Tobacco (Nicotiana tabacum L. cv. CB-1) seedlings were first germinated in soil trays and the grown under standard conditions for approximately 5-6 weeks, until they reached the seven-leaf stage. Subsequently, tobacco seedlings were transferred from soil trays to 1000 mL plastic boxes filled with a modified Hoagland nutrient solution. The nutrient solution consisted of the following components: 4.66 mM Ca(NO₃)₂·4H₂O, 1.41 mM KH₂PO₄, 4.98 mM KNO₃, 1.99 mM MgSO₄·7H₂O, 0.10 mM FeSO₄·7H₂O, 0.10 mM EDTA-2Na, 46.26 μM H₃BO₃, 9.10 μM MnCl₂·4H₂O, 0.77 μM ZnCl₂, 0.41 μM CuCl₂·2H₂O, and 0.13 μM Na₂MoO₄·2H₂O (Lu et al., 2023).

After a pre-culture of one week in growth chambers under a 12-hour light at 25°C/12-hour dark at 20°C, the well-grown tobacco seedlings were transferred to new plastic boxes and cultured in the modified Hoagland nutrient solutions with the Mg concentrations of Mg-deficient (0 mM), 1/4-normal Mg (0.50 mM), normal Mg (1.99 mM), and 4×-normal Mg (7.96 mM) (hereafter designated as Mg0, Mg1/4, Mg1, and Mg4), respectively. Tobacco seedlings were maintained in the growth chambers under the same light-dark cycle and temperature conditions as during the pre-culture phase. The liquid solutions in the plastic boxes were completely replaced every three days to ensure fresh nutrients for the seedlings. Each treatment group consisted of three replicates, with each replicate containing three seedlings.

The first fully expanded leaf of tobacco seedlings grown under different levels of Mg supply was sampled at 25 days after treatment (DAT). A total of 14 physiological parameters were measured, including the contents of Mg, chlorophyll a (Chla), chlorophyll b (Chlb), carotenoids (Car), soluble proteins (SP), and hydrogen peroxide (H₂O₂); the activities of ribulose-1,5-bisphosphate carboxylase (RuBPCase), acid invertase (AI), neutral invertase (NI), nitrate reductase (NR), superoxide dismutase (SOD), catalase (CAT), and peroxidase (POD); root vitality; as well as cell membrane permeability (CMP). Each experiment had three biological replicates.

The measurements were conducted following established methods as previously described (Lu et al., 2023). Specifically, the content of Mg was determined using inductively coupled plasma optical emission spectroscopy (ICP-OES) system (iCAP 7000 Series, Thermo Fisher Scientific, USA) (Musharraf et al., 2012). The contents of Chla, Chlb, and Car were measured using a spectrophotometer, as per the protocol by Lichtenthaler and Buschmann (2001); SP and H₂O₂ levels were quantified following the procedures described by Anderson et al. (1995) and by Clemensson-Lindell (1994), respectively; RuBPCase activity was measured following the procedure described by Leech et al. (1985), while AI, NI, NR, SOD, CAT, and POD activities were assessed using Zou’s methodologies (Zou, 2000); root vitality was determined using the triphenyltetrazolium chloride (TTC) method (Clemensson-Lindell, 1994); lastly, CMP of leaves was determined through electrolyte leakage, as described by Palta and Stadelmann (1997).

The first fully expanded leaf of tobacco seedlings grown in nutrient solutions with different levels of Mg supply was collected at 5, 15, and 25 DAT, respectively, for RNA-Seq analysis. Each treatment consisted of three independent biological replicates, with one seedling per replicate. Total RNA was extracted from finely ground leaf samples using TRIzol reagent (Thermo Fisher Scientific, China). The extracted RNA was treated with RNase-free DNase I (Thermo Fisher Scientific, China) to eliminate genomic DNA contamination. The RNA samples were then subjected to RNA-Seq analysis at Novogene, Beijing, China, using an Illumina Novaseq platform.

The resulting sequencing data was processed and were mapped to the reference genome of tobacco available at https://solgenomics.net/organism/Nicotiana_tabacum/genome, using the HISAT2 v2.0.5 program (Kim et al., 2015). The expression level of each gene was quantified based on the fragments per kilobase of transcript per million mapped reads (FPKM) value (Trapnell et al., 2010). Differential expression analysis was performed using the DESeq2 package in R platform (Love et al., 2014). Genes with a |log2 FC| ≥ 1 and the false discovery rate (FDR) ≤ 0.05 were considered as differentially expressed genes (DEGs).

The identified DEGs were subjected to GO analysis using the GOseq R packages (Young et al., 2010), and to KEGG pathway enrichment analysis using the KEGG Orthology-based Annotation System (KOBAS) software (Xie et al., 2011). GO terms or KEGG pathways with p ≤ 0.05 were considered for further assessment in this study.

Total RNAs were extracted from tobacco leaves using the TransZol Up kit (TransGen Biotech, China) and were treated with RNase-free DNase I (Takara, China) to eliminate any contaminating DNA. First-strand cDNA synthesis was then performed using a HiScript II 1st Strand cDNA Synthesis Kit (Vazyme, China). Quantitative real-time PCR (qRT-PCR) reactions were carried out on a CFX Connect Real-Time System (BIO-RAD, USA) with a SYBR qPCR Master Mix (Vazyme, China). Three replications were conducted for each sample. Internal control tests were conducted with the tobacco EF-1α gene (Schmidt and Delaney, 2010). Relative expression values were calculated using the 2^-ΔΔCT method (Livak and Schmittgen, 2001). The transcriptional profiles of 11 DEGs involved in Mg distribution and antioxidative regulation were validated using qRT-PCR. The specific primers used for the qRT-PCR analysis can be referenced in Supplementary Table S1 .

The open reading fragment (ORF) sequences of the target genes were obtained from the Solanaceae Genomics Network (https://solgenomics.net/ftp/genomes/Nicotiana_tabacum/edwards_et_al_2017/). Subcellular localization constructs for the homologous NtMGT7 (Nitab4.5_0003331g0150) and NtMGT9 (Nitab4.5_0000436g0030), and the homologous NtMHX1 (Nitab4.5_0005805g0040), were created by amplifying and inserting the ORFs of these genes into pCS-NGFP (Qu et al., 2021) to fuse with the GFP gene. The resulting constructs were then introduced into the Agrobacterium tumefaciens strain GV3101. The GV3101 bacteria carrying these GFP-fusion constructs, a mitochondria ScCOX4-DsRed marker construct or a tonoplast Osγ-TIP-DsRed marker construct (Chen et al., 2019) were cultured in liquid yeast extract peptone media supplemented with kanamycin (50 μg/ml) and rifampicin (50 μg/ml). Suspensions of the GV3101 bacteria containing the GFP-fusion constructs were co-infiltrated with the DsRed marker constructs into leaves of 4-week-old Nicotinana benthamiana plants grown in a growth chamber at 25°C under a 16/8 h light/dark cycle. The infiltrated plants were maintained in darkness at 25°C for 3 days. The infiltrated leaf tissues were collected and used to isolate protoplasts. Fluorescence microscopy was conducted on the N. benthamiana protoplasts using a Leica DMi8 Laser Scanning Confocal microscope (STELLARIS 5, Leica, Germany) with Excitation/emission wavelengths 488/535 nm for green fluorescence, and 552/610 nm for red fluorescence.

Tobacco seedlings were grown in modified Hoagland’s nutrient solutions supplied with four different levels of Mg (Mg0, Mg1/4, Mg1, and Mg4, respectively). At 5 DAT, no significant morphological differences were observed among the seedlings (data not shown). However, by 15 DAT, the Mg0 seedlings showed slight growth stunting, and by 25 DAT, they exhibited symptoms of leaf chlorosis and curling ( Figure 1A ). In contrast, the Mg1/4, Mg1, and Mg4 seedlings did not show noticeable morphological differences at either 15 or 25 DAT ( Figure 1A ).

At 25 DAT, the Mg content and eight physiological traits related to photosynthesis, carbon, and nitrogen metabolism in leaves of tobacco seedlings grown under different Mg levels were assessed. The Mg content in the Mg0 seedlings were notably lower compared to that in the Mg1/4, Mg1, and Mg4 seedlings ( Figure 1B ). The contents of three essential photosynthetic pigments (Chla, Chlb, and Car) in the Mg0 seedlings were significantly lower compared to those in the Mg1/4, Mg1, and Mg4 seedlings ( Figure 1C ). Similarly, the content of SP, an important osmoregulatory substance, was notably reduced in the Mg0 seedlings than in the Mg1/4, Mg1, and Mg4 seedlings ( Figure 1D ). In contrast, the activity of RuBPCase, a key C3 enzyme responsible for carbon fixation, was markedly higher in the Mg0 seedlings compared to the Mg1/4, Mg1, and Mg4 seedlings ( Figure 1E ). Furthermore, the activities of AI and NI, both involved in carbon metabolism, were significantly reduced in the Mg0 seedlings compared to the Mg1/4, Mg1, and Mg4 seedlings ( Figures 1F, G ). Similarly, the activity of NR, a crucial enzyme in nitrogen metabolism, was lower in the Mg0 seedlings than in the Mg1/4 and Mg4 seedlings ( Figure 1H ). These results demonstrated that a deficiency of Mg causes severe physiological disorders and inhibits the growth of tobacco seedlings.

To investigate the molecular responses of tobacco plants to different levels of Mg supply, leaves from the Mg0, Mg1/4, Mg1, and Mg4 seedlings were collected at 5, 15, and 25 DAT, respectively, for RNA-Seq analysis. A total of 11,267 DEGs were identified in the Mg0, Mg1/4, and/or Mg4 seedlings compared to the Mg1 seedlings ( Supplementary Table S2 ). While only a small number of DEGs were identified at 5 DAT (41 in Mg0, 38 in Mg1/4, and 13 in Mg4), a significantly higher number of DEGs were identified at 15 DAT (4,357 in Mg0, 910 in Mg1/4, and 482 in Mg4), and 25 DAT (6,376 in Mg0, 303 in Mg1/4, and 2,440 in Mg4), respectively ( Figure 2A ; Supplementary Table S3 ). Heatmap analysis revealed similar transcriptomic profiles for the Mg0, Mg1/4, Mg1, and Mg4 seedlings at 5 DAT, but more distinct differences emerged at 15 and 25 DAT, especially between the Mg0 seedlings and the other groups ( Figure 2B ).

Additionally, an upset analysis was conducted on the DEGs identified in the Mg0, Mg1/4, and/or Mg4 seedlings compared to the Mg1 seedlings to gain deeper insights into the regulation patterns of these DEGs in tobacco seedlings grown under various Mg levels over different growth durations ( Figure 2C ). For example, approximately 2,132 and 4,405 DEGs were specifically regulated in the Mg0 seedlings at 15 and 25 DAT, respectively, while 1,022 DEGs were co-regulated in the Mg0 seedlings at both 15 DAT and 25 DAT. In contrast, only 11 DEGs were specifically regulated in the Mg0 seedlings at 5 DAT, with eight DEGs co-regulated at both 5 DAT and 15 DAT, and five DEGs co-regulated at both 5 DAT and 25 DAT ( Figure 2C ). The results highlight that a significant number of genes were markedly induced in tobacco seedlings in response to Mg deficiency after 15 days of growth, with certain DEGs showing relatively long-term regulation.

Plants have developed mechanisms to transport and distribute Mg to maintain optimal cellular levels (Tang and Luan, 2017). Mg²⁺ transporters (MGTs) play essential roles in Mg uptake, transport and distribution (Yan et al., 2018). Among the DEGs identified, one MGT7 homologous gene (Nitab4.5_0003331g0150) and one MGT9 homologous gene (Nitab4.5_0000436g0030) were found to be down-regulated in the Mg0 seedlings at 25 DAT ( Figure 3A ). Mg²⁺/H⁺ exchanger (MHX) has been shown to facilitate Mg²⁺ influx into the vacuole (Shaul et al., 1999). A homologous MHX1 gene (Nitab4.5_0005805g0040) was also identified as a DEG that was down-regulated in the Mg0 seedlings at 25 DAT ( Figure 3B ). To validate the transcriptome results, qRT-PCR analysis was conducted on the homologous NtMGT7, NtMGT9, and NtMHX1 gene in the Mg0, Mg1/4, Mg1, and Mg4 seedlings at 25 DAT. The results indicated a significant down-regulation of all three genes in the Mg0 seedlings, consistent with the RNA-Seq data ( Figures 3C-E ).

The subcellular localization of MGTs and MHX is closely associated with their roles in the biological processes occurring within respective organelles. To investigate the subcellular localization of the homologous NtMGT7, NtMGT9 and NtMHX1 proteins, transient expression of the NtMGT7-GFP, NtMGT9-GFP, and NtMHX1-GFP fusions was conducted in N. benthamiana cells. The results demonstrated that the green fluorescence of NtMGT7-GFP, and NtMGT9-GFP co-localized with the mitochondria ScCOX4-DsRed marker ( Figure 3F ), indicating that both NtMGT7 and NtMGT9 are localized in the mitochondria. On the other hand, the green fluorescence of NtMHX1-GFP merged with the tonoplast Osγ-TIP-DsRed marker ( Figure 3F ), suggesting that NtMHX1 is localized to the vacuole membrane. These observations provide insights into the specific subcellular locations of NtMGT7, NtMGT9 and NtMHX1 and their respective roles within the cell.

To further understand the molecular mechanisms underlying tobacco’s response to different levels of Mg supply, a KEGG analysis was conducted on the DEGs from the Mg0, Mg1/4, and Mg4 seedlings compared to the Mg1 seedlings. The analysis revealed that both the up- and down-regulated DEGs in the Mg0, Mg1/4, and Mg4 seedlings were enriched in diverse pathways, particularly at 15 and 25 DAT. Significantly, multiple KEGG pathways associated with photosynthesis were prominently highlighted, including photosynthesis, photosynthesis-antenna proteins, and carbon fixation in photosynthetic organisms ( Supplementary Table S4 ).

The KEGG photosynthesis pathway was significantly enriched among the up-regulated DEGs in the Mg0 seedlings at 15 DAT, as well as among the down-regulated DEGs in the Mg0 seedlings and the up-regulated DEGs in the Mg4 seedlings at 25 DAT ( Supplementary Table S4 ). The DEGs encoding proteins related to processes involved in the KEGG photosynthesis pathway, including photosystem II (PSII), photosystem I (PSI), photosynthetic electron transport, and F-type ATPase, were summarized ( Figure 4A ). A total of 45 DEGs related to the photosynthesis pathway were identified, with 32 up-regulated in the Mg0 seedlings at 15 DAT, nine down-regulated in the Mg0 seedlings at 25 DAT, and nine up-regulated in the Mg4 seedlings at 25 DAT ( Figure 4B ).

The KEGG pathway of photosynthesis-antenna proteins was significantly enriched among the up-regulated DEGs in the Mg4 seedlings at 25 DAT ( Supplementary Table S4 ). There were four DEGs related to the light-harvesting chlorophyll protein complex, all of which were up-regulated in the Mg4 seedlings at 25 DAT ( Figures 5A, B ). Additionally, the KEGG pathway of carbon fixation in photosynthetic organisms was significantly enriched among the up-regulated DEGs in the Mg0 seedlings at 15 DAT, and the down-regulated DEGs in the Mg0 seedlings at 25 DAT ( Supplementary Table S4 ). A total of 26 DEGs related to processes (C4-Dicarboxylic acid cycle and reductive pentose phosphate cycle) involved in the carbon fixation in the photosynthetic organisms pathway were identified within this pathway. Of these, 15 DEGs were up-regulated in the Mg0 seedlings at 15 DAT, while 12 were down-regulated in the Mg0 seedlings at 25 DAT ( Figures 6A, B ).

Plant leaves that are deficient in Mg are highly photosensitive, leading to an over-saturation of photosynthetic electron transport system. Under highly reduced condition, electrons could pass-on to O₂, resulting in the generation of O₂ ^- and other reactive oxygen species (ROS) (Grossman and Takahashi, 2001). At 25 DAT, the root vitality (measured by the triphenylmethyl hydrazone (TTF) content), the relative level of CMP, H₂O₂ content, and the activities of the enzymatic scavengers SOD, POD, and CAT in leaves of tobacco seedlings grown under different Mg levels were measured. The root vitality was significantly lower in the Mg0 seedlings compared to that in the Mg1/4, Mg1, and Mg4 seedlings ( Figure 7A ), and the CMP level and H₂O₂ content were notably higher in the Mg0 seedlings compared to the others ( Figures 7B, C ). These results indicate that Mg deficiency cause oxidative stress in tobacco seedlings. As for antioxidant enzymes, the activity of SOD was significantly lower in the Mg0 seedlings compared to the Mg1 and Mg4 seedlings ( Figure 7D ). Similarly, CAT activity was notably lower in the Mg0 seedlings compared to the Mg1/4, Mg1, and Mg4 seedlings ( Figure 7E ). In contrast, POD activity in the Mg0 seedlings were significantly higher than that in the Mg1/4, Mg1, and Mg4 seedlings ( Figure 7F ).

We further characterized the DEGs related to antioxidant responses. Six SOD homologous genes were identified among the DEGs, including three copper/zinc superoxide dismutase genes (CSDs) and three iron superoxide dismutase genes (FSDs) ( Figure 7G ). Surprisingly, despite the lower SOD activity detected in the Mg0 seedlings at 25 DAT, all six SOD homologous genes were found to be up-regulated in the Mg0 seedlings at 15 DAT. Furthermore, one CAT2 homologous gene (Nitab4.5_0000588g0030) was identified as a DEGs that was up-regulated in the Mg1/4 and Mg4 seedlings at 25 DAT ( Figure 7G ). Additionally, 10 POD homologous genes were identified among the DEGs, with the majority (seven out of 10) being up-regulated in the Mg0 seedlings at 25 DAT ( Figure 7G ), consistent with the significantly higher POD activity detected in the Mg0 seedlings at 25 DAT. Moreover, eight peroxiredoxin (Prx) homologous genes were identified among the DEGs, including three peroxiredoxin Q genes (Prx Q), two peroxiredoxin II E genes (Prx II E), and three 2-Cys peroxiredoxin B genes (2-Cys Prx B). All eight Prx homologous genes were up-regulated in the Mg0 seedlings at 15 DAT ( Figure 7G ). The transcriptional profiles of the DEGs NtCSD2 (Nitab4.5_0004871g0010) and NtFSD3 (Nitab4.5_0000071g0190) in the Mg0, Mg1/4, Mg1, and Mg4 seedlings at 15 DAT, and NtPOD25 (Nitab4.5_0003858g0010) and NtPOD52 (Nitab4.5_0002731g0010) in the Mg0, Mg1/4, Mg1, and Mg4 seedlings at 25 DAT were validated using real-time RT-PCR. The results showed a significant up-regulation of these four genes in the Mg0 seedlings compared to the Mg1 seedlings, and the results were in line with the RNA-Seq data ( Figures 7H-K ). In addition to the SOD, CAT2, POD, and Prx homologous genes, a total of 44 glutathione S-transferase (GST) homologous genes were identified among the DEGs. These included three NtGSTU7, 21 NtGSTU8, three NtGSTU9, two NtGSTU10, six NtGSTU19, five NtGSTU25, two NtGSTF8, one NtGSTF11, and one NtGSTL3 ( Figure 8A ). The majority of these GST homologous genes (39 out of 44) were up-regulated in the Mg0 seedlings at 15 and/or 25 DAT. The transcriptional profiles of the DEGs NtGST8 (Nitab4.5_0000008g0130), NtGST8 (Nitab4.5_0017813g0010), NtGST10 (Nitab4.5_0000026g0030), and NtGSTF8 (Nitab4.5_0011583g0010) in the Mg0, Mg1/4, Mg1, and Mg4 seedlings at 25 DAT were validated by real-time RT-PCR. The results showed a significant up-regulation of these four NtGST homologous genes in the Mg0 seedlings, consistent with the RNA-Seq data ( Figures 8B-E ).