Characterization of SaeIF1 isolation from the eukaryotic translation factor SUI1 family in cadmium hyperaccumulator Sedum alfredii

Abstract

The hyperaccumulator of Sedum alfredii has the extraordinary ability to hyperaccumulate cadmium (Cd) in shoots. To investigate its underlying molecular mechanisms of Cd hyperaccumulation, a cDNA library was generated from leaf tissues of S. alfredii. SaeIF1, belonging to the eukaryotic protein translation factor SUI1 family, was identified by screening Cd-sensitive yeast transformants with this library. The full-length cDNA of SaeIF1 has 582 bp and encodes a predicted protein with 120 amino acids. Transient expression assays showed subcellular localization of SaeIF1 in the cytoplasm. SaeIF1 was constitutively and highly expressed in roots and shoots of hyperaccumulator of S. alfredii, while its transcript levels showed over 100-fold higher expression in the hyperaccumulator of S. alfredii relative to the tissues of a non-hyperaccumulating ecotype of S. alfredii. However, over-expression of SaeIF1 in yeast cells increased Cd accumulation, but conferred more Cd sensitivity. Transgenic Arabidopsis thaliana expressing SaeIF1 accumulated more Cd in roots and shoots without changes in the ratio of Cd content in shoots and roots, but were more sensitive to Cd stress than wild-type. Both special and general roles of SaeIF1 in Cd uptake, transportation and detoxification are discussed, and might be responsible for the hyperaccumulation characteristics of S. alfredii.

Main conclusion

Cadmium-sensitive yeast screening resulted in the isolation of protein translation factor SaeIF1 from hyperaccumulator Sedum alfredii which has both general and special regulatory roles in controlling cadmium accumulation.

Introduction

Cadmium (Cd) is a metabolically non-essential heavy metal with high solubility and persistence in nature because of the absence of degradation. High concentrations of Cd in soil can inhibit plant growth and interfere with the physiological and biochemical processes of photosynthesis and photorespiration, which inevitably decrease agricultural yields (Satarug et al. 2003; Lee et al. 2010). Many crop plants easily absorb and accumulate Cd, and threaten human health via the food chain. It has been reported that Cd uptake through the diet can lead to a plethora of disease states including cancer, arthritis, and kidney failure (Satarug et al. 2003; Godt et al. 2006). Historically, the itai-itai disease was caused by Cd poisoning from long-term consumption of Cd-contaminated water and crops in Japan (Inaba et al. 2005). Cadmium pollution has become a worldwide environmental problem and has attracted extensive attention from governments and international scientific communities (Nriagu and Pacyna 1988; Satarug et al. 2003; Nordberg 2004).

Phytoremediation employs green plants to absorb and sequester pollutants from the environment, and is considered to be an effective, economical and environmentally friendly method to tackle Cd pollution (Clemens et al. 2002). Hyperaccumulator plants are powerful tools for studying and developing phytoremediation because of their extraordinary abilities to hypertolerate and hyperaccumulate heavy metals in shoots (Krämer 2010; Yang et al. 2018). Arabidopsis halleri, Thlaspi caerulescens and Sedum alfredii are well-known natural Cd hyperaccumulators (Baker et al. 1994; Bert et al. 2002; Yang et al. 2004; Zhou and Qiu 2005; Reeves et al. 2018). Physiological and molecular investigations indicate that Cd hypertolerance and hyperaccumulation are associated with efficient Cd transportation and detoxification processes in hyperaccumulator species (Papoyan and Kochian 2004; Ueno et al. 2011; Zhang et al. 2016b). TcZNT1 encodes a putative Zn and Cd transporter and has been shown to mediate low-affinity Cd uptake in T. caerulescens (Pence et al. 2000). After Cd uptake by roots, the highly expressed transporter HMA4 in root xylem parenchyma plays an essential role in loading Cd into xylem for transportation to shoot tissues of A. halleri and T. caerulescens (Bernard et al. 2004; Hanikenne et al. 2008). Downregulation of AhHMA4 expression in A. halleri results in enhanced Cd sensitivity and reduced Cd accumulation in leaves (Hanikenne et al. 2008). HMA3 is a tonoplast-localized Cd transporter, and plays a significant role in Cd sequestration into vacuoles for detoxification in T. caerulescens, A. halleri and S. alfredii (Becher et al. 2004; Zhao et al. 2006; Ueno et al. 2011; Liu et al. 2017). Knowledge of Cd hyperaccumulation and hypertolerance mechanisms is fundamental for the development of transgenic plants that might be implemented in phytoremediation.

Unlike A. halleri and T. caerulescens, the hyperaccumulator of S. alfredii belongs to the Crassulaceae family (Yang et al. 2004). It can tolerate up to 200 μM Cd without obvious toxicity, and can accumulate approximately 11,000 mg Cd kg⁻¹ dry weight (DW) in aerial tissues (Yang et al. 2004; Zhou and Qiu 2005). Most Cd in the leaves of hyperaccumulator of S. alfredii is localized in the cell walls (Zhang et al. 2010; Peng et al. 2017). Our previous work showed that Cd hyperaccumulation and hypertolerance in the hyperaccumulator of S. alfredii was associated with active Cd efflux of the roots and shoots (Zhang et al. 2016b). The up-regulated Cd extruder HMA2 was selected as a key determinant in the hyperaccumulator of S. alfredii that results in efficient Cd translocation from roots to shoots and its aerial hyperaccumulation (Zhang et al. 2016b). In this study, we screened a cDNA library of the hyperaccumulator of S. alfredii for potential genes involved in the processes of Cd hyperaccumulation, and identified a new gene that has high homology with the translation initiation factor SUI1 (eIF1).

Materials and methods

Plant materials, growth conditions, and treatments

The Zn/Cd hyperaccumulator of S. alfredii was collected from an old Pb/Zn mining area in China, and was previously defined as the hyperaccumulating (HP) ecotype of S. alfredii (Yang et al. 2001; Long et al. 2004; Yang et al. 2004). Another kind of S. alfredii from a non-polluted tea garden in China showed much more sensitive to Pb/Zn/Cd stresses and less contents of Pb/Zn/Cd in tissues, and was previously defined as the non-hyperaccumulating (NHP) ecotype of S. alfredii (Yang et al. 2001; He et al. 2002; Ni et al. 2004). Similar-sized shoots of HP and NHP of S. alfredii were grown hydroponically to initiate new root development, as previously described (Zhang et al. 2016b). For gene expression analysis, the HP and NHP S. alfredii plants were treated with 100 μM CdCl₂ for 7 d. The control plants were cultured without Cd treatments. Each treatment had three replicates. After Cd treatments, the leaf, stem and root tissues were collected and quickly frozen in liquid nitrogen.

cDNA library screening

After treatment with 200 μM CdCl₂ for 15 d, the leaf tissues of HP S. alfredii were collected to construct cDNA libraries in pYES2 using the SMART cDNA library construction kit (CLONTECH, Palo Alto, CA, USA). The library was transformed into the BY4741-derived met15 opt1 yeast strain (MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 ΔOPT1) by the lithium acetate method (Zhang et al. 2016a). In one method, the tolerant yeast transformants were selected on minimal synthetic defined medium (SD) containing YNB (yeast nitrogen base with ammonium sulfate), 2% galactose and 1% raffinose, essential amino acids without uracil (URA) and 80 μM CdCl₂. In a second method, the yeast transformant cells first grew on SD minus uracil (SD-URA) plates without Cd and then were transferred onto new SD-URA plates containing 40 μM CdCl₂. The sensitive transformant colonies were identified by comparing the different appearance of colonies on the plates with and without Cd. Both the tolerant and sensitive yeast transformants were confirmed by further selection, and were used to extract plasmids for sequencing inserts.

Among the candidate genes after library screening, one insert was predicted to encode a 120-amino acid polypeptide that shares homology with the translation initiation factor SUI1 family protein, and was renamed here as SaeIF1. The full length of SaeIF1 was amplified through 3′ and 5′ RACE using the Smart RACE cDNA amplification kit (Clontech, CA, USA) with specific primers (Suppl. Table S1). The amino acid sequence of the SaeIF1 gene was aligned using DNAMAN software version 8.0.

Expression analysis of SaeIF1 in S. alfredii

Total RNA was extracted from HP and NHP S. alfredii tissues using the Spectrum^™ Plant Total RNA Kit (Sigma-Aldrich, US), treated with TURBO DNA-free^™ DNase (Thermo-Fisher Scientific, Vilnius, Lithuania), and reverse-transcribed using the First-Strand cDNA Synthesis Kit (GE Healthcare, UK). The Ultra SYBR Mixture with ROXII (CWBIO) was used for quantitative real-time RT-PCR on an ABI PRISM 7900HT fluorescent quantitative PCR machine. Actin2 and SaeIF1 were amplified using specific primers (Suppl. Table S1). PCR amplification conditions were 95 °C for 10 min, then 40 cycles of 95 °C for 30 s, 58 °C for 30 s, and 72 °C for 20 s. Levels of transcript were normalized against the house-keeping gene Actin2 (Gao et al. 2013).

The full-length of SaeIF1 was amplified and subcloned into pET-28a-SUMO using specific primers (Suppl. Table S1). The construct was transformed into E. coli Rosetta (DE3) to induce the expression of SaeIF1 protein, which were extracted and purified by His-bind resin (Novagen) (Chen et al. 2020). The purified protein was used to generate polyclonal rabbit antibodies. Tissues of HP and NHP S. alfredii were ground in mortars with liquid nitrogen to extract total proteins using buffer containing 175 mM Tris-HCl pH 7.5, 5% SDS, 5 mM EDTA, 5% glycerol, 10 mM DTT, 1 mM PMSF, 0.1% (v/v) Triton X-100. Proteins were separated by 15% SDS-PAGE and transferred to nitrocellulose filters (Millipore). The SaeIF1 protein was detected with antibodies (1:500 dilution), and visualized with a goat anti-rabbit alkaline phosphatase antibody (Invitrogen) using nitroblue tetrazolium and 5-bromo-4-chloro-3-indolylphosphate as substrates.

Subcellular localization of SaeIF1

The full-length cDNA of SaeIF1 without the stop codon was amplified by Phusion Hot Start II DNA Polymerase (Thermo Scientific, USA) using specific primers (Suppl. Table S1). The PCR product was subcloned into the pFF19-GFP vector under control of the cauliflower mosaic virus (CaMV) 35S promoter. Tender leaves of Nicotiana tabacum were cut into small squares and digested at 37 °C for 90 min in 10 ml of enzyme solution containing 1.0% (w/v) cellulase R-10 (Yakult, Japan), 0.25% (w/v) macerozyme R-10 (Yakult, Japan), 0.6 M mannitol, 25 mM MESKOH, pH 5.6. The released protoplasts were transfected with 50 μg salmon sperm carrier DNA and 30 μg plasmid pFF19-SaeIF1-GFP or pFF19-GFP by the polyethylene glycol method (Negrutiu et al. 1987; Damm et al. 1989; Abel and Theologis 1994). The transfected protoplasts were incubated for 24 h at 22 °C, and the fluorescence signals were imaged by confocal microscopy.

Heterologous expression of SaeIF1 in Saccharomyces cerevisiae

The full-length cDNA of SaeIF1 was amplified using specific primers (Suppl. Table S1) and cloned into the pYES2 vector (Invitrogen, USA). The met15 opt1 yeast strain was transformed with pYES2-SaeIF1 or pYES2 by the lithium acetate method (Zhang et al. 2016a). Single colonies were picked and cultured in liquid SD-URA medium overnight on a shaker at 220 rpm at 28 °C until OD₆₀₀ reached 1.0. Portions of the cultures were serially diluted with sterile H₂O to OD₆₀₀ 0.1, 0.01, 0.001, then 10 μL of each solution was dropped on SD-URA plates with Cd concentrations ranging from 0 to 120 μM, followed by culture at 28 °C for 4 days. To further evaluate the growth curves of yeasts in liquid medium, single colonies were picked and cultured in liquid SD-URA medium with 40 μM or 80 μM CdCl₂. The OD₆₀₀ of culture was determined at 12-h intervals. For Cd-uptake assays, yeast cells were cultured to logarithmic phase and then treated with 40 μM CdCl₂ for 24 h. Yeast cells were collected and washed three times with 20 mM Na₂-EDTA and deionized water, oven-dried at 65 °C, and digested in HNO₃. Clear samples were subsequently diluted with ultrapure water and Cd concentrations were determined using an Inductively Coupled Plasma Optical Emission Spectrometer (ICP-OES, Perkin Elmer Optima 3000 DV, USA).

Overexpression of SaeF1 in Arabidopsis and cadmium stress treatment

The full-length cDNA of SaeIF1 was amplified with specific primers (Suppl. Table S1), and subcloned into the modified binary vector pBI121 under the CaMV35S promoter. The Columbia-0 (Col-0) ecotype of Arabidopsis thaliana was transformed with pBI121-SaeIF1 by the floral dip method (Clough and Bent 1998). Transgenic homozygous lines were selected with kanamycin (25 μg mL⁻¹) on 1/2-strength Murashige and Skoog (MS) plates. RT-PCR was carried out to investigate the expression of SaeIF1 in transgenic A. thaliana. After germination for 2 d, wild-type and three T3 homozygous lines (OV-1, OV-2, and OV-3) were transferred to 1/2 MS plates containing 60 μM CdCl₂ then grown in a growth chamber for two weeks at 22 °C with 16 h light/8 h dark and light intensity of 400 μmol photons m⁻² s⁻¹. The seedlings were photographed, and the primary root length of each seedling was measured using Image J.

Sterilized seeds of Col-0 and SaeIF1 overexpression lines were germinated and grown on 1/2 MS media. After one week, the seedlings were hydroponically cultured in 1/2 Hoagland solution (Hoagland and Arnon 1950) for three weeks, and then received the next 7 days’ treatments with 40 μM CdCl₂. The plants were washed three times with 20 mM Na₂-EDTA and deionized water, and were separated into root and shoot tissues. All the tissues were oven-dried at 65 °C and digested in HNO₃. Clear samples were subsequently diluted with ultrapure water and Cd concentrations were determined by ICP-OES as above.

Statistical analysis

The assays of gene expression, root elongation and Cd accumulation were conducted in three independent experiments with at least three replications. Student’s t-test (P < 0.05) analysis was used to detect differences.

Results

Isolation of SaeIF1 from the eukaryotic translation factor SUI1 family in S. alfredii

More than 30,000 yeast colonies transformed with the HP S. alfredii cDNA library grew on SD-URA plates. Among them, 20 sensitive colonies were unable to grow when transferred onto SD-URA plates with 40 μM Cd. In addition, 72 tolerant colonies were identified when the yeast transformants with the HP S. alfredii cDNA library were selected on SD-URA plates with 80 μM Cd. The candidate genes for Cd tolerance in yeast cells encoded phytochelatin synthase 1, N-methyltransferase, ribosomal proteins, ferredoxin and some unknown proteins (Table 1). The genes conferring Cd sensitivity in yeast cells were predicted to encode hydrolases, elongation factors and translation-related proteins (Table 1). One gene among the Cd-sensitivity genes was annotated as the translation initiation factor SUI1 family protein, and was renamed as S. alfredii eukaryotic initiation factor 1 (SaeIF1), which will participate in the regulation of protein synthesis. To our knowledge, no studies were reported about the mechanisms of protein synthesis regulation in cadmium accumulation in hyperaccumulators, although many studies have showed high levels of transcripts in hyperaccumulators (Hammond et al. 2006; Weber et al. 2006; Zhang et al. 2016b; Yang et al. 2017). Thus, SaeIF1 was chosen for further study to explore new avenues of the molecular mechanisms of Cd hyperaccumulation in plants.

Table 1.

Candidate genes from screening cDNA library of Sedum alfredii.

Accession number	Gene Description
The genes conferring Cd tolerance in yeast cells
MK044851	Phytochelatin synthetase 1
MT594410	Stress enhanced protein
MT594416	14-3-3 protein
MT594404	Cupredoxin superfamily protein
MT594405	Acyl-coa N-acyltransferases (NAT) superfamily protein
MT594406	Slow green1 protein
MT594412	Ferredoxin
MT594427	NADH-ubiquinone oxidoreductase
MT594413	Transcription factor PCL1
MT594408	SWI/SNF-related matrix-associated actin-dependent regulator
MT594417	Ubiquitin-conjugating enzyme 10
MT594423	Probable calcium-binding protein CML41
MT594424	Signal peptidase
MT594415	Annexin D2-like
MT594407	Ribose-5-phosphate isomerase 3
MT594411	60S ribosomal protein
MT594426	Ribosomal protein
MT594419	Uncharacterized protein
The genes conferring Cd sensitivity in yeast cells
MN944398	Protein translation initiation factor 1
MT594421	Elongation factor P
MT594422	Alpha/beta-hydrolases superfamily protein

The full-length cDNA of SaeIF1 contains 582 bp, and encodes a predicted protein with 120 amino acids. SaeIF1 contains the conserved motifs found in eIF1s from other eukaryotes, including fourteen predicted rRNA binding sites in the SUI1-like superfamily domains (Fig. 1). There were 98% sequence identical of coding regions between SaeIF1 from HP S. alfredii and its homologue gene from NHP S. alfredii (Suppl. Fig. S2). Multiple sequence alignments indicated that SaeIF1 has a relatively high similarity to the eIF1s from other plant species. It is estimated that SaeIF1 shows 66% to 68% sequence similarities to eIF1 homologs from A. thaliana, A. halleri, Brassica napus, Zea mays, and Nicotiana tabacum (Fig. 1).

Fig. 1.

Sequence alignment of eIF1 proteins from different plant species. The alignment was performed using DNAMAN software version 8.0 among the deduced amino acid sequences of SaeIF1 from hyperaccumulator Sedum alfredii (accession no. MN_944398), BneIF1 (XP_013751102) from Brassica napus, CmeIF1 (XP_016903023) from Cucumis melo, ZmeIF1 (NP_001150984) from Zea mays, BdeIF1 (XP_003564393) from Brachypodium distachyon, IneIF1 (XP_019191100) from Ipomoea nil, LjeIF1 (Lj3g3v1010910) from Lotus japonicus, CreIF1 (XP_006304318) from Capsella rubella, OseIF1 (XP_015639366) from Oryza sativa, AteIF1 (NP_001332503) from Arabidopsis thaliana and AheIF1 (gene ID: g09362, sequence from Ensembl Plants) from Arabidopsis helleri. Identical residues are in shown black and similar residues are shaded. The eIF1 SUI1-like superfamily conserved domain for SaeIF1 is underlined. Potential rRNA binding sites are boxed.

To investigate the subcellular location of SaeIF1, the C-terminus of SaeIF1 was fused to green fluorescent protein (GFP) and used for transient expression in tobacco protoplasts. Fluorescence of GFP alone (control) was observed throughout the cytoplasm and the nucleus (Fig. 2a–d). In contrast, transient expression of SaeIF1-GFP fusion protein was observed in tobacco cytoplasm only, suggesting that SaeIF1 was located in the cytoplasm (Fig. 2e–h).

Fig. 2.

Subcellular localization of SaeIF1. (a–d) Tobacco mesophyll protoplasts expressing GFP. (e–h) Tobacco mesophyll protoplasts expressing fusion protein SaeIF1–GFP. (a and e) GFP fluorescence; (b and f) autofluorescence of chlorophyll; (c and g) merge of GFP fluorescence and chlorophyll fluorescence; (d and h) bright field image. Scale bar = 5 μm.

Expression profiles of SaeIF1 in S. alfredii

SaeIF1 was constitutively expressed in roots, stems, and leaves of the HP S. alfredii (Fig. 3a–c). The expression levels of the SaeIF1 gene was not significantly affected in any of the tissues after Cd treatments (Fig. 3a–c). Although the homolog of SaeIF1 was also constitutively expressed in all the tissues of NHP S. alfredii, the SaeIF1 was expressed over 100-fold more highly in HP than in NHP S. alfredii, regardless of Cd treatments (Fig. 3a–c). Using western blot analysis, the SaeIF1 protein was readily detected in leaf, stem and root tissues of HP S. alfredii (Fig. 3d–f). The SaeIF1 protein showed no obvious changes in abundance after Cd treatment in any of the tissues of the HP S. alfredii, relative to tissues without Cd treatments (Fig. 3d–f). However, the SaeIF1 homolog was hardly detected in tissues of NHP S. alfredii, regardless of Cd treatment (Fig. 3d–f). Thus, detection at both the RNA and protein levels indicated expression of SaeIF1 was significantly up-regulated in HP S. alfredii compared to NHP S. alfredii.

Fig. 3.

Effect of Cd treatment on expression of SaeIF1 in different tissues of hyperaccumulating population (HP) and non-hyperaccumulating population (NHP) of Sedum alfredii. (a–c) Transcript detection of SaeIF1 expression. Data are means ± standard deviation (n = 3). Different letters on the error bars denote significant differences (P ≤ 0.05). (d–f) Western blot analysis of SaeIF1 protein content: “−” denotes samples without Cd treatment, “+” denotes samples with Cd treatment. Cd-treated plants received 100 μM Cd for 7 d.

SaeIF1 increased Cd accumulation but conferred Cd sensitivity in yeast cells

Both SaeIF1- and empty vector pYES2-transformed yeast cells grew well, and showed no significant difference on plates without Cd treatment (Fig. 4a). The growth curves suggested that the SaeIF1-transformed yeast cells grew slightly faster than pYES2-transformed yeast cells in liquid medium without Cd (Fig. 4b). Although Cd addition inhibited yeast growth, the yeast cells with SaeIF1 expression were more sensitive to Cd stress (Fig. 4c). The SaeIF1-transformed yeast cells did not grow in the presence of 80 μM Cd, while the pYES2-transformed yeast displayed a prolonged growth into the log phase after 80 hours. Consistently, the transgenic yeast cells with SaeIF1 grew much slower than the yeast cells transformed with the pYES2 vector under the lower Cd treatment (Fig. 4c). Unexpectedly, Cd accumulation in the yeast cells with SaeIF1 expression was about 4-fold higher after 24 h in 40 μM Cd than that in the pYES2-transformed yeast cells (Fig. 4d).

Fig. 4.

Cd tolerance and accumulation assays in yeast cells with expression of SaeIF1. (a) Yeast transformant cells harboring the empty vector pYES2 (EV) or expressing SaeIF1 were grown on plates with 0–100 μM Cd concentrations for 4 d. (b) Growth curves of EV- and SaeIF1-transformed yeast in liquid media without Cd treatment. (c) Growth curves of EV- and SaeIF1-transformed yeast in liquid media with 40 μM or 80 μM Cd. (d) Cd contents in yeast transformants with EV or SaeIF1 in liquid media with 40 μM Cd for 24 h. Different letters on the error bars denote significant differences (P ≤ 0.05). Data represent means ± standard deviation (n = 3).

Transgenic A. thaliana expressing SaeIF1 showed increased Cd accumulation but decreased Cd tolerance

Three independent SaeIF1-transgenic homozygous lines were available to test the function of SaeIF1 in Arabidopsis plants. After germination on plates, the root growth of wild-type and SaeIF1-overexpressing lines were similar in the absence of Cd treatment (Fig. 5a and 5c). But the growth of both the wild-type and SaeIF1 transgenic A. thaliana was significantly inhibited after addition of 60 μM Cd to the medium (Fig. 5b). Furthermore, the root lengths of SaeIF1-overexpressing lines were shorter than the wild-type under Cd treatment (Fig. 5b and 5d). Cd contents were measured in different tissues of plants exposed to 40 μM Cd in hydroponic culture for 7 d. The wild-type A. thaliana accumulated 0.9 mg Cd g⁻¹ DW and 0.4 mg Cd g⁻¹ DW in roots and shoots, respectively (Fig. 5e), while a maximum of around 1.2 mg Cd g⁻¹ DW in roots and 0.6 mg Cd g⁻¹ DW in shoots was measured in the SaeIF1 transgenic A. thaliana plants. These results imply that the overexpression of SaeIF1 decreased Cd tolerance, but increased root accumulation of Cd by 40%, and shoot Cd accumulation by 27% in the transgenic A. thaliana plants (Fig. 5e). However, the ratios of the content of Cd in shoots and roots were almost same in both WT and transgenic A. thaliana plants, at around 0.5 (Fig. 5f).

Fig. 5.

Cd tolerance and accumulation assays in SaeIF1 transgenic Arabidopsis. Wild-type (WT) and SaeIF1-overexpressing transgenic seeds (OV-1 to OV-3) were germinated and grown for 7 d in 1/2 MS medium without Cd (a) or with 60 μM Cd (b). Root length of WT and transgenic seedlings without Cd (c) or with 60 μM Cd (d) was measured after 7 d. (e) Cd contents were measured in roots and shoots of WT and transgenic plants treated with 40 μM Cd for 7 d. (f) Ratio of shoot-to-root content of Cd. Data are means ± standard deviation (n = 3). Different letters on the error bars denote significant differences (P ≤ 0.05).

Discussion

Screening a cDNA library provides a powerful and efficient strategy to obtain functional genes in plant studies. Several key genes (e.g. PCS1 and HMA4) participating in heavy metal tolerance have been identified in the last two decades by selecting tolerant yeast colonies via library screening. TaPCS1 was obtained by screening a wheat root cDNA library, and encodes phytochelatin synthase 1 (PCS1), which catalyzes the production of phytochelatins as the major ligands for cellular Cd complexation (Clemens et al. 1999). TcHMA4 was identified by screening a cDNA library from T. caerulescens, and has been found to encode a plasma membrane Cd extruder, which was later proved to be a key determinant for Cd hypertolerance and hyperaccumulation, by effectively loading Cd into xylem tissues and transporting Cd from roots to shoots (Papoyan and Kochian 2004). In the present study, we intended to identify new genes by screening tolerant yeast colonies for Cd hypertolerance, and screening sensitive yeast colonies for Cd hyperaccumulation, using a cDNA library from the hyperaccumulator S. alfredii. Using this approach, PCS1 and other 17 genes were screened as putative Cd tolerance genes and three genes were selected as Cd sensitivity genes (Table 1). Among these 21 genes, just PCS1 has previously been studied extensively in relation to phytochelatin synthesis to detoxify Cd (Clemens et al. 1999; Cobbett, 2000). Other candidate genes with known functions encoded ferredoxin, stress enhanced protein, and proteins participating in protein synthesis and signal transduction (Table 1), but their underlying mechanisms in response to Cd stress in plants were not clear. Rather, our screening of candidate genes would explore new avenues to extend our understanding of the mechanisms of Cd hyperaccumulation in plants.

Amongst the aforementioned genes, we chose to study in detail the role of SaeIF1. Overexpression of eIF1 from rice and Leymus chinensis was reported to enhance salt tolerance in transgenic plants (Diédhiou et al. 2008; Sun and Hong 2013). This suggested a protective role of eIF1 in salt stress acclimation in plants. The eIF1 protein was reported to be involved in start codon recognition during translation initiation for protein synthesis, ensuring proper amounts of protein were obtained. In outline, eIF1 bound the 40S ribosomal subunit near the P-site, and stimulated the assembly of eIF2-GTP-Met-tRNA_i complex on the 40S subunit together with eIF1A, eIF3 and eIF5, resulting in the formation of a pre-initiation complex (PIC) (Jackson et al., 2010; Hinnebusch 2014; Browning and Bailey-Serres, 2015). eIF1 worked in cooperation with eIF1A, and stabilized the conducive open conformation of 40S subunits to facilitate PIC scanning. A hypothesis was further proposed that eIF1 prevented the full P-site engagement of Met-tRNA_i in 40S subunits at non-AUG codons until PIC scanning achieved attachment to the initiation codon. This state prevented the closed conformational rearrangement of the 40S subunit and enabled it to completely accommodate Met-tRNA_i in the P-site and release eIF1, finally producing an 80S ribosome for protein synthesis (Jackson et al., 2010; Hinnebusch 2014; Browning and Bailey-Serres, 2015). In the present study, heteroexpression of SaeIF1 from HP S. alfredii decreased Cd tolerance but increased Cd accumulation in transgenic yeast and Arabidopsis (Fig. 4 and Fig. 5). These results not only expanded our knowledge of the function of eIF1 in plants, but also indicated the potential significance of translation initiation (involving eIF1) in the regulation of Cd accumulation in plants.

Cd hyperaccumulation in plants is achieved by the synergistic processes of uptake, translocation, and subsequent detoxification (Verbruggen et al. 2009). Heterologous expression of SaeIF1 significantly increased Cd contents but decreased Cd tolerance in the unicellular organism of transgenic yeast cells (Fig. 4), suggesting its potential role in enhancing protein synthesis for Cd uptake. This hypothesis was further supported by the accumulation of higher Cd content in transgenic Arabidopsis plants (Fig. 5). In fact, SaeIF1 was expressed much higher at both transcriptional and translational levels in roots of HP S. alfredi than NHP S. alfredii (Fig. 3a, 3d). Consistent with that finding, the rate of Cd uptake was much higher in the former roots than in the latter roots (Lu et al., 2009).

Hyperaccumulators have unique mechanisms to tolerate heavy metals. Previous work showed Cd efflux in the cells of HP S. alfredii, as well as in cells of another hyperaccumulator T. caerulescens, but it was absent in the cells of non-hyperaccumulators (Bernard et al. 2004; Zhang et al. 2016b). Cd efflux not only played important roles in efficient Cd transportation, but also decreased Cd accumulation within the cells and made the cell wall the major site of Cd storage in HP S. alfredii (Zhang et al. 2010; Zhang et al. 2016b; Peng et al. 2017). This special characteristic would confer Cd tolerance in HP S. alfredii, but would be lacking as a mechanism to detoxify Cd in transgenic yeast or Arabidopsis. Besides, the genes of ZIP4, HMA2 and HMA3, encoding key transporters for Cd uptake, translocation and vacuolar sequestration, were transcribed with much stronger levels in HP S. alfredii than in NHP S. alfredii (Zhang et al. 2016b; Peng et al. 2017; Yang et al. 2017). Expression studies including RNA-seq and proteomics experiments also showed that a plethora of other functional genes were transcribed and translated at much higher levels in HP relative to NHP Sedum plants (Suppl. Tables S2 and S3; Zhang et al. 2016b; Peng et al. 2017; Yang et al. 2017; Zhang et al. 2017). This was also the case in the other Cd hyperaccumulators (Weber et al. 2004; Hammond et al. 2006; Weber et al. 2006). Transcriptomic analysis indicated expression of one eIF1 homologue gene in HP S. alfredii (Suppl. Fig. S1; Yang et al. 2017). In combination with unique characteristics including high levels of transcripts and Cd efflux, it could be expected that strong expression of SaeIF1 generally facilitated translation control of these plenty of transcripts into proper amounts of proteins, which efficiently functioned in Cd uptake, transportation and sequestration for its hypertolerance and hyperaccumulation in HP S. alfredii.

Unlike the up-regulated roles of SaeIF1 in Cd uptake, expression of SaeIF1 alone was probably not enough to enhance translation of transporters for Cd detoxification and translocation from low abundance transcripts, as Cd tolerance decreased in SaeIF1-transgenic organisms and root-to-shoot Cd translocation were almost same between WT and transgenic Arabidopsis (Fig. 4 and Fig. 5). The eIF1-integrated PIC might be modified to induce selective recruitment of ribosomes for specific mRNA translation under stress conditions (Sonenberg and Hinnebusch, 2009; Spriggs et al., 2010). Some other regulatory factors might be expected to direct eIF1-integrated PIC to enhance protein translation specifically for Cd uptake, while not acting on synthesis of proteins in Cd translocation and detoxification. Thus, the detailed special mechanisms of SaeIF1 function would be exciting to study in the future, as it might advance our knowledge of the post-transcriptional regulation of Cd hyperaccumulation in plants.

In summary, screening for Cd-sensitive yeast resulted in the isolation of SaeIF1 from the hyperaccumulator of S. alfredii. SaeIF1 belongs to the eukaryotic translation initiation factor SUI1 family and was highly expressed in all the tissues of hyperaccumulator of S. alfredii, regardless of Cd treatments. Heteroexpression of SaeIF1 enhanced Cd accumulation but conferred higher Cd sensitivity both in transgenic yeast cells and Arabidopsis. SaeIF1 likely played special roles in enhancing protein synthesis for Cd uptake, and also worked in general roles in enhancing translation of high levels of transcripts for Cd efficient transportation and detoxification in hyperaccumulator of S. alfredii. Present work shed light on the potential significance of translation control with eIF1 participation for Cd hyperaccumulation in hyperaccumulator. Further study of this underlying mechanism will help elucidate the link between increased gene transcription and Cd hyperaccumulation.

Supplementary Material

Supplement file and supplemental figures

Suppl. Fig. S1. Sequence alignment of SaeIF1s. SaeIF1 was obtained in the present study. The sequence of SaeIF1Y was obtained from RNA-seq data of hyperaccumulator Sedum alfredii (Yang et al. 2017, Front Plant Sci 8: 425). The alignment was performed using DNAMAN software version 8.0, and showed identical sequences of the open reading frames.

Suppl. Fig. S2. Sequence alignment of eIF1 from HP and NHP Sedum alfredii. The nucleotide and amino acid sequences of SaeIF1h and SaeIF1n were isolated from the HP and NHP of S. alfredii, respectively. (a) Sequence alignment of eIF1 nucleotide sequence from HP and NHP S. alfredii. (b) Sequence alignment of eIF1 proteins from HP and NHP S. alfredii. The alignment was performed using DNAMAN software version 8.0. Identical residues are in black and similar residues are shaded.

Suppl. Table S1. Primers used in this study.

Suppl. Table S2. RNA-seq analysis of expression of representative genes in hyperaccumulating populations (HP) compared to non-hyperaccumulating populations (NHP) of Sedum plants without (control) or with Cd treatment (Cd).

Suppl. Table S3. Relative contents of representative proteins in the leaves of HP S. alfredii in comparison to NHP S. alfredii.

Abbreviations

Cd

cadmium

HP

hyperaccumulating population

NHP

non-hyperaccumulating population

GFP

green fluorescent protein

OV

over-expression transgenic plants