Extraction and purification and microcapsulation of sea buckthorn seed proanthocyanidins using deep eutectic solvent-assisted extraction

Abstract

This study aimed to develop a green deep eutectic solvents (DES) extraction strategy to recover proanthocyanidins (PAs) from sea buckthorn (Hippophae rhamnoides) seeds, a followed by composite biopolymer microencapsulation to overcome the poor stability of PAs. Five choline chloride-based DES systems were screened, and the choline chloride/1,4-butanediol binary DES containing 50% (v/v) water exhibited the highest extraction efficiency. Single-factor experiments combined with Box–Behnken response surface methodology optimized the extraction parameters: solid-to-liquid ratio of 41:1 mg/mL, extraction temperature of 61 °C, and extraction duration of 3.3 h, delivering a maximum PA yield of 48.025 mg/g seed powder. After purification via D101 macroporous resin, the purity of PA-rich powder reached 69.17%, and LC-MS/MS characterization confirmed the presence of epigallocatechin monomers and dimeric proanthocyanidin B1–B4 in the extract. In vitro antioxidant assays quantified the radical-scavenging capacity of extracted PAs as 4.61 μmol TE/mg (DPPH), 5.53 μmol TE/mg (ABTS) and 6.24 μmol TE/mg (FRAP). To improve thermal and digestive stability, sodium alginate-carboxymethyl cellulose composite microcapsules were fabricated by ionic gelation, achieving an encapsulation efficiency of 89.015%. TGA analysis verified that the composite wall significantly delayed thermal degradation of PAs; in simulated gastric fluid, free PAs reached 53.64% cumulative release after 180 min, while encapsulated PAs only released 21.99%, and the microcapsules exhibited sustained intestinal release up to 45.65% at 180 min. This integrated DES extraction and composite microencapsulation route realizes high-value utilization of sea buckthorn seed waste and provides stable, gastrointestinal-targeted PA ingredients for functional food applications.

Graphical abstract

Highlights

• •Choline chloride-1,4-butanediol DES achieves the highest proanthocyanidin (PA) extraction rate from sea buckthorn seeds.
• •Sea buckthorn seed PAs have potential as natural antioxidants and hypoglycaemic agents.
• •Cross-linking of core-wall materials in PAs microcapsules improves PA transport, preservation and intestinal‑fluid release performance.

1. Introduction

Hippophae rhamnoides L., a deciduous shrub indigenous to the Xinjiang region of China, is widely distributed across Asia, Europe and Canada (Ciesarová et al., 2020). Its fruits are abundant bioactive metabolites of bioactive constituents, including carbohydrates, polysaccharides, polyphenols, vitamins, and amino acids (Yu et al., 2024). These compounds possess multiple biological functions, such as antioxidant (Teng, He, Hong, Xie, & Zha, 2024), anticarcinogenic (J. Zhao et al., 2024), anti-inflammatory (Ciesarová et al., 2020), antimicrobial (Netreba et al., 2024), and hepatoprotective (Dubey, Shukla, Shukla, & Singh, 2024) effects, thereby positioning sea buckthorn as a subject of considerable scientific interest.

Among these constituents, proanthocyanidins (PAs)—oligomeric and polymeric flavonoid polyphenols—are particularly owing to their prominent health-promoting bioactivities for their potential health benefits. PAs naturally accumulate in sea buckthorn fruit peel, seeds, and stems (Liu et al., 2023). Nevertheless, existing investigations overwhelmingly concentrate on PAs isolated from sea buckthorn pulp, while seed, leaf and peel fractions are largely overlooked. These PA-rich by-products are routinely discarded during industrial fruit processing, leading to severe resource waste and low comprehensive utilization efficiency of sea buckthorn processing residues(Zeng et al., 2023). This formed the first major limitation of current research: researchers ignore the valorization potential of seed-derived PAs, and lack targeted green extraction strategies for this waste stream.

PAs are well-documented for their potent antioxidant properties (Hosoda et al., 2018), alongside cardiovascular protective effects(Rao, 2018), anti-obesity potential via pancreatic lipase inhibition (McDougall, Kulkarni, & Stewart, 2009), and enhancement of microvascular integrity (Duthie, Duthie, & Kyle, 2000; Shi, Yu, Pohorly, & Kakuda, 2003). However, the extraction of PAs from plant matrices is challenging due to the recalcitrance of plantcell walls and the lability of polyphenols. While conventional methods such as high-pressure extraction (Huang, Wu, Lu, Shyu, & Wang, 2017), supercritical CO₂ extraction(del Valle, 2015), and ultrasonic-assisted extraction(Tiwari, 2015), have been adopted for polyphenol recovery, yet they suffer from inherent drawbacks: high equipment investment, large organic solvent consumption, high energy input and secondary environmental pollution risks.

Deep eutectic solvents (DESs) have recently emerged as eco-friendly, low-cost and biodegradable alternatives for phytochemical extraction (Hansen et al., 2020; Oliveira, Pereiro, Rebelo, & Marrucho, 2013). Typically formed through hydrogen bond between a hydrogen bond donor (HBD) and a hydrogen bond acceptor (HBA), DES are often composed of biodegradable, low-cost, and environmentally friendly components (Smith, Abbott, & Ryder, 2014) (Radošević et al., 2015). Despite wide applications of DESs in extracting anthocyanins and phenolic acids, previous studies rarely develop customized DESs systems to selectively extract PAs from sea buckthorn seeds. This constitutes the second critical research limitation: no optimized DES extraction protocol is available for sea buckthorn seed PAs, and the extraction performance of tailored DESs for this specific PA source remains uncharacterized.

Despite their potent bioactivity, the practical application of PAs is constrained by limited water aqueous solubility, poor stability under environmental, processing and gastrointestinal conditions, and rapid metabolism (El Asbahani et al., 2015) (Estevinho, Ramos, & Rocha, 2015; Ge, Dong, Zhu, Zhang, & Li, 2015). Microencapsulation, a technology involving the entrapment of an active core within a protective wall matrix, has been widely adopted to address these challenges for various polyphenols (Bah, Bilal, & Wang, 2020; (Keller & Sottos, 2006). Encapsulation can shield the core material from oxidation, pH fluctuations, and thermal degradation, while also enabling controlled release in the gastrointestinal tract.

However, microencapsulation systems relying on a single wall material may exhibit compromised mechanical strength or suboptimal barrier properties (Ramdhan, Ching, Prakash, & Bhandari, 2020; Song et al., 2022). This study, therefore, employs a composite wall material system comprising sodium alginate (a plant-derived polysaccharide of α-L-guluronic and β-D-mannuronic acids) and carboxymethyl cellulose (a water-soluble cellulose derivative rich in carboxyl and hydroxyl groups). This specific combination, leveraging synergistic effects such as enhanced mechanical strength and complexation ability, has not been previously applied for the encapsulation of sea buckthorn seed PAs. At present, an integrated green manufacturing route combining DES extraction and composite biopolymer microencapsulation for sea buckthorn seed PAs has not been reported.

Collectively, three interconnected research gaps and unresolved scientific problems can be summarized from the above limitations: (i) The waste sea buckthorn seed, a rich PA source, lacks efficient green extraction technology, and most studies only focus on pulp PAs without exploiting processing by-products;(ii) There is no customized DES extraction system optimized for sea buckthorn seed PAs, and the extraction efficiency, recovery rate and protective effect of DES on seed PAs remain unclear;

(iii) Single-wall microcapsules fail to stabilize PAs effectively, and alginate-CMC composite microencapsulation has not been combined with DES extraction to improve the storage stability and gastrointestinal controlled release of seed PAs.

Based on the above unresolved scientific issues, we put forward the core research hypothesis of this study: Customized biodegradable DESs can efficiently disrupt sea buckthorn seed cell walls and protect labile PAs from degradation to achieve high PA recovery; furthermore, alginate-CMC composite microcapsules can significantly improve the water solubility, thermal and digestive stability of DES-extracted seed PAs, and realize sustained release in simulated gastrointestinal fluids.

Accordingly, the dual objectives of this work were proposed to fill the above research gaps and verify the proposed hypothesis: (i) to screen and optimize a tailored DES extraction system to maximize the recovery yield of PAs from discarded sea buckthorn seeds; (ii) to fabricate alginate-carboxymethyl cellulose composite microcapsules loaded with DES-extracted seed PAs, and systematically characterize their physicochemical properties, storage stability and in vitro gastrointestinal release behavior. This integrated strategy simultaneously addresses the resource waste of sea buckthorn industrial residues and the stability defects of PAs, providing a sustainable technical scheme for the high-value utilization of sea buckthorn by-products and advancing the application of PAs in functional foods and nutraceuticals.

2. Materials and methods

2.1. Materials

Sea buckthorn seeds were obtain from Xiaxu-Dongri Sea Buckthorn Foods Co. Ltd.. Petroleum ether and sodium carboxymethyl cellulose were obtained from Tianjin Kemi Chemical Reagent Co. Ltd.. Catechins, vanillin, choline chloride, 2,2′-diazobis-3-ethylbenzothiazoline-6-sulfonic acid (ABTS⁺), TPTZ, Ferric chloride and l-ascorbic acid were purchased from Shanghai McLean Biochemical Technology Co. Ltd.. acetate, Formic acid, acetonitrile (chromatographic grade), 1,1-diphenyl-2-trinitrophenylhydrazine (DPPH•), pepsin, trypsin, 2,4,6-tripyridin-2-yl-1,3,5-triazine and porcine bile salt were supplied from Shanghai Yuanye Biotechnology Co., Ltd.. Hydrogen peroxide, salicylic acid, malonic acid and potassium ferricyanide were purchased from Chengdu Kelon Chemical Co. Ltd. Unless otherwise stated, all chemicals and reagents used in this study were of analytical grade.

2.2. Processing of Sea buckthorn Seeds

Remove impurities from the sea buckthorn seeds, grind them in a grinder, sieve through a 50-mesh screen, degrease them with diethyl ether, set aside the degreased sea buckthorn seed powder, dry it in an electric constant-temperature oven for 5 h, and set it aside for later use.

2.3. Determination of the optimal hydrogen bond donor

The quantification of PAs was performed using the vanillin–HCl colorimetric assay, as described by Sun, Ricardo-da-Silva, and Spranger (Sun, Ricardo-da-Silva, & Spranger, 1998). A calibration curve was constructed using catechins as the reference standard (Y = 2.5838× + 0.0626, R² = 0.9991), where Y is absorbance at 500 nm and x is the concentration of catechins (mg/mL). To identify the optimal hydrogen bond donor (HBD) for DES extraction, choline chloride (ChCl) served as the hydrogen bond acceptor (HBA), Five different HBDs-1,4-butanediol (1,4-BD), citric acid (CA), lactic acid (LA), urea (UR), and Malonic acid (MA)- were synthesized by mixing ChCl with each HBD at a specific molar ratio(da Silva et al., 2020; Dai, Van Spronsen, Witkamp, Verpoorte, & Choi, 2013). The mixtures were stirred in a water bath maintained at 37° and dissolved until no crystals formed and no solid residue remained., and water was incorporated to achieve final concentration of 30%–70% (v/v) to modulate solvent polarity and viscosity. Extraction was conducted at 30 °C using the water bath, and the PAs yield from sea buckthorn seed powder was assessed for each DES system.

2.4. Optimization of extraction conditions

Following the selection of the most effective DES, single-factor experiments were conducted to investigate the influence of key extraction parameters. Based on the papers by Ran (Ran et al., 2025), Yang (Yang et al., 2023), and Kilic (Kilic-Pekgözlü, Kurt, Ceylan, & Çiklaçifci, 2025), three parameters were set: solid-to-liquid ratio (10,1–90:1, mg/mL), extraction time (1–5 h), and extraction temperature (30–70 °C). Based on these preliminary result, a three-factor, three-level Box–Behnken design (BBD) was employed using the Design-expert 13 software (Stat-Ease Inc. Minneapolis, MN, USA) to optimize extraction conditions via response surface methodology (RSM). The independent variables and their coded levels are presented in Table 2. Experimental data were fitted to a second-order polynomial model, and analysis of variance (ANOVA) was performed to evaluate model significance and adequacy.

2.5. Compositional analysis of the extract

2.5.1. Isolation and purification of PAs

PAs extracted under optimal conditions were purified D101 macroporous adsorption resin. The resin was pre-conditioned by sequential soaking in anhydrous ethanol, 5% HCl and 2% NaOH, followed by rinsing with deionized water until neutral. The solution is neutral. The crude extract was dissolved in 40% ethanol, filtered, and loaded onto the column (pH 3–5) at a concentration of 3.0 mg/mL and a flow rate of 1.25 BV/h. Gradient elution was performed using 60% ethanol until PAs were no longer detectable in the eluent. The purified fraction was concentrated and freeze-dried to obtain a PA-rich powder. Quantitative analysis revealed a PA content of 69.17% in the purified extract, calculated using the formula:

$$C_x=\frac{A_x}{A_s}\times C_s\times F$$

where: C_x = Concentration of the target analyte in the sample under test. A_x = Response value of the target analyte in the sample under test (typically peak area or peak height in LC-MS). A_s = Response value of the target analyte in the standard sample. C_s = Concentration of the standard sample. F = Potential dilution factor, recovery rate correction factor.

2.5.2. Determination of PAs using liquid chromatography-mass spectrometry

Qualitative characterization of PAs in the purified extract was performed using liquid chromatography– mass spectrometry (LC-MS/MS) following the method of Shoji et al. (2003) (Shoji et al., 2003). Chromatographic separation was conducted on a Waters ACQUITY BEH C18 (100 × 2.1 mm 1.7 μm). The column was maintained at a constant temperature of 40 °C. The mobile phases comprised 0.1% formic acid in water (A) and acetonitrile (B). The gradient elution program was detailed in Table 3. Mass spectrometric detection was employed an electrospray ionization (ESI) source operating in positive ion mode. Parameters were set as follows: ion spray voltage, +50 eV; sheath gas, 45 Arb; auxiliary gas, 10 Arb; curtain gas, 2 Arb; auxiliary gas temperature, 300 °C; capillary transfer tube temperature, 350 °C. Sample volume, 20 μL; flow rate, 0.3 mL/min; detection wavelength, 280 nm.

2.6. Antioxidant activity analysis

Antioxidant capacity was assessed using supplementary assays (DPPH, ABTS, and FRAP), guided by established mechanisms and assay limits. Results are expressed in μmol Trolox equivalents per gram of dry weight (μmol TE/g DW). Each assay was baseline-corrected using a blank sample, and both standards and samples were analyzed within the linear range.

2.6.1. DPPH radical scavenging activity

The DPPH assay was conducted following a modified protocol (Nakagawa et al., 2021) Briefly, 1 mL of sample or Trolox standard was mixed with 1 mL of DPPH methanolic solution and incubated in the dark for 30 min. Absorbance was then measured at 520 nm. The Trolox standard curve was obtained as y = 0.0457× + 0.0366 (R² = 0.997).

2.6.2. ABTS⁺ radical scavenging activity

The ABTS assay was performed with minor modifications(Ali, Almagribi, & Al-Rashidi, 2016). A 7.4 mM ABTS solution was reacted with 2.6 mM potassium persulfate for 16 h in the dark. The resulting ABTS⁺ solution was diluted with phosphate buffer (0.2 M, pH 7.4) to an absorbance of 0.70 ± 0.02 at 734 nm. For the assay, 1 mL of sample was mixed with 3 mL of the diluted ABTS⁺ solution, and absorbance was recorded at 734 nm. The Trolox standard curve was y = 0.0611× + 0.0216 (R² = 0.998).

2.6.3. Reducing power (FARP)

The FRAP assay was performed as previously described (Kiss et al., 2025; Mažeikienė, Dringelytė, & Burokienė, 2026). The FRAP reagent was freshly prepared by mixing 25 mL of acetate buffer (pH 3.6) with 10 mL of 20 mM FeCl₃ and 10 mL of 10 mM TPTZ. An aliquot of 1 mL of sample was combined with 6.5 mL of ultrapure water and 225 μL of FRAP reagent. Following incubation at 37 °C for 8 min, absorbance was measured at 595 nm. The Trolox standard curve was y = 0.0574× + 0.0712 (R² = 0.998). All antioxidant assays were performed in triplicate, with l-ascorbic acid serving as the positive control.

2.7. Microcapsulation

2.7.1. Fabrication of microcapsules

Microcapsules were prepared using an ionic gelation method, as adapted from Zhao et al. (Q. Zhao, Gao, Jin, & Zhu, 2022). A wall material solution was formulated by mixing 2% (w/v) sodium alginate and 1.5% (w/v) carboxymethyl cellulose. The purified PA extract (core material) was dispersed into the wall material solution and stirred for 1 h to ensure homogeneity. The mixture was then extruded through a syringe into a 2% (w/v) calcium chloride solution (solidification bath) to form microcapsules. The resulting microcapsules were stirred for an additional hour, rinsed with deionized water, and freeze-dried. The dried product was stored at −20 °C. The encapsulation efficiency (EE) was determined to be 89.015% using the formula:

$$\mathit{EE}\left(\%\right)=\frac{W_{\mathit{total}}-W_{\mathit{free}}}{W_{\mathit{total}}}\ast 100\%$$

where EE is Encapsulation Efficiency. W_total is Total amount (or mass) of core material added. W_free is amount (or mass) of non-encapsulated (free) core material.

2.7.2. Morphological characterization by scanning electron microscopy (SEM)

The surface morphology of the PAs microcapsules was examined using a scanning electron microscope (Quattro S, Thermo Fisher Scientific, Waltham, Massachusetts, USA). Images were captured at accelerating voltages ranging from 1 to 30 kV at various magnifications to assess the structural features and surface characteristics (de Souza et al., 2018).

2.7.3. Thermogravimetric analysis (TGA)

The thermal stability of free PAs and PAs-loaded microcapsules was evaluated using a Thermogravimetric Analyser TGA2 (Mettler Toledo Group, Zurich, Switzerland).Samples were placed in alumina crucibles and heated from 30 to 500 °C at a rate of 10 °C/min under a nitrogen atmosphere (Li, Zhang, Zhao, Lv, & Liu, 2024). Thermograms were recorded to evaluate thermal degradation behavior and the protective effect of encapsulation.

2.7.4. Fourier transform infrared spectroscopy (FTIR)

FTIR analysis was conducted to investigate the potential interactions between the core and wall materials. Following the procedure described by Sheng et al. (Sheng et al., 2021), Spectra of free PAs, PA-loaded microcapsules, and blank (SA-CMC) microcapsules were recorded using an FTIR spectrometer. Samples were ground with potassium bromide (KBr) and analyzed in the range of 4000–500 cm⁻¹ with 32 scans at a resolution of 4 cm⁻¹.

2.7.5. In vitro gastrointestinal digestion

An in vitro digestion model simulating gastric and intestinal phases was employed, based on the standardized INFOGEST protocol with minor modifications ((Dupont, Lavoisier, Ménard, Nebbia, & Morzel, 2023; Minekus et al., 2014)). For the gastric phase, samples were acidified to pH 2.0, mixed with pepsin, and incubated at 37 °C for 2 h with agitation. For the subsequent intestinal phase, the pH was adjusted to 7.0, and simulated intestinal fluid containing electrolytes, pancreatin, and bile salts was added. Digestion was continued for an additional 2 h. At predetermined intervals, aliquots were withdrawn and heated to inactivate enzymes. The concentration of released PAs in the supernatant was measured spectrophotometrically, and the cumulative release rate was calculated as:

$$\mathit{Release\; rate}\left(\%\right)=\frac{W_r}{W_t}\times 100\%$$

where W_r is the mass of released PAs, and W_t is the total mass of proanthocyanidins loaded in the sample.2.8 Data analysis.

All experiments were conducted in triplicate, and the results are expressed as mean ± standard deviation (SD). Statistical significance was determined using one-way analysis of variance (ANOVA), followed by Tukey's multiple comparison test for post hoc analysis. A p-value of <0.05 was considered statistically significant. Data analysis was carried out using SPSS 22.0 (SPSS Inc.， Chicago， IL， United States)and graphical visualizations were prepared using Origin 2024 (OriginLab, Northampton, Massachusetts, USA).

3. Results and discussion

3.1. Selection of the optimum DES

The extraction efficiency of PAs from sea buckthorn seeds varied as a function of water content across different DESs, exhibiting a characteristic bell-shaped curve (Fig. 1). Water content is a critical determinant of DES physicochemical properties, including viscosity, polarity, and surface tension (Kivela et al., 2022). (Vilková, Płotka-Wasylka, & Andruch, 2020). High viscosity impedes mass transfer, thereby limiting extraction efficiency (Dai, Witkamp, Verpoorte, & Choi, 2015). The addition of moderate amount of water reduces DES viscosity, enhancing solute diffusivity. However, excessive water content can dilutes the effective solvent concentration and disrupt the hydrogen-bonding network crucial for solvent-solute interactions, ultimately diminishing extraction yield (Dabetić, Todorović, Panić, Radojčić Redovniković, & Šobajić, 2020). Consequently, the extraction yield of bioactive compounds shows a bell-shaped curve in response to increasing water content, which aligns with the findings of this study. Furthermore, while increased water content enhances the polarity of DES, potentially improving the extraction of polar compounds, maximum extraction is not solely dependent on polarity matching. Other intermolecular interactions between solvent and analyte also play a role (Aydin, Yilmaz, & Soylak, 2018). Thus, optimizing water content is essential to balance viscosity reduction and extraction power.

Fig. 1.

Extraction of PAs from sea buckthorn seeds by different deep eutectic solvents.

In existing research, the use of organic reagents for PAs extraction is a classical method. As demonstrated in Li's(Jia et al., 2024) study, DES extraction efficiency for PAs was found comparable to acetone and superior to traditional solvents such as methanol and ethanol. The hydroxyl groups in phenolic compounds can form hydrogen bonds with the HBD and HBA components in DES, creating an interaction network. Together, they enhance the polarity of deep eutectic solvents (DES), improving interactions with substances like sugars, polyols, organic acids, and choline chloride, thereby promoting the extraction process(Nguyen, Nguyen, & Kha, 2025). The hydrogen bonding interactions between DES and polyphenols are very strong, stronger even than those between water and solutes, and their bonding strength exceeds the electrostatic forces between other polyphenols(Zhou, Fakayode, & Li, 2023). When hydrogen bonding is enhanced, phenolic compounds exhibit higher solubility in DES. This occurs because dipole-dipole and ion-dipole interactions weaken the binding forces of phenolic components within the plant matrix, facilitating their uniform dissolution in DES(Zannou & Koca, 2022).Mechanistically, the water-content-dependent extraction efficiency of PAs can be interpreted by three competing effects:

(i) Viscosity–mass transfer trade-off. At low water content (<10%, w/w), high DES viscosity restricts solvent penetration into seed matrix pores, limiting PA release(Dai et al., 2015). Moderate water addition disrupts inter-DES hydrogen bonds between HBA and HBD, reducing viscosity and enhancing diffusivity (Kivela et al., 2022).

(ii) Solvent–solute hydrogen bond competition. Water molecules compete with PAs for DES hydrogen-bonding sites. At optimal hydration, DES retains sufficient HBD/HBA groups to form bidentate hydrogen bonds with PA phenolic –OH groups, which is stronger than water–PA interactions(Zhou et al., 2023). Excess water (>40%) hydrates DES components, shifting the equilibrium from DES–PA toward water–PA and water–DES, thus weakening extraction(Dabetić et al., 2020).

(iii) Polarity and solvation selectivity. While higher water content increases DES polarity (lower log P), maximum PA extraction is not purely polarity-driven. Florindo et al., (Florindo, Oliveira, Rebelo, Fernandes, & Marrucho, 2014) and DJ (Ramón & Guillena, 2020)demonstrated that dipole–dipole and ion–dipole interactions between DES chloride anions and PA aromatic rings contribute significantly to solubility, which diminishes at high water dilution.

In this study, all five ChCl-based DES systems exhibited significantly higher PA extraction absorbance than 75% ethanol (p < 0.05). The ChCl/1,4-butanediol DES with 50% water achieved the maximum absorbance of 0.255 ± 0.008, which was 201.2% higher than the ethanol control group (absorbance ≈ 0.084). The second-best system was malonic acid DES at 60% water (0.228 ± 0.005), 10.6% lower than the optimal DES; urea DES with 70% water only reached 0.11, showing nearly 56.9% lower extraction capacity. All five ChCl-based DES systems exhibited significantly higher PA extraction absorbance than 75% ethanol (p < 0.05). The ChCl/1,4-butanediol DES with 50% water achieved the maximum absorbance of 0.255 ± 0.008, which was 201.2% higher than the ethanol control group (absorbance ≈ 0.084). The second-best system was malonic acid DES at 60% water (0.228 ± 0.005), 10.6% lower than the optimal DES; urea DES with 70% water only reached 0.11, showing nearly 56.9% lower extraction capacity.

The results indicate that the type of hydrogen bond donor in DES and the water content have a highly significant effect on the extraction efficiency of PAs (p < 0.05). Among all tested systems, the 1,4-butanediol-based DES exhibited the best extraction performance, with the PAs extraction efficiency reaching a peak (0.255 ± 0.008) in the 50% water content group, which was significantly higher than that of other DES systems and the conventional 75% ethanol control group (p < 0.05). Malonic acid-based DES (60% water content group, 0.228 ± 0.005) and lactic acid-based DES (50% water content group, 0.214 ± 0.004) exhibited the next-highest extraction efficiencies, while the overall extraction capabilities of urea- and citric acid-based DES were significantly lower. With some high-water-content groups (e.g., the urea 70% water content group) exhibiting an absorbance of only around 0.11, indicating extremely poor extraction results. Compared to the conventional 75% ethanol extraction method, the PAs extraction efficiency of the 1,4-butanediol-based DES increased by over 200%, validating the significant application potential of DES as a green extraction solvent in the extraction of natural phenolic compounds. This study identified the optimal DES extraction system for PAs from sea buckthorn seeds, providing a solid experimental basis for the development of a green and efficient extraction process.

3.2. Determination of optimal extraction conditions

Based on preliminary screening, the choline chloride–1,4-butanediol DES containing 50% water as selected for further optimization. Single-factor experiments (Fig. 2) indicated that PA yield increased with solid-to-liquid ratio up to a maximum at 30:1 (mg/mL), suggesting optimal mass transfer and solvent availability. Similarly, yield peaked at an extraction time of 3 h and a temperature of 60 °C. Prolonged extraction times or elevated temperatures (70 °C) resulted in diminished yields, likely attributable to the degradation of thermolabile PAs (Qu, Shi, Li, Pan, & Venkitasamy, 2014). (Qu et al., 2014).

Fig. 2.

One-factor experiment on proanthocyanin extraction.

Response surface methodology was employed to further refine these parameters (Fig. 3). The experimental design and corresponding results are presented in Table 4. The data were fitted to a second-order polynomial model:

$$\mathrm{Y}=-231.38183+446.93281\mathrm{A}+7.1720\text{3B}+30.8155\text{9C}-2.8825\text{6AB}-25.10062\mathrm{AC}+0.060773\mathrm{BCE}-2336.4296\text{9A}2–0.058611\mathrm{B}2–5.2878\text{4C2}$$

Fig. 3.

Response surface and contour plots extracted from PAs.

Analysis of variance data from Table 5 was further interpreted to evaluate the significance of individual variables and their interaction terms. The F-value reflects the magnitude of the influence, whereas the p-value indicates the statistical significance. For the linear terms, A (solid-to-liquid ratio), B (extraction time), and C (extraction temperature) all exhibited significant effects on PA yield (p < 0.05). By comparing F values, the order of factor influence was determined as A > C > B, suggesting that the solid-to-liquid ratio exerted the most prominent effect on PA extraction yield, followed by extraction temperature and extraction duration.

Regarding pairwise interaction terms, AC showed extremely high significance (p < 0.001) with a large F-value, indicating a strong interaction between solid-to-liquid ratio and extraction temperature, which can also be visually supported by the steep curved surface and closed elliptical contour lines in Fig. 3B. In contrast, the p-value of term AB was greater than 0.05, demonstrating that the interaction between solid-to-liquid ratio and extraction time was not statistically significant; Fig. 3A also presents relatively rounded contour features consistent with weak interaction. The BC interaction was non-significant (p > 0.05), implying extraction time and temperature had negligible combined effects on PA yield (Fig. 3C). All quadratic terms A², B² and C² reached significant levels (p < 0.05), revealing obvious nonlinear relationships between these variables and PA yield.

From a mechanistic perspective, the strong interaction between solid-to-liquid ratio and temperature can be explained by DES viscosity-mass-transfer behavior: higher temperature reduces DES viscosity and improves mass transfer efficiency, but this promoting effect is dependent on sufficient solvent dosage. Excessively high temperature will accelerate thermal degradation of thermolabile sea buckthorn-seed proanthocyanidins, thus resulting in a maximum yield within a suitable parameter window. The shape of response-surface and contour plots further visualizes these competitive effects among extraction parameters.

ANOVA results (Table 5) further confirmed the quadratic regression model possessed extremely high fitting goodness with R² = 0.9964 and adjusted R² = 0.9917, while the lack-of-fit p-value was 0.7337 (p > 0.05), proving the model could accurately predict PA yield. The theoretical maximum PA yield was 48.025 mg/g under 41:1 mg/mL, 61.11 °C, 3.29 h. Validation experiments conducted under adjusted practical conditions (41,1 ratio, 61 °C, 3.3 h) yielded an actual PA content of 47.591 mg/g, confirming the model's reliability with a relative error of only 0.90%.

3.3. Identification of PAs in sea buckthorn seeds

The purified PA powder used for LC-MS detection contained 69.17% total proanthocyanidins, which was obtained after D101 macroporous resin separation and freeze-drying. Mass spectra confirmed m/z 291.08/292.08 monomers and m/z 579.15–581.15 B1–B4 dimers, whereas trimer (m/z 833, 849) and tetramer (m/z 1155) signals were undetectable, indicating monomers and dimers accounted for almost all detectable PAs in the seed extract.

LC-MS analysis of the purified extract confirmed the presence of monomeric and dimeric PAs. Comparison with catechin and epicatechin standards revealed corresponding peaks at m/z 291.08 and 292.08 in the extract, confirming the presence of these monomeric flavan-3-ols (Fig. 4). More specifically, Furthermore, analysis using a PA B1-B4 standard confirmed the presence of these dimeric isomers in the extract, as evidenced by distinct peaks at m/z 579.15, 580.15, and 581.15 (Fig. 4C, D). These findings are consistent with prior characterizations of PAs in other plant sources (Cuevas-Rodriguez et al., 2010) (Gu et al., 2003). While DES composed of choline chloride and 1,4-butanediol was thus validated as effective for extracting PAs from sea buckthorn seeds. Further characterization was carried out using PAs B1-B4 standard (Fig. 4-C). Since PAs B1-B4 are dimeric isomers, their separation during analysis is inherently difficult. However, three distinct peaks at m/z 579.15, 580.15, and 581.15 were observed in both the standard and the sea buckthorn extract spectra (Fig. 4-D), confirming the presence of PAs B1–B4. These findings were consistent with Gu et al. (2003), confirming that choline chloride-1,4-butanediol is effective in extracting PAs B1-B4 from sea buckthorn seeds. Additionally, a PAs dimer consisting of gallic acid, catechin, and epicatechin was identified, aligning with previous studies (Zhang et al., 2017).

Fig. 4.

Mass spectra for the detection of catechins and epicatechins in standards and sea buckthorn seed extracts. A – Catechin and epicatechin standards; B - Catechin and epicatechin in extracts. C-PAs B1-B4 standard; D-PAs B1-B4 in sea buckthorn seed extracts.

While Gu et al.(Gu et al., 2003), identified a PAs trimer at m/z 833 and m/z 849, and González-Manzano, Santos-Buelga, Pérez-Alonso, Rivas-Gonzalo, and Escribano-Bailón (2006) (González-Manzano et al., 2006) reported a tetramer at m/z 1155, no clear peaks for these larger PAs were found in this study. (Fig. 4). This suggests that sea buckthorn seed PAs are primarily composed of monomers and dimers.

3.4. Antioxidant capacity of PAs in sea buckthorn seeds

The PAs extracts obtained from sea buckthorn seeds were freeze-dried and the resulting powder was subjected to various antioxidant assays (Table. 7).The results showed that the antioxidant activities of PAs in the DPPH, TEAC, and FRAP systems were 4.61, 5.53, and 6.24 μM TE/mg, respectively, while those of ascorbic acid were 5.56, 5.69, and 7.96 μM TE/mg. Overall, The extracted PAs presented DPPH = 4.61 μmol TE/mg, ABTS = 5.53 μmol TE/mg, FRAP = 6.24 μmol TE/mg; positive ascorbic acid control recorded 5.56, 5.69, 7.96 μmol TE/mg respectively. The PA sample was 17.1% lower in DPPH activity, 2.8% lower in ABTS and 21.6% lower in FRAP than ascorbic acid, yet it retained strong reducing and radical-scavenging capacity across three independent test systems. However, PAs demonstrated significant free radical scavenging and iron ion reduction capabilities in all three antioxidant systems based on different principles. The activity measured by the FRAP method was the highest, indicating that PAs possess strong reductive antioxidant activity.

From a methodological perspective, the three models exhibit fundamental differences in their mechanisms of action: The FRAP method is based on the reaction where the Fe³⁺-TPTZ complex is reduced to the blue-violet Fe²⁺-TPTZ under acidic conditions (pH 3.6), directly reflecting the total reducing capacity of the substance and showing high sensitivity to hydrogen-donating antioxidants such as phenols(Gulcin, 2020; Shahidi & Samarasinghe, 2025); The DPPH assay measures the sample's ability to scavenge the stable organic radical DPPH, reflecting its hydrogen atom transfer capacity; however, it has response limitations for certain polyphenolic compounds with significant steric hindrance; The TEAC (ABTS⁺) method, based on the scavenging reaction of the ABTS⁺ cation radical, covers both water-soluble and lipid-soluble antioxidants, providing a more comprehensive reflection of total radical scavenging capacity. The pattern where these two test substances exhibit the highest activity in the FRAP method and the lowest in the DPPH method aligns closely with the underlying methodological principles(Thaipong, Boonprakob, Crosby, Cisneros-Zevallos, & Byrne, 2006).

As natural polyphenolic compounds, the antioxidant activity of PAs primarily stems from the hydrogen-donating capacity of the numerous phenolic hydroxyl groups in their molecular structure and their metal ion chelation properties. Although their activity is slightly lower than that of ascorbic acid, as natural plant extracts, they offer higher application safety and stability in the food and health supplement sectors, presenting promising prospects for development and application.

As expected, the positive control ascorbic acid showed slightly higher values (5.56, 5.69, and 7.96 μM TE/mg). To contextualize these results within existing literature, our DES extracted PAs showed comparable antioxidant activity to previously reported values. Fan et al. (Fan, Ding, & Gu, 2007)demonstrated that water-acetone (3,7) extracts of sea buckthorn seeds exhibited strong DPPH radical-scavenging activity, which was attributed to the high content of proanthocyanidins (mean degree of polymerization 12.2). Similarly, Dabetić et al. (Dabetić et al., 2020) reported that DES extraction (ChCl-citric acid) from grape seeds yielded polyphenol extracts with antioxidant activity comparable to or even exceeding that of conventional organic solvents. This comparison demonstrates that DES extraction not only preserves. But in some aspects enhances the antioxidant capacity of PAs relative to both conventional solvents and other DES systems. This result is consistent with the findings of Bagchi et al. (Bagchi, Swaroop, Preuss, & Bagchi, 2014).

3.5. SEM

As shown in Fig. 5, the SEM images of the PAs microcapsules reveal that the microcapsules are generally elliptical in shape, with no visible tears or holes. This morphology may result from mechanical compression between particles during the production process. Surface folds were observed on the microcapsule exterior and were likely caused by particle shrinkage during the freeze-drying process, consistent with the findings of Rosenberg (1985) (Rosenberg, Kopelman, & Talmon, 1985). These surface folds increase the specific surface area, potentially enhancing the microcapsule's ability to retain and protect the encapsulated PAs. The embedding efficiency of the PAs microcapsules was calculated to be 89.015%, indicating a highly successful encapsulation process with significant retention of PAs. Furthermore, Fig. 6 A1 and A2 demonstrate that the microcapsules possess a well-defined internal structure, effectively encapsulating the dispersed PAs. The PAs, initially in a dispersed state, were transformed into regular and uniform solid particles, within the microcapsules. The homogeneous distribution of PAs suggests that the microcapsules system provides structural stability and effective protection, supporting its potential for PAs preservation and controlled release.

Fig. 5.

Morphological structure of scanning electron microscope at different magnifications. A: PAs microcapsules B: PAs.

Fig. 6.

FTIR plots of PAs and PAs microcapsules.

3.6. FTIR

Molecular interactions between free PAs, PAs microcapsule and SA-CMC were evaluated by analyzing Fourier transform infrared spectroscopy (Fig. 6). The three curves exhibit distinct similarities and differences in the characteristic absorption regions, intuitively reflecting the interactions between the core and shell materials as well as the formation of the microcapsule structure. Near 3300 cm⁻¹, PAs exhibit a broad, intense absorption peak caused by the O—H stretching vibration of phenolic hydroxyl groups(de Souza et al., 2018). while empty microcapsules exhibit a stronger and broader absorption feature due to intermolecular hydrogen bonding among polysaccharides. The absorption peak intensity of the microcapsule samples in this region lies between the two, with a flatter peak shape, indicating that hydrogen bonding interactions form between the phenolic hydroxyl groups of PAs and the hydroxyl groups of the wall material polysaccharides, while the encapsulation effect of the wall material weakens the hydrogen bonding between PAs molecules. The absorption peak at 2926 cm⁻¹ corresponds to the C—H stretching vibration of phenolic compounds(Keresztury, Billes, Kubinyi, & Sundius, 1998). All three curves exhibit distinct absorption; the microcapsule curve represents the superposition of signals from the core and shell materials, with no significant shift in peak intensity, confirming that the hydrocarbon backbone structure remains stable in the system.

The absorption peak at 1600 cm⁻¹ is attributed to the aromatic ring C C stretching vibration in PAs(Abdel-Gaber, Ezzat, & Mohamed, 2022), while in empty microcapsules it primarily corresponds to the asymmetric stretching vibration of the carboxyl group (-COO⁻)(P. Zhao et al., 2020). The microcapsule curve exhibits a superposition of these two signals at this wavelength, with an absorption intensity significantly higher than that of PAs, further confirming the effective composite formation between the shell and core materials. The absorption peaks at 1450 cm⁻¹ and 1345 cm⁻¹ both exhibit broadening and intensity reshaping in the microcapsule system, resulting from the combined effects of the PAs C—O vibration and the polysaccharide carboxyl and hydroxyl vibrations(Sowers, Stuckey, & Sparks, 2018). Particularly notable is the 1020 cm⁻¹ region, where PAs exhibit only weak absorption, whereas empty microcapsules display a strong absorption peak at this wavelength, corresponding to the stretching vibration of the polysaccharide C-O-C glycosidic bond(Holden, McAinsh, Taylor, Beckett, & Martin, 2024). The microcapsule spectrum also exhibits strong absorption characteristics, showing a high degree of consistency with the trend observed in the empty microcapsules. This fully demonstrates that the microcapsule wall structure is intact and that the PAs have been successfully encapsulated within the sodium alginate–carboxymethyl cellulose matrix, which is consistent with the findings of Song (Song et al., 2022) and others. Overall, no new characteristic peaks appeared in the microcapsule spectrum, proving the non-covalent hydrogen-bond interaction between SA-CMC wall and PA core did not destroy the polyphenol skeleton. The core and shell materials formed a stable complex through non-covalent interactions, providing strong spectroscopic evidence for the encapsulation efficacy.

3.7. Thermogravimetric analysis (TGA)

The thermal stability of free PAs, blank SA-CMC microcapsules, and PAs-loaded SA-CMC microcapsules was evaluated via thermogravimetric analysis (Fig. 7) over a temperature range of 25–600 °C, as illustrated in Fig. X. Free PAs exhibited a single-stage, rapid thermal degradation profile, with a sharp mass loss occurring between 100 and 150 °C, resulting in a residual char yield of only ∼8% at 600 °C, which is consistent with the well-documented poor thermal stability of PAs due to their polyphenolic structure and susceptibility to oxidative degradation at moderate temperatures(Lee, 2010). In contrast, the blank SA-CMC microcapsules displayed a two-step, gradual thermal decomposition process: an initial minor mass loss (<5%) below 100 °C attributed to the removal of adsorbed moisture, followed by a major degradation stage between 100 and 300 °C corresponding to the pyrolysis of the polysaccharide backbone of SA and CMC, ultimately achieving a high residual char content of ∼52% at 600 °C, indicative of the excellent thermal barrier properties of the SA-CMC wall material(Li et al., 2024).

Fig. 7.

Schematic TGA curves of PAs and PAs microcapsule.

Notably, the TGA curve of PAs-loaded microcapsules showed no distinct rapid degradation peak characteristic of free PAs, and its thermal decomposition behavior closely mirrored that of the blank microcapsules, with a significantly delayed and attenuated mass loss compared to free PAs. The residual char yield of PAs microcapsules (∼26% at 600 °C) fell between that of free PAs and blank microcapsules, confirming the successful encapsulation of PAs within the SA-CMC matrix. These findings demonstrate that the SA-CMC wall material effectively acts as a physical barrier, significantly improving the thermal stability of PAs by shielding the core material from direct heat exposure, which is a critical advantage for the application of PAs in food and pharmaceutical systems where thermal processing is involved(Sheng et al., 2021).

3.8. In vitro release of PAs from microparticles

Fig. 8A illustrates the release kinetics of PAs under simulated gastric conditions. Free PAs showed rapid dissolution in simulated gastric fluid (SGF), reaching 29.58% release within the first 30 min In contrast, microencapsulated PAs exhibited a significantly lower release rate of only 2.86% demonstrating the barrier effect of the microcapsule matrix. This delay release is primarily attributed to the pH-sensitive nature of sodium alginate. At low pH levels, the protonation of carboxyl groups in the alginate chains leads to reduced hydrophilicity and solubility, thereby restricting water penetration and swelling of the matrix (Bannikova, Rasumova, Evteev, Evdokimov, & Kasapis, 2017). (Bannikova et al., 2017). Furthermore, intermolecular hydrogen bonding between sodium alginate, carboxymethyl cellulose and PAs likely contributes to structural rigidity and limited diffusion of the active compound (Nwabor, Singh, Marlina, & Voravuthikunchai, 2020). Although the release of PAs from microcapsules gradually increased beyond 90 min, it remained consistently lower than that of the unencapsulated form throughout the experimental period. After 180 min, the cumulative release from unembedded PAs reached 53.64%, while the release from microcapsules was 21.99%, clearly demonstrating a controlled and sustained-release effect. These findings emphasize that microencapsulation effectively shields PAs from gastric degradation, making it a promising strategy for enhancing the stability and bioavailability of sensitive polyphenolic compounds during oral administration.

Fig. 8.

In vitro release profiles of PAs and microencapsulated PAs in simulated gastric fluid (SGF) and simulated intestinal fluid (SIF).

The release behavior of PAs from microcapsules in simulated intestinal fluid is depicted in Fig. 8B. The data demonstrate a positive correlation between time and PA release for both encapsulated and unencapsulated forms. However, while the release rate of free PAs gradually declined over time, the release from microcapsules showed a steady increase, suggesting a sustained-release profile in the intestinal environment. When compared with gastric conditions (as shown in Fig. 8A), the release rate of microencapsulated PAs was notably higher in intestinal fluid, which can be due attributed to the neutral to mildly alkaline pH of the intestinal medium. In contrast, the highly acidic gastric environment (pH < 2.0) promotes the aggregation of alginate-based polymers, effectively reducing the permeability of the microcapsule matrix and restricting PA release (Sarkar, Horne, & Singh, 2010). This protective function helps preserve the integrity of PAs during gastric transit and facilitates their targeted release in the intestine. At 180 min, free PAs exhibited a cumulative release of 39.57%, showing signs of plateauing, whereas microcapsules reached a higher release rate of 45.65%, with a continued upward trend. This indicates that the encapsulation technique not only protects the bioactive compounds in the acidic gastric phase but also enhances their availability in the intestinal environment, which is crucial for maximizing absorption and bioactivity.

These findings support the potential of microencapsulation as a delivery system for controlled intestinal release of PAs, aligning with previous studies that report improved core material stability in gastric conditions and increased bioavailability during intestinal digestion.

From the results of two in vitro simulated digestion experiments, it is evident that unencapsulated PAs exhibited a rapid release rate during the initial early stages of digestion, which then gradually plateaued over time. In contrast, microencapsulated PAs demonstrated a more controlled and nearly linearly release profile throughout the digestive process. This sustained release behavior of microcapsules ensures that the active components of PAs are steadily released, within the intestinal tract, potentially enhancing their bioavailability and effective absorption in the human body. This conclusion lays a certain theoretical foundation for the microencapsulation of PAs. These findings underscore the functional advantage of microencapsulation in protecting bioactive compounds during gastric transit and promoting targeted release in the intestine. Therefore, this study provides a solid theoretical basis for the application of microencapsulation technology in improving the stability and bioefficacy of PAs in functional foods or nutraceutical delivery systems.

4. Conclusion

This study successfully developed an integrated green process combining deep eutectic solvent extraction and sodium alginate-carboxymethyl cellulose composite microencapsulation to valorize discarded sea buckthorn seed by-products. The optimal choline chloride/1,4-butanediol DES with 50% (v/v) water yielded a maximum proanthocyanidin (PA) content of 48.025 mg/g under optimized extraction conditions (solid-liquid ratio 41:1 mg/mL, 61 °C, 3.3 h). After purification via D101 macroporous resin, the PA-rich powder reached a purity of 69.17%, which mainly comprised catechin, epicatechin monomers and PA B1–B4 dimers, with DPPH, ABTS and FRAP antioxidant activities of 4.61, 5.53 and 6.24 μmol TE/mg, respectively. Ionic gelation prepared composite microcapsules achieved an encapsulation efficiency of 89.015%. TGA quantification verified that the SA-CMC matrix drastically delayed PA thermal degradation, with microcapsule residual char (26%) falling between free PAs (8%) and blank wall materials (52%). In vitro digestion tests demonstrated microcapsules reduced gastric PA premature release by 31.65% (21.99% vs. 53.64% at 180 min) and delivered a higher intestinal cumulative release of 45.65% compared with free PAs (39.57%).

Despite the prominent advantages of this DES-microcapsule system, two practical bottlenecks limit its industrial translation: the relatively high viscosity of the ChCl/1,4-butanediol DES restricts continuous large-scale extraction, and the freeze-drying step for microcapsule preparation elevates overall production costs. For future research, three promising directions are proposed. First, novel low-viscosity natural DES combinations (e.g., polyol-organic acid ternary systems) can be screened and coupled with ultrasonic or microwave auxiliary technologies to boost continuous extraction efficiency and cut solvent dosage. Second, cheaper alternative wall materials such as agricultural waste-derived polysaccharides (corn fiber, rice bran cellulose) can be blended with SA-CMC to lower microcapsule manufacturing costs while retaining barrier performance. Third, in vivo animal trials are required to systematically evaluate the bioavailability, tissue distribution and long-term biosafety of encapsulated sea buckthorn seed PAs, laying preclinical foundations for its application as functional food additives or mild oral nutritional supplements. Moreover, this integrated green strategy can be extended to recover polyphenolic bioactive from other fruit seed processing residues, providing a universal technical reference for circular utilization of agro-industrial waste resources.

CRediT authorship contribution statement

Yuxuan Sun: Writing – review & editing, Writing – original draft, Validation, Software, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Dong Sun: Visualization, Validation, Supervision. Mehraj Ahmad: Writing – review & editing, Supervision, Resources. Xuan Li: Supervision, Software, Investigation. Meigui Huang: Visualization, Project administration, Methodology. Xiufang Bi: Writing – review & editing, Writing – original draft, Validation. Lu Han: Visualization, Validation, Resources, Project administration, Investigation. Gang Hao: Writing – review & editing, Writing – original draft, Methodology, Funding acquisition.

Ethics declaration

The author(s) declare(s) that the study does not involve humans nor animal subjects.

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.

Data availability

Data will be made available on request.