Two Furanosesterterpenoids from the Sponge Luffariella variabilis

Abstract

Two new sesterterpenoids, 1 and 2, were isolated from the sponge Luffariella variabilis. Their planar structures were characterized with spectroscopic analyses. The sole chiral center of compound 1 was elucidated as 12R by comparing observed and calculated optical rotation values. The configurations of compound 2 were determined by NMR and electronic circular dichroism (ECD) studies. Furthermore, compound 2 showed cytotoxicity at IC₅₀ 1.0 µM against NBT-T2 cells.

1. Introduction

Among sessile marine macroorganisms, sponges have remained the richest sources of new bioactive molecules [1]. Those belonging to the family Thorectidae are known as sources of sesterterpenoids and other molecules [2,3,4,5,6,7,8,9,10,11,12]. The title sponge Luffariella variabilis, a representative species of this family, has been studied by several research groups and found to be prolific in secondary metabolites [2,3,4,7,11,12]. Of the metabolites, manoalide was originally reported as an antibiotic by Scheuer [2] from a Palauan specimen. It was found to have additional biological activity such as anti-inflammatory [3], inhibitory activity against phospholipase A2 [4] and NS3 helicase of hepatitis C virus [5], and has been distributed as a reagent for life science studies. Additional members of sesterterpenoids with a similar molecular scaffold have been reported from different collections of sponges of the same genus [6,7,8,9,10,11,12]. Here, we report structures of two new sesterterpenoids isolated from a specimen of L. variabilis collected at a deeper coral reef.

2. Results and Discussion

2.1. Structure Elucidation

As a part of collaborative efforts to search for antiviral molecules [13], we constructed a small library of Okinawan marine invertebrates. Although the lipophilic extract of the title sponge did not inhibit the growth of target viruses, it showed moderate cytotoxicity with a unique ¹H-NMR spectrum. It was then sequentially separated with silica gel columns to give compounds 1 and 2 as glassy or oily substances.

The molecular formula of compound 1 was determined as C₂₅H₃₀O₅ by observing its sodiated ion at m/z 433.19909 (∆ + 1.2 ppm). As eight out of eleven degrees of unsaturation can be accounted for by the presence of one carbonyl (1715 cm⁻¹, δ_C 167.0 s) and seven double bonds (14 signals in the range from δ_C 156.4 to 106.0), the molecule was determined to contain three rings. Of the double bonds, two were incorporated into a β-substituted furan (δ_H 7.34 brs, 7.21 brs, 6.26 brs; δ_C 142.7 d, 138.9 d, 124.5 s, 110.9 d), while another two were assigned to a disubstituted furan moiety (δ_H 7.96 brs, 6.38 brs; δ_C 156.4 s, 147.7 d, 119.0 s, 106.0 d) after NMR inspection (Figures S1 and S2). Using 2D NMR (Figures S3–S5) analyses, three other double bonds were involved in the following three substructures: (a) –CH₂–CH₂–CH=C< (δ_H 5.33 m; δ_C 139.8 s, 119.2 d), (b) –CH₂–CH(O–)–CH=C< (δ_H 5.24 m; δ_C 139.7 s, 124.9 d), and (c) –CH₂–CH₂–CH=C< (δ_H 5.24 m; δ_C 130.9 s, 127.2 d) (Figure 1). The remaining pieces of the molecule were an oxymethylene (δ_H 4.39 d, 4.20 d; δ_C 68.3), an sp³ hybridized methylene (δ_H 3.28 s; δ_C 38.1), and two vinyl methyls (δ_H 1.70 brs, 1.60 brs; δc_C 16.7, 15.9).

The oxymethylene group was incorporated into a five-membered ether ring connecting the substructures a and b, as shown by the HMBC cross peaks: H-22/C-12,14,15, H-12/C-22, and H-13/C-14,15,22. The sp³ hybridized methylene at δ_H 3.28 was flanked by the disubstituted furan and an olefin connected to substructure c by observing H-5/C-3,4,6,7,24 and H-24/C-5,6,7. Finally, additional correlations of H-23/C-9,10,11, H-17/C-19,21, and H-19,21/C-17 (Figure 1) allowed us to connect all of the above structural units except for the carbonyl group. The carbonyl group was at first judged as an ester from the chemical shift, however, it gave a methyl ester 3 (m/z 447.21457; δ_H 3.81 s; 1731 cm⁻¹, Figure S6) on treatment with TMSCHN₂, confirming it as a carboxylic acid. Both the ¹H and ¹³C-NMR signals (δ_H 7.96; δ_C 147.7) for C-25 shifted more downfield than other furan α-positions such as C-21 (δ_H 7.21; δ_C 138.9), likely caused by conjugation with the carbonyl group. These chemical shifts are also comparable to those of β-carboxylic furans [14]. The double bond geometry at C-6 and C-10 was assigned as E by the chemical shifts of the vinyl methyls (δ_C 15.9 (C-24) and 16.7 (C-23)) and nuclear Overhauser effects (NOEs) (H-5/H-7, H-9/H-11, H-23/H-12), while the configuration at C-14 was determined as Z with an NOE (H-13/H-15). With all the NMR signals assigned, the planar structure was determined, as depicted in Figure 2 and

Table 1. ¹³C and ¹H-NMR data of compounds 1 and 2 in CDCl₃.

No.	1				2
	13C		1H (J in Hz)		13C		1H (J in Hz)
1	167.0	s	-	-	166.5	s	-	-
2	119.0	s	-	-	128.2	s	-	-
3	106.0	d	-	6.38 brs	140.0	d	-	6.63 brd (6.7)
4	156.4	s	-	-	23.8	t	a	2.34 m
-	-	-	-	-	-	-	b	2.21 ddd (3.1, 6.7, 17.7)
5	38.1	t	-	3.28 s (2H)	81.9	d	-	4.30 dd (3.1, 12.9)
6	130.9	s	-	-	38.9	s	-	-
7	127.2	d	-	5.24 m	31.8	t	α	1.56 m
-	-	-	-	-	-	-	β	1.30 m
8	26.3	t	-	2.16 m (2H)	17.4	t	-	1.51 m (2H)
9	39.2	t	-	2.06 m (2H)	37.8	t	α	0.97 dt (3.1, 12.6)
-	-	-	-	-	-	-	β	1.67 m
10	139.7	s	-	-	38.9	s	-	-
11	124.9	d	-	5.24 m	46.6	d	-	1.73 dd (2.6, 11.3)
12	75.9	d	-	4.56 dt (8.4, 5.9)	17.5	t	α	1.55 m
-	-	-	-	-	-	-	β	1.47 m
13	39.4	t	a	2.58 dd (5.9, 15.0)	41.6	t	α	1.65 m
-	-	-	b	2.23 m	-	-	β	1.70 m
14	139.8	s	-	-	73.0	s	-	-
15	119.2	d	-	5.33 m	58.8	d	-	1.05 brs
16	29.9	t	-	2.16 m (2H)	26.0	t	a	1.68 m
-	-	-	-	-	-	-	b	1.54 m
17	24.6	t	-	2.48 t (7.5) (2H)	28.4	t	a	2.44 m
-	-	-	-	-	-	-	b	2.39 m
18	124.5	s	-	-	125.4	s	-	-
19	110.9	d	-	6.26 brs	110.9	d	-	6.29 brs
20	142.7	d	-	7.34 brs	142.6	d	-	7.35 brt (1.5)
21	138.9	d	-	7.21 brs	138.6	d	-	7.25 brs
22	68.3	t	a	4.39 d (13.0)	30.5	q	-	1.18 s
-	-	-	b	4.20 d (13.0)	-	-	-	-
23	16.7	q	-	1.70 brs	15.9	q	-	1.03 s
24	15.9	q	-	1.60 brs	17.4	q	-	0.87 s
25	147.7	d	-	7.96 brs	16.9	q	-	1.92 s

.

To elucidate the absolute configuration, we calculated the specific rotation values of (12S)-1 and (12R)-1 with molecular dynamics (MD) simulations and time-dependent density functional theory (TDDFT). After generating 5000 structures from a 500-ns MD trajectory for each chiral molecule in CH₂Cl₂ solutions, we calculated the specific rotation values with the TDDFT method. The resulting averaged optical rotation values of (12S)-1 and (12R)-1 were calculated as −9.6 and +9.6 (Figure 3). Since the natural 1 presented a specific rotation value of [α]_D +11, it is likely to have 12R configuration.

Compound 2 showed a sodiated ion at m/z 423.25046 (∆ − 1.6 ppm) in high resolution electrospray ionization mass spectrometry (HRESIMS), establishing the molecular formula C₂₅H₃₆O₄ with eight degrees of unsaturation. As in compound 1, a β-substituted furan moiety was deduced from the signals for α-positions at δ_H 7.35, 7.25; δ_C 142.6, 138.6 and for β-positions at δ_H 6.29; δ_C 125.4, 110.9 (Figures S7 and S8). The presence of an α,β-unsaturated δ-lactone moiety was inferred by observing an ester carbonyl (1705 cm⁻¹; δ_H 166.5 s, C-1), conjugated olefinic signals (δ_H 6.63 brd; δ_C 128.2, 140.0, C-2,3), allylic methylene protons at δ_H 2.34 m and 2.21 ddd (H-4a,4b), and an oxymethine signal at δ_H 4.30; δ_C 81.9 (C-5). The lactone moiety was also confirmed by HMBC correlations (Figure S10) from a vinyl methyl at δ_H 1.92 s (H-25) to C-1,2,3 (δ_C 166.5, 128.2, 140.0) and from the vinyl proton H-3 to C-1,4,5,25 and COSY cross peaks (H-5/H-4a,4b and H-3/H-4a,25, Figure S11) and by a comparison of NMR data with those of schisanlactone B and colossolactone G [15,16]. The remaining two degrees of unsaturation are assignable to two additional rings.

By analyzing the HMBC correlations from three singlet methyls at δ_H 1.18 (H-22), 1.03 (H-23), and 0.87 (H-24), it was straightforward to construct a partial structure starting from C-24 via C-6,7,11,10,9,15,14,13 to C-22. The two remaining methylenes at C-8 (δ_C 17.4) and C-12 (δ_C 17.5) were involved in the bicyclic portion by Heteronuclear Single Quantum Correlation-Total Correlation Spectroscopy (HSQC-TOCSY, Figure S12) correlations (H-7α C-8,9, H-11/C-12,13). Therefore, the decaline ring was confirmed. Two allylic methylene protons at δ_H 2.44 and 2.39 (H-17a,17b) were assigned next to the furan by observing HMBC correlations to furan carbons (H-17a,17b/C-18,19,21) and another methylene at C-16 (H-16/C-14,15,17,18). Since the δ-lactone moiety was confirmed to be attached to C-6 by observing an HMBC cross peak (H-24/C-5), the whole planar structure of compound 2 was determined, as depicted in Figure 4.

The absolute configuration of the δ-lactone was determined as R on the basis of its positive Cotton effect at 253 nm (∆ε + 27.6, Figure S13), which sign is the same as that of (+)-parasorbic acid with the same δ-lactone moiety [17]. The bicyclic portion takes the trans-decaline ring system with chair conformation by observing NOEs for H-15/H-11,22 and H-23/H-24 (Figure S14). Therefore, two candidate structures, 2a and 2b, are possible for the absolute configuration of compound 2. In both cases, there is steric hindrance between the methylene protons at C-4 and at C-12. In fact, we could not observe any NOE between the methylene protons, confirming the restricted rotation between C-5/C-6. The DFT calculation also showed that one conformer is dominant with respect to the rotation between C-5/C-6 (Figure 5). The observed NOE between H-4b/H-24 and the absence of NOEs between H-4a,4b/H-11 support the structure 2a, but not 2b. In conclusion, the configurations of compound 2 were elucidated as 5R, 6R, 10S, 11R, 14S, and 15R.

2.2. Biological Activity

Compounds 1–3 were tested against cultured NBT-T2 cells by MTT assay. Compounds 1 and 2 showed cytotoxicity at IC₅₀ 47.5 and 1.0 µM, while compound 3 did not.

3. Conclusions

Two new furanosesterterpenoids 1 and 2 were isolated from the Okinawan sponge L. variabillis, and their absolute stereochemistry was elucidated with spectroscopic analyses and calculation. Of the two molecules, compound 2 may be the first member with this skeleton. Specimens of L. variabilis collected at shallower coral reefs in Okinawa usually contain manaoalide as the major constituent, while the current specimen collected from a deeper reef at 60 m contained new sesterterpenoids without manoalide. We have previously encountered metabolite variation with depth [18], and the current work may be another example of this.

4. Materials and Methods

4.1. General Experimental Procedure

NMR spectra were recorded on a Bruker AVANCE III 500 spectrometer (Billerica, MA, USA). ESI mass spectra were taken on a Jeol JMS-T 100LP mass spectrometer (Tokyo, Japan). HPLC separation was done on a unit equipped with a Shimadzu LC-10AD pump (Kyoto, Japan), a Shimadzu SPD-10A UV detector, a Shodex RI-101 refractive index detector (Tokyo, Japan), and a Nacalai Cosmosil 5SL-II column (4.6 × 250 mm) (Kyoto, Japan). Specific rotation was observed on a Jasco P-1010 polarimeter (Tokyo, Japan). FTIR spectrum was measured on a Jasco FT/IR-300 spectrophotometer. ECD spectrum was obtained on a Jasco J-820 spectropolarimeter.

4.2. Specimen

The specimen was collected by hand using a trimix rebreather at a depth of 60 m from a reef near Iriomote Island, Okinawa, in July 2013, and kept frozen until extraction. The specimen was identified as Luffariella variabilis by one of the authors (NJdV) and deposited at the Naturalis Biodiversity Center with the code RMNH POR 8677.

4.3. Extraction and Isolation

After thawing, the sponge (57 g, wet) was cut and macerated ca. 12 h in acetone (150 mL), three times. The acetone solution was filtered and concentrated under vacuum, and the residue was partitioned between EtOAc and H₂O. The organic layer gave an extract (770 mg) after concentration. A portion (260 mg) of the extract was separated on a Merck silica gel 60 (63–200 µm) column using solvents stepwise, namely, n-hexane, n-hexane-EtOAc (9-1, 7-3, 1-1), EtOAc, EtOAc-MeOH (1-1), and MeOH, to give eight fractions. A portion (18.2 mg) of the seventh fraction (54.4 mg) was passed through a silica Sep-Pak^® 3 cc Vac cartridge (200 mg) with EtOAc to remove polar impurities, and its first fraction (14.6 mg) was subjected to HPLC (Cosmosil 5SL-II, n-hexane-EtOAc, 2-3) to give compound 1 (2.0 mg, 3.3% from the extract). A further amount of compound 1 was obtained with similar treatment. A portion (12.7 mg) of the fourth fraction was purified via HPLC (Cosmosil 5SL-II, n-hexane-EtOAc, 4-1) to give compound 2 (7.0 mg, 6.5%).

Compound 1: Glass. ¹H and ¹³C-NMR (CDCl₃) see Table 1. HRESIMS m/z 433.19909, calcd. for C₂₅H₃₀O₅Na 433.19854 (∆ + 1.2 ppm). FTIR (neat) 2923, 1715, 1695, 1549, 1295, 1198, 767 cm⁻¹. [α] + 11 (c 0.14, CH₂Cl₂).

Compound 2: Colorless oil. ¹H and ¹³C-NMR (CDCl₃) see Table 1. HRESIMS m/z 423.25046, calcd. for C₂₅H₃₆O₄Na 423.25113 (∆ − 1.6 ppm). FTIR (neat) 3498, 2928, 1705, 1243, 1136, 781 cm⁻¹. ECD (MeOH) nm (∆ε) 253 nm (+27.6). [α] + 68 (c 0.18, MeOH). UV (MeOH) 212 nm (logε 4.1).

4.4. Methylation of Compound 1 to 3

Eight drops of a solution of TMSCHN₂ in hexane were added to compound 1 (0.2 mg) in MeOH (200 µL). After allowing the mixture to stand for 5 min at room temperature, it was concentrated with N₂ flow to give 0.2 mg of the ester 3.

Compound 3: HRESIMS m/z 447.21457, calcd. for C₂₆H₃₂O₅Na 447.21419 (∆ +0.8 ppm). FTIR (neat) 1731 cm⁻¹. ¹H-NMR (CDCl₃) δ 7.88 (1H, s, H-25), 7.34 (1H, bs, H-20), 7.21 (1H, bs, H-21), 6.35 (1H, bs, H-3), 6.27 (1H, bs, H-19), 5.33 (1H, m, H-15), 5.23 (1H, m, H-7), 5.23 (1H, m, H-11), 4.55 (1H, dt, J = 5.8, 8.7 Hz, H-12), 4.38 (1H, d, J = 13.3 Hz, H-22a), 4.20 (1h, d, J = 13.3 Hz, H-22b), 3.81 (3H, s, OMe), 3.28 (2H, s, H-5), 2.58 (1H, dd, J = 5.5, 15.3 Hz, H-13a), 2.48 (2H, t, J = 7.4 Hz, H-17), 2.22 (1H, m, H-13b), 2.16 (2H, m, H-8), 2.16 (2H, m, H-16), 2.05 (2H, m, H-9), 1.70 (3H, s, H-23), 1.60 (3H, s, H-24).

4.5. Cytotoxicity Assay

NBT-T2 cells were incubated at 37 °C under 5% CO₂ for 24 h in a 96-well plate containing 100 µL of Dulbecco’s Modified Eagle’s Medium (DMEM). An aliquot (1 µL) of DMSO solution of compounds 1, 2, and 3 was dispensed into each well in triplicate for each concentration and the cells were incubated for 48 h. Doxorubicin was used as a positive control, showing cytotoxicity at IC₅₀ 0.5 µM. A 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) solution was prepared by dissolving 5 mg of MTT in 1 mL of phosphate-buffered saline (PBS), and the solution was added to 10 mL of DMEM. After removing the culture media and dispensing 100 µL of the MTT solution, the wells were incubated for additional 3 h. The solution was then removed and 100 µL of DMSO was added to each well. The absorbance of each well was measured at 570 nm using a microplate reader. IC₅₀ values were obtained by analyzing the data on Kaleida Graph software (version 4.1, Synergy Software, Reading, PA, USA).

4.6. Calculation of Specific Rotation Value

We first performed the MD simulation to sample various conformations of compound 1 in CH₂Cl₂ solution. The system consisted of one compound 1 and 3375 CH₂Cl₂ molecules. A cubic unit cell with the periodic boundary condition was employed, using a box length of 72.5 Å. The general Amber force field (GAFF) [19] was used for compound 1 and CH₂Cl₂. Long-range electrostatic interactions were treated with the particle mesh Ewald method. Bonds involving hydrogen were constrained by using the SHAKE method. The equations of motion were integrated using the leapfrog algorithm with a time step of 1 fs at a temperature of 300 K. The MD simulation was performed for 500 ns. Next, 5000 geometries of compound 1 taken from the MD trajectory were optimized with the DFT method at the IEFPCM(CH₂Cl₂)/B3LYP/6-31G(d) level. Finally, specific rotation values were calculated with the TDDFT method at the IEFPCM(CH₂Cl₂)/B3LYP/aug-cc-pVDZ level [20]. We performed this procedure for both 12S and 12R configurations. The distributions of specific rotation values for (12S)-1 and (12R)-1 were almost antisymmetric (Figure 1), indicating that the number of samples was sufficient. The specific rotation values averaged over 5000 samples were −12.6 and +6.5 for (12S)-1 and (12R)-1, respectively. Further averaging the values over the two configurations, the values of (12S)-1 and (12R)-1 were −9.6 and +9.6, respectively. The MD simulations and TDDFT calculations were performed using the AMBER 12 [21] and Gaussian 09 [22] program packages at the Research Center for Computational Science, Okazaki, Japan.