Anchoring plant metallothioneins to the inner face of the plasma membrane of Saccharomyces cerevisiae cells leads to heavy metal accumulation

In this study we engineered yeast cells armed for heavy metal accumulation by targeting plant metallothioneins to the inner face of the yeast plasma membrane. Metallothioneins (MTs) are cysteine-rich proteins involved in the buffering of excess metal ions, especially Cu(I), Zn(II) or Cd(II). The cDNAs of seven Arabidopsis thaliana MTs (AtMT1a, AtMT1c, AtMT2a, AtMT2b, AtMT3, AtMT4a and AtMT4b) and four Noccaea caerulescens MTs (NcMT1, NcMT2a, NcMT2b and NcMT3) were each translationally fused to the C-terminus of a myristoylation green fluorescent protein variant (myrGFP) and expressed in Saccharomyces cerevisiae cells. The myrGFP cassette introduced a yeast myristoylation sequence which allowed directional targeting to the cytosolic face of the plasma membrane along with direct monitoring of the intracellular localization of the recombinant protein by fluorescence microscopy. The yeast strains expressing plant MTs were investigated against an array of heavy metals in order to identify strains which exhibit the (hyper)accumulation phenotype without developing toxicity symptoms. Among the transgenic strains which could accumulate Cu(II), Zn(II) or Cd(II), but also non-canonical metal ions, such as Co(II), Mn(II) or Ni(II), myrGFP-NcMT3 qualified as the best candidate for bioremediation applications, thanks to the robust growth accompanied by significant accumulative capacity.


Introduction
Heavy metals such as Cu(II), Mn(II), Ni(II), Co(II) and Zn(II) are of unique importance for life, being included in the shortlist of trace elements which are essential for the activity of various biomolecules, particularly enzymes. These elements are taken up by organisms from their environment through a variety of high or low affinity transport systems, which constantly nigrum MTs were shown to develop both Cd(II) tolerance and increased Cd(II) adsorption, exhibiting higher affinity for Cd(II) than for Cu(II) or Hg(II), along with a remarkable capacity to concentrate ultra-traces of Cd(II) at the cell surface [32].
In the present study, we addressed the possibility to obtain heavy metal hyperaccumulating S. cerevisiae by engineering cells towards producing plant MTs targeted to the inner face of the yeast plasma membrane. We hypothesized that the engineered yeast cells would accumulate heavy metals thanks to cation sequestration by the MTs attached to the cytosolic face of the membrane. The accumulative capacity of the engineered yeasts was tested under two conditions: (1) physiological, when "traces" of Co(II), Cu(II), Mn(II), Ni(II), Zn(II) and the nonessential Cd(II) were simultaneously present in the incubation medium, or (2) "tolerable excess", when growth media were supplemented with individual metal ions introduced at the highest concentration that did not significantly affect cell viability. Under both conditions we identified strains which could accumulate Cu(II), Zn(II) or Cd(II), but also the MT-noncannonical Co(II), Mn(II) or Ni(II).

Cloning plant MT cDNAs
Total RNA was extracted from the A. thaliana accession Col-0 and the N. caerulescens accession La Calamine with the Spectrum Plant Total RNA kit (Sigma-Aldrich, Saint Louis, USA) as described by the supplier. An on-column DNase digestion was performed using the RNase-Free DNase Set (Qiagen, Hilden, Germany) to eliminate genomic DNA. Total RNA was quantified using a NanoDrop ND-1000 (Nanodrop, Delaware, USA) and 1 μg total RNA was used to synthesize cDNA with the QuantiTect Reverse Transcription Kit (Qiagen, Hilden, Germany). The full-length coding sequence (minus the codon for the initial methionine) of seven A. thaliana and four N. caerulescens MTs was amplified from the respective cDNA with TaKaRa Ex Taq (Takara Bio Inc, Otsu, Japan) polymerase using the primers listed in S1 Table. The amplified products were purified from agarose gel using the Wizard SV Gel and PCR Clean-Up System (Promega Corporation, Madison, USA), cloned into the pCRII-TOPO vector by TOPO TA cloning (Life Technologies, Carlsbad, USA) and verified by sequencing (S1 Table). All cloned MT sequences were identical with those reported in NCBI, with the exception of NcMT1, which contained a 24bp in-frame insert generating an additional NCGCGSSC close to the N-terminus and a single nucleotide polymorphism leading to a W to C change. These sequence features also exist in AtMT1a and AtMT1c. The sequence cloned for NcMT3 contained two nucleotide polymorphisms leading to a Q to S change at position 18 when comparing to the NcMT3 sequence published in NCBI. For each of the AtMT4s different protein sequences are reported in NCBI; only one of each was cloned in the present study. The amino acid sequence alignment of the cloned MTs is presented in S1 Fig. The MT-encoding inserts were excised from pCRII-TOPO using the restriction sites introduced by the primers during PCR amplification (S1 Table) and inserted into the same restriction sites of the yeast plasmid pGREG596 (GAL1::myrGFP, URA3, HIS3) [33], purchased from EUROSCARF (European S. cerevisiae Archive for Functional Analysis, www.euroscarf.de). This directional cloning allowed the removal of the HIS3 gene from pGREG596 and introduced the plant MT sequences in frame downstream of the myrGFP gene (S2 Table). The resulting yeast plasmids harboring plant MT ORFs (minus initial methionine) are presented in Table 1.
Yeast cells were transformed by the "quick and dirty" method [34] and the transformants were selected on SD-Ura agar plates. Correct transformation was double-checked by colony PCR [33].

Yeast strains and culture conditions
The S. cerevisiae strain used in this study was BY4741 (MATa; his3Δ1; leu2Δ0; met15Δ0; ura3Δ0), obtained from EUROSCARF. Cell storage, growth media and cell manipulation were done as described [35]. Transformed strains were maintained and grown in standard SD-Ura (synthetic media with dextrose lacking uracil, for plasmid selection) or SGal-Ura (synthetic media with galactose, for transgene induction). Minimal defined media (MMe) were virtually metal free and were supplemented with the essential elements and with controlled metal concentrations prior to use. For solid media, 2% agar was used. For growth improvement, all the synthetic media were supplemented with 40 mg/L leucine [36].

Metal accumulation by growing cells
Overnight pre-cultures were prepared as described above. At time 0, cells were shifted to SGal-Ura and incubated 4 h for transgene induction (30˚C, 200 rpm). From this point, cells were used in two separate ways: (1) To determine metal accumulation under low-metal conditions, cells were washed and shifted to MMeGal-Ura (containing 2 μM MnCl 2 and ZnCl 2 ; 1 μM CuCl 2 , CoCl 2 , NiCl 2 and CdCl 2 ) and incubated for 10 h (30˚C, 200 rpm) before being harvested for metal assay. (2) To determine metal accumulation under high concentration conditions, metal ions were added from sterile stocks to the indicated concentrations and incubated for 2 h (30˚C, 200 rpm). To measure the metal accumulated, cells were harvested by centrifugation (1 min, 5000 rpm, 4˚C) and washed three times with ice-cold 10 mM 2-(N-morpholino)ethanesulfonic acid (MES)-Tris buffer, pH 6.0. Cells were finally suspended in deionized water (10 8 cells/mL). For metal assay, cells were digested with 65% ultrapure HNO 3 (Merck, Darmstadt, Germany). Metal analysis was done using an instrument with a single collector, quadrupole inductively coupled plasma with mass spectrometry (ICP-MS, Perkin-Elmer ELAN DRC-e, Concord, Canada) against Multielement ICP Calibration Standard 3, matrix 5% HNO 3 (Perkin Elmer Pure Plus). The metal cellular content was normalized to total cellular proteins, assayed spectrophotometrically as described [37]. Values were expressed as the mean ± standard deviation of triplicate determinations (technical triplicates) of three independent yeast transformants (biological replicates).

myrGFP-MT localization
For the detection of myrGFP-MT fluorescence, the transformed cells were grown overnight in SD-Ura and manipulated as described above. At time 0, cells were shifted to SGal-Ura for transgene induction and grown for 2-4 hours before being visualized by fluorescence microscopy. Live cells were examined with an Olympus fluorescent microscope system (Olympus BX53, Tokyo, Japan) equipped with a HBO-100 mercury lamp and an Olympus DP73 camera.
To detect the GFP signals, a GFP filter set (excitation filter 460-480, dichromatic mirror 585, emission filter 495-540) was used. The microscopic photographs were processed using the CellSens Dimension V1 imaging software (Olympus, Tokyo, Japan). For each strain, one representative image is shown.

Isolation of mRNA and reverse transcription-PCR (RT-PCR)
Yeast transformants grown as described above were shifted to SGal-Ura for transgene induction. Four hours after the galactose shift, cells were harvested for RNA isolation. RNA was isolated using the RiboPure™ RNA Purification Kit for yeast (Ambion™, Thermo Fisher Scientific) following the manufacturer's instructions. Approximately 50 ng RNA were used for each RT-PCR, using AccessQuick™ RT-PCR System (Promega Corporation, Madison, USA). The primers used to detect the expression of the transgene were presented in S3

Reproducibility of the results and statistics
All experiments were repeated, independently, on three different transformants. For each individual experiment values were expressed as the mean ± standard deviation (SD) of triplicate determinations (technical triplicates) on three independent colonies of yeast transformants (biological replicates). For visual experiments the observed trends were fully consistent among the independent experiments and a representative example is shown. One sample t test was used for the statistical analysis of each strain compared with strain myrGFP or myrGFP-Cup1 under the specific conditions. Asterisks indicate the level of significance: Ã p < 0.05, ÃÃ p < 0.01, and ÃÃÃ p < 0.001.

Expression in yeast cells of myrGFP C-tagged with plant MTs
After entering the cell, free metal ions are buffered by a variety of anions present in the cytosol, such as phosphate and polyphosphate or by the metal-induced native MTs. If in excess, metal ions bind non-specifically to biomolecules, such as proteins, altering their function. We speculated that the presence of metal-binding ligands adjacent to the plasma membrane would ensure cation sequestration immediately after crossing the plasma membrane, thus preventing the potential deleterious effects of the free cations. To test this hypothesis, the cDNA of seven Arabidopsis thaliana MTs (AtMT1a, AtMT1c, AtMT2a, AtMT2b, AtMT3, AtMT4a and AtMT4b) and four Noccaea caerulescens MTs (NcMT1, NcMT2a, NcMT2b and NcMT3) (S1 Table) were translationally fused to the C-terminus of a chimeric protein consisting of an Nterminal myristoylation sequence followed by the green fluorescent protein (myrGFP)(S2 Table). Proteins bearing this myristoylation sequence (MGCTVSTQTI) are targeted preponderantly to the plasma membrane [38]. In our case the expressed myrGFP-MT fusion proteins would be anchored by myristoylation to the inner leaflet of the lipid bilayer, while the GFP domain would act as a fluorescent linker, leaving the MT domain free to face the cytosol. As the primary goal of our study was to manipulate yeast cells towards increased accumulation of metal ions without developing the metal-related toxicity symptoms, we hypothesized that the metal ions would lose their innate toxicity if they encountered a metal binding protein (MT) immediately after penetrating the cell.
To test this possibility, we made use of a vector (pGREG596, [33]) that contains the DNA cassette for myrGFP under the control of the galactose-inducible GAL1 yeast promoter [39] which allows the expression of a downstream gene solely when cells are grown on media containing galactose as carbon source (and not glucose).
The laboratory strain BY4741 was transformed with the pGRD-myrGFP::MTx plasmid series (Table 1); the transgenic strains were selected on SD-Ura medium and checked by colony PCR. Early log phase pre-cultures in liquid SD-Ura were washed and shifted to galactosecontaining selective medium (SGal-Ura) for the induction of myrGFP::MTx expression. The transcription of the chimeric constructs was checked by reverse transcription PCR (RT-PCR) using primers that flanked the GFP C-terminus and the fragment situated approximately 60 bp downstream the GFP::MTx fusion (S3 Table). The RT-PCR amplicons analyzed by agarose gel electrophoresis revealed that all the chimeric myrGFP::MTx genes were transcribed ( Fig 1A).
The production of fusion proteins was checked by fluorescence microscopy on cell samples taken every two hours from the galactose shift. It was noted that cells expressing myrGFP-MTx exhibited bright green fluorescence 4 hours after galactose shift (Fig 1B). The fluorescence concentrated at the cell periphery, suggesting that most of the myrGFP-MTx produced was targeted to the plasma membrane ( Fig 1B). Fluorescence was detected when incubation in SGal-Ura continued overnight (checked up to 24 hours).
To overrule the eventual toxicity of myrGFP-MTx expression, its effect on the growth of the transgenic lines was determined (Fig 2). Compared with the control myrGFP, the expression of myrGFP-MT x was not deleterious to cell growth, with the exception of myrGFP-AtMT3. In this case, albeit viable (S2A Fig), the cells grew slowly, having an unusually high doubling time. In contrast, cells expressing myrGFP-NcMT2a, myrGFP-NcMT2b, and especially myrGFP-NcMT3 showed more robust growth compared to cells expressing control myrGFP (Fig 2).

Accumulation of metals by yeast cells expressing myrGFP-MTx
Accumulation from media with low metal concentration. We first checked whether anchoring any of the plant MTs to the plasma membrane would trigger metal accumulation from media containing physiological concentrations of heavy metals. For this purpose MMe, a variant of the synthetic minimal medium for yeast [35] was used. MMe contained "traces" of the essential heavy metals (2 μM Mn(II) and Zn(II), 1 μM Cu(II), Co(II) and Ni(II), final concentration each). Since MTs are known to bind Cd(II), this non-essential metal was also used at 1 μM in MMe. The metal accumulated by cells expressing various myrGFP-MTx was determined by ICP-MS on samples collected 10 hours after shifting the cells to MMe-Ura/Gal, and the results are presented in Table 2.
The data revealed that all transgenic lines accumulated one or more metals from the MMe when compared to control myrGFP and this increased accumulation could be attributed to MT expression. Practically all transgenic lines accumulated Cu(II) significantly more than the control myrGFP strain (p < 0.05), with the highest accumulation being recorded for the strain expressing myrGFP-AtMT1c. This did not come as a surprise, since all MT have a natural tendency to coordinate Cu(I) [40]. Functionally, the yeast MT, Cup1, is classified as the most restrictive Cu(I)-thionein [40], but the Cu(II) accumulation for myrGFP-AtMT1c was almost doubled in comparison to myrGFP-Cup1. Interestingly, myrGFP-AtMT1a differs from myrGFP-AtMT1c by only two amino acids (S1B Fig), and although it accumulated less Cu(II) than the latter, it clearly accumulated more Cu(II) than myrGFP-Cup1. A similar behavior was noticed for myrGFP-NcMT1, and also for myrGFP-AtMT4a/b. Zn(II), another metal preferred by MTs, significantly accumulated in cells expressing myrGFP-NcMT2a/2b and myrGFP-NcMT3 (more than 100 times compared to cells expressing control myrGFP). In contrast, accumulation of Cd(II), chemically similar to Zn(II), occurred only in cells expressing myrGFP-MTx of subfamily 3 and it was rather modest, Cd(II) concentration in the medium being too low to be taken up efficiently by the innate yeast metal transporters. Remarkably Co (II), which is not a natural substrate of MTs, significantly accumulated in strains expressing myrGFP-AtMT4a/4b and myrGFP-NcMT2a/2b, the latter two also accumulating Mn(II). Notably, the sensitive strain expressing myrGFP-AtMT3 accumulated significant amounts of Cu(II), Zn(II) and Cd(II) but also Co(II) and Mn(II), a possible explanation for the poor growth of this strain. Among the metals present in the medium, Ni(II) was least accumulated by all transgenic strains, and this was in contrast to the chemically-similar Co(II). Further studies are needed to establish whether the low Ni(II) accumulation was determined by its low affinity to MTs, or simply by the low Ni(II) translocation across the plasma membrane. Nevertheless, the accumulation of Ni(II) by myrGFP-NcMT2a/2b and myrGFP-NcMT3, even though modest, was more than 10 times higher when compared to the control myrGFP.
Accumulation from media supplemented with surplus heavy metals. For bioremediation purposes, it is important to obtain strains which accumulate metals under high concentration conditions, without developing toxicity symptoms. Therefore, we further determined the accumulative potential of our transgenic myrGFP-MTx strains under higher-than-normal metal concentrations. It was noticed that metal accumulation in galactose-containing media reached a plateau following 1-2 h of metal exposure; this is why metal accumulation by cells expressing myrGFP-MTx was determined 2 hours after the shift to metal-surplus media. The Early-log phase yeast cells transformed with the pGRD series (Table 1) were grown and shifted to MMe-Ura/Gal as described in Materials and methods.
The minimal medium contained several essential metals (2 μM Mn(II) and Zn(II); 1 μM Co(II), Cu(II), Ni(II) and Cu(II)) and the non-essential metal Cd(II) (1μM). After 10 hours' incubation (30˚C, 200 rpm) the cells were harvested and processed for multielemental analysis (ICP-MS). Each determination was done in triplicate on approximately 10 8 cells from three different transformants. Results are given as mean ± standard deviation. The transgenic strains accumulated metal ions significantly more than the control strain myrGFP, in most cases. Values significantly higher than the amount of the metal accumulated by myrGFP-Cup1, according to one sample t test, are shown in bold letters and marked with asterisks: *p < 0.05, highest metal concentrations which were not deleterious to the control strain myrGFP were selected for the accumulation assays under metal surplus. These concentrations were set to 0.5 mM for Co(II), Cu(II) and Ni(II), 1 mM for Zn(II) and Mn(II), and 0.05 mM for Cd(II) and corresponded to the maximum concentrations which allowed a cell viability that was higher than 75% (S2A Fig), following exposure to metal for 2 hours (the time necessary to reach the metal accumulation plateau) or 20 hours (the time necessary to reach stationary phase). At these concentrations, the half-life (the time after which more than 50% of the cells lose viability) of all transgenic strains was longer than 24 hours (data not shown). At higher metal concentrations, cells expressing myrGFP-MTx had higher viabilities than control myrGFP (S2B Fig), suggesting that MTs expression had a protective effect against metal surplus. The growth of the transgenic cells in the presence of surplus metals was determined after 20 h of exposure, a time which allowed cells to adapt to stress conditions and to proliferate. Metal accumulation determined after 20 h exposure did not differ significantly from the values obtained for accumulation after 2 h exposure (data not shown). The metals investigated are presented further in alphabetical order, and not based on their innate affinity for MTs.
Cd(II) It was noted that all strains expressing myrGFP-MTx grew considerably better in the presence of excess Cd(II) than the control myrGFP, except myrGFP-NcMT2a (Fig 3A). This observation suggested that the very toxic Cd(II) ions are sequestered by the transgenic MTs in a less toxic form. Cd(II) accumulated significantly in strains expressing myrGFP-AtMT3, myrGFP-NcMT1, myrGFP-NcMT2a/b and myrGFP-NcMT3 (Fig 3B), but the expected combination for a potential bioremediator (robust growth and high accumulation) was only obtained for myrGFP-NcMT1, myrGFP-NcMT2b and myrGFP-NcMT3. Even though the maximum accumulation was recorded for myrGFP-NcMT2a (Fig 3B), the growth of this strain was affected by the surplus Cd(II) (Fig 3A). The best tolerance/accumulation ratio was noted for the myrGFP-NcMT2b strain, making it a good candidate for further studies on Cd(II) bioremediation.
Co(II) Co(II) impaired the growth of cells expressing myrGFP-Cup1, myrGFP-AtMTs of subfamilies 1, 2 and 3 and myrGFP-NcMT1 more than that of the myrGFP control, but not of myrGFP -AtMT4a/b, myrGFP-NcMT2a/b or myrGFP-NcMT3 (Fig 4A). As in the case of the accumulation from MMe, containing "trace" metal concentrations, the pairs myrGFP-AtMT4a/4b and myrGFP-NcMT2a/2b also accumulated Co(II) under surplus condition ( Fig  4B). However, the accumulation of Co(II) under the surplus condition was not significantly higher than that under the low-concentration condition, suggesting that a saturation limit was reached in both cases.
Maximum Co(II) accumulation was recorded for myrGFP-NcMT3 (Fig 4B), which also showed robust growth under high Co(II) conditions (Fig 4A), making this strain a good candidate for Co(II) bioremediation.
Cu(II) The growth of the yeast cells expressing myrGFP-MTx was not significantly altered by Cu(II) addition when compared to the control myrGFP, with the exception of myrGFP-NcMT2a/b and myrGFP-NcMT3 whose expression supported better growth under Cu(II) surplus ( Fig 5A). All transgenic strains accumulated more Cu(II) than the control strain expressing myrGFP (Fig 5B). Nevertheless, the only strain which accumulated significantly more Cu (II) under high concentration conditions compared to accumulation under trace Cu(II) conditions was strain myrGFP-NcMT3 (Fig 5B), which also showed robust growth under the same conditions ( Fig 5A). The cell growth was determined for each strain spectrophotometrically (OD 600 ) 20 h after adding the metal salt. B. Co(II) accumulation. Cells were exposed to Co(II) for 2 hours (30˚C, 200 rpm) before being processed for metal assay by ICP-MS. Accumulated metal was normalized to cell total protein. Values are mean ± standard deviation of three independent data. Asterisks indicate that the mean of the myrGFP-MTx strain was significantly different from the mean of the myrGFP control under the same conditions, according to one sample t test. *p < 0.05, **p < 0.01, ***p < 0.001. Cu(II) accumulation could be also visualized with naked eye in transgenic cells grown on solid media, as the cells changed color following prolonged incubation (Fig 5C).
Mn(II) As seen in Fig 6A, Mn(II) hardly affected the growth of most transgenic cells expressing myrGFP-MTx in comparison to myrGFP (Fig 6A). Mn(II) is not a natural substrate of MTs, but it was significantly taken up from normal medium by strains expressing myrGFP-NcMT2a/2b (Table 2). Under surplus conditions these two strains also accumulated Mn(II), but the accumulation was not significantly higher compared to the low concentration conditions, probably due to saturation (Fig 6B).
Although Mn(II) accumulation was noticed in most of the other transgenic strains tested (Fig 6B), the accumulation under high Mn(II) did not differ significantly when compared to the low concentration conditions. Ni(II) Under a high Ni(II) concentration only cells expressing myrGFP-NcMT1, myrGFP-NcMT2a and myrGFP-NcMT3 grew better than the myrGFP control strain (Fig 7A). As Ni(II) is known as one of the most recalcitrant metals in terms of removal from contaminated sites The cell growth was determined for each strain spectrophotometrically (OD 600 ) 20 h after adding the metal salt. B. Cu(II) accumulation. Cells were exposed to Cu(II) for 2 hours (30˚C, 200 rpm) before being processed for metal assay by ICP-MS. Accumulated metal was normalized to cell total protein. Values are mean ± standard deviation of three independent data. Asterisks indicate that the mean of the myrGFP-MTx strain was significantly different from the mean of the myrGFP control under the same conditions, according to one sample t test. *p < 0.05, **p < 0.01, ***p < 0.001.C. Cu(II) accumulation from solid medium. Cells were 10-fold serially diluted in a multi-well plate and stamped with a replicator on SGal-Ura plates containing the specified concentrations of Cu(II). Plates were photographed after 6 days' incubation at 30˚C. One representative plate is shown out of three similar experiments.
https://doi.org/10.1371/journal.pone.0178393.g005 [1], obtaining an efficient Ni(II)-hyperaccumulating organism would represent a major asset for bioremediation techniques. Ni(II) accumulation by myrGFP-MTx under normal conditions was rather modest, the highest value being recorded for the strain expressing myrGFP-NcMT2b (Table 2). This strain also accumulated most Ni(II) under surplus conditions ( Fig  7B), but its growth was not robust (Fig 7A). In fact, it was interesting to notice the apparent opposite behavior towards Ni(II) of cells expressing myrGFP-NcMT2a and myrGFP-NcMT2b (Fig 7A and 7B). Whether the small changes in the amino acid sequence (S1C Fig) are responsible for the different accumulation capacity of the two transgenic MTs are issues to be investigated in the future.
Although the strain expressing myrGFP-NcMT3 accumulated less Ni(II) than the strain expressing myrGFP-NcMT2b (Fig 7B), the former is a more suitable candidate for further studies due to its robust growth under high Ni(II) (Fig 7A). Moreover, it was noted that cells expressing myrGFP-NcMT3 gradually turned green after prolonged incubation on solid medium supplemented with Ni(II), probably due to continuous accumulation of this cation (Fig 7C).
Zn(II) It was intriguing to notice that under surplus conditions, Zn(II) accumulation by myrGFP-MTx strains did not differ significantly from Zn(II) accumulation under normal concentration conditions (data not shown). Zn(II) is an essential element supplied from the environment by two plasma membrane transporters: Zrt1 (high affinity) and Zrt2 (low affinity), both upregulated by low cellular Zn(II) [41]. It is tempting to speculate that in cells expressing myrGFP-MTx under "normal" conditions (i. e., MMe), the Zn(II) ions entering the cell were readily sequestered by the transgenic MTs, resulting in intracellular Zn(II) depletion and triggering Zrt1/2 activation. The result of this series of events would be an unusually high accumulation of Zn(II) by strains such as myrGFP-AtMT3, myrGFP-NcMT2a/b and myrGFP-NcMT3  Table 2). The Zn(II) accumulation would continue until myrGFP-MTx reached saturation, explaining why increasing the Zn(II) concentration in the growth environment would have only a modest effect upon Zn(II) accumulation.

Conclusions
Saccharomyces cerevisiae has often been used to express plant methallothioneins (MTs) either for functional elucidation through complementation studies [24,[28][29][30][31] or for their metalbinding abilities [32,[42][43][44][45]. While most of the studies concerning heterologous expression of MTs in yeast focus on canonical MT substrates, our work encompassed a greater variety of both natural and non-canonical MT substrates, all in parallel settings. Aiming at obtaining heavy metal accumulating cells we expressed seven MTs from the non-accumulator plant A. thaliana and four MTs from the hyperaccumulator plant N. caerulescens in S. cerevisiae. Unlike any other study so far, the MTs were targeted to the inner face of the plasma membrane via a The cell growth was determined for each strain spectrophotometrically (OD 600 ) 20 h after adding the metal. B. Ni(II) accumulation. Cells were exposed to Ni(II) for 2 hours (30˚C, 200 rpm) before being processed for metal assay by ICP-MS. Accumulated metal was normalized to cell total protein. Values are mean ± standard deviation of three independent data. Asterisks indicate that the mean of the myrGFP-MT line is significantly different from the mean of the myrGFP control under the same conditions, according to one sample t test. *p < 0.05, **p < 0.01. C. Ni(II) accumulation from solid medium. Cells were 10-fold serially diluted in a multi-well plate and stamped with a replicator on SGal-Ura plates containing 2 mM NiCl 2 . Plates were photographed after 6 days' incubation at 30˚C. One representative plate is shown out of three similar experiments.
https://doi.org/10.1371/journal.pone.0178393.g007 myristoylation sequence that was N-fused to GFP. This approach has the advantage that the osmotically-free metal ions that reach the cytosol are readily sequestered by the anchored MTs, increasing the chances to obtain a hyperaccumulating strain with increased tolerance to otherwise toxic heavy metals.
With the exception of myrGFP-AtMT3, the expression of none of the other myrGFP-MTx affected yeast growth negatively, allowing high accumulation of metals from environments containing either low or high metal concentrations. It is of note that, in most cases, the transgenic strains which expressed these plant MTs had higher accumulative power than the strain expressing myrGFP-Cup1 (yeast MT).
All transgenic strains expressing myrGFP-MTx accumulated Cu(II) under low concentration conditions and the possibility to use these strains to extract similar ions such as Ag(I) is under investigation. Of note are also the constructs which allowed the accumulation from "normal" media of metals less usual in MT-related studies: Co(II) (myrGFP-AtMT4a/b, myrGFP-NcMT2a/b) or Mn(II) (myrGFP-NcMT2a/b). The reasons for such preferences are now being considered by molecular modelling, starting from the structural differences between MT classes, but also from the differences between MTs within the same subfamily.
It was interesting to notice that compared to myrGFP and myrGFP-Cup1 we obtained one poorly growing (myrGFP-AtMT3) and one robust growing strain (myrGFP-NcMT3), both belonging to the same family of plant MTs. Remarkably, myrGFP-AtMT3 accumulated almost indiscriminately all the metal ions tested under low concentration conditions, a possible explanation for its reduced growth rates. The viability of this strain was however unusually high, as cells survived for days under high metal surplus without proliferating, but exhibiting continuous fluorescence at the plasma membrane level (data not shown). Due to these characteristics the myrGFP-AtMT3 strain is an interesting candidate for further studies. In contrast, myrGFP-NcMT3 seemed the most suitable strain for bioremediation studies, as it exhibited robust growth, metal tolerance and high accumulative capacity under high concentration conditions for the metals tested in this study. Early log phase growing cells transformed with pGRD-myrGFP::MTx series were shifted to SGal-Ura for transgene induction. Four hours after the galactose shift, MeCl 2 was added at the indicated concentration and samples were harvested for viability assay by methylene blue staining. Cell viability was expressed as percentage of live cells within a whole population. Viability was examined for at least 300 cells from one biological replicate. Viable cells were colorless, and dead cells were blue. Values are means ± standard deviation of three independent data. A. Maximum metal concentrations which did not significantly alter cell viability. B. At higher metal concentrations, viability of cells expressing myrGFP-MTx is less affected compared to control myrGFP. (TIF) S1 Table. Primers used to clone plant metallothionein cDNAs. The primers were used to amplify plant MTs cDNAs from Arabidopsis thaliana (At) or Noccaea caerulescens (Nc) cDNAs. The MT gene CUP1 from Saccharomyces cerevisiae (Sc) is intronless and was amplified from genomic DNA. The primers used introduced restriction sites suitable for subsequent subcloning of the amplified fragments into yeast vectors. To amplify the myrGFP control, pGREG596 was used as template. The map of plasmid pGREG596 can be found at (http://web. uni-frankfurt.de/fb15/mikro/euroscarf/data/pGREG.html).