Solid-phase oligosaccharide tagging (SPOT): Validation on glycolipid-derived structures

Article abstract of DOI:10.1002/anie.200600642

(Chemical Equation Presented) SPOT check: A simple technique for labeling oligosaccharides consists of capturing the sugars on beads functionalized with hydroxylamine groups bearing cleavable linkers. After sequential pipetting in and out of reagents, tag

Full text of DOI:10.1002/anie.200600642

Angewandte
Chemie
Oligosaccharide Analysis
DOI: 10.1002/anie.200600642
Solid-Phase Oligosaccharide Tagging (SPOT):
Validation on Glycolipid-Derived Structures**
Anders Lohse, Rita Martins, Malene Ryborg Jørgensen,
and Ole Hindsgaul*
Glycosylation is the most abundant form of post-translational
modification of proteins. Over 50% of all protein sequences
are glycosylated in eukaryotic systems.[1] Glycosylated lipids
are also widespread in animal cells: they constitute up to 5–
10% of the cell membrane content and are responsible for a
wide array of pathological disorders.[2]
The structural analysis, or sequencing, of the oligosac-
charide chains of glycoproteins and glycolipids is much more
complex than that of DNA or proteins, as the molecules are
generally branched and the linkage between sugar residues
introduces a new a or b stereocenter. The major analytical
approaches[3] rely on the detection of oligosaccharides either
by mass spectrometry (MS) or pulsed amperometry. Oligo-
saccharides released from glycoproteins or glycolipids by
chemical or enzymatic methods are also frequently labeled at
the reducing end to afford increased sensitivity of detection.[4]
The insertion of fluorescent tags followed by analysis by
chromatography or electrophoresis is common practice. The
above approaches, however, remain complex endeavors that
require a combination of sophisticated equipment and a high
degree of expertise in both sample handling and data analysis.
There would be clear advantages in any technique that
could sequester a reducing oligosaccharide from solution and
allow its further manipulation (including tagging) while
immobilized on a solid phase, for example, on beads or
glass surfaces such as slides. This is the rationale behind solid-
phase oligosaccharide tagging (SPOT) that would involve
only pipetting and washing. The advantages are:
*
Once the oligosaccharide is “captured” it remains cova-
lently bound to the beads, so there would be no sample loss
during subsequent manipulations.
*
Large excesses of tagging (or other) reagents could be used
to drive reactions to completion, as they would be washed
away without sample loss or dilution.
*
If the immobilized oligosaccharide remained sterically
accessible to enzymes and lectins (or antibodies), biolog-
ical oligosaccharide recognition and enzyme-assisted
[*] Dr. A. Lohse, Dr. R. Martins, Dr. M. R. Jørgensen, Prof. O. Hindsgaul
Carlsberg Laboratory
2500 Valby, Copenhagen (Denmark)
Fax: (+45)3327-4708
E-mail: hindsgaul@crc.dk
[**] This work was supported by the Danish Technical Research Council
(STVF).
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2006, 45, 4167 –4172
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4167

Communications
sequencing procedures already in use in glycomics could
be applied.
If a cleavable linker was present in the capture molecule, a
“processed” (for example, fluorescently tagged) oligosac-
charide could be released into solution in a small volume
for further analysis.
oxime nor the immobilized sugar OH groups react. Reduction
of the oxime double bond in C would then give the N,O-
alkylhydroxylamine D, now with a reactive sp3 nitrogen atom.
Tagging of the active nitrogen atom in D can be accomplished
under conditions in which the sugar OH groups do not react
(for example, by using aryl isothiocyanates) to give E. Base
treatment of E then releases the tagged sugar F into solution
for analysis. In the process, the immobilized oligosacchar-
ide(s) in B–E should be detectable by lectins or antibodies,
thus yielding potentially valuable information on their
structure(s).
The process described has, as its key step, the selective
inactivation of excess capture groups prior to reduction and
insertion of a tag on the sugar. In this way, a single tag is added
only to the sugar molecule, even though an excess of tagging
agent is used. The excess tagging agent, and indeed all
reagents used, are easily washed away from the immobilized
sugar, which thereby remains in a small volume.
*
Recently, Nishimura and co-workers[5] showed that poly-
meric beads functionalized with hydroxylamine groups could
be used to capture reducing oligosaccharides from complex
samples. After washing away the non-carbohydrate debris,
the oligosaccharide could be released back into solution for
further analysis. This convenient cleanup procedure, termed
“glyco-blotting”, yielded superior MS data for use in frag-
mentation sequencing. Shin and co-workers[6] later showed
that oligosaccharides could be captured on glass array slides
functionalized with hydrazide or hydroxylamine groups, and
detected on the slides using fluorescently tagged lectins.
Herein, we show that a system can be devised where solid
supports bearing hydroxylamine groups can capture oligo-
saccharides from solution. Furthermore, the immobilized
product can be efficiently chemically manipulated while on
the solid phase, and then released back into solution for
analysis. We demonstrate the SPOT process on glycolipid-
derived structures.
The feasibility of the SPOT process was evaluated by
using two simple, model reducing oligosaccharides derived
from glycolipids (Scheme 2): the linear tetrasaccharide lacto-
The chemistry described in Scheme 1 was developed for
SPOT. Structure A shows benzyl hydroxylamine capture
Scheme 2. Structures of LNT, LNFP, and fluorescently tagged products
of LNT. TMR=tetramethylrhodamine.
N-tetraose (LNT) and its fucosylated product lacto-N-fuco-
pentaose (LNFP). These specific structures were selected
because they are commercially available in pure form, as are
the glycosidases required to cleave their terminal sugar
residues for eventual enzyme-assisted sequencing.
Two different solid supports were investigated in the
evaluation of SPOT: PEGA-1900,[8] a polyethylene glycol–
polyacrylamide resin that swells in both organic and aqueous
solvents and has functionalizable amino groups incorporated
in the range of 0.5 mmolgÀ1; and controlled-pore glass (CPG),
which comprises nonswellable, rigid particles having pores of
size > 1400 Š and near 30 mmolgÀ1 incorporated amino
groups. The PEGA beads should be more useful when a
high density of capture groups is required, whereas the CPG
beads would be favored when an immobilized oligosaccharide
should be accessible to a protein probe. For these reasons, the
Scheme 1. The chemistry underlying SPOT.
groups attached through a cleavable ester linkage to a solid
support, optionally through a spacer molecule. An appropri-
ate spacer would be needed if access of a captured sugar
molecule to a protein active site, such as that of a glycosidase
or lectin, was required. Incubation of A with a solution
containing a reducing sugar results in its capture to form an
oxime (B), along with an excess of unreacted capture groups.
The oxime is expected to be a mixture of tautomers including
cyclic forms (not shown).[7] The excess unreacted hydroxyl-
amines in B could now be capped (for example, by reaction
with an amine-reactive compound such as acetic anhydride
(Ac2O)) to give C, under conditions in which neither the
4168
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 4167 –4172

Angewandte
Chemie
amino groups on the PEGA beads were functionalized
directly with the cleavable hydroxylamine capture groups
(no spacer) to give SPOT-PEGA beads (Scheme 3), whereas
the silylpropylaminated CPG was functionalized with a 40-
Figure 1. Capillary electropherogram of crude LNT-TMR. X denotes a
non-carbohydrate side product.
Scheme 3. Structures of solid supports, showing the capture group,
cleavable linker, and optional spacer.
We next examined the suitability of SPOT-PEGA beads
for the insertion of a tag that would be of use in MS. We chose
the p-bromophenyl (BrPh) group, as the unique isotope
pattern (79Br/81Br 1:1) of the bromine atom can confirm that a
product is indeed labeled. Incorporation of a bromine label
also allows the very efficient “editing” of mass spectra as a
result of the mass defect properties of this atom.[9] We
therefore tagged the intermediate D (Scheme 1) with p-
bromophenyl isothiocyanate, washed the beads, cleaved the
linker with an aqueous base, and neutralized the sample as
before. The removal of salts from the product LNT-BrPh
(Figure 2) prior to MS analysis was effected by adsorption on
C-18 resin, washing with water, and elution with an aqueous
solution of MeOH. The mass spectrum of the crude product is
shown in Figure 2 and confirms the quality of the tagged
sample.
atom polyethylene glycol derived spacer prior to the incor-
poration of the capture groups to give SPOT-CPG beads
(Scheme 3). Experimental details are provided in the Sup-
porting Information.
Examination of the optimal conditions for the capture of
oligosaccharides showed that the conditions for > 95%
capture differed between the two bead types. For SPOT-
PEGA beads, dimethyl sulfoxide/acetic acid (DMSO/AcOH,
7:3), 708C, and reaction for 3 h were found to be optimal.
Capping was achieved with acetic anhydride/methanol
(Ac2O/MeOH; 1 h) and reduction was effected with pyri-
dine/borane (Pyr-BH3; 10 min). For SPOT-CPG beads, incu-
bation overnight at 558C in aqueous citrate/phosphate buffer
(pH 5.0) proved ideal for the capture process. The capping
and reduction steps were performed as for the SPOT-PEGA
beads. Extensive washing of the beads was performed
between steps, with the beads placed in syringes with frits
on a vacuum manifold.
In a typical experiment, the capture of LNT (70 mg) on
SPOT-PEGA beads (5 mg, ca. 30 mL) at 708C for 3 h was
followed by capping, reduction, fluorescence tagging with
tetramethylrhodamine-5-isothiocyanate (TMR-NCS), and
cleavage with an aqueous solution of LiOH. The brilliant
red solution obtained was neutralized with aqueous AcOH
and diluted to a final volume of 3 mL. An aliquot of this
sample was further diluted 100 to 1000-fold with water, and a
22-nL portion was injected for micellar capillary electro-
phoresis (CE) with conventional (non-laser) fluorescence
detection at 570 nm. A portion of about 10À8 of the sample
was used for the analysis.
The electropherogram of the SPOT-labeled product is
shown in Figure 1. A highly efficient chemical sequence is
indicated, with the LNT-TMR (Scheme 2) eluting near 16 min
(the structure was verified by MS and was independently
confirmed by solution synthesis), and the only significant
impurity (marked X, Figure 1) being a peak at 23 min. This
latter impurity was isolated by HPLC, was found by MS to
have a m/z of 624, and has the structure shown for LNT-TMR
but with the sugar replaced by a methyl group. This
compound arises from the scavenging of trace contaminating
formaldehyde from the air and/or the solvents used. It does
not interfere in the present analysis, but efforts are under way
to minimize its formation.
Figure 2. Structure and mass spectrum of LNT-BrPh, with the 79Br and
81Br isotopes highlighted in the inset.
We then set up a “dummy run” to evaluate a potential
dual application of SPOT in both fluorescence tagging and
CE analysis as just described, and in a differential glyco-
mics[10] setting where the concentration of the same com-
pound in two different samples would be estimated by MS
using stable, isotope-encoded tagging species (Figure 3). For
this experiment, we chemically synthesized TMR-NCS with
two trideuteromethyl groups ([D6]TMR-NCS). Separate
samples containing LNT at concentrations of 1.0 and 0.5 mm
Angew. Chem. Int. Ed. 2006, 45, 4167 –4172
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 4169

Communications
Figure 3. Flow chart for estimating the relative concentrations of LNT
in two different samples, by both CE and MS analysis.
were run through the SPOT process (Scheme 1), each using
5 mg of SPOT-PEGA beads. The more concentrated sample
was labeled with TMR-NCS and the other by using [D6]TMR-
NCS. CE analysis of the two cleaved and neutralized samples
revealed identical retention times for LNT-TMR and LNT-
[D6]TMR (see Scheme 2 for structures). As expected, the
integral of the peak for LNT-[D6]TMR was half that of the
unlabeled compound (Figure 4).
Figure 5. Electrospray mass spectrum of a mixture of equal volumes of
the LNT solutions labeled as in Figure 3. The inset highlights the
isotope patterns for nonlabeled and [D6]-labeled LNT-TMR.
Figure 6. Capillary electropherogram of a mixture of LNT and LNFP
SPOT-labeled with TMR.
1H NMR spectrum of the mixture used for capture and
tagging revealed that the compounds were in fact present in
the ratio 100:79 (data not shown). The conclusion is therefore
that the SPOT process does not significantly discriminate
between the oligosaccharides of different size used here in the
capture, subsequent chemistry, and analysis steps. The LNFP
sample was most likely contaminated with salt and/or
moisture, or an error in weighing was made.
Figure 4. Capillary electropherograms of the LNT samples labeled as in
Figure 3.
Finally, we evaluated the entire SPOT process on a real-
life glycolipid, the brain-derived ganglioside GM1 (Scheme 4).
The ganglio series of glycolipids are known to be good
substrates for the commercial recombinant endoglycocera-
midase II (Takara Bio), and the ceramide was indeed
completely cleaved in an overnight incubation. The reaction
mixture was lyophilized, the residue was redissolved in
DMSO/AcOH, SPOT-PEGA beads were added, and the
entire sequence of reactions developed on LNTwas repeated.
The CE and MS of the product GM1-TMR are shown in
Figure 7. The quality of the product validates the SPOT
process.
We have demonstrated that reducing oligosaccharides can
be captured on both swellable resins like PEGA, which are
water and organic solvent compatible, and on rigid surfaces
such as glass (CPG). Herein, immobilized hydroxylamine
groups have been used as the capture reagents to produce
immobilized oximes. Following a sequence of capping,
reduction, and tagging with aryl isothiocyanates, a single tag
is installed at the reducing end. Excess tagging (and other)
Identical volumes of each sample were then mixed
together, following the flow diagram of Figure 3, and the
resulting solution was desalted on C-18 resin prior to analysis
by electrospray MS. CE analysis of the mixture (Figure 4, top
trace) confirmed their expected co-elution. The mass spec-
trum obtained in the negative-ion mode is shown in Figure 5,
and unambiguously reports the relative abundance of LNT in
the two investigated samples.
To show that the composition of a mixture of oligosac-
charides could be quantitated by the SPOT method, an
equimolar mixture of LNTand LNFP was prepared by careful
weighing of the commercial compounds. This time, SPOT-
CPG beads were used. Surprisingly, the electropherogram of
the product obtained after SPOT labeling with TMR-NCS
showed that the compounds were present in the ratio LNT/
LNFP 100:74 (Figure 6). We initially thought that either the
labeling of the smaller oligosaccharide was more efficient, or
that the excitation/emission properties of the labeled com-
pounds differed. However, the fully relaxed 600-MHz
4170
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 4167 –4172

Angewandte
Chemie
ide can therefore be rapidly derivatized with several different
tags, and analyzed by a variety of different techniques.
The results presented are clearly those of a model study,
designed only to test the feasibility of the chemistry used for
capture, capping, reduction, tagging, and release. The central
consideration for the chemistries developed is that, in real-life
situations, very little oligosaccharide may be available for
analysis. This means that the concentration of oligosaccharide
in solution might be very low. Thus, it is essential that the
concentration of immobilized capture groups is high for a
bimolecular reaction to occur, and these capture groups must
therefore be present in a large excess. In all of the experi-
ments presented, the hydroxylamine capture groups were in
fact present in at least a tenfold molar excess of the target
oligosaccharide. As a result, it was necessary to be able to
inactivate this excess prior to the tagging step, which would,
for the same reasons, require a large excess of tagging agent.
The sequence of chemical reactions developed proved to be
highly efficient at achieving these objectives.
Scheme 4. Sequence of reactions for SPOT labeling of ganglioside
GM1.
As presented herein, the capture and tagging of oligosac-
charides using the SPOT process is highly efficient (> 90%)
with SPOT-PEGA beads down to the range of 1 nmol
oligosaccharide/mg beads. The lower level of capture group
incorporation in SPOT-CPG requires a tenfold higher con-
centration of oligosaccharide to approach the same efficiency.
The release of tagged oligosaccharide from the beads was
judged to be complete under the saponification conditions
used, as all of the visible red color was removed from CPG-
SPOT beads.
There are certainly many challenges that remain before
SPOT can be applied to real-life situations, especially to crude
samples such as cell or tissue extracts. However, these are the
same challenges that exist today for solution labeling
techniques: complex mixtures of glycans will be present and
their release from glycoproteins or glycolipids, by either
enzymatic or chemical means, must be quantitative, as should
their labeling. Furthermore, the labeling of N-linked oligo-
saccharides at the reducing end GlcNAc residue is more
difficult than at the reducing Glc residue used herein. An
additional challenge will be learning to eliminate trace
compounds containing aldehydes and ketones, whether
present in the biological samples or in the solvents and
reagents used. These same compounds are present in the
solution labeling techniques used today, but they are either
removed by “sample processing” prior to labeling or react and
become labeled. These undesired reaction products, which
include excess labeling agent, are then removed in subsequent
sample processing steps with attendant potential losses of
oligosaccharide components. It is hoped that such sample
processing steps, with their potential selective sample losses,
can be avoided with the further development of the SPOT
technique.
Figure 7. Characterization of the SPOT-labeled GM1-TMR obtained as
described in Scheme 4. Top: capillary electropherogram. Bottom:
electrospray mass spectrum.
reagent is removed by simple washing. As a cleavable linker is
installed between the capture group and the solid support, the
tagged oligosaccharide can be released into solution for
analysis by a variety of techniques. The types of tags used
were selected to permit detection by fluorimetry and/or MS.
The two key advantages of SPOT are: 1) the simplicity of
all the manipulations—which involve only pipetting and
washing, and require no prior experience in the handling of
oligosaccharides—means that valuable oligosaccharides
cannot be lost once they are reduced after immobilization;
and 2) a single protocol is used for labeling and release,
irrespective of the tag used. The same captured oligosacchar-
Received: February 17, 2006
Published online: May 19, 2006
Keywords: fluorescent probes · glycolipids · glycosylation ·
.
isotopic labeling · oligosaccharides
Angew. Chem. Int. Ed. 2006, 45, 4167 –4172
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
4171

Communications
[1]R. Apweiler, H. Hermjakob, N. Sharon, Biochim.Biophys.Acta
1999, 1473, 4 – 8.
[2]T. Kolter, K. Sandhoff, Angew.Chem. 1999, 111, 1632 – 1670;
Angew.Chem.Int.Ed. 1999, 38, 1532 – 1568.
[3]For recent reviews on oligosaccharide analysis, see: a) P. M.
Rudd, R. A. Dwek, Curr.Opin.Biotechnol. 1997, 8, 488 – 497;
b) A. Dell, H. R. Morris, Science 2001, 291, 2351 – 2356; c) Z.
Shriver, S. Raguram, R. Sasisekharan, Nat.Rev.Drug Discovery
2004, 3, 863 – 873; d) M. V. Novotny, Y. Mechref, J.Sep.Sci. 2005,
28, 1956 – 1968; e) W. Morelle, J. C. Michalski, Curr.Anal.Chem.
2005, 1, 29 – 57.
[4]For recent reviews on oligosaccharide tagging in solution, see:
a) Y. Mechrev, M. Novotny, Chem.Rev. 2002, 102, 321 – 369;
b) N. V. Shilova, N. V. Bovin, Russ.J.Bioorg.Chem.
2003, 29,
309 – 324.
[5]S. Nishimura, K. Niikura, M. Kurogochi, T. Matsushita, M.
Fumoto, H. Hinou, R. Kamitani, H. Nakagawa, K. Deguchi, N.
Miura, K. Monde, H. Kondo, Angew.Chem. 2005, 117, 93 – 98;
Angew.Chem.Int.Ed. 2005, 44, 91 – 96.
[6]M. Lee, I. Shin, Org.Lett. 2005, 7, 4269 – 4272.
[7]a) W. von Deyn, W. S. York, P. Albersheim, A. G. Darvill,
Carbohydr.Res. 1990, 201, 135 – 144; b) M. Pauly, W. S. York,
R. Guillen, P. Albersheim, A. G. Darvill, Carbohydr.Res. 1996,
282, 1 – 12; c) S. L. Ramsay, C. Freeman, P. B. Grace, J. W.
Redmond, J. K. MacLeod, Carbohydr.Res. 2001, 333, 59 – 71.
[8]R. M. Ferreras, J. M. Delaisse, N. T. Foged, M. P. Meldal, J.Pept.
Sci. 1998, 4, 195 – 210.
[9]M. P. Hall, S. Ashrafi, I. Obegi, R. Petesch, J. N. Peterson, L. V.
Schneider, J.Mass Spectrom. 2003, 38, 809 – 816.
[10]For recent reviews on differential proteomics, see: a) L. Mon-
teoliva, J. P. Aalbar, Briefings Funct.Genomics Proteomics 2004,
3, 220 – 239; b) S. Julka, F. E. Regnier, Briefings Funct.Genomics
Proteomics 2005, 4, 158 – 177; c) L. V. Schneider, M. P. Hall,
Drug Discovery Today 2005, 10, 353 – 363. For examples of
stable-isotope tagging in differential glycomics, see: d) H.
Zhang, X.-J. Li, D. B. Martin, R. Aebersold, Nat.Biotechnol.
2003, 21, 660 – 666; e) J. Yuan, N. Hashii, N. Kawasaki, S. Itoh, T.
Kawanishi, T. Hayakawa, J.Chromatogr.A 2005, 1067, 145 – 152.
4172
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 4167 –4172