Automated Assembly of Starch and Glycogen Polysaccharides

Article abstract of DOI:10.1021/jacs.1c02188

Polysaccharides are Nature's most abundant biomaterials essential for plant cell wall construction and energy storage. Seemingly minor structural differences result in entirely different functions: cellulose, a β (1-4) linked glucose polymer, forms fibrils that can support large trees, while amylose, an α (1-4) linked glucose polymer forms soft hollow fibers used for energy storage. A detailed understanding of polysaccharide structures requires pure materials that cannot be isolated from natural sources. Automated Glycan Assembly provides quick access to trans-linked glycans analogues of cellulose, but the stereoselective installation of multiple cis-glycosidic linkages present in amylose has not been possible to date. Here, we identify thioglycoside building blocks with different protecting group patterns that, in concert with temperature and solvent control, achieve excellent stereoselectivity during the synthesis of linear and branched α-glucan polymers with up to 20 cis-glycosidic linkages. The molecules prepared with the new method will serve as probes to understand the biosynthesis and the structure of α-glucans.

Full text of DOI:10.1021/jacs.1c02188

pubs.acs.org/JACS
Article
Automated Assembly of Starch and Glycogen Polysaccharides
Yuntao Zhu, Martina Delbianco, and Peter H. Seeberger*
Cite This: J. Am. Chem. Soc. 2021, 143, 9758−9768
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Polysaccharides are Nature’s most abundant biomate-
rials essential for plant cell wall construction and energy storage.
Seemingly minor structural differences result in entirely different
functions: cellulose, a β (1−4) linked glucose polymer, forms fibrils
that can support large trees, while amylose, an α (1−4) linked glucose
polymer forms soft hollow fibers used for energy storage. A detailed
understanding of polysaccharide structures requires pure materials that
cannot be isolated from natural sources. Automated Glycan Assembly
provides quick access to trans-linked glycans analogues of cellulose, but
the stereoselective installation of multiple cis-glycosidic linkages present
in amylose has not been possible to date. Here, we identify
thioglycoside building blocks with different protecting group patterns
that, in concert with temperature and solvent control, achieve excellent
stereoselectivity during the synthesis of linear and branched α-glucan polymers with up to 20 cis-glycosidic linkages. The molecules
prepared with the new method will serve as probes to understand the biosynthesis and the structure of α-glucans.
methods to prepare cis-glycosides involving remote assistance,
the use of chiral auxiliaries, or catalysts as well as substrate
control combined with solvent and temperature effects has
been developed (Figure 1B).16−34 Still, the synthesis of
oligosaccharides with multiple cis-glycosidic linkages has been
a major challenge to manual syntheses that relied on block
couplings to prepare amylose and amylopectin sequences as
long as 10-mers.7,35,36
Automated Glycan Assembly (AGA) has been used to
prepare polysaccharides as large as 151-mers,37 a variety of
biologically important complex glycans containing different
monomers as well as natural and unnatural carbohydrate
materials.5,38−45 AGA requires high yielding and completely
selective glycosylation reactions as none of the intermediates,
but only the products of the synthesis are purified.46 The
separation of full-length oligosaccharide side-products that
differ from the desired product only in the stereochemistry of
one linkage is very difficult if not impossible. Therefore, very
few oligosaccharides containing cis-glycosidic linkages have
been prepared by AGA, exploiting building blocks with remote
ester groups in the 3- and 6-positions.47−49
INTRODUCTION
■
Polysaccharides are built from monosaccharide units con-
nected through glycosidic bonds with different regio- and
stereochemistry. Glucose-based polymers are the most
abundant biomaterials on earth, but differ greatly in structure
and function depending on whether they are β (1−4) linked as
in cellulose, or α (1−4) linked as in starch and glycogen
(Figure 1A). Cellulose (1) is a linear polymer that unlike
starch does not coil or branch, but adopts an extended and stiff
rodlike conformation.1,2 Hydrogen bonding between parallel
chains form microfibrils with high tensile strength that are part
of the polysaccharide matrix in the cell wall.1−5 Starch is a
mixture of about one-quarter strictly linear amylose (2) and
about three-quarters α (1−6) branched amylopectin (3).6,7
Glycogen (4), the polymer that stores energy in the form of
glucose in animals, is highly branched.8 Amylose exists in a
disordered amorphous conformation or can adopt two
different helical conformations that can host other mole-
cules.6,7,9−12 Amylopectin is linear and does not crystallize as
well as the long linear chains of amylose.6,7,10,13
The molecular level understanding of cellulose structure
formation has benefitted from synthetic glycans and single
molecule glycan imaging.14,15 In order to understand the
factors contributing to the structure of amylose, amylopectin,
and glycogen, pure oligo- and polysaccharides containing
exclusively cis-glycosidic linkages are needed. The stereo-
controlled synthesis of linear or branched oligosaccharides has
proven extremely challenging since glycosidic bond formation
cannot rely on participating neighboring protecting groups as
for trans-glycoside construction (Figure 1B). A host of
Here, we describe a simple strategy based on thioglycoside
building blocks with protecting group patterns that, together
Received: February 25, 2021
Published: June 11, 2021
© 2021 The Authors. Published by
American Chemical Society
https://doi.org/10.1021/jacs.1c02188
J. Am. Chem. Soc. 2021, 143, 9758−9768
9758

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 1. Chemical structure, synthetic strategies, and Automated Glycan Assembly of starch and glycogen. (A) Chemical structure of β (1−4)
linked polyglucoside cellulose (1), linear α (1−4) linked polyglucoside amylose (2), α (1−4) linked polyglucoside with α (1−6) branched α (1−4)
linked polyglucoside side chains, amylopectin (3), and highly branched glycogen (4). (B) Participating neighboring protecting group ensures trans-
glycoside construction, stereocontrolled cis-glycosidic bond formation is challenging: uncontrolled glycosylations result in anomeric mixtures.
Typical solutions: Solvent assistance,19 bulky protecting group assistance,50 ester remote assistance,26 chiral-auxiliary-mediation,18 macrocyclic bis-
thiourea catalysis.27 Strategy developed in this work for α (1−4) glucan AGA with simple chemistry. (C) Schematic view of Automated Glycan
Assembly and the building blocks used to prepare cis-linked polyglucosides. (D) Synthetic amylose, and amylopectin oligo- and polysaccharides
(5−11) prepared by AGA.
9759
https://doi.org/10.1021/jacs.1c02188
J. Am. Chem. Soc. 2021, 143, 9758−9768

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
a
Table 1. Glucose Building Blocks and Solution Phase Glycosylations
entry
building block
acceptor
ratio (α/β)
yield (α + β)
entry
building block
acceptor
ratio (α/β)
yield (α + β)
1
2
3
4
5
6
7
8
12
13
17
20
22
23
25
26
27
28
29
30
31
32
i-PrOH
i-PrOH
i-PrOH
i-PrOH
i-PrOH
i-PrOH
i-PrOH
i-PrOH
i-PrOH
i-PrOH
i-PrOH
i-PrOH
i-PrOH
i-PrOH
1:2.9
2.3:1
2.5:1
1.7:1
2.2:1
3.5:1
1.9:1
4.5:1
6.4:1
4.8:1
7.1:1
2.4:1
1:1.1
1:2.6
93%
95%
92%
93%
90%
89%
90%
84%
88%
87%
74%
76%
96%
94%
15
16
17
18
19
20
21
22
23
24
25
26
27
28
12
13
17
20
22
23
25
26
27
28
29
30
31
32
33
33
33
33
33
33
33
33
33
33
33
33
33
33
1.5:1
3.8:1
5.5:1
6.8:1
5.6:1
6.2:1
7.5:1
9.8:1
>10:1
>10:1
>10:1
>10:1
3.9:1
2.7:1
83%
75%
69%
72%
68%
67%
44%
78%
62%
83%
64%
80%
75%
81%
9
10
11
12
13
14
a
(A) The glucose thioglycoside building blocks were prepared from a common precursor as described in Figure S1. (B) Glycosylations using
isopropanol and monosaccharide as nucleophiles. The ratio of α/β anomers were quantified by NMR, yield represented the isolated yield of α- and
β-anomers (complete results in Table S1).
with temperature and solvent control, allows for excellent cis-
selectivity during AGA of α-glucans (Figure 1C). A series of
glucans resembling linear or branched natural starch structures
with an α (1−4) backbone and α (1−6) side chains were
prepared to illustrate the power of the approach (Figure 1D).
Branched 20-mer α-glucan 10, the largest full cis-linked
branched glycan ever made, was obtained within 50 h AGA
time.
during glycosylation (Figure 1B). Remote assistance of a 6-
acetyl (Ac) protecting group is not sufficient to prepare α (1−
4) glucan oligosaccharides with high cis-selectivity and good
isolated yield.49 In search for building blocks that maximize cis-
selectivity and yield during AGA, a comprehensive evaluation
of different C3 and C6 ester groups and their influence on α
(1−4) glucan formation was undertaken (see Table S1). An
electron rich, nonparticipating benzyl (Bn) ether group was
selected as a permanent protecting group on the C2 hydroxyl
group to allow for cis-glycoside formation with good reactivity.
Thioglycosides 13−25 carry a C3 benzyl ether as well as
different esters at C6, while building blocks 26−30 carry ester
protection groups both on C3 and C6 (see Table 1A).
Additional building blocks with ester at C3 but benzyl
protection at C6-OH were also shown (Table S1, compound
110−112) as comparison. Thioglycosides 12 and 31−32,
substituted with a fluorine or an ether at C6, respectively, were
RESULTS AND DISCUSSION
■
Building Block Design. Thioglycosides with temporary
fluorenylmethoxycarbonyl (Fmoc) protecting groups have
proven as useful building blocks for AGA (Table 1A).40
Thus, Fmoc protection to mask the C4 hydroxyl group
required for chain elongation was selected. The orientation of
C3 and C6 functional groups on the glucopyranose ring makes
them suitable remote assistant groups to block the β-face
9760
https://doi.org/10.1021/jacs.1c02188
J. Am. Chem. Soc. 2021, 143, 9758−9768

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Table 2. Screening of Glycosylation Conditions by AGA
entry
building block
R1
R2
product
selectivity
yield
1
2
3
4
5
6
13
17
23
27
28
29
OAc
OPiv
O(p-NO2Bz)
OPiv
OBz
OBn
OBn
OBn
OPiv
OBz
36
37
38
39
40
41
0.35
0.93
0.68
40%
4%
14%
3%
17%
19%
a
>0.95
0.74
a
O(p-NO2Bz)
O(p-NO2Bz)
0.45
a
The p-nitrobenzoyl group present in building blocks 23 and 29 is not stable to UV irradiation. Therefore, the data was calculated based on
deprotected products.
prepared for comparison. All building blocks were prepared
from a single precursor using stratified methods in order to
minimize experimental work (see Figure S1).
disubstituted building blocks 26−30 were tested (entries 8−
12). All of these compounds improved cis-glycoside formation.
With initial information concerning the selectivity of
different thioglycoside building blocks in hand, we next
investigated the influence of the reaction temperature on the
product ratio (Table S2). Reactions at a constant temperature
revealed that glycosylations at higher temperatures gave better
stereoselectivity, as previously reported.28 However, activation
temperatures above 0 °C may result in lower yields (Table S2,
entry 3 vs 4). Solvent is another important factor to control the
stereochemistry during glycosylation. Ethers are the most
common alpha directing solvents.16,19,28,49 We tested our
system in several different ethers (Table S3, entries 1−5).
Although most of the selected ethers were helpful for
producing cis-glycoside, the poor solubility of our building
blocks and NIS makes these solvents not compatible with
AGA. Dioxane is a good choice for use in the AGA system, but
a balance between the solubility of the activator and the
melting point of the reaction mixture needs to be maintained.
For AGA we dissolved the activator in a DCM−dioxane
mixture (2:1), and mixed it with the same volume of building
block DCM solution (DCM−dioxane 5:1 during reaction).39
We tested several typical building blocks in a 5:1 DCM−
dioxane mixture and proved that this condition can be used to
increase the cis-product formation(Table S3, entries 1 and 6−
14).
Next, the influence of the nucleophile on the outcome of the
glycosylation was explored, by analyzing the glycosylation of
the secondary hydroxyl group at the C4 position of
monosaccharide 33 (entries 15−28).55 EWG and ester remote
assistance contribute both to the formation of cis-linked
disaccharide. Higher cis-selectivity was obtained with the more
sterically hindered monosaccharide-acceptor compared to
isopropanol, in particular when building blocks with bulky
pivaloyl (Piv) 17 (entry 3 vs 17), triphenylacetyl 25 (entry 7 vs
21) protecting groups or disubstituted thioglycosides 27−30
(entries 9−12 vs 23−26) were used. For sterically more
demanding nucleophiles, the protecting group pattern on the
glycosylating agent might have an important impact on the
stereochemical outcome of the glycosylation. Bulky protecting
Solution-Phase Glycosylation. The differentially pro-
tected thioglycoside building blocks were used to glycosylate
different nucleophiles to screen the influence of protecting
groups on α-selectivity (Table 1B and Table S1). Initially, all
building blocks were activated with N-iodosuccinimide
(NIS)−triflic acid (TfOH) in DCM for 5 min at −15 °C
followed by 60 min at 0 °C to glycosylate isopropanol (i-
PrOH) as a model for a secondary hydroxyl group.
Glycosylations involving 12−25 (entries 1−7 in Table 1B,
and entries 1−14 in Table S1) as well as 31−32 (entries 13
and 14) suggest a relationship between electronic and steric
properties of the C6 protecting groups and the α/β ratio of the
glycoside products. An electron withdrawing group (EWG) on
C6 improves cis-selectivity. Thioglycosides 12 (entry 1) and 32
(entry 14) that carry nonparticipating electron donating
groups (EDGs), are less α-selective than building blocks 13−
25 (entries 2−7, and entries 2−14 in Table S1) and 31 (entry
13). Esterification on 6-OH(entries 2−7, and entries 2−14 in
Table S1) resulted in higher stereoselectivity than deoxyfluori-
nation (entry 13), even though ester groups are less electron
withdrawing than fluorine. These results further supported the
ester remote assistance hypothesis during glycosylation.16,26,49
The argument whether the ester group really stabilized the
intermediate by forming a covalent glycosyl dioxolenium ion
during glycosylations has not been settled, although this
intermediate can be detected under extreme conditions.51−54
We observed that 6-OH esterification alone does not ensure
excellent selectivity during α (1−4) glycosylations. Selectivities
for the glycosylation of isopropanol ranged from 1.5−3:1 (α/
β) as changes in the electronic and steric properties of ester
groups had little effect (entries 2−7, and entries 2−14 in Table
S1). Additional experiments employing 3-OH esterified, 6-OH
benzylated building blocks (entries 22−24 in Table S1) agreed
with our previous findings that C3 esters are an EWG that can
provide remote assistance, although usually less helpful than
C6 esters.49 To increase the selectivity further, 3,6-
9761
https://doi.org/10.1021/jacs.1c02188
J. Am. Chem. Soc. 2021, 143, 9758−9768

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 2. AGA of amylose, amylopectin, and glycogen α-glucans. (A) AGA of amylose oligosaccharides 5−7. (B) AGA of amylopectin
polysaccharides 8−10. (C) AGA of glycogen oligosaccharide 11.
groups had been used in the solution phase synthesis to
introduce cis-linked glucose residues, and an α-glucan decamer
was prepared recently.50,56 Overall, selectivities exceeding 10:1
(α/β) were achieved when all factors favoring cis-glycoside
formation were combined using building blocks 27−30
carrying 3,6-diesters for AGA.
Glycosylations on Solid Support. Using the optimized
solution-phase conditions, glycosylations were carried out on a
Merrifield resin functionalized with the 5-aminopentanol
photolabile linker (35) (Table 2). The coupling cycle involved
an acidic wash with trimethylsilyl trifluoromethanesulfonate
(TMSOTf), followed by NIS-TfOH induced glycosylation
(−20 °C for 5 min → 0 °C for 60 min), capping of any
unreacted nucleophile with acetic anhydride (Ac2O) and
methanesulfonic acid (MsOH), and cleavage of the temporary
Fmoc protective group with triethylamine (TEA). After
completion of each synthesis, the product was released from
the resin by irradiation with UV light. The α (1−4) glucose
pentasaccharide containing a β-glycosidic linkage between the
first unit and the amino pentanol linker served as a model to
study the cis-selectivity on both C4 secondary and C6 primary
hydroxyl groups.49
The products 36−41 were characterized by analytical HPLC
and were purified in protected form to remove any deletion
sequences. All pentamer products were collected. The isolated
yield of the combined pentamers is reported independently of
the stereochemistry of the linkages. The selectivity is calculated
as the ratio of the amount of desired product containing four
α-linkages to all pentamer products, as measured by NMR
spectroscopy (for details, see Supporting Information). For
protected glucans in this study, the chemical shift of α-linked
anomeric proton is usually > 4.8 ppm and anomeric carbon <
100 ppm; for β-linked anomeric proton, the chemical shift is
usually < 4.7 ppm and anomeric carbon > 100 ppm.
In search for a suitable building block for α-glucan AGA,
both yield and cis-selectivity were taken into consideration.
The results for thioglycoside 13 were in good agreement with
our earlier findings (Table 2, entry 1).49 Despite the promising
9762
https://doi.org/10.1021/jacs.1c02188
J. Am. Chem. Soc. 2021, 143, 9758−9768

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 3. HPLC analysis of protected 16-mer 45, 20-mer 48, and 14-mer 49. Left side, crude mixture after AGA; right side, target compound after
purification. (A) 16-mer 45; (B) 20-mer 48; (C) 14-mer 49. Signal was collected through ELSD.
yield obtained with monoacetylated 13, only 35% of the
pentamer products had the desired stereochemistry (entry 1).
The pivaloyl group at 6-OH (17 and 27) resulted in good cis-
selectivity on solid phase (entries 2 and 4). However, bulky
esters at C3 or C6 lowered the reactivity of 17 and 27 both as
glycosylating agents and nucleophiles to render them
unsuitable for the assembly of longer oligosaccharides. Building
blocks 23 and 28 showed similar yields and cis-selectivity
(entries 3 and 5). The trans-linked side-products arose mainly
from the glycosylation of primary hydroxyl group of 34. Di(p-
nitrobenzoyl) protected building block 29 showed a slightly
lower cis-selectivity (entry 6). Additionally, the p-nitrobenzoyl
group was not stable under UV irradiation and complicated
purification and characterization of the protected oligomers.
3,6-Dibenzoylated thioglycoside building block 28 emerged as
the best choice considering selectivity and yield (entry 5).
Similar good cis-selectivity was observed recently with
galactose building blocks bearing multiple benzoyl (Bz)
groups.57
AGA of Starch and Glycogen α-Glucans. On the basis
of the preliminary AGA results described above, the reaction
conditions were optimized in the context of the synthesis of a
series of α-glucans (Figure 2). The 3,6-di-O-benzyolated
thioglycoside 28 carrying a temporary 4-OFmoc protecting
group was employed to assemble the α (1−4) glycan
backbone. α (1−6) Glycan branching was installed using
building block 30 with benzoate ester at C3 and a 4-OFmoc
and a 6-OLev temporary protecting groups. A double
glycosylation cycle (−20 °C for 5 min → 0 °C for 60 min,
twice) with thioglycoside 28 was used for each elongation to
increase yields and minimize the formation of deletion
sequences. Traceless linker (42)42 was cleaved by photolysis
9763
https://doi.org/10.1021/jacs.1c02188
J. Am. Chem. Soc. 2021, 143, 9758−9768

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
1
Figure 4. Comparison of H NMR spectra of natural starch (a), glycogen (b), glucose (c), and synthetic α-glucans (d−j).
and removed after global deprotection to release the natural
starch/glycogen α-glucan. Linear amylose oligomer α (1−4)
glucan, including tetramer 5, octamer 6, and 16-mer 7, were
prepared with good yield and selectivity (Figure 2A, see
Supporting Information for full characterization). The HPLC
analysis of the crude protected 16-mer 45 after AGA displayed
the (partially capped) deletion sequences as the record of
glycan elongation (Figure 3A). No trans-glycosylic linkages
were observed by NMR spectroscopy, indicating that even the
first glycosylation between 28 and linker 42 proceeded with
excellent cis-selectivity.
To extend the AGA scope to natural amylopectin and
glycogen structures, the selective installation of α (1−6)
branching points within the α (1−4) backbone had to be
developed. Glycosylation of the C6 primary hydroxyl group of
34 using building block 28 did not result in satisfactory cis-
selectivity (Table 2). Two synthetic strategies were tested in
the context of the syntheses of 8−11, as the order of Fmoc and
Lev cleavage may influence yield and selectivity. In Strategy A,
after incorporation of building block 30, the Fmoc group was
cleaved and one glycosylation was performed in the presence
of the levaloyl group. Then, Lev was cleaved and the C6
9764
https://doi.org/10.1021/jacs.1c02188
J. Am. Chem. Soc. 2021, 143, 9758−9768

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 5. I2−KI staining of glucose, natural starch/glycogen and synthetic α-glucans.
hydroxyl group was glycosylated (Figure S2A). After the
branching point was introduced, α (1−4) elongation of
backbone and side chain was achieved simultaneously.
Tetramer 8 and heptamer 9 were prepared using this strategy
(Figure 2B). Protected heptamer 47 was fully analyzed using
HPLC and HSQC NMR to evaluate the cis-selectivity of
Strategy A. After AGA, the crude product was analyzed with
HPLC and all the detectable side-products were assigned
according to MALDI-TOF MS (Figure S2B). Interestingly,
most of the deletion sequences were capped by benzoyl groups
that may result from the acetylated deletion sequences
(capping step) through the ester exchange reaction. The MS
signal of the desired heptamer was only observed for the main
peak (red square) and its surrounding area, together with the
signal for some deletion sequences (black square). The main
product proved to be the desired fully cis-linked heptamer 47
by HSQC NMR (Figure S2C). The mixture of deletion
sequences was also analyzed by HSQC NMR showing no
evidence of β linkage (Figure S2D). These results strongly
suggest that Strategy A is a highly selective method for the
installation of α (1−6) linkages in our synthesis. Glycosylation
at C4-OH may increase the glycan bulk for the subsequent
glycosylation of the C6 hydroxyl group, thereby promoting cis-
selectivity.
Strategy B was tested for trimer 50 (Figure S3) which not
only allowed for the formation of different patterns during
glycan chain elongation, but also gave a much higher yield
compared to Strategy A and greatly simplified the synthesis of
longer amylopectins. After glycosylation with building block
30, cleavage of the levulinic ester was followed by glycosylation
on C6-OH in the presence of 4-OFmoc (Figure S3A). After
Fmoc cleavage, the C4 hydroxyl groups were glycosylated.
Only one main fraction was observed by HPLC analysis of the
crude product (Figure S3B) that proved to be the desired cis-
linked trimer by HSQC NMR analysis (Figure S3C). The
higher selectivity observed for this reaction, compared to the
glycosylation of 34, could result from the electron withdrawing
4-OFmoc in 30, instead of an electron donating benzyl ether in
34 that renders it less nucleophilic.
chain on large glycan may not be as efficient using the standard
glycosylation condition.
The assembly of a more challenging, hyperbranched
glycogen model 11 was undertaken. Up to the secondary
branching points, the assembly followed the procedures
established above. Two secondary branching points was
introduced with one double glycosylation, but in order to
add four glucose units in one pot at the end, two double
glycosylation cycles were required. The HPLC analysis of the
crude product after AGA indicated that the deletion sequences
were mainly caused by incomplete multiple-point glycosylation
(Figure 3C).
Since the end-stage deletion sequences partially overlap with
the product peak, three HPLC purifications were required for
48 and 49 (see Supporting Information, Method B3 in AGA
part). The first purification yielded pure fraction (I) (purity
>90% indicated by MS) and fraction (II) (purity >50%). The
second purification of fraction (II) yielded fraction (III)
(purity >90%), before the third purification of combined
fractions (I) and (III) gave the product in high purity. After
deprotection, both 10 and 11 were obtained in good yield and
stereochemical purity, as confirmed by HSQC NMR (see
Supporting Information for full characterization). The entire
AGA process to access the protected 20-mer 48 took about 50
h and for protected 14-mer 49 took only 30 h.
Comparison of Synthetic α-Glucans and Natural
Starch/Glycogen. The properties of the synthetic α-glucan
were compared to natural potato starch and glycogen from
bovine liver. All characteristic signals in the NMR spectra of
the synthetic glucans exactly match those of the natural
compounds (Figure 4). Except for glucose, all compounds
show signals from 5.30 to 5.50 ppm that represent the α (1−4)
anomeric protons of the essential starch/glycogen linkages.
Signals at 5.24, 4.67, and 3.29 ppm represent the α, β anomeric
protons and the β H2 at the reducing end that are not
observed in natural starch and glycogen. The signals at 4.99−
4.96 ppm in starch and glycogen represent the α (1−6)
anomeric protons of the glycan branching points. The signals
at 3.45−3.39 ppm represent the C4 protons of the non-
reducing end. These signals are indicative of the “degree of
branching” of starch/glycogen polysaccharides with strong
signals (compared to the signals of reducing end protons) for
glycogen and 14-mer 11. In contrast to the broad signals for
heterogeneous natural polysaccharides, all homogeneous
synthetic compounds show sharp NMR signals. Our data
Amylopectin 20-mer 10 was assembled using Strategy B
(Figure 2B). After HPLC purification (Figure 3B), we
obtained the protected 20-mer in 24% yield. The major side
products are (partially capped) deletion sequences formed
during chain extension after the α (1−6) branching point was
installed. The simultaneous elongation of backbone and side
9765
https://doi.org/10.1021/jacs.1c02188
J. Am. Chem. Soc. 2021, 143, 9758−9768

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Authors
highlight the importance of synthetic compounds as standard
to characterize naturally sourced compounds.
Yuntao Zhu − Max Planck Institute for Colloids and
Interfaces, 14476 Potsdam, Germany; orcid.org/0000-
0002-6732-6529
Martina Delbianco − Max Planck Institute for Colloids and
Interfaces, 14476 Potsdam, Germany; orcid.org/0000-
0002-4580-9597
The iodine−starch test is a simple, but characteristic
detection method for α (1−4) glucans.9 Although the details
of this interaction was not entirely understood, an iodine−
iodide complex is known to embed into the helix formed by α
(1−4) glucan.58,59 With increasing glucan chain length, more
complex is bound, causing a color change from yellow, orange,
red-brown, to purple and dark blue. Amylose tetramer 5,
amylopectin tetramer 8, and heptamer 9 show a light yellow
color similar to glucose. Amylose octamer 6, amylopectin 20-
mer 10, and glycogen 14-mer 11 result in a deep yellow color,
indicative of weak interactions between iodide complex and
glucan. Amylose 16-mer 7 gave a red-brown color similar to
glycogen indicating that the α (1−4) glucan chain has the
required length to strongly interact with the iodine−iodide
complex. The long α (1−4) glucan backbone of natural starch
is poorly soluble at room temperature and has a very strong
bonding with the complex, indicated by the dark blue color
(Figure 5). Our results match literature reports that six
continuous α (1−4) glucose residues are needed to form the
repeating unit of helical structure, which can assemble with
iodine complex nicely.60,61
Complete contact information is available at:
https://pubs.acs.org/10.1021/jacs.1c02188
Funding
This work is supported by the Max Planck Society.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank the Max Planck Society financial support. We thank
Sabrina Leichnitz and Giulio Fittolani for the help in collection
of HRMS data, Theodore Tyrikos-Ergas and Zhouxiang Zhao
for the help during resin preparation, and Elena Shanina for the
help in collection of NMR spectrum for amylopectin 20-mer.
ABBREVIATIONS
■
CONCLUSIONS
■
AGA, automated glycan assembly; Ac, acetyl; Fmoc,
fluorenylmethoxycarbonyl; Bn, benzyl; Bz, benzoyl; NIS, N-
iodosuccinimide; TfOH, triflic acid; DCM, dichloromethane;
i-PrOH, isopropanol; EWG, electron withdrawing group;
EDG, electron donating group; Piv, pivaloyl; TMSOTf,
trimethylsilyl trifluoromethanesulfonate; Ac2O, acetic anhy-
dride; MsOH, methanesulfonic acid; TEA, trimethylamine;
Lev, levaloyl; HPLC, high-performance liquid chromatogra-
phy; HRMS, high-resolution mass spectrometry; MALDI-TOF
MS, matrix-assisted laser desorption ionization−time-of-flight
mass spectrometry; ELSD, vaporative light scattering detector
On the basis of a comprehensive study of 21 differentially
protected thioglycoside building blocks, glycosylation con-
ditions that result in high yielding cis-selective glycosidic bond
formation were developed. With amylose and amylopectin
oligo- and polysaccharides as targets, 3,6-dibenzoylated glucose
thioglycoside 28 provided the perfect balance between cis-
selectivity and reactivity during AGA required to prepare long
α (1−4) glycans. Differentially protected thioglycoside 30 was
the key to α (1−6) branching with excellent cis-selectivity and
good yield using AGA. The 20-mer amylopectin polysacchar-
ide 10 is the largest entirely cis-linked branched carbohydrate
assembled by chemical synthesis to date. NMR studies and
iodine-stain tests, confirmed that the synthetic glucans share
common structural properties with natural starch and
glycogen. This simple and efficient AGA method provides
convenient access to large, well-defined cis-linked starch and
glycogen polysaccharides for biological and material science
investigations. Future improvements on the instrumentation
hardware and the chemistry will shorten assembly times.
Increased glycosylation efficiencies will save building blocks,
reduce deletion sequences, and facilitate HPLC purification.
REFERENCES
■
(1) O’Sullivan, A. C. Cellulose: the structure slowly unravels.
Cellulose 1997, 4 (3), 173−207.
(2) Habibi, Y.; Lucia, L. A.; Rojas, O. J. Cellulose nanocrystals:
chemistry, self-assembly, and applications. Chem. Rev. 2010, 110 (6),
3479−3500.
(3) Nishiyama, Y.; Langan, P.; Chanzy, H. Crystal structure and
hydrogen-bonding system in cellulose Iβ from synchrotron X-ray and
neutron fiber diffraction. J. Am. Chem. Soc. 2002, 124 (31), 9074−
9082.
(4) Cosgrove, D. J. Growth of the plant cell wall. Nat. Rev. Mol. Cell
Biol. 2005, 6 (11), 850.
(5) Yu, Y.; Tyrikos-Ergas, T.; Zhu, Y.; Fittolani, G.; Bordoni, V.;
ASSOCIATED CONTENT
■
Singhal, A.; Fair, R. J.; Grafmuller, A.; Seeberger, P. H.; Delbianco, M.
̈
sı
* Supporting Information
Systematic hydrogen-bond manipulations to establish polysaccharide
structure−property correlations. Angew. Chem. 2019, 131 (37),
13261−13266.
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/jacs.1c02188.
(6) Buleon, A.; Colonna, P.; Planchot, V.; Ball, S. Starch granules:
structure and biosynthesis. Int. J. Biol. Macromol. 1998, 23 (2), 85−
112.
Experimental procedures, additional data, and spectral
data for all reactions and compounds (PDF)
(7) Damager, I.; Engelsen, S. B.; Blennow, A.; Lindberg Møller, B.;
Motawia, M. S. First principles insight into the α-glucan structures of
starch: their synthesis, conformation, and hydration. Chem. Rev. 2010,
110 (4), 2049−2080.
(8) Manners, D. J. Recent developments in our understanding of
glycogen structure. Carbohydr. Polym. 1991, 16 (1), 37−82.
(9) Bates, F. L.; French, D.; Rundle, R. Amylose and amylopectin
content of starches determined by their iodine complex formation. J.
Am. Chem. Soc. 1943, 65 (2), 142−148.
AUTHOR INFORMATION
■
Corresponding Author
Peter H. Seeberger − Max Planck Institute for Colloids and
Interfaces, 14476 Potsdam, Germany; Institute for Chemistry
and Biochemistry, Freie Universität Berlin, 14195 Berlin,
Germany; orcid.org/0000-0003-3394-8466;
Email: peter.seeberger@mpikg.mpg.de
9766
https://doi.org/10.1021/jacs.1c02188
J. Am. Chem. Soc. 2021, 143, 9758−9768

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
(10) Tester, R. F.; Karkalas, J.; Qi, X. Starchcomposition, fine
structure and architecture. J. Cereal Sci. 2004, 39 (2), 151−165.
(11) Popov, D.; Buléon, A.; Burghammer, M.; Chanzy, H.;
Montesanti, N.; Putaux, J.-L.; Potocki-Veronese, G.; Riekel, C.
Crystal structure of A-amylose: A revisit from synchrotron micro-
diffraction analysis of single crystals. Macromolecules 2009, 42 (4),
1167−1174.
(12) Putseys, J.; Lamberts, L.; Delcour, J. Amylose-inclusion
complexes: Formation, identity and physico-chemical properties. J.
Cereal Sci. 2010, 51 (3), 238−247.
(29) Wang, L.; Overkleeft, H. S.; van der Marel, G. A.; Codée, J. D.
Reagent controlled stereoselective synthesis of α-glucans. J. Am. Chem.
Soc. 2018, 140 (13), 4632−4638.
(30) van der Vorm, S.; van Hengst, J. M.; Bakker, M.; Overkleeft, H.
S.; van der Marel, G. A.; Codée, J. D. Mapping the relationship
between glycosyl acceptor reactivity and glycosylation stereo-
selectivity. Angew. Chem. 2018, 130 (27), 8372−8376.
(31) Yao, H.; Vu, M. D.; Liu, X.-W. Recent advances in reagent-
controlled stereoselective/stereospecific glycosylation. Carbohydr. Res.
2019, 473, 72−81.
(13) Jane, J.; Chen, Y.; Lee, L.; McPherson, A.; Wong, K.;
Radosavljevic, M.; Kasemsuwan, T. Effects of amylopectin branch
chain length and amylose content on the gelatinization and pasting
properties of starch. Cereal Chem. 1999, 76 (5), 629−637.
(14) Wu, X.; Delbianco, M.; Anggara, K.; Michnowicz, T.; Pardo-
Vargas, A.; Bharate, P.; Sen, S.; Pristl, M.; Rauschenbach, S.;
Schlickum, U.; Abb, S.; Seeberger, P. H.; Kern, K. Imaging single
glycans. Nature 2020, 582 (7812), 375−378.
(15) Anggara, K.; Zhu, Y.; Delbianco, M.; Rauschenbach, S.; Abb, S.;
Seeberger, P. H.; Kern, K. Exploring the molecular conformation
space by soft molecule−surface collision. J. Am. Chem. Soc. 2020, 142
(51), 21420−21427.
(16) Demchenko, A. V.; Rousson, E.; Boons, G.-J. Stereoselective 1,
2-cis-galactosylation assisted by remote neighboring group partic-
ipation and solvent effects. Tetrahedron Lett. 1999, 40 (36), 6523−
6526.
(17) Demchenko, A. V. 1, 2-cis-O-Glycosylation: methods, strategies,
principles. Curr. Org. Chem. 2003, 7 (1), 35−79.
(32) Hettikankanamalage, A. A.; Lassfolk, R.; Ekholm, F. S.; Leino,
R.; Crich, D. Mechanisms of stereodirecting participation and ester
migration from near and far in glycosylation and related reactions.
Chem. Rev. 2020, 120 (15), 7104−7151.
(33) Li, Q.; Levi, S. M.; Jacobsen, E. N. Highly selective β-
mannosylations and β-rhamnosylations catalyzed by bis-thiourea. J.
Am. Chem. Soc. 2020, 142 (27), 11865−11872.
(34) Mayfield, A. B.; Metternich, J. B.; Trotta, A. H.; Jacobsen, E. N.
Stereospecific furanosylations catalyzed by bis-thiourea hydrogen-
bond donors. J. Am. Chem. Soc. 2020, 142 (8), 4061−4069.
(35) Damager, I.; Olsen, C. E.; Møller, B. L.; Motawia, M. S.
Chemical synthesis of 6‴-α-maltotriosyl-maltohexaose as substrate for
enzymes in starch biosynthesis and degradation. Carbohydr. Res. 1999,
320 (1−2), 19−30.
(36) Damager, I.; Jensen, M. T.; Olsen, C. E.; Blennow, A.; Møller,
B. L.; Svensson, B.; Motawia, M. S. Chemical synthesis of a dual
branched malto-decaose: A potential substrate for α-amylases.
ChemBioChem 2005, 6 (7), 1224−1233.
(37) Joseph, A. A.; Pardo-Vargas, A.; Seeberger, P. H. Total
synthesis of polysaccharides by automated glycan assembly. J. Am.
Chem. Soc. 2020, 142 (19), 8561−8564.
(38) Plante, O. J.; Palmacci, E. R.; Seeberger, P. H. Automated
Solid-Phase Synthesis of Oligosaccharides. Science 2001, 291 (5508),
1523−1527.
(18) Boltje, T. J.; Kim, J.-H.; Park, J.; Boons, G.-J. Chiral-auxiliary-
mediated 1, 2-cis-glycosylations for the solid-supported synthesis of a
biologically important branched α-glucan. Nat. Chem. 2010, 2 (7),
552−557.
(19) Satoh, H.; Hansen, H. S.; Manabe, S.; van Gunsteren, W. F.;
Hunenberger, P. H. Theoretical investigation of solvent effects on
̈
(39) Delbianco, M.; Kononov, A.; Poveda, A.; Yu, Y.; Diercks, T.;
Jiménez-Barbero, J.; Seeberger, P. H. Well-defined oligo-and
polysaccharides as ideal probes for structural studies. J. Am. Chem.
Soc. 2018, 140 (16), 5421−5426.
(40) Guberman, M. n.; Seeberger, P. H. Automated glycan assembly:
A perspective. J. Am. Chem. Soc. 2019, 141 (14), 5581−5592.
(41) Guberman, M.; Bräutigam, M.; Seeberger, P. H. Automated
glycan assembly of Lewis type I and II oligosaccharide antigens. Chem.
Sci. 2019, 10 (21), 5634−5640.
glycosylation reactions: stereoselectivity controlled by preferential
conformations of the intermediate oxacarbenium-counterion complex.
J. Chem. Theory Comput. 2010, 6 (6), 1783−1797.
(20) Aubry, S.; Sasaki, K.; Sharma, I.; Crich, D. Influence of
protecting groups on the reactivity and selectivity of glycosylation:
chemistry of the 4,6-O-benzylidene protected mannopyranosyl donors
and related species. In Reactivity Tuning in Oligosaccharide Assembly;
Springer, 2010; pp 141−188.
(21) Lu, S. R.; Lai, Y. H.; Chen, J. H.; Liu, C. Y.; Mong, K. K. T.
Dimethylformamide: an unusual glycosylation modulator. Angew.
Chem., Int. Ed. 2011, 50 (32), 7315−7320.
(22) Yasomanee, J. P.; Demchenko, A. V. Effect of remote picolinyl
and picoloyl substituents on the stereoselectivity of chemical
glycosylation. J. Am. Chem. Soc. 2012, 134 (49), 20097−20102.
(23) Yasomanee, J. P.; Demchenko, A. V. Hydrogen bond mediated
aglycone delivery: Synthesis of linear and branched α-glucans. Angew.
Chem. 2014, 126 (39), 10621−10624.
(24) Baek, J. Y.; Kwon, H.-W.; Myung, S. J.; Park, J. J.; Kim, M. Y.;
Rathwell, D. C.; Jeon, H. B.; Seeberger, P. H.; Kim, K. S. Directing
effect by remote electron-withdrawing protecting groups at O-3 or O-
4 position of donors in glucosylations and galactosylations.
Tetrahedron 2015, 71 (33), 5315−5320.
(25) Nigudkar, S. S.; Demchenko, A. V. Stereocontrolled 1, 2-cis
glycosylation as the driving force of progress in synthetic carbohydrate
chemistry. Chem. Sci. 2015, 6 (5), 2687−2704.
(26) Komarova, B. S.; Tsvetkov, Y. E.; Nifantiev, N. E. Design of α-
selective glycopyranosyl donors relying on remote anchimeric
assistance. Chem. Rec. 2016, 16 (1), 488−506.
(27) Park, Y.; Harper, K. C.; Kuhl, N.; Kwan, E. E.; Liu, R. Y.;
Jacobsen, E. N. Macrocyclic bis-thioureas catalyze stereospecific
glycosylation reactions. Science 2017, 355 (6321), 162−166.
(28) Chatterjee, S.; Moon, S.; Hentschel, F.; Gilmore, K.; Seeberger,
P. H. An empirical understanding of the glycosylation reaction. J. Am.
Chem. Soc. 2018, 140 (38), 11942−11953.
(42) Le Mai Hoang, K.; Pardo-Vargas, A.; Zhu, Y.; Yu, Y.; Loria, M.;
Delbianco, M.; Seeberger, P. H. Traceless photolabile linker expedites
the chemical synthesis of complex oligosaccharides by automated
glycan assembly. J. Am. Chem. Soc. 2019, 141 (22), 9079−9086.
(43) Pardo-Vargas, A.; Bharate, P.; Delbianco, M.; Seeberger, P. H.
Automated glycan assembly of arabinomannan oligosaccharides from
Mycobacterium tuberculosis. Beilstein J. Org. Chem. 2019, 15 (1),
2936−2940.
(44) Zhu, Y.; Tyrikos-Ergas, T.; Schiefelbein, K.; Grafmuller, A.;
̈
Seeberger, P. H.; Delbianco, M. Automated access to well-defined
ionic oligosaccharides. Org. Biomol. Chem. 2020, 18 (7), 1349−1353.
(45) Tyrikos-Ergas, T.; Bordoni, V.; Fittolani, G.; Chaube, M.;
Grafmuller, A.; Seeberger, P. H.; Delbianco, M. Systematic structural
̈
characterization of chitooligosaccharides enabled by Automated
Glycan Assembly. Chem. - Eur. J. 2021, 27 (7), 2321−2325.
(46) Seeberger, P. H. The logic of automated glycan assembly. Acc.
Chem. Res. 2015, 48 (5), 1450−1463.
(47) Werz, D. B.; Castagner, B.; Seeberger, P. H. Automated
synthesis of the tumor-associated carbohydrate antigens Gb-3 and
Globo-H: incorporation of α-galactosidic linkages. J. Am. Chem. Soc.
2007, 129 (10), 2770−2771.
(48) Walvoort, M. T.; van den Elst, H.; Plante, O. J.; Kröck, L.;
Seeberger, P. H.; Overkleeft, H. S.; van der Marel, G. A.; Codée, J. D.
Automated solid-phase synthesis of β-mannuronic acid alginates.
Angew. Chem., Int. Ed. 2012, 51 (18), 4393−4396.
9767
https://doi.org/10.1021/jacs.1c02188
J. Am. Chem. Soc. 2021, 143, 9758−9768

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
(49) Hahm, H. S.; Hurevich, M.; Seeberger, P. H. Automated
assembly of oligosaccharides containing multiple cis-glycosidic
linkages. Nat. Commun. 2016, 7 (1), 1−8.
(50) Boons, G.-J.; Bowers, S.; Coe, D. M. Trityl ethers in
oligosaccharide synthesis: A novel strategy for the convergent
assembly of oligosaccharides. Tetrahedron Lett. 1997, 38 (21),
3773−3776.
(51) Elferink, H.; Severijnen, M. E.; Martens, J.; Mensink, R. A.;
Berden, G.; Oomens, J.; Rutjes, F. P.; Rijs, A. M.; Boltje, T. J. Direct
experimental characterization of glycosyl cations by infrared ion
spectroscopy. J. Am. Chem. Soc. 2018, 140 (19), 6034−6038.
(52) Marianski, M.; Mucha, E.; Greis, K.; Moon, S.; Pardo, A.;
Kirschbaum, C.; Thomas, D. A.; Meijer, G.; Helden, G.; Gilmore, K.;
Seeberger, P. H.; Pagel, K. Remote participation during glycosylation
reactions of galactose building blocks: direct evidence from cryogenic
vibrational spectroscopy. Angew. Chem., Int. Ed. 2020, 59 (15), 6166−
6171.
(53) Hansen, T.; Elferink, H.; van Hengst, J. M. A.; Houthuijs, K. J.;
Remmerswaal, W. A.; Kromm, A.; Berden, G.; van der Vorm, S.; Rijs,
A. M.; Overkleeft, H. S.; Filippov, D. V.; Rutjes, F. P. J. T.; van der
Marel, G. A.; Martens, J.; Oomens, J.; Codee, J. D. C.; Boltje, T. J.
Characterization of glycosyl dioxolenium ions and their role in
glycosylation reactions. Nat. Commun. 2020, 11 (1), 1−9.
(54) Crich, D. En route to the transformation of glycoscience: A
chemist’s perspective on internal and external crossroads in
glycochemistry. J. Am. Chem. Soc. 2021, 143 (1), 17−34.
(55) Yu, Y.; Gim, S.; Kim, D.; Arnon, Z. A.; Gazit, E.; Seeberger, P.
H.; Delbianco, M. Oligosaccharides Self-Assemble and Show Intrinsic
Optical Properties. J. Am. Chem. Soc. 2019, 141 (12), 4833−4838.
(56) Zhao, Q.; Zhang, H.; Zhang, Y.; Zhou, S.; Gao, J.
Stereoselective synthesis of a branched α-decaglucan. Org. Biomol.
Chem. 2020, 18 (33), 6549−6557.
(57) Shadrick, M.; Singh, Y.; Demchenko, A. V. Stereocontrolled α-
galactosylation under cooperative catalysis. J. Org. Chem. 2020, 85
(24), 15936−15944.
(58) Teitelbaum, R. C.; Ruby, S. L.; Marks, T. J. On the structure of
starch-iodine. J. Am. Chem. Soc. 1978, 100 (10), 3215−3217.
(59) Madhu, S.; Evans, H. A.; Doan-Nguyen, V. V.; Labram, J. G.;
Wu, G.; Chabinyc, M. L.; Seshadri, R.; Wudl, F. Infinite polyiodide
chains in the pyrroloperylene−iodine complex: insights into the
starch−iodine and perylene−iodine complexes. Angew. Chem., Int. Ed.
2016, 55 (28), 8032−8035.
(60) Hinrichs, W.; Bu
Pfannemuller, B.; Saenger, W. An amylose antiparallel double helix at
̈
ttner, G.; Steifa, M.; Betzel, C.; Zabel, V.;
̈
atomic resolution. Science 1987, 238 (4824), 205−208.
(61) Fittolani, G.; Seeberger, P. H.; Delbianco, M. Helical
polysaccharides. Pept. Sci. 2020, 112 (1), e24124.
9768
https://doi.org/10.1021/jacs.1c02188
J. Am. Chem. Soc. 2021, 143, 9758−9768