Dissection of Molecular Processes and Genetic Architecture Underlying Iron and Zinc Homeostasis for Biofortification: From Model Plants to Common Wheat

The micronutrients iron (Fe) and zinc (Zn) are not only essential for plant survival and proliferation but are crucial for human health. Increasing Fe and Zn levels in edible parts of plants, known as biofortification, is seen a sustainable approach to alleviate micronutrient deficiency in humans. Wheat, as one of the leading staple foods worldwide, is recognized as a prioritized choice for Fe and Zn biofortification. However, to date, limited molecular and physiological mechanisms have been elucidated for Fe and Zn homeostasis in wheat. The expanding molecular understanding of Fe and Zn homeostasis in model plants is providing invaluable resources to biofortify wheat. Recent advancements in NGS (next generation sequencing) technologies coupled with improved wheat genome assembly and high-throughput genotyping platforms have initiated a revolution in resources and approaches for wheat genetic investigations and breeding. Here, we summarize molecular processes and genes involved in Fe and Zn homeostasis in the model plants Arabidopsis and rice, identify their orthologs in the wheat genome, and relate them to known wheat Fe/Zn QTL (quantitative trait locus/loci) based on physical positions. The current study provides the first inventory of the genes regulating grain Fe and Zn homeostasis in wheat, which will benefit gene discovery and breeding, and thereby accelerate the release of Fe- and Zn-enriched wheats.


Introduction
A multitude of trace metals, such as iron (Fe), zinc (Zn), copper (Cu), and manganese (Mn), are essential micronutrients for both plant growth and human health due to their pivotal roles in key metabolic reactions and myriad biochemical functions [1,2]. Among these minerals, Fe and Zn are thought to have vital biological relevance and clinical significance for global public health [3]. In plants, Fe is responsible for regulating photosynthesis, mitochondrial respiration, and protein stability, and is one of the cofactors in electron or oxygen transfer [1,4,5]. Zn is a key cofactor for protein function (e.g., zinc finger proteins), and is required for numerous plant physiological processes such as photosynthesis, signal transduction, and maintenance of CO 2 levels in mesophyll [1,6,7]. With Fe and Zn deficiency, plants frequently show chlorosis and reduced photosynthesis, resulting in significantly reduced yield [8,9]. Millions of people worldwide suffer from micronutrient malnutrition, with, most prevalently, approximately two billion people with Fe and/or Zn deficiencies [10]. About 40% of children and pregnant women globally tend to be impacted by iron-deficiency anemia, which affects cognitive development and decreases immunity function [11,12]. Zn deficiency results in stunted growth and mental retardation in children worldwide [13]. Among the three primary interventions to surmount micronutrient malnutrition, biofortification, which increases the intrinsic nutritional content in edible crop plants, is thought to be the most sustainable method [14].
Wheat (Triticum aestivum L.) is one of the most widely grown and consumed crops, and is therefore recognized as an appropriate target for biofortification [15]. However, most common wheat cultivars have suboptimal concentrations of Fe and Zn [16]. Side-by-side comparisons of high-and low-yielding cultivars of crops in the past few decades have provided strong evidence that there can be trade-offs between the yield and content of some nutrients caused by breeding objectives highly focused on yield rather than other traits, such as micronutrient content [17][18][19][20][21]. Hence, it is imperative for scientists to re-introduce previously neglected current wheat breeding gene pools and increase grain Fe/Zn to an appropriate level. To achieve this, it is essential to identify as many as Fe/Zn-related QTL (quantitative trait locus/loci) or genes for wheat biofortification.
Although limited molecular progress has been made to understand Fe and Zn regulatory networks in wheat, substantial numbers of genes involved in Fe and Zn homeostasis in model plants such as the non-graminaceous plant Arabidopsis thaliana and graminaceous plant rice (Oryza sativa L.) have been extensively studied [22,23]. Comparative genomics approaches have facilitated the identification of putative orthologous genes as potential candidates for wheat Fe and Zn homeostasis [24,25]. This approach became particularly attractive when the fully annotated wheat reference genome of Chinese Spring RefSeq v.1.0 was released [26]. Concurrent with rapidly evolving high-throughput and cost-effective genotyping systems (e.g., array-based and sequencing-based platforms), a large number of Fe and Zn loci have been excavated through QTL mapping and association analysis [27,28]. Moreover, high-throughput phenotyping technologies, such as inductively coupled plasma-mass spectrometry (ICP-MS), X-ray fluorescence spectrometry, and newly emerged synchrotron-based X-ray fluorescence microscopy, are also making rapid advances, complementing deep genetic analysis and increasing the efficiency of Fe/Zn gene discovery in hexaploid wheat [29,30].
In this review, we will firstly recapitulate extant molecular processes in Fe and Zn homeostasis in the model plants Arabidopsis and rice, and summarize the genes implicated in Fe and Zn modulation. Secondly, sequences of the above genes were retrieved in BLAST analysis to obtain orthologous genes in wheat; these genes and previously located QTL for wheat Fe and Zn metabolism were jointly assembled on wheat chromosomes and stable genetic loci were identified based on QTL-rich clusters (QRCs). Thirdly, different strategies for wheat Fe and Zn biofortification are proposed by leveraging genetic basis.

Molecular Mechanisms Regulating Fe and Zn Homeostasis in Model Plants
Despite high abundance in the crust of the earth, Fe is less accessible to plants as it primarily exists as insoluble ferric oxides, especially in aerobic conditions at neutral or alkaline pH [31]. In response to the limited Fe availability in soil and potential Fe deficiency, plants have developed delicate mechanisms to tightly regulate Fe homeostasis, including Fe acquisition, translocation, and storage. Zn is also a crucial element and serves as a cofactor for over 1200 proteins in plants [32]. Zn absorption is a complex physiological trait mainly governed by Zn transporters and metal chelators of plant systems. Notably, previous literature has shown that many transporters, chelators, and key regulatory genes involved in Zn homeostasis simultaneously took part in Fe regulation in plants [33][34][35][36]. To gain a better understanding of Fe and Zn homeostasis in wheat, we recapitulate the extant knowledge of the molecular mechanisms modulating Fe and Zn homeostasis in the model plants Arabidopsis and rice. About 172 genes, of which Arabidopsis and rice have 110 and 62, respectively, are summarized in Table S1. These provided useful information for the identification of orthologous genes and the dissection of molecular mechanisms underlying Fe and Zn homeostasis in wheat.

Xylem and Phloem Transport of Fe and Zn
Once Fe and Zn are loaded into the xylem, they can be transferred to phloem for further long-distance transport and remobilization towards the sink organs. In the acidic xylem (pH = 5.0-6.5), Fe is transported mainly as metal chelates and Zn is in the form of both free ions and metal chelates [62]. Although detailed mechanisms of Fe/Zn xylem-to-phloem transport remain largely unknown, their long-distance transport and remobilization at the source tissues through the phloem are partly revealed. In Arabidopsis, oligo peptide transporter 3 (OPT3) contributes to regulating the transportation of a shoot-to-root Fe signal, and redistributing Fe to developing tissues through the phloem [63,64]. In rice, the predominant Fe chelates in the phloem are Fe (II)-NA and Fe (II)-DMA [65]. OsYSL2 may be related to Fe (II)-NA translocation through the phloem [66,67]. OsYSL18 and OsYSL16 are responsible for Fe (II)-DMA translocation via the phloem [68][69][70]. OsYSL9 transports both Fe (II)-NA and Fe (II)-DMA in the rice phloem sap [53]. In graminaceous plants, stem nodes are of importance as a hub for the distribution of Zn and possibly Fe from xylem to phloem to different above-ground organs [71,72]. OsZIP3/OsHMA2 and YSL family transporters/OsHMA9 are responsible for Zn xylem-to-phloem transfer and phloem-to-organs long-distance transport and remobilization in rice, respectively [59,60,[72][73][74].

Storage of Fe and Zn in Seeds
After long-distance transport and remobilization, the terminal destination of Fe and Zn in major crops is thought to be seeds. Plants use two major approaches for Fe storage; that is, sequestration into vacuoles and sequestration into ferritin, mainly in the form of ferrous Fe complexes [5]. The proportion of total Fe stored in vacuoles and ferritin differs among plant species [75]. Approximately 50% of the total Fe in Arabidopsis seeds is present in vacuolar globoids in endodermal cells, whereas less than 5% is deposited as ferritin [76,77]. Meanwhile, seeds of legumes such as soybean, chick pea, and lentil are especially rich in ferritin. In the grains of cereals such as rice and wheat, Fe is mainly present in vacuoles in the aleurone layer and the seed coat.
Fe-NA complexes are transported into Arabidopsis seeds by YSL1 and YSL3 transporters [78,79]. The storage of Fe in seed vacuoles is mediated by vacuolar iron transporter 1 (VIT1) [80]. NRAMP3 and NRAMP4 contribute to the export of Fe from the vacuole to the cytosol, an essential step for seed germination [81,82]. VIT family genes, including OsVIT1 and OsVIT2, similarly play a crucial role in Fe sequestration into rice vacuoles [83,84], and NRAMPs are responsible for Fe transport from the vacuoles [85].
Fe and Zn in wheat are concentrated in small vacuoles in the aleurone layer and sub-aleurone layer, where they localize with phosphorus, most likely in the form of phytic acid (InsP 6 ) (PA). Some Fe and Zn are also located in the starchy endosperm [86,87]. Fe and Zn, as complexes with PA, can hardly be assimilated by humans [88]. Hence, a feasible way for wheat biofortification is to simultaneously increase Fe/Zn contents and their bioavailability (e.g., decrease the PA content or increase phytase activity) [87,89].

Transcriptional Regulation and Other Factors
An efficient modulation network that senses the status of Fe and Zn in plants, and alters homeostasis accordingly, is vital to ensure that sufficient Fe and Zn reach the appropriate target tissues. Over the past 15 years, substantial numbers of transcription factors (TFs) involved in Fe homeostasis were identified in Arabidopsis [90]. FER-like iron deficiency-induced transcription factor (FIT) is the core transcriptional regulator of Fe homeostasis [91,92]. FIT is expressed only in roots and it positively regulates key Fe uptake genes [93]. Various proteins have been identified as interacting directly or indirectly with FIT and an overview of the FIT-regulated network in Fe homeostasis was systematically reviewed by Wu and Ling [94]. Another bHLH TF, POPEYE (PYE), as a negative regulator of NAS4 and FRO3 in Arabidopsis, also plays an important role in Fe homeostasis independently of FIT [95]. Notably, 17 of 133 bHLH TFs present in Arabidopsis have been reported to regulate Fe homeostasis [90,96]. MYB family TFs MYB10 and MYB72 participate in regulating Fe acquisition and distribution [97][98][99]. WRK46 and two MYB-CC TFs, phosphate starvation response 1 (PHR1) and PHR1-like (PHL1), were implicated in Fe translocation and storage, respectively [100,101]. In addition, a small gene family of hemerythrin E3 ubiquitin ligases, such as BRUTUS (BTS), is suggested to sense Fe signals and act as negative regulators of Fe homeostasis [102]. Recently, BTS-like 1 (BTSL1) and BTS-like 2 (BTSL2) were reported to negatively regulate Fe uptake by targeting FIT in Arabidopsis [103]. A novel peptide family termed iron man (IMA) or Fe-uptake-inducing peptide (FEP) positively regulates Fe-deficiency responses in Arabidopsis [104,105].
OsbHLH156 was recently reported to partner with IRO 2 by localizing IRO 2 to the nucleus [117,118].
Several Fe-related transporters and TFs in plants function redundantly and may fine-tune the homeostasis [5]. Some TFs involved in Fe homeostasis can be controlled epigenetically. For instance, SHK1-binding protein 1 (SKB1) catalyzes the symmetric dimethylation of histone H4R3 (H4R3sme2), which in turn suppresses expression of bHLH38, 39, 100, and 101 [119]. Meanwhile, thus far, only a few TFs regulating Zn homeostasis have been reported in Arabidopsis and rice. Arabidopsis TFs bZIP19 and bZIP23 regulate several ZIP transporters [120,121]. NA and PS synthesis genes and their regulators involved in Fe homeostasis are also thought to function in Zn homeostasis, due to the important role of NA and PSs in Fe and Zn uptake and transport in plants [23].
Apart from TFs, large numbers of plant hormones, small signaling molecules, kinases, and other factors also participate in Fe-deficiency response and Fe homeostasis. Various proteins responsible for the transduction of phytohormones (e.g., ethylene) and signaling molecules (e.g., NO) are able to interplay with FIT/FIT-binding complexes and regulate their activities [122,123]. Ethylene, gibberellin (GA), auxin, and NO positively affect the activity of FIT, whereas cytokinins, abscisic acid (ABA), jasmonic (JA), brassinosteroids (BRs), and H 2 O 2 suppress Fe-deficiency responses [122,123]. It is worth mentioning that many signal pathways involve cross-talk, and therefore phytohormones and signaling molecules probably influence Fe homeostasis through more than one pathway [124,125]. Recently, Ca 2+ , as a second messenger to transduce environmental cues, was implicated in Fe homeostasis, with calcium B-like interacting protein kinases 23 (CIPK23) and CIPK11 positively regulating Fe-deficiency responses [126,127]. Moreover, FIT-binding protein (FBP) is involved in the balance of Fe and Zn homeostasis in Arabidopsis [128]. In addition, the circadian clock regulator Mediator 16 (MED16) and MED25 modulate Fe uptake by either directly or indirectly interplaying with FIT [129,130].

Genetic Architecture Underlying Fe and Zn Homeostasis in Wheat
Recent advances in sequencing technologies have enabled the development of improved wheat genome assembly and high-throughput genotyping platforms, significantly accelerating forward genetic approaches for wheat gene discovery [131]. Considerable numbers of Fe/Zn-related loci were identified by QTL mapping and genome-wide association studies (GWASs) [132,133]. Simultaneously, reverse genetic approaches such as comparative genomics promote Fe/Zn gene discovery, and some wheat orthologous genes have been cloned recently [134,135]. Here, we used all curated Fe/Zn-related genes from Arabidopsis and rice for the identification of wheat orthologs. We summarized the known Fe/Zn-related QTL or genes in wheat and assembled them on wheat chromosomes according to their physical locations. Stable genetic loci for Fe and Zn homeostasis were then identified based on QRCs.

Identification of Orthologous Wheat Genes
In total, 172 Fe/Zn-related gene sequences from Arabidopsis and rice were already used in BLAST analysis against the wheat database in EnsemblPlants and 436 wheat orthologous genes were identified (Table S2). Due to the hexaploid nature of common wheat with the A, B, and D sub-genomes, most query genes from model plants have three copies in wheat. Some query genes, mainly from the FRO, YSL, FER, NARMP, and ZIP gene families have similar gene sequences and correspond to the same wheat genes. After removal of duplications, we finally selected 254 genes with identities greater than 60% for further analysis (Table S3). Orthologous genes were identified on all 21 wheat chromosomes ( Figure 1). Among these genes, 82, 86, and 86 were located in the wheat A, B, and D sub-genomes, respectively. Homologous group 2 chromosomes (2A, 2B, and 2D) had the maximum number of genes (58,~23%), followed by group 5 (39,~15%) and 3 (36,~14%), and the minimum number was in group 6 chromosomes (28,~11%). At the individual chromosome level, 2B had the greatest number of genes (21,~8%), indicating its significance for Fe/Zn homeostasis. In total, 172 Fe/Zn-related gene sequences from Arabidopsis and rice were already used in BLAST analysis against the wheat database in EnsemblPlants and 436 wheat orthologous genes were identified (Table S2). Due to the hexaploid nature of common wheat with the A, B, and D sub-genomes, most query genes from model plants have three copies in wheat. Some query genes, mainly from the FRO, YSL, FER, NARMP, and ZIP gene families have similar gene sequences and correspond to the same wheat genes. After removal of duplications, we finally selected 254 genes with identities greater than 60% for further analysis (Table S3). Orthologous genes were identified on all 21 wheat chromosomes ( Figure 1). Among these genes, 82, 86, and 86 were located in the wheat A, B, and D sub-genomes, respectively. Homologous group 2 chromosomes (2A, 2B, and 2D) had the maximum number of genes (58, ~23%), followed by group 5 (39, ~15%) and 3 (36, ~14%), and the minimum number was in group 6 chromosomes (28, ~11%). At the individual chromosome level, 2B had the greatest number of genes (21, ~8%), indicating its significance for Fe/Zn homeostasis.

Cloned Genes for Fe and Zn Homeostasis in Wheat
Prior to the release of the Chinese Spring reference genome sequence, it was relatively difficult to isolate causal genes by map-based cloning in wheat due to its hexaploidy, enormous genome size (~16 Gb), and abundance of repetitive sequences (>85%) [26]. The first Fe/Zn-related gene isolated was Gpc-B1 derived from T. turgidum ssp. diccocoides and characterized as NAM-B1, a NAC family TF. It accelerates senescence and increases remobilization of Fe, Zn, and nitrogen compounds from source organs to the grain [136][137][138]. The frequency of the functional allele is rare in modern Australian cultivars [139] and zero in Chinese wheats [140]. Introgression of the functional Gpc-B1 allele improved wheat grain Fe and Zn content, but tended to cause some undesirable traits, such as senescence, reduced panicle number, and low grain weight [137,[141][142][143][144].
Comparative genomics provides an efficient way to identify and clone orthologous genes with conserved functions [145]. Six genes representing 14 loci have been characterized to affect wheat grain Fe/Zn (Table S4; Figure 1). By homologous cloning, two VIT paralogs, TaVIT1 and TaVIT2, were identified in wheat, and occur as three copies on wheat group 2 and 5 chromosomes, respectively [146]. TaVIT2 is responsible for vacuolar Fe transport in the endosperm and was confirmed to be effective in wheat biofortification [146]. Cell number regulator 2 (TaCNR2) and TaCNR5 identified in wheat contain similar conserved domains as plant cadmium resistance proteins (PCR) [134,135]. Furthermore, TaCNR2 and TaCNR5 were characterized to affect heavy metal transport and accumulation in plant seeds [134,135]. In addition, wheat ABCC transporter TaABCC13 and inositol pentakisphosphate kinase (TaIPK1), with copies in each sub-genome, were reported to be involved in PA biosynthesis [147][148][149]. RNAi-mediated downregulation of TaABCC13 caused reduced PA without significantly affecting Fe and Zn concentrations in wheat grains, whereas silencing of TaIPK1 significantly decreased PA levels while enhancing grain Fe and Zn contents [148,149].
Notably, "Green Revolution" genes Rht-B1b and Rht-D1b not only reduced plant height and influence grain yield, but also had negative effects on grain micronutrient concentrations in wheat [17,19,[150][151][152][153][154]. Additionally, the photoperiod insensitivity alleles (Ppd1) and possibly flowering time regulatory alleles (Vrn1) were associated with wheat micronutrient concentration, indicating that the development timing also affects micronutrient traits (Table S4) [155]. Doubtlessly, more Fe/Zn-related genes are yet to be cloned and their allelic variations affecting gene expression need to be determined. The development of gene-based markers will greatly benefit marker-assisted selection of Fe/Zn-enriched wheat [27,156].

Genetic Loci Detected by Linkage Mapping and Association Mapping in Wheat
About 120 Fe/Zn-related QTL have been mapped on most wheat chromosomes (Table S5) [132,141,[157][158][159][160][161][162][163][164][165][166][167][168][169][170][171][172][173]. These are shown in Figure 1, using the physical locations of the closest linked markers or midpoints of the respective genetic intervals. In addition, more than 20 relatively stable genetic loci for Fe or/and Zn have been reported from GWASs (Table S5; Figure 1) [133,[174][175][176][177][178][179]. Among them, three Fe/Zn loci match well with those reported from QTL linkage mapping, indicating a degree of reliability. When all the genetic loci detected by QTL mapping and GWASs were placed in Figure 1, the distribution patterns at the chromosome level were different. Among the three wheat sub-genomes, B had the largest number (67) of Fe/Zn-related QTL, followed by A (49), and D (24). Group 2 chromosomes had the largest number (26), whereas group 6 chromosomes had the least (11). These findings were partially consistent with the distribution patterns of orthologous genes as mentioned above, and again indicated that group 2 chromosomes carried multiple genes affecting Fe/Zn homeostasis, whereas the contributions by group 6 chromosomes were relatively minor. It was also not unexpected to find 12 pleiotropic QTL associated with Fe and Zn because the same transporters and chelators, such as IRT1 and NA, are shared for both Fe and Zn homeostasis.

Reliable Genetic Loci for Fe and Zn Homeostasis Based on QTL-Rich Clusters (QRCs)
Fe and Zn homeostasis in plants are complex traits controlled by a great number of QTL and characterized by large effect of genotype × environment interactions [180]. These environmental factors mainly include cross-talk between Fe/Zn and other cations, including carbon (C), nitrogen (N), phosphorus (P), and potassium (K); different environmental conditions, such as soil type, water, and future climatic changes; and field management, such as foliar or soil Fe/Zn fertilization [31,181]. Rational N fertilization increases grain Zn content by N-Zn synergism, whereas P input generally decreases Zn by P-Zn antagonism [181]. Increasing Zn supply to wheat by soil or foliar fertilization can effectively improve Zn concentrations in grains [182]. Considering the large impact of environment factors and genotype × environment interactions on Fe/Zn phenotypes, it is appropriate to conduct multi-environment experiments in fields with sufficient Fe/Zn fertilization for mining Fe/Zn genetic loci. Near isogenic lines (NILs) are ideal materials to reduce confounded interactions of genotype × environment and to investigate the reliable loci [137].
The inheritance of Fe/Zn traits is frequently disrupted by epistatic interactions among multiple QTL/genes [168]. However, the mapping of genetic loci for Fe/Zn traits to the same or close chromosome regions in different studies is indicative of true and stable genetic effects. To identify such regions, we defined QRCs as the regions of <10 Mb including two or more QTL for the same trait from different studies [183]. We investigated distributions of Fe/Zn QTL across chromosomes and identified 14 QRCs based on the above criteria (Table 1; Figure 1). Only four QRCs were specific for Fe or Zn, and other ten were related to both Fe and Zn, indicating that these QRCs may harbor the same or very closely linked candidate genes for Fe and Zn homeostasis. Considering the 12 QTL associated with both Fe and Zn described above, further fine mapping and isolation of functional genes located in these QRCs will benefit the understanding of molecular mechanisms underlying Fe/Zn traits in wheat. The chromosomal distributions of QRCs show that five are located in sub-genome A, and seven are located in sub-genome B, whereas only two are in D. Wheat sub-genome D tends to harbor limited molecular markers so that Fe/Zn loci are identified much less often than those in sub-genomes A and B. However, surprisingly, wheat orthologs of genes for Fe/Zn are found to be abundant in the D sub-genome (86 out of 254), suggesting potential to exploit Fe/Zn genes on the D sub-genome and to further investigate their polymorphisms. In comparison with other agronomic traits, such as yield or disease resistance, the number of studies on Fe and Zn is much smaller and identification of related QRCs is relatively limited [183,184]. QGZn.iari-7B, QGFe.cau-7B, QGFe.huj-7B GFe, GZn a Intervals of QTL-rich clusters (QRCs) were defined according to IWGSC RefSeq v1.0. b Some loci were designated in this study because they were not named in the references. Cloned genes in QRCs are underlined. c "GFe" and "GZn" represent grain Fe concentration and grain Zn concentration, respectively.
Many wheat orthologs (41,~15%) identified in this study were closely (<10 Mb) linked with Fe/Zn QTL discovered in the linkage and association mapping (Figure 1), indicating their probable candidate gene roles. Once validated, these genes could be immediately implemented in breeding for biofortification. Orthologous genes not linked with known QTL are probably monomorphic in most wheat cultivars, or a QTL has not yet been identified in those regions ( Figure 1). Conversely, the many individual QTL found in the regions without known candidate genes might represent unique genes in wheat. Isolation of these genes and dissection of their molecular mechanisms will be extremely important for wheat biofortification. Using GWAS and QTL mapping approaches, we recently identified 20 new Fe/Zn loci, mostly from Chinese wheat genotypes, and seven of them were very close to candidate orthologs. One candidate gene on chromosome 3AL, as an ortholog of OsPEZ, an Fe uptake-related gene, is worthy of further validation. We also identified 15 loci that were probably the same as the known QTL, and at least eight have gene candidates mapped as shown in Figure 1 (our unpublished data). Thus, the current study provides a roadmap for the continuous discovery of wheat Fe/Zn-related genes.

Genetic Biofortification to Enrich Grain Fe and Zn in Wheat
Conventional breeding has attained few successes in enriching Fe and Zn in wheat grain. Biofortification harnessing molecular biotechnology and genetic engineering promises a multi-dimensional avenue to enhance the required micronutrients in the grains [185][186][187]. Using orthologs of Fe/Zn homeostasis genes from model plants and altering their expression in staple crops have offered valuable insights into genetic biofortification. Nevertheless, in comparison to other staple crops (e.g., rice and maize), the practice has appeared to be less to develop Fe/Zn-rich grains in wheat using transgenic and biotechnology approaches [87,146,149,[188][189][190]. Here, we emphasize different research paths to leverage genetic architecture for Fe/Zn, which enable fast release of Fe/Zn-biofortified wheats

Transgene and Gene Editing Approaches
The 172 genes from Arabidopsis and rice for Fe/Zn homeostasis are mainly involved in Fe/Zn uptake, translocation, storage, and regulation. In the process of Fe/Zn uptake, ferric reductases, proton ATPases, and NA/DMA synthases play key roles, and the corresponding candidate genes are to be proved and used for increasing plant Fe/Zn contents. For Fe/Zn translocation, many key transporters, such as FRD3, FPN1, the ZIP family, and the YSL family, are predominantly responsible for homeostatic step. In wheat, Fe and Zn are mainly deposited in the aleurone layer, whereas some Fe and Zn are also located in the starchy endosperm [86,87]. Hence, whole-grain flour has a higher amount of Fe and Zn than white flour, but improved Fe and Zn could also be achieved in white flour by using key genes/proteins, including ferritin, VIT, and the NRAMP family, which are expressed in endosperm. Alternatively, specific TFs and other Fe/Zn regulation factors can also be considered as potential target genes for biofortification as long as they confer no or little yield penalty. (Figure 2).

Transgene and Gene Editing Approaches
The 172 genes from Arabidopsis and rice for Fe/Zn homeostasis are mainly involved in Fe/Zn uptake, translocation, storage, and regulation. In the process of Fe/Zn uptake, ferric reductases, proton ATPases, and NA/DMA synthases play key roles, and the corresponding candidate genes are to be proved and used for increasing plant Fe/Zn contents. For Fe/Zn translocation, many key transporters, such as FRD3, FPN1, the ZIP family, and the YSL family, are predominantly responsible for homeostatic step. In wheat, Fe and Zn are mainly deposited in the aleurone layer, whereas some Fe and Zn are also located in the starchy endosperm [86,87]. Hence, whole-grain flour has a higher amount of Fe and Zn than white flour, but improved Fe and Zn could also be achieved in white flour by using key genes/proteins, including ferritin, VIT, and the NRAMP family, which are expressed in endosperm. Alternatively, specific TFs and other Fe/Zn regulation factors can also be considered as potential target genes for biofortification as long as they confer no or little yield penalty.
In principle, any of the above proteins/genes can be a target for biofortification. The transformation of individual or combined key Fe/Zn-related genes from model plants to wheat is demonstrated to be an effective strategy for Fe/Zn biofortification in wheat. A successful example was overexpression of OsNAS2 or/and PvFERRITIN in wheat, which has seen significantly higher Fe concentration of grains compared with the control [87,190]. Another case was that enriched grain Fe content was achieved by overexpression of wheat FERRITIN and TaVIT2 [146,189]. In principle, any of the above proteins/genes can be a target for biofortification. The transformation of individual or combined key Fe/Zn-related genes from model plants to wheat is demonstrated to be an effective strategy for Fe/Zn biofortification in wheat. A successful example was overexpression of OsNAS2 or/and PvFERRITIN in wheat, which has seen significantly higher Fe concentration of grains compared with the control [87,190]. Another case was that enriched grain Fe content was achieved by overexpression of wheat FERRITIN and TaVIT2 [146,189].
Other than the transgene approach, gene editing could be a more powerful tool to make predicted changes in target genes by deactivating, generating functional alleles, replacing mutant alleles, or creating site-specific transgene integration [191]. Despite a wide range of applications (e.g., yield traits and pre-harvest sprouting resistance) already reported in wheat genome editing, there is no report for editing in Fe/Zn-related genes [192,193]. Gene editing tools, especially the emerging plant prime editing (PPE), enable precise gene modification, including single base substitution and multiple base insertion/deletion, and thus are able to offer potential for wheat Fe/Zn biofortification [194,195]. Gpc-B1 could serve as a starting point for candidate gene editing since removal of 1-bp in the exon in certain wheat genotypes can recover the expected function of NAC, as higher content of Fe/Zn was detected in grains.

Marker-Assisted Selection (MAS) and Genomic Selection
Although transgene and gene editing approaches can effectively and precisely increase intrinsic grain Fe/Zn content, their applications are seriously restricted by public acceptance and government regulations [196]. Marker-assisted selection is dependent on the inheritance of a high-Fe/Zn phenotype, together with a particular genetic marker, and has been widely used in wheat breeding. In the previous section, we summarized, in total, 140 wheat Fe/Zn-related genetic loci detected by linkage mapping and association mapping, as well as 14 QRCs defined as stable and reliable Fe/Zn-related clusters. Close linkage of markers for these genetic loci can be further confirmed irrespective of whether marker-trait associations were outcomes of bi-parental hybrid populations or large natural populations. Thereafter, the confirmed molecular markers, or ideally functional markers for causal genes, will be used in marker-assisted breeding for high Fe and Zn. Gpc-B1, as an example, has been exploited for wheat grain Fe/Zn enhancement following marker-assisted breeding. Interestingly, most lines carrying Gpc-B1 showed significantly higher Fe and Zn contents [197]. Instead of incorporating only significant markers into a selection model in MAS breeding, genomic selection (GS) uses genomically estimated breeding values (GEBVs) developed by all markers covering the entire genome. Similar to MAS, GS can be carried out at any location, including non-target environments so that the breeding process can be greatly accelerated at lower cost [198]. GS could serve as a potential method for Fe/Zn biofortification but one important factor affecting GS accuracy is whether the markers of key QTL/genes are incorporated in the model as a fixed effect [199]. Therefore, a continuing effort in Fe/Zn gene discovery and marker development will not only facilitate breeding for biofortification but will also improve the accuracy of molecular tools for future wheat improvement.

Fast Tracking the Use of Orthologous Genes in Wheat Genetics and Breeding
In this review, 254 Fe/Zn-related gene orthologs were identified in the wheat genome. Utilizing genotyping by targeted sequencing (GBTS), a targeted sequence-capture strategy, these genes can be amplified and sequenced for identification of polymorphic sites in large populations with natural variations, followed by verification of associations with wheat Fe/Zn traits [27,200,201]. Orthologous genes closely linked to known QTL can be given priority. Additionally, the orthologous wheat genes can be functionally validated using reverse genetics approaches, such as heterologous expression, silencing, or knockout, and overexpression [202]. It is interesting that some orthologous genes (e.g., TraesCS2D02G373800 and TraesCS2D02G373900 on chromosome 2D) are present as tandem duplications, since a recent study reported that some copy number variations (CNVs) may be important in wheat stress resilience [203]. Polymorphism in CNVs cannot be detected using SNP chips, thus these genes might not be identified despite potential regulatory functions in quantitative traits. GBTS or gene expression profiling could reveal such polymorphisms.
The described wheat Fe/Zn orthologous genes are potentially related to Fe and Zn homeostasis, and provide valuable resources for marker-assisted selection and/or for gene manipulation. The increasing identification/characterization of Fe/Zn-related genes will promote the elucidation of regulatory mechanisms underlying Fe/Zn homeostasis of wheat. Moreover, the discovery of potential genes with pleiotropic effects on improving both nutrients and yield is of great significance for breeding more nutritious wheats with high yield, and potentially higher economic returns for farmers.

Conclusions and Future Prospects
The pace of genetic and genomics research has accelerated with the availability of increasing amounts of genomics information for common wheat. The current review exhibits the molecular processes and genetic landscape of Fe and Zn traits through referential and comparative molecular mechanisms from model plants, provides the first inventory genes or loci regulating wheat grain Fe and Zn homeostasis from diverse approaches, and identifies reliable genetic loci based on the fully annotated wheat reference genome. Molecular marker technology will be an efficient tool for breeding Fe/Zn-biofortified cultivars. Allelic variations of substantial numbers of Fe and Zn candidate genes need to be detected, and the development and establishment of high-throughput and cost-effective genotyping platforms (e.g., GBTS genotyping platform) are of great importance for practical high-Fe and -Zn breeding. The screening and pyramiding of superior allelic variants will be the next stage of development for wheat Fe/Zn biofortification. However, it is unsuitable to pyramid too many high-Fe and -Zn genes simultaneously, since the latent metal toxicity is detrimental to wheat growth. In model plants, a large number of transporters, TFs, plant hormones, small signaling molecules, kinases, and other factors were revealed to be involved in Fe/Zn homeostasis. The understanding of the above factors and regulatory mechanisms is providing resources to elucidate the molecular regulatory networks in wheat with the current limited molecular progress. Multi-omics methods (e.g., genomics, metabolomics, transcriptomics, and bioinformatics) and emerging biotechnologies will promote the deciphering of the mechanisms regulating Fe/Zn uptake, translocation, and storage in wheat, contributing to the application of transgene and gene editing approaches in breeding programs.