LjMOT1, a high‐affinity molybdate transporter from Lotus japonicus, is essential for molybdate uptake, but not for the delivery to nodules

Summary

Molybdenum (Mo) is an essential nutrient for plants, and is required for nitrogenase activity of legumes. However, the pathways of Mo uptake from soils and then delivery to the nodules have not been characterized in legumes. In this study, we characterized a high‐affinity Mo transporter (LjMOT1) from Lotus japonicus. Mo concentrations in an ethyl methanesulfonate–mutagenized line (ljmot1) decreased by 70–95% compared with wild‐type (WT). By comparing the DNA sequences of four AtMOT1 homologs between mutant and WT lines, one point mutation was found in LjMOT1, which altered Trp²⁹² to a stop codon; no mutation was found in the other homologous genes. The phenotype of Mo concentrations in F₂ progeny from ljmot1 and WT crosses were associated with genotypes of LjMOT1. Introduction of endogenous LjMOT1 to ljmot1 restored Mo accumulation to approximately 60–70% of the WT. Yeast expressing LjMOT1 exhibited high Mo uptake activity, and the K_m was 182 nm. LjMOT1 was expressed mainly in roots, and its expression was not affected by Mo supply or rhizobium inoculation. Although Mo accumulation in the nodules of ljmot1 was significantly lower than that of WT, it was still high enough for normal nodulation and nitrogenase activity, even for cotyledons‐removed ljmot1 plants grown under low Mo conditions, in this case the plant growth was significantly inhibited by Mo deficiency. Our results suggest that LjMOT1 is an essential Mo transporter in L. japonicus for Mo uptake from the soil and growth, but is not for Mo delivery to the nodules.

Introduction

Molybdenum (Mo) is an essential micronutrient for nearly all living organisms (Arnon and Stout, 1939; Johnson et al., 1980; Turnlund, 2002), and is taken up by living organisms mostly as molybdate. However, molybdate is biologically inactive and must be complexed with pterin to form Mo‐cofactor (Moco) or with an iron–sulfur cluster to form iron–molybdenum cofactor, which is unique in bacterial nitrogenases (Burgess and Lowe, 1996; Mendel, 1997, 2007). Mocos are used as electron donors and/or acceptors in Mo‐requiring enzymes (molybdoenzymes), such as nitrate reductase, nitrogenase, aldehyde oxidase, sulfite oxidase, xanthine oxidase, and mitochondrial amidoxime reductase (Schwarz and Mendel, 2006; Mendel, 2009; Bjornsson et al., 2015). These enzymes catalyze diverse oxidation–reduction reactions and play key roles in the assimilation and biogeochemical cycles of carbon, nitrogen, and sulfur (Mendel and Schwarz, 1999; Zimmer and Mendel, 1999; Bittner, 2014). Mo deficiency in plants results in a phenotype commonly referred to as ‘whiptail’, which includes mottled lesions on and rolling of the leaves, as well as wilting of leaf edges (Hewitt and Bolle‐Jones, 1952; Gupta, 1997; Nautiyal and Chatterjee, 2004; Kaiser et al., 2005).

Mo is of particular importance in legumes and is required by enzymes involved in the nitrogen assimilation process, such as those involved in nitrate reduction, nitrogen fixation, and transport of nitrogen compounds (Mulder, 1948; Anderson and Spencer, 1950; Srivastava, 1997). Mo‐containing nitrate reductase catalyzes the reduction of nitrate (NO₃⁻) into nitrite (NO₂⁻), which is the first and essential step in nitrate assimilation (Campbell, 1999; Delgado et al., 2003). Mo‐containing nitrogenase catalyzes the reduction of atmospheric nitrogen (N₂) into ammonia (NH₃) in nodules of legumes (Burgess and Lowe, 1996; Lawson and Smith, 2002; Stüeken et al., 2015). In addition, xanthine dehydrogenase/oxidase, another molybdoenzyme, is thought to be required for the mobilization and export of fixed nitrogen out of the nodules (Kaiser et al., 2005). These studies suggest that Mo is required for nitrogen fixation and mobilization in nodules of legumes, and Mo is considered to be more important for legumes than for non‐legumes (Bambara and Ndakidemi, 2010). Mo application benefits the growth of soybeans and other legumes, particularly in soils with low pH (Parker and Harris, 1977; Adams, 1997).

For organisms to synthesize molybdoenzymes, they must take up Mo from the growth medium. Mo transporters are well characterized in Escherichia coli, where Mo transport is mediated by the high‐affinity ABC‐type transport system encoded by modABC genes (Grunden and Shanmugam, 1997; Self et al., 2001; Hollenstein et al., 2007). Although ABC transporters are present in eukaryotes, ABC‐type Mo‐specific transporters have not been identified in eukaryotes (Kaiser et al., 2005). Therefore, in eukaryotes, Mo is thought to be transported by other transporters, such as those involved in sulfate and/or phosphate uptake (Heuwinkel et al., 1992; Alhendawi et al., 2005; Shinmachi et al., 2010; DeTar et al., 2015). Mo transporters in Chlamydomonas reinhardtii (MoT1 and MoT2) and Arabidopsis thaliana (AtMOT1 and AtMOT2) were shown to transport Mo with high affinity (Tejada‐Jiménez et al., 2007, 2011; Tomatsu et al., 2007; Baxter et al., 2008; Li et al., 2009; Gasber et al., 2011). AtMOT1 belongs to the sulfate transporter superfamily and is essential for A. thaliana's uptake of Mo from soil (Tomatsu et al., 2007; Baxter et al., 2008). The T‐DNA mutant of AtMOT1 accumulates significantly less Mo in roots and shoots than does the wild‐type (WT), and shows Mo deficiency symptoms when grown under low Mo conditions (Tomatsu et al., 2007; Ide et al., 2011). Recently, a putative molybdate transporter LjMOT1 was reported may be involved in Mo transport in Lotus japonicas, since the Mo concentration in shoots was positively correlated to the expression of LjMOT1 (Gao et al., 2016). However, the Mo transport activity of the putative LjMOT1 was not tested either in yeast or in planta, and the effects on the process of nodule symbiosis was not investigated either (Gao et al., 2016).

In the present study, we identified a high‐affinity Mo transporter, LjMOT1, through an ethyl methanesulfonate‐mutagenized line. LjMOT1, belonging to the sulfate transporter superfamily and expressing mainly in the epidermis and endodermis of roots. The Mo transport activity of LjMOT1 was characterized by expressing in yeast and planta, including the mutant ljmot1, the F2 progeny from ljmot1 and WT crosses, and LjMOT1 complemented plants. The effects of LjMOT1 on the process of nodule symbiosis were also investigated. Our results demonstrated that LjMOT1 is essential for L. japonicus to take up Mo from the soil. However, disruption of this transporter did not significantly affect the process of nodule formation and nitrogen fixation.

Results

An L. japonicus mutant with low Mo accumulation

The low Mo‐accumulating mutant (the 4–22 line) was obtained by ionomic screening of an ethyl methanesulfonate (EMS)‐mutagenized M₂ population of L. japonicus MG20 (Chen et al., 2009). Because this mutant has a mutation in LjMOT1, as described below, it is referred as ljmot1 hereafter. When plants were grown in vermiculite supplied with 0.1 μm Mo, the Mo concentrations in shoots and roots of the mutant decreased by about 90% compared with the corresponding tissues of WT; in the nodules, the concentration decreased by approximately 75% (Figure 1a). To further investigate the patterns of Mo accumulation in the mutant, plants were grown on agar plates supplied with various Mo concentrations (0, 0.17, 1, and 10 μm; Figure 1b). The results showed that Mo accumulation in the mutant was about 5–10% that in WT when the plants were supplied with 0, 0.17, or 1 μm Mo, and about 30% that in WT when the plants were supplied with 10 μm Mo. After maturation of plants grown in vermiculite supplied with 0.1 μm Mo, Mo concentration in seeds was determined, results showed that Mo concentration in the seeds of WT was about three‐fold higher than that in the seeds of ljmot1 (Figure 1c). The accumulation of other elements, i.e., Mg, K, Mn, Fe, Cu, and Zn, was similar between the mutant and WT (Figure 1d).

Figure 1

Accumulation of Mo and other elements in ljmot1 and MG20.

(a) Mo concentrations in seedlings grown in vermiculite irrigated with half‐strength B&D solution plus 0.17 μm Mo. Plants were harvested 1 month after inoculation with Mesorhizobium loti. Different letters indicate significant differences of the average Mo concentration between ljmot1 and the WT wild‐type (P < 0.01); small letters a, b: nodules; capital letters A, B: roots; small letters a′, b′: shoots.

(b) Mo concentrations in 10‐day‐old seedling grown on agar plates containing half‐strength B&D solution, 2 mm KNO₃, and 0, 0.17, 1, or 10 μm Mo. Asterisks (**) indicate significant difference of the average Mo concentration between ljmot1 and the WT wild‐type (P < 0.01).

(c) Mo concentration in seeds. Three days after transplanting to vermiculite, the seedlings were inoculated with M. loti. Four weeks after inoculation, the seeds were harvested to determine Mo concentration. ^a,bDifferent letters indicate significant differences between ljmot1 and the WT wild‐type (P < 0.01).

(d) Concentration of Mo and other elements in ljmot1 relative to the WT line. Seedlings were grown in hydroponic solution for 2 weeks (half‐strength B&D solution plus 2 mm KNO₃, and 0.17 μm Mo).

Data are means ± standard deviation (SD) (n = 4). Asterisks (**) indicate significant difference of the average Mo concentration between ljmot1 and the WT wild‐type (P < 0.01).

Identification of the causal mutant gene responsible for low Mo accumulation

To identify the candidate gene responsible for the low Mo‐accumulating phenotype in the mutant, an AtMOT1 homolog search was performed against the L. japonicus genome database (http://www.kazusa.jp/lotus/index.html). The search identified four homologs: LjMOT1, LjMOT2, LjMOT3, and LjMOT4 (Figure S1). By comparing the DNA sequences of these homologs in the mutant and WT, a single nucleotide substitution was found at the second exon of LjMOT1, which resulted in the substitution of a stop codon for the amino acid Trp²⁹² (W²⁹²Stop; Figure 2a). No mutation was found in the other three LjMOT genes in the mutant.

Figure 2

Identification of the causal gene of the low Mo‐accumulating mutant, ljmot1.

(a) Exon–intron structure of the LjMOT1 gene. Exons are shown as black boxes. The mutation was found in the second exon of LjMOT1 (arrow), where the TGG codon encoding W²⁹² was changed to a TGA stop codon.

(b) Mo concentration of F2 plants. F2 progeny and parent plants were grown in hydroponic solution for 2 weeks (half‐strength B&D solution plus 2 mm KNO₃, and 0.17 μm Mo). Data are shown as boxplots, in which the boundary of boxes indicate the 25th and 75th percentile, the line within the box marks the median, error bars indicate the 10th and 90th percentiles, points show individual plants. ^a,b,cDifferent letters indicate significant differences of the average Mo concentration of each genotype (P < 0.01).

To confirm that LjMOT1 is the causal gene for low Mo accumulation of ljmot1, the relationship between phenotype and genotype was investigated. F₂ progeny from crosses between the mutant and WT was established, and the genotypes and Mo concentrations of 78 F₂ seedlings were determined. Among the 78 F₂ plants, 20 plants were homozygous for WT LjMOT1, 37 plants had heterozygous LjMOT1, and 21 plants were homozygous for the mutate allele, ljmot1, at a ratio of approximately 1:2:1 (Figure 2b). F₂ progeny with homozygous LjMOT1 mutant and WT genotypes had Mo concentrations similar to the respective parental ljmot1 and WT plants, and the Mo concentrations of heterozygous plants were intermediate between those of the mutant and WT. A significant linkage between the LjMOT1 genotype and Mo concentration was observed in the F₂ progeny. This result indicated that the mutation was semi‐dominant, and that the low Mo‐accumulating phenotype in the mutant is likely caused by a single mutation in the LjMOT1 gene.

LjMOT1 partially complements Mo accumulation in ljmot1

To further confirm that mutation in the LjMOT1 gene was responsible for the low Mo‐accumulating phenotype, the mutant plants were transformed with a WT LjMOT1 genomic fragment, including the putative promoter and terminator regions. Five independent hygromycin‐resistant transgenic lines were obtained, two of which did not produce enough seeds for subsequent Mo exposure experiments. Therefore, the following genotyping and Mo exposure experiments were conducted with the three lines that produced enough seeds.

To verify integration of the WT LjMOT1 gene into the mutant plants and ensure that it was expressed in these three transgenic lines, the mRNA genotype of LjMOT1 in each transgenic line was determined using dCAPS markers with cDNA as a template. In the polymerase chain reaction (PCR) product from each transgenic line, mutated and WT LjMOT1 products were detected, indicating that the WT LjMOT1 was expressed successfully in these three independent transgenic lines. However, LjMOT1 in all these three lines are heterozygous, no homozygous line was obtained (Figure 3a).

Figure 3

LjMOT1 partially complemented the low Mo‐accumulating phenotype in the mutant.

(a) Introduction of LjMOT1 to the mutant was confirmed by dCAPS marker analysis. The shorter band indicates wild‐type LjMOT1, longer bands indicates mutated LjMOT1, and both bands indicate transgenic heterozygous lines. M indicates DNA marker, and C1, C2, C3 indicate complemented lines.

(b) Mo concentrations in the wild‐type, mutant, and complemented plants. Plants were grown in hydroponic solution for 2 weeks (half‐strength B&D solution plus 2 mm KNO₃, and 0.17 μm Mo). Data are means ± standard deviation (SD) (n = 4). Different letters indicate significant differences (P < 0.01); small letters a–c: roots; capital letters A–C: shoots.

The transgenic, mutant, and WT lines were grown in hydroponic solution containing 0.17 μm Mo for 2 weeks, after which Mo concentrations were compared (Figure 3b). The concentrations in the roots and shoots of transgenic lines were intermediate between those of the mutant and WT plants. The hierarchy of Mo accumulation among the different genotypes of LjMOT1 (WT > LjMOT1 transgenic lines > ljmot1) was comparable to that observed in the F₂ population. These data further suggested that LjMOT1 is the causal gene for low Mo accumulation of ljmot1.

Mo transporter activity of LjMOT1 in yeast

To examine Mo transport activity, LjMOT1, ljmot1 (W²⁹²Stop), and AtMOT1 were introduced into Saccharomyces cerevisiae using the pYES2 expression vector, which allows expression of the inserted gene under control of the galactose‐inducible GAL1 promoter. After growing to mid‐log phase, the transformed yeast cells were treated with 170 nM Mo for 30 min, after which Mo concentrations in cells were determined (Figure 4a). When grown with glucose, all transformants exhibited low Mo accumulation. When the glucose was replaced with galactose in the medium, the yeast cells carrying LjMOT1 showed 100‐fold higher Mo concentrations than those carrying empty vector or LjMOT1 (W²⁹²Stop). These results indicated that LjMOT1 is a Mo uptake transporter and that the mutation of Trp²⁹² into a stop codon results in loss of Mo uptake activity.

Figure 4

Mo transport activity of LjMOT1 in Saccharomyces cerevisiae

(a) Mo concentrations in yeast cells carrying different constructions. Yeast cells were incubated in selective media containing glucose or galactose supplemented with 170 nm Mo for 30 min. Different letters indicate significant differences (P < 0.01); small letters a–c: glucose; capital letters A–C: galactose.

(b) Kinetic Mo uptake of yeast cells expressing LjMOT1. Yeast cells were incubated in selective medium containing galactose with 56, 84, 112, 168, or 679 nm Mo for 15 min.

(c) Mo uptake kinetics parameter of LjMOT1 expressed in yeast cells. The K_m was 182 nm, and V_max was 4.83 μg g⁻¹ DW min⁻¹. Data are means ± standard deviation (SD) (n = 3).

The Mo uptake kinetics of LjMOT1 was investigated by incubating yeast cells carrying LjMOT1 in galactose‐containing medium with various concentrations of Mo for 15 min (Figure 4b). Mo uptake activity followed Michaelis–Menten kinetics (Figure 4b). Double‐reciprocal plotting showed that the K_m for Mo uptake was 182 nm, and the V_max was 4.8 μg g⁻¹ DW min⁻¹ (r = 0.998; Figure 4c).

Regulation of LjMOT1 expression

To investigate the regulation of LjMOT1 expression, real‐time PCR was performed using different plant tissues grown under different Mo conditions. First, to study the effects of Mo supply on the expression of LjMOT1, mRNA of LjMOT1 was quantified in roots and shoots from plants treated with different Mo concentrations; namely, 0, 0.17, 1, and 10 μm. LjMOT1 was expressed mainly in roots, where expression was about 20 times greater than in shoots, and the expression of LjMOT1 in roots was not significantly induced by the conditions examined (Figure 5a). In shoots, the expression of LjMOT1 was statistically significantly induced by high Mo supply (1 and 10 μm Mo) (P < 0.01). Second, to examine the effects of rhizobium inoculation on LjMOT1 gene expression, the relative amounts of LjMOT1 transcripts in roots and nodules were determined at different stages of nodule development (Figure 5b). The expression of LjMOT1 in roots was about 10‐fold greater than that in nodules, and the expression of LjMOT1 in roots or nodules was not significantly affected by nodule development (Figure 5b).

Figure 5

Regulation of LjMOT1 expression.

(a) Relative expression of LjMOT1 in Lotus treated with 0, 0.17, 1, or 10 μm Mo for 48 h. The mRNA levels were normalized to that of the ubiquitin gene and are expressed as ratios relative to the value observed in shoots treated with 0 μm Mo. Different letters indicate significant differences (P < 0.01); small letters a, b: roots; capital letters A, B: shoots. Data are means ± standard deviation (SD) (n = 4).

(b) Relative expression of LjMOT1 in roots and nodules sampled at different days after inoculation with M. loti. Uninoc indicates roots which were not inoculated. The mRNA levels were normalized to that of the ubiquitin gene and are expressed as ratios relative to the value observed in un‐inoculated roots. Data are means ± SD (n = 4).

(c–e) GUS activity in transgenic roots and nodules carrying a LjMOT1‐promoter‐GUS construct. Root and nodule (c), root section (d) and nodules section (e) are shown. Scale bars: (c) 1 mm; (d) 50 μm; (e) 200 μm. [Colour figure can be viewed at wileyonlinelibrary.com].

Because LjMOT1 is expressed mainly in roots, the tissue and cell specificity of its expression was examined in roots and nodules (Figure 5c). The construct containing the putative promoter region (−2106 to −1 from the first ATG) and terminator region (+1 to +940 from the stop codon) was fused with E. coli β‐glucuronidase (GUS) and introduced into L. japonicus roots using the hairy‐root transformation system. GUS activity was detected in roots and nodules (Figure 5c). In roots, high GUS activity was detected in the exodermis and endodermis, and the stele region of roots, probably in xylem vessels (Figure 5d). In nodules, GUS activity was detected in the epidermis, outer cortex, and vascular bundle, but not in infected cells (Figure 5e).

Plant growth phenotype, nodulation, and acetylene reduction activity of ljmot1

The growth of ljmot1 was indistinguishable from that of WT when plants were grown in vermiculite, a hydroponic system, and a sealed container supported with a urethane form (Figure S3). In addition, under all growth conditions, the growth of ljmot1 and WT was not significantly affected by Mo supply (Figure S3).

To investigate the effects of LjMOT1 mutation on nitrogen‐fixing symbiosis, nodule morphology was compared between ljmot1 and WT. The ljmot1 mutant formed nodules under no Mo addition (−Mo) and normal Mo addition (+Mo) conditions. Nodules that formed on the mutant were similar to those that formed on WT in terms of features including size, color, and weight under all growth conditions tested (Figures S3 and S4a, b).

Acetylene reduction activity (ARA) was measured to evaluate nitrogenase activity. The ARAs of ljmot1 and WT were measured 14, 21, and 28 days after inoculation with M. loti (Figure S4c). Our results showed that nitrogenase activity of nodules formed on ljmot1 was similar to that of WT nodules.

Removal of cotyledons affects growth, but not nodulation, of ljmot1 under −Mo conditions

Mo stored in L. japonicus seeds may be utilized for plant growth, especially at the early seedling stage, which may be sufficient for initial growth even without Mo supplementation. To investigate the effects of Mo in seeds on plant growth, cotyledons were removed from seedlings at 10 days after germination to avoid possible Mo supply from cotyledons. Under −Mo conditions, cotyledon removal significantly inhibited the growth of ljmot1 and WT compared with cotyledon retention (Figure 6, see also Figure S3). When the cotyledons were removed, without Mo supply could significantly inhibited the growth of both ljmot1 and WT, i.e., the fresh weights of plants grown without Mo addition were decreased significantly compared with those of plants grown under 0.17 μm Mo treatment (Figure 6a, b). In addition, when the cotyledons were removed, the growth of ljmot1 was significantly suppressed compared with that of WT under both Mo and +Mo conditions, (Figure 6a, b). However, the growth inhibition under low Mo conditions and phenotype differences between ljmot1 and WT were not observed when the cotyledons were not removed (Figure S3).

Figure 6

Growth phenotype, ARA activity and Mo concentration of cotyledons‐removed ljmot1 and WT plants.

(a–d) The plants were grown in a sealed container supporting with urethane form watered with half‐strength B&D solution supplemented with (0.17 μm) or without (0 μm Mo) Mo. After 21 days inoculation plant growth and nodules (a), fresh weight (b), ARA activity (c), Mo concentration (d) were measured. Data are means ± standard deviation (SD) (n = 6). [Colour figure can be viewed at wileyonlinelibrary.com].

Although the growth of ljmot1 without cotyledons was significantly inhibited, the process of nodulation was not significantly affected by LjMOT1 mutation. Both ljmot1 and WT formed nodules on their roots irrespective of Mo conditions, and the morphology of nodules of ljmot1 was similar to that of WT nodules (Figure 6a). In addition, no significant difference in ARA was observed between ljmot1 and WT under low or normal Mo conditions (Figure 6c). Mo concentrations in nodules, roots, and shoots were significantly lower in ljmot1 than in WT (Figure 6d).

Discussion

Mo is an essential nutrient for plants and is of particular importance for legumes (Anderson and Spencer, 1950; Gupta, 1997). However, the Mo transport system in leguminous plants has not been described previously. Chen et al. (2009) isolated one low Mo‐accumulating mutant of L. japonicus (line 4‐22) using ionomic screening. Mo accumulation in the mutant decreased by 70–95% in the whole plant, including nodules, roots, shoots (Figures 1a, b and 6d), and seeds (Figure 1c). To identify the causal gene for the low Mo‐accumulating mutant, we compared DNA sequences of AtMOT1 homologs in mutant and WT lines. A single nucleotide substitution was found at the second exon of LjMOT1 (Figure 2a), and no mutation was found in the other three AtMOT1‐like genes. To verify that LjMOT1 is the causal gene for low Mo accumulation, the mutant plants were transformed with a WT LjMOT1 genomic fragment, including the putative promoter and terminator regions, and results showed that heterozygous complementation partially restored Mo accumulation in mutant plants (Figure 3b). In addition, in F₂ progeny from crosses between the mutant and WT, the Mo‐accumulating phenotype was clearly associated with the genotype of LjMOT1 (Figure 2b). These results suggest that LjMOT1 is the causal gene for the low Mo‐accumulating mutant, and the major gene controlling Mo accumulation in L. japonicus.

LjMOT1 showed high affinity for Mo uptake, with a K_m of 182 nm (Figure 4). This low K_m value is similar to the level observed in AtMOT1 (Tomatsu et al., 2007), which suggests that the high affinity of LjMOT1 for Mo may be essential to obtain scarce Mo from soil environments. This explains why the LjMOT1 mutation resulted in a more significant decrease in Mo accumulation under low Mo treatments than under high Mo treatments; i.e., under low Mo conditions, Mo concentrations in ljmot1 decreased by 90% compared with WT, while under high Mo conditions (10 μm Mo treatment), Mo concentrations decreased by about 70% (Figure 1b). These results indicate that under high Mo conditions, low‐affinity Mo uptake transporter(s) should be present in L. japonicus. In Stylosanthes hamata, the sulfate transporter SHST1 transports Mo (Fitzpatrick et al., 2008). Shinmachi et al. (2010) also suggested that sulfate transporters are involved in Mo uptake of wheat plants. Therefore, under high Mo conditions, sulfate transporter(s) or homologs of LjMOT1 may be involved in Mo uptake and distribution in L. japonicus.

Although mutation in LjMOT1 resulted in a 70–95% reduction in Mo accumulation in plant roots and shoots, ljmot1 plants showed no characteristic symptom of Mo deficiency, such as pale‐green to yellow leaves or reduced biomass production, even for long‐term growth under Mo‐deficient conditions (Figure S3). In contrast, the A. thaliana mot1‐1 mutant showed yellow leaves, retarded growth, and a significant decrease in fresh weight under low Mo conditions (Tomatsu et al., 2007; Ide et al., 2011). This phenotypic difference between ljmot1 and mot1‐1 may be due to the large size of L. japonicus seeds, which may contain sufficient Mo for the growth of young seedlings (Figure 1c). To reduce the influence of Mo stored in seeds on growth, seedlings without cotyledons were exposed to low and normal Mo conditions. In this case, significant growth reduction of ljmot1 was observed by comparing with WT, both under low and high Mo conditions (Figure 6a, b). These results suggest that Mo from large seeds of ljmot1 can satisfy plant growth needs even under low Mo conditions.

Mo is involved in nitrogen assimilation in plants, and Mo application increased the number and weight of nodules, as well as NA and NR enzyme activity, in soybean and hairy vetch (Kanaan et al., 2013; Alam et al., 2015). Therefore, the effects of LjMOT1 mutation on processes of nodulation and nitrogenase activity were investigated. However, we found that ljmot1 exhibited no significant defect in the processes of nodulation or ARA (representing nitrogenase activity) compared with WT (Figures S3 and S4). Even for ljmot1 plants without cotyledons grown under low Mo conditions, the processes of nodulation and nitrogenase activity were not significantly affected (Figure 6a, c), although plant growth was inhibited under these conditions (Figure 6a, b). These results indicate that LjMOT1 is not essential for the process of symbiosis in L. japonicus. However, this does not mean that Mo is not essential for symbiosis, because we found that Mo accumulation in nodules of ljmot1 remained relatively high enough for normal nodulation and nitrogenase activity, even under low Mo conditions, which was about 10‐fold higher than that in roots and shoots (Figures 1a and 6d). This result suggests that nodules have the priority of Mo accumulation compared with other parts of plants in L. japonicus when growing under low Mo conditions. These findings are in agreement with those of Brodrick and Giller (1991), who reported that Mo can be translocated from roots and shoots to nodules for nitrogen fixation when it is scarce in plants.

Under low Mo conditions, LjMOT1 mutated plants exhibited high Mo accumulation in nodules and normal processes of nodulation and nitrogenase activity. These results advance our understanding of Mo distribution in L. japonicus, i.e., the pathway of Mo delivery to nodules is different from that of Mo delivery to plant bodies. LjMOT1 is involved in Mo delivery to plant bodies, but not delivery to the symbiosome directly. This result agrees with the expression pattern of LjMOT1, which is mainly localized at the exodermis, endodermis and stele of roots (Figure 5c, d), function as Mo uptake transporter. While LjMOT1 is undetectable in infected cells of nodules (Figure 5e), in the epidermis of nodules, the expression of LjMOT1 was less than one‐tenth of that in roots (Figure 5b), and its expression was not affected by rhizobium inoculation or nodule development (Figure 5b). All these patterns of Mo accumulation and gene expression in nodules suggest that LjMOT1 should not be the direct transporter for Mo loading to nodules, and other transporter(s) is (are) required. However, the mechanism of Mo transport into the cortex and across the symbiosome membranes enclosing endosymbionts remains to be investigated. Four AtMOT1 homologs, including LjMOT1, are listed in the available L. japonicus genome database. MOT2 from Arabidopsis has been demonstrated as a vacuolar membrane Mo transporter, which performs Mo delivery to the vacuoles (Gasber et al., 2011). So LjMOT2 could be involved in Mo delivery to the vacuoles as AtMOT2. BLAST search against the NCBI database using the amino acid sequence of LjMOT3 or LjMOT4 as queries showed that most of the genes with high similarity are from legumes. This conservation indicate that LjMOT3 or LjMOT4 may be involved in Mo delivery to the nodules. Future investigation and characterization of these homologs in L. japonicus would increase our understanding of Mo metabolism in legumes.

Experimental procedures

Plant culture and treatment

Seeds of WT L. japonicus MG20 Miyakojima and EMS mutant line 4–22 were kindly provided by Dr Mitsuru Osaki. For germination, seeds were scarified with sandpaper, surface‐sterilized with bleach, and then washed with sterile water. After sterilization, the seeds were soaked in sterile water and shaken gently at room temperature for 24 h. The seeds were then sown on agar (0.8%) plates containing half‐strength B&D solution (Broughton and Dilworth, 1971) plus 2 mm KNO₃ with various Mo concentrations (0, 0.17, 1, or 10 μm). After 10 days, the plants were harvested to determine their Mo concentrations.

For growth under controlled Mo supply conditions, seeds were germinated on agar plates containing half‐strength B&D solution without Mo for 10 days. Cotyledons were removed from seedlings to limit Mo supply from cotyledons 10 days after germination. The seedlings were then transferred to a sealed container, where the plants were supported with a urethane form and watered with half‐strength B&D solution without (0 μm) or with (0.17 μm) Mo. M. loti MAFF303099 was inoculated when plants were transferred to the container. After 21 days of inoculation, the plants were used for nodule morphology observation and Mo concentration determination.

The plants were grown in incubators (28/20°C day/night temperature, 16 h light/8 h dark, light intensity 120 μmol m⁻² sec⁻¹, relative humidity 70%).

Searching for and sequencing putative Mo transporters in L. japonicus

To identify the candidate causal gene for low Mo accumulation in the mutant, the L. japonicus genome database (http://www.kazusa.jp/lotus/index.html; Sato et al., 2008) was searched for putative Mo transporter genes by performing a BLAST search using the amino acid sequence of AtMOT1 as a query. Protein sequences were aligned using the ClustalW2 program (http://www.ebi.ac.uk/Tools/clustalw2/index.html).

The open reading frame fragments of these putative Mo transporters were amplified by PCR from genomic DNA isolated from L. japonicus leaves using the following primers: 5′‐CCTAACATTGACTTTACAACCCC‐3′ and 5′‐GAACCCCAAACAAGCCCTCAA‐3′ for LjMOT1; 5′‐GCACCACACAAAAGTGTTCC‐3′ and 5′‐GCCATTCTCTCACACACTACC‐3′ for LjMOT2; 5′‐GTGTTCTCCATGCTGACTCTAG‐3′ and 5′‐GAGGTAACATGCACCCAAACAC‐3′ for LjMOT3; and 5′‐GGTTGATTTACCTTCCTAACACTG‐3′ and 5′‐GGTACCAGTAACAAAAGCCAGC‐3′ for LjMOT4. The genomic of L. japonicus leaves was isolated by using an DNA isolation kit (Omega, Japan) according to the manufacturer's instructions, and the PCR conditions were 95°C for 5 min, followed by 35 cycles of 95°C for 30 sec, 58°C for 30 sec, and 72°C for 30 sec. The PCR products were sequenced in forward and reverse orientations, and the sequencing results were compared between mutant and WT.

Segregation analysis of F2 progeny from crosses between ljmot1 and WT

Homozygous ljmot1 were crossed with WT, and the resulting F₁ plants were self‐crossed to generate F₂ seeds. About 100 F₂ seeds were germinated on agar plates containing half‐strength B&D solution plus 2 mm KNO₃ and 0.17 μm Mo. After 1 week, genomic DNA was isolated from leaves of F₂ seedlings. Genotyping analysis of LjMOT1 in the F₂ population was performed using the dCAPS marker: 5′‐GTGATGAAATTCTCTAGACATCCTTG‐3′ (forward) and 5′‐CCAACCAAGTTCATCAAACCAACAGTCAC‐3′ (reverse). The underlined C indicates an introduced mutation to generate a StyI site (CCWWGG) only in WT LjMOT1 (Figure S2). The dCAPS markers were designed using the dCAPs FINDER 2.0 program (http://helix.wustl.edu/dcaps/dcaps.html). The PCR products were digested with StyI, after which the digested solutions were electrophoresed on a 3% agarose gel.

After determination of the genotype, F₂ seedlings, along with the parental WT and mutant plants, were transferred to half‐strength B&D hydroponic solution containing 2 mM KNO₃ and 0.17 μm Mo. The hydroponic solution was aerated continuously and renewed every 3 days. After 2 weeks, the plants were harvested and the Mo concentrations were determined.

Complementation of the ljmot1 mutant by the LjMOT1 genome

A 7252 base pair (bp) genomic DNA fragment containing putative LjMOT1 promoter and terminator regions (from 3369 bp before the start codon to 2003 bp after the stop codon) was amplified by PCR using the following primers: 5′‐CACCCTAAGTTTTAGACAAAATGTCTAAGAAGTC‐3′ and 5′‐ ATAATTTGGACCAATGAAAATGCATTAATG‐3′. The PCR fragment was first cloned into the pENTR/D‐TOPO vector and then subcloned into the pGWB1 binary vector (Nakagawa et al., 2007) using Gateway cloning technology (Invitrogen, Carlsbad, CA, USA). The resulting plasmid was then introduced into Agrobacterium tumefaciens EHA101. Agrobacterium infection and calli induction were performed as described by Lombari et al. (2003).

Seeds of T1 transgenic, WT, and ljmot1 lines were germinated on agar plates (half‐strength B&D solution, 2 mm KNO₃, 0.17 μm Mo). After 1 week, total RNA was isolated from the leaves of T1 seedlings and reverse‐transcribed into cDNA. To confirm introduction of the LjMOT1 gene in the mutant, PCR was performed using cDNA as template and the dCAPS markers used in the genotype assay of the F₂ population. After StyI digestion, the transgenic plants expressing LjMOT1, along with the mutant and WT plants, were transferred to hydroponic solution. After growing in hydroponic solution for 2 weeks, the roots and shoots were harvested to determine Mo concentrations.

Heterologous expression of LjMOT1 in Saccharomyces cerevisiae and Mo transport activity assay

Yeast expression plasmids were constructed using a homologous recombination method (Ma et al., 1987). WT and mutated LjMOT1 were amplified from cDNA of WT or ljmot1 plants using the following primers: 5′‐GCTGTAATACGACTCACTATAGGGAATATTAAGCTTGGTCACCATGGCTAACCAAAACCCTCCTTC‐3′ and 5′‐ GCATGCTCGAGCGGCCGCCAGTGTGATGGATATCTGCATGGGGCTTTTCTGAGTCCAAATAG‐3′. AtMOT1 was amplified using the following primers: 5′‐GCTGTAATACGACTCACTATAGGGAATATTAAGCTTGGTCACCATGGAGTCTCAGTCTCAGAG‐3′ and 5′‐ GCATGCTCGAGCGGCCGCCAGTGTGATGGATATCTGTCAAGCATGTTCACCGGATTG‐3′ (underlines indicate the homologous recombination regions). The PCR fragments and pYES2 vector (Invitrogen) digested with SacI and EcoRI were introduced simultaneously into S. cerevisiae (BY4741). The empty vector was used as a negative control. After transformation, the yeast cells were grown on SD‐Ura plates to obtain transformants, after which the inserted fragments in the transformants were sequenced.

For Mo uptake activity, yeast cells carrying different constructions were cultured to the mid‐log phase in Mo‐free selective medium containing glucose (SD) or galactose (SG) (Sherman, 1991), and then incubated for 30 min in the same medium containing 170 nm Mo. For kinetic analysis of Mo uptake, cells expressing LjMOT1 were cultured to mid‐log phase in Mo‐free SG medium and then incubated for 15 min in SG medium supplemented with 56, 84, 112, 168, or 679 nm Mo. After Mo treatment, cells were harvested and washed twice with ice‐cold deionized water. The yeast cells were re‐suspended in 1 mL ice‐cold deionized water and boiled at 100°C for 40 min. After cooling, the boiled solutions were centrifuged and the supernatants and pellets were used for the measurement of Mo concentrations.

Quantitative RT‐PCR and histochemical analysis of LjMOT1

To investigate LjMOT1 expression under different Mo supply conditions, surface‐sterilized seeds of WT were germinated on half‐strength B&D solution containing 2 mm KNO₃ solidified with agar (0.8%). After 10 days, the seedlings were transferred to hydroponic pots containing the same nutrient solution supplemented with 0, 0.17, 1, or 10 μm Mo. The nutrient solutions were aerated continuously. After 48 h, the plant roots and shoots were harvested for RNA extraction.

For the M. loti inoculation experiments, surface‐sterilized seeds of WT were germinated on water agar (0.8%) plates. After 4 days, the seedlings were transplanted to vermiculite supplied with half‐strength B&D medium without nitrogen or Mo. Three days after transplanting, the plants were inoculated with M. loti MAFF303099. Total RNAs were isolated from roots sampled 4 and 8 days after inoculation, from nodules sampled 12, 17, and 22 days after inoculation, and from roots that were not inoculated.

Total RNA was prepared using the RNeasy Plant Mini Kit (Qiagen, Valencia, CA, USA), and DNase treatment was performed with an RNase‐free DNase set (Qiagen). Total RNA (500 ng) was reverse‐transcribed into cDNA using the PrimeScript RT reagent kit (Takara, Ohtsu, Japan) with oligo(dT)₁₆ primer. The cDNA was diluted 10‐fold and used for real‐time PCR analysis with Dice (TaKaRa) using SYBR Premix Ex TaqII (TaKaRa). The primers used for LjMOT1 were 5′‐GGGATGAGCCACTTTCTACT‐3′ and 5′‐AGGAACTTGAGTGCAGTGTG‐3′. The primers used for the ubiquitin gene (internal standard) were 5′‐ TTCACCTTGTGCTCCGTCTTC‐3′ and 5′‐ AACAACAGCACACACAGCCAATCC‐3′.

For promoter GUS analysis, the putative promoter and terminator regions of LjMOT1 were fused with the GUS reporter gene. The LjMOT1 promoter and terminator fragments were amplified by PCR from genomic DNA using the following primers: 5′‐ GGTACCCACCATCATTCAATCACAAC‐3′ and 5′‐ GGATCCGGAAGGAAGACTGTGTTTTT‐3′ for the promoter (underlines indicate KpnI and BamHI sites); and 5′‐CTGCAGCTCTGGAAGAAAAGGAACAT‐3′ and 5′‐ AAGCTTTATGTGTGAGTGTTGGAGTG‐3′ for the terminator (underlines indicate PstI and HindIII sites). The resulting DNA fragments were digested with the indicated restriction enzymes and inserted into a pC1300GFP vector digested with the same enzymes. The gusA gene was amplified from pENTR‐gus (Invitrogen) using the following primers: 5′‐GGATCCATGGTCCGTCCTGTAGAA‐3′ and 5′‐GTCGACTTATTGTTTGCCTCCCTG‐3′ (underlines indicate BamHI and SalI sites). The BamHI/SalI‐digested DNA fragment was then inserted into a pC1300GFP vector between the introduced LjMOT1 promoter and terminator fragments. The fused construct (proLjMOT1‐gusA‐tLjMOT1) was introduced into L. japonicus ecotype Gifu by Agrobacterium rhizogenes–mediated hairy‐root transformation (Diaz et al., 2005). Transgenic roots (or roots with nodules) were incubated with GUS staining solution (2 mm 5‐bromo‐4‐chloro‐3‐indolyl‐β‐d‐glucuronide, 5 mm potassium ferricyanide, 5 mm potassium ferrocyanide, 100 mm sodium phosphate; pH  7.0) for 16 h. Transgenic nodules were embedded in 5% agar and sectioned at 60‐μm thickness with a micro‐slicer (VT1200S; Leica, Heerbrugg, Switzerland), followed by incubation in staining solution for 16 h. The stained roots and nodules were observed with a light microscope.

Measurement of acetylene reduction activity

Surface‐sterilized seeds of WT and ljmot1 were germinated on water agar (0.8%) plates. After 4 days, the seedlings were transplanted to vermiculite and supplied with half‐strength B&D solution without nitrogen or Mo. Three days after transplanting, the plants were inoculated with M. loti MAFF303099. ARA was measured 14, 21, and 28 days after inoculation with M. loti, as described by Kumagai et al. (2007). Briefly, intact plants were placed in 12‐ml vials containing 10% (v/v) acetylene and incubated at 25°C. After 20 min incubation, ethylene was measured by gas chromatography (GC‐8A; Shimadzu, Kyoto, Japan).

Measurement of Mo concentration

In total, 0.2 g finely ground plant samples or total dried yeast cells were weighed into a quartz glass tube, to which 2.5 ml of high purity nitric acid was added and left to stand overnight. Then the samples was digested on a block digester at 100°C for 1 h, 120°C for 1 h, and 140°C until the acid was evaporated and the digests were dissolved in 30 ml 0.08 N HNO₃. The concentrations of Mo and other nutrients in the samples were determined by inductively coupled plasma mass spectrometry (model SPQ9700; SII, Chiba, Japan).

Data analysis

Data were analyzed by analysis of variance, followed by comparisons between means using the least significant difference.

Accession numbers

Sequence data from this article can be found in the L. japonicus protein database (miyakogusa.jp, http://www.kazusa.jp/lotus/index.html, database of build 2.5) under the following accession numbers: LjMOT1, CM2163.80; LjMOT2, CM0226.90; LjMOT3, CM0760.220; LjMOT4, CM2163.20.

Acknowledgements

We would like to thank Dr Tsuyoshi Nakagawa of Shimane University for providing pGWB1 and Dr Mitsuru Osaki of Hokkaido University for providing seeds of WT L. japonicus, WT Miyakojima, and the EMS mutant. This work was supported in part by a grant from the Japan Society for the Promotion of Science (to GD) and a Grant‐in‐Aid for Scientific Research, Kiban S (No. 25221202 to TF).

Conflicts of interest

The authors declare no conflicts of interest.