Binding affinity of surface functionalized gold nanoparticles
Ryan D. Ross, Ryan K. Roeder
Department of Aerospace and Mechanical Engineering, Bioengineering Graduate Program, University of Notre Dame,
Notre Dame, Indiana 46556
Received 14 October 2010; revised 29 April 2011; accepted 11 May 2011
Published online 25 July 2011 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/jbm.a.33165
Abstract: Gold nanoparticles (Au NPs) have been investigated
for a number of biomedical applications, including drug and
gene delivery vehicles, thermal ablation therapy, diagnostic
sensors, and imaging contrast agents. Surface functionaliza-
tion with molecular groups exhibiting calcium affinity can ena-
ble targeted delivery of Au NPs to calcified tissue, including
damaged bone tissue. Therefore, the objective of this study
was to investigate the binding affinity of functionalized Au
NPs for targeted delivery to bone mineral, using hydroxyapa-
tite (HA) crystals as a synthetic analog in vitro. Au NPs were
synthesized to a mean particle size of 10–15 nm and surface
functionalized with either L-glutamic acid, 2-aminoethylphos-
phonic acid, or alendronate, which exhibit a primary amine for
binding gold oppositecarboxylate,
bisphosphonate groups, respectively, for targeting calcium.
Bisphosphonate functionalized Au NPs exhibited the most
rapid binding kinetics and greatest binding affinity to HA, fol-
lowed by glutamic acid and phosphonic acid. All functional
groups reached complete binding after 24 h. Equilibrium bind-
ing constants in de-ionized water, determined by nonlinear
regression of Langmuir isotherms, were 3.40, 0.69, and 0.25
mg/L for bisphosphonate, carboxylate, and phosphonate func-
tionalized Au NPs, respectively. Functionalized Au NPs exhib-
ited lower overall binding in fetal bovine serum compared to
de-ionized water, but relative differences between functional
groups were similar. V
C 2011 Wiley Periodicals, Inc. J Biomed Mater
Res Part A: 99A: 58–66, 2011.
Key Words: gold nanoparticles, hydroxyapatite, binding affin-
ity, targeted delivery, contrast agent, bisphosphonate
How to cite this article: Ross RD, Roeder RK. 2011. Binding affinity of surface functionalized gold nanoparticles to hydroxyapatite.
J Biomed Mater Res Part A 2011:99A:58–66.
Gold nanoparticles (Au NPs) have been investigated for a
number of biomedical applications,1,2including drug and
gene delivery vehicles,3,4thermal ablation therapies,4,5diag-
nostic sensors,6,7and imaging contrast agents.8–11The util-
ity of Au NPs in these applications is derived from multiple
factors including the ease of synthesizing monodispersed
nanoparticles, colloidal stability, biocompatibility, surface
plasmon resonance, and high X-ray attenuation. Moreover,
Au NPs are readily surface functionalized with molecular
groups for cellular targeting and delivery, as well as enhanc-
ing colloidal stability. Surface functionalization with molecu-
lar groups exhibiting calcium affinity could enable targeted
delivery of Au NPs to mineralized tissue, including fatigue
microdamage in bone tissue,12as well as calcifications in
soft tissue associated with cancer or musculoskeletal injury.
Microdamage accumulates in bone tissue during repetitive
loading in the form of microcracks and diffuse damage.13,14
The accumulation of microdamage has been implicated in clini-
cal fracture susceptibility, including fatigue fractures in active
individuals, such as military recruits, fragility fractures in the
elderly, and the effects of long-term antiresorptive treatments
for osteoporosis.13,14However, the role of microdamage in clin-
ical bone fragility remains poorly understood due, in part, to
limitations in available methods for detection and imaging.14
Current methods for imaging microdamage in bone tissue
are limited to histological sections, which are inherently two-
dimensional (2-D), destructive, invasive, and tedious.15The
first and most common method has been en bloc staining
with basic fuchsin in ethanol.15–17Basic fuchsin exhibits pro-
tein affinity, but primarily acts as a space-occupying dye.15
Therefore, microdamage must be distinguished from non-
specific staining under observation in transmitted light or
epifluorescent microscopy. More recently, fluorophores, such
as xylenol and calcein, which employ iminodiacetate moieties
for calcium-chelation, have been used as damage-specific
labels for epifluorescent microscopy.15,18,19Moreover, se-
quential labeling with fluorophores exhibiting differing bind-
ing affinity enabled the measurement of spatiotemporal varia-
tion in damage events, including crack propagation18,19and
modes of loading.20Additionally, calcium-specific photoin-
duced electrontransfer sensormoleculeshave been
Correspondence to: R. K. Roeder; e-mail: email@example.com
Contract grant sponsors: US Army Medical Research and Materiel Command (W81XWH-06-1-0196) though the Peer Reviewed Medical Research
C 2011 WILEY PERIODICALS, INC.
investigated for the ability to ‘‘switch-on’’ fluorescence upon
chelation to calcium ions in damaged tissue.21,22However,
both light and epifluorescent microscopy are inherently 2-D.
Three-dimensional (3-D) spatial information can only be
ascertained from serial sectioning,23,24which is destructive
and tedious. Laser scanning confocal microscopy has been
used to provide limited depth of field for 3-D imaging of dam-
age in histological sections labeled with fluorophores.25,26
X-ray computed tomography has been proposed for non-
destructive and 3-D imaging of damage using contrast
agents with higher X-ray attenuation than bone tissue. The
presence, spatial variation, and accumulation of microdam-
age in both cortical and trabecular bone specimens was
detected using micro-computed tomography (micro-CT) af-
ter staining tissue with a precipitated barium sulfate
(BaSO4) contrast agent.27–29While this technique enabled
non-destructive and 3-D detection of microdamage in vitro,
the precipitated BaSO4 stain was not damage-specific and
the staining solutions were not biocompatible. Iodinated
molecules were investigated as a damaged-specific X-ray
contrast agent,30,31but were only detected using micro-CT
when precipitated in powder form.
Functionalized Au NPs were recently investigated as a
targeted X-ray contrast agent for labeling microdamage in
bone tissue.12Gold exhibits greater X-ray attenuation than
other commonly used X-ray contrast agents containing bar-
ium and iodine.32Au NPs exhibited high X-ray attenuation
and biocompatibility as an intravascular X-ray contrast agent
in mice after intravenous administration.9–11Functionalized
Au NPs also exhibit relatively high water solubility and low
viscosity compared to iodinated molecular contrast agents.
Finally, gold surfaces are readily functionalized through
adsorption of thiols33–35and amines,36–39which can be
used to attach molecules which target calcium ions on bone
mineral crystals exposed on the surface of microcracks.
The objective of this study was to investigate the bind-
ing affinity of functionalized Au NPs for targeted delivery to
bone mineral, using hydroxyapatite crystals as a synthetic
analog in vitro. Au NPs were surface functionalized with ei-
ther L-glutamic acid, 2-aminoethylphosphonic acid, or alen-
dronate (Fig. 1). Glutamic acid is a naturally occurring
amino acid with known affinity for calcium phosphates.40,41
Moreover, glutamic acid functionalized Au NPs were previ-
ously shown to target damaged bone tissue.12However,
phosphonate and particularly bisphosphonate groups are
known to exhibit strong binding affinity to calcium phos-
Bisphosphonates, such as alendronate, are
widely used clinically to suppress bone resorption as a
pharmacological treatment for osteoporosis and other bone-
Gold nanoparticle synthesis and functionalization
Au NPs were synthesized to a mean particle diameter of
10–15 nm using the citrate reduction method.45Briefly, 0.1
g HAuCl4?3H2O (?99.9%, Aldrich) was added to 400 mL of
de-ionized (DI) water and the solution was boiled vigo-
rously while stirring. A 1% solution of trisodium citrate
dehydrate (ACS reagent, >99.0%, Sigma) was added to the
boiling solution at a mass ratio of 5:1 HAuCl4 to sodium
citrate and left boiling for an additional 20 min. The solu-
tion volume was then adjusted to a total of 500 mL using
DI water. The resulting Au NP solution had a gold concen-
tration of ?0.5 mM and a wine red color.
Au NPs were prepared for functionalization by first
removing excess citrate ions. 1.5 mL 2% polyvinyl alcohol
(PVA 10–98, MW¼ 61,000, Fluka) was added to 24 mL Au
NP solution, followed by 0.6 g ion exchange resin (Amberlite
MB-150, Sigma). The resulting solution was stirred over-
night and subsequently filtered (Grade 3, Whatman) to
remove spent ion exchange resin. Au NPs were surface func-
tionalized by adding 1 mL of a 0.01M solution of either
L-glutamic acid (?99.5%, Fluka), 2-aminoethylphosphonic
acid (99%, Aldrich), or alendronate sodium trihydrate
(?97%, Sigma) (Fig. 1). The solution was left to equilibrate
under mild stirring overnight. Excess functionalization mole-
cules were removed by dialysis (Spectra/Por, MWCO ¼
3500, Spectrum Laboratories) against DI water for a total of
3 days, changing the water solution at least twice daily.
The mean particle diameter and particle size distribu-
tions were measured before and after functionalization by
transmission electron microscopy (TEM, Hitachi H-600) at 75
kV accelerating voltage. TEM specimens were prepared by
dropping a solution of functionalized Au NPs onto carbon-
coated grids and evaporating the solvent. The particle diame-
ter and aspect ratio were characterized by the mean and
standard deviation from measurements on a total of 100 par-
ticles per functional group. Ultraviolet–visible (UV–vis) spec-
tra (Varian Cary 3 spectrophotometer) were collected before
and after functionalization to verify colloidal stability and rel-
ative concentrations of the particles within solution.
Surface functionalization of Au NPs was qualitatively
verified using diffuse reflectance infrared Fourier transform
FIGURE 1. Gold nanoparticles (Au NPs) were surface functionalized
L-glutamic acid, (b) 2-aminoethylphosphonic acid, or (c)
alendronate, which exhibit a primary amine for binding to gold surfaces
opposite (a) carboxylate, (b) phosphonate, or (c) bisphosphonate func-
tional groups, respectively, for binding to calcium on bone mineral crys-
tals exposed within damaged bone tissue. [Color figure can be viewed
in the online issue, which is available at wileyonlinelibrary.com.]
JOURNAL OF BIOMEDICAL MATERIALS RESEARCH A|OCT 2011 VOL 99A, ISSUE 159
spectroscopy (DRIFTS, Bruker Tensor 27) at 4 cm?1with
64 total scans. Particles were functionalized and dialyzed as
outlined above, except without the presence of PVA and
without the use of ion exchange resin prior to functionaliza-
tion and dialysis, in order to prevent PVA from interfering
with the detection of functional groups. After dialysis, func-
tionalized Au NPs were collected by evaporation at 40?C
and 25 in Hg below atmospheric pressure. The dried pow-
ders were diluted in KBr (IR grade, 99þ%, Acros Organics)
at a ratio of ?1:200 by mass. Fourier transform infrared
(FTIR) spectra were collected over frequencies of 4000–500
cm?1and normalized to the background spectra for KBr
alone. FTIR peaks used to identify functional groups on Au
NPs were determined by comparison to corresponding FTIR
spectra for stock solutions of the functional molecules.
The amount of the three functional molecules (Fig. 1)
adsorbed to Au NP surfaces was quantitatively measured
using electrospray ionization mass spectrometry (ESI-MS).
After the overnight functionalization step described above, 1
mL aliquots of the surface functionalized Au NP solution
were removed and centrifuged at ?24,000g for 30 min to
separate nanoparticles from the supernatant solution which
included unbound functional molecules. The concentration
of unbound functional molecules remaining in supernatant
solutions was measured using ESI-MS (Bruker micro-TOF-
QII) with calibration by serial dilution of functional mole-
cule solutions. Glutamic acid and phosphonic acid samples
were measured in positive ion mode, while bisphosphonate
samples were measured in negative ion mode, with a capil-
lary voltage of 63.8–4.5 kV, a nebulizer gas pressure of 0.4
bar, a dry gas flow rate of 4.0 L/min, and a continuous sam-
ple flow rate of 4 lL/min. Spectra were collected over a
range of 50–3000 m/z at a rate of 5000 s?1for a period of
5 min. The amount adsorbed to Au NP surfaces was meas-
ured in triplicate and reported as the mean (6 standard
deviation) in lmol functional group per mg gold.
Binding affinity of functionalized Au NPs
Two separate binding experiments were performed to investi-
gate the binding affinity of functionalized Au NPs to bone min-
eral, using hydroxyapatite (HA) crystals as a synthetic analog.
Calcium-deficient HA single crystal whiskers were synthesized
using thechelate decomposition method, as describedin detail
elsewhere.46As-synthesized HA crystals exhibited a mean
length of ?18 lm and mean width of ?2 lm.46The specific
surface area of the crystals was 5.63 m2/g as measured by
Brunauer–Emmett–Teller (BET) N2adsorption (Autosorb-1,
The first set of experiments investigated the kinetics of
binding. HA whiskers (10 6 0.1 mg) were added to DI water,
followed by a measured volume of functionalized Au NP so-
lution, for a final volume of 15 mL containing a gold concen-
tration of 0.1 mg/L. Solutions were placed onto a test tube
rotator and allowed to incubate for 0.25, 0.5, 1, 2, 4, 8, 12,
and 24 h. After the predetermined time, solutions were cen-
trifuged at ?700g for 2 min to separate HA crystals and
bound Au NPs from Au NPs remaining in solution. Percent
binding was defined as the concentration (mg Au/L) of Au
NPs remaining in the supernatant solution after binding,
subtracted from the initial concentration of Au NPs in solu-
tion, and divided by the initial concentration of Au NPs.
Binding of functionalized Au NPs to the surface of HA crys-
tals was verified by TEM after dispersing the collected HA
crystals onto carbon-coated grids.
A second set of experiments investigated the effect of
concentration on binding affinity by plotting the binding
isotherm for each functional group. The binding affinity of
functionalized Au NPs to HA crystals was determined sepa-
rately in DI water and 10% fetal bovine serum (FBS,
Omega Scientific). The binding affinity of as-synthesized, ci-
trate stabilized Au NPs was also measured as a control.
Experiments were performed using the same methods out-
lined above for the binding kinetics, except the incubation
time was held constant at 4 h and the initial gold concen-
tration was varied. The kinetics of binding revealed that af-
ter 4 h 30–80% of Au NPs were bound to HA crystals
depending on the functional group. This incubation time
was also comparable to previous studies investigating
Binding isotherms were plotted as the amount of gold
bound per mass of HA crystals added, V (mg Au/g HA), ver-
sus the initial gold concentration, [S] (mg Au/L), for both DI
water and FBS. Plotted binding data were modeled as a
K þ ½S?
where V is the amount of functionalized Au NPs bound per
mass of HA crystals (mg/g), Vmaxis the maximum surface
binding (mg/g), [S] is the initial concentration of gold (mg/L),
and K is the equilibrium binding constant (mg/L). The equi-
librium binding constant, K, and maximum binding of func-
tionalized Au NPs on HA crystals, Vmax, were measured from
Langmuir isotherms using nonlinear least squares regression.
The gold concentration in control and supernatant solu-
tions was measured using inductively coupled plasma-
opticalemissionspectroscopy (ICP-OES, Optima
PerkinElmer). Solutions were acidified to 2% v/v HCl prior
to analysis. Calibration curves were created by diluting cer-
tified standard gold solutions (SPEX CertiPrep). All binding
tests were performed at least in triplicate, reporting the
mean and first standard deviation.
RESULTS AND DISCUSSION
Characterization of functionalized Au NPs
Functionalized Au NPs were spherical and relatively mono-
dispersed (Fig. 2). There was no apparent change in particle
size or morphology between as-synthesized and functional-
ized Au NPs. The mean (6standard deviation) particle diam-
eter was 13.4 (1.2) nm for as-synthesized Au NPs, with an as-
pect ratio of 1.1 (0.1). The mean (6standard deviation)
particle diameter was 11.4 (1.4), 13.6 (1.4), and 12.8 (1.6)
nm for glutamic acid [Fig. 2(a)], phosphonic acid [Fig. 2(b)],
and bisphosphonate [Fig. 2(c)] functionalized Au NPs,
60 ROSS AND ROEDER HYDROXYAPATITE BINDING AFFINITY OF FUNCTIONALIZED GOLD NANOPARTICLES
respectively, with a mean (6standard deviation) aspect ratio
of 1.1 (0.1) for each group. The plasmon resonance peak
remained constant at ?530 nm for functionalized Au NPs,
indicating that the colloidal suspension remained stable after
functionalization (Fig. 3). The plasmon resonance band
broadened slightly upon surface functionalization, due to the
adsorption of amine groups.
Gold surface adsorption was expected to occur through
the terminal amine group present in each functional mole-
cule (Fig. 1). Previous studies have reported that the
amine–gold surface interaction is facilitated by a weak cova-
lent bond and an electrostatic complex between protonated
amine groups and chloroaurate ions.36–39Adsorption of ter-
minal amines facilitated the availability of carboxylate,
phosphonate, and bisphosphonate groups for binding to cal-
cium on HA crystals, which was qualitatively verified by
FTIR (Fig. 4). Characteristic peaks for primary amines
(3400–3500 cm?1, NAH stretch; 1500–1650 cm?1, NAH
scissoring) were apparent for stock solutions of functional
molecules, but were not apparent for functionalized Au
NPs. A broad hydroxyl peak (2500–3600 cm?1, OAH
stretch) was present in all spectra. As-synthesized, citrate
stabilized Au NPs and phosphonic acid functionalized Au
NPs also exhibited peaks corresponding to alkyl groups
(2850 and 2920 cm?1, CAH stretch) which were not appa-
rent for glutamic acid and bisphosphonate functionalized
Au NPs. Carbonyl peaks (?1600 and 1710–1730 cm?1,
C¼ ¼O stretch) were most prominent for glutamic acid func-
tionalized and as-synthesized, citrate stabilized Au NPs.
Phosphonic acid and bisphosphonate functionalized Au NPs
(1220–1260 cm?1, P¼ ¼O stretch; 2500–2700 cm?1, POAH
stretch). The location of these peaks were similar to a pre-
vious study on phosphonic acid functionalized Au NPs.47
Taken together, the FTIR spectra suggested that Au NPs
exhibited the intended terminal surface functionality, but
did not quantify the amount of functional molecules
adsorbed to Au NP surfaces.
FIGURE 2. Particle size distributions and representative TEM micro-
graphs of (a) glutamic acid (GA), (b) phosphonic acid (PA), and (c)
bisphosphonate (BP) functionalized Au NPs. The scale bar in each
micrograph represents 100 nm. The mean (6standard deviation) parti-
cle diameter was 11.40 (1.36), 13.57 (1.43), and 12.79 (1.58) nm for
GA-, PA-, and BP-Au NPs, respectively.
FIGURE 3. UV–vis spectra of as-synthesized Au NPs compared to glu-
tamic acid (GA), phosphonic acid (PA), and bisphosphonate (BP) func-
tionalized Au NPs, showing a consistent plasmon resonance peak at
?530 nm, which indicated that the particle size did not change and
the colloid was stable after surface functionalization.
JOURNAL OF BIOMEDICAL MATERIALS RESEARCH A|OCT 2011 VOL 99A, ISSUE 161
The mean (6 standard deviation) amount of functional
molecules adsorbed to glutamic acid, phosphonic acid, and
bisphosphonate functionalized Au NP surfaces was 5.01
(0.02), 4.96 (0.02), and 3.25 (0.14) lmol/mg Au, respec-
tively, as measured by ESI-MS. Therefore, the surface den-
sity of glutamic acid and phosphonic acid was similar and
?50% greater than bisphosphonate. The lower surface den-
sity of bisphosphonate molecules adsorbed to Au NP surfa-
ces was likely due to steric interactions given the larger size
of bisphosphonate molecules compared to glutamic acid and
phosphonic acid (Fig. 1).
Functionalized Au NP binding to HA crystals
Binding of functionalized Au NPs to HA crystal surfaces was
confirmed by direct observation in TEM (Fig. 5). The number
of Au NPs bound to HA crystal surfaces was comparatively
greatest for bisphosphonate functionalized Au NPs followed
by glutamic acid and phosphonic acid. Bisphosphonate func-
tionalized Au NPs exhibited more rapid binding kinetics com-
pared to glutamic acid or phosphonic acid functionalized Au
NPs, reaching complete binding after 8 h in DI water (Fig. 6).
Glutamic acid and phosphonic acid functionalized Au NPs
exhibited similar binding kinetics, reaching near complete
binding by 24 h in DI water. Thus, kinetic tests suggested dif-
ferences in binding affinity between different functional
groups, requiring further binding experiments.
Binding affinity isotherms for functionalized Au NPs
exhibitedlarge differencesbetween functionalgroups
FIGURE 4. FTIR spectra of as-synthesized, citrate stabilized Au NPs
comparedto glutamicacid (GA),
bisphosphonate (BP) functionalized Au NPs, confirming the presence
of functional groups on gold surfaces. As-synthesized, citrate stabi-
lized Au NPs and GA-Au NPs exhibited carboxylate functionality,
while PA- and BP-Au NPs exhibited phosphonate functionality. Note
that all peaks were verified by comparison to corresponding FTIR
spectra for stock solutions of the functional molecules.
phosphonic acid(PA), and
FIGURE 5. TEM micrographs showing (a) glutamic acid (GA), (b) phosphonic acid (PA), and (c) bisphosphonate (BP) functionalized Au NPs
bound to the surface of HA crystals after 4 h in a solution containing a gold concentration of 0.1 mg/L in DI water.
FIGURE 6. Binding kinetics of glutamic acid (GA), phosphonic acid
(PA), and bisphosphonate (BP) functionalized Au NPs to HA crystals
in DI water, showing the mean mass percent of Au NPs bound as a
function of incubation time. Error bars show one standard deviation.
Percent binding was defined as the concentration (mg Au/L) of Au
NPs remaining in the supernatant solution after binding, subtracted
from the initial concentration of Au NPs in solution, and divided by
the initial concentration of Au NPs.
62 ROSS AND ROEDERHYDROXYAPATITE BINDING AFFINITY OF FUNCTIONALIZED GOLD NANOPARTICLES
(Fig. 7). Bisphosphonate functionalized Au NPs exhibited the
greatest binding affinity to the surface of HA crystals, as evi-
denced by a comparatively greater amount of gold bound
per HA (mg/g) and greater saturation concentration. Bind-
ing isotherms for glutamic acid [Fig. 7(a)], phosphonic acid
[Fig. 7(b)], and bisphosphonate [Fig. 7(c)] functionalized Au
NPs in DI water exhibited correlation coefficients (R2) of
0.88, 0.69, and 0.95, respectively, using a Langmuir model
(Table I). The equilibrium binding constant, K, and the maxi-
mum surface binding, Vmax, were greatest for bisphospho-
nate functionalized Au NPs followed by glutamic acid and
phosphonic acid (Table I). The maximum surface binding
was also normalized by the specific surface area of the HA
crystals to obtain a measure of the maximum binding of
FIGURE 7. Binding isotherms for (a) glutamic acid (GA), (b) phosphonic acid (PA), and (c) bisphosphonate (BP) functionalized Au NPs in both DI
water and FBS. Note that the scale of axes is changed in (c) due to the significantly greater binding affinity of BP-Au NPs. Experimental data for
binding in DI water were fit by a Langmuir model using Eq. (1) with the binding constants in Table I. Binding in FBS did not exhibit equilibrium
binding due to competition of soluble species present within serum solutions.
TABLE I. The Equilibrium Binding Constant, K (mg/L),
Maximum HA Surface Binding, Vmax(mg Au/g HA), and
Maximum HA Surface Binding Normalized to the Specific
Surface Area of HA Crystals, V?
Glutamic Acid (GA), Phosphonic Acid (PA), and
Bisphosphonate (BP) Functionalized Au NPs Compared to
As-Synthesized (Citrate Stabilized) Au NPs
max(mg Au/m2HA), for
Constants were determined by nonlinear least squares regression
of Langmuir isotherms.
JOURNAL OF BIOMEDICAL MATERIALS RESEARCH A|OCT 2011 VOL 99A, ISSUE 1 63
functionalized Au NPs per surface area of HA, V?
m2HA) (Table I).
As-synthesized Au NPs exhibited significantly lower
binding affinity than the lowest of functionalized Au NPs in
DI water (Table I) and therefore a relatively weak correla-
tion to Langmuir kinetics (R2¼ 0.75) (Fig. 8). Therefore,
differences in binding affinity between functional groups
were attributed to the calcium-affinity of the functional
group and not a simple electrostatic interaction.
The high binding affinity of bisphosphonate functional-
ized Au NPs was expected and consistent with previous
studies investigating bisphosphonate functionalized model
proteins, such as albumin,42–44and poly(D,L-lactide-co-glyco-
lide) nanoparticles.48Other previous studies reported that
molecules conjugated with eight aspartic acid or glutamic
acid residues exhibited greater HA binding affinity com-
pared to alendronate or pamidronate.49,50However, these
results were likely due to a greater number of calcium bind-
ing sites per molecule compared to the single binding site
present in a bisphosphonate molecule, and not necessarily
due to the relative binding affinity of the functional groups
themselves. In this study, alendronate functionalized Au NPs
exhibited a six-fold greater binding affinity to HA compared
to glutamic acid despite having ?50% fewer molecules
adsorbed to Au NP surfaces (lmol/mg Au).
The overall binding affinity of functionalized Au NPs was
decreased in FBS compared to DI water, but relative differen-
ces in binding affinity between functional groups remained
consistent. However, the binding affinity of functionalized Au
NPs in FBS was not able to be fit using any of the common
adsorption models, including linear, Langmuir, Freundlich, or
Brunauer, Emmett and Teller (BET) isotherms. Non-equilib-
rium binding behavior in serum solutions was also previ-
ously observed by Uludag et al.42Non-equilibrium binding
of functionalized Au NPs to HA crystals in FBS may be attrib-
uted to a number of possible competitive binding effects,
including (1) competition between soluble calcium ions and
HA surfaces for binding functionalized Au NPs, (2) competi-
tion of soluble proteins and functionalized Au NPs for HA
surfaces, (3) interactions between bound Au NPs through
shared interactions with either soluble calcium or soluble
proteins, (4) place-exchange reactions between functional
groups and thiol or amine containing proteins in solution, or
(5) some combination of these mechanisms. Note that serum
proteins have been shown to rapidly coat Au NP surfaces,
resulting in a change in the hydrodynamic diameter and zeta
potential,51which could contribute to blocking functional
group binding sites.
Implications for targeted delivery
A damage-specific X-ray contrast agent could enable non-
invasive and 3-D imaging of fatigue microdamage in bone.
The results of this study suggest that, of the functional
groups investigated, bisphosphonate functionalized Au NPs
would provide the highest surface density (cf., Table I) of
Au NPs labeling damaged bone tissue and, therefore, the
greatest contrast enhancement in X-ray tomography. This
will be verified in future work. Fatigue microcracks in
human cortical bone specimens labeled by bisphosphonate
functionalized Au NPs were able to be detected by synchro-
tron X-ray tomography in a preliminary study.52On the
other hand, release and clearance of the contrast agent in
clinical use might benefit from functional groups with inter-
mediate binding affinity, similar to L-glutamic acid in this
study. For example, c-carboxyglutamic acid is a post-transla-
tionally modified amino acid which has been identified in
proteins with hydroxyapatite affinity.53Further work will be
required to determine the appropriate balance between the
contrast enhancement and clearance of the agent. The sig-
nificance of the present study was in demonstrating the
ability to tailor and quantify the binding affinity for these
An additional concern for the use of bisphosphonate
functionalized Au NPs as an imaging contrast agent is the
pharmacological activity of bisphosphonates, which are com-
monly used to treat osteoporosis by inhibiting the cellular
remodeling process to prevent additional loss of bone mass.
The mechanism by which bisphosphonates are able to in-
hibit osteoclast activity is directly related to the presence of
nitrogen.54Nitrogen-containing bisphosphonates, such as
alendronate, have been shown to selectively inhibit a spe-
cific enzymatic pathway necessary for osteoclast function,
while bisphosphonates that do not contain nitrogen are
incorporated into toxic ATP analogs that cause apoptosis.54
However, the nitrogen in alendronate is in the form of a pri-
mary amine, which was bound to the surface of Au NPs.
Therefore, the pharmacological activity may have been
passivated. The effects of amine binding to gold surfaces on
cellular activity is unknown and will depend on potential
intracellular place-exchange reactions between thiol and
FIGURE 8. Binding isotherm for as-synthesized Au NPs in DI water.
Experimental data was fit by a Langmuir model using Eq. (1) with the
binding constants in Table I. The low binding affinity relative to func-
tionalized Au NPs (Fig. 7) verified that binding was due to the cal-
cium-specific functional groups rather than electrostatic interactions.
Note that the scale of axes is different than Figure 7 due to a rela-
tively low binding affinity.
64 ROSS AND ROEDERHYDROXYAPATITE BINDING AFFINITY OF FUNCTIONALIZED GOLD NANOPARTICLES
Amongst the expansive body of research for active tar-
geting of Au NPs for biological applications,1–4there has not,
to our knowledge, been prior investigation on the binding af-
finity of functionalized Au NPs for mineralized tissue. While
the main motivation for this study was in the use of Au NPs
as a targeted contrast agent for labeling damaged bone
tissue, other applications are readily envisioned for the
detection of calcifications in soft tissue associated with can-
cer or musculoskeletal injury, and as drug delivery vehicles.
Au NPs were synthesized to a mean particle diameter of 10–
15 nm and surface functionalized with either L-glutamic
acid, 2-aminoethylphosphonic acid, or alendronate, which ex-
hibit a primary amine for gold surface adsorption opposite
carboxylate, phosphonate, or bisphosphonate groups, respec-
tively, for targeting calcium. Bisphosphonate functionalized
Au NPs exhibited the most rapid binding kinetics and great-
est binding affinity to HA, followed by glutamic acid and
phosphonic acid. Functionalized Au NPs exhibited lower
overall binding in FBS compared to de-ionized water, but rel-
ative differences between functional groups were similar.
The high binding affinity of bisphosphonate functionalized
Au NPs to HA could lead to improved labeling of damage
bone tissue and enhanced-contrast in X-ray tomography.
This research was supported by the U.S. Army Medical
Research and Materiel Command (W81XWH-06-1-0196)
through the Peer Reviewed Medical Research Program
(PR054672). The authors acknowledge the Universityof Notre
Dame Center for Environmental Science and Technology
(CEST) for the use of ICP-OES and UV–vis spectroscopy, the
Mass Spectrometry and Proteomics Facility and William C.
Boggess for the use of ESI-MS, and the Notre Dame Integrated
Imaging Facility (NDIIF) for the use of TEM. Timothy L. Conrad
is gratefully acknowledged for preparing the HA crystals used
1. Sperling RA, Rivera Gil P, Zhang F, Zanella M, Parak WJ. Biologi-
cal applications of gold nanoparticles. Chem Soc Rev 2008;37:
2. Huang X, Jain PK, El-Sayed IH, El-Sayed MA. Gold nanoparticles:
Interesting optical properties and recent applications in cancer
diagnostics and therapy. Nanomedicine 2007;2:681–693.
3. Ghosh P, Han G, De M, Kim CK, Rotello VM. Gold nanoparticles
in delivery applications. Adv Drug Deliv Rev 2008;60:1307–1315.
4. Pissuwan D, Valenzuela SM, Cortie MB. Therapeutic possibilities
of plasmonically heated gold nanoparticles. Trends Biotechnol
5. Hirsch LR, Stafford RJ, Bankson JA, Sershen SR, Rivera B, Price
RE, Hazle JD, Halas NJ, West JL. Nanoshell-mediated near-infra-
red thermal therapy of tumors under magnetic resonance guid-
ance. Proc Nat Acad Sci 2003;100:13540–13554.
6. Elghanian R, Storhoff JJ, Mucic RC, Letsinger RL, Mirkin CA.
Selective colorimetric detection of polynucleotides based on the
distance-dependent optical properties of gold nanoparticles. Sci-
7. Mo ¨ller R, Fritzche W. Metal nanoparticle-based detection for DNA
analysis. Curr Pharm Biotechnol 2007;8:274–285.
8. Hayat MA, editor. Colloidal Gold: Principles Methods and Applica-
tions, Vol. 1. San Diego: Academic Press; 1989.
9. Hainfeld JF, Slatkin DN, Focella TM, Smilowitz HM. Gold nanopar-
ticles: A new X-ray contrast agent. Br J Radiol 2006;79:248–253.
10. Cai Q-Y, Kim SH, Choi KS, Kim SY, Byun SJ, Kim KW, Park SH,
Juhng SK, Yoon KH. Colloidal gold nanoparticles as a blood-pool
contrast agent for X-ray computed tomography in mice. Invest
11. Kim D, Park S, Lee JH, Jeong YY, Jon S. Antibiofouling polymer-
coated gold nanoparticles as a contrast agent for in vivo X-ray
computed tomography imaging. J Am Chem Soc 2007;129:
12. Zhang Z, Ross RD, Roeder RK. Preparation of functionalized gold
nanoparticles as a targeted X-ray contrast agent for damaged
bone tissue. Nanoscale 2010;2:582–586.
13. Burr DB, Forwood MR, Fyhrie DP, Martin RB, Schaffler MB, Turner
CH. Bone microdamage and skeletal fragility in osteoporotic and
stress fractures. J Bone Miner Res 1997;12:6–15.
14. Chapurlat BD, Delmas PD. Bone microdamage: A clinical perspec-
tive. Osteoporos Int 2009;20:1299–1308.
15. Lee TC, Mohsin S, Taylor D, Parkesh R, Gunnlaugsson T, O’Brien
FJ, Giehl M, Gowin W. Detecting microdamage in bone. J Anat
16. Frost HM. Presence of microscopic cracks in vivo in bone. Henry
Ford Hosp Med Bull 1960;8:25–35.
17. Burr DB, Hooser M. Alterations to the en bloc basic fuchsin stain-
ing protocol for the demonstration of microdamage produced in
vivo. Bone 1995;17(4):431–433.
18. O’Brien FJ, Taylor D, Lee TC. An improved labelling technique for
monitoring microcrack growth in compact bone. J Biomech 2002;
19. O’Brien FJ, Taylor D, Lee TC. Microcrack accumulation at different
intervals during fatigue testing of compact bone. J Biomech 2003;
20. Wang X, Niebur GL. Microdamage propagation in trabecular bone
due to changes in loading mode. J Biomech 2006;39:781–790.
21. Parkesh R, Mohsin S, Lee TC, Gunnlaugsson T. Histological, spec-
troscopic, and surface analysis of microdamage in bone: Toward
real-time analysis using fluorescent sensors. Chem Mater 2007;19:
22. Parkesh R, Lee TC, Gunnlaugsson T. Fluorescence imaging of
bone cracks (microdamage) using visibly emitting 1,8-naphthali-
mide-based PET sensors. Tetrahedron Lett 2009;50:4114–4116.
23. Mohsin S, O’Brien FJ, Lee TC. Microcracks in compact bone: A
three-dimensional view. J Anat 2006;209:119–124.
24. Bigley RF, Singh M, Hernandez CJ, Kazakia GJ, Martin RB, Kea-
veny TM. Validity of serial milling-based imaging system for
microdamage quantification. Bone 2008;42:212–216.
25. Zioupos P, Currey JD. The extent of microcracking and the morphol-
ogy of microcracks in damaged bone. J Mater Sci 1994;29:978–986.
26. O’Brien FJ, Taylor D, Dickson GR, Lee TC. Visualization of three-
dimensional microcracks in compact bone. J Anat 2000;197:413–420.
27. Wang X, Masse DB, Leng H, Hess KP, Ross RD, Roeder RK, Nie-
bur GL. Detection of trabecular bone microdamage by micro-com-
puted tomography. J Biomech 2007;40:3397–3403.
28. Leng H, Wang X, Ross RD, Niebur GL, Roeder RK. Micro-com-
puted tomography of fatigue microdamage in cortical bone using
a barium sulfate contrast agent. J Mech Behav Biomed Mater
29. Landrigan MD, Li J, Turnbull TL, Burr DB, Niebur GL, Roeder RK. Con-
trast-enhanced micro-computed tomography of fatigue microdam-
age accumulation in human cortical bone. Bone 2011;48:443–450.
30. Parkesh R, Gowin W, Lee TC, Gunnlaugsson T. Synthesis and
agents for bone structure and microdamage analysis. Org Biomol
31. Parkesh R, Lee TC, Gunnlaugsson T, Gowin W. Microdamage in
bone: Surface analysis and radiological detection. J Biomech
32. Hubbell JH, Seltzer SM. Table of X-ray Mass Attenuation Coeffi-
cients and Mass Energy-Absorption Coefficients (version 1.4). Gai-
thersburg, MD: National Institute of Standards and Technology;
2004. Available at http://physics.nist.gov/xaamdi.
33. Giersig M, Mulvaney P. Preparation of ordered colloid monolayers
by electrophoretic deposition. Langmuir 1993;9:3408–3413.
JOURNAL OF BIOMEDICAL MATERIALS RESEARCH A|OCT 2011 VOL 99A, ISSUE 165
34. Brust M, Walker M, Bethell D, Schiffrin DJ, Whyman R. Synthesis Download full-text
of thiol-derivated gold nanoparticles in a 2-phase liquid-liquid
system. J Chem Soc Chem Commun 1994;7:801–802.
35. Brust M, Fink J, Bethell D, Schiffrin DJ, Kiely C. Synthesis and
reactions of functionalised gold nanoparticles. J Chem Soc Chem
Commun 1995: 1655–1656.
36. Leff DV, Brandt L, Heath JR. Synthesis and characterization of
hydrophobic, organically-soluble gold nanocrystals functionalized
with primary amines. Langmuir 1996;12:4723–2730.
37. Kumar A, Mandal S, Selvakannan PR, Pasricha R, Mandale AB,
Sastry M. Investigation into the interaction between surface-
bound alkylamines and gold nanoparticles. Langmuir 2003;19:
38. Selvakannan PR, Mandal S, Phadtare S, Pasricha R, Sastry M.
Capping of gold Nanoparticles by the amino acid lysine renders
them water-dispersable. Langmuir 2003;19:3545–3549.
39. Aslam M, Fu L, Su M, Vijayamohanan K, Dravid VP. Novel one-
step synthesis of amine-stabilized aqueous colloidal gold nano-
particles. J Mater Chem 2004:1795–1997.
40. Bolander ME, Young MF, Fisher LW, Yamada Y, Termine JD.
Osteonectin cDNA sequence reveals potential binding regions
for calcium and hydroxyapatite and shows homologies with
both a basement membrane protein (SPARC) and a serine
proteinase inhibitor (ovomucoid). Proc Natl Acad Sci 1988;85:
41. Fujisawa R, Wada Y, Nodasaka Y, Kuboki Y. Acidic amino acid-
rich sequences as binding sites of osteonectin to hydroxyapatite
crystals. Biochim Biophys Acta 1996;1292:53–60.
42. Uludag H, Kousinioris N, Gao T, Kantoci D. Bisphosphonate con-
jugation to proteins as a means to impart bone affinity. Biotech-
nol Prog 2000;16:258–267.
43. Bansal G, Wright JEI, Zhang S, Zernicke RF, Uludag H. Imparting
mineral affinity to proteins with thiol-labile disulfide linkages. J
Biomed Mater Res A 2005;74:618–628.
44. Wright JEI, Gittens SA, Bansal G, Kitov PI, Sindrey D, Kucharski
C, Uludag H. A comparison of mineral affinity of bisphosphonate-
protein conjugates constructed with disulfide and thioether link-
ages. Biomaterials 2006;26:769–784.
45. Turkevich J, Stevenson PC, Hillier J. A study of the nucleation
and growth processes in the synthesis of colloidal gold. Discuss
Faraday Soc 1951;11:55–75.
46. Roeder RK, Converse GL, Leng H, Yue W. Kinetic effects on hy-
droxyapatite whiskers synthesized by the chelate decomposition
method. J Am Ceram Soc 2006;89:2096–2104.
47. Fiurasek P, Reven L. Phosphonic and sulfonic acid-functionalized
gold nanoparticles: A solid-state NMR study. Langmuir 2007;23:
48. Choi SW, Kim JH. Design of surface-modified poly(D,L-lactide-co-
glycolide) nanoparticles for targeted drug delivery to bone. J Con-
trol Release 2007;122:24–30.
49. Wang D, Miller S, Sima M, Kopeckova ´ P, Kopecek J. Synthesis
and evaluation of water-soluble polymeric bone-targeting drug
delivery systems. Bioconjug Chem 2003;14:853–859.
50. Murphy MB, Hartgerink JD, Goepferich A, Mikos AG. Synthesis
and in vitro hydroxyapatite binding of peptides conjugated to cal-
cium-binding moieties. Biomacromolecules 2007;8:2237–2243.
51. Alkilany AM, Nagaria PK, Hexel CR, Shaw TJ, Murphy CJ, Wyatt
MD. Cellular uptake and cytotoxicity of gold nanorods: Molecular
origin of cytotoxicity and surface effects. Small 2009;5:701–708.
52. Ross RD, Roeder RK. Functionalized gold nanoparticles for tar-
geted labeling of damaged bone tissue in X-ray tomography.
Trans Orthop Res Soc 2010;35:1368.
53. Price PA, Otsuka AS, Poser JW, Kristaponis J, Raman N. Charac-
terization of a gamma-carboxyglutamic acid-containing protein
from bone. Proc Natl Acad Sci USA 1976;73:1447–1451.
54. Reszka AA, Roda GA. Bisphosphonate mechanism of action. Curr
Rheumatol Rep 2003;5:65–74.
66 ROSS AND ROEDERHYDROXYAPATITE BINDING AFFINITY OF FUNCTIONALIZED GOLD NANOPARTICLES