Toward Nutrient-Rich Rice: Biofortification through Mineral Accumulation and Low Phytic Acid Content

Abstract

Nicotianamine synthases regulate the biosynthesis of two mineral chelators key for rice grain biofortification, nicotianamine (NA) and 2′-deoxymugineic acid (DMA). We produced transgenic rice expressing OsNAS2 under the OsRCc3 promoter (RcN2) and mutated OsLpa1 in these (lpa1 RcN2) and the wild-type (lpa1) to enhance essential mineral accumulation in grains while lowering phytic acid (PA) levels. NA and DMA contents were higher in the brown grains of RcN2, lpa1 RcN2, and lpa1 plants. The grains of these lines accumulated increased levels of essential minerals, with lpa1 RcN2 exhibiting the greatest increases. PA-bound iron and zinc levels were lower in lpa1 and lpa1 RcN2 grains, while NA- and DMA-bound iron and zinc levels were higher in RcN2, lpa1 RcN2, and lpa1 grains. Field-grown RcN2 plants showed no significant growth penalties, unlike lpa1 and lpa1 RcN2. Furthermore, lpa1 and lpa1 RcN2 grains were chalky, a property that facilitates rice flour production.

1. Introduction

Rice (Oryza sativa) grains are the primary source of dietary calories and essential micronutrients for over 50% of the human population. This section of humanity mainly includes people from low- and middle-income countries with limited access to other nutrient-rich foods. Although they consume sufficient calories, they suffer from varying levels of malnutrition due to inadequate intake of essential micronutrients, causing “hidden hunger”. This phenomenon is more severe among those who prefer white or polished rice over brown rice, as polishing removes the micronutrient-rich embryo and aleurone layer of the grain. , In fact, polished rice grains can lose up to 97% of their iron (Fe) and 45% of their zinc (Zn) contents compared to brown rice grains. , An average of 372 million children aged six months to five years and 1.2 billion nonpregnant adult women are reported to be malnourished due to a deficiency in at least one of the micronutrients Fe, Zn, or vitamin A, with about 50% of affected individuals residing in Asia. Insufficient intake of micronutrients, including vitamins and minerals, can have severe health consequences, such as birth defects, compromised immunity, poor growth and development, intellectual disabilities, and even death. , Pregnant women and children are the most vulnerable groups, as they require higher mineral and vitamin intake for normal growth and physiological functions. ,

Addressing this global health problem through dietary diversification remains a challenge in developing countries, where various development programs such as vitamin and mineral supplementation campaigns often face shortcomings due to a lack of funds and inadequate social infrastructure. Thus, direct improvement of the nutritional quality of staple food crops through biofortification provides a necessary and effective long-term solution. ,, Raising the micronutrient content of staple food crops is known as biofortification, which can be achieved through agronomic or genetic approaches. − Genetic biofortification, using conventional breeding or various biotechnological methods, has shown promising results in sustainably enhancing micronutrient content. ,, Conventional breeding is limited by the gene pool of the target crop species, whereas biotechnological approaches can incorporate a wider range of genetic resources, including those from other species, to enhance crop micronutrient content more efficiently and in a more targeted manner. Therefore, these approaches are more promising for rice biofortification. − Many micronutrient content–related genes have been deployed for this purpose. For example, the rice NICOTIANAMINE SYNTHASE (NAS) genes OsNAS1, OsNAS2, and OsNAS3 (hereafter NAS1, NAS2, and NAS3) have been used successfully to raise the contents of various mineral elements in rice grains, particularly Fe and Zn. −

NAS enzymes biosynthesize nicotianamine (NA), a metal chelator essential for mineral translocation, distribution, and metal homeostasis in plants. NA can chelate Fe and Zn as well as other minerals such as copper (Cu) and manganese (Mn). − NA is a precursor for the biosynthesis of the phytosiderophore 2′-deoxymugineic acid (DMA), which is catalyzed by nicotianamine aminotransferase (NAAT) and DMA synthase (DMAS). Like NA, DMA can chelate metals such as Fe and Zn, and both NA and DMA are secreted by plant roots. ,, Notably, when greater amounts of NA and DMA are secreted into the rhizosphere, the chelators can bind to more metals and ultimately enhance metal translocation to rice grains. , For instance, activation-tagged mutants of NAS2 and NAS3 with higher expression of either gene accumulate more Fe, Zn, and other essential minerals in their grains, and these minerals have higher bioavailability, along with much higher NA and DMA levels, compared to wild-type (WT) grains. − Simultaneous overexpression of NAS2 and NAS3 resulted in the greatest improvements in Fe, Zn, NA, and DMA contents in brown grains compared to the WT and the single activation-tagged lines. Overexpression of each of the three NAS genes individually under the control of the cauliflower mosaic virus (CaMV) 35S promoter has also been reported to significantly raise Fe and Zn levels in grains. , Importantly, these studies emphasize that NAS2 is more effective than NAS1 or NAS3 for Fe and Zn biofortification of rice grains. , Other studies have also utilized NAS2 in combination with other metal homeostasis–related genes, with NAS2 expression driven by constitutive or tissue-specific promoters significantly enhancing mineral levels in rice grains. For instance, transgenic indica rice plants engineered with NAS2 driven by the 35S promoter and the soybean (Glycine max) Ferritin gene SferH-1 driven by an endosperm-specific promoter accumulated more Fe and Zn in polished grains with little to no yield loss during field trials. Moreover, transgenic indica rice plants with phloem-specific expression of NAS2 combined with endosperm-specific expression of a Ferritin gene from Pisum sativum achieved significantly enhanced Fe and Zn contents in polished grains without any significant yield loss, although these micronutrients reached higher levels when the phloem-specific promoter was replaced by a constitutive promoter. However, NAS2 has not yet been singly the subject of tissue-specific expression to efficiently biofortify rice without noticeable growth penalties.

The OsRCc3 promoter is a tissue-specific promoter that primarily controls gene expression in roots. Transgenic plants overexpressing OsNAC10 under the control of the OsRCc3 promoter showed greater tolerance for drought, salinity, and cold stresses at the vegetative stage. Moreover, these plants exhibited significantly higher grain yield under drought conditions compared to control plants and those constitutively overexpressing OsNAC10 under the control of GOS2 promoter. Although the OsRCc3 promoter mainly drives gene expression in roots, its transcriptional activity was also observed in the phloem cells of nodes and internodes. Transgenic plants expressing the vacuolar arsenic (As) sequestration genes Yeast Cadmium Factor 1 (ScYCF1) from fission yeast (Schizosaccharomyces pombe) or the rice C-type ATP-binding cassette transporter gene OsABCC1 under the control of the OsRCc3 promoter accumulated significantly lower As levels in their grains without any growth penalty. The tissue-specific OsRCc3 promoter is therefore a promising tool for engineering transgenic rice plants with enhanced nutrient accumulation for grain biofortification without compromising overall plant growth, while also limiting the accumulation of toxic elements in grains.

Beyond being consumed as cooked brown or white rice grains, rice is ground to make flour that is well recognized for its versatility in culinary applications. As a gluten-free and hypoallergenic substitute for wheat flour, it has become increasingly popular among individuals with dietary restrictions or sensitivities. , Rice flour–based foods are also gaining popularity among the general population due to changing dietary preferences, mainly in developed countries. Outside of Asia, where rice remains a dietary cornerstone, rice flour–based products have seen robust uptake in Europe and North America, especially in the form of gluten-free breads, pastas, and snacks. This growth is driven by health-conscious consumers and those with gluten sensitivities, as well as ongoing processing innovations that improve the texture and shelf life of rice flour. , Biofortified rice grains enriched with essential minerals will benefit populations who consume rice flour–based foods. However, the hardness of rice grains presents a significant obstacle in producing high-quality flour with desirable characteristics, such as fine particle size and minimal starch damage. By contrast, the endosperm of chalky grains consists of loosely packed starch granules, lowering grain hardness and making them less susceptible to starch damage during milling, thereby facilitating the production of finer flour. Rice flour with lower starch damage (4–6%) and finer particle size (75–100 μm) have been identified as key factors in producing bread with desirable characteristics to enhance consumer acceptance. , The use of additives such as, hydrocolloids and other gelling agents further enhances dough rheology, bread texture and overall consumer acceptability of gluten-free breads made with rice flour.

Chalky grains offer a cost-effective means of producing high-quality rice flour via simple dry milling and do not require the higher capital investment and operational complexity needed for wet milling, jet milling, or cryogenic milling techniques, which are associated with high energy use, wastewater disposal challenges, and additional drying or sterilization steps. Incorporating the chalkiness trait into biofortified grains can thus be a useful approach to producing rice flour with enhanced mineral content. Notably, a rice low phytic acid (lpa) mutant was reported to produce chalky grains. Phytic acid (PA) is an antinutrient that acts as a strong chelator of metal ions, limiting the bioavailability of essential minerals and remaining a significant challenge to effective biofortification. , Therefore, lower PA levels would ensure greater bioavailability of essential minerals in rice grains. For instance, a point mutation in Low phytic acid 1 (OsLpa1, hereafter Lpa1; LOC_Os02g57400), also known as 2-PHOSPHOGLYCERATE KINASE (2-PGK) resulted in a significantly lower PA content in the grains of the mutant, together with greater chalkiness compared to WT grains. Other mutant alleles of this gene, namely, KBNT lpa 1–1, DR1331–2, and Os-lpa-XQZ-1, also result in lower amounts of PA in their grains. − In addition, the mutants KBNT lpa 1–1 and DR1331–2 in the japonica background showed PA levels lower by 28% and 48% relative to their WT, respectively, whereas Os-lpa-XQZ-1, in an indica background, exhibited about a 35% drop in PA content compared to WT grains, along with higher levels of essential minerals such as Fe and Zn. ,

In this study, we implemented a biotechnological approach to produce plants expressing NAS2 in a tissue-specific manner in RcN2, plants lacking Lpa1 function in lpa1, and plants combining RcN2 and lpa1 in lpa1 RcN2. We used the rice tissue-specific promoter OsRCc3 ^35,37 to produce RcN2 plants, and CRISPR/Cas9-mediated gene editing to generate lpa1 plants. Our aim in mutating Lpa1 was to promote grain chalkiness and diminish PA content to facilitate the production of biofortified rice flour with higher mineral content. We grew all plants under field conditions and subjected them to phenotypic inspection for any growth defects while assessing the mineral contents of their polished and brown grains.

2. Materials and Methods

2.1. Plant Materials and Growth Conditions

The rice cultivar Dongjin (Oryza sativa L. ssp. japonica) was used as the WT in this study. RcN2, lpa1, and lpa1 RcN2 plants were generated and used in this study. Plants were grown in a growth chamber at 28 °C/20 °C (day/night) under a 14-h light/10-h dark photoperiod with 70–80% relative humidity. During the summer season, plants were cultivated in a designated paddy field for genetically modified rice at Kyung Hee University, Republic of Korea, under natural conditions. Field-grown plants were used for phenotypic analysis and for measuring the contents of grain minerals, PA, NA, and DMA.

2.2. Generation of RcN2 Plants

The full-length coding sequence of NAS2 was cloned into the pENTR/D-TOPO entry vector, and the correct clone was confirmed by colony PCR and Sanger sequencing. The resulting entry clone was then recombined with a destination vector (pCAMBIA3301 as the vector backbone) containing the OsRCc3 promoter via LR recombination using the Gateway cloning system. The resulting construct was confirmed by colony PCR and Sanger sequencing and introduced into Agrobacterium (Agrobacterium tumefaciens) strain LBA4404. The BAR gene was used as the selection marker in plants. Transgenic rice plants were generated by Agrobacterium-mediated transformation of seed-derived calli. A total of 12 independent transgenic plants were generated and subjected to expression analysis. Two transgenic plants exhibiting the highest NAS2 expression levels were chosen for further study, and progeny from the T₂ generation onward were used in all subsequent experiments.

2.3. Generation of lpa1 and lpa1 RcN2 Plants

To generate lpa1 plants using CRISPR/Cas9, a 20-nucleotide target sequence was selected using the CRISPRdirect web tool (https://crispr.dbcls.jp/). The selected target sequence was cloned into the pOs-sgRNA entry vector and subsequently recombined into the pH-Ubi-cas9–7 destination vector using the Gateway system. , The final construct was confirmed by colony PCR and Sanger sequencing, and introduced into Agrobacterium LBA4404. The hygromycin resistance gene was used as the selection marker in plants. lpa1 RcN2 plants were generated by Agrobacterium-mediated cotransformation of seed-derived calli with a mixture of LBA4404 strains each harboring the RcN2 and lpa1 constructs. A total of nine independent lpa1 plants and 14 independent lpa1 RcN2 plants were generated. Two lpa1 plants exhibiting homozygous mutations, and two lpa1 RcN2 plants exhibiting the highest NAS2 expression levels and homozygous mutations were selected for further study. Progeny from the T₂ generation onward were used in all subsequent experiments.

2.4. RNA Isolation and RT-qPCR Analysis

Total RNA was isolated from the roots of two-week-old WT, transgenic, and NAS2 activation-tagged mutant (2D) seedlings grown in Yoshida solution using a FastPure Universal Plant Total RNA Isolation Kit (Vazyme Biotech, Nanjing, China). Yoshida solution was prepared as described in Lee et al., adjusted to pH 5.7, and replaced every 2 days. First-strand cDNA was synthesized from 1 μg of total RNA using ReverTra Ace qPCR RT Master Mix (Toyobo, Osaka, Japan). qPCR was performed using Prime Q-Master Mix 2X (GeNetBio, Daejeon, South Korea) on a Rotor-Gene Q real-time PCR cycler (Qiagen, Hilden, Germany). The rice UBIQUITIN5 gene (OsUBQ5; LOC_Os01g22490) was used as an internal control. The relative expression levels were calculated using the 2^–ΔΔCT method. Primer sequences used for RT-qPCR are listed in Supplementary Table S1.

2.5. Measurement of PA Content in Brown Grains

PA content was measured using a K-PHYT assay kit (Megazyme Inc., Bray, Ireland), following the manufacturer’s instructions. Briefly, mature brown grains were finely ground and extracted in 0.66 M HCl overnight at room temperature. One milliliter of each extract was centrifuged at 13,000 rpm for 10 min, and the supernatant was used in the colorimetric assay.

2.6. Measurement of NA and DMA Contents in Brown Grains

NA and DMA contents were quantified according to methods described by Callahan et al., and Yamaguchi and Uchida with modifications. Twenty milligrams of ground brown grain powder was extracted in 5 mM ethylenediaminetetraacetic acid (EDTA) (Fujifilm Wako Co., Osaka, Japan) to release NA and DMA from metal complexes. The samples were sonicated (15 min), heated at 80 °C for 15 min, centrifuged at 14,500 rpm for 5 min, and filtered through a PTFE membrane before LC–MS/MS analysis. Quantification was performed using an LC–MS/MS system consisting of an HPLC (Ultimate 3000; Thermo Fisher Scientific, Waltham, MA, USA); a Scherzo SW-C18 column (Imtakt Corp., Kyoto, Japan), and a TSQ Quantum Ultra mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA).

2.7. Measurement of Mineral Contents in Brown and Polished Grains

Brown rice grains were polished using a Kett Pearlest grain polisher (Kett Electric Laboratory, Tokyo, Japan) for 10 s to obtain white rice grains. Brown and polished grains were separately subjected to acid digestion in 1.5 mL of 65% (w/v) nitric acid at 150 °C for 10 h. The concentrations of mineral elements were determined using inductively coupled plasma mass spectrometry (ICP–MS) (7900X; Agilent Technologies).

2.8. Size-Exclusion Chromatography Analysis

Size-exclusion chromatography (SEC) was performed using a Dionex UltiMate 3000 HPLC system (Thermo Fisher Scientific, Bremen, Germany) with a Superdex Peptide HR 10/30 column (300 mm × 10 mm i.d.; GE Healthcare BioScience AB, Sweden). SEC coupled with ICP–MS was conducted using an Agilent Model 7700s ICP–MS with a hydrogen collision cell (Agilent 7700, Tokyo, Japan). The ICP–MS conditions were optimized daily for maximum sensitivity and minimal interference using standard built-in software. The H₂ collision cell mode was employed to exclude interferences. HPLC coupled with electrospray ionization mass spectrometry (ESI–MS) was carried out using an Orbitrap Fusion Lumos mass spectrometer (Thermo Fisher Scientific, Bremen, Germany), operated with SEC chromatography in both positive and negative ion modes, with electrospray voltages set to 4 kV and 2.5 kV, respectively. The ion transfer tube and vaporizer temperature were maintained at 325 and 300 °C, respectively. For SEC–ESI–MS analysis, 100 mg of brown rice grain powder was extracted with 200 μL 5 mM ammonium acetate buffer (pH 6.2) by vortexing for 2 min, followed by centrifugation at 13,200 rpm for 3 min. The resulting supernatant was used for SEC–ICP–MS and SEC–ESI–MS analysis. For SEC–ICP–MS, 40 μL of each supernatant was injected; for SEC–ESI–MS, 25 μL was injected. In both cases, 5 mM ammonium acetate buffer (pH 6.2) was used as the mobile phase at a flow rate of 0.7 mL/min. To improve sensitivity in SEC–ESI–MS, a low-pressure flow divider (PEEK valve) was used to slow the flow to 0.150 mL/min. ESI–MS spectra were acquired over an m/z range of 150–1500 (or 130–1500) with the resolution set to 240,000.

2.9. Scanning Electron Microscopy (SEM)

SEM was conducted to observe starch granules in the endosperm of WT and transgenic grains. Brown grains were sectioned transversely to expose the endosperm and were then coated with gold. Samples were imaged using a LEO supra 55 field emission SEM (Genesis 2000, Carl Zeiss, Germany).

2.10. Statistical Analysis

All analyses were performed in triplicate with data obtained from a minimum of three biological replicates (n). All statistical analyses were performed using Minitab Statistical Software version 18.1 (Minitab LLC, State College, PA, USA). Data are presented as mean ± standard error of the mean (SEM). Differences between two groups were assessed using an unpaired Student’s t-test. A p-value of <0.05 was considered statistically significant

3. Results

3.1. Generation of RcN2 Plants and Incorporation of lpa1

NAS2 catalyzes the biosynthesis of NA, which facilitates the production of DMA; NA and DMA are both secreted by rice roots. ,, In this study, we used the OsRCc3 promoter , to overexpress NAS2 in rice roots to generate RcN2, thereby enhancing the mineral content of grains (Figure a). To generate chalky rice grains with low PA content, we also used the CRISPR/Cas9 tool to mutate Lpa1, producing lpa1 mutant plants with low PA content. We also generated lpa1 plants carrying the RcN2 transgene via cotransformation, with the aim of enhancing grain mineral content and introduce chalkiness to facilitate flour production (Figure b).

1.

Generation of lines expressing NAS2 in roots driven by OsRCc3 promoter and harboring a mutation of Lpa1. (a) Diagram of the transgene used to generate plants expressing NAS2 in a tissue-specific manner. The resulting lines are designated RcN2 (for OsRCc3pro:NAS2). LB, left border; RB, right border; BAR, selection marker conferring resistance to Basta. OsRCc3 is a tissue-specific promoter predominantly active in roots. (b) Diagram of the Lpa1 locus (OsLpa1, LOC_Os02g57400) showing three alternative splicing forms, and the CRISPR/Cas9 target site and mutation patterns in the mutant plants. Gray boxes indicate untranslated regions (UTRs), black boxes indicate exons, and solid horizontal lines indicate introns. RcN2, transgenic plants harboring the OsRCc3pro:NAS2 transgene, lpa1, CRISPR/Cas9-generated knockout mutants of Lpa1, lpa1 RcN2, lpa1 mutants in the RcN2 background. (c) Relative NAS2 expression levels in the roots of WT and transgenic seedlings. OsUBQ5 was used as the internal control. Values are presented as means ± SEM (n = 3). Statistical significance was assessed by two-tailed, unpaired Student’s t-test. ***p < 0.001.

We checked NAS2 expression levels in the roots of WT, RcN2, lpa1, and lpa1 RcN2 seedlings (Figure c). We observed significantly higher NAS2 transcript levels in the roots of RcN2 and lpa1 RcN2 seedlings compared to those of WT, while NAS2 was expressed at similar levels in WT and the two Lpa1 knockout lines, lpa1–1, and lpa1–2. NAS2 expression was highest in the roots of RcN2–2 seedlings, reaching levels about 11-fold higher than in WT. NAS2 expression levels were about 5.7-fold, 7-fold, and 5.8-fold higher in the roots of RcN2–1, lpa1 RcN2–1, and lpa1 RcN2–2 seedlings relative to the WT, respectively. Additionally, when we examined NAS2 expression in the roots of 2D, its expression level was about 4-fold higher than in WT (Figure S1).

3.2. PA, NA, and DMA Contents in Mature Brown Grains

Previous studies have shown that lpa1 mutants have lower grain PA content, by about 28–48% in mutants in a japonica background and about 35% in mutants in an indica background. , To verify the new lpa1 mutants generated in this study, we measured the PA content of mature brown grains from WT and transgenic plants (Figure a). The PA contents of grains from the RcN2–1 and RcN2–2 lines were not significantly different from those of WT grains. However, we observed a significant drop in PA amounts in the grains of our new lpa1 mutants. For instance, the mature brown grains of lpa1–1, lpa1–2, lpa1 RcN2–1, and lpa1 RcN2–2 contained about 38%, 30%, 41%, and 38% less PA, respectively, than those of the WT (Figure a).

2.

Contents of PA, NA, and DMA in mature brown grains of WT and transgenic plants. Contents of PA (g/100 g) (a), NA (μg/g) (b), and DMA (μg/g) (c). Values are presented as means ± SEM (n = 3). Statistical significance was assessed by two-tailed, unpaired Student’s t-test. *p < 0.05; **p < 0.01; ***p < 0.001.

Rice roots secrete NA and DMA, which chelate various metal ions in the soil, and the resulting NA–metal and DMA–metal complexes are eventually translocated into the grains. We thus measured NA and DMA contents in mature grains. Compared to WT grains, the grains of all other genotypes contained significantly higher amounts of NA and DMA (Figure b,c). For instance, grains of RcN2–1, RcN2–2, lpa1 RcN2–1, and lpa1 RcN2–2 contained about 16.0-, 16.2-, 26.0-, and 18.1-fold higher levels of NA, respectively, than in WT grains (Figure b). Similarly, DMA levels in the grains of RcN2–1, RcN2–2, lpa1 RcN2–1, and lpa1 RcN2–2 plants were about 3.5-, 3.5-, 4.0-, and 3.8-fold higher, respectively, than those of WT (Figure c). Interestingly, grains of lpa1–1 and lpa1–2 plants also contained slightly higher amounts of NA (1.7- and 3.1-fold, respectively) and DMA (1.7- and 1.2-fold, respectively) compared to WT grains (Figure b,c), suggesting that low PA in grains may enhance NA and DMA accumulation.

3.3. Mineral Contents Are Enhanced in Brown and Polished Grains of RcN2 and lpa1 RcN2 Plants

Overexpression of NAS2 enhances the NA and DMA contents of rice grains, leading to higher contents of minerals, especially Fe and Zn, in more bioavailable forms. ,− , Similarly, the grains of lpa1 mutants accumulate higher Fe and Zn levels. , We measured the contents of several essential micro-minerals (Figure S2) and macro-minerals (Figure S3) in brown grains from WT and transgenic plants. The levels of Fe, Zn, manganese (Mn), and copper (Cu) were higher in the brown grains of all transgenic lines compared than in WT (Figure S2).

Notably, brown grains from the lpa1 RcN2–1 line exhibited the highest accumulation of these microminerals compared to those of WT and the other transgenic lines. For instance, Fe levels in the brown grains of RcN2–1, RcN2–2, lpa1–1, lpa1–2, lpa1 RcN2–1, and lpa1 RcN2–2 lines were about 1.6-, 1.8-, 1.2-, 1.3-, 2.7-, and 2.1-fold higher, respectively, than those in the brown grains of WT (Figure S2a). Zn levels were about 1.7-, 1.8-, 1.1-, 1.2-, 2.1-, and 2.0-fold higher, respectively (Figure S2b). Mn levels were 1.3-, 1.3-, 1.1-, 1.2-, 1.6-, and 1.3-fold higher, respectively (Figure S2c) and Cu levels were 1.2-, 1.3-, 1.1-, 1.1-, 1.6-, and 1.5-fold higher, respectively (Figure S2d). For the macro-minerals, magnesium (Mg) and potassium (K), all genotypes showed 1.1- to 1.4-fold higher levels compared to the brown grains of WT, with lpa1 RcN2–1 showing the highest accumulation (Figure S3b,d).

Since polished rice grains are preferred over brown grains for consumption, we also measured the amounts of the same nutrients in polished grains (Figures and S3). As with brown grains, the polished grains of lpa1 RcN2–1 also contained the highest amounts of micro-minerals and macro-minerals compared to those of WT and other lines (Figures and S3). For instance, Fe levels in the polished grains of RcN2–1, RcN2–2, lpa1–1, lpa1–2, lpa1 RcN2–1, and lpa1 RcN2–2 lines were about 2.8-, 3.6-, 1.5-, 1.8-, 3.9-, and 3.3-fold higher, respectively, than those of WT (Figure a). Likewise, Zn levels were 2.3-, 2.2-, 1.1-, 1.5-, 2.7-, and 2.3-fold higher, respectively (Figure b). Mn levels were 1.7-, 2.1-, 1.6-, 1.5-, 2.3-, and 2.2-fold higher, respectively (Figure c). Cu levels were 1.4-, 1.4-, 1.2-, 1.0-, 1.9-, and 1.7-fold higher, respectively (Figure d).

3.

Polished grains of RcN2, lpa1, and lpa1 RcN2 transgenic plants accumulate more essential micro-minerals. (a–d) Contents of Fe (a), Zn (b), Mn (c), and Cu (d) in polished grains from WT, RcN2, lpa1, and lpa1 RcN2 transgenic plants (in μg/g). Values are presented as means ± SEM (n = 3). Statistical significance was assessed by two-tailed, unpaired Student’s t-test. *p < 0.05; **p < 0.01; ***p < 0.001.

Moreover, the polished grains of RcN2 and lpa1 lines contained similar and significantly higher levels of Mg and K compared to those of WT. Notably, polished grains of lpa1 RcN2 had the highest levels of Mg and K among all lines tested (Figure S3a,c). Additionally, we measured the As content of polished and brown grains for WT and transgenic lines. We did not detect significant differences in As levels across genotypes (Figure S3e,f). These results suggest that brown and polished grains from RcN2 lines both contain higher levels of essential macro- and micro-minerals, and that knocking out Lpa1 in the RcN2 background further enhances mineral accumulation beyond that of RcN2 alone.

3.4. Measurement of Metal–Ligand Complexes by SEC–ICP–MS

To determine the ligands that bind to metal cations present in rice water extracts from various plant lines, we conducted SEC coupled with ICP–MS and ESI–MS. We examined a range of elements to obtain a comprehensive understanding of metal speciation across genotypes. The chromatograms obtained from SEC–ICP–MS analysis of the WT are shown with the most prominent peaks labeled 1 through 9 (Figure S4). Chromatograms for Fe and Zn from the SEC–ICP–MS analysis of WT and lpa1 RcN2–2 are shown in Figure S5. The quantification of metal–chelate complexes is given in Figure .

4.

Measurement of metal–ligand complexes by SEC–ICP–MS. Percent signal intensities of high-molecular-weight (HMW) PA–Fe (a), HMW PA–Zn (b), NA–Fe + DMA–Fe (c), and NA–Zn + DMA–Zn (d) complexes in SEC–ICP–MS peaks from ammoniumacetate buffer extracts (pH 6.2) of mature brown grains from WT and transgenic plants. Values are presented as means ± SEM (n = 3). Statistical significance was assessed by two-tailed, unpaired Student’s t-test. *p < 0.05; **p < 0.01.

Following the methodology of Flis et al., the mass spectra recorded by SEC–ESI–MS at the peak retention time enabled the identification of the smallest ligands (peaks 6 to 9). Peak 6 (retention time of 14.2 min) corresponds to PA, likely in its monomeric form, as indicated by its m/z of 328.92340 for its unbound form, [PA – 2H]^2–, and m/z 355.88306 for its Fe-bound form, [PA + Fe­(II) – 4H]^2–. We predominantly detected phosphorus (P), Fe, aluminum (Al), Zn, Mn, and calcium (Ca) in peak 6 as well, consistent with the presence of phytate and metal–phytate complexes. Peak 7 (18.3 min) contained citrate (Citr) at m/z 191.01974, [Citr – H]⁻. While we did not characterize Citr–metal complexes here due to their low concentrations, the presence of Citr and metal signals (Fe, Zn, Al, Mn, and Ca) support the involvement of Citr in metal binding. Peak 8 (20 min) was associated with DMA, detected at m/z 305.13520, [DMA + H]⁺, forming complexes with Fe­(III), Zn, and Cu­(II). Peak 9 (21.5 min) corresponds to NA at m/z 304.15071, a known Zn chelator. Although we detected only trace amounts of an NA–Fe­(II) complex at m/z 356.05505, [NA + Fe­(II) – H]⁺, NA likely plays a significant role in chelating Fe­(II) in rice grains, as this complex is unstable during chromatography. The Fe signal in peak 9 suggested the presence of other organic acids, as its peak intensity did not correlate with NA levels. We also detected P-containing compounds, possibly phosphate and glycerophosphorylcholine, in peaks 8 and 9; however, their presence did not appear to correlate directly with metal complex formation. Notably, their abundance was higher in the grains of lpa1 and lpa1 RcN2, suggesting the involvement of alternative P-containing metabolites in the absence of phytate.

We identified peaks 1 to 5 based on elemental composition, retention time (corresponding to molecular weight), and variation across plant lines. Peak 1 (retention time of 9.2 min) contained a substantial amount of P, whose levels were significantly lower in the grains of lpa1 and lpa1 RcN2. Several metals, including Fe, Al, Zn, Mn, Cu, and Ca, as well as P, followed similar trends, suggesting the presence of high-molecular-weight (HMW) phytate-containing complexes (>20 kDa). Peak 2 (9.9 min) corresponds to molecules in the 14–20-kDa range that primarily bind to Fe, Zn, and Mn. These compounds are likely albumin-type (Alb-like) proteins based on their size, metal-binding properties, and presence in rice extracts. Peak 3 (11.2 min) revealed minor compounds, likely proteins that bind to Fe and Cu, as well as a molecule binding to strontium (Sr), Ca, and barium (Ba). With a molecular weight of about 10–12 kDa and a distinct metal-binding profile, this compound is likely rhamnogalacturonan II (RG-II). Peaks 4 (12 min) and 5 (12.5 min) correspond to molecules that predominantly bind to Zn and Cu, respectively. These compounds, with molecular weights ranging from 6 to 10 kDa and low affinity for cadmium (Cd), are likely class I and class II metallothioneins (MT). Considering only the pool of stable and soluble Fe complexes present in the water extracts and Fe chromatographic peak areas, “HMW PA–Fe + PA–Fe”, “NA–Fe + DMA–Fe”, and “other detected ligands–Fe” represent 70%, 15%, and 15% of the total soluble Fe in WT, respectively, and 19%, 60%, and 21% in lpa1 RcN2–2, respectively (Figure S5).

The HMW phytate-containing complexes (peak 1) represented phytate-bound metal ions, which exhibit low bioavailability due to their insolubility and resistance to digestion. Fe and Zn levels associated with this fraction were significantly lower in the grains of lpa1 and lpa1 RcN2 lines than in those of WT (Figure a,b). By contrast, metal levels associated with NA and DMA showed a moderate rise in their Fe levels and markedly higher Zn levels in the grains of RcN2 and lpa1 RcN2 lines compared to those of WT (Figure c,d). This latter fraction corresponds to highly bioavailable metal ions, suggesting that these nutrients are more likely to be readily absorbed from the grains of RcN2 and lpa1 RcN2 lines in the intestinal tract.

3.5. Expression of Metal Homeostasis-Related Genes is Enhanced in lpa1 and RcN2 Lines

Overexpression of NAS2 leads to higher transcript levels of various metal homeostasis–related genes in roots, facilitating the transport of NA–metal and DMA–metal complexes to the aerial parts of plants, especially to grains. − We performed RT-qPCR to assess the expression levels of different metal homeostasis–related genes in the roots of seedlings from WT and other lines (Figure ). NA AMINOTRANSFERASE 1 (OsNAAT1) and DMA SYNTHASE 1 (OsDMAS1) function downstream of NAS2 to synthesize DMA, while TRANSPORTER OF MUGINEIC ACID FAMILY PHYTOSIDEROPHORE 1 (OsTOM1) facilitates DMA secretion into the rhizosphere. The expression levels of the three corresponding genes were significantly higher in the roots of RcN2–1, RcN2–2, lpa1 RcN2–1, and lpa1 RcN2–2 seedlings than in those of WT (Figure a–c). IRON-REGULATED TRANSPORTER 1 (OsIRT1) takes up Fe from the soil. OsIRT1 expression was higher in RcN2–1, RcN2–2, lpa1 RcN2–1, and lpa1 RcN2–2 than in WT (Figure d). YELLOW STRIPE-LIKE 2 (OsYSL2) facilitates the transport of NA–Fe and NA–Mn complexes within the plant. OsYSL2 expression was also higher in all RcN2 and lpa1 RcN2 lines than in WT (Figure e). COPPER TRANSPORTER 1 (OsCOPT1) is involved in Cu uptake and distribution, while ZRT/IRT-LIKE PROTEIN 4 (OsZIP4) and OsZIP5 are crucial Zn transporters responsible for Zn uptake and internal distribution in rice. , The expression of the corresponding genes was significantly higher in RcN2–1, RcN2–2, lpa1 RcN2–1, and lpa1 RcN2–2 than in WT (Figure f–h). By contrast, the two lpa1 lines showed expression levels similar to or slightly higher than those of WT for these genes (Figure a-h). We also examined the expression of MAGNESIUM TRANSPORTER 1 (OsMGT1) and HIGH-AFFINITY POTASSIUM TRANSPORTER 5 (OsHAK5) in the roots of all genotypes (Figure S6). OsMGT1 is involved in Mg uptake by roots, and OsHAK5 is critical for K⁺ uptake. The expression levels of their genes were significantly higher in all RcN2, lpa1, and lpa1 RcN2 lines compared to WT. These results suggest that the OsRCc3 promoter controlled NAS2 and mutation of Lpa1 induce the expression of various metal homeostasis–related genes in the roots of rice seedlings.

5.

Expression analysis of metal homeostasis–related genes in the roots of WT and transgenic seedlings. Relative expression levels of OsDMAS1 (a), OsNAAT1 (b), OsTOM1 (c), OsIRT1 (d), OsYSL2 (e), OsCOPT1 (f), OsZIP4 (g), and OsZIP5 (h). OsUBQ5 was used as the internal control. Values are presented as means ± SEM (n = 3). Statistical significance was assessed by two-tailed, unpaired Student’s t-test. *p < 0.05; **p < 0.01; ***p < 0.001.

3.6. Evaluation of the Agronomic Traits of RcN2, lpa1, and lpa1 RcN2 Plants

We evaluated various agronomic traits of field-grown WT, RcN2, lpa1, and lpa1 RcN2 plants (Figures and ). RcN2 plants showed culm lengths similar to those of WT, whereas lpa1 and lpa1 RcN2 plants had significantly shorter culms (Figure a,b). Specifically, culm lengths were shorter by 9.1%, 15.4%, 5.4%, and 3.4% in lpa1–1, lpa1–2, lpa1 RcN2–1, and lpa1 RcN2–2 plants, respectively, relative to those of WT (Figure b). Similarly, the number of tillers per plant was lower by 22.6%, 32.1%, 29.8%, and 25.0% in the same order compared to WT (Figure c), with no significant difference between the tiller number of WT and RcN2 plants. The panicles of lpa1 plants were shorter than those of WT, while those of WT, RcN2, and lpa1 RcN2 plants were comparable (Figure d). Total grain number per panicle was unchanged in RcN2 compared to WT, but was lower by 16.6%, 19.4%, 18.8%, and 20.7% in lpa1–1, lpa1–2, lpa1 RcN2–1, and lpa1 RcN2–2 plants, respectively (Figure e). We observed similar trends for seed-setting rate and hundred-grain weight (Figure f,g). RcN2 plants did not differ significantly from WT for these traits. The seed-setting rate was lower by 29.8%, 34.9%, 22.2%, and 28.4%, while the hundred-grain weight was lower by 15.9%, 11.3%, 20.1%, and 16.4%, in lpa1–1, lpa1–2, lpa1 RcN2–1, and lpa1 RcN2–2 respectively, compared to WT.

6.

Agronomic trait analysis of field-grown WT and transgenic plants. (a) Representative images of mature WT and transgenic plants. Scale bars, 10 cm. From left to right, WT, RcN2–1, RcN2–2, lpa1–1, lpa1–2, lpa1 RcN2–1, and lpa1 RcN2–2. (b) Culm length (cm) at maturity. (c) Number of tillers per plant. (d) Panicle length (cm). (e) Total number of grains per panicle. (f) Seed-setting rate (%). (g) Hundred-grain weight (g). Values are presented as means ± SEM (n = 6). Statistical significance was assessed by two-tailed, unpaired Student’s t-test. *p < 0.05; **p < 0.01; ***p < 0.001.

7.

Grain phenotypes from WT and transgenic plants. (a) Representative photographs of dehulled grains. Scale bars, 1 cm. The right image was captured on a backlit glass panel to visualize the chalky areas. (b) Grain length (cm). (c) Grain width (cm). (d) Percentage of chalky area per grain. Values are presented as means ± SEM (n = 6). Statistical significance was assessed by two-tailed, unpaired Student’s t-test. **p < 0.01; ***p < 0.001. (e–k) Representative scanning electron microscopy (SEM) images of starch granules in the endosperm of WT (e), RcN2-1 (f), RcN2-2 (g), lpa1–1 (h), lpa1–2 (i), lpa1 RcN2–1 (j), and lpa1 RcN2–2 (k). Scale bars, 10 μm.

We also evaluated grain phenotypes (Figure ). The grains were on average longer (by 3.5–10.2%) and thinner (by 1.5–11.3%) in RcN2, lpa1, and lpa1 RcN2 lines compared to WT grains (Figure a–c). These changes were less pronounced in RcN2 lines than in lpa1 or lpa1 RcN2 lines. We also measured total chalky grain area (Figure a,d). The grains of RcN2–1 and RcN2–2 lines showed little chalkiness (4.9% and 6.0%), not significantly different from that seen in WT grains (3.1%). By contrast, grains from lpa1–1, lpa1–2, lpa1 RcN2–1, and lpa1 RcN2–2 lines were highly chalky, with 44.2%, 57.1%, 50.2%, and 58.3% chalkiness, respectively (Figure d). To determine whether chalkiness was associated with starch granule structure, we analyzed the endosperm of grains from WT, RcN2, lpa1, and lpa1 RcN2 lines by scanning electron microscopy (SEM) (Figure e-k). In line with the chalkiness data, the endosperm of lpa1 and lpa1 RcN2 grains showed loosely packed spherical starch granules, unlike those of WT and RcN2 grains. These data indicate that RcN2 plants are agronomically similar to the WT and their grains show no increase in chalkiness, whereas lpa1 and lpa1 RcN2 plants exhibit significant growth penalties and produce highly chalky grains.

4. Discussion

4.1. A Combined Strategy Using Tissue-Specific NAS2 Expression and Lower PA Levels Enables Efficient Mineral Biofortification in Rice Grain

Biofortification is a promising strategy to combat malnutrition resulting from insufficient intake of essential micronutrients and has demonstrated success in improving health outcomes across various populations. In this study, we expressed NAS2 in roots with the tissue-specific OsRCc3 promoter to efficiently biofortify rice grains. To further facilitate biofortification by lowering the PA content and introducing chalkiness to rice grains to assist with rice flour production, we knocked out Lpa1 into the RcN2 background. Tissue-specific NAS2 expression resulted in higher relative NAS2 transcript levels in the roots of RcN2 and lpa1 RcN2 seedlings compared to those of WT. Notably, NAS2 was expressed at higher levels in the roots of RcN2 and lpa1 RcN2 seedlings than those of the 2D. PA content was significantly lower in mature brown grains of lpa1 and lpa1 RcN2 plants, with drops ranging from 31% to 41% compared to WT grains, whereas RcN2 and WT grains did not differ in their PA contents. Previous studies have reported similar declines in the PA contents of lpa1 mutants in japonica and indica backgrounds, ranging from 28% to 49%. ,, When we measured the NA and DMA contents of grains from WT and other lines, we established that NA and DMA accumulated to higher levels in the mature brown grains of RcN2 and lpa1 RcN2 plants relative to WT grains. These results are in agreement with previous studies utilizing NAS2 overexpression that have reported higher levels of NA and DMA in mature grains of the transgenic plants compared to WT. − ,,

NAS2 synthesizes NA, which is the precursor to DMA. NAS2 expression was driven by the OsRCc3 promoter, which is primarily active in root cortical cells and the phloem companion cells of nodes. , NA and DMA produced in root cortical cells bind Fe and Zn taken up from the soil and contribute to xylem-mediated transport to shoots via transporters, while NA in phloem companion cells in nodes facilitates the translocation of metals chelated to NA and DMA to grains. The heterologous expression of HvNAS1 from barley (Hordeum vulgare) in rice was shown to increase NA and DMA contents in the roots and shoots, supporting the idea that NAS2 overexpression would lead to elevated NA and DMA pools. As we only quantified NA and DMA contents in brown grains, and not in the roots or phloem tissue, the possible contribution of root-derived or phloem-derived NA and DMA to grain accumulation remains speculative and will require targeted follow-up studies, such as root and phloem profiling of NA and DMA levels or isotope tracing. We observed greater accumulation of NA and DMA in the grains of lpa1 mutants compared to WT; similarly, grains of lpa1 RcN2 lines also showed a slight increase in their NA and DMA contents compared to RcN2 grains. Since lpa1 and lpa1 RcN2 grains contained lower amounts of PA, NA and DMA would face less competition from PA to form complexes with metal ions, thereby contributing to a higher accumulation of NA and DMA complexed to metal ions.

PA and DMA both form strong Fe­(III) complexes, and potentiometric data indicate that certain phytate species can exhibit conditional stability constants exceeding those of DMA under specific pH and ionic strength conditions. , In this study, we detected Fe­(III) associated with both PA and DMA. A similar association was reported in maize (Zea mays), suggesting that Fe speciation in grains is determined by thermodynamic stability, kinetic factors, ligand accessibility (including steric and solubility constraints for PA), and subcellular compartmentalization. While PA likely serves as a storage form of Fe, DMA may represent a more dynamic and readily mobilizable Fe reservoir, even in the presence of PA.⁷⁶ In line with this idea, lpa1 grains had greater amounts of various minerals than WT, a phenotype that was even more pronounced in lpa1 RcN2 grains. Lower PA levels are therefore associated with higher mineral contents in crop grains, especially raising the bioavailability of minerals for humans and other monogastric animals. , Additionally, the modest rise in NA and DMA levels along with those of Fe and Zn observed in lpa1 grains likely reflects direct or compensatory effects due to improved metal mobility, induced metal transporters, changes in chelator metabolism, and altered seed loading dynamics. Similarly, the grains of KBNT lpa1–1 plants were reported to contain higher Zn levels compared to WT, whereas the contents of Fe, Mg, and K were almost identical between the two genotypes. Another study using an lpa1 mutant, Os-lpa-XQZ-1, reported significantly higher amounts of Fe, Zn, and Ca compared to WT grains. Importantly, polished RcN2 and lpa1 RcN2 grains contained significantly more Fe and Zn than polished WT grains, reaching the established targets of 15 μg/g Fe and 28 μg/g Zn for efficiently biofortified white rice. In polished grains of RcN2 and lpa1 RcN2, Fe and Zn levels were comparable to those reported for other NAS2-overexpressing lines. Constitutive overexpression of NAS2 under the 35S promoter yielded 4- and 2-fold enhancements of Fe and Zn levels in polished grains compared to those of WT. Likewise, 2D polished grains displayed 3- and 2.7-fold higher contents of Fe and Zn. , In an indica rice cultivar coexpressing 35S-driven NAS2 and endosperm-specific promoter-driven GmFer-H1, polished grains showed 7.5- and 3.8-fold higher levels in Fe and Zn, respectively, relative to the WT. While the above-mentioned studies focused mainly on Fe and Zn, we observed significantly greater accumulation of other essential micro-minerals as well, such as Mn and Cu. We hypothesize that the higher amounts of Mn and Cu in the polished grains of the transgenic lines are due to the enhanced NA and DMA production of these plants. NA and DMA can form complexes with Mn and Cu, in addition to Fe and Zn, − and the higher NA and DMA contents of the transgenic lines should help take up more NA–metal and DMA–metal complexes from the soil and relocate them to the grains. Notably, the amounts of these trace metals were found significantly higher after polishing the brown grains of transgenic lines, particularly Fe and Mn amounts. This finding suggests that the translocation of NA–metal and DMA–metal complexes predominantly occur into the grain endosperm.

Additionally, the levels of NA and DMA bound to Fe and Zn were higher in the brown grains of transgenic lines relative to those of WT, as measured by SEC–ICP–MS analysis, with the highest levels observed in lpa1 RcN2 grains. By contrast, PA-bound Fe and Zn accumulated to lower levels in lpa1 and lpa1 RcN2 grains. Higher levels of NA-bound and DMA-bound Fe and Zn, together with the lower levels of their PA-bound forms, would be likely to result in improved bioavailability of Fe and Zn. It has been reported that the grains of NAS2 overexpressing plants contain higher Fe and Zn levels in the endosperm with greater bioavailability. , Bioavailability of the enhanced Fe and Zn in the 2D grains was confirmed by two independent mouse-feeding studies. , An in vitro Caco-2 cell assay was also employed which further corroborated the increased Fe bioavailability from the grains of NAS2 overexpressing plants. PA–Fe­(III) complexes exhibit poor bioavailability due to their high stability, low solubility at intestinal pH, and the absence of endogenous phytase activity in the human gastrointestinal tract. Consequently, Fe remains tightly bound and is largely inaccessible for absorption. By contrast, Fe­(II/III) complexes with NA and DMA are water-soluble and form moderately stable complexes that do not precipitate under physiological conditions. Recently, Murata et al. demonstrated that NA–Fe complexes can be absorbed directly in the human small intestine via the proton-coupled amino acid transporter PAT1. Together, these findings suggest that the higher bioavailability of NA–Fe and DMA–Fe arises from favorable solubility and lability and from active transport mechanisms that are absent for PA–Fe. Furthermore, the lower levels of PA–Fe observed in RcN2 grains may reflect the enhanced NA and DMA contents of these grains, and consequently, their higher competition for binding to Fe with PA.

Moreover, we measured higher Mg and K contents in the brown and polished grains of lpa1, RcN2, and lpa1 RcN2 plants compared to WT. Macro-minerals such as Mg and K are present in higher amounts than micro-minerals and are essential in animal diets due to their roles in enzyme activation, muscle contraction, nerve function, and fluid balance. , NA binds weakly to Mg in vitro. However, there is no evidence that biologically relevant NA–Mg or DMA–Mg complexes form in planta, nor that NA or DMA chelate K. Thus, future studies should evaluate whether such complexes form at physiological concentrations and whether NA or DMA influence Mg and K homeostasis indirectly. Enhanced NA and DMA contents may improve the efficiency of metal transporters, thereby affecting Mg and K uptake to maintain ionic balance. , We did measure higher expression levels for several metal homeostasis–related genes in the roots of RcN2 and lpa1 RcN2 seedlings. Overexpression of NAS2 enhances DMA production via the action of OsNAAT1 and OsDMAS1, while OsTOM1 facilitates the secretion of DMA into the soil. OsIRT1 and OsYSL2 are key genes regulating Fe levels in rice plants. Specifically, OsIRT1 facilitates Fe uptake from the soil, while OsYSL2 aids in the translocation of NA–Fe and NA–Mn complexes. , OsCOPT1 is involved in Cu uptake and distribution, while OsZIP4 and OsZIP5 are important for Zn uptake and distribution. , These genes were expressed at higher levels in the roots of RcN2 and lpa1 RcN2 seedlings, which may contribute to greater and more efficient transport of Fe, Zn, Mn, and Cu to the grains later during plant development. Additionally, the expression levels of OsMGT1 and OsHAK5, involved in Mg and K uptake, , were also higher in RcN2, lpa1, and lpa1 RcN2 seedlings, which may reflect physiological adjustments to maintain ionic balance in response to altered metal homeostasis.

4.2. RcN2 and lpa1 RcN2 Lines Provide Practical Applications for Nutritional Rice Improvement and Flour Use

Transgenic rice plants overexpressing NAS2 have been reported to be compromised in their agronomic traits, such as lower yield and altered growth patterns. While overexpression enhances the accumulation of essential minerals in grains, it may also lead to imbalances in nutrient homeostasis and metabolic processes, thus adversely affecting development and productivity. , Growth penalties may also result from Fe-mediated oxidative stress. Importantly, field-grown RcN2 plants did not show any noticeable growth penalty compared to the WT. Culm length, tiller number, panicle length, grain number, grain weight, and fertility were similar between RcN2 and WT. The OsRCc3 promoter-driven expression of NAS2 may therefore promote optimal mineral distribution and limit or prevent mineral imbalance. In a japonica background, constitutive overexpression of NAS2 negatively influenced agronomic traits, whereas transgenic plants in an indica background exhibited minimal or no significant yield loss. , We produced the RcN2 plants used in this study in the Dongjin japonica background and did not observe any significant negative effects on their agronomic traits. The OsRCc3 promoter is active in nodes and internodes in addition to roots. In rice, nodes play a central role in directing minerals taken up by roots into the phloem for delivery to developing grains. , Thus, enhanced NA levels in nodes may facilitate efficient loading of NA–metal and DMA–metal complexes into the phloem. Furthermore, transporter genes expressed in nodes can help prevent the accumulation of toxic metals. Notably, RcN2, lpa1, and lpa1 RcN2 grains did not accumulate more As than WT grains. A previous study using OsRCc3 promoter-driven ScYCF1 and OsABCC1 showed lower As accumulation without growth penalty in rice.

Rice lpa1 mutants have lower grain PA content but often exhibit poor agronomic traits, resulting in yield losses. ,,, The lpa1 and lpa1 RcN2 plants showed significant growth penalty compared to WT and RcN2 plants. PA serves as a major storage form of P, and lower PA levels lead to greater levels of free inorganic phosphate, which can affect germination, seedling vigor, and development. , Despite their yield loss, these lpa1 and lpa1 RcN2 plants offer nutritional benefits by enhancing mineral bioavailability and contributing to lower P-related water pollution. Other lpa1 mutant alleles exhibit clear agronomic penalties. Kishor et al. showed that plants with mutation of Lpa1 are shorter, with fewer tillers, and lower grain yield. The KBNT lpa1–1 and DR1331–2 display a 5–10% drop in grain dry weight. Likewise, mutants in INOSITOL 1,3,4-TRISPHOSPHATE 5/6-KINASE 6 (OsITPK6) or MULTIDRUG RESISTANCE PROTEIN 5 (OsMRP5) accumulate substantially lower PA levels but suffer from lighter grains, delayed seedling vigor, and lower yields. , By contrast, several lpa lines have been engineered without detectable growth or yield defects. Seed-specific RNA interference (RNAi)-mediated knockdown of OsIPK1 or MYO-INOSITOL-1-PHOSPHATE SYNTHASE (OsMIPS), and the Tos17 insertion mutants in SULTR-LIKE P DISTRIBUTION TRANSPORTER (SPDT) maintain normal agronomic performance. − Combining RcN2 with these lpa mutants may be a promising strategy to further enhance nutritional benefits while minimizing yield penalties. Importantly, only lpa1 mutants are characterized by a chalky endosperm phenotype, a potentially advantageous feature for rice flour processing not observed in other lpa backgrounds. We observed 44–58% chalky area in lpa1 and lpa1 RcN2 grains, in contrast to less than 4% in WT. An SEM analysis confirmed that these grains contained loosely packed spherical starch granules. Although chalky rice grains are generally considered less desirable for consumption, they offer advantages for rice flour production. The softer grain texture makes grinding easier and limits starch damage, improving flour quality. ,

In conclusion, RcN2 and lpa1 RcN2 plants showed efficient biofortification of essential macro- and micro-minerals in their polished and brown grains. Importantly, field-grown RcN2 plants showed no significant agronomic penalty compared to WT and represent a promising genetic resource for breeding biofortified rice grains. Enhanced chalkiness facilitates energy-efficient milling and produces a finer flour with lower starch damage, ideally suited for gluten-free food applications, while significantly lower PA levels markedly improve mineral bioavailability. Although grain yield is lower in lpa1 RcN2, the observed chalkiness and enhanced grain mineral accumulation suggest a potential to offset yield losses through improved processing and nutrient delivery. However, quantitative milling and economic assessments are required to confirm this. Biosafety and regulatory assessments, including the characterization of transformation events, transgene copy number, insertion site, off-target analyses, and full compositional testing, should be completed prior to any confined trials or release, in accordance with country-specific regulatory requirements. Translation activities, including introgression of the traits into elite cultivars, multienvironment performance trials, quantitative milling and flour assessments, and small sensory studies, will support the translation of this approach into functional biofortified rice flour–based products.

The data that support the findings of this study are available in the supporting material of this article.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jafc.5c06403.

• Supplementary Table 1. Primers used for RT-qPCR. Figure S1. Relative NAS2 expression levels in the roots of WT and 2D seedlings. Figure S2. Mature brown grains of RcN2 and lpa1 RcN2 plants contain more essential micro-minerals than WT. Figure S3. Amounts of various minerals in mature polished and brown grains of WT and transgenic plants. Figure S4. Chromatograms of ³¹P, ⁵⁵Mn, ⁵⁶Fe, ⁶³Cu, ⁶⁴Zn, and ⁸⁸Sr obtained by SEC–ICP–MS analysis of water extracts from mature brown grains of WT plants. Figure S5. Chromatograms of ⁵⁶Fe and ⁶⁴Zn for water extracts from mature brown grains of WT and lpa1 RcN2–2 transgenic mutant plants. Figure S6. Expression analysis of metal homeostasis–related genes in the roots of WT and transgenic mutant seedlings (PDF)

#.

M.M.R. and D.-Y.S. contributed equally to this work. J.-S.J. conceived the study; J.-S.J. and M.M.R. designed the experiments; M.M.R., D.-Y.S., M.M.R., C.Y.H., A.O., T.B., M.S.A., L.O., and W.-Y.S. performed the experiments; M.M.R., D.-Y.S., M.S.A., L.O., W.-Y.S., S.L., G.A. and J.-S.J. analyzed the data and wrote the manuscript. All authors have read and approved the final manuscript.

This work was supported by the Rural Development Administration (RS-2024–00322378) and the National Research Foundation of Korea (2023R1A2C1003142).

The authors declare no competing financial interest.

Published as part of Journal of Agricultural and Food Chemistry special issue “Synthetic Biology in Food Production”.