Accumulation and translocation of selenium in a soil–rice system in Guangxi, China

ABSTRACT

Selenium (Se) is an essential element for the human body. Humans meet their basic Se requirements primarily through food consumption. The largest Se-rich contiguous area in China is located in the southeast of Guangxi. In this study, the Se accumulation and translocation in the Guangxi rice–soil system was assessed. Organically bound Se (O-Se) was found to be the main type of soil Se (72.5%), and had a good linear relationship with soil total Se (TSe) and soil organic matter (SOM); the three forms were closely related and thus, affected each other. The contents of Se in the various plant organs were highly correlated. The roots had the greatest Se concentration, followed by the stems, grains, and lastly the leaves. The distribution coefficient of grain/stem (0.74) was greater than that of grain/leaf (0.54), suggesting that the migration of Se in plants may have apical dominance and/or reproductive advantages. Ligand-exchangeable Se, water-soluble Se, and residual Se had little effect on the accumulation of Se in the soil and plants. The soil S content showed no significant correlation with the various types of soil Se and TSe, but showed a low correlation with the Se content in various plant organs. SOM played a special role in the translocation of Se between the soil and plants, which affected the absorption and accumulation of Se by plants. SOM can affect the content of O-Se and, consequently, of TSe in the soil. SOM can also directly or indirectly affect the Se content in various plant organs, especially the Se concentration in rice grains.

KEY WORDS:

• Se
• rice
• SOM
• soil
• translocation

1. Introduction

Numerous studies have demonstrated the importance of Se to human health. Notably, excessive Se intake can cause staggering disease (Gissel-Nielsen et al. 1984), and Se deficiency causes Keshan and Kashin–Beck diseases (Shao and Zheng 2008), although it also contributes to the prevention of cancer and cardiovascular disease and has anti-aging effects (Schrauzer and Surai 2009). The recommended Se daily intake is 55 μg for humans (Navarro-Alarcon and Cabrera-Vique 2008); however, the abundance of Se in the earth’s crust is only 0.05–0.09 mg/kg (Wang and Gao 2001). Therefore, there are approximately 0.5–1 billion Se-deficient people worldwide. In China, nearly two-thirds of planting soils are Se-deficient, leading to a lack of Se intake among the local population (Combs 2001; Zhang et al. 2014a). Therefore, the Se content of food crops needs to be increased through the external administration of this element, as the most effective and safe way to supplement Se to locals is through the consumption of Se-rich organic foods (Finley and Davis 2001).

Although Se is a beneficial element for plants, excessive Se supplementation can cause plant poisoning (Han et al. 2013). The consumption of plants, which take up the Se present in the soil, is the primary source of Se in the human body (Rayman 2008). Se absorption by plants is related to many factors, such as the content of Se in soil, the types of plants, and the physical and chemical properties of the soil (Antanaitis et al. 2008). As rice is the main food crop in China, several studies have focused on the Se absorption by rice plants and enrichment characteristics of Se in the soil (Qin et al. 2013; Zhang et al. 2014c). The distribution of Se in the earth’s crust is extremely uneven (Tolu et al. 2014); Se-enriched areas can be found in the United Kingdom (Bullock et al. 2018), Japan (Yanai, Mizuhara, and Yamada 2015; Yamada et al. 2009), and some areas in Saudi Arabia (Al-Farhud, Al-Sewailem, and Usman 2017), Poland (Patorczyk-Pytlik and Kulczycki 2009) have low amounts of Se. China has Se-enriched (Zhu and Zheng 1999) as well as low Se areas (Chen et al. 1991). However, the Se content in plants is determined by the proportion of Se in the soil rather than by the total amount of Se in the soil (He et al. 1996). Plants have different absorption mechanisms for different types of Se during the various stages of growth. The transformation of Se in the soil is affected by various factors, such as pH, S, and total organic carbon (Wang et al. 2017, 2018, 2019; Dinh et al. 2017).

According to the Guangxi Zhuang Autonomous Region government’s report, the geochemical survey of 23 cities (counties and districts) performed by the Chinese geological and mining department revealed that Guangxi currently has almost 30,000 km² of Se-rich soil, which represents the largest contiguous Se-rich soil area in China (Tang 2019). This was due to the carbonate rock formations with high Se content in the Devonian, Carboniferous, and Permian (Liu, Chen, and Pan et al. 2018). The content of Se in the soil is not currently changing over time, but gradually decreases from the southeast (0.81 mg/kg) to the northwest (0.44 mg/kg) of Guangxi (Ban and Ding 1992). The Se-enriched status of Enshi in China is garnering increasing attention in the research community, although its Se-contaminated area is relatively large, and human and livestock Se poisoning incidents are frequent (Chang et al. 2019; Qin, Zhu, and Su 2012; Qin et al. 2013). In comparison, Guangxi, has moderate soil Se content, which avoids excessive or lack of Se in rice, and is more conducive to meet the human health demands for Se. Therefore, it is of great practical significance to study the distribution and translocation of Se in soil–rice systems in this area. To the best of our knowledge, this is the first report on the migration of Se in the soil–rice system of Guangxi, China. Currently, a unified understanding of the correlation between different types of Se in soil and Se content in plants is not yet available, as few studies have focused on the correlation between the Se content in various plant organs and the different types of Se in the soil. In particular, the Se types that are mainly absorbed by rice plants are still debated, and there are no relevant reports on the Se migration in the soil–rice plant system of Guangxi. Therefore, the primary objectives of this study are to: (i) reveal the Se migration and transformation rules between the environment and organisms, as well as their influencing factors, and (ii) clarify the influence of various types of Se in the soil on the absorption of rice plants.

Materials and Methods

2. Sample collection and preparation

The samples were collected in July 2020, which corresponds to the mature season of early rice. The study area has a subtropical monsoon climate, where the average annual rainfall is 1,500–1,800 mm and the annual average temperature is 21–24°C. The 16 sampling sites were located in rice fields in southeast Guangxi, which corresponds to the overlap between the main area of high-quality rice production in Guangxi and the Se-rich soil area, as shown in Figure 1 (109.09°–110.50° E, 21.66°–23.39° N; further information on sampling sites is listed in Table S1). All sampling sites had lowland paddy soil with a deep tillage of approximately 20 cm and the rice plants had been grown for more than five years. The soil texture was clay loam or clay. During the rice maturity period, multi-point sampling was adopted and three rice plots with an area of 2–3 m² were selected for each sampling site. The 10 rice samples were collected from each rice plot and mixed together; the same was performed on 10 soil samples of the corresponding cultivated 0–20 cm soil layer (Williams et al. 2009). The rice samples were separated into roots, stems, leaves, and grains, washed with distilled deionized water (dd H₂O), fixed in an oven at 95°C for 30 mins, and dried at 60°C until a constant weight was reached. The dried samples were then transformed into powder with a high-speed crusher (BJ-800A, Baijie Electric Co., Ltd., Hangzhou, China), and subsequently sealed and stored in a polyethylene plastic sample bag for later use. Soil samples were air-dried, and impurities were removed. Samples were subsequently crushed with a wooden hammer, and passed through a 100-mesh (0.149 mm) nylon sieve before bagging for storage. On average, 91% (by weight) of each soil sample can be sieved smoothly, suggesting the 100-mesh sieve is suitable for the following analysis of the soil samples.

Figure 1. Overview of the study area and sampling locations. The sampling points from the southwest to the northeast are numbered 1–16

3. Chemical analysis

Se analysis was performed at the Guangxi Key Laboratory of Agricultural Resources Chemistry and Biotechnology. The national standard soil sample GBW07407 and plant sample GBW10020 were used for quality control analysis. The total Se (TSe), organically bound Se (O-Se), ligand-exchangeable Se (L-Se), water-soluble Se (W-Se), and residual Se (R-Se) in soil were measured. The recovery rate of all samples was close to 100%. The methodological procedures of Zhu et al. (2008) were used to determine the Se content in soil and plant samples. Briefly, 0.50–2.00 g of plant powder or soil sample were weighted into a clean microwave digestion instrument (TANK, Sineo Microwave Chemistry Technology Co., LTD, Shanghai, China) inner tank, after which 10 mL of HNO₃ and 2 mL of H₂O₂ were added to decompose the sample. After the digestion was complete, the solution was transferred into a 50 mL beaker with a washing bottle, where 0.5 mL of HClO₄ were added; the sample was then dissolved on a hot plate until the HClO₄ evaporated and diluted to 50 mL with dd H₂O. A total of 5 mL of the solution were collected and the Se content was determined with a liquid chromatography-atomic fluorescence spectrometer (LC-AF 7590, EWAI, Beijing) (Zhu et al. 2008). The description and determination method of Se types in soil is provided in Table 1.

Table 1. The description and determination of the Se types in paddy soil

Se fractions	Describe	Method*
W-Se	W-Se mainly includes Se (VI), a small amount of Se (IV) and soluble organic Se. W-Se is the easiest selenium to be absorbed by plants, and it is considered as bioavailable selenium in agriculture (Ge et al. 2000).	Twenty five mL of dd H2O was added to the sample, ultrasonic extraction for 30 mins, centrifuga at 4000 rpm for 8 mins, the supernatant was filtered through a 0.45 μm filter membrane. Samples need to be filtered (0.45 μm) before each determination of Se concentration.
L-Se	L-Se mainly refers to the tetravalent selenite ion, which is related to water-containing oxides and the surface adsorption of clay and humus (Zhang et al. 2014b).	Add 8 mL of 0.1 M KH2PO4 solution to the test tube containing residual solids in the previous step, shake vigorously for 3 mins, shake at 200 rpm for 4 h, centrifuge at 4000 rpm for 20 mins, and transfer the supernatant to a new test tube, the Se concentration was determined.
O-Se	Se species that are bound to SOM (Zhang et al. 2014b).	Add 6 ml of 0.1 M NaHO solution to the test tube containing the residual solids in the previous step. Ultrasound for 30 mins, heat in a water bath at 90°C for 2 h, cool to room temperature, and centrifuge at 4000 rpm for 15 mins. Add 5 mL of HCl, dilute to 25 mL, and analyze the Se concentration.
R-Se	R-Se mainly exists in silicate minerals, which is not easily destroyed by acid-base reaction and is not available to plants (Ge et al. 2000).	Add 10 mL of HNO3 and 5 mL of HClO3, heat (170°C) until the sample turns white, and dilute the supernatant to 25 mL with HCl before testing.

* method was modified from Qin, Zhu, and Su (2012)

4. Soil properties

After mixing the soil with water (1:2.5), a pH meter (Leici PHS-3 C, INESA Scientific Instrument Co., Ltd, Shanghai) was used to determine the soil pH.

The K₂Cr₂O₇ volumetric method was used for the determination of the soil organic matter (SOM) content (Wang et al. 2012). The total soil S content was determined using the Mg(NO₃)₂ oxidation-BaSO₄ turbidimetric method. The neutralization titration method was used to determine soil CaCO₃ content: a total of 20 mL of 0.5 M HCl solution were added to 3 g of air-dried soil samples to release the generated CO₂, after which two drops of phenolphthalein indicator were added, and the remaining HCl was titrated with 0.25 M NaOH until the solution turned reddish; the soil CaCO₃ content was then calculated based on the used amount of NaOH solution.

5. Statistical analysis

A one-way analysis of variance was performed using statistical software SPSS v. 17.0, and the correlation coefficient and significance probability of linear regression fitting were calculated according to Pearson correlation analysis. Origin 2018 software was used for graphic production.

Results

6. Soil characteristics

Soil pH is a comprehensive reflection of soil physical and chemical properties, and it influences the chemical types of absorbed Se. The variation of the paddy soil’s pH was relatively small (5.19–6.22), with an average of 5.53 ± 0.30, which represents a weakly acidic soil. The SOM content varied greatly, ranging from 1.21% to 3.92%, with an average of 2.44 ± 0.17% and a coefficient of variation of 0.28. The total S and CaCO₃ contents in the soil were 110.03–591.54 mg/kg and 13.33–59.33 g/kg, respectively. The Se content in the surface paddy soil was 0.11–0.91 mg/kg, with an average of 0.52 ± 0.05 mg/kg. The soil characteristics of this study area and the comparison with those of different Se-rich areas are listed in Table 2.

Table 2. Comparison of soil characteristics between this study and other selenium-rich regions

Area	pH	CaCO3 (g/kg)	SOM (%)	Total sulfur (mg/kg)	TSe (mg/kg)	Crop type	Reference
Guangxi, China	5.19–6.22	13.33–59.33	1.21–3.92	110.03–591.54	0.11–0.91	Rice	This study
Hubei (Enshi), China	4.80–4.06		2.76–4.64	130.00–590.00	0.85–11.46	Rice	Chang et al. (2019)
Shaanxi (Ziyang), China	5.80–6.00	5.20–6.30	1.50–2.00		9.8–19.8	Maize	Wang et al. (2012)
Guangdong, China	5.17–5.48		0.65–2.75		0.28–0.41		Zhang et al. (2008)
Japan	3.96–7.29				0.05–2.80		Yamada et al. (2009)
Punjab, India*	8.05	22.90–46.80	0.53–0.97		1.31–2.41	wheat, rice	Dhillon and Dhillon (2000)
California, USA	6.90–8.80		1.03–1.55		0.07–1.17		Abrams, Burau, and Zasoski (1990)
Malawi	5.20–7.90		1.08–2.64		0.12–0.36	Maize	Chilimba et al. (2012)
Iran					0.11–0.45	wheat, date, pistachio	Nazemi et al. (2012)

* only refers to the soil characteristics of the 0–30 cm soil layer

7. Distribution and translocation of Se in rice plant tissues

The Se content of rice roots was 0.02–0.46 mg/kg (Figure 2), which was generally lower than the soil Se content, and the root-soil distribution coefficient was between 0.04–1.41, indicating that rice roots have a low accumulation ability from the soil. The amount of Se in the aboveground parts of rice plants was generally lower than that of the root system. The Se content was greater in the roots (59.4%), followed by the stems (18.1%), grains (12.6%), and leaves (9.8%) (Figure 2). The Se content in the shoots only accounted for 27.5%–68.8% of the total Se content of the plant. The root/stem distribution coefficient was 0.16–0.84 (Table 3), with an average value of 0.43, which was slightly higher than the root/soil distribution coefficient (0.41). This suggests that after the different rice roots absorb Se from the soil, the ability to translocation it to the aboveground plant parts is quite different. The average grain/stem distribution coefficient was 0.74, indicating that the Se translocation capacity between grains and stems is greater than that between stems and roots.

Table 3. Relationship between the distribution of Se content in different parts of rice plant tissues and soil

ratio	ADC	range	CV	SE
root/soil	0.41	0.04–1.41	0.81	0.02
stem/root	0.43	0.16–0.84	0.53	0.01
leaf/stem	0.54	0.18–0.78	0.27	0.01
grain/stem	0.74	0.47–1.04	0.21	0.01

ADC: average distribution coefficient; CV: coefficient of variation; SE: standard error

Figure 2. Distribution and concentration of Se in rice plant tissues. IQR: interquartile range

8. Types and potential influencing factors of soil Se content

The TSe in the soil was 0.11–0.73 mg/kg, and the average proportions of the four types of Se in TSe were: 72.5% O-Se (0.39 mg/kg), 3.6% L-Se (0.19 mg/kg), 1.0% W-Se (0.005 mg/kg), and 22.9% R-Se (0.12 mg/kg) (Figure 3). O-Se was the most represented type of Se in 93.8% of soil samples, accounting for more than half of the total amount, being much higher than other types. Among the four types of Se in the soil, the correlation coefficient between O-Se and TSe was high (r = 0.87, p < 0.01) (Table 4), indicating that the O-Se in the soil directly affects the TSe content and that the enrichment of Se is closely related to the activities of SOM.

Table 4. Pearson correlation matrix of soil properties, Se concentration in rice plant tissues, and soil Se type in Guangxi

	Total S	pH	CaCO3	TSe	O-Se	L-Se	W-Se	R-Se	Root	Stem	Leaf	Grain
SOM	0.360	0.649**	0.068	0.701**	0.726**	−0.393	−0.310	0.188	0.487	0.457	0.403	0.650**
Total S		−0.168	−0.273	−0.006	0.086	−0.208	−0.403	−0.122	0.196	0.005	−0.049	0.135
pH			0.209	0.633**	0.584*	−0.090	−0.051	0.257	0.473	0.350	0.449	0.440
CaCO3				0.191	0.335	0.053	0.354	−0.212	0.032	−0.307	−0.200	−0.322
TSe					0.868**	−0.111	0.059	0.492	0.430	0.466	0.506*	0.428
O-Se						−0.052	0.222	−0.001	0.629**	0.545*	0.534*	0.468
L-Se							0.440	−0.249	−0.042	−0.240	−0.285	−0.424
W-Se								−0.339	0.198	0.125	0.167	−0.227
R-Se									−0.228	0.008	0.108	0.093
Root										0.577*	0.588*	0.532*
Stem											0.865**	0.833**
Leaf												0.792**

* significant at the 0.05 level; ** significant at the 0.01 level

Figure 3. SOM and content of each type of Se in paddy soil of the 16 sampling sites

9. Discussion

The soil Se content is generally used to classify the soil type. According to Zhang et al. (2014c) and Zhang et al. (2014b), who proposed that soil with a Se content of 0.4–3.0 mg/kg is considered Se-rich, 87.5% of our samples were Se-rich soil, as their Se content exceeded 0.40 mg/kg. Although this value is lower than the soil Se level (0.85–11.46 mg/kg) in the ‘Se capital’ in the Enshi region (Chang et al. 2019), it is higher than the average soil Se in China (0.29 mg/kg) (Chen et al. 1991). The Se content in the roots (59.4%) was greater than the sum of the Se content in other plant parts, indicating that Se accumulates in the roots and is not easy to translocation, which is consistent with the findings of previous studies (Chang et al. 2019; Huang et al. 2015). Sun et al. (2010) reported that the Se content in grains in Enshi is between 2.2 and 8.4 mg/kg, which is much higher than the Se content of 0.01–0.07 mg/kg found in Guangxi. Since food items with a Se content exceeding 0.3 mg/kg represent a risk for poisoning (Chang et al. 2019), our findings demonstrate that the rice planted in the study area has good edible value.

The root/stem distribution coefficient (0.43) was higher than the root/soil (0.41), but lower than the grain/stem distribution coefficient (0.74). This may have been caused by the presence of Se in the form of selenite in the roots. Se in acid (Elrashidi et al. 1987) and hypoxic soils (such as flooded rice fields) (Zhu et al. 2009) mainly exists in the form of selenite, which is often converted into selenomethionine, its oxide, selenomethylcysteine, and other derivatives; these types of substances mainly concentrate in the roots, as their translocation to the aboveground plant tissues is difficult (Li, McGrath, and Zhao 2008). In addition, the average and median Se content in rice grains was higher than that of leaves, but slightly lower than that of stems, resulting in a grain/stem distribution coefficient (0.74) greater than that of grain/leaf (0.54), indicating that the translocation of Se may have apical dominance or reproductive development advantages. Sun et al. (2010) reported that the average grain/stem distribution coefficient of rice was 0.75 in four villages in Hubei, China, ranging between 0.28 and 1.59, which is consistent with our results. This is much lower than the average grain/stem distribution coefficient of 2 found in the Se-polluted area of Yutangba, Enshi (Zhu et al. 2008). Eiche et al. (2015) found that the Se concentration in wheat and mustard plant tissues was 15–30 times that in the soil, which indicates that the accumulation levels of Se in plants are varied. This variation may be explained by two factors: different crops have different Se accumulation mechanisms, and soil Se pollution increases the ability of plants to accumulate Se; this requires further research. Moreover, the Se concentration in the stem tissues was significantly positively correlated with that in leaves (R² = 0.75, p < 0.01) and grains (R² = 0.69, p < 0.01) (Figure 4B), had different distribution coefficients, and the correlation between the Se concentration in stems and leaves was slightly higher than that between stems and grains.

Figure 4. (A) Relationship between TSe, SOM, and O-Se in paddy soil. (B) Relationship between Se content in leaf, grain and stem in rice tissues

The paddy soil in this study area is acidic under anaerobic conditions, and the soil pH value had a significant correlation with the content of TSe (r = 0.63, p < 0.01) and O-Se (r = 0.58, p < 0.05). Eich-Greatorex et al. (2007) reported that soil pH had a negative correlation with the Se content of wheat grown in peat soil, while the Se content of wheat grown in loam soil had a positive correlation. This may be related to the increased plant resistance provided by Se. In environments that are unsuitable for plant growth, Se helps ensure the normal growth and development of plants, as well as the absorption and utilization of minerals, while enhancing their resistance to biotic and abiotic stresses (Reynolds and Pilon-Smits 2018). Although soils with acidic and anaerobic conditions have a stronger ability to fix Se, plants present a lower Se absorption rate (Johnson et al. 2000; Cao et al. 2001; Sharma et al. 2015). The translocation of Se from plant roots to aboveground parts is mainly achieved through the xylem (Zhu et al. 2009; Li, McGrath, and Zhao 2008; Wolswinkel 1992). However, most nutrients are imported into the seeds from the source leaves along the phloem pathway (Zhang et al. 2007; Patrick and Offler 2001). As such, like other nutrients, the accumulation of Se in rice grains may be primarily carried out through the phloem. In other words, during grain development, Se compounds are transported to the developing rice seeds through the phloem and combined with the protein or starch granules (Moore et al. 2010). In this study, the correlation between the Se content in leaves and grains also reached an extremely significant level (r = 0.65, p < 0.01), indicating that leaves directly or indirectly participated in the process of Se accumulation in grains.

The changes between L-Se, W-Se, and R-Se within the TSe were extremely small, and the correlation was not significant. It shows that the W-Se in most soils in Guangxi is basically in a state of dissolution equilibrium, while SOM pellets in the soil have a strong adsorption capacity for Se. The proportion of R-Se in the TSe was second only to O-Se, indicating that a considerable amount of Se is attached to the mineral lattice and cannot be absorbed by crops. Bisbjerg (1972) found a strong correlation between W-Se and crop absorption in pot experiments, but not in field experiments. In this study, the W-Se content was 0.005 mg/kg, which accounts for 1% of TSe. This value is close to the average level in farmland soils in China (0.006 mg/kg) (Tan et al. 2002), while the proportion of W-Se is almost equal to that of Enshi (1.14%) (Chang et al. 2019). O-Se in soil was significantly positively correlated with TSe (R² = 0.75, p < 0.01) and SOM (R² = 0.53, p < 0.01) (Figure 4A), suggesting that the concentration of O-Se determines the content of TSe in the soil, and that SOM has a major impact on the concentration of O-Se and TSe, which is similar to the findings of previous reports (Johnsson 1991; Li et al. 2017). O-Se is the main potential source of soil Se for plants. When the balance between the original Se types in the soil is destroyed, O-Se can be partially released as a fraction from the Se complex bound by SOM (Li et al. 2017). SOM has a dual role in regulating the bioavailability of soil Se: it reduces it through fixation, which causes plants to reduce Se accumulation, and provides more Se sources through the decomposition of O-Se, which is in line with the dynamic balance of O-Se Se fractionation and decomposition in natural soils (Li et al. 2017).

SOM represents a variety of soil organic compounds containing C, and has two main effects on the soil Se content: SOM itself contains organic Se and it acts on the fixation of Se in the soil, directly affecting the Se concentration in soils with high SOM. SOM not only provides a source of Se, but also increases its fixation, thus increasing the soil Se content. In a high pH environment, the hydrolysis of SOM causes an increase in the soluble Se content (Yanai, Mizuhara, and Yamada 2015). In this study, the correlation between SOM, pH, TSe, and O-Se was significant (Table 4), which confirms this theory. The Se leached in acid soils is primarily of the O-Se type, and the large accumulation of SOM reduces the availability of Se to crops (Gustafsson and Johnsson 1992). This also explains the low Se accumulation capacity of rice plants in the study area. SOM had a very significant correlation with soil TSe (r = 0.70, p < 0.01), which is consistent with the results of previous studies (Bujdoš, Kubová, and Streško 2000; Eich-Greatorex et al. 2007) because, under natural conditions, SOM often acts as an environmental host of anions and fixes Se in the soil (Gustafsson et al. 1993), resulting in the adsorption of Se by the soil. SOM also had significant correlations with pH, TSe, O-Se, and grains; the correlation with O-Se was the highest (r = 0.73, p < 0.01), indicating that the soil SOM content in the study area has a positive effect on the concentration of TSe and O-Se, which corroborates the findings of previous studies (Wang et al. 2016; Martens and Suarez 1997). Although the concentration of TSe in soil is high in some areas, the large fixation due to high SOM content (6–10%) results in a low bioavailability of Se in soil, which affects the absorption of soil Se by plants (Wang and Gao 2001; Floor et al. 2011; Johnson et al. 2000). In this study, the increase of SOM content did not inhibit the plants absorption of Se, but displayed a weak positive correlation with the Se content in the roots (r = 0.30) (Figure S1). This is related to the low SOM content (1.21–3.92%) in the study area.

The correlation between the content of R-Se, W-Se, L-Se, and the other types of Se was not significant, as was that between the content of Se in the various plant tissues. This differs from the situation reported in the Enshi area (Chang et al. 2019). In addition, O-Se not only had a significant correlation with TSe, but also affected the Se content in the roots (r = 0.63, p < 0.01), stems (r = 0.55, p < 0.05), and leaves (r = 0.53, p < 0.05) of rice plants (Table 4), although the Se content in grains did not reach a significant level. Our findings indicated a significant positive correlation between the Se concentration in various plant tissues and that between the stems and leaves was extremely significant (r = 0.87, p < 0.01). In terms of crop quality, the correlation between the Se concentration in the grains and in stems was the highest (r = 0.83, p < 0.01), while the correlations with soil pH, TSe, and O-Se did not reach significant levels. Interestingly, there was a significant correlation between the Se content in the grains and SOM. The correlation between total soil S, CaCO₃, and the various types of Se did not reach a significant level. There was a significant positive correlation between the W-Se and L-Se contents and the O-Se content in maize soils in Shaanxi, China, while the R-Se content had no significant correlation (Wang et al. 2012). In this study, L-Se, W-Se, and R-Se had no significant correlation with other types of soil Se and the Se content in plant tissues, which shows the difference in the relationship between the concentrations of various types of Se in dryland and paddy soils.

9. Conclusion

The soil in the study area was weakly acidic, with a Se content of 0.11–0.95 mg/kg, and 87.5% of the soil samples were Se-rich. The O-Se was the main type of Se, and there was a significant correlation between the soil pH, SOM, TSe, and O-Se. Our findings differ from those of previous studies, as the L-Se, W-Se, and R-Se had little effect on the accumulation of Se in the soil and plants. We also observed that the translocation of Se in plants has reproductive advantages. This study thoroughly explored the distribution characteristics and migration rules of Se in a soil–rice system and provided a scientific basis for the development and utilization of Se resources in Guangxi, as well as for the development of a Se-enriched high-quality rice industry.

Acknowledgments

The authors thank the financial support provided by the Guangxi Science and Technology Program Project of China (GuiKe AA17202037, GuiKe AD19245169, GuiKe AD18281072), PhD. Startup Project of Yulin Normal College (G2020ZK13) and Yulin Science, and Technology Program Project (Yushikeneng 20194301, Yushike 20204038, Yushikete 202020001).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed here.

Additional information

Funding

This work was supported by the Yulin Science and Technology Program Project [Yushikeneng 20194301, Yushike 20204038, 202020001]; Ph.D. Startup Project of Yulin Normal College [G2020ZK13]; Guangxi Science and Technology Program Project of China [GuiKe AA17202037, GuiKe AD19245169, GuiKe AD182810].