Structure and Interactions of Human IgG-Fc - DiVA...
Transcript of Structure and Interactions of Human IgG-Fc - DiVA...
Linköping Studies in Science and Technology
Dissertation No. 1361
Structure and Interactions of Human IgG-Fc
Daniel Kanmert
Department of Physics, Chemistry and Biology
Linköping University, Sweden
Linköping 2011
1
Cover: The structure of an IgG antibody (PBD entry: 1IGT)
During the course of the research underlying this thesis, Daniel Kanmert was
enrolled in Forum Scientium, a multidisciplinary doctoral programme at
Linköping University, Sweden.
© Copyright 2011 Daniel Kanmert, unless otherwise noted
Kanmert, Daniel
Structure and Interactions of Human IgG-Fc
ISBN: 978-91-7393-241-7
ISSN: 0345-7524
Linköping Studies in Science and Technology, Dissertation No. 1361
Electronic publication: http://www.ep.liu.se
Printed in Sweden by LiU-Tryck, 2011.
2
Acceptance is freedom
3
4
v
Abstract
This thesis involves structure and interaction studies of the Fc fragment of
human IgG. For this purpose, hIgG-Fc of different subclasses were cloned and
expressed in the eukaryotic host Pichia pastoris, where relevant protein
modification at the post-translational level can be obtained.
Sometimes, changes in pH, temperature and salt concentration or addition of
moderate amounts of denaturants to a protein solution are associated with the
protein forming non-natively folded states, such as the molten globule or the A
state. IgG and some parts thereof are capable of forming another, so called
alternatively folded state, usually induced by acidification in the presence of
anions. This state is in many aspects related to the molten globule and the A
state but with distinguishing properties related mainly to chemical stability and
formation of oligomeric structures. The first part of this thesis describes two
different alternatively folded states of hIgG-Fc of subclass 4. One of them was
induced by decreasing the pH of the protein solution. Observed structural
changes were highly dependent on the concentration of sodium chloride. The
alternatively folded protein showed drastic changes in its secondary structure
compared to the native protein and significant tertiary structure was lost.
Moreover, it displayed an apparently increased chemical stability and had
surface exposed hydrophobic patches resulting in the formation of higher order
assemblies. In addition, it was shown for the first time that thermal induction of
an alternatively folded state is also possible, with similar, but not identical,
properties as the acid-induced state. Heat incubation for 20 hours at neutral pH
and at a physiological salt concentration further resulted in the formation of
protein aggregates. The dye Congo red had affinity for these aggregates, and
when viewed under polarized light, it showed green birefringence. They also
displayed binding of Thioflavin T and had a typical fibril appearance in the
transmission electron microscope. Hence, the formed aggregates share key
properties with structures constituting amyloid.
The second part of this thesis is focused on interactions of the Fc-fragment with
respect to both Fcγ-receptors on monocytes and the IgG autoantibody
rheumatoid factor. Immune complexes and their binding to Fcγ-receptors are of
pathogenic interest to rheumatoid arthritis. A surface mimic presenting full IgG
5
vi
molecules was designed as an in vitro immune complex model. Utilizing self-
assembled monolayers composed of alkanethiolates with different chemical
functionalities, the lateral IgG density could be tuned, enabling control of
monocyte interaction with the surface. Importantly, the IgG molecules were
homogeneously oriented to expose the Fc-fragment. The protein repellent
properties of these surfaces ensured that only differences in IgG concentration
determined variations in cellular adhesion. In a separate study the specificities of
IgG rheumatoid factor with respect to the different subclasses of hIgG-Fc were
investigated, using sera from patients with early rheumatoid arthritis. Strikingly
high IgG-RF reactivity against hIgG2-Fc was observed, together with raised
levels against hIgG1-Fc and hIgG4-Fc. No reactivity against hIgG3-Fc was
found.
6
vii
Populärvetenskaplig sammanfattning
Människokroppen består till cirka 70 procent av vatten. Resten, den så kallade
torrsubstansen, består av en lång rad olika ämnen. Ungefär 20 procent av dessa
utgörs av proteiner. Proteiner ansvarar för många olika funktioner, bland annat
transport av syre och uppbyggnad av kroppens alla organ. En viktig grupp av
proteiner är antikroppar, så kallade immunglobuliner. De utgör en del av
kroppens immunförsvar, ett system av olika proteiner och celler som har till
uppgift att identifiera och eliminera främmande substanser, till exempel
bakterier, virus och skadade celler.
Grundläggande för att ett protein ska kunna utföra sin funktion är att det kan
anta en specifik struktur. Olika typer av proteiner har olika strukturer. Den här
avhandlingen berör en del, ett fragment, av den mest förekommande
immunoglobulinen i kroppen (IgG) och har involverat undersökningar av
proteinets grundläggande förmåga att anta olika typer av strukturer. Om man
ändrar de fysiologiska betingelser som krävs för att proteinet ska existera i sin
rätta form, kan man tvinga det att anta strukturer med andra egenskaper. Sådana
inducerades i detta fall antingen genom att surgöra proteinprovet eller genom att
värma det. De alternativa proteinformerna karaktäriserades sedan med avseende
på sina biofysikaliska egenskaper och det visade sig att proteinet i dessa former
bildade lösliga proteinaggregat. Detta är intressant ur grundforskningssynpunkt
och bidrar till en ökad förståelse för proteiners förmåga att ändra sin struktur till
en som inte är funktionell, utan kanske rent av skadlig för kroppen. Vid
uppvärmning av IgG-fragmentet i 20 timmar bildades fibrösa strukturer. Sådana
fibrer, bildade från olika proteiner, förekommer i amyloida sjukdomar, till
exempel Alzheimers sjukdom och Parkinsons sjukdom. Fiberbildning från vissa
immunoglobulindelar är inblandad i relaterade sjukdomar. Denna avhandling
bidrar således till en ökad förståelse för vilka delar av immunoglobuliner som
kan forma dessa strukturer och möjligtvis vara inblandade i sjukdom.
Avhandlingen inkluderar även studier av hur IgG-fragmentet är delaktigt i
interaktioner med andra proteiner. När immunglobuliner interagerar med det
främmande ämne det är specifikt för så bildas så kallade immunkomplex. Dessa
har förmågan att reglera immunförsvarets reaktion på den främmande
substansen. Immunoglobuliner som ordnades med en väldefinierad orientering
7
viii
på en plan yta användes här för att skapa ett artificiellt immunkomplex ämnat
för studier av aktivering av vissa av immunförsvarets celler. Tätheten av
proteinet på ytan styrde antalet celler som interagerade och förankrade sig till
ytan.
Vid så kallade autoimmuna sjukdomar bildas antikroppar som är reaktiva mot
kroppsegna ämnen istället för främmande. Sådana antikroppar kallas för
autoantikroppar. En del av avhandlingen behandlar analys av en viss typ av
autoantikroppar, reumatoid faktor, involverade i sjukdomen ledgångsreumatism.
De är inblandade i komplexa interaktioner och det krävs mycket noggrant
utprovade metoder för att kunna detektera dem på ett pålitligt sätt. Ett test
utformades här för att undersöka delar av denna komplexitet. Resultatet visar i
huvudsak att hos en patientgrupp med tidig ledgångsreumatism binder
reumatoid faktor i olika grad till olika varianter av det studerade IgG-fragmentet.
8
ix
List of publications
This thesis is based on the following papers, which are referred to in the text by
their Roman numerals (I-V).
I D. Kanmert, A-C. Brorsson, B-H. Jonsson and K. Enander. “Thermal
Induction of an Alternatively Folded State of Human IgG-Fc”,
Biochemistry 2011, In press: doi 10.1021/b101549n.
II D. Kanmert and K. Enander. “Human IgG-Fc Forms an Alternatively
Folded State at Low pH”, In manuscript.
III D. Kanmert, A-C. Brorsson and K. Enander. “In Vitro Amyloid Fibril
Formation of Human IgG-Fc”, Submitted.
IV D. Kanmert, H. Enocsson, J. Wetterö, A. Kastbom, T. Skogh and K.
Enander. “Designed Surface with Tunable IgG Density as an in Vitro
Model for Immune Complex Mediated Stimulation of Leukocytes”,
Langmuir 2010, 26(5), 3493-3497.
V D. Kanmert, A. Kastbom, G. Almroth, T. Skogh, K. Enander and J.
Wetterö. “IgG Rheumatoid Factor Against the Four Human Fc-gamma
Subclasses in Early Rheumatoid Arthritis (the Swedish TIRA Project)”,
Submitted.
9
x
Contribution report
D.K. is the main author of the appended papers. D.K. planned and performed all
experiments, with the following exceptions: analytical ultracentrifugation (paper
I and II) was performed by B-H.J.; cell cultures and fluorescence microscopy
imaging (paper IV) were performed jointly by D.K. and H.E.; ELISA (paper V)
was developed by D.K. and performed by G.A.
10
xi
Publications not included in the thesis
VI K. Enander, L. Choulier, A. L. Olsson, D. A. Yushchenko, D. Kanmert,
A. S. Klymchenko, A. P. Demchenko Y. Mély and D. Altschuh. “A
Peptide-Based, Ratiometric Biosensor Construct for Direct
Fluorescence Detection of a Protein Analyte”, Bioconjugate Chemistry
2008, 19, 1864-1870.
VII O. Andersson, H. Nikkinen, D. Kanmert and K. Enander. “A Multiple-
Ligand Approach to Extending the Dynamic Range of Analyte
Quantification in Protein Microarrays”, Biosensors and Bioelectronics
2009, 24, 2458-2464.
11
12
xiii
Abbreviations
AFS Alternatively folded state
CD Circular dichroism
CD20 Cluster of differentiation 20
CHX Immunoglobulin constant heavy domain (X = 1, 2, 3)
ELISA Enzyme-linked immunosorbent assay
E. coli Escherichia coli
Fab Antigen binding fragment
FcR Fc receptor
GdmHCl Guanidinium hydrochloride
hIgGX-Fc Crystallizable fragment of human IgG (X = 1, 2, 3, 4)
IgX Immunoglobulin of isotype X (X = A, D, E, G, M, Y)
PBS Phosphate-buffered saline
P. pastoris Pichia pastoris
RA Rheumatoid arthritis
RF Rheumatoid factor
SAM Self-assembled monolayer
SPR Surface plasmon resonance
TEM Transmission electron microscopy
ThT Thioflavin T
UV Ultraviolet
VH Immunoglobulin variable heavy domain
VL Immunoglobulin variable light domain
13
14
xv
Contents
Preface ................................................................................................................ 1
1. Introduction ................................................................................................... 3
1.1 Protein structure......................................................................................... 3 1.2 Protein aggregation and amyloid fibrils..................................................... 8 1.3 Antibody structure ................................................................................... 10 1.4 Antibodies in disease ............................................................................... 13
1.4.1 Autoimmunity and rheumatoid factor............................................... 13 1.4.2 Antibody light (AL) and antibody heavy (AH) amyloidosis ............ 14
2. Alternatively folded states of IgG-Fc ......................................................... 17
2.1 Molten globules ....................................................................................... 17 2.2 Folded states at low pH............................................................................ 18 2.3 Alternatively folded states in human IgG-Fc........................................... 19
2.3.1 Thermal induction of an alternatively folded state in hIgG4-Fc....... 19 2.3.2 Acid induction of an alternatively folded state in hIgG4-Fc............. 23
2.4 In vitro amyloid fibril formation of human IgG-Fc ................................. 28 2.4.1 Acid-induced aggregation of hIgG4-Fc ............................................ 33
2.5 Summary.................................................................................................. 34
3. IgG-Fc interactions in disease..................................................................... 37
3.1 Fc receptors.............................................................................................. 37 3.2 IgG presenting surface as an in vitro immune complex-model ............... 38
3.2.1 Interaction between rituximab and a CD20-derived peptide epitope 38 3.2.2 Surface design and characterization.................................................. 40 3.2.3 Cell adhesion propensity for immune complex-model surfaces ....... 41
3.3 Assays for IgG rheumatoid factor............................................................ 42 3.4 IgG-RF subclass specificity in early rheumatoid arthritis........................ 44
3.4.1 IgM/IgA depletion and its effect on IgG-RF binding ....................... 45 3.4.2 IgG-RF subclass specific tests .......................................................... 47
4. Methodology................................................................................................. 49
4.1 Protein expression in Pichia pastoris...................................................... 49 4.2 Biochemical characterization techniques................................................. 51
15
xvi
4.2.1 CD spectroscopy............................................................................... 51 4.2.2 Fluorescence spectroscopy ............................................................... 52
4.3 ELISA...................................................................................................... 54 4.4 Amyloid fibril characterization techniques.............................................. 54
4.4.1 Thioflavin T binding......................................................................... 54 4.4.2 Congo red birefringence ................................................................... 55 4.4.3 Transmission electron microscopy ................................................... 56
4.5 Surface preparation and characterization................................................. 57 4.5.1 Self-assembled monolayers .............................................................. 57 4.5.2 Surface characterization by wettability, ellipsometry and SPR ........ 58
References ........................................................................................................ 61
Acknowledgement............................................................................................ 69
16
1
Preface
________________________________________________________________
This thesis is focused on structure and interactions of the Fc-fragment of one of
the most abundant proteins in human serum, IgG. This protein is important in
the body due to its ability to recognize foreign substances and its specific
binding action, initializing the removal of the foreign substance from circulation.
However, immunoglobulin function may also go awry, and become a cause of
disease.
The appended papers I – III describe some of the folding properties of the Fc-
fragment of human IgG. By changing specific conditions and destabilizing the
native structure, drastic conformational effects are observed. Paper IV describes
the use IgG in the design of artificial immune complexes, while paper V
presents the binding characteristics of the autoantibody rheumatoid factor
regarding its interactions with the different subclasses of the Fc fragment.
During the course of the research underlying this thesis I have had the great
fortune of working in an environment where multidisciplinarity colors the daily
experience. The work presented here was performed at the interface between
biochemistry, medicine, and molecular physics. Accordingly, the summarizing
chapters of this thesis should reflect this multidisciplinarity and are intended to
provide the reader with background and an overview of the appended papers.
17
18
3
1. Introduction
________________________________________________________________
1.1 Protein structure
Proteins are one of the largest, most versatile and complex class of biological
molecules. The correct folding of every protein within the cell is essential for
living systems to function. Proteins are fundamental for life and display a vast
number of functional properties. They can store and transport a wide variety of
molecules, ranging from large entities like DNA and RNA down to the size of
an electron. Proteins are vital for cell communication, cell adhesion and immune
responses and also crucial for eye-sight, smell and hearing. Proteins are
responsible for movement in living organisms; the compression and expansion
of muscles for example are due to forces in systems dominated by the protein
actin. Even the very synthesis of proteins, from gene expression to covalent
linkage of amino acids, is controlled by proteins.
Although being complex molecules in their folded conformations, proteins are
built up by a fairly simple condensation reaction. Amino acids constitute the
building blocks of a protein and they are linked together forming a rigid peptide
bond between the carboxyl group of one amino acid and the amino group of
another (Figure 1.1). Water is lost in every coupling step and each amino acid in
a protein is thus referred to as a residue. Part of the peptide backbone displays
the characteristics of a double bond as a result from the delocalized electrons
involving the carbonyl group and the amino nitrogen. This resonance is
19
4
responsible for the rigidity of the peptide bond. However, the bonds involving
the α-carbon have some rotational freedom, thus enabling the protein backbone
to be dynamic.
H2N
R1
O
OHH2N
R2
O
OHH2N
HN
OH
R1
O R2
O
H2Oαα
Figure 1.1 A peptide bond is the covalent chemical bond formed between two amino acids
when the carboxyl group of one molecule reacts with the amino group on the other
molecule, thereby releasing water.
The 20 naturally occurring amino acids constituting a protein all have different
properties distinguished by the functionality of the side-chains (Figure 1.1, R1
and R2). They can be divided into three main groups: polar, non-polar and
charged. When the amino acids are combined in different sequences, it gives rise
to polypeptides which are virtually infinite in variety. This first level of
complexity in the architecture of a protein is called the primary structure1
(Figure 1.2).
Figure 1.2 The primary structure is the sequence of amino acid residues constituting a
protein.
The next level of structure is the local folding of polypeptide backbone segments.
This is termed the secondary structure1 and it is stabilized by hydrogen-bonds
between amino hydrogens (donors) and carboxyl oxygens (acceptors) in the
backbone. The specific amino acid sequence determines the torsion angles Φ
and Ψ within the amino acid residues in a polypeptide. Repeating values of the
20
5
torsion angles along the polypeptide backbone results in regular structure.
Consecutive values of -60° and -50° of Φ and Ψ respectively, causes the
backbone to spiral and adopt an α-helical conformation (Figure 1.3, left). Each
completed turn in the helix corresponds to 3.6 amino acid residues and stretches
5.4 Å per turn. Hydrogen bonding occurs between the carboxyl oxygen on
residue n and the amino hydrogen on residue n+4, and the tight packing of
atoms within the helix allows for favorable van der Waals interactions. The side
chains protrude from the α-helix, thus eliminating steric interference with the
helical backbone.
Figure 1.3 Secondary structure elements (α-helix to the left and β-sheet to the right) in
proteins is a result from hydrogen bonds between repetitive backbone constituents. (PBD
entries: 2GP82 and 3KG03, respectively)
Another regular conformation of a polypeptide backbone is the β-sheet (Figure
1.3, right), with Φ and Ψ values of -110° to -140° and +110° to +135°
respectively. The basic unit of this structural element is the β-strand; an
extended stretch of amino acid residues, only marginally stable. When
incorporated into a β-sheet, favorable hydrogen bonding between the carboxyl
oxygens on one strand and the amino hydrogens on an adjacent strand stabilizes
the structure. Depending on the direction of the N- and C-terminals in the
polypeptide chains, the flanking β-strands can be parallel or antiparallel. The
side chain of each amino acid residue is rotated by 180° with respect to the
preceding one. A difference compared to the α-helix is that interactions between
the amino acid residue side chains may take place among those that are distant
in primary structure, whereas in the α-helix interactions usually take places
21
6
between amino acid residue side chains in close vicinity. In addition to the
regular elements found in the secondary structures, proteins can also contain
highly structured and stable elements lacking regularity. These elements often
form loops on the protein surface. Even though most proteins adopt stable
conformations, naturally unfolded proteins exist4, and can for instance fold upon
binding to an interactant5.
As a protein folds, the complexity of interactions increases dramatically. The
overall topology of the folded polypeptide chain is termed the tertiary structure6
(Figure 1.4, left). The tertiary structure is stabilized by several factors, including
hydrophobic clustering, ionic bonds, hydrogen bonds as well as van der Waals
interactions and covalent bridges. In an aqueous environment, soluble proteins
display their lowest free energy, i.e. the most stable conformation, when the
hydrophobic parts are buried and hydrophilic and charged parts are exposed on
the surface. In the unfolded state, water molecules are specifically ordered
around the polypeptide chain, coating the hydrophobic residues.
Figure 1.4 (Left) Tertiary structure of an immunoglobulin fold where the β-strands are
aligned into a sandwich shape (PBD entry: 1TEN7). (Right) The sum of the conformational
entropy, the enthalpy of internal interactions and the entropy of water exclusion yields in a
negative total free energy of folding of soluble proteins.
When the polypeptide chain folds the hydrophobic amino acid residues will
interact with each other by van der Waals interactions and are masked by other
parts of the protein, thus water is being excluded (hydrophobic effect)8. As the
22
7
protein forms its native structure, conformational entropy is lost (the factor
driving the protein to its unfolded state) and must be compensated for. Internal
interactions, including van der Waals interactions, (the enthalpy driving the
protein to its folded state) and favorable entropy contributions from water
exclusion result in the decrease of free energy (∆G) of folding of soluble
proteins8 (Figure 1.4, right) according to:
STHG ∆−∆=∆ (1.1)
Proteins can be divided into different structural and functional domains.
Different domains may be structurally diverse and display a variety of functions.
Association of protein units into complexes of more than one polypeptide chain
leads to the final level of protein organization, the quaternary structure6,9
(Figure 1.5). Such interactions can be either homotypic (when identical
polypeptides interact) or heterotypic (when different polypeptides interact),
leading to multi subunit proteins. The inter-chain interactions in multi subunit
proteins are stabilized by the same interactions as in the tertiary structure:
hydrophobic interactions, ionic bonds, hydrogen bonds, van der Waals
interactions and covalent bridges.
Figure 1.5 Ribbon representation of protein quaternary structure (PBD entry: 1IGT10). The
IgG molecule consists of four polypeptide chains; two heavy chains (light-blue and light-
orange) and two light chains (blue and orange).
23
8
1.2 Protein aggregation and amyloid fibrils
The folding of proteins into unique three-dimensional structures is central to
their biological functions in the body. However, the folded state of a protein is
transient and partial unfolding is constantly occurring. The partially unfolded
proteins are, due to exposure of groups otherwise hidden in the native structure,
susceptible to intermolecular interactions with other partially unfolded protein
units. Such interactions may in turn prevent the partially unfolded protein from
returning to the native state. Protein misfolding and aggregation may be an
extension of these folding actions11. Misfolding in a disordered manner leads to
the formation of amorphous aggregates, from which the protein is recruited to
the cellular degradation system. In vitro, amorphous aggregates may precipitate
from solution when they become large. In recombinant protein production,
formation of amorphous aggregates could be the mechanism behind
accumulation of inclusion bodies, an often occurring problem when expressing
mammalian proteins in bacterial expression systems.
Another pathway of aggregation occurs in a highly ordered manner, to form
amyloid fibrils. These structures are characterized by an extensive cross β-sheet
structure and are stabilized by hydrogen bonds between carbonyl and amide
groups in the peptide backbone12. Many proteins have been shown to be capable
of amyloid formation and all amyloid fibrils share the same overall morphology.
It is therefore believed that a vast number of proteins can form this state given
the appropriate conditions11,13. A general pathway for amyloid fibril formation is
ill ustrated in figure 1.6. The definition of in vivo amyloid (Nomenclature
Committee of the International Society of Amyloidosis) declares that there has
to be tissue deposits, with a distinctive fibril appearance under the transmission
electron microscope and a display of a characteristic X-ray diffraction pattern.
Further the deposits must have affinity for the dye Congo red and when viewed
under polarized light, it must show green birefringence. Significant binding to
ThT is also a common feature for amyloid fibrils14.
Mature amyloid fibrils are typically unbranched structures up to a few µm in
length and 5-10 nm in width. Moreover, they are very resistant to proteolytic
cleavage and highly tolerant toward denaturants.
24
9
Figure 1.6 Illustration of a pathway for amyloid fibril formation. Intermediate states can form
aggregated species that may be structurally amorphous and/or further assemble into
amyloid fibrils via a prefibrillar state.
The very first amyloidogenic protein to be observed was the antibody light
chain15, isolated from patients suffering from multiple myeloma. Since then,
many different proteins have been found to form amyloid fibrils in vivo.
Diseases related to amyloid deposits are classified either as localized or systemic.
In the localized type, the amyloidogenic protein is produced at the site where the
deposition occurs, whereas in systemic amyloidosis, the protein production is
distant from the deposits16.
25
10
1.3 Antibody structure
Antibodies17 (i.e. immunoglobulins) are globular proteins whose unique function
is to specifically bind to foreign substances invading the host organism, which
induces complex cascade-reactions with the final aim of eliminating the foreign
invaders. Antibodies are a part of the humoral immune system and are produced
by B lymphocytes upon stimulation. Due to great variation on the genetic level,
humans can generate around ten billion unique antibodies, each capable of
binding to individual epitopes.
The antibody structure1,18 comprises four individual units: two heavy chains
(Figure 1.7A, blue) and two light chains (Figure 1.7A, orange). The polypeptide
chains fold into consecutive structural domains, called the immunoglobulin fold
(Figure 1.4, left). This topology is composed by seven antiparallel β-strands
forming a sandwich-like structure. The fold occurs twice in the light chain and
four times in the heavy chain. The chains further assemble into a Y-shaped
quaternary structure (Figure 1.5 and 1.7A), which results in the antibody being
divided in three main parts: two Fab-regions and one Fc-region. Fab is short for
antigen binding fragment and is composed of the light chain (with the two
domains CL and VL) and the CH1 and the VH domains of the heavy chain. This
region is responsible for binding to different antigens through its hypervariable
parts. The base of the Y-shape corresponds to the Fc (crystallizable fragment)
and is the region which binds for example to cell-bound Fc-receptors and
components of the complement system. The subsequent events after such
interactions promote the elimination of the Ig-bound antigen.
Although the general structure of all antibodies is very similar, they are
produced in different varieties known as isotypes. In man, five Ig isotypes
(classes) exist: IgA, IgD, IgE, IgG and IgM. Human IgG is further divided into
four subclasses (IgG1, IgG2, IgG3 and IgG4) and IgA into IgA1 and IgA2. The
diversity in isotype and subclass gives rise to distinct functional properties and
roles in the immune system. The basic structural unit of an antibody is a
monomer, as for IgD, IgE and IgG. Higher order associations exist as well;
secretory IgA is a dimer, i.e. two monomers interconnected by a joining chain
and with a secretory component attached. IgM is produced as a pentamer, also
containing a joining chain.
26
11
A
B
Figure 1.7 (A) Schematic view of the main structural features of an IgG antibody. The
heavy chain is composed of the variable domain (VH) and three constant domains (CH1,
CH2, CH3), with N-linked glycans in the CH2 (black circles). The light chain is composed of
one variable domain (VL) and one constant domain (CL). The hinge region separates the Fc
and Fab fragments and inter- and intramolecular disulfide bonds (black lines) stabilize the
protein structure. (B) The number of intermolecular disulfide bonds in the hinge region
varies depending on IgG subclass.
27
12
An antibody may undergo several post-translational modifications, including
variations in glycosylation19, deamidation20, isomerization21 and oxidation22 and
also other, more global structural alterations, e.g. proteolysis23.
The IgG class is the most abundant in serum and constitutes ~80% of the total
Ig-content. The four subclasses of IgG are numbered according to the amount of
each subclass found on average in serum with [IgG1] > [IgG2] > [IgG3] >
[IgG4]. The four subclasses are similar in their primary structure. However,
small differences in sequence give rise to distinctly different properties. IgG1
and especially IgG3 are capable of activating complement, while IgG2 activates
only moderately and IgG4 not at all. Further, IgG of subclasses 1, 3 and 4 can
cross the placenta and are important for the fetus. There is also a large variation
in affinity in binding to FcγRs on phagocytic cells; IgG1 and IgG4 bind with
high affinity, IgG4 with intermediate affinity and IgG2 very weakly.
The most obvious structural differences between the four subclasses are related
to the size of the hinge region and the number of intermolecular disulfide bonds.
IgG3 has a four-fold longer hinge region compared to the other subclasses, with
eleven disulfide bonds (Figure 1.7B). IgG1, IgG2 and IgG4 have two, four and
four disulfide bonds, respectively. An interesting and remarkable feature of
IgG4 is its possibility of chain-exchange, resulting in an antibody capable of
binding to two different antigens at the same time24.
Specific antibodies are produced by injecting an animal with the antigen of
interest. The antibodies formed as result of the immunization process are
polyclonal, i.e. displaying heterogeneity in binding pattern against the antigen.
Monoclonal antibodies specific for a precise epitope on an antigen is produced
by isolating antibody-secreting lymphocytes from the animal fusing them with a
myeloma cell line and isolating a clone producing identical antibodies. The use
of both polyclonal and monoclonal antibodies is fundamental in research, e.g. in
immunoassays, affinity purification, western blotting and immunohistological
staining.
28
13
1.4 Antibodies in disease
1.4.1 Autoimmunity and rheumatoid factor
The immune system17 is a marvellously complex organization of different cell
types, antibodies and protein networks that constantly governs the body to find
and destroy foreign invaders. Being responsible for recognizing foreign antigens,
proper antibody function is essential, along with correct, subsequent immune
responses. Adaptive immune reactions occur as a result of the presentation of
foreign molecules (heteroantigens) to adaptive immune cells (T-cells and B-
cells). The subsequent T- and B-cell responses can lead to neutralization and/or
elimination of the foreign material, for instance due to specific antibody
formation against the heteroantigens and often, but not always, including
temporary inflammation. Early in life, the adaptive immune system develops
specific tolerance to autoantigens, i.e. “self-molecules” expressed in the own
body. If the immunological tolerance is lost, autoantibodies may be formed,
resulting in immune reactions against autoantigens. This may in turn lead to
reactivity against the body’s own tissues and result in chronic autoimmune
disease.
An example of a chronic inflammatory autoimmune disease is rheumatoid
arthritis (RA). When insufficiently treated, this disease causes permanent tissue
damage, joint deformities and functional impairments, resulting in a reduced
quality of life. Premature death, mainly due to cardiovascular disease, may be
another serious outcome25-28. This disease is associated with the occurrence in
serum of a family of autoantibodies called rheumatoid factor (RF) which is used
as a marker to classify RA-patients into distinct disease groups29 and was one of
the first autoantibodies to be described30. RF is reactive to the Fc portion of IgG
and can be of different isotypes: IgA-RF, IgG-RF and IgM-RF.
Characteristic antigenic epitopes for RF are located in the cleft between the CH2
and CH3 domains in IgG-Fc31. The first high resolution structure ever published
on an autoantibody in complex with its autoantigen is that of human IgG4-Fc
and a monoclonal IgM-RF Fab32 (Figure 1.8). The region in IgG-Fc involved in
RF binding is very similar to those involved in binding of the bacterial proteins
Staphylococcal Protein A and Streptococcal Protein G. The shared binding site
with RF explains the possibility of these proteins to inhibit RF activity and it has
29
14
also been speculated that the origin of RF may be due to anti-idiotypic networks
generating bacterial immunoglobulin-binding protein mimics33. Other RF target-
sequences have also been identified via epitope-mapping34-36 but they do not
posses any special auto-immunological potential and are overlapping with, or
flanking, epitopes for antibodies raised by immunization with IgG.
Figure 1.8 Cartoon of human IgG4-Fc (orange, one half of the biological unit) in complex
with a monoclonal IgM-RF Fab (blue). Spheres represent the binding interface, located in
the cleft between the CH2 and the CH3 domains. (PDB entry: 1ADQ32)
1.4.2 Antibody light (AL) and antibody heavy (AH) amyloidosis
Antibody light chain (AL) amyloidosis is the most common type of systemic
amyloidosis, and an extensive number of amyloidogenic light chain sequences
have been published37,38. In AL amyloidosis the light chain or fragments thereof
are produced by plasma cells in the bone marrow. Both λ and κ chains can form
amyloid with the λ-type being of higher prevalence in an approximate ratio of
3:1. It is difficult and time-consuming to properly diagnose this disease; renal
dysfunction (kidney failure), hepatomegaly (enlarged liver) and heart failure are
the primary symptoms39-41. In contrast to the vast clinical and molecular
30
15
characterization of AL amyloidosis, only a small number of antibody heavy
chain (AH) amyloidosis cases have been described42-50. This type of amyloidosis
is associated with the same symptoms as in AL amyloidosis. To most part,
sequence identification has confirmed that (part of) the VH and CH3 domains are
engaged in this type of disease43-45,47-50.
31
16
32
17
2. Alternatively folded states of IgG-Fc
________________________________________________________________
2.1 Molten globules
As every individual polypeptide chain undergoes the folding process, it may
pass through a nearly infinite number of possible conformations in order to
finally fold into its native structure. Throughout this process proteins can form
folded, non-native structures with lifetimes sufficiently long to make them
amenable for characterization. The molten globule is one of the most intensively
studied of these intermediates51. It can be defined as an ensemble of non-
natively folded conformations, with an average degree of secondary structure
similar to that of the native protein, yet loose packing within the hydrophobic
core results in hydrophobic residues with greater solvent exposure52,53. Tertiary
contacts are poor due to the lack of tight packing of amino acid residue side
chains and the overall radius of gyration is increased by about 20% compared to
that of the natively folded protein51. Further, the molten globule state displays
low stability toward chemical denaturation and unfolds in non-cooperative
transitions. Induction of the molten globule ensemble in vitro often occurs
through changes in temperature or pH or by exposure to relatively low
concentrations of a denaturant, leading to destabilization of the native state.
33
18
2.2 Folded states at low pH
When a protein is incubated at low pH, pH-sensitive amino acid side chains are
protonated which can have a drastic effect on the native protein structure, thus
leading to denaturation54. The cause for denaturation is electrostatic repulsion
among the charged side chains. When acidifying a protein solution at a low ionic
strength, the resulting conformation is often a randomly coiled unfolded
structure, although it has been shown that acid-induced denaturation often gives
a less unfolded protein than chemical denaturation induced with e.g. GdmHCl or
urea54. This may be due to the force of electrostatic repulsion being weaker than
the forces stabilizing the tertiary structure, which includes mainly hydrophobic
interactions (including van der Waals interactions), ionic bonds and disulfide
bonds. However, due to the ability of anions to decrease the overall charge
repulsion at low pH, addition of various stabilizing salts may induce a folded,
non-native conformation55,56. Under these conditions, a molten globule is often
observed, but formation of so called A states have also been reported57,58. The A
state shares many properties with the molten globule in that it displays a higher
exposure of hydrophobic residues compared to the native state and chemically
(and thermally) unfolds in non-cooperative transitions. It has been claimed that
different A states differ substantially in compactness of structure, and the same
protein can form a diverse set of A states57. Importantly, the A state is not
always globular in structure. Many proteins have been characterized based on
their propensity to form this state. For example, it has been demonstrated that
different anions can induce three different A states in the Stapylococcal
nuclease57. These different A states vary in size and compactness and display
hydrophobic patches to a different extent. The native secondary structure is
preserved to 50, 70 and ~100%, respectively, although lack of tertiary contacts
characterizes all three states57. Similarly, three different A states have been
observed in apomyoglobulin58.
Another folded state induced by low pH is the so called alternatively folded state
(AFS). The AFS is so far restricted to IgG-related proteins and murine IgG was
one of the first proteins described to form this folded structure59. Later, more
proteins derived from IgG (Fab, VL, VH, CH1 and CH3) have been shown to
adopt the AFS60-63. The AFS is a folded non-native structure distinct from both
the classical molten globule and the A state. The alternatively folded IgG or
IgG-derived proteins share many proerties with proteins forming a molten
34
19
globule or an A state. For instance, they display hydrophobic patches as a result
from a less compact protein conformation, along with loss of rigid tertiary
structural contacts. However, both the molten globule and the A state are less
stable than the native state, reflected in the lower tolerance toward chemically
induced unfolding which often occurs in non-cooperative transitions. In contrast,
the AFS displays a higher chemical stability compared to the native state and,
importantly, unfolding from the AFS occurs in cooperative transitions. The A
state and in particular the molten globule display secondary structure similar to
the native state, while it is dramatically changed in the AFS. In addition, the
formation of the AFS is often accompanied by drastic changes in quaternary
structure through the formation of oligomers.
AFS formation has previously only been described under acidic conditions;
however, this thesis expands the view of when IgG-derived proteins can form
this structural state. Paper I describes the AFS of hIgG4-Fc, induced by heating
to 75°C at neutral pH and at a physiological salt concentration. Section 2.3.1
presents the essential properties of the thermally induced AFS of hIgG4-Fc.
Paper II presents another AFS of hIgG4-Fc induced by the classical conditions
involving acidification in the presence of NaCl. The acid-induced AFS is
described in section 2.3.2.
2.3 Alternatively folded states in human IgG-Fc
2.3.1 Thermal induction of an alternatively folded state in hIgG4-Fc
Heating of all-β-sheet proteins may result in significant structural changes on the
secondary and tertiary levels64. In some cases heating causes structural
rearrangements into an α-helical structure which can be both reversible and
irreversible65,66. Heating of hIgG4-Fc results in total disruption of the native
structure reflected at secondary, tertiary and quaternary levels. The far-UV CD
spectrum of native hIgG4-Fc displays a dip in negative ellipticity at 216 nm that
is characteristic of β-sheet proteins67 (Figure 2.1B, solid line). When monitoring
the mean residue ellipticity at 216 nm during temperature ramping, a dramatic
increase in negative ellipticity between 60°C 75°C (Figure 2.1A) indicates a
cooperative structural rearrangement that is not equivalent with the formation of
a randomly coiled structure. Rather, it can be interpreted as an increase in β-
35
20
sheet structure, as evident from the full far-UV spectrum at 75°C (Figure 2.1B,
dashed line). The structural changes induced by heating can be regarded as an
irreversible process. This is evident from the spectrum in figure 2.1B (dotted
line) of the sample re-equilibrated at 25°C which shows the same overall shape,
although slightly decreased in negative ellipticity.
A
B
Figure 2.1 Far-UV CD analysis of hIgG4-Fc. (A) Mean residue ellipticity monitored at 216
nm during temperature ramping. (B) Full far-UV CD spectra recorded ay 25°C (solid line),
after equilibration at 75°C (dashed line) and after re-equilibration at 25°C (dotted line).
Because of the large number of tryptophan residues evenly distributed
throughout hIgG4-Fc, monitoring of the changes in intrinsic fluorescence is a
36
21
sensitive way to probe global structural alterations on the tertiary level. The red
shift of the wavelength at maximum intensity was monitored during temperature
ramping (Figure 2.2). It was revealed that heating is associated with the
formation of a less compact protein conformation where tryptophan residues are
accessed by solvent molecules. When thermal denaturation monitored by a
peptide backbone sensitive technique (far-UV CD) was compared to monitoring
with an environmentally sensitive technique (intrinsic tryptophan fluorescence)
it was revealed, as expected, that the tertiary structural contacts are affected
before the secondary structure. The midpoint of transition for denaturation of the
peptide backbone is shifted by 8°C toward higher temperatures.
Figure 2.2 Wavelength at maximum intensity of the intrinsic tryptophan fluorescence from a
sample of hIgG4-Fc monitored during temperature ramping. Solid line represents least-
square fit to a two-state model.
To investigate whether the structural changes in secondary and tertiary structure
were associated with hydrophobic patches being exposed on the protein surface,
binding of the probe ANS was investigated. When ANS binds to hydrophobic
regions in proteins, its emission intensity increases, along with a characteristic
blue shift of the wavelength at maximum intensity of the fluorescence signal68.
Indeed, when heating hIgG4-Fc the ANS fluorescence starts to increase at
around 50°C to reach a maximum at 68°C (Figure 2.3).
37
22
Figure 2.3 ANS fluorescence emission from a sample of hIgG4-Fc monitored during
temperature ramping.
An important feature that distinguishes the AFS from a classical molten globule
is that it displays an apparently higher stability compared to the native protein,
as reflected in the increased tolerance toward GdmHCl unfolding (Figure 2.4).
Importantly, hIgG4-Fc (heat treated for 10 min at 75°C) unfolds in a cooperative
transition, although somewhat less cooperatively compared to the native protein.
Another notable aspect of this structural state is that heat-treated hIgG4-Fc
forms higher order assemblies, dominated by an equilibrium between
monomeric and heptameric species, with 20% of the protein in heptameric form
at the given conditions. It is reasonable to expect that the solvent-exposure of
hydrophobic residues is responsible for interchain interactions between
individual hIgG-Fc units, thus promoting oligomerization.
In summary, heating of hIgG4-Fc to 75°C involves the formation of a state
which structurally shares many of the characteristics of the AFS rather than a
molten globule or a randomly coiled unfolded structure. This section has
focused on hIgG-Fc of subclass 4. Indeed, initial analysis of the ability of other
subclasses to form this structural state in terms of far-UV CD spectral behavior
at elevated temperatures suggests that this may not be a general feature among
all subclasses. A further subclass-wide characterization is warranted to explain
the determinants for AFS-formation at elevated temperatures.
38
23
Figure 2.4 Stability of hIgG4-Fc toward GdmHCl-induced unfolding. Unfolding transitions
monitored with intrinsic tryptophan fluorescence of the native protein (□) and heat-treated
protein (75°C for 10 min) after re-equilibration at 25°C (▲). Solid lines represent least
square fits.
2.3.2 Acid induction of an alternatively folded state in hIgG4-Fc
The AFS of the murine IgG1 antibody domain CH3 has been extensively
characterized by Buchner and colleagues69. This protein is known to form an
AFS upon acidification in the presence of anions. The CH3 domain is composed
of an antiparallel β-sheet immunoglobulin fold stabilized by an intra-molecular
disulfide bridge. Native CH3 is a dimer with a molecular weight around 26,000
Daltons and has a typical β-sheet appearance in far-UV CD. Upon acidification
of the protein in solution at a low ionic strength, CH3 completely unfolds into a
random structure. However, the addition of anions induces drastic structural
rearrangements in the protein, yielding a conformation with distinctly different
secondary structure when compared with the native state. Further, the AFS of
murine CH3 involves the formation of a defined oligomeric state of a dodeca- to
tetradecamers and importantly, it displays an apparently higher stability toward
chemically induced unfolding, proceeding in cooperative transitions. Because
CH3, which constitutes on half of IgG-Fc, forms an AFS at acidic pH it was
relevant to investigate the entire Fc-fragment in this respect.
Paper II describes a second AFS of hIgG4-Fc, beside the thermally induced state
described in section 2.3.1. Presented here is the formation of an AFS by
acidification in the presence of anions. This AFS also shares key properties with
39
24
the states previously described for the murine CH3 and other IgG-derived
proteins60. This section gives an overview of the most important structural
characteristics.
When decreasing the pH from 7.4 to 2.0 at a physiological salt concentration
there is a pronounced blue-shift of the minimum of the negative ellipticity in the
far-UV CD spectrum from 216 nm to 206-208 nm (Figure 2.5, dashed line). It is
evident that this reflects a structural rearrangement that is distinct from the
formation of a random coil, as well as from the native structure (Figure 2.5,
solid line). The thermally induced AFS of hIgG4-Fc forms in an irreversible
process, while the CH3 dimer displays a fully reversible transition62.
Reversibility was investigated for the acid induced AFS of hIgG4-Fc; after re-
equilibration at 7.4 only partial refolding to the native state was observed
(Figure 2.5, dotted line). The difference in reversibility when comparing hIgG4-
Fc and CH3 may be explained by the difference in size of the two proteins.
Possibly, it may be more difficult to find the refolding pathway in the larger Fc-
fragment.
Figure 2.5 Far-UV CD analysis of hIgG4-Fc. Spectra recorded in PBS pH 7.4 (solid line),
after dialysis to PBS pH 2.0 (dashed line) and after dialysis back to PBS pH 7.4 (dotted
line).
Acidification of hIgG4-Fc in the absence of NaCl results in what appears to be
an acid-unfolded randomly coiled structure, as judged by the far-UV CD
spectrum (not shown). The structural transition to the AFS is highly dependent
40
25
on anion concentration. Titration with NaCl reveals that AFS formation occurs
in a broad interval of concentrations, with a midpoint of transition at 28 mM
(Figure 2.6A). In contrast, with the CH3 domain the corresponding transition is
highly cooperative and has a midpoint at ~50 mM NaCl62. Structural changes are
fast and occur immediately during stepwise addition of NaCl.
A
B
Figure 2.6 (A) Titration of hIgG4-Fc at pH 2.0 with NaCl monitored by Far-UV CD. Inset
shows spectra (mean residue ellipticity versus wavelength) at 140, 50, 10 and 0 mM NaCl
(from top). (B) pH dependence of the formation of the alternatively folded hIgG4-Fc at 140
mM NaCl. Inset shows spectra (mean residue ellipticity versus wavelength) at pH 7.5, 3.5
and 2.0 (from top).
41
26
Upon decreasing the pH of hIgG4-Fc solutions with 140 mM NaCl (Figure
2.6B), the protein retains its native structure until pH 4.5, followed by a
cooperative structural transition with similarities to that of the CH3 domain. The
midpoint of transition occurs at pH 3.3 and at pH values ≤ 2.0 the protein has
fully adopted the AFS. For the CH3 domain however, this is observed at slightly
higher pH values (with a midpoint at pH 4.0 and a fully developed AFS at pH ≤
3.8).
To further investigate structural changes on the tertiary level, analysis of
environmental changes in the vicinity of the eight tryptophan residues upon
acidification was performed. At pH 2.0 in the presence of NaCl there is a slight
red shift of the wavelength at maximum intensity compared with the emission
spectrum of native hIgG4-Fc (Figure 2.7). This indicates the formation of a less
compact protein structure where loose packing of the amino acid residue side
chains result in solvent exposure of tryptophans. Tryptophan residues in native
proteins are often quenched internally by local contacts with either disulfide
bridges or each other70. In hIgG4-Fc, four of the tryptophan residues are in the
proximity of disulfide bridges32. Physical separation of these groups resulting in
reduced quenching is the probable explanation for the accompanying increase in
fluorescence intensity observed at pH 2.0
Figure 2.7 Intrinsic tryptophan fluorescence from hIgG4-Fc in PBS pH 7.4 (solid line) and
PBS pH 2.0 (dashed line).
42
27
In line with a less compact protein structure, addition of ANS to hIgG4-Fc under
AFS forming conditions clearly demonstrates significant binding of the probe to
the protein in this state (Figure 2.8, dashed line). At these instrumentation
settings no ANS binding to either the native protein or to the chemically
unfolded protein is observed (Figure 2.8, solid and dash-dot lines, respectively).
This implies extensive exposure of hydrophobic patches on the surface of
hIgG4-Fc at pH 2.0.
Figure 2.8 ANS emission spectra recorded in PBS pH 7.4 (solid line), PBS pH 2.0 (dashed
line) and in 5 M GdmHCl (dash-dot line).
Acidification of proteins in general can have different effects on the structural
conformation and function71. Some proteins lose all their native contacts and
unfold into a random structure, while some are sufficiently stable to retain the
native conformation. Others adopt non-native conformations with or without
preserved activity72-74. At pH 2.0, hIgG4-Fc displays many of the properties
observed for the A state and the classical molten globule, but significant
differences exist. The apparent increase in chemical stability, as deduced form
data in figure 2.9, is an important feature distinguishing the present state from
the A state. Oligomerization (resulting in a dimer – dodecamer equilibrium), in
all probability attributable to interactions between exposed hydrophobic regions
among individual hIgG4-Fc units, further shows that this protein fulfils the
criteria of forming an AFS at pH 2.0 in the presence of NaCl.
43
28
Figure 2.9 Stability of hIgG4-Fc toward GdmHCl-induced unfolding monitored by intrinsic
tryptophan fluorescence. Unfolding transitions of the native protein at pH 7.4 (□, same
dataset as in figure 2.4) and at pH 2.0 (▲) are shown. Solid lines represent least square fits.
2.4 In vitro amyloid fibril formation of human IgG-Fc
As described in section 2.3.1, thermal denaturation of hIgG4-Fc results in the
formation of an AFS. Heating of proteins has been associated with the formation
of amyloid structures and certain antibody fragments have been identified in
different kinds of amyloidosis. The possibility of hIgG4-Fc being capable of
amyloid fibril formation was therefore investigated in paper III. This chapter
presents the effects of prolonged incubation of hIgG4-Fc at 75°C, along with an
investigation of the amyloidogenic properties of this protein.
Due to the extensive characterization of hIgG4-Fc regarding AFS-formation in
Papers I and II, this subclass was at special focus also in Paper III on the
formation of amyloid fibrils. In vitro formation of amyloid fibrils has been
claimed to be a generic property of proteins11 and is not only restricted to those
related to amyloidosis13. It often involves destabilization of the native protein by
low pH, heating or exposure to moderate amounts of denaturant75. In vitro
amyloid formation of antibody light chain fragments has been obtained at both
neutral and acidic pH and with or without the addition of denaturants, typically
at 37°C76-81. Agitation is sometimes a prerequisite. Here, amyloid formation was
obtained by quiescent incubation at neutral pH and at a physiological salt
concentration at 75°C.
44
29
Binding of ThT to hIgG4-Fc (Figure 2.10) was monitored during quiescent
incubation at 75°C and the binding profile greatly resembles that of the common
lag-growth-plateau trajectories displayed by many amyloidogenic proteins82-84.
The lag-phase during the first 20 min of incubation is followed by a relatively
long growth-phase to finally reach a plateau after ~20 h. It is likely that
nucleation within the first 20 min is the rate limiting step.
Figure 2.10 ThT fluorescence from a sample of hIgG4-Fc and ThT monitored during
incubation at 75°C. The first data point correspond s to the ThT signal for the native protein
before heating.
Concurrently with the analysis of ThT binding, TEM was used to investigate the
morphologies of the different aggregates forming upon heating. During the
growth-phase there is great variation in morphology. Already in the period of
0.5 – 2 h mature fibrils along with smaller oligomeric species (Figure 2.11B) are
present but more amorphous assemblies (Figure 2.11A) and large ribbon-like
structures (Figure 2.11C) can also be observed. Reaching the plateau, at 20 h,
more homogeneous structures are obtained, with mature fibrils being the
dominating species (Figure 2.11D). The fibrils formed after 20 h further
assemble into twisted conformations made up by several fibril units (Figure
2.11E). Interestingly, the early formed aggregates with apparent morphological
disorder mature into fibrils and are no longer visible in the TEM.
45
30
Figure 2.11 Transmission electron micrographs of hIgG4-Fc after incubation at 75°C for
0.5-2 h (A-C) and 20 h (D-E). (Bars: A-B: 200 nm, C: 2 µm, D-E: 500 nm.)
The significant increase in ThT-binding to hIgG4-Fc after 20 h incubation
compared to binding to the native protein (Figure 2.10) classifies the fibrils as
thioflavinophilic, a characteristic property of amyloid structures85. Although not
being considered one of the “defining” criteria for amyloid, ThT binding is still
a common method to probe its formation. More established in amyloid
characterization is green birefringence upon staining with Congo red85. After 20
h incubation, green birefringence is indeed observed in the polarization
microscope (Figure 2.12), thus qualifying the hIgG4-Fc to be denoted amyloid.
Figure 2.12 Micrographs of hIgG4-Fc stained with Congo red after incubation at 75°C for
20 h with aligned (A) and crossed (B) polarizers. (Bars: 10 µm).
46
31
The sequence and native three-dimensional structure of the different hIgG-Fc
are to a great extent very similar. Interestingly however, heating of the four
different subclasses of hIgG-Fc results in subclass-specific changes of the
secondary structure. This is reflected in their far-UV CD spectra (Figure 2.13A).
A
B
Figure 2.13 (A) Far-UV CD analysis of of hIgG1-Fc (solid line), hIgG2-Fc (dashed line),
hIgG3-Fc (dotted line) and hIgG4-Fc (dash-dot line) after heating to 75° and re-equilibration
at 25°C. (B) Mean residue ellipticity at 216 nm of hIgG1-Fc (black), hIgG2-Fc (dark grey),
hIgG3-Fc (medium grey) and hIgG4-Fc (light grey) in PBS pH 7.4 monitored during
temperature ramping.
47
32
The spectrum of heat-treated hIgG4-Fc displays a dip at 214 nm while the
spectra of hIgG2-Fc and hIgG3-Fc display dips at 206 nm and 203 nm,
respectively, indicating a more pronounced contribution from the random coil
structure. The spectrum of heat-treated hIgG1-Fc does not display a well-
defined minimum between 200 and 230 nm. When monitoring the thermal
denaturation of the different hIgG-Fc subclasses at 216 nm (Figure 2.13B), there
is a dramatic increase in negative ellipticity between 60°C and 75°C for hIgG-Fc
of subclasses 2-4 while the same transition is shifted ~10°C toward higher
temperatures for hIgG1-Fc.
When subclasses 1-3 were subjected to quiescent incubation at 75°C for 20 h
fibrils of hIgG2-Fc and hIgG3-Fc were observed (Figure 2.14C and D,
respectively). Neither fibrils nor disordered aggregates could be detected for
hIgG1-Fc under these conditions (Figure 2.14A). The profile of the far-UV CD
denaturation curve for this particular subclass is shifted 10°C toward higher
temperature (Figure 2.13B, black line). Indeed, quiescent incubation at 85°C
during 20 h results in fibril formation (Figure 2.14B). The hIgG1-Fc fibrils are
in the 0.5 µm range and shorter than those obtained with hIgG-Fc of subclasses
2-4 which are up to three to four µm in length. Despite the differences in
structure upon heating of these proteins, all subclasses are de facto capable of
forming fibrils.
Figure 2.14 Transmission electron micrographs of hIgG1-Fc after incubation at 75°C (A)
and at 85°C (B) and hIgG2-Fc (C) and hIgG3-Fc (D) a fter incubation at 75°C. All samples
were incubated for 20 h. (Bars: 500 nm.)
48
33
2.4.1 Acid-induced aggregation of hIgG4-Fc
As both heating and acidification of hIgG4-Fc is associated with the formation
of an AFS, and as prolonged heating results in amyloid fibrils, the tendency of
hIgG4-Fc folded into the acid-induced AFS to form amyloid fibrils was
investigated. In vitro amyloid fibrils of IgG light chain fragments have been
obtained at low pH, at 37°C and with agitation. Indeed, when incubating hIgG4-
Fc under these conditions, thioflavinophilic (Figure 2.15B) and congophilic
structures (Figure 2.15C-D) are formed after eight hours. In accordance with the
pH-dependence of AFS-formation monitored by far-UV CD (Figure 2.6B), the
ThT binding to hIgG4-Fc increases with lower pH values (Figure 2.15A). In
contrast to heat-incubation (20 h) which results in mature amyloid fibrils,
incubation at low pH, 37°C with agitation does not result in fibrillar structures
(Figure 2.15E). Rather, the formed aggregates can be considered pre-fibrillar in
nature.
Figure 2.15 ThT binding to hIgG4-Fc after incubation at different pH values (A) and during
prolonged incubation in PBS pH 2.0 with agitation at 37°C (B). Solid lines represent least
square fits. Micrographs of hIgG4-Fc stained with Congo red with open (C) and crossed (D)
polarizers (Bars: 10 µm) and a transmission electron micrograph (E) (Bar: 200 nm), all
recorded after 8 h incubation, are also shown.
49
E
34
2.5 Summary
The IgG molecule is an important protein in the host defense of mammalian
species. Being a part of the humoral immune system it relies on proper folding
and stability to be able to perform its function. The different human IgG
subclasses are to a great extent similar in sequence as well as in the secondary,
tertiary and quaternary structures. The physical properties and chemical stability
of antibodies has been extensively studied for many years. Of special interest to
this thesis, it has been shown that they may undergo aggregation and pH and
thermally induced conformational changes.
Heating of proteins may result in unfolding; however, many proteins undergo
drastic structural rearrangements at the secondary and tertiary levels. Similarly,
acidification of protein solutions often results in the adoption of non-native
folded states. These conformations often share key properties with the classic
descriptions of the molten globule and anion-induced A states. In contrast,
exposure of IgG and parts thereof to acidic conditions leads to another state first
described for the intact IgG molecule, the AFS59. This state is distinguished
from the classical molten globule and the A state by its remarkable stability
against chemical denaturation which often proceeds in cooperative transitions.
Further, oligomerization of AFS folded proteins is often observed.
Paper I describes for the first time that acidification in the presence of anions is
not an absolute prerequisite of AFS-formation, but that this state can also be
induced by heating at neutral pH. Paper II contributes to further mapping of
which individual IgG-domains and fragments that can form an AFS under the
previously studied acidic conditions. Even though analysis of intact IgG, IgG-
Fab and most of the IgG domains has resulted in a general knowledge on the
fundamentals of the AFS, the entire Fc-fragment has never been investigated in
that respect. Interestingly, hIgG4-Fc is capable of forming at least two
structurally distinct alternative folds, the properties of which are in agreement
with what has been reported for AFS by Buchner and colleagues60,62. The
thermally induced AFS of hIgG4-Fc displays a well-defined dip in its far-UV
CD spectrum at 214 nm, which indicates structural similarities on the secondary
level with the AFS for murine CH3, CH1, VH, VL and Fab, that also have their
dips around the same wavelength60,62. The dip in far-UV CD for the intact
murine IgG molecule at pH 2 in the presence of anions is slightly more blue-
50
35
shifted in comparison59, indicating more contribution from the random coil. The
acid induced AFS of hIgG4-Fc shows an even further blue-shift, with a dip at
206-208 nm.
Table 2.1 Summary of data showing conformational differences in hIgG4-Fc.
Dip in far-UV CD spectrum
Tryptophan λmax
Cma
Oligomeric stateb
Native 216 nm 338 nm 1.5 M monomer
Heat induced AFS 214 nm 348 nm 1.9 M monomer, 7-mer
Acid induced AFS 206-208 nm 343 nm 2.2 M dimer, 12-mer
astability paramater obtained from GdmHCl-induced unfolding, bmonomer refers to the dimer of the Fc-fragment.
The effect of AFS formation on the tertiary level is even more protein specific.
VL and CH1 display a wavelength of maximum intrinsic tryptophan fluorescence
at ~350 nm, in line with the lack of tertiary contacts determined from their near-
UV CD spectra60. The near-UV CD signals are to a great extent lost during AFS
formation of CH3 as well, and the intrinsic fluorescence displays its maximum at
343 nm, which is in accordance with the acid induced AFS of hIgG4-Fc. The
thermally induced AFS is red shifted further, to 348 nm, indicative of the
tryptophan residues being more solvent exposed. Interestingly with VH, the
alterations in the tertiary structure are different compared to the other cases. The
near-UV CD signals indicate significant tertiary contacts present at pH 2 in the
presence of anions. Further, the tryptophan fluorescence spectrum does not shift,
but displays maximum intensity at 338 nm60, a wavelength associated with
natively folded proteins.
Increase in apparent stability toward chemical denaturants is a general property
of an AFS. This is also the case with hIgG4-Fc as revealed by the higher Cm-
values obtained from the unfolding curves. Notably, the increase is more
pronounced for the acid induced AFS. As unfolding probably involves
dissociation of oligomers, this may be explained by the fact that acid-incubated
hIgG4-Fc is oligomerized to a much lager extent than the heat-incubated protein.
51
36
Formation of exposed hydrophobic patches in all described AFS-folded proteins
so far is associated with the assembly of different oligomeric states involving
~5-40 monomeric units. Investigation of the two alternative folds of hIgG4-Fc
revealed that in PBS at pH 2 there is an equilibrium between dimer and
dodecamer (with 60% of the protein in the dodecameric form) while the
thermally induced AFS is associated with a monomer-heptamer equilibrium
(with 20% of the protein as heptamers). Changes in the quaternary structure of
many IgG light chain fragments have been associated with the formation of
higher order assemblies. Amyloid fibrils of light chains are deposited in various
organs of patients with AL amyloidosis. A significant number of light chain
sequences have been accounted for as amyloidogenic, both in vivo and in vitro.
In contrast, only a few heavy chain fragments have been isolated from patients,
and AH amyloidosis is rare. Paper III reports that in vitro fibrillation is also
possible for the Fc-fragment of all human IgG subclasses. The findings in Paper
III broaden the view on which parts of the IgG molecule that can undergo
fibrillation. Considering the fact that IgG-CH3 has been found in in vivo amyloid
deposits, Paper III is an important contribution which invites further studies on
the possibility of involvement of the Fc-fragment in disease.
52
37
3. IgG-Fc interactions in disease
________________________________________________________________
3.1 Fc receptors
The opsonization of an antigen by immunoglobulins is one of the important
processes underlying the removal of an antigen from circulation in the body.
Cells expressing specific Fc-receptors (FcRs) on their surface are responsible for
this immunological response by binding to the Fc part of the immunoglobulin17.
This process initiates e.g. phagocytosis of the foreign antigens recognized by
antibodies. Besides phagocytosis of antibody-opsonized particles or endocytosis
of antibody-antigen immune complexes, other mechanisms are activated, e.g.
plasma cell antibody production, antibody dependent cell-mediated cytotoxicity
(ADCC) and secretion of pro-inflammatory cytokines that in turn activate other
components of the immune system. Ligation of some FcRs can, on the contrary,
exert down-regulatory/anti-inflammatory effects.
Immunoglobulins are recognized by FcRs in an isotype specific manner; IgG is
recognized by FcγRs, IgA by FcαRs and IgE by FcεRs86. Hence, in the
interaction of an immune cell with IgG-opsonized particles and soluble antigens,
FcγRs are engaged. FcγRs are type I transmembrane proteins comprising two or
three immunoglobulin domains and different cells of the immune system express
different FcRs. Three classes of human FcγRs have been characterized: the high-
affinity FcγRI and the low/medium-affinity FcγRII and FcγRIII. Both of the
latter FcγRs can further be divided into two subclasses: a and b. FcγRIIIa is
53
38
found on monocytes, macrophages and natural killer cells. In contrast to all the
other FcγRs which are activating/pro-inflammatory, FcγRIIb (e.g. on B-cells and
other antigen-presenting cells) is inactivating/anti-inflammatory.
3.2 IgG presenting surface as an in vitro immune complex-model
There is currently a renewed interest in antibodies, immune complexes (ICs),
and their specific receptors in relation to etiopathogenesis and treatment of
autoimmune diseases, e.g. RA86,87. Despite the fundamental importance of IgG-
IC/FcγR interactions, most in vitro IC-models (e.g. heat-aggregated IgG) show
limitations in reproducibility and control of size. Paper IV describes a designed
surface where the IgG density can be controlled. This surface acts as an in vitro
model for IC-mediated stimulation of leukocytes. A molecular system was
designed to allow the presentation of IgG-Fc to cell surface FcγRs on monocytes
and to investigate whether the IgG density affected cell adhesion (Figure 3.1).
Figure 3.1 Schematic illustration of the surface design for the analysis of IgG-density-
dependence of monocyte adhesion. The monoclonal IgG1 antibody rituximab was
immobilized at different densities to a mixed SAM of oligo(ethylene glycol)-containing
alkane thiolates on gold via a CD20-derived peptide linked to neutravidin.
3.2.1 Interaction between rituximab and a CD20-derived peptide epitope
The chimeric monoclonal antibody Rituximab is a therapeutic antibody which
comprises human IgG1 constant regions and mouse variable regions. Its target is
CD20, a membrane protein located on the surface of B-cells. CD20 has no
known natural function, but it is believed that it forms a calcium channel and it
has been shown that it associates with MHC class II molecules88. Rituximab is
54
39
used to treat different types of lymphomas and various autoimmune diseases89,90.
Binder et al. described the CD20 epitope recognized by rituximab91. Phage
display libraries were used to select peptides binding to the antibody. The
epitope was identified to a large extracellular loop comprising residues 163-187.
Figure 3.2 (right) shows an illustration of the proposed topology of the
extracellular domain. An intra-molecular disulfide has been found to be of great
importance for the interaction88. Du et al. presented the crystal structure of the
complex of rituximab-Fab and the CD20 epitope peptide corresponding to
amino acids 163-18792. The peptide indeed adopts a cyclic conformation and the
recognition and binding is greatly affected by the 170ANPS173 motif being deeply
embedded in a pocket of the antibody surface.
Figure 3.2 (Left) Sensorgrams showing the interaction between rituximab and immobilized
pCD20ox (solid line) and pCD20red (dashed line). (Right) Illustration of CD20 as presented
on the extracellular region of a B-cell membrane. Amino acids marked in grey indicate the
residues in the synthetic epitope-peptide (pCD20) and those marked in blue constitute the
ANPS binding motif. Cysteine residues marked in yellow form an intramolecular disulfide
bond.
In Paper IV, a biotinylated peptide with reduced cysteine residues (pCD20red)
was synthesized on the solid phase. This peptide comprises the linear epitope 170ANPS173 and the two cysteine residues (167 and 183) that form a disulfide in
CD20. To ensure functional stability of the IC-mimic surfaces a high affinity
interaction between rituximab and the surface bound peptide epitope was desired.
Blasco et al. argued, regarding the oxidation state of pCD20, that a disulfide
bond does not influence rituximab binding93, while others have demonstrated the
55
40
importance of an intact disulfide bond for rituximab binding to full length
CD2088. SPR analysis of rituximab binding to pCD20 at different states of
oxidation revealed that 400% higher degree of complex formation was observed
with the oxidized peptide (at the chosen rituximab concentration and comparable
immobilization levels of the two peptides). Figure 3.2 (left) shows a much
slower post-injection phase for the oxidized peptide compared to the reduced
peptide, indicative of a slower dissociation of the rituximab complex with that
peptide. An explanation for the increase in binding strength for the oxidized
peptide could be the unique cyclic structure that pCD20 adopts when oxidized.
The rituximab/pCD20-complex presented by Du et al. indeed shows that this
loop-motif is deeply embedded in the rituximab-Fab hypervariable region92. The
reduced peptide would clearly display more degrees of rotational freedom, as a
result of a more open structure, which would probably have an effect on the
binding strength of rituximab. The peptide pCD20ox was thus used in the IC-
mimic surface design.
3.2.2 Surface design and characterization
SAMs on gold provide physical and chemical flexibility in the design of model
surfaces mimicking biological interfaces94,95. In Paper IV mixed SAMs of
oligo(ethylene glycol)-containing carboxyl- and methoxy-terminated alkane
thiolates on gold were used as a biologically inert anchoring layer for
immobilization. The repetitive ethylene glycol units are associated with
attractive protein repellent properties desired in surfaces used in biological
systems94,96. The carboxyl group was used to covalently attach the biotin-
binding protein neutravidin. The lateral distribution of neutravidin was
controlled by varying the fraction of the carboxyl-terminated thiolate in the
mixed SAM.
The carboxyl groups on the different SAMs were activated using 1-ethyl-3-(3-
dimethylaminopropyl) carbodiimide and N-hydroxysuccinimide, followed by
immobilization of neutravidin. The peptide pCD20ox was linked to neutravidin
(il lustrated in Figure 3.1, bottom) followed by subsequent injection of rituximab
into the flow chamber. When describing the surface concentration of rituximab,
references are based on the percentage of carboxyl-terminated alkane thiol in the
incubation solution when preparing the SAMs. Percentages analyzed were 0, 2.5,
5, 10, 20, 40, 80 and 100%. The corresponding surface fraction of the respective
56
41
thiolate may deviate from this number97. The surfaces were characterized
regarding their capability of binding rituximab. Wettability measurements of the
SAMs yielded a static contact angle of around 50° for the surfaces with the
lowest fraction of carboxyl-terminated thiolates. However, as the carboxyl-
content on the surfaces increased, the contact angles decreased to reach a
minimum at around 35° (Figure 3.3).
Figure 3.3 Static contact angles (∆) and rituximab SPR response (□) versus the solution
fraction of acid-terminated thiol.
Low contact angles with water are related to hydrophilic surfaces, and should be
expected for surfaces with high surface content of carboxyl groups. To further
evaluate the binding capacity of rituximab to the surfaces, the complete
molecular assembly on the surfaces was monitored by SPR. The rituximab
responses at the different surfaces are shown in figure 3.3. There is a strong
negative correlation between contact angle and bound rituximab. Interestingly,
very low amounts of rituximab was found on surfaces with low carboxyl content,
most likely due to insufficient electrostatic attraction of the protein to the surface.
Between relative carboxyl concentrations of 10% and 40%, tuning of rituximab
surface concentration is possible.
3.2.3 Cell adhesion propensity for immune complex-model surfaces
To verify the capability of the surfaces to function as an in vitro IC-model, the
surface adhesion of human monocytes (cell line U937) was investigated. This
particular cell line has been shown to express FcγRs98-100. Incubation of the
57
42
different surfaces in cell cultures for 24 h revealed a quantitative cell adhesion
maximum on surfaces with 20-40% carboxyl-terminated thiolate (Figure 3.4).
Figure 3.4 Monocytes (U937) adhered to surfaces presenting human IgG at different
densities. Percentages relate to the solution fraction of acid terminated thiol in the SAM.
The bound cells were visualized by staining for filamentous actin. (Bars = 10 µm)
This maximum occurs in the dynamic region where tuning of the IgG density is
possible (Figure 3.3). An explanation to this phenomenon could be that the
lateral distribution of FcγRs expressed on the monocytes matches the density of
presented IgG-Fc on the IC-model surfaces. This would then allow for optimal
binding possibilities for FcγRs to IgG-Fc. From almost non-existing cellular
adhesion to control surfaces lacking rituximab it can be concluded that it is in
fact the exposure of IgG-Fc and not surface wettability per se that facilitates
cellular adhesion. Differences in surface properties, e.g. wettability, have been
found to affect mammalian cells and their behavior at related interfaces101.
However, the protein repellent property of the ethylene glycol-containing SAMs
used in Paper IV eliminates this risk. This IC-model, mimicking an IgG-
opsonized surface, allows for highly controllable experiments where adjustable
lateral distribution of IgG is desired and can be of potential use in studies on the
importance of FcγR-mediated immune response in chronic immune-mediated
diseases.
3.3 Assays for IgG rheumatoid factor
Patients suffering from the chronic inflammatory disease RA are diagnosed
according to criteria established by the World Health Organization International
Classification of Diagnoses (ICD-10). Occurrence of autoantibodies to
58
43
citrullinated proteins/peptides (ACPA) is highly specific for RA and predicts
poor disease outcome102-106. A positive test for rheumatoid factor is also a
prognostic and diagnostic seromarker for RA107-109. Patients are denoted
seropositive or seronegative based on the outcome in an RF test. The clinically
most widespread RF assay, detection of anti-IgG antibodies agglutinating IgG-
sensitized particles, mainly detects IgM-class antibodies. Also, there exist
subclass specific RF tests, straightforwardly performed for IgM/IgA-RF.
Intriguingly, even though IgG-RF is of special pathogenic interest110,111, IgG-RF
specific tests remains technically challenging and require careful assay design.
Tests targeting different isotypes of RF using radioimmunoassay or enzyme
immunoassays were introduced in the 1970s112-114. In 1977, Carson et al.
developed a radioimmunoassy to assess IgG-RF using a heterogeneous pool of
hIgG-Fc fragments as target antigen112. IgG-RF also has reactivity for rabbit
IgG-Fc115. To avoid false positive results due to interaction between the
detection antibody and the hIgG-Fc antigen as well as IgG-RF interaction with
the Fc-fragment of the rabbit detection antibody, F(ab’)2 fragments directed
toward the human Fab γ-chain was used for revelation. Later, McDougal et al.
refined this concept using a monoclonal anti-Fd antibody instead115. An
alternative approach has been to use plates coated with bovine serum albumin
(BSA) as a capture molecule for rabbit anti-BSA antibodies116. Here, false
positives may arise from recruitment of anti-BSA antibodies in the serum
samples. This risk was eliminated by blocking the anti-BSA reactivity by the
addition of BSA to the buffers. However, the formation of BSA-anti-BSA
complexes in the samples may in turn neutralize the RF activity and may also
serve as substrate for “false” IgG-RF reactivity. Elson et al. studied the subclass
specificity of IgG-RF in a study where they used heterogeneous pools of hIgG1-
Fc and hIgG2-Fc and intact hIgG3 and hIgG4 as the target antigens117. The
results showed significantly raised IgG-RF levels against hIgG1-Fc and hIgG2-
Fc and IgG-RF binding decreased according to: IgG2>IgG1>IgG4>IgG3.
However, their assay involved BSA used as blocking agent. As illustrated in
Figure 3.5, false negatives (A) as well as false positives (B) may also be induced
by the presence of IgM-RF or IgA-RF, either by blocking the solid-phase IgG-
Fc (A) or by cross-linking the adsorbed IgG-Fc with soluble IgG-ICs in the
patient serum (B).
59
44
Figure 3.5 Schematic picture of a false negative IgG-RF test due to blocking effect by IgM-
RF (A) and a false positive IgG-RF test due to IgM-RF mediated cross-linking of adsorbed
IgG-Fc antigen and soluble IgG-containing immune complexes (ICs) in patient serum (B).
3.4 IgG-RF subclass specificity in early rheumatoid arthritis
Paper V describes data aimed at targeting the subclass-specificity (IgG1-4) of
IgG-RF in a population of clinically well-characterized patients with
seropositive and seronegative (as judged by particle agglutinating RF and
ELISA) early RA. Sera from patients included in the prospective Swedish early
arthritis study “TIRA-1”103 were analyzed. Inclusion sera from 20 patients
seropositive regarding IgM/IgA-RF and with a positive test for IgG anti-CCP
were included, together with 20 patients with negative results in these tests. Sera
from 20 healthy blood-donors were used as reference.
In the ELISA presented in Paper V (Figure 3.6), hIgG1-Fc, hIgG2-Fc, hIgG3-Fc
and hIgG4-Fc were used as the antigens in four separate assays for IgG-RF. To
avoid possible complications due to the presence of IgG anti-BSA antibodies in
the sera, Tween-20 alone was used for blocking of non-coated parts in the
microtitre plates. Chicken IgY anti-human IgG-Fab (κ- and λ-light chain specific)
was used as the primary antibody. IgY is an attractive alternative to e.g. rabbit
antibodies due to the fact that they do not display cross-reactivity with human
IgG118, thus eliminating the risk of false positive results involving non-RF
interactions with the primary antibody. Data demonstrated significantly raised
IgG-RF levels against subclasses 1, 2 and 4 (Figure 3.7) in RF(+)/anti-CCP(+)
patients compared to RF(-)/anti-CCP(-) patients.
60
45
Figure 3.6 Illustration of the IgG-RF ELISA. The solid support is coated with hIgG-Fc of
subclasses 1, 2, 3, or 4, followed by the addition of the serum sample containing the
analyte (IgG-RF). Chicken IgY anti-human IgG-Fab is used as primary antibody and a
biotinylated donkey IgG anti-chicken IgY is used as secondary antibody. Streptavidin
conjugated to alkaline phosphatase is used for revelation.
No reactivity was found against hIgG3-Fc, in line with most previous findings
that IgG3 does not attract RFs117,119, although Tokano et al. have reported a
distinct subgroup with IgG-RF reactivity against IgG3120. Interestingly, IgG-RF
reaction against hIgG2-Fc was four-fold higher compared to the reaction with
hIgG1-Fc and hIgG4-Fc. Significant correlations were observed between IgG-
RF reactivity against hIgG2-Fc and IgM/IgA-RF levels. No significant
correlations were found regarding IgG-RF against the other hIgG-Fc subclasses
and IgA/M-RF levels.
3.4.1 IgM/IgA depletion and its effect on IgG-RF binding
The complexity of RF activity is an important aspect to consider when
investigating IgG-RF levels in serum. As illustrated in figure 3.5, the presence
of IgM-RF (and IgA-RF) may result in either false positive or false negative
results when assessing IgG-RF. If there is an effect and to what extent it
influences the outcome in IgG-RF analysis has not yet carefully been
investigated. Depletion of IgM has been performed in a number of studies but
61
46
Figure 3.7 IgG-RF ELISA with different human IgG-Fc subclasses used as antigen. Black
bars represent the average signal expressed as % change relative to controls for the
patient-group denoted seropositive, i.e. RF(+) as well as anti-CCP(+). Gray bars represent
the average signal expressed as % change relative to control for the patient-group denoted
seronegative, i.e. RF(-) as well as anti-CCP(-). % change relative to controls values were
calculated by normalization to a control-group consisting of 20 healthy blood-donors (NS =
not significant, ** = p < 0.01, *** = p < 0.001). Error bars represent the 95% confidence
interval (n = 20).
the results have not been properly discussed115,121. Within this thesis, a protocol
for efficient IgM/IgA-depletion was established, where F(ab’)2 fragments of
rabbit anti-human IgM and rabbit anti-human IgA were covalently immobilized
to an affinity gel support. IgG-RF has been shown to bind to rabbit IgG-Fc and
in order to avoid cross reactivity resulting in depletion of the analyte itself (IgG-
RF), the intact capture antibodies were digested with pepsin and purified prior to
immobilization. Non-specific adsorption to affinity supports is a common
problem in protein purification. In this case a carefully chosen support and high
salt buffers including Tween 20 and EDTA, with the purpose of e.g. inhibiting
interference from complement proteins on the affinity support, were used to
minimize these risks. The ELISA presented in Paper V, in combination with
pull-down assays and western blot, was used to investigate the effect of
IgM/IgA-depletion by comparing depleted and non-depleted sera from six
seropositive RA patients. Preliminary results (not shown), involving only few
patients, show a slight tendency towards increased OD-values with the depleted
serum samples. This could indicate a possible blocking effect of IgG-RF caused
by the presence of IgM/IgA-RF (Figure 3.5A). A complete investigation
including more patients from both seropositive and seronegative groups as well
as healthy blood donors is needed to be able to draw any conclusions of the
62
47
impact of other isotypes in IgG-RF assays. It is also critical to adjust for possible
changes in total IgG concentration in the serum samples upon depletion.
3.4.2 IgG-RF subclass specific tests
Previously, assessments of IgG-RF reactivity against human IgG have utilized
isolated, intact human IgG and in some cases isolated hIgG-Fc (subclasses 1 and
2) as target antigen112,115,117,119. Paper V describes the utility of recombinant
hIgG-Fc. An advantage with using the Fc-fragment alone is that no other parts
of IgG are present to interfere with IgG-RF binding. Moreover, using
recombinant Fc-fragments allows for molecular control of the antigens, i.e.
molecular heterogeneity, which can be a source of limited assay reproducibility,
is eliminated. The possibility of glyco-engineering in P. pastoris122,123 allows for
a careful characterization of the impact on IgG-RF reactivity from IgG-Fc
glycosylation.
It would be interesting to profile the IgG-RF pool in early RA patients with
respect to both specificity and subclass. A two-dimensional array is envisioned
(the concept is illustrated in figure 3.8) where the different subclasses of hIgG-
Fc would serve as the antigen and an application of subclass-specific primary
antibodies would allow for profiling of the IgG-RF pool regarding their subclass.
Results could then be presented as an individual patient IgG-RF pattern (Figure
3.8, right) and analyzed for correlations with disease-progress and outcome.
Figure 3.8 Schematic illustration of a two-dimensional array concept for profiling of IgG-RF
specificity and subclass. A panel of the different hIgG-Fc could be used as the sources of
antigen to assess IgG-RF specificity with respect to hIgG subclass (presented in Paper V),
while subclass-specific primary antibodies would allow characterization of the subclass
profile of IgG-RF.
63
48
64
49
4. Methodology
________________________________________________________________
4.1 Protein expression in Pichia pastoris
P. pastoris is a very useful expression system for the production of a broad
variety of recombinant proteins124. It was developed in the early 1980s by the
American company Phillips Petroleum, in collaboration with the Salk Institute,
La Jolla, CA. Together they developed strains, vectors and methods for genetic
modification of P. pastoris. In the mid 1990s the expression system became
public; figure 4.1 (left) shows the almost linear increase of P. pastoris-related
research since then.
Figure 4.1 (Left) Diagram showing the number of P. pastoris-related publications during the
last 30 years. (Right) General diagram of a P. pastoris expression vector, containing both E.
coli and P. pastoris antibiotic selection marker genes, a promoter gene, a 3’AOX1 fragment
for transcription termination and a multiple cloning site (MCS) for insertion of a gene of
choice.
65
50
P. pastoris is a singled-cell, methylotropic microorganism and can survive and
grow on methanol as its sole carbon source125. This species of yeast bears great
similarity to Saccharomyces cerevisiae, which together with the bacterium E.
coli belongs to two of the most characterized model systems in modern
microbiology. Yeasts are eukaryotic, and like higher eukaryotes capable of
performing post-translational modifications such as glycosylation, disulfide
bond formation and proteolytic processing.
The expression of a foreign gene in P. pastoris requires tree general steps: (i)
insertion of the gene into an expression vector. The basic features of such a
vector are illustrated in figure 4.1 (right). All P. pastoris expression vectors have
been designed as shuttle vectors, i.e. containing an origin of replication and a
selection marker gene for plasmid maintenance in E. coli. (ii) Introduction of the
expression vector into the P. pastoris genome, thus creating an expression strain.
Due to the expression vector being integrated into the genome in a homologous
recombination event, the vector needs to be linearized prior to the reaction. (iii)
The clones obtained after transformation requires screening for potential
expression strains for the foreign gene product. A wide variety of expression
vectors are available; the protein can be expressed either intracellularly or
secreted into the growth medium. Expression can be either induced when the
appropriate cell density has been reached, or performed constitutively, allowing
the gene product to accumulate already from inoculation.
The P. pastoris N-glycosylation pathway is only partly homologous to the
pathway in human cells. In the Golgi apparatus, the human cells synthesize
complex oligosaccharides, whereas P. pastoris produces high mannose
structures, with up to 40 residues. This hyperglycosylation is not desired when
expressing a foreign glycoprotein. There are however ways of producing
glycoproteins with humanized N-linked glycans, involving co-expression of the
glycoprotein with enzymes tagged to be retained in the endoplasmic reticulum,
where they perform glycan engineering, thus removing outer mannose residues
and trimming the glycostructure. In vivo synthesis of mammalian-like, hybrid-
type N-glycans123 and optimization of humanized glycoengineered proteins have
been reported for several proteins, amongst else human IgG122.
66
51
4.2 Biochemical characterization techniques
A wide variety of techniques have been used to characterize the biochemical
properties of the proteins in this thesis, including analytical ultra centrifugation,
matrix-assisted laser desorption/ionization time-of-flight mass spectrometry and
western blot. This section will focus on the two most important methods: CD
and fluorescence spectroscopy.
4.2.1 CD spectroscopy
CD spectroscopy measures ellipticity as a function of wavelength and is a
frequently used method to probe secondary and tertiary structure in a protein or
a peptide126. A difference in absorption of right-handed and left-handed
circularly polarized light, due to the chirality of the chromophores, is used to
calculate the ellipticity. In far-UV CD, amide bonds constitute the main
chromophore in a protein or a peptide. They absorb below 240 nm and give rise
to distinctive differences in the spectrum if they are located in α-helix, β-sheet or
random coil secondary structure elements. The CD spectra in the far-UV region
(< 240 nm) of different proteins typically display pronounced minima dependent
on the secondary structure elements (random coil: 198 nm, α-helix: 208 nm and
222 nm, β-sheet: 216 nm)127. The most important chromophore when studying
absorption at wavelengths between 240 nm and 340 nm (near-UV CD) is the
indole side chain of tryptophan, although the aromatic rings of tyrosine and
phenylalanine also contribute to the spectrum. Measurements in this region
provide a fingerprint of the tertiary structure of a protein. Despite its low
resolution, CD spectroscopy is still an informative and quantitative structural
technique128,129.
The difference in extinction coefficients (∆ε) for the absorption of right-handed
(R) and left-handed (L) circularly polarized light is proportional to the
difference in absorption (∆A) according to the Lambert-Beer relation:
ε∆⋅⋅=∆ clA (4.1)
where l is the optical path length and c is the sample concentration. The
difference in amplitude between the two polarization components, due to
unequal adsorption of the right-handed and the left-handed light, results in the
67
52
transmitted light being elliptically polarized. CD is thus expressed as ellipticity
(θ) which is the angle between the major and minor axis of the ellipse described
by the vector sum of the R and L components. θ is related to ∆A according to:
AA ∆⋅≈∆⋅⋅= 98.32
4
18010ln
πθ (4.2)
For proteins, the data is often reported as mean residue ellipticity ([θ]), which is
related to θ according to:
[ ]cl
mrw
⋅⋅⋅=
10
θθ (4.3)
where θ is the observed ellipticity (deg), mrw is the mean residue weight (g/mol),
l is the optical path length (cm) and c is the protein concentration (g/ml).
4.2.2 Fluorescence spectroscopy
Fluorescence130 is a phenomenon where a fluorophore absorbs light of a given
wavelength within its absorption spectrum and is excited to a higher energy
level, followed by emission of photons at longer wavelengths. The technique is
commonly used in the biological, biochemical and biophysical sciences for its
simplicity and sensitivity.
The process of fluorescence occurs in three steps and is often illustrated by a
Jablonski diagram (Figure 4.2, left). When a fluorophore absorbs light within its
absorption spectrum, it is excited to a higher energy state. Following absorption
of light, a process called internal conversion occurs; part of the energy in the
excited state is lost due to rotational and vibrational relaxation. Upon return to
the ground state, there is an emission of a photon which is detected as
fluorescence. As a result of internal conversion, the emitted light is always of a
lower energy, i.e. longer wavelengths, than the absorbed light. This is called the
Stokes shift (illustrated in figure 4.2, right); this shift is fundamental to the
sensitivity of fluorescence techniques as it allows detection of the emission
photons without the interference of the excitation photons.
68
53
Figure 4.2 Simplified Jablonski diagram (left) illustrating the three-step process of
fluorescence. A fluorophore is excited to a higher energy state by a photon of the energy
hνex, supplied by an external source (1). The excited state exists for a short time, during
which the fluorophore undergoes conformational changes and part of the energy is
dissipated through rotational and vibrational relaxation (2). As the fluorophore returns to its
ground state, a photon of the energy hνem is emitted (3). The difference in energy or
wavelength between hνex and hνem is called the Stokes shift (right).
Fluorescence emission spectra can be strongly affected by the medium
surrounding the fluorophore. Environmental factors that influence the
fluorescence include solvent polarity, pH and the concentration and closeness of
extrinsic and intrinsic quenchers. When the polarity of the environment
increases, the excited state becomes more stable compared to when the
fluorophore is located in a more non-polar environment. This leads to a redshift
of the fluorescence emission, i.e. emission of lower energy shifted towards
longer wavelengths. The opposite is called blueshift and occurs when the
surrounding polarity decreases.
Proteins can contain three aromatic amino acids that act as intrinsic fluorophores:
tryptophan, tyrosine and phenylalanine. These amino acids have distinct
absorption and emission spectra, but the fluorescence spectrum of a protein
greatly resembles that of tryptophan because of its dominant UV absorption. The
wavelength of fluorescence intensity maximum of tryptophan is highly sensitive
to environmental and solvent effects. It moves towards longer wavelengths as
the polarity surrounding the tryptophan increases, e.g. as a result of unfolding.
Intrinsic fluorescence spectroscopy is often used to study the conformation,
folding and interactions of proteins. In this work, steady-state fluorescence
spectroscopy and microscopy was used, utilizing intrinsic tryptophan
fluorescence and the extrinsic probes ANS and Alexa Fluor 488.
69
54
4.3 ELISA
Among the most commonly used methods for the detection of biomolecules in
research and diagnostics are antibody-based immunoassays. The specificity and
sensitivity of such assays to most part rely on the binding characteristics of the
antibodies used and also on the choice of detection method. One of the most
widespread methods is ELISA, enzyme-linked immunosorbent assay. ELISA
can be divided in three major types: direct, indirect and sandwich formats. In a
direct ELISA the analyte is adsorbed to a solid support and its presence is
confirmed by a detection antibody raised against it. The detection antibody is
labeled with an enzyme. The final step involves addition of a substrate for the
enzyme followed by detection of the enzymatic product, usually by changes in
optical density. Direct detection is not widely in ELISA but is the typical format
for immunohistochemistry. In the indirect format a secondary antibody
specifically directed against the detection antibody is added to enhance the
signal. The secondary antibody is furthermore labeled with the enzyme. The
sandwich format includes coating of the solid support with a capture antibody
targeting the analyte of interest, followed by addition of the detection antibody
and the enzyme-labeled secondary antibody. This method is called a sandwich
assay because the analyte is bound between the capture and detection antibodies.
In this thesis, an indirect sandwich ELISA was used, where the solid support
was coated by the antigen for the analyte (IgG-RF) and indirect detection was
performed in three steps (Figure 3.6).
4.4 Amyloid fibril characterization techniques
The presence of amyloid deposits in tissue, or the confirmation of in vitro fibrils,
can be probed by observing characteristic properties of these structures85:
spectroscopic observation of increased β-sheet content, thioflavinophilic and
congophilic properties of the material studied, characteristic x-ray diffraction
patterns, and fibril appearance in the TEM. This section will focus on the three
methods used in this thesis: ThT binding, Congo red birefringence and TEM
imaging.
4.4.1 Thioflavin T binding
ThT is a fluorescent dye first introduced for the detection of amyloid in 1959131.
The binding mode of ThT to amyloid fibrils is not completely elucidated, but
when specifically bound to the β-pleated sheets of the fibrils, its fluorescence
70
55
emission spectrum is characteristically blue-shifted along with a significant
increase in intensity132,133. ThT is a small, positively charged benzothiazole
compound (Figure 4.3). It contains a hydrophobic part with a dimethylamino
group attached to a phenyl ring, linked to a more hydrophilic part containing
nitrogen and sulfur. The amphipathic property of the molecule allows for the
formation of micelles in aqueous solutions. It has been postulated that the
thiazole nitrogen forms hydrogen bonds with hydroxyl groups on the material
studied, thus contributing to its specific binding to amyloid structures134. ThT
binding is often used to monitor the kinetics of fibril formation, and significant
binding of ThT compared to controls denotes the aggregates studied as
thioflavinophilic, a characteristic property of amyloid.
N
SN
Cl
Figure 4.3 The structure of ThT
4.4.2 Congo red birefringence
The dye Congo red was first synthesized in 1884 as a direct cotton dye and
introduced for staining of amyloid in the 1920s. Since then it has become the
most universally used staining method for demonstration of amyloid135. The
symmetrical Congo red molecule is a sulfonated azo-dye with a hydrophobic
biphenyl center in between the negatively charged ends (Figure 4.3). When
Congo red is specifically bound to amyloid fibrils it displays birefringence if
exposed to polarized light136.
NN
NN
NH2
SO
O
O
Na
Na
S
H2N
O O
O
Figure 4.4 The structure of Congo red
71
56
The refractive index of a substance depends on the velocity of light transmitted
through the specimen. Materials like e.g. glass are considered isotropic, i.e.
displaying the same refractive indices regardless of the wavelength of the light
or the angle of incidence. Other materials that change refractive index with
wavelength or angle of the incident light are anisotropic. If anisotropic materials
are subjected to polarized light, they display birefringence, i.e. different colors
due to changes in the refraction of the light.
The binding property of Congo red to amyloid is not fully clarified; however,
when bound and aligned to amyloid fibrils the molecule changes from being
isotropic to anisotropic. Congo red stains with a red color when examining a
specimen using bright field microscopy. If the specimen is placed in between the
light source and the analyzer arranged perpendicularly to polarized light, only
light altered by the specimen would become visible, against a dark background.
An assay for amyloid using Congo red is considered positive when green
birefringence is detected under polarized light, and the aggregates studied can
thus be regarded as congophilic136. In vitro generated amyloid fibrils and in vivo
amyloid fibrils stained for ex vivo have the same Congo red binding properties137.
4.4.3 Transmission electron microscopy
In TEM138, a beam of electrons is passed through a specimen to produce an
image of the analyzed sample. The electrons ejected from the electron gun are
focused and oriented by electromagnets to interact with the sample. As they hit
the sample, some electrons transmit through the sample (onto a phosphorescent
microscope screen, where the image is produced), whereas some are lost due to
absorption or scattering. The contrast depends on the number of transmitted
electrons. Biological specimens are carbon-based and display low electron-
scattering power; thus, resulting in many electrons being transmitted, giving
poor contrast. However, if few electrons are transmitted, this will produce a
shadow with high contrast on the screen. The scattering of electrons is related to
the atomic number of the specimen. A high atomic number is equivalent to a
high density and good electron-scattering power. A common method to enhance
contrast in TEM is to use negative staining. The stain used in this thesis is
aqueous uranyl acetate, which binds with high affinity to biological compounds
and owes its contrast-enhancing properties to its high atomic number. Negative
staining is illustrated in figure 4.5.
72
57
Figure 4.5 Schematic illustration of negative staining used to enhance contrast of biological
specimens. The sample is adsorbed to carbon-coated cupper grids pre-charged with UV-
light (A) followed by addition of staining solution, which completely embeds the specimen
(B). Subsequent washing should be performed in such a way that the stain surrounds the
specimen (C). In the TEM, the adsorbed material will appear white against the dark-colored,
adjacent stain.
4.5 Surface preparation and characterization
4.5.1 Self-assembled monolayers
The formation self-assembled monolayers139 (SAMs) on surfaces allows for the
preparation of well defined molecular structures exposing a variety of chemical
functionalities. A SAM is a two-dimensional, thin film spontaneously assembled
on a surface. In this thesis, SAMs of alkane thiolates on gold was used. The
kinetics of alkane thiol adsorption is a two step process (illustrated in figure 4.6).
The initial adsorption is very fast and occurs within the first minute. The second
organization step occurs more slowly and takes up to several hours. The bond
formed between gold and thiol is very strong and the thiolates will occupy every
available binding site on the surface, thus giving a well-defined layer of
molecules. The tail group of the alkanethiols, exposed on the SAM surface can
comprise functional units for e.g. immobilization of biomolecules.
Figure 4.6 Schematic illustration of the preparation of an alkane thiolate self-assembled
monolayer on a planar gold surface. The surface is placed in a thiol solution; the thiols will
immediately adsorb forming a covalent bond between sulfur and gold. The organization
process into a well-defined monolayer is much slower and usually takes 12-24 h.
73
58
To allow for variation of the surface properties, two or more thiols can be co-
adsorbed, forming a so called mixed SAM. This could be important when for
example the lateral distribution between functional groups needs to be controlled,
e.g. to prevent steric hindrance between immobilized biomolecules97,140-142. A
mixed SAM from two oligo(ethylene glycol)-containing alkane thiols (Figure
4.7) was used in this thesis. SAMs formed from alkane thiols with multiple
ethylene glycol units are very attractive in e.g. biological assays due to their
protein-repellent properties, ensuring low non-specific binding of proteins and
cells94,96.
HS NH
OO
O
7
O
10
OH
O
HS NH
OO
O
7 9
O
Figure 4.7 Structures of the alkane thiols used in this thesis. Both molecules have ten
methylene units in the alkyl chain, an amide bond that provides lateral stability of the SAM
and multiple ethylene glycol units. The molecule with twelve ethylene glycol units has a
carboxyl group for immobilization of proteins and the molecule with ten ethylene glycol units
was used as a “filler molecule”.
4.5.2 Surface characterization by wettability, ellipsometry and SPR
Several surface-sensitive techniques can be used to characterize the physical,
structural and functional properties of a two dimensional surface. Here, three of
the methods used in this thesis will be described briefly: contact angle
goniometry, null ellipsometry and surface plasmon resonance.
Contact angle goniometry
The wettability, i.e. the global hydrophobicity and hydrophilicity properties of a
SAM can be determined by contact angle goniometry143. When a droplet of
water is placed on a surface it will spread more or less depending on the polarity
of the surface. The contact angle (θ) is defined as the angle between the surface
and the tangent of the droplet (figure 4.8); the higher the contact angle, the more
hydrophobic the surface. The use of contact angle goniometry reveals properties
74
59
of the outermost parts of a SAM and provides information on the structure and
organization of the alkane thiolates on the gold surface.
Figure 4.8 Contact angle goniometry is used to determine the wettability of a SAM. A water
droplet on a hydrophilic surface (A) will spread, whereas a water droplet on a more
hydrophobic surface (B) will, as a result of higher surface tension, give an increase in
contact angle.
Null ellipsometry
Null ellipsometry is an optical technique which provides information on the
thickness of a thin molecular film deposited on a planar surface.144 It is based on
changes in the polarization state of light when it is reflected toward a surface.
Monochromatic light passes a polarizer and a compensator to become
elliptically polarized. Upon surface reflection, a phase shift (∆) and a change in
amplitude (Ψ) are induced. The reflected light passes through the analyzer and
finally reaches the detector where the light intensity is monitored (figure 4.9).
The polarizer and the analyzer are adjusted until the reflected light is
extinguished. The values of ∆ and Ψ are used to calculate the thickness of the
molecular film covering the surface145.
Figure 4.9 Schematic illustration describing the setup of a polarizer-compensator-sample-
analyzer null ellipsometer.
75
60
Surface plasmon resonance
Biosensing based on the phenomenon of surface plasmon resonance (SPR) was
described for the first time in 1983146. Since then the SPR technology has
evolved to a powerful tool to study specificity, affinity, kinetics and
thermodynamics of biomolecular interactions. The technique does not require
labeling but it relies on the immobilization of one of the interaction partners, the
ligand, to a noble metal surface. The other interaction partner, the analyte, is
passed over the surface in a micro fluidic system at a constant flow and the
binding event is monitored in real time147.
A surface plasmon is an electron density wave that propagates along the
interface of a noble metal surface and the ambient medium. When light is totally
internally reflected towards a thin gold film (50 nm) at a certain angle of
incidence, θsp, part of the light can penetrate the metal and excite the surface
plasmon. When this resonance phenomenon occurs, the energy of the photons
couples to the plasmon and the intensity of the reflected light is decreased. This
is observed as a characteristic dip in a graph where the reflected light intensity is
plotted versus the angle of incidence. θsp is related to the refractive index of the
ambient medium. When a biomolecule binds to the surface the refractive index
of the ambient changes and gives rise to an angular shift of the dip, ∆θsp. In a
Biacore (GE Healthcare, Uppsala, Sweden) setup, ∆θsp is monitored as so called
response units (RU) as a function of time. One RU corresponds to a ∆θsp =
0.0001°, which can be translated into a change in mass density on the surface of
1 pg/mm2.
Figure 4.9 Typical setup for SPR. When analyte molecules in the flow channel bind to
immobilized ligand molecules on the metal surface the refractive index of the ambient
medium changes, giving rise to a change in θsp.
76
61
References
________________________________________________________________
(1) Branden, C.; Tooze, J. Introduction to Protein Structure; 2nd ed.; Garland Publishing,
Inc.: New York, 1999.
(2) Sun, Y.; Parker, M. H.; Weigele, P.; Casjens, S.; Prevelige Jr, P. E.; Krishna, N. R.
Journal of Molecular Biology 2000, 297, 1195-1202.
(3) Grocholski, T.; Koskiniemi, H.; Lindqvist, Y.; Mäntsälä, P.; Niemi, J.; Schneider, G.
Biochemistry 2010, 49, 934-944.
(4) Tompa, P. Trends in Biochemical Sciences 2002, 27, 527-533.
(5) Dyson, H. J.; Wright, P. E. Nat Rev Mol Cell Biol 2005, 6, 197-208.
(6) Creighton, T. E. Proteins: Structures and Molecular Properties; 2nd ed.; W.H. Freeman
and Company: New York, 1993.
(7) Leahy, D. J.; Hendrickson, W. A.; Aukhil, I.; Erickson, H. P. Science 1992, 258, 987-991.
(8) Privalov, P. L. In Protein Folding; W.H. Freeman and Company: New York, 1992.
(9) Fersht, A. Structure and Mechanism in Protein Sceince; W.H. Freeman and Company:
New York, 1999.
(10) Harris, L. J.; Larson, S. B.; Hasel, K. W.; McPherson, A. Biochemistry 1997, 36, 1581-
1597.
(11) Dobson, C. M. Nature 2003, 426, 884-890.
(12) Fändrich, M. Cellular and Molecular Life Sciences 2007, 64, 2066-2078.
(13) Dobson, C. M. Trends in Biochemical Sciences 1999, 24, 329-332.
(14) Westermark, G. T.; Johnson, K. H.; Westermark, P.; Ronald, W. In Methods in
Enzymology; Academic Press: 1999; Vol. Volume 309, p 3-25.
(15) Bence Jones, H. Philosophical Transactions of the Royal Society of London 1848, 138,
55-62.
(16) Westermark, P.; Benson, M. D.; Buxbaum, J. N.; Cohen, A. S.; Frangione, B.; Ikeda, S. I.;
Masters, C. L.; Merlini, G.; Saraiva, M. J.; Sipe, J. D. Amyloid 2002, 9, 197-200.
(17) Parham, P. The Immune System; 3rd ed.; Garland Science, Taylor & Francis Group:
London, 2009.
77
62
(18) Zouali, M. Antibodies; John Wiley & Sons, Ltd, 2001.
(19) Wright, A.; Morrison, S. L. Trends in Biotechnology 1997, 15, 26-32.
(20) Tsai, P. K.; Bruner, M. W.; Irwin, J. I.; Ip, C. C. Y.; Oliver, C. N.; Nelson, R. W.; Volkin,
D. B.; Middaugh, C. R. Pharmaceutical Research 1993, 10, 1580-1586.
(21) Zhang, W.; Czupryn, M. J. Journal of Pharmaceutical and Biomedical Analysis 2003, 30,
1479-1490.
(22) Lam, X. M.; Yang, J. Y.; Cleland, J. L. Journal of Pharmaceutical Sciences 1997, 86,
1250-1255.
(23) Cordoba, A. J.; Shyong, B. J.; Breen, D.; Harris, R. J. Journal of Chromatography B:
Analytical Technologies in the Biomedical and Life Sciences 2005, 818, 115-121.
(24) van der Neut Kolfschoten, M.; Schuurman, J.; Losen, M.; Bleeker, W. K.; Martínez-
Martínez, P.; Vermeulen, E.; den Bleker, T. H.; Wiegman, L.; Vink, T.; Aarden, L. A.;
De Baets, M. H.; van de Winkel, J. G. J.; Aalberse, R. C.; Parren, P. W. H. I. Science
2007, 317, 1554-1557.
(25) Björk, M. A.; Thyberg, I. S. M.; Skogh, T.; Gerdle, B. U. C. Journal of Rheumatology
2007, 34, 296-302.
(26) Sokka, T.; Abelson, B.; Pincus, T. Clinical and Experimental Rheumatology 2008, 26,
S35-61.
(27) Thyberg, I.; Skogh, T.; Hass, U. A. M.; Gerdle, B. Journal of Rehabilitation Medicine
2005, 37, 159-165.
(28) Scott, D. L.; Wolfe, F.; Huizinga, T. W. J. The Lancet 2010, 376, 1094-1108.
(29) Aletaha, D.; Neogi, T.; Silman, A. J.; Funovits, J.; Felson, D. T.; Bingham, C. O.;
Birnbaum, N. S.; Burmester, G. R.; Bykerk, V. P.; Cohen, M. D.; Combe, B.;
Costenbader, K. H.; Dougados, M.; Emery, P.; Ferraccioli, G.; Hazes, J. M. W.; Hobbs,
K.; Huizinga, T. W. J.; Kavanaugh, A.; Kay, J.; Kvien, T. K.; Laing, T.; Mease, P.;
Ménard, H. A.; Moreland, L. W.; Naden, R. L.; Pincus, T.; Smolen, J. S.; Stanislawska-
Biernat, E.; Symmons, D.; Tak, P. P.; Upchurch, K. S.; Vencovský, J.; Wolfe, F.; Hawker,
G. Anneals of the Rheumatic Diseases 2010, 69, 1580-1588.
(30) Waaler, E. Acta Pathologica, Microbiologica et Immunologica Scandinavica 1940, 17,
172-178.
(31) Artandi, S. E.; Calame, K. L.; Morrison, S. L.; Bonagura, V. R. Proceedings of the
National Academy of Sciences of the United States of America 1992, 89, 94-98.
(32) Corper, A. L.; Sohi, M. K.; Bonagura, V. R.; Steinitz, M.; Jefferis, R.; Feinstein, A.;
Beale, D.; Taussig, M. J.; Sutton, B. J. Nature Structural Biology 1997, 4, 374-381.
(33) Oppliger, I. R.; Nardella, F. A.; Stone, G. C.; Mannik, M. Journal of Experimental
Medecine 1987, 166, 702-710.
(34) Peterson, C.; Malone, C. C.; Williams Jr, R. C. Molecular Immunology 1995, 32, 57-75.
(35) Williams Jr, R. C.; Malone, C. C. Scandinavian Journal of Immunology 1994, 40, 443-
456.
(36) Williams Jr, R. C.; Malone, C. C.; Kolaskar, A. S.; Kulkarni-Kale, U. Molecular
Immunology 1997, 34, 543-556.
(37) Bellotti, V.; Mangione, P.; Merlini, G. Journal of Structural Biology 2000, 130, 280-289.
(38) Stevens, F. J. In Amyloid proteins. The beta sheet conformation and disease.; Sipe, J. D.,
Ed.; WILEY-VCH Verlag: Weinheim, 2005.
(39) Kyle, R. A.; Gertz, M. A. Seminars in Hematology 1995, 32, 45-59.
78
63
(40) Falk, R. H. Circulation 2005, 112, 2047-2060.
(41) Park, M. A.; Mueller, P. S.; Kyle, R. A.; Larson, D. R.; Plevak, M. F.; Gertz, M. A.
Medicine 2003, 82, 291-298.
(42) Copeland, J. N.; Kouides, P. A.; Grieff, M.; Nadasdy, T. American Journal of Surgical
Pathology 2003, 27, 1477-1482.
(43) Eulitz, M.; Weiss, D. T.; Solomon, A. Proceedings of the National Academy of Sciences
of the United States of America 1990, 87, 6542-6546.
(44) Gono, T.; Yazaki, M.; Fushimi, T.; Suzuki, T.; Uehara, T.; Sano, K.; Kametani, F.; Ito,
N.; Matsushita, M.; Nakamura, S.; Hoshii, Y.; Matsuda, M.; Ikeda, S. American Journal
of Kidney Diseases 2006, 47, 908-914.
(45) Kaplan, B.; Martin, B. M.; Boykov, O.; Gal, R.; Pras, M.; Shechtman, I.; Saute, M.;
Kramer, M. R. Virchows Archiv 2005, 447, 756-761.
(46) Mai, H. L.; Sheikh-Hamad, D.; Herrera, G. A.; Gu, X.; Truong, L. D. American Journal
of Surgical Pathology 2003, 27, 541-545.
(47) Miyazaki, D.; Yazaki, M.; Gono, T.; Kametani, F.; Tsuchiya, A.; Matsuda, M.; Takenaka,
Y.; Hoshii, Y.; Ikeda, S. I. Amyloid-Journal of Protein Folding Disorders 2008, 15, 125-
128.
(48) Solomon, A.; Weiss, D. T.; Murphy, C. American Journal of Hematology 1994, 45, 171-
176.
(49) Tan, S. Y.; Murdoch, I. E.; Sullivan, T. J.; Wright, J. E.; Truong, O.; Hsuan, J. J.;
Hawkins, P. N.; Pepys, M. B. Clinical Science 1994, 87, 487-491.
(50) Yazaki, M.; Fushimi, T.; Tokuda, T.; Kametani, F.; Yamamoto, K.; Matsuda, M.;
Shimojo, H.; Hoshii, Y.; Higuchi, K.; Ikeda, S. American Journal of Kidney Diseases
2004, 43.
(51) Arai, M.; Kuwajima, K. Advances in Protein Chemistry 2000, 53, 209-282.
(52) Goto, Y.; Fink, A. L. Biochemistry 1989, 28, 945-952.
(53) Ohgushi, M.; Wada, A. Febs Letters 1983, 164, 21-24.
(54) Fink, A. L.; Calciano, L. J.; Goto, Y.; Kurotsu, T.; Palleros, D. R. Biochemistry 1994, 33,
12504-12511.
(55) Goto, Y.; Calciano, L. J.; Fink, A. L. Proceedings of the National Academy of Sciences of
the United States of America 1990, 87, 573-577.
(56) Goto, Y.; Takahashi, N.; Fink, A. L. Biochemistry 1990, 29, 3480-3488.
(57) Uversky, V. N.; Karnoup, A. S.; Segel, D. J.; Seshadri, S.; Doniach, S.; Fink, A. L.
Journal of Molecular Biology 1998, 278, 879-894.
(58) Fink, A. L.; Oberg, K. A.; Seshadri, S. Folding & Design 1998, 3, 19-25.
(59) Buchner, J.; Renner, M.; Lilie, H.; Hinz, H. J.; Jaenicke, R.; Kiefhaber, T.; Rudolph, R.
Biochemistry 1991, 30, 6922-6929.
(60) Feige, M. J.; Simpson, E. R.; Herold, E. M.; Bepperling, A.; Heger, K.; Buchner, J.
Journal of Molecular Biology 2010, 399, 719-730.
(61) Lilie, H.; Buchner, J. Febs Letters 1995, 362, 43-46.
(62) Thies, M. J. W.; Kammermeier, R.; Richter, K.; Buchner, J. Journal of Molecular
Biology 2001, 309, 1077-1085.
(63) Welfle, K.; Misselwitz, R.; Hausdorf, G.; Hohne, W.; Welfle, H. Biochimica Et
Biophysica Acta-Protein Structure and Molecular Enzymology 1999, 1431, 120-131.
79
64
(64) Roychaudhuri, R.; Sarath, G.; Zeece, M.; Markwell, J. Archives of Biochemistry and
Biophysics 2003, 412, 20-26.
(65) Narhi, L. O.; Philo, J. S.; Li, T. S.; Zhang, M.; Samal, B.; Arakawa, T. Biochemistry 1996,
35, 11447-11453.
(66) Yang, J. J.; Yang, H. D.; Ye, Y. M.; Hopkins, H.; Hastings, G. Cell Biochemistry and
Biophysics 2002, 36, 1-18.
(67) Greenfield, N. J. Methods Enzymology 2004, 383, 282-317.
(68) Semisotnov, G. V.; Rodionova, N. A.; Razgulyaev, O. I.; Uversky, V. N.; Gripas, A. F.;
Gilmanshin, R. I. Biopolymers 1991, 31, 119-128.
(69) Thies, M. J. W.; Kammermeier, R.; Richter, K.; Buchner, J. Journal of Molecular
Biology 2001, 309, 1077-1085.
(70) Cowgill, R. W. Biochimica et Biophysica Acta - Protein Structure 1967, 140, 37-44.
(71) Tanford, C. Advances in Protein Chemistry 1968, 23, 121-282.
(72) Rudolph, R.; Siebendritt, R.; Nesslauer, G.; Sharma, A. K.; Jaenicke, R. Proceedings of
the National Academy of Sciences of the United States of America 1990, 87, 4625-4629.
(73) Kiefhaber, T.; Schmid, F. X.; Renner, M.; Hinz, H. J.; Hahn, U.; Quaas, R. Biochemistry
1990, 29, 8250-8257.
(74) Jiang, J. X.; London, E. Journal of Biological Chemistry 1990, 265, 8636-8641.
(75) Chiti, F.; Webster, P.; Taddei, N.; Clark, A.; Stefani, M.; Ramponi, G.; Dobson, C. M.
Proceedings of the National Academy of Sciences of the United States of America 1999,
96, 3590-3594.
(76) Raffen, R.; Dieckman, L. J.; Szpunar, M.; Wunschl, C.; Pokkuluri, P. R.; Dave, P.;
Stevens, P. W.; Cai, X. Y.; Schiffer, M.; Stevens, F. J. Protein Science 1999, 8, 509-517.
(77) Solomon, A.; Weiss, D. T.; Murphy, C. L.; Hrncic, R.; Wall, J. S.; Schell, M.
Proceedings of the National Academy of Sciences of the United States of America 1998,
95, 9547-9551.
(78) Wall, J.; Schell, M.; Murphy, C.; Hrncic, R.; Stevens, F. J.; Solomon, A. Biochemistry
1999, 38, 14101-14108.
(79) Glenner, G. G.; Ein, D.; Eanes, E. D.; Bladen, H. A.; Terry, W.; Page, D. L. Science 1971,
174, 712-714.
(80) Hurle, M. R.; Helms, L. R.; Li, L.; Chan, W. N.; Wetzel, R. Proceedings of the National
Academy of Sciences of the United States of America 1994, 91, 5446-5450.
(81) Khurana, R.; Gillespie, J. R.; Talapatra, A.; Minert, L. J.; Ionescu-Zanetti, C.; Millett, I.;
Fink, A. L. Biochemistry 2001, 40, 3525-3535.
(82) Harper, J. D.; Lansbury, P. T. Annual Review of Biochemistry 1997, 66, 385-407.
(83) Hong, D. P.; Ahmad, A.; Fink, A. L. Biochemistry 2006, 45, 9342-9353.
(84) Marshall, K. E.; Serpell, L. C. Biochemical Society Transactions 2009, 37, 671-676.
(85) Nilsson, M. R. Methods 2004, 34, 151-160.
(86) Kleinau, S. Current Pharmaceutical Design 2003, 9, 1861-1870.
(87) Nandakumar, K. S. Cell and Tissue Research 2009, 1-8.
(88) Ernst, J. A.; Li, H.; Kim, H. S.; Nakamura, G. R.; Yansura, D. G.; Vandlen, R. L.
Biochemistry 2005, 44, 15150-8.
(89) Boye, J.; Elter, T.; Engert, A. Anneals of Oncology 2003, 14, 520-35.
(90) Cohen, Y.; Nagler, A. Autoimmunity Reviews 2004, 3, 21-9.
(91) Binder, M.; Otto, F.; Mertelsmann, R.; Veelken, H.; Trepel, M. Blood 2006, 108, 1975-8.
80
65
(92) Du, J.; Wang, H.; Zhong, C.; Peng, B.; Zhang, M.; Li, B.; Huo, S.; Guo, Y.; Ding, J.
Journal of Biological Chemistry 2007, 282, 15073-80.
(93) Blasco, H.; Lalmanach, G.; Godat, E.; Maurel, M. C.; Canepa, S.; Belghazi, M.; Paintaud,
G.; Degenne, D.; Chatelut, E.; Cartron, G.; Le Guellec, C. Journal of Immunological
Methods 2007, 325, 127-39.
(94) Love, J. C.; Estroff, L. A.; Kriebel, J. K.; Nuzzo, R. G.; Whitesides, G. M. Chemical
Reviews 2005, 105, 1103-1169.
(95) Wetterö, J.; Hellerstedt, T.; Nygren, P.; Broo, K.; Aili, D.; Liedberg, B.; Magnusson, K.
E. Langmuir 2008, 24, 6803-6811.
(96) Prime, K. L.; Whitesides, G. M. Science 1991, 252, 1164-7.
(97) Bain, C. D.; Whitesides, G. M. Journal of the American Chemical Society 1989, 111,
7164-7175.
(98) Ichiyama, T.; Ueno, Y.; Hasegawa, M.; Niimi, A.; Matsubara, T.; Furukawa, S. Naunyn-
Schmiedeberg's Archives of Pharmacology 2004, 369, 428-433.
(99) Kontny, U.; Kurane, I.; Ennis, F. A. Journal of Virology 1988, 62, 3928-3933.
(100) Nambu, M.; Morita, M.; Watanabe, H.; Uenoyama, Y.; Kim, K. M.; Tanaka, M.; Iwai, Y.;
Kimata, H.; Mayumi, M.; Mikawa, H. Journal of Immunology 1989, 143, 4158-4165.
(101) Arima, Y.; Iwata, H. Biomaterials 2007, 28, 3074-3082.
(102) Forslind, K.; Ahlmén, M.; Eberhardt, K.; Hafström, I.; Svensson, B. Anneals of the
Rheumatic Diseases 2004, 63, 1090-1095.
(103) Kastbom, A.; Strandberg, G.; Lindroos, A.; Skogh, T. Annals of the Rheumatic Diseases
2004, 63, 1085-1089.
(104) Rönnelid, J.; Wick, M. C.; Lampa, J.; Lindblad, S.; Nordmark, B.; Klareskog, L.; van
Vollenhoven, R. F. Annals of the Rheumatic Diseases 2005, 64, 1744-1749.
(105) Skogh, T. Arthitis research and Therapy 2005, 7, 230-232.
(106) Van Venrooij, W. J.; Van Beers, J. J. B. C.; Pruijn, G. J. M. Annals of the New York
Academy of Sciences 2008, 1143, 268-285.
(107) Sihvonen, S.; Korpela, M.; Mustila, A.; Mustonen, J. The Journal of Rheumatology 2005,
32, 2089-2094.
(108) Vittecoq, O.; Pouplin, S.; Krzanowska, K.; Jouen-Beades, F.; Ménard, J. F.; Gayet, A.;
Daragon, A.; Tron, F.; Le Loët, X. Rheumatology 2003, 42, 939-946.
(109) Westwood, O. M. R.; Nelson, P. N.; Hay, F. C. Rheumatology 2006, 45, 379-385.
(110) Berglin, E.; Johansson, T.; Sundin, U.; Jidell, E.; Wadell, G.; Hallmans, G.; Rantapää-
Dahlqvist, S. Ann. Rheum. Dis. 2006, 65, 453-458.
(111) Houssien, D. A.; Jonsson, T.; Davies, E.; Scott, D. L. J. Rheum. 1997, 24, 2119-2122.
(112) Carson, D. A.; Lawrance, S.; Catalano, M. A. Journal of Immunology 1977, 119, 295-300.
(113) Hay, F. C.; Nineham, L. J.; Roitt, I. M. British Medical Journal 1975, 3, 203-204.
(114) Maiolini, R.; Ferrua, B.; Quaranta, J. F. Journal of Immunological Methods 1978, 20, 25-
34.
(115) McDougal, J. S.; Kennedy, M. S.; Hubbard, M.; McDuffie, F. C.; Moore, D. D.; Aloisio,
C. H.; Phillips, D. J.; Roubey, R. A.; Reimer, C. B. Journal of Laboratory and Clinical
Medicine 1985, 106, 80-87.
(116) Rantapää-Dahlqvist, S.; De Jong, B. A. W.; Berglin, E.; Hallmans, G.; Wadell, G.;
Stenlund, H.; Sundin, U.; Van Venrooij, W. J. Arthritis and Rheumatism 2003, 48, 2741-
2749.
81
66
(117) Elson, C. J.; Carter, S. D.; Scott, D. G. I.; Bacon, P. A.; Lowe, J. Rheumatology
International 1985, 5, 175-179.
(118) Ambrosius, H.; Hadge, D. Veterinary Immunology and Immunopathology 1987, 17, 57-
67.
(119) Bonagura, V. R.; Artandi, S. E.; Davidson, A.; Randen, I.; Agostino, N.; Thompson, K.;
Natvig, J. B.; Morrison, S. L. Journal of Immunology 1993, 151, 3840-3852.
(120) Tokano, Y.; Arai, S.; Hashimoto, H.; Hirose, S.; Yagita, H.; Nakauchi, H.; Kumagai, Y.;
Arata, Y.; Okumura, K. Autoimmunity 1989, 5, 107-114.
(121) Carson, D. A.; Lawrance, S.; Catalano, M. A. J. Immun. 1977, 119, 295-300.
(122) Li, H.; Sethuraman, N.; Stadheim, T. A.; Zha, D.; Prinz, B.; Ballew, N.; Bobrowicz, P.;
Choi, B.-K.; Cook, W. J.; Cukan, M.; Houston-Cummings, N. R.; Davidson, R.; Gong, B.;
Hamilton, S. R.; Hoopes, J. P.; Jiang, Y.; Kim, N.; Mansfield, R.; Nett, J. H.; Rios, S.;
Strawbridge, R.; Wildt, S.; Gerngross, T. U. Nature Biotechnology 2006, 24, 210-215.
(123) Vervecken, W.; Kaigorodov, V.; Callewaert, N.; Geysens, S.; De Vusser, K.; Contreras,
R. Applied and Environmental Microbiology 2004, 70, 2639-2646.
(124) Cereghino, J. L.; Cregg, J. M. FEMS Microbiology Reviews 2000, 24, 45-66.
(125) Hartner, F. S.; Glieder, A. Microbial Cell Factories 2006, 5, 39.
(126) Holzwarth, G.; Doty, P. Journal of the American Chemical Society 1965, 87, 218-228.
(127) Greenfield, N. J.; Fasman, G. D. Biochemistry 1969, 8, 4108-4116.
(128) Johnson Jr., W. C. Proteins 1990, 7, 205-214.
(129) Kelly, S. M.; Price, N. C. Current Protein & Peptide Science 2000, 1, 349-384.
(130) Lakowicz, J. R. Principles of fluorescence spectroscopy; Third ed.; Springer: New York,
2006.
(131) Vassar, P. S.; Culling, C. F. Archives of Pathology 1959, 68, 487-498.
(132) Levine, H. Archives of Biochemistry and Biophysics 1997, 342, 306-316.
(133) LeVine, H. Amyloid 1995, 2, 1-6.
(134) Kelényi, G. Journal of Histochemistry and Cytochemistry 1967, 15, 172-180.
(135) Steensma, D. P. Archives of Pathology and Laboratory Medicine 2001, 125, 250-252.
(136) Ladewig, P. Nature 1945, 156, 81-82.
(137) Glenner, G. G.; Eanes, E. D.; Bladen, H. A. Journal of Histochemistry and Cytochemistry
1974, 22, 1141-1158.
(138) Williams, D. B.; Carter, C. B. Transmission Electron Microscopy; Springer
Science+Business Media, Inc: New York, 1996.
(139) Ulman, A. Chemical Reviews 1996, 96, 1533-1554.
(140) Bain, C. D.; Evall, J.; Whitesides, G. M. Journal of the American Chemical Society 1989,
111, 7155-7164.
(141) Nelson, K. E.; Gamble, L.; Jung, L. S.; Boeckl, M. S.; Naeemi, E.; Golledge, S. L.;
Sasaki, T.; Castner, D. G.; Campbell, C. T.; Stayton, P. S. Langmuir 2001, 17, 2807-2816.
(142) Pale-Grosdemange, C.; Simon, E. S.; Prime, K. L.; Whitesides, G. M. Journal of the
American Chemical Society 1991, 113, 12-20.
(143) Adamson, A. W. Physical Chemistry of Surfaces; 5th ed.; John Wiley & Sons, Inc: New
York, 1990.
(144) Azzam, R. M. A.; Bashara, N. M. Ellipsometry and polarized light; North Holland
Publishing Company: Amsterdam, 1977.
(145) McCrackin, F. L.; NBS Technical Note 479: Washington DC, 1969.
82
67
(146) Liedberg, B.; Nylander, C.; Lundström, I. Sensors and Actuators 1983, 4, 299-304.
(147) Liedberg, B.; Johansen, K. In Methods in Biotechnology; Humana Press Inc: Totowa,
1998; Vol. 7.
83
68
84
69
Acknowledgement
________________________________________________________________
My ten or so years in Linköping have come to an end. I consider them the formative years of my life,
a time during which I’ve grown – both personally and professionally. I’ve met friends for life and
gained from experiences that helped me become a happier and wiser person. During my time as a
graduate student I’ve had the fortune of working with a number of very gifted and talented people.
I would like to extend my gratitude to my advisor Karin Enander. These past five years have been a
journey for the both of us and I consider it a privilege to have been your first graduate student.
You’ve helped me to come to an understanding of what research really is all about. With great
enthusiasm and trust in my abilities you supported me through my work and your immense skill in
giving constructive feedback is something I value and respect. We both like to talk and I’ve truly
enjoyed our many discussions about science, but also on the subject of other interesting matters in
life.
To Bo Liedberg, thank you for accepting me as a graduate student at the Division of Molecular
Physics. You helped Karin and me establish a fruitful collaboration with AIR/Rheumatology at the
University Hospital and elegantly handed over the responsibility of driving our research forward, at
our own speed and in the directions we chose.
Thomas Skogh and Jonas Wetterö have been my two assistant advisors throughout the years. Thank
you very much for always making me feel welcome at AIR, for supporting me and giving valuable
feedback on my manuscripts. Thank you Alf Kastbom for helpful comments on my work. To Helena
Enocsson for good company in the lab. I’m very grateful to Gunnel Almroth for the incredible work
you did! Thank you for sharing your lab experience with me. Every time I was at AIR I left filled
with energy, mostly due to the humor and the positive attitudes all of you possess.
To Nalle Jonsson and Anki Brorsson, for invaluable help with everything from cloning to biophysics.
Thank you for showing interest in my work and also for good co-authorship.
85
70
Stefan Klintström is the right man in the right position. Thank you for the work you are doing for the
graduate students in Forum Scientium. Your care and support is very much appreciated. It was
always helpful having someone from the ‘outside’ giving his view on your research.
The Divisions of Biochemistry and Molecular Biotechnology hold a lot of wonderful people. Ina
Caesar brought color to work and in her own special way supported me with “try and relax now,
because it will only get worse”. It was always nice having someone else around all those late nights.
Thank you Patricia Wennerstrand and Sofie Nyström, for helping me in the lab and for sometimes
setting your own work aside in my favor. You guys always left your door slightly ajar if I ever
needed to come in for advice, to laugh about something or sometimes even to cry a little. Your
‘confession-booth’ is much underestimated. Thank you Linda Helmfors, for helping me with
statistics and graphical layouts, and also for your witty sense of humor. Thanks for letting me harass
you, and Karin Magnusson, with green-red-what-brown-no way-red-green - gosh, I guess it must not
have been easy having color-blind me in the office next door! A lot of people have contributed to a
friendly atmosphere, both on and off work: Leffe Johansson, Andreas Åslund, Timmy Fyrner, Roger
Gabrielsson, Jeff Mason, Robert Selegård, Veronica Sandgren, Sara Helander, Alexandra Ahlner,
Patrik Lundström, Mattias Tengdelius, Annica Johnsson, Maria Jonson, Martin Karlsson, Gabriel
Gustafsson, Lotta Tollstoy Tegler, Anna-Lena Göransson and other past and present members of the
Divisions of Biochemistry, Molecular Biotechnology and Organic chemistry, especially Johan
Viljanen, Lan Bui, Jenny Larsson Viljanen and Patrik Nygren. To Marcus Bäck, for interesting
discussions and for being a good listener. Thank you Jutta Speda for nice company in the lab, but
mostly, for being a good and supportive friend. To Maria Carlsson, for keeping track of us and for
taking care of the labs (well, I guess I’m [partly] to blame for the constant mess). Thank you Uno
Carlsson, Per Hammarström, Peter Nilsson, Magdalena Svensson and Lars-Göran Mårtensson for
valuable comments on my work.
When I first realized that perhaps research was something for me, I was doing my masters in Maria
Sunnerhagen’s group. You supported me in my choice of doing a PhD and helped me a lot by
opening up doors for me. I’ve valued the concern and interest you’ve shown for my work. Thank
you Cissi Andrésen, for very much appreciated help in the lab, clever ideas, for your impressive
patience, late night scones-breaks and also, for your friendship. Thanks Janosch Hennig, for
interesting discussions and all the fun and late nights out.
To my first and only master student, Caroline Göransson. I’m very impressed by your never-ending
struggle with the mutagenesis, before we with sad faces had to realize that the project was a no
success. It was nice having your company in the lab though, and you helped me a lot moving
forward with all the cloning.
To my fellow graduate students in the physics building: Daniel Aili, Olle Andersson, Tobias Ekblad
and Emma Ericsson - thank you for all the nice discussions and help with things I’d never really
heard of before I started my work. Thanks to Annica Myrskog for introducing me to the surface prep
techniques and for always reminding me: gloves on, don’t breath and don’t touch anything (oh we
had fun, though). Thank you Cissi Vahlberg for your encouraging words. Thanks to all other, past
and present members of the Molecular Physics group, for good company, interesting discussions and
happy times on lab-retreats and bbqs.
86
71
Thank you Pia Blomstedt and Susanne Andersson, for helping me with administrative issues, for
dealing with my messy paperwork, but most importantly for being the kind persons you are, always
greeting me with a smile.
Thank you Rozalyn and Daniel Simon, for taking care of me and for giving me delicious food when I
in the middle of writing this thesis almost forgot to eat. To Rozalyn again, Alma Åslund, Isa
Lindgren and Selma Bajramovic, you all helped me to realize exactly what’s important in life. Thank
you for support, love and friendship.
Things got much more fun when Magnus Elfwing came back into town. Thanks for trying to pass on
your smile to a [sometimes] grumpy guy like me. Thanks also for the fun and late nights talking,
over too many glasses of red wine.
To my dear friends Lena Forsgren, Kristina Rönnebjerg, Sofia Williams, Hanna Tidblom, Sophia
Fransson, Lina Fälth, Jennie Fälth and Anna Måhl. You all mean the world to me, lots of love!
Professional success is great, but it doesn’t comfort you at times when you need it the most. To my
family, especially my parents Birgitta and Sten, for never-ending love, understanding and support.
Now on to new and exciting adventures!
Daniel Kanmert
Linköping in February 2011
87