Direct N-glycan profiling in the presence of tryptic peptides on MALDI-TOF by controlled ion enhancement and suppression upon glycan-selective derivatization

Article abstract of DOI:10.1021/ac0492766

Even though the formidably laborious and time-consuming nature of oligosaccharide analysis limits certain attempts to analyze the glycosylation profile, the significant elucidation of carbohydrate modifications is largely dependent on it. Aiming to substa

Full text of DOI:10.1021/ac0492766

Anal. Chem. 2004, 76, 6989-6997
Direct N-Glycan Profiling in the Presence of Tryptic
Peptides on MALDI-TOF by Controlled Ion
Enhancement and Suppression upon
Glycan-Selective Derivatization
Yasuro Shinohara,* Jun-ichi Furukawa, Kenichi Niikura, Nobuaki Miura, and Shin-Ichiro Nishimura*
Division of Biological Sciences, Graduate School of Science, Frontier Research Center for Post-Genomic Science and
Technology, Hokkaido University, Sapporo 001-0021, Japan
reached the same degree of sophistication as peptides. This is
because analyzing carbohydrates is much more difficult due to
the problems imposed by structural complexities and their
inherent difficulty in detecting chromatography or mass spec-
trometry in their native state. Due to its simplicity, speed of
analysis and comparatively increased sensitivity, matrix-assisted
laser desorption/ionization (MALDI) mass spectrometry has
proven to have several advantages over other analytical techniques
when analyzing carbohydrates.5-7 However, the relatively low
response to oligosaccharides (compared, for instance, peptides)
often requires derivatization of oligosaccharides where highly
sensitive detection is required.
Although a number of derivatization procedures have been
evaluated and were proven to be efficient in improving detection
sensitivity,8-10 there are still vast challenges for glycosylation
analysis on both naturally occurring and recombinant proteins.
Substantial improvement in sample preparation is eagerly expected
in particular. Analyzing oligosaccharides is a formidably laborious
and time-consuming task. For the sensitive detection of oligosac-
charides, they need to be liberated from any protein(s) or peptide-
(s), extensively purified from other molecular species, derivatized
with a labeling reagent, and must be purified often from the excess
reagent. The purification step(s) is/are typically achieved by
chromatography. Indeed, many sample preparation steps are
hardly integratable as an automated format, and so they raise a
substantial risk to lose trace components in the sample. These
inconveniences are further pronounced for the analysis of sialy-
lated oligosaccharides, which have been implicated in a number
of important biological processes. Sialylated oligosaccharides
analyzed in MALDI-TOF often cause metastable fragmentation,
Even though the formidably laborious and time-consum-
ing nature of oligosaccharide analysis limits certain at-
tempts to analyze the glycosylation profile, the significant
elucidation of carbohydrate modifications is largely de-
pendent on it. Aiming to substantially improve the sample
preparation procedure, a novel protocol allowing glycan-
specific detection in the presence of other species, such
as tryptic peptides, on MALDI-TOF was proposed and
then evaluated. The new protocol is based on the concept
that the desorption/ionization efficiency of glycans could
be selectively and substantially enhanced while drastically
suppressing the other ion species upon glycan-selective
derivatization. A series of known and novel labeling
reagents, all of which carry hydrazide functionality to allow
glycan-specific derivatization, were prepared and evalu-
ated in terms of their abilities to enhance the detection
sensitivity of glycans, suppress ions of other contaminants
(e.g., peptides), and detect acidic oligosaccharides. Sev-
eral novel reagents that possess hydrophobic residue(s)
together with quaternary ammonium/pyridinium or guani-
dino functionalities significantly enhanced the detection
sensitivity of oligosaccharides. When enzymatically de-
glycosylated tryptic ovalbumin digest was directly deriva-
tized by these reagents and subjected to MALDI-TOF
analysis without any prior purification, we observed that
a single type of analyte ion (labeled glycan) could suppress
a large majority of peptide ions while allowing a low-
femtomole level detection of oligosaccharides. The efficacy
of this approach was further evaluated using several other
model glycoproteins, including r1-acid glycoprotein that
contains a variety of sialylated oligosaccharides.
(3) Groulet, A.; Dorvillius, M.; Pelegrin, A.; Barbet, J.; Baty, D. Int. J. Cancer
2002, 100, 367-374.
Glycoprotein oligosaccharides play a vital role in biological
processes, such as stability, protein conformation, intra- and
intercell signaling, and a binding affinity to and a specificity for
other biomolecules.1,2 It is also becoming evident that glycosylation
can affect the safety and efficiency of the therapeutic glycopro-
teins.3,4 However, the analysis of constituent glycans has not
(4) Elliott, S.; Lorenzini, T.; Asher, S.; Aoki, K.; Brankow, D.; Buck, L.; Busse,
L.; Chang, D.; Fuller, J.; Grant, J.; Hernday, N.; Hokum, M.; Hu, S.; Knudten,
A.; Levin, N.; Komorowski, R.; Martin, F.; Navarro, R.; Osslund, T.; Rogers,
G.; Rogers, N.; Trail, G.; Egrie, J. Nat. Biotechnol. 2003, 21, 414-421.
(5) Harvey, D. J. Mass Spectrom. Rev. 1999, 18, 349-450.
(6) Zaia, J. Mass Spectrom. Rev. 2004, 23, 161-227.
(7) Kurogochi, M.; Matsushita, K.; Nishimura, S.-I. Angew. Chem., Int. Ed. 2004,
43, 4071-4075.
(8) Okamoto, M.; Takahashi, K.; Doi, T.; Takimoto, Y. Anal. Chem. 1997, 69,
2919-2926.
* To whom correspondence should be addressed. Phone and Fax:
+81 11 706 9038. E-mails: yshinohara@glyco.sci.hokudai.ac.jp and shin@
glyco.sci.hokudai.ac.jp.
(1) Helenius, A.; Aebi, M. Science 2001, 291, 2364-2369.
(2) Lowe, J. B. Cell 2001, 104, 809-812.
(9) Harvey, D. J. J. Am. Soc. Mass Spectrom. 2000, 11, 900-915.
(10) Mechref, Y.; Novotny, M. V. Chem. Rev. 2002, 102, 321-369.
10.1021/ac0492766 CCC: $27.50 © 2004 American Chemical Society
Published on Web 10/30/2004
Analytical Chemistry, Vol. 76, No. 23, December 1, 2004 6989

and those analyzed in a positive ion mode have yielded a mixture
of cation adducts ions, giving rise to a complex mass profile.
Although permethylation10 and methyl esterification of sialic acid12
have been demonstrated to be quite useful in stabilizing these
groups and rendering them chemically equivalent to neutral
oligosaccharides, they have also been proven to lower the
throughput.
EXPERIMENTAL SECTION
Materials and Reagents. A complex-type N-linked oligosac-
charide (Gal(â1-4)GlcNAc(â1-2)Man(R1-6)(Gal(â1-4)GlcNAc-
(â1-2)Man(R1-3))Man(â1-4)GlcNAc(â1-4)GlcNAc, NA2) was
purchased from Prozyme (San Leandro, CA). Girard’s reagents
T and P were acquired from Tokyo Kasei Kogyo (Tokyo,
Japan). (3-(2-(Methoxycarbonyl)ethyl)-2-indolyl)trimethylammo-
nium iodide and 3-Et-1-(2-(2-methoxycarbonylmethyl-1H-indol-3-
yl)-ethyl)-pyridinium perchlorate were ordered from Sigma-Aldrich
Aiming at partially filling these gaps, our attention was directed
toward utilizing the phenomenon known as “ion suppression” in
a mass spectrometric analysis.13,14 Ion suppression is a competition
effect in ionization when the sample contains a high concentration
of salts, or (excess of) other analyte(s) that can be ionized in the
operating condition. For instance, peak intensities of peptide
mixtures, such as those obtained by tryptic digestion, differ
significantly. This is considered attributable not only to the
intrinsic properties of peptides, but also to ion suppression.15,16
The magnitude of the observed suppression strongly depends on
both nature and the amount of suppressing peptide. Due to this
phenomenon, the direct analysis of unseparated mixtures is a
particularly difficult exercise, since the suppression could have
been caused by other analytes in the mixture, as well as by
contaminants introduced during isolation and workup procedures.
Thus, the negative aspects of ion suppression have quickly become
a major concern. Consequently, many research interests have
been directed to reducing this unfavorable phenomenon to enable
the detection of as many peptide ions as possible, either by
derivatization17-19 or by thorough purification.20,21 Our interest is
to see whether specific ion species could be almost completely
embossed in a complex mixture if the ionization efficiency of the
ions of interest could be selectively and substantially enhanced.
In this article, a series of known and novel labeling reagents,
all of which carry a hydrazide functionality to allow glycan-specific
derivatization in a complex mixture, were evaluated in terms of
their abilities to enhance the ionization efficiency of labeled
glycans, suppress ions of other contaminants (e.g. peptides), and
detect sialylated oligosaccharides. Using glycoprotein tryptic
digests followed by enzymatic N-glycan liberation as a model case,
we observed that a single type of analyte ion (labeled glycan) could
suppress a large majority of peptide ions and other contaminants.
In other words, the N-glycan profiling in an unpurified tryptic
digest could be achieved without any purification steps. The
detection of sialylated oligosaccharides in a complex mixture will
also be addressed.
(St. Louis, MO). NR-benzoyl-
was purchased from Sigma Chemical (St. Louis, MO),
L
-arginine methyl ester bicarbonate
-arginine
L
methyl ester dihydrochloride was acquired from Fluka (Milwau-
kee, WI), and (Z)-Phe-Arg methyl ester was acquired from Chem-
Impex International (Wood Dale, IL). Ovalbumin (grade VI),
asialofetuin (type II) from fetal calf serum, ribonuclease B from
bovine pancreas, and human R1-acid glycoprotein were purchased
from Sigma Chemicals. 2,5-Dihydroxybenzoic acid (DHB) and
R-cyano-4-hydroxycinnamic acid (HCCA) were obtained from
Bruker Daltonics (Billerica, MA). Recombinant PNGase F (E.C.
3.2.2.18) was procured from Roche Diagnostics (Penzberg,
Germany), and sequencing grade-modified trypsin (EC 3.4.21.4)
was purchased from Promega Corp (Madison, WI).
Synthesis of Reagents. (3-(2-(Hydrazinocarbonyl)ethyl)-2-
indolyl)trimethylammonium (HEITA). (3-(2-(Methoxycarbonyl)-
ethyl)-2-indolyl)trimethylammonium iodide (21.7 mg, 55.9 µmol)
was dissolved in methanol (2.8 mL), and hydrazine monohydrate
(0.5 mL, 10.3 mmol) was added. The reaction mixture was stirred
for 12 h at room temperature. The titled compound was isolated
by reversed-phase chromatography with 50 mM ammonium
1
bicarbonate/acetonitrile (80:20, vol) at a yield of ∼72%; H NMR
(500 MHz, DMSO-d6) δ 7.57-6.97 (m, 4H, Ph), 4.13 (m, 2H,
CONHNH2); positive ion mode MALDI-TOF m/z 261.3 (M+).
3-Et-1-(2-(2-hydrazinocarbonylmethyl-1H-indol-3-yl)-ethyl)-pyri-
dinium (EHIEP). 3-Et-1-(2-(2-methoxycarbonylmethyl-1H-indol-3-
yl)-ethyl)-pyridinium perchlorate (11.1 mg, 26.2 µmol) was dis-
solved in methanol (0.99 mL), and hydrazine monohydrate (0.28
mL, 5.9 mmol) was added. The reaction mixture was stirred for
12 h at room temperature. The titled compound was isolated by
reversed-phase chromatography with 50 mM ammonium bicar-
1
bonate/acetonitrile (80:20, vol) at a yield of ∼74%; H NMR (500
MHz, DMSO-d6) δ 4.76 (t, 2H, J ) 6.5 Hz, CH2N), 4.31 (bs, 2H,
CONHNH2); positive ion mode MALDI-TOF m/z 323.0 (M+).
L
-Arginine Hydrazide (Rh). L-Arginine methyl ester dihydro-
chloride (RO-Me) was prepared according to the previously
described procedure with minor modifications.22 RO-Me (5.0 mg,
19.1 µmol) was dissolved in methanol (0.96 mL), and hydrazine
monohydrate (0.35 mL, 7.3 mmol) was added. The reaction
mixture was stirred for 6 h at room temperature. The titled
compound was isolated using reversed-phase chromatography
with 50 mM ammonium bicarbonate/acetonitrile (86:14, vol) at a
yield of ∼76%; positive ion mode MALDI-TOF m/z 189.1 (M +
H)+.
(11) Dell, A.; Morris, H. R. Science 2001, 291, 2351-2356.
(12) Powell, A. K.; Harvey, D. J. Rapid Commun. Mass Spectrom. 1996, 10, 1027-
1032.
(13) Knochenmuss, R.; Stortelder, A.; Breuker, K.; Zenobi, R. J. Mass Spectrom.
2000, 35, 1237-1245.
(14) Annesley, T. M. Clin. Chem. 2003, 49, 1041-1044.
(15) Burkitt, W. I.; Giannakopulos, A. E.; Sideridou, F.; Bashir, S.; Derrick, P. J.
Aust. J. Chem. 2003, 56, 369-377.
(16) Kratzer, R.; Eckerskorn, C.; Karas, M.; Lottspeich, F. Electrophoresis 1998,
19, 1910-1919.
(17) Brancia, F. L.; Oliver, S. G.; Gaskell, S. J. Rapid Commun. Mass Spectrom.
NR-Benzoyl- -arginine Hydrazide (BRh). NR-benzoyl-
L L-arginine
2000, 14, 2070-2073.
methyl ester bicarbonate (4.5 mg, 12.7 µmol) was dissolved in
methanol (0.63 mL), and hydrazine monohydrate (0.23 mL, 4.7
mmol) was added. The reaction mixture was stirred for 6 h at
(18) Beardsley, R. L.; Reilly, J. P. Anal. Chem. 2002, 74, 1884-1890.
(19) Keough, T.; Youngquist, R. S.; Lacey, M. P. Anal. Chem. 2003, 75, 156A-
165A.
(20) Preisler, J.; Hu, P.; Rejtar, T.; Karger, B. L. Anal. Chem. 2000, 72, 4785-
4795.
(21) Chen, V. C.; Cheng, K.; Ens, W.; Standing, K. G.; Nagy, J. I.; Perreault, H.
Anal. Chem. 2004, 76, 1189-1196.
(22) Wang, H. K.; Liu, S. Y.; Hwang, K. M.; Taylor, G.; Lee, K. H. Bioorg. Med.
Chem. 1994, 2, 1397-1402.
6990 Analytical Chemistry, Vol. 76, No. 23, December 1, 2004

room temperature. The titled compound was isolated by reversed-
phase chromatography with 50 mM ammonium bicarbonate/
4.29 (m, 2H, CONHNH2); positive ion mode MALDI-TOF m/z
509.2 (M + H)+.
NR-(Benzyloxycarbonyl)tyrosinylarginine Hydrazide (ZYRh). To
a solution of NR-(benzyloxycarbonyl)tyrosine (788 mg, 2.5 mmol)
in dry THF (6 mL), we added N-methylmorpholine (330 µL, 3.0
mmol) and isobutyl chloroformate (393 µL, 3.0 mmol) at -20 °C.
The mixture was heated to room temperature and stirred for 15
min and was followed by the addition of a solution of the arginine
methyl ester dihydrochloride (784 mg, 3.0 mmol) and NaHCO3
(252 mg, 3.0 mmol) in H2O (3 mL) at 0 °C. The reaction mixture
was stirred for 1 h. Solvents were removed under reduced
pressure. The residue was chromatographed on a column of silica
gel using 5:1 CHCl3/MeOH to give (Z)-Tyr-Arg methyl ester (0.8
1
acetonitrile (80:20, vol) at a yield of ∼87%; H NMR (500 MHz,
DMSO-d6) δ 8.31-7.37 (m, 5H, Ph), δ 4.42 (m, 1H, R-CH), 4.24
(bs, 2H, CONHNH2); positive ion mode MALDI-TOF m/z 293.2
(M + H)+.
NR-(Benzyloxycarbonyl)arginine Hydrazide (ZRh). NaHCO3 (168
mg, 2 mmol) and Z chloride (313 µL, 1 mmol) was added to a
solution of R-OMe (261 mg, 1.0 mmol) in H2O (5 mL). The
mixture was stirred for 12 h at room temperature. The reaction
mixture was extracted with CHCl3, and the organic layer was
washed with saturated aqueous NaHCO3 and brine. The extraction
was concentrated, and the residue was purified by silica gel
chromatography, eluting it with CHCl3/MeOH (5:1) to give 0.3 g
1
g, 67%); H NMR (500 MHz, CD3OD) δ 8.53-7.09 (m, 9H, Ph),
1
4.95 (d, 1H, J ) 12.9 Hz, 1/2 × CH2Ph), 4.92 (d, 1H, J ) 12.9 Hz,
1/2 × CH2Ph), 4.23 (m, 2H, Tyr R- and Arg R-CH), 3.62 (s, 3H,
COOCH3). (Z)-Tyr-Arg methyl ester hydrochloride (5.2 mg,
10.7µmol) was dissolved in methanol (0.54 mL), and hydrazine
monohydrate (0.2 mL, 4.0 mmol) was added. The reaction mixture
was stirred for 6 h at room temperature. The titled compound
was isolated by reversed-phase chromatography with 50 mM
ammonium bicarbonate/acetonitrile (66:34, vol) at a yield of ∼66%;
1H NMR (500 MHz, DMSO-d6) δ 7.35-6.63 (m, 9H, Ph), 4.97 (d,
1H, J ) 13.4 Hz, 1/2 × CH2Ph), 4.92 (d, 1H, J ) 12.9 Hz, 1/2 ×
CH2Ph), 4.37 (m, 2H, CONHNH2), 4.20 (m, 2H, Tyr R- and Arg
R-CH); positive ion mode MALDI-TOF m/z 486.2 (M + H)+.
Digestion of Glycoproteins with Trypsin. Digestion of
glycoproteins with trypsin was performed according to the
procedure previously described.23 Briefly, glycoproteins were
solubilized with 0.1% (w/v) RapiGest SF (Waters, Milford, MA)
solution buffered with 50 mM ammonium bicarbonate (pH of 7.8)
at a concentration of 15 µM each. They were digested with 5 µg
of modified trypsin at 37 °C for 1 h. Heating the reaction mixture
at 100 °C for 3 min terminated the reaction.
Deglycosylation of Tryptic Digests by PNGase F. The
tryptic digest of each glycoprotein was incubated by directly
adding PNGase F (0.5 U) to the tryptic glycoprotein digest at 37
°C overnight. The reaction mixture was kept at -20 °C until
further analysis.
Derivatization of Standard Oligosaccharides with Labeling
Reagents. The derivatization reaction was modified from the
procedure for hydrazone formation, as previously described.24,25
For standard derivatization, 10 µL of each labeling reagent (200
µM) in 80% methanol and 1 µL of 10 µM oligosaccharide (NA2)
were added to 10 µL of 50 mM ammonium bicarbonate. The
mixture was then incubated at 90 °C for 1 h. After cooling the
reaction tube on ice, the reaction mixture was accordingly mixed
with DHB (10 g/L in 30% acetonitrile), and 0.5 µL of the resulting
mixture was subjected to MALDI-TOF mass analysis without any
purification.
(31% yield) of (Z)-arginine methyl ester ((Z)R-OMe); H NMR
(500 MHz, CDCl3) δ 7.27 (m, 5H, Ph), 5.08 (d, 1H, J ) 12.1 Hz,
1/2 × CH2Ph), 5.03 (d, 1H, J ) 12.1 Hz, 1/2 × CH2Ph), 4.23 (bs,
1H, R-CH), 3.67 (s, 3H, COOCH3). (Z)R-OMe (2.9 mg, 9.0µmol)
was dissolved in methanol (0.45 mL), and hydrazine monohydrate
(0.17 mL, 3.4 mmol) was added. The reaction mixture was stirred
for 6 h at room temperature. The titled compound was isolated
by reversed-phase chromatography with 50 mM ammonium
1
bicarbonate/acetonitrile (66:34, vol) at a yield of ∼68%; H NMR
(500 MHz, DMSO-d6) δ 7.37 (m, 5H, Ph), 5.01 (s, 2H, CH2Ph),
4.22 (bs, 2H, CONHNH2), 4.09 (bs, 1H, R-CH); positive ion mode
MALDI-TOF m/z 323.2 (M + H)+.
NR-(Benzyloxycarbonylphenylalanylarginine Hydrazide (ZFRh).
(Z)-Phe-Arg methyl ester, hydrochloride (2.7 mg, 5.3µmol) was
dissolved in methanol (0.27 mL), and hydrazine monohydrate (0.1
mL, 2.1 mmol) was added. The reaction mixture was stirred for
6 h at room temperature. The titled compound was isolated using
reversed-phase chromatography with 50 mM ammonium bicar-
1
bonate/acetonitrile (66:34, vol) at a yield of ∼77%; H NMR (500
MHz, DMSO-d6) δ 7.37 (m, 10H, Ph), 4.95 (d, 1H, J ) 13.0 Hz,
1/2 × CH2Ph), 4.91 (d, 1H, J ) 12.6 Hz, 1/2 × CH2Ph), 4.28 (m,
2H, Phe R- and Arg R-CH), 4.25 (bs, 2H, CONHNH2); positive
ion mode MALDI-TOF m/z 470.1 (M + H)+.
NR-(Benzyloxycarbonyl)tryptophanylarginine Hydrazide (ZWRh).
To a solution of NR-(benzyloxycarbonyl)tryptophan (846 mg, 2.5
mmol) in dry THF (6 mL), we added N-methylmorpholine (330
µL, 3.0 mmol) and isobutyl chloroformate (393µL, 3.0 mmol) at
-20 °C. The mixture was heated to room temperature and stirred
for 15 min, followed by the addition of a solution of the arginine
methyl ester dihydrochloride (784 mg, 3.0 mmol) and NaHCO3
(252 mg, 3.0 mmol) in H2O (3 mL) at 0 °C. The reaction mixture
was stirred for 1 h. Solvents were removed under reduced
pressure. The residue was chromatographed on a column of silica
gel using 5:1 CHCl3/MeOH to give (Z)-Trp-Arg methyl ester (1.1
1
g, 87%); H NMR (500 MHz, CD3OD) δ 8.55-7.18 (m, 10H, Ph
and indole), 4.93 (s, 2H, CH2Ph), 4.33 (m, 2H, Trp R- and Arg
R-CH), 3.63 (s, 3H, COOCH3). (Z)-Trp-Arg methyl ester (5.4 mg,
10.6µmol) was dissolved in methanol (0.53 mL), and hydrazine
monohydrate (0.2 mL, 4.0 mmol) was added. The reaction mixture
was stirred for 6 h at room temperature. The titled compound
was isolated using reversed-phase chromatography with 50 mM
ammonium bicarbonate/acetonitrile (66:34, vol) at a yield of ∼76%;
1H NMR (500 MHz, DMSO-d6) δ 8.13-6.95 (m, 10H, Ph and
indole), 4.94 (s, 2H, CH2Ph), 4.29 (m, 2H, Trp R- and Arg R-CH),
Derivatization of Deglycosylated Tryptic Digests with
Labeling Reagents. Typically, 1 µL of deglycosylated tryptic
digest of ovalbumin, asialofetuin, ribonuclease B, or R1-acid
(23) Yu, Y. Q.; Gilar, M.; Lee, P. J.; Bouvier, E. S.; Gebler, J. C. Anal. Chem.
2003, 75, 6023-6028.
(24) Shinohara, Y.; Sota, H.; Kim, F.; Shimizu, M.; Gotoh, M.; Tosu, M.;
Hasegawa, Y. J. Biochem. (Tokyo) 1995, 117, 1076-1082.
(25) Shinohara, Y.; Sota, H.; Gotoh, M.; Hasebe, M.; Tosu, M.; Nakao, J.;
Hasegawa, Y.; Shiga, M. Anal. Chem. 1996, 68, 2573-2579.
Analytical Chemistry, Vol. 76, No. 23, December 1, 2004 6991

Figure 1. Structures of the reagents used in this study: (a) quaternary ammonium/pyridinium series and (b) guanidino series.
glycoprotein (15 µM) was mixed with 10 µL of each labeling
reagent (200 µM) in 80% methanol and 9 µL of 50 mM ammonium
bicarbonate. The mixture was incubated at 90 °C for 1 h. After
cooling the reaction tube on ice, 1 µL of the reaction mixture was
mixed with 9 µL of DHB (10 g/L in 30% acetonitrile), and 0.5 µL
of the resulting mixture was subjected to MALDI-TOF mass
analysis without any purification.
MALDI-TOF Mass Spectrometry. Mass measurements were
carried out using an Ultraflex TOF mass spectrometer equipped
with a pulsed ion extraction system (Bruker-Daltonik GmbH,
Bremen, Germany). Ions were generated by a pulsed 337-nm
nitrogen laser. The laser power was adjusted to a level slightly
above the threshold for the formation of ions. All of the spectra
were obtained using a reflectron mode with an acceleration voltage
of 25 kV, a reflector voltage of 26.3 kV, and a pulsed ion extraction
of 160 ns in the positive ion mode. These were the results of the
signal averaging of 100 laser shots. Typically, three spectra of one
spot measured at different positions were averaged. Good spec-
trum quality of each derivatized glycan was typically obtained from
the peripheral region of the DHB crystal. Acidic oligosaccharides
were also analyzed at 20 kV without the reflectron in the positive
ion mode.
achieve selective labeling of the oligosaccharides in the presence
of other species, such as tryptic peptides, we employed the
hydrazide functionality as the carbohydrate reactive center. The
formation of hydrazone has advantages for the microscale analysis
of oligosaccharides, since the substituted hydrazide couples to
the oligosaccharide with high efficiency by simply heating the
oligosaccharide with some fold of reagent molar excess.24,25
Moreover, it does not necessarily require the purification step prior
to MS analysis.26 Since they are proven to be efficient in improving
the sensitivity, we employed quaternary ammonium/pyridinium
centers as ionization promoting functionalities (e.g., Girard’s
reagent T (GT),26 trimethyl-(p-aminophenyl) ammonium (TMA-
PA)27). In addition to two commercially available compounds, GT
and Girard’s reagent P (GP), two novel reagents, (3-(2-(hydrazi-
nocarbonyl) ethyl)-2-indolyl)trimethylammonium (HEITA) and
3-Et-1-(2-(2-hydrazinocarbonylmethyl-1H-indol-3-yl)-ethyl)-pyridin-
ium (EHIEP), were synthesized for this purpose (see the
Experimental Section). The other class of ionization promoting
functionality used in this study was the guanidino center, since it
was also ionized with high efficiency.28 Six reagents that possess
both hydrazide and guanidino functionalities, Rh, BRh, ZRh, ZFRh,
ZWRh, and ZYRh, were synthesized for this purpose (see the
Experimental Section).
The detection limits for the derivatized oligosaccharides were
estimated by depositing each carbohydrate derivative on the target
in decreasing amounts by serial dilution and analyzing them by
MALDI-TOF MS, as described above.
(26) Naven, T. J. P.; Harvey, D. J. Rapid Commun. Mass Spectrom. 1996, 10,
829-834.
(27) Okamoto, M.; Takahashi, K.; Doi, T. Rapid Commun. Mass Spectrom. 1995,
RESULTS AND DISCUSSION
9, 641-643.
Basis of the Selection of Reagents Used. The structures of
(28) Krause, E.; Wenschuh, H.; Jungblut, P. R. Anal. Chem. 1999, 71, 4160-
the reagents tested in this study are summarized in Figure 1. To
4165.
6992 Analytical Chemistry, Vol. 76, No. 23, December 1, 2004

Figure 2. MALDI-TOF spectra of the derivatized asialobiantennary N-glycan (NA2) by (a) GT, (b) GP, (c) HEITA, (d) EHIEP, (e) Rh, (f) BRh,
(g) ZRh, (h) ZFRh, (i) ZWRh, and (j) ZYRh. The sample amounts loaded on the target were 50 fmol for GT, GP, HEITA, Rh, and ZRh and 4.5
fmol for EHIEP, BRh, ZFRh, ZWRh and ZYRh. The asterisk shown in (g) indicates the byproduct. All spectra were acquired with DHB as matrix.
Evaluation of Enhanced Detection Sensitivity upon De-
rivatization. All of the reagents properly reacted with a model
oligosaccharide, asialobiantennary N-glycan (NA2), to give ex-
pected hydrazones under the conditions described in the Experi-
mental Section. As shown in Figure 2a-d, derivatives of GT, GP,
HEITA, and EHIEP were detected solely as M+ at m/z 1754.9,
1774.8, 1883.8, and 1945.9, respectively, on positive reflectron
mode MALDI-TOF analysis using DHB as a matrix. M+ ions were
the only ion species detected that were consistent with the
previous observations for those derivatives of other reagents
having a quaternary ammonium center,26,27 and this may be
explained by the presence of a permanent positive charge. On
the other hand, derivatives of Rh, BRh, ZRh, ZFRh, ZWRh, and
ZYRh were detected as [M + H]+ at m/z 1811.8, 1915.8, 1945.9,
2092.8, 2131.8, and 2108.8, respectively (Figure 2e-j).
tive cationic charge by reacting with Girard’s reagent T, TMAPA,
and the dimethylation of the reductive amination product of sugars
with benzylamine29 equally increased the sensitivity about an order
of magnitude, the detection limits of the derivatives of GT, GP,
HEITA, and EHIEP differed considerably while they all carried
constitutive cationic charge. Likewise, the detection limits of
derivatized NA2 by the guanidino series reagents differed ∼10-
fold: the derivatized NA2 by Rh, BRh, ZRh, ZFRh, ZYRh, and
ZWRh could be detected down to 10, 1, 5, 1, 1, and 1 fmol,
respectively, loaded on the target. Derivatization of NA2 using
ZRh tended to give a relatively intense signal of the byproduct at
90 mu below the molecular ion [MH]+, which may be attributable
to the cleavage of the benzyl group. Krause et al. demonstrated
that signals from arginine-containing peptides are generally
stronger than lysine-containing peptides in a tryptic digest,28 so a
recent flurry of research has been directed toward increasing mass
spectral signal intensities from lysine-containing peptides,17,18 for
example, modifying lysine residues with O-methylisourea to form
more basic homoarginine residues. The current study has revealed
The approximate detection limits for each derivatized oligosac-
charide was determined by acquiring the spectra of the derivatized
NA2 in decreasing amount using serial dilution. The derivatized
NA2 by GT, GP, HEITA, and EHIEP could be detected down to
20, 10, 7.5, and 1 fmol, respectively, loaded on the target. Despite
the previous observations that derivatives incorporating a constitu-
(29) Broberg, S.; Broberg, A.; Duus, J. O. Rapid Commun. Mass Spectrom. 2000,
14, 1801-1805.
Analytical Chemistry, Vol. 76, No. 23, December 1, 2004 6993

Figure 3. MALDI-TOF spectra obtained for the mixed reaction mixture of derivatized asialobiantennary N-glycan (N2) by (a) GT, GP, HEITA,
and EHIEP and (b) Rh, BRh, ZRh, ZFRh, ZWRh, and ZYRh. The sample amounts loaded onto the target were (a) 114 fmol and 10 fmol, each.
All spectra were acquired with DHB as matrix.
that the incorporation of hydrophobic residue(s), such as phenyl,
indole, and benzoyl groups together with the guanidino group,
further increase detection sensitivity. This observation is partly
supported by the recent publication of Baumgart et al. that the
signal-enhancement impact of peptides depends on the amino acid
compositions as well as on the identity of adjacent amino acids to
arginine.30
digest gave plural new signals while retaining most of the peptides’
signals at comparable intensities. Six of the new signals were
identical to the molecular ions ([M + Na]+) of the previously
reported most abundant N-glycans of ovalbumin33,34 (see Figure
4b).
When an aliquot of deglycosylated tryptic ovalbumin digest
was directly derivatized with a 70-fold excess of GT, GP, HEITA,
or EHIEP and subjected to MALDI-TOF analysis by simply mixing
the reaction mixture with DHB, the resulting mass spectra differed
significantly, depending on the reagent used (Figure 4c-f). The
result obtained from the EHIEP derivatives gave the most
remarkable result. Almost all of the peptides’ ions disappeared,
and all of the m/z values of the newly appeared ions upon EHIEP
derivatization were identical to those of previously reported
N-glycans (mainly high-mannose and hybrid structures). For the
derivatives of GT, GP, and HEITA, peptide ion suppression was
modest. Many of the peptides’ ions were still visible, and only a
limited number of the derivatized glycans were detectable. Upon
EHIEP derivatization, the detection of major ovalbumin oligosac-
charides could be achieved at an amount equivalent to 10 fmol of
glycoprotein by serial dilution (data not shown), which is
consistent with the estimated detection limit of the EHIEP-
derivatized glycan.
Among the guanidino series reagents, derivatization with BRh,
ZFRh, ZWRh, and ZYRh allowed for oligosaccharide-specific ion
embossment under the presence of tryptic peptides of ovalbumin
(Figure 4h, j, k, and l). Derivatization with ZRh also allowed for
the detection of the glycan ions, though the ability to suppress
the peptides’ ions was behind those of EHIEP, BRh, ZFRh, ZWRh,
and ZYRh (Figure 4i). Rh derivatization barely allowed the
detection of a few major glycans among the peptides’ signals
(Figure 4g). Thus, the specific glycan ion embossment ability in
a complex mixture appeared to correlate well with the relative
sensitivity that the derivatization by each reagent could achieve.
To elucidate whether the observed peptides’ ion suppression
phenomenon is attributable to the presence of excess reagent or
the presence of the labeled glycan species, the deglycosylated
tryptic digest of ovalbumin was mixed with a 100-fold excess of
EHIEP, and an aliquot was analyzed by MALDI-TOF. It was
confirmed that oligosaccharides do not react with EHIEP unless
the reaction mixture is heated. Almost all of the peptides’ ions
disappeared on the spectrum, indicating that the observed pep-
To further compare the relative sensitivities of the different
reagent derivatives, each derivatization reaction mixture by GT,
GP, HEITA, and EHIEP was mixed in equal quantities and
analyzed using MALDI-TOF to compare the relative intensity of
each derivative under exactly the same analytical conditions.
Derivative of EHIEP was predominantly detected with slight signal
intensities of the derivatives of GP and HEITA. GT derivative was
not observed (Figure 3a). Note that the derivatives of GT, GP,
and HEITA were clearly detectable at this concentration when
they were measured independently. Likewise, the reaction mix-
tures by Rh, BRh, ZRh, ZFRh, ZWRh, and ZYRh were mixed and
analyzed by MALDI-TOF. Derivative ions of BRh, ZFRh, ZWRh,
and ZYRh were dominantly observed at comparable intensities
with a lesser signal intensity of the ZRh adduct (Figure 3b). Rh
derivative was not even visible, while it should be detectable at
this concentration when it is measured independently. This may
be explained by ion suppression, in which glycan derivatives
having higher sensitivities could suppress those having lesser
sensitivities.
Evaluation of the Peptides’ Ion-Suppressing Ability of
Each Reagent. Ovalbumin tryptic digest was used as one of the
model studies. Following the tryptic digestion of ovalbumin, it was
further digested with PNGase F in the same vial to remove the
N-glycans and was subjected to MALDI-TOF analysis using DHB
as matrix. Since glycoproteins widely differ in their susceptibility
to enzymatic deglycosylation,31 tryptic digestion prior to PNGase
F digestion has been commonly used to ensure complete N-glycan
liberation.
As shown in Figure 4a, tryptic digest gave reasonable peptide
mass fingerprints in the MALDI-TOF analysis, whereas a better
quality peptide mass fingerprint (PMF) result was obtained with
CHCA (data not shown). This is often the first choice matrix for
PMF due to the predominantly higher mass peak intensities when
compared to DHB.32 Additional PNGase F digestion of the tryptic
(30) Baumgart, S.; Lindner, Y.; Kuhne, R.; Oberemm, A.; Wenschuh, H.; Krause,
E. Rapid Commun. Mass Spectrom. 2004, 18, 863-868.
(31) Tarentino, A. L.; Gomez, C. M.; Plummer, T. H., Jr. Biochemistry 1985, 24,
4665-4671.
(33) Duffin, K. L.; Welply, J. K.; Huang, E.; Henion, J. D. Anal. Chem. 1992, 64,
1440-1448.
(32) Thiede, B.; Siejak, F.; Dimmler, C.; Jungblut, P. R.; Rudel, T. Electrophoresis
2000, 21, 2713-2720.
(34) Harvey, D. J.; Wing, D. R.; Ku¨ster, B.; Wilson, I. B. H. J. Am. Soc. Mass
Spectrom. 2000, 11, 564-571.
6994 Analytical Chemistry, Vol. 76, No. 23, December 1, 2004

Figure 4. MALDI-TOF spectra of (a) tryptic digest of ovalbumin, (b) tryptic and PNGase F digest of ovalbumin, and direct derivatives of
deglycosylated tryptic digest of ovalbumin by (c) GT, (d) GP, (e) HEITA, (f) EHIEP, (g) Rh, (h) BRh, (i) ZRh, (j) ZFRh, (k) ZWRh, and (l) ZYRh.
The sample amounts loaded on the target were 750 fmol equiv to glycoprotein for (a) and (b), and 170 fmol equiv to glycoprotein for c-l. All
spectra were acquired with DHB as matrix. Ions labeled with O in (a) are ovalbumin peptides. 1, (Hex)3(HexNAc)2; 2, (Hex)4(HexNAc)2; 3,
(Hex)3(HexNAc)3; 4, (Hex)5(HexNAc)2; 5, (Hex)4(HexNAc)3; 6, (Hex)3(HexNAc)4; 7, (Hex)6(HexNAc)2; 8, (Hex)5(HexNAc)3; 9, (Hex)4(HexNAc)4;
10, (Hex)3(HexNAc)5; 11, (Hex)5(HexNAc)4; 12, (Hex)4(HexNAc)5; 13, (Hex)7(HexNAc)2; 14, (Hex)5(HexNAc)5; and 15, (Hex)6(HexNAc)5.
tides’ ion suppression is attributable to the presence of excess
reagent. On the other hand, when the reaction mixture was
analyzed after the removal of unreacted excess EHIEP, most of
the peptides’ ions were observed together with labeled glycans
(data not shown). These results indicated that efficient peptides’
ion suppression could be realized only when excess reagents were
present in the sample.
The detection of EHIEP-labeled oligosaccharide was found to
be quite tolerant of the large excess presence of EHIEP. The
presence of a large EHIEP excess did not give any detection
sensitivity reduction of EHIEP-labeled NA2, even when a 2000-
fold excess of EHIEP was present (data not shown). This was
also true for BRh, ZFRh, ZWR, and ZYRh. We thus concluded
that a higher reagent/sample molar ratio was advantageous for
an efficient oligosaccharide-specific ion embossment on the
MALDI-TOF analysis. To the authors’ knowledge, this is the first
successful example demonstrating the direct N-glycan profiling
in an unpurified glycoprotein tryptic digest without any purification
steps.
It has been considered that the high sensitivity on MALDI is
due to the high efficiencies during both the solid-gas phase
transition and ionization processes. Despite the fact that it is widely
used in many fields, the theories and models of MALDI are still
crude.35 The detailed mechanism of the observed highly efficient
ion suppressing ability needs to be further elucidated.
Applicability to Other Glycoproteins, Including Those
That Carry Sialylated Oligosaccharides. To further evaluate
selective oligosaccharide detection under the presence of the
tryptic peptides, this protocol was applied to other glycoproteins,
asialofetuin, ribonuclease B, and R1-acid glycoprotein. The de-
(35) Karas, M.; Kruger, R. Chem. Rev. 2003, 103, 427-440.
Analytical Chemistry, Vol. 76, No. 23, December 1, 2004 6995

Figure 5. MALDI-TOF spectra of derivatized tryptic and PNGase F digest of (A) asialofetuin, (B) ribonuclease B, and (C) R1-acid glycoprotein
by EHIEP (a) and ZFRh (b). The sample amounts loaded onto the target were 170 fmol equiv to glycoprotein. All spectra were acquired with
DHB as matrix. Ions with numbers are identical to (A) 1, (Hex)2(HexNAc)2 + (Man)3(GlcNAc)2; 2, (Hex)3(HexNAc)3 + (Man)3(GlcNAc)2; (B) 1,
(Hex)2 + (Man)3(GlcNAc)2; 2, (Hex)3 + (Man)3(GlcNAc)2; 3, (Hex)4 + (Man)3(GlcNAc)2; 4, (Hex)5 + (Man)3(GlcNAc)2; 5, (Hex)6 + (Man)3-
(GlcNAc)2; (C) 1, (Hex)2(HexNAc)2(Neu5Ac)1 + (Man)3(GlcNAc)2; 2, (Hex)2(HexNAc)2(Neu5Ac)2 + (Man)3(GlcNAc)2; 3, (Hex)3(HexNAc)3(Neu5Ac)2
+ (Man)3(GlcNAc)2; 4, (Hex)3(HexNAc)3(Neu5Ac)3 + (Man)3(GlcNAc)2; 5, (Hex)4(HexNAc)4(Neu5Ac)2 + (Man)3(GlcNAc)2; 6, (Hex)3(HexNAc)3-
(dHex)1(Neu5Ac)3 + (Man)3(GlcNAc)2; 7, (Hex)4(HexNAc)4(Neu5Ac)3 + (Man)3(GlcNAc)2; 8, (Hex)4(HexNAc)4(Neu5Ac)4 + (Man)3(GlcNAc)2.
glycosylated tryptic digest of each glycoprotein was derivatized
with 130-fold excesses of EHIEP and ZFRh, and the reaction
mixture was directly subjected to MALDI-TOF analysis using DHB
as the matrix (Figure 5). EHIEP derivatization allowed for the
detection of all of the previously reported oligosaccharides of
asialofetuin36 and ribonuclease B,37 and the peptides’ signal were
largely suppressed, although a few relatively intense contaminant
signals were observed around m/z 2400 for ribonuclease B.
However, regarding R1-acid glycoprotein, EHIEP barely allowed
for the detection of a few derivatized glycan ions, whereas the
peptides’ ion suppression was fairly successful. Unlike asialofetuin
and ribonuclease B, almost all of the oligosaccharides of R1-acid
glycoprotein are reported to be sialylated.38 Although labeling
reagents having constitutive cationic charge have been reported
to successfully improve the detection sensitivities of oligosaccha-
rides, to the authors’ knowledge, their application had been limited
to neutral oligosaccharides. Considering that oligosaccharide
derivatives of reagents having constitutive cationic charge (e.g.,
Girard’s reagents T and P, HEITA, EHIEP) are detected solely
as M+ on the MALDI-TOF analysis, the reduction of the net
positive charge due to the incorporation of negative charge(s) from
sialic acid may drastically decrease desorption/ionization ef-
ficiency.
spectrum obtained for R1-acid glycoprotein was both qualitatively
and quantitatively consistent with that of the previous report,
which was obtained after converting sialic acid into methyl ester
by its reaction with methyl iodide.38 However, the relative signal
intensity of each sialylated oligosaccharide could somewhat vary
among the spectra acquired from firing the laser at different
locations on the target. Careful use of threshold laser irradiance
was required to detect ions that correspond to intact sialylated
oligosaccharides in a positive reflectron mode. On the other hand,
in a positive linear mode, the signal detection of the ZFRh
derivatives of intact sialylated oligosaccharides was much less
sensitive to the laser irradiance power. The relative signal intensity
of each sialylated oligosaccharide was fairly stable among spectra
acquired from firing the laser at different locations on the target
as far as peripheral regions of the crystal were chosen (data not
shown). Despite the previous observation that sialylated oligosac-
charides analyzed with DHB in either the linear or reflectron
positive mode often lose sialic acid, and these problems are further
aggravated as the number of sialic acid residues incorporated into
the oligosaccharide increases,39 ZFR derivatives appeared to be
quite stable through ionization under linear mode and may even
be detectable under reflector mode. This may be explained by
the highly improved ionization/desorption efficiency upon deriva-
tization that allowed labeled glycans to be detected with a much
lower laser power, and hence, the internal energy supplied by
the laser may be reduced. This superior nature observed for ZFRh
is also true for BRh, ZWRh, and ZYRh.
In contrast, ZFRh derivatization allowed for the detection of
oligosaccharides for all of the model proteins used, regardless of
the presence or absence of sialic acid. Sialylated oligosaccharides,
detected solely as [M + H]+, were detected for neutral oligosac-
charides. The detection of the major sialylated oligosaccharides
of R1-acid glycoprotein could be achieved at an amount equivalent
to 10 fmol of glycoprotein loaded on the target by serial dilution,
suggesting that the detection sensitivity of sialylated oligosaccha-
ride is comparable to that of a neutral oligosaccharide. The mass
Securing a series of labeling reagents could provide some
practically useful applications. For instance, when more than one
reagent is used for a single sample, one can be assured that the
detected signals originate from oligosaccharides by comparing
the m/z differences of the detected signals of each spectrum with
those of the molecular weights of the employed labeling reagents.
This is especially true for glycans that are present at lower
(36) Yet, M. G.; Chin, C. C.; Wold, F. J. Biol. Chem. 1988, 263, 111-117.
(37) Fu, D.; Chen, L.; O’Neill, R. A. Carbohydr. Res. 1994, 261, 173-186.
(38) Ku¨ster, B.; Hunter, A. P.; Wheeler, S. F.; Dwek, R. A.; Harvey, D. J.
Electrophoresis 1998, 19, 1950-1959.
(39) Papac, D. I.; Wong, A.; Jones, A. J. Anal. Chem. 1996, 68, 3215-3223.
6996 Analytical Chemistry, Vol. 76, No. 23, December 1, 2004

concentrations and with molecular weights that fall intermediate
to those of the more abundant glycans.
ity and was necessary to achieve significant ion suppression of
other contaminants, such as peptides. Moreover, the derivatization
of oligosaccharides with ZFRh and related compounds could allow
for the sensitive detection of oligosaccharides, regardless of the
presence or absence of sialic acid.
Mass spectrometry has in the past decade become accepted
as a key analytical technique among biology and biochemistry
laboratories. Due to its inherent limitation in the direct analysis
of complex mixtures, mass spectrometry has often been adopted
to combine sophisticated separation technologies, that is, LC, GC,
gel electrophoresis, and capillary electrophoresis, prior to analysis.
Though the central role of such a separation prior to the detection
and evaluation of analytes will be indisputable, the concept to
regulating ion enhancement and suppression upon derivatization
may provide a new way to help reduce the number of sample
preparation steps.
The current protocol did not allow for the entire suppression
of the ions of other species (e.g., a few intense signals observed
for ribonuclease B and several weak signals observed for asialo-
fetuin). These ions would complicate the N-glycan profiling of the
real sample in which the glycosylation characteristics are un-
known; however, the concept to selectively enhance the ioniza-
tion/desorption efficiency of glycans in a crude mixture upon
glycan-selective derivatization and, hence, to minimize the sample
preparation steps have been proven. To integrate the described
protocol with the current proteome analysis workflow, further
study is currently in progress in our laboratory.
CONCLUSION
In this article, we proposed and demonstrated a novel concept
that glycan ions could be highly selectively embossed on MALDI-
TOF, even without the need for a sample cleanup under the
condition that the ionization efficiency of glycan is selectively and
substantially enhanced upon glycan-specific derivatization (hy-
drazone formation). We prepared a series of novel derivatization
reagents that possess either quaternary ammonium/pyridinium
or guanidino functionalities as ionization-promoting groups. The
ability to enhance the signal is dependent not only on the presence
of these functionalities, but also on the other part of the reagent
structure. The incorporation of hydrophobic residue, such as the
phenyl, indole, and benzoyl groups, further increased the sensitiv-
ACKNOWLEDGMENT
This work was supported in part by the National Project on
Functional Glycoconjugate Research Aimed at Developing for New
Industry from the Ministry of Education, Science, Sports and
Culture of Japan.
Received for review May 18, 2004. Accepted August 27,
2004.
AC0492766
Analytical Chemistry, Vol. 76, No. 23, December 1, 2004 6997