Total Synthesis of Three Families of Natural Antibiotics: Anthrabenzoxocinones, Fasamycins/Naphthacemycins, and Benastatins

Abstract

Open AccessCCS ChemistryRESEARCH ARTICLE1 Dec 2020Total Synthesis of Three Families of Natural Antibiotics: Anthrabenzoxocinones, Fasamycins/Naphthacemycins, and Benastatins Delu Jiang†, Kunyun Xin†, Baochao Yang, Yanyu Chen, Quan Zhang, Haibing He and Shuanhu Gao Delu Jiang† Shanghai Key Laboratory of Green Chemistry and Chemical Processes, School of Chemistry and Molecular Engineering, East China Normal University, Shanghai 200062 , Kunyun Xin† Shanghai Key Laboratory of Green Chemistry and Chemical Processes, School of Chemistry and Molecular Engineering, East China Normal University, Shanghai 200062 , Baochao Yang Shanghai Key Laboratory of Green Chemistry and Chemical Processes, School of Chemistry and Molecular Engineering, East China Normal University, Shanghai 200062 , Yanyu Chen Shanghai Key Laboratory of Green Chemistry and Chemical Processes, School of Chemistry and Molecular Engineering, East China Normal University, Shanghai 200062 , Quan Zhang Shanghai Key Laboratory of Green Chemistry and Chemical Processes, School of Chemistry and Molecular Engineering, East China Normal University, Shanghai 200062 , Haibing He Shanghai Engineering Research Center of Molecular Therapeutics and New Drug Development, East China Normal University, Shanghai 200062 and Shuanhu Gao *Corresponding author: E-mail Address: [email protected] Shanghai Key Laboratory of Green Chemistry and Chemical Processes, School of Chemistry and Molecular Engineering, East China Normal University, Shanghai 200062 Shanghai Engineering Research Center of Molecular Therapeutics and New Drug Development, East China Normal University, Shanghai 200062 https://doi.org/10.31635/ccschem.020.202000151 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesTrack Citations ShareFacebookTwitterLinked InEmail Due to the critical bacterial resistance and antibiotic crisis, the discovery of new antibiotics is an urgent need in the clinic than ever. Naturally occurring antibiotics have proven to be an indispensable source of the development of new antibacterial agents. Herein, we report the total synthesis of three families of biogenetically related natural antibiotics, including anthrabenzoxocinones (ABXs), fasamycins/naphthacemycins, and benastatins. The synthesis featured divergent and convergent approaches, which enabled efficient construction of the basic polycyclic skeletons in 6–10 steps on a large-scale, followed by a collective synthesis of 14 natural products and their corresponding analogs. The core scaffold of gem-dimethyl-anthracenone, a naturally occurring type II fatty acid-specific condensation enzyme (FabF-specific) antibiotic pharmacophore, was forged via a Ti(Oi-Pr)4-mediated photoenolization/Diels–Alder (PEDA) reaction between 2-isopropyl benzaldehyde and a variety of enones. A scale-up of the PEDA reaction was facilitated in an assembled continuous-flow reactor, which allowed us to overcome the issues associated with batch photochemistry. Subsequently, the synthetic natural antibiotics and their analogs would be utilized in structure–activity relationships (SAR) and mechanism studies, which should enable the discovery of new and leading antibiotic compounds. Download figure Download PowerPoint Introduction Antibiotics discovery was the most important medical revolution of the 20th century and have since improved quality of life and prolonged lifespan significantly.1–8 Since the discovery of penicillin,9,10 small-molecule research has generated numerous clinically useful antibiotics, but the emergence of antibiotic resistance has invoked an urgent need for new compounds than ever. For example, the occurrence of methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant Enterococcus (VRE) are now of major concern in hospitals. Naturally occurring antibiotics such as penicillin G ( 1),9–10 chlortetracycline ( 2),11–13 erythromycin A ( 3)14–16 (Figure 1a), and vancomycin ( 4)17–19 have proven to be indispensable in the development of new antibacterials. Recently, three families of aromatic polyketides, including anthrabenzoxocinones (ABXs),20–24 fasamycins/naphthacemycins,25–30 and benastatins31–36 (Figure 1), have attracted our attention because of their highly potent activities against MRSA and VRE.37 As biogenetically related natural products, these antibiotics were isolated from the same natural source as Streptomyces strains, which share a common tricyclic gem-dimethyl-anthracenone substructure (A–B–C ring) in the polycyclic skeleton (Figure 1b). Biosynthetic studies demonstrated that a methyltransferase (abxM) is responsible for the formation of the gem-dimethyl scaffold.28,29 Figure 1 | The antibiotic drugs and the three families of natural antibiotics polyketides. Download figure Download PowerPoint In 2005, Merck researchers reported that bischloro-anthrabenzoxocinone (BABX) ( 5) inhibited the elongation condensing enzymes (FabF) involved in biosynthesis of type II fatty acids (FASII) essential for bacterial cell viability,20 and therefore, were a potential target for new antibiotics. Additional three FASII inhibitors, including platensimycin,38 phomallenic acid C,39 and plantencin,40 were subsequently discovered via high-throughput screening of the broth extracts derived from cultured bacteria. In 2012, Brady and co-workers28 reported that, similar to BABX, fasamycin A ( 6) inhibited FASII in vitro with IC50 = 50 μg/mL and FabF with 80 μg/mL. In silico docking studies suggested that the chloro-gem-dimethyl-anthracenone substructure in all known FASII inhibitors were capable of binding the FabF active site in the same position and orientation, thereby, identifying the scaffold as a naturally occurring FabF-specific antibiotic pharmacophore.28 The limited amount of material and structural diversities that could be obtained from the natural source hampered further efforts to investigate the antibiotic potential of these compounds. Chemical synthesis is an indispensable approach to provide the diverse target molecules and in adequate amounts. However, synthetic studies toward these three families of natural antibiotics are lacking to date. As part of our ongoing efforts to achieve the total synthesis of bioactive polyketides and understand their medicinal chemistry,41,42 we intended to develop a unified and flexible strategy to prepare efficiently, these three families of natural antibiotics and their analogs in adequate amounts, followed by functional studies to determine their performances. We achieved this goal by designing a highly divergent and convergent approach to build the core gem-dimethyl-anthracenone substructure from various readily available building blocks such as 8–12 (Figure 1c). Our strategy was based on a key Ti(Oi-Pr)4-mediated photoenolization/Diels–Alder (PEDA) reaction for efficient construction of basic polycyclic skeletons on a large scale in 6–10 steps, followed by the demonstration of collective synthesis of 14 natural products and their corresponding analogs. Conventional strategies to construct gem-dimethyl quaternary centers include the reductive Heck reaction,43 radical cyclization,44 and cation-mediated cyclization45. However, these reactions always require harsh reaction conditions and tolerate a narrow range of functional groups. Hauser46,47 and Staunton/Weinreb annulations48–50 are commonly used to synthesize anthraquinones, which involve tandem Michael addition/Dieckmann cyclization. Taking methyl 2-isopropyl-4,6-dimethoxybenzoate 13 and cyclohexenone 11 as model substrates, we investigated the Staunton/Weinreb reaction by screening various bases and solvents (Scheme 1a and Supporting Information Table S1). Given that only starting materials were recovered after all these reactions, we assumed that the tertiary anion species ( I or II, Scheme 1a) formed during the reaction acted as a nonnucleophilic base, and therefore, failed to undergo Michael addition to enone 11. Scheme 1 | Methods for the formation of the gem-dimethyl group and PEDA reaction. Download figure Download PowerPoint We were intrigued by the possibilities to use the PEDA reaction to form the gem-dimethyl-anthracenone structures. The PEDA reaction was initiated by the activation of carbonyl group with UV light, which underwent a single-electron transfer via an excited-state species to generate the highly active diene hydroxy-o-quinodimethane.51–57 Subsequent Diels–Alder cycloaddition with dienophiles yields various highly functionalized polycyclic molecules. In principle, photoenolization of 2-isopropylbenzaldehyde gives rise to a stereohindered hydroxy-o-quinodimethane III bearing two additional methyl groups on the terminal olefin (Scheme 1b). However, this transient diene species shows low reactivities in the following cycloaddition due to stereo effects. Therefore, we proposed to address this issue by adding Lewis acid, such as Ti(Oi-Pr)4, to chelate the diene species and form a stabilized Ti-complex, which should facilitate the following Diels–Alder reaction with cyclohexenone from the endo-direction via transition state IV, generating gem-dimethyl-anthracenol V. Then, a subsequent oxidative aromatization would easily afford gem-dimethyl-anthracenones. We tested the PEDA reaction using electron-rich 4,6-dimethoxy-2-isopropylbenzaldehyde 8a and electron-deficient cyclohexenone 11 as model substrates. After screening the photolytic conditions, we found that stoichiometric amounts of Ti(Oi-Pr)4 promoted intermolecular 4+2 cycloaddition successfully under 300 nm UV light, giving the desired anthracenol 15 as a single diastereomer, as well as its dehydration enone in combined yield of 67% (see Supporting Information Table S2). These results were in good agreement with our previous studies and observation.51 Also, we found that the ratio of 8a to 11 was tunable, which might have been useful if the availability of one of the synthetic materials became limited, especially, in the case of natural products synthesis. To our knowledge, this is the first study that has employed PEDA reaction to build the polycyclic molecules bearing the gem-dimethyl groups, which offers a new methodology for retrosynthetic analysis. Scalability has long been problematic of photochemistry due to the Bouguer−Lambert−Beer limitation on photon transport. In the past decade, continuous-flow photochemistry has been developed successfully to address this issue, and has been utilized by researchers in academia and industry.58 This eco-friendly technology provides new solution in process chemistry with advantages of shorter reaction times, higher specificities, easier scalability, and safer operations. To scale-up the PEDA reaction, we assembled a continuous-flow reactor using crocheted fluorinated ethylene propylene (FEP) tubing, placed inside a Rayonet chamber with 16 UV lamps, which maximized the photon flux hitting the flow solution (Scheme 1c). Then we used a peristaltic pump to control the flow rate, and performed the flow PEDA reaction on a larger scale, obtaining a better yield (see Supporting Information Table S3). With this system, >2 g of cycloadduct 15 (89% yield) was prepared easily using a flow rate of 1 mL/min. This new method might prove more efficient than conventional methods for synthesizing polycyclic anthracenols bearing gem-dimethyl groups. It would also extend the usefulness of the PEDA reaction to achieve a new retrosynthetic disconnection. Results and Discussion Synthesis of ABX-type antibiotics and analogs We selected ABX-type antibiotics as the first target molecules. To date, more than 20 members of this family have been isolated from different actinomycetes (Scheme 2a and Supporting Information Figures S1 and S4),20–24 and shown to contain a strained hexacyclic aromatic skeleton, which includes a chloro-gem-dimethyl-anthracenone (A–B–C ring) and a unique bridged ketal scaffold (D–E ring) at C-6, composed of a hydroxyl group at C-16 and a phenol group at C-5. The synthetic challenge to reach the goal of efficiency and diversity relied on the rational ring forming sequence and selective chlorination. Accordingly, we planned a highly convergent approach by using the Ti-promoted PEDA reaction to install directly the hexacyclic skeleton, which could serve as an advanced intermediate to prepare ABXs and their analogs. Scheme 2 | Total synthesis of the ABX-type antibiotics ABX-A, -C, -E, -F, -H, and BABX, as well as analogs. DMP: Dess–Martin periodinane, IBX: 2-iodoxybenzoic acid, DDQ: 2,3-dichloro-5,6-dicyano-1,4-benzoquinone, TBAF: tetrabutylammonium fluoride. Download figure Download PowerPoint Based on this proposition, our synthesis started from the preparation of the right fragment of the D–E–F ring bearing a sensitive ketal moiety (Scheme 2b). Heating a neat solution of silyl enol diene 16 and unsaturated aldehyde 17 initiated the intermolecular Diels–Alder reaction,59 which gave rise to the highly substituted cyclohexenol C ring 18 in 60% yield as a mixture of two diastereomers at C-15 (d.r. = 1.5∶1). Lithium bromide exchange of the F ring 19 formed aryllithium reagent, which underwent nucleophilic addition of 18 to yield benzylic alcohol 20a. The reaction was quenched directly with 2N hydrochloric acid at 50 °C, giving rise to the ketal 21 in 57% yield, which was oxidized with Dess–Martin periodinane to afford the desired enone 9 as a single diastereomer. The relative stereochemistry of ketal 9 was confirmed by X-ray diffraction analysis. We reasoned that the bridged ketal scaffold was generated stereospecifically through formation of C6–O and C16–O bonds in a cascade process involving acid-promoted dehydration of the benzylic alcohol, removal of ketal/MOM protecting groups, formation of o-quinone methide species, and an intramolecular hetero-Diels–Alder reaction ( 20a→ 20b→ 21). We then investigated the Ti-promoted PEDA reaction with enone 9 and 2-isopropylbenzaldehyde 8b. Under optimal conditions, UV irradiation (300 nm) of a flowing solution of 9 and 8b in the presence of Ti(Oi-Pr)4 in the continuous-flow reactor gave the desired adduct 22 in 89% yield (endo:exo = 3∶1). This adduct contained the basic hexacyclic skeleton as the natural molecules but in a lower oxidation state. The fact that the sensitive ketal group remained unaltered during this reaction highlighted the mild conditions of the PEDA reaction and its synthetic potentials. Next the benzylic alcohol group was oxidized by 2-iodoxybenzoic acid (IBX) to a carbonyl group,60 followed by oxidative aromatization with 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ), giving the gem-dimethyl-anthracenone 23. This six-step synthesis enabled the efficient preparation of the core structure of ABX-type antibiotics on a gram scale. Significantly, four rings (B–C–D–E rings) were forged through three Diels–Alder reactions in this sequence. Then we turned our attention toward finishing the synthesis of ABX-type antibiotics and analogs using 23 as a precursor (Scheme 2c). Three phenols on the A and F ring were masked using different protecting groups: methyl (Me), benzyl (Bn), and triisopropylsilyl (TIPS) group, which facilitated selective removal operations. Lewis-acid-mediated (magnesium iodide)61 removal of methyl group, followed by the deprotection of TIPS and Bn group gave ABX-A ( 27; Supporting Information Tables S5 and S6) in good yield over three steps. Palladium-catalyzed hydrogenation of 23 afforded the desired phenol, which was protected as its methoxyl group to give compound 24. Repeating the procedure used to prepare 25 afforded ABX-E ( 26; Supporting Information Tables S9 and S10) in 94% yield over two steps from 24. To realize regioselective chlorination of this aromatic skeleton without overchlorination62–68 was challenging because the multiple potential reactive sites showed only subtle electronic differences. Thus, to achieve this, we deprotonated phenol 25 with n-butyl lithium to generate the oxyanion species whose ortho position is more reactive. Tert-butyl hypochlorite (t-BuOCl) was added as a chlorine source to install the mono- and dichlorination products, which were transformed into a mixture of ABX-F ( 28; Supporting Information Table S7) and ABX-H ( 29; Supporting Information Table S8), respectively, after deprotection of Bn group. Following the same procedure, ABX-E ( 26) was converted efficiently to a mixture of ABX-C ( 30; Supporting Information Figures S11 and S12) and BABX ( 5; Supporting Information Tables S13 and S14) in 31% and 27% yields, respectively. Synthesis of fasamycins, naphthacemycins, and analogs After the discovery of fasamycins A and B by Brady and co-workers,28,29 new biogenetically related members of this subgroups of antibiotics were isolated successively from different Streptomyces strains. In 2017, Wilkinson and coworkers30 reported the isolation of fasamycins C–E and formicamycins A–M from a new Streptomyces species, Streptomyces formicae derived from the African fungus-growing plant-ant Tetraponera penzigi. Subsequently, Ōmura and coworkers26,27 discovered 17 naphthacemycins (A1–A11, B1–B4, and C1–C2) from a cultured broth of Streptomyces sp. KB-3346-5 (Scheme 3a and Supporting Information Figure S2). Sharing the same pharmacophore of chloro-gem-dimethyl-anthracenone substructure, formicamycins and naphthacemycins also showed growth inhibition activities of the clinically relevant pathogens MRSA and VRE ( Supporting Information Figure S4). It has been reported that formicamycins, especially, poly-halogenated congeners containing up to four chlorine atoms, were more potent against MRSA and VRE than fasamycins, which might be attributable to an increased lipophilicity and an enhanced ability to cross the bacterial cell membrane. Significantly, naphthacemycins increased imipenem activities 100–500 times against MRSA at a concentration of 0.5 μg/mL, and naphthacemycins A4–A11 showed MIC50 values of 1–4 μg/mL against 22 MRSA strains.25–30 Accordingly, their high potential as new chemotherapeutic antibiotics agent attracted the researchers attention to carry out the chemical synthesis and medical chemistry of these antibiotics.69,70 Scheme 3 | Total synthesis of fasamycins/naphthacemycins and analogs. Me3SiNTf2: N-(trimethylsilyl)bis(trifluoromethanesulfonyl)imide, TMSOTf: trimethylsilyl trifluoromethanesulfonate, LDA: lithium diisopropylamide, TMSCl: chlorotrimethlsilane. Download figure Download PowerPoint This family of natural antibiotics can be further classified into three subgroups based on structural features ( Supporting Information Figure S2), subgroup 1: fasamycins A–E and naphthacemycins B1–B4 possessed a pentacyclic aromatic skeleton including a chloro-gem-dimethyl-anthracenone (A–B–C ring) and a chiral axis of the C6–C7 bond connecting D and E rings; subgroup 2: formicamycin had the same scaffold with fasamycins except a nonaromatic C-ring and two additional chiral centers at C-10 and C-19; subgroup 3: naphthacemycin A series had a unique quinone C ring and C series has a semiquinone structure with an all carbon center at C-21. To better understand the structure–activity relationship (SAR), we selected fasamycins/naphthacemycins as target molecules, which had similar aromatic skeleton with ABXs and benastatins. The synthesis of this family of molecules was divided into two stages: construction of the pentacyclic skeleton (Scheme 3b) and selective chlorination (Scheme 3c). Silyl enol diene 31 and cyclohexenone 11 underwent a Lewis-acid-mediated Diels–Alder reaction to furnish tricyclic 32 (C–D–E ring) in 60% yield as a single diastereomer.71 Selective protection of carbonyl group on D ring as its ketal using bis-TMS ether reagent is done in the presence of catalytic TMSOTf, which was converted to enone 33 through a sequential Saegusa–Ito oxidation. The relative stereochemistry of the ketal 33 was assigned based on X-ray diffraction analysis. Then enone 33 underwent an intermolecular PEDA reaction with 2-isopropylbenzaldehyde 8b to install the pentacyclic ring skeleton. Photolysis of a solution of 33 and 8b with Ti(Oi-Pr)4 in dioxane in the continuous-flow reactor afforded the required cycloadduct 34 in 84% yield as a single diastereomer. The stereochemistry of 34 was confirmed by analogy with a structurally related adduct (see Supporting Information), which revealed that the PEDA reaction was controlled by a Ti-chelated transition state like IV (Scheme 1b) and occurred stereospecifically through an endo-direction. We envisioned that 34 could be used as a precursor to prepare these three subgroups of fasamycin/naphthacemycins natural antibiotics by selective functionalization. In order to target fasamycins and naphthacemycins, the oxidation state of B–C–D ring of 34 was adjusted by a sequence of oxidation steps. The benzylic alcohol was oxidized sequentially with IBX and then DDQ to give 35 with an aromatic C ring in good yield over two steps. After removal of the ketal group on C-23, we tried to aromatize the D ring using conventional oxidative conditions, but this either overoxidized 35 or led to recovery of the starting ketone. Extensive screening of reaction conditions showed that the Pd-catalyzed aerobic dehydrogenation, developed by Stahl and transformed into phenol in 89% Thus, we employed this to 31 to with a gem-dimethyl-anthracenone core on the gram scale, which was used successfully as an advanced intermediate to prepare fasamycins, naphthacemycins, and their analogs. of all protecting groups in generated naphthacemycin ( Supporting Information Tables and via naphthacemycin ( Supporting Information Tables and in good yield Scheme 3c). Selective of site on the D ring was by using its group as a group. of with formed an oxyanion species, which with to give followed by into naphthacemycin ( Supporting Information Tables and and analogs by of deprotection steps. In these steps, we first masked the phenol on the D ring as a group and two phenols on the A ring to yield compound which was to investigate the selective chlorination (see Supporting Information Table S4). the same conditions as in the preparation of mono- and A ring products and were After removal of the protecting groups on the D and E naphthacemycin ( and fasamycin B ( Supporting Information Tables and were achieved in good yields and We also found that E ring of could be using to form naphthacemycin ( after the removal of protecting groups on E Synthesis of and analogs In and reported the isolation of from the broth of Streptomyces sp. (Scheme and Supporting Information Figure S3). of B generated also known as studies showed that A and B only growth of including methicillin-resistant but also inhibited the of Supporting Information Figure and were as of the revealed that A also and cell in and these were to be by the inhibition of In the and of the of the enzymes that synthesize revealed that one by generated the gem-dimethyl group. of these enzymes were used to prepare which had In the the gem-dimethyl scaffold also oxidation and Scheme | Total synthesis of and analogs. acid, Download figure Download PowerPoint the highly or hexacyclic rings containing gem-dimethyl-anthracenone ring) and In and between the acid on and the on of the E ring gives rise to an or F We a convergent strategy involving the and rings to generate hexacyclic skeleton with an which could to and (Scheme 15 was used to prepare the aromatic IBX oxidation of the benzylic alcohol, followed by oxidative aromatization with in the presence of and gave To prepare the was to with and to give the desired in yield over three steps. the method of and the ring in was stereospecifically via a process using generating the ring in After the selective and terminal was obtained in yield, over two steps. A the two 50 and to afford whose was via Pd-catalyzed hydrogenation followed by selective the as the precursor of ring an solution of to UV nm) in the presence of formed the hexacyclic core in yield, whose structure was confirmed by X-ray diffraction This ring through radical with of the This synthetic sequence the basic hexacyclic efficiently on a large scale. To generate ( we and of in the presence of then the on the phenol (Scheme We confirmed the structure using X-ray diffraction ( Supporting Information Tables and with and then the to afforded the intermediate which was converted to G ( Supporting Information Tables and in We prepared F ( which contained the same skeleton but had a bond in the C ring and a higher oxidation state than G by hydrogenation of using to afford the in of the C ring via formed which was converted readily to F ( Supporting Information Tables and the same procedure used in we then employed the same radical reaction to prepare pentacyclic bearing an on We started from 50 and terminal which we prepared from in two steps (see in Supporting by the of the same including and selective the key precursor in three steps. After the selective and radical pentacyclic was obtained and its structure was confirmed by X-ray diffraction which was converted readily to B ( Supporting Information Tables and via the The of these synthetic were in good agreement with the corresponding of the natural Total synthesis of three families of biogenetically related natural antibiotics, including (ABXs), fasamycins/naphthacemycins, and