Concentration dependence of glycosylation outcome: A clue to reproducibility and understanding the reasons behind

Article abstract of DOI:10.1002/ejoc.201101613

Changes in the concentration of reagents (0.009-0.2 M) have been shown to dramatically effect the yield and stereoselectivity of glycosylation with a sialic acid based glycosyl donor in a complex nonlinear manner that correlates with changes in the structures of the supramers of the reagents. The yield of disaccharide gradually increases with concentration and levels off at concentrations of glycosyl donor higher than 69 mM. The ratio of anomers is very high at some concentrations (α/β ≈ 20:1), moderate (α/β ≈ 8:1) or very low (α/β ≈ 4:1) at others. The formation of mixed supramers of glycosyl donor and glycosyl acceptor at concentrations exceeding 69 mM was detected by polarimetry and laser light scattering. Depending on the concentration, the reagents can form different supramers with distinct chemical properties. The changes in supramer structure (indicated by the red vertical arrows) correlate with the glycosylation yield (1) and stereoselectivity (2) and can be detected by physical methods. Copyright

Full text of DOI:10.1002/ejoc.201101613

FULL PAPER
DOI: 10.1002/ejoc.201101613
Concentration Dependence of Glycosylation Outcome: A Clue to
Reproducibility and Understanding the Reasons Behind
Leonid O. Kononov,*[a] Nelly N. Malysheva,[a] Anna V. Orlova,[a] Alexander I. Zinin,[a]
Tatiana V. Laptinskaya,[b] Elena G. Kononova,[c] and Natalya G. Kolotyrkina[a]
Dedicated to the memory of the late Professor Leon V. Backinowsky
Keywords: Glycosylation / Sialic acids / Stereoselectivity / Aggregation
Changes in the concentration of reagents (0.009–0.2
M
) have
levels off at concentrations of glycosyl donor higher than
69 m . The ratio of anomers is very high at some concentra-
tions (α/β ≈ 20:1), moderate (α/β ≈ 8:1) or very low (α/β ≈ 4:1)
been shown to dramatically effect the yield and stereoselecti-
vity of glycosylation with a sialic acid based glycosyl donor
in a complex nonlinear manner that correlates with changes
in the structures of the supramers of the reagents. The yield
of disaccharide gradually increases with concentration and
M
at others. The formation of mixed supramers of glycosyl do-
nor and glycosyl acceptor at concentrations exceeding 69 m
was detected by polarimetry and laser light scattering.
M
ride separately,[4] the use of a sialyl donor with suitable tem-
porary protection at N(5) is generally considered more rea-
sonable because in this case the number of nontrivial sialyl-
ation steps is minimized.[2d,2e] One of the practical ap-
proaches to libraries of sialo-oligosaccharides, which com-
prise the (α-2Ǟ3)-intersaccharidic linkage, with almost any
N substituent from the single precursor involves the use of
a sialyl-(α-2Ǟ3)-galactose building block with a removable
protecting group at the N(5) of the sialic acid residue.[5]
In our own work, during the development of a prepara-
tive synthesis of the sialyl-galactose building block 6[6]
(Scheme 1) we faced a problem of reproducibility of yield
and stereoselectivity of sialylation (1b + 5 Ǟ 6). Especially
annoying results were obtained when we tried to change the
concentrations of the reagents. Clearly, this reaction re-
quired optimization, which could be accomplished in a
number of ways.[7] We are trying to develop rational ap-
proaches to the design and optimization of glycosylation
experiments. These approaches are based on the recently
introduced supramer concept,[8] which emphasizes the im-
portance of supramolecular aggregation in the reaction
mixture leading to the formation of supramers[9] (supra-
molecular isomers), which are differently arranged supra-
molecular assemblies of the same molecular entities. Ac-
cording to this concept, molecular structures of reactants
and reaction conditions (solvent, temperature, concentra-
tion, presence of “nonreacting” compounds, etc.) would de-
termine the aggregation type and the spatial arrangement
(“structure”) of the supramers formed in each particular
case. Modification of the protecting groups or other nonre-
acting functional groups in the molecule could modulate
Introduction
Sialic acid containing glycoconjugates are involved in a
wide range of biological phenomena ranging from cell–cell
adhesion and mobility to oncogenesis and recognition by
viruses and bacteria.[1] Therefore the synthesis and the bio-
medical investigation of sialic acid containing glycoconju-
gates, oligosaccharides, and their analogues is a very impor-
tant area of research aimed at understanding their bio-
logical roles and determining their therapeutic importance.
For this reason, tremendous effort has been made to de-
velop efficient methods for the synthesis of sialo-oligosac-
charides.[2,3] Sialic acids are attached to other carbohydrates
by means of a glycosylation reaction called sialylation.[2a]
Although substantial progress has recently been achieved in
the synthesis of sialo-oligosaccharides,[3] poor predictability
and reproducibility of yield and stereoselectivity are still
typical of the sialylation reaction. The problem becomes
even more complicated when both N-acetyl- and N-
glycolyl-substituted sialo-oligosaccharides are required. Al-
though it is possible to synthesize each type of oligosaccha-
[a] N. K. Kochetkov Laboratory of Carbohydrate Chemistry, N. D.
Zelinsky Institute of Organic Chemistry,
Leninsky prospect 47, 119991 Moscow, Russian Federation
Fax: +7-499-137-6148
E-mail: kononov@ioc.ac.ru
leonid.kononov@gmail.com
[b] Faculty of Physics, M. V. Lomonosov Moscow State University,
Leninskie Gory, 119992 Moscow, Russian Federation
[c] Laboratory of Molecular Spectroscopy, A. N. Nesmeyanov
Institute of Organo-Element Compounds,
Vavilova 28, 119991 Moscow, Russian Federation
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201101613.
1926
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2012, 1926–1934

Concentration Dependence of Glycosylation Outcome
the aggregation type and the structure of the supramers
formed from the same molecular scaffold. Supramers with
different structures or compositions are expected to react
differently. The accessibility of the reaction center in the
supramers present would determine the apparent (macro-
scopic) reactivity and the outcome of a reaction [product
yield and reaction (stereo/regio-)selectivity]. The supramer
approach has been shown to be useful to explain, predict,
and discover a series of unexpected phenomena[8,10] and led
to the development of a novel sialyl donor, N,N-diacetylsia-
lyl chloride,[11] with improved glycosylating properties.
Figure 1. Concentration-induced rearrangement of supramers in a
solution of glycosyl donor 1a in MeCN (a), and the yield of disac-
charide 3 (MeCN, NIS–TfOH, –40 °C, 3 h) at different concentra-
tions of 1a (b): 1 (circles, dashed line): Amount of “free” NH
groups in a solution of glycosyl donor 1a in MeCN (ca. 25 °C)
calculated as the ratio of the absorptions of the NH band at
3365 cm–1 and the amide C=O band at 1676 cm–1 [A(NH free)/
A(C=O amide)], the intensity of the latter band being an “internal
standard”; 2 (squares, solid line): specific optical rotation ([α]D) of
a solution of 1a in MeCN (averaged over a range of 23–26 °C).
The rectangle denotes the concentration range of the supramer re-
Scheme 1. Structures of sialic acid glycosyl donors 1a,b [R = Ac
(a), TFA (b)], glycosyl acceptors 2 and 5, disaccharides 3 and 6,
and glycals 4a,b. Reagents and conditions: a) NIS-TfOH, MeCN,
–40 °C.
At present, little is known about supramers and their
structures. Nevertheless, by studying solutions of a typical
sialyl donor 1a by IR spectroscopy and polarimetry it was
possible to detect changes in the structures of hydrogen-
bonded supramers of sialyl donor 1a upon a change in the
concentration.[8b] This supramer rearrangement occurs in a
very narrow concentration range and is detectable by polar-
imetry and IR spectroscopy (Figure 1, a). Not surprisingly,
the yield of disaccharide 3, formed from this glycosyl donor
1a by the reaction 1a + 2 Ǟ 3, shown in Scheme 1, also ex-
periences discontinuity in the same concentration range
(Figure 1, b) in which supramers with different structures
and reactivities are formed.[8b]
There are two important messages from Figure 1. The
first is that polarimetry and IR spectroscopy are very sensi-
tive[12,13] to supramolecular aggregation and the second is
that critical points on the plots of optical rotation or the
intensity of a relevant band (e.g., NH vibrations) in the IR
spectrum against concentration can give us information
concerning changes in the structures of supramers and
hence changes in their reactivity and outcome of glycosyl-
ation. Thus, by studying the concentration dependence of
the optical rotation or the intensity of bands in the IR spec-
tra of solutions of glycosyl donor in the reaction solvent,
one may find ranges of concentration in which anomalies
in chemical reactivity could be expected.
arrangement. Adapted from ref.[8b]
.
of solutions of the glycosyl donor by polarimetry, IR spec-
troscopy, or other appropriate physical methods should be
attempted. Analysis of the corresponding concentration
plots may reveal discontinuities. The corresponding critical
points would serve as guidelines for choosing concentra-
tions for glycosylation. When performing chemical experi-
ments, it is important to include critical concentrations and
at least one intermediate different critical concentrations or
between them.
Herein we report the use of the supramer approach for
the optimization of the troublesome sialylation reaction
1b + 5 Ǟ 6 (Scheme 1). This proof-of-principle study has
led to the discovery of completely unprecedented results,
which are described below.
Results and Discussion
Thus, the following scenario for the optimization of gly-
The first task was to find the critical concentrations in
cosylation can be proposed, which is based on the supramer solutions of glycosyl donor 1b[15] in MeCN, which was the
approach. Before embarking on any glycosylation, a study reaction solvent. However, unlike in the previous study of
Eur. J. Org. Chem. 2012, 1926–1934
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1927

L. O. Kononov et al.
FULL PAPER
the glycosyl donor 1a with the AcNH group at C(5) (see
Figure 1, a),[8b] an attempt to use IR spectroscopy to moni-
tor changes in hydrogen-bond-mediated aggregation in
solutions of glycosyl donor 1b with the TFANH group at
C(5) was unsuccessful. Although changes in the relative in-
tensities of “free” and “bound” NH vibrations and in the
positions and intensities of carbonyl vibration bands were
clearly visible upon changing the concentration of glycosyl
donor 1b, their quantitative assessment was found difficult
due to substantial overlap of the NH and solvent bands.
Then we measured the optical rotations of a series of
solutions of the glycosyl donor 1b in MeCN. Two maxima
and one minimum are observed in the plot of specific rota-
tion against concentration of glycosyl donor (Figure 2, b).
To simplify further comparison, these concentrations (50,
69, and 103 mm) are marked with thick arrows, which are
also present in the following figures. These arrows indicate
the critical concentrations of the glycosyl donor solution,
as detected by polarimetry. It is important to emphasize
that the shape of the optical rotation plot clearly indicates
the formation of supramolecular aggregates in solution and
their interconversion upon changes in concentration.
After studying the properties of solutions of glycosyl do-
nor 1b we performed a series of chemical experiments at
critical points and at some intermediate concentrations
(Figure 2, a and Table 1). We found that by changing the
concentrations of reagents one can dramatically modulate
the yield and stereoselectivity of sialylation (Figure 2, a).
The concentration dependence of the yield in this case
(dashed line, Figure 2, a, 1) is strikingly similar to that pre-
viously obtained[8b] for the glycosyl donor 1a with the
AcNH group at C(5) (Figure 1, b). The yield gradually in-
creases with the concentration of the glycosyl donor 1b with
the TFANH group at C(5) and levels off after the critical
concentration of 69 mm. The concentration dependence of
the stereoselectivity (solid line, Figure 2, a, 2) is even more
complex. Note that discontinuities of this line correspond
well to the critical concentrations detected by polarimetry
Figure 2. Outcome of sialylation (a) and its correlation with optical
(see Figure 2, b), which confirms the validity of our ap- rotation (b,c) and DLS data (d) for solutions at different concentra-
tions: (a) 1 (squares, dashed line): Yield of disaccharide 6; 2 (circles,
proach. In the particular case of a sialylation reaction
(1b + 5 Ǟ 6), the ratio of anomers of the glycosylation
solid line): anomeric ratio (α/β) of 6. (b) Specific optical rotation
([α]2D8) of sialyl donor 1b in MeCN. (c) Optical rotation (α2D8) in
product 6 can be very high at some concentrations (α/β ≈
MeCN; 3 (circles, dashed line): the sum of rotations of 1b and 5
20:1), moderate (α/β ≈ 8:1), or very low (α/β ≈ 4:1) at others,
in all cases the equatorial glycoside 6[16] dominating due to
the “nitrile effect”[17] (see Figure 2, a and Table 1).[18]
However, more careful examination of this plot suggests
that at high concentrations the situation is more complex
measured separately; 4 (squares, solid line): optical rotation of 1:1
mixtures of 1b and 5. (d) Comparison of the average hydrodynamic
radii (Rh) of light-scattering particles in MeCN: 5 (squares, solid
line): 1:1 mixtures of 1b and 5; 6 (triangles, dashed line): solution
of glycosyl acceptor 5; 7 (circles, dotted line): solution of glycosyl
donor 1b. Vertical thick arrows indicate critical concentrations (50,
than we initially assumed. The second maximum (at 69, and 103 mm). Shaded area marks the concentration range in
103 mm) in the optical rotation plot (Figure 2, b), unlike
which mixed supramers {1b + 5} exist.
the first one (at 50 mm), corresponds to a poorly resolved
discontinuity in the stereoselectivity plot (Figure 2, a) rather
than to a pronounced maximum. For this reason, we de-
We studied the aggregation in 1:1 mixtures of glycosyl
cided to include the second component of the reaction, that donor 1b and glycosyl acceptor 5 in MeCN by polarimetry
is, the glycosyl acceptor 5,[19] in our supramer model. This (Figure 2, c). The solid line (Figure 2, c, 4) represents the
idea is quite natural because there are a number of exam- observed optical rotation of this mixture in MeCN. The
ples reported in the literature in which the nature of the dashed line (Figure 2, c, 3) corresponds to the calculated
glycosyl acceptor influences the stereoselectivity of glyco- sum of the optical rotation values of glycosyl donor 1b and
sylation.[20,21]
glycosyl acceptor 5 measured separately. It can clearly be
1928
www.eurjoc.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2012, 1926–1934

Concentration Dependence of Glycosylation Outcome
Table 1. Conditions and products of glycosylation.[a]
Entry Conc. Reaction Yield of Yield of α iso- Anomeric ratio
[mM]
time
6 [%][b] mer of 6 [%][c]
for 6 (α/β)[d]
1
2
3
4
5
6
7
8
9
18
35
50
69
86
103
120
199
50
138 h
4 h
2.25 h
1 h
45 min
1 h
40 min
45 min
40 min
1 h
33.6
54.1
57.1
66.7
70.8
70.5
71.2
71.8
71.3
28.1
26.8
44.9
51.0
60.6
64.5
64.3
64.7
66.2
66.0
12.2
3.9:1[e]
8.2:1
7.3:1
18.8:1
6.4:1
8.6:1
12.3:1
12.1:1
19.9:1
1:1.3[e]
9
10[f]
[a] Reagents and conditions: 1 equiv. 1b, 1 equiv. 5, MeCN, NIS–
TfOH, 3 Å molecular sieves, –40 °C. The reaction was quenched
after complete consumption of the glycosyl donor (TLC control).
[b] The disaccharide fraction was isolated by gel chromatography
on Bio-Beads S-X3 (toluene) and then the isomers were separated
by silica gel column chromatography (the sum of the isolated yields
of the α and β isomers of 6 is given). [c] As for [b] (the isolated
yield of the α isomer of 6 is given). [d] Determined from the 1H
NMR spectroscopic data for the disaccharide fraction isolated by
gel chromatography on Bio-Beads S-X3 (toluene). [e] Approximate
value obtained by weighing the purified α and β isomers of 6. [f]
The reaction was performed in CH2Cl2.
seen that at the critical concentration of 69 mm, these two
lines become separated. This means that at higher concen-
trations these two compounds are no longer independent
and do interact. We believe that this deviation from additiv-
ity at high concentrations is related to the formation of
mixed supramers (heterosupramers) {1b + 5} comprising
molecules of both reagents.
To confirm independently the existence of supramers in
our system we used dynamic light scattering (DLS).[22] Typ-
ical correlation functions and size distributions of light-
scattering particles in solutions of the glycosyl donor 1b,
glycosyl acceptor 5, and a 1:1 mixture 1b and 5 in MeCN
at one of the concentrations studied (69 mm) are shown in
Figure 3.[24] The plot of hydrodynamic radii of light-scat-
tering particles in these solutions against concentration of
solute (Figure 2, d) also indicates the formation of mixed
supramers {1b + 5} of sialyl donor 1b and glycosyl acceptor
5 at concentrations exceeding 69 mm. It can clearly be seen
in this plot that in the high concentration range, marked
Figure 3. Correlation functions and intensity-weighted size distri-
butions in solution of the glycosyl donor 1b (a), glycosyl acceptor
5 (b), and their 1:1 mixture (c) in MeCN at 69 mm concentration.
with a shaded background, the hydrodynamic radius of the crease in the yields of products of the side-reactions that
light-scattering particles in the mixture of glycosyl donor 1b usually accompany the sialylation reaction, the major one
and glycosyl acceptor 5 (solid line, Figure 2, d, 5) is larger being glycal 4b[25] formed by elimination. In other words, a
than the hydrodynamic radii of particles in solutions of the change in the supramer structure of glycosyl donor 1b in-
individual components (dashed and dotted lines, Figure 2, duced by a change in concentration resulted in a change in
d, 6 and 7). Thus, the existence of mixed supramers {1b + the chemoselectivity of the reaction, that is, the chemical
5} at high concentrations may be considered established.
properties of 1b have changed. This is not surprising from
It is important to stress that in the high concentration the supramer point of view. As mentioned earlier, changes
range (at concentrations exceeding 69 mm), when mixed su- in the structures of the supramers can modulate the accessi-
pramers {1b + 5} exist (Figure 2, c and d), the yield of di- bility of the reaction center for the nucleophile and hence
saccharide 6 does not depend on concentration although it the selectivity of the reaction, leading to either the substitu-
clearly does at lower concentrations, when homosupramers tion (SN1) product (disaccharide 6) or the elimination (E1)
{1b} exist (Figure 2, a, 1). Because all the reactions were product (glycal 4b). Apparently, unimolecular elimination
quenched after the complete consumption of the glycosyl begins to dominate when substitution (glycosylation) is dis-
donor, the decrease in disaccharide yield reflects the in- favored.
Eur. J. Org. Chem. 2012, 1926–1934
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1929

L. O. Kononov et al.
FULL PAPER
Although the influence of concentration on the stereose-
lectivity of the glycosylation is very complex in this particu-
lar case (Figure 2, a, 2), it can be rationally discussed by
using the supramer approach. One can even say that with-
out knowing the critical concentrations, which were deter-
mined by the supramer approach, we would not be able
to reveal this complex picture. This complex concentration
dependence of the stereoselectivity of sialylation is very im-
portant from both practical and mechanistic points of view.
From a practical point of view, during the optimization,
without knowing the critical concentrations it is very easy
to miss good conditions, under which the ratio of anomers
is very high (α/β ≈ 20:1). The theoretical implications are
even more important. It is commonly believed that sialyl-
ation with thioglycoside glycosyl donors is a multistep pro-
cess that proceeds via the formation of an oxacarbenium
ion (A, Scheme 2),[2,26,27] which may be stabilized in MeCN
by the solvent molecules to form the so-called “nitrilium
intermediate” B.[2,17] Because α-selective sialylation can be
achieved only under kinetically controlled conditions (as
the β anomer is more thermodynamically stable due to the
anomeric effect),[2] the stereoselectivity of sialylation essen-
tially reflects the different preferences for the attack of a
nucleophile from different diastereotopic faces of a glycosyl
donor (or, more correctly, species generated from it upon
activation). Current views on the origin of the stereoselecti-
vity of sialylation in MeCN (and other glycosylation reac-
Scheme 2. Simplified, generally accepted mechanism of stereocon-
trol during sialylation (thioglycoside 1a shown as an example) in
MeCN, which implies an increase in α stereoselectivity upon di-
lution.
tions with glycosyl donors without stereocontrolling par- faces of the anomeric center in a sialic acid supramer by the
ticipating substituents) predict that the proportion of equa- neighboring molecules that comprise the supramer would
torial glycoside[16] should be higher in dilute solutions, considerably influence the stereoselectivity of glycosylation.
which is attributed to increased solvent participation upon It is not unexpected that dramatic changes in supramer
dilution.[28] However, the practical validation of such a pre- structure upon changes in concentration, as evidenced by
diction is accompanied by unexpected difficulties. Indeed, the polarimetry and DLS data presented herein, are ac-
it is virtually impossible to perform glycosylation at every companied by significant and correlated changes in the ster-
feasible concentration. We found that the results of the ex- eoselectivity of sialylation (Figure 2). Interestingly, the sup-
perimental verification depend on the set of concentrations ramer approach is better correlated with the recently intro-
used, as can be seen from Figure 2 (a) and Table 1. De- duced[31] “conformer and counterion distribution hypothe-
pending on the choice of concentrations one can obtain re- sis” (which emphasizes the conformational preferences of a
sults that either support this view (entries 4 and 5 or 4 and glycosyl donor as the major factor influencing the stereose-
6, or 4, 7, and 8 in Table 1) or contradict it[29] (entries 1, 3, lectivity of glycosylation) rather than with the commonly
6, 7, and 9). It is even possible to “prove” that stereoselecti- accepted “solvent-coordination hypothesis” (which ratio-
vity almost does not depend on concentration[30] by the ap- nalizes, e.g., the “nitrile effect”[17] by assuming preferential
propriate selection of experimental concentrations (en- coordination of acetonitrile to the reactive cation on the β
tries 2, 3, and 6). It looks like almost any theory and view side of the anomeric carbon in sialic acid derivatives[16]). It
can find experimental support. We wish to stress that this is important that the latter explanation (see Scheme 2) does
paradoxical situation can emerge only if one does not con- not leave any room for the effects of concentration and the
sider the existence of critical concentrations, which indicate presence of glycosyl acceptor, which were revealed in this
changes in supramer structure, and their relevance to the study. Our optical rotation data (Figure 2, b and c) suggest
outcome of glycosylation. Because previous studies[28,30] of dramatic changes in the molecular conformations of both
the influence of concentration on the outcome of glycosyl- reagents (1b and 5) upon changes in concentration.[12]
ation reactions did not pay any attention to the choice of These changes might modify the barriers of the pathways
experimental set of concentrations, it is only natural that leading to different conformers of the true reacting species
the results obtained earlier often seem to be confusing.
(glycosyl cation-like intermediates), hence their populations
The supramer approach recognizes all the established in the reaction mixtures and the overall stereoselectivity of
factors that could influence the stereoselectivity of sialyl- glycosylation.
ation at the molecular level.[2] In fact, our results confirm
One has to comment on the apparent discrepancy of the
the existence of the stereocontrolling “nitrile effect”.[17,18] discovery in this study of the complex concentration depen-
In addition, we believe that selective screening of one of the dence of stereoselectivity of the sialylation of glycosyl ac-
1930
www.eurjoc.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2012, 1926–1934

Concentration Dependence of Glycosylation Outcome
ceptor 5 with sialyl donor 1b, the α/β ratio ranging from residues (with identical protecting group pattern) involved
4:1 to 20:1 (see Scheme 1 and Figure 2, a, 2), with our own in the actual chemical reaction, may (or may not) be prob-
previous results[8b] in which the stereoselectivity of sialyl- lematic or even impossible. This situation is expected from
ation of the glycosyl acceptor 2 with sialyl donor 1a (see the supramer point of view. Larger oligosaccharides neces-
Scheme 1) at the end of the reaction was almost indepen- sarily have additional sites capable of intermolecular inter-
dent of concentration (α/β ≈ 7:1–10:1 after 3 h of reaction). actions[35] and the supramers formed from such oligosac-
Later, we showed that considerable changes in the stereose- charides (unlike much smaller monosaccharides) may be ar-
lectivity of sialylation may occur during the course of the ranged in such a way that the reaction center (anomeric
NIH–TfOH-promoted glycosylation (in some cases the α/β center in the case of glycosylation) is partially shielded
ratio approached 27:1 after 15 min of reaction) and related (leading to modified selectivity) or completely blocked (giv-
them to changes in the supramer structure of the glycosyl ing no reaction at all) by the neighboring molecules com-
donor induced by changes in the concentration of the suc- prising the supramer. The supramer approach might be use-
cinimide formed during the reaction.[8c] In our opinion, the ful in interpreting other examples (non-carbohydrate) in
main difference in these two cases lies in the noticeable dif- which a model reaction in fact does not model anything at
ference in reactivity between the glycosyl donor 1a with an all or when difficult and unexpected reactivities of func-
AcNH group at C(5) and glycosyl donor 1b with a TFANH tional groups are observed.[36]
group at C(5), the latter being much more reactive. This
The results obtained in this study suggest that at different
leads to shorter reaction times with 1b (40–60 min is concentrations, glycosyl donors form different supramers,
enough for complete consumption of the starting thiogly- which apparently have very distinct chemical properties. In
coside 1b in the high concentration range; see Table 1).[32] other words, we can now claim that the same chemical com-
The high stereoselectivities achieved in this study suggest pound can react differently depending on the type of sup-
that during this period of time no significant rearrangement ramer formed. One can even say that the chemical proper-
of supramers, induced by succinimide, takes place. This sug- ties are a feature of a supramer rather than that of an iso-
gests that a minimum time of around 30 min is required for lated molecule.[37] This unusual notion may have important
supramer rearrangement induced by succinimide, which is implications and consequences, which are yet to be com-
in agreement with our previous results.[8c]
pletely understood, in other areas of chemistry.
Caution should be exercised when transferring our find-
ings to other systems. This proof-of-principle study repre-
sents the situation in the particular system of glycosyl do-
nor 1b and glycosyl acceptor 5 in MeCN. The results ob-
tained for other combinations of glycosyl donor and gly-
cosyl acceptor (or in other solvents) may be different; the
supramer approach even predicts that they should be dif-
ferent because supramers with different structures, and
hence chemical properties, can be formed in these cases.
More systems should be studied and analyzed by using the
supramer approach before any substantiated generaliza-
tions on the influence of concentration (and other factors
including the molecular structures of reagents) on the out-
come of glycosylation can be made.
This necessarily pessimistic view is supported by the gen-
erally accepted fact that currently it is almost impossible to
predict the outcome of a specific glycosylation.[33] In our
opinion, at the present level of development, only the sup-
ramer approach allows rational discussion of the influence
of the nature of the glycosyl acceptor on the yield and
stereoselectivity of glycosylation.[2h,20,21b,27] The commonly
accepted concept of double diastereoselection,[21] although
helpful in some cases, cannot treat adequately the some-
times pronounced influence of a change in protecting group
pattern (especially at positions remote from the reaction
center, for example, in aglycon of an oligosaccharide gly-
cosyl acceptor[20c]) on the same molecular scaffold.[20] This
is especially true for glycosylations involving large oligosac-
charide glycosyl donors and acceptors. Even though a
model reaction with monosaccharide derivatives may pro-
ceed smoothly, reaction of the corresponding oligosac-
Conclusions
The results obtained in this study can be summarized in
the following way. A change in the concentration of rea-
gents was found to influence the outcome of glycosylation
in a complex way. Depending on the concentration, mole-
cules of reagents form fundamentally different supramers,
which can be distinguished by physical methods and differ
in their chemical properties. A better understanding of the
real reasons that determine the stereoselectivity of glycosyl-
ation can be obtained by consideration of the nature and
supramolecular structures of all the components of the re-
action mixture rather than by analysis of the molecular
structure of the glycosyl donor alone. The commonly ac-
cepted view of the origin of the stereoselectivity of sialyl-
ation in MeCN, which relies on solvent participation, is not
supported by the results obtained herein. The results ob-
tained in this study allow a fresh look at the problems of
reproducibility of glycosylation.
Experimental Section
General: The reactions were performed with commercial reagents.
Solvents for reactions were distilled and purified before use accord-
ing to standard procedures. HPLC far-UV-grade acetonitrile
(99.9%, water content Ͻ0.02%, Acros Organics, Cat. No.
268260010) was used for optical rotation and DLS measurements.
TLC was carried out on silica gel 60 F254 plates (Merck); spots
were visualized under UV light and by heating the plates after im-
mersion in a 1:10 (v/v) mixture of 85% aqueous H3PO4 and 95%
charide derivatives, with formally the same monosaccharide EtOH. NMR spectra of CDCl3 solutions were recorded with a
Eur. J. Org. Chem. 2012, 1926–1934
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1931

L. O. Kononov et al.
FULL PAPER
Bruker AVANCE-600 spectrometer. 1H and 13C NMR chemical
shifts are given relative to residual CHCl3 (δ = 7.27 and 77.0 ppm).
Signals were assigned through 2D NMR (COSY, HSQC, HMBC)
and DEPT-135 experiments. Anomeric configurations of sialic
acid derivatives were determined by measurement of the
3JC-1,3ax-H coupling constants[38] using the J-HMBC experiment.
HRMS (electrospray ionization, ESI) spectra were recorded with a
Bruker micrOTOF II mass spectrometer for 2ϫ10–5 m solutions in
MeCN. IR spectra of amide 1b and 1:1 mixtures of 1b and alcohol
5 in MeCN (concentration range 9–199 mm) were obtained with a
Bruker IFS 25 FTIR spectrometer. Optical rotations were mea-
sured for filtered (0.45 μm) solutions of amide 1b and 1:1 mixtures
of 1b and alcohol 5 in MeCN (concentration range 9–199 mm; ex-
act concentrations correspond to the concentrations listed in
Table 1) with a PU-07 automatic polarimeter (Russia) at 28 °C in
a jacketed cell (10 cm length). The temperature was maintained
with an accuracy of Ϯ0.2 °C. Each measurement at a particular
concentration was repeated 10 times on one day and then repeated
again on another day (10 times), and then averaged and plotted
against concentration (see Figure 2, b, Figure 2, c). The standard
deviations were calculated by using the Student distribution (95%
probability) and did not exceed 1% for either observed (αD) or
specific ([α]D) rotation values. DLS experiments were performed on
filtered (0.45 μm) solutions of amide 1b, alcohol 5, and 1:1 mixtures
of 1b and 5 in MeCN (concentration range 9–199 mm; exact con-
centrations correspond to the concentrations listed in Table 1) at
24 °C with an ALV Correlation Goniometer System 5000/6010
(Langen, Germany). To obtain intensity correlation functions
[g2(τ)], data were averaged over 20 independent measurements (the
total collection time was 40 min for each point on the graph in
Figure 2, d) and then processed by using the CONTIN algorithm
to calculate contributions to the scattered intensity from particles
of each observable size (the so-called “intensity-weighted size dis-
tribution”,[23] see Figure 3) and hydrodynamic radii (Rh) of light-
scattering particles (see Figure 2, d), which were calculated at the
maxima of intensity-weighted size distributions.
and β-3eq-H (δ = 2.77 ppm) of the Neu5Ac residue were used].
Later eluted fractions contained Neu5TFA glycal 4b[25] and finally
unreacted alcohol 5. A base-line separation of all mentioned com-
ponents was repeatedly achieved. The disaccharide fraction was
purified by chromatography on a silica gel 60 column to give pure
α- and β-linked isomers of disaccharide 6 (for the yields, see Table 1
and Figure 2, a, 1; all yields were calculated with respect to glycosyl
donor 1b). TLC data: Rf = 0.40 (1b), 0.20 (4b), 0.63 (5), 0.36 (α-
6), 0.60 (β-6) (benzene/acetone, 9:1 v/v).
Disaccharide α-6: [α]2D7 = –14.7 (c = 2.0, CH2Cl2). 1H NMR
(600 MHz, CDCl3): δ = 7.44–7.41 (m, 2 H, Ph), 7.36–7.25 (m, 13
H, Ph), 7.05–7.01 (m, 2 H, MeOC6H4), 6.82–6.79 (m, 2 H,
MeOC6H4), 6.58–6.52 (m, 1 H, NH), 5.46 (ddd, J8,7 = 7.7, J8,9
=
5.6, J8,9Ј = 2.4 Hz, 1 H, 8-H Neu), 5.29 (dd, J7,6 = 2.0, J7,8 = 7.7 Hz,
1 H, 7-H Neu), 5.06 (ddd, J4,3ax = 12.1, J4,3eq = 4.8, J4,5 = 10.4 Hz,
1 H, 4-H Neu), 4.97 (d, J1,2 = 7.6 Hz, 1 H, 1-H Gal), 4.95 (d, J =
11.8 Hz, 1 H, PhCH), 4.90 (d, J = 11.6 Hz, 1 H, PhCH), 4.84 (d,
J = 11.8 Hz, 1 H, PhCH), 4.54 (d, J = 11.6 Hz, 1 H, PhCH), 4.51
(d, J = 11.6 Hz, 1 H, PhCH), 4.45 (d, J = 11.6 Hz, 1 H, PhCH),
4.42 (dd, J9Ј,8 = 2.4, J9a,9b = 12.6 Hz, 1 H, 9b-H Neu), 4.25 (dd,
J3,2 = 9.9, J3,4 = 2.9 Hz, 1 H, 3-H Gal), 4.08 (dd, J6,5 = 10.7, J6,7
= 2.0 Hz, 1 H, 6-H Neu), 4.02–3.95 (m, 1 H, 5-H Neu), 3.99 (dd,
J9,8 = 5.6, J9a,9b = 12.6 Hz, 1 H, 9a-H Neu), 3.95 (dd, J2,1 = 7.6,
J2,3 = 9.9 Hz, 1 H, 2-H Gal), 3.78 (s, 3 H, CH3OC6H4), 3.77–3.72
(m, 2 H, 6b-H Gal, 5-H Gal), 3.75 (s, 3 H, CO2Me), 3.71 (d, J4,3
= 2.9 Hz, 1 H, 4-H Gal), 3.70–3.65 (m, 1 H, 6a-H Gal), 2.59 (dd,
J3eq,3ax = 13.2, J3eq,4 = 4.8 Hz, 1 H, 3eq-H Neu), 2.13 (s, 3 H, Ac),
2.04 (dd, J3ax,3eq = 13.2, J3ax,4 = 12.1 Hz, 1 H, 3ax-H Neu), 2.00
(s, 3 H, Ac), 1.97 (s, 3 H, Ac), 1.94 (s, 3 H, Ac) ppm. 13C NMR
(150 MHz, CDCl3): δ = 170.76, 170.44, 170.04, 169.78 (4 MeCO),
3
168.02 (C-1 Neu; JC-1,3ax-H = 6.9 Hz, J-HMBC data), 157.54 (q, J
= 38.0 Hz, CF3CO), 155.12, 151.70 (C-1, C-4 MeOC6H4), 138.97,
138.90, 138.09 (3 Cquat Ph), 128.31, 128.08, 128.06, 127.94, 127.68,
127.59, 127.35, 127.26 (Ph), 118.40 [C-2(6) or C-3(5) MeOC6H4],
115.40 (q, J = 288.0 Hz, CF3CO), 114.43 [C-3(5) or C-2(6) Me-
OC6H4], 102.91 (C-1 Gal), 98.59 (C-2 Neu), 77.45 (C-2 Gal), 76.32
(C-3 Gal), 76.08 (C-4 Gal), 74.91 (2 PhCH2O), 73.48 (PhCH2O),
73.34 (C-5 Gal), 71.64 (C-6 Neu), 69.16 (C-8 Neu), 68.63 (C-6 Gal),
68.50 (C-4 Neu), 67.25 (C-7 Neu), 61.99 (C-9 Neu), 55.63
(CH3OC6H4), 52.91 (CO2CH3), 50.15 (C-5 Neu), 36.62 (C-3 Neu),
21.09, 20.61, 20.49, 20.37 (4 CH3CO) ppm. HRMS (ESI): calcd.
for C54H60F3NNaO19 [M + Na]+ 1106.3604; found 1106.3636.
Typical Glycosylation Procedure: A mixture of thioglycoside 1b[15]
(132.0 mg, 0.206 mmol, 1 equiv.) and alcohol 5[19] (115.2 mg,
0.206 mmol, 1 equiv.) was dried in vacuo for 2 h and then anhy-
drous MeCN (3 mL, distilled from P2O5, stored over 3 Å molecular
sieves) was added under argon. Freshly activated (220 °C, 6 h, in
vacuo) powdered 3 Å molecular sieves (300 mg, Fluka; 100 mg per
1 mL of solvent) were added to the resulting solution and the reac-
tion flask was flushed with argon. The suspension was stirred under
argon at room temp. for 1 h and then cooled to –40 °C (liquid N2/
Disaccharide β-6: [α]2D4 = –13.5 (c = 2.0, CH2Cl2). 1H NMR
(600 MHz, CDCl3): δ = 7.51–7.24 (m, 15 H, 3 Ph), 7.04–7.01 (m,
2 H, MeOC6H4), 6.82–6.78 (m, 2 H, MeOC6H4), 5.22–5.18 (m, 2
MeCN bath). Solid NIS (70.4 mg, 0.312 mmol, 1.5 equiv. per H, 8-H Neu, NH), 5.15 (dd, J7,6 = 2.4, J7,8 = 2.4 Hz, 1 H, 7-H
1 equiv. donor) was added followed by TfOH. Only the minimum
amount of TfOH (2–5 μL) required to generate persistent color was
added. The reaction mixture was stirred under argon at –40 °C un-
til complete consumption of the starting thioglycoside 1b (TLC
monitoring; the time is specified in Table 1), then diluted with
CHCl3 (20 mL), and filtered through a pad of Celite. The solids
Neu), 5.05 (dd, J9b,8 = 2.3, J9a,9b = 12.2 Hz, 1 H, 9b-H Neu), 5.00
(ddd, J4,3ax = 11.6, J4,3eq = 4.9, J4,5 = 10.2 Hz, 1 H, 4-H Neu), 4.98
(d, J = 11.3 Hz, 1 H, PhCH), 4.95 (d, J1,2 = 7.5 Hz, 1 H, 1-H Gal),
4.90 (d, J = 13.0 Hz, 1 H, PhCH), 4.84–4.81 (m, 2 H, 2 PhCH),
4.61–4.55 (m, 2 H, PhCH2), 4.25 (dd, J3,2 = 9.9, J3,4 = 2.0 Hz, 1
H, 3-H Gal), 4.06 (dd, J2,3 = 9.9, J2,1 = 7.5 Hz, 1 H, 2-H Gal),
were thoroughly washed with CHCl3 (100 mL) and the filtrate was 4.03–3.91 (m, 5 H, 9a-H Neu, 4-H Gal, 5-H Neu, 6-H Neu, 5-H
successively washed with 20% aqueous Na2S2O3 (2ϫ50 mL) and Gal), 3.82 (dd, J6a,6b = 9.6, J6b,5 = 6.3 Hz, 1 H, 6b-H Gal), 3.79–
water (2ϫ50 mL), filtered through a plug of cotton wool, and con- 3.75 (m, 1 H, 6a-H Gal), 3.78 (s, 3 H, CH3OC6H4), 3.41 (s, 3 H,
centrated. The residue was dissolved in toluene (2 mL) and sepa-
rated by gel chromatography on a column (50ϫ2.5 cm) with Bio-
Beads S-X3 (200–400 mesh, Bio-Rad) using toluene as the eluent
and a differential refractometer (Knauer) as the detector. The first
eluted fraction contained disaccharide 6, which was analyzed by
NMR spectroscopy to obtain the anomeric ratios α/β [see Table 1
and Figure 2, a, 2; to determine the ratio of anomers of disaccha-
ride 6 the integral intensities of signals of α-3eq-H (δ = 2.59 ppm)
CO2Me), 2.77 (dd, J3eq,3ax = 13.7, J3eq,4 = 4.9 Hz, 1 H, 3eq-H Neu),
2.08 (s, 3 H, Ac), 2.06 (s, 3 H, Ac), 2.01 (s, 6 H, 2 Ac), 1.87 (dd,
J3ax,3eq = 13.7, J3ax,4 = 11.6 Hz, 1 H, 3ax-H Neu) ppm. 13C NMR
(150 MHz, CDCl3): δ = 170.64, 170.52, 170.11, 169.71 (4 MeCO),
3
166.53 (C-1 Neu; JC-1,3ax-H = 2.5 Hz, J-HMBC data), 157.53 (q, J
= 38.1 Hz, CF3CO), 155.14, 151.46 (C-1, C-4 MeOC6H4), 139.27,
138.28, 137.95 (3 Cquat Ph), 128.73, 128.59, 128.33, 128.21, 127.81,
127.77, 127.68, 127.40, 126.78 (Ph), 118.09 [C-2(6) or C-3(5) Me-
1932
www.eurjoc.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2012, 1926–1934

Concentration Dependence of Glycosylation Outcome
[9]
This term has recently been coined to describe stereoisomerism
at the supramolecular level (in the crystal): M. Czugler, N. Ba-
thori, CrystEngComm 2004, 6, 494–503.
Y.-J. Wang, J. Jia, Z.-Y. Gu, F.-F. Liang, R.-C. Li, M.-H. Hu-
ang, C.-S. Xu, J.-X. Zhang, Y. Men, G.-W. Xing, Carbohydr.
Res. 2011, 346, 1271–1276.
OC6H4], 115.24 (q, J = 287.0 Hz, CF3CO), 114.48 [C-3(5) or C-
2(6) MeOC6H4], 102.82 (C-1 Gal), 99.51 (C-2 Neu), 77.61 (C-4
Gal), 77.23 (C-2 Gal), 76.16 (C-3 Gal), 75.05 (PhCH2O), 73.98 (C-
6 Neu), 73.87 (PhCH2O), 73.51 (PhCH2O), 72.57 (C-8 Neu), 71.47
(C-5 Gal), 68.88 (C-6 Gal), 68.67 (C-7 Neu), 67.98 (C-4 Neu), 62.38
(C-9 Neu), 55.61 (CH3OC6H4), 52.54 (CO2CH3), 49.53 (C-5 Neu),
37.25 (C-3 Neu), 20.88, 20.72, 20.51, 20.49 (4 CH3CO) ppm. HR
MS (ESI): calcd. for C54H60F3NNaO19 [M + Na]+ 1106.3604;
found 1106.3589.
[10]
[11]
[12]
A. V. Orlova, A. M. Shpirt, N. Y. Kulikova, L. O. Kononov,
Carbohydr. Res. 2010, 345, 721–730.
The high sensitivity of polarimetry is not surprising because
even subtle conformational changes are known to induce dra-
matic changes in specific rotation values (see ref.[14a]). If these
changes are concentration dependent, they are apparently re-
lated to the aggregation of solute molecules (see ref.[14b]), which
leads to the formation of supramers comprising molecules dif-
fering in conformations or mode (see ref.[14c]) of supramolec-
ular arrangement. Another feature of polarimetry is that al-
most everyone can use it. The instruments are relatively inex-
pensive and available in every carbohydrate laboratory. For this
reason, polarimetry is currently the main instrumental method
capable of detecting changes in supramer structure.
Optical rotation was shown to be a highly sensitive tool for
probing the state of solution, see: L. O. Kononov, D. E. Tsvet-
kov, A. V. Orlova, Russ. Chem. Bull. 2002, 51, 1337–1338.
a) S. M. Wilson, K. B. Wiberg, M. J. Murphy, P. H. Vaccaro,
Chirality 2008, 20, 357–369; b) M.-R. Goldsmith, N. Jayasu-
riya, D. N. Beratan, P. Wipf, J. Am. Chem. Soc. 2003, 125,
15696–15697; c) M. Suarez, N. Branda, J.-M. Lehn, A. Decian,
J. Fischer, Helv. Chim. Acta 1998, 81, 1–13.
Supporting Information (see footnote on the first page of this arti-
cle): Copies of 1H and 13C NMR spectra for disaccharides α-6 and
β-6.
Acknowledgments
The authors are grateful to V. I. Bragin (NMR Center of GNIIKh-
TEOS) for recording the NMR spectra with a Bruker AVANCE
600 spectrometer. This work was supported by the Russian Foun-
dation for Basic Research (project numbers 08-03-00839 and 11-
03-00918).
[13]
[14]
[1] a) R. Schauer, Biochemistry of Sialic Acid Diversity, in: Carbo-
hydrates in Chemistry and Biology, part II, Biology of Saccha-
rides, vol. 3, Biosynthesis and Degradation of Glycoconjugates
(Eds.: B. Ernst, G. W. Hart, P. Sinaÿ), Wiley-VCH, Weinheim,
Germany, 2000, pp. 227–244; b) T. Angata, A. Varki, Chem.
Rev. 2002, 102, 439–469; c) A. Varki, Nature 2007, 446, 1023–
1029.
[15]
[16]
H. Tanaka, M. Adachi, T. Takahashi, Chem. Eur. J. 2005, 11,
849–862.
Owing to the peculiarities of carbohydrate nomenclature, the
equatorial glycoside corresponds to the α anomer for sialic ac-
ids whereas, for example, in the glucose series the β anomer is
equatorial: a) Pure Appl. Chem. 1996, 68, 1919–2008; b) http://
www.chem.qmul.ac.uk/iupac/2carb; c) http://www.iupac.org/
publications/pac/1996/pdf/6810x1919.pdf.
For the nitrile effect in glycosylation, see: a) J.-R. Pougny, P.
Sinaÿ, Tetrahedron Lett. 1976, 17, 4073–4076; b) R. R.
Schmidt, E. Rücker, Tetrahedron Lett. 1980, 21, 1421–1424; c)
T. Murase, A. Kameyama, K. P. R. Kartha, H. Ishida, M.
Kiso, A. Hasegawa, J. Carbohydr. Chem. 1989, 8, 265–283; d)
A. J. Ratcliffe, B. Fraser-Reid, J. Chem. Soc. Perkin Trans. 1
1990, 747–750; e) R. R. Schmidt, M. Behrendt, A. Toepfer,
Synlett 1990, 694–696; f) J. Braccini, C. Derouet, J. Esnault,
C. H. de Penhoat, J.-M. Mallet, V. Michon, P. Sinaÿ, Car-
bohydr. Res. 1993, 246, 23–41.
When the glycosylation was performed in CH2Cl2, in line with
expectations, the amount of the axial β anomer of 6 substan-
tially increased (α/β = 1:1.3; see entry 10 in Table 1).
H. G. Bazin, Y. Du, T. Polat, R. J. Linhardt, J. Org. Chem.
1999, 64, 7254–7259.
For some examples of the influence of the nature of protecting
groups in the glycosyl acceptor on the yield and stereoselecti-
vity of glycosylation, see: a) Y. J. Lee, K. Lee, E. H. Jung, H. B.
Jeon, K. S. Kim, Org. Lett. 2005, 7, 3263–3266; b) S. Hanash-
ima, K. Sato, Y. Ito, Y. Yamaguchi, Eur. J. Org. Chem. 2009,
4215–4220; c) A. V. Kornilov, A. A. Sherman, L. O. Kononov,
A. S. Shashkov, N. E. NifantЈev, Carbohydr. Res. 2000, 329,
717–730.
[2] For recent reviews, see: a) G.-J. Boons, A. V. Demchenko,
Chem. Rev. 2000, 100, 4539–4565; b) D. K. Ress, R. J. Lin-
hardt, Curr. Org. Synth. 2004, 1, 31–46; c) H. Ando, A. Im-
amura, Trends Glycosci. Glycotechnol. 2004, 16, 293–303; d) C.
DeMeo, in: Frontiers in Modern Carbohydrate Chemistry (Ed.:
A. V. Demchenko), ACS Symposium Series 960, American
Chemical Society, Washington, DC, 2007, pp. 118–131; e) C.
De Meo, U. Priyadarshani, Carbohydr. Res. 2008, 343, 1540–
1552; f) H. Ando, Trends Glycosci. Glycotechnol. 2008, 20, 141–
158; g) S. Hanashima, Trends Glycosci. Glycotechnol. 2011, 23,
111–121; h) D. Crich, J. Org. Chem. 2011, 76, 9193–9209.
[3] a) Y. Uchinashi, M. Nagasaki, J. Zhou, K. Tanaka, K. Fukase,
Org. Biomol. Chem. 2011, 9, 7243–7248, and references cited
therein; b) K.-C. Chu, C.-T. Ren, C.-P. Lu, C.-H. Hsu, T.-H.
Sun, J.-L. Han, B. Pal, T.-A. Chao, Y.-F. Lin, S.-H. Wu, C.-H.
Wong, C.-Y. Wu, Angew. Chem. Int. Ed. 2011, 50, 9391–9395,
and references cited therein; c) B. N. Harris, P. P. Patel, C. P.
Gobble, M. J. Stark, C. De Meo, Eur. J. Org. Chem. 2011,
4023–4027, and references cited therein.
[17]
[18]
[19]
[20]
[4] D. Crich, B. Wu, Org. Lett. 2008, 10, 4033–4035, and references
cited therein.
[5] a) A. A. Sherman, O. N. Yudina, B. S. Komarova, Y. E. Tsvet-
kov, S. Iacobelli, N. E. Nifantiev, Synthesis 2005, 1783–1788;
b) S. Hanashima, B. Castagner, D. Esposito, T. Nokami, P. H.
Seeberger, Org. Lett. 2007, 9, 1777–1779, and references cited
therein.
[6] L. O. Kononov et al., unpublished results.
[7] For an example of the successful use of multivariative experi-
mental design for the optimization of glycosylation, see: F.
Stazi, G. Palmisano, M. Turconi, S. Clini, M. Santagostino, J.
Org. Chem. 2004, 69, 1097–1103.
[8] a) L. O. Kononov, N. N. Malysheva, E. G. Kononova, O. G.
Garkusha, Russ. Chem. Bull. 2006, 55, 1311–1313; b) L. O.
Kononov, N. N. Malysheva, E. G. Kononova, A. V. Orlova,
Eur. J. Org. Chem. 2008, 3251–3255; c) L. O. Kononov, N. N.
Malysheva, A. V. Orlova, Eur. J. Org. Chem. 2009, 611–616,
and references cited therein; d) A. V. Orlova, L. O. Kononov,
B. G. Kimel, I. B. Sivaev, V. I. Bregadze, Appl. Organomet.
Chem. 2006, 20, 416–420.
[21]
[22]
a) N. M. Spijker, C. A. A. van Boeckel, Angew. Chem. 1991,
103, 179; Angew. Chem. Int. Ed. Engl. 1991, 30, 180–183; b) L.
Bohe, D. Crich, Trends Glycosci. Glycotechnol. 2010, 22, 1–15,
and references cited therein.
DLS – also known as quasi-elastic light scattering (QELS) or
photon correlation spectroscopy (PCS) – is concerned with the
investigation of the time correlation of scattered photons (for
details see, for example, ref.[23]). The apparent size (“hydrody-
namic radius”, Rh) of a scattering particle (e.g., supramer) de-
termined by DLS is by definition the radius of a hypothetical
hard sphere that diffuses with the same speed as the particle
Eur. J. Org. Chem. 2012, 1926–1934
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1933

L. O. Kononov et al.
FULL PAPER
under examination. In fact, DLS measures the relaxation time
of a time autocorrelation function g2(τ), or the correlation
time, τ0, which is assumed to be inversely proportional to the
translational diffusion coefficient, D, and proportional to the
hydrodynamic radius Rh (τ0 ~ Rh ~ 1/D). Note that the changes
in supramer structure may not be always accompanied by
change in supramer size. In such a case, DLS may not be able
to provide information on supramers.
[32]
[33]
A similar reaction time (1 h at –35 °C) was reported in the pub-
lication (ref.[15]) describing glycosylations with 1b.
A statement made in 2000 that “glycosylation chemistry is …
not routine, predictable or generally accessible” (ref.[34]) is still
valid.
[34]
[35]
B. Davis, J. Chem. Soc. Perkin Trans. 1 2000, 2137–2160.
The supramer concept considers remote functional or protect-
ing groups as the sites capable of intermolecular interactions,
which may determine the structure of the supramer formed in
each particular case and hence its reactivity. For example, a
significant and unexpected influence of a remote functional
group in aglycon on the reactivity of the hydroxy group of a
trisaccharide glycosyl acceptor in glycosylation during the
chemical synthesis of HNK-1 pentasaccharide (ref.[20c]) can
easily be rationalized by the supramer approach. Owing to the
presence of an extra hydrogen-bonding acceptor site (azide),
which allows the formation of more tightly packed supramers
with a reaction center buried within the supramer core, 2-az-
idoethyl glycoside was much less reactive than the correspond-
ing allyl glycoside of an otherwise identical trisaccharide (gly-
cosylation was five times slower and gave a lower yield of the
pentasaccharide). Similarly, consideration of the possibility of
an additional hydrogen bond formed by an extra amide group
in a spacer allowed a rational explanation of the dramatic de-
pendence of hydrolytic stability of closo-carborane–lactose neo-
glycoconjugates on the nature of the spacer (ref.[8d]).
[23] a) B. Chu, Laser Light Scattering: Basic Principles and Prac-
tice, Academic Press, Boston, MA, 1991, p. 343 ; b) B. J. Berne,
R. Pecora, Dynamic Light Scattering with Applications to
Chemistry, Biology, and Physics, Dover, New York, 2000, p.
376; c) http://www.malvern.com/common/downloads/cam-
paign/MRK656–01.pdf.
[24] Note the bimodal size distribution in solutions of glycosyl ac-
ceptor 5 (Figure 3, b), the slower mode (millisecond range of
correlation time) corresponds to very large supramers with Rh
of 30–60 nm. The slower mode disappears in mixtures of gly-
cosyl donor 1b and glycosyl acceptor 5. For this reason, all
further discussion is limited to the faster mode (microsecond
range of correlation time).
[25] P. Rota, P. Allevi, R. Mattina, M. Anastasia, Org. Biomol.
Chem. 2010, 8, 3771–3776.
[26] D. M. Whitfield, Adv. Carbohydr. Chem. Biochem. 2009, 62, 83–
159.
[27] See the relevant discussions in recent reviews on the mechanism
of chemical glycosylation: a) L. K. Mydock, A. V. Demchenko,
Org. Biomol. Chem. 2010, 8, 497–510; b) L. Bohe, D. Crich, C.
R. Chim. 2011, 14, 3–16.
[28] C.-S. Chao, C.-W. Li, M.-C. Chen, S.-S. Chang, K.-K. T.
Mong, Chem. Eur. J. 2009, 15, 10972–10982.
[29] Our results, taken as a whole, are at variance with a recent
report (see ref.[28]) that an increase in concentration of glycosyl
donor in nitrile solvents leads to lower selectivity for the forma-
tion of equatorial glycosides in the hexose series (see also
Scheme 2).
[30] For a typical example, see: C. Liu, M. R. Richards, T. L.
Lowary, J. Org. Chem. 2010, 75, 4992–5007.
[31] H. Satoh, H. S. Hansen, S. Manabe, W. F. van Gunsteren, P. H.
Hünenberger, J. Chem. Theory Comput. 2010, 6, 1783–1797,
and references cited therein.
[36] a) M. A. Sierra, M. C. de la Torre, Angew. Chem. 2000, 112,
1628; Angew. Chem. Int. Ed. 2000, 39, 1538–1559; b) M. A.
Sierra, M. C. de la Torre, Dead Ends and Detours: Direct Ways
to Successful Total Synthesis, Wiley-VCH, Weinheim, Ger-
many, 2004, p. 290.
[37] This view is consistent with our recent finding that the estima-
tion of the relative reactivity of sialyl donors under competitive
conditions may be incorrect (see ref.[8a,8b]).
[38] a) H. Hori, T. Nakajima, Y. Nishida, H. Ohrui, H. Meguro,
Tetrahedron Lett. 1988, 29, 6317–6320; b) S. Prytulla, J. Laut-
erwein, M. Klessinger, J. Thiem, Carbohydr. Res. 1991, 215,
345–349.
Received: November 7, 2011
Published Online: February 13, 2012
1934
www.eurjoc.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2012, 1926–1934