Golgi-Localized OsFPN1 is Involved in Co and Ni Transport and Their Detoxification in Rice

Cobalt (Co) and nickel (Ni) are beneficial and essential elements for plants, respectively, with the latter required for urease activity, which hydrolyzes urea into ammonium in plants. However, excess Co and Ni are toxic to plants and their transport mechanisms in rice are unclear. Here, we analyzed an ethyl methanesulfonate (EMS)-mutagenized rice mutant, 1187_n, with increased Co and Ni contents in its brown rice and shoots. 1187_n has a mutation in OsFPN1, which was correlated with a high Co and Ni phenotype in F2 crosses between the parental line and mutant. In addition, CRISPR/Cas9 mutants exhibited a phenotype similar to that of 1187_n, demonstrating that OsFPN1 is the causal gene of the mutant. In addition to the high Co and Ni in brown rice and shoots, the mutant also exhibited high Co and Ni concentrations in the xylem sap, but low concentrations in the roots, suggesting that OsFPN1 is involved in the root-to-shoot translocation of Co and Ni. The growth of 1187_n and CRISPR/Cas9 lines were suppressed under high Co and Ni condition, indicating OsFPN1 is required for the normal growth under high Co and Ni. An OsFPN1-green fluorescent protein (GFP) fusion protein was localized to the Golgi apparatus. Yeast carrying GFP-OsFPN1 increased sensitivity to high Co contents and decreased Co and Ni accumulation. These results suggest that OsFPN1 can transport Co and Ni and is vital detoxification in rice.


Background
Co and Ni are rare elements of the earth's crust and usually compete with Fe when absorbed by organisms (Buschow 1971;Salnikow et al. 2004;Morrissey et al. 2009). Co is a beneficial element that improves growth in certain plants, particularly leguminous plants (Karlengen et al. 2013). However, excess Co leads to black patches in tomato fruit (Chatterjee and Chatterjee 2005), and chlorosis in the young leaves of French bean (Phaseolus vulgaris L.), groundnut mung bean (Vigna radiate), and tomato (Lycopersicon esculentum) plants by causing reduced chlorophyll a, b, and carotene concentrations (Liu et al. 2000;Gopal et al. 2003;Chatterjee et al. 2006). Ni is an essential micronutrient in plants that is present in the active site of urease, an enzyme that hydrolyzes urea into ammonium in plant tissues (Eskew et al. 1983(Eskew et al. , 1984Kerby et al. 1997). In addition to urease, Ni forms a part of other metalloenzymes, such as superoxide dismutase, glyoxalase, and hydrogenase. Similar to Co, excess Ni affects plant germination and growth. For example, the application of 1.5 mM of Ni decreased pigeon pea (Cajanus cajan (L.) Millspaugh) germination by approximately 20% (Aggarwal et al. 1990; Rao and Sresty 2000). Additionally, the growth of cabbage (Brassica oleracea L var. capitate) was inhibited by 0.5 mM of Ni (Pandey and Sharma 2002).
Ferroportin (FPN), also referred to as iron-regulated gene1 (IREG1) or metal transporter protein1, was first identified in hypotransferrinemic mice, where it was involved in Fe absorption in the duodenum (McKie et al. 2000) and Fe recycling in macrophages (Muckenthaler et al. 2008). Arabidopsis thaliana contains three homologous genes of mammalian FPN, i.e., AtFPN1, AtFPN2, and AtFPN3. AtFPN1 encodes a plasma membrane-localized exporter of Co and Fe, which is expressed in the vasculature of the root and shoot and is involved in the loading of these elements from the pericycle to the xylem. (Morrissey et al. 2009). The expression of AtFPN1 is not regulated by Fe (Morrissey et al. 2009), while AtFPN2 and AtFPN3 is induced by Fe starvation (Morrissey et al. 2009;Kim et al. 2021). AtFPN2 is localized in the tonoplast and is expressed in the epidermis and cortex of the root (Morrissey et al. 2009). The AtFPN2 mutant causes high Co and Ni accumulation in the shoots, but low accumulation in the roots, indicating that it sequesters Co and Ni into vacuoles in the root for detoxification (Schaaf et al. 2006;Morrissey et al. 2009). AtFPN3, which has 20% similarity to AtFPN1 and AtFPN2, is localized in chloroplasts/mitochondria function as Fe exporter and essential for Fe homeostasis (Kim et al. 2021). In addition to Fe homeostasis, AtFPN3, also known as MAR1, also played a role in controlling the entry of antibiotics into chloroplasts . In the hyperaccumulator Psychotria gabriellae, the expression of PgFPN1 in mRNA expression was higher than that in non-accumulator species (Merlot et al. 2014). The overexpression of PgFPN1 in Arabidopsis enhances its tolerance to high levels of Ni (Merlot et al. 2014). In monocotyledons, buckwheat (Fagopyrum esculentum Moench) FPN1 (FeFPN1) is involved in internal Al 3+ detoxification (Yokosho et al. 2016). FeFPN1 is expressed in the roots and is greatly upregulated by Al 3+ , but not by Fe, unlike AtFPN2 (Schaaf et al. 2006;Morrissey et al. 2009). The overexpression of FeFPN1 in Arabidopsis confers Al tolerance, suggesting that it is involved in Al detoxification in buckwheat (Yokosho et al. 2016). However, the function of putative FPNs in rice remains unclear.
In this study, we conducted ionome screening of an ethyl methanesulfonate (EMS)-mutagenized rice M 2 population and identified a mutant with high Co and Ni contents in the grain and shoot. The causal gene encoding FPN is OsFPN1, which is localized to the Golgi apparatus and mediates Co and Ni transport. The mutants exhibited sensitivity to high levels of Co and Ni, indicating that OsFPN1 is necessary for tolerance to these elements.

1187_n Exhibited Increased Co and Ni Contents in Brown Rice, Shoot, and Xylem Sap
To identify the genes regulating mineral transport in rice (O. sativa cv. Hitomebore), ionome screening of the EMS-mutagenized M 2 population was conducted (Tanaka et al. 2016). We isolated a mutant, 1187_n, which exhibited high Co and Ni concentrations in brown rice grown in paddy fields during 2013 and 2016 (Additional file 1: Fig. S1, Fig. 1A). Additionally, 1187_n exhibited high Co and Ni concentrations in shoots when the plants were grown in soil (Fig. 1B) or the Kimura B hydroponic solution (Ma et al. 2001) (Fig. 1C). Additionally, increased Co and Ni concentrations were observed in the xylem sap (Kan et al. 2019) of 1187_n grown under hydroponic conditions (Fig. 1D). These results indicated that 1187_n has defect in root-to-shoot transport of Co and Ni.

Sensitivity of 1187_n to High Co and Ni
To observe the effects of Co and Ni on the growth of 1187_n, the plants were grown in Kimura B solution supplemented with various concentrations of Co or Ni for three weeks. The seedlings were cut at the shootroot junction to divide them into shoots and roots, and the length and dry weight of the shoots and roots were measured. The shoot and root growth of 1187_n were more suppressed under applied Co or Ni concentrations of 1 and 10 μM compared to the wild type (Fig. 2). These results indicate that 1187_n is sensitive to high levels of Co and Ni.

OsFPN1 is the Causal Gene of 1187_n
The phenotype of increased Co concentration in the shoot and sensitivity to Co and Ni of 1187_n is similar to the phenotype of the fpn2 mutant of A. thaliana (Morrissey et al. 2009). Thus, we hypothesized that the FPN2 homolog in rice was the causal gene of 1187_n. Rice contains only one homolog, Os06g0560000 (hereafter named OsFPN1) (Additional file 1: Fig. S2). Sequencing analysis of OsFPN1 in 1187_n revealed that 1187_n had a mutation (T to G) in the second exon, which caused Leu 181 to Gln substitution (Fig. 3A). The OsFPN1 protein was predicted to be composed of nine transmembrane domains (Additional file 1: Fig. S3A). Leu 181 was located between transmembrane domains 4 and 5, and was highly conserved among plant and human FPN homologs (Additional file 1: Fig. S3).
To confirm whether OsFPN1 was the gene responsible for the increased Co and Ni in 1187_n, ionome analysis of F 2 crosses between 1187_n and wild type was conducted in seedlings grown under hydroponic conditions (Fig. 3B). The Co and Ni concentrations were positively correlated and segregated with a ratio of 56:20 (wild type:mutant phenotype), which fits the 3:1 ratio (chisquared test, χ 2 = 0.853), indicating that increased levels of both Co and Ni were caused by a single recessive gene (Fig. 3B). Next, to observe the correlation between the ionome phenotype and mutation in OsFPN1, we designed a dCAPS marker to detect the mutation, and the F 2 plants used for ICP-MS analysis underwent genotyping (Fig. 3B). All F 2 seedlings with homozygous Co and Ni accumulation in Hitomebore and 1187_n. A Co and Ni concentrations in the grains of Hitomebore and 1187_n grown in a paddy field (n = 3-4). B Co and Ni concentrations in the shoots of Hitomebore and 1187_n cultivated in soil (n = 3-4). C Co and Ni concentrations in the shoots of Hitomebore and 1187_n cultivated in a hydroponic culture (n = 11-12). D Co and Ni concentrations in the xylem sap of Hitomebore and 1187_n cultivated in a hydroponic culture (n = 9-11). Xylem sap was collected for 4 h. For A, B data are the same as those in Figure S1. For B-D 21-d-old seedlings were harvested for the experiments. Data represents the mean ± SD. Student's t-test, *p < 0.05; **p < 0.01 conditions. Plants were grown in hydroponic culture for three weeks. Data represent the means ± SD (n = 8-11). Student's t-test, *p < 0.05; **p < 0.01

Fig. 3
Responsibility of OsFPN1 for the phenotype of 1187_n. A Exon intron structure of OsFPN1. Gray boxes, dark boxes, and lines represent the untranslated region, exon, and intron, respectively. OsFPN1 of 1187_n has a nonsynonymous mutation of T to G [leucine (L) is replaced with glutamine (G)]. Osfpn1-2 and osfpn1-3 are CRISPR/Cas 9 mutants with 1 bp and 26 bp deletions at the first exon of OsFPN1, respectively. WT (Nipponbare) is the background of the mutants. B Scatterplot based on the Co and Ni concentrations in the shoots of three-week-old seedlings of Hitomebore, 1187_n, and F2. F2_Wild, F2_Hetero, and F2_Mutant represent the wild-type, heterozygous, and mutant homozygous genotypes of F2, respectively. C Co and Ni concentrations in the shoots of Hitomebore, osfpn1-1, Nipponbare, osfpn1-2, and osfpn1-3. Plants were grown in Kimura B solution supplied with 10 trace elements. Data represent the means ± SD (n = 3-12). Student's t-test (compared to Hitomebore or Nipponbare). **p < 0.01 mutations exhibited increased Co and Ni contents, while the wild-type homozygous and heterozygous F 2 seedlings exhibited low Ni and Co contents (Fig. 3B). These results indicate the co-segregation between the Co and Ni concentrations and the genotype, suggesting that OsFPN1 is the causal gene of 1187_n. Hereafter, we refer to1187_n as osfpn1-1.
To demonstrate that OsFPN1 is the causal gene, two independent CRISPR/Cas9 lines, osfpn1-2 and osfpn1-3, were generated in Nipponbare as it was difficult to transform Hitomebore. osfpn1-2 and osfpn1-3 have 1 bp and 26 bp deletions, respectively. First, we conducted ionome analysis in seedlings, and found that osfpn1-1, osfpn1-2, and osfpn1-3 exhibited higher Co and Ni contents in their shoots than the wild-type plants (Fig. 3C). Taken together with the correlation between the phenotype and genotype (Fig. 3B), these data demonstrate that OsFPN1 is the causal gene of 1187_n.

Osfpn1 Mutants Were Sensitive to High Co and Ni
To investigate whether the sensitivity of osfpn1-1 (1187_n) is also caused by the mutation in OsFPN1, we tested the tolerance of osfpn1-1 and osfpn1-2, together with their wild type, to Co and Ni (Fig. 4). Without Co or Ni, the shoot and root growth of osfpn1-1 and osfpn1-2 were similar to that of the wild type. Under 10 μM Ni, the shoot and root growth of osfpn1-1 and osfpn1-2 were inhibited. These results indicated that OsFPN1 is important for high Co and Ni tolerance.

OsFPN1 is Localized to the Golgi Apparatus
OsFPN1 is a membrane protein with nine putative transmembrane domains (Additional file 1: Fig. S2).
To determine the subcellular localization of OsFPN1, OsFPN1-GFP or GFP-OsFPN1 were driven by the CaMV 35S RNA promoter and transiently expressed in protoplasts prepared from rice leaf sheaths. In the protoplast expressing OsFPN1, a dot-like structure was observed in both the N-and C-terminal fusion proteins. The dotlike structure was co-localized with the Golgi marker OsCTL1 (Wu et al. 2012) (Fig. 5). Therefore, OsFPN1-GFP was localized to the Golgi apparatus.

OsFPN1 can Transport Co and Ni in Yeast
To test whether OsFPN1 can transport Co and Ni, GFP-OsFPN1 was expressed in Co-and Ni-sensitive yeast cot1 mutants (Schaaf et al. 2006;Morrissey et al. 2009). COT1p is a Zn and Co transporter localized in vacuoles, and cot1 mutants are sensitive to high Co concentrations (Conklin et al. 1992). In the following experiments, COT1 was used as a positive control and GFP-fused COT1 (GFP-COT1) was introduced into the yeast. GFP fluorescence was observed in the yeast transfected with GFP-COT1 and GFP-OsFPN1, suggesting that the genes were translated (Fig. 5A). Based on the pattern of GFP, GFP-COT1 could be localized to the vacuole, which is consistent with a previous report (Morrissey et al. 2009), and GFP-OsFPN1 was observed in the dot-like structure and cell periphery (Fig. 6A).
To determine the Co tolerance of cot1 carrying each construct, the yeasts were spotted onto SD-uracil free media containing a high (1 mM) Co concentration. Consistent with a previous report (Morrissey et al. 2009), COT1-GFP conferred tolerance to 1 mM of Co (Fig. 6B). However, the expression of OsFPN1-GFP increased Co sensitivity when compared to the empty vector (Fig. 6B). We also determined the Co concentrations of yeast carrying GFP-OsFPN1 after incubation in liquid media. The yeast carrying GFP-OsFPN1 exhibited a lower Co concentration than that carrying the empty vector, while it was higher in the yeast carrying GFP-COT1 (Fig. 6C). These results suggest that OsFPN1 can transport Co in yeast. We also determined the Ni concentration in the yeasts after incubation with SD-uracil free media containing Ni, and the yeast carrying GFP-OsFPN1 exhibited a lower Ni content than that carrying the empty vector (Fig. 6C), suggesting that OsFPN1 can transport both Ni and Co in yeast.

OsFPN1 is Involved in Co and Ni Translocation, Particularly Under Low-Fe Conditions
In Arabidopsis, two genes are involved in Co and Ni transport: AtFPN1 and AtFPN2 (Morrissey et al. 2009). The mRNA expression of FPN2 is upregulated by low levels of Fe (Morrissey et al. 2009), whereas that of AtFPN1 is not (Morrissey et al. 2009). AtFPN2 can sequester Co and Ni into vacuoles in roots, while AtFPN1 can export Co to the shoot. To confirm whether OsFPN1 mRNA is regulated by Fe, Co, and Ni, rice was grown under 0, 9, and 90 μM of Fe, or 0 and 50 μM of Co or Ni in Kimura B solution and the OsFPN1 mRNA levels were quantified. Unlike AtFPN2, the mRNA expression of OsFPN1 was not altered by Fe, Co, or Ni (Fig. 7A).
In Arabidopsis, the difference in the Co concentration in the shoots of the wild type and fpn2 increased under low-Fe conditions (Morrissey et al. 2009). To investigate whether Fe affected Co and Ni accumulation between the wild type and osfpn1-1, we analyzed the accumulation of Co and Ni in their shoots and roots under various Fe conditions. There was no large difference in growth under the tested conditions (Additional file 1: Fig. S5). Co was higher in osfpn1-1 under most of the Fe conditions than in the wild type (Fig. 7B). However, the opposite scenario was observed in the roots. The difference in Co accumulation in the shoots between the wild type and osfpn1-1/1187_n was larger under 0 μM Fe conditions than under normal Fe (90 µM) conditions, as well as the difference in the Co accumulation in roots (Fig. 7B). A similar pattern was observed for Ni (Fig. 7C). These results indicated that OsFPN1 is involved in the root-to-shoot translocation of Co and Ni, particularly under low-Fe conditions.

Discussion
Through ionome screening, we isolated osfpn1-1 and identified OsFPN1 as a gene responsible for Co and Ni accumulation, as well as for tolerance to high Co or Ni conditions. It has been reported that plant FPN has a wide substrate specificity. For instance, Arabidopsis FPN2 and poplar PgIREG1 can transport Co or Ni (Morrissey et al. 2009;Merlot et al. 2014;Yokosho et al. 2016). In buckwheat, FeIREG1 is involved in Al detoxification (Yokosho et al. 2016). In this study, heterologous expression of GFP-OsFPN1 in yeast increased the sensitivity to high concentrations of Co (Fig. 6B). Furthermore, the yeast experiment suggested the Co and Ni transport activities of OsFPN1 (Fig. 6C). These results suggest that OsFPN1 can transport Co and Ni in rice.
In the yeast experiments, the yeast carrying OsFPN1 exhibited increased sensitivity to high Co and accumulated less Co and Ni than the empty vector. In a previous report on Arabidopsis FPN, it was suggested that FPN proteins could be transported from the cytosol to the outside of the cell or into the vacuole (Morrissey et al. 2009). OsFPN1 in yeast is localized to a dot-like structure, as well as the plasma membrane (Fig. 6A). Considering the direction of transport and plasma membrane localization, OsFPN1 transports Co 2+ and Ni 2+ to the outside of the cell, leading to low Co 2+ and Ni 2+ contents (Fig. 6C). In contrast, the yeast exhibited high Co sensitivity (Fig. 6B). This is contradictory to the result of the yeast Co and Ni concentrations because, if Co is exported to the outside of the cell, the yeast would exhibit tolerance, similar to the COT1-expressing yeast. The reason for this is currently unknown, but may be due to the high accumulation of Co in a dot-like organelle, such as the Golgi. Inside the organelle, the Co 2+ and Ni 2+ contents would be sufficiently high to inhibit the reaction, leading to growth inhibition.
OsFPN1 was localized to the Golgi when transiently expressed in protoplasts prepared from rice (Fig. 5). In Arabidopsis, FPN1 and FPN2 are localized to the plasma membrane and vacuole, respectively (Schaaf et al. 2006;Morrissey et al. 2009). In our experiments, both C-terminal and N-terminal fusion proteins were localized to the Golgi in rice, and GFP-OsFPN1 exhibited Co and Ni transport activity in yeast (Figs. 5, 6). These results indicate that GFP-OsFPN1 is functional and that the fusion protein reflects the function and localization of endogenous OsFPN1. In plants, there are currently no reports indicating that the Golgi are a storage site of Co. It has been reported that Co highly accumulates in the perinuclear fraction of human keratinocytes, and can accumulate in the endoplasmic reticulum or Golgi (Ortega et al. 2009). Therefore, the Golgi apparatus might be a storage site for Co in rice.

Conclusion
Our study shows that OsFPN1 mediates Co and Ni transport, and is involved in the detoxification of Co and Ni. From the viewpoint of application, a high Co content in grains and leaves would be useful for the biofortification of grasses as it would resolve Co deficiencies in sheep and cattle.

Plant Materials and Growth Condition
The ionome mutant, hereafter termed 1187_n, was isolated from EMS-mutagenized Oryza sativa cv. Hitomebore in a previous study (Tanaka et al. 2016

Element Determination
The plants were separated into shoots and roots and washed with ultrapure water. After drying at 70 °C for 3 d, the dry weights of the shoots and roots were measured and digested with HNO 3 . After dissolving the digested samples in 0.08 M HNO 3 containing 2 ppb of In (internal control), the element concentrations were determined by Inductively coupled plasma-mass spectrometry (ICP-MS) (Agilent, Agilent 7800).

Genotyping of F 2 Crosses Between 1187_n and Hitomebore
1187_n was backcrossed with wild-type Hitomebore, and a self-fertilized F 2 population was used to confirm the correlation between the phenotype (Co and Ni) and genotype. Genotyping was conducted using dCAPS markers and primers Nos. 1 and 2 (Additional file 1: Table S1).

Plasmid Constructs and Transformation
For the subcellular localization of OsFPN1, coding sequences (CDS) were amplified with primers Nos. 7 and 8 for GFP-OsFPN1, or Nos. 7 and 9 for OsFPN1-GFP (Additional file 1: Table S1). The amplified fragment was ligated into the EcoRI and XhoI sites of the pENTR2Bdual entry vector (Life Technologies) using a DNA ligation kit (Ligation High ver.2, TOYOBO). CDS was cloned into pMDC45 or pMDC85 vectors (Curtis and Grossniklaus 2003) using LR clonase (Life Technologies). For OsCTL1-red fluorescent protein (RFP), a Golgi marker, CDS was amplified using primers Nos. 10 and 11 (Additional file 1: Table S1). The CDS was ligated into the XbaI and SalI sites of the pENTR2B-dual entry vector using a DNA ligation kit (Ligation High ver.2, TOYOBO). CDS was cloned into the destination vector pGWB554 (Nakagawa et al. 2009), which contains RFP, using LR clonase (Life Technologies).
To generate CRISPR/Cas9 mutant lines of OsFPN1, 20-nt (primers Nos. 14 and 15; Additional file 1: Table S1) targeting the gene was annealed and introduced into the pU6gRNA vector (Mikami et al. 2015). pU6gRNA was digested with AscI and PacI, and the digested fragment was inserted into the AscI and PacI sites of pZH_OsU6g-RNA_MMCas9 (Mikami et al. 2015). The construct was transformed into Agrobacterium tumefaciens EHA105 and used for rice transformation (Ozawa 2012). Oryza sativa cv. Nipponbare was used for transformation in all experiments, and mutations in OsFPN1 CRISPR/Cas9 mutants were identified using primers Nos. 16 and 17 by direct sequencing (Additional file 1: Table S1).
To generate an expression vector expressing GFP-COT1p in the yeast experiments, the GFP fragment was amplified with primers Nos. 18 and 19, and the CDS of COT1 was amplified with primers Nos. 12 and 13 (Additional file 1: Table S1). GFP and COT1 fragments were directly cloned into the XhoI and SalI sites of pKT10myc (Tanaka et al. 1990) using an In-Fusion ® HD Cloning Kit (TOYOBO). GFP-OsFPN1 was amplified with primers Nos. 20 and 21 (Additional file 1: Table S1) from pMDC45 carrying GFP-OsFPN1, and the fragment was cloned into the EcoRI and SalI sites of pKT10-myc using a DNA ligation Kit (Ligation High ver.2, TOYOBO).

Confocal Laser Scanning Microscopy Observation
Plasmids of GFP-OsFPN1 or OsFPN1-GFP were cotransformed with OsCTL1-RFP into 10-d-old rice protoplasts prepared from leaf sheaths grown in 0.5 × MS medium following the protocol described by Zhang (Zhang et al. 2011). After incubation in W5 solution (154 mM NaCl, 125 mM CaCl 2 , 5 mM KCl, and 2 mM MES at pH 5.7) at 25 °C for 16-20 h (Wu et al. 2012), the fluorescence of GFP and RFP was observed by confocal microscopy (FV1000, Olympus). The excitation and emission wavelengths were set at 473 nm and 500-525 nm for GFP and 559 and 600-660 nm for RFP, respectively.

Yeast Growth Assay and Element Uptake Experiments
The plasmids were introduced into Co-and Ni-sensitive mutants, cot1 (MATa,ura3Δ0,leu2Δ0,his3Δ1,met15Δ0,and YOR316c::kanMX4) (Conklin et al. 1992). After transformation, two independent colonies were isolated and used in the experiments. The yeast was cultivated overnight in SD uracil-free medium, and the OD 600 was adjusted to 1.0. After a series of dilutions, the yeast was spotted onto SD uracil-free media supplemented with 1-mM Co or without Co and incubated at 30 °C.
To determine the elements in yeast, the yeast was incubated overnight and adjusted to the same optical density (OD 600 = 0.3), and then incubated in SD uracil-free liquid medium supplemented with 50 μM of Co or 50 μM of Ni for 1 h. After incubation, the yeast cells were washed three times with cold ultrapure water and dried overnight at 70 °C. The dried cells were then digested with HNO 3 at 70 °C for 1 h. After dilution with 0.08-M HNO 3 containing 2 ppb of In, the samples then underwent ICP-MS analysis.

RNA Extraction and Expression Analysis
The total RNA was extracted from 10-d-old seedlings using an RNeasy Plant Mini Kit (Qiagen) following the manufacturer's instructions. Approximately 500 ng of the total RNA was reverse-transcribed to cDNA using a Pri-meScript RT reagent kit (Takara), which was then diluted 10 times and used for quantitative (q) PCR using TB Green ® Premix Ex TaqTM II (Tli RNaseH Plus, Takara). OsActin1 (Zhang et al. 2004) was used as an internal control for qPCR. Nos. 3 and 4, and Nos. 5 and 6 primers were used for the qPCR of OsFPN1 and OsActin1, respectively (Additional file 1: Table S1).

Statistical Analysis
The sample size is described in the figure legend and statistical analysis was performed with GraphPad Prism software (https:// www. graph pad. com/ scien tific-softw are/ prism/).