Aniline: A Catalyst for Sialic Acid Detection
Jennifer J. Kohler*[a]
In an era in which much of life-science
research seems to sprint ahead by ex-
ploiting postgenomic, systems-wide ap-
proaches, the glycobiologist’s progress is
often stymied by the nearly insurmount-
able hurdles imposed by techniques and
reagents that seem a generation out-of-
date. In a recent report in Nature Meth-
ods,the Paulson and Dawson groups
from the Scripps Research Institute de-
scribe how simple addition of an appro-
priate catalyst transforms an old glycan-
detection method, thus paving the way
for a variety of glycomic applications.
Some scientists choose to focus on
glycobiology, while others have glycobi-
ology thrust upon them. Increasingly,
biologists of all stripes—immunologists,
developmental biologists, cancer biolo-
gists—are making discoveries that point
to glycosylation state as a key regulator
of the activity or localization of a mole-
cule of interest. As these scientists seek
to follow up on their observations, their
next discovery might be about the chal-
lenges associated with glycobiology re-
search. The goal is often to understand
the relationship between a glycoconju-
gate’s activity and its glycosylation state.
But tracking the abundance, location, dy-
namics, and function of individual glyco-
forms is a significant undertaking: label-
ing glycans is not as simple as adding a
green fluorescent protein (GFP) tag. In-
stead, the options generally include pro-
tein- and chemical-detection reagents.
Glycan detection by lectins and anti-
bodies remains an important analytical
approach, as it has for over 40 years.
Lectins, proteins that recognize glycan
structures, are naturally occurring sen-
sors that first became available in puri-
fied forms in the 1960sand have since
been widely used to profile glycosylation
patterns, most recently in microarray
format.[3–5]Unfortunately, data obtained
by using lectins can be difficult to inter-
pret because lectins typically suffer from
poor specificities and low affinities. Anti-
bodies also serve as important glycan-
detection reagents—indeed, many of
the clusters of differentiation (CDs) used
to define immune cells are based on
topes—yet glycan-recognizing antibod-
ies can be difficult to produce and their
specificities are often context depen-
In the 1990s, metabolic oligosacchar-
ide engineering was introduced as an al-
ternative to these detection methods: by
culturing cells with sugars that contain
an abiotic functional group, one can
handle into cell-surface glycans.The
handle—typically an azide or alkyne—is
capable of participating in chemoselec-
tive reactions to add fluorescent or
biotin tags to specific glycans. Metabolic
groups has the advantages of providing
very specific detection—there is little
cross-reactivity—and relying on covalent
bond formation. But because this tech-
nique requires that cells be cultured
with the sugar precursor prior to label-
ing, it cannot easily be applied to clinical
Rather than introducing an unnatural
functional group into glycans, Paulson
and Dawson exploited the intrinsic reac-
tivity of naturally occurring sialic acid.
Sialic acid is the general name for a class
sugars that includes N-acetylneuraminic
acid (NeuAc) and N-glycolylneuraminic
acid (NeuGc). Sialic acids have a polyol
side chain that usually includes hydroxyl
groups at C-7, C-8, and C-9. This side
chain is particularly sensitive to oxidation
by periodate,which oxidizes vicinal
diols to aldehydes that can subsequently
react with appropriate amines to form
imines. Periodate oxidation followed by
Schiff base formation is the basis of the
widely used periodic acid–Schiff (PAS)
histology stain,which detects many
different carbohydrates. But under suita-
bly mild conditions, periodate selectively
oxidizes sialic acids to their C-7 aldehyde
counterparts, while the vicinal diols in
other sugars remain intact (Scheme 1).[10,11]
Because aldehydes and ketones are
otherwise absent from the mammalian
cell surface, this reaction introduces a
unique functional group that can be
chemoselectively tagged by treatment
with an aminooxy (sometimes referred
to as hydroxylamine) reagent to yield an
oxime adduct, or with a hydrazine to
yield a hydrazone. Indeed, in 1975,
Weber and Hof used mild periodate oxi-
dation followed by treatment with dan-
sylhydrazine to fluorescently label sialic
acid residues in glycolipids, glycopepti-
des, and glycoproteinsand, more re-
cently, de Bank et al. used a similar reac-
tion to fluorescently label sialic acids on
the surface of living myoblasts.
The oxime and hydrazone ligation re-
actions also find wider use in a variety
of bioconjugation reactions, including
peptide ligations.By relying on these
same ligation reactions, specific intro-
duction of aldehydes or ketones has
enabled site-specific protein labeling in
vitro,labeling of proteins on the bac-
terial cell surface,and metabolic label-
ing of glycans on the mammalian cell
surface.In addition, a genetically en-
coded “aldehyde-tag” has been devel-
oped to facilitate labeling of recombi-
The ability of ketones and aldehydes
to participate in chemoselective reac-
tions offers a conceptually simple way to
selectively tag these functional groups.
But, in practice, the chemistry of the re-
action imposes limitations. The reaction
[a] Dr. J. J. Kohler
Division of Translational Research
Department of Internal Medicine
University of Texas Southwestern Medical
Dallas, TX 75390-9185 (USA)
ChemBioChem 2009, 10, 2147–2150 ? 2009 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim
between an aldehyde and an a-effect ni-
trogen (in an aminooxy or hydrazine re-
agent) is optimal under slightly acidic
conditions (pH 5–6) and even then pro-
ceeds at only a modest rate. The slow re-
action kinetics can be partially overcome
by using high concentrations of the ami-
nooxy or hydrazine reagent, but quanti-
tative labeling remains elusive. The slug-
gish kinetics and pH dependence of
these reactions result from the require-
ment for protonation of the carbonyl
oxygen.[19,20]Lowering the pH of the re-
action mixture increases the concentra-
tion of protonated carbonyl, but simulta-
neously decreases the reactivity of the
aminooxy or hydrazine nucleophile by
protonating that molecule as well, so it
is difficult to achieve rapid kinetics at
The ingredient that has been lacking
in these reactions is an appropriate nu-
unique ability to fill this role. Aniline cat-
alyzes oxime ligations through a transi-
mination mechanism (Scheme 2). Like a-
effect nitrogens, aniline reacts with pro-
tonated carbonyls to form imines, but
the pKaof these aniline imines is such
that they are significantly protonated at
the pH of the reaction mixture. Thus, the
protonated aniline Schiff base is poised
to react rapidly with an aminooxy or hy-
drazine reagent, forming an oxime or hy-
drazone product that does not readily
re-react with aniline. The ability of aniline
to function as a catalyst in these types of
reactions was first reported in 1961,
but has only recently been applied to
The new advance in the recent Nature
Methods report is the recognition that
aniline catalysis could be used to accel-
erate reaction of oxidized sialic acids
with aminooxy detection reagents in a
cellular setting. The addition of aniline
results in a tenfold increase in ligation
efficiency, as measured by flow cytome-
try, and transforms a mediocre labeling
reaction into highly effective one. The
striking effect of aniline can be visualized
by immunoblot detection, where barely
perceptible bands become distinct, and
bright. Paulson and colleagues dub their
Scheme 1. Covalent labeling of sialylated glycoconjugates. Mild periodate treatment oxidizes the C-7 position of sialic acids to an aldehyde (red circle), while
stronger oxidation conditions also oxidize vicinal diols present in other sugars. Subsequent treatment with aminooxy detection reagents results in oxime liga-
tion products, while the less-stable hydrazone ligation products can be formed by treatment with hydrazine detection reagents. Commercially available ami-
nooxy and hydrazine detection reagents come in biotin- or fluorophore-functionalized versions and have the potential to be used for a variety of applica-
? 2009 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim ChemBioChem 2009, 10, 2147–2150
Sialic Acid Detection
new protocol PAL (periodate oxidation
perhaps in a nod to the tried-and-true
PAS histology stain.
All of the reagents necessary to imple-
ment the PAL method are commercially
available. Researchers can choose be-
tween aminooxy-biotin or one of several
aminooxy-fluorophores, depending on
the intended application. Because the
reaction proceeds more efficiently with
aniline, the researcher can use a ten- to
50-fold lower concentration of the ami-
nooxy reagent than is typically used in
the uncatalyzed oxime ligation, thus re-
sulting in a substantial cost savings. Al-
though not examined in this paper, ani-
line is also known to catalyze the reac-
tion of aldehydes with hydrazine to form
hydrazone ligation products.Since hy-
drazine-detection reagents are some-
what less expensive than their aminooxy
counterparts, the researcher might con-
sider using them instead. However, hy-
drazone products are less stable than
oximes,so, for some applications, the
added expense of the aminooxy reagent
One technical concern about the im-
plementation of this technique is the se-
lectivity of periodate oxidation.It is
imperative that sialic acids are oxidized
with high efficiency, while other vicinal
diols are unaffected. The Nature Methods
authors demonstrate that they have ach-
ieved such conditions by conducting
control experiments with cell lines that
lack sialic acid. Essentially no labeling is
observed unless the growth medium is
supplemented with sialic acid. Research-
ers who wish to apply this technique to
their own samples and cell lines will
need to conduct appropriate control
experiments to confirm that they have
found the sweet spot where sialic acid is
oxidized while other sugars are not. It is
also worth noting that the structure of
sialic acids endows them with unique re-
activity; devising methods to chemose-
lectively label naturally occurring sugars
other than sialic acid promises to be a
much more challenging task.
comes at a time when many new roles
are being identified for sialic acid and
sialosides. For example, recent studies
show that the NeuGc form of sialic acid
mediates susceptibility to infection by
Shiga-toxigenic E. coli,that the tropism
of influenza virus is determined by the
context in which sialic acid is presented
by host cells,that sialylation of intra-
venously administered immunoglobulin
is responsible for its anti-inflammatory
activity,that elevated expression of a
human sialyltransferase (ST6GalNAc5) is
a reliable predictor of the ability of
breast cancers to metastasize to the
brain,and that expression of another
sialyltransferase (ST3GaL5) is essential for
the development of hearing in mice.
The PAL method has the potential to fa-
cilitate future work in all of these areas.
The most immediate application for
aniline-catalyzed sialic acid labeling is
likely to be in glycoproteomics experi-
ments aimed at profiling the “sialome”
of particular cells or tissue types. Previ-
ous glycoproteomics experiments have
made use of extensive periodate oxida-
tion followed by hydrazone ligation,a
nonselective approach that results in the
isolation of all glycoproteins, not just the
sialylated ones. Alternatively, lectin affini-
ty chromography (LAC) has been used
to enrich sialylated proteins, but is ham-
pered by the relatively poor affinities of
sialic acid binding lectins.The ability
to tag sialylated proteins covalently and
with high efficiency will bring the glyco-
proteomic analysis of sialylated mole-
cules into the 21st century.
Financial support from the University of
Texas Southwestern Medical Center, the
0834336), and the Welch Foundation (I-
1686) is gratefully acknowledged. J.J.K.
thanks Joseph Ready and Seok-Ho Yu for
comments on the manuscript.
Scheme 2. Aniline catalysis of oxime ligations. The uncatalyzed oxime ligation begins with addition of
an aminooxy reagent to an aldehyde, forming a carbinolamine. This initial step is slow because it re-
quires protonation of the aldehyde, but not the aminooxy group. Next, the carbinolamine undergoes
dehydration to form the oxime ligation product. The oxime ligation can be catalyzed by excess aniline,
which reacts with the aldehyde to form an imine. The aniline imine is readily protonated, enabling it to
react with aminooxy reagent. Subsequent loss of aniline leads to formation of the stable oxime prod-
ChemBioChem 2009, 10, 2147–2150? 2009 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim
J. J. Kohler
hydrazones · oximes · sialic acids
 Y. Zeng, T. N. C. Ramya, A. Dirksen, P. E.
Dawson, J. C. Paulson, Nat. Methods 2009, 6,
 B. B. Agrawal, I. J. Goldstein, Biochem. J.
1965, 96, 23c.
 K. T. Pilobello, D. E. Slawek, L. K. Mahal, Proc.
Natl. Acad. Sci. USA 2007, 104, 11534.
 S. C. Tao, Y. Li, J. B. Zhou, J. Qian, R. L.
Schnaar, Y. Zhang, I. J. Goldstein, H. Zhu, J. P.
Schneck, Glycobiology 2008, 18, 761.
 H. Tateno, N. Uchiyama, A. Kuno, A. Togaya-
chi, T. Sato, H. Narimatsu, J. Hirabayashi, Gly-
cobiology 2007, 17, 1138.
 R. Schwartz-Albiez, B. Kniep, Cell. Immunol.
2005, 236, 48.
 D. H. Dube, C. R. Bertozzi, Curr. Opin. Chem.
Biol. 2003, 7, 616.
 R. G. Spiro, J. Biol. Chem. 1964, 239, 567.
 R. D. Lillie, Stain Technol. 1951, 26, 103.
 G. W. Jourdian, L. Dean, S. Roseman, J. Biol.
Chem. 1971, 246, 430.
 T.-H. Liao, P. M. Gallop, O. O. Blumenfeld, J.
Biol. Chem. 1973, 248, 8247.
 P. Weber, L. Hof, Biochem. Biophys. Res.
Commun. 1975, 65, 1298.
 P. A. De Bank, B. Kellam, D. A. Kendall, K. M.
Shakesheff, Biotechnol. Bioeng. 2003, 81, 800.
 J. Shao, J. P. Tam, J. Am. Chem. Soc. 1995,
 V. W. Cornish, K. M. Hahn, P. G. Schultz, J. Am.
Chem. Soc. 1996, 118, 8150.
 Z. W. Zhang, B. A. C. Smith, L. Wang, A.
Brock, C. Cho, P. G. Schultz, Biochemistry
2003, 42, 6735.
 L. K. Mahal, K. J. Yarema, C. R. Bertozzi, Sci-
ence 1997, 276, 1125.
 P. Wu, W. Q. Shui, B. L. Carlson, N. Hu, D.
Rabuka, J. Lee, C. R. Bertozzi, Proc. Natl. Acad.
Sci. USA 2009, 106, 3000.
 S. Rosenberg, S. M. Silver, J. M. Sayer, W. P.
Jencks, J. Am. Chem. Soc. 1974, 96, 7986.
 J. M. Sayer, B. Pinsky, A. Schonbrunn, W.
Washtien, J. Am. Chem. Soc. 1974, 96, 7998.
 E. H. Cordes, W. P. Jencks, J. Am. Chem. Soc.
1962, 84, 826.
 A. Dirksen, T. M. Hackeng, P. E. Dawson,
Angew. Chem. 2006, 118, 7743; Angew. Chem.
Int. Ed. 2006, 45, 7581.
 A. Dirksen, S. Dirksen, T. M. Hackeng, P. E.
Dawson, J. Am. Chem. Soc. 2006, 128, 15602.
 J. Kalia, R. T. Raines, Angew. Chem. 2008, 120,
7633; Angew. Chem. Int. Ed. 2008, 47, 7523.
 J. Needham, P. E. Reid, D. A. Owen, Histo-
chem. J. 1991, 23, 290.
 E. Byres, A. W. Paton, J. C. Paton, J. C. Lofling,
D. F. Smith, M. C. J. Wilce, U. M. Talbot, D. C.
Chong, H. Yu, S. S. Huang, X. Chen, N. M.
Varki, A. Varki, J. Rossjohn, T. Beddoe, Nature
2008, 456, 648.
 J. Stevens, O. Blixt, L. Glaser, J. K. Tauben-
berger, P. Palese, J. C. Paulson, I. A. Wilson, J.
Mol. Biol. 2006, 355, 1143.
 Y. Kaneko, F. Nimmerjahn, E. V. Ravetch, Sci-
ence 2006, 313, 670.
 P. D. Bos, X. H. F. Zhang, C. Nadal, W. P. Shu,
R. R. Gomis, D. X. Nguyen, A. J. Minn, M. J.
van de Vijver, W. L. Gerald, J. A. Foekens, J.
Massague, Nature 2009, 459, 1005.
 M. Yoshikawa, S. Go, K. Takasaki, Y. Kakazu,
M. Ohashi, M. Nagafuku, K. Kabayama, J. Se-
kimoto, S. Suzuki, K. Takaiwa, T. Kimitsuki, N.
Matsumoto, S. Komune, D. Kamei, M. Saito,
M. Fujiwara, K. Iwasaki, J. Inokuchi, Proc. Natl.
Acad. Sci. USA 2009, 106, 9483.
 H. Zhang, X. J. Li, D. B. Martin, R. Aebersold,
Nat. Biotechnol. 2003, 21, 660.
 J. Zhao, D. M. Simeone, D. Heidt, M. A. An-
derson, D. M. Lubman, J. Proteome Res. 2006,
Received: July 1, 2009
Published online on July 27, 2009
? 2009 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim ChemBioChem 2009, 10, 2147–2150
Sialic Acid Detection