The genomic sequence, protein sequence and genomic annotation file of the foxtail millet material ‘xiaomi’ were obtained from the Multi-omics Database for Setaria italica (MDSi, http://foxtail-millet.biocloud.net/) (Yang et al., 2020). The genomic information of barley (Hordeum vulgare), sorghum (Sorghum bicolor), maize, rice, Arabidopsis, and potato (Solanum tuberosum) was downloaded from the Ensemble Plants database (http://plants.ensembl.org/index.html). The ZIP proteins in barley were obtained from published papers and were used together with ZIP proteins from rice and Arabidopsis as seed sequences for BLASTP alignment (BLAST+, version 2.12.0) of the protein databases of foxtail millet, sorghum, maize, and potato (Deng et al., 2022). Protein sequences with a similarity greater than 50% and an e-value less than 10^-5 were considered as candidates for ZIP. The conserved domain (PF02535) of all candidate genes was screened using the HMMsearch tool (version 3.3.2) with a threshold of an e-value less than 10^-10, and only those candidate genes containing the complete domain were retained. The putative candidate ZIPs were obtained by manually removing different transcripts of the same gene. Subsequently, all ZIPs were renamed according to their positions on the chromosome, such as SiZIP1, SiZIP2.

The basic information of these ZIP genes and the proteins they encode was extracted from the genomic annotation files. The information such as protein molecular weight and isoelectric point, transmembrane domains and subcellular localization was obtained respectively from the computer pI/MW tool in the ExPASy database (https://web.expasy.org/compute_pi/), the TMHMM Server 2.0 (http://www.cbs.dtu.dk/services/TMHMM/), and Cell-PLoc 2.0 (http://www.csbio.sjtu.edu.cn/bioinf/plant-multi/).

To investigate the expansion pattern of the ZIP gene family in foxtail millet, the Multiple Collinearity Scan toolkit (MCscanX) was employed to analyze the duplication relationship of ZIP genes in foxtail millet. Meanwhile, for the duplicated gene pairs, the ratio of non-synonymous mutations (Ka) to synonymous mutations (Ks), Ka/Ks, was calculated using KaKs_Calculator 3.0. Finally, the SiZIP genes and their duplication relationships on the foxtail millet chromosomes were visualized using the TBtools (version 2.119) (Zhang, 2022; Chen et al., 2023a).

To clarify the relationships among the members of the ZIP gene family in foxtail millet, the protein sequences of ZIP transporters from six species were selected to construct a phylogenetic tree together with SiZIPs, including the monocotyledonous gramineous crops rice, sorghum, maize, barley, and the dicotyledonous plants Arabidopsis and potato. Multiple sequence alignment was performed using MEGA 6.06, and the maximum likelihood method was used to construct the phylogenetic tree. The phylogenetic tree was visualized by Interactive Tree of Life (iTOL v6, https://itol.embl.de/).

To analyze the similarities and differences among different SiZIP transporters, the Motif-based sequence analysis Server (MEME version 5.5.1, http://meme-suite.org) was used to predict the conserved motifs of these SiZIP proteins. The motif width was set from 6-200aa, the number of predicted motifs was 10, and they could be displayed repeatedly. The gene structure information of these SiZIP genes was extracted from the genomic annotation file. MEGA 6.06 was used for multiple sequence alignment of all SiZIP sequences and a phylogenetic tree was constructed based on the maximum likelihood method. TBtools was used to visualize protein conserved domains, gene structure, and phylogenetic trees (Sievers et al., 2011). Meanwhile, Escript 3.0 was used to visualize the results of multiple sequence alignment of SiZIP proteins and analyze the relationship between conserved motifs and transmembrane (TM) regions.

To elucidate the sequence characteristics of the promoter regions of SiZIPs, we used TBtools to extract the sequences 2000 bp upstream of the start codon of these genes for cis-acting element prediction. All sequences were uploaded to the PlantCare website (version 1.0, http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) (Yang et al., 2023). After the prediction was completed, the cis-acting elements in the promoter regions were counted and classified to study the response patterns of SiZIPs.

The spatiotemporal expression profiles of SiZIPs in different tissues during the whole growth period of foxtail millet were obtained through the foxtail millet multi-omics website (MDSi, version 1.0, http://foxtail-millet.biocloud.net/home). These datasets included RNA-seq data of various tissues, including roots, stems, leaves, flowers, and grains of the foxtail millet model material ‘xiaomi’ from the seedling stage to the mature stage. The expression levels of all genes were standardized by TPM values.

To investigate the responses of SiZIPs to Cd, Fe, and K ions, the genomic sequencing material ‘xiaomi’ was selected for seedling experiment and subjected to stresses of Cd, Fe, and K. The detailed steps for seedling planting and Cd treatment at three different concentrations (5 µM, Cd1; 10 µM, Cd2, 30 µM, Cd3) were referred to our previous article (Yang et al., 2023). As for Fe and K stresses, five treatments were set by controlling the concentrations of Fe and K ions in Hoagland’s nutrient solution: normal Fe and K (CK), Fe deficiency (0 µM, LFe), high Fe (600 µM, HFe), low K (0.1 mM, LK) and high K (10 mM, HK). The seedlings and roots were rapidly frozen in liquid nitrogen and sent to PERSONAL GENE TECHNOLOGY Co., Ltd. (Nanjing, China) for transcriptome sequencing. cDNA library construction and transcriptome sequencing were performed on the BGISEQ500 sequencing platform and followed the company’s standard sequencing and analysis procedures (Yang et al., 2023). The sequencing raw data was deposited in the NCBI Sequence Read Archive (SRA) with the accession number PRJNA1146895. Meanwhile, the seedlings and roots after Cd treatments were also used for ionome determination of 9 metal ions including Cd, Fe, Zn, Mn, Cu, Mg, Ca, K, Na and Mo on the Thermo Fisher Scientific iCAP TQ ICPMS/MS platform.

Furthermore, to clarify the expression characteristics of SiZIPs under salt stress, cold stress and drought stress, we added gene expression datasets obtained from some public databases and unpublished transcriptomes, including the transcriptomes of foxtail millet seedlings after treatment with 150 mM NaCl for 0.5 h and 2 h (unpublished), the transcriptomes of foxtail millet seedlings after cold stress for 1 h, 3 h and 6 h (SRA accession: PRJNA343268), and the transcriptomes of foxtail millet seedlings during the day/night after long-term drought (unpublished). The gene expression levels under all these different treatments were standardized using TPM values (Transcripts Per Kilobase of exon model per Million mapped reads).

To explore the sequence variation and expression level variation of SiZIPs in natural populations and further evaluate the role of SiZIPs in the quality improvement of foxtail millet. We analyzed the previously completed genomic resequencing data of 360 foxtail millet genotypes and the transcriptome data of grains at the filling stage (Li et al., 2022), and obtained the sequence variation information and expression data of SiZIPs in 360 genotypes based on manual screening.

Considering the great progress of rice in the study of gene functions and the widespread convergent evolutionary characteristics of gramineous crops, a foxtail millet-rice homology comparison strategy was used to further analyze the functions of SiZIPs. The orthologs of SiZIPs in rice were detected by MCscanX and the phylogenetic tree. On this basis, we conducted an in-depth and detailed literature survey to search for the reported functions of OsZIPs. Additionally, the expression patterns of these genes across different tissues were obtained from the Rice Expression Database (https://rapdb.dna.affrc.go.jp/) (Chen et al., 2023b). By comparing the expression characteristics of ZIP genes in foxtail millet and rice, along with the reported functions of OsZIP genes, we inferred the potential functions of SiZIP genes. In addition, the PPI (protein-protein interaction) networks of SiZIPs were predicted through the STRING database (Search Tool for the Retrieval of Interacting Genes/Proteins, version 12.0, https://cn.string-db.org/).

Through local BLASTP and after eliminating sequences of different transcripts from the same gene as well as those containing incomplete conserved domains, 14, 14, 12, and 12 ZIP genes were identified in foxtail millet, sorghum, maize, potato, respectively. Compared with 13 in rice and 15 in Arabidopsis, the number of ZIPs did not exhibit significant disparities between monocotyledons and dicotyledons, and no conspicuous uneven expansion was discerned among different species.

Based on their locations on the chromosomes, the SiZIPs were designated as SiZIP1 to SiZIP14 and the detailed information was listed in Table 1 . Overall, SiZIPs are relatively short with lengths ranging from 1062 bp for SiZIP3 to 2637 bp for SiZIP8. Except for SiZIP8 with 10 introns, the number of introns in other SiZIPs varies from 1 to 3. The 14 SiZIP proteins range in size from a minimum of 271 amino acids (aa) to a maximum of 474 aa and the corresponding molecular weights range from a minimum of 28678.23 Da to a maximum of 43731.21 Da. As typical transmembrane transporters, all SiZIPs contain varying numbers of TM regions with the number of TMs ranging from 4 to 9. The subcellular localization results based on Loctree 3 and Cell-PLoc indicated that all SiZIPs were localized on the membrane. In general, SiZIPs exhibited rich diversity in terms of gene structure, protein physicochemical properties, and TM regions.

All 14 SiZIPs were unevenly distributed across 7 chromosomes with chromosomes 3, 7, and 9 each having the highest number of 3 genes. Chromosome 6 contains 2 genes and the remaining genes are distributed on chromosomes 2, 4, and 5 ( Supplementary Figure S1 ). Overall, SiZIPs were mainly located at both ends of the chromosomes, which in line with the distribution characteristics of foxtail millet genes on chromosomes. Gene duplication analysis showed that 8 SiZIPs participated in tandem duplication (TD) and segmental duplication (SD) events, forming two TD gene pairs (SiZIP3 & SiZIP4, TD1; SiZIP12 & SiZIP13, TD2) and two SD gene pairs (SiZIP2 & SiZIP5, SD1; SiZIP7 & SiZIP10, SD2). Although the Ka/Ks values of all duplicated gene pairs were less than 1, the TD gene pairs (0.32 and 0.43) had higher Ka/Ks values than the SD gene pairs (0.09 and 0.24). These results indicated that both TD and SD have contributed to the expansion of ZIP gene family in foxtail millet.

To clarify the phylogenetic relationships of the ZIP family in different species, we constructed a phylogenetic tree based on protein sequences for 99 ZIP proteins from seven species (dicots: Arabidopsis and potato; monocots: foxtail millet, sorghum, maize, rice, and barley). The results showed that all ZIP proteins were divided into four major clades, Class I-IV, and all clades contained ZIP transporters from dicots and monocots, indicating that the ZIP gene family was formed before the differentiation of dicots and monocots ( Figure 1 ).

Class I consisted of 27 ZIPs, including 3 SiZIPs of foxtail millet, 3 OsZIPs of rice, 3 ZmZIPs of maize, 3 SbZIPs of sorghum, 10 HvZIPs of barley, 3 AtZIPs and 2 StZIPs of potato, forming three orthologous clusters (OCs, OC1–3) ( Table 2 ). Except for the significant expansion of HvZIPs in Class I, ZIPs of other species were relatively conserved in monocots and dicots. Class II contained 18 ZIPs from different species, including 2 SiZIPs, 2 OsZIPs, 2 ZmZIPs, 3 SbZIPs, 2 HvZIPs, 4 AtZIPs and 3 StZIPs. Although the number of ZIPs of different species in Class II did not show a significant difference, ZIPs of monocots and dicots in Class II were located in different branches. Furthermore, Class II included multiple tandem duplicate gene pairs of monocots and dicots, which confirmed that the ZIP genes in Class II were formed by independent duplication after the differentiation of monocots and dicots (OC 4-1 and OC 4-2). Class III comprised 22 ZIP proteins, including 3 SiZIPs, 3 OsZIPs, 3 ZmZIPs, 3 SbZIPs, 3 HvZIPs, 5 AtZIPs and 2 StZIPs, forming an OC (OC5) containing ZIP proteins from seven species and two OCs (OC6-1 and OC6-2) consisting of five monocotyledonous ZIP proteins. Moreover, the tandem duplication of AtZIPs in Class III also promotes the expansion of ZIP gene family in Arabidopsis. Class IV consisted of 32 ZIPs. Besides including one orthologous gene cluster OC 7 composed of ZIPs of monocots and dicots, it also contained 4 homologous gene clusters (OC 8-1 to OC 8-4) specific to monocots. The remaining ZIPs in Arabidopsis and potato were formed after the divergence of monocots and dicots. Overall, ZIP proteins, apart from those of Class I being highly conserved, displayed a certain level of unequal duplication and expansion in monocots and dicots in other classes, forming multiple ZIP clusters unique to monocots.

To compare the characteristics of different SiZIP genes, we analyzed their gene structures and protein conserved domains ( Figure 2 ). Overall, SiZIP genes within the same class exhibit relative conservation in gene structure. For instance, SiZIP3, SiZIP4, SiZIP9 and SiZIP1 in Class IV contained two introns, while SiZIP2 and SiZIP5 in Class III had three introns. However, some SiZIP genes also displayed some variations, such as SiZIP6 and SiZIP14. Furthermore, the two genes of paralogous gene pairs also showed characteristics of the coexistence of similarity in conservation and diversity in gene structure. For example, the gene structures of SiZIP2 and SiZIP5 (SD1) were similar, while SiZIP7 and SiZIP10 (SD2) showed differences. In terms of the conserved domains, some motifs were conserved in all SiZIP transporters, such as Motif 5, Motif 3 and Motif 6, while some motifs were specific in some SiZIPs. For instance, Motif 9 only exists in SiZIP6 and SiZIP14. Regarding paralogous gene pairs, the conserved domains of SiZIP12 and SiZIP13 (TD2) were basically the same, while SiZIP10 lacked three motifs compared to SiZIP7. In general, SiZIPs maintained a certain level of conservation in both gene structure and protein conserved domains, while also generating novel variations, especially among paralogous gene pairs.

The relationship between the TM regions and conserved motifs of these SiZIP proteins was further compared ( Figure 3 ). The results showed that a large number of conserved motifs highly coincided with the TM regions, such as Motif 5 - TM 2, Motif 4 - TM 4, Motif 6 - TM 7 and Motif 7 - TM 9. Some other motifs contained TM regions and some intra- and extra-membrane structures. For example, Motif 3 contained TM 3 and the intra-membrane part, and Motif 1 contained TM regions TM 5 and TM 6 and the extra-membrane part in the middle. The absence of conserved motifs in SiZIP proteins was directly related to the number of their TM regions.

We predicted and analyzed the cis-acting elements in the promoter regions of all SiZIP genes ( Figure 4 ). The promoter regions of SiZIPs mainly contained five types of cis-acting elements: hormone response, light response, MYB transcription factor binding, stress response and tissue-specific expression. The hormone response elements included abscisic acid response, auxin response, gibberellin response, MeJA response and salicylic acid response elements. The abscisic acid response elements and MeJA response elements were the most abundant, followed by gibberellin response elements and salicylic acid response elements, indicating that these SiZIP genes are regulated by plant growth regulators to varying degrees. Multiple MYB transcription factor binding sites were predicted in the promoter regions of SiZIPs such as SiZIP13 and SiZIP9, suggesting that these genes may be regulated by MYB transcription factors. In addition, the stress response elements included various stresses such as anaerobic induction, drought, low temperature, and salt, confirming that these genes might be involved in the response to various osmotic stresses. Finally, The presence of tissue-specific expression elements, including endosperm-, root-, meristem-, and seed-specific expression elements, was likely a significant factor contributing to the diversity of tissue-specific expression in SiZIPs.

SiZIPs exhibited diverse expression patterns throughout the whole growth period of foxtail millet, demonstrating clear temporal and spatial specificity ( Figure 5 ; Supplementary Figure S2 ). Overall, the expression patterns of SiZIPs could be broadly classified into three categories. The first category was highly expressed during the germination and the subsequent seedling growth, with SiZIP9, SiZIP4 and SiZIP13 specifically expressed in germinating seeds and SiZIP6 expressed in both germinating seeds and seedlings. The second category was primarily expressed in panicles and leaves during the heading and flowering stages, such as SiZIP8 and SiZIP3. The third category was expressed in various tissues during the grain-filling stage, such as SiZIP14 and SiZIP7. Regarding tissue-specific expression, SiZIP10 was specifically expressed in stems, SiZIP5 and SiZIP1 were primarily expressed in leaves, SiZIP12 was expressed in panicles during multiple stages, and SiZIP2, SiZIP7 and SiZIP11 were highly expressed in roots. Furthermore, paralogous gene pairs also exhibited differences in expression patterns. For example, in SD1, SiZIP2 was expressed in multiple tissues including roots, stems and leaves, while its paralog SiZIP5 was primarily expressed in leaves. In TD1, SiZIP3 was primarily expressed in panicles, while its paralog SiZIP4 was specifically expressed in germinating seeds. The diverse temporal and spatial expression patterns exhibited by different members of the SiZIP gene family, along with the emergence of novel expression patterns in newly duplicated genes, contributed significantly to the functional diversification of SiZIPs.

SiZIPs exhibited distinct expression patterns in different tissues in response to varying concentrations of Cd, Fe, and K stress ( Figure 6 ). Except for SiZIP10, SiZIP8 and SiZIP11, which did not respond to Cd, Fe and K treatments, the expression of the remaining 11 SiZIPs was inhibited or induced by different metal ions. As important Zn and Fe transporters, 8 genes (SiZIP2, SiZIP5, SiZIP7, SiZIP3, SiZIP4, SiZIP13, SiZIP1 and SiZIP6) showed significantly upregulated expression under high Fe treatment. Interestingly, some SiZIPs were induced to be expressed by both high Fe and high K simultaneously. While SiZIP12 and SiZIP9 were significantly suppressed under different concentrations of Fe treatment but significantly upregulated under different K treatments. In addition, the expression patterns of duplicated genes under different metal treatments were also diverse. For example, SiZIP7 in SD2 was significantly induced under high Fe and high K treatments, while SiZIP10 did not respond to any metal ions. In TD1, SiZIP3 was expressed in both roots and leaves and was not significantly induced by K, while SiZIP4 was only expressed in roots and was significantly induced and upregulated by K. Overall, the SiZIP genes exhibited diverse responses and expression patterns under different ionic stresses.

To clarify the correlation between SiZIP gene expression and metal ion transport, we measured the contents of 10 metal ions (Cd, K, Fe, Ca, Na, Mg, Mn, Cu, Zn and Mo) in the roots and leaves of foxtail millet seedlings after Cd stresses ( Supplementary Figure S3A ). Among these ions, besides Cd, Fe (318.9-657.6 mg/kg), Ca (345.0-1880.4 mg/kg), K (2249.8-14230.1 mg/kg) and Na (31.5-193.1 mg/kg) all showed certain differences between Cd treatments and CK. The correlation analysis revealed significant correlations between the expression levels of different SiZIPs in various tissues and the contents of these metal ions. Specifically, SiZIP4, SiZIP6, SiZIP7 and SiZIP9 showed significant positive and negative with Fe (at least 0.57) and K (0.81-0.90) content, respectively, which directly correlated with their induced expression under Fe and K stresses ( Supplementary Figure S3B ). SiZIP12, which was induced by K, exhibited a significant negative correlation (0.85) with K content. Although the dataset used for correlation analysis was not large, it suggested to a certain extent that SiZIPs were closely related to the transport and accumulation of different metal ions in foxtail millet, especially Fe.

In addition to analyzing the response of SiZIPs to metal ions, we have also analyzed their response to various abiotic stresses ( Figure 7 ). Except for SiZIP6, all other SiZIPs were induced to up-regulate their expression under salt stress. The weak response of SiZIP6 to Na and K and its strong up-regulation under Fe and Cd stresses suggested that it might be involved in the transport of Fe and Cd in foxtail millet simultaneously. Multiple SiZIPs, including SiZIP3, SiZIP10, SiZIP7, SiZIP14 and SiZIP8, were induced to up-regulate their expression under cold stress, especially in the early stage (1 h). SiZIP13, SiZIP3, SiZIP11, SiZIP5 and SiZIP1 were induced to up-regulate their expression after drought stress. The results indicated that these SiZIPs might enhance the tolerance of seedlings to abiotic stresses by altering the transport and distribution of metal ions in foxtail millet and regulating the osmotic potential inside and outside the membrane under abiotic stresses. Furthermore, The PPI results showed that except for a certain degree of interaction between SiZIP13, SiZIP8, SiZIP11, and SiZIP12, the functions of other SiZIP proteins were relatively independent ( Figure 8 ). However, interactions between SiZIPs and MTPs (Metal Tolerance Protein), potassium channel protein, EXO (Exocyst subunit Exo70 family protein), calumenin protein, other Zn transporters, and MYB transcription factors were discovered. For example, SiZIP8 interacted with ZTP29, MTP1 and calumenin-B, SiZIP13 interacted with MYB4 and EXO70B1. This suggested that SiZIPs might play a role in metal transport in conjunction with other metal transporters and exosome related proteins, and these transporters were also regulated by transcription factors.

A total of 1718 genetic variation loci were identified among 360 foxtail millet genotypes for 14 SiZIP genes, including 1525 SNPs and 193 Indels ( Table 3 ). Among them, 54 variations were non-synonymous SNPs, and 235 variations were located in the promoter regions. To explore the expression variations of SiZIPs in different genotypes, we investigated the expression levels of SiZIPs in grains during the filling stage across 360 genotypes ( Supplementary Figure S4 ) (Li et al., 2022). The expression levels of SiZIP genes showed significant differences among 360 accessions. For example, SiZIP14 was expressed at a low level (TPM < 25) in most foxtail millet genotypes, but up to a high level in a few genotypes (TPM = 80). The results confirmed the extensive sequence and expression variations of SiZIPs in the natural foxtail millet population, suggesting their genetic selection potential in foxtail millet breeding.

Extensive studies have confirmed the existence of convergent selection among gramineae crops (Chen et al., 2022). Based on the analysis of orthologous relationships and expression characteristics between different species, several important yield-related genes have been discovered in foxtail millet and wheat (Yang et al., 2021a; Liu et al., 2022). Rice, as a model crop in the family of gramineae, has undergone intensive research on the functions of multiple OsZIPs, while SiZIPs have not been reported yet. Therefore, we preliminarily predicted the functions of SiZIPs through literature investigation and comparison of the expression patterns of orthologous ZIPs in foxtail millet and rice ( Table 4 ). Out of the 14 SiZIP genes, 12 orthologs have been identified in rice, and 8 of them exhibited similar expression patterns in foxtail millet and rice, such as SiZIP10, SiZIP3 and SiZIP2. SiZIP10 was specifically highly expressed in the stems of foxtail millet, while its ortholog OsZIP3 in rice is reported to be responsible for Zn unloading and distribution in the enlarged vascular bundle xylem. SiZIP3 was specifically highly expressed in the panicles of foxtail millet at multiple stages, and its ortholog OsZIP5 was mainly expressed in the panicles and participates in Zn homeostasis. SiZIP2 was highly expressed in the roots, stems, and panicles of foxtail millet, and its ortholog OsZIP7 has been proven to be highly expressed in the same tissues in rice, responsible for the loading of Zn/Cd in the root xylem and their transfer from the stem node vascular bundles to the grains. The highly conserved expression patterns of ZIP genes in foxtail millet and rice suggested that the functions of these genes in gramineae, at least in foxtail millet and rice, were relatively conserved. In addition to sharing similar tissue expression patterns with their orthologs in rice, some SiZIP genes also expressed in other tissues. For example, SiZIP12, besides highly expressing in the roots for Zn/Fe absorption and transport similar to its ortholog OsIRT1, also exhibited high expression in the panicles, suggesting that it may have evolved new functions in foxtail millet while maintaining its essential roles. Moreover, apart from conserved and partially conserved genes, there were a few SiZIP genes that exhibited completely different expression patterns compared to their orthologs in rice. For instance, SiZIP5 was specifically highly expressed in the leaves of foxtail millet throughout all developmental stages, whereas its ortholog OsZIP12 was primarily expressed in the grains. It suggested that the functions of ZIP12 may differ in foxtail millet and rice.

Based on the homology relationships and expression patterns of ZIPs between foxtail millet, rice, and Arabidopsis, we predicted the functions of most SiZIP genes. Ultimately, we created a preliminary model diagram of the SiZIP genes in foxtail millet, integrating gene function predictions and expression characteristics in response to metal ions ( Figure 9 ). According to the expression levels of the genes in different tissues and previous studies, the functions of the SiZIP genes were mainly classified into four categories: 1) absorption and efflux of metal ions such as Zn, Fe, and Cd in roots; 2) loading, unloading, and distribution of metal ions in stems; 3) intracellular homeostasis and utilization of metal ions in tissues and organs like leaves; 4) accumulation of metal ions in grains. The localization of conserved SiZIP genes in different tissues and cells was also depicted as precisely as possible, such as SiZIP2 (OsZIP7) likely localized in the root stele, and SiZIP7 (OsZIP4) potentially positioned in the stem or phloem of leaf axillary meristems. Additionally, the expression changes of ZIP genes under non-stress conditions of Fe, Cd, K, and Na were also characterized. In addition to having significant responses to different concentrations of Fe and Cd, the SiZIP genes also exhibited general responses to K and Na, which was likely closely related to the complex synergistic and antagonistic relationships between different metal ions and their transporters.