Fitoterapia 103 (2015) 136–142
Contents lists available at ScienceDirect
Fitoterapia journal homepage: www.elsevier.com/locate/fitote
Lignans from the root of Paeonia lactiflora and their anti-β-amyloid aggregation activities Xiao Liu a, Ming-hua Yang a,⁎, Xiao-bing Wang a, Sai-sai Xie a, Zhong-rui Li a, Dong-hynu Kim b, Jun-seong Park b, Ling-yi Kong a,⁎⁎ a State Key Laboratory of Natural Medicines, Department of Natural Medicinal Chemistry, China Pharmaceutical University, 24 Tong Jia Xiang, Nanjing 210009, People's Republic of China b Amore-Pacific R&D Center, 314-1, Bora-dong, Giheung-gu, Yongin-si, Gyeonggi-Do, 446-729, South Korea
a r t i c l e
i n f o
Article history: Received 18 January 2015 Accepted in revised form 7 March 2015 Available online 27 March 2015 Keywords: Paeonia lactiflora Lignan Enantiomeric separation β-amyloid aggregation
a b s t r a c t Four new neolignans (1–4), together with two known lignans (5 and 6), were isolated from the root of Paeonia lactiflora. Compounds 1 and 2 were two racemates and were separated by chiral high performance liquid chromatography (HPLC) to give all of the four stereoisomeric forms sharing a common planar structure. Compounds 3 and 4 were two neolignan glycoside diastereomers but interestingly appeared to be enantiomers: they had the extremely similar 1H and 13C NMR spectra and had to be solved only by chiral HPLC. Their structures were determined by spectroscopic analysis, including 1D and 2D NMR, HRESIMS and electronic circular dichroism experiments. All compounds were evaluated for their inhibitory effects on β-amyloid aggregation, and the optical pure compound 2b was found to show the optimal Aβ1–42 aggregation inhibition potency (81.1% at 20 μM). In addition, despite large amount of chemical studies performed on genus Paeonia, the lignans were reported for the first time. © 2015 Published by Elsevier B.V.
1. Introduction Paeonia lactiflora (Ranunculaceae), widely distributed in China, Korea, and Russia, is a herb with ornate flowers blooming in spring. The root of P. lactiflora (Chi-Shao and Bai-Shao) is one of the most important crude drugs used as antioxidant [1,2], analgesic [3], and anti-inflammatory [4] agent in Asia. It is commonly used as a remedy for cardiovascular disorders, extravagated blood, and rheumatoid arthritis [5–8]. Modern pharmacological research demonstrates that P. lactiflora, with effects of neuroprotectivity and ameliorating cognitive impairment, is a potential therapeutic strategy for the treatment of Alzheimer′s disease (AD) [9–11]. Monoterpenoid glycosides, the main ingredient of the genus Paeonia, exhibited the above
⁎ Corresponding author. Tel.: +86 25 8618 5039. ⁎⁎ Corresponding author. Tel./fax: +86 25 8327 1405. E-mail addresses: [email protected] (M. Yang), [email protected] (L. Kong).
http://dx.doi.org/10.1016/j.fitote.2015.03.011 0367-326X/© 2015 Published by Elsevier B.V.
protective effect and ameliorated oxidative stress damage occurred in the pathological courses of AD [10,11]. In addition, triterpenoids, flavonoids, phenols, and tannins which were related to several different bioactivities have also been isolated from Paeonia [12]. In our recent search for bioactive constituents from P. lactiflora, lignans (Fig. 1) including two pairs of new enantiomeric 8-O-4′ neolignans (1a and 1b, 2a and 2b), two new benzofuran neoligans (3 and 4), and two known lignans (5 and 6) were obtained from this genus for the first time. Their structures were elucidated on the basis of spectroscopic methods including 1D and 2D NMR, and HRMS, and their stereochemistry was determined by acid hydrolysis and electronic circular dichroism (ECD) experiments. Although these lignans did not have obvious neuroprotective activities in hydrogen peroxide-induced PC12 cell damage bioassay, they had good inhibitory potency toward self-induced Aβ1–42 aggregation (56.2–81.1% at 20 μM). According to the “amyloid hypothesis” [13], the production and accumulation of oligomeric β-amyloid (Aβ) aggregates in the brain are a central
X. Liu et al. / Fitoterapia 103 (2015) 136–142
137
Fig. 1. Structures of compounds 1–6.
event in the pathogenesis of AD. Thus, the discovery of these lignans indicated a new potential approach of P. lactiflora to treat AD. Herein, we discuss the purification, structure determination, and Aβ aggregation inhibitory effects of these isolates.
A voucher specimen (No. 2013-PL) has been deposited at the Department of Natural Medicinal Chemistry, China Pharmaceutical University.
2. Experimental
2.3. Extraction and isolation
2.1. General
The air-dried root of P. lactiflora (10.0 kg) was powdered and extracted with 60% EtOH for three times affording dark residue (4.1 kg) after solvent evaporation under reduced pressure. The residue was subjected to D-101 macroporous resin column chromatography (CC) with a step gradient of EtOH– H2O solvent system (10:90, 30:70, 50:50, and 95:5, v/v) to yield four fractions (A–D). Fr. C (210 g) was then subjected to silica gel CC eluted by a step gradient of CH2Cl2–MeOH solvent system (50:1, 20:1, 10:1, 5:1, and 1:1, v/v) to obtain eight subfractions (C1–C8). Fr. C4 was further separated via silica gel CC, ODS CC, and MPLC successively to produce six subfractions (C4.2.1.1–C4.2.1.6). Fr. C4.2.1.2 was subjected to preparative HPLC with MeOH–H2O (40:60, v/v) to yield 5 (4.6 mg). Fr. C4.2.1.4 and Fr. C4.2.1.5 were applied to preparative HPLC with MeOH–H2O (45:55, v/v) to acquire 1 (8.2 mg) and 2 (3.9 mg), respectively. In addition, direct enantiomeric separation of compounds 1 and 2 was achieved by preparative chiral HPLC with 2-propanol–n-hexane (25:75, v/v), affording compounds 1a (3.0 mg) and 1b (3.9 mg), 2a (1.8 mg) and 2b (1.4 mg), respectively. Fr. C4.2.1.3 was chromatographed over Sephadex LH-20 with MeOH to produce five subfractions (C4.2.1.3.1– C4.2.1.3.5). Fr. C4.2.1.3.1 was then subjected to preparative HPLC with MeOH–H2O (45:55, v/v) to afford 6 (5.0 mg). Fr. C6 was chromatographed over MCI, ODS, silica gel and Sephadex LH-20 respectively to obtain Fr. C6.2.6.5.3 and Fr. C6.2.6.5.3 was further purified by preparative HPLC (MeOH–H2O, 40:60, v/v) to obtain the mixture (6.1 mg) of 3 and 4. Compounds 3 (3.2 mg) and 4 (2.4 mg) were separated using preparative chiral chromatography with 2-propanol–n-hexane (32:68, v/v).
Optical rotations were determined with a JASCO P-1020 polarimeter. ECD spectra were carried out on a JASCO-810 spectropolarimeter. UV spectra were obtained on a Shimadzu UV-2501 spectrophotometer. IR (KBr disks) spectra were measured by a Bruker Tensor 27 spectrometer. NMR spectra were recorded on a Bruker Avance III NMR (1H: 500 MHz, 13C: 125 MHz) instrument at 300 K, with TMS as internal standard. HRESIMS was carried out on an Agilent UPLC-Q-TOF (6520B) mass spectrometer. Fluorescence was measured on SpectraMax M5 (Molecular Devices, Sunnyvale, CA, USA) multi-mode plate reader. Chiral HPLC was done with a Daicel Chiralpak AD-H column (250 × 4.6 mm). Preparative HPLC was carried out using a SHMADZU LC-6A series instrument with a Shim-park RP-C18 column (200 × 20 mm) and a SHMADZU SPD-20A detector. Chiralpak AD-H preparative columns (250 × 10 mm) were purchased from Daicel Chemical Ltd. (Shanghai, China). Silica gel (Qingdao Haiyang Chemical Co., Ltd., Qingdao, China), RP-C18 (40–63 μm, Fuji, Japan), Sephadex LH-20 (Pharmacia, USA) and macroporous resin D-101 (poresize B 13–14 nm, 26– 60 mesh, Qingdao Haiyang Chemical Co. Ltd., China) were used for column chromatography. 2.2. Plant material The root of P. lactiflora was collected from Sinwol-ri, Yuchimyeon, Jangheung-gun, Jeollanam-do, Korea in March 2013 and authenticated by Professor Mian Zhang, Department of Natural Medicinal Chemistry, China Pharmaceutical University.
138
X. Liu et al. / Fitoterapia 103 (2015) 136–142
Compound 1: colorless gum; UV (MeOH) λmax (log ε) 208 (4.46), 280 (3.81) nm; IR (KBr) νmax 3456, 1634, 1513, 1384, 1274, 1033, 639 cm−1; 1H NMR and 13C NMR (CD3OD) data see Table 1; HRESIMS m/z 371.1464 [M + Na]+ (calcd for C19H24NaO6, 371.1465). 1a: [α]25 D +3.3 (c 0.12, MeOH); CD (2.0 × 10−4, MeOH) λmax (Δε) 239 (+2.25) nm. 1b: [α]25 D −4.7 (c 0.14, MeOH); CD (2.0 × 10−4, MeOH) λmax (Δε) 239 (−2.75) nm. Compound 2: pale yellow gum; UV (MeOH) λmax (log ε) 208 (4.46), 280 (3.81) nm; IR (KBr) νmax 3450, 1638, 1384, 1035, 669 cm−1; 1H NMR and 13C NMR (CD3OD) data see Table 1; HRESIMS m/z 366.1912 [M + NH4]+ (calcd for C19H28NO6, 366.1911). 2a: [α]25 D −21.7 (c 0.06, MeOH); CD (2.0 × 10−4, MeOH) λmax (Δε) 231 (−3.45) nm. 2b: [α]25 D +16.0 (c 0.04, MeOH); CD (1.5 × 10−4, MeOH) λmax (Δε) 232 (+2.83) nm. Compound 3: white, amorphous powder; [α]25 D −17.5 (c 0.10, MeOH); UV (MeOH) λmax (log ε) 208 (4.31), 280 (3.32) nm; CD (1.5 × 10−4, MeOH) Δε 219 nm 0.58, 238 nm −0.64; IR (KBr) νmax 3449, 1638, 1399, 1072, 669 cm−1; 1H NMR and 13 C NMR (CD3OD) data see Table 1; HRESIMS m/z 524.2485 [M + NH4]+ (calcd for C26H38NO10, 524.2490). Compound 4: white, amorphous powder; [α]25 D −11.5 (c 0.11, MeOH); UV (MeOH) λmax (log ε) 208 (4.31), 280 (3.32) nm; CD (1.5 × 10−4 MeOH) Δε 220 nm −2.19, Δε 245 nm 0.72; IR (KBr)
νmax 3447, 1638, 1398, 1074, 669 cm−1; 1H NMR and 13C NMR (CD3OD) data see Table 1; HRESIMS m/z 524.2491 [M + NH4]+ (calcd for C26H38NO10, 524.2490). 2.4. Acid hydrolysis and HPLC analysis of 3 and 4 The mixture of compounds 3 and 4 (1.0 mg) was dissolved in 2 N HCl (2.0 mL) at 100 °C for 4 h. After removal of the solvent, the residue was partitioned between EtOAc and H2O. The vacuum dried H2O layer was reacted in 1 mL pyridine with 1 mg/mL L-cysteine methyl ester at 60 °C for 1 h, followed by adding o-tolylisothiocyanate (2.0 μL) for an additional hour. The residue given by concentration was then dissolved in MeOH and analyzed by HPLC, with isocratic elution of CH3OH– H2O (50:50, v/v) for 30 min at a flow rate of 1.0 mL/min. Dglucose was detected in the acid hydrolysate by comparing the retention time of its derivatives with that of the authentic Dglucose derivative (tR 18.28 min) prepared in the same manner [14]. 2.5. Inhibition of self-induced Aβ aggregation Inhibition of self-induced Aβ1–42 aggregation was measured using a thioflavin T (ThT)-binding assay [15]. Hexafluoroisopropanol (HFIP) pretreated Aβ1–42 samples were resolubilized with dimethyl sulfoxide (DMSO) and diluted with 50 mM phosphate buffer (pH 7.4) to give a 25 μM solution. All compounds were firstly prepared in DMSO at a concentration of 200 μM. Then 1 μL of each compound and 9 μL of 25 mM Aβ1–42 sample were added to the well of black, opaque Corning 96well plates and the samples were mixed by gentle trapping. The
Table 1 1 H NMR (500 MHz) and 13C NMR (125 MHz) data for 1–4 in CD3OD. Position
1 δH(multi, J in Hz)
1 2 3 4 5 6 7 8 9 1′ 2′ 3′ 4′ 5′ 6′ 7′ 8′ 9′ CH3O-3 CH3O-3′ 1″ 2″ 3″ 4″ 5″ 6″
7.04 (d, 1.5)
6.77 (d, 8.1) 6.83 (dd, 8.1,1.5) 4.85 (d, 3.5) 4.39 (dq, 6.2, 3.5) 1.16 (d, 6.2) 6.69 (d, 1.8)
6.83 (d, 8,2) 6.60 (dd, 8.2, 1.8) 2.56 (t, 7.5) 1.79 (m) 3.55 (t, 6.5) 3.85 (s)
2 δC 133.6 111.5 148.8 146.9 115.8 120.5 76.3 81.5 13.8 137.8 117.0 149.1 144.7 117.8 120.6 32.5 35.5 62.3 56.4
δH(multi, J in Hz) 7.02 (d, 1.8)
6.80 (d, 8.1) 6.86 (dd, 8.1,1.8) 4.64 (d, 6.8) 4.22 (dq, 6.8, 6.5) 1.17 (d, 6.5) 6.74 (d, 2.0)
6.87 (d, 8.1) 6.62 (dd, 8.1, 2.0) 2.60 (t. 7.5) 1.82 (m) 3.59 (t, 6.5) 3.87 (s)
3 and 4 δC 134.1 111.7 149.0 147.4 115.9 121.1 78.7 83.1 17.2 138.4 117.2 149.4 145.4 118.8 120.5 32.6 35.5 62.3 56.4
δH(multi, J in Hz) 7.09 (d, 1.6)
7.17 (d, 8.3) 6.97 (dd, 8.3,1.6) 5.11 (d, 8.8) 3.37 (m) 1.38 (d, 6.8) 6.71 (s)
6.65 (s) 2.64 (t, 7.5) 1.82 (m) 3.57 (t, 6.5) 3.86 (s) 3.85 (s) 4.92 (d,7.3) 3.50 (m) 3.41 (m) 3.41 (m) 3.41 (m) 3.69 (m) 3.87 (m)
δC 137.1 111.7 151.1 148.0 118.1 120.1 94.1 47.2 18.4 137.3 113.9 145.2 146.7 134.4 117.0 32.9 35.8 62.2 56.8 56.8 102.9 75.0 77.9 71.4 78.2 62.6
X. Liu et al. / Fitoterapia 103 (2015) 136–142
139
final concentration of each compound was 20 μM and was prepared in independent triplicates. The solvent control was also included. Plates were covered to minimize evaporation and incubated in the dark at 37 °C for 48 h. After incubation, 200 μL of 5 μM ThT in 50 mM glycine–NaOH buffer (pH 8.0) was added to each well. Fluorescence was measured with excitation and emission wavelengths at 446 nm and 490 nm, respectively. The fluorescence intensities were compared and the inhibition percent was calculated by the following formula: 100 − (IFi / IFo × 100) where IFi and IFo were the fluorescence intensities obtained for Aβ1–42 in the presence and in the absence of inhibitor, respectively. 2.6. Docking study Molecular modeling calculations and docking studies were performed using Molecular Operating Environment (MOE) software version 2008.10 (Chemical Computing Group, Montreal, Canada). The X-ray crystal structure of Aβ1–42 (PDB 1IYT) used in the docking study was obtained from the Protein Data Bank. Heteroatoms and water molecules in the PDB file were removed, and all hydrogen atoms were subsequently added to the protein. The enzyme and the partial charges were calculated with Amber99 force field. Protonate states of the enzyme at pH 7 were obtained by following the Protonate 3D protocol in which all configurations were set as default. Compounds 1a, 1b, 2a and 2b were drawn in MOE with all hydrogen atoms added. During the docking procedure, poses of compounds 1a, 1b, 2a and 2b were initially generated by Triangle Matcher method, and scored with London dG function. Thirty poses of each compound, which were fine tune with the force field refinement scheme, were dedicated to the next refinement procedure. The best 10 poses of molecules were retained and scored. After docking, the geometry of resulting complex was studied using the MOE's pose viewer utility. 3. Results and discussion Compound 1 was obtained as colorless gum. Its molecular formula C19H24O6 was determined from the HRESIMS quasimolecular ion peak [M + Na]+ at m/z 371.1464 (calcd for C19H24NaO6, 371.1465). The IR absorbance bands at 1634 and 1513 cm−1 and UV bands at λmax 208 and 280 nm suggested the existence of benzene moieties. The 1H NMR spectrum showed signals of two ABX substituted aromatic rings [δH 7.04 (d, J = 1.5 Hz), 6.83 (dd, J = 8.1, 1.5 Hz), 6.77 (d, J = 8.1Hz ) and δH 6.83 (d, J = 8.2 Hz), 6.69 (d, J = 1.8 Hz), 6.60 (dd, J = 8.2, 1.8 Hz)], two oxygenated methine [δH 4.85 (d, J = 3.9 Hz), 4.39 (dq, J = 6.2, 3.9 Hz)], one methoxyl group (δH 3.85, s), one oxidated methylene [δH 3.55 (t, J = 6.5 Hz)], two methylenes [δH 2.56 (t, J = 7.5 Hz) and 1.79 (m)] and one doublet methyl [δH 1.16 (d, J = 6.2 Hz)]. With the help of HSQC, its 13C NMR also showed 19 carbon signals corresponding for the relative groups above, indicating the lignan nature of 1. In HMBC spectrum, correlations of H-8 with C-1, C-9 and C-4′, of H-7 with C-2, C-6 and C-9, and H-7′ with C-2′, C-6′ and C-9′ established that 1 is an 8-O-4′ neolignan possessing one methoxyl group and two phenolic hydroxyl groups. Detailed analysis of the HMBC correlations (Fig. 2), along with comparison of the NMR data revealed the similar structure of 1 with erythro-1-(4-hydroxy-3-methoxyphenyl)-2-[2-hydroxy-4-(3-
Fig. 2. Selected HMBC (H → C) correlations of 1.
hydroxypropyl)phenoxy]-1,3-propanediol [16,17], but the hydroxymethyl of C-9 in the above known compound was reduced as methyl in 1. In addition, the relative configuration of C-7 and C-8 was considered to be erythro due to the small coupling constant (J = 3.5 Hz) between H-7 and H-8 [18]. Its ECD spectrum did not give obvious cotton effect, indicating the racemic nature of 1. To confirm, a chiral HPLC–CD coupling analysis was carried out, and it clearly gave two phasecontrasted chromatographic peaks which further were prepared as optical pure 1a and 1b. The positive cotton effect at ca. 239 nm in the ECD spectrum of 1a indicated the 7R and 8S configurations, while 1b was confirmed as 7S and 8R configurations due to the negative cotton effect [19,20]. Based on quantum mechanical time-dependent density functional theory (TDDFT), ECD calculations were further used to confirm their absolute configurations, which matched well with the experimental ones (Fig. 3). Thus, the structure of 1a was established as (7R, 8S)-3′,4,7,9′-tetrahydroxy-3-methoxy-8-O4′-neolignan and 1b was identified as (7S, 8R)-3′,4,7,9′tetrahydroxy-3-methoxy-8-O-4′-neolignan. Compound 2, an isomer of 1, also had a molecular formula of C19H24O6 on the basis of a quasi-molecular ion at m/z 366.1912 [M + NH4]+ (calcd for C19H28NO6, 366.1911). Similar
Fig. 3. Calculated and experimental CD spectra of 1a and 1b.
140
X. Liu et al. / Fitoterapia 103 (2015) 136–142
1
H and 13C NMR spectra combined with 2D NMR analysis confirmed the same planar structure as 1. But compared with 1, the lager coupling constant of H-7/H-8 (J = 6.8 Hz) and the downfield-shift of C-7, C-8, and C-9 revealed the threo relationship of C-7 and C-8 in 2. Compound 2 was also optically inactive and resolved into a pair of enantiomers 2a and 2b. The cotton effects at ca. 232 nm (negative for 2a and positive for 2b) also confirmed the absolute configurations as 7R, 8R for 2a and 7S, 8S for 2b, respectively [19,20]. Consequently, 2a and 2b were identified as (7R, 8R)-3′,4,7,9′-tetrahydroxy-3-methoxy8-O-4′-neolignan and (7S, 8S)-3′,4,7,9′-tetrahydroxy-3methoxy-8-O-4′-neolignan. Compounds 3 and 4 were obtained as a mixture at first, and showed the same chromatographic behavior in both HPLC and TLC analysis. In HRESIMS, the mixture only displayed one quasi-molecular ion peak, [M + NH4]+ at m/z 524.2489 (calcd for C26H38NO10, 524.2490), consistent with the molecular formula C26H34O10. The 1H and 13C NMR spectra also deceptively gave data as a pure compound which structurally resembled to the known compound urolignoside [21], a dihydrobenzofurantype neolignan glycoside. Careful analysis of the 1H and 13C NMR data and HMBC correlations (Fig. 4) revealed the only difference between them was that the hydroxymethyl of C-9 in urolignoside was reduced to a methyl in the mixture. As for the stereochemistry, the anomeric proton showed the large coupling constant (J = 7.3 Hz), establishing the β-glucosidic linkage, while acid hydrolysis of the mixture afforded Dglucose based on HPLC analysis of its chiral derivatives [14]. The relative configuration of C-7 and C-8 was assigned as trans by the coupling constant (J = 8.8 Hz) [22,23]. The ECD spectrum of the mixture was then measured, but surprisingly gave no obvious cotton effect, suggesting the mixture nature. Reanalysis of the 13C NMR spectrum revealed the both double but extremely close signals for C-2 (δc 94.07, 94.11) and C-7 (δc 111.66, 111.71) (Fig. S16, Supplementary data), which further confirms the above speculation. The highly congruent NMR data and the determined D-glucose indicated the presence of enantiometric aglycones in the mixture [24]. Finally, the separation was achieved by chiral HPLC, affording compounds 3 and 4. In the ECD spectrum of 3, the negative cotton effect at ca. 240 nm and the positive cotton effect at ca. 220 nm confirmed 7R, 8R-configurations. Meanwhile, compared with 3, the almost symmetrical ECD curve of 4 confirmed its absolute configurations as 7S, 8S-configurations [21,24,25]. Thus, 3 and 4 were
elucidated as (7R, 8R)-9-dehydroxy-vladinol F-4-O-β-D-glucoside and (7S, 8S)-9-dehydroxy-vladinol F-4-O-β-D-glucoside, respectively. Two known compounds were identified as (+)isolariciresinol (5) [26], rel-(2α, 3β)-7-O-methylcedrusin (6) [27,28] by comparing their spectroscopic data with published values. All the compounds were evaluated for their inhibitions on self-induced Aβ1–42 aggregation through a ThT-based fluorometric assay [14], with resveratrol as a reference compound. The results (Table 2) showed that all the compounds exhibited good inhibition of Aβ1–42 aggregation (56.2–81.1% at 20 μM) relative to that of resveratrol (70.8% at 20 μM). Although the four compounds from 1a to 2b had the same planar structure, they did have different effects in this bioassay. Structurally, the threo 7,8-diol compounds were more efficient than the erythro ones, and the 7S/8S configuration (2b) showed the most effective inhibition (81.1% at 20 μM). Docking study (Fig. 5) was carried out to explain this interesting result. As shown in Fig. 5, 3′-OH, 7-OH of compound 2b were bound to the His6, Glu3 residues of Aβ1–42 via hydrogen bond interaction, respectively, and the benzene ring of compound 2b was bound to the Tyr10 of Aβ1–42 via Π–Π stacking interaction. Meanwhile, compounds from 1a to 2a showed only two interactions including one or no hydrogen bond interaction with Aβ1–42. Moreover, the hydrogen bond interaction played more important roles than Π–Π stacking interaction in the stability of the complex [29]. Thus, the docking study was well consistent with the results of ThT measurement, and strongly proved that compound 2b showed the optimal Aβ1–42 aggregation inhibition. Compound 4 showed similar Aβ1–42 aggregation inhibition (80.1% at 20 μM) with compound 2b, although they had different lignan parent nucleus. It is generally accepted that the accumulation of Aβ aggregates is a characteristic of AD and the aggregates have neurotoxic effects [12]. Thus, these lignans especially 2b and 4 might be promising lead compounds for further research on AD treatment. Acknowledgments This research work was funded by the National Natural Science Foundation of China (81073009), the Program for Changjiang Scholars and Innovative Research Team in University (IRT1193), the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD), the Youth Fund Project of the Basic Research Program of Jiangsu Province Table 2 Inhibition of Aβ1–42 self-induced aggregation.
Fig. 4. Selected HMBC (H → C) correlations of 3 and 4.
Compound
Inhibition of Aβ1–42 aggregation (%)a
Resveratrol 1a 1b 2a 2b 3 4 5 6
70.8 69.1 61.1 74.6 81.1 74.9 80.1 56.2 67.6
± ± ± ± ± ± ± ± ±
1.6 4.7 3.2 4.3 1.0 4.5 3.8 0.7 6.5
a The thioflavin-T fluorescence method was used. Values are expressed as mean ± SD from at least two independent measurements. All values were obtained at 20 μM concentration of the tested compounds.
X. Liu et al. / Fitoterapia 103 (2015) 136–142
141
Fig. 5. Docking models of compounds 2b (A and B), 2a (C), 1a (D) and 1b (E) with Aβ1–42 (PDB code 1IYT) generated with MOE. (A, C, D and E) 2D schematic diagrams of docking models of compounds from 1a to 2b with Aβ1–42. (B) 3D docking model of compound 2b with Aβ1–42.
(Natural Science Foundation, BK20130651), and the Fundamental Research Funds for the Central Universities (JKQZ2013016). Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.fitote.2015.03.011. References [1] Yoshikawa M, Uchida E, Kawaguchi A, Kitagawa I, Yamahara J. Galloyloxypaeoniflorin, suffruticoside-A, suffruticoside-B, suffruticoside-C, and suffruticoside-D, 5 new antioxidative glycosides, and suffruticoside-E, a paeonol glycoside, from Chinese Moutan cortex. Chem Pharm Bull 1992; 40:2248–50. [2] Matsuda H, Ohta T, Kawaguchi A, Yoshikawa M. Bioactive constituents of Chinese natural medicines. VI. Moutan cortex. (2): structures and radical scavenging effects of suffruticosides A, B, C, D, and E and galloyloxypaeoniflorin. Chem Pharm Bull 2001;49:69–72. [3] Murakami N, SaKa M, Shimada H, Matsuda H, Yamahara J, Yoshikawa M. New bioactive monoterpene glycosides from Paeoniae radix. Chem Pharm Bull 1996;44:1279–81. [4] Xu HM, Wei W, Jia XY. Effects and mechanisms of total glucosides of paeony on adjuvant arthritis in rats. J Ethnopharmacol 2007;109:442–8. [5] State Administration of Traditional Chinese Medicine. Zhong hua ben cao. Shanghai: Shanghai Science and Technology Publishing Company; 1999 2104–5. [6] Yu HY, Liu MG, Liu DN, Shang GW, Wang Y, Qi C, et al. Antinociceptive effects of systemic paeoniflorin on bee venom-induced various phenotypes of nociception and hypersensitivity. Pharmacol Biochem Behav 2007;88:131–40. [7] Lin JP, Xiao LB, Ouyang GL, Shen Y, Huo RF, Zhou Z, et al. Total glucosides of paeony inhibits Th1/Th17 cells via decreasing dendritic cells activation in rheumatoid arthritis. Cell Immunol 2012;280:156–63. [8] Ding HY, Lin HC, Teng CM, Wu YC. Phytochemical and pharmacological studies on Chinese Paeonia species. J Chin Chem Soc 2000;47:381–8.
[9] Zhong SZ, Ge QH, Qu R, Li Q, Ma SP. Paeonol attenuates neurotoxicity and ameliorates cognitive impairment induced by D-galactose in ICR mice. J Neurol Sci 2009;277:58–64. [10] Wang HB, Gu WF, Chu WJ, Zhang S, Tang XC, Qin GW. Monoterpene glucosides from Paeonia lactiflora. J Nat Prod 2009;72:1321–4. [11] Lee SM, Yoon MY, Park HR. Protective effects of Paeonia lactiflora Pall on hydrogen peroxide-induced apoptosis in PC12 cells. Biosci Biotechnol Biochem 2008;72:1272–7. [12] Wu SH, Wu DG, Chen YW. Chemical constituents and bioactivities of plants from the genus Paeonia. Chem Biodivers 2010;7:90–104. [13] Hardy J, Selkoe DJ. The amyloid hypothesis of Alzheimer's disease: progress and problems on the road to therapeutics. Science 2002;297: 353–6. [14] Lv HW, Zhu MD, Luo JG, Kong LY. Antihyperglycemic glucosylated coumaroyltyramine derivatives from Teucrium viscidum. J Nat Prod 2014; 77:200–5. [15] Reinke AA, Gestwicki JE. Structure–activity relationships of amyloid betaaggregation inhibitors based on curcumin: influence of linker length and flexibility. Chem Biol Drug Des 2007;3:206–15. [16] Ouyang F, Liu Y, Li R, Li L, Wang NL, Yao XS. Five lignans and an iridoid from Sambucus williamsii. Chin J Nat Med 2011;9:0026–9. [17] Fang JM, Lee CK, Cheng YS. Lignans from leaves of Juniperus-chinensis. Phytochemistry 1992;31:3659–61. [18] Xue CB, Chai DW, Jin XJ, Bi YR, Yao XJ, Wu WS, et al. Triterpenes and neolignans from the roots of Nannoglottis carpesioides. Phytochemistry 2011;72:1804–13. [19] Liao SG, Wu Y, Yue JM. Lignans from Wikstroemia hainanensis. Helv Chim Acta 2006;89:73–80. [20] Greca MD, Molinaro A, Monaco P, Previter L. Neolignans from Arum Ztalzcum. Phytochemistry 1994;35:777–9. [21] Shen YC, Hsieh PW, Kuo YH. Neolignan glucosides from Jasminsum Urophyllun. Phytochemistry 1998;48:719–23. [22] Kim CS, Kwon OW, Kim SY, Lee KR. Bioactive lignans from the trunk of Abies holophylla. J Nat Prod 2013;76:2131–5. [23] Valcic S, Montenegro G, Timmermann BN. Lignans from Chilean propolis. J Nat Prod 1998;61:771–5. [24] Matsuda N, Sato H, Yaoita Y, Kikuchi M. Isolation and absolute structures enantiometric glycones from the of the leaves of Vthurnum awabuki K Koch. Chem Pharm Bull 1996;44:1122–3.
142
X. Liu et al. / Fitoterapia 103 (2015) 136–142
[25] Nakanishi T, Iida N, Inatomi Y, Murata H, Inada A, Murata J, et al. Neolignan and flavonoid glycosides in Juniperus communis var Depressa. Phytochemistry 2004;65:207–13. [26] Jutiviboonsuk A, Zhang HJ, Tan GT, Ma CY, Van Hung N, Cuong NM, et al. Bioactive constituents from roots of Bursera tonkinensis. Phytochemistry 2005;66:2745–51. [27] Song CW, Wang SM, Zhou LL, Hou FF, Wang KJ, Han QB, et al. Isolation and identification of compounds responsible for antioxidant capacity of Euryale ferox seeds. J Agric Food Chem 2011;59:1199–204.
[28] Seidel V, Bailleul F, Waterman PG. Novel oligorhamnosides from the stembark of Cleistopholis glauca. J Nat Prod 2000;63:6–11. [29] Jiang N, Li SY, Xie SS, Li ZR, Wang K, Wang XB, et al. Design, synthesis and evaluation of multifunctional salphen derivatives for the treatment of Alzheimer's disease. Eur J Med Chem 2014;87:540–51.
Contents lists available at ScienceDirect
Fitoterapia journal homepage: www.elsevier.com/locate/fitote
Lignans from the root of Paeonia lactiflora and their anti-β-amyloid aggregation activities Xiao Liu a, Ming-hua Yang a,⁎, Xiao-bing Wang a, Sai-sai Xie a, Zhong-rui Li a, Dong-hynu Kim b, Jun-seong Park b, Ling-yi Kong a,⁎⁎ a State Key Laboratory of Natural Medicines, Department of Natural Medicinal Chemistry, China Pharmaceutical University, 24 Tong Jia Xiang, Nanjing 210009, People's Republic of China b Amore-Pacific R&D Center, 314-1, Bora-dong, Giheung-gu, Yongin-si, Gyeonggi-Do, 446-729, South Korea
a r t i c l e
i n f o
Article history: Received 18 January 2015 Accepted in revised form 7 March 2015 Available online 27 March 2015 Keywords: Paeonia lactiflora Lignan Enantiomeric separation β-amyloid aggregation
a b s t r a c t Four new neolignans (1–4), together with two known lignans (5 and 6), were isolated from the root of Paeonia lactiflora. Compounds 1 and 2 were two racemates and were separated by chiral high performance liquid chromatography (HPLC) to give all of the four stereoisomeric forms sharing a common planar structure. Compounds 3 and 4 were two neolignan glycoside diastereomers but interestingly appeared to be enantiomers: they had the extremely similar 1H and 13C NMR spectra and had to be solved only by chiral HPLC. Their structures were determined by spectroscopic analysis, including 1D and 2D NMR, HRESIMS and electronic circular dichroism experiments. All compounds were evaluated for their inhibitory effects on β-amyloid aggregation, and the optical pure compound 2b was found to show the optimal Aβ1–42 aggregation inhibition potency (81.1% at 20 μM). In addition, despite large amount of chemical studies performed on genus Paeonia, the lignans were reported for the first time. © 2015 Published by Elsevier B.V.
1. Introduction Paeonia lactiflora (Ranunculaceae), widely distributed in China, Korea, and Russia, is a herb with ornate flowers blooming in spring. The root of P. lactiflora (Chi-Shao and Bai-Shao) is one of the most important crude drugs used as antioxidant [1,2], analgesic [3], and anti-inflammatory [4] agent in Asia. It is commonly used as a remedy for cardiovascular disorders, extravagated blood, and rheumatoid arthritis [5–8]. Modern pharmacological research demonstrates that P. lactiflora, with effects of neuroprotectivity and ameliorating cognitive impairment, is a potential therapeutic strategy for the treatment of Alzheimer′s disease (AD) [9–11]. Monoterpenoid glycosides, the main ingredient of the genus Paeonia, exhibited the above
⁎ Corresponding author. Tel.: +86 25 8618 5039. ⁎⁎ Corresponding author. Tel./fax: +86 25 8327 1405. E-mail addresses: [email protected] (M. Yang), [email protected] (L. Kong).
http://dx.doi.org/10.1016/j.fitote.2015.03.011 0367-326X/© 2015 Published by Elsevier B.V.
protective effect and ameliorated oxidative stress damage occurred in the pathological courses of AD [10,11]. In addition, triterpenoids, flavonoids, phenols, and tannins which were related to several different bioactivities have also been isolated from Paeonia [12]. In our recent search for bioactive constituents from P. lactiflora, lignans (Fig. 1) including two pairs of new enantiomeric 8-O-4′ neolignans (1a and 1b, 2a and 2b), two new benzofuran neoligans (3 and 4), and two known lignans (5 and 6) were obtained from this genus for the first time. Their structures were elucidated on the basis of spectroscopic methods including 1D and 2D NMR, and HRMS, and their stereochemistry was determined by acid hydrolysis and electronic circular dichroism (ECD) experiments. Although these lignans did not have obvious neuroprotective activities in hydrogen peroxide-induced PC12 cell damage bioassay, they had good inhibitory potency toward self-induced Aβ1–42 aggregation (56.2–81.1% at 20 μM). According to the “amyloid hypothesis” [13], the production and accumulation of oligomeric β-amyloid (Aβ) aggregates in the brain are a central
X. Liu et al. / Fitoterapia 103 (2015) 136–142
137
Fig. 1. Structures of compounds 1–6.
event in the pathogenesis of AD. Thus, the discovery of these lignans indicated a new potential approach of P. lactiflora to treat AD. Herein, we discuss the purification, structure determination, and Aβ aggregation inhibitory effects of these isolates.
A voucher specimen (No. 2013-PL) has been deposited at the Department of Natural Medicinal Chemistry, China Pharmaceutical University.
2. Experimental
2.3. Extraction and isolation
2.1. General
The air-dried root of P. lactiflora (10.0 kg) was powdered and extracted with 60% EtOH for three times affording dark residue (4.1 kg) after solvent evaporation under reduced pressure. The residue was subjected to D-101 macroporous resin column chromatography (CC) with a step gradient of EtOH– H2O solvent system (10:90, 30:70, 50:50, and 95:5, v/v) to yield four fractions (A–D). Fr. C (210 g) was then subjected to silica gel CC eluted by a step gradient of CH2Cl2–MeOH solvent system (50:1, 20:1, 10:1, 5:1, and 1:1, v/v) to obtain eight subfractions (C1–C8). Fr. C4 was further separated via silica gel CC, ODS CC, and MPLC successively to produce six subfractions (C4.2.1.1–C4.2.1.6). Fr. C4.2.1.2 was subjected to preparative HPLC with MeOH–H2O (40:60, v/v) to yield 5 (4.6 mg). Fr. C4.2.1.4 and Fr. C4.2.1.5 were applied to preparative HPLC with MeOH–H2O (45:55, v/v) to acquire 1 (8.2 mg) and 2 (3.9 mg), respectively. In addition, direct enantiomeric separation of compounds 1 and 2 was achieved by preparative chiral HPLC with 2-propanol–n-hexane (25:75, v/v), affording compounds 1a (3.0 mg) and 1b (3.9 mg), 2a (1.8 mg) and 2b (1.4 mg), respectively. Fr. C4.2.1.3 was chromatographed over Sephadex LH-20 with MeOH to produce five subfractions (C4.2.1.3.1– C4.2.1.3.5). Fr. C4.2.1.3.1 was then subjected to preparative HPLC with MeOH–H2O (45:55, v/v) to afford 6 (5.0 mg). Fr. C6 was chromatographed over MCI, ODS, silica gel and Sephadex LH-20 respectively to obtain Fr. C6.2.6.5.3 and Fr. C6.2.6.5.3 was further purified by preparative HPLC (MeOH–H2O, 40:60, v/v) to obtain the mixture (6.1 mg) of 3 and 4. Compounds 3 (3.2 mg) and 4 (2.4 mg) were separated using preparative chiral chromatography with 2-propanol–n-hexane (32:68, v/v).
Optical rotations were determined with a JASCO P-1020 polarimeter. ECD spectra were carried out on a JASCO-810 spectropolarimeter. UV spectra were obtained on a Shimadzu UV-2501 spectrophotometer. IR (KBr disks) spectra were measured by a Bruker Tensor 27 spectrometer. NMR spectra were recorded on a Bruker Avance III NMR (1H: 500 MHz, 13C: 125 MHz) instrument at 300 K, with TMS as internal standard. HRESIMS was carried out on an Agilent UPLC-Q-TOF (6520B) mass spectrometer. Fluorescence was measured on SpectraMax M5 (Molecular Devices, Sunnyvale, CA, USA) multi-mode plate reader. Chiral HPLC was done with a Daicel Chiralpak AD-H column (250 × 4.6 mm). Preparative HPLC was carried out using a SHMADZU LC-6A series instrument with a Shim-park RP-C18 column (200 × 20 mm) and a SHMADZU SPD-20A detector. Chiralpak AD-H preparative columns (250 × 10 mm) were purchased from Daicel Chemical Ltd. (Shanghai, China). Silica gel (Qingdao Haiyang Chemical Co., Ltd., Qingdao, China), RP-C18 (40–63 μm, Fuji, Japan), Sephadex LH-20 (Pharmacia, USA) and macroporous resin D-101 (poresize B 13–14 nm, 26– 60 mesh, Qingdao Haiyang Chemical Co. Ltd., China) were used for column chromatography. 2.2. Plant material The root of P. lactiflora was collected from Sinwol-ri, Yuchimyeon, Jangheung-gun, Jeollanam-do, Korea in March 2013 and authenticated by Professor Mian Zhang, Department of Natural Medicinal Chemistry, China Pharmaceutical University.
138
X. Liu et al. / Fitoterapia 103 (2015) 136–142
Compound 1: colorless gum; UV (MeOH) λmax (log ε) 208 (4.46), 280 (3.81) nm; IR (KBr) νmax 3456, 1634, 1513, 1384, 1274, 1033, 639 cm−1; 1H NMR and 13C NMR (CD3OD) data see Table 1; HRESIMS m/z 371.1464 [M + Na]+ (calcd for C19H24NaO6, 371.1465). 1a: [α]25 D +3.3 (c 0.12, MeOH); CD (2.0 × 10−4, MeOH) λmax (Δε) 239 (+2.25) nm. 1b: [α]25 D −4.7 (c 0.14, MeOH); CD (2.0 × 10−4, MeOH) λmax (Δε) 239 (−2.75) nm. Compound 2: pale yellow gum; UV (MeOH) λmax (log ε) 208 (4.46), 280 (3.81) nm; IR (KBr) νmax 3450, 1638, 1384, 1035, 669 cm−1; 1H NMR and 13C NMR (CD3OD) data see Table 1; HRESIMS m/z 366.1912 [M + NH4]+ (calcd for C19H28NO6, 366.1911). 2a: [α]25 D −21.7 (c 0.06, MeOH); CD (2.0 × 10−4, MeOH) λmax (Δε) 231 (−3.45) nm. 2b: [α]25 D +16.0 (c 0.04, MeOH); CD (1.5 × 10−4, MeOH) λmax (Δε) 232 (+2.83) nm. Compound 3: white, amorphous powder; [α]25 D −17.5 (c 0.10, MeOH); UV (MeOH) λmax (log ε) 208 (4.31), 280 (3.32) nm; CD (1.5 × 10−4, MeOH) Δε 219 nm 0.58, 238 nm −0.64; IR (KBr) νmax 3449, 1638, 1399, 1072, 669 cm−1; 1H NMR and 13 C NMR (CD3OD) data see Table 1; HRESIMS m/z 524.2485 [M + NH4]+ (calcd for C26H38NO10, 524.2490). Compound 4: white, amorphous powder; [α]25 D −11.5 (c 0.11, MeOH); UV (MeOH) λmax (log ε) 208 (4.31), 280 (3.32) nm; CD (1.5 × 10−4 MeOH) Δε 220 nm −2.19, Δε 245 nm 0.72; IR (KBr)
νmax 3447, 1638, 1398, 1074, 669 cm−1; 1H NMR and 13C NMR (CD3OD) data see Table 1; HRESIMS m/z 524.2491 [M + NH4]+ (calcd for C26H38NO10, 524.2490). 2.4. Acid hydrolysis and HPLC analysis of 3 and 4 The mixture of compounds 3 and 4 (1.0 mg) was dissolved in 2 N HCl (2.0 mL) at 100 °C for 4 h. After removal of the solvent, the residue was partitioned between EtOAc and H2O. The vacuum dried H2O layer was reacted in 1 mL pyridine with 1 mg/mL L-cysteine methyl ester at 60 °C for 1 h, followed by adding o-tolylisothiocyanate (2.0 μL) for an additional hour. The residue given by concentration was then dissolved in MeOH and analyzed by HPLC, with isocratic elution of CH3OH– H2O (50:50, v/v) for 30 min at a flow rate of 1.0 mL/min. Dglucose was detected in the acid hydrolysate by comparing the retention time of its derivatives with that of the authentic Dglucose derivative (tR 18.28 min) prepared in the same manner [14]. 2.5. Inhibition of self-induced Aβ aggregation Inhibition of self-induced Aβ1–42 aggregation was measured using a thioflavin T (ThT)-binding assay [15]. Hexafluoroisopropanol (HFIP) pretreated Aβ1–42 samples were resolubilized with dimethyl sulfoxide (DMSO) and diluted with 50 mM phosphate buffer (pH 7.4) to give a 25 μM solution. All compounds were firstly prepared in DMSO at a concentration of 200 μM. Then 1 μL of each compound and 9 μL of 25 mM Aβ1–42 sample were added to the well of black, opaque Corning 96well plates and the samples were mixed by gentle trapping. The
Table 1 1 H NMR (500 MHz) and 13C NMR (125 MHz) data for 1–4 in CD3OD. Position
1 δH(multi, J in Hz)
1 2 3 4 5 6 7 8 9 1′ 2′ 3′ 4′ 5′ 6′ 7′ 8′ 9′ CH3O-3 CH3O-3′ 1″ 2″ 3″ 4″ 5″ 6″
7.04 (d, 1.5)
6.77 (d, 8.1) 6.83 (dd, 8.1,1.5) 4.85 (d, 3.5) 4.39 (dq, 6.2, 3.5) 1.16 (d, 6.2) 6.69 (d, 1.8)
6.83 (d, 8,2) 6.60 (dd, 8.2, 1.8) 2.56 (t, 7.5) 1.79 (m) 3.55 (t, 6.5) 3.85 (s)
2 δC 133.6 111.5 148.8 146.9 115.8 120.5 76.3 81.5 13.8 137.8 117.0 149.1 144.7 117.8 120.6 32.5 35.5 62.3 56.4
δH(multi, J in Hz) 7.02 (d, 1.8)
6.80 (d, 8.1) 6.86 (dd, 8.1,1.8) 4.64 (d, 6.8) 4.22 (dq, 6.8, 6.5) 1.17 (d, 6.5) 6.74 (d, 2.0)
6.87 (d, 8.1) 6.62 (dd, 8.1, 2.0) 2.60 (t. 7.5) 1.82 (m) 3.59 (t, 6.5) 3.87 (s)
3 and 4 δC 134.1 111.7 149.0 147.4 115.9 121.1 78.7 83.1 17.2 138.4 117.2 149.4 145.4 118.8 120.5 32.6 35.5 62.3 56.4
δH(multi, J in Hz) 7.09 (d, 1.6)
7.17 (d, 8.3) 6.97 (dd, 8.3,1.6) 5.11 (d, 8.8) 3.37 (m) 1.38 (d, 6.8) 6.71 (s)
6.65 (s) 2.64 (t, 7.5) 1.82 (m) 3.57 (t, 6.5) 3.86 (s) 3.85 (s) 4.92 (d,7.3) 3.50 (m) 3.41 (m) 3.41 (m) 3.41 (m) 3.69 (m) 3.87 (m)
δC 137.1 111.7 151.1 148.0 118.1 120.1 94.1 47.2 18.4 137.3 113.9 145.2 146.7 134.4 117.0 32.9 35.8 62.2 56.8 56.8 102.9 75.0 77.9 71.4 78.2 62.6
X. Liu et al. / Fitoterapia 103 (2015) 136–142
139
final concentration of each compound was 20 μM and was prepared in independent triplicates. The solvent control was also included. Plates were covered to minimize evaporation and incubated in the dark at 37 °C for 48 h. After incubation, 200 μL of 5 μM ThT in 50 mM glycine–NaOH buffer (pH 8.0) was added to each well. Fluorescence was measured with excitation and emission wavelengths at 446 nm and 490 nm, respectively. The fluorescence intensities were compared and the inhibition percent was calculated by the following formula: 100 − (IFi / IFo × 100) where IFi and IFo were the fluorescence intensities obtained for Aβ1–42 in the presence and in the absence of inhibitor, respectively. 2.6. Docking study Molecular modeling calculations and docking studies were performed using Molecular Operating Environment (MOE) software version 2008.10 (Chemical Computing Group, Montreal, Canada). The X-ray crystal structure of Aβ1–42 (PDB 1IYT) used in the docking study was obtained from the Protein Data Bank. Heteroatoms and water molecules in the PDB file were removed, and all hydrogen atoms were subsequently added to the protein. The enzyme and the partial charges were calculated with Amber99 force field. Protonate states of the enzyme at pH 7 were obtained by following the Protonate 3D protocol in which all configurations were set as default. Compounds 1a, 1b, 2a and 2b were drawn in MOE with all hydrogen atoms added. During the docking procedure, poses of compounds 1a, 1b, 2a and 2b were initially generated by Triangle Matcher method, and scored with London dG function. Thirty poses of each compound, which were fine tune with the force field refinement scheme, were dedicated to the next refinement procedure. The best 10 poses of molecules were retained and scored. After docking, the geometry of resulting complex was studied using the MOE's pose viewer utility. 3. Results and discussion Compound 1 was obtained as colorless gum. Its molecular formula C19H24O6 was determined from the HRESIMS quasimolecular ion peak [M + Na]+ at m/z 371.1464 (calcd for C19H24NaO6, 371.1465). The IR absorbance bands at 1634 and 1513 cm−1 and UV bands at λmax 208 and 280 nm suggested the existence of benzene moieties. The 1H NMR spectrum showed signals of two ABX substituted aromatic rings [δH 7.04 (d, J = 1.5 Hz), 6.83 (dd, J = 8.1, 1.5 Hz), 6.77 (d, J = 8.1Hz ) and δH 6.83 (d, J = 8.2 Hz), 6.69 (d, J = 1.8 Hz), 6.60 (dd, J = 8.2, 1.8 Hz)], two oxygenated methine [δH 4.85 (d, J = 3.9 Hz), 4.39 (dq, J = 6.2, 3.9 Hz)], one methoxyl group (δH 3.85, s), one oxidated methylene [δH 3.55 (t, J = 6.5 Hz)], two methylenes [δH 2.56 (t, J = 7.5 Hz) and 1.79 (m)] and one doublet methyl [δH 1.16 (d, J = 6.2 Hz)]. With the help of HSQC, its 13C NMR also showed 19 carbon signals corresponding for the relative groups above, indicating the lignan nature of 1. In HMBC spectrum, correlations of H-8 with C-1, C-9 and C-4′, of H-7 with C-2, C-6 and C-9, and H-7′ with C-2′, C-6′ and C-9′ established that 1 is an 8-O-4′ neolignan possessing one methoxyl group and two phenolic hydroxyl groups. Detailed analysis of the HMBC correlations (Fig. 2), along with comparison of the NMR data revealed the similar structure of 1 with erythro-1-(4-hydroxy-3-methoxyphenyl)-2-[2-hydroxy-4-(3-
Fig. 2. Selected HMBC (H → C) correlations of 1.
hydroxypropyl)phenoxy]-1,3-propanediol [16,17], but the hydroxymethyl of C-9 in the above known compound was reduced as methyl in 1. In addition, the relative configuration of C-7 and C-8 was considered to be erythro due to the small coupling constant (J = 3.5 Hz) between H-7 and H-8 [18]. Its ECD spectrum did not give obvious cotton effect, indicating the racemic nature of 1. To confirm, a chiral HPLC–CD coupling analysis was carried out, and it clearly gave two phasecontrasted chromatographic peaks which further were prepared as optical pure 1a and 1b. The positive cotton effect at ca. 239 nm in the ECD spectrum of 1a indicated the 7R and 8S configurations, while 1b was confirmed as 7S and 8R configurations due to the negative cotton effect [19,20]. Based on quantum mechanical time-dependent density functional theory (TDDFT), ECD calculations were further used to confirm their absolute configurations, which matched well with the experimental ones (Fig. 3). Thus, the structure of 1a was established as (7R, 8S)-3′,4,7,9′-tetrahydroxy-3-methoxy-8-O4′-neolignan and 1b was identified as (7S, 8R)-3′,4,7,9′tetrahydroxy-3-methoxy-8-O-4′-neolignan. Compound 2, an isomer of 1, also had a molecular formula of C19H24O6 on the basis of a quasi-molecular ion at m/z 366.1912 [M + NH4]+ (calcd for C19H28NO6, 366.1911). Similar
Fig. 3. Calculated and experimental CD spectra of 1a and 1b.
140
X. Liu et al. / Fitoterapia 103 (2015) 136–142
1
H and 13C NMR spectra combined with 2D NMR analysis confirmed the same planar structure as 1. But compared with 1, the lager coupling constant of H-7/H-8 (J = 6.8 Hz) and the downfield-shift of C-7, C-8, and C-9 revealed the threo relationship of C-7 and C-8 in 2. Compound 2 was also optically inactive and resolved into a pair of enantiomers 2a and 2b. The cotton effects at ca. 232 nm (negative for 2a and positive for 2b) also confirmed the absolute configurations as 7R, 8R for 2a and 7S, 8S for 2b, respectively [19,20]. Consequently, 2a and 2b were identified as (7R, 8R)-3′,4,7,9′-tetrahydroxy-3-methoxy8-O-4′-neolignan and (7S, 8S)-3′,4,7,9′-tetrahydroxy-3methoxy-8-O-4′-neolignan. Compounds 3 and 4 were obtained as a mixture at first, and showed the same chromatographic behavior in both HPLC and TLC analysis. In HRESIMS, the mixture only displayed one quasi-molecular ion peak, [M + NH4]+ at m/z 524.2489 (calcd for C26H38NO10, 524.2490), consistent with the molecular formula C26H34O10. The 1H and 13C NMR spectra also deceptively gave data as a pure compound which structurally resembled to the known compound urolignoside [21], a dihydrobenzofurantype neolignan glycoside. Careful analysis of the 1H and 13C NMR data and HMBC correlations (Fig. 4) revealed the only difference between them was that the hydroxymethyl of C-9 in urolignoside was reduced to a methyl in the mixture. As for the stereochemistry, the anomeric proton showed the large coupling constant (J = 7.3 Hz), establishing the β-glucosidic linkage, while acid hydrolysis of the mixture afforded Dglucose based on HPLC analysis of its chiral derivatives [14]. The relative configuration of C-7 and C-8 was assigned as trans by the coupling constant (J = 8.8 Hz) [22,23]. The ECD spectrum of the mixture was then measured, but surprisingly gave no obvious cotton effect, suggesting the mixture nature. Reanalysis of the 13C NMR spectrum revealed the both double but extremely close signals for C-2 (δc 94.07, 94.11) and C-7 (δc 111.66, 111.71) (Fig. S16, Supplementary data), which further confirms the above speculation. The highly congruent NMR data and the determined D-glucose indicated the presence of enantiometric aglycones in the mixture [24]. Finally, the separation was achieved by chiral HPLC, affording compounds 3 and 4. In the ECD spectrum of 3, the negative cotton effect at ca. 240 nm and the positive cotton effect at ca. 220 nm confirmed 7R, 8R-configurations. Meanwhile, compared with 3, the almost symmetrical ECD curve of 4 confirmed its absolute configurations as 7S, 8S-configurations [21,24,25]. Thus, 3 and 4 were
elucidated as (7R, 8R)-9-dehydroxy-vladinol F-4-O-β-D-glucoside and (7S, 8S)-9-dehydroxy-vladinol F-4-O-β-D-glucoside, respectively. Two known compounds were identified as (+)isolariciresinol (5) [26], rel-(2α, 3β)-7-O-methylcedrusin (6) [27,28] by comparing their spectroscopic data with published values. All the compounds were evaluated for their inhibitions on self-induced Aβ1–42 aggregation through a ThT-based fluorometric assay [14], with resveratrol as a reference compound. The results (Table 2) showed that all the compounds exhibited good inhibition of Aβ1–42 aggregation (56.2–81.1% at 20 μM) relative to that of resveratrol (70.8% at 20 μM). Although the four compounds from 1a to 2b had the same planar structure, they did have different effects in this bioassay. Structurally, the threo 7,8-diol compounds were more efficient than the erythro ones, and the 7S/8S configuration (2b) showed the most effective inhibition (81.1% at 20 μM). Docking study (Fig. 5) was carried out to explain this interesting result. As shown in Fig. 5, 3′-OH, 7-OH of compound 2b were bound to the His6, Glu3 residues of Aβ1–42 via hydrogen bond interaction, respectively, and the benzene ring of compound 2b was bound to the Tyr10 of Aβ1–42 via Π–Π stacking interaction. Meanwhile, compounds from 1a to 2a showed only two interactions including one or no hydrogen bond interaction with Aβ1–42. Moreover, the hydrogen bond interaction played more important roles than Π–Π stacking interaction in the stability of the complex [29]. Thus, the docking study was well consistent with the results of ThT measurement, and strongly proved that compound 2b showed the optimal Aβ1–42 aggregation inhibition. Compound 4 showed similar Aβ1–42 aggregation inhibition (80.1% at 20 μM) with compound 2b, although they had different lignan parent nucleus. It is generally accepted that the accumulation of Aβ aggregates is a characteristic of AD and the aggregates have neurotoxic effects [12]. Thus, these lignans especially 2b and 4 might be promising lead compounds for further research on AD treatment. Acknowledgments This research work was funded by the National Natural Science Foundation of China (81073009), the Program for Changjiang Scholars and Innovative Research Team in University (IRT1193), the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD), the Youth Fund Project of the Basic Research Program of Jiangsu Province Table 2 Inhibition of Aβ1–42 self-induced aggregation.
Fig. 4. Selected HMBC (H → C) correlations of 3 and 4.
Compound
Inhibition of Aβ1–42 aggregation (%)a
Resveratrol 1a 1b 2a 2b 3 4 5 6
70.8 69.1 61.1 74.6 81.1 74.9 80.1 56.2 67.6
± ± ± ± ± ± ± ± ±
1.6 4.7 3.2 4.3 1.0 4.5 3.8 0.7 6.5
a The thioflavin-T fluorescence method was used. Values are expressed as mean ± SD from at least two independent measurements. All values were obtained at 20 μM concentration of the tested compounds.
X. Liu et al. / Fitoterapia 103 (2015) 136–142
141
Fig. 5. Docking models of compounds 2b (A and B), 2a (C), 1a (D) and 1b (E) with Aβ1–42 (PDB code 1IYT) generated with MOE. (A, C, D and E) 2D schematic diagrams of docking models of compounds from 1a to 2b with Aβ1–42. (B) 3D docking model of compound 2b with Aβ1–42.
(Natural Science Foundation, BK20130651), and the Fundamental Research Funds for the Central Universities (JKQZ2013016). Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.fitote.2015.03.011. References [1] Yoshikawa M, Uchida E, Kawaguchi A, Kitagawa I, Yamahara J. Galloyloxypaeoniflorin, suffruticoside-A, suffruticoside-B, suffruticoside-C, and suffruticoside-D, 5 new antioxidative glycosides, and suffruticoside-E, a paeonol glycoside, from Chinese Moutan cortex. Chem Pharm Bull 1992; 40:2248–50. [2] Matsuda H, Ohta T, Kawaguchi A, Yoshikawa M. Bioactive constituents of Chinese natural medicines. VI. Moutan cortex. (2): structures and radical scavenging effects of suffruticosides A, B, C, D, and E and galloyloxypaeoniflorin. Chem Pharm Bull 2001;49:69–72. [3] Murakami N, SaKa M, Shimada H, Matsuda H, Yamahara J, Yoshikawa M. New bioactive monoterpene glycosides from Paeoniae radix. Chem Pharm Bull 1996;44:1279–81. [4] Xu HM, Wei W, Jia XY. Effects and mechanisms of total glucosides of paeony on adjuvant arthritis in rats. J Ethnopharmacol 2007;109:442–8. [5] State Administration of Traditional Chinese Medicine. Zhong hua ben cao. Shanghai: Shanghai Science and Technology Publishing Company; 1999 2104–5. [6] Yu HY, Liu MG, Liu DN, Shang GW, Wang Y, Qi C, et al. Antinociceptive effects of systemic paeoniflorin on bee venom-induced various phenotypes of nociception and hypersensitivity. Pharmacol Biochem Behav 2007;88:131–40. [7] Lin JP, Xiao LB, Ouyang GL, Shen Y, Huo RF, Zhou Z, et al. Total glucosides of paeony inhibits Th1/Th17 cells via decreasing dendritic cells activation in rheumatoid arthritis. Cell Immunol 2012;280:156–63. [8] Ding HY, Lin HC, Teng CM, Wu YC. Phytochemical and pharmacological studies on Chinese Paeonia species. J Chin Chem Soc 2000;47:381–8.
[9] Zhong SZ, Ge QH, Qu R, Li Q, Ma SP. Paeonol attenuates neurotoxicity and ameliorates cognitive impairment induced by D-galactose in ICR mice. J Neurol Sci 2009;277:58–64. [10] Wang HB, Gu WF, Chu WJ, Zhang S, Tang XC, Qin GW. Monoterpene glucosides from Paeonia lactiflora. J Nat Prod 2009;72:1321–4. [11] Lee SM, Yoon MY, Park HR. Protective effects of Paeonia lactiflora Pall on hydrogen peroxide-induced apoptosis in PC12 cells. Biosci Biotechnol Biochem 2008;72:1272–7. [12] Wu SH, Wu DG, Chen YW. Chemical constituents and bioactivities of plants from the genus Paeonia. Chem Biodivers 2010;7:90–104. [13] Hardy J, Selkoe DJ. The amyloid hypothesis of Alzheimer's disease: progress and problems on the road to therapeutics. Science 2002;297: 353–6. [14] Lv HW, Zhu MD, Luo JG, Kong LY. Antihyperglycemic glucosylated coumaroyltyramine derivatives from Teucrium viscidum. J Nat Prod 2014; 77:200–5. [15] Reinke AA, Gestwicki JE. Structure–activity relationships of amyloid betaaggregation inhibitors based on curcumin: influence of linker length and flexibility. Chem Biol Drug Des 2007;3:206–15. [16] Ouyang F, Liu Y, Li R, Li L, Wang NL, Yao XS. Five lignans and an iridoid from Sambucus williamsii. Chin J Nat Med 2011;9:0026–9. [17] Fang JM, Lee CK, Cheng YS. Lignans from leaves of Juniperus-chinensis. Phytochemistry 1992;31:3659–61. [18] Xue CB, Chai DW, Jin XJ, Bi YR, Yao XJ, Wu WS, et al. Triterpenes and neolignans from the roots of Nannoglottis carpesioides. Phytochemistry 2011;72:1804–13. [19] Liao SG, Wu Y, Yue JM. Lignans from Wikstroemia hainanensis. Helv Chim Acta 2006;89:73–80. [20] Greca MD, Molinaro A, Monaco P, Previter L. Neolignans from Arum Ztalzcum. Phytochemistry 1994;35:777–9. [21] Shen YC, Hsieh PW, Kuo YH. Neolignan glucosides from Jasminsum Urophyllun. Phytochemistry 1998;48:719–23. [22] Kim CS, Kwon OW, Kim SY, Lee KR. Bioactive lignans from the trunk of Abies holophylla. J Nat Prod 2013;76:2131–5. [23] Valcic S, Montenegro G, Timmermann BN. Lignans from Chilean propolis. J Nat Prod 1998;61:771–5. [24] Matsuda N, Sato H, Yaoita Y, Kikuchi M. Isolation and absolute structures enantiometric glycones from the of the leaves of Vthurnum awabuki K Koch. Chem Pharm Bull 1996;44:1122–3.
142
X. Liu et al. / Fitoterapia 103 (2015) 136–142
[25] Nakanishi T, Iida N, Inatomi Y, Murata H, Inada A, Murata J, et al. Neolignan and flavonoid glycosides in Juniperus communis var Depressa. Phytochemistry 2004;65:207–13. [26] Jutiviboonsuk A, Zhang HJ, Tan GT, Ma CY, Van Hung N, Cuong NM, et al. Bioactive constituents from roots of Bursera tonkinensis. Phytochemistry 2005;66:2745–51. [27] Song CW, Wang SM, Zhou LL, Hou FF, Wang KJ, Han QB, et al. Isolation and identification of compounds responsible for antioxidant capacity of Euryale ferox seeds. J Agric Food Chem 2011;59:1199–204.
[28] Seidel V, Bailleul F, Waterman PG. Novel oligorhamnosides from the stembark of Cleistopholis glauca. J Nat Prod 2000;63:6–11. [29] Jiang N, Li SY, Xie SS, Li ZR, Wang K, Wang XB, et al. Design, synthesis and evaluation of multifunctional salphen derivatives for the treatment of Alzheimer's disease. Eur J Med Chem 2014;87:540–51.
