Seco-Tetracenomycins from the Marine-Derived Actinomycete Saccharothrix sp. 10-10

Six new tetracenomycin congeners, saccharothrixones E–I (1–5) and 13-de-O-methyltetracenomycin X (6), were isolated from the rare marine-derived actinomycete Saccharothrix sp. 10-10. Their structures were elucidated by spectroscopic analysis and time-dependent density functional theory (TDDFT)-electronic circular dichroism (ECD) calculations. Saccharothrixones G (3) and H (4) are the first examples of tetracenomycins featuring a novel ring-A-cleaved chromophore. Saccharothrixone I (5) was determined to be a seco-tetracenomycin derivative with ring-B cleavage. The new structural characteristics, highlighted by different oxidations at C-5 and cleavages in rings A and B, enrich the structural diversity of tetracenomycins and provide evidence for tetracenomycin biosynthesis. Analysis of the structure–activity relationship of these compounds confirmed the importance of the planarity of the naphthacenequinone chromophore and the methylation of the polar carboxy groups for tetracenomycin cytotoxicity.


Introduction
Aromatic polyketides constitute a large group of structurally diverse natural products biosynthesized by type II polyketide synthases (PKS II) [1]. Many of these natural products have been widely used as antibacterial, antifungal, and anticancer agents [2,3]. Based on the polyphenolic ring systems and their biosynthetic pathways, bacterial aromatic polyketides are classified as anthracyclines, angucyclines, aureolic acids, tetracyclines, tetracenomycins (Tcms), benzoisochromanequinones, and pentangular polyphenols [4]. Tcms, which have been isolated from Streptomyces glaucescens and Streptomyces olivaceus, represent a separate group of aromatic polyketides featuring a tetracyclic naphthacenequinone chromophore with highly hydroxylated cyclohexenone moiety, and exhibit moderate antibacterial and antitumor activities [5]. Terrestrial and marine actinomycetes are particularly rich sources of bioactive PKS II metabolites. With the advent of molecular tools and advances in biosynthetic-mechanism research, genetic-level investigation of bacterial aromatic polyketides has become possible [6]. PCR-based genetic screening has become a useful approach to identify novel metabolites with desired structural characteristics [7,8].
As part of our screening program for new antibiotics from marine-derived microorganisms [9][10][11], we previously identified Tcm X (7) and four Tcm analogs, saccharothrixones A, B (8), C (9), and D, from the rare actinomycete Saccharothrix sp. 10-10 by PCR screening [12,13]. To investigate the structural diversity of Tcms produced by strain 10-10 and, thereby, to explore structure-activity relationships, we further analyzed the LC-MS data of other fractions of the culture extracts and identified six new Tcm analogs, saccharothrixones E-I (1)(2)(3)(4)(5), and 13-de-O-methyltetracenomycin X (6, Figure 1). Saccharothrixone E (1) and F (2) were identified as 5-de-oxo-5-hydroxy derivatives of Tcm X and C, respectively. Saccharothrixones G (3) and H (4) were C-4 epimers of seco-tetracenomycins that featured a unique ring-A-cleaved chromophore. This paper describes the isolation and structural characterization of the six new tetracenomycin congeners 1-6, as well as the structure-activity relationship of their cytotoxicity.
Mar. Drugs 2018, 16, x FOR PEER REVIEW 2 of 10 As part of our screening program for new antibiotics from marine-derived microorganisms [9][10][11], we previously identified Tcm X (7) and four Tcm analogs, saccharothrixones A, B (8), C (9), and D, from the rare actinomycete Saccharothrix sp. 10-10 by PCR screening [12,13]. To investigate the structural diversity of Tcms produced by strain 10-10 and, thereby, to explore structure-activity relationships, we further analyzed the LC-MS data of other fractions of the culture extracts and identified six new Tcm analogs, saccharothrixones E-I (1)(2)(3)(4)(5), and 13-de-O-methyltetracenomycin X (6, Figure 1). Saccharothrixone E (1) and F (2) were identified as 5-de-oxo-5-hydroxy derivatives of Tcm X and C, respectively. Saccharothrixones G (3) and H (4) were C-4 epimers of secotetracenomycins that featured a unique ring-A-cleaved chromophore. This paper describes the isolation and structural characterization of the six new tetracenomycin congeners 1-6, as well as the structure-activity relationship of their cytotoxicity.

Results
Saccharothrixone E (1) was isolated as a yellow powder. Its molecular formula was determined to be C24H24O11 by HRESIMS, which is 2 mass units higher than Tcm X (7). The UV spectrum of 1 exhibited absorption maxima at 266, 277, and 382 nm. The 1 H NMR spectrum in acetone-d6 (Table 1) displayed characteristic signals for two aromatic protons (δH 7.44 (s, H-6) and 7.20 (s, H-7)), an olefinic proton (δH 5.61 (s, H-2)), and two oxygenated methine protons (δH 4.77 (brs, H-5) and 4.62 (brs, H-4)). In addition, an olefinic methyl and four methoxy signals were observed at δH 2.79-3.94 ppm. In the 1 H NMR spectrum recorded in DMSO-d6 (Table S1), characteristic feature was the presence of four exchangeable protons at δH 14.70 (brs), 6.07 (brs), 5.47 (d), and 5.23 (s). These spectroscopic data implied that the structure of 1 is closely related to 7. The 13 C NMR (Table 2) and HSQC spectra of compound 1 revealed the presence of 24 carbons, and also indicated its close similarity to 7 (Table  S2). One of the three ketonic carbonyl resonances observed for 7 was missing in 1, and, instead, an Obearing methine carbon signal was observed at δC 69.6. This indicated that one of the ketone groups in 7 was replaced by a hydroxy-substituted carbon in 1. A detailed comparison of the NMR data of 1 and 7 revealed the structural similarities in their A, C, and D rings.

Results
Saccharothrixone E (1) was isolated as a yellow powder. Its molecular formula was determined to be C 24 H 24 O 11 by HRESIMS, which is 2 mass units higher than Tcm X (7). The UV spectrum of 1 exhibited absorption maxima at 266, 277, and 382 nm. The 1 H NMR spectrum in acetone-d 6 ( Table 1) displayed characteristic signals for two aromatic protons (δ H 7.44 (s, H-6) and 7.20 (s, H-7)), an olefinic proton (δ H 5.61 (s, H-2)), and two oxygenated methine protons (δ H 4.77 (brs, H-5) and 4.62 (brs, H-4)). In addition, an olefinic methyl and four methoxy signals were observed at δ H 2.79-3.94 ppm. In the 1 H NMR spectrum recorded in DMSO-d 6 (Table S1), characteristic feature was the presence of four exchangeable protons at δ H 14.70 (brs), 6.07 (brs), 5.47 (d), and 5.23 (s). These spectroscopic data implied that the structure of 1 is closely related to 7. The 13 C NMR (Table 2) and HSQC spectra of compound 1 revealed the presence of 24 carbons, and also indicated its close similarity to 7 (Table S2). One of the three ketonic carbonyl resonances observed for 7 was missing in 1, and, instead, an O-bearing methine carbon signal was observed at δ C 69.6. This indicated that one of the ketone groups in 7 was replaced by a hydroxy-substituted carbon in 1. A detailed comparison of the NMR data of 1 and 7 revealed the structural similarities in their A, C, and D rings.  The HMBC correlations ( Figure 2) from H-7 to C-8 (δ C 158.1), C-9, C-10 (δ C 137.8), C-10a, and C-13 (δ C 168.6), and from H-6 to C-5a, C-6a, C-10a, C-11 (δ C 166.7), and C-11a confirmed that the naphthalene ring substitution pattern of 1 is identical to that of 7. The correlation from H-2 to C-1 (δ C 193.6) and the relatively weaker 4 J correlation from H-6 to C-12 (δ C 202.2) observed in the HMBC spectrum suggested that the ketone groups at C-1 and C-12 that are characteristic for tetracenomycins remained intact in 1. The strong 3 J HMBC correlation from H-6 to C-5 (δ C 69.6) indicated that the carbonyl group at C-5 in 7 was replaced by an oxygenated methine group in 1. The remaining HMBC correlations confirmed structure 1 as 5-de-oxo-5-hydroxytetracenomycin X.
Mar. Drugs 2018, 16, x FOR PEER REVIEW 3 of 10 The HMBC correlations ( Figure 2) from H-7 to C-8 (δC 158.1), C-9, C-10 (δC 137.8), C-10a, and C-13 (δC 168.6), and from H-6 to C-5a, C-6a, C-10a, C-11 (δC 166.7), and C-11a confirmed that the naphthalene ring substitution pattern of 1 is identical to that of 7. The correlation from H-2 to C-1 (δC 193.6) and the relatively weaker 4 J correlation from H-6 to C-12 (δC 202.2) observed in the HMBC spectrum suggested that the ketone groups at C-1 and C-12 that are characteristic for tetracenomycins remained intact in 1. The strong 3 J HMBC correlation from H-6 to C-5 (δC 69.6) indicated that the carbonyl group at C-5 in 7 was replaced by an oxygenated methine group in 1. The remaining HMBC correlations confirmed structure 1 as 5-de-oxo-5-hydroxytetracenomycin X. The relative configuration of compound 1 was determined by interpretation of the ROESY data. The ROESY cross-peaks ( Figure 3) of OH-4a with OH-4, OH-5, and 12a-OCH3, and of OH-4 with OH-5 indicated that they were all located on the same β-face of the ring. The ROESY correlations of H-5 with H-4 and H-2 indicated that H-4 and H-5 were on the α-face. Since the relative structure of ring A in 1 is consistent with those in Tcms C [5] and X (7) [14], it was deduced that compound 1 has the same absolute configuration as 7. The absolute configuration of 1 was further confirmed using timedependent density functional theory (TDDFT)-electronic circular dichroism (ECD) calculations [15]. The ECD spectra of the possible isomers of 1 obtained by geometry optimization were generated using TDDFT calculations at the B3LYP/6-311++G(d,p) level. The ECD spectrum calculated for the The relative configuration of compound 1 was determined by interpretation of the ROESY data. The ROESY cross-peaks ( Figure 3) of OH-4a with OH-4, OH-5, and 12a-OCH 3 , and of OH-4 with OH-5 indicated that they were all located on the same β-face of the ring. The ROESY correlations of H-5 with H-4 and H-2 indicated that H-4 and H-5 were on the α-face. Since the relative structure of ring A in 1 is consistent with those in Tcms C [5] and X (7) [14], it was deduced that compound 1 has the same absolute configuration as 7. The absolute configuration of 1 was further confirmed using time-dependent density functional theory (TDDFT)-electronic circular dichroism (ECD) calculations [15]. The ECD spectra of the possible isomers of 1 obtained by geometry optimization were generated using TDDFT calculations at the B3LYP/6-311++G(d,p) level. The ECD spectrum calculated for the 4S,4aR,5S,12aR-1a isomer was in good agreement with the experimental ECD curve ( Figure 4). Consequently, the absolute configuration of 1 was firmly assigned as 4S,4aR,5S,12aR.
Mar. Drugs 2018, 16, x FOR PEER REVIEW 4 of 10 4S,4aR,5S,12aR-1a isomer was in good agreement with the experimental ECD curve ( Figure 4). Consequently, the absolute configuration of 1 was firmly assigned as 4S,4aR,5S,12aR.  The molecular formula of saccharothrixone F (2) was deduced to be C23H22O11 by HRESIMS, one CH2 unit less than that of 1. The UV spectrum of 2 displayed absorption bands almost identical to those of 1, indicating that 1 and 2 have the same chromophore. A comparison of the 1 H and 13 C NMR data of 1 and 2 revealed the absence of one of the four O-methyl groups of 1 in 2. As compared with that of 1, the resonance for C-12a in 2 was shifted by Δδ −5.3. These data suggested that 2 is the 12ade-O-methyl analogue of 1. This was further confirmed by 2D NMR experiments. The HMBC displayed cross-peaks from the three O-methyl protons at δH 3.88, 3.96, and 3.90 to C-3 (δC 176.6), C-8 (δC 158.0), and C-13 (δC 168.6), respectively, indicating that the methoxy groups were located at C-3, C-8, and C-13. The HMBC correlations of H-2 with C-12a (δC 82.5) confirmed the replacement of the 12a-O-methyl in 1 by a hydroxy group in 2. The relative and absolute configurations of 2 were determined to be the same as 1 based on the similarity of the ROESY correlations and ECD spectra. This was also supported by TDDFT-ECD calculations for 2 (Figure 4). Saccharothrixone G (3) was isolated as a yellow powder. The molecular formula was determined to be C24H24O11 by HRESIMS, which is the same as that of 1. Its 1 H and 13 C NMR data (Tables 1 and 2) differed significantly from those of 1, suggesting a substantial structural change. Analysis of the 1 H NMR data measured in DMSO-d6 (Table S1) indicated that one of the four exchangeable proton signals in 1 disappeared in 3. Comparison of their 13 C NMR/HSQC spectra revealed that the nonprotonated carbon (C-12a) in 1 was replaced by an O-bearing methine group in 3, suggesting that 4S,4aR,5S,12aR-1a isomer was in good agreement with the experimental ECD curve (Figure 4). Consequently, the absolute configuration of 1 was firmly assigned as 4S,4aR,5S,12aR.  The molecular formula of saccharothrixone F (2) was deduced to be C23H22O11 by HRESIMS, one CH2 unit less than that of 1. The UV spectrum of 2 displayed absorption bands almost identical to those of 1, indicating that 1 and 2 have the same chromophore. A comparison of the 1 H and 13 C NMR data of 1 and 2 revealed the absence of one of the four O-methyl groups of 1 in 2. As compared with that of 1, the resonance for C-12a in 2 was shifted by Δδ −5.3. These data suggested that 2 is the 12ade-O-methyl analogue of 1. This was further confirmed by 2D NMR experiments. The HMBC displayed cross-peaks from the three O-methyl protons at δH 3.88, 3.96, and 3.90 to C-3 (δC 176.6), C-8 (δC 158.0), and C-13 (δC 168.6), respectively, indicating that the methoxy groups were located at C-3, C-8, and C-13. The HMBC correlations of H-2 with C-12a (δC 82.5) confirmed the replacement of the 12a-O-methyl in 1 by a hydroxy group in 2. The relative and absolute configurations of 2 were determined to be the same as 1 based on the similarity of the ROESY correlations and ECD spectra. This was also supported by TDDFT-ECD calculations for 2 (Figure 4). Saccharothrixone G (3) was isolated as a yellow powder. The molecular formula was determined to be C24H24O11 by HRESIMS, which is the same as that of 1. Its 1 H and 13 C NMR data (Tables 1 and 2) differed significantly from those of 1, suggesting a substantial structural change. Analysis of the 1 H NMR data measured in DMSO-d6 (Table S1) indicated that one of the four exchangeable proton signals in 1 disappeared in 3. Comparison of their 13 C NMR/HSQC spectra revealed that the nonprotonated carbon (C-12a) in 1 was replaced by an O-bearing methine group in 3, suggesting that The molecular formula of saccharothrixone F (2) was deduced to be C 23 H 22 O 11 by HRESIMS, one CH 2 unit less than that of 1. The UV spectrum of 2 displayed absorption bands almost identical to those of 1, indicating that 1 and 2 have the same chromophore. A comparison of the 1 H and 13 C NMR data of 1 and 2 revealed the absence of one of the four O-methyl groups of 1 in 2. As compared with that of 1, the resonance for C-12a in 2 was shifted by ∆δ −5.3. These data suggested that 2 is the 12a-de-O-methyl analogue of 1. This was further confirmed by 2D NMR experiments. The HMBC displayed cross-peaks from the three O-methyl protons at δ H 3.88, 3.96, and 3.90 to C-3 (δ C 176.6), C-8 (δ C 158.0), and C-13 (δ C 168.6), respectively, indicating that the methoxy groups were located at C-3, C-8, and C-13. The HMBC correlations of H-2 with C-12a (δ C 82.5) confirmed the replacement of the 12a-O-methyl in 1 by a hydroxy group in 2. The relative and absolute configurations of 2 were determined to be the same as 1 based on the similarity of the ROESY correlations and ECD spectra. This was also supported by TDDFT-ECD calculations for 2 (Figure 4). Saccharothrixone G (3) was isolated as a yellow powder. The molecular formula was determined to be C 24 H 24 O 11 by HRESIMS, which is the same as that of 1. Its 1 H and 13 C NMR data (Tables 1 and 2) differed significantly from those of 1, suggesting a substantial structural change. Analysis of the 1 H NMR data measured in DMSO-d 6 (Table S1) indicated that one of the four exchangeable proton signals in 1 disappeared in 3. Comparison of their 13 C NMR/HSQC spectra revealed that the nonprotonated carbon (C-12a) in 1 was replaced by an O-bearing methine group in 3, suggesting that the fused A and B rings were cleaved at C-12a. HMBC correlations of OH-4a with C-4a and C-12a, and OH-5 with C-5, indicated that the free hydroxy groups at C-4a and C-5 remained in 3, whereas the free hydroxy group at C-4 was missing. Key HMBC correlations from H-4 to the carbonyl carbon C-1 (δ C 172.2) suggested that the oxygenated C-4 was connected to C-1 via an oxygen atom to form a lactone ring. Finally, HMBC correlations from H-4 to C-4a, C-5, and C-12a indicated the connectivity of C-4 and C-4a, thus completing the establishment of the planar structure of 3.
The relative configuration of 3 was established by analysis of the NOE correlations. The NOE enhancement between H-12a and H-5 revealed their 1,3-diaxial positions and a half-chair conformation for ring B (Figure 3). The NOE correlations of H-4 with H-5 and H-12a indicated that the methine group C-4 was in an equatorial orientation in ring B while OH-4a was axial. The ECD spectra of four possible isomers were then calculated using the TDDFT-ECD method. The calculated curves for 4S,4aR,5S,12aS-3a ( Figure 5) at both CAM-B3LYP/TZVP and WB97XD/6-311++G(d,p) levels were in good agreement with the experimental spectrum, thus determining the absolute configuration of 3 to be 4S,4aR,5S,12aS.
Mar. Drugs 2018, 16, x FOR PEER REVIEW 5 of 10 the fused A and B rings were cleaved at C-12a. HMBC correlations of OH-4a with C-4a and C-12a, and OH-5 with C-5, indicated that the free hydroxy groups at C-4a and C-5 remained in 3, whereas the free hydroxy group at C-4 was missing. Key HMBC correlations from H-4 to the carbonyl carbon C-1 (δC 172.2) suggested that the oxygenated C-4 was connected to C-1 via an oxygen atom to form a lactone ring. Finally, HMBC correlations from H-4 to C-4a, C-5, and C-12a indicated the connectivity of C-4 and C-4a, thus completing the establishment of the planar structure of 3. The relative configuration of 3 was established by analysis of the NOE correlations. The NOE enhancement between H-12a and H-5 revealed their 1,3-diaxial positions and a half-chair conformation for ring B (Figure 3). The NOE correlations of H-4 with H-5 and H-12a indicated that the methine group C-4 was in an equatorial orientation in ring B while OH-4a was axial. The ECD spectra of four possible isomers were then calculated using the TDDFT-ECD method. The calculated curves for 4S,4aR,5S,12aS-3a ( Figure 5) at both CAM-B3LYP/TZVP and WB97XD/6-311++G(d,p) levels were in good agreement with the experimental spectrum, thus determining the absolute configuration of 3 to be 4S,4aR,5S,12aS. Saccharothrixone H (4) was determined to be a diastereoisomer of 3 based on their identical molecular formula (C24H24O11) and their similar NMR data. This was verified by HMBC correlations (Figure 2). In addition, similar correlations ( Figure 3) observed in the ROESY spectra of 4 revealed that the relative configuration of ring B in 4 was also identical to that of 3. This was suggestive of a 4epimer of 3. The calculated ECD spectrum of 4R,4aR,5S,12aS-3c fit well with the experimental spectrum of 4. Therefore, the structure of 4 was assigned and named saccharothrixone H.
Saccharothrixone I (5) was obtained as a white powder. The HRESIMS data established its molecular composition as C24H24O11, identical with the molecular formulae of 3, 4, and saccharothrixones B (8) and C (9) [12]. The UV, 1 H NMR, and 13 C NMR spectra were very similar to those of 8 (Table S3) except for the chemical shifts of the protons and carbon signals in rings A and B, indicating that 5 was a diastereoisomer of 8. The 2D NMR data of 5 ( Figure 2) confirmed it had the same planar structure as 8. The ROESY correlation between H-4 and H-12a indicated a 1,3-diaxial interaction. ROESY correlations of H-5 with H-4 and H-12a illustrated an equatorial orientation of the methine group C-5 in ring A, and thus an axial position for OH-4a. Therefore, the relative configuration of ring A in 5 was determined to be identical to that of 8, suggesting that 5 is a 5-epimer of 8. The ECD spectrum of 5 exhibited negative Cotton effects (CEs) at 262 nm and between 299 and 350 nm, and a positive CE at 244 nm. The observed negative CE around 262 nm ascribed for the n→π* transition of the α,β-unsaturated γ-lactone [16], was opposite to that of 8 ( Figure S1), indicating that Saccharothrixone H (4) was determined to be a diastereoisomer of 3 based on their identical molecular formula (C 24 H 24 O 11 ) and their similar NMR data. This was verified by HMBC correlations (Figure 2). In addition, similar correlations ( Figure 3) observed in the ROESY spectra of 4 revealed that the relative configuration of ring B in 4 was also identical to that of 3. This was suggestive of a 4-epimer of 3. The calculated ECD spectrum of 4R,4aR,5S,12aS-3c fit well with the experimental spectrum of 4. Therefore, the structure of 4 was assigned and named saccharothrixone H.
Saccharothrixone I (5) was obtained as a white powder. The HRESIMS data established its molecular composition as C 24 H 24 O 11 , identical with the molecular formulae of 3, 4, and saccharothrixones B (8) and C (9) [12]. The UV, 1 H NMR, and 13 C NMR spectra were very similar to those of 8 (Table S3) except for the chemical shifts of the protons and carbon signals in rings A and B, indicating that 5 was a diastereoisomer of 8. The 2D NMR data of 5 ( Figure 2) confirmed it had the same planar structure as 8. The ROESY correlation between H-4 and H-12a indicated a 1,3-diaxial interaction. ROESY correlations of H-5 with H-4 and H-12a illustrated an equatorial orientation of the methine group C-5 in ring A, and thus an axial position for OH-4a. Therefore, the relative configuration of ring A in 5 was determined to be identical to that of 8, suggesting that 5 is a 5-epimer of 8. The ECD spectrum of 5 exhibited negative Cotton effects (CEs) at 262 nm and between 299 and 350 nm, and a positive CE at 244 nm. The observed negative CE around 262 nm ascribed for the n→π* transition of the α,β-unsaturated γ-lactone [16], was opposite to that of 8 ( Figure S1), indicating that the configuration of C-5 in the lactone ring B of 5 was opposite to that of 8. This was further confirmed by the TDDFT-ECD calculation, which showed good agreement of the calculated spectrum for 4S,4aR,5R,12aR-5a ( Figure 6) with the experimental curve. Therefore, the absolute configuration of 5 was assigned as 4S,4aR,5R,12aR.
Mar. Drugs 2018, 16, x FOR PEER REVIEW 6 of 10 the configuration of C-5 in the lactone ring B of 5 was opposite to that of 8. This was further confirmed by the TDDFT-ECD calculation, which showed good agreement of the calculated spectrum for 4S,4aR,5R,12aR-5a ( Figure 6) with the experimental curve. Therefore, the absolute configuration of 5 was assigned as 4S,4aR,5R,12aR. Figure 6. Comparison of experimental curve of 5 and calculated ECD spectra for 4S,4aR,5R,12aR-5a and 4R,4aS,5S,12aS-5b.
Compound 6 has the molecular formula C23H20O11, one CH2 unit less than Tcm X (7), as determined by HRESIMS and NMR data. The NMR data of 6 were similar to those of 7 except for the absence of one methoxy group signal in 6. The HMBC correlations from the O-methyl protons at δH 3.80, 4.01, and 3.56 to C-3 (δC 174.8), C-8 (δC 159.3), and C-12a (δC 89.0), respectively, located the methoxy groups at C-3, C-8 and C-12a, suggesting that the methoxy group at C-13 in 7 was absent in 6. Therefore, the structure of 6 was determined to be 13-de-O-methyltetracenomycin X.
The identification of compounds 1 and 2 supports our previously proposed biosynthetic pathway for the ring-B cleaved tetracenomycin derivatives saccharothrixones A-C [12]. In that biosynthetic pathway, saccharothrixones A-C are derived from the intermediate 2 and its epimer of C-5, which undergo an intramolecular nucleophilic addition from OH-5 to 12-oxo group and simultaneous cleavage of ring B. Similarly, the nucleophilic addition from OH-4 to 1-oxo group of 1 would result in a divergent pathway committed to the formation of saccharothrixones G (3) and H (4) (Scheme S1) [12]. The coisolation of the different C-4, C-5, and C-12a epimers of secotetracenomycins indicated that the cleavage of rings A and B by the intramolecular nucleophilic addition was not stereocontrolled.
In a previous study [12], we reported that Tcm X and its isomer, saccharothrixone D, showed moderate cytotoxicity (5.4-20.8 µM) against the HepG2, MCF-7, and K562 human cancer cell lines, whereas the B-ring-cleaved derivatives saccharothrixones A-C were inactive at a concentration of 100 µM. To further evaluate the structure-activity relationship of these tetracenomycin congeners, we examined the cytotoxicity of compounds 1-6 against the cancer cell lines mentioned above. All the compounds were found to be inactive at 100 µM. The action mechanism of tetracenomycins was assumed to be intercalation with DNA, requiring flat structural moiety to move between the base pairs [17]. These results confirmed that the naphthacenequinone chromophore and the planarity of the molecules are vital for their cytotoxicity. The cleaved naphthacenequinone chromophore with a large substituent (ring A) in saccharothrixones A-C and G-I (3)(4)(5) probably blocks the intercalation with DNA, causing loss of cytotoxicity. The replacement of the ketone group by the hydroxy group at C-5 in compounds 1 and 2 changes the planarity of ring B, which could lower the effectiveness of intercalation. Compound 6, which differs from Tcm X only at the C-13 substituent (carboxy vs. methoxycarbonyl), showed no effect at 100 µM, indicating that the free carboxy group caused a significantly negative interaction with the DNA. Rohr et al. previously reported that tetracenomycin derivatives with a free hydroxy group at C-8 (elloramycinone) or C-12a (Tcm C) were less active than Tcm X, which has methoxy groups at C-8 and C-12a [17]. Our results further confirmed the Figure 6. Comparison of experimental curve of 5 and calculated ECD spectra for 4S,4aR,5R,12aR-5a and 4R,4aS,5S,12aS-5b.
Compound 6 has the molecular formula C 23 H 20 O 11 , one CH 2 unit less than Tcm X (7), as determined by HRESIMS and NMR data. The NMR data of 6 were similar to those of 7 except for the absence of one methoxy group signal in 6. The HMBC correlations from the O-methyl protons at δ H 3.80, 4.01, and 3.56 to C-3 (δ C 174.8), C-8 (δ C 159.3), and C-12a (δ C 89.0), respectively, located the methoxy groups at C-3, C-8 and C-12a, suggesting that the methoxy group at C-13 in 7 was absent in 6. Therefore, the structure of 6 was determined to be 13-de-O-methyltetracenomycin X.
The identification of compounds 1 and 2 supports our previously proposed biosynthetic pathway for the ring-B cleaved tetracenomycin derivatives saccharothrixones A-C [12]. In that biosynthetic pathway, saccharothrixones A-C are derived from the intermediate 2 and its epimer of C-5, which undergo an intramolecular nucleophilic addition from OH-5 to 12-oxo group and simultaneous cleavage of ring B. Similarly, the nucleophilic addition from OH-4 to 1-oxo group of 1 would result in a divergent pathway committed to the formation of saccharothrixones G (3) and H (4) (Scheme S1) [12]. The coisolation of the different C-4, C-5, and C-12a epimers of seco-tetracenomycins indicated that the cleavage of rings A and B by the intramolecular nucleophilic addition was not stereocontrolled.
In a previous study [12], we reported that Tcm X and its isomer, saccharothrixone D, showed moderate cytotoxicity (5.4-20.8 µM) against the HepG2, MCF-7, and K562 human cancer cell lines, whereas the B-ring-cleaved derivatives saccharothrixones A-C were inactive at a concentration of 100 µM. To further evaluate the structure-activity relationship of these tetracenomycin congeners, we examined the cytotoxicity of compounds 1-6 against the cancer cell lines mentioned above. All the compounds were found to be inactive at 100 µM. The action mechanism of tetracenomycins was assumed to be intercalation with DNA, requiring flat structural moiety to move between the base pairs [17]. These results confirmed that the naphthacenequinone chromophore and the planarity of the molecules are vital for their cytotoxicity. The cleaved naphthacenequinone chromophore with a large substituent (ring A) in saccharothrixones A-C and G-I (3)(4)(5) probably blocks the intercalation with DNA, causing loss of cytotoxicity. The replacement of the ketone group by the hydroxy group at C-5 in compounds 1 and 2 changes the planarity of ring B, which could lower the effectiveness of intercalation. Compound 6, which differs from Tcm X only at the C-13 substituent (carboxy vs. methoxycarbonyl), showed no effect at 100 µM, indicating that the free carboxy group caused a significantly negative interaction with the DNA. Rohr et al. previously reported that tetracenomycin derivatives with a free hydroxy group at C-8 (elloramycinone) or C-12a (Tcm C) were less active than Tcm X, which has methoxy groups at C-8 and C-12a [17]. Our results further confirmed the importance of the methylation of the polar hydroxy and carboxy groups for their cytotoxic activity. In additions, 1-6 were also evaluated for antibacterial activity, but were found to be inactive (MIC > 64 µg/mL). data, Table 2

Biological Assays
The cytotoxicities of the tested compounds against the human cancer cells HepG2 (hepatocellular carcinoma), MCF-7 (breast adenocarcinoma), and K562 (leukemia) were evaluated by the sulforhodamine B (SRB) assay as described previously [12]. The antibacterial assay was performed by using the agar dilution method [12].

Conclusions
Six new tetracenomycin derivatives including three seco-tetracenomycins were isolated from the rare marine-derived actinomycete Saccharothrix sp. 10-10. Saccharothrixones G (3) and H (4) are the first examples of tetracenomycins featuring a novel ring-A-cleaved chromophore. Saccharothrixone I (5), together with previously identified saccharothrixones A-C from the same cultures, are the only seco-tetracenomycins with a cleaved ring-B skeleton isolated from microbial natural products. This finding not only adds diversity of tetracenomycins, but also provides evidence for tetracenomycin-related polyketides biosynthesis. The structure-activity relationship study indicated that the planarity of the chromophore and methylation of the polar carboxy groups are important for tetracenomycin cytotoxicity.