Chemical profiling and antioxidants screening from natural products: using CiNingJi as an example

Abstract

Natural products with good antioxidative properties have been paid increased attention globally. However, due to its chemical complexity, it is difficult to find out its antioxidative compounds. Herein, the chemical profiling and antioxidant capacity of CiNingJi (CNJ) were analyzed, as an example. By using UHPLC-Q-TOF/MS, a total of 82 compounds were tentatively deduced. Furthermore, its free radical scavenging capacity was assessed by different in vitro spectrophotometric-based assays. The result showed that one ingredient, Rosa roxburghii, plays a critical role in its antioxidant activity. In addition, 18 potential antioxidants were screened out in CNJ by comparing the difference of it with and without DPPH reaction. They were identified mainly as catechin, ellagic acid, kajiichigoside F1, and their derivatives or isomers. With the further quantification of major found antioxidants, our results may provide some knowledge on predicting the antioxidative compounds of natural products.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10068-022-01049-4.

Introduction

As more and more people realize that foods and their derived products containing bioactive compounds can bring benefits to human health recently, the interest of health-conscious consumers towards them is growing considerably. These products, which were known as nutraceuticals and functional foods containing enriched phytochemical extracts from herbs or traditional Chinese medicines (TCMs), are now widely available on the market (Espin et al., 2007; Gonzalez-Sarrias et al., 2013). With their bioactive components, there are many healthy properties were reported about them, including antioxidant, cardioprotective, anticancer, and hypolipidemic (Chang et al., 2020; Chen et al., 2019; Kalhori et al., 2021; Parham et al., 2020). Among them, antioxidant protection effects were reported against various diseases or illnesses relevant to oxidative stress, such as aging, cardiovascular diseases, cancer, Alzheimer’s disease, hypertension, and inflammation (Ayvaz et al., 2018; Grzesik et al., 2018; Jiang et al., 2018; Mansoori et al., 2021; Zhou et al., 2020). Therefore, foods, fruits, medicinal plants, or nutraceuticals and pharmaceutical agents that contained natural antioxidants have drawn increased attention globally. However, due to the chemical complexities of natural products, their antioxidant activities are difficult to acquire, as well as the “real” antioxidative compounds  that might contribute to their properties. Herein, the chemical profiling and antioxidative capacity of CiNingJi beverage (CNJ) were analyzed, as an example.

CNJ is a newly developed healthy product, which mainly consisted of Rosa roxburghii, Citrus limon, and Malus domestica (apple). As the main and principal material in CNJ, R. roxburghii is an excellent plant resource that has the concomitant function of both medicine and foodstuff. For now, phytochemical research have shown that it has lots of bioactive compounds, such as phenolics, flavonoids, organic acids, triterpenes, amino acids, etc. (Hou et al., 2020; Liu et al., 2016). However, finding out the compounds with good antioxidative properties in R. roxburghii is still not available, neither is CNJ.

Therefore, in order to acquire the antioxidative capacity of natural products and antioxidative compounds, we used CNJ as an example. First, the chemical profile of CNJ was elucidated by using ultrahigh performance liquid chromatography quadrupole time-of-flight mass spectrometry (UHPLC-Q-TOF/MS). Then, the antioxidant activities of CNJ and its ingredients were evaluated with different methods, such as 1,1-Diphenyl-2-picrylhydrazyl (DPPH), 2,2′-azinobis-(3-ethylbenzthiazoline-6-sulfonic acid) (ABTS) scavenging abilities, and ferric-reducing antioxidant power (FRAP). And the potential antioxidant compounds were found based on LC-MS coupled with pre-column DPPH reaction. Finally, major antioxidants in CNJ were also quantified. For the first time, the chemical components of CNJ were illuminated, and the antioxidant profiles were screened and quantified. With more and more attention on the natural antioxidants, it might be useful for predicting the antioxidative compounds of other nutraceuticals or pharmaceutical agents, which contain medicinal plants/food.

Materials and methods

Materials and reagents

Methanol and acetonitrile of LC-MS grade were purchased from Anaqua Chemicals Supply (Houston, TX, USA), and formic acid (LC-MS grade) was provided by Fisher Scientific Laboratories, Inc. (Pittsburgh, PA, USA). Ultrapure water (18.2 MΩ) was produced by a Direct-Q® 8 UV–R Milli-Q purification system (Millipore, MA, USA).

Reference substances, (+)-catechin (25), isoquercitrin (51), rutin (57), quercetin (64), and kajiichigoside F1 (71) were obtained from Shenzhen hiboled century Biotechnology Co. Ltd. (Shenzhen, China), while L-ascorbic acid (1), procyanidin B1 (23), ellagic acid (48) and hyperoside (49) were supplied by Guangzhou zhuanyan Biological Technology Co. Ltd. (Guangzhou, China). Chemical structures of these standards were shown in Figure S1, and their purities were found more than 98.0% by HPLC analysis. 2, 4, 6-tripyridyl-s-triazine (TPTZ), DPPH, potassium persulfate, and ABTS were purchased from Sigma Aldrich (Shanghai, China). Internal standard (IS) substance, 2-chloro-phenylalanine, was purchased from Bide Pharmatech Ltd. (Shanghai, China).

The extracts of R. roxburghii, C. limon, M. domestica and CNJ beverage (batch number: ZX 20210120-010) were provided by Wanlaoji Pharmaceutical Co., Ltd. (Guangzhou, China). The proportion of each raw material in CNJ is R. roxburghii: C. limon: M. domestica = 1.7: 2.3: 1.5 (by weight).

Preparation of solutions

Standard solutions for calibration All 9 reference standards were accurately weighted and prepared in methanol-water (5:5, v/v) as stock solutions (mg/mL): L-ascorbic acid (6.07), procyanidin B1 (10.10), (+)-catechin (10.03), ellagic acid (9.95), hyperoside (10.27), isoquercitrin (10.21), rutin (10.12), quercetin (10.03), kajiichigoside F1 (10.97). An appropriate volume of each stock solution was mixed to yield the mixed stock solution. IS working solution was also obtained with a concentration of 10 µg/mL. Then, working standard solutions were obtained by diluting of mixed stock solution to provide a series of working standard solutions for calibration (Table 1). Working solutions with low, middle, and high concentrations were used as quality controls (QCs) for the investigation of accuracy, precision, and stability. All standard solutions were stored at 4 °C until used, and filtered through a nylon membrane filter (0.22 µm, Phenomenex, CA, USA) before UHPLC-Q-TOF/MS analysis.

Table 1.

Regression equations, linear ranges of nine antioxidants and their contents in CNJ and its ingredients (mean ± SD) (µg/mL)

No.	Analyte	Measured value [ion form]	Regression equation	r	Linear range (µg/mL)	AP	NM	CL	CNJ
1	L-ascorbic acid	175.0248 [M − H]−	y = 0.0115x + 0.0328	0.988	12.00-2000	23.87 ± 4.35	24.81 ± 5.71	1819.63 ± 463.46	329.44 ± 86.24
23	Procyanidin B1	577.1351 [M − H]−	y = 0.1077x + 0.5244	0.991	0.20-200	ND	ND	164.81 ± 67.84	5.64 ± 3.14
25	(+)-Catechin	289.0717 [M − H]−	y = 0.1761x + 0.2162	0.999	0.20-40	ND	ND	77.06 ± 22.19	8.21 ± 3.71
48	Ellagic acid	300.9990 [M − H]−	y = 0.0378x + 0.5683	0.996	0.40-200	0.09 ± 0.04	ND	75.41 ± 8.20	4.48 ± 1.07
49	Hyperoside	463.0882 [M − H]−	y = 0.0935x + 1.2593	0.997	0.05-40	ND	0.37 ± 0.18	1.02 ± 0.17	0.75 ± 0.26
51	Isoquercitrin	463.0882 [M − H]−	y = 0.1176x + 1.2437	0.992	0.05-40	ND	0.41 ± 0.07	1.56 ± 0.43	0.94 ± 0.27
57	Rutin	609.1461 [M − H]−	y = 0.1409x + 1.1948	0.994	0.20-200	ND	0.08 ± 0.01	0.10 ± 0.04	0.06 ± 0.03
64	Quercerin	301.0354 [M − H]−	y = 0.144x + 1.9332	0.990	0.40-200	0.09 ± 0.02	ND	ND	ND
71	Kajiichigoside F1	695.4012 [M + COOH]−	y = 0.3951x + 0.5426	0.999	0.20-200	ND	ND	156.87 ± 20.85	5.19 ± 2.71

CL: R. roxburghii; NM: C. limon; AP: M. domestica. ND, not detected

Sample solutions After samples warmed to room temperature, an appropriate amount of IS was added to obtain a certain concentration of IS (10 µg/mL) (Lin et al., 2018). And then, they were filtered through membrane filters for analysis. As for antioxidant assays, the extract of all ingredients and CNJ were concentrated to same dry weight (88 mg/mL) and diluted to appropriate concentration with distilled water.

Solutions for antioxidant assays DPPH radical solution was prepared by dissolving an accurately weighed DPPH sample (24.6 mg) into a 100 mL brown volumetric flask right with methanol before the experiments and the whole procedure was protected from light. Meanwhile, 5 mL of ABTS (7 mM) was mixed with 88 µL of potassium persulfate (140 mM) to produce ABTS⁺ stock solution, and it was incubated in the dark at room temperature for 16 h. And FRAP stock solution contained 2.5 mL of a 10 mM TPTZ solution in 40 mM HCl, 2.5 mL of 20 mM FeCl₃ and 25 mL of 0.3 M acetate buffer, pH 3.6. It was prepared daily and kept in the dark at 37 °C before use.

UHPLC-Q-TOF/MS analysis

For the UHPLC-ESI-MS/MS analysis, all samples were analyzed using a Nexera X2 Shimadzu UHPLC system coupled with an AB SCIEX TripleTOF® 5600 mass spectrometer. The chromatographic separation was conducted using Waters ACQUITY UHPLC® BEH C18 column (2.1 × 100 mm, 1.7 µm). The optimal mobile phase consisted of 0.1% formic acid-containing water (A) and 0.1% formic acid-containing acetonitrile (B) at a flow rate of 0.3 mL/min with 40 °C. The solvent gradient was set as follows: 0–0.5 min, 5% B; 0.5–5 min, 5–10% B; 5–10 min, 10–25% B; 10–18 min, 25–35% B; 18–20 min, 35–80% B; 20–20.5 min, 80–95% B; 20.5–28 min, 95% B. The autosampler was set at 6 °C, and the sample injection volume was set at 2 µL.

The mass spectrometer coupled with an electrospray ionization (ESI) source was operated in both positive (POS) and negative (NEG) ion mode. The data was collected using information-dependent acquisition (IDA) mode with mass range m/z 100–1200 for TOF-MS scan, and m/z 50–1200 for TOF-MS/MS scan. After optimization, the nebulizer gas, heater gas, and curtain gas were set at 50, 50, and 35 L/min, respectively. The ion spray voltage and declustering potential in POS mode were set at 5.5 kV and 80 V, while they were at −4.5 kV and −80 V in NEG mode. And for the IDA experiments, the collision energies were set at 35 eV and −35 eV, and collision energy spread was set at 15 eV. The experiments were run with 150 ms accumulation time for TOF-MS and 50 ms accumulation time for TOF-MS/MS. Meanwhile, the accurate MS and MS/MS measurements were acquired with the Automated Calibration Delivery System. As for quantification, the column and conditions were the same with described qualitative method except for acquisition only in NEG mode. And the information for quantified compounds was shown in Table 1.

Antioxidant assays

DPPH free radical scavenging activity DPPH scavenging activity was assessed as described with little revise (Bao et al., 2018). Briefly, different volumes of samples were diluted with methanol to 500 µL, and then mixed with DPPH solution (500 µL). Mixtures were left to react in the dark for 30 min at 25 °C. After incubation, the absorbance (A_sample) was measured at 517 nm using Multiskan GO Microplate Reader (Thermo Scientific, USA). L-ascorbic acid was used as the positive control, with different concentrations for standard curve. The DPPH radical scavenging activity of sample was expressed as mg of L-ascorbic acid equivalents (AAE) per gram of dried sample (mg AAE/g). In addition, to determine the IC₅₀ (50% inhibition) of samples on DPPH, five different concentrations were used. And, the distilled water instead of sample was made as blank control (A_control). Antioxidant activity of DPPH scavenging was calculated as follow: (%) inhibition = [(A_control − A_sample) / A_control)] ×100%. IC₅₀ value was determined to be the effective concentration at which DPPH radicals were inhibited by 50%.

ABTS free radical scavenging activity ABTS radical scavenging activity was assayed according to the reported method (Sridhar and Charles, 2019). First, ABTS⁺ stock solution was diluted with ethanol to get a stable absorbance of about 0.700 ± 0.005 at 734 nm. Afterward, 100 µL of appropriately diluted sample was added to 200 µL of ABTS⁺ solution in a 96-well plate. Then, the mixture was incubated in the dark at room temperature for 30 min and determined the absorbance at 734 nm. Ethanol was selected as blank control, and concentrations of 0 to 50 µg/mL L-ascorbic acid dissolved in water were made to construct standard curve. The scavenging activities of sample against ABTS⁺ were also expressed as µmol AAE/g. IC₅₀ value was calculated similarly with DPPH scavenging method.

Ferric reducing/antioxidant power assay The FRAP assay was carried out as previously described (Zhu et al., 2015). 50 µL of properly diluted sample was added to 250 µL of freshly prepared FRAP reagent in a 96-well plate and mixed thoroughly. The mixture was incubated at 37 °C for 10 min, and then measured at 593 nm. The FRAP value was calculated as millimoles of Fe²⁺ equivalents per 100 g of sample (mmol Fe²⁺ equiv/100 g) based on a calibration curve plotted using FeSO₄·7H₂O as standard curve.

Screening out of antioxidants by pre-column UHPLC-Q-TOF/MS

DPPH assay combined with UHPLC-Q-TOF/MS was employed to rapidly screen out the antioxidants in extracts. Briefly, after the incubation of sample with DPPH solution, UHPLC-Q-TOF/MS was employed to further screen out the antioxidant profiles and identify the antioxidant components by detecting changes of characteristic peaks compared with the corresponding blank sample (methanol + sample, without DPPH reaction). 2 µL aliquot of the sample with and without DPPH reaction were analyzed under the same conditions as described before. The scavenging percentage of target compound was calculated according to the following formula: [(PA_sample − PA_{sample+ DPPH}) / PA_sample] × 100%.

Quantitative analysis of major antioxidants

The calibration curves for nine antioxidative compounds was constructed by the plotting peak area ratios (y) of standard/IS versus corresponding concentrations (x). Six replicates of QCs at three different levels (low, medium, high) were analyzed within the same day for intra-day variation, and with three consecutive days for inter-day precision and accuracy. The recovery for these nine compounds were also determined by analyzing them in CNJ beverage spiked with QCs. And, the limit of quantification (LOQ) was estimated as the lowest concentration of the calibration curve with acceptable relative error (≤ 20%) and RSD (≤ 20%), as well as signal-to-noise ratio (≥ 10).

Data analysis

UHPLC-MS data was collected using Analyst TF 1.7.1 software and processed by Peakview (AB Sciex, MA, USA) software. As for antioxidant assays, all the tests were performed in triplicate, and results were analyzed and expressed as mean ± standard deviation using Graph Pad Prism 5.0 (La Jolla, CA, USA).

Results and discussion

Identification of Major Constituents in CNJ

Due to the complexity of the natural plant, the structure types and amounts of chemical compounds in CNJ were numerous. Therefore, before data processing, an in-house formula database consisting of name, molecular formula, chemical structure, and accurate mass of the compounds in individual ingredients of CNJ was established by searching from databases, such as TCMSP (https://www.old.tcmsp-e.com/tcmsp.php), PubMed (http://www.ncbi.nlm.nih.gov/pubmed), CNKI (http://www.cnki.net).

Because of the numerous compounds of natural products and their different chemical properties, both positive and negative ion modes were performed for MS analysis and the typical total ion chromatography (TIC) profiles of CNJ in the positive and negative ion mode were presented in Fig. 1. A total of 82 compounds were discovered, including 26 phenolic acids, 9 organic acids, 26 flavonoids, 11 triterpenoids, 6 other and 4 unknown compounds (Table 2). Among them, 9 compounds were unambiguously characterized by comparing with the reference standards. The other compounds were tentatively identified based on their retention time, mass accuracy of precursor ions, the fragmentation pathways from MS/MS spectra, reported literatures and established in-house database of CNJ. The specific chemical structures of these compounds were illustrated in Figure S1, and the exact ESI-MS/MS mass spectrum of these identified compounds was shown in Supporting Information.

Fig. 1.

Representative total ion chromatograms of the CNJ sample in positive and negative ion mode

Table 2.

Identification of the chemical constituents in CNJ by using UHPLC − ESI-Q-TOF/MS analysis

No.	Rt (min)	Molecular Formula	[M + H]+	(+) MS/MS	[M - H]−	(-) MS/MS	Identification	Source
1a	0.74	C6H8O6			175.0231	157.0120, 127.0018, 115.0014, 87.0068	L-ascorbic acid	CL
2	0.87	C5H10O6			165.0394	147.0287, 129.0191, 75.0088	Ribonic acid or Xylonic acid	CL, AP
3	0.88	C7H12O6			191.0561	173.0446, 127.0381, 85.0724	Quinic acid	CL, NM
4	0.93	C12H22O11	365.1023	203.0519, 185.0299	341.1082	179.0567, 119.0354, 89.0249	6-O-α-D-Galactopyranosyl-D-glucose or Sucrose or isomer	CL
5	0.99	C4H8O5			135.0290	117.0194, 89.0246, 75.0092	Threonic acid	CL, NM, AP
6	1.33	C6H8O7			191.0173	173.0071, 129.0166, 111.0066, 85.0276	Citric acid	CL, NM, AP
7	1.33	C4H6O5			133.0132	115.0013, 89.0223, 71.0117	Malic acid	AP, CL
8	1.33	C6H8O7			191.0176	173.0066, 154.9971, 111.0063, 87.0064	Isocitric acid	CL, NM, AP
9	1.57	C9H8O3	165.0546	147.0389,123.0389,119.0441			p-Coumaric acid	AP
10	2.69	C7H6O4			153.0191	123.0449, 109.0296, 81.0351,	Protocatechuic acid	AP, CL
11	1.84	C13H16O10			331.0657	169.0127, 125.0225,	Galloyl-glucose	CL
12	1.88	C7H6O5			169.0145	151.0018, 125.0243, 107.0133, 97.0293	Gallic acid	CL
13	2.66	C9H11NO2	166.0868	120.0813,103.0548			Phenylalanine	NM, AP
14	2.71	C45H38O18	867.2125	579.1475,289.0709	865.2067	713.1536, 695.1464, 575.1241, 407.0800, 287.0586	Procyanidin C1 or isomer	CL
15	2.91	C34H24O22	785.0833	767.0756, 465.0666, 321.0249, 303.0133, 277.0338	783.0701	401.0629, 300.9999, 275.0203	Pedunculaginb or Casuariin	CL
16	3.63	C9H8O5			195.0295	179.0346, 165.0198, 151.0405, 123.0455	3, 4, 5-Trihydroxycinnamic acid	CL
17	4.13	C27H22O18	635.0535	465.0350, 321.0038, 277.0165	633.0732	481.0625, 300.9987, 169.0138	Galloyl-HHDP-glucose or Strictinin	CL
18	4.24	C8H12O7	221.0538	185.0328, 157.0391, 138.9937, 111.0005	219.0531	157.0509, 111.0089, 87.0088	1-Ethyl Citrate	NM
19	4.65	C41H28O27			951.0767	907.0887, 783.0721, 300.9998	Geraniin or isomer	CL
20	4.77	C34H24O22	785.0833	767.0756, 465.0666, 321.0249, 303.0133, 277.0338	783.0713	401.0623, 300.9996, 275.0197	Pedunculaginb or Casuariin	CL
21	4.79	C41H28O27	953.0893	935.0157, 633.0293, 615.0203, 470.9874, 427.0006, 277.0165			Geraniin or isomer	CL
22	4.85	C15H18O9			341.0879	179.04	Caffeic acid hexoside	CL
23a	5.29	C30H26O12	579.1492	427.1009, 409.0905, 287.0540, 247.0593, 163.0375	577.1360	407.0776, 339.0887, 289.0720, 245.0824, 161.0241, 125.0245	Procyanidin B1	CL
24	5.77	C45H38O18	867.2125	579.1475,289.0709	865.2067	739.1741, 713.1604, 577.1408, 575.1241, 407.0800, 287.0586	Procyanidin C1 or isomer	CL
25a	5.88	C15H14O6	291.0865	207.0639, 165.0537, 147.0427, 139.0372, 123.0424	289.072, 579.1507	245.0834, 203.0724, 151.0408, 123.0456	(+)-catechin	CL
26	6.26	C21H32O10	445.2070	265.1419, 247.1320, 229.1204	443.1926	237.1516, 119.0355, 101.0224	Penstemide	NM
27	6.41	C27H22O18	635.0882	465.0350, 321.0038, 277.0165	633.0732	300.9982, 275.0195	Galloyl-HHDP-glucose or Strictinin	CL
28	6.50	C30H26O11	563.1499	291.0865,287.0556,273.0762	561.1420	289.0727, 245.0841	Fisetinidol-(4α,8)-catechin or isomer	CL
29	6.94	C30H26O11	563.1502	393.0964, 291.0865,287.0556,273.0762	561.1401	435.1067, 407.0756, 289.0712, 273.0760, 125.0244	Fisetinidol-(4α,8)-catechin or isomer	CL
30	7.24	C13H8O8			291.0149	247.0240, 219.0294, 191.0344, 175.0407	Brevifolincarboxylic acid	CL
31	7.86	C35H30O15	691.1658	539.0836, 521.0745, 359.0530, 301.0502, 239.0398	689.1522	537.1070, 519.0976, 399.0742, 289.0733, 237.0422	Unknown	CL
32	7.89	C21H10O13			469.0050	299.9910, 270.9899,	Flavogallonic acid	CL
33	7.97	C41H28O27	953.0893	917.0072, 615.0208, 452.9780, 277.0163			Geraniin or isomer	CL
34	8.02	C21H10O13			469.0050	299.9910, 270.9899,	Flavogallonic acid	CL
35	8.19	C20H18O9	403.1029	285.0755, 263.0543, 251.0535, 151.0373, 139.0374	401.0878	301.0696, 289.0699, 203.0703, 151.0394, 137.0236, 109.029	(Epi)catechin derivative	CL
36	8.23	C30H26O12	579.1495	561.1374,453.1176,409.0909,291.0861,287.0546			Procyanidin B2	CL
37	8.30	C20H14O14	479.0459	302.9928	477.0311	301.0007	Ellagic acid glucuronide	CL
38	8.38	C27H22O18			633.0737	481.0609, 463.0512, 343.0098, 300.9990, 275.0192, 169.0136, 125.0237	Galloyl-HHDP-glucose or Strictinin	CL
39	8.63	C14H24O10	353.1251	177.1010, 159.0909, 141.0084, 85.0586	351.1299	249.0619, 175.0266, 113.0245, 101.0610	3-Digitalose-quinic acid	AP
40	8.83	C19H30O8	387.2012	207.1366, 189.1269, 161.1315, 149.0948, 123.0793	385.1886	223.1355, 205.1253, 153.0932	Roseoside	CL
41	8.91	C27H30O15	595.1659	379.0800, 271.0621	593.1510	473.1084, 383.0796, 353.0671	Apigenin-6,8- di-C-glucoside	NM
42	9.11	C37H30O16	731.1193	579.0702, 441.0537, 411.0808, 271.0430, 153.0082	729.1458	407.0780, 289.0729, 169.0149, 125.0251	Gallocatechin-(4a, 8)-catechin	CL
43	9.21	C27H30O15			593.1551	285.04	Luteolin-7-O-rutinoside	NM
44	9.22	C41H28O26	937.0403	767.0235, 465.0373, 447.0262, 277.0162	935.0804, 467.0360	633.0739, 300.9988	Galloylpedunculagin	CL
45	9.30	C30H26O12	579.1161	561.1400,427.1025,409.0918,291.0861,289.0703			Procyanidin B3	CL
46	9.43	C28H32O16	625.1772	607.1631,589.1553,571.1418	623.1618	503.1187, 413.0869, 383.0763	Chrysoeriol 6,8-di-C-glucoside (stellarin-2)	NM
47	9.82	C20H16O12	449.0467	302.99	447.0573, 895.1224	299.99	Ellagic acid-4-O-rhamnoside	CL
48a	9.91	C14H6O8			300.9997	283.9963, 229.0142, 201.0192, 173.0245, 145.0296	Ellagic acid	CL
49a	10.36	C21H20O12	465.1027	303.0501	463.0907	300.0294, 271.0269, 255.0313	Que-3-gal* (Hyperoside)	CL
50	10.40	C27H32O15	597.1812	451.1236,435.1288,289.0704	595.1667	287.0562, 151.0039	Eriodictyol 7-O-rutinoside (Eriocitrin)	NM
51a	10.44	C21H20O12	465.1027	303.0494	463.0912	300.0298, 271.0272, 255.0312	Isoquercitrin	CL
52	10.49	C30H26O13			593.1303	447.0952, 285.0328, 255.0298, 145.0302	Tiliroside	NM
53	10.51	C23H36O2	345.2813	155.1330, 137.1232, 81.0645			Unknown	NM
54	10.99	C27H28O16			607.1309	463.0876, 300.0274, 271.0246, 151.0040	Quercetin 3-O-HMGlc	CL
55	11.51	C27H28O15			591.1358	529.1361, 489.1044, 447.0942, 285.0418	Kaempferol 3-O-HMGal	CL
56	11.70	C28H32O15	609.1814	463.1224,301.0705			Diosmin	NM
57a	11.78	C27H30O16	611.1620	303.07	609.1825	301.07	Rutin	NM
58	11.88	C29H32O17	653.1338	347.05	651.1572	589.1613, 549.1284, 507.1178, 345.0637	Limocitrin-HMG-glucoside	NM
59	11.91	C27H28O15			591.1358	529.1361, 489.1044, 447.0942, 285.0418	Kaempferol 3-O-HMGlc	CL
60	11.98	C16H18O12			401.0778	357.0866, 313.0975	Carboxyl derivative	CL, NM, AP
61	12.04	C30H34O18	683.1819	377.0870	681.1727	579.1342, 537.1236, 375.0715, 360.0479	Limocitrol 3-O-HMGlc	NM
62	12.57	C21H28O9	425.1808	389.1586, 371.1487, 245.1171, 233.1172, 227.1053, 215.1062	423.1689	279.1258, 205.1257, 139.0778	Grandidentatine	CL
63	13.56	C16H24O7	346.1671	311.1295,177.0285, 165.1166, 135.1170	327.1450	267.1264, 237.1147, 207.1049, 117.0206	Rhododendrin or Perilloside B or isomer	NM
64a	13.60	C15H10O7	303.0456	287.0735, 269.0628, 203.0204, 147.0344	301.0355	257.1551, 203.0169, 178.9988, 151.0036, 121.0295	Quercetin	CL
65	13.85	C16H16O6	305.1010	203.0337, 159.0415, 147.0440			Oxypeucedanin hydrate	NM
66	13.99	C16H24O7	346.1648	311.1487, 265.1263, 135.1170	327.1452	267.1250, 249.1136, 207.1035, 165.0931, 147.0809	Rhododendrin or Perilloside B or isomer	NM
67	14.28	C22H21NO5	380.1500	317.0815, 299.0727, 233.0296, 231.0140, 175.0273			Unknown	NM
68	15.61	C16H24O7	346.1648	135.11	327.1449	267.1262, 249.1150, 207.1044, 165.0938, 117.0204	Rhododendrin or Perilloside B or isomer	NM
69	16.27	C36H58O10	651.3716, 673.3924	489.3475, 471.3373,453.3276, 407.3232, 247.1645, 223.1646, 205.1546	695.4028	649.4010, 487.3446	Kajiichigoside F1 isomer	CL
70	16.81	C13H6N2O	207.0455	192.0281, 164.0351, 149.0125, 121.0556			Unknown	NM
71a	16.81	C36H58O10	673.3924, 651.4158	489.3566, 471.3455, 453.3352, 425.3403, 407.3297, 205.1579	695.4026	649.3975, 487.3436, 207.0524	Kajiichigoside F1	CL
72	17.01	C36H58O10	673.3913	511.3363,185.0419	695.4029	649.3975, 487.3447	Kajiichigoside F1 isomer	CL
73	17.01	C30H48O5	489.3576	471.3472,453.3364,425.3404,407.3297,389.3215,201.1639			Euscaphic acid or tormentic acid or isomer	CL
74	17.59	C36H58O10	673.3919	489.3563, 471.3452, 453.3344, 425.3403, 407.3287, 205.1574	695.4026	649.3987, 487.3442	Kajiichigoside F1 isomer	CL
75	18.62	C36H56O10			693.3854	647.3844, 485.3265	Periplocoside O	CL
76	20.48	C14H14O6	279.0787	201.0333, 173.0399, 149.0130			Columbianetin or isomer	CL, NM, AP
77	20.97	C36H56O9	633.3631	471.3156, 453.3058, 425.3134, 407.3035, 247.1528, 205.1458	631.3847, 677.3915	469.3318, 425.3425	Glycyrrhetinic acid-3-O-monoglucose	CL
78	21.24	C30H46O5			485.3272	467.3164, 425.3049, 375.3079	2,19-Dihydroxy-3-oxo-urs-12-en-28-oic acid isomer	CL
79	21.25	C18H14O6	327.0836	287.0899,203.0306			Rugosaflavonoid A	CL, NM, AP
80	21.37	C30H48O5	489.3581	471.3487, 453.3358,425.3417,407.3316,389.3215, 205.1582, 201.1631	487.3429	469.3320, 443.3518, 407.3309, 371.2963	Euscaphic acid	CL
81	21.86	C30H46O5			485.3272	467.3156, 425.3061, 375.3079	2,19-Dihydroxy-3-oxo-urs-12-en-28-oic acid	CL
82	22.17	C30H46O4			469.3310	451.3211, 407.3318	2a,3b-Dihydroxylup-20(29)-en-28-oic acid	CL

^a Compounds confirmed with references. CL: R. roxburghii; NM: C. limon; AP: M. domestica. HM, 3-hydroxy-3-methylglutaryl-; Rha, rhamnose; Rut, rutinoside; Gal, galactoside

Flavonoids A total of 26 flavonoids were found in CNJ, including 10 flavonones and 16 flavones, which were mainly originating from R. roxburghii, and C. limon. Among them, compounds 23, 25, 51, 57, and 64 were uniquely assigned as procyanidin B1, catechin, isoquercitrin, rutin, and quercetin, respectively, by comparing with reference substances. With the characteristic fragment ions of reference compounds, it can be learned that: Since most flavonones are derivatives of catechin, they can be preliminarily characterized by specific fragmentation ions in MS/MS spectra, such as m/z 291 in positive mode, and m/z 289, 245 in negative mode. Meanwhile, flavones have similar fragmentation pathways by the neutral loss of glucosyl (162 Da), rhamnosyl (146 Da), or rutinosyl (308 Da). Compounds 36 and 45 shared the same pseudo-molecular ion (m/z 579.14), and similar product ions (m/z 561.13, 409.09, 291.08) with reference 23, procyanidin B1. Thus, they were constitutional isomers of procyanidin B1. According to the referring retention time of these isomers found in R. roxburghii (Liu et al., 2016), they were identified as procyanidin B2 and B3.

Besides the neutral loss of sugar from the cleavage of glycosidic bond, dehydration, successive losses of CO from phenolic hydroxyl groups and ketone group, and Retro-Diels-Alder (RDA) fragmentation are also the most possible fragmentation pathways for flavones (Zhang et al., 2015). For example, compounds 49, 51, 54, and 57 all showed fragment ion at m/z 301.04 in negative mode, which is the same as deprotonated ion of quercetin (65). Thus, they might all share the same aglycone. Compound 49 gave a [M − H]⁻ ion at m/z 463.09 with the molecular formula C₂₁H₂₀O_12, indicating the constitutional isomer of isoquercitrin (51). Since isoquercitrin was a glycoside of quercetin with glucose, compound 49 might be glycoside with galactose-hyperoside (Porter et al., 2012). Compound 54 showed [M − H]⁻ ion at m/z 607.13, and fragment ion at m/z 463.09, 301.03, 151.01. Considering the 3-hydroxy-3-methyl-glutaryl (HM) specialized found in Rosa family (Porter et al., 2012), these fragment ions reasonably corresponded to the loss of an HM part ([M + H₂O − HM − H]⁻), further loss of glucose [M + 2H₂O − HM − Glc − H]⁻, and left A ring after RDA fragmentation of aglycon (Liu et al., 2016). Similarly, compounds 43, 50, 52, 55, 56, 58, 59 and 61 were annotated as luteolin 7-O-Rut, eriodictyol 7-O-Rut, tiliroside (Wang et al., 2020), kaempferol 3-O-HMGal, diosmin (Zhang et al., 2018), limocitrin-HMGlc, kaempferol 3-O-HMGlc, and limocitrol 3-O-HMGlc (Ledesma-Escobar et al., 2015), respectively. Two flavone C-glycosides, which were tentatively identified as apigenin 6,8- di-C-Glc (41) (Hwang and Ma, 2018) and chrysoeriol 6,8-di-C-Glc (46) (Guccione et al., 2016), were also discovered in C. limon and CNJ.

Phenolic acids 26 phenolic acids were found in CNJ, and primarily (21) originated from R. roxburghii (Table 2). Among them, compound 48 was unambiguously identified as ellagic acid with the authentic standard. Due to the fragment ions at m/z 300.99 ([M − H]⁻ of ellagic acid), compounds 37 and 47 were both proposed as ellagic acid derivatives. Compound 37 ([M−H]⁻, m/z 477.03) was then characterized as ellagic acid glucuronide as it yielded a fragment ion at m/z 301.00 (ellagic acid) by the loss of glucuronide (176.03). And compound 47 was assigned as ellagic acid-4-O-Rha due to the loss of rhamnose (146.06) (Zeng et al., 2017). Due to the presence of carboxyl, carbonyl, or hydroxyl groups in structures, loss of CO₂, CO and H₂O could be found in fragment ions of phenolic acids. Compounds 9, 10, 12, 16, and 30 yielded [M−H₂O−H]⁻ and [M−CO₂−H]⁻ fragment ions in MS² spectra. Based on their pseudo-molecular ions and literatures, these compounds were identified as p-coumaric acid, protocatechuic acid, gallic acid, 3, 4, 5-trihydroxycinnamic acid (Liu et al., 2016), and brevifolincarboxylic acid (Yisimayili et al., 2019), respectively.

Compound 11 was tentatively assigned as galloyl-glucose, because of its characteristic fragment ions (m/z 169.01, 125.02) of gallic acid and neutral loss of glucose (162.05). Compounds 17, 27, and 38 exhibited deprotonated molecules [M−H]⁻ at m/z 633.07 and produced base peak ion at m/z 300.99 in MS² spectrum, which suggested an ellagic acid unit released. Finally, they were identified as galloyl-HHDP-glucose, or strictinin, or isomers (Xu et al., 2018). And, compounds 32 and 34 produced ions at m/z 300.99 in MS² spectrum, which was also characterized as an ellagic acid unit and corresponding to neutral loss of 168 Da. Thus, they were identified as flavogallonic acid or isomer (Zeng et al., 2017). Moreover, compound 44 was tentatively identified as galloylpedunculagin according to its [M − H]⁻ at m/z 935.08 and fragment ions at m/z 765.05 [M − H − gallic acid]⁻, 633.07 [M − H − galloyl dimer]⁻, 463.05 [M − H − galloyl dimer − gallic acid]⁻, 300.99 [galloyl dimer]⁻ (Yang et al., 2020).

Triterpenoids The triterpenoids in CNJ were all originated from R. roxburghii, and most of them can be classified as euscaphic acids. Thus, after the prominent loss of H₂O or CO₂, the cleavage of D ring could be observed in the fragment ions of these compounds. Compounds 69, 71, 72, and 74 all showed the same pseudo-molecular ions [M − H]⁻ at m/z 695.40, indicating their chemical formula C₃₆H₅₈O₁₀. Among them, compound 71 was identified as kajiichigoside F1 by comparison of its retention time and MS/MS spectra with the standard. And the others also generated similar fragment ions with kajiichigoside F1, at m/z 649.39, 487.34 by the neutral losses of HCOOH and glucose groups. Thus, they were identified as kajiichigoside F1 isomers.

Compounds 73 and 80 displayed [M + H]⁺ ion at m/z 489.35 and [M − H]⁻ ion at m/z 487.34, corresponding to C₃₀H₄₈O₅. In MS² spectra, they generated fragment ions at m/z 471.34/469.33 and 425.34/423.32 by continuous losses of H₂O and HCOOH group. Finally, they were identified as euscaphic acid or its isomers by referring to the literature (Liu et al., 2016). With a similar fragmentation pathway, compound 82 was assigned as 2α, 3β-Dihydroxylup-20(29)-en-28-oic acid. Compounds 78 and 81 had the same molecular formula of C₃₀H₄₆O₅. Their fragment ions of m/z 467.31 and 425.30 indicated that they were isomers. By comparing their retention time on C18 column, compound 81 was identified as 2α, 19β-Dihydroxy-3-oxo-urs-12-en-28-oic acid, and the other one was its isomer (Liu et al., 2016).

Organic acids 9 organic acids were successfully identified in CNJ using negative ion mode and from three ingredients of CNJ. The characteristic fragment ions of these organic acids were [M − H − H₂O]⁻, [M – H − CO₂]⁻, and [M − HCOOH]⁻. Compound 1 with a deprotonated molecule at m/z 175.02, which produced ions at m/z 157.01, 129.01, 113.02, by neutral loss of H₂O, HCOOH, CO₂ groups, was identified as L-ascorbic acid by comparing with standard. According to the exact mass of their deprotonated ions and MS² fragment ions in previous reports, compounds 2, 3, 5, and 7 were tentatively identified as ribonic acid or xylonic acid, quinic acid, threonic acid, and malic acid, respectively (Liu et al., 2016; Zheng et al., 2020). In addition, compounds 6 and 8, which were isomers of each other, were acknowledged as citric acid, isocitric acid based on their retention time difference.

Others Compound 13 showed quasi-molecular ion at m/z 166.08 with the molecular formula of C₉H₁₁NO₂. Fragment ions at m/z 149.05, 131.04, 120.08 correspond to the neutral losses of NH₃, H₂O, HCOOH groups. Finally, it was identified as phenylalanine (Liu et al., 2016). Compound 65 obtained [M + H]⁺ ion at ions at m/z 305.10. With its biological source, Citrus genus which contained many coumarins, the high-intensity fragment m/z 203.03 indicated a neutral loss of the substituent [M + H − C₅H₁₀O₂]⁺, and the sequential cleavage to lose molecules of CO produced the typical fragments m/z 147.04 [203-2CO]⁺, therefore, it was tentatively identified as oxypeucedanin hydrate (Zhang et al., 2018).

Antioxidant activities of CNJ and its ingredients

Due to the complex nature of phytochemicals and different mechanism-based methods, the antioxidant capacities of CNJ and its ingredients were measured by three commonly used methods: DPPH, ABTS scavenging activity, and FRAP. Results of these evaluations were expressed as L-ascorbic acid equivalents and IC₅₀ values, which are shown in Table 3. The scavenging activity of free radicals determined by DPPH and ABTS⁺ assays showed that R. roxburghii and CNJ had higher AAEs and lower IC₅₀ values, indicating they had better antioxidant activities. And, the scavenging effects on DPPH radicals (Figure S20) suggested that extracts of C. limon and M. domestica scavenged DPPH radical in a dose-dependent way, while with the same concentration, R. roxburghii and CNJ almost tended to peak value. Therefore, as an important ingredient in CNJ, R. roxburghii might contribute more to the antioxidant activity of CNJ. With regard to the ferric reducing capacities, the trend was almost the same as that of the DPPH and ABTS assay. The result revealed that the order of antioxidant potency was: R. roxburghii > CNJ beverage > C. limon > M. domestica.

Table 3.

DPPH free-radical scavenging capacity of CNJ and its ingredients

Sample	DPPH		ABTS		FRAP (mmol Fe2+/100 g)
	(mg AAE/g)	IC50 (mg/mL)	(mg AAE/g)	IC50 (mg/mL)
CNJ	1.35 ± 0.38	11.3 ± 0.93	1.92 ± 0.93	20.5 ± 8.18	68.39 ± 12.51
CL	3.05 ± 0.21	4.5 ± 1.42	3.84 ± 1.06	8.3 ± 2.74	103.57 ± 37.26
NM	0.13 ± 0.06	73.5 ± 4.68	0.19 ± 0.10	79.4 ± 16.22	11.05 ± 3.08
AP	0.08 ± 0.03	80.1 ± 1.21	0.15 ± 0.07	91.3 ± 10.79	1.71 ± 0.73
L-Ascorbic acid	nt	0.026 ± 0.009	nt	0.035 ± 0.011	1882.72 ± 202.91

All values are expressed as mean ± SD of three independent measurements; AAE, L- ascorbic acid equivalents; nt, not tested

Screening potential antioxidants by pre-column UHPLC-DPPH

To find out the potential antioxidants in CNJ, the pre-column DPPH reaction coupled with UHPLC was used. After reaction, the antioxidants would be oxidized, leading to a structural change of them. Based on this, the peak intensity of the antioxidants would obviously decrease in the UHPLC chromatogram after with DPPH reaction. The UHPLC chromatograms of CNJ with or without DPPH-treated were shown in Fig. 2. The compounds with peak intensity decreased by more than 50% were considered as potential antioxidants. Finally, 18 potential antioxidant compounds (Table 4) were screened out and identified. Of these, six (procyanidin B1, (+)-catechin, catechin derivative, gallocatechin-(4α, 8)-catechin, rutin, limocitrin-HMGlc) are flavonoids; seven (mainly are ellagic acid and its derivatives, rhododendrin isomers) are phenolic acids; four (kajiichigoside F1 and its isomers) are triterpenoids; and one, penstemide. Moreover, 12 potential antioxidative compounds were originated from R. roxburghii, while 7 were from C. limon. The result further proved that R. roxburghii contributes more to the antioxidant activity of CNJ compared with others, and the selected compounds by pre-column UHPLC-DPPH might play an important role in its antioxidant activity.

Fig. 2.

Comparison of the chromatograms of the CNJ treated with DPPH-methanol solution and methanol in negative ion mode

Table 4.

Potential antioxidative compounds founded in CNJ

No.	Name	Intensity reduced (%)
23	Procyanidin B1	67
25	(+)-Catechin	52
26	Penstemide	64
35	Catechin derivative	71
37	Ellagic acid glucuronide	74
42	Gallocatechin-(4, 8)-catechin	75
47	Ellagic acid 4-O-rhamnoside	58
48	Ellagic acid	53
57	Rutin	71
58	Limocitrin-HMGlc	78
60	Carboxyl derivative	71
63	Rhododendrin or isomer	79
66	Rhododendrin or isomer	60
68	Rhododendrin or isomer	59
69	Kajiichigoside F1 isomer	64
71	Kajiichigoside F1	62
72	Kajiichigoside F1 isomer	70
74	Kajiichigoside F1 isomer	68

Contents of major antioxidants in CNJ and its ingredients

With the discovery of potential antioxidative compounds in CNJ, they should be quantified as much as possible to make sure its antioxidant activity. In this study, five selected potential antioxidants and four other compounds in CNJ were determined and quantified. The results of intra- and inter-day precision, accuracy, and recoveries Table S1 and S2 suggested this quantification method was validated with good reliability and sensitivity.

The content of these nine compounds in CNJ and its ingredients is shown in Table 1. From the results, the contents of most analytes in R. roxburghii were higher than other ingredients, which further proved its good antioxidant activity. And it was worthy that, comparing with other compounds, procyanidin B1 (23), (+)-catechin (25), ellagic acid (48), and kajiichigoside F1 (71) not only had good antioxidative capacities (decreased more than 50% after DPPH-treated, Table 4), but also had a higher amount in both CNJ (5.64, 8.21, 4.48, and 5.19 µg/mL) and R. roxburghii (164.81, 77.06, 75.41, and 156.87 µg/mL) (Table 1). Thus, these four compounds might contribute more to the antioxidant activity. By contrast, even though rutin also had a large decrease after DPPH reaction (71%), due to its low content (0.06 and 0.10 µg/mL in CNJ and R. roxburghii), it might contribute less. As the commonly used positive control, L-ascorbic acid was also quantified despite no significant decrease after the reaction. And L-ascorbic acid, vitamin C, did have much higher contents in both CNJ and R. roxburghii (329.44 and 1819.63 µg/mL, Table 1).

As we know, the components of natural products are not only highly complex, but also with wide scales of content. Furthermore, there is still a lack of knowledge on the antioxidant capacities of most components, especially those unidentified compounds. Therefore, it is really difficult to find out its “real” antioxidative compounds of natural products or antioxidants. In this study, the chemical profiling and antioxidant capacity of CNJ were analyzed, as an example. Although there is still a lack of information on antioxidative capacities (IC₅₀) of most compounds, the degree to which they react with DPPH (percentage of peak intensity decrease after reaction) can somehow explain it. With the chemical identification, and further quantitation, it may provide some knowledge on predicting the antioxidative compounds of natural products.

Finally, for the first time, the chemical composition profile of CNJ beverage was analyzed, as well as its antioxidant activity. By using UHPLC-Q-TOF/MS analysis, a total of 82 compounds were identified or tentatively deduced, mainly consisting of flavonoids, phenolic acids, triterpenoids, and organic acids. And, the antioxidant capacities of CNJ and its ingredients were assessed by different mechanism-based assays. The result suggested that the ingredient, R. roxburghii, plays a critical role in the antioxidant activity of CNJ. Furthermore, 18 potential antioxidants in CNJ were screened out by pre-column DPPH-UHPLC. Based on the identification, these antioxidants are considered mainly as catechin and its derivatives, ellagic acid and its derivatives, and kajiichigoside F1 or isomers. This study not only provides information on the chemical composition of CNJ and its antioxidant activity but also most importantly, by further quantification of these antioxidants in CNJ, it helps us find out the “real antioxidative compounds” in it. This study may provide a practical strategy for rapid screening out antioxidants in natural plants or foods and a better understanding of their antioxidative activities.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Author contributions

Yida Zhang, Peiyan Zheng and Guanyu Yan contributed equally to this work. Yida Zhang analyzed the data, interpreted the results, and drafted the manuscript. Guanyu Yan performed the experiments. Yue Zhuo provided guidance for experiments. Peiyan Zheng critically revised the manuscript. Jian-lin Wu and Baoqing Sun conceived and designed the study. All authors read and approved the final manuscript.

Declarations

Conflict of interest

The authors declare no conflict of interest.