Sugar-decorated hydroxyapatite: An inorganic material bioactivated with carbohydrates

Article abstract of DOI:10.1016/j.carres.2011.04.044

An efficient method for the direct and covalent decoration of granules of nanostructured apatite with a sample monosaccharide is presented; the hydroxyapatite material was directly functionalised with a short azido-containing spacer arm, to which α-propargyl glucopyranoside has been chemoselectively ligated by Huisgen-type cycloaddition. The 'glycosylated' hydroxypatite was characterised by its ability to interact with glucose recognising lectins.

Full text of DOI:10.1016/j.carres.2011.04.044

Carbohydrate Research 346 (2011) 1564–1568
Contents lists available at ScienceDirect
Carbohydrate Research
journal homepage: www.elsevier.com/locate/carres
Sugar-decorated hydroxyapatite: an inorganic material bioactivated
with carbohydrates
Laura Russo a, Elena Landi b, Anna Tampieri b, Antonino Natalello a, Silvia M. Doglia a, Luca Gabrielli a,
a,
Laura Cipolla a, , Francesco Nicotra
⇑
⇑
a Department of Biotechnology and Biosciences, University of Milano-Bicocca, P.za della Scienza 2, 20126 Milano, Italy
b Institute of Science and Technology for Ceramics, National Research Council, via Granarolo 64, 48018 Faenza, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
An efficient method for the direct and covalent decoration of granules of nanostructured apatite with a
sample monosaccharide is presented; the hydroxyapatite material was directly functionalised with a
Received 1 February 2011
Received in revised form 22 April 2011
Accepted 27 April 2011
short azido-containing spacer arm, to which
a-propargyl glucopyranoside has been chemoselectively
ligated by Huisgen-type cycloaddition. The ‘glycosylated’ hydroxypatite was characterised by its ability
to interact with glucose recognising lectins.
Available online 3 May 2011
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
Propargyl glycosides
Click chemistry
Hydroxyapatite
Biomaterials
Lectin
1. Introduction
ics, such as hydroxyapatite (HA), have been widely investigated
due to their biological and chemical similarity to the inorganic
All known living organisms present a collection of free or cova-
lently linked carbohydrates usually referred to as glycans.1,2 Some
of these complex structures decorate the surface of cells in the
form of glycoconjugates, while others are secreted into the extra-
cellular matrix (ECM); in both cases, they mediate an extremely
wide variety of processes required for life including structural sup-
port, protection, recognition, localisation, and information/nutrient
transfer.
The extracellular matrix is a coacervate of glycosaminoglycans
(GAGs) and proteins with various mechanical and signalling func-
tions. For example, heparan sulphate (HS), a significant sulphated
GAG, binds to several proteins of biological relevance3 having a
considerable functional role in a wide array of physiological pro-
cesses,4 including cell proliferation and migration, as well as angi-
ogenesis, wound healing, and inflammatory responses.5
Tissue engineering is a promising branch for the treatment of a
broad variety of diseases by providing an alternative for donor
tissues and organs by artificial tissue substitutes. Since cells
isolated from host tissues may not have the ability to assemble into
functional tissue, scaffold materials must be designed in order to
support cell attachment and provide necessary signals for cell
growth and organisation. For bone tissue regeneration, bioceram-
phases of bones and teeth.6,7 However, despite its proven osteo-
conductivity, stoichiometric HA does not possess any specific bio-
activity and is thus unable to cross-talk with cells.
A key point for bone tissue regeneration is the effective inter-
play and integration between the inorganic matrix, that is the syn-
thetic hydroxyapatite (bone surrogate) and the organic matrix, the
ECM. For this reason, the upgrading of this inorganic material by
the addition of biological cues, ameliorating specific cell-biomate-
rial interactions and eventually triggering the desired biological re-
sponse appear extremely interesting.8 Immobilisation of
biomolecules on material surfaces enables both localisation and
retention of molecules at the cell–biomaterial interface.
Thus, any factors favouring positive interactions between
hydroxyapatite and ECM in general, and carbohydrates in particu-
lar, are highly desirable. Based on these considerations, as a proof
of principle, in the present work, we wish to report an efficient
method for the direct and covalent decoration of granules of nano-
structured apatite with a sample monosaccharide, in order to open
the way to innovative biomaterials for bone tissue regeneration.9,10
2. Results and discussion
The effective immobilisation of biomolecules on the surface of
bioceramics is still challenging, due to the requirement of a
sufficiently strong and specific affinity with the material surface,
and of site-directed immobilisation, and finally of maintenance of
⇑
Corresponding authors. Tel.: +39 02 64482152; fax: +39 02 64483565 (L.C.).
E-mail addresses: laura.cipolla@unimib.it (L. Cipolla), Francesco.nicotra@
unimib.it (F. Nicotra).
0008-6215/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2011.04.044

L. Russo et al. / Carbohydrate Research 346 (2011) 1564–1568
1565
the biological activity. Most immobilisation methods developed to
date involve modification or coating of the inorganic material with
appropriate substances in order to immobilise the proteins by
physical adsorption via van der Waals, hydrophobic or electrostatic
forces, or chemical bonding.11
Covalent linkage of bioactive molecules to material surface is a
valid alternative strategy in order to allow a sufficiently strong and
specific immobilisation of biomolecules with the surface itself; in
addition covalent bonding may permit site-directed immobilisa-
tion and preservation of specific conformation and exposition to
control biological responses.
Figure 1. (A) HA granules functionalised with the dansyl group; (B) HA granules as
negative control.
Thus, we envisaged the possibility of taking advantage of the
hydroxyl groups of hydroxyapatite granules having dimensions
in the range 400–600 microns12 for the covalent linkage of a sam-
ple monosaccharide through a chemoselective ligation, via the azi-
do-alkyne Huisgen cycloaddition, as illustrated in Scheme 1. As a
preliminary test to assess the reactivity of hydroxyl groups of
hydroxyapatite, a simple reaction between HA and dansyl chloride
was performed (Scheme 1).
Nanostructured biomimetic powder of carbonate substituted
apatite (CHA) synthesised by a wet chemical route, was used to
prepare granules of materials, having size in the range 400–600
microns: the powder particles were agglomerated using a wet
based process avoiding any additions of chemicals; then after dry-
ing at 80 °C the granules have been sieved and the specified frac-
tion was collected. The granules are crystallographically pure: no
secondary crystalline phases are detected by XRD besides nano-
cristalline apatite (data not shown). The carbonate ions are substi-
tuting in the phosphate site of hydroxyapatite (B-type CHA) as
determined by Fourier transform infrared (FTIR) spectroscopy
(data not shown), in amount to 6 wt % (determined by TGA). The
carbonation is thus in the range of the contents of the biological
apatite.
In order to assess the reactivity of HA, reaction with dansyl
chloride was performed suspending HA granules in dry THF
(2 mL) containing 0.25 M dansyl chloride and 5% triethylamine,
and kept under stirring for 4 h (Scheme 1A).
Figure 2. IR analysis of HA granules functionalised with the dansyl group (HA-
dansyl) compared to unfunctionalised HA (HA).
The dansyl functionalisation of HA was confirmed by UV analy-
sis at 254 nm (Fig. 1), and FTIR spectroscopy in attenuated total
reflection (ATR) mode (Fig. 2).
All the spectra reported in Figure 2 are dominated by the
1017 cmÀ1 absorption peak due to the phosphate (PO43À) vibration
of HA. Other absorption bands of HA are observed at 3570 cmÀ1
(stretching of structural OH), couple at 1455 and 1415 cmÀ1
(stretching of B-type carbonate), 962 cmÀ1 (phosphate), and at
872 cmÀ1 (bending of B-type carbonate). The spectrum of HA func-
tionalised directly with dansyl chloride (HA-dansyl, Fig. 2) displays
three evident peaks at 2979, 2946 and 2880 cmÀ1 that are due to
the CH3 of the dansyl group. To better characterise the HA decora-
tion, we performed the second derivative analysis of the measured
spectra (a resolution enhancement mathematical procedure, Fig. 2
inset): in the case of HA-dansyl, the infrared response of sulfonic
ester leads to the well resolved peak at 1380 cmÀ1 (Fig. 2, HA-dan-
syl inset). All together, therefore, these results underline the suc-
cess of our hydroxyapatite biodecoration strategy.
With these results in our hands, we then proceeded for the
chemoselective ligation of the carbohydrate moiety. In order to en-
sure a minimal spacing between HA surface and the monosaccha-
ride we used a short PEG arm, possessing an azido group at one end
for the Huisgen cycloaddition, and a carboxy group on the other
side for the covalent bond to the HA hydroxyls. Hence, the trieth-
ylene glycol was desymmetrised first by monosubstitution with
the azido group via the monomesylate 2, (Scheme 2), as already de-
scribed by Bertozzi et al.,13 then oxidised at the remaining OH by
TEMPO, affording 4.14
O
S
OH
O
O
a)
HA
granules
A.
B.
HA
HA-dansyl
granules
HA
N
O
O
N3
OH
O
O
b)
HA
granules
HA
granules
HA-N3
c)
OH
O
HO
HO
N
HO
N
O
O
N
O
O
O
HA
granules
HA-Glc
Bioceramic functionalisation has been performed by esterifica-
tion reaction between the OH groups of hydroxyapatite and COOH
groups of the glycol, upon in situ activation of 4 by N,N-diisopro-
Scheme 1. Reagents and conditions: (A) (a) dansyl chloride, THF, triethylamine,
4 h; (B) (b) activated 4, dry THF, 12 h; (c)
a-propargyl glucoside, CuSO4Á5H2O,
sodium ascorbate, H2O, 24 h.

1566
L. Russo et al. / Carbohydrate Research 346 (2011) 1564–1568
b)
O
OH
O
OMs
a)
O
O
N3
HO
HO
HO
2
1
3
c)
O
N3
HO
4
Scheme 2. Preparation of the difunctional PEG linker. Reagents and conditions: (a) MsCl, Et3N dry Et2O; (b) NaN3, EtOH; (c) NaHCO3 (0.5 M in H2O), KBr (0.5 M inH2O), TEMPO
(0.1 M in CH3CN) NaOCl (0.35 M).
pylcarbodiimide (DIC), in dry THF (Scheme 2). Finally, the HA was
added to the THF solution of activated 4, and the suspension stirred
for 12 h at room temperature. After that, the HA was filtered and
washed vigorously with THF, MilliQ water and acetone. In order
to avoid the non-specific adsorption of PEG on the HA surface,
the samples were soaked in MilliQ water for 1 h at room tempera-
ture and then washed again, with MilliQ water and acetone.
The obtained HA-N3 was submitted to Huisgen cycloaddition
with an alkyne. The Huisgen cycloaddition,15 that is the 1,3-dipolar
cycloaddition between azides and alkynes yielding irreversibly a
stable 1,4-disubstituted 1,2,3-triazole offers several advantages
for the biodecoration step of hydroxyapatite: (i) the cycloaddition
reaction is characterised by high versatility as well as high specific-
ity and chemoselectivity in the presence of a wide variety of sur-
rounding functional groups and solvents (ranging from both
protic and aprotic organic solvent to water); (ii) these reactions
are featured by a high thermodynamic driving force (generally
above 20 kcal molÀ1), thus allowing the use of mild reaction condi-
tions and short reaction times, leading to quantitative coupling
yields at room temperature;16 (iii) it guarantees the desired spatial
orientation of the biomolecule on the material surfaces. Cu(I)-cat-
alysed Huisgen-type cycloadditions have been employed in a vari-
ety of applications in the field of material science,17,18 such as the
preparation of highly functionalised macromolecules19 and the
surface immobilisation of (bio)macromolecules and polymers.20,21
However, to the best of our knowledge, no examples of Huisgen
cycloaddition have been employed to date for the ‘biodecoration’
of bioceramics. As the alkyne partner, in order to give also biolog-
Figure 3. ATR/FTIR spectra of hydroxyapatite samples: plain HA (dotted line), of
HA-Glc (dashed line), and of HA-Glc incubated with lectin (full line) are reported
after normalisation at the B-type carbonate stretching peak at ꢀ1455 cmÀ1. The
absorption spectrum of free lectin (dashed-dotted line) is also reported for
comparison. The second derivatives of the above absorption spectra are reported
in the inset.
ical relevance to the method, propargyl
used as model substrate.
a-D-glucopyranoside was
Hence, the propargyl
a-glucoside was prepared by straightfor-
ward Fischer glycosylation catalysed by H2SO4 supported on silica
gel and ultrasounds on the unprotected monosaccharide22 and
coupled to the azido group by chemoselective click chemistry.
Click reaction was performed using 0.03 M stock solutions of
hydroxyapatite leads to an increased absorption in the 1700–
1600 cmÀ1 region, where the protein Amide I band occurs. To better
resolve the protein contribution, the second derivatives of the
absorption spectra were performed. In particular, the two new peaks
at ꢀ1624 and at ꢀ1694 cmÀ1, observed in HA-Glc only after incuba-
tion with lectin, can be assigned to the protein b-sheet secondary
structures, in agreement with the FTIR spectrum of the free lectin.23
propargyl 2-
a-D
-glucopyranoside, CuSO4Á5H2O and sodium ascor-
bate in milliQ water. The HA-N3 was suspended in the saccharide
stock solution (2 mL), and a mixture of cupric sulfate solution
(0.100 mL, 0.003 mmol, 5% in respect to the saccharide), and ascor-
bate solution (0.300 mL, 0.009 mmol, 15 mol % in respect to the
saccharide) was added and stirred for 24 h.
3. Materials and methods
The ‘glycosylated’ carbonated hydroxypatite HA-Glc was charac-
terised by its ability to interact with glucose recognising lectins, that
is, Concanavalin A. While usually lectins can be easily detected
by fluorescence methods, in the present case this technique had
to be discarded. In fact HA itself is emitting in wide range of
wavelengths, over-imposing theemissionsignalsoffluorescentlyla-
belled lectins. Thus, selective binding of Concanavalin A was con-
firmed for the ‘glucosylated’ hydroxyapatite by FTIR spectroscopy
and compared to the unfunctionalised material (Fig. 3). In Figure 3
the ATR/FTIR spectra of hydroxyapatite HA, HA-Glc, and HA-Glc
incubated with lectin were compared with that of the free lectin,
in the 1450–1750 cmÀ1 spectral range. The functionalisation of
General: Chemicals were purchased by Sigma-Aldrich or Merck
and used without any further purification. Concanavalin A was
purchased by Invitrogen. Solvents were dried over molecular
sieves, for at least 24 h prior to use. When dry conditions were re-
quired, the reaction was performed under Ar atmosphere.
3.1. Carbonate hydroxyapatite granules
Carbonate substituted apatite powder with a specific surface
area value of 36 m2/g, has been synthesised through a wet synthesis
based on drop-wise addition of
a phosphoric acid solution

L. Russo et al. / Carbohydrate Research 346 (2011) 1564–1568
1567
containing 88.8 g H3PO4 (Aldrich, 85 wt % pure) in 600 ml of
3.4. Lectin binding assay
distilled water and, simultaneously, of a sodium hydrogencarbon-
ate solution containing 13.3 g NaHCO3 (Merck-Shuchardt, 99.7%
pure) in 400 ml of distilled water to a stirred calcium hydroxide dis-
persion containing 100 g Ca(OH)2 (Aldrich, 95% pure) in 700 ml of
distilled water. The amounts of reagents were chosen in order to re-
spect the Ca/P molar ratio 1.67 of the stoichiometric hydroxyapatite
but, for the presence of sodium hydrogencarbonate, a competition
for entering in the apatitic cell structure between phosphate and
carbonate groups was set up.
Glucosylated HA (HA-Glc) was exposed to a solution containing
lectin (20
g mLÀ1) in phosphate-buffered saline (PBS) for 60 min
at room temperature followed by successive rinses in PBS, phos-
phate-buffered saline diluted to 50% v/v with deionised water,
and twice with deionised water.25
l
3.5. FTIR characterisation
During the reaction process (taking 3–4 h), the temperature was
kept at 40 °C under mechanical stirring. Once the reaction has been
completed, the suspension was maintained at the same temperature
of40 °C andstirred for24 h and, finally, aged for 24 h. The precipitate
was washed three times with deionised water, freeze-dried and
FTIR spectra of the samples were collected in attenuated total
reflection (ATR) using a diamond single reflection device (Golden
Gate, Specac, USA). Indeed, solid samples can be better analyzed
in ATR mode when they can be placed in close contact with the
ATR element. In particular diamond element is the preferred choice
for most applications on solid samples, due to its robustness and its
high penetration depth that leads to a high sensitivity. On the other
finally sieved at 150 lm.
The obtained powder has been processed into granules having
dimensions in the range 400–600 microns, to obtain a material
usable as bone filler for dental and orthopaedic applications (sim-
ilar granules made of stoichiometric HA are already commercia-
lised by Finceramica Faenza Spa, Italy).
The ICP analysis gave a Ca/P molar ratio of 1.87, which is higher
than the stoichiometric value (1.67), that proves that carbonate
ions entered the HA lattice replacing phosphate ions leading to
the so-called B-type carbonation. In fact A-type carbonation
(occurring when carbonate ions replace hydroxyl ions) does not
change Ca/P molar ratio in comparison to stoichiometric hydroxy-
apatite.24 At higher temperatures carbonate ions decompose
causing a weight loss due to CO2 elimination, that allows to esti-
mate the starting carbonation of the HA granules as about
5.5 wt %. This value, that is in agreement with those indirectly
found by the other analysis, is in the range of the contents of the
biological apatite (2–8 wt %).
side, diamond is, unfortunately, not transparent around 2100 cmÀ1
,
where the azido group absorbs. However, even if the choice of dia-
mond made our ATR device unsuitable for azido group detection, its
high sensitivity in the Amide I spectral region enabled us to study
lectin binding, thanks to the IR response of the protein that is spe-
cific of its backbone conformations.
The Varian 670-IR (Varian Australia Pty Ltd, Mulgrave VIC,
Australia) spectrometer—equipped with a nitrogen cooled mercury
cadmium telluride detector and an air purging system—was
employed under the following conditions: 25 kHz scan speed,
2 cmÀ1 spectral resolution, 512 scan co-additions, and triangular
apodisation. The second derivatives of the absorption spectra were
calculated following the Savitsky–Golay procedure (5 points), after
11 points of binomial smoothing of the spectra, using the GRAMS/
32 software (Galactic Industries Corporation, Salem, NH, USA).
Negative peaks in the second derivative correspond to maxima of
the measured absorption spectra.
3.2. Synthesis
4. Conclusions
Difunctionalised triethylene glycol was synthesised according
to Bertozzi et al.13
This work presents a novel method for the ‘biodecoration’ of
hydroxyapatite. The conjugation step, mediated by the Huisgen
cycloaddition allows the chemoselective ligation of a model mono-
saccharide to an inorganic biomimetic material; the method herein
described can be extended to chemically defined short fragment of
ECM polysaccharides (i.e., hyaluronic acid or heparan sulphate),
thus improving the bioactivity of hydroxyapatite-based biomateri-
als for bone tissue regeneration.
The propargyl a-glucoside was obtained in one step following a
recently published procedure on fully deprotected glucose.22 Acti-
vation of 4 (79 mg, 0.42 mmol) was performed by DIC (185 mg,
1.47 mmol) in dry THF (1.23 mL) for 2 h. The solution containing
activated 4 was used directly for HA functionalisation.
3.3. HA decoration
Hydroxyapatite granules (0.2 g) were suspended in the solution
containing activated 4 in dry THF and kept stirring for 12 h at room
temperature; the quantity of 4 was calculated in order to have
0.24 mol of reactant per 100 g of HA granules.12 Finally, the solvent
was removed by filtration and HA washed vigorously with THF,
MilliQ water and acetone.
Acknowledgements
We gratefully acknowledge MIUR, under project FIRB
RBPO68JL9 and FONDAZIONE CARIPLO, project 2008/3175 for
financial support.
Click reaction was performed using 0.03 M stock solutions of
propargyl
a-
D-glucopyranoside, CuSO4Á5H2O and sodium ascorbate
References
in milliQ water. The cupric sulfate solution (0.100 mL, 0.003 mmol,
5% in respect to the saccharide) and the ascorbate solution
(0.300 mL, 0.009 mmol, 15 mol % in respect to the saccharide) were
pre-mixed and stirred until obtaining yellow solution. The yellow
solution was then added to the HA-N3 suspended in the saccharide
solution (2 mL, 0.06 mmol). The reaction was stirred at room tem-
perature for 24 h, then the solvents removed by filtration and the
material washed vigorously with MilliQ water and acetone. In or-
der to avoid the non-specific adsorption of 4 on the HA surface,
the samples were soaked in MilliQ water for 1 h at room tempera-
ture and then washed again, with MilliQ water and acetone, and fi-
nally dried at room temperature.
1. Werz, D. B.; Ranzinger, R.; Herget, S.; Adibekian, A.; von der Lieth, C.-W.;
Seeberger, P. H. ACS Chem. Biol. 2007, 2, 685–691.
2. Ohtsubo, K.; Marth, J. D. Cell 2006, 126, 855–867.
3. Shipp, E. L.; Hsieh-Wilson, L. C. Chem. Biol. 2007, 14, 195–208.
4. Linhardt, R. J.; Kerns, R. J.; Vlahov, I. R. Biomedical Functions and Biotechnology of
Natural and Artificial Polymers; ATL Press, 1996. pp 46–62.
5. Christopher, R.; Parish, C. R. Nat. Rev. Immunol. 2006, 6, 633–643.
6. Hench, L. L. J. Am. Ceram. Soc. 1991, 74, 1487–1510.
7. Hench, L. L.; Wilson, J. In An Introduction to Bioceramics; World Scientific:
Singapore, 1993; Vol. 1,
8. Sawyer, A. A.; Weeks, D. M.; Kelpke, S. S.; McCracken, M. S.; Bellis, S. L.
Biomaterials 2005, 26, 7046–7056.
9. Hench, L. L.; Polak, J. M. Science 2002, 295, 1014–1017.
10. Vallet-Regí, M. Chem. Eng. J. 2007, 137, 1–3.

1568
L. Russo et al. / Carbohydrate Research 346 (2011) 1564–1568
11. Zurlinden, K.; Laub, M.; Jennissen, H. P. Mat wiss u Werkstoff tech 2005, 36, 820–
827.
20. Cummins, D.; Duxbury, C. J.; Quaedflieg, P. J. L. M.; Magusin, P. C. M. M.; Koning,
C. E.; Heise, A. Soft Matter 2009, 5, 804–811.
12. Baeza, A.; Izquierdo-Barba, I.; Vallet-Regí, M. Acta Biomater. 2010, 6, 743–749.
13. Bertozzi, C. R.; Bednarski, M. D. J. Org. Chem. 1991, 56, 4326–4329.
14. Bragd, P. L.; van Bekkum, H.; Besemer, A. C. Top. Catal. 2004, 27, 49–66.
15. Moses, J. E.; Moorhouse, A. D. Chem. Soc. Rev. 2007, 36, 1249–1262.
16. Tornoe, C. W.; Christensen, C.; Meldal, M. J. Org. Chem. 2002, 67, 3057–3064.
17. Nandivada, H.; Jiang, X.; Lahann, J. Adv. Mater. 2007, 19, 2197–2208.
18. Fournier, D.; Hoogenboom, R.; Schubert, U. S. Chem. Soc. Rev. 2007, 36, 1369–
1380.
21. Dirks, A. J.; Cornelissen, J. J. L. M.; van Delft, F. L.; van Hest, J. C. M.; Nolte, R. J.
M.; Rowan, A. E.; Rutjes, F. P. J. T. QSAR Comb. Sci. 2007, 26, 1200–1210.
22. Shaikh, N.; Russo, L.; Cipolla, L.; Nicotra, F. Mol. Diver. 2010. doi:10.1007/
s11030-010-9281-9282.
23. Susi, H.; Byler, D. M. Methods Enzymol. 1986, 130, 290–311.
24. Le Geros, R. Z. Calcium Phosphates In Oral Biology and Medicine In Monographs
In Oral Science; Myers Karger, H., Ed.; AG publisher: Basel, 1991; Vol. 15, pp 82–
107.
19. Damiron, D.; Desorme, M.; Ostaci, R. V.; Akhrass, S.; Hamaide, T.;
Drockenmuller, E. J. Polym. Sci., Part A: Polym. Chem. 2009, 47, 3803–3813.
25. Harris, L. G.; Schofield, W. C. E.; Doores, K. J.; Davis, B. G.; Badyal, J. P. S. J. Am.
Chem. Soc. 2009, 131, 7755–7761.