Extraction, characterization and antioxidant activity of free and bound polyphenols from Undaria pinnatifida

Abstract

Free polyphenols (FP) and bound polyphenols (BP) were extracted from Undaria pinnatifida, with processes optimized by response surface methodology. BP accounted for approximately 65% of total polyphenols. UHPLC-QTOF-MS/MS tentatively identified 28 phenolic compounds in BP and 3 in FP, with characteristic brown algal polyphenols, such as eckol-type phlorotannin derivatives and highly oxidized phlorotannin dimers, exclusively detected in BP. They showed stronger DPPH radical scavenging capacity than FP, even surpassing tea polyphenols at 3.5 mg/mL. These findings demonstrate that alkaline extraction is essential for accessing polyphenolic resources otherwise inaccessible through ethanol extraction, and BP possess potential as natural antioxidants. This study may provide a reference for the comprehensive utilization of UP polyphenolic resources.

Highlights

• •It is the first time to simultaneously extract free and bound polyphenols from U. pinnatifida.
• •LC-MS/MS was used to characterize free and bound polyphenols.
• •The substances identified by combined polyphenols are more abundant than free polyphenols.
• •Bound polyphenols from U. pinnatifida are good natural antioxidants.

1. Introduction

Undaria pinnatifida (UP) is a one-year-old economically significant seaweed from the family Undariaceae, primarily distributed along the eastern coasts of China, Japan, and South Korea, typically found from the low tide line to depths of approximately 10 m. This type of algae can reach heights of 50–150 cm and exhibits vigorous growth during the winter and spring seasons. From late spring to early summer, auricular sporangia are formed at its base (Catalina et al., 2026). As a vital marine biological resource, UP is rich in natural active substances, particularly polyphenols (Zhang et al., 2020). Its structural characteristics differ significantly from ortho-phenolic compounds found in terrestrial plants, with phloroglucinol serving as the characteristic skeleton.

As an essential active component in UP (Wang et al., 2018), polyphenols are mainly classified into two categories: free polyphenols (FP) and bound polyphenols (BP) (Wei et al., 2025). These polyphenolic types collectively represent the key activity value of UP polyphenols due to their synergistic interactions. FP primarily occurs in the food matrix as monomeric units, demonstrating a high degree of solubility in water and easily dissolving in both water and organic solvents. Conversely, BP is generally located within the matrix of plant cell walls, where it associates with macromolecules such as cellulose, structural proteins, and polysaccharides of the cell wall through covalent bonding, which can take the form of ester, ether, or C—C linkages. BP may appear in a variety of structures, including monomers, dimers, or oligomers, complicating their extraction process (Peng et al., 2024). The concentration of BP tends to be highest in the residues remaining after the hydrolysis of free polyphenols, often overlooked because it is not released during this process, leading to the wastage of valuable resources. Effective recycling of these BP can not only fully exploit the potential value of UP but also maximize resource utilization. Existing studies indicate that the content of BP in plants is not only significantly higher than that of FP but also exhibits superior biological activity (Qiu et al., 2026).

Currently, numerous reports have been published regarding the extraction, purification, and biological activities of polyphenols found in algae, grains, and fruits and vegetables (Murdiono et al., 2025). Among these, research on the polyphenols of UP primarily concentrates on FP and its content determination, while studies related to BP are comparatively scarce. Consequently, this study systematically investigates the optimal extraction conditions for both FP and BP, employing petroleum ether-ethyl acetate extraction and purification techniques to obtain solid samples via vacuum freeze-drying. We compared the phytochemical components of the two forms of polyphenols, including polysaccharides, proteins, total flavonoids, total phenols, and brown algae polyphenols, and analyzed their composition using LC-MS/MS technology. The results from in vitro antioxidant experiments demonstrated that the antioxidant activity of BP was significantly superior to that of FP, thereby enhancing the overall utilization value of polyphenols in UP. The comprehensive development of polyphenols from UP, particularly the recycling of BP, may have potential applications in the fields of food, medicine, and cosmetics. However, these applications require further experimental validation, particularly in vivo studies. This research not only facilitates the deep processing of UP resources and increases product added value but also effectively addresses the issue of resource waste, achieving efficient utilization of these resources.

2. Materials and methods

2.1. Materials

Dried Undaria pinnatifida (UP) was purchased from Xiapu County, Fujian Province, China. The dried sample was ground and sieved through a 40-mesh screen (particle size <0.425 mm). The resulting powder was sealed and stored in a cool, dry place until further use. Phloroglucinol (purity 99%) was purchased from Shanghai Yien Chemical Technology Co., Ltd. (Shanghai, China). 2,4-Dimethoxybenzaldehyde (DMBA, purity 98%) was obtained from Shanghai McLean Biochemical Technology Co., Ltd. (Shanghai, China). Gallic acid (3,4,5-trihydroxybenzoic acid, purity 98%) and 1,1-diphenyl-2-picrylhydrazyl (DPPH, purity 97%) were supplied by Hefei Bomei Biotechnology Co., Ltd. (Hefei, China). All other solvents and reagents were of analytical grade and purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). They were used as received without further purification.

2.2. Single factor experimental and response surface optimization of free polyphenols

Ultrasonic-assisted ethanol extraction (Zhang et al., 2026) was employed to investigate the effects of ultrasonic temperature (set at 40, 45, 50, 55, and 60 °C), ultrasonic extraction time (ranging from 40 to 80 min), liquid-solid ratio (between 20 and 60 mL of solvent per gram of solid), ultrasonic power density (ranging from 30 to 50 W/L), and the volume fraction of ethanol used (spanning from 40% to 80%) on the extraction process of FP, with FP content serving as the primary index. Building upon the findings obtained from the single-factor experimental, the Box-Behnken response surface method was employed to further refine and optimize the extraction process. A three-factor, three-level experimental framework was established to investigate how the ethanol volume fraction (X₁), ultrasonic time (X₂), and ultrasonic power density (X₃) influence FP content (Y₁).

2.3. Single factor experimental and response surface optimization of bound polyphenols

The residue of UP, following the extraction of FP under optimal process conditions, was subjected to further treatment (Liang et al., 2025). The effects of various factors on the extraction process of BP were investigated through single-factor experimental, including NaOH concentration (0.04, 0.08, 0.12, 0.16, 0.20 mol/L), ultrasonic temperature (40, 50, 60, 70, 80 °C), ultrasonic time (60, 70, 80, 90, 100 min), liquid-solid ratio (30, 40, 50, 60, 70 mL/g), and ultrasonic power density (30, 35, 40, 45, 50 W/L). After analyzing the outcomes from the single-factor experiments, response surface methodology was utilized to create an optimized test involving three factors at three levels. This research particularly investigated how NaOH concentration (Z₁), liquid-solid ratio (Z₂), and ultrasonic time (Z₃) affected the BP content (Y₂).

2.4. Determination of TPC

Based on the Folin-Ciocalteu method (Torres et al., 2024), 1 mL of the sample solution was taken, and 1.25 mL of Folin-Ciocalteu reagent was added. After 3 min, 4 mL of a 0.1 g/mL anhydrous sodium carbonate solution was introduced. The reaction mixture was allowed to stand in the dark for 60 min, after which the absorbance was measured at 760 nm. The findings were reported in terms of gallic acid equivalents (μg/mg GAE/g DW) for each gram of dry mass.

2.5. Separation and purification of polyphenols

The UP filtrates, encompassing both FP and BP, were sourced by refining extraction parameters and utilizing high-speed centrifugation. Initially, concentration for the FP filtrate was performed using a rotary evaporator set at 50 °C and 60 rpm until a predetermined volume was achieved. This was succeeded by three consecutive extractions with equal amounts of petroleum ether, followed by three extractions with the same volume of ethyl acetate. The rotary evaporator then facilitated the removal of the organic solvents. The concentrated solution was subsequently freeze-dried in a vacuum freeze dryer maintained at −25 °C and 50 Pa for 36 h. In the case of the BP filtrate, its pH was modified to 3.0 prior to concentration through rotary evaporation at 55 °C and 60 rpm, with the following procedures mirroring those employed for the FP filtrate.

2.6. Identification of main phytochemical components

2.6.1. Total flavonoid content

The determination of total flavonoid content in the samples was carried out through colorimetric analysis (Zheng et al., 2024). First, a specific quantity of the sample was accurately weighed and then dissolved in 60% ethanol, followed by ultrasonic extraction for one hour. Following this step, 2 mL of the filtrate was mixed with 500 μL of a 5% sodium nitrite solution and allowed to react for 5 min. Afterward, 500 μL of a 10% aluminum nitrate solution was introduced, and following a 6-min reaction period, 4 mL of a 4% sodium hydroxide solution was added. To conclude, 3 mL of a 60% ethanol solution was incorporated, and the absorbance was assessed at a wavelength of 510 nm utilizing a microplate reader. The quantitative analysis of total flavonoid content was performed using rutin as the standard via the external standard method. The results were expressed as milligrams of rutin equivalent (RE) per gram of dry extract (DE).

2.6.2. Polysaccharide content

The determination of polysaccharide content was carried out through the phenol‑sulfuric acid method (Rover et al., 2013). First, the sample was extracted ultrasonically with an adequate volume of distilled water for one hour, and the resulting filtrate was set aside for subsequent analysis. At the same time, a glucose standard solution was prepared, and a control group was formed by replacing the sample with an equivalent volume of distilled water. Then, individual test tubes received 1 mL each of the sample solution, the standard solution, and the control solution. Afterward, 1 mL of 5% phenol solution was added to each test tube, followed by the gradual addition of 5 mL of concentrated sulfuric acid while keeping the mixture in an ice bath. The resulting mixture was incubated in a boiling water bath for a duration of 20 min. Once cooled to room temperature, colorimetric analysis was performed using a microplate reader set to a wavelength of 490 nm. The results were expressed as mg GE/g DE.

2.6.3. Protein content

The determination of protein content in the samples was carried out utilizing the Bradford method (Kayeye et al., 2026). This method involves the observation that free Coomassie Brilliant Blue G-250 displays a reddish-brown hue in acidic conditions, changing to blue upon protein binding; the blue color's intensity correlates positively with the protein concentration. The steps for processing the samples adhered to the same protocol used for the determination of polysaccharide content. At the same time, a set of standard crystalline bovine serum albumin solutions was prepared. After the staining solution was added, absorbance was assessed at 595 nm with a touch-screen microplate reader, within a timeframe of 2 to 5 min. By inserting the absorbance values from the samples into the standard curve created from the standard protein solutions, the protein content can be accurately calculated.The results were expressed as mg BSA/g DE.

2.6.4. Phlorotannins content

The content of phlorotannins in brown algae was evaluated through the DMBA method (He et al., 2026). To prepare the working reagent, equal amounts of 2% (v/v) DMBA solution and 6% (v/v) hydrochloric acid solution were first mixed with glacial acetic acid. Next, 25 μL of the sample solution combined with 125 μL of the working reagent were placed in a 96-well plate and mixed thoroughly. The mixture was then allowed to react in darkness for 60 min, after which absorbance was recorded at a wavelength of 520 nm with a microplate reader. The findings were reported as milligrams of phlorotannin equivalent (PGE) per gram of dry extract (DE).

2.7. Characterization of polyphenols from Undaria pinnatifida

Ultra-high performance liquid chromatography coupled with quadrupole time-of-flight mass spectrometry (UHPLC-QTOF, Agilent 1290 Infinity II-6546, USA) was utilized for conducting the analysis. The chromatographic column used for this purpose was Eclipse Plus C18 (1.8 μm, 50 × 2.1 mm). The mobile phase was prepared by mixing a 0.1% formic acid solution (designated as A) with methanol (designated as B). A gradient elution method was applied, beginning with an initial composition of 95% A, which was then reduced to 5% A over time. The sample was prepared by dissolving it in mobile phase B and subsequently filtered through a microporous membrane with a pore size of 0.22 μm. An injection volume of 2 μL was employed. The conditions for chromatography were established as follows: a flow rate of 0.3 mL/min, a column temperature maintained at 30 °C, and a detection wavelength set at 280 nm. For mass spectrometric analysis, a G6546 A system (Dual AJS ESI) was utilized, with parameters set to nitrogen atomization pressure at 45 psi, a drying gas flow of 11 L/min (heated to 350 °C), a sheath gas flow also at 11 L/min (at 350 °C), a capillary voltage set to 3.5 kV, fragmentation voltage fixed at 135 V, and nozzle voltage at 1000 V. Full scans were conducted in negative ion mode over the range of m/z 50–2000 at a frequency of 3 times/s, while MS/MS analysis applied three collision energies: 10, 20, and 40 eV for automatic fragmentation. Data acquisition was finalized in negative ion mode, with the MassHunter workstation software (B.06.00 qualitative version) being utilized for data processing and peak identification.

2.7.1. Compound identification criteria

Compound identification was performed according to the following criteria. First, the molecular formula of each compound was tentatively assigned by accurate mass measurement of the deprotonated molecular ion [M − H]⁻ (or molecular ion [M]•⁻ for highly conjugated phlorotannin structures). A mass accuracy threshold of <5 ppm was applied for tentative identification. For compounds in complex matrices where background interference was present, an extended threshold of ≤10 ppm was applied. Second, identification was supported by matching the observed MS/MS fragmentation patterns with those reported in the literature and publicly available databases (e.g., MassBank, GNPS, and Metlin). Based on the Metabolomics Standards Initiative (MSI) guidelines(Sumner et al., 2007), confidence levels were assigned as follows: Level 1 (confirmed by authentic standards), Level 2 (putatively annotated based on accurate precursor mass and diagnostic MS/MS fragments), and Level 3 (putatively characterized based on molecular formula only). All tentatively identified compounds were assigned MSI Level 2 confidence based on accurate precursor ion mass and diagnostic MS/MS fragments, in accordance with the MSI guidelines. This confidence level is well-established for putative annotation in untargeted metabolomics studies when multiple identification criteria are combined. Detailed identification data, including retention times, accurate masses, mass errors, molecular formulas, diagnostic MS/MS fragment ions, and assigned confidence levels for each compound, are summarized in Table 5, Table 6, respectively. Representative MS/MS spectra are provided in the Supplementary Materials (Figs. S1–S6).

Table 5.

Tentative identification of phenolic compounds in Undaria pinnatifida free polyphenols by UHPLC-QTOF-MS/MS in negative ion mode.

NO.	RT(min)	[M-H]- Real m/z	[M-H]-Theory m/z	Error (ppm)	Formula	DBE	MS/MS	Proposed compound	Confidence level
Phenolic acids and derivatives
1	4.741	121.0285	121.0290	−4.13	C7H6O2	5	77	Benzoic acid	Level 2
2	6.597	237.0036	237.0041	−2.1	C10H6O7	8	193, 149, 121, 105	Benzene polycarboxylic acid derivative	Level 2
Lipophilic phenolic derivatives
3	11.781	317.1388	317.1389	−0.32	C18H22O5	8	273, 229, 125	Phenolic dicarboxylic acid derivative	Level 2

RT: retention time in total ion chromatogram.

DBE: double bond equivalent.

[M-H]⁻: deprotonated molecular ion.

Error (ppm) = [(Real m/z – Theory m/z) / Theory m/z] × 10⁶.

Confidence levels were assigned according to the Metabolomics Standards Initiative (MSI) guidelines: Level 2, tentatively identified based on accurate precursor ion mass and diagnostic MS/MS fragments.

Table 6.

Tentative identification of phenolic compounds in Undaria pinnatifida bound polyphenols by UHPLC-QTOF-MS/MS in negative ion mode.

NO.	RT(min)	[M-H]− Real m/z	[M-H]− Theory m/z	Error (ppm)	Formula	DBE	MS/MS	Proposed compound	Confidence level
Low-molecular-weight phenolic acids
1	1.892	137.0234	137.0244	−7.4	C7H6O3	5	109, 93	dihydroxybenzoic acid derivative	Level 2
2	2.947	153.0546	153.0557	−7.3	C8H10O3	4	109, 81	dimethylphloroglucinol derivative	Level 2
3	4.376	167.0706	167.0714	−4.6	C9H12O3	4	123, 105	C3-alkylphloroglucinol	Level 2
4	5.905	195.1019	195.1027	−3.9	C11H16O3	4	151, 133	C5-alkylphloroglucinol	Level 2
5	6.451	181.0862	181.0870	−4.5	C10H14O3	4	137	C4-alkylphloroglucinol	Level 2
Alkylphloroglucinol derivatives
6	5.104	227.0919	227.0925	−2.6	C11H16O5	4	183, 165	carboxyalkylphloroglucinol derivative	Level 2
7	6.225	241.1074	241.1081	−3.1	C12H18O5	4	197, 181	acetoxyalkylphloroglucinol derivative	Level 2
8	6.314	357.1913	357.1919	−1.6	C18H30O7	4	321, 183, 127	highly oxidized C12-alkylphloroglucinol derivative	Level 2
9	6.969	383.2064	383.2075	−2.9	C20H32O7	5	365, 281, 219	highly oxidized C14-alkenylphloroglucinol	Level 2
10	7.028	321.1705	321.1707	−0.8	C18H26O5	6	303, 287	alkoxyphloroglucinol (C12)	Level 2
11	7.083	207.1017	207.1027	−4.7	C12H16O3	5	189, 163	alkylhydroxybenzoic acid derivative	Level 2
12	7.325	341.1962	341.1970	−2.2	C18H30O6	4	323, 297	carboxyhydroxy-C12-alkylphloroglucinol	Level 2
13	7.335	367.2118	367.2126	−2.2	C20H32O6	5	335, 306, 262	methoxyhydroxy-C14-alkylphloroglucinol	Level 2
14	7.384	209.1178	209.1183	−2.5	C12H18O3	4	191, 167, 149	C₆-alkylphloroglucinol	Level 2
15	7.560	323.1858	323.1864	−1.9	C18H28O5	5	305, 261, 227	hydroxy-C12-alkylphloroglucinol	Level 2
16	7.575	385.2227	385.2232	−1.2	C20H34O7	4	350, 295, 239	highly oxidized alkylphloroglucinol derivative	Level 2
17	7.797	283.1548	283.1551	−1.1	C15H24O5	4	265, 239, 221	C9-alkylphloroglucinol derivative	Level 2
18	7.958	325.2011	325.2012	−0.3	C18H30O5	4	289, 249, 209	dihydroxy-C12-alkylphloroglucinol	Level 2
19	8.092	369.2277	369.2282	−1.5	C20H34O6	4	351, 325, 241	carboxy-C14-alkylphloroglucinol	Level 2
20	8.180	267.1591	267.1602	−4.1	C15H24O4	4	249, 223, 205	hydroxycarboxy-C9-alkylphloroglucinol	Level 2
21	8.342	351.2171	351.2177	−1.7	C20H32O5	5	333, 295, 235	hydroxy-C14-alkenylphloroglucinol	Level 2
22	8.592	311.1851	311.1864	−4.2	C17H28O5	4	293, 267, 249	hydroxycarboxy-C11-alkylphloroglucinol	Level 2
23	8.656	353.2329	353.2333	−1.3	C20H34O5	4	335, 277, 235	hydroxy-C14-alkylphloroglucinol	Level 2
Phlorotannin derivatives
24	6.535	237.0035	237.0041	−2.4	C10H6O7	8	193, 149	highly oxidized phloroglucinol derivative	Level 2
25	9.896	436.0435	436.0436	−0.2	C22H12O10	17	401, 333, 249, 160, 125	eckol-type phlorotannin derivative*	Level 2
26	10.518	416.0736	416.0743	−1.7	C20H16O10	13	398, 342, 272, 237	highly oxidized phlorotannin dimer*	Level 2
27	11.453	585.2130	585.2125	0.9	C34H34O9	18	140	phlorotannin derivative	Level 2
28	11.663	317.1385	317.1389	−1.3	C18H22O5	8	273, 229	highly unsaturated C12-alkylphloroglucinol	Level 2

RT: retention time in total ion chromatogram.

DBE: double bond equivalent.

[M-H]⁻: deprotonated molecular ion.

Error (ppm) = [(Real m/z – Theory m/z) / Theory m/z] × 10⁶.

Confidence levels were assigned according to the Metabolomics Standards Initiative (MSI) guidelines: Level 2, tentatively identified based on accurate precursor ion mass and diagnostic MS/MS fragments.

MS/MS fragments are reported as nominal masses. Due to the complexity of the bound phenolic matrix, mass errors up to ±10 ppm were considered acceptable for tentative identification of compounds in complex matrices.

*Compounds 25 and 26 were detected as molecular ions [M]•⁻ instead of [M-H]⁻ due to their highly conjugated phlorotannin structures.

2.8. Antioxidant activity in vitro

2.8.1. DPPH free radical scavenging activity

The assessment of the scavenging capacity for DPPH free radicals was carried out in accordance with the procedures described in previous studies (El Hafez et al., 2022). Equal volumes (100 μL each) of the sample solution and 0.2 mM DPPH solution were combined and allowed to react in a dark environment for 30 min. Following this period, absorbance at a wavelength of 517 nm was recorded using a microplate reader. To mitigate any interference, a blank solution (consisting of 60% ethanol and DPPH) and a control solution (composed of anhydrous ethanol and the sample) were simultaneously prepared.

$${\mathrm{Y}}_{\mathrm{a}}\left(\%\right)=\left(1-\frac{{\mathrm{OD}}_{\text{sample}}-{\mathrm{OD}}_{\text{control}}}{{\mathrm{OD}}_{\text{blank}}}\right)\times 100\%$$ (1)

2.8.2. Hydroxyl radical scavenging ability

The scavenging capacity of hydroxyl radicals was assessed following a methodology outlined in previous research (He et al., 2022). In this process, 50 μL of the sample solution was mixed with 50 μL of a 9 mM FeSO₄ solution and 50 μL of a 10 mM salicylic acid ethanol solution in 96-well plates. This mixture was thoroughly combined, after which 50 μL of an 8.8 mM H₂O₂ solution was introduced. Following a reaction period of 20 min, the absorbance at 510 nm was measured and recorded as A_i. For the blank group (labeled as A_c), the sample solution was substituted with 60% ethanol, whereas distilled water replaced the H₂O₂ solution to act as the control group (identified as A_j). Ultimately, the clearance rate (Y_b) was calculated.

$$Y_b\left(\%\right)=\left(1-\frac{A_i-A_j}{A_c}\right)\times 100\%$$ (2)

2.8.3. Superoxide anion free radical scavenging ability

The assessment of the scavenging ability of superoxide anion radicals was performed according to established methodologies found in the literature (Lu et al., 2018). Specifically, 100 μL of Tris-HCl buffer was mixed with 100 μL of the sample solution and allowed to react in the dark at 25 °C for a duration of 10 min. Subsequently, 50 μL of pyrogallol solution was introduced and thoroughly mixed to evaluate the absorbance (A) at a wavelength of 420 nm. For the blank control group (A₀), the sample solution was replaced with 60% ethanol, while for the negative control group (A₁), the pyrogallol solution was substituted with distilled water. The clearance rate (Y_c) was subsequently calculated based on the obtained results.

$$Y_c\left(\%\right)=\left(1-\frac{A-A_1}{A_0}\right)\times 100\%$$ (3)

2.8.4. Total antioxidant capacity (FRAP method)

The experiment was conducted utilizing the iron ion reduction ability assay kit from Nanjing Jiancheng Bioengineering Institute. Reagents were added in accordance with the specified requirements, mixed thoroughly, and incubated at 37 °C for a duration of 3 to 5 min. Afterward, the optical density (OD) for each well was assessed using a microplate reader at a wavelength of 593 nm.

2.9. Statistical analysis

To guarantee data reliability, three parallel repetition groups were utilized for all experiments, and the outcomes were presented as the mean ± standard deviation. The differences in response values among the groups were analyzed statistically using ANOVA through SPSS 20.0 software. A P-value of less than 0.05 was deemed statistically significant. The comparison of phytochemical components between free and bound polyphenol extracts was analyzed using independent samples t-test, with the same significance level.

3. Results

3.1. Extraction process of free polyphenols

3.1.1. Single-factor experiments

Single-factor experimental results are shown in Fig. 1A–E. The effects of volume fraction of ethanol, ultrasonic time, power density of ultrasound, and ultrasonic temperature on FP content exhibited an initial increase followed by a decrease, whereas the effect of liquid-solid ratio showed an initial increase followed by a plateau. The maximum FP content was achieved at 50% ethanol (462.44 μg GAE/g DW), 70 min (431.51 μg GAE/g DW), 40 W/L (389.16 μg GAE/g DW), and 55 °C (444.12 μg GAE/g DW), respectively; further increases in these parameters led to a decline in FP content. When the liquid-solid ratio was increased to 40 mL/g, FP content reached 449.57 μg GAE/g DW, with no significant changes observed upon further increasing the ratio to 50 and 60 mL/g.

Fig. 1.

Single-factor effects on free (A-E) and bound (F-J) polyphenol contents from Undaria pinnatifida. Data are mean ± SD (n = 3).

3.1.2. Response surface optimization experiment

3.1.2.1. Fitting model

Response surface optimization was conducted using ethanol volume fraction (X₁), ultrasonic time (X₂), and ultrasonic power density (X₃) as independent variables, with FP content (Y₁) as the response value. The experimental design and results are shown in Table 1. Based on multiple regression analysis of the experimental data (Table 2), a second-order polynomial equation was established as follows:

Table 1.

Response surface design optimization test and results of free polyphenol content.

No.	Variable levels			Responses
	X1 (ethanol, %)	X2 (ultrasonic time, min)	X3 (ultrasonic power density, W/L)	Y1 (FP content, μg GAE/g DW)
1	0 (50)	−1 (60)	−1 (35)	693.59
2	0 (50)	+1 (80)	−1 (35)	619.24
3	0 (50)	+1 (80)	+1 (45)	841.93
4	+1 (60)	0 (70)	−1 (35)	534.30
5	−1 (40)	−1 (60)	0 (40)	760.51
6	0 (50)	0 (70)	0 (40)	1356.69
7	+1 (60)	+1 (80)	0 (40)	751.91
8	+1 (60)	−1 (60)	0 (40)	1123.87
9	−1 (40)	+1 (80)	0 (40)	774.76
10	−1 (40)	0 (70)	+1 (45)	693.33
11	0 (50)	0 (70)	0 (40)	1440.15
12	+1 (60)	0 (70)	+1 (45)	806.06
13	0 (50)	0 (70)	0 (40)	1305.55
14	0 (50)	0 (70)	0 (40)	1311.65
15	−1 (40)	0 (70)	−1 (35)	426.72
16	0 (50)	0 (70)	0 (40)	1253.64
17	0 (50)	−1 (60)	+1 (45)	950.08

Table 2.

Analysis of variance of free polyphenol content response surface regression equation.

Source	Sum of Squares	df	Mean Square	F-value	p-value	significant
Model	1,570,000	9	174,500	47.66	< 0.0001	**
X1	39,315.16	1	39,315.16	10.74	0.0135	*
X2	36,477.95	1	36,477.95	9.96	0.0160	*
X3	129,400	1	129,400	35.35	0.0006	**
X1X2	37,289.35	1	37,289.35	10.18	0.0152	*
X1X3	6.62	1	6.62	0.0018	0.9673	ns
X2X3	285.54	1	285.54	0.0780	0.7881	ns
X12	433,700	1	433,700	118.45	< 0.0001	**
X22	107,600	1	107,600	29.38	0.0010	*
X32	665,300	1	665,300	181.70	< 0.0001	**
Residual	25,629.96	7	3661.42
Lack of Fit	6081.73	3	2027.24	0.4148	0.7522	ns
Pure Error	19,548.24	4	4887.06
Cor Total	1,596,000	16

Note: P < 0.01 is extremely significant, expressed as **; p < 0.05 was significant, expressed as *; p > 0.05 was not significant, expressed as ns.

Y₁ = − 44,657.79 + 394.51 × ₁ + 272.05 × ₂ + 1307.96 × ₃–0.97X₁X₂ + 0.026X₁X₃–0.17X₂X₃–3.21 × ₁²–1.60 × ₂²–15.90 × ₃².

The coefficients of the linear terms indicated that all three factors positively contributed to FP content, with the order of influence being ultrasonic power density (X₃) > ethanol volume fraction (X₁) > ultrasonic time (X₂). All quadratic coefficients were negative, indicating that the model had a maximum point within the experimental range. The interaction coefficients between X₁X₂ and X₂X₃ were negative, while the X₁X₃ interaction was negligible. The model was statistically significant (p < 0.05), and the lack of fit was not significant (p > 0.05), indicating good adequacy of the model.

3.1.2.2. Optimization and validation of the model

Based on the regression model, the predicted optimal extraction conditions were 58% ethanol, 67 min ultrasonic time, 42 W/L power density, 40:1 mL/g liquid-solid ratio, and 55 °C ultrasonic temperature. Under these conditions, three repeated validation experiments were performed, and the experimental FP content was 1317.76 ± 114.30 μg GAE/g DW, which was close to the predicted value of 1301.610 μg GAE/g DW.

3.2. Extraction process of bound polyphenols

3.2.1. Single-factor experiments

Single-factor experimental results are shown in Fig. 1F–J. The effects of liquid-solid ratio, NaOH concentration, and ultrasonic time on BP content exhibited an initial increase followed by a decrease, whereas the effects of ultrasonic temperature and power density of ultrasound showed a continuous increase within the tested range. BP content reached peaks at 60 mL/g (1.72 mg GAE/g DW), 0.16 mol/L (2.70 mg GAE/g DW), and 90 min (3.69 mg GAE/g DW), respectively; ultrasonic temperature and ultrasonic power density increased to 4.86 mg GAE/g DW at 80 °C and 3.58 mg GAE/g DW at 50 W/L, respectively.

3.2.2. Response surface optimization experiment

3.2.2.1. Fitting model

NaOH concentration (Z₁), liquid-solid ratio (Z₂), and ultrasonic time (Z₃) were selected as independent variables, with BP content (Y₂) as the response value. The experimental design and results are shown in Table 3. Based on the analysis of variance (ANOVA) results in Table 4, a second-order polynomial equation was established as follows:

Table 3.

Response surface design optimization test and results of bound polyphenol content.

No.	Variable levels			Responses
	Z1 (NaOH concentration, mol/L)	Z2 (liquid-solid ratio, mL/g)	Z3 (ultrasonic time, min)	Y2 (BP content, mg GAE/g DW)
1	0 (0.16)	0 (60)	0 (90)	4.67
2	−1 (0.12)	0 (60)	−1 (80)	3.36
3	+1 (0.20)	+1 (70)	0 (90)	3.87
4	0 (0.16)	+1 (70)	−1 (80)	3.30
5	+1 (0.20)	0 (60)	−1 (80)	3.57
6	0 (0.16)	−1 (50)	−1 (80)	3.36
7	0 (0.16)	+1 (70)	+1 (100)	3.55
8	+1 (0.20)	0 (60)	+1 (100)	3.76
9	−1 (0.12)	0 (60)	+1 (100)	3.40
10	+1 (0.20)	−1 (50)	0 (90)	3.75
11	0 (0.16)	0 (60)	0 (90)	4.79
12	0 (0.16)	−1 (50)	+1 (100)	3.65
13	−1 (0.12)	+1 (70)	0 (90)	3.32
14	−1 (0.12)	−1 (50)	0 (90)	3.87
15	0 (0.16)	0 (60)	0 (90)	4.84
16	0 (0.16)	0 (60)	0 (90)	4.72
17	0 (0.16)	0 (60)	0 (90)	4.67

Table 4.

Analysis of variance of bound polyphenol content response surface regression equation.

Source	Sum of Squares	df	Mean Square	F-value	p-value	significant
Model	5.55	9	0.6168	103.11	< 0.0001	**
Z1	0.1251	1	0.1251	20.91	0.0026	**
Z2	0.0452	1	0.0452	7.56	0.0285	*
Z3	0.0713	1	0.0713	11.93	0.0106	*
Z1Z2	0.1119	1	0.1119	18.71	0.0035	**
Z1Z3	0.005	1	0.005	0.8368	0.3907	ns
Z2Z3	0.0005	1	0.0005	0.0846	0.7795	ns
Z12	1.05	1	1.05	176.2	< 0.0001	**
Z22	1.31	1	1.31	218.19	< 0.0001	**
Z32	2.31	1	2.31	385.51	< 0.0001	**
Residual	0.0419	7	0.006
Lack of Fit	0.0235	3	0.0078	1.71	0.3024	ns
Pure Error	0.0184	4	0.0046
Cor Total	5.59	16

Note: P < 0.01 is extremely significant, expressed as **; p < 0.05 was significant, expressed as *; p > 0.05 was not significant, expressed as ns.

Y₂ = −79.45 + 70.14Z₁ + 0.60Z₂ + 1.33Z₃ + 0.42Z₁Z₂ + 0.09Z₁Z₃–0.000112Z₂Z₃–312.70Z₁²–0.0056Z₂²–0.007Z₃².

All linear coefficients were positive, while all quadratic coefficients were negative. The quadratic coefficient of NaOH concentration (Z₁) was the largest in magnitude (−312.70), followed by ultrasonic time (Z₃) and liquid-solid ratio (Z₂). The model was highly significant (p < 0.0001) with a non-significant lack of fit (p = 0.3024). All three quadratic terms reached highly significant levels (p < 0.0001), with F values of 176.2, 218.19, and 385.51, respectively.

3.2.2.2. Optimization and validation of the model

The optimal extraction conditions predicted by the model were 0.176 mol/L NaOH, 68:1 mL/g liquid-solid ratio, and 97 min ultrasonic time. Validation experiments were conducted in triplicate under these conditions, and the experimental BP content was 4.46 ± 0.52 mg GAE/g DW, which corresponded well with the predicted value of 4.28 mg GAE/g DW.

3.3. Phytochemical composition of free and bound polyphenol extracts

The free polysaccharide content (201.46 ± 4.37 mg GE/g DE) and free total flavonoid content (81.78 ± 2.62 mg RE/g DE) were both significantly higher than their respective bound forms (121.46 ± 2.20 mg GE/g DE and 45.88 ± 0.85 mg RE/g DE) (p < 0.001). In contrast, the bound protein content (214.12 ± 3.16 mg BSA/g DE) and bound polyphenol content (295.67 ± 2.52 mg GAE/g DE) were both significantly higher than their respective free forms (175.94 ± 5.17 mg BSA/g DE and 160.51 ± 3.48 mg GAE/g DE) (p < 0.001). The free phlorotannin content (23.62 ± 0.91 mg PGE/g DE) was slightly higher than the bound form (21.95 ± 0.43 mg PGE/g DE) (p = 0.045).

3.4. Identification of phenolic compounds in free and bound polyphenols

The tentative identification results of phenolic compounds in free and bound polyphenol extracts are shown in Table 5, Table 6. A total of 3 compounds were tentatively identified in the FP extract, including 2 phenolic acids and their derivatives (benzoic acid, benzene polycarboxylic acid derivative) and 1 lipophilic phenolic derivative (phenolic dicarboxylic acid derivative). A total of 28 compounds were tentatively identified in the BP extract, which were classified into three major groups: low-molecular-weight phenolic acids (5 compounds), alkylphloroglucinol derivatives (18 compounds), and phlorotannin derivatives (5 compounds). Among these, the alkylphloroglucinol derivatives and phlorotannin derivatives (e.g., eckol-type phlorotannin derivative, highly oxidized phlorotannin dimer) were characteristic phenolic compounds of brown algae and were not detected in the free fraction.

3.5. In vitro antioxidant activities of free and bound polyphenols

The in vitro antioxidant activities of FP, BP, and tea polyphenols (TP, positive control) were evaluated using four different methods: DPPH, hydroxyl radical, superoxide anion scavenging, and Fe³⁺ reducing capacity. The results are shown in Fig. 2. At a concentration of 3.5 mg/mL, the DPPH scavenging assay (Fig. 2A) showed that the scavenging activity of BP (94.98%) was higher than that of FP (41.52%) and was comparable to that of TP. In the hydroxyl radical scavenging assay (Fig. 2B), FP showed a higher scavenging rate (59.95%) than BP (45.48%). In the superoxide anion scavenging assay (Fig. 2C), the scavenging rates of FP, BP, and TP were 82.84%, 80.92%, and 85.82%, respectively. However, in the Fe³⁺ reducing capacity assay (Fig. 2D), the reducing capacities of FP and BP were 0.25 mM and 0.41 mM, respectively, which were lower than that of TP (11.89 mM).

Fig. 2.

Evaluation of antioxidant capacity of polyphenols at different concentrations in vitro. A. DPPH radical scavenging capacity; B. Hydroxyl radical scavenging capacity; C. Superoxide anion radical scavenging capacity. D. Fe³⁺ reducing capacity.

4. Discussion

4.1. Extraction processes of free and bound polyphenols

Through single-factor experiments and response surface optimization, the optimal extraction conditions for FP were determined as: ethanol volume fraction of 58%, ultrasonic time of 67 min, ultrasonic power density of 42 W/L, ultrasonic temperature of 55 °C, and liquid-to-solid ratio of 40:1 mL/g. The moderate polarity of the ethanol-water system is considered to be capable of matching the polarity of various phenolic compounds, thereby promoting their dissolution (Huo et al., 2025). Moderate ultrasonic treatment enhanced cell wall disruption and mass transfer through cavitation effects, while excessive ultrasonic energy may lead to oxidative degradation of polyphenols(Dong et al., 2020; Liu et al., 2022). An increase in temperature to 55 °C reduced solvent viscosity and improved diffusion coefficients, thereby enhancing mass transfer; however, further increases in temperature may accelerate thermal degradation of polyphenols(Huo et al., 2025).

Based on single-factor experiments and response surface optimization, the optimal extraction conditions for BP were determined as: NaOH concentration of 0.176 mol/L, liquid-to-solid ratio of 68:1 mL/g, ultrasonic time of 97 min, ultrasonic temperature of 80 °C, and ultrasonic power density of 50 W/L. The alkaline environment is considered to be capable of cleaving ester and ether bonds between polyphenols and cell wall polysaccharides or proteins, thereby facilitating the release of BP(Stefanello et al., 2018; Zhang et al., 2021). However, excessive NaOH concentration may lead to oxidative polymerization and degradation of polyphenols(Huang et al., 2025). A liquid-to-solid ratio of 68:1 mL/g ensured sufficient solvent volume for mass transfer, while an ultrasonic time of 97 min enhanced cavitation-mediated cell wall disruption(Zhong et al., 2022); however, further prolongation may cause thermal degradation(Bhagya Raj & Dash, 2020; Kumar et al., 2021). In addition, the optimal ultrasonic temperature of 80 °C is considered to be capable of reducing solvent viscosity and improving diffusion coefficients, thereby enhancing alkaline hydrolysis efficiency(Huo et al., 2025).

The simultaneous extraction of FP and BP promotes the high-value utilization of polyphenolic resources in Undaria pinnatifida, providing a foundation for subsequent studies.

4.2. Phytochemical composition and phenolic identification of free and bound polyphenols

Significant differences were observed in the phytochemical composition of FP and BP extracts, reflecting their different forms of existence in Undaria pinnatifida. Free polysaccharides and free flavonoids were more readily released during ethanol extraction, whereas bound proteins and bound polyphenols were more tightly associated with cell wall matrices, which could be effectively released by the alkaline hydrolysis employed in this study. Bound polyphenols accounted for approximately 65% of the total extractable polyphenols, highlighting that the majority of polyphenolic resources in UP exist in the bound form.

Tentative identification by UHPLC-QTOF-MS/MS revealed a striking difference in the composition of phenolic compounds between FP and BP. Only 3 phenolic compounds were tentatively identified in the FP extract, all of which were common phenolic acids. In contrast, 28 phenolic compounds were tentatively identified in the BP extract, including eckol-type phlorotannin derivatives and highly oxidized phlorotannin dimers. These differences indicate that ethanol extraction provides a limited diversity of phenolic compounds, whereas alkaline extraction enables the recovery of a broader range of polyphenols, particularly those characteristic of brown algae. Therefore, the alkaline extraction process established in this study enables the recovery of previously overlooked polyphenolic resources, providing a foundation for further investigation of these compounds in UP.

4.3. In vitro antioxidant activities of free and bound polyphenols

The in vitro antioxidant activities of FP and BP were evaluated using four different assays. In the DPPH scavenging assay, BP exhibited a higher scavenging activity than FP, even surpassing the level of TP at a polyphenol concentration of 3.5 mg/mL, which may be attributed to the richer phenolic composition of BP, particularly the characteristic brown algal polyphenols, such as eckol-type phlorotannin derivatives, endowed it with stronger antioxidant capacity. However, in the hydroxyl radical scavenging assay, FP outperformed BP. The contrasting results of FP and BP across different assays suggest that they exhibit different antioxidant capacities depending on the assay employed, possibly related to differences in their chemical composition. In the superoxide anion scavenging assay, FP and BP exhibited comparable scavenging rates at a polyphenol concentration of 3.5 mg/mL, both approaching the level of TP, indicating that both forms of polyphenols are effective in scavenging superoxide anions. Notably, in the Fe³⁺ reducing capacity assay, both FP and BP exhibited low reducing capacities, which were significantly lower than that of TP. These results indicate that FP and BP exhibit distinct antioxidant capacities across different assays, which may be attributed to their different phenolic compositions.

4.4. Limitations and future directions

Despite the aforementioned results, several limitations should be acknowledged. First, the compounds in free and bound extracts were annotated as tentative identifications (MSI Level 2) based on comparison of precursor ion masses and MS/MS fragmentation patterns with spectral databases and published literature, without confirmation using authentic standards or quantitative determination. Second, the antioxidant activity was evaluated exclusively using chemical in vitro assays (e.g., DPPH, hydroxyl radical, superoxide anion scavenging, and Fe³⁺ reducing capacity); the results only reflect the chemical antioxidant capacity and cannot be extrapolated to in vivo biological activity, health benefits, or nutritional functionality. Third, the ultrasound-assisted extraction parameters were optimized at the laboratory scale, and the feasibility of industrial scale-up requires systematic evaluation.

Nevertheless, bound polyphenols derived from Undaria pinnatifida show promising potential as natural food antioxidants, demonstrating strong DPPH radical scavenging capacity and rich phenolic composition. Future research will focus on: (i) confirmation and quantitative determination of major metabolites using authentic standards; (ii) in vivo antioxidant activity evaluation and toxicological safety assessment; and (iii) systematic investigation of the stability, sensory compatibility, and optimal dosage of the extracts in actual food matrices.

5. Conclusion

In this study, FP and BP were extracted from Undaria pinnatifida, with the extraction processes optimized through single-factor experiments and response surface methodology. Phytochemical composition analysis revealed that the majority of polyphenols in UP existed in the bound form, accounting for approximately 65% of the total. Tentative Phenolic compound identification further showed that BP contained a substantially higher diversity of phenolic compounds than FP, with characteristic brown algal polyphenols, such as eckol-type phlorotannin derivatives, exclusively detected in BP. The in vitro antioxidant activities of FP, BP, and TP were evaluated using four assays: DPPH, hydroxyl radical, superoxide anion, and Fe³⁺ reducing capacity. BP exhibited stronger DPPH scavenging capacity than FP, and both FP and BP showed comparable superoxide anion scavenging capacities. However, both FP and BP exhibited low Fe³⁺ reducing capacities, which were significantly lower than that of TP, and BP showed weaker hydroxyl radical scavenging capacity than FP. These results indicate that the antioxidant activity of FP and BP is method-dependent. Therefore, alkaline extraction is essential for accessing the polyphenolic resources in UP that are otherwise inaccessible through ethanol extraction, and BP possess potential as natural antioxidants. Further investigation is required to elucidate their specific mechanisms of action and in vivo bioavailability.

CRediT authorship contribution statement

Qiuling Lao: Writing – original draft, Validation, Software, Project administration, Conceptualization. Jianing Zheng: Project administration, Formal analysis. Weiqin Yang: Investigation, Data curation. Chunhui Luo: Writing – review & editing, Formal analysis. Lei Guo: Writing – review & editing, Project administration, Funding acquisition.

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.

Appendix A. Supplementary data

Supplementary material

Data availability

Data will be made available on request.