Hyperbranched polyester supported disaccharide synthesis: The effect of loading level on glycosylation efficiency

Article abstract of DOI:10.1055/s-2005-869863

The efficiency of glycosylation of glycosyl acceptors supported on a hyperbranched polyester (Boltorn H-50) by an activated trichloroacetamidate donor was investigated as a function of the polymer loading. Loading of the hyperbranched polyester with an o-

Full text of DOI:10.1055/s-2005-869863

LETTER
1567
Hyperbranched Polyester Supported Disaccharide Synthesis: The Effect of
Loading Level on Glycosylation Efficiency
Hyperbranched
r
P
olyeste
i
r
Supp
c
orted
D
isaccharid
A
e
Synthesis ssen B. Kantchev, Jon R. Parquette**
The Ohio State University, Department of Chemistry, Columbus, OH 43210, USA
E-mail: parquett@chemistry.ohio-state.edu
Received 26 April 2005
mers suggests enhanced reagent accessibility of the termi-
Abstract: The efficiency of glycosylation of glycosyl acceptors
supported on a hyperbranched polyester (Boltorn H-50) by an acti-
vated trichloroacetamidate donor was investigated as a function of
the polymer loading. Loading of the hyperbranched polyester with
nal groups. Accordingly, branched materials provide a
potential for increased loading levels compared with sup-
ports based on strictly linear systems. The high synthetic
an o-nitrobenzyl photolabile linker at 25% loading capacity (corre- investment associated with preparation of dendrimers
sponding to 1.36 mmol/g) and using five equivalents of the glycosyl
renders their utilization as supports prohibitively expen-
donor was optimal for quantitative transformation (>98%) at both
sive. To provide a low-cost alternative to dendrimers, we
the mono- and disaccharide stage. The disaccharide product ob-
reported the first example of the use of a soluble, hyper-
tained under these optimized conditions was isolated in 16% yield
branched polyester [Boltorn H40 and H50 (1),8 Figure 1]
over five steps.
to support the synthesis of oligosaccharides.9 In addition,
Key words: carbohydrates, glycosylations, polymers, dendrimers,
cluster glycoside synthesis based on Boltorn H20 and H30
solid-phase
polymer cores was recently reported.10 The focus of this
work is to determine the optimal/maximal loading level of
this hyperbranched support for the synthesis of oligosac-
Cross-linked polystyrene resins, first developed in the
charides via the trichloroacetamidate glycosyl coupling
context of solid-phase peptide synthesis,1 have gained in-
protocol.
creased popularity for synthesis of other classes of organic
compounds.2 The facile removal of excess reagents and
HO
OH
HO
O
soluble by-products from the polymer-attached product(s)
by simple filtration constitutes the main advantage of the
resin supports, enabling high-throughput and automated
synthetic protocols. Limitations of these resins include
solvent-dependent swelling behavior, nonlinear reaction
kinetic profiles that hamper the adaptation of synthetic
protocols and analytical techniques to heterogeneous con-
ditions. Linear soluble polymers such as poly(ethylene-
glycol) (PEG) have been employed as alternatives to the
resin supports to circumvent some of these limitations.3
Soluble supports allow reactions to be conducted in solu-
tion, which provides a kinetic advantage and facilitates
the characterization of polymer-bound intermediates by
solution-phase spectroscopy. Precipitation, size-exclusion
chromatography (SEC) or membrane filtration separates
the supported intermediates from soluble reagents. How-
ever, the intrinsically low levels attainable with many
linear soluble polymers, such as in PEG, limit this strategy
to small-scale applications. To address this limitation,
dendrimers,4 hyperbranched polymers5 and dendronized
solid supports have been investigated as high-loading sup-
ports for organic synthesis.6 Dendrimers and hyper-
branched polymers possess an (x – 1)n + 1 (n = degree of
polymerization) number of terminal groups for an ABx-
type repeat unit.7 Furthermore, the putatively globular
structure of most dendrimers and hyperbranched poly-
OH
HO
O
O
O
HO
O
O
OH
OH
O
O
O
O
O
O
O
O
O
O
O
O
HO
HO
O
O
O
O
OH
O
O
OH
HO
OH
HO
HO
Figure 1 Idealized depiction of Boltorn hyperbranched polymers.
The efficiency of the trichloroacetamidate glycosylation
protocol11 was explored as a function of the loading of the
hyperbranched support and the stoichiometry of glycosyl
donor. Accordingly, four polymer samples (5–8) were
functionalized with a photolabile linker at loading levels
ranging from 25% to 100% (Scheme 1). The photolabile
linkage moiety, 4, was prepared by THP-protection of
methyl-4-hydroxymethyl-3-nitrobenzoate12 (2) and hy-
drolysis with KOH. The THP-protected carboxylic acid 4
was then loaded (EDCI, cat. DMAP) onto the Boltorn
H-50 polymer (1) in 25%, 50%, 75% and 100% of its
theoretical loading capacity (8.8 mmol/g), followed by
peracetylation of the remaining hydroxyl groups with
acetyl chloride. The THP protecting group was removed
with cat. HCl in methanol–dichloromethane. Polymers 5–
7 (25–75% loading) were purified by precipitation into
SYNLETT 2005, No. 10, pp 1567–1570
Advanced online publication: 07.06.2005
DOI: 10.1055/s-2005-869863; Art ID: S04405ST
© Georg Thieme Verlag Stuttgart · New York
1
7
.0
6
.2
0
0
5

1568
E. A. B. Kantchev, J. R. Parquette
LETTER
OH
O
OAc
O
OAc
methanol; however, the fully loaded polymer, 8, exhibited
solubility in methanol and, therefore, was precipitated
into CH2Cl2. The linker-polymer conjugates loaded at
25% (5) and 50% (6) showed good agreement between
calculated and observed loading values (Table 1), sug-
gesting complete incorporation of the linker. However,
the polymers loaded at 75% (7) and 100% (8) afforded
loading levels considerably less than the theoretical value.
The lower efficiency of linker incorporation at higher
loading levels may be a consequence of differing steric
accessibilities that exist among the terminal hydroxyl
groups.
a, b
d
HO
HO
HO
TBDMSO
TBDMSO
O
AcO
AcO
AcO
AcO
95%
66%
OR
OH
O
CCl3
NH
9
10 R = Ac
11 R = H
12
c
quant.
Scheme 2 Reagents and conditions: (a) TBDMSCl, cat. DMAP,
pyridine; b) (CH3COO)2O; (c) NH2NH2·AcOH, DMF–THF (1:2); d)
CCl3CN, cat. NaH, CH2Cl2.
6-TBDMS methyl protons at d = 0.0–0.2 ppm in CDCl3
(Table 2). The monoglycosylated products 17–20 were
then capped with acetic anhydride in pyridine to block all
unreacted hydroxyl groups. Treatment with HF·pyridine
in THF cleaved the 6-TBDMS group affording acceptors
21–24 after partitioning between water and ethyl acetate.
Subsequent glycosylation of 21–24 with donor 12 afford-
ed polymer-bound disaccharides 25–28 after precipitation
into methanol.
COOCH3
COOCH3
a
b
HO
THPO
96%
92%
NO2
2
NO2
3
HO
O
1
O
COOH
HO
OAc
O
O
RO
AcO
AcO
HO
c
THPO
OAc
O
O
NO2
d
5–8
NO2
4
12
O
O
OAc
HO
a
Scheme 1 Reagents and conditions: (a) DHP, cat. TsOH, CH2Cl2;
(b) KOH, H2O–THF, then dil. H2SO4; (c) EDCI, cat. DMAP, THF–
pyridine, then excess CH3COCl; (d) cat. HCl, MeOH–CH2Cl2.
NO2
OAc
NO2
13–20 R = TBDMS
21–24 R = H
5–8
b, c
OAc
O
TBSO
AcO
AcO
Table 1 Loaded Linker-Polymer Conjugates
Compound Capacity (%)
Yield (%)a
Loading (mmol/g)b
1.36 (1.32)
OAc
O
O
O
12
O
AcO
AcO
5
6
7
8
25
68
38
49
25
a
O
OAc
50
1.55 (1.49)
25–28
NO2
75
2.20 (2.91)
OAc
O
TBSO
AcO
AcO
100
2.34 (3.41)
OAc
O
a Isolated yields of loaded polymer obtained by precipitation into
MeOH (5–7) or CH2Cl2 (8).
c, 12, a, d
12, a
O
5
17
AcO
AcO
OH
32%
50%
b Determined by 1H NMR using DMF as an internal standard. The the-
oretical loading was calculated at the loading capacity indicated based
on 8.8 mmol/g OH groups in the Boltorn H50 polymer assuming
100% coupling and acetyl capping efficiency.
29
Scheme 3 Reagents and conditions: (a) cat. BF3·OEt2, THF; (b)
(CH3COO)2O (capping); (c) HF·pyridine, THF; (d) hn (350 nm),
THF.
The polymer-supported acceptors 5–8 were each glycosy-
lated with trichloroacetamidate 12 to determine the effect
of loading level on coupling efficiency. Accordingly, do-
nor 12 was prepared by selective silylation of D-mannose
at C-6 followed by acetylation of the remaining hydroxyl
groups (Scheme 2). Subsequent cleavage of the anomeric
acetyl group with hydrazine monoacetate and reaction
with trichloroacetonitrile provided trichloroacetamidate
12 in 66% yield.
Table 2 presents the efficiency for each glycosylation step
as a function of the loading capacity of the support and the
stoichiometry of the glycosyl donor. Several points are
noteworthy: (1) At a 25% loading level (n = 1), whereas
the maximal efficiency occurs using five equivalents of
donor, the greatest increase in efficiency takes place going
from one (43%) to two (81%) equivalents of donor (en-
tries 1 and 2); (2) Increasing the loading capacity from
25% to 100% afforded a corresponding linear decrease in
efficiency from 98% to 33% (entries 5–8); (3) The second
glycosylation step (n = 2) experienced the most sig-
nificant decrease in efficiency upon increasing from 25%
to 50%, but exhibited similar efficiencies at 75% and
100% loading levels (entries 11 and 12). The decrease in
Exposure of polymers 5–8 to trichloroacetamidate 12 in
the presence of BF3·OEt2 in THF at room temperature re-
sulted in complete consumption of 12 within 15 minutes
(Scheme 3). Following precipitation of the polymer into
methanol, the glycosylation efficiencies were determined
by 1H NMR on the support using the normalized ratio of
the o-nitrobenzyl linker protons at d = 7.6–8.7 ppm to the
Synlett 2005, No. 10, 1567–1570 © Thieme Stuttgart · New York

LETTER
Hyperbranched Polyester Supported Disaccharide Synthesis
1569
Table 2 Glycosylation Efficiency as a Function of Loading Levela
TBDMSO
OAc
O
O
AcO
AcO
O
O
OAc
n
NO2
glycosylation product (13–20, 25–28)
Entry
1
Acceptor
n
1
1
1
1
1
1
1
1
2
2
2
2
Loading (mmol/g)
25% (1.36)
25% (1.36)
25% (1.36)
25% (1.36)
25% (1.36)
50% (1.55)
75% (2.20)
100% (2.34)
25% (n.d.)
12 (Equiv)
Efficiencyb
0.43
Product
13
5
5
1
2
3
4
5
5
5
5
5
5
5
5
2
0.81
14
3
5
0.88
15
4
5
0.92
16
5
5
0.98
17
6
6
0.68
18
7
7
0.48
19
8
8
0.33
20
9
21
22
23
24
1.00c
0.40c
0.25c
0.24c
25
10
11
12
50% (n.d.)
26
75% (n.d.)
27
100% (n.d.)
28
a Purified by precipitation into MeOH, centrifugation, and drying in vacuum.
b Determined by 1H NMR in CDCl3.
c Yield from 5–8 over two glycosylation steps (Scheme 3).
glycosylation efficiency at very high loading levels is, in monosaccharide product 17 was recovered on preparative
part, a consequence of the imperfect branching of hyper- scale (Scheme 3)16 due to the partial solubility of the ace-
branched polymers. A recent detailed study on the molec- tate protected, polymer-bound monosaccharide in metha-
ular mass analysis of lower generation Boltorn polyesters nol.17 However, deprotection and glycosylation of 17 with
using SEC in polar solvents has shown that Boltorn H-40 trichloroacetamidate donor 12, followed by irradiation at
has a higher MW (6700 vs. 5100) and wider PDI (2.6 vs. 350 nm for 48 hours in THF afforded dissacharide 29 as a
1.8) compared to the values provided by the manu- mixture of equilibrating anomers in 32% yield (3 steps).18
facturer.13 Accordingly, in contrast to a perfect dendritic The major anomer was assigned to be a,a based on the
structure, the glycosyl acceptors would not be uniformly 1JC–H coupling constants of the anomeric carbons.19
presented at the periphery of the polymer. At higher load-
In conclusion, we determined the efficiency of glycosyla-
ing levels, the less accessible internal sites react less effi-
tion of Boltorn H-50 supported glycosyl acceptors as a
ciently with glycosyl donors. Similarly, the yield of
function of both the polymer loading and equivalents of
chemical reactions conducted on dendrimeric supports
activated donor. The combination of 25% loading level
depends on the dendrimer generation. For example, Brad-
(1.36 mmol/g) and 5 equivalents glycosyl donor was opti-
ley and coworkers found that complete Fmoc-deprotec-
mal for quantitative transformation (>98%) at both the
tion of the terminal aminogroups of polyurea dendrimer-
mono- and disaccharide stage.
derivatized, high loading polystyrene beads is slower at
higher dendrimer generation.14 Lee-Ruff et al. observed
Acknowledgment
that photochemical coupling of cyclobutanones to poly-
hydroxy dendritic PEG supports proceeds with lower effi-
ciency at higher generation or increased branching.15
This work was supported by The Ohio State University through a
University Seed Grant and graduate assistantship to E. A. K. The
generous gift of large samples of Boltorn polymers from the manu-
facturer, Perstorp polyols, is greatly appreciated.
Although quantitative conversion to 17 was observed by
NMR at the 25% loading level, only 50% yield of the
Synlett 2005, No. 10, 1567–1570 © Thieme Stuttgart · New York

1570
E. A. B. Kantchev, J. R. Parquette
LETTER
160 mol%) was added and the solution was stirred for 30
References
min. The solution was concentrated in vacuo and then
precipitated with MeOH (50 mL). After centrifugation,
decantation of the supernatant, rinsing of the polymer mass
with MeOH (3 × 10 mL) and drying under high vacuum,
compound 17 (0.275 g, 0.23 mmol, 50%) was obtained. 1H
NMR (500 MHz, CDCl3): d = 0.08 (s, 6 H), 0.91 (s, 9 H),
4.95–5.42 (m, 5 H), 7.95 (br s, 1 H), 8.32 (br s, 1 H), 8.67 (br
s, 1H).
Compound 25. To a solution of 17 (0.275 g, 0.225 mmol,
100 mol%) in THF, HF·pyridine (0.5 mL, 0.5 g, 5.0 mmol,
222 mol%) was added and the solution stirred for 12 h. The
mixture was diluted with EtOAc (15 mL) and washed with
H2O (5 mL), 10% H2SO4 (5 mL) and sat. NaHCO3 (5 mL).
After drying (MgSO4) and evaporation of the solvent,
compound 21 was obtained as yellow foam and used directly
for the next step. The monosaccharide donor 21 was
glycosylated using the glycosyl donor 12 (0.64 g, 1.13
mmol, 500 mol%), BF3·OEt2 (45 mL, 51 mg, 0.36 mmol, 160
mol%), and dry THF (2.5 mL) as described for 17 above.
The polymer-immobilized disaccharide 25 was isolated after
precipitation into MeOH (25 mL). 1H NMR (500 MHz,
CDCl3): d = 0.09 (s, 3 H), 0.10 (s, 3 H), 1.28 (s, 9 H), 4.80–
5.45 (m, 10 H), 7.95 (br s, 1 H), 8.32 (br s, 1 H), 8.67 (br s,
1 H).
(1) Fmoc Solid Phase Peptide Synthesis: A Practical Approach;
Chan, W. C.; White, P. D., Eds.; Oxford University Press:
Oxford UK, 2000.
(2) (a) Dörwald, F. Z. Organic Synthesis on Solid Phase, 2nd
ed.; Wiley-VCH: New York, 2002. (b) Seneci, P. Solid
Phase Synthesis and Combinatorial Technologies; Wiley
Interscience: New York, 2002.
(3) Reviews: (a) Bergbreiter, D. E. Curr. Opin. Drug Discovery
Dev. 2001, 4, 736. (b) Osburn, P. L.; Bergbreiter, D. E.
Prog. Polym. Sci. 2001, 26, 2015. (c) Toy, P. H.; Janda, K.
D. Acc. Chem. Res. 2000, 33, 546. (d) Gravert, D. J.; Janda,
K. D. Chem. Rev. 1997, 97, 489.
(4) General introduction to dendrimers: Newkome, G. R.;
Moorefield, C. N.; Vögtle, F. Dendrimers and Dendrons:
Concepts, Syntheses, Applications; Wiley-VCH: New York,
2001.
(5) (a) Star and Hyperbranched Polymers; Mishra, M. K.;
Kobayashi, S., Eds.; Plastic Engineering Series 53, Marcel
Dekker: New York, 1999. (b) Frechet, J. M.; Hawker, C. J.
React. Funct. Polym. 1995, 26, 127.
(6) Reviews: (a) Gebbink, R. J. M.; Klein, K.; Cornelis, A.; Van
Klink, G. P. M.; Van Koten, G. Rev. Mol. Biotechnol. 2002,
90, 183. (b) Haag, R. Chem.–Eur. J. 2001, 7, 327.
(7) Kim, Y. H.; Webster, O. Hyperbranched Polymers, In Star
and Hyperbranched Polymers, 53; Mishra, M. K.;
Kobayashi, S., Eds.; Plastic Engineering Series 53, Marcel
Dekker: New York, 1999, 201.
(8) (a) Manufacturer specifications for Boltorn H-40: 64
terminal OH groups, nominal MW 7,316 and PDI 1.8;
Boltorn H-50: 128 terminal OH groups, nominal MW 14,600
and PDI 2.0. (b) Preparation of Boltorn polyesters:
Malmstroem, E.; Johansson, M.; Hult, A. Macromolecules
1995, 28, 1698.
(9) Kantchev, A. B.; Parquette, J. R. Tetrahedron Lett. 1999, 40,
8049.
(10) Arce, E.; Nieto, P. M.; Diaz, V.; Castro, R. G.; Bernad, A.;
Rojo, J. Bioconjugate Chem. 2003, 14, 817.
(11) Reviews: (a) Knerr, L.; Schmidt, R. R. In Solid Support
Oligosaccharide Synthesis and Combinatorial
Carbohydrate Libraries; Seeberger, P. H., Ed.; John Wiley
and Sons: New York, 2001, 67–98. (b) Schmidt, R. R.;
Kinzy, W. Adv. Carbohydr. Chem. Biochem. 1994, 50, 21.
(12) Zehavi, U.; Sadeh, S.; Herchman, M. Carbohydr. Res. 1983,
124, 23.
(13) Zagar, E.; Zigon, M. J. Chromatogr. A 2004, 1034, 77.
(14) Lebreton, S.; Newcombe, N.; Bradley, M. Tetrahedron Lett.
2002, 43, 2479.
Disaccharide 29. Compound 25 (as obtained from the
previous step) was dissolved in dry THF (2.5 mL) in a quartz
vessel, purged with Ar for 10 min, then irradiated with UV
light (350 nm) for 48 h. The solution was evaporated and the
crude mixture applied directly onto a silica gel column. After
chromatography (hexanes–EtOAc, 2:1), compound 29 (50
mg, 0.072 mmol, 32% from 21) was isolated as colorless oil.
Major anomer (a,a): 1H NMR (400 MHz, CDCl3): d = –0.04
(s, 3 H), –0.02 (s, 3 H), 0.84 (s, 9 H), 1.91 (s, 3 H), 1.92 (s, 3
H), 1.96 (s, 3 H), 1.98 (s, 3 H), 2.05 (s, 3 H), 2.08 (s, 3 H),
3.48 (m, 1 H), 3.62–3,82 (m, 5 H), 4.16 (m, 1 H), 4.75 (d,
J = 1.6 Hz, 1 H), 5.13–5.28 (m, 5 H). 13C NMR (100 MHz,
CDCl3): d = –5.01, –4.95, 18.8, 21.1, 21.1, 21.2, 21.2, 21.2,
21.2, 26.3, 62.9, 66.9, 67.2, 67.7, 69.3, 69.6, 69.7, 70.1, 70.4,
71.9, 92.5 (1JC–H = 172.6 Hz), 97.7 (1JC–H = 170.1 Hz),
170.1, 170.2, 170.3, 170.5, 170.6, 170.7. Anal. Calcd for
C30H17O17Si: C, 50.84; H, 6.83. Found: C, 50.72; H, 6.64.
(17) In contrast, benzyl-protected oligosaccharides loaded on
acetate-capped Boltorn polyesters at comparable loading
levels were completely insoluble in MeOH (ref. 9). Such
behavior is in agreement with the notion that the terminal
groups largely determine the solubility properties of
dendrimers and hyperbranched polymers.
(18) (a) In addition to losses during precipitation into MeOH,
cleavage of the o-nitrobenzyl linker with UV light generally
suffers from low efficiency. See: Holmes, C. P. J. Org.
Chem. 1997, 62, 2370. (b) The disaccharide yield is
comparable with yields obtained previously utilizing the UV
light cleavage protocol (ref. 9).
(15) Zhong, J.-H.; Fishman, A.; Lee-Ruff, E. Org. Lett. 2002, 4,
4415.
(16) Representative Experimental Procedures.
Compound 17. The glycosyl acceptor 5 (25% loading
capacity, 1.36 mmol/g; 0.33 g, 0.45 mmol, 100 mol%) and
the glycosyl donor 12 (1.27 g, 2.25 mmol, 500 mol%) were
dissolved in dry THF (4.5 mL). After a homogeneous
solution had formed, BF3·OEt2 (90 mL, 102 mg, 0.72 mmol,
(19) Barresi, F.; Hindsgaul, O. Can. J. Chem. 1994, 72, 1447.
Synlett 2005, No. 10, 1567–1570 © Thieme Stuttgart · New York