The fungal Xylariaceae family includes more than 16 genera and 130 species (Sánchez-Ballesteros et al., 2000) and has been extensively investigated for the chemo diversity and biological activity of their metabolites (Stadler et al., 2006, 2008). Among the 16 genera reported, Hypoxylon with 14 species is largely distributed in various marine and terrestrial habitats, and producing a large variety of bioactive compounds among which cohaerins (Quang et al., 2005a; Surup et al., 2013), daldinins and daldinones (Quang et al., 2004; Gu et al., 2007), cytochalasin (Espada et al., 1997), fragiformin (Stadler et al., 2006), mitorubrinols (Quang et al., 2005b), hypoxylonols (Fukai et al., 2012), hypoxylans (Kuhnert et al., 2015a), hypoxyvermelhotins (Kuhnert et al., 2014), rickenyls (Kuhnert et al., 2015b), rutilins (Quang et al., 2005b), carneic acids (Quang et al., 2006), hymatoxins (Bodo et al., 1987; Borgschulte et al., 1991), malettinins (Angawi et al., 2005), hypoxysordarin (Daferner et al., 1999), lenormandins (Kuhnert et al., 2015c), nodulisporic acids (Bills et al., 2012), schweinitzin A (Linh et al., 2014), truncatones (Sudarman et al., 2016), macrocyclic polyesters 15G256 family (Schlingmann et al., 2002), and sporothriolide (Krohn et al., 1994; Surup et al., 2014; Cao et al., 2016).

Sporothriolide belongs to the furofurandione family of natural compounds first published in 1994 (Krohn et al., 1994). This compound exhibits antifungal activity and benefits from substantial synthetic efforts (Sharma and Krishnnudu, 1995; Yu et al., 2001; Fernandes and Ingle, 2009; Ishihara et al., 2014). The name sporothriolide is related to Sporothrix sp. Hektoen and Perkins (strain 700), from which this compound was first isolated. Sporothrix genus belongs to a different ascomycete family, ophiostomataceae. The first report on sporothriolide in 1994 detailed both the structure and bioactivity of this product (Krohn et al., 1994). It shows that Sporothrix produces sporothriolide, dihydrosporothriolide, as well as various sporothriolide analogs with different side-chain length (canadensolide, discosiolide, avenaciolide, ethiosolide). The authors reported also the antifungal/herbicidal activities of these compounds (Krohn et al., 1994) (Figure 1).

Twenty years later, sporothriolide and dihydrosporothriolide were isolated from Hypoxylon monticulosum together with three monocyclic acid precursors: sporothric acid, isosporothric acid and dihydroisosporothric acid (Figure 1) (Surup et al., 2014). More recently, sporothriolide was isolated from Nodulisporium sp., an anamorph of Hypoxylon, and the herbicidal activity was confirmed (Cao et al., 2016).

In our previous contribution, we added to the scarce sporothriolide family two new compounds, deoxysporothric acid and a new complex architecture sporochartine A, combining sporothriolide and trienylfuranol A moieties. The trienylfuranol A was recently isolated from a different Hypoxylon submoniticulosum (Burgess et al., 2017).

In the present work we report four new sporochartines B to E. Their structures were elucidated using 1D and 2D NMR, HRMS, IR and comparison with sporochartine A data, for which the absolute configuration was previously established by X-Ray analysis. A Diels-Alderase type reaction is probably involved in the biosynthesis of the five isolated sporochartines, as discussed below.

The human cancer cell-lines cytotoxicity bioassay shows that the conformation of sporochartines has an impact on the biological activity.

According to our previous work, sporochartine A (1) was obtained after 5 days cultivation of H. monticulosum CLL-205 in PDB broth (Leman-Loubière et al., 2017). By extending the cultivation of the same microorganism by a further 4 days, the ethyl actetate extract gives the HPLC chromatogram presented in Figure 2.

Sporochartines B–D were isolated as white powders by flash chromatography followed by semi-preparative HPLC. They had similar [M+H]⁺ HRESIMS molecular weights, molecular formula C₂₄H₃₄O₆ and IR spectra compared with sporochartine A (1) (Table 2) (Leman-Loubière et al., 2017). The ¹H and ¹³C NMR spectra of sporochartines B–D were similar to those of compound 1 (Tables 1, 2). Optical rotations [α]25D, IR bands and HRESIMS are reported in Table 2.

COSY and HMBC spectra, confirmed that sporochartines A–D had the same connectivities supporting similar planar scaffold (Figure 3). In addition, the common coupling constant of 15.4–15.6 Hz between H-18 and H-19 confirmed that the double bond C-18/C-19 is in E configuration.

Based on the previously reported absolute configuration of sporochartine A (1) and ROE correlations, we deduced the absolute configuration of sporochartines B–D (2–4) (Figure 4).

The common ROE correlations between H-2 and H-5 and between H-5 and H-6 requiring a cis orientation of these three protons was found in the sporochartine A–D. Therefore, the stereochemistry of the sporothriolide moiety was identical. Moreover, based on ROE correlations between H-20 and H-21 and between H-21 and H-23, the strerochemistry of the tetrahydrofurane moiety is also a common feature in sporochartines A–D.

For sporochartine B (2) (Figure 4), we did not observe ROE correlations between H-17 and H-2 and between H-17 and H-14b as in sporochartine A (1), while a new correlation is observed between H-17 and H-13. This data suggests that the carbon C-17 have opposite stereochemistry compared to 1 supporting a 3S,17R configuration of 2 (instead of 3S,17S in 1).

For sporochartine C (3) (Figure 4), the H-17/H-2 and H-17/H-14b correlations observed in sporochartine A (1) are absent. In addition, we observed a correlation between H-17 and H-13 and H-2 and H-14a in compound 3. Based on this data, we suggest that compound 3 has a 3R,17S configuration.

Sporochartine D (4) (Figure 4) conserved the correlations between H-17 and H-2 and between H-17 and H-14b reported for sporochartine A (1). Furthermore, the correlation between H-17 and H-13 is absent in both 4 and 1. In 4 we have an additional correlation between H-13 and H-2, absent in 1. These observations support the conclusion that 4 is the 3R,17R isomer of 1.

A new compound referred as sporochartine E (5) was also isolated as a white powder. Compound 5 has the same molecular formula C₂₄H₃₄O₆ as compound 1, deduced from HRESIMS m/z [M+H]⁺ 419.2433. Here again we have eight degrees of unsaturation accounting for two γ-lactones, two double bonds, one six-membered cycle moiety and one tetrahydrofurane moiety.

Compound 5 has a terminal methylene group (at δ_C 119.6, δ_H 5.27 and δ_H 5.21) while the tetrahydrofuran moiety connected to C-19 in 1 is connected to C-14 in 5.

Based on COSY correlations (Figure 5), the sporothriolide moiety was the same in compound 5 as in 1. Moreover, COSY correlations from H-13 to H-19 through H-14 (δ_H 3.25), H-15 (δ_H 6.16), H-16 (δ_H 5.56), H-17 (δ_H 2.80) and H-18 (δ_H 5.65) together with HMBC correlation between H-13 and C-3 and H-17 and C-3 formed a cyclohexane fragment like in 1. Finally, by using the COSY correlations from H-24 (δ_H 1.34) to H-20 (δ_H 3.22), through H-23 (δ_H 3.96), H-22 (δ_H 1.53 and 2.39) and H-21 (δ_H 4.28) we deduce the tetrahydrofuran moiety. The HMBC correlations between H-20 and C-14 and C-15 allowed us to connect this tetrahydrofurane moiety to the sp³ methine C-14.

The absolute configuration of sporochartine E (5) was suggested using ROE correlations compared to the absolute configuration of sporochartine A (1) (Figure 6).

ROE correlations between H-2 and H-5/H-6 in the sporothriolide moiety and the ROE correlation between H-20 and H-21 and between H-23 and H-21 in the tetrahydrofuran moiety indicated a similar to that in 1.

Sporochartine E (5), showed a correlation between H-17 and H-2, like in compounds 1 and 4. H-2 also exhibited a correlation with H-13b but not with H-14. This suggests that C-3 and C-17 has the same relative configuration than 4. For C-14, we observed ROE correlations between H-14 and H-13a, H-13a, and H-17 and H-14 and H-24, suggesting a 3R, 14S, 17S configuration for compound 5.

Based on the structure of sporothriolide and the recently reported trienylfuranol A isolated from H. submoniticulosum, we suggested a hypothetic biosynthetic pathway of sporochartines, involving a “spiro” Diels-Alderase reaction as shown in Figure 7 (Klas et al., 2015; Byrne et al., 2016). The possibility of a non-enzymatic catalysis was excluded as reported previously (Leman-Loubière et al., 2017).

The cytotxicity of sporochartines was evaluated on three human cancer cell lines, HCT-116 (human colon carcinoma), PC-3 (prostate cancer cell lines) and MCF-7 (breast cancer cell line). The results presented in Table 3 are highly contrasting but nevertheless clearly indicate the impact of sporochartine conformation on the bioassay results.

Thus, sporochartine C (3) is toxic against the three cell lines with IC₅₀ ranging from 7.2 to 21.5 μM. In contast, sporochartine A (1) is totally inactive at concentrations higher than 100 μM. This may be due to the substantial difference in the spatial conformation of compounds 1 and 3 (Figure 4).

The lower IC₅₀ values were recorded for different sporochartines and against different cell lines, sporochartine B (2) for MCF-7 (2.28 μM), sporochartine C (3) for HCT-116 (7.2 μM) and sporochartine E (5) for PC-3 (5.96 μM).

Our future efforts will focus on the cytotoxic profile, biosynthesis and synthesis of sporochartines. The cytotoxicity profile reveals a non-cytotoxic sporochartine A (1), a large spectrum cytotoxic sporochartine C (3) and more cell line specific sporochartines B (2), D (4), and E (5). This finding merits future investigation on the mechanisms of action of these new scaffolds of cytotoxic compounds.

The biosynthesis of sporochartines, and the biosynthesis of its two moieties, sporothriolide and trienylfuranol A are still unknown. This opens new and promising opportunities for the discovery of novel biosynthetic microbial clusters.

Finally, having in hand hundreds of milligrams of sporothriolide, the hemi-synthesis of sporochartines is currently in progress based on a final Diels-Alder connection. The selectivity of the chemical catalysis and the proportion of different isomers will be compared to the microbial counterpart. According to our expertise in biocatalysis-based chemodiversification of natural compounds (Adelin et al., 2011; Martins et al., 2015), sporothriolide will be submitted to a panel of microorganisms in order to pursue the enrichment of sporothriolide related compounds.

Detailed 1D and 2DNMR, MS and IR spectra of sporochartines are available free of charge via the Internet at http://pubs.acs.org.

CL-L: microbiologie chemistry; GL: chemistry; CD: invetebrate investigation; JO: head of the team and science manager.