MALDI-TOF mass spectrometric analysis of enzyme activity and lectin trapping on an array of N-glycans

Article abstract of DOI:10.1002/anie.201006304

Active arrays: Complex lipid-tagged oligosaccharides, including large multiantennary species, can be efficiently immobilized on self-assembled monolayers of alkyl mercaptans . These arrays can be used to follow the action of a galactosyltransferase (GalT) and a hydrolase. The utility of the system for the selective trapping and identification of a lectin from a complex mixture was also demonstrated.

Full text of DOI:10.1002/anie.201006304

DOI: 10.1002/anie.201006304
Glycan Arrays
MALDI-TOF Mass Spectrometric Analysis of Enzyme Activity and
Lectin Trapping on an Array of N-Glycans**
Antonio Sanchez-Ruiz, Sonia Serna, Nerea Ruiz, Manuel Martin-Lomas, and Niels-
Christian Reichardt*
Glycan microarrays are an established platform for the high-
throughput screening of substrate specificities of carbohy-
drate-binding proteins and processing enzymes.[1] The most
common detection method, which is based on fluorescently
tagged lectins, can only give a measure of binding specificity,
as quantification is often compromised by the specific lectin
affinity.[1a,b] Therefore, a label-free technique that could
overcome these analytical problems is needed for the analysis
of array spot compositions.
MALDI-TOF-based assays on surface-bound carbohy-
drates have been reported. Mrksich and co-workers have
shown that this technique can be used to study enzyme
activity or to trap affinity ligands on biofunctionalized self-
assembled monolayers (SAMs) of oligoethylene glycols on
gold surfaces.[2] Unfortunately, the covalent attachment of
glycans to the monolayer requires high ligand concentrations
that are not suitable when working with complex oligosac-
charides.[3,4] More recently, Wong, Siuzdak, and co-workers
studied 2,3-sialyltransferase and b-galactosidase activity on
fluorous-tagged lactose immobilized on a perfluorinated
surface.[5] The difficult preparation of the nanostructured
surface and the multistep tagging procedure, however, render
this approach less appealing for routine and high-throughput
use in on-chip mass spectrometric analysis of enzyme ac-
tivity.
and noncovalently immobilized on the MALDI plate by
insertion into a self-assembled alkylthiolate monolayer. This
setup was tested in a series of glycomics applications.
The facile preparation of both a hydrophobic MALDI
plate and tagged ligands, as well as the general applicability
for large oligosaccharides make this method stand out from
other surface-based MALDI-TOF approaches. Hydrophobic
surfaces have been applied to MALDI-TOF-based proteo-
mics for sample desalting,[7] but, to the best of our knowledge,
these surfaces have not been applied to the oriented
immobilization of lipid-tagged biomolecules for surface
MALDI-TOF analysis.
We chose a commercially available gold-coated MALDI
sample plate as surface to prepare a hydrophobic self-
assembling monolayer of 1-undecanethiol by following a
published procedure.[8] Analysis of the monolayer under
standard MALDI-TOF conditions showed only peaks corre-
sponding to known matrix ions,[9] while no disulfide or
thiolate ions were detected (Figure 1). The hydrophobically
tagged carbohydrate ligands 1–10 used in this study were
prepared from synthetic N-glycan structures[10] (2–8) by
conjugation with stearic acid or from commercial reducing
sugars by reductive amination (1, 9, and 10). Glycan arrays
(see the Supporting Information) were made by spotting the
conjugates 2–10 onto individual wells of the hydrophobic
sample plate, drying, and rinsing the plates with water to
remove unbound material. MALDI-TOF analysis using 2,4,6-
trihydroxy-acetyophenone (THAP) as matrix showed strong
ion signal intensities with signal-to-noise values of typically
100 or higher for all compounds; these values are comparable
to those reported by Siuzdak and co-workers.[5] Ions were
detected as sodium adducts with a detection limit of around
5 picomol. The maximum surface capacity for glycan immo-
bilization was determined by deposition of increasing
amounts of conjugate 2, washing, and measurement of the
signal for 2 normalized to an internal standard. Surface
saturation was reached after deposition of around 2 nmol of
conjugate 2 per well, which translates to a surface concen-
tration of 0.2 mmolmmÀ2.
Inspired by the flexible and mobile organization of
glycolipids in lipid bilayers,[6] we present herein a novel
strategy for the surface-based MALDI-TOF analysis of
glycan arrays that is conceptually a fusion of the approaches
of Wong, Siuzdak, Mrksich, and their respective co-work-
ers.[2a,5] Glycans are functionalized with a lipid tag (Scheme 1)
[*] Dr. A. Sanchez-Ruiz, Dr. S. Serna, Dr. N. Ruiz,
Prof. Dr. M. Martin-Lomas, Dr. N.-C. Reichardt
Biofunctional Nanomaterials Unit, CICbiomaGUNE
Paseo Miramon 182, 20009 San Sebastian (Spain)
Fax: (+34)943-005-314
E-mail: nreichardt@cicbiomagune.es
Prof. Dr. M. Martin-Lomas
CIBER-BBN, San Sebastian (Spain)
To avoid the unnecessary waste of valuable analytes,
glycans were spotted at half the saturation concentration
without compromising signal intensity. The stability of the
immobilized glycans to repeated washing with aqueous
buffers, water, or organic solvents was then determined.
Tagged glycans typically resisted 3–5 wash cycles of 1 minute
duration without significant reduction of signal intensity, and
the intensity of the model glycan 2 was essentially unchanged
after 60 seconds of continuous sonication (Figure 1d). How-
ever, the glycans were completely removed from the surface
[**] We thank Dr. Juan Etxebarria for help with enzymatic extension and
lectin incubation and Dr. Eduardo Junceda (Madrid) for a gift of
E. coli. expressing a a-1,6-fucosyltransferase from Rhizobium sp.
Funding from Ministerio de Ciencia e Innovaciꢀn (grant CTQ2008-
04444/BQA), the Government of the Basque Country, an Etortek
grant, and the European Union (ITN-Euroglycoarrays grant) is
gratefully acknowledged.
Supporting information (including experimental procedures and
MALDI-TOF mass spectra) is available for this article is available on
the WWW under http://dx.doi.org/10.1002/anie.201006304.
Angew. Chem. Int. Ed. 2011, 50, 1801 –1804
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1801

Communications
Scheme 1. a) Tagged glycan structures employed in this study. b) Pictogram representation of the glycans.
after washing with organic solvents. Since enzymatic reactions
are usually carried out in buffered aqueous solutions, the
immobilization method seemed suitable for our purposes.
Thus an array of immobilized glycans was tested for three
glycomics applications: substrate screening of glycosyltrans-
ferases, measurement of glycosidase activities,[2a,5] and lectin
identification.
The first trials for the enzymatic galactosylation of the
immobilized glycans 2–8 with bovine b-1,4-galactosyltransfer-
ase (GalT) and UDP-galactose (UDP-Gal) with addition of
30% glycerol to avoid evaporation showed incomplete
conversion even after prolonged reaction times. When the
blocking sugar 1 and a sample glycan were spotted in a ratio of
10:1, however, complete conversion of all glycans was
observed. We assume that the major function of the blocking
sugar is to cover the exposed hydrophobic patches and to
dilute the ligand surface density, thereby improving enzyme
accessibility. We were able to reduce the blocking sugar/
glycan ratio to 2:1 to increase the signal strength while
maintaining complete conversion in the enzyme reaction.
Enzyme activity could be detected as soon as 10 minutes
after incubation, but full conversion of multiantennary
glycans, which is a requisite for on-chip enzymatic prepara-
tion of glycan arrays,[10] took up to 60 hours. The resulting
array of fully galactosylated glycans was then used to follow
the action of a b-1,4-galactosidase from Aspergillus oryzae
(Figure 2). After 1.5 hours, almost complete hydrolysis of
galactose residues was observed for most of the glycans with
1802
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 1801 –1804

Figure 2. a) Scheme of enzymatic processing on surface-bound glycans.
b) Transformations performed on glycan 7. c) Observed m/z values of C18-
tagged glycans after treatment with b-1,4-galctosyltransferase (60 h, 378C)
and b-galactosidase (1.5 h, 308C).
functionalized with tagged glycan 9 was incubated with wheat
germ agglutinin (WGA) for 40 minutes at room temperature,
rinsed with water to remove nonspecifically bound lectin, and
analyzed. MALDI-TOF analysis of 9 bound in the well
showed peaks for the single- and double-charged lectin at
17160 Da and 8705 Da, respectively (Figure 3c), while no
signal for the lectin was observed from a control well
functionalized with maltopentaose derivative 10. This inter-
action could be detected down to a WGA concentration of
130 nm.
As a second lectin, we tested Phytolacca americana lectin
(PAL), which specifically recognizes the N-acetylglucosamine
trimer. Immobilized glycoconjugate 9 was incubated with the
lectin, washed, and analyzed by MALDI-TOF MS. The
observed signals at 13795 Da, 9343 Da, and 9130 Da corre-
Figure 1. a) Formation of glycan-functionalized surfaces for MALDI-
TOF MS analysis. b) MALDI-TOF spectrum of immobilized trianten-
nary glycan 7 after washing. c) Saturation graph of the MALDI plate for
chitobiose. d) Desorption graph of chitobiose adsorbed on the MALDI
plate. A=peak area.
the exception of glycans 3 and 7, which showed only partial
hydrolysis of galactose residues (Figure 2b).
To explore the utility of this platform for lectin identifi-
cation, we incubated sample wells functionalized with differ-
ent tagged glycans (2, galactosylated 8, 9, and 10 as control)
and investigated binding of three purified lectins, a mixture
with different specificities, and the selective trapping of a
lectin from a cell lysate (Figure 3). A well of the sample plate
Angew. Chem. Int. Ed. 2011, 50, 1801 –1804
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1803

Communications
specificity of the interaction. In a related experiment,
immobilized chitobiose 2 was incubated with a lysate from
an E. coli culture spiked with WGA at 580 nm, washed, and
analyzed. Again, WGA was detected with good signal
intensity only on the chitobiose-functionalized wells.
In summary, we have developed a simple and efficient
platform for the immobilization and analysis of surface-bound
complex oligosaccharides. With this system, it was possible to
follow the action of a glycosyltransferase and a glycosidase on
surface-bound multiantennary N-glycans. This technique will
allow optimization of reaction conditions for the enzymatic
on-chip elongation of surface-bound glycans to increase
structural diversity on glycan microarrays.[10] In addition, it
has been shown that this method is sensitive enough to
observe the specific but weak interaction between immobi-
lized glycans and lectins in various cases.
Received: October 7, 2010
Published online: January 14, 2011
Keywords: glycan arrays · hydrolases · monolayers ·
.
oligosaccharides · transferases
[1] a) P. H. Liang, C. Y. Wu, W. A. Greenberg, C. H. Wong, Curr.
Opin. Chem. Biol. 2008, 12, 86 – 92; b) N. Laurent, J. Voglmeir,
S. L. Flitsch, Chem. Commun. 2008, 4400 – 4412; c) T. Horlacher,
P. H. Seeberger, Chem. Soc. Rev. 2008, 37, 1414 – 1422.
[2] a) J. Su, M. Mrksich, Angew. Chem. 2002, 114, 4909 – 4912;
Angew. Chem. Int. Ed. 2002, 41, 4715 – 4718; b) L. Ban, M.
Mrksich, Angew. Chem. 2008, 120, 3444 – 3447; Angew. Chem.
Int. Ed. 2008, 47, 3396 – 3399; c) J. Li, P. S. Thiara, M. Mrksich,
Langmuir 2007, 23, 11826 – 11835; d) J. Su, M. Mrksich, Lang-
muir 2003, 19, 4867 – 4870.
Figure 3. a) WGA binding to a well functionalized with chitopentaose 9
(blue trace) and control on surface functionalized with 10 (green
trace). b) PAL binding to 9 (blue) and control with 10 (green). c) ECA
lectin binding fully galactosylated glycan 8. (blue) and control with
glycan 8 (green). d) Specific trapping of WGA lectin in presence of a
fourfold excess of ECA. (blue) and control for ECA binding on well
containing glycan 2 (green).
spond to the subunits of the Pa-4/PL-C and PL-D1/PL-D2
isoforms.
[3] Z. L. Zhi, N. Laurent, A. K. Powel, R. Karamanska, M. Fais, J.
Voglmeir, A. Wright, J. M. Blackburn, P. R. Crocker, D. A.
Russell, S. Flitsch, R. A. Field, J. E. Turnbull, ChemBioChem
2008, 9, 1568 – 1575.
[4] A. Sanchez, N.-C. Reichardt, unpublished results.
[5] T. R. Northen, J. C. Lee, L. Hoang, J. Raymond, D. R. Hwang,
S. M. Yannone, C. H. Wong, G. Siuzdak, Proc. Natl. Acad. Sci.
USA 2008, 105, 3678 – 3683.
[6] Y. Fang, A. G. Frutos, J. Lahiri, Langmuir 2003, 19, 1500 – 1505.
[7] a) A. H. Brockman, B. S. Dodd, N. Orlando, Anal. Chem. 1997,
69, 4716 – 4720; b) Y.-C. Tyan, M.-H. Yang, P.-C. Liao, J.-D. Liao,
S.-B. Jong, C.-Y. Liu, M.-C. Wang, M. Grunze, Int. J. Mass
Spectrom. 2007, 262, 67 – 72.
[8] M. I. Bethencourt, L. O. Srisombat, P. Chinwangso, T. R. Lee,
Langmuir 2009, 25, 1265 – 1271.
[9] W. E. Wallace, M. A. Arnould, R. Knochenmuss, Int. J. Mass
Spectrom. 2005, 242, 13 – 22.
[10] S. Serna, J. Etxebarria, N. Ruiz, M. Martin-Lomas, N.-C.
Reichardt, Chem. Eur. J. 2010, 16, 13163 – 13175.
When a galactose-binding lectin from Erythrina cristagalli
(ECA) was studied on immobilized fully galactosylated
glycan 8, broad peaks for the single-charged (27825 Da) and
double-charged (14401 Da and 13443 Da) a and b subunits
were observed. However, no peaks corresponding to ECA
were observed in a control experiment with maltopentaose
10.
To assess the ability of our platform to specifically trap
and identify a lectin from a mixture of proteins, wells with
immobilized chitobiose 12 were incubated with a 4:1 ECA/
WGA mixture to compensate for the differences in ionization
efficiency. MALDI-TOF analysis showed strong peaks for
single- and double-charged WGA at 17160 kDa and
8705 kDa, respectively. Immobilized maltopentaose 10
showed no binding to the lectin mix, thus demonstrating the
1804 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 1801 –1804