Glycosylation with N-acetyl glycosamine donors using catalytic iron(III) triflate: from microwave batch chemistry to a scalable continuous-flow process

Abstract

Efficient and highly selective glycosylation reactions of peracetylated β-D-N-acetyl gluco- and galactosamine are described using catalytic iron(III) triflate under microwave conditions or in a continuous flow process. Simple β-glycosides and β-(1 → 6), β-(1 → 2) and β-(1 → 3) linked disaccharides bearing various protecting groups were obtained in high yields. Insights into the glycosylation mechanism are discussed.

---

Introduction

Numerous natural glycoconjugates (oligomeric structures and small molecules)^1,2 contain N-acetylated-D-glucosamine residues connected through a 1,2-trans linkage. They are implicated in important biological systems, such as structural polysaccharides (chitin), circulating signaling molecules (chitooligosaccharides³ and lipo-chitooligosaccharides^4,5), tumor markers (sialyl-Lewis X), anticoagulants (heparin), glycoproteins (multi-antennary complex type N-glycans)⁶ or as an essential part of small molecules for various bioactivities.⁷ In all cases these structures are difficult to obtain from natural sources. The main challenge of these syntheses is the glycosidic linkage formation through glycosylation, one of the most studied reactions in organic synthesis,⁸ especially in the case of N-acetyl D-glucosamine.

Under glycosylation conditions, the 2-acetamido group in sugar donors bearing various leaving groups at C-1 forms a rather stable 1,2-O,N-oxazoline, which must be opened by acceptors under appropriate conditions to give the trans glycoside.

Numerous β-selective glycosylation methods^8,9 have been developed using elaborate glucosamine donors possessing temporary participating groups^10–12 of the 2-amino function such as the well-known phthaloyl,^13,14 trichloroethoxycarbonyl,¹⁵ trichloro- and trifluoroacetyl (TCA and TFA) groups,^16,17 and more recently, the N-acetyl-2,3-oxazolidinone group.^18,19 The appropriate leaving groups at C-1 are the trichloroacetimidate,²⁰ phosphite,²¹ or thio groups.²² The reactions generally proceed at low temperature with high yields but require separate steps for the introduction of the protecting groups and the post-coupling conversion to the 2-acetamido substituent found in natural products. To date, these methods have been most commonly used for the synthesis of glycoconjugates bearing an N-acetylated-D-glucosamine residue.

Glycosylation with glycosyl acetate donors²³ is a straightforward alternative to the above methods using donors bearing complex leaving groups at the anomeric position.⁹ It involves a direct acid-catalyzed exchange of the anomeric oxygen to provide the glycosidic acetal. Recently, stoichiometric cupric salts (CuCl₂, CuBr₂),²⁴ 30 mol% Yb(OTf)₃,²⁵ 15 mol% rare earth metal triflates [Sc(OTf)₃, Sm(OTf)₃, La(OTf)₃, Dy(OTf)₃, Nd(OTf)₃],^26,27 H₂SO₄-silica under microwave conditions,²⁸ and TsOH²⁹ were used as promoters in the synthesis of glycosides of N-acetyl D-glucosamine (GlcNAc), directly or via the isolated 1,2-O,N-oxazolines. Activation using FeCl₃ was also previously described for anomeric ester donors incorporating a C-2 amide functionality (N-acetyl, N-phthaloyl, N-chloroacetyl glycosyl acetate donors)^30,31via the oxazolinium cations. It was also reported for other donors having a C-2 ester participatory group^32–34 that react via the 1,2-acyloxonium ion. It involved a large excess of both FeCl₃ and glycosyl donors producing, in the case of fluorogenic and serine acceptors, rather the α-anomer under anomerization conditions.³⁵

Mild conditions using triflates of rare earth metals were previously reported.^9,26,27 Iron^36–39 has a number of advantages over other metals typically used in catalysis since it is cheap, non-toxic, environmentally friendly and abundant. In carbohydrate chemistry,⁴⁰ iron(III) triflate has only been utilized in a few instances: oxidative C–C bond cleavage,⁴¹ thioglycosylation of peracetylated glycosides⁴² and type I Ferrier rearrangement of 2,4,6-tri-O-acetyl-D-glucal.⁴³ Over the past few years, our laboratory has developed several step-saving options that have significantly shortened the synthetic route to bioactive glycoconjugates.^44–49 Along these lines, we recently communicated⁵⁰ the glycosylation of the stable and commercially available glucosaminyl donor 1β using, as the activator, catalytic amounts of stable and non-hygroscopic Fe(OTf)₃·6.2DMSO.⁵¹ We present here a full account of this glycosylation: the catalysis design, the scope and limitations of the method, the scale-up using flow chemistry and some mechanistic elements.

Results and discussion

Catalytic system design

For optimization conditions of the glycosylation, glucose derivative 3⁵² was selected as a test sugar acceptor and the results are presented in Table 1. The glycosylation reaction of donor 1β, prepared by acetylation of 1,3,4,6-tetra-O-acetyl-2-amino-2-deoxy-β-D-glucopyranose hydrochloride,⁵³ was very slow at r.t. (entries 1 and 2, Table 1, 12–38%) and required heating under refluxing conditions for 84 hours in CH₂Cl₂ to furnish 4 in good yields with both Fe(OTf)₃·6.2DMSO and Fe(OTf)₃ (entries 5 and 6, Table 1, 86–87%). The same range of yields was also obtained using microwave irradiation at 110 °C for 45 min (entries 7 and 8, Table 1, 89–93%). At r.t., the catalyst Fe(OTf)₃·6.2DMSO was less efficient than Fe(OTf)₃ (entries 1 and 2, Table 1) and addition of 2,4,6-tri-tert-butylpyrimidine (TTBP) blocked the reaction (entries 3 and 4, Table 1). This did not occur under microwave irradiation at 110 °C (entries 7 and 8, Table 1). In previous experiments,⁵⁰ we established that glycosylation of oxazoline 2 produced, similarly, the β-1,6 disaccharide using the Fe(OTf)₃ solvate (15 mol%) under microwave irradiation and that no reaction occurred with the more stable donor 1α (Fig. 1).

Fig. 1 Glucosaminyl donor 1α and oxazoline 2.

Table 1 Optimization of iron triflate-catalyzed glycosylation using donor 1β and acceptor 3 with Fe(OTf)₃·6.2DMSO and Fe(OTf)₃

Entry	Catalyst (15 mol%)	TTBP^a	Temperature, time	Yield^b (product 4)
a TTBP = 2,4,6-tri-tert-butylpyrimidine. b Yield after silica gel chromatography. c Microwave irradiation (Anton Paar device). d No reaction.
1	Fe(OTf)3·6.2DMSO	—	r.t., 96 h	12%
2	Fe(OTf)3	—	r.t., 96 h	38%
3	Fe(OTf)3·6.2DMSO	2 equiv.	r.t., 96 h	nr^d
4	Fe(OTf)3	2 equiv.	r.t., 96 h	nr^d
5	Fe(OTf)3·6.2DMSO	—	Reflux, 84 h	87%
6	Fe(OTf)3	—	Reflux, 84 h	86%
7	Fe(OTf)3·6.2DMSO	2 equiv.	110 °C^c, 45 min	89%
8	Fe(OTf)3	2 equiv.	110 °C^c, 45 min	93%

---

Table 2 shows that Fe(OTf)₃·6.2DMSO or Fe(OTf)₃ was superior to other Fe(III) salts (FeI₃, FeCl₃, Fe(NTf₂)₃·6.2DMSO) (entries 1, 4 vs. 3, 5, 13, Table 2, 90–92% vs. 31–59%), Sc(OTf)₃ (entry 2, Table 2, 62% in our hands),²⁶ and acidic conditions (TfOH) (entries 6 and 7, Table 2, 47–70%). The addition of TTBP (2 equiv.) (entries 7 and 8, Table 1 and entry 8, Table 2, 89–98%) optimized the procedure. Using another base such as 2,6-lutidine with Fe(OTf)₃·6.2DMSO was inefficient to carry out the transformation. It is interesting to note that in dichloromethane, the Fe(OTf)₃ solvate was not soluble at the onset of the reaction while the complex became soluble in the final medium. The dissolving of the Fe(III) salts occurred in acetonitrile but the yield of glycosylation decreased (entries 9 vs. 8, Table 2, 43 vs. 89–98%). The use of a mixture of CH₂Cl₂–CH₃CN (7 : 3) or CHCl₃–CH₃CN (7 : 3) provided a soluble mixture all along the reaction course with only a slight decrease in the glycosylation yield (entries 10, 11 vs. 8, Table 2, 76–80 vs. 89–98%). This enabled the development of the reaction using a millifluidic device (see below). Interestingly, under our optimized conditions, Bi(OTf)₃, an alternative cheap, non-toxic, environmentally friendly and abundant metal complex^54–56 already described for the glycosylation of sialyl acetates,⁵⁷ proved to be as effective as Fe(OTf)₃·6.2DMSO (entries 12 vs. 8, Table 2, 88 vs. 89–98%).

Table 2 Comparison of the iron triflate-catalyzed glycosylation of glucosaminyl donor 1β with other catalytic systems under microwave irradiation

Entry	Catalyst (mol%)	TTBP (equiv.)	Solvent	Time	Yield %^c
a Under microwave irradiation at 80 °C (CEM Discover). b Under microwave irradiation at 110 °C (Anton Paar device). c Yield after silica gel chromatography. d The yield varies with the device used (CEM device 80 °C or Anton Paar device 110 °C).
1	Fe(OTf)3·6.2DMSO^a (15)	—	CH2Cl2	30 min	92%
2	Sc(OTf)3^a (15)	—	CH2Cl2	180 min	62% (75%)^26
3	Fe/I2^a (20)	—	CH2Cl2	60 min	59%
4	Fe(OTf)3^a (15)	—	CH2Cl2	30 min	90%
5	FeCl3^a (15)	—	CH2Cl2	30 min	31%
6	TfOH^a 0.45 equiv.	—	CH2Cl2	30 min	47%
7	TfOH^a 0.45 equiv.	2	CH2Cl2	30 min	70%
8	Fe(OTf)3·6.2DMSO^a^or^b (15)	2	CH2Cl2	45 min	89–98%^d
9	Fe(OTf)3·6.2DMSO^b (15)	2	CH3CN	45 min	43%
10	Fe(OTf)3·6.2DMSO^b (15)	2	CH2Cl2–CH3CN 7 : 3	45 min	80%
11	Fe(OTf)3·6.2DMSO^b (15)	2	CHCl3–CH3CN 7 : 3	45 min	76%
12	Bi(OTf)3^a (15)	2	CH2Cl2	45 min	88%
13	Fe(NTf2)3·6.2DMSO^a (15)	—	CH2Cl2	30 min	51%

---

Using an excess of the reactive benzyl alcohol acceptor with the commercially available Fe(OTf)₃ without TTBP (entry 5, Table 3), a large amount of α-anomer 6³⁵ was produced (α/β, 3 : 7). This was also observed but to a lesser extent with Fe(OTf)₃·6.2DMSO (entry 7, Table 3, α/β 1 : 9). Proceeding with an excess of the donor (2 equiv.) and/or adding TTBP with the catalyst Fe(OTf)₃ or Fe(OTf)₃·6.2DMSO prevented this α-anomerization to occur²⁶ (entries 1–4 and 6, Table 3).

Table 3 Comparison of the Fe(OTf)₃ and Fe(OTf)₃·6.2DMSO catalysts in the glycosylation of benzyl alcohol 5

Entry	D : A (equiv.)	Catalyst (15 mol%)	TTBP (equiv.)	Yield %^b (ratio α/β)^c
a Under microwave irradiation with an Anton Paar device. b Yield after silica gel chromatography. c Ratio determined by ^1H NMR.
1	2 : 1	Fe(OTf)3·6.2DMSO^a	2	95% (<5 : 95)
2	2 : 1	Fe(OTf)3·6.2DMSO^a	—	96% (<5 : 95)
3	2 : 1	Fe(OTf)3^a	2	95% (<5 : 95)
4	2 : 1	Fe(OTf)3^a	—	89% (5 : 95)
5	1 : 2	Fe(OTf)3^a	—	79% (30 : 70)
6	1 : 2	Fe(OTf)3^a	2	97% (<5 : 95)
7	1 : 2	Fe(OTf)3·6.2DMSO^a	—	77% (10 : 90)

---

Iron triflate-catalyzed glycosylation under microwave irradiation

The scope of the β-glycosylation was evaluated with different acceptors using glycosyl donor 1β (Table 4). Our conditions led to an efficient glycosylation with highly reactive acceptors BnOH (entries 1, 3 and 6, Table 3, 95–97%) and 4-ClBnOH (entry 1, Table 4, 95%). Glycosylation of 2-chloroacetic acid provided a poor yield of the α/β-anomeric esters 10 (entry 2, 21%, α/β, 4/1, Table 4). The use of TTBP enabled the glycosylation of silylated or benzylidene acceptors (compounds 11, 13, 15 and 17) without degradation (entries 3–6, Table 4) with the recovered acceptor. For instance, the β(1 → 3) linked disaccharide 14 was obtained in 74% yield from donor 1β (entry 4, Table 4). This method could also be applied to the efficient formation of β(1 → 2) linked disaccharides 16 and 18 (entries 5 and 6, Table 4, 61–53%) with almost quantitative recovery of the acceptor. The reaction was tested in the synthesis of a β-1,4-glycosidic linkage between two D-glucopyranosyl units (donor 1β and acceptors of the glucose and glucosamine series 19, 21, 23, entries 7–10, Table 4). Very moderate yields were obtained (20–26%) (entries 7, 9 and 10, Table 4), with however a quantitative recovery of the acceptor (entry 7, Table 4). In the glucosaminyl series with a phthaloyl group at the C2 position, a 3-O-acetyl (compound 21) or 3-O-benzyl group (compound 23) (entries 9 and 10, Table 4) furnished the same amount of β-1,4-disaccharide 22 or 24 (23–25%). The N-acetyl-2,3-oxazolidinone⁵⁸ acceptor 25 or the 1,6-anhydro acceptor 27⁵⁹ (entries 11 and 12, Table 4), developed to enhance the nucleophilicity of the hydroxyl group at the C4 position, gave only traces of the glycosylation product 26 or no glycosylation. This could be due to the degradation of these acceptors or products in the reaction mixture. However, optimization by proceeding at higher concentration (0.65 M in acceptor) led to a slight increase in the yield of glycoside 20 to 37% (entry 8, Table 4).

Table 4 Scope of the acceptor for the iron triflate-catalyzed glycosylation using donor (D) 1β with 15 mol% of Fe(OTf)₃·6.2DMSO and 2 equiv. of TTBP in CH₂Cl₂ under microwave irradiation

Entry	Acceptor (A)		Product		D : A	[A] (M)	Time	Yield %^a
a Yield after silica gel chromatography. b 110 °C (Anton Paar device). c 70–80 °C (CEM device); for details see ESI. d 20 mol%, Fe(OTf)3; no TTBP. e Yield based on the recovered acceptor. f Donor was recovered as a mixture of anomers (1α/1β = 1/1; 24% combined yield). g No TTBP.
1		7		8	2 : 1	0.065	30 min	95%^b
2		9		10	2.5 : 1	0.065	45 min	21% (α/β 8 : 2)^c^,^d
3		11		12	2 : 1	0.065	45 min	76%^c
4		13		14	2 : 1	0.065	45 min	74%^b
5		15		16	2 : 1	0.065	1 h	61%^b (90%)^e
6		17		18	2 : 1	0.065	1 h	53%^b
7		19		20	2 : 1	0.065	3 h	20–26%^b^,^c^,^f (>95%)^e
8		19		20	2 : 1	0.65	3 h	37%^b (75%)^e
9		21		22	5 : 1	0.065	11 h	23%^c^,^g
10		23		24	1 : 2	0.065	3 h	25%^c^,^g
11		25		26	2 : 1	0.065	3 h	7%^c
12		27		28	2 : 1	0.065	45 min	0%^c

---

The glycosylation scope was then evaluated with various N-acetyl-D-glucosamine donors (10, 2, 29, 32) in the formation of β-1,6 and β-1,4-glycosidic linkages between two D-glucopyranosyl units using glycosyl acceptors 3, 19 and 21 (Table 5). Compared with donor 1β, the replacement of the anomeric acetate group by a chloroacetate group or the acetates at the 3,4,6-positions by benzyl groups had no significant effect on the glycosylation (entries 1–3, Table 5). However, the benzylidene donor 32 failed to give the expected β-(1 → 6) linked disaccharide (entry 4, Table 5). This result is in accordance with the stereoelectronic effect of the 4,6-O-benzylidene acetals of pyranosides stabilizing the C–O bond at the anomeric center.⁶⁰ Oxazoline 2 furnished the β-(1 → 4) linked disaccharide 20 in poor yield (13%) (entry 5, Table 5). Glycosylation with the commercially available N-acetyl D-galactosamine donor 33 gave results similar to those of the N-acetyl D-glucosamine donor 1β in the formation of β-1,6; β-1,3; β-1,2 and β-1,4-glycosidic linkages (95–26%, entries 6–10, Table 5, versus (89–98%)–(20–26%), entry 8, Table 2 and entries 4–7, Table 4) with a quantitative recovery of acceptors. Under our harsh reaction conditions (microwave irradiation at 80–110 °C), variations of the oxygen protecting group at the 1, 3, 4 and 6 positions in donor or acceptor had no effect on the disaccharide yield and the course of the reaction. With our device for microwave irradiation, the iron triflate-catalyzed glycosylation scale-up was limited to the use of a 30 mL reactor versus a 10 mL reactor. This change induced a slight decrease in the yield (77% vs. 89% for 4, and 89% vs. 95% for 6β) probably as a consequence of the impaired heat transfer.

Table 5 Scope of N-acetyl D-glycosamine donors (2 equiv.) for the iron triflate-catalyzed glycosylation under microwave irradiation

Entry	Donor	Acceptor	Product (yield %)^a
a Yield after silica gel chromatography. 15 mol% of Fe(OTf)3·6.2DMSO and TTBP (2 equiv.) in CH2Cl2. b 45–180 min, 80 °C, CEM Discover®; for details see ESI. c No reaction. d 30–180 min, 110 °C, Anton Paar Monowave 300®; for details see ESI. e Yield based on the recovered acceptor.
1	10	3	4 (73%)^b
2	29 (α/β 1/2)	3	30 (86%)^b
3	29	21	31 (14%)^b
4	32 (α/β 1/1)	3	nr^b^,^c
5	2	19	20 (13%)^b
6	33	3	34 (95%)^d
7	33	13	35 (75%)^d
8	33	15	36 (63%)^d (>95%)^e
9	33	17	37 (55%)^d
10	33	19	38 (26%)^d (>95%)^e

---

Flow chemistry

The above limitation can be overcome by transposing the reaction in flow chemistry.^61,62 It has been demonstrated that micro- or minifluidic flow devices fitted with a backpressure regulator mimic high temperatures and pressures attainable in a sealed-vessel microwave chemistry batch experiment. Flow chemistry has already been used for glycosylation with success.^63–65 The major limitation was the low solubility of donor 1β that required the use of a mixture of solvents (CH₂Cl₂–acetonitrile or CHCl₃–acetonitrile) which induced a yield decrease under microwave irradiation (76–80% vs. 89–98%, entries 10 and 11 vs. 8, Table 2). In this study a Vapourtec R4-Unit was used as a millifluidic system. This system suppressed the tendency to block and does not limit the flow capacity observed with micro reactors when preparing substantial amounts of the product.⁶⁶ The formation of disaccharides 4, 20 and benzyl glycosides 6β and 8 was studied using donor 1β and acceptor 3, 5, 7 or 19.

The use of TTBP dramatically slowed down the process and decreased the yield of the reaction (entries 1 vs. 2, Table 6, 25 vs. 62%). A slight decrease of the yield was also observed with a decrease of the loading of Fe(OTf)₃·6.2DMSO (entries 3 vs. 2, Table 6, 51 vs. 62%). The optimized temperature of the reactor was 110 °C (entries 2 vs. 4, Table 6, 62 vs. 44%) and higher temperatures increased degradation. A higher pressure (33 vs. 25 bar) associated with a longer residence time (70 vs. 45 min) and a more concentrated reaction mixture in the acceptor (0.15 M) with an excess of donor 1β gave a 78% yield of 4 with high recovery of the unreacted acceptor. This was also obtained with the commercially available Fe(OTf)₃ which provided a 75% yield of 4 (entry 8, Table 6). The same yield range was obtained for benzyl glycosides 6β (77%, entry 9, Table 6) and 8 (75%, entry 12, Table 6). An excess of benzyl alcohol 5 (2 equiv./1β) decreased the yield of 6β (62%, entry 10, Table 6) without the formation of 6α as observed under microwave heating. A residence time of only 30 min, more practical for a g-scale production, allowed to maintain an acceptable yield of 6β (73%, entry 11, Table 6) as well with the Fe(OTf)₃ catalyst (77%). Our conditions were ineffective for the formation of the β-1,4-glycosidic linkage (<10%, entry 13, Table 6).

Table 6 Optimization conditions of the iron triflate-catalyzed glycosylation in flow chemistry using an injection loop

Entry	Acceptor (equiv.)	Catalyst (mol%)	TTBP	1β: equiv.; concentration	Pressure	Temperature	Residence time	Product (yield %)^a
a Yield after silica gel chromatography. b Yield based on the recovered acceptor.
1	3 (1)	Fe(OTf)3·6.2DMSO (15)	2 equiv.	2; 0.1 M	25 bar	110 °C	45 min	4 (25%)
2	3 (1)	Fe(OTf)3·6.2DMSO (15)	—	2; 0.1 M	25 bar	110 °C	45 min	4 (62%)
3	3 (1)	Fe(OTf)3·6.2DMSO (10)	—	2; 0.1 M	25 bar	110 °C	45 min	4 (51%)
4	3 (1)	Fe(OTf)3·6.2DMSO (15)	—	2; 0.1 M	25 bar	100 °C	45 min	4 (44%)
5	3 (1)	Fe(OTf)3·6.2DMSO (15)	—	2; 0.1 M	33 bar	110 °C	45 min	4 (70%)
6	3 (1)	Fe(OTf)3·6.2DMSO (15)	—	1; 0.1 M	33 bar	110 °C	45 min	4 (45%)
7	3 (1)	Fe(OTf)3·6.2DMSO (15)	—	2; 0.15 M	33 bar	110 °C	70 min	4 (74–78%) (86%)^b
8	3 (1)	Fe(OTf)3 (15)	—	2; 0.15 M	33 bar	110 °C	70 min	4 (75%)
9	5 (1)	Fe(OTf)3·6.2DMSO (15)	—	2; 0.15 M	33 bar	110 °C	45 min	6β (77%)
10	5 (2)	Fe(OTf)3·6.2DMSO (15)	—	1; 0.15 M	33 bar	110 °C	45 min	6β (62%)
11	5 (1)	Fe(OTf)3·6.2DMSO (15)	—	2; 0.15 M	33 bar	110 °C	30 min	6β (73%)
12	7 (1)	Fe(OTf)3·6.2DMSO (15)	—	2; 0.15 M	33 bar	110 °C	45 min	8 (75%)
13	19 (1)	Fe(OTf)3·6.2DMSO (15)	—	2; 0.15 M	33 bar	110 °C	70 min	20 (<10%)

---

Extending the glycosylation reaction to a continuous flow process without further changes proceeded with good yields (75–78%) using chloroform instead of dichloromethane. Due to its high volatility, dichloromethane was not suitable with our flow chemistry device for long injection times. This procedure delivered 2 g (2.52 mmol) of disaccharide 4 with a 50 mL injected volume (Fig. 2).

Fig. 2 Flow system used for glycosylation reactions under optimized conditions.

Mechanistic studies

Starting from GlcNAc glycosyl donors for the one-step synthesis of β-D-GlcNAc glycopyranosides, oxazolinium ion C (Scheme 1) is expected to be the intermediate, justifying the high β-stereoselectivity.^1,9 However, the glycosylation results with the less nucleophilic 4-OH acceptor 19 and oxazoline 2 compared to β-acetate 1 (13% yield, entry 5, Table 5vs. 20–26% yield, entry 7, Table 4) were different. This suggested that the reaction may not proceed via this intermediate and another route to the glycoside may operate. To study this possibility, glycosylation with iron(III) triflate was examined by modulating the electronic and/or the steric properties of the N-substituent in D-glucosaminyl donors 39–47. This was done by choosing the glycosylation of primary alcohol 3 under the optimized conditions (entry 1, Table 7) as a reference glycosylation reaction. Similar to 1β, formyl amide 39 and tolyl amide 40 (entries 3 and 4, Table 7) provided the expected glycosides 48 and 49, while carbamate 41, trichloroacetamide 42, trifluoroacetamide 43, phthalimide 46 or pivaloyl amide 45 (entries 6, 7, 9, 13 and 11, Table 7) were completely ineffective or significantly less effective (chloroacetamide 44, entry 10, Table 7). In the case of the diacetamide 47, one acetyl group was transferred to the acceptor providing 58, without detecting the formation of the disaccharide (entries 14, Table 7). These negative results should be compared with 2-deoxy-2-trichloroacetamido¹⁷ and 2-deoxy-2-trifluoroacetamido^10,67,68 derivatives equipped with a good leaving group at the anomeric carbon (e.g., trichloroacetimidate). When activated with appropriate promoters (e.g., Me₃SiOTf), they are known to be good glycosyl donors through the formation of the oxazolinium ion intermediate.¹⁷ In the absence of nucleophile, oxazolines 2, 56 and 57 (entries 2, 8 and 12, Table 7) were not detected except for oxazoline 55 from the tolylamide 40 (entry 5, Table 7). These experiments suggest that glycosylation would proceed through an alternative intermediate and not necessarily through the oxazolinium ion. Glycosylation may require a pre-complexation of the catalyst by a suitable amide group such as the acetamide present in 1β (see A, Scheme 1), the tolyl amide in 40 or the formyl amide in 39 before the activation of the anomeric acetate occurs. Effective amide pre-complexation of Fe(OTf)₃·6.2DMSO may be partially or totally prevented for electronic reasons (NHTCA, donor 42; NHTFA, donor 43; NHAcCl, donor 44) or steric grounds (NHPiv, donor 45; NPhth, donor 46), thus preventing glycoside formation as experimentally observed. Alpha–ionic pair B from 1β, instead of oxazolinium ion C, would then encourage the glycosylation to take place from the β face by shielding the α face.

Scheme 1 Possible mechanism for the iron triflate-catalyzed glycosylation.

Table 7 Iron triflate-catalyzed glycosylation of 3 using 2 equiv. of donors 1β and 39–47 under microwave irradiation

Entry	Donor	Acceptor	Product (yield %)^a
a Yield after silica gel chromatography. b Reaction performed in the presence of TTBP (2 equiv.) with 15 mol% of Fe(OTf)3·6.2DMSO in CH2Cl2/CEM device 80 °C, 45–60 min. c 1β/1α ratio of 4/1. d No reaction. e Inseparable mixture with the donor, conversion determined by ^1H NMR. f Traces detected by UPLC-MS/DAD. g Reaction performed in the presence of TTBP (2 equiv.)/Anton Paar device 110 °C, 45 min.
1	1β	3	4 (98%)^b
2	1β	None	1α–1β (100%)^c
3	39 R^1 = R^2 = H	3	48 (50%)^b
4	40 R^1 = PhCH3, R^2 = H	3	49 (82%)^b
5	40 R^1 = PhCH3, R^2 = H	None	55 (62%)^b
6	41 R^1 = OCH2Ph, R^2 = H	3	nr^b^,^d
7	42 R^1 = CCl3, R^2 = H	3	50 (<5%)^b
8	42 R^1 = CCl3, R^2 = H	None	nr^b^,^d
9	43 R^1 = CF3, R^2 = H	3	nr^b^,^d
10	44 R^1 = CH2Cl, R^2 = H	3	51 (30%)^b^,^e
11	45 R^1 = C(CH3)3, R^2 = H	3	52 (<5%)^b^,^f
12	45 R^1 = C(CH3)3, R^2 = H	None	nr^b^,^d
13	46 R^1 = R^2 = Phth	3	53 (<5%)^f^,^g
14	47 R^1 = CH3, R^2 = COCH3	3	58 ^69 (40%)^b

---

It is noteworthy that the reaction scale-up in the preparation of disaccharide 4 allowed the isolation of a small amount of oxazoline 2, suggesting a partial contribution of the oxazolinium ion C in the formation of the glycoside. A possible equilibrium between B and C could be envisioned depending on the nature of the group R¹ favoring one or the other.

Conclusion

This novel catalytic glycosylation using peracetylated β-GlcNAc 1β and β-GalNAc 33 as glycosyl donors with Fe(III) triflate and TTBP is effective in the direct synthesis of β-GlcNAc and β-GalNAc glycosides but has not yet been efficient using less nucleophilic sugar acceptors. Our results suggest a possible mechanism which proceeds mostly by intermediates not involving the unique oxazolinium ion. We have demonstrated that the Fe(III) triflate glycosylation conducted under microwave irradiation is amenable to flow chemistry without requiring the presence of TTBP.

Experimental section

Typical procedure for microwave-assisted glycosylation

The donor 1β (50 mg, 0.128 mmol, 2 equiv.), TTBP (32 mg, 0.129 mmol, 2 equiv.) and Fe(OTf)₃·6.2DMSO (10 mg, 0.010 mmol, 15 mol%) were added to the acceptor 3 (30 mg, 0.065 mmol, 1 equiv.) in an oven-dried, argon-purged microwave vial equipped with a magnetic stirring bar. Everything was flushed under argon and dry CH₂Cl₂ (1 mL) was added. After sealing the vial, the reaction mixture was heated to 110 °C under microwave irradiation for 45 min (1 minute ramp time from room temperature to 110 °C and 45 min hold time at 110 °C, stirring set at 800 rpm). The reaction mixture was diluted with CH₂Cl₂ (20 mL) and washed with a saturated aqueous solution of NaHCO₃ (10 mL). The aqueous layer was extracted with CH₂Cl₂ (3 × 10 mL) and the combined organic layers were washed with brine, dried over Na₂SO₄, filtered and concentrated under reduced pressure. The crude product was purified by flash chromatography on silica gel (heptane–EtOAc 5 : 5 to 0 : 1) to afford the pure product 4 (46 mg, 89%, white amorphous solid).

Typical procedure for glycosylations under continuous flow conditions

The donor 1β (2.92 g, 7.50 mmol, 2 equiv.) and Fe(OTf)₃·6.2DMSO (556 mg, 0.56 mmol, 15 mol%) were added to the acceptor 3 (1.74 g, 3.75 mmol, 1 equiv.) in an oven-dried, argon purged vial equipped with a magnetic stirring bar. A dry mixture of chloroform–acetonitrile 7 : 3 (50 mL) was added and the reaction mixture was stirred and sonicated for a few minutes (until complete homogenisation). After setting up and drying the whole flow system with dry chloroform–acetonitrile 7 : 3, the pump was primed and the reaction mixture (contained in an argon overpressured vial) is pumped into two 10 mL stainless steel reactors in series, heated at 110 °C with a flow rate of 0.286 mL min⁻¹ (corresponding to a residence time of 70 min). The system pressure, controlled with a back pressure regulator, was fixed at 33 bars and the reaction mixture was finally collected into a single receptor. The reaction mixture was diluted with dichloromethane (250 mL) and washed with a saturated aqueous solution of NaHCO₃ (100 mL). The aqueous layer was extracted with CH₂Cl₂ (4 × 100 mL) and the combined organic layers were washed with brine (100 mL), dried over Na₂SO₄, filtered and evaporated under reduced pressure. The crude product was purified by flash chromatography on silica gel (heptane–EtOAc 5 : 5 to 0 : 1) to afford the pure product 4 (2.00 g, 78%, white amorphous solid).

Acknowledgements

The authors thank R. Beau for her comments on the manuscript. We are grateful to the Institut de Chimie des Substances Naturelles (ICSN), the Institut Universitaire de France (IUF) for the financial support of this study. The CHARM3AT Labex program is also acknowledged for its support.

Notes and references

1. P. Sinaÿ, B. Ernst and G. Hart, Carbohydrates in Chemistry and Biology, Wiley VCH, 2000.
2. A. Varki, J. D. Esko, R. D. Cummings, H. H. Freeze, P. Stanley, C. R. Bertozzi, G. W. Hart and M. E. Etzler, Essentials of Glycobiology, Cold Spring Harbour Laboratory Press, Cold Spring Harbor, NY, 2nd edn, 2009.
3. N. Shibuya and E. Minami, Physiol. Mol. Plant Pathol., 2001, 59, 223–233.
4. J. Dénarié, F. Debellé and J. C. Promé, Annu. Rev. Biochem., 1996, 65, 503–535.
5. F. Maillet, V. Poinsot, O. André, V. Puech-Pagès, A. Haouy, M. Gueunier, L. Cromer, D. Giraudet, D. Formey, A. Niebel, E. A. Martinez, H. Driguez, G. Bécard and J. Dénarié, Nature, 2011, 469, 58–63.
6. Z. Wang, Z. S. Chinoy, S. G. Ambre, W. Peng, R. McBride, R. P. de Vries, J. Glushka, J. C. Paulson and G. J. Boons, Science, 2013, 341, 379–383.
7. M. E. Jung and P. Koch, Org. Lett., 2011, 13, 3710–3713.
8. X. M. Zhu and R. R. Schmidt, Angew. Chem., Int. Ed., 2009, 48, 1900–1934.
9. A. V. Demchenko, Handbook of Chemical Glycosylation, Wiley-VCH, 2008.
10. A. F. G. Bongat and A. V. Demchenko, Carbohydr. Res., 2007, 342, 374–406.
11. R. Enugala, L. C. Carvalho, M. J. Dias Pires and M. M. Marques, Chem. – Asian J., 2012, 7, 2482–2501.
12. Y. Yang and B. Yu, Tetrahedron, 2014, 70, 1023–1046.
13. R. U. Lemieux, T. Takeda and B. Chung, Abstr. Pap. Am. Chem. Soc., 1976, 6–6.
14. J. Banoub, P. Boullanger and D. Lafont, Chem. Rev., 1992, 92, 1167–1195.
15. W. Dullenkopf, J. C. Castro-Palomino, L. Manzoni and R. R. Schmidt, Carbohydr. Res., 1996, 296, 135–147.
16. M. L. Wolfrom and H. B. Bhat, J. Org. Chem., 1967, 32, 1821–1823.
17. G. Blatter, J.-M. Beau and J.-C. Jacquinet, Carbohydr. Res., 1994, 260, 189–202.
18. Y. Geng, L.-H. Zhang and X.-S. Ye, Chem. Commun., 2008, 597–599.
19. J. D. M. Olsson, L. Eriksson, M. Lahmann and S. Oscarson, J. Org. Chem., 2008, 73, 7181–7188.
20. M. R. E. Aly, E.-S. I. Ibrahim, E. S. H. El Ashry and R. R. Schmidt, Carbohydr. Res., 2001, 331, 129–142.
21. R. Arihara, S. Nakamura and S. Hashimoto, Angew. Chem., Int. Ed., 2005, 44, 2245–2249.
22. S. Yamago, T. Yamada, T. Maruyama and J. Yoshida, Angew. Chem., Int. Ed., 2004, 43, 2145–2148.
23. B. Helferich and E. Schmitz-Hillebrecht, Ber. Dtsch. Chem. Ges., 1933, 66, 378–383.
24. V. Wittmann and D. Lennartz, Eur. J. Org. Chem., 2002, 1363–1367.
25. C. F. Crasto and G. B. Jones, Tetrahedron Lett., 2004, 45, 4891–4894.
26. H. Christensen, M. S. Christiansen, J. Petersen and H. H. Jensen, Org. Biomol. Chem., 2008, 6, 3276–3283.
27. J. Krag, M. S. Christiansen, J. G. Petersen and H. H. Jensen, Carbohydr. Res., 2010, 345, 872–879.
28. S. Mandal, N. Sharma and B. Mukhopadhyay, Synlett, 2009, 3111–3114.
29. Y. Cai, C.-C. Ling and D. R. Bundle, Org. Lett., 2005, 7, 4021–4024.
30. M. Kiso and L. Anderson, Carbohydr. Res., 1985, 136, 309–323.
31. F. Dasgupta and L. Anderson, Carbohydr. Res., 1990, 202, 239–255.
32. S. K. Chatterjee and P. Nuhn, Chem. Commun., 1998, 1729–1730.
33. S. Koto, M. Hirooka, T. Tashiro, M. Sakashita, M. Hatachi, T. Kono, M. Shimizu, N. Yoshida, S. Kurasawa, N. Sakuma, S. Sawazaki, A. Takeuchi, N. Shoya and E. Nakamura, Carbohydr. Res., 2004, 339, 2415–2424.
34. J. Seibel, L. Hillringhaus and R. Moraru, Carbohydr. Res., 2005, 340, 507–511.
35. G. H. Wei, X. Lv and Y. Du, Carbohydr. Res., 2008, 343, 3096–3099.
36. B. D. Sherry and A. Fürstner, Acc. Chem. Res., 2008, 41, 1500–1511.
37. D. D. Diaz, P. O. Miranda, J. I. Padron and V. S. Martin, Curr. Org. Chem., 2006, 10, 457–476.
38. C. Bolm, J. Legros, J. Le Paih and L. Zani, Chem. Rev., 2004, 104, 6217–6254.
39. E. B. Bauer, Curr. Org. Chem., 2008, 12, 1341–1369.
40. J.-M. Beau, Y. Bourdreux, F.-D. Boyer, S. Norsikian, D. Urban, G. Doisneau, B. Vauzeilles, A. Gouasmat, A. Lemétais, A. Mathieu, J.-F. Soulé, A. Stévenin and A. Xolin, in Carbohydrate Chemistry, The Royal Society of Chemistry, 2014, vol. 40, pp. 118–139.
41. S. Ichikawa, I. Tomita, A. Hosaka and T. Sato, Bull. Chem. Soc. Jpn., 1988, 61, 513–520.
42. S. S. Weng, Tetrahedron Lett., 2009, 50, 6414–6417.
43. P. Chen and S. Wang, Tetrahedron, 2012, 68, 5356–5362.
44. N. Grenouillat, B. Vauzeilles, J. J. Bono, E. Samain and J.-M. Beau, Angew. Chem., Int. Ed., 2004, 43, 4644–4646.
45. A. Français, D. Urban and J.-M. Beau, Angew. Chem., Int. Ed., 2007, 46, 8662–8665.
46. Y. Bourdreux, A. Lemétais, D. Urban and J.-M. Beau, Chem. Commun., 2011, 47, 2646–2648.
47. J.-F. Soulé, A. Mathieu, S. Norsikian and J.-M. Beau, Org. Lett., 2010, 12, 5322–5325.
48. S. Zameo, B. Vauzeilles and J.-M. Beau, Angew. Chem., Int. Ed., 2005, 44, 965–969.
49. A. Malapelle, Z. Abdallah, G. Doisneau and J.-M. Beau, Angew. Chem., Int. Ed., 2006, 45, 6016–6020.
50. A. Stévenin, F.-D. Boyer and J.-M. Beau, Eur. J. Org. Chem., 2012, 1699–1702.
51. S. Antoniotti and E. Duňach, Chem. Commun., 2008, 993–995.
52. T. Ishikawa, Y. Shimizu, T. Kudoh and S. Saito, Org. Lett., 2003, 5, 3879–3882.
53. H. Myszka, D. Bednarczyk, M. Najder and W. Kaca, Carbohydr. Res., 2003, 338, 133–141.
54. H. Gaspard-Iloughmane and C. Le Roux, Eur. J. Org. Chem., 2004, 2517–2532.
55. V. Mandadapu, F. Wu and A. I. Day, Org. Lett., 2014, 16, 1275–1277.
56. J. R. Desmurs, M. Labrouillère, C. Le Roux, H. Gaspard, A. Laporterie and J. Dubac, Tetrahedron Lett., 1997, 38, 8871–8874.
57. K. Ikeda, Y. Torisawa, T. Nishi, J. Minamikawa, K. Tanaka and M. Sato, Bioorg. Med. Chem., 2003, 11, 3073–3076.
58. D. Crich and A. U. Vinod, J. Org. Chem., 2005, 70, 1291–1296.
59. D. Tailler, J.-C. Jacquinet and J.-M. Beau, J. Chem. Soc., Chem. Commun., 1994, 1827–1828.
60. H. H. Jensen, L. U. Nordstrom and M. Bols, J. Am. Chem. Soc., 2004, 126, 9205–9213.
61. T. N. Glasnov and C. O. Kappe, Chem. – Eur. J., 2011, 17, 11956–11968.
62. K. S. Elvira, X. Casadevall i Solvas, R. C. Wootton and A. J. deMello, Nat. Chem., 2013, 5, 905–915.
63. D. M. Ratner, E. R. Murphy, M. Jhunjhunwala, D. A. Snyder, K. F. Jensen and P. H. Seeberger, Chem. Commun., 2005, 578–580.
64. F. R. Carrel, K. Geyer, J. D. C. Codee and P. H. Seeberger, Org. Lett., 2007, 9, 2285–2288.
65. D. T. McQuade and P. H. Seeberger, J. Org. Chem., 2013, 78, 6384–6389.
66. J. Wegner, S. Ceylan and A. Kirschning, Chem. Commun., 2011, 47, 4583–4592.
67. L. G. Weaver, Y. Singh, J. T. Blanchfield and P. L. Burn, Carbohydr. Res., 2013, 371, 68–76.
68. D. J. Silva, H. Wang, N. M. Allanson, R. K. Jain and M. J. Sofia, J. Org. Chem., 1999, 64, 5926–5929.
69. M. Giordano, A. Iadonisi and A. Pastore, Eur. J. Org. Chem., 2013, 3137–3147.

---

Footnotes

† Dedicated to Professor Max Malacria on the occasion of his 65th birthday.

‡ Electronic supplementary information (ESI) available: Preparation, ¹H and ¹³C NMR spectra for novel compounds. See DOI: 10.1039/c4qo00183d

---