New Cladiellin-Type Diterpenoids from the South China Sea Soft Coral Cladiella krempfi: Structures and Molecular Docking Analysis in EGFRs

Two new cladiellin-type diterpenoids (1 and 2) and four known related compounds 3–6, were isolated from the South China Sea soft coral Cladiella krempfi. Compound 2 is the third example of cladiellins of an unusual peroxy group in the C-6 position in C. krempfi. The structures and absolute configurations of the new compounds were established by extensive spectroscopic analysis, X-ray diffraction, and/or chemical correlation. In bioassay, all the compounds were evaluated for cytotoxicity and epidermal growth factor receptor (EGFR) inhibitory activity. A molecular docking experiment was conducted to study the structure–activity relationship of cladiellin-type diterpenoids on EGFR inhibitory activity.


Introduction
Soft corals of the genus Cladiella (order Alcyonacea, family Alcyoniidae) are widely distributed in the tropical Indo-Pacific area [1]. They are rich sources of diverse and complex diterpenoids, especially 2,11-cyclized cembranoids, which are divided into three related structural classes: cladiellins (ether formation between C-2 and C-9, also known as eunicellins [2]), briarellins (ether formation between C-2/C-9 and C-3/C- 20), and asbestinins (converted from the briarellin scaffold through a 11,12-methyl shift) [3]. A literature survey indicated that cladiellins isolated from C. krempfi, usually with signal ether bridge (C2-C9) [4,5], also have different oxidation patterns, such as two ether bridges (C-2/C-9 and C-3/C-7 or C-2/C-9 and C-12/C- 19), or the esterification of C-19 [6][7][8]. Some of these diterpenoids also showed broad biological activities, such as anti-inflammatory [9], antifouling [10], antiproliferative [4], and cytotoxic effects [11]. Particularly, sclerophytin A showed significant cytotoxicity against mouse lymphocytic leukemia L1210 cells at a concentration as low as 1 ng/mL ( Figure 1) [12]. Moreover, epidermal growth factor receptor (EGFR), related to the inhibition of tumor cell proliferation, angiogenesis, tumor invasion, metastasis, and apoptosis, is the only target protein found so far that cladiellin-type diterpenoids play a role in. For instance, Sayed and co-workers reported that pachycladin A exhibited promising EGFR inhibitory activity with an IC 50 value of 0.5 µM (Figure 1) [13]. Therefore, the further investigation of cladiellins have attracted chemists and pharmacologists from all over the world [13,14]. sclerophytin A and pachycladin A, such as new and bioactive cladiellins, C. krempfi was collected off Ximao Island, Hainan Province, China and chemically investigated, resulting in the isolation and characterization of two new cladiellin-type diterpenoids, namely litophynols C and D (1 and 2) ( Figure 1) and four known ones (3−6). Herein, we report the isolation, structure elucidation, biological evaluation, and the preliminary structure-activity relationship (SAR) analysis of the isolates, assisted by the molecular docking experiment.
Litophynol C (1) was isolated as an optically active colorless oil. Its molecular formula was established as C26H40O7 by HR-ESIMS (m/z 487.2671 [M + Na] + , calcd. 487.2666), indicating seven degrees of unsaturation. The 13 C NMR, DEPT, and HSQC spectra of 1 revealed 26 carbon signals including five sp 3 methyls, five sp 3 methylenes, nine sp 3 methines (five oxygenated ones at δC 91.1, 79.2, 83.7, 71.5, and 68.4), one oxygenated sp 3 quaternary carbon (δC 84.7), two sp 2 methylene, and four sp 2 quaternary carbons (two ester carbonyls at δC 172.7 and 170.8). The diagnostic 1 H and 13 C NMR resonances, as well as coupling constants of the connected protons (Table 1), indicated the presence of two In our ongoing efforts on the discovery of novel and bioactive marine natural products, several cladiellins have been isolated from soft corals [4,5,15]. In order to obtain more sclerophytin A and pachycladin A, such as new and bioactive cladiellins, C. krempfi was collected off Ximao Island, Hainan Province, China and chemically investigated, resulting in the isolation and characterization of two new cladiellin-type diterpenoids, namely litophynols C and D (1 and 2) ( Figure 1) and four known ones (3)(4)(5)(6). Herein, we report the isolation, structure elucidation, biological evaluation, and the preliminary structure-activity relationship (SAR) analysis of the isolates, assisted by the molecular docking experiment.
The HMBC correlations from H-10 to C-11/C-12, from H 2 -17 to C-10/C-11/C-12, and from H-12 to C-14 revealed the presence of a cyclohexane ring (ring A) with a terminal double bond at C-11 and an isopropyl group at C-14. The cross peaks from H 3 -15 to C-2/C-3/C-4 and from H-2 to C-4/C-9/C-10/C-14 revealed that ring A and fragment c were connected via C-2 and C-3. Additionally, the cross peaks from H 2 -16 to C-6/C-7/C-8 suggested a cyclodecane ring fused with ring A at C-1 and C-10. Furthermore, the strong correlation from H-2 to C-9 and the remaining degree of unsaturation all indicated an ether bridge between C-2 and C-9, which divided the cyclodecane ring into a tetrahydrofuran ring B and an oxocane ring C. Finally, the acetyl group at C-8 was deduced by the HMBC correlation from H-8 to C-1 and from H 3 -2 to C-1 , which also suggested that the connection of the rest butyryl group should be at C-3. Thus, the planar structure of 1 was determined as shown in Figure 1. The HMBC correlations from H-10 to C-11/C-12, from H2-17 to C-10/C-11/C-12, and from H-12 to C-14 revealed the presence of a cyclohexane ring (ring A) with a terminal double bond at C-11 and an isopropyl group at C-14. The cross peaks from H3-15 to C-2/C-3/C-4 and from H-2 to C-4/C-9/C-10/C-14 revealed that ring A and fragment c were connected via C-2 and C-3. Additionally, the cross peaks from H2-16 to C-6/C-7/C-8 suggested a cyclodecane ring fused with ring A at C-1 and C-10. Furthermore, the strong correlation from H-2 to C-9 and the remaining degree of unsaturation all indicated an ether bridge between C-2 and C-9, which divided the cyclodecane ring into a tetrahydrofuran ring B and an oxocane ring C. Finally, the acetyl group at C-8 was deduced by the HMBC correlation from H-8 to C-1′′ and from H3-2′′ to C-1′′, which also suggested that the connection of the rest butyryl group should be at C-3. Thus, the planar structure of 1 was determined as shown in Figure 1.
In order to confirm the absolute configuration of 1, we tried with much effort and fortunately obtained its single crystal from the recrystallization of 1 in methanol, which allowed a successful performance of X-ray crystallography study using Cu Kα radiation (λ = 1.54178 Å). Analysis of the X-ray data unambiguously confirmed the planar structure The relative configuration of 1 was established by analysis of its NOESY spectrum ( Figure 2). As the orientation of H-1 in all the reported cladiellin-type diterpenoids was determined as β [17], that of compound 1 was also arbitrarily assigned as s β-configuration. The NOE correlation of H-1/H-10 and H-10/H-8 suggested that H-8 and H-10 were all β-configurations. In addition, the cross peaks of H-6/H 3 -15, H 3 -15/H-2, H-2/H-9, and H-9/H-14 indicated the α-configuration in H-6, Me-15, H-2, H-9, and H-14. Above all, the relative configuration of 1 was ambiguously deduced as 1R*, 2R*, 3R*, 6R*, 8R*, 9S*, 10R*, and 14R*. Due to the confusing NOE correlations between adjacent H signals, the orientation of the rest of H-12 was still hard to define.
In order to confirm the absolute configuration of 1, we tried with much effort and fortunately obtained its single crystal from the recrystallization of 1 in methanol, which allowed a successful performance of X-ray crystallography study using Cu Kα radiation (λ = 1.54178 Å). Analysis of the X-ray data unambiguously confirmed the planar structure of 1, and determined its absolute configuration as 1R, 2R, 3R, 6R, 8R, 9S, 10R, 12S, 14R [Flack parameter was 0.04(13)] ( Figure 3). Thus, the structure of 1 was determined, namely litophynol C ( Figure 1).
Litophynol D (2) was isolated as an optically active colorless oil. Its molecular formula was established as C 26 H 40 O 8 by HR-ESIMS (m/z 503.2618 [M + Na] + , calcd. 503.2615), indicating the presence of seven degrees of unsaturation. The 13 C NMR, DEPT, and HSQC spectra of 2 revealed 26 carbon signals, including five sp 3 methyls, five sp 3 methylenes, nine sp 3 methines, one oxygenated sp 3 quaternary carbon, two sp 2 methylene, and four sp 2 quaternary carbons. In fact, as shown in Table 1, the NMR data of 2 were extremely close to those of 1, indicating that they were structure analogs. After the careful comparison of 1D and 2D NMR spectra, the main differences between 1 and 2 were the deshielding of C-6 from δ C/H 68.4/4.71 in 1 to 81.6/4.93 in 2. The chemical shifts in the C-6 surrounding carbons also changed slightly, e.g., C-5 from δ C/H 35.9/1.74, 2.20 in 1 to 30.4/1.50, 2.16 in 2; C-7 from δ C 149.4 in 1 to 145.0 in 2, which strongly implied the hydroxyl at C-6 in 1 was replaced by the hydroperoxide group in 2. The 16 mass units more of molecular weight in 2 further supported our determination. The detailed 2D NMR analyses, as shown in Figure 2, confirmed the planar structure and relative configuration of 2.  Figure 3). Thus, the structure of 1 was determined, namely litophynol C (Figure 1). Litophynol D (2) was isolated as an optically active colorless oil. Its molecular formula was established as C26H40O8 by HR-ESIMS (m/z 503.2618 [M + Na] + , calcd. 503.2615), indicating the presence of seven degrees of unsaturation. The 13 C NMR, DEPT, and HSQC spectra of 2 revealed 26 carbon signals, including five sp 3 methyls, five sp 3 methylenes, nine sp 3 methines, one oxygenated sp 3 quaternary carbon, two sp 2 methylene, and four sp 2 quaternary carbons. In fact, as shown in Table 1, the NMR data of 2 were extremely close to those of 1, indicating that they were structure analogs. After the careful comparison of 1D and 2D NMR spectra, the main differences between 1 and 2 were the deshielding of C-6 from δC/H 68. To further confirm our assignment, triphenylphosphine was used to convert the hydroperoxide group of 2 into the corresponding alcohol in CDCl3 (Scheme 1). The reduction of 2 resulted in a compound identical [ 1 H NMR ( Figure 4) and MS data ( Figure S30)] to 1. Thus, the absolute configuration of 2 was unambiguously determined to be the same as that of 1 (Figure 1). To further confirm our assignment, triphenylphosphine was used to convert the hydroperoxide group of 2 into the corresponding alcohol in CDCl 3 (Scheme 1). The reduction of 2 resulted in a compound identical [ 1 H NMR ( Figure 4) and MS data ( Figure S30)] to 1. Thus, the absolute configuration of 2 was unambiguously determined to be the same as that of 1 (Figure 1).  Considering the interesting anti-cancer activities displayed by sclerophytin A and pachycladin A [12,13], all isolates were used to carry out the tests of cytotoxic effects on A549 (human lung cancer) tumor cells and EGFR inhibitory activities; however, none of them showed obvious bioactivities (IC50 > 20 M). By comparing the structures of our isolated compounds with those of the most bioactive, sclerophytin A and pachycladin A, it seemed that the esterification of the hydroxy groups, the site of the esterification, and the length of the ester influenced the activity. Therefore, the selective hydrolysis of the isolated compounds might result in some sclerophytin A and pachycladin A, which could be anti-cancer drug candidates. However, the scarcity of the isolates prevented further chemical correlation. We then tried to do an intensive molecular docking analysis to study the SAR of these types of molecules on EGFRs, aiming to give an insight into future structural modification.

Mar. Drugs
On the basis of the speculation on the SAR, molecular docking analysis was performed using all the isolated compounds, together with sclerophytin A and pachycladin A, on EGFRs. Since the structures of litophynol C (1) and pachycladin A were more similar than other isolated compounds, litophynol C (1) was selected for the detailed comparison with sclerophytin A and pachycladin A ( Figure 5, the docking results of compounds 2−6 are added in Figure S31 and the binding energy of all the docking molecules are listed in Table S2 in Supplemetary Materials). The highly resolved EGFR crystal structure (PDB codes: 5X2A with a resolution of 1.85Å) was used to investigate the possible binding modes of the three compounds within the catalytic domain of the EGFR by means of the Discovery Studio software ( Figure 5). Pachycladin A occupied the same region as the Considering the interesting anti-cancer activities displayed by sclerophytin A and pachycladin A [12,13], all isolates were used to carry out the tests of cytotoxic effects on A549 (human lung cancer) tumor cells and EGFR inhibitory activities; however, none of them showed obvious bioactivities (IC 50 > 20 µM). By comparing the structures of our isolated compounds with those of the most bioactive, sclerophytin A and pachycladin A, it seemed that the esterification of the hydroxy groups, the site of the esterification, and the length of the ester influenced the activity. Therefore, the selective hydrolysis of the isolated compounds might result in some sclerophytin A and pachycladin A, which could be anti-cancer drug candidates. However, the scarcity of the isolates prevented further chemical correlation. We then tried to do an intensive molecular docking analysis to study the SAR of these types of molecules on EGFRs, aiming to give an insight into future structural modification.
On the basis of the speculation on the SAR, molecular docking analysis was performed using all the isolated compounds, together with sclerophytin A and pachycladin A, on EGFRs. Since the structures of litophynol C (1) and pachycladin A were more similar than other isolated compounds, litophynol C (1) was selected for the detailed comparison with sclerophytin A and pachycladin A ( Figure 5, the docking results of compounds 2-6 are added in Figure S31 and the binding energy of all the docking molecules are listed in Table S2 in Supplemetary Materials). The highly resolved EGFR crystal structure (PDB codes: 5X2A with a resolution of 1.85Å) was used to investigate the possible binding modes of the three compounds within the catalytic domain of the EGFR by means of the Discovery Studio software ( Figure 5). Pachycladin A occupied the same region as the ligand of 7XO, which is a member of the SKLB compound series targeting EGFR activation [18], and its C-3 acetate participated in hydrogen bonds with Lys745, Asn842, and Asp855 of the EGFR crystal structures 5X2A ( Figure 5B, upper row). Interestingly, Lys745 lay in the active site of the kinase. Sclerophytin A, Ser720, and ARG841 were combined by three hydrogen bonds, while the C-6 hydroxy of litophynol C (1) participated in hydrogen bond with Asp800, which did not lie in the active site ( Figure 5A,C, upper rows). Furthermore, the C-3 acetate and C-11 butyrate of pachycladin A fully occupied the hydrophobic pocket, which promoted Van der Waals interactions with Gly719, Ser720, Gly721, Val726, Ala743, Gly796, Leu799, Asp800, and Thr854 ( Figure 5B, mid and lower row). The lower binding affinity of compound 1, compared to that of pachycladin A (Table 2), could be deemed to be related to the position and the length of the ester, which might influence the interaction with the hydrophobic pocket and explained why our isolates did not show obvious EGFR inhibitory activity. Although sclerophytin A has been previously reported to be super cytotoxic against mouse lymphocytic leukemia L1210 cells, it displayed less binding affinity than pachycladin A, with lower binding energy ( Table 2), indicating that it might be less effective on non-small cell lung cancer; however, further validation and target fishing of this would be worthwhile. the C-3 acetate and C-11 butyrate of pachycladin A fully occupied the hydrophobic pocket, which promoted Van der Waals interactions with Gly719, Ser720, Gly721, Val726, Ala743, Gly796, Leu799, Asp800, and Thr854 ( Figure 5B, mid and lower row). The lower binding affinity of compound 1, compared to that of pachycladin A (Table 2), could be deemed to be related to the position and the length of the ester, which might influence the interaction with the hydrophobic pocket and explained why our isolates did not show obvious EGFR inhibitory activity. Although sclerophytin A has been previously reported to be super cytotoxic against mouse lymphocytic leukemia L1210 cells, it displayed less binding affinity than pachycladin A, with lower binding energy ( Table 2), indicating that it might be less effective on non-small cell lung cancer; however, further validation and target fishing of this would be worthwhile. Figure 5. In silico binding mode of sclerophytin A, pachycladin A, and litophynol C (1) at EGFR kinase crystal structure 5X2A: upper row-the transparent protein surface, in light grey color, and three compounds shown as sticks with atoms colored C cyan, N blue, O red, and H white, are shown to emphasize the clear combination of hydrogen bonds within the target pocket; middle row-surfaces of 5X2A with combined compounds; lower row-two-dimensional ligand interaction

Biological Material
The soft coral specimen, identified as C. krempfin by Prof. Xiu-Bao Li from Hainan University, was collected along the coast of Ximao Island (east longitude 109 • 22 5.9" and northern latitude 17 • 110 49.7"), Hainan Province, China in 2019, at a depth of 15 m. A voucher specimen (NO. 19XD-6) is available for inspection at the Shanghai Institute of Materia Media, Chinese Academy of Sciences.

EGFR Activity Assays
Six compounds were screened on EGFRs using staurosporine obtained from Med-ChemExpress (Cat. No. HY-15141; Lot. No. 41248) as the reference compound. In the bioassay for the EGFRs, testing kinases were performed in 1× kinase base buffer containing 50 mM HEPES (pH 7.5) and 0.0015% Brij-35, and stop buffer containing 100 mM HEPES (pH 7.5), 0.0015% Brij-35, 0.2% Coating Reagent #3, and 50 mM EDTA. For the compound preparing part, compounds were diluted to 50X of the final desired highest inhibitor concentration in the reaction by 100% DMSO, and 100 µL of the compound dilution was transferred to a well in a 96-well plate. Additionally, 100 µL of 100% DMSO was added to two empty wells for a no-compound control and a no-enzyme control in the same 96-well plate, which was marked as the source plate. Then, 10 µL of compound from the source plate was transferred to a new 96-well plate as the intermediate plate where each well contained 90 µL of 1× kinase buffer. The compounds were mixed in an intermediate plate for 10 min on a shaker. For the assay plate, 5 µL of each well from the 96-well intermediate plate was transferred to a 384-well plate in duplicates. For the kinase reaction, 2.5× enzyme solution and 2.5× peptide solution were prepared. Moreover, 5 µL of compound were settled in the assay plate with 10% DMSO. Additionally, 10 µL of 2.5× enzyme solution was added to each well of the 384-well assay plate that was incubated at room temperature for 10 min. Then, 10 µL of 2.5× peptide solution was added to each well of the 384-well that was incubated at 28 • C for 30 min and stopped by 25 µL stop buffer. The percentage of inhibition was estimated using max-conversion divided by the max-min, and inhibition values were converted by convert conversion values from the caliper program. IC50 values were obtained from fitting the data in the XLfit excel add-in version 4.3.1.

Anti-Tumor Assays
Anti-tumor assays were carried out using A549 (Human lung cancer) tumor cells (ATCC, #CCL-185), following a previously described procedure for a modification of the MTT colorimetric method [21,22], with 5-fluorouracil used as the positive control. To measure the anti-proliferative activity of tested compounds, three concentrations with three replications were performed on the cell line.

Molecular Docking
The cocrystal structure of an EGFR (PDB code 5X2A) was obtained from the RCSB Protein Data Bank. The water molecules were removed in Discovery Studio (DS), while the binding site of sclerophytin A, pachycladin A, and litophynol C (1) on the EGFR was consistent with that of 7XO. Moreover, the combined spherical area was x = 3.789, y = 17.478, z = −29.936, radius = 10.85. A clean protein tool was used to remove protein conformation, supplement incomplete amino acid disability, and refine the hydrogen, and the receptor was prepared for docking. Ligands were sketched in the ChemBioDraw program and uploaded to DS. Furthermore, optimized ligands were obtained through prepared ligand tools and minimized ligand tools. The molecular docking module (CDOCKER) was used for docking, and the simulation with the optimal situation of -CDOCKER ENERGY and -CDOCKER INTERACTION ENERGY scores was analyzed and visualized in Pymol [23,24]. Before calculating the binding energies, all ligands were minimized by an in situ ligand minimization tool. Under the calculated binding-energies tool, Poisson-Boltzamann with non-polar surface area (PBSA) was used to calculate the binding energy (salt concentration = 0.15, rest parameters were default).

Conclusions
In summary, this was the first detailed chemical investigation of C. krempfi from Ximao island in South China Sea. Two new cladiellin-type diterpenoids, litophynols C and D (1 and 2), and four known related compounds 3-6 were isolated and fully elucidated. The stereochemistry of the new compounds was unambiguously determined by extensive spectroscopic analysis, X-ray diffraction analysis, and/or chemical correlation. The discovery of 1 and 2 expanded the diversity and complexity of marine diterpenes. In bioassay, the results obtained were somewhat disappointing since none of the tested compounds are bioactive, while a preliminary SAR of these type of molecules, both isolated and previously reported ones, were studied with the assistance of molecular docking analysis. The above evidence indicated that the most bioactive compound, sclerophytin A, should be regarded as the starting material for further structural modification. It is worth mentioning that this molecule was totally synthesized by Yang and co-workers [10], which enabled further structural modification, and our biological study and SAR analysis, together with those previously reported by Sayed and co-workers [13,25], gave some information to avoid the synthesis of inactive compounds, such as compounds 1-6 related esters.