Synthesis of Natural O-Linked Carba-Disaccharides, (+)- and (−)-Pericosine E, and Their Analogues as α-Glucosidase Inhibitors

Pericosine E (6), a metabolite of Periconia byssoides OUPS-N133 was originally isolated from the sea hare Aplysia kurodai, which exists as an enantiomeric mixture in nature. The enantiospecific syntheses of both enantiomers of Periconia byssoides OUPS-N133 has been achieved, along with six stereoisomers, using a common simple synthetic strategy. For these efficient syntheses, highly regio- and steroselective processes for the preparation of bromohydrin and anti-epoxide intermediates were applied. In order to access the unique O-linked carbadisaccharide structure, coupling of chlorohydrin as a donor and anti-epoxide as an acceptor was achieved using catalytic BF3·Et2O. Most of the synthesized compounds exhibited selectively significant inhibitory activity against α-glycosidase derived from yeast. The strongest analog showed almost 50 times the activity of the positive control, deoxynojirimycin.


Introduction
As the first WHO Global report says "422 million adults were living with diabetes in 2014" and "diabetes caused 1.5 million death in 2012", conquering diabetes or obesity is one of the most serious problems facing the humankind [1]. Discovering new potent α-glucosidase inhibitors is one way that pharmacists can contribute to resolving this problem. Successful and well-known examples include the clinically used anti-diabetes drugs acarbose, voglibose, and miglitol. Acarbose and voglibose molecules contain carbasugar moieties, whereas miglitol has an azasugar structure. Most candidates for glycosidase inhibitors are nitrogen-containing molecules, such as azasugars, bicyclic molecules with a nitrogen atom at the juncture, and N-linked pseudo-oligosaccharides [2]. Recent studies on thiosugar-containing α-glucosidase inhibitors have also made progress [3]. However, there have been only a few reports on pseudo-oligosaccharides constructed of only carbasugars [4][5][6][7][8]. Shing and Hudlicky independently synthesized such molecules, providing a new class of unique glycosidase inhibitors [4][5][6][7]. As this background shows, the synthetic study of carba-oligosaccharides as potential glycosidase inhibitors is a challenging step into a new research area for diabetes drugs discovery.
Among the pericosine family, pericosine E (6) is extremely unique, containing an O-linked carbadisaccharide structure between the pericosine-A-like moiety and pericosine-B-like moiety with the opposite absolute configurations ( Figure 1). Indeed, (−)-pericosine E (6) has the structure of (−)-pericosine A and that of (+)-pericosine B linked together. To our knowledge, pericosine E (6) is the only example of a natural O-linked carbadisaccharide to date [29]. Furthermore, natural pericosine E (6) was reported to exist as an enantiomeric mixture [17]. As the stereochemistry is highly complicated, the synthesis of 6 and its analogs presents an exciting challenge toward molecules with biological activity, such as glycosidase inhibitory activities. A part of this work, the first total synthesis of (−)-6, was recently published in preliminary form [14]. The synthetic strategy and newly developed technologies in the previous paper could be applied to the as yet unsynthesized natural minor enantiomer (+)-pericosine E, along with its analogs. Herein, we describe the enantiospecific syntheses of both enantiomers and six diastereomers of 6, and their glycosidase inhibitory activities.

Retrosynthetic Strategy
Based on our previous synthetic route for pericosines A-C, we envisioned the retrosynthesis of (−)-6 as summarized in Scheme 1 [13,14]. In this strategy, we illustrated the absolute configuration of 6 as being made up of (−)-pericosine A (1) and (+)-pericosine B (2). This structural pattern is

Retrosynthetic Strategy
Based on our previous synthetic route for pericosines A-C, we envisioned the retrosynthesis of (−)-6 as summarized in Scheme 1 [13,14]. In this strategy, we illustrated the absolute configuration of 6 as being made up of (−)-pericosine A (1) and (+)-pericosine B (2). This structural pattern is denoted as (−pA, +pB)-type hereafter for better understanding of the puzzling stereochemistry in pericosine E and its analogs. The former corresponds with donor chlorohydrin 9, while the latter corresponds with acceptor anti-epoxide 10, in the key condensation reaction. As mentioned above, donor 9 could be derived from syn-epoxide 11, which corresponds with a pericoxide proven to be the precursor of pericosine A in the culture of Tolypocladium sp. [27]. Therefore, our strategy might be biomimetic. Since both enantiomers of common intermediates of unstable diene (13) are available from commercially available (−)-quinic acid or (−)-shikimic acid [14,15,30], the synthesis of (+)-6, (+pA, −pB)-type, was also possible using essentially the same approach. Furthermore, coupling chlorohydrin (−)-9 with anti-epoxide (−)-10 will lead a new analog, (−pA, −pB)-type, which might correspond to undiscovered natural products.

Preparation of Both Enantiomers of Cholohydrin and Anti-Epoxide
In order to achieve these total syntheses effectively, two innovations were required: the regioand steroselective bromohydrination of unstable diene 13 to bromohydrin 12, and the epoxidation of 13 to give 10. The former reaction was carried out with N-bromosuccinimide (NBS) in an acetonitrile-H2O (3:2) solvent system using a 5 mg/mL substrate concentration, while the latter was realized by the addition of 13 to methyl (1,1,1-trifluoromethyl)dioxirane, TFDO) [31,32], prepared in situ at −15 °C in H2O-1,1,1-trifluoroacetone (1:1), affording 10 exclusively. When TFDO was generated and reacted with 13 at 0 °C, the product ratio of 10 and its regioisomer, which was inseparable from 10, was ca. 15:1. Gradual and careful addition of Oxone ® at −15 °C to a H2Otrifluoroacetone solvent system for in situ generation of TFDO was required in this process, otherwise the inseparable regioisomer was present in the product [12,28,33]. Details of reaction condition optimization can be found in our previous communication [15].

Preparation of Both Enantiomers of Cholohydrin and Anti-Epoxide
In order to achieve these total syntheses effectively, two innovations were required: the regio-and steroselective bromohydrination of unstable diene 13 to bromohydrin 12, and the epoxidation of 13 to give 10. The former reaction was carried out with N-bromosuccinimide (NBS) in an acetonitrile-H 2 O (3:2) solvent system using a 5 mg/mL substrate concentration, while the latter was realized by the addition of 13 to methyl (1,1,1-trifluoromethyl)dioxirane, TFDO) [31,32], prepared in situ at −15 • C in H 2 O-1,1,1-trifluoroacetone (1:1), affording 10 exclusively. When TFDO was generated and reacted with 13 at 0 • C, the product ratio of 10 and its regioisomer, which was inseparable from 10, was ca. 15:1. Gradual and careful addition of Oxone ® at −15 • C to a H 2 O-trifluoroacetone solvent system for in situ generation of TFDO was required in this process, otherwise the inseparable regioisomer was present in the product [12,28,33]. Details of reaction condition optimization can be found in our previous communication [15].
Analog (+)-22, which corresponds to the (+pA, −pB)-type, and its epimer (+)-23 were also synthesized from (+)-9 and (+)-10 (Scheme 3a). Ether formation between (+)-9 and (+)-10 to give (+)-24, followed by Dess-Martin oxidation and NaBH 4 reduction, proceeded to afford (+)-25 in a similar fashion to the synthesis of 6. However, the final deprotection of (+)-24 or (+)-25 at room temperature did not occur, even after a prolonged reaction time. Thus, MW-aided deprotection was applied to this step (100 In order to extend this scheme to a variety of stereoisomeric analogs, such as (+pA, +pC), (+pA, −pC), (−pA, +pC), and (−pA, −pC)-type of compounds, we attempted the challenging coupling of syn,syn-epoxide (−)-11 with alcohols 15 and (+)/(−)-9 under various conditions. Unfortunately, none of the trials gave the desired O-linked carba-disaccharides derived from 11, despite the epoxide being consumed. These failures were contrary to our expectations for 11 as an acceptor molecule, as it seemed that 11 would react with alcohols more easily than 10 due to the steric demand of 11. The reason for condensations not occurring with 11 is still not known.

Evaluation of Inhibitory Activities against Glycosidases
Eight samples synthesized here were used in a biological assay against three glycosidases, namely α-glucosidase (yeast), β-glucosidase (sweet almond), and α-mannosidase (Jack bean) [36]. The results are shown in Table 1. Six carba-disaccharides of eight samples showed significant α-glucosidase inhibition, but no inhibition was observed against β-glucosidase or α-mannosidase. Naturally

Conclusions
We have achieved the first total synthesis of both enantiomers of pericosine E (6), which are metabolites of Periconia byssoides OUPS-N133 isolated from the sea hare Aplysia kurodai. The total synthesis elucidated the absolute configuration of the naturally preferred enantiomer to be (−)-6. Using this simple and efficient strategy, pericosine E and seven further stereoisomers were prepared. Almost all the synthesized chlorine-containing O-linked carbadisaccharides showed significant inhibitory activity against only α-glycosidase derived from yeast. In general, compounds containing a (−)-pericosine B-like portion as an acceptor showed better activity. The most potent compound, (−)-22, exhibited ca. 50 times the α-glycosidase inhibitory activity of DNJ, a positive control. These results suggest that O-linked carba-oligosaccharides based on pericosine E are promising seeds for a new class of diabetes drugs. Further study of the syntheses of O-linked carbadisaccharides without chlorine is ongoing. The anti-glycosidase assay will elucidate the role of the chlorine atom.
To a solution of NaBH 4 (13.

Synthesis of (−)-21
To a solution of alcohol (−)-8 (21.6 mg, 0.038 mmol) in MeOH (0.2 mL), TFA (1.8 mL) was added dropwise at 0 • C. After stirring for 3 days at RT, the reaction mixture was concentrated under vacuum to afford white crystals. The product was purified by preparative TLC (MeOH-CH 2

Synthesis of (+)-25
To a solution of (+)-24 (170 mg, 0.30 mmol) in CH 2 Cl 2 (5 mL), DMP (190 mg, 0.39 mmol) was added at 0 • C, and the mixture was stirred for 4 h at RT. The reaction was quenched by the addition of sat. aq. Na 2 S 2 SO 4 and sat. aq. NaHCO 3 (10 mL) and extracted with TBME (3 × 20 mL). The combined organic layers were washed with brine (30 mL), dried over MgSO 4 , filtered, and the solvent was removed under reduced pressure to give a crude residue (180 mg). Without purification, the residue was taken up in methanol (5 mL) and the methanol solution (1 mL) of NaBH 4 (11 mg, 0.29 mmol) was added in four portions at 0 • C. After 30 min, the reaction mixture was quenched by the addition of sat aq. NH 4 Cl (30 mL), extracted with CH 2 Cl 2 (3 × 30 mL). The organic layer was washed with brine, dried over MgSO 4 , filtered, and the solvent was removed under reduced pressure to give a crude residue, which was purified by silica gel column chromatography (Hexane-EtOAc, 3:1) to afford (+)-25 (100 mg, 59% in two steps) as an oil.

Microwave-Aided Deprotection toward (+)-22
To a methanol solution (0.2 mL) of 25 (23.7 mg, 0.042 mmol) in a microwave vial, TFA (1.8 mL) was added at 0 • C. The vial was sealed and irradiated in the MW reactor at 100 • C for 30 min. After cooling, the reaction mixture was condensed under reduced pressure to give a crude residue, which was purified by silica gel column chromatography (MeOH-CH 2  After incubation for 20 min at 37 • C, the reaction was interrupted by the addition of 0.5 M sodium carbonate (100 µL). The amount of p-nitrophenol liberated was measured colorimetrically at 400 nm (optical density at 400 nm: ODtest). The inhibition rates (%) were calculated using the formula 100 − 100 × (ODtest−ODblank)/(control ODtest−control ODblank) and the IC 50 values were obtained from the inhibition curves. Assays for β-glucosidase and α-mannosidase were carried out as outlined above using p-nitrophenyl β-D-glucopyranoside and α-D-mannopyranoside as the substrates. The IC 50 values are shown in Table 1.
Assays on β-glucosidase and α-mannnosidase inhibition of synthesized carbadisaccharides 6 and 21-23 were carried out in a similar fashion.