HPLC analysis was performed using an 1100 series HPLC system (binary pump: G1312A, auto sampler: A1329A, column compartment: G1316A, UV detector: G1314A, Agilent Technologies, Inc., Santa Clara, CA, USA) equipped with an

Inertsil ODS-3 column (3 μm, 2.1 × 150 mm) (GL Science Inc., Tokyo, Japan). Open column chromatography for compound isolation was performed using 100 g of Sephadex LH-20 (GE Healthcare, Chicago, IL, USA) in a 3.5 × 80 cm glass column. Preparative MPLC was performed on a Shimadzu LC-20AT and SPD-10AV instrument equipped with an ULTRA PACK ODS-S-50B column (50 μm, 26 × 300 mm) (Yamazen Corp., Osaka, Japan). Preparative HPLC was performed on a Shimadzu LC-20AT and SPD-10AV instrument equipped with a GL Science InertSustain C18 column (5 μm, 10 × 250 mm). NMR spectra were recorded on an ECA-600II instrument (¹H: 600.17 MHz; ¹³C: 150.91 MHz) (JEOL, Tokyo, Japan). Chemical shifts for ¹H- and ¹³C-NMR are given in parts per million (δ) relative to residual solvent signals ([δ_H 7.26]/[δ_C 77.0] for CDCl₃, [δ_H 2.49]/[δ_C 39.5] for DMSO-d₆ as internal standards). Mass spectra were measured on a JEOL JMS-T100LP or JMS-700. Circular dichroism (CD) spectra were recorded on a J-820 spectropolarimeter (Jasco). Liquid cultures were fermented in a Taitec Bio-Shaker BR-300LF (Taitec Corp., Saitama, Japan). Rice cultures were fermented in a Sanyo MIR-554 incubator (Sanyo Electric Co., Ltd., Osaka, Japan).

The C. yanagawaensis MAFF 247556 isolate was obtained from ears of Zoysia japonica.

Clavicepsyanagawaensis was pre-cultured on potato dextrose agar (PDA, Nissui Pharmaceutical Co., Ltd., Tokyo, Japan). Fermentation was conducted using M102 medium (sucrose 30 g, malt extract 20 g, peptone 2 g, yeast extract 1 g, MgSO₄-7H₂O 0.5 g, KCl 0.5 g, KH₂PO₄ 1.0 g, purified water 1 L, NaOH [pH 6.0]) and rice medium (rice: 750 g, tap water: 450 mL).

First, C. yanagawaensis was pre-cultured on PDA for 21 days at 25˚C, after which two pieces of cultured agar (1 × 2 cm²) were transferred into 500-mL Erlenmeyer flasks containing 200 mL of M102 medium and grown at 25˚C with shaking at 150 rpm for 7 days. Next, 10-mL aliquots of cultured M102 medium were transferred into 500-mL Roux flasks containing rice medium (each flask contained 150 g of medium as dry rice) and grown at 25˚C for 28 days.

Each rice culture of C. yanagawaensis was extracted with 2.5 L of methanol, and then the methanol extract was evaporated to dryness in a vacuum evaporator. The crude methanol extract (9.83 g) was suspended in 300 mL of water and extracted three times with an equal volume of ethyl acetate. The resulting ethyl acetate extract (2.1 g) was obtained by evaporation under vacuum, suspended in 300 mL of n-hexane, and extracted three times with an equal volume of acetonitrile. The solvent was removed under vacuum, and the acetonitrile extract (1461.5 mg) was separated into 11 fractions (A-K) by open column chromatography using Sephadex LH-20. The mobile phases were n-hexane:chloroform (1:4; 200 mL), chloroform:acetone (3:2; 200 mL and 1:4; 200 mL), acetone (200 mL), and MeOH (500 mL). Fraction E (136.4 mg) was purified by HPLC on an ODS column eluted with 50% acetonitrile to obtain 1 (3.5 mg) as a yellow amorphous powder. Fraction F (135.3 mg) was separated by MPLC on an ODS column eluted with 50% acetonitrile to obtain 7 fractions. Fraction F-5 (19.3 mg) was purified by HPLC on an ODS column eluted with 45% acetonitrile to obtain 2 (2.1 mg) as a yellow amorphous powder. Fraction C (77.5 mg) was purified by HPLC on an ODS column eluted with 50% acetonitrile to obtain 3 (27.3 mg) as a colorless oil. Fraction I (43.4 mg) was separated by MPLC on an ODS column eluted with 50% acetonitrile to obtain 7 fractions. Fraction I-5 (10.7 mg) was purified by HPLC on an ODS column eluted with 50% acetonitrile to obtain 3,4-dihydroxy-5-(3-methyl-2-buten-1-yl)benzoic acid (4; 7.9 mg) as a white amorphous powder [11]. Fraction B (162.6 mg) was separated by MPLC on an ODS column eluted with 50% acetonitrile to obtain 4 fractions. Fraction B-2 (34.6 mg) was purified by HPLC on an ODS column eluted with 50% acetonitrile to obtain methyl veratrate (5; 16.2 mg) as a white amorphous powder [12]. The absence of EAs in the extracts was confirmed by HPLC analysis (data not shown).

Compound 1: yellow amorphous powder. [ α ] D 24 −351.3 (c 0.1, CHCl₃). UV (MeOH) λ_max 204 nm (logε 5.57), 335 (logε 5.38). ¹H- and ¹³C-NMR data, see Table 1. HR-ESI-MS (negative ion) m/z [M-H]⁻ 653.1514 (calcd for C₃₂H₂₉O₁₅: 653.1506).

Compound 2: yellow amorphous powder. [ α ] D 24 −222.02 (c 0.1, CHCl₃). UV (MeOH) λ_max 204 nm (logε 5.49), 335 (logε 5.23). ¹H- and ¹³C-NMR data, see Table 1. HR-ESI-MS (negative ion) m/z [M-H]⁻ 639.1328 (calcd for C₃₁H₂₇O_15: 639.1350).

Compound 3: colorless oil. UV (MeOH) λ_max 206 nm (logε 4.69), 273 (logε 0.67). ¹H- and ¹³C-NMR data, see Table 2. HR-EI-MS (positive ion) m/z [M]⁺ 236.1413 (calcd for C₁₄H₂₀O₃: 236.1412).

Human hepatoma cells (HepG2) were obtained from the National Institutes of Bio-medical Innovation, Health, and Nutrition (Osaka, Japan). Human promyelocytic leukemia cells (HL60), human pancreatic carcinoma cells (PANC-1), and human glioblastoma multiforme tumor cells (T98G) were obtained from Riken BioResource Research Center (Ibaraki, Japan). Human colon adenocarcinoma cells (HT29) were obtained from KAC Co., Ltd. (Kyoto, Japan). Cells were cultured in DMEM (Nacalai Tesque, Inc., Kyoto, Japan) or RPMI 1640 (Nacalai Tesque, Inc.) supplemented with 10% heat-inactivated fetal bovine serum (Merck, Darmstadt, Germany) and 1% Antibiotic-Antimycotic Mixed Stock Solution (Nacalai Tesque, Inc.) at 37˚C in a humid incubator containing ambient air supplemented with 5% CO₂.

Cell viability was measured using an MTT Cell Count kit (Nacalai Tesque, Inc.) according to the manufacturer’s instructions. Briefly, cells were seeded in a 96-well plate at 5.0 × 10³ cells per well; each experiment was performed with four replicates. After 48 h, the growth medium was replaced with fresh medium supplemented with test compounds (at concentrations ranging from 6.25 to 100 μM), and the plates were cultured for another 72 h. Next, 10 μL of MTT solution was added to each well, and the plates were incubated at 37˚C for 3 h and then treated with 100 μL/well of solubilization solution to dissolve the formazan crystals. The absorbance at 570 nm was measured using a microplate reader (As One Corp., Osaka, Japan). Cell viability is expressed as the relative percentage of absorbance in the experimental groups normalized to that of the negative control (no added compound) group; where appropriate, outliers were excluded based on identification using the Smirnov-Grubbs test.

Statistical analysis was performed using the multcomp package in R, version 4.1.0 [13]. Where appropriate, values are expressed as mean ± standard deviation (SD). Comparisons to the respective negative controls were conducted using Dunnett’s multiple comparisons test. Values of p < 0.05 were considered statistically significant.

Compound 1 { [ α ] D 24 −351.3 (c 0.1, CHCl₃)} was obtained as a yellow amorphous powder solid. Compound 1 was revealed to have the molecular formula C₃₂H₃₀O₁₅ [m/z 653.1514, [M-H]⁻, calcd 653.1506] based on HR-ESI-MS data. The ¹H-NMR spectrum of 1 in CDCl₃ (Table 1) showed four doublet aromatic protons [δ_H 7.57 (d, J = 8.7 Hz), 7.47 (d, J = 8.7 Hz), 6.68 (d, J = 8.7 Hz), 6.63 (d, J = 8.7 Hz)] and suggested the presence of two benzene rings with ortho coupling. The ¹H-NMR spectrum also showed two doublet methyl protons [δ_H 1.22 (d, J = 6.6 Hz) and δ_H 1.17 (d, J = 5.8 Hz)], two oxygen-bearing methyl protons (δ_H 3.68 and 3.72), and two oxygen-bearing methine protons [δ_H 4.50 (d, J = 10.7 Hz) and δ_H 3.93 (d, J = 11.6 Hz)]. The ¹³C-NMR spectrum showed three ketone carbonyl (δ_C 198.9, 191.8, 187.1) and three sp³ quaternary carbons, four sp³ methyl carbons, two sp³ methylene carbons, two sp³ methine carbons, two oxygen-bearing sp³ methine carbons, and 16 sp² carbons (estimated benzene rings, an olefin carbon, an oxygen-bearing olefin carbon or ester carbons).

Next, the planar structure of 1 was elucidated through detailed analyses of 2D NMR data, including ¹H-¹H correlation spectroscopy (COSY), heteronuclear single quantum coherence (HSQC), and heteronuclear multiple bond correlation (HMBC) spectra in CDCl₃ (Figure 2 and Figures S2-S6). The ¹H-¹H COSY correlations for H-3 (δ_H 7.57)/H-4 (δ_H 6.68) and HMBC correlations from H-3 to

C-1 (δ_C 160.6) and C-4a (δ_C 157.4), from H-4 to C-2 (δ_C 118.5), C-4a and C-9a (δ_C 106.6), and from OH-1 (δ_H 11.77) to C-1, C-2, and C-9a suggested a benzene ring with a hydroxy group at the 1-position and an oxygen attached at the 4a-position. The ¹H-¹H COSY correlations for H-11 (δ_H 1.22)/H-6 (δ_H 2.07), H-6 (δ_H 2.07)/H-7(δ_H 2.49, 2.94)/H-5 (δ_H 4.50) and HMBC correlations from H-7 to C-8 (δ_C 198.9) and C-8a (δ_C 71. 7) and from H-5 to C-10a (δ_C 89.4) suggested a cyclohexanone ring with a methyl group at the 6-position, a hydroxy group at the 5-position, and a carbonyl group at the 8-position. In addition, HMBC correlation from the H-5 methine and H-13 methyl to C-12 (δ_C 167.9) suggested partial structure A with a carboxymethyl group at the 10a-position (Figure 2).

Similarly, the ¹H-¹H COSY correlations for H-3’ (δ_H 7.47)/H-4’ (δ_H 6.63) and HMBC correlations from H-3’ to C-1’ (δ_C 159.2), from H-4’ to C-2’ (δ_C 117.4), C-4a’ (δ_C 158.3), C-9a’ (δ_C 106.9), and C-9’ (δ_C 187.1) and from OH-1’ (δ_H 11.73) to C-1’, C-2’, and C-9a’ suggested another benzene ring with a hydroxy group at the 1’-position and an oxygen attached at the 4a’-position. The ¹H-¹H COSY correlations for H-11’ (δ_H 1.17)/H-6’ (δ_H 2.42), H-6’/H-7’ (δ_H 2.32, 2.74), and H-5’ (δ_H 3.93) and HMBC correlations from H-7’ to C-8’ (δ_C 177.5) and C-8a’ (δ_C 101.5), from H-5’ to C-10a’ (δ_C 84.8), and from OH-8’ (δ_H 13.76) to C-7’ (δ_C 36.2), C-8’, and C-8a’ suggested a cyclohexene ring with a methyl group at the 6’-position and a hydroxy group at the 5’- and 8’-positions. In addition, HMBC correlation from the H-5’ methine and H-13’ methyl to C-12’ (δ_C 170.3) suggested partial structure B. The HMBC correlations from H-3 to C-2’ and from H3’ to C-2 suggested a 2-2’ bond between partial structures A and B.

These results and UV spectra (λ_max 204, 264 and 335 nm) suggested that 1 is a secalonic acid analogue as a dimer of tetrahydroxanthone derivatives. The ¹H- and ¹³C-NMR data indicated that 1 was similar to the secalonic acid analogue F-7 (6), except for the coupling constant at the 5-position (1: d, 10.7 Hz, F-7 (6): brs). Therefore, we considered 1 to be a stereoisomer of F-7 (6) (Figure 2) [14].

The configuration of 1 was also examined. The 5’-H/6’-H coupling constant of 1 was 11.6 Hz, similar to that of F-7 (6, 11.0 Hz), and these two hydrogens had a trans-diaxial configuration, but 5-H/6-H differed significantly from that of F-7 (6; brs) at 10.7 Hz. Thus, the 5-H/6-H configuration of 1 was suggested as trans, different from that of F-7 (6).

Next, we examined the configuration at the 10a- and 10a’-positions of 1. In the case of secalonic acids, it has been reported that the Cotton around 330 nm in the CD spectrum correlates with the configuration at the 10a- and 10a’-positions [14] [15] [16] [17] [18]. The CD spectrum of 1, which showed a negative Cotton around 333 nm, suggested that 1 has a configuration of 10aR and 10a’S, which differs from that of F-7 (6; 10aS, 10a’R) (Figure 3).

The configurations at positions 5, 6, and 10a and 5’, 6’, and 10a’ of 1 were estimated from the biosynthetic reaction of secalonic acids. The tetrahydroxanthone skeleton, the monomer unit of secalonic acids, is formed by spontaneous cyclization of the biosynthetic intermediate diphenylmethanone analogue A (FigureS7). During this spontaneous cyclization reaction, the carboxymethyl group at position 10a and the methyl group at position 11 are enantioselectively placed in the trans configuration due to 1,3-axial pseudoaxial interaction [19]

[20]. Therefore, the configuration of 1 was 5S, 6R, 10aR, 5’S, 6’R, and 10a’S, and the structure of 1 was suggested to be (-)-5-epi-F-7 (Figure 1).

Compound 2 { [ α ] D 24 −222.02 (c 0.1, CHCl₃)} was obtained as a yellow amorphous powder solid. Compound 2 was revealed to have the molecular formula C₃₁H₂₈O₁₅ [m/z 639.1328, [M-H]⁻, calcd 639.1350] by HR-ESI-MS data. From the molecular formula and UV spectra (λ_max 204 and 335 nm), 2 was estimated to be a secalonic acid analogue with a molecular weight lower by CH₂ than compound 1. Comparing ¹H- and ¹³C-NMR spectra in CDCl₃ (Table 1, Figure 4 and Figures S8-S12) with those of 1, the 13-position methyl disappeared, and the carbon at the 8-position shifted downfield. Therefore, the structure of 2 was suggested to be a lactone ring formed between positions 10a and 8 of 1. This result was supported by detailed analysis of the 2D NMR spectra (Figure 4).

The configurations at 5-H/6-H and 5’-H/6’-H of 2 were trans-diaxial due to similarity to the coupling constants of 1. The CD spectra of 2 showed a negative Cotton around 333 nm (Figure 5). The configuration at the chiral center of 2 was estimated as 5S, 6R, 8R, 10aR, 5’S, 6’R, and 10a’S, and the structure of 2 was suggested as shown in Figure 1.

Compound 3 was obtained as a colorless oil. Compound 3 was revealed to have the molecular formula C₁₄H₂₀O₃ [m/z 236.1413, [M]⁺, calcd. 236.1412] based on HR-EI-MS data. The ¹H-NMR spectrum of 3 in DMSO-d₆ (Table 2) showed two doublet aromatic protons [δ_H 6.81 (d, J = 1.4 Hz), 6.66 (d, J = 1.4 Hz)] and suggested the presence of a 1,3,4,5-tetrasubstituted benzene ring with meta coupling. One olefine proton (δ_H 5.18) suggested the presence of a trisubstituted olefinic moiety. Two oxygen-bearing methyl protons (δ_H 3.65 and 3.75) suggested two methoxy groups. The ¹H-NMR spectrum also showed two broadened singlet methyl protons (δ_H 1.67 and δ_H 1.68), one oxygen-bearing methylene proton [δ_H 4.37 (d, J = 5.5 Hz)], and a doublet methylene-attached olefin [δ_H 3.23 (d, J = 7.6 Hz)]. The ¹³C-NMR spectrum showed two sp³ methyl carbons, two oxygen-bearing sp³ methyl carbons, one sp³ methylene carbon and one oxygen-bearing sp³ methylene carbon, and eight sp² carbons (estimating a

benzene ring and two olefin carbons).

The gross structure of 3 was elucidated through detailed analyses of 2D NMR data, including ¹H-¹H COSY, HSQC, and HMBC spectra in DMSO-d₆ (Figure 6 and Figures S13-S17). The ¹H-¹H COSY correlations for OH-12 (δ_H 5.09)/H-12

(δ_H 4.37) and HMBC correlations from H-4 (δ_H 6.81) to C-2 (δ_C 145.0), C-3 (δ_C 152.1), C-5 (δ_C 138.1), C-6 (δ_C 119.2), and C-12 (δ_C 62.9), from H-6 (δ_H 6.66) to C-2, C-4 (δ_C 108.9), and C-12, and from H-12 to C-4, C-5, and C-6 suggested a benzene ring with a hydroxy methyl group at the 5-position. The HMBC correlations from H-13 (δ_H 3.75) to C-3 and from H-14 (δ_H 3.65) to C-2 suggested methoxy groups at the 2- and 3-positions. The ¹H-¹H COSY correlations for H-7 (δ_H 3.23)/H-8 (δ_H 5.18) and HMBC correlations from H-11 (δ_H 1.68) to C-10 (δ_C 25.5) and from H-10 (δ_H 1.67), C-8 (δ_C 123.2), C-9 (δ_C 131.2), and C-11 (δ_C 17.6) suggested the presence of a dimethylallyl group. The HMBC correlations from H-7 (δ_H 3.23) to C-1 (δ_C 134.1), C-2, and C-6 and from H-6 to C-7 (δ_C 28.1) suggested that the dimethylallyl group was attached at the 1-position. Therefore, the structure of 3 was determined as a 3-methoxy derivative (Figure 1) of methyl bigutol isolated from the mycoparasite Verticilliumbiguttatum (Table S1) [21].

MTT assays were performed using HepG2, HL60, HT29, PANC-1, and T98G cell lines to investigate the cytotoxicity of the isolated compounds against cancer cells (Figure 7). Compound 1 decreased the viability of HepG2, HT29, PANC-1, and T98G at 100 μM, especially T98G at 12.5 μM in a concentration-dependent manner. Compound 2 slightly decreased the viability in HepG2 and PANC-1 cells at 100 μM. Compound 3 decreased the viability of HepG2, HL60, HT29, and PANC-1 at 100 μM, especially HL60 at 6.25 μM in a concentration-dependent manner. Compound 4 decreased the viability of HepG2, HL60, and PANC-1 cells at 100 μM, reducing the viability of PANC-1 cells to <50% relative to control cells. Compound 5 at 100 μM did not affect the viability of any of the cancer cell lines examined.