M Bloem en Thesis Final

Immunomagnetic separation of bacteria by iron oxide
nanoparticles
Maarten Bloemen
Examination committee: Dissertation presented in partial

Prof. dr. M. Smet, chair fulfillment of the requirements for
Prof. dr. T. Verbiest, supervisor the degree of Doctor
dr. N. Geukens, co-supervisor in Science
Prof. dr. A. Gils, co-supervisor
Prof. dr. G. Koeckelberghs
Prof. dr. S. De Feyter
Prof. dr. F. Ollevier
Prof. dr. M. Ameloot
(UHasselt)
May 2015
© 2015 KU Leuven – Faculty of Science
Uitgegeven in eigen beheer, Maarten Bloemen , Celestijnenlaan 200D box 2425, B-3001 Heverlee (Belgium)
Alle rechten voorbehouden. Niets uit deze uitgave mag worden vermenigvuldigd en/of openbaar gemaakt worden
door middel van druk, fotokopie, microfilm, elektronisch of op welke andere wijze ook zonder voorafgaande
schriftelijke toestemming van de uitgever.
All rights reserved. No part of the publication may be reproduced in any form by print, photoprint, microfilm,
electronic or any other means without written permission from the publisher.
The story of science is often told as a series of eureka moments.
But the truth is that influence, passion and sheer blind chance
have played equally significant parts.
Michael Mosley
Preface
The booklet in front of you is the result of my great PhD adventure. It was a
combination of many bold ideas, successes (and as many failures), perseverance,
persistence, enthusiasm and the occasional disappointment when a chemical
reaction was not cooperating. Even though there is only one name on the cover,
this PhD was a combined effort by many people, who contributed scientifically,
practically or mentally. I would like to seize this opportunity to thank everyone
who helped, influenced or supported me.
First and foremost, I want to thank Prof. Thierry Verbiest for being a terrific
promotor and supervisor. I fully enjoyed the early morning coffee moments,
brainstorming sessions and scientific discussions (even though you were mostly
right in the end). Your enthusiasm and never-ending optimism created a great
working environment. Thank you for giving me the freedom to pursue my own
ideas and the possibility to see the world by visiting conferences abroad. I
honestly cannot imagine a better promotor!
I was given the unique opportunity to conduct multidisciplinary research,
combining chemistry with pharmacology and biology. It goes without saying
that my limited knowledge in these new fields had to be compensated. Therefor
I would like to thank my co-supervisors Prof. Ann Gils and dr. Nick Geukens
for their support, mentoring, insights and for providing me with (ridiculously)
large amounts of antibodies whenever I needed them. The synergy between
our research has proven to be very valuable. I would also like to express my
gratitude for funding the first year of my PhD.
While performing experiments away from the safe haven of chemistry, I received
(read: I needed) a lot of practical assistance and guidance. On the pharmacology
side, Miet Peeters and Thomas Van Stappen helped me out regularly. Thank
you for teaching me the basics of ELISAs and protein purification, and answering
all my questions (over and over again). On the biology side, Carla Denis and
dr. Louise Vanysacker introduced me into the world of bacteria. Special thanks
i
ii PREFACE
to Carla for doing all preparations, guidance and performing the DNA-related
measurements!
Collaborations with other research groups allowed me to use their scientific
equipment and discuss the results with experts in their respective fields. Special
thanks to the groups of Prof. Luc De Meester, Prof. Guy Koeckelberghs, Prof.
Koen Binnemans, Prof. Tatjana Parac-Vogt , Prof. Carmen Bartic, Prof. Bart
Goderis, Prof. Wim Annaert and Prof. Jeroen Lammertyn. More specific, I
am grateful for all assistance by dr. Pieter Willot, Frederic "Fred" Monnaie,
David Dupont, Sophie Carron, Vincent Goovaerts, Matthias Ceulemans, David
Debruyne, dr. Marlies Van De Wouwer, Arun Kumar Tharkeshwar and dr.
Karel Knez. Not to forget, my exemplary master student Ben, for all the hard
(and perhaps repetitive) work in the lab.
I also would like to express my gratitude to the jury members: Prof. Mario
Smet (KU Leuven), Prof. Guy Koeckelberghs (KU Leuven), Prof. Steven de
Feyter (KU Leuven), Prof. Frans Ollevier (KU Leuven) and Prof. Marcel
Ameloot (UHasselt). Thank you for reading my dissertation and providing me
with comments and remarks, that greatly improved the manuscript.
Scientific research is dependent on funding, so I want to thank the Agency for
Innovation by Science and Technology (IWT) Flanders for awarding me a grant
that funded my PhD.
Thanks to all of my LCBD colleagues for the fun and laughter, lab trips, cake-
time (for every possible reason) and coffee breaks. Special thanks to PJ for
being my fellow lab responsible and making our lab a safer and cleaner place,
and to Rik for turning my ideas into working hardware and the swift repairs. Of
course I cannot forget my office buddies: Ward, the other Maarten, Stefaan and
Stefan; for making our office the best (and most productive) in the building. We
clearly had a shared preference for a certain music genre (at least on Friday).
Luckily there is more to life than just working and studying. Throughout my
time in Leuven I had the pleasure of sharing countless evenings with my friends
Bram, Bellie (the other Maarten) en Kevin. The 2-for-1 pizza and game nights
were legendary!
I also would like to thank the Rebus-clan for the sportive evenings, BBQs and
bar visits. And of course all my friends from Geel and surroundings, for the
weekend entertainment and great New Year’s eve parties.
To my parents, thank you for supporting, stimulating and encouraging me
throughout my studies and for always showing great interest in my work.
Driving back and forth to Leuven on numerous occasions was a tedious task
but it was worth it in the end. Special mention to my zusje and broertje, for all
PREFACE iii
the joking around, dinners and mental support.

And last but definitely not least, my girlfriend Sara, thank you for supporting
me, sharing my happiness and frustration, and providing me with lots of hugs
and love. My tendency to get up as early as possible, "nerdy" talking, total
lack of patience (seriously, none whatsoever) and exaggerated never-ending
happiness were sometimes hard to bear, but you never complained. You are
awesome!
Thank you all!!!!
Maarten
Abstract
The development of multifunctional nanomaterials has proven to be beneficial

for many applications. Nanoparticles in particular have been intensively
investigated, which resulted in large advances in particle synthesis and control
of their properties. From a biomedical point of view, iron oxide nanoparticles
are especially valuable thanks to their biocompatibility and low toxicity. Their
magnetic properties are already applied in medical imaging, hyperthermia for
cancer therapy and advanced drug delivery techniques.
In this dissertation we present the development of multifunctional iron oxide
nanoparticles that can be used for the magnetic isolation of bacteria from an
aqueous solution, as a pre-concentration technique.
The synthesis and characterization of the core nanoparticle is described in the
first part of this work. An enhanced method of surface functionalization was
developed, that enabled us to efficiently coat the particles with multiple different
ligands. We applied a covalent surface modification method to improve the
robustness and long-term stability and storage of the material. The influence
of the different ligands on the colloidal stability was investigated, providing a
solid base for further development.
In order to provide sufficient colloidal stability to the nanoparticles, a new
ligand was designed and synthesized. By modifying a poly(ethylene glycol)
chain with two different functional groups, a new heterobifunctional ligand
was produced. Thanks to the presence of the water soluble backbone, the
functionalized nanoparticles showed excellent stability in various complex
environments. Moreover, we covalently coupled antibodies to these particles and
confirmed that their activity was retained, which is crucial for their applicability.
The synthesis protocol to form these bioconjugated nanoparticles is described
in the second part of this work.
Legionella pneumophila bacteria were chosen as the model target for magnetic
isolation experiments. These bacteria are known to cause various diseases,
v
vi ABSTRACT
including severe types of pneumonia. Consequently their detection is mandatory

by law in all public water systems. We investigated by flow cytometry
measurements, whether or not the bioconjugated nanoparticles are capable
of isolating these bacteria from an aqueous solution. The results indicated that
the developed nanoparticles can very efficiently isolate these targeted organisms.
Furthermore, we were able to capture L. pneumophila from a mixture of bacteria,
which underlines the potential and robustness of the proposed methodology.
The results showed that the new heterobifunctional ligand can efficiently stabilize
nanoparticles and allows for subsequent modification with antibodies. We
developed a nanoparticle platform for highly efficient magnetic isolation of
bacteria in complex environments. By choosing the correct antibodies, a wide
range of proteins or organisms can be targeted. This makes the methodology
highly valuable for monitoring water quality or for magnetic purifications in
complex media.
Beknopte samenvatting
De ontwikkeling van multifunctionele nanomaterialen is voordelig gebleken voor

vele toepassingen. Nanopartikels in het bijzonder zijn intensief onderzocht, wat
geleid heeft tot grote vooruitgang in de productie van partikels en controle
over hun eigenschappen. Vanuit een biomedisch standpunt zijn ijzeroxide
nanopartikels bijzonder waardevol dankzij hun biocompatibiliteit en lage
toxiciteit. Hun magnetische eigenschappen worden al benut in medische
beeldvorming, hyperthermie voor kankertherapie en geavanceerde afgifte van
geneesmiddelen.
In dit proefschrift presenteren we de ontwikkeling van een multifunctioneel
ijzeroxide nanopartikel dat kan gebruikt worden voor de magnetische scheiding
van bacteriën uit een waterige oplossing, als een preconcentratie techniek.
De aanmaak en karakterisatie van de nanopartikels is beschreven in het eerste
deel van dit werk. We ontwikkelden een verbeterde oppervlaktefunctionalisatie
methode die het mogelijk maakte om de partikels efficiënt te coaten met
verschillende liganden. Dankzij het gebruik van een covalente oppervlaktemodi-
ficatie werd de robuustheid en stabiliteit op lange termijn van het materiaal
verbeterd. De invloed van de verschillende liganden op de colloïdale stabiliteit
werd onderzocht, wat een solide basis vormde voor verdere ontwikkelingen.
Om de nanopartikels voldoende colloïdale stabiliteit te bieden, werd een
nieuwe ligand ontworpen en aangemaakt. Door het modificeren van een
poly(ethyleenglycol)-keten met twee verschillende functionele groepen werd
een nieuwe heterobifunctionele ligand gevormd. Dankzij de aanwezigheid
van de wateroplosbare keten werd de stabiliteit van de gefunctionaliseerde
nanopartikels in complexe omgevingen sterk verbeterd. Daarenboven koppelden
we antilichamen via een covalente binding aan de partikels en werd hun activiteit
gecontroleerd, wat cruciaal is voor hun verdere toepasbaarheid. Het protocol
voor de aanmaak van deze biogeconjugeerde nanopartikels wordt beschreven in
het tweede deel van dit werk.
vii
viii BEKNOPTE SAMENVATTING
Legionella pneumophila bacteriën werden als modelorganismen gekozen voor de

magnetische isolatie experimenten. Deze bacteriën zijn gekende veroorzakers
van verschillende ziektes, inclusief gevaarlijke vormen van longonstekingen.
Bijgevolg is hun detectie wettelijk verplicht in alle publieke watersystemen.
We onderzochten met behulp van flowcytometrie of de biogeconjugeerde
nanopartikels al dan niet in staat zijn om deze bacteriën aan te trekken uit een
waterige oplossing. De resultaten gaven aan dat de ontwikkelde nanopartikels
zeer efficiënt deze organismen kunnen isoleren. Daarenboven werd Legionella
pneumophila ook geïsoleerd uit een mengsel van bacteriën, wat het potentieel
van de ontwikkelde methodologie onderlijnt.
De resultaten toonden aan dat de nieuwe heterobifunctionele ligand na-
nopartikels efficiënt kan stabiliseren en dat een verdere modificatie met
antilichamen mogelijk is. We ontwikkelden een nanopartikelplatform voor
een zeer efficiënte scheiding van bacteriën uit een complexe omgeving. Door het
zorgvuldig selecteren van antilichamen, kan een breed spectrum aan proteïnes
of micro-organismen als doelwit gekozen worden. Dit maak de methodologie
zeer waardevol voor de controle van de waterkwaliteit of voor magnetische
opzuiveringen uit complexe media.
Abbreviations
Ab antibodies
Ag antigen
CDI N,N’-carbonyl diimidazole
CFU colony forming units
CMC 1-cyclohexyl-3-(2-morpholinoethyl)
carbodiimide
CTAB cetyltrimethyl ammonium bromide
DCC dicyclohexyl carbodiimide
DIC diisopropyl carbodiimide
DMAP 4-dimethylaminopyridine
DMPAP 2,2-dimethoxy-2-phenyl-acetophenone
EDC 1-ethyl-3-(3-dimethylaminopropyl)-
carbodiimide hydrochloride
ELISA enzyme-linked immunosorbent assay
FCM flow cytometry
FTIR Fourier transform infrared
GFP green fluorescent protein
Ig immunoglobulins
IgG immunoglobulin G
L. pneumophila Legionella pneumophila
MES 2-(N-morpholino)-ethanesulfonic acid
MRI magnetic resonance imaging
NHS N-hydroxy succinimide
nm nanometer
NP nanoparticles
PAI-1 plasminogen activator inhibitor-1
PEG polyethylene glycol
PFA paraformaldehyde
PVA polyvinyl alcohol
qPCR quantitative polymerase chain reaction
ix
x ABBREVIATIONS
RFP red fluorescent protein

SN supernatant
SPR surface plasmon resonance
TAFI Thrombin Activatable Fibrinolysis inhibitor
TEM transmission electron microscopy
TEOS tetraethyl orthosilicate
TFA trifluoroacetic acid
TRIS 2-amino-2-hydroxymethyl-propane-1,3-diol
VSM vibrating sample magnetometry
XRD X-ray powder diffraction
Contents
Abstract v
Abbreviations x
Contents xi
1 General introduction to bioconjugated nanoparticles 1

1.1 Magnetic nanoparticles . . . . . . . . . . . . . . . . . . . . . . . 2
1.1.1 General introduction to magnetism . . . . . . . . . . . . 3
1.1.2 Superparamagnetism . . . . . . . . . . . . . . . . . . . . 4
1.2 Synthesis of iron oxide nanoparticles . . . . . . . . . . . . . . . 5
1.3 Surface functionalization . . . . . . . . . . . . . . . . . . . . . . 9
1.4 Colloidal stability . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.5 Protein coronas . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.6 Antibodies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.7 Coupling chemistry . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.8 Legionella pneumophila . . . . . . . . . . . . . . . . . . . . . . 24
1.9 Objectives and outline . . . . . . . . . . . . . . . . . . . . . . . 27
2 Functionalization of oleic acid-coated iron oxide nanoparticles 31
xi
xii CONTENTS
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.2 Materials and methods . . . . . . . . . . . . . . . . . . . . . . . 33
2.3 Results & Discussion . . . . . . . . . . . . . . . . . . . . . . . . 35
Characterization of the core nanoparticle . . . . . . . . . . . . 35
Colloidal stability . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.5 Supporting information . . . . . . . . . . . . . . . . . . . . . . 44
3 Heterobifunctional PEG ligands for bioconjugation reactions 49

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
Ligand Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
Bioconjugation . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Activity assessment . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
4 Two-step directional surface modification via protected siloxanes 63

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
Nanoparticle multilayers . . . . . . . . . . . . . . . . . . . . . . 72
4.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
CONTENTS xiii
5 Design of nanoparticles for efficient magnetic isolation of aquatic

pathogens 85
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
Synthesis and characterization of the bioconjugates . . . . . . . 91
Magnetic isolation results . . . . . . . . . . . . . . . . . . . . . 92
5.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
6 Conclusions and outlook 101

6.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
6.2 Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
A Selected experimental techniques 107

A.1 Enzyme-linked immunosorbent assay (ELISA) . . . . . . . . . . 107
A.2 Flow cytometry (FCM) . . . . . . . . . . . . . . . . . . . . . . 109
A.3 Quantitative polymerase chain reaction (qPCR) . . . . . . . . . 110
Health, safety and environment 134
List of publications 137

Chapter 1
General introduction to
bioconjugated nanoparticles
The continuous quest in materials development has gained momentum in the past
decades thanks to the introduction of nanomaterials. This new field of research
lies on the boundary between the scale of atomic or quantum phenomena and
bulk materials. As a result of significant investments during the past decade and
a broad commercial applicability, the field of nanotechnology has matured to
the point of showing significant potential to improve efficiencies and accelerate
progress in biomedical, manufacturing, space and energy applications.1
To be categorised as a nanomaterial, at least one dimension of the material
should be in the 1-100 nanometer (nm) range, while its shape can be planar,
rod-like or spherical. At this scale, the properties of the material often differ
from its bulk characteristics. Multiple properties are known to change with
the size of the object, such as the melting point, conductivity, magnetism and
optical properties.2,3 Figure 1.1 compares the size of nanomaterials (yellow
region) with common other objects.
The first uses of nanotechnology were surprisingly not in recent ages, but date
back to the 4th century and the Roman Empire. At that point, nanoparticles
(NP) were (unknowingly) added to materials to alter their appearance under
different lighting conditions.4 During the past decades a whole new range of
nanomaterials have been developed, going from carbon nanotubes, nanorods
and nanoparticles with different morphologies, fullerenes and graphene to
complex DNA-scaffolds.5–8 Moreover, multiple techniques have been invented
that enable researchers to modify a material on the nanoscale, such as atomic
1
2 GENERAL INTRODUCTION TO BIOCONJUGATED NANOPARTICLES
Figure 1.1: One dimension of a nanomaterial should be in the 1-100 nm range. As a

comparison, other typical sizes of everyday objects and organisms are shown.
layer deposition, contact printing, photolithography or etching.9–11

In this dissertation, the focus lies on the development of a multifunctional
nanoparticle, based on a superparamagnetic iron oxide (magnetite) core, with
custom-designed ligands and proteins for specific magnetic separation of targeted
bacteria or proteins. This type of nanomaterial is considered to be a prime
candidate for future applications in biomedical imaging or drug delivery. A
model nanoparticle-based system for the magnetic separation and quantification
of aqueous pathogens was designed and tested thoroughly for its efficiency and
reproducibility. The system described here will provide a generic platform, that
combines multiple highly valued properties like magnetism, colloidal stability in
complex environments and target specificity.
The following sections will first give a broad overview of different important
properties of superparamagnetic NP as well as possible synthetic routes and
functionalization approaches. Secondly, the biomedical and biological aspects
of the research will be situated to provide sufficient background for the reader.
To finish the introduction, the research goals and scope of this dissertation will
be described.
1.1 Magnetic nanoparticles
A large variety of NP with magnetic properties have been described in scientific

literature, but they all have atoms with a large magnetic moment incorporated in
their crystal structure. Magnetic NP with different compositions and phases have
been produced, such as iron oxides (Fe3 O4 , γ-Fe2 O3 ), spinel-type ferromagnets
(MnFe2 O4 , MgFe2 O4 ), pure metals (Fe, Co) and alloys (FePt).12–15 Even though
MAGNETIC NANOPARTICLES 3
all of these NP have interesting characteristics, including large saturation

magnetization, coercivity or biocompatibility, only iron oxides will be discussed
in this dissertation. Particularly Fe3 O4 (magnetite) is a promising candidate
for biomedical and biological research applications. This type of NP can be
produced at a very large scale with a well-defined size and shape, has excellent
magnetic properties and shows only minor toxicity.15–17
1.1.1 General introduction to magnetism
The phenomenon of magnetism is electronic in nature, since it is related to

the orbital motion of electric charges. The applied magnetic field strength (see
equation 1.1) is represented by H, while the magnetic moment (per volume) is
noted as M, also known as the magnetization. The parameter that links the
applied field and the induced moment, is the volume magnetic susceptibility
(χ), which is a measure for how effectively a field induces a magnetic dipole in
a material per unit volume.18
M = χH (1.1)
The sign and size of the susceptibility (χ) can be used to differentiate between
different forms of magnetism. In a diamagnetic material, χ is negative and small,
as can be seen in Figure 1.2, where the magnetic moments try to oppose the
applied magnetic field. All substances show diamagnetic behaviour, but since
this effect is very small, any other magnetic response will oppose it. Typical
examples are bismuth, water or lead. In a paramagnetic material (see Figure
1.2), χ is positive but small, meaning that the magnetic moments align to the
external field, but the response is low. These moments are caused by unpaired
electrons in the substance. Examples are liquid oxygen, copper sulphate or iron
chloride. For a ferromagnetic substance, that have a permanent net magnetic
moment, χ is positive and large, resulting in a strong attraction towards an
external field. Typical substances are iron, nickel and cobalt.18
Iron oxide nanoparticles can have different morphologies, such as hematite
(α-Fe2 O3 ), maghemite (γ-Fe2 O3 ) or magnetite (Fe3 O4 ). The latter differs from
most other iron oxides since it has different oxidation states of the same atom
in its crystal structure, called a spinel. For magnetite in particular, the full
formula should be Fe2+ 1 Fe2 O4 .
3+ 18,19
Moreover, the Fe3+ atoms are located in
the tetrahedral and octahedral lattice sites, while Fe2+ is only in the octahedral
sites, which is called an inverse spinel. This unusual arrangement causes the
magnetic moments of Fe3+ to cancel each other out (tetrahedral and octahedral
sites have opposite spins), leaving only the moment of Fe2+ contributing to
Figure 1.2: Schematic illustration of magnetic materials and their responses to an

external field. Atomic spins are shown in circles, while magnetic domains are shown
in rectangles. Ferri- is a special case of ferromagnetism, where some spins have an
opposite direction, partially cancelling the net magnetic moment. Superparamagnetic
materials have only one domain and sufficient thermal energy to flip the magnetic
moment.
the overall magnetic moment. As shown in Figure 1.2, this is an example of a

ferrimagnetic material.
1.1.2 Superparamagnetism
When the dimensions of a particle become sufficiently small, the energy needed
to sustain multiple magnetic domains becomes higher than the energy to sustain
just one domain. This results in a single magnetic domain in the particle where
all spins behave as one superposed magnetic moment (see Figure 1.2).
Superparamagnetism is caused by magnetic anisotropy in a material, which can
be seen as a preferred direction of electron spin alignment (crystallographic)
and hence preferred direction of magnetization. If sufficient energy is supplied,
the direction of magnetization can reverse along this axis. This effect is typical
for small nanoparticles, that have only one magnetic domain. The time required
to reverse the spins, the Néel relaxation time (τ ), depends on the anisotropy
constant (K), the particle volume (V), the Boltzmann constant (k) and the
temperature (T) (see Equation 1.2). When temperature increases, the activation
energy between the two reversed directions becomes smaller, and hence the
SYNTHESIS OF IRON OXIDE NANOPARTICLES 5
reversal will occur faster. At lower temperatures, on the contrary, the energy
hurdle becomes too high to overcome. At a certain point this freezes the
magnetic moments in a certain direction, giving the material a ferromagnetic
appearance. This temperature is called the blocking temperature.19–22
KV
τ ∼ exp (1.2)
kT
1.2 Synthesis of iron oxide nanoparticles
Multiple synthetic protocols and approaches have been published for the
synthesis of magnetite (Fe3 O4 ) nanoparticles. These include co-precipitation,
polyol and hydrothermal reactions, micro-emulsions, laser ablation, thermal
decomposition, flow-injection and electrospray syntheses. The properties of the
most common methods are compared in Table 1.1.15,23–26
Table 1.1: Comparison of the most common methods for the synthesis of magnetite
nanoparticles. Thermal decomposition was preferred because of the large scale and
high quality nanoparticles it produces.
Method co-precipitation microemulsion hydrothermal thermal decomposition

Conditions air air high pressure air/inert gas
T (°C) 20-95 20-50 180-220 180-350
Solvent water water in oil water-alcohols long hydrocarbons
Distribution broad medium narrow very narrow
Size 2-100 1-50 10-800 3-50
Shape control not good good very good very good
Scale (gram) large (0.1-2) small (0.1-0.3) small (0.1-0.5) large (1-40)
The following section will discuss the basic theory of nanoparticle growth and the
reaction mechanism of the thermal decomposition method in more detail, since
it is capable of producing high quality monodisperse spherical nanoparticles at
a large scale and is the only method used in this dissertation.
Nucleation and growth
During the very first stages of any crystallization process, nucleation occurs.
Although the theoretical understanding of the phenomenon is well developed, the
experimental data are hard to obtain because these very small nuclei, consisting
of just few atoms, are formed only in the beginning and immediately grow
or disappear again. These nuclei can be introduced externally or be formed
in situ in solution. The first process, where nucleation is induced by foreign
seeds is called heterogeneous nucleation. The latter process, where a single

liquid phase forms the nuclei spontaneously, is called homogeneous nucleation.
This homogeneous nucleation has a profound influence on the obtained size
distribution of the NP. When the formation of nuclei is happening randomly in
time, all particles have a different growth history (∼ age). This will result in
uncontrolled growth and a broad size distribution. However, when all nucleation
occurs at once, the particles will have similar ages and will grow simultaneously.
Such rapid nucleation generally results in a monodisperse set of NP. Even
though a heterogeneous nucleation can achieve similar results, the preparation
of highly uniform small seeds is difficult and hence not very popular.27,28
Figure 1.3: The LaMer model is often used to explain the principle of supersaturation
and nucleation. When the concentration of monomers reaches the critical
supersaturation limit, nucleation will start. During the nucleation stage, the nuclei
for further NP growth are formed. After the supersaturation has lowered, the growth
regime starts (shown at the right), during which the particle size is focused, resulting
in a monodisperse set of NP. The size distribution width shows the transformation of
a polydisperse intermediate product to a monodisperse end product.29
During homogeneous nucleation (the in situ formation of nuclei) a large energy

barrier has to be overcome, which is achieved by the critical supersaturation
level of starting products, also called monomers (see Figure 1.3, LaMer
model).15,27,29,30 An advantage of this process is that when nucleation and
initial growth have started, the supersaturation level quickly lowers, which
terminates the nucleation automatically. This largely reduces the time frame in
SYNTHESIS OF IRON OXIDE NANOPARTICLES 7
which nucleation occurs and is therefore known as burst nucleation. Although

this system approaches the ideal case of a single nucleation event, it is not
a guarantee for a narrow size distribution. During the burst nucleation, the
growth of the particles already starts, due to the supersaturation. At the end
of the nucleation stage, a series of NP with different growth histories is present
in solution (see Figure 1.3). Nevertheless the process of nucleation is crucial
to consume the starting products and eliminate unwanted nucleation at later
stages in the reaction. As long as the duration of the process is much shorter
than the growth process, the NP will have a similar age and a narrow size
distribution can be achieved during the next step of the nanoparticle formation.
The reaction on the surface of the growing NP, turning dissolved monomers into
solid mass, is reversible. A dissolved monomer precipitates onto the surface of
a particle, thereby turning into a solid. This process, called diffusion controlled
growth, occurs when the concentration of dissolved monomers is much larger
than the solubility of the particle itself. When this concentration lowers, the
mass transport direction can reverse, dissolving the particle. At the point where
no monomers are present in solution anymore, small particles will be consumed
by large particles. This phenomenon is know as Oswald ripening and is very
important in nanoparticle synthesis. It is caused by the difference in chemical
potential between small and large NP, with smaller NP having a higher chemical
potential (being more unstable). When the monomer concentration becomes
low, the reversed mass transport reaction dissolves small nanoparticle thereby
releasing monomers, while the more stable large nanoparticles keep growing by
these released monomers.27
Thermal decomposition
A large variety of NP can be produced by the thermal decomposition of

organometallic precursors in high boiling organic solvents, with the addition
of surfactants. Thanks to the successful synthesis of semiconductor NP
by this method, similar methods for the production of magnetic NP were
developed. Possible organometallic precursors are based on metal complexes of
acetylacetonate (acac), N-nitrosophenylhydroxylamine (cup), carbonyls, oleic
acid, other fatty acids or oleylamine. Besides the choice of the precursor, the
solvent, extra surfactants, the reaction temperature, the reaction time and the
ageing period have a profound influence on the size and shape of the obtained
NP.15,28,30,31 Jana et al. reported as one of the first that the synthesis of metal
oxide NP was possible by thermal decomposition of metal fatty acid complexes.
They proved that the concepts, derived from semiconductor NP syntheses, were
valid for metal oxides.31,32 Six months later later, Park et al. published a
protocol for a very large scale synthesis of magnetite nanoparticles.12 They
Figure 1.4: The decomposition of iron-oleate at high temperature happens via a

radical process. Carbondioxide gas is released from the oleate molecules in one of
these reaction steps. (Figure adapted from Perez et al.)34
used iron-oleate as the organometallic precursor, which was decomposed in

1-octadecene (as solvent) and oleic acid (as surfactant) at 320°C. Later on, the
decomposition of iron pentacarbonyl was described, which is known as one
of the best syntheses for monodisperse iron NP.33 A second step is necessary
however to oxidize the NP further to magnetite. Nevertheless the latter method
is used less often due to the toxicity of the organometallic precursor. In this
dissertation, the thermal decomposition of iron-oleate, as described by Park et
al. was used to produce the magnetic NP.12
The so-called ultra-large synthesis of iron oxide NP is based on the decomposition
of Fe3+ (oleate)3 . The exact reaction mechanism is not know yet.35 Since the
magnetite particles have Fe3+ and Fe2+ in their crystal lattice, a reduction
reaction has to occur. Perez et al. have published the most complete description
so far.34 They found out that the heating of the oleate precursor results in
the formation of radicals (see Figure 1.4). These reactions provide the oxygen
for the oxide formation as well as the reduction of Fe3+ to Fe2+ . The radical
byproducts can combine to form ketones, that were detected at different stages
in the reaction.34 Several different opinions about the influence of temperature
can be found in literature. Perez et al. stated that the first decomposition step
of the oleate precursor already starts at 200°C.34 Kwon et al. on the other hand,
reported that this reaction occurs at 275-300°C.28 Both reports agree on the
radical mechanism and the production of CO2 during the reaction.
SURFACE FUNCTIONALIZATION 9
After their synthesis, the NP have a coating of oleic acid molecules, which are
very apolar (oleic acid is a component of olive oil). The particles only disperse
in apolar solvents (hexane, heptane, chloroform, dichloromethane, ...) at this
point. For further use in aqueous applications, the surface has to be modified
with a polar moiety. The following section will give an overview of different
functionalization strategies and their main properties.
1.3 Surface functionalization
Figure 1.5: Overview of different ways to functionalize iron oxide NP: organic ligands
(to form mono- of bilayers), polymers or inorganic shells.
Coating the surface of a nanoparticle is generally performed to improve the

chemical or physical stability. Some NP, such as elemental Fe, Co or metal
alloys, are very sensitive to oxidation by air. A surface coating protects the
crystal from the environment and sustains the properties of the core. The
physical stability of a nanoparticle should be regarded at a larger scale: the
colloidal stability in the medium and prevention of aggregation. In general, the
coating strategies can be divided into two groups: organic or inorganic shells
(see Figure 1.5). The organic shells include monomeric surfactants, polymers,
double layers or proteins. Inorganic shells are usually based on silica (SiO2 ),
carbon, metals (Au, Ag) or oxides.15 The process of changing the initial ligand
with a new molecule or shell is called ligand exchange. As an alternative, NP
can be embedded in a matrix, such as a polymer, silica or composites. This
prevents aggregation and oxidation, but the particles are fixed spatially and
this limits their applicability.15
Polymer coatings
The most common coatings described in literature are dextran, carboxymethy-

lated dextran, sulfonated styrene, starch, polyvinyl alcohol (PVA) and
polyethylene glycol (PEG).23,36
Dextran is a polymer, consisting of polysaccharides with varying chain length
and branching and is often used because of its biocompatibility. It interacts
via chelation or hydrogen bonding with the surface of iron oxide. Although
these hydrogen bonds are weak, the polymer molecules have numerous hydroxyl
groups and the total bonding energy hence becomes very large. PEG and
PVA are very hydrophilic and water-soluble polymers. These coatings are
biocompatible and provide the NP with excellent solubility and stability in
aqueous environments.
Even though the resulting colloidal properties of these polymer-coated NP are
often very good, the lack of control over the shell thickness and density of
the layer makes this approach less favourable.37 Most methods are based on
a non-covalent interaction between the polymer and the surface, which limits
the robustness and applicability. Physisorbed drug molecules can be lost before
they reach the target site. Another issue arises when the coating of the particle
is removed, since this results in aggregation and consequently adverse effects
on cells or tissue.38 These problems have been addressed by cross-linking the
polymer layer, which increases the robustness but complicates the synthesis
of the nanoparticles. Moreover, controlling the presence and availability of
functional groups on the polymer layer is difficult and generally requires more
advanced methods such as grafting-from or grafting-to.37,38
Inorganic shells
The introduction of inorganic shells on iron oxide NP can be performed from

two different points of view: protect the magnetic core from the surrounding
environment or introduce a new property (such as plasmonic or catalytic
SURFACE FUNCTIONALIZATION 11
materials). The first is often achieved by the growth of a silica shell on the
surface, which is inert and forms a dense protective layer. The latter, introducing
a new property can be accomplished by the growth of a golden or silver shell.
These materials have interesting properties such as plasmons, catalytic or
antimicrobial activity.39–41 An important advantage of inorganic shell formation
is the high degree of control that, compared to polymers, can be achieved over
the shell thickness and shape.
Silica coatings are well-established for magnetic NP, since it prevents aggregation,
improves the chemical stability and lowers the toxicity by ion leaching. Moreover
the shell is negatively charged, stabilizing the NP by coulombic repulsion. Three
methods are developed to growth the shell: the Stöber process, silicic acid and
emulsions methods.42–44 The Stöber process is very well-known and relies on
the hydrolysis and polycondensation of a silica precursor such as tetraethyl
orthosilicate (TEOS).42 Emulsions (micelles or inverse micelles) are of particular
interest for modification of apolar iron oxide NP, since the shell is usually
grown directly on top of the present oleic acid surfactant. In this case, extra
surfactants are added to make the NP disperse better, such as Igepal CO-520
or cetyltrimethyl ammonium bromide (CTAB).45 However, this complicates the
workup of the resulting core-shell NP. More complex silica can be produced using
a large excess of surfactants, resulting in mesoporous silica. These structures
have a very large surface thanks to the pores and can be used for drug delivery.45
Coating iron oxide NP with a gold layer is not trivial since the two crystal
structures do not match well. Recently, progress has been made by performing
the reaction at higher temperatures, starting from an organometallic precursor
(gold acetate) in an apolar medium.46 Other approaches are often done
in aqueous environments by chloroauric acid or seeded growth.41,47 Post-
functionalization by thiol-containing molecules is straightforward, thanks to the
strong bond between thiols and gold. Moreover, the surface is chemically inert
and has interesting optical properties.48
Monomers and bilayers
Small organic molecules can also be used to form stabilizing layers on iron oxide
NP. They can be chemically anchored or physically adsorbed to the surface, in a
single or double layer. Well-known functional groups that interact strongly with
the surface by forming a coordination bond are carboxylic acids, phosphates
or catechols. Another class of molecules, siloxanes, are capable of forming a
covalent bond with the nanoparticle’s surface.23,49
Functionalization by a coordination bonding molecule is one of the most
straightforward methods. Citric acid (citrate) is a widely used ligand in this
regard. Two of its carboxylates interact with the surface, leaving the third
exposed to the solvent, providing a negative charge to the particle.50 However,
this type of bond is labile and is prone to break by increasing temperature,
extreme pH or the presence of other ligands. Similarly, phosphonic acid
(phosphonate) forms a coordination bond, albeit stronger than by a carboxylic
acid.51 This makes it possible to replace for instance oleic acid on the surface in
a single step without complex two phase systems.52 Those ligands show good
chemically stability at neutral pH during several weeks.
Catechols-containing ligands are a recent development in surface chemistry.53–55
These 1,2-dihydroxybenzene derivatives showed an exceptionally high com-
plexation constant with metal ions (log K ≈ 30-40), making the bond almost
irreversible.56,57 The concept of using catechols as highly versatile ligands
is derived form nature itself, since the molecule is present in the adhesive
proteins secreted by mussels for attachment to wet surfaces.55 The large
variety of commercially available catechol-containing molecules (dopamine, 3,4-
dihydroxyhydrocinnamic acid, . . . ) has caused a large increase of publications
about these ligands in recent years. A minor disadvantage of dopamine is its
tendency to polymerize into polydopamine. This issue was addressed by the
introduction of a nitro group onto the phenyl ring.58
Alternatively, the already existing oleic acid layer can be used to form a double
layer. To achieve this, amphiphilic molecules (phospholipids) are added to the
freshly prepared apolar NP, which will encapsulate the particle and form a
double layer. This is stabilized by the hydrophobic interactions between the
oleic acid and phospholipids.59
Siloxanes
Some organic molecules, called silanes, have silicon atoms incorporated in their
structure. A subgroup of these are the organosilanes, which have one or more
organic groups. Siloxanes are member of this organosilane group, but they have
at least one alkoxy bond present in their molecular structure. These molecules
are of particular interest for the functionalization of iron oxide NP, since
they can form a covalent bond with the surface, while introducing functional
groups. They are frequently used for the modification of oxide surfaces (glass,
metal oxides, . . . ) and hence a large variety of these siloxanes is commercially
available.60–64
In the presence of water, one of the alkoxy groups (usually methoxy or ethoxy)
splits off by an acid- or base-catalysed reaction (see Figure 1.6). This so-called
silanol molecule can polymerize with another silanol molecule via a condensation
COLLOIDAL STABILITY 13
Figure 1.6: The alkoxy siloxane molecule is first hydrolyzed to a silanol molecule by
an acid- or base-catalyzed reaction. This molecule can polymerize or react directly
with the hydroxyls on the surface.
reaction or directly interact with the hydroxyl groups on the surface of the
nanoparticle via a similar condensation reaction.65
1.4 Colloidal stability
In a colloidal system, with particles of similar sizes, several interaction forces have
an influence on the overall stability of these particles. These forces, responsible
for the behavior of NP in solution, usually are the electrostatic repulsion and
van der Waals attraction. Moreover, hydration and steric repulsion can also
have a significant impact if less charges are present. In the following section,
the concepts that are applicable to the NP presented in this dissertation, will
be discussed further in detail.
DLVO theory
The most common way of stabilizing NP is by making use of electrostatic

repulsion. This repulsion occurs when particles have a similar charge and is
very effective in preventing particle aggregation. In Figure 1.7, this electrostatic
repulsion is shown by the red curve. It quickly grows at smaller distance due to
the close proximity of the charges. Besides this repulsion, an attractive force is
also present in colloidal systems, the van der Waals forces. These dipole related
forces are smaller and are more pronounced at smaller distances between the
dipoles. It is the sum of the Keesom, Debye and London forces, respectively
the attraction between two permanent dipoles, a permanent and induced dipole
Figure 1.7: The DLVO theory (for a system with a high salt concentration) describes
the sum of the van der Waals attraction and electrostatic repulsion. The primary
minimum represent an irreversible aggregate, while the secondary minimum shows
a flocculated state, which is reversible. An energy barrier has to be overcome to
reach this irreversible aggregated state, which is related to the collision speed of the
particles.
and the spontaneous oscillations of electrons that create temporary dipoles.

Usually in colloidal systems, the London forces are most important. In Figure
1.7, the van der Waals force is shown by the green curve. The sum of both
(see Figure 1.7, black curve) essentially explains the behavior of a colloidal
system and can be used to determine whether it will be stable in time. This
concept was developed by Derjaguin, Landau, Verwey and Overbeek and was
hence called the DLVO theory.66,67 It provides a framework that considers
the total interaction energy of two NP to determine the stability of colloids.
When two particles with similar charges approach each other, a small attractive
force is present due to the van der Waals forces. This results in a minimum
(secondary) where the particles are loosely packed together, called flocculation.
So far the interactions are still reversible. If the particles would come even
closer together, the repulsive force, due to their charges, creates a energy barrier
that the particles have to cross. If the collision energy (from Brownian motion)
is large enough, the particles can cross the barrier and end up in the primary
minimum. This is the point of irreversible aggregation and minimal energy, and
often results in precipitation. If the particles are highly charged, the barrier
becomes very high and this would ensure colloidal stability since the particles
PROTEIN CORONAS 15
cannot end up in the primary minimum. This DLVO theory is dependent on

many different parameters that determine the initial attraction of repulsion
of two particles. For instance, at high charge densities, the overall force is
repulsive, since the interaction is dominated by the double layer contribution.
At small charge densities, the interaction is dominated by the attractive van
der Waals force. In this work, the relation between a functional group and its
zeta potential is discussed in Chapter 3, while in other chapters stability will
be achieved by steric repulsion rather than electrostatic repulsion.
Zeta Potential
Most oxide materials are charged by nature, since the presence of hydroxyls on
their surface makes them partially pH responsive. This charge attracts ions with
an opposite charge, that stick to the surface of the particle, forming the Stern
layer. A second more diffuse layer is also formed, that has more mobile charges.
The potential, caused by the surface charges, at the edge of the Stern layer is
called the Stern potential, while the potential at the edge of the diffuse layer
(slipping plane) is called the zeta potential. The latter is an important concept
in colloid physics, since it gives a relatively good indication of the colloidal
stability of a charged particle. It can be measured in an adapted dynamic
light scattering experiment where the electrophoretic mobility of the particles
is translated into a zeta potential value. If an absolute value higher than 25-30
milliVolts (mV) is measured, the particles are said to be colloidally stable in
time.68 If the absolute value is lower than 25 mV, the system is probably not
stable and aggregation will occur in time. However, this theory does not take
any other repulsive forces, such as steric hindrance, into account. Therefore the
value of the zeta potential is only important for highly charged particles, without
presence of steric stabilizers. In this work, steric repulsion is far more important
than electrostatic repulsion, hence the measurement of the zeta potential was
often omitted due to its limited contribution. For example, a PEG coated
nanoparticle will have a very low potential due to the absence of charged groups,
insinuating colloidal instability, while these particles show exceptional stability
in time. Nevertheless, this concept is generally accepted as one of the only ways
to measure the stability of a nanoparticle dispersion.
1.5 Protein coronas
When NP enter a biological fluid (for instance blood), their surface is instantly
covered by a layer of proteins, which is called a protein corona (see Figure 1.9).
This newly formed corona alters the interfacial properties, size and aggregation
Figure 1.8: A negatively charged nanoparticle attracts a layer of positive ions, which
are almost strongly bound, to its surface, called the Stern layer. The next layer is
more diffuse and partially compensates the charge of the inner layers. The edge of
the layer is called the slipping plane, which is the point where the zetapotential is
measured.
PROTEIN CORONAS 17
Figure 1.9: Two different protein layers are spontaneously formed around the
nanoparticle: a hard corona on the surface and a soft corona as a second layer,
which is more dynamic.
behavior of the nanoparticle. Moreover the NP gain a biological entity that

is different form the initial synthetic identity. Their biological identity will
determine whether NP can cross certain barriers or membranes and how the
interaction with other biomolecules will occur.69–73
Two different types of coronae are formed around a nanoparticle: a hard and a
soft one. The hard corona is defined as the first layer of proteins that are strongly
interacting with the surface. The soft corona is the second layer of proteins,
that have no direct contact with the surface, but have solely protein-protein
interactions. Both layers show different kinetic behavior, with the soft corona
having a faster exchange of proteins, while this is rather slow for the hard
corona. The initial protein layer formation is thermodynamically driven, since
it lowers the surface free energy of the nanoparticle. The protein composition
of this layer depends fully on the concentration and nature of proteins in the
surrounding liquid. For instance in blood, the high concentration of serum
albumin will initially cause a high presence in the hard corona (kinetically
driven). But in time, less abundant proteins with higher affinity and slower
kinetics might displace them (thermodynamically driven). Nevertheless, the
protein corona is what a cell sees of a nanoparticle since it cannot directly
interact with the nanoparticle’s core and is hence a very important concept in
nanoparticle science.71,74
1.6 Antibodies
The human body is defended by an adaptive immune system, that has evolved
to recognize a great variety of antigens from bacteria, viruses or other pathogens.
The antigen-recognizing molecules (on B-cells) are called the immunoglobulins
(Ig) (or antibodies (Ab)) and are produced in a vast range of antigen specificities,
even though each B-cell produces Ig of just one single specificity. They come
in two forms, membrane-bound or secreted (which are structurally different),
depending on the type of B-cell. These Ab serve two functions: binding to the
molecules (from the pathogen) that initiated the immune response and to rally
other cells and molecules to the pathogen to destroy it. For example, an antibody
can bind to a virus, making it ineffective and marking it for destruction.75
Structure
Ab are roughly Y-shaped proteins that have three portions of equal size,
connected by a flexible linkage, called the hinge. The twofold purpose of
an antibody is visible in its structure (see Figure 1.10). One part specifically
binds the target antigen while another part influences recognition and removal
mechanisms. The first part varies greatly between different Ab, and is hence
called the variable region (VL and VH ), which is a substructure of the antigen
binding region (Fragment antigen binding, Fab). The Fab fragment is build up
from a complete light chain and part of the heavy chain. The large variability
makes that Ab can bind to a wide range of pathogens and virtually any structure
can be recognized. The second part, that can activate removal mechanisms, has
less variation and is hence called the constant region. It was originally observed
to crystallize easily and was called the fragment crystallizable (Fc). Five main
classes of these constant fragments exist, which divides the Ab into classes,
namely IgG, IgM, IgD, IgA and IgE. The immunoglobulin G (IgG) class is
the most abundant in the human body.75 A single pathogen can have many
possible binding sites for Ab, called epitopes. As these Ab are produced by
many different cells (1 cell = 1 type of antibody), they are called polyclonal.
They are all targeting the same pathogen, but have a different variable region
and bind to different epitopes. While this is advantageous for fighting infections
ANTIBODIES 19
Figure 1.10: An antibody is roughly Y-shaped and has different regions with separate
functions. The bottom Fc (Fragment crystallizable) part has no antigen binding
properties, but is responsible for the activation of removal mechanisms. The top Fab
(Fragment antigen binding) part is responsible for the interaction with the antigen
and is build up from the complete light chain and part of the heavy chain. In between
the two part, a flexible region is present, called the hinge.
in nature, the heterogeneity sometimes hinders their applicability in research.

The following section will give more insight in the production of antibodies.75
Production
Two main types of Ab can be obtained: mono- and polyclonal Ab. The latter
were discussed in the previous paragraph, being a set of Ig produced by different
B-cells, all targeting the same antigen but on different epitopes. Monoclonal
Ab on the other hand represent a set of Ig that all target the same epitope and
are produced by the same cell. This major breakthrough in the research and
development of Ig was accomplished by Köhler and Milstein in 1975.76 They
fused the B-cells (that produce the Ab, but do not divide indefinitely) with
immortal myeloma tumor cells. This resulting single hybrid cell is known as a
hybridoma cell, which produces Ab and can divide indefinitely and grow well
in cultures. These cells however are clones of each other and hence all secrete
the same antibody, which is monoclonal.75,77,78 Alternatively, the hybridoma
cells can be injected into the abdomen of a mouse, where they will multiply and
produce Ab, which is know as the mouse ascites method.
Table 1.2: Comparison of monoclonal and polyclonal antibodies, including their

advantages and drawbacks.75,78
Monoclonal Polyclonal
Production in vitro (hydridoma cells) mostly in vivo
Target one epitope multiple epitopes
Cost expensive relatively cheap
Technology highly complex relatively easy
Production time long (> 6 months) short (2-3 months)
Batch to batch identical large variability
Scale very large limited
Cross-reactivity very low possible, depending on
level of purification
Polyclonal antibodies on the other hand are acquired from the blood of an
immunized animal. In short, a rabbit or goat is injected with a pathogen (or
antigen (Ag)), after which the immune system is given time to produce an
immune response in the form of Ab. After a second (or even third) boost with
extra antigens, the blood from the animal is collected and purified.
Purification of Ab is usually performed via affinity column chromatography with
proteins that strongly bind Ab, for instance protein A (Staphylococcus aureus),
G (Streptococcus spp.) or other Ab. Depending on the starting material, derived
from an in vitro culture of hybridoma cells or serum of an immunized animal,
this process results in monoclonal Ab or the whole IgG fraction present in the
serum. This fraction can later be purified further to collect all the Ab targeting
a certain antigen. Again affinity column chromatography can be used to perform
this purification step.77,79–82 Instead, polyclonal Ab can also be prepared by
mixing several monoclonal Ab together.
The main differences between the two types of Ab, mono- and polyclonal, are
summed up in Table 1.2.
1.7 Coupling chemistry
The covalent attachment of proteins onto nanoparticles is a widely investigated

research domain. Hundreds of methods have been described in scientific
COUPLING CHEMISTRY 21
literature, including zero-length, homobifunctional, heterobifunctional and even

trifunctional cross-linkers. The following section will give a brief overview of
the zero-length coupling options (no extra atoms after coupling) that require no
modification of the protein itself, since this approach was used in this dissertation.
A complete overview can be found in the book Bioconjugate Techniques written
by G. T. Hermanson.83
The smallest possible coupling between two entities is via zero-length cross-
linkers. These molecules are capable of connecting two molecules, without
leaving behind extra atoms, intervening linkers or spacers. Definitely in the
case of antibody coupling, the presence of linker molecules can induce possibly
unwanted cross-reactivity. Moreover, the modification of Ab prior to coupling,
such as cleavage of the disulfide bonds or oxidation of their carbohydrates, alters
their structure and can have a negative impact on the affinity and stability of
the protein.84,85 Hence, the most common way of cross-linking Ab to a substrate
is by means of their already present functional groups: carboxylic acids or
amines.
Carbodiimides
Carbodiimides are by far the most well-known mediators for the formation of
amide linkages between amines and carboxylic acids. Similarly they can form
phosphoramide bonds between phosphates and amines. The carbodiimides can
be divided into two classes: water soluble or insoluble. The first are of course
the most common choice for coupling reactions involving proteins, since these
are only stable in aqueous environments. The water insoluble carbodiimides
on the other hand are frequently selected for peptide synthesis or coupling of
molecules in organic solvents.
Essentially all carbodiimides react via the same reaction pathway, as shown
in Figure 1.11. The diimide reacts with a carboxylic acid to form a highly
reactive O-acylisourea intermediate. This molecule will then react with a
primary amine to form an amide bond. Other nucleophiles, such as thiols
or hydroxyls, can also attack the O-acylisourea molecule. This implies that
hydrolysis by water is a major competing reaction, forming an isourea, setting
the carboxylic acid free again, resulting in a net loss of diimide starting product
(see Figure 1.11).86 In aqueous environments, EDC is the most commonly used
reagent. Another possibility is 1-cyclohexyl-3-(2-morpholinoethyl) carbodiimide
(CMC). For organic solvents, dicyclohexyl carbodiimide (DCC) and diisopropyl
carbodiimide (DIC) are often selected.83
The reaction between the O-acylisourea intermediate and an amine is
unfortunately slow and hydrolysis in aqueous solutions is prone. If the amine
Figure 1.11: A typical reaction between a carbodiimide (1-ethyl-3-(3-

dimethylaminopropyl)-carbodiimide hydrochloride (EDC)) and a carboxylic acid,
starts with the formation of a O-isoacylurea molecule. This can reaction directly
with an amine to form the amide product. Usually N-hydroxy succinimide (NHS) is
added to form a NHS-ester, which is more stable. This ester will then react with the
amine to form the amide product. An important side reaction is the hydrolysis of the
intermediate, resulting in a net loss of EDC.
does not find the activated carboxylate before hydrolysis occurs, the coupling
will fail. This situation is especially occurring in coupling reactions with proteins,
since those are dissolved in water at relatively low concentrations. To solve
this issue, NHS can be added to the reaction. This way a reactive NHS-ester
can be formed (see Figure 1.11), that can react rapidly with amines. Moreover,
this ester hydrolyses very slowly, allowing to purify the product at this stage.
Overall, the addition of NHS to a coupling reaction increases the efficiency more
than 20-fold, as investigated by Staros et al.87 This improved efficiency has
the unexpected consequence that the coupling might be too efficient and hence
causes excessive cross-linking (and precipitation). Scaling the amount of EDC
and NHS will solve this problem.
In this dissertation, carboxylic acid groups on the surface of the NP were coupled
to amines on the Ab. The opposite reaction (with amines on the NP) could
also have been performed. However this has a major disadvantage: proteins
have both carboxylic acids and amines; activation of the carboxylic acids would
COUPLING CHEMISTRY 23
Figure 1.12: CDI reacts with a carboxylic acid group and form a N-acylimidazole with
the liberation of CO2 as the driving force. Next, this reacts with an amine to form
the final amide bond.
lead to severe protein-protein cross-linking. The NP were designed to have only

carboxylic acids and no amines, which can safely be activated by EDC and
NHS. An extra advantage of the usage of EDC is the introduction of charges,
which might help to stabilize the functionalized NP.
N,N’-carbonyl diimidazole
N,N’-carbonyl diimidazole (CDI) is very efficient activator of carboxylic acids.

It is capable of forming a zero-length bond between the acid and an amine, or a
one-carbon-lenght bond between two amines (as a ureum), or a hydroxyl and an
amine (as a N-alkyl carbamate). A carboxylic acid reacts with CDI and forms
a N-acylimidazole compound that is highly reactive. This formation is very
efficient and produces high yield thanks to the liberation of carbon dioxide as
the driving force. The activated carboxylate can react with amines or hydroxyls
to form amides or esters, respectively. Figure 1.12 shows the reaction of CDI to
form an amide.83,88
The major disadvantage of CDI is that it is extremely sensitive to the presence
of water. Hydrolysis will break down CDI instantly to CO2 and imidazole.
Therefore all reactions should be performed in dry solvents. An advantage is
the very rapid activation of carboxylic acids, which also releases characteristic
gas bubbles, which indicate a successful reaction. All excesses of CDI should
be removed from the reaction mixture before the addition of the amine, since
side reactions can occur at that point. Adding a small amount of water can be
used for this. Hydrolysis of the activated carboxylate by water is a possible
side reaction, but occurs much slower than the nucleophilic attack by a primary
amine.88
Figure 1.13: Fluorescence image of L. pneumophila bacteria (recombinant strain

expressing the green fluorescent protein (GFP)), which are around 2 µm long and 0.5
µm wide. The rodlike shape is clearly visible.
1.8 Legionella pneumophila
The aim of this dissertation was to develop an NP-Ab system that targets
Legionella pneumophila (L. pneumophila). These bacteria have a large social
relevance and their detection in various aqueous environments (f.i. swimming
pools, waste and cooling water) is mandatory by law. The following section will
give more insight in the properties of these bacteria as well as their prevalence
and detection.
Ecology of L. pneumophila
The Legionella bacteria is a Gram-negative, aerobic, rodlike organism that is

approximately 0.5 µm wide and 2 µm long (see Figure 1.13). It is ubiquitously
present in fresh water and is the largest genus of bacteria that survives by
parasitizing on free living single-celled protozoa.89 They are remarkably acid-
tolerant (able to withstand pH 2) and have been isolated from environmental
LEGIONELLA PNEUMOPHILA 25
sources with pH of 2 to 8.90,91 Moreover they have been found in sources as

diverse as water on plants in rainforests, groundwater and even seawater.92
Legionella bacteria thrive at elevated temperature (30-40°C), in water with high
(in)organic content and possible host protozoa.93 This explains why the bacteria
are so often found in artificial habitats, for instance warm water systems such as
whirlpools, fountains, showers or air-conditioning units.91 If more than 100,000
colony forming units (CFU) are found in 1 liter of water sample (in Belgium),
the facility has to be closed until levels have dropped under 10,000 again.94,95
This mandatory detection of the bacteria makes detection and pre-concentration
techniques highly important.
The exact proliferation mechanism of L. pneumophila in protozoan hosts is
complicated but is based on overcoming three major hurdles. First, the bacteria
are phagocytosed by the host into a phagosome, which usually becomes a
very inhospitable environment (phagolysosome). Secondly, they need nutrition
and sustenance for growth, which has to be acquired through the vesicle that
contains them. Thirdly, the surrounding phagosomes restrict the available space
for continuous growth. The L. pneumophila bacteria overcome these issues by
hijacking the host secretory system and membrane traffic (by dot/icm genes).96
They camouflage their surrounding vesicle to obtain a steady supply of nutrients
from the host and continue to grow.
Legionellosis
An outbreak of a severe type of pneumonia during an American Legion

Convention in Philadelphia led to the discovery and description of the
Legionnaires’ disease, named after the participants of the conference, in 1976.97
More than 42 species of Legionella have been identified since the first outbreak.
A less severe form of the disease is called the Pontiac fever. The Legionnaires’
disease causes flu like symptoms, such as fever, headache, chills and muscle pains.
It can cause large fatality rates of 40-80% among patients with lowered resistance
(f.i. in hospitals). It can be treated by antibiotics such as clarithromycin and
azithromycin.91
Detection methods
For the detection of the L. pneumophila bacteria, 2 important steps should be

considered: pre-concentration and the quantification itself. Since a standardized
water sample for official Legionella enumeration has a volume of one liter, a
pre-concentration step is crucial. No quantification technique is capable of
handling one liter samples directly and taking only small aliquots is not allowed
by law.94 Usually the sample is centrifuged or filtered to collect the harmful
pathogens. However both techniques have serious drawbacks inherently related
to their mechanism. Centrifugation can obtain high recovery values, but is not
target specific and is very inconvenient for large sample volumes making it a
very time consuming approach. Filtration can handle these large volumes, but
is not specific either and has a profound influence on the cell viability and has
issues with complex (dirty) samples.98 Another technique, immunomagnetic
separation, makes use of Ab and magnetic NP to have a target specific magnetic
attraction of bacteria from aqueous environments. It is capable of achieving a
purification from large volumes, has only limited effect on the cell viability and
can be used in complex samples. This method will be developed further for the
L. pneumophila bacteria is this dissertation.
For the quantification of the bacteria, multiple techniques have been developed:
cell cultures, flow cytometry (FCM), enzyme-linked immunosorbent assay
(ELISA), quantitative polymerase chain reaction (qPCR), hybridization (DNA
or RNA), surface plasmon resonance (SPR) sensors or antibody staining.99,100
Cell cultures are still considered as the golden standard for the enumeration
of bacteria. However, the method is very time consuming, slow (7-10 days)
and has issues with the presence of other bacteria if they grow faster than
the target.101 FCM makes use of the specific fluorescent signal and size of
organisms for detection and quantification. Unfortunately the technique is
expensive and impurities might hinder a correct enumeration.101,102 qPCR
is based on the polymerase chain reaction, in combination with fluorescent
tags, for the quantification of target DNA (or RNA is the case of RT-qPCR).
It is a fast method, with a low detection limit, but inhibitors (that are
often present in an environmental water sample, f.i. metal ions, surfactants
or polysaccharides) are a recurring problem. DNA- or RNA-hybridization
approaches utilize complementary strands of DNA or RNA to detect the target
bacteria. It is a technique with a high specificity, but can only handle relatively
low concentrations and free DNA or RNA can result in false positive signals.
SPR sensors measure the shift in the plasmon resonance of gold to determine
the concentration of an analyte. It is very sensitive but can have problems
with aspecific adsorption and false positive signals.103 Antibody staining on
the other hand employs fluorescent antibody conjugates to selectively stain
target organisms. Cross-reactivity is recurring issue with this technique and
pre-concentration of the sample is advisable.100
The detection methods that were used for this dissertation will be explained in
more detail in the appendix.
OBJECTIVES AND OUTLINE 27
1.9 Objectives and outline
Objectives and strategy
The increasing interest in functional nanomaterials has led to their implemen-

tation in various biomedical applications, either as a therapeutic agent or for
diagnostic purposes.104–108 NP with magnetic properties are well-known for their
potential in magnetic resonance imaging (MRI), for which commercialization has
been achieved (Feridex, Endorem). However, applications regarding selective
magnetic purification are more scarce in scientific literature. These studies are
often approached from either the synthetic or biomedical side, lacking an optimal
design on the other side. A thorough nanoparticle synthesis and functionalization
is often combined with a less comprehensive separation study and vice versa.
Moreover the complexity of the syntheses often makes implementation in
applications rather hard. In this dissertation a complete nanoparticle platform
for magnetic purification of targeted organisms will be described that pays
attention to all relevant properties, ranging from the nanoparticle’s synthesis
to their functionalization and testing their performance in several relevant
applications, with an extra focus on reproducibility and low cost.
Three major material properties were identified as being crucial for the
development of a NP-based magnetic separation platform. First, the
synthesized core nanoparticle should be well-defined, in terms of size and shape.
Moreover the reduction of batch-to-batch difference, and hence an improved
reproducibility was an important rationale. Secondly, the available surface
modification procedures of these particles had to be improved to meet two
crucial requirements: a fast and highly reproducible functionalization method
that allowed covalent bonding of the ligand to the surface. Thirdly, a new ligand
had to be designed that combined all properties necessary for the intended
applications. The presence of a functional group and improved colloidal stability,
combined with the possibility to covalently bind to the surface of iron oxide were
considered the primary requirements.
The synthetic method for nanoparticle production was chosen from dozens
of possible approaches. Despite the complex setup and high temperatures,
thermal decomposition of iron-oleate was selected mostly for its scalability, price,
reproducibility and high nanoparticle quality.12 For the surface functionalization
less information was found in literature. Covalent surface modification is hardly
described for oleic acid-coated magnetite NP and the available procedures were
rather unreproducible during the initial tests. Therefore a new method was
developed that was capable of introducing siloxane ligands covalently on the
surface of iron oxide. The ligands required for the successful development of
the nanoparticle platform, needed several properties, such as improved colloidal

stability of the NP and the possibility to perform subsequent coupling chemistry,
combined with the presence of a siloxane group for covalent bonding to the
surface. Since this product was not commercially available, a new synthesis
was developed, keeping in mind that it should be highly reproducible, but also
cost-effective (in terms of starting products) and easy to perform. Thiol-ene
click chemistry was used in this regard and proved to be a valuable approach
towards the synthesis of a highly effective new ligand.
Outline
The second chapter will explain in more detail how the NP were synthesized
and introduces the new functionalization procedure that was developed. Several
characterization techniques were used at this stage to fully characterize the
core material. The surface modification procedure was tested thoroughly by
introducing multiple different siloxane molecules onto the surface of iron oxide
NP. Moreover their colloidal stability in buffer systems was investigated, which
was correlated with their zeta potential.
Because the commercially available siloxane ligands do not have the required
functionalities, a new ligand molecule was designed and synthesized. Chapter
three gives an overview about the synthetic details and how the ligand was
introduced onto the NP. Furthermore, Ab were coupled to the molecule to
underline its potential in biomedical applications. ELISA was introduced at this
point as a new approach to quantify the amount of coupled proteins on the NP.
The activity of these bioconjugated particles was tested by SPR measurements.
Functionalization of metal oxide surfaces by siloxanes is a well-established
approach, even though some side reactions are generally ignored. Since
functional groups on the siloxanes can interact with the surface itself, correct
orientation is hard to control. In the fourth chapter a generalized methodology
is presented that fully solves the issues related to the usage of functional siloxanes.
Protective groups were used to ensure that no unwanted interactions could occur.
To underline the potential of this approach, six different functional groups were
introduced onto iron oxide NP. It was observed that the colloidal stability of
the resulting product is far better compared to the traditional functionalization
methods.
To prove that the developed ligands and in a broader sense the NP are capable
of performing efficient and selective magnetic separations, we designed an
experiment that supports this hypothesis. L. pneumophila bacteria were
magnetically separated from an aqueous solution by making use of particles,
OBJECTIVES AND OUTLINE 29
conjugated with Ab that target the bacteria. These experiments are explained
in chapter five.
In the appendix, three techniques that are not standard in chemical sciences
will be explained to support the reader.
The presented results summarize the development of a generic nanoparticle-
based magnetic separation platform with excellent colloidal stability in complex
environments, while being cost-effective and straightforward to perform.
Chapter 2
Functionalization of oleic
acid-coated iron oxide
nanoparticles
This chapter is based on the following publication with minor modifications:
Bloemen M, Brullot W, Luong TT, Geukens N, Gils A, and Verbiest T

Improved functionalization of oleic acid-coated iron oxide nanoparticles for
biomedical applications
Journal of Nanoparticle Research 2012; 14: 1100–1109
Dr. W. Brullot performed the VSM experiments and analyzed the results. T. T.
Luong helped with the zeta potential measurements.
31
32 FUNCTIONALIZATION OF OLEIC ACID-COATED IRON OXIDE NANOPARTICLES
Abstract
Superparamagnetic iron oxide NP can provide multiple benefits for biomedical
applications in aqueous environments, such as magnetic separation or MRI. To increase
the colloidal stability and allow subsequent reactions, the introduction of hydrophilic
functional groups onto the particles’ surface is essential. During this process, the
original coating is exchanged by preferably covalently bonded ligands, such as trialkoxy
silanes. The duration of the silane exchange reaction, which commonly takes more
than 24 hours, is an important drawback for this approach. In this paper we present
a novel method, which introduces ultrasonication as an energy source to dramatically
accelerate this process, resulting in high quality water-dispersible NP, around 10 nm
in size. To prove the generic character, different functional groups were introduced
on the surface, including polyethylene glycol chains, carboxylic acid, amine and thiol
groups. Their colloidal stability in various aqueous buffer solutions as well as human
plasma and serum was investigated to allow implementation in biomedical and sensing
applications.
2.1 Introduction
For many years, iron oxide NP have been the subject of intensive research. These
cost-effective and non-toxic particles are used nowadays in many applications, such
as magnetic storage, drug delivery, biosensing, magnetic separation and contrast
reagents for imaging techniques.23,109 A common method to make such particles is co-
precipitation, where iron (II+) and (III+) ions are dissolved in water and precipitated
using ammonia or sodium hydroxide.110 Drawbacks are the poor monodispersity and
irregular shape of these particles. Recently, multiple methods have been published to
synthesize very monodisperse NP in organic solvents. Sun et al. made high quality
magnetite nanocrystals with small size distribution by thermal decomposition of iron
(III) acetylacetonate in phenyl ether.14 A similar method, reported by Park et al.,
uses iron oleate as a precursor and oleic acid as a capping agent.12 This results in NP
with a hydrophobic coating, since the polar end groups are attached to the surface.
Capping agents such as oleic acid are often used because they form a protective
monolayer, which is strongly bonded. This is necessary for making monodisperse and
highly uniform NP.111,112 For biomedical applications in aqueous environments, this
hydrophobic coating has to be replaced with a hydrophilic coating. The so-called
ligand exchange is well-known for noble metal NP where for instance thiol groups
attach strongly to the surface, thereby forming monolayers by self-assembly. A similar
approach is possible for iron oxide NP by using polymers or α-cyclodextrin.113–115
Since these layers are often not covalently bonded to the surface, high ionic strength
or extreme pH conditions might alter their interaction. An elegant alternative is
silane chemistry, which is based on the reactivity of silanol molecules, formed by
the hydrolization of alkoxy silane.116,117 A high degree of control and reproducibility
is possible, when the appropriate reaction conditions are chosen. The possibility
to introduce a large variety of functional groups onto the surface of magnetic NP
MATERIALS AND METHODS 33
makes this approach very valuable. Other advantages are the high stability and
density of the formed silicon oxide layer. However, introducing silane molecules onto
the surface of oleic acid-stabilized NP has only been scarcely reported so far. De
Palma et al. described a method, which uses hexane as a solvent and acetic acid
as a catalyst to form the reactive silanol molecules.118 Larsen et al. published a
protocol with toluene as the solvent and water being the catalyst; triethylamine was
added to facilitate the reaction.119 Kohler et al. also managed to introduce silanes
but pre-treated the oleic acid coating with a mixture of 1M ammonium hydroxide
in 1-butanol.120 These methods have serious drawbacks, e.g. the extensive reaction
time (24-72h) or pre-treatment procedure. In this chapter we present an improved
nanoparticle functionalization method. The obtained superparamagnetic NP are
thoroughly characterized by transmission electron microscopy (TEM), X-ray powder
diffraction (XRD) and vibrating sample magnetometry (VSM). Infrared spectroscopy
is used to confirm the presence of the functional groups on the surface after reaction
with silane molecules. By performing this reaction in an ultrasonication bath, the
reaction time is greatly reduced, while avoiding cross-linking thus maintaining the
monodispersity. The colloidal stability of the resulting NP was extensively tested
in different aqueous media at several pH’s as well as in human serum and plasma,
which demonstrates their applicability in biomedical applications. A large variety of
functional groups were introduced to the surface, proving the generic character of the
method.
2.2 Materials and methods
Materials
Sodium oleate and iron (III) chloride hexahydrate (97%) were obtained from Sigma
Aldrich, ethanol (absolute) and oleic acid from VWR and heptane and toluene from
Fisher Scientific. Triethyl amine was ordered at Janssen Chimica. Acetone was
purchased at Chem Lab. Methoxy (polyethyleneoxy) propyl trimethoxy silane (90%,
9-12 PE-units), 3-Mercapto propyl trimethoxy silane (95%), N-(Trimethoxysilylpropyl)
ethylenediamine triacetic acid trisodium salt (45%) and 3-Amino propyl trimethoxy
silane (97%) were obtained from ABCR. 1-octadecene (90%, technical grade) was
purchased at Acros.
Synthesis of NP
Superparamagnetic iron oxide nanoparticles were prepared using the method published
by Park et al., with minor modifications.12 It consists of two separate reactions, first
preparing an iron-oleate precursor, which is later on transformed into iron oxide
nanocrystals. For the synthesis of the precursor, sodium oleate (36.5 g, 120 mmol) and
iron (III) chloride hexahydrate (36.5 g, 120 mmol) were dissolved in a mixture of 80 mL
ethanol, 60 mL MilliQ water and 140 mL heptane. This mixture was heated to reflux
at 70°C, for 4 hours under an argon atmosphere. Afterwards the upper heptane layer,
which contains the iron-oleate, was separated using a separatory funnel and washed
three times with 40 mL MilliQ water. As a final step, the heptane was evaporated
using a rotavapor, resulting in a dark brown waxy solid. The iron oxide NP synthesis
starts with mixing iron-oleate (36 g, 40 mmol) with oleic acid (5.7 g, 20 mmol) and
200 g of 1-octadecene in a 500 mL three-neck flask. This mixture was first heated to
100°C for 5 min to evaporate all remaining heptane. After fitting a reflux cooler, the
mixture was heated further to 320°C and kept at that temperature for 30 min. Around
250°C, the decarboxylation of the oleate starts, producing a large amount of CO2 gas.
Afterwards the reaction mixture is cooled down to room temperature by removing
the heat source. 500 mL of ethanol is added to precipitate the freshly prepared NP.
Separation was done by centrifugation, after which the particles were washed three
times with ethanol. After drying, the nanocrystals were dispersed in heptane (with
one drop of oleic acid) in high concentration (100 mg/mL) for long-term storage.
Functionalization
The new protocol presented here was partially based on a method published by Larsen
et al., but with important modifications.119 In a typical functionalization experiment,
100 mg of iron oxide nanoparticles (in heptane, stock solution) were mixed with 50 mL
of toluene. To this mixture, 2.5 mL of triethylamine, 0.05 mL of MilliQ water and 0.5
mL of the desired silane were added. The beaker was then placed in an ultrasonication
bath for 5 hours. The temperature of the water inside this bath was kept at 50°C
during the reaction. Afterwards, the volume of the reaction mixture was doubled by
adding heptane to the solution, 50 mL in this case. The mixture was placed on a
magnet to precipitate the functionalized nanoparticles. The supernatant was decanted
and the particles were washed 3 times with acetone and precipitated by a magnet.
After drying under reduced pressure for 15 min, the sample was weighed and dissolved
in MilliQ water or the appropriate medium.
Equipment & Characterization
The ultrasonication bath used in the particle functionalization was a Bransonic

Model 5510 sonicator with a capacity of 10 liters. The built-in heating was never
used. TEM measurements were performed on a 80 kV Zeiss EM-900 using 300 mesh
Formvar coated copper grids. Distribution data were calculated using ImageJ. Oleic
acid coated nanoparticles were dispersed in heptane and deposited onto the grid.
Fourier transform infrared (FTIR) spectra were obtained by a Bruker Alpha FTIR
spectrometer, equipped with a Platinum ATR module. Dynamic light scattering and
zeta potentials were measured on a Brookhaven 90plus particle analyzer. The internal
detector was positioned at 90 degrees. UV-VIS Spectrometry was performed on a
Perkin Elmer Lambda 900 spectrometer. VSM experiments were conducted on a VSM
Maglab setup from Oxford Instruments. XRD spectra were obtained in reflection
(Bragg-Brentano geometry) by a Rigaku Rotaflex diffractometer, fitted with a Rigaku
RESULTS & DISCUSSION 35
RU-200B rotating Cu-anode (λ = 1. 54 Å) at a power of 4 kW. The diffracted X-rays

were collected after Ni-filtering on a scintillation counter. Samples were deposited
on a glass microscope slide from solution. Samples, to test the colloidal stability,
were prepared by adding a concentrated nanoparticle dispersion (in water) to the
appropriate medium. The protocol for collection of human plasma and serum is
described in the supporting information at the end of this chapter. Absorbance values
were measured at a wavelength of 1000 nm.
2.3 Results & Discussion
Characterization of the core nanoparticle
Efficient surface modification of superparamagnetic NP is crucial for their application

in biomedicine. Recent advances in high temperature syntheses already improved the
shape and monodispersity of the core. As mentioned before, these particles are only
soluble in apolar solvents, because of their oleic acid coating. To change the polarity of
the layer to being hydrophilic, a ligand-exchange is essential. We preferred interaction
with alkoxy silanes to polymers because of the covalent bond formation, resulting in
more versatile and robust nanoparticles. The reaction mechanism is shown in Figure
2.1. After formation of the active silanol molecule, it reacts with the surface OH
groups of the iron oxide nanoparticle. This results in the formation of a Fe-O-Si bond.
Metal oxides are known to have reactive hydroxyl groups present on their surface,
caused by the adsorption of water.19 This is similar to functionalization of silicon
oxide substrates.62 The density and structure of the shell depends largely on the
reaction conditions, being a combination of linear polymerization, network formation
and covalent attachment to the iron oxide surface. Although chlorosilanes would react
faster than alkoxy silanes, we opted for the latter. Chlorosilanes release hydrogen
chloride upon reaction, which is incompatible with iron oxide. On top of that, their
reactivity is often too high to allow sufficient control of the reaction. The nanoparticle
synthesis published by Park et al. allowed us to make high quality spherical NP at
large scale.12 Figure 2.2 shows a TEM image of the oleic acid coated NP. As can be
seen in the inset, the nanoparticle distribution is narrow. A mean size of 9.3 nm was
calculated by a Gaussian fit, with a spread of ±1.6 nm (one sigma).
Since superparamagnetism is an important property of the NP, VSM measurements

were performed. Figure 2.3 shows the data of the oleic acid coated particles. The
applied field was varied between four and minus four Tesla, while recording the remnant
magnetization. No coercivity or magnetic remanence is observed, which is typical for
superparamagnetic NP. The hysteresis loop can be fitted by a Langevin function to
deduct the size of the magnetic core. In this case a core diameter of 10.5 nm was
obtained. This value is comparable to the size determined by TEM measurements.
The crystal structure of the oleic acid coated iron oxide NP was determined via
XRD (step size = 0.05°, dwell time = 40 sec). As Figure 2.4 indicates, the spectrum
Figure 2.1: Overview of the chemical reactions during the functionalization of iron
oxide NP with silanes. The formation of the silanol molecule occurs by reaction with
water. Subsequent polycondensation renders a silane network on the surface of the
nanoparticle.
closely resembles the reference spectrum of magnetite. However since the difference
between maghemite and magnetite is very subtle in XRD, no conclusions about
the exact composition can be drawn. Therefore, we can only state that the crystals
produced in the synthesis are superparamagnetic iron oxide NP, consisting of magnetite,
maghemite or a mixture of both. By using the Scherrer equation, the crystal size
can be derived from the peak broadening in the spectrum.26 Using MDI Jade, the
(400) peak was fitted with a pseudo-Voight function after polynomial background
subtraction, resulting in a size of 9.3±0.7 nm. This value corresponds very well
with the size determined by TEM measurements. Compared to previously reported
functionalization methods, the method presented here has several advantages. (1) The
reaction time is reduced to 5 hours compared to 24 or 72 h as described by Larsen
et al. and De Palma et al.118,119 (2) By using a sonicator, crosslinking of particles
during the reaction is strongly reduced thereby maintaining the monodispersity of
the NP. In this case the exact mechanism is unknown but in general, sonication of
a solution introduces microbubbles, which subsequently implode. These implosions
generate high temperatures inside and around the cavity, as well as a shock wave upon
collapse. The surrounding liquid quickly disperses the heat, allowing the use of fragile
organic materials.121,122 (3) Another advantage is the elevated concentration of NP
(2 mg/mL) during the synthesis, which reduces the amount of solvent needed. The
functionalization procedure was performed with four different trialkoxy silanes. PEG,
carboxylic acid, amine and thiol groups were introduced on the surface for various
reasons. Thiol and amine groups are excellent anchor points for subsequent coating
with a gold layer, by e.g. reduction of a gold containing salt. This is particularly
useful in biomedical applications where the plasmonic response of gold is used to
heat the environment or to release drugs at a specific location.48,104 PEG chains on
Figure 2.2: TEM image of oleic acid-stabilized iron oxide NP. The scale bar represents
20 nm. The inset shows the size distribution: 9.3±1.6 nm
the surface, on the other hand, provide the NP with excellent dispersibility in water.
These particles can be used to form magnetic fluids for applications like magnetic
hyperthermia or thermo-ablation.123 The introduction of carboxylic acid and amine
groups can be utilized for bioconjugation of proteins to the NP. Typical chemical
coupling reagents like EDC-NHS or glutaraldehyde, link these functional groups to
the target protein via an amide or imine bond. These bioconjugates are often used for
detection and magnetic separation.
Successful modification of the NP’ surface was confirmed by FTIR measurements.

Table 2.1 gives an overview of the vibrations measured on the oleic acid and silane
coated particles. For the unmodified iron oxide cores, the characteristic bands of
the asymmetric stretching, symmetric stretching and scissoring of CH2 are visible at
2919, 2850 and 1436 cm−1 , respectively.112,124,125 The stretching of the C=O double
bond is clearly shown by the peak at 1709 cm−1 . This was expected since the peak
resembles free oleic acid, indicating that a significant amount of unbound surfactant
is present, which can be related to the storage conditions. The presence of a peak
at 1462 cm−1 , coming from in-plane OH bending, supports this idea. Zhang et al.
reported that the C=O peak shifts to 1541 and 1639 cm−1 when the molecule is
attached to a ferrite surface.112 The carboxylic acid groups are then present in a
COO− conformation. If the spectrum is magnified, these vibrations are also visible
Figure 2.3: VSM signal of the NP, showing a hysteresis curve. The data points are
fitted by a Langevin function to determine the magnetic core size.
Table 2.1: Overview of the different vibrations, related to the different coatings around
the nanoparticle. The original spectra can be found in the supporting information
(see Figures 2.6, 2.7, 2.8, 2.9, 2.10) at the end of this chapter.
Surface of Fe3 O4 IR vibrations (cm−1 )

Oleic acid 3005 (HC=), 2919 (CH2 ), 2850 (CH2 ), 1709 (C=O), 1635 (COO− ),
1541 (COO− ), 1462 (OH), 1436 (CH2 ), 598 (Fe-O)
COOH silane 3600-3000 (OH), 2932 (CH2 ), 1612 (COO− ), 1452 (CH2 ), 1396 (COO− ),
1113 & 1089 & 1007 (Si-O), 585 (Fe-O)
PEG silane 3600-3000 (OH), 2860 (PEG CH2 ), 1643 (H2 O), 1454 & 1349 & 1297
& 1250 & 1047 & 947 (CH2 -O-CH2 ), 1198 (O-CH3 ), 620 (Fe-O)
NH2 silane 3004 (OH & NH2 ), 2922 (CH2 ), 2850 (CH3 ), 1543 (NH+
3 ),
1400 (CH3 COOH), 1224 (Si-C), 1073 (Si-O-R), 773 (NH2 ), 617 (Fe-O)
SH silane 3600-2500 (OH & CH2 ), 2600-2550 (SH), 1645 (H2 O), 1430 (CH2 ),
1035 (Si-O), 590 (Fe-O)
Figure 2.4: X-ray powder diffraction spectrum of the oleic acid-coated iron oxide
NP. The solid and dashed vertical drop-down lines represent the peak positions of
a reference magnetite and maghemite spectrum, respectively (AMCSD 0007824 and
0007899).
in our case at 1541 and 1635 cm−1 , although they are fairly small. The wavenumber
separation (of 94 cm−1 ) between those two peaks is an indication for how the oleate
and the iron atoms on the surface interact. Because the difference is smaller than
110 cm−1 , a chelating bidentate interaction can be derived from the spectrum. The
iron oxide core itself shows a characteristic vibration at 598 cm−1 , related to the Fe-O
bond.112 The presence of carboxylic acid groups, after functionalization, is proven
by the vibrations at 1612 and 1396 cm−1 . These correspond to the asymmetric and
symmetric stretching of the COO− group, respectively. Other important features are
the 1113, 1089 and 1007 cm−1 bands due to the stretching of the Si-O bond.126,127
The PEG silane shows distinct bands caused by the ether functions in the chain,
to which several peaks between 1454 and 947 cm−1 can be attributed. The high
hydrophilicity of the PEG chain is expressed by the presence of a water peak in
the spectrum. Even after extensive drying this peak remains, indicating that the
water is trapped by hydrogen bonding.118 Several papers already reported that more
information about the structure of the PEG layer on the nanoparticle’s surface can be
derived from the FTIR spectrum. Compared to these reports, the conformation of
the PEG layer is partially crystalline and partially amorphous. Small shoulders at
Figure 2.5: Zeta potential values for PEG, carboxylic acid and amine coated
nanoparticles in various pH solutions. Every data point was derived from 10
measurements by the software.
1244, 1460 and 1470 cm−1 are present in the spectrum, indicating crystalline parts in
the coating. On the other hand, also the amorphous bands are visible, at 948, 1140
and 1250 cm−1 .128,129 For the amino silane coated nanoparticles, the most important
peaks are visible at 3004, 1543 and 773 cm−1 corresponding to the protonated amines.
Because this protonation was done by adding acetic acid, a peak at 1400 cm−1 appears.
Mercapto silane coated particles show a broad band above 2500 cm−1 , consisting of a
combination of OH, CH2 and SH vibrations. Also a small peak caused by the presence
of water is visible. Typical peaks for Si-O en Fe-O also appear, at 1035 and 590 cm−1 ,
respectively.
Colloidal stability
Further experiments were only conducted on the PEG, COOH and amine coatings,
since the thiol particles were only stable in water for a very short time, regardless
the pH. In general, the stability of NP dispersed in aqueous media can be expressed
by the zeta potential. Theoretically, it refers to the potential difference between the
slipping plane in the electronic double layer and the bulk potential. If the potential
Table 2.2: Overview of the stability of the nanoparticles in buffer solutions: (+) stands
for excellent dispersibility and stability in time (minimum 1 week), (±) corresponds
to colloidal solutions that are stable for <4 days, (-) stands for dispersions which are
stable for only a short period of time (between 5 and 10 hours). Concentration of
nanoparticles in solution was 0.25 mg/mL. Criterion for stability was the absence of
visible aggregation or precipitate.
Coating CH3 COOH/CH3 COONa CH3 COOH/CH3 COONa MES/HCl

100% pH 4 pH 5 pH 6
0.1 M 0.05 M 0.025 M 0.1 M 0.05 M 0.025 M 0.1 M 0.05 M 0.025 M
NH2 + + + + + + ± ± ±
PEG + + + ± ± ± ± ± ±
COOH ± + + + + + + + +
Coating NaH2 PO4 /Na2 HPO4 TRIS/HCl Glycine/NaOH

100% pH 7 pH 8 pH 9
0.1 M 0.05 M 0.025 M 0.1 M 0.05 M 0.025 M 0.1 M 0.05 M 0.025 M
NH2 - - - - - - - - -
PEG + + + + + + + + +
COOH + + + + + + + + +
has an absolute value higher than 25-30 mV it is generally accepted that the particles
are electrostatically stable.68 Although colloidal stability is related to electrostatic and
steric repulsion, zeta potential measurements usually give a good indication. Figure
2.5 shows the combined data of the PEG-, carboxylic acid- and amine-coated NP. A
clear downward trend for the zeta potential of NH2 is visible when the pH increases.
This can be related to the lowering of the surface charge due to deprotonation of
the amine at high pH. A similar but reverse trend can be observed for COOH, since
the acid becomes protonated at low pH, thereby losing its negative charge. The zeta
potential would eventually approach zero at a pH lower than 4. For PEG-coated
particles the zeta potential shows similar behavior, even though polyethylene glycol
chains have no pH responsive groups. The incorporation of ions into the PEG layer
can explain this trend, taking into consideration that sodium hydroxide and hydrogen
chloride were added to adjust the pH of the solution.130,131 . Nevertheless, the PEG
coated particles were stable over the entire pH range, indicating that their stability is
caused by steric repulsion rather than electrostatic repulsion. For the COOH-coated
nanoparticles, solutions of pH 4 and 5 showed extensive aggregation and consequent
precipitation. A similar effect was observed for the amine-coated particles, this time
for dispersions of pH 7 to 9. The profound impact of the pH on the stability of
COOH and NH2 coatings shows that electrostatic repulsion is crucial for these types
of dispersions.118,124,125
To study the effect of buffer solutions onto the colloidal stability, various dispersions
with different salt concentrations were prepared. To our knowledge, no extensive
studies have been conducted on magnetic nanoparticle dispersions in different buffer
media at different pH’s and concentrations.132,133 . Nevertheless this information can
be very valuable for subsequent reactions or applications of the nanoparticles. Five
commonly used buffer reagents were chosen, covering the entire pH range between 4
and 9. These included sodium acetate, 2-(N-morpholino)-ethanesulfonic acid (MES),
sodium hydrogen phosphate, 2-amino-2-hydroxymethyl-propane-1,3-diol (TRIS) and
glycine. The concentration of the reagents ranged from 0.1 M to 0.025 M; while
the particle concentration was fixed at 0.25 mg/mL. Table 2.2 provides an overview
of the stability in different buffer solutions, each having 3 different concentrations.
Carboxylic acid coated NP show excellent stability in different buffers above pH 4,
which can be related to the presence of charged carboxylate groups on their surface,
providing sufficient electrostatic repulsion. For the NH2 coated particles, the instability
can be explained by the lack of charged functional groups. This is partially caused
by the pH, which is too high for sufficient stabilization. On top of that, the high
ionic strength and ion size largely influence the zeta potential. These results are in
perfect agreement with the zeta potential measurements and the observed correlation
between the presence of surface charges and colloidal stability. The particles with
polyethylene glycol chains on their surface show good stability in phosphate, glycine
and TRIS based buffer solutions. The instability in MES and acetate buffer however
is somewhat unexpected, since the zeta potential measurements showed that PEG
provides mainly steric hindrance, rather than electrostatic repulsion. Ion adsorption
onto the coated surface is expected to be the cause of this instability.
If superparamagnetic NP are used as for example a MRI contrast reagent, colloidal

stability in (human) serum or plasma is crucial. To prove the value of the presented
functionalization method, nanoparticle dispersions in both serum and plasma were
prepared and their stability was monitored in time by absorption measurements (see
supporting information at the end of this chapter). Human blood has a pH of 7.4,
which will strongly influence the stability of the nanoparticles as was shown in previous
paragraphs. Similar as for the dispersions in buffer solutions, three different coatings
were tested, respectively amine, carboxylic acid and PEG functional groups. These
NP were dispersed in serum and plasma at a concentration of 1 and 0.25 mg/mL.
In accordance to the previous results, amine groups on the surface cannot provide
sufficient electrostatic repulsion at pH 7.4, which resulted in a rapid decline of the
absorbance. On the contrary, carboxylic acid groups and PEG chains should be able
to provide sufficient stabilization in these conditions. This was observed in both cases,
since the absorbance remained constant during the entire experiment, which lasted 48
hours. These results indicate that the PEG and carboxylic acid coated nanoparticles
can be of great importance for future in vivo experiments.
An important remark about the use of NP in biomedical applications is their possible

toxicity towards cells. Although many conflicting results were published about the
toxicity of superparamagnetic iron oxide NP, a study by Mahmoudi et al. showed
that the surface coating as well as the cell type itself has a large influence on the
possible toxic effects.17,117 They state that the introduction of a functional surface
coating lowers the toxicity. Tartaj et al. reported that the size, shape and magnetic
dipole moment of the particle also play a role in in vivo experiments.134 Properties
of nanoparticles like longer sedimentation times, higher surfaces areas and smaller
magnetic dipole-dipole interactions might facilitate their use. Nevertheless a study of
the adverse effects towards their environment is necessary for every specific application.
CONCLUSIONS 43
2.4 Conclusions
Oleic acid-coated NP were functionalized by reaction with trialkoxy silanes. The
reaction takes place in an ultrasonication bath, which reduces the reaction time
and prevents crosslinking. The successful coating procedure of the NP’ surface was
proven by FTIR measurements. Multiple techniques (TEM, XRD, VSM) proved the
composition of the magnetic core. The obtained functionalized NP can be dispersed
in various aqueous environments, including human serum and plasma. Their stability
under these conditions was addressed by zeta potential and absorbance measurements,
showing a strong relation between the colloidal stability and the pH of the solution.
Although PEG coated NP also exhibit this dependency, steric hindrance is expected
to be more prominent here. In general, the generic method described here allows
the introduction of various functional groups on the surface of the NP. This is
particularly useful for subsequent coupling reactions to fluorescent probes, proteins or
substrates. Therefore, we believe that this type of superparamagnetic NP can be of
major importance for future research and applications in biomedicine.
2.5 Supporting information
Protocol human plasma and serum
Human serum and plasma were obtained by the following method: a blood sample
from a healthy volunteer was collected in heparinized tubes (BD Vacutainer Systems)
and spun at room temperature at a speed of 1300 g for 20 minutes in a swinging
bucket centrifuge, with plasma harvested and stored at -20°C until assayed. For serum
preparation, a blood sample from the same healthy volunteer was collected in SST
Serum Separation Tubes (BD Vacutainer Systems), inverted five times and allowed to
clot at room temperature for 30 minutes before centrifugation in a swinging bucket
centrifuge for 20 minutes at 1300 g; the resultant serum was collected and stored at
-20°C until assayed.
FTIR spectra
Following figures show the FTIR spectra of the coated NP. Oleic acid (Figure 2.6),
carboxylic acid (Figure 2.7), PEG (Figure 2.8), amine (Figure 2.9) and thiol (Figure
2.10) coatings were measured.
Figure 2.6: Infrared spectrum of oleic acid coated NP.

SUPPORTING INFORMATION 45
Figure 2.7: Infrared spectrum of NP coated with carboxylic acid groups.
Figure 2.8: Infrared spectrum of NP coated with PEG chains.

Figure 2.9: Infrared spectrum of NP coated with amine groups.
Figure 2.10: Infrared spectrum of NP coated with thiol groups.

Absorbance spectra
The following figures combine the absorbance data of the amine (Figure 2.11),
carboxylic acid (Figure 2.12) and PEG (Figure 2.13) coated NP respectively. The
concentration of the NP was 1 or 0.25 mg/mL in serum or plasma, as indicated in
the legend. Absorbance values were measured at a wavelength of 1000 nm. At pH
7.4, the amine functionalized NP are not colloidally stable, resulting in a decrease in
absorbance over time. The COOH and PEG coated particles on the other hand show
excellent stability in these complex environments.
Figure 2.11: Absorbance data of the amine coated NP. The concentration of NP in
serum or plasma was 1 or 0.25 mg/mL.
Figure 2.12: Absorbance data of the carboxylic acid coated NP. The concentration of
NP in serum or plasma was 1 or 0.25 mg/mL.
Figure 2.13: Absorbance data of the NP coated with PEG chains. The concentration
of NP in serum or plasma was 1 or 0.25 mg/mL.
Chapter 3
Heterobifunctional PEG
ligands for bioconjugation
reactions
Bloemen M, Van Stappen T, Willot P, Lammertyn J, Koeckelberghs G, Geukens N,

Gils A, and Verbiest T
Heterobifunctional PEG ligands for bioconjugation reactions on iron oxide
nanoparticles
PloS ONE 2014; 9: e109475
T. Van Stappen performed the enzyme-linked immunosorbent assays and analyzed the
results. Dr. P. Willot and Prof. dr. G. Koeckelberghs helped with the interpretation
of the NMR and mass spectrometry results. Prof. dr. J. Lammertyn and his team
were responsible for the SPR measurements and the data analysis.
49
50 HETEROBIFUNCTIONAL PEG LIGANDS FOR BIOCONJUGATION REACTIONS
Abstract
Ever since iron oxide nanoparticles have been recognized as promising scaffolds
for biomedical applications, their surface functionalization has become even more
important. We report the synthesis of a novel polyethylene glycol-based ligand that
combines multiple advantageous properties for these applications. The ligand is
covalently bound to the surface via a siloxane group, while its polyethylene glycol
backbone significantly improves the colloidal stability of the particle in complex
environments. End-capping the molecule with a carboxylic acid, introduces a variety
of coupling chemistry possibilities. In this study an antibody targeting plasminogen
activator inhibitor-1 was coupled to the surface and its presence and binding activity
was assessed by enzyme-linked immunosorbent assay and surface plasmon resonance
experiments. The results indicate that the ligand has high potential towards biomedical
applications where colloidal stability and advanced functionality is crucial.
3.1 Introduction
The potential of iron oxide NP in biomedical applications is widely recognized: they
can act as MRI contrast agents, superparamagnetic carriers for drugs or are used
in hyperthermia treatments.6,48,134–137 By improving the synthesis of these particles,
their quality and availability has largely increased.12,14,33,138–140 When NP are used
in biomedical applications, two requirements are often necessary. First, their colloidal
stability in complex environments is crucial. If the particles become unstable in
for instance blood, they will precipitate, possibly triggering severe inflammatory
responses.141–143 Secondly, they should possess accessible anchor points for molecules
or proteins to be coupled onto. This allows NP to selectively interact with certain
targets or to carry drugs close to a desired location. However, functionalization of
their surface has proven to be non-trivial. Although multiple different approaches
have been developed, most of them lack a certain degree of control.23 Coating their
surface with functional polymers is a straightforward method, but has cross-linking
issues and allows little control over the thickness of the layer and orientation of
functional groups.144 Since they are not covalently attached to the surface, they could
potentially detach, which would make the particles precipitate. Growing an additional
silica layer on the iron oxide core, on the other hand, has several advantages: the
shell thickness can be well controlled and it is chemically inert.145 However, the
diameter of such NP increases by several nanometers, which is often not desired for
biomedical applications.146 This problem was circumvented by the introduction of
functional siloxane molecules on iron oxide NP . They also form a silicon dioxide
shell, albeit very thin, and they contain a functional group, which can have several
advantages or uses later on.119,147 Even though multiple variants of these silanes are
commercially available, they often do not have the desired structure or properties.
This can easily be related to the complicated handling of siloxane molecules. Since
they react with water and are relatively intolerant to heat, modification reactions
have to be limited in time and workup. Tucker-Schwartz et al. recently published an
easy method to avoid this direct modification of the siloxanes, by adopting thiol-ene
click chemistry.148 Their approach allows to synthesize a very complex molecule first
and attach a siloxane group as the final step. Click chemistry is a concept rather
than a specific reaction, which comprises fast reactions with very high yields and
non-aggressive by-products.149,150 In addition the reaction should be modular and
have relatively simple reaction conditions. Very well-known examples are copper
mediated azide-alkyne cycloadditions, thiol-ene and Diels-Alder reactions.150,151 We
developed a new ligand, based on a PEG backbone, and transformed it into a siloxane
by straightforward thiol-ene click chemistry. By modifying the end-group of the
backbone, functional groups were easily introduced onto the nanoparticle’s surface.
The high purity and straightforward synthesis of the ligand makes this method very
valuable for large scale and reproducible functionalization of iron oxide NP. This
universal method requires only basic knowledge of organic chemistry and can be widely
applicable by scientists without a substantial chemistry background. To investigate the
full potential of the ligand, several Ab were coupled to its anchor groups (carboxylic
acids) and their activity was assessed via fiber optic SPR experiments. As a model
system, an antibody (MA-33H1F7) targeting the serine protease inhibitor (serpin)
plasminogen activator inhibitor-1 (PAI-1) protein was selected.152 This protein is an
important factor in the plasminogen-plasmin system since it inhibits plasminogen
activators, tissue-type plasminogen activator and urokinase, which are involved in clot
formation and degradation processes in blood.153 These Ab were coupled to the NP
by using popular EDC-NHS chemistry and their presence was investigated by ELISAs.
To assess their potential in biomedical applications, their colloidal stability was tested
in undiluted human plasma and serum. The results indicate that the developed ligand
has high potential because of its elegant synthesis, its positive influence on the colloidal
stability of the nanoparticle as well as its properties for antibody coupling chemistry.
Materials
2,2-dimethoxy-2-phenylacetophenone (DMPAP, 99%), 4-dimethylamino pyridine

(DMAP, >99%), succinic anhydride (>99%), 1-Ethyl-3-(3-dimethylaminopropyl)
carbodiimide (EDC) and mercaptopropyl trimethoxysilane (95%) were purchased from
Sigma Aldrich. Allyl-PEG10-OH was obtained from Polysciences, Inc. Triethylamine
was ordered at Janssen Chimica. N-hydroxy succinimide (NHS, 98+%) was obtained
from Alfa Aesar. 2-(N-morpholino)ethanesulfonic acid monohydrate (MES) was
purchased at Fluka. The monoclonal antibodies (host: mouse) used in this study are
MA-33H1F7 (target: human PAI-1/t-PA complex) and MA-T12D11 (target: human
Thrombin Activatable Fibrinolysis inhibitor (TAFI)), supplied by the Therapeutic
and Diagnostic Antibodies group of the KU Leuven.152 All ultrasonication steps were
performed in a Branson 5510 sonicator bath. FTIR spectra were measured using a
Bruker Alpha FT-IR spectrometer equipped with a Platinum ATR module.
Synthesis of carboxylic acid-terminated PEG
In a 50 ml flask, allyl-PEG10-OH (1 eq, 4,00 mmol, 1.992 g) was mixed with succinic
anhydride (1.1 eq, 4.40 mmol, 440 mg) and 4-dimethylaminopyridine (DMAP) (0.02
eq, 0.08 mmol, 9.7 mg). This mixture was stirred and heated to 50°C for 16
hours. The resulting product was purified twice by precipitation in cold diethyl
ether, centrifugation and drying in vacuum. 1 H NMR (300 MHz, CDCl3 ): δ (ppm)
2.65 (s, 4H), 3.6-3.7 (m, 38H), 4.02 (d, 2H), 4.26 (t, 2H), 5.16-5.30 (m, 2H), 5.8-6.0
(m, 1H). 13 C NMR (75 MHz, CDCl3 ): δ (ppm) 29.2, 29.5, 63.8, 68.9, 69.3, 70.5, 72.2,
117.1, 134.7, 172.1. MS (chemical ionization, isobutane): m/z = 499 (ester fragment,
M+ - C4 O3 H), 101 (ester fragment, M+ - C23 O11 H46 )
Thiol-ene click chemistry
To form the siloxane-terminated PEG molecule, allyl-terminated PEG (mixture of

modified and unmodified, 1 mmol) was mixed with (3-mercaptopropyl) trimethoxysi-
lane (1 eq, 1 mmol, 185.7 µL) and 2,2-dimethoxy-2-phenyl-acetophenone (DMPAP)
(0.05 eq, 0.05 mmol, 12.8 mg). This mixture was stirred for 1 hour in a UV chamber,
equipped with 3 LEDs (365 nm, output power 200 mW). If smaller quantities are used,
a small amount of chloroform can be added to improve the stirring.148 The product
was used without further purification. 1 H NMR (300 MHz, CDCl3 ): δ (ppm) 0.76 (t,
2H), 1.70 (m, 2H), 1.85(m, 2H), 2.55 (m, 4H), 2.64 (s, 4H), 3.57 (s, 9H), 3.5-3.8 (m,
40H), 4.26 (t, 2H)
Nanoparticle surface functionalization
The synthesis of iron oxide NP as well as the introduction of siloxanes onto their surface
was performed as reported in Chapter 2. In general, 1 mmol of siloxanes is mixed with
100 mg of Fe3 O4 NP in 50 mL of toluene. To this mixture 2.5 mL of triethylamine and
50 µL of water are added. The solution was placed in a ultrasonication bath for 5 hours,
after which 50 mL of heptane was added to precipitate the particles. Afterwards, they
were attracted magnetically and washed 3 times with acetone. Finally the particles
were dried in vacuum and dispersed in MilliQ water (with a concentration up to
20mg/mL).
Protein coupling
The concentrated nanoparticle solution was diluted in 50 mM MES buffer, pH 5.5,

to reach a final concentration of 3 mg/mL. 0.75 mg EDC and 0.75 mg NHS was
added to 1ml of this solution and shaken for 20 minutes to activate the carboxylic
acids. The antibodies were diluted in 2 mL of the same MES buffer after which
both solutions were mixed and shaken for 1 hour. To separate the particles from the
solution, a Miltenyi Biotech MS magnetic column was used. After rinsing the column
with MilliQ water, the nanoparticle dispersion was run through the column, which was
placed inside a circular magnet. The column was washed 2 times with 1mL of sodium
phosphate buffer (20 mM, pH 7). To elute the particles, the column was removed
from the magnet and 1mL of phosphate buffer was used as eluent.
ELISA
In the ELISA assay, recombinant PAI-1 is coated on the plate and free binding sites
are blocked with bovine serum albumin. Samples are applied in different dilutions as
well as a standard curve of MA-33H1F7.152 After incubation, horseradish peroxidase
(HRP) conjugated rabbit anti-mouse IgG (Sanbio B.V., Uden, The Netherlands) is
applied, followed by an o-phenylenediamine (OPD) induced colorimetric reaction. The
intensity of the color is directly correlated with the amount of bound MA-33H1F7.
Sample values are calculated using the standard curve.
SPR measurements
An optical fiber was first coated with a gold layer, which was subsequently covered
with a self-assembling monolayer (SAM). This thiol- and carboxyl-terminated molecule
was obtained from Dojindo molecular technologies. The SAM was activated by a
solution containing 0.4 M EDC and 0.1 M NHS in a 50 mM MES buffer (pH 6.0) for
20 minutes. Afterwards the fiber was brought into a solution containing the antigen,
PAI-1 (24 µg/mL) for 25 minutes. Finally the fiber was transferred into a blocking
solution (0.1 % tween and 0.05 % BSA). All subsequent experiments were performed
with 1 mg/mL nanoparticle solutions.
Ligand Design
The novel PEG-siloxane ligand was designed bearing two important characteristics in
mind: having one accessible functional group and providing excellent colloidal stability
to the nanoparticle. To ensure the first property, a PEG-oligomer end-capped with an
allyl functionality was modified with succinic anhydride. This reaction was performed
without solvent, since the anhydride dissolves in PEG at elevated temperatures. DMAP
was added, as a catalyst, to speed up the reaction.154 The available hydroxyl group at
the end of the PEG chain reacts with the anhydride, resulting in a free carboxylic
acid (see Figure 3.1). This product was purified once by precipitating it in diethyl
ether, which removed traces of the catalyst and excess anhydride.
Figure 3.1: Allyl-terminated polyethylene glycol was modified by reaction with succinic
anhydride. DMAP catalyzes this reaction. Subsequently the allyl functionality is
reacted with a thiol-containing siloxane molecule, by thiol-ene click chemistry, which
yields the final carboxylic acid-terminated PEG-siloxane.
The second step of the ligand synthesis involved a click chemistry reaction. We opted
for this approach, since working with siloxanes is difficult. They react with moisture
and are not resistant to prolonged heating.148 The thiol-ene click chemistry, on the
other hand, is fast and takes place at room temperature. Another great advantage
of this approach is that the final siloxane molecule can be added directly to the
functionalization solution, without additional workup. Any traces of the radical
initiator or its by-products are inert in this reaction. Even though the functionalized
PEG molecule could provide sufficient steric hindrance, which ensures colloidal stability
of the nanoparticle, we chose to mix modified and unmodified PEG siloxanes during
the functionalization step.147 Thus, the nanoparticle is covered with a complete PEG
shell, where the modified chains are sterically available, since they are longer. From
the FTIR data (data in the supporting information at the end of this chapter, Figures
S1 and S2) was derived, that even though the chain length of the ligand is sufficiently
short to enhance the stacking of the molecules (crystalline domains), a small percentage
is coiled (amorphous domains).147 Because only a small part of the ligands have a
carboxylic acid functionality, the overall pH sensitivity is reduced. The end result
(idealized) is shown in Figure 3.2: carboxylic acid groups are now available as anchor
points for future reactions. By FTIR measurements, the presence of modified PEG
chains was observed (data in the supporting information at the end of this chapter,
Figures S1 and S2). All further experiments were conducted on nanoparticles coated
with 90% unmodified and 10% modified PEG siloxanes (molar percentages). These
functionalized NP show excellent colloidal stability in multiple different environments.
Similarly to Chapter 2, we tested the stability in undiluted human serum and plasma
(data in the supporting information at the end of this chapter, Figure S3).147 The
nanoparticles (8.6±0.6 nm, TEM data in the supporting information at the end of
this chapter, Figure S5), coated with mixed siloxanes, clearly showed the properties
of both PEG and carboxylic acids. In particular they show enhanced stability in pH
ranges above 5, where the carboxylic acids are charged (picture in the supporting
information at the end of this chapter, Figure S4). In media like serum or plasma, no
precipitation was observed, even after 25 hours at room temperature without agitation.
NP were also coated with 100% modified PEG siloxanes, but these particles had
significantly lower colloidal stability in these acidic environments (pH 5-6), due to the
Figure 3.2: The available carboxylic acid groups are activated with EDC-NHS
chemistry. The resulting NHS ester reacts with amine groups of the antibody in
a MES buffer. Finally the particles are recovered from the supernatant by a magnetic
column.
lack of stabilizing charges. Steric stabilization by the PEG chains was not sufficient in
this case, since to much carboxylic acids were present. We therefore decided to focus
on nanoparticles with mixed siloxane coatings.
Bioconjugation
Covalent attachment of the selected Ab onto the NP was performed via a standard
EDC-NHS coupling.155 The mechanism is based on the activation of the carboxylic
acid with EDC, which forms an unstable acylisourea intermediate. This intermediate
reacts with NHS to form a stable ester that exhibits enhanced stability in aqueous
environments. Although this extra step is not strictly necessary, it greatly improves
the binding efficiency, by reducing the occurrence of side reactions on the acylisourea
intermediate. All reactions were performed in a slightly acidic buffer (MES 50mM,
pH5.5), which improves the final coupling reaction on two domains. First, the low
pH enhances the activation of the carboxylic acid by EDC.83 Secondly, the formed
NHS ester has substantially lower hydrolysis rates below pH 7.83 Further protein
crosslinking (second, third, . . . layer) is reduced by the slow reaction rate of the
partially protonated amines.83 A slower reaction rate was preferred in this procedure,
since the formation of a protein corona is also a thermodynamically favorable process.70
A higher reaction rate could result in coating the NP with multiple layers of proteins
and crosslinking between different NP . Afterwards the conjugated NP were purified by
a magnetic column, which has a very large surface area, since normal attraction with a
magnet was too time-consuming. This was necessary because of the excellent colloidal
stability of the NP in the buffer, which dramatically slows down the attraction rate.
If the NP were precipitated by a highly concentrated salt solution, it was difficult to
redisperse them afterwards. Using a magnetic column also enabled us to wash the
particles while they were retained on the column. After removing the magnetic field
from the column, the particles were easily collected by eluting with a PBS buffer.
In this study, we opted for two different Ab: MA-33H1F7, targeting PAI-1, and
MA-T12D11, targeting TAFI, as the negative control.152,156
Multiple methods are available to determine the concentration of proteins on NP;

however not all are appropriate when iron oxide is involved. Colorimetric methods like
Figure 3.3: If EDC-NHS coupling reagents are added to the mixture of NP and Ab,
slightly more proteins are retained on the Ab. This indicates that a small level of
crosslinking occurs. When a large amount of Ab (without coupling reagents) is added,
no significant difference is observed, which shows that only a hard corona remains on
the NP after washing. All error bars are shown as the percentage error on the total
value.
the Bradford assay are influenced by the strong light absorption of the black NP , which
makes the results difficult to interpret.157 FTIR spectroscopy can only confirm the
presence of proteins but is not appropriate for assessing the concentration. We opted
for an ELISA assay in this case, whereby the remaining proteins in the supernatant and
washing fractions were determined. This way, the amount of proteins on the surface of
the NP can easily be calculated. As a comparison, NP and Ab were also mixed in the
absence of coupling reagents, so only aspecific adsorption could occur (protein corona
formation).70,158 Hence this would set a benchmark for the protein concentration of
the hard corona formation (without possible protein crosslinking).73,159 When the
concentration of proteins was increased ten times (see Figure 3.3), the amount of
adsorbed proteins does not change significantly. This indicates that the washing steps
remove all proteins, except the hard corona, which is more strongly attached to the
surface.73,159,160 Since all three experiments without coupling reagents (including error
bars) give a similar value, we learned that the hard protein corona corresponds to
15-20 micrograms of proteins per milligram nanoparticles, which is similar to literature
for particles of comparable size and shape.106,161–165 When we added the coupling
Figure 3.4: A multimode optical fiber was coated with a gold layer, a self-assembling
monolayer (SAM) and the appropriate Ag (PAI-1), as shown in the inset. The SPR
shift caused by the NP, coated with MA-33H1F7 or MA-T12D11, is clearly visible.
reagents (25 + EDC-NHS), we obtained a result that was comparable, but slightly
higher in value. This indicates that a small amount of crosslinking is occurring, which
can be expected for EDC-NHS reactions involving proteins. However the coating
of the nanoparticles is close to the optimal value (solely the hard corona), which
underlines the quality of the coating and the coupling procedure.
Activity assessment
Although an ELISA can determine the loading capacity, it is incapable of assessing

the activity of the coupled antibodies on the spherical nanoparticles. To investigate
the ability of the Ab to recognize their Ag, the nanoconjugates were brought into
contact with a PAI-1-coated SPR optical fiber. By looking at the shift in the plasmon
wavelength, the interaction between Ab and Ag can be assessed. A standard multimode
optical fiber was coated with a gold layer and a self-assembling monolayer with
carboxylic acid end-groups.166 To these groups, PAI-1 was coupled via EDC-NHS
chemistry (Figure 3.4). When the nanoparticles were brought into contact with the
fiber, they induced a shift in the plasmon resonance relative to their binding efficiency.
The binding, as a whole, is the sum of two separate interactions: the protein corona
effect and the antibody-antigen (Ab-Ag) bond. The first is caused by the aspecific
interaction between the Ab and the Ag, similar to the formation of a second (soft)
corona, this cannot be avoided and hence is viewed as a background in the signal.
The latter, however, is specific for each Ab-Ag couple. In this experiment, one can
clearly see the difference in SPR shift, caused by the Ab-Ag interaction. An extra shift
of more than 8 nm was measured by the SPR-fiber setup when comparing the NP ,
coated with MA-33H1F7 (targeting PAI-1) or MA-T12D11 (targeting TAFI). This
result ensures that, although the Ab are coupled in a random fashion, their activity
is retained and they are still partially sterically accessible. We hypothesize that a
large fraction of the Ab indeed lose their activity due to an unfavorable direction of
bonding. However, the strongly curved, large surface of the NP allows a high overall
antibody loading capacity that partially compensates for the losses in activity. Future
experiments will focus on employing a more directional coupling strategy, which will
give us more insight in this complex relation.
The excellent colloidal stability of the NP, coated with the PEG-ligand, will allow
to use the particles for various biomedical applications. Since the ratio of functional
ligands can easily be adjusted, a library of mixed-monolayer nanoparticles can be
synthesized for future experiments. Similarly, the core size of the NP can be varied,
to control the overall size of the bioconjugates. This can have an important influence
on their cell uptake or retention time in vivo.167,168 Moreover, they can serve as a
platform for the bioconjugation of proteins for multiple applications like selective
magnetic separation or MRI contrast agents.
3.4 Conclusions
In order to fully customize the surface coating of iron oxide NP, a PEG building block
was modified with carboxylic acid groups and afterwards attached to a siloxane via
thiol-ene click chemistry. These ligands were introduced onto the NP’ surface, which
significantly improved the colloidal stability in complex environments. To prove their
added functionality, Ab were coupled to the carboxylic acid end-groups. An ELISA
was performed to indirectly determine the amount of coupled proteins, while SPR
experiments confirmed their activity. Because these ligands provide excellent colloidal
stability and can also act as an anchor point for coupling via a simple modification,
they have high potential in future nanoparticle design for biomedical applications.
Figure 3.5: Scheme of the thiol-ene click chemistry reaction, after the formation of
a thiyl radical by the initiator, a propagation step forms the bond to the alkene. A
chain transfer reaction forms the new radical on another thiol-containing molecule.
An ideal thiol-ene reaction is an alternation of a radical propagation and chain transfer

reaction. The propagation step involves the addition of a thiyl radical to a double
bond, while the chain transfer step involves the abstraction of a hydrogen radical
from the thiol by the carbon-centered radical. Ideally, no polymerization reactions,
chain growth by propagation of the carbon radical onto another double bond, occur.
It has been proven that the thiol-ene kinetics are largely dependent on the electron
density of the double bond and steric hindrance.151,169,170 Because of the fast kinetics,
limited side reactions and high yields, the thiol-ene reaction is categorized under
click chemistry reactions. Other typical examples are copper mediated azide-alkyne
cycloadditions, Diels-Alder reactions or ring openings of strained heterocycles.150
FTIR spectra
Figure 3.6: FTIR spectrum of the original allyl-PEG10 -OH ligand and the modified
version. The ester peak at 1725 cm−1 is clearly visible after the ring opening of the
anhydride, while the –OH peak around 3500 cm−1 disappears.
Figure 3.7: FTIR spectrum of the allyl-PEG10 -COOH ligand, the oleic acid-coated
and the modified iron oxide NP. The ester peak is still clearly visible at 1725 cm−1 ,
as well as the different PEG vibrations between 1250 and 1500 cm−1 . The presence of
the iron oxide NP is confirmed by the Fe-O and Si-O vibrations at respectively, 590
and 1100 cm−1 .
Colloidal stability data
Figure 3.8: Absorbance of nanoparticle dispersions in plasma and serum. To verify

the stability of the functionalized NP in complex environments; the absorbance of
dispersions in plasma and serum was measured at 1000 nm. The particles were
dispersed at 1 mg/mL and the absorbance was monitored for 25 hours. A significant
decrease of the absorbance would indicate colloidal instability and precipitation of the
NP.
Figure 3.9: The colloidal stability of the nanoparticle dispersions is excellent, even
after 1 year of storage. The samples (5 mg/mL in water, pH 7) shown above have the
following coatings (molar percentages): A, 100% PEG10 -OH; B 10% PEG10 -COOH
90% PEG10 -OH; C, 25% PEG10 -COOH 75% PEG10 -OH; D, 50% PEG10 -COOH 50%
PEG10 -OH
TEM data
Figure 3.10: TEM image of the iron oxide NP (8.6±0.6 nm). Their size was determined
by ImageJ software.
Chapter 4
Two-step directional surface

modification via protected
siloxanes
Bloemen M, Sutens B, Brullot W, Gils A, Geukens N and Verbiest T

Two-step directional surface modification of iron oxide nanoparticles via protected
siloxanes
ChemPlusChem 2014; 80: 50-53
B. Sutens helped with the click chemistry reactions and particle functionalization. Dr.
W. Brullot performed the nanoparticle multilayer experiments and interpreted the
results.
63
64 TWO-STEP DIRECTIONAL SURFACE MODIFICATION VIA PROTECTED SILOXANES
Abstract
Successful surface modification of iron oxide NP is crucial for their usage in applications.
However, ligand exchange methods introducing siloxanes have a few drawbacks. We
herein present a novel approach for surface modification of iron oxide NP by making
use of siloxanes, synthesized by thiol-ene click chemistry, with protected functional
groups. Afterwards the ligands were deprotected to liberate their functionality. This
approach solves the issues related to ligand orientation and colloidal stability during
surface modification.
4.1 Introduction
There is an increased interest in the development of iron oxide NP with customized
surface functionalization, which can be used in magnetic storage devices, ferrofluids,
drug delivery, contrast agents or hyperthermia applications.26,104,109,171–174 Multiple
protocols have already been published for obtaining high quality monodisperse spherical
NP.12,14,175,176 These particles are often coated with apolar alkyl chains, like oleic
acid or oleylamine, that serve as surfactants during the synthesis. However, for most
applications they have to be replaced by other ligands, which make the NP dispersible
in water or introduce functional groups. This so-called ligand exchange process can
be performed with different molecules. Polymers and phospholipid double layers
are often used, since they are very versatile and can be easily introduced.144,177
Catechol-containing molecules are another option, but their synthesis is rather
complicated.55,178,179 Moreover, these ligands are not covalently coupled to the surface.
Valuable alternatives are silica or siloxane shells that are very robust and can be
well defined. Since siloxane molecules can have functional groups, they are often the
molecules of choice for functionalization of metal oxide surfaces.120,127,147,180 However,
these functionalities induce a few issues for iron oxide NP surface modification. Firstly,
the colloidal stability of the NP during the modification process is crucial; if NP
precipitate from the reaction solution this will have a negative effect on the coating
procedure’s performance. Secondly, multiple popular functional groups (carboxylic
acids, amines, . . . ) can also interact with the surface of the NP itself. This way, the
molecule can be oriented upside-down, with the siloxane groups pointing outwards,
which results in crosslinking and excessive layer thickness.181–183 Another challenging
property of siloxanes is their tendency to crosslink through the presence of small
amounts of water or prolonged exposure to heat. This requires the reactions to be
limited in time and influences the available workup methods. Tucker-Schwartz et
al. reported the synthesis of siloxanes by using thiol-ene click chemistry, which can
produce complex molecules on a large scale with minimal impurities.148
In this chapter, we made use of thiol-ene click chemistry to synthesize various different
siloxanes with protected functional groups. These were introduced onto the surface
of iron oxide NP and deprotected later on. This novel approach solves the issues of
wrong ligand orientations and colloidal stability issues. By adopting the thiol-ene
click reactions; we were able to synthesize various siloxanes that are not commercially
available, without elaborate organic synthesis steps. Initially apolar NP could be
dispersed in an aqueous phase after coating with protected polar siloxanes and
subsequent deprotection of their functional groups. The presented concept is generic
and is in no way limited to the ligands that are discussed in this manuscript.
Finally, the amine functionalized NP were used as a building block in a layer-by-layer

synthesis of nanoparticle multilayers to prove their efficiency in these complex systems.
Such magnetic-plasmonic nanoparticle multilayers can be used for fundamental research
to unravel interactions at the nanoscale (magnetoplasmonics) and for applications
in sensing, optical switching and designing optical components. To be used for such
research, nanocomposites need to have large nanoparticle filling fractions, a minimum
of potentially disturbing organic content and partial transparency.184 Layer-by-layer
nanocomposites are mainly synthesized using polymers or polyelectrolytes as ‘glue’
interlayers between nanoparticle layers, which inherently limits nanoparticle filling
fractions and can introduce unwanted noise. Recently, a novel synthesis method for such
nanoparticle multilayers was developed that uses short bifunctional molecular linkers to
connect the subsequent nanoparticle layers.184 No polymers or polyelectrolytes are used
in this method. To produce gold-magnetite nanoparticle multilayers, aminosiloxanes
are ideal molecules as they can strongly bind to gold through the amino functionality
and to magnetite through the siloxane moiety.
Materials
Allyl α-D-galactopyranoside (97%), 4-dimethylaminopyridine (DMAP, >99%), tert-

butyl acrylate (98%), 2,2-dimethoxy-2-phenylacetophenone (DMPAP, 99%), tert-butyl
N-allylcarbamate (98%), acrolein dimethyl acetal (98%), mercaptopropyltrimethoxysi-
lane (95%), trimethoxy(7-octen-1-yl)silane (80%, technical grade) and trifluoroacetic
acid (TFA, 99%) were purchased from Sigma Aldrich. S-allyl thiopropionate (99%)
was bought from ABCR GmbH & Co. 1-thioacetic acid (98%) was obtained from
Acros Organics. Triethylamine (99%) and tetrabutyl ammonium hydrogen sulphate
(98%) were purchased from Janssen Chimica. Acetic anhydride (pro analysis) and
sodium hydroxide were obtained from Riedel de Haën. All ultrasonication steps were
performed in a Branson 5510 sonicator bath. FTIR spectra were measured using a
Bruker Alpha FT-IR spectrometer equipped with a Platinum ATR module. CHN
data were obtained by a CE Instruments EA-1110 elemental analyzer.
Acetylation of allyl α-D-galactopyranoside

The acetylation of the sugar starts with adding allyl α-D-galactopyranoside (440 mg,
2 mmol) to a mixture of 4-dimethylaminopyridine (12.2 mg, 0.1 mmol) and acetic
anhydride (3.78 mL, 40 mmol). This solution was stirred overnight at a temperature of
80°C. The resulting mixture was added to an excess dichloromethane in a separatory
funnel and washed three times with 40 mL MilliQ water. Afterwards, the solution
was washed with 50 mL of 1M NaOH and potassium carbonate until the pH was
neutral. The resulting solution was dried with magnesium sulfate and filtered. The
product (620 mg, 83 %) was obtained after removing the dichloromethane under
reduced pressure. 1 H NMR (300 MHz, CDCl3 ): δ (ppm) 2-2.2 (s, 12H), 4.15-4.25 (m,
5H), 5.1-5.5 (m, 6H), 5.8-6.0 (m-1H).
The synthesis of the functionalized trimethoxysiloxanes was done in a UV chamber,

equipped with 3 LEDs (365 nm, output power 200 mW), following the method
published by Tucker-Schwartz et al.148 All products were used in the subsequent
nanoparticle functionalization protocol without further purification.
Siloxane-acrylate (1)
For the synthesis of the siloxane-acrylate, tert-butyl acrylate (73.24 µL, 0.5 mmol)
was added to 3-mercaptopropyltrimethoxysilane (93 µL, 0.5 mmol), 2,2-dimethoxy-
2-phenylacetophenone (12.8 mg, 0.05 mmol) and 0.5 mL chloroform. This mixture
was placed next to a UV-light for 1 hour on top of a stirring plate. 1 H NMR (300
MHz, CDCl3 ): δ (ppm) 0.78 (t, 2H), 1.46 (s, 9H), 1.59 (m, 2H), 2.5 (t, 2H), 2.6 (t,
2H), 2.75 (t, 2H), 3.57 (s, 9H).
Siloxane-carbamate (2)
Tert-butyl N-allylcarbamate (78,6 mg, 0.5 mmol) was added to a mixture

of 3-mercaptopropyltrimethoxysilane (93 µL, 0.5 mmol) and 2,2-dimethoxy-2-
phenylacetophenone (12.8 mg, 0.05 mmol) in 0.5 mL chloroform. This mixture
was placed next to a UV-light for 1 hour on top of a stirring plate. 1 H NMR (300
MHz, CDCl3 ): δ (ppm) 0.75 (t, 2H), 1.49 (s, 9H), 1.65 (m, 2H), 1.74 (m, 2H), 2.55 (t,
4H), 3.2 (t, 2H), 3.57 (s, 9H).
Siloxane-acrolein (3)
The synthesis of the siloxane-acrolein starts with adding acrolein dimethyl acetal (59.24
µL, 0.5 mmol) to 3-mercaptopropyltrimethoxysilane (93 µL, 0.5 mmol), 2,2-dimethoxy-
2-phenylacetophenone (6.4 mg, 0.05 mmol) and 0.5 mL chloroform. Secondly, the
mixture was placed next to a UV-light for 1 hour on top of a stirring plate. 1 H NMR
(300 MHz, CDCl3 ): δ (ppm) 0.76 (t, 2H), 1.7 (m, 2H), 1.89 (q, 2H), 2.56 (t, 4H), 3.34
(s, 6H), 3.57 (s, 9H), 4.5 (t, 1H).
Siloxane-thiol-propionate (4)
S-allyl thiopropionate (67.5 µL, 0.5 mmol) was added to a mixture of 3-mercaptopropyl
trimethoxysilane (93 µL, 0.5 mmol) and 2,2-dimethoxy-2-phenylacetophenone (6.4 mg,
0.025 mmol) in 0.5 mL chloroform. This mixture was placed next to a UV-light for 1
hour on top of a stirring plate. 1 H NMR (300 MHz, CDCl3 ): δ (ppm) 0.76 (t, 2H),
1.18 (t, 3H), 1.7 (m, 2H), 1.85 (m, 2H), 2.56 (m, 6H), 2.95 (t, 2H), 3.57 (s, 9H).
Siloxane-C8-thiol-acetate (5)
1-thioacetic acid (35.23 µL, 0.5 mmol) was added to trimethoxy(7-octen-1-yl)silane

(125.35 µL, 0.5 mmol) and 2,2-dimethoxy-2-phenylacetophenone (6.4 mg, 0.025 mmol)
in 0,5 ml chloroform. This mixture was placed next to UV-light for 1 hour on top of a
stirring plate. 1 H NMR (300 MHz, CDCl3 ): δ (ppm) 0.64 (t, 2H), 1.3-1.5 (m, 10H),
1.58 (m, 2H), 2.36 (s, 2H), 2.85 (t, 2H), 3.57 (s, 9H).
Siloxane-acetylated-sugar (6)
Tetra-acetyl allyl α-D-galactopyranoside (194 mg, 0.5 mmol) was added to a

mixture of 3-mercaptopropyltrimethoxysilane (93 µL, 0,5 mmol) and 2,2-dimethoxy-2-
phenylacetophenone (6.4 mg, 0.025 mmol) in 0.5 ml chloroform. This mixture was
placed next to a UV-light for 1 hour on top of a stirring plate. 1 H NMR (300 MHz,
CDCl3 ): δ (ppm) 0.76 (t, 2H), 1.72 (m, 2H), 1.85 (m, 2H), 2-2.2 (s, 12H), 2.56 (t, 2H),
2.6 (t, 2H), 3.52 (t, 2H), 3.57 (s, 9H), 3.8 (m, 1H), 4.15+4.25 (t, 2H), 5.1-5.5 (m, 4H).
Nanoparticle synthesis and functionalization
The NP were synthesized and functionalized, following the methods described in

chapter 2 and 3. The functionalized NP were dispersed in the appropriate solvent
(chloroform or THF, for deprotection) with a concentration of 5 mg/mL.
Deprotection of the functional group
The deprotection of the functionalized particles can be divided in two different methods.
The first method includes the deprotection of the tert-butyl acrylate (1), tert-butyl
N-allylcarbamate (2) and the acrolein dimethyl acetal (3) with a 1:1 mixture of
trifluoroacetic acid (TFA) and dichloromethane. Secondly, grinded sodium hydroxide
and tetrabutyl ammonium hydrogen sulphate were used to deprotect the thiopropionate
(4), the thioacetate (5) and the tetra-acetyl α-D-galactopyranoside (6), as reported
by Crouch et al. with minor modifications.185
TFA/CH2 Cl2
5 mg functionalized particles (in CHCl3 , 5 mg/mL) were added to a flask with 1.5
mL dichloromethane and 2.5 mL trifluoroacetic acid. The flask was put on a shaker
overnight at room temperature. The particles were then magnetically separated and
washed three times with acetone. Finally the NP were stored in MilliQ water, with a
concentration of 5mg/mL.
Grinded NaOH/Bu4 NHSO4
For the deprotection with grinded NaOH and Bu4 NHSO4 , 1 mg functionalized particles
(in THF, 5 mg/mL) were mixed with grinded NaOH (16 mg, 0.4 mmol), Bu4 NHSO4
(3.2 mg, 0.08 mmol) and 1 mL tetrahydrofuran. The resulting solution was placed
in the ultrasonication bath for 5 hours. Afterwards, the particles were magnetically
separated and washed three times with THF and acetone. Finally the NP were stored
in MilliQ water, with a concentration of 5 mg/mL.
Layer-by-layer
Nanoparticle multilayers incorporating gold and amine-functionalized magnetite NP

on glass substrates were synthesized as published elsewhere.184 Briefly stated, first gold
NP with a size of approximately 9.2 nm and a strong plasmon resonance centered on
530 nm were produced using an aqueous citrate reduction procedure.186 Synthesis of
the nanoparticle multilayers started by cleaning glass microscope slides (16x16 mm2)
with NoChromix cleaning solution (7 g/100 mL NoChromix in MilliQ water, add 100
mL sulfuric acid 97%) for 1 h. These cleaned substrates were then functionalized
with amino propyl trimethoxy silane (1 vol% in MeOH, 1 h) and rinsed with MeOH
and water. A first metal nanoparticle layer was added by putting the functionalized
substrate in 10 mL of a metal NP dispersion while shaking (350 rpm) for 1.5 h and
rinsing with water and MeOH afterwards. By putting the sample in 10 mL of a 0.4
mg/mL dispersion of functionalized iron oxide NP in MeOH for 1 h, a layer of these
NP could be added on top of the gold nanoparticle layer. After rinsing the sample
with methanol and water, adding extra nanoparticle layers was possible by repeating
the previous steps.
Methodology
To rule out unwanted interactions between the functional group of siloxane molecules
and the surface of the iron oxide NP, we developed a strategy that involves the synthesis
Figure 4.1: Six different siloxanes, with protected groups, were synthesized by click
chemistry for subsequent introduction on the surface of iron oxide NP. Later on, the
ligands were deprotected to make the desired functional group available. This results
in free carboxylic acids (1), amines (2), aldehydes (3), thiols (4, 5) or sugars (6) on
the surface of the nanoparticle. All experimental information can be found in the
supporting information at the end of this chapter.
of protected siloxanes and their subsequent deprotection after surface functionalization

(see Figure 4.1). This method has a three major advantages: first, it protects functional
groups that can interact directly with the surface, like carboxylic acids (1) or amines
(2).181–183,187,188 Hence, only the siloxane side of the molecule can interact with the
surface of the nanoparticle, which ensures proper directionality. Secondly, groups
that would not survive the functionalization procedure in the ultrasonicator (such as
aldehydes) can be introduced (3). They would hydrolyze, by the presence of water at
elevated temperatures. Thirdly, it is possible to introduce functionalities that provide
insufficient colloidal stability (due to polarity) during the functionalization step, like
thiols (4, 5) or sugars (6). The ligand exchange reaction was performed in toluene,
since this solvent has the perfect water content (0.3-0.4%) for silanization reactions.189
Unfortunately, this method is limited to siloxanes that dissolve properly in toluene
and provide sufficient colloidal stability to the nanoparticle. Our approach eliminates
the restrictions due to polarity of the ligands, because most protective groups, such
as tert-butyl esters, are apolar (see Figure 4.1). The apolar groups stabilize the
particle in toluene, which improves the functionalization by keeping the nanoparticle
in dispersion (avoiding precipitation).
To our knowledge, none of the siloxanes shown in Figure 4.1 is commercially available,
so we opted for the straightforward approach of thiol-ene click chemistry to synthesize
the different compounds. Because this click reaction is fast and gives high yields, side
reactions of the trimethoxy groups are limited.151 Multiple different molecules with
an allyl functionality are readily available since they are vinyl-monomers, which are
often used in polymerization reactions.
Figure 4.2: The transfer from the organic (lower phase) to the aqueous phase (upper
phase) is clearly visible after the deprotection of the ligands on the NP.
Deprotection of all ligands was achieved by typical protocols with minor modifications,
more specific the duration of the reaction.190 Similar to dendrimer chemistry, long
reaction times are necessary to ensure that full deprotection has occurred.191 This is
caused by the limited rotational freedom of the ligands and the corresponding steric
hindrance. Therefore, we prolonged the TFA and sodium hydroxide deprotection
protocols to 16 and 5 hours respectively. The deprotection reaction for compounds 4, 5,
6 was performed in an ultrasonicator, which raises the temperature to approximately
55°C and improves the solubility of all reagents. Figure 4.2 shows the phase transfer
that occurs after the ligands are deprotected. After the initial ligand exchange, the NP
are apolar and hence they only disperse in the bottom organic phase. The successful
introduction of the ligands on the surface is supported by FTIR measurements (data
in supporting information at the end of this chapter). The characteristic peaks of
the different functional groups are clearly visible in the spectra of the NP samples.
After deprotection, the NP become polar and can be dispersed in the upper aqueous
phase. Since no particles are visible in the organic phase after deprotection, we
can conclude that the polarity of the particles has switched from apolar to polar,
indicating a successful deprotection. CHN analyses also supported the deprotection
of the functional groups (data in supporting information at the end of this chapter,
Table 4.1). However this approach is no guarantee for colloidal stability of the NP
in the aqueous phase. Aldehydes (3) and thiols (4, 5) possess no charges (or can
form disulfide bridges) in a pH range between 4 and 10 and they are therefore very
susceptible to precipitation. This is visible in Figure 4.2 as a slightly turbid aqueous
phase.
NP with charged carboxylic acids and amines are often used in nanotoxicity experiments
to test the effect of the surface charges on the cellular uptake.17 Thiol-coated NP, on
the other hand, are valuable for their interaction with noble metals, like gold or silver.
Sugar molecules are biocompatible; which makes these sugar-coated NP interesting for
biomedical applications. Unlike typical sugar-containing polymers like dextran, the
presented approach results in a very thin surface layer, with a directional bonding.192
Figure 4.3: Due to plasmon hybridization the plasmon resonance broadens and
red-shifts with each added nanoparticle double layer (NDL). The absorbance at
maximal intensity increases linearly with the number of added nanoparticle layers,
demonstrating that the layer-by-layer procedure was successful (inset).
Nanoparticle multilayers
To demonstrate their efficiency in more complex systems, amine-functionalized iron

oxide NP, together with gold NP were incorporated in nanoparticle multilayers. To
produce gold-magnetite nanoparticle multilayers, amino-siloxanes are ideal molecules
as they can strongly bind to gold through the amino functionality and to magnetite
through the siloxane moiety. Such multilayers, in which nanoparticle layers are layer-
by-layer constructed through short molecular linkers, have large nanoparticle filling
fractions and interesting linear, nonlinear and magneto-optical properties.184 They
can be applied for research into e.g. plasmon influenced (magneto-) optical effects and
active plasmonics through magnetism.
When adding multiple gold and magnetite nanoparticle layers, the plasmon resonance
of gold NP broadens and red-shifts due to plasmon hybridization (see Figure 4.3).184
The linear increase of the absorbance at maximal intensity demonstrates the success
of the layer-by-layer process. In contrast to previously used amine-functionalized iron
oxide NP, the NP presented in this work are monodisperse and have a well defined
surface functionalization. This is an asset for layer-by-layer processes as it might
improve packing of subsequent nanoparticle multilayers.
CONCLUSIONS 73
4.4 Conclusions
We developed a new approach for efficient functionalization of iron oxide NP by using
protected siloxanes as ligands. Six different protected ligands, with a large variety of
functional groups (including carboxylic acids, amines, thiols, aldehydes and sugars),
were prepared by thiol-ene click chemistry and subsequently introduced onto the
surface of the NP by a ligand exchange reaction. By protecting the functional group,
a directional bonding of the ligand on the iron oxide surface was ensured. Moreover,
this approach allows the introduction of very polar or reactive functional groups.
After deprotection of these groups, all samples were dispersible in aqueous solutions,
indicating a successful surface functionalization. To highlight the potential of the NP,
they were incorporated in nanoparticle multilayers, a very promising material where
well-defined surface properties are crucial.
FTIR spectra of synthesized compounds
This section combines all FTIR spectra of the synthesized protected siloxanes. These
spectra serve as a reference for subsequent nanoparticle functionalization.
Figure 4.4: Spectrum of allyl α-D-galactopyranoside and the acetylated end-product.

The broad OH vibrations at 3200-3500 cm−1 are clearly replaced by the ester peak at
1720 cm−1 .
Figure 4.5: Spectrum of the tert-butyl acrylate and the siloxane-acrylate (1) end-
product. Characteristic peaks of this product are the Si-O vibration at 1100 cm−1
and the ester peak at 1720 cm−1 .
Figure 4.6: Spectrum of the tert-butyl N-allylcarbamate and the siloxane-carbamate

(2) end-product. Characteristic peaks of this product are the Si-O vibration at 1100
cm−1 , the ester peak at 1720 cm−1 and the amine band at 3100-3500 cm−1 .
Figure 4.7: Spectrum of the acrolein dimethyl acetale and the siloxane-acrolein (3)
end-product.
Figure 4.8: Spectrum of the siloxane-thiol-propionate (4) compound. S-allyl

thiopropionate was not measured because of its odor and the related safety issues.
Characteristic peaks of this product are the Si-O vibration at 1100 cm−1 and the ester
peak at 1720 cm−1 .
Figure 4.9: Spectrum of the siloxane-C8-thiol-acetate (5) compound. 1-thioacetic acid

was not measured because of its odor and the related safety issues. Characteristic
peaks of this product are the Si-O vibration at 1100 cm−1 , the ester peak at 1720
cm−1 and the C-H vibrations at 2920-2930 cm−1 .
Figure 4.10: Spectrum of the tetra-acetyl allyl α-D-galactopyranoside and the siloxane-
sugar (6) end-product. Characteristic peaks of this product are the Si-O vibration at
1100 cm−1 and the ester peak of the four protective acetyl-esters at 1720 cm−1 .
FTIR spectra of functionalized nanoparticles
This section shows all spectra of the protected siloxanes and the NP samples,
functionalized with those protected siloxanes. In all spectra, a substantial overlap
between both spectra can be observed, indicating a successful functionalization of the
NP. Characteristic vibrations of iron oxide can be seen at 580-620 cm−1 (Fe-O).
Figure 4.11: Spectrum of siloxane-acrylate (1) functionalized iron oxide NP.

Figure 4.12: Spectrum of siloxane-carbamate (2) functionalized iron oxide NP.
Figure 4.13: Spectrum of siloxane-acrolein (3) functionalized iron oxide NP.

Figure 4.14: Spectrum of siloxane-thiol-propionate (4) functionalized iron oxide NP.
Figure 4.15: Spectrum of siloxane-C8-thiol-acetate (5) functionalized iron oxide NP.

Figure 4.16: Spectrum of siloxane-acetylated-sugar (6) functionalized iron oxide NP.
1
H-NMR spectra of synthesized compounds
In this section, all 1 H-NMR spectra of the synthesized protected siloxanes are shown.
The peaks of the double bond (5.0-6.5 ppm) disappear after the click chemistry,
indicating a successful reaction. Due to safety restrictions, the precursors for the
protected thiol siloxanes (4, 5) were not measured. We refer the reader to a
manufacturer’s website, where these spectra can be found.
S-allyl thiopropionate: http://www.sigmaaldrich.com/catalog/product/aldrich/w332909
1-thioacetic acid: http://www.sigmaaldrich.com/catalog/product/aldrich/t30805

CHN analysis derived data
Table 4.1: Comparison of the weight percentage of surface groups on the functionalized
nanoparticle. After deprotection, the overall percentage lowers, since carbon and
hydrogen atoms are removed in this process. The thiol ligands (4, 5) show an opposite
trend, but this can be attributed to the formation of thiolate-tetrabutylammonium
ion pairs (caused by deprotection process and sample preparation), hence largely
increasing the carbon content. Even though a similar deprotection protocol is followed
for the sugar ligand (6), this effect is absent here since no ion pairs can be formed.
Sample name (Fe3 O4 -ligand) Surface group w% (protected) Surface group w% (deprotected)
Fe3 O4 -1 24.98 9.12
Fe3 O4 -2 43.30 12.66
Fe3 O4 -3 20.68 15.59
Fe3 O4 -4 24.00 32.96
Fe3 O4 -5 23.99 37.15
Fe3 O4 -6 62.34 38.32
Chapter 5
Design of nanoparticles for

efficient magnetic isolation of
aquatic pathogens
Selective magnetic purification by nanoparticles

Y Y
Y Y Y
YY
Y
Y
YY
Y
YY
Y
Y Y
YY
YY
N
Y
YY
Y
Y Y Y Y
Y Y
YY
Y
YY
Y
Y
Y
Y Y
YY
Y
YY
S
Y
Y Y
Y Y
YY
Y
YY
Y
Y
Y
Y
Y
YY
Y
YY
Y
Y Y
Y Y
YY
Y
YY
Y
Y Y Y
YY
Y
Y Y
YY
Y
Y
Y
YY
Y
Antibody-modified iron oxide nanoparticles for efficient magnetic isolation and flow
cytometric determination of L. pneumophila
Bloemen M, Denis C, Peeters M, De Meester L, Gils A, Geukens N and Verbiest T
Microchimica Acta 2015; accepted
C. Denis prepared the cell cultures and performed the DNA extraction and qPCR
experiments. M. Peeters provided assistance during the antibody purification process.
85
86 DESIGN OF NANOPARTICLES FOR EFFICIENT MAGNETIC ISOLATION OF AQUATIC PATHOGENS
Abstract
Multifunctional nanoparticles are becoming increasingly important since they combine
interesting properties like fluorescence, magnetism or plasmonic effects. We designed
a superparamagnetic nanoparticle that is capable of selectively separating targeted
bacteria from an aqueous solution. By covalently functionalizing the surface of
the particle with a heterobifunctional polyethylene glycol ligand, functional groups
for antibody coupling were introduced. Polyclonal antibodies, targeting Legionella
pneumophila bacteria, were coupled to the particles and these bioconjugates were
tested for their magnetic separation properties. Flow cytometry measurements showed
high efficiency in this regard. Furthermore the targeted bacteria were purified from
a complex mixture of microorganisms. The results indicate that the presented
multifunctional nanoparticles are capable of selectively attracting pathogens from
a complex mixture, with a high efficiency, which can be of great interest for pre-
concentration protocols in water quality monitoring.
5.1 Introduction
Iron oxide NP have become popular since they were applied in various fields like
MRI, hyperthermia or drug delivery carriers.45,171,173,193,194 For most applications,
magnetite (Fe3 O4 ) is the preferred species of iron oxide, because of its interesting
magnetic properties.26 If the NP size ranges between 5 and 25 nm, they exhibit
superparamagnetic behavior. This is highly valuable because it combines the large
magnetic moment of a ferromagnetic material, with the convenient handling of
a paramagnet, which has no net moment without an external magnetic field.195
Consequently these NP are excellent materials for performing magnetic separation of
target ions, molecules or cells from solution.196–198 However two more properties of
the material are crucial for these applications: their colloidal stability in the medium
and their selectivity towards the target. To achieve the first requirement, multiple
solutions have been reported. Typically the surface of the NP is coated with a
polymer, silica or organic ligands, to introduce functionalities and improve its colloidal
properties.41,144,179,199 Recently we reported a strategy to efficiently functionalize the
surface of iron oxide NP with a heterobifunctional PEG ligand and couple Ab to
this layer.200 Depending on the target: organic ligands, Ab or nanobodies with a
high affinity can be introduced to acquire selective targeting.108,198,201–205 Xu et al.
reported that these NP-Ab bioconjugates can help in the separation of cancer cells
from blood.198 However, to our knowledge, no references are present in literature of
studies describing a full cell separation study with multiple cell detection techniques.
In this manuscript we focus on the magnetic separation of Legionella pneumophila
serogroup 1 bacteria from aqueous solutions. These gram-negative bacteria are a wide-
spread problem in cooling towers, air-conditioning systems, fountains and showers.206
They are known to cause the Legionnaires’ disease or legionellosis, a serious form of
pneumonia.207 Their detection is mandatory in all publicly accessible water system,
like swimming pools, but also in cooling circuits and wastewater. Different methods,
such as filtration, centrifugation or immunomagnetic isolation by microparticles, are

currently available to separate these bacteria from their aqueous environment for
subsequent quantification.98,208–210 However all these methods have some specific
drawbacks. Filtration has a profound influence on cell viability and centrifugation is
less appropriate for large volumes. Moreover these methods are not target specific;
so all organisms are retained, including possible inhibitors that might complicate
further quantification. Immunomagnetic isolation by microparticles is target specific,
but the efficiency is low and cell damage might occur due to the localized excessive
magnetic forces.210 After isolation, quantification of the bacteria can be performed
by traditional plate counting or quantitative real-time polymerase chain reactions
(qPCR). Even though the culturing method is still considered as the standard, it is very
time-consuming and inconvenient. qPCR is substantially faster but is very susceptible
to inhibitors, such as metal salts, surfactants or polysaccharides.211 In comparison
to microparticles, the proposed NP have a more than 100 times larger surface to
volume ratio, which allows binding more Ab to the nanoparticle’s surface. Moreover
more particles can interact with the cells, resulting in a larger net magnetic moment,
hence lowering separation time. On top of that, the magnetic forces are more evenly
spread across the cell’s surface, which will increase the amount of intact organisms
that can be collected. In this manuscript we report the design of a functionalized iron
oxide (Fe3 O4 ) nanoparticle with customized PEG ligands and Ab. These particles
were tested extensively on their ability to specifically attract and separate Legionella
pneumophila serogroup 1 bacteria from an aqueous solution. To underline the efficiency,
the targeted bacteria were also separated from a mixture of bacteria (L. pneumophila
and Escherichia coli). Negative control experiments reported in literature often make
use of NP without a targeting ligand (f.i. an antibody). The recent scientific discussions
about protein coronas have shown that the interaction between the nanoparticle’s
surface and another entity is largely dependent on its coating.159 Therefor we opted
for a negative control composed of iron oxide NP, conjugated with a non-Legionella
targeting antibody (targeting murine PAI-1).
Materials
Allyl-PEG10 -OH was obtained from Polysciences, Inc. DMPAP (99%), DMAP
(99%), EDC, succinic anhydride (99%) and mercaptopropyltrimethoxysilane (95%)
were purchased from Sigma Aldrich. NHS (98+%) was purchased from Alfa Aesar.
Triethylamine was obtained from Janssen Chimica. MES was purchased at Fluka. All
ultrasonication steps were performed in a Branson 5510 sonicator bath.
Antibody purification
“New Zealand White” rabbits were immunized with Legionella pneumophila serogroup 1
and murine PAI-1.212 After collecting the serum, the full IgG fraction was purified using
protein A ProSep beads.213 These polyclonal fractions were used without additional
enrichment towards the target protein and denoted as Ab that target L. pneumophila
(pAb Leg) and Ab that target murine PAI-1 (pAb mur PAI-1; negative control).
Synthesis of the ligands, nanoparticles and functionalization

procedure
The heterobifunctional PEG ligand was synthesized as described in Chapter 3.200

Allyl-PEG10 -OH (1.1 eq, 4.00 mmol, 1.992 g) was mixed with succinic anhydride (1.1
eq, 4.40 mmol, 440 mg) and 4-dimethylaminopyridine (DMAP) (0.02 eq, 0.08 mmol,
9.7 mg). This mixture was stirred and heated to 50°C for 16 hours. The product (2)
was purified twice by precipitation in cold diethyl ether, centrifugation and drying
in vacuum. 1 H NMR (300 MHz, CDCl3 ): δ (ppm) 2.65 (s, 4H), 3.55–3.75 (m, 38H),
4.02 (d, 2H), 4.26 (t, 2H), 5.15–5.32 (m, 2H), 5.85–5.95 (m, 1H). 13 C NMR (75 MHz,
CDCl3 ): δ (ppm) 29.2, 29.5, 63.8, 68.9, 69.3, 70.5, 72.2, 117.1, 134.7, 172.1. MS
(chemical ionization): m/z = 499, 101.
To introduce the siloxane onto the heterobifunctional PEG molecule, allyl-terminated

PEG (1 mmol in total, mixture of modified (0.1 mmol) and unmodified (0.9 mmol))
was mixed with (3-mercaptopropyl) trimethoxysilane (1 eq, 1 mmol, 185.7 µL) and
2,2-dimethoxy-2-phenylacetophenone (DMPAP, 0.05 eq, 0.05 mmol, 12.8 mg). This
mixture was stirred during 1 hour in a UV reactor, equipped with 3 LEDs (365 nm,
output power 200 mW). The product (3) was used without further purification. 1 H
NMR (300 MHz, CDCl3 ): δ (ppm) 0.76 (t, 2H), 1.70 (m, 2H), 1.85(m, 2H), 2.55 (m,
4H), 2.64 (s, 4H), 3.57 (s, 9H), 3.55–3.75 (m, 40H), 4.26 (t, 2H). The synthesis of
iron oxide NP and coating their surface with siloxanes was performed as described
in Chapter 2 and 3.200,214 In general, 1 mmol of siloxanes was mixed with 100 mg
of Fe3 O4 NP in 50 mL of toluene. To this mixture triethylamine (2.5 mL) and 50
µL of water were added. The solution was placed in an ultrasonication bath for 5
hours, after which 50 mL of heptane was added to precipitate the particles (dispersion
becomes turbid). Afterwards, they were attracted magnetically and washed 3 times
with acetone. Finally the particles were dried in vacuum and dispersed in MilliQ water
(with a concentration up to 20 mg/mL).
Antibody coupling
The concentrated nanoparticle solution was diluted in 50 mM MES buffer, pH 5.5, to

reach a final concentration of 3 mg/mL. 0.75 mg EDC and 0.75 mg NHS was added
to 1 mL of this solution and shaken for 20 minutes to activate the carboxylic acids.
The antibodies (75 µg) were diluted in 2 mL of the same MES buffer after which
both solutions were mixed and shaken for 1 hour. To separate the particles from the
solution, a Miltenyi Biotech MS magnetic column was used. After rinsing the column
with MilliQ water, the nanoparticle dispersion was loaded onto the column, which was
placed inside a circular NdFeB magnet. The column was washed 2 times with 1 mL
of sodium phosphate buffer (20 mM, pH 7). To elute the particles, the column was
removed from the magnet and 0.5mL of phosphate buffer and subsequently 0.5 mL of
MilliQ were used as eluent.
Bacterial strains and growth conditions
Recombinant strains of E. coli and L. pneumophila serogroup 1 expressing the red

fluorescent protein (RFP) and green fluorescent protein (GFP) genes respectively,
were used to test strain specificity and efficacy of the iron oxide nanoparticles. The
transformed bacteria were cultivated in the presence of appropriate antibiotics (10 µg
of kanamycin/mL for RFP and 5 µg/mL chloramphenicol for GFP) to ensure plasmid
maintenance. Briefly, Legionella bacteria were cultured by standard procedures at
a temperature of 37°C on buffered yeast extract agar containing α-ketoglutarate
(BCYE-α) supplemented with L-cysteine and ferric pyrophosphate.215 E. coli bacteria
(LMG2092T, RFP labelled and kindly given by Prof. N. Boon, UGent, Belgium) were
grown in Luria broth (LB) medium supplemented with kanamycin (10 mg/mL) and
incubated overnight at 30°C.
Magnetic separation of bacteria
Separate L. pneumophila and E. coli solutions were prepared by spiking Ringer’s

Solution (Oxoid) with scraped bacteria from the culture plates following the preparation
of 10-fold dilution series. In general, 4 different concentrations of Legionella bacteria
(1E7, 1E6, 1E5 and 1E4 cells/mL) were tested in either monospecies suspensions or
mixed suspensions with added E. coli bacteria. In general, 300 µL of each bacteria
solution was mixed with 250 µL of NP solution (0.5 mg NP). Then, this mixture was
placed on a rotator at room temperature for 1 hour and afterwards separated by a
magnetic column. The supernatant was collected and the column was washed with 1
mL of MilliQ water. To elute the NP and bacteria, the column was removed from the
magnet and 1 mL of MilliQ was used as an eluent (see Figure 5.3 B-C).
Flow cytometry
Cell numbers of L. pneumophila and E. coli were determined by flow cytometry using
an Attune®Acoustic Focusing Cytometer (Life Technologies, Gent, Belgium) equipped
with a 488 nm laser, a forward scatter (FSC) diode detector, and a photomultiplier
tube (PMT) SSC detector. The instrument was checked for stable fluidic alignment
using Performance tracking beads (Life Technologies). Bacterial fractions expressing
GFP and RFP fluorescence were detected using, respectively, the BL-1 (530/30 nm)
and BL-3 (640 LP) detector. The different fractions of the separation experiment were
fixed with 4% paraformaldehyde (PFA) and were diluted 100-fold. The collection rate
of the instrument was 25 µL/min and a total of 50 µL was analyzed per sample.
Quantitative PCR
For the specific detection of L. pneumophila, a TaqMan qPCR designed by P. Declerck

and J. Behets was used.216 DNA was extracted using the QIAGEN DNeasy Blood
& Tissue Kit. The primers were based on the macrophage infectivity potentiator
(mip) gene and were: LPQF (5’-TTCATTTGYTGYTCGGTTAAAGC) and LPQR
(5’-AWTGGCTAAAGGCATGCAAGAC). The mip-specific TaqMan probe was 5’-
GCGCCACTCATAG labeled with a 6-carboxyfluorescein (FAM) reporter dye at the
5’ end and a non-fluorescent quencher at the 3’ end. The probe was conjugated to a
minor groove binder (MGB) to improve real-time PCR specificity and sensitivity.
Figure 5.1: The synthesized iron oxide (magnetite) nanoparticles are 8.6±0.6nm in
size and are spherical.
Figure 5.2: The allyl-PEG10 molecule is modified first by an anhydride ring opening
on the hydroxyl group, resulting in a carboxylic acid group. Secondly the allyl group
is used in a thiol-ene click chemistry reaction to attach the siloxane.
Synthesis and characterization of the bioconjugates
The procedure to synthesize the nanoparticles was carefully selected for its ability
to produce particles at a large scale with good monodispersity.12 Even though these
characteristics are not strictly necessary for magnetic separation experiments, it
does improve the reproducibility of surface functionalization and antibody coupling.
The iron oxide nanoparticles (magnetite, Fe3 O4 ) that were prepared are 8.6 nm
wide in diameter, with a narrow size distribution of 0.6 nm (see transmission electron
microscopy data, Figure 5.1). The large scale at which these nanoparticle are produced
(>10 grams) ensures that batch-to-batch differences during experiments can be fully
excluded.
Even though the functionalization of iron oxides is a well-known research topic, multiple
recent advances have been reported.174 We choose siloxane surface chemistry in this
regard, since it provides the particles with several important properties. First of all,
the ligand coating is covalently attached to the surface, which improves its resistance to
the environment, merely extreme pH, heat or high ionic strength. Secondly, the ligand
can be designed with a specific application in mind. In this case, a PEG backbone was
preferred for its excellent solubility in aqueous environments. Moreover, the ligand
was altered to have one functional carboxylic acid end-group, which concentration
on the nanoparticle can be tailored by mixing with unaltered ligands. As shown in
Figure 5.2, the ligand is prepared by a two-step reaction involving an anhydride ring
opening reaction and a thiol-ene click chemistry reaction. The latter is very convenient
for siloxane chemistry since it occurs fast and without notable side reactions. The
product can be used without any further purification, which reduces the chance of
cross-linking and hydrolysis of the siloxane group.
The presence of a carboxylic acid group on the ligand implies that pH will have an
influence on the charge of the functionalized nanoparticle. To reduce this potential
issue, the modified ligand was mixed with unmodified ligands (siloxane-PEG10 ) during
the nanoparticle functionalization procedure. A fixed ratio of 10% modified to 90%
unmodified ligands was maintained throughout the experiments. Since the subsequent
coupling to antibodies will use the present functional groups to form amide bonds,
the influence of pH will be limited, which is necessary to ensure colloidal stability in
complex environments (such as buffer solutions).
After functionalization of the NP’ surface, Ab were covalently coupled onto their
carboxylic acid groups. A well-known EDC-NHS approach was preferred for its
reproducibility and simplicity.83 One major drawback of this method is the non-
directional bonding that occurs. Since a protein contains multiple free amines, the
orientation of the protein can hardly be controlled. Hence a high percentage of the Ab
will lose their activity because of a non-optimal orientation (as shown in Figure 5.3
A). The polyclonal Ab used in this study are the whole IgG fraction of leporine serum,
derived from L. pneumophila immunized rabbits. This IgG usually contains 1 to 5%
of target specific Ab. It is technically possible to purify this further to a monospecific
polyclonal fraction, but this is very time consuming. Moreover, the target species are
bacteria, which makes the purification even more complex. It would require a column
coated with whole L. pneumophila bacteria, not just membrane proteins, to ensure
that no epitopes would be ignored. Therefore, in this study, the full IgG fraction was
used during the coupling experiments, maximizing the range of targeted epitopes. We
argue that the gain of having monospecific polyclonal Ab is small if the bonding is
non-directional anyway. A directional (but much more complex) bonding strategy
could definitely profit from this on the other hand.83,217
Magnetic isolation results
The bioconjugated iron oxide NP were added to monospecies spiked L. pneumophila

solutions (1E7, 1E6, 1E5 and 1E4 cells/mL) and incubated for 1 hour while rotating
the test tubes. No precipitation was observed, which is a strong indication that the
custom functionalized surface is coping well with the complex environment and is
keeping the particle colloidally stable. The dispersion was loaded onto a magnetic
column (inside a circular magnet) and the supernatant was collected. After a washing
step, the column was removed from the magnet and eluted to collect the nanoparticle-
bacteria complexes. This fraction was split to perform all different characterization
techniques.
Flow cytometry was selected as the main characterization technique because of its
ability to quantify fluorescently labeled bacteria with high accuracy. Since the L.
pneumophila strain was labeled with a GFP marker, detection and quantification was
straightforward. The results show that as expected the type of antibody, coupled
to the nanoparticle, has a profound influence on the magnetic separation behavior.
Table 5.1 shows the amount of bacteria found in the different fractions that were
magnetically separated from an initial solution of 5E6 bacteria per milliliter. The
presence of Ab targeting the bacteria clearly has a positive influence on the separation
capabilities of the NP. In the supernatant of the negative control (pAb mur PAI-1),
the concentration of cells is more than 150 times higher. While the opposite can be
seen in the NP-cell fraction, where the negative control performs more than 10 times
Figure 5.3: Schematic overview of the cell separation procedure. First, the
functionalized NP were conjugated to the corresponding Ab via an amide bond,
induced by EDC-NHS coupling chemistry (A). Secondly, L. pneumophila solutions were
prepared (B), which were brought into contact with the NP and magnetically separated
(C). Finally, the different collected fractions were characterized by flowcytometry and
qPCR (D).
worse. Moreover, the final fraction is underestimated by the flow cytometer, since
small clusters of bacteria are formed after interaction with NP. Hence multiple cells
are shown as one event only. Experimental data revealed that the underestimation is
approximately a factor 2-3 (see supporting information, Figure 5.5). The background
of the flow cytometer, used in these experiments, is approximately 3E3 cells/ml; this
number was not subtracted from the experimental values, since this was a systematic
error and the influence on the end results is minimal. However, due to this substantial
background signal, the limit of detection is limited to 104 cells/mL. Using other
quantification techniques can drastically improve this limit of detection. An overview
of different purification methods and their respective recoveries can be found in the
Supporting information, Table 5.3. These results show that the NP are capable of
selectively capturing target bacteria from an aqueous solution, even though the Ab
are coupled non-directionally and are not monospecific.
To further underline the capabilities of the functionalized NP, a separation experiment

was performed on a mixed set of bacteria. This would give information about the
specificity of the NP and the behavior of the negative control. In an experiment
with only one bacterial species present, the negative control will always detect this
bacterium, albeit aspecifically, since no other targets are present. By carefully selecting
Figure 5.4: The separation capabilities of the nanoparticles are still active in a mixed
bacteria environment. L. pneumophila bacteria (GFP-labeled) were mixed with E. coli
bacteria (RFP-labeled) at a 65/35 ratio (1.4E5/7.4E4 cells/mL). The flow cytometry
measurements show that the RFP-labeled E. coli bacteria remain in the supernatant
fraction of the pAb Leg coated NP (upper row), while the L. pneumophila bacteria
are present in the NP-cell fraction. The negative control NP (pAb mur PAI-1, bottom
row) show a different behavior: both species are visible in the supernatant fraction,
with only minor presence in the NP-cell fraction.
Table 5.1: The type of antibody has a profound influence on the magnetic separation
behavior. This dataset was measured by flow cytometry, based on 3 replicates. A
large amount of the bacteria is present in the supernatant (SN) of the negative control
(NP-pAb mur PAI-1), while the final fraction (NP-cell elution) in comparison only
contains a few percent. The opposite can be seen for the bioconjugates that target L.
pneumophila (NP-pAb Leg): a small number of cells can be found in the supernatant,
while a large number is present in the final fraction. The ratio of the two types of
conjugates is presented to clarify the difference. The full data set can be found in the
supporting information (Table 5.2).
Collected cells/mL Supernatant Washing step NP-cell elution

NP-pAb Leg 1.1E4 (2.3%) 5.0E3 (1.1%) 4.4E5 (96.6%)
NP-pAb mur PAI-1 1.9E6 (82.7%) 3.5E5 (15.5%) 4.0E4 (1.1%)
Ratio 0.0057 0.014 11
the two different bacteria (L. pneumophila GFP-labeled and E. coli RFP-labeled),
we were able to discriminate them in the FCM plots and investigate the overall
performance. Figure 5.4 summarizes the data of these experiments. The FCM plots of
the supernatants are shown in the left column, while the NP-cell fractions are shown
in the right column. The washing steps were omitted for clarity. In the nanoparticle
samples coated with pAb Leg (upper row), only a minor presence of L. pneumophila
in the supernatant is shown, while a high concentration is present in the NP-cell
fraction. E. coli on the other hand stays in the supernatant and is hardly present
in the NP-cell fraction. We noticed a small leak of green fluorescence into the red
fluorescence detector, which accounts for a significant portion of the visible dots in the
red ellipse. Keeping this background in mind, more than 90 percent of the bacteria
were successfully separated. Nevertheless, this clearly underlines the target-specific
magnetic separation that is occurring. In the negative control samples (bottom row),
both bacterial species are mainly present in the supernatant, but only minimally in the
NP-cell fraction. We can therefore conclude that the negative control nanoparticles
only interact aspecifically with the bacteria, as expected.
The presented nanoparticle platform is capable of selectively attracting bacterial

species from a complex mixture. Moreover the ligand design is straightforward and
thanks to the usage of siloxanes, the ligand is covalently bound to the surface of
the NP. Even though the coupling of Ab is non-directional, this did not hamper the
properties of the NP. The selection of polyclonal Ab as targeting moieties broadened
the range of epitopes to which the NP could attach. The magnetic separation strategy
strongly reduces the presence of surfactants or metal ions that are present in complex
environments, which are known inhibitors for sensitive quantification techniques like
qPCR.218,219 We believe that this proof of concept can be translated to a wide
series of diagnostic applications as a fast and efficient pre-concentration step. Future
experiments will optimize the antibody coupling and make use of monospecific Ab to
further enhance the efficiency of the system.
5.4 Conclusions
Superparamagnetic iron oxide NP were functionalized with a heterobifunctional
polyethylene glycol ligand and subsequently bioconjugated with Ab. The ligand has a
siloxane moiety on one end and one carboxylic group, introduced by an anhydride
ring opening, on the other end. We used thiol-ene click chemistry for the siloxane
modification, since this type of reactions is fast and gives high yield without side
reactions. After surface functionalization of the NP, Ab were coupled to the surface.
These bioconjugates were added to L. pneumophila bacteria and after magnetic
separation, the different fractions were investigated by flow cytometry and qPCR.
An enrichment of bacteria was visible in the eluted fraction of the column, showing
that the nanoparticle efficiently interact with the targeted species. The negative
control showed a more than 10-fold lower enrichment, caused by aspecific adsorption.
Moreover, the targeted species were also attracted from a mixture of bacteria, where a
similar enrichment was obtained. These results indicate that the surface functionalized
bioconjugates can be used for magnetic separation of bacteria from complex solutions,
with great efficiency, even though the Ab are coupled non-directionally.
Table 5.2: Summary of all flow cytometry data, measured on the different sets of
spiked samples. All values have as unit: cells/mL. The different Ab, coupled to the NP,
are shown as: +, pAb Leg; -, pAb mur PAI-1. The flow cytometer underestimates the
amount of cells in the NP-cell fraction, since small clusters of cells are formed, that are
only counted as one event. The apparatus has a background signal of approximately
3E3 counts, but this value was not subtracted from the results since it is a systematic
error.
Reference SN + SN - Wash + Wash - NP-cell + NP-cell -

sample 1 5.8E+4 7.0E+3 6.8E+3 7.8E+3 6.4E+3 1.9E+4 9.0E+3
sample 2 5.0E+4 6.0E+3 9.8E+3 5.4E+3 8.8E+3 1.9E+4 7.6E+3
sample 3 4.3E+4 6.4E+3 1.1E+4 6.4E+3 8.8E+3 2.1E+4 1.1E+4
average 5.0E+4 6.5E+3 9.2E+3 6.5E+3 8.0E+3 2.0E+4 9.2E+3

sample 1 5.1E+4 6.6E+3 2.1E+4 4.2E+3 1.1E+4 1.8E+4 1.1E+4
sample 2 5.6E+4 7.8E+3 3.0E+4 1.3E+4 8.4E+3 1.8E+4 1.0E+4
sample 3 4.6E+4 7.8E+3 3.0E+4 6.6E+3 1.1E+4 2.6E+4 8.2E+3
average 5.1E+4 7.4E+3 2.7E+4 8.1E+3 1.0E+4 2.1E+4 9.8E+3

sample 1 3.9E+5 4.2E+3 2.0E+5 4.0E+3 3.6E+4 4.6E+4 7.0E+3
sample 2 4.5E+5 1.0E+4 2.0E+5 3.0E+3 3.4E+4 5.0E+4 8.8E+3
sample 3 4.2E+5 3.2E+3 1.8E+5 4.0E+3 2.9E+4 4.3E+4 1.2E+4
average 4.2E+5 5.8E+3 1.9E+5 3.7E+3 3.3E+4 4.6E+4 9.3E+3

sample 1 4.5E+6 2.0E+4 1.8E+6 3.8E+3 3.0E+5 3.7E+5 3.1E+4
sample 2 4.4E+6 6.2E+3 1.9E+6 5.8E+3 3.2E+5 4.6E+5 5.4E+4
sample 3 4.3E+6 5.8E+3 1.9E+6 5.4E+3 4.4E+5 4.9E+5 3.5E+4
average 4.4E+6 1.1E+4 1.9E+6 5.0E+3 3.5E+5 4.4E+5 4.0E+4
Figure 5.5: Comparison of flow cytometry and qPCR data of the NP-cell fractions.
Since the FCM underestimates the amount of cells due to possible cluster formation,
DNA data were collected as comparison. All values have as unit: cells/mL. The
different antibodies, coupled to the nanoparticles, are shown as: +, pAb Leg; -,
pAb mur PAI-1. The FCM data (green and red curve) clearly show the detection
limit of the apparatus at lower cell concentrations and the underestimation at higher
concentration. However, the ratio of magnetically separated cells by the different
NP-Ab conjugates remains similar regardless the detection technique.
Table 5.3: Comparison of different pre-concentration methods for L. pneumophila and their respective quantification methods.
SUPPORTING INFORMATION
An improvement of immunomagnetic separation over centrifugation and filtration is clearly visible. Individual cases are difficult
to compare since many parameters influence the overall detection efficiency. The quantification method has a profound influence
of the limit of detection, with qPCR being considerably better than FCM. For a complete review of bacterial separation methods,
we direct to reader to a very complete review by Stevens et al.220 Abbreviations: FCM, flow cytometry; EFM, epifluorescent
microscopy; qPCR, quantitative polymerase chain reaction
Method Recovery (percentage) Reference Quantification method Sample

Centrifugation 13.8 (3.800 x g) [221] Cell cultures Spiked tap water
31.9 (8.150 x g) [221] Cell cultures Spiked tap water
10.7 (12.000 x g) [222] FCM Metalworking cooling fluids
13.5 (12.000 x g) [222] EFM Metalworking cooling fluids
43 (3.800 x g) [223] PCR Spiked tap water
49 (8.150 x g) [223] PCR Spiked tap water
Filtration 52.6 (flat membrane) [221] Cell cultures Spiked tap water
13.1 (cast membrane) [221] Cell cultures Spiked tap water
55-81 [223] PCR Spiked tap water
Immunoseparation 12.7 (Dynabeads) [222] FCM Metalworking cooling fluids
29.7 (Dynabeads) [222] EFM Metalworking cooling fluids
46-71 (Dynabeads) [210] qPCR Distilled water
16-74 (Dynabeads) [210] qPCR Potable water
7-89 (Dynabeads) [210] qPCR Cooling tower water
57-96 (nanoparticles) this work FCM Spiked distilled water
99
Chapter 6
Conclusions and outlook
6.1 Conclusions
Multifunctional iron oxide NP have already shown tremendous potential in a wide
range of applications. From a biomedical point of view, their biocompatibility and
low toxicity are important assets. Besides their usage in medical imaging techniques,
such as MRI, drug delivery and hyperthermia are important fields of current research.
In this dissertation, the development of a reliable, low cost and modular nanoparticle-
based magnetic separation platform was the main purpose. Three crucial material
properties were identified beforehand and served as the basis for the design of the
functional nanomaterial platform. (1) The synthesis of the core nanoparticle should
be reproducible, scalable and low cost. A reduction of batch-to-batch differences
was considered the main goal in this regard. (2) Functionalization of the surface of
the nanoparticle should be performed by a covalently attached ligand to enhance
to robustness of the final nanomaterial. (3) The colloidal stability of nanoparticles
in complex environments is often an issue which might negatively influence the
applicability. A heterobifunctional ligand that enhances the colloidal stability and has
a group for subsequent coupling is required.
Chapter 2 focused on the synthesis and functionalization of iron oxide NP, with
special attention to the characterization of the core material. A synthetic procedure
for the large scale synthesis of the core nanoparticle was selected from the abundantly
available protocols in literature. More specifically, the method reported by Park et
al., based on the thermal decomposition of iron(III+) oleate, is capable of producing
high quality magnetite NP at an ultra large scale (up to 20-40 grams).12 However,
the nanoparticle’s surface is coated by oleic acid after the synthesis, which is a very
apolar molecule that is not easily replaced by another ligand. Due to the absence
of reliable surface functionalization methods in scientific literature, a new procedure
101
102 CONCLUSIONS AND OUTLOOK
was developed that is still one of the fastest available methods for covalent surface
modification of oleic acid-coated NP by trialkoxy silanes. The use of an ultrasonication
bath largely reduced the reaction time, by efficiently removing the initial ligand while
keeping the particles separated spatially. A wide range of siloxanes was tested and
the results showed that the proposed protocol is capable of introducing multiple
different varieties onto the iron oxide particles. Moreover the colloidal stability of the
functionalized NP was tested in complex environments, such as serum and plasma.
The non-charged PEG and positively charged carboxylic acids showed an incredible
stability during several days, which was very promising for further development.
After the successful development of the functionalization procedure, the research was
focused onto the required heterobifunctional PEG ligand. Even though thousands of
different siloxane molecules are commercially available, no ligand with all preferred
properties (heterobifunctional based on PEG) is available. Chapter 3 describes the
design and synthesis of the envisioned molecule that incorporates a siloxane at one end,
a well-defined PEG chain and a single functional group at the other end. Since the
modification of native PEG (with 2 hydroxyl groups) would be a less-favourable choice,
an allyl-PEG-OH backbone was bought that allows a straightforward modification.
The hydroxyl group at the end was used for a ring-opening reaction of succinic
anhydride, that resulted in the formation of a carboxylic acid. The allyl group on the
other side was attached to siloxane moiety by using thiol-ene click chemistry. This
approach made it possible to produce the ligand at a large scale, with minimal reaction
steps. As stated before, an elaborate synthetic protocol would largely reduce the
applicability due to cost issues. To prove the added value of the developed ligand, Ab
were covalently coupled to the carboxylic acid groups and their surface concentration
and activity was investigated. The results indicated that the coupling of proteins
resulted in the formation of an almost perfect monolayer, with minimal cross-linking.
Furthermore, the Ab still retained their properties, which was investigated by SPR
experiments, that showed an enhanced interaction with an Ag-coated fiber.
The complex interaction between ligands (with multiple functional groups) and metal
oxides surfaces can result in a wrongful orientation or unwanted side reactions. To
provide a general solution to this (commonly ignored) issue, a strategy was developed
that is based on surface functionalization with protected functional siloxanes and
subsequent deprotection on the nanoparticle. In Chapter 4 we show that this approach
allows to functionalize the surface of iron oxide nanoparticles with different functional
groups, such as amines, carboxylic acids, aldehydes, thiol or sugars. This elegant
and easy method can be transposed onto virtually any (complex) ligand and might
drastically enhance the quality of surface modifications of metal oxide surfaces in
general, even though we showed the methodology on iron oxide NP.
Finally the applicability of the developed heterobifunctional ligand was tested

thoroughly in a magnetic separation experiment with a high social relevance. Polyclonal
Ab (full IgG fraction) were covalently coupled to the magnetic NP to make them target
specific. More detailed, Legionella pneumophila bacteria were selectively magnetically
purified from an aqueous environment. The results indicate that a very high efficiency
can be achieved, even though the covalent coupling with the antibody is non-directional
OUTLOOK 103
and the Ab are not monospecific. A substantial aspecific signal (10%) was observed in
the negative control experiment. To further investigate this potential issue, the target
was separated from a mixture of two bacteria, L. pneumophila and E. coli. Flow
cytometry measurements indicate that the target-specific NP are capable of selectively
attracting a single type of bacteria. The negative control on the other hand showed
no specificity, which is expected for aspecific interactions.
The results, presented in this dissertation, show that a nanoparticle-based platform

was developed, that is capable of selectively targeting a bacterial species in complex
environments. We believe that the cost-effective nature and simplicity of the design
will enhance the broad applicibility of the presented methodology.
6.2 Outlook
Even though we succeeded to develop a fully working magnetic separation methodology,
several enhancements could improve the overall performance. The following section
will focus on the fine-tuning of the presented approach, as well as other applications
that could be developed from the core nanomaterial.
The most apparent modification that could improve the selectivity of the NP are
the selected Ab and the method of covalent coupling. The latter is performed via
a non-directional amide bond, that results in a potential loss of activity of the Ab
since the antigen binding region can be blocked sterically. A few methods have been
reported in scientific literature that are able to bind Ab directionally onto a surface.
Well-known examples are the use of protein A or G as an intermediate layer. These
proteins have a strong affinity for the Fc fragment of the antibody, which forces
the Fab region to point outwards (see Section 1.6). Another option comprises the
chemical modification of the antibody itself, either via a tag, or via the oxidation and
modification of the inner carbohydrates. However Abraham et al. have shown that
this could alter the affinity of the antibody, which could have a negative influence on
the envisioned application.85
Besides the method of covalent coupling, the properties of the ligand itself could also
have an influence on the separation efficiency. It has been reported by Stefanick et
al. that the length of the linker molecule has a large effect on the rotational freedom
of the coupled proteins.224,225 They determined that the optimal length of a PEG
linker was between 6 and 18 monomers. If the chain is too short, the functional group
is buried in the stabilizing layer and hence cannot bind an antibody. If the chain
becomes too long, the PEG chain can sterically hinder the association of the antibody
to its antigen. A shorter linker will probably have a more linear conformation, which
will allow some rotational freedom but will still present the antibody correctly to the
surrounding environment. The optimal linker length will also be influenced by the size
of the NP, since it alters the final surface area and consequently the size of the protein
corona. Furthermore this surface area will influence the aspecific adsorption behavior
104 CONCLUSIONS AND OUTLOOK
of proteins and contaminants in the surrounding medium. It is therefor recommended

to perform a thorough evaluation of a range of linker lengths and coupling chemistries.
The selection of Ab could definitely improve the performance of the presented NP. If
the polyclonal Ab would be purified to monospecific polyclonal Ab, a 20-fold increase
in the selectivity could be obtained. The full IgG fraction is expected to contain only
1-5% of useful Ab. However, this upgrade would only make sense if the coupling
procedure is improved first. The use of monoclonal Ab would ensure a high affinity,
but they only target one epitope. In complex water samples this single type of epitope
might not be accessible (due to pH, contaminants, ...), resulting in a potential loss of
efficiency. Hence the usage of monospecific polyclonal Ab, targeting a broad range of
epitopes, is regarded as the best option for future development.
When looking further into the production costs of the proposed product, it is clear
that the development (and production) of Ab has the largest influence on the final
price. The PEG-coated iron oxide NP have a raw material cost of less than 10
euro/gram. The Ab on the other hand are far more expensive. Taking the radius of
the nanoparticle and its coating into account, it can be estimated that 10-15 Ab could
be coupled on each nanoparticle (by an ideal directional coupling of the Fc fragment
onto the surface).
For polyclonal Ab, the production is fast and the costs are rather low. It can be
estimated that a batch of NP, coated with non-monospecific polyclonal Ab, would
cost approximately 1,000 euro/gram. Keeping in mind that only a few milligrams
are needed for each separation, this is an interesting option. However, the batch size
of polyclonal Ab is limited (approximately 25 mg, since they are collected from the
blood of an animal), so batch to batch differences can be expected. This limits the
long-term applicability, but it can be sufficient for niche markets. If the polyclonal
IgG fraction is purified further to a monospecific batch, we estimate that the price
increases by at least a factor 10, but this depends largely on the concentration of
monospecific Ab in the initial batch (usually 1-5%).
For monoclonal Ab, the initial development is slow and the costs are high (10,000
euro for creating the hybridoma cell lines). However the production can be scaled
up (>100 mg), which significantly lowers the overall price per milligram of Ab. We
estimate that the price of the product (NP-mAb) is 3,000-4,000 euro/gram. Even
though this is more expensive than the polyclonal version, it is a better long-term
option because all Ab are identical and their affinity is guaranteed. To extend the
range of recognized and targeted epitopes, different monoclonal Ab can be mixed.
On the applications side, a wide range of opportunities is available. Thanks to the

excellent colloidal stability of the NP in complex environments, such as waste water
or blood, a logical next step would be their application in water quality monitoring or
biomedical research. Since the properties of collected environmental water samples can
be very diverse in terms of pH or ionic strength, the robustness and enhanced colloidal
stability of the developed nanoparticles could be very advantageous. Detection of
multiple different organisms (Salmonella spp., E. coli or Vibrio cholerae) is mandatory
by law in different parts of the world and hence has a large market potential. The
OUTLOOK 105
proposed method of magnetic pre-concentration can be combined with almost any

detection technique.
The selective purification of proteins (rather than bacteria) could also be of great
interest. The targeting moiety in this regard could still be an antibody, but also a
minor modification of the PEG ligand to introduce a different functional group could
be an option. By using the strategy shown in Chapter 4, very complex ligands can be
introduced onto the nanoparticle’s surface.
In contrast to the diagnostic applicability, imaging is a very important future research

domain. If the functional coating of the particles is enhanced by the addition of
a dye, the resulting fluorescent signal will enable a straightforward localization in
cells or tissue. Multiple approaches are possible in this regards: attaching the dye
directly to a modified PEG ligand or developing a siloxane version of the required dye.
Combined with the selective targeting by antibodies this could be a very valuable
imaging approach.
Similarly, the iron oxide NP show a strong MRI response, which could be exploited
further by the addition of dyes or Ab, to produce for instance a tumor-targeting
bimodal contrast agent. It has been reported that doping the core nanoparticle with
atoms such as cobalt or zinc can largely enhance the MRI contrast. In this context,
the possible toxicity of the magnetite NP should be tested thoroughly. Even though
they are FDA (US Food and Drug Administration) approved and their toxicity is
expected to be low, every new surface coating or dopant atom could trigger negative
effects in in-vivo situations.
In conclusion, the presented functionalized nanoparticles have a tremendous

applicability in various biomedical areas, which will lead to many future research
projects.
Appendix A
Selected experimental
techniques
A.1 Enzyme-linked immunosorbent assay (ELISA)

During an immune response, the concentration of Ab is of crucial importance (next
to the specificity, isotype and affinity). Determination of the concentration of Ab or
Ag can be done by an immunoassay. One of the most used formats is called ELISA.
It was developed by the Swedish researchers Perlmann and Engvall, in 1971.226 The
technique exist in multiple variations, but is generally based on the concept of an
antigen binding to its specific antibody in combination with an enzyme-mediated color
change to detect the presence of proteins, peptides, hormones or Ab.227 The following
section will give an overview of the different possible approaches (see Figure A.1).
The key principle of ELISA is the direct or indirect detection of an antigen by coating
it directly onto the surface of the well, or by coating an antibody with high affinity
for the antigen to the well. For a direct ELISA (ee Figure A.1) the well is coated with
the sample, that has to be analyzed. Typically a blocking solution of noninteracting
proteins is added as an intermediate step, to block uncoated areas. Finally a labeled
antibody, with high affinity for the antigen, is added, that induces the colorimetric
change. An indirect ELISA follows the same approach, but uses an unlabeled primary
antibody first. By addition of a secondary label anti-antibody, the quantification is
possible. The main disadvantage of this approach (direct and indirect) is that the
immobilization of the antigen is not specific. Any other proteins in the sample might
adhere to the well and interfere with the assay.226
The overcome this issue, the sandwich ELISA was developed. Now an antibody with
affinity for the target is coated onto the wells. After addition of the sample and
107
108 SELECTED EXPERIMENTAL TECHNIQUES
Figure A.1: Overview of the different types of ELISA.
sufficient washing, a labeled antibody can be used for the colorimetric quantification
reaction. The antigen is sandwiched between the two Ab, hence the name of the
technique. The purity of the sample is less important in this regard, which simplifies
the assay.
A fourth ELISA approach is based on competition between the antigen in the sample
and antigens coated onto the wells. First an antibody is added to the sample, which
will capture the antigens. If this solution is added to the well, the excess Ab will bind
to the well, but not the ones that already interacted with the target in the sample.
Absence of color indicates the presence of the antigen for this approach. The main
advantage is that the technique is very sensitive even in complex mixtures.227
Figure A.2: Example of an ELISA well plate. Dilution series are made in the vertical
direction to improve the concentration determination.
In this dissertation, the Ag to be quantified was an antibody. Its target was coated on
the plate (e.g. PAI-1, see Chapter 3). The sample containing Ab towards PAI-1 was
applied. A secondary labeled antibody targeting the Ab was added and a colorimetric
reaction was started. Standard curves on the 96-well plate were used to deduct the
FLOW CYTOMETRY (FCM) 109
concentration from the absorbance measurement of the colored solutions (see Figure
A.2).
A.2 Flow cytometry (FCM)

Cytometry is defined as the measurement of physical and chemical characteristics
of particles or cells. In FCM, these cells or particles pass through a narrow channel
which separates them, so discrete signals can be measured. The obtained scattering
and fluorescence data allow to study multiple different types of cells at the same time.
Figure A.3: The principle of FCM is based on the illumination of a focused stream
of sample by a laser source. The forward and side scattering, as well as the different
fluorescent signals are detected. The inset shows a typical graph with different cell
populations.
Two types of signals are usually recorded during FCM measurements: scattering
and fluorescence. The fluorescence signal can originate from labels or dyes and is
often used to identify certain types of cells. By separating the signal into different
wavelength regions (blue, green, red, ...), cell types or populations can be differentiated
(see Figure A.3). The scattering signal is usually split into two parts, the forward
and side scattering. The forward scattering is caused by the scattering of light by the
surface of the cells and is roughly proportional to the cell size. Hence small cells will
have low values, while large cells will have high values. The side scattering on the
110 SELECTED EXPERIMENTAL TECHNIQUES
other hand gives more information about the structural complexity inside the cell and
the granularity.228
In this dissertation, FCM was used to quantify bacteria in the different fraction
we obtained after magnetic isolation, see Chapter 5. Even though a substantial
background signal limited the detection limit, the technique is able to discriminate
between different populations of bacteria, which was a strong asset for the magnetic
isolation experiments.
A.3 Quantitative polymerase chain reaction (qPCR)

During a polymerase chain reaction which relies on thermal cycling, DNA is copied at
an exponential rate. Figure A.4 shows the process of DNA amplification. The double
stranded DNA is heated until the hydrogen bonds are disrupted and the two strands
separate, which is called melting. Afterwards the temperature is lowered to enable the
primers (short DNA fragments) to bind to the strands. Note that they bind at the
5’ (phosphate group) end of the single strand and will extend at their 3’ (hydroxyl
group) end. The polymerase enzyme will now extend the strand to obtain the initial
double stranded DNA. Since the amount of DNA is doubled in every step, one can
obtain large amounts of DNA in a short amount of time. The selectivity of PCR is
related to the use of specific primers that specifically bind to the targeted region in
the DNA. Multiple inhibitors, such as metal ions, surfactants or polysaccharides can
interfere with the reaction and their removal is crucial to obtain good PCR results.211
qPCR or realtime PCR (not to be mistaken with reverse transcriptase PCR (RT-PCR)
which amplifies RNA) is based on the general PCR reaction but uses a fluorescent
signal to quantify in situ the amount of present DNA. This can be achieved by using
non-specific intercalating dyes or target-specific fluorescent-labeled probes. A very
often used, inexpensive and easy to use non-specific dye is SYBR Green, that will bind
to double stranded DNA molecules. However, it can also bind to primer-dimers and
contaminations, hence overestimation of the DNA concentration can occur. Target-
specific alternatives are TaqMan probes or molecular beacons. The first consist of a
fluorescent marker and a quencher in the same probe. During DNA amplification,
the probe is cut by the polymerase enzyme, releasing the fluorescent marker from
the quencher and generating a signal. The molecular beacons are hairpin-like probes
where the fluorophore and quencher are in close proximity. When the probe binds to
the target DNA, the fluorophore is distantiated from the quencher which results in a
fluorescent signal.211,229,230
The fluorescent signal of the DNA amplification is generally plotted (logarithmic)

against the number of cycles. A threshold level is set above the background, but still
in the linear part of the curve (amplification is exponential). The cycle number at
which an amplification plot crosses this threshold fluorescence level is called the Ct or
threshold cycle. This Ct value can be correlated to the initial concentration of DNA
in the sample. A large amount of starting material will result in a rapid amplification
QUANTITATIVE POLYMERASE CHAIN REACTION (QPCR) 111
Figure A.4: During a PCR cycle, double stranded DNA is melted at elevated
temperatures to obtain two single strands. After primer annealing, the extending
process elongates the strand, until a double stranded DNA molecule is obtained. This
process is repeated multiple times until concentration is sufficient for quantification.
and a small Ct value. A small amount on the other hand will take much longer to
reach the threshold level and hence more cycles are required, increasing Ct . Standard
curves are used to correlate the threshold cycle to the initial DNA concentration.231
Bibliography
1. OECD/NNI. Assessing the Economic Impact of Nanotechnology. Tech.

rep. 2013:81.
2. Mattoussi H, Palui G, and Na HB. Luminescent quantum dots as
platforms for probing in vitro and in vivo biological processes. Advanced
drug delivery reviews 2012;64:138–166.
3. Inasawa S, Sugiyama M, and Yamaguchi Y. Laser-induced shape
transformation of gold nanoparticles below the melting point: the effect of
surface melting. The journal of physical chemistry. B 2005;109:3104–3011.
4. Merali Z. This 1,600-Year-Old Goblet Shows that the Romans Were
Nanotechnology Pioneers. 2013.
5. Whitesides GM. The ’right’ size in nanobiotechnology. Nature biotech-
nology 2003;21:1161–1165.
6. Sapsford KE, Algar WR, Berti L, Gemmill KB, Casey BJ, Oh E,
Stewart MH, and Medintz IL. Functionalizing nanoparticles with
biological molecules: developing chemistries that facilitate nanotechnology.
Chemical reviews 2013;113:1904–2074.
7. Montellano A, Da Ros T, Bianco A, and Prato M. Fullerene C60 as a
multifunctional system for drug and gene delivery. Nanoscale 2011;3:4035–
4041.
8. Warner MG and Hutchison JE. Linear assemblies of nanoparticles electro-
statically organized on DNA scaffolds. Nature materials 2003;2:272–277.
9. Langereis E, Heil SBS, Knoops HCM, Keuning W, Sanden MCM van de,
and Kessels WMM. In situ spectroscopic ellipsometry as a versatile tool
for studying atomic layer deposition. Journal of Physics D: Applied
Physics 2009;42:073001.
10. Pang C, Lee C, and Suh KY. Recent advances in flexible sensors for
wearable and implantable devices. Journal of Applied Polymer Science
2013;130:1429–1441.
113
114 BIBLIOGRAPHY
11. Finne R and Klein D. A Water-Amine-Complexing Agent System for

Etching Silicon. Journal of The Electrochemical Society 1967;114:965–970.
12. Park J, An K, Hwang Y, Park JG, Noh HJ, Kim JY, Park JH, Hwang NM,
and Hyeon T. Ultra-large-scale syntheses of monodisperse nanocrystals.
Nature materials 2004;3:891–895.
13. Jun Yw, Lee JH, and Cheon J. Chemical design of nanoparticle probes
for high-performance magnetic resonance imaging. Angewandte Chemie
(International ed. in English) 2008;47:5122–5135.
14. Sun S, Zeng H, Robinson DB, Raoux S, Rice PM, Wang SX, and Li G.
Monodisperse MFe2O4 (M = Fe, Co, Mn) nanoparticles. Journal of the
American Chemical Society 2004;126:273–279.
15. Lu AH, Salabas EL, and Schüth F. Magnetic nanoparticles: synthesis,
protection, functionalization, and application. Angewandte Chemie
16. Jokerst JV, Lobovkina T, Zare RN, and Gambhir SS. Nanoparticle
PEGylation for imaging and therapy. Nanomedicine (London, England)
2011;6:715–728.
17. Mahmoudi M, Laurent S, Shokrgozar MA, and Hosseinkhani M. Toxicity
evaluations of superparamagnetic iron oxide nanoparticles: cell "vision"
versus physicochemical properties of nanoparticles. ACS nano 2011;5:7263–
7276.
18. Schmid G and Klabunde K. Nanoscale materials in chemistry. Vol. 3.
2001.
19. Cornell RM and Schwertmann U. The Iron Oxides. Weinheim, FRG:
Wiley-VCH Verlag GmbH & Co. KGaA, 2003. doi: 10.1002/3527602097.
20. Chatterjee J, Haik Y, and Chen CJ. Size dependent magnetic properties
of iron oxide nanoparticles. Journal of Magnetism and Magnetic Materials
2003;257:113–118.
21. Bean CP and Livingston JD. Superparamagnetism. Journal of Applied
Physics 1959;30:120S–129S.
22. Knobel M, Nunes W, Socolovsky L, De Biasi E, Vargas J, and Denardin J.
Superparamagnetism and Other Magnetic Features in Granular Materials:
A Review on Ideal and Real Systems. Journal of Nanoscience and
Nanotechnology 2008;8:2836–2857.
23. Laurent S, Forge D, Port M, Roch A, Robic C, Vander Elst L, and
Muller RNR. Magnetic iron oxide nanoparticles: synthesis, stabilization,
vectorization, physicochemical characterizations, and biological applica-
tions. Chemical reviews 2008;108:2064–2110.
BIBLIOGRAPHY 115
24. Willard M, Kurihara L, Carpenter E, Calvin S, and Harris V.

Chemically prepared magnetic nanoparticles. International Materials
Reviews 2004;49:125–170.
25. Jolivet JP, Chanéac C, and Tronc E. Iron oxide chemistry. From molecular
clusters to extended solid networks. Chemical communications 2004:481–
487.
26. Brullot W, Reddy N, Wouters J, Valev V, Goderis B, Vermant J, and
Verbiest T. Versatile ferrofluids based on polyethylene glycol coated
iron oxide nanoparticles. Journal of Magnetism and Magnetic Materials
2012;324:1919–1925.
27. Kwon SG and Hyeon T. Formation Mechanisms of Uniform Nanocrystals
via Hot-Injection and Heat-Up Methods. Small 2011:2685–2702.
28. Kwon SG, Piao Y, Park J, Angappane S, Jo Y, Hwang NM, Park JG, and
Hyeon T. Kinetics of Monodisperse Iron Oxide Nanocrystal Formation
by “Heating-Up” Process. Journal of the American Chemical Society
2007;129:12571–12584.
29. LaMer V and Dinegar R. Theory, Production and Mechanism of Formation
of Monodispersed Hydrosols. Journal of the American Chemical Society
1950;72:4847–4854.
30. Xia Y, Xiong Y, Lim B, and Skrabalak SE. Shape-controlled synthesis of
metal nanocrystals: simple chemistry meets complex physics? Angewandte
Chemie (International ed. in English) 2009;48:60–103.
31. Jana N, Chen Y, and Peng X. Size-and shape-controlled magnetic (Cr,
Mn, Fe, Co, Ni) oxide nanocrystals via a simple and general approach.
Chem. Mater 2004;16:3931–3935.
32. Peng X, Wickham J, and Alivisatos A. Kinetics of II-VI and III-V colloidal
semiconductor nanocrystal growth:focusing of size distributions. Journal
of the American Chemical Society 1998;120:5343–5344.
33. Park J, Lee E, Hwang NM, Kang M, Kim SC, Hwang Y, Park JG, Noh HJ,
Kim JY, Park JH, and Hyeon T. One-nanometer-scale size-controlled
synthesis of monodisperse magnetic iron oxide nanoparticles. Angewandte
34. Pérez N, López-Calahorra F, Labarta A, and Batlle X. Reduction of iron
by decarboxylation in the formation of magnetite nanoparticles. Physical
chemistry chemical physics : PCCP 2011;13:19485–19489.
35. Thanh NTK, Maclean N, and Mahiddine S. Mechanisms of nucleation and
growth of nanoparticles in solution. Chemical reviews 2014;114:7610–7630.
116 BIBLIOGRAPHY
36. Wu W, He Q, and Jiang C. Magnetic iron oxide nanoparticles: synthesis

and surface functionalization strategies. Nanoscale research letters
2008;3:397–415.
37. Turcheniuk K, Tarasevych AV, Kukhar VP, Boukherroub R, and
Szunerits S. Recent advances in surface chemistry strategies for the
fabrication of functional iron oxide based magnetic nanoparticles.
Nanoscale 2013;5:10729–10752.
38. Muthiah M, Park IK, and Cho CS. Surface modification of iron oxide
nanoparticles by biocompatible polymers for tissue imaging and targeting.
Biotechnology Advances 2013;31:1224–1236.
39. Liu Z, Zhao B, Shi Y, Guo C, Yang H, and Li Z. Novel nonenzymatic hy-
drogen peroxide sensor based on iron oxide-silver hybrid submicrospheres.
Talanta 2010;81:1650–1654.
40. Prucek R, Tuček J, Kilianová M, Panáček A, Kvítek L, Filip J,
Kolář M, Tománková K, and Zbořil R. The targeted antibacterial and
antifungal properties of magnetic nanocomposite of iron oxide and silver
nanoparticles. Biomaterials 2011;32:4704–4713.
41. Ji X, Shao R, Elliott AM, Stafford RJ, Esparza-Coss E, Liang G, Luo ZP,
Park K, Markert JT, and Li C. Bifunctional Gold Nanoshells with
a Superparamagnetic Iron Oxide-Silica Core Suitable for Both MR
Imaging and Photothermal Therapy. The journal of physical chemistry.
C, Nanomaterials and interfaces 2007;111:6245–6251.
42. Stöber W, Fink A, and Bohn E. Controlled growth of monodisperse silica
spheres in the micron size range. Journal of colloid and interface science
1968;26:62–69.
43. Butterworth M, Bell S, Armes S, and Simpson A. Synthesis and
characterization of polypyrrole-magnetite-silica particles. Journal of
colloid and interface science 1996;183:91–99.
44. Ding HL, Zhang YX, Wang S, Xu JM, Xu SC, and Li GH. Fe3O4-
SiO2 Core/Shell Nanoparticles: The Silica Coating Regulations with a
Single Core for Different Core Sizes and Shell Thicknesses. Chemistry of
Materials 2012;24:4572–4580.
45. Kim J, Kim HS, Lee N, Kim T, Kim H, Yu T, Song IC, Moon WK, and
Hyeon T. Multifunctional uniform nanoparticles composed of a magnetite
nanocrystal core and a mesoporous silica shell for magnetic resonance
and fluorescence imaging and for drug delivery. Angewandte Chemie
BIBLIOGRAPHY 117
46. Chiang IC and Chen DH. Structural characterization and self-assembly

into superlattices of iron oxide-gold core-shell nanoparticles synthe-
sized via a high-temperature organometallic route. Nanotechnology
2009;20:015602.
47. Iglesias-Silva E, Vilas-Vilela J, López-Quintela M, Rivas J, Rodríguez M,
and León L. Synthesis of gold-coated iron oxide nanoparticles. Journal of
Non-Crystalline Solids 2010;356:1233–1235.
48. Brullot W, Valev VK, and Verbiest T. Magnetic-plasmonic nanoparticles
for the life sciences: calculated optical properties of hybrid structures.
Nanomedicine: Nanotechnology, Biology and Medicine 2011;8:559–568.
49. Zhou L, Yuan J, and Wei Y. Core–shell structural iron oxide hybrid
nanoparticles: from controlled synthesis to biomedical applications.
Journal of Materials Chemistry 2011;21:2823–2840.
50. Liu J, Sun Z, Deng Y, Zou Y, Li C, Guo X, Xiong L, Gao Y, Li F, and
Zhao D. Highly Water-Dispersible Biocompatible Magnetite Particles
with Low Cytotoxicity Stabilized by Citrate Groups. Angewandte Chemie
2009;121:5989–5993.
51. Yee C, Kataby G, Ulman A, Prozorov T, White H, King A, Rafailovich M,
Sokolov J, and Gedanken A. Self-Assembled Monolayers of Alkanesulfonic
and -phosphonic Acids on Amorphous Iron Oxide Nanoparticles. Langmuir
1999;15:7111–7115.
52. Smolensky ED, Park HYE, Berquó TS, and Pierre VC. Surface func-
tionalization of magnetic iron oxide nanoparticles for MRI applications -
effect of anchoring group and ligand exchange protocol. Contrast media
& molecular imaging 2011;6:189–199.
53. Xu C, Xu K, Gu H, Zheng R, Liu H, Zhang X, Guo Z, and Xu B.
Dopamine as a robust anchor to immobilize functional molecules on
the iron oxide shell of magnetic nanoparticles. Journal of the American
Chemical Society 2004;126:9938–9939.
54. Zürcher S, Wäckerlin D, Bethuel Y, Malisova B, Textor M, Tosatti
S, and Gademann K. Biomimetic surface modifications based on the
cyanobacterial iron chelator anachelin. Journal of the American Chemical
Society 2006;128:1064–1065.
55. Lee H, Dellatore SM, Miller WM, and Messersmith PB. Mussel-inspired
surface chemistry for multifunctional coatings. Science 2007;318:426–430.
56. Holten-Andersen N, Harrington MJ, Birkedal H, Lee BP, Messersmith PB,
Lee KYC, and Waite JH. pH-induced metal-ligand cross-links inspired
by mussel yield self-healing polymer networks with near-covalent elastic
moduli. Proceedings of the National Academy of Sciences of the United
States of America 2011;108:2651–2655.
118 BIBLIOGRAPHY
57. Taylor S, Luther G, and Waite J. Polarographic and Spectrophotometric

Investigation of Iron(III) Complexation to 3,4-Dihydroxyphenylalanine-
Containing Peptides and Proteins from Mytilus edulis. Inorganic
Chemistry 1994;19:5819–5824.
58. Amstad E, Gehring AU, Fischer Hk, Nagaiyanallur VV, Hahner G, Textor
M, and Reimhult E. Influence of Electronegative Substituents on the
Binding Affinity of Catechol-Derived Anchors to Fe3O4 Nanoparticles.
The Journal of Physical Chemistry C 2011;115:683–691.
59. Shtykova E, Huang X, Remmes N, Baxter D, Stein B, Dragnea B, Svergun
D, and Bronstein L. Structure and Properties of Iron Oxide Nanoparticles
Encapsulated by Phospholipids with Poly(ethylene glycol) Tails. Journal
of Physical Chemistry C 2007;111:18078–18086.
60. Alonso JM, Reichel A, Piehler J, and Campo A del. Photopatterned
surfaces for site-specific and functional immobilization of proteins.
Langmuir 2008;24:448–457.
61. Bantignies JL, Vellutini L, Maurin D, Hermet P, Dieudonné P, Wong Chi
Man M, Bartlett JR, Bied C, Sauvajol JL, and Moreau JJE. Insights into
the self-directed structuring of hybrid organic-inorganic silicas through
infrared studies. The Journal of Physical Chemistry B 2006;110:15797–
15802.
62. Fadeev A and McCarthy T. Self-Assembly Is Not the Only Reaction
Possible between Alkyltrichlorosilanes and Surfaces: Monomolecular and
Oligomeric Covalently Attached Layers of Dichloro- and Trichloroalkylsi-
lanes on Silicon. Langmuir 2000;16:7268–7274.
63. Demirci S and Caykara T. Formation of dicarboxylic acid-terminated
monolayers on silicon wafer surface. Surface Science 2010;604:649–653.
64. Sanchez C, Rozes L, Ribot F, Laberty-Robert C, Grosso D, Sassoye C,
Boissiere C, and Nicole L. Chimie douce: A land of opportunities for the
designed construction of functional inorganic and hybrid organic-inorganic
nanomaterials. Comptes Rendus Chimie 2010;13:3–39.
65. Nagarale RK, Shin W, and Singh PK. Progress in ionic organic-inorganic
composite membranes for fuel cell applications. Polymer Chemistry
2010;1:388–408.
66. Derjaguin B and Landau L. Theory of the stability of strongly charged
lyophobic sols and of the adhesion of strongly charged particles in solutions
of electrolytes. Acta Phys. Chim. USSR 1941;14:633–662.
67. Verwey E and Overbeek J. Theory of stability of lyophobic colloids.
Elsevier, 1948.
68. Xu R. Particle Characterization: Light Scattering Methods. Kluwer
Academic Publishers, 2001:403.
BIBLIOGRAPHY 119
69. Cedervall T, Lynch I, Lindman S, Berggard T, Thulin E, Nilsson H,

Dawson KA, and Linse S. Understanding the nanoparticle-protein corona
using methods to quantify exchange rates and affinities of proteins for
nanoparticles. Proceedings of the National Academy of Sciences of the
United States of America 2007;104:2050–2055.
70. Nel AE, Madler L, Velegol D, Xia T, Hoek EMV, Somasundaran
P, Klaessig F, Castranova V, and Thompson M. Understanding
biophysicochemical interactions at the nano-bio interface. Nature
materials 2009;8:543–557.
71. Lynch I, Cedervall T, Lundqvist M, Cabaleiro-Lago C, Linse S, and
Dawson KA. The nanoparticle-protein complex as a biological entity; a
complex fluids and surface science challenge for the 21st century. Advances
in colloid and interface science 2007;134-135:167–174.
72. Walkey CD and Chan WCW. Understanding and controlling the
interaction of nanomaterials with proteins in a physiological environment.
Chemical Society reviews 2012;41:2780–2799.
73. Lynch I and Dawson K. Protein-nanoparticle interactions. Nano Today
2008;3:40–47.
74. De M, You CC, Srivastava S, and Rotello VM. Biomimetic interactions
of proteins with functionalized nanoparticles: a thermodynamic study.
Journal of the American Chemical Society 2007;129:10747–10753.
75. Janeway C, Travers P, Walport M, and Shlomchik M. Immuno Biology:
the immune system in health and disease. Garland Science, 2005:823.
76. Köhler G and Milstein C. Continuous cultures of fused cells secreting
antibody of predefined specificity. Nature 1975;256:495–497.
77. Leenaars M and Hendriksen C. Critical Steps in the Production of
Polyclonal and Monoclonal Antibodies: Evaluation and Recommendations.
ILAR journal 2005;46:269–279.
78. Kennett R. Hybridomas: a new dimension in biological analyses. In Vitro
1981;17:1036–1050.
79. Huse K, Böhme HJ, and Scholz GH. Purification of antibodies by
affinity chromatography. Journal of biochemical and biophysical methods
2002;51:217–231.
80. Cooper HM and Paterson Y. Production of polyclonal antisera. Current
protocols in neuroscience 2009;5:5.5.1–5.5.10.
81. Hjelm B, Forsström B, Igel U, Johannesson H, Stadler C, Lundberg E,
Ponten F, Sjöberg A, Rockberg J, Schwenk JM, Nilsson P, Johansson C,
and Uhlén M. Generation of monospecific antibodies based on affinity
capture of polyclonal antibodies. Protein science 2011;20:1824–1835.
120 BIBLIOGRAPHY
82. Low D, O’Leary R, and Pujar NS. Future of antibody purification. Journal
of chromatography B 2007;848:48–63.
83. Hermanson G. Bioconjugate techniques. 2nd. Elsevier, 2008:1202.
84. Andersen B, Abele D, and Vannier W. Effects of mild periodate oxidation
on antibodies. The Journal of Immunology 1966;97:913–924.
85. Abraham R, Moller D, and Gabel D. The influence of periodate
oxidation on monoclonal antibody avidity and immunoreactivity. Journal
of immunological methods 1991;144:77–86.
86. Nakajima N and Ikada Y. Mechanism of amide formation by carbodiimide
for bioconjugation in aqueous media. Bioconjugate chemistry 1995;6:123–
130.
87. Staros JV, Wright RW, and Swingle DM. Enhancement by N-hydroxysulfo-
succinimide of water-soluble carbodiimide-mediated coupling reactions.
Analytical Biochemistry 1986;156:220–222.
88. Montalbetti CA and Falque V. Amide bond formation and peptide
coupling. Tetrahedron 2005;61:10827–10852.
89. Fields BS. The molecular ecology of legionellae. Trends in microbiology
1996;4:286–290.
90. Sheehan KB, Henson JM, and Ferris MJ. Legionella species diversity in
an acidic biofilm community in Yellowstone National Park. Applied and
environmental microbiology 2005;71:507–511.
91. World Health Organization (WHO). Legionella and the prevention of
legionellosis. Tech. rep. 2007.
92. Ortiz-Roque C and Hazen T. Abundance and distribution of Legionel-
laceae in Puerto Rican waters. Applied and environmental microbiology
1987;53:2231–2236.
93. Steinert M, Hentschel U, and Hacker J. Legionella pneumophila: an
aquatic microbe goes astray. FEMS microbiology reviews 2002;26:149–162.
94. Vlaamse Overheid. Besluit van de Vlaamse Regering betreffende de
preventie van de veteranenziekte op publiek toegankelijke plaatsen.
Belgisch Staatsblad 2007:23822–23830.
95. Calders R. Legionella pneumophila in vlaanderen presentatie. Tech. rep.
2008.
96. Vincent C and Vogel JP. The Dot/Icm Type IVB Secretion System of
Legionella. Legionella: Molecular Microbiology. 2008.
97. Fraser DW, Tsai TR, Orenstein W, Parkin WE, Beecham HJ, Sharrar RG,
Harris J, Mallison GF, Martin SM, McDade JE, Shepard CC, and
Brachman PS. Legionnaires’ disease: description of an epidemic of
pneumonia. The New England journal of medicine 1977;297:1189–1197.
BIBLIOGRAPHY 121
98. Pernin P, Pélandakis M, Rouby Y, and A. Comparative recoveries of

Naegleria fowleri amoebae from seeded river water by filtration and
centrifugation. Applied and environmental microbiology 1998;64:955–959.
99. Lazcka O, Del Campo FJ, and Muñoz FX. Pathogen detection:
a perspective of traditional methods and biosensors. Biosensors &
bioelectronics 2007;22:1205–1217.
100. Maiwald M, Helbig JH, and Lück PC. Laboratory methods for the
diagnosis of Legionella infections. Journal of Microbiological Methods
1998;33:59–79.
101. Hoefel D, Grooby W, Monis P, Andrews S, and CP. Enumeration of water-
borne bacteria using viability assays and flow cytometry: a comparison to
culture-based techniques. journal of microbiological methods 2003;55:585
–597.
102. Füchslin HP, Kötzsch S, Keserue HA, and Egli T. Rapid and quantitative
detection of Legionella pneumophila applying immunomagnetic separation
and flow cytometry. Cytometry. Part A : the journal of the International
Society for Analytical Cytology 2010;77:264–274.
103. Grow AE, Wood LL, Claycomb JL, and Thompson PA. New biochip
technology for label-free detection of pathogens and their toxins. Journal
of Microbiological Methods 2003;53:221–233.
104. Janib SM, Moses AS, and MacKay JA. Imaging and drug delivery using
theranostic nanoparticles. Advanced drug delivery reviews 2010;62:1052–
1063.
105. Ma J, Fan Q, Wang L, Jia N, Gu Z, and Shen H. Synthesis of magnetic
and fluorescent bifunctional nanocomposites and their applications in
detection of lung cancer cells in humans. Talanta 2010;81:1162–1169.
106. Tran DT, Chen CL, and Chang JS. Immobilization of Burkholderia sp.
lipase on a ferric silica nanocomposite for biodiesel production. Journal
of biotechnology 2012;158:112–119.
107. Wang Z, Wang J, Yue T, Yuan Y, Cai R, and Niu C. Immunomagnetic
separation combined with polymerase chain reaction for the detection of
Alicyclobacillus acidoterrestris in apple juice. PloS one 2013;8:e82376.
108. Chalasani R and Vasudevan S. Cyclodextrin functionalized magnetic
iron oxide nanocrystals: a host-carrier for magnetic separation of non-
polar molecules and arsenic from aqueous media. Journal of Materials
Chemistry 2012;22:14925–14931.
109. Speliotis D. Magnetic recording beyond the first 100 Years. Journal of
Magnetism and Magnetic Materials 1999;193:29–35.
122 BIBLIOGRAPHY
110. Sun J, Zhou S, Hou P, and Yang Y. Synthesis and characterization

of biocompatible Fe3O4 nanoparticles. Journal of Biomedical Materials
Research Part A 2006:333–341.
111. Bronstein LM, Huang X, Retrum J, Schmucker A, Pink M, Stein BD, and
Dragnea B. Influence of Iron Oleate Complex Structure on Iron Oxide
Nanoparticle Formation. Chemistry of Materials 2007;19:3624–3632.
112. Zhang L, He R, and Gu H. Oleic acid coating on the monodisperse
magnetite nanoparticles. Applied Surface Science 2006;253:2611–2617.
113. Fauconnier N, Pons J, Roger J, and Bee A. Thiolation of Maghemite
Nanoparticles by Dimercaptosuccinic Acid. Journal of colloid and interface
science 1997;194:427–433.
114. Wang Y, Wong JF, Teng X, Lin XZ, and Yang H. “Pulling” Nanoparticles
into Water: Phase Transfer of Oleic Acid Stabilized Monodisperse
Nanoparticles into Aqueous Solutions of a-Cyclodextrin. Nano Letters
2003;3:1555–1559.
115. Park JW, Bae KH, Kim C, and Park TG. Clustered magnetite nanocrys-
tals cross-linked with PEI for efficient siRNA delivery. Biomacromolecules
2011;12:457–465.
116. Soderholm KJ and Shang SW. Molecular Orientation of Silane at the
Surface of Colloidal Silica. Journal of Dental Research 1993;72:1050–1054.
117. Zhang Y, Kohler N, and Zhang M. Surface modification of superparamag-
netic magnetite nanoparticles and their intracellular uptake. Biomaterials
2002;23:1553–1561.
118. De Palma R, Peeters S, Van Bael M, Van den Rul H, Bonroy K, Laureyn
W, Mullens J, Borghs G, and Maes G. Silane ligand exchange to make
hydrophobic superparamagnetic nanoparticles water-dispersible. Chem.
Mater 2007;19:1821–1831.
119. Larsen EKU, Nielsen T, Wittenborn T, Birkedal H, Vorup-Jensen T,
Jakobsen MH, Ostergaard L, Horsman MR, Besenbacher F, Howard
Ka, and Kjems Jr. Size-Dependent Accumulation of PEGylated Silane-
Coated Magnetic Iron Oxide Nanoparticles in Murine Tumors. ACS nano
2009;3:1947–1951.
120. Kohler N, Fryxell GE, and Zhang M. A bifunctional poly(ethylene glycol)
silane immobilized on metallic oxide-based nanoparticles for conjugation
with cell targeting agents. Journal of the American Chemical Society
2004;126:7206–7211.
121. Mason T and Lorimer J. Applied Sonochemistry: Uses of power ultrasound
in chemistry and Processing. WILEY-VCH Verlag Weinheim, 2002.
BIBLIOGRAPHY 123
122. Morel AL, Nikitenko SI, Gionnet K, Wattiaux A, Lai-Kee-Him J,

Labrugere C, Chevalier B, Deleris G, Petibois C, Brisson A, and Simonoff
M. Sonochemical approach to the synthesis of Fe(3)O(4)@SiO(2) core-
shell nanoparticles with tunable properties. ACS nano 2008;2:847–856.
123. Prasad NK, Rathinasamy K, Panda D, and Bahadur D. Mechanism
of cell death induced by magnetic hyperthermia with nanoparticles of
γ-MnxFe2–xO3 synthesized by a single step process. Journal of Materials
Chemistry 2007;17:5042–5051.
124. Chen ZP, Xu RZ, Zhang Y, and Gu N. Effects of Proteins from
Culture Medium on Surface Property of Silanes- Functionalized Magnetic
Nanoparticles. Nanoscale research letters 2008;4:204–209.
125. Chen ZP, Zhang Y, Xu K, Xu RZ, Liu JW, and Gu N. Stability of Hy-
drophilic Magnetic Nanoparticles Under Biologically Relevant Conditions.
Journal of Nanoscience and Nanotechnology 2008;8:6260–6265.
126. Can K, Ozmen M, and Ersoz M. Immobilization of albumin on
aminosilane modified superparamagnetic magnetite nanoparticles and its
characterization. Colloids and surfaces. B, Biointerfaces 2009;71:154–159.
127. Ma M, Zhang Y, Yu W, Shen Hy, Zhang Hq, and Gu N. Preparation
and characterization of magnetite nanoparticles coated by amino silane.
Colloids and Surfaces A: Physicochemical and Engineering Aspects
2003;212:219–226.
128. Harder P, Grunze M, Dahint R, Whitesides G, and Laibinis P. Molecular
Conformation in Oligo(ethylene glycol)-Terminated Self-Assembled
Monolayers on Gold and Silver Surfaces Determines Their Ability
To Resist Protein Adsorption. The Journal of Physical Chemistry B
1998;102:426–436.
129. Valiokas R, Svedhem S, and Svensson S. Self-assembled monolayers
of oligo (ethylene glycol)-terminated and amide group containing
alkanethiolates on gold. Langmuir 1999;15:3390–3394.
130. Schweiss R, Welzel P, Werner C, and Knoll W. Dissociation of surface
functional groups and preferential adsorption of ions on self-assembled
monolayers assessed by streaming potential and streaming current
measurements. Langmuir 2001;17:4304–4311.
131. Kreuzer HJ, Wang RLC, and Grunze M. Hydroxide ion adsorption on
self-assembled monolayers. Journal of the American Chemical Society
2003;125:8384–8389.
132. Barreto Ja, Matterna M, Graham B, Stephan H, and Spiccia L. Synthesis,
colloidal stability and 64Cu labeling of iron oxide nanoparticles bearing
different macrocyclic ligands. New Journal of Chemistry 2011;35:2705–
2712.
124 BIBLIOGRAPHY
133. Bhattacharya D, Sahu SK, Banerjee I, Das M, Mishra D, Maiti TK, and
Pramanik P. Synthesis, characterization, and in vitro biological evaluation
of highly stable diversely functionalized superparamagnetic iron oxide
nanoparticles. Journal of Nanoparticle Research 2011;13:4173–4188.
134. Tartaj P, Morales MDP, Veintemillas-Verdaguer S, Gonzalez-Carreno T,
and Serna CJ. The preparation of magnetic nanoparticles for applications
in biomedicine. Journal of Physics D: Applied Physics 2003;36:182–197.
135. Montenegro JM, Grazu V, Sukhanova A, Agarwal S, Fuente JM de
la, Nabiev I, Greiner A, and Parak WJ. Controlled antibody/(bio-)
conjugation of inorganic nanoparticles for targeted delivery. Advanced
drug delivery reviews 2013;65:677–688.
136. Jeong U, Teng X, Wang Y, Yang H, and Xia Y. Superparamagnetic
Colloids: Controlled Synthesis and Niche Applications. Advanced
Materials 2007;19:33–60.
137. Lee N and Hyeon T. Designed synthesis of uniformly sized iron oxide
nanoparticles for efficient magnetic resonance imaging contrast agents.
138. Park J, Joo J, Kwon SG, Jang Y, and Hyeon T. Synthesis of monodisperse
spherical nanocrystals. Angewandte Chemie (International ed. in English)
2007;46:4630–4660.
139. Wang X, Zhuang J, Peng Q, and Li Y. A general strategy for nanocrystal
synthesis. Nature 2005;437:121–124.
140. Lin Fh, Chen W, Liao YH, Doong Ra, and Li Y. Effective approach for
the synthesis of monodisperse magnetic nanocrystals and M-Fe3O4 (M
= Ag, Au, Pt, Pd) heterostructures. Nano Research 2011;4:1223–1232.
141. Park EJ, Kim H, Kim Y, Yi J, Choi K, and Park K. Inflammatory
responses may be induced by a single intratracheal instillation of iron
nanoparticles in mice. Toxicology 2010;275:65–71.
142. Mendoza A, Torres-Hernandez JA, Ault JG, Pedersen-Lane JH, Gao
D, and Lawrence DA. Silica nanoparticles induce oxidative stress and
inflammation of human peripheral blood mononuclear cells. Cell stress &
chaperones 2014;19:777–790.
143. Trickler WJ, Lantz SM, Murdock RC, Schrand AM, Robinson BL,
Newport GD, Schlager JJ, Oldenburg SJ, Paule MG, Slikker W,
Hussain SM, and Ali SF. Silver nanoparticle induced blood-brain barrier
inflammation and increased permeability in primary rat brain microvessel
endothelial cells. Toxicological sciences : an official journal of the Society
of Toxicology 2010;118:160–170.
BIBLIOGRAPHY 125
144. Petri-Fink A, Steitz B, Finka A, Salaklang J, and Hofmann H. Effect of

cell media on polymer coated superparamagnetic iron oxide nanoparticles
(SPIONs): colloidal stability, cytotoxicity, and cellular uptake studies.
European journal of pharmaceutics and biopharmaceutics 2008;68:129–
137.
145. Gerion D, Pinaud F, Williams SC, Parak WJ, Zanchet D, Weiss S, and
Alivisatos AP. Synthesis and Properties of Biocompatible Water-Soluble
Silica-Coated CdSe/ZnS Semiconductor Quantum Dots. The Journal of
Physical Chemistry B 2001;105:8861–8871.
146. Jiang W, Kim BYS, Rutka JT, and Chan WCW. Nanoparticle-mediated
cellular response is size-dependent. Nature nanotechnology 2008;3:145–
150.
147. Bloemen M, Brullot W, Denis C, Vanysacker L, and Verbiest T. Core-shell
nanoparticles as enhanced probes for imaging applications. Proceedings
of the Society of Photo-Optical Instrumentation Engineers: Vol. 8427
2012;8427. Ed. by Popp J, Drexler W, Tuchin VV, and Matthews DL.
148. Tucker-Schwartz AK, Farrell RA, and Garrell RL. Thiol-ene click reaction
as a general route to functional trialkoxysilanes for surface coating appli-
cations. Journal of the American Chemical Society 2011;133:11026–11029.
149. Moses JE and Moorhouse AD. The growing applications of click chemistry.
150. Kolb HC, Finn MG, and Sharpless KB. Click Chemistry: Diverse Chemical
Function from a Few Good Reactions. Angewandte Chemie (International
ed. in English) 2001;40:2004–2021.
151. Hoyle CE and Bowman CN. Thiol-ene click chemistry. Angewandte
152. Bijnens AP, Gils A, Knockaert I, Stassen JM, and Declerck PJ.
Importance of the Hinge Region between alpha -Helix F and the Main
Part of Serpins, Based upon Identification of the Epitope of Plasminogen
Activator Inhibitor Type 1 Neutralizing Antibodies. Journal of Biological
Chemistry 2000;275:6375–6380.
153. Declerck PJ and Gils A. Three decades of research on plasminogen
activator inhibitor-1: a multifaceted serpin. Seminars in thrombosis and
hemostasis 2013;39:356–364.
154. Spivey AC and Arseniyadis S. Nucleophilic catalysis by 4-(dialkylamino)
pyridines revisited–the search for optimal reactivity and selectivity.
Angewandte Chemie (International ed. in English) 2004;43:5436–5441.
155. Dixit CK, Vashist SK, MacCraith BD, and O’Kennedy R. Multisubstrate-
compatible ELISA procedures for rapid and high-sensitivity immunoas-
says. Nature protocols 2011;6:439–445.
126 BIBLIOGRAPHY
156. Gils A, Ceresa E, Macovei AM, Marx PF, Peeters M, Compernolle G, and
Declerck PJ. Modulation of TAFI function through different pathways–
implications for the development of TAFI inhibitors. Journal of thrombosis
and haemostasis : JTH 2005;3:2745–2753.
157. Bradford MM. A rapid and sensitive method for the quantitation of
microgram quantities of protein utilizing the principle of protein-dye
binding. Analytical Biochemistry 1976;72:248–254.
158. Casals E, Pfaller T, Duschl A, Oostingh GJ, and Puntes V. Time evolution
of the nanoparticle protein corona. ACS nano 2010;4:3623–3632.
159. Lundqvist M, Stigler J, Elia G, Lynch I, Cedervall T, and Dawson KA.
Nanoparticle size and surface properties determine the protein corona with
possible implications for biological impacts. Proceedings of the National
Academy of Sciences of the United States of America 2008;105:14265–
14270.
160. Cui M, Liu R, Deng Z, Ge G, Liu Y, and Xie L. Quantitative study of
protein coronas on gold nanoparticles with different surface modifications.
Nano Research 2014;7:345–352.
161. Zhang S, Zou L, Zhang D, Pang X, Yang H, and Xu Y. GoldMag nanopar-
ticles with core/shell structure: characterization and application in MR
molecular imaging. Journal of Nanoparticle Research 2011;13:3867–3876.
162. Ranjbakhsh E, Bordbar a.K., Abbasi M, Khosropour a.R., and Shams E.
Enhancement of stability and catalytic activity of immobilized lipase on
silica-coated modified magnetite nanoparticles. Chemical Engineering
Journal 2012;179:272–276.
163. Shi B, Wang Y, Ren J, Liu X, Zhang Y, Guo Y, Guo Y, and
Lu G. Superparamagnetic aminopropyl-functionalized silica core-shell
microspheres as magnetically separable carriers for immobilization of
penicillin G acylase. Journal of Molecular Catalysis B: Enzymatic
2010;63:50–56.
164. Cui Y, Li Y, Yang Y, Liu X, Lei L, Zhou L, and Pan F. Facile synthe-
sis of amino-silane modified superparamagnetic Fe3O4 nanoparticles
and application for lipase immobilization. Journal of biotechnology
2010;150:171–174.
165. Cao H, He J, Deng L, and Gao X. Fabrication of cyclodextrin-
functionalized superparamagnetic Fe3O4/amino-silane core–shell nanopar-
ticles via layer-by-layer method. Applied Surface Science 2009;255:7974–
7980.
166. Pollet J, Delport F, Janssen KPF, Jans K, Maes G, Pfeiffer H, Wevers M,
and Lammertyn J. Fiber optic SPR biosensing of DNA hybridization and
DNA-protein interactions. Biosensors & bioelectronics 2009;25:864–869.
BIBLIOGRAPHY 127
167. Ge Y, Zhang Y, Xia J, Ma M, He S, Nie F, and Gu N. Effect of surface

charge and agglomerate degree of magnetic iron oxide nanoparticles
on KB cellular uptake in vitro. Colloids and surfaces. B, Biointerfaces
2009;73:294–301.
168. Osaka T, Nakanishi T, Shanmugam S, Takahama S, and Zhang H. Effect
of surface charge of magnetite nanoparticles on their internalization into
breast cancer and umbilical vein endothelial cells. Colloids and surfaces.
B, Biointerfaces 2009;71:325–330.
169. Hoyle CE, Lee TY, and Roper T. Thiol-enes: Chemistry of the past
with promise for the future. Journal of Polymer Science Part A: Polymer
Chemistry 2004;42:5301–5338.
170. Northrop BH and Coffey RN. Thiol-ene click chemistry: computational
and kinetic analysis of the influence of alkene functionality. Journal of
the American Chemical Society 2012;134:13804–13817.
171. Mahon E, Salvati A, Baldelli Bombelli F, Lynch I, and Dawson KA.
Designing the nanoparticle-biomolecule interface for "targeting and
therapeutic delivery". Journal of controlled release : official journal of the
Controlled Release Society 2012;161:164–174.
172. Noh SH, Na W, Jang JT, Lee JH, Lee EJ, Moon SH, Lim Y,
Shin JS, and Cheon J. Nanoscale magnetism control via surface and
exchange anisotropy for optimized ferrimagnetic hysteresis. Nano letters
2012;12:3716–3721.
173. Babes L, Denizot B, Tanguy G, Le Jeune J, and Jallet P. Synthesis of
Iron Oxide Nanoparticles Used as MRI Contrast Agents: A Parametric
Study. Journal of colloid and interface science 1999;212:474–482.
174. Schladt TD, Schneider K, Schild H, and Tremel W. Synthesis and bio-
functionalization of magnetic nanoparticles for medical diagnosis and
treatment. Dalton transactions 2011;40:6315–6343.
175. Yu W, Falkner J, Yavuz C, and Colvin V. Synthesis of monodisperse iron
oxide nanocrystals by thermal decomposition of iron carboxylate salts.
Chem. Commun 2004;20:2306–2307.
176. Sun S and Zeng H. Size-controlled synthesis of magnetite nanoparticles.
Journal of the American Chemical Society 2002;124:8204–8205.
177. Harisinghani MG, Barentsz J, Hahn PF, Deserno WN, Tabatabaei S,
Hulsbergen van de Kaa C, Rosette J de la, and Weissleder R. Noninvasive
detection of clinically occult lymph-node metastases in prostate cancer.
The New England Journal of Medicine 2003;348:2491–2499.
178. Bloemen M, Debruyne D, Demeyer PJ, Clays K, Gils A, Geukens N,
Bartic C, and Verbiest T. Catechols as ligands for CdSe–ZnS quantum
dots. RSC Advances 2014;4:10208–10211.
128 BIBLIOGRAPHY
179. Na HB, Palui G, Rosenberg JT, Ji X, Grant SC, and Mattoussi H.

Multidentate catechol-based polyethylene glycol oligomers provide
enhanced stability and biocompatibility to iron oxide nanoparticles. ACS
nano 2012;6:389–399.
180. Daniels M and Francis L. Silane Adsorption Behavior, Microstructure,
and Properties of Glycidoxypropyltrimethoxysilane-Modified Colloidal
Silica Coatings. Journal of colloid and interface science 1998;205:191–200.
181. Boerio FJ, Armogan L, and Cheng S. The structure of γ-aminopropyl
triethoxysilane films on iron mirrors. Journal of Colloid and Interface
Science 1980;73:416–424.
182. Vandenberg ET, Bertilsson L, Liedberg B, Udval K, Erlandsson R, Elwing
H, and Lundström I. Structure of 3-Aminopropyl Triethoxy Silane on
Silicon Oxide. Journal of Colloid and Interface Science 1991;147:103–118.
183. Gu RA, Chen H, Liu GK, and Ren B. Study on Films of Organofunctional
Silanes on Transition Metal Surfaces- The Structure of Aminopropy-
ltrimethoxysilane Films on Nickel Electrodes. Acta Chimica Sinica
2003;61:1550–1555.
184. Brullot W, Strobbe R, Bynens M, Bloemen M, Demeyer PJ, Vanderlinden
W, De Feyter S, Valev VK, and Verbiest T. Layer-by-Layer synthesis and
tunable optical properties of hybrid magnetic–plasmonic nanocomposites
using short bifunctional molecular linkers. Materials Letters 2014;118:99–
102.
185. Crouch DR, Burger JS, Zietek KA, Cadwallader AB, Bedison JE, and
Smielewska MM. Removal of Acyl Protecting Groups Using Solid NaOH
and a Phase Transfer Catalyst. Synlett 2003;7:991–992.
186. Keating CD, Musick MD, Keefe MH, and Natan MJ. Kinetics and
Thermodynamics of Au Colloid Monolayer Self-Assembly: Undergraduate
Experiments in Surface and Nanomaterials Chemistry. Journal of
Chemical Education 1999;76:949–955.
187. Golub A, Zubenko A, and Zhmud B. APTES Modified Silica Gels : The
Structure of the Surface Layer. Journal of colloid and interface science
1996;487:482–487.
188. Ung T, Liz-Marzán L, and Mulvaney P. Controlled method for silica
coating of silver colloids. Influence of coating on the rate of chemical
reactions. Langmuir 1998;7463:3740–3748.
BIBLIOGRAPHY 129
189. Maczynski A, Shaw DG, Goral M, Wisniewska-goclowska B, Skrzecz

A, Owczarek I, Blazej K, Haulait-Pirson MC, Hefter GT, Kapuka F,
Maczynska Z, Szafranski A, and Young CL. IUPAC-NIST Solubility
Data Series. 81. Hydrocarbons with Water and Seawater-Revised and
Updated. Part 5. C7 Hydrocarbons with Water and Heavy Water. Journal
of Physical and Chemical Reference Data 2005;34:1399–1487.
190. Wuts PGM and Greene TW. Greene’s Protective Groups in Organic
Synthesis. 4th. Hoboken, NJ, USA: John Wiley & Sons, Inc., 2006. doi:
10.1002/0470053488.
191. Pittelkow M and Christensen JrB. Convergent synthesis of internally
branched PAMAM dendrimers. Organic letters 2005;7:1295–1298.
192. Jung C and Jacobs P. Physical and chemical properties of superparamag-
netic iron oxide MR contrast agents: ferumoxides, ferumoxtran, ferumoxsil.
Magnetic resonance imaging 1995;13:661–674.
193. Lartigue L, Innocenti C, Kalaivani T, Awwad A, Sanchez Duque MDM,
Guari Y, Larionova J, Guérin C, Montero JLG, Barragan-Montero V,
Arosio P, Lascialfari A, Gatteschi D, and Sangregorio C. Water-dispersible
sugar-coated iron oxide nanoparticles. An evaluation of their relaxometric
and magnetic hyperthermia properties. Journal of the American Chemical
Society 2011;133:10459–10472.
194. Larson TA, Bankson J, Aaron J, and Sokolov K. Hybrid plasmonic
magnetic nanoparticles as molecular specific agents for MRI/optical
imaging and photothermal therapy of cancer cells. Nanotechnology
2007;18:325101.
195. Bedanta S and Kleemann W. Supermagnetism. Journal of Physics D:
Applied Physics 2009;42:013001.
196. Yang H, Qu L, Wimbrow AN, Jiang X, and Sun Y. Rapid detection
of Listeria monocytogenes by nanoparticle-based immunomagnetic
separation and real-time PCR. International journal of food microbiology
2007;118:132–138.
197. Shahbazi-Gahrouei D, Abdolahi M, Zarkesh-Esfahani SH, Laurent S,
Sermeus C, and Gruettner C. Functionalized magnetic nanoparticles for
the detection and quantitative analysis of cell surface antigen. BioMed
research international 2013;2013:349408.
198. Xu H, Aguilar ZP, Yang L, Kuang M, Duan H, Xiong Y, Wei H, and Wang
A. Antibody conjugated magnetic iron oxide nanoparticles for cancer cell
separation in fresh whole blood. Biomaterials 2011;32:9758–9765.
130 BIBLIOGRAPHY
199. Gillich T, Acikgöz C, Isa L, Schlüter aD, Spencer ND, and Textor M.
PEG-stabilized core-shell nanoparticles: impact of linear versus dendritic
polymer shell architecture on colloidal properties and the reversibility of
temperature-induced aggregation. ACS nano 2013;7:316–329.
200. Bloemen M, Van Stappen T, Willot P, Lammertyn J, Koeckelberghs G,
Geukens N, Gils A, and Verbiest T. Heterobifunctional PEG ligands
for bioconjugation reactions on iron oxide nanoparticles. PloS ONE
2014;9:e109475.
201. Koh I, Wang X, Varughese B, Isaacs L, Ehrman SH, and English DS.
Magnetic iron oxide nanoparticles for biorecognition: evaluation of surface
coverage and activity. The journal of physical chemistry. B 2006;110:1553–
1558.
202. Dupont D, Brullot W, Bloemen M, Verbiest T, and Binnemans K.
Selective uptake of rare earths from aqueous solutions by EDTA-
functionalized magnetic and nonmagnetic nanoparticles. ACS applied
materials & interfaces 2014;6:4980–4988.
203. Dupont D, Luyten J, Bloemen M, Verbiest T, and Binnemans K. Acid-
Stable Magnetic Core–Shell Nanoparticles for the Separation of Rare
Earths. Industrial & Engineering Chemistry Research 2014;53:15222–
15229.
204. Horák D, Babic M, Macková H, and Benes MJ. Preparation and properties
of magnetic nano- and microsized particles for biological and environmen-
tal separations. Journal of separation science 2007;30:1751–1772.
205. Chalmers JJ, Xiong Y, Jin X, Shao M, Tong X, Farag S, and
Zborowski M. Quantification of non-specific binding of magnetic micro-
and nanoparticles using cell tracking velocimetry: Implication for
magnetic cell separation and detection. Biotechnology and bioengineering
2010;105:1078–1093.
206. Atlas RM. Legionella: from environmental habitats to disease pathology,
detection and control. Environmental Microbiology 1999;1:283–293.
207. Vogel JP and Isberg RR. Cell biology of Legionella pneumophila. Current
opinion in microbiology 1999;2:30–34.
208. Behets J, Seghi F, Declerck P, Verelst L, Duvivier L, Van Damme A,
and Ollevier F. Detection of Naegleria spp. and Naegleria fowleri: a
comparison of flagellation tests, ELISA and PCR. Water science and
technology : a journal of the International Association on Water Pollution
Research 2003;47:117–122.
BIBLIOGRAPHY 131
209. Behets J, Declerck P, Delaedt Y, Verelst L, and F. Quantitative

Detection and Differentiation of Free-Living Amoeba Species Using
SYBR Green–Based Real-Time PCR Melting Curve Analysis. Current
microbiology 2006;53:506–509.
210. Yáñez MA, Carrasco-Serrano C, Barberá VM, and Catalán V. Quan-
titative detection of Legionella pneumophila in water samples by
immunomagnetic purification and real-time PCR amplification of the
dotA gene. Applied and environmental microbiology 2005;71:3433–3441.
211. Demeke T and Jenkins GR. Influence of DNA extraction methods,
PCR inhibitors and quantification methods on real-time PCR assay
of biotechnology-derived traits. Analytical and bioanalytical chemistry
2010;396:1977–1990.
212. Van De Craen B, Scroyen I, Vranckx C, Compernolle G, Lijnen HR,
Declerck PJ, and Gils A. Maximal PAI-1 inhibition in vivo requires
neutralizing antibodies that recognize and inhibit glycosylated PAI-1.
Thrombosis research 2012;129:e126–133.
213. Van Stappen T, Brouwers E, Tops S, Geukens N, Vermeire S, Declerck PJ,
and Gils A. Generation of a highly specific monoclonal anti-infliximab
antibody for harmonization of TNF-coated infliximab assays. Therapeutic
drug monitoring 2014.
214. Bloemen M, Brullot W, Luong TT, Geukens N, Gils A, and Verbiest T.
biomedical applications. Journal of Nanoparticle Research 2012;14:1100–
1109.
215. Köhler R, Bubert A, Goebel W, Steinert M, Hacker J, and Bubert B.
Expression and use of the green fluorescent protein as a reporter system
in Legionella pneumophila. MGG - Molecular and General Genetics
2000;262:1060–1069.
216. Behets J, Declerck P, Delaedt Y, Creemers B, and Ollevier F. Development
and evaluation of a Taqman duplex real-time PCR quantification method
for reliable enumeration of Legionella pneumophila in water samples.
Journal of microbiological methods 2007;68:137–44.
217. Rao SV, Anderson KW, and Bachas LG. Oriented immobilization of
proteins. Mikrochimica Acta 1998;128:127–143.
218. Fluit AC, Torensma R, Visser MJ, Aarsman CJ, Poppelier MJ, Keller BH,
Klapwijk P, and Verhoef J. Detection of Listeria monocytogenes in cheese
with the magnetic immuno-polymerase chain reaction assay. Applied and
environmental microbiology 1993;59:1289–1293.
132 BIBLIOGRAPHY
219. Coklin T, Farber JM, Parrington LJ, Bin Kingombe CI, Ross WH,
and Dixon BR. Immunomagnetic separation significantly improves the
sensitivity of polymerase chain reaction in detecting Giardia duodenalis
and Cryptosporidium spp. in dairy cattle. Journal of veterinary diagnostic
investigation 2011;23:260–267.
220. Stevens Ka and Jaykus LA. Bacterial separation and concentration from
complex sample matrices: a review. Critical reviews in microbiology
2004;30:7–24.
221. Boulanger Ca and Edelstein PH. Precision and accuracy of recovery of
Legionella pneumophila from seeded tap water by filtration and centrifu-
gation. Applied and environmental microbiology 1995;61:1805–1809.
222. Chang SC, Anderson TI, Bahrman SE, Gruden CL, Khijniak AI,
and Adriaens P. Comparing recovering efficiency of immunomagnetic
separation and centrifugation of mycobacteria in metalworking fluids.
Journal of Industrial Microbiology and Biotechnology 2005;32:629–638.
223. Villari P, Motti E, Farullo C, and Torre I. Comparison of conventional
culture and PCR methods for the detection of Legionella pneumophila in
water. Letters in applied microbiology 1998;27:106–110.
224. Stefanick JF, Ashley JD, Kiziltepe T, and Bilgicer B. A systematic
analysis of peptide linker length and liposomal polyethylene glycol coating
on cellular uptake of peptide-targeted liposomes. ACS Nano 2013;7:2935–
2947.
225. Stefanick JF, Ashley JD, and Bilgicer B. Enhanced cellular uptake of
peptide-targeted nanoparticles through increased peptide hydrophilic-
ity and optimized ethylene glycol peptide-linker length. ACS Nano
2013;7:8115–8127.
226. Engvall E and Perlmann P. Enzyme-linked immunosorbent assay
(ELISA) quantitative assay of immunoglobulin G. Molecular Immunology
1971;8:871–874.
227. Gan SD and Patel KR. Enzyme immunoassay and enzyme-linked im-
munosorbent assay. The Journal of investigative dermatology 2013;133:e12.
228. Applied Biosystems. Attune Acoustic Focusing Cytometer Training Guide.
Tech. rep. 2010.
229. Huggett J, Dheda K, Bustin S, and Zumla A. Real-time RT-PCR
normalisation; strategies and considerations. Genes and immunity
2005;6:279–284.
230. Wong ML and Medrano JF. Real-time PCR for mRNA quantitation.
BioTechniques 2005;39:75–85.
231. Agilent Technologies, Introduction to Quantitative PCR. Tech. rep. 2010.
BIBLIOGRAPHY 133
232. Liu G, Gao J, Ai H, and Chen X. Applications and potential toxicity of

magnetic iron oxide nanoparticles. Small 2013;9:1533–1545.
233. Murray AR, Kisin E, Inman A, Young SH, Muhammed M, Burks T,
Uheida A, Tkach A, Waltz M, Castranova V, Fadeel B, Kagan VE,
Riviere JE, Monteiro-Riviere N, and Shvedova Aa. Oxidative Stress and
Dermal Toxicity of Iron Oxide Nanoparticles In Vitro. Cell Biochemistry
and Biophysics 2013;67:461–476.
234. Ling D and Hyeon T. Chemical design of biocompatible iron oxide
nanoparticles for medical applications. Small 2013;9:1450–1466.
235. Gupta AK and Wells S. Surface-Modified Superparamagnetic Nanoparti-
cles for Drug Delivery: Preparation, Characterization, and Cytotoxicity
Studies. IEEE Transactions on Nanobioscience 2004;3:66–73.
236. Wan S, Huang J, Guo M, Zhang H, Cao Y, Yan H, and Liu K.
Biocompatible superparamagnetic iron oxide nanoparticle dispersions
stabilized with poly(ethylene glycol)-oligo(aspartic acid) hybrids. Journal
of Biomedical Materials Research - Part A 2007;80:946–954.
237. Yu M, Huang S, Yu KJ, and Clyne AM. Dextran and polymer polyethylene
glycol (PEG) coating reduce both 5 and 30 nm iron oxide nanoparticle
cytotoxicity in 2D and 3D cell culture. International Journal of Molecular
Sciences 2012;13:5554–5570.
Health, safety and
environment
During the experiments, conducted for this dissertation, all health and safety guidelines
and regulations from the HSE department of the KU Leuven were implemented. Risk
analyses and proper waste management were evaluated before every experiment and
were submitted to the appropriate channels.
The following sections will summarize several specific cases where special precautions
should be taken.
Nanoparticle synthesis by thermal decomposition
As stated in Chapter 2, the synthesis of iron oxide NP by the thermal decomposition

methods is conducted at elevated temperatures, above 300°C. Special personal
protection equipment, such as heat resistant gloves, should be worn when manipulating
the reaction at this stage. Furthermore the reaction is known to produce CO2 gas,
which can cause a pressure buildup in the system. It is advised to keep the setup well
ventilated and allow the excess pressure to escape. Finally, it is necessary to allow the
hot mixture to cool down to room temperature before the purification is started.
Nanoparticle toxicity
The scientific community is still debating whether or not NP are toxic. Since a general
consensus has not been achieved so far, every nanoparticle system is being evaluated
separately. In the case of iron oxide, a clear difference has been observed between the
toxicity of the core and the different coatings.
Bare iron oxide is known to generate reactive oxygen species by the Fenton reaction,
during which hydrogen peroxide is turned into hydroxyl (HO•) and hydroperoxyl
(HOO•) radicals. These radical species can easily induce damage to proteins or DNA
which can result in a decline in physiological function and cell death.232–234
134
BIBLIOGRAPHY 135
The coating of a nanoparticle has a profound influence on the toxicity towards cells
and tissue. PEG-coated particles generally perform well during toxicity tests, showing
only a minor decrease in cell viability.235,236 Mahmoudi et al. reported that particles
with carboxylic acids show lower toxicity than amine-coated NP, but the effect is
largely dependent on the cell type. Yu et al. wrote that the size of the nanoparticle
and even the type of study (2D versus 3D) can have an influence on the observed
toxicity.237 Therefor it is advised to perform a complete toxicology study before in vivo
experiments are conducted. Based on the available literature the expected toxic effects
are small, but the lack of general consensus and understanding of the phenomenon
requires that every type of nanoparticle is tested separately.
During the presented experiments, the particles were always manipulated in solution,
to reduce any risks of dust formation and inhalation. Secondly, appropriate personal
protection equipment, such as gloves, were worn at every step.
Legionella pneumophila
The KB of August 04, 1996 states that Legionella pneumophila is a class 2 pathogen
and should be handled accordingly in a class 2 lab. A full risk analysis should be
made before the start of the experiments and exposure should be minimized at all
times. Moreover, proper disinfection of equipment (for instance with ethanol) was
performed.
List of publications
Articles in internationally reviewed academic journals

10. Carron S, Bloemen M, Vander Elst L, Laurent S, Verbiest T and Parac-Vogt
T. Potential Theranostic and Multimodal Iron Oxide Nanoparticles Decorated
with Rhenium-Bipyridine and -Phenanthroline Complex. Journal of Materials
Chemistry B 2015; Accepted.
9. Bloemen M, Denis C, Peeters M, De Meester L, Gils A, Geukens N and

Verbiest T. Antibody-modified iron oxide nanoparticles for efficient magnetic
isolation and flow cytometric determination of L. pneumophila. Microchimica
Acta 2015; Accepted.
8. Bloemen M, Van Stappen T, Willot P, Lammertyn J, Koeckelberghs G,

Geukens N, Gils A and Verbiest T. Heterobifunctional PEG Ligands for
Bioconjugation Reactions on Iron Oxide Nanoparticles. PLoS ONE 2014;
9: e109475.
7. Bloemen M, Sutens B, Brullot W, Gils A, Geukens N and Verbiest T. Two-Step

Directional Surface Modification of Iron Oxide Nanoparticles with Protected
Siloxanes. Chempluschem 2014; 80: 50-53.
6. Dupont D, Luyten J, Bloemen M, Verbiest T and Binnemans K. Acid-Stable

Magnetic Core-Shell Nanoparticles for the Separation of Rare Earths. Industrial
& Engineering Chemistry Research 2014; 53: 15222–15229.
5. Brullot W, Strobbe R, Bynens M, Bloemen M, Demeyer P-J, Vanderlinden

W, De Feyter S, Valev V and Verbiest T. Layer-by-Layer synthesis and tunable
optical properties of hybrid magnetic–plasmonic nanocomposites using short
bifunctional molecular linkers. Materials Letters 2014; 118: 99–102.
4. Bloemen M, Vandendriessche S, Goovaerts V, Brullot W, Vanbel M, Carron

S, Geukens N, Parac-Vogt T and Verbiest T. Synthesis and Characterization of
137
138 LIST OF PUBLICATIONS
Holmium-Doped Iron Oxide Nanoparticles. Materials 2014; 7: 1155–1164.
3. Bloemen M, Debruyne D, Demeyer P-J, Clays K, Gils A, Geukens N, Bartic

C and Verbiest T. Catechols as ligands for CdSe-ZnS quantum dots. RSC
Advances 2014; 4: 10208–10211.
2. Dupont D, Brullot W, Bloemen M, Verbiest T and Binnemans K. Selective

uptake of rare earths from aqueous solutions by EDTA-functionalized magnetic
and nonmagnetic nanoparticles. ACS Applied Materials & Interfaces 2014; 6:
4980–4988.
1. Bloemen M, Brullot W, Luong TT, Geukens N, Gils A and Verbiest T.

biomedical applications. Journal of Nanoparticle Research 2012; 14: 1100–1109.
Papers at international scientific conferences and symposia,

published in full in proceedings
5. Bloemen M, Denis C, Van Stappen T, De Meester L, Geukens N, Gils A and
Verbiest T. Multifunctional iron oxide nanoparticles for biomedical applications.
Proceedings SPIE 2015; Vol. 9338: 933816.
4. Van Stappen T, Lub J, Bloemen M, Geukens N, Spasic D, Delport F,

Lammertyn J, Verbiest T and Ann Gils. Transferability of antibody pairs
from ELISA to fiber optic-surface plasmon resonance for infliximab detection.
Proceedings SPIE 2015; Vol. 9317: 931705.
3. Bloemen M, Brullot W, Denis C, Vanysacker L and Verbiest T. Core-shell

nanoparticles as enhanced probes for imaging applications. Proceedings SPIE
2012; Vol. 8427: 84272Q.
2. Demeyer P-J, Bloemen M, Verbiest T and Clays K. Tuning the properties of

colloidal magneto-photonic crystals by controlled infiltration with superparam-
agnetic magnetite nanoparticles. Proceedings SPIE 2012; Vol. 8425: 84251R.
1. Demeyer P-J, Bloemen M, Verbiest T and Clays K. Engineering of colloidal

magneto-photonic crystals by infiltration with superparamagnetic magnetite
nanoparticles. Proceedings symposium IEEE Photonics Society Benelux 2011;
Ghent, Belgium:pp. 13-16.
LIST OF PUBLICATIONS 139
Meeting abstracts, presented at international scientific confer-

ences and symposia, published or not published in proceedings
or journals
6. Bloemen M, Van Stappen T, Knez K, Lammertyn J, Geukens N, Gils A,
Verbiest T. Multifunctional nanoparticles for biomedical applications. ISN2A.
Lisbon, Portugal, 20-22 January 2014.
5. Bloemen M, Van Stappen T, Lammertyn J, Geukens N, Gils A, Verbiest T.
Multifunctional iron oxide nanoparticles for biomedical applications. ISACS15
Challenges in nanoscience. San Diego, USA, 17-20 August 2014.
4. Bloemen M, Van Stappen T, Geukens N, Gils A, Verbiest T. Multifunctional
iron oxide nanoparticles for biomedical applications. Nanomedicine UK.
Edinburgh, UK, 26-27 March 2014.
3. Dupont D, Luyten J, Brullot W, Bloemen M, Verbiest T and Binnemans
K. Extraction and separation of rare earths with functionalized magnetic
nanoparticles. 1st European Rare Earth Resources Conference (ERES 2014).
Milos (Greece), 4-7 September 2014.
2. Bloemen M, Vandendriessche S, Brullot W, Verbiest T. Synthesis and
properties of holmium doped iron oxide nanoparticles. Collaborative Conference
on Materials Research (CCMR). Jeju Island, South Korea, 24-28 June 2013.
1. Brullot W, Strobbe R, Bynens M, Bloemen M, Demeyer P-J, Valev V and
Verbiest T. Hybrid Nanoparticle Networks (HyNANs): A modular approach to
multifunctional nanomaterials. Collaborative Conference on Materials Research
(CCMR) 2013. Jeju Island, South Korea, 24-28 June 2013, Abstract No. 521.

M Bloem en Thesis Final

Uploaded by

Copyright:

Available Formats

M Bloem en Thesis Final

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

M Bloem en Thesis Final

Uploaded by

Copyright:

Available Formats

Immunomagnetic separation of bacteria by iron oxide

Examination committee: Dissertation presented in partial

the joking around, dinners and mental support.

The development of multifunctional nanomaterials has proven to be beneficial

including severe types of pneumonia. Consequently their detection is mandatory

De ontwikkeling van multifunctionele nanomaterialen is voordelig gebleken voor

Legionella pneumophila bacteriën werden als modelorganismen gekozen voor de

RFP red fluorescent protein

1 General introduction to bioconjugated nanoparticles 1

2 Functionalization of oleic acid-coated iron oxide nanoparticles 31

3 Heterobifunctional PEG ligands for bioconjugation reactions 49

4 Two-step directional surface modification via protected siloxanes 63

5 Design of nanoparticles for efficient magnetic isolation of aquatic

6 Conclusions and outlook 101

A Selected experimental techniques 107

Health, safety and environment 134

List of publications 137

Figure 1.1: One dimension of a nanomaterial should be in the 1-100 nm range. As a

layer deposition, contact printing, photolithography or etching.9–11

1.1 Magnetic nanoparticles

A large variety of NP with magnetic properties have been described in scientific

all of these NP have interesting characteristics, including large saturation

1.1.1 General introduction to magnetism

The phenomenon of magnetism is electronic in nature, since it is related to

Figure 1.2: Schematic illustration of magnetic materials and their responses to an

the overall magnetic moment. As shown in Figure 1.2, this is an example of a

1.2 Synthesis of iron oxide nanoparticles

Method co-precipitation microemulsion hydrothermal thermal decomposition

Nucleation and growth

seeds is called heterogeneous nucleation. The latter process, where a single

During homogeneous nucleation (the in situ formation of nuclei) a large energy

which nucleation occurs and is therefore known as burst nucleation. Although

A large variety of NP can be produced by the thermal decomposition of

Figure 1.4: The decomposition of iron-oleate at high temperature happens via a

used iron-oleate as the organometallic precursor, which was decomposed in

1.3 Surface functionalization

Coating the surface of a nanoparticle is generally performed to improve the

The most common coatings described in literature are dextran, carboxymethy-

The introduction of inorganic shells on iron oxide NP can be performed from

Monomers and bilayers

1.4 Colloidal stability

The most common way of stabilizing NP is by making use of electrostatic

and the spontaneous oscillations of electrons that create temporary dipoles.

cannot end up in the primary minimum. This DLVO theory is dependent on

1.5 Protein coronas

behavior of the nanoparticle. Moreover the NP gain a biological entity that

in nature, the heterogeneity sometimes hinders their applicability in research.

Table 1.2: Comparison of monoclonal and polyclonal antibodies, including their

1.7 Coupling chemistry

The covalent attachment of proteins onto nanoparticles is a widely investigated

literature, including zero-length, homobifunctional, heterobifunctional and even

Figure 1.11: A typical reaction between a carbodiimide (1-ethyl-3-(3-

lead to severe protein-protein cross-linking. The NP were designed to have only

N,N’-carbonyl diimidazole (CDI) is very efficient activator of carboxylic acids.

Figure 1.13: Fluorescence image of L. pneumophila bacteria (recombinant strain

1.8 Legionella pneumophila

The Legionella bacteria is a Gram-negative, aerobic, rodlike organism that is

sources with pH of 2 to 8.90,91 Moreover they have been found in sources as

An outbreak of a severe type of pneumonia during an American Legion