Three New Malyngamides from the Marine Cyanobacterium Moorea producens

Three new compounds of the malyngamide series, 6,8-di-O-acetylmalyngamide 2 (1), 6-O-acetylmalyngamide 2 (2), and N-demethyl-isomalyngamide I (3), were isolated from the marine cyanobacterium Moorea producens. Their structures were determined by spectroscopic analysis and chemical derivatization and degradation. These compounds stimulated glucose uptake in cultured L6 myotubes. In particular, 6,8-di-O-acetylmalyngamide 2 (1) showed potent activity and activated adenosine monophosphate-activated protein kinase (AMPK).


Introduction
The ocean covers more than 70% of the Earth's surface and hosts huge biological and chemical diversity. Because marine environmental conditions are quite different from terrestrial ones, natural products from marine organisms have unique structures and biological activities. Marine cyanobacteria, in particular, are known to produce various secondary metabolites and have been recognized as a source of pharmaceutical lead compounds [1][2][3]. For example, bisebromoamide, isolated from Lyngbya sp., showed potent cytotoxicity against HeLa S 3 cells [4]. Bisebromoamide inhibited the phosphorylation of extracellular signal-regulated protein kinase (ERK) and was identified as an actin filament stabilizer [5]. Meanwhile, hoshinolactam was found to possess both a cyclopropane ring and γ-lactam ring, and exhibited potent antitrypanosomal activity without cytotoxicity against human fetal lung fibroblast MRC-5 cells [6]. The malyngamide series of natural products have been isolated from various marine filamentous cyanobacteria. Malynagmide A, the first compound of this group, was isolated from Lyngbya majuscule in 1979 [7]; since then, over 30 malyngamide analogs have been isolated [8]. As part of our ongoing effort to identify novel bioactive natural products, we have focused on the constituents of marine cyanobacteria and isolated odoamide [9,10] and odobromoamide [11]. We recently discovered three new malyngamides, 6,8-di-O-acetylmalyngamide 2 (1), 6-O-acetylmalyngamide 2 (2), and N-demethyl-isomalyngamide I (3), from the Okinawan cyanobacterium belonging to the genus Moorea producens. (Figure 1). Herein, we report the isolation, structure determination, and biological evaluation of these compounds.
The structure of the remaining C14H19ClNO6 unit was determined as follows. COSY correlation between H-1/NH and HMBC correlations between H-1/C-2, H-3/C-2, and H-4/C-2 led to the partial structure C-1 to C-4 containing the chloromethylene moiety ( Figure 2b). The geometry of the vinyl chloride was determined to be E by NOESY correlation between H-3/H-1. Additionally, COSY correlations between H-6/H-7 and H-7/H-8, and HMBC correlations between H-4/C-5, H-4/C-9, H-6/C-5, and H-8/C-9, allowed the assignment of the cyclohexanone ring. The chemical shifts of H-6 (δH 5.41), C-6 (δC 72.4), 7 and HMBC correlation from H-6 to the quaternary carbon (δC 170.3) connected the acetyl group (δH 2.16, δC 20.8) to C-6 via the oxygen atom. Similarly, another acetyl group (δH 2.19, δC 21.3) was connected to C-8 via the oxygen atom. The remaining methyl group (δH 1.26, δC 23.5) and OH group (δH 5.69) were connected to C-9 by HMBC correlations between H-10/C-9 and OH/C-9 ( Figure 2c). The connections of these partial structures (Figure 2a-c) were determined on the basis of HMBC data. The NH proton showed a cross peak to the C-1' carbonyl carbon, and H-4 correlated with the C-2 olefinic carbon. Thus, the gross structure of 1 was determined to be that shown in Figure 3a.
Compound 2 was obtained as a colorless oil. The molecular formula of 2 was determined to be C27H44ClNO7 on the basis of the 13 C NMR spectrum (27 carbon signals, Figure S8) and HRESIMS (m/z 552.2692 [M + Na] + , calcd. 552.2699). The 1 H NMR features of 2 ( Figure S7) were very similar to those of 1, but there was only one signal from an acetyl proton. In the 13 C NMR spectrum of 2, there was The marine cyanobacterium M. producens (30 g, wet weight) was collected at Bise, Okinawa Prefecture, Japan, and extracted with methanol. The extract was filtered and concentrated, and then the residue was partitioned between EtOAc and H 2 O. The organic layer was further partitioned between 90% aqueous MeOH and n-hexane. The material obtained from the 90% aqueous portion was fractionated by octadecylsilyl (ODS) column chromatography and subjected to reversed-phase high-performance liquid chromatography (HPLC) to give compound 1 (12.0 mg), 2 (26.8 mg), and 3 (12.6 mg).
The structure of the remaining C 14 H 19 ClNO 6 unit was determined as follows. COSY correlation between H-1/NH and HMBC correlations between H-1/C-2, H-3/C-2, and H-4/C-2 led to the partial structure C-1 to C-4 containing the chloromethylene moiety ( Figure 2b). The geometry of the vinyl chloride was determined to be E by NOESY correlation between H-3/H-1. Additionally, COSY correlations between H-6/H-7 and H-7/H-8, and HMBC correlations between H-4/C-5, H-4/C-9, H-6/C-5, and H-8/C-9, allowed the assignment of the cyclohexanone ring. The chemical shifts of H-6 (δ H 5.41), C-6 (δ C 72.4), 7 and HMBC correlation from H-6 to the quaternary carbon (δ C 170.3) connected the acetyl group (δ H 2.16, δ C 20.8) to C-6 via the oxygen atom. Similarly, another acetyl group (δ H 2.19, δ C 21.3) was connected to C-8 via the oxygen atom. The remaining methyl group (δ H 1.26, δ C 23.5) and OH group (δ H 5.69) were connected to C-9 by HMBC correlations between H-10/C-9 and OH/C-9 ( Figure 2c). The connections of these partial structures (Figure 2a-c) were determined on the basis of HMBC data. The NH proton showed a cross peak to the C-1' carbonyl carbon, and H-4 correlated with the C-2 olefinic carbon. Thus, the gross structure of 1 was determined to be that shown in Figure 3a.
Compound 2 was obtained as a colorless oil. The molecular formula of 2 was determined to be C 27 H 44 ClNO 7 on the basis of the 13 C NMR spectrum (27 carbon signals, Figure S8) and HRESIMS (m/z 552.2692 [M + Na] + , calcd. 552.2699). The 1 H NMR features of 2 ( Figure S7) were very similar to those of 1, but there was only one signal from an acetyl proton. In the 13 C NMR spectrum of 2, there was one less carbonyl carbon compared with that of 1, and thus, compound 2 was thought to be a deacetylated version of compound 1. COSY correlations between H-6/H-7, H-7/H-8, and H-8/OH, and HMBC correlations between H-6/C-5 and H-8/C-9 revealed the position of the acetyl group at C-6. Therefore, the gross structure of 2 was established to be that depicted in Figure 3b. one less carbonyl carbon compared with that of 1, and thus, compound 2 was thought to be a deacetylated version of compound 1. COSY correlations between H-6/H-7, H-7/H-8, and H-8/OH, and HMBC correlations between H-6/C-5 and H-8/C-9 revealed the position of the acetyl group at C-6. Therefore, the gross structure of 2 was established to be that depicted in Figure 3b.   The relative structures of the cyclohexanone rings in 1 and 2 were determined by NOESY experiments. The NOESY spectrum of 1 indicated that H-4, H-6, and 8-OAc were in axial positions of the ring (Figure 4a). NOESY correlations between H-4/H 3 -10 and 8-OAc/H 3 -10 indicated the methyl group at C-9 was in an equatorial position, and thus, the relative configuration of the ring moiety in 1 was 4R*, 6S*, 8S*, and 9S*. The relative configuration of the cyclohexanone ring in 2 was revealed to be 4R*, 6S*, 8S*, and 9S* by NOESY correlations (Figure 4b), and was the same as that 1.  The absolute stereochemistries of the cyclohexanone rings in 1 and 2 were determined as follows. Compound 2 was treated with (R)-and (S)-MTPACl to give (S)-and (R)-MTPA esters, respectively. The 1 H NMR chemical shifts of these esters ( Figures S25 and S26) were assigned on the basis of the COSY spectrum. Calculation of the Δδ(S−R) values ( Figure 5) revealed that C-8 existed in the S configuration [16], and the absolute configuration of the ring moiety in 2 was therefore determined to be 4R, 6S, 8S, and 9S. Compound 2 was derivatized with Ac2O to give an acetylated derivative of compound 2. The optical rotation value of this compound ([α] 25 D +5.5) was identical to that of 1 ([α] 24 D +4.8). Thus, the absolute configuration of the cyclohexanone ring in 1 was determined to be 4R, 6S, 8S, and 9S.
To confirm the absolute configuration of C-7' in 2, compound 2 was hydrolyzed under basic conditions to yield lyngbic acid. The optical rotation of the product ([α] 26 D −10.8) was comparable to the reported value for 7(S)-methoxytetradec-4(E)-enoic acid ([α] 26 D −11.1 [17]), thus establishing the S configuration at the C-7' position in 2. Because the 1 H NMR spectrum ( Figure S24) and optical rotation of the acetylated compound of 2 were identical to those of 1 described above, the absolute configuration of C-7' in 1 was determined to be S. Therefore, the complete stereostructures of compound 1 and 2 were established to be those shown in Figure 1.   The absolute stereochemistries of the cyclohexanone rings in 1 and 2 were determined as follows. Compound 2 was treated with (R)-and (S)-MTPACl to give (S)-and (R)-MTPA esters, respectively. The 1 H NMR chemical shifts of these esters ( Figures S25 and S26) were assigned on the basis of the COSY spectrum. Calculation of the Δδ(S−R) values ( Figure 5) revealed that C-8 existed in the S configuration [16], and the absolute configuration of the ring moiety in 2 was therefore determined to be 4R, 6S, 8S, and 9S. Compound 2 was derivatized with Ac2O to give an acetylated derivative of compound 2. The optical rotation value of this compound ([α] 25 D +5.5) was identical to that of 1 ([α] 24 D +4.8). Thus, the absolute configuration of the cyclohexanone ring in 1 was determined to be 4R, 6S, 8S, and 9S.
To confirm the absolute configuration of C-7' in 2, compound 2 was hydrolyzed under basic conditions to yield lyngbic acid. The optical rotation of the product ([α] 26 D −10.8) was comparable to the reported value for 7(S)-methoxytetradec-4(E)-enoic acid ([α] 26 D −11.1 [17]), thus establishing the S configuration at the C-7' position in 2. Because the 1 H NMR spectrum ( Figure S24) and optical rotation of the acetylated compound of 2 were identical to those of 1 described above, the absolute configuration of C-7' in 1 was determined to be S. Therefore, the complete stereostructures of compound 1 and 2 were established to be those shown in Figure 1.  The absolute stereochemistries of the cyclohexanone rings in 1 and 2 were determined as follows. Compound 2 was treated with (R)-and (S)-MTPACl to give (S)-and (R)-MTPA esters, respectively. The 1 H NMR chemical shifts of these esters ( Figures S25 and S26) were assigned on the basis of the COSY spectrum. Calculation of the ∆δ (S−R) values ( Figure 5) revealed that C-8 existed in the S configuration [16], and the absolute configuration of the ring moiety in 2 was therefore determined to be 4R, 6S, 8S, and 9S. Compound 2 was derivatized with Ac 2 O to give an acetylated derivative of compound 2. The optical rotation value of this compound ([α] 25 D +5.5) was identical to that of 1 ([α] 24 D +4.8). Thus, the absolute configuration of the cyclohexanone ring in 1 was determined to be 4R, 6S, 8S, and 9S.
To confirm the absolute configuration of C-7' in 2, compound 2 was hydrolyzed under basic conditions to yield lyngbic acid. The optical rotation of the product ([α] 26 D −10.8) was comparable to the reported value for 7(S)-methoxytetradec-4(E)-enoic acid ([α] 26 D −11.1 [17]), thus establishing the S configuration at the C-7' position in 2. Because the 1 H NMR spectrum ( Figure S24) and optical rotation of the acetylated compound of 2 were identical to those of 1 described above, the absolute configuration of C-7' in 1 was determined to be S. Therefore, the complete stereostructures of compound 1 and 2 were established to be those shown in Figure 1. conditions to yield lyngbic acid. The optical rotation of the product ([α] 26 D −10.8) was comparable to the reported value for 7(S)-methoxytetradec-4(E)-enoic acid ([α] 26 D −11.1 [17]), thus establishing the S configuration at the C-7' position in 2. Because the 1 H NMR spectrum ( Figure S24) and optical rotation of the acetylated compound of 2 were identical to those of 1 described above, the absolute configuration of C-7' in 1 was determined to be S. Therefore, the complete stereostructures of compound 1 and 2 were established to be those shown in Figure 1.

N-Demethyl-isomalyngamide I (3)
Compound 3 was obtained as a colorless oil. The molecular formula of 3 was determined to be C25H40ClNO5 on the basis of the 13 C NMR spectrum (25 carbon signals, Figure S14) and HRESIMS

N-Demethyl-isomalyngamide I (3)
Compound 3 was obtained as a colorless oil. The molecular formula of 3 was determined to be C 25 H 40 ClNO 5 on the basis of the 13 C NMR spectrum (25 carbon signals, Figure S14) and HRESIMS (m/z 460.2692 [M + H] + , calcd. 470.2668). The NMR data for 3 are summarized in Table 2. The 1D and 2D NMR spectra of 3 revealed that it had a lyngbic acid moiety (Figure 6a), like compound 1 and 2. The structure of the remaining C 10 H 13 ClNO 3 unit was determined as follows. COSY correlation between H-1/NH and HMBC correlations between H-1/C-2, H-1/C-4, and H-3/C-2 led to the partial structure C-1 to C-4 containing a chloromethylene moiety (Figure 6b). Because the geometry of the vinyl chloride could not be determined by NOESY experiments, we conducted HSQMBC NMR experiments [18]. We observed a 6.8 Hz 3 J coupling from H-3 to C-1 and 4.2 Hz 3 J coupling from H-3 to C-4. These coupling constants and comparison with the reported results for malyngamide R [14] revealed the E geometry of this double bond. Additionally, COSY correlations between H-10/H-6, H-6/H-7, H-7/H-8, and H-8/H-9, and HMBC correlations between H-6/C-5 and H-9/C-4 led to the partial structure from C-5 to C-4 ( Figure 6c). The chemical shifts of H-7 (δ H 3.81) and C-7 (δ C 68.7) were consistent with the presence of a hydroxy group at C-7. The remaining component and the high-field chemical shifts of C-4 and C-9 (δ C 61.4 and 62.1, respectively) indicated the presence of an epoxy group at C-4 and C-9 [19,20]. Although HMBC correlations between H-6/C-5 and H-9/C-5 were not observed, C-4 and C-5 should be connected considering the degree of unsaturation of 3. Thus, it became apparent that compound 3 had a cyclohexanone ring, and the gross structure of 3 was determined to be that displayed in Figure 6d. The relative structure of the cyclohexanone ring in 3 was determined by a NOESY experiment. The NOESY spectrum of 3 ( Figure S18) indicated that H-6 and H-8b were in axial positions of the ring. The coupling constant of H-9 (2.6 Hz) and NOESY correlations between H-8b/H-7 and H-8b/H-9 indicated that H-7 and H-9 were in equatorial positions of the ring. Therefore, the relative configuration of the ring moiety in 3 was deduced to be 4S*, 6R*, 7R*, and 9S* (Figure 7).  The relative structure of the cyclohexanone ring in 3 was determined by a NOESY experiment. The NOESY spectrum of 3 ( Figure S18) indicated that H-6 and H-8b were in axial positions of the ring. The coupling constant of H-9 (2.6 Hz) and NOESY correlations between H-8b/H-7 and H-8b/H-9 indicated that H-7 and H-9 were in equatorial positions of the ring. Therefore, the relative configuration of the ring moiety in 3 was deduced to be 4S*, 6R*, 7R*, and 9S* (Figure 7). The relative structure of the cyclohexanone ring in 3 was determined by a NOESY experiment. The NOESY spectrum of 3 ( Figure S18) indicated that H-6 and H-8b were in axial positions of the ring. The coupling constant of H-9 (2.6 Hz) and NOESY correlations between H-8b/H-7 and H-8b/H-9 indicated that H-7 and H-9 were in equatorial positions of the ring. Therefore, the relative configuration of the ring moiety in 3 was deduced to be 4S*, 6R*, 7R*, and 9S* (Figure 7). To determine the absolute configuration of C-7, 3 was treated with MTPACl. However, the MTPA ester of 3 was not obtained; compound 4 was obtained instead (Figure 8). Accordingly, the absolute stereochemistry of the cyclohexanone ring in 3 was not determined. The absolute configuration of C-7' was established to be S using the same method as described above for 2.  To determine the absolute configuration of C-7, 3 was treated with MTPACl. However, the MTPA ester of 3 was not obtained; compound 4 was obtained instead (Figure 8). Accordingly, the absolute stereochemistry of the cyclohexanone ring in 3 was not determined. The absolute configuration of C-7' was established to be S using the same method as described above for 2. To determine the absolute configuration of C-7, 3 was treated with MTPACl. However, the MTPA ester of 3 was not obtained; compound 4 was obtained instead (Figure 8). Accordingly, the absolute stereochemistry of the cyclohexanone ring in 3 was not determined. The absolute configuration of C-7' was established to be S using the same method as described above for 2. Structurally similar compounds, malyngamide I [21] and 8-epi-malyngamide C [20], have been reported. The formation of the carbon framework of compound 3 is predicted to proceed in a similar fashion to those of 8-epi-malyngamide C and jamaicamides, and the branching methyl group C-6 should originate from methionine [12].

Biological Activities
The biological activities of compounds 1, 2, and 3 were evaluated using a glucose uptake assay in cultured L6 myotubes. Compound 1 stimulated glucose uptake in a dose-dependent and insulinindependent manner, and compounds 2 and 3 showed weak activity for glucose uptake (Figure 9). To confirm the involvement of AMP-activated protein kinase (AMPK), which increases insulin- Structurally similar compounds, malyngamide I [21] and 8-epi-malyngamide C [20], have been reported. The formation of the carbon framework of compound 3 is predicted to proceed in a similar fashion to those of 8-epi-malyngamide C and jamaicamides, and the branching methyl group C-6 should originate from methionine [12].

Biological Activities
The biological activities of compounds 1, 2, and 3 were evaluated using a glucose uptake assay in cultured L6 myotubes. Compound 1 stimulated glucose uptake in a dose-dependent and insulin-independent manner, and compounds 2 and 3 showed weak activity for glucose uptake ( Figure 9). To confirm the involvement of AMP-activated protein kinase (AMPK), which increases insulin-independent glucose uptake in skeletal muscle [22,23], we examined the effect of compound C, a selective AMPK inhibitor and performed western blotting with anti-AMPK and anti-phosphorylated AMPK (p-AMPK) antibodies. Compound C markedly lowered glucose uptake stimulated by compound 1 in cultured L6 myotubes (Figure 10a). The expression of p-AMPK (the activated form of AMPK) increased in the cells treated with 40 µM of 1 (Figure 10b). These results indicate that compound 1 stimulated glucose uptake in cultured L6 myotubes via the AMPK pathway, regulating cellular metabolism.
Mar. Drugs 2017, 15, 367 7 of 12 independent glucose uptake in skeletal muscle [22,23], we examined the effect of compound C, a selective AMPK inhibitor and performed western blotting with anti-AMPK and anti-phosphorylated AMPK (p-AMPK) antibodies. Compound C markedly lowered glucose uptake stimulated by compound 1 in cultured L6 myotubes (Figure 10a). The expression of p-AMPK (the activated form of AMPK) increased in the cells treated with 40 μM of 1 (Figure 10b). These results indicate that compound 1 stimulated glucose uptake in cultured L6 myotubes via the AMPK pathway, regulating cellular metabolism.  . The effect of compounds 1, 2, and 3 on glucose uptake in cultured L6 myotubes. Cells were preincubated in Krebs-Henseleit-HEPES buffer (KHH buffer) without glucose for 2 h. They were then incubated in KHH buffer containing 5 mM glucose with the indicated concentrations of compounds for 16 h. Glucose uptake was measured using a Glucose C-II Test kit. Values are the mean ± SD of quadruplicate determinations.

Base Hydrolysis of Compounds 2 and 3
Compound 2 (7.3 mg) was dissolved in a 5.0 mL solution of 10% KOH in 80% aqueous EtOH and refluxed for 14 h. The reaction mixture was concentrated, and the residue was partitioned between EtOAc and H 2 O. The organic layer was subjected to reversed-phase HPLC [Cosmosil 5C 18 -AR-II (10 mm × 250 mm), 80% MeOH with 0.1% TFA at 5.0 mL/min, and UV detection at 215 nm] to yield lyngbic acid (1.6 mg). Using the same procedure as described above, lyngbic acid (4.7 mg) ester was obtained from 3 (13.5 mg).

Determination of Glucose Uptake
L6 myotubes were incubated in filter-sterilized Krebs-Henseleit buffer (1.2 mM MgSO 4 , 1.2 mM KH 2 PO 4 , 4.7 mK KCl, 119 mM NaCl, 2.5 mM CaCl 2 ·2H 2 O, and 25 mM NaHCO 3 , pH 7.4) containing 0.1% bovine serum albumin (BSA), 10 mM HEPES, and 2 mM sodium pyruvate (KHH buffer) for 2 h. The myotubes were then cultured in KHH buffer containing 5 mM glucose with or without compounds 1, 2, and 3 (10-40 µM) for 16 h and without or with compound C (30 µM), an AMPK inhibitor for 6 h. Nepodin [25] was used as a positive control, and DMSO alone was used as a negative control. The concentrations of glucose remaining in KHH buffer were determined by a commercial assay kit (Glucose CII-Test Wako) and a microplate reader at 490 nm. The amounts of glucose uptake by myotubes were calculated from the differences in glucose concentrations between before and after culture.

Western Blotting
L6 myotubes were lysed in Blue Loading Buffer for 1 min after washing with ice-cold PBS. The lysates were sonicated for 10 s, boiled at 100 • C for 10 min and centrifuged at 15,000 rpm for 5 min. The protein concentrations of the supernatants were determined by a commercial assay kit (RC DC Protein Assay, Bio-Rad laboratories Inc., Hercules, CA, USA). Equal amounts