Direct synthesis of β-N-glycosides by the reductive glycosylation of azides with protected and native carbohydrate donors

Article abstract of DOI:10.1002/anie.201301264

A simple and straightforward method for the stereocontrolled synthesis of β-linked N-glycosides uses alkyl and aryl azides as the nitrogen source. The N-glycosides are formed in high yields and with high βselectivities (typically >70 % yield, >15:1 β:α selectivity). This approach is also amenable to the synthesis of N-glycosylated amino acids and peptides (see example, Fmoc=9-fluorenylmethoxycarbonyl). Copyright

Full text of DOI:10.1002/anie.201301264

.
Angewandte
Communications
DOI: 10.1002/anie.201301264
Glycosylation
Direct Synthesis of b-N-Glycosides by the Reductive Glycosylation of
Azides with Protected and Native Carbohydrate Donors**
Jianbin Zheng, Kaveri Balan Urkalan, and Seth B. Herzon*
The b-N-glycoside linkage is embedded within structurally
diverse natural products such as the anthraquinone antibiotics
(e.g. mycorhodin),[1] indigo glycosides (e.g. akashin C),[2] and
the ansamycin antibiotics (e.g. ansacarbamitocin A;
Scheme 1).[3,4] This linkage is also found in a large number
of glycopeptides, which exhibit a broad spectrum of biological
functions;[5] erythropoietin (EPO) is a well-known example.
Reported methods for the synthesis of N-linked glycoside
bonds include the functionalization of glycosyl azides[6] and
the condensation of ammonia,[7,8] N,O-dialkylhydroxyl-
amines,[9] and acyl hydrazides[10] with reducing sugars. We
describe herein a simple and complementary method that
proceeds by mild thermolysis of alkyl and aryl azides in the
presence of reducing sugars and a tertiary phosphine; it is
essentially an aza-Wittig reaction in which a carbohydrate is
used as a latent carbonyl group [Eq. (1)]. Despite the
simplicity of this approach, to our knowledge, a single
report describing the condensation of a polyfluorinated
iminophosphorane and lactose[11] stands as the only direct
precedent for this reaction.
The condensation of ammonia and N,O-dialkyl hydroxyl-
amines with unactivated carbohydrates is well-developed, and
this method has found great utility in the synthesis of
glycoproteins[12] and natural product glycoconjugate li-
braries.[9a–c] However, the reaction of ammonia with a reducing
sugar requires long reaction times and a large excess of
ammonia to drive the process to completion, thus rendering
this process unsuitable for the synthesis of N-glycosides
incorporating precious amine fragments. The condensation of
N,O-dialkyl hydroxylamines with carbohydrates is more
efficient, but necessarily provides an unnatural (neoglyco-
side) linkage. The method we report herein is characterized
by short reaction times, a small (0.5 equiv) excess of reagent,
and high selectivity for the b-N-glycoside product.
The reaction between an azide, phosphine, and carbohy-
drate to form an N-glycoside and phosphine oxide may
proceed by several pathways. Regardless, the overall trans-
formation is formally a Staudinger reduction–aza-Wittig
sequence and we therefore began by evaluating the ability
of various phosphines to promote the coupling of benzyl azide
(1a) and O-allyl-N,N-dimethyl-d-pyrrolosamine (2a,
Table 1).[13] Products derived from the protected 2,6-dideoxy-
glycoside 2a can be readily purified by flash-column chro-
matography, and exhibit well-resolved resonances in their
1H NMR spectra, which facilitated the characterization of the
products and optimization of the reaction conditions. Heating
a mixture of 1a (1.5 equiv) and 2a in the presence of
triphenylphosphine (1.5 equiv) in tetrahydrofuran as the
solvent resulted in the formation of the N-glycoside 3a in
24% yield and with a > 15:1 selectivity for the b anomer
(1H NMR spectroscopic analysis; JH1-H2ax = 10.5 Hz). The
reaction did not proceed at lower temperatures (248C).
When the more reactive dimethylphenylphosphine was
employed, the product 3a was isolated in 56% yield
(entry 2). Reasoning that protic acids might promote the
pyranose–hydroxyaldehyde isomerization, a variety of acidic
additives were evaluated. The addition of para-toluenesul-
fonic acid (5 mol%) increased the yield of the N-glycoside 3a
Scheme 1. Representative natural products that contain the b-N-glyco-
side linkage.
[*] Dr. J. Zheng, Dr. K. B. Urkalan, Dr. S. B. Herzon
Department of Chemistry, Yale University
225 Prospect Street, New Haven, CT 06520-8107 (USA)
E-mail: seth.herzon@yale.edu
Homepage: http://www.chem.yale.edu/faculty/herzon.html
Dr. J. Zheng
Shanghai Key Laboratory of New Drug Design, School of Pharmacy,
East China University of Science and Technology (P.R. China)
[**] Financial support from the National Institute of General Medical
Sciences (R01GM090000), the Searle Scholars Program, Shanghai
Committee of Science and Technology (11DZ2260600, Shanghai
Municipal Education Commission Fellowship to J.Z.), and Yale
University is gratefully acknowledged. We thank Dr. Seann Mulcahy
for preliminary experiments and Dr. Le Li for helpful discussions.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201301264.
6068
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 6068 –6071

Angewandte
Chemie
Table 1: Optimization of the reductive glycosylation.[a]
Table 2: Scope of the reductive glycosylation reaction.[a]
Entry Azide
Carbohydrate donor
Yield
[%][b]
Entry
Phosphine
Additive
Solvent
Yield [%] (b/a)[b]
1
2
3
4
5
PPh3
–
–
THF
THF
THF
THF
DMF
24 (>15:1)
56 (>15:1)
90 (>15:1)
88 (>15:1)
43 (>15:1)
1
2
BnN3 1a
PMBN3 1b
90 3a
72 3b
P(CH3)2Ph
P(CH3)2Ph
P(CH3)3
PTSA (5 mol%)
PTSA (5 mol%)
PTSA (5 mol%)
3
78 3c
1c
P(CH3)2Ph
4
5
84 3d
83 3e
[a] Conditions: Benzyl azide (1a, 1.50 equiv), 2a (1 equiv), phosphine
(1.50 equiv), 558C, 12 h. [b] Yield of isolated product after purification by
flash-column chromatography. THF=tetrahydrofuran, PTSA=p-tolu-
enesulfonic acid, DMF=N,N-dimethylformamide, PMB=para-meth-
oxybenzyl, CBz=benzyloxycarbonyl.
1d
PhN3 1e
6
77 3 f
1 f
7
78 3g
to 90% (entry 3). Trimethylphosphine was also effective as
the reductant (88% yield, entry 4). The product was obtained
in lower yield when N,N-dimethylformamide was employed
as the solvent (43%, entry 5).
1g
8
BnN3 1a
88 3h
This reductive N-glycosylation reaction has proven to be
relatively general in nature (Table 2). As shown in entries 1–4,
primary alkyl azides coupled with 2a in high yields (72–90%).
Phenyl azide (1e) also formed the corresponding N-phenyl-
N-glycoside 3e in 83% yield (entry 5), thus establishing that
this method is not limited to alkyl azide substrates. Ethyl
azidoacetate (1 f) and w-azidolysine methyl ester (1g) both
coupled with 2a in high yields (77 and 78%, entries 6 and 7,
respectively). Benzyl azide (1a) also coupled efficiently with
O-allyl-l-oleandrose (2b, 88%; entry 8). 2,3,5-Tribenzyl-d-
ribose (2c) underwent conjugation with para-methoxybenzyl
azide (1b) in 73% yield (entry 9). N-Acetyl-d-glucosamine
(2d) and the fully deprotected carbohydrates d-glucose (2e)
and maltose (2 f) were shown to couple with benzyl and
phenyl azide in yields of 72–82% (entries 10–13). The product
3i, derived from ribose (entry 9), was formed with a b:a
selectivity of 3:2; in all other cases, the b anomer was obtained
with > 15:1 selectivity (1H NMR spectroscopic analysis).
The results in Table 2 demonstrate that b-linked N-
glycosides are readily formed from simple azides and
reducing sugars. It was of interest to determine if the
method is also amenable to the synthesis of N-linked
glycopeptides and, in particular, to the formation of the b-
Asn-N-acetylglucosamine linkage, which is abundant in
natural glycoproteins.[5] Toward this end, we envisioned
a sequence comprising the N-glycosylation of an ammonia
equivalent, followed by aspartylation.[12,14] After much exper-
imentation, we found that heating trimethylsilylmethyl azide
(1h)[15] and N-acetyl-d-glucosamine (2d) with dimethylphe-
nylphosphine resulted in the formation of the b-linked N-
glycoside 3n in 75% yield as a 30:1 mixture of b and
a anomers (Scheme 2). Benzoylation of 3n (benzoyl chloride)
afforded the benzamide 4a in 85% yield. Finally, the
trimethylsilylmethyl substituent was cleaved by treatment
with ceric ammonium nitrate in acetonitrile (90%).[16] The
benzamide 5a was obtained without detectable erosion of the
anomeric site (1H NMR spectroscopic analysis). This
73 3i
3:2
b:a
9
PMBN3 1b
BnN3 1a
10[c]
76 3j
11[c]
12[c]
BnN3 1a
PhN3 1e
82 3k
78 3l
13[c]
BnN3 1a
72 3m
[a] Conditions: Azide (1 equiv), PhP(CH3)2 (1.00 equiv), carbohydrate
(1.50 equiv), PTSA (5 mol%), THF, 558C, 12 h. All reactions were
conducted on a 0.5 mmol scale. [b] Yield of isolated product after
purification by flash-column chromatography. b:a selectivity >15:1,
unless otherwise noted. [c] Reaction conducted with 1 equiv carbohy-
drate, 1.50 equiv azide, and 1.50 equiv PhP(CH3)2, THF, 558C, 12 h.
PMB=p-methoxybenzyl; CBz=benzyloxycarbonyl.
sequence provides a complementary and highly practical
alternative to the reaction of carbohydrates with ammonia, as
primary N-glycosides are notoriously difficult to purify and
are prone to epimerization at the anomeric position.
This approach was readily adapted toward amino acid
coupling reactions by employing the corresponding amino
acid mixed anhydride in the acylation step. A selection of
glycosylated amino acids and peptides prepared in this way
are shown in Table 3. As a consequence of their high
polarities, the intermediates in these sequences were
Angew. Chem. Int. Ed. 2013, 52, 6068 –6071
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 6069

.
Angewandte
Communications
direct introduction of disaccharides (5e, 61%), which sug-
gests its application in more complex settings.
Although glycosyl azides have been widely used for the
synthesis of N-glycosides,[6] the reductive condensation of
azides with unactivated sugars does not appear to have been
developed. We have disclosed herein a simple and straight-
forward method for the stereocontrolled synthesis of b-linked
N-glycosides by employing alkyl and aryl azides as the sources
of nitrogen atoms. The method is compatible with protected
and deprotected carbohydrate donors, and does not require
activation of the anomeric position. A simple, high-yielding
procedure for the synthesis of b-Asn-linked glycosides has
been developed, and this method may be of great utility in the
synthesis of complex glycopeptides. Efforts to fully delineate
the scope of the azide and carbohydrate components, as well
as applications in natural products synthesis, are underway.
Scheme 2. Synthesis of the N-benzoyl-N-glycoside 5a. TMS=trimethyl-
silyl, Bz=benzoyl, THF=tetrahydrofuran, CAN=ceric ammonium
nitrate.
Experimental Section
Table 3: Glycosylated amino acids and peptides prepared by the
sequence outlined in Scheme 2.[a]
Representative experimental procedure (Table 2, entry 1): Benzyl
azide (1a; 67.0 mg, 500 mmol, 1 equiv) and THF (300 mL) were added
in sequence to an oven-dried 4 mL vial equipped with a teflon-lined
cap and magnetic stir bar. Dimethylphenylphosphine (70.0 mg,
500 mmol, 1.00 equiv) was added dropwise by syringe over 5 min.
The reaction vessel was sealed under an atmosphere of argon and the
mixture stirred for 30 min at 248C. A solution of 2a (161 mg,
750 mmol, 1.50 equiv) in THF (100 mL) and para-toluenesulfonic acid
(5.0 mg, 25.0 mmol, 0.05 equiv) were then added in sequence. The vial
was sealed under argon and then placed in an oil bath that had been
preheated to 558C. The reaction mixture was stirred and heated for
12 h at 558C before cooling the product mixture to 238C. The cooled
product mixture was concentrated to dryness and the residue
obtained was purified by flash-column chromatography (deactivated
with 1% triethylamine/5% acetone in hexanes; eluting with 5%
acetone in hexanes) to provide the N-glycoside 3a as a clear, colorless
oil (two runs: 139 mg and 133 mg, average yield = 90%).
Entry
1
Product
Yield [%][b]
76
2
3
75
70
Received: February 13, 2013
Published online: April 22, 2013
Keywords: aspartylation · glycosides · glycosylation ·
.
Wittig reactions
4
61
[1] a) U. Takeda, T. Okada, M. Takagi, S. Gomi, J. Itoh, M. Sezaki,
M. Ito, S. Miyadoh, T. Shomura, J. Antibiot. 1988, 41, 417; b) U.
Takeda, T. Okada, M. Takagi, S. Gomi, J. Itoh, M. Sezaki, M. Ito,
S. Miyadoh, T. Shomura, J. Antibiot. 1988, 41, 425.
[2] a) R. P. Maskey, I. Grꢀn-Wollny, H. H. Fiebig, H. Laatsch,
Angew. Chem. 2002, 114, 623; Angew. Chem. Int. Ed. 2002, 41,
597; b) R. P. Maskey, I. Grꢀn-Wollny, H. Laatsch, Nat. Prod. Res.
2005, 19, 137.
[a] Conditions: Step 1: 1h (1.50 equiv), PhP(CH3)2 (1.50 equiv), 2d or 2 f
(1 equiv), THF, 558C. Step 2: ClCO2iBu (1.20 equiv), amino acid
(1.10 equiv), TEA (1.20 equiv), THF, ꢀ40!248C. Step 3: CAN
(5.00 equiv), CH3CN/H2O (30:1, v/v), MW, 908C, 10 min. [b] Yield of
isolated product after purification by flash-column chromatography. b:a
selectivity >15:1. TEA=triethylamine, MW=microwave, Bn=benzyl,
Fmoc=9-fluorenylmethoxycarbonyl.
[3] C. E. Snipes, D. O. Duebelbeis, M. Olson, D. R. Hahn, W. H.
Dent, J. R. Gilbert, T. L. Werk, G. E. Davis, R. Lee-Lu, P. R.
Graupner, J. Nat. Prod. 2007, 70, 1578.
[4] For additional examples, see a) S. Omura, Y. Iwai, A. Hirano, A.
Nakagawa, J. Awaya, H. Tsuchiya, Y. Takahashi, R. Masuma, J.
Antibiot. 1977, 30, 275; b) D. E. Nettleton, T. W. Doyle, B.
Krishnan, G. K. Matsumoto, J. Clardy, Tetrahedron Lett. 1985,
26, 4011; c) R. Bonjouklian, T. A. Smitka, L. E. Doolin, R. M.
Molloy, M. Debono, S. A. Shaffer, R. E. Moore, J. B. Stewart,
G. M. L. Patterson, Tetrahedron 1991, 47, 7739; d) S. Matsunaga,
N. Fusetani, Y. Kato, H. Hirota, J. Am. Chem. Soc. 1991, 113,
9690.
employed without purification, although they may be isolated
by reverse-phase chromatography, if desired. This approach
led to the anomerically pure glucosamine derivatives 5b and
5c in high overall yields (76% and 75% for 5b and 5c,
respectively). The Asn-linked tripeptide was also obtained in
excellent yield and with complete b selectivity (5d, 70%).
Finally, we discovered that the method is amenable to the
6070 www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 6068 –6071

Angewandte
Chemie
[5] For selected reviews, see a) R. G. Spiro, Glycobiology 2002, 12,
43R; b) D. P. Gamblin, E. M. Scanlan, B. G. Davis, Chem. Rev.
2009, 109, 131; c) C. Kan, S. J. Danishefsky, Tetrahedron 2009, 65,
9047; d) A. Larkin, B. Imperiali, Biochemistry 2011, 50, 4411.
[6] For selected examples, see a) M. Mizuno, I. Muramoto, K.
Kobayashi, H. Yaginuma, T. Inazu, Synthesis 1999, 162; b) Y. He,
R. J. Hinklin, J. Chang, L. L. Kiessling, Org. Lett. 2004, 6, 4479;
c) K. J. Doores, Y. Mimura, R. A. Dwek, P. M. Rudd, T. Elliott,
B. G. Davis, Chem. Commun. 2006, 1401; for a recent review, see
d) Z. Gyçrgydeꢁk, J. Thiem, Adv. Carbohydr. Chem. Biochem.
2006, 60, 103.
Hoffmann, J. S. Thorson, J. Am. Chem. Soc. 2006, 128, 14224;
c) P. Peltier-Pain, S. C. Timmons, A. Grandemange, E. Benoit,
J. S. Thorson, ChemMedChem 2011, 6, 1347; for reviews, see
d) B. R. Griffith, J. M. Langenhan, J. S. Thorson, Curr. Opin.
Biotechnol. 2005, 16, 622; e) J. M. Langenhan, B. R. Griffith, J. S.
Thorson, J. Nat. Prod. 2005, 68, 1696.
[10] S. p. Peluso, B. Imperiali, Tetrahedron Lett. 2001, 42, 2085.
[11] M. El Ghoul, B. Escoula, I. Rico, A. Lattes, J. Fluorine Chem.
1992, 59, 107.
[12] For selected examples, see a) P. Wang, J. Zhu, Y. Yuan, S. J.
Danishefsky, J. Am. Chem. Soc. 2009, 131, 16669; b) P. Nagorny,
N. Sane, B. Fasching, B. Aussedat, S. J. Danishefsky, Angew.
Chem. 2012, 124, 999; Angew. Chem. Int. Ed. 2012, 51, 975; c) P.
Wang, X. Li, J. Zhu, J. Chen, Y. Yuan, X. Wu, S. J. Danishefsky, J.
Am. Chem. Soc. 2011, 133, 1597.
[7] a) L. M. Likhosherstov, O. S. Novikova, V. A. Derevitskaja,
N. K. Kochetkov, Carbohydr. Res. 1986, 146, C1; b) L. Urge, E.
Kollat, M. Hollosi, I. Laczko, K. Wroblewski, J. Thurin, L.
Otvos, Jr., Tetrahedron Lett. 1991, 32, 3445.
[8] The condensation of allyl- and benzylamine (10 equiv, 4 days
reaction time) with unactivated carbohydrates has been
reported: L. Cipolla, L. Lay, F. Nicotra, C. Pangrazio, L.
Panza, Tetrahedron 1995, 51, 4679.
[9] a) J. M. Langenhan, N. R. Peters, I. A. Guzei, F. M. Hoffmann,
J. S. Thorson, Proc. Natl. Acad. Sci. USA 2005, 102, 12305; b) A.
Ahmed, N. R. Peters, M. K. Fitzgerald, J. A. Watson, Jr., F. M.
[13] S. L. Gholap, C. M. Woo, P. C. Ravikumar, S. B. Herzon, Org.
Lett. 2009, 11, 4322.
[14] S. T. Anisfeld, P. T. Lansbury, J. Org. Chem. 1990, 55, 5560.
[15] K. Nishiyama, N. Tanaka, J. Chem. Soc. Chem. Commun. 1983,
1322.
[16] C. Palomo, J. M. Aizpurua, M. Legido, A. Mielgo, R. Galarza,
Chem. Eur. J. 1997, 3, 1432.
Angew. Chem. Int. Ed. 2013, 52, 6068 –6071
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6071