DHA was obtained from the Kunming Pharmaceutical Group (Chongqing, China, batch number C00220160402). The compound had purity ≥ 99%. Methanol, acetonitrile, and formic acid were HPLC grade and purchased from Fisher (Geel, Belgium). Other chemicals in this work were of analytical grade and purchased from Beijing Chemical Works (Beijing, China). Silicone filler, silica gel GF₂₅₄ thin layer plates, and silica gel GF₂₅₄ thin layer preparation plates were purchased from Qingdao Marine Chemical Group Corporation (Qingdao, China). Silica gel (200–300 mesh, Qingdao Marine Chemical Group Corporation, Qingdao, China), Silica Flash Column 330 g, and Chormatorex (FujiSilysia Chemical, Japan) were used for column chromatography. Water was prepared using a Milli-Q system operating at 18.2 MΩ (Millipore, Bedford, MA, United States). All other chemicals used were purchased from Fisher Scientific or Beijing Chemical Works, and were of the highest purity available.

The UPLC-ESI-Q-TOF-MS^E system consisted of a Waters ACQUITY I-class UPLC and Xevo G2-XS Q-TOF Mass Spectrometer (Waters, Manchester, United Kingdom). ¹H (600 MHz), ¹³C (150 MHz), and NMR spectra were recorded on a Bruker AV 600 spectrometer (Bruker Corporation, Fallanden, Switzerland) with TMS as an internal reference. X-ray diffraction experiments were performed using a Bruker D8 venture diffractometer with Cu Kα radiation. (Crystals of metabolites were obtained from EtOAc and acetone by slow evaporation of solvent at room temperature.)

C. elegans CICC 40250 (MT1) was obtained from the China Center of Industrial Culture Collection (Beijing, China). Culture and biotransformation experiments were conducted in a medium composed of 20 g Sabouraud dextrose broth (Oxoid, Basingstoke, United Kingdom), 10 g peptone (Solarbio, Beijing, China), 15 g sucrose (Solarbio, Beijing, China), and 1,000 ml deionized water.

Two-stage fermentation was employed in this study (Zhan et al., 2017). The substrate was dissolved in MeOH at a concentration of 25 mg/ml, and 2 ml of the solution was added into each flask after the second fermentation stage, producing a DHA concentration of 0.5 mg/ml. Cultures were incubated at 28°C and shaken at 180 rpm for 14 days. Mycelia and broth were then separated by filtration. The filtrate was extracted with EtOAc (1:1, v/v) three times, and the extract was evaporated under vacuum to produce a brown residue.

The UPLC-ESI-QTOF-MS^E system consisted of a Waters ACQUITY I-class UPLC and Xevo G2-XS Q-TOF mass spectrometer equipped with electrospray ionization (ESI) source. Chromatographic separation was achieved using an ACQUITY UPLC BEH C18 (2.1 mm × 100 mm, 1.7 µm, Waters). The mobile phase consisted of solvent A (H₂O containing 0.1% formic acid, v/v) and solvent B (acetonitrile containing 0.1% formic acid, v/v). The gradient program for biosamples included three segments: 5–100% B from 0 to 15 min; followed by 100–5% B from 15 to 17 min; and a postrun of 3 min for column equilibration. The flow rate was 0.4 ml/min, and the temperature was 30°C throughout the analysis.

The MS was operated in positive ionization mode across a scan range of m/z 50 to 1,000, with a scan time of 0.2 s. Source parameters: source temperature 150°C, cone gas 50 L/h, desolations temperature 450°C, and desolation gas flow 800 L/h. Argon (99.95%) was used for collision-induced dissociation, and N₂ was used as the drift gas. The low collision energy was set to 6 eV, and the high collision energy was ramped from 12 to 25 eV. MSE analysis was employed for simultaneous acquisition of the exact mass of small molecules at high and low collision energies.

All data processing was performed using UNIFI 1.9 (Waters, Manchester, United Kingdom). Components were identified using the following 3D peak detection features: low-energy limits of 150 and high-energy limits of 20, isotope clustering, and high-to-low energy association within a 0.5 fraction of the chromatographic and drift peak width, with a mass accuracy of ±2 mDa. The maximum number of allowed fragment ions per match was set to 10.

Scale-up fermentations were also carried out according to the standard two-stage procedure. Under the same temperature-controlled shaking conditions, each 2000 ml flask contained 500 ml Sabouraud glucose liquid medium, and DHA (6 g) was dissolved in MeOH at a concentration of 25 mg/ml and evenly distributed in all flasks. After incubation for 14 days, mycelia and broth were separated by filtration. The filtrate was extracted with EtOAc (1:1, v/v) three times, and the extract was evaporated under vacuum to produce a brown residue. The EtOAc extract was fractionated by a Silica Flash Column using a gradient solvent system of petroleum ether-EtOAc to afford 43 fractions (Fr. 1-Fr. 43) according to TLC analysis.

Fr. 15 (0.5 g) was subjected to a Chromatorex column and eluted with a gradient solvent system of petroleum ether-EtOAc to afford compound M1 (100 mg). Fr. 30 was separated by recrystallization to obtain compound M9 (20 mg). Fr. 40 (1 g) was separated over a CHORMATOREX column and eluted with a gradient solvent system of petroleum ether-acetone to give 14 subfractions. Fr. 40-2 was subjected to thin layer preparation plates to obtain compound M8 (80 mg).

Compounds M1, M8, and M9 were diluted in DMSO and examined for antimalarial activity in vitro against the Pf. 3D7 clone. Artemisinin was used as a positive control drug. Inhibition of plasmodium proliferation by different concentrations of transformed products was determined according to a certain formula, and the in vitro antimalarial activity of the drug was evaluated with the IC₅₀.

The microbial transformation products were identified with UNIFI platform based on chromatographic and mass spectral characteristics provided by reference substances DHA, such as retention time, exact mass, quasi-molecular ions, in-source fragments, characteristic fragments. A total of nine products were found from C. elegans 40250, classified into hydroxylated DHAs (DHA + O), hydroxylated and dehydrogenated DHAs (DHA + O-H2), hydroxylated and dehydrated DHAs (DHA + O-H2-O), deoxygenated DHA (DHA-O), and dehydrated DHA (DHA-H2-O). All identified products were listed in Table 1. Extracted ion chromatograms and MS^E spectra of DHA biotransformation products were shown in Figure 1. Also extracted ion chromatograms and MS^E spectra of its products were shown in Figures 2–10, respectively. The response strength of nine products, which could partly estimate the transform yield of compounds, is shown in Supplementary Figure S1.

DHA, the reference substance, was eluted at 7.27 min. Its molecular ion [M + Na]⁺ was observed at m/z 307.1509. The fragment ions at m/z 267.1590 and 249.1485 were generated by the successive loss of water from m/z 307.1509. The fragment ions at m/z 221.1538 represented the loss of HCOOH from m/z 267.1590. The fragment ions at m/z 203.1432 resulted from the loss of water from m/z 221.1538. Extracted ion chromatograms and mass spectrometry (MS^E) spectra of DHA were shown in Figure 1.

Hydroxylated DHAs (DHA + O, M1-M4) were detected at 3.89, 4.31, 4.97, and 5.14 min. Their molecular ions [M + Na]⁺ were observed at m/z 323.1457, 323.1461, 323.1455, and 323.1450, respectively. They exhibited the same fragment ions. The fragment ions at m/z 283, 265, and 247 were generated by the successive loss of water from m/z 323. The fragment ions at m/z 219 represented the loss of HCOOH from m/z 265. The fragment ions at m/z 201 resulted from the loss of water from m/z 219. Extracted ion chromatograms and mass spectrometry (MS^E) spectra of M1-M4 were shown in Figures 2–5, respectively.

Hydroxylated and dehydrogenated DHA (DHA + O-H2, M5) was detected at 4.17 min. Its molecular ion [M + Na]⁺ was observed at m/z 321.1294. The fragment ions at m/z 249 and 231 were generated by the successive loss of water from m/z 267. The fragment ions at m/z 221 represented the loss of HCOOH from m/z 267 or the loss of CO from m/z 249. The fragment ions at m/z 203 resulted from the loss of water from m/z 221. Extracted ion chromatograms and mass spectrometry (MS^E) spectra of M5 were shown in Figure 6.

Hydroxylated and dehydrated DHAs (DHA + O-H2O, M6-M7) were detected at 3.41 and 4.42 min, respectively. The molecular ions [M + H]⁺ were observed at m/z 283.1554. Consistent with DHA, these products showed a series of fragment ions resulting from the loss of H2O, CO, or HCOOH, such as m/z 265, 247, 237, 219, and 201. The fragment ions at m/z 265 and 247 were generated by the successive loss of water from m/z 283. The fragment ions at m/z 237 represented the loss of CO from m/z 265. Extracted ion chromatograms and mass spectrometry (MS^E) spectra of M6 and M7 were shown in Figures 6–8, respectively.

Deoxygenated DHA (DHA-O, M8) and dehydrated DHA (DHA-H2O, M9) were detected at 7.16 and 8.65 min, respectively. Their molecular ions [M + Na]⁺ were observed at m/z 291.1564 and m/z 289.1405, respectively. They showed similar fragment ions except with a 2 Da (a hydrogen atom) mass shift, such as m/z 251, m/z 249, m/z 233, m/z 231, m/z 205, m/z 203, m/z 187, and m/z 185. The fragment ions at m/z 251 and m/z 249 were generated by the successive loss of water from m/z 291 and m/z 289, respectively. The fragment ions at m/z 205 and m/z 203 represented the loss of CO from m/z 233 and m/z 231, respectively. Their ions exhibited the same mechanism resulting from the loss of CO or H2O. Extracted ion chromatograms and mass spectrometry (MS^E) spectra of M8 and M9 were shown in Figures 9 and 10, respectively.

Furthermore, as a semiquantitative comparative chromatography analysis, the relative contents of modified dihydroartemisinins (M1–M9) were evaluated by mass spectral response abundance data. As a result, M1 had the highest content of all compounds, followed by M8 and M9. And the numerical results of response provided the guidance for further isolation. The histogram of the response of all DHA products is shown in Supplementary Figure S1.

Compound M1 colorless needles (from EtOAc and acetone). ESI-HRMS calcd. for [M + Na]⁺, C₁₅H₂₄O₆Na, 323.1471; found m/z 323.1457. ¹H and ¹³C NMR data are shown in Table 2. The HSQC spectrum showed correlations of H-12 (δ 5.39) with C-12 (δ 90.4), C10-OH (δ 4.64) with C-10 (δ 93.7), and H-7 (δ 2.96) with C-7 (δ 72.6), which supported compound M1 as a hydroxylation product. The HMBC spectrum showed correlations of C10-OH (δ 4.64) with C-8a (δ 43.9), C-7 (δ 72.6) and C-8 (δ 31.2), along with the correlations of C-7 (δ 72.6) with H-14 (δ 0.94) and H-15 (δ 1.28). These spectral dates suggested that the hydroxylation occurred at C-7. Moreover, the absolute configuration of the hydroxylated group in the stereogenic carbon center was determined to be β-OH by single-crystal X-ray diffraction analysis (Scheme 2). A single crystal was obtained from EtOAc and acetone by slow evaporation of the solvent at room temperature. A colorless block-shaped crystal with dimensions of 0.20 × 0.15 × 0.10 mm³ was mounted. Data were collected using a Bruker APEX-II CCD diffractometer operating at T = 150.0 K and measured and scans using CuKa radiation. The maximum resolution achieved was = 72.120° (0.81 Å). The final completeness is 98.90% out to 72.120° in. Crystal data: C₁₅H₂₄O_6, Mr = 300.34, triclinic, P1, a = 9.6563 (6) Å, b = 9.9540 (6) Å, c = 15.6044 (9) Å, α = 93.832 (3)°, β = 90.860 (3)°, γ = 102.948 (3)°, V = 1,457.78 (15) Å3, Z = 4, Z' = 4, (CuKα) = 0.875, 28521 reflections measured, 10327 unique (Rint = 0.0363), which were used in all calculations. The final wR2 was 0.1013 (all data), and R1 was 0.0383 (I ≥ 2 (I)). Crystallographic data of compound M1 have been deposited to CCDC (http://www.ccdc.cam.ac.uk, No. CCDC 2039245). Thus, the structure of compound M1 was elucidated as 7β-hydroxydihydroartemisinin.

Compound M8 white powder (EtOAc). ESI-HRMS calcd. for [M + Na]⁺, C₁₅H₂₄O₄Na, 291.1572; found m/z 291.1564. ¹H and ¹³C NMR data are shown in Table 2. The data supported the compound M8 was 1-deoxydihydroartemisinin (Han and Lee, 2017).

Compound M9 white powder (EtOAc). ESI-HRMS calcd. for [M + H]⁺, C₁₅H₂₂O₄H, 267.1596; found m/z 267.1589. HR-ESI-MS: [M + H]⁺ at m/z at 267.1589 (calcd. for C₁₅H₂₂O₄). ¹H and ¹³C NMR data are shown in Table 2. The data supported the compound M9 was 1-deoxyartemisinin (Lee et al., 1989).

Based on the analysis results of DHA metabolites in vivo, the three products from C. elegans 40250 were compared with in vivo metabolites by UPLC-ESI-Q-TOF-MS^E. The retention time and the mass behavior, such as exact mass, quasi-molecular ions, in-source fragments, and characteristic fragments, are compared to confirm the consistency. The results revealed that M1, M8, and M9 were the common metabolites of DHA in erythrocyte. The extracted ion chromatograms and mass spectrometry (MS^E) spectra of products M1, M8, and M9 and the metabolites of DHA in red blood cells were shown in Figure 11.

The positive control drug artemisinin exhibited in vitro antimalarial activity against Pf. 3D7 with an IC₅₀ (50% inhibition concentration) value of 11 nM. Compound M1 was indicated to exhibit in vitro antimalarial activity against Pf. 3D7 with an IC₅₀ value of 133 nM. Compound M8 and M9 showed no activity against Pf. 3D7.