Degradation Profiling of Nardosinone at High Temperature and in Simulated Gastric and Intestinal Fluids

Nardosinone, a predominant bioactive product from Nardostachys jatamansi DC, is well-known for its promising therapeutic applications, such as being used as a drug on anti-inflammatory, antidepressant, cardioprotective, anti-neuroinflammatory, anti-arrhythmic, anti-periodontitis, etc. However, its stability under varying environmental conditions and its degradation products remain unclear. In this study, four main degradation products, including two previously undescribed compounds [2–deoxokanshone M (64.23%) and 2–deoxokanshone L (1.10%)] and two known compounds [desoxo-narchinol A (2.17%) and isonardosinone (3.44%)], were firstly afforded from the refluxed products of nardosinone in boiling water; their structures were identified using an analysis of the extensive NMR and X–ray diffraction data and the simulation and comparison of electronic circular dichroism spectra. Compared with nardosinone, 2–deoxokanshone M exhibited potent vasodilatory activity without any of the significant anti-neuroinflammatory activity that nardosinone contains. Secondly, UPLC–PDA and UHPLC–DAD/Q–TOF MS analyses on the degradation patterns of nardosinone revealed that nardosinone degraded more easily under high temperatures and in simulated gastric fluid compared with the simulated intestinal fluid. A plausible degradation pathway of nardosinone was finally proposed using nardosinonediol as the initial intermediate and involved multiple chemical reactions, including peroxy ring-opening, keto–enol tautomerization, oxidation, isopropyl cleavage, and pinacol rearrangement. Our findings may supply certain guidance and scientific evidence for the quality control and reasonable application of nardosinone-related products.

Nardosinone (CAS No: 23720-80-1; molecular formula: C 15 H 22 O 3 ; molecular weight: 250.1569) is a colorless crystal or a white amorphous powder with a chemical structure of nardosinane-type sesquiterpenoid that features a unique peroxide bridge between C-7 and C-11 [1]. Nardosinone can be derived from gansongone, an aristolane-type sesquiterpenoid that has been isolated from Nardostachys plants by exposure to air oxygen or through a liquid phase autoxidation (such as 1 O 2 , 3 O 2 , and H 2 O 2 ) in the presence of different phenolic reagents (e.g., phenols, quercetin, and naphthol) under dark conditions [19]. The intrinsic five-membered peroxy ring in nardosinone is unstable and is prone to having a ringopening reaction occur [20]. A key intermediate metabolic pathway may be undergone by degradation or structural modification of nardosinone. Liu et al. found that when dissolved in ethanolic solution, nardosinone is largely stable in an alkaline environment; in contrast, it rapidly degrades in acidic and high-temperature conditions. In a solid state, nardosinone is relatively more stable under the conditions of high temperature and high humidity compared with its unstable and readily biodegradable properties in the condition of strong light [21]. Furthermore, in a boiling methanol solution [20], nardosinone was shown to degrade into C 12 -norsesquiterpenoids (e.g., desoxo-narchinol A) or other nardosinanetype C 15 -sesquiterpenoids (e.g., nardosinonediol, isonardosinone), which have also been isolated from the same plant of N. jatamansi.
Given the above-mentioned unstable property of nardosinone, it is questionable whether nardosinone exerts the therapeutic effects in a prototype form or not for the medicinal herbs and patent prescriptions that have nardosinone as the main component. Thus, the chemical unstable property is suggested to require adequate consideration in future in vivo investigations on the pharmacokinetic characteristics of nardosinone when orally administered. In light of this, two pharmacokinetic studies of nardosinone were reported using intravenous injection [22] and oral ingestion [23] as the routes of administration. However, to date, the metabolites of nardosinone have not been characterized and investigated in depth.
In light of this, our study aimed to isolate and identify the main degradation products of nardosinone after being refluxed in boiling water by combining several modern chromatographic separations with extensive spectroscopic analyses. Moreover, we attempted to profile the plausible degradation pattern of nardosinone by means of ultra-high performance liquid chromatography-photo-diode array (UPLC-PDA) and ultra-high performance liquid chromatography-diode array detector-quadrupole time-of-flight mass spectrometry (UHPLC-DAD/Q-TOF MS). The overall goal was to provide guidance and scientific evidence for future studies on the quality control and reasonable application of nardosinone-related products. The workflow of the entire study is presented in Figure 1.

Evaluation of the Vasodilatory and Anti-Neuroinflammatory Activities of Nardosinone and the Main Degradation Products
Previous studies have revealed that nardosinone possesses significant antineuroinflammatory [10] and vasodilatory (IC 50 254.59 µM) effects [27]. In this study, these two activities of nardosinone and its primary degradation product (2-deoxokanshone M, 1) were evaluated based on our previously reported methods [26,27]. As shown in Figure 7, nardosinone could significantly inhibit the inflammatory factor NO production of the LPS-induced neuroinflammation model in BV-2 microglia cells, which is consistent with previous results reported (IC 50 : 37.82-74.21 µM) [10] and compared with minocycline (IC 50 = 23.69 ± 2.01 µM) [28]. At the same time, 2-deoxokanshone M (1) failed to exhibit any inhibitory effect (Figure 7a,b). Other main degradation products, isonardosinone and desoxo-narchinol A, have been reported to exhibit significant anti-neuroinflammatory activities with IC 50 values of 37.82-74.21 µM [10] and 3.48 ± 0.47 µM [29], respectively. As presented in Figure 7c, 2-deoxokanshone M displayed a vasorelaxation activity that was more potent than nardosinone. Specifically, the vasodilation rate of nardosinone was 74.57% at a concentration of 400 µM, while that of 2-deoxokanshone M already reached 80.39% at a lower concentration of 75 µM.
Given these results, it appears that the degradation mentioned above is beneficial for nardosinone in the treatment of vasodilation-related diseases, exemplified by hypertension [27], and is unfavorable for nardosinone in the aid of neuroinflammation-associated neurodegenerative disorders such as Alzheimer's disease, parkinsonian syndromes, and amyotrophic lateral sclerosis [30].

Degradation of Nardosinone in Different Conditions
As shown in Figure 8, the UPLC analysis result of the refluxed products of nardosinone in slightly boiling water for 0, 30, 60, and 120 min indicated that nardosinone (5) was rather unstable in hot water and degraded entirely after 2 h of refluxing in boiling water. The main degradation products were 2-deoxokanshone M (1), desoxo-narchinol A (2), 2-deoxokanshone L (3), and isonardosinone (4). Hence, compounds 1-5 were selected as the marker compounds ( Figure 9c for 60, 48, and 12 h, respectively. Moreover, as shown in Figure 10, the influence of incubation time on the incubation products of nardosinone under three different environments (HT, SGF, and SIF) was studied. Notably, the degradation profile of nardosinone incubated in 50% aqueous methanol (80 • C) for 108 h was consistent with that of nardosinone refluxed in boiling water for 2 h, but markedly different from those in SGF and SIF for 108 h. Desoxo-narchinol A (2) and 2-deoxokanshone L (3) were the common degradation products in the HT, SGF, and SIF conditions. Isonardosinone (4) could not be detected in the incubation products of nardosinone in SGF. At the same time, 2-deoxokanshone M (1) could be found in the degradation products of nardosinone incubated in SIF for 12 h, where it then gradually disappeared with the incubation time.

Evaluation of the Vasodilatory and Anti-Neuroinflammatory Activities of Nardosinone and the Main Degradation Products
Previous studies have revealed that nardosinone possesses significant anti-neuroinflammatory [10] and vasodilatory (IC50 254.59 µM) effects [27]. In this study, these two activities of nardosinone and its primary degradation product (2-deoxokanshone M, 1) were evaluated based on our previously reported methods [26,27]. As shown in Figure 7, nardosinone could significantly inhibit the inflammatory factor NO production of the LPS-induced neuroinflammation model in BV-2 microglia cells, which is consistent with previous results reported (IC50: 37.82-74.21 µM) [10] and compared with minocycline (IC50 = 23.69 ± 2.01 µM) [28]. At the same time, 2-deoxokanshone M (1) failed to exhibit any inhibitory effect (Figure 7a,b). Other main degradation products, isonardosinone and desoxo-narchinol A, have been reported to exhibit significant anti-neuroinflammatory activities with IC50 values of 37.82-74.21 µM [10] and 3.48 ± 0.47 µM [29], respectively. As presented in Figure 7c, 2-deoxokanshone M displayed a vasorelaxation activity that was more potent than nardosinone. Specifically, the vasodilation rate of nardosinone was 74.57% at a concentration of 400 µM, while that of 2-deoxokanshone M already reached 80.39% at a lower concentration of 75 µM. Given these results, it appears that the degradation mentioned above is beneficial for nardosinone in the treatment of vasodilation-related diseases, exemplified by hypertension [27], and is unfavorable for nardosinone in the aid of neuroinflammation-associated neurodegenerative disorders such as Alzheimer's disease, parkinsonian syndromes, and amyotrophic lateral sclerosis [30].

Degradation of Nardosinone in Different Conditions
As shown in Figure 8, the UPLC analysis result of the refluxed products of nardosinone in slightly boiling water for 0, 30, 60, and 120 min indicated that nardosinone (5) was rather unstable in hot water and degraded entirely after 2 h of refluxing in boiling water. The main degradation products were 2-deoxokanshone M (1), desoxo-narchinol A (2), 2deoxokanshone L (3), and isonardosinone (4). Hence, compounds 1-5 were selected as the marker compounds (Figure 9c) for further UPLC profiling of the degradation of nardosinone under a high-temperature (in 50% aqueous methanol at 80 °C, HT) condition and in simulated gastric (SGF) and intestinal (SIF) fluids in vitro. In consideration of the UV absorption wavelengths of the five marker compounds (Figure 9a), a wavelength of 270 nm was chosen. Figure 9b displays the UPLC stacked chromatograms of the five standard solutions and different incubation products of nardosinone in HT, SGF, and SIF conditions for 60, 48, and 12 h, respectively. Moreover, as shown in Figure 10, the influence of incubation time on the incubation products of nardosinone under three different environments (HT, SGF, and SIF) was studied. Notably, the degradation profile of nardosinone incubated in 50% aqueous methanol (80 °C) for 108 h was consistent with that of nardosinone refluxed in boiling water for 2 h, but markedly different from those in SGF and SIF for 108 h. Desoxo-narchinol A (2) and 2-deoxokanshone L (3) were the common degradation products in the HT, SGF, and SIF conditions. Isonardosinone (4) could not be detected in the incubation products of nardosinone in SGF. At the same time, 2-deoxokanshone M     During the 108-h degradation process of nardosinone in HT, SGF, and SIF, the concentrations of nardosinone and the four main products changed along with the incubation time, as shown in Figure 11. Nardosinone degraded significantly faster in HT and SGF than in SIF (Figure 11a), and nardosinone degraded entirely after 108 h of incubation in HT and SGF instead of a 79.66% degradation in SIF. In the beginning of the HT condition (Figure 11b), nardosinone rapidly degraded at a rate of 3.74% per hour within 24 h, then slowly degraded at a rate of 0.27% per hour, finally disappearing within 60 h. 2-Deoxokanshone M continuously increased within 84 h and then tended to obviously decrease During the 108-h degradation process of nardosinone in HT, SGF, and SIF, the concentrations of nardosinone and the four main products changed along with the incubation time, as shown in Figure 11. Nardosinone degraded significantly faster in HT and SGF than in SIF (Figure 11a), and nardosinone degraded entirely after 108 h of incubation in HT and SGF instead of a 79.66% degradation in SIF. In the beginning of the HT condition (Figure 11b), nardosinone rapidly degraded at a rate of 3.74% per hour within 24 h, then slowly degraded at a rate of 0.27% per hour, finally disappearing within 60 h. 2-Deoxokanshone M continuously increased within 84 h and then tended to obviously decrease in the remaining time. 2-Deoxokanshone L and isonardosinone were constantly increased during the whole degradation process, while desoxo-narchinol A was increased at a very slow speed with extremely low concentrations. In SGF (Figure 11c), nardosinone degraded at a rate of 1.15% per hour within 84 h, then slowly at a rate of 0.14% per hour, finally reaching the full degradation within 108 h; Desoxo-narchinol A firstly increased and then decreased; 2-Deoxokanshone M continuously increased; 2-Deoxokanshone L increased rather slowly in this process, which led to the speculation that conversion into this compound may occur in the clue of sustained temperature. In SIF (Figure 11d), nardosinone rapidly degraded at a rate of 4.26% per hour within 12 h and then slowly at a rate of nearly 0.27% per hour; nardosinone would not degrade completely even after 108 h of incubation. Isonardosinone sharply increased within 5 h, and then its concentration remained largely unchanged in the remaining time; desoxo-narchinol A constantly increased in this process, while 2-deoxokanshone M also tended to firstly increase and later decrease until it was imperceptibly beyond the detection of UPLC-PDA analysis. 2-deoxokanshone L increased at a very slow similar rate, similarly to that of the SGF condition. As mentioned above, the production of 2-deoxokanshone M increased and then decreased in 50% MeOH under high temperature, indicating that this product may still be unstable and tended to be further degraded or transformed into other derivatives. Therefore, the preliminary stability test of this compound was undertaken as follows: an accurately weighed amount of 20.0 mg of 2-deoxokanshone M was dissolved in 40 mL of 50% aqueous methanol and SGF-A and then incubated at 80 °C and 37 °C, respectively, for 48 h. Some incubation products with retention times of 25.90 min and 26.32 min were gradually afforded under high temperature in 50% MeOH, while no products were found in the SGF-A incubation (Supplementary Figure S3). Moreover, when 2-deoxokanshone M was subjected to boiling water and refluxed for 2 h without methanol (Supplementary Figure S4a), there was no degradation of transformation undertaken instead. Compared with nardosinone (Figure 8), 2-deoxokanshone M showed significantly high stability in boiling water (Supplementary Figure S4b). Combining the findings mentioned above with As mentioned above, the production of 2-deoxokanshone M increased and then decreased in 50% MeOH under high temperature, indicating that this product may still be unstable and tended to be further degraded or transformed into other derivatives. Therefore, the preliminary stability test of this compound was undertaken as follows: an accurately weighed amount of 20.0 mg of 2-deoxokanshone M was dissolved in 40 mL of 50% aqueous methanol and SGF-A and then incubated at 80 • C and 37 • C, respectively, for 48 h. Some incubation products with retention times of 25.90 min and 26.32 min were gradually afforded under high temperature in 50% MeOH, while no products were found in the SGF-A incubation (Supplementary Figure S3). Moreover, when 2-deoxokanshone M was subjected to boiling water and refluxed for 2 h without methanol (Supplementary Figure S4a), there was no degradation of transformation undertaken instead. Compared with nardosinone (Figure 8), 2-deoxokanshone M showed significantly high stability in boiling water (Supplementary Figure S4b). Combining the findings mentioned above with the NMR data ( Supplementary Figures S5-S7), it could be speculated that the instability of 2-deoxokanshone M in 50% MeOH at high temperature may be attributed to the ketoenol tautomerization and methylation of the enol group due to the presence of MeOH; however, 2-deoxokanshone M seemed to be inert and insensitive to temperature only. Thus, 2-deoxokanshone M (1a) was proposed to be simultaneously transformed into three other products (1b, 2a, and 2b), as deduced by the quadruple δ C-8 peaks in the 13 C NMR spectrum ( Figure 12) of 2-deoxokanshone M in CD 3 OD. And, two compounds (2a, and 2b) furtherly afforded two methylation products (3a and 3b) when incubated in 50% MeOH at a high temperature of 80 • C for 48 h. The concentration ratio of compounds 1, 2, and 3 was approximately 7:4:1 based on the integral values of δ H-8 in the 1 H NMR spectrum of 2-deoxokanshone M in CD 3 OD before and after heating in hot aqueous methanol ( Figure 13).  In brief, nardosinone was more unstable under conditions of HT and SGF than in SIF. Isonardosinone was not detected in the incubation products of nardosinone in SGF, and 2-deoxokanshone L was more likely to be generated by continuous heating and seemed to be inert to pH variations. 2-Deoxokanshone M was relatively stable only in boiling water and was prone to causing a methylation of the enol group in hot aqueous methanol.
Furthermore, the degradation of nardosinone was investigated under the conditions of SGF-A (without pepsin) and SGF-B (without HCl), as illustrated in Supplementary  Figures S8 and S9. As a result, nardosinone degraded faster in an acidic environment than in a condition of solely pepsin, suggesting that nardosinone was more susceptible to the pH change.

UHPLC-DAD/Q-TOF MS Analysis and Proposed Degradation Pathway of Nardosinone
The degradation products of nardosinone in HT, SGF, SIF, SGF-A, and SGF-B were further analyzed using a UHPLC-DAD/Q-TOF MS analysis in the positive mode, with the total ion current (TIC) chromatograms being presented in Figure 14 and the detailed MS information being listed in Supplementary Table S1. Five compounds were identified based on the MS fragments, as presented in Supplementary Figures S10-S14, and the proposed fragmentation pathway was speculated accordingly, as presented in Supplementary Figures S15-S19. It is worth mentioning that nardosinonediol was characterized by UHPLC-DAD/Q-TOF MS through an MS comparison with that of the standard compound. Nardosinonediol was found to be only present in the incubation products of nardosinone in SIF.  All identified compounds belong to the nardosinane-type sesquiterpenoids; therefore, they share some common ionic fragments in the similar fragmentation pathways. Take nardosinone as an example ( Figure 15): A quasi-molecular ion at m/z 251.1646 [M+H] + has been detected, and it is predisposed to lose an H 2 O (m/z 18) due to the instability of the peroxy ring, resulting in a characteristic fragment ion at m/z 233.1569. The presence of the carbonyl group in the structure leads to a further loss of the CO group (m/z 28) to afford an ion at m/z 205.1588. More importantly, the inverse Diels-Alder (DA) reaction is readily available for the specialized structure of nardosinane-type sesquiterpenoid, and the ion at m/z 233.1569 could further produce a characteristic fragment ion at m/z 191.1429 in the clue of cleavage of C 3 H 6 (m/z 42). In addition, losses of    Based on the above-mentioned findings, a plausible degradation mechanism has been proposed, as shown in Figure 16. Initially, the five-membered peroxy ring in nardosinone (5) [8] makes it prone to a ring-opening reaction, which directly converts it to nardosinonediol [8]. Nardosinonediol remains generally stable in a neutral medium and is readily oxygenated to 7-oxonardosinone [8] or dehydrated to form kanshone A [8] in acidic and/or high-temperature conditions. This might be why nardosinonediol cannot be detected in the degradation products of nardosinone in HT and SGF except for in SIF. Kanshone A could undergo an epoxidation reaction to form isonardosinone (4) [8]. The epoxy three-membered ring in isonardosinone (4) is highly unstable in an aqueous acidic solution; it is readily exposed to a ring-opening reaction followed by a pinacol rearrangement and transforms into 7-oxonardosinone. This might explain why isonardosinone (4) cannot be found in the incubation products of nardosinone in SGF and SGF-A. Kanshone A was subject to a dehydration reaction to provide desoxo-nardosinanone H [8] or a cleavage of C 3 H 6 to generate desoxo-narchinol A (2) [8]. 7-Oxonardosinone was susceptible to a keto-enol interconversion to form 2-deoxokanshone L (3) with an unsaturated hydroxyl group, which was followed by a cleavage of C 3 H 6 O to afford 2-deoxokanshone M (1). It is noteworthy that all of the compounds mentioned above have been reported to be isolated from the same plant of N. jatamansi except for the novel compounds, namely 2-deoxokanshone M (1) and 2-deoxokanshone L (3). Based on the above-mentioned findings, a plausible degradation mechanism has been proposed, as shown in Figure 16. Initially, the five-membered peroxy ring in nardosinone (5) [8] makes it prone to a ring-opening reaction, which directly converts it to nardosinonediol [8]. Nardosinonediol remains generally stable in a neutral medium and is readily oxygenated to 7-oxonardosinone [8] or dehydrated to form kanshone A [8] in acidic and/or high-temperature conditions. This might be why nardosinonediol cannot be detected in the degradation products of nardosinone in HT and SGF except for in SIF. Kanshone A could undergo an epoxidation reaction to form isonardosinone (4) [8]. The epoxy three-membered ring in isonardosinone (4) is highly unstable in an aqueous acidic solution; it is readily exposed to a ring-opening reaction followed by a pinacol rearrangement and transforms into 7-oxonardosinone. This might explain why isonardosinone (4) cannot be found in the incubation products of nardosinone in SGF and SGF-A. Kanshone A was subject to a dehydration reaction to provide desoxo-nardosinanone H [8] or a cleavage of C3H6 to generate desoxo-narchinol A (2) [8]. 7-Oxonardosinone was susceptible to a keto-enol interconversion to form 2-deoxokanshone L (3) with an unsaturated hydroxyl group, which was followed by a cleavage of C3H6O to afford 2-deoxokanshone M (1). It is noteworthy that all of the compounds mentioned above have been reported to be isolated from the same plant of N. jatamansi except for the novel compounds, namely 2-deoxokanshone M (1) and 2-deoxokanshone L (3).

General Experimental Procedures, Reagents, and Materials
Chromatographic eluting fractions were monitored using a TLC analysis using silica gel 60 GF254 plates (5 mm × 10 mm, 10 mm × 20 mm, and 20 mm × 20 mm, 20 µm, Merck, Darmstadt, Germany). The spots were visualized by heating the plates after spraying them with 10% H 2 SO 4 in ethanol. The 1D/2D NMR spectra were recorded on a Bruker AV-III spectrometer (600 MHz, Bruker, Zurich, Switzerland). Optical rotations were measured using a Rudolph AUTOPOL V polarimeter (Rudolph Research Analytical, Hackettstown, NJ, USA). ECD data were collected using a JASCO-815 ECD spectrometer (JASCO Co., Ltd., Tokyo, Japan). The HR-ESIMS data were obtained on a UPLC-Q Exactive Orbitrap Mass system (Thermo Scientific, Waltham, MA, USA). UV spectra were scanned on an Agilent Cary 60 UV-Vis spectrophotometer (Agilent, Palo Alto, CA, USA), and IR spectra were obtained on a Varian 640 FT-IR spectrometer (Thermo Fisher Scientific, Waltham, MA, USA). The X-ray diffraction experiment was performed using a Rigaku Xtalab P200 diffractometer (Rigaku Co., Ltd., Tokyo, Japan). The melting point was supplied by an SGW ® X-4 microscopic melting point apparatus (Shanghai PRECISION Scientific Instrument Co., Ltd., Shanghai, China). Mobile phases were filtered through 0.45 µm microporous membranes and ultrasonically degassed before analysis.
For the profiling of the degradation products, monobasic potassium phosphate, porcine pepsin, and trypsin were purchased from Shanghai Yuanye Bio-Technology Co., Ltd., Shanghai, China. HCl and NaOH were obtained from Tianjin Damao Chemical Reagent Co., Ltd., Tianjin, China. Watsons Water was provided by Guangzhou Watson's Food & Beverage Co., Ltd., Guangzhou, China. Chromatographic-grade acetonitrile and formic acid were obtained from Thermo Fisher Scientific, Waltham, MA, USA. Other solvents that were used in this work were of analytical grade and were purchased from Concord Technology Co., Ltd., Tianjin, China. Nardosinone was prepared in our lab with a UPLC purity of above 98%, as detailed in Supplementary Figure S20

NMR Spectroscopy
The NMR spectra were recorded on a Bruker AV-III spectrometer equipped with 5 mm probes at 298 K. Chemical shifts were provided on the δ scale and referenced to TMS at 0.00 ppm for proton and carbon. The coupling constants (J) are in hertz. The pulse conditions are presented in Supplementary Table S2.
ECD Calculation of (1 and 3): As reported in our previous study [31], stochastic conformational searches were firstly performed using CONFLEX 8 software to afford the low-energy conformers. Geometry optimizations and the frequency pre-calculations of the low-energy conformers were finished at the B3LYP/6-311+g (2d, p) basis set level in MeOH. Using the optimized conformers of compound 1 and 3, 100 excitation states at the B3LYP/6-311+g (2d, p) level were calculated using an IEFPCM solvent model in MeOH with a half bandwidth of 0.45 eV. The calculation results of different conformers were Boltzmann averaged to simulate the ECD spectra after UV correction, which were finally extracted using SpecDis 1.70.1 software.

Evaluation of the Vasodilatory Activity
Using the method previously described [27], the intact endothelium aorta was used to measure the vasorelaxation activity of the isolated compounds.

Evaluation of the Anti-Neuroinflammatory Activity
As reported in a previous report [31], a CCK-8 assay was used to evaluate the effects of the isolated degradation products on the cell viabilities of BV-2 microglial cells, and a Griess method [32] was applied to evaluate their activities on the NO production in LPS-simulated BV-2 cells, with minocycline being selected as the positive drug.

Quantitative UPLC-PDA Analysis
A Waters Acquity UPLC ® H class plus system (Waters Corporation, Milford, MA, USA), equipped with a column heater, sample manager, quaternary solvent manager, and photo-diode array (PDA) detector, was employed to undertake the chromatographic separation using an Acquity UPLC BEH C 18 column (2.1 mm × 100 mm, 1.7 µm) at 35 • C. The mobile phase comprised acetonitrile (A) and 0.1% formic acid aqueous solution (B). The samples were detected at a wavelength of 270 nm with the flow rate set at 0.3 mL/min and an injection volume of 3 µL, and the gradient program was set as follows: 0-22 min, 18-26% (A); 22-30 min, 26-95% (A), respectively.
For quantitative analysis, linearity was evaluated by analyzing six different concentrations of the standard solutions (nardosinone, isonardosinone, desoxo-narchinol A, 2-deoxokanshone L, and 2-deoxokanshone M). Each linearity sample was injected in triplicate. The calibration curve was constructed as a linear regression analysis of the peak area (Y-axis) versus the corresponding concentration (X-axis). The limits of detection (LOD) and quantification (LOQ) of the tested compounds were calculated as concentrations where the signal-noise ratio (S/N) was about 3 and 10, respectively. The calibration plot was linear for 5 evaluation components, and a linear relationship with the r 2 values being greater than 0.9995 was observed in all cases. The linear relationship, LOD, and LOQ values are detailed in Table 3.

UHPLC-DAD/Q-TOF MS Analysis
The analysis was performed on an Agilent 1260 Infinity II UHPLC system equipped with an Agilent 6550 iFunnel Q-TOF MS detector (Agilent Technologies, Palo Alto, CA, USA). The chromatographic condition of UHPLC was employed as described in Section 3.6. For the MS analysis, nitrogen was used as drying gas at a flow rate of 12 L/min. The gas temperature and sheath gas temperature were fixed at 200 • C and 350 • C, respectively. The nebulizer was 40 psi. The capillary voltage and nozzle voltage were set as 3.0 kV and 1.5 kV, respectively. The spectra were recorded in the m/z 50-1200 range for a full scan.

Preparation of Simulated Gastrointestinal Fluids
Simulated gastric fluid (SGF) and simulated intestinal fluid (SIF) in vitro conditions were prepared according to the Chinese Pharmacopoeia (2020 Edition, Part IV) [33]. For SGF, 10 g of pepsin was dissolved in 16.4 mL of HCl and a sufficient volume of water to form up to 1000 mL with a pH of approximately at 1.3. For SIF, 6.8 g of monobasic potassium phosphate was dissolved in 500 mL of water with a pH adjusted to 6.8 using 0.1 mol/L NaOH solution. An amount of 10 g of trypsin was dissolved in an appropriate amount of water, and the two solutions were mixed; the final solution volume formed up to 1000 mL with water. In addition, simulated gastric fluid A (SGF-A, without pepsin) and B (SGF-B, without HCl) were also prepared to investigate the specific factors (acidic or not) affecting the degradation of nardosinone compared with that in SGF.

Incubations of Nardosinone at High-Temperature Condition and in Simulated Gastrointestinal Fluids
An accurately weighed amount of 20.0 mg of nardosinone was dissolved in 40 mL of 50% aqueous methanol, and the solution was then incubated at 80 • C for 108 h. Similarly, each accurately weighed 20.0 mg amount of nardosinone was dissolved in 40 mL of SGF, SIF, SGF-A, and SGF-B fluids, respectively, and incubated at 37 • C for 108 h. It is worth noting that due to the poor water solubility of nardosinone, it does not dissolve in simulated gastrointestinal fluids at room temperature; instead, it is in a suspended state. Therefore, an equal volume of methanol was added to dissolve and configure the sample before further tests or analyses. Moreover, during the incubations, the concentrations of nardosinone, isonardosinone, desoxo-narchinol A, 2-deoxokanshone L, and 2-deoxokanshone M were calculated based on the regression equation at each point of time (0, 2, 4, 6, 8, 10, 12, 24, 48, 60, 72, 84, 96, and 108 h) using the above-established UPLC method. The stability study of nardosinone was repeated three times to guarantee the reliability of the results. In addition, trivalent 1.0 mg of nardosinone was dissolved and incubated in 2 mL of 50% aqueous methanol (80 • C), SGF (37 • C), and SIF (37 • C), respectively, until it completely degraded to afford the incubation end products before being diluted five times with methanol for the UHPLC-DAD/Q-TOF MS analysis.

Statistical Analysis
The data were presented as the mean ± standard errors of the mean (SEM). The statistical differences were analyzed using GraphPad Prism 8.0 software using a one-way ANOVA test. Values of p < 0.05 were considered statistically significant.

Conclusions
Four main degradation products of nardosinone in refluxing boiling water were isolated and identified as 2-deoxokanshone M, desoxo-narchinol A, 2-deoxokanshone L, and isonardosinone. Among them, 2-deoxokanshone M represents a new C 12 norsesquiterpenoid featuring an α, β-unsaturated enol group in its structure, and it possesses significant vasodilatory activity without any anti-neuroinflammatory activity. Nardosinone is more susceptible to temperature and pH changes, and the degradation products slightly vary in high-temperature and simulated gastric and intestinal fluids conditions. Furthermore, the transformation pathway of 2-deoxokanshone M was reasonably deduced, and the MS fragmentation pattern of nardosinone and its plausible degradation pathway were proposed accordingly. Our study preliminarily sheds light on the stability of nardosinone. Future investigations are still needed for the analysis on the metabolites of nardosinone and the components that are absorbed in the blood. This study of the degradation products and patterns of nardosinone will facilitate the identification of the components of N. jatamansi extracts or fractions and the elucidation of dynamic mechanisms, and these findings may contribute to providing reasonable guidance for the preparation, production, and quality control of nardosinone, N. jatamansi, and the prepared patent Chinese medicines that contain N. jatamansi.   Figure S20: The separation process diagram of nardosinone; Table S1: Characterization of nardosinone and its degradation products in three different conditions using UHPLC-DAD/ESI-Q-TOF MS; Table S2: The detailed parameters of the NMR measurements.