Main characteristic metabolites with xanthine oxidase inhibitory activities from Yunnan Coffea arabica husks

Abstract

Coffee “husks”, as primary by-products whose inadequate disposal causes environmental pollution, are rich in bioactive ingredients. The in-depth exploration and practical application of these by-products remain insufficient. A comprehensive investigation was performed with the combination of LC-MSⁿ-IT-TOF analysis, antioxidant and xanthine oxidase (XO) inhibitory assays. A total of 55 main characteristic compounds were identified and their detailed mass spectrometry cleavage patterns were elucidated. Chlorogenic acid analogs as key bioactive constituents comprised over 66% of the identified phenols, which spontaneously bind to XO, with binding energy ranging from −4.46 to −3.32 kcal/mol. The antioxidant activities of di-substituted chlorogenic acids were significantly higher than those of mono-substituted molecules. Flavonoids from the “husks” with XO inhibitory activities were attributed to the flavone skeleton, in which the presence of a stable chromone group was vital. These results provided novel perspectives for discovery of XO inhibitors from widely discarded natural by-products like coffee “husks”.

Highlights

• •The chemical profiles of coffee “husks” with different primary treatment were elucidated.
• •The in vitro antioxidant and XO inhibitory activities of 11 coffee “husks” were determined.
• •A detailed mass spectrometry cleavage patterns of main characteristic compounds were displayed.
• •The relationship between in vitro bioactivities and molecular docking results of chlorogenic acid analogs was discussed.

1. Introduction

With an estimated production of 10 million metric tons per year, coffee is one of the most traded commodities in the world (ICO, 2023). Coffea arabica contributes approximately 60% of global coffee production (Salojärvi et al., 2024; Scalabrin et al., 2024). Yunnan Province is considered an ideal region for C. arabica cultivation in China, contributing approximately 99% of the nation's total coffee output (Bi, Yu, et al., 2023). The production of coffee by-products will increase with the expansion of coffee bean processing, such as coffee “husks”, which account for about 45–50% of the coffee cherry and have an annual yield of about 160,000 tons (Li et al., 2023). Coffee “husks”, consisting of pulp, mucilage, dried skin and parchment, are largely discarded into the environment, potentially causing soil damage and pollution through inadequate disposal (Oliveira et al., 2024; Qin et al., 2024). Despite the high yield of these by-products in coffee production, the utilization prospects of coffee “husks” are seriously underestimated.

Coffee “husks” are abundant in phenolic substances such as chlorogenic acids, p-coumaric acid and ferulic acid, which are strongly correlated with antioxidant activity and have been extensively validated as potent free radical scavengers (Abreu et al., 2024; Silva et al., 2024). Studies have shown that these phenolic compounds exhibited anti-inflammatory, antibacterial, anticancer, cardiovascular-protective and xanthine oxidase (XO) inhibitory activities (Freitas et al., 2024; Grzelczyk et al., 2024; Makiso et al., 2023; Mehmood et al., 2022; Rodrigues da Silva et al., 2022). Among them, chlorogenic acid analogs are compounds formed by the esterification of quinic acid with trans-cinnamic acids (mainly caffeic, ferulic and coumaric acids), which represent one of the most abundant and widely distributed classes of phenols in coffee (Asamenew et al., 2019; Nemzer et al., 2021). In addition to chlorogenic acid analogs, flavonoids, alkaloids, polysaccharides and diterpenoids are critical to the potential exploitation of coffee “husks” (Arya et al., 2022; Nzekoue et al., 2020; Tripathi et al., 2025). Quercetin, as a representative flavonoid in coffee “husks”, exhibited antifungal, anticarcinogenic and anti-hyperuricemia effects (Batiha et al., 2020; Li et al., 2024; Nzekoue et al., 2020). Diterpenoids in coffee displayed hepatoprotective and anticancer activities (Al-Romaima et al., 2024). Hence, coffee “husks”, rich in valuable natural compounds, could serve as bioactive sources with eco-friendly characteristics.

Our preliminary research indicated that phenolic compounds were the primary constituents in Yunnan C. arabica, and different primary processing methods influenced the composition and content of its secondary metabolites, thereby affecting coffee quality (Wan, Li, et al., 2024; Wan, Wang, et al., 2024; Wang, Li, et al., 2025; Wang, Wang, et al., 2025). Phenolic compounds such as chlorogenic acid analogs and flavonoids were reported to possess potential for inhibiting XO activity (Li et al., 2024; Mehmood et al., 2022). Therefore, investigating the non-volatile components in coffee “husks” and identifying bioactive phenolic substances within them hold significant potential for developing natural and low-toxicity XO inhibitors.

To further understand the primary bioactive metabolites in C. arabica “husks” from Yunnan, the antioxidant and XO inhibitory activities of 11 coffee “husks” with three primary processing treatments were determined in this study, and four samples with the best potential were screened out (sun-exposed: K4 and K6, washed: K2 and K8). A detailed investigation involving LC-MSⁿ-IT-TOF exploration, antioxidant (DPPH•, ABTS^•+, FRAP) and XO inhibitory activities was performed. More importantly, the relationship between molecular docking within specific chlorogenic acid analogs and their antioxidant, XO inhibitory capacities was discussed. This study is expected to provide new insights into the development and utilization of the C. arabica “husks” from Yunnan.

2. Materials and methods

2.1. Samples

Coffee (C. arabica) from different primary processing treatments (sun-exposed, washed and honey-treated) were collected from different latitudes coffee plantations (low latitudes: Pu'er, 22.33°N, 99.58°E; mid-latitudes: Baoshan, 25.11°N, 99.17°E; high latitudes: Dali, 25.83°N, 100.58°E) in Yunnan Province. Coffee beans of uniform maturity were manually selected based on appearance, color and hardness. Subsequently, the peel was manually removed in a clean environment, and any adhering pulp was thoroughly cleaned off. Finally, the “husks” were stored in a dry and shade-proof place for use. Eleven coffee “husks” samples were analyzed, and the specific sample information was shown in Table S1. 100.00 g of dried coffee “husks” powder was weighed and mixed with 300 mL of 70% ethanol-water solution at a solid-to-liquid ratio of 3:1 (mL/g). Ultrasonic extraction was performed at room temperature (25 ± 2 °C) for 30 min per cycle, and the process was repeated three times. After each extraction, the mixture was filtered through medium-speed qualitative filter paper. The three filtrates were combined and concentrated under reduced pressure at 40 °C until dry to yield the crude extract. This crude extract was then dissolved and diluted into sample solutions of varying concentrations for use in antioxidant activity assays (DPPH•, ABTS^•+, FRAP) and XO inhibitory activity evaluation. Corresponding sample solutions were also prepared for LC-MSⁿ-IT-TOF analysis. Compounds requiring testing were first dissolved in a 70% methanol aqueous solution. The dissolution process was conducted at room temperature with ultrasonic treatment until the solution became clear and transparent. After complete dissolution, the compounds were diluted to the target concentration using PBS (pH = 7.6). All stock solutions were stored in the dark at 4 °C and used within 24 h.

2.2. Chemicals and reagents

1, 1-Diphenyl-2-picrylhydrazyl (DPPH), 2, 2′-azino-bis (3-ethylbenzothiazoline-6-sulfonic acid) (ABTS), 2, 4, 6-tris (2-pyridyl)-s-triazine (TPTZ), 6-hydroxy-2, 5, 7, 8-tetramethylchromane-2-carboxylic acid (Trolox), xanthine oxidase (XO), xanthine and allopurinol were purchased from Aladdin (Shanghai, China). Analytical reagent of acetic acid, hydrochloric acid, absolute ethanol, dimethyl sulfoxide and methanol were purchased from Titan Technology Co., Ltd. (Shanghai, China). Ferric chloride hexahydrate, sodium acetate trihydrate, ferrous sulfate heptahydrate, sodium hydroxide and potassium persulfate were obtained from Solarbio Science & Technology Co., Ltd. (Beijing, China). Potassium dihydrogen phosphate, ethylene diamine tetraacetic acid, and dipotassium hydrogenphosphate were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China).

2.3. Main compound composition by LC-MSⁿ-IT-TOF

The main compounds composition analyses were performed on high-performance liquid chromatography system (HPLC, 1290 Infinity, Agilent Technologies, Inc., USA) equipped with an Agilent ZORBAX SB-C18 column (4.60 × 250 mm, 5 μm). Mobile phase A consisted of water, while mobile phase B was 100% methanol, with a flow rate of 1.00 mL/min. Chromatographic analysis was conducted using a gradient elution program with phases A and B. The elution conditions were as follows: 0–30 min, 10%-100% B; 30–35 min, 100% B. The injection volume for each analysis was 2.00 μL. Mass spectrometry analysis was performed in automatic mode using multi-stage mass scanning. Mass spectrometry parameters were set as follows: ion source voltage at −3.50 kV, detection voltage at 1.60 kV, full scan mass range set to 100–1000 Da, drying gas pressure at 100.00 kPa, nebulizer gas (N₂) flow rate at 0.50 L/min, CDL temperature at 200.00 °C, with both the ion source heater block and the ion source heater maintained at 200 °C, equipment temperature: 40.00 °C, relative collision energy: 50%, ion trap time: 30 ms, precursor ion isolation width: 3.00 Da, ion selection time: 20 ms, collision-induced dissociation time: 30 ms, collision energy: 50%, collision gas: 50%, scan range: m/z 100–1000. Molecular formula prediction was performed using Qualitative MassHunter Analysis software.

2.4. Determination of antioxidant activity in vitro

DPPH• and ABTS^•+ radical scavenging abilities and ferric reducing antioxidant power (FRAP) assays were performed in vitro according to the previous reports (Wan, Li, et al., 2024; Wan, Wang, et al., 2024). For DPPH• radical scavenging capacity assay, 0.10 mL of sample extract was mixed with 3.90 mL of freshly made DPPH• solution (0.075 mmol/L) and incubated at room temperature for 30 min in darkness. The absorbance was measured at 515 nm, and Trolox was used as a reference standard (0.00, 0.03, 0.06, 0.09, 0.12, 0.15, 0.18, 0.21, 0.24, 0.27 and 0.30 mg/mL). The inhibition rate was calculated using the following formula:

$$\text{Inhibition rate}\left(\%\right)=\left[\left(\mathrm{Ab}-\mathrm{As}\right)/\mathrm{Ab}\right]\times 100$$

where As and Ab represent the absorbance of samples and blank at 515 nm, respectively. The DPPH• scavenging activity was expressed as milligrams of Trolox equivalents per gram sample extracts (mg TE/g Ex) using the linear regression value obtained from the Trolox calibration curve.

For ABTS^•+ radical scavenging assay, the ABTS (7.00 mmol/L) and K₂S₂O₈ (2.50 mmol/L) were mixed in equal volume and stored in the dark for 12–16 h at room temperature to obtain ABTS^•+ base solution. The ABTS^•+ working solution was obtained by diluting the ABTS^•+ base solution with absolute ethanol to obtain an absorbance value of 0.70 ± 0.02 at 734 nm. The 0.20 mL sample was mixed with 3.80 mL ABTS^•+ working solution, and the absorbance was measured at 734 nm after 6 min reaction in the dark at room temperature. Trolox was used as a reference standard (0.00, 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09 and 0.10 mg/mL). The inhibition rate was calculated by the following formula:

$$\text{Inhibition rate}\left(\%\right)=\left[\left(\mathit{Ab}-\mathit{As}\right)/\mathit{Ab}\right]\times 100$$

where As and Ab represent the absorbance of samples and blank at 734 nm, respectively. The ABTS^•+ scavenging activity was expressed as milligrams of Trolox equivalents per gram sample extracts (mg TE/g Ex) using the linear regression value obtained from the Trolox calibration curve.

For FRAP assay, 10.00 mmol/L TPTZ was prepared with 40.00 mmol/L HCl. The FRAP working solution was prepared by mixing 5.00 mL TPTZ (10.00 mmol/L), 5.00 mL FeCl₃ (20.00 mmol/L) and 50.00 mL CH₃COONa buffer solution (pH = 3.6). 0.10 mL sample, 0.30 mL water and 3.00 mL FRAP working solution were mixed and placed in a water bath at 37 °C for 30 min. The absorbance was measured at 595 nm. The standard curve was made with FeSO₄ solution (0.00, 0.20, 0.40, 0.60, 0.80, 1.00, 1.20, 1.40 and 1.60 × 10⁻⁴ mmol). The reducing power was calculated as follows:

$$\mathrm{As}-\mathrm{Ab}=0.6572\times c\left({\mathit{Fe}}^{2+}\right)-0.0131\left(R^2=0.9970\right)$$

where As and Ab represent the absorbance of samples and blank at 595 nm, respectively. The FRAP of sample was expressed as mmol of FeSO₄ equivalents per gram sample extracts (mmol Fe²⁺/g Ex).

2.5. Determination of XO inhibitory activity in vitro

The XO inhibitory assay was performed as follows (Xie et al., 2020): The sample solution (0.04 mL) was mixed with freshly prepared xanthine oxidase solution (0.08 mL, final concentration 8.00 U/L) and incubated at 20 °C for 35 min. Then the substrate xanthine solution (0.08 mL, final concentration 0.20 mmol/L) was added and incubated at 37 °C for 15 min. The absorbance was measured at 295 nm, with allopurinol (0.50 mmol/L) serving as the positive control. The inhibition rate was calculated using the following formula:

$$\text{Inhibition rate}I\left(\%\right)=1-\left[\left(\mathrm{As}-\mathrm{A}{\mathrm{b}}_1\right)/\left(\text{An}-\mathrm{A}{\mathrm{b}}_2\right)\right]\times 100$$

As and Ab₁ represent the absorbance of samples and sample blanks, respectively. An and Ab₂ represent the absorbance of normal group and blanks, respectively.

2.6. Molecular docking of XO inhibitory activity

The 3D crystal structure of XO (PDB ID: 1FIQ) was obtained from the Protein Data Bank (https://www.rcsb.org/), and the protein macromolecules were processed (including removal of water molecules and cofactors) using PyMOL software. The small molecule structure of the target ligand was downloaded from PubChem (https://pubchem.ncbi.nlm.nih.gov/) and energy-minimized using Chem3D. Preprocessing such as hydrogen addition to protein macromolecules and to ligand small molecules was performed using AutoDockTools (v1.5.6). After saving them as PDBQT files, the molecular docking analysis was continued using AutoDock. The optimal conformation with the lowest binding energy was exported and saved as PDBQT files, and the docking results were visualized using Discovery Studio 2021.

2.7. Statistical analysis

Antioxidant data were analyzed using Microsoft Office Excel. The correlation heatmap and Venn diagram were drawn using R 4.4.1. Statistical analysis was performed using IBM SPSS Statistics 27 software to assess significance (p < 0.05), with differences between groups evaluated by one-way analysis of variance (ANOVA). When ANOVA results indicated significant differences, Tukey's post-hoc multiple comparison test was further applied. All chemical structures were drawn using ChemDraw software (version 20.0). The results of the antioxidant assays were visualized and a model was constructed to calculate the half maximal inhibitory concentration (IC₅₀) using GraphPad Prism 9.5.0. The molecular docking was carried out using AutoDock and the docking results were visualized by PyMOL and Discovery Studio.

3. Results and discussion

3.1. Antioxidant and XO inhibitory activities of coffee “husks” in vitro

Antioxidant and XO inhibitory activities showed a positive correlation in studies in which the composition of the substance was rich in both flavonoids and phenolic compounds (Kumar et al., 2024; Orhan & Deniz, 2021). To accurately identify the most active coffee “husks” and the key bioactive compounds in them, antioxidant and XO inhibitory activities were determined for 11 coffee “husks” subjected to three different primary processing methods in the main production areas of Yunnan.

XO inhibitory tests involving 11 coffee “husks” were carried out in vitro, and only K2 (XO inhibition rate, 29.14%) and K4 (XO inhibition rate, 19.71%) showed XO inhibitory activity at the concentration of 4.0 mg/mL (Table S2). Within 11 coffee “husks”, samples K2, K4, K6 and K8 exhibited outstanding ABTS^•+ and DPPH• radical scavenging capacities and FRAP antioxidant activities (Fig. 1). Among them, the washed sample K2 displayed the strongest activity across all assays: DPPH• (77.48 mg TE/g Ex), ABTS^•+ (143.16 mg TE/g Ex), FRAP (1.69 mmol Fe²⁺/g Ex) and XO inhibition rate (29.14%) (Table S2 and S3). The chemical profiles of four samples with the best antioxidant effects were further investigated.

Fig. 1.

Antioxidant activity of 11 coffee “husks” samples with different primary treatments. Note: Different letters in the same column indicate significant differences (p < 0.05).

3.2. Main compound composition of coffee “husks” identified by LC-MSⁿ-IT-TOF

A total of 55 compounds were detected and identified by LC-MSⁿ-IT-TOF, including 4 amino acids, 3 organic acids, 7 benzoic acids and derivatives, 9 cinnamic acids and derivatives, 5 other phenols, 11 flavonoids and derivatives, 8 terpenoids, 3 steroids, 2 fatty acids and 3 others (Table 1). In negative ion mode, the base peak chromatograms (BPC) and their compound assignment of the four representative samples were presented in Fig. 2.

Table 1.

MS fragmentation data of compounds in coffee “husks” based on LC-MSⁿ-IT-TOF analysis.

No.	Molecular formula	Fragment ions (m/z)	Adducts and Fragment ions
Amino acids
1	C4H7NO4	132.030[M-H] −	115.004[M-H2O] −, 88.041[M-H-CO2] −
4	C5H7NO3	128.036[M-H] −	111.008[M-H2O] −, 84.045[M-H-CO2] −
8	C6H13NO2	130.087[M-H] −	147.029[M-H + H2O] −, 115.003[M-OH] −, 87.009[M-CO2] −
9	C9H11NO2	164.072[M-H] −	147.044[M-H2O] −, 129.018[M-2H2O] −
Organic acids
3	C7H12O6	191.056[M-H] −	173.043[M-H-H2O] −,146.044[M-H-COOH] −, 127.038[M-2H-COOH-H2O] −
5	C6H8O7	191.020[M-H] −	173.008[[M-H-H2O] −
7	C4H4O4	115.004[M-H] −	147.029[M + OCH3] −, 71.014[M- COOOH]] −
Benzoic acids and derivatives
6	C7H6O2	121.030[M-H] −	95.050, 77.039, 69.035, 58.030
10	C8H8O4	167.035[M-H] −	137.024[M-OCH3] −, 123.045[M-COOH] −
12	C7H6O4	153.019[M-H] −	109.030[M-COOH] −, 120.022[M-2OH] −,135.045[M-H-H2O]−
26	C7H8O2	123.045[M-H] −	159.064[M-H-2H2O] −, 89.024[M-H-2OH] −
27	C9H10O3	165.056[M-H] −	147.045[M-H-H2O] −, 119.049[M-2H-CO2] −, 103.055[M-H-OCH2CH3-OH] −
36	C7H6O3	137.024[M-H] −	109.030[M-CHO] −, 119.947[M-H2O] −, 119.035[M-H-H2O] −
39	C8H8O3	151.040[M-H] −	119.036[M-H-OCH3] −, 107.05[M-COOH] −
Cinnamic acids and derivatives
11	C16H18O9	353.089[M-H] −	377.085[M + H + Na] +, 179.035, 191.056,161.025, 145.029, 135.045
17	C16H18O9	353.089[M-H] −	377.085[M + H + Na] +, 179.035, 191.056, 161.025, 145.029, 135.045
18	C16H18O9	353.089[M-H] −	377.085[M + H + Na] +, 179.035, 191.056, 161.025, 145.029, 135.045
22	C17H20O9	367.102[M-H] −	191.056, 173.046, 191.056
28	C25H24O12	515.117[M-H] −	353.087, 173.046, 191.056, 179.046, 303.050, 135.045, 163.040
29	C25H24O12	515.117[M-H] −	353.087, 173.046, 191.056, 179.046, 303.050, 135.045, 163.040
31	C25H24O12	515.117[M-H] −	353.087, 173.046, 191.056, 179.046, 303.050, 135.045, 163.040
33	C11H12O4	207.066[M-H] −	173.045[M-2OH] −, 135.045, 179.035, 161.026, 191.056, 119.049
42	C10H10O3	177.056[M-H] −	163.040[M-CH3] −, 146.982, 107.050
Other phenols
13	C9H6O3	161.024[M-H] -	179.034[M-H + 18] −, 135.044[M-CO] −,101.024[M-OH-CO2] −
14	C17H26O4	353.197[M-H + CH3COOH] −	397.181[M-H + COOH] −,
16	C11H8O3	233.045[M + COOH] −	161.023[M-CO] −
20	C8H8O2	135.045[M-H] −	167.034[M + OCH3] −, 108.021[M-CO] −
41	C19H30O4	321.206[M-H] −	365.197[M-2H + COOH] −,303.198[M-H-H2O] −
Flavonoids and derivatives
19	C16H22O8	401.144[M-H + CH3COOH] −	179.035, 135.045, 327.069, 253.071, 269.103
21	C19H18O11	421.076[M-H] −	403.066[M-H-H2O] −, 259.024, 152.012, 108.022, 331.045, 301.035
23	C32H38O19	771.196[M + COOH] −	433.109, 161.023, 179.0327
24	C26H36O11	523.219[M-H] −	460.198[M-2OH-OCH3] −, 363.182[M-Glc + H2O] −, 179.056
25	C27H30O17	625.140[M-H] −	300.028, 191.056, 146.967, 245.081
30	C27H30O16	609.146[M-H] −	633.143[M + Na] +, 251.109, 229.109, 163.039, 487.085
34	C27H30O15	593.149[M-H] −	146.966, 69.041
35	C21H20O11	447.092[M-H] −	367.211, 351.217, 284.032, 195.066, 165.055, 151.039, 137.024
40	C15H10O7	301.035[M-H] −	107.050, 135.044, 151.044, 178.998, 273.040
44	C15H10O6	285.040[M-H] −	107.050, 119.947, 207.065, 225.000
47	C15H12O4	255.066[M-H] −	123.088, 136.016, 151.033, 213.055
Terpenoids
43	C20H30O5	349.202[M-H] −	309.111, 265.084, 251.177, 167.069, 138.066
45	C20H28O4	331.192[M-H] −	161.057, 133.098, 107.082, 81.068, 69.042, 55.052
48	C10H16O	211.133[M-H + CH3COOH] −	229.144 [M-H + CH3COOH + H2O] −,121.105, 109.104, 95.087
50	C10H18O	213.149[M-H + CH3COOH] −	107.086, 55.055, 69.045, 81.071, 95.086
51	C16H30O2	313.237[M-H + CH3COOH] −	185.118, 157.011, 109.102, 95.087, 81.071, 69.071, 55.055
53	C30H48O6	503.336[M-H] −	527.334 [M + Na] +, 157.011, 429.169, 341.206, 317.180, 201.164
54	C30H48O5	487.342[M-H] −	109.100, 69.045, 55.054
55	C16H28O	295.228[M-H + CH3COOH] −	205.160, 135.117, 121.102, 107.087, 95.087, 58.066, 55.055
Steroids
37	C18H26O2	273.187[M-H] −	319.192[M + COOH] −, 317.176[M + COOH-H] −, 255.176[M-H-H2O] −, 215.144, 191.057, 165.057, 137.025
38	C35H54O13	727.353[M + COOH] −	643.300
46	C19H28O2	333.207[M + COOH] −	271.21 [M + H-H2O] +, 253.199[M + H-2H2O] +,159.120, 145.104
Fatty acids
49	C14H28O2	287.223[M-H + CH3COOH] −	199.134[M-CHO] −
52	C10H18O3	185.117[M-H] −	170.003[M-H-CH3] −, 151.039[M-H-2OH] −
Others
2	C5H10O5	195.051[M + COOH] −	113.021[M-H-2H2O] −
15	C7H8N4O2	179.057[M-H] −	164.033[M-CH3] −, 151.040[M-CO] −
32	C10H10O4	193.051[M-H] −	163.040[M-OCH3] −, 133.023[M-2OCH3] −

Note: “–” is not detected.

Fig. 2.

Base peak chromatograms of coffee “husks” in ESI negative ion mode.

The mass spectrometric data of compounds 1–55 were obtained from LC-MSⁿ-IT-TOF experiments in both positive and negative ESI modes. The accurate masses and corresponding assigned elemental compositions of the product ions were summarized in Table 1.

3.2.1. Benzoic acids and derivatives

As shown in Table 1, a total of seven benzoic acids and their derivatives were identified after analysis of mass fragmentation spectra in the negative ion mode: p-hydroxybenzaldehyde (6), vanillic acid (10), protocatechuic acid (12), p-hydroxybenzyl alcohol (26), ethyl vanillin (27), protocatechuic aldehyde (36) and 3-methoxybenzoic acid (39).

In the single-stage mass experiments, using full-scan acquisition (m/z 100–1000) in negative ion mode, the deprotonated molecular ion m/z 167.035 [M-H] ⁻ was readily observed for vanillic acid (10). The molecular ions m/z 167.035 [M-H] ^– of vanillic acid (10) lost a methoxide to produce m/z 137.024 [M-OCH₃] ⁻, and further decarboxylated to produce the fragment ion m/z 123.045 [M-H] ⁻ for p-hydroxybenzyl alcohol (26). The deprotonated molecular ions m/z 153.019 [M-H] ⁻ and m/z 137.024 [M-H] ⁻ were detected for protocatechuic acid (12) and protocatechuic aldehyde (36), which were deprotonated from a carboxyl and an aldehyde group, respectively, and the fragment ions produced at m/z 109.030 were characteristic of these compounds (Nemzer et al., 2021). In the negative ion mode, protocatechuic acid (12) produced fragment ion at m/z 120.022, indicating the loss of two hydroxyl groups, and a fragment ion at m/z 135.045, resulting from the dehydration of the deprotonated molecular [M-H] ⁻ ion. Protocatechuic aldehyde (36), detected at m/z 119.947, yielded a dehydrated fragment ion [M-H₂O] ⁻, and the dehydrated deprotonated ion at m/z 119.035 [M-H-H₂O] ⁻ was formed due to the loss of H₂O from the [M-H] ⁻ ion. Ethyl vanillin (27), detected at m/z 165.056 [M-H] ⁻, generated dehydrated fragment ion at m/z 147.045 [M-H-H₂O] ⁻ and a decarboxylated fragment ion at m/z 119.049 [M-H-CO₂] ⁻. Moreover, fragment ions at m/z 103.055 [M-H-OCH₂CH₃-OH] ⁻ were detected in the mass spectra. Based on literature reports (Song et al., 2023), peak 27 was identified as ethyl vanillin. The mass spectra showed 3-methoxybenzoic acid (39) fragment ions at m/z 151.040 [M-H] ⁻, m/z 119.036 [M-H-OCH₃] ⁻, and m/z 107.050 [M-COOH] ⁻. To further understand the mass spectrometry pattern of benzoic acid derivatives, the cleavage pattern of ethyl vanillin (27) was shown in Supplementary material (Fig. S1A).

3.2.2. Cinnamic acids and derivatives

A total of 9 cinnamic acids and their derivatives were identified in the negative ion mode (Table 1). A detailed illustration of the cleavage mode of 3-O-caffeoylquinic acid (11) was demonstrated in Fig. S1(B and C). Three major peaks (11, 17 and 18) were extracted from the BPC of coffee “husks” at m/z 353.089 [M-H] ⁻ (Fig. 2). These isomers showed identical mass fragmentation pattern among each other and with authentic standard compounds (3-, 5- and 4-O-caffeoylquinic acid, respectively). The dropped caffeic acid fragment ion showed [M-H] ⁻ at m/z 179.035 and was observed to undergo decarboxylation to produce m/z 135.045 [M-COOH] ⁻. The ion at m/z 191.056 [M-H] ⁻ was detected in quinic acid fragmentation. In the positive ion mode, the protonated ions [M + H] ⁺ yielded a fragment ion at m/z 377.085 [M + H + Na] ⁺. The characteristic ions of caffeoyl moiety at m/z 163.040 [M + H-H₂O] ⁺ and m/z 145.029 [M + H-2H₂O] ⁺ were detected due to the loss of H₂O from the released caffeic acid fragment ions. According to literature reports (Clifford et al., 2008; Jaiswal et al., 2014), the position of acylation and structural orientation of hydroxycinnamic acids in the quinic moiety determined the hydrophobicity of these compounds, the elution order of these analogs on reversed phase chromatographic column was as follows: trans-1CQA < trans-3CQA < cis-3CQA < trans-5CQA < trans-4CQA < < cis-5CQA < < < cis-4CQA. Meanwhile, referring to the related literature (Asamenew et al., 2019), peaks 11, 17 and 18 were identified as 3-O-caffeoylquinic acid, 5-O-caffeoylquinic acid and 4-O-caffeoylquinic acid, respectively.

Three isomers (peaks 28, 29 and 31) of di-caffeoylquinic acid (di-CQA) were extracted from the BPC of coffee “husks” at m/z 515.117 [M-H] ⁻ (Fig. 2). In the negative ion mode, the most abundant fragment ions at m/z 353.087 and 173.046 corresponded to the loss of a caffeoyl moiety from the parent ion and the quinic acid fragment at m/z 191.056 [M-H] ⁻ due to the loss of H₂O, respectively. In addition, the dropped caffeic acid fragment ion showed [M-H] ⁻ at m/z 179.035, and was observed to undergo decarboxylation to produce m/z 135.045 [M-COOH] ⁻. In the positive ion mode, two characteristic ions of di-CQA at m/z 163.040 and m/z 303.050 were recognized. The former originated from dehydration of protonated caffeic acid, while the latter derived from protonated mono-substituted caffeoylquinic acid formed by the removal of H₂O and two hydroxyl groups. According to the related literature (Clifford et al., 2008; Jaiswal et al., 2014), the sequence of elution for the di-CQA isomers has been determined as: 1, 3 < 1, 4 < 3, 4 < 5, 3–1, 5(co-eluted) < 4, 5-diCQA. Meanwhile, in conjunction with relevant literature (Asamenew et al., 2019), peaks 28, 29 and 31 were unambiguously assigned as 3, 4-di-O-caffeoylquinic acid, 3, 5-di-O-caffeoylquinic acid and 4, 5-di-O-caffeoylquinic acid, respectively.

In the negative ion mode, the fragment ion m/z 367.102 [M-H] ⁻ from 3-O-caffeoylquinic acid methyl ester (22) was observed, along with fragments at m/z 191.056 and 173.046 generated through demethylation and dehydration of the methylated quinic acid moiety, respectively. The deprotonated ion [M − H] ⁻ of caffeic acid ethyl ester (33) was observed at m/z 207.066, with a fragment at m/z 173.045 resulting from the sequential loss of two hydroxyl groups. Additionally, several minor fragments such as m/z 101.061, 108.022, 119.049, 135.045, 161.026, 179.035 and 191.056 were detected. The deprotonated molecular ion of 4-methoxycinnamic acid (42) at m/z 177.056 [M − H] ⁻ was detected, and the fragment at m/z 163.040 [M-CH₃] ⁻, corresponding to the loss of a methyl group from 4-methoxycinnamic acid (42), was easily recognized. Additionally, fragments at m/z 146.982 and 107.050 were attributed to the loss of a methoxy group and acrylic acid, respectively.

3.2.3. Flavonoids and derivatives

In the current study, we identified 11 flavonoids and their derivatives in both positive and negative ion modes (Table 1). In the negative ion mode, the fragment ion of coniferin (19) at m/z 401.144 [M-H + CH₃COOH] ⁻ was observed. The cleavage of the glycosidic bond between glucose and coniferol yielded the deprotonated glucosyl fragment at m/z 179.035 [M-H] ⁻. Concurrently, the coniferol moiety underwent sequential loss of a methyl and methoxy groups, generating the fragment ion at m/z 135.045. The primary fragmentation patterns of coniferol without glycosidic bond cleavage were detected, characterized by ions at m/z 327.069, 253.071 and 269.103, resulting from the cleavage of the methoxy groups and allyl alcohol.

Mangiferin (21) showed [M-H] ⁻ at m/z 421.076, dehydrated fragment ion at m/z 403.066 [M-H-H₂O] ⁻, and underwent deglucosylation to produce m/z 259.024. Moreover, multiple ring-opening-derived fragments were observed, including those generated via cleavage of the parent core structure (m/z 152.012 and 108.022) and the glucosyl residue (m/z 331.045 and 301.035), as illustrated in Fig. S1(D).

Secoisolariciresinol-β-D-glucoside (24) exhibited [M-H] ⁻ at m/z 523.219, with loss of two hydroxyl groups and one methoxy group producing fragment ions at m/z 460.198 [M-2OH-OCH₃] ⁻, and underwent loss of glucose with additional H₂O to yield m/z 363.182 [M-Glc + H₂O] ⁻, while the cleaved glucose moiety appeared as a characteristic fragment ion at m/z 179.056.

Quercetin-3-O-sophoroside (25) was detected as the [M-H] ⁻ ion at m/z 625.140. The observed fragment at m/z 300.028 corresponded to the sequential loss of two glucosyl units. Subsequently, the fragment ion m/z 191.056 was formed by the loss of the catechol moiety from the flavonoid parent nucleus, and ring-opening of the core structure generated m/z 146.967. Alternatively, cleavage of a single glucosyl residue and ring-opening of the flavonoid skeleton produced a fragment at m/z 245.081.

Kaempferol-3-O-gentiobioside (30) showed [M-H] ⁻ at m/z 609.146. Glycosidic bond cleavage followed by dehydration generated a fragment ion at m/z 251.109, while ring-opening of the flavonoid skeleton produced another fragment at m/z 229.109. In the positive ion mode of ESI, the sodium adducts [M + Na] ⁺ were observed at m/z 633.143, with a fragment at m/z 163.039 arising from hydroxyl group elimination after glucose residue cleavage. Additionally, partial ring-opening of the flavonoid skeleton yielded a fragment at m/z 487.085.

Vicenin-2 (34) was detected with the [M-H] ⁻ ion at m/z 593.149, as well as the fragment at m/z 146.966, generated by ring-opening of the flavonoid core structure. Orientin (35) displayed [M-H] ⁻ at m/z 447.092, and all other detected fragment ions (m/z 367.211, m/z 351.217, m/z 284.032, m/z 195.066, m/z 165.055, m/z 151.039 and m/z 137.024) were formed by cleavage or ring-opening of the three carbocycles in the flavonoid parent skeleton. In addition, quercetin (40), fisetin (44) and pinocembrin (47) showed the deprotonated molecular ions [M-H] ⁻ at m/z 301.035, m/z 285.040 and m/z 255.066, respectively.

3.2.4. Terpenoids

A total of 8 terpenoids were identified in the positive and negative ion modes (Table 1). The fragmentation pattern of andrographolide (43) was illustrated in Fig. S1(E). (±)-camphor (48) was observed the [M-H + CH₃COOH] ⁻ ion at m/z 211.133 and yielded a dehydrated fragment ion at m/z 229.144 [M-H + CH₃COOH + H₂O] ⁻. Characteristic ions of (±)-camphor (48) were detected at m/z 121.105, 109.104, 95.087, 81.073, 67.057 and 55.057 in the positive mode. Similarly, linalool (50) exhibited the [M-H + CH₃COOH] ⁻ at m/z 213.149 and the characteristic ions at m/z 55.055, 69.045, 81.071 and 95.086 were detected.

Andrographolide (43) and dehydroandrographolide (45) showed the deprotonated molecular ions [M-H] ⁻ at m/z 349.202 and m/z 331.192, respectively. In the positive mode, the characteristic fragment ions of andrographolide (43) at m/z 309.111, 265.084, 251.177, 167.069 and 138.066 were detected. Similarly, dehydroandrographolide (45) exhibited characteristic fragment ions at m/z 161.057, 133.098, 107.082, 81.068, 69.042 and 55.052.

Sclareol glycol (51) and madecassic acid (53) were detected as the acetate adduct at m/z 313.237 [M + CH₃COO-H] ⁻ and the deprotonated molecule at m/z 503.336 [M-H] ⁻, respectively. In the positive mode, characteristic ions of sclareol glycol (51) were observed at m/z 109.102, 95.087, 81.071, 69.071, and 55.055. Madecassic acid (53) displayed characteristic fragment ions at m/z 527.334, 429.169, 341.206, 317.180, 201.164 and 135.022. In addition, alisol F (54) and ambroxane (55) were identified with fragment at m/z 487.342 [M-H] ⁻ and m/z 295.228 [M-H + CH₃COOH] ⁻, respectively.

3.2.5. Steroids

Nandrolone (37) showed [M-H] ⁻ at m/z 273.187 and dehydrated fragment ions at m/z 255.176 [M-H-H₂O] ⁻. The carboxylation adducts of nandrolone (37) at m/z 319.192 [M + CO₂] ⁻ underwent dehydrogenation to form a double bond, producing the fragment ion at m/z 317.176. Furthermore, several open-loop fragments were observed, including m/z 215.144, 191.057, 165.057 and 137.025. A detailed fragment ion analysis of nandrolone (37) was demonstrated in Fig. S1(F). Despite the structural similarity between testosterone (46) and nandrolone (37), differing solely in the absence of a methyl group at C19, their fragmentation mass spectra demonstrate significant differences under negative ESI mode. Testosterone (46) was detected with the [M + COOH] ⁻ ion at m/z 333.207. In the positive mode, the adduct ions of testosterone (46) at m/z 271.210 [M + H-H₂O] ⁺ and m/z 253.199 [M + H-2H₂O] ⁺ were recognized due to the loss of H₂O from the parent ions. And other fragment ions of testosterone (46) at m/z 159.120, 145.104, 121.105, 107.089, 95.090 and 81.074 were also detected.

The identification of the aforementioned metabolites in this study was primarily based on database matching and fragment ion analysis. Future research should employ standard samples to conduct further precise confirmation and absolute quantitative validation of these key metabolites, thereby enhancing the certainty of the identification results.

3.3. Key characteristic compounds from washed and sun-exposed of coffee “husks”

After accurate qualitative analysis of key phytochemicals, the chromatograms peak areas of the compounds in the “husks” of the four coffees were normalized (Table S4). Among these substances, cinnamic acid and its derivatives were the most abundant category in the four coffee “husks” samples (K2, K4, K6 and K8). Based on relative abundance, they accounted for 59.88%, 31.60%, 59.12% and 57.68% of the identified compounds, respectively. This class of compounds was primarily composed of chlorogenic acid analogs, providing abundant precursor substances for the potential antioxidant capacity of coffee “husks” (Kobylińska et al., 2025; Murai & Matsuda, 2023; Rojas-González et al., 2022). The result aligned with the conclusion from prior research that “chlorogenic acid analogues were the primary phenolic compounds in coffee” (Kulapichitr et al., 2022), which further confirmed the central role of these compounds in C. arabica.

In terms of primary processing treatment, sun-exposed “husks” contained more amino acids and organic acids, while washed “husks” harbored (p < 0.05) significantly more benzoic acid and derivatives, terpenoids, steroids and fatty acids. The concentration of quinic acid was affected by the primary processing method, with higher levels found in sun-exposed beans compared to washed and honey-treated ones (Toledo et al., 2016). This was supported by the results of coffee “husks” in the present study, which further validated that sun exposure processing increased the concentration of quinic acid. Our previous studies also revealed higher levels of organic acids and amino acids in sun-exposed coffee beans compared to washed and honey-treated beans (Wang, Wang, et al., 2025). This finding indicated that sun exposure provided ample time and conditions for protein degradation into amino acids and the conversion of sugars into various organic acids, ultimately allowing these flavor compounds to accumulate in the beans and “husks”. Furthermore, for coffee “husks”, the high carbohydrate content in the pulp and pectin layer of sun-exposed samples served as a degradation substrate, converting into simple organic acids (Bi, Zhang, et al., 2023), which also contributed to the higher organic acid content in the “husks”.

There have been studies related to fatty acids in coffee “husks”, but reports exploring them from the point of view of primary processing methods were relatively rare. Research has demonstrated that lipids (including steroid and fatty acids) in green coffee beans are lower in sun-exposed and honey-treated beans but highest in washed beans (Kitzberger et al., 2020). This finding of the present study corroborated that report to some extent and added to the information on the variations of fatty acids and steroids content in coffee “husks” under different methods of primary processing.

The above research confirmed our earlier findings that the initial accumulation of secondary metabolites in coffee “husks” during primary processing also indirectly resulted in the enrichment and transformation of compounds in coffee beans, which supported the penetration of chemicals from the “husks” to the coffee beans during the primary processing (Wan, Li, et al., 2024; Wan, Wang, et al., 2024; Wang, Li, et al., 2025; Wang, Wang, et al., 2025).

The results of the comprehensive profiling of the compounds were summarized by means of the Venn diagram shown in Fig. 3. There were 14 common components, which were mainly composed of cinnamic acid and derivatives (11, 17, 18, 22, 28, 29 and 31). L-aspartic acid (1), L-pyroglutamic acid (4), L-phenylalanine (9), vanillic acid (10), vaccarin (23), quercetin-3-O-sophoroside (25) and dimethyl phthalate (32) were present only in the sun-exposed “husks” (K4 and K6), and they could be characteristic metabolites of sun-exposed coffee “husks”. Vanillic acid (10), a metabolite derived from aromatic compounds like vanillin, was synthesized through a multi-step biosynthetic pathway involving multiple enzymatic reactions and exhibited a variety of beneficial biological effects, such as antioxidant, antibacterial and tyrosinase inhibition activities (Ali et al., 2021; Xu et al., 2024). Vaccarin (23) and quercetin-3-O-sophoroside (25) are characterized as a flavonoid and flavonol, respectively, and vaccarin (23) has a wide range of biological activities, including alleviating osteoarthritis and septic cardiomyopathy (Gan et al., 2025; Zhu et al., 2024). Quercetin-3-O-sophoroside (25), a glycoside derivative of quercetin, was shown to be an inhibitor of calcium channels with promising applications in biology (Zhuang et al., 2016).

Fig. 3.

Venn diagrams of key characteristic compounds of coffee “husks”.

Dimethyl phthalate (32), commonly used as a plasticizer and flavor fixative, was generally considered to originate from chemical synthesis rather than natural sources, and was therefore classified as an organic contaminant and excluded as a characteristic metabolite (Wan, Li, et al., 2024). Consequently, vanillic acid (10), vaccarin (23) and quercetin-3-O-sophoroside (25) were finally identified as characteristic metabolites of the sun-exposed “husks”. 8-Gingerol (41) and pinocembrin (47) were selected as characteristic compounds for washed coffee “husks”. Among them, pinocembrin (47) was a dihydroflavonoid with extensive activities involving anti-inflammatory, antimicrobial and antioxidant effects (Elbatreek et al., 2023).

3.4. Correlation between activity and compounds

With the aim of identifying specific constituents responsible for the observed antioxidant and XO inhibitory activities in the coffee “husks”, the correlation between activities and constituents in the 4 candidates (K2, K4, K6 and K8) was analyzed in detail.

There was a high degree of concordance (r > 0.90) between the three antioxidant activities and the XO inhibitory activity (Fig. 4). Studies showed that natural products with strong antioxidant capacity could also serve as potential XO inhibitors (Sui et al., 2021). Correlation analysis showed that a total of 15 compounds were positively correlated with both antioxidant activities (DPPH•, ABTS^•+ and FRAP) and XO inhibitory activity (Fig. 4). Among them, 3-O-caffeoylquinic acid (11), 5-O-caffeoylquinic acid (17), 4-O-caffeoylquinic acid (18), 3-O-caffeoylquinic acid methyl ester (22), 3,4-O-dicaffeoylquinic acid (28), 3,5-O-dicaffeoylquinic acid (29), 4,5-O-dicaffeoylquinic acid (31), ambroxane (55) and nandrolone (37) were present in all coffee “husks” samples. 3-O-caffeoylquinic acid (11), 5-O-caffeoylquinic acid (17), 4-O-caffeoylquinic acid (18), 3,4-O-dicaffeoylquinic acid (28), 3,5-O-dicaffeoylquinic acid (29) and 4,5-O-dicaffeoylquinic acid (31) have all been reported to possess antioxidant activity (Han et al., 2024; Li et al., 2018; Makori et al., 2021). Additionally, 4-O-caffeoylquinic acid (18), 3,4-O-dicaffeoylquinic acid (28), 3,5-O-dicaffeoylquinic acid (29) and 4,5-O-dicaffeoylquinic acid (31) have shown XO inhibitory activity, with IC₅₀ values ranging from 26.74 to 70.23 μmol/L (Mehmood et al., 2022). Collectively, the antioxidant and XO inhibitory activities were strongly correlated with chlorogenic acid analogs in the coffee “husks”, which were identified as the contributors responsible for the bioactivities of the four coffee “husks” samples (Song et al., 2024; Wan et al., 2021). In this study, the K2 sample exhibited the highest relative content of chlorogenic acids, providing direct evidence for its superior biological activity.

Fig. 4.

Heatmap of correlation between antioxidant and XO inhibitory activities and non-volatile components. Note: The correlation analysis was performed using Pearson's correlation coefficient (r), calculated based on all samples (n = 55).

As shown in Fig. 4, seven compounds exhibited significant correlations (r > 0.70) with both antioxidant activity and XO inhibitory activity: p-hydroxybenzaldehyde (6), plumbagin (16), caffeic acid ethyl ester (33), fisetin (44), (±)-camphor (48), linalool (50) and sclareol glycol (51). These compounds exhibited broad biological activities; plumbagin (16) and linalool (50) have been demonstrated to possess antioxidant, antibacterial, anti-inflammatory and anticancer effects (Mączka et al., 2022; Petrocelli et al., 2023). Caffeic acid ethyl ester (33) could mitigate drug toxicity by inhibiting Escherichia coli β-glucuronidase and alleviate lung injury through anti-inflammatory pathways (Huang et al., 2025; Li et al., 2020). Fisetin (44), a flavonoid compound, improved hyperuricemic nephropathy by regulating renal uric acid transporters (Ren et al., 2021). These compounds have been identified as characteristic metabolites of washed “husks” K2, indicating that they are likely the key contributors to its outstanding antioxidant and XO inhibitory capabilities.

3.5. Antioxidant and XO inhibitory activities of chlorogenic acid analogs in vitro

Analysis of the relative content of substances in coffee and correlation of activities with non-volatile metabolites revealed that chlorogenic acid analogs including three mono-substituted chlorogenic acids (3-, 4- and 5-O-caffeoylquinic acids) and di-substituted chlorogenic acids (3, 4-, 3, 5- and 4, 5-di-O-caffeoylquinic acids) were key phytochemical components for potential antioxidant and XO inhibitory properties in the four samples. To further understand the effect of such substances on coffee “husks” activity, six chlorogenic acid analogs (11, 17, 18, 28, 29 and 31) identified in this study were tested in vitro for their antioxidant and XO inhibitory activities, using allopurinol and quercetin as positive controls.

All six compounds demonstrated desirable antioxidant activity (Table 2). The three antioxidant activities of di-substituted chlorogenic acids were significantly higher than those of mono-substituted chlorogenic acids (p < 0.05), closely approaching the activity levels of positive controls. Meanwhile, 4,5-di-O-caffeoylquinic acid (31) had the best DPPH• and ABTS^•+ scavenging ability with IC₅₀ values of 0.317 and 0.103 mmol/L, respectively, and 3,4-di-O-caffeoylquinic acid (28) exhibited the best FRAP reducing power (7.276 mmol Fe²⁺/mmol). Among the mono-substituted chlorogenic acids, 3-O-caffeoylquinic acid (11) displayed the best antioxidant activity. In the XO inhibitory activity assay in vitro, 4-O-caffeoylquinic acid (18) possessed the highest XO inhibition rate of 29.438% at a concentration of 0.5 mmol/L. Simultaneously, the di-O-caffeoylquinic acid with a C-4 substitution showed superior XO-inhibitory activity, indicating that the position of caffeic acid substitution affected the XO inhibitory activity of chlorogenic acids, and that the advantage of the C-4 substitution was evident in both mono- and di-substituted chlorogenic acids. Future research should focus on chlorogenic acid derivatives in coffee “husks”, conducting precise quantitative activity evaluations and exploring their structure-activity relationships in depth.

Table 2.

DPPH•, ABTS^•+, FRAP antioxidant and XO inhibitory activity of compounds in vitro.

Compounds	DPPH•	ABTS•+	FRAP	XO
	IC50 (mmol/L)	IC50 (mmol/L)	(mmol Fe2+/mmol)	I (%)
3-O-caffeoylquinic acid (11)	0.626 ± 0.008c	0.211 ± 0.000c	4.973 ± 0.048e	27.134 ± 0.459c
4-O-caffeoylquinic acid (18)	0.865 ± 0.002a	0.256 ± 0.008a	3.687 ± 0.049g	29.438 ± 1.838b
5-O-caffeoylquinic acid (17)	0.817 ± 0.017b	0.225 ± 0.008b	3.989 ± 0.011f	18.014 ± 0.159f
3,4-Di-O-caffeoylquinic acid (28)	0.432 ± 0.002d	0.130 ± 0.004d	7.276 ± 0.022b	24.992 ± 0.695cd
3,5-Di-O-caffeoylquinic acid (29)	0.347 ± 0.003e	0.128 ± 0.002d	6.770 ± 0.008c	21.973 ± 0.981e
4,5-Di-O-caffeoylquinic acid (31)	0.317 ± 0.004f	0.103 ± 0.002e	6.344 ± 0.114d	23.337 ± 1.471de
Quercetin	0.239 ± 0.003g	0.078 ± 0.002f	8.081 ± 0.071a	91.090 ± 0.337a

Values are expressed as means ± Standard Deviation (n = 3). Note: different letters within the same column denote significant differences (p < 0.05) based on one-way ANOVA and Tukey's HSD post-hoc test, and the XO inhibitory activity of all standards was determined at a concentration of 0.50 mmol/L.

3.6. Molecular docking of XO inhibitory activities

The interaction between XO and ligands was studied using molecular docking simulations with allopurinol and quercetin as positive controls. In order to further investigate the inhibitory ability of these compounds against XO at the molecular level, molecular docking simulations of 6 chlorogenic acid compounds and 11 potentially active compounds in coffee “husks” were conducted.

Binding energy is one of the most important parameters in molecular docking that determines the stability of protein complexes and the binding specificity of ligand-protein adducts, and the lowest binding energy implies good affinity and inhibitory effects on protein activity (Limanto et al., 2019). Additionally, the inhibition constant (Ki) plays a vital role in determining ligand-protein interactions, with smaller values indicating tighter ligand-protein binding and stronger inhibition of protein activity. Molecular docking results showed that although the binding free energies of these six chlorogenic acid molecules (−4.46 to −3.32 kcal/mol) were significantly higher than those of the positive control quercetin (−6.0 kcal/mol) and allopurinol (−5.11 kcal/mol) (Table S5), they all spontaneously bound to XO (Table S5). Meanwhile, among the mono-substituted chlorogenic acids, 5-O-caffeoylquinic acid (17) displayed the greatest XO suppression ability.

The results of the interaction of chlorogenic acid analogs with XO were shown in Fig. 5, where 5-O-caffeoylquinic acid (17) formed three hydrogen bonds with residues LYS249, VAL259 and LEU404 on the B chain of XO, and generated hydrophobic interactions with GLU402 and ILE353 on the B chain. The above-mentioned hydrogen bonds and the hydrophobic interactions on the benzene ring stabilized the three moieties of 5-O-caffeoylquinic acid (17) within the XO active site, enhancing its structural stability and resulting in superior XO inhibitory activity. Furthermore, both 3-O-caffeoylquinic acid (11) and 4-O-caffeoylquinic acid (18) formed four common hydrogen bonds with XO residues GLN423, LYS422, LYS433, and LYS1228. In summary, the hydrogen bonding and the hydrophobic interaction on the benzene ring stabilized the three moieties of mono-substituted chlorogenic acids in the XO active site, thereby enhancing structural stability.

Fig. 5.

The interactions of (A) allopurinol, (B) 3-O-caffeoylquinic acid (11), (C) 5-O-caffeoylquinic acid (17), (D) 4-O-caffeoylquinic acid (18), (E) 3, 4-di-O-caffeoylquinic acid (28), (F) 3, 5-di-O-caffeoylquinic acid (29) and (G) 4, 5-di-O-caffeoylquinic acid (31) with XO.

Among the di-substituted chlorogenic acids, 3, 5-di-O-caffeoylquinic acid (29) showed the strongest XO inhibitory activity (Table S5 and Fig. 5). It formed three hydrogen bonds with residues ARG60 and ASP59 on the A chain of XO and TRP283 on chain B, along with additional interactions with TYR58 and HIS82 on chain A, and PRO224 and LYS225 on chain B. Thus, 3, 5-di-O-caffeoylquinic acid exhibited stronger hydrophobic interactions in addition to forming hydrogen bonds, leading to the formation of stable spatial structures that demonstrated better XO inhibitory activity.

Molecular docking results showed (Table S5 and Fig. 6) that the binding energies and Ki values of plumbagin (16), fisetin (44) and pinocembrin (47) were lower than those of allopurinol and quercetin, suggesting that these compounds could theoretically inhibit XO effectively. Meanwhile, we found that plumbagin (16), which retains only the chromone core, exhibited the lowest binding energy and Ki value with XO. This suggested that the chromone skeleton may be a crucial structural feature for inhibitory activity, rather than the entire flavonoid core. Studies have also confirmed that flavonoids containing the chromone structure (such as flavones, flavonols and isoflavones) generally exhibit stronger inhibitory activity compared to those lacking this skeleton (e.g., flavanones), primarily due to their A-ring and C-ring structures resembling purine, enabling more efficient binding to the purine-binding pocket of XO (Lin et al., 2015; Sianipar et al., 2022). Therefore, we hypothesized that the effective retention of the chromone moiety may also contribute to the XO inhibitory activity of flavonoids, with the specific mechanism requiring further validation through more precise multidimensional activity studies.

Fig. 6.

The interactions of (A) quercetin (40), (B) plumbagin (16), (C) fisetin (44) and (D) pinocembrin (47) with XO.

The substituents on different rings of flavonoids (including the position, number and specific types of substituents) also affected the activity of the substances. Both quercetin (40) and fisetin (44) possess a C C double bond, and both contain a hydroxyl group on the A ring. Previous studies have indicated that the C C double bond, which maintains planarity in the flavonoid structure, is a key factor in inhibiting XO activity. Hydroxyl groups at positions C7 and C5 could form favorable hydrogen bonds and interactions with the active site, thereby enhancing the inhibitory effect (Sianipar et al., 2022; Ullah et al., 2024; Xue et al., 2023). These structural features collectively explained the lower binding energies and Ki values of quercetin (40) and fisetin (44) for XO.

Pinocembrin (47) possessed potent XO inhibitory activity, with a binding pattern to XO similar to that of febuxostat, a clinically used XO inhibitor (Fu et al., 2019). In contrast, fisetin (44) exhibited lower binding free energy and Ki values toward XO. This may be attributed to the presence of a C C double bond in its structure and the dual hydroxyl substituents on the B-ring, which spatially formed stronger hydrophobic interactions, such as pi-alkyl interactions with VAL1011, LEU873 and LEU1014 on the C chain of XO. Previous studies indicated that the hydroxyl group at C4 on the B-ring contributes to enhanced XO inhibitory activity, potentially explaining why fisetin (44) exhibits stronger activity than pinocembrin (47). However, other literature reported that the hydroxyl group at C3 on the A-ring, when extended into the hydrophobic region of the active site, may induce structural instability and weaken inhibitory effects (Xue et al., 2023). Thus, the influence of hydroxyl groups on the flavonoid aglycone on activity is complex. Furthermore, the accuracy of molecular docking predictions was limited by factors such as the approximation of scoring functions, protein flexibility, and solvent effects (Saikia & Bordoloi, 2019; Tao et al., 2020). In this study, although compounds such as quercetin (44), pinocembrin (47) and plumbagin (16) exhibited strong binding energies with XO, their actual inhibitory activity required experimental validation.

In summary, the XO inhibitory activity demonstrated by coffee “husks” extracts likely stems from the synergistic interaction of components such as chlorogenic acid analogues and flavonoids, rather than being dominated by a single high-affinity compound.

4. Conclusions

To recognize the connection between chemical profiles and bioactive potential of the most widely discarded coffee “husks” from Yunnan, the main characteristic metabolites together with their detailed mass spectrometry cleavage patterns, their bioactivities in vitro and molecular docking testing were elucidated in the present study. Among them, chlorogenic acid analogs, as key bioactive constituents, comprised over 66% of the identified phenolic compounds. The antioxidant activities of di-substituted chlorogenic acids were significantly higher than those of mono-substituted chlorogenic acids (p < 0.05). And these compounds spontaneously bound to XO. More specifically, the enhanced hydrophobic interactions of the di-substituted chlorogenic acids with XO were enhanced along with hydrogen bond formation, which promoted the stable spatial conformation and ultimately demonstrated enhanced XO-inhibiting potency. Furthermore, flavonoids and terpenoids could be important for the prominent antioxidant and XO inhibitory activities of washed “husks”. Among them, flavonoids and their derivatives such as pinocembrin (47), fisetin (44) and plumbagin (16) exhibited potentials in XO inhibition. And this advantage was primarily attributed to the flavone skeleton, in which the presence of a stable chromone group was vital.

Collectively, chlorogenic acid analogs served as key bioactive constituents in coffee “husks” of C. arabica from Yunnan with prominent antioxidant and potential XO inhibitory activities, providing new insights into the development of the natural antioxidants and XO inhibitors and need further exploration.

CRediT authorship contribution statement

Hong Wang: Writing – original draft, Methodology, Investigation, Data curation. Li Wan: Methodology, Data curation. Yu Wang: Investigation, Formal analysis. Yan Li: Methodology, Investigation, Data curation. Wenli Zhao: Investigation, Formal analysis. Wenjuan Liang: Writing – review & editing, Project administration, Funding acquisition, Conceptualization.

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.