Determination of Ten Flavonoids in the Raw and Fermented Fructus Aurantii by Quantitative Analysis of Multicomponents via a Single Marker (QAMS) Based on UPLC

Fermented Fructus Aurantii (FFA) is widely used in South China for the treatment of functional dyspepsia. Naringin, neohesperidin, and other flavonoids are the main pharmacodynamic components of FFA. A new method is presented for the simultaneous determination of 10 flavonoids (including flavonoid glycosides and aglycones) in FFA using the quantitative analysis of multicomponents via a single marker (QAMS) approach and is used to investigate changes in flavonoids during fermentation. The viability and precision of QAMS were validated against the ultrahigh-performance liquid chromatography (UPLC), with various UPLC instruments and chromatographic conditions being evaluated. Differences between raw Fructus Aurantii (RFA) and FFA were examined using orthogonal partial least squares discrimination analysis (OPLS-DA) and content determination. The influence of various fermentation conditions on flavonoids was also investigated. There were no appreciable differences between the QAMS and the external standard method (ESM), demonstrating that QAMS is an improved method for the determination of FA and FFA. FFA and RFA can be readily distinguished based on OPLS-DA chemometric modelling and the corresponding chromatograms. In addition, the flavonoid changes after fermentation. Fermentation considerably reduced the contents of flavonoid glycosides, while increasing hesperidin-7-O-glucoside and flavonoid aglycones. Moreover, fermentation conditions impact multiple flavonoids in FA, so controlling these conditions is necessary for the quality control of fermented FA products. This QAMS approach is useful for detecting numerous components in RFA and FFA simply, quickly, and efficiently, thus strengthening the quality control of FA and its fermented products.


Introduction
Fructus Aurantiic (FA), also known as Zhiqiao in China, is derived from the fruit of the Citrus aurantium L. plant and its cultivars [1]. It is a famous and popularly used herbal medicine that is widely used around the world, particularly in China, Japan, India, and Vietnam. FA is often used for the clinical treatment of stomach distension, gastrointestinal food retention, and uterine prolapse [2,3]. Many studies have shown that FA contains various active ingredients, with favonoids being the most active [4,5]. Naringin and neohesperidin are the signifcant favonoids in FA and were selected for quantitative analysis in the quality control of FA for the 2020 edition of Te Chinese Pharmacopoeia [1]. In South China (e.g., Guangdong, Hong Kong, and Macao), fermented Fructus Aurantii (FFA) is more widely used because of its improved efcacy in the treatment of functional dyspepsia [6]. Whilst our previous research found that favonoids are also the main active components in FFA, the contents of naringin and neohesperidin were lower, indicating that other favonoids may have pharmacodynamic roles. A comprehensive analysis of the favonoid content will be valuable in the evaluation of FFA quality.
Te curative efects of traditional Chinese medicines (TCM) are related to their complex chemical components [7]. Te chemical composition of FA is more complex following fermentation, so quantitative examination of only two favonoids is insufcient to comprehensively evaluate its quality. While multicomponent quality control methods are desirable for Chinese herbal medicines with multiple targets, they can be problematic. However, the quantitative analysis of multicomponents via a single marker (QAMS) method can be useful [8]. QAMS is an analytical method that can simultaneously monitor multiple analytes via the determination of a cheap and readily available reference compound. Tis widely used approach signifcantly alleviates the defciency and high cost of reference compounds [9].
Tis study investigated the viability and applicability of the QAMS method. Ten favonoids that exhibit apparent changes after fermentation were selected and accurately quantifed using ultrahigh-performance liquid chromatography (UPLC). Changes in these compounds due to fermentation were compared using the external standard method (ESM) and the newly developed QAMS method. Te QAMS approach can shorten the detection time to 30 min, improve efciency, and reduce analytical costs [10][11][12][13][14][15]. Moreover, fermentation conditions afect microbial metabolism during the fermentation of FA and, consequently, alter the contents of various chemical components. Tis study used the QAMS method to simultaneously determine the contents of the ten favonoids in FA and FFA ( Figure 1) under diferent fermentation conditions, namely, fermentation time, temperature, humidity, and soaking time and provide a foundation for the optimization of FFA production. Tis study also includes a preliminary exploration of the changes in favonoids during the fermentation of FA and a scientifc basis for the quality control and evaluation of FA and its processed products (FFA).

Materials and Chemicals.
Te Fructus Aurantii samples were procured from the Guangzhou Zhixin Group and authenticated by Prof. Danyan Zhang from the Department of Resources and Identifcation of Chinese Herbal Medicine at Guangzhou University of Traditional Chinese Medicine. After determination, the contents of naringin and neohesperidin in FA were no less than 4% and 3%, respectively, meeting the requirements of the Chinese Pharmacopoeia [1]. Phosphoric acid of chromatographic quality was acquired from the Guangzhou Chemical Reagent Co., Ltd. (Guangzhou, China). Te experiment employed ultrapure water. Shanghai Yuanye Co., Ltd. provided the following compounds: naringin (no: YJ77D9F001), hesperidin (no: P06D9F77001), neohesperidin (no: C05F4Y2), naringenin (no: YJ0603HA13), and hesperetin (no: C03F6Y1). Te following compounds were bought from Chengdu Pu Si Biotechnology Co., Ltd.: narirutin (no: PS011543), eriocitrin (no: PS010198), neoeriocitrin (no: PS010420), hesperidin-7-O-glucoside (no: PS020721), and poncirin (no: PS010580). Te purity of 10 standards exceeded 98% (the structure of the ten favonoids are shown in Figure 1) (the picture of FFA and FA are shown in Supporting Information Figure S1).

Instruments and Conditions.
Two UPLC systems were used in this research: a Shimadzu LC-20A series UPLC system (Shimadzu, Japan) and a Waters Acquity UPLC system (Waters, USA). Both systems had an autosampler, an online degasser, a photodiode array detector, binary pumps, and a column heater. For sample separation, the following three analytical columns were utilized: Waters UPLC BEH C 18 (2.1 mm × 100 mm, 1.6 μm), Phenomenex LC C 18 (2.1 mm × 100 mm, 1.6 μm), and Shimadzu C 18 (2.1 mm × 100 mm, 1.6 μm). Te mobile phase included the following two components: mobile phase A, which was a 0.1% aqueous solution of phosphoric acid, and aqueous solution B, which contained acetonitrile. Te elution conditions followed a solvent gradient, with 5% (B) used for the frst 2 min, followed by a gradual increase to 26% (B) over the next 8 min. Te concentration of B was increased to 27.7% between 10 and 20 min and then to 55% between 20 and 26 min. Between 26 and 28 minutes, the concentration of B was reduced to 20% and fnally to 5% between 28 and 29 min, before being held constant at 5% for the fnal min. Te mobile phase fowed at a rate of 0.3 mL/min throughout the entire detection procedure. A column heater was used to maintain the analysis column at 35°C. Te detection wavelength was confgured to 283 nm for the UPLC analysis, while the detection volume of the sample was injected into 2 μL.

Processing of FFA Samples.
Te FFA samples were prepared using the fermentation processing methods outlined in the 1984 edition of the Traditional Chinese Medicine Processing Code of Guangdong Province [6] (the processing steps are shown in Supporting Information Figure S2). Te basic operations are as follows: Te FA was weighed, with approximately 40 g per batch, and soaked in water for a specifed duration of time (2 h, 4 h, 6 h, 8 h, or 10 h). Ten, the water was poured out, the soaked FA was placed into a breathable and leaky transparent bag, and the fermentation was performed in a culture incubator with a certain temperature (22°C, 27°C, 32°C, 37°C, or 42°C) and humidity (50%, 60%, 60%, 70%, 80%, 80%, or 90%). After fermentation for a certain period of time (2 d, 3 d, 4 d, 5 d, or 6 d), the FFA samples in the incubator were removed, cleaned, sliced, and dried, and then, the corresponding FFA samples were collected (the sample information is shown in Table 1)

Preparation of Standard and Sample Solutions.
To prepare a series of mixed reference results with varying concentrations for UPLC analysis, ten diferent concentrations of the reference solution were created. Subsequently, different volumes of the reference solution were drawn and mixed before being diluted with methanol. Tis process resulted in a range of mixed reference solutions with varying concentrations, which were also analyzed using UPLC.
Te RFA or FFA samples were dried and then powdered to a particle size of 850 μm. A 0.500 g portion of the powder was extracted by heating under refux with 45 mL of methanol for 1.5 h. After refuxing and heating, the weight was replenished by adding methanol and shaking the liquid evenly. Te resulting supernatant fuid was fltered through a 0.22 μm membrane for UPLC examination.

Te Principle of Quantitative Analysis of Multicomponents via a Single Marker Method.
Te guiding principle of the QAMS method states that the component's content is positively proportional to the detector's response value within a specifc linear range [8]. When many components are measured at once, as shown in (1), one of the typical components can be chosen as the internal standard to calculate the relative correction factors (RCF) (f k/m ) between the internal and external components. Based on the f k/m of the components to be evaluated and the internal standard component, an equation can be used to calculate the contents of other components in sample (2) [12].
Te calculation of RCF is as follows: Te calculations of the favonoids' contents are performed as follows: In this study, W k and W m were used to represent the concentrations of naringin and other favonoids presented in both the FFA samples and reference solutions, while A k and A m represent the corresponding peak areas of naringin and 2.6. Te Quantifcation of FFA Samples. Each FFA sample was measured three times using both the quantitative analysis of mixture standards (QAMS) and the external standard method (ESM) to determine the contents of 10 favonoids. To compare the diferences in results between the two methods, the standard method diference (SMD) was derived using the following equation, as previously explained [16]: Here, W ESM represents the contents of the components measured using the ESM method, while W QAMS represents the contents of the components measured using the QAMS method.

Data Processing and Multivariate Statistical Analysis.
To classify the samples, the orthogonal partial least squares discrimination analysis (OPLS-DA) method was employed to maximize the covariance between the independent variables X and the response variables Y. Te variable infuence on projection (VIP) scores were used to evaluate the discriminatory capacity of each observable variable, with variables possessing a VIP score >1 deemed as potential marker compounds for distinguishing various groupings. All OPLS-DA analyses were conducted by utilizing SIMCA 14.1 software (Umetrics, Umea, Sweden).

Optimization of the Preparation of Sample Solutions.
To optimize the extraction of the 10 favonoids from the FFA samples, various refux extraction times (60, 90, and 120 min) were evaluated. Te lowest extraction efciency was at 60 min, and there was a minimal diference between 90 and 120 min, but higher extraction efciency was attained at 90 min (Supporting Information Figure S3).

UPLC Method Validation.
Samples for the UPLC method validation were pretreated and analyzed as described above, and linearity, the limit of quantifcation, and the limit of detection were determined. As shown in Table S1, the regression correlation coefcients of all the 10 favonoids were above 0.9990, indicating a satisfactory linearity of the calibration curves within the range of contents considered appropriate for quantitative analysis. Te 10 favonoids' average recoveries ranged from 96.07 to 104.32%, whereas RSDs' average recoveries were between 0.02% and 2.74%. Te RSD values of 10 favonoids ranged from 1.03% to 1.22% in the repeatability test. Te stability of the sample solutions was verifed over 24 hours at 4°C, with results ranging from 1.22% to 2.58% across the time points of 0, 4, 8, 12, 16, 20, and 24 h. Tese fndings indicate that the FFA sample solutions were highly stable for up to 24 hours. Tus, this UPLC method demonstrates acceptable recovery, precision, stability, and repeatability for the reliable determination of all 10 favonoids (Supporting Information  Table S1). RCFs were initially calculated based on the peak area ratio and concentrations of naringin and other favonoids in the mixed reference standard for the simultaneous determination of the ten favonoids using QAMS. Te RCFs of the nine favonoids relative to naringin are shown in Table 2. Te RCFs displayed excellent accuracy, with RSDs between 0.11% and 2.93%.

Evaluation of the Durability and System Applicability of Quantitative Analysis of Multicomponents by a Single
Marker. Te impact of varied fow rates, chromatographic columns, and column temperatures on the RCFs was examined to appraise the stability and durability of the QAMS method. Tree columns of Waters UPLC BEH C 18 , Phenomenex LC C 18 , and Shimadzu C 18 columns were used for analysis on a Shimadzu LC-20A UPLC system and a Waters Acquity UPLC system, respectively. Te RCFs determined using diferent instruments and columns exhibited RSDs below 3%. Te Shimadzu LC-20A UPLC system equipped with a Waters BEH C 18 was utilized to assess the infuences of column temperature (30, 35, and 40°C) and fow rate (0.1, 0.2, and 0.3 mL/min). Te RCFs evaluated using various column temperatures and fow rates had RSDs of less than 3% and 2%, respectively. Tus, instrument, column, column temperature, and fow rate had no signifcant impact on the RCFs, which also showed good reproducibility (Supporting Information Table S2).

Te Location of Target Chromatographic Peaks.
Te precise determination of target peak locations using a single reference is still a signifcant problem for QAMS. In order to solve this problem, the concept of relative retention time is proposed to accurately identify the desired chromatographic peak, as outlined in the following equation [17]: Here, t k and t m are the retention times of naringin and other favonoids under test, respectively.
Two diferent UPLC devices were used to evaluate the relative retention times of three chromatographic columns. Te fndings revealed that their RSD values for the relative retention times of all components were less than 3%, indicating that they could be utilized to locate the peak of all tested components (Supporting Information Table S3).

Consistency Assessment of QAMS and ESM Results.
Te concentrations of the 10 favonoids were determined in 20 FFA samples from various regions using both the ESM and QAMS (Table 3). Te accuracy of QAMS was expressed as the SMD value by comparing the analytical results. SMD ranged from 0% to 2.8% (Supporting Information Table S4), which demonstrates that it is feasible to simultaneously quantify these 10 favonoids in FFA samples using QAMS. Figures 2 and 3 compare the 10 favonoids in RFA and FFA samples produced under various fermentation conditions. Tere were seven identifable peaks in RFA and 10 in the FFA samples, thus revealing three new peaks ( Figure 2). Te peak heights of favonoid glycosides were signifcantly higher in RFA than in FFA, while favonoid aglycones were distinctly lower in RFA than in FFA. Figure 3 shows that the contents of the seven favonoid glycosides were higher in RFA than in all the FFA samples, and RFA contained almost no, or only a few, favonoid aglycones. Tus, fermentation reduced the favonoid glycosides and increased the favonoid aglycones, and it is speculated that the three favonoid aglycones were produced by the process of fermentation (Supporting Information  Table S5-S8).

Comparison of the Content and Quantity of Flavonoids in RFA and FFA.
Tese data demonstrate the abundant diferences in the chemical composition of RFA and FFA. Multivariate statistical analysis (OPLS-DA) was applied to characterize and visualize these diferences arising from fermentation. (RFA and FFA). OPLS-DA was used to discriminate between FA samples before and after fermentation based on the   Studies have shown that naringin and hesperidin are not easily absorbed from Chinese herbal medicine [18]. However, processing can transform these compounds into single glycosides or aglycones, signifcantly improving their bioavailability and absorption by the human body [19][20][21]. Fermentation of FA can produce secondary glycosides such as naringenin-7-O-glucoside and hesperidin-7-O-glucoside, as well as signifcantly increase the contents of naringenin and hesperetin. It is speculated that some favonoid glycosides are degraded to aglycones or secondary glycosides by intracellular or extracellular enzymes secreted by microorganisms.

Multivariate Statistical Analysis of Flavonoids Composition Changes in FA before and after Fermentation
Tus, consistent with the contents analysis of favonoids, OPLS-DA provides further evidence that fermentation   afects the composition of FA, resulting in diferences between RFA and FFA. Fermentation conditions impact FFA composition, and further analysis of diferent conditions is warranted. Figures 3 and 5 show that fermentation temperature, humidity, time, and soaking time afect the chemical composition of FFA. Flavonoid contents are reduced when the fermentation time reaches four days. Fermentation temperatures over 37°C or below 27°C afect the fermentation process and the production of new compounds. Similarly, when the fermentation humidity is low, the growth of microorganisms is inhibited, thus impacting the fermentation process and the production of new compounds. Tese trends in the composition of favonoids in FFA processed under diferent conditions provide a scientifc basis for the optimization of the fermentation process (Supporting Information Table S5∼S8).

Conclusions
Te study of TCMs requires comprehensive analytical methods. Te present study establishes a QAMS method for the determination of 10 favonoids in FFA. Tis method is shown to be efcient, reliable, and suitable for the evaluation of FFA quality. Tese 10 favonoids were determined in FFA and FA to explore the changes arising from the fermentation process. Fermentation conditions (temperature, humidity, and time) afect the favonoid contents. Fermentation results in a considerable decrease in favonoid glycosides, while hesperidin-7-O-glucoside and favonoid aglycones increase. Te QAMS method developed in this study will make the quality assessment of FFA more feasible and efcient and will provide a basis for process optimization.

Abbreviations
FA: Fructus Aurantii RFA: Raw Fructus Aurantii FFA: Fermented Fructus Aurantii DAD: Diode array detection ESM: External standard method QAMS: Quantitative analysis of multicomponents by a single marker RCF: Relative correction factor SMD: Standard method diference TCM: Traditional Chinese medicine OPLS-DA: Orthogonal partial least squares discrimination analysis.

Data Availability
Te data used to support the fndings of this study are available from the corresponding author upon request.

Conflicts of Interest
Te authors declare that they have no conficts of interest.

Supplementary Materials
Supplementary Materials are about "repeatability of the correction factor," "structure of 10 favonoids," and "comparison of components under diferent fermentation conditions." Figure S1: Pictures of Fructus Aurantii and fermented Fructus Aurantii. Figure S2: Processing steps of Lingnan Special Decoction Pieces "Processed Fructus aurantia." Figure S3: Efect of extraction time on extraction yields of the contents of ten components. Table S1: Te regression equations, LOD, LOQ, precision, recovery, repeatability, and stability for the determination of ten components. Table S2: Efects of diferent instruments, columns, column temperatures, and fow rates on RCFs.