Glycosidation with a Disarmed Glycosyl Iodide: Promotion and Scope

Article abstract of DOI:10.1021/ol035475k

(Equation presented) Glucuronyl iodide 1 has been studied in detail as a "disarmed" glycosyl donor. In a model reaction, using N-iodosuccinimide (NIS) as a promoter and 2-phenylethanol as acceptor, best results were obtained using NIS with I₂,

Full text of DOI:10.1021/ol035475k

ORGANIC
LETTERS
2003
Vol. 5, No. 24
4545-4548
Glycosidation with a Disarmed Glycosyl
Iodide: Promotion and Scope
Jennifer A. Perrie,† John R. Harding,‡ Clare King,‡ Deborah Sinnott,† and
,†
Andrew V. Stachulski*
The Robert Robinson Laboratories, Department of Chemistry, UniVersity of LiVerpool,
LiVerpool L 69 7 ZD, UK, and Drug Metabolism and Pharmacokinetics Department,
AstraZeneca UK, Ltd., Mereside, Alderley Park, Macclesfield, Cheshire,
SK10 4 TG, UK
stachuls@liV.ac.uk
Received August 5, 2003
ABSTRACT
Glucuronyl iodide 1 has been studied in detail as a “disarmed” glycosyl donor. In a model reaction, using N-iodosuccinimide (NIS) as a
promoter and 2-phenylethanol as acceptor, best results were obtained using NIS with I2, followed by trimethylsilyltrifluoromethanesulfonate
(TMSOTf). When a series of primary and secondary alcohols was glycosylated using these conditions, yields of 60−83% of â-glucuronides
were obtained. Various “nonheavy” metal salts also effectively catalyzed the model reaction but led to significant amounts of r-product with
less reactive secondary alcohols.
Glycosidation is a fundamental process in organic chemistry,1
and there is continuing interest in the development of new
methodology in this area. Of late there has been renewed
interest in glycosyl iodides as glycosyl donors, generally in
the “armed” series (typically benzyl ether-protected sugars).2,3
We recently reported4 that the “disarmed” glycosyl iodide 1
derived from glucuronolactone was a highly stable derivative
of well-defined crystal structure and commented briefly on
its potential in glycosidation. We now present a detailed study
of the glycosyl donor ability of 1, using mainly iodine-based5
reagents and excluding heavy metals. Iodosugar 1 has proved
to be a good and flexible donor, reacting under mild
conditions (0°-20 °C) more efficiently than the correspond-
ing bromosugar and with a wider range of substrates,
including hindered and deactivated alcohols.
Although iodine alone was a satisfactory catalyst for the
glucuronidation of 3-O-pivaloyl morphine by 1 (55% yield),4
this procedure proved to lack generality when applied to a
range of alcohols. We investigated NIS, which had given
good results with the corresponding glucuronyl bromide,6
in the reaction of 1 with 2-phenylethanol 2 (Scheme 1) for
primary evaluation.
† University of Liverpool.
Reaction of 1 with 2 (1.5 equiv) promoted by NIS (1.1
equiv) and I2 (0.25 equiv) in 1,2-dichloroethane at 0-20 °C
in the presence of 3 Å molecular sieves7 gave (81%) a
mixture of the desired â-glucuronide 3 and ortho ester 4 upon
complete reaction of 1; in this series, such ortho esters proved
to be highly stable.8
‡ AstraZeneca UK, Ltd.
(1) For example: Paulsen, H. Angew. Chem., Int. Ed. Engl. 1982, 21,
155-173. Boons, G. J. Tetrahedron 1996, 52, 1095-1121. Davis, B. G. J.
Chem. Soc., Perkin Trans. 1 2000, 2137-2160.
(2) Gervay, J.; Hadd, M. J. J. Org. Chem. 1997, 62, 6961-6967.
(3) (a) Gervay, J. Glycosyl Iodides in Organic Synthesis. In Organic
Synthesis: Theory and Applications; JAI Press, Inc.: 1998; Vol. 4, pp 121-
153. (b) Gervay-Hague, J.; Lam, S. N. Carbohydr. Res. 2002, 337, 1953-
1965. (c) Gervay-Hague, J.; Lam, S. N. Org. Lett. 2002, 4, 2039-2042.
(d) Gervay-Hague, J.; Ying, L. Q. Carbohydr. Res. 2003, 338, 835-841.
(4) Bickley, J.; Cottrell, J. A.; Field, R. A.; Harding, J. R.; Ferguson, J.
R.; Hughes, D. L.; Kartha, K. P. R.; Law, J. L.; Scheinmann, F.; Stachulski,
A. V. Chem. Commun. 2003, 1266-7.
(5) (a) Kartha, K. P. R.; Aloui, M.; Field, R. A. Tetrahedron Lett. 1996,
37, 8807-8810. (b) Kartha, K. P. R.; Karkkainen, T. S.; Marsh , S. J.;
Field, R. A. Synlett 2001, 260-262.
(6) Stachulski, A. V. Tetrahedron Lett. 2001, 42, 6611-6614.
10.1021/ol035475k CCC: $25.00 © 2003 American Chemical Society
Published on Web 10/30/2003

All yields were from 60 to 83%: the somewhat lower yield
for benzyl alcohol is mainly due to some oxidation to
benzaldehyde. The relative proportions of ortho ester and
directly formed glucuronide varied for different alcohols, and
in general it was simpler to add TMSOTf (0.5 equiv) in one
portion at -15 °C upon complete reaction of 1: rearrange-
ment of ortho ester was complete in about 2 h in all cases.
We included steroidal alcohols (entries 5 and 6) because
steroidal glucuronides are of great importance in metabolism
and as analytical standards.
Scheme 1. Glycosidation of 2-Phenylethanola
The steroidal results are noteworthy and should be
compared with those obtained using the corresponding
imidate donor. Thus, the known tetrapivaloate 59 (Figure 3)
a Reagents and conditions: (a) NIS, I2, 0-20 °C, then Lewis
acid; see text.
Addition of I2 was beneficial: NIS alone gave poorer
yields and significantly more degradation. Simply stirring 1
and 2 together with the sieves in DCE gave no reaction. The
use of 0.8 g of 3 Å sieves per 0.2 mmol of 1 was important
to reduce degradation and give clean â-product. The mixture
of 3 and 4 was treated with either BF3‚Et2O or TMSOTf to
effect rearrangement of 4 to 3; the latter proved to be slightly
higher-yielding, though in either case a small amount of
degradation resulted. Marginally the best procedure involved
the addition of TMSOTf (0.5 equiv) from the start, followed
by an additional 0.25 equiv when 1 had wholly reacted,
affording 3 in 76% yield. 1,2-Dichloroethane (DCE) was
clearly the best solvent: acetonitrile led mainly to degrada-
tion products.
Figure 1. Chemical structure of glycosyl iodide (“donor”) 1.
was subjected to anomeric deacylation (N2H4-AcOH, DMF)
to afford the hemiacetal 6 in 74% yield. Treatment of 6 with
Cl3CCN-K2CO3 in CH2Cl2 then gave imidate 710 (87%)
previously prepared via the glucuronyl bromide 8. We
employed 7 for the first time in glucuronidation of steroidal
alcohols and, using 1.2 equiv 7 with TMSOTf catalysis,
obtained the â-glucuronide esters derived from epiandros-
terone and 3-benzoylestradiol in 36 and 65% yields, respec-
tively; cf. Table 1, entries 5 and 6. Clearly, iodosugar 1
performs very well here and takes one step fewer to prepare
than 7.
Having established the reliability of the NIS-I2-TMSOTf
method, we then evaluated 1 as a donor for a range of
alcohols, with the results summarized (Table 1: total reaction
The results with monosaccharide acceptors (Table 1,
entries 7 and 8) show clear potential for oligosaccharide
synthesis: in particular, the glucose tetraacetate (entry 8),
regarded as a poor acceptor, still gives a very acceptable
result (62%).
In contrast to glycosidation with bromide 8 + NIS,6
secondary alcohols now react at a good rate (6 h or less in
all cases studied) in highly acceptable yields without large
excesses of promoters. For the more nucleophilic alcohols
(viz. entries 1-4, 7, and 8), no R-anomer could be detected
Table 1. Glucuronidation of Various Alcoholsa Using 1 +
NIS/I2/TMSOTfb in the Presence of 3 Å Molecular Sieves
entry
alcohol
yield (%) â: R ratio
1
2
3
4
5
6
7
8
Ph(CH2)2OH
PhCH2OH
c-C6H11OH
3-pentanol
epiandrosterone
3-O-benzoylestradiol
76c
60d
67
77
65
70
83
62
â only
â only
â only
â only
â:R ) 96:4
â:R ) 96:4
â only
di-O-isopropylidene-D-galactose
â-D-glucose-1,2,3,4-tetraacetate
â only
a Performed with 1.5 equiv of acceptor (alcohol) in each case. b NIS (1.1
equiv) and I2 (0.25 equiv) were added at 0 °C initially, then TMSOTf (0.5
equiv) was added at -15 °C when all 1 had reacted. c TMSOTf added in
two portions: see text. d Some oxidation to benzaldehyde seen.
time of 6 h in all cases). In all cases, we used just 1.5 equiv
of acceptor, not a large excess, which would be unrealistic
for more complex alcohols.
Figure 2. Chemical structure of ortho ester 4.
Org. Lett., Vol. 5, No. 24, 2003
(7) Spherical 3 Å molecular sieves of 1/16 in. diameter were used,
activated by drying at 130 °C.
4546

range of transition metals and eVen lanthanides (normally
regarded as oxophilic, not halophilic: entries 6 and 7) are
now effective and give over 50% yields: the results with
FeCl3 (entry 1) and CuCl (entry 3) are particularly good.
We emphasize that, apart from entries 1 and 3, these results
are unoptimized but refer to pure, chromatographed product.
“Harder” metal cations, as in entries 8-10, are less effective.
The use of LiClO4 (entry 10) in conjunction with an in situ-
formed “armed” glycosyl iodide was reported before.14
Our early results with this method gave variable R/â ratios,
but we discovered that on using a minimum quantity of 3 Å
molecular sieves (0.8 g per 0.2 mmol of 1), the anomer 3â
was cleanly produced (<2% 3R by NMR): cf. the NIS
method above. No significant amounts of ortho ester were
seen in this reaction mode.
Figure 3. Glucuronate Intermediates.
even in the NMR of crude product: for the less reactive
steroidal secondary alcohols (entries 5 and 6), about 4% of
the R-product was detected.11 As the â-products are ano-
merically stable under the reaction conditions, we attribute
this small “slippage” to R/â glycosyl iodide exchange prior
to glycosidation. Small (<5%) quantities of R-anomers are
not a concern, being readily crystallized out.
The FeCl3-I2 + 1 method was then applied to the same
range of acceptors as in Table 1 with the results shown in
Table 3.
Table 3. Glucuronidation of Various Alcoholsa Using 1 +
b
FeCl3/I2 with 3 Å Molecular Sieves
We also investigated various “nonheavy” metal salts as
promoters, with or without I2 present. The use of Zn halides
with a glucuronyl bromide but with no added I2 had been
reported previously,12 and a few other metal salts have been
used in conjunction with glucosyl bromides.13 The results
for the model reaction, Scheme 1, are summarized in Table
2.
entry
alcohol
yield (%) R: â ratio
1
2
3
4
5
6
7
8
9
Ph(CH2) 2OH
79
â only
â only
â only
â only
â:R ) 9:1
â only
â:R ) 9:1
â only
â only
â only
PhCH2OH
c-C6H11OH
3-pentanol
76
67
70
42
epiandrosterone
epiandrosterone
3-O-benzoylestradiol
di-O-isopropylidene-D-galactose
â-D-glucose-1,2,3,4-tetraacetate
monosaccharide 9
64c
68
57
10d
61c
a
Table 2. Glucuronidation of 2 with 1 Using MXn ( I2
10
entry
promoter
I2 equivb time, comments yield of 3 (%)
a Performed with 1.5 equiv of acceptor (alcohol) in each case. b FeCl3
(1.1 equiv), I2 (1.5 equiv), 20 h at 20 °C in 1,2-dichloroethane. c Replacing
FeCl3 with CuCl. d See text.
1
2
3
4
5
6
7
8
9
FeCl3
1.5
0
1.5
1.5
1.5
0
1.5
0
1.5
1.5
24 h, complete
24 h, complete
5 h, complete
26 h, complete
26 h, complete
26 h, incomplete
72 h, incomplete
50 h, complete
96 h, incomplete
72 h, incomplete
79
65
88
66
62
58
47
43
19
13
ZnCl2
CuCl
NiCl2
c
With primary alcohols, yields were very comparable to
those seen using NIS-I2-TMSOTf, and clean â-products
resulted. However, when this method was used for less
reactive secondary steroidal alcohols, entries 5 and 7,
anomeric mixtures containing 10% R-anomer resulted. We
suggest that the entrainment of halide ion by the sieves,
which inhibits anomeric halide exchange, is not complete
and therefore the less nucleophilic alcohols no longer give
clean â-products.15 The results with CuCl (entries 6 and 10)
suggest that appropriate selection of metal salt may, at least
in some cases, eliminate R/â exchange: the faster rate seen
for CuCl over FeCl3 in the model reaction (Scheme 1 and
Table 2), however, is not maintained for the steroidal
alcohols.
c
NiI2
CeCl3 (2 equiv)
Yb(OTf)3
Sc(OTf)3
MgI2
10 LiClO4
a Using 1.1 equiv of metal halide and 1.5 equiv of I2 unless otherwise
stated, in the presence of 3 Å molecular sieves (for quantity, see text) at 20
°C in DCE. b For entries 1, 3, 4, 5, 7, 9, and 10 the reaction without I2 was
impracticably slow. c Using 4 Å sieves (same quantity).
In contrast to the traditional Koenigs-Knorr reaction,
where Ag, Hg, and Cd are used almost exclusively, a wide
(8) Ortho ester derived from 2-trimethylsilylethanol and 1 was fully
characterized: in particular, δH (CDCl3) 4.29 (1 H, approximately t, 2-H)
and 5.91 (1 H, d, J ) 4.8 Hz, 1-H).
(12) Rukhman, I.; Nisnevich, G.; Gutman, A. L. Tetrahedron 2001, 57,
1083-1092.
(9) Vlahov, J.; Snatzke, G. Liebigs Ann. Chem. 1983, 570-574.
(10) (a) Kornilov, A. V.; Kononov, L. O.; Zatonskii, G. V.; Shashkov,
A. S.; Nifant’ev, N. E. Russ. J. Bioorg. Chem. (Engl.) 1997, 23, 597-607.
(b) Suzuki, T.; Mabuchi, K.; Fukazawa, N. Bioorg. Chem. Med. Lett. 1999,
9, 659-662.
(13) (a) ZnO (or a preformed Zn alkoxide): Gurudutt, K. N.; Rao, L. J.
M.; Rao, S.; Srinivas, S. Carbohydr. Res. 1996, 285, 159-165. (b) InCl3:
Mukherjee, D.; Ray, P. K.; Chowdhury, U. S. Tetrahedron 2001, 57, 7701-
7704. (c) Sn(OTf)2: Lubineau, A.; Malleron, A. Tetrahedron Lett. 1985,
26, 1713-1716.
(11) For example, the R-product in the epiandrosterone series can be
distinguished by δH (CDCl3), inter alia, 4.40 (d, J ) 9.9 Hz, 5-H), 4.78
(dd, J ) 9.8 and 3.9 Hz, 2-H), and 5.57 (t, J ) 9.8 Hz).
(14) Schmid, U.; Waldmann, H. Tetrahedron Lett. 1996, 37, 3837-3841.
(15) A referee suggested that the R-anomers may be formed by
rearrangement from the ortho ester stage.
Org. Lett., Vol. 5, No. 24, 2003
4547

and slightly superior yields to the metal halide procedure,
in shorter reaction times.
Scheme 2. Disaccharide Synthesis, Anomeric Glycosidea
The low yield obtained with glucose 1,2,3,4-tetraacetate,
Table 3, entry 9, is the result of anomeric ester exchange
and related side reactions (ZnCl2 gave a similar result).
However, the conditions are quite compatible with an
anomeric glycoside as shown in Scheme 2 (using CuCl-
I2), which also illustrates a useful selective C(6)-reduction
of glucuronides of type 3. Here again, the disaccharide yield
is unoptimized but the â-selectivity is complete, affording
1
<2% R-anomer as indicated by H NMR; the NIS-I2-
TMSOTf procedure was equally effective here: cf. Table
1, entry 8.
It is well-known that glucuronates, ceteris paribus, have
the lowest donor reactivity of all pyranose sugars.16 By
inference, disarmed glycosyl iodides of other pyranoses are
likely to show good donor ability. We believe that glycosyl
iodide 1 shows considerable potential as a donor and will
be of interest to other workers in the field.
Acknowledgment. We thank the EPSRC and Astra
Zeneca for maintenance grants (to J.A.P. and D.S.), as well
as Prof. R. A. Field (University of East Anglia) and Dr. R.
Whyman (University of Liverpool) for valuable comments.
a Reagents and conditions: (a) NaBH4, THF, MeOH, 0 °C, 3 h,
81%; (b) 1, CuCl, I2, 3 Å MS, (CH2Cl)2, 0-20 °C, 24 h, 61%.
OL035475K
The NIS/I2/TMSOTf method is therefore on balance our
procedure of choice, as it affords essentially clean â-products
(16) Mueller, T.; Schneider, R.; Schmidt, R. R. Tetrahedron Lett. 1994,
35, 4763-4766.
4548
Org. Lett., Vol. 5, No. 24, 2003