Tuning reactivity of glycosyl imidinium intermediate for 2-azido-2-deoxyglycosyl donors in α-glycosidic bond formation

Article abstract of DOI:10.1021/ol402519c

The chemical properties of nucleophile additives were investigated in a modulated glycosylation context. N-Formylmorpholine (NFM) was found to be an effective modulator for glycosylation with less reactive 2-azido-2- deoxythioglucosyl and thiogalactosyl d

Full text of DOI:10.1021/ol402519c

ORGANIC
LETTERS
XXXX
Vol. XX, No. XX
000–000
Tuning Reactivity of Glycosyl Imidinium
Intermediate for 2‑Azido-2-deoxyglycosyl
Donors in r‑Glycosidic Bond Formation
Arun B. Ingle, Chin-Sheng Chao, Wei-Cheng Hung, and Kwok-Kong Tony Mong*
Applied Chemistry Department, National Chiao Tung University, 1001 Ta Hsueh Road,
Taiwan 300, ROC
tmong@mail.nctu.edu.tw
Received September 2, 2013
ABSTRACT
The chemical properties of nucleophile additives were investigated in a modulated glycosylation context. N-Formylmorpholine (NFM) was found to
be an effective modulator for glycosylation with less reactive 2-azido-2-deoxythioglucosyl and thiogalactosyl donors.
Different glycosylation approaches havebeendeveloped
to achieve pure and structurally defined oligosaccharide
compounds.1À3 Oligosaccharide synthesis involves regio-
and stereospecific assembly of glycosyl building units.
Such requirements are usually fulfilled by designing gly-
cosyl building blocks with suitable reactivity profiles and
stereodirecting functional groups.4À7
The reactivity of glycosyl building blocks is generally
modified by protecting functions. For example, a glycosyl
donor with acyl-protecting functions is less reactive in gly-
cosylation than a glycosyl donor carrying benzyl ether pro-
tection.8 This electronic effect was elaborated by Frasier-
Reid et al. in relation to the seminal armed-disarmed
glycosylation concept.9 Alternatively, external nucleo-
philes can be used to modulate the reactivity of the glycosyl
donor. A bromide nucleophile was exploited as a catalyst
in the conversion of R-glycosyl bromide to the more
reactive β-glycosyl bromide for glycosylation coupling.10
β-Glucosyl onium adducts prepared from substitution of
R-glucosyl bromide with dimethyl sulfide (Me2S), triethy-
lamine (Et3N), and phosphine (Ph3P) nucleophiles exhib-
ited different glycosylation reactivity.11 A difference in
reactivity was also observed for glycosyl sulfonium ad-
ducts prepared from different thioether nucleophiles.12,13
In 2011, we described an R-selective glycosylation
method14 that used dimethylformamide (DMF) to trap
glycosyl oxacarbenium ions as glycosyl imidinium ad-
ducts. Nucleophilic displacement of the glycosyl imidi-
nium adducts with an acceptor furnishes a glycosylation
product with good R-selectivity. We coined the term
“DMF-modulated glycosylation” for this method. Later
studies found that this method was impractically slow for
(1) Seeberger, P.-H. Chem. Soc. Rev. 2008, 37, 19–28.
(2) Hsu, C.-H.; Hung, S.-C.; Wu, C.-Y.; Wong, C.-H. Angew. Chem.,
Int. Ed. 2011, 50, 11872–11923.
(3) Ranade, S. C.; Demchenko, A. V. J. Carbohydr. Chem. 2013, 32,
1–43.
(4) Kerns, R.; Wei, P. ASC Symp. Ser. 2012, 235–263.
(5) Enugala, R.; Carvalho, L. C. R.; Pires, M. J. D.; Marques,
M. M. B. Chem.;Asian J. 2012, 7, 2482–2501.
(6) Manabe, S.; Ito, Y. Curr. Bioactive Compd. 2008, 4, 258–281.
(7) Yang, L.; Qin, Q.; Ye, X.-S. Asian J. Org. Chem. 2013, 2, 30–49.
(8) Zhang, Z.; Ollmann, I. R.; Ye, X.-S.; Wischnat, R.; Baasov, T.;
Wong, C.-H. J. Am. Chem. Soc. 1999, 121, 734–753.
(9) (a) Fraser-Reid, B.; Wu, Z.; Udodong, U.; Ottoson, H. J. Org.
Chem. 1990, 55, 6068–6070. (b) Fraser-Reid, B.; Wu, Z.; Andrews,
C. W.; Skowrinski, E. J. Am. Chem. Soc. 1991, 113, 1434–1435.
(10) Lemieux, R. U.; Hendriks, K. B.; Stick, R. V.; James, K. J. Am.
Chem. Soc. 1975, 97, 4056–4062.
(11) West, A.; Schuerch, C. J. Am. Chem. Soc. 1973, 95, 1333–1335.
(12) Nokami, T.; Shibuya, A.; Manabe, S.; Ito, Y.; Yoshida, J.
Chem.;Eur. J. 2009, 15, 2252–2255.
(13) Nokami, T. Trends Glycosci. Glycotechnol. 2012, 24, 203–214.
(14) Lu, S.-R.; Lai, Y.-H.; Chen, J.-H.; Liu, C.-Y.; Mong, K.-K. T.
Angew. Chem., Int. Ed. 2011, 50, 7315–7320.
r
10.1021/ol402519c
XXXX American Chemical Society

glycosylations of secondary glycosyl acceptors with
2-azido-2-deoxy-glycosyl donors. Because 2-azido-2-deoxy-
glycosyl donors are widely used for R-glycoside forma-
tion,4,5,15 a more effective modulated glycosylation method
is necessary.
Scheme 1. Glycosylation with Nucleophile (Nu) Additives
Because glycosyl imidinium adducts are the key in
modulated glycosylation, we speculated that their reactiv-
ity might be tuned by variation of the nucleophile additive.
Accordingly, we designed experiments to evaluate the
properties of different nucleophiles in the modulated gly-
cosylation context using model glycosyl substrates 1, 2,
and 3 (Scheme 1, Table 1). At the outset, thiogalactosyl
donor 1 undertaken in the modulated glycosylation pro-
cedure was coupled with galactosyl acceptor 3.14 In the
present study, N,N-diisopropylformamide (DIPF), N-for-
mylpiperidine (NFP), N-formylmorpholine (NFM), di-
methylacetamide (DMA),16 tetramethylurea(TMU),17 tri-
phenylphosphine oxide (TPP),18 and diphenyl sulfoxide
(DPSO)19 were used as the nucleophiles for evaluation
(Table 1, entries 1À9).
In general, formamide compounds offer higher R-selec-
tivity of glycosylation than do nonformamide compounds
(entries 1À3 vs 4À6). No glycosylation product was ob-
tained when DPSO was applied. Among the formamide
compounds examined, DIPF offered the highest R-selec-
tivity of glycosylation, but the longest reaction time of 6 h
was required (entry 4). In contrast, NFM-modulated
glycosylation could be completed within 2 h, but the
R-selectivity obtained was moderate (entry 6). Next, we
increased the amount of DIPF and NFM to 4.0 equiv
(entries 7 and 8).
Although the DIFP-modulated glycosylation produced
an excellent 36:1 R/β ratio of disaccharide 4, the reaction
yield dropped from 74% to68%, and the reaction time was
prolonged from 6 to >10 h (entry 4 vs 7). In contrast, the
reaction time and yield of the NFM-modulated glycosyla-
tions were similar at 2.0 and 4.0 equiv of NFM, and to
our delight, the R/β ratio of 4 was improved from 5:1 (at
2 equiv) to 13:1 (at 4 equiv). Along this line, we increased
the amount of NFM to 16.0 equiv, and the R/β ratio was
further raised to 39:1 (entry 9).20
Table 1. Results of Glycosylation of Galactosyl Acceptor 3 with
Thiogalactoside 1 or 2-Azido-2-deoxythiogalactoside 2
nucleophile
entry donor
(equiv)
time (h) product, yield (%), R/βa
1
1
1
1
1
1
1
1
1
1
2
2
2
2
DMA (2)
TMU (2)
TPP (2)
5
4, 77, 3:1
4, 70, 1:1
4, 83, 1.2:1
4, 74, 10:1
4, 77, 8:1
4, 85, 5:1
4, 68, 36:1
4, 92, 13:1
4, 92, 39:1
5, 90, 4:1
5, 92, 8.5:1
5, 90, 11:1
5, 91, 13:1
2
6
3
2
4
DIPF (2)
NFP (2)
NFM (2)
DIPF (4)
NFM (4)
NFM (16)
NFM (2)
NFM (4)
NFM (8)
NFM (16)
6
5
3
6
2
7
>10b
2
8
9
2
10
11
12
13
2
5
5
6
a R/β of 4 and 5 were determined by HPLC analysis. b The reaction
was not complete, and ca. 10À20% of acceptor 3 was recovered.
These evaluation studies convinced us to apply NFM as
the modulator for glycosylations with 2-azido-2-deoxy-
thioglycosyl donors. To validate the reaction procedure,
2-azido-2-deoxythiogalactosyl donor 2 was activated with
NIS and TMSOTf in the presence of 2.0, 4.0, 8.0, and
16.0 equiv of NFM, followed by coupling with the
galactosyl acceptor 3 (entries 10À13). The glycosylations
were complete in ca. 5À6 h withexcellent 91% yield despite
the stoichiometric amount of NFM, and the highest 13:1
R/β ratio of glycosylation product 5 was obtained with
16.0 equiv of NFM.
After validating the glycosylation procedure, we applied
the low-temperature NMR spectroscopy method to probe
the glycosyl imidinium adducts. Thus, 2-azido-2-deoxy
thiogalactoside 6 in CDCl3 was activated with NIS and
TMSOTf promoters in the presence of 2.0 equiv of the
DMF, DIPF, or NFM nucleophile, and the preactivation
mixture was taken for NMR analysis.
(15) (a) Paulsen, H.; Kalar, C.; Stenzel, W. Chem. Ber. 1978, 111,
2358–2369. (b) Lemieux, R. U.; Ratcliffe, R. M. Can. J. Chem. 1979, 57,
1244–1251.
(16) (a) Koto, S.; Morishima, N.; Owa, M.; Zen, S. Carbohydr. Res.
1984, 130, 73–83. (b) Koto, S.; Asami, K.; Hirooka, M.; Nagura, K.;
Takiazawa, M.; Yamamoto, S.; Okamoto, N.; Sato, M.; Tajima, H.;
Yoshida, T.; Nonaka, N.; Sato, T.; Zen, S.; Yago, K.; Tomonaga, F.
Bull. Chem. Soc. Jpn. 1999, 72, 765–777.
(17) (a) Jiaang, W.-T.; Hsiao, K.-F.; Chen, S.-T.; Wang, K.-T.
Synlett 1999, 1687–1690. (b) Jiaang, W.-T.; Chang, M.-Y.; Tseng,
P.-H.; Chen, S.-T. Tetrahedron Lett. 2000, 41, 3727–3730.
(18) Kobashiy, Y.; Mukaiyama, T. Chem. Lett. 2004, 33, 874–875.
(19) Crich, D.; Li, W. Org. Lett. 2006, 8, 959–962.
A set of signals at chemical shifts (δ) of 6.18, 6.30, and
6.26 ppm were found in the H NMR spectra of the
1
preactivation mixture (Figure 1). Referring to spectro-
scopic data of known R-imidinium adducts,14 the signals
(20) A further increase in NFM did not improve the R-selectivity of
the reaction.
1
were assigned to the R-anomeric H signals (H1R) of the
B
Org. Lett., Vol. XX, No. XX, XXXX

DMFÀ, DIPFÀ, and NFMÀimidinium adducts 7, 8, and
9, respectively (Scheme 2). From HSQC- and HMBC-
NMR analysis, the corresponding R-anomeric 13C signals
of 7, 8, and 9 were also identified (see the Supporting
Information).
Scheme 2. Production and NMR Detection of Formamide
Imidinium Adducts 7À9
Figure 1. (a) Selected 1H spectrum of DMFÀimidinium adduct
1
7. (b) Selected H spectrum of DIPFÀimidinium adduct 8. (c)
Selected 1H spectrum of NFMÀimidinium adduct 9.
(see Scheme S1 in the Supporting Information). In pre-
activation, glycosyl donor is activated to form glycosyl
oxacarbenium ions, which are trapped by formamide
molecules and converted to a mixture of R- and β-glycosyl
imidinium adducts.25 The subsequent reaction of the
β-imidinium adduct with the acceptor furnishes the
R-glycosylation product. Such a mechanism is consistent
with CurtinÀHammett kinetics.26
1
Closer examination of the H spectrum of the DIPFÀ
imidinium adduct 8 revealed a smaller signal at chemical
shifts (δ) of 5.56 ppm (Figure 1b). Driven by curiosity, we
magnified the spectra of DMFÀ and NFMÀimidinium
adducts 7 and 9. As such, two tiny but distinctive signals
emerged at chemical shifts (δ) of 5.47 and 5.50 ppm
Next, we investigated the scope of the NFM-modulated
glycosylation. The 2-azido-2-deoxythioglucoside (GlcN3)
11 and 2-azido-2-deoxythiogalactosides (GalN3) 2 and 10
were coupled with glycosyl acceptors 12À14, and serine
acceptor 15 using the NFM-modulated glycosylation
procedure (Scheme 3, Table 2). For comparison, the
NFM-modulated glycosylation (method A) and the low-
concentration glycosylation (LCG) procedures (method B)
were performed in parallel.27 In the LCG method, a 1:2:1
CH2Cl2ÀCH3CNÀEtCN solvent mixture was used to
obtain a high β-selectivity in glycosylation.
3
(Figure 1a and c). The JH/H coupling constants of these
signals lie between 7.0 and 8.5 Hz, implicating a β-config-
uration if they were anomeric protons.21 In addition, the
corresponding 13C signals of these protons were also
identified through HSQC- and HMBC-NMR analysis.22
On the basis of the NMR data and the literature,23
we assigned these H and 13C signals as the anomeric
1
1H and 13C signals of the β-imidinium adducts 7À9.
Additional support of the assignment was given by the
formyl 1H signals of R/β-imidinium adducts (at ca.
δ = 9 ppm).
The glycosylations of the acceptors 12aÀ15 with the
GlcN3 donor 11 under the NFM modulation conditions
furnished the disaccharides 16À19 in yields of 70À80%. The
R/β ratios of the glycosylation products spanned from 11:1
to 19:1 (entries 1, 3, 5, and 7). Note that the disaccharide 18
is a repeating unit in heparin sulfate oligomers.28 With the
exception of the acceptor 14, under the LCG conditions, the
glycosylations of the acceptors 12a, 13, and 15 with the same
GlcN3 donor gave the β-anomers of the disaccharides 16,
17, and 19 with high selectivity (entries 2, 4, and 8).
1
From the ratio of the R/β-anomeric H signals, the
equilibrium constants (Keq) of 0.07, 0.34, and 0.01 were
determined for formamideÀimidinium adducts 7, 8, and 9,
respectively.12,13,24 The concentration of β-imidinium ad-
duct 9 was the lowest among the three β-imidinium
adducts, implying that the β-imidinium adduct 9 is also
the least stable species in modulated glycosylation.
With this new NMR information, we propose the fol-
lowing mechanism of the formamide-modulated glycosylation
(21) Bock, K.; Pedersen, C. J. Chem. Soc., Perkin Trans. 2 1974, 293–
297.
(22) HSQC and/or HMBC correlation NMR spectra of the imidi-
nium adducts 7À9 are provided in the Supporting Information.
(23) Lin, Y.-H.; Ghosh, B.; Mong, K.-K. T. Chem. Commun. 2012,
48, 10910–10912.
(26) Seeman, J. I. Chem. Rev. 1983, 83, 83–134.
(27) Chao, C.-S.; Li, C.-W.; Chen, M.-C.; Chang, S.-S.; Mong, K.-K. T.
Chem.;Eur. J. 2009, 15, 10972–10982.
(28) Bishop, J. R.; Schuksz, M.; Esko, J. D. Nature 2007, 446, 1030–
1037.
(24) Park, J.; Kawatkar, S.; Kim, J. H.; Boons, G. J. Org. Lett. 2007,
9, 1959–1962.
(29) When NIS and TMSOTf were used for the glycosylation of 15
with 11, the reaction was sluggish. Therefore, we applied NIS and TfOH
as glycosylation promoters, which gave a cleaner reaction.
ꢀ
(25) Bohe, L.; Crich, D. C. R. Chim. 2011, 14, 3–16.
Org. Lett., Vol. XX, No. XX, XXXX
C

Table 2. NFM-Modulated Glycosylation and Low Concentration
Glycosylation (LCG) with 2-Azido-2deoxythioglycosyl Donors
2, 10, and 11
Scheme 3. Glycosylation Studies of Donors 2, 10, and 11 with
Acceptors 12À15
methoda
entry donor/acceptor (NFM equiv) T (°C)
time (h), product, yield (%),
R:βb
1
11/12a
11/12a
11/13
11/13
11/14
11/14
11/15
11/15
10/12a
10/12a
10/13
10/13
2/14
A (16)
B
18, À5
16, 70, 19:1
16, 76, 1:16
17, 84, 19:1
17, 60, 1:16
18, 75, 16:1
18, 74, 1:2
2
2, À60
12, À10
1.5, À60
12, À5
3
A (16)
B
4
5
A (16)
B
A (16)c
Bc
6
5, À60
12, À5
7
19, 81, 11:1
19, 74, 1:9
8
1, À60
17, 0
9
A (16)
B
20, 83, 13:1
20, 80, 1:19
21, 82, 32:1
21, 70, 1:19
22, 89, R only
22, 85, 1:10
23, 90, R only
23, 79, 1:12
24, 60, 15:1
25, 65, R-only
10
11
12
13
14
15
16
17
18
1, À60
3, À10
2.5, À60
24, À10
1, À60
18, À10
1.5, À60
18, À10
7, À5 to 0
A (4)
B
A (16)
B
2/14
2/15
A (16)
B
2/15
11/12b
10/12b
A (8)
A (16)
a Method A refers to NFM-modulated glycosylation, and method
B refers to LCG. b The R/β ratios of 16À23 were determined by
HPLC analysis except for 24 and 25. c TfOH was used as an acid
promoter.29
Concerning the glycosylations of the acceptors 12aÀ15
with GalN3 donors 2 and 10, the disaccharides 20, 21, 22,
and 23 were furnished with high R-selectivity under the
NFM modulation conditions (entries 9, 11, 13, and 15).
The R/β ratios of the disaccharides spanned from 13:1 to
R-exclusive. Under LCG conditions, the glycosylations of
the same set of glycosyl acceptors with the GalN3 donors pro-
duced the disaccharides 20, 21, 22, and 23 in good yields,
but the selectivity of the reactions was reversed (entries 10,
12, 14, and 16). We also examined the glycosylation of
thioglucosyl acceptor 12b with 2-azido-2-deoxythioglycosyl
donors 10 and 11 (entries 17 and 18). The glycosylation
furnished the disaccharide thioglycosides 24 and 25 in
acceptable yields with high R-selectivity.
primary glycosyl acceptors with GlcN3 donor is inadequate
under the NFM modulation conditions (unpublished
data).
In this study, we evaluated the chemical properties of
nucleophiles in a modulated glycosylation context and
identified N-formyl morpholine (NFM) as a reactive mod-
ulator for less reactive2-azido-2-deoxythioglycosyl donors.
Acknowledgment. We thank the National Science
Council of Taiwan (NSC 102-2113-M-009-009) and the
Center for Interdisciplinary Science of NCTU for support.
Thanks to Dr. C.S. Chao for initial exploration on NFM
modulation.
We observed that the R-selectivity of glycosylation is
generally higher for the GalN3 donor than for the GlcN3
donor. For example, in modulated glycosylation of 13 with
GalN3 10, 4.0 equiv of NFM was enough to afford an
excellent R/β ratio of 32:1 (entry 11). However the glyco-
sylation of 13 with GlcN3 donor 11 required 16.0 equiv of
NFM, although the R/β ratio of the product was slightly
lower (entry 3). Thus far, the selectivity of glycosylation of
Supporting Information Available. Synthetic procedures,
HPLC chromatographs for determination of R-toβ-anomer
ratio in Tables 1 and 2, and data for new compounds. This
material is available free of charge via the Internet at
http://pubs.acs.org.
The authors declare no competing financial interest.
D
Org. Lett., Vol. XX, No. XX, XXXX