Cytotoxic trichothecene derivatives from Trichothecium sp. DWS815

1 Introduction

Trichothecenes, a family of sesquiterpenoids produced by various fungal genera such as Aspergillus, Cephalosporium, Fusarium, Myrothecium, Stachybotrys, Verticimonosporium and Trichothecium [1, 2], are common contaminants of grains [3]. This family is divided into two main groups, simple and macrocyclic trichothecenes [4]. The simple trichothecenes share a core ring structure and exhibit cytotoxicity by inhibiting eukaryotic protein synthesis and disrupting mitochondrial and ribosomal functions [5–8]. The 12,13-epoxytrichothec-9-ene (EPT) is cited as the cause of this toxicity [9], and structural diversity within this family arises from various substitution patterns on the EPT core. Recent years have witnessed the continuous discovery of trichothecenes from the genus of Trichothecium [10–13].

Building upon the long-standing focus on discovering the structurally novel and biologically active compounds from soil fungi [14–17], our previous chemical investigation of Myrothecium verrucaria PA 57 led to the discovery of novel macrocyclic trichothecenes exhibiting cytotoxic activities [18]. In our persistent endeavor to mine the novel types of trichothecene natural products with higher cytotoxic activities against cancer cell lines and lower cytotoxic activities toward normal cell lines, the fungal strain Trichothecium sp. DWS815 was isolated from a soil sample collected in Daweishan National Forest Park, Hunan Province, China. A comprehensive LC–MS analysis of the fungal fermentation extract revealed a metabolite profile abundant in simple trichothecenes, suggesting the strain's potential to produce novel trichothecene derivatives. Chemical investigation of this strain led to the isolation of ten previously undescribed trichothecene derivatives, trichotheciumones A (1) and B (2), trichothecinoside A (3), and trichothecrotocins T–Z (4–10), together with three new natural products (11–13) and thirteen known compounds (14–26) (Fig. 1). The structures and absolute configurations of all new compounds were established through spectroscopic analysis and quantum chemistry calculations of electronic circular dichroism (ECD). This paper details the purification, structure elucidation, and cytotoxicity of these compounds.

Open image in new window

2 Results and discussion

Compound 1 was obtained as a white powder. The molecular formula was determined to be C₁₉H₂₆O₇ based on HRESIMS at m/z 367.1758 [M + H]⁺ (calcd for C₁₉H₂₇O₇, 367.1751), corresponding to seven degrees of unsaturation. The ¹H NMR spectrum (Table 1) of 1 indicated the presence of four methyl groups at δ_H 0.98 (s, H₃-14), 1.02 (s, H₃-13), 2.17 (dd, J = 7.2, 1.8 Hz, H₃-4'), and 2.17 (s, H₃-15); four methylene groups at δ_H 2.13/2.53 (ddd, J = 15.8, 4.9, 3.4 Hz, H_a-2; dd, J = 15.8, 8.0 Hz, H_b-2), 2.34/2.89 (d, J = 15.1 Hz, H_a-6; d, J = 15.1 Hz, H_b-6), 2.44/3.10 (dd, J = 16.6, 2.3 Hz, H_a-9; dd, J = 16.6, 9.1 Hz, H_b-9), and 4.21/4.30 (d, J = 13.9 Hz, H_a-12; d, J = 13.9 Hz, H_b-12); five methines including two cis-olefinic methines at δ_H 5.77 (dq, J = 11.4, 1.8 Hz, H-2') and 6.45 (dq, J = 11.4, 7.2 Hz, H-3') as well as three oxygenated methines at δ_H 4.01 (d, J = 9.1 Hz, H-10), 4.45 (d, J = 4.8 Hz, H-1), and 5.63 (dd, J = 8.0, 3.4 Hz, H-3). The ¹³C NMR (Table 4) and DEPT spectra confirmed the presence of 19 carbon resonances assignable to four methyl groups (δ_C 7.4, 15.6, 20.6, 31.5), four sp³ methylene groups including one downfield oxygenated carbon (δ_C 36.5, 39.4, 42.0, 66.8), three sp³ oxygenated methine carbons (δ_C 72.3, 75.3, 81.7), three non-protonated sp³ carbons including one oxygenated carbon (δ_C 43.1, 53.4, 76.5), two sp² olefinic carbons (δ_C 119.3, 147.9), one ketone carbonyl carbon (δ_C 206.5), and two ester carbonyl carbons (δ_C 164.8, 173.6).

Key ¹H–¹H COSY correlations (H-1/H-2/H-3, H-10/H-9, and H-2'/H-3'/H-4') established three independent spin–spin coupling systems (Fig. 2). A crotonyl group was connected to C-3 by HMBC correlations from H-3 and H-2' to C-1' as well as from H-4' to C-2' and C-3'. Further detailed 1D and 2D spectra analysis and comparison with a known co-isolated compound, trichothecene analogue (15) [19], indicated compound 1 bore a resemblance with the trichothecene central skeleton. The B and C rings in 1 were constructed by HMBC correlations from H-1 to C-10 and from H₃-13 to C-3, C-5 and C-11. An acetonyl group was attached to C-10 in the B ring based on the HMBC correlations from H-10, H-9 and H₃-15 to the carbonyl carbon C-8 (δ_C 206.5) and from H₃-15 to C-9 and ¹H–¹H COSY correlation between H-10 and H-9. Moreover, HMBC correlations from H₂-12 to C-11 and C-7, from H₃-14 to C-4, C-10 and C-6, together with a long-range (⁴J_CH) correlation from H₃-14 to C-7 confirmed a lactone bridge between C-5 and C-11 across the B ring. The aforementioned planar structure assignment confirmed that compound 1 possessed a 7,8-seco-trichothecene core, representing the first reported isolation of this specific scaffold. NOESY correlations of H-3/H-10/H₃-14, as well as H-1/H_a-12 and H₃-13/H_b-12, indicated that H-3, H-10, and H₃-14 were α-oriented, while H-1, H₃-13, and OH-11 resided on the opposite face. This relative configuration of 1 aligned with the biosynthetic origin of trichothecene sesquiterpenoids. The absolute configuration was verified based on the agreement between the experimental and calculated ECD spectra. Finally, the absolute configuration was determined to be 1R,3R,4S,5R,10R,11S (Fig. 4). Thus, 1 was established to be trichotheciumone A, a 7,8-seco-trichothecene.

Open image in new window

Compound 2 was obtained as a white powder with the molecular formula of C₁₈H₂₄O₇, determined by HRESIMS at m/z 353.1599 [M + H]⁺ (calcd for C₁₈H₂₅O₇, 353.1595) and corresponding to seven degrees of unsaturation. The NMR spectrum for 2 (Tables 1 and 4) showed similarities to those of 1. The key structural difference was the absence of a methylene unit in 2 at the bridged lactone moiety, suggesting it to be a previously undescribed 7,8-seco-7-nor-trichothecene. The B and C rings, along with the crotonyl group and the acetonyl group were assigned by comparison of the key NMR signals (¹H–¹H COSY, HSQC, and HMBC correlations) with those of 1, as depicted in Fig. 2. The bridged lactone moiety was assigned between C-5 and C-10 across the B ring according to the critical HMBC correlations from H-9, H-11, and H₃-13 to C-6 and from H₂-11 to C-10 and C-6. Moreover, NOESY correlations of H₃-12/H-3/H-9/H₃-13, and H₃-12/H_a-11, and H-1/H_b-11 supported the relative configuration shown in Fig. 3. The relative configuration of 2 was consistent with the biosynthetic origin of trichothecene sesquiterpenoids. By comparing the experimental and calculated ECD spectra (Fig. 4), the absolute configuration of 2 was defined as 1R,3R,4S,5R,9R,10S. Consequently, 2 was assigned as depicted and named trichotheciumone B, a 7,8-seco-7-nor-trichothecene.

Open image in new window

Open image in new window

Compound 3 was obtained as a white powder, displaying an exact mass of m/z 499.2519 [M + H]⁺ (calcd for C₂₅H₃₉O₁₀, 499.2538) in HRESIMS, corresponding to a molecular formula of C₂₅H₃₈O₁₀ with eight degrees of unsaturation. Analysis of the NMR data in Tables 1 and 4 indicated that 3 was deemed as a derivative of the trichothecenes. The ¹³C NMR spectrum displayed six characteristic signals indicative of a glycosyl unit (δ_C 61.8, 67.5, 70.9, 74.1, 75.6, 102.1). The mannose unit was further evidenced by comparison with the literature data (δ_C 61.2, 66.6, 71.1, 74.1, 75.9, 100.1) and δ_H 3.18 (brs), 3.53 (brs), 3.86 (m), 3.87 (brs), 3.97 (brs), and 4.44 (d, J = 0.7 Hz) in CDCl₃ [20]. The mannosidic linkage of 1''-O-7 was revealed by a key HMBC correlation from H-1'' to C-7. The β-configuration of the mannose unit was confirmed by the following evidences. Two broad singlets of the anomeric proton H-1'' and H-2'', as well as the critical and obvious NOESY correlations of H-1'' with H-3'' and H-5'', provided compelling evidence for axial situation of anomeric proton H-1'', H-3'', and H-5'' and equatorial situation of H-2''. Based on the NOESY correlations of H₃-13/H-3/H-6/H₃-14 and analysis in light of biosynthetic origin, the relative configuration of chiral centers C-1, C-3, C-4, C-5, C-10, and C-11 was established. Meanwhile, the overall relative configuration of 3 was determined by interpretation of key NOESY correlations and GIAO ¹³C NMR calculations with the STS protocol performed for eight isomers (3a–3h) (Table S4). The absolute configuration of 3, including the D-mannose moiety, was determined by ECD calculations using the time-dependent density functional theory (TDDFT) method. As shown in Fig. 4, the calculated ECD curve of 3a (1R,3R,4S,5R,7S,8R,10R,11S-β-D-mannopyranoside) matched well with the experimental ECD curve. Therefore, the absolute configuration of 3 was assigned as 1R,3R,4S,5R,7S,8R,10R,11S-β-D-mannopyranoside, and it was named trichothecinoside A.

Compound 4 was obtained as a white powder, having the molecular formula C₁₆H₂₄O₅ based on the protonated ion peak at m/z 297.1700 [M + H]⁺ (calcd for C₁₆H₂₅O₅, 297.1697), corresponding to five degrees of unsaturation. Analysis of 1D data of 4 (Tables 1 and 4) revealed close resemblance to the co-isolated new natural product isocrotocol B (11) [21], indicating a shared trichothecene backbone. A key structural distinction was the presence of an oxygenated methyl group at C-7 in 4. Further analysis of the 2D data confirmed that the remaining structure of 4 were the same as that of 11. The absolute configuration was defined to be 1R,3R,4S,5R,6R,7R,10R,11S by comparison of the experimental ECD spectrum with calculated one (Fig. 4). Therefore, 4 was established as depicted and named trichothecrotocin T.

Compound 5 was obtained as a white powder, exhibiting the molecular formula C₂₇H₃₂O₇, based on the protonated molecule peak at m/z 469.2211 [M + H]⁺ (calcd for C ₂₇H₃₃O₇, 469.2221), implying twelve degrees of unsaturation. Analysis of the 1D data (Tables 2 and 4) indicated a structure similar to that of co-isolated compound trichothecrotocin O (16), differing by the replacement of a methoxy group in 16 with a phenylacetoxy moiety at the C-7 position in 5. The proposed structure was supported by ¹H–¹H COSY correlations for H-4''/H-5''/H-6''/H-7''/H-8'', as well as HMBC correlations from H-7 to C-1'' and from H₂-2'' to C-1'', C-3'', C-4'', and C-8''. Key NOESY correlations of H₃-13/H-3/H-10/H₃-14 revealed the same relative configuration as 16, although the stereochemistry at C-7 remained unassigned. To resolve this, GIAO ¹³C NMR calculations with the STS protocol were conducted for two diastereomers 5a and 5b (Table S9). The calculated data for 5a matched the experimental data, thereby establishing the relative configuration. Furthermore, the agreement between experimental ECD spectrum and the calculated spectrum for 5 confirmed absolute configuration as 1R,3R,4S,5R,6R,7R,10R,11S (Fig. 4). Therefore, the structure of 5 was validated and assigned the name trichothecrotocin U.

Compounds 6 and 7 were each obtained as a white powder, presenting the same molecular formula C₁₉H₂₆O₅, based on their HRESIMS data at m/z 335.1851 [M + H]⁺ (calcd for C₁₉H₂₇O₅, 335.1853) and 335.1863 [M + H]⁺ (calcd for C₁₉H₂₇O₅, 335.1853), respectively, corresponding to seven degrees of unsaturation. Analysis of 1D data for 6 and 7 (Tables 2 and 4) indicated close similarity to those of a new co-isolated natural product dihydrotrichothecolone (12), differing from 12 by the presence of the crotonyl moiety. The connectivity of this moiety was confirmed by the ¹H–¹H COSY data of H-2'/H-3'/H-4' and HMBC correlations from H-3 and H-2' to C-1' and from H-4' to C-2' and C-3'. The relative configuration of 6 was deemed as being identical to that of 12 through key NOESY correlations as depicted in Fig. 2. The α-orientation of the C-15 was assigned by comparing experimental data with GIAO ¹³C NMR calculations (STS protocol) performed for two diastereomers 6a and 6b (Table S11). A key distinction between 6 and 7 is the geometry of crotonyl group, which was ascertained as Z in 7, assigned based on the coupling constant (J = 11.4 Hz) between H-2' and H-3'. The relative stereochemistry of 7 was elucidated using the same procedure as for 6. Finally, the absolute configurations of both 6 and 7 were defined to be 1R,3R,4S,5R,8R,10R,11S by comparing their experimental and calculated ECD spectra (Fig. 4). Consequently, compounds 6 and 7 were named trichothecrotocin V and trichothecrotocin W, respectively.

Compound 8 was obtained as a white powder, possessing the molecular formula C₁₉H₂₄O₅ based on the proton adduct at m/z 333.1708 [M + H]⁺ (calcd for C₁₉H₂₅O₅, 333.1697) in its HRESIMS, indicating eight degrees of unsaturation. The 1D (Tables 2 and 4) and 2D data disclosed that the structure of 8 was structurally similar to that of a co-isolated known compound crotocin (22), differing only in the E geometry of the crotonyl group. The geometric distinction between 8 and 22 paralleled that observed between 6 and 7, as supported by the large coupling constant (J = 15.6 Hz) between H-2' and H-3' in 8. NOESY correlations of H₃-13/H-3/H-10/H₃-14/H-6 suggested their co-facial orientation, which facilitated the determination of the relative configuration. The calculated ECD curve of 1R,3R,4S,5R,6R,7S,10R,11S was in good agreement with the experimental data (Fig. 4). Thus, 8 was assigned as depicted and named trichothecrotocin X.

Compounds 9 and 10 were each obtained as a white powder. Compound 9 possessed a molecular formula of C₁₅H₂₀O₃ based on the proton adduct in its HRESIMS at m/z 249.1482 [M + H]⁺ (calcd for C₁₅H₂₁O₃, 249.1485), corresponding to six degrees of unsaturation. The molecular formula for 10 was determined to be C₁₉H₂₄O₅ by HRESIMS based on the sodium adduct at m/z 339.1563 [M + H]⁺ (calcd for C₁₉H₂₄O₅Na, 339.1567), corresponding to eight degrees of unsaturation. Analysis of 1D data for both compounds 9 and 10 (Tables 3 and 4) exhibited close similarity to those of a co-isolated known compound 7-dehydro-8-dehydroxytrichothecinol B (24). For compound 9, the key structural distinction was the absence of a crotonyl group at C-3. Comprehensive analysis of 2D data confirmed that the remaining planar structure and the relative configuration of 9 were identical to those of 24. The calculated ECD curve for isomer 9a (1R,3R,4S,5R,10R,11S) showed excellent agreement with the experimental data (Fig. 4). For compound 10, the key structural distinction was only in the E geometry of the crotonyl group, supported by HMBC correlations from H-3 to C-1' and from H-4' to C-2' and C-3' as well as the large coupling constant (J = 15.5 Hz) between C-2′ and C-3′, confirming crotonyl group with trans-olefinic double bond. NOESY correlations observed for 10 were similar to those of 9, suggesting an identical relative configuration. Furthermore, the experimental ECD spectrum of 10 (Fig. 4) closely matched that of 9 and was consistent with the calculated spectrum for isomer 10a (1R,3R,4S,5R,10R,11S). Therefore, 9 and 10 were named trichothecrotocin Y and trichothecrotocin Z, respectively.

Compounds 11, 12, and epi-trichothecinol B (13) were isolated from this fungus as new natural products for the first time. The absolute configuration of compound 11 was first determined and the ECD spectra of 13 was provided in Figure S143. The known compounds 14–26 were characterized via analysis of their HRESIMS and 1D data and comparison with literature data. They were identified as trichothecrotocin N (14) [12], trichothecene analogue (15) [19], trichothecrotocin O (16) [12], trichothecrotocin Q (17) [12], trichothecolone (18) [22], trichodermol (19) [23], 8-deoxy-trichothecin (20) [22], trichodermin (21) [24], crotocin (22) [21], crotocol (23) [21], 7-dehydro-8-dehydroxytrichothecinol B (24) [25], cyclonerodiol (25) [26], and cyclonerodiol C (26) [27].

Trichotheciumone A (1) and B (2) possess unusual sesquiterpenoid skeletons based on the trichothecene core. Although structurally distinct, compounds 1 and 2 are likely biosynthesized from the same precursor, trichodiene, a key intermediate in trichothecene biosynthesis [4, 28]. A putative biosynthetic pathway for 1 and 2 is proposed in Scheme 1. The pathway suggests that cleavage of the C-7–C-8 bond in trichothecin, followed by oxidation, would yield the crucial intermediate ⅰ, which belongs to the rare 7,8-seco-trichothecene class. A subsequent nucleophilic attack of the carboxylate anion on the C-12 epoxide would form compound 1, representing the first reported trichothecene-type sesquiterpenoid with an opened A-ring. Alternatively, crotocin (22) could be converted to intermediate 15 via nucleophilic ring-opening of the 6,7-epoxide by a hydroxide ion. Further cleavage of the C-7–C-8 bond in 15, followed by oxidation, would lead to intermediate ⅱ. The oxidative decarboxylation of ⅱ would then yield compound 2, which is the first reported example of a rare A-ring-seco-nor-trichothecene-type sesquiterpenoid.

Open image in new window

The cytotoxic activities of the isolated compounds were evaluated against human colon carcinoma HCT116 cell line, mouse breast cancer 4T1 cell line, and human liver cancer MHCC97H cell line, with one strain of human gastric epithelial GES-1 cells serving as a normal cell lines control, using the CCK-8 method. Seven compounds (7, 8, 10, 13, 17, 18, 20, 24) exhibited significant anti-proliferative activity (Fig. 5A, B). Among them, compound 17 was the most potent compound with IC₅₀ of 0.35 μM, 0.52 μM, 0.57 μM against HCT116, 4T1, and MHCC97H cells respectively (Table 5). Cell cycle assays revealed that new compounds 7 and 8 arrested HCT116 cells at the G2/M phase. Treatment with 0, 1, and 2 μM of compound 7 increased the G2/M population compared with the control (control: 5.25%, 1 μM: 17.22%, 2 μM: 32.42%) and decreased G0/G1 population (control: 64.79%, 1 μM: 52.01%, 2 μM: 24.61%) (Fig. 6A, B). Treatment with 0, 1, and 2 μM of compounds 8 increased the G2/M population compared with the control (control: 4.41%, 1 μM: 7.14%, 2 μM: 12.31%) and decreased G0/G1 population (control: 62.78%, 1 μM: 60.54%, 2 μM: 50.59%) (Fig. 6C, D). However, the test compounds exhibited no observable selectivity in inhibiting proliferation between these three cancer cell lines and GES-1 cell line. Structure–activity relationship (SAR) analysis indicated that the configuration of the C-3 crotonyl group influences toxicity. Specifically, changing the crotonyl group at C-3 from the Z-configuration to the E-configuration results in decreased toxicity, a Z-configuration (as in compounds 7 and 24) conferred greater potency than an E-configuration (as in compounds 6 and 10).

Open image in new window

Open image in new window

3 Experimental section

3.1 General experimental procedures

Optical rotations were measured on a Rudolph Research Analytical Autopol IV automatic polarimeter. UV spectra were recorded using a Cary 300 spectrometer (Agilent Technologies, Santa Clara, CA, USA). IR spectra were recorded on a Shimadzu Fourier Transform Infrared Spectrometer using KBr pellets. Circular Dichroism (CD) spectra were obtained on a Chirascan^TM-plus Circular Dichroism spectrometer. High-resolution electrospray ionization mass spectrometry (HRESIMS) spectra were acquired with an Agilent 6500 series Q-TOF mass spectrometer (Agilent Technologies, Singapore). NMR spectra were recorded on Bruker Avance Ⅲ HD 600 MHz and 500 MHz spectrometer with tetramethylsilane (TMS) serving as the internal standard. Column chromatography (CC) was performed using macro-porous adsorbent resin D101 (Tianjin Haoju Resin Technology Co., Ltd., Tianjin, China), silica gel (200–300 mesh, Qingdao Marine Chemical, Qingdao, China), and Sephadex LH-20 (GE Healthcare, Uppsala, Sweden). MCI GEL^TM (Mitsubishi Chemical Corporation, Japan). Semipreparative ZORBAX SB-C18 (Agilent, 5 μm, 9.4 mm × 150 mm), Supersil ODS-B (Elite, 10 μm, 20.0 mm × 250 mm), Supersil ODS2 (Elite, 10 μm, 20.0 mm × 250 mm), Ultimate XB-CN (Welch, 5 μm, 10.0 mm × 250 mm) and preparative SinoPark C18 (Elite, 10 μm, 20.0 mm × 250 mm) were analyzed by an EClassical P3500 prep-HPLC system and a DAD detector (EClassical P3500, Dalian Elite Analytical Instruments Co., Ltd., Dalian, China). The progress of the experiments was monitored by pre-coated silica gel GF254 plates (Qingdao Marine Chemical). Spots were visualized under UV light (254 or 365 nm) or by spraying with 10% H₂SO₄ in EtOH followed by heating. All other chemicals used in this study were of analytical grade.

3.2 Fungal material

The fungal strain Trichothecium sp. DWS815 was isolated from a forest soil sample collected from Daweishan National Forest Park, Hunan Province. The strain was identified as Trichothecium sp. (GenBank Accession No. ON386019.1) based on their ITS sequences and the phylogenetic analysis (Tables S24 and S25). A voucher specimen of this strain has been deposited at the Xiangya School of Pharmaceutical Sciences, Central South University, Changsha, China.

3.3 Fermentation, extraction and isolation

This fungus was initially cultivated on Sabouraud Dextrose Agar (SDA) medium (containing 20 g/L dextrose, 10 g/L peptone, 1 g/L yeast extract powder, 15 g/L agar and water) for 8 days at 25 ℃. Subsequently, the mycelia on agar plugs were cut and transferred to the sterile rice fermentation medium. The medium was prepared with 100 g of rice and 80 mL of water in 500 mL Erlenmeyer flasks. For large-scale fermentation, a total of 14 kg of rice was distributed into 139 flasks and inoculated with this fungus. The fermentation proceeded at 25 ℃ in the dark for 21 days.

The rice was extracted with ethyl acetate at room temperature 10 times, and the solvent was evaporated under vacuum to yield an extract (170 g). The extract was absorbed onto 500 g macro-porous adsorbent resin D101 and eluted stepwise with water, 80% aqueous MeOH, and 100% MeOH. The 80% aqueous MeOH fraction was collected and condensed to yield 26 g residue. The 80% aqueous MeOH fraction was then fractionated using silica gel CC eluted with a step gradient of Petroleum ether–EtOAc (v/v, 100:0, 20:1, 15:1, 10:1, 5:1, 2:1, 1:1) and EtOAc–MeOH (v/v, 100:0, 10:1, 5:1, 1:1, 0:100) to provide nine fractions (A~I).

The Fr.A was subjected to MCI to give five subfractions (Fr.A1~Fr.A5). Fr.A2 was separated using semipreparative Supersil ODS-B (CH₃CN: H₂O, 55:45) to afford compound 24 (t_R = 20.9 min, 22.5 mg) and Fr.A2-1. Fr.A2-1 was fractionated using Supersil ODS-B (MeOH: H₂O, 57:43) to afford compounds 7 (t_R = 36.9 min, 3.5 mg) and 21 (t_R = 40.1 min, 1.9 mg). Fr.A4 was separated by semipreparative ZORBAX SB-C18 (CH₃CN: H₂O, 49:51) to obtain compound 20 (t_R = 23.9 min, 50.2 mg). Fr.B was subjected to Sephadex LH-20 (MeOH) to give five subfractions (Fr.B1~Fr.B5). Fr.B3 was isolated on the preparative SinoPark C18 (MeOH: H₂O, 62:38) to obtain four fractions (Fr.B3-1~Fr.B3-4). Fr.B3-1 was purified using semipreparative Supersil ODS2 (CH₃CN: H₂O, 28:72) to afford compounds 1 (t_R = 21.1 min, 0.3 mg) and 2 (t_R = 29.2 min, 0.6 mg). Fr.B3-4 was purified using semipreparative Supersil ODS-B (MeOH: H₂O, 53:47) to yield compound 17 (t_R = 47.0 min, 24.5 mg). Fr.C was applied to semipreparative ZORBAX SB-C18 with a step gradient of CH₃CN-H₂O (v/v, from 20:100 to 50:50 in 35 min, flow speed: 3 mL/min) to afford compounds 22 (t_R = 19.8 min, 1.6 g), 16 (t_R = 17.7 min, 4.9 mg), 8 (t_R = 19.4 min, 6.8 mg), and 5 (t_R = 32.8 min, 6.58 mg). Fr.D was subjected to Sephadex LH-20 (MeOH) and then was further fractionated using semipreparative Ultimate XB-CN (CH₃CN: H₂O, 19:81) to give 19 (t_R = 14.6 min, 1.1 mg), 18 (t_R = 9.5 min, 1.9 mg), 9 (t_R = 12.3 min, 1.5 mg), 13 (t_R = 28.9 min, 8.4 mg), 26 (t_R = 20.0 min, 3.5 mg), 25 (t_R = 24.9 min, 9.9 mg), and 12 (t_R = 17.0 min, 2.3 mg). Fr.E was subjected to Sephadex LH-20 (MeOH) and then was purified using semipreparative Supersil ODS-B (MeOH: H₂O, 26:74) to afford 14 (t_R = 15.9 min, 1.4 mg), 15 (t_R = 12.2 min, 41.9 mg), 4 (t_R = 10.7 min, 3.09 mg), 23 (t_R = 23.5 min, 3.3 mg), and 11 (t_R = 8.7 min, 1.6 mg). Fr.F was subjected to Sephadex LH-20 (MeOH) and then was purified using semipreparative ZORBAX SB-C18 with a step gradient of CH₃CN–H₂O (v/v, from 5:95 to 21:79 in 100 min, flow speed: 3 mL/min) to yield 3 (t_R = 81.6 min, 1.3 mg). Fr.I was also chromatographed on Sephadex LH-20 (MeOH) and purified using semipreparative Ultimate XB-CN with a step gradient of CH₃CN–H₂O (v/v, from 25:75 to 45:55 in 40 min, flow speed: 3 mL/min) to give compounds 6 (t_R = 29.8 min, 0.5 mg) and 10 (t_R = 35.1 min, 1.9 mg).

Trichotheciumone A (1). White powder; [α]_D²⁵+ 37.5 (c 0.04, MeOH); UV (MeOH) λ_max (log ε): 210 (3.48) nm; IR (KBr) ν_max: 3442, 2979, 2950, 2923, 1714, 1647, 1294, 1178, 1115, 1028, 817 cm⁻¹; CD (c 1.19 mM, MeOH): 205 nm (Δε –4.84), 227 nm (Δε + 4.38), 269 nm (Δε –1.91); for ¹³C NMR (CDCl₃, 150 MHz) and ¹H NMR (CDCl₃, 600 MHz) data, see Tables 1 and 4; HRESIMS (positive) m/z 367.1758 [M + H]⁺ (calcd for C₁₉H₂₇O₇, 367.1751, Δ + 1.9064 ppm).

Trichotheciumone B (2). White powder; [α]_D²⁵ + 264.86 (c 0.037, MeOH); UV (MeOH) λ_max (log ε): 210 (2.88) nm; IR (KBr) ν_max: 3434, 2924, 2855, 2788, 1718, 1578, 1435, 1384, 1180 cm⁻¹; CD (c 1.05 mM, MeOH): 200 nm (Δε + 5.29), 226 nm (Δε –2.83), 270 nm (Δε –0.51); for ¹³C NMR (CDCl₃, 150 MHz) and ¹H NMR (CDCl₃, 600 MHz) data, see Tables 1 and 4; HRESIMS (positive) m/z 353.1599 [M + H]⁺ (calcd for C₁₈H₂₅O₇, 353.1595, Δ + 1.1326 ppm).

Trichothecinoside A (3). White powder; [α]_D²⁵ –2.0 (c 0.05, MeOH); UV (MeOH) λ_max (log ε): 200 (3.17) nm; IR (KBr) ν_max: 3420, 2962, 2935, 2876, 1715, 1649, 1383, 1291, 1184, 1083, 1027, 959 cm⁻¹; CD (c 1.00 mM, MeOH): 217 nm (Δε –7.13); for ¹³C NMR (CDCl₃, 150 MHz) and ¹H NMR (CDCl₃, 600 MHz) data, see Tables 1 and 4; HRESIMS (positive) m/z 499.2519 [M + H]⁺ (calcd for C₂₅H₃₉O₁₀, 499.2538 Δ –3.8057 ppm).

Trichothecrotocin T (4). White powder; [α]_D²⁵ –12.5 (c 0.037, MeOH); UV (MeOH) λ_max (log ε): 200 (2.82) nm; IR (KBr) ν_max: 3459, 3371, 2981, 2932, 2859, 1628, 1457, 1135, 1125, 1098, 1051 cm⁻¹; CD (c 1.35 mM, MeOH): 200 nm (Δε –28.13); for ¹³C NMR (CDCl₃, 150 MHz) and ¹H NMR (CDCl₃, 600 MHz) data, see Tables 1 and 4; HRESIMS (positive) m/z 297.1700 [M + H]⁺ (calcd for C₁₆H₂₅O₅, 297.1697, Δ + 1.0009 ppm).

Trichothecrotocin U (5). White powder; [α]_D²⁵ –10.9 (c 0.064, MeOH); UV (MeOH) λ_max (log ε): 200 (3.21) nm; IR (KBr) ν_max: 3434, 2978, 2946, 2922, 2859, 1737, 1714, 1647, 1436, 1248, 1179, 1133, 1049 cm⁻¹; CD (c 1.37 mM, MeOH): 200 nm (Δε –30.89), 210 nm (Δε + 0.19), 225 nm (Δε –8.52); for ¹³C NMR (CDCl₃, 150 MHz) and ¹H NMR (CDCl₃, 600 MHz) data, see Tables 2 and 4; HRESIMS (positive) m/z 469.2211 [M + H]⁺ (calcd for C ₂₇H₃₃O₇, 469.2221, Δ –2.1312 ppm).

Trichothecrotocin V (6). White powder; [α]_D²⁵+ 12.5 (c 0.008, MeOH); UV (MeOH) λ_max (log ε): 210 (2.94) nm; IR (KBr) ν_max: 2971, 2928, 2855, 1717, 1576, 1445, 1384, 1183, 1081, 971 cm⁻¹; CD (c 1.50 mM, MeOH): 208 nm (Δε –11.61), 291 nm (Δε + 0.92); for ¹³C NMR (CDCl₃, 150 MHz) and ¹H NMR (CDCl₃, 600 MHz) data, see Tables 2 and 4; HRESIMS (positive) m/z 335.1851 [M + H]⁺ (calcd for C₁₉H₂₇O₅, 335.1853, Δ –0.5967 ppm).

Trichothecrotocin W (7). white powder; [α]_D²⁵ –6.98 (c 0.043, MeOH); UV (MeOH) λ_max (log ε): 210 (3.08) nm; IR (KBr) ν_max: 2959, 2928, 1716, 1647, 1466, 1366, 1287, 1183, 1081, 966, 816 cm⁻¹; CD (c 1.29 mM, MeOH): 213 nm (Δε –4.32), 293 nm (Δε + 1.46); for ¹³C NMR (CDCl₃, 150 MHz) and ¹H NMR (CDCl₃, 600 MHz) data, see Tables 2 and 4; HRESIMS (positive) m/z 335.1863 [M + H]⁺ (calcd for C₁₉H₂₇O₅, 335.1853, Δ + 2.9834 ppm).

Trichothecrotocin X (8). White powder; [α]_D²⁵ –16.67 (c 0.024, MeOH); UV (MeOH) λ_max (log ε): 208 (3.29) nm; IR (KBr) ν_max: 2965, 2925, 2857, 1716, 1602, 1384, 1185, 1083, 969 cm⁻¹; CD (c 0.72 mM, MeOH): 203 nm (Δε –13.10), 230 nm (Δε + 0.50); for ¹³C NMR (CDCl₃, 125 MHz) and ¹H NMR (CDCl₃, 500 MHz) data, see Tables 2 and 4; HRESIMS (positive) m/z 333.1708 [M + H]⁺ (calcd for C₁₉H₂₅O₅, 333.1697, Δ + 3.3016 ppm).

Trichothecrotocin Y (9). White powder; [α]_D²⁵ –51.35 (c 0.037, MeOH); UV (MeOH) λ_max (log ε): 200 (2.60) nm; IR (KBr) ν_max: 3452, 2965, 2923, 2852, 1664, 1623, 1593, 1447, 1422, 1281, 1185, 1074, 955 cm⁻¹; CD (c 1.49 mM, MeOH): 256 nm (Δε –43.19); for ¹³C NMR (CDCl₃, 125 MHz) and ¹H NMR (CDCl₃, 500 MHz) data, see Tables 3 and 4; HRESIMS (positive) m/z 249.1482 [M + H]⁺ (calcd for C₁₅H₂₁O₃, 249.1485, Δ –1.2041 ppm).

Trichothecrotocin Z (10). White powder; [α]_D²⁵ –63.16 (c 0.019, MeOH); UV (MeOH) λ_max (log ε): 210 (3.15) nm; IR (KBr) ν_max: 2917, 2852, 1732, 1717, 1606, 1590, 1384, 1189, 1101, 1082, 1064, 1028, 957 cm⁻¹; CD (c 0.60 mM, MeOH): 226 nm (Δε –13.12); for ¹³C NMR (dimethyl sulfoxide-d₆, 150 MHz) and ¹H NMR (dimethyl sulfoxide-d₆, 600 MHz) data, see Tables 3 and 4; HRESIMS (positive) m/z 339.1563 [M + H]⁺ (calcd for C₁₉H₂₄O₅Na, 339.1567, Δ –1.1793 ppm).

Isocrotocol B (11). White powder; [α]_D²⁵ –18.0 (c 0.05, MeOH); UV (MeOH) λ_max (log ε): 200 (2.71) nm; IR (KBr) ν_max: 3507, 3435, 3332, 2978, 2939, 2876, 1637, 1452, 1321, 1135, 1100, 1055, 973, 898, 817 cm⁻¹; CD (c 1.77 mM, MeOH): 221 nm (Δε + 1.42); for ¹³C NMR (CDCl₃, 150 MHz) and ¹H NMR (CDCl₃, 600 MHz) data, see Tables 3 and 4; HRESIMS (positive) m/z 283.1542 [M + H]⁺ (calcd for C₁₅H₂₃O₄, 283.1540, Δ + 0.7063 ppm).

Trichothecolone (12). White powder; [α]_D²⁵ –5.22 (c 0.23, MeOH); UV (MeOH) λ_max (log ε): 200 (2.02) nm; IR (KBr) ν_max: 3502, 2982, 2966, 2931, 1715, 1458, 1280, 1081, 957 cm⁻¹; CD (c 8.64 mM, MeOH): 200 nm (Δε –71.20), 292 nm (Δε + 15.78); for ¹³C NMR (methanol-d₄, 150 MHz) and ¹H NMR (methanol-d₄, 600 MHz) data, see Tables 3 and 4; HRESIMS (positive) m/z 267.1585 [M + H]⁺ (calcd for C₁₅H₂₃O₄, 267.1591 Δ –2.2458 ppm).

Epi-trichothecinol B (13). White powder; [α]_D²⁵+ 87.5 (c 0.04, MeOH); UV (MeOH) λ_max (log ε): 200 (2.71) nm; IR (KBr) ν_max: 3468, 2985, 2962, 2938, 1712, 1643, 1434, 1376, 1313, 1245, 1175, 1078, 1034, 996, 972, 962 cm⁻¹; CD (c 1.77 mM, MeOH): 221 nm (Δε + 1.42); for ¹³C NMR (CDCl₃, 150 MHz) and ¹H NMR (CDCl₃, 600 MHz) data, see Tables 3 and 4; HRESIMS (positive) m/z 335.1856 [M + H]⁺ (calcd for C₁₉H₂₇O₅, 335.1853, Δ + 0.8950 ppm).

3.4 Quantum chemical calculations

Conformer–rotamer ensemble sampling tool (CREST) was used to search the conformational space of compounds at the GFN0 level of theory in the gas phase [29, 30], followed by optimization at the GFN2-xTB level to avoid the errors caused by poor initial geometry [31], with a 4 kcal/mol energy window to remove high-energy conformers. Optimization and frequency calculation of each conformer were performed at the r²-SCAN-3c level of theory in the conductor-like polarizable continuum model (CPCM) to ensure they were true local minima on the potential energy surface [32, 33]. Under STS protocol [34], DFT GIAO ¹³C NMR calculation was calculated at the ωB97x-D/6-31G* (IEFPCM, chloroform) level or ωB97x-D/6-31G* (IEFPCM, methanol) level of theory. Time-dependent density-functional theory (TDDFT) ECD calculations were performed at the B3LYP/TZVP (IEFPCM, methanol) level of theory with the consideration of solvent effects. The calculated shielding tensors and ECD curves of conformers were Boltzmann averaged based on Gibbs free energy. In the TDDFT-ECD calculations, 40 excited states were calculated for each conformer. The ECD curves of all conformers were Boltzmann averaged based on Gibbs free energy calculated at the r²SCAN-3c level of theory and generated by Multiwfn [35, 36]. All DFT calculations were performed by Gaussian 16 or ORCA 6.0.0 software packages [37–39]. The DFT optimized geometry data, relative energies, and conformational population of all calculated structures were provided in the supplemental material. The 3D molecular structures were rendered using VMD (1.9.3) [40]. Details are provided in the Supplementary Information.

3.5 Cell lines and culture conditions

The cells of GES-1, HCT116, 4T1 and MHCC97H were cultivated in DMEM medium enriched with 10% fetal bovine serum and 1% penicillin–streptomycin. Cultures were maintained at 37 ℃ in a humidified environment with 5% CO₂.

3.6 Cell viability analysis

Cell viability was determined with CCK-8 method [18]. All cell lines were plated into 96-well plates at a concentration of 10,000 cells per well and incubated for an additional 24 h. Then, the cells were treated with concentrations 5 μM and 50 μM of the test compounds and positive control (doxorubicin) for a duration of 48 h. In addition to the experimental groups, blank controls (wells with CCK-8 solution but no cells) and negative controls (wells with vehicle) were included. After treatment, 10 μL of CCK-8 reagent was added to each well and incubated for 1 h. The optical density at 450 nm was then measured on a microplate reader (Feyond-A300, allsheng, Hangzhou, China). The IC₅₀ value represents the half of maximal inhibitory concentration. The cells of HCT116, 4T1, and MHCC97H were treated with a concentration gradient of selected compounds (7, 8, 10, 13, 17, 18, 20, and 24) to obtain IC₅₀ values, using either the range of (0.04 μM–10 μM) or (0.5 μM–50 μM).

3.7 Cell cycle assays

The cell cycle distribution was appraised, using the Propidium Iodide (PI) Flow Cytometry Kit, followed by the flow Cytometry analysis [12]. Briefly, HCT116 cells were planted on 6-well plates and incubated overnight. Next, the cells were treated with 7 and 8 at concentrations of 0, 1 and 2 μM, and incubated for 24 h in triplicate manner. Cells were harvested and washed with cold phosphate buffered saline (PBS), and fixed overnight with ice-cold 70% ethanol. Then, the ethanol was removed, and the cells were rinsed with PBS and stained with the DNA fluorochrome PI at room temperature in the dark for 20 min. CYTEK^TM NL-3000 flow cytometer was then used to test samples.

4 Conclusion

In summary, ten previously undescribed trichothecenes, three new natural products and thirteen known compounds were identified by chemical investigation of the EtOAc extract from rice cultures of the fungus Trichothecium sp. DWS815. Notably, trichotheciumones A (1) and B (2) are novel trichothecene skeleton characterized by the cleavage of the A ring found in classical trichothecenes and the formation of a lactone bridge across the B ring. Trichothecinoside A (3), trichothecrotocin V (6), and trichothecrotocin W (7) are unusual natural trichothecene characterized by a hydrogenated EPT moiety. The isolated compounds were evaluated for cytotoxic activity against three human cancer cell lines (HCT116, 4T1, MHCC97H) and one normal cell line (GES-1). Among them, compound 17 was the most potent compound with IC₅₀ of 0.35 μM, 0.52 μM, 0.57 μM against HCT116, 4T1, and MHCC97H cells respectively and new compounds 7 and 8 were found to induce G2/M phase cell cycle arrest in HCT116 colorectal cancer cells, thereby inhibiting cell proliferation. This work not only expands the structural diversity of trichothecene derivatives but also provides a foundation for future studies on structure–activity relationships and mode of action.

Notes

Acknowledgements

The authors thank the Institute for Advanced Study, Central South University for NMR measurement and Center for Medical Research and Innovation, the First Hospital of Hunan University of Chinese Medicine for MS measurement.

Author contributions

Peng-Ju Xu: Writing–original draft, Methodology, Investigation. Ai-Lin Liang: Methodology. Wen-Yu Lu: Methodology. Hong-Ping Long: Methodology. Qing-Hui Xiao: Methodology. Qi-An Chen: Methodology. Meng-Lan Hu: Methodology. Ngoc Nhu Thao Nguyen: Methodology. Shao Liu: Methodology. Jing Li: Writing – review & editing, Investigation, Funding acquisition. Wen-Xuan Wang: Writing–review & editing, Funding acquisition, Conceptualization. All authors read and approved the final manuscript.

Funding

The work was financially supported by National Natural Science Foundation of China (No. 82173713), Natural Science Foundation of Hunan Province (No. 2024JJ8127), Changsha Municipal Natural Science Foundation (No. kq2403041), Hunan Provincial Administration of Traditional Chinese Medicine (No. B2023059), and Fundamental Research Funds for the Central Universities of Central South University (No. 2024ZZTS0990, 2025ZZTS0913).

Data availability

All data generated or analyzed during this study are included within the article and its supplementary information file.

Declarations

Competing interests

The authors declare no conflict of financial interest.