Results and Discussion
Preparation of (S)-Decursinol (1)For the semi-synthesis of new decursin derivatives, natural pure (S)-decursinol (1) as a main scaffold was prepared from the root extract of Angelica gigas Nakai (purchased at Gyungdong market, Seoul, Republic of Korea) followed by its alkaline hydrolysis.11) The dried root of Angelica gigas Nakai (1 kg) was extracted with EtOH. The concentrated extract was partitioned between water and ether. The ether layer was hydrolyzed with 10% NaOH to afford pure (S)-decursinol (1, 31.5 g, 3.15%), whose spectral data was identical to the previously reported data11) and its optical purity (>93% ee) was determined by the chiral HPLC method (Supplementary Fig. S1).
Synthesis of Both Enantiomeric (S)-Decursin and (R)-Decursin DerivativesAs shown in Chart 1, (S)-decursinol (1) was coupled with various acid by 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC)-mediated reaction to afford a series of (S)-decursin derivatives. The commercially unavailable carboxylic acids (7, 9, and 11) were synthesized according to the reported methods: 3-(E)-(Pyridin-2-yl)acrylic acid (7) was easily prepared in 21% yield from the reaction of pyridine-2-carboxaldehyde (6) with malonic acid catalyzed by piperidine under the reflux condition of pyridine solvent using Knoevenagel condensation.12) 3-(1H-Imidazol-1-yl)propanoic acid (9) was obtained in 81% yield using conjugate addition of imidazole to acrylic acid under the reflux condition of N,N-dimethylformamide (DMF) solvent. (E)-4-(N, N-Dimethylsulfamoyl)cinnamic acid (11) was synthesized in 2 steps (60% overall yield) using chlorosulfonylation of cinnamic acid (10) followed by the amination. The esterification of (S)-decursinol with each carboxylic acid (total 9 compounds) by EDC-mediated coupling provided the corresponding (S)-decursin derivatives (S)-2a–i in 39–85% yields.13)
Chart 1. Reagents and Conditions
(a) R-CO2H, EDC-HCl, DMAP, CH2Cl2, 0 °C to room temperature (r.t.), 18–85%; (b) PPh3, CCl4, CH3CN, 60 °C, 84%; (c) Jacobsen’s (R, R)-salen-Mn (III) catalyst, 15% NaOCl/0.05 M Na2HPO4/CH2Cl2, pH 11.3, 0 °C, 39%, 96% ee; (d) BF3·OEt2, NaCNBH3, THF, 0 °C, 99%; (e) R-CO2H, EDC-HCl, 4-dimethylaminopyridine (DMAP), CH2Cl2, 0 °C to r.t., 18–62%; (e) Jacobsen’s (S, S)-salen-Mn (III) catalyst, 15% NaOCl/0.05 M Na2HPO4/CH2Cl2, pH 11.3, 0 °C, 50%, 95% ee; (f) malonic acid, piperidine, pyridine, reflux, 21%; (g) acrylic acid, DMF, reflux, 83%; (h) ClSO3H, 0 °C to r.t., 79%; (i) 40% dimethylamine in H2O, 0 °C to r.t., 76%.
In other to investigate the stereochemistry-activity relationship at chiral center of decursin derivatives, we decided to synthesize unnatural (R)-decursinol (5, ent-1) by featuring three steps: dehydration of (S)-decursinol (1), asymmetric epoxidation reaction using Jacobsen’s salen-Mn (III) catalyst,14) and subsequent nucleophilic ring opening of epoxide. That is, (S)-decursinol (1) was dehydrated using Appel reaction (Ph3P and CCl4) in refluxing CH3CN15) to afford xanthyletin (3) in 84% yield for asymmetric epoxidation. We have used Jacobsen’s (R,R)-salen-Mn (III) catalyst in order to prepare chiral α-xanthyletin oxide 4 as a major epoxide using chiral discrimination strategy by reagent.14) Xanthyletin (3) was converted to the desired chiral α-xanthyletin oxide 4 in 39% yield in the presence of Jacobsen’s (R,R)-salen-Mn (III) catalyst with 15% NaOCl according to the reported procedure.16) The absolute configuration of α-xanthyletin oxide 4 was confirmed based on the previously reported results17,18) and its enantiometic excess (96% ee) was determined using chiral HPLC (Supplementary Figs. S2, S3). The regio- and stereoselective reduction of α-epoxide 4 by using NaBH3CN with BF3·OEt2 in tetrahydrofuran (THF) at 0 °C gave (R)-decursinol (5, ent-1) in 99% yield.17) The esterification of (R)-decursinol (5) with each carboxylic acid gave the corresponding (R)-decursin derivatives (R)-2a-i in 18–62% yields by the same EDC-mediated coupling. Xanthyletin (3) was again converted to the desired chiral β-xanthyletin oxide ent-4 in 50% yield (95% ee, Supplementary Figs. S4, S5) under the same condition using enantiomeric Jacobsen’s (S,S)-salen-Mn (III) catalyst. This β-xanthyletin oxide (ent-4) was again treated with NaBH3CN with BF3·OEt2 in THF at 0 °C gave a natural (S)-decursinol (1) in quantitative yield, which could be also used for the preparation of (S)-decursin derivatives. The structures of both enantiomeric (S)-decursin and (R)-decursin derivatives were summarized in Table 1 with their isolated yields.
Table 1. Structures and Biological Activities of Both Enantiomeric Decursin Derivative Pairs
a) Isolated yield; b) Value is average of three independent experiments; c) Determined using chiral HPLC (Supplementary Fig. S7); d) Determined using chiral HPLC (supplementary Fig. S10).
On the hand other, the direct stereoinversion of (S)-decursinol (1) into (R)-decursinol (5, ent-1) was failed via the typical Mitsunobu reaction (AcOH, Ph3P, and DEAD condition)19) and this failure might be due to the steric hindrance of geminal dimethyl groups at C-2′ position of (S)-decursinol (1).
Antiproliferative ActivityThe cytotoxicity of synthesized decursin derivatives was evaluated against human lung adenocarcinoma A549 cell line via 3-(4,5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) assay.20) Five doses of each compound (12.5, 25, 51, 100, and 200 µM) were evaluated in order to obtain its dose-dependent cytotoxic effect. The IC50 values of all synthetic derivatives were inserted in Table 1. At a glance, both enantiomeric decursin derivatives exhibited a broad spectrum of cytotoxicity against A549 cancer cell line. Among them, compound (S)-2d (MRC-D-004, Supplementary Figs. S6–S8) showed the most potent activity (IC50: 14.03 µM) against A549 cancer cells followed by a sequence of (S)-2g (IC50: 80.66 µM) and (S)-2f (IC50: 113.14 µM) as compared with the reference, decursin as a natural compound (IC50: 43.55 µM). The brief structure–activity relationship of all decursin derivatives was analyzed. The following structural features can be related to potent cytotoxic activity: (i) a double bond between ester group and aryl ring was better linker than a single bond (that is, acryloyl moiety > propionyl moiety: 2a–c series > 2d–i series), (ii) (S)-decursin derivatives generally exhibited more potency than (R)-derivatives and (iii) a smaller aromatic ring size was more active than aromatic ring size (2d, 2f–2g series vs. 2e, 2h–2i series). Considering these structural features, therefore, compound (S)-2d with (E)-(furan-3-yl)acryloyl group exhibited the most potency (IC50: 14.03 µM), which was 10-fold more potent than its enantiomer (R)-2d (IC50: 151.59 µM, Supplementary Figs. S9, S10) against A549 cancer cells. In order to further confirm the key feature of a double bond between ester group and aryl ring, compound (S)-2j with 3-(furan-3-yl)propionyl group was newly prepared, exhibiting significant decrease of cytotoxicity effect on A549 cancer cells (IC50: 114.15 µM) as compared with (S)-2d (IC50: 14.03 µM) as shown in Table 1.
Western Blotting AnalysisNext, we investigated the molecular mechanism for cytotoxic effect of (S)-2d (MRC-D-004) on A549 cancer cells. A549 cells were cultured with various concentrations of (S)-2d (12.5, 25, 50 µM). DMSO and decursin at the same concentrations were served as a vehicle and a positive control, respectively. After 24 h, Western blotting of A549 cell lysate was performed to identify specific proteins associated with the mechanism for cytotoxic effect of (S)-2d. It was recently reported that decursin inhibited tumor progression in gastric cancer and head/neck squamous cell carcinoma by down-regulating CXC chemokine receptor 7 (CXCR7)/ atypical chemokine receptor 3 (ACKR3) expression.21,22) In addition, the expression level of CXCR7/ACKR3 in A549 cancer cells was relatively high compared to the normal cells.23) Based on these reported literatures, therefore, we firstly decided to investigate whether decursin or (S)-2d could suppress CXCR7/ACKR3 overexpression in A549 cancer cells. As a result, both compounds could not inhibit CXCR7/ACKR3 overexpression in A549 lung cancer cells regardless of their doses as shown in Fig. 2A, while decursin suppresses CXCR7/ACKR3 overexpression in gastric cancer and head/neck squamous cell carcinoma.
On the other hand, Haura’s group reported that Janus kinase 1 (JAK1) signaling is responsible for signal transducer and activator of transcription activation 3 (STAT3) in lung cancer cells and loss of STAT3 activity via direct JAK1 inhibition or indirect inhibition by eliminating IL-6 signaling resulted in reduced cell proliferation and apoptosis of lung cancer cells.24) Out Western blotting analysis also indicated 60% decrease in p-JAK1/JAK1 ratio (Figs. 2B, C) and >95% decrease in p-STAT3/STAT3 ratio (Figs. 2B, D) without significant change in JAK1 and STAT3 levels upon A549 cancer cells with treatment of (S)-2d at 50 µM concentration compare with the vehicle control. Furthermore, (S)-2d inhibit more strongly the phosphorylation of JAK1 and STAT3 than decursin at the same concentration (Figs. 2C, D), while it had no effect in both expression and phosphorylation of JAK2. Therefore, this Western blotting analysis implicates the plausible mechanism that (S)-2d (MRC-D-004) exhibited significant cytotoxicity through inhibition of JAK1/STAT3 signaling pathway in A549 lung cancer cells.
Cell Cycle Arrest and ApoptosisIt was recently reported that inhibition of STAT3 signaling induces cell cycle arrest and apoptosis of lung cancer.25) The cell cycle arrest effect of decursin or (S)-2d (MRC-D-004) on A549 cells was further investigated by employing Western blotting assays. Along with activation of p27Kip1 which is a member of the universal cyclin-dependent kinase inhibitor (CDKI) family that arrest the cell cycle at the G1 phase, CDK2, cyclin D1, cyclin D3 and p-cdc2 were markedly decreased by (S)-2d compared with decursin (Fig. 3A). In addition, the cleavages of caspase-8, -9 and PARP were more significantly observed by (S)-2d than decursin in A549 cells (Fig. 3B). Collectively, these overall results suggest that (S)-2d (MRC-D-004) inhibits the proliferation, induces cell cycle arrest and promotes apoptotic cell death in A549 lung cancer in vitro via JAK1/STAT3 axis.
Molecular Docking AnalysisMolecular docking simulation was conducted to rationalize the structure–activity relationship of selected compounds such as (S)-2d (MRC-D-004), (R)-2d, (S)-2j, and decursin and their plausible mode of action for anti-proliferative activity against A549 cancer cells. For this goal, 3D crystal structures of JAK1 (PDB code: 3EYG)26) and STAT3 (PDB code: 6NJS)27) complexes were retrieved from Protein Data Bank. In this study, molecular docking calculations were then performed using the Maestro Version 12.5, Release 2020-3, Platform Windows-x64.28) The structure of JAK1 with ligand (3EYG) was downloaded and refined. Docking studies were carried out using Extra Precision (XP) docking module of Glide. The best pose of each compound with the highest Docking Score was selected and summarized in Table 2 together with amino acid residues forming interaction with each pose, respectively. Among them, the top ranked compound was (S)-2d which showed a high docking score of −7.037 by forming interaction with Leu959, His885 and Arg1007 (Figs. 4A, C). The carbonyl oxygen atom of pyranocoumarin ring was formed hydrogen bond with Leu959 residue at a distance 2.1 Å in the active site of JAK1. The furan ring stabilized the binding pose by forming a Pi–Pi stacking interaction with side chain of His885 and Pi–cation interaction with Arg1007 (Figs. 4A, C). (R)-2d and decursin formed same hydrogen bonds with Leu959 and Glu883, but (R)-2d exhibited the lowest docking score (−5.544). The reason for the relatively low docking score of (R)-2d could be explained by the furan ring being located outside the binding pocket, leading to a lack of stabilization by His885 and Arg1007 (Fig. 4A). (S)-2j exhibits a distinct docking pose compared to the other compounds (S)-2d, (R)-2d and decursin. The docking analysis revealed that (S)-2j demonstrated a distinct docking pose in comparison to the other compounds, (S)-2d, (R)-2d, and decursin (Fig. 4B). While (S)-2j maintained similar hydrogen bond interactions, pi-pi stacking, and pi-cation interactions as observed in (S)-2d, a notable difference was observed in the orientation of the pyranocoumarin ring. In the case of (S)-2j, the pyranocoumarin ring adopted an opposite orientation, leading to a clash with Glu957. Unlike (S)-2d with a double bond, furthermore, the sp3 chain of (S)-2j occupied the inside of the binding pocket. This positioning resulted in unfavorable interactions with Asp1021 and Asn1008 (Fig. 4B). These findings highlighted the unique structural features of (S)-2j in its binding to JAK1, suggesting potential implications for its lower biological activity than (S)-2d.
Table 2. Docking Analysis of Selected Compounds in JAK1 (PDB Code: 3EYG) and STAT3 (PDB Code: 6NJS)
| Compound | Docking scorea) | Interaction in the active site of JAK1 kinase domain |
|---|
| JAK1 | STAT3 | Good interaction residueb) | Contacted or Crushed residuec) |
|---|
| (S)-2d (MRC-D-004) | −7.037 | −2.087 | Leu959, His885, Arg1007 | — |
| (R)-2d | −5.544 | −2.187 | Leu959, Glu883 | — |
| (S)-2j | −6.916 | −0.943 | Leu959, His885, Arg1007 | — |
| Decursin | −5.756 | −2.888 | Leu959, Glu883 | Glu957, ASN1008, ASP1021 |
a) Ligand docking score was calculated by Schrödinger software; b) Amino acid residues involved in hydrogen bonds and Pi interactions between the pose and JAK1; c) Amino acid residues involved contacted or crushed with compounds.
Regard to the second docking studies against STAT3 according to the same protocol, all compounds exhibited nearly identical and low Docking Scores, regardless of their structures and stereochemistries as shown in Table 2. The discrepancy between virtual and real experimental values for (S)-2d and (R)-2d could be inferred based on 3D crystal structure of STAT3-lignad complex in which the binding site of ligand is located on the outer exposed pocket of STAT3 protein (Supplementary Fig. S11). Therefore, the chiral centers of both enantiomers, (S)-2d and (R)-2d maybe have a relatively less impact on their binding affinities to STAT3 protein as compared to the inner binding pocket of JAK1 protein. This overall combination of Western blotting data and docking data strongly suggests a plausible mechanism that instead of direct inhibition of STAT3 activation, (S)-2d as a JAK1 inhibitor could suppress the STAT3 phosphorylation, resulting in its strong anti-proliferative effect on A549 cancer cell. That is, (S)-2d (MRC-D-004) as a novel inhibitor of JAK1 would exhibit anti-proliferative activity against A549 lung cancer cells through the JAK1/STAT3 axis.
Experimental
ChemistryGeneral Experimental ProceduresAnalytical TLC was performed on silica gel precoated on glass-backed plates (Fluka Kieselgel 60F254, Merck, Burlington, MA, U.S.A.). An UV light (λ = 254 nm) was used for the detection. Flash chromatography was performed on silica gel 60 (particle size 230–400 mesh, Merck). Commercially available reactants were supplied by Sigma-Aldrich (St. Louis, MO, U.S.A.) and used without further purification. Melting points were measured on a MEL-TEMP ® 3.0 capillary melting point apparatus (Sigma-Aldrich). Optical rotations were obtained with an Autopol III automatic polarimeter (Rudolph Research Analytical). 1H-NMR spectra on a Bruker Avance DXR 400 (400 MHz) spectrometer (Bruker, Billerica, MA, U.S.A.) and Jeol JNM-ECZR 500 MHz spectrometer (Tokyo, Japan) were recorded in CDCl3. Chemical shifts were reported in δ (ppm) units relative to tetramethylsilane (TMS). High resolution (HR) MS were recorded on a Thermo Scientific ™ Q-Exactive mass spectrometer (ionization mode: ESI+) (Thermo Fisher Scientific, Waltham, MA, U.S.A.). Enantiomeric excess (% ee) was determined by high performance liquid chromatography (Agilent 1100; Santa Clara, CA, U.S.A.) with a chiral column (Chiralpak AD-H or Chiralpak IK-3 or Chiralpak IB-N3).
Preparation of (S)-Decursinol (1)The dried root of Angelica gigas (1.0 kg) was extracted three times with EtOH (2.0 L) at 80 °C for 24 h and then filtered. The concentrated extract was partitioned between water (1 L) and ether (1 L). The concentrated ether layer was hydrolyzed with 500 mL of 10% NaOH in a mixed solvent (1 L) of THF/H2O (v/v = 1 : 1) for 24 h. The reaction mixture was acidified to pH 2 with conc. HCl and extracted with CH2Cl2. The combined CH2Cl2 layer was dried with MgSO4, purified with activated charcoal, concentrated and triturated with a small amount of CH2Cl2 to afford pure (S)-decursinol (1, 31.5 g, 3.15%, >93% ee), whose spectral data was identical to that of authentic compound11,29): [α]22D +10.1 (c 0.9, CHCl3); 1H-NMR (400 MHz, CDCl3) δ: 7.57 (1H, d, J = 9.5 Hz), 7.17 (1H, s), 6.77 (1H, s), 6.22 (d, J = 9.5 Hz, 1H), 3.87 (1H, dd, J = 4.6, 5.8 Hz, 1H), 3.11 (1H, dd, J = 4.6, 17.1 Hz, 1H), 2.84 (dd, J = 5.8, 17.1 Hz, 1H), 1.39 (s, 3H), 1.37 (s, 3H), 1.98 (s, 1H).
Preparation of Xanthylein (3)(S)-Decursinol (1, 1 g, 4.06 mmol, 1 equivalent (equiv.)) and triphenylphosphine (2.66 g, 10.15 mmol, 2.5 equiv.) was dissolved in a mixed solvent of CH3CN and CCl4 (14 mL, 1 : 1 (v/v)). The reaction mixture was stirred for 3 h at 60 °C, concentrated under reduced pressure and purified by flash silica gel column chromatography on (EtOAc : Hex = 1 : 3) to afford a xanthylein (3, 0.78 g, 84%), whose spectral data was identical to that of authentic compound30): 1H-NMR (400 MHz, CDCl3) δ: 7.60 (1H, d, J = 9.2 Hz), 7.07 (1H, s), 6.74 (1H, s), 6.36 (1H, d, J = 10.0 Hz), 6.24 (1H, d, J = 9.2 Hz), 5.71 (1H, d, J = 10.0 Hz), 1.48 (6H, s).
Preparation of (E)-3-(Pyridin-2-yl)acrylic Acid (7)Malonic acid (0.49 g, 4.67 mmol) and piperidine (0.05 mL, 0.47 mmol, 0.1 equiv.) was dissolved in pyridine (0.5 mL) and slowly followed by an addition of 2-pyridinecarboxaldehyde (6, 0.44 mL, 4.67 mmol, 1 eq). The reaction mixture was heated at reflux for 2.5 h. After the complete reaction on TLC, the reaction mixture was cooled to room temperature and treated with EtOAc and water. The separated EtOAc was dried over MgSO4, concentrated under reduced pressure and triturated with CH2Cl2 to provide a targets compound (7, 0.14 g, 21%): 1H-NMR (400 MHz, DMSO-d6) δ: 12.64 (1H, br s), 8.64 (4H, d, J = 4.0 Hz), 7.88–7.84 (1H, m), 7.33 (1H, d, J = 8.0 Hz), 7.60 (1H, d, J = 15.6 Hz), 7.41-7.39 (1H, m), 6.83 (1H, d, J = 15.6 Hz).
Preparation of 3-(1H-imidazol-1-yl)propanoic Acid (9)The solution of midazole (8, 1 g, 14.69 mmol) and acrylic acid (1.11 mL, 16.16 mmol, 1.1 equiv.) in DMF (20 mL) was heated at reflux for 2.5 h. After the complete reaction on TLC, the reaction mixture was triturated with EtOAc to provide a solid, which was washed with acetone to afford a pure target compound (9, 1.72 g, 83%): 1H-NMR (400 MHz, DMSO-d6) δ: 12.52 (1H, br s), 7.60 (1H, s), 7.16 (1H, s), 6.86 (1H, s), 4.16 (2H, t, J = 6.8 Hz), 2.72 (2H, t, J = 6.8 Hz).
Preparation of (E)-3-(4-(N,N-Dimethylsulfamoyl)phenyl)acrylic Acid (11)Chlorosulfonic acid (3.6 mL, 54.00 mmol, 8 equiv.) was cooled to 0 °C and slowly treated with trans-cinammic acid (10, 1.00 g, 6.75 mmol). After stirring at room temperature for 5 h, the reaction mixture was treated with water and filtered. The filtered cake was washed with CH2Cl2 to give (E)-3-(4-(chlorosulfonyl)phenyl)acrylic acid (1.32 g, 79%), which was treated with dimethylamine solution 40% wt. % in H2O (6.1 mL, 48.2 mmol, 9 eq) at 0 °C. The reaction mixture was stirred for 17 h and acidified to pH 4 with conc. HCl. The resultant solid was filtered and wished with CH2Cl2 to give a target compound (11, 1.02 g, 76%): 1H-NMR (400 MHz, DMSO-d6) δ: 12.65 (1H, br s), 7.96 (2H, d, J = 8.0 Hz), 7.75 (2H, d, J = 8.0 Hz), 7.67 (1H, d, J = 16.0 Hz), 6.71 (1H, d, J = 16.0 Hz), 2.62 (6H, s).
Preparation of Chiral α-Xanthyletin Oxide (4)The mixed solution of 15% NaOCl (33 mL) and 0.05 M NaHPO4 (13 mL) was adjusted to pH 11.3 by using 1 M NaOH or 1 M HCl and then treated with a solution of xanthylein (3, 0.7 g, 3.07 mmol, 1 equiv.) and (R,R)-Jacobsen’s catalyst (39 mg, 0.06 mmol, 0.02 eq) in CH2Cl2 (8 mL) at 0 °C. The reaction mixture was stirred at 0 °C into room temperature for 72 h and poured into a mixture of CH2Cl2 and H2O. The separated organic layer was dried over MgSO4, concentrated under reduced pressure and purified with flash silica gel column chromatography (EtOAc : Hex = 1 : 3) to provide a target compound (4, 0.29 g, 39%, 96% ee: Supplementary Figs. S2, S3), whose spectral data was identical to the previously reported data31): 1H-NMR (400 MHz, CDCl3) δ: 7.65 (1H, d, J = 9.6 Hz), 7.48 (1H, s), 6.79 (1H, s), 6.30 (1H, d, J = 9.6 Hz), 3.99 (1H, d, J = 4.4 Hz), 3.56 (1H, d, J = 4.4 Hz), 1.63 (3H, s), 1.33 (3H, s).
Preparation of Chiral β-Xanthyletin Oxide (Entipode-4)By using (S,S)-Jacobsen’s catalyst and xanthyletin, the same procedure for chiral α-xanthyletin oxide (4) was carried out to provide a target compound (ent-4, 50%, 95% ee: Supplementary Figs. S4, S5), whose spectral data was identical to that of chiral α-xanthyletin oxide (4.96 min) except for its different retention time (4.15 min) on chiral HPLC: 1H-NMR (400 MHz, CDCl3) δ: 7.65 (1H, d, J = 9.6 Hz), 7.48 (1H, s), 6.79 (1H, s), 6.30 (1H, d, J = 9.6 Hz), 3.99 (1H, d, J = 4.4 Hz), 3.56 (1H, d, J = 4.4 Hz), 1.63 (3H, s), 1.32 (3H, s).
Preparation of Unnatural (R)-Decursinol (5)Chiral α-xanthyletin oxide (4, 0.2 g, 0.82 mmol) and NaCNBH3 (57 mg, 0.90 mmol, 1.1 equiv.) was dissolved in THF (15 mL) and cooled to 0 °C. To the solution was added BF3OEt2 (0.10 mL, 0.82 mmol, 1 equiv.) and the reaction mixture was further stirred for 10 min. After the complete reaction on TLC, the reaction mixture was pour poured into a mixture of EtOAc and H2O. The separated organic layer was dried over MgSO4, concentrated under reduced pressure and purified with flash silica gel column chromatography (only EtOAc) to provide a target compound (5, 0.2 g, 99%), whose spectral data was identical to that of natural (S)-decursinol (1) except for the (−)-sign of optical rotation28): [α]22D −10.4 (c 1.0, CHCl3); 1H-NMR (400 MHz, CDCl3) δ: 7.60 (1H, d, J = 9.6 Hz), 7.20 (1H, s), 6.80 (1H, s), 6.24 (1H, d, J = 9.6 Hz), 3.89 (1H, q, J = 5.6 Hz), 3.14 (1H, dd, J = 16.8 Hz and 4.8 Hz), 2.86 (1H, dd, J = 16.8 Hz and 4.8 Hz), 1.85 (1H, d, J = 6.8 Hz), 1.41 (3H, s), 1.38 (3H, s).
Preparation of (S)-Decursinol (1)By chiral β-xanthyletin oxide, the same procedure for chiral α-xanthyletin oxide (4) was carried out to provide a target compound [(S)-decursinol (1), >99%], whose spectral data was identical to that of natural (S)-decursinol (1).
Representative Preparation of (S)-2d (MRC-D-004)To a stirred solution of (S)-decursinol (1, 0.10 g, 0.41 mmol), (E)-3-(furan-3-yl)acrylic acid (0.06 g, 0.45 mmol, 1.1 eq) and DMAP (0.02 g, 0.20 mmol, 0.5 eq) in dried CH2Cl2 (25 mL) was added EDC·HCl (0.09 g, 0.49 mmol, 1.2 equiv.) at 0 °C and the resultant reaction mixture was further stirred for 19 h at room temperature. After the complete reaction on TLC, the reaction mixture was treated with distilled water (30 mL). The separated organic layer was dried over MgSO4, concentrated under reduced pressure and purified with flash silica gel column chromatography (EA : Hex = 1 : 3) to provide a solid, which was triturated with Ether/Hex to a pure target compound [(S)-2d (MRC-D-004), 0.09 g, 59%, 92% ee]: Supplementary Fig. S6–S8 of supplementary data): mp 128–130 °C; 1H-NMR (400 MHz, CDCl3) δ: 7.66 (1H, s), 7.60–7.56 (2H, m), 7.43 (1H, s), 7.19 (1H, s), 6.84 (1H, s), 6.57 (1H, s), 6.25 (1H, d, J = 10.0 Hz), 6.15 (1H, d, J = 15.6 Hz), 5.19 (1H, t, J = 4.3 Hz), 3.25 (1H, dd, J = 16.0 and 4.3 Hz), 2.94 (1H, dd, J = 16.0 and 4.3 Hz), 1.44 (3H, s), 1.40 (3H, s); 13C-NMR (100 MHz, CDCl3) 23.4, 24.9, 27.9, 70.0, 76.7, 104.8, 107.3, 112.9, 113.4, 115.7, 117.1, 122.4, 128.8, 135.8, 143.2, 144.6, 144.9, 154.2, 156.4, 161.3, 166.3; HRMS (ESI+): m/z calcd for C21H18O6 [M + H]+ 367.1176, found 367.1183.
Representative Preparation of (R)-2d (MRC-D-17)By using (R)-decursinol, the same procedure for (S)-2d (MRC-D-004), was carried out to provide a target compound [(R)-2d, 33%, 96% ee: Supplementary Fig. S9–S10 of supplementary data], whose spectral data was identical to that of (S)-2d (MRC-D-004: 5.96 min) except for its different retention time (5.48 min) on chiral HPLC: mp 129–132 °C; 1H-NMR (400 MHz, CDCl3) δ: 7.66 (1H, s), 7.62–7.56 (2H, m), 7.43 (1H, s), 7.19 (1H, s), 6.84 (1H, s), 6.57 (1H, s), 6.25 (1H, d, J = 9.6 Hz), 6.15 (1H, d, J = 16.0 Hz), 5.19 (1H, t, J = 4.4 Hz), 3.25 (1H, dd, J = 16.8 and 4.4 Hz), 2.94 (1H, dd, J = 16.8 and 4.4 Hz), 1.44 (3H, s), 1.40 (3H, s); 13C-NMR (125 MHz, CDCl3) δ: 23.5, 25.0, 28.0, 70.1, 76.8, 104.9, 107.4, 113.0, 113.5, 115.8, 117.2, 122.5, 128.8, 135.9, 143.3, 144.7, 145.0, 154.3, 156.5, 161.4, 166.4; HRMS (ESI+): m/z calcd for C21H19O6 [M + H]+ 367.1182, found 367.1175.
The structures, 1H-NMR, 13C-NMR, HRMS data, and % yields of the other decursin derivatives were provided in Supplementary Materials.
BiochemistryCell CultureA549 cells derived from human non-small lung cancer were obtained from American Type Culture Collection (ATCC, Manassas, VA, U.S.A.). A549 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) medium (LM011-05, Welgene, South Korea) supplemented with 10% fetal bovine serum (FBS; 16000-044, Gibco, Amarillo, TX, U.S.A.) and 1% penicillin–streptomycin (LS 202-02, Welgene, South Korea). A549 cells maintained in a conventional humidified incubator set at 37 °C in an atmosphere of 5% CO2.
Cell Viability AssayCells treated with the various concentrations of drugs for 24 h were determined using 2,5-diphenyl-2H-tetrazolium bromide (MTT; M2128, Sigma-Aldrich) assay. Cells were seeded in 96 well plate and incubated in incubator. After being treated with decursin or decursinol derivative (12.5, 25, 50, 100, and 200 µM) for 48 h, the cells were treated with MTT for 2 h. Next, the cells were treated with MTT for 24 h and measured at 590 nm using microplate reader. Then, MTT lysis buffer was prepared as a solution of 20% sodium dodecyl sulfate (SDS; Cat. 74255, Sigma-Aldrich) after mixing N, N-dimethylformamide (D4551, Sigma-Aldrich) and water in a 1 : 1 ratio. IC50 values for measured cell viability were calculated in an online IC50 calculator webtool (IC50 Calculator, AAT Bioqeust, Pleasanton, CA, USA, https://www.aatbio.com/tools/ic50-calculator).20,32)
Western BlottingA549 cells were lysed using RIPA buffer including protease inhibitor, Na3VO4, NaF, DTT and PMSF and incubated for 30 min at ice. The protein concentration was measured using a Brad-Ford assay dye kit (Bio-Rad Protein Assay Dye Reagent Concentrate, #5000006, Bio-Rad, Hercules, CA, U.S.A.). The proteins were separated by SDS-PAGE gel and transferred to PVDF membranes (Millipore, Bedford, MA, U.S.A.). The membranes were blocked with 5% skim milk. The blocked membranes were incubated with primary antibody for CXCR7 (PA3-069, 1 : 5000, Thermo Fisher Scientific), GAPDH (5174, 1 : 1000, CST, Danvers, MA, U.S.A.), p-JAK1 (3331, 1 : 1000, CST, U.S.A.), JAK1 (3332, 1 : 1000, CST), p-JAK2 (4406, 1 : 1000, CST), JAK2 (3230, 1 : 1000, CST), p-STAT3 (9145, 1 : 1000, CST), STAT3 (4904, 1 : 1000, CST), CDK2 (2546, 1 : 1000, CST), Cyclin D1 (2978, 1 : 1000, CST), Cyclin D3 (2936, 1 : 1000, CST), p27 KIP1 (3686, 1 : 1000, CST), p-cdc2 (4539, 1 : 1000, CST), Cleaved caspase 8 (9496, 1 : 1000, CST), Cleaved caspase 9 (7237, 1 : 1000, CST) and PARP (9542, 1 : 1000, CST) All primary antibodies were sued with overnight incubation at 4 °C. Next, after being incubated secondary antibodies for 1 h, the membranes were treated with ECL kit (Lumi-pico, DG-WP250, Dogen, South Korea).
Molecular Docking AnalysisMaestro version 12.5: Molecular docking analysis was performed using Maestro version 12.5 and the accompanying software Schrödinger Release 2020-3.28) Two X-ray crystal structures of JAK1 (PDB code: 3EYG)26) and STAT3 (PDB code: 6NJS)27) complexes were retrieved from Protein Data Bank. The refined protein structure was prepared using Protein Preparation Wizard, resulting in the protein structure without ligand. The ligand was prepared by LigPrep module, which is a utility of Schrödinger software suit that combines tool for generating 3D structure from 2D structure. The Glide grid file was generated using the receptor grid generation. This process involves selecting the co-crystallized ligand located at the active site of the protein to automatically determines its X, Y, and Z coordinates. Glide utilizes the position and size of the ligand to calculate a default size and centroid for the region for which grids will be calculated. The Ligand docking tool was applied for the calculation of docking score and interactions between the ligand and protein. Flexible ligand docking was performed using the Extra Precision (XP) feature of Glide module. (S)-2d [MRC-D-004], (R)-2d, (S)-2j, and decursin were run for the comparison. The best pose of each compound with the highest docking score was selected and analyzed for its binding mode including hydrogen bond.
