Pre- and postfunctionalized self-assembled π-conjugated fluorescent organic nanoparticles for dual targeting

Article abstract of DOI:10.1021/ja2075345

There is currently a high demand for novel approaches to engineer fluorescent nanoparticles with precise surface properties suitable for various applications, including imaging and sensing. To this end, we report a facile and highly reproducible one-step method for generating functionalized fluorescent organic nanoparticles via self-assembly of prefunctionalized π-conjugated oligomers. The engineered design of the nonionic amphiphilic oligomers enables the introduction of different ligands at the extremities of inert ethylene glycol side chains without interfering with the self-assembly process. The intrinsic fluorescence of the nanoparticles permits the measurement of their surface properties and binding to dye-labeled target molecules via Foerster resonance energy transfer (FRET). Co-assembly of differently functionalized oligomers is also demonstrated, which enables the tuning of ligand composition and density. Furthermore, nanoparticle prefunctionalization has been combined with subsequent postmodification of azide-bearing oligomers via click chemistry. This allows for expanding ligand diversity at two independent stages in the nanoparticle fabrication process. The practicability of the different methods entails greater control over surface functionality. Through labeling with different ligands, selective binding of proteins, bacteria, and functionalized beads to the nanoparticles has been achieved. This, in combination with the absence of unspecific adsorption, clearly demonstrates the broad potential of these nanoparticles for selective targeting and sequestration. Therefore, controlled bifunctionalization of fluorescent π-conjugated oligomer nanoparticles represents a novel approach with high applicability to multitargeted imaging and sensing in biology and medicine.

Full text of DOI:10.1021/ja2075345

ARTICLE
pubs.acs.org/JACS
Pre- and Postfunctionalized Self-Assembled π-Conjugated
Fluorescent Organic Nanoparticles for Dual Targeting
Katja Petkau,† Adrien Kaeser,‡ Irꢀen Fischer,‡ Luc Brunsveld,*,† and Albertus P. H. J. Schenning*,‡
†Laboratory of Chemical Biology and ‡Laboratory of Functional Organic Materials and Devices, Eindhoven University of Technology,
Den Dolech 2, 5612 AZ Eindhoven, The Netherlands
S Supporting Information
b
ABSTRACT: There is currently a high demand for novel
approaches to engineer fluorescent nanoparticles with precise
surface properties suitable for various applications, including
imaging and sensing. To this end, we report a facile and highly
reproducible one-step method for generating functionalized
fluorescent organic nanoparticles via self-assembly of prefunc-
tionalized π-conjugated oligomers. The engineered design of
the nonionic amphiphilic oligomers enables the introduction of
different ligands at the extremities of inert ethylene glycol side
chains without interfering with the self-assembly process. The
intrinsic fluorescence of the nanoparticles permits the measure-
ment of their surface properties and binding to dye-labeled target molecules via F€orster resonance energy transfer (FRET). Co-
assembly of differently functionalized oligomers is also demonstrated, which enables the tuning of ligand composition and density.
Furthermore, nanoparticle prefunctionalization has been combined with subsequent postmodification of azide-bearing oligomers
via click chemistry. This allows for expanding ligand diversity at two independent stages in the nanoparticle fabrication process. The
practicability of the different methods entails greater control over surface functionality. Through labeling with different ligands,
selective binding of proteins, bacteria, and functionalized beads to the nanoparticles has been achieved. This, in combination with
the absence of unspecific adsorption, clearly demonstrates the broad potential of these nanoparticles for selective targeting and
sequestration. Therefore, controlled bifunctionalization of fluorescent π-conjugated oligomer nanoparticles represents a novel
approach with high applicability to multitargeted imaging and sensing in biology and medicine.
’ INTRODUCTION
based on self-assembled lipid amphiphiles, are prepared from
prefunctionalized building blocks.8 However, sensitive targeting
ligands like proteins are attached to liposomes via postfunctionali-
zation.9 Additionally, in order to construct fluorescent liposomes,
dye molecules are introduced, for example, via the polar head
groups of the lipids on the outside of the liposome.
Fluorescent organic nanoparticles based on π-conjugated
systems combine the features of inorganic nanoparticles and
liposomes in being self-assembling amphiphiles with high fluor-
escence brightness and photostability.10 The functionalization of
these particles with biomolecules, which will be crucial for their
use in sensing and imaging applications, has scarcely been
reported. π-Conjugated polymer dots have been coassembled
with polymers bearing carboxylic acids as attachment points for
postfunctionalization by McNeill, Chiu, and co-workers. In this
way, the nanoparticle surface was functionalized on the one hand
with ethylene glycol chains to provide a biocompatible layer that
minimizes nonspecific adsorption and on the other hand with
biomolecules to afford specific targeting.11 Interestingly, these
particles were used as probes for in vivo cell labeling and tumor
The ability to actively target specific receptors and, as a result,
cells is a prerequisite for the use of fluorescent nanoparticles as
molecular imaging probes, diagnostics, and therapeutic delivery
vehicles.1 Targeting is typically achieved through surface func-
tionalization of the nanoparticles with high-affinity ligands, such
as small molecules, peptides, and antibodies. Multiple ligands can
be attached, leading to high binding affinities due to multivalent
interactions, which are necessary to achieve maximal selectivity in
diagnostics.2 This at the same time requires chemical control
over the ligand functionalization of nanoparticle surfaces via pre-
or postfunctionalization.
Several classes of nanoparticles have been developed in the last
few decades starting from the first liposomes in the 1960s3 via
inorganic quantum dots4 to organic nanoparticles.5 Because of
their intrinsic fluorescence with high quantum yields, especially
quantum dots are currently widely used as fluorescent imaging
probes.6 Having no intrinsic aqueous solubility, these inorganic
nanoparticles require a relatively thick encapsulation layer to
ensure biocompatibility and stability. These so-called caps not
only solubilize and protect the inorganic core, but as well carry
reactive groups, which serve as points of covalent attachment for
biomolecules via postfunctionalization.7 Liposomes, which are
Received: August 10, 2011
Published: September 13, 2011
r
2011 American Chemical Society
17063
dx.doi.org/10.1021/ja2075345 J. Am. Chem. Soc. 2011, 133, 17063–17071
|

Journal of the American Chemical Society
ARTICLE
Figure 1. (a) Chemical structures of amphiphiles used in this study and a photograph of a 0.3 μM nanoparticle solution under UV light (360 nm). (b)
(i) Generation of nanoparticles prefunctionalized with mannose using mixtures of 1 and 2. (ii) Generation of nanoparticles prefunctionalized with azides
using mixtures of 1 and 3 and subsequent postfunctionalization via copper catalyzed azideꢀalkyne cycloaddition with alkyne derivatives of either
mannose or biotin. (iii) Generation of bifunctional nanoparticles containing azides and mannose using mixtures of 2 and 3 and subsequent
postfunctionalization via copper catalyzed azideꢀalkyne cycloaddition with alkyne derived biotin yielding dual targeting mannose and biotin labeled
nanoparticles.
imaging.11,12 Using a large variation of π-conjugated oligomers,
an extensive pioneering structural study has been performed by
Lee and co-workers. Mannose prefunctionalized oligomers were
designed to assemble into bioactive nanostructures of different
shapes showing enhanced binding affinities due to multivalency.13
Although the nanoparticles were not fluorescent, the encapsula-
tion of a dye was possible, which could even trigger changes in the
structure of these dynamic aggregates.14
Recently, we have reported on self-assembled fluorescent
nanoparticles based on π-conjugated fluorene oligomers in
water.15 Via the coassembly of the different fluorene-based
amphiphiles, fluorescent nanoparticles with tunable emission
colors covering the entire visible range (including white)16 were
obtained. We now report on a facile and flexible approach to
prepare (multi)functional fluorescent organic nanoparticles for
(multi)targeted imaging. Via the coassembly of synthetically
accessible prefunctionalized π-conjugated oligomers a library
of functionalized nanoparticles was generated. The amphiphilic
oligomers were on the one hand prefunctionalized with ligands
for biological targeting, such as mannose, and on the other hand
with azides to allow postfunctionalization of preformed nano-
particles (Figure 1) via the copper-catalyzed azideꢀalkyne
cycloaddition (CuAAC) reaction. Multitargeting of the fluo-
rescent organic nanoparticles was demonstrated through func-
tionalization of the same nanoparticle with different ligands,
which were recognized by their corresponding protein partners.
The nanoparticles were capable of selective binding to bacteria
and to magnetic beads, which led to nanoparticleꢀtarget
clustering and, in the case of magnetic beads, to complete
removal from solution. Our results reveal synthetic pathways
and self-assembly protocols to prepare tunable functionalized
fluorescent multivalent nanoparticles for biological multitarget
imaging applications.
’ RESULTS AND DISCUSSION
Design and Synthesis. The studied amphiphilic fluorene
oligomers 1ꢀ3 (Scheme 1) consist of the same π-conjugated
core of two fluorene units connected by a benzothiadiazole
linker. The hydrophobic chromophore has been decorated with
gallic acid derivatives bearing at one end alkyl tails and ethylene
glycol chains at the other. These hydrophilic ethylene glycol
chains are not only necessary for phase segregation but should
also prevent unspecific adsorption of biological matter to the
surface of the resulting nanoparticles.17 Furthermore, polyethy-
lene glycols have an excellent water solubility and low toxicity
and are now extensively used in clinical applications.18
To functionalize the surface of the nanoparticles, our
strategy was to introduce ligands at the periphery of hydro-
philic ethylene glycol chains. This should ensure the exposure
of the ligands to the aqueous environment.19,20 Two func-
tionalities, a mannose and an azide group, were selected for
the prefunctionalization of the oligomers (Figure 1, com-
pounds 2 and 3, respectively). Mannose is used to evaluate
the binding to biological structures such as Escherichia coli
bacteria21 and the lectin concanavalin A (ConA),22 while the
17064
dx.doi.org/10.1021/ja2075345 |J. Am. Chem. Soc. 2011, 133, 17063–17071

Journal of the American Chemical Society
ARTICLE
Scheme 1. Synthesis of the Mannose and Azide Amphiphiles 2 and 3a
a Reagents and conditions: (a) CuSO4, sodium ascorbate, H2O/tBuOH, 1:1, rt, 56%; (b) Et3N, CH2Cl2, 0 °C to rt, 26%; (c) HBTU, DIPEA, DMF, 40 °C,
47%; (d) (i) (COCl)2, Et3N, CH2Cl2, 0 °C to rt, quant., (ii) Et3N, CH2Cl2, 0 °C to rt, 55%.
Figure 2. (a) Absorption and emission spectra of self-assembled NP, azide-50-NP, and mannose-50-NP in water and of molecularly dissolved
amphiphile 1 in THF (amphiphile concentration for absorption measurement 3 μM, for emission spectra 1.5 μM). (b) Hydrodynamic diameter of
azide-50-NP (0.3 μM in phosphate buffer) measured by dynamic light scattering. (c) Transmission electron microscopy image of azide-50-NP. Scale
bar represents 200 nm.
azide moiety allows the preparation of nanoparticles for
postfunctionalization via the copper catalyzed azideꢀalkyne
cycloaddition reaction.23
chromophore 10 with 11, an acid chloride of gallic acid
derivatized with hydrophobic alkyl chains, yielding 12, which
was used for the preparation of functionalized amphiphiles 2
and 3. For the final acylation step, the azide-functionalized
gallic acid derivative 5 was transformed into an acid chloride
in situ using oxalyl chloride and reacted with 12. The mannose-
functionalized gallic acid derivative 7 was preactivated using
O-benzotriazole-N,N,N0,N0-tetramethyluronium hexafluoropho-
sphate (HBTU) and subsequently coupled to 12. Both activation
strategies led to the desired amphiphiles. (Scheme 1 and Supporting
Information).
The synthesis of amphiphile 1 is based on our earlier reported
method.15 For the synthesis of amphiphile 2 and 3, the azide-
functionalized gallic acid derivative 5 was synthesized in a
multistep process, as described previously.19 Reacting this build-
ing block with propargyl mannose 6 in the presence of copper
sulfate and sodium ascorbate yielded 7. Suzuki coupling of
benzothiadiazole dibromo compound 9 with aminofluorene
boronic ester 8 was followed by acylation of the resulting diamino
17065
dx.doi.org/10.1021/ja2075345 |J. Am. Chem. Soc. 2011, 133, 17063–17071

Journal of the American Chemical Society
ARTICLE
Figure 3. (a) Photograph of washed E. coli bacteria under UV illumination after incubation with (left) azide-50-NP and (right) mannose-50-NP (both
3 μM). (b) Hydrodynamic diameter of mannose-50-NPs (0.3 μM) before and after the addition of the lectin ConA (2 nM) measured by dynamic light
scattering in phosphate buffer (with 0.1 mM CaCl2 and 0.1 mM MnCl2) at pH = 7.0.
Formation and Characterization of Prefunctionalized Na-
noparticles. Nanoparticles from the π-conjugated oligomers
were prepared using the reprecipitation method24 via rapid
injection of 15 μL of a 1 mM THF stock solution of the oligomer,
or mixtures of oligomers, into 5 mL of water (Figure 1b). In this
way, by premixing known volumes of THF stock solutions of the
oligomers 1ꢀ3 before injection, nanoparticles with equal emis-
sion colors but different functional group compositions were
generated and their properties studied (see also Tables S1 and
S2, Supporting Information). Nonfunctionalized nanoparticles
(NP) of 1 only and functionalized nanoparticles containing
50% of 1 and either 50% of azide amphiphile 3 (azide-50-NP)
or 50% of mannose amphiphile 2 (mannose-50-NP) were also
prepared. Dynamic light scattering (DLS) measurements
showed for all nanoparticles a similar average hydrodynamic
radius of 40ꢀ50 nm, which is in good agreement with the size of
the spherical nanoparticles observed with transmission electron
microscopy (TEM) (Figure 2). The oligomers thus self-assem-
ble into spherical nanoparticles, and the TEM images suggest an
amorphous internal morphology. The NP, azide-50-NP, and
mannose-50-NP systems all showed identical absorption max-
ima at around 334 and 430 nm and an emission maximum at
542 nm similar to the previously reported nanoparticles based
on self-assembled bolaamphiphilic fluorene benzothiadiazole
derivatives15 (Figure 2a). Additionally, the different nanopar-
ticles all had high quantum yields between 80 and 90% (Tables
S1 and S2, Supporting Information). These results show that
the introduction of mannose and azide functionalities into the
π-conjugated oligomers does not alter the size and optical and
fluorescence properties of the nanoparticles. Most likely the
self-assembly and optical properties are dominated by the
core structure of these amphiphiles, which is not influenced by
the introduced functionalities, since these compounds contain
the same chromophore, and the hydrophilic/hydrophobic ratio is
hardly changed.
that the mannose-functionalized particles bind to the bacteria. At
the same time, the incubation of E. coli with azide-50-NP did not
stain the bacterial pellet (Figure 3a). Additionally, the hydro-
dynamic radius of these nanoparticles before and after the
addition of mannose binding lectin ConA was measured with
DLS (Figure 3b). The addition of ConA to mannose-50-NPs
significantly increased the hydrodynamic radius of the particles in
solution, presumably via the cross-linking of the nanoparticles
upon binding to the tetrameric lectin. In contrast, no increase in
the size of azide-50-NPs was observed via DLS upon addition of
ConA (Figure S2, Supporting Information), indicating that the
azide functionality does not lead to unspecific binding. Our
results show, therefore, that targeting organic nanoparticles can
be prepared via a simple coassembly process. Targeting is solely
induced through the appending ligand, and unspecific adsorp-
tion is prevented by the shielding ethylene glycol chains. At the
same time, the introduction of a ligand at the extremities of the
π-conjugated oligomers does not affect the optical and self-
assembly properties.
Control over Nanoparticle Ligand Density. An important
feature of nanoparticles generated via self-assembly of prefunc-
tionalized building blocks should be the facile and reproducible
control over ligand density.26 To investigate the effect of
nanoparticle ligand densities on ConA binding, particles con-
taining different amounts of mannose amphiphile were subse-
quently prepared, ranging from 10% (mannose-10-NP) to 90%
(mannose-90-NP) of 2 (Figure 4a and Supporting Informa-
tion, Figure S1 and Tables S1 and S2).27 To determine the
influence of ligand density on protein binding, F€orster reso-
nance energy transfer (FRET) studies were performed using
dye-labeled ConA. Alexa Fluor 633 (AF633) was the dye of
choice, as its absorbance spectrum overlaps with the emission
spectrum of the nanoparticles. The binding of ConA to
mannose on the nanoparticle surface should bring the nano-
particle and the dye-labeled protein in close proximity and
enable energy transfer from the fluorene benzothiadiazole
donor chromophore to the acceptor dye AF633. Thus, addition
of AF633-labeled ConA (ConA-AF633) indeed resulted in a
decrease in the emission of mannose-NP (λ = 542 nm) and the
appearance of AF633 acceptor emission (λ = 648 nm)
To investigate the bioavailability of the introduced mannose,
mannose-50-NPs were incubated with E. coli bacteria, since it is
known that mannose specifically binds to the FimH receptor of
E. coli bacteria.25 After incubation and washing, the bacterial
pellet was yellow fluorescent under UV illumination, showing
17066
dx.doi.org/10.1021/ja2075345 |J. Am. Chem. Soc. 2011, 133, 17063–17071

Journal of the American Chemical Society
ARTICLE
Figure 4. (a) Generation of prefunctionalized mannose nanoparticles with different densities of mannose on the nanoparticle surface using mixtures of
1 and 2. (b) Change in emission spectra of mannose-90-NP upon addition of Alexa Fluor 633 labeled concanavalin A (ConA-AF633; 0.8 ꢀ15 nM)
caused by energy transfer between nanoparticles and ConA-AF633. (c) Change in FRET ratio of different nanoparticles upon addition of ConA-AF633
shown as the mean value of at least three measurements. The determined nanoparticleꢀprotein energy transfer ratio was constant for different batches of
nanoparticles prepared from different stock solutions, underlining the reproducibility of the generated ligand densities. See Figure S3 (Supporting
Information) for all titration curves. (d) Change in FRET ratio of mannose-50-NPs upon addition of ConA-AF633 without and in the presence of
125 μM methyl mannose. All spectra measured in phosphate buffer (with 0.1 mM CaCl2 and 0.1 mM MnCl2) at pH 7.0 using a nanoparticle
concentration of 0.3 μM and an excitation wavelength of 350 nm.
(Figure 4b and Supporting Information, Figure S3).25 In
contrast, when AF633-labeled ConA was added to azide-50-
NP, no energy transfer was observed, indicating that the
mannose molecules attached to the nanoparticles are crucial
for binding to ConA (vida supra). Interestingly, a clear correla-
tion between the energy transfer and the percentage of used
mannose amphiphile 2 could be observed, with the highest FRET
ratio in the case of mannose-90-NPs (Figure 4c). A higher
mannose density leads to a higher recruitment of proteins to the
nanoparticle surface, resulting in an increase in FRET. The
tuning of ligand density during the nanoparticle preparation
therefore modulates the extent of protein binding to the
nanoparticles.
Remarkably, FRET studies in the absence and presence of a
500-fold excess of competing methyl mannose did not lead to
changes in FRET ratio (Figure 4d). Methyl mannose is thus not
capable of competing with the nanoparticles for ConA binding,
showing the strength of lectin ConA binding to the nanoparti-
cles. This indicates that a high ligand density of a low affinity
ligand (mannose) leads to increased effective binding affinities of
the nanoparticles.28 This multivalency effect has not been observed
for molecular dissolved mannose-containing π-conjugated
oligomers29 and shows the advantage of using self-assembled
nanoparticles based on π-conjugated oligomers. The ability to
control and tune the ligand density in a single preparation step
opens up the possibility to screen for optimal surface functio-
nalization for a given application in a simple and reliable way.
Postfunctionalization via AzideꢀAlkyne Cycloaddition.
Nanoparticles with surface accessible azides enable the introduc-
tion of diverse ligands via the copper(I)-catalyzed azideꢀalkyne
cycloaddition (CuAAC) reaction after nanoparticle formation
(Figure 1b,ii).11,30 We have first postfunctionalized the azide
nanoparticles, azide-50-NP, with mannose to compare the degree
of functionalization via the pre- and postfunctionalization methods.
The cycloaddition reaction of azide nanoparticles was per-
formed in degassed water using copper(II) sulfate. The copper
salts did not cause any significant quenching of nanoparticle
fluorescence (Figure S4, Supporting Information).11 The excess
17067
dx.doi.org/10.1021/ja2075345 |J. Am. Chem. Soc. 2011, 133, 17063–17071

Journal of the American Chemical Society
ARTICLE
of ligand and more importantly the cytotoxic copper salts were
removed either by dialysis or PD-10 Sephadex desalting column,
both purification methods leading to the same results (Figure S5,
Supporting Information). The functionalization was confirmed
by mass spectrometry (MALDI-ToF), where, in addition to the
mass of the azide amphiphile 3, the mass of the mannose-
functionalized amphiphile 2 could be detected (Figure S6,
Supporting Information). After CuAAC functionalization of
azide-50-NPs with propargyl mannose 6, changes in the FRET
ratio of the postfunctionalized nanoparticles upon addition of
ConA-AF633 were measured and correlated with changes in the
FRET ratios of prefunctionalized mannose-50-NPs. The post-
functionalization was successful as observed by the energy
transfer from nanoparticle to the dye-labeled protein. When
compared with the prefunctionalized mannose-50-NPs, the
FRET experiments revealed a significantly lower degree of ligand
density for the postfunctionalization strategy (Figure 5). On the
basis of the observed FRET ratios for the prefunctionalized
mannose nanoparticles (Figure 4c), less than 10% of the azide
moieties in the azide-50-NPs was functionalized with mannose
using typical aqueous CuAAC conditions. The lower degree of
functionalization via the postfunctionalization method compared
with the prefunctionalization is not surprising, as the accessibility
and therewith the degree of amphiphile functionalization will be
significantly lower for preformed nanoparticles.
Subsequently, biotin was chosen as a second biomolecule to
investigate the postfunctionalization of the azide nanoparticles.
With a suitable alkyne derivative, biotin could also be successfully
introduced onto the nanoparticle surface via CuAAC reaction at
room temperature (Figure 1b,ii). The functionalization was
Figure 5. Changes in FRET ratio upon addition of ConA-AF633 to
prefunctionalized mannose-50-NPs, azide-50-NPs, or azide-50-NPs
postfunctionalized with mannose,31 in phosphate buffer (with 0.1 mM
CaCl2 and 0.1 mM MnCl2) at pH = 7.0 using an excitation wavelength of
350 nm. The nanoparticle concentration was 0.3 μM.
Figure 7. Photographs of azide-50-NP and biotin-NP under visible
light (left) and UV light (right) after incubation with 150 μL of magnetic
streptavidin beads (for preparation see the Supporting Information)
followed by magnetic separation after 40 min of incubation.
Figure 6. Emission spectra of (a) azide-50-NPs and (b) biotin-NPs before and after addition of 40 nM Alexa Fluor 633 labeled streptavidin
(SA-AF633) showing that only the presence of biotin induces energy transfer. The insets show photographs of the corresponding nanoparticle solutions
before (left) and after (right) the addition of 40 nM SA-AF633 under illumination with UV light. Energy transfer from biotin-NP to SA-AF633 leads to a
strong color change. Spectra were measured in phosphate buffer (with 0.1 mM CaCl2 and 0.1 mM MnCl2) at pH 7.0 using an excitation wavelength of
350 nm. The concentration of nanoparticles was 0.3 μM. See Figure S10 (Supporting Information) for complete titration curves.
17068
dx.doi.org/10.1021/ja2075345 |J. Am. Chem. Soc. 2011, 133, 17063–17071

Journal of the American Chemical Society
ARTICLE
Figure 8. Emission spectra of dual-targeting-NPs (0.3 μM), which contain both biotin and mannose upon addition of SA-AF633 (a; 1ꢀ5 nM) and of
ConA-AF633 (b; 0.8ꢀ15 nM) measured in phosphate buffer at pH 7.0 using an excitation wavelength of 350 nm. Energy transfer was observed for both
ligands.
confirmed by MALDI-ToF, where both the azide amphiphile 3
and the biotinylated amphiphile were detected. (Figure S7,
Supporting Information). Furthermore, DLS and TEM measure-
ments showed no significant changes in the size and shape of the
biotin-functionalized nanoparticles (biotin-NPs) compared with
azide-50-NP, as still spherical particles with a somewhat larger
hydrodynamic radius of 60 nm were observed (Figure S8,
Supporting Information). FRET studies between the postfunc-
tionalized biotin-NPs and Alexa Fluor 633 labeled streptavidin
(SA-AF633), revealed energy transfer from the fluorene ben-
zothiadiazole donor chromophore to the acceptor dye AF633.
No energy-transfer was observed for nonreacted azide-50-NPs
(Figure 6). The increase in acceptor emission upon binding to
the nanoparticles led to a strong color change of the nanoparticle
solution when illuminated with UV light, allowing simple optical
detection (Figure 6, insets). The addition of the lectin ConA-
AF633 did not lead to energy transfer (Figure S9, Supporting
Information), suggesting that nonspecific binding does not
occur. The energy transfer efficiency was higher for the strepta-
vidin-biotin-NP system than for the concanavalin A-mannose-
NP system (Figure 6 versus Figure 4). This might be explained
by the shorter distance between the biotin-NP and the Alexa
Fluor 633 dye attached to streptavidin, as SA is smaller than
ConA and additionally possesses a stronger binding constant to
its corresponding ligand.
The isolation of self-assembled postfunctionalized biotin-NPs
by magnetic streptavidin beads was also investigated.32 After
incubation with these beads and subsequent extraction using a
magnet, the solution of azide-50-NPs remained fluorescent,
suggesting no unspecific binding to the beads. In contrast, the
fluorescent biotin-NPs were not visible in solution, indicating
that the nanoparticles are effectively sequestered from solution
(Figure 7).33 The lack of fluorescence in solution simultaneously
underlines the stability of these self-assembled nanoparticles.
This result shows the ability to extract nanoparticles from
solution by magnetic beads. This should not only enable the
separation of specifically modified nanoparticles in solution
but also control circulation times in vivo via an avidin-induced
clearance strategy.34
which can be used for dual targeting with possible synergistic
effects.35 For this purpose, we combined our pre- and postfunc-
tionalization strategies. First, via self-assembly of prefunctionalized
oligomers, bifunctional-NPs consisting of 50% mannose-
functionalized amphiphile 2 and 50% azide-functionalized am-
phiphile 3 were prepared (Figure 1b,iii). To evaluate the surface
availability of azides for subsequent bioconjugation, the cycload-
dition reaction was performed using the alkyne derivative of biotin,
yielding mannose- and biotin-containing dual-targeting-NPs.
The dual targeting properties of these particles were evaluated
by measuring the energy transfer between these dual-target-
ing-NPs and both ConA-AF633 and SA-AF633. FRET studies
showed that the dual-targeting-NPs bind both to ConA and
to SA (Figure 8). When compared with the pure biotin-NPs
and mannose-50-NPs, a similar binding behavior was observed
(Figure 8 versus Figures 4 and 6), proving the same effective
introduction and availability of the targeting ligands. Our
results show that a portfolio of pre- and postfunctionalization
strategies significantly broadens the flexibility of these nano-
particles in terms of ligand modification and in multitargeting
applications.
’ CONCLUSIONS
We have demonstrated facile methods to generate (multi)-
targeted fluorescent nanoparticles based on the self-assembly of
πꢀconjugated amphiphilic oligomers in water. Azide and man-
nose groups were introduced at the periphery of the ethylene
glycol chains of the amphiphile. Using these building blocks, a
library of differently prefunctionalized nanoparticles was gener-
ated. Additionally, the nanoparticle prefunctionalization was
combined with subsequent postfuctionalization of azide-bearing
nanoparticles via the copper-catalyzed azideꢀalkyne cycloaddi-
tion (CuAAC) reaction, thus expanding the ligand diversity at
two independent stages in the nanoparticle fabrication process.
The intrinsic fluorescence of these organic nanoparticles enables
the simple evaluation of recognition events and surface properties
via FRET, showing, for example, their capability and potential for
multitargeting. Decoration of nanoparticles with specific ligands
induced the selective binding of proteins, bacteria, and magnetic
beads to the nanoparticles, which, in combinationwiththe absence
Dual-Targeting Nanoparticles. The self-assembly of π-con-
jugated oligomers enables direct access to bifunctional nanoparticles,
17069
dx.doi.org/10.1021/ja2075345 |J. Am. Chem. Soc. 2011, 133, 17063–17071

Journal of the American Chemical Society
ARTICLE
of unspecific adsorption and their stability, proves their broad
potential as selective biological targeting tools. The successfully
combined pre- and postfunctionalization of fluorescent oligomer
nanoparticles represents a novel opportunity to apply these versa-
tilenanoparticleswithreadilytailored opticalproperties to imaging
applications in biology and medicine.
Pu, K.-Y.; Liu, B.; Bazan, G. C.Chem. Mater. 2011, 23, 501–515. (i) Phillips,
R. L.; Kim, I.-B.; Tolbert, L. M.; Bunz, U. H. F. J. Am. Chem. Soc. 2008,
130, 6952–6954. For reviews on π-conjugated systems with biomolecules
to control self-assembly, see:(j) Jatsch, A.; Schillinger, E.-K.; Schmid, S.;
B€auerle, P. J. Mater. Chem. 2010, 20, 3563–3578. (k) Ruiz-Carretero, A.;
Janssen, P. G. A.; Kaeser, A.; Schenning, A. P. H. J. Chem. Commun. 2011,
47, 4340–4347.
(11) (a) Wu, C.; Schneider, T.; Zeigler, M.; Yu, J.; Schiro, P. G.;
Burnham, D. R.; McNeill, J. D.; Chiu, D. T. J. Am. Chem. Soc. 2010,
132, 15410–15417. (b) Wu, C.; Jin, Y.; Schneider, T.; Burnham, D. R.;
Smith, P. B.; Chiu, D. T. Angew. Chem., Int. Ed. 2010, 49, 9436–9440.
(12) Wu, C.; Hansen, S. J.; Hou, Q.; Yu, J.; Zeigler, M.; Jin, Y.;
Burnham, D. R.; McNeill, J. D.; Olson, J. M.; Chiu, D. T. Angew. Chem.,
Int. Ed. 2011, 50, 3430–3434.
’ ASSOCIATED CONTENT
S
Supporting Information. Materials, methods, detailed
b
experimental procedures (synthesis and biological assays), com-
pound characterization, including copies of NMR spectral data
and supplementary figures. This material is available free of
charge via the Internet at http://pubs.acs.org.
(13) Kim, B.-S.; Hong, D.-J.; Bae, J.; Lee, M. J. Am. Chem. Soc. 2005,
127, 16333–16337.
(14) Ryu, J.-H.; Lee, E.; Lim, Y.-beom; Lee, M. J. Am. Chem. Soc.
2007, 129, 4808–4814.
’ AUTHOR INFORMATION
(15) Abbel, R.; van der Weegen, R.; Meijer, E. W.; Schenning, A. P.
H. J. Chem. Commun. 2009, 1697–1699.
(16) Abbel, R.; van der Weegen, R.; Pisula, W.; Surin, M.; Leclꢁere, P.;
Lazzaroni, R.; Meijer, E. W.; Schenning, A. P. H. J. Chem.—Eur. J. 2009,
15, 9737–9746.
(17) Zalipsky, S. Adv. Drug Delivery Rev. 1995, 16, 157–182.
(18) (a) Caliceti, P.; Veronese, F. M. Adv. Drug Delivery Rev. 2003,
55, 1261–1277. (b) Karakoti, A. S.; Das, S.; Thevuthasan, S.; Seal, S.
Angew. Chem., Int. Ed. 2011, 50, 1980–1994.
(19) M€uller, M. K.; Brunsveld, L. Angew. Chem., Int. Ed. 2009, 48,
2921–2924.
Corresponding Author
a.p.h.j.schenning@tue.nl; l.brunsveld@tue.nl
’ ACKNOWLEDGMENT
We thank Dr. Pol Besenius for help with TEM studies and Ralf
Bovee for MALDI-ToF MS measurements. This research has
been made possible by a ERC grant 204554ꢀSupraChemBio and
The Netherlands Foundation for Scientific Research (NWO) by a
VICI grant.
(20) Ryu, J.-H.; Hong, D.-J.; Lee, M. Chem. Commun. 2008,
1043–1054.
’ REFERENCES
(21) Karlsson, K. A. Biochem. Soc. Trans. 1999, 27, 471–474.
(22) Concanavalin A as a tool, Bittiger, H.; Schnebli, H. P., Ed.; John
Wiley & Sons: New York, 1976.
(23) (a) Rostovtsev, V. V.; Green, L. G.; Fokin, V. V.; Sharpless, K. B.
Angew. Chem., Int. Ed. 2002, 41, 2596–2599. (b) Tornøe, C. W.;
Christensen, C.; Meldal, M. J. Org. Chem. 2002, 67, 3057–3064.
(24) Kasai, H.; Nalwa, H. S.; Oikawa, H.; Okada, S.; Matsuda, H.;
Minami, N.; Kakuta, A.; Ono, K.; Mukoh, A.; Nakanishi, H. Jpn. J. Appl.
Phys. 1992, 31, L1132–L1134.
(1) (a) Davis, M. E.; Chen, Z.; Shin, D. M. Nat. Rev. Drug Discovery
2008, 7, 771–782. (b) Subbiah, R.; Veerapandian, M.; Yun, K. S. Curr.
Med. Chem. 2010, 17, 4559–4577. (c) Shi, J.; Votruba, A. R.; Farokhzad,
O. C.; Langer, R. Nano Lett. 2010, 10, 3223–3230.
(2) (a) Zhou, Y.; Drummond, D. C.; Zou, H.; Hayes, M. E.; Adams,
G. P.; Kirpotin, D. B.; Marks, J. D. J. Mol. Biol. 2007, 371, 934–947. (b)
Montet, X.; Funovics, M.; Montet-Abou, K.; Weissleder, R.; Josephson,
L. J. Med. Chem. 2006, 49, 6087–6093.
(3) Bangham, A. D.; Horne, R. W. J. Mol. Biol. 1964, 8, 660–668.
(4) (a) Medintz, I. L.; Uyeda, H. T.; Goldman, E. R.; Mattoussi, H.
Nat. Mater. 2005, 4, 435–446. (b) Zrazhevskiy, P.; Sena, M.; Gao, X.
Chem. Soc. Rev. 2010, 39, 4326–4354.
(5) (a) Wu, C.; Bull, B.; Szymanski, C.; Christensen, K.; McNeill, J.
ACS Nano 2008, 2, 2415–2423. (b) Tuncel, D.; Demir, H. V. Nanoscale
2010, 2, 484–494. (c) Pecher, J.; Mecking, S. Chem. Rev. 2010, 110,
6260–6279.
(25) Disney, M. D.; Zheng, J.; Swager, T. M.; Seeberger, P. H. J. Am.
Chem. Soc. 2004, 126, 13343–13346.
(26) Gu, F.; Zhang, L.; Teply, B. A.; Mann, N.; Wang, A.; Radovic-
Moreno, A. F.; Langer, R.; Farokhzad, O. C. Proc. Natl. Acad. Sci. U. S. A.
2008, 105, 2586–2591.
(27) DLS and TEM studies revealed aggregation for nanoparticles
consisting of pure (100%) 2 only, probably through intermolecular
hydrogen bonding of carbohydrates (see ref 10j). Reduction to 90% was
found to be enough to overcome unspecific aggregation, as evidenced
by DLS.
(6) Biju, V.; Itoh, T.; Ishikawa, M. Chem. Soc. Rev. 2010, 39,
3031–3056.
(28) This strong multivalent binding manifested itself as well during
the incubation with E. coli, which led to a strong cross-linking of the
bacterial pellet.
(29) (a) Kim, I.-K.; Wilson, J. N.; Bunz, U. H. F. Chem. Commun.
2005, 1273–1275. (b) Schmid, S.; Mishra, A.; B€auerle, P. Chem.
Commun. 2011, 47, 1324–1326.
(7) Oh, J. K. J. Mater. Chem. 2010, 20, 8433–8445.
(8) Mulder, W. J. M.; Strijkers, G. J.; van Tilborg, G. A. F.; Cormode,
D. P.; Fayad, Z. A.; Nicolay, K. Acc. Chem. Res. 2009, 42, 904–914.
(9) Strijkers, G. J.; Kluza, E.; Tilborg, G. A. F.; Schaft, D. W. J.;
Griffioen, A. W.; Mulder, W. J. M.; Nicolay, K. Angiogenesis 2010,
13, 161–173.
(30) (a) Cutler, J. I.; Zheng, D.; Xu, X.; Giljohann, D. A.; Mirkin,
C. A. Nano Lett. 2010, 10, 1477–1480. (b) van Dongen, S. F. M.; Nallani,
M.; Schoffelen, S.; Cornelissen, J. J. L. M.; Nolte, R. J. M.; van Hest,
J. C. M. Macromol. Rapid Commun. 2008, 29, 321–325.
(31) The figures for postfunctionalized nanoparticles represent
typical curves obtained after postfunctionalization. In contrast to
reproducible ligand densities achieved via the prefunctionalization
approach, the reproducibility of ligand densities via the postfunctiona-
lization approach was less.
(10) For some articles on π-conjugated nanoparticles, see: (a)
Moon, J. H.; McDaniel, W.; MacLean, P.; Hancock, L. F. Angew. Chem.,
Int. Ed. 2007, 46, 8223–8225. (b) Wu, C.; Szymanski, C.; Cain, Z.;
McNeill, J. J. Am. Chem. Soc. 2007, 129, 12904–12905. (c) Kaeser, A.;
Schenning, A. P. H. J. Adv. Mater. 2010, 22, 2985–2997. For reviews and
articles about molecularly dissolved π-conjugated polymers and oligo-
mers for imaging and sensing, see:(d) Feng, X.; Liu, L.; Wang, S.; Zhu, D.
Chem. Soc. Rev. 2010, 39, 2411–2419. (e) Herland, A.; Ingan€as, O.
Macromol. Rapid Commun. 2007, 28, 1703–1713. (f) Thomas, S. W.;
Joly, G. D.; Swager, T. M. Chem. Rev. 2007, 107, 1339–1386. (g) Xue, C.;
Velayudham, S.; Johnson, S.; Saha, R.; Smith, A.; Brewer, W.; Murthy, P.;
Bagley, S. T.; Liu, H. Chem.—Eur. J. 2009, 15, 2289–2295. (h) Duarte, A.;
(32) Haukanes, B.-I.; Kvam, C. Nat. Biotechnol. 1993, 11, 60–63.
(33) The intrinsic nanoparticle fluorescence can be used to deter-
mine the optimal biotin-NPs/magnetic streptavidin beads ratio to
17070
dx.doi.org/10.1021/ja2075345 |J. Am. Chem. Soc. 2011, 133, 17063–17071

Journal of the American Chemical Society
ARTICLE
completely remove the nanoparticles from solution (Figure S11, Sup-
porting Information).
(34) Kobayashi, H.; Kawamoto, S.; Star, R. A.; Waldmann, T. A.;
Brechbiel, M. W.; Choyke, P. L. Bioconjugate Chem. 2003, 14, 1044–1047.
(35) Kluza, E.; van der Schaft, D. W. J.; Hautvast, P. A. I.; Mulder,
W. J. M.; Mayo, K. H.; Griffioen, A. W.; Strijkers, G. J.; Nicolay, K. Nano
Lett. 2010, 10, 52–58.
17071
dx.doi.org/10.1021/ja2075345 |J. Am. Chem. Soc. 2011, 133, 17063–17071