Quality evaluation of Lonicerae Japonicae Flos and Lonicerae Flos based on simultaneous determination of multiple bioactive constituents combined with multivariate statistical analysis.

INTRODUCTION: Lonicerae Japonicae Flos (LJF) and Lonicerae Flos (LF) belong to different genera of Caprifoliaceae. They have been historically utilised as herbal medicine to treat various diseases. However, the comprehensive assessment of them still remains a challenge.
OBJECTIVE: To develop a comprehensive method of ultra-fast liquid chromatography-tandem triple quadrupole mass spectrometry (UFLC-QTRAP-MS/MS) coupled with multivariate statistical analysis for the quality evaluation and reveal differential components of LJF and LF.
METHODOLOGY: A validated UFLC-QTRAP-MS/MS method was established for simultaneous determination of 50 constituents, including 12 organic acids, 12 flavonoids, 6 iridoids, 3 saponins, 13 amino acids and 4 nucleosides. The obtained data were employed to multivariate statistical analysis. Principal component anlysis (PCA) and partial least squares determinant analysis (PLS-DA) were performed to classify and reveal differential components of samples; grey relational analysis (GRA) was introduced to assess the samples according to the contents of 50 constituents by calculating the relative correlation degree of each sample.
RESULTS: Fifty constituents were simultaneously determined of LJF and LF. Based on obtained data, PCA and PLS-DA were easy to distinguish samples and the classification of the samples was related to 11 chemical constituents. GRA implied the quality of LJF was better, and that the flower buds were superior to the flowers. Moreover, organic acids are the main components of samples.
CONCLUSION: This study not only established a method of simultaneous determination of multiple bioactive constituents in LJF and LF, but provided comprehensive information on the quality control of them. The developed method is conducive to distinguish orthologues or paralogues of them, and supply the support for "heterologous effects".

INTRODUCTION

Lonicerae Japonicae Flos (LJF) and Lonicerae Flos (LF), commonly known as Jin‐yin‐hua and shan‐yin‐hua in China, respectively, are traditional Chinese medicines (TCMs) with widespread use in the medicine industry. More than 500 prescriptions contain LJF for treatment of carbuncles, furuncles, erysipelas, sores, swelling and affections caused by exopathogenic wind‐heat, epidemic febrile diseases at the early period. Modern pharmacological research has confirmed its high medicinal value, such as anti‐inflammatory, , , anti‐bacterial, anti‐oxidant, hepatoprotective, anti‐viral and other biological activities. Meanwhile, clinical practice in recent years also showed LJF possesses a preventive effect on the severe acute respiratory syndrome (SARS) virus and H1N1 influenza virus. , In China, there is a notion of “Let food be thy medicine and medicine be thy food”. LJF and LF also have been used to make tea, food, and beverages due to their heat‐clearing and health care properties.

LJF and LF have been documented as independent projects according to plant morphology, medicinal properties and chemical composition in Chinese Pharmacopoeia since 2005. Thunb. is determined as the only plant source of LJF. Whereas, LF has four sources of germ plasm including Lonicera macranthoides Hand.‐Mazz., L. hypoglauca Miq., DC. and L. fulvotomentosa Hsu et S.C. Cheng in current Chinese Pharmacopoeia. They are typical Chinese medicines of “heterologous effect”. Numerous literature has reported that LF has similar pharmacological activities as LJF, , only with a marginal difference in the strength of the drug effect. But until now, there is still no recognised theory to explain this phenomenon. Therefore, owing to the close proximity of plant species, the similar appearance and function, the quality standards and mutual substitution of LF and LJF still remain controversial.

Chemical composition is the basis of the pharmacological action of medicinal materials. In recent years, phytochemical studies have revealed that more than 200 components have been identified from , its main constituents include essential oils, flavonoids, organic acid, iridoids and saponins,12, 13 which modulate multiple functions. For instance, the volatile oil possesses antifungal activity; flavonoids and organic acid have strong inhibitory activity against multiple pathogens, anti‐oxidant, and anti‐tumour; iridoids and saponins also have shown excellent activity of anti‐tumour, anti‐inflammation, antioxidant and hepatoprotective; amino acid is an essential nutrient that is beneficial to human body; it exhibits a variety of physiological activities, including anti‐platelet aggregation, antioxidant and immunity enhancement; nucleosides are the major components of nucleic acids. Some nucleosides and their derivatives have significant physiological functions, for example inosine can be used to treat acute and chronic hepatitis, rheumatic heart disease. Therefore, it is certain that excellent clinical efficacy arises from the synergistic effects of complicated chemical components. In Chinese Pharmacopoeia 2015 edition, chlorogenic acid and luteoloside have been used as biomarkers to characterise the quality of LJF. Notably, some other medicinal plants have also been found to contain high contents of chlorogenic acid and luteoloside, and to determine single or several bioactive compounds in herbal medicines is one‐sided with respect to the wholeness of TCM meaning multi‐components at multi‐targets. Only chlorogenic acid and luteoloside might not completely inflect the quality of LJF, and owing to the mixing of different medicinal parts, unconsciously mislabelled or confused of LJF and LF, the quality and safety of LJF and LF cannot be guaranteed. Therefore, an effective and reliable method is necessary to be established to evaluate the quality of LJF and LF. Meanwhile, it is meaningful to find the distinguishing chemical markers and identify which part (the flower bud or the flower) is the optimum quality of LJF.

Many methodologies have been reported to control the quality of LJF and LF. For instance, high‐performance liquid chromatography with ultraviolet detector (HPLC‐UV) was used for determination of phenolic acid; , , HPLC with evaporative light scattering detector (HPLC‐ELSD) was employed for quantitative saponins and iridoid glucosides. , , Nevertheless, the earlier methods only could detect an individual active ingredient which made the assay less efficient. HPLC‐diode array detector (DAD)‐ELSD , was reported to detect flavonoids, iridoids, phenolic acids, and saponins in LJF and LF. However, there are still several shortcomings such as the relatively low sensitivity of ELSD and the inaccuracy of chromatographic peaks only determined by retention time. With the development of analytical technology, HPLC‐DAD‐electrospray ionisation mass spectrometry (ESI‐MS) was proposed to analyse multiple types of bioactive constituents. , It has been reported that HPLC‐ESI‐time‐of‐flight (TOF)‐MS was used to quantify 32 bioactive compounds in Lonicera species. But these methods mainly focus on quantification. Nowadays, ultra‐fast liquid chromatography‐tandem triple quadrupole mass spectrometry (UFLC‐QTRAP‐MS/MS) has been used to identify the content of TCM due to its high sensitivity and informativity. ,

In the present article, we attempted to develop a comprehensive and reliable method of UFLC‐QTRAP‐MS/MS coupled with multivariate statistical analysis for the quality evaluation of LJF and LF. This method could simultaneously detect 50 constituents, including 12 organic acids, 12 flavonoids, 6 iridoids, 3 saponins, 13 amino acids and 4 nucleosides in 35 batches of LJF and LF samples from different habitats and commercial herbs. Moreover, principal component analysis (PCA) and partial least squares discriminant analysis (PLS‐DA) was performed to distinguish the LJF and LF. Grey relational analysis (GRA) was introduced to assess the quality of LJF and LF according to the contents of the tested constituents. Our study not only established a method of simultaneous determination of multiple bioactive constituents of LJF and LF, but provided comprehensive information on the quality control of them. The developed method is conducive to distinguish orthologues or paralogues of them, and provide the support for “heterologous effects”. To the best of our knowledge, our study is the most comprehensive methodology on the quantitative comparative analysis of LJF and LF, which has a high scientific and reference value.

MATERIALS AND METHODS

Chemicals, reagents, and materials

Fifty chemical standards including chlorogenic acid (1), neochlorogenic acid (2), cryptochlorogenic acid (3), 3,5‐O‐dicaffeoylquinic acid (4), 3,4‐O‐dicaffeoylquinic acid (5), 4,5‐O‐dicaffeoylquinic acid (6), 1,3‐O‐dicaffeoylquinic acid (7), caffeic acid (8), quinic acid (9), protocatechuic acid (10), ferulic acid (11), 4,5‐O‐dicaffeoylquinic acid methyl ester (12), rutin (13), hyperoside (14), luteoloside (15), luteolin (16), rhoifolin (17), diosmetin (18), apigenin (19), kaempferol (20), astragalin (21), lonicerin (22), kaempferol‐3‐O‐rutinoside (23), isoquercitrin (24), sweroside (25), secologanic acid (26), loganin (27), secoxyloganin (28), loganin acid (29), morroniside (30), macranthoidin A (31), dipsacoside B (32), akebia saponin D (33), l‐alanine (34), l‐serine (35), l‐proline (36), l‐valine (37), l‐threonine (38), l‐isoleucine (39), l‐leucine (40), l‐aspartic acid (41), l‐glutanmate (42), l‐lysine (43), l‐histidine (44), l‐phenylalanine (45), l‐arginine (46), cytidine (47), uridine (48), adenosine (49) and inosine (50) were detected in this experiment. The purity of all standard components was more than or equal to 98%. The structures of the standard substances are shown in Supporting Information Figure S1. Among them, 9, 13, 14, 21 and 24 were purchased from the Control of Pharmaceutical and Biological Products (Beijing, China); 18–20, 23, 25 and 33 were offered by Chengdu Chroma Biotechnology Co. Ltd (Sichuan, China); 1–5, 8, 11, 27 and 34–50 were obtained from Shanghai Yuanye Biotechnology Co. Ltd (Shanghai, China); 6, 7 and 10 were received from Chengdu Prefa Technology Development Co. Ltd (Sichuan, China); 12, 15–17, 22, 26, and 29–31 were provided by Liangwei Chemical Reagent Co. Ltd (Nanjing, China); 28 and 32 were acquired from Nanjing Jingzhu Biotechnology Co. Ltd (Nanjing, China).

In the experiment, chromatographic grade methanol and acetonitrile were purchased from Merck (Darmstadt, Germany); other analytical grade solvents were purchased from Liangwei Chemical Reagent Co. Ltd. Ultrapure water was obtained in Milli‐Q purifying system (Millipore, Bedford, MA, USA); samples were collected in 2018, samples 1–10 are LF (Lonicera macranthoides Hand.‐Mazz.), samples 11–21 are flower buds of Lonicerae Japonicae and samples 22–35 are flowers of Lonicerae Japonicae. Detailed information on these samples is listed in Table 1. The botanical origins of the materials were identified by one of our authors, Professor Xunhong Liu. Voucher specimens were deposited in the Herbarium of Pharmacy, Nanjing University of Chinese Medicine, China.

TABLE 1

Information of Lonicerae Japonicae Flos (LJF) and Lonicerae Flos (LF)

Species	Sample number	Batch number	Habits	Origin
Lonicerae Flos (Lonicera macranthoides Hand.‐Mazz.)	1	2018110306	Hunan	Shaodong Lianqiao
	2	20181103061	Hunan	Shaodong Lianqiao
	3	2018110701	Hunan	Longhui
	4	20181107011	Hunan	Longhui
	5	2018110305	Shanxi	Shangluo Danfeng
	6	20181103051	Shanxi	Shangluo Danfeng
	7	2018110308	Hunan	Shaodong Lianqiao
	8	20181103081	Hunan	Shaodong Lianqiao
	9	2018110703	Hunan	Longhui
	10	20181107031	Hunan	Longhui
Lonicerae Japonicae Flos	11	20181108	Shandong	Local collection
	12	2018110302	Hebei	Juluxian Gouqijinyinhua market
	13	2018110506	Henan	Fengqiu
	14	2018110502	Shandong	Linyi
	15	180401	Henan	Anhui YaoZhiyuan TCM decoction co., ltd.
	16	2018110603	Henan	Fengqiu
	17	2018110604	Henan	Fengqiu
	18	C16011901	Shandong	Zhejiang Yetongren pharmaceutical co., ltd.
	19	2018110503	Shandong	Linyi
	20	20181106011	Henan	Fengqiu
	21	20181106012	Henan	Fengqiu
	22	2018110505	Shandong	Linyi
	23	2018110302	Hebei	Juluxian Gouqijinyinhua market
	24	20181103021	Hebei	Juluxian Gouqijinyinhua market
	25	20181103022	Hebei	Juluxian Gouqijinyinhua market
	26	2018110301	Hebei	Juluxian Gouqijinyinhua market
	27	1803151	Shandong	Suzhou Boyuan pharmaceutical industry
	28	20181103011	Hebei	Juluxian Gouqijinyinhua market
	29	2018110504	Shandong	Linyi
	30	20181109		Local collection
	31	2018110303	Hebei	Juluxian Gouqijinyinhua market
	32	20181107	Shandong	Local herbal medicine market
	33	170802	Shandong	Bozhou Beshixin traditional Chinese medicine slice co., ltd.
	34	20181106	Shandong	Local herbal medicine market
	35	171116	Shandong	Shanghai medicine holdings Yixing co., ltd.

Samples 1–10 are LF (Lonicera macranthoides Hand.‐Mazz.); samples 11–21 are Flower buds Lonicerae Japonicae; samples 22–35 are Flowers Lonicerae Japonicae.

Preparation of standard solutions

Fifty standard substances were prepared by dissolving in 70% methanol, and their concentrations were as follows: (1) 10.05 mg/mL, (2) 1.04 mg/mL, (3) 5.13 mg/mL, (4) 1.24 mg/mL, (5) 1.01 mg/mL, (6) 10.09 mg/mL, (7) 10.06 mg/mL, (8) 5.05 mg/mL, (9) 0.99 mg/mL, (10) 10.01 mg/mL, (11) 1.02 mg/mL, (12) 1.05 mg/mL, (13) 1.33 mg/mL, (14) 1.04 mg/mL;, (15) 1.01 mg/mL, (16) 5.09 mg/mL, (17) 1.28 mg/mL, (18) 1.02 mg/mL, (19) 1.05 mg/mL, (20) 1.00 mg/mL, (21) 1.17 mg/mL, (22) 1.01 mg/mL, (23) 1.01 mg/mL, (24) 1.01 mg/mL, (25) 5.08 mg/mL, (26) 5.04 mg/mL, (27) 5.08 mg/mL, (28) 5.07 mg/mL, (29) 5.07 mg/mL, (30) 5.10 mg/mL, (31) 5.04 mg/mL, (32) 10.09 mg/mL, (33) 4.99 mg/mL, (34) 5.09 mg/mL, (35) 5.09 mg/mL, (36) 5.09 mg/mL, (37) 1.32 mg/mL, (38) 1.06 mg/mL, (39) 1.03 mg/mL, (40) 1.06 mg/mL, (41) 1.04 mg/mL, (42) 1.07 mg/mL, (43) 1.00 mg/mL, (44) 1.00 mg/mL, (45) 1.01 mg/mL, (46) 1.01 mg/mL, (47) 1.38 mg/mL, (48) 0.68 mg/mL, (49) 1.03 mg/mL, (50) 1.59 mg/mL. The mixed standard stock solution containing 50 standard substances was serially diluted with 70% (v/v) methanol to require concentrations for the establishment of the calibration curves. All solutions were stored at 4°C, and then filtered through 0.22 μm membranes (Jinteng laboratory equipment, Tianjin, China) before analysis.

Preparation of sample solutions

All of the samples were pulverised and sieved through the 50‐mesh. The accurately weighed powder (1.0 g) was extracted by ultrasonication in 40 mL 70% methanol for 45 min, then cooled at room temperature; the same solution was used to replenish the extraction system upon solvent loss because of volatilisation. The mixture was centrifuged at 12000 rpm for 10 min, then filtered through a 0.22 μm membrane prior to analysis.

UFLC‐QTRAP‐MS/MS instrumentation and conditions

All samples were analysed using UFLC system (SHIMADZUDGU Corp., Kyoto, Japan) with a triple quadrupole‐linear ion trap mass spectrometer (QTRAP‐5500) (AB SCIEX, Framingham, MA, USA). The separation was performed using the X Bridge ®C18 (4.6 mm × 100 mm, 3.5 μm) (Waters, Wexford, Ireland). The mobile phase was composed of 0.2% aqueous formic acid (A) and acetonitrile with 0.2% formic acid (B) at the flow rate of 0.8 mL/min. The gradient elution as follows: 0–5 min: 2% B; 5–10 min: 2–13% B; 10–12 min: 13% B; 12–17 min: 13–25% B; 17–25 min: 25–33% B; 25–27 min: 33–35% B; 27–29 min: 35–50% B; 29–31 min: 50–95% B. The re‐equilibration time was 4 min, and the injection volume was 1 μL. The ESI‐MS spectra were acquired in the multiple reaction monitoring (MRM) mode under both positive and negative ion modes. The conditions for the ESI‐MS analysis were set as follows: pressure of nebuliser of MS, 4500 V (positive) and −4500 V (negative); gas temperature 550°C; GS1 flow 55 L/min; GS2 flow 55 L/min; CUR flow 40 L/min; all MS data were analysed by the Analyst 1.6.2 software (AB SCIEX, Framingham, MA, USA).

Validation of UFLC‐QTRAP‐MS/MS method

The proposed method was validated according to linearity and range, the limit of detection (LOD), limit of quantitation (LOQ), precision (intra‐day and inter‐day), repeatability, stability, accuracy, and matrix effect. The mixed standard stocked solution containing 50 reference substances were serially diluted with 70% methanol to require concentrations for the establishment of calibration curves. The LODs and LOQs of these analytes under the present chromatographic conditions were determined at signal‐to‐noise (S/N) ratio equalled to 3 and 10, respectively. Precision was evaluated by injecting six replicates the same sample solution within one day and three days. Six independent sample solutions from the same sample were analysed to ensure the repeatability. The same sample solution was injected at 0, 2, 4, 8, 12, and 48 h to evaluate the stability of the instrument. A recovery test was used to check the accuracy of the method. The matrix effect was evaluated by the slope (slope matrix/slope solvent) comparison method.

Multivariate statistical analysis

PCA is used to visualise similarities or differences in multivariate data, which is an unsupervised pattern recognition technique. It is a method to pass multiple variables through a linear transformation to select fewer important variables and has been widely used in the differentiation and identification of medicinal materials. In order to observe the classification of LJF and LF, the data of 50 analytes were used to carry out PCA using SIMCA‐P 13.0 software (version 13.0, Umetrics AB, Umea, Sweden).

To find out the different chemical composition between JLF and JF, supervised PLS‐DA was performed by SIMCA‐P 13.0 software (version 13.0, Umetrics AB).

GRA is an effective and quantitatively comparative analysis method, which uses grey correlation degree to describe the strength, size, and order of the relationship between factors based on the sample data of each factor. It was introduced to assess the quality of LJF and LF based on the contents of 50 analytes by calculating the relative correlation degree of each sample.

RESULTS AND DISCUSSION

Optimisation of extraction conditions

In order to obtain quantitative extraction, various factors are utilised and optimised including extraction solvent (100% (v/v) methanol, 70% (v/v) methanol, 25% (v/v) methanol), time (60 min, 45 min, 20 min), method (ultrasonic and refluxing), and the solvent‐to‐sample ratios (100:1, 40:1, 25:1 (v/w)). The results showed that the extraction efficiency of samples in 70% (v/v) methanol solution at room temperature is optimal, and suitable solvent‐to‐sample ratio and ultrasonic time were 40:1 and 45 min, respectively.

Optimisation of UFLC‐QTRAP‐MS/MS conditions

Chromatographic conditions were optimised. Four common important parameters including three types of columns [X Bridge ®C18 (4.6 mm × 100 mm, 3.5 μm), Agilent ZORBAX SB C18 column (250 mm × 4.6 mm, 5 μm) (Agilent, Palo Alto, CA, USA), Thermo Acclaim TM RSLC 120 C18 (150 mm × 2.1 mm,2.2 μm) (Thermo Scientific, Waltham, MA, USA)], different mobile phases (water/acetonitrile, water/methanol, 0.1% aqueous formic acid/acetonitrile, 0.2% aqueous formic acid/0.2% formic acid acetonitrile), flow rate (0.3 mL/min, 0.8 mL/min, 1.0 mL/min), and column temperatures (25, 30, and 35°C) were investigated. Considering the strong hydrophilicity of organic acids, amino acids and nucleosides, the column of X Bridge ®C18 (4.6 mm × 100 mm, 3.5 μm) was chosen. Water/acetonitrile system showed more powerful resolution than water/methanol system. Furthermore, when the mobile phase was added with formic acid, the shape and symmetry of chromatographic peak of organic acids were significantly improved. In addition, the column temperature also affects the chromatographic separation. As a result, 0.2% aqueous formic acid/0.2% formic acid acetonitrile at the flow rate of 0.8 mL/min at 30°C on X Bridge ®C18 (4.6 mm × 100 mm, 3.5 μm) was selected and applied.

In order to effectively distinguish isomers by high resolution mass spectrometry and secondary mass spectrometry fragments, the individual solutions of all standard compounds (100 ng/mL in 70% (v/v) methanol) were injected into the ESI source in the positive and negative ion modes to get more suitable de‐clustered voltage (DP) and collision energy (CE) parameters. The most abundant fragment ions were chosen as MRM transition from MS/MS spectrum; after trial and error inspection, most constituents had a good response in the negative ion mode; only kaempferol‐3‐O‐rutinoside, amino acids, and nucleosides respond better in positive ion mode. Optimal values of 50 analytes are summarised in Table 2 and the chromatograms with MRM mode are presented in Figure 1. However, as shown in Table 2 chlorogenic acid, neochlorogenic acid and cryptochlorogenic acid as isomer; 3,5‐O‐dicaffeoylquinic acid, 3,4‐O‐dicaffeoylquinic acid, 4,5‐O‐dicaffeoylquinic acid, and 1,3‐O‐dicaffeoylquinic acid as isomer; hyperoside, isoquercitrin as isomer; luteoloside, astragalin as isomer; l‐leucine, l‐isoleucine as isomer; lonicerin, kaempferol‐3‐O‐rutinoside as isomer; they have the same precursor ion‐product ion pairs, respectively. Except for lonicerin, kaempferol‐3‐O‐rutinoside can be discriminated between different ion modes; other isomer reference substances were sequentially injected into QTRAP‐MS/MS to determine the compound based on the different retention time.

TABLE 2

Optimised mass spectrometric parameters of 50 compounds

	Name	CAS no.	Formula	t R (min)	MV	MRM (precursor→product)	DP (V)	CE (eV)
1	Chlorogenic acid	327–97‐9	C16H18O9	18.75	354.31	305.01/125	−35	−20
2	Neochlorogenic acid	906–33‐2	C16H18O9	17.64	354.31	305.01/125	−80	−26
3	Cryptochlorogenic acid	905–99‐7	C16H18O9	19.86	354.31	305.01/125	−95	−20
4	3,5‐O‐Dicaffeoylquinic acid	2450‐53‐5	C25H24O12	20.15	516.45	515.1/191	−85	−22
5	3,4‐O‐Dicaffeoylquinic acid	14534–61‐3	C25H24O12	20.13	516.45	514.989/353	−80	−26
6	4,5‐O‐Dicaffeoylquinic acid	57378–72‐0	C25H24O12	20.27	516.45	515.1/191	−75	−24
7	1,3‐O‐Dicaffeoylquinic acid	19870–46‐3	C25H24O12	20.3	516.45	514.980/190.979	−95	−24
8	Caffeic acid	331–39‐5	C9H8O4	19.45	180.16	179.03/134.6	−125	−20
9	Quinic acid	77–95‐2	C7H12O6	18.75	192.17	191.099/84.981	−195	−28
10	Protocatechuic acid	99–50‐3	C7H6O4	12.99	154.12	152.9/109	−85	−16
11	Ferulic acid	1135‐24‐6	C10H10O4	23.89	194.18	193.017/134	−50	−10
12	4,5‐O‐Dicaffeoylquinic acid methyl ester	114637–83‐1	C26H26O12	29.5	530.47	529.194/135.001	−85	−42
13	Rutin	153–18‐4	C27H30O16	22.06	610.52	609.06/300	−245	−46
14	Hyperoside	482–36‐0	C21H20O12	22.74	464.38	463.003/299.9	−160	−36
15	Luteoloside	5373‐11‐5	C21H20O11	23.03	448.38	447.117/284.963	−300	−36
16	Luteolin	491–70‐3	C15H10O6	29.53	286.24	285.086/132.980	−170	−40
17	Rhoifolin	17306–46‐6	C27H30O14	24.82	578.52	577.185/268.958	−65	−46
18	Diosmetin	520–34‐3	C16H12O6	30.88	300.26	298.938/283.929	−215	−30
19	Apigenin	520–36‐5	C15H10O5	30.74	270.24	268.8/116.9	−129	−40
20	Kaempferol	520–18‐3	C15H10O6	30.88	286.24	285.0/116.9	−120	−36
21	Astragalin	480–10‐4	C21H20O11	24.4	448.38	447.1/283.9	−100	−36
22	Lonicerin	25694–72‐8	C27H30O15	23	594.52	593.146/283.984	−200	−54
23	Kaempferol‐3‐O‐rutinoside	17650–84‐9	C27H30O15	23.59	610.52	595/287.2	36	25
24	Isoquercitrin	482–35‐9	C21H20O12	22.5	464.38	463.015/300	−180	−36
25	Sweroside	14215–86‐2	C16H22O9	20.14	358.34	357.213/124.985	−65	−20
26	Secologanic acid	60077–46‐5	C16H22O10	19.17	376.36	357.107/212.956	−170	−22
27	Loganin	18524–94‐2	C17H26O10	19.72	390.38	389.262/226.980	−40	−12
28	Secoxyloganin	58822–47‐2	C17H24O11	21.1	404.37	403.219/120.973	−135	−32
29	Loganin acid	22255–40‐9	C16H24O10	18.07	376.36	375.107/212.956	−170	−22
30	Morroniside	25406–64‐8	C17H26O11	18.53	406.38	405.235/243	−100	−14
31	Macranthoidin A	140360–29‐8	C59H96O27	30.05	1237.4	1235.464/911.43	−235	−48
32	Dipsacoside B	33289–85‐9	C53H86O22	30.19	1075.2	1073.466/749.384	−250	−40
33	Akebia saponin D	39524–08‐8	C47H76O18	30.32	929.1	927.445/603.34	−155	−50
34	l‐Alanine	56–41‐7	C3H7NO2	1.38	89.09	90.06/44.02	100	10
35	l‐Serine	56–45‐1	C 3 H 7 NO 3	1.38	105.09	106.05/59.99	100	8
36	l‐Proline	147–85‐3	C 5 H 9 NO 2	1.65	115.13	116.07/70.02	68	10
37	l‐Valine	72–18‐4	C5H11NO2	2.32	117.15	118.09/72.06	100	10
38	l‐Threonine	72–19‐5	C4H9NO3	1.38	119.12	120.07/74	100	20
39	l‐Isoleucine	73–32‐5	C6H13NO2	4.96	131.17	132.1/86.05	64	10
40	l‐Leucine	61–90‐5	C 6 H 13 NO 2	5.4	131.17	132.1/86.05	100	16
41	l‐Aspartic acid	56–84‐8	C 4 H 7 NO 4	1.38	133.1	134.05/87.96	59	10
42	l‐Glutamate	138–18‐1	C5H7NO4	1.24	145.11	147.08/83.92	100	16
43	l‐Lysine	56–87‐1	C6H14N2O2	1.25	146.19	147.11/83.91	100	14
44	l‐Histidine	71–00‐1	C 6 H 9 N 3 O 2	1.24	155.15	156.08/110.03	100	16
45	l‐Phenylalanine	63–91‐2	C 9 H 11 NO 2	13	165.19	166.1/120.05	100	14
46	l‐Arginine	74–79‐3	C6H14N4O2	1.36	174.2	175.12/70.02	100	18
47	Cytidine	65–46‐3	C 9 H 13 N 3 O 5	1.65	243.22	244.09/112	61	10
48	Uridine	58–96‐8	C 9 H 12 N 2 O 6	4.25	244.2	244.896/113	10	13
49	Adenosine	58–61‐7	C 10 H 13 N 5 O 4	6.73	267.24	268.1/136.07	86	23
50	Inosine	58–63‐9	C 10 H 12 N 4 O 5	9.62	268.22	269/137.07	46	15

FIGURE 1

Multiple‐reaction monitoring (MRM) chromatogram of 50 compounds [Colour figure can be viewed at wileyonlinelibrary.com]

Analytical method validation

Validation results of the method are shown in Table 3. All calibration curves showed good linearity (r > 0.9990) within the test range. The method also provided satisfactory sensitivity for all analytes; the LODs and LOQs ranged 0.005–139.226 ng/mL and 0.015–417.797 ng/mL, respectively. Percentage relative standard deviation (RSD%) values of intra‐day, inter‐day, repeatability, stability test of the 50 analytes were all less than 5%. The mean recoveries fell between 94.2% and 104.62%, with the RSD% values less than 4.79%, and the slope ratio values of the matrix curve to pure solution curve were between 0.94 and 1.05. All the earlier‐mentioned results demonstrated the credibility of the developed method.

TABLE 3

Regression equation, limit of detection (LOD), limit of quantitation (LOQ), intra‐ and inter‐day precision, repeatability, stability, recovery and matrix effect of 50 compounds

Name	Regression equation	r	Linear range (ng/mL)	LOD (ng/mL)	LOQ (ng/mL)	Precision		Repeatability	Stability	Recovery		Matrix effect
						Intra‐day (RSD%; n = 6)	Inter‐day (RSD%; n = 3)	(RSD %; n = 6)	(RSD%; n = 6)	Mean	RSD%
Chlorogenic acid	y = 845x + 4.77e5	0.9999	0.604–755000	0.195	0.585	1.78	1.79	1.55	3.33	99.70	0.23	1.01
Neochlorogenic acid	y = 115x + 1.13e5	0.9992	3.01–376000	0.388	1.163	1.83	1.71	0.88	1.61	96.70	3.97	1.00
Cryptochlorogenic acid	y = 878x + 6.7e3	0.9993	0.316–79000	0.090	0.271	4.04	4.31	2.42	2.47	100.66	1.62	1.04
3,5‐O‐Dicaffeoylquinic acid	y = 1.02e3 x + 1.41e5	0.9996	77.6–388000	14.483	43.449	2.63	2.79	0.72	1.62	98.70	2.65	0.98
3,4‐O‐Dicaffeoylquinic acid	y = 602x + 1.45e5	0.9995	7.63–38200	2.094	6.282	3.95	4.39	3.14	4.94	100.67	3.65	1.02
4,5‐O‐Dicaffeoylquinic acid	y = 483x + 2.1e4	0.9998	45–28100	7.918	23.755	3.27	3.64	1.16	1.62	98.40	2.12	1.01
1,3‐O‐Dicaffeoylquinic acid	y = 5.78e3 x + 5.62e4	0.9996	42–2630	6.971	20.913	3.44	3.28	2.82	3.16	100.92	3.43	1.04
Caffeic acid	y = 2.49e3 x + 1.69e5	0.9998	5.76–7200	1.169	3.506	2.86	3.17	3.29	4.83	101.17	2.05	1.03
Quinic acid	y = 1.19e3 x + 3.05e5	0.9996	19.6–245000	5.131	15.392	2.73	2.38	1.69	4.18	100.20	2.48	1.05
Protocatechuic acid	y = 6.13e3 x + 1.95e5	0.9994	7.68–960	0.780	2.341	2.50	2.30	4.09	3.85	99.50	2.82	0.96
Ferulic acid	y = 121x + 3.68e3	0.999	11.2–1400	2.814	8.441	4.21	3.96	3.69	3.58	102.08	4.88	0.99
4,5‐O‐Dicaffeoylquinic acid methyl ester	y = 9.67e3 x – 4.3e4	0.9996	5.6–700	1.319	3.958	3.79	4.10	1.78	1.78	98.20	3.48	0.97
Rutin	y = 2.5e3 x + 3.14e5	0.9998	0.979–12200	0.036	0.108	2.91	3.21	3.31	3.46	95.90	3.99	1.03
Hyperoside	y = 4.32e3 x + 1.77e5	0.9995	0.265–3313	0.058	0.174	1.69	1.55	3.21	3.22	100.32	3.82	1.03
Luteoloside	y = 1.07e3 x + 1.34e5	1	0.664–41500	0.056	0.169	1.03	0.92	4.64	3.78	99.16	2.95	1.01
Luteolin	y = 100x + 2.47e3	0.9995	124–15600	10.150	30.450	2.28	2.22	1.34	4.09	100.29	4.35	1.00
Rhoifolin	y = 1.08e4 x + 572	0.9999	1.3–162	0.280	0.840	2.53	2.75	5.20	4.64	99.82	3.14	0.97
Diosmetin	y = 1.15e4 x – 9.93e3	0.9993	1.72–214	0.146	0.438	1.09	0.78	4.63	2.31	100.93	2.95	0.99
Apigenin	y = 2.17e4 x – 9.13e4	0.9995	25.4–318	4.188	12.563	3.54	3.83	3.39	3.47	103.09	3.98	0.98
Kaempferol	y = 191x + 250	0.9997	2.02–101	0.565	1.695	4.28	4.68	3.59	3.32	101.27	3.01	0.95
Astragalin	y = 2.17e4 x + 8.73e4	0.9999	0.0353–883	0.005	0.015	3.51	3.92	3.09	3.51	100.17	3.57	0.97
Lonicerin	y = 1.11e3 x + 4.25e3	0.9994	22.4–14000	0.024	0.073	3.45	3.85	2.95	2.74	100.74	2.80	1.01
Kaempferol‐3‐O‐rutinoside	y = 3.13e3 x + 9.11e4	0.9999	0.632–7900	0.039	0.117	2.31	2.44	3.79	3.36	99.82	3.18	0.98
Isoquercitrin	y = 4.33e3 x + 1.4e5	0.9996	2.65–3310	0.195	0.584	1.68	1.54	4.97	2.91	99.68	2.12	1.02
Sweroside	y = 34.9x + 1.56e3	0.9996	0.841–52600	0.231	0.692	3.54	3.68	3.97	4.99	101.78	2.92	1.00
Secologanic acid	y = 609x + 3.48e4	0.9999	1.45–182000	0.072	0.217	3.02	3.37	4.87	2.19	99.32	4.05	1.03
Loganin	y = 3.95x + 1.72e3	0.999	14.1–35200	4.179	12.536	2.26	2.43	1.83	1.18	103.36	3.89	0.99
Secoxyloganin	y = 638x + 1.32e5	0.999	0.189–47400	0.052	0.157	4.19	4.67	2.47	4.68	98.42	3.14	1.02
Loganin acid	y = 1.11e3 x – 1.98e5	0.9993	441.2–22100	139.266	417.797	2.95	2.54	4.60	4.55	100.32	2.60	1.03
Morroniside	y = 10.4x + 8.21e3	0.999	41.1–514000	7.867	23.600	2.11	1.83	4.32	3.03	101.75	4.18	0.98
Macranthoidin A	y = 193x + 9.76e3	0.9997	18.1–22600	2.989	8.967	2.72	1.94	2.43	2.70	103.97	3.66	1.01
Dipsacoside B	y = 8.94x + 336	0.9999	3.34–20900	0.745	2.236	2.25	2.37	3.18	3.84	99.81	0.92	1.01
Akebia saponin D	y = 6.02x + 8.06e3	0.9996	16.6–207000	1.539	4.618	3.10	1.65	1.46	1.05	102.90	4.10	0.94
l‐Alanine	y = 1.43e3 x + 1.78e4	0.9997	8.47–10600	2.228	6.684	3.07	2.60	4.50	3.43	101.09	4.66	1.03
l‐Serine	y = 351x + 5.93e4	0.9997	9.16–11500	1.342	4.027	2.71	2.54	3.31	3.40	99.84	1.96	1.02
l‐Proline	y = 1.77e3 x + 3.04e4	0.9996	0.381–23800	0.093	0.279	2.75	2.86	2.84	1.13	102.36	2.19	1.01
l‐Valine	y = 6.36e3 x + 2.17e5	0.9995	2.51–1570	0.215	0.644	2.96	3.19	1.49	3.33	98.52	3.85	1.04
l‐Threonine	y = 1.75e3 x – 9.1e4	0.9996	111–13900	22.325	66.975	2.83	1.46	3.20	2.83	96.47	4.28	0.97
l‐Isoleucine	y = 5.15e3 x + 1.92e5	0.9998	2.88–3600	0.422	1.265	0.67	0.63	1.27	2.54	99.49	2.45	1.00
l‐Leucine	y = 9.64e3 x + 4.72e5	0.9997	3.38–4230	0.711	2.134	1.21	1.28	4.23	3.05	103.97	2.36	1.01
l‐Aspartic acid	y = 660x + 2.38e4	0.9991	4.6–57500	0.835	2.506	1.49	1.66	4.18	4.27	99.27	2.96	0.96
l‐Glutamate	y = 1.64e3 x – 2.82e5	0.9998	474–29600	136.177	408.530	0.69	0.74	1.71	1.53	96.93	3.96	1.04
l‐Lysine	y = 2.32e3 x – 2.01e5	0.9993	147–12600	36.781	110.342	0.58	0.65	2.15	2.20	95.67	3.73	0.99
l‐Histidine	y = 7.11e3 x + 1.65e5	0.9999	59.1–7390	7.656	22.968	1.67	1.45	3.55	3.62	104.62	4.79	1.05
l‐Phenylalanine	y = 1.7e4 x + 2.55e5	0.9997	25.6–3200	0.550	1.650	1.78	1.28	3.16	2.36	97.56	3.06	1.03
l‐Arginine	y = 6.54e3 x + 1.21e5	0.9996	0.241–1510	0.043	0.128	1.25	1.19	4.54	4.50	100.17	1.98	1.00
Cytidine	y = 4.72e4 x – 1.94e4	0.9995	1.84–230	0.496	1.488	3.39	3.20	4.69	3.99	94.20	3.90	1.01
Uridine	y = 861x + 7.0e3	0.9997	40.4–5050	7.124	21.372	3.20	3.53	4.19	3.26	100.95	2.55	1.01
Adenosine	y = 5.28e4 x + 3.05e5	0.9994	6.31–789	0.945	2.834	1.45	1.41	1.89	3.29	97.76	3.70	0.98
Inosine	y = 7.98e3 x + 2.12e4	0.9991	8.78–220	1.847	5.542	1.73	1.87	3.95	3.93	101.52	4.10	0.95

Simultaneous quantitation of constituents of sample

Contents of 50 constituents were determined using UFLC‐QTRAP‐MS/MS method in 35 batches of samples. The contents of quantitative determination are shown in Supporting Information Table S1. Our data from different samples elucidated that the chemicals were significantly different between LJF and LF. The contents of protocatechuic acid, rutin, loganin, and morroniside were too low to be detected in LF. The flavonoids of luteoloside, lonicerin, kaempferol‐3‐O‐rutinoside and the iridoids of sweroside, secologanic acid were also in low contents compared with LJF. However, the contents of macranthoidin A, akebia saponin D and dipsacoside B were much higher than LJF. Among them, dipsacoside B reached up to 70687.6 μg/g, therefore, it was considered as a characteristic component of LF in current Chinese Pharmacopoeia. Content percentages of six types of component (organic acids, flavonoids, iridoids, saponins, amino acids, and nucleosides) in 35 batches of samples are shown in Figure 2. The results indicated that the contents have a significant variation in different samples. It is clear that the bioactive components of organic acids account for the largest proportion. Noteworthy, the literature , also suggest that phenolic acids are regarded as one of the main anti‐inflammatory and anti‐bacterial active ingredients. This conclusion might explain the efficacy of “qing‐re‐jie‐du” of LJF and LF. Moreover, regardless of the organic acid, flavonoids or iridoids, the average contents in the flower buds were higher than that of the flowers.

FIGURE 2

Content percentages of six types of component in 35 batches of samples [Colour figure can be viewed at wileyonlinelibrary.com]

Multivariate statistical analysis of samples

PCA and PLS‐DA of samples

To classify and differentiate LJF and LF in chemical composition, PCA and PLS‐DA were performed. The PCA scores plot are displayed in Figure 3(a). The two principal components were set to PC1 and PC2 as variables. Obviously, the samples were divided into two groups. LJF was distributed over the PC1 negative axis; LF was scattered in the PC1 positive axis. Furthermore, the PC1 and PC2 described 69.3% and 13.8% of variability between samples, respectively. This indicated that LJF and LF were significantly different in chemical composition. PLS‐DA was used to extend to find out the constituents contributing to the differences between LJF and LF. In the PLS‐DA scores plot (Figure 3b), LJF and LF were separated into two clusters; the flower buds and flowers of LJF also were divided. The flowers were distributed over the PC2 positive axis, the flower buds were mostly scattered in the PC2 negative axis. In the plot PLS‐DA loading (Figure 3c) and the variable influence on projection (VIP) (Figure 3d), the large load value (VIP > 1) can be regarded as a marker component that contributes greatly to the classification of these samples. l‐Proline, l‐serine, l‐glutamate, chlorogenic acid, 1,3‐O‐dicaffeoylquinic acid, protocatechuic acid, secoxyloganin, secologanic acid, morroniside, macranthoidin A and dipsacoside B could be considered as chemical markers for the sample classification.

FIGURE 3

Multivariate statistical analysis of 35 batches of samples. (a) Principal component analysis (PCA) scores plot: triangles represent Lonicerae Flos (LF) samples, squares represent Lonicerae Japonicae Flos (LJF) samples. (b) Partial least squares discriminant analysis (PLS‐DA) scores plot: triangles represent LF samples, yellow squares represent flowers samples, blue squares represent flower buds samples (LJF). (c) PLS‐DA loading plot: each green squares represent one compound. (d) VIP [Colour figure can be viewed at wileyonlinelibrary.com]

GRA of samples

GRA was carried out to evaluate the quality of samples based on the contents of 50 constituents. Grey comprehensive evaluation values (r ) and quality ranking are listed in Table 4. It could be seen that the quality of LJF is better than LF; the quality ranking of flower buds mostly ahead of the flowers, except for samples 11, 15 and 17. It may be related to different habits and origin. In general, the quality of the flower buds is superior to that of the flowers. A previous study has also shown that the flower buds have the highest medical value. Our results add further support for this conclusion.

TABLE 4

Quality sequencing of samples

Sample	r i	Quality ranking	Sample	r i	Quality ranking
1	0.3567	33	19	0.4321	2
2	0.3778	25	20	0.4130	8
3	0.3693	31	21	0.4088	11
4	0.3977	18	22	0.3767	26
5	0.3752	27	23	0.4094	10
6	0.3745	28	24	0.3676	32
7	0.3897	21	25	0.3705	29
8	0.4070	13	26	0.4202	4
9	0.3865	22	27	0.4134	7
10	0.4018	16	28	0.3947	19
11	0.3809	23	29	0.3695	30
12	0.4479	1	30	0.3789	24
13	0.4168	5	31	0.4050	14
14	0.4124	9	32	0.4032	15
15	0.3935	20	33	0.3548	35
16	0.4142	6	34	0.3550	34
17	0.4083	12	35	0.3978	17
18	0.4247	3

In summary, a sensitive and comprehensive method employing UFLC‐QTRAP‐MS/MS coupled with multivariate statistical analyses was established to evaluate the quality of LJF and LF. A total number of 50 constituents, including 12 organic acids, 12 flavonoids, 6 iridoids, 3 saponins, 13 amino acids, and 4 nucleosides were simultaneously assayed in LJF and LF. Using either PCA or PLS‐DA, it was clear to differentiate LJF and LF, and the sample classification is closely related to 11 different chemical constituents, such as, l‐proline, l‐serine, l‐glutamate, chlorogenic acid, 1,3‐O‐dicaffeoylquinic acid, protocatechuic acid, secoxyloganin, secologanic acid, morroniside, macranthoidin A and dipsacoside B. The GRA assay and the data of content determination also demonstrated the quality of LJF was better than that of LF, and that of the flower buds was superior to the flowers. Therefore, we suggest that LJF should be harvested at the bud stage in order to improve the clinical medicinal value. In addition, organic acids were the main components in LJF and LF. Overall, our study not only established a method of simultaneous determination of multiple types of bioactive constituents of LJF and LF, but provided comprehensive information on the quality control of them. The developed method is conducive to distinguish orthologues or paralogues of them, and provide the support for “heterologous effects”.

Figure S1 Chemical structures of 50 compounds in Lonicerae Japonicae Flos (LJF) and Lonicerae Flos (LF)

Click here for additional data file.

Table S1 Contents (ug/g) of 50 constituents in samples of Lonicerae Japonicae Flos (LJF) and Lonicerae Flos (LF)

Click here for additional data file.