Synthetic Study of Natural Metabolites Containing a Benzo[c]oxepine Skeleton: Heterocornol C and D

A versatile strategy for the enantioselective synthesis of a benzo[c]oxepine structural core containing natural secondary metabolites was developed. The key steps of the synthetic approach include ring-closing alkene metathesis for seven-member ring construction, the Suzuki–Miyaura cross-coupling reaction for the installation of the double bond and Katsuki–Sharpless asymmetric epoxidation for the introduction of chiral centers. The first total synthesis and absolute configuration assignment of heterocornol D (3a) were achieved. Four stereoisomers, 3a, ent-3a, 3b and ent-3b, of this natural polyketide were prepared, starting with 2,6-dihydroxy benzoic acid and divinyl carbinol. The absolute and relative configuration of heterocornol D was assigned via single-crystal X-ray analysis. The extension of the described synthetic approach is further presented with the synthesis of heterocornol C by applying the ether group reduction method to the lactone.


Introduction
Marine organisms are interesting sources of new biologically and therapeutically active compounds [1]. Over the last several decades, many polyketide-type secondary metabolites of fungal origin have been evaluated [2]. Polyketides containing a salicylic/benzylic alcohol structural fragment have been shown to exhibit diverse biological activities [3][4][5][6][7]. Recently, Huang, Han and coworkers [8] described twelve new salicylaldehydetype polyketide derivatives obtained from a marine-sponge-associated fungus Pestalotiopsis heterocornis. Among these compounds, bicyclic benzo[c]oxepines such as heterocornols C, D, G (1-3) and H (4) were identified (Figure 1).  The structure of natural compounds 1-4 was determined based on detailed spectroscopic data. The configuration of the C-3 and C-1 centers in heterocornol H (4) were identified as anti by comparing the 13 C NMR data of the putative structure of xylarinol B [8,9]. In addition, all the isolated metabolites were evaluated via MTT assay for their cytotoxic activity against four human cancer cell lines (BGC-823, H460, PC-3 and SMMC-7721). Compounds 1, 2 and 4 showed moderate cytotoxicity, with IC 50 values in the 15-100 µM range, with adriamycin assayed as a positive control. The antifungal properties of the isolated derivatives appeared to be dependent on the presence of the pent-4-ene-2,3-diol fragment. Derivatives 1 and 4 showed moderate antifungal activity against Candida parapsilosis and Cryptococcus neofromans at a concentration of 100 µg/mL. The promising results of the preliminary SAR study of P. heterocosnis metabolites encouraged us to develop a synthetic strategy for these natural compounds containing a benzo[c]oxepine skeleton.

Results and Discussion
To determine the relative and absolute configurations of the isolated heterocornols (1-4), we propose a flexible strategy for the synthesis of all possible stereoisomers of 3. Herein, we report a synthetic route for the construction of the benzo[c]oxepine skeleton. The synthesis of all stereoisomers of heterocornol D (3) (3a, 3b, ent-3a and ent-3b) with a relative C3-C1 syn-/anti-configuration is described. The utilized strategy is depicted in Scheme 1. The structure of natural compounds 1-4 was determined based on detailed spectroscopic data. The configuration of the C-3 and C-1′ centers in heterocornol H (4) were identified as anti by comparing the 13 C NMR data of the putative structure of xylarinol B [8,9]. In addition, all the isolated metabolites were evaluated via MTT assay for their cytotoxic activity against four human cancer cell lines (BGC-823, H460, PC-3 and SMMC-7721). Compounds 1, 2 and 4 showed moderate cytotoxicity, with IC50 values in the 15-100 μM range, with adriamycin assayed as a positive control. The antifungal properties of the isolated derivatives appeared to be dependent on the presence of the pent-4-ene-2,3-diol fragment. Derivatives 1 and 4 showed moderate antifungal activity against Candida parapsilosis and Cryptococcus neofromans at a concentration of 100 μg/mL. The promising results of the preliminary SAR study of P. heterocosnis metabolites encouraged us to develop a synthetic strategy for these natural compounds containing a benzo[c]oxepine skeleton.

Results and Discussion
To determine the relative and absolute configurations of the isolated heterocornols (1-4), we propose a flexible strategy for the synthesis of all possible stereoisomers of 3. Herein, we report a synthetic route for the construction of the benzo[c]oxepine skeleton. The synthesis of all stereoisomers of heterocornol D (3) (3a, 3b, ent-3a and ent-3b) with a relative C3-C1′ syn-/anti-configuration is described. The utilized strategy is depicted in Scheme 1. Scheme 1. Retrosynthetic analysis of the heterocornols C and D.
In general, the key operation in synthesis is the construction of an oxepine-1-one ring via ring-closing metathesis. The substrate, optically pure diene 16, having the defined configuration of the alkyl fragment, can be obtained via the re-esterification of the aromatic ester 8 with partially protected diol 14. The following styrene 8 can be prepared via Suzuki-Miyaura coupling starting with triflate 7, which, in turn, is available from 2,6-dihydroxy-benzoic acid 5. The applicability of this strategy in the preparation of all stereoisomers of the targeted natural products is ensured through the Sharpless asymmetric epoxidation of divinyl carbinol 9 [10,11] and the eventual Mitsunobu inversion of the epoxide 10, affording both diastereomers syn and anti, respectively.
In addition, the proposed strategy could be utilized in the preparation of other natural heterocornols, (1), (2) and (4), involving selective reduction of the lactone 3, the introduction of the prenyl group at carbon C-6 and/or hydrogenation of the C4-C5 double bond. Consequently, styrene 8 was readily prepared from commercially available 2,6-dihydroxy-benzoic acid 5 in three steps, using the literature protocols (Scheme 2) [12]. In general, the key operation in synthesis is the construction of an oxepine-1-one ring via ring-closing metathesis. The substrate, optically pure diene 16, having the defined configuration of the alkyl fragment, can be obtained via the re-esterification of the aromatic ester 8 with partially protected diol 14. The following styrene 8 can be prepared via Suzuki-Miyaura coupling starting with triflate 7, which, in turn, is available from 2,6-dihydroxybenzoic acid 5. The applicability of this strategy in the preparation of all stereoisomers of the targeted natural products is ensured through the Sharpless asymmetric epoxidation of divinyl carbinol 9 [10,11] and the eventual Mitsunobu inversion of the epoxide 10, affording both diastereomers syn and anti, respectively.
In addition, the proposed strategy could be utilized in the preparation of other natural heterocornols, (1), (2) and (4), involving selective reduction of the lactone 3, the introduction of the prenyl group at carbon C-6 and/or hydrogenation of the C4-C5 double bond. Consequently, styrene 8 was readily prepared from commercially available 2,6-dihydroxy-benzoic acid 5 in three steps, using the literature protocols (Scheme 2) [12]. Int. J. Mol. Sci. 2023, 24, x FOR PEER REVIEW 3 of 9 Scheme 2. Synthesis of styrene 8.
With both key fragments in hand, the aromatic and alkyl fragments were coupled using a re-esterification reaction. In this manner, the reaction of the previously prepared styrene 8 with the corresponding alcohols 14 and sodium hydride in THF at 0 • C for 2 h provided the benzoates 16 in good yields (16a/80%, ent-16a/79%, 16b/85%, ent-16b/71%), after work-up and purification using MPLC (Scheme 4a,b). The key reaction for the construction of the benzoxepinone skeleton was ring-closing alkene metathesis. Thus, the treatment of the dienes 16 with the Grubbs catalyst (second generation, 0.13 equiv) in toluene at reflux for 24 h provided benzoxepine-1-ones 17a, ent-17a, 17b and ent-17b in 73%, 82%, 71% and 84% yields, respectively. The final deprotection of the methoxymethyl group via acidic hydrolysis using Dowex in iso-propanol furnished the target heterocornols ( With both key fragments in hand, the aromatic and alkyl fragments were coupled using a re-esterification reaction. In this manner, the reaction of the previously prepared styrene 8 with the corresponding alcohols 14 and sodium hydride in THF at 0 °C for 2 h provided the benzoates 16 in good yields (16a/80%, ent-16a/79%, 16b/85%, ent-16b/71%), after work-up and purification using MPLC (Scheme 4a,b). The key reaction for the construction of the benzoxepinone skeleton was ring-closing alkene metathesis. Thus, the treatment of the dienes 16 with the Grubbs catalyst (second generation, 0.13 equiv) in toluene at reflux for 24 h provided benzoxepine -1-ones 17a, ent-17a, 17b and ent-17b in 73%, 82%, 71% and 84% yields, respectively. The final deprotection of the methoxymethyl group via acidic hydrolysis using Dowex in iso-propanol furnished the target heterocornols (  The following NMR analysis of the prepared final compounds showed very good agreement of the 1 H and 13 C spectral data for the synthetic derivatives 3a and ent-3a with those obtained from the natural source, confirming the constitution of heterocornol D (3). The relative and absolute configurations of natural compound 3 were confirmed in a single-crystal X-ray analysis. See Supplementary Materials for X-ray data ( Figures S1-S3, Tables S1 and S2). The colorless needles suitable for single-crystal X-ray analysis were isolated after MPLC purification and recrystallisation of the intermediate ent-17a (EtOAc/nheptane = 1/25). Accordingly, the benzo[c]oxepin-1-one structure with 1′,3-anti-alignment and (1′R,3S)-absolute configuration of the compound ent-17a were confirmed (Figure 2) [16]. Considering the positive sign of the rotation of the plane of polarized light for the stereoisomer 3a, the structure of the natural heterocornol D (3) was assigned as having 1′,3-anti-alignment. The (1′S,3R)-absolute configuration was also established, despite the Scheme 4. Synthesis of heterocornol D (3a and its isomers, 3b, ent-3a and ent-3b).
The following NMR analysis of the prepared final compounds showed very good agreement of the 1 H and 13 C spectral data for the synthetic derivatives 3a and ent-3a with those obtained from the natural source, confirming the constitution of heterocornol D (3). The relative and absolute configurations of natural compound 3 were confirmed in a single-crystal X-ray analysis. See Supplementary Materials for X-ray data (Figures S1-S3, Tables S1 and S2). The colorless needles suitable for single-crystal X-ray analysis were isolated after MPLC purification and recrystallisation of the intermediate ent-17a (EtOAc/n-heptane = 1/25). Accordingly, the benzo[c]oxepin-1-one structure with 1 ,3-antialignment and (1 R,3S)-absolute configuration of the compound ent-17a were confirmed ( Figure 2) [16]. Considering the positive sign of the rotation of the plane of polarized light for the stereoisomer 3a, the structure of the natural heterocornol D (3) was assigned as having 1 ,3-anti-alignment. The (1 S,3R)-absolute configuration was also established, despite the fact that the value of specific rotation α] 25  Moreover, the compound ent-17a was evaluated for a three-dimensional Hirshfeld surface. Furthermore, the calculated electrostatic potential on the Hirshfeld surface is shown in Figure 3.  Moreover, the compound ent-17a was evaluated for a three-dimensional Hirshfeld surface. Furthermore, the calculated electrostatic potential on the Hirshfeld surface is shown in Figure 3. Moreover, the compound ent-17a was evaluated for a three-dimensional Hirshfeld surface. Furthermore, the calculated electrostatic potential on the Hirshfeld surface is shown in Figure 3.  The deep-red spots on the d norm Hirshfeld surfaces indicate close contact interactions, which are mainly responsible for the significant intermolecular C-H···O hydrogen-bonding interactions. The red and blue colors on the electrostatic potential of the Hirshfeld surface show the donor (red) and acceptor (blue) parts of the molecule for the formation of intermolecular interactions. The results of the electrostatic potential on the Hirshfeld surface confirm the existence of C-H···O hydrogen-bonding interactions in the crystal structure of ent-17a. The analysis of the experimentally obtained Hirshfeld surface of ent-17a directly indicates the capacity for intermolecular interactions in future docking analysis.
To further extend the applicability of the described synthetic approach, an additional investigation for the synthesis of heterocornol C (1) was carried out. Since the relative and absolute configurations of natural compound 1 have not been described in the literature, we aimed to prepare both syn and anti-isomers of heterocornol C (1). Thus, synthesis was proposed utilizing the lactone reduction method, starting with the synthetic heterocornol D (3a). After trialing a few methods and optimization steps, we succeeded in the reduction of the lactone group of 3a to the ether 1 using a modified literature method (this is an extension to substrates having free OH groups) [17]. The two-step procedure, utilizing potassium tetrakis[3,5-bis(trifluoromethyl)phenyl]borate as a catalyst and phenylsilane as a reducing agent, following treatment of the reaction mixture with TBAF, provided natural compound 1 in a 38% yield (Scheme 5).
17a directly indicates the capacity for intermolecular interactions in future docking analysis.
To further extend the applicability of the described synthetic approach, an additional investigation for the synthesis of heterocornol C (1) was carried out. Since the relative and absolute configurations of natural compound 1 have not been described in the literature, we aimed to prepare both syn and anti-isomers of heterocornol C (1). Thus, synthesis was proposed utilizing the lactone reduction method, starting with the synthetic heterocornol D (3a). After trialing a few methods and optimization steps, we succeeded in the reduction of the lactone group of 3a to the ether 1 using a modified literature method (this is an extension to substrates having free OH groups) [17]. The two-step procedure, utilizing potassium tetrakis[3,5-bis(trifluoromethyl)phenyl]borate as a catalyst and phenylsilane as a reducing agent, following treatment of the reaction mixture with TBAF, provided natural compound 1 in a 38% yield (Scheme 5). Subsequently, the configuration of natural 1 was established through a comparison of the NMR data and optical rotation of synthetic 1 with the reported data of the isolated natural compound. See Supplementary Materials for NMR ( Figures S4-S69, Tables S3 and  S4). The sign and value of specific rotation [ ] 25 + 57.5 (c 0.3, MeOH) of the synthetic sample 1 and perfect match of the NMR spectra confirmed the anti-1′S,3R configuration of the natural heterocornol C (1) (see [8]: [ ] 20 + 80 (c 0.30, MeOH)).
To reliably prove the relative configuration of 1, we also prepared its diastereomer 18b in a 36% yield by applying the same synthetic procedure. The comparison of the NMR data of the synthetic syn-isomer 18b with the natural 1 data showed large discrepancies in the shifts of the oxepine ring and C-1′ and C-3 proton signals in the 1 H NMR spectrum, thus confirming the anti-alignment of the substituents of natural 1. Subsequently, the configuration of natural 1 was established through a comparison of the NMR data and optical rotation of synthetic 1 with the reported data of the isolated natural compound. See Supplementary Materials for NMR ( Figures S4-S69, Tables S3 and S4). The sign and value of specific rotation α] 25 D + 57.5 (c 0.3, MeOH) of the synthetic sample 1 and perfect match of the NMR spectra confirmed the anti-1 S,3R configuration of the natural heterocornol C (1) (see [8]: α] 20 D + 80 (c 0.30, MeOH)). To reliably prove the relative configuration of 1, we also prepared its diastereomer 18b in a 36% yield by applying the same synthetic procedure. The comparison of the NMR data of the synthetic syn-isomer 18b with the natural 1 data showed large discrepancies in the shifts of the oxepine ring and C-1 and C-3 proton signals in the 1 H NMR spectrum, thus confirming the anti-alignment of the substituents of natural 1.
Melting points were obtained using a Boecius apparatus and are uncorrected. Optical rotations were measured with a JASCO P-2000 polarimeter and are given in units of 10 −1 deg·cm 2 ·g −1 . FTIR spectra were obtained using a Nicolet 5700 spectrometer (Thermo Electron, currently: Thermo Fisher Scientific, Waltham, MA, USA) equipped with a Smart Orbit (diamond crystal ATR) accessory using the reflectance technique (400-4000 cm −1 ). 1 H and 13 C NMR spectra were recorded using either a 300 (75) MHz Unity Inova or a 600 (151) MHz VNMRS spectrometer from Varian. Standard chemical shifts are referenced to the corresponding solvent residual peaks (CDCl 3 : δ H 7.26 ppm, δ C 77.16 ppm; CD 3 OD: δ H 3.31 ppm, δ H 49.00 ppm; DMSO-d 6 : δ H 2.50 ppm, δ H 39.52 ppm) or tetramethylsilane (TMS) as an internal standard. High-resolution mass spectra (HRMS) were recorded with an OrbitrapVelos mass spectrometer (Thermo Scientific, currently: Thermo Fisher Scientific, Waltham, MA, USA) with a heated electrospray ionization (HESI) source. The mass spectrometer was operated with full scan (50-2000 amu) in the positive or negative FT mode (at a resolution of 100,000). The analyte was dissolved in MeOH and infused via a syringe pump at a rate of 5 mL/min. The heated capillary was maintained at 275 • C with a source heater temperature of 50 • C, and the sheath, auxiliary, and sweep gases were used at 10, 5, and 0 units, respectively. The source voltage was set to 3.5 kV. Flash column liquid chromatography (FLC) was undertaken on silica gel Kieselgel 60 (40-63 µm, 230-400 mesh), and analytical thin-layer chromatography (TLC) was performed on aluminum plates precoated with either 0.2 mm (DC-Alufolien, Merck Life Science, Bratislava, Slovakia) or 0.25 mm silica gel 60 F254 (ALUGRAM SIL G/UV254, Macherey-Nagel, Dueren, Germany). The compounds were visualized using UV fluorescence and by dipping the plates into an aqueous H 2 SO 4 solution of cerium sulfate/ammonium molybdate, followed by charring with a heat gun.
The data collection and cell refinement of ent-17a were performed with a SuperNova diffractometer using a CCD detector Atlas S2 and a micro-focus sealed tube with mirrorcollimated CuKα radiation (λ = 1.54184 Å). The structure was solved using the Superflip program and refined using the full-matrix least-squares procedure of the Independent Atom Model (IAM) with Shelxl (ver. 2018/3) [18,19]. The Hirshfeld Atom Refinement (HAR) method was carried out using the IAM model as a starting point. The wave function was calculated using ORCA 4.2.0 software with the basic set def2-TZVP and hybrid exchangecorrelation functional PBE0 [20][21][22]. The least-squares refinement of the HAR model was then carried out with Olex2.refine (ver. 1.5-alpha) [23]. The NoSpherA2 implementation of HAR was used for tailor-made aspherical atomic factors calculated on-the-fly using a Hirshfeld-partitioned electron density [24]. For the HAR approach, all hydrogen atoms were accurately anisotropic, using restraints on the X-H distances of the neutron structures [25]. The structure was drawn using the OLEX2 package [26]. The absolute configuration of the ent-17a stereoisomer was determined using the Parsons and Hooft methods [27,28]. The software CrystalExplorer [29] (version 21.5) was used to calculate the Hirshfeld surface, electrostatic potentials, and associated fingerprint plots [30][31][32].

Conclusions
In summary, we developed a synthetic strategy for the enantioselective construction of a benzo[c]oxepine framework of natural Pestalotiopsis heterocornis metabolites. The key features of the strategy employ the ring-closing metathesis of diene for the assembly of the oxepinone ring and the Sharpless asymmetric epoxidation of allyl alcohol. The applicability of this approach was demonstrated through the synthesis of naturally occurring heterocornol D. Four stereoisomers of this secondary metabolite were prepared from prochiral divinyl carbinol and 2,2-dihydroxy benzoic acid. The absolute configuration of the natural stereoisomer was established based on the single-crystal X-ray analysis of the intermediate ent-17a. The (R)-9-hydroxy-3-((S)-1 -hydroxyethyl)benzo[c]oxepin-1(3H)-one (+)-3a was identified as the naturally occurring heterocornol D (3) based on the comparison of the specific rotation and NMR spectroscopic data. In addition, the synthesis of heterocornol C (1) was achieved by extending the synthetic route using the lactone group reduction method with the prepared heterocornol D (3). Finally, the presented approach can provide access to other natural benzo[c]oxepine-derived polyketides that show improved properties for their evaluation as potential drugs.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/ijms241210331/s1. All experimental procedures and analytical data for all compounds, copies of 1 H and 13 C NMR spectra, HPLC analysis of compounds and crystallographic data are included.

Conflicts of Interest:
The authors declare no conflict of interest.