Structurally diverse terpenoids and phenolic compounds from compound Ruteng, a Tibetan medicine

2.1. Experimental procedures

A detailed description of the general experimental procedures can be found in our previous report [9] and provided in the supplementary materials.

2.2. Plant materials and CRT preparation

The plant materials used in this study were sourced from two primary locations in China. Boswellia carteri Birdw. and Senna obtusifolia (L.) were obtained from Mayinglong Pharmaceutical Group Co., Ltd, Hubei province. Concurrently, Tinospora sinensis (Lour.), Terminalia chebula Retz., Phlomoides rotata (Benth. ex Hook.f.), and Abelmoschus manihot (L.) were collected from Xining City, Qinghai province. Furthermore, Tanacetum tatsienense (Bureau & Franch.) was procured from Lhasa City, Xizang Autonomous Region. These botanical specimens were identified and classified by Professor Xinqiao Liu of the School of Pharmaceutical Sciences, South-Central Minzu University. A voucher specimen for each plant has been deposited at the Herbarium of the School of Pharmaceutical Sciences, South-Central Minzu University, under the accession number from 20180808 to 20180814 (arranged alphabetically by Latin name). CRT was prepared following the methodology described in the literature [4].

2.3. Extraction and isolation

The ethanol-extracted product of CRT (3.4 kg) was subjected to a sequential reflux process using petroleum ether (PE), ethyl acetate (EtOAc), and methanol. For each solvent, extraction was performed three times, with the duration decreasing from 3 to 2 to 1 h across the successive cycles. The procedure yielded 1122 g of PE extract, alongside 381 and 571 g of EtOAc and MeOH extracts, respectively. A portion of the EtOAc extract (309 g) was subjected to silica gel column chromatography (SGCC), semi-preparative high-performance liquid chromatography (semi-prep HPLC), and recrystallization to obtain compounds. From the SGCC using a PE–EtOAc gradient, seven subfractions (Fr. 1-Fr. 7) were obtained, and compound 28 (200 mg) was isolated through recrystallization. Fr. 2 was separated via SGCC, followed by semi-prep HPLC to yield compounds 4 (4.8 mg), 17 (2.1 mg), 19 (25.7 mg), 20 (43.6 mg) and 29 (7.6 mg). Fr.3 was subjected to SGCC and eluted using PE–EtOAc mixtures, and was further purified via semi-prep HPLC to afford compounds 8 (1.8 mg), 11 (5.2 mg), 12 (1.6 mg), 18 (5.4 mg), 21 (7.3 mg), 22 (1.3 mg), 23 (9.3 mg), 26 (5.8 mg), and 27 (0.6 mg). Crystals of compounds 9 (797.8 mg) and 13 (3 g) were collected from Fr. 4. The residual filtrate from Fr. 4 was purified via medium pressure liquid chromatography, SGCC (eluted by PE–EtOAc and CH₂Cl₂–EtOAc), and semi-prep HPLC to yield compounds 1 (20.4 mg), 2 (9.7 mg), 3 (20.9 mg), 5 (8.8 mg), 6 (5.7 mg), 7 (3.6 mg), 10 (19.2 mg), 14 (22.9 mg), 15 (8.2 mg), 16 (16.0 mg), 24 (16.6 mg) and 25 (3.8 mg). The detailed isolation procedure is presented in the supplementary materials (Tables S1-S22).

2.5. NMR and electronic circular dichroism (ECD) calculations

NMR and ECD calculations were performed using the gauge-independent atomic orbital (GIAO) and time-dependent density functional theory (TD-DFT) methods, respectively, which are provided in the supplementary materials. To accurately determine the structure of complex natural products, a robust method that integrates GIAO-NMR calculations, DP4+ analysis, and the MAE_{ΔΔ δ} technique has been developed. These methods have successfully resolved several problems in the identification of natural product structures, which could not be solved using conventional methods, particularly, the accurate and efficient discrimination of a set of stereoisomers [10,11].

3.1. Structural identification of isolated compounds

In this section, all 29 compounds are classified in the order of 10 new compounds (1-8, 17 and 18) and 19 known compounds (9-16, 19-29). The structural identification of the new compounds were discussed first. Compounds 9-16, 19-29 were known standards and thus described later in the known-compounds list.

Compounds 1 and 2 were identified as white amorphous powders. Their molecular formulas of C₃₀H₄₈O₄ were confirmed by the positive ion peak [M + Na]⁺ at m/z 495.34431 and m/z 495.34433, respectively, in the high-resolution electrospray ionization mass spectrometry (HR–ESI–MS) spectra (calcd. for C₃₀H₄₈O₄Na⁺, 495.34448). The analysis revealed that both compounds under investigation were isomers, and each exhibited seven indices of hydrogen deficiency (IHDs) corresponding to four rings and two double bonds and one carboxyl group. The data obtained from the ¹H NMR spectroscopy for both compounds (Table S1) indicated distinct signals corresponding to seven single-methyl groups. For compound 1, the observed chemical shifts included δ_H 1.51 (6H, s, CH₃-26 and CH₃-27), 1.25 (3H, s), 1.04 (3H, s), 1.01 (3H, s), 0.98 (3H, s), and 0.93 (3H, s). Similarly, compound 2 exhibited signals δ_H 1.51 (6H, s, CH₃-26 and CH₃-27), 1.27 (3H, s), 1.12 (3H, s), 1.07 (3H, s), 0.99 (3H, s), and 0.98 (3H, s). In addition to the methyl group signals, both compounds contained one trans double bond. For compound 1, the relevant signals were found at δ_H 6.19 (1H, m), 6.13 (1H, d, J = 15.6 Hz), whereas compound 2 showed similar signals at δ_H 6.22 (1H, m), 6.15 (1H, d, J = 15.0 Hz). Moreover, compounds 1 and 2 featured an oxymethine at δ_H 3.66 (1H, br s) and δ_H 3.47 (1H, dd, J = 11.4, 4.8 Hz), respectively. The ¹³C NMR and distortionless enhancement by polarization transfer (DEPT) data for compounds 1 and 2 indicated 30 carbon signals assigned to 7 methyls, 9 methylenes, 6 methines (including one oxymethine and two sp² methines), and 8 nonprotonated carbons (comprising one carboxyl, two sp² quaternary carbons, and one oxygenated tertiary carbon; Table S3). In addition to two double bonds and one carboxyl group, it is necessary for compounds 1 and 2 to have four rings to meet the index of hydrogen deficiency (IHD) requirements. Based on the aforementioned data, compounds 1 and 2 can be classified as tirucallane-type triterpenoids [12,13]. A thorough examination of the NMR data for compound 1, comparing it with 3α, 24, 25-trihydroxytirucall-8-en-21-oic acid, revealed that compound 1 likely had the core structure typical of tirucallane-type triterpenoids, which is defined by the presence of a 3α-hydroxy group, a Δ⁸⁽⁹⁾ double bond, and a carboxyl at C-20 [14]. Heteronuclear multiple bond correlations (HMBCs, Figure 2) of CH₃-28 and CH₃-29/C-3 (δ_C75.4); C-4 (δ_C38.7); and C-5 (δ_C45.6), CH₃-19/C-9 (δ_C135.7), and CH₃-30/C-8 (δ_C133.6) further corroborated this assumption. The main distinction between compound 1 and 3α, 24, 25-trihydroxytirucall-8-en-21-oic acid was the dehydration of the 24-hydroxy group in the latter, which led to the formation of a double bond at Δ²³⁽²⁴⁾, resulting in compound 1. The HMBC correlations of CH₃-26 and CH₃-27/C-23 (δ_C124.4), C-24 (δ_C142.4), and C-25 (δ_C70.2) confirmed this assignment. The double bond at Δ²³⁽²⁴⁾ was assigned an E-configuration based on the large coupling constant observed for H-23/H-24 (J = 15.6 Hz). This assignment is consistent with the typical coupling constants reported for such olefins, where E-isomers exhibit J values around 15–16 Hz, while Z-isomers are typically near 7–8 Hz [15,16]. Comparison of the 1D NMR spectra of compound 2 and compound 1 indicated that they were two C-3 epimers. Owing to the γ-gauche effect, the chemical shift (δ_C) of C-1, C-3, and C-5 in compound 2 were found to be downfield +5.4, +3.1, and +6.3 ppm, respectively, compared to compound 1, suggesting that the 3-OH group is axial in compound 1 and equatorial in compound 2. The configurations of compounds 1 and 2 were validated using the rotating frame overhauser effect spectroscopy (ROESY, Figure 3) method. Observed correlations included CH₃-28 with H-5, CH₃-30 with H-17, CH₃-18 with H-20, and CH₃-19 with CH₃-29, suggesting that the relative configurations of these compounds were consistent with the features of the tirucallane-type skeleton. Taken together, the chemical structures were determined to be 3α,25-dihydroxy-tirucall-8,23E-dien-21-oic acid (compound 1) and 3β,25-dihydroxy-tirucall-8,23E-dien-21-oic acid (compound 2).

Close

Figure 2.

HMBC correlations for compounds 1, 3, 5, 6, 8, 17 and 18.

Export to PPT

Close

Figure 3.

Key ROSEY correlations for compounds 1-8 and 18.

Export to PPT

Compound 3, obtained as a white amorphous powder, was assigned molecular formula C₃₀H₄₆O₄ based on HR–ESI–MS spectrum, which showed a positive ion peak [M + Na]⁺ at m/z 493.32886 (calcd. for C₃₀H₄₆O₄Na⁺, 493.32883), indicating a total of eight IHDs (four rings, two double bonds, one carbonyl group and one carboxyl group). The NMR data of compound 3 were compared with those of 1 and 2, indicating that they share an identical carbon framework, with the primary distinction being the functional group present at the C-3 position. Compound 3 contains a ketone group, whereas compounds 1 and 2 have a hydroxyl group. The presence of the ketone group was unequivocally demonstrated by its characteristic carbon signal at δ_C 217.1 ppm, an assignment confirmed by HMBC correlations involving CH₃-28 and CH₃-29 with C-3 (δ_C 217.1). Examination of the NMR data and ROESY spectrum revealed that the relative configuration of compound 3 matched that of compounds 1 and 2. Thus, the chemical structure of compound 3 was conclusively identified as 3-oxo-25-hydroxy- tirucall-8,23E-dien-21-oic acid.

Compound 4, isolated as a colorless oil, was determined to have the molecular formula C₃₀H₄₆O₂ as evidenced by HR–ESI–MS (positive ion detection), with [M + Na]⁺ detected at m/z 461.33893 (calcd. for C₃₀H₄₆O₂Na⁺, 461.33900), suggesting the presence of eight IHDs (four rings, two double bonds, one carbonyl group and one aldehyde group). Analysis of the 1D NMR spectra established that compound 4 retains the tetracyclic nucleus skeleton of compound 3. However, a notable difference was observed in the C-17 side chain. Furthermore, the C-17 side chain of compound 4 was found to be consistent with that of boscartene J, comprising an aldehyde group at C-20 and a Δ²⁴⁽²⁵⁾ double bond [17]. The relative configuration of the tirucallane-type skeleton of compound 4 was identical to that of boscartene J, as indicated by the ROESY spectrum. The structure of compound 4 was confirmed as 3-oxo-tirucalla-8,24-dien-21-al.

Compounds 5 and 6 exhibited a consistent planar structure with the same molecular formula, C₃₀H₄₆O₅, evidenced by the positive ion peak [M + Na]⁺ at m/z 509.32404 and m/z 509.32401 in the HR–ESI–MS spectra (calcd. for C₃₀H₄₆O₅Na⁺, 509.32375). The 1D NMR data for compounds 5 and 6 (Tables S2 and S3) closely resembled those of 3-keto-tirucall-8,24-dien-21-oic acid [18], implying the same carbon skeleton featuring a ketone at the C-3 and a Δ⁸⁽⁹⁾ double bond within the tetracyclic nucleus. This inference was validated by HMBC correlations involving CH₃-28 and CH₃-29 with C-3, CH₃-19 with C-9, and CH₃-30 with C-8. The only difference was the presence of a 1-peroxyhydroxy-2-methyl-2-propenyl at the C-23 position in compounds 5 and 6, with a 2-methyl-2-propenyl found at the same position in 3-keto-tirucall-8,24-dien-21-oic acid. The presence of this distinct side chain was confirmed through HMBC correlations involving CH₃-27 with C-25, C-26, and C-24. Further analysis of the chemical shifts observed from C-23 to C-27 in compounds 5 and 6, when compared with the corresponding values found in meliasenines U and V, provided additional validation for the peroxyhydroxy group at the C-24 position [19]. Previous research has shown that such a peroxyhydroxy group can be formed by UV-induced oxidation of the double bond in the side chain [20]. Accordingly, compounds 5 and 6 can be classified as having planar configurations, specifically identified as 3-oxo-24-peroxyhydroxy tirucall-8,25-dien-21-oic acid. Moreover, the ¹³C NMR spectral data for compound 5 indicated that its characteristics were nearly identical to those of compound 6, with only a minor ¹³C chemical shift variation observed around C-24 on the C-17 side chain, suggesting that they are C-24 epimers. Except for the relative configuration at C-24, the configurations of the remaining chiral centers in compounds 5 and 6 were specified to match the configuration of compound 3 based on the ROESY spectra. To define the configurations of C-24 in compounds 5 and 6, the MAE_{ΔΔ δ} approach was used to establish their specific configurations [11,21,22]. In ranking analysis, the small MAE_{ΔΔ δ} values of ranking indicated that compound A best matched compound 6 and that compound B best matched compound 5 (Figure 4, Table S22). The chemical structure of compound 5 was determined as 3-oxo-24R-peroxyhydroxy tirucall-8,25-dien-21-oic acid, whereas compound 6 was identified as 3-oxo-24S-peroxyhydroxy tirucall-8,25-dien-21-oic acid.

Close

Figure 4.

MAE_{ΔΔ δ} parameters of compounds 5 and 6 based on calculated and experimental NMR data.

Export to PPT

Compound 7 was characterized as a colorless oil with the molecular formula C₃₂H₅₂O₆. This identification was confirmed via mass spectrometry, where a positive ion peak corresponding to [M + Na]⁺ was observed at m/z 555.36548, closely aligning with the theoretical value for C₃₂H₅₂O₆Na⁺ (555.36561). Moreover, the ¹³C NMR spectroscopic data for compound 7 closely matched those of the previously studied 3α,24,25-trihydroxytirucall-8-en-21-oic [14], except for the presence of an additional acetyl group. In addition, the C-3 chemical shift (δ_C) was observed to move downfield by +2.7 ppm compared to compound 1. These observations collectively enabled the identification of the planar configuration of compound 7 as a 3-O-acetylated derivative of 3α,24,25-trihydroxytirucall-8-en-21-oic. The small coupling constants observed between H-2 and H-3, and the ROESY correlation involving CH₃-29 (δ_H 0.84) with H-3 (δ_H 4.87) implied spatial proximity, such configuration was consistent with a known tirucallane-type triterpenoid containing α-orientated acetoxy group at C-3 (Boscartene A) [17]. The orientation of 24-OH remained uncertain, while the configurations of the other stereocenters were confirmed to be consistent with those of 3α,24,25-trihydroxytirucall-8-en-21-oic by the ROESY spectrum. Based on information obtained from previous studies [14,19,23], the arrangement of C-24 within the 24,25-vicinal diol present in the C-17 side chain can be inferred by analyzing the ¹H NMR results, which indicate that the chemical shift variation of H-24 (Δδ_{R- S} = δ_R - δ_S) is ∼0.05 ppm. An analysis of the H-24 chemical shift for compound 7 in relation to that of 24R,25-dihydroxy-3-oxo-tirucalla-8-en-21-oic acid revealed that both possessed the same configuration at C-24 [13]. Moreover, the calculated ¹³C chemical shifts for 24R were consistent with the experimental data, demonstrating high correlation coefficients (R²) (close to 0.998), a corrected mean absolute deviation (CMAD) of ∼1.56 ppm, a root-mean-square deviation (RMSD) of ∼2.0 ppm, and a corrected largest absolute deviation (CLAD) of 4.86 ppm when compared with 24S (Table S20). Accordingly, the chemical structure of compound 7 was characterized as 24R,25-dihydroxy-3α-acetoxy-tirucalla-8-en-21-oic acid.

Compound 8 was isolated as a white amorphous powder, with the molecular formula determined to be C₃₂H₄₈O₄ based on the HR–ESI–MS at m/z 519.34375 [M + Na]⁺ (calcd. for C₃₂H₄₈O₄Na⁺, 519.34448), indicating 9 IHDs (four rings, two double bonds, one γ-lactone ring, and two ester carbonyl groups). The ¹H and ¹³C NMR data confirmed that compound 8 bears a strong resemblance to compound 7 in its 3α-acetoxy-Δ⁸⁽⁹⁾ -tetracyclic skeleton. A major structural divergence, however, was located in the C-17 side chain. The carbon signals associated with the C-17 side chain include two methyls (δ_C 18.7 and 26.0), one oxymethine (δ_C 75.4), one methine (δ_C 41.6), one methylene (δ_C 34.8), one tri-substituted double bond (δ_C 123.3 and 139.9), and one carboxyl (δ_C 179.2). The existence of nine IHDs indicates that a ring persists on the C-17 side chain. An analysis of the ¹H–¹H COSY spectrum for H₂-20 (δ_H 2.74), H-22 (δ_H 2.34, 1.88), H-23 (δ_H 5.04), and H-24 (δ_H 5.22), combined with the chemical shifts of C-23 and C-21 at δ_C 75.4 and δ_C 179.2, respectively, revealed the creation of a γ-lactone ring between C-21 and C-23 [24]. In addition, ROESY correlations involving CH₃-18 with H-20, H-20 with H-23, and CH₃-30 with H-17 further elucidated the configuration of the γ-lactone as 20S, 23R. Consequently, the proposed structure for compound 8 is 3α-acetoxy-tirucalla-8,24-dien-21,23-olide.

Compound 17, recognized as a colorless oil, was assigned the molecular formula C₂₀H₃₀O, which was confirmed by the positive ion peak at m/z 287.23686 [M + H]⁺ (calcd. for C₂₀H₃₁O, 287.23694). Analysis of the spectroscopic data and comparison with previous report revealed that compound 17 is likely a diterpenoid belonging to the verticillane series [25]. The 1D NMR data (Table S4) for compound 17 were largely identical to those of (1R,7E,11Z)-verticilla-4(20),7,11-triene [25], except that the methylene at C-12 is replaced by a ketone. Evidence for this substitution came from the HMBC correlation CH₃-20 (δ_H 1.89) and H-14 (δ_H 2.37) with C-12 (δ_C 198.4). Furthermore, ROESY experiments established the E-configuration of Δ⁶⁽⁷⁾ via correlation between H-6 (δ_H 5.19) with H-8 (δ_H 2.60) and the Z-configuration of Δ¹⁰⁽¹¹⁾ via the correlation between H_a-9 (δ_H 2.51) and CH₃-20 (δ_H 1.89). The computed and experimental ECD spectra (Figure 5) were well matched, confirming the absolute configuration of compound 17 as (14R,6E,10Z)-12-oxo-verticilla-3(18),6,10-triene.

Close

Figure 5.

Experimental ECD and calculated ECD spectra of compound 17.

Export to PPT

Compound 18 was obtained as a colorless oil. Its molecular formula was determined to be C₂₀H₃₄O based on the positive-ion HR-ESI-MS peak at m/z 273.25757 [M-H₂O+H]⁺ (calcd for C₂₀H₃₃⁺, 273.25768), a fragmentation pattern consistent with tertiary alcohols [26]. This formula corresponds to four IHDs (three rings and one double bond). The ¹H NMR spectrum exhibited an olefinic proton at δ_H 5.10 (1H, t, J = 7.2 Hz), as well as five methyl groups at δ_H 1.68 (3H, s), 1.61 (3H, s), 1.16 (3H, s), 1.01 (3H, s), and 0.95 (3H, d, J = 6.6 Hz). In addition, two cyclopropyl protons were observed at δ_H 0.64 (1H, m) and 0.15 (1H, t, J = 9.6 Hz). The ¹³C NMR and DEPT spectra for compound 18 showed 20 distinct carbon resonances, which comprised five methyls, six methylenes, six methines (including one sp² carbon), and three nonprotonated carbons (which encompassed one sp² quaternary carbon and one oxygenated tertiary carbon). Aside from the detected double bond, the remaining IHDs indicated a tricyclic framework for compound 18. A bicyclo[5.3.0]decane structure containing two methyls at C-3 and C-7, along with a hydroxyl at C-7, was confirmed through the analysis of ¹H–¹H COSY relationships involving H-6/H-5/H-4/H-3/H-2/H-1/H-10/H₂-9/H₂-8, CH₃-18/H-3, and H-6/H-2, along with the HMBC correlations of CH₃-18 with C-2, C-3, and C-4, and CH₃-19 with C-6, C-7, and C-8. Examination of both 1D and 2D NMR data showed that the remaining carbon resonances included a methyl (δ_C 13.6), a quaternary carbon (δ_C 22.6), and a 4-methyl-3-pentenyl group (δ_C 43.7, 25.5, 125.2, 131.1, 25.9, 17.8). Moreover, the HMBC spectrum revealed the correlations of CH₃-20 with C-1, C-10, and C-11, establishing a bicyclo[5.1.0]octane framework bearing a methyl and a 4-methyl-3-pentenyl group at C-11, respectively. The chemical shifts of C-1, C-10, and C-11 observed in the upfield region further supported the presence of the cyclopropane moiety. Therefore, the planar structure of compound 18 was found to be consistent with that of cneorubin U [27]. Cneorubin U is classified under the aromandendrane framework, characterized by the presence of the trans-bicyclo[5.3.0]decane and cis-bicyclo[5.1.0]octane structures, with ¹H chemical shifts for the cyclopropyl protons observed at δ_H 0.40-0.75 ppm. In contrast, the ¹H NMR chemical shifts for the cyclopropyl protons in compound 18 were recorded at δ_H 0.64 (1H, m) and 0.15 (1H, t, J = 9.6 Hz). This indicates that compound 18 contains an alloaromadendrane framework featuring cis-arrangements between H-2 and H-6 and H-1 and H-10 [28]. The ROESY correlation observed between CH₃-20 (δ_H 1.01) and H-2 (δ_H 1.86), combined with the lack of correlation of CH₃-20 with H-1 and H-10, indicated a trans arrangement between H-2 and H-10. Consequently, H-2, H-6, and CH₃-20 were identified as β-oriented, whereas H-1 and H-10 were categorized as α-oriented. The orientations of CH₃-18 as α and CH₃-19 as β were confirmed by ROESY correlations with CH₃-18 (δ_H 0.95) with H-1 (δ_H 0.15) and CH₃-19 (δ_H 1.16) with H-6 (δ_H 1.81). This assignment was corroborated by ¹³C–NMR calculations (Figure 6), which demonstrated small values for CMAD, CLAD, and RMSD with 1.62, 3.65, and 1.99, respectively, and an R² value of ∼ 0.9966. Therefore, compound 18 was deduced to be the C-6 epimer of cneorubin U, designated as 7α-hydroxy-12-(3-methyl-2-butenyl) alloaromadendrane.

Close

Figure 6 (a,b).

Linear regression fitting of computed ¹³ C and ¹ H-NMR chemical shifts of compound 18 with the experimental values.

Export to PPT

Compounds 9-16, 20-29 were identified as known structures. This identification was achieved through direct comparison of their NMR data with those reported in the references. They were identified as 3-keto-tirucall-8,24-dien-21-oic acid (9) [13], 3α-hydroxy-tirucalla-8,24-dien-21-oic acid (10) [29], isoflindissone lactone (11) [17], Δ⁹⁽¹¹⁾-nimolinone (12) [30], 3α-O-acetyl-11-keto-β-boswellic acid (13) [31], β-boswellic acid (14) [32], 9, 11-dehydro-β-boswellic acid (15) [31], α-boswellic acid (16) [26], nephthenol (19) [33], incensole oxide acetate (20) [34], incensole (21) [34], populeuphrine A (22) [35], isoincensole acetate (23) [34], aurantio-obtusin (24) [36], obtusin (25) [36], chrysophanol (26) [37], physcion (27) [38], gallic acid (28) [39], and bakutrol (29) [40]. In addition, a detailed data comparison between the newly identified compounds (1-8, 17 and 18) and their structurally similar analogues is also provided in the supplementary data (Tables S5-S9).

3.2. Source analysis of identified compounds from CRT

CRT is derived from the Shiwei Ruxiang Capsule, as documented in the classic Tibetan medical book “Four Volumes of Medical Treatment” [41]. This formulation is characterized by its concise and precise, comprising seven Tibetan medicines, with frankincenses, specifically resins from Boswellia trees, as the key herb. Prior investigations in phytochemistry focusing on frankincense have revealed that ∼97% of the primary constituents are terpenoids. These include various subclasses, including diterpenoids (e.g., cembranes, verticillanes, abietanes, and prenylaromadendranes) and triterpenoids (e.g. lupanes, dammaranes, ursanes, oleananes, and tirucallanes) [42]. Drawing on these research results, we propose that the terpenoids discovered in this study largely derive from frankincense. Among them, compounds 1–12 represent tirucallane-type triterpenoid, compounds 13–15 are classified as ursane-type triterpenoid, compound 16 pertains to oleanane-type triterpenoid, compound 17 is identified as a verticillane-type diterpenoid, compound 18 corresponds to prenylalloaromadendrane-type diterpenoid, and compounds 19–23 are cembrane-type diterpenoid. Notably, the prenylalloaromadendrane-type diterpenoid has been isolated from frankincense for the first time.

To date, researchers have isolated 31 triterpenoids of the tirucallane type from this resin. Nevertheless, finding a peroxyhydroxyl on the C-17 side chain of these triterpenoids is uncommon in natural products. To our knowledge, this study is the first to report natural products of this structural type from frankincense [42]. Except for compounds 12 and 22, the remaining known terpenoids have all been reported in frankincense [42-44]. The biosynthetic pathway of all tirucallane-type triterpenoids is presented in Scheme S1. Farnesyl pyrophosphate undergoes a head-to-tail combination, catalyzed by squalene synthase, to form squalene, which is subsequently converted into 2,3-epoxysqualene by squalene epoxidase. Under the catalysis of oxidosqualene cyclase, the key intermediate 5α-triucalla-8,24-dien-3α-ol, which was also isolated from frankincense, is formed. This compound undergoes a series of reactions, including oxidation, reduction, dehydration, hydration, and acylation, which link the newly isolated tirucallane-type triterpenoids with the known compounds, further indicating that these new tirucallane-type triterpenoids are derived from frankincense. The isolated diterpenoids are derived from (E, E, E)-geranylgeranyl pyrophosphate (GGPP) and undergo three distinct cyclization pathways, yielding three structurally different diterpene types (Scheme S2). Initially, GGPP loses its pyrophosphate group, becoming ionized, followed by cyclization at the C(1)/C(14) and C(10)/C(15) positions to form key intermediate I. Intermediate I then undergoes proton elimination, double-bond isomerization, and oxidation to yield compound 17 [45,46]. Macrocyclic diterpenes are formed via GGPP cyclization at the C(1)/C(14) position, generating carbocation intermediate II, which subsequently undergoes hydration to produce compound 19. Intermediate II undergoes dihydroxylation at the Δ¹⁰⁽¹¹⁾ double bond, followed by ether bond formation between C-11 and C-14 to yield compound 21. This compound undergoes acetylation, epoxidation, and epoxide ring-opening reactions to yield compounds 20, 22, and 23, respectively [47]. GGPP undergoes C(1)/C(10) cyclization to form carbocation III, which then undergoes proton elimination and double-bond isomerization to form intermediate IV. Subsequently, under acid catalysis, consecutive cyclization occurs between C(1)/C(11) and C(2)/C(6), ultimately resulting in the formation of compound 18 via a hydration reaction. Compounds 24–27 are categorized as anthraquinones and are characteristic molecular markers derived from S. obtusifolia [48]. Compound 28 typically comes from T. chebula, which is known for its high content of phenolic compounds [49], and has also been identified in T. sinensis as a quality marker [50]. Compound 29 is initially isolated from Psoralea glandulosa [40], though its source remained undetermined.

Among all isolated compounds, KTDA (9) and AKBA (13), being the most prevalent ones [4] (with 13.99 mg/g and 34.04 mg/g for KTDA and AKBA, respectively), exhibited the most effective anti-inflammatory properties. KTDA reduced the production of nitric oxide by inhibiting the phosphorylation of c-Jun N-terminal kinase, p38, and protein kinase B (Akt) proteins to suppress the mitogen-activated protein kinase (MAPK) and Akt pathways, which resulted in the downregulation of the mRNA expression of interleukin (IL)-1β and IL-6 at the molecular-level regulation of inflammatory responses. AKBA also suppressed the activities of enzymes, such as 5-lipoxygenase, human leukocyte elastase, IκB kinase, thereby reducing the expression levels of proinflammatory cytokine tumor necrosis factor-α (TNF-α) and IL-6 and increasing the expression level of anti-inflammatory cytokine IL-10 [51]. Thus, the two compounds can be two quality marker substances for CRT, which can be used to control the quality of TMCs, and to ensure the effectiveness, safety, and stability of TMCs.

TMCs are distinguished by their complex compositions, which undergo extensive transformations during preparation and within the human body. Since B. carteri is one of the most important components of Tibetan medicine CRT, and its terpenoid constituents are generally reported in the EtOAc fraction [42,51], focusing on this extract for compound isolation was both a direct and effective strategy. However, we isolated only 29 compounds from CRT, which is an extremely limited part of the complex chemical composition. Our emphasis on medium-polarity constituents resulted in the omission of key components, such as flavonoid glycosides from T. tatsienense, the other main medicinal component of CRT. These components may exist in the high-polarity fraction, like the methanol extract. The newly identified compounds may come from the original components inherent in Tibetan medicine, or emerge from novel substances formed during the preparation process. This is especially relevant for compounds 5 and 6, which contain a peroxide group and can be produced under UV light and oxygen. Due to insufficient compound quantities, a more comprehensive study of the transformations within CRT, re-evaluation of bioactivity, structure–chemistry relationships and trace of the botanical origin or chemotaxonomic significance were not pursued in this study but warrant further exploration in the future.