Within marine natural products, polyether metabolites encompass a unique class displaying a great diversity of structures and a broad array of bioactivities, acting as enzymatic inhibitors, ion channel blockers, stimulating neurotransmitters release or showing potent cytotoxic activities [1–4]. One important group of squalene-derived polyethers, has been isolated from red algae of the genus Laurencia and Chondria, (Rhodomelaceae family), sponges of the Axinellidae family, and from some mollusks [3–5]. Undoubtedly, the main source of these metabolites is the red alga Laurencia viridis, endemic algae of the Canary Islands that grow on basaltic rocks in the lower intertidal zone at early spring when the coastal temperature is about 18 °C [5,6]. Dehydrothyrsiferol (1), the major metabolite of this class, has shown important pharmacologic properties such as potent cytotoxic effects, protein phosphatase type 2A inhibition and integrin antagonist activity [7–11].

Further investigation of the minor constituents present in the fresh alga L. viridis led to the isolation of four novel polyethers, iubol (2), 22-hydroxy-15(28)-dehydrovenustatriol (3), 1,2-dehydropseudodehydrothyrsiferol (4) and secodehydrothyrsiferol (5). Herein, we describe the isolation and structure determination of these compounds as well as their cytotoxic effects in several cancer cell lines (Figure 1).

Iubol (2), was isolated as an amorphous white solid, [α]²⁵ _D +14.8 (c 0.27, CHCl₃). The molecular formula, C₃₀H₅₁O₆Br, determined by ESI-HRMS (m/z 609.3679/611.3607, [M + Na]⁺), turned out to be identical to that found for dehydrothyrsiferol (1) [7,12]. The structure of 2 was mainly determined from the analysis of its NMR data. Interpretation of its ¹H and ¹³C NMR spectra together with the examination of the HSQC experiment showed that 2 consists of seven methyls, eleven methylenes, six methines and six quaternary carbons. Analysis of the ¹H–¹H COSY spectra enabled us to set five independent proton-proton spin systems, in a similar way to that found in other polyethers isolated from L. viridis [5,7,8]. Specifically, the fragment C-1→C-18 of 2 closely resembles that of 1, with the presence of the same four spin systems: Fragment I [H-3 (δ_H 3.89), H₂-4 (δ_H 2.10/2.24), H₂-5 (δ_H 1.53/1.80)]; Fragment II [H-7 (δ_H 3.08), H₂-8 (δ_H 1.45/1.72), H₂-9 (δ_H 1.50/1.77)]; Fragment III [H-11 (δ_H 3.42), H₂-12 (δ_H 1.63/1.78), H₂-13 (δ_H 1.82/2.03), H-14 (δ_H 4.27)] and Fragment IV [H₂-16 (δ_H 2.12/2.46), H₂-17 (δ_H 1.38), H-18 (δ_H 3.30)] (Figure 2; Table 1). On the other hand, notable differences were found towards the terminal moiety of the molecule (C-19→C-24), where the COSY spectrum revealed coupling between the characteristic proton H-22 (δ_H 3.38) and H₂-21 (δ_H 1.75) and these sequentially to both H₂-20 (δ_H 1.50/1.79). Finally, the structure of 2 was unambiguously assigned by the HMBC correlations. Thus, significant correlations between methyl H₃-29 (δ_H 1.18) with C-18 (δ_C 77.0), C-19 (δ_C 76.1) and C-20 (δ_C 27.9), as well as those of methyls H₃-24 (δ_H 1.24) and H₃-30 (δ_H 1.25) with C-22 (δ_C 75.2) and C-23 (δ_C 75.7), were observed (Figure 2).

The high coincidence between the ¹H and ¹³C NMR data of the C-1→C-14 fragment in 2 with those observed for 1 indicated that both compounds share the same relative configuration at C-3, C-6, C-7, C-10, C-11 and C-14 (Table 1). This conclusion was further supported by the analysis of the NOESY experiment, where identical correlations between 1 and 2 were observed. In particular, those connecting H-3 with H₃-1 (δ_H 1.26), H-4α (δ_H 2.10) and H-5α (δ_H 1.53); H₃-25 (δ_H 1.39) with H₃-26 (δ_H 1.20) as well as H-11 with H-7 and H-14 together with those of H₃-27 (δ_H 1.22) with H-8β (δ_H 1.45) and H-12β (δ_H 1.63) secured the relative configuration of this fragment. With regard to the C-18→C-24 fragment, when H-22 (δ_H 3.38, dd, J = 5.1, 10.5 Hz) was selectively irradiated in the 1D-gNOESYexperiment, only H₃-29 and H₃-24 showed enhancements indicating that all protons are on the same side of the tetrahydropyran ring. From these coupling constant values and NOESY correlations we concluded that only a twist-boat conformation with 19R*, 22R* relative configuration is possible for the D ring, as shown in Figure 3.

From a biogenetic viewpoint, most of the polyoxygenated squalene-derived ethers isolated from Laurencia species seem to arise from a common precursor: the (10R,11R)-squalene 10,11-epoxide isolated from L. okamurai. This compound can evolve to (6S,7S,10R,11R,14R,15R,18S,19S)-squalene tetraepoxide as a common intermediate [5]. Therefore, the new tetrahydropyran ring found in 2 could be biosynthesised from the diepoxide fragment 18S, 19S, 22R by protonation of the (20R)-epoxide followed by hydroxylation at C-23 yielding 18S, 19S-epoxy-20R-hydroxy as intermediate. The subsequent cyclization process with configuration inversion at C-19 would give the 18S*, 19R*, 22R* relative configuration as illustrated in Figure 4. Importantly, this hypothesis is consistent with the experimental data used to proffer the relative configuration of C-19 and C-22 offering simultaneously a solution for the configuration of C-18.

22-Hydroxy-15(28)-dehydrovenustatriol (3), was obtained as an amorphous white solid, [α]²⁵ _D +15.2 (c 0.07, CHCl₃). The molecular formula of 3, C₃₀H₅₁O₇Br, deduced by ESI-HRMS (m/z 625.2695/627.2723, [M + Na]⁺), indicated the presence of an additional oxygen atom with respect to dehydrothyrsiferol (1) [7,12]. Comparison of the ¹H and ¹³C NMR data of 3 with those reported for 1 clearly revealed that 3 shares the same carbon skeleton with 1, but contains an additional oxygen atom at the C-19→C-24 moiety (Table 2). Thus, a detailed analysis of the ¹H–¹H COSY and HSQC spectra confirmed the existence of the same four ¹H–¹H spin systems (Fragments I–IV) observed in the C-1→C-18 portion of 1 (Figure 5). However, in contrast with 1, the terminal moiety of the molecule contains only a four-proton spin system formed by H₂-20 (δ_H 2.00) and H₂-21 (δ_H 1.80/1.91), missing the characteristic H-22 signal. Finally, we used the HMBC experiment to connect the new and characteristic ¹³C signal of C-22 (δ_C 112.5) with H₂-20 (δ_H 2.00), H₃-24 (δ_H 1.29) and H₃-30 (δ_H 1.31). In addition those correlations observed between H₃-29 (δ_H 1.44) and C-18 (δ_C 84.8), C-19 (δ_C 86.9) and C-20 (δ_C 29.3) allowed us to build up the structure of the C-18→C-24 portion of 3.

Comparison of the ³J_H–H coupling constants, ¹H and ¹³C chemical shifts together with a detailed analysis of the NOESY experiment established the relative configuration of the C-1→C-14 fragment in 3 as identical to dehydrothyrsiferol (1). However, the absence of NOE correlations between H₃-29 with H₃-24 and H₃-30, excluded any definitive conclusion although it clearly suggests a trans configuration for this oxolane ring. The relative configuration of this moiety can be proposed making use of a biogenetic hypothesis. Thus, the presence of a hemiacetal group at C-22 can be rationalized via an intermediate with a carbonyl group at C-22 and by the subsequent cyclization as represented in Figure 4. Again, assuming that the biogenesis starts with a common precursor for all the oxaqualenoids isolated from Laurencia, the most plausible configuration for the stereogenic centers of this fragment should be 18R*, 19S* and 22R*.

Two additional related compounds, 1,2-dehydropseudodehydrothyrsiferol (4) and secodehydrothyrsiferol (5) were isolated in this study. Analysis of their ESI-HRMS spectra allowed us to establish their molecular formulas as C₃₀H₅₀O₆ and C₃₀H₅₀O₈ respectively. Comparison of the ¹H and ¹³C chemical shifts of 1,2-dehydropseudodehydrothyrsiferol (4) with other related compounds revealed great similarities with those reported for pseudodehydrothyrsiferol (6) except at the C-1→C-3 portion of the molecule (Tables 2 and 3). In fact, analysis of the 2D NMR spectra led us to the same carbon skeleton. However, the observation of HMBC correlations between the methylene protons H₂-1 (δ_H 4.77/4.99) with C-2 (δ_C 145.3), C-3 (δ_C 83.2) and C-25 (δ_C 17.2) clearly indicated that the differences between 4 and 6 were due to the dehydration of the hydroxyl group attached to C-2 in 6, that yielded an exocyclic double bond at C-1→C-2 in 4 (Figure 6). With regard to secodehydrothyrsiferol (5), the NMR data showed changes with respect to 6 around the C-1→C-6 fragment. The ¹H NMR spectrum revealed the existence of two new methyl groups as well as one methylene located at α-carbonyl positions. In addition, two new carbonyl groups (one ketone and one ester) were easily detected from the ¹³C NMR spectrum. Therefore, the structure of 5 was determined on the basis of the connectivity observed between H₂-4 (δ_H 2.50) and H₂-5 (δ_H 2.16/2.26) in the COSY experiment together with the HMBC correlations of H₃-25 (δ_H 2.15) and H₂-4 with C-3 (δ_C 208.5), those of H₃-1 (δ_H 1.98) with C-2 (δ_C 170.1) and C-6 (δ_C 84.1) as well as those of H₃-26 (δ_H 1.39) with C-5 (δ_C 29.3), C-6 (δ_C 84.1) and C-7 (δ_C 80.0). Comparison of the ¹H and ¹³C chemical shifts, ³J_H–H coupling constants, together with an analysis of the ROESY experiments of 4 and 5 showed very similar results compared to those of compound 6, and therefore the same relative configuration is proposed for the new compounds. An important feature in 4 and 5 is the absence of the bromine atom in ring A as it is common in the thyrsiferol or venustatriol series. Thus, the biogenetic origin of both compounds could be rationalized by attack of the oxane oxygen to the C-3 position that bears a bromine atom via a 5-exo process, giving an oxonium ion intermediate that would evolve as shown in Figure 7 to yield both metabolites [5]. The biogenetic origin of 5 could be rationalized by a Wagner-Meerwein rearrangement. These rearrangements have been confirmed experimentally in many cases and are almost always consistent with the participation of an oxonium ion intermediate [13,14]. In order to confirm the previous proposal, we induced this rearrangement in dehydrothyrsiferol (1). Therefore, 1 was treated with AgNO₃ using either methanol:water (1:1) or anhydrous acetonitrile as solvents. In the first case, using a methanol:water (1:1) mixture, the reaction yielded pseudodehydrothyrsiferol (6) as well as the methoxy derivatives at C-2 and C-3 positions. The relative yields of this reaction could be modulated changing the methanol:water ratio. On the other hand, when an aprotic solvent was used, three compounds showing m/z of 506 were identified by LC-MS although the main product was 1,2-dehydropseudodehydrothyrsiferol (4) [15,16].

This class of polyether squalene natural products shows a variety of biological properties where its cytotoxicity against breast cancer cells is highly significant [5,9,17]. Therefore, the in vitro cytostatic activity of 1–5 was assessed by XTT assays [18], using several human cancer cell lines, including Jurkat (human T-cell acute leukemia), MM144 (human multiple myeloma), HeLa (human cervical carcinoma), and CADO-ES1 (human Ewing’s sarcoma). As shown in Table 4, Jurkat leukemic cells were the most sensitive cells to the tested polyether compounds. In particular, iubol (2), 22-hydroxy-15(28)-dehydrovenustatriol (3), and secodehydrothyrsiferol (5) showed the highest effectiveness against this cell line (IC₅₀, 2.0–3.5 μM). The capacity of the above compounds to inhibit cell proliferation was likely due to their ability to induce apoptosis, as assessed by the appearance of a sub-G₁/G₀ subpopulation in cell cycle analysis, indicative of DNA breakdown [19], following incubation of Jurkat cells with 5 (Figure 8).

Interestingly, the substitution of an H atom in 1 by a hydroxyl group in 3 at C-22 highly potentiated the cytostatic activity (Table 4). It is also noteworthy that all the above compounds were active against the CADO-ES1 human Ewing’s sarcoma cell line (Table 4) in the range of 10–12 μM, suggesting that this type of compound could be suitable in the search for therapeutic lead compounds against both sarcoma and T-cell leukemia malignancies. Doxorubicin, used as a positive control rendered IC₅₀ values in the order of 10⁻⁷–10⁻⁸ M (data not shown).