Dual Biological Potential: Anticancer and Antibacterial Activities of Bis‐α, ω‐Bifunctional Sphingolipids from the Marine Sponge Oceanapia ramsayi

Abstract

Marine sponges of the genus Oceanapia are prolific producers of bioactive secondary metabolites. Four known bifunctional sphingolipids are isolated: isorhizochalin (1), rhizochalin (2), rhizochalin C (3), and oceanapiside (4) from Oceanapia ramsayi sponges. These compounds are derivatized to form aglycones (1b‐4b), perbenzoates (1c–4c), and bis‐oxazoline (2e). The compounds and their derivatives are investigated for their cytotoxicity against three breast cancer cell lines (HCC70, MDA‐MB‐231, MCF7). All four sphingolipids are cytotoxic with significant selectivity indices (SI) for the three breast cancer cells. Compound 1 has high selectivity against the HCC70 cell line, 4 against the MDA‐MB‐231 cells, and aglycon 4b against both the MDA‐MB‐231 and MCF7 cell lines, respectively, demonstrating the functional importance of the sugar units and terminal amino‐alcohol stereochemistry. Compounds 1–4 also show promising antibacterial activity against Staphylococcus aureus, Acinetobacter baumannii, Pseudomonas aeruginosa, Escherichia coli, and the multiple antibiotic‐resistant Klebsiella pneumoniae test strains, with compound 2 exhibiting the greatest activity. Derivatization diminished potency, indicating that the sugar moiety and stereochemistry are functionally important for antibacterial activity.

Preparative work on Oceanapia ramsayi sponge specimens collected from the northeast South African coast (A) result in the isolation of four bifunctional sphingolipids (center). Bioassays of these compounds and their derivatives (B) show antimicrobial activity against bacterial pathogens and selective cytotoxicity to breast cancer tissue culture cell lines (C).

1. Introduction

Natural products (NPs) continue to play a crucial role in the development of new therapeutics for various ailments, with the majority of contemporary drugs either derived from or inspired by natural sources.^{[ 1} , ² , ^{3 ]} Marine ecosystems, with their largely uncharacterized biodiversity and complex chemical ecology, represent an important reservoir of bioactive small molecules providing scaffolds^{[ 4 ]} for the development of new therapeutics. Several marine sponge species are known to be rich sources of chemical and biologically active sphingolipids. Species of the genus Oceanapia and Rhizochalina are distinguished by their ability to synthesize a variety of complex bioactive compounds. Among these are bifunctional sphingolipids, specifically bis‐α,ω‐bifunctional sphingolipids, that have gathered attention due to their amphipathic nature and exhibit diverse biological properties, including antibacterial and anticancer activities.^{[ 5} , ⁶ , ⁷ , ⁸ , ^{9 ]}

Several α,ω‐bipolar sphingolipid NPs have been previously isolated from the marine sponges Oceanapia ramsayi, Oceanapia seychellensis, Oceanapia phillipensis, and Rhizochalina incrustata. These include rhizochalin, oceanapiside, the oceanalin series, and their corresponding peracetates.^{[ 10} , ¹¹ , ¹² , ^{13 ]} While rhizochalin was first isolated as an antibacterial agent against the Gram‐positive bacterium Staphylococcus aureus,^{[ 11 ]} subsequent studies have focused on its anticancer activity. Sphingolipid metabolism is closely linked to cell fate, regulating processes such as signal transduction, proliferation, differentiation, apoptosis, and cell membrane integrity.^{[ 14 ]} Given the structural resemblance of marine sponge‐derived sphingolipids to ceramide analogs, which are key mediators of apoptosis in cancer cells, several studies have focused on the anticancer activity of these bifunctional sphingolipids.^{[ 5} , ¹⁵ , ^{16 ]} For instance, rhizochalin and its aglycone, rhizochalinin, exhibit potent cytotoxicity against the mouse epidermal cell line (JB6 P + Cl41), human cervical cancer cell line (HeLa), human acute monocytic leukemia cells (THP‐1), and human acute promyelocytic leukemia cells (HL‐60), with rhizochalinin also activating AMP‐activated protein kinase (AMPK) in HT‐29 colon cancer cells.^{[ 7} , ^{9 ]} Additionally, rhizochalinin A and C selectively target THP‐1 leukemia cells, while isorhizochalinin, a stereochemical variant, displays strong cytotoxicity against HL‐60 and THP‐1 cells.^{[ 6} , ^{17 ]} Furthermore, structure‐activity relationship (SAR) studies on rhizochalin and synthetic analogs against castration resistant prostate cancer showed differential cytotoxicity, antiproliferative, proapoptotic, autophagy, and androgen receptor signaling inhibition of the compounds, with analogs of rhizochalin, rhizochalinin, and 18‐hydroxyrhizochalinin selected for further development.^{[ 18 ]}

Here, we extend the range of antimicrobial activity of the four sponge‐derived α,ω‐bipolar sphingolipids and their derivatives using SAR to shed light on the mechanism of action in bacteria and cancer cells.

2. Results and Discussion

2.1. Chemistry

Preparative chromatography efforts, guided by biological activities of the intermediate fractions, focused on specimens of the sponge O. ramsayi (MOZ95‐019 and TIC2022‐101) collected from Ponta do Ouro, Mozambique, in 1995 and KwaZulu‐Natal, South Africa, in 2022, respectively. These specimens’ methanol (MeOH) extracts were subjected to Sephadex LH‐20 column chromatography and reversed‐phase solid‐phase extraction. This was followed by reversed‐phase semipreparative high‐performance liquid chromatography (HPLC) purification. We isolated four known bifunctional sphingolipids, namely isorhizochalin (1)^{[ 17 ]} rhizochalin (2),^{[ 19 ]} rhizochalin C (3),^{[ 6 ]} and oceanapiside (4)^{[ 12 ]} (Figure  1 ).

Figure 1.

Isolated bifunctional sphingolipids 1–4.

The structures of compounds 1–4 were elucidated using 1D and 2D NMR spectra (SI 6–9, Supporting Information), complemented with Fourier transform infrared spectroscopy (FTIR) and circular dichroism (CD) spectroscopy, and mass spectrometry (MS) (SI 1–14, Supporting Information). The sugar residues in the α, ω‐bis‐amino alcohol structure, featuring a β‐sugar moiety, were identified using ¹H and ¹³C NMR chemical shifts, supported by existing literature.^{[ 6} , ¹² , ^{19 ]} The heteronuclear single quantum correlation (HSQC) spectrum revealed diastereotopic protons at the C‐6′ in compounds 1 and 2. DEPT and ¹³C (SI‐6 and 7, Supporting Information) spectra analysis showed that the anomeric carbon δ of 1 and 2 were displayed at δ_C 103.4 and 103.7 ppm, respectively. The heteronuclear multiple‐bond correlation (HMBC) cross‐peak correlation ³ J coupling of the anomeric proton (H‐1′) with C‐3 at δ 4.22, 79.2 ppm and δ 4.27, 80.1 ppm, confirmed that the sugar unit of compounds 1 and 2 is a β‐galactoside bonded to C‐3.^{[ 13 ]} In addition, NOE correlations via space between H‐1′ to H‐2 of 1 and 2 supported the proposed configuration of the β‐galactoside bonded to C‐3.^{[ 19 ]} Therefore, compounds 1 and 2 show a resonance pseudo‐dimeric amino alcohol^{[ 6 ]} and C‐2 epimers.^{[ 19 ]} The FTIR analysis (SI‐3 and 4, Supporting Information) verified the observed ¹³C‐NMR chemical shift signal at 214.5 ppm, corresponding to the stretching vibration of the ketone functional group at 1640 cm⁻¹. The deshielded ¹H‐NMR δ _H at 2.32 ppm (dd, J = 7.3, 2.5, 4H) for H‐17 and H‐19 may be attributed to the anisotropic effect exerted by the nearby (—C=O).

The HR‐ESI‐MS spectrum of compound 1 shows a base peak of ESI‐MS m/z [M + 2H]²⁺ 317.2570 (100%) corresponding to the molecular formula (MF) C₃₄H₇₀N₂O₈. The singly charged ion [M + H]⁺ was also observed at m/z 633.5074, matching C₃₄H₆₉N₂O₈ (SI‐1, Supporting Information). Compound 2 indicated the base peak [M + 2H]²⁺ m/z 317.2605 (100%), which agreed with the MF C₃₄H₇₀N₂O₈, and the singly charged ion was observed at m/z 633.5043, matching the MF C₃₄H₆₉N₂O₈ (SI‐1, Supporting Information). Compound 3 was also a β‐galactoside bonded to C‐3, differing from compounds 1 and 2 in the attachment of the hydroxy group to C‐1. However, compound 4 contains a β‐glucopyranose sugar moiety bonded to C‐3 (Figure 1). The β‐configuration of the sugar moiety was confirmed by the ³ J HMBC correlation between the anomeric proton and C‐3 of the sphingoid base. The HSQC spectrum revealed diastereotopic protons at the C‐1 and C‐6′ in compounds 3 and 4 (SI‐8 and 9, Supporting Information). The anomeric carbon of compound 4 exhibits a slightly deshielded signal at δ _C 104.9 ppm compared to compound 3 at δ _C 102.5 ppm, which can be attributed to the axial proton at C‐4′. DEPT‐135 and ¹³C NMR verified one methyl at 14.9 and 12.4 ppm in compounds 3 and 4, respectively. In compounds 3 and 4, ² J HMBC inferred that a ketone group is flanked by two CH₂ groups of C‐17 and C‐19 (SI‐8 and 9, Supporting Information).^{[ 6 ]} A ketone functional group was confirmed by the FTIR spectrum in the range of 1650–1702 cm⁻¹ (SI‐3 and 4, Supporting Information).

The exact locations of the carbonyl group in compounds 1–4 were determined using in situ decomposition followed by MS analysis, and the data were in line with the previous report.^{[ 20 ]} The CD spectroscopy (SI‐5, Supporting Information) confirmed that the stereochemistry of compounds 1–4 is C‐2 epimers.^{[ 19} , ^{20 ]} The HR‐ESI‐MS showed that compound 3 is a doubly protonated molecular ion [M + 2H]²⁺ at m/z 325.225 (100%) corresponding to the MF C₃₄H₇₀N₂O₉ and a singly charged m/z [M + H]⁺ at 649.438 (8.3%), matrix‐assisted laser desorption/ionization ‐ time‐of‐flight mass spectrometry produced a [M + Na]⁺ ion at m/z 671.419 (1.55%, C₃₄H₆₉N₂O₉Na) (SI‐2, Supporting Information). Compound 4 has a base peak of [M + 2H]²⁺ at m/z 325.715 (100%) with MF C₃₄H₇₀N₂O₉. It is a doubly charged ion with an m/z [M + H]+ of 650.432, corresponding to the MF C₃₄H₆₉N₂O₉ (1.99%) (SI‐2, Supporting Information).

The outcome of our preparative efforts was unsurprising, as bifunctional sphingolipids are well‐known chemotaxonomic markers of the sponge genus Oceanapia. These compounds are reported to exhibit both antimicrobial and anticancer activities. The bifunctional sphingolipids (1–4) were further derivatized into 13 additional compounds using the previously described methodology.^{[ 20 ]} The derivatization of compound 2 serves as an example of the typical process (Figure  2 ), with the resulting derivatives confirmed through MS (SI‐10 and 12, Supporting Information) and ¹H NMR (SI 11–13, Supporting Information). Compound 2 was acetylated in the presence of pyridine, yielding intermediate rhizochalinin peracetate (2a). This intermediate underwent methanolysis in the presence of HCl to produce rhizochalinin (2b). Additionally, compound 2a was subjected to Baeyer‐Villiger oxidation, resulting in compound 2d, which was used in decomposition studies. Compound 2b was cyclized in the presence of benzoyl chloride and pyridine, producing bis‐oxazoline (2e). While these compounds have been previously reported, their biological activities toward bacteria are interrogated for the first time here.

Figure 2.

Sample derivatization process of bifunctional sphingolipids 1–4.

2.2. Biological Activities

2.2.1. Anticancer Activities

Bifunctional sphingolipids, particularly rhizochalin and its analogs, along with their derivatives, are renowned for their cytotoxic activities. These compounds have shown promising results against various cancer cell lines, including uterine, cervical, colon cancer, and leukemia.^{[ 8} , ^{9 ]} However, their effects on breast cancer cell lines remain underexplored. Initially, we investigated compounds 1–4 and their aglycones (1b–4b); perbenzoate (1c‐4c) and bis‐oxazoline (2e) derivatives for cytotoxicity against Homo sapiens HeLa cells. All four NPs (1–4) were active, with compound 2 being the most cytotoxic (IC₅₀: 7.11 μM), followed by compound 1 (IC₅₀: 12.45 μM), and compounds 3 and 4 IC₅₀ values: 22.81 and 24.90 μM, respectively (SI‐15, Supporting Information). Of the derivatives, only compound 2b retained some activity (IC₅₀: 22.50 μM), and with IC₅₀ > 50 μM for the remaining eight derivatives (Table  1 ). We assessed the cytotoxic activity of the compounds 1–4 and 1b–4b against the three breast cell lines: African American triple‐negative breast cancer cell line (HCC70), Caucasian triple‐negative breast cancer cell line (MDA‐MB‐231), the hormone receptor‐positive Caucasian breast cancer cell line (MCF7), and a nontumorigenic breast epithelial cell line (MCF12A). The IC₅₀ values and selectivity indices (SI) were determined to evaluate the potential as novel anticancer agents.

Table 1.

Cytotoxic activities of test compounds against breast cancer cell lines.

Compound	Cancer cell lines (IC50, μM)a)
	HCC70	SI	MDA‐MB231	SI	MCF7	SI	MCF12A
1	2.85 ± 1.19	7.8	5.99 ± 1.26	3.7	6.91 ± 1.29	3.2	22.10 ± 1.26
2	1.94 ± 1.13	2.2	4.59 ± 1.22	1.0	2.43 ± 1.17	1.8	4.25 ± 1.28
3	5.77 ± 1.19	2.4	2.15 ± 1.27	6.5	6.16 ± 1.18	2.3	13.97 ± 1.14
4	8.75 ± 1.15	2.3	1.96 ± 1.14	10.1	5.77 ± 1.10	3.4	19.80 ± 1.21
1b	47.0 ± 1.4	2.0	93.22 ± 1.2	1.0	34.85 ± 1.4	2.7	95.83 ± 1.3
2b	4.97 ± 1.2	4.2	6.78 ± 1.3	3.1	21.42 ± 1.2	1.0	20.83 ± 1.3
3b	6.43 ± 1.1	4.5	6.78 ± 1.2	4.3	26.37 ± 1.2	1.1	29.03 ± 1.3
4b	117.2 ± 1.4	0.6	6.36 ± 1.4	10.4	8.39 ± 1.39	7.9	65.93 ± 1.3
Cisplatin	2.6 ± 1.2	4.5	18.1 ± 1.2	0.6	4.4 ± 1.2	2.5	11.2 ± 1.3

^a)Data represent mean ± standard deveiations (n = 3).

The data revealed that the compounds 1–4 and their analogs (1b–4b) exhibit varying cytotoxic activity, equivalent to or in some cases higher than the control, cisplatin (Table 1). Compound 1 demonstrated cytotoxicity against all three cancer cell lines, with IC₅₀ values spanning from 2.85 ± 1.19 to 6.91 ± 1.29 μM, while activity against the control cell line (MCF12A) was three to sevenfold lower compared to that of cancer cell lines (SI values: 7.8 for HCC70, 3.2 for MCF7, and 3.7 for MDA‐MB231). Compound 2, on the other hand, was cytotoxic to all cell lines, with a poor selectivity profile, except for HCC70 cells (SI = 2.2). Compound 3 was more than twofold less cytotoxic against the nontumorigenic epithelial cell lines, with SI values ranging from 2.3 to 6.5. Compound 4 demonstrated high selectivity toward the MDA‐MB‐231 cancer cell line (SI = 10.1), while being only twofold and threefold more cytotoxic against the HCC70 and MCF7, cell lines, respectively, compared to the control cell line. Two aglycons of compounds 1 and 4, namely 1b and 4b, were also screened for cytotoxic activity against all the cell lines, while 1b was essentially inactive and 4b which was not active against HCC70, retained activity, and was highly selective against the MDA‐MB231 and MCF7 cell lines (SI 10.5 and 7.9, respectively). Notably, the selectivity of 4b against the MCF7 cells was twofold higher than its parent compound 4, and its IC₅₀ was at least threefold lower, and its selectivity was 17‐fold greater than cisplatin. The observed increase in activity could be based just on the molar additive effects of two adjacent amino alcohol groups that perhaps suggest an intramolecular synergism.^{[ 21 ]} This synergistic effect is likely in the inhibition of multiple receptor sites by compound 4 and its analog 4b. The cleavage of β‐O‐galactosyl from compound 2 enhanced its activity nearly twofold against HCC70 and MDA‐MB231, perhaps increasing its lipophilicity.^{[ 21 ]} The SAR of compounds 1–4 perbenzoates (1c–4c) showed a lack of cytotoxic activity, suggesting the role of the sugar moiety and terminal amino‐alcohol functional groups (SI 15, Supporting Information). Additionally, cyclization of aglycone 2b into bis‐oxazoline (2e) derivative (Figure 2) also resulted in a loss of activity. This highlights the importance of the amino‐alcohol unit and, perhaps, the role of stereochemistry as a key structural requirement (SI 15, Supporting Information). Overall, the SAR study of compound 2 and its derivatives, 1b, 2b, and 2e, revealed the importance of sugar units, the terminal amino‐alcohol, and its stereochemistry. Furthermore, the data confirmed that the C‐18 ketone (—C=O) does not contribute to the cytotoxic activity in this case, as evidenced by derivatives 1c–4c and 2e. These findings warrant further investigation into the mechanisms of action of the active compounds and potential development as targeted therapies for breast cancer, particularly for challenging subtypes such as triple‐negative breast cancer. The data also suggests that these compounds could potentially offer better therapeutic indices compared to traditional chemotherapeutics such as cisplatin.

2.2.2. Antibacterial Activities

The two O. ramsayi specimens were initially selected for preparative work during screening of crude extracts for antibacterial activity against E. coli and S. aureus. The antibacterial activities of compounds 1–4 and their derivatives (1b–4b, 1c–4c, and 2e) were evaluated against a panel of ESKAPE pathogens, including S. aureus, A. baumannii, P. aeruginosa, and E. coli. We also screened compounds 1–3, 1b, and 2b against strains of K. pneumoniae, including the drug‐resistant K. pneumoniae ST512‐K30BO isolate (Table  2 and SI‐15, Supporting Information). Active compounds that showed inhibitory activity at a concentration of 50 μM were further assessed to determine their minimal inhibitory concentrations (MIC).

Table 2.

Antibacterial activities of sphingolipids from O. ramsayi and their derivatives.

Compound			MIC [μM]a)
	S. aureus	E. coli	P. aeruginosa	A. baumannii	K. pneumoniae
1	3.88 ± 0.05	10.87 ± 0.75	6.47 ± 0.14	12.55 ± 0.16	7.55 ± 2.97
2	3.64 ± 0.47	6.15 ± 0.04	3.84 ± 0.82	6.29 ± 0.02	8.52 ± 3.10
3	11.34 ± 0.89	22.76 ± 0.38	12.23 ± 2.24	14.66 ± 11.54	5.85 ± 0.40
4	11.85 ± 0.68	14.64 ± 0.11	23.52 ± 0.33	31.60 ± 7.26	ND
1b	>50	>50	>50	>50	>50
2b	24.7 ± 0.25	49.1 ± 3.42	>50	55.7 ± 0.85	23.78 ± 0.15
3b	>50	>50	>50	>50	>50
4b	>50	>50	>50	>50	>50
1c	>50	>50	>50	>50	>50
2c	>50	>50	>50	>50	>50
3c	>50	>50	>50	>50	>50
4c	>50	>50	>50	>50	>50
2e	>50	>50	>50	>50	>50
Nalidixic acid	–	0.47 ± 0.22	0.17 ± 0.01	1.28 ± 0.92	28.83 ± 1.09
Novobiocin	0.002 ± 0.002	–	–	–	–

^a)Key: ND: not done (Data represent mean ± standard deviations (n = 3).

Compounds 1–4 were active against all bacteria strains tested, with MIC values ranging from 3.64 ± 0.47 to 31.60 ± 7.26 μM (Table 2). Among these, compound 2 demonstrated the most potent activity across all tested strains, followed closely by compound 1. This difference in activity may be attributed to variations in stereochemistry (SI‐5, Supporting Information)^{[ 22 ]} rather than β‐O‐galactosylated moiety. Notably, compounds 1 and 2 were active against P. aeruginosa (6.47 ± 0.14 and 3.84 ± 0.82 μM, respectively) and the drug‐resistant K. pneumoniae ST512 K130BO strain (7.55 ± 2.97 and 8.52 ± 3.10 μM, respectively). Conversely, β‐O‐galactosylated compound 3 exhibited weaker antibacterial activity, being over two times less potent than 1 and 2. The increased polarity from the C₁—OH group may influence its interactions with cell membranes. The β‐O‐glucosyl substituted compound 4 showed weaker antibacterial activity than β‐O‐galactosylated substituted compounds 1–3 (Table 2). Only derivative 2b retained some antibacterial activity, with an MIC of 24.7 ± 0.25 μM against S. aureus and 23.78 ± 0.15 μM against K. pneumoniae (Table 2), while the other derivatives were inactive. The stereochemistry at C₂ (SI‐5, Supporting Information) in derivatives 1b and 2b appears to have a greater influence on the antibacterial activity than the sugar moiety, as observed in parent compounds 1 and 2 against E. coli, P. aeruginosa, and A. baumannii (Table 2 and Figure 2). This data extends the antimicrobial activity of bis‐α,ω‐bifunctional sphingolipids for the Gram‐negative ESKAPE pathogens, including drug‐resistant strain K. pneumonia, and the sugar moiety is important for antimicrobial activity, because derivatizing compounds 1–4 leads to diminished antibacterial activity, and it may increase bioavailability (Table 2; SI‐15, Supporting Information).

2.3. Time‐kill Kinetics

To better understand their mechanism of action, we investigated the time‐kill kinetics of compounds 1–4 against E. coli and S. aureus. E. coli and S. aureus were exposed to concentrations equivalent to one and fourfold MICs of each compound for a period ranging from 30 min to 5 h postexposure to assess the time required to achieve bacterial kill. To distinguish between bacteriostatic and bactericidal effects, we used the minimum bactericidal concentration/MIC (MBC/MIC) ratio, where a ratio of ≤4 indicates bactericidal activity and a ratio of >4 indicates bacteriostatic activity.^{[ 23 ]} All compounds 1–4 had similar kill profiles, with bactericidal activity against both E. coli and S. aureus. For E. coli (Figure  3 ), the compounds induced rapid killing within 30 min and reached bactericidal endpoints at both 1x and 4x MICs within 1–2 h, a characteristic typical of agents targeting bacterial cell membranes.^{[ 24} , ^{25 ]} Against S. aureus (Figure 3), compounds 1–4 were not as active.

Figure 3.

Time‐kill kinetics of compounds 1–4 at 1× (19.6 μM) and 4× (78.4 μM) MIC against A) E. coli and at 1× (15 and 16.6 μM) MIC and 4× (60 and 66.4 μM) against B) S. aureus. Nalidixic acid at 1× (94.7 μM) and 4× (378.8 μM) MIC and novobiocin at 1× (0.03 μM) and 4× (0.12 μM) MIC were used as positive controls.

The time‐kill kinetics showed that significant killing at 4x MIC was observed after 2 h, and the bactericidal endpoint was achieved within 3 h. Surprisingly, at 1x MIC, complete bactericidal endpoints were not achieved, but rather a reduction in cell recovery, suggesting the potential arrest of cell proliferation rather than outright killing.

The differential time‐kill activities with E. coli versus S. aureus suggest that these compounds may employ distinct mechanisms of action. In Gram‐negative and Gram‐positive bacteria. Notably, compounds 1–4 were more potent than the controls, nalidixic acid and novobiocin, typically classified as bactericidal and bacteriostatic agents, respectively.^{[ 26} , ^{27 ]} Unlike compounds 1–4, the controls at 1x and 4x MICs did not achieve bactericidal endpoints, showing instead a rapid regrowth of cells, particularly at 1x MIC. This regrowth may be attributed to the reversal of antibacterial activity upon removal of the test compounds or selective amplification of less susceptible bacterial sub‐populations.

2.4. Membrane Disruption Activities

Antibacterial agents with rapid bactericidal activities often target bacterial cell membranes.^{[ 28 ]} Therefore, we assessed the membrane disruption activity of compounds 1–4 against E. coli and S. aureus. After incubation of 1 h, compounds 1–4 caused membrane damage in over 50% of E. coli cells, suggesting that the observed bacterial activity may be partially due to membrane disruption. In contrast, exposure to S. aureus cultures to compounds 1–4 resulted in minimal membrane damage after 1 h (Figure  4 ). Considering the time kill for S. aureus (Figure 3B) and the concentration of compounds used in this assay are five to tenfold their MICs, it is unlikely that membrane disruption is the mechanism of antibacterial activity in S. aureus. This finding further supports the likelihood of different antibacterial mechanisms in Gram‐negative and Gram‐positive bacteria.

Figure 4.

Bacterial membrane disruption potential of compounds 1–4. Data are presented as mean ± standard deviations (n = 3).

3. Conclusion

In this study, we investigated the anticancer activities of four sponge‐derived α,ω‐bipolar sphingolipid NPs, demonstrating selective cytotoxic activity against triple‐negative breast cancer cell lines. In addition, these compounds exhibit promising antimicrobial activity against five ESKAPE pathogens, notably A. baumannii and K. pneumoniae, both of which are priority antimicrobial resistance pathogens of concern in low and middle‐income countries. This study also demonstrated the importance of sugar moiety, amino‐alcohol units, and stereochemistry as key structural requirements for bioactivity and selectivity. The cytotoxicity to mammalian cells precludes the development of these NPs as antimicrobial drugs; however, understanding their mechanism of action provides opportunities for discovering new targets for antimicrobial drug development.

4. Experimental Section

4.1.

4.1.1.

General Experimental Procedure

Preparative column chromatography was conducted using Sephadex LH‐20 (Sigma Aldrich, Johannesburg, South Africa), and solid‐phase extraction was done using the Sep‐Pak Vac 35 cc 10 g C‐18 cartridges (Water, Milford, MA, USA). Semipreparative HPLC was performed with a Waters 1525 binary pump system, equipped with a Waters 2487 dual λ absorbance detector and an XBridge C₁₈‐RP‐Shield column (10 × 150 mm, 5 μm; Waters, Milford, MA, USA), using solvent mixtures of water (Merck Millipore, Johannesburg, South Africa) and acetonitrile (MeCN) (Liquid Chromatography‐Mass Spectrometry (LC‐MS) grade; Merck Millipore, Johannesburg, South Africa), both supplemented with 0.1% formic acid (FA) (Sigma‐Aldrich, Johannesburg, South Africa). The purity of subfractions of derivatives was analyzed using a thin‐layer chromatography (TLC) plate coated with silica gel (60 F254, 20 × 20 cm, Merck KGaA, Darmstadt, Germany). MS analyses were carried out on a Bruker Compact QToF mass spectrometer with an electrospray ionization probe (Bruker, Bremen, Germany). Organic extract analysis was conducted in LC‐MS/MS mode using a Dionex UHPLC (Thermo Fisher Scientific, Sunnyvale, CA, USA) equipped with a Kinetex polar C‐18 column (3.0 × 100 mm, 2.6 μm; Phenomenex, Torrance, CA, USA) at a flow rate of 0.300 mL min⁻¹, utilizing the same solvent mixtures as semipreparative HPLC. IR spectra were recorded on a Perkin‐Elmer Spectrum 100 FTIR spectrometer (Perkin‐Elmer, Shelton, CT, USA). NMR spectra were acquired on a Bruker Avance DRX‐600 MHz spectrometer (Bruker, Rheinstetten, Germany), using methanol‐d4 (Merck Millipore, Johannesburg, South Africa) or deuterium oxide (D₂O) (Sigma‐Aldrich, Johannesburg, South Africa) as solvents. Solvents pyridine (Merck, ACS, Reag. Ph Eur), acetic anhydride (Sigma‐Aldrich, ≥99.5%, MerckGaA, Darmstadt, Germany), 5% HCl (2 M, 99.9% Riedel‐de Haen, France), dichloromethane (DCM, Riedel‐de Haen, France), and chemicals 4‐dimethylamino pyridine (DMAP, WGK Germany, Sigma‐Aldrich), benzoic acid (assay 99%, UniLABR, SAAR Chem, South Africa), and N, N′‐dicyclohexylcarbodiimide (DCC, Sigma‐Aldrich) were used during derivatization.

Sample Collection and Specimen Identification

The sponge specimens used in this study, MOZ95‐019 and TIC2022‐101, were collected from the Southern African coast. Specimen MOZ95‐019 was collected at a depth of 13 m by Philip Coetzee in 1995, near Ponta Do Ouro, Mozambique (coordinates: −26.77958°S, 32.8997°E). Specimen TIC2022‐101 was collected at a depth of 49–55 m by Dr. Gwynneth Matcher in 2022, from KwaZulu‐Natal, South Africa (coordinates: −28.9845°S, 32.71766°E). Both specimens were morphologically identified as belonging to the species Oceanapia ramsayi by Professor Shirley Parker‐Nance. The identification was based on the general characteristics of the sponge, the distribution of oscula and porefields, the partitioning of the choanosome, and the shape of the spicules. The voucher specimens are deposited at the South African Institute for Aquatic Biodiversity (SAIAB) in Grahamstown, South Africa.

Extraction, Isolation, and Structural Elucidation

The MeOH extract of MOZ95‐019 was sourced from the MDC extract library^{[ 29 ]} deposited at the Department of Biochemistry, Microbiology and Bioinformatics, Rhodes University, in the Marine Natural Product Research Lab. The extract of MOZ95‐019 was subjected to bioassay‐guided purification. A portion of this extract (7.3 g) was dissolved in 80% aqueous MeOH and separated into five fractions on a C₁₈ SPE cartridge (H₂O: MeOH, 80:20, 60:40, 50:50, 25:75, and 0:100). A portion of the active fraction (F2: 1 g) was further fractionated into fifteen subfractions through semiprep C₁₈UV‐HPLC (step gradient method, utilizing H₂O‐MeCN ranging from 90% aqueous MeCN to 70% MeCN + 0.1% v/v FA). Subfractions 8 and 9 were active and identified as isorhizochalin (1, 452.7 mg) and rhizochalin (2, 256.1 mg), respectively.

Frozen sponge material of TIC2022‐101 (86.4 g) freeze‐dried was exhaustively extracted with HPLC‐grade MeOH using a cold maceration technique,^{[ 30 ]} yielding 8.1 g of extract. A portion of the extract (1.5 g) was dissolved in MeOH and adsorbed to Sephadex‐LH20, then eluted with 100% MeOH, yielding five fractions. Fraction 2 (644.7 mg) was biologically active and further fractionated into eight subfractions using C₁₈ UV‐HPLC (step gradient method, utilizing H₂O‐MeCN ranging from 80% aqueous MeCN to 70% MeCN + 0.1% v/v FA). Among the subfractions, fraction 2 was biologically active and was identified as rhizochalin C (3, 56 mg). Furthermore, 1.3 g of crude extract was dissolved in 80% aqueous MeOH and separated into 15 fractions on C₁₈UV‐HPLC using the isocratic method (50% aqueous MeCN + 0.1% FA). This resulted in oceanapiside (4, 20.7 mg). Compounds 1–4 were characterized by 1D and 2D NMR, CD, and FTIR spectroscopies and MS data accompanied by literature reports.

Derivatization

Compounds 1–4 were further derivatized according to published methods^{[ 6} , ^{17 ]} to yield aglycons: isorhizochalinin (1b), rhizochalinin (2b), rhizochalinin C (3b), and oceanin (4b); tetraperbenzoyl isorhizochalinin (1c), tetraperbenzoyl rhizochalinin (2c), pentaperbenzoyl rhizochalinin C (3c), and pentaperbenzoyl oceanin (4c). Additionally, compound 2 was derivatized to bis‐oxazoline (2e), Figure 2. To provide structural evidence for compounds 1–4, we investigated the stereochemistry using the derivatives 1c–4c, and the data are consistent with published reports.^{[ 17} , ¹⁹ , ^{20 ]} Furthermore, the ketone functional group of compounds 1–4 was studied using peracetate derivatives 1a–4a employing in situ MS and confirmed the position of the ketone group to be at C‐18 as reported previously.^{[ 19} , ^{20 ]}

Biological Activity Screening: Cytotoxicity Activities

A resazurin assay was used to determine cytotoxicity and the resulting half‐maximal inhibitory concentrations (IC₅₀) values in the Caucasian MDA‐MB‐231 triple‐negative breast cancer (TNBC) cell line (ATCC: HTB‐26), African‐American HCC70 TNBC cell line (ATCC: CRL‐2315), the Caucasian MCF‐7 hormone receptor‐positive (ER+, PR+, HER‐2‐) breast cancer cell line (ATCC: HTB‐22) and a nontumorigenic breast epithelial cell line MCF12A (ATCC: CRL‐10782) when treated with compounds at a concentration range from 0.08 to 250.00 μM or a vehicle control of, 2% v/v DMSO for 96 h at 37 °C and 9% CO₂.^{[ 31 ]} The cells were seeded in a 96‐well plate at a density of 5.0 × 10³ cells/well and allowed to attach overnight at 37 °C and 9% CO₂. After a 96‐h treatment, 0.54 nM resazurin solution was added to the cells and incubated for 2–4 h at 37 °C in 9% CO₂. Fluorescence readings of the resulting solutions were obtained on the Synergy MX microplate reader (BioTek Instruments Inc., USA) with the excitation and emission wavelengths set at 560 and 590 nm, respectively. The experiment was performed in triplicate, and the data were analyzed using GraphPad Prism version 8.0.2, with IC₅₀ values determined by nonlinear regression.

Antibacterial Activities

Antibacterial activity was assessed using the resazurin assay method.^{[ 32 ]} Single colonies of P. aeruginosa ATCC 27853, S. aureus ATCC 12600, and E. coli ATCC 10536 were cultured overnight at 37 °C in Miller's modified Luria broth (LB) medium, while A. baumannii ATCC 19606 and K. pneumoniae ST512‐K30BO were cultured in Tryptic Soy Broth (TSB). Each strain was subsequently grown to log phase at 37 °C. To examine the antibacterial activities of the test fractions/compounds, a 96‐well microtiter plate was utilized. The MICs of the test compounds were investigated at twofold dilutions (100–0.78 μM). In each well, 20 μL of compowas 160 μL of LB broth, and 20 μL of mid‐log‐phase bacterial culture were combined to achieve a final volume of 200 μL. The bacterial suspensions were 1:1000 for P. aeruginosa, E. coli, and A. baumannii, K. pneumoniae, and 1:500 for S. aureus. The treatments were incubated at 37 °C for 6 h, followed by the addition of 20 μL of 0.13 mg mL⁻¹ resazurin and further incubated for 15–30 min. Fluorescence readings were taken at an excitation wavelength of 560 nm and an emission wavelength of 590 nm using a multiwell plate reader (TECAN Infinite M Nano+). MICs were conducted in duplicate, with novobiocin and nalidixic acid serving as positive controls. The MIC values were calculated using GraphPad Prism version 8.0.2, employing the nonlinear regression analysis.

Time of Kill Kinetics

Time‐kill assays against E. coli ATCC 10536 and S. aureus ATCC 12600 were conducted according to the protocol outlined by^{[ 33 ]} with modifications. Single colonies of bacteria were cultured in LB broth at 37 °C overnight. Fifty microliters of overnight cultures were inoculated in 5 mL LB broth, and cultures were grown to mid‐log phase. Twenty microliters of the mid‐log phase cultures were plated in 96‐well plates, and the final concentration of bacterial suspension in a 200 μL volume reaction mixture was ≈105 CFU mL⁻¹. Compounds and controls (nalidixic acid and novobiocin) were adjusted to a final concentration of 1x and 4x MIC. The 96‐well plates were then incubated at 37 °C. At time points 0.5, 1, 2, 3, 4, and 5 h, 200 μL from each treatment were transferred into a 1.5 mL microcentrifuge tube, centrifuged at 9200 × g for 2 min, and the supernatant discarded. The cells were resuspended in 1 mL of 10 mM MgSO₄, and 10‐fold serial dilutions were prepared as needed. 10 μL of each suspension was plated on LB agar and incubated at 37 °C for 16 h. The rate of killing was determined by visual inspection, where total inhibition of cell growth for each treatment was compared to the untreated control.

Membrane Disruption Assay

The Live/Dead BacLight bacterial viability test was carried out to determine whether the compounds interfere with the physiological integrity of bacterial cell membranes. The tests were carried out against E. coli and S. aureus using the Invitrogen LIVE/DEAD BacLightTM Bacterial Viability Kit (L7012) according to the manufacturer's instructions. Three hundred microliters of E. coli or S. aureus overnight cultures were inoculated in 30 mL of nutrient broth and grown to late log phase on a shaker at 37 °C, followed by centrifugation at 10,000 × g for 10 min to pellet the bacterial cells. The pellet was resuspended in 2 mL of 0.85% NaCl, and 1 mL of this suspension was added to 20 mL of 0.85% NaCl (for Live bacteria) or 20 mL of 70% isopropyl alcohol (for killed bacteria). For the test compounds and DMSO control, 50 μL of bacterial suspension was added to 0.95 mL of test compounds in 0.85% NaCl, resulting in a final concentration of 50 μM. The contents were incubated at room temperature for 1 h (mixing every 15 min), then centrifuged at 10,000 × g for 10 min. The pellets were washed with 20 mL of 0.85% NaCl for the live/dead controls and 1 mL for the pellets treated with the compounds, then centrifuged at 10,000 × g for 10 min. The pellets were resuspended in 10 mL of 0.85% NaCl for the live/dead controls and 0.5 mL for the pellets treated with the compounds. Exactly 300 μL of these suspensions were transferred to the 96‐well plates, and OD₆₇₀ readings were taken using a microtiter plate reader (Tecan, Infinite 200 PRO, USA). E. coli bacterial suspensions (live, killed, and treated) were adjusted to 2 × 10⁸ bacteria mL⁻¹ (0.06 OD₆₇₀), and the suspension of S. aureus was adjusted 10‐fold less than E. coli suspensions (2 × 10⁷, 0.3 OD₆₇₀). In order to generate a standard curve, various proportions of live (untreated cells)/dead (cells treated with absolute isopropanol) cells were prepared as follows: 0:100, 10:90, 50:50, 90:10, and 100:0 (%). Exactly 100 μL of adjusted bacterial suspensions were transferred to the flat‐bottomed 96‐well microtiter plates in duplicates, followed by 100 μL of 2x staining solution (mixture of SYTOTM 9 and propidium iodide stains), and the contents were further incubated at room temperature in the dark for 15 min. The excitation/emission maxima of SYTO 9 stain were recorded at 485/530 nm (green), and propidium iodide stain was recorded at 485/630 nm (red) using the microtiter plate reader (Tecan, Infinite 200 PRO, USA). The ratio of green/red fluorescence was calculated (RatioG/R = SYTO 9 readings/Propidium iodide readings), and a linear graph of RG/R versus live bacteria was plotted. The regression formula generated from the standard curve was used to calculate the percentage of live bacteria treated with the compounds. A reduction in the viability of cells signified that the test compounds elicit antibacterial activity by damaging the cell membrane.

Supporting Information

Supporting Information related to this article can be found in the online version at Abdissa et al. 2025 Supplementary information.docx.

Conflict of Interest

The authors declare no conflict of interest.

Supporting information

Supplementary Material

Keneni Dele Abdissa, Njanje Idris, Kalinski Jarmo‐Charles Julian, Isaacs Michelle, Swart Tarryn, Davison Candace, de la Mare Jo‐Anne, Dorrington Rosemary Ann, ChemMedChem, 2025, 0, e202500392. 10.1002/cmdc.202500392

Data Availability Statement

The data that support the findings of this study are available in the supplementary material of this article.