Glycan analysis via derivatization with a fluorogenic pyrylium dye

Article abstract of DOI:10.1016/j.carres.2012.02.016

The expansion of glycomics analysis is reliant upon the development of robust, routine methods for carbohydrate characterization. Simple protocols to derivatize sugars with functionality that facilitate analysis - chromophores, fluorophores, charges, ionizable groups - are therefore necessary. Here we describe a method for the labeling of oligosaccharide mixtures with a fluorogenic pyrylium dye to enable analysis by capillary electrophoresis (CE) and matrix assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-ToF-MS). The unreacted free dye, Py-1, is effectively non-fluorescent but when conjugated to the analyte it displays strong fluorescence at 600-640 nm. Removal of excess dye following labeling is not required prior to analysis unlike for many traditional oligosaccharide labels. Labeling is achieved in two steps; the oligosaccharide mixtures are first functionalized with an ethylenediamine moiety via reductive amination at the reducing-end sugar, then the remaining free primary amine is reacted with the pyrylium dye (Py-1) under basic conditions to form a pyridinium ion. We have labeled mixtures of maltooligosaccharides and observed good peak separation in CE analysis using a SDS/?borate pH 9.3 running buffer. Excellent sensitivity in MALDI-ToF-MS analysis enabled detection of oligosaccharides with up to 58 glucose units.

Full text of DOI:10.1016/j.carres.2012.02.016

Carbohydrate Research 352 (2012) 94–100
Contents lists available at SciVerse ScienceDirect
Carbohydrate Research
journal homepage: www.elsevier.com/locate/carres
Glycan analysis via derivatization with a fluorogenic pyrylium dye
Sine A. Johannesen a, Sophie R. Beeren a, Dennis Blank b, Byung Y. Yang c, Rudolf Geyer b, Ole Hindsgaul a,
⇑
a Carlsberg Laboratory, Gamle Carlsberg Vej 10, Copenhagen V DK-1799, Denmark
b Institute of Biochemistry, Faculty of Medicine, Justus-Liebig-University of Giessen, Friedrichstrasse 24, 35392 Giessen, Germany
c Grain Processing Corporation, 1600 Oregon St. Muscatine, IA 52761, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
The expansion of glycomics analysis is reliant upon the development of robust, routine methods for car-
bohydrate characterization. Simple protocols to derivatize sugars with functionality that facilitate anal-
ysis—chromophores, fluorophores, charges, ionizable groups—are therefore necessary. Here we describe a
method for the labeling of oligosaccharide mixtures with a fluorogenic pyrylium dye to enable analysis by
capillary electrophoresis (CE) and matrix assisted laser desorption/ionization time-of-flight mass spec-
trometry (MALDI-ToF-MS). The unreacted free dye, Py-1, is effectively non-fluorescent but when conju-
gated to the analyte it displays strong fluorescence at 600–640 nm. Removal of excess dye following
labeling is not required prior to analysis unlike for many traditional oligosaccharide labels. Labeling is
achieved in two steps; the oligosaccharide mixtures are first functionalized with an ethylenediamine
moiety via reductive amination at the reducing-end sugar, then the remaining free primary amine is
reacted with the pyrylium dye (Py-1) under basic conditions to form a pyridinium ion. We have labeled
mixtures of maltooligosaccharides and observed good peak separation in CE analysis using a SDS/borate
pH 9.3 running buffer. Excellent sensitivity in MALDI-ToF-MS analysis enabled detection of oligosaccha-
rides with up to 58 glucose units.
Received 30 November 2011
Received in revised form 13 February 2012
Accepted 15 February 2012
Available online 24 February 2012
Dedicated to Professor Hans Paulsen on the
occasion of his 90th birthday.
Keywords:
Carbohydrate analysis
Fluorogenic dye
Capillary electrophoresis
MALDI-ToF-MS
Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction
dyes, themselves, are non-fluorescent but become fluorescent only
after reaction with the analyte.9–11 Purification of fluorogenically-
Research in glycobiology is dependent upon the availability of
efficient and reliable analytical techniques to quantify and charac-
terize carbohydrates.1–3 Since the carbohydrates themselves lack
chromophores or fluorophores, and have poor ionization efficiency,
derivatization is generally required for analysis using HPLC or
capillary electrophoresis (CE), and for detection by mass spectrom-
etry.4–8 Derivatization at the reducing-end aldehyde (hemi-acetal)
ensures that each oligosaccharide receives precisely one label,
enabling direct quantification of glycan mixtures. Here we describe
a two step method for the functionalization of oligosaccharides
with a fluorogenic pyrylium dye (Py-1) that generates a highly
fluorescent sugar conjugate and incorporates a permanent positive
charge, thus allowing sensitive detection by CE and matrix assisted
laser desorption/ionization time-of-flight mass spectrometry
(MALDI-ToF-MS).
labeled samples, therefore, or careful washing of a fluorogenical-
ly-stained gel to minimize background fluorescence, is not
required before analysis.
Pyrylium salts selectively react with primary amines to form
pyridinium salts and have been used to functionalize proteins with
various labels via reaction with the e
-amino group of lysine.11–13
Py-1, recently developed by Wolfbeis as a fluorogenic protein stain
for gel electrophoresis, is blue and virtually non-fluorescent.11
Upon reaction with a primary amine it forms a red pyridinium con-
jugate (kabs. max. ꢀ 500 nm) with strong fluorescence emission at
600–640 nm (Fig. 1). Commercially sold as Chromeo P503, Py-1
has been effectively utilized as a stain for SDS–PAGE separation
of proteins,11,14 to label proteins for highly sensitive quantitative
analysis using CE,15–19 as well as to investigate protein–aptamer
complexation by means of kinetic capillary electrophoresis
(KCE).20
In this article we explore, for the first time, the use of Py-1 to
label, characterize, and profile oligosaccharide mixtures. In an easy,
two-step process, the oligosaccharides are first equipped with a
primary amino group, by reaction with one end of ethylenedia-
mine, leaving a free primary amine, which is subsequently reacted
with Py-1 (Scheme 1). Though not made use of in the present
study, the remaining secondary amine is available for introduction
of additional functionalities such as stable isotope tags for mass
While there are many different fluorescent labels that have
been developed for oligosaccharide labeling using, for example,
reductive amination or hydrazone formation at the reducing-end
aldehyde as a means of conjugation,5,6 a common limitation is
the need to remove the large quantities of excess dye required
for the derivatization before analyses can be made. Fluorogenic
⇑
Corresponding author. Tel.: +45 33275382; fax: +45 33274708.
E-mail address: Ole.Hindsgaul@carlsberglab.dk (O. Hindsgaul).
0008-6215/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2012.02.016

S. A. Johannesen et al. / Carbohydrate Research 352 (2012) 94–100
95
-
-
Maltohexaose was reacted with ethylenediamine in MeOH in
the presence of CH3COOH and NaBH3CN. The reductive amination
proceeds via imine formation with the open-chain aldehyde form
of the reducing-end sugar. As the aldehyde form of the reducing-
end sugar is typically only present in small amounts—the pyranose
form being the significantly more stable tautomer—a large excess
of ethylenediamine is required to drive the equilibrium toward
imine formation. This large excess helps to promote the reaction
in the case of large oligosaccharides, where the reducing-end sugar
may be less reactive, while preventing dimerization.
BF4
X
N
(a)
R
O
R
NH2
+
N
N
Py-1
(b)
Transformation of maltohexaose to the glycamine, maltohexa-
2A
1
Py-1A
ose-NHCH2CH2NH2 (1), was confirmed by 1H NMR spectroscopy
(Fig. 2). Upon reductive amination, the signals for the
a and b ano-
2F
meric protons of the reducing-end sugar (H1 and H1b) disappear.
a
0.8
0.6
0.4
0.2
0
This is accompanied by the significant upfield shift (0.2 ppm) of the
signal for the anomeric proton on the glucose residue adjacent to
the reducing-end sugar (H10). At neutral to basic pH, the product
reacts reversibly with carbon dioxide to generate a carbamic acid.
This species may be observed in the 1H NMR spectrum in slow ex-
change with the free amine (Supplementary data, Fig. S1). Upon
acidification, the spectrum is simplified and a single species—the
protonated amine—is observed, as seen in Figure 2a.
Maltohexaose-NHCH2CH2NH2 (1) was isolated by careful pre-
cipitation following dropwise addition of ethanol to the partially-
evaporated reaction mixture. Removal of all excess ethylenedia-
mine salt at this stage is crucial, as it can react with the pyrylium
dye in the next step and would therefore give misleading results
when the fluorescence of the labeled solution is measured directly.
The methylene protons of the ethylene diamine salt, if present, ap-
pear in the 1H NMR spectrum as a singlet at 3.45 ppm.
A, F
Py-1F
350
450
550
650
750
(nm)
Figure 1. (a) Conjugation of Py-1 to a primary amine via pyridinium ion formation
(Xꢁ is the counter-ion and depends upon the reaction condition and purification
method) and (b) normalized UV–vis absorbance spectra (—) and fluorescence
emission spectra (kex. = 500 nm) (---) of Py-1-maltohexaose (2) and pyrylium dye,
Py-1.
The reductive amination with ethylene diamine was then per-
formed on various mixtures of oligosaccharides, including maltool-
igosaccharides (primarily
a
1–4 linked glucose residues) and more
1–6 linked) starch derivatives. The
highly branched ( 1–4 and
a
a
reaction may be performed in MeOH/DMSO mixtures or in H2O
depending on the solubility of the substrate; larger, more highly
branched glycans are less soluble in MeOH mixtures. Smaller oligo-
saccharides, should be isolated from the reaction mixture by precip-
itation with EtOH, but larger oligosaccharides are preferentially
precipitated using MeOH. Excess ethylenediamine has a tendency
to precipitate with the larger oligosaccharides but re-precipitation
from H2O with MeOH and sonication of the precipitate in MeOH,
are effective methods of purification.
spectrometry or biotin for capture. Herein, we evaluated the use of
Py-1 for analysis of maltooligosaccharides using CE and MALDI-
ToF-MS.
2. Results and discussion
2.1. Labeling of oligosaccharides
2.1.1. Reductive amination with ethylenediamine
The first step in the conjugation of oligosaccharides with Py-1 is
the reductive amination of the reducing-end aldehyde (hemi-ace-
tal) with ethylenediamine. We first tested the labeling method
using maltohexaose, as a model compound, before moving to mix-
tures of larger oligosaccharides (Scheme 1).
2.1.2. Conjugation with the pyrylium dye (Py-1)
Reaction of pyrylium dye (Py-1) with ethylenediamine-func-
tionalized oligosaccharides under basic conditions readily gener-
ates fluorescent pyridinium-labeled sugars (Scheme 1). The
OH
OH
O
O
OH
Glc4
O
Glc4
Glc4
O
O
O
CH3COO-
a
H
H
H
OH
HO
HO
HO
OH
OH
OH
OH
OH
OH
OH
OH
OH
O
O
O
H
N
O
O
OH
HO
HO
HO
+
NH3
OH
1
OH
Glc4
-
BF4
O
O
H1'
O
HO
OH
O 4
H
N
1
OH
CH3COO-
+
6
5
b
OH
2OH1
8
HO
N
N
3
12
7
13 14
15
16
9
10
11
Py-1
2
N
Scheme 1. Reagents and conditions: (a) ethylenediamine, NaBH3CN, CH3COOH, MeOH; (b) Et3N, DMSO, MeOH, MeCN.

96
S. A. Johannesen et al. / Carbohydrate Research 352 (2012) 94–100
mine-functionalized sugars with Py-1 confers a permanent posi-
(a)
(b)
tive charge. Samples of maltohexaose (Fig. 3a), and mixtures of
polydisperse maltooligosaccharides, with varying degrees of poly-
merization (DP) (mixtures A and B) (Fig. 3b and c), were labeled
with Py-1 and analyzed.
Analysis by MALDI-ToF-MS was performed without the need for
work-up or purification following the labeling reaction. Aliquots of
reaction mixture were diluted with H2O then mixed with the ma-
trix solution and applied to the plate. Recrystallization on the plate,
H1 -H1
H1 -H1
H1
by addition of 1 lL H2O, was then carried out, prior to analysis. A
H1
H1
Figure 2. 1H NMR spectra (300 K): (a) maltohexaose-NHCH2CH2NH2 (1) in 1%
CH3COOH in D2O and (b) maltohexaose in D2O.
reaction is preferentially performed in non-aqueous solvent at
room temperature, but also proceeds in H2O upon heating.
We first tested the reaction on maltohexaose-NHCH2CH2NH2
(1). The glycamine was dissolved in a mixture of DMSO and MeOH.
Triethylamine was added, followed by a fivefold excess of pyrylium
dye (Py-1) as a suspension in MeCN. The reaction mixture quickly
turned from deep blue to red. A control experiment showed that in
the absence of amine, the solution turns instead from blue to green
and then yellow. Conversion of the amine to the pyridinium ion
could be monitored by ESI-MS and it was observed that the reac-
tion proceeded readily at room temperature and appeared com-
plete within 2 h. The reaction was subsequently performed on
mixtures of oligosaccharides. Where solubility permitted, the glyc-
amines were dissolved in mixtures of DMSO and MeOH. However,
for larger oligosaccharides, for example, partially hydrolyzed b-
limit dextrin mixtures, the aminated glycans were dissolved in
H2O, a larger excess of dye was added and the reaction mixture
was heated for 2 h at 60 °C.
Purification of the labeled sugars was not necessary prior to
analysis by MALDI-ToF-MS, CE, or direct fluorescence spectros-
copy. We found, however, that if samples were to be stored, or
used for further study, isolation of the product from the reaction
mixture was important to prevent decomposition over longer peri-
ods of time. Dilution of the reaction mixtures in H2O, therefore, al-
lowed the removal of some precipitated excess dye, as well as
insoluble side products. The remaining dye could then be extracted
with CH2Cl2 following acidification or alternatively, the diluted
reaction was concentrated to dryness, then resuspended in a small
amount of H2O, then centrifuged, or filtered to remove the insolu-
ble dye. Further purification by chromatography using Sep-pak C18
and gradient elution with 0.2% CH3COOH H2O/MeCN mixtures re-
moved the triethylammonium salt, remaining excess dye, and
impurities from the labeled oligosaccharide mixtures.
2.2. Analysis of Py-1-labeled oligosaccharides
2.2.1. MALDI-ToF-MS
MALDI-ToF-MS is a powerful tool for the analysis of high-mass
molecules and mixtures of polymers. However, the analysis of neu-
tral oligosaccharides using this technique is hindered by their low
ionization potential and poor proton affinity.5–8 Chemical modifi-
cation, in particular via the introduction of a permanent charge,
can significantly increase the detectable mass range and sensitivity
of analysis using MALDI-ToF-MS.21 Conjugation of ethylenedia-
Figure 3. MALDI-ToF-MS analysis of Py-1-labeled oligosaccharides: (a) Py-1-
maltohexaose (2), (b) maltooligosaccharide mixture A (from MALTRIN M040, Grain
Processing Corporation) and (c) maltooligosaccharide mixture B (from maltodex-
trin, Aldrich). In (c) are shown several overlaid analyses; the laser intensity was
increased to observe higher-mass oligosaccharides. Selected peaks are labeled with
the DP of the labeled oligosaccharide.

S. A. Johannesen et al. / Carbohydrate Research 352 (2012) 94–100
97
variety of matrices were tested and the best results were obtained
using 4-hydroxy- -cyanocinnamic acid (HCCA) as a saturated
(a)
a
solution in H2O/EtOH (1:1). We found that it was important not
to add additional acid to the matrix mixture—HCCA is often used
as a saturated solution in 0.1% TFA H2O/MeCN (1:1)—as this led
to acid-induced decomposition of the analyte on the plate.
The positive-ion MALDI mass spectrum of Py-1-maltohexaose
(2) shows a single peak at 1322.6 m/z corresponding to the [M]+
ion. The maltooligosaccharide mixtures were observed as a series
of single peaks separated by mass intervals of 162, which is the
mass difference expected for each additional glucose residue in
an oligosaccharide. In the higher DP maltooligosaccharide mixture
(mixture B) it was possible to convincingly observe labeled oligo-
saccharides with up to 58 glucose monomers, which corresponds
to [M]+ = 9746 m/z (Fig. 3c). The permanent positive charge on
the label avoids the complication of Na+ and K+ adducts, simplify-
ing the spectrum and improving sensitivity. The sensitivity of the
technique was tested on Py-1-maltohexaose (2) and we found that
the analysis could be made with as little as 25 pg (18 fmol) of com-
pound added to the target.
Py-1-maltohexaose (2)
(b)
While MALDI-ToF-MS is very useful for identifying the Py-1-la-
beled oligosaccharides present in a mixture, it cannot be used to
profile the size distributions of mixtures of maltooligosaccharides.
The analysis will typically be biased toward smaller oligosaccha-
rides, which are more prone to laser desorption, and furthermore,
the inherently inhomogeneous nature of the target means that the
distribution of oligosaccharides identified will not necessarily be
representative of the composition of the whole mixture. An addi-
tional analysis technique is required.
Time / min
20
15
Figure 4. Capillary electropherograms of Py-1-labeled oligosaccharide mixture (A)
in the absence (a) and presence (b) of Py-1-labeled maltohexaose (2).
labeled oligosaccharides should be possible without the need for
reaction workup and the removal of excess dye. In order to evalu-
ate the applicability of Py-1 in fluorescence assays a test of linear-
ity was conducted using maltohexose-NHCH2CH2NH2 (1).
A series of solutions of 1 in DMSO/MeOH, ranging in concentra-
tion from 8 to 160 lM (0.02–0.1 mg of sugar) and a control con-
2.2.2. Capillary electrophoresis
Capillary electrophoresis (CE) provides a fast, and efficient
method to analyze the size distribution of labeled oligosaccharide
mixtures. Where the oligosaccharides are labeled with a fluores-
cent tag, the technique can also be highly sensitive. We found that
Py-1-labeled maltooligosaccharides displayed good electropho-
retic properties when using a borate (pH 9.3) running buffer con-
taining an anionic surfactant, specifically, sodium dodecylsulfate
(SDS), and could be analyzed at concentrations as low as 50 nM
using our Laser Induced Fluorescence (LIF) Detector (Supplemen-
tary data Fig. S2). As the oligosaccharides are labeled at the reduc-
ing-end, there can be only one label per molecule, which means
that the peak areas in the electropherogram accurately reflect the
relative molar concentrations of the different length
oligosaccharides.
taining no sugar, were reacted with Py-1 in DMF in the presence
of Et3N and heated to 60 °C for 2 h. After approximately 10-fold
dilution with H2O, and no further work up or purification, these
solutions were analyzed using a fluorescence microplate reader
with kex./kem. = 495/635 nm. We observed that, despite the fluoro-
genic nature of the Py-1 label, the control sample, containing no
sugar, gave rise to some background fluorescence. The same has
been observed by Dovichi in the analysis of Py-1-labeled a-lactal-
bumin and most-likely is produced by hydrolysis/degradation
products from the unreacted dye.17 By subtracting this signal from
the measured fluorescence of the Py-1-labeled sugars, however, we
were able to observe a linear correlation between concentration of
1 and the fluorescence at 635 nm over a 1–20 lM concentration
range (Fig. 5). The experiment was replicated several times, and
A sample of labeled maltooligosaccharides (mixture A) taken di-
rectly from the labeling reaction mixture, diluted with H2O (1 mg
oligosaccharide mixture/10 mL H2O) and centrifuged to remove
precipitates, was analyzed by CE (Fig. 4a). The solution was then
spiked with Py-1-maltohexaose (2) as a mobility marker in order
to determine the degree of polymerization of the oligosaccharides
present (Fig. 4b). In the electropherograms, the larger sugars elute
first. They show well separated peaks for the smaller sugars and
distinguishable peaks up to a degree of polymerization of 21. The
last peak in the electropherogram (20.5 min) is due to labeling of
the trace amounts of ethylenediamine present in the reductively
aminated sugar mixture. Since Py-1 is a fluorogenic dye, and the
unconjugated dye shows almost no fluorescence above 570 nm,
the presence of excess dye in the analyte does not adversely affect
the analysis by CE.
700
600
500
400
300
200
100
0
0
5
10
[1] (µM)
15
20
2.2.3. Quantitative analysis by direct fluorescence measurement
The advantage of Py-1 over more commonly employed fluores-
cent labels results from the very large difference in fluorescence
intensity between the conjugate pyridinium ion and the free
pyrylium dye (Fig. 1); direct, quantitative fluorescence analysis of
Figure 5. Fluorescence (kex./kem. = 495/635 nm) measurements made directly on
three different series of solutions of maltohexaose-NHCH2CH2NH2 (1) labeled with
Py-1.

98
S. A. Johannesen et al. / Carbohydrate Research 352 (2012) 94–100
within this range, linear regression analysis gave R2 >0.99 on each
occasion.
pyridinium formation facilitates analysis by MALDI-ToF-MS. The
labeling method is targeted toward larger oligosaccharides,
which are more easily isolated by precipitation, and since the la-
beled oligosaccharides can be effectively separated by CE, Py-1 is
a good candidate for the analysis of oligosaccharide using a CE-
MS22 set-up.
It should be noted, however, that direct quantitative analysis of
oligosaccharides will only be accurate if all excess ethylenediamine
has been removed from the mixture following reductive amination
and prior to Py-1 conjugation, as any ethylenediamine present will
also be labeled and give an artificially high fluorescence recording.
Direct fluorescence measurements are therefore most appropriate
for small oligosaccharides (DP <10). Care must be taken that all
ethylenediamine is removed, and direct fluorescence analysis is
best accompanied by CE analysis for profiling and quantification
of larger oligosaccharides.
4. Experimental
4.1. General methods
Sodium cyanoborohydride, triethylamine, ethylenediamine,
and maltohexaose (94%) were obtained from Sigma–Aldrich. The
Py-1 dye was purchased from Active Motif Chromeon GmbH,
Tegernheim, Germany. Maltooligosaccharide mixtures were pre-
pared as detailed below from: (A) MALTRIN M040, donated by
the Grain Processing Corporation, Muscatine, IA, USA and (B) mal-
todextrin (DE = 16.5–19.5), acquired from Aldrich. Solvents used
were all HPLC-grade and were dried over molecular sieves where
specified. D2O was obtained from Cambridge Isotope Laboratory.
NMR spectra were recorded on a Bruker Avance III 400 MHz
spectrometer or Bruker 500 MHz CryoProbe instrument operating
at 300 K unless otherwise specified. Spectra are internally refer-
enced to the solvent residue. Spectra were processed using Bruker
Topspin 2.0. Assignments of 1H and 13C NMR signals were based on
COSY and HSQC spectra.
UV–vis spectra were acquired at room temperature using either
a Perkin Elmer precisely Lambda 45 UV/Vis spectrometer, or for
more concentrated solutions a Saveen Werner Nanodrop Spectro-
photometer. Fluorescence spectra were recorded on a Perkin Elmer
Luminescence Spectrometer L5 50 B operating at room tempera-
ture. High-resolution electrospray ionization mass spectrometry
was performed on a Bruker MicroTOF-Q II instrument using HCO-
ONa clusters as the internal standard.
2.3. Label stability
Py-1 has been developed primarily as a label for proteins to
facilitate analysis using various techniques, such as SDS–PAGE
and CE. For these techniques, the label need only be stable long en-
ough for the analysis to be completed. We have shown oligosac-
charides can be derivatized with Py-1 in order to carry out
analysis using CE and MALDI. However, we were also interested
to investigate the stability of the Py-1-conjugates and to assess
whether isolated samples could be employed in biological studies.
We tested the stability of Py-1 conjugates in acidic aqueous
solutions. Solutions of Py-1-labeled maltohexaose (2) (0.07 mM)
in sodium citrate buffer (90 mM) at pH 2, 3, and 4, and in milli-Q
H2O were analyzed by UV/Vis spectroscopy, then incubated at
37 °C overnight, and reanalyzed (Supplementary data, Fig. S3). It
was found that while the UV–vis spectrum of the sample in
milli-Q H2O remained constant, in the acidic buffered solutions,
the absorbance at ꢀ500 nm decreased significantly during the
24 h period: the lower the pH, the greater the decrease in absor-
bance. This change could also be observed by the naked eye in
the decoloration of the red solutions; the solution at pH 2 was al-
most colorless after 24 h at 37 °C. The solutions were analyzed by
MALDI-ToF-MS and we observed that the decomposition was
accompanied by a decrease in the [M]+ ion at 1322 m/z, and the
appearance of new ions with 4 mass units lower (Supplementary
data, Fig. S4). In another experiment, the decomposition of two
samples of Py-1-maltohexaose (1 mM), one in D2O and the other
in 1% CH3COOH in D2O, maintained at 37 °C over a period of days,
was monitored by 1H NMR spectroscopy (Supplementary data,
Figs. S5 and S6). While the faster decomposition in the acidic solu-
tion was obvious, we were unable to identify the degradation prod-
uct/s. Wolfbeis has previously reported that fixation of Py-1-
labeled proteins separated on SDS–PAGE using a standard 5:4:1
(v/v/v) mixture of MeOH, H2O and CH3COOH caused the intensity
of the fluorescently-labeled proteins to decrease.14 This observa-
tion is consistent with the acid-induced decomposition/decoloriza-
tion we have encountered.
4.2. Preparation of maltooligosaccharide mixtures (A and B)
Maltooligosaccharide mixture A was obtained from MALTRIN
M040 by fractional dissolution in EtOH/H2O mixtures. One hun-
dred grams of MALTRIN M040 was suspended, with mechanical
stirring, in 80% (v/v) EtOH/H2O (1200 mL) at ambient temperature
for 16 h. The mixture was sedimented and decanted. The superna-
tant was evaporated in vacuo to give Fraction 1 (5.9 g). The solid
collected after decanting was re-suspended in 71% (v/v) EtOH/
H2O (1500 mL) at room temperature for 16 h. The suspension
was sedimented, decanted, and the supernatant evaporated in va-
cuo to give Fraction 2 (8.10 g). Further fractionation was carried out
to give: Fraction 3 (3.1 g, from 1000 mL 65% (v/v) EtOH/H2O); Frac-
tion 4 (9.8 g from 50% (v/v) EtOH/H2O) and Fraction 5 (68.5 g, as the
remaining insoluble material).
The degree of polymerization and extent of 1–6 branching in
the various fractions were determined by methylation analysis
performed following reduction of the samples with NaBH4. Per-
O-methylation was carried out using the NaOH/DMSO/CH3I system
described by Ciucanu and Kerek.23 The routine processes of hydro-
lysis (2 M TFA), reduction (NaBD4), and acetylation (Ac2O–pyri-
dine) were then performed followed by GC-MSD analysis to
detect the partially methylated acetate alditols. Fraction 2, used
in this study, was determined to have an average size of 12 glucose
residues and only 1–2% 1–6 branches.
From these experiments, we concluded that Py-1-conjugates
are unstable in acidic solution, and labeled samples should be
stored in the freezer as solids, evaporated from neutral solutions.
Given the acid sensitivity of the Py-1 conjugates, we recommend
its use primarily for fast CE or MS analysis of oligosaccharide anal-
ysis, for which purpose, we have found it to be very well suited.
3. Conclusion
Oligosaccharide labeling using Py-1 provides a fast, effective,
and sensitive means to analyze and profile oligosaccharide mix-
tures by MALDI-ToF-MS and CE. Labeling with this fluorogenic
pyrylium dye eliminates the need to remove the excess dye used,
prior to fluorescence analysis, since the free dye is non-fluores-
cent. The introduction of a permanent positive charge upon
Maltooligosaccharide mixture B was produced from maltodex-
trin (dextrose equivalent 16.5–19.5). Maltodextrin (200 mg) was
dissolved in H2O (10 mL), then filtered through a Vivaspin 3000
Mw cut-off filter, using
a non-fixed angle centrifuge with
2000 rpm for 99 min. The upper fraction was diluted and re-filtered

S. A. Johannesen et al. / Carbohydrate Research 352 (2012) 94–100
99
five times. The resulting upper fraction was subsequently labeled
with Py-1.
the solid washed with MeOH (10 mL). Purification of the product
was achieved by reprecipitation with MeOH (20 mL) from H2O
(1.5 mL) and subsequent suspension and sonication of the aminat-
ed sugars in MeOH. The solid was then collected and dissolved in
1% CH3COOH (4 mL), and then evaporated in vacuo.
4.3. Py-1 labeling of maltohexaose
4.3.1. Maltohexaose-NHCH2CH2NH2 (1)
The aminated glycan mixture (1 mg) was dissolved in H2O
To a solution of maltohexaose (50 mg, 50
lmol) in a mixture of
(200
lL). Et3N (10
lL, 70
lmol) was added to the solution followed
MeOH (95 mL), acetic acid (1.15 mL) and ethylenediamine
(0.65 mL, 9.7 mmol) was slowly added a solution of NaBH3CN
(500 mg, 8.0 mmol) in MeOH (5 mL). The solution was refluxed
(60 °C) for 3 h then the solvent was reduced to approximately
5 mL in vacuo. EtOH (45 mL) was added dropwise to precipitate
the product as a fine white powder. The mixture was centrifuged,
the supernatant was decanted, and the solid was washed with
EtOH (4 ꢂ 10 mL). The pure product was obtained after reprecipi-
tation from H2O (0.2 mL) with EtOH (10 mL), and washing with
EtOH (5 mL). The precipitate was then dissolved in 0.2% CH3COOH
(1 mL) and the solvent evaporated in vacuo to give product 1 as a
colorless solid (22 mg, 40%). 1H NMR (400 MHz, 1% CH3COOH in
D2O): d 5.45 (br s, 4H, H100-H100000), 5.18 (br s, 1H, H10), 4.25 (br s,
2H, CH2N), 4.10–3.36 (m); 13C NMR (100 MHz, 1% CH3COOH in
D2O): d 100.2 (C10), 100.1 (C100-C100000), 81.9, 77.5, 77.5, 77.4, 77.3,
73.8, 73.7, 73.4, 73.2, 72.9, 72.2, 72.0, 71.9, 71.7, 71.5, 69.8, 67.4,
62.7, 61.0, 60.9, 60.8, 57.9, 51.3, 44.9, 35.9.
by a suspension of Py-1 (2.0 mg, 5
l
mol) in MeCN (200 L). The
l
reaction mixture was stirred at 60 °C for 2 h (during which time
the initially blue solution turned bright red with some yellow pre-
cipitate). The mixture was then diluted to 10 mL with H2O, centri-
fuged, and decanted. The mixture can be analyzed directly at this
point, or otherwise purified in the same manner as described for
Py-1-maltohexaose (2).
4.5. Analysis
4.5.1. MALDI-ToF-MS
A saturated solution of 4-hydroxy-a-cyanocinnamic acid in 50%
ethanol in water (13.3 mg/mL) was prepared as the matrix. An ali-
quot of labeling reaction mixture was then diluted with H2O (typ-
ically 1/100 dilution gave the best results). The matrix solution
(20
and vortexed. This solution (0.5
dried, and then dried in a vacuum. H2O (1
l
L) was combined with the diluted analyte solution (20
L) was added to the target, air-
L) was added to the
lL)
l
l
4.3.2. Py-1-maltohexaose (2)
spotted sample, which was again air-dried and vacuum-dried be-
Amine 1 (5.0 mg, 4.6
DMSO (0.35 mL), MeOH (5 mL) and Et3N (50
the solution was added a suspension of Py-1 (10 mg, 25
l
mol) was suspended in a mixture of
L, 0.36 mmol). To
mol), in
fore analysis.
l
MALDI-ToF-MS was performed on a Bruker Daltonics Microflex
instrument operating in reflectron mode. A 340 nm laser was used
and mass spectra were typically accumulated from 75 laser shots.
Spectra were generally acquired over a 4000 m/z range, and the
laser power was increased to obtain a signal for larger sugars.
l
MeCN (1 mL). The reaction mixture was stirred at room tempera-
ture for 3 h (during which time the initially blue solution turned
deep red), then diluted with H2O (47 mL), centrifuged, to sediment
a yellow/green solid, and decanted. The supernatant was evapo-
rated to dryness in vacuo then resuspended in H2O (4 mL), centri-
fuged, to sediment a black/dark blue solid, then decanted. The
supernatant was filtered, diluted with H2O (6 mL), and purified
by careful chromatography (Waters Sep-pak C18 plus: 2% MeCN
in 0.2% CH3COOH (aq) to 16% MeCN in 0.2% CH3COOH (aq)) to give
Py-1-labeled maltohexaose (2) as red solid (3.9 mg, 61%). UV–vis
(D2O) kabs. max. = 500 nm, kem. max. = 615 nm; 1H NMR (400 MHz,
D2O): d 7.72 (s, 2H, H10 or H13), 7.66 (d, 2H, 3J = 15.6 Hz, H11 or
H12), 7.29 (s, 2H, H13 or H10), 6.96 (d, 2H, 3J = 15.6 Hz, H12 or
H11), 5.45 (m, 4H, H100–H100000), 5.15 (br s, 1H, H10), 4.17–3.63 (m,
35H), 3.49 (dd apparent t, 1H, 3J = 9.3 Hz), 3.43 (br m, 2H, H7),
3.32 (t, 4H, 3J = 5.5 Hz, H16), 3.17 (br m, 2H, H1), 2.86 (s, 6H, H9),
2.84 (t, 4H, 3J = 6.1 Hz, H14), 2.02 (m, 4H, H15). (Note that the res-
idue for H8 occurs under the suppressed H2O signal); 13C NMR
(125 MHz, D2O): d 181.16, 153.75, 145.54, 141.74 (C12 or C11),
127.46 (C13 or C10), 122.84, 122.62 (C10 or C13), 122.44, 116.53
(C11 or C12), 100.28 (C10), 99.9–99.6 (C100–C100000), 81.55, 73.33,
73.29, 73.21, 71.69, 71.48, 71.21, 71.18, 69.27, 62.17, 60.40,
60.32, 57.37, 51.05 (C1), 49.69 (C16), 48.22 (C8), 45.07 (C7),
26.89 (C14), 20.94 (C15), 20.40 (C9); HR-MS (ESI): m/
z = 1322.5785 [M]+ (calcd for C59H92N3O30: 1322.5760), 661.7921
[M+H]2+ (calcd for C59H93N3O30: 661.7916).
4.5.2. Capillary electrophoresis
Capillary electrophoresis was carried out on the labeling reac-
tion mixture diluted with H2O (1 mg oligosaccharide/10 mL
H2O). CE analysis was performed on a PrinCE 560 (Prince Tech-
nologies, Netherlands) capillary electrophoresis system, equipped
with a fused-silica capillary (Prince Technologies, Netherlands) of
75 lm I.D. and a total length of 85 cm (60 cm effective length).
The running buffer was borate buffer (50 mM, pH 9.3) containing
150 mM SDS and the samples were analyzed at 25 kV, with 10 s
injection (50 mbar), outlet = cathode. Detection was carried out
by on-column measurement of the fluorescence using either an
Argos 250B FL-detector (FLUX Instruments, Switzerland) with
excitation and emission cut-off filters (kex./kem. = 546/570 nm)
or
a Picometrics Evalution LIF detector with excitation at
532 nm.
4.5.3. Direct fluorescence analysis
A stock solution of maltohexaose-NHCH2CH2NH2 (1) was pre-
pared in dry DMSO (1.8 mM, 500
charged with 50, 40, 30, 20, 10, 2.5, or 0
DMSO (0, 10, 20, 30, 40, 47.5 or 50 L) was added to obtain a total
volume of 50 L in each vial. To each solution was then added dry
MeOH (500 L), Et3N (1 L), and a solution of Py-1 in DMF (7.5 L,
l
L). Seven reacti-vials were
l
L of this solution. Dry
l
l
l
l
l
4.4. Py-1-labeling of mixtures of maltooligosaccharides with
higher DP
25 mM). The solutions were stirred at 60 °C for 2 h, then cooled to
room temperature, and diluted to 5 mL with H2O. Six aliquots of
each solution (200 lL) were transferred to a Black clear-bottomed
To a solution of the oligosaccharide mixture (40 mg) in H2O
(3 mL) was added CH3COOH (2.5 mL) and ethylenediamine
(0.4 mL) followed by a solution of NaBH3CN (400 mg) in H2O
(0.4 mL). The solution was then stirred at 60 °C for 3 h. The reaction
mixture was evaporated in vacuo to reduce the volume by half
before precipitation by addition of MeOH (20 mL). The mixture
was sedimented by centrifuging, the supernatant decanted, and
microplate for analysis.
Fluorescence measurements were performed using a Spectra-
Max Gemini EM dual scanning microplate spectrofluorometer
(Molecular Devices Corporation, California) with excitation at
485 nm, cut-off above 495 nm, and emission measured at
635 nm. Twenty readings per well were acquired with bottom
read.

100
S. A. Johannesen et al. / Carbohydrate Research 352 (2012) 94–100
10. Udenfriend, S.; Stein, S.; Böhlen, P.; Dairman, W.; Leimgruber, W.; Weigele, M.
Supplementary data
Science 1972, 178, 871–872.
11. Wetzl, B. K.; Yarmoluk, S. M.; Craig, D. B.; Wolfbeis, O. S. Angew. Chem., Int. Ed.
2004, 43, 5400–5402.
12. O’Leary, M. H.; Samberg, G. A. J. Am. Chem. Soc. 1971, 93, 3530–3532.
13. Caro, B.; Le Guen-Robin, F.; Salmain, M.; Jaouen, G. Tetrahedron 2000, 56, 257–
263.
Supplementary data (stability data, CE sensitivity measure-
ments and NMR titrations and spectra) associated with this article
can be found, in the online version, at doi:10.1016/
j.carres.2012.02.016.
14. Meier, R. J.; Steiner, M.-S.; Duerkop, A.; Wolfbeis, O. S. Anal. Chem. 2008, 80,
6274–6279.
15. Craig, D. B.; Wetzl, B. K.; Duerkop, A.; Wolfbeis, O. S. Electrophoresis 2005, 26,
2208–2213.
References
16. Wojcik, R.; Swearingen, K. E.; Dickerson, J. A.; Turner, E. H.; Ramsay, L. M.;
Dovichi, N. J. J. Chromatogr., A 2008, 1194, 243–248.
17. Turner, E. H.; Dickerson, J. A.; Ramsay, L. M.; Swearingen, K. E.; Wojcik, R.;
Dovichi, N. J. J. Chromatogr., A 2008, 1194, 253–256.
18. Ramsay, L. M.; Dickerson, J. A.; Dovichi, N. J. Electrophoresis 2009, 30, 297–302.
19. Dickerson, J. A.; Ramsay, L. M.; Dada, O. O.; Cermak, N.; Dovichi, N. J.
Electrophoresis 2010, 31, 2650–2654.
20. de Jong, S.; Krylov, S. N. Anal. Chem. 2011, 83, 6330–6335.
21. Gouw, J. W.; Burgers, P. C.; Trikoupis, M. A.; Terlouw, J. K. Rapid Commun. Mass
Spectrom. 2002, 16, 905–912.
1. Krishnamoorthy, L.; Mahal, L. K. ACS Chem. Biol. 2009, 4, 715–732.
2. Novotny, M. V.; Mechref, Y. J. Sep. Sci. 2005, 28, 1956–1968.
3. Mechref, Y.; Novotny, M. V. Chem. Rev. 2002, 102, 321–369.
4. Anumula, K. R. Anal. Biochem. 2006, 350, 1–23.
5. Ruhaak, L. R.; Zauner, G.; Huhn, C.; Bruggink, C.; Deedler, A. M.; Wuhrer, M.
Anal. Bioanal. Chem. 2010, 397, 3457–3481.
6. Pabst, M.; Kolarich, D.; Pöltl, G.; Dalik, T.; Lubec, G.; Hofinger, A.; Altmann, F.
Anal. Biochem. 2009, 384, 263–273.
7. Lamaria, F. N.; Kuhnb, R.; Karamanosa, N. K. J. Chromatogr., B 2003, 793, 15–36.
8. Harvey, D. J. Int. J. Mass Spectrom. 2003, 226, 1–35.
22. Mechref, Y.; Novotny, M. Mass Spectrom. Rev. 2009, 28, 207–222.
23. Ciucanu, I.; Kerek, F. Carbohydr. Res. 1984, 131, 209–217.
9. Mohr, G. J. Anal. Bioanal. Chem. 2006, 386, 1201–1214.