Pharmacokinetic and Metabolism Studies of Curculigoside C by UPLC-MS/MS and UPLC-QTOF-MS

Pharmacokinetic and metabolism studies were carried out on curculigoside C (CC), a natural product with good antioxidant and neuroprotective effects, with the purpose of investigating the effects of the hydroxyl group at C-3′ in curculigoside. A rapid and sensitive method with UPLC-MS was developed and fully validated for the first time in the pharmacokinetic analysis for quantification of CC in rat plasma. The assay was linear (R2 > 0.9984) over the concentration range of 1–2500 ng/mL, with the lower limit of quantification (LLOQ) being 1 ng/mL. The intra-day and inter-day precision (expressed as relative standard deviation, RSD) ranged from 4.10% to 5.51% and 5.24% to 6.81%, respectively. The accuracy (relative error, RE) ranged from −3.28% to 0.56% and −5.83% to −1.44%, respectively. The recoveries ranged from 92.14% to 95.22%. This method was then applied to a pharmacokinetic study of rats after intragastric administration of 15, 30 and 60 mg/kg CC. The results revealed that CC exhibited rapid oral absorption (Tmax = 0.106 h, 0.111 h, and 0.111 h, respectively), high elimination (t1/2 = 2.022 h, 2.061 h, and 2.048 h, respectively) and low absolute bioavailability (2.01, 2.13, and 2.39%, respectively). Furthermore, an investigation on the metabolism of CC was performed by UPLC-QTOF-MSE. Twelve metabolites of CC from plasma, bile, urine and faeces of rats were confirmed. The main metabolic pathways of CC, which involve dehydration, glucosylation, desaturation, formylation, cysteine conjugation, demethylation and sulfonation, were profiled. In conclusion, this research has developed a sensitive quantitative method and demonstrated the metabolism of CC in vivo.

Curculigoside C (CC) is the natural product of C-3 hydroxylation of curculigoside A [16]. It was reported that CC has better antioxidant and neuroprotective effects that the parent compound [17,18]. For example, in the assay on hydroxyl radicals produced by H 2 O 2 /Fe 2+ , CC exhibited more significant scavenging effects than curculigoside A, and the scavenging effect of CC was comparable with that of pigallocatechin gallate, a known antioxidant [17]. Moreover, in the 1,1-diphenyl-2-picrylhydrazyl (DPPH) radical scavenging assay, CC also showed better antioxidant activity than curculigoside A, and had a comparable IC 50 value to vitamin C, another well-known antioxidant [18]. Furthermore, in the evaluation of salvaging SY5Y cell death induced by H 2 O 2 , CC displayed better neuroprotective effect than curculigoside A [18]. The discussion about the structure-activity relationship showed that the presence of OH at C-3 and the two vicinal oxygen-bearing groups at the benzene ring might be the functional groups [18]. It is known that the pharmacokinetic parameters are necessary for the evaluation of active ingredients [19,20]. To the best of our knowledge, there are no reports on pharmacokinetic assays of CC. In the present research, CC was used for pharmacokinetic and metabolism studies. Due to the high sensitivity and rapid quantification [21,22], an ultra-high performance liquid chromatography combined with tandem quadrupole mass spectrometry (UPLC-MS/MS) method was developed and validated for the determination of CC in rat plasma. This method was then applied to pharmacokinetic study of CC in rats. In addition, the metabolic characteristics of CC were also studied to elucidate the dynamic process of CC in rats. The aims were to investigate the pharmacokinetics and metabolic characteristics of CC with a hydroxyl group at C-3 . In summary, this exploration provides a sensitive quantitative assay method and illustrated the pharmacokinetic characteristics of CC in rats.

Method Development
Aiming at increasing the extraction recovery and minimizing the matrix effect, different sample pre-treatment methods such as solid-phase extraction, liquid-liquid extraction, or protein precipitation were comparatively investigated. As a consequence, the optimized one-step protein precipitation using methanol was selected. Several additions were tested to improve the sensitivity of the CC and curculigoside B (IS), 0.1% formic acid was chosen. No significant signal diminution or enhancement was found under the current conditions used in this research. The IS has a similar structure, chromatographic behavior and extraction efficiency to those of CC and both of them showed strong mass responses in positive ESI mode. ESI + showed that CC (MW: 482.1) and IS (MW: 452.1) formed predominately molecular ions [M + Na] + at m/z 505.1 and mainly deprotonated molecular ions [M + H] + at m/z 453.1 in full-scan product ion spectra, respectively. Precursor ion and product ions were selected according to the stability and ion response. Mass spectrometric scanning was operated on multiple reaction monitoring (MRM) using follow monitored transitions: m/z 505.1 → 221.0 for CC, m/z 453.1 → 291.1 for IS, respectively. The f ragmentation pathways of CC and the IS are shown in Figure 1.

Method Validation
The detection of CC (RT, 0.98 min) and IS (RT, 1.44 min) showed high selectivity. No significant interference of endogenous plasma components was observed at the retention time of CC and the IS. Typical chromatograms of blank plasma (A), blank plasma spiked with CC at 100 ng/mL and IS at 100 ng/mL (B), and plasma sample collected from a rat at 0.083 h after an oral dose of 60 mg/kg CCwere shown in Figure 2.

Method Validation
The detection of CC (RT, 0.98 min) and IS (RT, 1.44 min) showed high selectivity. No significant interference of endogenous plasma components was observed at the retention time of CC and the IS. Typical chromatograms of blank plasma (A), blank plasma spiked with CC at 100 ng/mL and IS at 100 ng/mL (B), and plasma sample collected from a rat at 0.083 h after an oral dose of 60 mg/kg CCwere shown in Figure 2.

Method Validation
The detection of CC (RT, 0.98 min) and IS (RT, 1.44 min) showed high selectivity. No significant interference of endogenous plasma components was observed at the retention time of CC and the IS. Typical chromatograms of blank plasma (A), blank plasma spiked with CC at 100 ng/mL and IS at 100 ng/mL (B), and plasma sample collected from a rat at 0.083 h after an oral dose of 60 mg/kg CCwere shown in Figure 2.  The calibration curve of CC in rat plasma indicated a high linearity in the range of 1-2500 ng/mL with a correlation coefficient greater than 0.99 (r 2 = 0.9984). There was no significant carry-over under the assay conditions. The LLOQ was 1 ng/mL, which was sufficient for the detection of CC in pharmacokinetic study. The intra-day and inter-day precision (RSD, %) ranged from 4.10% to 5.51% and 5.24% to 6.81%, respectively. The intra-day and inter-day accuracy (RE, %) ranged from −3.28% to 0.56% and −5.83% to −1.44%, respectively ( Table 1). The variation of the IS measured value was less than 10%. The best extraction recovery of CC was obtained using methanol as the protein-precipitating agent. The recoveries of CC at 3, 200 and 2000 ng/mL were 95.22 ± 5.64%, 92.82 ± 8.74% and 92.14 ± 3.45%, respectively. The recovery of the IS was 94.40 ± 4.23%. The results showed that the preparation efficiency of CC and IS in this present study was acceptable.
The matrix effect for CC was evaluated by analyzing three concentrations of QC plasma samples (3,200, 2000 ng/mL). While the matrix effect for IS was evaluated with a single concentration (100 ng/mL). The average matrix effect values were 94.80 ± 4.06%, 91.49 ± 6.68% and 91.73 ± 3.67% for CC at the low, medium and high QC concentrations, respectively. The matrix effect on IS turned out to be 92.94 ± 5.91% at the tested concentration.
After being placed at 25 °C for 4 h or at −20 °C for two weeks, or undergoing three freeze-thaw (−20 °C to 25 °C) cycles, the results of stability (Table 2) demonstrated that CC was stable. The calibration curve of CC in rat plasma indicated a high linearity in the range of 1-2500 ng/mL with a correlation coefficient greater than 0.99 (r 2 = 0.9984). There was no significant carry-over under the assay conditions. The LLOQ was 1 ng/mL, which was sufficient for the detection of CC in pharmacokinetic study. The intra-day and inter-day precision (RSD, %) ranged from 4.10% to 5.51% and 5.24% to 6.81%, respectively. The intra-day and inter-day accuracy (RE, %) ranged from −3.28% to 0.56% and −5.83% to −1.44%, respectively ( Table 1). The variation of the IS measured value was less than 10%. The best extraction recovery of CC was obtained using methanol as the protein-precipitating agent. The recoveries of CC at 3, 200 and 2000 ng/mL were 95.22 ± 5.64%, 92.82 ± 8.74% and 92.14 ± 3.45%, respectively. The recovery of the IS was 94.40 ± 4.23%. The results showed that the preparation efficiency of CC and IS in this present study was acceptable.
The matrix effect for CC was evaluated by analyzing three concentrations of QC plasma samples (3,200, 2000 ng/mL). While the matrix effect for IS was evaluated with a single concentration (100 ng/mL). The average matrix effect values were 94.80 ± 4.06%, 91.49 ± 6.68% and 91.73 ± 3.67% for CC at the low, medium and high QC concentrations, respectively. The matrix effect on IS turned out to be 92.94 ± 5.91% at the tested concentration.
After being placed at 25 • C for 4 h or at −20 • C for two weeks, or undergoing three freeze-thaw (−20 • C to 25 • C) cycles, the results of stability (Table 2) demonstrated that CC was stable.

Pharmacokinetic Study
The developed approach was successfully applied to our pharmacokinetic studies after oral administration of 15, 30, 60 mg/kg or intravenous injection of 2.0 mg/kg CC in rats (n = 6), respectively. All of the data were calculated with the DAS 3.0. statistical software (Shanghai Bojia Pharmatech Co. Ltd., Shanghai, China). The mean plasma concentration versus time curves and the major pharmacokinetic parameters were shown in Figure 3 and Table 3, respectively. It could be concluded that CC was quickly cleared (clearance, CL, 4.87 ± 0.83 L/h/kg; elimination half-life, t 1/2 , 1.15 ± 0.20 h) and had high extravascular distribution (apparent volume of distribution, V d , 8.12 ± 1.97 L/kg) after intravenous administration. As for the three doses of CC after oral administration, the absorption rate of CC was also ultrafast as demonstrated by the fact that CC was detected in plasma at the first blood sampling time (0.05 h) and reached the maximum concentration at 0.083 h. On the other hand, it was concluded that CC was also eliminated rapidly with the t 1/2 from 2.02 h to 2.06 h and CL in the range of 201.90-234.19 L/h/kg through the gastrointestinal tract. Whereas the parameter of V d in the range of 585.34-722.61 L/kg inferred that CC distributed in tissues extensively after gastrointestinal administration. The higher CL values and the rapid T max demonstrated that the oral absorption of CC was poor. Furthermore, the poor absorption as well as the first pass metabolism led to the low AUC values and the poor oral bioavailability (F, 2.01%, 2.13%, and 2.39%, respectively). These observations suggested that this validated analytical LC-MS/MS method was suitable and sufficient for pharmacokinetic study of CC. In short, the characteristic pharmacokinetic properties of CC were rapid oral absorption, high clearance and poor absolute bioavailability. And the short t 1/2 revealed that CC was easily metabolized in vivo. The above characteristics are similar to those of most of the phenolic glycosides reported in the literature [14].

Metabolites Identification of CC
For the determination of fragmentation patterns, the reference compounds of CC were applied for the main MS/MS fragments in UPLC-QTOF-MSE. As Table 4  In the UPLC elution, the retention time of CC was 6.80 min, and the molecular ion in the mass spectrum is m/z 505.1316. Samples of CC-dosed plasma, bile, urine and feces were analyzed in parallel with blank controls through the UPLC-Q/TOF-MS method. Despite the lack of standards for metabolites, their structures could be evaluated on the basis of retention times and mass spectrometry patterns between CC and its product ions.
Based on certain rules, mass accuracy (± 5 ppm), nitrogen rule, isotopic pattern, and doublebond equivalents for instance, the most likely molecular formulas of metabolites were examined. Moreover, according to the MS/MS fragmentation, the tentative chemical structures were determined and common metabolic pathways were profiled. Twelve metabolites including seven phase I metabolites and five phase II metabolites are shown in Figure 4.

Metabolites Identification of CC
For the determination of fragmentation patterns, the reference compounds of CC were applied for the main MS/MS fragments in UPLC-QTOF-MSE. As Table 4  In the UPLC elution, the retention time of CC was 6.80 min, and the molecular ion in the mass spectrum is m/z 505.1316. Samples of CC-dosed plasma, bile, urine and feces were analyzed in parallel with blank controls through the UPLC-Q/TOF-MS method. Despite the lack of standards for metabolites, their structures could be evaluated on the basis of retention times and mass spectrometry patterns between CC and its product ions.
Based on certain rules, mass accuracy (±5 ppm), nitrogen rule, isotopic pattern, and double-bond equivalents for instance, the most likely molecular formulas of metabolites were examined. Moreover, according to the MS/MS fragmentation, the tentative chemical structures were determined and common metabolic pathways were profiled. Twelve metabolites including seven phase I metabolites and five phase II metabolites are shown in Figure 4.   13 , 28 Da higher than CC, which indicated that it was a formylated product of CC. Abundant daughter ions at m/z 331.0715 in the MS 2 spectrum, which were 28 Da higher than the fragment ion at m/z 303.0952 of CC, support the notion that decarboxylation had occured to the parent nucleus. The fragment ion at m/z 331.0715 was conjectured to be the formylated product of the typical fragment of CC at m/z 303.0952. In conclusion, an effective strategy using UPLC-Q-TOF-MSE coupled with the UNIFI 1.7.0 software (Waters, Manchester, UK) for quick characterising and identificating metabolites of CC was developed. In our study, the pathways of Phase I metabolites included dehydration, hydrolysis, deglycosylation, desaturation, and for Phase II metabolites they were sulfation, glucosylation, cysteine conjugation, formylation, demethylation. These results could provide a theoretical foundation for understanding the pharmacological effects and metabolic processes of CC.

Discussion
Targeting at exploring the characteristics of the phenolic glycosides, the pharmacokinetics and metabolism of CC were investigated. First of all, a UPLC-MS/MS method with high sensitivity (1 ng/mL) and a short running time (3.5 min), was validated and used for the pharmacokinetic study. Secondly, the pharmacokinetic analysis elucidated that CC was rapidly absorbed and quickly eliminated, which is similar to curculigoside A, but CC showed better bioavailability (2.39%) than that of curculigoside A (0.38%) [15], which might be the reason for the better activity of CC. Finally, the major metabolites of Phase I and Phase II reactions and the proposed metabolic pathways of CC were both profiled. Furthermore, the results of metabolite identification demonstrated that CC was mainly metabolized in the liver, since all metabolites could be found in bile, urine and feces samples. Considering the results acquired in this research, we deduce that the poor oral bioavailability of CC in rats might be related to hepatic microsomes. Since CC acted as an antioxidant, it is worth mentioning the quinone metabolites, which were not identified under the screening conditions used in the samples of healthy rats. The reason might be as follows: in order to obtain the main metabolites with high response and good mass accuracy, the metabolites with worse mass accuracy (out of the range of ±5 ppm) and low response (below 5000) were filtered firstly before the identification of metabolites. During this step, quinones might be filtered due to their low content. In future studies, the model rats with peroxidation injury may be used to investigate possible quinone metabolites. Besides the common pathways of curculigoside A, such as dehydration and deglycosylation, four other metabolic pathways including formylation, desaturation, cysteine conjugation and sulfonation also had been discovered in CC. More importantly, the -OH at C-3 still existed in all the metabolites, which further confirmed that the -OH at C-3 might be the functional group of CC, but the low absolute bioavailability might limit its further application. Future studies should focus on structure modification to improve the bioavailability.

Materials and Reagents
CC (purity: 99.1%) was isolated from Curculiginis rhizoma in our laboratory and identified by HR-MS and Nuclear Magnetic Resonance (NMR) spectroscopy. Its purity was determined by using high performance liquid chromatography (HPLC). Internal standard (IS, Curculigoside B) was provided by Wuhan Tian Zhi Biotechnology Co., Ltd. (Hubei, China). 3-Hydroxy-2,6-dimethoxybenzoic acid was obtained from HE Chemical Co., Ltd. (Jiangsu, China). The chemical structures of CC and Curculigoside B are shown in Figure 5. UPLC-MS pure grade methanol and acetonitrile were purchased from Fisher (Geel, Belgium). Formic acid was acquired from the Sigma-Aldrich Company (St. Louis, MI, USA). Deionized water was purified with a Millipore water purification system (Millipore, Billerica, MA, USA). All other chemicals were of analytical grade. Blank rat plasma samples (drug-free and anti-coagulated with heparin sodium) were prepared by our group.
Molecules 2018, 23, x 10 of 14 also had been discovered in CC. More importantly, the -OH at C-3′ still existed in all the metabolites, which further confirmed that the -OH at C-3′ might be the functional group of CC, but the low absolute bioavailability might limit its further application. Future studies should focus on structure modification to improve the bioavailability.

Materials and Reagents
CC (purity: 99.1%) was isolated from Curculiginis rhizoma in our laboratory and identified by HR-MS and Nuclear Magnetic Resonance (NMR) spectroscopy. Its purity was determined by using high performance liquid chromatography (HPLC). Internal standard (IS, Curculigoside B) was provided by Wuhan Tian Zhi Biotechnology Co., Ltd. (Hubei, China). 3-Hydroxy-2,6dimethoxybenzoic acid was obtained from HE Chemical Co., Ltd. (Jiangsu, China). The chemical structures of CC and Curculigoside B are shown in Figure 5. UPLC-MS pure grade methanol and acetonitrile were purchased from Fisher (Geel, Belgium). Formic acid was acquired from the Sigma-Aldrich Company (St. Louis, MI, USA). Deionized water was purified with a Millipore water purification system (Millipore, Billerica, MA, USA). All other chemicals were of analytical grade. Blank rat plasma samples (drug-free and anti-coagulated with heparin sodium) were prepared by our group.

Animals and Drug Administration
All animal experiments were approved (permit number: 20180057) by the Review Committee of Animal Care and Use of Jilin University according to ethical principles for animal use and care. Wistar rats (200 ± 20 g) were purchased from Changchun Yisi Laboratory Animal Co. Ltd. (Changchun, China), and were housed under standard conditions in a controlled breeding room (12 h light/dark cycle, temperature: 22 ± 2 °C, relative humidity: 55 ± 5%). All rats were fed with standard laboratory food and water ad libitum, except during a fast period prior to the experiments. For the pharmacokinetic study, the rats were divided into four groups (n = 6, 3 males and 3 females): (1) CC (15 mg/kg, i.g.), (2) CC (30 mg/kg, i.g.), (3) CC (60 mg/kg, i.g.) and (4) CC (2.0 mg/kg, i.v.). While for the metabolism study, CC (60 mg/kg, i.g.) was gastrointestinally administrated to achieve higher concentrations of the metabolites.

Pharmacokinetic Study
The stock solution of CC and IS was prepared by dissolving the accurately weighed reference compound in methanol at 1.00 mg/mL, respectively. The blank rat plasma was prepared as follows: blood samples were collected from the abdominal aorta immediately and incubated at 37 °C for half an hour, then a 15 min centrifugation at 4000 rpm was carried out to separate the plasma, the supernatant was kept at −20 °C until analysis. After dilutions were spiked with blank rat plasma, the calibration curve for CC was plotted to produce the points equivalent to 1, 2.5, 5, 25, 50, 250, 500, 2500

Animals and Drug Administration
All animal experiments were approved (permit number: 20180057) by the Review Committee of Animal Care and Use of Jilin University according to ethical principles for animal use and care. Wistar rats (200 ± 20 g) were purchased from Changchun Yisi Laboratory Animal Co. Ltd. (Changchun, China), and were housed under standard conditions in a controlled breeding room (12 h light/dark cycle, temperature: 22 ± 2 • C, relative humidity: 55 ± 5%). All rats were fed with standard laboratory food and water ad libitum, except during a fast period prior to the experiments. For the pharmacokinetic study, the rats were divided into four groups (n = 6, 3 males and 3 females): (1) CC (15 mg/kg, i.g.), (2) CC (30 mg/kg, i.g.), (3) CC (60 mg/kg, i.g.) and (4) CC (2.0 mg/kg, i.v.). While for the metabolism study, CC (60 mg/kg, i.g.) was gastrointestinally administrated to achieve higher concentrations of the metabolites.

Pharmacokinetic Study
The stock solution of CC and IS was prepared by dissolving the accurately weighed reference compound in methanol at 1.00 mg/mL, respectively. The blank rat plasma was prepared as follows: blood samples were collected from the abdominal aorta immediately and incubated at 37 • C for half an hour, then a 15 min centrifugation at 4000 rpm was carried out to separate the plasma, the supernatant was kept at −20 • C until analysis. After dilutions were spiked with blank rat plasma, the calibration curve for CC was plotted to produce the points equivalent to 1, 2.5, 5, 25, 50, 250, 500, 2500 ng/mL. With the same treatment as calibration standards, 3, 200 and 2000 ng/mL of quality control (QC) samples were prepared independently. All samples were prepared freshly before analysis.
All samples were kept at −20 • C before analysis. Frozen plasma samples were thawed to room temperature and vortexed. CC was extracted from plasma only by single step precipitation method. Both 50 µL of each rat plasma sample and 500 µL of methanol containing 100 ng/mL IS were transferred into 2.0 mL microcentrifuge tubes, and then extracted by vortexing for 3 min to deproteinize the endogenous protein. After a 10 min centrifugation step at 10,000 rpm and 4 • C, the supernatant was collected and transferred into vials, and 2 µL of it was injected into UPLC-MS system for quantification. All prepared samples were stored in the autosampler at 10 • C until injection.

Metabolism Study
After intragastric administration, blood samples were collected at 0.25, 0.5, 0.75, 1, 1.5, 2, 2.5, 3, 6, 12 h. After vortexing 500 µL of pooled plasma sample with 2 mL of methanol, the mixtures were centrifuged at 10,000 rpm and 4 • C for 10 min to obtain the clear supernatant, which was evaporated with N 2 . Then, the residue was reconstituted in 100 µL of methanol.
To obtain bile samples, a plastic cannula was surgically inserted into the bile ducts to collect the bile through an abdominal incision made after rats were anesthetized with intraperitoneal injection of urethane (1.0 g/kg). Blank bile was collected for 2 h before dosed, and bile samples were collected for 12 h after administration. While the blank and dosed samples of urine and feces were obtained respectively with metabolic cages equipped with separator prior to drug administration or 18 h after oral administration. The mixture was centrifuged for 10 min at 10,000 rpm and 4 • C after diluting pooled urine and bile sample with triple volume of methanol. The obtained supernatant was evaporated by N 2 at room temperature after transferred into another tube. The dried residue of urine or bile was dissolved in 50 µL of methonal, respectively. Finally, the supernatant being centrifuged for 10 min at 10,000 rpm and 4 min was applied to UPLC-MS/MS analysis.
Faeces samples were dried and crushed into powder before being stored. Each 10 mg of faeces powder was added with 1 mL of methanol and was extracted for 30 min using an ultrasonic ice-water bath. After centrifugation at 10,000 rpm and 4 • C for 10 min, the supernatant was transferred to a clean tube and then evaporated to dryness under a stream of N 2 . The residue reconstituted with 500 µL methanol was centrifuged at 10,000 rpm and 4 • C for 10 min. All biological samples were stored at −20 • C.

LC-MS/MS Conditions
Quantification of CC was performed on an Acquity UPLC unit coupled with a XEVO TQ-S mass spectrometer equipped with an electrospray ionization (ESI) source (Waters Co.). Separation was achieved using a Waters BEH C18 UPLC column (2.1 mm × 50 mm, 1.7 µm) at 40 • C. The optimized method used binary gradient mobile phases with 0.1% formic acid in acetonitrile as mobile phase A and 0.1% formic aqueous solution as mobile phase B (0-2 min, 20% → 30% A; 2-2.5 min, 30% → 20% A; 2.5-3.5 min, 20% A). A flow rate of 0.3 mL/min was used with 2 µL of injection volume. Chromatography of the CC and IS was performed within 3.5 min. CC was quantified using multiple reaction monitoring (MRM) in the positive ESI mode and the optimized MS parameters were set as follows: capillary voltage floating at 3200 V; 150 • C for source temperature; 300 • C for desolvation temperature; 150 L/h for flow of cone gas; 800 L/h for flow of desolvation gas; 0.15 mL/min for flow of collision gas; nebulizer gas flow at 7 bar; cone voltage at 70 and 48 V for CC and IS, respectively; Collision energy at 20 and 18 eV for CC and IS, respectively. Data acquisition and processing were operated through Masslynx V4.1 workstation.

UPLC-QTOF/MS Conditions
Metabolic analysis was conducted on a Waters ACQUITY UPLC System coupled with a Xevo G2-S Q-TOF mass spectrometer in ESI + mode. A Waters UPLC BEH C 18 column (2.1 mm × 50 mm, 1.7 µm) was used with the following parameter settings: 30 • C for the column temperature; 0.4 mL/min for flow rate; 10 µL for injection volume; 4 • C for the autosampler temperature; and the mobile phase consisted of 0.1% formic aqueous solution (A) and 0.1% formic acid in acetonitrile (B) in proportions adjusted through a gradient elution programme as follows: 0-2 min, 10% B; 2-26 min, 10% → 100% B; 26-28 min, 100% B; 28-28.1 min, 100% → 10% B, 28.1-30 min, 10% B. The following optimized conditions were carried out: 2.6 kV for capillary voltage; 40 V for cone voltage; 120 • C for source temperature; 300 • C for desolvation temperature; 50 L/h for the flow of cone gas; 800 L/h for the flow of desolvation gas. MSE acquisition mode is an intelligent approach, which collects exact-mass precursor and product ion information from a single injection in a data-independent manner. It could parallel alternate scans at either low collision energy to obtain precursor ion information, or high collision energy to obtain full-scan accurate mass fragment, precursor ion and neutral loss information. In doing so, it excludes false positive results; requires no "knowledge" of the ions to be fragmented; runs parallel precursor, product ion, and neutral loss analyses; and streamlines data interrogation and reporting. In MS E mode, the collision energy of low energy function and high energy function was set at 6 V and 20-40 V, respectively. The mass spectrometer was calibrated over a range of 100-1000 Da with sodium formate to ensure mass accuracy. Leucine-enkephalin (m/z 556.2771) was used as the lockmass at a concentration of 200 ng/mL and flow rate of 10 µL/min. Masslynx V4.1 workstation in continuum mode was used for data collection. Metabolic characterization of CC was analyzed using UNIFI 1.7.0 software (Waters, Manchester, UK).

Method Validation
According to the Bioanalytical Method Validation Guideline (Chinese Pharmacopoeia 2015, Vol. 4) and Drug Non-Clinical Pharmacokinetic Study Technical Guideline (China Food And Drug Administration 2014), specificity, accuracy and precision, linearity, LLOQ, extraction recovery and matrix effect had been validated in this present study.
Specificity was investigated by analyzing chromatograms of blank rat plasma, plasma spiked with CC and IS, and experimental plasma sample following dosing of CC. Linearity was analyzed through weighted regression (1/x 2 ) of peak area ratios (y) of CC to IS versus nominal concentration (x) in plasma. The blank plasma has been run after the determination of upper concentration (2500 ng/mL) of the standard curve aiming at determine the carry-over in the assay condition. Precision and accuracy were evaluated by analyzing three concentration levels (3,200 and 2000 ng/mL) of QC samples. The intra-, inter-day precision and accuracy were assessed by determining at three QC levels (3,200 and 2000 ng/mL) in replicates of six QC samples. The precision was expressed as relative standard deviation (RSD, %) and the accuracy was expressed as relative error (RE, %). Both of the values for QC concentrations within 15% were acceptable. The lowest limit of quantification (LLOQ) was the lowest concentration on calibration curve, with acceptable precision (RSD ≤ 20%) and accuracy (RE ≤ ±20%). The extraction recovery of CC from rat plasma at three QC concentrations was assessed by comparing peak areas of the QC samples pre-spiked in blank plasma (A) with those of post-extracted blank plasma spiked at the same concentration (B). The matrix effect of CC from rat plasma was evaluated by comparing the peak areas of post-extracted spiked rat plasma (B) with those of equivalent concentrations of the pure authentic standard at three QC concentrations (C). The recovery and matrix effect of IS were evaluated at 100 ng/mL in the same way as CC. Stability in plasma samples was investigated by analyzing QC samples (n = 6) under various storage conditions: (1) at room temperature for 4 h; (2) at −20 • C for two weeksl; (3) subjected to three complete freeze/thaw cycles (from −20 • C to room temperature) on consecutive days; (4) stored in the stock solutions (4 • C) and in methanol with plastic autosampler vials for 16 h at 10 • C.

Pharmacokinetic Study
The concentration versus the time curve was plotted. The pharmacokinetic parameters, including the maximum plasma concentration (C max ) and the time to reach the peak concentration (T max ) were obtained in rat plasma after intragastric administration, elimination half-life (t 1/2 ), area under the plasma concentration-time curve (AUC), were analyzed through a non-compartmental pharmacokinetic analysis carried out by Drug and Statistics (DAS) 3.0 pharmacokinetic software programme (Mathematical Pharmacology Professional Committee of China, Shanghai, China). The results were presented as the mean ± standard deviation. Absolute bioavailability (F%) was determined through the equation (AUC p.o. × Dose i.v. )/(AUC i.v. × Dose p.o. ) × 100%.

Conclusions
In the present assay, the pharmacokinetics and metabolism of curculigoside C (CC) were investigated for the first time. In the pharmacokinetic analysis, a rapid, sensitive and reproducible UPLC-ESI-MS/MS quantification method was developed and validated for determination of CC in rat plasma samples. The sample preparation process was simple and the analysis time was just 3.5 min. The method was then successfully applied to the pharmacokinetic study after intravenous administration of 2.0 mg/kg CC and oral administration of 15, 30 and 60 mg/kg CC to rats, respectively. The results showed that CC exhibited rapid oral absorption (T max = 0.11 h), high elimination (t 1/2 = 2.01 h) and poor absolute bioavailability (2.39%). A metabolic investigation of CC was conducted through UPLC-QTOF-MS E . The major metabolites and metabolic pathways were all characterized. In general, CC, another phenolic glycoside in comparation with curculigoside A, has the similar pharmacokinetic parameters and metabolic pathways, but better absolute bioavailability. These findings provide foundation for pharmacological research and development of CC.