Cationic Gold(I)-Catalyzed Glycosylation with Glycosyl N-1,1-Dimethylpropargyl Carbamate donors

Abstract

A mild and efficient cationic gold(I)-catalyzed O-glycosylation methodology involving the use of bench-stable glycosyl N-1,1-dimethylpropargyl carbamate donors has been developed. In the presence of 1–2 mol% [tris(2,4-di-tert-butylphenyl)phosphite]gold(I) chloride and 5 mol% silver triflate, both “armed” and “disarmed” glycosyl N-1,1-dimethylpropargyl carbamate donors react with various sugar acceptors at room temperature to afford corresponding glycosides in good to excellent yields. These glycosyl N-1,1-dimethylpropargyl carbamates are found to be orthogonal to regular phenyl thioglycoside donors. The utilization of this method has been demonstrated in the synthesis of a trisaccharide.

Carbohydrate molecules play essential roles in numerous biological and cellular processes.¹ Extensive studies of their biological functions require the availability of sufficient amounts of pure and structurally well-defined carbohydrate structures, which demands the development of efficient glycosylation methods and strategies as their isolation is usually very challenging due to the scarcity and heterogeneous nature. Chemical glycosylation reactions catalyzed by transition metal complexes² or organocatalysts³ have recently been explored and demonstrated great success in stereoselective synthesis of complex oligosaccharides and glycoconjugates. In contrast to “traditional” protocols involving stoichiometric amounts of promoters, catalytic glycosylations oftentimes contribute alternative “green”, mild, and orthogonal approaches for stereoselective construction of challenging glycosidic linkages.

Since the Hotha group reported the first gold-catalyzed glycosylation in 2006,⁴ catalytic construction of glycosidic linkages by gold complexes has been quite popular due to its selective high affinity for triple bond, high efficiency and mild conditions.⁵ For instance, the Yu group developed an efficient cationic gold(I)-catalyzed glycosylation method using glycosyl o-alkynylbenzoate donors 1 which has been utilized in the synthesis of various types of glycosides (a, Scheme 1).^5b,6,7 In addition, the Hotha group disclosed an interesting cationic gold(I)-catalyzed glycosylation involving glycosyl ethynylcyclohexyl carbonate donors 3 (b, Scheme 1).⁸ Recently, our group reported a cationic gold(I)-catalyzed glycosylation using bench-stable glycosyl S-3-butynyl thiocarbonate donors,⁹ based on our previous discovery of glycosyl S-3-butynyl thioglycosides in gold-catalyzed glycosylation.¹⁰ However, fully substituted sugarderived glycosyl donors bearing these sulfur-containing leaving groups^9,10 were found to be less reactive than those glycosyl o-alkynylbenzoates^5b or ethynylcyclohexyl carbonate donors⁸ towards cationic gold(I) catalysis.

Scheme 1.

Representative gold-catalyzed glycosylations.

Therefore, we speculated that by replacing ethynylcyclohexyl carbonate⁸ with N-1,1-dimethylpropargyl carbamate in the leaving group, the reactivity of glycosyl donors (cf. 4) toward cationic gold(I) catalysis may be improved due to the more electron-rich nature of the carbamate over the corresponding carbonate (c, Scheme 1). Hopefully, a low loading of costly gold(I) catalyst may be sufficient for activation of glycosyl N-1,1-dimethylpropargyl carbamates 4, which would potentially improve the practicality and cost-effectiveness of this new glycosylation method.

Bench-stable glycosyl N-propargyl and N-1,1-dimethylpropargyl carbamate donors can be readily prepared from sugar lactols 5 and the corresponding propargyl isocyanates. In our hands, treatment of protected sugar lactols with N-propargyl isocyanate 6 or N-1,1-dimethylpropargyl isocyanate 7 in the presence of 0.5 eq. of Cs₂CO₃ in dichloromethane from 0 ^oC to room temperature afforded the corresponding desired glycosyl N-1,1-dimethylpropargyl carbamate donors 8-13 in high yields, respectively (Table 1).

Table 1.

Synthesis of glycosyl N-propargyl and N-1,1-dimethylpropargyl carbamate donors.

With these glycosyl N-propargyl and N-1,1-dimethylpropargyl carbamate donors available, we next evaluated their reactivity towards cationic gold(I) catalysis for glycosylation. As shown in Table 2, in the presence of 1 mol% of Ph₃PAuCl and 5 mol% AgOTf,¹¹ “armed” 2,3,4,6-tetra-O-benzyl-d-glucosyl N-1,1-dimethylpropargyl carbamate donor 9 reacted with primary alcohol acceptor 14 to afford the corresponding glycosides 15 in 81% yield in 1 hour (entry 2), while d-glucosyl N-propargyl carbamate donor 8 showed low reactivity (entry 1). This was not unexpected as cyclization can be accelerated by the well-known Thorpe-Ingold effect.¹² It was found that Ph₃PAuCl or silver triflate independently were unable to catalyze this reaction. Changing the ligand from PPh₃ to more electron-deficient (4-CF3-Ph)3P and tris(2,4-di-tert-butylphenyl)phosphite in the cationic gold(I) complex slightly improved the yield to 84% (entry 3) and 88% (entry 4), respectively. When the loading of [tris(2,4-di-tert-butylphenyl)phosphite]gold(I) chloride catalyst was increased to 2 mol%, the reaction was completed in 10 minutes and the yield was improved to 98% (entry 5). Other gold catalysts including (IPr)AuCl, PicAuCl2, Ph3PAuNTf2 were found to be less effective (entries 6–8).

Table 2.

Optimization of Conditions.^a

entry	glycosyl donor	reaction conditions	product/yieldb(α/β)c
1	8	1 mol% Ph3PAuCl, 5 mol% AgOTf, 1 h	15, low conversion
2	9	1 mol% Ph3PAuCl, 5 mol% AgOTf, 1 h	15, 81% (1.1/1)
3	9	1 mol% (4-CF3Ph)3PAuCl, 5 mol% AgOTf, 1 h	15, 84% (1.1/1)
4	9	1 mol% L*AuCl, 5 mol% AgOTf, 1 h	15, 88% (1.1/1)
5	9	2 mol% L*AuCl, 5 mol% AgOTf, 10 min	15, 98% (1.1/1)
6	9	2 mol% (IPr)AuCl, 5 mol% AgOTf, 18 h	15, 79% (1.1/1)
7	9	2 mol% PicAuCl2, 5 mol% AgOTf, 18 h	15, 15% (1.1/1)
8	9	2 mol% Ph3PAuNTf2, 18 h	15, 79% (1/1.2)
9	10	1 mol% L*AuCl, 5 mol% AgOTf, 10 min	16, 99% (β only)

^aReactions were performed using 0.2 mmol of acceptor 14, 0.24 mmol of donor 8 (or 9 or 10, 1.2 equiv.), gold and/or silver catalyst in 1 mL anhydrous CH₂Cl₂ at room temperature.

^bisolated yield.

^cα/β ratio was determined by ¹H NMR analysis.

L*AuCl = [Tris(2,4-di-tert-butylphenyl)phosphite]gold(I) chloride; (IPr)AuCl = 1,3-Bis(2,6-diisopropylphenyl-imidazol-2-ylidene)gold(I) chloride; PicAuCl_{2 =} Dichloro(2-pyridinecarboxylato)gold.

Interestingly, for glycosylation of “disarmed” 2,3,4,6-tetra-O-benzoyl-d-glucosyl N-1,1-dimethylpropargyl carbamate donor 10, 1 mol% of [tris(2,4-di-tert-butylphenyl)phosphite]gold(I) chloride catalyst was found to be sufficient and glycosides 16 was obtained in 99% yield in 10 minutes (β only, entry 9). These results are exciting as 1–2 mol% gold catalyst is sufficient for glycosylation involving glycosyl N-1,1-dimethylpropargyl carbamate donors, which makes this methodology practical. In addition, a 4,4-dimethyl-5-methyleneoxazolidin-2-one 17¹³ by-product was successfully isolated and characterized.

With this optimal condition developed, we next investigated the reaction scope using glycosyl N-1,1-dimethylpropargyl carbamate donors (9–13) and acceptors (14, 18–20). In general, “armed” and “disarmed” D-glucose, D-galactose, and L-rhamnose-derived glycosyl N-1,1-dimethylpropargyl carbamate donors reacted with various primary and secondary alcohol acceptors to afford the corresponding desired glycosides (22–36) in good to excellent yields (Table 3). As previously mentioned, 2 mol% of gold catalyst was used for activation of “armed” donors (9, 11 and 12) while 1 mol% gold catalyst was sufficient for activation of “disarmed” donors (10 and 13). It was also found that glycosylation with less reactive acceptors, such as 18¹⁴ and 20, appeared to afford corresponding glycosides in slightly lower yields.

Table 3.

Cationic gold(I)-catalyzed glycosylation with glycosyl N-1,1-dimethylpropargyl carbamate donors.^a,b

^a Reactions were performed using 0.2 mmol of acceptor, 0.24 mmol of donor (1.2 equiv.), 1–2 mol% L*AuCl and 5 mol% AgOTf in 1 mL anhydrous CH₂Cl₂ at room temperature for 10 minutes. ^b isolated yield (α/β ratio was determined by ¹H NMR analysis). ^c 2 mol% L*AuCl was used. ^d 1 mol% L*AuCl was used. L*AuCl = [Tris(2,4-di-tert-butylphenyl)phosphite]gold chloride.

Based on the fact that 4,4-dimethyl-5-methyleneoxazolidin-2-one by-product 17¹³ was identified as the by-product, a plausible mechanism is proposed for this cationic gold(I)-catalyzed glycosylation. As shown in Scheme 2, activation of the alkyne functionality of N-1,1-dimethylpropargyl carbamate donor 4 by cationic gold(I) catalyst followed by 5-exo-dig attack of the carbonyl oxygen atom provides the activated species B (via intermediate A). Subsequent cleavage of the glycosidic bond of B results in the formation of oxocarbenium ion C and oxazolidinone-derived alkenylgold(I) complex D (path a). Glycosylation of C with the alcohol acceptor furnishes the desired glycosides 2 and a molecule of triflic acid. Proto-deauration of complex D leads to the formation of 4,4-dimethyl-5-methyleneoxazolidin-2-one by-product 17 and regeneration of the cationic gold(I) catalyst. Alternatively, complex B may lose a molecule of triflic acid followed by proto-deauration to produce 2-glycosyloxy oxazoline intermediate E and regenerate the cationic gold(I) catalyst.¹⁵ Coordination of AgOTf to the nitrogen atom in E may form complex F which undergoes an rearrangement via a six-membered ring transition state to afford glycosyl triflate H and a silver complex G. Glycosyl triflate H may equilibrate with oxocarbenium ion C or react with the alcohol acceptor directly to produce the desired glycosides 2 and a molecule of triflic acid. Protonolysis of complex G may lead to the formation of by-product 17 and regeneration of silver triflate.

Scheme 2.

Proposed mechanism.

This glycosylation method was next demonstrated in the synthesis of a trisaccharide. As shown in Scheme 3, cationic gold-catalyzed glycosylation of “disarmed” 2,3,4,6-tetra-O-benzoyl-D-glucosyl N-1,1-dimethylpropargyl carbamate donor 10 and phenyl 2,3,6-tri-O-benzyl-1-thio-β-D-glucoside 37¹⁶ afforded β-linked disaccharide 38 in 88% yield (β only). This result indicated that glycosyl N-1,1-dimethylpropargyl carbamate donors are orthogonal to phenyl thioglycosides. Disaccharide thioglycoside 38 was then employed as a glycosyl donor which reacted with primary alcohol acceptor 14 under traditional glycosylation protocols (NIS/TfOH) to produce the trisaccharide 39 in 87% yield and moderate anomeric selectivity.

Scheme 3.

Synthesis of a trisaccharide.

In conclusion, we have developed a mild and efficient cationic gold(I)-catalyzed glycosylation involving readily available bench stable glycosyl N-1,1-dimethylpropargyl carbamate donors. A very low loading of costly gold(I) catalyst (1–2 mol%) is generally sufficient for activation of glycosyl N-1,1-dimethylpropargyl carbamate donors. This methodology tolerates both “armed” and “disarmed” glycosyl donors and various sugar acceptors and affords the corresponding glycosides in good to excellent yields. In addition, glycosyl N-1,1-dimethylpropargyl carbamate donors are found to be orthogonal to regular phenyl thioglycoside donors under gold catalysis. The utilization of this newly developed method has been demonstrated in the synthesis of a trisaccharide. Further application of this cationic gold(I)-catalyzed glycosylation to the synthesis of other biologically relevant carbohydrate molecules is currently underway and will be reported in due course.