Two Onnamide Analogs from the Marine Sponge Theonella conica: Evaluation of Geometric Effects in the Polyene Systems on Biological Activity

Two previously unreported onnamide analogs, 2Z- and 6Z-onnamides A (1 and 2), were isolated from the marine sponge Theonella conica collected at Amami-Oshima Is., Kagoshima Prefecture, Japan. Structures of compounds 1 and 2 were elucidated by spectral analysis. Structure–activity relationships (SARs) for effects on histone modifications and cytotoxicity against HeLa and P388 cells were characterized. The geometry in the polyene systems of onnamides affected the histone modification levels and cytotoxicity.


Introduction
Marine invertebrates are a rich source of compounds with unique structures and biological activities. Marine invertebrates are composed of the phyla Porifera, Cnidaria, Mollusca, Echinodermata, Chordata and so forth. Marine sponges of the genus Theonella are well known for their rich secondary metabolites, including nonribosomal peptides, polyketides and terpenoids [1][2][3][4][5][6]. Onnamides or theopederins share a common core skeleton produced via polyketide and nonribosomal peptide biosynthetic pathways [7][8][9], and form a group with distinct structures and potent cytotoxicity against cancer cell lines. In this study, the isolation and structure elucidation of two unreported onnamide analogs, 2Z-and 6Z-onnamides A (1 and 2), are described, as well as known analogs 3-6.
Histone modifications play a crucial role in the epigenetic control of gene expression [10][11][12], and perturbations in this gene-switching system are related to chronic diseases such as cancer [13]. From this viewpoint, we have developed an in vitro cell-based assay system for evaluating the effects of compounds on multiple histone modifications in parallel [14]. Among 3750 extracts of marine organisms tested [14], the hydrophobic extract prepared from a marine sponge T. conica collected at Amami-Oshima Is., Kagoshima Prefecture, Japan, markedly enhanced the levels of trimethylated histone H3 lysine 27 (H3K27me3) and reduced the level of acetylated H4 lysine 5 (H4K5ac). Bioassay-guided isolation allowed us to identify onnamide analogs as active components in this marine sponge. Structure-activity relationships (SARs) for the activity of controlling histone modifications as well as for cytotoxicity were examined using six analogs (1)(2)(3)(4)(5)(6). As the result, we observed that some of the histone modification changes occurred at the lower concentration than IC 50 values for cytotoxicity. This finding imply different modes of action for these two biological activities that, however, were not fully supported because of the different sensitivity of HeLa cells to the compounds in the respective assays. Therefore, our conclusion was that the activity in control of histone modification and cytotoxicity changes depending on the geometric isomerism in the side chains, but the different modes of action between these biological activities are not confirmed. Identification of the target gene expressions controlled by the histone modifications affected by onnamides should provide new insights into the underlying mechanisms of action for onnamides. A detailed investigation is currently underway.
The distinct difference between 1 and 3 was observed in the side chain, as observed by the downfield-shifted signals for the polyene protons (H-2 to H-7). The overlapping signals of H-2 and H-3 were assigned based on the coupling constants of J 2,3 (11.3 Hz), J 4,5 (15.0 Hz) and J 6,7 (15.0 Hz), which were obtained by processing data using a modified apodization function [16] (Figure 1). The geometry of the double bond between H-2 and H-3 in 1 was deduced as cis, confirming 2Z-onnamide A as 1. The distinct difference between 1 and 3 was observed in the side chain, as o by the downfield-shifted signals for the polyene protons (H-2 to H-7). The over signals of H-2 and H-3 were assigned based on the coupling constants of J2,3 (11.3 (15.0 Hz) and J6,7 (15.0 Hz), which were obtained by processing data using a m apodization function [16] (Figure 1). The geometry of the double bond between H-3 in 1 was deduced as cis, confirming 2Z-onnamide A as 1. The molecular formula of 6Z-onnamide A (2, C39H63N5O12) was also determin the same as that of 3 by the positive-mode HRESIMS (m/z 794.4530 [M+H] + , c C39H64N5O12 794.4546, Δ−2.0 ppm). The 1 H NMR spectrum of 2 measured in CD3 superimposable with that of 3 (Table 2), except for the cis geometry in Δ 6,7 ( Figure 2 indicated that 2 is 6Z-onnamide A.  The molecular formula of 6Z-onnamide A (2, C 39 H 63 N 5 O 12 ) was also determined to be the same as that of 3 by the positive-mode HRESIMS (m/z 794.4530 [M + H] + , calcd for C 39 H 64 N 5 O 12 794.4546, ∆ − 2.0 ppm). The 1 H NMR spectrum of 2 measured in CD 3 OD was superimposable with that of 3 (Table 2), except for the cis geometry in ∆ 6,7 (Figure 2), which indicated that 2 is 6Z-onnamide A.    The stereochemistry of 2Z-and 6Z-onnamides A (1 and 2) was assigned to the 11R, 13R, 15S, 16R, 17S, 18S, 21S, 22R, 25R, 26R and 2 S configuration, which is the same as onnamide A (3) and accounts for the biosynthetic pathway [8]. Analysis of the 1 H NMR spectrum of 3 in CD 3 OD revealed that 3 undergoes photoisomerization by light irradiation to produce 1, 2 and 4Z-onnamideA (4) ( Table S2). Calyculin and marinomycin analogs have been reported to undergo photoisomerization at their tetraene moieties [17,18]. Thus, we deduced that 1 and 2 are likewise artificially photoisomerized at the polyene moiety of 3. However, it remains unclear whether the geometric isomerism occurs in living T. conica by sunlight or in the laboratory under artificial light. Metabolomic analysis using freshly collected T. conica should provide an answer to this issue.
Compounds 1-6 showed significant cytotoxicity against HeLa cells with IC 50 values of 38-540 nM (Table 3). SARs study using HeLa cells revealed that compound 5 with the reduced C21-C22 single bond is as potent as onnamide A (3), whereas 6 with a shorter side chain showed weaker activity (×1/8.2). The positions of the cis-trans isomerism in the side chains of onnamides also affects cytotoxicity: compound 4 was 1.7-fold more potent than 3, whereas 1 and 2 were 2.5-fold less cytotoxic than 3 (Figure 3). of 38-540 nM (Table 3). SARs study using HeLa cells revealed that compound 5 w reduced C21-C22 single bond is as potent as onnamide A (3), whereas 6 with a side chain showed weaker activity (×1/8.2). The positions of the cis-trans isomerism side chains of onnamides also affects cytotoxicity: compound 4 was 1.7-fold more than 3, whereas 1 and 2 were 2.5-fold less cytotoxic than 3 (Figure 3).  Subsequently, the effects on histone modifications by onnamide A (3) were gated using 16 monoclonal antibodies specific to each histone modification (Figu The results revealed that 3 altered 13 histone modifications at concentrations of 70 nM, that are higher than the IC50 value (66 nM). Compound 3 enhanced the leve methylated histone H3 lysine 4, 27 and 36 (H3K4me3, H3K27me3 and H3K36m reduced the level of acetylated H4 lysine 5 (H4K5ac) at the lowest concentration Figure 4). All four histone modifications are related to cytotoxicity [19][20][21][22][23][24]. As H4 Subsequently, the effects on histone modifications by onnamide A (3) were investigated using 16 monoclonal antibodies specific to each histone modification ( Figure S18). The results revealed that 3 altered 13 histone modifications at concentrations of 70 and 140 nM, that are higher than the IC 50 value (66 nM). Compound 3 enhanced the levels of trimethylated histone H3 lysine 4, 27 and 36 (H3K4me3, H3K27me3 and H3K36me3) and reduced the level of acetylated H4 lysine 5 (H4K5ac) at the lowest concentration (35 nM, Figure 4). All four histone modifications are related to cytotoxicity [19][20][21][22][23][24]. As H4K5ac is associated with DNA replication in the cell cycle, its decrease is consistent with the cell cycle arrest induced by 3 [14].
We also compared the effects of analogs 1-6 on histone modifications of H3K4me3, H3K27me3, H3K36me3 and H4K5ac at the same concentrations (35, 70 and 140 nM, Figure S19). Compound 6 administered at 70 nM (about 8 times less than the IC 50 value of 540 nM) induced changes in the levels of the four histone modifications. However, enhanced cytotoxicity was also observed for 6 in this system ( Figure S20), suggesting that the effects on histone modifications may be as the result of the cytotoxic effects by 6.
Molecules 2023, 28, 2524 6 of 10 associated with DNA replication in the cell cycle, its decrease is consistent with the cell cycle arrest induced by 3 [14]. We also compared the effects of analogs 1-6 on histone modifications of H3K4me3, H3K27me3, H3K36me3 and H4K5ac at the same concentrations (35, 70 and 140 nM, Figure  S19). Compound 6 administered at 70 nM (about 8 times less than the IC50 value of 540 nM) induced changes in the levels of the four histone modifications. However, enhanced cytotoxicity was also observed for 6 in this system ( Figure S20), suggesting that the effects on histone modifications may be as the result of the cytotoxic effects by 6.
Onnamide A (3) and anisomycin [25,26] were reported to inhibit protein synthesis and to elicit ribotoxic stress response (RSR) [27]. RSR is induced in response to ribosomal impairment in the mitogen-activated protein kinase (MAPK)-mediated inflammatory signaling cascade. It includes activation of stress-activated protein kinases (SAPKs), such as p38 and c-Jun N-terminal kinase (JNK), and eventually causes cell death [28,29]. Moreover, activation of SAPKs by anisomycin was reported to be independent of protein synthesis inhibition [25], implying the possible explanation for the difference in cytotoxicity by 3 and anisomycin.
Anisomycin and onnamide A (3) bind to different sites on the ribosome (anisomycin binds to the A site of the ribosome, whereas 3 binds to the E site) [30,31]. The different signaling pathways caused by different binding sites likely explain the weaker cytotoxicity of anisomycin [27]. We had expected that the effects on histone modification could reflect the difference in cytotoxicity, but the similar effects of anisomycin on histone modifications to those of 3 were observed ( Figure S21).
In this study, we could confirm that both cytotoxicity and control of histone modifications change depend on the geometric isomerism in the side chains. Evaluating the effects on the control of histone modifications can be effective way to distinguish modes of action by two different types of cytotoxic compounds, although in this case, the largely overlapping mechanisms by anisomycin and onnamides in cytotoxicity hampered us from clearly distinguishing them.  Onnamide A (3) and anisomycin [25,26] were reported to inhibit protein synthesis and to elicit ribotoxic stress response (RSR) [27]. RSR is induced in response to ribosomal impairment in the mitogen-activated protein kinase (MAPK)-mediated inflammatory signaling cascade. It includes activation of stress-activated protein kinases (SAPKs), such as p38 and c-Jun N-terminal kinase (JNK), and eventually causes cell death [28,29]. Moreover, activation of SAPKs by anisomycin was reported to be independent of protein synthesis inhibition [25], implying the possible explanation for the difference in cytotoxicity by 3 and anisomycin.
Anisomycin and onnamide A (3) bind to different sites on the ribosome (anisomycin binds to the A site of the ribosome, whereas 3 binds to the E site) [30,31]. The different signaling pathways caused by different binding sites likely explain the weaker cytotoxicity of anisomycin [27]. We had expected that the effects on histone modification could reflect the difference in cytotoxicity, but the similar effects of anisomycin on histone modifications to those of 3 were observed ( Figure S21).
In this study, we could confirm that both cytotoxicity and control of histone modifications change depend on the geometric isomerism in the side chains. Evaluating the effects on the control of histone modifications can be effective way to distinguish modes of action by two different types of cytotoxic compounds, although in this case, the largely overlapping mechanisms by anisomycin and onnamides in cytotoxicity hampered us from clearly distinguishing them.

General Experimental Procedures
NMR spectra were recorded on an Avance (400 MHz) spectrometer (Bruker Corporation, Billerica, MA, USA). 1 H and 13 C NMR chemical shifts were referenced to the solvent peaks, δ H 3.31 and δ C 49.15 for CD 3 OD (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan). HRESI-MS spectra were measured on a Triple TOF 4600 (AB Sciex Pte. Ltd., Tokyo, Japan) in the positive mode. Optical rotation was determined on a DIP-1000 digital polarimeter (JASCO Corporation, Tokyo, Japan) in CH 3 OH. UV spectrum was recorded using a UV-1800 spectrophotometer (Shimadzu Corporation, Kyoto, Japan). IR spectrum was measured on a JIR-WINSPEC50 spectrometer(JEOL Ltd., Tokyo, Japan). Fluorescent images were obtained with an IX70 microscope equipped by DP72 (Olympus Corporation, Tokyo, Japan).

Biological Material
The same T. conica specimens as previous work [32,33] were used in this study. T. conica was collected by hand using SCUBA, Amami-Oshima Is., Kagoshima Prefecture, Japan (N 28 • 06.82 , E 129 • 21.09 ) in June 2007. The sample was immediately frozen and kept at −25 • C until extraction.

Isolation
The frozen sponge specimen (1020 g wet wt.) was extracted with CH 3 OH (1 L × 5), and the combined extract was evaporated in vacuo. The concentrated extract was suspended in H 2 O and extracted with CHCl 3 , then n-C 4 H 9 OH. The CHCl 3 and n-C 4 H 9 OH layers were combined and subjected to the Kupchan procedure [15] yielding n-hexane, CHCl 3 and aqueous CH 3 OH layers [32,33]. CHCl 3 layer was concentrated to dryness and then separated by ODS flash chromatography (CH 3 Table S1.

Cytotoxic Test
Cytotoxicity test was conducted as previously reported [34]. Briefly, HeLa cells in DMEM or P388 cells in RPMI (cell concentration, 10,000 cells/mL, 200 µL) were added to each well of 96-well microplates and kept in the incubator at 37 • C under an atmosphere of 5% CO 2 . After 24 h, samples in DMSO with various concentrations of onnamides (1-6) were added to each well. After 72 h cultivation, to each well was added 50 µL of 3-(4,5-dimethyl-2-thiazoyl)-2,5-diphenyl-2H tetrazolium bromide (MTT) saline solution (1 mg/mL, FUJIFILM Wako Pure Chemical Corporation) and they were then kept in the incubator at 37 • C under an atmosphere of 5% CO 2 . After 4 h, medium was removed by aspiration and 150 µL of DMSO was added to each well to lyse cells. Concentration of the reduced MTT was quantified, measuring the absorbance at 650 nm to estimate IC 50 values.

Histone Modification Assay
Assay of histone modification levels was performed by an immunofluorescence using a previously reported method [14], with some modifications. Briefly, HeLa cells were incubated under the medium containing the sample for 20 h and then immunostained. Cells were fixed with 4% paraformaldehyde in PBS for 10 min, 1% Triton X-100 in PBS for 20 min, and blocked in Blocking One-P (Nacalai Tesque Inc., Kyoto, Japan) for 20 min and then incubated in Alexa Fluor 488 (Thermo Fisher Scientific, Inc., Waltham, MA, USA) or Cy3 (Thermo Fisher Scientific, Inc.)-labeled antibodies against each histone modification (1:1000, Monoclonal Antibody Institute, Nagano, Japan) for 2 h with Hoechst 33342 (1:2000, Dojindo Laboratories Co., Ltd., Kumamoto, Japan). Fluorescent images were obtained with a microscope. The relative fluorescence intensity was digitalized using CellProfiler TM software 3.0.0 [35] against those of the control wells. Anisomycin was purchased from FUJIFILM Wako Pure Chemical Corporation.

Conclusions
In conclusion, new onnamide analogs 2Z-and 6Z-onnamides A (1 and 2) and four known onnamides (3)(4)(5)(6) were isolated from the marine sponge T. conica, and their structures were elucidated by MS and NMR spectral analyses. The combined effects on histone modifications and cytotoxicities by these compounds revealed that their modes of action follow those of anisomycin. Identifying modulations in the expression patterns of target genes caused by onnamides via changes in histone modifications should provide new insights into the underlying mechanisms of onnamide activity. A detailed investigation of these mechanisms is currently underway.