Chemical Engineering of Small Affinity Proteins
Joel Lindgren
KTH Royal Institute of Technology
School of Biotechnology
Stockholm 2014
© Joel Lindgren
Stockholm, 2014
KTH Royal Institute of Technology
School of Biotechnology
AlbaNova University Center
SE-106 91 Stockholm
Sweden
Printed by Universitetsservice US-AB
Drottning Kristinas väg 53B
SE-106 91 Stockholm
Sweden
ISBN 978-91-7595-004-4
TRITA-BIO Report 2014:3
ISSN 1654-2312
iii
Joel Lindgren (2014): Chemical Engineering of Small Affinity Proteins.
School of Biotechnology, KTH Royal Institute of Technology, Stockholm, Sweden.
Abstract
Small robust affinity proteins have shown great potential for use in therapy, in vivo diagnostics, and
various biotechnological applications. However, the affinity proteins often need to be modified or
functionalized to be successful in many of these applications. The use of chemical synthesis for the
production of the proteins can allow for site-directed functionalization not achievable by recombinant
routes, including incorporation of unnatural building blocks. This thesis focuses on chemical engineering
of Affibody molecules and an albumin binding domain (ABD), which both are three-helix bundle proteins
of 58 and 46 amino acids, respectively, possible to synthesize using solid phase peptide synthesis (SPPS).
In the first project, an alternative synthetic route for Affibody molecules using a fragment condensation
approach was investigated. This was achieved by using native chemical ligation (NCL) for the
condensation reaction, yielding a native peptide bond at the site of ligation. The constant third helix of
Affibody molecules enables a combinatorial approach for the preparation of a panel of different Affibody
molecules, demonstrated by the synthesis of three different Affibody molecules using the same helix 3
(paper I).
In the next two projects, an Affibody molecule targeting the amyloid-beta peptide, involved in Alzheimer’s
disease, was engineered. Initially the N-terminus of the Affibody molecule was shortened resulting in a
considerably higher synthetic yield and higher binding affinity to the target peptide (paper II). This
improved variant of the Affibody molecule was then further engineered in the next project, where a
fluorescently silent variant was developed and successfully used as a tool to lock the amyloid-beta peptide
in a β-hairpin conformation during studies of copper binding using fluorescence spectroscopy (paper III).
In the last two projects, synthetic variants of ABD, interesting for use as in vivo half-life extending partners
to therapeutic proteins, were engineered. In the first project the possibility to covalently link a bioactive
peptide, GLP-1, to the domain was investigated. This was achieved by site-specific thioether bridgemediated cross-linking of the molecules via a polyethylene glycol (PEG)-based spacer. The conjugate
showed retained high binding affinity to human serum albumin (HSA) and a biological activity
comparable to a reference GLP-1 peptide (paper IV). In the last project, the possibility to increase the
proteolytic stability of ABD through intramolecular cross-linking, to facilitate its use in e.g. oral drug
delivery applications, was investigated. A tethered variant of ABD showed increased thermal stability and a
considerably higher proteolytic stability towards pepsin, trypsin and chymotrypsin, three important
proteases found in the gastrointestinal (GI) tract (paper V).
Taken together, the work presented in this thesis illustrates the potential of using chemical synthesis
approaches in protein engineering.
Keywords: Affibody molecules, albumin binding domain, ligation, protein synthesis, solid phase peptide
synthesis
iv
v
List of publications
This thesis is based upon the following five papers, which are referred to in the
text by their Roman numerals (I-V) The papers are included in the Appendix
of this thesis and are reproduced with permission from the copyright holders.
I.
Lindgren J., Ekblad C., Abrahmsén L. and Eriksson Karlström A.
(2012). A native chemical ligation approach for combinatorial assembly
of Affibody molecules. ChemBioChem 13(7): 1024-1031.
II.
Lindgren J., Wahlström A., Danielsson J., Markova N., Ekblad C.,
Gräslund A., Abrahmsén L., Eriksson Karlström A. and Wärmländer
S. K. (2010). N-terminal engineering of amyloid-beta-binding Affibody
molecules yields improved chemical synthesis and higher binding affinity.
Protein Sci 19(12): 2319-2329.
III.
Lindgren J., Segerfeldt P., Sholts S. B., Gräslund A., Eriksson
Karlström A. and Wärmländer S. K. (2013). Engineered nonfluorescent Affibody molecules facilitate studies of the amyloid-beta
(Abeta) peptide in monomeric form: low pH was found to reduce
Abeta/Cu(II) binding affinity. J Inorg Biochem 120: 18-23.
IV.
Lindgren J., Refai E., Zaitsev S., Abrahmsén L., Berggren P-O. and
Eriksson Karlström A. (2014). A GLP-1 receptor agonist conjugated to
an albumin-binding domain for extended half-life. Submitted
V.
Lindgren J. and Eriksson Karlström A. (2014). Intramolecular
thioether cross-linking of therapeutic proteins to increase proteolytic
stability. Submitted
vi
Contents
Introduction 1 Proteins................................................................................................................ 2 1.1 Protein therapeutics ........................................................................................ 6 1.2 Protein engineering ........................................................................................ 7 1.2.1 Protein engineering by rational design .................................................... 7 1.2.2 Protein library technology ...................................................................... 8 1.3 Affinity proteins ............................................................................................. 9 1.3.1 Affibody molecules............................................................................... 11 1.3.2 Albumin binding domain (ABD) ......................................................... 14 2 Protein synthesis ................................................................................................ 17 2.1 The pre-solid phase peptide synthesis era of chemical protein synthesis ......... 18 2.2 Solid phase peptide synthesis (SPPS) ............................................................ 19 2.3 Peptide ligation strategies.............................................................................. 24 2.3.1 Native chemical ligation (NCL) ........................................................... 24 2.3.2 Traceless Staudinger ligation ................................................................ 26 2.3.3 Expressed protein ligation (EPL) .......................................................... 26 2.3.4 Enzyme-catalyzed ligation .................................................................... 27 2.3.5 Non-native ligation methods ................................................................ 27 Present investigation 3 Aim of the thesis ................................................................................................ 30 4 Paper I – A native chemical ligation approach for combinatorial assembly of
Affibody molecules .................................................................................................. 32 5 Paper II – N-terminal engineering of amyloid-beta-binding Affibody molecules
yields improved chemical synthesis and higher binding affinity ................................ 37 vii
6 Paper III – Engineered non-fluorescent Affibody molecules facilitate studies of the
amyloid-beta (Aβ) peptide in monomeric form: low pH was found to reduce
Aβ/Cu(II) binding affinity ....................................................................................... 42 7 Paper IV – A GLP-1 receptor agonist conjugated to an albumin-binding domain
for extended half-life ................................................................................................ 45 8 Paper V – Intramolecular thioether cross-linking of therapeutic proteins to
increase proteolytic stability ..................................................................................... 50 9 Concluding remarks ........................................................................................... 56 Populärvetenskaplig sammanfattning ....................................................................... 59 Acknowledgments.................................................................................................... 63 References................................................................................................................ 65 viii
Introduction
Chemical Engineering of Small Affinity Proteins
1 Proteins
Every living organism, including bacteria, fungi, plants and animals, is
dependent on proteins and they perform a vast amount of distinct functions
both inside and outside of cells. Many proteins are enzymes, which catalyze
chemical reactions. One example is the conversion of carbon dioxide and
water to carbohydrates in plants, algae, and cyanobacteria using the energy
from sunlight. This reaction is catalyzed by a series of enzymes and all life on
earth is dependent on this process. Other proteins act as molecular motors
(e.g. myosin in muscles), which make it possible for us to move by converting
stored energy to motion. Antibodies are another important protein class which
is in charge of recognizing invaders in our body, and without antibodies an
ordinary cold could be lethal. A fourth example is the protein hemoglobin,
further discussed below, which is responsible for transporting oxygen from the
lungs to the rest of our body. This very small fraction of mentioned functions
gives a clue of the importance of proteins. Indeed, the word Protein is derived
from the Greek word proteios meaning “of the first rank”, proposed for the
first time by the Swedish scientist Jöns Jacob Berzelius in 1838. (Vickery,
1950) Proteins truly are very important for every living organism, and also
highly abundant, comprising approximately 60% of the dry mass of a
mammalian cell. (Alberts, 2002)
Proteins are typically built up from an ensemble of 20 naturally occurring
amino acids joined together by peptide bonds to form polypeptides (see figure
1), and polypeptides with a defined secondary structure are commonly
referred to as proteins. Each amino acid has a carboxyl group, an amino group
and a side chain specific for the amino acid, which gives each of the 20 amino
acids its specific properties. Even though the number of different residues
(amino acids) is relatively low, they can be combined in a very large number of
different ways. Polypeptides and proteins vary greatly in size, from very large
proteins (e.g. the muscle protein titin with more than 34,000 residues (Li,
2012), to very small peptides consisting of only a few amino acids, e.g. the
2
Joel Lindgren
R2
O
H2 N
R1
N
H
OH
O
O
O
H3C
OH
Alanine (Ala, A)
H3C
OH
CH3
NH2
NH2
HN
NH2
O
O
NH
OH
Arginine (Arg, R)
H2N
OH
O
O
OH
O
Asparagine (Asn, N)
H3C
S
OH
O
O
OH
O
Aspartic acid (Asp, D)
OH
Phenylalanine (Phe, F)
NH2
NH2
O
HS
Methionine (Met, M)
NH2
NH2
HO
Lysine (Lys, K)
NH2
NH2
H2N
Leucine (Leu, L)
OH
O
H
N
Cysteine (Cys, C)
OH
Proline (Pro, P)
NH2
O
O
HO
O
OH
Glutamic acid (Glu, E)
HO
CH3
O
OH
O
Glutamine (Gln, Q)
OH
Glycine (Gly, G)
NH
O
OH
Histidine (His, H)
NH2
OH
CH3
OH
Isoleucine (Ile, I)
Tyrosine (Tyr, Y)
NH2
HO
O
NH2
Tryptophan (Trp, W)
NH2
O
H3C
OH
O
NH2
CH3
Valine (Val, V)
NH2
O
N
H
OH
O
HO
NH2
N
Serine (Ser, S)
NH2
NH2
NH2
OH
O
H3C
OH
Valine (Val, V)
NH2
Figure 1. The chemical structure of a dipeptide (top) and the 20 natural amino acids (bottom) are shown. In
the dipeptide, the peptide bond is marked with a dashed square and R1 and R2 represent the two amino acid
side chains. For the 20 amino acids the amino group, carboxyl group, and alpha carbon, common for all
different amino acids, are shown in grey while the unique side chains are shown in black. Note that glycine has
no side chain other than a hydrogen atom and that the side chain of proline is connected to the amine nitrogen.
3
Chemical Engineering of Small Affinity Proteins
bioactive polypeptide angiotensin-(1-7) built up by only seven residues.
(Santos, 2003) These are extremes, while a protein in nature typically consists
of approximately 300 amino acids. (Brocchieri, 2005) In humans there are
approximately 20,000-25,000 different protein-encoding genes and the
proteins are often further processed by post-translational modifications
increasing the estimated number of unique protein molecules in humans to
50,000-500,000. (Uhlén, 2005)
The primary structure of proteins, i.e. the amino acid sequence, is determined
by the gene sequence stored in the DNA of every cell, together with possible
post-translational modifications, including enzymatic cleavage of the protein.
The usually long string of amino acids often forms secondary structure
segments where hydrogen bonds between the C=O and the N-H groups of
different peptide bonds play an important role. Two major forms of secondary
structure are alpha-helices (Pauling, 1951) and beta-sheets (Pauling, 1951)
often connected by more flexible loops or turns. In most proteins the formed
secondary structures subsequently fold in what is called tertiary structures of
which several can assemble into quaternary structures, exemplified by
hemoglobin in figure 2.
While hydrogen bonds within the protein backbone are most important for
the secondary structure formation, the side chains are most important for
forming the tertiary and quaternary structure of a protein. Also in these higher
levels of protein structure hydrogen bonds are involved, but other interactions
based on hydrophobic effects (exclusion of water from hydrophobic parts of
the protein), electrostatic effects (ionic bonds), and van der Waals interactions
are often more important.
Many proteins are involved in interactions with other molecules, e.g. other
proteins or small molecules. Such proteins can be responsible for sensing the
environment and make the cells, and finally the whole organism, respond to
changes. One example are the proteins acting as receptors for binding the
hormone melatonin, a hormone secreted in higher levels when it is dark,
4
Joel Lindgren
which induces a number of reactions which together make us feel more sleepy.
(Witt-Enderby, 2003) The action of melatonin receptors can be used to treat
people with sleeping problems or to more quickly get rid of jet-lag, by simply
administering melatonin to activate the receptors and induce sleepiness.
V H LTPEE K SAVTALW G K
Figure 2. The different levels of protein structure illustrated using hemoglobin. Top left:
The primary structure (the amino acid sequence) of a part of the protein is shown. Top
right: The secondary structure of a part of one domain is shown, here as an alpha-helix.
Bottom left: The tertiary structure, where the different folded secondary structure segments
are packed against each other to create a defined 3D-structured domain, is illustrated.
Bottom right: Here the complete structure of hemoglobin is shown, including the
quaternary structure built up by two pairs of identical domains, i.e. four domains in total.
The four heme groups, responsible for the oxygen binding, are shown in cyan, green, orange
and red.
Other proteins are responsible for us feeling pain and developing an
inflammatory response, and one well known example is cyclooxygenase
(COX), an enzyme which induces an inflammatory response when the
immune system has recognized an intruder and is involved in pain signaling.
5
Chemical Engineering of Small Affinity Proteins
Even though the pain and inflammatory response help the body to fight the
intruder, these effects sometimes need to be reduced. This can be done by
blocking the activity of COX, by for example taking the drug aspirin (acetyl
salicylic acid), which irreversibly inhibits the action of COX and thereby
exhibits an anti-inflammatory effect and reduces the experienced pain. (Vane,
2003)
The two examples illustrate how proteins can be activated or blocked to
induce a pharmacological effect. As a matter of fact, the vast majority of all
drugs that are used today target proteins, where G-protein coupled receptors
(GPCRs) is the most common class of target proteins. (Rask-Andersen, 2011)
1.1 Protein therapeutics
Proteins themselves can also be used as drugs, either as replacement therapy
when an endogenous protein is missing or present at reduced levels, or by
administration of other proteins having a pharmacological effect. Two classical
examples of replacement therapy are insulin, used by diabetic patients, and
human growth hormone, used by patients with growth hormone deficiency.
Protein drugs, or protein therapeutics, specifically generated for targeting
other types of diseases, have during recent years gained interest and taken
some focus from the classical low molecular weight drugs. Protein therapeutics
often show excellent properties such as high potency and selectivity, and thus
often show a lower risk of severe side effects compared to low molecular
weight drugs. Moreover, protein-based drugs tend to have a faster clinical
development time, which is a very costly part of the drug development
process. (Leader, 2008;Carter, 2011;Dimitrov, 2012)
However, protein therapeutics are typically not compatible with oral delivery,
and are instead generally administered though injection or infusion. An
additional problem with oral delivery of proteins is the often large size; they
often have too large hydrodynamic radii to be efficiently absorbed into the
blood stream from the intestine, but by the aid of additives and different
formulations uptake of small- to medium-sized proteins such as the 51 amino
6
Joel Lindgren
acid long protein insulin have been reported. (Yun, 2013; Park, 2011;
Swaminathan, 2012; Renukuntla, 2013)
The importance of understanding protein function and the growing
importance of proteins as therapeutics have resulted in a massive amount of
ongoing work in the protein research field.
1.2 Protein engineering
Nature has been modifying and improving proteins for their functions for
billions of years through what we often refer to as evolution. By naturally
occurring changes in the protein-encoding sequence, such as point mutations,
insertions and deletions, together with natural selection, proteins have been
able to evolve and contribute to the host´s ability to adapt to changes in the
environment. This process has been going on in nature for billions of years
and is responsible for that life has progressed on earth. In that time
perspective, we as humans have only very recently learnt how to modify
proteins to better fit our needs by what we refer to as protein engineering.
Such modifications often involve amino acid substitutions, deletions or
insertions, but also the creation of fusion proteins and the introduction of
various protein tags, e.g. His-tags for affinity purification.
1.2.1 Protein engineering by rational design
Rational protein engineering often relies on available detailed structural
information about a protein and defined amino acids are typically targeted for
mutations. Rational protein engineering is also referred to as knowledge-based
engineering, because already known information about the protein is used to
decide what part of a protein, or even which specific amino acid, to target and
mutate or modify.
In 1951 the amino acid sequence of a protein (insulin) was determined for the
first time, which was an important milestone in protein engineering. (Sanger,
1951;Sanger, 1951) Seven years later, in 1958, the first high resolution 3-D
7
Chemical Engineering of Small Affinity Proteins
structure of a protein (myoglobin) was determined using X-ray
crystallography. (Kendrew, 1958) The exact amino acid sequence together
with a high resolution 3-D structure of the protein of interest can make it
possible to identify important parts and key residues in that particular protein.
However, the toolbox to manipulate and modify specific residues in a
controlled and systematic manner became available in the late 1970s, when
general recombinant DNA technology together with techniques for sitedirected mutagenesis and the polymerase chain reaction (PCR) were
introduced, making it possible to mutate codons for single amino acids in a
relatively straightforward manner. (Hutchison, 1978;Mullis, 1986) Since the
1980s much research has been performed using the rational approach. One
early example is the work done on subtilisin, a protein commonly used in
washing detergents. This protein has been extensively studied and engineered
for increased catalytic efficiency and enhanced stability to high temperature,
high and low pH, and oxidative inactivation. (Wells, 1988)
Even though rational protein engineering can be very useful, it is often hard to
predict the effect certain mutations will have on function, structure and
stability of the protein. Typically, many amino acids are involved in
interacting with other molecules, and these are commonly dependent on the
overall structure of the protein for correct positioning. One single amino acid
substitution can have a huge effect on the structure and function of a protein.
In an engineered domain from the streptococcal protein G (SPG), a single
mutation was shown to promote the protein to adopt a completely different
fold, with both the secondary and tertiary structures of large parts of the
protein are changed. (He, 2012) It can be argued that this is a non-natural
protein domain and perhaps not very representative for most proteins, but it
still indicates how fragile proteins can be.
1.2.2 Protein library technology
Protein library technology relies on the creation of large repertoires of protein
variants (libraries) from which particular members can be identified via
8
Joel Lindgren
screening or selection based on desired properties, e.g. improved catalysis,
increased stability, or high affinity to a target protein.
This approach can be seen as a mimic of the natural selection process in
nature, but at a much shorter time span. Typically there has to be a physical
link between the phenotype, which is the protein in this case, and the
genotype, which is the DNA or RNA sequence encoding for that protein.
Moreover, an efficient screening or selection system has to be used to be able
to fish out and identify proteins with the desired properties. There are many
different types of selection systems available today. Cell-based systems, such as
phage display, bacterial display and yeast display, all rely on living cells to
generate the link between phenotype and genotype (Smith, 1985;Gai,
2007;Löfblom, 2011), while the cell-free systems, such as ribosome display
and mRNA display, are truly in vitro systems with no need of living host cells.
(Lipovsek, 2004)
Even when a library with many billion different proteins is used, typically only
a fraction of the theoretical diversity is covered. If for example 13 positions are
randomized, and all 20 amino acids are allowed in each position, there are
theoretically 1320 different possible protein variants. Practically, the library size
is often limited to ~109 members. (Reetz, 2008) To circumvent this potential
problem one can introduce additional diversity during the selection process,
which even further mimics evolution, e.g. by using error prone PCR.
(Cadwell, 1992) Today libraries are often constructed based on prior
information about protein function, structure and sequence, in a semi-rational
approach. Constraints on which amino acid positions that are varied and
which amino acids that are allowed in those positions may reduce the
theoretical library size dramatically and often yield libraries with a higher
degree of functional proteins. (Lutz, 2010)
1.3 Affinity proteins
As already mentioned, many proteins recognize and interact with other
molecules. Some proteins are specialized in recognizing and binding very
9
Chemical Engineering of Small Affinity Proteins
tightly to only one, or a group of, other target molecule. It is said that this
type of proteins has high affinity and selectivity for that target; thus such
proteins are often referred to as affinity proteins. The most well-known class is
antibodies, or immunoglobulins (Ig), which is the golden standard when it
comes to affinity proteins. Antibodies are naturally occurring and vital for the
immune system of humans and higher vertebrates, but they have also been
used for a long time in man-made applications as research reagents and later
also for in vivo diagnostics and therapy of cancer. (Carter, 2001) Antibodies,
and variants of antibodies, are still the most widely used and generally most
successful affinity protein, but there are today a number of non-antibody
based classes of affinity proteins, or alternative scaffold proteins as they are
often called. Some selected alternative scaffold proteins are listed in table 1.
(Friedman, 2009;Löfblom, 2011;Ruigrok, 2011) Another interesting
approach is the combination of designed polypeptide scaffolds to which small
molecules can be attached to create selective high affinity binders. (Tegler,
2011)
Common for most alternative scaffolds is that they are relatively small and
stable proteins that fold independently and have high solubility in water. The
scaffold should have a high tolerance towards mutations of certain amino
acids, without compromising the structure of the protein, which is a
prerequisite to be able to create large libraries of well-folded proteins. Two
affinity proteins central for this thesis are the Affibody molecule and the
albumin binding domain (ABD), which I will introduce in section 1.3.1 and
1.3.2, respectively.
10
Joel Lindgren
Table 1. Examples of alternative scaffold proteins.
Name
Number of
amino acids
MW (kDa)
Number of
disulfide bonds
Binding motif
Selected references
Adnectin
94
10
-
Loops
Lipovsek, 2011
Affibody
58
7
-
Helices
Löfblom, 2010
Anticalin
≈170
≈20
0-2
Loops
Schonfeld, 2009
DARPin
67 + n x 33
14 + n x 3
-
Helices + loops
Binz, 2004
Knottin
≈ 30
3
3-4
Loop
Schmidtko, 2010
Kunitz domain
58
7
3
Loop
Williams, 2003
Peptide aptamer
~100
~20
1
Loop
Kunz, 2006
1.3.1 Affibody molecules
One important class of non-antibody derived affinity proteins is the Affibody
molecule, derived from one of the immunoglobulin-binding domains of
staphylococcal protein A (SPA). SPA has five homologous domains that are
individually folded and all separately capable of binding immunoglobulins.
One of the domains, the 58 residue B-domain, had earlier been optimized for
use as affinity fusion partner through amino acid substitutions to yield the
very stable Z-domain. (Nilsson, 1987) This Z-domain folds in three antiparallel α-helices that form a compact three-helix bundle with a hydrophobic
core (illustrated in figure 4). (Wahlberg, 2003;Lendel, 2006;Lindborg, 2013)
11
Chemical Engineering of Small Affinity Proteins
Side view
Side view
Top view
Top view
Side view
Top view
Figure 3. The structure of an Affibody molecule is shown, here illustrated by the original Z
domain (pdb: 1H0T). The side chains of the thirteen amino acid positions in helices 1 and
2, which are randomized to create large Affibody molecule libraries, are shown in green.
To construct unbiased Affibody molecule libraries, thirteen surface-exposed
amino acids in helices one and two are typically randomized. (Nord,
1995;Nord, 2001;Wikman, 2004;Grönwall, 2007;Grimm, 2011;Kronqvist,
2011;Lindborg, 2011) Nine of the thirteen residues are amino acids known to
be important in binding to IgG, thereby the affinity to IgG has been removed
for the absolute majority of the different library members. (Deisenhofer,
1981;Cedergren, 1993) A further optimized scaffold has been created by
substituting additional nine residues not involved in the binding interface.
(Feldwisch, 2010) From libraries based on either the original or the second
generation scaffold, a large number of Affibody molecules with specificity to
different targets have been selected and used in a large variety of applications,
some of which are shown in figure 4. (Löfblom, 2010)
Affibody molecules typically fold quickly and are highly stable, and are small
relative to most other affinity proteins, including IgG, which has more than
12
Joel Lindgren
20 times higher molecular weight (7 kDa for Affibody molecules compared to
150 kDa for IgG). The small size makes it possible to synthesize Affibody
molecules using solid phase peptide synthesis (SPPS), an advantage discussed
further in the following chapters. The high stability and quick folding and refolding, together with the small size and relatively simple production, make
the Affibody molecules highly successful in biotechnological applications,
including affinity separation of antibodies (in e.g. MabSelect SuRe sold by GE
Healthcare). The small size gives Affibody molecules fast biodistribution and
tissue penetration, together with a rapid clearance of unbound molecules,
which are ideal properties for use as imaging agents in vivo, for e.g. cancer
diagnostics. (Tolmachev, 2008)
SPIO
SPECT
Enzyme fusion
Conjugated
fluorescence
Fc fusion
PET
MRI
Nuclide
Fused
fluorescent
protein
Bacterial toxin
Viral
Fusion
proteins
Nanoparticle
Particles
Optical
Liposome
Receptor
Enzyme
Blocking
Blocking
Cell-Cell
interaction
Detection
Enzyme fusion
IP
Microarray, FRET
Separation
Conjugations
Enzyme
Protein
ligand
MabSelect SuRe
[GeHealthcare]
Chemotoxin
Radiotherapy
Figure 4. Examples of applications of Affibody molecules (figure adopted from Löfblom et
al. 2010 with permission of the copyright holder).
13
Chemical Engineering of Small Affinity Proteins
1.3.2 Albumin binding domain (ABD)
The second affinity protein central for this thesis is the albumin binding
domain (ABD). Similarly to the Affibody molecule, ABD is also derived from
a bacterial protein, namely streptococcal protein G (SPG) from strain G148.
SPG contains three homologous immunoglobulin-binding domains and three
homologous albumin-binding domains. (Olsson, 1987) The third albuminbinding domain of protein G from the streptococcal strain G148 (abbreviated
G148-GA3), hereafter simply referred to as ABD, has been extensively
studied. The solution structure of ABD was solved using NMR, which showed
that also this protein folds as a three-helix bundle (figure 5). (Johansson,
2002)
Side view
Top view
Figure 5. The structure of the albumin binding domain (ABD) studied in this thesis is
shown, here illustrated by the structure of the last 46 amino acids in of G148-GA3 (pdb:
1GJT) The residues important for albumin binding are highlighted in red (based on
mutational analysis (Linhult, 2002))
The binding strength of proteins is often illustrated and compared by their
equilibrium dissociation constants (KD) to their target proteins. ABD has a KD
of approximately 5 nM to human serum albumin (HSA), which is a rather
high affinity, comparable to the affinity many biotechnologically used
14
Joel Lindgren
monoclonal antibodies (mAbs) have to their targets. (Linhult, 2002) By
introducing point mutations in ABD and comparing the affinity to HSA,
important amino acids were identified and it was determined that the binding
to HSA is predominantly restricted to the second helix and the loops flanking
it. (Linhult, 2002) The affinity to HSA has been improved significantly by
creating libraries and using phage display to select high affinity variants of
ABD, resulting in a variant denoted ABD035 with a KD of approximately 120
fM to HSA. (Jonsson, 2008) Furthermore, other variants based on ABD035
have also been created and used for half-life extension, discussed below. (Frejd,
2012)
1.3.2.1
ABD for half-life extension of biotherapeutics
Both IgG and HSA have remarkably long in vivo circulation half-lives, with a
half-life of approximately three weeks. Because of their large size, neither IgG
(150 kDa), nor HSA (67 kDa) are renally cleared. Moreover, the neonatal Fc
receptor (FcRn) binds endocytosed IgG and albumin and transports them to
the outside of the cell, where they are released back into the blood circulation.
The two proteins are hereby rescued from degradation in lysosomes and hence
have very long in vivo circulation half-lives. (Anderson, 2006)
It has been shown that it is possible to take advantage of the long half-life of
albumin by genetically fusing a protein of interest to albumin and thereby
getting a longer circulation half-life of the partner protein. One example where
this approach is used is in Albiglutide, a fusion protein of HSA and a dimer of
the 30 amino acid peptide GLP-1. (Bush, 2009)
An alternative approach to direct fusion is to use a high affinity binder to
albumin as gene fusion or conjugate partner to the protein of interest.
Albumin, being the most abundant protein in blood with a concentration of
~40 g/L (0.6 mM), will then likely form a non-covalent, but tightly bound
complex, with the protein conjugate in vivo and thereby potentially increase
the half-life of the protein of choice (illustrated in figure 6). One example of
this approach was the genetic fusion of an Affibody molecule dimer with an
15
Chemical Engineering of Small Affinity Proteins
engineered variant of ABD. The ABD–Affibody molecule fusion protein
showed similar in vivo half-life as the albumin control, thereby indicating that
the ABD fusion protein is rescued from degradation by FcRn as a complex
together with albumin. (Andersen, 2011)
The possibility to control the circulation time for a protein used in vivo can be
very useful. For some applications a fast in vivo clearance is desired (e.g. in
cancer diagnostics using molecular imaging) while for other applications a
long circulation time is desired (e.g. for many protein therapeutics).
(1)
Bloodstream pH 7.4
(5)
(2)
Inside of cell
(3)
(4)
Recycling
endosome
Sorting
endosome
pH 5-6
Albumin
ABD
(6)
Fusion partner
Lysosome
Other serum protein
FcRn
Figure 6. Schematic illustration of FcRn-mediated rescue of an ABD-fusion protein. ABD
with its fusion partner bind the highly abundant albumin (~40 mg/ml) in the bloodstream
(1). Protein are taken up by hematopoietic or endothelial cells by endocytosis (2). In the
sorting endosome the pH is lowered and albumin together with associated proteins bind to
membrane-bound FcRn (3). Recycling endosomes then transport the complex back to the
cell surface (4), and upon exposure to physiological pH albumin, and proteins bound to
albumin, are released from the FcRn and recycled back to the bloodstream (5). Proteins not
bound to FcRn in the sorting endosome will be degraded in the lysosome (6).
16
Joel Lindgren
2 Protein synthesis
Proteins can essentially be produced by two different methods, through
recombinant production or by chemical synthesis. In recombinant production
the protein producing machinery of host cells is used, where cells from
bacteria, yeast, insects or mammals are grown under controlled conditions.
The gene encoding for the protein of interest is inserted into the host cells,
and the protein of interest can often be produced in high yield at a relatively
low cost. After some time the cells are typically harvested and the protein of
interest purified from the complex mixture of host cell proteins.
Recombinantly produced proteins can then be modified using bioconjugation
chemistry, e.g. by coupling to cysteine residues, using maleimide chemistry
(Monji, 1978), or to primary amines (lysine residues and the N-terminus),
using succinimidyl esters.
The alternative production route is to use chemical synthesis, which has many
advantages compared to recombinant production in host cells. Using chemical
synthesis expands the possible building blocks beyond the twenty DNAencoded amino acids, enabling the use of non-natural amino acids to change
the properties of the protein or to introduce new functionalities. It also makes
it possible to build branched proteins and introduce non-amino acid building
blocks, such as polyethylene glycol (PEG), carbohydrates, and lipids. Using
chemical protein synthesis it is possible to site-specifically label a protein with
e.g. fluorescent probes, chelators for radiolabeling, or spin labels for electron
paramagnetic resonance (EPR) (Altenbach, 1990) still with an atom-to-atom
precision, which is an advantage for many applications, but particularly
important when using high resolution spectroscopic methods. Chemical
synthesis makes it possible to engineer proteins to perform better in a
particular application; as an example it is possible to extend the in vivo half-life
of a protein by increasing the hydrodynamic radius in a controlled manner by
site-specific PEGylation. (Jevsevar, 2010) Furthermore, recombinant
production of proteins in gram-negative bacteria, including E. coli, often yields
a product with contaminants such as lipopolysaccharides (LPS, often referred
17
Chemical Engineering of Small Affinity Proteins
to as endotoxins). These can be toxic if injected into the bloodstream of a
patient, even in very small amounts, and LPS can often be challenging to
remove in downstream processing of protein therapeutics. (Bito, 1977;Liu,
1997)
Chemical protein synthesis indisputably opens doors closed in classical
molecular biology, and the importance of the research field is demonstrated by
the several Nobel laureates exploring different aspects of protein synthesis,
some mentioned below.
2.1 The pre-solid phase peptide synthesis era of chemical
protein synthesis
The quest to be able to synthesize proteins chemically was initiated by Emil
Fischer, when he in 1901 was able to link two amino acids together with a
peptide bond. He used the simplest amino acid glycine to make glycyl-glycine,
and named it a dipeptide; thus the word peptide was introduced. (Fischer,
1901) Interestingly, Fischer was awarded the Nobel Prize in chemistry only
one year later, not for his work on peptides, but instead for previous work on
sugar and purine synthesis.
In the reaction when a new peptide bond is formed, the primary amine of one
amino acid attacks the carbonyl carbon of another amino acid, forming a new
covalent bond between the nitrogen and the carbonyl carbon (see figure 7).
Fischer made a homodipeptide and therefore did not need to control which
amino acid that was supposed to act as a nucleophile and which to act as an
electrophile. The coupling of two different amino acids would be more
problematic since both amino acids have a primary amine, acting as a
nucleophile, and a carboxyl group to be attacked. This problem was solved in
1932 when Bergmann and Zervas introduced the Nα-protecting group
benzyloxycarbonyl (abbreviated Cbz or Z), thereby making it possible to
protect the primary amine on one amino acid to prevent that residue to act as
a nucleophile. (Bergmann, 1932) To drive the reaction in the desired direction
18
Joel Lindgren
various activation strategies of the carboxyl group have been used, making the
carbonyl carbon more electrophilic and thus more reactive. In 1965 Sheehan
and co-workers reported the use of carbodiimides for activation of the
carboxyl group, which still today is used for some applications. (Sheehan,
1965) Many of the twenty amino acids have functional groups on the side
chains that can participate in chemical reactions and lead to unwanted side
reactions during the synthesis. Protecting groups for the different amino acid
side chains were early introduced, but have undergone further development to
function better under different conditions. (Isidro-Llobet, 2009)
O
H2N
O
OH
+
H2N
R1
OH
R2
O
AG
+
H2N
OH
N
H
PG
NH
R1
+
H2O
O
O
OH
R2
R2
R1
O
PG NH
R1
O
H2N
R2
N
H
OH
+
H AG
O
Figure 7. Top scheme: The reaction scheme for the coupling of two amino acids is shown.
In nature this reaction is catalyzed by the ribosome. Bottom scheme: The use of an amino
protecting group (PG) and a carboxyl activating group (AG) is illustrated.
After the early attempts, continuing research in the field further developed
peptide synthesis, and one major breakthrough was in 1953, when du
Vigneaud and his colleagues synthesized the first biologically active peptide.
(duVigneaud, 1953) The peptide was the nine amino acid long peptide
hormone oxytocin, and du Vigneaud was two years later awarded the Nobel
Prize in chemistry for the achievement.
2.2 Solid phase peptide synthesis (SPPS)
Up to the early 1960s, all peptide synthesis was performed using classical
organic chemistry in solution, which is time-consuming and very laborious,
and associated with severe solubility problems of the fully protected peptides.
19
Chemical Engineering of Small Affinity Proteins
In 1963 Bruce Merrifield published a paper that would truly revolutionize the
world of peptide synthesis. In that paper Merrifield founded the concept of
solid phase peptide synthesis (SPPS) that gave him the Nobel Prize in
chemistry in 1984 (the basic concepts of SPPS are illustrated in figure 7).
(Merrifield, 1963)
The key feature of SPPS is that the growing peptide being synthesized is
anchored to a solid support (resin) via the carboxyl group of the C-terminal
amino acid, thus permitting the use of excess reagents in each coupling step,
which then can be removed by simple filtration. The alpha-amine of each
subsequent amino acid is reversibly protected when it is added to the peptide
immobilized on the resin. Merrifield initially used the Cbz protecting group,
but the harsh conditions needed for deprotection were not ideal. Merrifield
quickly developed another protecting group, the tert-butyloxycarbonyl (tBoc), which could be cleaved under milder conditions, thus increasing the
synthesis yield. (Merrifield, 1964) The next amino acid to be incorporated is
added in excess, with an Nα-protecting group and an activated with an
electron-withdrawing group, making the carbonyl more reactive to
nucleophilic attack by the amino group of the previously coupled residue.
After coupling of the amino acid, any unreacted reagents can easily be filtered
away and the product bound to the solid support can be carefully washed. To
avoid polypeptides that have deletions, meaning that they miss one or several
amino acids in the middle of the sequence, irreversible capping of potential
unreacted primary amines is usually performed after the coupling step. This
can be done using acetic anhydride, which efficiently caps all free amines and
hinders them to react with other activated amino acids later in the synthesis.
This procedure yields truncated peptides instead of polypeptides with
deletions, which typically are harder to separate from the desired, full-length
product.
Next, the N-terminal protecting group can be removed and another amino
acid coupled, and this is repeated until the peptide chain is finished. Finally,
the polypeptide can be released from the solid support and the side chain
20
Joel Lindgren
O
H
T-PG N
HO
AG
Linker
Resin
Rn
PG
Attachment of first amino acid to the solid support
O
H
T-PG N
O
Rn
Linker
Resin
PG
Repeat for each amino acid
Deprotection of the temporary
amino protecting group
O
H2N
O
Linker
Resin
Rn
PG
T-PG
O
H
N
AG
Coupling of next amino acid
Rn-1
PG
PG
T-PG
N
H
Rn-1 H
N
O
O
Linker
Resin
Rn
O
PG
Final deprotection and removal of
side chain protecting groups
Cleavage from the solid support
AG
Carboxyl activating group
R
O
T-PG
PG
Temporary amino protecting group
Side chain protecting group
H2N
R1
N
H
O
H
N
O
OH
Rn
n-2
Figure 8. The basic principles of solid phase peptide synthesis are illustrated.
21
Chemical Engineering of Small Affinity Proteins
protecting groups removed; usually this is done in a one-pot reaction, yielding
the fully unprotected product in solution.
However, using Cbz or t-Boc as N-terminal protection requires repetitive acid
treatments and the use of acids with varying strengths to be able to selectively
deprotect the N-terminus or the side chains, or release the polypeptide from
the solid support. Typically one has to use extremely strong and toxic acids,
e.g. hydrofluoric acid (HF), to cleave the peptide from the solid support,
which is not only dangerous to handle, but can also lead to undesired side
reactions. Another approach, introduced by Carpino and Han in 1972, makes
it possible to use much milder reaction conditions. They developed the baselabile 9-fluorenylmethyloxycarbonyl (Fmoc) group that can be used for
protection of the alpha-amino group of amino acids. (Carpino, 1972) This
method, often referred to as Fmoc chemistry, is the most commonly used
approach in SPPS today. The Fmoc protecting group is removed by base
treatment (typically 20% piperidine), while the removal of the side chain
protecting groups and the cleavage from the solid support are achieved by
treatment with 95% triflouroacetic acid (TFA). Furthermore, today there is a
number of orthogonal protecting groups available, meaning that they can be
selectively deprotected, and thereby exposing only selected functional site(s)
for further modification. (Albericio, 2000;Isidro-Llobet, 2009)
Another important factor for successful protein synthesis is the activation of
the carboxyl group with an electron-withdrawing group, making the carbonyl
carbon more electrophilic and more susceptible to a nucleophilic attack by the
amino group of the previously coupled amino acid. The already mentioned
carbodiimides were first developed, where the C-terminal carboxylic acid is
replaced with a highly reactive O-acylisourea. (Sheehan, 1965) These coupling
reagents are very efficient, but can cause racemization of the chiral alpha
carbon, which potentially gives the polypeptide undesired properties.
Relatively soon the combination of carbodiimides and additives in the form of
triazoles such as 1-hydroxybenzotriazole (HOBt) was introduced, yielding an
active ester, which gives a reduced risk of racemization. (Künig, 1970) Some
22
Joel Lindgren
years later new coupling reagents were developed which made it possible to
activate the carboxyl group without the use of carbodiimides, e.g. 2-(1Hbenzotriazol-1-yl)-1,1,3,3,-tetramethyluronium
hexafluorophosphate
(HBTU). (Fields, 1991) There is today a large number of other commonly
used activators for amino acid coupling and SPPS. (Albericio,
2004;Pattabiraman, 2011)
The relatively simple and repetitive chemistry used in SPPS makes it well
suited for automation, and there are a number of different commercial peptide
synthesizers, for small scale parallel synthesis to larger scale synthesis. By using
Fmoc chemistry there is no need for special lab equipment other than
materials commonly found in biotechnological laboratories.
Even though SPPS is a very attractive method for the generation of proteins, it
does have some limitations, with the most severe one being the limitation in
protein length possible to synthesize, as mentioned previously. Due to the
iterative process and many coupling and deprotection steps in SPPS, the
synthesis yield of each reaction has to be very high to give a reasonable over all
synthesis yield. The theoretical total yield for a relatively short protein of 50
amino acids would only be 60% even if the yield is 99% in each step, while it
would drop to below 8% with a 95% yield in each step. In practice, the
coupling yield is not constant but instead differs between different coupling
steps and regions in the sequence, e.g. some amino acids are harder to couple
due to steric hindrance, and some regions of a sequence can be difficult due to
self-association and aggregation of the resin-bound peptide. Even though
methods to overcome these problems have been developed, such as low
loading resins and pseudo-prolines (Mutter, 1995;Valente, 2005), peptide
sequences of over 100 amino acids have been hard to synthesize with
reasonable yields. (Albericio, 2004;Kent, 2009) Given that the average protein
is approximately 300 amino acids long, other approaches have to be
considered to be able to produce a synthetic full-length protein.
23
Chemical Engineering of Small Affinity Proteins
2.3 Peptide ligation strategies
One approach to circumvent the size-limitation of SPPS can be to fuse smaller
peptide fragments to create a full-length protein. An ideal ligation method
should be selective and possible to use with unprotected side chains without
any side reactions, it should be possible to perform at mild reaction
conditions, and it should yield a native peptide bond at the site of ligation.
Some selected ligation methods are described in this section.
2.3.1 Native chemical ligation (NCL)
A significant contribution to the field of peptide ligation has been made by the
group of Stephen Kent. Interestingly, Kent did his post-doctoral work in the
lab of the founder of SPPS, the legendary Bruce Merrifield. The first ligation
approach suggested from the Kent group relied on the formation of a nonamide bond between two peptide fragments. (Schnolzer, 1992) However,
already two years later the same group reported a ligation method that would
yield a native peptide bond at the site of ligation, which they named “Native
Chemical Ligation” (NCL). (Dawson, 1994) The principle of NCL is
illustrated in figure 9, and described here in short. NCL relies of the reaction
between two polypeptides, one carrying a C-terminal thioester (Fragment 1),
and the other having an N-terminal cysteine (Fragment 2). The first step of
the ligation is a nucleophilic attack from the side chain of the N-terminal
cysteine residue in fragment 2 to the C-terminal carbonyl of fragment 1,
yielding a thioester linkage between the fragments. This orients the Nterminal amine of fragment 2 in a favorable position for another nucleophilic
attack on the same carbonyl of fragment 1, resulting in the second,
irreversible, step of the ligation, the S-to-N acyl shift, which yields a native
peptide bond at the site of ligation. NCL can be performed in aqueous
solution at neutral pH with fully unprotected peptides, and it is even possible
to have other cysteine residues in the peptides without them interfering with
the ligation reaction. The requirement of a C-terminal thioester was initially a
limitation of NCL when using Fmoc chemistry since the thioester is not stable
24
Joel Lindgren
during the repetitive base treatments used used for Fmoc deprotection.
However, today many methods for forming the thioester after completed
synthesis have been developed, where the use of a diaminobenzoyl linker is
especially straightforward. (Blanco-Canosa, 2008)
The high selectivity and efficient ligation reaction of NCL has made it very
popular and widely used. The limitation of the demand of a cysteine, which
are relatively rare in natural proteins, at the site of ligation has been
circumvented by the possibility to convert the cysteine to the more common
alanine by desulfurization after the ligation reaction. (Pentelute, 2007)
One interesting example where NCL has proved very successful is for the
synthesis of the 166 amino acid erythropoietin (EPO) glycosylated at all of the
four native glycosylation sites, resulting in a fully synthetic wild type EPO.
(Wang, 2013)
O
O
H2N
Fragment 1
S
H2N
R
O
NH
Fragment 2
OH
HS
Transthioesterification
(reversible)
O
Fragment 1
H2N
S
O
NH
H2N
Fragment 2
OH
O
S
N acyl shift
(irreversible)
O
H2N
Fragment 1
O
SH
NH
NH
Native peptide bond
Fragment 2
OH
O
Figure 9. The principle of the native chemical ligation (NCL) reaction is shown.
25
Chemical Engineering of Small Affinity Proteins
2.3.2 Traceless Staudinger ligation
Similarly as NCL, also the traceless Staudinger reaction yields a native peptide
bond at the site of ligation without any residual atoms. (Nilsson, 2000;Saxon,
2000) An advantage of the traceless Staudinger ligation is that it is not
restricted to any special amino acid residues. The method relies on the ligation
of one fragment having a C-terminal phosphinothioester and another
fragment with an N-terminal azide.
The traceless Staudinger ligation has not yet been used as much as NCL and
the practical significance of the method needs to be further verified. However,
there are some examples where the method has been fruitful. The ligation
method has for example been used to synthesize a number of different
glycoproteins (Liu, 2006) and to prepare small cyclic peptides.
(Kleineweischede, 2008) The method is well suited to be used together with
NCL since the chemistries are orthogonal, as exemplified by the synthesis of
the 124 amino acid ribonuclease A (RNase A) with an isotope-labeled proline
in position 114 as a probe for NMR studies of the protein. (Nilsson, 2003) In
this example the traceless Staudinger ligation was performed with the peptide
still attached to the solid phase.
2.3.3 Expressed protein ligation (EPL)
Expressed Protein Ligation (EPL) combines classical molecular biotechnology
with protein synthesis in an ingenious way. The method makes it possible to
combine a recombinantly produced protein with a synthetic peptide yielding a
native peptide bond at the site of ligation. (Muir, 1998) The method takes
advantage of the naturally occurring protein splicing, which is a posttranslational modification that cleaves out a protein fragment (an intein). The
desired protein gene is simply cloned into a special vector containing the
sequence of an intein. After expression and purification of the fusion protein,
the ligation reaction is performed by the addition of the peptide of interest,
containing an N-terminal cysteine, and a thiol catalyst (e.g. thiophenol),
resulting in the formation of a C-terminal thioester, which rapidly undergoes
26
Joel Lindgren
NCL. EPL has been used to chemically modify proteins in a large number of
ways, including introduction of fluorescent probes, post-translational
modifications, stable isotopes, and unnatural amino acids. (Muir, 2003)
2.3.4 Enzyme-catalyzed ligation
Another method to combine recombinant production with SPPS is by protein
ligation catalyzed by enzymes. Two examples of well-studied enzymes are
Subtiligase and Sortase A, which can ligate two proteins, or peptides, in a
clean reaction resulting in a native peptide bond. (Abrahamsen, 1991;Mao,
2004) For Sortase A-mediated ligation, the C-terminal protein fragment needs
to have an N-terminal glycine tag (two or three glycine residues), while the Nterminal fragment has to include a C-terminal LPXTGG-motif (X is an
arbitrary amino acid). Sortase-mediated ligation can thus also be used with
two synthetic peptides or two recombinantly produced proteins since no
chemical modifications are needed. However, the product will have an LPXTGGG- sequence between the ligated fragments. Sortase A-mediated
ligation can also be used to prepare backbone-cyclized proteins. (Proft, 2010)
2.3.5 Non-native ligation methods
There are a number of other ligation methods available that do not form a
native peptide bond at the site of ligation, e.g. oxime-, hydrazone-, and
thiazolidine ligation that all rely on the reaction of an aldehyde with a
hydroxylamine, a hydrazine or an aminothiol, respectively. (Shao, 1995)
Cycloaddition reactions (“Click chemistry”) are other popular methods used
by protein chemists to ligate two peptide fragments. (Kolb, 2001)
The selective reaction between thiols and haloacetyl groups is a popular and
efficient method used for peptide ligation. (Lu, 1991) The resulting bond is a
thioether which is stable and insensitive to reducing conditions, as compared
to disulfides which can be reduced and undergo thiol/disulfide exchange.
Using SPPS, haloacetyl groups can straightforwardly be site-specifically
introduced in polypeptides, e.g. on the side chain of lysine, and be used for
27
Chemical Engineering of Small Affinity Proteins
conjugation to thiol-bearing proteins, peptides or other biomolecules.
Thioalkylation can also be used for intramolecular crosslinking, by the
reaction of an introduced haloacetyl group with the side chain of cysteine.
(Ekblad, 2009)
28
Present investigation
Chemical Engineering of Small Affinity Proteins
3 Aim of the thesis
The overall aim of this thesis was to investigate if rational design and chemical
protein engineering could be used to improve the properties of the affinity
proteins ABD and Affibody molecules.
Paper I
The aim of this study was to investigate if native chemical ligation could be
used to synthesize Affibody molecules, and to develop a synthetic scheme
enabling a combinatorial assembly of Affibody molecules.
Paper II
The aim of this study was to investigate if the dimeric Affibody molecule
targeting the amyloid-beta peptide, relevant for the development of
Alzheimer’s disease, could be prepared by SPPS and if the synthetic yield
could be increased by N-terminal engineering without reducing the affinity to
the amyloid-beta peptide.
Paper III
The aim of this study was to investigate if an Affibody molecule devoid of
intrinsic fluorescence targeting amyloid-beta peptide could be developed for
use as a research tool to study the copper binding of the amyloid-beta peptide
by fluorescence spectroscopy.
Paper IV
The aim of this study was to investigate the feasibility of covalently linking
ABD to GLP-1, a bioactive peptide proven to be very beneficial for diabetic
patients, for potentially increased in vivo half-life and less frequent
administration.
30
Joel Lindgren
Paper V
The aim of this study was to investigate if intra molecular cross-linking of
ABD could enhance the protease stability of the protein, as a step towards
orally available ABD fusion proteins.
31
Chemical Engineering of Small Affinity Proteins
4 Paper I – A native chemical ligation approach for
combinatorial assembly of Affibody molecules
Depending on the application of the affinity protein, various modifications or
functional groups often need to be introduced. A fluorescent probe can be
introduced for in vitro visualization and if the protein should be used for in
vivo diagnostics, labeling with a radionuclide can be performed by
introduction of a chelator. If the affinity molecule instead is intended for use
in therapy, it may be conjugated to an effector molecule such as a cytotoxic
group. By using SPPS for the production of the affinity protein, site-specific
introduction of such reporter groups or effector molecules may be facilitated.
Moreover, several modifications can site-specifically be introduced in the same
molecule, which is generally hard to achieve with a recombinantly produced
protein.
As previously discussed in the introduction of this thesis, proteins are often
larger than 50-60 amino acids and thus challenging to produce with
reasonable yields using standard SPPS. Nevertheless, the 58-residue large
Affibody molecules have previously been successfully synthesized using direct
SPPS. (Renberg, 2005;Engfeldt, 2007;Ekblad, 2009;Heskamp, 2012) In these
examples the Affibody molecules were also site-specifically labeled with
reporter groups such as biotin moieties, fluorophores, and radionuclide
chelators for use in both in vitro and in vivo applications. However, the
synthesis yield of the Affibody molecules is highly sequence-dependent and
the synthesis often needs to be optimized for each Affibody molecule variant,
and the use of expensive pseudo-prolines is sometimes required to get
acceptable yields. A classical method to increase the synthesis yield in organic
chemistry is to use parallel synthesis of smaller sub-fragments and later
combine them to give the desired molecule. The same approach is vital also
for protein synthesis, where the shorter fragments are synthesized with high
yield followed by an efficient condensation reaction where the fragments are
joined together to form the full-length protein.
32
Joel Lindgren
Figure 10. Schematic illustration of the combinatorial assembly approach using NCL for
preparation of Affibody molecules. The target selectivity of Affibody molecules is determined
by the N-terminal amino acid sequence found in helices 1 and 2 (I). The C-terminal part
contains helix 3 (II), and is constant in all Affibody molecules. Various modifications can
readily be introduced in helix 3 at this stage. After the peptides are cleaved from the solid
support (III), NCL is used to yield the full-length Affibody molecules (IV).
In this study we wanted to investigate if the NCL method (Dawson, 1994)
could be used to produce Affibody molecules using a fragment condensation
approach. Affibody molecules consist of one variable part (helices 1 and 2),
determining the target binding specificity, and one constant part (helix 3),
needed for protein stability, making a fragment condensation approach very
appealing. It would enable combinatorial assembly of the protein, where the
same, or different, constant part could be ligated to different variable parts,
thus creating full-length proteins with different specificities and overall
properties. Thus, if different batches of constant parts are modified with
reporter groups or effector molecules, a panel of Affibody molecules with
33
Chemical Engineering of Small Affinity Proteins
different specificities and various modifications can readily be prepared, as
illustrated by the schematic figure above (figure 10).
The flexible loop between helix 2 and helix 3 was chosen as NCL site; more
specifically the peptide bond between glutamine 40 and serine 41 of the
original Affibody scaffold. To enable for the NCL two substitutions were
introduced; in position 40 a glutamine to glycine substitution for increased
flexibility and minimal steric hindrance, and in position 41 a serine to cysteine
substitution for the introduction of the required 1,2-aminothiol.
The generation of the constant C-terminal fragment with the N-terminal
cysteine was straightforward and the synthetic yield of this fragment was high
with few side products. The last amino acid, lysine 58 in the original Affibody
molecule scaffold, was introduced with a 4-methyltrityl (Mtt) protecting
group that could be selectively deprotected to give a free amine used for sitespecific conjugation to reporter groups.
The N-terminal fragments containing the C-terminal thioester were more
challenging to prepare. Since the thioester is not stabile to the repeated Fmoc
deprotections, the thioester has to be formed at the end of the synthesis.
Initially three different methods for thioester formation were evaluated; the
approaches of using a “safety catch” sulfamylbutyryl linker (Quaderer, 2001)
or an aryl hydrazine support (Camarero, 2004) were not successful in our
hands, while the use of a diaminobenzoyl linker (Blanco-Canosa, 2008)
worked very well. Using this linker, a peptide with a C-terminal N-acylbenzimidazolinone (Nbz) group was prepared and cleaved from the resin. The
peptide-Nbz is stable at low pH and under the conditions normally used
during workup and purification of peptides. However, in aqueous solution at
neutral pH the Nbz group is quickly cleaved by thiolysis and replaced by a
thioester, which can be used in NCL. The formation of the thioester was done
in situ during the NCL by the addition of the thiol catalyst 4mercaptophenylacetic acid (MPAA).
34
Joel Lindgren
The NCL reaction proved to be very selective and with few side reactions
other than the competing hydrolysis of the thioester. The reaction was
complete after approximately 3 hours incubation, illustrated by figure 11, and
after overnight incubation typically no adducts or by-products could be
detected, thus providing flexibility in the practical work procedure. The very
clean reaction also makes it possible to use unpurified (crude) starting
materials in the ligation reaction, making the preparation time significantly
shorter and less laborious.
A / mAU
400
300
200
100
0
400
300
200
100
0
MPAA
A
0
10
20
30
40
50
40
50
MPAA
B
Z(41-58)
0
ZHER2(1-40)-Nbz
Z(41-58)
10
20
t / min
ZHER2(1-58)
30
Figure 11. A typical ligation reaction between purified starting materials ZHER2(1–40)Nbz and Z(41–58) analyzed using RP-HPLC is shown. The chromatograms show the
absorbance at 220 nm for the ligation mixture at A) t=0, and B) t=3 h. MPAA (4mercaptophenylacetic acid) is the thiol catalyst used in the reaction.
Three different Affibody molecules were prepared using the NCL approach,
ZIgG(1-58) targeting IgG, ZInsulin(1-58) targeting insulin, and ZHER2(1-58)
targeting human epidermal growth factor receptor 2 (HER2). The three
Affibody molecules were analyzed using surface plasmon resonance (SPR),
where the binding to the respective target proteins was verified and the
binding affinity (KD) to each respective target correlated well with previously
reported values.
35
Chemical Engineering of Small Affinity Proteins
The constant C-terminal fragment was synthesized in very high yield,
providing a perfect starting point for further modifications. This was
exemplified by the site-specific introduction of the fluorescent probe (5-(6)carboxyfluorescein (CF)) on the side chain of lysine in position 58. This
labeled fragment was ligated to the HER2-specific helices 1 and 2 to yield the
full-length labeled Affibody molecule ZHER2(1-58)-CF. The function of this
labeled Affibody molecule was verified using fluorescence microscopy, which
clearly showed the expected staining of a HER2-overexpressing cell line and
not the negative control cell line.
To conclude, in this paper we describe a method to use NCL for the synthesis
of Affibody molecules, enabling a combinatorial assembly approach that
makes it straightforward to prepare a panel of site-specifically modified
Affibody molecules in relatively high yield.
36
Joel Lindgren
5 Paper II – N-terminal engineering of amyloid-betabinding Affibody molecules yields improved chemical
synthesis and higher binding affinity
Still today more than one hundred years after Alzheimer’s disease (AD) was
first described, the exact cause of the disease remains unknown. However, at
present, the short and aggregation-prone amyloid-beta (Aβ) peptides are
believed to be the most important factor for the onset and development of
AD. (Gilbert, 2013) Aβ peptides vary in length between 39 and 43 residues,
where Aβ (1-40) is the most abundant while Aβ (1-42) is more aggregationprone and generally considered to be the most toxic peptide. The Aβ peptides
were identified in the insoluble tangles (fibrils) in the post mortem brains of
patients diagnosed with AD for the first time almost thirty years ago. (Masters,
1985) It was long believed that the fibrils caused the disease, while today it is
generally believed that soluble oligomers of Aβ peptides are the most toxic
species. (Gilbert, 2013) Nevertheless, the exact mechanism and cause of the
disease is still unknown, and therefore more knowledge about the Aβ peptides
is needed.
Previously an Aβ peptide binding Affibody molecule, denoted ZAβ3, was
selected using phage display. (Grönwall, 2007) ZAβ3 has been shown to bind to
Aβ peptides with high affinity and was demonstrated to have the capacity to
both inhibit aggregation of Aβ peptides and dissolve previously formed fibrils.
(Hoyer, 2008;Luheshi, 2010) Furthermore, ZAβ3 has in fruit fly models shown
potential to attenuate the neurotoxic effect of Aβ peptides. (Luheshi, 2010)
The three-dimensional structure of the Aβ (1-40)/ZAβ3 complex was solved
using NMR and it was seen that the Affibody molecule binds the peptide as a
disulfide-linked homodimer. (Hoyer, 2008) It was also seen that the normal
three-helix bundle structure of the Affibody molecule was not retained for this
protein, but instead that the major part of the first helix is unstructured, while
the rest of it forms a beta strand when binding to Aβ peptides. In complex
with ZAβ3 the Aβ (1-40) peptide forms an anti-parallel β-hairpin flanked by a
37
Chemical Engineering of Small Affinity Proteins
β-strand of each Affibody molecule monomer, together forming a fourstranded anti-parallel β-sheet (see figure 12). Interestingly, the Aβ peptide is
locked in a β-hairpin conformation, which is very similar to the conformation
found in fibrils. Moreover, when the Aβ peptide is locked by the Affibody
molecule it cannot aggregate and form oligomers or fibrils.
In this study we wanted to investigate if homodimeric Affibody molecules
targeting the Aβ peptide could be prepared using SPPS, which would facilitate
site-specific modifications of the Affibody molecule for possible future
applications. We also wanted to examine if the unstructured N-terminal part
of the Affibody molecule ZAβ3 could be omitted with retained binding affinity
to the Aβ peptide, and if this N-terminal truncation would result in a higher
synthesis yield.
Based on the three-dimensional solution structure of the Aβ (1-40)/ZAβ3
complex, six N-terminally truncated variants of ZAβ3 were designed and
synthesized using standard SPPS, according to figure 12. The shortest variant
was truncated in the region that forms a beta strand when binding to the Aβ
peptide, and was thus not expected to be able to bind the peptide.
Figure 12. To the left is the 3D-structure of Aβ (1-40) peptide (in red) in complex with
the Affibody molecule dimer ZAβ3 (in dark and light blue) (based on pdb-file: 2OTK). To
the right are the sequences of the ZAβ3(1-58) Affibody molecule. The arrows indicate the
sites of truncation giving rise to the six different variants.
The synthesis yield was considerably higher for the shorter variants, where
ZAβ3(12-58), ZAβ3(15-58) and ZAβ3(18-58) had a synthetic yield of
approximately 30%, compared to 8% for the full-length ZAβ3(1-58). The
38
Joel Lindgren
different ZAβ3 variants were then dimerized and the corresponding
homodimers were isolated (hereinafter all ZAβ3 Affibody molecules referred to
in the text are understood to be disulfide-linked dimers).
Relative response (RU)
Surface plasmon resonance (SPR) was used to investigate the Aβ (1-40)
peptide binding affinity of the ZAβ3 variants (see figure 13). These experiments
showed that the shortest variant, ZAβ3(18-58), as expected had lost its affinity
to Aβ peptide, while both ZAβ3(15-58) and ZAβ3(12-58) showed high binding
affinities with KD values of 0.5 nM and 0.7 nM, respectively. This is actually
more than ten times higher affinities than the affinity determined for the fulllength ZAβ3(1-58) variant, which showed a KD of 9.5 nM. The higher binding
affinity for the shorter variants is related to an approximately ten-fold faster
on-rate compared to the full-length variant, together with essentially
unaffected off-rate kinetics.
60
50
40
ZA3(15-58)
30
ZA3(12-58)
20
10
ZA3(1-58)
ZA3(18-58)
0
0
200
400
600
800
1000
1200
1400
1600
Time (s)
Figure 13. SPR sensorgram showing the responses when different Affibody homodimers
were injected over a chip surface harboring Aβ (1–40) peptides. Injected concentrations:
ZAβ3(18–58) 200 nM; Z Aβ3 (15–58) 20 nM; Z Aβ33(12–58) 20 nM; Z Aβ33(1–58) 20
nM.
Circular dichroism spectroscopy was also used to study changes in the
secondary structure content upon titration with Aβ (1-40) peptide. These
experiments showed that there is a uniform decrease in signal when titrating
Aβ peptide to the full length ZAβ3(1-58), indicating a lower helical content in
the ZAβ3(1-58)/Aβ peptide complex than the Affibody molecule alone. This is
not as pronounced for the shorter variant ZAβ3(12-58) and especially not for
39
Chemical Engineering of Small Affinity Proteins
ZAβ3(15-58). These results indicate that the N-terminal part of the full-length
ZAβ3(1-58) is, at least partially, folded as an α-helix in the free state, but
unfolds upon binding to the Aβ peptides. In the shorter variants, ZAβ3(12-58),
and ZAβ3(15-58), the N-terminal part is deleted and thus no unfolding is
required for Aβ peptide binding, which is a plausible explanation for the faster
on-rate kinetics of the truncated variants. Again no difference was seen in the
signal from ZAβ3(18-58), alone or when Aβ peptide has been added to the
sample, again verifying the SPR results where ZAβ3(18-58) did not show any
binding to the peptide.
Isotope-labeled Aβ (1-40) peptide and 2D-NMR spectroscopy (1H15NHeteronuclear Single Quantum Correlation Nuclear Magnetic Resonance
(HSQC-NMR) was used to compare the three-dimensional structure of Aβ
peptide alone and together with ZAβ3(12-58), ZAβ3(15-58) or ZAβ3(18-58).
Once again no difference was seen when ZAβ3(18-58) was added to the Aβ
peptide, indicating that no interaction takes place between the peptide and
ZAβ3(18-58). On the contrary, clear shifts of the signals were seen when
ZAβ3(12-58) or ZAβ3(15-58) was added to the Aβ peptide, indicating that the
Affibody molecules bind the peptide and induce a structural change in the
peptide, which alter the environment for certain atoms and thereby inducing a
chemical shift that is recorded using NMR. The spectra from ZAβ3(12-58) and
ZAβ3(15-58) in complex with the Aβ peptide are similar, but not identical.
Furthermore, the spectrum from ZAβ3(12-58) is consistent with the previously
reported spectrum for the Aβ peptide in complex with the original ZAβ3
Affibody dimer, indicating a very similar structure of the Aβ peptide in
complex with ZAβ3(12-58) as in the previously determined structure in
complex with ZAβ3. (Hoyer, 2008)
In conclusion, by omitting the 11 or the 14 N-terminal amino acids of ZAβ3
the synthesis yield was increased almost four times and the binding affinity to
Aβ peptide was enhanced 10-20 times compared to the full-length ZAβ3(1-58).
The ZAβ3(12-58) binds the Aβ peptide similarly as the previously reported full-
40
Joel Lindgren
length ZAβ3 and locks the Aβ peptide in a biologically relevant β-hairpin
conformation, previously shown to efficiently prevent peptide aggregation.
41
Chemical Engineering of Small Affinity Proteins
6 Paper III – Engineered non-fluorescent Affibody
molecules facilitate studies of the amyloid-beta (Aβ) peptide
in monomeric form: low pH was found to reduce Aβ/Cu(II)
binding affinity
The aggregation of Aβ peptides is believed to be one important factor for the
onset and development of AD, described in more detail in the summary of
paper II. The aggregation of Aβ peptides is strongly pH-dependent, where a
low pH increases the aggregation rate. Inflammatory conditions, which lead to
physiological acidosis (low pH), are suggested to be involved in AD
progression. (Fuster-Matanzo, 2013) Therefore it is important to study the
behavior of the Aβ peptide with varying pH.
Furthermore, metal ions, such as copper, zinc and iron, are known to
accelerate Aβ peptide aggregation, and deposits of copper and zinc have been
found to co-localize with amyloid plaques in human AD brain tissue;
consequently metal ions are believed to be involved in AD progression.
(Gaggelli, 2006;Eskici, 2012) Previous copper binding studies using NMR
have shown that the predominant binding site is located to the N-terminal
part of the Aβ peptide, where three histidine residues (H6, H13 and H14)
together with the N-terminus bind the copper ion. (Danielsson, 2007)
Another method to study the Cu(II) binding of Aβ peptide is to use
fluorescence spectroscopy and measure the fluorescence from the single
fluorescent amino acid tyrosine 10 in the Aβ peptide. Cu(II) binding in the
proximity of the Tyr10 residue will quench the fluorescence, which makes it
possible to determine the binding affinity between Cu(II) and Aβ peptide by
titrating Cu(II) to the peptide.
In this study we wanted to investigate if an engineered variant of the Affibody
molecule ZAβ3 could be produced and functionally used to lock the Aβ (1-40)
peptide in the biologically relevant β-hairpin conformation without interfering
with the fluorescence spectroscopy measurements. For this experiment it is an
42
Joel Lindgren
advantage if the Affibody molecule is devoid of any fluorescent amino acids,
which otherwise would interfere with the fluorescence spectroscopy
measurements, and the original ZAβ3 contains one fluorescent amino acid, a
tyrosine residue in position 18.
Previous studies of the Aβ peptide in complex with ZAβ3 have shown that the
flexible N-terminus of the peptide is not involved in binding and can be
omitted with retained binding affinity. (Hoyer, 2008; paper II) Therefore we
used ZAβ3(12-58), the previously N-terminally engineered variant of ZAβ3, as a
starting point when designing three non-fluorescent Affibody molecules
targeting the Aβ peptide. Based on the results from the original phage display
selection of Affibody variants towards the Aβ peptide (Grönwall, 2007) we
chose to synthesize the homodimers of ZAβ3(12-58)Y18L and ZAβ3(1258)Y18R and the heterodimer ZAβ3(12-58)Y18L/R, since variants with leucine
or arginine in position 18 also were found to bind the Aβ peptide.
Using SPR it could be concluded that all three non-fluorescent variants of
ZAβ3(12-58) bind the Aβ peptide, both at physiological pH (pH 7.4) and
acidic pH (pH 5.0). By studying the off-rate and comparing the dissociation
phase of the sensorgrams of the different variants it was seen that ZAβ3(1258)Y18R/L showed the slowest off-rate at pH 7.4, while ZAβ3(12-58)Y18L and
ZAβ3(12-58)Y18R showed a slightly faster off-rate indicating a potentially
weaker binding. At pH 5.0 on the other hand, ZAβ3(12-58)Y18L showed the
slowest off-rate, indicating a higher affinity. Based on these results and the
more straightforward production route for the homodimer form, ZAβ3(1258)Y18L was the variant chosen for further studies. The dissociation constant
(KD) for ZAβ3(12-58)Y18L binding Aβ peptide was determined to 5.5 nM at
pH 7.4 and 0.55 nM at pH 5.0.
Isotope-labeled Aβ peptide was used to investigate the binding region of
Cu(II) to the peptide in complex with ZAβ3(12-58)Y18L using 2D-NMR
(1H15N-HSQC NMR). The data showed that all N-terminal cross-peaks
assigned in the spectrum without Cu(II) disappeared after addition of Cu(II).
On the other hand, a number of the assigned cross-peaks coming from the
43
Chemical Engineering of Small Affinity Proteins
region in the Aβ peptide that is known to bind the Affibody molecule were
unaffected by the addition of addition of Cu(II) and the assigned C-terminal
cross-peaks displayed reduced intensity, but were still visible in the presence of
Cu(II). Taken together, these results strongly indicate that Cu(II) binds the
N-terminal part of the Aβ peptide when the peptide is in complex with
ZAβ3(12-58)Y18L, similarly as previously shown for the free Aβ peptide.
Next we determined the binding constant between Cu(II) and free Aβ peptide
and Aβ peptide in complex with ZAβ3(12-58)Y18L, both at physiological (7.4)
and low (5.0) pH. This was done by titrating Cu(II) to the peptide, measuring
the fluorescence intensity, and calculating the KD from the titration curves.
For the free peptide the KD was determined to 0.86 μM at pH 7.4 and to 51
μM at pH 5.0. Thus, a more than 50-fold decreased affinity was observed at
pH 5.0 compared to at pH 7.4. This can possibly be related to protonation of
the histidine residues, with a theoretical pKa of 6, which are involved in
copper binding. When instead the Aβ peptide is in complex with ZAβ3(1258)Y18L, the KD for Cu(II) was determined to 5.4 μM at pH 7.4 and 198 μM
at pH 5.0. Again, there is approximately a 50-fold difference in affinity
between pH 7.4 and 5.0 consistent with the results for free peptide. There is a
reduced affinity, approximately five-fold increased KD, between Cu(II) and Aβ
peptide when the peptide is bound by ZAβ3(12-58)Y18L compared to free
peptide.
In conclusion, this paper showed that it is possible to use a fluorescently silent
variant of ZAβ3(12-58) to lock the Aβ peptide in a monomeric and biologically
relevant β-hairpin conformation to facilitate the use of fluorescence
spectroscopy to study copper binding of the Aβ peptide.
44
Joel Lindgren
7 Paper IV – A GLP-1 receptor agonist conjugated to an
albumin-binding domain for extended half-life
After we have eaten, insulin is released to the blood stream from the beta-cells,
which makes the liver, muscles and fat tissue take up glucose from the blood
for immediate use or storage as glycogen for future use. This process is
absolutely necessary for us to be able to live a normal life. Diabetic persons
either lack the ability to produce insulin (diabetes mellitus type 1) or have
developed insulin resistance (diabetes mellitus type 2). The classical treatment
for diabetes mellitus type 1 is by subcutaneous injection of insulin or insulin
derivatives, while diabetes mellitus type 2 sometimes can be improved by
increased exercise and dietary modification. However, diabetes mellitus type 2
often also demands medication by injected insulin. Even though insulin
injections have a long history, there are risks associated with this treatment, if
insulin is taken in too large dose there is a risk of hypoglycaemia, or too low
blood sugar, which can lead to unconsciousness. Since the frequency of type 2
diabetes has dramatically increased during the last couple of decades
(estimated 30 million in 1985 and estimated 217 million in 2005 (Smyth,
2006)), the need for safer and more convenient medications is desired.
Glucagon-like peptide 1 (GLP-1) is a peptide released from the gut in
response to food intake and it is involved in blood glucose regulation. GLP-1
promotes secretion of insulin from the beta cells and suppresses the pancreatic
secretion of glucagon, a peptide hormone that triggers the liver to convert its
glycogen to glucose. Moreover, the action of GLP-1 is glucose-dependent,
resulting in little risk of hypoglycaemia. (Baggio, 2007;Nauck, 2009) GLP-1
exists in two equipotent variants, the 30 amino acid GLP-1(7-37) and the 29
amino acid GLP-1(7-36)-NH2 with a C-terminal amide. However, the in vivo
half-life of these two peptides is less than two minutes, making it impractical
to use the endogenous peptides for therapy.
A number of different analogues of GLP-1 with enhanced in vivo circulation
half-lives have been developed and many show promising results. GLP-1
45
Chemical Engineering of Small Affinity Proteins
analogues have shown to be beneficial for patients suffering from type 2
diabetes since the peptide has shown potential to increase insulin levels,
decrease glucose levels, and restore glucose sensitivity of the beta cells. (Meier,
2012;Jones, 2013) The most successful GLP-1-based drug so far is Liraglutide
developed by Novo Nordisk (Denmark), which was approved for clinical use
in 2009 and sold for approximately $1.7 billion in 2012 alone. Liraglutide
binds non-covalently to albumin via a fatty acid introduced in the side chain
of lysine 34, thereby showing a dramatic 300-fold prolonged in vivo half-life
(13 hours) allowing for administration once daily. (Agerso, 2002;Elbrond,
2002;Juhl, 2002)
The aim of this study was to investigate the feasibility of covalently linking
ABD (introduced in section 1.3.2) to GLP-1, as a potential alternative strategy
to extend the in vivo half-life of GLP-1. By binding very strongly to HSA,
which has an in vivo half-life of approximately 3 weeks, the ABD-GLP-1
conjugate could potentially be suited for administration once a week rather
than once daily. A prerequisite for this approach to be successful is that the
ABD-GLP-1 conjugate shows a retained strong binding to HSA and an
agonistic effect on the GLP-1 receptor (GLP-1R).
An engineered variant of ABD (Jonsson, 2008;Frejd, 2012) was prepared by
standard SPPS with a lysine to cysteine mutation in position 14 to be used for
conjugation to GLP-1. The position was chosen based on the determined
three-dimensional X-ray crystal structure of a homologue ABD in complex
with HSA (Lejon, 2004) and previous studies where that position was mutated
and still showing retained HSA binding. (Alm, 2010)
GLP-1(7-37), with an alanine to glycine mutation in position 8 and a lysine to
arginine mutation in position 34 similar to Liraglutide, was synthesized using
SPPS. Lysine in position 26 was introduced with a 4-methyltrityl (Mtt)
protecting group, enabling site-specific introduction of a polyethylene glycol
(PEG)-based linker to be used as a handle in the conjugation to ABD (see
figure 14). PEG-based linkers with three different lengths were introduced,
corresponding to two, four, and six PEG units, to avoid potential steric
46
Joel Lindgren
hindrance and interference with either HSA binding or GLP-1R activation.
The linker was functionalized with a haloacetyl group, as a chemical handle
for the reaction with a thiol.
The ABD and functionalized GLP-1-PEG were cross-linked by the formation
of a thioether bond between cysteine 14 in ABD and the functionalized lysine
26 in GLP-1. The ABD-PEGn-GLP-1 conjugates were readily prepared and
the ligation reaction proved to be clean without the formation of by-products
(schematic illustration shown in figure 14).
AcCl
PEG based spacer
Lys26
Cys
ABD
GLP-1
GLP-1-(PEG)2-ABD
GLP-1-(PEG)4-ABD
GLP-1-(PEG)6-ABD
Figure 14. Schematic figure showing the ligation reaction and the three GLP-1-ABD
conjugates with varying linker lengths.
SPR was used to confirm retained HSA binding of the ABD-PEGn-GLP-1
conjugates. The conjugates were first ranked and compared to free ABD
reference protein by comparing the dissociation phases after injecting the
different ABD proteins over a surface harboring HSA. The experiments
showed that the off-rate kinetics of the three conjugates were similar to each
other and to the reference ABD. When determining the overall affinities to
47
Chemical Engineering of Small Affinity Proteins
HSA via both on-rate and off-rate kinetic measurements it became evident
that the conjugates displayed similar dissociation rate constants (kd), but that
the association rate constants (ka) were slightly lower for the ABD-PEGn-GLP1 conjugates than for free ABD. Nevertheless, the three conjugates still
showed a very high affinity, with overall KD values of approximately 200 pM
to HSA.
G
LP
-1
-
G
LP
-1
PE
G
2G
A
LP
B
D
-1
-P
EG
4G
A
LP
B
D
-1
-P
EG
6A
BD
0
P = 0.037
P = 0.044
400
P = 0.092
P = 0.120
200
0
G
G
LP
LP
-1
-1
-P
EG
2G
A
LP
B
D
-1
-P
EG
4G
A
LP
B
D
-1
-P
EG
6A
BD
200
11 mM Glucose
600
C
on
tro
l
400
Normalized Insulin secretion (%)
3 mM Glucose
600
Co
nt
ro
l
Normalized Insulin secretion (%)
To investigate if the ABD-PEGn-GLP-1 conjugates were functional and could
activate the GLP-1 receptor, insulin secretion from mouse pancreatic islets was
measured after addition of the conjugates. GLP-1 was used as a positive
control and the experiments were done in triplicates at low glucose
concentration (3 mM) and high glucose concentration (11 mM). No
significant increased insulin secretion was seen at low glucose concentration,
which was expected since the action of GLP-1 is glucose-dependent. At high
glucose concentration, all conjugates as well as free GLP-1 showed increased
insulin secretion compared to the control, although at different levels. The
increase was most significant for GLP-1 and the ABD-PEG2-GLP-1 variant.
Figure 15. Normalized insulin secretion from mouse islets at low (3 mM) glucose
concentration, and high (11 mM) glucose concentration. The mean values ± SEM are
shown.
48
Joel Lindgren
In conclusion, in this paper it was demonstrated possible to link a biologically
active peptide to “the backside” of ABD via a thioether-anchored PEG linker
and yield a conjugate with retained very high affinity to HSA and retained
biological activity of the conjugated peptide.
49
Chemical Engineering of Small Affinity Proteins
8 Paper V – Intramolecular thioether cross-linking of
therapeutic proteins to increase proteolytic stability
Compared to classical low molecular weight (LMW) drugs, protein
therapeutics potentially have a number of advantages, including higher
selectivity and potency, and have therefore gained interest and importance
during recent years. (Carter, 2011;Dimitrov, 2012) Protein-based therapeutics
have historically shown a lower risk of severe side effects and typically a faster
clinical development and approval time compared to classical LMW drugs.
(Leader, 2008;Reichert, 2003)
On the other hand, there are some drawbacks related to protein therapeutics,
such as an often high production cost and the risk of inducing an immune
response. (Jiskoot, 2012) Another significant disadvantage of protein drugs is
the incompatibility of most proteins with oral delivery of the therapeutic.
Proteins that we eat are degraded by the digestive system in the gastrointestinal
(GI) tract, where the low pH in the stomach and the proteases pepsin, trypsin
and chymotrypsin are responsible for most of the degradation, and the body
does not distinguish proteins intended as therapeutics from proteins from e.g.
a steak. Therefore protein therapeutics generally need to be administered by
injection or infusion, which can be inconvenient for the patient.
Even though the vast majority of natural proteins do not withstand the
conditions of the GI tract, there are exceptions. Two very stable protein classes
found in nature are cyclotides and cysteine-knot mini proteins, which both
have been carefully studied and engineered for various applications and have
proved to be possible to administer orally. (Werle, 2006;Craik, 2011)
Common for these protein classes is that they are stabilized by intramolecular
cross-linking by several disulfide bonds and cyclotides have additionally a
head-to-tail cyclic backbone. Intramolecular cross-linking have also shown
promising results for stabilizing peptides naturally devoid of intramolecular
disulfides by the introduction of thioether bridges. (Kluskens, 2009;Rink,
50
Joel Lindgren
2010) The thioether bond is potentially even more stable than the disulfide
since it is resistant to reducing conditions and thiol-disulfide exchange.
The aim of this study was to investigate if the proteolytic stability of small
therapeutic proteins can be enhanced by the introduction of intramolecular
thioether bridges. As a model we used the 46-residue ABD, described in the
introduction of this thesis, which binds HSA with high affinity. (Jonsson,
2008;Frejd, 2012)
A
ABD:
LAEAK
GVS
AKT
ALP
ABD_CL1:
LAEAK
GCS
AKT
AKP
ABD_CL2:
LAEAK
GVS
CKT
AKP
ABD_CL3:
CAEAK
GVS
AKT
AKP
ABD_CL12:
LAEAK
GCS
CKT
AKP
ABD_CL13:
CAEAK
GCS
AKT
AKP
B
O
Cl
SH
ABD
NH
C
3
2
3: Cys1 – Lys29
ABD
2: Lys5 – Cys28
1: Cys17 – Lys45
or
1: Cys17 – Cys45
O
NH
ABD
S
ABD
1
Figure 16. A) Schematic illustration of the design of the thioether cross-linked proteins,
with the cross-linking positions indicated by yellow, red, and blue. B) The thioalkylation
reaction between the side chain of cysteine and the chloroacetylated side chain of lysine. C)
Schematic representation of the three-helical protein, with the three cross-linking sites
indicated by colored circles connected by solid lines.
Five different ABD variants were designed to contain one or two
intramolecular cross-links according to figure 16. The cross-linking positions
were chosen based on the structure of ABD G148-GA3 (pdb file: 1GJT)
51
Chemical Engineering of Small Affinity Proteins
(Johansson, 2002) and designed to link the beginning of helix 1 to the loop
region between helices 2 and 3 (in ABD_CL2 and ABD_CL3) or the end of
helix 3 to the loop region between helices 1 and 2 (in ABD_CL1), or a
combination of the two (in ABD_CL12 and ABD_CL13).
The five ABD variants were synthesized by SPPS and prepared for
intramolecular cross-linking by the introduction of a cysteine and a lysine
residue that are in close proximity if the proteins fold as the expected threehelical bundle. The lysine residues intended to be used for cross-linking were
introduced with a 4-methyltrityl (Mtt) protecting group that selectively was
deprotected and functionalized with a chloroacetyl group. The cross-linking
reaction was initiated by dissolving the functionalized protein in buffer, where
the thiol group of the cysteine performs a nucleophilic attack on the α-carbon
of the chloroacetyl group and thus forming a thioether. The five cross-linked
ABD variants were all successfully prepared; the cross-linking reaction proved
to be were selective with few side reactions, especially for the three variants
containing one cross-link, consequently it was not necessary to purify the
linear proteins before the reaction.
The HSA binding of the five cross-linked proteins was analyzed using SPR
and compared to a reference ABD. The experiment showed that ABD_CL2
and ABD_CL12 had lost the ability to bind HSA, while ABD_CL13 showed
a considerably affected low affinity binding to HSA. On the other hand, both
ABD_CL3 and ABD_CL1 still showed high affinity binding to HSA. The
binding kinetics to HSA for ABD_CL1 and the ABD reference protein were
determined, and both displayed affinities in the low pM range, approximately
20 pM.
The secondary structure content of the cross-linked ABD variants was studied
using circular dichroism (CD) and compared to the ABD reference protein.
These experiments showed that ABD_CL1 and the reference protein both are
predominantly α-helical, showing the typical CD spectra with distinct local
minima at 208 and 221 nm. The four other cross-linked proteins do not show
52
Joel Lindgren
the characteristic CD spectra, indicating a partly or completely lost alphahelical secondary structure content for these variants.
The solution three-dimensional structure of ABD G148-GA3 shows that the
N-terminus of the protein is part of the first α-helix. This can possibly explain
the radical loss of helicity for all variants having a cross-link to that part of the
protein (i.e. all variants except ABD_CL1). Additionally, the last three amino
acids of the C-terminus seem unstructured, which potentially makes this part
of the protein less sensitive to modifications.
The melting temperatures for ABD_CL1 and the reference protein were
determined using CD measurements, where the Tm for ABD_CL1 (63°C)
was found to be 5°C higher than for the ABD reference protein (58°C),
indicating an enhanced thermal stability of ABD_CL1 compared to the
reference protein. None of the other cross-linked proteins showed a clear
transition and therefore no Tm could be calculated from the variable
temperature measurements.
Next the proteolytic stability of the cross-linked ABD variants was studied and
compared to the reference protein. Initial screening experiments showed a
considerably increased stability toward both pepsin and the combination of
trypsin and chymotrypsin for ABD_CL1 compared to the reference ABD,
while none of the other variants showed an improved stability under the
conditions used in these experiments.
The protease stability of ABD_CL1 and the reference protein was studied
more thoroughly by following the degradation over time when challenged
with pepsin or trypsin and chymotrypsin. The results from these experiments
clearly showed that the proteolytic stability of ABD_CL1 was substantially
enhanced, particularly against pepsin, visualized by figure 17. When
comparing the area under curve (AUC) of ABD_CL1 to the ABD reference
protein, it is evident that the total amount of intact protein present during the
60 minutes challenged with pepsin is at least 17 times higher for ABD_CL1,
53
Chemical Engineering of Small Affinity Proteins
ABD_CL1
100
ABD
50
0
0
20
40
Time (min)
60
Residual intact protein (%)
Residual intact protein (%)
and almost 5 times higher for ABD_CL1 when challenged with the
combination of trypsin and chymotrypsin.
ABD_CL1
100
ABD
50
0
0
20
40
60
Time (min)
Figure 17. Proteolytic degradation of ABD_CL1 and the ABD reference protein by pepsin
at pH 2.6, 37°C to the left, and by trypsin and chymotrypsin at pH 7.4, 37°C to the right.
A variant of ABD_CL1 compatible with recombinant production of the ABD
protein was also synthesized, having another cysteine in position 45 instead of
the lysine residue. By using a dibromide linker the two cysteine residues form
thioether bridges with the linker, thus cross-linking the ABD in the same
positions as ABD_CL1. This variant unfortunately showed lower proteolytic
stability than ABD_CL1, and only displayed slightly improved proteolytic
stability compared to the ABD reference protein. The Tm of the variant was
also decreased (Tm ~55°C) and found to be lower than the Tm of the ABD
reference protein. A possible explanation for the dissimilarity between this
variant and ABD_CL1, even though cross-linked in the same positions, can be
related to the structural difference of the linkers. Even though the linkers are
of approximately the same length, the investigated haloacetyl linkers are more
bulky and possible interfere with the protein folding, making the three-helical
bundle less compact and thus more susceptible for proteolytic degradation and
less thermostable.
Mass spectrometry (MS) was used to investigate the major degradation
products of the ABD reference protein. The peptide bond between leucine 42
and alanine 43 was rapidly cleaved when challenged with pepsin, and the
corresponding protein fragment is the dominating product after 3 minutes.
The fragment is then relatively rapidly further cleaved, especially between
54
Joel Lindgren
phenylalanine 20 and tyrosine 21. When the ABD reference protein instead is
challenged with the combination of trypsin and chymotrypsin, the most
susceptible peptide bond is between tyrosine 21 and lysine 22, which proved
to be the predominant cleavage site. Taken together, these results show that
the region Phe20-Tyr21-Lys22-Arg23 in the ABD reference protein is a
region especially sensitive to proteolysis by the three investigated enzymes. The
cross-linking in ABD_CL1 between Cys17 and Lys45 evidently suppresses the
proteolysis in this region, as well as between position 42 and 43, potentially by
sterically interfering with the access of the proteases to these susceptible
regions. In previous work, the amino acids in position 20-24 were found to be
important for albumin binding. (Linhult, 2002) This indicates that traditional
mutation of susceptible amino acids may not be feasible to use for this protein.
Moreover, the concept of using an intramolecular cross-linking possibly also
increases the stability towards a wider range of proteases compared to single
amino acid substitutions.
To conclude, in this study it was shown that it is possible to substantially
increase the proteolytic stability by the introduction an intramolecular crosslink. In this paper a tethered variant of ABD was presented, cross-linked by a
thioether bridge, resulting in increased thermostability and enhanced stability
towards pepsin, trypsin and chymotrypsin.
55
Chemical Engineering of Small Affinity Proteins
9 Concluding remarks
The five papers that this thesis is based upon all involve chemical synthesis of
the affinity proteins ABD or Affibody molecules. Both protein classes are
small and can therefore relatively easily be synthesized, which makes it possible
to chemically modify the proteins with acceptable total yield. Furthermore,
the small size and the high stability of the proteins make them highly
competitive to other affinity protein classes, including antibodies.
In the first paper, an alternative production route for Affibody molecules using
a fragment condensation approach was described. The method makes it
possible to prepare a large set of full-length Affibody molecules with different
C-terminal modifications from a set of pre-produced sub-fragments, which
can be very convenient if several variants need to be screened. In the second
and third papers, Affibody molecules targeting the Aβ peptide involved in AD
were studied. Firstly, the synthesis yield was considerably improved by
omitting the first 11 residues, but it somewhat surprisingly also increased the
binding affinity to the Aβ peptide at least ten times for ZAβ3(12-58). These
results were recently further explored in affinity maturation studies of the ZAβ3
Affibody molecule using staphylococcal display, where N-terminally truncated
variants were shown to have both increased surface expression level and
affinity to the Aβ peptide compared to full-length variants. (Lindberg, 2013)
In the third paper, the improved variant ZAβ3(12-58) was further modified,
and this study was facilitated by the increased synthesis yield. Moreover, the
increased affinity of ZAβ3(12-58) was highly beneficial as it could compensate
for the reduced (approximately ten-fold) affinity caused by the substitution of
the fluorescent tyrosine residue to create the fluorescently silent ZAβ3(1258)Y18L. In paper III it was also shown that it is possible to lock the Aβ
peptide in a biologically relevant β-hairpin conformation and thereby study
the peptide without any risk of aggregation, thus making it easier to study the
peptide in conditions that normally would be problematic due to
oligomerization and fibril formation. The possibility to synthesize sitespecifically modified Aβ-binding Affibody molecules in high yield gives
56
Joel Lindgren
improved opportunities to develop new methods for studying the Aβ peptide,
e.g. fluorimetric assays based on fluorescent-labeled Aβ-binding Affibody
molecules.
ABD was studied in the fourth and fifth papers. In paper IV we linked ABD
to the biologically active peptide GLP-1 via the linkage of two side chains.
Position 14 in ABD was chosen for conjugation since it is positioned on the
opposite side as the albumin-binding site, potentially resulting in low risk of
interference with albumin-binding. An advantage of using chemical synthesis
is that any position in ABD can be selected for conjugation, and as predicted,
conjugation to GLP-1 via this position did not reduce the albumin-binding
properties of the conjugate. The conjugation of biologically active proteins,
peptides, or even small molecules to ABD for increased circulation half-life
seems very promising. It would be very interesting to evaluate the ABD-GLP1 conjugates in vivo, to investigate both the biological effect, and the effect of
ABD conjugation on the in vivo circulation half-lives, in diabetic mouse
models. I believe that in the near future this, or similar, approach will be much
more widely spread. The approach has the potential to expand the portfolio of
methods used to tailor molecules as drugs that in themselves are impossible to
use due to poor in vivo properties, as for example GLP-1.
In paper V a further example of the use of chemical synthesis was illustrated
by the introduction of thioether bridges to improve the properties of ABD.
This resulted in a variant, ABD_CL1, with considerably enhanced proteolytic
stability. With the increasing interest of protein-based therapeutics, there is a
need for stabilizing proteins to withstand the harsh conditions related to oral
administration. ABD_CL1 shows substantially increased stability towards the
three most important proteases in the GI tract and the low pH in the stomach,
compared to the non-crosslinked ABD reference protein. Such an improved
stability can result in significantly higher bioavailability of a protein after oral
delivery, and the thioether cross-linking approach is possibly general and can
also be used for other protein-based therapeutics. With the use of alkylating
agents, such as dihalides or trihalides, intramolecular cross-linking
(investigated in paper V) of larger proteins, not possible to produce with
57
Chemical Engineering of Small Affinity Proteins
standard SPPS, becomes available. It also makes it possible to use selection
systems such as phage display to screen libraries of peptides or proteins
constrained by thioether bridges. (Angelini, 2012;Chen, 2013)
Taken together the work presented in this thesis demonstrates the potential of
using solid phase peptide synthesis and engineering by rational design to
modify small proteins.
Since protein therapeutics show such great potential, there is a need to develop
efficient production routes for such drugs. I believe that synthetic, and
especially semi-synthetic, production routes will be further developed to meet
the new needs. The possibility to improve the properties of proteins and
peptides by the introduction of unnatural amino acids, building blocks such as
PEG or lipids, and the introduction of stabilizing links such as intramolecular
thioether bonds, makes synthetic or semi-synthetic production of the protein
drug highly competitive compared to recombinant production.
58
Joel Lindgren
Populärvetenskaplig sammanfattning
Proteiner beskrivs ibland som cellers arbetare. De sköter de flesta funktionerna
i cellen, men är även ansvariga för mycket annat utanför själva cellerna. Hos
oss människor är det till exempel proteiner som plockar upp syre från
lungorna och transporterar ut det till resten av kroppen. En annan typ av
välkända proteiner är antikroppar tillverkade av vårt immunsystem och med
uppgift att känna igen ovälkomna gäster i våra kroppar, såsom t.ex. bakterier.
Antikropparna binder ofta hårt till den främmande molekylen och gör så att
vår kropp kan känna igen och oskadliggöra inkräktaren. Proteiner som binder
andra molekyler hårt, med hög s.k. affinitet, kallas för affinitetsproteiner och
kan vara väldigt användbara för många tillämpningar. Olika antikroppar,
riktade mot olika målmolekyler, används t. ex. världen över som verktyg i
laboratorier samt för diagnostik och behandling av t.ex. bröstcancer.
Idag finns det även flera typer av affinitetsproteiner som kan designas för att
binda till olika typer av molekyler; jag har jobbat med s.k. Affibody-molekyler
som kan ses som en typ av konstgjorda antikroppar. Genom olika metoder
kan Affibody-molekyler som binder med hög affinitet till ett speciellt protein
fiskas fram. T. ex. finns det Affibody-molekyler som binder till proteinet
HER2, som är överuttryckt på vissa cancerceller och som kan användas för att
diagnostisera cancerpatienter. En annan framtagen Affibody-molekyl binder
till peptiden amyloid-beta, som tros vara inblandad i utvecklingen av
Alzheimers sjukdom. Affibody-molekyler är mindre och mer stabila än
antikroppar vilket ger dem en mängd fördelar; de tränger t.ex. in bättre i
tumörer och kan ge snabbare svar om de används för cancerdiagnostik. För
många tillämpningar behöver proteinet märkas in med reportergrupper, t. ex.
färgämnen eller radioaktiva ämnen, som gör det möjligt att se var proteinet
har ansamlats i kroppen och på så vis använda Affibody-molekyler för
cancerdiagnostik.
Ett annat protein som jag har arbetat med är ett litet albuminbindande
protein, förkortat ABD, som binder stenhårt till albumin som är det mest
59
Chemical Engineering of Small Affinity Proteins
förekommande proteinet i blodet och som har en väldigt lång cirkulationstid i
kroppen. Tack vare att ABD binder till albumin åker också ABD runt i blodet
längre tid, och genom att sätta ihop ett läkemedel med ABD kan effekten av
läkemedlet förlängas. På så vis kan ABD användas för s.k.
halveringstidsförlängning av i stort sett vilka läkemedel som helst.
Både Affibody-molekyler och ABD är små proteiner, bestående av 58
respektive 46 byggstenar (s.k. aminosyror) vilket gör det möjligt att tillverka
dem genom kemisk syntes. Det underlättar och är i vissa fall det enda sättet för
att modifiera proteinerna på ett kontrollerat sätt och det möjliggör också
användandet av andra byggstenar än de 20 naturligt förekommande
aminosyrorna, vilket kan ge proteinet nya fördelaktiga egenskaper. Det är
också lättare att få proteinet helt rent om det produceras genom kemisk syntes,
jämfört med produktion i bakterier som är det vanligaste andra sättet. Det är
självklart en stor fördel om proteinet ska användas som läkemedel att det inte
finns några rester från bakterier kvar. Det är dock fortfarande inte helt lätt att
tillverka proteiner genom kemisk syntes på grund av de många syntesstegen,
men bättre och effektivare syntesmetoder utvecklas hela tiden. Att undersöka
möjligheten att genom kemisk syntes tillverka och förbättra Affibodymolekyler och ABD är en central del i denna avhandling.
I projekt I undersökte vi om det går producera Affibody-molekyler genom att
tillverka två separata fragment som sedan sätts ihop. De aminosyror som avgör
vad Affibody-molekylen binder till sitter i den första delen av proteinet,
medan den sista delen är konstant i alla Affibody-molekyler. Därför vore det
behändigt om man separat kunde göra en konstant del, med eventuella
reportergrupper, som kan sättas ihop med önskad första del av proteinet och
således bilda ett funktionellt protein. Resultaten visar att det är möjligt och vi
beskriver en ny metod för att effektivt producera olika Affibody-molekyler på
ett smidigt sätt genom kemisk syntes.
I projekt II och III har vi jobbat med Affibody-molekyler som binder amyloidbeta-peptiden, en peptid som kan bilda långa trådar (fibriller) som ofta hittas i
hjärnan hos avlidna personer som har haft Alzheimers sjukdom. I projekt II
60
Joel Lindgren
undersökte vi om det går göra just den Affibody-molekylen mindre så att den
blir enklare att tillverka genom kemisk syntes. Resultaten från denna studie
visade att det går göra proteinet mindre utan att förstöra funktionen, vilket gav
ökat syntetiskt utbyte. Affiniteten till amyloid-beta-peptiden visade sig inte
försämras, utan i stället har vår förkortade variant, som vi kallar ZAβ3(12-58),
minst tio gånger starkare bindning till amyloid-beta-peptiden. I projekt III
abetade vi vidare med den förbättrade Affibody-molekylen för att undersöka
om den går använda som ett verktyg för att låsa amyloid-beta-peptiden i en
speciell struktur och studera peptidens bindning till kopparjoner. Förhöjda
halter av koppar har hittats hos personer med Alzheimers sjukdom och det
spekuleras i om dessa påverkar amyloid-beta-peptiden och på så vis har
inverkan på sjukdomsförloppet. Resultaten från denna studie visar att den
modifierade Affibody-molekylen lämpar sig bra för att användas i studier av
amyloid-beta-peptiden och kan vara ett hjälpmedel i framtida forskning för att
undersöka hur Alzheimers sjukdom uppstår, något man idag inte vet helt
säkert.
I de sista två projekten har jag arbetat med proteinet ABD istället för
Affibody-molekyler. I projekt IV undersökte vi om det går sätta ihop ABD
med en peptid, GLP-1, som har visat sig ge positiv effekt på personer som har
typ 2-diabetes. GLP-1-peptiden bryts ned på mindre än två minuter i kroppen
och kan därför inte själv användas som läkemedel. Det finns ett liknande
läkemedel baserat på GLP-1 (Liraglutid) på marknaden som tas genom en
spruta en gång om dagen. Detta läkemedel sålde för över 11 miljarder kronor
under 2012, vilket visar vilken enorm marknad det finns. Vår variant skulle
potentiellt kunna reducera antalet injektioner från en om dagen till en per
vecka, något som självklart skulle vara fördelaktigt för användarna. Resultaten
från denna studie visar att ABD-GLP-1-molekylen fortfarande binder hårt till
albumin, och borde därför ha en lång halveringtid i kroppen. I våra försök
visar molekylen jämförbar verkan som GLP-1, vilket typer på att den
fortfarande är aktiv och borde ha samma biologiska effekt i kroppen som
GLP-1 eller läkemedlet Liraglutid.
61
Chemical Engineering of Small Affinity Proteins
Gemensamt för alla proteinläkemedel är att de inte kan tas som ett piller utan
behöver i stället injiceras (som t. ex. insulin). Det beror på att vår kropp inte
skiljer på proteiner från maten vi äter och proteiner som ätits för att fungera
som läkemedel, utan båda bryts effektivt ned i magen och tunntarmen och tas
upp som små byggstenar för att tillverka nya proteiner i kroppens celler. Det
forskas en hel del på att få fram proteinbaserade läkemedel som kan tas som
piller istället för att injiceras. Ett exempel på detta är att det danska
läkemedelsbolaget Novo Nordisk (samma bolag som tagit fram Liraglutid)
nyligen gick ut med att de satsar 23 miljarder kronor på att utveckla läkemedel
mot diabetes som kan tas i tablettform istället för med spruta. I projekt V
studerade vi möjligheten att göra ABD mer stabilt mot proteaser. Detta
gjordes genom att modifiera vissa aminosyror i ABD och låta dem bilda nya
bindningar till varandra. Resultaten från detta projekt visade att vår bästa
variant, som vi kallar för ABD_CL1, har avsevärt högre stabilitet mot de tre
viktigaste proteaserna i mag-tarmkanalen. De övertygande resultaten från
denna studie indikerar att sådan korslänkning kan användas för att stabilisera
proteiner och är ett steg mot ett proteinläkemedel som kan tas som ett piller.
I majoriteten av alla läkemedel som används idag, som t. ex. alvedon och
magnecyl, är den aktiva substansen inte proteiner utan s.k. små-molekyler.
Idag har dock mycket forskningsfokus skiftat från små molekyler till större
proteinbaserade läkemedel. Proteinläkemedel har ofta färre sidoeffekter och är
därför säkrare än små molekyler, de har historiskt sett också snabbare tagit sig
genom de väldigt kostsamma kliniska prövningar som görs innan ett
läkemedel blir godkänt, vilket självklart är en fördel för läkemedelsbolagen.
Det är dock fortfarande många problem som måste lösas innan den fulla
potentialen hos proteiner som läkemedel kan utnyttjas och proteinläkemedel
kan ta en större del av marknaden, där det faktum att de måste injiceras är ett
stort problem.
Sammantaget visar arbetet som behandlas i denna avhandling att det genom
kemisk syntes går modifiera små proteiner och optimera deras egenskaper för
att få dem att fungera bättre i de avsedda tillämpningarna.
62
Joel Lindgren
Acknowledgments
Då var avhandlingen snart färdig för tryck, det känns skönt. Det är många jag
vill tacka som gjort det möjligt för mig att att skriva denna bok och om allt går
vägen hämta ut min doktorstitel om någon månad.
Först och främst Amelie, tack för att du har varit en helt fantastisk handledare!
Du har från dag ett alltid tagit dig tid för frågor och diskussioner, även om du
många gånger egentligen haft alldeles för mycket att göra. Tack för en väldigt
trevlig och lärorik tid!
Min bihandledare Per-Åke, vi har inte direkt pratat varje vecka under de här
åren, men du har ändå alltid haft full koll på vad jag jobbat med och de inlägg
och förslag du har är alltid högst relevanta. Tack också för noggrann
genomläsning av boken!
Torbjörn, tack för att du kvalitetsgranskade avhandlingen!
Tack alla PIs på plan 3 för att ni, även om forskarvärlden generellt är ganska
pressande och stressig, möjliggör en skön och avslappnad atmosfär på jobbet!
Sen vill jag passa på att tacka VINNOVA och Vetenskapsrådet för
finansiering.
Projektledarna som jag samarbetat med från Affibody AB: Lars, Caroline och
Fredrik. Det är alltid lika roligt att komma till Affibody för möten med er. Ni
är alltid väl insatta och har input i världsklass till våra projekt. Tack!
Alla övriga som jag kommit i kontakt med på Affibody AB: Först Tove, du
var verkligen den bästa handledare jag kan tänka mig som exjobbare. Sen blev
jag kvar i proteinreningsgruppen med Pelle som chef, det var en trevlig tid,
även om det inte blev så länge. Tack Elin för att du tipsade mig om att Amelie
sökte en doktorand. Sen vill jag också tacka alla andra på Affibody AB för en
trevlig tid hos er, det är fortfarande alltid trevligt att komma och hälsa på och
köra CD.
63
Chemical Engineering of Small Affinity Proteins
Jag vill också tacka medförfattarna i amyloid-beta projekten, särskilt
Sebastian W, det har varit väldigt roligt att samarbeta med dig.
I would also like to thank the group of P-O Berggren at KI for a very
interesting collaboration within the ABD-GLP-1 project.
Vidare vill jag tacka alla alla ni som har jobbat på plan 3 under min
doktorandtid. Alla behövs för att skapa den härliga helhet som vi har där på
AlbaNova. Speciellt vill jag tacka de jag jobbat närmast, alla i syntesgruppen:
Andreas, Anna K, Anna P, Alexis, Danne, Cissi, Emily, Kristina, Melina,
Nima, Patrik, Peter, Sara och Viktor, för att ni har gjort arbetet extra roligt.
Sen vill jag också tacka alla andra vänner som gjort att jag kommit ihåg att
det finns en värld utanför forskningen på KTH.
Min familj med Mamma, Pappa, Malin och Erik som gett mig en trygg bra
uppväxt, utan den skulle jag aldrig skrivit den här boken. Tack!
Tack också Furbergs för att ni tagit emot mig med öppna armar, och för hjälp
med barnpassning då och då, det har underlättat arbetet.
Till sist och viktigast av alla, min nya lilla växande familj. Sara, tänk att det
redan gått drygt 9 år sedan vi var på bio tillsammans för första gången, det har
hänt en hel del sedan dess. Tack för att du är du! Viktor min son, du kom
som en härlig vitaminkick mitt i doktorandtiden. Tack för att du är helt
underbar och gör det väldigt enkelt att sätta saker och ting i perspektiv och
förstå vad som verkligen är viktigt. Sen till vårt ofödda lilla knyte, tack för att
du satte lite tidspress på mig, utan dig hade det säkert tagit lite längre tid att
bli färdig. Jag älskar er tre över allt annat!
64
Joel Lindgren
References
Abrahmsén, L., J. Tom, J. Burnier, K. A. Butcher, A. Kossiakoff and J. A.
Wells (1991). "Engineering subtilisin and its substrates for efficient ligation of
peptide bonds in aqueous solution." Biochemistry 30(17): 4151-4159.
Agerso, H., L. B. Jensen, B. Elbrond, P. Rolan and M. Zdravkovic (2002).
"The pharmacokinetics, pharmacodynamics, safety and tolerability of
NN2211, a new long-acting GLP-1 derivative, in healthy men." Diabetologia
45(2): 195-202.
Albericio, F. (2000). "Orthogonal protecting groups for N(alpha)-amino and
C-terminal carboxyl functions in solid-phase peptide synthesis." Biopolymers
55(2): 123-139.
Albericio, F. (2004). "Developments in peptide and amide synthesis." Curr
Opin Chem Biol 8(3): 211-221.
Alberts, B., A. Johnson, J. Lewis, M. Raff, K. Roberts and P. Walter (2002).
Molecular Biology of the Cell. New York, Garland Science.
Alm, T., L. Yderland, J. Nilvebrant, A. Halldin and S. Hober (2010). "A small
bispecific protein selected for orthogonal affinity purification." Biotechnol J
5(6): 605-617.
Altenbach, C., T. Marti, H. G. Khorana and W. L. Hubbell (1990).
"Transmembrane protein structure: spin labeling of bacteriorhodopsin
mutants." Science 248(4959): 1088-1092.
Andersen, J. T., R. Pehrson, V. Tolmachev, M. B. Daba, L. Abrahmsén and
C. Ekblad (2011). "Extending half-life by indirect targeting of the neonatal Fc
receptor (FcRn) using a minimal albumin binding domain." J Biol Chem
286(7): 5234-5241.
Anderson, C. L., C. Chaudhury, J. Kim, C. L. Bronson, M. A. Wani and S.
Mohanty (2006). "Perspective-- FcRn transports albumin: relevance to
immunology and medicine." Trends Immunol 27(7): 343-348.
65
Chemical Engineering of Small Affinity Proteins
Angelini, A., J. Morales-Sanfrutos, P. Diderich, S. Chen and C. Heinis
(2012). "Bicyclization and tethering to albumin yields long-acting peptide
antagonists." J Med Chem 55(22): 10187-10197.
Baggio, L. L. and D. J. Drucker (2007). "Biology of incretins: GLP-1 and
GIP." Gastroenterology 132(6): 2131-2157.
Bergmann, M. and L. Zervas (1932). "Über ein allgemeines Verfahren der
Peptid-Synthese." Ber. dtsch. Chem. Ges. A/B 65: 1192-1201.
Binz, H. K., P. Amstutz, A. Kohl, M. T. Stumpp, C. Briand, P. Forrer, M. G.
Grutter and A. Pluckthün (2004). "High-affinity binders selected from
designed ankyrin repeat protein libraries." Nat Biotechnol 22(5): 575-582.
Bito, L. Z. (1977). "Inflammatory effects of endotoxin-like contaminants in
commonly used protein preparations." Science 196(4285): 83-85.
Blanco-Canosa, J. B. and P. E. Dawson (2008). "An efficient Fmoc-SPPS
approach for the generation of thioester peptide precursors for use in native
chemical ligation." Angew Chem Int Ed Engl 47(36): 6851-6855.
Brocchieri, L. and S. Karlin (2005). "Protein length in eukaryotic and
prokaryotic proteomes." Nucleic Acids Res 33(10): 3390-3400.
Bush, M. A., J. E. Matthews, E. H. De Boever, R. L. Dobbins, R. J. Hodge, S.
E. Walker, M. C. Holland, M. Gutierrez and M. W. Stewart (2009). "Safety,
tolerability, pharmacodynamics and pharmacokinetics of albiglutide, a longacting glucagon-like peptide-1 mimetic, in healthy subjects." Diabetes Obes
Metab 11(5): 498-505.
Cadwell, R. C. and G. F. Joyce (1992). "Randomization of genes by PCR
mutagenesis." PCR Methods Appl 2(1): 28-33.
Camarero, J. A., B. J. Hackel, J. J. de Yoreo and A. R. Mitchell (2004).
"Fmoc-based synthesis of peptide alpha-thioesters using an aryl hydrazine
support." J Org Chem 69(12): 4145-4151.
Carpino, L. A. and G. Y. Han (1972). "9-Fluorenylmethoxycarbonyl aminoprotecting group." The Journal of Organic Chemistry 37(22): 3404-3409.
66
Joel Lindgren
Carter, P. (2001). "Improving the efficacy of antibody-based cancer
therapies." Nat Rev Cancer 1(2): 118-129.
Carter, P. J. (2011). "Introduction to current and future protein therapeutics:
a protein engineering perspective." Exp Cell Res 317(9): 1261-1269.
Cedergren, L., R. Andersson, B. Jansson, M. Uhlén and B. Nilsson (1993).
"Mutational analysis of the interaction between staphylococcal protein A and
human IgG1." Protein Eng 6(4): 441-448.
Chen, S., I. Rentero Rebollo, S. A. Buth, J. Morales-Sanfrutos, J. Touati, P.
G. Leiman and C. Heinis (2013). "Bicyclic peptide ligands pulled out of
cysteine-rich peptide libraries." J Am Chem Soc 135(17): 6562-6569.
Craik, D. J. and A. C. Conibear (2011). "The chemistry of cyclotides." J Org
Chem 76(12): 4805-4817.
Danielsson, J., R. Pierattelli, L. Banci and A. Gräslund (2007). "Highresolution NMR studies of the zinc-binding site of the Alzheimer's amyloid
beta-peptide." FEBS J 274(1): 46-59.
Dawson, P. E., T. W. Muir, I. Clark-Lewis and S. B. Kent (1994). "Synthesis
of proteins by native chemical ligation." Science 266(5186): 776-779.
Deisenhofer, J. (1981). "Crystallographic refinement and atomic models of a
human Fc fragment and its complex with fragment B of protein A from
Staphylococcus aureus at 2.9- and 2.8-A resolution." Biochemistry 20(9):
2361-2370.
Dimitrov, D. S. (2012). "Therapeutic proteins." Methods Mol Biol 899: 126.
Ekblad, T., V. Tolmachev, A. Orlova, C. Lendel, L. Abrahmsén and A. E.
Karlström (2009). "Synthesis and chemoselective intramolecular crosslinking
of a HER2-binding affibody." Biopolymers 92(2): 116-123.
Elbrond, B., G. Jakobsen, S. Larsen, H. Agerso, L. B. Jensen, P. Rolan, J.
Sturis, V. Hatorp and M. Zdravkovic (2002). "Pharmacokinetics,
pharmacodynamics, safety, and tolerability of a single-dose of NN2211, a
67
Chemical Engineering of Small Affinity Proteins
long-acting glucagon-like peptide 1 derivative, in healthy male subjects."
Diabetes Care 25(8): 1398-1404.
Engfeldt, T., A. Orlova, T. Tran, A. Bruskin, C. Widstrom, A. E. Karlström
and V. Tolmachev (2007). "Imaging of HER2-expressing tumours using a
synthetic Affibody molecule containing the 99mTc-chelating mercaptoacetylglycyl-glycyl-glycyl (MAG3) sequence." Eur J Nucl Med Mol Imaging 34(5):
722-733.
Eskici, G. and P. H. Axelsen (2012). "Copper and oxidative stress in the
pathogenesis of Alzheimer's disease." Biochemistry 51(32): 6289-6311.
Feldwisch, J., V. Tolmachev, C. Lendel, N. Herne, A. Sjöberg, B. Larsson, D.
Rosik, E. Lindqvist, G. Fant, I. Höidén-Guthenberg, J. Galli, P. Jonasson and
L. Abrahmsén (2010). "Design of an optimized scaffold for affibody
molecules." J Mol Biol 398(2): 232-247.
Fields, C. G., D. H. Lloyd, R. L. Macdonald, K. M. Otteson and R. L. Noble
(1991). "HBTU activation for automated Fmoc solid-phase peptide
synthesis." Pept Res 4(2): 95-101.
Frejd, F. Y. (2012). Half-Life Extension by Binding to Albumin through an
Albumin Binding Domain. Therapeutic Proteins: Strategies to Modulate
Their Plasma Half-Lives. R. Kontermann, Wiley-VCH Verlag GmbH & Co.
KGaA, Weinheim, Germany.
Friedman, M. and S. Ståhl (2009). "Engineered affinity proteins for tumourtargeting applications." Biotechnol Appl Biochem 53(Pt 1): 1-29.
Fuster-Matanzo, A., M. Llorens-Martin, F. Hernandez and J. Avila (2013).
"Role of neuroinflammation in adult neurogenesis and Alzheimer disease:
therapeutic approaches." Mediators Inflamm 2013: 260925.
Gaggelli, E., H. Kozlowski, D. Valensin and G. Valensin (2006). "Copper
homeostasis and neurodegenerative disorders (Alzheimer's, prion, and
Parkinson's diseases and amyotrophic lateral sclerosis)." Chemical Reviews
106(6): 1995-2044.
68
Joel Lindgren
Gai, S. A. and K. D. Wittrup (2007). "Yeast surface display for protein
engineering and characterization." Curr Opin Struct Biol 17(4): 467-473.
Gilbert, B. J. (2013). "The role of amyloid beta in the pathogenesis of
Alzheimer's disease." J Clin Pathol 66(5): 362-366.
Grimm, S., F. Yu and P. Å. Nygren (2011). "Ribosome display selection of a
murine IgG(1) Fab binding affibody molecule allowing species selective
recovery of monoclonal antibodies." Mol Biotechnol 48(3): 263-276.
Grönwall, C., A. Jonsson, S. Lindström, E. Gunneriusson, S. Ståhl and N.
Herne (2007). "Selection and characterization of Affibody ligands binding to
Alzheimer amyloid beta peptides." J Biotechnol 128(1): 162-183.
He, Y., Y. Chen, P. A. Alexander, P. N. Bryan and J. Orban (2012).
"Mutational tipping points for switching protein folds and functions."
Structure 20(2): 283-291.
Heskamp, S., P. Laverman, D. Rosik, F. Boschetti, W. T. van der Graaf, W. J.
Oyen, H. W. van Laarhoven, V. Tolmachev and O. C. Boerman (2012).
"Imaging of human epidermal growth factor receptor type 2 expression with
18F-labeled affibody molecule ZHER2:2395 in a mouse model for ovarian
cancer." J Nucl Med 53(1): 146-153.
Hoyer, W., C. Grönwall, A. Jonsson, S. Ståhl and T. Härd (2008).
"Stabilization of a beta-hairpin in monomeric Alzheimer's amyloid-beta
peptide inhibits amyloid formation." Proc Natl Acad Sci U S A 105(13):
5099-5104.
Hutchison, C. A., 3rd, S. Phillips, M. H. Edgell, S. Gillam, P. Jahnke and M.
Smith (1978). "Mutagenesis at a specific position in a DNA sequence." J Biol
Chem 253(18): 6551-6560.
Isidro-Llobet, A., M. Alvarez and F. Albericio (2009). "Amino acid-protecting
groups." Chem Rev 109(6): 2455-2504.
Jevsevar, S., M. Kunstelj and V. G. Porekar (2010). "PEGylation of
therapeutic proteins." Biotechnol J 5(1): 113-128.
69
Chemical Engineering of Small Affinity Proteins
Jiskoot, W., T. W. Randolph, D. B. Volkin, C. R. Middaugh, C. Schoneich,
G. Winter, W. Friess, D. J. Crommelin and J. F. Carpenter (2012). "Protein
instability and immunogenicity: roadblocks to clinical application of injectable
protein delivery systems for sustained release." J Pharm Sci 101(3): 946-954.
Johansson, M. U., I. M. Frick, H. Nilsson, P. J. Kraulis, S. Hober, P.
Jonasson, M. Linhult, P. Å. Nygren, M. Uhlén, L. Bjorck, T. Drakenberg, S.
Forsen and M. Wikstrom (2002). "Structure, specificity, and mode of
interaction for bacterial albumin-binding modules." J Biol Chem 277(10):
8114-8120.
Jones, L. H. and D. A. Price (2013). "Medicinal chemistry of glucagon-like
peptide receptor agonists." Prog Med Chem 52: 45-96.
Jonsson, A., J. Dogan, N. Herne, L. Abrahmsén and P. Å. Nygren (2008).
"Engineering of a femtomolar affinity binding protein to human serum
albumin." Protein Eng Des Sel 21(8): 515-527.
Juhl, C. B., M. Hollingdal, J. Sturis, G. Jakobsen, H. Agerso, J. Veldhuis, N.
Porksen and O. Schmitz (2002). "Bedtime administration of NN2211, a
long-acting GLP-1 derivative, substantially reduces fasting and postprandial
glycemia in type 2 diabetes." Diabetes 51(2): 424-429.
Kendrew, J. C., G. Bodo, H. M. Dintzis, R. G. Parrish, H. Wyckoff and D.
C. Phillips (1958). "A three-dimensional model of the myoglobin molecule
obtained by x-ray analysis." Nature 181(4610): 662-666.
Kent, S. B. (2009). "Total chemical synthesis of proteins." Chem Soc Rev
38(2): 338-351.
Kleineweischede, R. and C. P. Hackenberger (2008). "Chemoselective peptide
cyclization by traceless Staudinger ligation." Angew Chem Int Ed Engl
47(32): 5984-5988.
Kluskens, L. D., S. A. Nelemans, R. Rink, L. de Vries, A. Meter-Arkema, Y.
Wang, T. Walther, A. Kuipers, G. N. Moll and M. Haas (2009).
"Angiotensin-(1-7) with thioether bridge: an angiotensin-converting enzymeresistant, potent angiotensin-(1-7) analog." J Pharmacol Exp Ther 328(3):
849-854.
70
Joel Lindgren
Kolb, H. C., M. G. Finn and K. B. Sharpless (2001). "Click Chemistry:
Diverse Chemical Function from a Few Good Reactions." Angew Chem Int
Ed Engl 40(11): 2004-2021.
Kronqvist, N., M. Malm, L. Göstring, E. Gunneriusson, M. Nilsson, I.
Höidén Guthenberg, L. Gedda, F. Y. Frejd, S. Ståhl and J. Löfblom (2011).
"Combining phage and staphylococcal surface display for generation of
ErbB3-specific Affibody molecules." Protein Eng Des Sel 24(4): 385-396.
Künig, W. and R. Geiger (1970). "Eine neue Methode zur Synthese von
Peptiden: Aktivierung der Carboxylgruppe mit Dicyclohexylcarbodiirnid
unter Zusatz von 1-Hydrox y-benzotriazolen." Chem. Ber. 103: 788-789.
Kunz, C., C. Borghouts, C. Buerger and B. Groner (2006). "Peptide aptamers
with binding specificity for the intracellular domain of the ErbB2 receptor
interfere with AKT signaling and sensitize breast cancer cells to Taxol." Mol
Cancer Res 4(12): 983-998.
Leader, B., Q. J. Baca and D. E. Golan (2008). "Protein therapeutics: a
summary and pharmacological classification." Nat Rev Drug Discov 7(1): 2139.
Lejon, S., I. M. Frick, L. Björck, M. Wikström and S. Svensson (2004).
"Crystal structure and biological implications of a bacterial albumin binding
module in complex with human serum albumin." J Biol Chem 279(41):
42924-42928.
Lendel, C., J. Dogan and T. Härd (2006). "Structural basis for molecular
recognition in an affibody:affibody complex." J Mol Biol 359(5): 1293-1304.
Li, S., W. Guo, B. M. Schmitt and M. L. Greaser (2012). "Comprehensive
analysis of titin protein isoform and alternative splicing in normal and mutant
rats." J Cell Biochem 113(4): 1265-1273.
Lindberg, H., A. Johansson, T. Härd, S. Ståhl and J. Löfblom (2013).
"Staphylococcal display for combinatorial protein engineering of a head-to-tail
affibody dimer binding the Alzheimer amyloid-beta peptide." Biotechnol J
8(1): 139-145.
71
Chemical Engineering of Small Affinity Proteins
Lindborg, M., E. Cortez, I. Höidén-Guthenberg, E. Gunneriusson, E. von
Hage, F. Syud, M. Morrison, L. Abrahmsén, N. Herne, K. Pietras and F. Y.
Frejd (2011). "Engineered high-affinity affibody molecules targeting plateletderived growth factor receptor beta in vivo." J Mol Biol 407(2): 298-315.
Lindborg, M., A. Dubnovitsky, K. Olesen, T. Björkman, L. Abrahmsén, J.
Feldwisch and T. Härd (2013). "High-affinity binding to staphylococcal
protein A by an engineered dimeric Affibody molecule." Protein Eng Des Sel
26(10): 635-644.
Linhult, M., H. K. Binz, M. Uhlén and S. Hober (2002). "Mutational
analysis of the interaction between albumin-binding domain from
streptococcal protein G and human serum albumin." Protein Sci 11(2): 206213.
Lipovsek, D. (2011). "Adnectins: engineered target-binding protein
therapeutics." Protein Eng Des Sel 24(1-2): 3-9.
Lipovsek, D. and A. Pluckthün (2004). "In-vitro protein evolution by
ribosome display and mRNA display." J Immunol Methods 290(1-2): 51-67.
Liu, S., R. Tobias, S. McClure, G. Styba, Q. Shi and G. Jackowski (1997).
"Removal of endotoxin from recombinant protein preparations." Clin
Biochem 30(6): 455-463.
Löfblom, J. (2011). "Bacterial display in combinatorial protein engineering."
Biotechnol J 6(9): 1115-1129.
Löfblom, J., J. Feldwisch, V. Tolmachev, J. Carlsson, S. Ståhl and F. Y. Frejd
(2010). "Affibody molecules: engineered proteins for therapeutic, diagnostic
and biotechnological applications." FEBS Lett 584(12): 2670-2680.
Löfblom, J., F. Y. Frejd and S. Ståhl (2011). "Non-immunoglobulin based
protein scaffolds." Curr Opin Biotechnol 22(6): 843-848.
Lu, Y. A., P. Clavijo, M. Galantino, Z. Y. Shen, W. Liu and J. P. Tam (1991).
"Chemically unambiguous peptide immunogen: preparation, orientation and
antigenicity of purified peptide conjugated to the multiple antigen peptide
system." Mol Immunol 28(6): 623-630.
72
Joel Lindgren
Luheshi, L. M., W. Hoyer, T. P. de Barros, I. van Dijk Hard, A. C. Brorsson,
B. Macao, C. Persson, D. C. Crowther, D. A. Lomas, S. Ståhl, C. M. Dobson
and T. Härd (2010). "Sequestration of the Abeta peptide prevents toxicity and
promotes degradation in vivo." PLoS Biol 8(3): e1000334.
Lutz, S. (2010). "Beyond directed evolution--semi-rational
engineering and design." Curr Opin Biotechnol 21(6): 734-743.
protein
Mao, H., S. A. Hart, A. Schink and B. A. Pollok (2004). "Sortase-mediated
protein ligation: a new method for protein engineering." J Am Chem Soc
126(9): 2670-2671.
Masters, C. L., G. Multhaup, G. Simms, J. Pottgiesser, R. N. Martins and K.
Beyreuther (1985). "Neuronal origin of a cerebral amyloid: neurofibrillary
tangles of Alzheimer's disease contain the same protein as the amyloid of
plaque cores and blood vessels." EMBO J 4(11): 2757-2763.
Meier, J. J. (2012). "GLP-1 receptor agonists for individualized treatment of
type 2 diabetes mellitus." Nat Rev Endocrinol 8(12): 728-742.
Merrifield, R. B. (1963). "Solid Phase Peptide Synthesis. I. The Synthesis of a
Tetrapeptide." J Am Chem Soc 85(14): 2149-2154.
Merrifield, R. B. (1964). "Solid Phase Peptide Synthesis. II. The Synthesis of
Bradykinin." J Am Chem Soc 86(2): 304-305.
Monji, N., H. Malkus and A. Castro (1978). "Maleimide derivative of hapten
for coupling to enzyme: a new method in enzyme immunoassay." Biochem
Biophys Res Commun 85(2): 671-677.
Muir, T. W. (2003). "Semisynthesis of proteins by expressed protein ligation."
Annu Rev Biochem 72: 249-289.
Muir, T. W., D. Sondhi and P. A. Cole (1998). "Expressed protein ligation: a
general method for protein engineering." Proc Natl Acad Sci U S A 95(12):
6705-6710.
73
Chemical Engineering of Small Affinity Proteins
Mullis, K., F. Faloona, S. Scharf, R. Saiki, G. Horn and H. Erlich (1986).
"Specific enzymatic amplification of DNA in vitro: the polymerase chain
reaction." Cold Spring Harb Symp Quant Biol 51 Pt 1: 263-273.
Mutter, M., A. Nefzi, T. Sato, X. Sun, F. Wahl and T. Wohr (1995).
"Pseudo-prolines (psi Pro) for accessing "inaccessible" peptides." Pept Res
8(3): 145-153.
Nauck, M. A. (2009). "Unraveling the science of incretin biology." Am J Med
122(6 Suppl): S3-S10.
Nilsson, B. L., R. J. Hondal, M. B. Soellner and R. T. Raines (2003). "Protein
assembly by orthogonal chemical ligation methods." J Am Chem Soc 125(18):
5268-5269.
Nilsson, B. L., L. L. Kiessling and R. T. Raines (2000). "Staudinger ligation: a
peptide from a thioester and azide." Org Lett 2(13): 1939-1941.
Nord, K., J. Nilsson, B. Nilsson, M. Uhlén and P. Å. Nygren (1995). "A
combinatorial library of an alpha-helical bacterial receptor domain." Protein
Eng 8(6): 601-608.
Nord, K., O. Nord, M. Uhlén, B. Kelley, C. Ljungqvist and P. Å. Nygren
(2001). "Recombinant human factor VIII-specific affinity ligands selected
from phage-displayed combinatorial libraries of protein A." Eur J Biochem
268(15): 4269-4277.
Olsson, A., M. Eliasson, B. Guss, B. Nilsson, U. Hellman, M. Lindberg and
M. Uhlén (1987). "Structure and evolution of the repetitive gene encoding
streptococcal protein G." Eur J Biochem 168(2): 319-324.
Park, K., I. C. Kwon and K. Park (2011). "Oral protein delivery: Current
status and future prospect." Reactive and Functional Polymers 71(3): 280287.
Pattabiraman, V. R. and J. W. Bode (2011). "Rethinking amide bond
synthesis." Nature 480(7378): 471-479.
74
Joel Lindgren
Pauling, L. and R. B. Corey (1951). "Configuration of polypeptide chains."
Nature 168(4274): 550-551.
Pauling, L. and R. B. Corey (1951). "The pleated sheet, a new layer
configuration of polypeptide chains." Proc Natl Acad Sci U S A 37(5): 251256.
Pentelute, B. L. and S. B. Kent (2007). "Selective desulfurization of cysteine
in the presence of Cys(Acm) in polypeptides obtained by native chemical
ligation." Org Lett 9(4): 687-690.
Proft, T. (2010). "Sortase-mediated protein ligation: an emerging
biotechnology tool for protein modification and immobilisation." Biotechnol
Lett 32(1): 1-10.
Quaderer, R. and D. Hilvert (2001). "Improved synthesis of C-terminal
peptide thioesters on "safety-catch" resins using LiBr/THF." Org Lett 3(20):
3181-3184.
Rask-Andersen, M., M. S. Almen and H. B. Schioth (2011). "Trends in the
exploitation of novel drug targets." Nat Rev Drug Discov 10(8): 579-590.
Reetz, M. T., D. Kahakeaw and R. Lohmer (2008). "Addressing the numbers
problem in directed evolution." Chembiochem 9(11): 1797-1804.
Reichert, J. M. (2003). "Trends in development and approval times for new
therapeutics in the United States." Nat Rev Drug Discov 2(9): 695-702.
Renberg, B., I. Shiroyama, T. Engfeldt, P. K. Nygren and A. E. Karlström
(2005). "Affibody protein capture microarrays: synthesis and evaluation of
random and directed immobilization of affibody molecules." Anal Biochem
341(2): 334-343.
Renukuntla, J., A. D. Vadlapudi, A. Patel, S. H. Boddu and A. K. Mitra
(2013). "Approaches for enhancing oral bioavailability of peptides and
proteins." Int J Pharm 447(1-2): 75-93.
75
Chemical Engineering of Small Affinity Proteins
Rink, R., A. Arkema-Meter, I. Baudoin, E. Post, A. Kuipers, S. A. Nelemans,
M. H. Akanbi and G. N. Moll (2010). "To protect peptide pharmaceuticals
against peptidases." J Pharmacol Toxicol Methods 61(2): 210-218.
Ruigrok, V. J., M. Levisson, M. H. Eppink, H. Smidt and J. van der Oost
(2011). "Alternative affinity tools: more attractive than antibodies?" Biochem
J 436(1): 1-13.
Sanger, F. and H. Tuppy (1951). "The amino-acid sequence in the
phenylalanyl chain of insulin. 2. The investigation of peptides from enzymic
hydrolysates." Biochem J 49(4): 481-490.
Sanger, F. and H. Tuppy (1951). "The amino-acid sequence in the
phenylalanyl chain of insulin. I. The identification of lower peptides from
partial hydrolysates." Biochem J 49(4): 463-481.
Santos, R. A., A. C. Simoes e Silva, C. Maric, D. M. Silva, R. P. Machado, I.
de Buhr, S. Heringer-Walther, S. V. Pinheiro, M. T. Lopes, M. Bader, E. P.
Mendes, V. S. Lemos, M. J. Campagnole-Santos, H. P. Schultheiss, R. Speth
and T. Walther (2003). "Angiotensin-(1-7) is an endogenous ligand for the G
protein-coupled receptor Mas." Proc Natl Acad Sci U S A 100(14): 82588263.
Saxon, E., J. I. Armstrong and C. R. Bertozzi (2000). "A "traceless"
Staudinger ligation for the chemoselective synthesis of amide bonds." Org Lett
2(14): 2141-2143.
Schmidtko, A., J. Lotsch, R. Freynhagen and G. Geisslinger (2010).
"Ziconotide for treatment of severe chronic pain." Lancet 375(9725): 15691577.
Schnolzer, M. and S. B. Kent (1992). "Constructing proteins by dovetailing
unprotected synthetic peptides: backbone-engineered HIV protease." Science
256(5054): 221-225.
Schonfeld, D., G. Matschiner, L. Chatwell, S. Trentmann, H. Gille, M.
Hulsmeyer, N. Brown, P. M. Kaye, S. Schlehuber, A. M. Hohlbaum and A.
Skerra (2009). "An engineered lipocalin specific for CTLA-4 reveals a
76
Joel Lindgren
combining site with structural and conformational features similar to
antibodies." Proc Natl Acad Sci U S A 106(20): 8198-8203.
Shao, J. and J. P. Tam (1995). "Unprotected Peptides as Building Blocks for
the Synthesis of Peptide Dendrimers with Oxime, Hydrazone, and
Thiazolidine Linkages." J. Am. Chem. Soc. 117(14): 3893-3899.
Sheehan, J. C., J. Preston and P. A. Cruickshank (1965). "A Rapid Synthesis
of Oligopeptide Derivatives without Isolation of Intermediates." J Am Chem
Soc 87: 2492-2493.
Smith, G. P. (1985). "Filamentous fusion phage: novel expression vectors that
display cloned antigens on the virion surface." Science 228(4705): 1315-1317.
Smyth, S. and A. Heron (2006). "Diabetes and obesity: the twin epidemics."
Nat Med 12(1): 75-80.
Swaminathan, J. and C. Ehrhardt (2012). "Liposomal delivery of proteins and
peptides." Expert Opin Drug Deliv 9(12): 1489-1503.
Tegler, L. T., G. Nonglaton, F. Buttner, K. Caldwell, T. Christopeit, U. H.
Danielson, K. Fromell, T. Gossas, A. Larsson, P. Longati, T. Norberg, R.
Ramapanicker, J. Rydberg and L. Baltzer (2011). "Powerful protein binders
from designed polypeptides and small organic molecules--a general concept for
protein recognition." Angew Chem Int Ed Engl 50(8): 1823-1827.
Tolmachev, V. (2008). "Imaging of HER-2 overexpression in tumors for
guiding therapy." Curr Pharm Des 14(28): 2999-3019.
Uhlén, M., E. Bjorling, C. Agaton, C. A. Szigyarto, B. Amini, E. Andersen, A.
C. Andersson, P. Angelidou, A. Asplund, C. Asplund, L. Berglund, K.
Bergstrom, H. Brumer, D. Cerjan, M. Ekstrom, A. Elobeid, C. Eriksson, L.
Fagerberg, R. Falk, J. Fall, M. Forsberg, M. G. Bjorklund, K. Gumbel, A.
Halimi, I. Hallin, C. Hamsten, M. Hansson, M. Hedhammar, G. Hercules,
C. Kampf, K. Larsson, M. Lindskog, W. Lodewyckx, J. Lund, J. Lundeberg,
K. Magnusson, E. Malm, P. Nilsson, J. Odling, P. Oksvold, I. Olsson, E.
Oster, J. Ottosson, L. Paavilainen, A. Persson, R. Rimini, J. Rockberg, M.
Runeson, A. Sivertsson, A. Skollermo, J. Steen, M. Stenvall, F. Sterky, S.
Stromberg, M. Sundberg, H. Tegel, S. Tourle, E. Wahlund, A. Walden, J.
77
Chemical Engineering of Small Affinity Proteins
Wan, H. Wernerus, J. Westberg, K. Wester, U. Wrethagen, L. L. Xu, S.
Hober and F. Ponten (2005). "A human protein atlas for normal and cancer
tissues based on antibody proteomics." Mol Cell Proteomics 4(12): 19201932.
Wahlberg, E., C. Lendel, M. Helgstrand, P. Allard, V. Dincbas-Renqvist, A.
Hedqvist, H. Berglund, P. Å. Nygren and T. Härd (2003). "An affibody in
complex with a target protein: structure and coupled folding." Proc Natl Acad
Sci U S A 100(6): 3185-3190.
Valente, A. P., F. C. Almeida, C. R. Nakaie, S. Schreier, E. Crusca, Jr. and E.
M. Cilli (2005). "Study of the effect of the peptide loading and solvent system
in SPPS by HRMAS-NMR." J Pept Sci 11(9): 556-563.
Vane, J. R. and R. M. Botting (2003). "The mechanism of action of aspirin."
Thromb Res 110(5-6): 255-258.
Wang, P., S. Dong, J. H. Shieh, E. Peguero, R. Hendrickson, M. A. Moore
and S. J. Danishefsky (2013). "Erythropoietin derived by chemical synthesis."
Science 342(6164): 1357-1360.
Wells, J. A. and D. A. Estell (1988). "Subtilisin--an enzyme designed to be
engineered." Trends Biochem Sci 13(8): 291-297.
Werle, M., T. Schmitz, H. L. Huang, A. Wentzel, H. Kolmar and A.
Bernkop-Schnurch (2006). "The potential of cystine-knot microproteins as
novel pharmacophoric scaffolds in oral peptide drug delivery." J Drug Target
14(3): 137-146.
Vickery, H. B. (1950). "The origin of the word protein." Yale J Biol Med
22(5): 387-393.
Vigneaud, V. d., C. Ressler, C. J. M. Swan, C. W. Roberts, P. G. Katsoyannis
and S. Gordon (1953). "THE SYNTHESIS OF AN OCTAPEPTIDE
AMIDE WITH THE HORMONAL ACTIVITY OF OXYTOCIN." J Am
Chem Soc 75(19): 4878-4879.
78
Joel Lindgren
Wikman, M., A. C. Steffen, E. Gunneriusson, V. Tolmachev, G. P. Adams, J.
Carlsson and S. Ståhl (2004). "Selection and characterization of HER2/neubinding affibody ligands." Protein Eng Des Sel 17(5): 455-462.
Williams, A. and L. G. Baird (2003). "DX-88 and HAE: a developmental
perspective." Transfus Apher Sci 29(3): 255-258.
Witt-Enderby, P. A., J. Bennett, M. J. Jarzynka, S. Firestine and M. A. Melan
(2003). "Melatonin receptors and their regulation: biochemical and structural
mechanisms." Life Sci 72(20): 2183-2198.
Yun, Y., Y. W. Cho and K. Park (2013). "Nanoparticles for oral delivery:
targeted nanoparticles with peptidic ligands for oral protein delivery." Adv
Drug Deliv Rev 65(6): 822-832.
79