YSL clones with distinct pattern were assigned with unique symbols (Fig. 3). Results are summarized in Table 2. For example, colony numbers 4 and 19 (Fig. 3b) showed same pattern in Hha1 digestion (Fig. 3b, upper panel), but a different pattern in Msp1 digestion (Fig. 3b, lower panel). Clones having unique fingerprinting patterns were sequenced. The digestion patterns for rest of the submitted clones are presented in Supplementary Figs. S2 and S3. Table 3 shows the number of clones having the same fingerprints to the clones submitted to Genbank. Most clones were obtained more than once in the study indicating saturation in obtaining the ESTs.

Fig. 3
figure 3

Restriction digestion result for NRAMP and YSL clones of same size showing different digestion patterns. Plasmids were prepared from colonies showing the same insert sizes. The inserts were amplified with vector-specific primers from the plasmids and digested with HhaI and MspI. The lane numbers in the agarose gel correspond to the colony numbers. a From NF2–NR1 primer combinations, same-sized inserts were used for the molecular fingerprinting technique with HhaI (upper panel) and MspI (lower panel). Four distinct patterns were obtained from the HhaI digestion. Colony nos. 3, 5, 13, 17, 21, 24 (¤), colony nos. 9, 20, 26 ( ), colony no. 12 ( ), colony no. 27 (●). Two distinct patterns were obtained from the MspI digestion corresponding to colony nos. 3, 5, 9, 12, 13, 17, 20, 21, 24, 26 ( ) and colony no. 27 (θ). Colony nos. 12, 13, 26, and 27 coded for BjNRAMP3.8, BjNRAMP3.4, BjNRAMP4.2, and BjNRAMP3.9, respectively. b From YF1–YR1 primer combinations, same-sized inserts were used for the molecular fingerprinting technique, HhaI (upper panel) and MspI (lower panel). Four distinct patterns were obtained from the HhaI digestion corresponding to colony nos. 4, 15, 19 (Φ), colony nos. 5, 6, 18, 20 (†), colony no. 12 ( ) and colony nos. 13, 21 ( ). Seven distinct patterns were obtained from the MspI digestion corresponding to colony nos. 4, 15 (●), colony nos. 5, 6 ($), colony nos. 12, 21 (©), colony no. 13 (§), colony no. 18 (@), colony no. 19 (ŧ), and colony no. 20 (¢). Colony nos. 4, 5, 12, 13, 18, 19, 20, and 21 codes for BjYSL5.8, BjYSL8.1, BjYSL5.11, BjYSL5.1, BjYSL6.1, BjYSL2.1, BjYSL6.4 and BjYSL5.12, respectively

Full size image
Table 2 Fingerprinting data of representative BjNRAMP and BjYSL family members
Full size table
Table 3 Numbers of BjNRAMP and BjYSL clones showing same fingerprinting result after HhaI and MspI restriction digestion
Full size table

Sequence and phylogenetic analysis of BjNRAMP and BjYSL isoforms

NRAMP and YSL sequences obtained with different primer combinations were assembled in contigs, and the sequences were manually curated for sequence errors. Twenty-three sub members of NRAMPs were classified into three isoforms, and 27 YSL sub members were classified into four isoforms through Blastx analysis. Table 4 shows the cloned isoforms along with their accession numbers and the closest known homolog in Blastx and Blastn analysis. These showed high degree of sequence similarity in the nucleotide (up to 97%) and protein levels (94–99%) with their Arabidopsis and Noccaea (Thlaspi) counterparts. Since the technique used here was PCR based, differences in nucleotide levels were only taken to be significant when present in more than one clone obtained in independent PCR amplifications. The clones were aligned with their closest A. thaliana or N. caerulescens (Thlaspi caerulescens) homolog and with Brassica (taxid 3705) ESTs obtained by Blastn analysis. Any isolated clone with single base pair change was ignored. The clones were translated, and the proteins aligned using the AlignX program and phylogenetic trees were generated (Fig. 4a, b). Orthologs of all the Arabidopsis NRAMPs and YSLs other than the pollen- or flower-specific isoforms with the only exception of AtYSL2 (Table 4) were obtained. AtNRAMP1 and 6 sequences were omitted while designing primers in the initial study in order to reduce degeneracy. We obtained several Brassica orthologs of AtNRAMP1 and 6 (BjNRAMP 1.1-1.4 and BjNRAMP 6.1-6.4) using the same technique when suitable degenerate gene-specific primers were designed (Supplementary Fig. S4). BjNRAMP2 and BjNRAMP3 similar to AtNRAMP2 and NcNRAMP3 had maximum number (8 and 10, respectively) of sub members. Similarly, BjYSL5 ortholog of AtYSL5 had maximum number of sub members (15). We obtained full-length clones of BjNRAMP4.1, BjYSL6.1, and BjYSL6.4 genes. In silico translation indicated that these ESTs encoded functional proteins.

Table 4 Submitted clones with corresponding accession number and closest homolog
Full size table
Fig. 4
figure 4

Phylogenetic tree of BjNRAMP and BjYSL partial proteins. Phylogenetic trees were created using maximum parsimony method MEGA 4.1 beta with translated protein of all the BjNRAMP and BjYSL clones and their closest homolog (mentioned in Table 4). a Phylogeny of BjNRAMP members. b Phylogeny of BjYSL members

Full size image

Comparative analysis of BjNRAMP and BjYSL proteins with closest orthologs

Transmembrane domains and topology of the in silico translated BjNRAMP and BjYSL proteins were predicted by Aramemnon 6.2 as described in “Materials and methods”. Sub members for each isoform were aligned with their closest Arabidopsis ortholog. The boxed regions in each diagram show the cloned segment of the largest Brassica submember relative to the Arabidopsis counterpart. The encircled numbers indicate the positions of the amino acid residues in Arabidopsis/Noccaea full-length proteins that had undergone changes in the Brassica sub members (Fig. 5). Brassica-specific changes in amino acid residues were observed in both BjNRAMP and BjYSL proteins. The details of these changes in the transmembrane, extracellular, or cytoplasmic loops are summarized in the Supplementary Table 3. Changes in the transmembrane domains were mostly conservative and usually resulted in the replacement of a residue by another similar residue. Only nonconservative changes which can have important influence on the function of these proteins in B. juncea are discussed in this section. Overall, BjNRAMP isoforms within the cloned region exhibited less changes compared to BjYSL proteins. BjNRAMP2 sub members relative to its closest homolog AtNRAMP2 showed either no or conservative changes (Fig. 5a). Although only few, nonconservative changes did occur in the extracellular loops of BjNRAMP3 (R132G, N312D and F325D) and 4 (E259K and Q426H).

Fig. 5
figure 5figure 5

Protein domain analysis of BjNRAMP and BjYSL clones. All the BjNRAMP and BjYSL isoforms were translated and aligned with their closest homolog. TM domains of the translated BjNRAMP and BjYSL proteins are marked relative to their Arabidopsis closest homolog according to TM prediction using Aramemnon 6.2. Arabidopsis ortholog for BjNRAMP2, BjNRAMP3, and BjNRAMP4 were AtNRAMP2, AtNRAMP3, and AtNRAMP4, respectively. The same for BjYSL2, BjYSL5, BjYSL6, and BjYSL8 were AtYSL5, AtYSL5, AtYSL6, and AtYSL8, respectively. The encircled positions indicate the positions of the residue in the Arabidopsis/Noccaea ortholog that changed in B. juncea. Amino acid changes for a particular position between all the submembers of a given isoforms are separated by ‘/’. The predicted extracellular side of the protein is marked with ‘E’. a Domain analysis of all submitted members of BjNRAMP. b Domain analysis of all submitted members of BjYSL

Full size image

In contrast, marked changes were detected in the predicted extracellular loops of all BjYSLs relative to their closest orthologs (Fig. 5b). The cytoplasmic loops had relatively less changes. But the extracellular loop between transmembrane domains VI and VII of BjYSL6 and 8 and VII-VIII of BjYSL5 and 2, which were predicted to be the substrate binding site (Lubkowitz 2006) in these proteins, harbored most of the non-conservative substitutions. In BjYSL6, the non-conservative changes between TM VI–VII relative to AtYSL6 were S350N, R351S, R354H, T361A, D362N, D365V, S367D and F635Y. The 15 members of BjYSL5 (5.1–5.15) exhibited small truncations or additions in the extracellular substrate-binding loop. In addition, nonconservative changes were also seen in the cytoplasmic loops between TM VI and VII in these isoforms (D325A, N326D, and N616 R/K/Q). Other prominent changes included three proline residues that were distributed in different cytosolic loops (P328E, S329P, and H614P). BjYSL2 sub members showed very similar changes to BjYSL5 in the extracellular substrate binding sites and the intracellular loop between TM VI–VII and TM XII–XIII, but had additional nonconservative changes in the intracellular loop between TM VIII and IX (H437Q, H440D, I452T, and C453F). The substrate-binding loop in BjYSL8 sub members were considerably shorter than its ortholog AtYSL8 and contained T355M, I369S, S377K, K382R, A383T, G384S, K392E, D394P as the non-conservative changes.

Semi-quantitative RT-PCR analysis of chosen isoforms

Semi-quantitative, allele-specific PCR-based expression analysis was performed with chosen isoforms of BjNRAMP and BjYSL. The specificity of the primers used for the study was validated using all NRAMP or YSL clones as templates before using them in the expression analysis (data not shown). The expression levels of all the transcripts were normalized against the intensity of the Actin 2 transcript, and relative intensities were plotted for each sub members. The NRAMP submembers chosen for such expression were BjNRAMP3.5, BjNRAMP4.1, and BjNRAMP4.3, and the YSL submembers were BjYSL2.1, BjYSL5.8, and BjYSL6.1. Plants were grown under metal (iron, manganese, and zinc) deplete, replete and excess conditions, and the expression of chosen sub members were analyzed in the roots and shoots of the plants under these conditions (Fig. 6a, b). The expression patterns of representative results are graphically shown.

Fig. 6
figure 6

Semi-quantitative RT-PCR analysis of representative BjNRAMP and BjYSL clones. RNA was isolated from plants grown under control (C), iron deplete (−Fe), iron excess (Fe, 1 mM), manganese deplete (−Mn), manganese excess (Mn, 1 mM), zinc deplete (−Zn), and zinc excess (Zn, 0.2 mM) conditions. SQ RT-PCR analysis was done using primers listed in Supplementary Table 4. An actin transcript from B. juncea was used as internal control. Intensities of all bands from the representative data were densitometrically estimated by Image Quanta TL v2005, normalized against the intensity of the actin transcript, and the relative intensities were plotted by Graph pad prism version 4. a The expression of BjNRAMP3.5, BjNRAMP4.3, and BjNRAMP4.1. b The expression of BjYSL2.1, BjYSL5.8, and BjYSL6.1

Full size image

The transporter genes under investigation showed distinct metal- and tissue-specific expressions. BjNRAMP3.5 in root was expressed only under metal replete condition. The shoot expression of BjNRAMP3.5 however showed a gradual increase in response to increasing manganese and gradual decrease in response to increasing iron (Fig. 6a). Some of the BjNRAMP genes showed elevated expressions under metal deplete and excess conditions (biphasic expression) compared to control. BjNRAMP4.1 in shoot showed a constitutive expression under manganese treatment and biphasic expression with increasing iron. Interestingly, its behavior was quite different in roots, with a gradual increase in expression in response to iron stress, and a biphasic pattern seen in case of increased manganese (Fig. 6a). Biphasic pattern of expression was also observed for BjNRAMP4.3 in response to one or both metals.

Expressions of the YSL isoforms were studied under iron and zinc treatment. All the BjYSL clones (BjYSL2.1, BjYSL5.8, and BjYSL6.1) that were chosen in the SQ RT-PCR analysis showed relatively more variation in expression in root than shoot. BjYSL2.1 was constitutively expressed under all conditions tested. BjYSL5.8 showed maximum expression under zinc deficiency in the roots. On the other hand, BjYSL6.1 was expressed under iron deplete condition in shoots, and under control and zinc deplete conditions in roots (Fig. 6b). This different tissue-specific expression of each transporter in response to a specific metal can have important functional consequence.

Effect of lead and cadmium treatment

Brassica juncea is well known for its cadmium and lead tolerance property (Begonia et al. 1998; Heiss et al. 2003; Vassil et al. 1998; Zhu et al. 1999). B. juncea L. Czern 211000 used in this study also showed cadmium and lead tolerance in root–shoot assay (Fig. 7a, b) but were more tolerant towards lead than cadmium. This was apparent from the root length of the treated plants and the relatively higher accumulation of Pb in planta compared to cadmium (data not shown). EDTA that was added in the media to keep the lead in solution had somewhat deleterious effect of plant growth, especially on the growth of the lateral root. Despite all this, root length was maximum for the lead-treated plants (Fig. 7a). Since the BjYSL proteins exhibited significant changes, especially in the substrate-binding loop, the YSL genes were used for expression studies using SQ RT-PCR in plants treated with cadmium and lead. BjYSL6.1 specifically showed high levels of expression in the shoots upon cadmium treatment (Fig. 7c). Expression of BjYSL6.1, which showed root-specific expression under control condition, was downregulated in roots and upregulated in shoots in cadmium-treated plants. In contrast, expression of BjYSL5.8 was completely downregulated upon cadmium treatment. BjYSL5.8 showed highest expression in plant roots treated with lead-EDTA. Omitting EDTA precipitated the lead in the media, reduced accumulation of lead in the plant and also abolished the upregulation of BjYSL5.8 (data not shown) indicating a direct relation of the expression of this gene with lead.

Fig. 7
figure 7

Effect of lead and cadmium treatment. Plants were grown under control-EDTA/control or in presence of lead-EDTA (Pb 75 μM, EDTA 0.2 mM) and cadmium (Cd 10 μM) as described in “Materials and methods”. BjYSL6.1 and BjYSL5.8-specific primers were used for SQ-RT-PCR analysis. An actin transcript from B. juncea was used as internal control. a B. juncea plant grown under control-EDTA, Pb-EDTA (Pb 75 μM) and control, cadmium (Cd 10 μM) treatment. b Graph showing shoot, root lengths (Y axis on the left) and leaf area (Y axis on the right) of the control and metal-treated plants. c Expression of BjYSL6.1 and BjYSL5.8

Full size image

Discussion

Brassica juncea is an allopolyploid heavy metal tolerant and accumulator plant, which can tolerate, accumulate, and transport various essential and nonessential heavy metals. The relative ease of transforming B. juncea, coupled with its fast growing nature, makes it an ideal candidate for environmental biotechnology. The molecular mechanism of the broad-spectrum heavy metal tolerance and accumulation property of B. juncea is unknown. In this study, we attempted to clone the B. juncea orthologs of NRAMP and YSL gene family, given their importance in the sub cellular and intercellular metal transport and homeostasis in plants.

To achieve this, we coupled an ARDRA-based molecular fingerprinting technique with classical degenerate primer-based RT-PCR to clone and distinguish closely related transporter isoforms/alleles in a high throughput manner. Using this approach, we obtained partial clones of 23 NRAMP and 27 YSL genes from B. juncea. This method is relatively inexpensive and simple compared to family cloning techniques like Family Walker (Gilmartin 2002), which includes several steps of repeated PCR and adapter ligation. In addition, this technique was able to distinguish between sequences that had more than 90% identity at the nucleotide level and therefore could be useful in cloning genes from a family of interest in a high throughput manner from a relatively less known organism. Although it is difficult to claim with cent percent certainty that all possible isoforms were cloned in the study, we obtained many Brassica orthologs of the genes that were targeted while designing degenerate primers. Several clones with identical fingerprints were obtained indicating sufficient saturation in obtaining the members (Table 3). Random sequencing of independent clones having same pattern as BjYSL6.2, BjNRAMP2.4, and BjNRAMP3.3 showed identical sequences. There were also exceptions: BjNRAMP3.1, BjNRAMP3.2, BjNRAMP3.3, and BjNRAMP3.7 despite showing same pattern showed very similar but not identical sequences. Although it is not possible to claim all the changes in the isoforms to be completely free of sequence errors, yet many changes at nucleotide and protein levels are conserved across sub-isoforms as well in ESTs of Brassica (taxid 3705 for Blastn analysis) and indicated that the changes were Brassica specific and represented real isoforms/alleles. The frequent cutter enzymes were randomly chosen in this study to prove a concept. It is possible to bioinformatically predict frequent cutter enzymes most suitable for a gene family of interest so that members can be distinguished better. Given a set of sequences as input, these will produce products of discernible sizes as distinct fingerprints on an agarose gel. Considerations like this will ensure the detection of all ESTs with a better level of confidence at a given condition. Similar predictions are used to have suitable patterns in cDNA AFLP applications (Kivioja et al. 2005; Rombauts et al. 2003).

Certain members of BjNRAMPs or BjYSLs were found to have more sub members compared to others. For example, in case of BjYSLs, 55% of the total clones were variants of BjYSL5 (ortholog of AtYSL5). Such difference in EST abundance can have important functional consequences related to tolerance/accumulation property. Increased gene dosages of homeostatic genes that occur during interspecific crosses in evolution often have significant contribution towards tolerance. Such changes can lead to major perturbation in the transcriptome and alter tissue specificity and function of these genes (Leitch and Leitch 2008) leading to better endurance to harsh environment, for example heavy metal stressed conditions. In support of this view, role of increased gene dosage was recently indicated as one of the most important mechanisms leading to the evolution of metal hypertolerance and hyperaccumulating traits (Maestri et al. 2010). Estimation of copy number of the transporters based on the ESTs appeared to be difficult. A recent comparative mapping effort in B. juncea reported major rearrangements, duplications, and fusions during diversification of B. juncea genome relative to Arabidopsis thaliana (Panjabi et al. 2008). According to this study, the genome blocks where AtYSL6, AtYSL5, and AtYSL8 (similar to BjYSL6, BjYSL5, and BjYSL8) were located duplicated one, six, and four times, respectively, in B. juncea (Panjabi et al. 2008). Similarly, the blocks having AtNRAMP2 and AtNRAMP4 (similar to BjNRAMP2 and BjNRAMP4) were duplicated four and two times, respectively. Number of ESTs for BjYSL5 (15, AtYSL5 homolog) was the maximum that correlated with the maximum duplication of the block where its Arabidopsis homolog is located. However, a large region of B. juncea genome remained unmapped in this study that used syntenic relationship with A. thaliana. Copy numbers can also be underestimated as Arabidopsis itself has a reduced genome and is evolutionarily located in different clade altogether compared to B. juncea. BjNRAMP3 and BjYSL2 genes in fact showed more sequence similarity to N. caeurescens (T. caeurescens) isoforms that belongs to the same clade as B. juncea than Arabidopsis (Schranz et al. 2006, 2007). Analysis of the exons revealed that many of the members share one or more identical exons (Supplementary Figs. S6, S7) raising the possibility that these isoforms were alternatively spliced variants. This identified alterative splicing as a major mechanism using which EST abundance of these transporters was potentially expanded in B. juncea. For example, BjYSL6.2 shares first four exons of its cloned portion with BjYSL6.1 and the last exon with BjYSL6.4 (Supplementary Fig. S7b). Similar results were also seen for BjNRAMP isoforms (Supplementary Fig. S6). BjYSL6.1 and BjYSL6.4 coded for proteins with 99% similarity and differed only in the 3′UTR region. Role of these very similar proteins needs to be investigated in detail. Due to such high identity of these sequences, assembling the 3′ and 5′ ends during full length cloning of the genes to the partial sequences was very difficult. Incidentally, presence of more than one allelic form is common in many other metal hyper accumulators. gi|89520728 and gi|149688671 represent two allelic forms of TcNRAMP4 from Noccaea (Thlaspi) caerulescens. Kim et al. 2004 obtained several MTP clones with small changes from octaploid Thlaspi goesingense, which were allelic variants of the transporter. A similar expansion of MTP genes has also been reported for A. halleri (Willems et al. 2007). We did not detect B. juncea orthologs of pollen- and flower-specific isoforms under our experimental conditions. Orthologs of almost all A. thaliana NRAMP and YSL genes predicted to be expressed in the root and shoot tissue were obtained in the study. The only exception was AtYSL2, which is expressed both in shoots and roots of A. thaliana (DiDonato et al. 2004), but the B. juncea homolog was not cloned under the conditions tested.

Different isoforms of BjNRAMPs and BjYSLs showed distinct tissue-specific expressions in the plant. Depending on the tissue, their response towards different metals tested also varied considerably. Many of the BjNRAMP showed a biphasic pattern of expression in response to both manganese and iron showing more expression under deplete and excess condition. Such biphasic expression might be a protection mechanism against sudden influx of nonspecific heavy metals, which enters through opportunistic use of transporters. In yeast, such condition has been shown for the vacuolar transporter ZRC1 under zinc deficiency and is referred to as zinc shock (MacDiarmid et al. 2003; Zhao et al. 1998). TcNRAMP3 shows similar biphasic pattern in both zinc deficient and excess conditions (van de Mortel et al. 2006). Expression pattern of BjYSL isoforms showed more variation in expression. BjYSL6.1 but not BjYSL5.8 was found to be specifically upregulated in cadmium-treated shoots of the plants. BjYSL6.1 expressed in roots of normal plants, which upon cadmium treatment was downregulated in roots and elevated in shoots. BjYSL5.8 in contrast was upregulated in root of lead-treated plant. Whether these observations are of any importance related to cadmium and lead accumulation in plants remains to be tested. In a recent study, cadmium was found to be a poor substrate for ZmYS1 in maize, but role of other YSLs in the uptake and root-to-shoot transport of cadmium is not available. However, it is also possible that the upregulation of YSLs in response to lead and cadmium is a consequence of lead- and cadmium-induced iron deficiency (Meda et al. 2007).

Interestingly and somewhat expectedly, most of the prominent amino acid changes between BjNRAMPs and their orthologs in other species occurred in the loop regions between the transmembrane domains (Supplementary Table 3). These prominent changes in the intracellular loop regions could alter the capacity and specificity of substrate binding, while conserving the overall organization of the transporters, and lead to functionally separate metal binding behaviors between orthologs in different species. Consistent to the tonoplast localization of NRAMP proteins, such changes can have important consequences on vacuolar storage and release of metals. YSL orthologs showed the most significant changes in this respect. The extracellular loop predicted to be the substrate-binding site of YSL proteins showed many nonconservative Brassica-specific changes relative to Arabidopsis. How these changes can influence the substrate binding and influence metals transport ability of these proteins would be interesting to investigate in future. This work would be important in understanding the role of these transporters in homeostasis and root-to-shoot transport of metals, both essential and nonessential in B. juncea.