Cesium accumulation in nodules is involved in mitigating cesium transfer to shoot

The hydrogen explosions at TEPCO’s Fukushima Daiichi Nuclear Power Plant in 2011 resulted in the release and deposition of radioactive substances, including radiocesium (¹³⁷Cs), onto agricultural fields. Crops cultivated in decontaminated fields remain at risk of radioactive substances transfer to edible plant parts¹. The transfer of ¹³⁷Cs from soil to plants has become a significant environmental and agricultural concern due to its long half-life (ca. 30.2 years). Potassium (K) fertilization has been shown to effectively reduce cesium (Cs) uptake in various crops, including soybean, wheat, and rice^2,3,4. For rice and soybean cultivation, maintaining exchangeable K concentrations in soils (over 25 mg K₂O per 100 g soil) through cultivation has been recommended^5,6. This mitigation strategy is based on the chemical similarity between Cs⁺ and K⁺, thereby elevating soil K concentrations competitively inhibiting Cs uptake. It is well known that Cs is primarily taken up and translocated via K transport systems in plants. The high-affinity K⁺ transporter family KUP/HAK/KT has been reported to be involved in Cs⁺ transport⁷. In the case of Arabidopsis thaliana, AtHAK5 has been identified as a key transporter contributing to increased Cs uptake under K-deficient conditions⁸. In rice, OsHAK1 has also been associated with Cs transport^9,10. While the role of the soybean homolog GmHAK5 in K uptake and distribution in root was evaluated¹¹, the function of this transporter for Cs remains poorly understood. Regarding other Cs transport pathways, the Arabidopsis inward-rectifying K⁺ channel AKT1 has been suggested to be involved in Cs transport, although its role appears to be limited^12,13. Cs is also absorbed and distributed primarily in voltage-insensitive cation channels under K-sufficient conditions¹⁴.

Under equivalent soil K concentrations, soybean exhibits high Cs transfer to seeds compared with other crops, such as wheat and maize^15,16. Even among leguminous species, soybean tends to accumulate more ¹³⁷Cs than peanut¹⁷. It is considered that the higher Cs accumulation in soybean seed is due to high K demand. Additionally, the structural characteristics of soybean seeds facilitate Cs translocation into the edible portions¹⁸. Soybean continues to absorb K from soil until the full seed stage, and it has been reported that the ability of soybean to discriminate between K and ¹³⁷Cs declines from the flowering stage to seed development, resulting in increased ¹³⁷Cs uptake during this period⁴. K fertilization is highly effective to reduce Cs uptake and improve the production in soybean; however, due to the economic cost of fertilizers, continuous K application may not be a sustainable strategy to mitigate ¹³⁷Cs transfer from soil to plant. Therefore, other mitigation strategies in addition to K fertilization should be explored for effective risk management of Cs transfer in soybean.

To better understand Cs transfer, several studies have investigated the role of nonexchangeable K¹⁹, Cs adsorption to soil mineral nutrients²⁰, and the effects of cattle manure application²¹. While these studies focus on soil-based countermeasures, it is also necessary to explore plant-specific strategies for reducing Cs transfer. One potential approach in soybean is the use of symbiotic microorganisms. Soybean establishes symbiotic relationships with rhizobia and arbuscular mycorrhizal (AM) fungi, which primarily enhance nitrogen and phosphorus acquisition. The role of AM fungi in Cs transfer from soil to shoot remains inconsistent, with some studies showing Cs delivery to host plants from soil²², while others report limiting its movement to shoot due to its retention within fungal hyphae^23,24. Nodules, which share evolutionarily conserved common signaling pathways with AM fungi symbiosis²⁵, are reportedly more sensitive to salinity and heavy metal stress than roots or whole plants²⁶. In terms of the role of K in nodules, it is presumed to contribute to intercellular potential regulation, cell expansion, and energy supply^27,28. In rhizobia, a positive relationship between K levels and nitrogenase activity has been reported²⁹. These limited studies suggest that K plays a critical role in both nodule tissues and rhizobia. However, while the mechanisms of K uptake and distribution in nodules are not yet fully understood, the involvement of Cs in nodules remains even more obscure. Furthermore, both nodule formation and Cs transfer are inherently influenced by soil nutrient conditions. In particular, soil nitrogen and phosphorus conditions have been suggested to play a role in nodule formation and Cs dynamics^30,31,32,33. Thus, it is necessary to investigate how nodules affect Cs transfer under various soil conditions.

This study aimed to reveal the effects of nodules on the mechanism of Cs transfer to shoot. To achieve this goal, we conducted field experiments in 2021 and 2024 using three soybean genotypes with different nodulation abilities, under various fertilization treatments, to investigate the dynamics of monovalent cations in each plant organ and soil. We also attempted to conduct a spatial analysis of ¹³⁷Cs in root and nodule tissue to investigate the ¹³⁷Cs distribution pattern. We found that Cs levels were higher in nodules than in other plant organs and increased nodule formation was potentially associated with suppressing Cs translocation to shoot. Our findings suggest that nodules may play a key role in mitigating Cs transfer in addition to K fertilization, providing a potential strategy for ¹³⁷Cs risk management in leguminous crops.

Plant materials and growth conditions

Three soybean (Glycine max (L.) Merrill) genotypes with different nodulation abilities [a normal nodulating genotype (Enrei), a hyper-nodulating genotype (En-b0-1), and a non-nodulating genotype (En1282), all derived from Enrei as described by Hamaguchi et al.³⁴] were cultivated in a field at Hokkaido University in 2021 and 2024. These fields were established in 1914, with four fertilizer treatments, namely, complete fertilization (+ NPK), without nitrogen (− N), without phosphorus (− P), and without potassium (− K), applied continuously for over 100 years³⁵. Soil type is categorized as gray lowland soil. In the − N, − P, − K, and + NPK treatments, each plot size was 5.25 × 15.80 m. Moreover, the N, P, and K fertilizers were applied as ammonium sulfate, superphosphate, and potassium sulfate, respectively (100 kg N, P₂O₅, K₂O ha⁻¹). Three seeds were sown at 70 cm × 20 cm intervals and then thinned to two plants per hill. Plant organs (leaf, petiole, stem, pod, seed, and root) and bulk soil (the 0–15 cm soil in the inter-row space) were sampled at the flowering and maturity stages with three replications. Plant samples at the flowering stage were divided into leaf, petiole, stem, root, and nodule. Plant samples were then dried at 80 °C for 1 week, and their dry weights were measured. These dried plant samples were ground for mineral analysis. Soil samples were air-dried and passed through a sieve with a 2 mm diameter for mineral analysis.

Measurement of plant mineral

Plant samples were digested in 2 mL of 61% (w/v) HNO₃ (EL grade; Kanto Chemical, Tokyo, Japan) at 107.5 °C in a DigiPREP apparatus (SCP Science, Canada). After approximately 1.5 h, about 0.5 mL of H₂O₂ (EL grade; Kanto Chemical, Tokyo, Japan) was added, following by further digestion for another 15 min. The digested solution was cooled and filled with 10 mL with 2% HNO₃ in Milli-Q water. Cs, K, sodium (Na), and rubidium (Rb) were analyzed using inductively coupled plasma mass spectrometry (ICP-MS) (ELAN DRC-e; Perkin Elmer, Waltham, MA, USA). To determine δ¹⁵N in leaves, plant samples were weighed into tin capsules and analyzed using an isotope ratio mass spectrometer (Integra2, sercon, UK). The δ¹⁵N values were reported relative to atmospheric N₂.

Measurement of soil mineral

For the soil exchangeable cation, soil was extracted with 1 M ammonium acetate at a soil/solution ratio of 1:20, shaken for 1 h and filtered using an Advantec® quantitative filter paper No. 5C (Toyo Roshi Co., Ltd., Tokyo, Japan). The filtrates were measured using ICP-MS (ELAN DRC-e; PerkinElmer, Waltham, MA, USA) to determine the concentration of soil exchangeable cation. Total nitrogen concentration in soil samples was determined by Kjeldahl digestion. Available phosphorus concentrations were measured using the Bray II method³⁶. Soil pH (soil/Milli-Q water = 1:5, w/v) was measured with a pH meter (HORIBA Ltd., Kyoto, Japan).

Spatial imaging of ¹³⁷Cs in root and nodule tissues

To study the distribution of ¹³⁷Cs in the nodules and roots, germinated soybean (Glycine max cv. Enrei) seeds were cultivated hydroponically in 3-L plastic containers while aerating and inoculated with Bradyrhizobium japonicum USDA110 (obtained from the National Agriculture and Food Research Organization, NARO, Japan) at a concentration of 1 × 10⁷ colony-forming units mL⁻¹. The solution was an N-free half-strength modified Hoagland’s nutrient solution (3 mM K, 2 mM Ca, 1 mM P, 0.5 mM S, 0.5 mM Mg). The nutrient solution inoculated with rhizobia was replaced every 3 days. At 18 days after the start of hydroponic cultivation, 3.7 kBq mL^{−1 137}Cs was added to the nutrient solution immediately after 18 days of hydroponic cultivation without any pre-treatment, and the plants and rhizobia were allowed to absorb it for 24 h. After washing the root with pure water, the root with attached nodules was cut into approximately 1 cm pieces and immediately frozen by submerging in liquid nitrogen. Thereafter, the roots and sliced sections were prepared according to a previously described method^37,38. Each sample was embedded in an embedding medium, covered with an adhesive film (Cryo-Film transfer kit, Finetec, Tokyo, Japan), and sliced with a cryostat into serial sections of 5 µm thickness at − 20 °C. A BAS IP TR imaging plate (GE Healthcare) was stuck to the frozen section on the side directly opposite the attached film. The imaging plate (IP) with the samples was kept in the freezer at − 80 °C for 6 days. The image on the IP was scanned with an FLA5000 image analyzer (Fujifilm, Tokyo, Japan) at a resolution of 10 µm.

Statistical analyses

To compare dry weight and elemental distribution of plant organs, a three-way analysis of variance (ANOVA) was first performed to evaluate the effects of organ, fertilizer treatment, and genotype in R (version 4.4.1). However, strong effects of treatment and genotype masked organ-specific differences. To address this, a one-way ANOVA was subsequently conducted for organs within each treatment–genotype combination, followed by Tukey’s multiple comparison analysis. To assess effects of treatment and genotype by year, a two-way ANOVA was conducted, with Tukey’s multiple comparison analysis applied when interactions were significant. Regarding correlation analysis, we used the Spearman’s rank correlation to accommodate potential nonlinearity and non-normality. An analysis of covariance (ANCOVA) was conducted in Python (version 3.9.18) using the statsmodels library (version 0.14.0). Least‐squares means (LS‐means) for each genotype were calculated at the overall mean of covariates, and pairwise t-tests were then performed on those LS-means. False discovery rate correction was applied using the two-stage Benjamini–Hochberg method (BH).

Soil chemical properties under different fertilization treatments

To clarify the effect of nodule formation on cation transfer from soil to soybean, we investigated three soybean genotypes with different nodulation abilities (normal nodulating genotype: Enrei; hyper-nodulating genotype: En-b0-1; and non-nodulating genotype: En1282) grown under various fertilization treatments. Compared with the + NPK treatment, soil exchangeable K concentrations were significantly reduced under − K treatment consistently across both years and growth stages (Supplementary Table S1). Meanwhile, soil exchangeable Cs was highest under − K treatment. Soil exchangeable K was higher and Cs was lower in the − N and − P treatments than in the + NPK treatment. Other exchangeable monovalent cations in soil showed similar trends to Cs. Available phosphorus in soil was lowest under − P treatment, and soil total nitrogen tended to be lowest under − N treatment. Soil pH was significantly increased under − N treatment.

Growth of genotypes with different nodulation abilities under different soil nutrient treatments

In the normal nodulating genotype, shoot (combined leaf, petiole, and stem) dry weight was highest in the + NPK treatment in both 2021 and 2024, and decreased in the order of − N, − P, and − K treatment (Table 1). In contrast, both the hyper-nodulating and non-nodulating genotypes reduced shoot biomass in − N treatment. Across all genotypes and treatments, leaf dry weight was consistently the highest among organs, while nodule dry weight was the lowest (Fig. 1a,b). Under − K treatments, growth was consistently suppressed in all genotypes, with the number and dry weight of nodules being the lowest across treatments in both 2021 and 2024 (Table 1). Nodule formation showed no significant difference between the − P and + NPK treatments but was significantly increased under the − N treatment. Seed weight at maturity showed a similar tendency to shoot dry weight at the flowering stage, with a significant reduction in − K treatment observed particularly in the normal nodulating genotype. While overall biomass and yield were higher in 2021 compared to those in 2024, the effects of fertilizer treatments and genotypes were largely consistent across years.

Elemental distribution within plant in field experiment

A two-way ANOVA was conducted for shoot Cs and K concentrations to evaluate the effects of genotype and treatment on the elemental distribution to shoot. Significant effects on shoot Cs concentration were observed only in 2021. Cs levels in the hyper-nodulating genotype tended to be lower compared with those in the normal and non-nodulating genotypes, particularly in − K and + NPK treatment (Table 2). Across all genotypes, shoot Cs concentration was highest in − K and lowest in − N treatment. Shoot K concentrations consistently decreased in the − K treatment in both years across all genotypes. To investigate elemental distribution among organs, soybean samples at the flowering stage were separated into leaves, petioles, stems, roots, and nodules. In the normal nodulating genotype grown in 2021, Cs concentrations were highest in nodules across all treatments compared with those in other organs, with particularly pronounced accumulation in − K and + NPK treatment (Fig. 2a). A similar tendency was observed in the hyper-nodulating genotype, which also exhibited significantly increased concentrations of Cs in nodules under these treatments. Although one-way ANOVA detected no significant differences under –N, –P, and + NPK treatment in 2024, Cs concentrations tended to be higher in nodules (Fig. 2b). In 2021, nodules had significantly higher K concentrations than other organs in the − K treatment in both the normal and hyper-nodulating genotypes (Fig. 2c). However, this trend was not observed in 2024, showing no consistent distribution pattern (Fig. 2d). For Cs concentrations in shoot, leaves were the primary organ of Cs allocation. In contrast, K concentrations were significantly higher in petioles than in leaves in all treatments except − K in both years. At maturity, seed Cs concentrations increased in the − K treatment, while seed K concentrations decreased (Table 2). For other monovalent cations of organs at the flowering stage, Na significantly accumulated in roots in all genotypes in both 2021 and 2024 (Supplementary Fig. S1). In the − K treatment, Na concentration increased in nodules but remained lower than that in roots. Rb tended to be more concentrated in nodules than in other organs in the normal nodulating and hyper-nodulating genotypes in 2021. A similar tendency was also observed in the normal nodulating genotype under the − K treatment in 2024.

Distribution ratios among organs were calculated to evaluate the effect of nodules on the elemental distribution within plant. The proportion of Cs distributed to nodules was highest in the − N treatment in both 2021 and 2024 (Supplementary Fig. S2). Compared with the normal nodulating genotype, the hyper-nodulating genotype showed higher Cs proportions allocated to nodules across all treatments. K exhibited similar distribution patterns to Cs in both years. Spearman’s rank correlation coefficient (r) between the proportion of Cs allocated to nodules and that allocated to shoot was significantly negative (r =  − 0.77, P < 0.005) (Supplementary Fig. S3). Similar patterns were observed for K (r =  − 0.66, P < 0.005). In contrast, Na was predominantly distributed to roots across all treatments, and increased allocation to nodules did not result in the change of shoot Na level (Supplementary Figs. S2 and S3). Rb showed a distribution pattern similar to those of Cs and K (Supplementary Fig. S2), and a negative correlation was observed between nodule and shoot allocation ratios (r =  − 0.70, P < 0.005) (Supplementary Fig. S3).

Spatial localization of ¹³⁷Cs in root system

Although high concentrations of stable Cs in nodules were observed in field experiments, it is also necessary to confirm the distribution pattern of ¹³⁷Cs. A transverse section (Fig. 3a) and its corresponding autoradiograph (Fig. 3b) were superimposed to generate Fig. 3c. According to the overlaid image, a strong signal was observed in the nodule, particularly in the infection zone. In contrast, the signal in the main and lateral roots was very weak. These observations suggest that ¹³⁷Cs is taken up and accumulated in the nodule at 24 h after the start of absorption.

Relationship between soil minerals and plant minerals

Although the Cs distribution pattern within plant was clarified, a comprehensive understanding of the mineral nutrient dynamics requires not only a distribution analysis but also an investigation of nutrient uptake from the soil. We then investigated the relationship between mineral nutrients in the soil and those in soybean. A consistently significant positive correlation was observed between soil exchangeable Cs concentrations and Cs concentrations of both shoots and seeds in both 2021 and 2024 (Supplementary Fig. S4). In contrast, a significant negative correlation was found between soil exchangeable K concentrations at the flowering stage and shoot Cs concentrations in 2021 (r =  − 0.74, P < 0.005), with a similar tendency observed in 2024 (r =  − 0.79, P < 0.005) (Fig. 4a,b). In all three genotypes, soil exchangeable K was negatively correlated with shoot Cs. Given this strong inverse relationship, soil exchangeable K was included as a covariate when comparing shoot Cs among genotypes. To compare the effects of genotypes on shoot Cs after adjusting for soil exchangeable K, an ANCOVA (shoot Cs ~ exchangeable K + genotype) was performed for the hyper-nodulating, normal, and non-nodulating genotypes, followed by pairwise t-tests of the resulting LS-means. The exchangeable K concentrations adjusted shoot Cs concentrations of the hyper-nodulating genotype were significantly lower than that of both the normal and non-nodulating genotypes (FDR-corrected two-stage BH P < 0.05), whereas the exchangeable K concentrations adjusted shoot Cs concentrations of the normal and non-nodulating genotypes were not significantly different. At the maturity stage, significant negative correlations were found between soil exchangeable K concentrations and seed Cs concentrations in both 2021 (r =  − 0.84, P < 0.005) and 2024 (r =  − 0.88, P < 0.005) (Fig. 4c,d). However, ANCOVA did not show significant differences among genotypes in this relationship.

Relationship between nodulation and Cs translocation

The relationship between soil exchangeable K concentration and shoot Cs concentration varied among nodulation genotypes, indicating that nodule formation may influence Cs translocation to shoot. We then further investigated the relationship between nodulation traits and shoot Cs concentration at the flowering stage. Shoot Cs concentration decreased as the nodule dry weight increased (r =  − 0.55, P < 0.005), and a similar negative relationship was found with nodule number (r =  − 0.66, P < 0.005) (Fig. 5a,b). Shoot Cs concentration was negatively correlated with soil exchangeable K concentration (Fig. 4a,b), and both nodule number and nodule dry weight were significantly reduced under K-deficient conditions (Table 1). The relationship between nodule formation and shoot Cs concentration may be potentially confounded by soil K availability. To address this, partial correlation analysis was performed using soil exchangeable K concentration at the flowering stage as a confounding factor. While no significant relationship was found between shoot Cs concentration and nodule dry weight, a significant negative partial correlation was retained with nodule number (r =  − 0.37, P < 0.01). δ¹⁵N in leaves, which is used as an indirect indicator of nitrogen fixation, showed no significant correlation with shoot Cs concentration (Supplementary Fig. S5). In contrast, a significant correlation was observed between nodule Cs concentration and leaf δ¹⁵N, and this relationship remained significant after accounting for soil exchangeable K (Supplementary Fig. S6). Similar correlation analyses were conducted for other monovalent cations in shoot. K concentration showed significantly positive correlations with both nodule dry weight and nodule number (Supplementary Fig. S7). Meanwhile, Na concentration was not significantly correlated with either trait. Rb showed a pattern similar to Cs. It was not significantly correlated with nodule dry weight but was negatively correlated with nodule number (r =  − 0.49, P < 0.005). Regarding nitrogen fixation, no significant correlations were found between δ¹⁵N and other monovalent cation concentrations in leaves (Supplementary Fig. S5). For nodule monovalent cation concentrations, significant correlations were observed for Na and Rb, but not for K (Supplementary Fig. S6).

Root nodule symbiosis affects the uptake and transport systems not only for nitrogen but also for other mineral nutrients, such as K^39,40. K has been suggested to play a role in maintaining ion homeostasis and nodule development and nitrogenase activity⁴¹. Our study revealed that Cs consistently accumulated at higher concentrations in nodules at the flowering stage in both 2021 and 2024, whereas K levels in nodules exhibited an inconsistent trend (Fig. 2). Moreover, ¹³⁷Cs was mainly accumulated in nodules under root nodule symbiotic conditions (Fig. 3). As reported by Shinano et al.⁴², K is preferentially translocated to the shoot rather than Cs, possibly limiting Cs mobility and leading to Cs accumulation in the root system, particularly in nodules. Supporting this notion, Cs accumulation in nodules was significantly high under K-deficient conditions, indicating that limited K availability enhances Cs accumulation in nodules. Furthermore, ¹³⁷Cs was particularly enriched in the infected cells of nodule. K is essential for rhizobia to maintain osmotic balance⁴³. This suggests that Cs may also accumulate preferentially in bacteroids. However, Cs localization within nodules remains inconclusive due to the short-term observation and the overall K requirement of nodule tissues. To clarify this, further studies, such as merging Cs imaging with rhizobia staining or conducting experiments using artificial nodules, will be necessary. Additionally, because nutrient uptake and distribution mechanisms may differ between symbiotic and non-symbiotic conditions, results such as ¹³⁷Cs imaging should be regarded as specific to the root nodule symbiotic condition.

Increasing exchangeable K concentration in soil has been widely regarded as the most effective strategy for reducing Cs transfer, regardless of species. We also observed a significant negative correlation between soil exchangeable K and shoot Cs concentrations in soybean (Fig. 4). However, ANCOVA showed the different relationships of different nodulation genotypes, suggesting that nodule formation influences Cs uptake at the flowering stage. In nodules, K supports ATP supply and nitrogenase activity and is positively associated with carbon and nitrogen transport^44,45,46. In the hyper-nodulating genotype, increased total photosynthate allocation to nodules may result from enhanced nodulation in lateral root and reduced root elongation⁴⁷. Changes in the root system may have increased carbon costs in roots and altered nutrient uptake mechanisms, potentially affecting translocation and consequently reducing Cs translocation to shoot. This suggestion is supported by the different elemental distributions in roots and nodules observed in the hyper-nodulating genotype compared with those in other genotypes (Supplementary Fig. S2). In contrast, there were no significant differences among genotypes at the maturity stage, indicating that seed Cs levels remain homeostatic despite genotypic differences in Cs uptake at the flowering stage.

Furthermore, our results suggest that nodules may contribute to suppressing Cs translocation to shoot (Fig. 5). The high Cs concentrations observed in nodules (Fig. 2a,b) indicate that nodules may function as “buffer zones”. While heavy metals and alkali metals generally differ in terms of chemical properties and translocation mechanisms, nodules have been proposed to serve as buffer zones to mitigate heavy metals^48,49,50. We also found that increased nodule number was associated with increased shoot K concentrations (Supplementary Fig. S7), suggesting that preferential K translocation to the shoot may have indirectly promoted Cs accumulation in nodules. Notably, δ¹⁵N, one of nitrogen fixation activity indicators, suggested that nitrogen fixation may not contribute to reduced Cs translocation to shoot (Supplementary Fig. S5). In contrast, Cs accumulation in nodules did not appear to inhibit nitrogen fixation activity (Supplementary Fig. S6). Further studies should use more quantitative indicators of nitrogen fixation activity, such as acetylene reduction assays, to examine its relationship with Cs dynamics.

Elemental uptake and distribution in nodules primarily occur via two pathways, i.e. direct transport through the nodule epidermis and cortex⁵¹ and translocation from roots via vascular tissues into nodules⁵². Several K transporters, including HAK5, GORK, SKOR and KUP7, are expressed in the epidermal and vascular tissues^53,54,55,56. Interestingly, our study found that the suppression of Cs translocation was more strongly associated with nodule number than with nodule dry weight (Fig. 5). In soybean nodules, the vascular bundles extend radially around the infected zone^57,58. Therefore, the increase in the nodule number implies an expansion of the total epidermal and vascular tissue area. Treating transporter density per unit membrane area as approximately constant, this expansion broadens the expression domain of K transporters at these sites and, in turn, altering Cs translocation. Verification of this hypothesis requires functional analyses of transporters, including assessments of Cs selectivity, and spatial localization of expression in the relevant regions. As another hypothesis, our findings suggest that a single nodule does not accumulate Cs indefinitely. Soybean forms determinate nodules, where nodule development is driven by cell expansion rather than division⁵⁹. To facilitate cell expansion, intercellular potential balance needs to be maintained, potentially absorbing Cs in nodules during nodule development. However, in mature nodules, an elevated intracellular potential may limit further passive Cs transport. K transporters such as GmHAK5 and LjKUP, which are expressed in nodules, have been identified, but their involvement in active Cs sequestration within nodules remains unclear^60,61. To clarify these physiological mechanisms, it is necessary to investigate the subcellular localization of elements and gene expression within nodules spatiotemporally, in accordance with nodule developmental stages.

For other cations, Cs⁺, K⁺, Na⁺ and Rb⁺ are chemically similar and generally exhibit competitive uptake due to overlapping parts of transport pathways⁷. In this study, exchangeable Na and Rb concentrations in the soil increased under K-deficient conditions, like Cs (Table S1). Regarding Rb distribution within the plant, Rb concentrations in nodules tended to increase, similar to Cs, and nodule number appeared to be associated with reduced Rb translocation to shoot (Figs. 2, S1, S2, S3 and S7). These similar dynamics of Cs and Rb suggest the involvement of a common transport mechanism. Further studies are needed to identify nodule-localized transporters responsible for Cs and Rb uptake and to clarify their ion selectivity. In contrast, Na was predominantly distributed to the roots, and no significant relationship was observed between nodule traits and Na translocation patterns (Figs. 2, S1, S2, S3 and S7). In legumes, Na taken up from the soil is retained prior to crossing the endodermis, with reduced levels in xylem vessels, thereby limiting its upward movement⁶². Although recent work in rice suggested that Na indirectly influences Cs transport through Na transporters⁶³, Cs and Rb transport is thought to rely mainly on K transport systems⁷. This indicates that Na is likely distributed through a distinct mechanism from Cs and Rb. To elucidate differences in these distribution mechanisms, future studies of rhizobia-inoculated soybean should employ techniques such as isotope tracing and elemental mapping. Importantly, because exchangeable Na concentrations in soil differ substantially from those of Rb and Cs, potential ionic interactions and competition among these monovalent cations should be carefully considered.

Our study revealed the influence of nodules on the mechanism of Cs translocation to shoot. Through field experiments conducted using soybean genotypes with different nodulation abilities, comparison of elemental concentrations in each organ revealed that Cs was highly accumulated in nodules. Similarly, spatial mapping of ¹³⁷Cs in the root system revealed that ¹³⁷Cs was predominantly distributed in nodules. These results highlight the importance of nodules as key sites for Cs sequestration. This study also demonstrated a significant negative correlation between nodule number and shoot Cs concentration, suggesting that increased nodulation may suppress Cs translocation to shoot. These findings provide fundamental insights into the potential use of root nodule symbiosis as a strategy for mitigating Cs transfer to shoots. To further elucidate the mechanisms underlying Cs accumulation in nodules and its suppressed translocation to shoot, spatiotemporal analyses of elemental distribution and gene expression within nodules are needed.