Triple negative breast cancer (TNBC), a subgroup of breast cancer, is often found as high grade of invasive ductal carcinoma with aggressive behavior (O'Reilly et al., 2021). The incidence rate of TNBC accounts for approximately 15%–20% of breast cancers (Chen et al., 2021; Chowdhury et al., 2021). “Triple negative” is regarded as the absence of the expression of three receptors, estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor 2 receptor (HER2). It causes an about threefold shorter median overall survival (OS) comparing with other breast cancers (O'Reilly et al., 2021). In addition, the relapse and metastasis rate of TNBC is high, the relapse commonly found within 3 years, and the metastasis often occurs in visceral and brain (O'Reilly et al., 2021; Chen et al., 2021; Chowdhury et al., 2021). Due to the difficulties in treating TNBC, more potential molecules are needed.

Ganoderma is a traditional Chinese medicine and has been discovered numerous bioactivities as hypoglycemic effect, cardiovascular protection, anti-tumor, antioxidant, and brain injury prevention (Lin and Deng, 2019; Lin and Sun, 2019; Liu and Tie, 2019; Meng and Yang, 2019; Quan et al., 2019). All along, Ganoderma triterpenoids have been considered as the main active components with anti-tumor effects. In recent years, Ganoderma meroterpenoids, which process phenol moiety and terpene moiety, have been continuously excavated significant anti-tumor activities, such as toward human cancer cell lines (A549, KYSE30, BT549, and MDA-MB-231) (Qin et al., 2018; Cai et al., 2021; Zhang et al., 2021). For the purpose of discovering active agents toward TNBC from natural sources, eleven meroterpenoid dimers including ten novel ones were isolated from Ganoderma (Figure 1). This paper deals with their isolation, structural elucidation, and biological evaluation for cytotoxicity and cell migration inhibition in TNBC cells.

Dimercochlearlactone A (1) was determined to have a molecular formula as C₃₆H₄₂O₈ from its positive HRESIMS (m/z 625.2774 [M + Na]⁺, calcd for C₃₆H₄₂NaO₈, 625.2777). Two typical ABX spin systems (δ _H 7.14, d, J = 3.0 Hz, H-3; δ _H 6.98, dd, J = 8.9, 3.0 Hz, H-5; δ _H 6.80, d, J = 8.9 Hz, H-6; δ _H 7.10, d, J = 2.9 Hz, H-3′; δ _H 6.71, dd, J = 8.8, 2.9 Hz, H-5′; δ _H 6.93, d, J = 8.8 Hz, H-6′) were observed by its ¹H NMR data (Table 1). The ¹³C NMR (Table 2) and DEPT spectra show 36 carbon signals, which were assigned as 5 methyl, 7 methylene, 10 methine, and 14 nonprotonated carbons (10 aromatic including 4 oxygenated, 1 oxygenated aliphatic, 2 ketones, and 1 carbonyl). When consideration of the NMR data of the previous reported meroterpenoids (Qin et al., 2018), the above signals suggest that dimercochlearlactone A (1) might be a dimeric meroterpenoid. The structure of dimercochlearlactone A (1) contains two parts (Parts A and B in Figure 1), which were mainly determined by 2D NMR spectra. For the structure of part A, the ¹H-¹H COSY correlations from H₂-12 (δ _H 1.97 and 1.83) to H₂-11 (δ _H 1.86 and 1.74) and H-13 (δ _H 5.03), and from H₂-17 (δ _H 2.00) to H₂-16 (δ _H 1.91) and H-18 (δ _H 5.07), along with the HMBC correlations (Figure 2) from H₃-20 (δ _H 1.54) and H₃-21 (δ _H 1.63) to C-18 (δ _C 122.5) and C-19 (δ _C 130.6), from H₃-20 to C-21 (δ _C 25.3), from H₃-15 (δ _H 1.52) and H₂-16 to C-13 (δ _C 122.4) and C-14 (δ _C 135.5), from H₃-15 to C-16 (δ _C 39.1), and from H₂-12 to C-14 suggest the presence of two isoprenyl moieties in dimercochlearlactone A (1). In addition, the HMBC correlations from H-8 (δ _H 3.92 and 3.57) to C-7 (δ _C 199.1), C-9 (δ _C 88.2) and C-10 (δ _C 202.6) and from H₂-11 to C-9 and C-10 imply another isoprenyl residue. Furthermore, HMBC correlation between H-3 (δ _H 7.14) and C-7 indicates that the sesquiterpeniod moiety is attached to the ring A via C-7 to C-2. Thus, the structure of part A was determined as shown.

Part B was also elucidated by 2D NMR experiments (¹H-¹H COSY, HSQC and HMBC). The HMBC correlations from H-3′ (δ _H 7.10) to C-7′, from H-8′ (δ _H 8.77) to C-2′ (δ _C 122.5), C-7′ (δ _C 115.0), C-9 and C-10 and the above-mentioned HMBC correlations from H-8 to C-9 and C-10 not only imply the presence of ring B but also indicate that C-2′ is attached to the ring C. The structure of side chain in part B was confirmed by ¹H-¹H COSY correlations observed from H₂-11′(δ _H 2.26) to H₂-10′ (δ _H 2.51) and H-12′(δ _H 5.09) and HMBC correlations from H₃-14′ (δ _H 1.55) and H₃-15′ (δ _H 1.63) to C-12′ (δ _C 123.9) and C-13′ (δ _C 132.2), and from H₂-11′ and H₂-10′ to C-9′ (δ _C 171.2). Since the carboxyl group (C-9′) in the side chain of part B needs form an ester with the phenolic hydroxyl group to meet the molecular formula requirement, the position of the side chain in part B linkage can be determined by the following evidence. In the 2D NMR experiments, the HMBC correlations of 1-OH (δ _H 10.75)/C-1 (δ _C 153.0), 4-OH (δ _H 9.17)/C-4 (δ _C 149.4), and 4′-OH (δ _H 9.50)/C-4′ (δ _C 139.5) and ROESY correlations between 1-OH with H-6, 4-OH with H-3, and 4′-OH with H-6 led to the determination of the side chain linkage at the C-1′ position. This conclusion was further secured by the ROESY correlation (Figure 3) between H₂-10′ and H-6′. Thus, the 2D structure of 1 was assigned.

In the ROESY spectrum, the correlation between H-13 and H₂-16 demonstrates that the Δ ¹³⁽¹⁴⁾ double bond is E configuration. Dimercochlearlactone A (1) was found to be a racemate, the separation by chiral-phase HPLC afforded enantiomers (+)-dimercochlearlactone A (1) and (–)-dimercochlearlactone A (1). Computational ECD spectral methods at time-dependent density functional theory (TDDFT) were employed to define the absolute configurations of (+)-dimercochlearlactone A (1) and (–)-dimercochlearlactone A (1). Due to the structure flexibility of dimercochlearlactone A (1), the model compound (1a) was constructed to ECD calculations. The result showed that the experimental CD spectrum of (+)-dimercochlearlactone A (1) exhibited similar Cotton effects with calculated ECD spectrum (Figure 4) of (9R)-1a. Accordingly, the absolute configurations as 9R for (+)-dimercochlearlactone A (1) and 9S for (–)-dimercochlearlactone A (1) were determined.

The NMR data of dimercochlearlactone B (2) resemble those of dimercochlearlactone A (1) revealing that the structure of dimercochlearlactone B (2) similar to that of dimercochlearlactone A (1). Only difference appears at the their side chains, which is a 7-carbon side chain in dimercochlearlactone A (1) was attached an isopentenyl to form a 12-carbon side chain in dimercochlearlactone B (2), supporting by the ¹H-¹H COSY correlations from H₂-16′ (δ _H 2.00) to H₂-15′ (δ _H 1.92) and H-17′ (δ _H 5.04) and the HMBC correlations (Figure 2) from H₃-19′ (δ _H 1.54) and H₃-20′ (δ _H 1.62) to C-17′ (δ _C 123.9) and C-18′ (δ _C 130.0), from H₂-14′ (δ _H 1.56) to C-12′ (δ _C 122.2), C-13′ (δ _C 135.8) and C-15′ (δ _C 39.0). In the ROESY experiment, correlations between H-13 and H₂-16 and between H-12′ and H₂-15′ suggest that both double bonds Δ ¹³⁽¹⁴⁾ and Δ ^12′(13′) are E-from configurations (Figure 3). Racemic dimercochlearlactone B (2) was separated by chiral HPLC to yield (+)-dimercochlearlactone B (2) and (–)-dimercochlearlactone B (2). Their absolute configurations were deduced as 9R for (+)-dimercochlearlactone B (2) and 9S for (–)-dimercochlearlactone B (2) by using the above-mentioned ECD calculations (Figure 4).

Compounds 1 and 2 bear a same skeleton, which are different from the previously isolated Ganoderma meroterpenoids, a plausible pathway for the biogenesis of 2 was proposed (Figure 5). At first, fornicin C (Niu et al., 2006) undergoes a series of oxidation, ring formation, and reduction reactions to form intermediates A and B, respectively, which further form intermediate C via aldol condensation reaction. Intermediate C undergoes a reduction and substitution reaction to form D, which then give hemiacetal E via a substitution addition reaction. After a decarboxylation reaction, hemiacetal E can produce intermediate F. Finally, F undergoes intermediates G and H through dehydration and C-C bond cracking to form 2.

Dimercochlearlactone C (3) has the molecular formula C₄₂H₅₄O₇ (16 degrees of unsaturation) deduced by the HRESIMS analysis at m/z 669.3777 [M-H]^－(calcd for C₄₂H₅₃O₇, 669.3797). The ¹H NMR spectrum of dimercochlearlactone C (3) exhibits signals for two typical ABX systems (δ _H 7.16, d, J = 3.0 Hz, H-3; δ _H 7.06, dd, J = 8.9, 3.0 Hz, H-5; δ _H 7.01, d, J = 8.9 Hz, H-6; δ _H 6.57, d, J = 2.7 Hz, H-3′; δ _H 6.54, dd, J = 8.6, 2.7 Hz, H-5′; δ _H 6.51, d, J = 8.6 Hz, H-6′). It was found 42 carbon signals including 6 methyl, 11 methylene, 11 methine, and 14 nonprotonated carbons (11 aromatic including 4 oxygenated, 1 oxygenated aliphatic, 1 ketone, and 1 carbonyl) by analyzing its ¹³C NMR and DEPT spectra. Like compound dimercochlearlactone A (1), the NMR data of dimercochlearlactone C (3) suggest a meroterpenoid dimer. The data of part A are very similar to those of ganotheaecolumol A (Luo et al., 2018), differing in that C-20 is a hydroxymethylene in ganotheaecolumol A, while the same position in part A of dimercochlearlactone C (3) is a methyl group. This deduction is supported by the HMBC correlations between H₃-20 (δ _H 1.58) with C-18 (δ _C 125.5), C-19 (δ _C 132.5), and C-21 (δ _C 26.1).

The analysis of 2D NMR spectra (See Supplementary Figures S30–S33) of dimercochlearlactone C (3) reveals that the structure of part B is similar to that of ganomycin F (Cheng et al., 2018). Thus, there are four possibilities for the connection between part A and part B of dimercochlearlactone C (3), C-4-O-C-1′, C-4-O-C-4′, C-9-O-C-1′, and C-9-O-C-4′. In the ¹H NMR spectrum, signals of two phenolic hydroxyl groups (δ _H 9.51, s and δ _H 9.41, s) are observed, it could be concluded that the connections between part A and part B in dimercochlearlactone C (3) are C-9-O-C-1′ or C-9-O-C-4′. Furthermore, the ROESY correlations (observed in DMSO-d ₆) between 4-OH with H-3 and 4′-O with/H-3′ are indicative of phenolic hydroxyl attaching to C-4 and C-4′, which indicate that C-9 and C-1′ are connected via oxygen atom to form phenolic ester. The ROESY correlation between H-8′ (δ _H 5.13) and H₂-11′ (δ _H 2.14) suggests that the double bond Δ ^8′(9′) is Z-form configuration. Furthermore, ROESY correlations between H-13 with H₂-16 and H-13′ with H₂-16′ demonstrate that both Δ ¹³⁽¹⁴⁾ and Δ ^13′(14′) double bonds are E configuration. Racemic dimercochlearlactone C (3) was segregated into (+)-dimercochlearlactone C (3) and (–)-dimercochlearlactone C (3) by using chiral HPLC. Since the calculated ECD curve of (9S)-3a (Model structure) agrees with the experimental CD spectrum of (+)-dimercochlearlactone C (3) (Figure 4), the absolute configurations at the stereogenic center were established as 9S for (+)-dimercochlearlactone C (3) and 9R for (–)-dimercochlearlactone C (3).

The molecular formula of dimercochlearlactone D (4) was deduced as C₄₂H₅₂O₇ by its negative HRESIMS. In ¹H NMR spectrum of dimercochlearlactone D (4), the signals at (δ _H 6.70, d, J = 8.5 Hz, H-6; 6.61, d, J = 8.5 Hz, H-5; δ_H 8.00, d, J = 3.0 Hz, H-3′; δ_H 6.57, d, J = 3.0 Hz, H-5′) suggest that a 1,2,3,4-tetrasubstituted benzene ring and a 1,3,4,5-tetrasubstituted benzene ring in the structure of dimercochlearlactone D (4). The ¹³C NMR and DEPT spectra of dimercochlearlactone D (4) show 42 carbons including 6 methyl, 9 methylene, 12 methine, and 15 nonprotonated carbons (14 aromatic including 4 oxygenated and 1 carbonyl). The structure of dimercochlearlactone D (4) was mainly determined by 2D NMR spectra. The observation correlations from H₂-12 to H₂-11 and H-13, and from H₂-17 to H₂-16 and H-18 in ¹H-¹H COSY spectrum, along with the HMBC correlations from H₃-20 and H₃-21 to C-18 and C-19, from H₃-20 to C-21, from H₃-15 and H-16 to C-13 and C-14, and from H₃-15 to C-16 indicate the presence of two isoprenyl moieties. Another isoprenyl residue is supported by the observation of ¹H-¹H COSY correlation between H-7 with H-8 and the HMBC correlations from H-10 and H-11 to C-8 and C-9, from H-10 to C-11, and from H-7 to C-9. Further observation of HMBC correlations from H-7 and H-8 to C-2 suggests that C-7 is connected to C-2. Similarly, the ¹H-¹H COSY correlations from H₂-7′ to H-8′, from H₂-12′ to H₂-11′ and H-13′, and from H₂-17′ to H₂-16′ and H-18′ along with the HMBC correlations from H₃-20′, and H₃-21′ to C-18′ and C-19′, from H-20′ to C-21′, from H₃-15′ and H-16′ to C-13′ and C-14′, from H₃-15′ to C-16′, from H-10′ and H-11′ to C-8′ and C-9′, and from H-10′ to C-11′ indicates substructure consisting of three isoprenyl groups in part B of 4. Moreover, the correlation from H-6′ to C-7′ in HMBC spectrum suggests that the side chain and benzene ring of part B are linked via C-7′-C-6′. Finally, the two meroterpenoids are linked via C-2′-C-3 and C-8-O-C-1′ supported by the key HMBC correlations from H-3′ to C-2 and from H-8 to C-1′. The ROESY correlations from H-8 to H-10, from H-13 to H-16, and from H-13′ to H-16′ indicate that three double bonds (Δ ⁸⁽⁹⁾, Δ ¹³⁽¹⁴⁾, and Δ ^13′(14′)) are E-form, and that between H-8′ and H₂-11′ suggests Δ ^8′(9′) double bond is Z form. Racemic 4 was submitted to chiral HPLC to afford their enantiomers. The absolute configurations were determined to be 7S for (–)-4 and 7R for (+)-4 by using computational ECD methods (Figure 4).

The NMR data of dimercochlearlactone E (5) are similar to those of the known spirocochlealactone A (Qin et al., 2018). Careful analysis of their structures showed that compound 5 is formed by the reduction of nonprotonated carbon (δ _C 88.3) and a carbonyl group (δ _C 173.0) in spirocochlealactone A to a methine (δ _C 40.1) and an aldehyde group (δ _C 194.0), respectively. The ¹H-¹H COSY correlation between H-7′ and H-8′, and the HMBC correlations from H-7′ to C-8′ and C-9′, and from H-8′ to C-9′, C-10′ and C-11′ supports the above conclusions. In the ROESY spectrum, correlations from H-8′ to H-10′, from H-13′ to H-16′, and from H-13 to H-16 indicate that three double bonds (Δ ^8′(9′), Δ ^13′(14′), and Δ ¹³⁽¹⁴⁾) are E-form. Furthermore, the relative configuration of dimercochlearlactone E (5) was assigned as 9R*,7S*, gaining support from the ROESY correlation of Hb-8/H-7′. Dimercochlearlactone E (5) was also separated by chiral HPLC to afforded (+)-dimercochlearlactone E (5) and (–)-dimercochlearlactone E (5). Their absolute configurations were assigned as 9R,7S for (+)-dimercochlearlactone E (5) and 9S,7R for (–)-dimercochlearlactone E (5) by comparing their CD curves with the calculated ones. Thus, the structure of 5 was determined.

Dimercochlearlactone F (6) has the molecular formula C₄₂H₅₄O₇ based on HRESIMS analysis (m/z 783.3727 [M + CF₃COO]^－; calcd for 783.3725). The ¹H NMR spectrum of 6 contains signals for one typical ABX system (δ _H 7.11, d, J = 2.7 Hz, H-3; δ _H 6.99, dd, J = 8.9, 2.7 Hz, H-5; δ _H 6.80, d, J = 8.9 Hz, H-6) and a 1,2,4,5-tetrasubstituted benzene ring (δ_H 6.58, s, H-3′ and δ_H 7.08 s, H-5′). The ¹³C NMR and DEPT spectra contain the resonances for 42 carbons including 6 methyl, 11 methylene, 10 methine, and 15 nonprotonated carbons (12 aromatic including 4 oxygenated, 1 oxygenated aliphatic, 2 ketones). The above signals suggest that compound 6 is also a meroterpenoid dimer, and its structure consists of parts A and B. The substructure of part A in 6 is similar to that of part A in 1 as they have very similar NMR data. The substructure of part B is very similar to ganomycin F (Cheng et al., 2018). The difference is that the 3-position of the benzene ring in part A is connected to the other additional substructures, which is a hydrogen atom in ganomycin F. The HMBC correlation from H-6′ to C-10 suggests that C-10 is connected to C-5′. Although no HMBC correlations are observed to support C-9-O-C-3 fragment, the presence of ring A is confirmed due to the observation of characteristic chemical shift of C-4 (δ _C 166.2) in the benzene ring (Ring B). Same phenomenon was observed in other such kind of benzofuran structures, such as cochlearol I and spiroapplanatumines A–Q (Luo et al., 2017; Wang et al., 2019). The ROESY correlations from H-13′ to H-16′ and from H-13 to H-16 indicate that both double bonds (Δ ^13′(14′) and Δ ¹³⁽¹⁴⁾) are E-form. Further correlation from H-8′ to H-11′ observed in ROESY spectrum suggests a E-form double bond (Δ ^8′(9′)). Chiral separation by HPLC afforded (+)-6 and (–)-6. Their absolute configurations were determined as 9S and 9R, respectively, when comparing their experimental CD curves with the calculated ones. As a result, the structure of 6 was assigned.

Dimercochlearlactone G (7) has a molecular formula C₄₂H₅₂O₉ based on its HRESIMS, ¹³C NMR, and DEPT data. Two ABX aromatic coupling systems at δ _H 7.07 (d, J = 3.0 Hz, H-3), 7.01 (dd, J = 8.9, 3.0 Hz, H-5), 6.82 (d, J = 8.9 Hz, H-6), 6.35 (d, J = 2.9 Hz, H-3′), 6.57 (dd, J = 8.6, 2.9 Hz, H-5′), and 6.69 (d, J = 8.6 Hz, H-6′) were observed in its ¹H NMR spectrum. Its ¹³C NMR and DEPT spectra reveal the presence of 42 carbons ascribed to 6 methyl, 9 methylene, 12 methine, and 15 nonprotonated carbons. Analysis of the NMR data of 7 found that part A is similar to dayaolingzhiol K (Zhang et al., 2021). The main difference is that the chemical shift of C-19 is 71.5 ppm in dayaolingzhiol K, while in compound 7 the chemical shift of C-19′ is downfield to 84.5 ppm. In addition, the NMR data of part B resembles that of ganodercin A implying that they have similar structure (Qin et al., 2021). Since no tailing behavior is observed in TLC, the carboxyl group in the structure of part B must be esterified. The structure of 7 is further confirmed by ¹H NMR, HMBC, and ROESY spectra in DMSO-d ₆. In the ¹H NMR spectrum, there are four free phenolic hydroxyl signals at δ _H 11.20 (s, 1-OH), 9.18 (s, 4-OH), 9.21 (s, 1′-OH), and 8.77 (s, 4′-OH). The ROESY correlations from 1-OH to H-6, from 4-OH to H-3, from 1′-OH to H-6′, and from 4′-OH to H-3′ fix position of phenolic hydroxyl. Thus, it can be inferred that the carboxyl group can satisfy the requirement of molecular weight only when it is esterified with C-19′. This conjecture was further secured by the ROESY correlation between H₃-21′ and H₂-11 and the above-mentioned downfield chemical shift of C-19′. Therefore, the planar structure of 7 was deduced. The ROESY correlation between H-8 and H-11 suggests that the double bond Δ ⁸⁽⁹⁾ is Z-form configuration. Furthermore, ROESY correlations from H-13 to H-16 and from H-13′ to H-16′ imply Δ ¹³⁽¹⁴⁾ and Δ ^13′(14′) double bonds are both E-form. The absolute configurations of 7 were assigned as 7′R for (+)-7 and 7′S for (–)-7 by comparing experimental CD spectra with those of (+)- and (–)-zizhine A (Cao et al., 2016).

The molecular formula of dimercochlearlactone H (8) was assigned as C₄₂H₅₂O₆ by the analysis of its HRESIMS. Its ¹H NMR spectrum contains one ABX aromatic coupling system with the signals at δ _H 6.57 (d, J = 3.0 Hz, H-3), 6.88 (dd, J = 8.9, 3.0 Hz, H-5), and 7.12 (d, J = 8.9 Hz, H-6). The signals at δ _H 6.44 (d, J = 2.9 Hz, H-3′) and δ _H 6.77 (d, J = 2.9 Hz, H-5′) demonstrate the existence of a 1,3,4,5-tetrasubstituted benzene ring in 8. Analysis of its ¹³C NMR and DEPT spectra resulted in 42 carbons, including 6 methyl, 10 methylene, 11 methine, and 15 nonprotonated carbons (13 sp2 including 4 oxygenated, 1 oxygenated aliphatic, and 1 carbonyl). Compound 8 is a meroterpenoids dimer consisting of two parts. The data of part A resemble those of ganomycin I (El Dine et al., 2009); the difference is that C-7 (δ _C 81.4) is a nonprotonated carbon in 8. The HMBC correlations from H-3 and H-8 to C-7 and from H-8 to C-9 and C-10 support the structure of part A. The structure of part B is very similar to ganomycin F, with the difference appearing at C-2′ being a nonprotonated carbon. The HMBC correlations from H-3′ to C-7 and from H-8 to C-2′ suggest that C-7 is connected to C-2′. The molecular weight and the unsaturation of the molecule need to form another ring to be satisfied. There are three possible ring formations, which are C-1-O-C-1′ C-1′-O-C-10 or C-1-O-C-10. In the ¹H NMR spectrum (DMSO-d ₆), there are two free phenolic hydroxyl signals at δ _H 9.42 (s, 4-OH), 9.29 (s, 4′-OH). The ROESY correlations from 4-OH to H-3 and H-5 and from 4′-OH to H-3′ and H-5′ fix phenolic hydroxyl at C-4′ and C-4′, confirming the formation of ring C. Furthermore, the ROESY correlations from H-13′ to H-16′ and from H-13 to H-16 suggest that double bonds Δ ^13′(14′) and Δ ¹³⁽¹⁴⁾ are E-form. Additional ROESY correlation between H-8′ and H-11′ suggests a E-form double bond Δ ^8′(9′). The absolute configurations of 8 were determined to be 7R for (+)-8 and 7S for (–)-8 by using ECD calculation methods.

The molecular formula of 9 was specified as C₃₂H₃₄O₈ based on the analysis of its positive HRESIMS ([M + Na]⁺, m/z 569.2154, calcd 569.2151). The ¹H NMR spectrum exhibits two typical ABX spin systems δ _H 7.14, d, J = 2.9 Hz, H-3; 7.08, dd, J = 8.9, 2.9 Hz, H-5; 6.84 d, J = 8.9 Hz, H-6; δ _H 6.98, d, J = 8.9 Hz, H-6′; 6.85, dd, J = 8.9, 2.9 Hz, H-5′; 6.45, d, J = 2.9 Hz, H-3′. The ¹³C NMR and DEPT spectra of 9 display 4 methyl, 5 methylene, 9 olefifinic methine, 14 nonprotonated carbons (11 aromatic including 4 oxygenated, 2 carbonyl, and a ketone group). These data are very similar to those of spirocochlealactone B (Qin et al., 2018). The difference between 9 and spirocochlealactone B is that a sesquiterpenoid residue in spirocochlealactone B is disappeared instead of a monoterpenoid residue in 9. This inference is supported by the ¹H-¹H COSY correlations from H₂-12′ to H₂-11′ and H-13′ and the HMBC correlations from H₃-15′ and H₃-16′ to C-13′ and C-14′, from H-8′ to C-9′, C-10′ and C-11′, and from H-11′ to C-12′ and C-10′. The relative configurations at the chiral centers were determined to be 9S*,7′S* based on the observation of the ROESY correlation H-8′(δ _H 7.06)/Ha-11 (δ _H 2.21). The absolute configurations of 9 were determined as 9R,7′R for (+)-9 and 9S,7′S for (−)-9 by comparing the experimental CD spectra with those of (+)-spirocochlealactone B and (−)-spirocochlealactone B (Qin et al., 2018). Thus, the structure of 9 was identified and named as dimercochlearlactone I.

The molecular formula of compound dimercochlearlactone J (10) is similar to that of 9. Careful examination of 2D NMR (see Supplementary Figures S81–S84, S91−S94) data between 10 and 9 indicates that they have same planar structure. The observed ROESY correlation (see Supplementary Figure S94) of H-8′ (δ _H 7.16)/H-8 (δ _H 3.68, 3.07) in 10 suggests 9R*,7S′* relative configurations at chiral centers. The absolute configurations of 10 were determined as 9S,7′R for (+)-dimercochlearlactone J (10) and 9R,7′S for (−)-dimercochlearlactone J (10) by comparing their experimental CD spectra to (+)-spirocochlealactone A and (−)-spirocochlealactone A (Qin et al., 2018). Thus, the structure of 10 was assigned.

Compound 11 was identified as spirocochlealactone A by comparing its NMR and MS data with the literature data (Qin et al., 2018). This compound has been isolated by us from 10 kg of G. cochlear, and in this experiment, it was isolated from G. lucidum.

To investigate the anti-TNBC effects of the isolated compounds, we used the MDA-MB-231 cells for our analyses. All 22 dimer meroterpenoid enantiomers were evaluated for their suppressive effect toward MDA-MB-231 cells. It was found that (+)-2, (–)-2, (–)-3, (+)-11, and (–)-11 significantly decreased the cell viability in MDA-MB-231 cells (Figure 6A). Moreover, the morphological and density changes were observed in MDA-MB-231 cells upon exposure of the compounds (Supplementary Figure S98). To further analyze the effects of the compounds on MDA-MB-231 cells, the dose-response studies were performed. The cell viability of MDA-MB-231 cells was substantially inhibited by treatment with compounds for 48 h in a dose-dependent manner. As shown in Figure 6B, similar to the positive control (cisplatin), compound treatment dose-dependently inhibits MDA-MB-231 cells growth. The IC₅₀ of compounds for MDA-MB-231 cells are 28.18, 25.65, 11.16, 8.18, and 13.02 μM, respectively. The remaining compounds showed negligible inhibitory effects on cell viability at 20 μM (Figure 6A). Interestingly, although all the remaining compounds have rather low cytotoxicities toward MDA-MB-231 cells, ten of the isolates (+)-5, (–)-5, (+)-6, (–)-6, (+)-7, (–)-7, (+)-8, (–)-8, (+)-10, and (–)-10 significantly inhibit the migration ability of MDA-MB-231 cells (Figure 7), suggesting that they might be promising lead compounds for the development of anti-cancer drugs against metastasis of TNBC.

To conclude, this study resulted in the isolation of ten pairs of novel meroterpenid dimers and one pair of known compounds from Ganoderma species. Biological results revealed the importance of (+)-2, (–)-2, (–)-3, (+)-11, and (–)-11 in the development of the anti-TNBC drugs. Furthermore, (+)-5, (–)-5, (+)-6, (–)-6, (+)-7, (–)-7, (+)-8, (–)-8, (+)-10, and (–)-10 significantly inhibit the migration ability of MDA-MB-231 cells, thereby providing promising compounds for the development of anti-TNBC drugs.