Ulapualides C-E Isolated from a Hawaiian Hexabranchus sanguineus Egg Mass

Abstract

Three new ulapualides (3–5) were isolated from egg masses of the nudibranch Hexabranchus sanguineus. The structures of 3–5 were deduced by analyses of physical and spectroscopic data in comparisons with ulapualides A (1) and B (2). Ulapualide C demonstrated submicromolar cytotoxicity against select NCI cell lines (768-0, DU-145, MDA-MB-231, and A549) with the most potent activity against MDA-MB-231 (IC₅₀ 0.58 μM). Ulapualides A (1) and B (2) were two to four-fold more potent than 3.

Graphical Abstract

Originally isolated in 1986 from egg masses of the nudibranch Hexabranchus sanguineus, the ulapualides¹ derive their Hawaiian name from the flowerlike appearance of the Spanish dancer’s eggs underwater (ula=red and pua=flower). The ulapualides contain a characteristic tris-oxazole moiety and a long aliphatic tail similar to the mycalolides² and kabiramides.³ These compounds exhibit strong cytotoxicity which is proposed to protect the egg masses from predation. The cytotoxicity arises from the conserved aliphatic tail region that irreversibly binds to G-actin and behaves as a molecular mimic of the actin severing protein gelsolin.⁴ Taking advantage of this strong affinity allowed co-crystallization with G-actin that provided the absolute configuration of ulapualide A (1) via X-ray crystallography.^5,6

As part of our ongoing search for BACE1 inhibitors, preliminary data suggested the ulapualides in crude extracts bound to that enzyme. After purification, however, ulapualides A (1) and B (2) showed no significant inhibition of BACE1 when tested at concentrations up to 30 μM. Further chemical investigation of the extract led to the isolation and identification of three new analogs, ulapualides C-E (3–5), described within.

RESULTS AND DISCUSSION

Two egg masses from the west shore of Oahu (Electric Beach) were extracted with 1:1 MeOH:CH₂Cl₂ overnight three times. The resultant extract was dry loaded on C8 silica gel and subjected to a stepped gradient of MeOH:H₂O. The 75% MeOH fraction was separated by reversed-phase HPLC leading to the isolation of ulapualides A and B (1–2), along with ulapualide C (3) and a mixture of two other analogs. Subjecting this mixture to further HPLC using a different eluent afforded ulapualides D (4) and E (5). ¹H and ¹³C NMR data were collected on all isolated compounds (Tables 1 and 2, respectively), but insufficient sample was available for direct detection of ¹³C NMR for 4 and 5. Consequently, ¹³C chemical shifts were obtained from indirect HSQC and HMBC experiments for these two compounds.⁷

Table 1.

¹H NMR Chemical Shifts of Ulapualides C-E (3–5) in CDCl₃

position	Ulapualide C (3)	Ulapualide D (4)	Ulapualide E (5)
	δH (J in Hz)a [minor]b	δH (J in Hz)a [minor]b	δH (J in Hz)a [minor]b
2	2.45, m	2.49, m	2.46, m
3	4.23, septet (4.1)	4.17, m	4.23, septet (4.1)
4	1.65, m 1.55, m	1.72, m	1.72, m 1.60, m
5	1.95, m 1.79, m	1.25, m	1.96, m 1.76, m
6	2.55, m	1.76, m 1.41, m	2.57, m
7		4.07, br t (8.4)
8	3.10, m 2.51, dd (4.9, 16.3)	1.76, m 1.71, m	3.01, m 2.51, dd (15.5, 3.7)
9	3.41, m	2.98, m	3.42, m
9-Me	1.31, d (7.0)	1.32, d (7.0)	1.32, d (6.9)
11	7.41, d (1.0)	7.40, d (1.3)	7.40, s
14	8.06, s	8.08, s	8.06, s
17	8.06, s	8.06, s	8.06, s
19	6.38, d (16.0)	6.31, d (15.7)	6.39, d (16.5)
20	7.05, dt (16.0, 7.4)	7.09, ddd (15.3, 8.9, 6.0)	7.01, m
21	2.66, ddt (14.4, 4.6, 1.8) 2.44, m	2.70, m 2.36, dd (14.3, 8.2)	2.65, br d (14.5) 2.42, m
22	3.41, m	3.66, t (6.1)	3.40, m
22-OMe	3.37, s	3.38, s	3.37, s
23	1.85, m	1.61, t (7.1)	1.81, m
23-Me	0.91, d (7.0)	0.88, d (6.9)	0.87, d (6.9)
24	5.24, t (8.2)	5.26, t (9.6)	5.23, m
25	1.54, m	1.69, m 1.41, m	1.49, m
26	2.95, br d (9.4)	2.98, m	2.93, dt (8.8, 3.3)
26-OMe	3.30, s	3.31, s	3.28, s
27	1.71, m	1.70, m	1.72, m
27-Me	0.832, d (6.7) [0.828, d]	0.792, d (6.7) [0.788, d]	0.83, d (6.9) [0.82, d]
28	1.65, m 0.90, m	1.70, m 1.25, m	0.94, m
29	1.25, br s	2.44, m	1.57, m
30	3.44, m		5.10, t (7.2)
31	1.54, m	2.77, dd (8.8, 7.3)	1.81, m
31-Me	0.85, d (7.0) [0.84, d]	1.06, d (7.1) [1.05, d]	0.954, d (6.9) [0.947, d]
32	4.80, td (9.8, 2.9)	5.12, dd (8.9, 3.9) [5.13 dd]	4.75 dd (10.4, 2.4)
32-OAc	2.16, s [2.15, s]	2.00, s [1.99, s]	2.11, s [2.10, s]
33	2.56, m	2.48, m	2.53, m
33-Me	1.05, d (6.8) [1.04, d]	1.04, d (7.1) [1.03, d]	1.00, d (5.9) [0.99, d]
34	5.00, dd (14.1, 9.4)	4.98, td (13.6, 9.4)	4.96, dt (14., 9.6)
35	6.50, d (14.3)	6.49, d (14.0)	6.50, d (14.9)
35-NMe	3.02, s [3.06, s]	3.02, s [3.07, s]	3.02, s [3.06, s]
35-NCHO	8.29, s [8.07, s]	8.28, s [8.07, s]	8.29, s [8.07, s]
37			3.81, t (4.8)
37-OMe			3.45, s
38			3.86, br s

^a500 MHz.

^bProton signals in brackets are for the minor rotamer.

Table 2.

¹³C NMR Chemical Shifts of Ulapualides A-E (1–5) in CDCl₃

position	1		2		3		4		5
	δCa,b	type	δCa,b	type	δCa,b	type	δCa,c	type	δCa,c	type
1	172.5	C	172.5	C	172.5	C	172.6	C	172.4	C
2	43.0	CH2	43.0	CH2	43.0	CH2	43.4	CH2	42.8	CH2
3	68.1	CH	68.1	CH	68.0	CH	69.0	CH	68.3	CH
4	37.1	CH2	37.1	CH2	37.1	CH2	37.5	CH2	37.0	CH2
5	20.8	CH2	20.7	CH2	20.8	CH2	29.8	CH2	20.5	CH2
6	43.8	CH2	43.7	CH2	43.8	CH2	36.8	CH2	43.7	CH2
7	210.5	C	210.4	C	210.5	C	71.2	CH	210.0	C
8	47.8	CH2	47.7	CH2	47.7	CH2	45.4	CH2	47.8	CH2
9	27.4	CH	27.3	CH	27.3	CH	31.1	CH	27.4	CH
9-Me	19.6	CH3	19.6	CH3	19.7	CH3	19.3	CH3	19.4	CH3
10	146.3	C	146.3	C	146.3	C	146.8	C	146.1	C
11	133.5	CH	133.5	CH	133.5	CH	133.8	CH	133.2	CH
12	154.1	C	154.1	C	154.1	C	154.4	C	153.8	C
13	130.1	C	130.0	C	129.8	C	NDd		NDd
14	137.3	CH	137.3	CH	137.3	CH	137.0	CH	NDd
15	156.1	C	156.0	C	156.1	C	156.2	C	155.8	C
16	131.5	C	131.4	C	131.5	C	NDd		NDd
17	137.7	CH	137.7	CH	137.7	CH	137.3	CH	NDd
18	162.7	C	162.7	C	162.7	C	162.9	C	162.4	C
19	116.9	CH	116.8	CH	116.8	CH	115.9	CH	116.9	CH
20	139.9	CH	140.0	CH	140.1	CH	141.4	CH	139.7	CH
21	33.8	CH2	33.9	CH2	34.0	CH2	35.0	CH2	33.4	CH2
22	79.9	CH	79.9	CH	79.9	CH	78.3	CH	79.6	CH
22-OMe	57.7	CH3	57.7	CH3	57.8	CH3	58.1	CH3	57.6	CH3
23	40.3	CH	40.3	CH	40.4	CH	41.0	CH	40.1	CH
23-Me	9.0	CH3	8.9	CH3	9.0	CH3	9.0	CH3	9.0	CH3
24	72.8	CH	72.9	CH	72.9	CH	72.6	CH	72.6	CH
25	31.9	CH2	31.8	CH2	32.1	CH2	33.6	CH2	31.6	CH2
26	81.7	CH	81.6	CH	81.1	CH	81.6	CH	81.4	CH
26-OMe	58.0	CH3	57.9	CH3	58.1	CH3	58.0	CH3	57.8	CH3
27	34.4	CH	34.7	CH	35.5	CH	34.5	CH	34.4	CH
27-Me	15.4	CH3	15.5	CH3	15.4	CH3	15.7	CH3	15.6	CH3
28	25.0	CH2	27.0	CH2	28.4	CH2	25.2	CH2	26.6	CH2
29	39.8	CH2	30.4	CH2	29.7	CH2	39.8	CH2	30.3	CH2
30	211.6	C	73.03	CH	70.0	CH	211.6	C	72.7	CH
31	48.6	CH	37.5	CH	39.7	CH	48.6	CH	37.2	CH
31-Me	13.3	CH3	9.7	CH3	8.5	CH3	13.6	CH3	9.1	CH3
32	77.2	CH	79.8	CH	79.3	CH	77.2	CH	76.4	CH
32-OAc	170.0	C	170.6	C	172.5	C	170.0	C	171.1	C
	20.8	CH3	20.9	CH3	20.9	CH3	20.9	CH3	20.8	CH3
33	36.8	CH	36.9	CH	36.5	CH	36.9	CH	36.7	CH
33-Me	18.8	CH3	19.3	CH3	19.4	CH3	19.0	CH3	19.4	CH3
34	110.4	CH	110.2	CH	110.0	CH	110.5	CH	109.5	CH
35	129.4	CH	129.4	CH	129.4	CH	129.4	CH	129.4	CH
35-NMe	27.5	CH3	27.5	CH3	27.5	CH3	27.5	CH3	27.5	CH3
35-NCHO	162.1	CH	162.1	CH	162.1	CH	162.1	CH	161.9	CH
36			170.2	C					170.3	C
37			80.6	CH					81.4	CH
37-OMe			58.6	CH3					58.2	CH3
38			72.98	CH2					62.9	CH2
38-OMe			59.3	CH3

^aCarbon spectra recorded at 125 MHz.

^bResonances for minor rotamers can be found in the supporting information.

^cCarbon resonances extrapolated from HSQC and HMBC experiments.

^dNot detected.

All new compounds provided ¹H NMR spectra consistent with tris-oxazole macrolides. These characteristic signals included the presence of a minor rotamer in an approximate 2:1 ratio evidenced by doubling of the signals for the N-methyl group (δ_H 3.02 and 3.06; 35-NMe) and formamide proton (δ_H 8.29 and 8.07; 35-NCHO). This doubling arises from the slow interconversion of the s-cis- and s-trans-rotamers of tertiary amide and has been documented in molecules containing similar tail regions including tolytoxin,⁸ scytophycins,⁹ luminaolide,¹⁰ and rhizapodin.¹¹ Additionally, for compounds 3–5, the deshielded ¹H NMR signals for H-11, H-14, and H-17 (Table 1) were consistent with the three contiguous 2,4-disubstituted oxazole rings. Co-isolation of 1–2 from this egg mass supported the conclusion that 3–5 were ulapualide analogs. In addition, 1–5 contained the same UV chromophore with a maximum at 249 nm and produced a consistent pattern of ESI-MS cations including [M+H-H₂O]⁺, [M+H]⁺, [M+NH₄]⁺, and [M+Na]⁺ where the [M+H]⁺ were the least abundant.

High resolution mass spectrometry of 3 produced a protonated molecule at m/z 883.4700 [M+H]⁺. This datum is consistent with a molecular formula of C₄₆H₆₆N₄O₁₃ for the parent molecule. Comparison to the ¹H NMR spectrum of 1 indicated 3 contained two more protons, while the ¹³C NMR spectrum of 3 contained an additional sp³ oxygenated carbon δ_C 70.0 (C-30; 3) at the expense of a keto group at δ_C 211.6 (C-30; 1). HMBC correlations from the methyl group at position 31 to C-30 confirm the keto group on the aliphatic tail of 1 was reduced rather than the one in the macrolide ring.

High resolution mass spectrometry of 4 provided a protonated molecule at m/z 883.4662 [M+H]⁺ as well. Comparison between the NMR data of 4 and 1 showed an additional oxygenated sp³ carbon δ_C 71.2 (C-7; 4) and a corresponding proton signal at δ_H 4.07 (H-7; 4) again at the expense of a carbonyl resonance δ_C 210.5 (C-7; 1). COSY correlations from the methyl group at position 9 to H-9 and from H₂-8 to both H-9 and H-7 support the conclusion that C-7 was a secondary alcohol in 4.

High resolution mass spectrometry of 5 gave a protonated molecule at m/z 985.4976 [M+H]⁺ consistent with the molecular formula (C₅₀H₇₂N₄O₁₆) 14 amu smaller than 2. Comparison of the ¹³C and ¹H NMR spectra of 5 with 2 indicated that 5 contained one less methoxy group. Taken together, the data suggested 5 contained a hydroxy rather than methoxy group. HMBC correlations from the existing OMe groups in 5 pinpoint the placement of the methoxy groups and through deduction that C-38 was a primary alcohol.

The relative and absolute configurations of 3–5 could not be determined due to the limited amount of material available. In mass limited cases, configurational assignment based on carbon chemical shift comparisons with known analogs is a common tactic. In this case, synthetic studies have demonstrated that this approach is unreliable for distinguishing ulapualide diasteromers.¹² However, all of the tris-oxazole compounds characterized by X-ray crystallography have the same absolute configuration (with the omission of position 9),¹³ this suggests that those stereocenters are conserved via a shared biogenesis. In addition, the two other stereocenters in 5 are most likely 30R,37R consistent with 2 and the structurally-related mycalolides.¹⁴

With the exception of the kabiramides, also isolated from nudibranch eggs, all other tris-oxazole containing molecules were isolated from sponges suggesting that these molecules may be sequestered through predation. The support for this hypothesis includes observed predation of Halichondria sp. by Hexabranchus sanguineus, isolation of dihydrohalichondramide from both species, and sequestration of dihydrohalichondramide in the mantle and the digestive system/gonad of the nudibranch.¹⁵ That similar tris-oxazole compounds have been isolated from sponges of dissimilar phylogeny suggests the ultimate source may be microbial as was found with the bryostatins, metabolites originally isolated from the bryozoan Bugula neritina¹⁶ and later discovered to be a product of the symbiont Endobugula sertula.¹⁷ The symbiont E. sertula is vertically transmitted from bryozoan adult to embryo¹⁸ and provides chemical defenses for the otherwise unprotected larvae.

Compound 3 could be an artifact as hydrolysis of the O-methylated glycerate in 2 would yield the free alcohol 3. That significant amounts of 3 can be observed via LCMS in the original MeOH:CH₂Cl₂ extract, before buffers were used suggests this is not the case. Furthermore, treatment of 2 in a 2:1 mixture of MeOH:H₂O with 0.1% formic acid for seven days yielded no hydrolysis product detectable by LCMS.

Compounds 1–3 were tested against select cancer cell lines within the NCI 60-cell panel. These included human renal cell adenocarcinoma (768-0), prostate carcinoma (DU-145), mammary gland adenocarcinoma (MDA-MB-231) and lung carcinoma (A549) cell lines. Samples were tested in triplicate in two independent trials. For the second independent experiment, the concentration was determined through colorimetry using a standard curve made from the more abundant 2. As expected, low to sub-micromolar IC₅₀ values were observed for all samples (Table 3). The activities of 1–2 were similar in all cell-lines tested, while 3 was two to four times less potent. Removal of methyl or methoxy groups from the tail region or the introduction of charged groups is predicted to decrease binding affinity to G-actin which is dominated by hydrophobic interactions in the cleft between actin subdomains 1 and 3.⁵ The observed decrease in activity could be attributed to replacement of the modified glycerate or keto group with a more polar hydroxy group.

Table 3.

IC₅₀ Values of Ulapualides A-C (1–3) Against Select NCI Cell Lines (μM)

	768-0		DU-145		MDA-MB-231		A549
	Trial 1	Trial 2	Trial 1	Trial 2	Trial 1	Trial 2	Trial 1	Trial 2
Ulapualide A (1)	0.28	0.26	0.26	0.25	0.25	0.23	0.28	0.29
Ulapualide B (2)	0.26	0.66	0.28	0.34	0.29	0.29	0.33	0.26
Ulapualide C (3)	1.1	1.4	0.76	0.85	0.58	0.76	0.68	0.60

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. 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 Varian Unity Inova 500 MHz spectrometer. NMR spectra were referenced to the appropriate residual solvent signal (δ_H 7.26, δ_C 77.0 for CDCl₃). 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 on a Agilent 6210 TOF LC/MS using the ESI source in positive mode.

Collection and Isolation

The egg masses of Hexabranchus sanguineus were collected via SCUBA off the west shore of O’ahu at Electric Beach (21°21′14.3″N 158°07′49.7″W) at a depth of 20 feet on September 23^rd, 2011. The two egg masses collected were bright pink and were immediately extracted with 1:1 MeOH:CH₂Cl₂ three times overnight to yield 2.7 g of extract. The combined extract was then dry loaded on C8 silica gel and was subjected to a solid phase extraction procedure consisting of five steps of increasing MeOH:H₂O content (0%, 25%, 50%, 75%, 100%, and an isopropanol wash). The 75% MeOH fraction was subjected to reversed-phase HPLC on a Phenomenex column (Luna C8; 250x10 mm, 5 μ) using a flow rate of 2.8 mL/min and a concentration gradient of 50%-70% (CH₃CN:H₂O) over 30 min. Detection was by UV, using a PDA and monitoring at 249 nm. This afforded pure compounds ulapualide B (t_R = 24 min, 8 mg, 0.32% yield), ulapualide A (t_R = 21 min, 6 mg, 0.24% yield; [α]²²_D-39, c 0.20, MeOH), ulapualide C (3, t_R = 18.5 min, 1.6 mg, 0.064% yield) and a mixture of ulapualides D and E. This mixture was further purified using a C18 column (Luna; 100x4.6 mm, 5 μ) at a flow of 1.0 mL/min and using an isocratic system of 65% MeOH-H₂O to yield ulapualide D (4, t_R =20 min, 0.5 mg, 0.02% yield) and ulapualide E (5, t_R =24 min, 0.5 mg, 0.02% yield). Purity was assessed at 210 nm and determined to be 95.3%, 94.5%, and 95.6% for 1, 2 and 3 respectively, prior to biological testing.

Ulapualide C (3)

white amorphous powder; [α]²²_D-17 (c 0.32, MeOH); UV (MeOH) λ_max (log ε) 249 (4.5), 237 (4.6), 202 (4.5); IR ν_max 3600-3200 (br), 3162, 2964, 2934, 1720, 1691, 1655, 1597, 1458, 1241, 1086, 916 cm^-1. Tables 1 and 2 for NMR data; HRESI-TOFMS m/z 883.4700 [M+H]⁺ (calcd for C₄₆H₆₇N₄O₁₃, 883.4705).

Ulapualide D (4)

white amorphous powder; [α]²²_D-10 (c 0.08, MeOH); UV (MeOH) λ_max (log ε) 249 (4), 213 (5), 202 (5); Tables 1 and 2 for NMR data; HRESI-TOFMS m/z 883.4662 [M+H]⁺ (calcd for C₄₆H₆₇N₄O₁₃, 883.4705).

Ulapualide E (5)

white amorphous powder; [α]²²_D-23 (c 0.10, MeOH); UV (MeOH) λ_max (log ε) 249 (4), 213 (5), 202 (5); Tables 1 and 2 for NMR data; HRESI-TOFMS m/z 985.4976 [M+H]⁺ (calcd for C₅₀H₇₃N₄O₁₆, 985.5022).

Cytotoxicity Assay of Ulapualide A, B, and C

Human renal cell adenocarcinoma (768-0), prostate carcinoma (DU-145), mammary gland adenocarcinoma (MDA-MB-231) and lung carcinoma (A549) cell lines were maintained in RPME 1640 Medium (Gibco, REF: 11875-093) supplemented with 10% premium fetal bovine serum (Atlanta biological, Cat. No.: S11150) and 100 U/mL penicillin and 100 μg/mL streptomycin (Gibco, REF: 15140-122). The day before treatment, cancer cells were seeded with 4,000 cells per well into a 96-well tissue culture plate. Twenty hours post seeding, the serial diluted compounds were added to the cells for the cytotoxicity assay, and co-incubated at 37 °C with 5% CO₂ for 72 h. Then the medium with compounds were replaced with 1 × dye binding solution prepared according to the manufacturer’s instruction (CyQuant NF Cell Proliferation Assay Kit [C35006, Invitrogen]) and incubated at 37 °C with 5% CO₂ for 60 minutes. After that, cell viability data were collected with a PerkinElmer Multimode Plate Reader according to the manufacturer’s instruction. IC₅₀ curves were generated using GraphPad Prism 5.