Unique self-anhydride formation in the degradation of cytidine-5'- monophosphosialic acid (CMP-Neu5Ac) and cytidine-5'-diphosphosialic acid (CDP-Neu5Ac) and its application in CMP-sialic acid analogue synthesis

Article abstract of DOI:10.1002/chem.201003387

Sialyloligosaccharides are synthesised by various glycosyltransferases and sugar nucleotides. All of these nucleotides are diphosphate compounds except for cytidine-5'-monophosphosialic acid (CMP-Neu5Ac). To obtain an insight into why cytidine-5'-diphosphosialic acid (CDP-Neu5Ac) has not been used for the sialyltransferase reaction and why it is not found in biological organisms, the compound was synthesised. This synthesis provided the interesting finding that the carboxylic acid moiety of the sialic acid attacks the attached phosphate group. This interaction yields an activated anhydride between carboxylic acid and the phosphate group and leads to hydrolysis of the pyrophosphate linkage. The mechanism was demonstrated by stable isotope-labelling experiments. This finding suggested that CMP-Neu5Ac might also form the corresponding anhydride structure between carboxylic acid and phosphate, and this seems to be the reason why CMP-Neu5Ac is acid labile in relation to other sugar nucleotides. To confirm the role of the carboxylic acid, CMP-Neu5Ac derivatives in which the carboxylic acid moiety in the sialic acid was substituted with amide or ester groups were synthesised. These analogues clearly exhibited resistance to acid hydrolysis. This result indicated that the carboxylic acid of Neu5Ac is associated with its stability in solution. This finding also enabled the development of a novel chemical synthetic method for CMP-Neu5Ac and CMP-sialic acid derivatives. Copyright

Full text of DOI:10.1002/chem.201003387

FULL PAPER
DOI: 10.1002/chem.201003387
Unique Self-Anhydride Formation in the Degradation of Cytidine-5’-
monophosphosialic Acid (CMP-Neu5Ac) and Cytidine-5’-diphosphosialic
Acid (CDP-Neu5Ac) and its Application in CMP-sialic Acid Analogue
Synthesis
Yasuhiro Kajihara,*[a] Sachiko Nishigaki,[b] Daisuke Hanzawa,[b] Go Nakanishi,[b]
Ryo Okamoto,[b] and Naoki Yamamoto[b]
Abstract: Sialyloligosaccharides are
synthesised by various glycosyltransfer-
ases and sugar nucleotides. All of these
nucleotides are diphosphate com-
pounds except for cytidine-5’-mono-
phosphosialic acid (CMP-Neu5Ac). To
obtain an insight into why cytidine-5’-
diphosphosialic acid (CDP-Neu5Ac)
has not been used for the sialyltransfer-
ase reaction and why it is not found in
biological organisms, the compound
was synthesised. This synthesis provid-
ed the interesting finding that the car-
boxylic acid moiety of the sialic acid at-
tacks the attached phosphate group.
This interaction yields an activated an-
hydride between carboxylic acid and
the phosphate group and leads to hy-
drolysis of the pyrophosphate linkage.
The mechanism was demonstrated by
stable isotope-labelling experiments.
This finding suggested that CMP-
Neu5Ac might also form the corre-
sponding anhydride structure between
carboxylic acid and phosphate, and this
seems to be the reason why CMP-
Neu5Ac is acid labile in relation to
other sugar nucleotides. To confirm the
role of the carboxylic acid, CMP-
Neu5Ac derivatives in which the car-
boxylic acid moiety in the sialic acid
was substituted with amide or ester
groups were synthesised. These ana-
logues clearly exhibited resistance to
acid hydrolysis. This result indicated
that the carboxylic acid of Neu5Ac is
associated with its stability in solution.
This finding also enabled the develop-
ment of a novel chemical synthetic
method for CMP-Neu5Ac and CMP-
sialic acid derivatives.
Keywords: anhydrides · hydrolysis ·
isotopic
labeling
·
reaction mechanisms · sialic acids
Introduction
sugars, lipids and proteins, a process that involves the re-
lease of a pyrophosphate group, in the form of ADP, for ex-
ample.[1] The biosynthesis of oligosaccharides, which takes
place in the endoplasmic reticulum (ER) and the Golgi ap-
paratus,[2] also employs sugar nucleotides containing pyro-
phosphate as a leaving group. During the biosynthesis of oli-
gosaccharides undergoing sequential elongation, glycosyl-
transferases are responsible for the transfer of sugar resi-
dues from sugar nucleotides, such as uridine-5’ -diphospho-
The mechanisms of enzymatic reactions have been exten-
sively studied. Some enzymes utilise an active precursor,
such as a pyrophosphate group, together with metal ions,[1]
phosphoenolpyruvate, thioesters or acid anhydride to form
the transition state.[1] Of these precursors, pyrophosphate is
a frequently encountered biological leaving group and is ob-
served as a by-product in a wide range of reactions in a vari-
ety of biosynthetic pathways. Adenosine-5’ -triphosphate
(ATP) is an essential energy source for cells and it is also
used in the phosphorylation of natural compounds such as
galactose
(UDP-Gal),
UDP-GlcNAc,
guanosine-5’-
diphosphomannose (GDP-Man) and GDP-fucose, to the
corresponding oligosaccharyl acceptors through glycosyl
linkages.[2] All of these building blocks (Leloir donors)[3] are
diphosphate compounds, with the exception of cytidine-5’-
monophosphosialic acid (CMP-Neu5Ac).[2] We were inter-
ested in why only the monophosphate form of the sialic
acid, rather than cytidine-5’ -diphosphosialic acid (CDP-
Neu5Ac), is used as the activated precursor of glycoconju-
gate syntheses. CDP-Neu5Ac is not found in biological or-
ganisms, which suggested an investigation into the chemical
behaviour of CDP-Neu5Ac through chemical synthesis.
During this research, we found that the synthesised form of
CDP-Neu5Ac exhibits very interesting chemical behaviour,
displaying a self-activating process during its decomposition.
This interesting behaviour prompted us to study why CMP-
[a] Prof. Dr. Y. Kajihara
Department of Chemistry
Graduate School of Science, Osaka University
1-1 Machikaneyama, Toyonaka 560-0043 (Japan)
Fax : (+81)6850-5382
E-mail: kajihara@chem.sci.osaka-u.ac.jp
[b] S. Nishigaki, D. Hanzawa, G. Nakanishi, Dr. R. Okamoto,
Dr. N. Yamamoto
International Graduate School of Arts and Sciences
Yokohama City University
22-2, Seto, Kanazawa-ku, Yokohama 236-0027 (Japan)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201003387.
Chem. Eur. J. 2011, 17, 7645 – 7655
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7645

Neu5Ac is an acid-labile compound. This finding also led to
the discovery of a new chemical synthetic route to the bio-
logically important CMP-Neu5Ac and its derivatives. Here
we describe the chemistry and synthesis of active com-
pounds consisting of biologically important sugars nucleo-
tides based on a serendipitous finding of compound behav-
iour.
a result, CDP-Neu5Ac (6) could not be isolated, due to its
lability.
Chemical behaviour of CDP-Neu5Ac: The hydrolysis of the
pyrophosphate linkage was reproducible and was confirmed
by monitoring of the 1H and 31P NMR spectra (Figure 1).
CDP-Neu5Ac methyl ester (5) was dissolved in D2O and its
saponification was then examined by the addition of NaOD
(0.5m). Saponification was complete within the first 5 min
(Figure 1b and Figure 1g) and the pyrophosphate linkage
was then gradually hydrolysed to afford Neu5Ac-2-O-phos-
phate (10) and CMP (8), as shown in Figure 1c, d, e, h, i
and j. As far as was investigated, the CDP-Neu5Ac methyl
ester (5) was stable under neutral and weakly acidic condi-
tions, but was very labile under basic conditions because
saponification allowed conversion into the labile form 6. In
contrast, CDP-Neu5Ac (6) was found to be very labile
under both acidic and basic conditions (Figure S1 in the
Supporting Information).
Results and Discussion
Chemical synthesis of CDP-Neu5Ac: To obtain artificial
CDP-Neu5Ac (6, Scheme 1), its chemical synthesis by use of
a conventional imidazolide phosphate-coupling strategy was
examined.[4] The CMP-triacetyl derivative 1[5] was converted
into its imidazolide derivative 2 by treatment with N,N’-car-
bonyldiimidazole, and coupling between 2 and the Neu5Ac
derivative 3[6] was then performed to obtain the protected
CDP-Neu5Ac derivative 4.
Removal of the acetate and methyl ester with sodium
methoxide was then examined (Scheme 1). During this final
deprotection step, we unexpectedly first obtained CDP-
Neu5Ac methyl ester (5) as the main product (62% yield,
three steps). We then examined saponification of this
methyl ester. After saponification, we found that the pyro-
phosphate linkage in CDP-Neu5Ac (6) was easily hydro-
lysed to yield Neu5Ac-2-O-phosphate (10) and CMP (8). As
Demonstration of the hydrolysis of the pyrophosphate link-
age by an 18O-labelling technique: From the experimental
results shown in Scheme 1 and Figure 1, we propose a
unique mechanism of hydrolysis in which nucleophilic
attack of the carboxylic acid moiety on a phosphate group
triggers the cleavage of the pyrophosphate linkage. As
shown in Scheme 1, this proposed mechanism results in the
five-membered ring compound
7. If the five-membered ring is
essential for this hydrolysis re-
action, the oxygen of the car-
boxylic acid (shown in bold in
7) should migrate from the car-
boxylic acid to the phosphate
group through the intermediate
anhydride 9. Therefore, to in-
vestigate this mechanism, we
examined 18O labelling[7] of the
oxygen of the carboxylic acid
(Scheme 2). Before 18O label-
ling, we examined saponifica-
tion with 18O-labelled Na18OH
as a preliminary reaction. As
shown in Figure 1b (reaction
time: 5 min) and Scheme 3a
(the control experiment: 15!
16), saponification with 18O-la-
belled (45%) NaOH afforded
the Neu5Ac derivative 16 in
quantitative yield and the
13C NMR spectrum of which
confirmed that about 40–45%
Scheme 1. Synthesis of CDP-Neu5Ac methyl ester and CDP-Neu5Ac and its presumed hydrolysis reaction
mechanism to yield CMP (8) and Neu5Ac-2-O-phosphate (10) through migration of the oxygen shown in
bold. a) N,N’-Carbonyldiimidazole, DMF, RT; b) 3, DMF, RT; c) NaOMe, MeOH, RT, 62% (three steps); the
first experiment employed MeOH/H2O (1:1) conditions and this condition fortunately yielded CDP-Neu5Ac-
methyl ester; d) NaOH (0.5m or 0.3m; these conditions afforded CMP (8) and Neu5Ac-2-phosphate (10),
quantitatively).
of the oxygen of the carboxylic
acid was labelled with 18O
(Scheme 3a).
In view of these results, we
performed 18O-labelling experi-
7646
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 7645 – 7655

Self-Anhydride Formation in the Degradation of CMP-Neu5Ac and CDP-Neu5Ac
FULL PAPER
and the Neu5Ac-2-O-phosphate 14a via the anhydrides 13a
and 13b. Both the 31P and 13C NMR spectra of a solution
containing 14a showed approximately 20% of the phos-
phate group oxygen labelled with the 18O atoms (Sche-
me 2I). When the carboxylic acid in 12 was 45% labelled
with 18O atoms, the 18O atoms immediately equilibrated be-
tween the two oxygen atoms of the carboxylic acid. The
phosphate in the Neu5Ac-2-O-phosphate 14a was therefore
about 20% labelled by migration through 13b, and the car-
boxylic acid in 14a was also about 20% labelled through
13a (Scheme 2 (I) and (II); isotope shifts were estimated by
integration of these spectra). Because it was confirmed that
CDP-Neu5Ac methyl ester (5), Neu5Ac-2-O-phosphate (10,
data not shown) and simple sialic acid (17) were not labelled
with 18O in the buffer solution (Scheme 3b and c), we con-
cluded that there were no 18O-exchange reactions between
the substrates and the 18O-labelled solution used. In addition
to these findings, because almost no CDP-Neu5Ac decom-
posed during the first 5 min of saponification (Figure 1b),
we concluded that 45% of the 18O-labelled CDP-Neu5Ac
(11) was hydrolysed in a 2.3% 18O-labelled H2O solution
(50 mm NaOH solution) and these results, shown in
Scheme 2, clearly indicated that the 18O atoms had migrated
from the carboxylic acid group to the phosphate group via
the five-membered intermediate 13b. The reason for CDP-
Neu5Ac methyl ester (5) being stable under acidic condi-
tions is thought to be a protective effect conferred by the
methyl ester group, which interferes with nucleophilic attack
of carboxylate on the phosphate group.
18O-Labelling experiments with CDP-Neu5Ac under acidic
conditions: Having demonstrated the interaction between
the carboxylic acid and the phosphate group under basic
conditions (Scheme 2), we further investigated the hydroly-
sis mechanism under acidic conditions similar to the pH
value suitable for the sialyltransferase reaction (Schemes 4
and 5). Scheme 4 displays the 18O migration pathway from
the carboxylic acid of CDP-Neu5Ac (11) to the phosphate
groups in 14b and 14c, whereas Scheme 5 displays the reac-
tion pathway under the 18O-labelled acidic water conditions.
In the case shown in Scheme 4, saponification was per-
formed with use of 18O-labelled NaOH (45% 18O) and the
solution was then diluted with sodium cacodylate buffer
(pH 5.0, 500 mm) to adjust the pH of the solution and the
degree of 18O labelling. The pH and 18O labelling in the so-
lution containing CDP-Neu5Ac (11) were adjusted to
pH 5.7 and 2.5% 18O labelling. In this solution (pH 5.7 and
2.5% 18O-labelled H2O), CDP-Neu5Ac (11) was gradually
converted into the Neu5Ac-2-O-phosphates 14b and 14c.
After the completion of the hydrolysis, the 13C and 31P NMR
spectra of a solution containing the Neu5Ac-2-O-phosphates
14b and 14c were measured (Scheme 4I and II). The isotope
shifts due to 18O labelling observed in the 31P NMR and
13C NMR spectra were about 17 and 40%, respectively (iso-
tope shifts were estimated by integration of these spectra).
The product 14b appeared to originate from both 13a and
13b by attack of water on the phosphate group. Product
Figure 1. Monitoring of hydrolysis of the pyrophosphate linkage in CDP-
Neu5Ac (6) after saponification of the CDP-Neu5Ac methyl ester 5.
ments with CDP-Neu5Ac methyl ester (5). Saponification of
CDP-Neu5Ac methyl ester (5) with Na18OH (0.5m, 45%
18O labelling) yielded the 18O-labelled CDP-Neu5Ac 11
(Scheme 2), in which about 45% of the total oxygen in the
carboxylic acid was labelled. Because the 18O-labelled CDP-
Neu5Ac 11 is labile, the degree of 18O labelling could not be
determined efficiently. After saponification of 5 (5 min), the
labelled reaction solution was immediately diluted to give
18O-labelled (2.3%) content by 20-fold dilution in H216O.
After this treatment, the 18O-labelled CDP-Neu5Ac 12
(45% 18O labelling) was gradually converted into CMP (8)
Chem. Eur. J. 2011, 17, 7645 – 7655
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7647

Y. Kajihara et al.
Scheme 2. Monitoring of hydrolysis of the pyrophosphate linkage by the 18O-labelling method. Reagents and conditions: a) 18O-labelled (45%) NaOH
(0.5m), 5 min; b) dilution of the solution used under the conditions of (a) to generate 18O-labelled solution (2.3%). Saponification (conditions a) of
CDP-Neu5Ac methyl ester (5) was performed in 18O-labelled (45%) NaOH (0.5m, 80 mL). After 5 min, H216O (720 mL) and NaOD (50 mm, 800 mL)
were added and the reaction mixture was adjusted to H218O (2.3%) and NaOD (50 mm) solution. After 3 h, all of the CDP-Neu5Ac had been hydrolysed
and the degree of 18O labelling was estimated from the 31P and 13C NMR spectra (I and II). A small coupling constant in the 13C NMR signal (II) was ob-
served, as indicated by JC,P. Approximately 20% of the oxygen in the phosphate group and 20% of the oxygen in the carboxylic acid of 14a were la-
belled. After the labelling of the carboxylic acid of 11 by H218O (ca. 45%), half of the 18O atoms were equilibrated between the carbonyl oxygen and hy-
droxy oxygen of 12. Therefore, about 20% of the 18O atoms had migrated from the carboxylic acid to the phosphate group. The blacked-in oxygen indi-
cates 18O.
14c, though, appeared to be derived from 13b, through
attack on the carboxylic acid by water. From these NMR
data, it appeared that water attacked the phosphate groups
in 13a and 13b as the major reaction pathway. In the case
shown in Scheme 5, saponification was performed by use of
Na16OH and the solution conditions were then made acidic
with 18O-labelled (42%) sodium cacodylate buffer (pH 5.0,
500 mm) to adjust the pH of the solution and the degree of
18O labelling. The pH and 18O percentage in the solution
containing CDP-Neu5Ac (6) were adjusted to pH 5.7 and
42%. In this solution (pH 5.7 and 42% 18O degree), CDP-
Neu5Ac (6) was gradually converted into the Neu5Ac-2-O-
phosphates 14b and 14c. After isolation of the Neu5Ac-2-
O-phosphates 14b and 14c, the 13C and 31P NMR spectra
were measured (Scheme 5I and II). The isotope shifts due to
18O labelling observed in the 31P NMR and 13C NMR spectra
were about 40% and less than 10%, respectively (isotope
shift was estimated by the integration of these spectra). The
product 14b appears to originate from 9 through attack on
the carboxylic acid group by the 18O-labelled water. Product
14c, though, appears to originate from 9 through attack on
the phosphate group by the 18O-labelled water. From these
NMR data (Scheme 5I and II), water appeared to attack the
phosphate group of 9 as the major reaction pathway.
These isotope-labelling experiments under acidic condi-
tions also show that CDP-Neu5Ac is hydrolysed via a five-
membered ring intermediate, but that a water molecule at-
tacks the phosphate group in the anhydride intermediates
13a, 13b and 9 to give the Neu5Ac-2-O-phosphates 14b and
14c (Schemes 4 and 5), whereas under basic conditions a
water molecule attacks the carbonyl group in the anhydride
intermediates 13a, 13b and 9 (Scheme 2). These results are
clearly supported by both the 31P and 13C NMR data
(Schemes 2, 4 and 5I and II). As described above, the results
of these chemical analyses indicate that CDP-Neu5Ac (6) is
excessively labile due to the self-degradation process that
occurs through the formation of a five-membered ring. Al-
though there is thus no evidence as to whether CDP-
Neu5Ac (6) actually exists or not, these experimental results
do suggest a reason why it has not been found in biological
organisms.
7648
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 7645 – 7655

Self-Anhydride Formation in the Degradation of CMP-Neu5Ac and CDP-Neu5Ac
FULL PAPER
Insight into the mechanism of
hydrolysis of CMP-Neu5Ac:
The results of the isotope-la-
belling experiments investigat-
ing the chemical properties of
CDP-Neu5Ac enabled us to
obtain insights into the hydrol-
ysis mechanism of CMP-
Neu5Ac. It is known that
CMP-Neu5Ac is labile under
acidic conditions and that acid
hydrolysis
easily
affords
Neu5Ac and CMP.[8,9] CMP-
Neu5Ac appears to be more
labile than other Leloir donors,
which are relatively stable
under mildly acidic condi-
tions.[8] A hydrolysis mecha-
nism for CMP-Neu5Ac has
therefore been proposed,[9] but
the activation process during
hydrolysis has not been dem-
onstrated. In this context, we
Scheme 3. Control labelling experiments with the Neu5Ac derivative 15 and 17 under the 18O labelling solu-
tion conditions. a) 18O-labelled (40%) NaOH (0.5m) afforded 18O-labelled (40%) sialic acid derivative 16;
b) 18O-labelled (43%) NaOH (0.5m) did not afford the 18O-labelled sialic acid derivative 18, and 18O-labelled
(45%) sodium cacodylate buffer (pH 5.0) did not afford the 18O-labelled sialic acid derivative 19.
Scheme 4. Monitoring of hydrolysis by the 18O-labelling method. Reagents and conditions: a) 18O-labelled (45%) NaOH (0.5m); b) dilution of solution
used under conditions (a) to generate 18O-labelled (2.5%) solution (pH 5.7). After saponification of 5 with 18O-labelled (45%) NaOH (0.5m, 40 mL),
sodium cacodylate buffer solution (500 mm, 1280 mL, pH 5.0) was added and the pH value and the degree of 18O labelling were adjusted to 5.7 and 2.5%,
respectively. The solution was further incubated at room temperature. After hydrolysis of the pyrophosphate 11 (1 h), the 13C and 31P NMR spectra, (I)
and (II), of the mixture of 14b and 14c were measured to estimate the degree of 18O labelling. A small coupling constant in the 13C NMR signal, (II),
was observed as indicated by JC,P. The degrees of labelling of 14b and 14c corresponded to less than 10% of the oxygen in the phosphate group and
about 40% of the oxygen in the carboxylic acid. These data indicate that the water molecule (H216O) attacked the phosphate groups in 13a and 13b.
Blacked-in oxygen indicates 18O.
Chem. Eur. J. 2011, 17, 7645 – 7655
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7649

Y. Kajihara et al.
Scheme 5. Monitoring of hydrolysis by the 18O-labelling method. Reagents and conditions: a) NaOH (0.5m); b) dilution of the solution used to generate
18O-labelled (42%) sodium cacodylate buffer (pH 5.0, 500 mm). After saponification of CDP-Neu5Ac methyl ester (5) in NaOH (0.5m, 60 mL) for 5 min,
this solution was diluted with sodium cacodylate buffer (pH 5.0, 500 mm, 180 mL) and H218O (85%, 240 mL) to adjust both the pH (5.7) and the degree of
labelling (42.5% H218O). NMR analysis of the products was performed by the same procedures as described in Scheme 4. Approximately 40% of the
oxygen in the phosphate group of 14b and less than 10% of the oxygen in the carboxylic acid group of 14c were labelled.
propose a new hypothesis relating to the activation mecha-
nism for the hydrolysis of CMP-Neu5Ac. We hypothesise
that CMP-Neu5Ac also forms the corresponding five-mem-
bered intermediate 22 (Scheme 6) during the acid hydrolysis
reaction to yield Neu5Ac (26) and CMP (8). It appears that
the carboxylic acid attacks the phosphate group and then
the five-membered ring 22 is formed.
method.[10,11] These synthetic results are explained in
Scheme S1 in the Supporting Information. The stabilities of
the synthesised CMP-Neu5Ac analogues 29 and 30 under
acidic conditions were evaluated. Each substrate—CMP-
Neu5Ac, its methyl ester derivative 29 and its amide deriva-
tive 30—was dissolved in AcOH (60%) and the solution
was divided into two portions. The solutions were individu-
ally incubated at room temperature and at 508C for 30 min.
After these acidic treatments, the Neu5Ac 1-methyl ester
and Neu5Ac 1-amide were estimated by integration of the
Additional protonation at the carbonyl oxygen of 23
might also afford the oxocarbenium ion intermediate 25
through the intermediate 24 to afford Neu5Ac (26). During
this process, if the formation of the five-membered ring is
indeed essential, protection of the carboxylic acid should
result in a slowing down of the hydrolysis reaction velocity.
With regard to other hypothetical hydrolysis mechanisms,
one would be the protonation of the sp2 oxygen of the phos-
phate group 20 to generate the cation intermediate 24 and
subsequent generation of the oxocarbenium intermediate
25. If this simple hydrolysis mechanism were employed, the
carboxylic acid of Neu5Ac would not be involved in the hy-
drolysis reaction pathway, and CMP-Neu5Ac analogues
such as the methyl ester 29 and amide 30 derivatives
(Figure 2) should be hydrolysed by simple acid hydrolysis.
We therefore hypothesised that an amide or ester CMP-
Neu5Ac analogue might display resistance to acid hydroly-
sis, so we synthesised the CMP-Neu5Ac methyl ester and
amide derivatives 29 and 30.[10]
1
H-3eq signals in the H NMR spectra relative to those of the
remaining starting materials 29 and 30. The results are
shown in Figure 2, the vertical axis representing CMP-
Neu5Ac and its derivatives remaining after the acidic treat-
ment. As expected, the two derivatives 29 and 30 are both
much more stable than CMP-Neu5Ac, even under acidic
conditions at 508C (Figure 2). The room temperature acid
hydrolysis experiments did not result in hydrolysis of the
CMP-Neu5Ac methyl ester 29 or its amide 30, although
70% of the CMP-Neu5Ac was hydrolysed. The intermediate
24 shown in Scheme 6 to be generated by simple protona-
tion at the sp2 oxygen of the phosphate group in 20, the
CMP-Neu5Ac ester 29 and amide 30 derivatives would be
expected to undergo hydrolysis. Because the ester and
amide derivatives 29 and 30 did not exhibit hydrolysis,
simple activation through protonation at the sp2 oxygen of
the phosphate group in 20 seems unfeasible, so it is pre-
sumed that the five-membered ring intermediate is more
likely. In terms of the acetic acid (pKa =4.76[12,13]) used in
the experiments, the pH values of the solutions measured
during the course of the experiments were about 2–2.5. The
phosphate group (pKa =2.12[12,13]) is therefore able to adopt
Synthesis of the CMP-Neu5Ac methyl ester and amide de-
rivatives 29 and 30 and evaluation of their stabilities under
acidic conditions: For the syntheses of the CMP-Neu5Ac
methyl ester and amide derivatives 29 and 30 we employed
an amidite coupling strategy based on a previously reported
7650
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 7645 – 7655

Self-Anhydride Formation in the Degradation of CMP-Neu5Ac and CDP-Neu5Ac
FULL PAPER
proton can interact with the
sialoside oxygen as
membered ring in an intramo-
lecular catalytic manner
a five-
(Scheme 6: sialoside 27),[15]
whereas the esterified form of
the sialoside cannot undergo
protonation of the sialoside
oxygen. Therefore, the suscept-
ibility of the sialoside to acid
hydrolysis could be dependent
on the interaction with the car-
boxylic acid. In the case of
CMP-Neu5Ac, however, the
protonation of the phosphate
oxygen by the carboxylic acid
would seem to be unlikely, be-
cause of the requirement for
the seven-membered ring in-
termediate 28 (Scheme 6).
The five-membered ring for-
mation in CMP-Neu5Ac: The
results shown in Scheme 6 also
suggest that the susceptibility
of CMP-Neu5Ac (21) to hy-
drolysis is due to the formation
of the five-membered inter-
mediate 22. It was not possible
to demonstrate the existence
of the five-membered ring,
however, because the 18O
atoms do not migrate from the
carboxylate to the phosphate
group in this activation mecha-
Scheme 6. Proposed hydrolysis mechanism of CMP-Neu5Ac yielding the oxocarbenium ion intermediate 25.
two distinct forms, that is, the phosphate diester 20 and its
dissociated form 21 (Scheme 6) in a about a 1:1 ratio, and
an equilibrium state exists between the two forms. Mean-
while, a carboxylic acid such as 21 exists, rather than its dis-
sociated carboxylate form. The dissociated phosphate 21
therefore takes the proton from the carboxylic acid, and the
carboxylate of Neu5Ac can then attack the phosphorus
atoms of the phosphate form to adopt the five-membered
ring structure 22 (Scheme 6). This formation under weakly
acidic conditions might be suitable to afford the subsequent
oxocarbenium intermediate 25 through the cation intermedi-
ate 24 formed by the protonation of the carbonyl oxygen
(lactone 23) of the five-membered ring. We concluded that
this is the reason why CMP-Neu5Ac is hydrolysed under
weakly acidic conditions. Because acetic acid is weakly
acidic relative to phosphate, protonation at the sp2 oxygen
of the phosphate group in 20 may not occur. In the case of a
simple sialoside (the sialyloligosaccharide 27), it is known
that a sialoside linkage is more labile towards acid treatment
than other glycosyl bonds, but the ester form of sialoside ex-
hibited resistance to treatment trifluoroacetic acid (TFA).[14]
From these results we hypothesise that the carboxylic acid
nism (Scheme 6), as described in the case of CDP-Neu5Ac
(Schemes 2, 4 and 5), and so we attempted to demonstrate
the activation mechanism by another method, as shown in
Scheme 7. Because a water molecule attacks the phosphate
group of the five-membered intermediate under acidic con-
ditions as the major reaction pathway (Schemes 4 and 5), we
examined the formation of the anhydride derivative 32 from
the Neu5Ac-2-O-phosphate 31, and then added cytidine
(33) to generate CMP-Neu5Ac through attack by the 5-hy-
droxy group of cytidine on the phosphate group of 34
(Scheme 7). If this new synthetic route were to produce
CMP-Neu5Ac (21), this experiment would provide support
for the hypothesis that CMP-Neu5Ac can employ the five-
membered intermediates 22 (Scheme 6) and 34 (Scheme 7).
The experiment was performed and monitored by 31P NMR
spectroscopy (Scheme 7I–V). When the condensation re-
agents diisopropylcarbodiimide and hydroxybenzotriazole
(DICPI and HOBt) were added to the Neu5Ac-2-O-phos-
phate 31, a new 31P signal was observable at d=ꢀ2.4 ppm,
the signal reaching its maximum intensity after 1 h (Sche-
me 7II). This signal decreased concomitantly with an in-
crease in a new signal at d=ꢀ4.1 ppm (Scheme 7III). These
Chem. Eur. J. 2011, 17, 7645 – 7655
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7651

Y. Kajihara et al.
changes correspond to the addition of cytidine to give the
desired CMP-Neu5Ac derivative 35. Finally, saponification
yielded CMP-Neu5Ac (21), which was observable as a
signal at d=ꢀ4.3 ppm (Scheme 7V). After isolation it was
confirmed that the d=ꢀ4.1 and ꢀ3.0 ppm signals corre-
sponded to the CMP-Neu5Ac derivative 35 and the sub-
strate 31, respectively. It was also concluded that a further
signal at ꢀ2.4 ppm corresponded to the anhydride 32 or the
activated form of 31 (Scheme 8; intermediate A or B). The
1H NMR data for the product 21 were identical to those for
authentic CMP-Neu5Ac, and sialyl transfer activity to
LacNAc by rat liver a-2,6-sialyltransferase with this pure 21
was also demonstrated. An additional important result of
this experiment was with the use of the esterified form of
the Neu5Ac-2-O-phosphate derivative 3 rather than 31. As
we had expected, this compound 3 did not yield the desired
CMP-Neu5Ac product under the same conditions, which in-
dicated that the five-membered ring was indeed essential for
the formation of CMP-Neu5Ac.
This experiment thus not only demonstrated a new
method of synthesising CMP-Neu5Ac, but also provided
evidence that the formation of this five-membered ring does
not impose a structural strain on the CMP-Neu5Ac skele-
ton. From these results, we propose that CMP-Neu5Ac un-
derwent spontaneous self-degradation by the mechanism
shown in Scheme 6. In an investigation of the hydrolysis
mechanism of CMP-Neu5Ac, Horrenstainꢁs group conduct-
ed isotope experiments based on other possible hydrolysis
mechanisms.[9b,c,d] In their isotope experiments, isotope label-
Figure 2. The stabilities of the CMP-Neu5Ac methyl ester 29 and its
amide analogue 30 under acidic conditions. The vertical axis shows the
CMP-Neu5Ac derivative remaining after acid hydrolysis. Each substrate,
CMP-Neu5Ac, its methyl ester derivative 29 and its amide derivative 30,
was dissolved in AcOH (60%) and the solution was divided in two. The
solutions were individually incubated at room temperature (dark grey)
and at 508C (light gray) for 30 min. All compounds yielded Neu5Ac and
its derivatives (Neu5Ac 1-methyl ester and Neu5Ac 1-amide) and these
compounds were estimated from the 1H NMR spectra, by focusing on
each H-3eq signal. All results represent averages of at least two experi-
ments.
Scheme 7. Synthesis of CMP-Neu5Ac from the starting material 31 by the new synthetic method. Reagents and conditions: a) HOBt, diisopropylcarbo-
diimide, DMF, RT, cytidine (33), 4.5 h; b) NaOH (0.2m), RT, 15 h. I)–V) Monitoring of the reactions by 31P NMR spectroscopy. V) The 31P NMR spec-
trum of the reaction mixture after deprotection. Use of the Neu5Ac-2-O-phosphate derivative 3, rather than 31, did not yield the desired CMP-Neu5Ac
product under the same conditions. 1H NMR data for product 21 was identical to those for authentic CMP-Neu5Ac (see the Supporting Information).
Sialyl transfer activity to LacNAc by rat liver a-2,6-sialyltransferase was also demonstrated by use of this pure form of 21.
7652
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 7645 – 7655

Self-Anhydride Formation in the Degradation of CMP-Neu5Ac and CDP-Neu5Ac
FULL PAPER
tion of the oxygen of the phos-
phate group in 21 (Scheme 6)
is a trigger for the cleavage of
the C2-phosphate bond and
then the formation of the oxo-
carbenium intermediate 25, the
CMP-Neu5Ac derivatives 29
and 30 should easily afford
Neu5Ac (26) under acidic con-
ditions. Therefore, it is con-
cluded that the formation of
this five-membered ring pres-
ents a large advantage in the
activation of CMP-Neu5Ac.
Synthesis of CMP-Neu5Ac an-
alogues by the new synthetic
method: We examined the acti-
vation
process
occurring
during the hydrolysis of CMP-
Neu5Ac as shown in Scheme 7,
and this insight into the activa-
tion mechanism enabled us to
find a new synthetic route for
CMP-Neu5Ac. This synthetic
route has the advantage of
omitting the need of a protect-
ing group for cytidine.[5,11,17]
We therefore further investi-
gated the condensation reac-
tion with use of several re-
agents (Table S1 in the Sup-
porting Information). Of the
reagents used, PyBOP or diiso-
propylcarbodiimide conditions
afforded CMP-Neu5Ac in
moderate yields. However, the
condensation reaction should
be performed at a low temper-
ature (such as ꢀ158C) because
the products, such as 35, exhib-
ited instability at room temper-
ature and gradually decom-
posed. During this reaction, a
Scheme 8. Plausible reaction pathway after activation of the Neu5Ac-phosphate 31.
ling of the carbon atoms of the Neu5Ac-carboxylate group
revealed a large isotope effect during the hydrolysis of
CMP-Neu5Ac.[9b,c,d,16] This large isotope effect, once suitably
corrected,[16] was understood to be due to the fluctuation of
the carboxylic acid during hydrolysis. However, this isotope
effect also seems to support our hypothesis of the attack of
carboxylate so as to form a five-membered ring. The isotope
labelling at the carboxylic acid atoms seemed to slow down
the formation of the five-membered ring. The fact that the
CMP-Neu5Ac methyl ester and amide derivatives 29 and 30
exhibit resistance to acid hydrolysis (Figure 2) demonstrates
that the key step in the hydrolysis reaction is the formation
of the five-membered ring 22 (Scheme 6). If simple protona-
phosphate group can be activated by reagents such as
PyBOP and carbodiimide (intermediate A; Scheme 8), and
an intermolecular attack by HOBt or intramolecular attack
by carboxylate can then be the result (Scheme 8). The inter-
mediate activated by HOBt (intermediate B) seems to be
inactive in terms of reacting with cytidine alcohol. When the
esterified Neu5Ac-phosphate derivative 3 (Scheme 7) was
used, it did not afford the corresponding CMP-Neu5Ac de-
rivative. When this reaction was monitored by 31P NMR
spectroscopy and ESI mass spectroscopy, the corresponding
intermediate C (Scheme 8: m/z: 687.0; 31P NMR d, d=
ꢀ3.58 ppm) activated by HOBt was found, but we con-
firmed that this intermediate C did not convert into the cor-
Chem. Eur. J. 2011, 17, 7645 – 7655
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7653

Y. Kajihara et al.
concentration in vacuo, purification of the residue on a gel permeation
column (Sephadex G15, 1.8 i.d.ꢂ90 cm, H2O) yielded CDP-Neu5Ac
methyl ester (5, 15.0 mg, 62%). H NMR (400.13 MHz, HOD=4.78 ppm,
D2O): d=8.07 (d, J=7.6 Hz, 1H; H-6), 6.23 (d, 1H; H-5), 6.10 (d, J=
4.5 Hz, 1H; H-1’), 4.46 (dd, J=4.5, 4.5 Hz, 1H; H-3’), 4.41 (dd, 1H; H-
2’), 4.27 (d, J=10.3 Hz; H-6’’), 4.06 (dd, J=10.0, 10.0 Hz, 1H; H-5’’), 3.99
(ddd, J=8.7, 2.4, 6.4 Hz, 1H; H-8’’), 3.94 (s, 3H; COOMe), 3.71 (dd, J=
6.4, 11.7 Hz, 1H; H-9’’b), 3.58 (d, J=9.3 Hz, 1H; H-7’’), 2.63 (dd, J=4.8,
13.3 Hz, 1H; H-3’’eq), 2.14 (s, 3H; Ac), 1.85 ppm (ddd, J=5.4, 11.4,
11.4 Hz, 1H; H-3ax); 31P NMR (161.98 MHz, D2O, H3PO4 =0.00 ppm ex-
responding CMP-Neu5Ac derivative 35 and, furthermore,
that it remained in this state until workup. Indeed, activa-
tion of the Neu5Ac-phosphate derivative 3 (Scheme 1) by
the imidazolide method did not afford the desired CDP-
Neu5Ac derivative 4. We therefore examined an alternative
route, that is, the activation of CMP by imidazole to afford
2 and then the CDP-Neu5Ac derivative 4 (Scheme 1), and
this route was successful. From these observations we con-
cluded that the anhydride form 32 is consumed from the
equilibrium state existing between the intermediate B and
the anhydride form 32, and that this anhydride form affords
the CMP-Neu5Ac derivative 35 (Scheme 8). We also exam-
ined the synthesis of CMP-Neu5Ac derivatives under these
conditions, obtaining the CMP-5’’-NH2-sialic acid[18,19] deriv-
ative in moderate yield (Scheme S3 in the Supporting Infor-
mation). In the case of CMP-5’’ -NH2-sialic acid, the amino
group at the 5’’-position can be modified by various conden-
sation reactions.[18] This versatile CMP-sialic acid analogue
can therefore be converted into various other CMP-sialic
acid analogues and also used for the preparation of sialoside
analogues through the sialyltransferase reaction.[20]
1
ternal standard): d=ꢀ11.52 (d, Jp,o,p =20.3 Hz), ꢀ15.88 ppm (d, Jp,o,p
=
20.3 Hz); MS (FAB) calcd for C21H33N4Na2O19 P2: 753.09 [M+H]+;
found: 753.09.
Monitoring of saponification by NMR spectroscopy: Saponification of
CDP-Neu5Ac methyl ester (5) was monitored by measuring 1H and
31P NMR spectra (Figure 1). CDP-Neu5Ac methyl ester (5, 5.0 mg,
6.4 mmol) was dissolved in NaOD/D2O solution (0.5m), and the solution
was immediately analysed by NMR spectroscopy. Pure CDP-Neu5Ac (6,
Figure 1b and g) was obtained within 5 min in the presence of NaOD
(0.5m). In the same experiment with NaOD/D2O solution (0.03m), CDP
was not generated. CMP along with Neu5Ac-2-O-phosphate were the
sole products.
18O atom exchange experiment with CDP-Neu5Ac methyl ester under
basic conditions: (Scheme 2, 5!14a) Saponification of CDP-Neu5Ac
methyl ester (5, 5.0 mg, 6.6 mmol) was performed by the addition of
NaOD/D2O (1.0m, 40 mL) and H218O (90%, 40 mL). After 5 min, D216O
(720 mL) and NaOD (50 mm, 800 mL) were added and the reaction mix-
ture was adjusted to H218O (2.3%) and NaOD solution (50 mm). After
3 h, we confirmed that all CDP-Neu5Ac had hydrolysed and estimated
the degree of 18O labelling by measurement of 31P and 13C NMR spectra.
Approximately 20% of the oxygen in the phosphate group and 20% of
the oxygen in the carboxylic acid of 14a were labelled. After labelling of
carboxylic acid with H218O (45%), half of the 18O atoms were equilibrat-
ed between the carbonyl oxygen and the hydroxy oxygen in 12. There-
fore, about 20% of the 18O atoms migrated from the carboxylic acid to
the phosphate group in 14a. To monitor saponification and hydrolysis by
NMR spectroscopy, deuterated solution was used.
18O atom exchange experiment under acidic conditions: (Scheme 4, 5!
14b and 14c) NaOH (1m, 40 mL) was added to a solution of CDP-
Neu5Ac methyl ester (5, 5.0 mg, 6.6 mmol) in H218O (85%, 40 mL) and
the solution was incubated at room temperature. After 5 min, sodium ca-
codylate buffer solution (500 mm, 1280 mL, pH 5.0) was added and the
pH value and the degree of 18O labelling were adjusted to 5.7 and 2.5%,
respectively. The solution was further incubated at room temperature.
After 1 h, 13C and 31P NMR spectra were measured to estimate the
degree of 18O labelling. The degrees of labelling of 14b and 14c were less
than 10% of oxygen in the phosphate group (Scheme 4I) and about 40%
of oxygen in the carboxylic acid (Scheme 4II). These data indicate that
the water molecule (H216O) attacked the phosphate groups in 13a and
13b. Because CDP-Neu5Ac methyl ester was not decomposed and la-
belled under acidic conditions (Figure S1 in the Supporting Information),
the experiment shown in Scheme 4 (5!14b and 14c) indicates that the
carboxylic acid in CDP-Neu5Ac acted to adopt the anhydride structures
13a and 13b.
Conclusion
Because we were interested in determining why CDP-
Neu5Ac has not been reported in biological organisms, we
undertook synthetic studies of this compound. These studies
resulted in several new findings, including the interesting
chemical behaviour of CDP-Neu5Ac. Our studies with
stable isotope-labelling experiments clearly showed that the
carboxylic acid corresponding to the 1-position of Neu5Ac
attacks the phosphate group attached to the 2-position of
the sialic acid and that this interaction causes the self-degra-
dation of CDP-Neu5Ac to afford the Neu5Ac-phosphate 10
and CMP (8). This finding provides insight into the activa-
tion mechanism of CMP-Neu5Ac in acidic solution. CMP-
Neu5Ac also seems to undergo self-degradation via a phos-
phate and carboxylate anhydride intermediate, similarly to
CDP-Neu5Ac. This may explain why CMP-Neu5Ac is labile
in relation to the other Leloir donors. This finding enabled
the discovery of a convenient synthetic route to valuable
CMP-sialic acid derivatives.
Experimental Section
Saponification with H216O and then hydrolysis with H218O: (Scheme 5,
5!6!14b and 14c) CDP-Neu5Ac methyl ester (5, 5.0 mg, 6.6 mmol)
was dissolved in NaOH (0.5m, 60 mL) and the system was incubated for
5 min at room temperature to allow the saponification to proceed to
completion. Sodium cacodylate buffer (500 mm, pH 5.0, 180 mL) and
H218O (85%, 240 mL) were added to this solution, to adjust both the pH
(5.7) and the degree of labelling (final degree: 42.5% H218O). After 1 h,
all CDP-Neu5Ac was hydrolysed. The degree of labelling was estimated
by measurement of the 13C and 31P NMR spectra. Approximately 40% of
the oxygen in the phosphate groups (Scheme 5I) and less than 10% of
the oxygen in the carboxylic acid groups (Scheme 5II) of 14b and 14c
were labelled. This experiment also indicates that carboxylic acid of
CDP-Neu5Ac acted to adopt an anhydride structure 9.
Synthesis of CDP-Neu5Ac methyl ester (5): The Neu5Ac-2-O-phosphate
3[6] (30 mg, 32 mmol) and the CMP triacetate 1[5] (43 mg, 93 mmol) were
separately co-evaporated five times with DMF. N,N-Carbonyldiimidazole
(30 mg, 186 mmol) was added to a solution of 1 in DMF (1.8 mL) and
this mixture was stirred at room temperature. After 1 h, the mixture was
quenched by the addition of MeOH (50 mL) and then concentrated in
vacuo. A solution of the Neu5Ac-2-O-phosphate 3 (30 mg, 32 mmol) in
DMF (1.0 mL) was added to this residue and the mixture was stirred for
5 days at room temperature. After concentration in vacuo, a solution of
NaOMe (11.0 mg, 204 mmol) in MeOH (2.0 mL) was added to a solution
of the residue in MeOH (2.0 mL). The reaction mixture was stirred at
room temperature for 1 h and then neutralised with acetic acid. After
7654
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 7645 – 7655

Self-Anhydride Formation in the Degradation of CMP-Neu5Ac and CDP-Neu5Ac
FULL PAPER
Synthesis of CMP-Neu5Ac: (Scheme 7) The Neu5Ac-2-O-phosphate 31
(25 mg, 30 mmol) and cytidine (72.5 mg, 300 mmol) were dissolved in dry
DMF (3.0 mL) and the solution was concentrated in vacuo. This co-evap-
oration with DMF (3.0 mL) was repeated four times and the residue was
then dried in a desiccator for 12 h. Hydroxybenzotriazole (8.1 mg, 60 mL)
and diisopropylcarbodiimide (9.2 mL, 60 mmol) were added to a solution
of the residue in dry DMF (350 mL), and the mixture was stirred at room
temperature. After 4.5 h, NaOH (0.2m, 6.0 mL, 1.2 mmol) was added and
the solution was stirred at room temperature. After a further 15 h, the
mixture was lyophilised. Purification of the residue by gel permeation
column (Sephadex G-15, NH4OH (5 mm)) yielded CMP-Neu5Ac (21,
11 mg, 61%). 1H NMR data for this compound were identical to those
for authentic CMP-Neu5Ac. Sialyl transfer activity to LacNAc by rat
liver a-2,6-sialyltransferase was also demonstrated by use of this pure 21.
1H NMR (400.13 MHz, HOD=4.81, D2O): d=8.00 (d, J=7.6 Hz, 1H;
H-6), 6.18 (d, J=7.6 Hz, 1H; H-5), 6.04 (d, J=4.5 Hz, 1H; H-1’), 4.20 (d,
J=10.6 Hz, 1H; H-6’’), 4.12 (ddd, J=10.3, 10.6, 4.8 Hz, 1H; H-4’’), 4.01
(dd, J=10.5, 10.5 Hz, 1H; H-5’’), 4.00 (ddd, J=9.7, 6.6 , 2.4 Hz, 1H; H-
8’’), 3.93 (dd, J=11.7, 2.4 Hz, 1H; H-9’’b), 3.68 (dd, J=11.9, 6.7 Hz, 1H;
H-9’’a), 3.51 (d, J=9.7 Hz, 1H; H-7), 2.54 (dd, J=4.7, 13.4 Hz, 1H; H-
3’’eq), 2.10 (s, 3H; Ac), 1.71 ppm (ddd, J=11.4, 13.4, 5.6 Hz, 1H; H-3’’ax);
31P NMR (161.98 MHz, D2O, H3PO4 =0.00 ppm external standard): d=
ꢀ4.46 ppm.
[7] a) J. M. Risley, R. L. Van Etten, J. Am. Chem. Soc. 1979, 101, 252–
253; b) M. Cohn, A. Hu, Proc. Natl. Acad. Sci. USA 1978, 75, 200–
203; c) G. Lowe, B. S. Sproat, J. Chem. Soc. Perkin Trans. 1 1981,
1874–1878; d) G. D. Dotson, R. K. Dua, J. C. Clements, E. R.
Wooten, R. W. Woodard, J. Biol. Chem. 1995, 270, 13698–13705;
e) T. C. Bruice, A. Blasko, R. D. Arasasingham, J.-S. Kim, J. Am.
Chem. Soc. 1995, 117, 12070–12077.
[8] a) D. G. Comb, D. R. Watson, S. Roseman, J. Biol. Chem. 1966, 241,
5637–5642; b) C.-H. Lin, B. W. Murray, I. R. Ollmann, C.-H. Wong,
Biochemistry 1997, 36, 780–785.
[9] a) J.-M. Beau, R. Schauer, J. Haverkamp, J. P. Kamerling, L. Dor-
land, J. F. G. Vliegenthardt, Eur. J. Biochem. 1984, 140, 203–208;
this paper describes several plausible mechanisms: b) B. A. Horen-
stein, M. Bruner, J. Am. Chem. Soc. 1998, 120, 1357–1362; c) B. A.
Horenstein, J. Am. Chem. Soc. 1997, 119, 1101–1107; d) B. A. Hor-
enstein, M. Bruner, J. Am. Chem. Soc. 1996, 118, 10371–10379.
[10] Synthesis of CMP-Neu5Ac ester 19 and amide 20 derivatives is pro-
vided in the Supporting Information.
[11] Y. Kajihara, K. Koseki, T. Ebata, H. Kodama, H. Matsushita, H. Ha-
shimoto, J. Org. Chem. 1995, 60, 5732–5735.
[12] K. P. C. Vollhardt, N. E. Schore, Organic Chemistry: Structure and
Function, 5th ed., W. H. Freeman, New York, 2007, pp. 864–866.
[13] E. V. Anslyn, D. A. Dougherty, Modern Physical Organic Chemistry,
University Science books, California, 2006, pp. 259–292.
[14] N. Yamamoto, Y. Ohmori, T. Sakakibara, K. Sasaki, L. R. Juneja, Y.
Kajihara, Angew. Chem. 2003, 115, 2641–2644; Angew. Chem. Int.
Ed. 2003, 42, 2537–2540.
Acknowledgements
[15] Intramolecular catalytic acid hydrolysis: a) B. Capon, Tetrahedron
Lett. 1963, 4, 911–913; b) I. R. Greig, A. J. Kirby, J. Phys. Org.
Chem. 2004, 17, 498–506.
Y.K. thanks Aoi Kajihara for technical support and encouragement and
thanks Prof. Carlo Unverzagt (University Bayreuth) for discussions on
the instability of sialoside with respect to intramolecular catalysis. This
work was supported by the research fund of Yokohama City University.
[16] M. Bruner, B. A. Horenstein, Biochemistry 1998, 37, 289–297.
[17] a) S. Makino, Y. Ueno, M. Ishikawa, Y. Hayakawa, T. Hata, Tetrahe-
dron Lett. 1993, 34, 2775–2778; b) M. D. Chappell, R. L. Halcomb,
J. Am. Chem. Soc. 1997, 119, 3393–3394; c) Y. Hayakawa, R.
Kawai, A. Hirata, J. Sugimoto, M. Kataoka, A. Sakakura, M.
Hirose, R. Noyori, J. Am. Chem. Soc. 2001, 123, 8165–8176.
[18] G. Dufner, R. Schworer, B. Muller, R. R. Schmidt, Eur. J. Org.
Chem. 2000, 1467–1482.
[1] a) J. M. Berg, J. L. Tymoczko, L. Stryer, Biochemistry, 6th ed., W. H.
Freeman, New York, 2007; b) A. Fersht, Enzyme Structure and
Mechanism, 2nd ed., W. H. Freeman, New York, 1985; c) C. H. Frey,
J. E. Stuehr, J. Am. Chem. Soc. 1972, 94, 8898–8904.
[2] a) A. Varki, R. D. Cummings, J. D. Esko, H. H. Freeze, G. W. Hart,
J. Marth, Essentials of Glycobiology, Cold Spring Harbor Laborato-
ry Press, Cold Spring Harbor, 1999; b) A. Varki, R. D. Cummings,
J. D. Esko, H. H. Freeze, P. Stanley, C. R. Bertozzi, G. W. Hart,
M. E. Etzler, Essentials of Glycobiology, Cold Spring Harbor Labo-
ratory Press, Cold Spring Harbor, 2009.
[19] R. Brossmer, H. J. Gross, Methods Enzymol. 1994, 247, 177–193.
[20] a) Y. Kajihara, T. Kamitani, R. Sato, N. Kamei, T. Miyazaki, R.
Okamoto, T. Sakakibara, T. Tsuji, T. Yamamoto, Carbohydr. Res.
2007, 342, 1680–1688; b) H. Yu, J. Cheng, L. Ding, Z. Khedri, Y.
Chen, S. Chin, K. Lau, V. K. Tiwari, X. Chen, J. Am. Chem. Soc.
2009, 131, 18467–18477; c) H. Cao, Y. Li, K. Lau, S. Muthana, H.
Yu, J. Cheng, H. A. Chokhawala, G. Sugiarto, L. Zhang, X. Chen,
Org. Biomol. Chem. 2009, 7, 5137–5145; d) H. Cao, S. Muthana, Y.
Li, J. Cheng, X. Chen, Bioorg. Med. Chem. Lett. 2009, 19, 5869–
5871.
[3] L. F. Leloir, Science 1971, 172, 1299–1303.
[4] a) H. A. Staab, H. Schaller, F. Cramer, Angew. Chem. 1959, 71, 736;
b) A. Saleh, F. Compernolle, G. Janssen, Nucleosides Nucleotides
1995, 14, 689–692.
[5] T. Martin, R. R. Schmidt, Tetrahedron Lett. 1993, 34, 1765–1768.
[6] M. M. Sim, H. Kondo, C. H. Wong, J. Am. Chem. Soc. 1993, 115,
2260–2267.
Received: November 24, 2010
Published online: May 19, 2011
Chem. Eur. J. 2011, 17, 7645 – 7655
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7655