Synthesis of Several Cleistrioside and Cleistetroside Natural Products via a Divergent De Novo Asymmetric Approach

Abstract

The de novo asymmetric syntheses of several partially acylated dodecanyl tri- and tetra-rhamnoside natural products (cleistriosides-5 & 6 and cleistetrosides-2 to 7) have been achieved (19 to 24 steps). The divergent route requires the use of three or less protecting groups. The asymmetry was derived via Noyori reduction of an acylfuran. The rhamno-stereochemistry was installed by a diastereoselective palladium-catalyzed glycosylation, ketone reduction and dihydroxylation.

The quest to find new natural products with interesting biological activity has led to the discovery of several partially acylated dodecanyl tri- and tetra-rhamnoside natural products with significant antibacterial activity from Cleistopholis patens and glauca (Figure 1).¹ For instance, cleistrioside-5 (1) and cleistetroside-2 (3) have shown significant antimicrobial activity against several methicillin-resistant Staphylococcus aureus (e.g., ATCC 33592 0.5 µg/mL for 3 and 8 µg/mL for 1). Due to limited supplies of the isolated materials only two members (1 and 3) of this class of natural products have been extensively studied for biological activity and only cleistetroside-2 (3) has succumbed to total synthesis.² However, some related acylated rhamnoside natural products have possessed anticancer activity.³

Figure 1.

Targeted cleistriosides and cleistetrosides

As part of a program of de novo synthesis and study of biologically active partially acylated rhamnoside natural products,⁴ we became interested in their synthesis and biological study. Key to the synthetic efficiency of this approach is the use of a divergent strategy and the minimal use of protecting groups. Herein we describe our successful development of a divergent de novo approach to eight members of this class of natural products.

Retrosynthetically, we envisioned that the eight oligosaccharides could be efficiently prepared from a common trisaccharide intermediate 9 (Scheme 1), where the pyranone ring could be elaborated into the required rhamno-mono- and di-saccharide with the use of one or less protecting groups.⁵ In turn the key intermediate 9 could be stereoselectively prepared from achiral acetyl furan 10 via our de novo asymmetric approach. The route as envisioned should allow for the selective preparation of the eight target molecules (1–8) with the use of one acetonide and one or two chloroacetate groups.

Scheme 1.

Retrosynthetic analysis

Our synthesis started with the commercially available pyranone 11 (Scheme 2), which also can be synthesized from achiral acetyl furan 10 in three steps (71%).⁶ Exposure of the pyranone 11 and dodecyl alcohol to our typical Pd-catalyzed glycosylation conditions (2.5 mol % Pd₂(dba)₃•CHCl₃ and 10 mol % of PPh₃ in CH₂Cl₂ at 0 °C)⁷ produced pyranone 12 in 87% yield with complete α-selectivity. A Luche reduction and Upjohn dihydroxylation diastereoselectively produced a rhamno-triol, which was then regioselectively protected as acetonide 13 (81%, three steps). Using a similar 4-step sequence the free equatorial hydroxyl group at C4 in 13 was glycosylated. This began with our Pd-catalyzed glycosylation of 13, followed by NaBH₄ reduction, acetylation (Ac₂O/Py) of the resulting allylic alcohol and finally Upjohn dihydroxylation, which provided diol 14 (68% yield for the four steps).

Scheme 2.

Approach to the Disaccharide 14

With the diol 14 in hand, we then turned our attention to the regioselective glycosylation of C3 hydroxyl group of the diol 14 (Scheme 3). We first investigated a C2 protection/C3 glycosylation strategy but failed to find a suitable procedure to selectively install an acetate-orthogonal protecting group that would selectively react at the C2 position. So, we turned to the selective glycosylation of diol 14. Exposing 14 to our typical Pd-catalyzed glycosylation with pyranone 11 gave a 4:1 ratio of regioisomeric trisaccharides 16 and 17, with the undesired isomer 16 being the major product (71% of 16).

Scheme 3.

Regioselective glycosylation

Given that tin ethers were known to react with Pd-π-allyl intermediates and that the stannylene complex of 2,3-manno-diols were known to direct alkylation⁸/acylation⁹ and silylations¹⁰ to the C3 hydroxyl group, we decided to use the 2,3-O-cis-stannylene acetal 15 to try to switch the regioselectivity of the Pd-catalyzed glycosylation. In practice, we treated diol 14 with a slight excess of the in situ generated Bu₂Sn(OMe)₂ before proceeding with our normal Pd-catalyzed glycosylation procedure. To our delight, exposure of the stannylene acetal complex 15 and pyranone 11 to our typical conditions gave the desired glycosylation product 17 in 76% yield (1:7, 16/17). Subsequent chloroacetylation of the C2-hydroxyl group using chloroacetic anhydride in the presence of catalytic amount of DMAP in pyridine provided 9 in excellent yield (97%).

With access to the key intermediate 9 for our proposed divergent synthesis, we embarked on the synthesis of the two cleistriosides (1 and 2) via post-glycosylation/deprotection sequences (Schemes 4 and 5). As before, a sequential Luche reduction, acetylation and Upjohn dihydroxylation installed the rhamno-stereochemistry producing diol 18 in 77% yield (for 3 steps). Selective acetylation of the C2 axial hydroxyl group of diol 18 was successfully achieved using orthoester chemistry forming 19 (CH₃C(OCH₃)₃/p-TsOH then 90% AcOH/H₂O, 96%).¹¹ Removal of chloroacetyl group using thiourea (3 equiv) in the presence of NaHCO₃ (3.3 equiv.) and catalytic Bu₄NI,¹² followed by deprotection of the acetonide group using 80% AcOH/H₂O^2a furnished the target cleistrioside-5 (75% yield, two steps).

Scheme 4.

Approach to cleistrioside-5

Scheme 5.

Approach to cleistrioside-6

The same intermediate 9 was used to synthesize the other trisaccharide, cleistrioside-6 (Scheme 5). For example, switching a chloroacetylation for acetylation step in the post-glycosylation sequence of 9 gave 20 (66% yield, 3 steps). A subsequent per-acetylation (21 in 97%) and similar deprotection sequence generated the target cleistrioside-6 (89% yield, 2 steps).

We believed that the key intermediate 9 from these two syntheses could also be used for the further divergent synthesis of the desired cleistetrosides (3–8). This would require a further branching point at the above trisaccharides 19 and 20, which vary by an acetyl vs chloroacetyl group at C4 in the third sugar (Schemes 6 and 7).

Scheme 6.

Synthesis of the key intermediates of 22–26

Scheme 7.

Synthesis of the key intermediates 27

We first turned our attention to the synthesis of the five required protected cleistetrosides (22–26) from the pivotal intermediate 19 (Scheme 6). Sequential Pd-catalyzed glycosylation, NaBH₄ reduction and Upjohn dihydroxylation of 19 afforded the desired triol 22 (64% yield, three steps), which was then peracetylated to give 23 (96% yield). By incorporating an acylation step in the synthesis of 22 tetrasaccharide 25 was prepared from 19 (60% yield, four steps). Whereas, the incorporation of two additional steps (chloroacylation at C4 and orthoester mediated acylation of the axial alcohol at C2) gave the tetrasaccharide 24 (44% yield, 5 steps). Once again, using orthoester chemistry (CH₃C(OCH₃)₃/TsOH; AcOH/H₂O) on 25 allowed for the installation of a C2 acetyl group affording the desired C2/C4 diacetate 26 in good yield (93%).

We then turned our attention to the synthesis of the final required protected cleistetroside (27) from the remaining pivotal intermediate 20 (Scheme 7). As before, a selective acetylation of the C2 hydroxyl group of diol 20 using orthoester chemistry followed by sequential Pd-catalyzed glycosylation, NaBH₄ reduction, and Upjohn dihydroxylation generated triol 27 (58%, 4 steps).

Finally a nearly identical deprotection protocol that was used for the preparation of cleistriosides-5 and -6 (Schemes 4 and 5) could be used for the desired cleistetrosides (Scheme 8). Thus, exposure of all the protected tetrasaccharides 22–27 to the typical alkylative cyclization deprotection procedure ((NH₂)₂CS/NaHCO₃/Bu₄NI) cleanly removed all the chloroacetyl groups. Similarly, exposure of the products to 80% aqueous AcOH completed the total synthesis of all desired cleistetrosides in excellent yields. All the synthetic cleistriosides (1 and 2) and cleistetrosides (3–8) material had physical and spectral data that matched that reported in the literature (see: supporting information).

Scheme 8.

Completion of the total synthesis of cleistetrosides

In conclusion, a short and enantioselective total synthesis of two cleistriosides and six cleistetrosides has been achieved. The synthetic approach is amenable to the installation of all the sugar carbons for these biologically active natural products from achiral acetyl furan (10). The sugar absolute stereochemistry was installed by means of a highly enantioselective Noyori reduction and the anomeric stereochemistry by a Pd-catalyzed glycosylation. The remaining rhamno-stereochemistry was installed by highly diastereoselective post-glycosylation ketone reduction and alkene dihydroxylation. The synthesis also features an organotin mediated regioselective glycosylation achieving the desired reversal diol glycosylation regioselectivity. All the functionalities on the sugar moiety have been established via corresponding post-glycosylation transformations with only two protecting groups used. The route to the tetrasaccharide is comparable in terms of total steps to the previous approaches and is more amenable to the synthesis of all members of this family of oligosaccharide natural products. The preparation and further biological investigation of other analogues are ongoing and will be reported in due course.