Active ester functional single core magnetic nanostructures as a versatile immobilization matrix for effective bioseparation and catalysis

Article abstract of DOI:10.1021/bm901203z

Multifunctional nanocarriers for amino functional targets with a high density of accessible binding sites are obtained in a single polymerization step by grafting from copolymerization of an active ester monomer from superparamagnetic cores. As a result of the brush-like structure of the highly dispersed shell, the nano-objects exhibit an available capture capacity for amines that is found to be up to 2 orders of magnitude higher than for commercial magnetic beads, and the functional brush shell can serve as a template for many types of pendant functional groups and molecules. As comonomer, oligo(ethylene glycol) methacrylate allows for excellent water solubility at room temperature, biocompatibility, and thermoflocculation. We demonstrate the biorelated applicability of the hybrid nanoparticles by two different approaches. In the first approach, the immobilization of trypsin to the core-shell nanoparticles results in highly active, nanoparticulate biocatalysts that can easily be separated magnetically. Second, we demonstrate that the obtained nanoparticles are suitable for the effective labeling of cell membranes, opening a novel pathway for the easy and effective isolation of membrane proteins.

Full text of DOI:10.1021/bm901203z

Biomacromolecules 2010, 11, 635–642
635
Active Ester Functional Single Core Magnetic Nanostructures
as a Versatile Immobilization Matrix for Effective Bioseparation
and Catalysis
Thorsten Gelbrich,† Michael Reinartz,‡ and Annette M. Schmidt*,†
Institut fu¨r Organische Chemie und Makromolekulare Chemie, and Institut fu¨r Herz- und
Kreislaufphysiologie, Heinrich-Heine-Universita¨t Du¨sseldorf, Universita¨tsstr. 1,
D-40225 Du¨sseldorf, Germany
Received October 22, 2009; Revised Manuscript Received January 10, 2010
Multifunctional nanocarriers for amino functional targets with a high density of accessible binding sites are obtained
in a single polymerization step by grafting from copolymerization of an active ester monomer from
superparamagnetic cores. As a result of the brush-like structure of the highly dispersed shell, the nano-objects
exhibit an available capture capacity for amines that is found to be up to 2 orders of magnitude higher than for
commercial magnetic beads, and the functional brush shell can serve as a template for many types of pendant
functional groups and molecules. As comonomer, oligo(ethylene glycol) methacrylate allows for excellent water
solubility at room temperature, biocompatibility, and thermoflocculation. We demonstrate the biorelated applicability
of the hybrid nanoparticles by two different approaches. In the first approach, the immobilization of trypsin to the
core-shell nanoparticles results in highly active, nanoparticulate biocatalysts that can easily be separated
magnetically. Second, we demonstrate that the obtained nanoparticles are suitable for the effective labeling of
cell membranes, opening a novel pathway for the easy and effective isolation of membrane proteins.
Reversible thermoflocculation of magnetic colloids by en-
capsulation with thermoresponsive polymers as a strategy to
overcome this gap has been proposed already in the 90s.8 While
latexes or microgels containing magnetic nanoparticles are
relatively easy to disperse at the application temperature, they
possess agglomeration at a temperature range above or below,
caused by a thermosensitive solvation behavior of the polymer
component. As the latter, poly(N-isopropylacrylamide) PNiPAAm
is frequently employed,8-10 although the LCST-type flocculation
temperature of 32 °C does not allow good particle dispersion
at body temperature, and furthermore, the polymer tends to
unspecific protein adsorption. As an alternative, just recently,
beads based on acryloyl glycinamide copolymers with UCST
behavior that flocculate below 10 °C have been developed and
commercialized.11,12
Introduction
Using colloidal supports for bioactive species has enabled
important advances in biotechnology, medical diagnostics, drug
screening, and other areas of actual interest. In in vitro as well
as in in vivo environments, the substrate can introduce or
facilitate the feasibility to detect, quantify, localize, or separate
biologically active species.
In this respect, superparamagnetic carriers actually attract high
attention as they enable the visualization, manipulation, and
activation of labeled species by the use of properly designed
magnetic fields. Examples for the wide applicability of magnetic
colloids in the bioscience include contrast agents for MRI,
magnetic separation kits, and therapeutic approaches like
magnetic fluid hyperthermia and active drug delivery.1-6
To provide an increased versatility in terms of protein
interaction, thermal behavior, and particle mobility, hydrophilic
coatings with controlled polymer architecture and easily tunable
parameters are needed. The formation of a polymer brush on
the surface of single nanoparticles has proved to be a valuable
tool for the design of single-cored hybrid structures with tailored
dispersion behavior.9,13-19 The advances of a brush shell
architecture are a high density of end-tethered chains, a shell
thickness that is adjustable by the polymer arm length, and a
greatly increased accessibility of the chains by solvent molecules
and reactants as compared to latex beads or cross-linked
microgels. Magnetic polymer brushes with thermoflocculation
behavior have been reported for organic solvents by our
group.17-19 Lately, hydrophilic brush shells have been descri-
bed,20,21 for instance, prepared by a “grafting to” approach of
tailored copolymers from oligo(ethylene glycol) methacrylates
with adjustable and narrow flocculation temperature and low
unspecific adsorption.22,23
When designing an immobilization matrix, a range of factors
such as density and accessibility of binding sites, their microen-
vironment, and mobility strongly influence the intensity of
interaction and the effectiveness of the binding event.7 In this
respect, the question of size relevance is nontrivial. In general,
small carriers in the nanometer range are superior in terms of
accessible surface, their mode of interaction with cells and
proteins, and higher mobility with respect to diffusion and cell
uptake. In contrast, concerning the ease of detection and
separation, micrometer-sized solid supports have clear advan-
tages over smaller structures and show higher effectivity for
their isolation by centrifugation or filtering methods. In the case
of magnetic structures, micrometer-sized beads can possess a
higher magnetic moment, thus, enhancing magnetophoretic
mobility, magnetic contrast, and heatability in HF fields.
* To whom correspondence should be addressed. Phone: +49-211-
8114820. Fax: +49-211-15840. E-mail: schmidt.annette@uni-duesseldorf.de.
† Institut fu¨r Organische Chemie und Makromolekulare Chemie.
‡ Institut fu¨r Herz- und Kreislaufphysiologie.
Here, we report the design and application of single-cored
hybrid magnetic core-shell particles that carry a high number
10.1021/bm901203z
 2010 American Chemical Society
Published on Web 02/04/2010

636 Biomacromolecules, Vol. 11, No. 3, 2010
Gelbrich et al.
Table 1. Composition of the Investigated FeOx-P(OEGMA-co-SIMA) Core-Shell Nanoparticlesa
samplea
µP (%)
Ms (A·m-1
)
µMF (%)
dc (nm)
dh (nm)
Tc (°C)
FeOx-P(OEGMA-co-SIMA)25
FeOx-P(OEGMA-co-SIMA)20
FeOx-P(OEGMA-co-SIMA)15
FeOx-P(OEGMA-co-SIMA)10
FeOx-P(OEGMA-co-SIMA)5
40.4
49.9
41.3
52.1
43.9
2090
1520
2280
1490
1890
2.14
1.54
2.33
1.51
1.91
12.1
10.3
12.0
10.4
10.4
79
47
73
49
48
59.4
57.8
60.8
57.9
a Sample annotations: FeOx-P(OEGMA-co-SIMA)xy, with xy ) theoretical molar SIMA content in the copolymeric shell; µP ) mass content of copolymer
in the particles (EA); Ms ) saturation magnetization of saturated DMSO particle dispersions (VSM); µMF ) mass content of FeOx magnetic cores in
saturated DMSO dispersion (VSM); dc ) volume average core diameter (VSM); dh ) hydrodynamic diameter (DLS); Tc ) critical solution temperature
(cloud point photometry, CPP) of partly hydrolyzed particles.
Synthesis of CPTMS-Functionalized FeOx Nanoparticles. Mag-
netic FeOx nanoparticles are obtained by alkaline coprecipitation of
iron(II) and iron(III) chloride based on a method of Cabuil and
Massart.27 The particles are washed several times with deionized water
and then stirred for 5 min with 420 mL of 2 M nitric acid. After three
washing steps with nitric acid, the particle precipitate is resupended in
90 mL of 0.01 M citric acid to functionalize the particle surface with
citric acid. Again the nanoparticles are magnetically separated and
redispersed by the addition of aqueous tetramethyl ammonium hydrox-
ide (pH 7-8). For the surface modification, the magnetic fluid is added
dropwise to a stirred 1.8 mM solution of CPTMS in ethanol up to a
concentration of 1 g ·L-1 FeOx. After allowing the ligand exchange
reaction at ambient temperature overnight, the mixture is concentrated
on a rotary evaporator to a 10th of the original volume. The obtained
CPTMS functionalized nanoparticles are then magnetically separated
from the supernatant and washed five times with a 1:1 mixture of diethyl
ether and acetone. Finally, the particle precipitate is redispersed in
DMSO for the following surface initiated polymerization.
of reactive groups in a single polymerization step. The polymer
brush shells are prepared by a grafting from (surface-initiated)
atom transfer radical polymerization. By copolymerization of
a monomer with active ester functionality, we are able to
introduce a high density of accessible functional groups that
facilitate the nucleophilic attack of amine-terminated moieties.
This way, not only is the extra functionalization step avoided,
but the polymer becomes a template for many types of pendant
functional groups and molecules. While the controlled polym-
erization of active ester monomers in solution has frequently
been reported, there are few reports of those polymers grafted
from solid surfaces.24,25 We demonstrate that this strategy results
in a high specific number of binding sites, being up to two orders
of magnitude higher than for magnetic bead structures.
Their versatile application is demonstrated on two examples.
By reacting the active ester-functional moieties with trypsin,
we result in magnetically supported biocatalysts for protein
scission that provide quasi-homogeneous reaction conditions at
body temperature and easy separation at elevated temperatures.
Second, the successful capture of membrane proteins by active-
ester bearing particles has been demonstrated for endothelial
cells in vitro being of use for future applications of membrane
protein characterization.
ATR-IR spectra of dried functionalized FeOx nanoparticles show
several characteristic peaks that are also found in the spectrum of free
CPTMS. The vibrational absorption at 1604, 1480, 1259, and 1379
cm-1 and a broad signal between 1140 and 850 cm-1 confirm the surface
functionalization with CPTMS.
Surface Initiated ATR Copolymerization of OEGMA and
SIMA. Magnetic FeOx-P(OEGMA-co-SIMA) core-shell nanoparticles
are obtained by dissolving CuBr and bpy (1:2.5 mol/mol) in a DMSO-
based dispersion of CPTMS functionalized FeOx nanoparticles (35
mg·L-1 FeOx). OEGMA and SIMA (total monomer content 30 mass
%) are added to start the polymerization. The runs are based on two
initiator batches that have been analogously prepared as described
above, and the composition of the copolymers is chosen to achieve a
SIMA content in the polymer shell of 5, 10, and 20 mol % for one
initiator batch, and 15 and 25 mol % for the other initiator batch (Table
1). The reaction is carried out for 24 h under stirring at ambient
temperature. The obtained core-shell nanoparticles are transferred to
buffer-based dispersions by precipitation of the primary DMSO
dispersions in diethyl ether, magnetic separation and redispersion in
buffer.
Experimental Section
Analytical Methods and Instrumentation. DLS experiments and
zeta potential measurements are performed on a Malvern Zetasizer Nano
ZS at 25 °C. ATR-IR spectra are measured on a Nicolet 6700
spectrometer. Vibrating sample magnetometry (VSM) measurements
are implemented on an ADE Magnetics vibrating sample magnetometer
EV7. For UV/vis spectroscopy a Nicolet UV 540 spectroscope is used.
The phase behavior of aqueous particle dispersions is investigated on
a Tepper TP1 cloud point photometer at 1 K ·min-1 in HEPES buffer.
Elemental analyses are performed on a Perkin-Elmer 2400 CHN
analyzer. The polymer content is calculated through C content. TEM
pictures are taken on a Hitachi H 600.
Capture of Benzylamine. Five BzA solutions in HEPES-buffer are
prepared with concentrations between 20 mM and 2.5 mM. For each
of the investigated particle samples (FeOx-P(OEGMA-co-SIMA)20,
FeOx-P(OEGMA-co-SIMA)10, and FeOx-P(OEGMA-co-SIMA)5)
HEPES buffer-based dispersions with known particle concentrations
(FeOx-P(OEGMA-co-SIMA20, 34.9 mg ·L-1; FeOx-P(OEGMA-co-
Materials. Benzylamine (BzA; Janssen Chimica), NR-benzoyl-D,L-
arginin-4-nitroanilide hydrochloride (BAPNA; Sigma, 98%), 2,2′-
bipyridine (Aldrich, 99%), citric acid monohydrate (Gru¨ssing GmbH,
99,5%), copper(I) bromide (Aldrich, 98%), (4-(chloromethyl)phenyl)-
trimethoxysilane (CPTMS; ABCR, 95%), 1-(3-dimethylaminopropyl)-
3-ethylcarbodiimide hydrochloride (EDC; ABCR, 98%), iron(III)
chloride hexahydrate, iron(II) chloride tetrahydrate (Fluka, 98%),
N-hydroxysuccinimide (NHS; Aldrich, 98%), ninhydrin (Riedel-de-
Haen), oligo(ethylene glycol) methyl ether methacrylate (OEGMA;
Aldrich; Mn ) 290 g ·mol-1), tetramethylammoniumhydroxide aq.
(Aldrich, 25%), and trypsin type IX-S (Aldrich) are used as received.
Ethanol is obtained in technical grade and used after distillation.
Dimethylsulfoxide (DMSO) is dried by heating over calcium hydride
(Riedel-de-Haen) for two hours, followed by distillation under reduced
pressure. Succinimidyl methacrylate (SIMA) is synthesized by a method
by Gatz et al.26 HEPES buffer was prepared as follows: 11 mM HEPES
(Sigma), 140 mM NaCl (Merck), 4 mM KCl (Merck), and 10 mM
D(+)-glucose dissolved in deionized water.
SIMA10, 69.8 mg·L-1; FeOx-P(OEGMA-co-SIMA5, 139.6 mg·L-1
)
are prepared. For the binding experiments 0.5 mL of the respective
BzA solution is mixed with 1 mL of the particle dispersion and 0.5
mL of an activation solution composed of 18.73 mM EDC and 43.30
mM NHS and shaken overnight.
Quantitative Analysis of the BzA Residue by Ninhydrin
Reaction. To remove the strongly light-absorbing particles from the
reaction mixtures for UV analysis, the hybrids are magnetically
separated by a Milteny column or by heating above the LCST and the
use of a magnet. For the ninhydrin reaction, 1 mL of the respective
supernatant is mixed with 1 mL of a 40 mM ethanolic ninhydrin solution
in a pressure resistant vial and the vial is closed. The reaction is carried

Single Core Magnetic Nanostructures
Biomacromolecules, Vol. 11, No. 3, 2010 637
Figure 1. TEM images of (a) FeOx nanoparticles electrostatically stabilized by citric acid; (b) FeOx nanoparticles after surface-initiated ATRP.
out under stirring at 100 °C for 20 min. After cooling to ambient
temperature, the purple colored solutions are dissolved (1:10 and 1:5,
respectively), and the absorption at 568 nm is measured by UV/vis
spectroscopy.
Immobilization of Trypsin. Trypsin (30 mg) was dissolved in 6
mL HEPES buffer and mixed with 6 mL of a HEPES buffer-based
FeOx-P(OEGMA-co-SIMA)15 particle dispersion (µ(FeOx) ) 0.15
mass %). To allow reactivation of possibly hydrolyzed active-ester
functions, 6 mL of 2.21 µM EDC/NHS solution is added. The binding
reaction is carried out for 6 h at ambient temperature on a shaker. The
obtained trypsin functionalized particles were separated and washed
carefully with water to remove any residues of free trypsin and
redispersed in HEPES buffer.
More precisely, the synthesis of FeOx nanoparticles is carried
out by alkaline coprecipitation of iron(II) and iron(III) chloride
leading to magnetic colloids with an average diameter of about
10.5 nm. The size reproducibility in this synthesis is as good
as 5% for the number-average diameter, however, small
differences in the size distribution give a higher deviation for
the number-average value. By surface attachment of citric acid,
a stabilization of the nanoparticles in water is obtained due to
electrostatic repulsion.31 Figure 1a shows a TEM image of the
obtained small, spherical FeOx nanoparticles.
To functionalize the nanoparticle surface with benzylic
chlorine atoms for the initiation of ATRP, (p-chloromethyl)phe-
nyltrimethoxysilane (CPTMS) is attached via ligand exchange
to the particle surface and a cross-linked silica shell monolayer
is formed, caging the magnetic core due to condensation of the
silanol groups.31 The initiator density is calculated from
elemental analysis (EA) at 0.52 mmol ·g-1 and 0.64 mmol ·g-1,
respectively, for the two different batches employed in the
following polymerization runs, indicating good conditions for
an effective initiation and high grafting density.
OEGMA with an average number of ethylene glycol repeating
units of 4.4 and methoxy end group (Mn ) 290 g·mol-1) is
used as main monomer to generate a hydrophilic polymer shell
that shows a critical solution behavior in water leading to
magnetic polymer brushes with thermoresponsive dispersion
behavior at around 65 °C. The biocompatibility of poly(ethylene
glycol) derivates is helpful to obtain nanoparticles accepted for
use in in vitro biological systems.
The direct introduction of carboxy functions to the polymer
shell by surface-initiated ATRP involving (meth)acrylic acid is
hindered by catalyst poisoning, resulting in a loss of reaction
control.29 To overcome this, the protection of the carboxy group
is useful,32 and in our approach, we employed succinimidyl
methacrylate (SIMA) as a methacrylic acid derivative suitable
for ATRP.33-35 In model copolymerization experiments in
solution, we proved the copolymerization behavior of the two
monomers by analyzing the comonomer ratio and PDI at
different stages of the polymerization. Comonomer conversion
during polymerization was normally between 80 and 100%. The
results of the copolymerization experiment are particularly
important due to the fact that the respective hybrid particles
are not accessible to solution NMR. The results are summarized
in the Supporting Information. We found a constant comonomer
ratio throughout the course of the reaction in all runs and a
Determination of Immobilized Enzyme Kinetics and Acti-
vity. BAPNA is used as the model substrate. Four HEPES buffered
BAPNA solutions with concentrations between 2.0 and 0.5 mM, and
a 6.0 µM trypsin solution are prepared. The respective BAPNA solution
is added to a cuvette and mixed with 100 µL of FeOx-POEGMA-
trypsin nanoparticle dispersion or 50 µL trypsin solution. The resulting
concentrations of trypsin in the two systems are 9.82 × 10-8 mol·L-1
(free trypsin) and 1.66 × 10-7 mol·L-1 (FeOx-POEGMA-trypsin
particle dispersion). Directly after addition of the enzyme the change
in absorption at 410 nm is detected every 15 or 30 s over a period of
up to 20 min by UV spectroscopy.
Functionalization of Endothelial Cell Membranes. Cultured
endothelial cells (second passage) from human umbilical cord are
incubated with
a FeOx-P(OEGMA-co-SIMA)25 and a FeOx-
POEGMA nanoparticle dispersion (µ(FeOx) ) 0.12 mass %) for 5 min
in HEPES buffer at RT. Afterward, the cells are washed with HEPES
buffer and fixed for TEM imaging with a 1% paraformaldehyde and
1.25% glutaraldehyde solution in cacodylate buffer (0.1 M pH 7.4) for
one night at 4 °C. Fixation and embedding for TEM is carried out
according to standard procedures.28
Results and Discussion
Functional Magnetic Core-Shell Nanoparticle Syn-
thesis. Based on our recent results on the surface-engineering
of superparamagnetic nanoparticles with polymer brush
shells,18,19,29,30 we developed a synthetic pathway to active ester-
functional magnetic core shell nanoparticles. Starting from
citrate-stabilized, highly magnetic iron oxide particles (FeOx)
in water, we modify the particle surface with an ATRP initiator.
These macroinitiators are then used in the following surface
initiated ATRP to result in magnetic hybrid particles.

638 Biomacromolecules, Vol. 11, No. 3, 2010
Gelbrich et al.
Scheme 1. Synthesis of FeOx-P(OEGMA-co-SIMA) Magnetic Polymer Brush Particles by Surface-Initiated ATRP
narrow molar mass distribution at SIMA contents <40 mol %.
The copolymer composition is adjustable by the molar ratio of
the monomers (Supporting Information). Furthermore, the
succinimidyl ester function remains stable under the conditions
of polymerization, as proven by NMR and IR, and can be used
as an active ester in the amide formation for the nucleophilic
attack of primary amines,33-35 while OEGMA as the main
comonomer determines the solution properties. For model
copolymers up to 30 mol % SIMA fraction, we have found high
water solubility and a lower critical solution behavior, in which
the critical solution temperature Tc decreases with increasing
SIMA content.
the succinimidyl ester is still existent and neither hydrolyzed
nor deactivated.33 From the carbon content obtained by elemen-
tal analysis (EA), a mass content of copolymer between 40 and
52 mass % is calculated (Table 1).
The hydrodynamic diameter of the core-shell nanoobjects
in aqueous dispersion can be detected by dynamic light
scattering (DLS). Compared to electrostatic stabilized particles
and to CPTMS functionalized particles with a volume-average
hydrodynamic diameter dh of 19 and 26 nm, respectively, the
size clearly increases for P(OEGMA-co-SIMA) coated nano-
particles to values between 47 and 79 nm, resulting from the
formation of a polymer brush shell. It strikes that two of the
runs, that is FeOx-P(OEGMA-co-SIMA)25 and FeOx-
P(OEGMA-co-SIMA)15, show higher values for dh in comparison
to the other investigated samples, although the polymer fraction is
not differing too much from the other batches. We assume the
reason for that is the employment of a different initiator particle
batch in these runs. As observed in VSM, the cores employed here
show a higher volume-average diameter dc compared to the other
runs. The surface-modified particles thus provide a lower specific
surface and, therefore, a lower specific initiator functionality (see
above, at comparable initiation sites density of 4.9 µmol·m-2). At
a similar polymer content, this results in longer chains being fixed
on the cores with a weaker surface curvature and, therefore, leads
to a higher hydrodynamic radius.
Magnetic polymer brush particles containing active ester units
have been prepared by surface-initiated copolymerization of
SIMA and OEGMA from CPTMS-modified FeOx nanoparticles
via ATRP in dimethylsulfoxide (DMSO; Scheme 1).
After 24 h, black viscous DMSO-based magnetic fluids are
obtained. The success of the surface-initiated ATRP is quali-
tatively analyzed by transmission electron microscopy (TEM)
and ATR-IR spectroscopy. A TEM image (Figure 1b) of the
obtained nanoparticles visualizes strongly contrasting FeOx cores
surrounded by less contrasting polymer shells. The small
nanoparticles are separately covered with a polymer layer of
an average thickness of 3 nm independent of the core size.
ATR-IR spectra (Figure 2) of the dry FeOx-P(OEGMA-co-
SIMA) nanoparticles feature signals relating to the vibrational
absorption of polymeric methyl and methylene groups (υ )
2800-3050 cm-1), carbonyl double bond (υ ) 1722 cm-1),
C-O deformation (υ ) 1099 cm-1), and N-O deformation (υ
) 1025 cm-1) clearly reveals the presence of P(OEGMA-co-
SIMA). The three distinct peaks at υ ) 1807, 1778, and 1722
cm-1 are characteristic of the vibrational absorption of the three
carbonyl double bonds of the SIMA function, indicating that
The quasi-static magnetic properties of FeOx-P(OEGMA-
co-SIMA) dispersions are investigated by vibrating sample
magnetometry (VSM) experiments. Figure 3 shows a charac-
teristic magnetization loop. The sigmoidal shape with low
coercivity values (<0.6 kA ·m-1) is attributed to superparamag-
netic behavior of the core-shell nanoparticles. The saturation
magnetization obtained from the experiments indicates a FeOx
Figure 2. ATR-IR spectra of FeOx-P(OEGMA-co-SIMA)25; trypsin,
and trypsin-functional core-shell particles FeOx-POEGMA-trypsin
(dry powders).
Figure 3. VSM loop of saturated magnetic fluid based on FeOx-
P(OEGMA-co-SIMA)15 nanoparticles in DMSO (µMF ) 2.33 mass
%).

Single Core Magnetic Nanostructures
Biomacromolecules, Vol. 11, No. 3, 2010 639
and N-hydroxysuccinimide (NHS) is added to reactivate hy-
drolyzed SIMA functions. On completion of the binding reaction
overnight, the BzA functionalized nanoparticles are magnetically
separated from the carrier medium at T > LCST. There is no
significant change of the LCST behavior observed due to BzA
binding. The residual BzA in the carrier medium is quantified
by ninhydrin reaction to Ruhemanns purple and subsequent
detection by UV spectroscopy (Scheme 2).39,40
Figure 5a compares UV spectra of two supernatants and
magnetic amine separation by FeOx-P(OEGMA-co-SIMA)20
nanoparticles, and by FeOx-POEGMA nanoparticles without
active ester or carboxy groups in the polymeric shell, after
ninhydrin reaction. As a reference, the UV spectrum of the basic
2.5 mM BzA solution after ninhydrin reaction without magnetic
separation treatment is also shown. All spectra indicate the
presence of the adduct by the presence of two absorption
maxima at 400 and 568 nm for all samples. As evident, the
spectra of the reference BzA solution and the BzA solution after
magnetic separation with FeOx-POEGMA nanoparticles are
almost matching, indicating a comparable adduct concentration
in both samples and, consequently, the absence of unspecific
binding of BzA to the nano-objects that do not show active
ester groups.
Figure 4. Relative transmittance τ vs temperature T of (partly
hydrolyzed) FeOx-P(OEGMA-co-SIMA)20 (compact line) and FeOx-
POEGMA (dashed line) suspensions in water in CPP experiments.
content between 1.5 and 2.3 mass %, according to the Langevin
equation, and by using the initial susceptibility ꢀini from the
graphs slope at H ) 0, the volume-average core diameter can
be calculated to values between 10.3 and 12.1 nm (Table 1).36
Zeta potential measurements on aqueous particle dispersions
indicate low surface charges below +- 5 mV in a pH range
between 3 and 8, and a poor pH dependence. The results
therefore confirm the predominantly uncharged polymer shell
in this pH range. However, when the pH is raised further, the
zeta potential drops to -30 mV, indicating a hydrolysis of part
of the active ester groups to carboxylate units in accordance to
literature reports on the hydrolytic stability of n-succiminidyl
esters.24,25 (see Supporting Information).
In contrast, when active ester bearing FeOx-P(OEGMA-co-
SIMA)20 nanoparticles are employed in the separation, the UV
absorption after magnetic amine separation and ninhydrin
reaction is significantly reduced in comparison to the reference
solution.
In this context, the results of cloud point experiments
performed on more concentrated particle dispersions as shown
in Figure 4 have to be interpreted carefully. A sudden decrease
in the relative transmittance τ is detected between 56 and 64
°C, caused by the increasing turbidity of the dispersion due to
agglomeration and precipitation of the nanoparticles. This
process is fully reversible and can be attributed to the lower
critical solution temperature (LCST) of the OEGMA units of
the polymer shell in water.22 In contrast to what is found for
model copolymers (see Supporting Information), the optically
determined critical solution temperature Tc of the phase separa-
tion shows only a weak dependence on the initial SIMA content
of the copolymer shell (see Table 1). As earlier results indicate,
a similar thermoresponsive solution behavior of POEGMA
derivatives immobilized on flat or nanoparticular surfaces as
compared to linear soluted chains,22,37,38 we attribute this to
the occurrence of hydrolysis of a considerable part of the
hydrophobic active ester groups during sample preparation.
In the agglomerated state above the LCST, simple permanent
magnets with magnetic field gradients below 50 mT/cm are
sufficient to separate the magnetic polymer brush particles from
the carrier medium. We show below that this behavior is of
use for the easy magnetic separation of amino-functional probes
and magnetically labeled biomolecules. In this respect, it is of
interest to note that the agglomeration temperature can be
adjusted by copolymerization in a wide range including tem-
peratures acceptable for biomolecules and biological species.
We focus on this issue in our upcoming paper.
After suitable calibration, the absorption at the peak maximum
(568 nm) is used to extract the amount of residual BzA, cr,BzA
.
We performed a representative experimental series and used
the results to extract the particle functionality for three different
runs of particles. By plotting cr,BzA versus the initial BzA
concentration c0,BzA, normalized by the magnetite concentration
used for separation, cFeO (Figure 5b), we observe linear graphs
with an onset on the x cox ordinate that can be used to calculate
the particles functionality.
For small initial BzA concentrations c0,BzA, virtually no
residual agent cr,BzA is detected, indicating that below a certain
concentration nearly the full portion of BzA is captured by the
particles. For BzA concentrations higher than this value, a sharp
transition to a linear increase of cr,BzA with c0,BzA/cFeO , giving
x
cFeO as the slope, is detected. The intersections of the linear
x
fits with the x-axis give the maximum amount of BzA that can
be captured by the given content of hybrid particles. From the
observed behavior, we may reasonably consider this value to
be connected to the active ester functionality of the core-shell
particles. For the investigated samples FeOx-P(OEGMA-co-
SIMA)20,FeOx-P(OEGMA-co-SIMA)10,andFeOx-P(OEGMA-
co-SIMA)5, values for f of 29.5, 12.5, and 2.87 mmol ·g-1 are
obtained, reflecting the differing SIMA content in the polymeric
shell resulting from the polymerization conditions. In contrast,
commercially available, carboxy- or active ester bearing mag-
netic beads for magnetic separation typically show functional-
ities below 1 mmol ·g-1.41 Owed to their brush-like architecture,
our hybrid core shell nanoparticles possess a superior binding
capacity of up to two scales higher, combined with quasi-
homogeneous binding conditions in the highly solvated polymer
shell, and easy magnetic separation above the LCST temperature.
Magnetic Biocatalysts by Enzyme Immobilization. The
immobilization of biomacromolecules on magnetic carriers is
of interest for separable biocatalytic systems. For a proof of
principle study, we successfully immobilized trypsin as a model
enzyme on the shell of FeOx-P(OEGMA-co-SIMA)15 nano-
Determination of Particle Functionality by Amination. To
investigate the (protected) COOH functionality and the binding
capacity of FeOx-P(OEGMA-co-SIMA) nanoparticles for
primary amines, the capture of benzylamine (BzA) as a model
substance from aqueous solution is examined. For this purpose,
BzA solutions of different concentration are mixed with HEPES-
buffered FeOx-P(OEGMA-co-SIMA) particle dispersions and
stirred at room temperature. In addition, a solution of 1-(3-
dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (EDC)

640 Biomacromolecules, Vol. 11, No. 3, 2010
Gelbrich et al.
Figure 5. (a) UV spectra of residual BzA solution after ninhydrin reaction and magnetic amine separation by FeOx-P(OEGMA-co-SIMA)20
(dotted line) and FeOx-POEGMA (dashed line) nanoparticles in comparison to the primarily 2.5 mM BzA solution (compact line). (b) Concentration
of residual BzA in solution cr,BzA after magnetic amine separation vs quotient of initial BzA concentration c0,BzA and FeOx concentration cFeO
;
x
open triangles (compact line): FeOx-P(OEGMA-co-SIMA)20; open circles (dotted line): FeOx-P(OEGMA-co-SIMA)10; solid squares (dashed
line): FeOx-P(OEGMA-co-SIMA)5.
Scheme 2. Quantification of Amine Separation by FeOx-P(OEGMA-co-SIMA) Particles: (1) Separation of Magnetic Brush Particles at T >
LCST after Amination, and (2) Staining of Residual Amine in the Supernatant with Ninhydrin to Ruhemanns Purple and Quantification by UV
Spectroscopy
a
Scheme 3. Enzymatic Scission of BAPNA Catalyzed by Magnetically Labeled Trypsin (
)
a The reaction product nitroaniline can be quantified spectroscopically.
particles. Particles with intermediate SIMA functionality have
been chosen to ensure enough binding sites for amine capture,
yet avoiding particle cross-linking or agglomeration due to
multiple attachment of a single protein molecule by several
nano-objects. Figure 2 compares the ATR-IR spectra of trypsin-
functional nanoparticles FeOx-POEGMA-trypsin to free trypsin.
In both samples, we observe similar amide signals (υ ) 1620,
1578 cm-1), and also the NH-signal (υ ) 3284 cm-1) of the
trypsin peptide sequence is visible in both spectra. From the
nitrogen content obtained by EA, the amount of trypsin bound
to the polymer surface of the nanoparticles is calculated to 1.1
µmol·g-1 or 26 mg ·g-1. This loading is higher than that of
The catalytic activity of trypsin, a protease for hydrolysis of
specific peptide bonds (chain scission after the amino acids
arginine and lysine), is investigated for the particle-immobilized
trypsin compared to free trypsin. For this purpose, we use the
release of p-nitroaniline by the trypsin-mediated scission of NR-
benzoyl-D,L-arginine-4-nitroanilide (BAPNA) for the detection
of the catalytic properties (Scheme 3).43
In our kinetic experiments concerning the trypsin-catalyzed
scission reaction of BAPNA, the concentration of p-nitroaniline
is followed throughout the reaction by UV spectroscopy at the
absorption maximum of nitroaniline at 410 nm. An analogous
set of experiments was carried out with native trypsin solution,
as well as with FeOx-POEGMA-trypsin (particle-immobilized
trypsin) in aqueous dispersion (for details, see Experimental
Section).Asacontrolexperiment,theprimarilyFeOx-P(OEGMA-
co-SIMA)15 nanoparticle dispersion without trypsin bound to
the polymer shell is also used. Respective results for selected
experiments are shown in Figure 6a.
commercially available magnetic particles for the protein binding
42
with reported capacities between 1.5 mg ·g-1 and 20 mg ·g-1
.
On the other hand, it is comparably lower than calculated from
the active ester functionality of the hybrid particles reported
above. We ascribe this to the different molecules size between
BzA, employed for active ester group evaluation, and the
protein, compared to the particle size and the proposed free
volume between polymer chains. With a molar mass of 23300
g·mol-1, the molecule size of trypsin is in the same order as
Mn of the surface immobilized polymer chains. Additionally,
trypsin exhibits several amino groups within the protein
sequence so that several active ester bearing polymer arms bind
one trypsin molecule.
In all runs employing either trypsin or immobilized trypsin,
a linear increase of absorption with time can be detected in UV
experiments, while the control does not react with BAPNA. To
exclude possible trypsin leaching from the carriers, we continued
the data collection for a couple of minutes after magnetic

Single Core Magnetic Nanostructures
Biomacromolecules, Vol. 11, No. 3, 2010 641
Figure 6. (a) UV absorption (A-A0) at 410 nm over time t during reaction between BAPNA (2.0 mM) and FeOx-POEGMA-trypsin nanoparticles
(open circles) compared to free trypsin (solid squares) and FeOx-P(OEGMA-co-SIMA)15 (solid triangles) nanoparticles. (b) Double-reciprocal
plot of the rate of trypsin-catalyzed BAPNA scission, dcBAPNA/dt, and the BAPNA concentration, cBAPNA, according to Lineweaver and Burk; open
circles: FeOx-POEGMA-trypsin (c(trypsin) ) 1.66 × 10-7 mol·L-1); solid squares: free trypsin (c(trypsin) ) 9.82 × 10-8 mol·L-1).
Figure 7. (a) TEM image of a human endothelial cell after incubation with FeOx-P(OEGMA-co-SIMA)25 nanoparticles; (b) detail marked in (a)
showing nanoparticles attached to the cell surface; (c) magnification of an endothelial cell surface efficiently coated with nanoparticles.
separation of FeOx-POEGMA-trypsin nanoparticles from the
BAPNA solution. No further increase in adsorption was
detected.
immobilized trypsin, indicating no significant loss of enzyme
activity upon immobilization. A likely explanation for the
observed difference in the obtained Michaelis-Menten constant
KM is the hindered accessibility of the trypsin molecules bound
to the polymer brush shell to its substrate molecules. This effect
might be of even higher relevance for large molecules such as
proteins and will be addressed in future experiments.
The active ester-functional FeOx-P(OEGMA-co-SIMA) nano-
particles form covalent amide bonds with primary amines, and
can thus be used to selectively label aminofunctional compo-
nents of the exposed surface of a cell, for example, membrane
proteins, to simplify their isolation and identification. To verify
this concept, we performed labeling experiments of endothelial
cell surface membranes. Cultured endothelial cells from human
umbilical cord are incubated with FeOx-P(OEGMA-co-
SIMA)25 nanoparticles, and, as a control, with particles without
binding sites for amino groups (FeOx-POEGMA). In Figure
7, the successful binding is visualized by transmission electron
microscopy (TEM). The surface of endothelial cells that has
been exposed to a FeOx-P(OEGMA-co-SIMA)25 dispersion
shows an effective and dense labeling with nanoparticles after
several washing steps. In contrast, nonfunctional FeOx-
POEGMA particles do not attach to the cell surface (Supporting
Information).
The experiments have been performed with varying BAPNA
cBAPNA concentrations to obtain quantitative information on the
kinetic parameters.44 In Figure 6b, a double-reciprocal plot of
the rate dcBAPNA/dt is plotted against cBAPNA according to
Lineweaver and Burk.45 The obtained linear relationship
indicates that the trypsin-catalyzed scission of BAPNA follows
a Michaelis-Menten mechanism for both free and immobilized
trypsin. It is therefore possible to extract the kinetic parameters
from the Michaelis-Menten equation:
[S]
V ) V
(1)
maxKM + [S]
with V, conversion rate; Vmax, maximum conversion rate (at
infinite substrate concentration); KM, Michaelis-Menten equi-
librium constant; and [S], substrate concentration (here, cBAPNA).
The determination of KM and Vmax for the two systems is done
by the method of Eadie and Hofstee from a single-reciprocal
plot of dcBAPNA/dt against the quotient of dcBAPNA/dt divided by
46-48
cBAPNA
.
For free trypsin, KM is calculated at 1.39 × 10-3
Recently, we have shown that the labeled species can be
magnetically separated and readily isolated after cell lysis.
Digestion and identification of the species is subject to ongoing
examination and beyond the scope of this paper.
M and Vmax at 2.11 × 10-4 M·s-1, and for FeOx-POEGMA-
trypsin nanoparticles, KM ) 2.10 × 10-2 M and Vmax ) 3.45 ×
10-4 M·s-1 for the respective trypsin concentration used in the
experiment. To compare the enzymatic activity of free and
immobilized trypsin, the maximum specific activity actspez, given
by the ratio of Vmax and the trypsin concentration, was used in
the experiment. We obtain a specific activity actspez of 92.2
µmol·s-1 ·mg-1 for free trypsin and 89.1 µmol·s-1 ·mg-1 for
Conclusion
The results demonstrate that by single-step surface-initiated
copolymerization of active ester monomer SIMA with OEGMA,

642 Biomacromolecules, Vol. 11, No. 3, 2010
Gelbrich et al.
(10) Sun, Y.; Ding, X.; Zheng, Z.; Cheng, X.; Hu, X.; Peng, Y. Chem.
Commun. 2006, 2765–2767.
(11) Ohnishi, N.; Furukawa, H.; Hideyuki, H.; Wang, J.-M.; An, C.-I.;
Fukusaki, E.; Kataoka, K.; Ueno, K.; Kondo, A. BioNanotechnology
2006, 2, 43–49.
brush-coated magnetic particles are obtained that possess a high
number of accessible carboxy-analogue functional groups that
can be further modified with bioactive species with the purpose
of their manipulation or isolation. Here, we have shown that
functional group densities of up to 30 mmol per gram of
magnetite could be achieved that are accessible by small
molecule substrates.
The particles are easily dispersed in aqueous media up to 23
mg/mL magnetite. The thermoreversible solvation of the shell
is responsible for thermoflocculation, that results in a fast and
highly effective magnetic isolation of the hybrid structures above
the transition temperature.
As a result of the high functional group density, the particles
show superior behavior in in vitro biocatalysis and labeling
experiments. By immobilization of trypsin to the polymeric shell
of the obtained nanoparticles, magnetically supported biocata-
lysts are obtained. Kinetic experiments on the catalytic activity
via the trypsin mediated scission reaction of BAPNA reveal no
loss in specific enzyme activity when compared to soluble
trypsin. In addition, we have demonstrated the nanostructures’
potential for selective tagging and separation of the exposed
surface of endothelial cells in order to facilitate the isolation
and identification of membrane proteins. While unspecific
adsorption of nonfunctional particles was not observed, in vitro
experiments involving active ester group bearing core-shell
particles resulted in a dense population of particles at the
exposed cell membrane, attributed to the formation of covalent
(amide) bonds.
(12) http://www.magnabeat.com/e_index.html.
(13) Perruchot, C.; Khan, M. A.; Kamitsi, A.; Armes, S. P.; von Werne,
T.; Patten, T. E. Langmuir 2001, 17, 4479–4481.
(14) Chen, X. Y.; Randall, D. P.; Perruchot, C.; Watts, J. F.; Patten, T. E.;
von Werne, T.; Armes, S. P. J. Colloid Interface Sci. 2003, 257, 56–
64.
(15) Li, D.; Sheng, X.; Zhao, B. J. Am. Chem. Soc. 2005, 127, 6248–6256.
(16) Li, D.; Jones, G. L.; Dunlap, J. R.; Hua, F.; Zhao, B. Langmuir 2006,
22, 3344–3351.
(17) Schmidt, A. M. Macromol. Rapid Commun. 2005, 26, 93–97.
(18) Gelbrich, T.; Feyen, M.; Schmidt, A. M. Macromolecules 2006, 39,
3469–3472.
(19) Kaiser, A.; Gelbrich, T.; Schmidt, A. M. J. Phys.: Condens. Matter
2006, 18, 2563–2580.
(20) Wang, S.; Zhou, Y.; Guan, W.; Ding, B. Appl. Surf. Sci. 2008, 254,
5170–5174.
(21) Gou, Z.; Chen, Y.; Zhou, W.; Huang, Z.; Hu, Y.; Wan, M.; Bai, F.
Mater. Lett. 2008, 62, 4542–4544.
(22) Chanana, M.; Jahn, S.; Georgieva, R.; Lutz, J.-F.; Ba¨umler, H.; Wang,
D. Chem. Mater. 2009, 21, 1906–1914.
(23) Lutz, J.-F.; Stiller, S.; Hoth, A.; Kaufner, L.; Pison, U.; Cartier, R.
Biomacromolecules 2006, 7, 3132–3138.
(24) Murata, H.; Prucker, O.; Ru¨he, J. Macromolecules 2007, 40, 5497–
5503.
(25) Cullen, S. P.; Mandel, I. C.; Gopalan, P. Langmuir 2008, 24, 13701–
13709.
(26) Batz, H.-G.; Franzmann, G.; Ringsdorf, H. Angew. Chem. 1972, 24,
1189–1190.
(27) Massart, R.; Cabuil, V. J. Chem. Phys. 1987, 84, 967–973.
(28) Schimmel, G., Vogell, W., Eds. Methodensammlung der Elektronen-
mikroskopie; Wissenschaftliche Verlagsgesellschaft mbH: Stuttgart,
1983.
(29) Kaiser, A.; Schmidt, A. M. J. Phys. Chem. B 2008, 112, 1894–1998.
(30) Gelbrich, T.; Feyen, M.; Schmidt, A. M. Z. Phys. Chem. 2006, 220,
41–49.
(31) Frickel, N.; Messing, R.; Gelbrich, T.; Schmidt, A. M. Langmuir 2009,
DOI: 10.1021/la902904f.
(32) Matyjaszewski, K.; Davis, T. P. Handbook of Radical Polymerization;
WileyIntersience: New York, 2002.
(33) Rathforn, J. M.; Tew, G. N. Polymer 2008, 49, 1761–1769.
(34) Shunmugam, R.; Tew, G. N. J. Polym. Sci., Part A: Polym. Chem.
2005, 43, 5831–5843.
(35) Monge, S.; Haddelton, D. M. Eur. Polym. J. 2004, 40, 37–45.
(36) Odenbach, S. MagnetoViscous Effects in Ferrofluids; Springer-Verlag,
2002.
(37) Jones, A. M.; Glinel, K.; Oren, R.; Nysten, B.; Huck, W. T. S.
Macromolecules 2007, 40, 4403–4405.
The principle described here is applicable to the modification
with other biologically or catalytically relevant groups and will
therefore open new ways for the design of multifunctional hybrid
nanostructures with different property portfolios.
Acknowledgment. We thank Prof. H. Ritter for his steady
interest in our studies. Thanks to Dr. B. Emde for her support
with biological and technical questions. We thank Dr. Yan Lu
University of Bayreuth (Germany) for TEM images of the
particles. We gratefully acknowledge DFG for financial support
within the Emmy-Noether Program and the Priority Program
SPP 1259.
Supporting Information Available. Synthesis and properties
of P(OEGMA-co-SIMA) model copolymers, zeta potential
measurements, and TEM of cell labeling control experiment.
This material is available free of charge via the Internet at http://
pubs.acs.org.
(38) Lutz, J.-F.; Hoth, A. Macromolecules 2006, 39, 893–896.
(39) Sarin, K. V.; Kent, S. B. H.; Tam, J. P.; Merrifield, R. B. Anal.
Biochem. 1981, 117, 147–157.
(40) McCaldin, D. J. Chem. ReV. 1960, 60, 39–51.
(41) www.ademtech.com, adembeads, f ) 0.35 mmol·g-1; www.bangslabs.
com, BioMag, f ) 0.24 mol · g-1; www.chemagen.de, M-PVA
Magnetic beads activated AK11, f ) 0.39 mmol ·g-1; www.invitrogen.
References and Notes
(1) Hora´k, D.; Babie`, M.; Mackova´, H.; Benesˇ, M. J. J. Sep. Sci. 2007,
30, 1751–1772.
(2) Gu, H.; Xu, K.; Xu, C.; Xu, B. Chem. Commun. 2006, 941–949.
(3) Franzrep, M.; Siemann-Herzberg, M.; Hobley, T. J.; Thomas, O. R. T.
Appl. Microbiol. Biotechnol. 2006, 70, 505–516.
(4) Safarik, I.; Safarikova, M. BioMagn. Res. Technol. 2004, 2, 1–17.
(5) Pankhurst, Q. A.; Connolly, J.; Jones, S. K.; Dobson, J. J. Phys. D:
Appl. Phys. 2003, 36, R167–R181.
(6) Safarik, I.; Safarikova, M. J. Chromatogr., B 1999, 722, 33–53.
(7) Yalcy´n, A.; Damin, F.; Ozkumur, E.; di Carlo, G.; Goldberg, B. B.;
Chiari, M.; Unlu, M. S. Anal. Chem. 2009, 81, 625–630.
(8) Kondo, A.; Kamura, H.; Higashitani, K. Appl. Microbiol. Biotechnol.
1994, 44, 99–105.
com, Dynabeads MyOne, f ) 0.4-0.8 mmol ·g-1
.
(42) www.chemagen.de,M-PVA Magnetic beads activated AK11 protein
binding capacity, 8-20 mg ·g-1; www.micromod.de, nanomag-D
(streptavidin) streptavidin functionality, 1.5-2.0 mg ·g-1
.
(43) Erlanger, B. F.; Kokowsky, N.; Cohen, W. Arch. Biochem. Biophys.
1961, 95, 271–278.
(44) Michaelis, L.; Menten, M. L. Biochem. Z. 1913, 49, 333–369.
(45) Lineweaver, H.; Burk, D. J. Am. Chem. Soc. 1934, 56, 658–666.
(46) Hofstee, B. H. J. Nature 1959, 184, 1296–1298.
(47) Hofstee, B. H. J. Science 1952, 116, 329–331.
(48) Eadie, G. S. J. Biol. Chem. 1942, 146, 85–93.
(9) Lai, J. J.; Hoffman, J. M.; Ebara, M.; Hoffman, A. S.; Estourne‘s, C.;
Wattiaux, A.; Stayton, P. S. Langmuir 2007, 23, 7385–7391.
BM901203Z