Wild-type (WT) tobacco Nicotiana tabacum var. Xanthi and generated tobacco transgenic plants expressing pMDC163::promNtNRAMP3::GUS were used in the study. Seeds of the WT tobacco were obtained from the stock of the Institute of Biochemistry and Biophysics PAS (Warsaw, Poland; in 2002) and then propagated in the greenhouse of the University of Warsaw.

Plants were cultivated in a growth chamber at 23/16°C day/night, 40–50% humidity, 16 h photoperiod, and quantum flux density [photosynthetically active radiation (PAR)] 250 μMol m⁻² s⁻¹, fluorescent Flora tubes (Siemianowski et al., 2011).

Seeds were surface sterilized in 8% sodium hypochlorite (w/v) for 2 min, washed with distilled water, then germinated and grown on vertically positioned Petri dishes containing quarter-strength Knop's medium, 2% sucrose (w/v), and 1% agar (w/v). After 3 weeks seedlings were transferred to hydroponic conditions (five plants per 2.5 L pot) and further grown according to experiment-specific schemes presented below. In all experiments, the quarter-strength Knop's medium was applied as a reference control medium (Barabasz et al., 2010). All experiments were conducted in three independent biological replicates. The nutrient solution was renewed every third day.

To determine developmental regulation of NtNRAMP3, 3-week-old WT seedlings were transferred from the Petri dishes into the hydroponic control medium for up to 5.5 weeks. Plant material was collected at three developmental stages: (1) 4-week-old seedlings (3 weeks on plates and 1 week on hydroponics): whole roots and all leaves were collected separately; (2) 6-week-old plants (3 weeks on plates and 3 weeks on hydroponics): whole roots and all leaves were collected separately; (3) 9-week-old plants (3 weeks on plates and 6 weeks on hydroponics); the following plant parts were collected: (a) apical part of the root (3–4 cm from the tip), (b) basal part of the root (3–4 cm from the base), (c) stem (3 cm of the middle part), (d) young leaves (two leaves of minimum 0.5 cm length counting from the top), and (e) old leaves (two leaves counting from the base). Lateral roots were not included in the analysis. The material was pooled from a total of 30 (stage 1), 15 (stage 2), or 10 (stage 3) plants, frozen in the liquid nitrogen and stored at −80°C.

For assessment of metal status-dependent expression of NtNRAMP3, 3-week-old WT plants were further hydroponically grown on a control medium for 2 weeks. Then, they were exposed to different regimes: (i) metal excess (200 μM Fe or 100 μM Mn or 20 μM Co or 20 μM Cu or 30 μM Ni or 50 μM Zn or 4 μM Cd); (ii) metal deficiency (Fe or Mn or Co or Cu or Zn; metal was not added to the medium); (iii) control conditions. Metal was added to the control medium as ZnSO₄; Fe-EDTA; MnSO₄; CoCl₂; CuSO₄; NiCl₂ or CdCl₂. After 3 days, the following organs were collected: (i) 3–4 cm fragment of the apical part of the root; (ii) 3–4 cm fragment of the basal part of the root; (iii) leaves (2nd and 3rd leaf counting from the base of the stem) without petioles and major midribs. The material was pooled from a total of 10 plants, frozen in the liquid nitrogen, and stored at −80°C.

To compare the NtNRAMP3 expression between leaves of tobacco exposed to a long-term treatment of elevated Zn, 3-week-old tobacco WT seedlings were grown on a liquid control medium for 18 days, then for 3 weeks at 10, 50, and 200 μM Zn. Next, all leaves (without petioles and midvein) were removed from the stem and collected in groups of two (e.g., 1st and 2nd leaves counting from the base of the stem were the 1st pair; 3rd and 4th−2nd pair, etc.), frozen in the liquid nitrogen and stored in −80°C. The material was pooled from a total of 8 plants.

To examine the tissue-specific NtNRAMP3 expression patterns, transgenic plants expressing pMDC163::promNtNRAMP3::GUS were subjected to two different treatments. In the first experiment, three-week-old plants were transferred from agar plates to hydroponics and grown under control conditions for 7 days. Whole seedlings were used to determine GUS activity. In the second experiment, 5-week-old plants grown at control conditions (3 weeks on agar plates and 2 weeks on hydroponics) were exposed to 200 μM Zn for 3 days. In the end, the second leaf (counting from the base of the stem) was collected for GUS assay. Wild-type tobacco plants were used as a negative control.

A partial genomic sequence of NtNRAMP3 (previously annotated as NtNRAMP3-like) was identified, and NtNRAMP3 was reported as a candidate gene encoding metal transporter potentially involved in zinc transport in tobacco leaves (Papierniak et al., 2018). In brief, known nucleotide NRAMP sequences from A. thaliana were blasted against the N. tabacum genome (Sierro et al., 2013, 2014), which is deposited in GenBank, using NCBI (National Center for Biotechnology Information) BLASTn program (http://www.ncbi.nlm.nih.gov/blast). After screening with the FGENESH+ tool (SoftBerry, Mount Kisco, NY, United States), the full-length NtNRAMP3 putative sequence was identified within the scaffold AWOK01S026429.

The full-length open reading frame (ORF) of NtNRAMP3 (containing STOP codon or not) was amplified from cDNA transcribed from total RNA using appropriate primers (Table 1) by PCR with Phusion HF polymerase (Thermo Scientific). Then, the full-length NtNRAMP3 sequence was cloned into the pENTR/D-TOPO vector using the pENTR Directional TOPO Cloning Kit (Invitrogen). Finally, two constructs were obtained—pENTR/D-TOPO::NtNRAMP3 and pENTR/D-TOPO::NtNRAMP3-STOP (STOP codon was included).

To prepare the construct for subcellular localization, pENTR/D-TOPO::NtNRAMP3 was recombined with the pMDC43 vector (purchased from Arabidopsis Biological Resource Center, https://abrc.osu.edu/) using LR reaction (Gateway Technology, Invitrogen). The resulting plasmid pMDC43::GFP::NtNRAMP3 contained GFP linked at the N-terminus of NtNRAMP3. The fusion protein was expressed under 35S promoter control.

For yeast growth assay the vectors pAG426GAL::NtNRAMP3 were generated by LR recombination between pENTR/D-TOPO::NtNRAMP3 and pAG426GAL-ccdB-EGFP plasmid.

For a generation of the construct for GUS analysis, a 1,684 bp promoter region located upstream from the translation initiation codon of NtNRAMP3 was amplified from genomic DNA using appropriate primers (Table 1) by PCR with Phusion HF polymerase (Thermo Scientific) and inserted into the pENTR/D-TOPO vector using pENTR Directional TOPO Cloning Kit (Invitrogen). Then, the LR reaction was used to obtain the pMDC163::promNtNRAMP3::GUS construct.

The nucleotide sequence of NtNRAMP3 was translated to a protein sequence with the ExPASy translate tool (https://web.expasy.org/translate/). Then, to view predicted transmembrane domains of NtNRAMP3, the web-based software Protter (http://wlab.ethz.ch/protter/start/) that gathers protein features from various annotation sources, such as Uniprot, was used.

The NtNRAMP3 amino acid sequence was aligned to several chosen NRAMP amino acid sequences from other plant species (Arabidopsis thaliana, Solanum lycopersicum, Zea mays, Theobroma caccoa, and Nicotiana) identified in ARAMEMNON, Solgenomics, MaizeSequence, Phytozome, and NCBI, using Clustal Omega (https://www.ebi.ac.uk/Tools/msa/clustalo/). The phylogenetic analyses were conducted with the MEGA X program (https://www.megasoftware.net) (Kumar et al., 2018) using the maximum likelihood method with 1000 bootstrap replicates. The prediction of membrane-spanning regions was performed with Phobius software (https://phobius.sbc.su.se/) (Käll et al., 2004).

All primers used in this study were designed with the OligoAnalyzer tool (https://www.idtdna.com/pages/tools/oligoanalyzer) based on the sequences of NtNRAMP3 (primers for expression analysis and NtNRAMP3 ORF amplification) and scaffold AWOK01S026429 (primers for NtNRAMP3 promoter amplification). Primers sequences are listed in Table 1.

Regulatory elements within the NtNRAMP3 promoter region were identified with the program PlantCARE (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/).

The construct pMDC163::promNtNRAMP3::GUS was stably transformed into tobacco plants using GV3101 A. tumefaciens-mediated transformation protocol for tobacco leaf disks (described in Siemianowski et al., 2011). Transgenic plants were selected in the presence of hygromycin B (100 μg·ml⁻¹). The T1 heterozygous lines with a segregation ratio of 3:1 (hygromycin^toler: hygromycin^sensit) were self-pollinated to obtain the homozygous generation (T2). Developed independent homozygous lines were tested in a preliminary histochemical GUS assay in young 4-week-old seedlings, subsequently three lines were selected for more detailed analysis.

Total RNA extraction and quantitative real-time RT-PCR (qPCR) analysis were performed according to procedures described in Papierniak et al. (2018). In brief, Plant RNA Mini Kit (Syngen, Wrocław, Poland) was used for RNA isolation, then samples were digested with DNase I (Invitrogen, Waltham, MA, United States). The relative quantities of each transcript were calculated based on the comparative ΔCt (threshold cycle) method (Livak and Schmittgen, 2001). The NtPP2A (Nicotiana tabacum Protein Phosphatase 2 A) gene was used as an internal control, and its stability was measured in all plant tissue samples in the range of applied metal concentrations. At least three independent biological replicates were analyzed for each experiment. A minimum 2-fold change in relative gene expression level was considered significant.

In the study, two wild-type (DY1457 and BY4742) and three mutants (Δsmf1, Δzrt1, Δfet3fet4) Saccharomyces cerevisiae strains (Supplementary File S1) were transformed with the empty vector pAG426GAL or with pAG426GAL::NtNRAMP3 construct, according to lithium acetate method protocol (Gietz and Schiestl, 2007). Complementation and sensitivity tests were performed as described previously (Papierniak-Wygladala et al., 2020). Briefly, yeast cultures grown on liquid synthetic medium (SC-URA+GLU; yeast nitrogen base, amino acids without uracil, 2% glucose, pH 5.3), after setting the optical density (OD₆₀₀) to 0.2, were serially diluted (1.0; 0.1; 0.01 and 0.001), and then spotted on Petri dishes containing SC-URA+GAL solidified with 2% (w/v) agar and supplemented with required components (details below). Yeast growth was monitored for the next up to 10 days.

In complementation tests, for the Δfet3fet4 mutant, the restrictive medium was supplemented with (i) 10 or 30 μM FeCl₃, (ii) 30 or 100 μM ferric citrate, and/or (iii) 30 or 80 μM BPDS (bathophenanthrolinedisulfonic acid). For Δsmf1 to the medium 2.5 mM EGTA (Ethylene glycol-bis(β-aminoethyl ether)-N, N, N′, N′-tetraacetic acid) alone or with 0.1 mM MnCl₂ (pH was adjusted to 6.0 with 50 mM MES or SC-URA+GAL medium of pH 5.3 was applied) were added. The Δzrt1 mutant was assayed for growth on medium containing (i) 1.0 mM EDTA (Ethylenediaminetetraacetic acid), and/or (ii) 100, 500, 600, or 700 μM ZnCl₂. Wild-type strains transformed with the empty pAG426GAL vector were included as a reference control in all assays (DY1457 in experiments including Δfet3fet4 and BY4742 in experiments including Δsmf1 and Δzrt1). Mutant strains transformed with the empty pAG426GAL vector were included as a negative control in all experiments.

In sensitivity tests, the growth of DY1457 yeast transformed with pAG426GAL::NtNRAMP3 construct was monitored on plates containing SC-URA+GAL supplemented with (i) 5, 10, 20, 50, or 75 μM CdCl₂, (ii) 50, 100, 250, 500, or 750 μM CoCl₂, (iii) 1.0, 1.4, 2.0, 2.2, or 2.4 mM CuSO₄, (iv) 0.2, 3.0, 4.0, 4.5, or 5.0 mM FeCl₃, (v) 0.2, 2.5, 5, 7.5, or 10.0 mM MnCl₂, (vi) 0.2, 0.3, 0.4, 0.5, or 0.6 mM NiCl₂, (vii) 0.2, 1.0, 2.5, 5.0, or 7.5 mM ZnSO₄, and (viii) 4.0, 4.5, 5.0, 5.5, or 6.0 mM ZnSO₄ (pH was adjusted to 4.0, 5.0, or 6.0 with 50 mM MES). DY1457 transformed with the empty pAG426GAL vector was included as a control in all assays.

The subcellular localization of NtNRAMP3 was determined by monitoring the transient expression of GFP-NtNRAMP3 translational fusion product in tobacco leaf epidermal cells accordingly to the previously described protocol (Siemianowski et al., 2013; Papierniak et al., 2018). The GV3101 A. tumefaciens (C58C1, Rif^R; pMP90, Gm^R) transformed with pMDC43::GFP::NtNRAMP3 were used for infiltration of 7-week-old WT tobacco leaves. To visualize the cell wall, after 24–72 h from the infiltration leaf fragments were stained with 50 μM water solution of propidium iodide (PI) for 20 min. Then, GFP (excitation: 488 nm line of argon laser; emission: 500–560 nm) and PI (excitation: 543 nm; emission: 617 nm) signals were detected in Nikon A1 confocal laser scanning microscope (Melville, NY, USA).

The GUS assay in young seedlings (4-week-old whole plants) was performed according to Weremczuk et al. (2020). In brief, 4-week-old whole seedlings were fixed in 90% ice-cold acetone and washed in the 50 mM phosphate buffer pH 7.0 with 0.2 % Triton X-100 with infiltration. Then, the plant material was incubated in a reaction buffer (50 mM phosphate buffer pH 7.0 with 0.2 % Triton X-100 and 2 mM X-Gluc), at 37°C in the darkness for 2.5 h. Next, they were fixed in FAA (formalin-acetic acid-alcohol) for 30 min, dehydrated in increasing ethanol concentrations, and stored in 70% ethanol. To visualize expression sites of NtNRAMP3 within leaves of 5.5-week-old tobacco, we used the method described in Weremczuk et al. (2020). In brief, the determination of GUS activity in the leaves of 5.5-week-old tobacco was carried out on cross-sections of 130 μm thickness cut on a Vibratome (Leica VT1000S, Heidelberg, Germany). They were used for histochemical GUS staining, subsequently fixed in FAA for 30 min, dehydrated in increasing ethanol concentrations, and stored in 70% ethanol. Observations were performed with an OPTA-TECH microscope.

Earlier studies showed that the partial sequence of NtNRAMP3-like first identified in tobacco within the scaffold AWOK01S026429 was upregulated in the leaves by 200 μM Zn (Papierniak et al., 2018). Here, this sequence has been cloned. The full-length cDNA is comprised of 1,542 bp and encodes a protein of 514 amino acids (Figure 2; FGENESH+). Analyses of the nucleotide and amino acid sequences provided evidence that the identified sequence is NtNRAMP3.

First, phylogenetic analysis that included 32 NRAMP proteins from A. thaliana, S. lycopersicum, T. cacao, Z. mays, and Nicotiana species (N. attenuata, N. sylvestris, N. tomentosiformis, N. tabacum) showed that the newly identified protein was located on the same branch as other NRAMP3 proteins, with the closest relationship between NtomNRAMP3 and AtNRAMP3 (Figure 1). In general, the examined proteins were classified into four groups. Among them, NtNRAMP3 formed a separate clade with NaNRAMP3, NsNRAMP3, NtomNRAMP3, TheccNRAMP3, LeNRAMP3, AtNRAMP3, and AtNRAMP4, which was distinct from other NRAMPs (Figure 1, marked with bracket). Furthermore, among all NRAMP proteins included in the study, NtNRAMP3 was the most similar to NtomNRAMP3 (99.81%), and the highest identity of the amino acid sequences was observed within Nicotiana (98.82% and 98.23% between NtNRAMP3 and NaNRAMP3 and NsNRAMP3, respectively, Table 2). On the other hand, the lowest identity within this group was observed between NtNRAMP3 and AtNRAMP4 (73.81%).

Like many other NRAMP proteins, NtNRAMP3 consists of 12 transmembrane domains (TMDs) and has both N- and C-termini located intracellularly (Phobius, Käll et al., 2004; Protter, Omasits et al., 2014) (Figure 2; Supplementary File S2). In silico analysis based on Phobius and Protter suggested also that NtNRAMP3 has one N-glycosylation motif of amino acid position 505 located in the C-terminus.

Second, multiple amino acid sequence alignment showed high conservation between the NtNRAMP3 and examined NRAMP proteins (Supplementary File S2). NtNRAMP3 contains key sequences considered typical for NRAMP proteins. The consensus transport motif (CTM) [GQSSTIT(/A)G(/D)TYAGQY(/F)V(/I)MQ(/G/E)GFLD(/H/N)], which is the signature sequence of NRAMP family, was present in all examined proteins. Moreover, three highly conserved histidine residues (one within the region between II and III TMD and two within VI TMD) were identified.

Histidine residues occur at different locations in the NRAMP protein sequences. The location of some is highly conserved for all examined NRAMP proteins (as between II and III TMD and within VI TMD). Others are distributed in various locations within the sequence, however, the same for certain groups of plants. Thus, in the structure of NtNRAMP3 together with NtomNRAMP3, NaNRAMP3, NsNRAMP3, and LeNRAMP3, an additional histidine was present within CTM. Furthermore, additional histidine residues, H4 and H40, were present in the N-terminus of NtNRAMP3, and also within NtomNRAMP3, NaNRAMP3, and NsNRAMP3 (Supplementary File S2).

To determine the subcellular localization of NtNRAMP3 protein, the pMDC43::GFP::NtNRAMP3 construct was transiently expressed in tobacco leaf epidermal cells (Figure 3). The green fluorescence was present along the contours of the strongly folded cell walls of the lower epidermal cells (Figure 3A1) and overlapped with the red signal coming from the cell walls stained with propidium iodide (Figures 3B1,D1). Since the plasma membrane sticks to the primary cell wall, it is not distinguished by confocal light microscopy. Thus, the detected colocalization of both green and red signals indicates that NtNRAMP3 is localized in the plasma membrane. It is also known that the central vacuole does not enter the narrow projections of the epidermal cell, so the lack of a green signal at their bases additionally confirms the location of the tested protein in the plasma membrane (Siemianowski et al., 2013; Pighin et al., 2014; Barabasz et al., 2019). Autofluorescence was extremely low (Figures 3A3,D3). The green fluorescence of GFP alone was observed only in small circular structures within the cells (Figures 3A2,D2).

Natural resistance-associated macrophage proteins are typically involved in the transport of several metals, including Fe, Mn, Ni, Co, Cu, Zn, and Cd (Gunshin et al., 1997; Thomine et al., 2000; Mizuno et al., 2005; Wang et al., 2019). Here, to identify substrates for NtNRAMP3, we applied two types of yeast growth assay—sensitivity tests (comparison of the growth of WT yeast transformed with either empty pAG426GAL vector or the construct pAG426GAL::NtNRAMP3) and complementation tests (comparison of the growth of mutant yeast transformed with either empty pAG426GAL vector or the construct pAG426GAL::NtNRAMP3; as the reference WT yeast transformed with the empty vector was used). Accordingly, NtNRAMP3 expression was tested in the selected mutants (Δfet3fet4, Δsmf1, or Δzrt1) and WT (DY1457 or BY4742) yeast strains.

First, the sensitivity of strain DY1457 (WT) expressing pAG426GAL or pAG426GAL::NtNRAMP3 to various concentrations of Fe, Mn, Co, Cu, Ni, Cd, and Zn was compared. Expression of NtNRAMP3 led to inhibition of yeast growth in the presence of high concentrations of Fe (Figure 4A), Mn (Figure 4C), Co (Figure 4E), Cu (Figure 4F), Ni (Figure 4G), or Cd (Figure 4H). However, expression of NtNRAMP3 in strain DY1457 increased yeast sensitivity to Zn only in the presence of a strictly defined concentration—around 5.0 mM (Figure 4I). To learn more, the experiments were extended by the use of three pH values (4.0, 5.0, and 6.0 adjusted with 50 mM MES; the pH of the SC-URA+GAL medium was 5.3) and five concentrations of Zn (4.0; 4.5; 5.0; 5.5; 6.0 mM). Likewise, NtNRAMP3-expressing yeast grew slower with the smallest difference to the control at 4.0 mM Zn applied at all tested pH values (Figures 4J–L).

Next, yeast mutants defective in Fe and Mn uptake were used. The expression of NtNRAMP3 restored the growth of Δfet3fet4 (Figure 4B) and Δsmf1 (Figure 4D) under Fe- or Mn-limited conditions, respectively. In contrast, expression of NtNRAMP3 did not restore the growth of the Δzrt1 mutant under Zn-limited conditions induced with a strong chelator EDTA (Figure 4M), and the growth of Δzrt1 yeast carrying either the empty pAG426GAL vector or the construct (pAG426GAL::NtNRAMP3) was severely restricted.

These results indicate that NtNRAMP3 is an influx transporter for Fe, Mn, Co, Cu, Ni, and also toxic Cd. It also seems to transport Zn, but probably only under very specific environmental conditions/medium composition.

To determine the role of NtNRAMP3 in tobacco, the level of its expression in organs during vegetative development under control conditions was analyzed (Figure 5). In young 4-week-old and also in 6-week-old plants NtNRAMP3 was preferentially expressed in the developing leaves, and its expression level was 2- to 3-fold higher in leaves compared with roots. The pattern changed in older, 9-week-old plants. Studies carried out on the apical and basal parts of the root, the stem, and young and older leaves did not show any significant differences. In general, however, the level was moderately lower compared with high expression in the leaves of younger plants.

Knowing that NtNRAMP3 mediates the transport of Fe, Mn, Co, Ni, Cu, and Cd, subsequent experiments were performed to define how NtNRAMP3 is regulated by the status of these metals. The main question was whether it is actively involved in the plant's response to excess metals or the supply of metals upon their deficiency. The appearance of plants after exposure to the tested metals is illustrated in Supplementary File S3. Organs, such as apical and basal parts of the roots and leaves, were included in the study. The stability of the reference NtPP2A gene is shown in Supplementary File S4.

With some exceptions, the highest level of expression of NtNRAMP3 was detected under control conditions. In experiments with the deficiency of Fe, Mn, Co, or Cu, NtNRAMP3 expression decreased significantly in all tested organs compared with control conditions (Figures 6B,D,F,H). Considering Zn deficiency, NtNRAMP3 was lower only in the leaves (Figure 6J). Importantly, the transcript level also decreased after plants were exposed to high concentrations of metals (200 μM Fe or 100 μM Mn or 20 μM Co or 20 μM Cu or 30 μM Ni), excluding leaves and apical root parts after exposure to Co and Cu, respectively (Figures 6A,C,E,G,I). Also, no changes in the expression level were found after exposure to 50 μM Zn (Figure 6K), as well as in the presence of 4 μM Cd (Figure 6L).

A previous study showed that the expression of NtNRAMP3 in tobacco leaves went up in response to up to 4-day exposure to very high (200 μM) Zn, suggesting NtNRAMP3 involvement in plants' response to toxic zinc levels (Papierniak et al., 2018). To initially verify this possible function, a more detailed examination of the transcript level in consecutive pairs of leaves of long-term exposed tobacco plants to elevated concentrations of Zn was performed. Two key pieces of information were obtained in this experiment. The expression of NtNRAMP3 was strongly induced by high Zn concentration, however, in a manner dependent on the age/position on the stem, but it was not regulated in the youngest leaves (5th pair) (Figure 7). The highest upregulation resulted from exposure to 200 μM Zn in the 2nd and 3rd pair of leaves. The 21-day exposure to 50 μM Zn also resulted in enhanced expression, however, to a much lesser extent.

Taken together, our results indicate a general role of NtNRAMP3 in the regulation of the balance of several metals in plants grown under control conditions, with a possible specific role at exposure to extreme Zn-imposed stresses.

The new sequence was assigned to the NRAMP family based on the degree of homology to nucleotide and amino acid sequences characteristic of this family, and also on the structure of the protein. The new sequence was named NtNRAMP3 based on its highest homology to NRAMP3 sequences from other plant species (Figure 1; Table 2). Dendrogram analysis showed the closest relationship between NRAMP3 from all eight plant species used for the study, which formed a separate branch (Figure 1). The highest degree of homology was found between NtNRAMP3 and NRAMP3 proteins from other tobacco species (99.81–98.04%), as well as for NRAMP3 from other plants, such as S. lycopersicum or A. thaliana (91.14 and 77.18%, respectively) (Table 2).

Homology-based analysis revealed that the overall structure of NtNRAMP3 is conserved, except for the N-terminal and C-terminal regions. These include the presence of 12 TMDs, a glycosylation signal at the C-terminus, and the consensus transport motif (CTM) between VIII and IX TMD (Figure 2; Supplementary File S2), which are indicative of the NRAMP family (Nevo and Nelson, 2006).

Furthermore, the NtNRAMP3 sequence contains essential amino acid residues important for the transport function. These include histidines, which are also believed to contribute to the functional divergence of the NRAMP family (Chaloupka et al., 2005; Ihnatowicz et al., 2014). First, NtNRAMP3 possesses conserved His residues, one within the loop between the II and III TMD and two within VI TMD, including the Met-Pro-His motif (Supplementary File S2). Secondly, in most NRAMP proteins, there are also His residues in a variable arrangement throughout a sequence. Three histidines (H4, H33, and H40) were found in the NtNRAMP3 sequence in the N- terminus, and interestingly, was also present in NRAMP3 from five plant species with the highest homology to NtNRAMP3 (NtomNRAMP3, NaNRAMP3 and NsNRAMP3 proteins, and also in LeNRAMP3 and AtNRAMP3). Two His residues were identified within the loop between VI and VII TMD, in NtNRAMP3 and NtomNRAMP3, NaNRAMP3, and NsNRAMP3 (Supplementary File S2).

The NRAMP proteins carry ions toward the cytoplasm (Nevo and Nelson, 2006). Thus, if a protein is targeted to the plasma membrane it mediates metal uptake, and localization to the tonoplast or the membranes or the Trans-Golgi Network (TGN) indicates involvement in redistribution. Thus, the detected localization of the NtNRAMP3-GFP fusion protein in the plasma membrane (Figure 3) provided indirect evidence for the contribution of the NtNRAMP3 to the uptake of metals. Further pieces of evidence come from yeast experiments. We showed that the NtNRAMP3 protein functions as a Fe, Mn, Co, Cu, Ni, and Cd uptake transporter, as revealed by the functional complementation of the fet3fet4 and smf1 yeast mutants, and by the sensitivity tests demonstrating growth inhibition of WT yeast DY1457 expressing the NtNRAMP3 cDNA (relative to the WT expressing the empty vector), in the presence of elevated concentrations of all tested metals (Figure 4). However, when drawing conclusions, one must remember that sometimes the results of yeast complementation tests are not consistent with the subcellular localization of examined proteins. For example, AtZIP1 is localized to the vacuole of Arabidopsis protoplasts, but an expression of AtZIP1 complemented a yeast mutant defective in plasma membrane Zn uptake. Two possible explanations for this phenomenon have been proposed. First, Zn deficiency-inducible AtZIP1 might efflux Zn to the cytoplasm of transformed yeast cells which promotes growth on low Zn. Secondly, the plant transporter could be mislocalized to the yeast plasma membrane (Milner et al., 2013). Likewise, AtNRAMP3 and AtNRAMP4 were found to complement the fet3fet4 Fe uptake mutant although, as being targeted to the tonoplast in plant cells, they participate in metal redistribution (Thomine et al., 2000, 2003; Languar et al., 2005). In our studies, however, the result of the subcellular localization of NtNRAMP3 (plasma membrane, Figure 3) was consistent with the direction of transport resulting from yeast tests (uptake, Figure 4). This supports the conclusion that NtNRAMP3, as a protein with low substrate specificity, is capable of carrying several metals.

For comparison, similar to tobacco, the presence of NRAMP3 in rice was found in the plasma membrane, but AtNRAMP3, TcNRAMP3, and LeNRAMP3 from three other plant species (A. thaliana, T. caerulescens, and tomato) were localized in the tonoplast (Bereczky et al., 2003; Thomine et al., 2003; Oomen et al., 2009). The NRAMP proteins with predominant localization in the plasma membrane include NRAMP1 and NRAMP5. This was shown for most of the NRAMP1 proteins known to date (from A. thaliana, M. truncatula, N. tabacum, O. sativa, and Sedum alfredii) (Thomine et al., 2000; Takahashi et al., 2011; Sano et al., 2012; Tejada-Jiménez et al., 2015; Zhang et al., 2020), except for NRAMP1 from G. max or tomato, which was found in intracellular membranes (Bereczky et al., 2003; Qin et al., 2017). Similarly, known NRAMP5 proteins (from O. sativa, G. max, and H. vulgare) were also targeted in the plasma membrane (Sasaki et al., 2012; Qin et al., 2017).

Although NRAMPs are known as transporters mediating translocation of a broad range of metals (Nevo and Nelson, 2006), the ability of NtNRAMP3 to carry six possible substrates identified in this study (Figure 4) seems to be unique among plants for this class of the proteins. However, the identification of six substrates for NtNRAMP3 resulted from the large scope of yeast-based analyses performed. Until now, usually up to four metals have been tested as possible substrates for an examined protein. For comparison, it has been shown that within the NRAMP3 proteins, the substrate for OsNRAMP3 was Mn, but not Fe or Cd (Yamaji et al., 2013), for AtNRAMP3 or TcNRAMP3 Mn, Fe, Cd, but not Zn (Thomine et al., 2000; Languar et al., 2004; Oomen et al., 2009). Similarly, the number of metals transferred by NRAMP proteins from the other classes (NRAMP1-2, NRAMP4-6) is usually a maximum of 2 to 4 (Thomine et al., 2000; Bereczky et al., 2003; Languar et al., 2004, 2010; Yamaji et al., 2013; Zhang et al., 2020). TjNRAMP4 is an opposite example of an NRAMP protein with high substrate specificity, for which only Ni (but not Fe, Zn, Mn, or Cd) was identified as a substrate (Mizuno et al., 2005).

Considering these data, it seems that NtNRAMP3, capable of carrying up to six different metals, can play a unique role in tobacco. For the proper growth and development of a plant, it is necessary to maintain the correct concentration of many metals (both micro- and macro-elements) (Baxter, 2009; Sperotto et al., 2014; Che et al., 2018). This is due to the interconnected network of regulation of the metal-related genes, for which tissue-specific expression depends on the status of metals in the cells. Underlying mechanisms are poorly understood, especially when we take into account the simultaneous regulation of the level of various metals in a given cell, tissue, or organ. In general, these include the presence of a range of cis-acting elements present in the promoter region, as well as the specificity of the transcription factors. Important components of such regulation may also be genes encoding multi-metal transport proteins, such as NtNRAMP3.