NMR spectra (Bruker Avance 600 MHz, Bruker BioSpin AG, Fällanden, Switzerland), HRESIMS data (AB SCIEX X500R QTOF, AB SCIEX LLC, Framingham, MA, United States), and analytical HPLC (Thermo Scientific Dionex UltiMate 3000, Thermo Fisher Scientific Inc., Waltham, MA, United States) were obtained using the respective instruments. Preparative HPLC was performed on a SEP LC-52 system from Separation (Beijing) Technology Co., Ltd. (Beijing, China), equipped with a MWD UV detector. A YMC Pack ODS-A C18 column (250 × 10 mm, 5 μm, YMC Co., Ltd., Kyoto, Japan) was used for separation. Column chromatography was performed using aluminum oxide (Shanghai Macklin Biochemical Co., Ltd., Shanghai, China), silica gel (200–300 mesh; Qingdao Marine Chemical Inc., Qingdao, China), and ODS (Tianjin Bonna-Agela Technologies Co., Ltd., Tianjin, China). Chromatographic-grade methanol (Sigma-Aldrich Co., LLC, St. Louis, MO, United States) and other analytical-grade solvents (Tianjin Kermel Chemical Reagent Co., Ltd., Tianjin, China) were used.

The roots of T. regelii were collected on 7 July 2019 from Ji’an City, Jilin Province, China. The roots were identified by Professor Chao Wang and deposited at Dalian Medical University.

T. regelii roots (30 kg) were extracted with 70% ethanol using heat reflux extraction (3 × 2 h). Following ethanol evaporation, the residual aqueous solution was diluted with water and subjected to sequential liquid-liquid partitioning using petroleum ether, EtOAc, and n-BuOH. The resulting EtOAc extract was divided into two portions: 20 g and 400 g. The petroleum ether extract was fractionated into 18 fractions (SG1–SG18) via silica gel column chromatography, eluting with gradient EtOAc/CH₃OH mixtures (20:1, 10:1, 5:1, 2:1, 1:1, 1:100). Fraction SG14 was further purified by preparative HPLC at 3 mL/min using a 58:42 CH₃CN/H₂O mixture (0.03% TFA) as the eluent, yielding compound 21 (6.8 mg). The 20 g portion of the EtOAc extract was subjected to silica gel column chromatography and eluted with a gradient of CH₂Cl₂/CH₃OH (100:0, 50:1, 20:1, 10:1, 1:1, 1:5, 1:10, 0:100), resulting in the isolation of 8 fractions designated D1 to D8. Fractions D2 and D3 were combined and further separated via ODS medium-pressure column chromatography using a CH₃OH/H₂O gradient elution system, which afforded 14 subfractions labeled DR1 to DR14. Subfractions DR6 to DR12 were merged and further purified by preparative HPLC at a flow rate of 3 mL/min. The eluent used was a 50:50 mixture of CH₃OH and H₂O containing 0.03% trifluoroacetic acid (TFA). This purification step yielded eleven compounds, namely compounds 3 (4.5 mg), 7 (10.5 mg), 10 (5.0 mg), 12 (7.0 mg), 13 (4.2 mg), 14 (3.3 mg), 15 (4.2 mg), 18 (6.6 mg), 20 (12.5 mg), 19 (8.7 mg), and 22 (5.0 mg). The 400 g portion of the EtOAc extract was fractionated into 4 fractions by aluminum oxide column chromatography, using a series of EtOAc/CH₃OH eluent mixtures with gradient ratios of 100:1, 4:1, 1:1, and 1:100. Fraction 2 was further sub-fractionated into 26 fractions (designated E1–E26) via silica gel column chromatography, eluting with CH₂Cl₂/CH₃OH mixtures in gradient proportions of 100:1, 50:1, 15:1, 10:1, and 5:1. Subsequently, sub-fractions E5–E6 were subjected to additional silica gel column chromatography, employing petroleum ether/EtOAc mixtures (100:1, 20:1, 10:1, 8:1, 5:1, 2:1, 1:100) as the eluent, which resulted in the isolation of 24 sub-fractions (ED1–ED24). Sub-fractions ED16–ED24 were purified by preparative HPLC (equipped with an RP C18 column) at a flow rate of 3 mL/min. The eluent used was a 70:30 MeOH/H₂O mixture containing 0.03% TFA (v/v), yielding compounds 1 (1.4 mg), 5 (2.3 mg), and 6 (4.5 mg). Sub-fractions ED11–ED15 were similarly purified by preparative HPLC (RP C18 column, 3 mL/min) using an 80:20 MeOH/H₂O mixture (0.03% TFA, v/v) as the eluent, affording compounds 4 (5.6 mg), 16 (3.8 mg), and 17 (2.2 mg). Fractions E7–E16 were combined and further purified by preparative HPLC at a flow rate of 3 mL/min. The elution was performed with a 60:40 MeOH/H₂O mixture containing 0.03% TFA, yielding four compounds: 2 (18.3 mg), 8 (4.8 mg), 9 (13.6 mg), and 11 (5.5 mg).

The colorectal cancer cell lines HCT116 and HT29 used in this study were obtained from the American Type Culture Collection (ATCC) and maintained in stable passage in our laboratory. Cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM; Gibco) supplemented with 10% fetal bovine serum (FBS; Gibco) and 5% antibiotic-antimycotic solution (Solarbio). Cultures were maintained in a humidified incubator at 37 °C with 5% CO₂, and cell passaging was performed every 3 days.

The cytotoxic effect of compounds isolated from T. regelii were assessed using a cell viability assay, with detailed procedures as previously described (Chen, et al., 2025). Briefly, colorectal cancer cells in the logarithmic growth phase were seeded into 96-well plates at 3,000 cells per well. Following cell adhesion, the cells were treated with drug-supplemented medium or solvent control for 48 h, and cell proliferation in each group was measured via the CCK-8 assay.

HCT116 or HT29 were trypsinized to form single-cell suspensions and seeded into 6-well plates at 500–1,000 per well. The cells were then treated with medium supplemented with various concentrations of compound 2 or solvent control for 10–14 days. Following treatment, cells were fixed with a fixative solution (methanol:glacial acetic acid:water = 1:1:8) for 10 min, then stained with 1% crystal violet for 10 min. Excess crystal violet was washed away with PBS to visualize colony formation across all groups.

Cells were seeded in 6-well plates and treated with compound 2 for 48 h, after which the subsequent assay was conducted. First, the BODIPY 581/591 C11 probe (Biyuntian, Cat. No. S0043M) was diluted at a ratio of 1:1000 in DMEM medium to prepare the probe working solution. Next, the cells were trypsinized, centrifuged, and rinsed once with PBS; the PBS was then discarded. For each tube of cell pellets, 200 µL of the diluted probe solution was added, followed by incubation at 37 °C for 30 min. Post-incubation, the samples were centrifuged to remove the supernatant, and the cell pellets were washed twice with PBS. Finally, the cells were resuspended in 500 µL of PBS and analyzed via flow cytometry using a BD C6 Plus instrument, with the experimental results processed and analyzed using Flow Jo software.

Cells were seeded in 6-well plates and treated with compound 2 for 48 h prior to the subsequent assay. First, cells were harvested via centrifugation; after removing the culture medium, they were resuspended and thoroughly mixed in electron microscopy (EM) fixative, followed by fixation at 4 °C for 2–4 h. The supernatant was then discarded, and 0.1 M phosphate buffer (PB, pH 7.4) was added. The mixture was gently agitated for a 3-min rinse, then centrifuged, and this washing procedure was repeated three times. Next, the samples were embedded in a 1% agarose solution. After fixation with 1% osmium tetroxide at room temperature for 2 h (protected from light), the samples were rinsed three times with 0.1 M PB (pH 7.4), with each rinse step lasting 15 min. Dehydration was performed by sequentially immersing the samples in a graded ethanol series (30% → 50% → 70% → 80% → 95% → 100% → 100%), with each immersion lasting 20 min. This was followed by two 15-min soaks in 100% acetone. For infiltration, the samples were incubated in an embedding agent mixture (acetone:812 resin = 1:1) at 37 °C overnight, then soaked in pure 812 resin at 37 °C for 5–8 h. The samples were then placed into embedding molds, incubated in an oven at 37 °C overnight, and polymerized at 60 °C for 48 h. Ultrathin sections (60–80 nm) were cut, then stained with a saturated 2% uranyl acetate alcoholic solution for 8 min (protected from light). After staining, the sections were rinsed three times with 70% ethanol and three times with ultrapure water in sequence. Subsequently, they were stained with a 2.6% lead citrate solution for 8 min (shielded from carbon dioxide), rinsed three more times with ultrapure water, and dried at room temperature overnight. Finally, the sections were observed under a transmission electron microscope (JEM1400PLUS), and images were captured for further analysis.

Cells in the growth phase, which had been cultured on coverslips, were seeded into 24-well plates. Following treatment with compound 2 for 24 h, the subsequent assay was performed. Briefly, the original culture medium was aspirated, and the cells were rinsed twice with PBS. The DCFH-DA fluorescent probe (Biyuntian, Cat. No. S0033M) was diluted in PBS at a ratio of 1:1000, and 500 µL of this diluted probe solution was added to each well. After incubating at 37 °C for 30 min, the solution in each well was removed, and any residual diluted probe was washed away with PBS. An anti-fluorescence quenching agent was then added for coverslip mounting, and images were acquired using an inverted fluorescence microscope (Olympus, Model 1X73) to obtain the assay results.

HCT116 and HT29 cell pellets treated with compound 2 were collected by centrifugation. Total proteins were extracted using RIPA lysis buffer, with concentrations quantified via BCA assay. Following separation by SDS-PAGE, proteins were transferred to PVDF membranes. Membranes were probed with primary antibodies against SLC7A11 (AWA00502, Abiowell), SLC3A2 (AWA61227, Abiowell), GPX4 (AWA11352, Abiowell), and GAPDH (81640-5-RR, Proteintech), where GAPDH served as a loading control to ensure equal protein loading. Proteins were visualized by autoradiography and quantified using ImageJ software.

Compound 1 was obtained as a white solid. Its molecular formula was established as C₃₃H₄₈O₆, corresponding to 10 degrees of unsaturation, based on HRESIMS data showing a peak at m/z 541.3486 [M + H]⁺ (calcd. for C₃₃H₄₉O₆ ⁺, 541.3524) and supported by NMR analysis (Table 1). The ¹H NMR spectrum of 1 indicated the presence of eight methyl groups [δ _H 1.11 (s, H₃-23), 1.07 (s, H₃-24), 1.26 (s, H₃-25), 1.17 (s, H₃-26), 1.42 (s, H₃-27), 0.93 (s, H₃-28), 1.33 (s, H₃-30), 2.05 (s, OCOCH₃)], a methoxyl group [δ _H 3.68 (s, COOCH₃)], one oxymethine group [δ _H 4.87 (dd, J = 12.0, 4.8 Hz, H-22)], a trisubstituted double bond [δ _H 5.66 (s, H-12)] (Table 1). The ¹³C NMR spectrum displayed 33 carbon resonances, comprising two ketone carbonyls [δ _C 217.2 (C-3) and 199.3 (C-11)], two ester carbonyls [δ _C 177.4 (C-29) and 170.6 (22-OCOCH₃)], two olefinic carbons [δ _C 129.1 (C-12) and 167.2 (C-13)], and one oxygenated carbon [δ _C 76.5 (C-22)] (Table 1). Analysis of the 1D NMR data suggested that 1 is an oleanane-type triterpenoid with a planar structure closely resembling that of 22β-acetylglabric acid (Shen et al., 1995). However, key differences were observed: the hydroxyl group at C-3 and the COOH group at C-29 in 22β-acetylglabric acid were replaced by a carbonyl group and a COOCH₃ group, respectively, in 1 (Figure 2). These modifications were corroborated by 2D NMR experiments (Figure 2). In the HMBC spectrum, correlations between OCH₃-29 and C-29, together with the chemical shift of C-29 (δ _C 177.4), confirmed the presence of COOCH₃ group (Figure 2). Additionally, the ketone at C-3 was assigned based on HMBC correlations from H₃-23 and H₃-24 to C-3, along with its characteristic chemical shift (δ _C 217.2, C-3) (Figure 2). The ¹H–¹H COSY spectrum revealed several contiguous proton spin systems: H₂-1/H₂-2, H-5/H₂-6/H₂-7, H₂-15/H₂-16, H-18/H₂-19, and H₂-21/H-22 (Figure 2). The relative configuration of 1 was elucidated via NOESY experiments, which showed key correlations between H₃-24 and H₃-25, H-9 and H₃-27, H₃-26 and H₃-28, H-18 and H-22, and H₃-30 and H-22 (Figure 3). Accordingly, the structure of 1 was determined as methyl 3,11-dioxo-22α-acetoxy-olean-12-en-29-oate, as depicted in Figure 1.

Compound 2 was obtained as a white solid. Its molecular formula was determined to be C₂₁H₂₆O₄ based on HRESIMS analysis, which showed an ion peak atm/z 343.1898 [M + H]⁺ (calcd. for C₂₁H₂₇O₄ ⁺, 343.1904). The ¹H NMR spectrum of 2 (in CD₃OD) displayed signals consistent with a pentasubstituted aromatic ring [δ _H 6.54 (s, H-11)], an oxymethylene group [δ _H 4.93 (d, J = 16.2 Hz, H-19a) and 4.86 (d, J = 16.2 Hz, H-19b)], a methoxyl group [δ _H 3.76 (s, OCH₃-12)], an isopropyl unit [δ _H 3.50 (m, H-15), 1.26 (d, J = 7.2 Hz, H₃-16), and 1.27 (d, J = 7.2 Hz, H₃-17)], and an angular methyl group [δ _H 1.04 (s, H₃-20)] (Table 1). The ¹³C NMR and HSQC spectra revealed 21 carbon resonances (Table 1), classified into four methyls, five methylenes, three methines, and nine quaternary carbons. The NMR data of 2 closely resembled those of the known compound triptophenolide (Zeng et al., 2013), with a key difference: the hydrogen atom at C-12 in triptophenolide was replaced by a methoxyl group in 2. This substitution was confirmed by the HMBC correlation from OCH₃-12 to C-12, supported by the chemical shift of C-12 (δ _C 158.3) (Figure 2). Therefore, the structure of 2 was established as 12-methoxytriptophenolide.

Compound 3 was assigned the molecular formula C₂₀H₂₆O₄ based on HRESIMS data, which exhibited an ion peak at m/z 331.1887 [M + H]⁺ (calcd. for C₂₀H₂₇O₄ ⁺, 331.1904). The ¹H NMR spectrum of 3 (in CD₃OD) displayed signals corresponding to a pentasubstituted aromatic ring [δ _H 6.49 (s, H-12)], an isopropyl unit [δ _H 3.31 (m, H-15),1.14 (d, J = 6.6 Hz, H₃-16), and 1.14 (d, J = 6.6 Hz, H₃-17)], an oxymethine group [δ _H 3.70 (dd, J = 12.6, 3.6 Hz, H-1)], and two angular methyl groups [δ _H 1.26 (s, H₃-18) and 0.92 (s, H₃-20)] (Table 1). The ¹³C NMR spectrum revealed 20 carbon signals, including an oxygenated carbon (δ _C 99.6), a carboxyl carbon (δ _C 181.1), and an aromatic group (δ _C 106.9, 120.4, 134.9, 137.9, 146.7, 153.3) (Table 1). These spectroscopic features indicated that 3 is structurally related to (1R,5S,10S)-1,11-oxyferruginol (Zhao et al., 2015). Further comparison of the NMR data between 3 and (1R,5S,10S)-1,11-oxyferruginol revealed several key differences: the hydroxyl group at C-12 in the reference compound was replaced by a hydrogen atom in 3, the hydrogen at C-14 was replaced by a hydroxyl group, and the methyl group at C-19 was oxidized to a carboxyl group (δ _C 181.1, C-19) (Figure 2). In the HMBC spectrum, correlations from H-12 to C-9, C-11, and C-14, along with the chemical shifts of C-12 (δ _C 106.9) and C-14 (δ _C 146.7), supported the presence of a hydroxyl group at C-14 (Figure 2). Additionally, the carboxyl group at C-19 was confirmed by HMBC correlations from H-3 and H₃-18 to C-19 (Figure 2). The relative configurations of the carboxyl group and H-1 were elucidated as β and α-oriented, respectively, based on the correlations of H-1/H-5 and H-1/H₃-18 in the NOESY spectrum (Figure 3). Thus, the structure of compound 3 was determined to be 1,11-epoxy-13-isopropyl-14-hydroxy-podocarpa-8,11,13-trien-19-oic acid.

Additionally, nineteen known compounds (4−22) were isolated from T. regelii, and their structures were elucidated by comparison of NMR spectroscopic data with those reported in the literature: triptobenzene E (4) (Takaishi et al., 1997), isoneotriptophenolide (5) (Zeng et al., 2013), triptophenolide methyl ether (6) (Deng et al., 1982), triptophenolide (7) (Zeng et al., 2013), quinone-21 (8) (Morota et al., 1995a), triptoquinone D (9) (Shishido et al., 1994), triptoquinone C (10) (Shishido et al., 1994), triptoquinone B (11) (Shishido et al., 1994), triptobenzene H (12) (Li et al., 1997), triptinin-A (13) (Xu et al., 1997), triptobenzene A (14) (Jiang et al., 2019), triptonoterpenol (15) (Zeng et al., 2013), triptobenzene B (16) (Takaishi et al., 1997), wilforol E (17) (Morota et al., 1995a), wilforol F (18) (Morota et al., 1995a), 8,12-dienabieta-11,14-dione-19-acid (19) (Ma et al., 2010), triptoquinone A (20) (Takaishi et al., 1992), wilforol B (21) (Morota et al., 1995b), and 6-hydroxy-celastrol (22) (Zhang et al., 2014).

Previous studies have indicated that diterpenoids and triterpenoids from T. regelii possess antitumor potential (Fan et al., 2016; Fan et al., 2017; Jiang et al., 2018; Wang et al., 2023). Employing the human colorectal cancer cell line HT29 as the in vitro experimental model, we assessed the antitumor potential of the 22 isolated terpenoid compounds through cell viability assays, with particular emphasis on investigating the antiproliferative activity of the newly identified compounds. Our screening results indicated that among the three newly identified compounds (1–3), compound 2 exhibited moderate antiproliferative activity against colorectal cancer in vitro, whereas compounds 1 and 3 showed weak or no detectable activity under the same experimental conditions (Figure 4). Overall, compounds 22, 6, 8–10, 16, 17, 21, and 22 exhibited antiproliferative effects against colorectal cancer cells (Figure 4). To further characterize the antiproliferative potency of compound 2, we determined its IC₅₀ values in two colorectal cancer cell lines (HCT116 and HT29), as shown in Figure 5A. After 48 h of treatment, compound 2 exhibited moderate inhibitory effects on CRC cells HCT116 and HT29 proliferation, with IC₅₀ values of 53.95 μM and 78.06 μM in the two cell lines, respectively. Additionally, colony formation assays further confirmed that compound 2 displays acceptable anti-colorectal cancer activity under prolonged exposure, as illustrated in Figures 5B,C.

We were also interested in elucidating the mechanism by which compound 2 induces colorectal cancer cell death. Given that inducing ferroptosis is an important anti-tumor strategy, we first examined whether compound 2 inhibits tumor cell proliferation via the ferroptosis pathway (Chen et al., 2021; Zhou et al., 2024). Western blot analysis initially confirmed that compound 2 exerts anti-colorectal cancer effects by inducing ferroptosis (Figures 6A,B). To investigate the role of ROS-mediated lipid peroxidation and mitochondrial damage, key drivers of ferroptosis (Cao et al., 2023; Niu et al., 2021), we utilized an ROS probe (DCFH-DA for general ROS) and fluorescence imaging. This approach revealed that compound 2 promotes ROS production in colorectal cancer cells. (Figure 6C). Furthermore, flow cytometry with a lipid peroxidation probe confirmed that compound 2 induces lipid peroxidation (Figure 6D). Additionally, transmission electron microscopy demonstrated that compound 2 induces mitochondrial damage in colorectal cancer cells (Figure 6E). Taken together, these findings indicate that compound 2 exerts anti-colorectal cancer activity by inducing ferroptosis.