Differentiated extraction strategies for selenium speciation in hyperaccumulator and non-hyperaccumulator plants via orthogonal LC-AFS Authors: Yuhua Duan, Guangxuan Du, Tao Li, Hua Zhang, Zezhou Zhang, Linxi Yuan Published: 2026 ABSTRACT The growing demand for selenium (Se)-enriched products highlights the need for accurate and accessible Se speciation analysis. In this study, a cost-effective orthogonal liquid chromatography-atomic fluorescence spectrometry (LC-AFS) method was developed for the determination of seven Se species (SeO32−, SeO42−, SeLan, SeMet, SeCys2, MeSeCys, and γ-GluMeSeCys), providing an alternative to ICP-MS-based approaches. Differences in Se accumulation mechanisms between hyperaccumulator Cardamine hupingshanensis and non-hyperaccumulator wheat significantly affected extraction efficiency. The sequential aqueous-enzymatic extraction increased SeLan recovery in C. hupingshanensis from 5.06 to 143.74 μg Se g−1, a 28.41-fold increase, while anaerobic extraction enhanced SeMet recovery in wheat from 98.73 to 112.17 μg Se g−1. Consequently, matrix-specific strategies improved Se recoveries from 20.82% to 80.65% in C. hupingshanensis and from 34.11% to 90.02% in wheat. These findings highlight the importance of matrix-specific extraction for accurate Se speciation analysis. Introduction Selenium (Se) is an essential trace element that plays an important role in antioxidant defense systems and thyroid hormone metabolism (Rayman, 2012). However, Se deficiency remains a widespread health concern, particularly in regions with low soil Se levels and insufficient dietary Se intake (Shreenath et al., 2025). Consequently, Se-enriched foods, such as Se-enriched rice and wheat products, have attracted considerable attention as an effective strategy to alleviate Se deficiency. Notably, the chemical form of Se in these products is a critical factor influencing its bioavailability, biological activity, and safety (Sele et al., 2018). Numerous studies have shown that organic Se species generally exhibit higher bioavailability and lower toxicity than inorganic Se species (Lei et al., 2022), highlighting the importance of Se speciation analysis for the evaluation and development of Se-enriched foods. In recent years, plant-based Se biofortification has emerged as an important approach for increasing dietary Se intake through organic Se sources (Wu et al., 2015). However, Se metabolism differs substantially among plant species, resulting in distinct Se distribution patterns. In non-hyperaccumulator crops such as wheat and rice, inorganic Se is primarily assimilated into selenomethionine (SeMet) and selenocysteine (SeCys), which can be non-specifically incorporated into proteins (Fig. 1) (White, 2018). In contrast, Se hyperaccumulators, including Cardamine hupingshanensis (accumulating up to 8000 mg Se kg−1 dry matter (DM)), possess specialized metabolic pathways (Wang et al., 2024; Wu et al., 2020; Yuan et al., 2013). Rather than incorporating them into proteins, these plants predominantly convert Se into non-toxic free selenoamino acids (Se-AAs), such as methylselenocysteine (MeSeCys) and selenolanthionine (SeLan) (Fig. 1) (Both et al., 2020; White, 2018). Currently, driven by increasing demand for Se-enriched foods, both Se-biofortified crops and edible hyperaccumulator plants have been increasingly commercialized (Tangjaidee et al., 2023; Zang et al., 2021). Se-rich plant products, including Brazil nuts, rice, and corn-derived products, have been shown to enhance glutathione peroxidase (GPx) activity and increase serum Se concentrations, with beneficial effects demonstrated in human clinical trials (Giacosa et al., 2014; Stockler et al., 2010; Yuan et al., 2022). In addition, C. hupingshanensis has been investigated as a feed additive to produce Se-enriched meat and eggs (Yu et al., 2022). Moreover, Se-enriched bioactive compounds such as polyphenols, peptides, and polysaccharides may also contribute to the nutritional value and biological activity of Se-enriched foods (Chomchan et al., 2018; Ma et al., 2018). For example, Se-enriched peptides isolated from C. hupingshanensis have been reported to increase GPx activity and reduce the levels of tumor necrosis factor-α, interleukin-6, and interleukin-1β in human colon cells, suggesting anti-inflammatory and antioxidant properties (Zhang et al., 2022). However, there remains a lack of comprehensive understanding regarding the Se speciation in these products, with some reports even presenting contradictory findings (Peng et al., 2022).Fig. 1Compared the primary Se metabolic pathway in hyperaccumulator and non-hyperaccumulator plants.Fig. 1 Compared the primary Se metabolic pathway in hyperaccumulator and non-hyperaccumulator plants. Accurate Se speciation relies heavily on effective chromatographic separation. Strong anion exchange (SAX) chromatography coupled with inductively coupled plasma mass spectrometry (ICP-MS) or atomic fluorescence spectrometry (AFS) has been widely used for the determination of inorganic Se species (SeO42−, SeO32−) and several organic compounds, including SeCys2, MeSeCys, SeMet (Farooq et al., 2024; Ma et al., 2023; Wang, Kong, et al., 2022; Wu et al., 2020). However, this analytical approach does not adequately cover certain Se metabolites, particularly SeLan and γ-glutamylmethylselenocysteine (γ-GluMeSeCys). SeLan is a characteristic free selenoamino acid involved in Se storage and detoxification in C. hupingshanensis, whereas γ-GluMeSeCys is an important intermediate in plant Se metabolism. Failure to detect these species may lead to an underestimation of non-protein-bound organic Se and an incomplete characterization of plant Se speciation profiles. Furthermore, several studies have highlighted limitations of SAX-based separation, particularly for SeCys2 (Dernovics, 2024; Larsen et al., 2001). On SAX columns, SeCys2 typically elutes near the column void volume (approximately 2.5 min), indicating poor retention and a high risk of co-elution with matrix components. To overcome this limitation, orthogonal liquid chromatography (LC) systems integrating SAX and strong cation exchange (SCX) chromatography have been developed, primarily in combination with ICP-MS detection (Both et al., 2018; Larsen et al., 2001; Polatajko et al., 2005; Yathavakilla et al., 2005). However, the high cost of ICP-MS restricts its routine application in many food analysis laboratories (Castro et al., 2017; Chen & Belzile, 2010). In contrast, AFS provides a sensitive and cost-effective alternative but has largely been limited to SAX-based separations (Souza et al., 2022). Therefore, integrating orthogonal SAX/SCX chromatography with AFS offers a practical approach for expanding Se speciation coverage, improving the separation of organic Se species, and enhancing the accessibility of Se speciation analysis. Beyond instrumental configuration, sample extraction represents a critical yet often overlooked step in Se speciation analysis. In recent years, matrix-specific extraction strategies have gained increasing attention for improving the recovery of bioactive compounds from plant materials. For example, ultrasonic-assisted extraction has been applied to enhance the recovery of polyphenols from fruit peels by disrupting plant cell walls and facilitating the release of intracellular constituents (Tahir et al., 2024). Similarly, accurate Se speciation requires extraction protocols that are tailored to the chemical forms and accumulation characteristics of Se in different plant matrices. Conventional extraction approaches typically apply uniform procedures, such as standard enzymatic hydrolysis, regardless of matrix composition (Ma et al., 2022; Wang et al., 2024). However, this approach ignores the distinct Se metabolic pathways and storage forms among plant species. Hyperaccumulator plants predominantly store Se as soluble free Se-AAs, whereas non-hyperaccumulator plants store Se in protein-bound forms that require enzymatic release (Wang et al., 2019). Consequently, the use of inappropriate extraction procedures may result in the degradation, transformation, or incomplete recovery of Se species. For example, excessive enzymatic hydrolysis may compromise the stability of labile free Se-AAs, whereas insufficient extraction may hinder the efficient release of protein-bound Se. Therefore, developing differentiated extraction protocols tailored to the Se speciation characteristics of different plant matrices is essential for improving detection accuracy. To address these analytical challenges, this study systematically evaluated seven representative Se species in plants, including inorganic Se species (SeO32−, SeO42−) and organic Se species (SeLan, SeMet, SeCys2, MeSeCys, and γ-GluMeSeCys). By exploiting the complementary separation capabilities of SAX and SCX chromatography, SeO32− and SeO42− were separated using SAX chromatography, while five organic Se species were efficiently separated using SCX chromatography. The quantification of all seven Se species was subsequently achieved by LC-AFS, enabling sensitive and selective detection. Furthermore, recognizing the distinct Se accumulation and storage patterns in leaves of hyperaccumulator C. hupingshanensis and in grains of non-hyperaccumulator plants e.g., wheat, different extraction procedures were evaluated and optimized. This study aims to develop a comprehensive and accessible strategy for Se speciation analysis in diverse plant matrices, thereby supporting accurate nutritional assessment of Se-enriched products that will be produced from future Se biofortification strategies. Materials and methods Standards and chemicals Ultrapure water (18.3 MΩ·cm) was produced using a Mingche-D 24UV water purification system (Merck Millipore, Germany). Standards of selenious acid (SeO32−; ≥99.9%), selenic acid (SeO42−; ≥99.0%), methyl-selenocysteine (MeSeCys; ≥95.0%), selenomethionine (SeMet; ≥99.0%), and selenocystine (SeCys2; ≥98.0%) were purchased from the Sigma-Aldrich (Merck, Darmstadt, Germany). Selenolanthionine (SeLan; ≥98.0%) and γ-glutamylmethylselenocysteine (γ-GluMeSeCys; ≥98.0%) were synthesized by TargetMol (Boston, USA). Since SeOMet was not commercially available as a certified quantitative standard, a laboratory-prepared SeOMet reference solution was generated by H2O2 oxidation of SeMet according to previously reported methods (Wang, et al., 2022; Zainal et al., 1996). For the enzymatic extraction, tris-(hydroxymethyl)-aminomethane (Tris; ≥99.8%) and Protease XIV (≥3.5 units/mg solid) were obtained from the Sigma-Aldrich. Pyridine (≥99.5%), ammonium phosphate ((NH4)2HPO4; ≥99.0%), and formic acid (≥99.5%) were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China), used for liquid chromatography (LC) analysis. Sodium borohydride (NaBH4; 94.0%), potassium hydroxide (KOH; ≥85.0%), potassium iodide (KI; ≥99.0%) and hydrochloric acid (HCl; 36.0–38.0%) were also obtained from Sinopharm Chemical Reagent Co., Ltd. to generate the hydrogen gas (H2) as a reaction gas. The mixed gas of Argon (Ar; 99.9%) supplied by Air Liquide Gas Co., Ltd. (Shanghai, China) was used as the carrier gas for AFS measurements. Sample collection Seedlings from C. hupingshanensis (provided by Enshi Se-Run Material Engineering Technology Co., Ltd. (Enshi, China)) were transplanted into hydroponic solution containing 80 mg Se L−1 (as SeO32−) for 14 days. The nutrient solution was renewed once a week. Plants were grown under controlled conditions with day/night temperatures of 25/18 °C, relative humidity of approximately 70%, a light/dark period as 14/10 h, and a photosynthetic photon flux density of approximately 150 μmol m−2 s−1. After 14 days of cultivation, leaves were harvested, freeze-dried, and ground into powder for further experiments. Wheat grain samples were provided by Yangzhou University (Yangzhou, China). At the heading stage, wheat plants had been cultivated in soil supplemented with 40 mg Se kg−1 (as SeO42−) for 31 days. Upon receipt, the grains were washed with deionized water, freeze-dried, and milled. Determination of total selenium content Approximately 0.1 g of each leaf and grain sample was weighed in polytetrafluoroethylene (PTFE) digestion tubes, followed by the addition of 4.5 mL HNO3 and 1.5 mL H2O2. The samples were pre-digested on an 8HW-09C heating block (BOTONYC; Shanghai, China) at 80 °C for 30 min, and subsequently digested using a Mars-6 Xpress digestion system (CEM; Matthews, USA). The digestion program was ramped to 190 °C for 30 min and held at 190 °C for 10 min. After cooling to room temperature, the digests were quantitatively transferred into volumetric flasks and diluted to 50 mL with ultrapure water. The resulting solutions were further diluted as necessary to ensure that Se concentrations fell within the calibration range of ICP-MS (NexION 350X; PerkinElmer, USA). Total Se concentrations were determined by ICP-MS using the 78Se isotopes. Quantification was performed using external calibration standards containing 0, 1, 2, 5, 10, and 20 μg Se L−1. Rhodium (103Rh) was used as an internal standard and introduced online during ICP-MS analysis at a final concentration of 20 μg L−1. The certified reference material GBW07602 (GSV-1, Bush Branches and Leaves; certified Se concentration: 0.184 ± 0.013 μg Se g−1) was used for quality control (Wu et al., 2024). The recovery of the reference material was 92.93 ± 3.1%, and the limit of detection (LOD) was below 0.001 μg g−1. Proteolytic extraction Approximately 0.5 g of samples, 15 mg of Protease XIV and 10 mL of ultrapure water were added to 50 mL polypropylene centrifuge tubes. Three different assisted extraction procedures during enzymatic hydrolysis were evaluated as follows: 1) For water bath-assisted extraction, the mixture was incubated in a water bath shaker at 45 °C and 180 rpm for 6 h (Ma et al., 2022); 2) For ultrasound-assisted water bath extraction, samples were treated in an ultrasonic water bath at 37 °C for 45 min (Farooq et al., 2024); 3) For probe-assisted ultrasonic extraction, the mixture was sonicated for 60 s using a 3 mm ultrasonic probe (Cole-Parmer, USA) operated at 50% amplitude (Xiao et al., 2017). Following the respective extraction, the suspensions were centrifuged, and then the supernatants were filtered through 0.45 μm nylon syringe filters prior to Se species analysis. Water extraction The extraction procedure was adapted from the method described by Both et al. (Both et al., 2018). Briefly, 0.5 g of sample was mixed with 10 mL of ultrapure water and vortexed for 1 min. The suspension was subsequently extracted using a probe sonicator (Cole-Parmer, USA) at ambient temperature for two cycles of 1 min sonication. After extraction, the supernatants were collected by centrifugation (4000 ×g, 30 min) and combined. The pooled extract was filtered through a 0.45 μm membrane filter prior to analysis. Enzyme purification A 10 kDa microsep ultrafiltration device (Labselect®-15 mL, Anhui, China) was used for enzyme purification. Aliquots (10 mL) of the enzymatic extract or water extract were centrifuged through the ultrafiltration unit at 4 °C and 3000 rpm for 10 min. The filtrates were used for Se speciation analysis (Ma et al., 2022). Anaerobic operation All procedures involving direct exposure of samples to air were performed in an anaerobic glove box (IG1200/750TS, Vigor, Suzhou, China) under a nitrogen (N2) atmosphere, with the oxygen concentration maintained below 5 ppm. Prior to extraction, all solutions were purged with N2 for 30 min to remove dissolved oxygen. Degassed solutions, samples, and required laboratory materials were placed in the transfer chamber, which was evacuated to remove residual air before the materials were transferred into the glove box. Anaerobic enzymatic extraction of wheat grain For wheat, enzymatic hydrolysis was performed entirely inside the glove box using protease XIV and degassed water at 30 °C for 6 h. Only the centrifugation step was performed outside the glove box using tightly capped tubes sealed with Parafilm. The tubes remained unopened throughout centrifugation and were immediately returned to the glove box, where the supernatant was collected, filtered, and transferred into LC autosampler vials. Anaerobic sequential extraction of C. Hupingshanensis leaves For C. hupingshanensis, samples were mixed with degassed water inside the glove box, after which the extraction tubes were tightly capped and additionally sealed with Parafilm. The free Se fraction was extracted by ultrasound-assisted extraction (37 °C, 45 min) outside the glove box using sealed tubes. The tubes remained unopened during ultrasonic extraction and the subsequent centrifugation step, and were immediately returned to the glove box for collection of the water extract. The remaining residue was subsequently subjected to enzymatic hydrolysis with protease XIV and degassed water at 30 °C for 6 h under anaerobic conditions. After hydrolysis, the extraction tubes were again tightly capped, sealed with Parafilm, centrifuged outside the glove box without opening, and immediately returned to the glove box for supernatant collection and filtration. Following filtration, the final extracts were transferred into LC autosampler vials and sealed with caps and septa within the glove box. The sealed vials were subsequently removed and immediately analyzed by LC-AFS. Determination of selenium species Prior to analysis, sample extracts were filtered through a 0.45 μm membrane filter and diluted with ultrapure water to appropriate Se concentrations. 2 mL aliquot of each extract were transferred into a LC autosampler vial, and 100 μL were injected for analysis. Selenium species were separated using SAX and SCX columns coupled with AFS. For SAX chromatography, a Hamilton PRP-X100 SAX column (250 mm × 4.1 mm × 10 μm; Reno, NV, USA) was used. The column temperature was maintained at 25 °C, and the column was equilibrated with the initial mobile phase for 30 min before injection. Isocratic elution was performed using 40 mM ammonium phosphate buffer (buffer A, pH 6.0). The program was as follows: 0–10 min, 1 mL min−1, 100% A; 10–17 min, 1.2 mL min−1, 100% A. SeO32− and SeO42− were quantified using external five-point calibration curves. For SCX chromatography, a Zorbax 300-SCX column (150 mm × 4.6 mm × 5 μm; Agilent, USA) equipped with a matching guard column was used. The column temperature was maintained at 25 °C, and equilibrated with the initial mobile phase for 30 min before injection. Gradient elution was performed using pyridine formate buffers (pH 2.2; buffer A: 1 mM; buffer B: 40 mM) at a flow rate of 1.2 mL min−1. The program was as follows: 0–2 min, 100% A; 2–15 min, 90% A; 15–16 min, 100% B; 16–20 min, 100% B; 20–21 min, 100% A. γ-GluMeSeCys, MeSeCys, SeMet, SeLan and SeCys2 were quantified using external five-point calibration curves. The outlet of the LC column was directly connected to an AFS detector (SA-50, JiTian, Beijing, China). The column effluent was introduced into a gas-liquid separator, where volatile Se hydrides generated by the hydride-generation system were separated from the liquid phase and transported to the Ar/H2 flame atomizer for detection. Chromatographic data were acquired and processed using JiTian software (version 1.0.3.4). Blank solutions were periodically measured to check any memory effect and cross-contamination during measurements. Instrumental operating parameters are summarized in Table S1. The certified reference material of Se-enriched yeast SELM-1 (National Research Council Canada, Ottawa, Canada) was used for the method validation. The certified values were 2031 ± 70 mg kg−1 for total Se and 3190 ± 260 mg kg−1 for SeMet. Before measurement, SELM-1 was diluted tenfold to obtain Se concentrations comparable to those of the plant samples.Serecovery%=Sumof quantifiedSespeciesTotalSecontent×100% Statistical analysis All experiments were conducted in triplicate (n = 3), and the results are presented as mean ± standard deviation (SD). Statistical analyses were performed using SPSS Statistics 26.0.0.2 and GraphPad Prism 10.3.1. Comparisons between two groups were performed using Student's t-test, whereas differences among multiple groups were evaluated by one-way analysis of variance (ANOVA) followed by Tukey's multiple comparison test. Statistical significance was defined as p < 0.05. Results and discussion Separation of seven selenium species using an orthogonal LC-AFS system As illustrated in Fig. 2A, the standards SeLan, SeCys2, and a laboratory-prepared SeOMet reference solution were co-eluted at the same retention time on the SAX column, indicating a high risk of interference and inaccurate quantification of cationic organic Se species. These results highlight the necessity of incorporating an orthogonal chromatographic system, such as SCX chromatography, when analyzing organic Se species beyond the commonly targeted inorganic analytes. The SeOMet reference solution was used only for qualitative evaluation of chromatographic behavior because its exact concentration was not certified. Therefore, in this study, we utilized a SAX column for the separation of SeO42− and SeO32−, and a SCX column for the separation of five organic Se species (MeSeCys, SeMet, SeCys2, SeLan and γ-GluMeSeCys) (Fig. 2B), thereby ensuring improved specificity and accuracy in Se speciation.Fig. 2Optimization of chromatographic separation conditions for Se species. (A) Chromatograms of cationic organic Se species using a SAX column. (B) Schematic illustration of the orthogonal LC-AFS strategy developed in this study. (C) Chromatograms of four cationic organic Se species using a SCX column under different mobile-phase compositions. T1–70% buffer A; T2–75% buffer A, T3–80% buffer A, T4–85% buffer A, T5–90% buffer A. (D) Chromatogram of inorganic Se species (SeO32− and SeO42−) separated using the SAX column. (E) Chromatogram of organic Se species (γ-GluMeSeCys, MeSeCys, SeMet, SeLan, and SeCys2) separated using the SCX column. Peak assignments: 1-SeO32−; 2-SeO42−, 3-γ-GluMeSeCys; 4-MeSeCys; 5-SeMet, 6-SeLan; 7-SeCys2.Fig. 2 Optimization of chromatographic separation conditions for Se species. (A) Chromatograms of cationic organic Se species using a SAX column. (B) Schematic illustration of the orthogonal LC-AFS strategy developed in this study. (C) Chromatograms of four cationic organic Se species using a SCX column under different mobile-phase compositions. T1–70% buffer A; T2–75% buffer A, T3–80% buffer A, T4–85% buffer A, T5–90% buffer A. (D) Chromatogram of inorganic Se species (SeO32− and SeO42−) separated using the SAX column. (E) Chromatogram of organic Se species (γ-GluMeSeCys, MeSeCys, SeMet, SeLan, and SeCys2) separated using the SCX column. Peak assignments: 1-SeO32−; 2-SeO42−, 3-γ-GluMeSeCys; 4-MeSeCys; 5-SeMet, 6-SeLan; 7-SeCys2. The application of SCX chromatography coupled with AFS detection for Se speciation has been less frequently reported (Both et al., 2020; Yathavakilla et al., 2005). Initially, we tried the method reported by Both et al. (2020) to separate four organic Se species using a SCX column. However, under the reported chromatographic conditions, SeLan and SeCys2 were not effectively separated (Fig. 2C). This limitation may be associated with differences in detection systems, as previous studies predominantly employed ICP-MS, whereas the present study utilized LC-AFS. To overcome the insufficient separation of SeLan and SeCys2, we systematically optimized the chromatographic conditions by adjusting the proportion of the mobile phase. Optimal separation was achieved when the proportion of buffer A was adjusted to 90% (Fig. 2C). Building upon this optimized method, we further included γ-GluMeSeCys as an additional organic Se species. As shown in Fig. 2E, all five Se species were well separated and were rapidly eluted within 13 min. Given the well-established capability of SAX chromatography for separating inorganic Se species, the SAX method described in Section 2.8 was adopted without further modification (Farooq et al., 2024). As illustrated in Fig. 2D, two inorganic Se species were efficiently separated by the SAX chromatographic method, supporting its application in subsequent experiments. The analytical performance of the proposed LC-AFS method was evaluated in terms of selectivity, linearity, and precision. Selectivity was confirmed by peak separation and baseline chromatographic resolution. As shown in Table S2, five-point external calibration curves were established for individual Se species using species-specific concentration ranges: 10–100 ng mL−1 for SeO32− and SeO42−; 20–200 ng mL−1 for MeSeCys, γ-GluMeSeCys, and SeMet; and 40–400 ng mL−1 for SeCys2 and SeLan. Excellent linearity was obtained for all analytes, with coefficients of determination (R2) greater than 0.9950. The LODs, determined as the lowest concentration yielding a signal-to-noise ratio (S/N) of 3, were 0.38 ng mL−1 for SeO32−, 1.50 ng mL−1 for SeO42−, and ranged from 3.16 to 12.42 ng mL−1 for the five Se-AAs. The LOQs, defined by an S/N of 10, were 1.27 ng mL−1 for SeO32−, 5.00 ng mL−1 for SeO42−, and 10.53–41.40 ng mL−1 for the Se-AAs (Table S2). Method precision was assessed using 40 ng mL−1 mixed standard solution, with RSDs of retention time (n = 6) ranging from 0.06% to 0.84%, and RSDs of peak area from 0.87% to 4.14%. Overall, the optimized orthogonal LC-AFS method enabled rapid and reliable separation of both inorganic and organic Se species, providing a robust platform for Se speciation analysis in plant samples. Comparison of assisted enzymatic extraction strategies Efficient extraction is a prerequisite for the accurate quantification of Se species in plant materials (Xiao et al., 2017). Because a substantial proportion of Se may occur in protein-bound forms, extraction procedures must facilitate the release of these species while minimizing species degradation or interconversion during sample preparation (Gao et al., 2018; Yu et al., 2004). To improve extraction efficiency, three assisted enzymatic extraction methods, namely water bath shaking-assisted extraction (WB), ultrasonic water bath-assisted extraction (WU), and probe sonication-assisted extraction (PS), were evaluated (Fig. S1A). The extraction performance of these methods was assessed in both the Se hyperaccumulator C. hupingshanensis and the non-hyperaccumulator wheat using the orthogonal LC-AFS system (Table S3). In C. hupingshanensis, three organic Se species and one inorganic Se species were detected as the major Se forms (Fig. 3A). As shown in Fig. S1B, the concentrations of MeSeCys, SeCys2, and SeO32− among the three extraction methods did not exhibit significant differences, except that the WU method yielded higher SeCys2 concentrations compared to the WB and PS methods by 2.26 ± 1.89 and 2.89 ± 0.37 μg Se g−1, respectively (Fig. S1B). This observation may be attributed to the fact that the enzymatic hydrolysis method primarily enhances the extraction of protein-bound Se, whereas in hyperaccumulator plants, Se mainly exists as free Se-AAs. Therefore, assisted extraction strategies had a limited influence on Se recovery in C. hupingshanensis. The slight increase in SeCys2 content by the WU method suggests that different assisted methods may influence the release of protein-bound Se to varying extents. In contrast, the extraction method had a greater impact on Se recovery in wheat (Fig. 3B). As shown in Fig. S1C, the extraction method had more pronounced effects on the concentrations of SeMet and SeCys2. The WB method produced SeMet concentrations that were 6.61 ± 5.81 and 17.48 ± 3.79 μg Se g−1 higher than those obtained using the WU and PS methods, respectively. Similarly, SeCys2 concentrations obtained by the WB method exceeded those of the WU and PS methods by 5.58 ± 4.02 and 0.68 ± 5.21 μg Se g−1, respectively. In contrast, no significant differences were observed for MeSeCys and SeO42− among the three methods. These results suggest that optimization of assisted enzymatic extraction primarily enhances the release of protein-bound Se species, which is more relevant in non-hyperaccumulator plants. Accordingly, total Se recovery in wheat increased from 34.11 ± 11.83% (PS method) to 52.21 ± 3.98% (WB method), highlighting the importance of extraction optimization for accurate Se speciation analysis.Fig. 3Effect of assisted enzymatic extraction methods and extraction procedures on Se speciation and recovery. Se recoveries obtained using different assisted enzymatic extraction methods in (A) C. hupingshanensis and in (B) wheat. (C) Schematic illustration of the extraction procedure optimization strategy developed in this study. Effect of extraction procedure on Se recovery in (D) C. hupingshanensis and in (E) wheat. (F) Effect of extraction procedure on the concentration of Se species in C. hupingshanensis. All experiments were performed in triplicate (n = 3), and error bars represent standard deviation. Asterisks indicate significant differences between the indicated groups based on Student's t-test (p < 0.05).Fig. 3 Effect of assisted enzymatic extraction methods and extraction procedures on Se speciation and recovery. Se recoveries obtained using different assisted enzymatic extraction methods in (A) C. hupingshanensis and in (B) wheat. (C) Schematic illustration of the extraction procedure optimization strategy developed in this study. Effect of extraction procedure on Se recovery in (D) C. hupingshanensis and in (E) wheat. (F) Effect of extraction procedure on the concentration of Se species in C. hupingshanensis. All experiments were performed in triplicate (n = 3), and error bars represent standard deviation. Asterisks indicate significant differences between the indicated groups based on Student's t-test (p < 0.05). Development of a sequential extraction strategy In Section 3.2, the Se recovery from C. hupingshanensis was relatively low, with the highest value reaching only 23.26 ± 3.63% by the WU-assisted extraction method. Compared with previous studies, the recovery of Se species obtained in this study was considerably lower, and high SeCys2 concentrations in C. hupingshanensis were not observed as previously reported (Ma et al., 2023; Yuan et al., 2013). Previous studies have shown that approximately 60% of the total Se in C. hupingshanensis is water-extractable, of which nearly 40% occurs as SeLan (Both et al., 2018). However, our current results did not detect significant levels of SeLan either. To further investigate, we determined the Se content in the water extract of C. hupingshanensis and found that the water-soluble Se fraction accounted for as much as 72.06 ± 0.32% of the total Se content. This result suggests that a substantial proportion of the water-soluble Se, most likely present as free Se-AAs, might be lost during the enzymatic extraction. This loss may be associated with the complex environment during enzymatic hydrolysis, including changes in pH, ionic strength, and enzyme-substrate interactions, which could promote the degradation or transformation of labile free Se-AAs. To address this issue, we proposed a sequential extraction approach for both types of plant samples (Fig. 3C). Unlike the conventional one-step method, where the protease was directly added to the plant material for enzymatic extraction, we designed a two-step procedure. In this method, an initial water extraction using ultrapure water was performed to selectively extract free Se-AAs, followed by enzymatic hydrolysis of the residual pellet (representing protein-bound Se). Subsequently, both the water extract and the enzymatic hydrolysate were analyzed for Se speciation using LC-AFS. This stepwise extraction strategy aimed to minimize the loss of water-soluble Se fraction and improve the overall Se recovery and accuracy of Se speciation in plants. As shown in Fig. 3F, the inclusion of a water extraction step significantly increased the recovery of free Se-AAs, particularly MeSeCys and SeLan, with the most notable improvement observed for SeLan. Specifically, the two-step extraction method increased the concentration of SeLan from 5.06 ± 0.28 μg Se g−1 (one-step method) to 143.74 ± 2.70 μg Se g−1, with SeLan detected almost exclusively in the aqueous extract (Table S4). In contrast, the one-step method resulted in substantial losses of SeLan, and MeSeCys was also affected to a lesser extent. These findings suggested that the complex enzymatic hydrolysis environment following protease addition may interfere with the determination of free Se-AAs, particularly for SeLan, although the underlying mechanisms require further investigation. For Se hyperaccumulator plants, where Se predominantly exists as free Se-AAs, the adoption of a sequential extraction procedure is therefore essential. As illustrated in Fig. 3D, the total recovery of Se species in C. hupingshanensis increased markedly from 30.20 ± 2.55% (one-step method) to 70.97 ± 0.22% (two-step method) using the optimized method. In contrast, the optimization of the extraction procedure had minimal impact on Se speciation results in wheat (Fig. 3E). A slight decrease in MeSeCys concentration was observed, which can be attributed to physical losses during the collection of the aqueous phase (Fig. S1D). Overall, these results demonstrate that extraction procedures should be tailored to the Se accumulation characteristics within different plant matrices. For Se hyperaccumulator plants, incorporation of a water-extraction step markedly improves the recovery of free Se-AAs. In contrast, the one-step enzymatic method remains convenient and effective with negligible loss of Se species for non-hyperaccumulator plants. Optimization of enzymatic extraction conditions The efficiency of enzymatic extraction is strongly influenced by extraction conditions, particularly pH and temperature. In this study, enzymatic temperature and pH were selected as key optimization factors to maximize the recovery of Se species. Since pH might directly affect enzyme activity and, consequently, extraction efficiency, a range of extraction solutions was evaluated, including Tris-HCl buffers with pH values from 4.5 to 8.5 as well as ultrapure water. Based on the previous results, the two-step extraction method was applied to C. hupingshanensis, while the one-step method was used for wheat. For C. hupingshanensis, the effect of extraction solution pH on the Se recovery in the water extraction was also investigated. As shown in Fig. 4A and E, the pH of the water extraction had a pronounced effect on the concentration of SeLan. Notably, the use of Tris-HCl buffer as the extraction solution resulted in significant interference with SeLan, with concentrations decreasing from 157.30 ± 8.14 μg Se g−1 in water to 42.72 ± 3.74 μg Se g−1 in Tris-HCl buffer at pH 4.5 (Table S5). These results suggest that the recovery of SeLan was substantially reduced in Tris-HCl buffer compared with water. This effect may be associated with interactions between Tris-HCl buffer components and free Se-AAs, although the underlying mechanism remains unclear. In contrast, the use of Tris-HCl buffer at pH 4.5 led to increased concentrations of MeSeCys and γ-GluMeSeCys from the enzymatic hydrolysate, indicating that acidic conditions may improve the stability of these species during extraction (Fig. S2). For wheat, water was identified as the optimal extraction matrix (Fig. 4B). As shown in Fig. 4F, MeSeCys was only detected when water was used as the extraction matrix, and could not be detected under any Tris-HCl buffer conditions. Similarly, the concentration of SeMet decreased markedly as the extraction matrix changed from water to Tris-HCl buffer, with further decreases observed as pH increased from 4.5 (32.66 ± 3.23 μg Se g−1) to 8.5 (4.09 ± 0.47 μg Se g−1). These results showed that MeSeCys and SeMet in wheat are unstable in Tris-HCl buffer, further indicating that water was the most suitable extraction medium for the plant matrices in this study. Overall, these findings highlight the importance of solution selection for accurate Se speciation in different plants.Fig. 4Effect of extraction pH and extraction temperature on Se recovery and speciation. Comparison of Se recoveries at different extraction pH conditions in (A) C. hupingshanensis and in (B) wheat. Comparison of Se recoveries at different extraction temperatures in (C) C. hupingshanensis and in (D) wheat. Effect of extraction pH conditions on the concentration of Se species in (E) the water extraction of C. hupingshanensis and in (F) wheat. Effect of extraction temperature on the concentration of Se species in (G) the enzymatic hydrolysate of C. hupingshanensis and in (H) wheat. All experiments were performed in triplicate (n = 3), and error bars represent standard deviation. Different lowercase letters indicate significant differences among groups based on one-way ANOVA followed by Tukey's multiple comparison test (p < 0.05).Fig. 4 Effect of extraction pH and extraction temperature on Se recovery and speciation. Comparison of Se recoveries at different extraction pH conditions in (A) C. hupingshanensis and in (B) wheat. Comparison of Se recoveries at different extraction temperatures in (C) C. hupingshanensis and in (D) wheat. Effect of extraction pH conditions on the concentration of Se species in (E) the water extraction of C. hupingshanensis and in (F) wheat. Effect of extraction temperature on the concentration of Se species in (G) the enzymatic hydrolysate of C. hupingshanensis and in (H) wheat. All experiments were performed in triplicate (n = 3), and error bars represent standard deviation. Different lowercase letters indicate significant differences among groups based on one-way ANOVA followed by Tukey's multiple comparison test (p < 0.05). Enzymatic temperature is also an important factor in the process of enzymatic hydrolysis. In this study, extraction temperatures ranging from 25 °C to 50 °C were evaluated using water as the extraction matrix. As shown in Fig. 4C, the extraction temperature had minimal impact on Se speciation in C. hupingshanensis, likely due to the relatively low proportion of protein-bound Se in this hyperaccumulator, resulting in negligible effects on the overall recovery. As shown in Fig. 4G, only a slight effect of temperature on MeSeCys extraction was observed, with the highest extracted concentration (9.30 ± 0.54 μg Se g−1) at 30 °C. In contrast, lower extraction temperatures had a pronounced effect on the recovery of SeCys2 and SeMet in wheat samples (Fig. 4D). Specifically, the extraction concentration of SeMet increased from 90.19 ± 2.40 μg Se g−1 at 25 °C to 100.38 ± 16.44 μg Se g−1 at 30 °C, but declined sharply with further temperature increases, reaching only 11.98 ± 1.19 μg Se g−1 at 50 °C (Fig. 4H). Similarly, the concentration of SeCys2 rose from 25.90 ± 1.35 μg Se g−1 at 25 °C to 28.47 ± 3.42 μg Se g−1 at 30 °C, then decreased from 25.02 ± 5.50 μg Se g−1 at 37 °C to 6.18 ± 0.30 μg Se g−1 at 50 °C (Table S6). Although enzymatic hydrolysis at 37 °C has frequently been reported for Se extraction from plant materials, the highest recovery (86.05 ± 11.71%) in the present study was obtained at 30 °C. These results suggest that elevated extraction temperatures may adversely affect the recovery of protein-bound Se species, particularly SeMet and SeCys2. Excessively high temperatures can compromise the stability of Se compounds. Therefore, an extraction temperature of 30 °C was selected as the optimal condition for Se speciation in both C. hupingshanensis and wheat in subsequent experiments. Optimization of enzyme purification and anaerobic extraction conditions The enzymatic hydrolysates from plant samples are complex mixtures containing proteases and high-molecular-weight impurities that may interfere with the accurate quantification of low-molecular-weight amino acids (AAs) (Cunningham et al., 2022). Previous studies have suggested that for samples with complex matrices or analytical methods susceptible to protein interference, purification after extraction is necessary to ensure the precise determination of free AAs (Desmons et al., 2020). However, limited information is available regarding the influence of residual proteins on LC-AFS determination of free Se-AAs in different plant matrices. In this study, a filtration step was therefore introduced to remove Protease XIV and other high-molecular-weight components prior to LC-AFS analysis. As shown in Fig. 5A, the purification had little effect on Se quantification in C. hupingshanensis, with no significant differences observed (Table S7). This may be attributed to the predominance of water-soluble Se species and the relatively simple matrix composition of C. hupingshanensis leaves, resulting in minimal matrix interference (Fig. S3A and B). In contrast, purification of the enzymatic hydrolysate led to an increase in Se recovery from 81.05 ± 3.44% (unpurified) to 87.78 ± 7.66% (purified) in wheat grains (Fig. 5B). Further analysis (Fig. S3C) revealed that the concentrations of SeMet and SeCys2 were notably enhanced after purification, with SeMet increasing from 100.87 ± 6.11 μg Se g−1 (unpurified) to 108.05 ± 7.70 μg Se g−1 (purified), and SeCys2 from 23.88 ± 4.03 μg Se g−1 to 29.07 ± 0.24 μg Se g−1, respectively. These results indicate that the enzymatic hydrolysate of wheat grains is more complex and contains a higher proportion of macromolecular substances, which can interfere with Se speciation analysis. These results suggest that purification of enzymatic hydrolysates can improve Se recovery in wheat and other plant matrices with relatively complex compositions. Overall, the complexity of plant sample extractions and the sensitivity of detection methods should be carefully considered, and the necessity of enzymatic hydrolysate purification should be validated for specific plant species and analytical techniques in future studies.Fig. 5Effect of enzyme purification and oxygen conditions on Se recovery and speciation. Effect of enzyme purification on the Se recovery in (A) C. hupingshanensis and in (B) wheat. Effect of oxygen conditions on the Se recovery in (C) C. hupingshanensis and in (D) wheat. Effect of oxygen conditions on the concentration of Se species in (E) the water extraction and in (F) the enzymatic extraction of C. hupingshanensis. (G) Effect of oxygen conditions on the color of the enzymatic extraction in C. hupingshanensis and wheat. (H) Effect of oxygen conditions on the concentration of Se species in wheat. All experiments were performed in triplicate (n = 3), and error bars represent standard deviation. Asterisks indicate significant differences between the indicated groups based on Student's t-test (p < 0.05); no asterisk indicates no significant difference.Fig. 5 Effect of enzyme purification and oxygen conditions on Se recovery and speciation. Effect of enzyme purification on the Se recovery in (A) C. hupingshanensis and in (B) wheat. Effect of oxygen conditions on the Se recovery in (C) C. hupingshanensis and in (D) wheat. Effect of oxygen conditions on the concentration of Se species in (E) the water extraction and in (F) the enzymatic extraction of C. hupingshanensis. (G) Effect of oxygen conditions on the color of the enzymatic extraction in C. hupingshanensis and wheat. (H) Effect of oxygen conditions on the concentration of Se species in wheat. All experiments were performed in triplicate (n = 3), and error bars represent standard deviation. Asterisks indicate significant differences between the indicated groups based on Student's t-test (p < 0.05); no asterisk indicates no significant difference. Certain Se species, particularly reduced organic selenocompounds, are susceptible to oxidation and transformation during extraction in the presence of oxygen (Liu & Bei, 2010). Previous studies have shown that a portion of SeMet in yeast (approximately 5.5%) can be eluted as its oxidized form, SeOMet, a proportion that should not be overlooked (Polatajko et al., 2005). The formation of oxidized species during extraction may result in the loss of target organic Se compounds, highlighting the need to compare the extraction efficiency under aerobic and anaerobic conditions. Our results demonstrated that the recoveries of Se species in both C. hupingshanensis and wheat were improved under anaerobic conditions (Table S8). As shown in Fig. 5C, the total recovery of Se species in C. hupingshanensis increased from 71.57 ± 4.75% under aerobic conditions to 80.65 ± 8.51% under anaerobic conditions. Notably, the concentrations of free Se-AAs such as MeSeCys and SeLan in the water extract were increased in the absence of oxygen, while the extraction of inorganic SeO32− remained unaffected (Fig. 5E). Additionally, the concentration of γ-GluMeSeCys in the enzymatic hydrolysate increased under anaerobic conditions, whereas MeSeCys decreased, likely due to its increased distribution into the water extract (Fig. 5F). For wheat, the Se recovery in the enzymatic hydrolysate was markedly higher under anaerobic conditions, rising from 80.79 ± 2.80% to 90.02 ± 5.93% (Fig. 5D). Furthermore, a noticeable color change was observed in wheat hydrolysates, with aerobically processed extracts becoming darker than those prepared under anaerobic conditions (Fig. 5G). The SeMet content in wheat also increased substantially, from 98.73 ± 4.50 μg Se g−1 in the aerobic group to 112.17 ± 7.87 μg Se g−1 in the anaerobic conditions group, representing a 13.61% increase (Fig. 5H). Chromatographic analysis using SAX-AFS revealed that the peak area at 2.3 min, corresponding to SeOMet and SeCys2, was significantly reduced in the anaerobic group (2397.50 ± 649.83) compared to the aerobic group (8085.5 ± 443.75), indicating that oxidation of SeMet to SeOMet was substantially suppressed under anaerobic conditions. We note that ultrasound-assisted extraction of C. hupingshanensis and centrifugation were performed outside the glove box using tightly capped tubes additionally sealed with Parafilm. Although a small degree of oxygen exposure during these sealed procedures cannot be completely excluded, the observed improvement in Se recoveries and suppression of SeOMet formation suggest that oxidative transformation was substantially reduced under the adopted anaerobic extraction protocol. Nevertheless, implementing strict anaerobic conditions remains operationally challenging. Therefore, future studies may consider the use of reducing agents such as β-mercaptoethanol and dithiothreitol as practical alternatives to prevent oxidation during extraction, provided that their effects on the enzymatic process are thoroughly evaluated beforehand (Sentkowska & Pyrzynska, 2021). Comparison of the developed method with previous studies The performance of the optimized extraction strategy coupled with LC-AFS in our study was compared with previously reported approaches for Se speciation analysis in C. hupingshanensis and wheat (Table 1), with detailed extraction procedures provided in Table S9. Representative SAX-LC-AFS and SCX-LC-AFS chromatograms of plant extracts obtained using the optimized extraction procedures are provided in Fig. S4 and S5, respectively, supporting the chromatographic identification of the major Se species detected in this study. In C. hupingshanensis, significant differences in Se species profiles have been reported depending on the chromatographic separation method employed. Studies based on SAX chromatography mainly identified MeSeCys, SeCys2, SeO32−, and SeO42−, with SeCys2 reported as the predominant organic Se species. In contrast, SCX-based methods primarily detected MeSeCys, SeLan, and SeO32−, while little or no SeCys₂ was observed, and SeLan accounted for approximately 40% of the total Se content. Consistent with previous SCX-based studies, the present method revealed SeLan as a major Se species and detected only trace levels of SeCys2. Additionally, considerable amounts of MeSeCys were identified, particularly in the aqueous extracts, together with γ-GluMeSeCys, which has not previously been reported in C. hupingshanensis. The detection of γ-GluMeSeCys provides new insight into the Se metabolic network in this hyperaccumulator. As a γ-glutamylated derivative associated with the MeSeCys metabolic branch (Fig. 1), γ-GluMeSeCys may be related to Se storage, transport, or detoxification under high-Se conditions. Similar observations have been reported for the Se hyperaccumulator Stanleya pinnata, in which leaf trichomes contain predominantly MeSeCys (53%) and γ-GluMeSeCys (47%) (Freeman et al., 2006). Furthermore, MeSeCys can be converted into volatile dimethyl diselenide (DMDSe), a key component of Se volatilization pathways in S. pinnata and Astragalus species (Freeman & Bañuelos, 2011; Schiavon & Pilon-Smits, 2017). Although Se volatilization mechanisms in C. hupingshanensis remain poorly understood, the occurrence of γ-GluMeSeCys suggests that MeSeCys-derived storage and volatilization pathways deserve further investigation. In addition, several unidentified Se-containing peaks were observed in the SCX chromatograms (Fig. S5B), warranting further investigation and identification. These results highlight the unique advantages of the developed detection and extraction methods for the accurate analysis of Se species in C. hupingshanensis.Table 1Comparison of the developed method versus published methods in C. hupingshanensis and wheat (% of the total Se). -, no data; ND, not detected.PlantExtraction procedureSeparation methodDetection methodMeSeCys (%)γ-GluMeSeCys (%)SeLan (%)SeCys2 (%)SeMet (%)SeO32− (%)SeO42− (%)Se recovery (%)ReferenceCardaminehupingshanensisProtease XIV-assisted enzymatic hydrolysisSAXAFS13––78ND5–96(Yuan et al., 2013)Protease XIV-assisted enzymatic hydrolysisSAXAFS0.22––22–360.060.0613–2143–54(Rao et al., 2021)Protease E-assisted enzymatic hydrolysisSAXICP-MS2–24––54–831–151–23–1189–100(Wu et al., 2020)Protease E-assisted enzymatic hydrolysisRPICP-MS18–23––30–490.4–0.70.6–10.3–0.453–75(Ma et al., 2023)Water extraction assistedby ultrasonic probeSAX/SCX/RPICP-MS0.4–40NDTrace1ND88(Both et al., 2018)Sequential water extraction and protease XIV-assisted enzymatic extractionSAX/SCXAFS23347NDND7ND81This studyWheatProtease E-assisted enzymatic hydrolysisSAXAFSND––8–1570–871–32–493–97(Di et al., 2023)Protease XIV-assisted enzymatic hydrolysisSAXICP-MSND––4–1387–96Trace1–693–100(Wang, Zhou, et al., 2022)Protease XIV-assisted enzymatic hydrolysisSAXICP-MS2–3––15–2564–73ND4–886–96(Subirana et al., 2023)α-Amylase/protease-assisted extraction with ultrasonic treatmentSAX/SCX/RPICP-MS0–0.5–––72–85–2–670–89(Cubadda et al., 2010)Sequential water extraction and protease XIV-assisted enzymatic extractionHILICLC-MSFree Se-AAs (MeSeCys/SeMet): 7Conjugated Se-AAs (SeMet/SeCys/SeCys2): 60–7067–77(Montone et al., 2019)Protease XIV-assisted enzymatic extractionSAX/SCXAFS4NDND1666ND390This study Table 1Comparison of the developed method versus published methods in C. hupingshanensis and wheat (% of the total Se). -, no data; ND, not detected.PlantExtraction procedureSeparation methodDetection methodMeSeCys (%)γ-GluMeSeCys (%)SeLan (%)SeCys2 (%)SeMet (%)SeO32− (%)SeO42− (%)Se recovery (%)ReferenceCardaminehupingshanensisProtease XIV-assisted enzymatic hydrolysisSAXAFS13––78ND5–96(Yuan et al., 2013)Protease XIV-assisted enzymatic hydrolysisSAXAFS0.22––22–360.060.0613–2143–54(Rao et al., 2021)Protease E-assisted enzymatic hydrolysisSAXICP-MS2–24––54–831–151–23–1189–100(Wu et al., 2020)Protease E-assisted enzymatic hydrolysisRPICP-MS18–23––30–490.4–0.70.6–10.3–0.453–75(Ma et al., 2023)Water extraction assistedby ultrasonic probeSAX/SCX/RPICP-MS0.4–40NDTrace1ND88(Both et al., 2018)Sequential water extraction and protease XIV-assisted enzymatic extractionSAX/SCXAFS23347NDND7ND81This studyWheatProtease E-assisted enzymatic hydrolysisSAXAFSND––8–1570–871–32–493–97(Di et al., 2023)Protease XIV-assisted enzymatic hydrolysisSAXICP-MSND––4–1387–96Trace1–693–100(Wang, Zhou, et al., 2022)Protease XIV-assisted enzymatic hydrolysisSAXICP-MS2–3––15–2564–73ND4–886–96(Subirana et al., 2023)α-Amylase/protease-assisted extraction with ultrasonic treatmentSAX/SCX/RPICP-MS0–0.5–––72–85–2–670–89(Cubadda et al., 2010)Sequential water extraction and protease XIV-assisted enzymatic extractionHILICLC-MSFree Se-AAs (MeSeCys/SeMet): 7Conjugated Se-AAs (SeMet/SeCys/SeCys2): 60–7067–77(Montone et al., 2019)Protease XIV-assisted enzymatic extractionSAX/SCXAFS4NDND1666ND390This study Comparison of the developed method versus published methods in C. hupingshanensis and wheat (% of the total Se). -, no data; ND, not detected. For wheat, previous SAX-based studies mainly reported MeSeCys, SeCys2, SeO32−, and SeO42−, with SeMet as the dominant Se species (Di et al., 2023; Montone et al., 2019; Wang et al., 2022). Compared with previous studies, no additional Se species were identified in wheat using the developed Se speciation strategy. Nevertheless, it provided improved discrimination between SeCys2 and co-eluting compounds. As shown in Fig. 2A, SeOMet and SeCys2 exhibited overlapping retention behavior on SAX chromatography, suggesting that SeCys2 concentrations reported using SAX alone may have been overestimated. Small amounts of MeSeCys were also detected in wheat. Overall, these findings demonstrated the developed method provides improved specificity and accuracy for Se speciation analysis in both C. hupingshanensis and wheat. Conclusion In this study, a cost-effective orthogonal LC-AFS method was established for the separation and quantification of seven Se species, including SeO32−, SeO42−, SeLan, SeMet, SeCys2, MeSeCys, and γ-GluMeSeCys. The results further demonstrated that Se extraction should be tailored to the Se accumulation characteristics within different plant matrices. For the Se hyperaccumulator C. hupingshanensis, the optimized two-step extraction strategy achieved the highest Se recovery (80.65%), comprising an initial ultrasonic water-bath extraction, followed by shaking-assisted enzymatic hydrolysis at 30 °C under anaerobic conditions. For wheat, the optimized one-step extraction strategy resulted in a maximum Se recovery of 90.02%, using shaing-assisted enzymatic hydrolysis at 30 °C under anaerobic conditions. Overall, this study highlights that differentiated extraction is not merely a methodological option, but a fundamental prerequisite for reliable nutritional and therapeutic assessments of plant-based products. By ensuring an accurate and clear Se species profile, this work will promote a more accurate assessment of Se species contained within different plant-based Se products to meet the increasing market demands for safer, non-toxic, and more bioavailable organic Se for human utilization and consumption. CRediT authorship contribution statement Yuhua Duan: Writing – original draft, Visualization, Validation, Methodology, Investigation, Data curation. Guangxuan Du: Validation, Methodology. Tao Li: Writing – review & editing, Investigation. Hua Zhang: Writing – review & editing. Zezhou Zhang: Writing – review & editing. Linxi Yuan: Writing – review & editing, Supervision, Project administration, Funding acquisition, Conceptualization. Informed consent Not applicable. Ethical approval This article does not contain any studies with human participants or animals performed by any of the authors. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.