Myrmenaphthol A, Isolated From a Hawaiian Sponge of the Genus Myrmekioderma

Abstract

#Four compounds (1–4) were isolated from a Hawaiian sponge of the genus Myrmekioderma. Myrmenaphthol A (1) incorporates two unusual elements into an oxidized steroidal core: a naphthyl AB-ring system and a hydroxy group at C-2. Comparison of the experimental and predicted ECD spectra of 1 assigned an S configuration to the lone stereocenter (Δ_ESI = 0.75; similarity factor 0.8137). Known compounds, cinanthrenol A (2), 3,4-dihydroxypregna-5,17-diene-10,2-carbolactone (3), and 3,4-dihydroxypregna-5,20-diene-10,2-carbolactone (4) were also isolated. Despite literature reports of competitive inhibition at nM levels for 2, neither 2 nor the structurally related 1 showed any activity against estrogen receptors at the concentrations tested.

Graphical Abstract

Although there is large diversity of reported terpenoid and steroidal compounds isolated from marine invertebrates, few derivatives of the pregnane class of compounds have been reported^1–4 with the most recent being a set of pregnane-10,2-carbolactones isolated by our group.⁵ Oxidations of the A ring into phenol pregnane derivatives are even rarer, and lead to range of biological activities.^6,7 Initial screening of a Hawaiian sponge belonging to the genus Myrmekioderma, collected off the southern coast of Lanaʻi displayed moderate activity against Herpes Simplex Virus-1 (HSV-1) and strong activity against Vesicular Stomatitis Indiana virus (VSV) at 200 μg/mL. When extracted on a larger scale, all fractions failed to reproduce this activity. Therefore, a chemical investigation of the sponge was conducted resulting in the isolation of myrmenaphthol A (1) and known compounds 2-4.^8,5

Myrmenaphthol A (1) was isolated as an amorphous yellow powder after multiple rounds of chromatography of the 75% MeOH fraction. HRMS indicated the presence of a protonated molecule with m/z 307.1135. This datum is consistent with the molecular formula C₂₀H₁₈O₃, indicating 12 degrees of unsaturation. The¹³C NMR spectrum displayed 14 sp² non-carbonyl carbons and the HMBC spectrum displayed an additional sp² carbon with a chemical shift indicative of an α,β-unsaturated keto group (δ_C-16 190.9). Therefore, the molecule has seven double bonds, a keto group, and four rings to account for the degrees of unsaturation.

The structure of 1 was assembled by analyses of COSY and HMBC correlations. H-6 showed a COSY correlation to H-7 as well as HMBC correlations to C-4, −5, −7, −8, and −10 to help establish part of fragment 1. This fragment was completed on the basis of a COSY correlation from H-4 to H-3. A COSY correlation between H-11 and H-12, HMBC correlations from H-19 to C-13, −18, −17, and −16 and HMBC correlations from H₃-20 to C −12, −14, and −17 allowed the assembly of fragment 2.

Furthermore, HMBC correlations from H-1 to C-2, −3, −5 and −9 helped establish the phenol A ring while correlations from H-11 to C-8, −9, and −10 connected fragments 1 and 2 to form the naphthyl nucleus. Correlations from H-7 to C-14 extended the ring system to generate fragment 3. At this point, the only remaining pieces in our chemical inventory were a non-protonated carbon (δ_C 149.6, C-15), an oxygen, a hydrogen and one degree of unsaturation. Therefore, the only existing option was to insert an enol moiety between carbons 14 and 16 while closing the ring to create structure 1.

The relative configuration was determined through analysis of ROESY correlations. A ROESY correlation between H₃-20 and H-19 determined the configuration of the pregnane’s C-17 double bond to be E. Further ROESY correlations from H₃-20 to H-11b suggest a cis orientation and correlations from H₃-20 to H-12a but not H-12b suggest that the methyl group adopts a conformation in which it is pseudo-axial. This is supported by the MM2 minimized energy conformer shown in Figure 2.

Figure 2. Relative configuration of 1.

A) An NOE correlation from H₃-20 to H-19 is evidence that the configuration of the C-17 olefin is E. B) MM2 minimized energy conformer of the S enantiomer of 1 as seen looking down the carbonyl bond.

The absolute configuration of 1 was determined through ECD spectroscopy using TDDFT calculations, which have become a commonly used technique for determining absolute configuration of various natural products.⁹ Conformational analysis of 1, using Monte Carlo multiple minimization (MCMM) and the OPLS-2005 force field in MacroModel with subsequent optimization in Gaussian 09 at the M06–2X/6–31+G level,¹⁰ identified four conformers within 5 kcal/mol of the lowest energy conformer which differed only in the direction of the two oxygen-hydrogen bonds. Figure 2B shows one of the minimized energy conformers of the 13S-enantiomer looking down the carbonyl bond. The two conformers with the α-hydroxy proton in position to hydrogen bond to the carbonyl accounted for 99.7% of the conformers. The primary difference between these two conformers was the orientation of the O-H bond at C-2 (See Table S1). A comparison with the experimental NMR data revealed no obvious contradictions in terms of expected coupling constants or chemical shifts validating the major conformers. Time-dependent density functional theory (TDDFT)¹¹ calculations were performed using Gaussian09¹² at the B3LYP^13–16/6–31+G^{17, 18} and the B3LYP/aug-cc-pVDZ¹⁹ levels to predict ECD spectra for the conformers of the 13S enantiomer. Finally, the Boltzmann weighted ECD spectrum of these conformers, calculated using SpecDis,²⁰ was compared to the experimental data. These data (Figure 3) were in good agreement based on a Δ_ESI= 0.75²⁰ with similarity factors of 0.8137 and 0.0647 for the 13S- and 13R-enantiomers, respectively.

Figure 3:

Predicted and Observed ECD Spectra of 1.

In an effort to obtain more of 1 from adjacent fractions, inspection of the 50% MeOH fraction revealed multiple peaks with related chromophores at 280 nm. Collection of these peaks yielded 2 as a pure compound and both 3 and 4 as mixtures.

Compound 2 shared many of the same proton resonances as 1. Perhaps the greatest similarity was that 2 displayed resonances consistent with the 2-naphthol spin system. HRMS provided a protonated molecule with m/z of 291.1386 in agreement with the molecular formula of C₂₀H₁₈O₂ and 12 degrees of unsaturation. This coupled with two additional aromatic doublets facilitated our dereplication efforts to establish 2 as the known compound cinanthrenol A, previously isolated from a Cinachyrella sp sponge.⁸

. Based on HRMS and NMR data, 3 and 4 were determined to be 3,4-dihydroxypregna-5,20-diene-10,2-carbolactone (3) and 3,4-dihydroxypregna-5,17-diene-10,2-carbolactone (4) Compounds 3 and 4 had been previously isolated from Strongylophora sp. (Haplosclerida).⁵

Myrmenaphthol A (1) displays an uncommon naphthyl nucleus in a traditional steroid ring system, which is the first example of such a system isolated from a marine source. One of the few terrestrial examples of this moiety appears in equilenin,²¹ a potent estrogen receptor activator and a minor component of the FDA approved drug Premarin. A major difference between 1 and nearly all similar natural products lies in the position of the oxidation of the A ring. Equilenin, like many known biological hormones such as testosterone and estradiol, is oxidized at C-3 whereas 1 has oxidation on the 2 position. This rare oxidation of a steroid nucleus is only known to occur in cinanthrenol A, which was isolated simultaneously with 1.

Previous studies have shown 2 to have affinity for estrogen receptors. Cinanthrenol A (2)⁸ was formerly isolated from a sponge dredged from the depths of the East China Sea, Japan. The phenanthracene compound displayed cytotoxicity against P-388 and He-La cells with IC₅₀ values of 4.5 and 0.4 μg/mL, respectively.⁸ Its strongest affinity was to estrogen receptor (ER1) which displayed competitive inhibition against estradiol with an IC₅₀ of 10 nM⁸ in a radioactive receptor binding assay. Synthesis of ent-(+)-cinanthrenol A has recently been published confirming the absolute configuration of cinanthrenol A and providing a pathway to synthesizing structural analogues, but no biological testing was reported.²² Due to 1’s similar framework to 2, testing of 1 and 2 against ER1 was conducted in parallel. Surprisingly, no significant inhibition was displayed for either compound, with only 30% inhibition at 100 μM against the ER-α binding domain. Indeed, multiple rounds of testing in ER biochemical assays failed to show any significant activity in our hands for either 1 or 2 as agonist or antagonist despite the nM activity previously reported for 2.

EXPERIMENTAL SECTION:

General Experimental Procedures.

Optical rotations were measured on a Jasco DIP-370 digital polarimeter at the sodium line (589 nm). UV absorbances were measured on a Varian Cary 50 Bio UV-Vis Spectrophotometer. CD spectra were recorded on a Jasco J-815 CD spectrometer. IR spectroscopy was measured as a thin film on a CaF₂ disk using a Shimadzu IRAffinity-1 FTIR. ¹H, ¹³C NMR and 2D NMR experiments on the natural products were carried out on a Varian Unity Inova at 500 MHz spectrometer. NMR spectra were referenced to the appropriate residual solvent signal (δ_H 3.30, δ_C 49.0 for CD₃OD). The HSQC experiments were optimized for ¹J_C,H = 140 Hz and HMBC experiments for ³J_C,H = 7 Hz. High-resolution mass spectrometric data were obtained using the ESI source in positive mode on an Agilent 6545 QTOF. A binary gradient Shimadzu Prominence HPLC system with ELSD and PDA detectors were used for all separations.

Sponge Collection and Identification.

An undescribed species of Myrmekioderma (class Demospongiae, order Axinellida, family Heteroxyiidae) (450 g) was collected in 40–60 fsw off the southern shore of Lana’i at the 2^nd Cathedral (roughly 20°44’05.0”N, 156°55’25.7”W). The encrusting sponge has fingery projections and a smooth surface, produces clear mucus, cream colored in life and in spirit, and the interior has a pulpy consistency. The ectosomal skeleton is parchment-like and readily detachable, it consists of a tangential layer of a smaller category of strongyles (~265 × 9.8 μm). The choanosomal skeleton is largely confused with ascending tracts of larger strongyles (~555 × 9.5μm) among disorganized spicules including fine oxea/strongyles (~325 × 2.5μm) and abundant raphides in 2 sizes of trichodragmata (~52.5 and 80μm). All megascleres are smooth, and centrally curved or occasionally flexuous. The common Indo-Pacific species M. granulatum has larger megascleres and only one category of raphides. This specimen differs from another undescribed species of Myrmekioderma from Hawaii⁵ by producing mucus, having only smooth megascleres, and a smaller size second category of raphides.

Extraction and Isolation.

Lyophilized sponge (100g) was extracted with 1:1 MeOH:CH₂Cl₂ three times overnight to yield 5.77 g of organic extract. The combined extracts were then subjected to a modified Kupchan partitioning using only MeOH, hexanes, and CH₂Cl₂. The CH₂Cl₂ partition (177.5 mg) was dry loaded on C8 silica gel and was subjected to a solid phase extraction procedure consisting of four steps of increasing MeOH:H₂O content (0%, 25%, 50%, 75%, 100%), and an isopropanol wash. Both the 50% MeOH and the 75% MeOH fractions contained the compounds of interest. The 50% fraction was subjected to reversed-phase HPLC on a Phenomenex column (Luna C18; 250×10 mm, 5 μ) using a flow rate of 2.8 mL/min and a concentration gradient of 40%–80% (MeCN in H₂O) over 20 min. This afforded pure compounds 1 (t_R = 15.8 min, 0.6 mg, 0.01% yield) and 2 (t_R = 16.1 min, 1 mg, 0.02% yield), and 4 (t_R = 9 min, 2.2 mg, 0.04% yield) along with 3 (t_R = 8 min, 2.3 mg, 0.04% yield). Purity was assessed by NMR and determined to be 95%, 88%, 77% and 85% for 1, 2, 3, and 4, respectively.

Myrmenaphthol A (1): (0.6 mg, 0.01 % yield): yellow amorphous powder; [α]²²_D −79 (c 0.2, MeOH); ECD (0.05 mg/mL, MeOH), λ_max (Δε) 380 (−89.5), 333 (22.6), 288 (44.8), 262 (34.8), 242 (−69.7) and 217 (133.5) nm; UV (MeOH) λ_max (log ε) 362 (3.75), 240 (4.14), 203 (4.23) nm; IR (CaF₂) ν_max 3396, 1667, 1597, 1218 cm⁻¹; NMR data, See Table 1; HRESI-TOFMS m/z 307.1335 [M+H]⁺ (calcd for C₂₀H₁₉O₃, 307.1329).

Table 1:

NMR Spectroscopic Data for 1 (¹H 500 MHz, ¹³C 125 MHz, CD₃OD)

	1
Position	δC, Type	δHb (J in Hz)	COSY	HMBC (1H to 13C)	ROESY
1	106.8, CH	7.33, d (2.2)	H-3	2, 3, 5, 9	H2-11
2	157.1, C
3	119.5, CH	7.09, dd (8.8, 2.2)	H-1, H-4	2, 5	H-4
4	131.1, CH	7.70, d (8.8)	H-3	1, 2, 5, 6, 9, 10	H-3
5	129.6, C
6	127.2, CH	7.64, d (8.6)	H-7	4, 5, 7 , 8, 10	H-7
7	124.6, CH	8.07, d (8.6)	H-6	5, 8, 9, 10, 14	H-6
8	128.3, C
9	131.8, C
10	134.7, C
11a	24.8, CH2	3.32, m	H-12b	8, 9, 10, 12, 13	H-1, H-11
11b		3.22, m	H2-12	8, 9, 10, 12	H-1, H-11, H-20
12a	33.0, CH2	2.68, dd (13.0, 6.1)	H-11b, H-12b	9, 11, 13, 14	H-11b, H-12, H-19, H-20
12b		1.82, td (12.5, 6.4)	H2-11, H-12a	11, 13, 20	H-12a
13	40.7, C
14	143.1, C
15	149.6, C
16	190.9,a C
17	144.1, C
18	130.7, CH	6.63, q (7.5)	H-19	13, 16, 17, 19	H-19
19	14.8, CH3	2.05, d (7.5)	H-18	13, 16, 17, 18	H-12a, H-18, H-20
20	22.3, CH3	1.31, s		12, 13, 14	H-11b, H-12a, H-19

^aCarbon chemical shifts determined from the HMBC experiment

ER-alpha Assay.

The assay was conducted through Thermo Fisher Scientific’s SelectScreen biochemical nuclear receptor profiling service employing the LanthaScreen TR-FRET Coregulator assay and using 17-β-estradiol as the stimulant. In-house assessment of the ER inhibition was performed using Indigo Biosciences Human ERα, reporter assay system with the same positive control.

Computational Analysis.

Conformers within 5 kcal/mol of the lowest energy conformer were searched using the Monte Carlo multiple minimum (MCMM) method and the OPLS-2005 force field in MacroModel¹⁰ (Schrodinger Inc.). Each conformer within 5 kcal/mol of the lowest energy conformer was optimized in Gaussian09¹² at the M06–2X/6–31+G level and the geometries of all conformers with similar energies were checked for redundancy. Density functional theory (DFT) was used to perform calculations, which were carried out in Gaussian 09. All ground-state geometries were optimized at the B3LYP/6–31+G level. The same DFT level was employed to evaluate the effects of solvation in MeOH using the SCRF/PCM method.^{23, 24} TDDFT^25,11 calculations at the B3LYP^13–16/6–31+G^{17, 18} and the B3LYP/aug-cc-pVDZ¹⁹ levels were conducted to calculate the electronic excitation energies and rotational strengths in MeOH. A Boltzmann weighted ECD spectrum was calculated using SpecDis²⁰ (σ = 0.24eV at 5nm shift; scaling factor 0.2369) for comparison with the experimentally determined data recorded in methanol at 0.05 mg/mL.

Figure 1.

Key HMBC and COSY correlations of 1.