Structure Revision of Lyciumamides A–C and Their Hypoglycemic Activities

Abstract

Lyciumamides A–C (1a–3a) are three phenolamide dimers (neolignanamides) with unusual linkage types from the fruits of Lycium barbarum reported by Gao et al.[ GaoK., ; MaD., ; ChengY., ; TianX., ; LuY., ; DuX., ; TangH., ; ChenJ., . Three new dimers and two monomers of phenolic amides from the fruits of Lycium barbarum and their antioxidant activities. J. Agric. Food Chem. 2015, 63, 1067−1075. 10.1021/jf5049222. ]. Recently, Zadelhoff et al.[ van ZadelhoffA., ; de BruijnW. J. C., ; VinckenJ.-P., . Comment on “three new dimers and two monomers of phenolic amides from the fruits of Lycium barbarum and their antioxidant activities”. J. Agric. Food Chem. 2024. 72. 6781−6786. 10.1021/acs.jafc.3c08738. ] suggested more reasonable structures with common linkage types for lyciumamides A–C, identifying them as the known compounds cannabisin F (1), grossamide (2), and grossamide K (3), respectively, by comparing the NMR data in the literature. However, the NMR data discussed by Zadelhoff et al. were insufficient to support the corrected structures, and the relative configurations of 2 and 3 were not assigned. In this report, we revised the structures of lyciumamides A–C as cannabisin F (1), grossamide (2b), and grossamide K (3b), respectively, with 1D and 2D NMR experiments and ¹³C NMR calculations. The absolute configurations of 2b and 3b were also assigned by comparing the experimental and calculated ECD spectra. Compound 3b was shown to promote preadipocyte differentiation and exhibit agonistic activity for peroxisome proliferator-activated receptor in adipocytes.

1. Introduction

Plants in the Lycium genus (Solanaceae) are well-known to produce various alkaloids, particularly phenolamides that result from the conjugation of hydroxycinnamic acid derivatives with amines. These structurally diverse phenolamides have been shown to exhibit a wide range of pharmacological and health-promoting properties, including antioxidant, antiinflammatory, anticancer, and antimicrobial activities, and protective effects against metabolic syndrome and neurodegenerative diseases. Therefore, it is significant to search for bioactive phenolamides from Lycium plants to prevent and treat chronic diseases.

In 2015, Gao et al. reported three new neolignanamides, lyciumamides A–C (1a–3a) (Figure ), which are dimers of feruloyltyramine (FerTrm) consisting of a hydroxycinnamoyl and a tyramine moiety, from the fruits of Lycium barbarum. In this report, lyciumamide A (1a) was elucidated as a dimer consisting of a FerTrm and an 8-hydroxy-FerTrm connected via a 16-O-16′ linkage; lyciumamide B (2a) was proposed as a FerTrm dimer linked via both a 16-O-16′ and a C-5-C-8′ linkages; while lyciumamide C (3a) was identified as a FerTrm derivative fused with an oxetane ring. Later, in 2023, two of these compounds (1a and 2a) were also isolated from the same plant by our group, and we reported that lyciumamide A (1a) exhibited potent α-glucosidase and PPAR-γ agonistic activities. However, Zadelhoff et al. recently published a comment proposing different structures, 1–3 (Figure ), for lyciumamides A–C (1a–3a), respectively. The linkage types of structures 1–3 are much more common for neolignanamides than those of structures 1a–3a. Neolignanamides are generally coupled by two phenolamides via only the hydroxycinnamic acid moieties, whereas the linkages via the amine moieties of the structures 1a–3a are extremely rare, and their putative biosynthetic routes are also presumed to lack chemical rationale. Therefore, the structures proposed by Zadelhoff et al. (1–3) are more reasonable. However, the structure revision of 1–3 by Zadelhoff et al. was based only on the comparison of their NMR spectroscopic data (data from Gao et al., tested in methanol-d ₄) with those of the known compounds in the literature (tested in acetone-d ₆), and there was no further solid experimental evidence available to support the corrected structures 1–3. Furthermore, the HMBC correlations described by Zadelhoff et al. were all valid for the structures proposed by Gao et al. for each compound, so it is still unconvincing to address the structures with the currently published interpretation of the HMBC data. In addition, the relative configurations of C-7 and C-8 in the benzofuran moiety of the corrected structure 2 or 3 were not assigned by Zadelhoff et al., and whether the configurations of 2 and 3 match those of the previously reported compounds need further confirmation. In this study, based on the above unresolved issues, compounds 1–3 were reinvestigated with NMR experiments and the aid of computational calculations to confirm their structures.

1.

Structures of compounds 1a–3a and 1–3.

2. Experimental Section

2.1. General Experimental Procedures

Optical rotations were recorded on a Rudolph AP-IV polarimeter. Electronic circular dichroism (ECD) spectra were obtained on an Applied Photophysics Chirascan qCD spectropolarimeter. NMR spectra were acquired using Bruker Avance III 500 spectrometers or Bruker Ascend Evo 600 spectrometers. HRESIMS data were recorded on an AB SCIEX TripleTOF 6600 mass spectrometer. Preparative HPLC was carried out using a Sepuruisi LC-52 instrument with a UV200 detector (Beijing Sepurusi Scientific Co., Ltd., China) and a YMCPack ODS-A column (250 × 20 mm, 5 μm). The column chromatography (CC) was carried out on HP-20 macroporous resin (Mistsubishi Chemical Corporation, Tokyo, Japan), Sephadex LH-20 (Amersham Pharmacia Biotech AB, Sweden), ODS (45–70 μm, Merck), and silica gel (100–200 mesh, Qingdao Marine Chemical Inc., China). TLC was performed on silica gel GF₂₅₄ (Qingdao Marine Chemical, Inc. China). Chiral separations were conducted on Daicel Chiralpak IC columns (250 × 10 mm, 5 μm) from Daicel Chemcial Industries, Ltd. (Japan).

2.2. Plant Material

The fruits of L. barbarum were collected from Golmud City, Qinghai Province, China, in September 2017 and authenticated by Prof. Suiqing Chen, School of Pharmacy, Henan University of Chinese Medicine. Its voucher specimen (20170906A) has been deposited in the School of Pharmacy, Henan University of Chinese Medicine, Zhengzhou 450046, China.

2.3. Extraction and Isolation

The dried fruits of L. barbarum (50 kg) were refluxed with 95% EtOH (200 L, 3 × 2 h). Following the removal of EtOH under reduced pressure, the residue was dispersed in H₂O. The H₂O-solution was chromatographed on an HP-20 macroporous resin column (30 × 120 cm) and eluted with H₂O, 95% EtOH, and acetone, respectively. The 95% EtOH extract (1336 g) was suspended in H₂O, followed by successive partitioning with EtOAc and n-BuOH to yield EtOAc- (184 g) and n-BuOH (386 g) soluble extracts. The EtOAc extract was subjected to a silica gel CC, eluted with petroleum ether/EtOAc (100:1, 70:1, 50:1, 30:1, 20:1, 10:1, 5:1, and 1:1) and EtOAc/MeOH (30:1, 20:1, 10:1, 5:1, 1:1, 0:1) to afford 11 fractions (A–K) on the basis of TLC analysis. Fraction I (7.12 g) was subjected to an ODS column eluted with a gradient of MeOH/H₂O (10:90–90:10) to give five fractions (I1–I5). Fraction I4 (9.15 g) was chromatographed over Sephadex LH-20 eluting with MeOH to afford two fractions I4–1–I4–2. Fraction I4–2 was further fractionated by silica gel CC with CH₂Cl₂/MeOH (50:1 to 1:1) to yield four subfractions (I4–2–1–I4–2–4). Subfraction I4–2–4 (225 mg) was purified by preparative HPLC (55% MeOH/H₂O, 7 mL/min) to yield 3 (7.3 mg, t _R = 75.2 min). The isolation procedures of compounds 1 and 2 were reported in our previous literature. Compounds (+)-2 (0.56 mg, t _R = 21.5 min), (−)-2 (0.42 mg, t _R = 31.0 min), (+)-3 (1.02 mg, t _R = 24.0 min), (−)-3 (0.89 mg, t _R = 33.0 min) were separated by using a Daicel Chiralpak IC column and an elution mixture of n-hexane/2-isopropanol (3:2) at a flow rate of 2 mL/min (Figures S32 and S33).

(7′R,8′R)-2 ((+)-2). [α]_D ²⁵ +96.3 (c 0.05, MeOH); ECD (MeOH) λ_max (Δε) 230 (−0.37), 320 (0.23) nm.

(7′S,8′S)-2 ((−)-2). [α]_D ²⁵ −91.0 (c 0.04, MeOH); ECD (MeOH) λ_max (Δε) 230 (0.78), 300 (−0.43) nm.

(7′R,8′S)-3 ((+)-3). [α]_D ²⁵ +58.2 (c 0.11, MeOH); ECD (MeOH) λ_max (Δε) 210 (0.45), 240 (−1.15), 310 (1.37) nm.

(7′S,8′R)-3 ((−)-3). [α]_D ²⁵ −54.7 (c 0.11, MeOH); ECD (MeOH) λ_max (Δε) 210 (0.46), 240 (−1.16), 310 (1.39) nm.

2.4. Quantum-Chemical NMR and ECD Calculations

The methods and details for NMR calculations with DP4+ probability analyses and ECD calculations are provided in the Supporting Information.

2.5. Bioactivity Assays

The PPAR-γ agonistic activity assay for compound 3b and its enantiomers (+)-3b and (−)-3b, the assessment of 3T3-L1 adipocyte differentiation, and the expression of PPAR-γ downstream genes were performed as previously described.

3. Results and Discussion

3.1. Structure Revision of Lyciumamides A–C

Lyciumamide A (1a) was initially elucidated by Gao et al. as a FerTrm (Feruloyltramine) dimer formed via the 16-O-16′-linkage and with one hydroxyl group at C-8. The supposed 16-O-16′-linkage was only deduced from EIMS fragment analysis, and there were no HMBC correlations observed to support this conclusion. In addition, because methanol-d ₄ was used as the testing solvent by Gao et al., the signal of the hydroxyl group at C-8 cannot be directly confirmed. The conclusion of C-8 being substituted by a hydroxy group is deduced by the authors mainly based on EIMS fragment analysis, low-field chemical shifts of C-8 (δ_C 141.5), and HMBC correlations of H-7′/C-9′, H-2′/C-7′, and H-6′/C-7′. Recently, Zadelhoff et al. corrected the structure for compound 1a as the known compound cannabisin F (1), , mainly based on the analysis of their logical biosynthetic pathways in hypothesis and comparison of the literature data. In fact, similar to the structure of 1a, the C-8 of the corrected structure 1 (cannabisin F) also contains an oxygen-bearing substituent, which is completely valid for the evidence provided by Gao et al. to support the structure of 1a. So it still remains unclear to distinguish between 1a and 1 with currently published experimental data. To further confirm the structure of compound 1 experimentally, we retested its NMR spectrum in methanol-d ₄. In its HMBC spectrum (Figure S5), although we found weak ⁴ J _H–C correlations of H-7′ (δ_H 7.24)/C-4 (δ_C 147.5) and/or C-3′ (δ_C 148.9), as well as H-5 (δ_H 6.72)/C-8′ (δ_C 141.5) and/or C-7 (δ_C 141.1), it is difficult to determine whether the structure is 1a or 1 through the HMBC spectrum due to their close ¹³C chemical shifts. However, in the NOESY spectrum (Figures S7 and S8), clear NOESY correlations were observed between H-5 (δ_H 6.72) and H-2′ (δ_H 7.28), as well as H-6′ (δ_H 7.04) and H-7′ (δ_H 7.24), which are consistent with the corrected structure 1 (Figure ). Quantum calculations of ¹³C NMR chemical shifts have recently been widely used in structure elucidation and confirmation. Therefore, we performed calculations of ¹³C NMR chemical shifts on 1a and 1, respectively, and the results showed that 1 is a possible structure, based on the comparison of their correlation coefficients (R ²) between the calculated and experimental data from linear regression analysis, since the R ² of 1 (0.9959) is greater than the R ² of 1a (0.9852) (Figure ). In addition, compound 1 was subjected to NMR testing using acetone-d ₆ as the solvent, and its ¹H and ¹³C NMR data were consistent with the reported compound cannabisin F (Table ). From these results, the structure of lyciumamide A (1a) was revised to cannabisin F (1).

2.

Key NOESY correlations are consistent with the structure of 1 but not with that of 1a.

8.

¹³C NMR calculations of compounds 1a and 1, 2a–2c, and 3a–3c.

1. ¹H and ¹³C NMR Data of Compound 1 and Cannabisin F (δ in ppm, J in Hz).

	1		1		Cannabisin F
No.	δH	δC	δH	δC	δH	δC
1		131.9		131.6		131.7
2	7.22, s	112.5	7.34, brs	113.5	7.35, d (2.0)	113.6
3		150.4		150.1		150.2
4		147.5		148.8		148.8
5	6.72, d (8.3)	116.3	6.78, d (8.3)	115.9	6.78, d (8.4)	115.9
6	7.01, d (8.3)	122.2	7.13, dd (8.3, 1.5)	125.6	7.14, dd (8.4, 2.0)	125.5
7	7.45, d (15.7)	141.1	7.46, d (15.7)	139.7	7.46, d (15.7)	139.6
8	6.49, d (15.7)	121.0	6.58, d (15.7)	121.8	6.57, d (15.7)	122.0
9		168.7		163.4		163.3
11	3.47, overlap	42.6	3.47, m	41.9	3.49, dd (13.0, 6.0)	42.0
12	2.75, t (7.0)	35.7	2.75, t (7.0)	35.7	2.75, t (7.1)	35.7
13		131.2		131.0		131.1
14, 18	7.05, d (8.2)	130.7	7.05, d (8.3)	130.4	7.06, d (8.6)	130.5
15, 17	6.72, overlap	116.3	6.75, d (8.3)	116.0	6.75, d (8.6)	116.1
16		156.9		156.6		156.7
3-OCH3	3.90, s	56.4	3.68, s	55.9	3.69, s	56.0
1′		130.8		125.4		125.5
2′	7.28, s	113.7	7.25, d (1.5)	112.3	7.27, d (2.0)	112.3
3′		148.9		148.3		148.3
4′		149.5		147.1		147.1
5′	6.72, overlap	116.3	6.75, d (8.3)	115.1	6.75, d (8.4)	115.1
6′	7.04, overlap	126.3	7.02, dd (8.3, 1.5)	121.5	7.03, dd (8.4, 2.0)	121.5
7′	7.24, s	125.3	7.28, s	123.8	7.27, s	123.7
8′		141.5		142.2		142.3
9′		165.5		166.3		166.1
11′	3.46, overlap	42.2	3.49, m	41.8	3.46, dd (13.0, 6.0)	41.8
12′	2.64, t-like (7.0)	35.5	2.65, t (7.0)	35.5	2.65, t (7.1)	35.5
13′		130.7		130.7		130.8
14′, 18′	6.83, d (8.3)	130.7	6.90, d (8.3)	130.5	6.90, d (8.6)	130.5
15′, 17′	6.59, d (8.3)	116.3	6.69, d (8.3)	116.0	6.68, d (8.6)	116.1
16′		156.8		156.6		156.7
3′-OCH3	3.65, s	56.0	3.92, s	56.2	3.94, s	56.2

^{a 1}H NMR (500 MHz) and ¹³C NMR (125 MHz) measured in CD₃OD.

^{b 1}H NMR (500 MHz) and ¹³C NMR (125 MHz) measured in acetone-d ₆.

^{c 1}H NMR (500 MHz) and ¹³C NMR (125 MHz) measured in acetone-d ₆, data from literature.

Lyciumamide B (2a) is also a dimer composed of FerTrm groups. The structures of 2a and 1a reported by Gao et al. are quite similar, both consisting of two FerTrm groups connected by a 16-O-16′ ether bond, which was deduced only based on EIMS fragment analysis. Unlike 1a, the C-7′ of 2a was considered to be substituted by a hydroxy group, and the two FerTrm groups formed a complex cyclic structure through a C–C bond between C-5 and C-8′. The structure of 2a was corrected by Zadelhoff et al. as the known compound grossamide, a dimer of two FerTrm units forming a benzofuran-type lignanamide, which has significant differences from 2a proposed by Gao et al. However, the relative configurations at C-7′ and C-8′ were not determined by Zadelhoff et al. The inference of the C-5-C-8′ connection in 2a by Gao et al. was based on the HMBC correlations of H-8′/C-1′/C-4/C-6 and H-7′/C-2′/C-6′/C-9′/C-5. However, the above HMBC correlations were also applicable to the furan ring of 2. Futhermore, the key HMBC correlation H-7′/C-4 used by Zadelhoff et al., to correct the structure, was very weak and also was not detected in our HMBC spectrum (Figure S15). Through the NOESY spectrum, it is still impossible to distinguish between structures 2a and 2. We reexamined the NMR data and found that the two methylene protons at C-11 (or C-11′ or C-12 or C-12′) shared the same chemical shift (δ_H 3.47 for H₂-11 or H₂-11′, δ_H 2.76 for H₂-12 or H₂-12′), and H/C-14 and H/C-18 also shared the same chemical shift (same to H/C-14′ and H/C-18′, H/C-15 and H/C-17, H/C-15′ and H/C-17′), which is much more possible for 2, since it has no closed ring formation at 16-O-16′ and allows free rotation of the side chain. Calculations of ¹³C NMR chemical shifts were performed on compounds 2a and 2b (grossamide), respectively, and the results showed that the correlation coefficient (R ² = 0.9831) of 2a is smaller than that (R ² = 0.9960) of 2b, suggesting that the original structure (2a) of lyciumamide B is incorrect and that 2 is a possible structure (Figure ). However, Zadelhoff et al. did not provide the relative configurations for 2, and there are two compounds reported in the literatures with the same planar structure as grossamide, one is grossamide (2b), whose H-7′ and H-8′ relative configurations are trans (7′R*,8′R*), the other one is tribulusamide A (2c), whose H-7′ and H-8′ relative configurations are cis (7′R*,8’S*) (Figure ). To further determine the relative configurations of compound 2, we analyzed the differences in NOE effect between the two compounds using NOESY spectrum. The NOESY correlation of H-7′ (δ_H 5.89)/H-6′ (δ_H 6.76) is stronger than that of H-7′ (δ_H 5.89)/H-8′ (δ_H 4.14) (Figure S18), and the NOESY correlations H-8′/H-7′, H-8′/H-2′, and H-8′/H-6′ are similar in intensity (Figure S18). This is consistent with the 3D structure of 2b (trans-H-7′/H-8′), since the distance of H-7′ to H-8′ is farther than that of H-7′ to H-6′, and the distances of H-8′ to H-7′ and H-8′ to H-2′ (or H-8′ to H-6′) are similarIn contrast, in the 3D structure of 2c (cis-H-7′/H-8′), the intensity of NOESY correlations between H-7′/H-8′ and H-7′/H-6′ is proposed to be similar, and the NOESY correlation of H-8′/H-7′ is proposed to be stronger than that of H-8′)­H-2′ (or H-8′)­H-6′) (Figure ). Therefore, the relative configurations at C-7′ and C-8′ were suggested to be 7′R*8′R* (trans-H-7′/H-8′ of 2b). To further confirm above conclusion, we calculated the ¹³C NMR chemical shifts for the cis/trans isomers of compound 2 (2b and 2c). The results showed that the correlation coefficient R ² for the cis isomer was 0.9969 and for the trans isomer was 0.9960 (Figure ). Since the correlation coefficient R ² values of 2b and 2c are too close, the parameter of DP4+ probability analysis was then performed for the combination of ¹H and ¹³C NMR data. As a result, the DP4+ probability of 2b was 100%, indicating that the trans-H-7′/H-8′ was the most probable relative configuration, and compound 2 was thus confirmed as grossamide (2b) (Figure S39). Compound 2b showed a small specific optical rotation value (−1.7, c 0.05, MeOH) and a weak Cotton effect, suggesting the presence of a racemic mixture. The isolation of individual enantiomers, (+)-2b and (−)-2b, were subsequently isolated by chiral HPLC (Figure S32). By calculating the ECD spectra of trans (7′R,8′R and 7′S,8′S) and cis (7′R,8′S and 7′S,8′R) isomers of compound 2 and comparing them with the experimental ECD spectra, it can be seen that the calculated ECD curves of the trans isomer (2b) are consistent with the experimental ECD curves, while the trend of the calculated ECD curves of the cis isomer (2c) is opposite around 250–280 nm to the experimental ECD curves, indicating that the structure of compound 2 is 2b with the 7′R*8′R* configuration (Figure ). The absolute configurations of (+)-2b and (−)-2b were assigned as 7′R,8′R and 7′S,8′S, respectively, based on comparison of their experimental and calculated ECD spectra (Figure ). Moreover, by comparing the NMR spectra of compound 2 with those of grossamide and tribulusamide A reported in the literatures, it can be observed that, under the same solvent conditions, the NMR data of compound 2 are more consistent with those of grossamide (Table ).

3.

Possible diastereoisomers of compounds 2 (2b and 2c) and 3 (3b and 3c).

4.

Strong H-7′/H-6′ and weak H-7′/H-8′ NOESY correlations that are consistent with the structure of 2b but less possible with that of 2c.

5.

Experimental ECD spectra of (+)/(−)-2 and calculated ECD spectra of (+)/(−)-2b (left)/or (+)/(−)-2c (right) in MeOH.

2. ¹H and ¹³C NMR Data of Compound 2b, Grossamide, and Tribulusamide A (δ in ppm, J in Hz).

	2b		2b		Grossamide		Tribulusamide A
No.	δH	δC	δH	δC	δH	δC	δH	δC
1		130.5		129.5		129.5		130.4
2	7.12, s	113.3	7.10, s	111.1	7.09, s	111.1	6.69, s	113.1
3		146.1		145.6		145.5		146.0
4		151.3		150.6		150.5		151.2
5		129.4		129.3		129.3		129.4
6	6.74, overlap	118.1	6.42, s	119.0	6.42, s	119.0	6.71, s	118.1
7	7.43, d (15.7)	141.8	7.40, d (15.7)	141.3	7.41, d (15.7)	141.3	7.41, d (15.7)	141.8
8	6.38, d (15.7)	119.5	6.48, d (15.7)	119.2	6.48, d (15.7)	119.2	6.37, d (15.7)	120.0
9		169.0		167.4		167.4		169.0
11	3.47, overlap	42.2	3.54, m	42.1	3.54, m	42.1	3.86, s	42.6
12	2.76, overlap	35.3	ND	35.5	2.79, m	35.4	3.33, m	35.8
13		131.1		130.4		130.4		131.3
14, 18	7.03, d (8.4)	130.9	7.09, d (8.3)	130.6	7.09, d (8.3)	130.6	7.09, d (8.5)	130.9
15, 17	6.72, overlap	116.5	6.77, d (8.3)	116.1	6.77, d (8.3)	116.0	6.73–7.05	116.5
16		156.9		156.7		156.7		156.9
3-OCH 3	3.89, s	56.9	3.85, s	56.3	3.85, s	56.3	3.81, s	56.8
1′		132.6		132.6		132.6		132.6
2′	6.91, d (1.4)	110.6	6.99, s	110.6	6.99, s	110.6	6.73–7.05	110.5
3′		149.3		148.5		148.5		148.1
4′		148.2		147.7		147.7		149.3
5′	6.80, d (8.1)	116.4	6.82, overlap	115.8	6.82, s	115.8	6.73–7.05	116.4
6′	6.76, overlap	120.0	6.82, overlap	119.9	6.82, s	119.9		119.4
7′	5.89, d (8.3)	90.0	6.03, d (8.5)	88.9	6.03, d (8.5)	88.8	5.87, d (8.3)	89.9
8′	4.14, d (8.3)	58.8	4.17, d (8.5)	57.7	4.18, d (8.5)	57.6	4.13, d (8.3)	58.7
9′		172.9		170.2		170.1		172.9
11′	3.47, overlap	42.6	3.72, m	41.4	3.72, m	41.4	3.82, m	42.2
			3.39, m		3.39, m
12′	2.76, overlap	35.8	ND	34.6	2.87, m	34.5	2.74, m	35.3
					2.74, m
13′		130.8		130.9		130.4		131.1
14′, 18′	7.06, d (8.4)	130.7	7.09, d (8.4)	130.5	7.09, d (8.4)	130.5	6.73–7.05	130.8
15′, 17′	6.72, overlap	116.3	6.81, overlap	116.9	6.81, d (8.4)	116.9		116.3
16′		156.9		156.7		156.7		156.9
3′-OCH 3	3.82, s	56.5	3.82, s	56.3	3.82, s	56.3	3.47, s	56.4

^{a 1}H NMR (500 MHz) and ¹³C NMR (125 MHz) measured in CD₃OD.

^{b 1}H NMR (400 MHz) and ¹³C NMR (100 MHz) measured in acetone-d ₆, data from literature.

^{c 1}H NMR (600 MHz) and ¹³C NMR (150 MHz) measured in acetone-d ₆.

^{d 1}H NMR (300 MHz) and ¹³C NMR (75 MHz) measured in CD₃OD, data from literature.

^eThere is no data in the literature.

^fND: not detected.

Lyciumamide C (3a) was originally identified as a FerTrm derivative fused with an oxetane ring (3a, lyciumamide C) by Gao et al., while its revised structure (3) by Zadelhoff et al. was proposed to be a known compound, grossamide K¹², also a benzofuran-type lignanamide that lacks a set of tyramine groups compared to those of 2. The presence of an unusual oxetane group, which was connected to phenyl groups at C-5 and C-1′ in 3a, was proposed by Gao et al. based on the ¹H–¹H COSY correlations of H-7′/H-8′/H-9′ and the HMBC correlations of H-9′/C-7′/C-5, H-8′/C-1′/C-4/C-6, and H-7′/C-2′/C-6′/C-5, along with the upfield chemical shifts of H-7′, C-7′, H-9′, and C-9′. However, the 1D and 2D NMR data described by Gao et al. were valid for both 3a and 3. To provide more evidence, we reanalyzed the HMBC spectrum of compound 3 and found a strong HMBC correlation at δ_H 5.55/δ_C 151.3, which is consistent with the three-bond correlation of H-7′/C-4 in 3, while the original report showed a four-bond correlation in structure 3a (Figure S26). The relative configurations at C-7′ and C-8′ of 3 were determined as trans based on the NOESY correlations. The stronger NOESY correlation H-7′ (δ_H 5.55)/H-9’ (δ_H 3.83) and the weaker NOESY correlation H-7′ (δ_H 5.55)/H-8′ (δ_H 3.52) were consistent with the structure of 3b but less consistent with that of 3c (Figures and S28). To further prove the above conclusion, we performed calculations of ¹³C NMR chemical shifts on the original reported structure 3a, the trans isomer 3b (7′R*8’S*-3), and the cis isomer 3c (7′R*8’R*-3), with correlation coefficients of R ² = 0.9894 (3a), 0.9962 (3b), and 0.9958 (3c), respectively, and further DP4+ analysis showed a 100% probability for 3b, indicating that 3b (7′R*8’S*-3) is the correct structure (Figure ). Compound 3b showed a small specific optical rotation value (−0.9, c 0.21, MeOH), suggesting the presence of a racemic mixture, which was separated using a Chiralpak IC chiral column to obtain (+)-3b and (−)-3b, respectively (Figure S33). The absolute configurations of (+)-3b and (−)-3b were defined as 7′R,8′S and 7′S,8′R, respectively, by comparing their experimental and calculated ECD spectra (Figure ). In conclusion, compound 3 was confirmed as the known compound grossamide K based on the key HMBC correlation and NMR calculations, and the relative and absolute configurations were also assigned (Table ).

6.

Stronger H-7′/H-9′ and weaker H-7′/H-8′ NOESY correlations that are consistent with the structure of 3b but less possible with that of 3c.

7.

Calculated and experimental ECD spectra of (+)-3b and (−)-3b in MeOH.

3. ¹H and ¹³C NMR Data of Compound 3b and Grossamide K (δ in ppm, J in Hz).

	3b		3b		Grossamide K
No.	δH	δC	δH	δC	δH	δC
1		130.3		129.9		129.7
2	7.08, s	113.4	7.08, s	113.0	7.07, s	112.8
3		145.8		145.3		145.2
4		151.3		150.7		150.7
5		130.9		131.2		130.8
6	7.13, s	118.5	7.14, s	117.7	7.14, s	117.8
7	7.47, d (15.7)	142.0	7.47, d (15.6)	140.5	7.49, d (15.6)	140.7
8	6.43, d (15.7)	119.2	6.51, d (15.6)	120.1	6.54, d (15.6)	119.8
9		169.1		166.3		166.7
11	3.46, t (7.2)	42.5	3.49, m	41.8	3.56–3.50, m	42.0
12	2.75, t (7.2)	35.8	2.75, t (7.2)	35.7	2.75, t (7.0)	35.6
13		131.3		130.9		130.9
14, 18	7.05, d (8.0)	130.7	7.06, d (8.4)	130.5	7.06, d (8.7)	130.5
15, 17	6.71, d (8.0)	116.3	6.76, d (8.4)	116.0	6.76, d (8.7)	116.0
16		156.9		156.6		156.7
3-OCH3	3.88, s	56.8	3.87, s	56.4	3.85, s	56.3
1′		134.2		134.0		133.8
2′	6.94, s	110.6	7.04, d (1.9)	110.5	7.00, d (2.0)	110.5
3′		149.2		148.4		148.4
4′		147.7		147.4		147.4
5′	6.77, d (8.1)	116.2	6.82, d (8.0)	115.7	6.82, d (8.2)	115.7
6′	6.83, d (8.1)	119.8	6.88, dd (8.0, 1.9)	119.7	6.88, dd (8.2, 2.0)	119.6
7′	5.55, d (6.4)	89.7	5.59, d (6.3)	88.9	5.60, d (6.4)	88.9
8′	3.52, dt (6.4)	54.8	3.57, dt (6.3)	54.4	3.57, q (6.4)	54.3
9′	3.83, overlap	64.7	3.79–3.91, overlap	64.3	3.79–3.91, m	64.3
3′-OCH3	3.81, s	56.4	3.82, s	56.3	3.80, s	56.2

^{a 1}H NMR (500 MHz) and ¹³C NMR (125 MHz) measured in CD₃OD.

^{b 1}H NMR (500 MHz) and ¹³C NMR (125 MHz) measured in acetone-d ₆.

^{c 1}H NMR (300 MHz) and ¹³C NMR (75 MHz) measured in acetone-d ₆, data from literature.

3.2. Evaluation of Biological Activities of Compound 3b

The PPAR-γ agonistic activities of compounds 1 and 2 have been previously reported by our group Figure . Compound 3b and its enantiomers (+)-3b and (−)-3b were also assessed for their PPAR-γ agonistic activity, and the racemic mixture (3b) showed significant agonistic activity for PPAR-γ, while its enantiomers (+)-3b and (−)-3b showed weaker activities (Figure A). Therefore, a dose-dependent test for PPAR-γ agonistic activity was performed on the racemic mixture 3b, and the EC₅₀ value was measured to be 11.97 ± 0.75 μM [positive control rosiglitazone (RSG), 1.68 ± 0.02 μM] (Figure B).

9.

Compound 3b promotes the differentiation of 3T3-L1 cells. (A, B) The effects of compound 3b and its enantiomers (+)-3b and (−)-3b on the luciferase activity of PPAR-γ (n = 3). (C) Representative images of 3T3-L1 cells stained with Oil Red O on day 6 of differentiation. Ins + Dex, 10 μg/mL insulin + 1 μM dexamethasone. (D) The quantification of Oil Red O in 3T3-L1 cells (n = 3). (E, F) The effect of compound 3b on the expression of Glut-4 and Scd-1 mRNA during different days of 3T3-L1 differentiation (n = 3). Data are expressed as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001 vs DMSO. Ins: insulin; Dex: dexamethasone; RSG: rosiglitazone.

3T3-L1 mouse embryonic fibroblasts are widely used adipocyte models, and they differentiate into adipocytes after induction with insulin, dexamethasone, etc. , PPAR-γ agonists exert hypoglycemic effects by enhancing insulin sensitivity, promoting glucose uptake, and adipocyte differentiation. , We compared the effects of compound 3b and PPAR-γ agonist RSG on anterior 3T3-L1 differentiation into adipocytes. Oil Red O specifically recognizes lipid droplets and stains them red. As shown in Figure C,D, 3T3-L1 cells cultured in medium 1 partially differentiated into adipocytes containing small lipid droplets, while compound 3b promoted the differentiation of 3T3-L1 adipocytes in a dose-dependent manner and was more effective than RSG at concentrations of 5 and 10 μM (p < 0.001). The mRNA expression of PPAR-γ target genes involved in adipocyte predifferentiation, such as glucose transporter 4 (Glut-4) and stearoyl-CoA desaturase 1 (Scd-1), showed a time-dependent increase during the differentiation process (Figure E and F). In conclusion, the above results indicated that compound 3b is a natural PPAR-γ agonist and has the potential in the treatment of type 2 diabetes.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.5c07441.

• 1D and 2D NMR spectra, HRESIMS spectra of compounds 1–3; ECD spectra and Chiral HPLC separation of compounds 2b and 3b; ECD calculations of compounds (+)-2b, (−)-2b, (+)-3b, and (−)-3b; NMR calculations of compounds 1a and 1, 2a–2c, and 3a–3c (PDF)

The authors declare no competing financial interest.