Otteacumiene G (1), isolated as colorless crystals, was assigned the molecular formula of C19H20O4 by the HRESIMS at m/z 311.1287 [M − H]− (calculated for 311.1289), suggesting ten degrees of unsaturation. The IR spectrum showed a broad hydroxyl absorption band at 3426 cm−1 and aromatic ring absorption band at 1613, 1596, 1516, 1502 cm−1. The 13C NMR (DEPT) spectra (Table 2) exhibited 19 carbon signals, including two benzene rings (δC 146.6, C-1; 151.3, C-2; 118.4, C-3; 134.5, C-4; 120.7, C-5; 117.0, C-6 and 158.0, C-1′; 123.7, C-2′; 133.5, C-3′; 140.7, C-4′; 131.3, C-5′; 124.8, C-6′) and an oxygenated unsaturated heptane (δC 76.3, C-7; 76.9, C-8; 36.3, C-9; 128.9, C-10; 132.0, C-11; 37.1, C-12; 35.9, C-13). The key HMBC correlations from H-7 to C-3/C-4/C-5 and H2-13 to C-3′/C-4′/C-5′ suggested 1 was a diarylheptanoid derivative. By combining HRESIMS and comparing NMR (Tables 1 and 2), compound 1 was identified to share the same diarylether-type cyclic diarylheptanoid skeleton as the known otteacumiene C [7], with only variation occurring in the heptane chain, a hydroxyl group at C-7 (δC 76.3). Furthermore, this deduction was confirmed by key 1H-1H COSY correlations of H-7 (δH 4.33, d, J = 6.8 Hz)/H-8 (δH 3.34, m)/H2-9/H-10 (δH 4.84, overlapped)/H-11 (δH 5.15, dt, J = 15.3, 6.1 Hz) /H2-12/H2-13 (Fig. 2). Besides, the 1H NMR signals at δH 4.84 (H-10) and δH 5.15 (H-11) also suggested the existence of a trans double bond [9]. Consequently, the planar structure of 1 as a diarylether-type cyclic diarylheptanoid was unequivocally determined.

Key 1H–1H COSY and HMBC correlations of 1–7

The cotton effects were not observed in the ECD spectrum of 1 and it was separated into two peaks by chiral-phase column (Fig. S1). Thus, 1 was a racemic mixture [10]. Fortunately, the crystals of 1 (Fig. 3) were obtained, which showed the relative configuration of 1. Following the chiral resolution of this compound, the absolute configurations of the resultant (+)-1 and (−)-1 enantiomers were unequivocally established via quantum chemical calculation. The results demonstrated that the calculated ECD spectra of the enantiomeric pair exhibited excellent agreement with the experimental data, allowing unambiguous assignment of the absolute configurations as 7S,8S for (+)-1 and 7R,8R for (−)-1 (Fig. 4).

X-ray crystallographic structures for 1, 3 and 5

Comparison of experimental and calculated ECD spectra of compounds 1–4

Otteacumiene H (2), isolated as colorless crystals, had its molecular formula C19H20O4, which was evidenced by the HRESIMS at m/z 311.1287 [M − H]− (calculated for 311.1289). Compound 2 shared the same planar structure as 1, as established by the detailed 1H-1H COSY and HMBC correlations (Fig. 2). The trans-configuration of the double bond was confirmed by the coupling constant of the olefinic proton δH 5.15 (1H, dt, J = 15.3, 7.5 Hz, H-11). Comparative analysis of the 1H and 13C NMR data (Tables 1 and 2) revealed a key difference at the chiral center C-8 (δH 3.64, m, δC 75.0 in 2; δH 3.34, m, δC 76.9 in 1), suggesting that 2 was a stereoisomer of 1. As the vicinal dihydroxyl groups in compound 1 adopt a trans configuration, the corresponding groups in compound 2 were cis-configured. With only two chiral centers (C-7 and C-8) in this planar structure, there existed two relative configurations and two possible absolute configurations, excluding the already established 7S,8S and 7R,8R configurations of (+)-1 and (−)-1, respectively. Further analysis of the ECD spectrum and chiral HPLC data (Fig. S14) confirmed that compound 2 existed as a racemate. Through chiral resolution followed by comparison of experimental and calculated ECD curves, the absolute configurations of (+)-2 and (−)-2 were assigned as 7S,8R and 7R,8S, respectively (Fig. 4).

Otteacumiene I (3) was purified as colorless crystals. Its molecular formula was determined as C19H20O4 by its HRESIMS (m/z 311.1286, [M − H]−, calculated for 311.1289). A detailed comparison of the HRESIMS and NMR data (Tables 1 and 2) of 3 with those of 1 revealed that they were close structural analogs. The differences from 1 were the positional variations of hydroxyl groups (δC 72.4, C-11) and double bond (δC 134.7, C-9; 134.9, C-10) in 3, as confirmed by the key 1H-1H COSY correlations of H2-7 (δH 2.71, dd, J = 13.8, 3.1 Hz; 2.44, dd, J = 13.8, 9.5 Hz)/H-8 (δH 3.99, m) /H-9 (δH 5.16, dd, J = 15.8, 4.7 Hz)/H-10 (δH 5.51, dd, J = 15.8, 7.2 Hz)/H-11 (δH 4.04, m)/H2-12/H2-13 and HMBC correlations from H-7 to C-3 (δC 118.7), C-4 (δC 130.2), and C-5 (δC 123.8); from H2-13 to C-3′ (δC 132.4), C-4′ (δC 141.1), and C-5′ (δC 131.4) (Fig. 2).

The appearance of two peaks on chiral-phase column (Fig. S27) and the absence of an obvious cotton effect in the ECD spectrum indicated that compound 3 was a racemate. Furthermore, the relative configurations of 3 were proposed by the result of single-crystal X-ray diffraction (Fig. 3). After chiral separation of (±)-3, the absolute configurations of (+)-3 and (−)-3 were determined as 8R,11S and 8S,11R by comparing the experimental and calculated ECD curves (Fig. 4).

Otteacumiene J (4), a racemic mixture (Fig. S40), was purified as white powder. It had the same molecular formula as compound 3, as deduced by HRESIMS (m/z 311.1285, [M − H]−, calculated for 311.1289). Comparative analyses of 1H and 13C NMR data (Tables 1 and 2) between 3 and 4 revealed that 4 was an epimer of 3, which was supported by the changes in chemical shift of C-7 (δC 42.7), C-8 (δC 74.0), C-9 (δC 137.1), C-10 (δC 139.4), and C-11 (δC 73.1). The relative configurations of 3 were proposed by the result of single-crystal X-ray diffraction (Fig. 3), indicating that two hydroxyls (C-8 and C-11) were on the opposite side. Therefore, two hydroxyls (C-8 and C-11) of 4 placed the co-facial orientation. The ECD spectrum and chiral separation chromatogram (Fig. S40) of (±)-4 clearly demonstrated that it was a racemate. Compound (±)-4 was successfully separated into (+)-4 and (−)-4 enantiomers whose absolute configurations were determined as 8R,11R and 8S,11S, respectively (Fig. 4).

Otteacumiene K (5), colorless needle crystals, was deduced to have a molecular formula of C19H18O3 according to the [M − H]− ion at m/z 293.1178 (calcd. for C19H17O3, 293.1178) in negative HRESIMS, suggesting 11 degrees of unsaturation. By comparing its 1D NMR data (Tables 1 and 2) with strained cyclic diarylheptanoids tedarene B [11], the differences between 5 and tedarene B were that there were hydroxyl groups attached to C-1 (δC 153.7) on the benzene ring and to C-12 (δC 79.9) on the heptane chain, as proven by the key 1H-1H COSY correlations of H2-7 (δH 2.86, m, H-7a; 2.72, m, H-7b) /H2-8 (δH 2.22, m, H-8a; 1.98, m, H-8b) /H-9 (δH 3.35, m) /H-10 (δH 4.06, dd, J = 17.1, 7.1 Hz) /H-11 (δH 5.36, dd, J = 17.1, 7.6 Hz) /H-12 (δH 3.81, m) /H2-13 (δH 2.49, t, J = 12.4 Hz, H-13a; 3.10, dd, J = 12.4, 5.0 Hz, H-13b) and HMBC correlations from H-7 to C-3, C-4, and C-5; from H2-13 to C-3′, C-4′, and C-5′ (Fig. 2). The downfield aromatic quaternary carbon signals at C-2 (δC 125.3) and C-2′ (δC 120.8) suggested that 5 could be a biaryl-type cyclic diarylheptanoid derivative (Table 2), evidenced by key HMBC correlations from H-3′ (δH 6.38, d, J = 2.0 Hz) to C-2. Combining degrees of unsaturation and analysis of NMR (Tables 1 and 2), it was inferred that 5 possessed one additional cyclic system, proved by key HMBC correlations (Fig. 2) from H-9 to C-3. As a consequence, the planar structure of 5 was unambiguously determined. The result of single-crystal X-ray diffraction of 5 (Fig. 3) confirmed the relative configuration and also indicated that (±)-5 was a racemate. However, due to the limited amount of trace samples obtained, it was unfeasible to obtain optically pure enantiomers of (±)-5. Therefore, the absolute configurations of (±)-5 could not be determined.

Diarylheptanoids featuring a disubstituted heptane chain, isolated from O. acuminata, were frequently obtained as racemic mixtures. The pronounced conformational flexibility of these molecules posed challenges in definitively assigning their relative configurations using NOESY experiments. To address this, these enantiomers were subjected to chiral separation to obtain a pair of optically pure enantiomers. Subsequently, the absolute configurations of the enantiomers were determined through comparative analysis of experimental and calculated ECD spectra. Otteacumienes G–K (1–5) were identified as enantiomeric isomers. Based on structural analysis, we suppose a potential mechanism for the formation of these enantiomers: the double bond ∆(7,8) on the heptane chain of 1 and 2 underwent oxidative cyclization to form a ternary oxygen ring, followed by a ring-opening reaction that results in the generation of enantiomers.

Otteacumiene L (6) was isolated as a white powder and its molecular formula was determined as C25H28O8 by the HRESIMS at m/z 479.1682 [M + Na]+ (calculated for 479.1682), suggesting 12 degrees of unsaturation. The 1H and 13C NMR (Table 3) showed the presence of a β-glucopyranosyl part (δH 5.02, d, J = 7.9 Hz, H-1″) [12]. Analysis of the HRESIMS and 1D NMR data (Table 3) of compound 6 indicated that the aglycone of 6 shared identical planar structure with otteacumiene A, a known compound previously characterized by our group [8], which was confirmed by the key 1H-1H COSY correlations of H2-7 (δH 2.37, m)/H-8 (δH 4.24, t, J = 9.8 Hz) /H-9 (δH 5.40, t, J = 11.5 Hz) /H-10 (δH 5.91, t, J = 11.5 Hz) /H-11 (δH 5.32, dd, J = 15.3, 11.5 Hz) /H-12 (δH 6.07, dt, J = 15.3, 4.7 Hz) /H2-13 (δH 3.52, overlapped), accompanied by HMBC correlations from H-7 to C-3 (δC 118.7), C-4 (δC 137.7), and C-5 (δC 122.9); from H-13 to C-3′ (δC 132.5), C-4′ (δC 139.2), and C-5′ (δC 134.8) (Fig. 2). Besides, key HMBC correlations from H-1″ to C-1, combined with the chemical shift changes at C-1 (δC 145.7), C-2 (δC 154.2), C-4 (δC 137.7), and C-6 (δC 118.2) proved that the glycosyl group was attached to C-1 [13]. Consequently, the planar structure of 6 was determined.

In order to determine the absolute configuration of 6, it was synthesized by glycosylation. Using Ca(OH)2 as a base, otteacumiene A was chosen as a substrate to react with α-D-fluoroglucose in aqueous solvent at room temperature for 1 h (Scheme 1) [14]. Then, O-β-glc-otteacumiene A was purified by semi-preparative HPLC (MeOH/H2O, 50:50). Comparative analysis of the 1H/13C NMR and HRESIMS data, combined with superimposed ECD spectra, revealed that the aglycone of compound 6 had the same absolute configuration as otteacumiene A [8]. Therefore, this compound was (8S)-otteacumiene A 1-O-β-D-glucopyranoside, named otteacumiene L.

Glycosylation reaction of otteacumiene A

Otteacumiene M (7) was isolated as yellow oil and its molecular formula was determined as C22H22O7 by the HRESIMS at m/z 421.1261 [M + Na]+ (calc. for 421.1263), suggesting 12 degrees of unsaturation. Compound 7 was a new natural product. Its 1H NMR spectrum was completely identical to that of synthetic (E)-α-(3,4-methylenedioxybenzylidene)-β-(3,4,5-trimethoxybenzyl)-γ-butyrolactone [15], which was confirmed by the key COSY correlations of H2-9′/H-8′/H2-7′, associated with HMBC correlations from H-7 (δH 7.52, s) to C-2 (δC 108.5), C-6 (δC 126.8), C-8 (δC 126.0), and C-9 (δC 172.8); from H2-7′ (δH 3.01, dd, J = 14.9, 4.8 Hz, H-7′a; 2.63, dd, J = 14.9, 9.9 Hz, H-7′b) to C-8′ (δC 40.3), C-9′ (δC 70.2), C-8 (δC 126.0), C-1′ (δC 133.9), and C-2′/C-6′ (δC 106.1); and from -OCH2O- (δH 6.03, s) to C-3 (δC 149.5) and C-4 (δC 148.6) (Fig. 2). However, the absolute configuration of the synthetic product mentioned above was not determined. Subsequently, it was found that the known jatrophan (8′S, + 87) and isosuchilactone (8′R, − 83.3) were a pair of enantiomers [16]. By comparing their NMR data (Table 4), it was found that compound 7 was a methoxylated derivative of these two compounds at C-5′ (δC 153.6). These compounds had only one chiral center (C-8′). The absolute configuration of 7 was speculated to be 8′S by comparison of its rotation ([α] 20D + 27.7 (c 0.141, MeOH) with jatrophan (8′S, + 87) and isosuchilactone (8′R, − 83.3) [16].

The inhibitory activities of compounds 1, 2, 4, 6, and 7 against both α-glucosidase and PTP1B were evaluated. As summarized in Table 5, these compounds exhibited varying degrees of inhibitory activity against α-glucosidase at a concentration of 50 μM. However, none of the tested compounds displayed detectable PTP1B inhibitory effects. An analysis of the structures and α-glucosidase inhibitory activities was conducted by comparing the diarylheptanoids in this study with those previously reported by our group [8]. The results of this analysis revealed that the presence and position of hydroxyl groups on the heptane chain had an impact on their α-glucosidase inhibitory activities. This structural-activity relationship warrants further investigation to elucidate the precise molecular mechanisms underlying the observed bioactivity.