Influence of 3-Thio Substituents on Benzylidene-Directed Mannosylation. Isolation of a Bridged Pyridinium Ion and Effects of 3-O-Picolyl and 3-S-Picolyl Esters

Abstract

The influence on glycosyl selectivity of substituting oxygen for sulfur at the 3-position of 4,6-O-benzylidene-protected mannopyranosyl thioglycosides is reported and varies considerably according to the protecting group employed at the 3-position. The substitution of a thioether at the 3-position for the more usual 3-O-benzyl ether results in a significant loss of selectivity. The installation of a 3-S-picolinyl thioether results in a complex reaction mixture, from which a stable seven-membered bridged bicyclic pyridinium ion is isolated, while the corresponding 3-O-picolinyl ether affords a highly α-selective coupling reaction. A 3-O-picolyl ester provides excellent β-selectivity, while the analogous 3-S-picolyl thioester gives a highly α-selective reaction. The best β-selectivity is seen with a 3-deoxy-3-(2-pyridinyldisulfanyl) system. These observations are discussed in terms of the influence of the various substituents on the central glycosyl triflate – ion pair equilibrium.

Introduction

The use of the 4,6-O-benzylidene acetal as a stereo-directing protecting group for the formation of once elusive β-mannopyranosides has demonstrated broad generality in a wide range of glycosylation methods^[1] beyond the original mannosyl sulfoxide and thioglycosides.^[2] Surrogates such as the 4,6-O-polystyryl boronates and the 4,6-O-silylene acetals,^[3] and modified versions designed to afford the β-rhamnopyranosides (6-deoxymannopyranosides) by bespoke fragmentation methods, function analogously.^[1e] A consistent limitation of the method, however, has been its sensitivity to functionality at the 3-position (Figure 1) with 3-O-tert-butyldimethylsilyl, 3-deoxy, and 3-deoxy-3-fluoro substituted donors giving much reduced selectivity, while in the 3-amino-3-deoxy series only the benzylidene imine system that most closely approaches the standard benzyl ether gives useful selectivity.^[1e]

Figure 1.

The influence of the 3-position on the selectivity of 4,6-O-benzylidene-directed mannosylation.

We rationalize these substituent effects in terms of their influence on the key equilibrium between the β-selective covalent α-mannosyl triflate generated in situ on activation and the various less selective ion pairs with which it is in equilibrium (Scheme 1). The more electron-withdrawing substituents destabilize the ion pair and favor the more S_N2-like β-selective reactions. This is offset by steric and torsional considerations, which also impact the key equilibrium, with the best combination of electron-withdrawing and steric effects found with benzyl ethers.^[1e]

Scheme 1.

The key covalent triflate – ion pair equilibrium governing selectivity.

4,6-O-Benzylidene protected mannosyl donors carrying alkanoyl and aroyl esters at the 3-position display completely reversed selectivity to the corresponding 3-O-benzyl ethers and are highly α-selective,^[4] with the exception of the 3-O-(2-picolinoyl) esters, which are again highly β-selective.^[5] In this Article we explore the influence of thioether and thioester substitution at the 3-position on selectivity in 4,6-O-benzylidene-protected mannosylation reactions (Scheme 1, X= SR) with a view to subsequent desulfurization and so generation of 3-deoxy-β-mannopyranosides, which, while not themselves not known to exist in nature, we viewed as a stepping stone en route to the 3,6-dideoxy-β-mannopyranosides, or tyvelosides.^[6]

Results and Discussion

Donor Synthesis.

Synthesis of suitable donors began with 1,2;5,6-diacetone-α-D-allofuranose 1,^[7] which was triflated and subjected to displacement with a series of sodium thiolates giving the 3-thioglucofuranose derivatives 2–4 cleanly in good yield. After removal of the acetonides with Amberlyst in hot 1:1 dioxane:water, the residual thiosugar was subjected to the Shoda protocol^[8] to give the 3-thio-β-glucothiopyranosides 5–7 in moderate to good yield. Benzylidene acetal installation, giving 8–10, was followed by oxidation with the Dess-Martin periodinane and reduction with L-selectride to give the 3-thiomannopyranosides 11–13, which were finally benzylated to give the donors 14–16 (Scheme 2).

Scheme 2.

Synthesis of 4,6-O-benzylidene-protected 3-thio-D-mannopyranosyl donors 14–16.

A further series of 3-thiomannopyranosyl donors were obtained from p-tolyl 4,6-O-benzylidene-α-D-thiomannopyrano-side 17^[9] by selective 2-O-benzylation according to the Garegg protocol,^[10] as described previously for the analogous phenyl thiomannoside, to give 18.^[11] Inversion of configuration at the 3-position by the Lattrell-Dax protocol^[12] gave the desired altro-configured thioglycoside 19 in 35% yield, accompanied by 28% of the 3-deoxy-3-nitromannopyranosyl donor 20, which was characterized by ¹H and ¹³C NMR chemical shifts of 4.89 and 87.2 at the 3-position and by typical nitro group stretching absorptions at 1561 cm⁻¹ and 1368 cm⁻¹ in the IR spectrum. An improved 65% yield of the desired altropyranoside 19 was obtained by the 2-step protocol of Dess-Martin oxidation followed by reduction with L-selectride. Subsequent triflation followed by displacement with potassium thioacetate gave 33% of the thioester 21, with the yield reduced due to competing elimination to give 22 (Scheme 3).

Scheme 3.

Synthesis of the 3-thiomannopyranosyl thioester 21 and structures of the byproducts 20 and 22.

Deacetylation of 21 with sodium methoxide in methanol was then followed by in situ conversion to the picolinyl thioether, 24, the 2-pyridyldisulfanyl derivative 25, and the picolyl thioester 26 by standard means, in 57%, 81%, and 75% yield, respectively (Scheme 4). For comparison purposes we also prepared the 3-O-picolinyl ether 27, the corresponding picolinate ester 28, and the isonicotinate ester 29 by standard means from 18 as set out in Scheme 5.

Scheme 4.

Preparation of the 3-thiomannopyranosyl derivatives 24, 25 and 26.

Scheme 5.

Preparation of the mannopyranosyl derivatives 27, 28 and 29.

Glycosylation.

With a series of donors in hand we turned to glycosylation, selecting the widely used benzenesulfinyl piperidine (BSP)/triflic anhydride preactivation protocol^[13] for activation and the sterically hindered 2,4,6-tri-tert-butylpyrimidine (TTBP)^[14] as base to buffer the reaction medium. A concern at the outset was the potential reaction of typical thiophilic promoters of glycosylation reactions with the thioether, disulfide, and thioester functionality at the 3-position, especially with glycosyl disulfides having been previously reported as glycosyl donors^[15]. We anticipated, however, that the thiophilic species generated on activation of BSP by triflic anhydride would potentially react with the sulfur functionality at the 3-position to form sulfonium ion-like species, but that any such reaction would be reversible enabling eventual activation of the thioglycosides in the desired manner. Readily available^[16] methyl 2,3-O-isopropylidene-β-D-ribofuranoside 30 was selected as a model primary alcohol for coupling to all donors, while methyl 2,3-O-isopropylidene-α-L-rhamnopyranoside 31 and 1,2;4,5-di-O-isopropylidene-α-D-glucofuranose 32 were employed as model secondary alcohols for coupling to a more limited range of donors. Glycosylations were conducted at −60°C by addition of 2.0 equiv. of triflic anhydride to a 0.06 M solution of the donor and 1.2 equiv. of BSP and 1.0 equiv. of TTBP, followed after 0.5 h by addition of 2.0 equivalents of acceptor in the form of a 1 M solution, giving a final concentration of 0.05 M. Glycosylations were monitored by TLC and mass spectrometry, and when no further evolution was observed were quenched at −60°C by addition of triethylamine, warming to room temperature, and extraction. The anomeric ratio was determined in each case by integration of the ¹H NMR spectrum of the crude reaction mixture, and the pure products isolated by column chromatography over silica gel in the standard manner leading to the results presented in Table 1. Following the established pattern in the 4,6-O-benzylidene protected mannopyranosides, the β-isomers are immediately discernible by the chemical shift of their H5 resonance, which displays an apparent dt at ~3.4 ppm; in the α-anomers the corresponding resonance is shifted downfield to ~3.8 ppm. Confirmatory evidence is provided in each case by the anomeric ¹J_C,H heteronuclear coupling constants of 155–165 Hz for the β-anomers and 170–175 Hz for the α-anomers.^[17] The initial concern about the stability of the thioether functionality in the presence of other typical thiophilic activators was borne out when attempted coupling of donor 15 with acceptor 30 in the presence of the widely applied N-iodosuccinimide/trifluoromethanesulfonic acid combination resulted in a complex mixture from which neither recovered donor nor any glycoside could be isolated.

Table 1.

Glycosylation reactions.

Entry	Donor, X	Acceptor	Product, [%] yield, β:α ratio
1	14, p-Tol-S-	30
2	14, p-Tol-S-	31
3	14, p-Tol-S-	32
4	15, Bn-S-	30
6	16, p-MeOBn-S-	30	Complex mixture
7	24, o-NC5H4-CH2-S-	30
8	25, o-C5H4N-SS-	30
9	26, picolyl-S-	30
10	26, picolyl-S-	31
11	27, o-NC5H4-CH2-O-	30
12	28, picolyl-O-	30
13	28, picolyl-O-	31
14	29, isonicotinyl-O-	30

The selectivities observed with the 3-p-tolylthio donor 14 and the 3-benzylthio donor 15 were disappointingly low (Table 1, entries 1–5), clearly indicating that the presence of a substituent at the 3-position of comparable steric bulk to a benzyl ether or benzylidene imine is insufficient for meaningful β-selectivity. The moderate to good overall yields in these reactions, however, revealed the compatibility of the BSP/Tf₂O activating system with the presence of thioethers to be less than perfect, with unidentified byproducts being formed in numerous cases. The complex reaction mixture observed with the 3-deoxy-3-p-methyoxybenzyl thioether-based donor 16 (Table 1, entry 6) revealed the limit of this compatibility.

The disappointing results observed with donors 14 and 15 (Table 1, entries 1–5) led us to refocus our efforts and combine the notion of a 3-thio-based donor with Demchenko’s hydrogen bond assisted glycosylation protocol, in which mannopyranosyl and rhamnopyranosyl donors carrying either picolinyl ethers or picolyl esters and related groups at the 3-position have been shown to be very effective for the preparation of the corresponding β-glycosides.^[5a,b,18] In the event, the picolinyl thioether 24 afforded a complex crude reaction mixture whose ¹H NMR spectrum showed a substantial amount of an apparent pyridinium ion whose presence was apparent from a downfield anomeric doublet at δ 7.05 ppm (Figure 2). Chromatography over silica gel of this complex reaction mixture eventually afforded a pure sample of the α-glycoside 37 and of the bicyclic pyridinium triflate 38 in 20 and 10% yield, respectively (Table 1, entry 7). The isolated pyridinium triflate was identified by a number of features in its NMR spectra (Figure 2).

Figure 2.

Structure and diagnostic spectral features of bridged bicyclic pyridinium ion 38.

On coupling with the model primary acceptor 30 the 3-(2-pyridyldisulfanyl)-functionalized donor 25 gave a cleaner reaction mixture than the thioether 24 and was more β-selective (Table 1, entry 8). The 3-picolyl thioester-based donor 26, however, exhibited complete α-selectivity in its coupling reactions with the primary and a secondary glycosyl acceptor (Table 1, entries 9 and 10), paralleling earlier work in our laboratories with 4,6-O-benzylidene protected mannosyl donors carrying 3-O-benzoyl and 3-O-chloroacetyl derived donors,^[4] and more recent work from the Yang and Demchenko laboratories with other rhamnopyranosyl and mannopyranosyl donors bearing esters at the 3-position.^[5c,18] The modest β-selectivity observed with donors 24 and 25, and the α-selectivity manifested by donor 26 (Table 1, entries 7–10) led us to prepare and evaluate as donors the 3-O-picolinyl ether 27 and the corresponding ester 28. Surprisingly, the 3-O-picolinyl ether 27 afforded the α-glycoside 43 with high selectivity (Table 1, entry 11), while the 3-O-picolyl ester 28 reverted to the pattern familiar from the work of Demchenko and gave high β-selectivity with a primary acceptor and more modest β-selectivity with a secondary acceptor (Table 1, entries 12 and 13). Finally, to determine the influence of the position of the heterocyclic nitrogen in the picolyl ester 28 on selectivity we prepared the 3-O-isonicotinyl analog 29 and subjected it to coupling with the standard primary acceptor 30, leading to the observation of an α-selective reaction (Table 1, entry 14).

Desulfurization.

Albeit the 3-thio substitution was found to be inadequate in the β-mannosylation reaction, we briefly investigated desulfurization leading to the formation of 3-deoxymannopyransides. This was achieved with concomitant removal of all benzyl ethers and benzylidene acetals in good yield by stirring with (30 g/mmol) Raney nickel in (6:1) methanol:dichloromethane at 65°C for 20–24 h, followed by filtration, concentration and chromatographic purification (Table 2).

Table 2.

Desulfurization reactions.

Entry	Substrate, R	Product, [%] yield
1	33, Tol	46, 68%
2	36, Bn	46, 76%
3	39, o-C5H4N-SS-	46, 72%

Hydrogen-Bonding Capabilities of Pyridine Derivatives.

To probe the relative influence of sulfur and oxygen on the hydrogen bond accepting capabilities of picolinyl ethers and thioethers, as in donors 24 and 27, and the corresponding picolyl thioester and ester 26 and 28, we synthesized the model compounds 47–50 (Figure 3) and adapted the NMR-based method of Franz and coworkers.^[19] Thus, equimolar solutions of 47–50 in CD₂Cl₂ were mixed with 0.5 molar equivalents of a stock solution of pentafluorobenzoic acid and the ¹⁹F NMR spectra recorded at room temperature. The changes (Δδ) in the ¹⁹F chemical shifts in the various mixtures relative to pentafluorobenzoic acid at the same final concentration serves as a measure of the relative H-bond accepting abilities of the four acceptors (Table 3).

Figure 3.

Model compounds for study of relative H-bonding ability of picolinyl thiothers and ethers, and picolyl thioesters and esters.

Table 3.

¹⁹F Chemical shift changes (Δδ) for the o-, m- and p-fluorines of pentafluorobenzoic acid in CD₂Cl₂ on admixture with two equivalents of H-bond acceptor.

H-Bond Acceptor	Δδ o-F [ppm]	Δδ m-F [ppm]	Δδ p-F [ppm]
Thioether 47	3.55	1.46	5.96
Ether 48	3.53	1.44	5.85
Thioester 49	0.33	0.15	0.56
Ester 50	1.89	0.83	3.14

It is apparent from inspection of Table 3 that thioether 47 and ether 48 have very comparable H-bond accepting abilities and that both are better H-bond acceptors than ester 50 and especially thioester 49. This observation is consistent with the greater electron-withdrawing ability of thioester and ester moieties than of substituted methylene groups and indicates that the better H-bond assisted directing ability of picolyl esters sometimes observed in glycosylation reactions compared to picolinyl ethers^[5a] is not related to the relative H-bond accepting ability of the two functional groups. In other words, differences in the extent of hydrogen bonding to the heterocyclic nitrogen alone cannot explain the differences in selectivity between the picolinyl ether 27 and the picolyl ester 28 (Table 1, entries 11 and 12). Thioester 49 is a weaker hydrogen bond acceptor than ester 50, consistent with the greater electron-withdrawing ability of the thiocarboxyl group than the carboxyl group that is most readily apparent from the approximately 2 pKa units greater acidity of the α-hydrogens of thioesters and thiolactones than of the corresponding carboxylate esters and lactones.^[20]

Influence of Substituents on Selectivity.

The poor selectivity of the thioethers 14 and 15 (Table 1, entries 1–5) is readily explained by the lower electronegativity of sulfur compared to oxygen, combined with the moderately greater steric bulk of arylthiol ethers and alkylthio ethers relative to benzyloxy groups.^[21] These changes negatively impact the key equilibrium of Scheme 1 and disfavor the S_N2-like mechanism on the covalent α-mannosyl triflate.

The isolation and characterization of the seven-membered bridged bicyclic pyridinium ion 38 from the 3-picolinyl thioether 24 (Table 1, entry 7) suggests that stereo-directing participation plays at least a partial competing role in at least some of these reactions. Stable, isolable glycosyl pyridinium ions resulting from the intermolecular reaction of pyridine and its congeners with various glycosyl donors have a long history dating to the early work of Lemieux and coworkers,^[22] and feature prominently in the highly α-selective 1,10-phenanthro-line-catalyzed glycosylations described by the Nguyen laboratory as in ion 51 (Figure 4).^[23] Six-membered fused bicyclic α-and β-pyridinium ions 52 and 53 (Figure 4) were described by Demchenko and coworkers. The α-isomer 52 was found to afford highly β-selective glycosylation reactions, albeit with prolonged reaction times owing to their stability, while the β-isomer 53 was found to be inert and could be recovered unchanged from the reaction mixture.^[24] However, to our knowledge, ion 38 is the first such bridged bicyclic pyridinium ion to be isolated and characterized. While the formation of such an ion should lead to α-selectivity, its very stability, as evidenced by its elution from silica gel with a DCM:MeOH solvent mixture, precludes its use as an intermediate in rapid, efficient glycosylation. With the analogous and highly α-selective 3-O-picolinyl ether 27 (Table 1, entry 11) no evidence was found for the existence of a similarly stable bridged bicyclic pyridinium ion in the crude reaction mixture. Such an oxabicyclic ion 54 would be expected to be less stable than the isolated thiabicyclic ion 38 because i) the electronegative oxygen atom in the ether bridge destabilizes the positive charge more than the sulfur atom in the comparable thioether bridge, and ii) the presence of two C – O bonds rather than the longer C – S bonds affords a more strained system. Because of its lower stability the putative oxabicyclic ion 54 is more readily opened by the acceptor alcohol, resulting in a higher yield of the α-glycoside.

Figure 4.

Literature pyridinium ions 51–53 and putative bridged bicyclic ion 54 responsible for the α-selectivity of donor 27.

The seemingly minor change of replacing a methylene group in thioether 24 by a divalent sulfur atom in the disulfide 25 has a significant impact on selectivity as the disulfide 25 is the most β-selective of the 3-thio-substitutions studied (Table 1). We rationalize the increased selectivity of disulfide 25 with respect to the nominally isosteric picolinyl thioether 24 by invoking the increased electron-withdrawing ability of disulfides compared to sulfides, with hydrodisulfides being more acidic than comparable thiols.^[25] In other words, for the purposes of β-mannosylation, the electronegativity and steric bulk of the disulfide group is a closer approximation to benzyl ether than any of the other substituents studied here including the simple benzyl thioether. Stereo-directing participation by the pyridyl group is clearly not a major factor in the coupling reactions of the disulfide 25, presumably because of a combination of the electron-withdrawing nature of the disulfanyl group and its ground state conformation with a 90° torsion angle about the disulfide bond.^[26]

The contrast in selectivity between the highly α-selective 3-(2-picolyl) thioester-functionalized donor 26 and the β-selective 3-picolyl-based donor 28 is remarkable. The latter, at least with the primary acceptor 30 (Table 1, entry 12), is highly β-selective and follows the pattern established by the Demchenko laboratory and attributed by them to stereo-directing donor-acceptor hydrogen bonding.^[5] The thioester 26 on the other hand (Table 1, entry 9) reverts to the behavior observed with simple 3-O-acyl protected mannopyranosyl donors,^[27] as first observed in the 4,6-O-benzylidene series.^[4] The α-selectivity observed with the isonicotinyl ester 29 (Table 1, entry 14) is also consistent with the precedent for the 3-O-acyl mannosyl donors, and supports the hypothesis of donor-acceptor hydrogen bonding in the reactions of the picolyl ester 28. As demonstrated with the model compounds 49 and 50 (Table 3) a picolyl thioester is a weaker hydrogen bond acceptor than a picolyl ester, from which we conclude that Demchenko-like β-directed hydrogen bond assisted glycosylation is less likely with the thioester than with the ester. Thus, weaker hydrogen bonding coupled with the selectivity-minimizing 3-thio substitution pattern on the pyranose system can be advanced to explain the loss of β-selectivity on going from the picolyl ester to the picolyl thioester. These differences are however not sufficient to explain the very high α-selectivity observed with the picolyl thioester 26, which matches that seen with other simple esters at the 3-position. Based on the use of a 3-O-(tert-butyloxycarbonyl) ester as a mechanistic probe 55 (Figure 5), we have argued that stereo-directing participation through a bridged six-membered cyclic dioxacarbenium ion is not the origin of the extreme α-selectivity seen with the 3-O-acyl mannopyranosyl donors,^[28] favoring instead donor-acceptor hydrogen bonding involving the carbonyl group of the 3-O-ester in its ground state conformation (Scheme 6),^[29] as also proposed on computational grounds by Boons and coworkers for the analogous effect in the glucopyranosyl series.^[30] In this hypothesis the covalent α-triflate 56 and the acceptor alcohol are in equilibrium with a hydrogen bonded complex between the corresponding β-triflate and the alcohol 57 and or the corresponding contact ion pair 58, setting the stage for delivery of the alcohol to the α-face and the formation of the product 59. Others have argued strongly in favor of stereo-directing participation by the 3-O-esters,^[27,29,31] and Boltje and coworkers have provided NMR spectroscopic evidence for the formation of such a bridged ion 60 on activation of a 2,4,6-tri-O-methylmannopyranosyl donor, albeit in a strongly unfavorable equilibrium with a mannopyranosyl triflate 61 (Figure 6).^[32] However, the ¹C₄ inverted chair conformation assigned to the pyranose ring in the bridged bicyclic ion 60 (Figure 6) is clearly even less favored in the presence of the 4,6-O-benzylidene protecting group employed here and in the earlier literature examples of high α-selectivity in the mannopyranose series.^[4] Consequently, we favor the hypothesis of hydrogen bond-directed acceptor delivery to the α-face that is further facilitated in the thioester case by the reduced electron-drawing effect in the 3-thio series and the consequent greater population of the ion pairs (Scheme 6).

Figure 5.

Boc-based probe 55 for remote participation.

Scheme 6.

Hydrogen-bond directed hypothesis for the formation of α-glycosides from 3-O-acyl (X=O) and 3-S-acyl (X=S) mannopyranosyl donors.

Figure 6.

Bridged bicyclic dioxacarbenium ion 60 in equilibrium with a glycosyl triflate 61.

Conclusions

The 3-thio modification clearly destroys the β-selectivity of 4,6-O-benzylidene-directed mannopyranosyl donors in their couplings with receptor alcohols, whatever the nature of the 3-S-substituent, under the conditions typically employed in our laboratory. In this it joins the 3-deoxy, 3-deoxy-3-fluoro, and the 3-azido-3-deoxy, and 3-O-(tert-butyldimethylsilyl) modifications. The requirement for successful 4,6-O-benzylidene-directed O-mannopyranosylation under these conditions therefore remains the presence of an electron-withdrawing substituent at the 3-position of intermediate steric bulk, which is best met by a benzyl ether or by a benzylidene imine. High α-selectivity is observed with a 3-O-picolinyl ether, which is attributed to participation via a transient bridged bicyclic pyridinium ion: in the analogous 3-S-picolinyl thioether donor, the bridged bicyclic ion is sufficiently stable to survive chromatographic purification over silica gel, enabling full characterization of such an ion for the first time. A 3-S-picolyl thioester affords high α-selectivity in contrast to the corresponding highly β-selective 3-O-picolyl ester, which is attributed to the reduced hydrogen bond accepting capability of the aromatic base in the thioester.

Experimental Section

General Procedure A: General Procedure for Thioacetate Deprotection.

A solution of 21 (270 mg, 0.52 mmol) in 2.7 mL of dry MeOH was stirred with 3 eq. (0.78 mL) of NaOMe (2 M solution in MeOH) at room temperature for 0.5 h and then was poured into NH₄Cl solution. EtOAc was then added, and the layers were separated, the organic layer was then washed with water and brine. The organic layer was then dried over Na₂SO₄, filtered, and concentrated down, giving a light-yellow to amber colored foam. The crude thiol was then used in the proceeding reactions without further purification.

General Procedure B: General Procedure for Glycosylations Using BSP.

A 0.06 M solution of donor in dichloromethane was stirred at room temperature for 1 h with TTBP (1 eq.), BSP (1.2 eq.), and 4 Å molecular sieves (100 mg/mL). The solution was then cooled to −60°C, triflic anhydride was then added (1.5 M in DCM). After 0.5 h, acceptor (1.0 M in DCM, 2 eq.) was added to the reaction mixture, giving a total concentration of 0.05 M with respect to the donor. After 0.5 h, the reaction mixture was quenched with TEA at −60°C and was allowed to warm slowly to room temperature. The reaction mixture was diluted with DCM and filtered through a plug of Celite. The organic layer was then washed with water and brine and dried over Na₂SO₄. The organic layer was then filtered and concentrated down. Purification using silica gel flash chromatography (EtOAc:Hex gradient) gave a combination of anomers. Separation of anomers for characterization was done by HPLC.

General Procedure C: General Procedure for Raney Nickel Reductions.

A solution of disaccharide (0.022 mmol) in 1.5 mL of DCM was added to a suspension of 660 mg of Raney nickel (30 g/mmol) in 8.5 mL of MeOH and heated to reflux with stirring under a hydrogen atmosphere for 24 h, before it was cooled to room temperature, diluted with MeOH, and filtered through Celite. The fiter cake was washed with additional MeOH and the combined filtrate was dried over Na₂SO₄, filtered through cotton, and concentrated down giving a yellow-green residue. Purification by silica gel flash chromatography (0⟶10% MeOH in DCM) gave the desired deprotected disaccharide.

Supporting Information

(see footnote on the first page of this article). Copies of ¹H, ¹³C{¹H} and 2D NMR spectra for all new compounds.

Data Availability Statement

The data that support the findings of this study are available in the supplementary material of this article.