1 Introduction

About 30 species of Clausena (Rutaceae) are widely distributed in the world, and 10 of them exist in China. Clausena lansium (Lour.) Skeels is a fruit tree and distributes widely in south of China [1]. Its leaves and roots have been used as a folk herb for the treatment of cough, asthma, dermatological disease, viral hepatitis, and gastro-intestinal disease; and its seeds for treating acute and chronic gastrointestinal inflammation, and ulcer [2]. Caryophyllaceae-type cyclopeptides (CPs), carbazole alkaloids, coumarins, amides, and terpenoids have been isolated from C. lansium [3,4,5,6,7,8]. Among them, CPs are formed with the peptide bonds of protein or non-proten α-amino acid residues, which are homomonocyclopeptides with mainly five to twelve α-amino acid residues [9]. During this work, two new cyclic nonapeptides, named clausenlanins A (1) and B (2) (Fig. 1), were isolated from the roots and rhizomes of C. lansium. Because the 1H NMR signals are weak and severely overlapped taken at room temperature, variable temperature NMR experiments were performed [10]. In this paper, their separation and structure elucidation are described.

Fig. 1
figure 1

Structures of compounds 1 and 2 from C. lansium

2 Results and Discussion

Clausenlanin A (1) was obtained as an amorphous solid. Its molecular formula was shown as C50H91N9O9 by its negative HRESIMS ([M−H], 960.6876, calcd 960.6867), indicating the 10° of unsaturation. The IR spectrum exhibited the absorption bands at 3429 and 1661 cm–1 ascribable to NH and CO groups. The 1H and 13C NMR spectra of 1 in C5D5N (Table 1) displayed the characteristic signals of typical CPs.

Table 1 NMR data of compounds 1 and 2

The 1H NMR signals of the amino acid residues of 1, especially the signals of NH and α-H, were severely overlapped taken at room temperature. The significant improvement of the 1H NMR signals was observed by increasing the temperatures from 243 to 338 K. Finally a well-resolved 1H NMR spectrum with sharp proton signals (Fig. 2a) was obtained at 338 K in pyridine-d5. Then the assignment of the 1H NMR signals of the amino acid residues was obtained by analyzing the 1H-1H COSY spectrum, particularly amide proton NH and α-H signals. The corresponding 13C NMR assignments were determined on the basis of the HSQC and HMBC experiments, particularly α-C signals (Table 1). The 1H-NMR spectrum of 1 showed the presence of nine NH (δ H 8.91, 8.75, 8.73, 8.72, 8.52, 8.46, 8.36, 8.23, 8.10) and ten α-H (δ H 4.81, 4.77, 4.68, 4.59, 4.58, 4.57, 4.51, 4.46, 4.46, 3.85), respectively. The 13C-NMR spectrum of 1 displayed nine carbonyl CO signals at δ C 175.6, 174.7, 174.5, 174.4, 174.1, 173.9, 173.9, 173.5, 171.1, eight α-CH signals at δ C 61.2, 54.9, 54.8, 54.7, 54.5, 54.0, 53.7, 53.6, one α-CH2 signal at δ C 44.8. These data indicated that 1 might be a cyclic nonapeptide. Analysis of the HSQC, HMBC and COSY spectra revealed that 1 consisted of one glycine (δ H 4.51 and 3.85 (α-H2), 8.75 (NH); δ C 171.1 (CO), 44.8 (α-CH2)), and one isoleucine (δ H 4.46 (α-CH), 2.30 (β-CH), 1.39 and 1.87 (γ-CH2), 1.18 (γ-CH3), 0.92 (δ-CH3), 8.72 (NH); δ C 173.9 (CO), 61.2 (α-CH), 36.7 (β-CH), 26.7 (γ-CH2), 16.6 (γ-CH3), 11.5 (δ-CH3)). The remaining signals mentioned-above of seven NH and seven α-H signals, seven CO and seven α-C signals, and other signals including seven methylenes at δ C 40.0–40.6, seven methines at δ C 25.5–26.0, two kinds of fourteen methyls at δ C 23.5–23.8 and at δ C 22.1–22.6, indicated that 1 contained other seven leucines. Therefore 1 consisted of seven leucines, one glycine, and one isoleucine (Table 1; Fig. 3).

Fig. 2
figure 2

1H NMR of compounds 1 (a) and 2 (b) at different temperatures

Fig. 3
figure 3

Key HMBC, 1H-1H COSY, and ROESY/NOESY correlations of 1 and 2

The sequence of the nine amino acid residues in 1 was determined by analyzing the ROESY correlations between the α-H of one amino acid residue and the amide proton NH of the next amino acid residue (Fig. 3). The ROESY correlations of Gly1-αH/Leu2-NH, Leu2-αH/Ile3-NH, Ile3-αH/Leu4-NH, Leu4-αH/Leu5-NH, Leu5-αH/Leu6-NH, Leu6-αH/Leu7-NH, Leu7-αH/Lue8-NH, Leu8-αH/Leu9-NH, Leu9-αH/Gly1-NH indicated that the structure of 1 is cyclo-(Gly1-Leu2-Ile3-Leu4-Leu5-Leu6-Leu7-Leu8-Leu9). This sequence of 1 was confirmed by the fragment ion peaks at 962.96 [M+H]+, 849.83 [M+H−113]+, 736.72 [M+H−2*113]+, 623.64 [M+H−3*113]+, 510.52 [M+H− 4*113]+, 453.49 [M+H−4*113−57]+, 340.43 [M+H−4*113–57−113]+, 227.29 [M+H−4*113–57−2*113]+ in the positive ESIMSMS.

The absolute configuration of the amino acids of 1 was determined using the advanced Marfey’s method and LC–MS analysis [11, 12]. The results indicated that the absolute configurations of the amino acid residues (Leu and Ile) in 1 were the l-configuration (Table S1; Fig. 3). Therefore the structure of 1 is determined as cyclo-(Gly1-l-Leu2-l-Ile3-l-Leu4-l-Leu5-l-Leu6-l-Leu7-l-Leu8-l-Leu9).

Clausenlanin B (2) was obtained as an amorphous solid. Its molecular formula was shown as C49H89N9O9 by its negative HRESIMS ([M−H], 946.6723, calcd 946.6710), indicating the 10° of unsaturation. The IR spectrum exhibited the absorption bands at 3430 and 1661 cm–1 ascribable to NH and CO groups. The 1H and 13C NMR spectra of 2 in C5D5N (Table 1) displayed the characteristic signals of typical CPs.

The 1H NMR signals of the amino acid residues of 2, especially the signals of NH and α-H, were severely overlapped taken at room temperature. The significant improvement of the 1H NMR signals was observed by increasing the temperatures from 243 to 338 K. Finally a well-resolved 1H NMR spectrum with sharp proton signals (Fig. 2b) was obtained at 303 K in pyridine-d5. After compared all data of 2 with those of 1, the results indicated that 2 and 1 are very similar, and 2 might also be a cyclic nonapeptide too. The only difference is to be replaced the isoleucine residue in 1 by valine residue in 2. The assignment of the 1H and 13C NMR signals of the valine residue was obtained by analyzing the HSQC, HMBC and COSY spectra, i.e. δ H 4.42 (α-CH), 2.57 (β-CH), 1.18 and 1.19 (2*γ-CH3), 9.22 (NH); δ C 173.9 (CO), 62.6 (α-CH), 30.6 (β-CH), 20.3 and 20.2 (2*γ-CH3). Therefore 2 consisted of seven leucines, one glycine, and one valine (Table 1; Fig. 3).

The sequence of the nine amino acid residues in 2 was determined by analyzing NOESY correlations between the α-H of one amino acid residue and the amide proton NH of the next amino acid residue (Fig. 3). The NOESY correlations of Gly1-αH/Leu2-NH, Leu2-αH/Val3-NH, Val3-αH/Leu4-NH, Leu4-αH/Leu5-NH, Leu5-αH/Leu6-NH, Leu6-αH/Leu7-NH, Leu7-αH/Lue8-NH, Leu8-αH/Leu9-NH, Leu9-αH/Gly1-NH indicated that the structure of 2 is cyclo-(Gly1-Leu2-Val3-Leu4-Leu5-Leu6-Leu7-Leu8-Leu9). This sequence of 2 was confirmed by the fragment ion peaks at 948.86 [M+H]+, 835.70 [M+H−113]+, 722.56 [M+H−2*113]+, 609.54 [M+H−3*113]+, 496.39 [M+H−4*113]+, 383.41 [M+H−5*113]+, 270.24 [M+H−6*113]+ in the positive ESIMSMS.

The absolute configuration of 2 was determined using the advanced Marfey’s method and LC–MS analysis too [11, 12]. The results indicated that the absolute configurations of the amino acid residues (Leu and Val) in 2 were the l-configuration (Table S1; Fig. 3). Therefore the structure of 2 is determined as cyclo-(Gly1-l-Leu2-l-Val3-l-Leu4-l-Leu5-l-Leu6-l-Leu7-l-Leu8-l-Leu9).

3 Experimental

3.1 General Experimental Procedures

Optical rotations were obtained on a Jasco P-1020 polarimeter. IR spectra were measured on a Tensor 27 spectrometer with KBr pellets. UV spectra were obtained using a Shimadzu UV-2401PC spectrophotometer. 1D and 2D NMR spectra were performed on a Bruker AM-400 (1H: 400 MHz, 13C: 100 MHz) or Bruker AVANCE III-800 (1H: 800 MHz, 13C: 200 MHz). Chemical shifts were expressed in ppm with reference to the solvent signals. Mass spectra were measured on a Waters XEVO-TQD spectrometer or an Agilent 1290 UPLC/6540 Q-TOF spectrometer. Analytical or semi-preparative HPLC was performed on Agilent 1100 apparatus equipped with a UV detector and a SunFire OBD (Waters, 1.9 × 25 cm, 5 μm). Column chromatography was performed with silica gel (100–200 mesh and 200–300 mesh, Qingdao Yu-Min-Yuan Chemical Co. Ltd., Qingdao, P.R. China), MCI gel (CHP-20P, 70–150 μm, Mitsubishi Chemical Co., Japan) or Lichroprep RP-18 gel (40–63 mm, Merck, Darmstadt, Germany). Fractions were monitored by TLC (GF254, Qingdao Yu-Min-Yuan Chemical Co. Ltd., Qingdao, P.R. China), and the orange spots were visualized on the plate by spraying with 2% ninhydrin reagent, after hydrolyzed in an incubator (110 °C) for 30 min by concentrated HCl [13].

3.2 Plant Material

The roots and rhizomes of C. lansium were collected in Hekou, Yunnan Province, P. R. China, in September 2010, and identified by Prof. Yu-Min Shui, Kunming Institute of Botany, Chinese Academy of Sciences. A voucher specimen (No. 0599043) was deposited at the Herbarium of Kunming Institute of Botany, Chinese Academy of Sciences.

3.3 Extraction and Isolation

Air dried, powdered roots and rhizomes of C. lansium (27 kg) were extracted and refluxed with MeOH for three times each for 4 h (MeOH, 3*50 L). The extract was evaporated under reduced pressure to yield a dark brown residue (0.9 kg). The residue was suspended in MeOH/H2O (7:3, 3 L) and then partitioned with EtOAc (3*2 L). After removing solvent, the EtOAc-soluble part (406 g) was fractionated by silica gel (200–300 mesh) column chromatograph (CC) and eluted with CHCl3/MeOH (30:1–4:1) to afford six fractions (Fr.1–Fr.6), on the basis of TLC detection.

Fr.6 (77 g) was subjected to silica gel CC (CHCl3/acetone 15:1–7:3) to afford Fr.6.1–Fr.6.4. Fr.6.2 (14.3 g) was subjected to silica gel CC (PE/acetone 5:1), MPLC with MCI (MeOH/H2O 10:90–60:40), and MPLC with RP-18 (MeOH/H2O 5:95–70:30). Then, the fractions was further purified by silica gel CC (CHCl3/MeOH 50:1), subsequently to afford 1 (308 mg) and 2 (47 mg).

3.4 Clausenlanin A (1)

Amorphous powder; [α] 20.7D –154.6 (c 0.10, MeOH); UV (MeOH) λmax (log ε) 203.0 (4.61) nm; CD (MeOH) 203 (Δε–30.8); IR (KBr) νmax 3429, 2960, 1661, 1528, 584 cm–1; 1H (400 MHz) and 13C (100 MHz) NMR data, see Table 1; positive ESIMSMS m/z 962.96 [M+H]+, 849.83 [M+H−113]+, 736.72 [M+H−2*113]+, 623.64 [M+H−3*113]+, 510.52 [M+H−4*113]+, 453.49 [M+H−4*113−57]+, 340.43 [M+H−4*113−57−113]+, 227.29 [M+H−4*113−57−2*113]+; negative HRESIMS m/z 960.6876 [M−H], calcd for C50H91N9O9, 960.6867.

3.5 Clausenlanin B (2)

Amorphous powder; [α] 20.6D –73.69 (c 0.10, MeOH); UV (MeOH) λmax (log ε) 202.8 (4.40) nm; CD (MeOH) 203 (Δε–26.1); IR (KBr) νmax 3430, 2960, 1661, 1527, 584 cm–1; 1H (800 MHz) and 13C (200 MHz) NMR data, see Table 1; positive ESIMSMS m/z 948.86 [M+H]+, 835.70 [M+H−113]+, 722.56 [M+H−2*113]+, 609.54 [M+H−3*113]+, 496.39 [M+H−4*113]+, 383.41 [M+H−5*113]+, 270.24 [M+H−6*113]+; negative HRESIMS m/z 946.6723 [M−H], calcd for C49H89N9O9, 946.6710.

3.6 Advanced Marfey’s Method [11, 12]

The cyclic peptide (about 1.0 mg each) was dissolved in 6 N HCl (1 mL) and heated at 110 °C for 24 h. The hydrolyzate was evaporated to dryness, and the residue was re-dissolved in 100 μL of acetone. To each a half portion (50 μL) were added 20 μL of NaHCO3 (1 M) and 100 μL of Nα-(5-Fluoro-2,4-dinitrophenyl)-l-leucinamide (l-FDLA, 1% in acetone) or 50 μL of Nα-(5-Fluoro-2,4-dinitrophenyl)-l-leucinamide and 50 μL of Nα-(5-Fluoro-2,4-dinitrophenyl)-d-leucinamide (mixture of l-FDLA and d-FDLA, 1% in acetone), and the mixture was heated at 45 °C for 1.5 h. Reaction was cooled to room temperature, and then acidified with 2 N HCl (10 μL), dried and dissolved in 50% aqueous MeCN. 5 μL of each solution of FDLA derivatives were analyzed by LC/MS.

The analysis of the l- and d, l-FDLA (mixture of d- and l-FDLA) derivatives was performed using an Waters Sunfire C18 column (4.6*150 mm, 5 μm) maintained at 30 °C. Acetonitrile—0.1% HCOOH/H2O was used as the mobile phase under a linear gradient elution mode (acetonitrile, 28–60%, 50 min (compound 1); acetonitrile, 35–60%, 50 min (compound 2)) at a flow rate of 1 mL/min. A Waters Xevo-TQD mass spectrometer was used for detection in ESI mode. The capillary voltage was kept at 2.5 kV, and the ion source at 450 °C. Nitrogen gas was used as a sheath gas at 650 L/h. A mass range of m/z 100–2000 was scanned in 0.2 s.