A pair of new enantiomeric hybrid phthalide–adenines with a rare 5-oxa-1-azaspiro[3,4]octane moiety and two pairs of new enantiomeric hybrid paraethyl phenol–adenines from Ligusticum chuanxiong

2.1 General experimental procedures

Optical rotations and IR data were measured using an Anton Paar MCP 200 automatic polarimeter (Anton Paar GmbH, Graz, Austria) and an Agilent Cary 600 FT-IR microscope (Agilent Technologies Inc., CA, USA). ECD spectra were recorded using an Applied Photophysics Chirascan and Chirascan-plus circular dichroism spectrometer (Applied Photophysics ltd., Leatherhead, England). NMR spectra were obtained by a Bruker-AVIIIHD-600 spectrometer or a Bruker Avance NEO 600 spectrometer (Bruker Corporation, Billerica, MA, USA) with solvent peaks as reference standards. X-ray structure analyses were performed on a Bruker D8 QUEST diffractometer (Bruker Corporation, Billerica, MA, USA). HRESIMS measurements were carried out using a Waters Synapt G2 HDMS instrument (Waters Corporation, Milford, MA, USA) or a Bruker tims-TOF-MS instrument (Bruker Corporation, Billerica, MA, USA). TLC was performed using glass plates precoated with silica gel GF₂₅₄ (Qingdao Marine Chemical Inc., Qingdao, China). Column chromatography separation was performed using D-101-type macroporous adsorbent resin (Chengdu Kelong Chemical Reagent Factory, Chengdu, China), 732-type strong acid cation exchange resin (Chengdu Chron Chemicals Co., ltd., Chengdu, China), silica gel (200–300 mesh, Yantai Institute of Chemical Technology, Yantai, China), and Sephadex LH-20 (40–70 μm, Amersham Pharmacia Biotech AB, Uppsala, Sweden). Meanwhile, HPLC separations were carried out using an Alltech 426 instrument. Reversed-phase semipreparative HPLC analyses were conducted on a Zorbax SB-C₁₈ (250 × 9.4 mm², 5 μm) column, while normal-phase and reversed-phase enantioseparations were conducted on Daicel Chiralpak AD-H (250 × 4.6 mm², 5 μm) and Daicel Chiralpak IG (250 × 4.6 mm², 5 μm) columns, respectively. The uterine smooth muscle relaxant activity and vasorelaxant activity of the isolated compounds were assessed using a PL3508B6/C-V Panlab 8 Chamber Organ Bath System (AD Instruments, Australia). RAW264.7 cells were obtained from Obio Technology Corp (Changsha, China), while the lipopolysaccharide was obtained from Sigma-Aldrich (St. Louis, MO, United States). Curcumin was purchased from Sichuan Weikeqi Biological Technology Co., ltd. (Sichuan, China). Mouse IL-6 Elisa and Mouse TNF-α Elisa kits were purchased from Elabscience Biotechnology Co., ltd. (Wuhan, Hubei, China). Total RNA Isolation Kit, 2 × RT OR-Easy^TM Mix, and 2 × Real Time PCR Easy^TM MIX-SYBR Green I were obtained from Chengdu Foregene Biological Technology Co., ltd. (Chengdu, China).

2.2 Plant material

Ligusticum chuanxiong Hort. was obtained from NeoGreen Pharmaceutical Technology Development Co., ltd. (Pengzhou, China) in October 2016. The sample was identified by Dr. Jihai Gao and deposited in the School of Pharmacy at Chengdu University of Traditional Chinese Medicine (specimen: LC-20161030).

2.3 Extraction and isolation

The L. chuanxiong rhizome (100 kg) was decocted with H₂O three times (2.5 h each time). The residue (28.8 kg) obtained by evaporating the water extract was suspended in warm water and partitioned with n-BuOH. The n-BuOH portion (9.3 kg) was separated by column chromatography on D101 macroporous adsorption resin, and six fractions were eluted with H₂O, 10%, 30%, 50%, 70%, and 95% ethanol. The 30% EtOH fraction (701.0 g) was further separated by elution on strong acid cation exchange resin with H₂O, 75% EtOH, and 75% EtOH containing 2 M ammonia. Of the three major fractions obtained, the one eluted with 75% EtOH containing 2 M ammonia water (120.0 g) was further separated into nine fractions (F₁–F₉) by silica gel column chromatography using gradient elution with CH₂Cl₂/MeOH (100:0–0:100). Separation of the F₆ fraction (3.1 g) by RP-MPLC (gradient elution with 10:90–100:0 MeOH/H₂O) afforded nine subfractions (F_6-1–F_6-9).

The F_6-3 subfraction was further separated on a Sephadex LH-20 column (85% MeOH in H₂O), and seven subfractions F_6-3-1–F_6-3-7 were obtained. F_6-3-3 was successively purified by Sephadex LH-20 column chromatography (CH₂Cl₂/MeOH, 1:1), preparative TLC (CH₂Cl₂/MeOH, 9:1), and semipreparative HPLC (25% MeOH in H₂O) to yield compound 1 (1.5 mg). In addition, compounds 2 (6.3 mg) and 3 (3.2 mg) were obtained by purifying F_6-3-4 on a silica gel column (CH₂Cl₂/MeOH, 80:1–1:1), followed by semipreparative HPLC (40% MeOH in H₂O).

Since compounds 1–3 show no cotton effects in their ECD spectra, and since their specific optical rotations are close to zero, these compounds were subjected to HPLC enantioseparation on a Daicel Chiralpak IG column or a Daicel Chiralpak AD-H column. Accordingly, three pairs of enantiomers, (+)-1 (0.59 mg)/(−)-1 (0.61 mg), (+)-2 (2.60 mg)/(−)-2 (2.50 mg), and (+)-3 (1.54 mg)/(−)-3 (1.60 mg), were obtained (Table 1).

2.4 Physicochemical properties and spectroscopic data of compounds 1–3

2.5 X-ray crystallography data of compound 2

Compound 2 was dissolved with dichloromethane/methanol (4:6, V/V) and then slowly evaporated to a single crystal at 4℃. Crystal data for 2: C₁₄H₁₅N₅O₂·H₂O, M = 303.32, triclinic, a = 7.8867(2) Å, b = 8.9792(2) Å, c = 11.4775(3) Å, α = 81.829(10)°, β = 73.211(10)°, γ = 68.117(10)°, V = 721.54(3) ų, space group P-1, T = 293(2) K, Z = 2, μ(Cu Kα) = 0.845 mm⁻¹, 22,147 reflections measured, 2652 independent reflections (R_int = 0.0765), average redundancy 8.351, completeness = 99.5 %. Final R indices (I > 2σ(I)): R₁ = 0.0428, wR₂ = 0.1140. Final R indices (all data): R₁ = 0.0637, wR₂ = 0.1256. The goodness of fit on F² was 1.073. CCDC number: 2157588. The results of X-ray diffraction analysis showed that 2 is an enantiomeric racemate.

2.6 Bioactivity assays

3.1 Structure elucidation

Compound 1 was isolated as a white amorphous powder. Its HR-ESI-MS presents a positive quasi-molecular ion at m/z 344.1126 (calcd. for C₁₇H₁₅N₅O₂Na, 344.1123), which suggests that the molecular formula of this compound is C₁₇H₁₅N₅O₂ possessing 13 degrees of unsaturation. The ¹H NMR and ¹H–¹H COSY data (Table 2 and Fig. 2) suggest that compound 1 possesses an adenine moiety [δ_H 8.40 (1H, s, H-2) and 8.60 (1H, s, H-8)] (Li et al., 2009), an ortho-substituted benzene ring [δ_H 7.94 (1H, d, J = 7.2 Hz, H-7′), 7.90 (1H, t, J = 7.2 Hz, H-5′), 7.85 (1H, d, J = 7.2 Hz, H-4′), and 7.75 (1H, t, J = 7.2 Hz, H-6′)], and a substituted n-butyl unit [δ_H 3.20 (1H, dd, J = 15.6, 11.4 Hz, H-8′a), 2.57 (1H, dd, J = 15.6, 1.8 Hz, H-8′b), 4.97 (1H, m, H-9′), 2.37 (1H, m, H-10′a), 2.22 (1H, m, H-10′b), and 0.97 (3H, t, J = 7.2 Hz, H₃-11′)]. The ¹³C NMR (Table 2) and DEPT spectra show 17 carbon signals corresponding to the above units, with seven quaternary carbons, three of which (δ_C 160.7, 151.8, 113.9) are attributed to the adenine unit. The remaining four quaternary carbons include a carbonyl carbon (δ_C 169.9), two aromatic carbons (δ_C 151.8 and 126.4), and a carbon (δ_C 95.3) bonded to two hetero atoms. The resonance signals detailed above reveal that compound 1 is an unusual hybrid phthalide–adenine. The HMBC correlations (Fig. 2) of H-2 with C-4 and C-6; and H-8 with C-4 and C-5 confirm the presence of an adenine unit. Meanwhile, the n-butylphthalide fragment is evidenced by the HMBC correlations of H-4′ with C-3′, C-6′, and C-7a′; H-5′ with C-3a′ and C-7′; H-6′ with C-4′ and C-7a′; H-7′ with C-1′, C-3a′, and C-5′; and H-8′ with C-3′, C-3a′, C-9′, and C-10′, as well as by the ¹H–¹H COSY signals of H-4′/H-5′/H-6′/H-7′ and H-8′/H-9′/H-10′/H-11′. The adenine and butylphthalide units of 1 account for 12 degrees of unsaturation. The 13th degree of unsaturation suggests a ring linkage between the two units. In addition, compared to the common n-butylphthalide compound, C-3′ in compound 1 is changed to a quaternary carbon (δ_C 95.3) bonded to two hetero atoms, while C-9′ in compound 1 is changed to a CH group (δ_H 4.97, δ_C 56.5) linked to a N atom. Overall, the spectroscopic data indicate that a C-3′–N-10–C-9′ linkage between the adenine and butylphthalide units in 1 forms a four-membered aza-heterocycle, resulting in the formation of a rare 5-oxa-1-azaspiro[3,4]octane moiety. Based on these results, the planar structure of compound 1 is established.

Close

Fig. 2

Key ¹H–¹H COSY and HMBC correlations of compounds 1–3.

Export to PPT

The relative configuration of compound 1 was elucidated based on the analysis of NOESY cross-peaks. The NOESY correlations of H-4′ (δ_H 7.85)/H-8′a (δ_H 3.20) and H-8′a (δ_H 3.20)/H-10′a (δ_H 2.37) demonstrate that H-4′, H-8′a, and H-10′ have the same orientation (Fig. 3). Interestingly, no cotton effect is observed in ECD spectrum of compound 1, which suggests that this compound may be a racemic mixture. Indeed, chiral separation of 1 on a Daicel Chiralpak IG column yields (+)-1 and (−)-1 (Table 1), and based on the comparison of experimental and calculated ECD spectra, the absolute configurations of these enantiomers are 3′R,9′S and 3′S,9′R, respectively (Fig. 4). Hence, (+)-1 and (−)-1 were established as (+)-(3′R,9′S)-liguadenine A and (−)-(3′S,9′R)-liguadenine A, respectively.

Close

Fig. 3

Key NOESY correlations of compound 1.

Export to PPT

Close

Fig. 4

The experimental and calculated ECD spectra of compounds 1–3.

Export to PPT

Based on HR-ESI-MS analysis (m/z 308.1127 [M + Na]⁺, calcd. for C₁₄H₁₅N₅O₂Na, 308.1123), the molecular formula of compound 2 is C₁₄H₁₅N₅O₂. ¹H and ¹³C NMR analyses reveal that 2 is an adenine hybrid that comprises a 4-ethyl-2-methoxyphenol moiety with C-7′ substituted [δ_H 7.03 (1H, s, H-2′), 6.88 (1H, d, J = 7.8 Hz, H-6′), 6.75 (1H, d, J = 7.8 Hz, H-5′), 5.43 (1H, m, H-7′), and 1.60 (3H, d, J = 7.2 Hz, H-8′); δ_C 149.0, 146.7, 136.8, 119.7, 116.2, 111.2, 51.0, and 23.1], as well as a methoxy group [δ_H 3.84 (3H, s, OMe-3′); δ_C 56.4]. This is confirmed by the HMBC correlations of H-2′ with C-1′, C-3′, C-4′, and C-7′; OMe-3′ with C-3′; H-6′ with C-1′, C-2′, C-4′, and C-7′; and H₃-8′ with C-1′ and C-7′, together with the ¹H–¹H COSY correlations of H-5′/H-6′ and H-7′/H-8′ (Fig. 2). Although no HMBC correlation of H-7′/C-6 is observed, the adenine and 4-ethyl-2-methoxyphenol substructures must be connected by a C-6–N-10–C-7′ linkage in order to match the molecular formula and chemical shifts of C-6 and C-7′. This linkage is verified by X-single crystal diffraction analysis (Fig. 5). Considering that the specific optical rotation of 2 is zero, this compound was subjected to enantioseparation on a Daicel Chiralpak AD-H column (Table 1). As described for compound 1, ECD calculations were used to determine the absolute configuration of 2, and as shown in Fig. 4, the experimental ECD spectra of (+)-2 and (−)-2 are consistent with the calculated ECD spectra of (7′R)-2 and (7′S)-2, respectively. Therefore, compounds (+)-2 and (−)-2 are labelled (+)-(7′R)-liguadenine B and (−)-(7′S)-liguadenine B.

Close

Fig. 5

X-ray crystallographic structure of compound 2.

Export to PPT

The spectroscopic data of compound 3 are similar to those of 2. Indeed, HR-ESI-MS analysis reveals that the two compounds possess the same molecular formula, which indicates that they are isomers. The ¹H and ¹³C NMR data of 3 (Table 2) suggest that this compound is also a hybrid, and that it is composed of adenine and 4-ethyl-2-methoxyphenol moieties, like compound 2. However, the C-4, C-5, C-6, C-1′, C-7′, and C-8′ resonances in the ¹³C NMR spectrum of 3 are shifted by Δδ_C −0.9, +1.4, +2.6, −3.7, +4.4, and −1.9 ppm, respectively, compared to the same signals in the spectrum of 2. Therefore, the difference between compounds 3 and 2 lies in the linkage between the two moieties. Based on the HMBC correlations of H-7′ with C-4, C-8, C-1′, C-2′, C-6′, and C-8′, the adenine and 4-ethyl-2-methoxyphenol substructures in 3 are connected by a N-9–C-7′ linkage. The specific optical rotation and ECD spectrum of 3 suggest that it is a racemic mixture, and the absolute configurations of the separated (+)-3 and (−)-3 enantiomers are 7′R and 7′S, respectively, as determined by ECD calculations (Fig. 4). Therefore, compounds (+)-3 and (−)-3 are established as (+)-(7′R)-liguadenine C and (−)-(7′S)-liguadenine C, respectively.

3.2 Effects of compounds 1–3 on TNF-α and IL-6 levels in RAW 264.7 cells

First, all compounds showed no cytotoxic activity against RAW264.7 cells in the range of experimental concentrations. Then, the anti-inflammatory effects of the three pairs of enantiomeric alkaloids isolated from L. chuanxiong were assayed using LPS-stimulated RAW 264.7 macrophages, and the concentrations of inflammatory factors (TNF-α and IL-6) in the supernatant were detected by ELISA. Except for (+)-2, all compounds exhibit significant inhibitory activity against LPS-induced TNF-α production, with (+)-3 being the most active (Fig. 6). Compounds (+)-1/(−)-1 and (+)-3 also show significant inhibitory effects against the production of IL-6. Interestingly, the anti-inflammatory activity of (+)-3 is better than that of its enantiomer. In addition, a comparison of the anti-inflammatory effects of (±)-2 and (±)-3 indicates that the linkage between the adenine and paraethyl phenol moieties in these compounds plays an important role in anti-inflammatory activity (N-9 linkage > N-10 linkage). Curcumin was used as a positive control and showed a significant inhibitory effect on TNF and IL-6 release (Fig. 7).

Close

Fig. 6

Effects of compounds 1–3 on the levels of TNF-α (A) and IL-6 (B) in LPS-stimulated RAW 264.7 cells. The data represent the mean ± SD of three experiments. ^##P < 0.01 vs the control group; *p < 0.05, **p < 0.01 vs the LPS group.

Export to PPT

Close

Fig. 7

Effects of curcumin on the levels of TNF-α (A) and IL-6 (B) in LPS-stimulated RAW 264.7 cells. The data represent the mean ± SD of three experiments. ^##P < 0.01 vs the control group; **p < 0.01 vs the LPS group.

Export to PPT

3.3 Effect of compound (+)-3 on the mRNA expression of LPS-induced inflammatory cytokines in RAW264.7 cells

The effect of (+)-3, the most active compound, on the mRNA expression of LPS-induced inflammatory cytokines in RAW264.7 cells was further explored. As shown in Fig. 8, LPS prominently induces the mRNA expressions of TNF-α and IL-6 compared to the normal control. However, treatment with 50 μM (+)-3 significantly downregulates the IL-6 and TNF-α mRNA expressions induced by LPS. These results suggest that (+)-3 suppresses inflammatory response by downregulating the mRNA expressions of TNF-α and IL-6.

Close

Fig. 8

Effects of compound (+)-3 on the mRNA expression of TNF-α (A) and IL-6 (B) in LPS-stimulated RAW264.7 cells. The data represent the mean ± SD of three experiments. ^#P < 0.05, ^##P < 0.01 vs the control group; *p < 0.05, vs the LPS group.

Export to PPT

3.4 Other activity assay

None of the tested compounds showed inhibitory effects against OT-induced contractions of rat uterine smooth muscle, KCl-induced contractions of rat aorta rings, and ADP-induced rabbit platelet aggregation.