Dracaenogenins C and D, Two New 12(13→14)-Abeo-Spirostanols from the Red Resin of Dracaena cochinchinensis

Abstract

The red resin of Dracaena cochinchinensis (Lour.) S.C. Chen, known as Chinese dragon’s blood, is formed through metabolic reprogramming following trunk injury, during which the original steroidal saponins are depleted and transformed. To investigate the steroidal degradation intermediates in this process, a systematic phytochemical study was conducted on the resin from Yunnan Province, leading to the isolation of 14 steroidal constituents (2 new and 12 known). The two new compounds, dracaenogenins C (1) and D (2), were identified as rare 12(13→14)-abeo-spirostanol aglycones, with 2 representing an unusual C-14α-hydroxylated derivative. Their structures, including absolute configurations, were unambiguously determined by comprehensive spectroscopic analysis (1D and 2D NMR, HRESIMS) and single-crystal X-ray diffraction. Biogenetic analysis suggests that these unusual aglycones arise from the acid-catalyzed Wagner–Meerwein rearrangement of diosgenin-type saponins via C-18 angular methyl migration (C-10→C-13) and C-ring contraction, serving as rare catabolic intermediates trapped during the metabolic shift from saponin accumulation to polyphenol biosynthesis. Furthermore, cytotoxicity evaluation against HepG2 cells revealed that while the parent glycosylated saponins (e.g., dioscin and gracillin) exhibited significant toxicity, the rearranged aglycones (1, 2, and 3) and other degradation products were devoid of cytotoxicity, supporting a detoxification mechanism during resin formation.

1. Introduction

Dragon’s blood is a renowned traditional medicine of the Arabian Peninsula indigenous to Socotra Island in the Red Sea region. It is derived from the red resin secreted by the trunk of Dracaena cinnabari Balf.f. According to local legend, this crimson resin is regarded as the congealed blood shed by dragons in battle, hence the name “dragon’s blood” [1]. Approximately since the 6th century CE, dragon’s blood was introduced to China via the Silk Road as a valuable spice and medicine. Over the subsequent five centuries, it was gradually incorporated into the Traditional Chinese Medicine (TCM) system. In TCM clinical practice, dragon’s blood is renowned for its efficacy in promoting blood circulation, alleviating pain, stopping bleeding, and healing wounds. It is widely applied in the treatment of traumatic injuries, hemorrhage (both internal and external), and pain caused by blood stasis [2,3,4,5].

In the 1970s, to address the shortage of imported exotic medicines and identify domestic substitutes, the renowned botanist Professor Cai Xitao discovered Dracaena cochinchinensis (Lour.) S.C. Chen in the tropical rainforests of southern Yunnan, China. The resin secreted from the trunk of this species was identified as the closest substitute to authentic D. cinnabari dragon’s blood. Consequently, it was designated as “Chinese Dragon’s Blood” and was officially included in the Yunnan Provincial Drug Standard in 1974 [6]. This discovery not only ended China’s complete reliance on imported dragon’s blood but also sparked a surge of interest in the translational research of Chinese dragon’s blood among scholars worldwide.

Phytochemical investigations have revealed distinct chemical profiles between the fresh stems and the red resin of Dracaena cochinchinensis. The fresh stems are rich in steroidal saponins, including major C-25 epimeric pairs (dioscin, gracillin, protodioscin, protogracillin, etc.), pregnane glycosides (dracaenosides A–D) [7], and fourteen C27-steroidal saponins (dracaenosides E–R) predominantly possessing 1,3-dihydroxy-5-ene spirostanol skeletons with unique C-14 hydroxylation [8]. Notably, these saponins exist predominantly as C-25 epimeric pairs (R/S configurations), which are characteristic chemotaxonomic markers of the genus Dracaena.

Upon physical injury or external stimulation, the plant activates complex defense mechanisms, leading to dramatic metabolic reprogramming. Transcriptomic studies on related species, such as D. cambodiana, suggest that this process involves significant down-regulation of cytosolic MVA pathway genes and steroidal saponin biosynthesis, resulting in the depletion of steroidal saponins [9]. Concurrently, the biosynthetic flux is redirected toward the phenylpropanoid pathway. The expression of genes encoding key enzymes for flavonoid biosynthesis—such as chalcone synthase (CHS), chalcone isomerase (CHI), flavanone 3-hydroxylase (F3H), and dihydroflavonol 4-reductase (DFR)—is significantly up-regulated [10,11]. Consequently, the original steroidal saponins are largely depleted, and the red resin becomes enriched in polyphenolic compounds, including flavonoids, chalcones, and their oligomers, which account for over 50% of the resin by weight [3].

Previous phytochemical studies on the red resin led to the isolation of a novel 12(13→14)abeo-spirostanol derivative, dracaenogenin A, featuring a unique 12(13→14)abeospirosta-5,13(18)-diene-1β,3β,15α-triol skeleton, along with dracaenogenin B, identified as (25R)-spirost-5-ene-1β,3β,14α,15α-tetrol [12]. These compounds represented the first examples of such skeletal rearrangements in steroidal sapogenins, suggesting that steroidal saponins undergo acid-catalyzed degradation and structural modification during the formation of red resin.

To further elucidate the diversity of steroidal degradation intermediates in this biotransformation process, we conducted a systematic phytochemical reinvestigation of Chinese dragon’s blood from Yunnan Province. This study resulted in the isolation of two additional 12(13→14)abeo-spirostanol derivatives, named dracaenogenins C (1) and D (2). These rare abeo-spirostanol sapogenins represent potential intermediate products in the degradation pathway from steroidal saponins to the phenolic compounds that dominate mature resin. Herein, we report the isolation, structural elucidation, and cytotoxicity evaluation of these compounds, providing new insights into the molecular rearrangement of plant steroids during dragon’s blood formation and supporting the metabolic shift hypothesis at the chemical level.

2. Results

2.1. Structural Elucidation

The methanol extract of Dracaena cochinchinensis resin was subjected to systematic chromatographic separation over Toyopearl (Tosoh, Tokyo, Japan), HW-40C, silica gel, Sephadex (Cytiva, Marlborough, MA, USA) LH-20, reversed-phase C₁₈, and semi-preparative HPLC, affording fourteen steroidal constituents (1–14). Among these, compounds 1 and 2 were identified as new 12(13→14)abeo-spirostanol derivatives, named dracaenogenin C (1) and dracaenogenin D (2), respectively, based on detailed spectroscopic analysis (see Section 2.1).

The remaining twelve known compounds were identified as dracaenogenin A (3) [12], neoruscogenin (4) [13], namogenin B (5) [14], (25S)-namogenin A (6) [14], dracaenogenin B (7) [12], (20R,22S,25R)-1β,3β,14α,15α-spirost-5-ene-tetrol (8) [15], 25(S)-ruscogenin 1-O-β-D-glucopyranoside (9) [16], yamogenin 3-O-β-D-glucopyranoside (10) [17], (1β,3β)-1-hydroxyspirosta-5,25(27)-dien-3-yl β-D-glucopyranoside (11) [18], prosapogenin 2 (12) [19], dioscin (13) and gracillin (14) [8], by comparison of their NMR spectroscopic and mass spectrometric data with values reported in the literature. The structures of all isolated compounds are shown in Figure 1.

2.1.1. Dracaenogenin C (1)

Compound 1 was obtained as a white amorphous powder. Its molecular formula was determined as C₂₇H₄₂O₆ by positive-ion HR-ESI-MS, which showed a pseudomolecular ion peak at m/z 485.2894 [M + Na]⁺ (calcd. for C₂₇H₄₂O₆Na, 485.2894), consistent with seven degrees of unsaturation.

The ¹H NMR spectrum (pyridine-d₅, 600 MHz) displayed four methyl signals at δ_H 0.56 (d, J = 5.6 Hz, H-27), 1.28 (s, H-19), 1.37 (d, J = 7.0 Hz, H-21), and 1.59 (s, H-18), which correlated in the HSQC spectrum with carbon signals at δ_C 17.5 (C-27), 13.0 (C-19), 16.4 (C-21), and 22.8 (C-18), respectively. Four oxygenated methine protons were observed at δ_H 3.75 (dd, J = 11.9, 4.1 Hz, H-1), 3.96 (tt, J = 10.8, 4.8 Hz, H-3), 4.43 (d, J = 9.5 Hz, H-15), and 4.39 (m, H-16), corresponding to carbons at δ_C 78.3 (C-1), 68.8 (C-3), 77.7 (C-15), and 75.2 (C-16), respectively. Additionally, two oxygenated methylene protons appeared at δ_H 3.59 (m, Ha-26) and 3.54 (dd, J = 10.7, 3.7 Hz, Hb-26), correlating with δ_C 67.5 (C-26). Five aliphatic methine protons were assigned at δ_H 2.39 (td, J = 11.8, 5.0 Hz, H-8), 2.13 (td, J = 12.3, 5.2 Hz, H-9), 2.25 (m, H-17), 2.76 (m, H-20), and 1.53 (m, H-25). The ¹³C NMR spectrum further revealed one olefinic quaternary carbon at δ_C 140.8 (C-5) and four sp³-hybridized quaternary carbons at δ_C 44.8 (C-10), 92.6 (C-13), 54.2 (C-14), and 105.8 (C-22).

In the ¹H–¹H COSY spectrum, correlations were observed between Hb-2 (δ_H 2.23) and H-1/H-3. Starting from H-6, sequential correlations were detected with Ha-7 (δ_H 2.52), H-8 (δ_H 2.39), H-9 (δ_H 2.13), Ha-11 (δ_H 1.68), and Ha-12 (δ_H 2.62). Correlations from H-15 extended to H-16 (δ_H 4.39), H-17 (δ_H 2.25), H-20 (δ_H 2.76), and H-21 (δ_H 1.37). Starting from Ha-26 (δ_H 3.59), correlations were observed with H-25 (δ_H 1.53), H-27 (δ_H 0.56), Ha-24 (δ_H 2.54), and Ha-23 (δ_H 1.67). These data established four spin-coupling systems, as indicated by the bold lines in Figure 2.

HMBC correlations showed that H-18 correlated with C-12, C-13, C-14, and C-17; H-19 correlated with C-1, C-5, C-9, and C-10; H-21 correlated with C-17, C-20, and C-22; H-27 correlated with C-24, C-25, and C-26; H-4 correlated with C-3, C-5, and C-6; H-2 correlated with C-3; and H-8 correlated with C-13, C-14, and C-15. On the basis of these COSY and HMBC correlations, the planar structure of compound 1 was established as shown in Figure 2.

The relative configuration of 1 was established by NOESY spectroscopy(Figure 3). Correlations between H-1, H-3, and H-9 indicated these protons are α-oriented, while cross-peaks between H-8 and CH₃-19 confirmed the β-orientation of the angular methyl group at C-10. Critically, the observed NOESY correlation between CH₃-18 and H-17 indicated their spatial proximity, establishing the 13S configuration. Key NOESY correlations between H-15/H-20 and H-16/H-17 placed H-15 on the β-face and H-16 on the α-face. Additionally, correlations of H-20 with both H-15 and H-17 supported the 20S configuration. The characteristic high-field shift of CH₃-27 (δ_H 0.56, d, J = 5.6 Hz)—diagnostic for this stereochemistry—together with NOESY correlations between CH₃-27 and both Ha-23 and CH₃-18 protons, indicated the 25R configuration. The absolute configuration was unambiguously determined as (1R,3R,8R,9S,10R,13S,14R,15R,16R,17S,20S,22S,25R) by single-crystal X-ray diffraction analysis (The Flack parameter was −0.1(11)) (Figure 4).

Thus, the structure of compound 1 was determined as (1β,3β,15α,16β,25R)-12(13→14)abeo-spirost-5-ene-1,3,15,16-tetrol, and named dracaenogenin C.

2.1.2. Dracaenogenin D (2)

Compound 2 was isolated as a white powder. HRESIMS analysis in positive-ion mode gave a pseudomolecular ion peak at m/z 467.2793 [M + Na]⁺ (calcd for C₂₇H₄₀O₅Na, 467.2793), establishing the molecular formula as C₂₇H₄₀O₅.

Inspection of the ¹H NMR spectrum (pyridine-d₅, 600 MHz) revealed an exomethylene group resonating at δ_H 5.11 and 5.03 (J = 2.2 Hz each, H-18), along with three methyl signals at δ_H1.38 (s, H-19), 1.27 (d, J = 6.9 Hz, H-21), and 1.07 (d, J = 7.1 Hz, H-27). Notably, the chemical shift of H-27 (δ_H 1.07) appeared significantly downfield compared to that of dracaenogenin A [12] (δ_H 0.57), suggesting a different configuration at C-25. Four oxymethine protons appeared at δ_H 3.78 (dd, J = 11.9, 4.1 Hz, H-1), 3.99 (m, H-3), 4.31 (d, J = 6.8 Hz, H-15), and 4.67 (dd, J = 9.7, 6.5 Hz, H-16). An oxymethylene was observed at δ_H 4.07 (dd, J = 11.0, 2.8 Hz, Ha-26) and 3.36 (m, Hb-26).

The ¹³C NMR and DEPT spectra displayed 27 carbon signals, including two olefinic quaternary carbons at δ_C 141.3 (C-5) and 160.0 (C-13), an exomethylene carbon at δ_C 103.3 (C-18), and three sp³ quaternary carbons at δ_C 45.2 (C-10), 59.1 (C-14), and 109.9 (C-22). The oxygenated quaternary carbon at δ_C 59.1 (C-14) indicated hydroxylation at this position, consistent with the presence of a C-14 hydroxyl group in the 12(13→14)abeo framework.

Four separate spin systems were identified through ¹H–¹H COSY correlations: H-1/H-2/H-3; H-6/H-7/H-8/H-9/H-11/H-12; H-15/H-16/H-17/H-20/H-21; and H-23/H-24/H-25/H-26/H-27. HMBC correlations from H-18 (δ_H 5.11, 5.03) to C-12 (δ_C 34.2), C-13 (δ_C 160.0), C-14 (δ_C 59.1), and C-17 (δ_C 51.3) confirmed the 12(13→14)abeo-spirostane skeleton with a Δ¹³⁽¹⁸⁾-exomethylene moiety. Additional key HMBC correlations included H-19 to C-1, C-5, C-9, and C-10; H-8 to C-13 and C-14; H-18 to C-13, C-14, and C-17; and H-15 to C-12 and C-14.

The relative configuration was determined through NOESY experiments. Correlations between H-1, H-3, and H-9 placed these protons on the α-face, whereas cross-peaks between H-8, H-18, and H-19 established their β-orientation. H-15 showed NOE contacts to H-7 and H-20, indicating an α-orientation for H-15. H-17 correlated with H-16 and H-21, fixing the α-orientation of H-16 and H-17 and the β-orientation of the C-21 methyl. The β-orientation of H-27 was inferred from correlations between H-27, H-21, and Ha-23. The critical distinction from dracaenogenin A lies in the configuration at C-25. The H-27 resonance at δ_H 1.07 (d, J = 7.1 Hz) and the chemical shift of H-25 at δ_H 1.59, compared to δ_H 0.57 for H-27 in dracaenogenin A (25R), unambiguously established the 25S configuration for compound 2 [20].

Thus, the structure of compound 2 was determined as the C-25 epimer of dracaenogenin A, namely (25S)-12(13→14)abeo-spirosta-5,13(18)-diene-1β,3β,14α,15α-tetrol, and named dracaenogenin D.

2.2. Cytotoxicity Evaluation

All isolated compounds were evaluated for cytotoxicity against HepG2 human hepatocellular carcinoma cells at 30 μM using the MTT assay. As shown in Figure 5, the glycosylated steroidal saponins dioscin (13) and gracillin (14), together with compounds 9 and 11, exhibited potent cytotoxic activity. In marked contrast, the steroidal aglycones, including the 12(13→14)abeo-spirostanols dracaenogenin C (1) and dracaenogenin D (2), as well as other sapogenins (3–8, 10, 12), displayed negligible cytotoxicity under identical conditions.

These results suggest that glycosylated steroidal saponins present in the fresh stems of D. cochinchinensis possess inherent cytotoxic potential. The significant reduction or loss of cytotoxicity upon deglycosylation and skeletal rearrangement suggests that the biotransformation of saponins into abeo-spirostanol aglycones may function as a detoxification mechanism during red resin formation. This metabolic adaptation likely protects surrounding tissues from the phytotoxic effects of accumulated saponins while facilitating wound sealing through resin production.

3. Discussion

The fresh stems of D. cochinchinensis accumulate cytotoxic diosgenin-type saponins, notably dioscin (13) and gracillin (14), as the principal steroidal constituents. Following physical injury, localized tissue acidification activates endogenous glycosidases, promoting the enzymatic hydrolysis of glycosidic bonds to release steroidal sapogenin aglycones [21].

The liberated steroidal sapogenins undergo acid-catalyzed Wagner–Meerwein rearrangement under acidic microenvironments. The mechanism involves protonation of the Δ¹²-olefin to generate a carbocation intermediate at C-13. Stabilization of this high-energy species occurs via a 1,2-methyl shift, wherein the C-18 angular methyl group migrates from C-10 to C-13, accompanied by cleavage of the C-12–C-13 bond and concomitant formation of a new C-12–C-14 linkage [22,23]. This skeletal rearrangement yields the structurally novel 12(13→14)-abeo-spirostanol derivatives. From a biogenetic perspective, these compounds represent rare catabolic intermediates in the saponin degradation pathway rather than primary metabolites (Figure 6).

While normal cell lines would be ideal for studying toxicity on plant tissues, HepG2 cells are a standard model for general cytotoxicity evaluation. Saponins like dioscin are known to be broadly cytotoxic. The significant loss of activity against HepG2 suggests a reduction in general cellular toxicity, which aligns with the proposed detoxification mechanism in planta—the reduction of inherent phytotoxicity, protecting the surrounding plant tissues during resin formation.

Deglycosylation is known to reduce hemolytic activity and certain toxicities, but the Wagner–Meerwein rearrangement drastically alters the stereochemistry and shape of the steroid core (C/D ring junction), likely further diminishing interactions with cellular targets (e.g., membrane sterols).

4. Materials and Methods

4.1. General Experimental Procedures

Electrospray ionization mass spectrometry (ESI-MS) and high-resolution electrospray ionization mass spectrometry (HR-ESI-MS) were performed on an Agilent 6530 Series Q-TOF mass spectrometer (Agilent Technologies, Santa Clara, CA, USA). NMR spectra were recorded on a Bruker Avance III HD-600 spectrometer (Bruker, Ettlingen, Germany) operating at 600 MHz for ¹H and 150 MHz for ¹³C, respectively. Chemical shifts (δ) are expressed in parts per million (ppm) relative to tetramethylsilane (TMS) as an internal standard, and coupling constants (J) are reported in Hertz (Hz). Samples were dissolved in pyridine-d₅ (C₅D₅N) and measured at 25 °C.

Column chromatography (CC) was performed using silica gel (100–200 mesh or 200–300 mesh, Qingdao Marine Chemical Factory, Qingdao, China) and reversed-phase C₁₈ silica gel (40–63 μm, Merck, Darmstadt, Germany). Thin-layer chromatography (TLC) was carried out on pre-coated silica gel GF₂₅₄ plates (Qingdao Haiyang Chemical Co., Ltd., Qingdao, China). Spots were visualized by spraying with 5% anisaldehyde in ethanol (reagent A) or 10% sulfuric acid in ethanol solution, followed by heating at 105 °C.

Analytical and semi-preparative high-performance liquid chromatography (HPLC) were performed on a Waters 2695/2996 system (Waters, Milford, MA, USA) equipped with a photodiode array (PDA) detector. Analytical separations were conducted on a Waters Sunfire C₁₈ column (5 μm, 250 mm × 4.6 mm i.d.), while semi-preparative separations were performed on a COSMOSIL 5C₁₈-MS-II preparative column (10 mm i.d. × 250 mm, Nacalai Tesque, Kyoto, Japan). The flow rate and mobile phase gradients are specified below in the Section 4.3.

4.2. Plant Material

The red resin of Dracaena cochinchinensis (Lour.) S.C. Chen (Chinese Dragon’s Blood) was purchased from a local medicinal material market in Yunnan Province, China, in August 2023. The botanical identity was authenticated by Professor Haizhou Li (Kunming University of Science and Technology, Kunming, China). A voucher specimen (No. RDCC-2023-08) has been deposited in the Laboratory of Pharmaceutical Chemical Biology, School of Life Science and Technology, Kunming University of Science and Technology.

4.3. Extraction and Isolation

The powdered red resin of Dracaena cochinchinensis (450 g) was extracted with methanol (MeOH) (3 × 2.0 L) at room temperature by maceration. The combined extracts were filtered and concentrated under reduced pressure to yield a dark red methanol extract (432.0 g). The extract was suspended in MeOH (500 mL) and partitioned with n-hexane (3 × 100 mL) to remove non-polar impurities. The methanol-soluble fraction (414.9 g) was subjected to column chromatography over Toyopearl HW-40C resin (Tosoh, Tokyo, Japan), eluted sequentially with EtOH–H₂O (60:40 → 100:0, v/v) and then acetone–H₂O (50:50, v/v), to afford five fractions (Fr.1–Fr.5) based on TLC analysis.

Fraction 1 (20.8 g), enriched in steroidal compounds, was further purified by silica gel column chromatography eluted with petroleum ether–ethyl acetate (PE–EtOAc, 8:1 → 1:1, v/v) followed by dichloromethane–methanol (CH₂Cl₂–MeOH, 6:1 → 2:1, v/v), yielding thirteen subfractions (Fr.1-1 to Fr.1-13).

Fraction 1–9 (1.2 g) was subjected to reversed-phase C₁₈ chromatography (MeOH–H₂O, 50% → 100%) and further purified by silica gel chromatography (CH₂Cl₂–MeOH, 8:1), followed by semi-preparative HPLC (MeOH–H₂O, 72:28, v/v; flow rate: 3.0 mL/min; detection: 210 nm) to yield compound 1 (10 mg), compound 2 (3 mg), and compounds 3 (2 mg), 4 (5 mg), 5 (10 mg), 6 (10 mg), 7 (6 mg), and 8 (9 mg).

Fraction 1–10 (7.0 g) was purified by RP-C₁₈ and silica gel chromatography, followed by semi-preparative HPLC (MeOH–H₂O, 68:32, v/v; flow rate: 3.0 mL/min), affording compounds 9 (18 mg), 10 (20 mg), 11 (10 mg), and 12 (8 mg).

Fraction 1–11 (4.9 g) was separated by repeated RP-C₁₈ column chromatography (gradient elution: 40%→ 80% MeOH in H₂O) to yield compound 13 (100 mg) and 14 (80 mg).

4.3.1. Dracaenogenin C (1)

White amorphous powder (10 mg); HRESIMS (Positive) m/z 485.2894 [M + Na]⁺ (cal for C₂₇H₄₂O₆Na⁺, 485.2894); ¹H NMR (C₅D₅N, 600 MHz) and ¹³C NMR (C₅D₅N, 150 MHz) data in

Table 1. ¹H (600 MHz) and ¹³C (150 MHz) spectral data of compounds 1–2 (in pyridine d₅), δ in ppm, J in Hz.

Position	1		2
No.	δC, Type	δH (J in Hz)	δC, Type	δH (J in Hz)
1	78.3, CH	3.75, dd (11.9, 4.1)	78.1, CH	3.78, dd (11.9, 4.1)
2	42.8, CH2	2.56, dt (12.2, 5.7)	42.8, CH2	2.59, m
		2.23, m		2.29, m
3	68.8, CH	3.96, tt (10.8, 4.8)	68.8, CH	3.99, m
4	43.3, CH2	2.68, m	43.4, CH2	2.74, m
		2.60, m
5	140.8, C		141.3, C
6	126.8, CH	5.79, d (5.4)	125.5, CH	5.70, dd (5.4, 1.6)
7	31.0, CH2	2.76, m	27.7, CH2	2.53, m
		2.52, m		2.11, m
8	41.0, CH	2.39, td (11.8, 5.0)	42.0, CH	2.27, m
9	54.2, CH	2.13, td (12.3, 5.2)	53.4, CH	2.38, td (11.6, 6.4)
10	44.8, C		45.2, C
11	29.9, CH2	1.68, m	27.9, CH2	2.65, m
		1.49, m		1.90, m
12	28.3, CH2	2.62, dd (11.8, 5.2)	34.2, CH2	2.74, m
		1.55, m		1.56, m
13	92.6, C		160.0, C
14	54.2, C		59.1, C
15	77.7, CH	4.43, d (9.5)	80.8, CH	4.31, d (6.8)
16	75.2, CH	4.39, m	86.8, CH	4.67, dd (9.7, 6.5)
17	56.1, CH	2.25, m	51.3, CH	3.23, ddd (9.6, 7.2, 2.4)
18	22.8, CH3	1.59, s	103.3, CH2	5.11, d (2.2)
				5.03, d (2.2)
19	13.0, CH3	1.28, s	13.1, CH3	1.38, s
20	46.6, CH	2.76, m	50.4, CH	1.81, m
21	16.4, CH3	1.37, d (7.0)	14.4, CH3	1.27, d (6.9)
22	105.8, C		109.9, C
23	33.9, CH2	1.67, m	26.8, CH2	1.93, m
		1.55, m		1.41, m
24	30.0, CH	2.54, m	26.3, CH2	2.15, m
		1.67, m		1.39, m
25	30.8, CH	1.53, m	28.1, CH	1.59, m
26	67.5, CH2	3.59, m	65.8, CH2	4.07, dd (11.0, 2.8)
		3.54, dd (10.7, 3.7)		3.36, m
27	17.5, CH3	0.56, d (5.6)	16.6, CH3	1.07, d (7.1)

.

4.3.2. Dracaenogenin D (2)

White amorphous powder (3 mg); HRESIMS (Positive) m/z 467.2793 [M + Na]⁺ (cal for C₂₇H₄₀O₅Na⁺, 467.2793); ¹H NMR (C₅D₅N, 600 MHz) and ¹³C NMR (C₅D₅N, 150 MHz) data in Table 1.

4.3.3. X-Ray Crystal Structure Analysis of 1

Single crystals of compound 1 suitable for X-ray diffraction were obtained by slow evaporation from a mixed solvent of methanol and dichloromethane at room temperature. Crystal data were collected on a Bruker D8 Venture diffractometer (Bruker, Ettlingen, Germany) using Mo Kα radiation (λ = 0.71073 Å) at 100(2) K. The structure was solved by direct methods using SHELXT-2018 and refined by full-matrix least-squares on F² using SHELXL-2018/3. All non-hydrogen atoms were refined anisotropically. Hydrogen atoms were placed in calculated positions and refined using a riding model.

Crystal data for compound 1: C₂₇H₄₂O₆, M = 462.61, tetragonal, space group P4₃2₁2 (No. 96), a = 8.9330(8) Å, b = 8.9330(8) Å, c = 64.200(8) Å, α = β = γ = 90°, V = 5123.0(9) ų, Z = 8, Dc = 1.200 g/cm³, μ(Mo Kα) = 0.083 mm⁻¹. A total of 50,756 reflections were measured, of which 6356 were unique (Rint = 0.0911). The final R₁ values were 0.0550 [for I > 2 σ(I)] and 0.0685 (for all data). The final wR(F²) values were 0.1357 [for I > 2 σ(I)] and 0.1458 (for all data). The goodness-of-fit (S) on F² was 1.084. The Flack parameter was −0.1(11), confirming the absolute configuration. CCDC 2,528,261.

4.4. Screening of Cytotoxic Activity of the Compounds

The cytotoxic activities of the isolated compounds were evaluated against human hepatocellular carcinoma (HepG2) cells using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) colorimetric assay [24]. HepG2 cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin at 37 °C in a humidified atmosphere containing 5% CO₂.

Cells in the logarithmic growth phase were harvested using 0.25% trypsin-EDTA and seeded into 96-well microplates at a density of 1000 cells per well (100 μL of 10,000 cells/mL suspension). After 24 h incubation to allow cell attachment, the culture medium was replaced with 200 μL of fresh medium containing the test compounds at various concentrations (ranging from 1.56 to 100 μM, dissolved in DMSO with final concentration ≤ 0.1%). Control groups included: (1) solvent control (0.1% DMSO), (2) positive control (e.g., doxorubicin or cisplatin, if used), and (3) blank control (medium only). Each treatment was performed in quintuplicate (n = 5).

Following 72 h of incubation with the test compounds, 20 μL of MTT solution (5 mg/mL in phosphate-buffered saline, PBS) was added to each well, and the plates were incubated for an additional 4 h at 37 °C. The resulting formazan crystals were dissolved in 150 μL of DMSO (or isopropanol containing 0.04 N HCl), and the absorbance was measured at 490 nm using a microplate reader (BioTek, Shoreline, WA, USA). Cell viability was calculated as: (Absorbance_sample − Absorbance_blank)/(Absorbance_control − Absorbance_blank) × 100% [24].

5. Conclusions

In conclusion, this study reports the isolation and structural elucidation of two new 12(13→14)-abeo-spirostanol aglycones, dracaenogenins C (1) and D (2), alongside twelve known steroidal constituents from Dracaena cochinchinensis resin. Compound 2 represents a rare example of a naturally occurring C-14α-hydroxylated 12(13→14)-abeo-spirostanol, while 1 features oxygenation at C-15 and C-16.

Biogenetic analysis indicates that these unusual aglycones originate from the acid-catalyzed degradation of diosgenin-type steroidal saponins during resin formation. Following physical injury, localized acidification and enzymatic hydrolysis release sapogenin precursors, which subsequently undergo Wagner–Meerwein rearrangement featuring C-18 angular methyl migration (C-10→C-13) and C-ring contraction to forge the 12(13→14)-abeo framework.

Significantly, cytotoxicity assays revealed that the parent glycosylated saponins (e.g., 13 and 14) exhibited potent toxicity against HepG2 cells, whereas the rearranged abeo-aglycones (1, 2) were devoid of cytotoxicity. This supports a detoxification mechanism wherein the biotransformation of toxic saponins into non-toxic abeo-intermediates facilitates wound healing without compromising tissue viability.

These findings expand the chemical diversity of natural steroidal sapogenins and provide experimental evidence for the catabolic origin of abeo-spirostanols in plant resins. The discovery of these rearranged steroids offers new insights into the metabolic plasticity of D. cochinchinensis and the molecular mechanisms underlying dragon’s blood formation.