Oxidation of sialic acid using hydrogen peroxide as a new method to tune the reducing activity

Article abstract of DOI:10.1016/j.carres.2014.01.007

Functionalized sialic acids are useful intermediates to prepare a wide range of biological products. As they often occur at a non-reducing terminal of oligosaccharides, the most used technique to activate them is by periodate-mediated oxidation of their glycerol side chain. Here, we describe an alternative, non toxic, and environmentally-friendly method to activate the sialic acid residues by hydrogen peroxide oxidation. Four oxidative systems involving H₂O₂, EDTA, iron chloride, and UV light were studied and the products obtained were analyzed by LC-MS and NMR, before and after a derivatization reaction. At first, we observed, for each system, an irreversible decarbonylation reaction at the reducing end. Then, the decarbonylated sialic acid (DSA) was oxidized and fragmented into a mix of carbonyls and carboxyl acids, more or less fast according to the experimental conditions. Analysis of the reaction indicated an apparent radical mechanism and heterolytic alpha-hydroxy-hydroperoxide cleavages. The modest reducing activity was mainly explained as a consequence of over-oxidation reactions.

Full text of DOI:10.1016/j.carres.2014.01.007

Carbohydrate Research 386 (2014) 92–98
Contents lists available at ScienceDirect
Carbohydrate Research
journal homepage: www.elsevier.com/locate/carres
Note
Oxidation of sialic acid using hydrogen peroxide as a new method
to tune the reducing activity
b,
a
C. Neyra a,b, , J. Paladino , M. Le Borgne
⇑
⇑
a Université de Lyon, Université Lyon 1, Faculté de Pharmacie—ISPB, EA 4446 Biomolécules Cancer et Chimiorésistances, SFR Santé Lyon-Est CNRS UMS3453—INSERM US7, 8
Avenue Rockefeller, F-69373 Lyon Cedex 8, France
b Reaction and Coupling Chemistry Laboratory, MTech, Sanofi Pasteur, 31/33 quai Armand Barbès, 69250 Neuville-sur-Saône, France
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 4 October 2013
Received in revised form 13 December 2013
Accepted 8 January 2014
Available online 14 January 2014
Functionalized sialic acids are useful intermediates to prepare a wide range of biological products. As
they often occur at a non-reducing terminal of oligosaccharides, the most used technique to activate
them is by periodate-mediated oxidation of their glycerol side chain. Here, we describe an alternative,
non toxic, and environmentally-friendly method to activate the sialic acid residues by hydrogen peroxide
oxidation. Four oxidative systems involving H2O2, EDTA, iron chloride, and UV light were studied and the
products obtained were analyzed by LC–MS and NMR, before and after a derivatization reaction. At first,
we observed, for each system, an irreversible decarbonylation reaction at the reducing end. Then, the
decarbonylated sialic acid (DSA) was oxidized and fragmented into a mix of carbonyls and carboxyl acids,
more or less fast according to the experimental conditions. Analysis of the reaction indicated an apparent
radical mechanism and heterolytic alpha-hydroxy-hydroperoxide cleavages. The modest reducing activ-
ity was mainly explained as a consequence of over-oxidation reactions.
Keywords:
Sialic acid
N-Acetylneuraminic acid
Hydrogen peroxide
Oxidation
Decarbonylation
Reductive amination
Ó 2014 Elsevier Ltd. All rights reserved.
Sialic acids consist of a large family of more than fifty acety-
lated, methylated, or sulfated 9-carbon carboxylated alpha-keto
sugars.1 The most common member of this family of natural prod-
ucts is N-acetylneuraminic acid (NeuNAc), which has an acetyl
group attached to the 5-amino group. Its structure has been stud-
ied in detail in the literature2 (Fig. 1). It exits predominantly in the
b configuration.3
NeuNAc occurs in nature as carbohydrate chains of bacterial
and animal glycoproteins,4 glycolipids,5 and polysaccharides.6 Lo-
cated at the terminus of numerous cell-surface oligosaccharides,
NeuNAc has been used in a wide range of biomedical applications
such as vaccines7 or drug delivery systems.8,9 They are ideally posi-
tioned to be covalently attached to proteins (glycoconjugate vac-
cines), polyethylene glycol (pegylated glycoproteins), or cytotoxic
drugs (antibody–drug conjugates).
As most of the bioconjugations occur in aqueous solution, the
ways to activate a carbohydrate present on a biopolymer are lim-
ited. The ketone at the reducing end of the NeuNAc can be deriva-
tized, for example via a reductive amination.10 But since the
unreactive ring form dominates,3 the reactivity of this function is
relatively low. Furthermore, sialic acid (SA) often occurs at a termi-
nal non-reducing oligosaccharide making its ring not capable of
opening. Although alcohols are abundant in carbohydrates like
NeuNAc, their reactivity in water is extremely low. In order to
achieve a high conjugation reaction yield between this carbohy-
drate and another molecule, an activation step is necessary: num-
ber of known reactions exist to functionalize hydroxyl groups. A
convenient method is to introduce a carbonyl group via an oxida-
tion reaction. The most used technique is by periodate-mediated
oxidation of the adjacent hydroxyls at C-7, C-8, and C-9 of sialic
acid residues to obtain carbon–carbon bond cleavage and amine
reactive aldehydes.8,11 Other carbohydrate oxidation techniques
involve the use of an oxidant such as sodium hypochlorite (NaOCl)
to introduce carboxyl and carbonyl groups in polysaccharides.12
Combined with TEMPO ((2,2,6,6-tetramethyl-1-piperidin-1-yl)ox-
yl), it selectively oxidizes primary hydroxyl groups in carbohy-
drates.13 Such reactions with TEMPO and sialic acid14 have been
described leading to the formation of 9-carboxy-NeuNAc via sev-
eral reactions. One drawback of these reagents is the difficulty to
avoid over-oxidation of the produced carbonyl functions.
In a green chemistry context, it would be interesting to design
an inexpensive oxidation reaction capable to selectively activate
the SA moiety, at its non-reducing end, under mild reaction condi-
tions. Hydrogen peroxide (H2O2), an environmentally friendly oxi-
dant, might be an ideal candidate.
There have been a number of studies of carbohydrate oxidations
with hydrogen peroxide as oxidant, like the depolymerization
reaction of polysaccharides (chitosan,15 starch16 or hyaluronic
acid,17,18 Neisseria meningitidis capsular polysaccharides19,20), or
the degradation of mono and di-saccharides21,22 but only few
⇑
Corresponding authors. Tel.: +33 4 37668699; fax: +33 4 37656187.
E-mail addresses: christophe.neyra@sanofi.com (C. Neyra), joseph.paladino@
sanofi.com (J. Paladino).
0008-6215/$ - see front matter Ó 2014 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.carres.2014.01.007

C. Neyra et al. / Carbohydrate Research 386 (2014) 92–98
93
1
COOH
OH
OH
2
O
1COOH
OH
8
3
HO
6
9
8
H
H
7
HO
6
9
7
2
1COOH
4
OH
4
O
2
HO
OH
4
O
HO
5
6
H3CCOHN
HO
N
N
HO
5
H3C
3
HO
5
H3C
3
7
8
9
OH
OH
OH
O
O
Beta anomer
Alpha anomer
Keto form
Figure 1. Anomeric forms of N-acetylneuraminic acid.
works have used H2O2 oxidation as an activation reaction through
the introduction of carbonyl groups.23
paper, we report identification of various structures by LC–MS
and NMR.
Isbell and Frush24 have proposed many mechanisms to explain
the formation of various oxidation products formed from carbohy-
drates by the action of hydrogen peroxide. Under acid catalysis,
carbohydrates seem to be inert to H2O2. But in alkaline conditions,
in presence of traces of metal ions (as shown by Fenton25), or under
UV light,16 highly reactive species such as hydroperoxide anion,
hydroxyl radical or hydroperoxyl radical can be generated and re-
act with carbohydrates. While hydroperoxide anion will preferen-
tially attack the carbonyl groups like the reducing ends, the
radicals will be able to abstract hydrogen from C–H moieties.
When hydrogen peroxide is in a large excess, Isbell26 has postu-
lated that the oxidative degradation of an aldose and a ketose pro-
ceeds by a nucleophilic addition of a peroxide anion on the
reducing end of the sugar, followed by the fragmentation of the
intermediate adduct by a free radical or an ionic mechanism. This
1. Experimental procedures
1.1. Chemicals
N-Acetylneuraminic acid (NeuNAc) was purchased from Carbo-
synth. In this present work SA will refer to NeuNAc. All other re-
agent grade chemicals were purchased from Sigma-Aldrich. The
solutions were prepared with Millipore-quality water (Milli-Q
plus, Ultrapure water system, 18 M
X cm).
1.2. Sialic acid oxidation experiments
In all reactions, NeuNAc (120 mg) was dissolved in 20 mL of so-
dium acetate (50 mM, pH 6). Five different oxidative systems with
H2O2 were prepared.
alpha-hydroxy hydroperoxide (aHHP) cleavage gives formic acid
and the next lower aldose. Repetitions of this reaction lead to the
stepwise degradation of the aldose to formic acid.
The sialic acid solution was treated with the following entries:
experiment (a) an equimolar quantity of hydrogen peroxide;
experiment (b) a large excess of H2O2 (90 molar equiv); experi-
Such reactions have been done almost entirely with ‘simple’
carbohydrates such as glucose or fructose but not with more com-
plex sugars like carboxylated alpha-keto sugars. Various alpha-
keto acids (similar to the acyclic keto form of the sialic acid) have
been shown to react with a molar equiv of H2O2 at neutral pH via a
nucleophilic attack of the hydroperoxide anion at the carbonyl
group adjacent to the carboxylic group.27 For all the keto-acids
studied, the reaction with H2O2 gave rise to the formation of car-
boxylate anions of chain length one carbon shorter. The mecha-
nism described involved the nucleophilic attack followed by a
decarboxylation. A similar study has been done by Ijiama and col-
laborators28,29 where the reaction between SA and H2O2 was char-
acterized. They have found that an equimolar amount of H2O2 can
oxidize sialic acid at pH 6 and provide its decarbonylated product
via the same mechanism.
ment (c) 25
a large excess of H2O2 (90 molar equiv) was added to the resulting
mixture; experiment (d) 50 g of EDTA were mixed to the NeuNAc
lg of FeCl2 were mixed to the NeuNAc solution and
l
solution and a large excess of H2O2 (90 molar equiv) was added to
the resulting mixture; experiment (e) a large excess of H2O2
(90 molar equiv) was added, then the solution was exposed to
UV light (2 W/cm2) using a photochemical device (system Omni-
cureÒ S2000).
All the reactions with H2O2 were conducted in the dark at 66 °C,
during 6 h (except for the UV experiment, 2 h only) while main-
taining pH 6 with sodium hydroxide and hydrogen chloride.
The experiments with sodium periodate (1.1 molar equiv) were
carried out at 15 °C, pH 6, and stirred during 20 min and then
quenched with glycerol.
In order to understand the reaction mechanism of sialic acid
residues with hydrogen peroxide, N-acetyl neuraminic acid was
chosen as the simplest model to react with hydrogen peroxide.
These results would be very valuable in case of a future investiga-
tion on the oxidation of glycoside derivatives of SA, such as oligo-
sialic acids, with H2O2.
In this study we were interested in the degradation of NeuNAc
when using common oxidative systems containing excess of H2O2
(in the presence of iron, EDTA or under UV light). We identified the
main products by LC–MS and NMR of the obtained fragments and
compared the data with results obtained from a well known mech-
anism (sodium periodate oxidation).
The oxidized products were dialyzed three times against water
(0.1 kDa membrane) and lyophilized.
1.3. Derivatization procedures
The formed carbonyl functions were characterized by derivati-
zation with aniline. The lyophilized products were redissolved in
20 mL of sodium acetate (50 mM, pH 5). Aniline was added
(1.1 molar equiv) and the pH was maintained at 5. After 2 h at
25 °C, 1.6 molar equiv of sodium cyanoborohydride (NaBH3CN)
was added to reduce the imine intermediates. The solution was
stirred at pH 5 and 25 °C during 40 h.
The aim of this study was to assess the introduction of carbonyl
groups for each oxidative system. The reducing activity of these
new carbonyl groups was evaluated by incorporating aniline to
the oxidized fragments through a reductive amination. In this
Unoxidized NeuNAc was derivatized with aniline, as a control,
to compare the reactivity of the reducing end with the new carbon-
yls functions formed through the H2O2 and sodium periodate
reactions.

94
C. Neyra et al. / Carbohydrate Research 386 (2014) 92–98
1.4. Analysis
(see Supplementary data). The DSA being very stable in its acyclic
state, no lactone form was observed.
1.4.1. HILIC–LC–MS analyses
Oxidized and derivatized products were analyzed using HILIC–
HPLC (hydrophilic interaction liquid chromatography) with UV
detection (at 214 nm for the oxidized products and at 280 nm for
the aniline derivatized products) coupled to ESI-MS (electrospray
ionization mass spectrometry).
Samples were brought to an acetonitrile content of 80%. The
liquid chromatography was performed with a Waters AllianceÒ
system equipped with Diode Array Detector. A LunaÒ HILIC diol
2.2. Oxidations of SA with excess of H2O2
Four different experiments (b–e) were carried out using large
excess of H2O2 (90 molar equiv). For each of them, the first reaction
was the decarbonylation of the SA, as seen previously. In less than
12 min, SA was entirely converted to DSA (no more SA detectable
by LC–MS or NMR). Then, a large excess of H2O2 was still available
to further oxidize the DSA (see Supplementary data). Use of LC–MS
analysis showed the formation of selective oxidation products. The
ions m/z of the new products were extracted from the total ion
chromatograms and were compared.
column (length 150 mm, diameter 4.6 mm, particles 3
lm, Phe-
nomenex) was used. 10 L sample was injected into the column.
l
Fragments were separated with acetonitrile/ammonium formate
(80:20, v/v, pH 6) at a flow rate of 0.8 mL/min, during 25 min.
Detection was made through a MS/MS ion trap detector (Bruker
Daltonics HCT Plus) using an ESI ion source, operated under nega-
tive mode. The control of the chromatographic system and data
acquisition was achieved with the HyStar™ software Bruker Dal-
tonics incorporating the MS acquisition program, Esquire Control.
The four experiments (b–e) seemed to follow the same degrada-
tion pathway but with different kinetics.
2.2.1. Excess of H2O2 with Fe2+
During the experiment (c), the first peaks [MÀH]À observed by
LC–MS were those with m/z 278 (Fig. 2, peaks D, E, G). Then, these
peaks tended to decrease and new peaks appeared: three of the
new products formed were less hydrophilic than SA and DSA
(shorter retention times by HILIC) with m/z of 248, 218, and 188,
while three others were more hydrophilic (m/z 264, 234, and
204). After 5 h, a large amount of DSA was still present. Table 1
summarizes the spectral information of the detected peaks. Com-
pared to the control experiment, experiment (b) using H2O2 in ex-
cess without any treatment, the obtained products were the same
but their formation seemed to be slightly faster.
1.4.2. NMR measurement
To follow the oxidation of NeuNAc during the reaction, samples
were regularly taken from the solution, dialyzed against water to
remove the salts and the excess of H2O2, lyophilized, and redis-
solved in D2O.
1H and 13C NMR spectra were recorded at 30 °C with a Bruker
Avance II spectrometer at 500 MHz equipped with a cryoprobe.
The data were recorded using the Bruker TopSpin™ software. The
spectra were referenced with respect to an internal standard
TSP-d4 (sodium tetra deuterated trimethylsilyl propionate), at
0 ppm (1H) or À0.18 ppm (13C).
2.2.2. Excess of H2O2 with EDTA
When EDTA was present in the solution, experiment (d), almost
no oxidation of DSA occurred. The rise of peaks with a m/z of 278
[MÀH]À was much slower than experiments (b) and (c). No other
peaks were detected by LC–MS after 5 h.
2. Results and discussion
The mechanism of the hydrogen peroxide oxidation of carbohy-
drates has been thought to involve the highly reactive hydroxyl
radical HOÅ.21,22,30 In the present study, hydroxyl radicals were gen-
erated by the iron-catalyzed decomposition of H2O2 or by irradia-
tion under ultraviolet light of H2O2 (homolytic fission). The
experiments were repeated in the presence of EDTA (metal chelat-
ing agent) to suppress the influence of metal traces upon the oxi-
dative degradation.31
2.2.3. Excess of H2O2 under UV
The experiment (e), under UV light, was significantly faster than
the others. After 1 h, all DSA and the compounds with m/z of 278
disappeared. The reaction was stopped after 2 h in order not to to-
tally degrade the oxidized products. The five predominant prod-
ucts detected by HILIC were those with m/z [MÀH]À of 188, 218,
234, 204, and 264 (peaks A, B, H, I, and J, respectively, Fig. 2 and
Table 1).
On the basis of these results, we can assume that the first reac-
tion (decarbonylation of SA to DSA) would occur even without free
radicals.
In the next oxidation reactions, the rapid increase in the num-
ber of peaks observed for the three experiments without EDTA
was attributed to the reaction of DSA with hydroxyl radicals liber-
ated by the Fe2+/Fe3+ oxidation-reduction cycle (Fenton reaction)
or by the homolytic cleavage of H2O2 under UV light. A metal con-
tamination in the reaction mixture could explain the minor differ-
ences between the experiments with and without iron. The
2.1. Oxidation of SA with 1 molar equiv of H2O2
This experiment (a) confirmed the results from Ijima and col-
laborators. When SA and H2O2 were mixed together at a ratio 1:1
(at 66 °C, pH 6) a fast nucleophilic attack of H2O2 on the ketone
of the open ring form of the SA happened, followed by a Baeyer–
Villiger like rearrangement and an irreversible decarboxylation.
The reaction was completed in 52 min (Scheme 1) and gave an
acyclic acid, DSA (4-(acetyl-amino)-2,4-dideoxy-D-glycero-D-galac-
to-octonic acid), fully characterized by LC–MS and NMR
O
OH
OH
OH
HO
HO
OH
H2O2, 1 eq
-CO2
OH
COOH
HO
N
O
HO
N
HO
H3C
HO
H3C
O
O
1 (SA)
MW = 309.28 g/mol
2 (DSA)
MW = 281.26 g/mol
Scheme 1. Reaction scheme of the decarbonylation of SA.

C. Neyra et al. / Carbohydrate Research 386 (2014) 92–98
95
Figure 2. Extracted ion chromatograms of hydrogen peroxide oxidation of DSA in the presence of a trace amount of Fe2+
.
Table 1
Spectral information on H2O2 oxidative products from DSA in the presence of a trace amount of Fe2+
Peak
A
B
C
D
E
F
G
H
I
J
RT (min)
4.5
188
5.4
218
6.4
248
8.0
278
9.2
278
10.5
280
(DSA)
11.0
278
14.0
234
14.9
204
19.8
264
[MÀH]À (m/z)
presence of EDTA enhanced the quenching of radical hydroxyls,31
which severely delayed the peroxide reaction, but did not stop it
completely.
To summarize the hydrogen peroxide oxidation, the degrada-
tion products of SA and their relative concentrations observed by
HPLC, according to the oxidation conditions, are presented in
Scheme 4 and Table 2. Products 2, 4, 7, and 8 were fully character-
ized by 1H and 13C NMR (see Supplementary data).
Studies on H2O2 oxidation24 have shown that HOÅ can abstract
hydrogen C–H moieties of an alcohol group, forming a radical
intermediate, which then reacts with hydrogen peroxide affording
a carbonyl group. In a same way, hydroxyl group of DSA could be
oxidized to a carbonyl group, which would explain the formation
of products with a m/z [MÀH]À of 278. The hydrogen abstraction
by hydroxyl radical being little or not selective, the five hydroxyl
groups of DSA can potentially be oxidized. One of these reactions
is shown in Scheme 2.
2.3. Derivatization of the oxidized compounds
To confirm the presence of carbonyl groups, the degradation
compounds for each oxidation were derivatized by reductive ami-
nation with aniline under the same conditions previously
described and analyzed by LC–MS.
The further oxidation and degradation of these compounds
could generate a series of acids and carbonyl derivatives.24 A plau-
sible reaction pathway would involve the nucleophilic addition of
hydrogen peroxide to the carbonyl group, followed by a heterolytic
No reaction was observed between aniline and DSA or the
di-carboxylic acids (products 7, 8, and 9, Scheme 4), as this deriv-
atization method was limited to carbonyl compounds. These four
compounds were still present and visible by LC–MS at the end of
the reductive amination.
alpha-hydroxy-hydroperoxide (aHHP) cleavage of the adduct. This
would explain the formation of the products described in Table 1.
The possible mechanism of formation of products with a molecular
weight of 265 g/mol and 219 g/mol is depicted in Scheme 3.
A similar pathway could explain the formation of two other
aldehydes (MW = 189 and 249 g/mol) and two other carboxylic
acids (MW = 205 and 235 g/mol).
The aldehyde groups, almost completely hydrated to diols or
cyclic hemiacetals, gave rise to several characteristic signals in
the proton and carbon NMR spectra. The presence of carbonyl
groups was confirmed by the addition of the carbonyl reductive
agent NaBH4 which caused a marked decrease of these NMR
signals.
Four different structures absorbing at 280 nm (aniline absorp-
tion band) were identified (in different amounts according to the
H2O2 oxidation condition), corresponding to the derivatization of
the products 3, 4, 5, and 6. We were surprised to see that the deriv-
atization kinetic was different according to the aldehyde. Product 3
(279 g/mol) and product 5 (219 g/mol) were reacting almost
instantaneously with aniline and after NaBH3CN addition to form
the derivatized structures with a mass of 356 g/mol and 296 g/
mol, respectively, while it took more time for the other aldehydes
to react. The derivatization of product 4 (249 g/mol) lasted around
20 h. A cyclization of product 4 by nucleophilic addition reaction
between the carbonyl group and a hydroxyl group might happen
HO
HO
O
HO
OH
O
HO.
OH
O
H
HO.
HO
.
N
O
H
H2O2
OH
+
N
OH
HO
H
N
HO
O
OH
H
OH
OH
H2O
HO
OH
O
HO
HO
O
3
2 (DSA)
MW = 281 g/mol
MW = 279 g/mol
Scheme 2. Proposed reaction pathway for the formation of one of the products with a MW of 279 g/mol.

96
C. Neyra et al. / Carbohydrate Research 386 (2014) 92–98
H
O
OH
O
O
OH
HO
- OH-
HO
H
N
OH
HO
O
OH
O
O
H2O2
O
HO
H
O
H
N
OH
OH
O
HO
7
H
N
OH
OH
HO
MW = 265 g/mol
H2O2
O
HO
O
HO
3
O
O
- OH-
OH
MW = 279 g/mol
H
HO
O
OH
O
N
OH
H
O
HO
H
N
OH
OH
O
HO
5
O
MW = 219 g/mol
Scheme 3. Proposed mechanisms for the
aHHP cleavage and the formation of new carbonyl and carboxylic groups.
Decarbonylated Sialic Acid
Sialic Acid
HO
OH
O
HO
OH
O
Very fast
HO
H
HO
H
N
O
OH
OH
N
OH
H2O2
65°C
,
HO
OH
HO
O
O
2 (DSA)
MW = 281 g/mol
HO
HO
OH
HO
CO2
OH
O
OH
OH
HO
O
H
N
HO
OH
O
O
H
O
O
HO.
N
O
H
N
OH
OH
HO
HO
O
OH
HO
O
O
O
HO
OH
HO
OH
O
O
1 (SA)
O
HO
H
N
OH
OH
HO
HO
H
MW = 309 g/mol
H
N
OH
OH
OH
N
OH
HO
O
HO
O
O
O
O
O
HO
H
3
OH
N
OH
MW = 279 g/mol
HO
O
4
MW = 249 g/mol
H2O2
O
O
H
OH
N
OH
HO
HO
O
HO
O
O
O
HO
O
O
O
5
H
N
H
OH
OH
H
N
OH
N
OH
OH
OH
MW = 219 g/mol
HO
HO
HO
O
O
O
H2O2
7
9
8
MW = 265 g/mol
MW = 235 g/mol
MW = 205 g/mol
O
O
H
N
OH
HO
O
6
MW = 189 g/mol
Scheme 4. H2O2 degradation products of SA formed under various conditions.
and yield a six-membered ring (Fig. 3) more stable than a 7 or
5-membered ring.
This stable hemiacetal ring form could explain the slower deriv-
atization reaction compared to the other products.
To verify the product structures obtained after H2O2 oxidation
and derivatization, SA was decarbonylated with an excess of
H2O2 in the presence of EDTA (to inhibit the oxidation reaction).
Then, the produced DSA was submitted to periodate oxidation, a
well known reaction where sodium periodate specifically cleaved
at the adjacent hydroxyls of the non-reducing end (between car-
bon atoms 6, 7, and 8) of DSA to form three different aldehydes
(with a mass of 189, 219, and 249 g/mol). These three intermedi-
ates were derivatized with aniline under the conditions previously
used. The results (retention time and mass spectra) were similar
than those observed with products 4, 5, and 6 (Scheme 4). Thus,
this experiment confirmed the structure of the derivatized prod-
ucts obtained after the initial H2O2 oxidation.
The study (by HPLC at 280 nm) of the integrated peak areas of
aniline excess and derivatized products helped us to estimate the
derivatization yield of the carbonyl compounds according to the
conditions of the prior oxidation reactions. The results are pre-
sented in Table 3.

C. Neyra et al. / Carbohydrate Research 386 (2014) 92–98
97
Table 2
Degradation products of SA formed under various conditions and their relative concentration
Products
Reaction conditions
H2O2 equimolar
H2O2 exa
H2O2 exa + EDTA
H2O2 exa + FeCl2
H2O2 exa + UV
2 (DSA)
+++
0
0
0
0
0
0
0
++
++
+
0
0
+
+
+
+++
+
0
0
0
0
0
0
+
0
0
0
+
+
+
+
+
3
4
5
6
7
8
9
++
++
+
+
+
+
+
(‘+++’ indicates a compound present in higher concentration,‘0’ indicates an absent compound).
a
ex: excess.
OH
NH
O
HO
O
O
OH
O
HO
H
OH
N
OH
HO
O
HO
O
Figure 3. Proposed structures of product 4.
Table 3
Aniline consumed during each derivatization reactions
Oxidation
conditions
(a) H2O2
equimolar
(b) H2O2
excess
(c) H2O2
excess + EDTA
(d) H2O2
(e) H2O2
excess + UV
(f) H2O2
excess + NaIO4
(g) No oxidative
reagent
excess + Fe2+
Aniline consumed 0%
20%
8%
16%
8%
40%
31%
The integrated peak areas of the derivatized products ob-
tained after H2O2 oxidation were the highest for the experiments
(b) and (d) (excess of H2O2 with or without Fe2+). The amount of
consumed aniline was also the most important for these two
experiments (between 16% and 20%). As expected, the derivatiza-
tion yields were lower for the experiments (a) and (c) (one equi-
molar H2O2 and with EDTA), but we were surprised to see that
the consumed aniline for the reaction (e) (under UV) was almost
three times less important than for the experiments (b) and (d).
Probably during the H2O2 oxidation step, the quantity of very
reactive hydroxyl radicals was too important. Thus, the carbonyl
groups could undergo an over-oxidation resulting in unreactive
smaller compounds like formic acid or carboxylic acid. As a com-
parison, 31% of aniline was consumed during the derivatization
of the SA at its reducing end (without any prior oxidation
reaction).
The sodium periodate oxidized compounds showed a signifi-
cantly higher yield of derivatization as the aniline consumed was
at least twice higher than the other derivatizations.
These results demonstrated that the periodate oxidation of
DSA yielded a higher amount of carbonyl groups compared to
the hydrogen peroxide oxidation. The modest derivatization
yield of H2O2 oxidized compounds could be explained by an
over-oxidation phenomenon responsible for the formation of
underivatizable carboxylic acids. This fact was emphasized
3. Conclusion
In summary, we developed a new method to activate sialic acid
at its non-reducing end by generating carbonyl functions. Indeed,
the chemical modification, by hydrogen peroxide-mediated oxida-
tion, lead, via
aHPP cleavages, to different carboxyl and carbonyl
functions along the glycerol side chain of sialic acid (Scheme 4).
The reducing activity was relatively modest compared to periodate
mediated oxidation. It could be explained by the formation of
undesired carboxylic acids and by over-oxidation reactions result-
ing in unreactive smaller compounds, which lower the derivatiza-
tion yield.
Further research toward optimal oxidation conditions, such as
reaction temperature, H2O2 concentration, and pH, might control
SA reducing activity. Once these conditions were well defined, an-
other study could be realized with oligosaccharide derivatives of
SA.
Hydrogen peroxide appeared to be a non toxic and environmen-
tally-friendly alternative to the use of sodium periodate as an acti-
vation agent of sialic acid residues.
Acknowledgements
This work was carried out with the support of the Global Tech-
nology Innovation department of Sanofi Pasteur, Marcy l’Etoile
(H. Pinton, P. Baduel, and D. Speck). The authors would like to
thank J. R. Besse, P. Duverney, and P. Verlant from the structural
analysis laboratories of Sanofi, Neuville-sur-Saône for the technical
assistance in NMR and in mass spectrometry. We are grateful to
F. Arminjon, S. Hauser, and R. McMaster for useful discussions.
during the reaction under UV light, where
a large amount
of hydroxyl radicals were generated. It is also interesting to
note that the formation of carbonyl groups on SA residues
using H2O2 can be tuned by various oxidizing process
conditions.

98
C. Neyra et al. / Carbohydrate Research 386 (2014) 92–98
14. Kiefel, M. J.; Chopra, P.; Madge, P. D.; Szyczew, A.; Thomson, R. J.; Grice, I. D.;
Supplementary data
von Itzstein, M. Tetrahedron Lett. 2011, 52, 98–100.
15. Tian, F.; Liu, Y.; Hu, K.; Zhao, B. Carbohydr. Polym. 2004, 57, 31–37.
16. Harmon, R. E.; Gupta, S. K.; Johnson, J. Starch—Stärke 1971, 23, 347–349.
17. Jahn, M.; Baynes, J. W.; Spiteller, G. Carbohydr. Res. 1999, 321, 228–234.
18. Xu, F.; Maslanek, J.; Behnm, R. S. Int. J. Appl. Chem. 2010.
19. Ryall, R. European Patent 2002707551, 2003.
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.carres.2014.
01.007.
20. Cai, X.; Lei, Q. P.; Lamb, D. H.; Shannon, A.; Jacoby, J.; Kruk, J.; Kensinger, R. D.;
Ryall, R.; Zablackis, E.; Cash, P. Anal. Chem. 2004, 76, 7387–7390.
21. Schuchmann, M. N.; von Sonntag, C. J. Chem. Soc., Perkin Trans. 2 1977, 1958–
1963.
22. Manini, P.; La Pietra, P.; Panzella, L.; Napolitano, A.; d’Ischia, M. Carbohydr. Res.
2006, 341, 1828–1833.
23. Klein-Koerkamp, C.; Granet, R.; Zerrouki, R.; Villandier, N.; Jérôme, F.; Barrault,
J.; Krausz, P. Carbohydr. Polym. 2009, 78, 938–944.
24. Isbell, H. S.; Frush, H. L. Carbohydr. Res. 1987, 161, 181–193.
25. Fenton, H. J. H. J. Chem. Soc. Trans. 1894, 65, 899–910.
26. Carbohydrates in Solution; Isbell, H. S., Ed.Advances in Chemistry; American
Chemical Society, 1973; Vol. 117, pp 70–87.
27. Perera, A.; Parkes, H. G.; Herz, H.; Haycock, P.; Blake, D. R.; Grootveld, M. C. Free
Radical Res. 1997, 26, 145–157.
28. Iijima, R.; Takahashi, H.; Namme, R.; Ikegami, S.; Yamazaki, M. FEBS Lett. 2004,
561, 163–166.
29. Iijima, R.; Takahashi, H.; Ikegami, S.; Yamazaki, M. Biol. Pharm. Bull. 2007, 30,
580–582.
30. Gilbert, B. C.; King, D. M.; Thomas, C. B. Carbohydr. Res. 1984, 125, 217–235.
31. Li, L.; Abe, Y.; Kanagawa, K.; Shoji, T.; Mashino, T.; Mochizuki, M.; Tanaka, M.;
Miyata, N. Anal. Chim. Acta 2007, 599, 315–319.
References
1. Angata, T.; Varki, A. Chem. Rev. 2002, 102, 439–470.
2. Schauer, R. Angew. Chem. Int. Ed. Engl. 1973, 12, 127–138.
3. Klepach, T.; Carmichael, I.; Serianni, A. S. J. Am. Chem. Soc. 2008, 130, 11892–
11900.
4. Furthmayr, H. J. Supramol. Struct. 1978, 9, 79–95.
5. Smirnova, G. P.; Kochetkov, N. K. Biochim. Biophys. Acta 1980, 618, 486–495.
6. Lifely, M. R.; Nowicka, U. T.; Moreno, C. Carbohydr. Res. 1986, 156, 123–135.
7. Lindberg, A. A. Vaccine 1999, 17, S28–S36.
8. Zhou, Q.; Stefano, J. E.; Harrahy, J.; Finn, P.; Avila, L.; Kyazike, J.; Wei, R.; Van
Patten, S. M.; Gotschall, R.; Zheng, X.; Zhu, Y.; Edmunds, T.; Pan, C. Q.
Bioconjugate Chem. 2011, 22, 741–751.
9. Scheigetz, J.; Zamboni, R.; Bernstein, M. A.; Roy, B. Org. Prep. Proced. Int. 1995,
27, 637–644.
10. Che, F.-Y.; Shao, X.-X.; Wang, K.-Y.; Xia, Q.-C. Electrophoresis 1999, 20, 2930–
2937.
11. Lenten, L. V.; Ashwell, G. J. Biol. Chem. 1971, 246, 1889–1894.
12. Kuakpetoon, D.; Wang, Y.-J. Carbohydr. Res. 2006, 341, 1896–1915.
13. Bragd, P. L.; Besemer, A. C.; van Bekkum, H. Carbohydr. Polym. 2002, 49, 397–
406.