Identification and Comparison of Constituents of Aurantii Fructus and Aurantii Fructus Immaturus by UFLC-DAD-Triple TOF-MS/MS

Although Aurantii Fructus (AF) and Aurantii Fructus Immaturus (AFI) are both the fruits of the same rutaceae plant at different stages of growth, they exert similar yet distinct clinical effects. The chemical composition is crucial for quality control as well as therapeutic application. To address this concern, it is significant to evaluate the similarities and differences of the constituents in both AF and AFI. The extract of AF and AFI were comprehensively analyzed by ultra fast liquid chromatography-photodiode array detector-triple-time of flight-tandem mass spectrometry (UFLC-DAD-Triple TOF-MS/MS). Among the 40 compounds detected, 19 metabolites were detected in both the AF and AFI; whereas 13 compounds were only detected in AF and five constituents were exclusively detected in AFI. In particular, even in AFI, three compounds were only identified in AFI (Citrus aurantium’ L. and its cultivar). Among the 18 compounds confirmed by standard database, 13 compounds were reported in AF and AFI for the first time. Furthermore, the distinction was also revealed by the content of naringin, hesperidin, neohesperidin, and synephrine. The study directly contributed to the similarities and differences of AF and AFI. Herein, similarities and the differences in chemical profiles of AF and AFI could explain the current clinical applications.


Introduction
Aurantii Fructus (AF) and Aurantii Fructus Immaturus (AFI) are commonly applied medicinal herbs in Traditional Chinese Medicine (TCM) practice for thousands of years.In fact, AF and AFI are both the fruits of the same rutaceae plant.AF, harvested in July, is the dried mature fruit of Citrus aurantium' L. and its cultivar, while AFI is the dried immature fruit of Citrus aurantium' L. and its cultivar, or Citrus sinensis (L.) Osbeck collected from May to June.AF and AFI are collected at different stages of fruit growth with diverse clinical efficacy, thus are recorded in the Chinese Pharmacopoeia as two distinct medicinal materials [1].According to TCM theory, AF and AFI each have their own unique clinical applications.Although AF and AFI have common effects of regulating visceral functions, AF is always used to alleviate chest pain and improve gastrointestinal functions such as alleviating dyspepsia in a gentle yet efficient manner.AFI, compared to AF, expresses a relatively more rapid and robust way of action and is often employed to disperse severe abdominal distention and to eliminate phlegm etc. Herein, it is of great interest to investigate the underlying compounds present in AF and AFI collected from the same plant source that exhibit diverse pharmacological and clinical effects.
It is well-known that the efficacy of herbal medicines is significantly correlated with chemical composition.Thus, to compare the similarities and differences of AF/AFI comprehensively, it is necessary to evaluate their chemical composition in larger scale and accuracy.Some publications on the similarities and differences between AF and AFI have mainly focused on the quantification of several flavonoids or alkaloids by high-performance liquid chromatography-photodiode array detector (HPLC-DAD) or high-performance liquid chromatography-mass spectrometry (HPLC-MS), and the comparison of characteristic compounds in a discriminating model [2][3][4][5][6][7][8][9].Even if HPLC-DAD and HPLC-MS have been utilized to quantify constituents, they can hardly be used to identify unknown compounds.Moreover, nuclear magnetic resonance (NMR) has been used to identify unknown components in AFI after separation [10,11].However, to the best of our knowledge, NMR has not been applied to determine multiple constituents of AF and AFI simultaneously.Recently, high-resolution mass spectrometry, including ion trap mass spectrometry (ITIMS) and time of flight-mass spectrometry (TOF-MS) have been applied to comprehensively compare the chemical composition in AF or AFI [12][13][14].Nevertheless, high-resolution mass spectrometry alone cannot determine new chemical structures and reference standards must be used to confirm constituents.
In this study, the chemical composition of AF and AFI was systematically evaluated with reference standards and standard database by UFLC-DAD-Triple TOF-MS/MS.The chemical similarities and differences between AF, AFI (Citrus aurantium' L. and its cultivar) and AFI (Citrus sinensis (L.) Osbeck) were summarized.Furthermore, the distinction was also revealed by hierarchical cluster analysis (HCA) based on the characteristics of the content of naringin, hesperidin, neohesperidin and synephrine.The differences in the above chemical compositions probably explain different clinical effects.

Qualitative Comparison of Constituents in Aurantii Fructus (AF) and Aurantii Fructus Immaturus (AFI)
In this study, a standard database (AB SCIEX LibraryView, Version 1.0) was used to identify the compounds including accurate mass-to-charge ratios, the ratio between isotopic peaks, and product ion spectra [15].Through the database, compounds could be discriminated from their structural isomers.As summarized in Table S1, 40 compounds including 27 flavonoids, seven coumarins, four triterpenoids, an organic acid, and an alkaloid were detected (shown in Figure 1).Except for 11 standards, 18 compounds were identified in AF and AFI by standard database.Among the 18 compounds, 13 compounds were identified and reported in AF and AFI for the first time.Among the 13 compounds, limonin was detected in both AF and AFI; obacunone was detected in both AF and AFI (Citrus aurantium' L. and its cultivar); nicotiflorin, narcissoside, and pedunculoside were only detected in AFI; and apigenin-6,8-di-C-glucoside (vicenin-2), eupatilin, vitexicarpin, marmesin, xanthotoxol, an isomer of xanthotoxol, osthole, and nomilin were solely detected in AF (shown in Figure 2).+ where the losses of H 2 O, methyl groups (CH 3 ), methylamine (CH 3 NH 2 ), and carbon oxide (CO) suggested the existence of hydroxyl, phenolic hydroxyl, and methylamine groups.Fragment ions at m/z 77, 65 were the characteristic fragmentation of benzene.Based on the similarity in retention time and fragmentation patterns with reference standard, compound 2 was confirmed as synephrine.

Characterization of Flavonoids
The flavonoids in AF and AFI included flavones, flavonols, dihydroflavone, and their corresponding glycosides.The flavonoid glycosides tended to generate [M − H] − ions more than [M + H] + ions.After losing the glycosyl moiety, their major characteristic fragment ions could be observed as A 1 , B 1 , A 2 , and B 2 as shown in Figure 3.In addition, the intensity of A 1 , B 1 , A 2 , and B 2 was always higher than the others.Relative fragmentation pathways of flavonoids are presented in Figure 3. Furthermore, flavonoids tended to lose 28 Da (CO), 18 Da (H 2 O) and 15 Da (CH 3 ), suggesting the existence of phenolic hydroxyl and methyl groups.

Data Analysis
To evaluate the differences between AF and AFI, herarchical cluster analysis (HCA) was performed based on the characteristics of the contents of naringin, hesperidin, neohesperidin, and synephrine.The HCA results demonstrated significant differences in the form of a dendrogram calculated based on Ward's minimum variance method (shown in Figure 5).All samples could be clearly divided into two clusters: 11 batches of AFI derived from Citrus sinensis Osbeck were classified into Cluster I (left); and AFI and AF derived from Citrus aurantium L. and its cultivar were classified into Cluster II (right).Moreover, Cluster II could be divided into two further clusters according the harvest time, while all AFI were included in Cluster III and all AF were included in Cluster IV, which was consistent with the classification in the Chinese Pharmacopoeia.

•
Stability: The analytes were found to be stable for 48 h for both samples and standards (all RSD < 2%).
Moreover, the distinction between them was also revealed by the content of naringin, neohesperidin, hesperidin, and synephrine.The HCA results clearly showed differences in chemical profiles of different plant species as well as differences in harvesting times of the same species.
The differences in chemical profiles of AF and AFI could explain the current clinical applications.Herein our study shed light on the potential pharmacological effects of these identified compounds of AF and AFI that deserves further investigation in the future.

Reagents and Materials
Twelve batches of AFI (Citrus aurantium' L. and its cultivar) were collected from four different districts in China (six batches from Jiangxi province; one batch from Hunan province; two batches from Sichuan province; three batches from Zhejiang province) in 2016.Eleven batches of AFI (Citrus sinensis (L.) Osbeck) were harvested from three different districts in China (four batches from Jiangxi province; four batches from Hunan province; three batches from Sichuan province) in 2016.Twenty-four batches of AF (Citrus aurantium' L. and its cultivar) were collected from four different districts in China (ten batches from Jiangxi province; four batches from Hunan province; five batches from Sichuan province; five batches from Zhejiang province) in 2016.All samples were deposited in the Guangdong Key Laboratory of Plant Resources and identified by chief pharmacist Liwei Yang (Guangdong Institute for Food and Drug Control, Guangzhou, China).
Methanol of HPLC grade (Honeywell, Morris Plains, NJ, USA) and formic acid of HPLC grade (Sigma-Aldrich) were used.All water used was distilled and further purified by a Milli-Q system (Millipore, Milford, MA, USA).Other reagents used in the experiment were analytical grade.

Standard Solutions and Sample Preparation
The mixed solution of 11 standards for identification was prepared in 50% methanol at the concentration of 10 µg /mL for each compound.After 0.2 g of powdered sample was accurately weighed to a 100 mL glass-stoppered conical flask, 50 mL of 50% methanol (methanol:water = 1:1, v:v) was added.The filled flask was weighed with a precision of ±0.01 g, sonicated for 30 min (250 W, 40 kHz), allowed to cool, and adjusted to the initial weight by 50% methanol as needed.After taking the filtrate 10 mL precisely to 25 mL volumetric flask, 50% methanol was add to the mark.Then, the solution was filtered with a membrane filter (0.45 µm) to collect the successive filtrate.The filtrate was used as the sample solution.

Analysis by UFLC-DAD-Triple TOF-MS/MS
A sample solution was prepared using four methods including sonication in methanol (0.5 h), reflux by methanol (1.5 h), sonication in 50% methanol (methanol:water = 1:1, v:v) (0.5 h) and reflux by 50% methanol (1.5 h).Separately the sample solution was injected to determine the amount of naringin (%), hesperidin (%), and neohesperidin (%).The method of sonication in 50% methanol for 30 min was chosen in the experiment since its efficiency was the greatest.After the optimization of LC and MS conditions, a simple UFLC-DAD-Triple TOF-MS/MS method was developed.
Detections were performed by a hybrid triple quadrupole time-of-flight mass spectrometer (AB SCIEX Triple TOF™ 5600 plus, AB Sciex, Foster City, CA, USA) equipped with electrospray ionization (ESI) source, analyst software (PeakView, Version 2.1, AB Sciex, Shanghai, China) and a standard database (LibraryView, Version 1.0, AB Sciex).The standard database included information on 1213 compounds (molecular formula, name, accurate mass-to-charge ratios, the ratio between isotopic peaks and product ion spectrum).The standards in the standard database were purchased from the National Institute for Control of Biological and Pharmaceutical Products of China.After opening the data files in the software, all the compound information could be imported from the standard database.Then, the software automatically calculated similarity score.With an error less than 5 ppm and a score greater than 50% as screening criterion, 40 compounds were chosen to analyze.
The TOF-MS worked in a full scan mode and mass range was set at m/z 50-1200 in both positive and negative ion modes.The acquisition of MS/MS data was accomplished by the IDA (information-dependent acquisition) mode.The conditions of the mass spectrometer were as follows: ion source gas1 55 psi; ion source gas2 55 psi; curtain gas 35 psi; temperature 550 • C; ion spray voltage floating 5500 V (positive) or −4500 V (negative); collision energy 40 V and declustering potential 60 V. Nitrogen was used as the nebulizer and auxiliary gas.
The contents of naringin, hesperidin, neohesperidin and synephrine were defined as four characteristics in the analysis to differentiate and classify the AF and AFI samples.Hierarchical cluster analysis (HCA) of samples was performed by SIMCA-P software (version 13.0, Umetrics, Malmö, Sweden) to evaluate the differences between AF and AFI.After the contents of the four compounds in 24 batches of AF and 23 batches of AFI formed a dataset (shown in Table S3), the dataset was imported to the software.Then, the distances between 47 samples were calculated based on Ward's minimum variance method, and the results was presented as a dendrogram.

Methodology Validation
Methodology validation was established according to the Guidelines for the investigation and formulation of Chinese material medica monographs in Chinese Pharmacopeia 2015.
Specificity: Separately, an equal volume of the blank solvent, standard solution (naringin, neohesperidin, hesperidin, and synephrine) and sample solution (AF and AFI) were injected into the chromatograph, and the chromatograms were recorded.
Precision: Repeatability was assessed by injecting six sample solutions of the same batch into the instrument to calculate their contents.Intermediate precision was evaluated by preparing sample solutions independently in duplicate with the same sample by two operators, on two separate days, and injected into different instruments to calculate their contents, separately.
Accuracy: Sample solutions were prepared in three different amount levels (low, medium and high) for triplicate experiments at each level.The differences of the found amount and original amount divided by the spiked amount of naringin, hesperidin, neohesperidin, and synephrine were used to calculate the recovery.
Stability: The stability of the sample solution and standard solution was investigated.It was carried out by comparing the peak areas of naringin, hesperidin, neohesperidin, and synephrine in the chromatograph of the same sample solution and standard solution, after being stored at room temperature for different times (0, 2, 4, 6, 8, 12, 24, 48 h).Stability was evaluated by calculating the relative standard deviation (RSD) of area obtained.
Ruggedness: The ruggedness of the established method was evaluated by examining its stability with small variations of procedural parameters; the effects of different columns (Welch Ultimate XB-C 18 , 4.6 mm × 25 cm, 5 µm; Agilent Zorbax Eclipse Plus C 18 , 4.6 mm × 25 cm, 5 µm; Elite Hypersil ODS 2 , 4.6 mm × 25 cm, 5 µm) on the relative retention time of each characteristic peak and their content were investigated.

Supplementary Materials:
The following are available online.Table S1: Qualitative comparison of constituents in AF and AFI; Table S2: Qualitative information of constituents in AF and AFI; Table S3: Content of naringin, hesperidin, neohesperidin and synephrine in AF and AFI on the dried basis (n = 2, RSD < 2%).

Figure 2 .
Figure 2. The chemical similarities and differences between AF and AFI.

Figure 5 .
Figure 5. Evaluation of the differences between AF and AFI with hierarchical cluster analysis (HCA).

2. 8 .
Methodology Validation • Specificity: The integration peak in the chromatogram of the sample solution corresponded in time to the peak in the chromatogram of the standard solution.No such peak of that retention time appeared in the chromatogram of the solvent.• Linearity and range: The regression analysis was performed with a peak area integral value of Y, and reference substance injection volume X.The regression equations and linearity ranges were as follows: naringin: Y = 29.655X+ 0.0322, R 2 = 1.0000, 0.0789~2.3661µg; hesperidin: Y = 35.29X+ 0.0086, R 2 = 1.0000, 0.0077~0.2321µg; neohesperidin: Y = 31.137X+ 0.0157, R 2 = 1.0000, 0.0826~2.4765µg; synephrine: Y = 54.454X− 0.081, R 2 = 1.0000, 0.0156~0.623µg.All results above showed good linear relationship.• Precision: The repeatability (all RSD < 2%) and the intermediate precision of contents (all RAD < 2%) indicated high precision.