Synthesis of pyridyl disulfide-functionalized nanoparticles for conjugating thiol-containing small molecules, peptides, and proteins

Article abstract of DOI:10.1021/bc9004443

Previously we reported emulsion polymerization of propylene sulfide with Pluronic F127 as an emulsifier, yielding nanoparticles (NPs) in the 25 nm size range. Immunologically functional NPs were prepared by adding an antigen-Pluronic conjugate to the polymerization mixture (Reddy, S. T., et al. (2007) Nat. Biotechnol. 25, 1159). We sought a more flexible scheme for conjugation of antigens and other biomolecules to the NP surfaces that would allow for milder reaction conditions than achievable during the polymerization step. Here, we present the synthesis of such functionalizable NPs in the form of NPs that carry thiol-reactive groups, to which thiol-containing antigens (peptide or protein) or other biomolecules can be conjugated under mild conditions to yield immunofunctional NPs. The Pluronic-stabilized poly(propylene sulfide) (PPS) NPs with thiol-reactive pyridyl disulfide groups are prepared in two steps by (1) emulsion polymerization of propylene sulfide in the presence of a carboxylate-Pluronic and (2) reaction of the carboxylic acid groups on the NP surface with cysteamine pyridyl disulfide and a water-soluble carbodiimide reagent. We choose pyridyl disulfide groups to have a reduction-sensitive disulfide bond linking the antigen to the NP surface, allowing efficient release of antigen inside the cell in response to the reductive conditions within the endosome. The functionalizable NPs are characterized by proton NMR, dynamic light scattering (DLS), UV/vis spectroscopy, and transmission electron microscopy (TEM). Conjugation of small molecules and protein to the NP surface is presented.

Full text of DOI:10.1021/bc9004443

Bioconjugate Chem. 2010, 21, 653–662
653
Synthesis of Pyridyl Disulfide-Functionalized Nanoparticles for Conjugating
Thiol-Containing Small Molecules, Peptides, and Proteins
Andre´ J. van der Vlies,† Conlin P. O’Neil,† Urara Hasegawa,† Nathan Hammond,† and Jeffrey A. Hubbell*,†,‡
Institute of Bioengineering (IBI) and Institute of Chemical Sciences and Engineering (ISIC), Ecole Polytechnique Fe´de´rale de
Lausanne (EPFL), Lausanne CH-1015, Switzerland. Received October 10, 2009; Revised Manuscript Received March 25, 2010
Previously we reported emulsion polymerization of propylene sulfide with Pluronic F127 as an emulsifier, yielding
nanoparticles (NPs) in the 25 nm size range. Immunologically functional NPs were prepared by adding an
antigen-Pluronic conjugate to the polymerization mixture (Reddy, S. T., et al. (2007) Nat. Biotechnol. 25, 1159).
We sought a more flexible scheme for conjugation of antigens and other biomolecules to the NP surfaces that
would allow for milder reaction conditions than achievable during the polymerization step. Here, we present the
synthesis of such functionalizable NPs in the form of NPs that carry thiol-reactive groups, to which thiol-containing
antigens (peptide or protein) or other biomolecules can be conjugated under mild conditions to yield
immunofunctional NPs. The Pluronic-stabilized poly(propylene sulfide) (PPS) NPs with thiol-reactive pyridyl
disulfide groups are prepared in two steps by (1) emulsion polymerization of propylene sulfide in the presence of
a carboxylate-Pluronic and (2) reaction of the carboxylic acid groups on the NP surface with cysteamine pyridyl
disulfide and a water-soluble carbodiimide reagent. We choose pyridyl disulfide groups to have a reduction-
sensitive disulfide bond linking the antigen to the NP surface, allowing efficient release of antigen inside the cell
in response to the reductive conditions within the endosome. The functionalizable NPs are characterized by proton
NMR, dynamic light scattering (DLS), UV/vis spectroscopy, and transmission electron microscopy (TEM).
Conjugation of small molecules and protein to the NP surface is presented.
adjuvant delivery (5). Immobilizing an antigen onto or encap-
sulating it within the polymeric vehicle can protect the antigen
from enzymatic degradation, and the carrier itself can be
delivered intracellularly, enabling antigen processing and pre-
sentation to take place. Most work has focused on targeting
peripheral immature DCs using polymeric poly(lactide-co-
glycolide) (PLGA) microparticles (6-12) and polyelectrolyte
microcapsules (13). However, these cells must travel to the
draining lymph nodes to present the antigen to T cells. We
previously showed that immature DCs in the lymph node can
be targeted efficiently using ultrasmall nanoparticles (NPs) of
about 25 nm in diameter exploiting the interstitial flow from
the blood capillaries to the lymphatic capillaries and then
draining lymph nodes (14). In this way the most potent APCs
in the lymph node are targeted directly and very efficiently.
INTRODUCTION
In the field of immunotherapy, one aims to either elicit or
repress an immune response by directly affecting the immune
system. Antigen-presenting cells (APCs§) patrol the body for
antigens, the most powerful APCs being dendritic cells
(DCs) (1, 2). After taking up the antigen, the DCs move to the
draining lymph node to present peptides processed from the
antigen and stimulate naive CD4 and CD8 T cells for induction
of humoral and cell-mediated immunity (3). In order to induce
an adaptive immune response, the DC should mature before
presenting antigen to T cells. Maturation of DCs is induced by
the presence of a danger signal, also referred to as an adjuvant,
often selected as a ligand of Toll-like receptors (4).
On the basis of the above-mentioned needs, there has been
growing interest in the use of biomaterials in both antigen and
Poly(propylene sulfide) (PPS) NPs are synthesized by emul-
sion polymerization of propylene sulfide using Pluronic
F127 an emulsifier surfactant (15, 16). The NPs consist of a
hydrophobic rubbery cross-linked PPS core with a Pluronic-
coated surface. The hydrophobic PPS core allows for encapsula-
tion of hydrophobic drugs, should it be desired, and is stimulus-
responsive toward oxidation (16, 17). We showed that exposure
to hydrogen peroxide as a model intracellular oxidant leads to
oxidation of the thioether groups leading to sulfoxide and sulfone
structures, making the initial hydrophobic PPS core hydrophilic
and thus soluble (16). This allows the incorporated drug to be
released and offers a mechanism for particle clearance from the
body (17). To show that these NPs are not only taken up by
lymph-node-resident DCs but also induce adaptive immunity,
we used ovalbumin (OVA) as a model antigen. OVA-NPs,
prepared by incorporating an OVA-Pluronic conjugate into the
NP polymerization mixture, induced both a humoral and cell-
mediated immune response when injected into mice (18), the
terminal Pluronic hydroxyl groups on the surface of the NPs
activating complement to serve as an adjuvant (18).
* To whom correspondence should be addressed. Address: Labora-
tory for Regenerative Medicine and Pharmacobiology, Institute of
Bioengineering, Station 15, Ecole Polytechnique Fe´de´rale de Lausanne
(EPFL), Lausanne, CH-1015, Switzerland. E-mail: jeffrey.hubbell@
epfl.ch. Phone: +41 21 693 9681. Fax: +41 21 693 9665.
† Institute of Bioengineering (IBI).
‡ Institute of Chemical Sciences and Engineering (ISIC).
§Abbreviations: APC, antigen presenting cell; DBU, 1,8-diazabicyclo-
[5.4.0]undec-7-ene; DC, dendritic cell; DLS, dynamic light scattering;
DMF, dimethylformamide; DMSO, dimethyl sulfoxide; DTT, dithio-
threitol; EDC, N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide; HEPES,
hydroxyethylpiperazineethanesulfonic acid; HPLC, high performance
liquid chromatography; HRMS, high resolution mass spectra; NMR,
nuclear magnetic resonance; MeOH, methanol; MES, morpholineethane-
sulfonic acid; MWCO, molecular weight cut-off; NHS, N-hydroxysuc-
cinimide; NP, nanoparticle; OVA, ovalbumin; PLGA, poly(lactide-co-
glycolide); PO, propylene oxide; PPO, polypropylene oxide; PPS,
polypropylene sulfide; sulfo-NHS, N-hydroxysulfosuccinimide sodium
salt; TCEP ·HCl, tris(2-carboxyethyl)phosphine hydrochloride; TEM,
transmission electron microscopy; TFA, trifluoroacetic acid; TIS,
triisopropylsilane.
10.1021/bc9004443  2010 American Chemical Society
Published on Web 04/06/2010

654 Bioconjugate Chem., Vol. 21, No. 4, 2010
van der Vlies et al.
A limitation of the scheme described above is the need for
synthesizing a Pluronic-antigen conjugate for each antigen to
be bound to the NP surface. Furthermore, the antigen is linked
to the NP with a bond that is not cleavable under biological
conditions, which could be problematic for antigen release,
especially of peptides, within DCs. We therefore sought a
scheme by which to provide NPs with a reactive group that
would allow straightforward conjugation of antigens and further
allow release upon endocytosis by DCs. We choose thiol-
reactive pyridyl disulfide groups at the surface of the NP, as
antigens such as peptides and proteins can be easily engineered
to contain free cysteine groups. The formed disulfide bond
between the antigen and NP would allow efficient release of
antigen within the cell in response to the reductive conditions
within the endosome. This report describes the concept,
synthesis, and characterization of thiol-reactive, reductively
sensitive NPs and the conjugation with low molecular weight
compounds and proteins.
electrospay ion source and operating in the positive ionization
mode. The samples were diluted in methanol acidified with 0.1%
HCOOH and introduced into the mass spectrometer by infusion
at a flow rate of 20 mL/min. External calibration was carried
out with a solution of phosphoric acid at 0.01%. Data were
processed using the MassLynx 4.1 software.
Transmission Electron Microscopy. Carbon coated 400 mesh
copper grids (Agar Scientific, Stansted, U.K.) were prepared
under glow discharge, and 5 µL of NP solution was placed onto
the grid. After 60 s, the grid was washed with H2O and then
dried by blotting the side of the grid with filter paper. To stain,
5 µL of 2% uranyl acetate solution was added to the grid. After
30 s, the stain was blotted off the same as above and was
allowed to dry for 5 min at room temperature before analysis.
Microscope images were acquired on a Philips CM-12 TEM
(Koninklijke Philips Electronics, Eindhoven, The Netherlands).
Images were captured using a Gatan MSC CCD camera (Gatan
Inc., Grandchamp, France) and processed using Adobe Photo-
shop software (Adobe Systems, San Jose, CA). TEM images
were analyzed with Scion image analysis software, and size
results are given as the mean diameter ( SD and number of
particles measured in parentheses.
PreparatiVe HPLC. Purification by HPLC was done by
reverse phase HPLC chromatography using a C18 column.
Separation and collection were performed by UV and mass
directed software (Waters, Baden-Daettwil, Switzerland).
SDS-PAGE. The 10% gels were run at 300 V for 2 h. Protein
bands were visualized by UV.
Ellman Assay (19). Ellman solution was prepared by dis-
solving 4.0 mg of Ellman reagent in 1.0 mL of Ellman buffer
(1 mM EDTA, 100 mM phosphate, pH 8.0). Thiol concentra-
tions were measured by reacting 20 µL of Ellman solution with
20 µL of sample. After 15 min the mixture was diluted by the
addition of 960 µL of Ellman buffer and the absorbance at 412
nm read. Thiol concentrations were calculated from a standard
curve using thiomaleic acid. In the range 1-4 mM measure-
ments were done in a well plate using 200 µL of sample. For
thiol concentrations in the range 50-1000 µM, measurements
were done using a quartz cuvette.
EXPERIMENTAL PROCEDURES
Materials and Methods. Pluronic F127, hydroxyethylpip-
erazineethanesulfonic acid (HEPES) (Sigma), streptavidin-
Alexa488 (Invitrogen), (+)-D-biotin (Applichem), aldrithiol,
fluorescamine, trifluoracetic acid (TFA), triisopropylsilane (TIS),
phenol, 0.5 M NaOCH3 in methanol, propylene sulfide, 1,8-
diazabicyclo[5.4.0]undec-7-ene (DBU), N-hydroxysuccinimide
(NHS),Tris(2-carboxyethyl)phosphinehydrochloride(TCEP ·HCl),
iodoacetamide, potassium hydroxide, CaH2 (Aldrich), dithio-
threitol (DTT) (Biorad), Sephadex G50, Sepharose 6B, Sepharose
CL-6B (Sigma), Ellman reagent (Pierce), cysteamine hydro-
chloride, potassium carbonate, mercaptopropionic acid methyl
ester, triphenylphosphine, acetic acid, methanesulfonyl chloride,
sodium hydroxide, N-(3-dimethylaminopropyl)-N′-ethylcarbo-
diimide (EDC), N-hydroxysulfosuccinimide sodium salt (sulfo-
NHS), triphenylmethanol, BF3 ·Et2O (Fluka), and morpho-
lineethanesulfonic acid (MES) (Calbiochem) were used as
received. All chemicals relating to peptide synthesis were from
Novabiochem. Molecular sieves (3 Å, Fluka) were heated at
180 °C under reduced pressure for at least 16 h. Diethyl ether,
methanol, and toluene were used as received. Dimethylforma-
mide (DMF) was distilled over CaH2 under reduced pressure
at 40 °C and kept over molecular sieves. Dichloromethane was
dried over molecular sieves. Water was distilled twice or freed
from salt using the Milli-Q water system. All reactions were
done under argon (Messer) atmosphere.
Fluorescamine Assay (20). Fractions from the Sepharose 6B
and Sepharose CL-6B column (50 µL) were mixed with 100
µL of 100 mM HEPES, pH 8, and 20 µL of a solution of
fluorescamine (5 mg/mL) in acetone. The sample was reacted
at room temperature for 15 min and the fluorescence (λex
)
390 nm, λem ) 475 nm) read using black polystyrene Nunc
1H NMR. Spectra were measured on a Bruker 400 Ultrashield
spectrometer. NMR was performed in deuterated chloroform
(CDCl3), deuterated methanol (CD3OD), or deuterated dimethyl
sulfoxide (DMSO-d6). Peaks are referenced against residual
solvent at 7.26 ppm (CDCl3), 3.35 ppm (CH3OD), or 2.50 ppm
(DMSO-d6). For each sample 32 scans were collected and the
D1 was set at 10 s.
UV/Visible Spectroscopy. Spectra were obtained on a Shi-
madzu UV mini 1240 UV/visible spectrophotometer using 1.0
cm quartz cuvettes or a Tecan Safire2 well plate reader using
polystyrene well plates.
plates.
Pyridyl Disulfide Loading NPs. Pyridyl-NP solutions were
mixed 1:1 by volume with a 5 mM TCEP·HCl solution in 100
mM HEPES, pH 8, and reacted for 15 min. The absorbance at
340 nm was read, and after correction for the absorbance of
pyridyl-NP and TCEP ·HCl the concentration of pyridyl disulfide
calculated from a standard curve of pyridine thione (8) in
HEPES, pH 8. If the concentration was too high, NP solutions
were diluted with H2O.
Compound Synthesis. Pluronic Bromide (1). 1 was prepared
as reported before (21).
Dynamic Light Scattering. Data were obtained on a Malvern
Nanoseries Zetasizer instrument with 632 nm laser using PMMA
cuvettes by diluting NP solutions with PBS (10 mM, pH 7.4)
to a final NP concentration of 2 mg/mL. Sizes are given in nm
and are obtained as the Z-average by fitting the correlation
function using the cumulant method.
High Resolution Mass Spectrometry. Spectra were collected
by the Institute of Chemical Sciences and Engineering (ISIC)
at the Ecole Polytechnique Fe´de´rale de Lausanne (EPFL),
Switzerland. ESI-TOF-MS data were acquired on a Q-Tof
Ultima mass spectrometer (Waters) fitted with a LockSpray dual
Pluronic Mesylate (2). An amount of 23.6 g (1.86 mmol) of
Pluronic F127 was dissolved in 500 mL of toluene, and 100
mL of toluene was removed by azeotropic distillation using a
Dean-Stark trap. The solution was cooled to room temperature,
and 2.6 mL (18.6 mmol, 10 equiv) of triethylamine and 1.4
mL (18.6 mmol) of methanesulfonyl chloride was added. After
18 h the yellow suspension was filtered and the clear yellow
filtrate concentrated under reduced pressure. The residue was
dissolved in 100 mL of CH2Cl2, filtered over a SiO2 bed (1.0
cm (diameter) × 5.0 cm (height)), and the SiO2 was washed
with CH2Cl2 (3 × 50 mL). The filtrate was concentrated to a

Pyridyl Disulfide-Functionalized Nanoparticles
Bioconjugate Chem., Vol. 21, No. 4, 2010 655
Table 1. Dynamic Light Scattering Results for the Different NP
Formulationsa
volume of about 50 mL and precipitated in 1000 mL of Et2O.
The white precipitate was filtered off, washed with 100 mL of
Et2O, and dried under reduced pressure to yield 21.0 g (89%)
of white powder. The degree of functionalization was 100%
according to H NMR. H NMR (CDCl3): δ ) 4.35-4.38 (m,
CH2-O-SO2-), 3.60-3.67 (m, CH2, PEG), 3.47-3.58 (m, CH2,
PPO), 3.34-3.42 (m, CH, PPO), 3.07 (s, -SO2-CH3), 1.11-1.15
(CH3, PPO).
carboxylate-NP
pyridyl disulfide-NP
wt % COOH Z-avgb (SD) PDIc (SD) Z-Avgb (SD) PDIc (SD)
1
1
0
31.8 (0.80) 0.260 (0.003)
10
25
50
75
100
33.0 (0.31) 0.142 (0.006) 35.7 (0.13) 0.207 (0.005)
31.0 (0.24) 0.168 (0.013) 33.7 (0.71) 0.175 (0.003)
32.6 (0.08) 0.147 (0.004) 42.5 (0.30) 0.225 (0.009)
32.9 (0.18) 0.145 (0.006) 42.0 (0.15) 0.219 (0.002)
31.3 (0.22) 0.113 (0.001) 42.4 (0.17) 0.209 (0.004)
Pluronic Thioether Propionic Acid Methyl Ester (3). An
amount of 21.0 g (2) was dissolved in 100 mL of DMF, and to
the clear solution was added 2.29 g (16.5 mmol) of potassium
carbonate followed by 2 mL (16.5 mmol) of mercaptopropionic
acid methyl ester. The mixture was stirred for 20 h and
concentrated under reduced pressure. The residue was dissolved
in 100 mL of CH2Cl2 and mixed with 11 g of activated carbon.
The mixture was filtered over a filter cell and the residue washed
with CH2Cl2 (3 × 100 mL). The clear colorless filtrate was
reduced to about 50 mL by evaporation and precipitated in 1000
mL of Et2O. The white solid was filtered, washed with 100 mL
of Et2O, and dried under reduced pressure to yield 18.2 g (87%).
The degree of functionalization was 92% according to 1H NMR.
1H NMR (CDCl3): δ ) 3.70 (s, OCH3), 3.59-3.63 (m, CH2,
PEG), 3.46-3.58 (m, CH2, PPO), 3.36-3.44 (m, CH, PPO),
2.83 (t, J ) 7.6 Hz, -O-CH2-CH2-S-), 2.73 (t, J ) 7.2 Hz, -S-
CH2-CH2-CO-), 2.63 (t, J ) 7.2 Hz, -CH2-CO-), 1.12-1.16
(CH3, PPO).
a Values are the average of three measurements of the same sample.
b Size in nm as measured by the Z-average obtained by fitting the
correlation function using the cumulant method. c Polydispersity index
(PDI) calculated from fitting the correlation function using the cumulant
method.
2.74 (t, CH2, J ) 6.8 and 6.4 Hz), 2.51 (t, CH2, J ) 7.2 and 7.6
Hz), 1.14 (m, CH3, PPO).
Thioacetic Acid S-[3-[3-(3-Acetylsulfanylpropoxy)-2,2-bis-
(3-acetylsulfanylpropoxymethyl)propoxy]propyl] Ester (7). The
four-arm initiator was synthesized as reported (16).
NP Synthesis. This was adapted from that reported (16). An
amount of 10 mL of Milli-Q H2O was freed from oxygen by
three evacuation/purge (Ar) cycles, and under Ar flow the
hydroxyl- and carboxylate Pluronic (500 mg total) were added.
After all the Pluronic was dissolved the resulting solution was
again evacuated and purged with Ar three times, and under Ar
flow 400 µL (5.0 mmol) of propylene sulfide was added. The
emulsion was stirred for 30 min after which 24.3 mg (0.04
mmol) of the four-arm initiator 7, previously reacted for 15 min
with 322 µL of 0.5 M NaOCH3 (0.16 mmol) solution in MeOH,
was added under Ar. The mixture was stirred for another 15
min, and then an amount of 60 µL of DBU (0.4 mmol) was
added. The mixture was stirred for 24 h and exposed to air for
2 h to cross-link, and the remaining thiolates were quenched
by the addition of 47.5 mg (0.26 mmol) of iodoacetamide and
stirred for 2 h. Ellman assay of the NP solution confirmed the
absence of thiolate anions, and the mixture was dialyzed
(MWCO ) 100 kDa) against 5 L of Milli-Q water for at least
2 days with regular replacement of the water. After filtration
(0.22 µm filter) the size of the NPs was measured by dynamic
light scattering and a small volume was lyophilized to obtain
the NP concentration in mg/mL. NMR of the solid was used to
calculate the weight % of Pluronic in the NPs.
Pluronic Thioether Propionic Acid Sodium Salt (4). An
amount of 18.0 g (3) (2.83 mmol) was dissolved in 200 mL of
Milli-Q water, and to the clear solution added a solution of
0.57 g (14.1 mmol) of sodium hydroxide in 5 mL of Milli-Q
water. After 21 h the solution was transferred into a regenerated
cellulose dialysis tube, molecular weight cut-off (MWCO) of
3.5 kDa, and dialyzed against 5 L of Milli-Q for 29 h with
replacement of the water three times. The solution was lyoph-
ilized to yield 12.2 g (68%) of white solid. The degree of
functionalization according to 1H NMR was 95%. NMR
(CDCl3): δ ) 3.59-3.63 (m, CH2, PEG), 3.46-3.58 (m, CH2,
PPO), 3.35-3.44 (m, CH, PPO), 2.83 (t, J ) 6.4 Hz, -O-CH2-
CH2-S-), 2.73 (t, J ) 6.8 Hz, -S-CH2-CH2-CO-), 2.50, (t, J )
6.8 Hz, -CH2-CO-), 1.13-1.16 (CH3, PPO).
Pyridyl Disulfide Cysteamine HCl Salt (5). An amount of
0.57 g (0.5 mmol) of cysteamine ·HCl was added to a solution
of 3.3 g (15.0 mmol) of aldrithiol in 10 mL of methanol. The
yellow solution was stirred overnight and the product obtained
by precipitation in 500 mL of Et2O. After filtration, the residue
was dissolved in 15 mL of MeOH and again precipitated in
500 mL of Et2O. The white solid was dried under reduced
Pyridinethione (8). The ether fractions of 5 were combined
and concentrated under reduced pressure, and the residue was
dissolved in 10 mL of MeOH and reacted with 1.74 g (15.0
mmol) of cysteamine ·HCl overnight. The turbid mixture was
diluted with 100 mL of EtOAc and washed with 100 mL of
H2O. After the organic phase was dried over Na2SO4 the mixture
was filtered and the clear solution concentrated under reduced
pressure. The residue was purified by SiO2 column chromatog-
1
pressure to yield 0.94 g (84%). H NMR (DMSO-d6): δ )
8.51-8.50 (m, 1H), 8.30 (bs, 3H), 7.86-7.82 (m, 1H), 7.75 (d,
1H), 7.31-7.28 (m, 1H), 3.13-3.05 (m, 4H). HRMS [M +
H]+ calculated for C7H10N2S2 187.0364, found 187.0358.
1
raphy, Rf ) 0.46 (EtOAc). H NMR (CDCl3): δ ) 7.61-7.56
Pluronic Pyridyl Disulfide (6). Amounts of 0.25 g (0.02
mmol) of 4 and 22.8 mg (0.2 mmol) of NHS were dissolved in
5 mL of CH2Cl2. To the clear solution were added 38 mg (0.2
mmol) of EDC, 44 mg (0.2 mmol) of 5, and 38 µL (0.2 mmol)
of triethylamine, and the turbid mixture was stirred for 18 h at
room temperature. The yellow solution was concentrated in a
flow of N2 and the residue dissolved in 5 mL of H2O. The milky
solution was transferred to a dialysis bag (MWCO ) 3.5 kDa)
and dialyzed against 2 L of H2O for 1 day. The polymer was
recovered by lyophilization to yield 0.19 g (76%). The degree
of functionalization according to 1H NMR was 95%. 1H NMR
(CDCl3): δ ) 8.51 (m, CHaromat), 7.64 (m, CHaromat), 7.55 (m,
CHaromat), 7.33 (bs, NH), 7.13 (m, CHaromat), 3.69-3.59 (m, CH2,
PEG), 3.58-3.44 (m, CH2, PPO), 3.42-3.33 (m, CH, PPO),
2.94 (t, CH2, J ) 5.6 and 6.0 Hz), 2.88 (t, CH2, J ) 7.2 Hz),
(m, 1H), 7.43-7.39 (m, 1H), 6.81-6.78 (m, 1H), 1.79 (bs, 1H).
HRMS [M + H]+ calculated for C5H5NS 112.0221, found
112.0297.
Synthesis of Pyridyl Disulfide-NPs, General Procedure. To
the carboxylate-NP solution is added solid MES to a final
concentration of 100 mM and the pH adjusted to 4.0-4.5 with
a pH meter. To the clear solution are added pyridyl disulfide
cysteamine·HCl, sulfo-NHS, and EDC all at an excess of 20
equiv relative to the amount of Pluronic as calculated from
NMR. After reaction for 2-24 h the solution is dialyzed for at
least 2 days to remove unreacted reagents using MWCO )
3.5-10 kDa. The pyridyl disulfide-NPs in Table 1 and Figures
2-4 were made by reacting for 24 h and then purified using
MWCO ) 3.5 kDa dialysis tubing. Below follows the detailed
reaction for functionalization of 100 wt % carboxylate-NPs. To

656 Bioconjugate Chem., Vol. 21, No. 4, 2010
van der Vlies et al.
3.0 mL of 100 wt % carboxylate-NP solution (50.4 mg/mL),
corresponding to 28.7 mg/mL of Pluronic according to NMR,
was added 60 mg of MES (final concentration 102 mM), and
the pH was adjusted to 4.4. To the NP solution containing 86.1
mg of carboxylate-Pluronic (0.014 mmol of carboxylate groups)
were added 30.1 mg (0.14 mmol, 10 equiv) pyridyl disulfide
cysteamine·HCl salt, 30.1 mg of sulfo-NHS (0.14 mmol, 10
equiv), and 33.1 mg (0.17 mmol, 12 equiv) of EDC. After 22 h
the NP solutions were dialyzed against 5 L of Milli-Q water
(MWCO ) 3.5 kDa) for 2 days. After filtration through 0.22
µm filter, the size was measured by dynamic light scattering.
The concentration of the pyridyl disulfide groups was determined
by measuring the absorbance at 340 nm after reducing the NPs
with TCEP ·HCl. Lyophilization of a small sample yields the
concentration of the NP solution.
of biotin on the NPs was confirmed by the HABA/avidin
displacement assay (22). The number of biotin molecules per
NP was estimated to be approximately 800 (see Supporting
Information).
Ac-SDKDSLKCG-OH Peptide (11). Fmoc-Gly (0.65 mmol,
194.0 mg, 3 equiv) was coupled to NovaPEG Wang resin (250
mg, loading 0.87 mmol/g) with 134.6 mg (0.65 mmol, 3 equiv)
of DCC and 13.2 mg (0.11 mmol, 0.5 equiv) of DMAP in 10
mL of DMF. After the resin was washed with DMF (5 × 10
mL) the Fmoc group was removed with 20% piperidine in DMF
(10 mL) and after washing with DMF (5 × 10 mL) used to
couple the other amino acids using standard Fmoc-amino acid/
HBTU/HOBt/DIPEA (0.5 M/0.5 M/0.5 M/1.0 M, 5 equiv
solutions in DMF) double coupling Fmoc chemistry on a
Chemspeed peptide synthesizer. After Fmoc removal of the last
amino acid serine the N-terminus was acetylated by reaction
with acetic acid/HBTU/HOBt/DIPEA. The peptide was cleaved
from the resin using 4 mL of TFA/TIS/phenol/H2O, 35:2:2:1,
for 4 h and precipitated in diethyl ether to yield the crude peptide
that was purified by reverse phase HPLC to yield 71 mg (0.064
mmol, 29%) of white solid. HRMS [M + H]+ calculated for
C43H12N2O20S 1109.4785, found 1109.4803.
Peptide-NPs. A solution of the peptide was prepared at a thiol
concentration of 2.1 mM, and 100 µL of this solution was diluted
with 100 µL of 100 mM acetate buffer, pH 4, and added to 200
µL of pyridyl-NP solution (58.8 mg/mL, 25% NP) with a pyridyl
disulfide loading of 0.93 mM. The excess of thiol to pyridyl
disulfide was 10%. The reaction was followed by measuring
the absorbance at 340 nm in a well plate reader. After 2 h the
absorbance was stable and the peptide-NPs were purified by
gel filtration over Sepharose 6B beads eluting in water (200
µL sample and 10 mL bed volume). Presence of the peptide on
the NPs was measured with fluorescamine. The number of Ac-
SDKDSLKCG-OH peptide molecules per NP was estimated to
be approximately 160 (see Supporting Information).
Reduction of Peptide-NPs. An amount of 200 µL of purified
peptide-NPs was mixed with 10 µL of a solution of 57.2 mg/
mL TCEP ·HCl, resulting in a final concentration of 9.5 mM
reducing agent. The mixture was incubated for 2 h, and the
NPs were purified by gel filtration over Sepharose 6B beads
eluting in water (200 µL sample and 10 mL bed volume). As a
control, 200 µL of purified peptide-NPs mixed with 10 µL of
H2O was purified as well.
OVA-Functionalized NPs. OVA (12.5 mg/mL) in 10 mM PBS
was reduced with DTT at a final concentration of 32 mM for
4 h. Excess reducing agent was removed by gel filtration over
Sephadex G50 eluting with H2O. The fractions were analyzed
by Ellman assay prior to lyophilization to quantify the amount
of OVA; it was assumed that all cysteines were present in the
thiol form. Pyridyl disulfide-NPs with a pyridyl disulfide loading
of 547 µM (41 mg/mL, 25% NP) were added to 750 µL of
lyophilized 100 mM HEPES, pH 8.0, solution, and the clear
solution was added to 7.85 mg of lyophilized OVA resulting in
a final OVA concentration of 10.4 mg/mL (236 µM). The
reaction mixture was incubated overnight at 4 °C and purified
by gel filtration over Sepharose CL-6B beads eluting in water
(180 µL sample and 12 mL bed volume). As a control,
nonreduced OVA was reacted in the same way. Presence of
the OVA protein was measured with fluorescamine. The number
of OVA protein molecules per NP was estimated to be
approximately 75 (see Supporting Information).
Trityl Cysteamine (9). An amount of 11.4 g (100 mmol) of
cysteamine·HCl and 27.3 g (105 mmol) of triphenylmethanol
was added to 100 mL of DMF, and the mixture was heated at
60 °C for 15 min to obtain a clear-yellow solution. The solution
was cooled, 14 mL of BF3 ·Et2O (110 mmol) was added, and
the mixture was heated at 80 °C. After 14 h the mixture was
concentrated under reduced pressure and the oil was dissolved
in 500 mL of EtOAc. Upon standing a white solid precipitated
that was filtered off. The remaining clear EtOAc solution was
washed with saturated NaCl (aq) solution, causing more solid
to precipitate that was filtered off. Both solids were combined
and dried under reduced pressure, yielding 33.1 g (93 mmol,
93%) of tritylcysteamine·HCl salt. The free base was prepared
by treating 2.9 g (8.1 mmol) of the HCl salt with 150 mL of
saturated KOH (aq) solution and 500 mL of Et2O. The clear
layers were separated, and the organic phase was dried over
Na2SO4 and concentrated under reduced pressure. This yielded
after drying under reduced pressure 2.0 g (8.3 mmol, 78%) of
the free salt. NMR (CDCl3): δ ) 7.44 (m, 6H, CHaromat), 7.27
(m, 9H, CHaromat), 2.60 (t, 2H, CH2-N, J ) 6.4 and 6.6 Hz),
2.34 (t, 2H, CH2-S, J ) 6.4 and 6.6 Hz), 1.48 (br, 2H, NH2).
HRMS [M + H]+ calculated for C21H21NS 320.1473, found
320.1466.
Biotin Cysteamine (10). Amounts of 0.49 g (2.0 mmol) of
D-(+)-biotin, 0.47 g (2.0 mmol) of EDC, and 0.96 g (5.0 mmol)
of N-hydroxysuccinimide were suspended in 10 mL of DMF.
After the addition of 0.58 g (5.0 mmol) of 9 and 278 µL (2.0
mmol) of triethylamine the mixture was stirred for 2 days at
room temperature. The mixture was concentrated under reduced
pressure and taken up in 50 mL of EtOAc and washed with
H2O (3 × 50 mL). After drying over Na2SO4 and solvent
removal the residue was dissolved in a mixture of 10 mL of
CF3COOH and 10 mL of CH2Cl2, and 431 µL (2.1 mmol)
triisopropylsilane was added. After 4 h the volatiles were
removed in a flow of N2 and the residue was dissolved in 25
mL of EtOAc and extracted with H2O (3 × 10 mL). The
aqueous extracts were combined and lyophilized. The compound
was purified by HPLC to yield 0.091 g of white solid (15%
overall yield). 1H NMR (CD3OD): δ ) 4.53 (m, 1H), 4.34 (m,
1H), 3.37 (t, 2H, J ) 6.8 Hz), 3.25 (m, 1H), 2.97 (dd, 1H, J )
4.8 and 12.81 Hz), 2.74 (d, 1H, J ) 12.8 Hz), 2.64 (t, 2H, J )
7.2 Hz), 2.25 (t, 2H, J ) 7.2 Hz), 1.82-1.58 (m, 4H), 1.54-1.42
(m, 2H). HRMS [M + H]+ calculated for C12H21N3O2S2
304.1154, found 304.1157.
Biotin-NPs. Pyridyl disulfide-NPs (500 µL, 35.2 mg/mL, 75%
NP) with a pyridyl disulfide loading of 1.9 mM were diluted
with 500 µL of HEPES solution (200 mM, pH 8.0) and added
to 1.1 mg (3.6 µmol, 1.8 equiv) of biotin cysteamine. The
reaction was followed by measuring the absorbance at 340 nm
in a quartz cuvette. After 1 h the absorbance was stable and the
NPs were purified by dialysis (MWCO ) 10 kDa) against 5 L
of H2O for 24 h with regular replacement of water. The presence
RESULTS
In Scheme 1, the design concept of the functionalizable NPs
is shown. When a carboxylate-Pluronic is mixed with the
hydroxyl-Pluronic into the propylene sulfide emulsion, NPs can
be prepared with surface carboxylate groups. These carboxylate

Pyridyl Disulfide-Functionalized Nanoparticles
Bioconjugate Chem., Vol. 21, No. 4, 2010 657
Scheme 1. Synthesis of Funtionalizable Nanoparticlesa
a (a) Hydroxyl-Pluronic and carboxylate-Pluronic are used to make an emulsion with propylene sulfide. (b) To the emulsion is added four-arm
thiol initiator, and polymerization is started by addition of base. After polymerization, particles are exposed to air to cross-link, and the remaining
thiolates are end-capped with iodoacetamide. (c) The carboxylate-nanoparticles are reacted with pyridyl disulfide cysteamine (5) to yield thiol-
reactive functionalizable nanoparticles. (d) Conjugation of thiol-containing biotin (10), AcSDKDSLKCGOH peptide (11), and OVA protein gives
biological-functional nanoparticles.
Scheme 2. Synthesis of Carboxyl-Functionalized Pluronica
bance at 340 due to pyridinethione release. The degree of
functionalization was 100% on a mass basis.
The synthesized carboxylate-Pluronic was used for NP
synthesis. To gain information on how it would affect NP
synthesis, the weight ratio of carboxylate-Pluronic to hydroxyl-
Pluronic in the polymerization mixture was varied between 0%
and 100%. (For some applications, retention of surface hydroxyl
groups on the NPs is important (18).) DLS and TEM show that
the addition of carboxylate-Pluronic does not significantly affect
the size of NPs compared to the NPs with hydroxyl-Pluronic
alone, as can be seen from Table 1 and Figure 1, panels A and
B. However, the polydispersity index (PDI) was lower than for
NPs with hydroxyl-Pluronic alone, which is reflected by the
absence of the larger particles in 100% carboxylate-NPs (Figure
1B) and their presence in the hydroxyl-NPs (Figure 1A).
From NMR, the degree of polymerization of propylene sulfide
can be calculated by comparing the integral ratio after normal-
izing of the propylene sulfide CH3 group at 1.3 ppm and that
of the four-arm initiator methylene group -CH2- at 1.8 ppm.
Figure 2 shows the degree of polymerization as a function of
weight percentage carboxylate-Pluronic present in the polym-
erization mixture. The average degree of polymerization for the
different carboxylate-NPs is 63 ( 3 propylene sulfide monomer
units, which corresponds to 16 ( 1 units per initiator arm in
average. These values are quite similar to those (58 and 14 units,
respectively) calculated for hydroxyl-NPs. NMR also allows
calculation of the weight contributions of Pluronic, PPS, and
initiator to the total NP mass (see Supporting Information). From
Figure 3 it can be seen that the different carboxylate-NPs all
have a similar composition. The average values are 57 ( 2.1%
for Pluronic, 39 ( 2.0% for PPS, and 3.6 ( 0.2% for the
initiator. These values agree well with those found for hydroxyl-
Pluronic alone NPs (55%, 41%, and 4.1%, respectively).
a (a) SOBr2, toluene, reflux 4 h (1) or CH3SO2Cl, Et3N, toluene,
overnight (2); (b) thiopropionic acid methyl ester, K2CO3, DMF,
overnight; (c) NaOH (aq), overnight.
groups on the NP surface can then be converted into pyridyl
disulfide groups. The thiol-reactive pyridyl disulfide group
allows for conjugation with thiol-containing small molecules,
peptides, and proteins. The first step is synthesis of carboxylate
group-carrying NPs. To accomplish this, a Pluronic-carboxylate
was synthesized as shown in Scheme 2.
Having established that NPs can be made using the carboxy-
late-Pluronic, the next step was reacting them with pyridyl
disulfide cysteamine·HCl, obtained in a single step from
cysteamine ·HCl and aldrithiol in methanol, using standard EDC/
sulfo-NHS coupling chemistry at pH 4.0-4.5. The amount of
Pluronic was calculated from NMR data of lyophilized car-
boxylate-NPs. Since it was not clear what the amount of
carboxylate-Pluronic in the 10-75% carboxylate-NPs was,
calculations of coupling reagents were based on the total amount
Pluronic in the NP solutions (as measured by NMR), which
was similar for all carboxylate-NPs solutions. The number of
equivalents of EDC (24 equiv), sulfo-NHS (20 equiv), and
pyridyl disulfide cysteamine ·HCl (20 equiv) relative to Pluronic
was fixed, and the reactions were carried out at the same time.
After functionalization, the NPs were dialyzed and character-
ized by DLS (Table 1) and TEM (Figure 1C). The PDI slightly
increased, and TEM supports the DLS data, as larger particles
are present as well. NMR of the different NPs revealed the
presence of the pyridyl disulfide group by the four characteristic
signals of the pyridyl disulfide group as observed for the pyridyl
As the first step of Pluronic carboxylation, both hydroxyl
groups were converted into either the bromide by reaction with
thionyl bromide or the mesylate with methanesulfonyl chloride
in toluene. The degree of functionalization was 100% for both
the bromide and mesylate according to NMR, as no OH groups
were visible in deuterated DMSO. Harris et al. showed that the
OH group of PEG occurs as a triplet at δ ) 4.58 ppm in this
aprotic solvent and can be used to determine the degree of
functionalization reliably (23). Nucleophilic substitution of the
bromide or mesylate with mercaptomethyl propionate in the
presence of potassium carbonate in DMF afforded the corre-
sponding methyl ester in high yield. Hydrolysis of the ester in
aqueous NaOH yielded the carboxylate-Pluronic, as was shown
by the disappearance of the CH3 group at 3.7 ppm. The presence
of the carboxylate groups was furthermore confirmed by reaction
with pyridyl disulfide cysteamine ·HCl in the presence of EDC
and NHS (Scheme 3) to yield the corresponding pyridyl
disulfide-Pluronic. In addition, the degree of functionalization
was calculated by reducing the pyridyl disulfide groups with
TCEP·HCl in aqueous HEPES, pH 8, and reading the absor-

658 Bioconjugate Chem., Vol. 21, No. 4, 2010
van der Vlies et al.
Scheme 3. Synthesis of Pyridyl Disulfide-Functionalized Pluronica
a EDC, NHS, CH2Cl2, overnight.
Figure 3. Mass percentage of Pluronic (black), poly(propylene sulfide)
(gray), and initiator (white) of total NP weight as a function of weight
percentage carboxylate-Pluronic in the polymerization mixture as
calculated from NMR.
Figure 1. TEM images of aqueous nanoparticle (NP) suspensions. Size
( SD is given in nm, and number of particles measured is indicated in
parentheses: (A) hydroxyl-NPs, 17.4 ( 4.5 (691); (B) 100% carboxy-
late-NPs, 16.0 ( 3.6 (570); (C) 50% pyridyl disulfide-NPs, 16.2 ( 4.5
(986); (D) 75% biotin-NPs, 16.7 ( 5.3 (849). Some larger particles as
mentioned in the text are indicated by white arrows.
Figure 4. Number of pyridyl disulfide groups divided by the number
of repeating propylene oxide (PO) units (triangles) as calculated from
NMR from the ratio of pyridyl disulfide groups in µM (as measured
by release of pyridine thione after reduction with TCEP ·HCl) and NP
concentration in mg/mL. The pyridyl disulfide number (squares) is a
linear function of weight percentage carboxylate-Pluronic in the
polymerization mixture.
as measured with NMR shows that there is a linear relationship
between the feed ratio of carboxylate-Pluronic and the amount
of pyridyl disulfide-Pluronic present on the NP. However, the
reaction does not lead to full conversion of all carboxylate into
pyridyl disulfide groups, as can be calculated from the carboxy-
late-Pluronic feed ratio at 100%. By comparison of the
experimentally determined ratio of pyridyl disulfide groups to
propylene glycol groups with that theoretically calculated, about
57% of carboxylate groups were determined to have reacted. A
similar value of 61% is obtained by measuring the release of
pyridine thione after reduction with TCEP ·HCl and comparing
this concentration with that calculated from the NP concentration
and weight percentage of Pluronic in the NP solution. This low
functionalization degree is not due to loss and/or modification
of carboxylic acid groups during NP synthesis. Control experi-
ments with the carboxylate-Pluronic under conditions used for
Figure 2. Degree of polymerization of propylene sulfide (circles) and
number of propylene sulfide units per initiator arm (squares) as a
function of weight percentage carboxylate-Pluronic in the polymeri-
zation mixture.
disulfide-Pluronic (6). More importantly, the presence of a broad
signal at 7.3 ppm assigned to the NH of the amide bond, and
also present in the above-mentioned pyridyl disulfide-Pluronic,
confirmed successful conjugation (see Supporting Information).
In Figure 4 the number of pyridyl disulfide groups relative to
the number of propylene glycol repeating units on the Pluronic

Pyridyl Disulfide-Functionalized Nanoparticles
Bioconjugate Chem., Vol. 21, No. 4, 2010 659
Scheme 4. Synthesis of Thiolated Biotina
a (a) BF3 ·Et2O, DMF, 80°C, 16 h; (b) EDC, NHS, Et3N, 2 days; (c) CF3COOH/TIS, CH2Cl2, 4 h.
Table 2. Dynamic Light Scattering Results for the Different
Conjugated NPsa
compd
Z-Avgb (SD)
PDIc (SD)
biotin
peptide
OVA
36.4 (0.26)
36.7 (0.21)
56.0 (0.32)
0.201 (0.003)
0.209 (0.003)
0.267 (0.006)
a Values are the average of three measurements of the same sample.
b Size in nm as measured by the Z-average obtained by fitting the
correlation function using the cumulant method. c Polydispersity index
(PDI) calculated from fitting the correlation function using the cumulant
method.
polymerization but without the addition of propylene sulfide
did not lead to changes in the NMR spectrum (data not shown).
The same linear trend is observed if the total pyridyl disulfide
loading in µM is normalized by the NP concentration in mg/
mL, the pyridyl disulfide number (Figure 4). Both DLS and
TEM confirmed that the NPs kept their small size but that some
aggregation occurred as shown by the increase in PDI (Table
1) and the presence of bigger particles in TEM (Figure 1C).
With the pyridyl disulfide-NPs at hand, we wanted to show
that we could easily conjugate small molecules, peptides, and
proteins. We first considered conjugation of a thiol-containing
biotin to the NPs. In Scheme 4, the synthesis of this compound
is outlined. After purification by HPLC, the biotin cysteamine
was conjugated to pyridyl disulfide-NPs and purified by dialysis.
DLS (Table 2) and TEM (Figure 1D) both show that the NPs
kept their size. Successful conjugation was shown by adding
the biotin-NPs to streptavidin Alexa488 and loading the NP
solutions on a SDS-PAGE gel. From Figure 5 (lanes 4-6) it
can be seen that above a certain biotin/streptavidin molar ratio,
all streptavidin was bound to the NP surface and was no longer
able to move into the gel. The reductive sensitivity of the biotin-
NP disulfide linkage is clearly illustrated by running the gel
(lanes 8-10) in the presence of mercaptoethanol, a reducing
agent that leads to cleavage of the disulfide bond between biotin
and the NP.
To see the effect of molecular size on conjugation, a model
peptide was synthesized with a cysteine added at the C-terminus.
Conjugation of the peptide was done at pH 4, and excess peptide
was removed by gel filtration on Sepharose 6B beads. Figure
6A shows the elution profile for pyridyl disulfide-NP after
reducing the fractions with TCEP ·HCl and measuring the
absorbance at 340 nm due to pyridinethione release. As can be
seen from Figure 6A, the NPs elute as a sharp peak and DLS
confirmed the presence of NPs in the fractions. Figure 6B shows
the elution profile of peptide-NP after gel filtration as measured
after reaction of the lysine residues on the peptide with
fluorescamine and measuring the fluorescence intensity. As can
be seen, the peptide is bound to the NPs, as the elution profile
is practically the same as that measured for the pyridyl disulfide-
Figure 5. SDS-PAGE gel showing the interaction between biotin-
NPs and streptavidin-Alexa488: (lane 1) ladder; (lane 2) biotin-NPs
(15 µL); (lane 3) streptavidin-Alexa488 (10 µg); (lanes 4-6) strepta-
vidin-Alexa488 (10 µg) with increasing volumes (5, 10, and 15 µL) of
biotin-NPs; (lanes 7-10) same as for lanes 3-6 but in the presence of
ꢀ-mercaptoethanol.
NP. DLS (Table 2) shows that the NPs still have similar size
relative to the pyridyl disulfide-NP. To show that this peptide
is linked by a disulfide bond, the peptide-NPs were reduced
with TCEP ·HCl prior to gel filtration. In Figure 6C it can be
seen that the peak in Figure 5B is completely gone and that a
new peak can be observed at higher elution volume, which is
due to free peptide.
Having shown that a small molecule and a larger peptide can
be conjugated to the pyridyl disulfide-NPs, we were also
interested in conjugating proteins. We choose OVA, a 44 000
Da protein that after full reduction has six cysteine residues,
and we conjugated it to the NPs. From the elution profile shown
in Figure 7A, it can be seen that all OVA has been conjugated.
Reduction of OVA is necessary, as shown for the elution profile
of the reaction mixture of OVA that had not been reduced prior
to conjugation (Figure 7B). Clearly OVA elutes much later than
when bound to the NPs.
DISCUSSION
We previously reported the emulsion polymerization of
propylene sulfide/Pluronic F127 by anionic ring-opening po-
lymerization (16). The presence of the two hydroxyl groups of
the Pluronic F127 on the surface of the cross-linked polypro-
pylene sulfide core opens the possibility of introducing biological
functionality onto the NPs by modifying the R,ω hydroxyl
groups. However, this should be done prior to making the NPs,
as commonly used hydroxyl activation chemistries are not
compatible with the aqueous solution of the NPs. For this reason,
Pluronic F127 has been modified with reactive groups and

660 Bioconjugate Chem., Vol. 21, No. 4, 2010
van der Vlies et al.
Chart 1. Commonly Available Thiol-Reactive Groups (A) Vinyl-
sulfone, (B) Acrylate, (C) Maleimide, (D) Alkyl Bromide or Iodide,
(E) Pyridyl Disulfide
Chart 2. Reaction between Pyridyl Disulfide and a Thiol To Give
a Mixed Disulfide with Release of Pyridine Thione
the surface of the NP for binding to the NPs might be most
advantageous, as peptides and proteins can be easily engineered
to have a free thiol group. In this way with only one NP batch,
functionalizable NPs, one would have access to a number of
biological-functional NPs.
Figure 6. (A) Elution profile of pyridyl disulfide-NPs. Fractions were
reduced with TCEP ·HCl, and the absorbance at 340 nm due to pyridine
thione was measured. (B) Elution profile of peptide-NPs after purifica-
tion. (C) Elution profile peptide-NPs after reduction with TCEP ·HCl.
Fractions were reacted with fluorescamine, and the fluorescence
intensity was measured.
For the thiol-reactive group on the NP surface, there are
several options, as shown in Chart 1. First are the Michael
acceptors vinylsulfone (A), acrylate (B), and maleimide (C).
These groups react with thiols in a conjugate Michael addition
to form a thioether bond between the thiol and double bond. A
thioether bond is also formed when the thiol reacts in a
nucleophilic substitution reaction with bromo and iodo deriva-
tives (D). The disadvantage of these thiol-reactive groups is that
a physiologically irreversible thioether bond is formed. In
the case of antigen delivery, this might be problematic, as the
antigen is not released from the NP after being taken up by the
cell unless first proteolytically processed. Despite the fact that
the acrylate-based conjugate is cleavable by ester hydrolysis,
this group is problematic for long-term storage of the aqueous
antigen-NPs solution.
Another thiol-reactive group is the pyridyl disulfide group
(E) that reacts with thiols to give a disulfide bond and releases
pyridine thione as shown in Chart 2. This disulfide exchange
reaction is practically irreversible because of the formation of
the very stable pyridine thione which has a anomalously low
pKa value (25). This is also the reason why this reaction takes
place over a broad pH range between 4 and 8 (data not shown).
In addition to the broad pH range, conjugation using the pyridyl
disulfide group has two other advantages: the amount of protein
or peptide conjugated can be quantified by measuring the release
of pyridine thione, and the peptide or protein is linked to the
NP surface via a reduction-sensitive disulfide bond, which allows
a cellular mechanism for antigen release after endocytosis due
to the reductive potential early during endolysosomal processing.
This increase in reducing environment will result in cleavage
of the peptide/protein-NP disulfide bond and allow for release
of the peptide or protein from the NP to affect its biological
activity. This antigen release mechanism was not possible in
the original scheme using the vinylsulfone coupling; rather,
proteolytic release was required. For this reason we function-
alized the NPs with this pyridyl disulfide group. The most
straightforward approach is synthesis of a pyridyl disulfide-
Pluronic and adding this to the polymerization mixture to make
pyridyl disulfide-NPs. However, a problem is that the thiolate
species that are responsible for the polymerization of propylene
sulfide could also react with the pyridyl disulfide groups on the
Pluronic. Indeed, after synthesizing pyridyl disulfide-Pluronic
(21) and addition to the polymerization mixture, NMR showed
that the pyridyl disulfide group was not present after NP
Figure 7. Elution profile after reacting pyridyl disulfide-NPs at pH 8
with (A) reduced and (B) nonreduced ovalbumin. Fractions were
analyzed by addition of fluorescamine and measurement of the
fluorescence intensity.
reacted with molecules prior to nanoparticle synthesis to obtain
biologically functional nanoparticles. Previously we have modi-
fied Pluronic F127 with p-nitrophenol (17) and vinylsulfone (18)
groupsthatallowconjugationofpeptides(24)andproteins(16,18)
using amines and thiols.
Despite the successful applications described above, there
remains a disadvantage in our previous scheme, in that for
making biological-functional NPs each time the corresponding
functionalized Pluronic has to be synthesized, which may limit
the scope of use of the biological-functional NPs. Furthermore,
the polymerization conditions during NP synthesis might be
deleterious for the bioactivity of, for example, proteins for which
the conservation of sensitive native three-dimensional structure
is highly important. Indeed, in the case of the GOx-NPs, despite
careful optimization of the synthesis, more than 30% loss of
enzyme activity after NP synthesis was observed.
For these reasons, we sought to develop a more flexible
system that would allow for mild reaction conditions and broad
pH range to conjugate small molecules, peptides, and proteins
to the NP surface. We reasoned that a thiol-reactive group on

Pyridyl Disulfide-Functionalized Nanoparticles
Bioconjugate Chem., Vol. 21, No. 4, 2010 661
synthesis (data not shown). To circumvent this problem, i.e.,
introducing the thiol-reactive pyridyl disulfide group in the
presence of thiolate species, we came to the following approach
as outlined in Scheme 1. As the first step, carboxylate-func-
tionalized NPs are synthesized and, in the second step, func-
tionalized with a pyridyl disulfide containing a free amine to
introduce the thiol reactive groups. Note that we choose the
NPs to have a negative charge instead of positive (amines) to
avoid the potential toxic effects of cationic charges that could
arise from unreacted amine groups. The carboxylate-Pluronic
was therefore synthesized in three steps from Pluronic F127 as
outlined in Scheme 1. The use of the methyl ester was necessary,
as reaction of mercaptopropionic acid also led to substitution
at the carboxylic acid leading to an ester bond as was visible
in MES (aq) buffer (data not shown). The linear trend in Figure
4 suggests that the percentage of carboxylate-Pluronic on the
NP surface is the same as that in the polymerizing mixture but
that only about 60% can be functionalized. This finding enables
the control of surface density of pyridyl disulfide groups on
the NP surface by adjusting the feed ratio in the polymerization
mixture of carboxylate- to hydroxyl-Pluronic.
Having shown that pyridyl disulfide groups could be intro-
duced onto the NPs, it was important to show that these are
actually in contact with the aqueous phase and available for
conjugating with peptides and proteins. To explore this, we
synthesized a thiol-containing biotin molecule and reacted it
with the pyridyl disulfide-NPs. With this biotin molecule at both
ends of the Pluronic, it would also be possible to see whether
after conjugating biotin on the Pluronic ends the NP surface
would be still accessible for interaction with large proteins.
Biotin interacts with streptavidin, a 66 000 Da protein, to form
the strongest noncovalent bond known in nature. Indeed biotin
on the NPs is able to interact with the binding pockets of
streptavidin, as shown in Figure 5. This result shows that the
pyridyl disulfide groups on the NP surface are accessible in the
aqueous phase for conjugation and that streptavidin is able to
interact with the conjugated biotin on the NP surface. The biotin-
NPs are furthermore interesting from a staining point of view.
Having biotin bound to the NPs enables us the ability to probe
the presence of the NPs in tissues by cryosectioning and staining
with fluorescently labeled streptavidin.
For immunotherapy applications normally whole immuno-
genic proteins are used; however, they typically lead to release
of a limited number of antigenic peptides when processed within
DCs. For example the OVA protein yields two (SIINFEKL and
ISQAVHAAHAEINE) antigenic peptide sequences upon pro-
cessing by DCs (27). Since peptides are easily accessible by
solid phase peptide synthesis, they can be made to have a free
cysteine residue, enabling conjugation to NPs. As a model, we
synthesized a nine amino acid peptide containing a cysteine and
conjugated it to the NP. For purification of the peptide-NP, gel
filtration was used because it is time saving and a better
purification method compared to dialysis, especially with
increasing size of the molecule to be conjugated. The presence
of the peptide on the NPs was confirmed with DLS and
fluorescamine (20). Important in our concept of antigen-bearing
NPs is the reductive sensitivity of the antigen-NP disulfide bond.
For this reason, we exposed both the biotin-NPs and peptide-
NPs to a reducing agent to demonstrate sensitivity to reduction.
In the case of biotin, mercaptoethanol led to a cleavage of the
disulfide bond at room temperature as shown by SDS-PAGE
gel where the streptavidin was no longer bound to the NP surface
and freely moved into the gel. For the peptide-NP, addition of
TCEP·HCl also led to disulfide cleavage as could be shown
after gel filtration.
1
from H NMR by the occurrence of a triplet around 4.2 ppm.
After alkaline hydrolysis, the carboxylate-Pluronic was kept as
its sodium salt so as to not interfere during NPs synthesis by
lowering the pH. As the final step in characterizing the
carboxylate-Pluronic, the polymer was reacted with pyridyl
disulfide·HCl to give pyridyl disulfide-Pluronic. This polymer
would serve as a reference for characterization of the pyridyl
disulfide NPs.
With the carboxylate-Pluronic at hand, NPs were synthesized
by emulsion polymerization with different COOH-to-OH Plu-
ronic mass ratios. After exposure to air for 2 h to form disulfides,
the remaining free thiolate anions that were spatially too far
apart to form disulfides were blocked by the addition of
iodoacetamide. Previous work had shown that this is effective
in blocking thiolates (15) and capping was quantitative as
measured by Ellman assay. This blocking is important to avoid
a disulfide exchange reaction when the pyridyl disulfide groups
are introduced onto the NPs. From the DLS and TEM data, it
can be seen that the addition of carboxylate-Pluronic did not
significantly affect the size of the final NPs. The PDI, however,
dropped significantly (Table 1), suggesting that the negatively
charged NP surface stabilizes the particles, which can be
explained by DLVO theory, which states that like charge
stabilizes colloid particles (26). In addition, the carboxylate-
Pluronic did not influence the ring-opening polymerization of
propylene sulfide to a significant extent as shown by the degree
of polymerization and weight composition of the NPs. This
shows that the NP synthesis is not affected significantly by the
presence of the carboxylate-Pluronic in the polymerization
mixture. Because of overlap of the characteristic signals of the
carboxylate-Pluronic at 2.5, 2.7, and 2.8 ppm with the broad
CH and CH2 signals at 2.5-2.7 and 2.8-3.0 ppm of the PPS,
it was not possible to determine the percentage of carboxylate-
Pluronic present in the 10-75% carboxylate-NPs by NMR. For
this reason we used the conjugation with pyridyl disulfide
cysteamine·HCl as an indirect measurement to obtain insight
into how the carboxylate-Pluronic was incorporated into NP
surface as a function of the feed ratio in the polymerization
mixture. From the 100% carboxylate-NPs, it was calculated that
only about 60% functionalization was achieved, even though
the carboxylate-Pluronic alone could be functionalized almost
quantitatively. This lower degree of functionalization is not due
to loss of carboxylate groups as was shown in control experi-
ments. An explanation could be that some carboxylate groups
remain inaccessible during functionalization, though how that
would happen is unclear: interaction of the negatively charged
carboxylate group with the hydrophobic PPS is unlikely. It is
therefore more likely that the EDC/sulfo-NHS coupling chem-
istry itself limits the degree of functionalization because of its
inherent water sensitivity. Indeed model experiments in which
pyridyl disulfide cysteamine ·HCl was coupled (using NHS) to
the carboxylate-Pluronic that had been used in NP synthesis
led to almost quantitative conversions in CH2Cl2 but only 45%
Finally we wanted to show that it is also possible to conjugate
a larger protein. We choose OVA because it is a model antigen
that is used frequently in immunotherapy studies. OVA has four
free cysteine residues and one disulfide bond that would allow
easy conjugation using our functionalization scheme. The
problem is that these four cysteine residues are buried in the
native state and are not reactive toward iodoacetic acid at pH
8.2, an alkylation agent (28). For this reason we reduced OVA
to expose all six cysteine residues (four free and two in a
disulfide). This, however, leads to a potential problem of cross-
linking, as the NP surface is decorated with several pyridyl
disulfide groups. With careful optimization, we managed to
conjugate OVA to pyridyl disulfide-NPs. From Figure 6A it
can be seen that all OVA eluted in one peak and that free OVA
was not detected. DLS indicated an increase in size but the OVA
nanoparticles still remained small (Table 2). That practically

662 Bioconjugate Chem., Vol. 21, No. 4, 2010
van der Vlies et al.
(10) McKeever, U., Barman, S., Hao, T., Chambers, P., Song, S.,
Lunsford, L., Hsu, Y. Y., Roy, K., and Hedley, M. L. (2002)
Protective immune responses elicited in mice by immunization
with formulations of poly(lactide-co-glycolide) microparticles.
Vaccine 20, 1524–1531.
(11) Jiang, W., Gupta, R. K., Deshpande, M. C., and Schwendeman,
S. P. (2005) Biodegradable poly(lactic-co-glycolic acid) micro-
particles for injectable delivery of vaccine antigens. AdV. Drug
DeliVery ReV. 57, 391–410.
(12) Bennewitz, N. L., and Babensee, J. E. (2005) The effect of
the physical form of poly(lactic-co-glycolic acid) carriers on the
humoral immune response to co-delivered antigen. Biomaterials
26, 2991–2999.
(13) De Koker, S., Naessens, T., De Geest, B. G., Bogaert, P.,
Demeester, J., De Smedt, S., and Grooten, J. (2010) Biodegrad-
able polyelectrolyte microcapsules: antigen delivery tools with
Th17 skewing activity after pulmonary delivery. J. Immunol. 184,
203–211.
no free OVA was present is clearly illustrated in Figure 6B by
elution nonreduced OVA, which elutes at a much higher elution
volume and shows the need for reducing OVA prior to
conjugation. Indeed, if conjugation of reduced OVA is not
quantitative, the elution profile shows both peaks (data not
shown).
To conclude, we have presented a scheme by which to make
thiol-reactive Pluronic-stabilized poly(propylene sulfide) NPs
by incorporating a carboxylate-Pluronic into the hydroxyl-
Pluronic/propylene sulfide mixture that was subsequently reacted
with pyridyl disulfide cysteamine to yield pyridyl disulfide-NPs.
The polymerization and NP composition were not affected by
the presence of the carboxylate-Pluronic, and the number of
pyridyl disulfide groups on the NP surface could be controlled
by changing the weight ratio of carboxylate-Pluronic to hy-
droxyl-Pluronic in the mixture. We further showed that we could
conjugate biotin, a nine amino acid peptide, as well as the larger
protein OVA. Currently we are exploring the functionalizable
NPs scheme for immunological application by conjugating
antigenic peptides, proteins, and protein structures with adjuvant
activity.
(14) Reddy, S. T., Rehor, A., Schmoekel, H. G., Hubbell, J. A.,
and Swartz, M. A. (2006) In vivo targeting of dendritic cells in
lymph nodes with poly(propylene sulfide) nanoparticles. J.
Controlled Release 112, 26–34.
(15) Rehor, A., Tirelli, N., and Hubbell, J. A. (2002) A new living
emulsion polymerization mechanism: episulfide anionic polym-
erization. Macromolecules 35, 8688–8693.
(16) Rehor, A., Hubbell, J. A., and Tirelli, N. (2004) Oxidation-
sensitive polymeric nanoparticles. Langmuir 21, 411–417.
(17) Rehor, A., Botterhuis, N. E., Hubbell, J. A., Sommerdijk, N.,
and Tirelli, N. (2005) Glucose sensitivity through oxidation
responsiveness. An example of cascade-responsive nano-sensors.
J. Mater. Chem. 15, 4006–4009.
ACKNOWLEDGMENT
We thank the Competence Centre for Materials Science and
Technology (CCMX) of the ETH-Board Switzerland for partial
funding of this research.
(18) Reddy, S. T., van der Vlies, A. J., Simeoni, E., Angeli, V.,
Randolph, G. J., O’Neil, C. P., Lee, L. K., Swartz, M. A., and
Hubbell, J. A. (2007) Exploiting lymphatic transport and
complement activation in nanoparticle vaccines. Nat. Biotechnol.
25, 1159–1164.
(19) Ellman, G. L. (1959) Tissue sulfhydryl groups. Arch. Biochem.
Biophys. 82, 70–77.
Supporting Information Available: Calculation of NP
composition, NMR spectra of 6 and pyridyl disulfide-NP, and
estimation of numbers of biotin, peptide, and OVA molecules
per NP. This material is available free of charge via the Internet
at http://pubs.acs.org.
(20) Udenfriend, S., Stein, S., Bohlen, P., Dairman, W., Leimgru-
ber, W., and Weigele, M. (1972) Fluorescamine: a reagent for
assay of amino acids, peptides, proteins, and primary amines in
the picomole range. Science 178, 871–872.
(21) O’Neil, C. P., van der Vlies, A. J., Velluto, D., Wandrey, C.,
Demurtas, D., Dubochet, J., and Hubbell, J. A. (2009) Extra-
cellular matrix binding mixed micelles for drug delivery ap-
plications. J. Controlled Release 137, 146–151.
(22) Green, N. M. (1970) Spectrophotometric determination of
avidin and biotin. Methods Enzymol. 18 (Part 1), 418–424.
(23) Dust, J. M., Fang, Z. H., and Harris, J. M. (2002) Proton NMR
characterization of poly(ethylene glycols) and derivatives. Mac-
romolecules 23, 3742–3746.
(24) Rothenfluh, D. A., Bermudez, H., O’Neil, C. P., and Hubbell,
J. A. (2008) Biofunctional polymer nanoparticles for intra-
articular targeting and retention in cartilage. Nat. Mater. 7, 248–
254.
(25) Tatsumi, E., and Hirose, M. (1997) Highly ordered molten
globule-like state of ovalbumin at acidic pH: native-like frag-
mentation by protease and selective modification of Cys367 with
dithiodipyridine. J. Biochem. 122, 300–308.
(26) Verwey, E. J. W., Overbeek, J. T. G. (1948) Theory of the
Stability of Lyophobic Colloids Elsevier, New York.
(27) Burgdorf, S., Kautz, A., Bohnert, V., Knolle, P. A., and Kurts,
C. (2007) Distinct pathways of antigen uptake and intracellular
routing in CD4 and CD8 T cell activation. Science 316, 612–
616.
LITERATURE CITED
(1) Nestle, F. O., Banchereau, J., and Hart, D. (2001) Dendritic
cells: on the move from bench to bedside. Nat. Med. 7, 761–
765.
(2) Banchereau, J., and Steinman, R. M. (1998) Dendritic cells and
the control of immunity. Nature 392, 245–252.
(3) Randolph, G. J., Angeli, V., and Swartz, M. A. (2005) Dendritic-
cell trafficking to lymph nodes through lymphatic vessels. Nat.
ReV. Immunol. 5, 617–628.
(4) Gay, N. J., Gangloff, M., and Weber, A. N. R. (2006) Toll-
like receptors as molecular switches. Nat. ReV. Immunol. 6, 693–
698.
(5) Hubbell, J. A., Thomas, S. N., and Swartz, M. A. (2009)
Materials engineering for immunomodulation. Nature 462, 449–
460.
(6) Singh, M., Briones, M., Ott, G., and O’Hagan, D. (2000)
Cationic microparticles: a potent delivery system for DNA
vaccines. Proc. Natl. Acad. Sci. U.S.A. 97, 811–816.
(7) Newman, K. D., Samuel, J., and Kwon, G. (1998) Ovalbumin
peptide encapsulated in poly(D,L lactic-co-glycolic acid) micro-
spheres is capable of inducing a T helper type 1 immune
response. J. Controlled Release 54, 49–59.
(8) Newman, K. D., Sosnowski, D. L., Kwon, G. S., and Samuel,
J. (1998) Delivery of MUC1 mucin peptide by poly(D,L-lactic-
co-glycolic acid) microspheres induces type 1 T helper immune
responses. J. Pharm. Sci. 87, 1421–1427.
(9) Newman, K. D., Elamanchili, P., Kwon, G. S., and, and Samuel,
J. (2002) Uptake of poly(D,L-lactic-co-glycolic acid) microspheres
by antigen-presenting cells in vivo. J. Biomed. Mater. Res. 60,
480–486.
(28) Takahashi, N., and Hirose, M. (1990) Determination of
sulfhydryl groups and disulfide bonds in a protein by polyacry-
lamide gel electrophoresis. Anal. Biochem. 188, 359–365.
BC9004443