A quartz crystal microbalance method to study the terminal functionalization of glycosaminoglycans

Article abstract of DOI:10.1039/c4cc06905f

We demonstrate the quartz crystal microbalance as a novel method to quantify the reaction yields and stability of the terminal conjugation of chemically complex molecules. Oxime ligation is identified as a facile, broadly applicable method for the reducing-end conjugation of glycosaminoglycans that overcomes the limited stability and yield of popular hydrazone ligation.

Full text of DOI:10.1039/c4cc06905f

ChemComm
View Article Online
View Journal
COMMUNICATION
A quartz crystal microbalance method to study the
terminal functionalization of glycosaminoglycans†
Cite this:DOI: 10.1039/c4cc06905f
ab
ab
ab
cde
Dhruv Thakar, Elisa Migliorini, Liliane Coche-Guerente, Rabia Sadir,
cde
ab
ab
ab
Received 1st September 2014,
Accepted 16th October 2014
Hugues Lortat-Jacob,
Didier Boturyn, Olivier Renaudet, Pierre Labbe and
abfg
Ralf P. Richter*
DOI: 10.1039/c4cc06905f
www.rsc.org/chemcomm
We demonstrate the quartz crystal microbalance as a novel method to
Here, we demonstrate that quartz crystal microbalance with
quantify the reaction yields and stability of the terminal conjugation of dissipation monitoring (QCM-D), a surface-sensitive technique
chemically complex molecules. Oxime ligation is identified as a popular for biosensing applications, enables quantitative analysis
facile, broadly applicable method for the reducing-end conjugation of conjugation yields and stability with a few micrograms of GAGs
of glycosaminoglycans that overcomes the limited stability and yield of arbitrary complexity. To this end, we compared the biotinylation
of popular hydrazone ligation.
of GAGs via two different chemoselective ligation chemistries
7
(
Scheme 1A). Hydrazone ligation has become the most frequently
1,8–10,11
Linear polysaccharides known as glycosaminoglycans (GAGs) are
ubiquitous cell surface and extracellular matrix components and
fulfill crucial biological functions. Advanced screening applications
used strategy for GAG functionalization.
Oxime ligation (for
selected references, see ref. 12), on the other hand, has only rarely
13
been applied to polysaccharides, and to our knowledge not to GAGs.
1
,2
(
e.g. glycan microarrays ), functional molecular and cellular
3
4
assays, as well as biosensors and biomaterials require the
attachment of GAGs to surfaces or other scaffolds (e.g. with a
biotin that can be anchored to biotin-binding proteins). In this
regard, site-specific conjugation through the reducing end is
desirable, as it effectively mimics the cell surface presentation of
GAG motifs and avoids alteration of GAG–protein interactions by
chemical modifications along the GAG chain, or by surface-
5
imposed conformational or spatial constraints.
An important but still underestimated challenge with the
conjugation of GAG-derived oligosaccharides, and in particular
polymeric GAGs, is the characterization of the reaction products.
The often low isolated yields from natural sources, the GAG’s
hydrophilic nature and lack of a suitable chromophore, and the
GAG’s acidity, fragility, polydispersity and heterogeneous sulfation
make their characterization not readily amenable by NMR, reverse-
6
phase HPLC and mass spectrometry, respectively.
a
Universit ´e Grenoble Alpes, DCM, 38000 Grenoble, France
b
CNRS, DCM, 38000 Grenoble, France
c
Universit ´e Grenoble Alpes, Institut de Biologie Structurale (IBS), 38027 Grenoble,
France
d
CNRS, IBS, 38027 Grenoble, France
Scheme 1 (A) Strategies adopted for the biotinylation of GAGs at their
reducing end, exemplified with a selected HS. Conditions: (a) 1 (0.33 mM), 4
(10 mM), PBS, pH 7.4, RT, 48 h; (b) 1 (4 mM), 5 (3.4 mM), aniline (100 mM),
acetate buffer (100 mM), pH 4.5, 37 1C, 48 h. (B) Library of biotin conjugates
of the GAGs HS and HA of various sizes, prepared using oxime ligation. HS
consists of GlcA b(1 - 4) GlcNAc a(1 - 4) disaccharides; X can be either H
e
CEA, DSV, IBS, 38027 Grenoble, France
f
CIC biomaGUNE, Paseo Miramon 182, 20009 Donostia - San Sebastian, Spain.
E-mail: rrichter@cicbiomagune.es; Tel: +34 943 00 53 29
g
Max Planck Institute for Intelligent Systems, 70569 Stuttgart, Germany
†
Electronic supplementary information (ESI) available: Supplementary materials
and methods, synthesis of a precursor, supplementary figures and references. See
DOI: 10.1039/c4cc06905f
or SO
3 3
H; Y can be either Ac or SO H. HA is unsulfated and consists of GlcA
b(1 - 3) GlcNAc b(1 - 4) disaccharides. R = remaining GAG chain.
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun.

View Article Online
Communication
ChemComm
We show that conventional hydrazone ligation is inefficient, in
particular for long GAGs, due to a low yield, and confirm that
14,15
hydrazone conjugates are unstable in an aqueous environment.
In contrast, oxime ligation emerges as a facile, rapid and efficient
method that provides conjugates with higher stability and can be
broadly applied, i.e. for different GAG types and for oligosaccharides
as well as polymeric GAGs of high molecular weight.
The reducing terminus of GAGs can selectively react with
biotin derivatives presenting aminooxy or hydrazide groups to
provide oxime and hydrazone linkage, respectively. For oxime
ligation, the reaction was successful only by using aniline as a
1
6
catalyst as described previously. To reproduce a commonly
1
,8,10,11
reported protocol,
we did not use aniline for hydrazone
ligation, although this compound is also known to enhance this
1
7
reaction. We created a library of biotin-conjugates through
oxime ligation, containing GAGs of different types and defined
chain lengths (Scheme 1B): hyaluronic acid (HA; oligosaccharides
with 2 and 5 disaccharide units (dp4, dp10), and a 360 kDa
polysaccharide) and heparan sulfate (HS; dp6, dp8, dp10, dp12,
and a 12 kDa polysaccharide).
Fig. 1 Binding assay to study the yield and stability of hydrazone and oxime
conjugation. (A, B) QCM-D frequency shifts, Df, obtained for the specific
binding of 2 (A) and 3 (B), stored at 4 1C for variable times (as indicated), on
For b-HA_dp4 10, mass analysis (Fig. S1, ESI†) demonstrated
that one biotin molecule is attached per HA while NMR analysis
(Fig. S2, ESI†) confirmed that the biotin is attached at the reducing streptavidin-coated surfaces (schematically shown in the inset). Samples
end, forming E and Z oxime isomers in a 73/27 ratio. Notably, these were incubated for 10 min (from 2 to 12 min) at a total GAG concentration of
À1
5
0 mg mL , which was followed by exposure to buffer solution. (C, D)
conventional characterization methods did not provide useful
information on any of the other compounds produced.
Parametric plots of QCM-D dissipation shifts, DD (not previously shown) vs.
Df (from A and B, respectively). The plots inform about the evolution of the
HS film mechanical properties and morphology with increasing surface
To characterize the conjugation of more complex GAGs, we
exploited the high affinity of the streptavidin (SAv)–biotin coverage. All curves had comparable shapes, indicating that all compounds
8
,18
interaction,
and followed the binding of GAGs to surfaces tested generate comparable film morphologies at a given surface coverage.
displaying a SAv monolayer (Fig. 1B, inset) by QCM-D (Fig. S3,
ESI†). Fig. 1A and B illustrates the differences in the stability of
biotinylated HS (b-HS made from 1, B22 disaccharides) prepared including the dissipation shift DD, in terms of so-called DD vs.
through hydrazone and oxime ligation, 2 and 3, respectively, as a Df curves (Fig. 1C and D) revealed invariant curve shapes,
function of storage time (up to two months) at 4 1C. The binding of indicating that the film morphology (at a given surface coverage)
b-HS samples was evaluated using two different parameters: the was comparable for all compounds tested. A plausible explanation
QCM-D frequency shift at saturation (Dfsat; proportional to the areal is that the composition of the film in terms of the HS molecular
mass density of immobilized b-HS (Fig. S4, ESI†)) and the maximal conformation or size distribution is not altered by the degradation
binding rate (Df/Dt; proportional to the concentration of b-HS in or the type of conjugation. This implies that degradation of b-HS
the sample solution (Fig. S5A, ESI†)). The magnitudes of Dfsat and occurs exclusively through cleavage of the bond that links biotin to
Df/Dt decreased appreciably with increasing storage time for b-HS HS, which is due to the hydrolysis of hydrazone (and to a much
14
hydrazone (Fig. 1A), whereas only minor changes were observed for lesser extent oxime) under physiological conditions.
b-HS oxime (Fig. 1B). Moreover, the magnitudes of Df and Df/Dt We note in passing that the biotin released upon degradation
for 2 were lower than for 3, even when de-frozen aliquots were would occupy a fraction of the available binding site on any
used immediately. We argue that these effects are the result of biotin-capturing surface. This influences the surface density of
the release of biotin upon degradation of the conjugates, i.e. immobilized GAGs, and thus entails a limited reproducibility of
the hydrazone but not the oxime is already appreciably degraded surface functionalization. This shortcoming needs to be considered,
immediately after purification, and further degrades upon storage for example, in solid-phase molecular binding assays such as by
11,19
at 4 1C.
To understand this, we note that the QCM-D responses in
surface plasmon resonance.
For a quantitative analysis of reaction yields and degradation
Fig. 1 are exclusively due to binding of intact b-HS. Biotin-free rates, we determined steady-state binding rates (Fig. 2A and Fig. S5A,
HS did not bind (Fig. S3, ESI†) and free biotin (produced as the ESI†) through linear fits to appropriate portions of the binding
result of degradation) did not by itself give rise to a measurable curves in Fig. 1A and B. At steady state, binding is mass-transfer
signal (Fig. S6, ESI†). A decrease in the magnitude of Dfsat
,
limited, and with total HS concentrations and flow conditions
therefore, is consistent with partial occupancy of the surface remaining unchanged in our assay, any decrease in the binding
with b-HS and saturation of the remaining binding sites by free rate reflects a proportional decrease in the concentration of intact
2
0
biotin (Fig. S7, ESI†). Further analysis of the QCM-D responses, b-HS in the probed solution. From the binding rate of freshly
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2014

View Article Online
ChemComm
Communication
Fig. 2 (A) Quantification of active analyte concentrations from QCM-D
binding assays. The ratio of the slopes of the extended linear binding regimes
(black lines are linear fits), here shown by way of example for freshly de-frozen
b-HS hydrazone 2 (red line) and b-HS oxime 3 (blue line) (data from Fig. 1), is
proportional to the ratio of the analytes’ active concentrations. (B) Fraction of
intact b-HS as a function of storage time at 4 1C for 2 (red squares) and 3 (blue
circles), determined through comparison of binding rates (data from Fig. 1).
Error bars represent variations in the slopes observed for three independent
measurements under identical conditions. Freshly de-frozen b-HS oxime 3, for
which the active analyte concentration was determined independently (Fig. S5,
ESI†), was used as a reference.
Fig. 3 QCM-D binding assays for HS (A) and HA (B, C) of various sizes
(as indicated). All GAGs were conjugated with biotin via oxime ligation. Data
for 8 are also shown in (B) to facilitate comparison with 11. Incubation conditions
were as described in Fig. 1, except for 12, which was incubated from 5 to
1
10 min. (D–F) DD vs. Df curves corresponding to (A–C), respectively; the inset in
D) shows a part of the data at a higher magnification. The magnitude of the
slope of the DD vs. Df curves increases as a function of GAG size, indicating that
de-frozen 3 and through comparison with a reference molecule
of known concentration we could estimate the reaction yield of
oxime ligation (Fig. S5, ESI†); comparison of the binding rates
in Fig. 1 then yielded the fraction of intact b-HS in all probed
solutions (Fig. 2). The fraction of the biotinylated analyte in the
(
21
larger GAGs generate a softer film.
freshly de-frozen 3 was 54 Æ 8%. This value was confirmed by using conventional analytical techniques, in particular when
weighing 10 (which in contrast to larger GAGs could be readily the amount of sample is limited to a few micrograms. Moreover, we
separated from non-biotinylated GAGs; Fig. S1, ESI†). The have established oxime ligation as a facile, one-step method for the
activity of freshly defrozen 2 was almost 5-fold smaller than selective conjugation of GAGs at the reducing end. The method is
that of 3, indicating that reaction yields are considerably superior in yield and stability to the commonly used hydrazone
improved for oxime ligation (Fig. 2B). The 5-fold improvement ligation, and versatile in that it can be applied to GAGs of various
in yield was confirmed by dot-blot analysis (Fig. S8, ESI†). (most likely any) types and sizes. The methods should find broad
Moreover, the fraction of intact 2 decreased by another 5-fold over use, as tools in the glycosciences and in biotechnological applica-
6
0 days of storage at 4 1C, indicating substantial degradation, tions. In particular, the control over and stability of GAG conjugates
whereas 3 was only marginally degraded over the same time are crucial for the reliable preparation of GAG-functionalized
interval (Fig. 2B). surfaces and scaffolds for tissue engineering and fundamental
Concerning the GAGs of various chain lengths, a clear trend biological studies.
in the Df values at saturation (Fig. 3A for HS and Fig. 3B and C for We thank C. Travelet (Cermav, Grenoble, France) and
22
HA) and in the DD vs. Df curves (Fig. 3D–F) as a function of size G. Dubacheva (CIC biomaGUNE) for support with DLS and SE
confirmed that QCM-D curves are indeed sensitive to variations in measurements, respectively, E. Defrancq and N. Spinelli (DCM)
molecular weight. Differences in molecular weight, as small as one for oligonucleotide synthesis and R. Viv ´e s (IBS) for preparation
disaccharide for oligomeric GAGs, can be readily distinguished of HS oligomers. This work was supported by the Nanosciences
through DD vs. Df curves (Fig. 3D, inset). Thus, the DD vs. Df curves Foundation (Grenoble) through the Chair of Excellence Project
represent a useful tool that provides insight into sample composi- GAG2D (to R.P.R.) and a PhD scholarship (to D.T.), the NanoBio
tion. HA_dp10 8 bound to SAv monolayers with a response similar programme, the ICMG FR 2607 and the LabEx ARCANE (ANR-
although not identical to HS_dp10 11 (Fig. 3B and E). We propose 11-LABX-0003-01).
that the slightly stronger Df shift for 8 over 11 reflects an increased
film thickness, resulting from a stronger repulsion between the Notes and references
sulfated and thus more highly charged HS chains in 8. Notably,
1
E. Mercey, R. Sadir, E. Maillart, A. Roget, F. Baleux, H. Lortat-Jacob
and T. Livache, Anal. Chem., 2008, 80, 3476–3482.
even the long HA polymer 12 with B900 disaccharides could be
readily biotinylated (Fig. 3C), indicating that the polymer length
does not affect conjugation.
In summary, QCM-D together with a suitably functionalized
sensor surface has proven to be instrumental for the characteriza-
tion of conjugates made from chemically complex molecules such
as GAGs, providing information about reaction yields, sample
degradation and sample composition that is difficult to assess
2 W. Takada, M. Fukushima, P. Pothacharoen, P. Kongtawelert and
K. Sugahara, Anal. Biochem., 2013, 435, 123–130; E. Clo, O. Blixt and
K. J. Jensen, Eur. J. Org. Chem., 2010, 540–554; S. Park, J. C. Gildersleeve,
O. Blixt and I. Shin, Chem. Soc. Rev., 2013, 42, 4310–4326.
3 M. Morra, Biomacromolecules, 2005, 6, 1205–1223.
4
5
J. A. Burdick and G. D. Prestwich, Adv. Mater., 2011, 23, H41–56.
N. Altgarde, J. Becher, S. Moller, F. E. Weber, M. Schnabelrauch and
S. Svedhem, J. Colloid Interface Sci., 2013, 390, 258–266.
6 J. Zaia, Biomacromol. Mass Spectrom., 2005, 1, 3–36.
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun.

View Article Online
Communication
ChemComm
7
8
J. A. Prescher and C. R. Bertozzi, Nat. Chem. Biol., 2005, 1, 13–21.
N. Altgarde, E. Nileback, L. de Battice, I. Pashkuleva, R. L. Reis,
J. Becher, S. Moller, M. Schnabelrauch and S. Svedhem, Acta 15 A. A. Kale and V. P. Torchilin, Bioconjugate Chem., 2007, 18, 363–370;
14 J. Kalia and R. T. Raines, Angew. Chem., Int. Ed., 2008, 47,
7523–7526.
Biomater., 2013, 9, 8158–8166.
K. Godula and C. R. Bertozzi, J. Am. Chem. Soc., 2010, 132, 9963–9965.
B. Levrand, Y. Ruff, J.-M. Lehn and A. Herrmann, Chem. Commun.,
2006, 2965–2967.
9
1
0 H. Ichijo, N. Sugiura and K. Kimata, Polymers, 2013, 5, 254–268; 16 M. B. Thygesen, H. Munch, J. Sauer, E. Clo, M. R. Jorgensen,
T. Mori, T. Kodera, H. Yoshimine, Y. Kakuta, N. Sugiura, K. Kimata
and Y. Okahata, Chemistry, 2012, 18, 7388–7393.
O. Hindsgaul and K. J. Jensen, J. Org. Chem., 2010, 75, 1752–1755.
17 J. Y. Byeon, F. T. Limpoco and R. C. Bailey, Langmuir, 2010, 26,
15430–15435.
11 E. Saesen, S. Sarrazin, C. Laguri, R. Sadir, D. Maurin, A. Thomas,
A. Imberty and H. Lortat-Jacob, J. Am. Chem. Soc., 2013, 135, 9384–9390. 18 N. S. Baranova, E. Nileback, F. M. Haller, D. C. Briggs, S. Svedhem,
2 F. Peri, P. Dumy and M. Mutter, Tetrahedron, 1998, 54, 12269–12278; A. J. Day and R. P. Richter, J. Biol. Chem., 2011, 286, 25675–25686.
R. D. Goff and J. S. Thorson, Med. Chem. Commun., 2014, 5, 19 R. I. W. Osmond, W. C. Kett, S. E. Skett and D. R. Coombe, Anal.
1
1
036–1047; F. Lin, J. Zheng, J. Yu, J. Zhou and M. L. Becker,
Biochem., 2002, 310, 199–207.
20 W. T. Hermens, M. Benes, R. Richter and H. Speijer, Biotechnol.
Appl. Biochem., 2004, 39, 277–284.
Biomacromolecules, 2013, 14, 2857–2865.
3 T. J. Styslinger, N. Zhang, V. S. Bhatt, N. Pettit, A. F. Palmer and
1
P. G. Wang, J. Am. Chem. Soc., 2012, 134, 7507–7515; A. Richard, 21 I. Reviakine, D. Johannsmann and R. P. Richter, Anal. Chem., 2011,
A. Barras, A. B. Younes, N. Monfilliette-Dupont and P. Melnyk, 83, 8838–8848.
Bioconjugate Chem., 2008, 19, 1491–1495; T. T. Beaudette, J. A. 22 E. Migliorini, D. Thakar, R. Sadir, T. Pleiner, F. Baleux, H. Lortat-
Cohen, E. M. Bachelder, K. E. Broaders, J. L. Cohen, E. G. Engleman
and J. M. J. Frechet, J. Am. Chem. Soc., 2009, 131, 10360–10361.
Jacob, L. Coche-Guerente and R. P. Richter, Biomaterials, 2014, 35,
8903–8915.
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2014