3.1 General experimental procedures

The melting points of crystalline samples were determined on an X-5 digital micro-melting point apparatus and are presented without correction. UV spectra were obtained using a JASCO V-550 UV/VIS spectrophotometer, and optical rotations were measured on an Autopol JASCO P-1020 digital polarimeter. IR spectra were recorded on a JASCO FT/IR-480 Plus spectrometer using KBr pellets. NMR measurements, including 1D and 2D experiments, were performed on a Bruker Avance 600 spectrometer (600 MHz for 1H and 150 MHz for 13C) equipped with standard Bruker pulse programs. Chemical shifts (δ) were expressed in ppm with reference to residual solvent peaks. High-resolution electrospray ionization mass spectra (HR-ESI–MS) were acquired using a Waters Synapt G2 mass spectrometer. Single-crystal X-ray diffraction data were collected on an Agilent Gemini Ultra diffractometer fitted with a Cu-Kα radiation source (λ = 1.54178 Å) and a CCD area detector (Agilent Technologies, USA). HPLC analyses were conducted on a Shimadzu system (Shimadzu Corporation, Tokyo, Japan) equipped with a PDA detector, using analytical or preparative Waters XBridge RP18 columns (MeCN/H₂O or MeOH/H₂O as eluents). TLC analysis employed aluminium oxide 60 F₂₅₄ basic plates (Merck, China). The chromatographic materials used in this study included D-101 macroporous resin (Diaion, Shanghai, China), neutral alumina N (200–300 mesh; Aldrich, China), Sephadex LH-20 (25–100 μm; Fluka, Switzerland), MCI CHP20P gel (75–150 μm; Sigma-Aldrich, China), and ODS silica gel (50 μm; YMC, Japan). All solvents and reagents were of analytical grade and purchased from commercial suppliers in China.

3.2 Plant material

The roots of S. flavescens Ait. were collected in September 2019 from Xi’an, Shaanxi Province, China (34°46′–34°55′ N, 109°08′–109°16′ E). The botanical identity of the material was verified by Mr. Zhengqiu Mai, a senior herbal specialist at the Chinese Medicinal Material Co. (Guangzhou, China). A voucher specimen (No. SF-201909) is deposited at the Guangdong Provincial Key Laboratory of Veterinary Pharmaceutics Development and Safety Evaluation, College of Veterinary Medicine, South China Agricultural University, Guangzhou, China.

3.3 Extraction and isolation

The powdered roots of S. flavescens (25 kg) were extracted three times with 95% EtOH (48 h each), and the combined extracts were concentrated to yield a crude ethanol extract (~ 3.5 kg). Suspension of the crude extract in water followed by acidification to pH 3–4 with dilute HCl enabled the removal of neutral constituents through extraction with CHCl₃. Basification of the remaining aqueous phase to pH 9–10 with saturated NH₄OH, followed by extraction with CHCl₃, afforded the crude alkaloid extract (926 g, extraction coefficient: 3.7%). The alkaloid extract was chromatographed on a D-101 macroporous resin column eluted with graded EtOH–H₂O mixtures (10%–95%), yielding five fractions designated Fr.1–Fr.5. Fraction Fr.1 was dissolved in water and partitioned twice with EtOAc to remove lipophilic materials, and the aqueous-enriched portion was collected as Fr.1A (60.0 g). Chromatography of Fr.1A on neutral Al₂O₃ using CH₂Cl₂–MeOH (100:0 → 0:100, v/v, with 1% Et₂NH) yielded seven subfractions, Fr.1Aa–Fr.1Ag. Among these, Fr.1Ag was further purified on an MCI CHP20P column eluted with MeOH–H₂O–NH₄OH to obtain sixteen fractions (Fr.1Ag.1–Fr.1Ag.16). Fractions Fr.1Ag.1–Fr.1Ag.5 were successively subjected to Sephadex LH-20 chromatography (MeOH–H₂O, 1:1) to enrich minor alkaloids. The combined eluates were chromatographed on an ODS column with increasing proportions of MeOH–H₂O (5%–50%) to obtain Fr.1Ag.1–5. Final purification was carried out by semipreparative HPLC on a Waters XBridge™ Phenyl column with MeOH/H₂O or CH₃CN/H₂O (each containing 0.1% Et₂NH; isocratic or shallow gradients as appropriate), affording compounds 1 (88.8 mg, tR = 9.7 min), 2 (0.9 mg, tR = 6.4 min), 3 (54.2 mg, tR = 16.1 min), and 4 (33.9 mg, tR = 21.5 min). Compounds 5 (7.0 mg, tR = 31.0 min) and 6 (69.9 mg, tR = 19.0 min) were purified by preparative RP-HPLC using CH₃CN/H₂O/Et₂NH (15:85:0.01, v/v/v).

3.4 Spectroscopic data of the isolates

Sophflarine F (1): pale yellow crystals in CH3OH; m.p. 128–129 °C; [α] 25D ± 0 (c 0.01, CH3OH); UV (CH3OH) λmax (log ε) 223 (3.17), 237 (3.02), 291 (2.66), and 352 (2.61) nm; IR (KBr) νmax 3412, 3080, 2937, 2859, 1561, 1447, 1389, 1041 cm−1; 1H NMR (600 MHz, CD3OD) and 13C NMR (150 MHz, CD3OD), see Table 1; HRESIMS m/z 245.1282 [M + H]+ (calcd for C14H17N2O2, 245.1285).

Sophflarine G (2): pale yellow oil; [α] 25D ± 0 (c 0.01, CH3OH); UV (CH3OH) λmax (log ε) 200 (3.17), 223 (2.92), 290 (2.71) and 351 (2.61) nm; IR (KBr) νmax 3390, 2931, 2859, 1593,1511, 1415, 1036 cm−1; 1H NMR (600 MHz, CD3OD) and 13C NMR (150 MHz, CD3OD), see Table 1; HR-ESI–MS m/z: 261.1235 [M + H]+ (calcd for C14H17N2O3, 261.1234).

Sophflarine H (3): colorless crystals in CH3OH; mp 135–136 °C; [α] 25D ± 0 (c 0.01, CH3OH); UV (CH3OH) λmax (log ε) 201 (3.33) and 359 (2.91) nm; IR (KBr) νmax 3423, 2929, 2857, 1632, 1451, 1415, 1331, 1162 cm−11H NMR (600 MHz, CD3OD) and 13C NMR (150 MHz, CD3OD), see Table 3; HR-ESI–MS m/z: 279.1694 [M + H]+ (calcd for C15H23N2O3, 279.1703); ( +)-Sophflarine H: white amorphous powder; [α] 25D 130.7 (c 0.01, CH3OH); ECD (CH3OH) λmax (Δε) 233 (− 1.4), 357 (+ 2.9) nm; ( −)-Sophflarine H: white amorphous powder; [α] 25D − 135.5 (c 0.01, CH3OH); ECD (CH3OH) λmax (Δε) 233 (− 13.2), 357 (+ 28.5) nm.

Sophflarine I (4): brown oil; [α] 25D 3.3 (c 0.01, CH3OH); UV (CH3OH) λmax (log ε) 200 (3.18) and 359 (3.01) nm; ECD (CH3OH) λmax (Δε) 233 (− 1.4), 357 (+ 2.9) nm. IR (KBr) νmax 3443, 3287, 2922, 2854, 1629, 1470, 1411, 1330 cm−1; 1H NMR (600 MHz, CD3OD) and 13C NMR (150 MHz, CD3OD), see Table 3; HR-ESI–MS m/z: 295.1642 [M + H]+ (calcd for C15H23N2O4, 295.1652).

Sophflarine J (5): pale yellow oil; [α] 25D 6.3 (c 0.01, CH3OH); UV (CH3OH) λmax 200 (3.13), 308 (2.68) nm; ECD (CH3OH) λmax (Δε) 213 (+ 0.3), 297 (− 3.7), 340 (+ 1.7) nm. IR (KBr) νmax 2933, 2866, 1625, 1603, 1444, 1332 cm−1; 1H NMR (600 MHz, CD3OD) and 13C NMR (150 MHz, CD3OD), see Table 3; HR-ESI–MS m/z: 259.1439 [M + H]+ (calcd for C15H19N2O2, 259.1441).

Sophflarine K (6): brown oil; [α] 25D ± 0 (c 0.01, CH3OH); UV (CH3OH) λmax 205, 228, and 296 nm; IR (KBr) νmax 3443, 3287, 2922, 2854, 1629, 1470, 1451 cm−1; 1H NMR (600 MHz, CD3OD) and 13C NMR (150 MHz, CD3OD), see Table 3; HR-ESI–MS m/z: 277.1520 [M + H]+ (calcd for C15H21N2O3, 277.1547).

3.5 X-ray crystallographic analyses of compounds 1, 3, and 6

Single-crystal X-ray diffraction analyses of compounds 1, 3, and 6 were performed using an Agilent Gemini Ultra diffractometer equipped with a CCD area detector and employing Cu Kα radiation (λ = 1.54184 Å). Structural determination was conducted using direct methods via the SHELXT program, followed by full-matrix least-squares refinement on F2 using SHELXL or SHELXS within the Olex2 platform. The crystallographic information files (CIFs) for compounds 1, 3, and 6 have been deposited in the Cambridge Crystallographic Data Centre (CCDC, https://www.ccdc.cam.ac.uk/) under deposition numbers 2085308, 2,358,218, and 2,358,219, respectively. ORTEP representations of the crystal structures were generated using SHELXP. A summary of crystallographic data and refinement parameters is provided below:

Crystal data for sophflarine F (1): C14H20N2O4 (M = 280.32 g/mol), monoclinic, space group C2/c (no. 15), a = 50.0093(10) Å, b = 11.3394(2) Å, c = 14.5656(3) Å, β = 90.749(2)°, V = 8259.1(3) Å3, Z = 24, T = 149.99(10) K, μ(CuKα) = 0.823 mm−1, Dcalc = 1.353 g/cm3, 16,655 reflections measured (7.072° ≤ 2Θ ≤ 147.84°), 8107 unique (Rint = 0.0247, Rsigma = 0.0310) which were used in all calculations. The final R1 was 0.0754 (I > 2σ(I)) and wR2 was 0.1958 (all data). The goodness of fit on F2 was 1.080. CCDC number: 2085308.

Crystal data for sophflarine H (3): C15H22N2O4 (M = 258.30 g/mol): monoclinic, space group P21/n (no. 14), a = 8.340(2) Å, b = 14.179(4) Å, c = 12.361(3) Å, β = 98.75(3)°, V = 1444.8(7) Å3, Z = 4, T = 293(2) K, μ(CuKα) = 0.611 mm−1, Dcalc = 1.187 g/cm3, 5060 reflections measured (9.556° ≤ 2Θ ≤ 151.168°), 2839 unique (Rint = 0.0519, Rsigma = 0.0601) which were used in all calculations. The final R1 was 0.1255 (I > 2σ(I)) and wR2 was 0.3621 (all data). The goodness of fit on F2 was 0.865. CCDC number: 2358218.

Crystal data for sophflarine K (6): C15H23N2O4.5 (M = 303.35 g/mol): triclinic, space group P-1 (no. 2), a = 8.7462(8) Å, b = 11.2707(8) Å, c = 16.4281(11) Å, α = 81.481(6)°, β = 78.275(7)°, γ = 68.629(8)°, V = 1471.7(2) Å3, Z = 4, T = 293(2) K, μ(CuKα) = 0.836 mm−1, Dcalc = 1.369 g/cm3, 10,848 reflections measured (5.512° ≤ 2Θ ≤ 147.798°), 5777 unique (Rint = 0.0433, Rsigma = 0.0562) which were used in all calculations. The final R1 was 0.0958 (I > 2σ(I)) and wR2 was 0.2977 (all data). The goodness of fit on F2 was 1.076. CCDC number: 2358219.

3.6 Quantum chemical ECD and ORD calculations

Theoretical ECD calculations of compounds 3–4 and optical rotation calculations of compounds 2 5, and 6 were performed using the Gaussian 09 software package. Conformational analysis was initially conducted using the Sybyl-X 2.1.1 program with an energy cutoff of 10 kcal/mol relative to the global minimum. These conformers were further optimized at the DFT/B3LYP/6-31G(d,p) level. For ORD calculations, single-point optical rotation calculations were performed at the B3LYP level using different basis sets, including 6-31G(d), cc-pVDZ, and 6–311 +  + G(2d,p), based on the optimized conformers. The final optical rotation values reported in Table 3 represent Boltzmann-weighted averages derived from the corresponding conformer populations. ECD spectra were simulated from the first 50 singlet electronic transitions using Gaussian band shapes with a standard deviation of 0.33 eV [38]. Additional experimental details can be found in our previous report [6]. Optimized Cartesian coordinates, thermodynamic data, and conformer populations for compounds 2–6 are provided in the Supporting Information.

3.7 CuSO 4-induced inflammation and drug treatments

CuSO4-induced inflammation assay was performed based on our previous report. Briefly, 3 days post fertilization (dpf), healthy larvae were immersed in 20 μM CuSO4 solution containing sophflarines F− K (50 μM). The behavior of fluorescent neutrophils and macrophages were observed after the treatment for 2 h. Photographs of the inflammatory neutrophil migration results were taken with a microscope (MVX10, Olympus, Japan). Dexamethasone (Dex) group was used for positive control.

3.8 Antibiotic susceptibility and checkerboard assays

According to the Clinical and Laboratory Standards Institute (CLSI) guidelines, the minimum inhibitory concentrations (MICs) of the test compounds and colistin against E. coli ATCC 25922 (conventional QC strain) and E. coli BW25113-mcr-1 (plasmid-mediated colistin-resistant strain) were determined using the broth microdilution method. MIC was defined as the lowest drug concentration that completely inhibited visible bacterial growth. To evaluate compound-colistin interactions, fractional inhibitory concentration (FIC) indices were measured via checkerboard assay. Colistin underwent twofold serial dilution along the x-axis while compounds were diluted along the y-axis, generating a matrix of wells containing combinatorial concentrations. Bacterial suspensions of target strains were inoculated into each well (final density: 5 × 105 CFU mL−1). Following 18 h incubation at 37 °C, optical density was measured at 600 nm. The FIC index (FICI) is calculated by dividing the MIC of the drug combination by the MIC of the individual drugs. FIC indices were calculated as: synergism (FICI ≤ 0.5); additive (0.5 < FICI ≤ 1); indifferent (1 < FICI ≤ 2); and antagonism (FICI > 4).

$$FICI = \frac^ }}^ }} + \frac^ }}^ }}$$

3.9 Statistical analysis

Experiments were conducted in triplicate. Data are represented as means ± SEMs or mean ± SD. The statistical significance was evaluated using either a Student's t-test or one-way ANOVA for multiple comparisons, conducted via GraphPad Pro software. A P-value < 0.05 was set as the threshold for statistical significance.