Effect of foliar spraying of organic and inorganic selenium fertilizers during different growth stages on selenium accumulation and speciation in rice

Most crops are supplemented with selenium (Se) through the exogenous addition of inorganic Se fertilizer. There is a lack of in-depth research on organic Se fertilizers. Meanwhile, the dosage range between human Se deficiency and Se toxicity is narrow, so the Se content of agricultural products needs to be controlled within a reasonable interval. Se accumulation and speciation in rice were analyzed and compared during three different growth stages (late tillering stage, initiate heading stage, and full heading stage) using three Se fertilizers: selenite, fermented Se, and potassium selenocyanoacetate (Se-AAF) via the foliar application. The Se content in rice sprayed with organic Se fertilizer was controlled in the relatively safe range and met the human Se supplement requirement compared to the sprayed sodium selenite, which was too high of a dose. The percentage of organic Se and protein Se in brown rice was found to be similar in all three Se fertilizers. The highest organic Se content of 91.57% was found in the grain of rice at the full heading stage by spraying Se-AAF. The main Se species in the grain was selenomethionine (SeMet), which reached 80% of the total Se. Se-methyl selenocysteine(SeMeCys) was found only in Se-AAF. The grain quality showed that all three Se fertilizers increased the consistency of gelatinization. Appropriately delaying the spraying time and selecting organic Se fertilizer as the Se source can help to produce green and safe selenium-rich rice.


3
Vol:. (1234567890) 2010). Crops are the main source of the human intake of Se. As the staple food crop, the Se content in rice is directly related to the Se nutritional status of most people. Therefore, increasing Se concentration in rice by applying Se fertilizer is of great significance for improving the Se nutrition intake in human beings (Giacosa et al. 2014).
Se is an essential trace element for both animals and humans (Schwarz and Foltz 2009). The human body can assimilate Se to meet normal human needs through the intake of Se-enriched agricultural products. However, the human body has little demand for it, and agricultural products with excessive Se content can be harmful to the human body (Hatfield et al. 2014). Se holds a narrow threshold between beneficial and harmful effects on the body, with both inadequate and excessive intake of Se affecting one's health, which happens at the appropriate dose interval. If the intake is less than 40 µg/day, the human body is in a state of Se deficiency, but if it is more than 400 µg/day, it can lead to safety risks and even symptoms of Se toxicity (Fordyce 2007;Winkel et al. 2012). Therefore, deciding to have Se as a supplement needs strict control of its amount and time intervals according to the human serum levels. Also, effective measures are needed to reasonably regulate the Se content in crops within the scientific range. It is not enough to just supplement Se; it is also important to supplement it reasonably and precisely.
The main methods of Se biofortification include soil application of Se and foliar spraying of Se. Soil application results in significant Se wastage since 80-95% of selenate is likely to be lost due to irrigation or rainfall (Keskinen et al. 2011). Also, more than 80% of selenite is fixed by the soil only for a short term (Liu et al. 2015), with environmental impacts shown after Se enters the water bodies through surface runoff. Moreover, the success of soil application depends largely on the homogeneity of the physicochemical properties of the soil, including soil structure and soil pH (Hartfiel and Bahners 1988). Foliar spraying eliminates the transportation of Se from the roots to the ground and transports it directly from the leaves to the grains, which is more bioavailable than soil-applied Se. The Se enrichment effect by foliar spray is eight times higher than that of the soil-applied method (Ros et al. 2016). The soil application of selenate results in 70% of the Se getting concentrated in the stalk part of rice while less than 18% enriches the grain (Boldrin et al. 2013). However, the foliar application of selenate enriches the rice grain with 45% of Se (Deng et al. 2017). Hence, the foliar spraying method could promote the uptake and accumulation of Se in the edible part of plants (Fang et al. 2009).
Usually, Se fertilizer is divided into organic Se fertilizer and inorganic Se fertilizer. Among these, inorganic Se fertilizer is mainly selenite and selenate. Organic Se fertilizer is mostly compounded with Se and organic substances such as amino acid chelated Se fertilizer, humic acid Se fertilizer, etc. The absorption of inorganic Se by crops is difficult. Plants can absorb only up to 10% of Se from inorganic fertilizers, and the remaining Se poses a potential threat to the environment (Eich-Greatorex et al. 2007;Wang et al. 2018). Se accumulation by continuous Se fertilizer application causes the increment of soil destruction, the amount of harmful heavy metals in the soil and the soil toxicity, which would completely damage the soil capability to grow plants eventually. Furthermore, when plants absorb inorganic Se, the structure of SeCys and SeMet is similar to cysteine and methionine, which would be mistakenly integrated into the protein, affecting the formation of disulfide bonds, leading to plant damage (Dumont et al. 2006).
Recently, safe and oral taken Se tonic are synthesized such as Se-AAF (Liu et al. 2019). In order to test if Se-AAF is a safer, environmentally friendly and effective Se fertilizer for rice Se foliar biofortification, experiments are designed in this paper to do the evaluations. Three key rice growth stages were selected for the Se foliar spraying experiments, which included later tillering (LT), initiate heading (IH), and full heading (FH) stages. Three Se fertilizers [inorganic Se (selenite), organic Se (fermented Se and Se-AAF)] were compared in terms of total biomass, yield, and total Se content. Se speciation and molecule variations among different fertilizer treatments in the grain contents were analyzed by ICP-MS. Hydroponics culture analysis and field spraying experiments were also conducted to show that Se-AAF was a safer, better organic Se fertilizer in the field applications, providing a new and environmentally friendly way to improve Se-enriched rice.

Plant materials and phenotypic analysis during seedling treatments
The seeds of rice variety "LY9348" (Oryza sativa L.), purchased from Guoying Seed Industry Co., Ltd (Wuhan, Hubei, China), were submerged in 75% ethanol for 1 min and disinfected with 0.15% HgCl 2 for 1 min. After washing with sterile water six times, the seeds were germinated in Petri dishes in the dark for three days at 37 °C. Next, healthy and uniform seedlings were chosen and soaked in a series of Yoshida solutions (1 L, 6 plants per pot) containing 0, 20 µM selenite, and 20 µM Se-AAF for 2 weeks. The pH of the solution was adjusted to 5.5 by HCl or NaOH. Each treatment was replicated in three pots and the solution was changed every three days. Two weeks later, primary root length, shoot length, number of lateral roots, and fresh weight of the shoots were measured.
The foliar spray of Se was applied as selenite, fermented Se, or Se-AAF at 30 g ha −1 . The Se working fluid was prepared at 75 mg L −1 concentration and applied to the foliage of rice plants at the late tillering (LT), initial heading (IH), or full heading (FH) stage (followed the developmental progress in time series, the FH is later than LT and IH), and the control rice plants (CK) were sprayed with only distilled water [The Se application rate of 30g ha -1 amount referred to (Ekanayake et al. 2015), the Se foliar spray experiments design referred to (Deng et al. 2017)]. Both in Hainan and Ezhou, the experiment was performed with a randomized complete block design and in three replicates (Table 1). The size of the plot in Hainan was 25 m 2 (5 m×5 m), and in Ezhou, it was 20 m 2 (4 m×5 m).
Paddy transplantation, irrigation, and other rice farming practices were carried out based on the farmers' experience. The transplanting dates were February 2, 2018, and June 18, 2018, while the harvest dates were May 18, 2018, andSeptember 25, 2018, in Hainan and Ezhou, respectively.

Sample preparation
Eight rice plants were selected randomly from each plot and divided into four parts, including root, shoot, spike axis, and grain. After washing with distilled water, each part of the rice plant was ovendried at 60 °C to obtain a constant weight and subsequently grounded into powder for measurement of the Se content.
Three regions were randomly selected from each plot, and grains from 15 rice plants belonging to each region were sampled. After mixing, the husk was removed to separate the brown rice, whose Se content was measured. Additionally, the protein from the brown rice was extracted, and its Se concentration was measured. Later, the grain yield and total biomass from additional 100 rice plants per plot were also measured.

Measurement of Se concentration
Acid digestion was applied to all the rice samples. For total Se concentration: The rice samples were digested by adding HNO 3 -HClO 4 (Volume ratio = 9:1), with the temperature being maintained at approximately 180 °C. The digested solution was restored with 6 mol L −1 HCl, cooled, and filtered at a set volume. Organic Se concentration was determined using the cyclohexane extraction method (Sun et al. 2013). The follow-up was consistent with the measurement method of total Se concentration. The obtained data indicated inorganic Se concentration; however, the organic Se concentration could be obtained indirectly by subtracting the inorganic Se content from the total Se content.
To determine the protein Se concentration, 30.00 g of brown rice flour was first weighed and added to a 250 mL Erlenmeyer flask, to which 150 mL of 0.2% NaOH (aq) (brown rice powder and sodium hydroxide solution at a ratio of 1:5) was added. The solution was then stirred with the glass rod evenly and placed on a shaker at 40 °C. After shaking for 1 h at 100 rpm, the solution was centrifuged at 3800 rpm for 10 min. Then, the supernatant was collected, and the step was repeated to extract the complete protein. The two supernatants were pooled together, and 0.1 M HCl was added to it. The pH was adjusted to an isoelectric point of 4.8 (Souza et al. 2016). The mixture was placed in a refrigerator for 1 h at 4 °C and centrifuged at 3800 rpm for 15 min at 4 °C. The obtained white precipitate was the protein which was washed thrice with distilled water and centrifuged at 4000 rpm for 5 min to remove the impurities. Later, the precipitate was put in an oven (AFD-270L-200, AoFeiDa Instrument and Equipment Co.,230 Ltd., China) and baked for 36 h at 30 °C. The follow-up was consistent with the measurement method of total Se concentration.
The concentrations of total Se, organic Se and protein Se in the rice samples were measured using hydride generation-atomic fluorescence spectrometry (AFS-230E, Kechuanghaiguang Instruments Co., Beijing, China). The soil pH, alkali nitrogen, and total Se concentration were commissioned by the Hubei Provincial Geological Experimental Testing Center. The instrument used was AFS-820. The whole Se speciation analysis of brown rice referred to (Bañuelos et al. 2012).

Determination of the characteristics of the grain quality
To determine the amylose content (AC), gelatinization consistency (GC), gelatinization temperature (GT), and alkali spreading value (ASV), the brown rice powdered samples were scanned using the Near-Infrared Grain Analyzer Perten DA7250. Note that the sample volume is constant each time, and the test was repeated twice with each sample. According to the accuracy requirements, the average value for the samples was calculated if the ratio of the difference between the two tests resulted in an average value of less than 2%. If the requirements were not met, the samples were re-tested, and the average value was calculated.

Calculation methods
The Se recovery efficiency (%) of the whole plant and Se recovery efficiency (%) of the brown rice were based on the plant Se uptake. Details and procedures were described in studies of (Deng et al. 2017).
% RDA was based on the total Se content of 50 g brown rice, which is according to 55 g RDA for adults in the USA. Details were described in the studies of (Monsen 2000).

Statistical analysis
The statistical analysis was performed using Gen-Stat 18 (18th Edition, VSN International Ltd., Hemel Hempstead, UK) and SPSS 20.0. The data were presented as mean ± standard error (SE). The mean values were compared using the Least Significant Difference test at the 0.05 level of probability. The Principle component analysis (PCA), partial least squares-discriminate analysis (PLS-DA), and variable importance in projection (VIP) were conducted using the "Statistical Analysis" module on MetaboAnalyst (Xia and Wishart 2011).

The hydroponics experiment conducted for different Se treatments
To investigate whether the organic Se fertilizer was toxic to the seedlings, the hydroponics experiment was conducted. Without adding exogenous Se, the average primary root length of rice was 5.2 cm after culturing for two weeks. Upon addition of 20 µM selenite, the growth of rice was strongly inhibited (Table 2), shortening the length of the root to an average of 3 cm. Also, the total number of lateral roots increased significantly to an average number of three compared to the control. The length of the seedlings became significantly shorter, dropping nearly 4 cm. The fresh weight of the aerial part reduced by 1.5 times compared to the control. However, upon culturing rice with 20 µM Se-AAF, no difference was observed in terms of the root length, shoot length, shoot fresh weight, and the number of lateral roots compared to the control, indicated no inhibition on the growth of rice plant.
The above phenotypic data were used for PCA and PLS-DA analysis (Fig. 1). The sample dots in Fig. 2b were discriminated by PC1, accounting for 95.1% of the variation along with showing separate clustering trends among CK, 20 µM Se-AAF, and 20 µM selenite. The treatments were distributed almost separately on the first component (Fig. 2a). Moreover, component 1 suggested that seedling traits were similar between CK and 20 µM Se-AAF treatment. The high concentration of organic Se exhibited no toxicity to rice at the seedling stage. Additionally, based on the values of the variable importance in the projection (VIP) of Component 1, a relatively stringent Green crisscrosses represent seedlings with 20 µM selenite treatment while the red triangles represent the control experiment, and the blue multiplication sign represents 20 µM Se-AAF treatment, (b) 2d score plots of PCA. The green circle represents seedlings with 20 µM selenite sodium treatment, while the red circle represents the control experiment. The blue circle indicates 20 µM Se-AAF treatment, (c) VIP scores level of VIP cutoff of 1.0 was chosen and the shoot weight (fresh weight) was found to be suffered serious impact when treated with 20 µM selenite. The toxic effect of selenite was mainly manifested in the aerial part of the plants.

Grain yield and the total biomass
Without Se fertilizer, the total yield and total biomass were found to be 8.09 t ha -1 and 16.06 t ha -1 in Hainan and 7.52 t ha -1 and 16.59 t ha -1 in Ezhou.
Although the kind of Se fertilizer being sprayed on the plant did not affect the yield, spraying with Se-AAF was shown to increase the total biomass by 9.8%. However, the grain yield and total biomass did not change significantly with Se spraying time. An interaction was observed between Se and the environment in the case of total biomass (Table 3).
Total Se, organic Se, and protein Se concentrations and proportions in brown rice As shown in Fig. 2a and b, the order of total Se concentration in brown rice under different Se spraying periods was FH > IH > LT. The result indicated that when the Se spray time was delayed, the accumulation of Se in brown rice was increased. The Se concentration of Se-sprayed brown rice at FH all exceeded 200 μg kg −1 , reached the standard of Se-enriched rice (In the control case, the Se concentration of brown rice in Hainan and Ezhou was 62 μg kg −1 and 54 μg kg −1 , respectively). Upon spraying of Se fertilizer during the LT stage, the Se content of the brown rice was found to be higher than that of the husk. However, upon delaying the Se spraying time, an increase was observed in the overall Se level in both brown rice and husk, with the Se concentration in the husk showed a greater increase, which indicated a decrease in the distribution of Se in brown rice by delaying the Se spray. Irrespective of the kind of Se fertilizer being sprayed, delayed the Se spray accounted for more than 70% of the brown rice's Se content (Fig. 2c). Similar to the results of total Se, the organic Se and protein Se content in brown rice also showed a gradual increase in its content with a delay in Se spraying. Interestingly, although the Se content in selenite sprayed plants was about twice the amount compared to the organic Se fertilizer spray, no significant differences were found among different Se fertilizer treatments since organic Se was all detected as higher than 85%. It was worth mentioning that when Se fertilizer was sprayed during the FH stage, the organic Se ratio of Se-AAF reached up to 91.57% (Table 4). Protein Se also behaved similarly; however, the lowest protein Se content was found upon spraying Se during the FH stage. The proportion reached up to 34% (protein Se) regardless of the Se source or spraying stage.

Se speciation in brown rice
Since the highest Se content was found during the FH stage, brown rice from this stage was selected to determine the Se speciation. The HPLC-ICP-MS chromatogram was shown in Fig. 3. As shown in Fig. 4, upon no addition of exogenous Se, three Se forms were found in brown rice, including SeMet, with a proportion of 62.71% (SeMet concentration: 31.9 μg kg −1 ), followed by SeCys 2 with a proportion of 28.83% (SeCys 2 concentration: 14.7 μg kg −1 ). After being treated with Se fertilizer, the proportion of each Se speciation changed, increased the SeMet content by about 20% while decreased the proportion of SeCys2 to 11.88%, 5.90%, and 9.79% in selenite, fermented Se, and Se-AAF fertilizer sprayed plants, respectively. Also, the proportion of Se 4+ showed a significant decrease. The presence of Se 6+ and different unknown Se compounds were detected in brown rice treated with all Se fertilizers as well. It was interesting that the presence of SeMeCys was found only in brown rice under Se-AAF treatment at a percentage of 5.23%. Hence, the Se source could significantly influence its form within brown rice.
The grain quality characteristics and percentage recommended daily allowance of LY9348 The analysis of the grain quality characteristics showed that there was no regular change in the grain quality among the type of Se fertilizers used or the time at which Se type was sprayed (Table 5). However, the addition of exogenous Se generally improved the gelatinization consistency and the highest improvement was found by Se-AAF. The recommended daily intake (RDA) of Se for adults was used as a measure of Se levels. The RDA provided by 50 g of brown rice was shown in Table 6. The control plants were found to have only 5% Se content (no Se applied), which was not enough to provide the amount of daily intake of Se. The RDA was significantly improved by spraying exogenous Se and the improvement was higher in selenite treatment compared with the organic Se fertilizers (Table 6).

Se concentration/distribution in different parts of rice plants and Pearson's correlation analysis
Under normal conditions, the Se concentration in different rice parts followed a decreasing pattern and was as follows: root > shoot > spike axis > grain. The Se content in the grain was found to be as low as 0.05 mg kg -1 . After spraying the Se fertilizer, the Se content of the grain was greatly increased, with the highest content being increased by 20 times. Table 7 showed that no matter whether it was shoot, spike axis, or grains, the Se content order was LT < IH < FH, except for the root, which uniformly showed the highest Se concentration during the initial heading stages. The Se concentration in the grains during the FH stage was at least 2.7 times the concentration found at the LT stage, regardless of the Se fertilizer used. Without external Se application, Pearson's correlation analysis indicated the correlations among root, shoot, spike axis, and grain to be 0.964, 0.932, and 0.22, respectively, showed a decreasing trend (Table 8). Upon foliar spraying with Se, the correlations among root, shoot, spike axis, and grain were found to be 0.272, 0.721, and 0.834, respectively, showed an increasing trend (Table 9).
Upon sprayed fermented Se and Se-AAF, the correlation coefficient increased among root, shoot, spike axis, and grain. The peak value of the spike axis and grain was 0.962 and 0.971, respectively. However, the trends observed in various changes were totally different upon spraying selenite, which showed an initial increase and was followed by a decline. The correlation coefficient of spike axis-grain was only 0.885 (Table 10).

Recovery efficiency of Se in the whole plant and brown rice
The Se recovery of selenite at each stage was higher than that of the organic Se fertilizer. The ranking of Se recovery of brown rice was selenite > Se-AAF > fermented Se during the FH stage. Maintaining the same Se sources, the Se recovery efficiency of the whole plant and brown rice was increased by delaying the Se spraying time. Among them, the whole-plant Se recovery using the fermented Se at the FH stage was 4.56 times higher than that found at the LT stage while the Se recovery of the brown rice at the FH stage was 4.81 times higher than that found at the LT stage, indicated a significant effect of varying time in Se spraying on the Se recovery (Table 11).

Discussion
Lack of Se in human food supplies is a consensus in nutriology research, therefore, biofortification of Se in the food production chain is strengthened in the industrial trials in the past few decades (John 2008;Schiavon et al. 2020;Hossain et al. 2021). Selenite compounds are chemically manufactured to be used as the inorganic Se fertilizers for agricultural trials of Se biofortification. However, its' toxicity to the plants, animals and the application environments including soil and water bodies has also been well identified in the agricultural practices. For food safety and ecofriendly agricultural development, it is very important to find safer (less or no toxicity) and efficient Se fertilizer substitutions. For example, many organic Se fertilizers are produced by mixing selenite compounds with soybean or bean pulps or other organic residuals to go through a series of fermentation procedures (Wang et al. 2017;Ramkissoon et al. 2019). These treatments could somehow reduce the toxicity of pure   selenite compounds but not get rid of its toxicity completely. In this study, potassium selenocyanoacetate (Se-AAF) (Liu et al. 2019), which is an oral security organic Se and used as health care product for the human body, is tested as a safe and truly organic Se fertilizer in Se biofortification of rice for the first time.
Se-AAF was proven to be a safe and non-toxic rice Se fertilizer Hydroponic experiments showed that there were no significant differences between Se-AAF solution and normal cultivation solution in plant growth and development phenotypes such as root length, lateral root number, shoot length and shoot weight (Fig. 1a, Table 2), however, plant growth was inhibited in selenite treatment significantly which reflected the toxicity of selenite to plants. PCA analysis also categorised Se-AAF with normal cultivation in the same group but distinguished selenite treatment out (Fig. 1b). Among all the tested growth traits, shoot weight was severely impacted by selenite toxicity (Fig. 1c). Therefore, Se-AAF can be used as a safe and non-toxic rice Se fertilizer which was different from selenite.
Field experiments showed that Se-AAF was a better rice Se fertilizer to replace selenite First of all, yield and biomass were not significantly influenced by Se foliar spraying in different growth stages and locations. However, Se-AAF application showed no influence in gain yield but increased the total biomass by 9.8%, which was not influenced in the selenite and fermented Se treatments (Table 3). Secondly, total Se, organic Se and protein Se accumulation in rice grains were not influenced by locations but varied by fertilizer types and treated growth stages (Fig. 2). FH stage was the best stage for Se-AAF, the organic Se and protein Se reached 0.388 and 0.148 mg kg −1 separately. Although its'absolute amount was lower than the selenite treatment, the proportion of organic Se to total Se was higher in Se-AAF (91.57%) than in selenite (89.51%); the proportion of protein Se to total Se was lower in Se-AAF (34.70%) than in selenite (37.22%). Therefore, this indicated that there was a larger proportion of free organic Se than integrated organic Se presented in rice grains treated by Se-AAF than by selenite (Table 4). Se ingested by the human body can be   (Thiry et al. 2012). Therefore, it is important to not only analyze the total Se concentration in the edible parts of the plants but also identify the different types of Se forms available to avoid harmful effects on humans. Generally, the inorganic form of Se is the most harmful to the human body (Hatfield et al. 2014;Yin et al. 2019), while the organic Se compounds are better absorbed and utilized by humans. Organic Se and protein Se amount were found to be the highest in the Se-AAF treatment, which means there were more human usable organic Se existed in Se-AAF treatment compared with the other Se fertilizers. This is an important factor to determine the value of Se biofortification products.
Thirdly, different Se forms were present in plants, which mainly included selenite, selenate, SeMet, and selenocysteine. Most of the absorbed Se by plants from external sources were converted into SeMet and selenocysteine (Zhu et al. 2009). The Se species detected by ICP-MS showed both SeMet and SeCys molecules were detected in all Se fertilizer treatments. However, only in the Se-AAF treatment, SeMeCys, which was the most effective seleno-compound identified in the reduction of tumors so far (Carey et al. 2012;Ellis and Salt 2003), was detected but not in the other two Se treatments. These results were consistent with the previous findings (Hu et al. 2018;Liao et al. 2016) of SeMeCys being not detected in the rice sprayed with selenite. However, SeMeCys was only found in brown rice sprayed with Se-AAF, which could be an impactful way to provide human Se requirements ( Fig. 3 and Fig. 4) and largely increase the biofortification and consumption quality of Se rice products. Results also showed that Se-AAF treatment (organic Se), compared with fermented Se treatment (partial organic Se), was better in the Se absorption rate, Se species contents and Se distributions.
Fourthly, the grain quality such as gelatinization consistency was elevated in all Se fertilizers but higher in the Se-AAF treatment than in selenite. No significant differences in other checked grain quality traits among different fertilizer treatments in different growth stages (Table 5).
Finally, overall Se concentration in different growth stages ranged 310 -907 μg kg −1 by selenite, 163 -391 μg kg −1 by fermented Se and 158 -423 μg kg −1 by Se-AAF (Table 6). The maximum intake of Se cannot exceed 400 µg per day (Pedrero and Madrid 2009). Se-enriched rice is a dangerous "double-edged sword" (Brozmanová et al. 2010). Its'concentration must be controlled precisely to a certain threshold. According to the "health standard of Se limit in food", the Se content in the Se-rich processed rice should be between 0.04 and 0.30 mg kg −1 while the Se content of grain (finished grain) should be less than or equal to 0.30 mg kg −1 . If taken this criterion into consideration, selenite fertilizer was shown to be hard to control the final biofortification concentration since the maximum concentration of Se easily can be increased to 900 µg kg −1 , which was already on the verge of danger. However, the Se content in rice sprayed with organic Se fertilizers was higher than the criterion but lower than the red line, which means that they were sufficient to provide daily intake with no risk of excessive consumption.
Based on the five aspects analyzed above, Se-AAF was not only non-toxic to the plants, it also did not decrease the yield but increased the biomass, improved organic Se proportion, accumulated SeMeCys which was the valuable molecule to improve human health, elevated grain qualities and precisely controlled the Se total content into the required recommended daily allowance criteria. Therefore, Se-AAF is a better rice Se biofortification fertilizer to replace selenite.
Foliar spray was a more efficient application way for Se-AAF The foliar spray of Se fertilizer, other than direct soil application, was also proved to be efficient, less cost and eco-friendly way for Se application, especially for Se-AAF ( Fig. 1 and 2), which was consistent with previous studies (Wang et al. 2015;Ramkissoon et al. 2019). Se accumulation was largely influenced by the Se supply source and plant developmental stages. When Se was sprayed at the grain-filling stage, the Se content was twice as high as that at the booting stage in wheat grains (Galinha et al. 2015). Besides, the Se content of grains with selenate applied in the soil before sowing was 4.5 times higher than that of grains with selenate applied at the flowering stage (Paulo et al. 2013). Interestingly, the distribution of Se was largely improved by foliar spraying from staw through spike axis to grains, showing a decreasing concentration trend by Se-AAF but the pattern was more complicated in selenite treatment (Table 7). Pearson's correlation coefficients analysis showed that without foliar spraying, the Se accumulation in grains was tightly correlated with root (0.964) and shoot (0.932) but less correlated with spike axis (0.22) (Table 8). However, the situation was reversed. In foliar spraying, the Se accumulation in grains was tightly correlated with spike axis (0.834) but less correlated with shoot (0.721) and root (0.272) ( Table 9). The highest correlation between grains and spike axis was detected in Se-AAF (0.971) but the correlation was only 0.885 in selenite (Table 10), which suggested that the transportation of foliar sprayed Se from leaf through spike axis to grain was more smoothly and effective by Se-AAF than selenite.
The recovery efficiency of Se-AAF was lower than selenite but higher than fermented Se The highest Se recovery efficiency was detected in the FH stage than other stages in all Se fertilizer treatments both in the whole plant and in brown rice Se evaluations (Table 11). In FH, whole plant Se recovery efficiency was 41.6% in selenite, 21.2% in Se-AAF and 19.6% in fermented Se. Since the cost of Se-AAF is 6.5 times more than that of the selenite for 30 g ha −1 , it is important that dosage analysis of Se-AAF should be conducted to find an effective but economical concentration for spraying this fertilizer in the future.
Based on the above analysis, it is reliably to conclude that the usage of Se-AAF as Se fertilizer is a better choice to replace selenite or fermented organic Se fertilizers for the food Se supplementation since it is an eco-friendly, safer and effective Se fertilizer, which could provide more bioactive Se compounds to human health.

Conclusion
The distribution and accumulation of Se in rice plants are different for different Se fertilizers and different Se spraying time points. The hydroponics experiment proved that the Se-AAF fertilizer is safe and harmless, with no side effects on the growth of the plant. For all Se fertilizers, the appropriate Se spray time can be postponed to enrich the Se content in the grains, which indicates the full heading stage is the best stage to spray. All treatments shows that the grain reaches a Se concentration of more than 200 µg kg −1 . Irrespective of the total Se, protein Se, and organic Se accumulation are higher in the Se-AAF treatment. Among three tested Se fertilizers, Se-AAF treatment shows no harm to the plant development, no impact on grain yield but increment in the biomass, increased grain quality, and specific SeMeCys molecule concentration compared with the other two fertilizers. Hence, it can be concluded that the usage of Se-AAF is a better choice for rice Se biofortification, therefore, a better Se fertilizer supplementation of selenite. Data availability All data generated or analysed during this study are included in this published article.
Code availability Not applicable.

Conflicts of interest
The authors have no conflicts of interest to declare that are relevant to the content of this article.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.