Synthetic Studies toward (−)-Cleistenolide: Highly Stereoselective Synthesis of New γ-Lactone Subunits

This study describes the stereoselective synthesis of two new γ-lactones in 6 and 3 steps and 19 and 32% yield, respectively, directed toward the total synthesis of the natural product (−)-cleistenolide. The starting material was an enantiomerically pure diacetonide diol, derived from d-mannitol with the required stereocenters for (−)-cleistenolide synthesis. γ-Lactone syntheses were based on highly selective protection and deprotection of hydroxyls from d-mannitol. The formation of γ-lactone rings was the culmination of this approach, made possible by a Horner-Wadsworth-Emmons Z-olefination between diacetal aldehyde and ethyl 2-(bis(o-tolyloxy)phosphoryl)acetate to produce an unsaturated ester. The Z-isomer ester was highly favored in relation to the E-isomer (Z/E ratio of 94:6), allowing the formation of the γ-lactone ring under acid catalysis. This strategy precluded the use of chiral auxiliaries or catalysts for the control of stereocenters in the novel γ-lactones.


Introduction
(−)-Cleistenolide (1) is an exponent of the class of 5,6-dihydro-2H-pyran-2-ones. 1-3 It contains a d-lactone ring, a moiety found in the structure of several natural products with expressive antibiotic properties and anticancer activity against a broad spectrum of tumor cells. [4][5][6][7][8][9][10][11][12][13][14] (−)-1 and the polyoxygenated compound cleistodienol ( Figure 1) were isolated by Samwel et al. 15 in 2007 from Cleistochlamys kirkii (Benth.) Oliv. The plant species belongs to the family Annonaceae and originated from Tanzania and Mozambique. The use of C. kirkii extract in traditional medicine to treat infections, tuberculosis, and rheumatism 1,15-21 motivated researchers to investigate the phytochemical properties of the plant. (−)-1 showed in vitro antibacterial activity against Bacillus anthracis and Staphylococcus aureus and antifungal activity against Candida albicans. 15,17 Because of the pharmacological potential of (−)-1, its interesting chemical architecture, and its low availability in nature (only 200 mg can be extracted from 1 kg dry weight of plant material), 22 several researchers around the world have focused on its synthesis. Some studies 17,23 indicated that the α,β-unsaturated d-lactone ring of (−)-1 can act as a Michael acceptor, which further enhances its biological action.
The first total synthesis of (−)-1 was performed in 2010 by Schmidt et al. 19 using a mannitol-derived compound as starting material. The researchers obtained

Synthetic Studies toward (−)-Cleistenolide: Highly Stereoselective Synthesis of New γ-Lactone Subunits
(−)-1 in an overall yield of 18% by applying a ring-closing metathesis protocol to prepare the basic component of the d-lactone ring. 19 In the same year, Cai et al. 22 reported the stereoselective synthesis of (−)-1 in eight steps and 49% overall yield. The reaction started from the natural sugar d-arabinose. The authors 22 used Wittig olefination to form the α,β-unsaturated carbonyl unit, and Yamaguchi lactonization to afford the d-lactone ring.
In the studies of Babu et al., 20,21 Ramesh and Meshram, 24 and Karier et al., 25 the d-lactone ring of (−)-1 was also obtained by ring-closing metathesis. Ghogare et al. 26 and Reddy et al. 1,27 used the Still-Gennari protocol to obtain cis-olefin, a key intermediate that facilitated obtaining the lactone ring of (−)-1. Benedekovic et al. 28 carried out Wittig olefination at a low temperature to favor the formation of Z-alkenes, followed by lactonization under acid catalysis to form the piran-2-one ring.
In the above-mentioned synthetic studies, 19-21,24-28 several interesting conventional methodologies were applied, alone or in combination, to obtain the stereogenic centers found in the natural product. The current study reports an unprecedented approach for the stereoselective synthesis of two γ-lactones whose stereocenters originated from d-mannitol. The γ-lactones are considered advanced intermediates in the total synthesis of (−)-1. Their synthesis was made possible by using highly selective steps for protection and deprotection of hydroxyl groups in the starting material.

Experimental
General 1 H and 13 C nuclear magnetic resonance (NMR) spectra were recorded on a Bruker Avance spectrometer at 400 and 100 MHz, respectively, using CDCl 3 and CD 3 OD as solvents. Chemical shifts (d) are given in ppm and coupling constants (J) in Hz. 1 H NMR chemical shifts are reported relative to tetramethylsilane (TMS) in CDCl 3 and using the residual signal of CD 2 HOD (d = 2.50 ppm) in CD 3 OD as a reference. 13 C NMR chemical shifts were recorded using CDCl 3 (d = 77.2 ppm) and CD 3 OD (d = 49.2 ppm) signals as references. 1 H and 13 C NMR spectra of compounds can be found in the Supplementary Information (SI) section. Liquid chromatography-tandem mass spectrometry (LC-MS/MS) analyses were performed on a Nexera UHPLC system (Shimadzu) hyphenated to a maXis ETD high-resolution (HR) electrospray ionizationquadrupole time-of-flight (ESI-QTOF) mass spectrometer (Bruker). Infrared (IR) spectra were recorded on a Varian 660 FTIR spectrometer equipped with a diamond attenuated total reflectance (ATR) accessory as a thin film. Melting points were determined on an MQAPF-302 Microchemical apparatus and uncorrected. [α] D values were measured using an Anton Paar MCP 300 polarimeter equipped with a 589 nm wavelength sodium vapor lamp at the Natural Products Laboratory 610 of the Oswaldo Cruz Foundation, Instituto René Rachou, Belo Horizonte City, Minas Gerais State, Brazil. Solution concentrations are denoted as c (g per 100 mL), followed by the solvent used. Purifications were performed by conventional or flash column chromatography on silica gel or neutral alumina. Tetrahydrofuran (THF) was distilled from sodium metal and benzophenone ketyl under nitrogen. Dimethylformamide (DMF) and dichloromethane were distilled from CaH 2 . Acetonitrile was dried with anhydrous MgSO 4 , distilled, and stored under 3 Å molecular sieves. Methanol and ethanol were dried over Mg 0 and I 2 (cat.) under reflux until total consumption of Mg 0 , distilled, and stored under 3 Å molecular sieves. Thin layer chromatography (TLC) visualization was achieved under ultraviolet light (254 nm) or by spraying with KMnO 4 solution (1.0 g KMnO 4 , 6.66 g K 2 CO 3 , and 1.66 mL 5% KOH in 100 mL distilled water) and heating and/or resublimed iodine. All chemicals were used as received unless otherwise stated. Syntheses (4R,4'R,4"R,5'R)-2,2,2',2',2",2"-Hexamethyl-4,4',5',4"-tert-(1,3-dioxolane) (10) To a mixture of acetone (75 mL) and sulfuric acid (0.5 mL) was added d-mannitol (5.052 g, 27.77 mmol) in portions and stirred for 18 h at room temperature. The mixture was neutralized by addition of aqueous NH 4 OH (1.75 mL) and then Na 2 CO 3 (3.125 g, 29.48 mmol). The residue was filtered, and the filtrate was concentrated to a volume of 20 mL under vacuum. Then, the solution was immersed in an ice bath to precipitate the product as a white solid. The solid product was collected and recrystallized from acetone to give 10 as crystals in 76% yield, mp 65. 5  To a solution of triacetonide 10 (1.958 g, 6.85 mmol) in anhydrous methanol (78 mL) under an argon atmosphere and in an ice bath (0 °C) was added previously distilled acetyl chloride (28.03 mmol, 2.0 mL) dropwise. The mixture was stirred for 5 min at 0 °C, saturated aqueous K 2 CO 3 (50 mL) was added, and the residue was extracted with EtOAc (3 × 80 mL). The combined EtOAc layers were washed with brine, dried with anhydrous Na 2 SO 4 , filtered, and concentrated under vacuum. The residue was purified by column chromatography (SiO 2 , hexane/EtOAc 90:10 → 60:40) to give 9 as a colorless oil in 29% yield and 41% (806.3 mg) of recovered starting material (10).
Study of the selective deprotection of triacetonide 10 ( Table 1) Entries 1-9 Pyridinium p-toluenesulfonate (PPTS) (1.2 or 0.2 equiv) was added to a solution of triacetonide 10 (3.795 g, 12.57 mmol) in anhydrous methanol (73 mL) and stirred under the time and temperature conditions specified in entries 1-9 of Table 1. The mixture was neutralized by addition of NaHCO 3 (15.14 mmol), filtered, and the filtrate concentrated in vacuum. Then, water was added (70 mL), and the product extracted with EtOAc (3 × 70 mL). The combined EtOAc layers were washed with brine, dried with anhydrous Na 2 SO 4 , filtered, and concentrated under vacuum. The residue was purified by column chromatography (SiO 2 , hexane/EtOAc 90:10 → 60:40) to give 9 as a colorless oil in 39% and recover 47% of the starting material (10)  Iron(III) chloride hexahydrate (1 mol%) was added to a solution of triacetonide 10 (2.014 g, 6.67 mmol) in anhydrous CH 2 Cl 2 (100 mL) and stirred under the time and The mixture was neutralized by addition of saturated aqueous NaHCO 3 . The aqueous layer was extracted with CH 2 Cl 2 (3 × 100 mL), and the combined EtOAc layers dried over anhydrous Na 2 SO 4 , filtered, and concentrated under vacuum. The residue was purified by column chromatography (SiO 2 , hexane/EtOAc 90:10 → 60:40) to give 9 and recover the starting material (10).

Entry 13
Bismuth(III) chloride (5 mol%) was added to a solution of triacetonide 10 (3.457 g, 11.45 mmol) in anhydrous CH 3 CN (114 mL) and stirred for 10 min. The mixture was neutralized by addition of NaHCO 3 (1.145 mmol), filtered, and the filtrate was concentrated in vacuum. Then, water (60 mL) was added, and the product extracted with EtOAc (3 × 60 mL). The combined EtOAc layers were washed with brine, dried with anhydrous Na 2 SO 4 , filtered, and concentrated under vacuum. The yellowish residue was purified by column chromatography (SiO 2 , hexane/EtOAc 90:10 → 60:40) to give 9 in 23% yield and recover 4% of the starting material 10.

Ethyl 2-(bis(o-toluyloxy)phosphoryl)acetate (8)
To a two-port flask containing 5 mol L −1 PCl 3 (5 mL, 57.21 mmol), freshly distilled, in anhydrous CH 2 Cl 2 (6.44 mL) at 0 °C was added 10 mol L −1 t-BuOH (5.72 mL, 57.21 mmol) in anhydrous CH 2 Cl 2 (0.29 mL), dropwise. The resulting mixture was stirred at 0 °C for 1 h, producing bubbles of HCl gas as a byproduct, which was purged (bubbling in water) at short intervals. Then, a 5 mol L −1 solution of o-cresol (11.74 mL, 114 mmol) in anhydrous CH 2 Cl 2 (11.06 mL) was added dropwise over 30 min. The mixture was stirred at 0 °C for 30 min and at ambient temperature for a further 12 h. The solvent was removed by rotary evaporation under vacuum, generating diphenylphosphite as a yellowish residue. It was used for the next step without previous purification.
Study of the selective deprotection of 6 ( Table 2) Entries 1-2 PPTS (0.2 equiv) was added to a solution of diacetal ester 6 (0.375 g, 1.25 mmol) in anhydrous methanol (7.3 mL) and stirred at 35 °C under the conditions described in entries 1 and 2 of Table 2. The mixture was neutralized by addition of NaHCO 3 (1.5 mmol), filtered, and the filtrate concentrated in vacuum. Then, water (15 mL) was added, and the product extracted with EtOAc (3 × 15 mL). The combined EtOAc layers were washed with brine, dried with anhydrous Na 2 SO 4 , filtered, and concentrated under vacuum. The residue was purified by column chromatography (SiO 2 , hexane/EtOAc 90:10 → 60:40) to give 11 and recover 6.

Entry 3
p-Toluenesulfonic acid (1 mol%) was added to a solution of diacetal ester 6 (0.552 g, 1.84 mmol) in anhydrous methanol (5 mL) and stirred at room temperature for 16 h. The mixture was neutralized by addition of K 2 CO 3 (0.20 mmol) and concentrated under vacuum. Then, water (30 mL) was added, and the product extracted with EtOAc (3 × 30 mL). The combined EtOAc layers were washed with brine, dried with anhydrous Na 2 SO 4 , filtered, and concentrated under vacuum. The residue was purified by column chromatography (SiO 2 , hexane/EtOAc 90:10 → 60:40) to give 11 in 33% yield and recover 16% of 6.
In the present study, the synthesis of γ-lactone 3 began with the preparation of diol-diacetonide 9 as starting material. This compound (9) was synthesized for the first time from d-mannitol by L. F. Wiggins 31 in 1946 (Scheme 2). Several attempts were made to obtain diacetal diol 9 by selective deprotection of triacetonide 10 using the Wiggins method; however, results were unsatisfactory. Wiggins described obtaining a mixture of monoacetylated and diacetylated products in addition to the recovered starting material, which, in practice, constitutes a protocol that is difficult to reproduce, as evidenced by our results. Subsequently, we attempted an alternative procedure, described by Chandrasekhar et al., 33 for the selective deacetylation of a benzyl ether diacetonide using acetyl chloride (5 equiv and 5 min) in methanol at 0 °C. Acetal diol 9 was obtained in 29% yield after chromatographic purification, and 41% of the starting material was recovered. The method described by Reddy et al.,1,27 in which 60% AcOH is used, was also tested, but no satisfactory results were obtained.
Given the disappointing results of the preparation of diacetal diol 9, we found that additional studies with other deprotection reagents and different reaction conditions were necessary (Table 1). To the best of our knowledge, at present, there are no effective methods for obtaining diacetonide 9 from its triacetonide other than that reported by Wiggins. 31 Investigations started with the use of pyridinium p-toluenesulfonate (PPTS) 34 (1.2 equiv) in methanol. Reactions were carried out at room temperature and 35 °C (entries 1-5, Table 1). The highest yield was 34% (entry 5). The use of PPTS (0.2 equiv) at room temperature did not improve the results (entries 6 and 7). When the reaction was carried out at 35 °C for 3 h (entry 8), 9 was obtained in 39% yield (57% based on the recovered starting material). However, when this same reaction was extended to 6 h (entry 9), the yield decreased to 31%. The use of FeCl 3 ·6H 2 O 35 (entries 10-12), 5 mol% BiCl 3 (entry 13), 36 and Amberlyst-15 (entry 15) 37 at room temperature did not substantially improve the results. The conditions used to obtain 9 in 57% yield (entry 8) were selected as the most efficient, as the yield was much higher than that reported by Wiggins (28% yield based on recovered starting material).
After obtaining diol 9, the compound was treated with sodium periodate to afford the corresponding aldehyde (7) (Scheme 2). Compound 7 was then immediately subjected to HWE olefination with Ando 30 phosphonate 8 (ethyl 2-(bis(o-tolyloxy)phosphoryl)acetate), prepared in our laboratory, affording the expected Z-olefin ester 6 with high diastereoselectivity (Z/E, 94:6), as verified by 1 H NMR spectroscopy of a crude aliquot of the reaction ( Figure S10, SI section). After purification by neutral alumina column chromatography, diacetonide ester 6 was subjected to another step of selective deprotection using BiCl 3 in catalytic quantity (5 mol%) in acetonitrile 34 to form diol ester 11 in 57% yield with 20% of recovered starting material.
For improvements in the yield of 7, we carried out a new study of the selective deacetylation of substrate 6 ( Table 2). Table 2 shows that the best condition for selective deacetylation from 6 was the use of a 2 mol L −1 HCl aqueous solution in ethanol at room temperature for 9 h, affording diol 11 in 68% yield (82% based on recovered starting material) (entry 5). This result was slightly higher than that obtained with 5 mol% BiCl 3 in acetonitrile 36 (57% yield, Scheme 2), with the additional advantage of precluding the use of a toxic metal catalyst.
Unexpectedly, when diacetal 6 was treated with Amberlyst-15 37 (0.02 equiv) for 72 h in acetone/distilled water (6.6:3.4), it provided γ-lactone 15 in 63% yield (Scheme 3). We believe that total deacetylation of 6 and concomitant closure of the ring generated the intermediate γ-lactone triol 14. Acetalization of 14 in situ was possible because of the presence of acetone (used as solvent) between the neighboring primary and secondary hydroxyls, thus forming the ketal γ-lactone 15 in a single step.
Subsequent steps in the total synthesis of (−)-1 should be aimed at the silylation of hydroxyl groups in 3 and 15 to obtain the advanced trisilylated and monosylated intermediates 4 and 5, respectively (Scheme 1). Some attempts have already been made in this regard by using TMSCl and TMSOTf; however, the yields of the silylated products were very low and catalysts caused the degradation of the starting intermediates. Additional investigations of experimental procedures are necessary to assess the effects of changes in reaction temperature and reagent quantity and optimize reaction times. Further studies aimed at the total synthesis of d-lactone (−)-1 are ongoing in our laboratory.

Conclusions
A new approach to stereoselective synthesis allowed obtaining γ-lactones 3 and 15 in 19 and 32% overall yield, respectively. These lactones are key intermediates in the total synthesis of (−)-1 from diacetonide diol 9 (derived from d-mannitol, which is abundant in nature, commercially available, and low cost). The approach consisted of highly selective chemical manipulations that enabled the use of polyol 9 as a source of stereocenters for γ-lactones 3 and 15, which can then be transferred to (−)-1. Studies in this direction are ongoing in our laboratory, and the results will be published later.

Supplementary Information
Supplementary data ( 1 H, 13 C NMR, and mass spectra) are available free of charge at http://jbcs.sbq.org.br as PDF file.