Rare actinomycetes are attractive resources of secondary metabolites that show biological activities with unprecedented chemical structures.1, 2 Genus Saccharothrix is a rare actinomycete that has furnished an increasing number of bioactive metabolites.3, 4, 5, 6, 7 Our recent studies of Saccharothrix sp. A1506 have afforded saccharothriolides A–F (3–8) (Supplementary Figure S1).8, 9 Saccharothriolides possess unique phenyl-substituted 10-membered macrolide structures, and some exhibited moderate cytotoxicity to human fibrosarcoma HT1080 cells. Since they contain a variety of substituents at C-7, we expected the presence of ‘precursor A’, with an α,β-unsaturated ketone that can function as a Michael acceptor.8 However, to date, our attempts to isolate precursor A from culture broth have not been successful, probably because of its high reactivity. Instead, we have investigated the presence of precursor A by modifying the culture conditions. When we removed tryptophan from the culture media, precursor A was still not observed in the LC-MS analysis, and moreover, no production of saccharothriolides A (3) and B (4) was observed (Figure 1b). As expected, saccharothriolides A (3) and B (4) were generated when aryl amines, predicted metabolites of tryptophan, were added as nucleophiles to the culture media (Figure 1). Additionally, new congeners saccharothriolides G (1) and H (2) were obtained when we used 2-methoxyaniline (o-anisidine) as a nucleophile. Herein, we report the precursor-directed in situ synthesis (PDSS) of saccharothriolides including new congeners, and their biological activities.

Figure 1
figure 1

Structures of saccharothriolides A (3) and B (4) and their biosynthesis. (a) Plausible biosynthetic pathway of saccharothriolides. Precursor A is a Michael acceptor, to which anthranilic acid or 2-aminophenol attacks to yield saccharothriolide A (3) and B (4), respectively. (b) Production of saccharothriolides A (3) and B (4) in various culture conditions. LC-MS analyses were examined for crude extracts (PEGASIL ODS SP100, φ 3 × 250 mm, 0.2 ml min−1, 40% aq. MeCN with 0.1% formic acid). Total ion chromatogram (TIC) and ion chromatograms for 486.2128 (up) and 458.2179 (down) (corresponding to [M+H]+ for 3 and 4, respectively) are shown. A full color version of this figure is available at The Journal of Antibiotics journal online.

Full size image

Saccharothriolides A (3) and B (4) are expected to be Michael addition products of precursor A and the amino aryl groups, anthranilic acid and 2-aminophenol, respectively. The biosynthetic origin of anthranilic acid and 2-aminophenol have been proposed to be the tryptophan which is abundant in the culture medium.10, 11, 12 To avoid the effect of tryptophan on the production of main product 3, we supplemented tryptophan-free media with the nucleophilic reagents anthranilic acid (10 g l−1) or 2-aminophenol (10 g l−1). There were obvious changes in the LC-MS profile, suggesting the production of 3 or 4 (Figure 1). These results supported the presence of precursor A as a Michael acceptor, and indicated that we can explore in situ synthesis to obtain further saccharothriolide analogs by simply adding nucleophilic substituents.

Previous structure–activity relationship (SAR) studies using saccharothriolides A–F (3–8, Supplementary Figure S1) revealed the importance of the substituent at C-2″ on their cytotoxicity;8, 9 metabolites possessing an alcohol group showed activity, while those possessing a carboxylic acid were less potent. To investigate the necessity of the free hydroxy group, we planned to obtain 2-methoxyaniline-substituted saccharothriolide analogs by PDSS. Saccharothrix sp. A1506 was cultivated in 3 l of tryptophan-free medium for 4 days, followed by the addition of 2-methoxyaniline (10 g l−1) and acetone (1:1 v/v), then shaken at 4 °C for 1 day. The reaction mixture was extracted, and LC-MS-guided fractionation was carried out to yield two new analogs, saccharothriolide G (1) and its C-2 epimer saccharothriolide H (2) (Figure 2a).

Figure 2
figure 2

Structure elucidation of saccharothriolide G (1). (a) Chemical structures of saccharothriolides G (1) and H (2). (b) 1H-1H COSY (left, bold) correlations, and selected HMBC (left, arrow) and NOESY (right) correlations for saccharothriolide G (1). (c) Experimental ECD spectra of saccharothriolide G (1) (solid line) and saccharothriolide B (4) (dotted line).

Full size image

Saccharothriolide G (1) was obtained as a light-yellow oil with [α]20D –81.6 (c=0.31, MeOH). The molecular formula was determined to be C26H33NO7 by HR-ESI-MS (m/z 472.2321 [M+H]+, calcd 472.2335), revealing that its molecular size was 14 Da larger than saccharothriolide B (4). 1H and 13C NMR data resembled those of 4 (Supplementary Table S1, Supplementary Figure S2), except for the presence of a methoxy group (δH 3.92, δC 56.5) and chemical shift differences in the amino aryl signals. The downfield shift of C-2″ (δC 148.8 in 1, δC 146.0 in 4) and upfield shift of C-3″ (δC 111.4 in 1, δC 115.2 in 4) suggested that 1 possessed a methoxy group instead of a phenolic hydroxy group at C-2″. Installation of 2-methoxyaniline (o-anisidine) was unambiguously determined by HMBC correlations from H-6″ to C-2″, H-3″ to C-1″, and from the methoxy proton 2″-OCH3 to C-2″, along with the 1H-1H COSY correlations from H-3″ to H-6″. The planar structure of the molecule was deduced by the COSY, HMQC and HMBC data (Figure 2b). The 1H-1H COSY data revealed the presence of three spin systems: CH3-10/H-2/H-3/H-4/CH3-11, CH3-12/H-6/H-7/H-8/CH3-13, and H-3″/H-4″/H-5″/H-6″ (Figure 2b). HMBC correlations from CH3-11/CH3-12 to C-5 connected C-4 and C-6 through a ketone group. HMBC correlations from CH3-10 to carbonyl C-1, and from H-9 to CH3-13/C-1, combined with the down-field shifted chemical shift of H-9 (δH 5.49), connected C-2 and C-9 through an ester bond, leading to the formation of the 10-membered lactone ring (Figure 2b). The meta-disubstituted benzene ring was connected to C-9 based on the HMBC correlations from H-9 to the aromatic carbons C-2′/6′ and C-1′.

The relative stereochemistry of 1 was determined by the NOESY and 3JH-H data (Figure 2b). NOESY cross peaks between H-2 and H-4, and between H-3 and H-2/CH3-10/H-4/CH3-11, indicated that H-2, -3, and -4 are on the same α face. NOESY correlations between H-6 and CH3-13/CH3-11 indicated that H-6 and H-8 have α and β configurations, respectively. The NOESY correlations between H-7 and H-6/H-8/CH3-12/CH3-13 indicated that H-7 also has an α configuration, while the β-orientation of H-9 was revealed by the correlations between the aromatic proton H-6′ and H-8/CH3-13. This result was also supported by the similar 3JH-H values between 1 and 4. Thus, the relative configurations were deduced to be 2R*, 3R*, 4S*, 6R*, 7R*, 8R*, and 9S*. The absolute stereochemistry of 1 was determined by comparing the ECD spectra of 1 and 4. The ECD spectrum of 1 showed characteristic Cotton effects at 212 (Δɛ, −29.3), 252 (Δɛ, +15.9), and 296 (Δɛ, −6.5) nm, which overlapped well with the spectrum of 4 (Figure 2c). Thus, the absolute configuration of 1 was established to be 2R, 3R, 4S, 6R, 7R, 8R, and 9S.

Saccharothriolide H (2) was obtained as a light yellow oil with [α]20D –81.2 (c=0.05, MeOH). The molecular formula was determined to be C26H33NO7 by HR-ESI-MS (m/z 472.2325 [M+H]+, calcd 472.2335), being the same as that of saccharothriolide G (1). 1H and 13C NMR data of 2 resembled those of 1, while differences were observed for the chemical shifts of H-2, H-3, CH3-10, and CH3-11 in the right half of the lactone ring (Supplementary Table S1). Notably, the chemical shifts of the lactone ring were identical to those of saccharothriolide E (7), a C-2 epimer of 4. From these results, 2 was deduced to be an epimer of 1 at C-2. The absolute stereochemistry of 2 was proven to be 2S, 3R, 4S, 6R, 7R, 8R and 9S, the same as that of 7 on the basis of the coupling constants and optical rotation values.9

We previously investigated the SAR of saccharothriolides A–F (3–8) which revealed that substitution at C-7 affects the cytotoxicity against human fibrosarcoma HT1080 cells.8, 9 Saccharothriolides B (4) and E (7), both of which have a 2-aminophenol group at C-7, exhibited moderate cytotoxicity (IC50 values, 13.9 and 29.2 μM, respectively). When the C-7 substituent was anthranilic acid or a hydroxyl group, that is, saccharothriolides A (3), D (6) or C (5), compounds were inactive even at 100 μM. To further analyze the structure-activity relationship of these 10-membered macrolides, we examined the cytotoxicity of saccharothriolides G (1) and H (2) against HT1080 cells. Saccharothriolide G (1) exhibited weak activity (IC50 value, 53.5 μM), making it less potent than its C-2 epimer saccharothriolide H (2) (IC50 value, 24.8 μM). These results not only confirmed the importance of the phenolic hydroxyl group at C-2″ in the observed cytotoxicity, but also suggested that the cytotoxicity was sensitive to the stereochemistry of C-2.

In summary, two new Michael addition products, saccharothriolides G (1) and H (2), were synthesized in situ and isolated. Their chemical structures were spectroscopically determined and their cytotoxicities were evaluated to determine the importance of the free hydroxyl group at C-2″ and the stereochemistry of C-2. Further precursor-directed in situ synthesis (PDSS) and isolation of precursor A are ongoing in our laboratory.