Dianthiamides A–E, Proline-Containing Orbitides from Dianthus chinensis

Orbitides are plant-derived small cyclic peptides with a wide range of biological activities. Phytochemical investigation of the whole plants of Dianthus chinensis was performed with the aim to discover new bioactive orbitides. Five undescribed proline-containing orbitides, dianthiamides A–E (1–5), were isolated from a methanolic extract of Dianthus chinensis. Their structures were elucidated by extensive analysis of 1D and 2D NMR and HRESI–TOF–MS as well as ESI–MS/MS fragmentation data. The absolute configuration of the amino acid residues of compounds 1–5 was determined by Marfey’s method. All compounds were tested for their cytotoxic activity, and dianthiamide A (1) exhibited weak activity against A549 cell line with IC50 value of 47.9 μM.

In the HMBC and ROESY experiments, the cyclic feature and amino acid sequence 1 were elucidated by the correlations observed between the amino acid Hα and contin ous amide group (CONH). Therefore, the linear sequence of 1 was identified as G-F-L-P Pb-I-N. Also, the HMBC correlation from Gly-Hα (δH 3.40 and 3.55) to Asn-C=O (δC 172. as well as the ROESY correlation between Gly-NH (δH 8.70) and Asn-Hα (δH 4.18) esta lished the cyclic heptapeptide as cyclo-G-F-L-Pa-Pb-I-N ( Figure 2). The amino acid s quence of 1 was further confirmed by analysis of the ESI-MS/MS fragment ions. Presum

Results and Discussion
Dianthiamide A (1) was obtained as a yellow amorphous powder. Its molecular formula of C 37 Table 1), suggesting that 1 is a heptapeptide. Furthermore, HSQC, HMBC, and COSY spectra showed the identification of seven amino acid residues including phenylalanine (F), glycine (G), isoleucine (I), asparagine (N), leucine (L), and two prolines (P a and P b ).
In the HMBC and ROESY experiments, the cyclic feature and amino acid sequence of 1 were elucidated by the correlations observed between the amino acid Hα and continuous amide group (CONH). Therefore, the linear sequence of 1 was identified as G-F-L-P a -P b -I-N. Also, the HMBC correlation from Gly-Hα (δ H 3.40 and 3.55) to Asn-C=O (δ C 172.3) as well as the ROESY correlation between Gly-NH (δ H 8.70) and Asn-Hα (δ H 4.18) established the cyclic heptapeptide as cyclo-G-F-L-P a -P b -I-N ( Figure 2). The amino acid sequence of 1 was further confirmed by analysis of the ESI-MS/MS fragment ions. Presumably, though there were several ring-opening sites, it occurred at two preferred positions at Pro a 4 -Pro b 5 and Ile 6 -Asn 7 , respectively. Each the linear sequences, Pro b 5 -Ile 6 -Asn 7 -Gly 1 -Phe 2 -Leu 3 -Pro a 4 (b 7 P b P a ) and Asn 7 -Gly 1 -Phe 2 -Leu 3 -Pro a 4 -Pro b 5 -Ile 6 (b 7 NI) was certified by acylium ions (b n P b P a and b n NI) and after loss of CO (a n NI) at m/z 642 (b 6 P b P a ), 626 (b 6 NI), 529 (b 5 P b P a and b 5 NI), 404 (a 4 NI), 364 (b 4 P b P a -H 2 O), 319 (b 3 NI), and 211 (b 2 P b P a ), corresponding to the successive loss of amino acid residues ( Figure S9). The geometry of proline residues was assigned on the basis of the ∆δ Cβ-Cγ values and the presence of a ROE correlations between the proline Hα or Hδ and the Hα of previous amino acid. The ∆δ Cβ-Cγ value (3.2 ppm) of the Pro a 4 and the ROE correlation between the Hα (δ H 4.48) of Leu 3 and the Hδ (δ H 3.44) of Pro a 4 indicated that the amide bond in the Pro a 4 was a trans. However, the ∆δ Cβ-Cγ value (9.0 ppm) of the Pro b 5 and the ROE correlation between the Hα (δ H 4.45) of Pro a 4 and the Hα (δ H 4.55) of Pro b 5 indicated that the geometry of Pro b 5 was a cis (Figure 3) [21][22][23]. The absolute configuration of amino acid residues in 1 were identified as L configuration, which was deduced by acid hydrolysis and Marfey's derivatization, followed by HPLC analysis [24][25][26][27]. The N-α- ( Proa 4 was a trans. However, the ΔδCβ-Cγ value (9.0 ppm) of the Prob 5 and the ROE correlation between the Hα (δH 4.45) of Proa 4 and the Hα (δH 4.55) of Prob 5 indicated that the geometry of Prob 5 was a cis (Figure 3) [21][22][23]. The absolute configuration of amino acid residues in 1 were identified as L configuration, which was deduced by acid hydrolysis and Marfey's derivatization, followed by HPLC analysis [24][25][26][27]. The N-α- (2,4-  In the HMBC and ROESY experiments, the cyclic feature and amino acid sequence of 1 were elucidated by the correlations observed between the amino acid Hα and continuous amide group (CONH). Therefore, the linear sequence of 1 was identified as G-F-L-Pa-Pb-I-N. Also, the HMBC correlation from Gly-Hα (δH 3.40 and 3.55) to Asn-C=O (δC 172.3) as well as the ROESY correlation between Gly-NH (δH 8.70) and Asn-Hα (δH 4.18) estab-  Figure S9). The geometry of proline residues was assigned on the basis of the ΔδCβ-Cγ values and the presence of a ROE correlations between the proline Hα or Hδ and the Hα of previous amino acid. The ΔδCβ-Cγ value (3.2 ppm) of the Proa 4 and the ROE correlation between the Hα (δH 4.48) of Leu 3 and the Hδ (δH 3.44) of Proa 4 indicated that the amide bond in the Proa 4 was a trans. However, the ΔδCβ-Cγ value (9.0 ppm) of the Prob 5 and the ROE correlation between the Hα (δH 4.45) of Proa 4 and the Hα (δH 4.55) of Prob 5 indicated that the geometry of Prob 5 was a cis (Figure 3) [21][22][23]. The absolute configuration of amino acid residues in 1 were identified as L configuration, which was deduced by acid hydrolysis and Marfey's derivatization, followed by HPLC analysis [24][25][26][27]. The N-α- (2,4-   Dianthiamide B (2) was isolated as a yellow amorphous powder, the HRESI-TOF-MS data were consistent with the molecular formula C 32 H 44 N 6 O 7 (m/z 647.3161 [M + Na] + ; calcd for 647.3163). The 1 H, 13 C and HSQC NMR spectra of 2 showed 32 carbons, consisting of seven amide carbonyl carbons, six α-amino acid carbons, six aromatic carbons, two methines, seven methylenes, and four methyls (Table 2). HSQC, HMBC, and COSY spectra demonstrated the presence of six amino acid residues including aspartic acid (D), phenylalanine (F), glycine (G), isoleucine (I), leucine (L), and proline (P). Moreover, HMBC, COSY, and ROESY spectra indicated that the sequence and connectivity of the hexapeptide was cyclo-G-L-P-F-D-I (Figure 2). The HMBC correlations between Hα (δ H 4.19) of Ile 6 and two carbonyls (δ C 176.2 and 177.2) of Asp 5 showed that dehydration of NH-Ile 6 and COOH-Asp 5 formed an additional five membered ring system (pyrrolidine-2,5-dione). The amino acid sequence in 2 was further supported by the fragmentation pattern of ESI-MS/MS data, in which the preferred ring-opening of 2 occurred at the amide bond between leucine and proline ( Figure S19). The geometry of amide bond of Pro 3 residue in 2 was assigned the trans configuration, on the basis of the difference of the 13 C NMR chemical shift (∆δ Cβ-Cγ = 3.9 ppm) [21][22][23] as well as the ROE correlation between the Hα (δ H 4.51) of Leu 2 and the Hδ (δ H 3.68 and 3.42) of Pro 3 residue. In addition, the absolute configuration of six amino acid residues in 2 were all assigned as L, which was determined by HPLC analysis of the acid hydrolysate after Marfey's derivatization ( Figure S20). Therefore, dianthiamide B (2) was determined as cyclo-(Gly 1 -L-Leu 2 -L-trans-Pro 3 -L-Phe 4 -L-Asp 5 -L-Ile 6 ).  (Table 3). Full assignments of 1 H and 13 C NMR data for each amino acid residue were accomplished by combined analysis of COSY, HSQC, and HMBC spectra and suggested that 3 was composed of seven amino acid such as phenylalanine (F), glycine (G), isoleucine (I), two leucines (L a and L b ), proline (P), and serine (S) residue ( Table 3). The HMBC and ROESY spectra indicated that the amino acid sequence was cyclo-G-L a -S-P-F-I-L b (Figure 2), which was further confirmed by ESI-MS/MS fragmentation analysis ( Figure S29). The observed ∆δ Cβ-Cγ value (3.2 ppm) of the Pro 4 and the ROE correlation from the Hα (δ H 4.89) of Ser 3 to the Hδ (δ H 3.91 and 3.42) of Pro 4 revealed that the geometry of Pro 4 of 3 was a trans configuration [21][22][23]. The absolute configuration of 3 was determined by Marfey's method [24][25][26][27], which indicated that all the amino acids were L configuration ( Figure S30). Therefore, dianthiamide C (3) was confirmed as cyclo-(Gly 1 -L-Leu 2 -L-Ser 3 -L-trans-Pro 4 -L-Phe 5 -L-Ile 6 -L-Leu 7 ). . Detailed analyses of the 1D and 2D (COSY, HSQC, and HMBC) NMR data revealed that 4 was a octapeptide containing alanine (A), glycine (G), isoleucine (I), two prolines (P a and P b ) and three valines (V a , V b , and V c ) residues ( Table 4). The amino acid sequence of 4 was established as cyclo-G-A-V a -I-P a -V b -V c -P b by analysis of HMBC and ROESY data ( Figure 2). This conclusion was also supported by the ESI-MS/MS sequence analysis ( Figure S39) [21][22][23]. Moreover, the absolute configuration of 4 was assigned by Marfey's method [24][25][26][27], which indicated that all the amino acids had L configuration ( Figure S40). Therefore, dianthiamide D (4) Figure 2) as well as ESI-MS/MS sequence data ( Figure S49) demonstrated that the seven amino acid residues were phenylalanine (F), glycine (G), isoleucine (I), asparagine (N), leucine (L), proline (P), and threonine (T). The sequence of these amino acid residues was assigned as cyclo-G-N-P-L-T-I-F by the observed HMBC and ROESY data (Figure 2).
Recently, it has been reported that cyclic peptides isolated from the genus Dianthus exhibited cytotoxic activity against several cancer cell lines [5,6,18]. Therefore, all isolates were tested for their cytotoxic activity against human non-small cell lung cancer A549 and human stomach adenocarcinoma MKN-28 cells, with docetaxel as a positive control. However, dianthiamide A (1) only showed weak activity against A549 cell line with IC 50 value of 47.9 µM, and docetaxel was used as a positive control (IC 50 : 0.08 µM). The other compounds 2-5 were inactive against A549 and MKN-28 cells (IC 50 : >200 µM).

Plant Material
The dried whole plants of Dianthus chinensis L. (Caryophyllaceae) were purchased from Kyungdong herbal market in Seoul, Korea, in June 2014. A voucher specimen (CBNU-2014-06-DC) was authenticated by B.Y.H. and deposited at the Herbarium of the College of Pharmacy, Chungbuk National University, Korea.