Quality Evaluation of Decoction Pieces of Gardeniae Fructus Based on Qualitative Analysis of the HPLC Fingerprint and Triple-Q-TOF-MS/MS Combined with Quantitative Analysis of 12 Representative Components

In this study, quality evaluation (QE) of 40 batches of decoction pieces of Gardeniae Fructus (GF) produced by different manufacturers of herbal pieces was performed by qualitative analysis of the HPLC fingerprint and ultra-fast liquid chromatography (UFLC)-triple-Q-TOF-MS/MS combined with quantitative analysis of multiple components, which we established previously for QE of traditional medicine. First, HPLC fingerprints of 40 samples were determined, and the common peaks in the reference fingerprint were assigned. Second, the components of the common peaks in the HPLC fingerprints were identified by UFLC-triple-Q-TOF-MS/MS. Finally, the contents of the components confirmed by reference substances were measured. The results showed that there were 28 common peaks in the HPLC fingerprints of 40 samples. The components of these 28 common peaks were identified as 13 iridoids, 4 crocins, 7 monocyclic monoterpenoids, 3 organic acids, and 1 flavonoid. Of these, a total of 12 components, including 7 iridoids of geniposide, shanzhiside, geniposidic acid, deacetyl asperulosidic acid methyl ester, gardenoside, scandoside methyl ester, and genipin gentiobioside, 2 crocins such as crocin I and crocin II, 1 monocyclic monoterpenoid of jasminoside B, 1 organic acid of chlorogenic acid, and 1 flavonoid of rutin, were unambiguously identified by comparison with reference substances. There were certain differences in the contents of these 12 components among 40 samples. The geniposide content ranged from 37.917 to 72.216 mg/g, and the total content of the 7 iridoids ranged from 59.931 to 94.314 mg/g.


Introduction
Gardeniae Fructus (GF), the desiccative ripe fruit of Gardenia jasminoides Ellis (Rubiaceae), is a well-known and frequently used traditional medicine officially recorded in the Chinese and Japanese Pharmacopoeias [1][2][3]. To date, nearly 200 phytochemicals have been isolated and identified from GF [1,[4][5][6][7][8][9], which mainly include iridoids, crocins, monocyclic monoterpenoids, organic acids, and flavonoids [1,10]. e representative components of iridoids include geniposide, genipin gentiobioside, gardenoside, shanzhiside, deacetyl asperulosidic acid methyl ester (DAAEM), and nephrotoxicity [24][25][26]. In other words, GF and iridoids have both toxic and protective effects on the liver and kidney. erefore, strictly controlling the quality of GF is very important to ensure the safety and effectiveness of clinical medications. e original medicinal materials of GF can only be used in the clinic after being processed into decoction pieces [27]. Decoction pieces of GF (Figure 1(a)) are the product of original medicinal materials of GF after removing impurities and crushing (Figure 1(b)) [2]. e quality of GF decoction pieces is directly related to the safety and effectiveness of clinical medication. At present, there are some literature reports on the quality evaluation (QE) of original medicinal materials of GF [28][29][30][31][32], but there is no report on the QE of GF decoction pieces. erefore, in this study, QE of 40 batches of GF decoction pieces produced by different manufacturers of herbal pieces was performed by qualitative analysis of the HPLC fingerprint and ultra-fast liquid chromatography (UFLC)-triple-Q-TOF-MS/MS combined with quantitative analysis of multiple components, which we established previously for QE of traditional medicine [33].

Samples and Sample Preparation.
Forty batches of GF decoction pieces produced by different manufacturers of herbal pieces were purchased from different large TCM hospitals in China; the information on all 40 samples is given in Table 1.
GF decoction pieces were ground into powders before use. Powder samples (0.1 g) were weighed accurately and placed in a 50 mL brown volumetric flask; approximately, 49 mL of 50% (v/v) methanol was added; the mixture was then extracted by ultrasonication (200 W, 53 kHz) for 30 min. After cooling to room temperature, 50% (v/v) methanol was added for calibration of the volumetric flask and shaken well; the mixture was filtered through a 0.22 μm filter membrane, and the filtrate was taken as a sample solution.

Preparation of Reference Substance Solutions.
Twelve reference substance stock solutions with a concentration range of 0.1-4 mg/mL were prepared by accurately weighing appropriate amounts of 12 reference substances and dissolving them in 50% (v/v) methanol.
Appropriate amounts of each reference substance stock solution were precisely measured, mixed together, and diluted with 50% (v/v) methanol. us, the mixed reference substance solution for qualitative analysis in a concentration range of 1-67 μg/mL of each compound was prepared.
Working solution A in a concentration range of 3-381 μg/mL of each compound for quantitative analysis was prepared by the same method as that used in preparing the mixed reference substance solution for qualitative analysis. Working solutions B, C, and D were prepared by diluting working solution A with 50% methanol to 2, 5, and 10 times its initial volume, respectively.

Chromatographic Conditions for HPLC Fingerprint and
Quantitative Analysis. Determination of the HPLC fingerprint and quantitative analysis of 12 components were performed on an HPLC system equipped with an e2695 separation unit, a 2998 PDA detector, and an Empower 3 data processing system (Waters Corp., Milford, MA, USA). Chromatographic separation was performed on a Symmetry C 18 column (4.6 mm × 250 mm, 5 μm, Waters Corp., USA). e column was maintained at 30°C. Acetonitrile (A) and 0.1% (v/v) formic acid (B) were used as mobile phases using the following gradient elution program: 0-5 min, 2% A; 5-10 min, 2-5% A; 10-45 min, 5-15% A; 45-80 min, 15-40% A; 80-82 min, 40-98% A. e injection volume of sample solution was 30 μL at a flow rate of 1.0 mL/min. e wavelength for the determination of fingerprints and contents of the 7 iridoids, rutin, and jasminoside B was set at 254 nm, and those for the determination of contents of chlorogenic acid and the 2 crocins were set at 324 nm and 430 nm, respectively.

Validation of the HPLC Method for Fingerprint Analysis.
By using peak 11 (genipin gentiobioside) as the reference peak and the relative standard deviation (RSD) value of the relative peak area (RPA) and the average relative retention time (RRT) of the 28 common peaks as measurement values, the HPLC method for fingerprint determination was validated with precision, stability, and repeatability tests. e precision was determined by six replicate injections of the same sample (S1) solution. e stability test was performed by injecting the sample solution (S1) at 0, 6, 12, 18, 24, and 36 h after preparation. e repeatability was evaluated by six sample solutions prepared in parallel from S1.  reference, and common peaks in this reference chromatogram were assigned. e similarities between sample chromatograms and reference chromatogram were calculated using the abovementioned software.

Mass Spectrometry Conditions for UFLC-Triple-Q-TOF-MS/MS Analysis.
Identification of the components of common peaks in the HPLC fingerprint was performed on a UFLC-triple-Q-TOF-MS/MS system. Component separation was performed on a UFLC system (equipped with an LC-20AD XR quaternary pump, an SIL-20AC XR autosampler, and an SPD-M20 A DAD detector, Shimadzu, Kyoto, Japan) by using the same column with the same mobile phases and gradient conditions as mentioned above. e injection volumes of both the mixed reference substance solution and sample solution were 20 μL. After component separation by UFLC, a Triple TOF 4600 system (AB SCIEX, Framingham, USA) was employed to acquire mass spectra in the negative ion mode with a DuoSpray source. e mass spectrometric parameters were set as follows: curtain gas (CUR) 35 psi, nebulizer gas (gas (1)) 65 psi, heater gas (gas (2)) 65 psi, ion spray voltage 4500 V, and source temperature 550°C.
e TOFMS-IDA-10MS/MS method was used to obtain mass spectrometry data, and relevant parameters were set as follows: collision energy (CE) −10 eV, decluster potential (

Method Validation of the Quantitative Analysis.
e quantitative analysis method was validated by investigating the linear relationship, limit of detection (LOD), limit of quantitation (LOQ), precision, stability, repeatability, and recovery test of 12 components. e linear relationship was investigated by precisely injecting working solution A (10, 20, 30, and 40 μL) and working solutions B, C, and D (10 μL of each solution) into the HPLC system to calculate the regression equation, correlation coefficient, and linear range for all 12 components. After diluted, working solution D was injected into the HPLC system many times; LOQ and LOD were determined on the basis of signal-to-noise ratios of 10 : 1 and 3 : 1, respectively. Intraday precision, interday precision, and stability were assessed by RSDs of the peak areas of the 12 components. e intraday precision was determined by six consecutive injections of 30 μL working solution A, and the interday precision was determined by six replicate injections of 30 μL working solution A, twice per day over 3 consecutive days. e stability test was carried out by using the peak areas of the 7 iridoids, rutin, and jasminoside B at 254 nm, chlorogenic acid at 324 nm, and 2 crocins at 430 nm detected in Section 2.5 of the stability test. By calculating the contents of 12 components according to the peak areas of the 7 iridoids, rutin, and jasminoside B at 254 nm, chlorogenic acid at 324 nm, and 2 crocins at 430 nm detected in Section 2.5 of the repeatability test, and using the values of RSDs, the repeatability test was examined. For the recovery test, approximately 0.05 g S1 powder was precisely weighed, and 12 reference substance stock solutions were added at a sample/reference substance ratio of 1 : 1. Six sample solutions prepared in parallel by this method were analyzed, and the average recovery and RSDs of 12 components were calculated.

Validation of the Method for HPLC Fingerprint Analysis.
e RSDs of RPA and RRT for precision were no more than 4.56% and 0.14%, those for stability did not exceed 4.84% and 0.20%, and those for repeatability were less than 4.87% and 0.21%, respectively. e results met the fingerprinting quality standards for TCM injections [34]. Figure 2 and Table 1, 28 common peaks in the reference chromatogram were assigned. Similarities between the sample chromatograms and the reference chromatogram were all greater than 0.98.

Identification of the Common Peaks by Triple-Q-TOF-MS/MS.
A comparison between the negative ion mode and the positive ion mode revealed that the negative ion mode was much richer in information and thus was chosen for MS analysis. First, the total ion chromatograms of the sample and mixed reference substances ( Figure 3) were extracted using PeakView software. Second, the mass spectral data and dissociative rules of the reference substances were summarized, and it was revealed that the quasimolecular ion [M-H] − and/or [M+Cl] − could be selected as the precursor ions to generate MS/MS product ion spectra. Finally, the retention time, quasimolecular ion, and MS/MS fragmentation patterns were compared between samples and reference substances or those reported in the literature. Online retrieval was performed in the database of PubChem (http:// pubchem.ncbi.nlm.nih. Gov); therefore, the components of the 28 common peaks in the HPLC fingerprint were identified (the mass spectral data are given in Table 2, and the structures or possible structures of the components of 28 common peaks are shown in Figure 4).
S20 S19 S18 S17 S16 S15 S14 S13 S12 S11 S10 S9 S8 S7 S6 S5 S4 S3 S2 S1       Figure 5: HPLC chromatograms of the mixed reference substances (a) and sample (b). e number of peaks is the same as in Table 2. e abovementioned fragmentation pathways basically confirmed that the basic skeleton of the component of peak 16 was shanzhiside. erefore, peak 16 was identified as a component of shanzhiside substituted by sinapoyl at 6′-O, and a natural compound with this kind of structure was also found in the compound database PubChem. Referring to the names of the components of peaks 17 and 24, the component of peak 16 was temporarily named as 6′-transsinapoyl shanzhiside. To the best of our knowledge, this component was first detected in GF [1,[4][5][6][7][8][9]. Tables 3 and 4, the coefficient of determination values R 2 was greater than 0.9995, all RSDs of the intraday precision, interday precision, stability, and repeatability were less than 5%, the average recovery rates were 96.37-102.65%, and the RSDs were 1.80-4.30%. e above results met the requirements of the standard drug quality analysis method in the Chinese Pharmacopoeia [38].

Wavelength Selection for Quantitative Analysis of 12
Components. All 12 components could be detected at 254 nm, but the absorption of chlorogenic acid (peak 10) was stronger at 324 nm, and the absorptions of crocin I (peak 20) and crocin II (peak 25) were stronger at approximately 430 nm. erefore, a wavelength of 324 nm was selected for the detection of chlorogenic acid, and a wavelength of 430 nm was selected for the detection of crocin I and crocin II. e chromatograms of the mixed reference substances and sample are shown in Figure 5.

Contents of 12 Representative Components in 40 Samples.
As given in Table 5, there were certain differences in the contents of the 12 representative components among 40 samples, of which, the content of geniposide ranged from 37.917 to 72.216 mg/g, and the total content of the 7 iridoids ranged from 59.931 to 94.314 mg/g. Iridoids, especially geniposide, have both toxic and protective effects on the liver and kidney [16, 24-26, 39, 40]. It has been reported that the intragastric administration of 50 mg/kg/d bodyweight (human equivalent dose of 8 mg/kg/d bodyweight) geniposide in rats for 12 weeks can lead to liver and kidney damage [24]. According to this report, adults weighing 60 kg may suffer liver and kidney damage if they take 6 g or 10 g GF decoction pieces with a content of 80 mg/g or 48 mg/g every day for a long time.
e recommended clinical dose of GF decoction pieces is 6-10 g/d in the Chinese Pharmacopoeia [2]. Table 5 provides that the geniposide content in most batches of GF decoction pieces exceeded 48 mg/g. erefore, the content of representative components such as geniposide in GF decoction pieces should be measured before clinical use, and the dose of GF decoction pieces should be adjusted according to the content of these components to achieve a therapeutic effect and avoid adverse reactions.

Conclusion
In this study, QE of 40 batches of decoction pieces of GF produced by different manufacturers of herbal pieces is performed by qualitative analysis of the HPLC fingerprint and UFLC-triple-Q-TOF-MS/MS combined with quantitative analysis of multiple components, which we established previously for QE of traditional medicine. e results show that there are 28 common peaks in the HPLC fingerprints of 40 samples. e similarities between the sample chromatograms and reference chromatogram were higher. e components of these 28 common peaks are identified as 13 iridoids, 4 crocins, 7 monocyclic monoterpenoids, 3 organic acids, and 1 flavonoid. Of these, a total of 12 components, including the seven iridoids geniposide, shanzhiside, geniposidic acid, DAAMS, gardenoside, SME, and genipin gentiobioside, crocin I and crocin II, the monocyclic monoterpenoid jasminoside B, and the organic acid chlorogenic acid and the flavonoid rutin, were unambiguously identified by comparison with reference substances. ere were certain differences in the contents of these 12 components among 40 samples; the geniposide content ranged from 37.917 to 67.039 mg/g, the total content of the 7 iridoids ranged from 37.917 to 67.039 mg/g, and the total content of 7 iridoids ranged from 59.931 to 87.843 mg/g. e content of representative components, such as geniposide, in GF decoction pieces should be measured before clinical use, and the dose of GF decoction pieces should be adjusted according to the content of these components to achieve a therapeutic effect and avoid adverse reactions.

Data Availability
e data used to support the findings of this study are included within the article.

Conflicts of Interest
e authors declare that they have no conflicts of interest.