4.1 General experimental procedures

Column chromatography (CC) was performed with silica gel (100–200 and 200–300 mesh, Marine Chemical Industry, Qingdao, China) and ODS (50 μm, YMC Group, Kyoto, Japan). Semipreparative high performance liquid chromatography (HPLC) separations were carried out on a Shimadzu LC-40 HPLC system, equipped with a DAD detector, using a reversed-phase (RP) C18 ODS column (10ID × 250 mm and 4.6ID × 250 mm, Cosmosil 5C18-MS-II Packed column, Nacalai Tesque, Japan). HRESIMS data were collected in the positive-ion mode on a Bruker Maxis HD mass spectrometer (Bruker, Germany). UV and IR spectra were obtained on an Evolution 300 instrument (Thermo Scientific, MA, USA) and a Nicolet IS 10 spectrophotometer (Thermo Scientific, MA, USA), respectively. Optical rotations and CD spectra were acquired by a Rudolph AP-IV polarimeter (Rudolph, Hackettstown, NJ, USA) and a Chirascan qCD spectrometer (Applied Photophysics Ltd, Surrey, U.K), respectively. The NMR spectra were measured on a Bruker Advance III 500 spectrometer (Bruker, Germany), and chemical shifts were referenced to the residual acetone-d6 (δH 2.05/δC 29.84) signals. Incubator and microplate reader used in activity experiments were carbon dioxide incubator 3111 and Multiskan MK3 microplate reader (Thermo Scientific, USA).

4.2 Plant material

The aerial parts of Epimedium sagittatum Maxim. (Berberidaceae) were collected from Fenghui Epimedium herb GAP Base, Zhumadian, Henan Province, People’s Republic of China (GPS data: 114.496128, 33.018587), in September 2020. A voucher specimen (no. 20200960) was deposited in the Department of Pharmacy, Henan University of Chinese Medicine.

4.3 Extraction and isolation

The air-dried aerial parts of E. sagittatum (80 kg) were extracted with 70% EtOH in three cycles, resulting in the production of an extract weighing 6.5 kg. Then, the extract was suspended with distilled water and successively partitioned into CH2Cl2 and EtOAc sequentially to obtain CH2Cl2, EtOAc, and aqueous fractions. Subsequently, the CH2Cl2 fraction (2.1 kg) was subjected to silica gel column chromatography (CC, 100–200 mesh) and eluted with a petroleum ether/ethyl acetate mixture (v:v, 50:1–0:1) to obtain eight fractions (Fr. A–H).

Fraction F (13.7 g) was subjected to silica gel CC (200–300 mesh) using a petroleum ether/EtOAc elution gradient (v:v, 35:1–0:1) to afford fifteen subfractions, Fr. F1–F15. Then, Fr. F15 (1.8 g) was further separated by ODS CC (MeOH:H2O = 40:60–100:0) and semipreparative HPLC to obtained a mixture of 5a and 5b. This mixture was then separated using a Chiralpak IC column (5 μm, 10 × 250 mm, Daicel Chiral Technologies Co., Ltd., Japan) with elution using MeOH:H2O (v:v, 80:20) to yield enantiomers of 5a (1.3 mg, tR = 39.3 min) and 5b (1.2 mg, tR = 33.3 min).

Fraction G (100.0 g) was separated through silica gel CC (100–200 mesh) eluting with petroleum ether/EtOAc (v:v, 50:1–0:1) to yield eighteen subfractions, Fr. G1–G18. Then, Fr. G17 (65.0 g) was separated by silica gel CC (200–300 mesh) to obtain nineteen fractions, Fr. G17.1–G17.19. Subsequently, Fr. G17.15 (38.9 g) was separated by ODS CC (MeOH:H2O = 40:60–100:0) to obtain twelve fractions, Fr. G17.15.1–G17.15.12. Further purification of Fr. G17.15.9 by semipreparative HPLC yielded six subfractions, Fr. G17.15.9.1–G17.15.9.6. Subsequent purification of subfraction Fr. G17.15.9.5 through HPLC (MeCN-H2O, 60:40, 1 mL/min) afforded compound 1b (9.9 mg, tR 13.7 min), while compound 1a (5.6 mg, tR 14.0 min) was isolated from subfraction Fr. G17.15.9.6 under identical chromatographic conditions. Fr. G17.15.10 was isolated by ODS CC (MeOH:H2O = 50:50–100:0) to obtain Fr. G17.15.10.1–G17.15.10.13, and then Fr. G17.15.10.6 was purified by HPLC to yield Fr. G17.15.10.6.1–G17.15.10.6.7. Compounds 2a (2.2 mg, tR 42.2 min) and 2b (1.9 mg, tR 48.9 min) were isolated from Fr. G17.15.10.6.1 via HPLC (MeCN–H2O, 40:60, 2 mL/min). Fr. G17.15.10.6.2 was subjected to HPLC separation to obtain Fr. G17.15.10.6.2.1–G17.15.10.6.2.6. Compounds 4a (3.6 mg, tR 63.1 min) and 4b (2.1 mg, tR 59.1 min) were purified from Fr. G17.15.10.6.2.2 and Fr. G17.15.10.6.2.3, respectively. Fr. G17.15.10.6.2.4 was further separated by HPLC (MeCN–H2O, 60:40, 2 mL/min) into five subfractions, Fr. G17.15.10.6.2.4.1–G17.15.10.6.2.4.5, with the third and fifth fractions identified as 3a (1.0 mg, tR 58.8 min) and 6b (11.2 mg, tR 63.6 min), respectively. Compounds 3b (1.4 mg, tR 20.8 min) and 6a (9.0 mg, tR 19.3 min) were purified from Fr. G17.15.10.6.2.4.4 (MeOH–H2O, 80:20, 2 mL/min).

1a and 1b: UV (MeOH) λmax (log ε): 207 (4.24), 242 (3.89), 268 (3.82), 342 (3.96) nm; IR (νmax): 3400, 2976, 1653, 1616, 1477, 1439, 1359, 1166, 1031, 842 cm–1; 1H and 13C NMR data, see Tables 1 and 2.

(R)-epimesatine J (1a): Yellow amorphous powder; [α] 20D − 4 (c 0.4, MeOH); HRESIMS m/z 463.1725 [M + Na]+ (calcd. for C25H28O7Na, 463.1727).

(S)-epimesatine J (1b): Yellow amorphous powder; [α] 20D + 6 (c 0.4, MeOH); HRESIMS m/z 463.1713 [M + Na]+ (calcd. for C25H28O7Na, 463.1727).

2a and 2b: UV (MeOH) λmax (log ε): 208 (4.32), 240 (3.98), 268 (3.93), 341 (4.01) nm; IR (νmax): 3444, 2975, 1699, 1653, 1436, 1363, 1260, 1170, 1052 cm–1; 1H and 13C NMR data, see Tables 1 and 2.

(2″R,2\(\prime\prime\prime\)S)-epimesatine K (2a): Yellow amorphous powder; [α] 20D − 3 (c 0.2, MeOH); HRESIMS m/z 479.1665 [M + Na]+ (calcd. for C25H28O8Na, 479.1676).

(2″S,2\(\prime\prime\prime\)R)-epimesatine K (2b): Yellow amorphous powder; [α] 20D + 3 (c 0.2, MeOH); HRESIMS m/z 479.1669 [M + Na]+ (calcd. for C25H28O8Na, 479.1676).

3a and 3b: UV (MeOH) λmax (log ε): 204 (4.43), 238 (4.11), 269 (4.04), 344 (4.14) nm; IR (νmax): 3408, 2933, 1653, 1606, 1367, 1166, 1035, 843 cm–1; 1H and 13C NMR data, see Tables 1 and 2.

(2″R,2\(\prime\prime\prime\)R)-epimesatine L (3a): Yellow amorphous powder; [α] 20D − 12 (c 0.1, MeOH); HRESIMS m/z 461.1568 [M + Na]+ (calcd. for C25H26O7Na, 461.1571).

(2″S,2\(\prime\prime\prime\)S)-epimesatine L (3b): Yellow amorphous powder; [α] 20D + 11 (c 0.1, MeOH); HRESIMS m/z 461.1567 [M + Na]+ (calcd. for C25H26O7Na, 461.1571).

4a and 4b: UV (MeOH) λmax (log ε): 202 (4.34), 269 (3.96), 336 (4.05) nm; IR (νmax): 3388, 2932, 2860, 1654, 1608, 1365, 1167, 1048, 843 cm–1; 1H and 13C NMR data, see Tables 1 and 2.

(1″R,2″R,2\(\prime\prime\prime\)R)-epimesatine M (4a): Yellow amorphous powder; [α] 20D − 3 (c 0.3, MeOH); HRESIMS m/z 491.1672 [M + Na]+ (calcd. for C26H28O8Na, 491.1676).

(1″S,2″S,2\(\prime\prime\prime\)S)-epimesatine M (4b): Yellow amorphous powder; [α] 20D + 4 (c 0.2, MeOH); HRESIMS m/z 491.1675 [M + Na]+ (calcd. for C26H28O8Na, 491.1676).

5a and 5b: UV (MeOH) λmax (log ε): 203 (4.45), 268 (4.19), 335 (4.18) nm; IR (νmax): 3415, 1653, 1608, 1475, 1437, 1364, 1168, 1033, 981 cm–1; 1H and 13C NMR data, see Tables 1 and 2.

(1″S,2″S)-epimesatine N (5a): Yellow amorphous powder; [α] 20D − 7 (c 0.1, MeOH); HRESIMS m/z 475.1728 [M + Na]+ (calcd. for C26H28O7Na, 475.1727).

(1″R,2″R)-epimesatine N (5b): Yellow amorphous powder; [α] 20D + 10 (c 0.1, MeOH); HRESIMS m/z 475.1730 [M + Na]+ (calcd. for C26H28O7Na, 475.1727).

6a and 6b: UV (MeOH) λmax (log ε): 202 (4.50), 244 (4.15), 268 (4.08), 341 (4.26) nm; IR (νmax): 3368, 2975, 1655, 1614, 1475, 1439, 1362, 1166, 1044, 841 cm–1; 1H and 13C NMR data, see Tables 1 and 2.

(2″S,2\(\prime\prime\prime\)R)-epimesatine O (6a): Yellow amorphous powder; [α] 20D − 11 (c 0.4, MeOH); HRESIMS m/z 461.1570 [M + Na]+ (calcd. for C25H26O7Na, 461.1571).

(2″R,2\(\prime\prime\prime\)S)-epimesatine O (6b): Yellow amorphous powder; [α] 20D + 16 (c 0.5, MeOH); HRESIMS m/z 461.1568 [M + Na]+ (calcd. for C25H26O7Na, 461.1571).

4.4 Preparation of the Mo2(OAc)4 complex of compounds 1a, 1b, 2a, and 2b

Firstly, Mo2(OAc)4 (1.65 mg) was dissolved in DMSO (1 mL) at room temperature. Then, this solution was added to compound 1a (0.50 mg) and thoroughly mixed. The initial ECD spectrum was recorded immediately to establish a background absorption, followed by the recording of the complex-induced ECD spectra at ten minute intervals. The preparation of the Mo2(OAc)4-complex and ECD spectra of 1b, 2a, and 2b followed a similar procedure as that of 1a. The absolute configurations of these compounds were determined based on the Cotton effect observed in the complex-induced ECD spectra, in accordance with the helicity rule.

4.5 ECD calculations

The conformations of 2a/2b–6a/6b were analyzed by GMMX software using the MMFF94 force field. The geometry optimizations and predictions of the ECD spectra of the conformers were conducted through density functional theory (DFT) at the B3LYP/6-31G(d) level using the Gaussian 16W software [37]. The ECD curves were simulated by SpecDis software (version 1.71) based on the Boltzmann distribution theory [38].

4.6 Preparation of the Rh2(OCOCF3)4 complex of compounds 4a and 4b

Firstly, Rh2(OCOCF3)4 (1.20 mg) was dissolved in anhydrous CH2Cl2 (800 μL) at room temperature. Then, this solution was combined with compound 4a (0.63 mg) and thoroughly mixed. The first ECD spectrum was immediately recorded as a baseline, after which complex-induced ECD spectra were recorded at ten minute intervals until reaching a stable state. The preparation of the Rh2(OCOCF3)4-complex and ECD spectra of 4b (0.55 mg) followed a similar procedure as that of 4a. The absolute configurations of 4a and 4b were determined by the Cotton effect observed in the E band around 350 nm in the complex-induced ECD spectra.

4.7 Cytotoxicity assay4.7.1 Cell culture

Take frozen MCF-7 and MCF-10A cells (Shanghai Cell Bank, China) and melt them in a 37 °C water bath until the state of ice and water coexisted. Then, the mixture was centrifuged immediately at 1000 rpm for 5 min. Subsequently, discard the supernatant and transfer the cells to a DMEM medium containing 10% FBS (100 kU/L for both penicillin and streptomycin). Cultivate in a 37 °C constant temperature incubator containing 5% CO2 until the cells grow to 80% ~ 90% of the dish for passage. Replace fresh culture medium every 24 h.

4.7.2 Cell viability of isolated compounds on MCF-7 and MCF-10A cells using MTT method

The MCF-7 and MCF-10A cells were cultured until the logarithmic growth phase in an incubator containing 5% CO2 at 37 °C. Then, the cells were inoculated into a 96-well plate (E190236X, PerkinElmer, United States) at a density of 2 × 104 cells/mL and 200 μL per well. After 24 h of incubation, these wells were divided into normal control (CON) group and sample (10 μM) groups to culture for another 24 h. Subsequently, 20 μL of MTT (Solarbio life sciences, Beijing, China) solution (5 mg/mL) were added to each well and continue to culture for 4 h. Then, the culture medium was aspirated carefully and 150 μL of DMSO were added to each well to dissolve the blue-violet crystals. The OD values of each well were measured by a microplate reader at 490 nm, and the cell viability was calculated.

4.7.3 Cellular immunofluorescence

The assay for cellular immunofluorescence was performed in 96-well plates, with MCF-7 cells seeded at a density of 2 × 104 cells/mL. After 24 h, the cells were divided into control (CON) and various sample groups (10 μM) and cultured for an additional 24 h. The cells were then fixed with 4% paraformaldehyde for 15 min and permeabilized with 0.25% Triton X-100 for 10 min. Subsequently, 1% BSA was added for blocking for 30 min, after which the primary antibody Sphk1 (ab262697, Abcam) was introduced and incubated overnight at 4 °C. The cells were subsequently washed three times with PBST, counterstained with DAPI for 4 min, washed once with PBS, and then imaged using the OperettaCLS high-content imaging analysis system (Opera Phenix, PerkinElmer, United States).

4.7.4 Real time cellular analysis (RTCA)

Baseline measurements were conducted using culture medium. MCF-7 cells in logarithmic growth phase were seeded into E-plate plates at a density of 2 × 104 cells/mL per well. After 24 h of incubation, these wells were divided into control (CON) group and sample (1 μM, 5 μM, 10 μM, 20 μM, 50 μM, and 100 μM) groups, and their growth curves were continuously measured. After completion, the Sigmoidal dose–response (Vanable slope) algorithm was employed to determine the target time for calculating the IC50 values of the analyzed compounds. Docetaxel (Shanghai yuanye Bio-Technology Co., Ltd, Shanghai, China) was used as positive control.

4.7.5 Statistical analysis

The experimental data were expressed as mean ± standard deviation (SD) and analyzed by SPSS 26.0 software. One-way analysis of variance was used for comparison between groups. *P < 0.05 indicates a significant difference, while **P < 0.01 indicates an extremely significant difference.

4.8 Molecular docking

Molecular docking simulations were performed using the software AutoDock Tools (1.5.7). The crystal structures of Sphk1 was obtained from the RCSB Protein Data Bank (http://www.pdb.org/) and embellished through PyMOL (3.0.3) and the AutoDock Tools softwares.