Aleutianamine B, A Novel Chiral Sulfoxide Isolated from Latrunculia spp. and Generated via Semi-Synthesis, Exhibits Selectivity for Ovarian Cancer Cells

Abstract

We report a new optically active sulfoxide containing natural product isolated from an Alaskan Latrunculia spp. collection off the shores of the Aleutian Islands which we term aleutianamine B. It was identified and characterized in part via contemporary computational tools. Global Natural Products Social Molecular Networking (GNPS) analysis was a key tool used in identifying aleutianamine B in the crude extract. The relative configuration of the unique stereogenic sulfur atom was determined by DFT computational analysis. We also report the synthesis of aleutianamine B from aleutianamine. Aleutianamine B was evaluated against ovarian cancer and normal cell lines. This revealed that aleutianamine B retained submicromolar potency against A2780 cells while being ≥50-fold less potent on noncancerous cells; this was in stark contrast to the parent compound, aleutianamine, which displayed potent cytotoxicity against normal cell lines. Overall, the pyrroloiminoquinone (PIQ) class increasingly reveals unique selectivity and potency for tumor cell groups that are currently resistant to available chemotherapy, providing significant leads for therapeutics and biochemical probes.

Graphical Abstract

Part of the significance of the pyrroloiminoquinone (PIQ) class of natural products stems from their unique ability to selectively target specific tumor subgroups.^1,2 The PIQs encompass a wide array of sponge-derived marine natural products including not only the makaluvamines, discorhabdins, batzellines, damirones, but also the recently discovered aleutianamine (Figure 1).^3,4 The unique structural subtypes that define each class has increasingly been the topic of numerous total syntheses with improved synthetic strategies aimed at constructing the PIQ core common to these molecules given a renewed focus.^5–10 Nonetheless, the discovery and isolation of new PIQ-containing natural products from the crude sponge extracts still remains one of the most effective means for identifying new complex and bioactive molecules.^3,11,12 We recently reported the discovery of aleutianamine (1), a potent inhibitor of PANC-1 cancer cells (IC₅₀ = 25 nM), from Latrunculia austini collected in the Gulf of Alaska.¹³ The Aleutian Islands are host to a surprisingly biodiverse marine environment that has been molded by the rough northern pacific ocean and cold temperatures.^4,14–18 GNPS molecular ion network analysis¹⁹ of the crude ethanol extract from Alaskan Latrunculia spp. have revealed they are concentrated with a wealth of PIQ-containing natural products yet to be identified.¹³

Figure 1.

Structures of aleutianamine and aleutianamine B. The pyrroloiminoquinone motif is highlighted in red.

Our interest in this class of marine natural products stems from the unique and highly strained structure of aleutianamine (1). The structure of aleutianamine incorporates a bridged thioether and strained bromodiene embedded onto a PIQ scaffold creating a fused heptacyclic ring system. The unambiguous placement of the bromine atom was challenging by conventional spectroscopic means and was only made possible by the use of DP4+ probability analysis. Importantly, the structural assignment and bioactivity of aleutianamine has been validated by three total syntheses.^5,20,21 This highlights the power behind these computational tools that are easily adapted to artificial intelligence-driven analysis to greatly simplify and validate the structural assignment of complex molecules that are relatively devoid of hydrogen.^13,22–25

In continuation of our search for new complex PIQ-containing alkaloids we report the discovery, isolation, and characterization of aleutianamine B (2), a novel stereogenic sulfoxide PIQ. The application of density functional theory (DFT) computational probability analysis to confirm the structural assignment and solve the relative stereochemistry of the sulfoxide moiety present in this new PIQ alkaloid was essential. We also report the synthesis of aleutianamine B from aleutianamine. In addition, we describe the unique and highly selective bioactivity of aleutianamine B, exemplifying the potential of sulfoxide-containing natural products in cancer medicine.²⁶

RESULTS AND DISCUSSION

L. oparinae and L. hamanni²⁷ were collected off the coast of Adak, AK, during NOAA’s biannual bottom trawl survey of the Aleutian Islands. A mixture of L. oparinae and L. hamanni were extracted with ethanol, defatted, and analyzed by qTOF-LCMS. A MS/MS GNPS molecular ion network analysis was used to assist in the identification of molecular ions that represent new PIQ-containing natural products and visualized with Cytoscape. The primary PIQ-containing metabolites that were detected are discorhabdin C, 3-dihydrodiscorhabdin C, discorhabdin B, 3-dihydrodiscorhabdin B, and makaluvamine I. The PIQ cluster containing aleutianamine (m/z 397.99) was identified (Figure 2) and revealed two new brominated PIQ metabolites with parent masses of m/z 413.98 and 432.00 that clustered closely with 3-dihydrodiscorhabdin B. Based on mass differences, the two new brominated PIQs are related to 3-dihydrodiscorhabdin B as oxidation (+16 amu) and dehydration (−18 amu) products.

Figure 2.

GNPS molecular ion network analysis of the PIQ cluster from the crude ethanol extract of L. oparinae and L. hamanni. The size of the node is mapped to relative ion intensity. Thickness of the edge lines is mapped to cosine score value (i.e., thicker lines = higher cosine score). Numbers on the line are the mass unit difference between the two nodes. ID = 1cbba787516f4fb79860d39fb5008cfd.

To further explore the structure and bioactivity of these new molecules the ethanol extract was partitioned between hexane and aq. acid to remove the lipids, and the alkaloids were fractionated on silica gel stationary phase. The compound with a target mass of m/z 413.98 was isolated as a single diastereomer with the guide of mass directed fractionation and purification as an olive-green solid (14 mg) and had a molecular formula (MF) consistent with [C₁₈H₁₃BrN₃O₂S]⁺. Analysis of the ¹³C NMR spectrum confirmed the presence of 18 carbon atoms. Initial inspection of the ¹H NMR showed a striking resemblance to that of aleutianamine with the primary differences being in the chemical shifts of the thioaminal CH-8, the C-4 bridgehead methylene, and a vinyl CH-7. By comparing the MF with that of aleutianamine, and in light of the similarities and perturbations of the ¹H NMR spectrum (Table 1), we surmised that this new PIQ could be the stereogenic sulfoxide variant of aleutianamine. The infrared spectrum showed two strong bands at 1677 and 1202 cm⁻¹ indicating the presence of a carbonyl and possible sulfoxide. Specific rotation indicated that (−)-aleutianamine B was isolated.

Table 1.

¹H and ¹³C NMR Data of Aleutianamine B in DMSO-d₆ and Methanol-d₄

Atom	1H (600 MHz, DMSO-d6) δ ppm	13C (150 MHz, DMSO-d6) δ ppm	1H (600 MHz, methanol-d4) δ ppm	13C (150 MHz, methanol-d4) δ ppm
1	7.24 (s, 1H)	128.2	7.13 (s, 1H)	129.5
2		122.5		124.3
3	5.15 (dd, J = 3.0, 3.0 Hz, 1H)	61.8	5.08 (dd, J = 3.0 Hz, 3.0, 1H)	64.4
4a	2.56 (dd, J = 12.5, 3.0 Hz, 1H)	26.9	2.60 (dd, J = 12.7, 3.0 Hz, 1H)	28.6
4b	2.28 (d, J = 12.5 Hz, 3.0 1H)		2.43 (dd J = 12.7, 3.0 Hz, 1H)
5		53.5		55.7
6		143.9		145.6
7	5.36 (d, J = 3.9 Hz, 1H)	107.3	5.34 (d, J = 4.0 Hz, 1H)	108.1
8	5.58 (d, J = 3.9 Hz, 1H)	69.0	5.55 (d, J = 4.0 Hz, 1H)	71.0
10		94.3		96.3
11		166.7		167.5
12		124.0		124.3
14	7.31 (s, 1H)	126.6	7.12 (s, 1H)	127.0
15		118.0		119.6
16a	3.12 (ddd, J = 16.0, 13.0, 6.7 Hz, 1H)	19.8	3.31 (ddd, J = 16.4, 6.2, 2.7 Hz, 1H)	21.3
16b	3.05 (dd, J = 16.0, 6.6 Hz, 1H)		3.23 (ddd, J = 16.4, 12.4, 6.6 Hz, 1H)
17a	4.27 (dd, J = 14.4, 6.7 Hz, 1H)	52.3	4.27 (ddd, J = 11.2, 6.6, 2.7 Hz, 1H)	54.5
17b	4.10 (ddd, J = 14.4, 13.0, 6.6 Hz, 1H)		4.29 (ddd, J = 12.4, 11.2, 6.2 Hz, 1H)
19		149.5		152.7
20		121.9		119.6
21		137.4		139.8
NH	13.27 (brs, 1H)
NH	10.67 (brs, 1H)

We confirmed the connectivity of the proposed structure through analysis of both COSY and HMBC correlations (Figure 3). COSY established the connectivity of C-3–C-4, C-7–C-8, and C-16–C-17. Strong support for the PIQ substructure was provided from the analysis of the HMBC of H-14 which showed strong correlations to C-12, −15, and −20 that comprise the PIQ as well as two four-bond correlations to C-11 and −19 arising from the sp²-zigzag relationship. Quaternary carbon C-5, which is central to the tetracyclic 6,6,5,6-ring system, is a critical atom for the structural elucidation of the aleutianamine structure. Support for the centrality of C-5 for the tetracyclic fused ring system is provided from the observed correlations with H-1, −3, −4, and −7. The characteristic quaternary N atom was identified by the presence of the adjacent downfield shifted H-3 (5.15 ppm) which showed a correlation to the iminoquinone C-19 atom; it also showed strong correlations to C-1 and −2 which further supports its position as a connector to both western and eastern halves of the molecule. The methylene bridgehead atoms H-4a/b gave ³J correlations to C-2 and C-21 and a ⁴J correlation to C-2. Vinyl H-7 showed a strong three-bond correlation to C-1 as well as two weaker four-bond correlations to C-2 and C-10 arising due to the zigzag orientation. Finally, thioaminal H-8 showed strong correlations to C-6, −10, and −21 which further supports the 2D structural assignment to the aleutianamine class.

Figure 3.

Key COSY and HMBC correlations for the structural elucidation of the 2D structure of aleutianamine B.

With an assigned 2D structure we performed a ¹H and ¹³C atom-to-atom chemical shift comparison between the proposed new aleutianamine B and aleutianamine. Comparison between the δ¹³C of the two structures (Figure 4A) revealed a significant difference in δC-4, C-10 and C-21 (−15%, −40%, 32% difference, respectively) and more moderate differences between δC-5, C-8, and C-15 (9%, 7%, and 7% difference, respectively). A majority of the δ¹³C deviations are near the S bridge. An analysis of the δ¹H differences (Figure 4B) revealed significant shielding effects at H-4b and H-8 (+0.22 and +0.44 ppm difference in DMSO-d₆, respectively). The large deviations of δ H-4b, H-8, C-10, and C-21, all of which flank the S bridgehead, can best be explained by the γ-gauche effect of a sulfoxide.

Figure 4.

(A). Δ(δ¹³C) of experimentally measured in DMSO-d₆ atom-to-atom comparison of the corresponding C atom δ¹³C (ppm) between the structures of aleutianamine B and aleutianamine, Δ(δ¹³C) = δ¹³C_AleutB – δ¹³C_Aleut (ppm); (B) Average value of Δ(δ¹H) of experimentally measured in DMSO-d₆ and methanol-d₄ atom-to-atom comparison of the corresponding H atom δ¹H (ppm) of aleutianamine B and aleutianamine, Δ(δ¹H) = δ¹H_AleutB – δ¹H_Aleut (ppm); (C) DFT-calculated δ¹H and δ¹³C chemical shifts of aleutianamine B.

The origin of the γ-gauche effect has been studied computationally with respect to its impact on ¹³C NMR chemical shifts and can largely be explained either as a steric or a stereoelectronic effect with studies indicating the latter as more likely the predominating factor.^28–30 In the case of aleutianamine B, the sulfoxide is stereogenic and the spatial orientation of the sulfoxide would yield different and quantifiable spatial nuclear shielding effects as it applies to either a steric or stereoelectronic origin. In order to determine the relative configuration of the sulfoxide and confirm the overall chemical shift assignment we performed a DFT computational analysis of the ¹H and ¹³C chemical shifts (Figure 4C, see the Supporting Information (SI) for experimental details). Force field calculations for all possible stereoisomeric assignments converged exclusively to two configurations: (3R,5R,8S) with the sulfoxide as either R or S. Within this constraint, DFT-based ¹H and ¹³C chemical shift predictions showed high accuracy for both possibilities, suggesting that the relative configuration of (3R,5R,8S) is correct. Specifically, for the δ¹³C calculations, most carbon predictions were shown to be highly accurate, with the exception of C-21, which yielded an average deviation of 8.7 ppm. For conjugated π-systems such as this, it is not uncommon for DFT calculations to exhibit delocalization errors, that can result in δ¹³C prediction inaccuracies for one or two carbon atoms within that system.^31–34 The protons attached to carbon atoms next to the brominated carbon atom (H-3 and H-1) are also expected to yield inaccuracies due to relativistic effects, yet the ¹³C NMR predictions were accurate.

To assign the relative (R) or (S)-configuration of the sulfoxide, a closer analysis of the chemical shifts near the sulfoxide region of the molecule was carried out. Analysis of the ¹H chemical shifts strongly support an (R)-configured sulfoxide, as predictions were significantly more accurate when the sulfoxide oxygen atom was oriented toward H-4a and H-7. Lower chemical shift accuracy for nearby H-7 and H-8 was observed in both diastereomers. Overall, the chemical shift prediction for atoms H-1, −4a/b, −7, −8 was in stronger agreement for the (R)-sulfoxide (RMSD = 0.14) than the (S)-sulfoxide (RMSD = 0.32). Analysis of the carbon atoms surrounding the sulfur atom (C-1, −4–8, −10, and −21) suggested that the (S)-sulfoxide (RMSD = 3.71) more closely matched the experimental data than the (R)-sulfoxide (RMSD = 3.97). As mentioned earlier, this may be a result of inaccuracies due to relativistic effects or delocalization errors. Consequently, we sought further validation by examining computed Δ(δ¹H) and Δ(δ¹³C) differences between the two proposed sulfoxide diastereomers and aleutianamine and comparing those to the experimentally observed differences (see the SI for details). Our calculations revealed much stronger support for the relative assignment of the sulfoxide as (R) as the Δ(δ¹H) and Δ(δ¹³C) calculations in this region aligned more closely with the experimentally determined differences. Collectively, these computational results provide strong support for the (3R,5R,8S)-(R)-sulfoxide relative configuration.

The absolute configuration of aleutianamine B was assigned as (3R,5R,8S)-(R)-sulfoxide on the basis of electronic circular dichroism (ECD) comparisons of the experimental spectrum to the TDDFT-predicted spectra (Figure 5). The predicted spectrum for (3R,5R,8S)-(R)-sulfoxide is in strong agreement with the experimental spectrum.

Figure 5.

Overlay of predicted and experimental ECD spectra of aleutianamine B.

In addition, the semisynthetic conversion of naturally occurring aleutianamine to aleutianamine B was achieved by treatment of aleutianamine with a small excess of tert-butyl hydroperoxide (TBHP) in 10% isopropanol/H₂O in the presence of trifluoroacetic acid. This resulted in an isolated 48% yield of aleutianamine B with the (R)-sulfoxide as a single diastereomer as confirmed by ¹H NMR and UHPLC-qTOF analysis (Scheme 1).

Scheme 1.

Conversion of Aleutianamine to Aleutianamine B

We evaluated the activity of aleutianamine B on A2780 and SKOV3 ovarian cancer cell lines using the fluorometric Cell-Titer Blue Assay (Promega). Cells were treated with a single dose of aleutianamine B, vehicle control (0.1% DMSO), or 10% DMSO (positive control) and incubated for 5 days. In the assay, aleutianamine B had activity against both cancer cell lines with IC₅₀ values of 824 ± 251 nM for A2780 cells and 1.8 ± 0.14 μM for SKOV3 cells (Figure 6A). We also evaluated the activity of aleutianamine B for selectivity for cancerous over ‘normal’ tissue cells using Chinese hamster ovary (CHO) and HEK293 (kidney epithelial) cells as untransformed controls.³⁵ In both cases, the IC₅₀ was >50 μM (Figure 6A), revealing a striking difference in the therapeutic window when treating noncancerous cells with aleutianamine B compared with the ovarian cancer cells.

Figure 6.

Concentration–response curves of (A) aleutianamine B and (B) aleutianamine on A2780 (blue) and SKOV3 (orange) ovarian cancer cell lines. This is compared with the treatment of CHO (red squares) and HEK293 (green triangles) untransformed cells. (C) Concentration–response curves of aleutianamine B and aleutianamine compared with standard chemotherapy agents 5-fluorouracil (green circles) and cisplatin (purple squares). Points are plotted as %RFU of vehicle control (0.1% DMSO in HBSS). Data presented as mean ± SEM of ≥ 3 biological replicates comprised of technical duplicates.

Next, we compared this to the activity of aleutianamine at the same cell types (Figure 6B). As previously reported for PANC-1 cells, aleutianamine^13,36 is also potent in reducing the viability of A2780 cells (IC₅₀ = 62 ± 8.5 nM) and SKOV3 cells (IC₅₀ = 349 ± 45 nM). However, perhaps the most striking result was the lack of selectivity for the parent aleutianamine at untransformed CHO (IC₅₀ = 550 ± 171 nM) and HEK293 (IC₅₀ = 42 ± 13 nM) cells. Taken together, this demonstrates that although aleutianamine B is ~ 10-fold less potent than aleutianamine on similar cell types, it is much less cytotoxic to untransformed cell lines.

We also treated the A2780 cells with a single dose of 5-fluorouracil (5-FU) and cisplatin in order to compare aleutianamine B against typical small molecule chemotherapy drugs (Figure 6C). This revealed that 5-FU (IC₅₀ = 6.1 ± 0.80 μM) and cisplatin (IC₅₀ = 6.0 ± 0.85 μM) were ~ 7-fold less potent than aleutianamine B. 5-FU primarily exerts its cytotoxic effects through inhibition of thymidylate synthase and incorporation into RNA,^37,38 while cisplatin induces cell death via the formation of DNA cross-links that disrupt replication and transcription.^39,40 Despite their clinical utility, 5-FU and cisplatin are associated with nonselective toxicity, impacting not only malignant cells but also normal, rapidly proliferating tissues. The fact that aleutianamine B is not cytotoxic to normal epithelial cells tested above is encouraging for future studies, offering a possible avenue for targeted cancer therapy with reduced off-target effects.

In summary, we report the discovery, isolation and structural elucidation of aleutianamine B, a new PIQ marine alkaloid from Alaska’s deep ocean. Aleutianamine B is an example of an exceedingly rare class of natural products bearing a conformationally constrained, stereogenic sulfoxide. The sulfoxide manifests itself in the NMR spectroscopic data through a remarkable γ-gauche effect. The γ-gauche effect enabled its relative stereochemical assignment as (R) through the use of NMR computational analysis, and it could not have been as easily determined in any other nondestructive way. We also report the conversion of aleutianamine to aleutianamine B through a peroxide-mediated oxidation. Lastly, aleutianamine B exhibited remarkable selective cytotoxicity toward ovarian cancer cells over normal tissue cell lines. This preliminary assessment indicates aleutianamine B was equally effective at killing both p53-wild-type and p53-null cells. SKOV3 cells harbor a mutation in the TP53 gene, resulting in a complete loss of p53 protein expression,^40,41 whereas A2780 cells retain a functional TP53 gene and express wild-type p53.⁴¹ Aleutianamine B maintains comparable efficacy across both A2780 and SKOV3 cell lines, indicating that its possible mechanism of action might be independent of p53 status. If true, this type of p53-independent activity is particularly valuable in the context of drug-resistant tumors, where p53 mutations often underlie therapeutic failure.^42,43

Active pharmaceutical ingredients containing chiral, nonracemic sulfoxides are rare. There are several examples in use today, two of which are the proton pump inhibitor esomeprazole and vasodilator armodafinil.⁴⁴ The importance of chiral, nonracemic sulfoxides as biologically active compounds or metabolites for cancer therapeutics has been given scant attention, and new therapeutics are desperately needed to combat drug-resistant cancers. The PIQ class of alkaloids has now been expanded to include the chiral sulfoxide moeity. Our findings strongly support the therapeutic potential of the chiral, nonracemic sulfoxide pharmacophore. This calls attention to the need for improved synthetic methods that increase the access to these molecules and enable further investigation of this highly promising class of natural products.

EXPERIMENTAL SECTION

General Experimental Section

Reactions that required heating were carried with a stir-hot plate using a heated external oil bath. Isopropanol (HPLC grade), acetonitrile (HPLC grade), ethanol (200 proof, USP), and hexane (ACS grade) were purchased from Fisher and used without further purification. Water was purified through a Milliq purification system. ¹H NMR and ¹³C{¹H} NMR spectra were measured on an Agilent 600 DD2 (600 MHz/150 MHz) or Bruker (600 MHz/150 MHz) spectrometer at ambient temperature. Chemical shifts are reported in parts per million (ppm) and are referenced to the solvent (e.g., δ 7.26 for CHCl₃; δ 77.0 for CDCl₃). Multiplicities are indicated as follows: br (broadened), s (singlet), d (doublet), t (triplet), q (quartet), pent (pentet), sext (sextet), sept (septet), etc. or m (multiplet). Coupling constants (J) are reported in Hertz (Hz). Infrared (IR) spectra were recorded on a ThermoFisher Nicolet Summit FTIR spectrophotometer. Optical rotations ([α]_D) were measured on a Rudolph Autopol polarimeter and circular dichroism spectroscopy was carried out on a Chirascan V100 by Applied Photophysics. High performance liquid chromatography (HPLC) analyses and purification were performed using a Waters instrument using UV detection at 254 nm. High-resolution mass spectra (HRMS) were obtained with a Bruker qTOF and quaternary Elute system. Thin layer chromatography (TLC) was performed on glass plates, 250 μm, particle size 5–17 μm, pore size 60 Å. All reactions were monitored by TLC and analyzed under UV (254 and/or 365 nm) light and visualized using either PAA or KMnO4 stains. Silica gel flash column chromatography was performed on silica gel, 200–400 mesh or premium silica gel, 60 Å, 40–75 μm. Purity and homogeneity of all materials was determined by TLC, ¹H NMR, ¹³C{¹H} NMR, LCMS, and qNMR.

Collection, Taxonomic Identification, Extraction and Isolation

Latrunculia spp. were collected off shore of Adak, AK during NOAA’s 2024 biannual bottom trawl survey of the Aleutian Islands. The sponges were recovered following a trawl at depths of 150–180 m with bottom surface temperatures approximately 6–7 °C. Diagnostic anisodiscorhabds were extracted for analysis by dissolving subsamples of sponge in bleach. Anisodiscorhabds were compared to taxonomic references, resulting in the identification of the collected species as Latrunculia (Latrunculia) hamanni Kelly, Reiswig & Samaai, 2016 and Latrunculia (Uniannulata) oparinae Samaai & Krasokhin, 2002, which have been previously reported to have a sympatric distribution.²⁷

A mixture of wet, torn L. oparinae and L. hamanni was extracted by soaking with 200 proof ethanol (x3) for 24–36 h. The extract was decanted and filtered through Celite and concentrated to give ca. 350 g of a sticky tar. A small aliquot (ca. 25 mg) of the crude ethanol extract was passed through a short C18 cartridge with methanol and analyzed by qTOF-LCMS for the GNPS analysis.

The crude ethanol extract was defatted by partitioning between 6 M HCl and hexane and washing the aqueous phase with hexane (x3). The acidic aqueous phase was basified with solid NaHCO₃, extracted with 25% IPA/DCM (x5), and dried over anhydrous sodium sulfate. After decanting and filtering of the drying agent, the organic extracts were acidified with TFA and concentrated to dryness to give ca. 25 g of total alkaloids. The crude alkaloid extract was dry loaded onto silica gel and eluted with DCM then 5% → 10% → 20% → 100% MeOH/DCM doped with ca. 1% TFA. Approximately 3–5 column volumes of eluent were passed through per step and approximately 50 mL fractions were collected and combined (TLC guided). This amounted to a total of seven fractions. F1, 439 mg, F2, 303 mg, F3, 3.8 g, F4, 2.7 g (contained aleutianamine B), F5, 7.9 g, F6, 7.5 g, and F7, 2.3 g.

These silica gel fractions were analyzed by qTOF-LCMS for m/z 413.99. Fraction 4 was found to contain the target mass. This fraction was then subjected to reversed phased prep-HPLC using a Kinetex C18 250 × 21 mm, 15 mL/min with an elution program of: 2% ACN/H₂O (0.1% TFA) equilibration, then isocratic 5% ACN/H₂O (0.1% TFA) for 15 min, isocratic 10% ACN/H₂O (0.1% TFA) for 15 min, then isocratic 15% ACN/H₂O (0.1% TFA) for 20 min, then a linear ramp gradient to 35% ACN/H₂O (0.1% TFA) over 20 min. The fractions containing aleutianamine B were identified by LCMS analysis. Further purification was achieved by reversed phase prep-HPLC using a Kinetex C18 250 × 21 mm, 20 mL/min with a linear ramp gradient from 2% to 10% IPA/H₂O (0.1% TFA) over 35 min. Aleutianamine B (14 mg) was dried to a constant mass under high vacuum and was isolated as an olive-green solid hydrate (51% AleutB as determined by qNMR). [α]_D = − 170.3 (c 1.1 mg/mL, MeOH). IR (cm⁻¹, neat) 3666, 3152, 2923, 2851,1677, 1621, 1525, 1412, 1202, 1135. HRMS m/z calcd for [M⁺] [C₁₈H₁₃BrN₃O₂S]⁺, 413.9906; found 413.9889.

Oxidation of Aleutianamine (1) to Aleutianamine B (2)

A stirred solution of aleutianamine TFA salt (ca. 3.0 mg of hydrate, 6.8 μmol) in 10% IPA/H₂O (5 mL) was added 0.4 mL of 1% tert-butyl hydroperoxide solution in water and 3 drops of TFA from a glass Pasteur pipet. The reaction mixture was heated to 55 °C. Upon completion (TLC: 5–10% MeOH/DCM with 0.1%TFA) the reaction mixture was cooled to room temperature, quenched with sodium sulfite solution, and concentrated in vacuo. Aleutianamine B TFA was isolated by purification on reversed phase semipreparative HPLC (Kinetex C18 250 × 21 mm, linear gradient 2→35% ACN/H₂O with 0.1%TFA over 45 min, 20 mL/min) as an olive-green solid (ca. 1.5 mg, 48% yield).

Cell-Titer Blue Assay

HEK293 cells were sourced from ATCC (CRL-1573). A2780 and SKOV3 cells were a generous gift from Dr. Joe R. Delaney (MUSC). Cells were screened negative for mycoplasma contamination by monthly scheduled testing (LookOut Mycoplasma PCR Detection Kit; Sigma, cat. No. MP0035). Cells were cultured at 37 °C, with an atmosphere of 5% CO₂, in Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% FBS and penicillin (100 U/mL) and streptomycin (100 μg/mL).

Drug dilutions were prepared at a 10x concentration in drug buffer (Hanks balanced salt solution with Ca²⁺, Mg²⁺, 20 mM HEPES, 0.1% BSA, and 0.03% ascorbic acid). The Cell-Titer Blue Assay was performed in black-walled, clear-bottomed 96-well plates. In brief, cells were seeded (10,000 cells/well) in complete DMEM growth media (90 μL), and allowed to adhere to the plate in the incubator for 24 h. Then, 10 μL of drug solution was added, and the cells were incubated for 5 days at 37 °C/5% CO₂. Cell-Titer Blue dye (20 μL) was added, and the cells were incubated for 4 h, before fluorescence was read on a Promega GloMax Discover Plate Reader using the default Cell-Titer Blue protocol. Changes in fluorescence are represented as relative fluorescence units, and are plotted as %RFU of vehicle control (0.1% DMSO) after subtracting background (positive kill control, 20% DMSO). Concentration–response analysis was performed using sigmoidal curve fitting functions in GraphPad Prism.

DFT Chemical Shift and ECD Calculations

Chemical shift predictions were used to confirm the structure of aleutianamine. Structures for several possible configurations of aleutianamine were energy-minimized using the MMFF94⁴⁵ force field in Chem3D, followed by an extended mixed torsional/low-mode (MTLMOD) conformational search performed using the OPLS4⁴⁶ force field in the Schrödinger MacroModel software package. DFT calculations were performed in Gaussian 16; calculations for carbon atoms and protons, excluding carbon atoms bonded to sulfur or bromine, were performed using the DELTA50 methodology,⁴⁷ optimized for DMSO-d₆ (first reported in a separate natural product study).⁴⁸ For these calculations, ¹³C δ predictions were performed at the PCM-ωB97X-D/def2-SVP//PCM-B3LYP-D3/6–311G(d,p) level of theory, and ¹H δ predictions were performed at the PCM-WP04/6–311++G(2d,p)//PCM-B3LYP-D3/6–311G(d,p) level of theory, both of which utilized an integral equation formalism (IEF) polarizable continuum model (PCM) for DMSO. Linear scaling factors include a slope of −1.0099 and intercept of 196.0386 for ¹³C, and a slope of −1.0080 and intercept of 32.0998 for ¹H. Predicted chemical shifts were Boltzmann-weighted based on the total Gibbs free energy of each conformer. No imaginary frequencies were present within the frequency calculation results.

Relativistic effects of heavy atoms were accounted for through accompaniment of specialized DFT calculations for carbons bound to sulfur and bromine. For the carbon atoms bonded to sulfur, the WC04 functional was selected, as the test set used for optimization of the functional included several sulfur-containing molecules.⁴⁹ δ¹³C predictions for sulfur-bound carbons were performed at the WC04/6–31G(d)//B3LYP/6–31G(d,p) level of theory. Linear scaling factors include a slope of −0.9017 and intercept of 192.7573 for ¹³C, based on results from Pierens for WC04-based DMSO-d₆ predictions.⁵⁰ For the carbon atom bonded to bromine, calculation methods were adopted from Giesen.⁵¹ δ¹³C predictions were performed at the B3LYP/MIDI!//B3LYP/MIDI! level of theory. Linear scaling factors include a slope of −1.07 and intercept of 200.4 for ¹³C, specifically derived from brominated carbon atoms.

For ECD predictions, TDDFT calculations were carried out at the PCM-CAM-B3LYP/aug-cc-pVDZ//CAM-B3LYP-D3/aug-cc-pVDZ level of theory, with an IEFPCM for methanol and 100 excited states for the time-dependent calculations. Data were Boltzmann-weighted and processed in SpecDis,⁵² with a sigma broadening factor of 0.3 eV and a UV correction of +19 nm. The similarity factor for the (3R,5R,8S)-(R)-sulfoxide enantiomer was calculated to be 0.9213 after fitting. The Boltzmann populations for the conformer 1 and 2 are 87.47% and 12.53%, respectively.

Supplementary Material

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jnatprod.5c01094.

Reproductions of NMR, IR, HRMS, photos of spicules isolated for taxonomic determination, and additional details on the DFT analysis (PDF)

Data output file for the TD-DFT calculations (TXT)

Data output file for the TD-DFT calculations (TXT)

Data Availability Statement

The NMR experimental data of aleutianamine B has been deposited in the Natural Products Magnetic Resonances Database (NP-MRD; www.np-mrd.org) and is available under the accession number NP0351690.