Isolation of Bastadin-6-O-Sulfate and Expedient Purifications of Bastadins-4, −5 and −6 from Extracts of Ianthella basta

Abstract

Bastadin-6–34-O-sulfate ester (8) was isolated from methanol extracts of Ianthella basta. The structure of 8 was characterized by analysis of MS and NMR data, and conversion through acid hydrolysis, to the parent compound, bastadin-6, which was identical by HpLC, MS and NMR with an authentic sample. An improved procedure for procurement of pure samples of bastadins-4, −5 and −6 is described.

Graphical Abstract

1. Introduction

Bastadins are group of over 20 highly brominated natural products, formally derived from bromotyrosine (1), dibromotyrosine (2) or their corresponding tyramines, and classified by two hypothetical parent skeletons: ‘bastarane’ (ia)^† and its constitutional isomer, ‘isobastarane’ (iia). The first seven bastadins including the ring opened symmetrical dimer, bastadin-3 (3), and macrodilactams, bastadins-4 (4), −5 (5) and −6 (6), were characterized over 35 years ago by Kazlauskus and coworkers from samples of the Australian marine sponge Ianthella basta, Pallas [1a,b] and identified with modest antibiotic properties. They gained new significance in 1995 with the finding that bastadin-5 (5) and bastadin-6 (6, Figure 1) are potent agonists for the release of Ca²⁺ ions from stores within the sarcoplasmic reticulum (SR) though modulation of heterotetrameric megaprotein (2.5 MDa) Ca²⁺ ion-channel, RyR-1.[2] The most potent agonist, bastadin-5 (5) (EC₅₀ = 2.3 μM), was shown to promote release of Ca²⁺ through modulation of RyR-1[3,4] – the last stage of excitation-contraction coupling in mammalian striated muscle – through alteration of the gating kinetics of channel opening and closing.[2]

Figure 1.

Structures of bromotyrosines (1, 2), bastadins −3 (3), −4 (4), −5 (5), −6 (6), −19 (7) and bastadin-6-O-34-sulfate (8).

Although high-resolution cryo-EM structures of the RyR-1 complex have been reported recently,[5] the binding site of 5 is as yet unknown.

Two ‘channelopathies’ - malignant hyperthermia (MH)[6] and central core disease (CCD)[7] - are manifestations of different single-point autosomally dominant inherited mutations of the same gene, RYR1 in chromosome 19q13.1, that encodes RyR-1. MH is a ‘cryptic’ mutation that reveals itself during general anesthesia with certain inhalation anesthetics. The resulting condition ‒ runaway hyperthermia ‒ can be fatal to the patient if not treated immediately. CCD is congenital myopathy that is often debilitating and can lead to delayed ambulation and reduced vital capacity.

preliminary structure activity relationships of the bastadins have been described.[8] While 5 and 6 are almost equipotent RyR-1 agonists, bastadin-4 (4, 5,6-dehydrobastadin-5) is less potent by order of magnitude, while the constitutional isomer of 5 ‒ bastadin-19 (5) with an isobastarane skeleton - is essentially inactive (IC₅₀ > 100 μM).[2,19] Nevertheless, 4 is valuable because it can be converted into 5 or the isotopically labeled probe, 5-d (and, ostensibly, 5-t) through selective cationic reduction (Et₃SiH or Et₃SiD, TFA) as we demonstrated earlier.[10]

Compounds 5 and 6, and synthetic analogs inspired by their structures,[8] may be useful probes in the study of RyR-1 mutations, expanding our understanding of the mechanism of RyR-1 channel gating and Ca²⁺ release, and the possible development of therapeutic agents for the treatment of Ca²⁺ channel-related myopathies.

However, contemporary investigations into the pharmacology of 5 and 6 are limited by supply; both compounds only occur as minor components along with complex mixtures of bastadins and their O-sulfate esters. The natural products are not separable on silica chromatography and are typically won through tedious, multistep reversed-phase chromatography.[11] Total syntheses of 5[12a] and 6[12b,c] have been reported, but they involve lengthy series of linear steps and the overall yields are typically low. Thus, practical sources of 5 and 6 are required.

in our investigations of sources of new bastadins from samples of I. basta collected at two locations in separated oceans ‒ Guam in the Pacific and Exmouth Gulf, Western Australia, in the Indian Ocean ‒we discovered a new member of the series: bastadin-6-O-sulfate ester (8) which is the subject of this report. In addition, we describe a streamlined procedure for rapid procurement of the two most useful RyR-1 modulators, 5 and 6 from I. basta, and an improved protocol for conversion of 4 to the highly valued 5.

2. Experimental

2.1. General methods

Optical rotations were measured on a JASCO P-2000 at the D-double emission line of Na°. UV-vis spectra were measured on a JASCO V-630 spectrometer. FTIR spectra were collected on thin-film samples using a JASCO FTIR-4100 fitted with an ATR accessory (ZnSe plate). 1D NMR and inverse-detected 2D NMR spectra were measured on a Bruker Avance II NMR spectrometer with a 1.7 mm¹H{¹³C/¹⁵N} 600 MHz microcryoprobe. Other NMR spectra were measured on a JEOL ECA spectrometer equipped with a 5 mm¹H{¹³C} 500 MHz room-temperature probe.¹³C NMR spectra were measured using a Varian NMR spectrometer equipped with a 5 mm Xsens¹³C{¹H} 125 MHz cryoprobe. NMR spectra are referenced to residual solvent signals. High-resolution ESITOF analysis was carried out on an Agilent 1200 HPLC coupled to an Agilent 6230 TOFMS. Low-resolution MALDI MS measurements were made on a Bruker Biflex IV in a nitrobenzyl alcohol matrix. Low-resolution MS measurements were made on a Thermoelectron Accela UHPLC coupled to an MSQ single-quadrupole detector. Preparative, semi­preparative, and analytical HPLC were performed on a JASCO PU-2086 Plus system consisting of a dynamic mixer (MX-2080–32) with UV-VIS detector (UV-2075) operating at λ 250 nm. Automated flash chromatography was carried out with a Teledyne-Isco CombiFlash system with UV (λ 250 nm) and RI detection.

2.2. Animal Material

Two collections of the sponge Ianthella basta were made using scuba ‒ one in 2009 from Guam (09-IBC, sample courtesy of Peter Schupp, University of Oldenburg) and a second in 1993 from Bennett Shoal, Exmouth Gulf, Western Australia (93-07-101). The samples were stored at −20 °C until required.

2.3. Expedient Purification of Bastadins-5 and −6.

A typical isolation procedure is described here for Ianthella basta. Frozen lyophilized sponge (93-07-101, 92 g, dry wt.) was cut into smaller pieces (~5–10 cm) and extracted by slow stirring with CH₃OH at room temperature (2 × 600 mL, 12 h), and the combined extracts concentrated under reduced to half of the volume. The H₂O content was adjusted to 1:9 H₂O-CH₃OH, and the extract repeatedly partitioned against hexanes (2 × 400 mL) to give, after removal of solvent, ‘fraction A’ (1.41 g). The water content was re-adjusted to 2:3 H₂OCH₃OH and the solution partitioned against CH₂Cl₂ (2 × 400 mL) to give, after removal of solvent, ‘fraction B’ (2.49 g). MeOH was removed from the remaining aqueous-MeOH partition, under reduced pressure, and the remaining H₂O solution partitioned against n-BuOH (2 × 100 mL) to give ‘fraction C’ (600 mg). The residue of the aqueous phase constituted ‘fraction D’ (9.8 g).

Fraction B was purified by size-exclusion chromatography (Sephadex LH-20) with elution by MeOH to give six fractions (F1–6). grouped by TLC (silica F254, developed with 1:9 MeOH-CH₂Cl₂, UV visualization and staining with vanillin-H₂SO₄-EtOH, Figure 2). TLC of F-4, containing 5 and 6 (LCMS), gave a spot that characteristically stained green while F-6, containing mostly 4, stained a yellow or orange yellow (Figure 2). Further purification of F-4 was achieved by reversed-phase preparative HPLC (Phenomenex Kinetex C₁₈ column, 150 × 21.2 mm, 5μ, linear gradient, 50:50 H₂O-0.1% TFA: CH₃CN to 30:70 over 20 min, 13 mL/min flow rate, Figure 3). Pure bastadin-5 (5, 5.2 mg) eluted as a peak at 15.6 min, while pure bastadin-6 (6, 4.0 mg) eluted at 17.2 min. The peak eluting at 14.2 min, ‘X’, contained a mixture of bastadins-16[15] and −19[9] that could be separated under alternative HPLC conditions (RP C₁₈). Bastadin-4 (4, 11.0 mg) was the major component of fraction 6. The identity of all compounds were confirmed by comparisons of their MS and¹H NMR data with literature values.

Figure 2.

TLC of fractions F1–6 from Sephadex LH-20 of an extract of Ianthella basta (1:9 MeOH-CH₂Cl₂, visualization with vanillin-H₂SO₄). Fraction 4 contained 5 and 6. Fraction 6 (orange-yellow) largely pure 4.

Figure 3.

RP HPLC chromatogram of Sephadex LH-20 Fraction 4 from Ianthella basta (UV detection, λ 250 nm) ; See Experimental and Section 2.4 for conditions.). Retention times: X – a mixture of bastadins-16[15], and −19[9] (t_R = 14.2 min), bastadin-5 (5, t_R = 15.6 min), bastadin-6 (6, t_R = 17.2 min).

2.4. Bastadin-6–34-O-sulfate Ester (8)

The MeOH extract of I. basta from Guam was solvent-partitioned and a portion of the n-BuOH-soluble partition (‘fraction C’) was separated by automated low-pressure chromatography (silica, gradient elution, 0­50% MeOH-CH₂Cl₂ over 60 min) to give nine fractions. Fraction 7 (385 mg) was further separated by HPLC (Phenomenex Luna 5μ Phenyl-hexyl, 21.2 × 250 mm, 8.5 mL.min⁻¹, 70–100% MeOH-10 mM Na₂S0₄ over 60 min). Individual HPLC fractions were freed of salts and recovered by capture on C₁₈ solid-phase extraction cartridges, followed by washing with H₂O and elution of the organic compound with MeOH to give bastadin-6 (6, 4.5 mg), bastadin-5 (5, 3.2 mg) bastadin-4 (4, 3.8 mg) and a mixed fraction containing 8. The latter was further purified by reversed phase HPLC (Phenyl-hexyl, 4.0 mL.min⁻¹, 60–100% MeOH-H₂O over 15 min) and a similar recovery procedure, to provide pure 8 (4.5 mg). Colorless powder; FTIR (ATR, ZnSe plate): ν 1676, 1437, 1203, 1133, 841, 801, 723 cm⁻¹;¹H and¹³C NMR (CD₃OD), see Table 1 (see Supporting Information for NMR data in CD₃CN). MALDI-TOF m/z 1222.87 [M+Na]⁺ (calcd for C₃₄H₂₅⁷⁹Br₃⁸¹Br₃N₄Na₂O₁₁S 1222.61). EIMS m/z 497.8 (34%)/499.8 (99%)/501.8 (100)/503.8 (36), see Figure 4. HRESIMS m/z 1092.6943 [M−SO₃+H]⁺ (calcd for C₃₄H₂₇⁷⁹Br₆N₄O₈ 1092.6924).

Table 1–

¹H and¹³C NMR data for 8 (CD₃OD, 23 °C)

Atom	δ13ca 6	δ13c’b 8	δ1H (mult, J, integ.)c 8	HMBC 8 (13C–>1H)
1	27.3	27.2	3.70 (s, 2H)	36,38
2	151.5	150.9		1
3	163.0	164.0		1,5
4
5	38.4	40.5	3.45 (m, 2H)	6
6	33.9	33.9	2.79 (t, J=7.5 Hz, 2H)	5
7	140.1	140.0		5,6
8	133.6	133.4	7.59 (s, 1H)	6,10
9	117.4	117.5		8
10	146.1	147.0		8,10
11	117.6	117.5		10
12	133.6	133.4	7.59 (s, 1H)	6,8
13		–
14	144.6	144.8		19
15	141.6	142.0		17,19
16	110.2	110.2		17
17	126.2	126.0	7.06 (d, J=1.9 Hz, 1H)	19,20
18	130.7	130.8		20,21
19	111.7	112.0	6.30 (d, J=1.9 Hz, 1H)	17,20
20	32.7	33.5	2.7 (t, J=7.2 Hz, 2H)	17,19,21
21	40.4	39.1	3.40 (m, 2H)	20
22
23	163.3	164.4		21,25
24	150.4	150.4		25
25	28.7	28.3	3.80 (s, 2H)	27,31
26	137.6	137.2		25
27	133.2	133.2	7.55 (s, 1H)	25,31
28	117.1	117.6		27
29	146.1	147.9		27,31
30	117.1	117.6		31
31	133.2	133.2	7.55 (s, 1H)	25,27
32
33	144.8	150.5		38
34	141.8	137.2		36,38
35	109.8	119.0		36
36	126.8	135.2	7.21 (d, J=1.9 Hz. 1H)	1,38
37	128.1	127.0		1
38	112.6	113.8	6.30 (d, J=1.9 Hz, 1H)	1,36

^a,recorded in DMSO-d₆, from reference 12a.

^b,125 MHz.

^c,500 MHz.

Figure 4.

ESIMS fragmentation of 8

2.5. Acid Hydrolysis of 8

A mixture of 8 (0.5 mg) and 2M HCl (0.5 mL) was heated in a sealed tube at 50 °C for 30 minutes. The sample was cooled, dried under a stream of nitrogen and taken up in CD₃OD. The¹H NMR (500 MHz) of the solution was identical to that of authentic bastadin-6 (6). Analytical HPLC of 8 and authentic 6 (Dynamax Microsorb C₁₈ column, 10 × 250 mm, 4.0 mL.min⁻¹, gradient: 60–100% MeOH-H₂O over 15 min, then isocratic) gave retention times of 9.9 min and 15.2 min, respectively. HPLC of the hydrolysate of 8 also gave a single peak with retention time of 15.2 min, consistent with 6.

2.6. Cationic Reduction of Bastadin-4 to Bastadin-5.

Cationic reduction of bastadin-4 (4) to bastadin-5 (5) was carried out by an improved modification of our earlier reported procedure.[10] To a solution of 4 (3.8 mg, 3.7 μmol, 1.0 equiv) in CF₃COOH (0.5 mL) under an atmosphere of dry N₂ was added a solution of triethylsilane (5.5 μL, 37 μmol, 10 equiv) in CH₂Cl₂ (0.5 mL). The mixture was stirred vigorously for 30 min, dried under a stream of dry N₂, and the residue purified by reversed phase HPLC (Microsorb C₁₈, 3.0 mL.min⁻¹, 80:20 MeOH-H₂O+0.1% TFA) to afford recovered 4 (0.5 mg, 13%) and 5 (1.5 mg, 40%; 65% based on recovered starting material) identical with an authentic sample by MS and¹H NMR.

3. Results and Discussion

3.1. Structure Elucidation of 8

Compound 8 was purified from the Guamanian sample of Ianthella basta using a variation of the conventional protocols we have used in the past and reported elsewhere.[9] The formula of 8 C₃₄H₂₅Br₆N₄NaO₁₁S was assigned from mass spectrometric measurements [MALDI m/z 1222.87 [M+Na]⁺ calcd. 1222.61 for C₃₄H₂₅⁷⁹Br₃⁸¹Br₃N₄Na₂O₁₁S; HRESITOFMS m/z 1092.6943 [MSO₃+H]⁺). The isotope pattern reveals the presence of Br₆, the highest number of Br atoms that can be accommodated in a bastadin skeleton (e.g. 6). While the⁷⁹Br₃⁸¹Br₃ isotopomer of 6 is expected to show m/z 1120.67 [M+Na]⁺ the higher mass measured for 8 is reconciled by the difference ΔM= 101.94 [SO₃+Na-H]; therefore, 8 is an O-sulfate half-ester of 6.

The¹H NMR spectrum of 8 (Table 1) showed characteristic two-proton aryl signals for each of the symmetrical 3,5-dibromotyrosine-like spin systems (δ 7.59, s, 2H, H-8/H-12; δ 7.55, s, 2H, H-27/H-31). These data could only be accommodated by the aryl-ring substitution pattern found in the hexabromo-substituted, bastadin-6 (6), or its unreported iso-bastarane (iia) constitutional isomer. Assignment of the catechol ether linkages between the eastern and western hemispheres of structure 8 was secured by observation of two sets of HMBC correlations ((Table 1, Figure 2) recorded in different solvents (600 MHz), CD₃OD (Table 1) and CD₃CN (see Supporting Information) to resolve equivocal assignments from overlapping signals. Cross peaks were observed between H-21 (δ 3.40, m, 2H) and H-25 (δ 3.80, s, 2H) to the C-23 amide carbonyl at (δ 164.4, s), and between H-1 (δ 3.70, s, 2H) and H-5 (δ 3.45, m, 2H) and C-3 (d, 164.0, s).

The foregoing data for 8 support the same constitution as 6. The locations of Br substituents in the carbon skeleton of 8 were verified by electron-impact mass spectrometry (EIMS) that revealed loss of SO₃ and a major Br₃-containing fragment (m/z 497.82/499.8/501.82/503.82) arising from the previously noted [1] and characteristic double-heterolytic cleavage between the amide carbonyl and ketoxime C=N double bond (Figure 4). Additional fragment ions from sequential homolytic losses of Br (m/z 417.8/419.8/421.8 and 339.9/341.9) were also observed in the EI mass spectrum of 8. Therefore, 8 is a member of the ‘bastarane’ series, ia.

Confirmation of the structure of 8 followed from its acid hydrolysis (2M HCl aqueous MeOH, 50 °C, 30 min) which gave a compound that was identical with authentic 6 by¹H NMR[1] and HPLC retention time.

The structures of known bastadin sulfate half-esters have O-sulfate groups placed at either C-15 or C-34, or both.[9,13] In order to ascertain the position of the O -sulfate in 8, the¹³C NMR chemical shifts of the latter were compared with those of 6.[1,] As noted earlier by Ragan [14], Wright and co-workers[13] and others,[9] sulfation of a phenoxyl group leads to an upfield shift of the ipso¹³C signal by approximately Δδ 5 ppm and downfield shifts of ortho and para¹³C signals. The 13C NMR chemical shifts of bastadin-6 (6) were assigned by HMBC and HSQC and compared with those of 8 (Table 1 and Figure 5), and subjected to differential analysis. The C-34 quaternary carbon signal of 1 (δ 137.2, s) was shifted upfield (Δδ −4.6 ppm), while C-33 and C-35 appeared downfield (Δδ 5.7 and 9.2 ppm, respectively); therefore, the O-sulfate half-ester in 8 is positioned at C-34.

Figure 5.

HMBC data (500 MHz) of 8. Correlations (¹H −>¹³C) observed in CD₃CN (red dashed lines, obscured in CD₃OD), and those observed both CD₃OD and CD₃CN (solid lines).

3.2. Optimized Procurement of Bastadins-4–6

Given the lengthy and tedious procedure required to obtain 4–6 and 8 from the Guamanian sample of I. basta, we invested time to optimize the purification of the former desirable compounds and reduce the number of HPLC purification steps. A sample of I. basta collected in 1993 in Exmouth Gulf, Western Australia, was found to be devoid of 4 but contained 5 and 6 (LCMS). Since the former complicates the separations of the latter by preparative reversed phase HPLC, we investigated high-loading purification of MeoH extracts of this sponge. In the event, gel filtration (Sephadex LH-20, MeOH elution) of the solvent-partitioned ‘fraction-B’, with monitoring by LCMS, delivered a single late-eluting fraction containing 5 and 6. The latter fraction was separated in single HPLC step (reversed phase Phenyl-hexyl column, Phenomenex, H₂O-CH₃CN gradient) giving pure samples of 5 and 6.

Although 4 is the 5,6-dehydro-derivative of 5, conversion of 4 into 5 by catalytic hydrogenation is complicated by over-reduction and loss of Br through hydrogenolysis. Attempted homogenous catalytic hydrogenation of 4 with Wilkinson’s catalyst ((Ph₃P)₃RhCl, H₂ >60 atm) returned only starting material.[10] Eventually, we refined an optimized procedure for procurement of high-value 5; cationic reduction of 4 (Figure 6, Et₃SiH, CH₂Cl₂-CF₃COOH, vigorous stirring under nitrogen) gave 5 in 65% yield (based on recovered 4) after HPLC purification (Figure 5), which significantly improves over our earlier protocol.[10]

Figure 6.

Cationic reduction of 4 to 5.

3.3. Conclusions

The new compound, bastadin-6-34-0-sulfate ester (8) was isolated from a specimen of Ianthella basta collected in Guam. Refinement of a new purification protocol gave pure samples of highly-value bastadins-5 (5) and −6 (6) in two steps from a polar solvent-partitioned fraction. An improved conversion of 4 to 5 by cationic reduction facilitates access to this most potent RyR-1 agonist. Compound 8 undergoes acid hydrolysis to provide 6. Thus, these reactions of 4 and 8 follow convergent paths to deliver more of the high-value analogs 5 and 6, respectively.