Chalasoergodimer A (1) was isolated as a brown powder. Its molecular formula was determined to be C60H78N2O5, indicating 23 degrees of unsaturation, based on the HRESIMS ion peak at m/z 929.5818 [M + Na]⁺ (calcd for C₆₀H₇₈N₂O₅Na+, 929.5808). The 1H NMR and HSQC spectra (Table 1 and Fig. S7) revealed the presence of an indolyl moiety, characterized by four coupled aromatic protons at δH 7.40 (d, 1H, J = 7.8 Hz), 7.28 (d, 1H, J = 7.8 Hz), 7.14 (dd, 1H, J = 7.8, 7.8 Hz), and 7.11 (dd, 1H, J = 7.8, 7.8 Hz). Eight non-terminal olefinic protons were observed at δH 6.31 (dd, 1H, J = 13.8, 9.6 Hz), 6.26 (d, 1H, J = 9.6 Hz), 6.19 (d, 1H, J = 9.6 Hz), 5.98 (d, 1H, J = 9.6 Hz), 5.51 (d, 1H, J = 4.8 Hz), 5.39 (ddd, 1H, J = 14.4, 10.8, 2.4 Hz), and two signals at δH 5.24 (dd, 1H, J = 15.6, 7.2 Hz). Two terminal olefinic protons appeared at δH 5.32 (s, 1H) and 5.13 (s, 1H). Nine methyl groups were detected at δH 1.87 (s, 3H), 1.10 (d, 3H, J = 6.6 Hz), 1.06 (d, 3H, J = 6.6 Hz), 1.06 (d, 3H, J = 6.6 Hz), 0.96 (s, 3H), 0.95 (s, 3H), 0.94 (d, 3H, J = 7.8 Hz), 0.85 (d, 3H, J = 6.6 Hz), and 0.84 (d, 3H, J = 6.6 Hz). The 13C NMR (Table 1) and HSQC spectra displayed 60 carbon signals, including two ketone carbonyls (δC 208.0 and 203.7), one amide carbonyl (δC 173.6), 22 olefinic carbons, two oxygenated carbons (δC 71.5 and 70.2), and 33 aliphatic carbons. The planar structure of 1 was further confirmed by analysis of 1H-1H COSY and HMBC spectra (Fig. 2). The 1H-1H COSY correlations of H-4′/H-5′/H-6′/H-7′, together with HMBC correlations from NH-1′ to C-1′a, C-3′, C-3′a, from H-4′ to C-1′a, and from H-7′ to C-3′a, suggested that the indole moiety contained only five hydrogen atoms. The 1H-1H COSY correlations of H-2/H-3/H-4/H-5/H-11, H-7/H-8/H-13/H-14/H-15/H-16/H-17, and H-20/H-21/H-22, along with the HMBC correlations from NH-2 to C-1, C-9, from H-4 to C-6, C-23, from H-12 to C-5, C-7, from H-17 to C-19, C-25, and from H-21 to C-19, C-23, further established unit A as the C-2′ dehydrogenated derivative of chaetoglobosin Fex (9). The characteristic structure included amide group at C-1, carbonyl groups at C-19 and C-23, a terminal alkene at C-12, and a hydroxyl group at C-20. In the structure of unit B, 1H-1H COSY correlations of H-1′′/H-2′′/H-3′′/H-4′′, H-6′′/H-7′′, H-9′′/H-11′′/H-12′′, H-15′′/H-16′′/H-17′′/H-20′′/H-22′′/H-23′′/H-24′′/H-25′′/H-26′′, together with HMBC correlations from H-3′′ to C-5′′, from H-4′′ to C-6′′, C-10′′, from H-6′′ to C-8′′, from H-7′′ to C-9′′, C-14′′, and from H-18′′ to C-12′′, C-14′′, C-17′′, indicated that unit B possessed a ergosta-4,6,8(14),22-tetraen-3β-ol skeleton [30] substituted at C-3′′. Moreover, the key HMBC correlations from H-3′′ to C-2′, C-3′, along with the absence of a proton signal at C-2′ in the HSQC spectrum, suggested that the planar structure of 1 consisted of chaetoglobosin Fex and ergosta-4,6,8(14),22-tetraen-3β-ol moiety, connected through a C–C single bond between C-2′ and C-3′′, with the hydroxyl group at C-3′′ eliminated.

The 1H-1H COSY and key HMBC correlations of 1

To further evaluate the rationality of the planar structure of 1, gauge-independent atomic orbital (GIAO) calculations were performed to predict the 13C NMR chemical shifts of a model featuring a linkage between C-2′ of chaetoglobosin Fex and C-3′′ of ergosta-4,6,8(14),22-tetraen-3β-ol. As C-3′′ was a stereogenic center, two configurations were considered, and the 13C NMR chemical shifts of both were calculated. Linear regression analysis of the calculated versus experimental 13C chemical shifts, together with deviation analysis (Fig. 3), indicated that the C-2′/C-3′′ linkage pattern determined by 1D and 2D NMR data was reasonable. In this mode, the calculated 13C NMR data showed good agreement with the experimental results. For 1a, the correlation coefficient (R2) was 0.9984, and all absolute deviations (|∆δ|) were within 5.40 ppm. Moreover, the key positions C-2′ and C-3′′ exhibited minimal deviations, with |∆δ| values below 1.30 ppm, suggesting that 1 was more likely to adopt the 1a structure.

A Linear Regression analysis, DP4 analysis and subtraction of the experimental and calculated 13C NMR chemical shifts of 1a. B Linear Regression analysis, DP4 analysis and subtraction of the experimental and calculated 13C NMR chemical shifts of 1b

The relative configuration of unit A was comprehensively assessed through coupling constants, NOESY correlations, and chemical shifts. The coupling constant between H-13 and H-14 in unit A exceeded 12 Hz, supporting an E-configuration for the Δ13 double bond. The NOESY correlation between H-8 and H-14 further supported this assignment. The NOESY correlation between H-16 and H3-25 indicated that the Δ17 double bond also adopted an E-configuration. A series of NOESY correlations involving H-3/H3-11, H3-11/H-7, H-7/H-13, H-13/H-22a, and H-22a/H-20 suggested that these protons shared the same spatial orientation. Likewise, NOESY cross-peaks observed between H-4/H-8, H-8/H-14, and H-14/H-16 indicated that these protons also adopted a common orientation. To further assign the relative configuration at C-16 and C-20, the comparison of the NMR data (Table S1) between 1 and 9 revealed that the chemical shifts of H-20, C-16, and C-20 were generally similar (|ΔδH|≤ 0.1 ppm, |ΔδC|≤ 0.3 ppm). Comparison with 7 showed that the chemical shifts of H-16, H-20, C-16, and C-20 in 1 were similar to those in 7 (|ΔδH|≤ 0.06 ppm, |ΔδC|≤ 0.3 ppm). These similar chemical shifts suggested that 1, 7, and 9 shared closely related structures and configurations. Therefore, the relative configuration of unit A was assigned as 3S*, 4R*, 5S*, 7S*, 8R*, 16S*, and 20S*. The configuration of unit B in 1 was assessed based on coupling constant analysis and NOESY data. The large coupling constant between H-22′′ and H-23′′ suggested that the Δ22′′ double bond adopted an E-configuration. NOESY correlations between H-11′′b and both H3-18′′ and H3-19′′ implied that the methyl groups at C-10′′ and C-13′′ were located on the same face. The NOESY correlations of H-9″/H-12″b and H-12″a/H3-18″ indicated that H-9″ and H3-18″ were oriented on opposite faces. The NOESY correlations of H-15″a with H-17″, and H-15″b with H3-18″ indicated that H-17″ and H3-18″ were oriented on opposite faces. Due to the conformational flexibility of the side chain in unit B, the relative configurations at C-20′′ and C-24′′ could not be accurately determined based on NOESY data alone. Due to the lack of key NOESY correlations for C-3″, its relative configuration could not be reliably determined. Except for C-3″, C-20″, and C-24″, the relative configuration of unit B was determined as 9″R*, 10″R*, 13″R*, and 17″R*.

Among the reported chaetoglobosins derived from Chaetomium species [31,32,33,34], the configurations at C-3, C-4, C-8, C-9, and C-16 were relatively conserved as 3S, 4R, 8R, 9R, and 16S, respectively. When a methyl group was attached at C-5, it typically exhibited the S-configuration. Similarly, the presence of a hydroxyl group at C-7 was generally associated with the S-configuration. In most cases, the hydroxyl substitution at C-20 displayed an S-configuration. Only a few reports described an R-configuration at C-20, which notably affected the chemical shifts of C-20 and its neighboring atoms.

Based on the relative configuration of 1, as well as the characteristic structural analogy to related compounds derived from Chaetomium species, the absolute configuration of unit A was confirmed as 3S, 4R, 5S, 7S, 8R, 9R, 16S, and 20S. To determine the absolute configuration of unit B, its chemical shifts were compared (Table S2) with those of the monomer ergosta-4,6,8(14),22-tetraen-3β-ol [30]. The side chain from C-20′′ to C-28′′ exhibited closely matching chemical shifts (|ΔδH|≤ 0.1 ppm), indicating that the stereocenters at C-20′′ and C-24′′ likely possessed the same configurations. A similar comparison between 1 and 7 also showed comparable chemical shifts for the side chain in unit B (|ΔδH|≤ 0.1 ppm). Furthermore, a structural feature comparison of ergosterol and its derivatives derived from Chaetomium species revealed that those bearing chemically identical side chains consistently adopted R-configurations at C-20 and C-24, while S-configured analogues have not been widely reported in the literature [31,32,33,34]. Thus, both of the absolute configurations at C-20′′ and C-24′′ in unit B were tentatively assigned as R-configuration. Based on the relative configuration of unit B and structural analogies of the chiral centers in ergosterol and its derivatives derived from Chaetomium species, the absolute configuration of unit B, excluding C-3″, was assigned as 9′′R, 10′′R, 13′′R, 17′′R, 20′′R, and 24′′R.

To further evaluate the configuration at C-3′′, DP4 analysis was performed for the two possible isomers. The DP4 results (Fig. 3) indicated that the population of 1a was 100%, whereas that of 1b was 0%, suggesting that the configuration at C-3′′ in 1 was R configuration. Subsequently, density functional theory (DFT) calculations at the B3LYP/6–311 + G(d,p) level in methanol were carried out for (3S, 4R, 5S, 7S, 8R, 9R, 16S, 20S, 3′′R, 9′′R, 10′′R, 13′′R, 17′′R, 20′′R, and 24′′R)-1, and the calculated ECD spectrum was compared with the experimental one. The results (Fig. 4) showed that the calculated and experimental ECD spectra exhibited good agreement. The absolute configuration of 1 was thus assigned as 3S, 4R, 5S, 7S, 8R, 9R, 16S, 20S, 3′′R, 9′′R, 10′′R, 13′′R, 17′′R, 20′′R, and 24′′R.

A Key NOESY correlations and B experimental versus calculated ECD spectra of 1

Chalasoergodimer B (2) was isolated as a brown powder. The molecular formula was determined to be C60H78N2O5, indicating 23 degrees of unsaturation, based on the HRESIMS ion peak at m/z 929.5811 [M + Na]+ (C60H78N2O5Na+, calcd. for 929.5808). The NMR data of 2 (Table 1) closely resembled those of 1, except for differences in the indolyl moiety. Other signals, including two carbonyl groups, one amide group, two hydroxyl groups, and seven double bonds (including one terminal double bond), were consistent with those of 1. These findings suggested that 2 was also formed from chaetoglobosin Fex and ergosta-4,6,8(14),22-tetraen-3β-ol, but exhibited a distinct substitution pattern compared to 1. Comparison of the indole proton signals revealed an additional signal at δH 6.98 (s, 1H) in 2, and lacked the proton signal for NH-1′, indicating that the substitution occurred at N-1′ rather than C-2′. This was further supported by the key HMBC correlation from H-2′ to C-3′′ (Fig. 5), which confirmed that the substitution took place at the NH-1′ position of the indolyl unit. Except for the connecting site and its neighboring region, compounds 1 and 2 generally exhibited similar 1D and 2D NMR signals in other regions.

The 1H-1H COSY and key HMBC correlations of 2–5

Compound 2 was suggested to possess the same monomeric composition as 1, but with a different dimerization pattern. Given that 2 exhibited several identical NOESY correlations (Fig. 6), closely similar chemical shifts, and comparable coupling constants to those of 1, it was considered reasonable to infer that the difference in dimerization mode did not significantly affect the relative or absolute configurations of the structure, except for C-3″. Therefore, 2 was considered to share the same relative and absolute configurations as 1 except the dimerization sites. The absolute configuration of 2, excluding C-3″, was suggested as 3S, 4R, 5S, 7S, 8R, 9R, 16S, 20S, 9′′R, 10′′R, 13′′R, 17′′R, 20′′R, and 24′′R. The C-3′′ configuration of 2 was determined using the same approach as for 1. DP4 analysis was performed to evaluate the distribution of the two possible configurations at the C-3′′ stereocenter. The results (Fig. S31) indicated that 2 was assigned 100% to the 3′′R configuration and 0% to the 3′′S configuration. Subsequently, theoretical ECD spectra of (3S, 4R, 5S, 7S, 8R, 9R, 16S, 20S, 3′′R, 9′′R, 10′′R, 13′′R, 17′′R, 20′′R, and 24′′R)-2 was calculated and compared with the experimental ECD spectrum of 2. As shown in Fig. 7, the calculated spectrum of the (3S, 4R, 5S, 7S, 8R, 9R, 16S, 20S, 3′′R, 9′′R, 10′′R, 13′′R, 17′′R, 20′′R, and 24′′R)-2 showed better agreement with the experimental data, supporting the assignment of the absolute configuration of 2 as 3S, 4R, 5S, 7S, 8R, 9R, 16S, 20S, 3′′R, 9′′R, 10′′R, 13′′R, 17′′R, 20′′R, and 24′′R.

Key NOESY correlations of 2–5

Experimental versus calculated ECD spectra of 2–5

Chalasoergodimer C (3) was obtained as a white powder. The molecular formula was determined as C60H76N2O6, corresponding to 24 degrees of unsaturation, based on the HRESIMS ion peak at m/z 943.5598 [M + Na]+ (C60H76N2O6Na+, calcd. for 943.5601). The 1H, 13C, and HSQC spectra (Table 2 and Fig. S38) exhibited features characteristic of chaetoglobosin-ergosterol derivative hybrid. The NMR data revealed the presence of two terminal olefinic protons at δH 5.20 (s, 1H) and 5.42 (s, 1H), three carbonyl carbons at δC 198.4, 206.6, and 214.3, one amide carbonyl at δC 172.7. Compared to chaetoglobosin D (11), compound 3 lacked the C-21 double bond and the hydroxyl group at C-19 was replaced by a carbonyl group. Key HMBC correlations from H-21 to C-15′′ and from H-22 to C-6′′, C-8′′, and C-15′′, along with critical 1H-1H COSY cross-peaks between H-21/H-15′′ and H-22/H-7′′ (Fig. 5), supported a dimeric structure comprising chaetoglobosin D and 14-dehydroergosterol [35], connected via C-21 to C-15′′ and C-22 to C-7′′.

Based on the coupling constant between H-13 and H-14 and the NOESY correlation between H-8 and H-14, the double bond at C-13 was assigned as E-configuration. The NOESY correlation between H-16 and the methyl group H3-25 indicated that the double bond at C-16 was also E-configuration. The NOESY spectrum of 3 (Fig. 6) revealed key correlations, including H-3/H-7 and H-3/H-11, indicating that protons H-3, H-7, and H-11 were co-facial. Similarly, correlations of H-4/H-8, H-8/H-14, and H-14/H-16 supported the same orientation for protons H-4, H-8, and H-16. NOESY cross-peaks of H-1′′a/H-3′′, H-1′′a/H-9′′, H-22/H-9′′, H-9′′/H-12′′b, and H-12′′b/H-17′′ suggested that H-22, H-3′′, H-9′′, and H-17′′ shared the same spatial orientation. Additional correlations of H-7′′/H-15′′, H-7′′/H-19′′, H-11′′b/H-19′′, H-11′′b/H-18′′, and H-21/H-18′′ indicated that the protons at H-21, H-7′′, H-15′′, H-18′′, and H-19′′ were positioned on the same side. The coupling constants and NOESY correlations both indicated that 3 and 6 shared a highly similar configuration. Based on the established NOESY correlations and the similar chemical characteristics between 3 and 6, the relative configuration of 3 was assigned as 3S*, 4R*, 5S*, 7S*, 8R*, 9R*, 16S*, 21S*, 22S*, 3′′S*, 7′′S*, 9′′R*, 10′′R*, 13′′R*, 15′′R*, 17′′R*, 20′′R*, and 24′′R*.

Based on the relative configurations, and in combination with structural feature comparisons of chaetoglobosins and ergosterol derivatives derived from Chaetomium species, the absolute configuration of 3 was determined as 3S, 4R, 5S, 7S, 8R, 9R, 16S, 21S, 22S, 3′′S, 7′′S, 9′′R, 10′′R, 13′′R, 15′′R, 17′′R, 20′′R, and 24′′R. Subsequently, ECD calculations were performed on (3S, 4R, 5S, 7S, 8R, 9R, 16S, 21S, 22S, 3′′S, 7′′S, 9′′R, 10′′R, 13′′R, 15′′R, 17′′R, 20′′R, and 24′′R)-3, and the results were compared with the experimental ECD spectrum of 3. The comparison (Fig. 7) showed good agreement between the calculated and experimental data, supporting the assignment of the absolute configuration of 3 as 3S, 4R, 5S, 7S, 8R, 9R, 16S, 21S, 22S, 3′′S, 7′′S, 9′′R, 10′′R, 13′′R, 15′′R, 17′′R, 20′′R, and 24′′R.

Chalasoergodimer D (4) was isolated as a white powder. The molecular formula was determined as C60H76N2O6, corresponding to 24 degrees of unsaturation, based on the HRESIMS ion peak at m/z 943.5579 [M + Na]+ (C60H76N2O6Na+, calcd. for 943.5601). The 1H and 13C NMR signals (Table 2) of 4 were similar to 3. By comparison, the structure of 4 lacked the terminal alkene signal but displayed an additional methyl resonance at δH 1.74 (s, 3H). Key HMBC correlations from H-11 to C-4, C-5, and C-6, and from H-12 to C-5, C-6, and C-7 (Fig. 5), supported the presence of a double bond between C-5 and C-6, each substituted with a methyl group. Given the absence of significant differences in other structural elements, the planar structure of 4 was determined accordingly. Compound 4 exhibited highly similar NOESY correlations and coupling constants to those of 3. Since they shared the similar planar structure and chemical shifts overall, this suggested that they possessed similar configurations. Therefore, except for C-5, compound 4 was assigned the same relative configuration as 3. Based on the relative configurations, and in combination with structural feature comparisons of chaetoglobosins and ergosterol derivatives derived from Chaetomium species, the absolute configuration of 4 was suggested as 3S, 4R, 7S, 8R, 9R, 16S, 21S, 22S, 3′′S, 7′′S, 9′′R, 10′′R, 13′′R, 15′′R, 17′′R, 20′′R, and 24′′R. Moreover, the excellent agreement between the calculated and experimental ECD spectra (Fig. 7) supported the assignment of the absolute configuration of 4 as 3S, 4R, 7S, 8R, 9R, 16S, 21S, 22S, 3′′S, 7′′S, 9′′R, 10′′R, 13′′R, 15′′R, 17′′R, 20′′R, and 24′′R.

Chalasoergodimer E (5) was isolated as a white powder. The molecular formula was determined as C60H78N2O6, corresponding to 23 degrees of unsaturation, based on the HRESIMS ion peak at m/z 945.5745 [M + Na]+ (C60H78N2O6Na+, calcd. for 945.5758). Analysis of the 1D (Table 3) and 2D NMR data of 5 (Fig. 5), in comparison with 3, revealed that 5 and 3 shared highly similar structures, with the only difference being that 5 bore a hydroxyl group at C-19. Comparing of NOESY correlations (Fig. 6) and coupling constants confirmed that, except for C-19, the relative configuration of 5 was identical to that of 3. The NOESY correlations of H-19 with H-17, H-17 with H-15b, and H-15b with H3-24 indicated that H-19 and H3-24 were oriented on same side. In addition, the similar chemical shifts of H-19 in 5 and 6 suggested comparable chemical environments, which further supported the evaluation that the relative configuration of C-19 was likely the same in 5 and 6. Based on the relative configurations, and in combination with structural feature comparisons of chaetoglobosins and ergosterol derivatives derived from Chaetomium species, the absolute configuration of 5 was determined as 3S, 4R, 5S, 7S, 8R, 9R, 16S, 19S, 21S, 22S, 3′′S, 7′′S, 9′′R, 10′′R, 13′′R, 15′′R, 17′′R, 20′′R, and 24′′R. Moreover, the calculated ECD spectrum (Fig. 7) of (3S, 4R, 5S, 7S, 8R, 9R, 16S, 19R, 21S, 22S, 3′′S, 7′′S, 9′′R, 10′′R, 13′′R, 15′′R, 17′′R, 20′′R)-5 exhibited good agreement with the experimental data of 5, thereby confirmed its absolute configuration as 3S, 4R, 5S, 7S, 8R, 9R, 16S, 19R, 21S, 22S, 3′′S, 7′′S, 9′′R, 10′′R, 13′′R, 15′′R, 17′′R, 20′′R, and 24′′R.

Based on the structures of compounds 1–12 and the currently reported studies on chaetoglobosin heterodimers [27, 36, 37], a possible biosynthetic pathway was proposed (Scheme 1). Compound 1 was likely formed via a substitution reaction between chaetoglobosin Fex (9) and ergosta-4,6,8(14),22-tetraen-3β-ol. The electron-rich C-2′ position of the indole ring rendered it particularly susceptible to nucleophilic substitution by electron-deficient species. Compounds 2 and 7 were presumably generated through substitution at the HN-1′ position of the indole moiety. Notably, compounds 1 and 2, as well as 7, exhibited opposite configurations at C-3′, suggesting a possible unimolecular nucleophilic substitution mechanism. Given the structural similarity between 7 and 8, their dimerization was assumed to proceed via similar mechanisms. Compounds 5 and 6 likely arose from Diels–Alder cycloaddition reactions at C-21 and C-22 between chaetoglobosin D or B and 14-dehydroergosterol. In contrast, compounds 3 and 4 may originate from post-cyclization oxidation of 5 and 6.

The proposed biosynthetic pathways of 1–12

The possibility of an artificial reaction between chaetoglobosins and ergosterol analogues was evaluated. In these model experiments, cholesterol, which shared a similar structure with ergosta-4,6,8(14),22-tetraen-3β-ol, and chaetoglobosin Fex were used. Equimolar mixtures of the two compounds were stirred at 55 °C for 12 h in methanol:dichloromethane solvents (1:0, 1:1, and 0:1, v/v). Thin-layer chromatography (TLC) analysis of the resulting mixtures revealed no significant formation of new products. Due to the lack of available ergosterol derivatives structurally similar to 14-dehydroergosterol, further experimental evaluation of the Diels–Alder cycloaddition pathway could not be conducted. Given the rarity of enzymatic nucleophilic substitution involving alcohols at the C-1′ or C-2′ positions of indole rings, compounds 1, 2, 7 and 8 were more likely biosynthesized through non-enzymatic pathways. Although the Diels–Alder cycloaddition represents a well-known class of natural product cyclizations, only a limited number of enzymes capable of catalyzing this reaction have been identified. Thus, the formation of 3, 4, 5, and 6 were also presumed to proceed via non-enzymatic mechanisms.

Chaetoglobosins have been reported to exhibit antitumor activity [38]. Accordingly, the cytotoxic activities of 1–12 against A549 cells were evaluated (Table S10). Compounds 9–12 showed notable inhibitory effects, with IC50 values of 9.26, 9.14, 14.89, and 5.14 μM. Cisplatin was selected as a positive control, with the IC₅₀ of 2.39 μM, slightly lower than that of compound 12. In contrast, compounds 1–8 displayed no significant cytotoxicity, possibly due to the increased molecular weight and reduced polarity resulting from polymerization with ergosterol derivatives, which may have hindered their membrane permeability and subsequent bioactivity.