Embed Size (px)
Citation preview
Research Collection
Doctoral Thesis
The SH3 domain of fyn kinase as a scaffold for the generation ofnew binding proteins
Author(s): Grabulovski, Dragan
Publication Date: 2007
Permanent Link: https://doi.org/10.3929/ethz-a-005407897
Rights / License: In Copyright - Non-Commercial Use Permitted
This page was generated automatically upon download from the ETH Zurich Research Collection. For moreinformation please consult the Terms of use.
ETH Library
DISS. ETH Nr. 17216
The SH3 Domain of Fyn Kinase as a Scaffold for the Generation
of New Binding Proteins
A dissertation thesis submitted to the ETH Zürich to obtain the degree of
DOKTOR DER WISSENSCHAFTEN
presented by
Dragan Grabulovski
Eidg. Dipl. Apotheker, ETH Zürich
born 9.3.1978
citizen of Winterthur (ZH)
accepted on the recommendation of
Prof. Dr. Dario Neri, examiner
Prof. Dr. Gerd Folkers, co-examiner
2007
To my parents
2
Table of contents
SUMMARY
1 INTRODUCTION
1.1 Recombinant proteins as biopharmaceuticals
1.1.1 Monoclonal antibody based drugs
1.2 Small globular proteins as antibody substitutes
1.2.1 Single domain antibodies
1.2.2 Tenth type III domain of fibronectin
1.2.3 A-domains
1.2.4 Z-domain of staphylococcal protein A
1.2.5 Kunitz type domains
1.2.6 Ankyrin repeat proteins
1.2.7 Lipocalins
1.2.8 Other domains
1.2.9 Concluding remarks
1.3 The SH3 domain of Fyn kinase as a scaffold for the generation of new binding protei
1.3.1 Src homology-3 (SH3) domains
1.3.2 SH3 domains as scaffolds for the generation of engineered binding proteins
1.3.3 SH3 domain of Fyn kinase
1.4 Isolation of binding domains
1.4.1 General aspects
1.4.2 Iterative colony filter screening
1.4.3 Phage display
1.4.4 Yeast display
1.4.5 Ribosome display
1.4.6 Covalcnt DNA display
1.5 Serum albumin as a target protein
1.6 Vascular tumor targeting
1.6.1 Concept and definitions
1.6.2 Classes of vascular targets
1.6.3 Imaging applications of vascular tumor targeting
3
1.6.4 Therapy applications of vascular tumor targeting 61
1.7 Aim of this thesis 64
2 RESULTS 66
2.1 Expression of Fyn SH3 wild-type and Fyn SH3 mutants 66
2.1.1 Cloning and expression of Fyn SH3 wild-type 66
2.1.2 Test library designs and characterization 67
2.2 Phage library design, cloning and characterization 69
2.3 Generation of Fyn SH3 derived proteins binding to mouse serum albumin (MSA) 72
2.3.1 Phage display selections of Fyn SH3 mutants binding to MSA 72
2.3.2 Affinity maturation of G4 using phage display 73
2.3.3 Characterization of affinity matured MSA binders isolated by phage display 76
2.4 Construction of a novel Fyn SH3 phage library 79
2.4.1 Phage library design, cloning and characterization 79
2.5 Generation of Fyn SH3 derived proteins binding to the extra-domain-B of fibronectin (EDB) 82
2.5.1 Isolation and in vitro characterization of Fyn SH3 mutants specific to the EDB of fibronectin, a
marker of angiogenesis 82
2.5.2 Biodistribution studies in tumor bearing mice 89
2.6 Immunogenicity studies 91
3 DISCUSSION 93
4 MATERIALS AND METHODS 98
5 REFERENCES 112
6 CURRICULUM VITAE 144
7 ACKNOWLEDGEMENTS 148
4
Summary
Therapeutic proteins have become an increasingly important class of medicines during the
last two decades. After the first successes with recombinant proteins in replacement
therapies, monoclonal antibodies are currently representing the second wave of innovation
created by the biotechnological industry. Antibodies have proved useful as human
therapeutics because they bind to drug targets with high affinity and specificity and exhibit a
favorable phannacokinetic profile.
However, even blockbuster biologies suffer from certain drawbacks. Nature did not evolve
them for manufacture ex vivo and for this reason, difficulties in their production often arise.
Poor expression yields, low solubility or immunogenicity are sometimes problematic aspects
of protcin-bascd therapeutics. Additionally, thermal instability and aggregate formation arc
known to increase the immunogenic potential of therapeutic proteins. Therefore, there is a
growing interest in the development of stable and highly soluble proteins from human origin,
which may be less immunogenic in therapeutic applications and easier to produce in
microbial expression systems.
According to this rationale, an emerging field in pharmaceutical biotechnology is represented
by the generation of novel binding molecules based on small globular domains serving as
protein frameworks („scaffolds"). In this approach, certain amino acid residues of the surface
of a single domain are combinatorially mutated to produce a protein library, which can then
be screened for binding specificities of interest. Thus, the concept of a universal binding site
from the antibody structure is transferred to alternative protein frameworks with suitable
biophysical properties. On one hand, these new proteins will provide further insights into the
processes of molecular recognition, on other hand, they could also have commercial
applications, as therapeutic agents, diagnostic reagents or affinity ligands.
In this thesis, we describe the design, construction and characterization of a human Fyn SH3
phage library. After test selections and successful isolation of Fyn SH3 derived binders to
mouse serum albumin (MSA), a novel, optimized phage library was created, comprising 1.2 x
109 library members. Using this second library, wc present the isolation and in vitro
5
characterization of Fyn SH3 derived proteins binding to the extra-domain B of fibronectin
(EDB), a marker of angiogenesis. One specific binding clone, named D3, was further
evaluated and showed a remarkable ability to stain vascular structures in tumor sections.
Furthermore, quantitative biodistribution studies in tumor bearing mice revealed the ability of
D3 to selectively accumulate in the tumor. In contrast to human scFv antibody fragments
administered to mice, neither Fyn SH3wt nor the D3 mutant were immunogenic in mice, after
four i.v. injections.
The EDB-binding D3 protein opens new biomedical opportunities for the in vivo imaging of
solid tumors and for the delivery of toxic agents to the tumoral vasculature. In addition, the
single-pot library of Fyn SH3 mutants described in this thesis may represent a rich source of
useful reagents for many biochemical and biomedical applications as an alternative to more
conventional IgG-based immunochemical technologies.
6
Zusammenfassung
Während den letzten 20 Jahren sind therapeutische Proteine zu einer immer wichtigeren
Klasse von Medikamenten geworden. Nach den ersten Erfolgen mit Substitutionstherapien
repräsentieren nun monoklonale Antikörper die zweite Welle von innovativen
Medikamenten, welche aus den Aktivitäten der biotechnologischen Industrie hervorgegangen
sind. Antikörper menschlichen Ursprungs haben ihre Nützlichkeit als therapeutische Proteine
bewiesen, weil sie Zielmoleküle mit hoher Affinität und Spezifität binden können und über
ein vorteilhaftes pharmakokinetisches Profil verfügen.
Trotz den obengenannten positiven Eigenschaften weisen auch kommerziell sehr erfolgreiche
Protein-Arzneimittel Nachteile auf. In der Natur wurden Proteine nicht zur ex vivo
Herstellung entwickelt, weshalb viele menschliche Proteine nur schwer rekombinant zu
produzieren sind. Schlechte Expressionsausbeuten, niedrige Löslichkeit und das Potenzial, im
Patienten eine Immunantwort auszulösen sind Probleme, welche häufig im Zusammenhang
mit Biotherapeutika anzutreffen sind. Zudem ist bekannt, dass niedrige Stabilität sowie die
Bildung von molekularen Aggregaten das immunogene Potenzial von therapeutischen
Proteinen erhöhen. Deshalb besteht ein wachsendes Interesse an der Entwicklung von
stabilen, gut löslichen Proteinen menschlichen Ursprungs, die weniger immunogen und
einfacher in Mikroben zu exprimieren sind.
Ein aufstrebendes Gebiet in der pharmazeutischen Biotechnologie stellt daher die
Entwicklung von neuen bindenden Molekülen dar, basierend auf sogenannten kleinen,
globulären Domänen, die als Protein-Gerüste dienen. In diesem Ansatz, werden bestimmte
Aminosäuren auf der Oberfläche der kleinen Domäne randomisiert, um eine Bibliothek an
Varianten zu generieren, die dann auf gewünschte Bindungsspezifitäten geprüft werden
können. Somit wird das erfolgreiche Konzept der Antikörper (eine variable
Bindungsoberfläche und konstante Regionen) auf diese alternativen Gerüste übertragen, die
im Gegensatz zu den Antikörpern vorteilhaftere biophysikalische Eigenschaften aufweisen.
Einerseits können diese alternativen, bindenden Proteine neue Einsichten gewähren in die
molekularen Erkennungsprozessen von Protein-Protein Interaktionen, andererseits könnten
7
sie auch kommerziell genutzt werden, zum Beispiel als Therapeutika, Diagnostika oder als
Bindungsreagenzien in Laborexperimenten.
In dieser Arbeit werden das Design, die Konstruktion und die Charakterisierung einer
menschlichen Fyn SH3 Phagenbibliothek vorgestellt. Nach Test-Selektionen und
erfolgreicher Isolierung von Fyn SH3 Derivaten, die an Maus-Serum Albumin binden,
präsentieren wir eine neue, optimierte Bibliothek, die aus 1.2 x 109 Fyn SH3 Varianten
besteht. Aus dieser optimierten Bibliothek wurden Fyn SH3 Varianten isoliert und in vitro
charakterisiert, die an die Extra-Domäne B von Fibronektin binden, welche ein Angiogcnesc-
Marker ist. Ein bindender Klon, D3, wurde weiter untersucht und zeigte eine
aussergewöhnlich gute Färbung von vaskulären Strukturen in Tumor-Gewebsschnitten. Des
weiteren zeigten quantitative Biodistributionen in einem murinen Tumormodell, dass D3 sich
spezifisch im Tumorgwebe anreicherte. Im Gegensatz zu menschlichen
Antikörperfragmenten, waren nach vier intravenösen Injektionen weder die Fyn SH3
Domäne noch die D3 Variante immunogen in immunokompetenten Mäusen.
Das EDB-bindendc Protein D3 öffnet neue biomedizinische Möglichkeiten für die in vivo
Darstellung von soliden Tumoren und für die zielgerichtete Anreicherung von toxischen
Wirkstoffen im Tumorgewebe. Zusätzlich bietet die präsentierte Fyn SH3 Phagenbibliothek
womöglich eine reiche Quelle an nützlichen Molekülen für viele biochemische und
biomedizinischc Anwendungen, und stellt somit eine Alternative dar zu den herkömmlichen
immunochemischen Technologien, die auf Antikörper als Bindungspartner basieren.
8
1 Introduction
1.1 Recombinant proteins as biopharmaceuticals
The development of genetic engineering and progresses in molecular biology have allowed
the cloning of genes in living organisms such as bacteria and yeast for their efficient
expression. Large-scale production of many proteins has become possible, thus opening
several industrial applications, including the production of enzymes for food processing or
the synthesis of chemicals and medicinal products. Therapeutic proteins produced via such
methods have been coined biopharmaceuticals (1).
The first biopharmaceutical to gain marketing approval was that of „humulin" (recombinant
human insulin), initially approved in the United States in 1982 (2). In the intervening years,
the biopharmaceutical industry has grown rapidly: by 2006, 165 biopharmaceutical products
have gained approval, with a market size estimated at some $33 billion in 2004 (3) and
projected to reach $70 billion by the end of the decade (4,5). The majority of initially
approved biopharmaceuticals may be classified as „simple replacement proteins", i.e.
proteins displaying an identical amino acid sequence to a native human protein and
administered in order to replace or augment levels of that protein (e.g., insulin, growth
hormone or blood factors). More modern biopharmaceuticals have been engineered to tailor
their therapeutic properties. Such second-generation products display either an altered amino
acid sequence (achieved via protein engineering or site directed mutagenesis), an altered
glycocomponent (in the case of some glycosylated biopharmaceuticals) or a covalently
attached chemical moiety such as polyethylene glycol or toxins.
1.1.1 Monoclonal antibody based drugs
30 years after the advent of hybridoma technology (6), antibodies represent nowadays the
best established class of binding molecules that can be rapidly isolated with high affinity and
specificity to virtually any target. It is therefore not surprising that antibodies are routinely
9
used in biomedical research for analytical and separation purposes, arc essential ingredients
in many diagnostic procedures, and represent the fastest growing sector of pharmaceutical
biotechnology. By 2008, engineered antibodies and fragments thereof are predicted to
account for >30% of all revenues in the biotechnology market (7). Not only the high-affinity
and specificity, but also the variety of mechanism of actions (e.g., neutralization, interfering
with cell-signalling, antibody-dependent cellular toxicity and/or complement activation) and
favorable pharmacokinetic profiles contribute to the great success of this class of drug. Table
1.1 gives an overview of antibodies approved by FDA for therapy and diagnosis.
Product name Generic name Target Type Indication Year
Orthoclone OKT3 Muromonab CD3 murine IgG2a Transplant rejection 1986
OncoScint Satumomab pendetide TAG-72 murine IgG lllln Colorectal and ovarian cancer 1991
ReoPro Abciximab GP Ilb/IIIa chimeric Fab Cardiovascular 1994
CEA-Scan Arcitumomab CEA murine Fab 99mTc Coloretal cancer 1996
Prostascint Capromab pendetide PSMA murine IgGl lllln Prostata carcinoma 1996
Verluma Nofetumomab CD20 murine Fab 99mTc Small cell lung cancer 1996
Zenapax Daclizumab CD2S humanized IgGl Transplant rejection 1997
Rituxan Rituximab CD20 chimeric IgGl B-cell lymphoma (Non-Hodgkln) 1997
Simulect Basiliximab CD25 chimeric IgGl Transplant rejection 1998
Synagis Palivizumab RSV humanized IgGl RSV bronchiolitis 1998
Remicade Infliximab TNF chimeric IgGl Rheumatoid arthritis; Crohns disease 1998
Herceptin Trastuzumab HER2/neu humanized IgGl Breast cancer 1998
Mylotarg Gemtuzumab CD33 humanized IgG4 toxin conjug.
(Calicheamycin)
Acute myeloid leukaemia 2000
MabCampath Alemtuzumab CD52 humanized IgGl Chronic lymphatic leukaemia 2001
Zevalin Ibrltumomab CD20 murine IgGl 90Y B-cell lymphoma 2002
Humira Adalimumab TNF human IgGl Rheumatoid arthritis 2002
Xolair Omalizumab IgE humanized IgGl Asthma 2003
Bexxar Tositumomab CD20 murine IgGl 1311 B-cell lymphoma 2003
Raptiva Efalizumab CDlla human IgGl Psoriasis 2003
Avastin Bevacizumab VEGF humanized IgGl Colorectal cancer 2004
Erbitux Cetuximab EGFR chimeric IgGl Colorectal cancer 2004
Tysabri Natalizumab alpha4 integrin humanized IgG4 Multiple sclerosis 2004
Lucentis Ranibizumab VEGF humanized Fab fragment age related macular degeneration 2006
Vectiblx Panitumumab EGFR human IgG2 Metastatic colorectal cancer 2006
cancer
diagnostic
Table 1.1 List of antibodies approved by FDA for therapy or diagnosis.
1.2 Small globular proteins as antibody substitutes
Even antibody blockbusters suffer from certain drawbacks, such as the requirement for
mammalian cell production systems and dependence on disulfide bonds for stability. In
addition, some antibody fragments tend to aggregate and display limited solubility. Finally,
the success, and consequently the extensive use, of antibodies has led to a complicated patent
situation of antibody technologies and applications. Therefore, several research groups and
10
small and medium-sized companies have recently focused on the development of small
globular proteins as scaffolds for the generation of a novel class of versatile binding proteins.
As therapeutics, globular proteins with engineered binding properties might be particularly
interesting, (i) if the neutralization of a target protein is the desired pharmacological effect (in
contrast to a full length antibody, where the Fc portion may stimulate immune processes), (ii)
as fusion proteins, for the targeted delivery of bioactivc molecules to sites of disease, (iii) as
receptor-binding drugs, thus interfering with the cell-signalling and (iv) as enzyme inhibitors.
Common to all approaches of finding a suitable scaffold are the following steps (8-11):
- choosing a small protein (domain) that is well expressed in bacteria or yeast and has
good biophysical properties (stability, solubility)
- creating a library of protein mutants by introducing diversity at a contiguous patch of
the surface of the protein (e.g., loops)
- using a genotype-phenotype display system (such as phage or ribosome display, see
chapter 1.4) to select a range of binders to a therapeutic target of interest
- screening the binders obtained by the selection procedure for the desired biological
activity, e.g., by ELISA
Mutations introduced in the protein scaffold to produce diversity may compromise the three-
dimensional structure, stability and solubility of the protein scaffold, thus making the
isolation of protein binders based on other folding frameworks than the immunoglobulin fold
difficult. Nevertheless, more than 40 scaffolds have been described for the generation of new
protein binders (reviewed in (12)). Being mainly a topic of academic interest at first,
nowadays the development and use of non-antibody classes of proteins is being pursued by
small- and medium-sized biotechnology companies (13). The common basis for the
successful development of well designed protein drugs lies in the availability of both a well
behaving protein scaffold and an efficient selection or screening technology in order to find
suitable candidate proteins in the large repertoire created by mutagenesis. The next sections
11
will introduce selected classes of protein scaffolds that have been successfully used and
commercialized by small- and medium sized companies.
1.2.1 Single domain antibodies
Although not belonging to the strict definition of „antibody alternative protein frameworks",
single domain antibody fragments arc included as well in this chapter since they share a lot of
common properties (e.g., small domains with high expression yields in bacteria). Moreover,
they represent a valuable source for the generation of new binding proteins.
In a seminal early publication (14), mouse single variable domains (Vh) were shown to be
functional, and it was proposed that they could potentially target cryptic epitopes normally
hidden for whole antibodies and even for smaller fragments thereof. However, they were
described to be very sticky, poorly soluble and often prone to aggregation, because of their
exposed hydrophobic surface normally „capped" by the VL domain. Interest was recently
revived when it was discovered that at least two types of organisms, the camelids and
cartilaginous fish have evolved high-affinity V-like domains, mounted on an Fc-equivalent
constant domain framework as an integral and crucial component of their immune system
(15,16) (Fig. 1.1).
? £^ <C»
VhH domain V-NARdomam V,tJoman( ISkDa) ( 15kDal I ISfcDa)
0fl 5 ( sSkOa)
;\ 36 & \\$\#
(r) 0 J'JCamel Ig IgNAR igG
Figure 1.1 Schematic representation of different antibody formats. Intact „classic" IgG molecule,
the corresponding intact camel lg and shark IgNAR with their single domain fragments arc shown.
12
Unlike mouse Vh domains (14), camelid VhH (termed nanobodies) and shark V-NAR
domains are in general soluble and can be produced as stable in vitro reagents. However, for
in vivo administration, humanization (or deimmunization) may be crucial to reduce
immunogenicity. Human single domain antibodies (dAbs) would be even more preferable,
and very recently the problems of poor stability and solubility have been solved for some
human V domains by the identification and design of mutations that minimize the
hydrophobic interface (17,18). Moreover, in a recent publication of Jespers and co-workers
(19), aggregation-resistant domain antibodies could be directly selected on phage by heat
denaturation. Starting from a DP47d domain antibody (a typical human Vh dAb), which
unfolds irreversibly and forms aggregates if heated above 55°C, a repertoire containing
approximately 1 billion different mutants was cloned by diversification of the CDR loops.
The library was multivalently displayed on phage and after three rounds of heat denaturation
followed by selection on protein A (a ligand common to folded dAbs), 179 out of 200
colonies secreted dAb phage that retained more than 80% of protein A-binding activity after
heating (in contrast to the starting protein DP47d, which loses binding by a factor of 560 after
heating). Twenty clones were sequenced and revealed many unique dAb sequences with a
large variability in the CDR length and sequences, which shows that mutations located only
in the CDR loops of the dAbs are sufficient to confer resistance to aggregation. Interestingly,
the TM of the mutants was not higher than the one of the parental clone DP47d, which shows
that the selected dAbs are not heat stable, but can fold reversibly.
In summary, progress has been made towards the design of human dAbs which are soluble
and can be expressed in E.coli in acceptable yields. However, the disulfide bond in the
immunoglobulin fold of the dAb can lead (when fused to other proteins containing an
additional disulfide bonds or free cysteine residues) to the formation of misfoldcd and
inactive proteins. To circumvent this problem, in 2004 Wittrup and colleagues have
engineered a VL domain devoid of disulfide bonds that is stable and eficiently expressed in
the cytoplasm of yeast (20). Intracellular antibodies, also called intrabodies, are attracting
interest as tools to manipulate and study biological systems intracellularly.
The company having commercialized the use of single domain antibodies, Domantis Ltd.,
was recently acquired by GlaxoSmithKline for £230 million. Together with the acquisition of
13
Avidia by Amgen (see chapter 1.2.3), this trade-sales certainly reflects the growing interest of
big pharmaceutical companies in single domain protein scaffolds.
1.2.2 Tenth type III domain of fibronectin
The tenth type III domain of human fibronectin (10Fn3) is a 94 aminoacid residue structural
protein with an immunoglobulin-like fold that is used as an antibody-mimic scaffold. High
thermal stability (TM = 90° C) and solubility (> 15mg/ml), high expression yields in E.coli
and the lack of cysteines in its structure are attractive properties that make the 10Fn3 a good
scaffold candidate (21). Three of the solvent exposed loops in 10Fn3, named BC, DE and FG
are structurally analogous to the VH complementary-determining regions (CDR) (Fig. 1.2).
VHH 10Fn3
Figure 1.2 Structural comparison of a llama Vhh domain and the wild-type human I0Fn3
domain (21). Despite the lack of significant sequence identity, both domains fold into similar beta
sheet sandwiches. The CDRs of the VHh domain and the residues randomized in the 10Fn3 domain are
shown in color.
The first report on the use of the 10Fn3 domain as a protein with artificial binding sites was
published in 1998 (22). In that study, a library of 108 distinct 10Fn3 mutants was made by
deleting three residues in the FG loop and by randomizing five residues in the BC loop and
four residues in the FG loop. The library, displayed on phage, was used in affinity selection
experiments with immobilised ubiquitin as model target protein. After five rounds of panning
14
clones were randomly picked for sequencing. A clone designated as Ubi4-Fn3 dominated the
population of selected protein mutants and was subjected to further analysis. Ubi-Fn3 was
demonstrated to bind in phage enzyme-linked immunosorbent assay (ELISA) experiments.
However, when expressed as a single domain protein in E.coli, Ubi4-Fn3 exhibited low
solubility at neutral pH, unspecific binding to the dextran matrices of the size-exclusion
chromatography column and the dextran coated biosensor chips used for the subsequent
characterization of Ubi-Fn3. More importantly, the affinity was relatively low, the IC50
determined by competition ELISA was 5uM
In order to select 10Fn3 domains binding to a target of interest with improved properties than
Ubi4-Fn3, an alternative library design in combination with a fully in vitro selection system
(mRNA display (23)) was used for the isolation of 10Fn3 variants binding to tumor necrosis
factor-a (TNF-a) (21). Clones from the ninth and tenth round of selection were cloned in
E.coli, sequenced and expressed. Affinity constants were determined by incubating the in
vitro translated 35S-mcthionine labeled proteins with biotinylated TNF-a at different
concentrations. The solution containing the 10Fn3 mutants bound to TNF-a were aspirated by
vacuum onto a membrane coated with strcptavidin, therefore capturing protein complexes on
the membrane. Binding was analysed by measuring the radioactivity on the membrane, the
KD for selected mutants were found in the range of 1-24 nM. Further affinity maturation
procedures revealed KD values around 100 pM, the best being 20 pM, and analytical gel
filtration showed that the apparent molecular weight of purified, soluble wild-type (wt) and
mutant domains was consistent with the variants being monomeric. The reported affinities,
especially for TNF-a, are very high. However, these values should be evaluated critically
because first, the capture step of the biotinylated TNF-a-Fn3 protein complexes was
performed in a solid phase format, and second, TNF-a is homotrimeric protein, so avidity
effects may contribute to the high apparent affinity observed.
Other Fn3 variants were shown to recognize avß3-integrin expressed on cell surface and
inhibiting avß- dependent cell adhesion (24) and to bind to vascular endothelial growth
factor receptor 2 (VEGF-R2) (25,26). For all 10Fn3 variants reported so far, their
thermodynamic stability was worse than that of the wt protein. In addition, some VEGFR-R2
binding mutants tend to dimerize and/or to aggregate (25).
15
As a conclusion, it seems difficult to obtain high-affinity and stable proteins with this
scaffold. In addition, the use of engineered l0Fn3 in therapeutic applications may turn out to
be problematic: an immune response of the patient to the injected protein might generate
cross-reactive antibodies binding to the wild-type 10Fn3, which is ubiquituos in the human
body, thus leading to an immune attack in the patient and/or formation of circulating
complexes, with a potential kidney toxicity. However, in the near future we will learn more
about this scaffold, since a VEGF-R2 antagonist, termed „Angiocept CT-322", has just
entered into phase I clinical trials, in patients with advanced solid tumors and non-hodgkin's
lymphoma (www.clinicaltrials.gov).
1.2.3 A-domains
A family of A-domains (27,28) has been described to be suitable as a scaffold for the
generation of new binding proteins (29). A-domains occur as strings of multiple domains in
several cell-surface receptors (Fig. 1.3a). Domains of this family bind over 100 different
known targets, including small molecules, proteins and viruses (30,31). Such a target is
typically contacted by multiple A-domains with each domain binding independently to a
unique epitope, thereby generating avidity. Each of the 217 human A-domains comprises
approximately 35 amino acids and domains are separated by linkers that average five amino
acids in length. Native A-domains fold quickly and efficiently to a uniform, stable structure
mediated by calcium binding and disulfide formation. A conserved scaffold motif of only 12
amino acids is required for this common structure (32). Stemmer and co-workers have
designed a phage display library of A-domains, consisting of a conserved consensus sequence
in the scaffold motif and variable amino acids in different positions (Fig. 1.3b) (29). Like
other directed evolution technologies, the process the authors developed included multiple
recursive cycles, each consisting of library generation and screening in functional assays.
However, rather than mutagenizing the protein between cycles (e.g., for affinity maturation
pruposcs), they added a new pool of domains adjacent to the domain(s) selected in previous
rounds (Fig. 1.3 c).
16
Figure 1.3 Overview of avimer technology (29). (a) Domain organization of several human receptor
containing A-domains. Multivalent binding of several natural ligands to known combinations of A-
domains is also depicted. a-2M, a-2 macroglobulin; LRP, low density lipoprotein-related protein;
RAP, receptor-associated protein; VLDLR, very low density lipoprotein receptor; VLDL, very low
density lipoprotein; ApoE, apolipoprotein E; ApoBlOO, apolipoprotcin B100; LDL, low density
lipoprotein; PAI-1, plasminogen activator inhibitor 1; uPA, urokinase-type plasminogen activator, (b)
Fixed and variable positions of the A-domain library, as well as the disulfide topology, are indicated.
Each circle represents an amino acid position. Calcium-coordinating scaffold residues are yellow,
structural scaffold residues are blue, cysteine residues are red and variable positions are green. A
ribbon diagram of a prototypical A-domain structure is included (PDB ID 1AJJ). (c) Flow chart
depicting the sequential panning and screening referred to as „domain walking".
The end result is a single protein chain containing multiple domains, each of which represents
a separate function. The A-domains are therefore called „avimers", from avidity multimers. A
hetcrotctramer consisting of three IL-6 binding A-domains and an IgG binding domain
(named C326) showed a remarkable affinity (picomolar range) to its target and exhibited sub-
picomolar IC5os in cell-proliferation assays (29). Moreover, C326 completely abrogated
acute-phase protein induction by human IL-6 in mice, in a dose dependent manner,
suggesting that C326 inhibited IL-6 functions in vivo. A placebo-controlled phase I study of
C326 was enrolled in the United States in September 2006, with patients suffering from
Crohn's disease (www.clinicaltrials.gov). The completion of this study is expected in
September 2007, giving new insights about safety and pharmacokinetics of C326.
17
The technology of using the modular platform of A-domains was commercialized by Avidia
Inc., that was acquired by Amgen for 290$ million in October 2006, proving the interest and
significant potential of antibody alternatives in the field of pharmaceutical biotechnology.
1.2.4 Z-domain of staphylococcal protein A
The immunoglobulin binding domain of staphylococcal protein A (SPA) is widely used as an
immunochemical ligand for the purification or detection of antibodies and serves as affinity
tag in fusion proteins (33). The IgG binding domain of SPA consists of five highly
homologous thrce-hclix bundle domains of approximately 58 amino acid residues each.
Because SPA is known as a non-cysteine containing, highly soluble, proteolytically and
thermally stable protein, an engineered version of one of the SPA domains, the so-called Z-
domain (34) was chosen as a scaffold for the development of novel affinity proteins
designated as „affibodies" (Fig. 1.4). Utilizing structural data available for the complex
between a native SPA domain and the Fc fragment of human IgGl, 13 positions distributed
across two helices, located at the surface of the domain and involved in the Fc interaction,
were chosen for random mutagenesis, and subsequently, for the creation of phage libraries
(35). In 1997, from two medium sized libraries comprising about 4x10 individual clones,
binding proteins against three model target proteins (Taq DNA polymerase, human insulin,
and a human apolipoprotein A-l variant) were isolated, with binding affinities in the
micromolar range (36). In the meantime, affibodies have been isolated against a variety of
protein targets (e.g., human factor VIII (37), human immunoglobulin A (38) and CD28 (39)).
In a more recent paper (40), an affibody ligand binding to human epidermal growth factor
receptor 2 (Her2) was isolated with a dissociation constant (Kd) of 50 nmol/L. Dimcrization
of this affibody molecule resulted in improved target binding affinity (KD = 3 nmol/L), and
radioiodination of the dimer allowed selective targeting and imaging of Her2-expressing
xenografts in vivo (41). Further affinity maturation strategies led to an affibody molecule
with a dissociation constant in the range of 20 pmol/L and to better tumor targeting and
imaging results in the same tumor mouse model (42). The most recent paper describes fully
synthetic affibody molecules, site-specifically and homogenously conjugated with a DOTA
18
(l,4,7,10-tetra-azacyclododecane-N,N',N",N'"-tetraacetic acid) chelator, produced in a
single chemical process by peptide synthesis (43).
Figure 1.4 Structure of an in vitro evolved affibody in complex with protein Z. (44) (a) Structure
of the complex, the ZSPA-1 affibody in blue and protein Z in green. The ordered sulfate ions with
partial occupancy and putative magnesium ion from the mother liquor are also shown; however, they
do not seem to influence the interaction surface, (b) Superposition of the two molecules. Notice the
shift of helix 1 in the affibody (blue) compared with protein Z (green), (c) Electron density for all 13
mutated residues in the affibody; for clarity, only the electron density around the side chains is
displayed.
Importantly, first clinical data of molecular imaging of breast cancer metastases using In
radiolabeled anti-Her2 affibody molecules were presented (Next-Generation Protein
Therapeutics Conference, Basel, November 2006), showing high-resolution images of
metastases within hours. Although the data obtained from patients have not been published in
peer reviewed journals yet, it seems that the affibody molecule works well for imaging
applications. However, future clinical data will show whether affibody molecules arc good
enough for therapeutic applications, especially concerns about the immunogenic potential of
this class of proteins should be seriously considered, owing to its bacterial origin.
19
1.2.5 Kunitz type domains
Proteases have important biological functions, they have been identified as key regulators of
cellular processes such as ovulation, fertilization, wound healing, angiogenesis, apoptosis,
peptide hormone release, coagulation, and complement activation (45). Because unregulated
proteolysis in any of these processes will have undesirable effects, effective control of
protease activity is critical. Indeed, unregulated activity has been implicated in the
pathogenesis of many diseases (e.g., cancer, inflammatory diseases, chronic obstructive
pulmonary disease (COPD), chronic pancreatitis, muscular dystrophy and Alzheimer's
disease (46-51)), making proteases important target proteins. It is notable, that there are only
few marketed drugs (such as HIV protease inhibitors, angiotensin-converting enzyme
inhibitors, the proteasome inhibitor bortezomib and the very recently approved renin-
inhibitor aliskiren), since a key challenge in the discovery of new inhibitors is the ability to
identify drugs which are both potent and specific. Structural similarities within the active site
of most proteolytic enzyme families often result in a simultaneous inhibition of several
family members, which can lead to an unacceptable toxicity profile. One solution to this
problem might represent the use of endogenous, engineered protease inhibitors that make
more contacts with the target protease and therefore allow for tighter control over specificity.
One type of endogenous inhibitors include the kunitz domain inhibitors, e.g., bovine
pancreatic trypsin inhibitor (BPTI) (52).
In 1992, Ladner and co-workers produced a small library of mutants of kunitz domain BPTI,
displayed on phage and screened for the ability to bind to elastase (53), a serine protease that
is believed to play a causative role in a variety of lung diseases, including COPD and cystic
fibrosis (54). Isolates that were able to bind to elastase were expressed, purified and tested for
inhibition of elastase. The P-loop (sec Fig 1.5), the region that was varied in the library, of
the best inhibitor (K; ~ 1 pM) was transferred to a human kunitz domain without subsequent
loss of potency (55). The resulting engineered molecule, depelestat (Dyax/Debiopharm SA),
inhibited human neutrophile elastase (HNE) activity in the sputum of cystic fibrosis patients,
was demonstrated to be safe in an aerolized format in monkeys and no adverse effects were
reported in phase I clinical trials (45,56,57). Pharmacodynamic reports of a phase Ha clinical
trial have been shown to inhibit completely sputum HNE in 52.6 % of the patients and to
20
decrease interleukin-8 levels, a biomarker of inflammation, in the sputum of treated patients
(58).
PI Region
Second Loop-
Figure 1.5 Schematic representation of a representative kunitz type domain (LAC-D1). (59)
Varied positions are depicted in black, the PI and second loop positions are enclosed.
In 1996, another human kunitz domain, the lipoprotein-associatcd coagulation inhibitor
(LACI-D1), was used as a scaffold for the successful isolation of a very specific and potent
inhibitor of human kallikrein (60) (Fig. 1.5). Kallikrein is believed to be an important
mediator of hereditary angioedema (HAE), a rare disorder with attacks of edema in the
hands, face, feet, abdomen, and/or throat. This condition is caused by a genetic deficiency of
CI esterase inhibitor (Cl-Inh). In Europe, HAE is treated with plasma derived Cl-Inh which
attenuates attacks and may prove life saving, but Cl-Inh is expensive and requires the use of
pooled blood product (45). After an iterative selection and screening approach chosen by the
authors (60), the plasma kallikrein inhibitor DX-88 (Dyax Corp / Genzyme Corp) was
isolated and proved to be a potent inhibitor (Kj = 40 pM) with high selectivity (Table 1 in
(61)). The use of DX-88 in a Cl-Inh deficient mouse model demonstrated that DX-88 is
effective in preventing changes in vascular permeability (62). DX-88 is currently being used
to treat acute attacks of HAE in clinical studies, positive phase II trial results have been
reported (61) and double-blind, placebo-controlled phase III clinical trials have been enrolled
to assess the efficacy and safety of DX-88 for the treatment of acute attacks of HAE.
With respect to clinical development, kunitz type domains represent the most advanced
21
scaffold in this field. The strategy of taking a human protease inhibitor as a scaffold and
improving it by protein engineering works well for certain therapeutic applications. At the
same time, kunitz type domains have the limitation of binding only to proteases, and thus
they do not represent a universal source for the generation of binding proteins to a variety of
targets.
1.2.6 Ankyrin repeat proteins
Ankyrin repeat (AR) proteins, first isolated in mammalian erythrocytes, are involved in the
targeting, mechanical stabilization and orientation of membrane proteins to specialized
compartments within the plasma membrane and endoplasmic reticulum. Natural ankyrin
repeat proteins consist of many 33-amino-acid modules, each comprising a ß-turn and two
anti-parallel a-helices (63). In most known complexes, the ß-turn and the first a-hclix
mediate the interactions with the target, and different numbers of adjacent repeats are
involved in binding. The reported target binding affinities of natural AR proteins are in the
low nanomolar range (64). Ankyrin repeat proteins were built and diversified to create a
library from which ankyrin variants were selected binding to maltose-binding-protein and
two eukaryotic kinases (65-67). In the approach chosen by the authors, a consensus ankyrin
repeat module consisting of six diversified potential interaction residues and 27 framework
residues was designed based on sequence alignments and structural analyses (Fig. 1.6).
Varying numbers of this repeat module were cloned between capping repeats, which are
special terminal repeats of ankyrin domains shielding the hydrophobic core. Two libraries
were created with two and three, respectively, randomized ankyrin repeats in between an N-
terminal and a C-terminal cap.
22
Figure 1.6 Schematic representation of the library generation of designed ankyrin repeat
proteins (65). (upper part) Combinatorial libraries of ankyrin repeat proteins were designed by
assembling an N-terminal capping ankyrin (green), varying numbers of the designed ankyrin repeat
module (blue) and a C-terminal capping ankyrin (cyan); side chains of the randomized residues are
shown in red. (lower part) Ribbon representation of the selected MBP binding ankyrin repeat protein
(65) (colors as above). This binder was isolated from a library of N-tcrminal capping ankyrin repeat,
three designed ankyrin repeat modules and a C-terminal capping ankyrin repeat.
Using a library with more than 1010 individual members in combination with the ribosome
display selection methodology (see chapter 1.4.5), binding variants (termed DARPins) were
selected by performing four or five rounds of selection. Dissociation constants were
determined by surface plasmon resonance and found to be in the range of 2-20 nM against
these model target proteins. More recently, more ankyrin repeat variants have been isolated
against target proteins such as intracellular kinases, phosphotransferases and human
epidermal growth factor receptor 2 (Hcr2), the latter being an important target for cancer
therapy and diagnosis (66-68). However, in vivo data are not available yet, and, because the
DARPins were newly designed and arc not found in nature in this format, their immunogenic
potential should be investigated in detail.
23
1.2.7 Lipocalins
The lipocalins represent a family of functionally diverse, small proteins that comprise 160-
180 amino acid residues and have weak sequence homology but high similarity at the tertiary
structural level (69). Members of this family have important biological functions in a variety
of organisms, from bacteria to humans. The majority of lipocalins are responsible for the
storage and transport of compounds that have low solubility or are chemically sensitive, such
as vitamins, steroids and metabolic products (70). An example of human lipocalin is the
retinol binding protein (RBP), of which the 3D-structure has been elucidated by X-ray
cristallography (71). RBP transports the poorly soluble and oxidation-prone vitamin A from
the liver, where it is stored as a fatty acid ester, to several target tissues. Despite their
extremely poor sequence homology, the lipocalins share a structurally conserved ß-barrel as
their central folding motif, which consists of eight antiparallel ß-strands that are arranged in a
cylindrical manner (Fig 1.7). The binding specificity for low-molecular weight compounds is
well-characterized for many lipocalins, some having high ligand specificity whereas others
form complexes with a considerable range of lipophilic molecules (72). Typically, the ligand
affinities of lipocalins arc moderate with dissociation constants of approximately 1 uM,
which is consistent with their presumed function as a physiological buffer for the bound
substance; however, there are notable exceptions, e.g., the tick histamine-binding protein
(HBP)-2 with a KD of 1,7 nM (73). The insect bilin binding protein (BBP) from Pieris
brassicae served as a model lipocalin in initial studies to create artificial binding sites for
several ligands. Sixteen amino acid positions located within the four loops at the open end of
the ß-barrel and adjoining regions of the ß-strands were subjected to targeted random
mutagenesis (74). The resulting molecular random library with about 4 x 108 members was
subjected to selection towards several low-molecular weight compounds using phage display.
BBP variants, so-called „anticalins", that specifically recognize fluorescein, digoxigenin,
phthalic acid esters and doxorubicin were obtained, with affinities in the nM range (74-76).
In order to extend the concept of anticalins, several human lipocalins were subjected to
random mutagenesis generating libraries that enable recognition of macromolecular protein
targets. Because proteins have larger molecular dimensions than small compounds, they
cannot penetrate into the ligand-binding pocket of lipocalins. Consequently, side chains at
more exposed positions, close to the tips of the four loops at the open end of the ß-barrel,
24
were subjected to random mutagenesis. Following this strategy, a panel of anticalins based on
human lipocalins with specificities for several therapeutic targets such as cytotoxic T-
lymphocyte-associated antigen (CTLA-4), VEGF and other undisclosed targets
(www.pieris.biz) were isolated. However, detailed data about this work has not yet been
published and it has to be seen how these engineered proteins will behave in experiments
with animal models and, subsequently, in clinical trials.
Figure 1.7 General structure of human retinol-binding protein, a prototypic Hpocalin. Ribbon
diagram of the crystal structure of RBP (adapted from (72)) with the bound ligand retinol (magenta,
ball and stick representation). The eight antiparallel strands of the conserved ß-barrel structure are
shown in blue with labels A to H, and the four loops, which are highly variable among the lipocalin
family, arc colored red and numbered. The typical a-helix that is attached to the central ß-barrel in all
lipocalins, the loops at the closed end the N- and C- terminal peptide segments are shown in grey. The
three disulfide bonds of RBP are depicted in yellow.
1.2.8 Other domains
The preceeding sections introduced important domain scaffolds that have been
commercialized and for which published data are available. In a recent review of Binz et al.
(12) more than 40 scaffolds have been described for the generation of new binding proteins.
The majority of them have been used for research purposes only (e.g., probing the specificity
determinants of WW-domains to their ligands (77), elucidating target recognition rules of
SH3 domains (78,79) or identifying signal transduction pathways with the staphylococcal
nuclease as scaffold (80)). Other domains have been commercialized by companies for in
25
vitro applications (e.g., PDZ domains for the generation of high-affinity detection reagents
(Biotech Studio (81)) or for intracellular signaling interference applications (thioredoxin,
Aptanomics (82)). A few other scaffolds have also been commercialized by small- and
medium sized companies, however, no refereed publications are avaliable (human y-crystallin
and human ubiquitin by Seil Proteins, C-typc lectin domain by Borean Pharma, monoclonal
soluble T cell receptors by Avidex and a not disclosed scaffold by Amunix; informations
obtained from the corresponding websites.)
1.2.9 Concluding remarks
After reviewing the most important scaffolds described in the literature, it can be summarized
that an ideal scaffold should comprise the following features:
- fully human protein scaffold (reduced immunogenic potential)
- conserved amino acid sequence in different species from mouse to man
(immunogenic potential can be tested in preclinical experiments)
- high-expression yields in bacteria (cheap production)
- scaffold devoid of cysteine residues (site-directed introduction of cysteine residues for
chemical modification at a later stage possible; less folding problems if scaffold is
expressed as fusion protein with cysteine residues, e.g., cytokines)
- thermally stable and soluble (no formation of aggregates -> less immunogenic)
- wild-type domain involved in protein-protein interaction (starting point for a first
library design)
structure solved (further rational randomization possible)
26
- randomizations introduced do not disrupt overall folding and expression
Although some rational justification may be given for the choice of almost any scaffold
described so far, in our opinion the protein scaffolds used for the development of therapeutic
proteins introduced in chapter 1.2 are not perfectly suited for the generation of binding
proteins. None of the protein scaffolds discussed before meets all the requirements for an
optimal scaffold candidate, maybe with the exception of ubiquitin. All other scaffolds have at
least one deviation from the set of desired properties. For example, camelid single domain
antibody fragments (nanobodies), protein A derived binders (affibodies), A-domains and
ankyrin repeat proteins arc from non-human origin or are synthetically designed scaffolds; A-
domains, T cell receptor domains, human C-typc lectin domains, human single domain
antibody fragments, kunitz type domains, lipocalins and y-cristallins contain cysteine
residues; 10Fn3 variants may lead to devastating effects in human beings, if such a therapeutic
variant elicits an immune response and if the antibodies cross-react with the wild-type
domain, which is ubiquitously expressed in the body.
It is difficult to predict how these intrinsic disadvantages of the scaffolds translate into
complications and increased costs in clinical development, but starting with a scaffold that
meets all requirements described may be more advantageous.
27
1.3 The SH3 domain of Fyn kinase as a scaffold for the generation of new
binding proteins
1.3.1 Src homology-3 (SH3) domains
SH3 domains arc globular protein modules typically comprising 50-60 amino acid residues
found in many different proteins, particularly those involved in cellular signal transduction
(83) and in those involved in the regulation of intracellular dynamic processes occurring at
the plasma membrane, such as the organization of the cytoskeleton and the internalization of
membrane receptors (84). Two hundred seventy human SH3-containing proteins have been
identified, with a total of- 300 different human SH3 domains (85). Despite the variability in
their primary structures, crystal and solutions structures from several SH3 domains revealed a
completely conserved fold: a compact ß-barrcl, formed by two perpendicular, two or three
stranded ß-shects, and a single turn of 3io helix (Fig. 1.8) (86-90).
Figure 1.8 Crystal structure of the Src-SH3 domain (87). The structure shown is PDB codclSHG.
28
SH3 domains mediate inter- and intramolecular interactions by binding to ligands that
contain a region with a secondary structure known as the polyproline type II (PPII) helix.
These ligands can bind to SH3 domains in two opposite orientations and show typically a
VXXP core-binding motif, flanked by a basic residue (RXTPXTP or PXXPXK) (91-93).
However, examples of unconventional SH3 binding sites are also known (94,95). The regions
involved in the interaction with the proline-rich core of target sequences is a surface patch
rich in aromatic residues (forming the ,^YP grooves") and by various polar residues located in
the RT- and n-src-loops (Fig 1.9) (96,97). SH3 domains bind their targets with low-affinities
(KD = 1 to 200uM), a property that is probably necessary for finely regulated, transient
signaling interactions (98). However, most of the affinity measurements reported so far used
the SH3 domain and a minimal binding motif (such as ^PZXP^lR), so the „real" affinity
between a SH3 domain and its interacting protein carrying a PPII helix is considered to be
higher (85).
Figure 1.9 Structure of the Crk SH3-N domain in complex with a proline rich peptide (97). The
structure shown is PDB accession code 1CKB. The SH3 domain is depicted in ribbons with
secondary-structural elements shown in different colors and labeled. The bound peptide,
PPPALPPKKR, is shown in blue with side chains. Interface residues on the SH3 domain arc shown in
pink, for aromatic residues, and in light blue, for non-aromatic residues. The locations of the XP
grooves and specificity pocket on the SH3 domain arc identified by broken arrows.
29
1.3.2 SH3 domains as scaffolds for the generation of engineered binding proteins
Since it has become clear that also molecular contacts outside the PPII helix interface can
provide specificity and strength to SH3 binding (99), e.g., by contacting loop regions of SH3
domains, efforts have been made by using engineered SH3 variants to elucidate further the
specificity and affinity determinants of the interactions between SH3 domains and their
ligands.
Saksela and colleagues constructed a phage display library of ~108 Hck SH3 variants, where
six amino acid residues in the RT loop were randomized (79). Wild-type Hck is known to
bind to the HIV-1 Nef protein and to a Nef mutant (F90R) with an affinity of 250 nM or
2 uM, respectively (99). Performing phage display selections against HIV-1 Nef, the authors
could isolate individual Hck SH3 variants which bound to HIV-1 Nef with up to 40-fold
higher affinities than the parental Hck SH3 domain, indicating that the RT loop contributes to
the binding strength in this particular interaction. Interestingly, some of the selected clones
bound also well to the F90R Nef mutant (NefR90), while others were dependent on the
presence of the Nef Phe90 residue even more critically than wild-type Hck SH3, and showed
over a 100-fold decreased affinity towards NefR90. Conversely, Hck SH3 derived binders to
NefR90 could be isolated using NcfR90 as target in phage display selections, whose affinity
depended on structural determinants involving the side chain of Arg90 residue. However, no
binding proteins could be isolated to other epitopes than the polyproline motif of Nef.
In conclusion, the group of Saksela succeeded in isolating affinity matured Hck SH3 binding
proteins to the known Hck SH3 ligand HIV-1 Nef. The authors concluded that artificial SH3
domains in general, like the ones isolated against HIV-1 Nef, could be potent competitors for
natural SH3 interactions, an important feature in basic research of intracellular physiological
pathways.
Cesareni and co-workers designed a repertoire of- 107 SH3 domains by grafting 12 residues
that arc represented in the binding surfaces of natural SH3 domains onto the scaffold of
human Abi SH3 domain (78). This phage-displayed library was screened by affinity selection
for SH3 domains that bind to the synthetic peptides, APTYPPPLPP and LSSRPLPTLPSP,
which are peptide ligands for the human Abl or Src SH3 domains, respectively. By analyzing
30
the isolates, the authors identified three amino acid positions of the Abi SH3 scaffold that are
crucial for recognition specificities, favoring either the „Abl-peptide" 0r the „Src-peptide":
one amino acid in the RT loop, one in the n-Src-loop and a tryptophan residue in the ß4
strand. By further site-directed mutagenesis experiments with these 3 residues in the wild-
type Abi SH3 scaffold, the authors evolved evolution pathways of protein interactions (e.g.,
an Abi SH3 variant carrying the residues TNW in the 3 positions and specific for the Abl
peptide can be mutated to TCW without substantially changing its recognition properties
whereas a second substitution, TCL, causes a sudden shift in ligand preference to the Src
peptide). The authors created then a position-specific scoring matrix, based on the amino acid
frequencies observed in the domains selected from the artificial repertoire, in order to rank
natural SH3 domains according to their probability of binding to the 2 peptides. Twenty-nine
SH3 domains in the yeast proteome were predicted to bind to either of the two poly-proline
peptides; the predictions were challenged by ELISA binding assays, which revealed false
negative and false positive results, as the underlying scaffold has a stronger influence on
binding specificity than assumed by the authors. The authors concluded nonetheless that their
library can be used to select Abl SH3 variants capable of binding to intracellular targets with
a proline-rich motif, in order to interfere with signaling cascades.
In summary, SH3-derived proteins have been used so far only for the generation of binders
against known SH3 ligands, such as HIV-1 Nef protein or synthetic proline-rich peptides
(78,79).
1.3.3 SH3 domain ofFyn kinase
Fyn kinase was identified in 1986 as a new member of the Src family of tyrosine kinases
(100,101). Originally denoted Syn or Slk, this related novel protein has been renamed Fyn
and encodes a 59 kDa kinase. As a result of alternative splicing, the Fyn protein is expressed
as two isoforms, differing in approximately 50 amino acids in a region between their SH2
and kinase domain. One form is found in thymocytes, splenocytes and some hcmatolymphoid
cell lines (FynT), while the second form accumulates principally in the brain (102). Its
biological functions arc diverse and related to Fyn's ability to associate and to phosphorylate
31
a variety of intracellular signaling molecules. One of the best known functions of Fyn kinase
is the phosphorylation of SLAM (signaling lymphocyte activation molecule) in T cells,
inducing a signaling complex modulating interferon-y expression (103) (Fig. 1.10). The
interaction between Fyn and SLAM occurs via the SH3 domain of Fyn and the adaptor
protein SAP (SLAM associated protein), forming a ternary complex. Interestingly, Fyn SH3
associates with SAP through a surface-surface interaction that does not involve the canonical
PX¥P recognition motif of SH3 domains (94). Interacting regions of the Fyn SH3 domain
include the RT-loop, n-src loop and single residues of the ß3- and ß4 strands (94).
Like other SH3 domains, Fyn SH3 is composed of-60 amino acids [amino acids 83-145 of
the sequence reported by Rcfs. (100,101)] and shows the typical SH3 topology of two
perpendicular ß-sheets and a single turn of 3i0 helix.
With its biophysical properties and its primary structure, Fyn SH3 perfectly matches our
criteria for a scaffold to be used as alternative to antibodies: (i) it is expressed in bacteria at
high level in soluble, monomeric form (104), (ii) it is stable (TM: 70,5° C) (105), (iii) it docs
not contain any cysteine residues, (iv) it is from human origin and (v) its amino acid sequence
is conserved in mouse, rat, monkey (gibbon) and man. Moreover, its binding mode to SAP
has been resolved and indicates that this particular domain does not necessarily need a ?XXP
core binding motif, which is an important prerequisite to be well suited for the isolation of
binders against a variety of different target epitopes. Moreover, since the groups of Saksela
and Cesareni showed that phage display technology is suitable for the isolation of artificial
Hck - and Abl - SH3 domains, the probability that this established methodology might also
work for the Fyn SH3 domain was very high.
32
Inactive Fyn Activa Fyn with bound SAP
Fyn SH3 SAP SH2
Figure 1.10 The role of Fyn kinase in T cells (103). (a) Signal transduction through the SLAM
receptor is regulated by SAP. Fyn is kept in an auto-inhibited state in the absence of SAP.
Recruitment of Fyn to the SLAM receptor by SAP leads to the activation of the kinase and subsequent
phosphorylation of SLAM on multiple tyrosine residues. Phosphorylated tyrosine residues serve as
docking sites for the SH2 domain of SHIP (SH2-containing inositol phosphatase). Binding of SHIP to
SLAM initiates a signaling cascade that ultimately leads to modulation of IFN-y production in T cells,
(b) Structure of a ternary complex of Fyn SH3-SAP-SLAM-peptide. The Fyn SH3 domain, in green,
binds to SAP, in cyan, via surface-surface association. SAP engages the SLAM peptide, in magenta,
on the opposite surface using a groove conserved for SH2 domains. The structure shown is PDB code
1M27.
33
1.4 Isolation of binding domains
1.4.1 General aspects
The identification and isolation of proteins exhibiting desired binding properties from large
repertoires of mutant domains is a daunting task. Due to the large number of different clones,
individual members of the repertoire cannot be assayed one by one due to time limitations
and practical reasons. Therefore, methodologies have been developed which allow the
simultaneous evaluation of many different protein mutants in a short period of time.
Screening methodologies represent procedures where all members of the repertoire are
individually assessed in a biological assay in a highly parallel fashion, thus opening the
opportunity to screen large numbers of protein candidates for desired activities. The
researcher then chooses those proteins, which were positive in the screening assay.
In addition to the screening techniques, selection technologies are often used, which allow to
mimic the natural process of evolution in a test tube in the laboratory. Common to all
selection technologies is the linkage of the genetic information (genotype) with the encoded
polypeptide (phenotype). This allows the amplification of the genetic information of the
isolated protein mutants, leading to the survival of genotypes which code for favorable
polypeptides. By repeating the two steps of generating genetic diversity followed by selection
for a desired protein activity, proteins can be evolved in vitro.
1.4.2 Iterative colony filter screening
In colony filter screening, clones (typically <106) are simultaneously assayed for their ability
to generate the binding specificities of interest. A filter screening assay was described (106)
for the identification of the few clones secreting an antibody of given antigen specificity, out
of several thousand hybridoma clones. A modified version of this methodology was used for
the isolation of antibodies from a large library with millions of different clones (107) (Fig.
34
1.11). The authors used the naive ETH-2 antibody library (108), in the form of a pool of
antibody secreting bacteria, to isolate monoclonal antibodies specific for the extra domain-B
(EDB) of fibronectin (see chapter 1.6).
Bacterial culture
representing the
ETH-2 Library
Membrane A
INext Screening Cycle
Inductionuction
Membrane A fCT^ J» ^"^Filter sandwich < ,jTVy£jrH
Amplification
Ç.-3
Membrane B LC^ 4 tf '^
Colony
Picking
>
Detection
Membrane B C
Realignment of membrane A and B
Figure 1.11 Schematic representation of the iterative colony filter screening method (107).
Bacteria expressing potentially each a different antibody fragment were spread on a Durapore filter
membrane (A). On the filter, placed on a solid medium that represses expression of the antibody
fragments, colonies were visible after 8 h incubation at 37°C (here three different colonies are
depicted schematically). A second filter membrane (B) was coated with the antigen of interest
(represented by pyramids), laid on a solid medium capable of inducing the expression of scFvs, and in
contact with membrane A. Antibody fragments that diffused from membrane A and that bound to the
antigen (in red) were captured on membrane B, and could be detected by an enzymatic colorimetric
reaction. The corresponding colonies could be identified on membrane A, regrown and the procedure
could be repeated until single positive colonies producing monoclonal antibodies were identified.
One hundred million bacteria from the ETH-2 library, potentially expressing a different scFv
antibody fragment, were grown on a porous master filter until a lawn of small confluent
colonies was visible. The master filter was then laid on a capture filter embedded with the
antigen (an EDB containing fibronectin fragment), and bacteria on the master filter were
induced to express the recombinant antibodies by placing the filter membranes on solid
medium containing IPTG. Recombinant antibodies can diffuse from the master to the capture
35
membrane, where the ones specifically binding the antigen are trapped onto the capture filter.
A Flag-tag at the C-tcrminal extremity of the scFv antibody fragments allowed the detection
of their binding to the antigen coated-capture membrane by an enzymatic colorimetric
reaction. Due to the huge number of bacteria on the master filter, the first round of screening
did not allow the direct identification of single positive bacterial colonies, but only of positive
areas of confluent colonies that, by superposition of the two filters, could be rescued from the
master membrane and grown to perform a further round of screening. Three rounds of
screening were performed, plating 106 bacteria on a master filter for the second round and
1000 for the third round. At this stage, colonies on the master filter were no longer confluent
and approximately 30% of the plated colonies produced scFvs that bound to filter B,
therefore representing potential anti-EDB monoclonal antibody fragments (107). The positive
clones at the third round of filter screening were regrown and the binding specificity of the
corresponding antibodies was analysed by ELISA. 16% produced antibodies that bound
specifically to the ED-B domain of fibroncctin and the best clone, named scFv ME4C, was
further characterized. The affinity of scFv(ME4C) towards the ED-B domain of fibronectin
was measured by real-time interaction analysis using surface plasmon resonance detection,
yielding a KD = lxlO"7 M (107).
Two general methodologies based on filter-sandwich assays were described by Heinis et al.
(109) for isolating enzymatic activities from a large repertoire of protein variants expressed in
the cytoplasm of E. coli cells. The enzymes were released by the freezing and thawing of
bacterial colonies grown on a porous master filter and diffused to a second "reaction" filter
that closely contacted the master filter. Reaction substrates were immobilized either on the
filter or on the enzyme itself (which was, in the latter case, captured on the reaction filter).
The resulting products were detected with suitable affinity reagents. Biotin ligase was used as
a model enzyme to assess the performance of the two methodologies. Active enzymes were
released by the bacteria, locally biotinylated the immobilized target substrate peptide, and
allowed the sensitive and specific detection of individual catalytically active colonies.
36
1.4.3 Phage display
Phage display is a powerful methodology that allows the selection of a particular phenotype
(e.g. a ligand specific to a desired antigen) from repertoires of polypeptides displayed on
phage. It was originally described in 1985 by Smith (110), who reported the use of the non-
lytic filamentous bacteriophage fd for the display of specific binding peptides on the phage
coat. The power of the methodology was further enhanced by the groups of Winter (111) and
Wells (112) who demonstrated the display of functional folded proteins on the phage surface
(an antibody fragment and a hormone, respectively). The technology is based on the fact that
a polypeptide (capable of performing a function, typically the specific binding to an antigen
of interest) can be displayed on the phage surface by inserting the gene coding for the
polypeptide into the phage genome. Thus, the phage particle links phenotype and genotype
(Fig. 1.12).
^P~
antigen
^ — displayed polypeptide
<D jo !'C j
Ü J
-c in_ :
Figure 1.12 Phage displaying a binding protein as fusion protein of a minor coat protein pill.
It is possible to create repertoires of phage (phage display libraries) in which the proteins
displayed on each phage represent a population of different molecules with different
properties. If a phage particle is isolated by virtue of its phenotype displayed on the surface
(e.g. the binding specificity), the genetic information coding for the protein is co-isolated. As
an example, one can consider the selection of a binding specificity from a displayed protein
library. The library on phage is panned against an antigen of interest; unbound phage are
vr- pill
— genotype
37
discarded whereas specifically binding phage are enriched and amplified in bacteria. Several
rounds of selection can be performed (typically 2-4 rounds). As a consequence, even very
rare phenotypes present in large repertoires can be selected and amplified from a background
of phage carrying undesired phenotypes (Figure 1.13).
S *m ^m lj
Domain libraryon phage
Immobi
antig
Next selection
AMPLIFICATION
/Infection of E.coli
bacteria
Figure 1.13 Selection of binders from a phage display library. A library of proteins displayed on
the phage surface is used as input for the selection. Phage which display a binding protein (red and
light blue) are captured on immobilized target molecules, and after unbound phage are washed off,
bound phage can be eluted. The eluted phage population is then used to infect E.coli cells and can be
propagated in bacterial cultures. The resulting phage population is enriched with target-specific
binders and can be used for subsequent rounds of panning.
Phage particles are very stable, they remain infective when treated with acids, bases,
dénaturants and even proteases. These properties allow a variety of selective elution protocols
and have been used for applications other than selection for binding, such as the selection of
proteins with altered thermal stability (113,114), the selection of catalytically active enzymes
(115-117), or aggregation resistant domain antibodies selected on phage by heat denaturation
(19).
38
Filamentous phage particles, approximately 6 nm in diameter and 900 nm in length, are
covered by several thousand copies of a small major coat protein (pVIIl). Few copies of the
minor coat proteins pill and pVI are displayed at one extremity of the phage particle, while
pVII and pIX are present at the other extremity. The minor coat protein pill, the product of
gene III, is displayed in 3-5 copies and mediates the adsorption of the phage to the bacterial
pilus. Peptides and/or proteins have been displayed on phage as fusions with the coat proteins
pill (110,118), pVIII (119), pVII and pIX (120,121). Display of proteins encoded by a cDNA
library as carboxy-terminal fusion with the minor coat protein pVI has also been reported
(122). The first peptides and proteins were displayed on phage using phage vectors
(essentially the phage genome with suitable cloning sites for pVIII or pill fusions and an
antibiotic resistance gene). Phage vectors carry all the genetic information necessary for the
phage life cycle. Using phage vectors, most peptides and folded proteins can be displayed as
pill fusions, while only short peptides of 6-7 residues without cysteine give rise to functional
phage when displayed as pVIII fusions (123).
Phagemids, a more popular vector for display, are plasmid vectors that carry only the gene III
with appropriate cloning sites and a packaging signal as phage derived sequences. For the
production of functional phage particles, phagemid containing bacteria have to be
supcrinfcctcd with helper phage particles, which contain the complete phage genome.
Phagemid vectors encoding the polypeptidc-pIII fusion are preferentially packaged into the
phage particles, because helper phage used for superinfection (such as M13K07 or VCS-
M13) have a defective origin of replication, which also serves as packaging signal. The
resulting phage particles may incorporate either pill derived from the helper phage or the
polypeptide-pIII fusion, encoded by the phagemid. Depending on the type of phagemid,
growth conditions used and the nature of the polypeptide fused to pill, ratios of (polypeptide-
pIII): pill ranging between 1:5 and 1:10000 have been reported (114,118). Furthermore, the
proteolytic cleavage of protein-pIII fusions has been described, contributing to further
decreased levels of polypeptide-pIII fusions on the phage (111). Therefore, it can be
concluded that phage particles, obtained by using a phage vector, are polyvalent (i.e. 3 to 5
identical polypeptides displayed on one phage particle), whereas the use of phagemids often
delivers monovalent phage, which is instrumental for the isolation of high-affinity binders. If
desired, performing superinfections with hyperphage, a phage which lacks most of the gene
III in the genome, polyvalent phage can still be generated with phagemids (124).
39
Not only antibody fragments, but also other proteins and enzymes have been successfully
displayed and selected on phage, e.g., domain antibodies (dAbs) (19), lipocalins (69), A-
domains (29), fibronectin type III domains (22), Kunitz type domains (59), SH3 domains
(78,79) biotin ligase and trypsin (117). However, not every protein can be displayed on phage
as fusion to pill. Even proteins that fold well in bacteria are frequently displayed poorly on
filamentous phage. The main reasons for inefficient display on phage are proteolytic cleavage
of the pill fusions and poor incorporation of the pill fusion protein into the phage coat,
because of a competition with the pill protein of the helper phage. Low protein presentation
on phage might be also caused by premature cytoplasmic folding, leading to inefficient
translocation into the periplasm (125). In order to obtain a sufficient number of phage
particles displaying the fusion protein, high titers of phage arc needed, thus limiting the
diversity of protein mutants that can be accessed for selection. A strategy to improve the
display of proteins on phage was published by Jestin et al. (126). In this work, optimized
leader peptides were selected, which are cloned upstream of pill fusion proteins and serve as
a signal for the export of pill fusions to the bacterial periplasm. More recently, the group of
Plückthun showed that redirecting the pill fusion proteins to the cotranslational SRP pathway
(instead of the post-translational Sec pathway) using an appropriate signal sequence results in
an efficient display of otherwise display-refractory proteins (125).
An additional drawback of phage display is the need for transformation of bacterial cells by
electroporation when cloning a library because the limiting factor in making large primary
libraries is the efficiency of introduction of plasmid or phage DNA into bacteria. This limits
the size of phage display libraries to maximum 1010 - 1011 individual library members (127).
1.4.4 Yeast display
Yeast display of antibody fragments has proven to be an efficient methodology for the
directed evolution of single chain Fv (scFv) antibodies for increased affinity and thermal
stability and, more recently, for the display and screening of a non-immune scFv and immune
Fab libraries. Yeast display is compatible with fluorescent-activated cell sorting (FACS) and
enables the screening of mid-sized libraries replacing the needs for selections (128).
However, when working with large libraries, the huge number of cells cannot be handled
40
anymore by a flow cytometer. If this is the case, magnetic beads coated with the antigen are
used for the first or two rounds of selection followd by screenings with FACS (129). A big
advantage of yeast display is that affinities and further important parameters like expression
or stability can be determined without the need for subcloning, expression and purification
(130). In yeast display, recombinant proteins are fused to the Aga2 protein, which engages in
a disulfide bonded complex with the Agal protein on the surface of Saccharomyces
cerevisiae (128). By labeling surface displayed antibodies with fluorescent antigen, yeast
cells can be screened by FACS for desired binding properties (Fig. 1.14).
Figure 1.14 The scFv Aga2 fusion protein surface expression system. Agal is bound to a cell wall
glucan and connected by a disulfide bond to Aga2. The protein to be displayed is cloned in frame with
Aga2 protein. Using suitable antibodies, N-terminal hemagglutinin (HA) tag and C-terminal c-myc
tag allow the monitoring of fusion protein expression. By addition of labeled antigen, yeast cells
displaying antibody fragments binding the antigen can be isolated by affinity purification (e.g.
biotinylated antigen) or FACS (e.g. fluorescently labeled antigen).
By means of a peptide tag which is appended to the scFv (hemagglutinin and myc tag in Fig.
1.14), a second fluorescent probe can be used to detect the amount of displayed antibodies.
This enables the discrimination between good expression and bad expression clones, and
eventually the affinity ranking of clones during screening (see also Fig. 1.15).
41
Enrichment of antigen-binding clones can be achieved by multiple (typically four in the case
of antibody libraries) subsequent rounds of enrichment on a cell sorter. Sort gates are drawn
such that the best 0.05 -1% clones are collected, depending on the stage of enrichment. This
includes that a threshold is set for scFv surface display (1st fluorophore) and for antigen
binding (2nd fluorophore), and the best clones, i.e. clones with a high 'antigen binding to
surface display'-ratio above these two thresholds, are collected.
Yeast surface display combined with FACS allows to perform selections in solution and does
not rely on immobilisation on a solid support which often hamper efficient selections due to
avidity effects. Moreover, doing selections in solution allows the investigator to precisely
control the antigen concentration and consequently establish conditions which allow to
optimally distinguish between high-affinity and low-affinity antibody clones (131) and also
to identify subtle affinity improvements (132).
Antigen binding
No scFv expression
ScFv and no antigen binding
—— ScFv with high affinity antigen binding
—— ScFv with low affinity antigen binding
Figure 1.15 Flow cytometric analysis of yeast-displayed antibody fragments in a population
greatly enriched for antigen binding (129). Yeast cells were labeled with biotinylated
antigen/streptavidin-phycocrythrin (detection of bound antigen), and anti-c myc/ anti-mouse-FlTC
42
(monitoring of scFv display level). The subpopulation of yeast cells in the blue circle does not display
scFv fragments. Despite good display of scFv, the subpopulation in the yellow circle does not bind to
the antigen. The subpopulations in the red and in the green circles represent clones with high levels of
scFv display (meaning good expression) and antigen binding. However, the subpopulation in the
green circle represents binders with higher affinities than the red population.
Characterization of clones, such as determination of the dissociation constant (KD), the
kinetic dissociation rate (koir) and stability analysis, which is usually a laborious and slow
process, can be done with monoclonal antibody populations displayed on the yeast surface
with high reliability, without the need for subcloning or purification of antibodies. Affinities
are determined using equilibrium-based kinetic analysis by measuring the degree of binding
over a wide range of antigen concentrations by FACS. The kinetic dissociation rate kofr is
determined by competition assays using an excess of unlabeiled antigen as competitor (133).
There are several examples of antibodies with extremely high affinities, which have been
affinity matured by yeast display. Boder et al. (130) used a scFv fragment called 4-4-20
recognizing the hapten fluorescein for affinity maturation. Mutagenesis and screening was
repeated three times, resulting in affinity matured scFv antibodies with dissociation rates over
four orders of magnitude slower than that of the parental scFv 4-4-20, and slower than the
dissociation rate of the streptavidin-biotin complex.
However, yeast display has its limitations. First, FACS throughput is limited and therefore,
big libraries (above 107) are not suited for FACS screening. Large naïve yeast surface display
antibody libraries have been used for subsequent rounds of panning using first magnetic bead
capturing in order to reduce the complexity of clones, followed by FACS screening (129).
Nevertheless, the strength of the methodology clearly resides in FACS. In cases where the
antigen is not monovalent, strong avidity effects may come into play due to dense scFv
display on the yeast cell wall, making yeast display inappropriate for the isolation of high-
affinity antibodies. Moreover, yeast cells arc less resistant to harsh experimental conditions if
compared with filamentous phage.
43
1.4.5 Ribosome display
The screening and display methodologies mentioned so far all have in common that library
size is limited to ~1010 individual variants due to the initial transformation of living cells with
library DNA. In addition, the cloning of libraries in the size of 109 can consume a
considerable amount of time and work. In order to circumvent this problem, selection
techniques have been proposed which can take place fully in vitro. This not only opened the
possibility to work with very large repertoires (>1012 members), but also shortened
significantly the time needed to generate sequence diversity, thereby allowing to repeatedly
introduce new mutations after each round of selection.
Ribosome display was first described by Mattheakis and colleagues (134), who displayed
short peptides on polysomes. Considerable improvements were needed to allow the display
and selection of folded proteins (135), such as scFv antibody fragments with dissociation
constants as low as 10"11 M (136). Proper folding in vitro is supported by introducing an
unstructured spacer region to the C-tcrminal end of proteins. This peptide spacer fills the
ribosomal tunnel and provides some extra flexibility, thus allowing the protein of interest to
fold as an independent unit and bind to the target. Additionally, chaperones and protein
disulfide isomerases (for proteins depending on disulfide bonds) are added to the translation
reaction (137). In ribosome display, a DNA library encoding a repertoire of proteins is
transcribed in vitro, the mRNA is purified and finally used for in vitro translation. As the
mRNA lacks a stop codon, the ribosome stalls at the end of the mRNA, giving rise to a
ternary complex of mRNA, ribosome and functional protein. Thus, the ribosome links the
phenotype and genotype. The ribosomal complexes, which are stabilized by high
concentrations of magnesium ions and low temperature (138), are directly used for selections
on a ligand either immobilized on a surface or in solution.
The mRNA incorporated in the bound ribosomal complexes is eluted by addition of EDTA,
purified, reverse-transcribed and amplified by PCR. During the PCR step, the T7 promoter
and the Shine-Dalgarno sequence are reintroduced by appropriate primers. Therefore, the
PCR product can be directly used for further selection cycles. Ribosome display is
schematically depicted in Figure 1.16.
44
DNA fragment coding for
scFv library
reverse transcription,
PCR amplification
in vitro
>
transcription
mRNA
without STOP codon
in vitro
translation
purification of mRNA ribosome/mRNA complexcannot dissociate
12é
affinity capture
mRNA eluted with EDTA
nascent scFV
antigen immobilised on solid support
Figure 1.16 Schematic representation of a selection cycle of ribosome display. Linear DNA
fragments coding for a protein library (here scFv variants) are transcribed in vitro and purified before
subsequent translation in vitro. After having reached the end of the mRNA during translation, the
ribosome is unable to dissociate from the mRNA because the stop codon is missing. The resulting
ternary complex comprising the ribosome, mRNA and the nascent polypeptide can be stabilised by
high concentrations of magnesium ions and low temperature, thereby creating a stable linkage
between the mRNA (genotype) and the encoded polypeptide (phcnotype). Ribosomes displaying a
binding protein can be isolated by affinity selection on immobilised antigen, the selected mRNA
molecules can be eluted, and the genetic information is amplified by reverse transcription and PCR,
which allows to introduce further variability by error prone PCR.
Hancs and colleagues have used a large synthetic antibody library, HUCAL-1 (139), for the
selection of antibodies specific to bovine insulin (140). Six rounds of selection were
performed, and after each round, further diversity was introduced into the selected genes by
error prone PCR. This procedure mimics to a certain degree the process of somatic
hypermutation of antibodies in vivo.
Ribosome display has been shown to work especially well for affinity maturation of scFv
45
fragments. So far, two studies have been reported in which a given antibody was evolved to
higher affinity. In both cases, stringent off-rate selections combined with error-prone PCR
were used. An antibody fragment specific to fluorescein with a dissociation constants of
about 100 pM was generated by Jermutus and colleagues (138). Interestingly, the majority of
the 4 to 11 mutated residues which were present in selected antibodies were found to be
unlikely to contact the antigen. In a second study, the dissociation constant of a scFv
fragment specific to a peptide deriving from the transcription factor GCN4 was improved
from 40 to 5 pM (141). Libraries of antibody mutants were generated with error-prone PCR
and DNA shuffling, and selected for decreased off-rates. Crystallographic analysis of the
scFv in its antigen bound and free states showed that only a few mutations, which do not
make direct contact to the antigen, lead to the affinity improvement. These results might
suggest that the affinity optimization of very high affinity binders is achieved by modulating
existing interactions via subtle changes in the framework rather than introducing new
contacts.
In a recent publication, ribosome display was successfully used for the isolation of binders
based on the ankyrin repeat scaffold (see chapter 1.2.6 and Ref (65)) which bound their target
with nanomolar affinities.
Throughout the selection process, high Mg2+ concentrations and low temperature arc needed
in order to maintain the integrity of the ternary ribosome-mRNA-polypeptide complexes, and
therefore, selection conditions can only be varied to a certain extent, which might hamper the
isolation of proteins that are intended to be stable under harsh conditions, e.g. high
temperature or extremes of pH. Since mRNA is used as carrier of genetic information, care
should be taken to avoid contamination with RNAse molecules.
1.4.6 Covalent DNA display
Bertschinger and Neri recently proposed a technology called covalent DNA display (142) in
which a library of linear DNA molecules is co-packaged with an in vitro
transcription/translation mix in the compartments of a water-in-oil emulsion (143) (Figures
46
1.17 and 1.18). Experimental conditions are chosen so that, in most cases, one compartment
contains one DNA molecule. A covalent linkage between DNA and the encoded polypeptide
is achieved by using fusion proteins which contain the Hae III DNA-methyltransferase
domain of Haemophilus aegypticus, an enzyme which is able to form a covalent bond with
DNA fragments containing the sequence 5'-GGFC-3' (F = 5-fluoro-2'-deoxycytidine) (144).
M.Hae III fusion proteins expressed in each compartment of the water-in-oil emulsion are
extracted from the emulsion, and DNA molecules displaying a protein with the desired
binding specificity are selected from the pool of DNA-protein fusions by affinity selection.
The genetic information of the selected DNA-protein fusions is amplified by polymerase
chain reaction (PCR) and can be used either for a further round of selection or for cloning and
characterisation of the selected mutants.
water-in-oil emulsion
library of PCR fragmentscoding for M.Mae III fusion proteins
>*
subcloningand screening Î
amplification of geneticinformation by PCR
(introduction of mutations)
library of proteins covalentlybound to their encoding DNA
capture of binding proteinson immobilized antigen
Figure 1.17 Concept of the selection method. A library of linear DNA molecules is co-packaged
with an in vitro transcription/translation mix into a water-in-oil emulsion (1). Ideally, one
compartment contains one DNA molecule. The DNA molecules each code for a Hae III DNA-
methyltransferase fusion protein and contain a mechanism-based inhibitor for the covalent cross-
linking of the DNA-methyltransferase fusion proteins (see also Fig. 1.18). After in vitro expression
47
and formation of the DNA-protein complexes, the water phase is extracted from the emulsion (2) and
DNA molecules displaying a protein with desired binding properties are selected from the pool of
DNA-protein fusions by affinity panning (3). The genetic information of selected DNA-protein
fusions is amplified by PCR (4) and can either be used for a further round of selection (5) or for
cloning and characterisation of the selected mutants (6).
Figure 1.18 Enlarged view of a water compartment of the water-in-oil emulsion. The DNA
molecule is transcribed into mRNA, which is translated into a fusion protein consisting of two
domains: the N-tcrminal Hac III DNA-methyltransferase (yellow) and a C-terminal potential binding
domain (red). Owing to the catalytic activity of the Hae 111 DNA-mcthyltransferase, the fusion
proteins form a covalent bond via the modified methylation target sequence 5'-GGFC-3' (F = FdC =
5-fluorodeoxycytidinc) with their encoding DNA molecule, which is present in the same compartment
of the water-in-oil emulsion.
Bertschinger et al. demonstrated successful enrichment of DNA molecules on the basis of the
binding specificity of the protein they encode from an excess of DNA molecules coding for a
protein with irrelevant binding specificity (142): mixtures of DNA templates coding either
for M.Hae III-calmodulin (CaM) or M.Hae I1I-EDB (145) were prepared in such a way that
the ratio of M.Hae III-EDB /-CaM molecules was —1000. CaM-binding peptide was used to
enrich DNA coding for M.Hae III -CaM from the DNA mixtures, resulting in enrichment
factors in the range of 102.
48
The technology was further optimized by the authors, showing that the covalent cross-linkage
of the M.Hae III fusion proteins with their encoding DNA takes place within the water
compartment and that the covalent bond is indeed necessary for the efficient recovery of
DNA molecules coding for binding proteins. In addition, several steps of the selection
protocol were optimized, such as emulsion preparation, elution of the selected DNA
molecules and their subsequent PCR amplification. With the optimized protocols, we recently
successfully isolated affinity matured Fyn SH3 derivatives binding to mouse serum albumin
(MSA) (146).
Overall, covalent DNA display represents a useful addition to the protein engineer's toolbox
with the potential to deliver good-quality binding proteins for biomedical applications.
However, so far only model selection and affinity maturation experiments have been
performed. It remains to be seen, whether the technology is feasible for selections with naive
libraries of small globular proteins.
49
1.5 Serum albumin as a target protein
In this thesis, two different target proteins have been chosen, mouse serum albumin (MSA)
and the extra-domain B (EDB) of fibronectin, a marker of angiogenesis. The latter one is
described in the next chapter within the context of vascular tumor targeting.
Albumin (molecular mass ~ 67 kDa) is the most abundant protein in plasma, present at 50
mg/ml (600 uM), and has a half-life of 19 days in humans (147). It maintains plasma pH,
contributes to colloidal blood pressure, functions as a carrier of many metabolites and fatty
acids, and serves as a major drug transport protein in plasma. There are several major small
molecule binding sites in albumin that have been described. Warfarin is known to bind at site
I, benzodiazepines at site II and cardenolides and biliary acids at site III. In addition, there is
an important metal ion binding site.
Because the kidney generally filters out molecules below 60 kDa, efforts to reduce clearance
have focused on either increasing molecular size through protein fusions (e.g., fusion to
albumin or to the Fc portion of an IgG), glycosylation, the addition of polyethylene glycol
(PEG) or on albumin binding polypeptides (148).
The noncovalent association with albumin has been shown to extend the half-life of short¬
lived proteins. A recombinant fusion of the albumin binding domain from streptococcal
protein G to human complement receptor type I increased its half-life 3-fold to 5 h in rats
(149). In another example, when insulin was acylated with fatty acids to promote association
with albumin, a protracted effect was observed when injected subcutaneously in pigs (150).
Genentech recently engineered fusion proteins consisting of a Fab fragment of Herccptin®
and different albumin binding peptides, each of them differing in their affinity to albumin,
ranging from 0.04 to 2.5 uM (151). Reduced affinity for albumin correlated with a reduced-
half-life and higher clearance rates. Since it has been observed that size and half-life of tumor
targeting agents can have a dramatic effect on their ability to accumulate in the tumor (152),
the fine-tuning of albumin binding affinities may improve tumor to organ ratios for these
classes of ligands.
50
Beside pharmacokinetic aspects, albumin binding molecules can be used for bioseparation
purposes, since the identification of biomarkers from serum or plasma is often hindered by a
few proteins present at high concentrations, which may obscure less abundant proteins.
Moreover, the identification of these biomarkers is hindered not only by their low absolute
abundance (ng/mL serum or less) but also by the complex and heterogeneous nature of serum
and plasma, where a modest number of proteins (such as albumin, IgG, IgA, haptoglobin,
transferrin, a 1-antitrypsin and a2-macroglobulin) are present at high concentrations (mg/mL),
accounting for more than 85% of the total serum proteins (153). By probing a plasma sample
over an affinity resin conjugated to an albumin binding domain, efficient removal of albumin
can be achieved.
In this thesis, albumin served as target protein for model selections with libraries of the Fyn
SH3 domain.
51
1.6 Vascular tumor targeting
1.6.1 Concept and definitions
The treatment of solid tumors with most chemotherapeutic drugs relics on the expectation
that the drugs will preferentially kill rapidly dividing tumor cells, rather than normal cells.
However, the lack of selectivity towards tumor cells leads to toxicities in normal tissues with
enhanced proliferation rates, such as the bone marrow, gastrointestinal tract and hair follicles.
As the high interstitial pressure and the irregular vasculature of the tumor impairs the
accumulation of the active agents at the tumor site, the efficiency of conventional therapy is
further decreased (154). Moreover, the activity of multidrug resistance proteins minimizes
drug uptake and leads often to failure of the therapy (155). A promising approach to
circumvent the hurdles of tumor therapy is the emerging field of vascular targeting.
Vascular tumor targeting aims at the rapid and selective shutdown or damage of the
established tumor vasculature. Consequently, this strategy will lead to tumor cell death as the
blood supply to these cells has been cut off. Potential advantages of this strategy over
attacking tumor cells are:
- Occluding one blood vessel will trigger cell death of a large number of tumor cells
which depend on it for supply of nutrients and oxygen
- One single vascular targeting agent could in principle be used to treat a wide range of
solid tumors
- Endothelial cells and cells of the surrounding stroma are genetically more stable than
tumor cells, so therapy resistance is less likely to occur. Moreover, these cells are
usually easily accessible for any compound
52
A broader definition of vascular targeting, which is not limited to the thrombosis of neo-
vasculature, may be "the targeted delivery of bioactive agents to new blood vessels". The
schematic shown in Fig. 1.19 illustrates the overall concept.
Figure 1.19 Illustration of the concept of vascular targeting. The targeted drug is delivered
intravenously and homes to the tumor-induced antigen that might be cither on the endothelial cell or
in the perivascular space.
Vascular tumor targeting is only possible because the endothelium and the surrounding
stroma in tumors differ from that in normal tissue; this fact has long been known but only
recently have these differences begun to be characterized at the molecular level. Compared
with the vasculature in normal tissue, the tumor vasculature is strikingly disorganized and
tortuous (156-158). The flow of blood through the tumor capillaries is frequently sluggish,
and at times might be stationary or even experience a reversal in the direction of flow
(159,160). The microenvironment - including blood and the endothelium lining the vessels -
is profoundly hypoxic (161). In this environment, the endothelial cell proliferates rapidly and
contributes to active angiogenesis (162). The tumor is also nutrient starved, acidic and under
oxidative stress (163,164); it has been shown that the endothelial cell responds
transcriptionally to these stimuli and remodeling activities, giving rise to the production of
new proteins either on their surface or in the surrounding extracellular matrix. Some of this
specific markers can be used as targets for vascular targeting approaches.
53
1.6.2 Classes of vascular targets
Extra-Domain B (EDB) offibronectin
Fibronectin is a large glycoprotein that is present in large amounts in the plasma and tissues.
EDB is a 91-amino-acid type III homology domain that becomes inserted into the fibronectin
molecule under tissue-remodelling conditions by a mechanism of alternative splicing at the
level of the primary transcript (145). EDB is essentially undetectable in healthy adult
individuals whereas EDB containing fibronectin is abundant in many aggressive solid tumors
and displays cither predominantly vascular or diffuse stromal patterns of expression,
depending on the tumor type (165). Despite its very restricted expression, the function of
EDB does not seem to be indispensable, as mice lacking the EDB exon develop normally
(166). The EDB sequence is identical in mouse, rat, rabbit, dog, monkey and man. This
feature facilitates animal experiments in immunocompetent syngeneic settings, but has, so
far, prevented the isolation of anti-EDB antibodies using hybridoma technology, probably
due to tolerance.
A few years ago, using phage display technology (167) and other technologies (107) our
group in collaboration with the group of L. Zardi (Genova, Italy) succeeded to isolate a
number of human monoclonal antibodies to EDB (168-170). These include the high-affinity
human antibody LI9, which has been shown to efficiently localize to tumor blood vessels in
animal models (171-173) and in patients with cancer (174) following intravenous injection. A
large number of therapeutic derivatives of the L19 antibody have been produced and tested in
animals, including conjugates to fluorophores and photosensitizcrs (175,176), therapeutic
radionuclides (177,178), liposomes (179), procoagulant agents (180), cytokines (181-185),
enzymes (186) and other proteins (187,188). Importantly, the anti-EDB antibody LI9, in
homodimeric single-chain Fv format and labeled with iodine-123, has been studied in over 40
patients with cancer. The results obtained in the first 20 patients have recently been described
(174) and confirm the ability of the antibody to localize to tumor masses exhibiting rapid
growth. L19-intcrlcukin 2 (L19-IL2), L19-tumor necrosis factor (L19-TNF) and Small-
54
Immuno-Protein [SIP]-L19-I131 arc three therapeutic derivatives of the L19 antibody which
are currently in clinical development (189).
Large Tenascin-C isoforms
Tenascin-C, a polymorphic high molecular mass extracellular matrix glycoprotein exists in
several isoforms which are generated as a result of different patterns of alternative splicing in
the region between domains Al and D (190). These large isoforms of tenascin-C, containing
extra domains, have long been known to be tumor associated antigens. Though not
completely absent in healthy tissues, they show a more restricted pattern of expression
compared to the isoforms without extra domains (191). Especially the C domain of tenascin-
C shows the most restricted expression pattern: while being undetectable in normal human
tissues and only barely detectable in most carcinomas, it is extremely abundant in high grade
astrocytoma (grade III and glioblastoma) and in lung cancer, particularly around vascular
structures and proliferating cells (192,193).
A critical immunohistochemical analysis of the expression pattern of the different isoforms in
various cancer types is needed to evaluate their potential as targets for biomolecular
intervention. Radiolabeled derivatives of monoclonal antibodies to domains Al and D of
tenascin-C have been used for imaging and radioimmunotherapy in patients with cancer for
over a decade (194-198). The pattern of staining of these antibodies varies between different
tumors, the two extremes being a predominantly vascular and a diffuse stromal staining.
Recently, Silacci et al. (193) and Brack et al. (199) reported about the isolation of antibody
fragments specific for the domain C and the domain Al of tenascin-C, with excellent
quantitative biodistribution results in mouse xenograft U87 glioma models.
Phosphatidyl Serine
Phosphatidyl serine (PS) phospholipids are major components of the cell membrane which
are preferentially found in the inner leaflet of the lipid bilayer. However, under conditions of
cellular stress, apoptosis, platelet activation and endothelial-cell proliferation in tumors, PS
becomes exposed on the outer leaflet of the cell membrane (200,201). Annexin V and
55
monoclonal antibodies have been used to confirm the surface accessibility of the
phosphatidyl serine moiety on endothelial cells in vitro (for example, after treatment with
hydrogen peroxide) and in vivo (200,202). The impressive microscopic analysis of tumor
targeting performance by monoclonal antibodies to PS have not yet been complemented by a
quantitative biodistribution analysis, but the 9D2 and 3G4 antibodies displayed potent
antitumor activities even when used as naked antibody in rodent models of cancer. Recently
it could be shown that the vascular targeting antibody, 3G4, significantly enhances the
therapeutic efficacy of docetaxel against the growth and dissemination to the lungs of MDA-
MB-435 human breast tumors in mice without concomitant increase in host toxicity (203). A
chimeric version of 3G4 has been developed and is scheduled to enter clinical trials now.
Annexin A1
Annexins are cytosolic proteins that can associate with cell membranes in a calcium
dependent manner. Some annexins may translocate the lipid bilayer to the external cell
surface. Annexin Al was recognized as a tumor endothelial target by Schnitzer and co¬
workers (204). A monoclonal antibody to this antigen has been used for the
radioimmunoscintigraphic detection of solid tumor lesions in a rat model. Furthermore,
relatively low radioactive doses of the same antibody labeled with iodine-125 have shown a
therapeutic benefit in rats.
lntegrins
During vascular remodeling and angiogenesis, endothelial cells show increased expression of
several cell-surface molecules that potentiate cell invasion and proliferation. One such
molecule is the integrin-avp3, which has a key role in endothelial cell survival during
angiogenesis in vivo and which might serve as a target for therapeutic molecules, particularly
those that require internalization in endothelial cells. Monoclonal antibodies to the integrin-
avß3 have been shown to display anti-angiogenic activities and to preferentially stain tumor
56
blood vessels. However, expression of integrin-avß3 and other intcgrins has been reported in
several normal tissues.
Efforts to influence the biology of blood vessels by gene delivery have been pursued using
cationic nanoparticles coupled to an intcgrin-avß3 targeting ligand for the selective gene
delivery to angiogenic blood vessels, with a substantial therapeutic benefit in tumor bearing
mice (205).
A high-affinity humanized anti-avß3 antibody is in clinical development as an anti-
angiogenic therapeutic (206), however, so far its tumor-targeting performance in patients
with cancer has been unsatisfactory (207).
Anti-avß3 antibodies have been shown to preferentially localize to tumor blood vessels using
ex vivo fluorescence microscopy detection (208). Furthermore, a paramagnetic contrast agent
targeted to the LM609 monoclonal antibody, which is specific to avß3, has been described for
the in vivo imaging of angiogenesis using magnetic resonance (209).
Vascular endothelial growthfactors (VEGFs) and their receptors
VEGFs represent a class of proteins that mediate angiogenesis. The overexpression of
VEGFs and their receptors in tumors (204,210) makes them attractive targets. Especially, the
recent approval of the humanized anti-VEGF monoclonal antibody bevacizumab for first-line
cancer treatment (211,212) has highlighted the contribution of VEGF-A to cancer
progression. In addition, the selective localization of monoclonal antibodies to VEGF-A,
VEGF receptor 2 and the VEGF-A/VEGF receptor 2 complex has been studied (213-217).
The targeting efficiencies reported so far were modest, which possibly reflects kinetic
limitations in the targeting of low or medium abundance antigens, even when they arc readily
accessible to binding agents injected into the bloodstream (183).
Prostate-specific membrane antigen (PSMA)
PSMA is a membrane glycoprotein with proteolytic activity. It is predominantly expressed in
the prostate and scrum concentrations are often elevated in patients with prostate cancer
57
(218). The interest for vascular targeting applications of PSMA has been stimulated by the
observation that PSMA is over-expressed in the neo-vasculature of several solid tumors
(219,220), whereas expression around blood vessels in normal tissues is limited to breast,
kidney, duodenum and prostate. Specific antibody-derivatives to PSMA with alpha-emitting
radionuclides in rodent cancer models have been reported (221). More recently, the
radiolabeled antibody J591 has been used for tumor imaging (222) and is currently being
evaluated for therapeutic applications (223,224),
Endoglin
Endoglin (CD 105) is a transforming growth factor-ß (TGF-ß) co-receptor that is
overexpressed in tumor neovasculature (225,226). Even though immunohistochemical
analysis has shown endoglin expression in normal adult tissues (227,228), monoclonal
antibodies to endoglin have been used in biodistribution studies and for imaging in rodents
and dogs (229,230).
Nucleolin
Ruoslahti and colleagues reported about a 31-amino acid synthetic peptide (F3) that
accumulates in the nuclei of tumor endothelial cells and tumor cells (231). The cell surface
protein that is recognized by F3 was then identified as nucleolin (232). The internalization of
F3 takes place in a nuclcolin-dependent manner, as antinucleolin antibodies that were
previously injected inhibited F3 cellular uptake. The restricted expression pattern of nucleolin
and its ability to internalize binding agents make it an attractive target for directed tumor
therapy.
Otherpossible targets
Over the past few years, a range of antigens have been proposed as vascular targets for
imaging and for the selective delivery of drugs to the tumor.
58
Peptides to the endothelial antigen CD 13 have been isolated (233,234), and peptide fusions
with tumor necrosis factor (TNF) have shown to markedly increase the therapeutic ratio of
this biopharmaceutical (235), even at low doses (236).
Radiolabeled monoclonal antibodies against a CD44 isoform, which is a cell-surface
receptor, have shown impressive performance in tumor targeting experiments in animal
models (237).
Magic roundabout or ROB04, a member of the roundabout receptor family, is a promising
target as it is absent in adult tissues except at sites of angiogenesis (238). This gene product is
believed to be intimately involved in the development of the vasculature. Biodistribution
studies, e.g. with radiolabeled antibodies, will give more information about the suitability of
ROBO-4 as a vascular tumor target.
Developmental endothelial locus-1 (Del-1) is a unique integrin ligand produced by
endothelial cells, thus mediating autocrine signals. It has been reported to mediate
angiogenesis in the adult (239). Biodistribution studies are needed for the evaluation of the
suitability of Del-1 as a target.
Endothclial-specific protein disulfide isomerase (EndoPDI) is a hypoxia induced gene that is
predominantly expressed by the endothelium (240). Functionally, EndoPDI has been shown
to protect the endothelium from apoptosis during hypoxia-induced stress. Use of RNA
interference has shown to that EndoPDI is required for the folding of endothclial-protective
molecules, including endothclin-1 (EDN1), adrenomedullin (ADM) and CD105, which are
produced when endothelial cells are exposed to hypoxic stress. Inhibition or ablation of
EndoPDI should, therefore, make the hypoxic tumor endothelium more sensitive to
apoptosis, particularly if it is concurrently stressed by a cytotoxic drug or radiotherapy.
1.6.3 Imaging applications of vascular tumor targeting
One of the most straightforward biomedical applications of ligands capable of selective
localization around tumor vascular structures may reside in their use for the macroscopic in
vivo imaging of sites of disease. The visualization of the homing of ligands to vascular
structures imposes two main requirements: (i) the use of suitable chemical modification
which makes the ligand visible, and, (ii) the choice of macromolccular targets which are
59
abundant enough to counterbalance the small contribution of vascular structures to the overall
solid tumor mass.
In principle, four main chemical modification strategies can be envisaged for the molecular
imaging of angiogenesis-relatcd diseases:
- the radioactive labeling of ligands, for SPECT or PET imaging modalities (241 -243)
- the use of near-infrared (NIR) fluorophores, which can be detected using cpi-
illumination (170,175), optical coherence tomography (244) or diffuse optical
tomography methodologies (245,246)
- the use of ligands for the targeted delivery of microbubbles, to be used as contrast
agents for sonographic imaging applications (247,248)
- the use of magnetic nanoparticles (249)
The ability of radioactively or fluorescently labeled ligands to image tumors in vivo is by now
well established. For macroscopic imaging applications, NIR fluorescence imaging may be
limited by the reduced light penetration of tissues (176,250). However, fluorescently labeled
antibodies may facilitate the microscopic imaging of superficial lesions (e.g., using
endoscopic methods (251)) and of transparent structures within the body (e.g., angiogenesis-
related ocular disorders (176,252), Furthermore, the use of NIR dyes may open novel
angiographic imaging opportunities for solid tumors and other conditions, such as
atherosclerosis (246,252).
It is much more debatable whether ligand-modified microbubbles and nanoparticles can
efficiently extravasate and reach abluminal antigens in tumors and other diseases. Promising
tumor imaging results with antibody-quantum dot conjugates suggest that even large particles
can selectively accumulate at the tumor site. This contrasts with experience of our group,
suggesting that highly charged antibody derivatives (183,187,188) and large antibody
derivatives (253) may completely abrogate the tumor targeting ability of the parental
antibody in vivo, while retaining unperturbed antigen binding properties.
60
1.6.4 Therapy applications of vascular tumor targeting
The following discussion is limited to derivatives of tumor targeting antibodies, a class of
therapeutic agents for which a substantial amount of preclinical experimental data is now
available. But, some of the considerations made in this section may be applicable for other
classes of ligands, in particular for small globular proteins.
Figure 1.20 illustrates some of the main classes of antibody-derivatives which have been
considered for tumor targeting applications. Some of these strategies (e.g., antibody-
photosensitizer and antibody-pro-coagulant conjugates) appear to be ideally suited for
vascular targeting applications, since they may lead to intravascular blood coagulation,
causing an avalanche of tumor cell death (254,255).
IMAGING
with enhanced
effector functions
liposomes loaded
with toxtc moieties
THERAPYprocoagulant factors
photoiensitizer
conjugates
drug conjugates
Figure 1.20 Antibody derivatives which could be considered either for tumor imaging or therapy.
Many of the antibody functionalization strategics depicted in Figure 1.20 have been applied
both to full immunoglobulin and to antibody fragments. In most cases, smaller antibody
fragments are preferred because of easier expression, rapid blood clearance and lack of Fc
61
(which avoids the undesired targeting of bioactive moieties to cells bearing Fc receptors).
However, some therapeutic strategies such as antibody-drug conjugates, may benefit from the
long blood circulation times of antibodies in the IgG format.
Antibody fragments have successfully been coupled to cytokines (181-185), chemokines
(256), fluorophorcs and photosensitizers (175,176,255), drugs (257), pro-coagulant factors
(254), enzymes for pro-drug activation (258) radionuclides (259) and liposomes (260), but
also to more exotic functional moieties such as uranium-loaded ferritin for neutron capture
therapy (261).
Figure 1.21a shows the selective accumulation at the tumor site of the SIP-L19 antibody
labeled with the infrared fluorophorc Cy7. Figure 1.21b illustrates a striking therapeutic
effect, observed in orthotopic animal models of hepatocellular carcinoma, using the fusion
protein L19-IL2. As mentioned above, the vascular targeting properties of the L19 antibody,
specific to the EDB domain of fibronectin, can be used to dramatically improve the tumor
accumulation of the pro-inflammatory cytokine IL2, thus enhancing the therapeutic index of
this anti-cancer biopharmaceutical.
Most of the studies performed with antibody derivatives suggest that the tumor targeting
properties ofjudiciously chosen antibodies can substantially potentiate the therapeutic action
of the bioactive moiety chosen for antibody coupling.
62
Figure 1.21 (a) Selective accumulation of L19 in the tumor. Twenty four hours near-infrared image
of 129SvEv mouse subcutaneously grafted with F9 teratocarcinoma, i.v. injected with SIP(L19),
labeled with the infrared fluorophore Cy7. The tumor is indicated by an arrow, (b) The therapeutic
efficacy of L19-IL2 in the orthotopic HuH7 mouse model for hepatocellular carcinoma. HuH7
tumors (tumor volume: 70-100 mm3 prctrcatmcnt) were treated with two treatment cycles consisting
of 5 consecutive daily i.v. bolus injections of L19-IL2 or PROLEUKIN (free 1L2; two dose levels for
each drug) followed by two drug-free days. After the second treatment cycle, the animals were
sacrificed and tumor volumes were measured. Compared with the non-targeted IL2 (proleukin), LI 9-
IL2 showed a superior efficacy.
63
1.7 Aim of this thesis
The aim of my PhD thesis was the development of technologies for the isolation of single
domain binders based on the Fyn SH3 scaffold, with a special interest in the isolation of
binders to mouse serum albumin (MSA) and the extra-domain B of fibroncctin (EDB).
The creation of proteins that have new binding properties is an important goal in protein
engineering. Beside giving new insights into processes of molecular recognition, such binders
also have potential commercial applications, as therapeutic agents, diagnostic reagents or
affinity ligands (10).
In biotechnology and biomedical research, antibodies and fragments thereof are currently the
most widely used specific, high-affinity binding molecules; they exhibit a favorable
pharmacokinetic profile and can be generated against essentially any target either by
immunization or by using natural or rationally designed antibody libraries in vitro (36).
However, antibodies suffer from certain drawbacks, as for example relatively low expression
yields, requirement for an expensive mammalian cell production system, dependence on
disulfide bonds for stability and, in addition, some antibody fragments tend to aggregate (12).
There is a clear need for the creation of an alternative protein scaffold, that may lead to the
same affinity and specificity as antibodies without any of the above mentioned drawbacks.
An appropriate scaffold would be human, devoid of cysteine residues, well expressed in
bacteria or yeast and would have good biophysical properties (solubility, stability). As
previously discussed in more detail, the Fyn SH3 domain scaffold fulfills important criteria
for being an optimal scaffold candidate.
I have constructed libraries of mutants of the Fyn SH3 domain and performed selection
experiments with mouse serum albumin (MSA) and EDB, a marker of angiogenesis (145).
The characterization of the EDB-binding Fyn SH3 derivatives in vitro and in vivo has been an
important goal of the thesis. Moreover, the complete identity of Fyn SH3 in mouse, rat,
gibbon monkey and man has facilitated the demonstration of a low immunogenic potential of
Fyn SH3 mutants in immunocompetent mice.
64
Fyn SH3 derived EDB-binders may be considered as useful molecules for the targeted
delivery of bioactive agents to the tumor neo-vasculature in full analogy to derivatives of the
L19 antibody, which are currently being investigated in clinical trials (189). In addition, the
single-pot library of Fyn SH3 mutants described in this thesis may represent a rich source of
target protein binding specificities, thus, it may provide useful reagents for many biochemical
and biomedical applications as an alternative to more conventional IgG-based
immunochemical technologies.
65
2 Results
2.1 Expression of Fyn SH3 wild-type and Fyn SH3 mutants
2.1.1 Cloning and expression of Fyn SH3 wild-type
The gene encoding the Fyn SH3 wild-type domain [amino acid residues 83-145, numbering
according to the Fyn kinase sequence reported by Refs (100,101)] was amplified from a
human cDNA library and cloned into the bacterial expression vector pQE-12, appending a C-
terminal His tag. The 8,6 kDa recombinant protein was expressed in E.coli and purified from
cleared lysate by immobilised metal affinity chromatography (IMAC) using Ni-NTA resin.
The quality of the protein was assessed by SDS-PAGE. Expression yields were ~ 60 mg
protein per liter of bacterial culture in shake flasks (Fig. 2.1).
a b
M wt
Figure 2.1 Cloning of Fyn SH3 wild-type, (a) Schematic map of vector pQE-12, appending a
hcxahistidine tag to the C-terminus of the inserted protein, (b) SDS PAGE of the purified Fyn SH3
domain. The protein runs at the expected size of 8,6 kDa.
66
2.1.2 Test library designs and characterization
In order to introduce novel binding specificities into the Fyn SH3 domain, amino acid
residues in the RT- and n-Src-loop were randomized (Fig. 2.2).
G
GGA
V
GTG
T
ACA
L
CTC
F
TTT
V
GT3
A
GCC
L
CTT
Y
TAT
D
GAC
TAT
E
GAA
A
GCA
R
CGj
T
ACA
E
GAA
D
GAT
D
GAC
L
CTG
S
AGT
F
TTT
H
CAC
K G
GCA
B
GAA
K
AAA
F
TTT
Q
CAA
I
ATA
L
TTG
N
AAC
3
AGC
S
TCG
E
GAA
G
GGA
L
GAT
W
TGG
W
TGG
E
GAA
A
GCC
R
CGC
S
rcc
I.
TTC
T
ACA
T
ACT
n
GGA
F.
GAG ACA
G
ÜGT
Y
TAC
I
ATT
p
ccc
S
AGC
N
AAT
Y
TAT
V
GTG
A
GCT
P
CCA
V
GTT
D
GAC
C
TCT
I
ATC
0
CAG
lib 1
lib 2
lib 3
RT Fyn SH3 *>
Fyn SH3 I src J>
Fyn SH3 src
1.PCR
2. PCR assembly3. cloning into pQE-12
J
Figure 2.2 Fyn SH3 wt structure, amino acid sequence and library cloning strategy, (a), left
structure of the Fyn SH3 domain (PDB code 1M27) prepared with the software Pymol,
www.pymol.org .The RT loop is in green, and the n-Src loop in red. right DNA and amino acid
sequence of the Fyn SH3 domain. The loop regions are also indicated by the corresponding colors, (b)
Library cloning strategy: Mutations were introduced either in the RT (lib 1) or in the n-Src (lib 2) or
in both loop regions (lib 3) by PCR using partially degenerate primers. PCR fragments were then
assembled by PCR and cloned into the pQE-12 vector.
67
Since it was not clear a priori, whether mutations in these regions of the wild type domain
would have an influence on expression, three small test libraries of Fyn SH3 mutants were
constructed by combinatorial mutagenesis, using the Fyn SFB wt gene as template. In the
first library, 6 amino acids of the RT loop were randomized (residues 94-99), in the second
library the n-Src loop was extended from 4 to 6 randomized residues (residues 113-116, plus
2 residues of extension), and in the third library both loops were randomized simultaneously.
Figure 2.2 shows the structure and sequence of Fyn SH3 wt, the corresponding loops and the
library construction strategies.
Individual clone members of the libraries were randomly picked and grown in a 96-well
plate. After induction of protein expression, a dot blot experiment with the lysates revealed
that about 60% of the selected clones of library 1, ~ 90% of library 2 and 60% of library 3
expressed a detectable amount of soluble Fyn SH3 mutants (Fig. 2.3). These levels of
functional expression were considered as good enough for phage library constructions.
C
• • m • #
• • • •
• • ©+
• • •
• • • • •
• • •
Figure 2.3 Dot blot experiments of the test libraries. Dot blot analysis of induced lysates of
individual clone members. Fyn SH3 mutants were detected with anti-His-HRP antibody conjugate, (a)
About 60% of the clones of library 1 (RT loop randomized), (b) 90% of the clones of library 2 (n-Src
loop randomized), and (c) - 65% of the clones of library 3 (simultaneous randomization of RT and n-
Src loops) expressed a detectable amount of soluble protein. The positive control represents Fyn SH3
wt, the negative control bacterial growth medium (2xYT).
68
2.2 Phage library design, cloning and characterization
The previous section showed that at least ~ 60% of mutants with altered RT and n-Src loops
expressed a detectable amount of soluble protein. In a first attempt, we were interested in
creating a phage display library of Fyn SH3 variants with 6 randomized residues in the RT
loop (postions 94-99, Fig 2.4a), since Saksela and colleagues reported that the RT loop of the
Hck SH3 domain contributes to target binding (79). After PCR amplification of the
randomized gene fragment of library 1 (see previous section) and thereby introducing
appropiate restriction sites (Fig 2,4b), a total of 3.3 |ig of insert was ligated into 17 jig of
double-digested phagemid vector pHENl and electroporated into freshly prepared
electrocompetent E.coli TGI cells, resulting in a library containing 1.2 x 107 individual
clones (theoretical library size of 6.4 x 107).
RT Fyn SH3
1 amplification2 cloning into pHENl
amber
codonNcol Istotl +
pelB J~l \-\ myc tag [j gene pill
PHËN1
COIE1 oil i—\ Amp Ï—I M13orl
Figure 2.4 Visualization of the Fyn SH3 wt RT loop and library cloning strategy, (a) Structure of
Fyn SH3 wt domain (1M27); the 6 randomized amino acid residues of the RT-loop-libary are shown
in green (positions 94-99). (b) Library cloning strategy. The randomized fragments of the test library
1 were amplified by PCR and cloned into the pHENl vector (262).
The quality and functionality of the library were assessed by PCR colony screening, dot blot,
Western blot and DNA sequencing. PCR colony screening showed that 77 % of randomly
picked clones contained an insert of the correct size. As expected from the previous test-
libraries, the dot blot analysis of this newly created library revealed that - 62% of the
analyzed library members expressed soluble Fyn SH3 mutants (Fyn SH3m). The efficient
69
display of Fyn SH3m-pIH fusion proteins on the surface of filamentous phage is an absolute
prerequisite for performing phage display selections. The performance of the display of Fyn
SH3m-pIII fusion proteins was evaluated by Western blot, using an anti-pill antibody as
primary detecting reagent. From the ratio of intensities of the bands corresponding to pill or
Fyn SH3m-plll fusions (Fig 2.5), and considering that 3-5 copies of pill are present at the tip
of the phage, we estimated that approximately 1:4 to 1:7 phage particles display the Fyn
SH3m fusion protein on its surface. This level of display is very high and comparable to
recently described antibody scFv libraries (118).
12 3 4
- - 75kDa
FynSH3m-pllh-> J^n^p|||~->'^^^ —- 50kDa
0
Figure 2.5 Western blot analysis to evaluate the efficiency of the display of the Fyn SH3mutants
(Fyn SH3m)-pIII fusion protein on the surface of the phage particle. Different amounts of
purified phage were analyzed: lane 1, 4xl010 transforming units (t.u.); lane 2, 1x10 t.u.; lane 3,
2xl09 t.u.; lane 4 lxlO9 t.u. The protein pill and the Fyn SH3m-pIII fusion protein were detected with
a monoclonal anti pill antibody.
70
Fifteen randomly picked clones were sequenced, revealing that all amino acid sequences in
the RT loop region were diverse; only one clone was not in frame due to a single nucleotide
insertion (Table 2.1 ).
Table 2.1 Amino acid sequences of the loop regions of randomly selected FynSH3 mutants (RT-
loop library). Single amino acid codes are used according to standard IUPAC nomenclature.
Numbering is according to the Fyn kinase sequence reported by Semba et al. (101) and Kawakami et
al. (100). For the sake of completeness, the sequence of the n-Src loop is also shown.
Clone sequence RT-loop94-99
Fyn SH3 wt E A R T E D
RRT 1 S P M H I L
RRT 2 G Q L M L T
RRT 3 A L T S G P
RRT 4 C R P A K P
RRT 5 N N A A N K
RRT 6 S Q R R G A
RRT 7 K Q L P E V
RRT 8 L V E L H S
RRT 9 Q P P Q S A
RRT 10 H L M D S P
RRT 11 M S F L P P
RRT 12 R H L K V S
RRT 13 K S W V P T
RRT14* N H K M Q A
RRT 15 V R Y C V L
sequence n-Src-loop113-116
N S S E
N S S E
N S S E
N S S E
N S S E
N S S E
N S S E
N S S E
N S S E
N S S E
N S s E
N S s E
N S s E
N S s E
N S s E
N S s E
single nucleotide insertion (frame shift)
71
2.3 Generation of Fyn SH3 derived proteins binding to mouse serum
albumin (MSA)
2.3.1 Phage display selections of Fyn SH3 mutants binding to MSA
In order to evaluate the above described phage library (RT-loop randomized), commercially
available mouse serum albumin (MSA) coated on immunotubcs was used as a model target
protein for phage display selections. After three rounds of panning, 13 clones out of 59 (22%)
gave a positive signal in phage ELISA (Fig. 2.6). Sequence analysis of all 13 clones revealed
that 2 unique sequences were enriched (termed G4 and C4) (Table 2.2). In both sequences, 4
amino acid residues in the artificial RT-loop were conserved, suggesting that they are mainly
involved in the binding. Interestingly, the two glycine residues were encoded by different
codons (data not shown), indicating a selection pressure for the observed sequence.
1.6
1.2a>u
c(0
S 0.8oWn
«0.4
Clones
Figure 2.6 Phage ELISA after the third round of panning. Phage ELISA signals of 59 clones are
shown, 13 clones gave a positive signal.
72
Table 2.2 Isolated unique binding sequences after 3 rounds of panning against MSA. The amino
acid sequences of the loop regions are shown (for the sake of completeness, the sequence of the n-Src
loop is also shown). Interestingly, 4 amino acids are conserved in both sequences.
Clone sequence RT-loop sequence n-Src-loop94-99 113-116
Fyn SH3 wt EARTED NSSE
G4 YSAGFG NSSE
C4 RTAGFG NSSE
2.3.2 Affinity maturation of G4 using phage display
As shown in Figure 2.6, the binding of Fyn SH3 variants to MSA could be detected by phage
ELISA, but ELISA experiments performed with the soluble protein G4 (after subcloning the
sequence into the cytosolic expression vector pQE-12) failed to yield signals even at
concentrations > 150 liM (shown below). For this reason, we aimed at affinity maturing the
clone G4, by randomizing amino acid residues in the vicinity of the RT-loop. Starting from
the sequence coding for G4, a sublibrary was created (Sublibl) in which the n-Src loop was
extended from 4 to 6 randomized amino acid residues (see Fig. 2.7a, (positions 113-116, plus
2 residues of extension)), since the expression tests (chapter 2.1.2) showed that the n-Src loop
allows for randomizations. In addition, a second sublibrary was created (Sublib2) based on
the G4 sequence, in which the n-Src loop was randomized in 4 positions (113-116), and 2
other residues of the Fyn SH3 binding surface (Trpll9 and Tyrl32) (Fig. 2.7b). These two
residues were shown in literature to be also involved in protein-protein interactions (see Ref
(94)). As before, the randomizations were introduced by PCR using partially degenerate
primers. The resulting fragments were assembled by PCR, amplified, and cloned into the
phagemid vector pHEN 1. The cloning strategy for the construction of the affinity maturation
libraries is outlined in Figure 2.7 c. Two and a half u^g insert and 13.5u,g vector per sublibrary
were ligated and electroporated into freshly prepared electrocompetent TGI bacteria. The
obtained library sizes for Sublibl and Sublib2 were 2.0 x 107 and 4.0 x 107, respectively
(theoretical size: 6.4 x 107). The quality and functionality of the libraries were assessed by
73
PCR colony screening and dot blot. PCR colony screening showed that all of 10 randomly
picked clones (for each sublibrary) contained an insert of the correct size and a dot blot
analysis of both libraries revealed that ~ 62% of the analyzed library members expressed
soluble Fyn SH3 mutants (Fyn SH3m). Both affinity maturation libraries were subjected
separately to one round of phage panning, using MSA as target protein, that was previously
coated on immunotubes. Additionally, both sublibraries were mixed and submitted for
panning in a third immunotube. Elution conditions of phage in this immunotube were chosen
to favor binders with a long koff value, by adding a solution containing soluble MSA as a
competitor after the washing step. Subsequently, the solution containing MSA and binders
with a fast koff value was discarded and bound phage were eluted. After one round of affinity
selection, clones from all 3 affinity selection tubes were screened for binding by phage
ELISA. Several clones gave at least a twofold higher signal than the signal of the parental
clone G4 (summarized in Figure 2.8), suggesting that the affinity maturation strategy was
successful. In addition, it was observed that Sublibl worked better than Sublib2, since only
one clone isolated from Sublib2 gave a twofold higher signal than G4. Moreover, the ELISA
signals of the clones isolated from the affinity selection round with soluble competing MSA
were much higher as compared to the other two standard elution conditions, with single
colonies having a nine-fold higher value of the signal.
a b
p*e
Sublib 1
s* i T
Sublib 2
jy
1 PCH»»n*ty2 «mpliftcaiiofl
afnbw3 ctoningintopHENI
l\ -| niyctog |i|
pHEN1
colE1 on]—J Amp |—J M13
74
Figure 2.7 Affinity maturation libraries of G4. (a) The structure of Fyn SH3 is shown with its
randomized and extended region of the n-Src loop in red, giving raise to Sublib 1. The G4-coding
sequence in the RT-loop is kept constant, (b) The same as (a), except that in Sublib 2 the randomized
positions comprise the n-Src loop (not extended) and, in addition, the positions Trpl 19 and Tyrl32
arc randomized (all in red) (c) Library cloning strategy: (left) Mutations were introduced either in the
n-Src loop (Sublib 1) or (right) in the n-Src and at the positions Trpl 19 and Tyrl32 (Sublib2) by PCR
using partially degenerate primers. PCR fragments were then assembled by PCR and cloned into the
pHENl vector.
0Ü
(0n
2
1.5
1
ow 0.5Q
< n
Phage ELISA Sublibl
it i,i,iin
Clones
Phage ELISA Sublib2
2
c 1-5(U
£ 1o
S 0.5
<o JlOJL-^-aL ill 1—1 Il J.U1
^r
O
Clones
75
Phage ELISA Sublib 1 & 2 "koff"
oo
CO.n
o</>n
<
3
2
1.5
1
0.5
0
Clones
Figure 2.8 Phage ELISA screening after one round of panning with affinity maturation libraries
Sublibl and Sublib2. (a) Phage ELISA signals of different clones from the affinity maturation
Sublib 1 are shown. Nine clones gave at least a twofold higher signal than the parental clone G4. (b)
The same is shown for Sublib2, revealing that only one clone was significantly better binding to
MSA. The signal produced by G4 is also indicated, (c) Phage ELISA signals from the panning
method favoring longer koff values are shown. In all three figures „-„ means negative control (PBS
was added to the wells instead of phage).
2.3.3 Characterization of affinity matured MSA binders isolated by phage display
Six affinity matured Fyn SH3 variants (Bl, Ell, C6, C3, H9 andD12), which gave at least a
twofold higher signal than G4 in phage ELISA, were sequenced (Table 2.3) and subcloned
into the cytosolic expression vector pQE-12. Five out of 6 Fyn SH3 variants had a histidine
residue in position 3, other conserved residues were a proline in position 4 and a threonine in
position 6. The clones were expressed and purified by affinity chromatography. The purity of
the proteins was assessed by SDS-PAGE (Fig. 2.9). Expression yields varied between 0.4 and
42.2 mg/L of bacterial culture in shake flasks (Table 2.3). Using the purified proteins, the
binding specificities of the clones was examined by ELISA. Mouse serum albumin (MSA),
human serum albumin (HSA), rat serum albumin (RSA), bovine serum albumin (BSA),
ovalbumin (OVA) were coated on plastic, and after addition of the purified Fyn SH3 mutants,
76
bound proteins were detected with anti-His tag HRP antibody conjugate (Fig. 2.10). All
affinity matured clones clearly bound to MSA. Two of them (Bl and Ell) were very specific
and did not cross-react with any albumin of other species, whereas C3 bound to all tested
albumins. Binding to MSA of the parental clone G4 used for affinity maturation was not
detected, however, a weak signal was observed towards OVA. Fyn SH3 wt did not bind to
any of the tested target proteins.
wt Q4H9D12 E11 C3 B1 C6
10kDa%
Figure 2.9 SDS-PAGE of Fyn SH3 variants binding to MSA.
<DO
<0
£ 2o(0
<
0
CD
MSA
IHSA
RSA
iBSA
i Ovalbumin
B ÜM
T— CD CO O) CM tJ-
UJÜ Ü x
5 o 5CO
CO
Figure 2.10 Specificity ELISA of Fyn SH3 variants isolated after affinity selections.
77
Table 2.3 Sequences of MSA binding clones after affinity maturation selections. The amino acid
sequences of the loop regions are shown, including Fyn SH3 wt and G4. Expression yields are also
indicated (expression under non-optimized conditions in shake flasks).
Clone RT-loop94-99
n-Src-loop113-116
Expression
yield (mg/L)
Fyn SH3 wt E ARTE D N S S E - 60.0
G4 (parental) Y S A G F G N S S E - - 58.0
Bl Y S A G F G R P H P F T 0.4
Ell Y S A G F G S I H P N T 12.3
C6 Y S A G F G R P H R L A 5.3
C3 Y S A G F G C K H P M S 18.4
H9 Y S A G F G N V H Y K T 6.8
D12 Y S A G F G G K S P L T 42.2
The binding properties of the only specific binder with reasonable expression yields (Ell)
was further characterized by real-time interaction analysis on a MSA coated BIAcore chip
revealing an apparent dissociation constant (Kd) of 87 nM. The sensogram is shown in Fig.
2.11.
360 -.
^^ 3203
280'—'
?40<J>
3
200-
160-
£• 120-
S 80-
J3 40-
< 0-
-40-
-50
E11
—4.4/vM
2.2/vM
50 100 150 200
Time (s)
250 300 350
Figure 2.11 BIAcore sensogram of the MSA binding clone Ell. El 1 was injected at two different
concentrations on a MSA coated chip.
78
2.4 Construction of a novel Fyn SH3 phage library
2.4.1 Phage library design, cloning and characterization
The results described in the previous chapters showed that the isolation of Fyn SH3 variants
binding to MSA with an apparent K» of- 100 nM was possible. However, these results were
achieved from a „RT-loop-library" via an affinity maturation step, resulting in Fyn SH3
variants with mutated residues in the RT- and n-Src loops. In order to create a new, large and
highly diverse library of Fyn SH3 mutants, residues in the RT- and n-Src loops of the
molecule were combinatorially mutated simultaneously (Fig. 2.12a). For library construction
(Fig. 2.12b), the randomized gene fragment of library 3 (chapter 2.1.2) was amplified by
PCR, thereby introducing appropriate restriction sites. A total of 10 jag of insert was ligated
into 75 |ag of double digested phagemid vector pHENl and electroporatcd into freshly
prepared clcctrocompetcnt E.coli TGI cells, resulting in a library containing a total of 1.2 x
109 individual clones.
a b
<_2I_ fyn bHi ^fc'
«_
1 iinrpîff^alï. i
L! c Mtnq «ID tiMfcNt
wi&et
Nico?laiton
«Ott f«J M tryf tag t< gene ptl. i~^
MkN-
t»>fc ^ Of U—* Amp " MO
Figure 2.12 Fyn SM3 wt - library cloning strategy, (a) Structure of the FynSH3 domain (pdb file
1M27 prepared with the software "Pymol" (www.pymol.org)). The RT-loop is colored in green and
the Sre-loop in red. (b) Library cloning strategy: mutations were introduced in the RT- and n-src loop
regions by PCR using partially degenerate primers. PCR fragments were then assembled by PCR and
cloned into the pHhNl vector (262).
79
The quality and the functionality of the library were assessed by PCR colony screening, dot
blot and DNA sequencing. PCR colony screening showed that 30/30 randomly picked clones
from the library contained an insert of the correct size of approximately 400 bp (Fig. 2.13a).
A dot blot experiment revealed that more than 70% of the analyzed library clones expressed a
detectable amount of soluble Fyn SH3 mutants (Fig. 2.13b). Sequence analysis of 8 clones
showed that all inserts were in frame and that mutations were introduced in the correct region
of the sequence (Tabic 2.4).
M123456789 10 BirA-
600bp
400bp
200bp
M 11 121314151617 181920 BirA-
600bp
400bp
200bp
M 21 22 23 24 25 26 27 28 29 30 BirA -
600bp
400bp
200bp
80
P^
H :ÉI 'Hl • El*m W "HP ÜB w
w »
Figure 2.13 Characterization of the library, (a) PCR colony screening of 30 clones. All the tested
clones showed an insert with the correct size of approximately 400 bp. As control, a BirA insert (1200
bp) of a pHENl vector was amplified, (b) Dot Blot analysis of 58 induced supematants of individual
library clones. The soluble Fyn SH3 mutants were detected with the anti-myc-HRP antibody
conjugate. More than 70% of the clones express a detectable amount of soluble protein. The positive
control represents Fyn SH3 wt.
Table 2.4 Amino acid sequences in the loop regions of randomly selected sequences of FynSH3
mutants (RT / n-Src-loop library).
Clone sequence
94-9S
RT-loop sequence n
113-116
-Src-loop
Fyn SH3 wt E A R T E D N S S E
clone 1 T L Q L H C T T Y A Q I
clone 2 W D S Y D A P T K Y S S
clone 3 Y K P V T C P F S Q V F
clone 4 W R L R G D L M S C N T
clone 5 A R R G H T E M G H A W
clone 6 C L D I Q H D S P S A N
clone 7 L R K E L Q T E S D K L
clone 8 F Q V S A V A L G T Y Q
81
2.5 Generation of Fyn SH3 derived proteins binding to the extra-domain-
B of fibronectin (EDB)
2.5.1 Isolation and in vitro characterization of Fyn SH3 mutants specific to the EDB of
fibronectin, a marker of angiogenesis
Specific Fyn SH3 binding proteins to the extra-domain B (EDB) of fibronectin were isolated
from the novel Fyn SH3 phage display library described in chapter 2.4. The target protein
was biotinylated using sulfo-NHS-SS-biotin as biotinylation reagent prior to selections. The
use of biotinylated target protein increases the probability that the domain retains its native
conformation upon immobilisation on streptavidin-coated wells, compared to unspecific
absorption on plastic. Three rounds of panning with EDB as target and with about 1012 phage
particles were performed. The phage titers after each round were 7 x 106 (first round), 2 x 106
(second round) and 1 x 107 (third round). After the third round of panning, two unique clones
were enriched, named Bll and D3, giving a positive signal in a phage ELISA screening
experiment. The amino acid sequences of the loop regions arc shown in Table 2.5. Both
clones were further characterized and subcloned into a cytosolic expression vector, expressed
and purified (Fig. 2.14a). Both variants gave a strong binding signal (Fig. 2.14b) in a soluble
ELISA experiment, using the purified proteins as primary binding reagents.
M wt D3 B11
4
Û)
Hi g 3
HÜ -O 11111 Bill «ED-B2
1
10kDa ^êêê^^^^^^^^^^ o
.a
IMPBS
Figure 2.14 SDS PAGE and ELISA of Fyn SH3 binding proteins to EDB. (a) SDS-PAGE analyis
of Fyn SH3 wt, D3 and Bll is shown, (b) Using the purified proteins, a soluble ELISA was
performed against biotinylated EDB (without His tag) with anti-His HRP immunconjugate as
secondary binding reagent. (MPBS= no antigen, wells were only blocked with milk, PBS)
82
D3 was chosen for further characterization and cloned as a genetically fused homo-dimer in
order to increase the apparent affinity by avidity effects. The cloning strategy is depicted in
Fig. 2.15. The expression yields of all constructs ranged between 7 and 94 mg/L of bacterial
culture, under non-optimized conditions in shake flasks (Table 2.5).
BamHI
D3
D3
%
4>
^%
D3
1.PCR
2. PCR assembly
linkerD3
3. cloning into pQE-12
Figure 2.15 Cloning strategy for D3 dimer. The gene coding for D3 was used for two independent
PCRs. The fragments were assembled by PCR, yielding a D3 dimer with BamHl/Hindlll restriction
sites with a 14 amino acid linker (GGGGSGGGGSGGGG) between the two domains and a Myc and
hexa-His tag at the C terminal-end of the protein.
Table 2.5 Sequences of EDB binding clones. The amino acid sequences of the loop regions are
shown, including Fyn SH3 wt. Expression yields are also indicated (expression under non-optimized
conditions in shake flasks).
Clone RT-loop94-99
n-Src-loop113-116
Expression
yield (mg/L)
Fyn SH3 wt E A R T E D N S S E - - 60
D3 H A Q S G A K F G R G K 50
Bll R A Y H S T R F G R F K 7
D3 dimer as monomer as monomer 94
83
The purified constructs were analyzed by size exclusion chromatography (Fig. 2.16c). The
major peak showed a delayed elution time for D3 monomer compared to Fyn SH3 wt. Also,
the D3 dimer eluted later than expected, thus suggesting a smaller size. Therefore, to ensure
the correct molecular mass of the Fyn SH3 mutants, the chromatographic fractions under the
peak were submitted to SDS-PAGE (Fig. 2.16a, lanes 3 and 5), confirming the identity of the
protein in the chromatographic peak. In addition, the oligomeric state of D3 mutants
(monomer and dimer) was determined by dynamic light scattering, using the Fyn SH3 wt
domains as reference (Fyn SH3 wt dimer was cloned and expressed the same way as
described for D3 dimer). As it can be seen in Figure 2.17, the Fyn SH3 variants have the
same size as the wild-type proteins, confirming that the mutants are present in a monomeric
form, like the wild-type proteins.
In order to determine the specificity of D3, an ELISA experiment was performed using other
structurally related fibronectin type III domains and MSA as target proteins. MSA was
commercially available, the other domains were available in our lab as His-tagged proteins.
Since D3 monomer also carries a His-tag, the experiment was performed with D3 dimer,
which has a His- and a Myc-tag. Binding was detected with anti-Myc HRP antibody
conjugate. D3 dimer bound EDB in a highly specific manner and did not cross-react with any
of the 5 other structurally related fibronectin type III homology repeats of tenascin-C, nor to
mouse serum albumin (Fig. 2.16b).
The binding properties of both D3 contructs were analyzed by real-time interaction analysis
on a BIAcore chip coated with biotinylated EDB (Fig. 2.16d), revealing a dissociation
constant (KD) of 8,5 x 10"8 M for the monomer and an apparent KD of 4,5 x 10"9 M for the
dimer, at the antigen surface density used. In order to confirm the ability of D3 monomer to
specifically recognize EDB in the context of the adjacent fibronectin domains, we performed
BIAcore studies probing the biomolecular interaction of D3 with the recombinant EDB-
containing fibronectin fragment 7B89 (263) (Fig. 2.16e). This analysis revealed a KD value of
1,0 x 10"7 M, which is comparable to the Kd value measured for the interaction between D3
monomer and biotinylated EDB.
84
M 1 2 3 4 5
25kDa^_
15kD
10kD
3iu
I 2
8 1
//
38-
< 30
E,c 22o
g- 14
-2
12
3- io
<a
a
< 0-
3B
S< 30
E
"c 22O33
S- 14-OCD
g 6-
-2
Fyn SH3 wt
10 15 20
Retention Volume (ml)
D3 monomer
25
10 15 20
Retention Volume (ml)
D3 dimer
25
I
10 15 20 25
Retention Volume (ml)
30
30
30
200
160
120
80
40-
0
-40
-50
g- 200
ce*
160
120
80
40
< o-
40
-50
D3 monomer
50 150
Time (s)
03 dimer
50 150
Time (s)
-2fJM
O.SjiM
250 350
1/jMO.SfiM0.25/jM
250 350
Fragment 7B89 of flbronectln
160*
S 120.0C
—2)JM
1(,M
0.5»(M
Ê 80.e
& 40. \/^~S o.
30 130
Time (s)
230 280
Figure 2.16 Fyn SH3 derived EDB binding proteins, (a) SDS PAGE of Fyn SH3 wt (lanc 1), D3
monomer after purification (lane 2) and after collecting the peak of size exclusion chromatography
(SEC) (lanc 3), D3 dimer after purfication (lanc 4) and after SEC (lane 5). (b) Specificity of the anti-
85
EDB binding protein D3 (dimer). ELISA signals are shown on different fibronectin type III homology
repeats. EDB, extradomain B of fibronectin; h Al Tnc, human domain Al of tenascin-C; mu D Tnc,
murine domain D of tenascin-C; h C-D-6, human domains C, D and 6 of tenascin-C; MSA, mouse
serum albumin; no antigen, wells only blocked with PBS, milk, (c) Size exclusion chromatograms of
Fyn SH3 wt, D3 monomer and D3 dimer on a Superdex 75 column. For every construct, the major
peak corresponds to the appropriate molecular weight of the corresponding proteins, (d) BIAcore
sensograms of D3 monomer and dimer injected at different concentrations on a EDB coated chip, (e)
BIAcore sensogram of 7B89, a fibronectin fragment containing the EDB domain, injected at different
concentrations on a D3 coated chip.
Size Distribution by Volume
20t
15
Ê 10+
i
0.1
M
1 10
Size (dum)
100 1000
Size Distribution by volume
20 t
15
10
5--
0
01 10
Size (d.nm)
100 1000
Figure 2.17 Dynamic Light Scattering, (top) Size distributions of Fyn SH3 wt (red) and D3
monomer (green) are very similar, (bottom) The same observation was made when comparing Fyn
SH3 wt dimer (red) with D3 dimer (green).
86
Furthermore, immunohistofluorescence experiments performed with cryosections of F9
tcratocarcinoma confirmed the ability of D3 monomer to stain tumor neo-vascular structures,
in a fashion similar to the one of an a CD-31 antibody used as positive control (Fig. 2.18).
The staining of vascular structures was further confirmed by immunohistochemistry, using
D3 dimcr as primary binding reagent on F9 teratocarcinoma cryosections (Fig. 2.19).
D3 anti CD31
-Ctrl -Ctrl
Figure 2.18 Immunohistofluorescence on F9 teratocarcinoma sections. Left, the fluorescent
pattern of D3, right the staining of a-CD31 antibody. As controls, no binding molecule was added.
Bar represents 50um.
87
D3 dimer
D3 dimer
Figure 2.19 Immunihostochemistry with D3 dimer on F9 teratocarcinoma sections. D3 dimer
was used as primary binding reagent, followed by biotinylated anti-myc antibody and streptavidin-
phosphatase as secondary and tertiary reagents, {top panel) Purified D3 dimer solution (17uM).
{middlepanel) D3 diluted in PBS, corresponding to a 0.17uM solution, {bottompanel) In the negative
control, no D3 dimer was added.
88
2.5.2 Biodistribution studies in tumor bearing mice
We used mice bearing s.c. grafted F9 murine teratocarcinoma as a syngeneic tumor model for
the biodistribution analysis of the performance of D3 monomer and dimer in the molecular
targeting of angiogencsis. In order to demonstrate that selective tumor uptake was a
consequence of a specific EDB recognition, wc also studied the biodistribution properties of
Fyn SH3 wt as a protein of irrelevant binding specificity in the mouse. Twenty-four hours
* 1 OS
after i.v. injection of I-labeled proteins, animals were sacrificed, organs were excised and
weighed, and radioactivity was counted. The results expressed as % injected dose per gram of
tissue (%ID/g), are summarized in Figure 2.20. Only D3 monomer and D3 dimer selectively
accumulated in the tumor, while Fyn SH3 wt did not exhibit any preferential tumor uptake.
As expected, the tumor targeting performance of the avid dimeric EDB binder was superior
compared to the one of D3 monomer. Tumonorgan ratios ranged between 3.0 and 9.9 for D3
monomer, the values for D3 dimer were between 4.0 and 21.8.
D3 monomer
0.75
9 0.5
0.25
0I i i I i I iôï £ & / <^ ^ # # £
^
89
Fyn SH3 wt monomer
0.75
Q 0.5
0.25
0I. •.. i i
0> Ç- £• A. Q) Ä-
"^ / / / f *
,Q ^ ;>*' orOr- "^ r?
J>
O 2
D3 dimer
/ <# # J*
É
Figure 2.20 Biodistribution experiments of radiolabeled proteins in F9 tumor-bearing mice. The
results are expressed as % injected dose per gram of tissue (% ID/g) ± S.E. (top) D3 monomer, 24 h
after injection, (middle) Fyn SH3 wt monomer, 24 h after injection (bottom) D3 dimer, 24 h after
injection.
90
2.6 Immunogenicity studies
Because the Fyn SH3 sequence is identical in mouse and man, we used Svl29 mice to study
the immunogenic potential of Fyn SH3 derived proteins. Mice were injected i.v. 4 times
cither with Fyn SH3 wt or D3, every third day. One day after the 4th injection, blood samples
were analyzed by ELISA for the presence of an IgG response against the injected protein.
This injection schedule typically exhibits a strong immunogenic reaction, when human
proteins are administered to mice (182). We used a human scFv antibody fragment as
positive control which was expected to be immunogenic in mice. Indeed, mice had to be
sacrificed after the third injection because mice became sick and blood samples were taken at
that time point for the scFv group. ELISA analysis of the blood samples revealed that no
murine antibodies against Fyn SH3 wt nor against D3 were detectable (ELISA signals <
0.03), but all mice had developed antibodies against the human scFv antibody fragment
(Figure 2.21).
Fyn SH3wt
4-j3-
u
Co
^ ci
oCOn
<
1 -
A
1:4
— 1:10
1:50
I 1:100
5 -§- Ctrl -Ctrl
D3 monomer
4n<D
c 3 " 1:4CO
£ 2-o 1:50w 1..Q 1 11:100
<n.
1 2 3 4 5 Hv Ctrl -Ctrl
91
scFv
+ Ctrl -Ctrl
Figure 2.21 Immunogenicity ELISA. ELISA signals represent the presence of murine antibodies
after repeated i.V. injections in mice. Top, no anti Fyn SH3 wt were detectable in 5 mice (ELISA
signal < 0.03). Middle, no anti D3 antibodies were detectable in 5 mice (ELISA signal < 0.03).
Bottom, all 4 mice had developed anti human scFv antibodies. Positive controls represent anti-tag
antibodies of the coated target proteins.
92
3 Discussion
The selection efficiency of any synthetic protein library crucially depends on the biophysical
properties of the selected protein scaffold, as well as on the library design, size, quality and
use. We have chosen the Fyn SH3 domain as a scaffold for the generation of binding proteins
because of a number of attractive features: it is expressed in bacteria at high level in soluble,
monomeric form (104), it is stable (TM: 70,5° C) (105), it does not contain any cysteine
residues, it is of human origin, with an amino acid sequence which is completely conserved
from mouse to man. Moreover, the structural basis of its binding mode has been resolved
which indicates that this particular domain, unlike other SH3 domains, does not necessarily
need a PXYP core binding motif: an important prerequisite for the isolation of binders against
a variety of different target epitopes.
Novel binding specificities can be introduced into a protein scaffold by randomizing those
residues that arc involved in a protein-protein interaction of the wild-type protein (10). The
Fyn SH3 domain contains two flexible loops (RT and n-Src loop) to interact with other
proteins, and therefore residues in this region were chosen to be randomized for the design of
3 test libraries. Protein expression was not impaired for the majority of the tested clones in
any of the 3 libraries, as shown by dot blot analysis (Fig. 2.3). In a first attempt, we designed
and constructed a Fyn SH3 phage library with a randomized RT loop. We reasoned that the
randomization of solely this loop might be sufficient to yield high-affinity binding
derivatives, since the RT loop is known to contribute a lot to SH3-mediated protein
interactions (79,84). Model affinity selection experiments using mouse serum albumin
(MSA) as target protein led to the successful isolation of two unique MSA binding clones
(G4 and C4). Although it was possible to detect binding in a phage ELISA, ELISA
experiments performed with the soluble protein of one of these clones (G4) failed to yield
binding signals. This difference reflects the moderate binding affinities of the clone G4 from
the primary phage repertoires of 107 individual clone members. It also reflects the different
sensitivities of the assays, phage ELISA being more sensitive because of the amplified
detection of the phage by virtue of the 3000 pVIII coat protein copies of the phage 'tail'
and/or greater binding avidities afforded by multivalent display on phage.
93
In order to increase the affinity of G4 for MSA, two affinity maturation strategies of G4 were
performed. Both led to the isolation of specific binding clones with higher affinities. One
clone, Ell, showed an apparent dissociation constant of 87 nM. We write apparent
dissociation constant since the oligomeric state of El 1 (and other MSA binders) has not been
determined by size exclusion chromatography (SEC) (data not shown), because the MSA
binders eluted at later time points than Fyn SH3 wt. Some clones had an elution volume of
more than 30 ml (instead of the expected 14-15 ml for Fyn SH3 wt, see Fig. 2.16c),
corresponding to an apparent molecular weight < 1 kDa. The retardation of the clones in the
column might be due to unspecific interactions with the matrix of the column, and/or simply
a general characteristic of such small domains to elute at different time points, although
having a similar size. For example, several monomeric single domain antibodies (13 kDa)
were shown to have apparent molecular weights ranging from about 10 to 22 kDa when
analyzing their size exclusion chromatograms (Figure 2 in Ref. (19)). Mutants of the
fibronectin type III domain have also been shown to elute after the wild-type protein (22).
Dynamic light scattering is an appropriate experiment to examine the size and the oligomeric
state of the MSA binding Fyn SH3 variants and other mutants (see below). Importantly, from
the MSA project we learned that it was possible to isolate Fyn SH3 mutants binding
specifically to MSA with dissociation constants in the range of -100 nM, from libraries
containing Fyn SH3 variants with a randomized RT- and n-Src loop. From these results we
concluded that selections of binding proteins from naïve libraries containing Fyn SH3
derivatives with a randomized RT- and n-Src loop should be feasible. For these reasons, a
novel library was constructed, containing more than 1 billion individual Fyn SH3 clones.
One clone (D3), exhibited a specific recognition of the pharmaceutical target EDB domain of
fibronectin and an impressive ability to selectively localize at the tumor site in mice bearing
murine F9 teratocarcinomas. The EDB binder D3 élûtes as a single peak in size exclusion
chromatography (SEC), but at later time points compared to the Fyn SH3 wt.
Chromatographic fractions were submitted to SDS PAGE (Fig. 2.16a, lanes 3 and 5) and
BlAcore analysis (data not shown), confirming the identity of the protein in the
chromatographic peak. Dynamic light scattering experiments further confirmed the correct
size and monomeric nature of the D3 mutants (Fig 2.17). As discussed above, the delayed
elution in SEC may be due to interactions with the column matrix.
94
The dissociation constant (KD) of D3 monomer towards biotinylated EDB (8,5 x 10 M) is
satisfactory, considering that the D3 clone was isolated directly from a naïve library and was
not submitted to affinity maturations. Further mutagenesis ofjudiciously selected residues in
the vicinity of the antigen binding site may yield EDB binders of even higher affinity. A
similar KD value (1,0 x 10"7 M) was observed for the interaction between D3 monomer and
non-biotinylated 7B89, a fibronectin fragment containing the EDB domain. This finding
indicates that the presence of the adjacent domains 7, 8 and 9 of EDB within the fibronectin
molecule and the absence of biotin did not impair the binding affinity of D3. Importantly, D3
specifically recognized EDB and did not cross-react with other, structurally related proteins
as shown by ELISA (Fig. 2.16 b). For practical applications, the cloning and expression of a
D3 dimer was sufficient for yielding a good binding avidity to the antigen and good in vivo
biodistribution properties in tumor-bearing mice. Twenty-four hours after injection of D3
monomer in tumor bearing mice, tumor to organ ratios ranged between 3.0 and 9.9, whereas
the Fyn SH3 wt did not accumulate in the tumor (Figure 2.20). At the same time point, these
ratios for D3 dimer ranged between 4.0 and 21.8. The absolute tumor value of D3 dimer
reached 2,6% ID/g; a similar value (3.2% ID/g) was obtained in the same mouse model by
the anti-EDB human antibody fragment scFv(L19), in the non-covalent homodimcric format
(Table Vin(177)).
Immunogenicity of protein drugs should be carefully assessed for all the scaffolds intended
for therapy, as the final molecule has a framework with engineered regions, thus potentially
introducing novel B- and T-cell epitopes. Ultimately, the immunogenic potential of a protein
for therapeutic applications can only be studied in the clinical setting, because of
immunological differences among animal species. Because the Fyn SH3 sequence is identical
in mouse and man, we investigated the immunogenic potential Fyn SH3 wt, D3 and of a
human scFv antibody fragment in mice. After repeated i.v. injections no antibodies against
Fyn SH3 and D3 could be detected (ELISA signals < 0.03), while mice treated with scFv
exhibited a strong antibody reaction against the human protein (Fig. 2.21). We have chosen
an injection schedule, which in our experience typically exhibits a strong immunogenic
reaction, when human proteins arc administered to mice (182). The complete identity of Fyn
SH3 in mouse, rat, gibbon and man, encourages us to pursue clinical developments of D3
derivatives for the targeted delivery of bioactive agents to the tumor neo-vasculaturc of
95
patients with cancer, in full analogy to derivatives of the LI9 antibody, which are currently
being investigated in clinical trials.
The data presented in this thesis open the way on one hand the development of D3
derivatives as discussed above (e.g., D3 cytokine fusion proteins) and on other hand the use
of the Fyn SH3 domain as a scaffold for the generation of new binding proteins, in general. In
future, it will be important to perform selections against a variety of different target proteins
to prove that this single-pot library may provide useful reagents for many biochemical and
biomedical applications. Recently, phage ELISA positive clones have been obtained for
another 2 different target proteins, glutathione S transferase and tumor necrosis factor alpha
{unpublished data). These findings further suggest that the Fyn SH3 domain might be well
suited for the isolation of binders against many different target proteins. Targets containing
an exposed peptide stretch (e.g., extracellular loops of transmembrane proteins) might
represent particularly attractive targets, since SH3 domains are known to bind well to
(proline-rich) peptidic stretches of their corresponding natural target proteins (84).
Several other binding proteins based on scaffolds described in the literature (e.g., ankyrin
repeat proteins, afiibody molecules, kunitz type domains etc.) have dissociation constants in
the low picomolar range towards their targets (see Introduction). In addition, the biophysical
properties of some derivatives have been described to be very favorable (e.g., ankyrin repeat
proteins and avimcrs) (29,65). Proving that such high affinities and good biophysical
properties can be achieved also with the Fyn SH3 domain is a very important goal. Until
now, we have been able to isolate binding proteins exhibiting dissociation constants between
26 nM (MSA-binder isolated by Bertschinger et al. (146) using covalent DNA display and
the affinity maturation Sublibl described in this thesis) and 85 nM with clone D3 against
EDB. Applying different affinity maturation approaches of selected Fyn SH3 variants, such
as randomizing residues in the vicinity of the loops, error-prone PCR, randomization of
certain residues within the loops and the use of more stringent selection protocols will show
whether higher affinities can be achieved. If needed, the biophysical properties of Fyn SH3
clinical candidates can be engineered as well, e.g., in full analogy to what Jespers et al. (19)
have impressively shown for domain antibodies: randomization of residues in the CDRs can
yield aggregation resistant domain antibodies using phage display technology in combination
with heat denaturation steps. The same strategy - randomizing selected residues of the Fyn
96
SH3 variant and performing selection strategies for the desired effect- might also work for
the Fyn SH3 domain.
Overall, the presented data in this thesis encourage us to pursue the development of this
promising domain as a scaffold for the generation of new binding proteins, thus providing
useful reagents for many biochemical and therapeutic applications as an alternative to more
conventional IgG-based immunochemical technologies.
97
4 Materials and methods
Fyn SH3 wt cloning, expression andpurification
The gene encoding the Fyn SH3 (amino acid residues are numbered according to the
sequence reported by (101) and (100)) domain was amplified from a human cDNA library
[human fetal MTC panel, brain (#K1425-1) (AMS Biotechnology (Europe) Ltd.,
Switzerland)] using the primers hFynSH3 domain ba [5'-AT CGC GGA TCC GGA GTG
ACA CTC TTT GTG GCC CTT TAT-3'] and hFynSFD domain fo [5'-GA AGA TCT CTG
GAT AGA GTC AAC TGG AGC CAC ATA-3'] (all primers were purchased from Operon
Biotechnologies). The resulting DNA fragment was cloned in the vector pQE-12 (Qiagen)
using the restriction sites BamHI and Bglll (New England Biolabs). The correct insertion was
verified by DNA sequencing (Big Dye Terminator v3.1 Cycle Sequencing kit; ABI PRISM
3130 Genetic Analyzer; Applied Biosystems). After transformation of TGI cells, colonies of
the constructs were inoculated in 100 ml 2xYT medium containing 100 ng/ml ampicillin and
0.1 % (w/v) glucose and grown at 37°C in a rotary shaker at 200 r.p.m. at an OD (600 nm) of
0.6, protein expression was induced by the addition of 1 mM IPTG (Applichem, Germany).
After 16 hours at 30°C in a rotary shaker (200 r.p.m,), the bacterial cells were harvested by
centrifugation and resuspended in 4 ml lysis buffer (50 mM NaH2P04, 300 mM NaCl, 10
mM imidazole, pH = 8.0). 1 mg/ml lysozyme was added and the cells were incubated on ice
for 30 min. After cell lysis by sonication, the lysates were centrifuged for 25 min at 11'000
r.p.m. 1 ml of Ni -NTA slurry (Qiagen) was added to the cleared lysate to capture the 6xHis
tagged proteins (incubation at 4°C for 1 h while shaking on a rotary shaker). The resin was
washed 2 times with 5 ml wash buffer (50 mM NaH2P04, 300 mM NaCl, 20 mM imidazole,
pH = 8.0) and the protein was eluted with 2 ml elution buffer (50 mM NaH2P04, 300 mM
NaCl, 250 mM imidazole, pH = 8.0). After elution, proteins were dialyzed against PBS. SDS
PAGE (Invitrogen) analysis was performed with 20ul of protein solution.
98
Cloning ofthe test libraries and expression tests (dot blot)
Library cloning strategy is depicted in Fig. 2.2. The gene encoding Fyn SH3 wt was used as a
template for all three libraries.
First library (RT loop randomized):
Two independent PCRs were performed. The first was done with the primer EDBHis ba [5'-
GAA AAG TGC CAC CTG ACG TCT AA-3'] and the partially degenerate primer FYN-
SH3-LOOPRT fo [5'-TTC TCC TTT GTG AAA ACT CAG GTC MNN MNN MNN MNN
MNN MNN ATA GTC ATA AAG GGC CAC AAA GAG-3'], the second with the primer
FYN-SH3-LOOPRT ba [5'-GAC CTG AGT TTT CAC AAA GGA GAA-3'] and EDBHis fo
[5'-CGGTCTGGTTATAGGTACA TTGAGC-3']. Full-length FynSH3 mutants were
assembled by PCR.
Second library (n-Src loop extended and randomized):
Two independent PCRs were performed. The first was done with the primer EDBHis ba [5'-
GAA AAG TGC CAC CTG ACG TCT AA-3'] and the partially degenerate primer FYN-
SH3LOOP2ND fo [5'-GGA GCG GGC TTC CCA CCA ATC TCC MNN MNN MNN
MNN MNN MNN CAA TAT TTG AAA TTT TTC TCC TTT-3'], the second with the
primer FYN-SH3-LOOP2ND ba [5'-GGAGATTGGTGGGAAGCCCGCT-3'] and EDBHis
fo [5'-CGGTCTGGTTATAGGTACA TTGAGC-3']. Full-length FynSH3 mutants were
assembled by PCR.
Third library (RT and n-Src loop randomized, the latter extended):
Three independent PCRs were performed. The first was done with the primer EDBHis ba [5'-
GAA AAG TGC CAC CTG ACG TCT AA-3'] and the partially degenerate primer FYN-
SH3-LOOPRT fo [5'-TTC TCC TTT GTG AAA ACT CAG GTC MNN MNN MNN MNN
MNN MNN ATA GTC ATA AAG GGC CAC AAA GAG-3'], the second with the primer
FYN-SH3-LOOPRT ba [5'-GAC CTG AGT TTT CAC AAA GGA GAA-3'] and the
partially degenerate primer FYN-SH3LOOP2ND fo [5'-GGA GCG GGC TTC CCA CCA
ATC TCC MNN MNN MNN MNN MNN MNN CAA TAT TTG AAA TTT TTC TCC
TTT-3'], whereas the third was done with the primer FYN-SH3-LOOP2ND ba [5'-
GGAGATTGGTGGGAAGCCCGCT-3'] and EDBHis fo [5'-
CGGTCTGGTTATAGGTACA TTGAGC-3']. Full-length FynSH3 mutants were assembled
by PCR.
99
The randomized inserts of all three libraries were independently double digested with
EcoRI/Bglll (New England Biolabs) and cloned into EcoRI/Bglll -digested pQE-12 vector.
The resulting ligation products were clcctroporated into electrocompetent Escherichia coli
TGI cells.
The percentage of clones expressing soluble Fyn SH3 variants was determined by dot blot
analysis of bacterial cell lysates (ELIFA system; Perbio) using anti-His-HRP antibody
conjugate (Roche) as detecting agent. Peroxidase activity was detected using the ECL plus
Western blotting detection system (Amersham Biosciences).
Bacterial cell lysates were prepared the following way: Transformed bacteria of all three test
libraries (see above) were plated on agar plates containing 100 ^ig/ml ampicillin and 1%
(w/v) glucose and grown overnight in an incubator at 37°C. The next day, colonies were
picked from the agar plate and grown in a round bottom 96-well plate (Nunc, cat. no.
163320) in 200 (xl 2xYT medium containing 100 ug/ml ampicillin and 0.1% (w/v) glucose.
Protein expression was induced after growth for 3 hours at 37°C and 200 r.p.m. by adding 1
mM IPTG (Applichem, Germany). Proteins were expressed overnight in a rotary shaker (200
r.p.m., 30°C). Subsequently, the 96-well plate was centrifugea at 1800g for 10 min and the
supernatant was discarded. The bacterial pellets were resuspended in 60 ^il lysis buffer (50
mM NaH2P04, 300 mM NaCl, 10 mM imidazole, pH = 8.0) containing 1 mg/ml lysozyme
and left for 30 min on ice. Afterwards, the bacterial cells were lysed by sonication in a water
bath (6 bursts à 10 sec) and then centrifuged at 1800g for 10 min. The supernatant (1:4
diluted in PBS) was used for the dot blot experiment.
Construction ofthe RT-loop library
The first test library described in the previous section was amplified using SH3(NcoI) ba [5'-
CAT GCC ATG GGC GGA GTG ACA CTC TTT GTG GCC CTT TAT-3'] and SH3(NotI)
fo [5'-TTT TCC TTT TGC GGC CGC CTG GAT AGA GTC AAC TGG AGC CAC ATA-
3']. The randomized insert was double digested with NcoI/NotI (New England Biolabs) and
cloned into NcoI/Notl-digested pHENl phagemid vector (262). The resulting ligation product
was electroporated into electrocompetent Escherichia coli TGI cells according to Viti et al.
(108), thereby obtaining a library of 1,2 x 107 Fyn SH3 mutant clones. The library was stored
100
as glycerol stocks, rescued and used for phage production according to standard protocols
(108).
RT-loop library characterization
A total of 10 clones were tested by PCR using the primers LMB31ong ba [5'-CAG GAA
ACA GCT ATG ACC ATG ATT AC-3'] and FDseqlong fo [5'-GAC GTT AGT AAA TGA
ATT TTC TGT ATG AGG-3'] to verify the correct size of the insert. 15 clones were selected
at random and sequenced as described before. The percentage of clones expressing soluble
Fyn SH3 mutants was determined by dot blot analysis of bacterial supernatants (ELIFA
system; Perbio) using anti-myc-HRP antibody conjugate (Roche) as detecting agent.
Peroxidase activity was detected using the ECL plus Western blotting detection system
(Amersham Biosciences). The display of the Fyn SH3 mutant - pill fusion protein on the
phage surface was evaluated by western blot. Different amounts of purified phage were
loaded on SDS gel and then transferred to NC membrane (Protran BA 85; Schleicher &
Schuel, Dassel, Germany). As detecting reagents anti-pill mouse mAB (MoBiTec, Göttingen,
Germany) and anti-mouse HRP immunoglobulins were used. Peroxidase activity was
detected using the ECL plus Western blotting detection system (Amersham Biosciences).
Phage display selections (RTloop library) against MSA andELISA screening
Unless stated otherwise, growth media, helper phage, and general protocols were essentially
used as the protocols for antibody phage display described in (108).
An immunotube (Nunc) was coated with MSA (Sigma) at a concentration of 100ug/ml in
PBS, overnight at room temperature. The immunotube was then rinsed with PBS and blocked
for 2 h at room temperature with 2% skimmed milk in PBS (MPBS 2%). After rinsing with
PBS, >1012 phage particles in MPBS 2% were added to the immunotube. The immunotube
was first incubated on a shaker for 30 min and then for 1.5 h standing upright at room
temperature. Unbound phage were washed away by rinsing the immunotube ten times with
PBS 0.1% Tween 20 and ten times with PBS. The bound phage was eluted specifically with
101
1% MSA solution in PBS for 30 min. The eluted phage was used to infect exponentially
growing E.coli TGI. Phage production was performed according to standard protocols (108).
After three rounds of panning, monoclonal phage ELISA was performed essentially as
described (264). Individual colonies from the plates of phage-infected TGI were inoculated
into 200ul 2xYT-, lOOug/ml ampicillin (Applichem), 1% Glucose (Sigma) (2xYT-AMP-
GLU) into 96-well plates (Nunclon Surface, Nunc). The plate was incubated for 3 h at
37°C in a shaker incubator. After removing 40ul of each well for a glycerol stock, 40ul
2xYT-AMP-GLU containing 4xl09 t.u. of helper phage were added to each well and the plate
was incubated at 37°C for 30 min. After spinning the plate at 1800g for 10 min and aspirating
the supernatants off, the bacterial pellet was resuspended in 200ul of 2xYT, 100ug/ml
ampicillin, 33,3ug/ml kanamycin (Applichem) and grown overnight at 30°C. After spinning,
phage supernatants were used for ELISA screening using biotinylated MSA as target. MSA
was biotinylated using EZ-Link Sulfo-NHS-SS-Biotin (Perbio) according to
manufacturer's instructions and was captured (concentration: ~3 x 10"7 M) on streptavidin
coated wells (StreptaWells, High Bind, Roche). After blocking with PBS, 2% Milk (Rapilait,
Migros, Switzerland), 20 ul PBS, 10% Milk and 80 ul of phage supernatants were applied.
After incubating for 1 hour and washing, detection was made with anti-M13-HRP antibody
conjugate (GE Healthcare). Peroxidase activity was detected by adding BM blue POD
substrate (Roche) and the reaction was stopped adding 1 M H2SO4. Absorbance was
measured using the VersaMax microplate reader (OD45onm - ODesonm)- The DNA of positive
clones was sequenced as described above.
Construction and characterization ofthe affinity maturation libraries Suhlibl and Sublib2
One MSA binding clone G4 was affinity matured using its sequence and creating 2
sublibrarics (Sublibl and Sublib2). Out of the glycerol stock of G4 described above, a single
colony was picked for the PCR-amplification of the G4 sequence using the primers
LMB31ong ba [5'-CAG GAA ACA GCT ATG ACC ATG ATT AC-3'] and FDseqlong fo
[5'-GAC GTT AGT AAA TGA ATT TTC TGT ATG AGG-3'], resulting in the template for
the libraries.
102
Sublibl:
Two independent PCRs were performed. The first was done with the primer LMB31ong ba
[5'-CAG GAA ACA GCT ATG ACC ATG ATT AC-3'] and the partially degenerate primer
FYN-SH3LOOP2ND fo [5'-GGA GCG GGC TTC CCA CCA ATC TCC MNN MNN MNN
MNN MNN MNN CAA TAT TTG AAA TTT TTC TCC TTT-3'], the second with the
primer FYN-SH3-LOOP2ND ba [5'-GGAGATTGGTGGGAAGCCCGCT-3'] and
FDseqlong fo [5'-GAC GTT AGT AAA TGA ATT TTC TGT ATG AGG-3']. Full-length
FynSH3 mutants were assembled by PCR.
Sublib2:
Two independent PCRs were performed. The first was done with the primer LMB31ong ba
[5'-CAG GAA ACA GCT ATG ACC ATG ATT AC-3'] and the partially degenerate primer
Fyn-SH3_sub_l 13-16/119/132fo [5' CC AGT TGT CAA GGA GCG GGC TTC CCA MNN
ATC TCC MNN MNN MNN MNN CAA TAT TTG AAA TTT TTC TCC 3'], the second
with the primer Fyn-SH3_sub_113-116/119/132ba [5' GCC CGC TCC TTG ACA ACT
GGA GAG ACA GGT NNK ATT CCC AGC AAT 3'] and FDseqlong fo [5'-GAC GTT
AGT AAA TGA ATT TTC TGT ATG AGG-3']. Full-length FynSH3 mutants were
assembled by PCR.
The resulting inserts of both sub-libraries were further amplified using LMB31ong ba [5'-
CAG GAA ACA GCT ATG ACC ATG ATT AC-3'] and FDseqlong fo [5'-GAC GTT AGT
AAA TGA ATT TTC TGT ATG AGG-3']. After double digestion with NcoI/NotI (New
England Biolabs) and cloning into NcoI/Notl-digested pHENl phagemid vector (262), the
resulting ligation products of both sublibraries were clectroporated into electrocompetent
Escherichia coli TGI cells according to Viti et al. (108), thereby obtaining a library of 2.0 x
107 (Sublibl) and 4.0 x 107 (Sublib2) Fyn SH3 mutant clones. The library was stored as
glycerol stocks, rescued and used for phage production according to standard protocols (108).
For the characterization of the libraries, PCR colony screening and dot blot experiments were
performed as described above under the section ,JtT-loop library characterization ".
103
Phage display selections (Sublibl and Sublib2) against MSA and ELISA screening
The affinity maturation selections and screenings were essentially performed as described in
„Phage display selections (RT loop library) against MSA and ELISA screening". For each
sublibrary, an immunotube (Nunc) was used. In addition, a third immunotube was used for
mixing the 2 sublibraries. Except for the third immunotube, bound phage was eluted in 1 ml
of lOOmM triethylamine solution and inverting the tube for 5 min. Triethylaminc was
neutralized by adding 0.5ml 1 M Tris-Hcl pH 7.4. Bound phage in the third tube were
incubated for 5 min with a luM MSA solution in PBS, washed again with PBS, 0.1% Tween
20 and PBS, and then eluted specifically with a 0.5% MSA solution on PBS (30min). The
eluted phage of all three immunotubes were used to infect exponentially growing E.coli TGI,
separately. Phage production was performed according to standard protocols (108). After
three rounds of panning, monoclonal phage ELISA was performed as described before. The
DNA of positive clones was sequenced as described above.
Subcloning, expression andpurification ofaffinity matured MSA binders
For expression of selected affinity matured MSA binding clones, their sequence was
subcloned into the pQE-12 vector (Qiagen). Out of the glycerol stocks prepared during the
ELISA screening as described above, a single colony was picked for the PCR-amplification
of the sequences using the primers LMB31ong ba [5'-CAG GAA ACA GCT ATG ACC ATG
ATT AC-3'] and FDseqlong fo [5'-GAC GTT AGT AAA TGA ATT TTC TGT ATG AGG-
3'], resulting in PCR-product that was used for another PCR using the primers hFynSH3
domain ba [5'-AT CGC GGA TCC GGA GTG ACA CTC TTT GTG GCC CTT TAT-3']
and FYN- SH3_cloning_into_pQE12 fo [5' ATC CCA AGC TTA GTG ATG GTG ATG
GTG ATG ACC CTG GAT AGA GTC AAC TGG AGC CAC 3'], introducing
BamHI/Hindlll restrictions sites and a hexa-His tag. All clones were BamHI/Hindlll (New
England Biolabs) double-digested and ligated into BamHI/Hindlll double-digested pQE-12
vector. After transformation of TGI cells, colonies of the constructs were induced for protein
expression, lysed and purified as described before for the Fyn SH3 wt protein.
104
Albumin specificity ELISA
Fifty ul of the purified proteins Bl (2 uM), Ell (65uM), C6 (28uM), C3 (lOOuM), H9
(35uM), D12 (220uM), G4 (300uM) and FynSH3 wt (37uM) were separately added together
with 50ul of MPBS 4% into the wells of a MaxiSorp plate (Nunc) that had been previously
coated with lOOul of a MSA, HSA, RSA, BSA and Ovalbumin (OVA) solution (all
purchased from Sigma; coating concentration: 100(ig/ml in PBS). After 2h at RT, the plate
was washed 3 times with PBS, followed by the addition of lOOul anti His-HRP
immunoconjugate (Sigma) diluted 1:1000 in 2% MPBS. The 96-well plate was left for 1 h at
RT and then washed 3 times with PBS, 0.1% Tween 20 and 3 times with PBS. Colorimetric
detection was made by the addition of lOOul of BM Blue POD substrate (Roche) and the
reaction was stopped with 60ul 1 M H2SO4. Absorbancc was measured using the VersaMax
microplate reader (OD45onm - OD650nm).
Surfaceplasmon resonance experiments with MSA
Affinity measurements were performed using a BIAcore3000 instrument (Biacore). For the
interaction analysis between MSA and El 1, a CM5 chip (Biacore) was used with 18680 RU
MSA immobilized. The running buffer was PBS, 0.1% NaN3, Surfactant P20 (Biacore). The
interactions were measured at a flow of 20ul/min and injections of different concentrations of
Ell. All kinetic data of the interaction (separate kon/koff) were evaluated using BIAevaluation
3.2RC1.
Construction and characterization ofthe RT- / n-Src loop library
The third test library described in the previous section was amplified using SH3(NcoI) ba [5'-
CAT GCC ATG GGC GGA GTG ACA CTC TTT GTG GCC CTT TAT-3'] and SH3(Notl)
fo [5'-TTT TCC TTT TGC GGC CGC CTG GAT AGA GTC AAC TGG AGC CAC ATA-
3']. The randomized insert was double digested with NcoI/NotI (New England Biolabs) and
cloned into NcoI/Notl-digested pHENl phagemid vector (262). The resulting ligation product
105
was electroporated into electrocompetent Escherichia coli TGI cells according to Viti et al.
(108), thereby obtaining a library of 1,2 x 109 Fyn SH3 mutant clones. The library was stored
as glycerol stocks, rescued and used for phage production according to standard protocols
(108).
The quality and functionality of this new library were assessed by DNA sequencing (8
clones) PCR colony screening (30 clones) and dot blot experiment (58 clones), performed as
described above under the section ,JlT-loop library characterization ".
Phage display selections (RT- /n-Src loop library) against EDB andELISA screening
The EDB domain of fibronectin (169), was expressed and purified as previously described.
EDB was biotinylated using EZ-Link Sulfo-NHS-SS-Biotin (Perbio) according to
manufacturer's instructions. The biotinylated EDB (final concentration 10"6 M) was captured
on avidin- (first and third round) or streptavidin-coated (second round) Maxisorp wells
(Nunc). After blocking with 1% BSA (Sigma) in PBS, 1ml containing 1012 transforming units
(t.u.) of Fyn SH3 phage library in 1% BSA were added to the wells (aliquoted in 8 wells) and
incubated for 45 minutes at room temperature on a shaker. The wells were rinsed 10 times
with PBS, 0,1% Twccn20 (Sigma) and 10 times with PBS. Bound phage was eluted by
incubation with lOOmM triethylamine (Fluka) for 5 min. The eluted phage was used to infect
exponentially growing E.coli TGI. Phage production was performed according to standard
protocols (108). After three rounds of panning, monoclonal phage ELISA was performed as
described above. In ELISA screening, biotinylated EDB (10~6 M) was captured on
streptavidin coated wells (StreptaWells, High Bind, Roche) and after blocking with PBS, 2%
Milk (Rapilait, Migros, Switzerland), 20 ul PBS, 10% Milk and 80 ul of phage supernatants
were applied. After incubating for 1 hour and washing, detection was made with anti-M13-
HRP antibody conjugate (GE Healthcare). Peroxidase activity was detected by adding BM
blue POD substrate (Roche) and the reaction was stopped adding 1 M H2S04. Absorbance
was measured using the VersaMax microplate reader (OD45onm - ODösonm)- The DNA of
positive clones was sequenced as described above.
106
Subcloning, expression andpurification ofEDB binders
For expression of Bll and D3, their sequence was subcloned into the pQE-12 vector
(Qiagen) as described before for the MSA binders.
ELISA ofpurified Bll andD3 proteins
Fifty ul of the purified proteins Bll (10 uM) and D3 (170uM), were separately added
together with 50ul of MPBS 4% into strcptavidin coated wells (StreptaWells, High Bind,
Roche) that had been coated with biotinylated EDB (without His tag) (10~6 M) or MPBS 4%.
After incubating for 1 hour and washing, detection was made with the addition of lOOul anti
His-HRP immunoconjugate (Sigma). The 96-well plate was left for 1 h at RT and then
washed 3 times with PBS, 0.1% Tween 20 and 3 times with PBS. Colorimetric detection was
made by the addition of lOOul of BM Blue POD substrate (Roche) and the reaction was
stopped with 60ul 1 M H2S04.
Cloning, expression andpurification ofD3 andFyn SH3 wt dimer
Using the D3 sequence as template (first PCR product obtained when cloning D3 into the
pQE-12 vector, see above) two independent PCRs were performed: in the first one, the
primers 52. ba [5'-GAC TAA CGA GAT CGC GGA TCC GGA GTG ACA CTC TTT GTG
GCC CTT TAT-3'] and 47. fo [5'-TCC GCC ACC GCC AGA GCC ACC TCC GCC TGA
ACC GCC TCC ACC CTG GAT AGA GTC AAC TGG AGC CAC-3'] and in the second
PCR the primers 48. ba [5'-GGT GGA GGC GGT TCA GGC GGA GGT GGC TCT GGC
GGT GGC GGA GGA GTG ACA CTC TTT GTG GCC CTT TAT-3'] and 51. fo [5'-ATC
CCA AGC TTA GTG ATG GTG ATG GTG ATG CAG ATC CTC TTC TGA GAT GAG
TTT TTG TTC ACC CTG GAT AGA GTC AAC TGG AGC CAC-3'] were used. The two
fragments were PCR assembled yielding a D3 dimer with BamHI/Hindlll restriction sites,
with a 14 amino acid linker (GGGGSGGGGSGGGG) between the two domains and a myc
and hexa-His tag at the C-terminal end of the protein. The insert was BamHI/Hindlll (New
107
England Biolabs) double-digested and ligated into BamHI/Hindlll double-digested pQE-12
vector. After transformation of TGI cells, expression and purification was made as described
before for the MSA binders.
For the cloning of Fyn SH3 wt, the same procedure was made using the Fyn SH3 wt as
template.
Specificity ELISA (EDB)
The biotinylated target proteins EDB (169), human domain Al of Tenascin C (199), murine
domain C of Tenascin C (199), human domains C-D-6 of Tenascin C were available in our
laboratory and expressed as (His)6 tagged proteins in E.coli TGI from pQE-12 based
expression vector, purified and biotinylated as previously described. Mouse serum albumin
was purchased from Sigma and biotinylated using EZ-Link Sulfo-NHS-SS-Biotin (Perbio)
according to manufacturer's instructions. Target proteins were added to streptavidin coated
wells (StreptaWells, High Bind, Roche) and after blocking with PBS, 2% Milk (Rapilait,
Migros, Switzerland), 50 ul PBS, 4% Milk and 50 ul of purified D3 dimer (3uM) was added.
After incubating for 1 hour and washing, detection was made with anti-myc-HRP antibody
conjugate (Roche). Peroxidase activity was detected by adding BM blue POD substrate
(Roche) and the reaction was stopped adding 1 M H2SO4. Absorbance was measured using
the VersaMax microplate reader (OD45onm - ODesonm)-
Size exclusion chromatography (SEC)
SEC of purified Fyn SH3 wt, D3 monomer, D3 dimer was performed on an ÄKTA FPLC
system using a Superdex 15 column (Amersham Biosciences).
Dynamic Light Scattering
Dynamic Light Scattering measurements were performed using the Zetasizcr Nano-ZS
instrument (Malvern), according to predefined settings for protein size measurements. The
108
results represent average sizes of 30 (Fyn SH3 wt and D3 monomer) and 60 (Fyn SH3 wt and
D3 dimer) independent runs, respectively.
Surface plasmon resonance experiments (EDB)
Affinity measurements were performed using a BIAcore3000 instrument (Biacore). For the
interaction analysis between biotinylated EDB and D3 monomer and D3 dimer, a streptavidin
SA chip (Biacore) was used with 600 RU biotinylated EDB (169) immobilized. The running
buffer was PBS, 0.1% NaN% Surfactant P20 (Biacore). The interactions were measured at a
flow of 20ul/min and injections of different concentrations of D3 monomer and dimer. For
the interaction analysis between D3 monomer and 7B89, a fibronectin fragment containing
the EDB domain, ((263), kindly provided by Philogen, Italy) a CM5 chip (Biacore) was used
with 3300 RU D3 monomer immobilized. The running buffer was HBS-EP (Biacore). The
interactions were measured at a flow of 20ul/min and injections of different concentrations of
7B89. All kinetic data of the interaction (separate kon/koff) were evaluated using
BIAevaluation3.2RCl.
Immunohistofluorescence
Female Svl29 mice (Charles River) received subcutaneous injection of 3 x 106 F9
teratocarcinoma cells in the right flank. After excising, the tumors were embedded in
cryoembedding compound (Microm), frozen in chilled isopentane and stored at -80°C. Then,
10|im sections were cut, fixed with ice cold acetone and double fluorescence staining for
EDB and CD31 was performed. D3 monomer and rat antimouse CD31 (BD Pharmingen)
were used as primary binding reagents. As secondary detection antibodies, we used for EDB
anti His Alexa Fluor 488 conjugate (Qiagen) and for CD31 donkey anti rat Alexa 594
(Molecular Probes). Slides were mounted with Glycergel mounting medium (Dako) and
analyzed with a Zeiss Axioskop 2 mot plus.
109
Immunohistochemistry
Immunohistochcmistry was performed using the same cryoscctions as described above. Pre-
blocked D3 dimer (17 \xM or 170 nM) in 2% BSA and biotinylated anti-Myc antibody were
used as primary binding reagents, followed by the Streptavidin-Alkaline-Phosphatase
complex (Biospa). Fast Red TR Salt (Sigma) was used to develop the red staining. Slides
were eounterstained with hematoxylin (Sigma), mounted with Glycergel mounting medium
(Dako) and analyzed with a Zeiss Axiovert S100 TV microscope.
Radioiodination ofFyn SH3 wt, D3 monomer andD3 dimer
125ug -200|ig Protein were combined with 200nCi of 125I (Amersham) and with filtered
Chloramin T (Sigma) solution (5mg/ml; 0.25^ig Chloramin T per (ig of protein was used) for
2 min followed by separation from unincorporated iodine using a PD-10 disposable gel
filtration column (Pharmacia). The binding activity after labeling was evaluated by loading
an aliquot of radiolabeled sample onto 500 ul of EDB-Sepharose resin (170) on a pasteur
pipette, followed by radioactive counting of the flow-through, wash and eluate fractions.
Binding reactivity, defined as the ratio between the counts of the cluted protein and the sum
of the counts (flow-through, wash, eluate and column) was 87% for the D3 monomer and
between 58 and 68 % for the dimer (36% for the non-binding Fyn SH3 wt).
Biodistributions oftumor-bearing mice injected with radiolabeled Fyn SH3 wt, D3 monomer
and D3 dimer
Biodistribution studies were performed under a license (Tumor targeting, Bewilligung
198/2005) issued to D.N. by the Veterinäramt des Kantons Zürich. F9 murine
teratocarcinomas were implanted as described previously (169,170) in 129Sv mice (8-12
weeks old, female). 125I-labeled protein (5.6 - 10(ig (9 - 13 jiCi)) in lOOul of saline solution,
radiolabeled on the same day, was injected i.v. Mice were sacrificed at 4 and 24 h (Fyn SH3
wt only 24 h) and organs were weighed and radioactively counted. 3 animals were used for
each time point (except Fyn SH3 wt 24h and D3 monomer 4h: 4 animals). Targeting results
110
of representative organs are expressed as %TD of protein /g of tissue (± SE) and tumonorgan
ratios were determined.
Immunogenicity ofFyn SH3 wt, D3 and a scFv antibodyfragment
Svl29 mice were injected 4 times (every third day) with 20 ug of Fyn SH3 wt or D3,
respectively (5 mice per group). One day after the 4th injection blood samples were taken for
examining the presence or absence of murine anti-Fyn SH3wt and anti-Fyn SH3 D3
antibodies (75-150ul blood per mice, diluted in 50 ul Heparin (Bichsel, Switzerland)). As a
positive control, 4 mice were injected at equal time points and equal molar dosages (= 60 jig)
with the human scFv antibody fragment clone El (anti human a2-macroglobulin described in
(153); expression and purification as described in (153)). However, because mice became
sick after the third injection of scFv fragments, they were sacrificed and blood samples were
taken after the third injection. Blood samples were analyzed in ELISA: Maxisorp wells
(Nunc) were coated overnight with lOOul of antigen solution (Fyn SH3 wt: 20ug/ml, D3:
20^ig/ml or scFv: 60ug/ml in PBS). After washing and blocking, 50 ul of PBS, 4% Milk and
50 ul of blood samples (1:4, 1:10, 1;50 and 1:100 diluted in PBS) were added and incubated
for 1 hour at room temperature. After washing, the presence of murine antibodies in the blood
samples were detected by adding anti mouse IgG HRP antibody conjugate (Sigma). As
controls for the coating efficiency of the antigens, anti His HRP antibody conjugate (Sigma)
(Fyn SH3 wt and D3) or anti-myc HRP antibody conjugate (Roche) were added. The
negative controls were performed using the anti mouse IgG HRP antibody conjugate without
adding any blood samples. Peroxidase activity was detected by adding BM blue POD
substrate (Roche) and the reaction was stopped adding 1 M H2SO4. Absorbance was
measured using the VersaMax microplate reader (OD45onm - ODesonm)-
111
References
Walsh, G. (2002) Biopharmaceuticals and biotechnology medicines: an issue of
nomenclature. Eur JPharm Sei, 15, 135-138.
Johnson, LS. (1983) Human insulin from recombinant DNA technology. Science,
219, 632-637.
Lawrence, S. (2005) Biotech drug market steadily expands. Nat Biotechnol, 23, 1466.
Pavlou, A.K. and Belsey, M.J. (2005) The therapeutic antibodies market to 2008. Eur
JPharm Biopharm, 59, 389-396.
Pavlou, A.K. and Reichert, J.M. (2004) Recombinant protein therapeutics—success
rates, market trends and values to 2010. Nat Biotechnol, 22, 1513-1519.
Köhler, G. and Milstein, C. (1975) Continuous cultures of fused cells secreting
antibody of predefined specificity. Nature, 256, 495-497.
Holliger, P. and Hudson, P.J. (2005) Engineered antibody fragments and the rise of
single domains. Nat Biotechnol, 23, 1126-1136.
Nygren, P.A. and Uhlen, M. (1997) Scaffolds for engineering novel binding sites in
proteins. Curr Opin Struct Biol, 1, 463-469.
Binz, H.K. and Pluckthun, A. (2005) Engineered proteins as specific binding reagents.
Curr Opin Biotechnol, 16, 459-469.
Smith, G. (1998) Patch engineering: a general approach for creating proteins that have
new binding activities. Trends Biochem Sei, 23, 457-460.
Tomlinson, I.M. (2004) Next-generation protein drugs. Nat Biotechnol, 22, 521-522.
112
12. Binz, H.K., Amstutz, P. and Pluckthun, A. (2005) Engineering novel binding proteins
from nonimmunoglobulin domains. Nat Biotechnol, 23, 1257-1268.
13. Hey, T., Fiedler, E., Rudolph, R. and Fiedler, M. (2005) Artificial, non-antibody
binding proteins for pharmaceutical and industrial applications. Trends Biotechnol,
23,514-522.
14. Ward, E.S., Gussow, D., Griffiths, A.D., Jones, P.T. and Winter, G. (1989) Binding
activities of a repertoire of single immunoglobulin variable domains secreted from
Escherichia coli. Nature, 341, 544-546.
15. De Genst, E., Handelberg, F., Van Meirhaeghe, A., Vynck, S., Loris, R., Wyns, L.
and Muyldcrmans, S. (2004) Chemical basis for the affinity maturation of a camel
single domain antibody. J Biol Chem, 279, 53593-53601.
16. Dooley, H. and Flajnik, M.F. (2005) Shark immunity bites back: affinity maturation
and memory response in the nurse shark, Ginglymostoma cirratum. Eur JImmunol,
35, 936-945.
17. Dottorini, T., Vaughan, CK., Walsh, M.A., LoSurdo, P. and Sollazzo, M. (2004)
Crystal structure of a human VH: requirements for maintaining a monomeric
fragment. Biochemistry, 43, 622-628.
18. Jespers, L., Schon, O., James, L.C., Veprintsev, D. and Winter, G. (2004) Crystal
structure of HEL4, a soluble, refoldable human V(H) single domain with a germ-line
scaffold. J Mol Biol, 337, 893-903.
19. Jespers, L., Schon, O., Famm, K. and Winter, G. (2004) Aggregation-resistant domain
antibodies selected on phage by heat denaturation. Nat Biotechnol, 22, 1161-1165.
20. Colby, D.W., Garg, P., Holden, T., Chao, G., Webster, J.M., Messer, A., Ingram,
V.M. and Wittrup, K.D. (2004) Development of a human light chain variable domain
(V(L)) intracellular antibody specific for the amino terminus of huntingtin via yeast
surface display. J Mol Biol, 342, 901-912.
113
21. Xu, L., Aha, P., Gu, K., Kuimelis, R.G., Kurz, M., Lam, T., Lim, A.C., Liu, H.,
Lohse, P.A., Sun, L. et al. (2002) Directed evolution of high-affinity antibody mimics
using mRNA display. Chem Biol, 9, 933-942.
22. Koide, A., Bailey, C.W., Huang, X. and Koide, S. (1998) The fibronectin type III
domain as a scaffold for novel binding proteins. J Mol Biol, 284, 1141-1151.
23. Wilson, D.S., Keefe, A.D. and Szostak, J.W. (2001) The use ofmRNA display to
select high-affinity protein-binding peptides. Proc Natl Acad Sei USA, 98, 3750-
3755.
24. Richards, J., Miller, M„ Abend, J., Koide, A., Koide, S. and Dewhurst, S. (2003)
Engineered fibronectin type III domain with a RGDWXE sequence binds with
enhanced affinity and specificity to human alphavbeta3 integrin. J Mol Biol, 326,
1475-1488.
25. Parker, M.H., Chen, Y., Danehy, F., Dufu, K., Ekstrom, J., Getmanova, E.,
Gokemeijer, J., Xu, L. and Lipovsek, D. (2005) Antibody mimics based on human
fibronectin type three domain engineered for thermostability and high-affinity binding
to vascular endothelial growth factor receptor two. Protein Eng Des Sel, 18, 435-444.
26. Getmanova, E.V., Chen, Y., Bloom, L., Gokemeijer, J., Shamah, S., Warikoo, V.,
Wang, J., Ling, V. and Sun, L. (2006) Antagonists to human and mouse vascular
endothelial growth factor receptor 2 generated by directed protein evolution in vitro.
Chem Biol, 13, 549-556.
27. Huang, W., Dolmer, K. and Gettins, P.G. (1999) NMR solution structure of
complement-like repeat CR8 from the low density lipoprotein receptor-related
protein. JBiol Chem, IIA, 14130-14136.
28. North, C.L. and Blacklow, S.C. (1999) Structural independence of ligand-binding
modules five and six of the LDL receptor. Biochemistry, 38, 3926-3935.
114
29. Silverman, J., Liu, Q., Bakker, A., To, W., Duguay, A., Alba, B.M., Smith, R., Rivas,
A., Li, P., Le, H. et al. (2005) Multivalent avimer proteins evolved by exon shuffling
of a family ofhuman receptor domains. Nat Biotechnol, 23, 1556-1561.
30. Gliemann, J. (1998) Receptors of the low density lipoprotein (LDL) receptor family
in man. Multiple functions of the large family members via interaction with complex
ligands. Biol Chem, 379, 951-964.
31. Krieger, M. and Herz, J. (1994) Structures and functions of multiligand lipoprotein
receptors: macrophage scavenger receptors and LDL receptor-related protein (LRP).
Annu Rev Biochem, 63, 601-637.
32. Koduri, V. and Blacklow, S.C. (2001) Folding determinants of LDL receptor type A
modules. Biochemistry, 40, 12801-12807.
33. Uhlen, M., Forsberg, G., Moks, T., Hartmanis, M. and Nilsson, B. (1992) Fusion
proteins in biotechnology. Curr Opin Biotechnol, 3, 363-369.
34. Nilsson, B., Moks, T., Jansson, B., Abrahmsen, L., Elmblad, A., Holmgren, E.,
Henrichson, C, Jones, T.A. and Uhlen, M. (1987) A synthetic IgG-binding domain
based on staphylococcal protein A. Protein Eng, 1, 107-113.
35. Nord, K., Nilsson, J., Nilsson, B., Uhlen, M. and Nygren, P.A. (1995) A
combinatorial library of an alpha-helical bacterial receptor domain. Protein Eng, 8,
601-608.
36. Nord, K., Gunncriusson, E., Ringdahl, J., Stahl, S., Uhlen, M. and Nygren, P.A.
(1997) Binding proteins selected from combinatorial libraries of an alpha-helical
bacterial receptor domain. Nat Biotechnol, 15, 772-777.
37. Nord, K„ Nord, O., Uhlen, M., Kelley, B., Ljungqvist, C. and Nygren, P.A. (2001 )
Recombinant human factor VHI-specific affinity ligands selected from phage-
displayed combinatorial libraries of protein A. Eur JBiochem, 268, 4269-4277.
115
38. Ronnmark, J., Gronlund, H., Uhlen, M. and Nygren, P.A. (2002) Human
immunoglobulin A (IgA)-specific ligands from combinatorial engineering of protein
A. Eur JBiochem, 269, 2647-2655.
39. Sandstrom, K., Xu, Z., Forsberg, G. and Nygren, P.A. (2003) Inhibition of the CD28-
CD80 co-stimulation signal by a CD28-binding affibody ligand developed by
combinatorial protein engineering. Protein Eng, 16, 691-697.
40. Wikman, M, Steffen, A.C., Gunncriusson, E., Tolmachev, V., Adams, G.P.,
Carlsson, J. and Stahl, S. (2004) Selection and characterization of HER2/neu-binding
affibody ligands. Protein Eng Des Sel, 17, 455-462.
41. Steffen, A.C., Orlova, A., Wikman, M., Nilsson, F.Y., Stahl, S., Adams, G.P.,
Tolmachev, V. and Carlsson, J. (2006) Affibody-mediated tumour targeting of HER-2
expressing xenografts in mice. Eur J Nucl Med Mol Imaging, 33, 631-638.
42. Orlova, A., Magnusson, M., Eriksson, T.L., Nilsson, M., Larsson, B., Hoiden-
Guthenberg, I., Widstrom, C, Carlsson, J., Tolmachev, V., Stahl, S. et al. (2006)
Tumor imaging using a picomolar affinity HER2 binding affibody molecule. Cancer
Res, 66, 4339-4348.
43. Orlova, A., Tolmachev, V., Pehrson, R., Lindborg, M., Tran, T., Sandstrom, M.,
Nilsson, F.Y., Wennborg, A., Abrahmsen, L. and Feldwisch, J. (2007) Synthetic
Affibody Molecules: A Novel Class of Affinity Ligands for Molecular Imaging of
HER2-Expressing Malignant Tumors. Cancer Res, 67, 2178-2186.
44. Hogbom, M., Eklund, M., Nygren, P.A. and Nordlund, P. (2003) Structural basis for
recognition by an in vitro evolved affibody. Proc Natl Acad Sei USA, 100, 3191-
3196.
45. Nixon, A.E. and Wood, CR. (2006) Engineered protein inhibitors of proteases. Curr
Opin Drug Discov Devel, 9, 261-268.
116
46. Egeblad, M. and Werb, Z. (2002) New functions for the matrix metalloproteinases in
cancer progression. Nat Rev Cancer, 2, 161-174.
47. Parks, W.C., Wilson, C.L. and Lopez-Boado, Y.S. (2004) Matrix metalloproteinases
as modulators of inflammation and innate immunity. Nat Rev Immunol, 4, 617-629.
48. Shapiro, S.D. (2002) Proteinases in chronic obstructive pulmonary disease. Biochem
Soc Trans, 30, 98-102.
49. Naruse, S., Kitagawa, M. and Ishiguro, H. (1999) Molecular understanding of chronic
pancreatitis: a perspective on the future. Mol Med Today, 5, 493-499.
50. Nunan, J. and Small, D.H. (2002) Proteolytic processing of the amyloid-beta protein
precursor of Alzheimer's disease. Essays Biochem, 38, 37-49.
51. Laval, S.H. and Bushby, K.M. (2004) Limb-girdle muscular dystrophies—from
genetics to molecular pathology. Neuropathol Appl Neurobiol, 30, 91-105.
52. Ascenzi, P., Bocedi, A., Bolognesi, M., Spallarossa, A., Coletta, M., Dc Cristofaro, R.
and Menegatti, E. (2003) The bovine basic pancreatic trypsin inhibitor (Kunitz
inhibitor): a milestone protein. Curr Protein Pept Sei, 4, 231-251.
53. Roberts, B.L., Markland, W., Ley, A.C., Kent, R.B., White, D.W., Guterman, S.K.
and Ladner, R.C. (1992) Directed evolution of a protein: selection of potent
neutrophil elastasc inhibitors displayed on M13 fusion phage. Proc Natl Acad Sei US
A, 89, 2429-2433.
54. Hautamaki, R.D., Kobayashi, D.K., Senior, R.M. and Shapiro, S.D. (1997)
Requirement for macrophage elastase for cigarette smoke-induced emphysema in
mice. Science, 277, 2002-2004.
55. Dyax Corp. (Ley AC, L.R., Guterman SK, Roberts Bruce L, Markland W, Ken RB).
(1997) Engineered human-derived kunitz domains that inhibit human neutrophil
elastase. US056633143 patent.
117
56. Delacourt, C, Herigault, S., Delclaux, C, Poncin, A., Levamc, M., Harf, A.,
Saudubray, F. and Lafuma, C. (2002) Protection against acute lung injury by
intravenous or intratracheal pretreatment with EPI-HNE-4, a new potent neutrophil
elastase inhibitor. Am J Respir Cell Mol Biol, 26, 290-297.
57. Grimbert, D., Vecellio, L., Delepine, P., Attucci, S., Boissinot, E., Poncin, A.,
Gauthier, F., Valat, C, Saudubray, F., Antonioz, P. et al. (2003) Characteristics of
EPI-hNE4 aerosol: a new elastase inhibitor for treatment of cystic fibrosis. JAerosol
Med, 16, 121-129.
58. Saudubray F, L.A., Durieu I, Wallaert B, Sayah H, Assael B, Cazzola G, Dubray C,
Poncin A, Lafuma C, Williams T. (2003) Phase Ha clinical study of a new human
neutrophil elastase inhibitor (HNE), EPI-HNE4 (DX-890), with repeated
administration by inhalation in adult cystic fibrosis patients. J Cystic Fibrosis, 2
(Suppl 1), A85.
59. Ley, A.C., Markland, W. and Ladner, R.C. (1996) Obtaining a family of high-affinity,
high-specificity protein inhibitors of plasmin and plasma kallikrein. Mol Divers, 2,
119-124.
60. Markland, W., Ley, A.C. and Ladner, R.C. (1996) Iterative optimization ofhigh-
affinity protease inhibitors using phage display. 2. Plasma kallikrein and thrombin.
Biochemistry, 35, 8058-8067.
61. Williams, A. and Baird, L.G. (2003) DX-88 and HAE: a developmental perspective.
Transfus Apher Sei, 29, 255-258.
62. Han, E.D., MacFarlane, R.C., Mulligan, A.N., Scafidi, J. and Davis, A.E., 3rd. (2002)
Increased vascular permeability in CI inhibitor-deficient mice mediated by the
bradykinin type 2 receptor. J Clin Invest, 109, 1057-1063.
63. Sedgwick, S.G. and Smerdon, S.J. (1999) The ankyrin repeat: a diversity of
interactions on a common structural framework. Trends Biochem Sei, 24, 311-316.
118
64. Suzuki, F., Goto, M., Sawa, C, Ito, S., Watanabe, H., Sawada, J. and Handa, H.
(1998) Functional interactions of transcription factor human GA-binding protein
subunits. J Biol Chem, 273, 29302-29308.
65. Binz, H.K., Amstutz, P., Kohl, A., Stumpp, M.T., Briand, C, Forrer, P., Grutter, M.G.
and Pluckthun, A. (2004) High-affinity binders selected from designed ankyrin repeat
protein libraries. Nat Biotechnol, 22, 575-582.
66. Amstutz, P., Binz, H.K., Parizek, P., Stumpp, M.T., Kohl, A., Grutter, M.G., Forrer,
P. and Pluckthun, A. (2005) Intracellular kinase inhibitors selected from
combinatorial libraries of designed ankyrin repeat proteins. JBiol Chem, 280, 24715-
24722.
67. Zahnd, C, Pecorari, F., Straumann, N., Wyler, E. and Pluckthun, A. (2006) Selection
and characterization of Her2 binding-designed ankyrin repeat proteins. J Biol Chem,
281,35167-35175.
68. Kohl, A., Amstutz, P., Parizek, P., Binz, H.K., Briand, C, Capitani, G., Forrer, P.,
Pluckthun, A. and Grutter, M.G. (2005) Allosteric inhibition of aminoglycoside
phosphotransferase by a designed ankyrin repeat protein. Structure, 13, 1131-1141.
69. Skerra, A. (2000) Lipocalins as a scaffold. Biochim Biophys Acta, 1482, 337-350.
70. Flower, D.R. (1995) Multiple molecular recognition properties of the lipocalin protein
family. J Mol Recognit, 8, 185-195.
71. Cowan, S.W., Newcomer, M.E. and Jones, T.A. (1990) Crystallographic refinement
of human serum retinol binding protein at 2A resolution. Proteins, 8, 44-61.
72. Schlehuber, S. and Skerra, A. (2005) Lipocalins in drug discovery: from natural
ligand-binding proteins to "anticalins". DrugDiscov Today, 10, 23-33.
119
73. Paesen, G.C., Adams, P.L., Harlos, K., Nuttall, P.A. and Stuart, D.I. (1999) Tick
histamine-binding proteins: isolation, cloning, and three-dimensional structure. Mol
Cell, 3, 661-671.
74. Beste, G., Schmidt, F.S., Stibora, T. and Skerra, A. (1999) Small antibody-like
proteins with prescribed ligand specificities derived from the lipocalin fold. Proc Natl
Acad Sei USA, 96,1898-1903.
75. Schlehuber, S., Beste, G. and Skerra, A. (2000) A novel type of receptor protein,
based on the lipocalin scaffold, with specificity for digoxigenin. J Mol Biol, 297,
1105-1120.
76. Mercader, J.V. and Skerra, A. (2002) Generation of anticalins with specificity for a
nonsymmetric phthalic acid ester. Anal Biochem, 308, 269-277.
77. Dalby, P.A., Hocss, R.H. and DeGrado, W.F. (2000) Evolution of binding affinity in a
WW domain probed by phage display. Protein Sei, 9, 2366-2376.
78. Parmi, S., Dente, L. and Cesareni, G. (2002) In vitro evolution of recognition
specificity mediated by SH3 domains reveals target recognition rules. JBiol Chem,
277,21666-2167r4.
79. Hiipakka, M., Poikonen, K. and Saksela, K. (1999) SH3 domains with high affinity
and engineered ligand specificities targeted to HIV-1 Nef. JMol Biol, 293,1097-
1106.
80. Norman, T.C., Smith, D.L., Sorger, P.K., Drees, B.L., O'Rourke, S.M., Hughes, T.R.,
Roberts, C.J., Friend, S.H., Fields, S. and Murray, A.W. (1999) Genetic selection of
peptide inhibitors of biological pathways. Science, 285, 591-595.
81. Ferrer, M., Maiolo, J., Kratz, P., Jackowski, J.L., Murphy, D.J., Dclagrave, S. and
Inglese, J. (2005) Directed evolution ofPDZ variants to generate high-affinity
detection reagents. Protein Eng Des Sel, 18, 165-173.
120
82. Kunz, C, Borghouts, C, Buerger, C. and Groner, B. (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, 983-
998.
83. Cohen, G.B., Ren, R. and Baltimore, D. (1995) Modular binding domains in signal
transduction proteins. Cell, 80, 237-248.
84. Musacchio, A. (2002) How SH3 domains recognize proline. Adv Protein Chem, 61,
211-268.
85. Karkkainen, S., Hiipakka, M., Wang, J.H., Kleino, I., Vaha-Jaakkola, M., Renkema,
G.H., Liss, M., Wagner, R. and Sakscla, K. (2006) Identification of preferred protein
interactions by phage-display of the human Src homology-3 proteome. EMBO Rep, 7,
186-191.
86. Yu, H., Rosen, M.K., Shin, T.B., Scidel-Dugan, C, Brugge, J.S. and Schreiber, S.L.
(1992) Solution structure of the SH3 domain of Src and identification of its ligand-
binding site. Science, 258, 1665-1668.
87. Musacchio, A., Noble, M., Pauptit, R., Wierenga, R. and Saraste, M. (1992) Crystal
structure of a Src-homology 3 (SH3) domain. Nature, 359, 851-855.
88. Kohda, D., Hatanaka, H., Odaka, M., Mandiyan, V., Ullrich, A., Schlessinger, J. and
Inagaki, F. (1993) Solution structure of the SH3 domain of phospholipase C-gamma.
Cell, 11, 953-960.
89. Koyama, S., Yu, FL, Dalgarno, D.C., Shin, T.B., Zydowsky, L.D. and Schreiber, S.L.
(1993) Structure of the PI3K SH3 domain and analysis of the SH3 family. Cell, 11,
945-952.
90. Noble, M.E., Musacchio, A., Saraste, M., Courtneidge, S.A. and Wierenga, R.K.
(1993) Crystal structure of the SH3 domain in human Fyn; comparison of the three-
121
dimensional structures of SH3 domains in tyrosine kinases and spectrin. Embo J, 12,
2617-2624.
91. Feng, S., Chen, J.K., Yu, H., Simon, J.A. and Schreiber, SX. (1994) Two binding
orientations for peptides to the Src SH3 domain: development of a general model for
SH3-ligand interactions. Science, 266, 1241-1247.
92. Mayer, B.J. (2001) SH3 domains: complexity in moderation. J Cell Sei, 114, 1253-
1263.
93. Ren, R., Mayer, B.J., Cicchetti, P. and Baltimore, D. (1993) Identification of a ten-
amino acid prolinc-rich SH3 binding site. Science, 259, 1157-1161.
94. Chan, B., Lanyi, A., Song, H.K., Griesbach, J., Simarro-Grande, M., Poy, F., Howie,
D., Sumegi, J., Terhorst, C. and Eck, M.J. (2003) SAP couples Fyn to SLAM immune
receptors. Nat Cell Biol, 5, 155-160.
95. Xu, W., Harrison, S.C. and Eck, M.J. (1997) Three-dimensional structure of the
tyrosine kinase c-Src. Nature, 385, 595-602.
96. Larson, S.M. and Davidson, A.R. (2000) The identification of conserved interactions
within the SH3 domain by alignment of sequences and structures. Protein Sei, 9,
2170-2180.
97. Wu, X., Knudsen, B., Feller, S.M., Zheng, J., Sali, A., Cowburn, D., Hanafusa, H. and
Kuriyan, J. (1995) Structural basis for the specific interaction of lysine-containing
prolinc-rich peptides with the N-terminal SH3 domain of c-Crk. Structure, 3, 215-
226.
98. Nguyen, J.T., Turck, C.W., Cohen, F.E., Zuckermann, R.N. and Lim, W.A. (1998)
Exploiting the basis of proline recognition by SH3 and WW domains: design ofN-
substituted inhibitors. Science, 282, 2088-2092.
122
99. Lee, C.H., Leung, B., Lemmon, M.A., Zheng, J., Cowburn, D., Kuriyan, J. and
Sakscla, K. (1995) A single amino acid in the SH3 domain of Hck determines its high
affinity and specificity in binding to H1V-1 Nef protein. Embo J, 14, 5006-5015.
100. Kawakami, T., Pennington, C.Y. and Robbins, K.C. (1986) Isolation and oncogenic
potential of a novel human src-likc gene. Mol Cell Biol, 6, 4195-4201.
101. Semba, K., Nishizawa, M., Miyajima, N., Yoshida, M.C., Sukegawa, J., Yamanashi,
Y., Sasaki, M., Yamamoto, T. and Toyoshima, K.. (1986) yes-related protooncogene,
syn, belongs to the protcin-tyrosine kinase family. Proc Natl Acad Sei USA, 83,
5459-5463.
102. Cooke, M.P. and Perlmutter, R.M. (1989) Expression of a novel form of the fyn
proto-oncogene in hematopoietic cells. New Biol, 1, 66-74.
103. Li, S.S. (2005) Specificity and versatility of SH3 and other proline-recognition
domains: structural basis and implications for cellular signal transduction. Biochem J,
390, 641-653.
104. Morton, C.J., Pugh, D.J., Brown, EX., Kahmann, J.D., Renzoni, D.A. and Campbell,
I.D. (1996) Solution structure and peptide binding of the SH3 domain from human
Fyn. Structure, 4, 705-714.
105. Filimonov, V.V., Azuaga, A.I., Viguera, A.R., Serrano, L. and Mateo, P.L. (1999) A
thermodynamic analysis of a family of small globular proteins: SH3 domains.
Biophys Chem, 11, 195-208.
106. Gherardi, E., Pannell, R. and Milstein, C. (1990) A single-step procedure for cloning
and selection of antibody-secreting hybridomas. JImmunol Methods, 126, 61-68.
107. Giovannoni, L., Viti, F., Zardi, L. and Neri, D. (2001) Isolation of anti-angiogenesis
antibodies from a large combinatorial repertoire by colony filter screening. Nucleic
Acids Res, 29, E27.
123
108. Viti, F., Nilsson, F., Demartis, S., Huber, A. and Neri, D. (2000) Design and use of
phage display libraries for the selection of antibodies and enzymes. Methods Enzymol,
326, 480-505.
109. Heinis, C, Melkko, S., Demartis, S. and Neri, D. (2002) Two general methods for the
isolation of enzyme activities by colony filter screening. Chem Biol, 9, 383-390.
110. Smith, G.P. (1985) Filamentous fusion phage: novel expression vectors that display
cloned antigens on the virion surface. Science, 228, 1315-1317.
111. McCafferty, J., Griffiths, A.D., Winter, G. and Chiswell, D.J. ( 1990) Phage
antibodies: filamentous phage displaying antibody variable domains. Nature, 348,
552-554.
112. Lowman, H.B., Bass, S.H., Simpson, N. and Wells, J.A. (1991) Selecting high-
affinity binding proteins by monovalent phage display. Biochemistry, 30, 10832-
10838.
113. Bothmann, H. and Pluckthun, A. (1998) Selection for a periplasmic factor improving
phage display and functional periplasmic expression. Nat Biotechnol, 16, 376-380.
114. Kristensen, P. and Winter, G. (1998) Proteolytic selection for protein folding using
filamentous bacteriophages. Fold Des, 3, 321-328.
115. Pederscn, H., Holder, S., Sutherlin, D.P., Schwitter, U., King, D.S. and Schultz, P.G.
(1998) A method for directed evolution and functional cloning of enzymes. Proc Natl
Acad Sei USA, 95, 10523-10528.
116. Demartis, S., Huber, A., Viti, F., Lozzi, L., Giovannoni, L., Neri, P., Winter, G. and
Neri, D. (1999) A strategy for the isolation of catalytic activities from repertoires of
enzymes displayed on phage. JMol Biol, 286, 617-633.
124
117. Heinis, C, Huber, A., Dcmartis, S., Bertschinger, J., Melkko, S., Lozzi, L., Neri, P.
and Neri, D. (2001) Selection of catalytically active biotin ligasc and trypsin mutants
by phage display. Protein Eng, 14, 1043-1052.
118. Silacci, M., Brack, S., Schirm, G., Marlind, J., Ettorre, A., Merlo, A., Viti, F. and
Neri, D. (2005) Design, construction, and characterization of a large synthetic human
antibody phage display library. Proteomics, 5, 2340-2350.
119. Greenwood, J., Willis, A.E. and Perham, R.N. (1991) Multiple display of foreign
peptides on a filamentous bacteriophage. Peptides from Plasmodium falciparum
circumsporozoite protein as antigens. J Mol Biol, 220, 821-827.
120. Gao, C., Mao, S., Lo, C.H., Wirsching, P., Lerner, R.A. and Janda, K.D. (1999)
Making artificial antibodies: a format for phage display of combinatorial
heterodimeric arrays. Proc Natl Acad Sei USA, 96, 6025-6030.
121. Gao, C, Mao, S„ Kaufmann, G., Wirsching, P., Lerner, R.A. and Janda, K.D. (2002)
A method for the generation of combinatorial antibody libraries using pIX phage
display. Proc Natl Acad Sei U SA, 99, 12612-12616.
122. Jespers, L.S., Messens, J.H., De Keyser, A., Eeckhout, D., Van den Brande, I.,
Gansemans, Y.G., Lauwereys, M.J., Vlasuk, G.P. and Stanssens, P.E. (1995) Surface
expression and ligand-based selection of cDNAs fused to filamentous phage gene VI.
Biotechnology (N Y), 13, 378-382.
123. Iannolo, G., Mincnkova, O., Petruzzelli, R. and Cesareni, G. (1995) Modifying
filamentous phage capsid: limits in the size of the major capsid protein. J Mol Biol,
248, 835-844.
124. Rondot, S., Koch, J., Breitling, F. and Dubcl, S. (2001) A helper phage to improve
single-chain antibody presentation in phage display. Nat Biotechno1,19, 75-78.
125
125. Steiner, D., Forrer, P., Stumpp, M.T. and Pluckthun, A. (2006) Signal sequences
directing cotranslational translocation expand the range of proteins amenable to phage
display. Nat Biotechnol, 24, 823-831.
126. Jestin, J.L., Volioti, G. and Winter, G. (2001) Improving the display of proteins on
filamentous phage. Res Microbiol, 152, 187-191.
127. Rauchenberger, R., Borges, E., Thomassen-Wolf, E., Rom, E., Adar, R., Yaniv, Y.,
Malka, M., Chumakov, I., Kötzer, S., Resnitzky, D. et al (2003) Human
combinatorial Fab library yielding specific and functional antibodies against the
human fibroblast growth factor receptor 3. JBiol Chem, 278, 38194-38205.
128. Boder, E.T. and Wittrup, K.D. (1997) Yeast surface display for screening
combinatorial polypeptide libraries. Nat Biotechnol, 15, 553-557.
129. Feldhaus, M.J. and Siegel, R.W. (2004) Yeast display of antibody fragments: a
discovery and characterization platform. JImmunol Methods, 290, 69-80.
130. Boder, E.T. and Wittrup, K.D. (2000) Yeast surface display for directed evolution of
protein expression, affinity, and stability. Methods Enzymol, 328, 430-444.
131. Boder, E.T. and Wittrup, K.D. ( 1998) Optimal screening of surface-displayed
polypeptide libraries. Biotechnol Prog, 14, 55-62.
132. VanAntwerp, J.J. and Wittrup, K.D. (2000) Fine affinity discrimination by yeast
surface display and flow cytometry. Biotechnol Prog, 16, 31-37.
133. Colby, D.W., Kellogg, B.A., Graff, C.P., Yeung, Y.A., Swcrs, J.S. and Wittrup, K.D.
(2004) Engineering antibody affinity by yeast surface display. Methods Enzymol, 388,
348-358.
134. Mattheakis, L.C., Bhatt, R.R. and Dower, W.J. (1994) An in vitro polysome display
system for identifying ligands from very large peptide libraries. Proc Natl Acad Sei U
SA, 91, 9022-9026.
126
135. Hanes, J. and Pluckthun, A. (1997) In vitro selection and evolution of functional
proteins by using ribosome display. Proc Natl Acad Sei USA, 94, 4937-4942.
136. Hanes, J., Jermutus, L., Weber-Bornhauser, S., Bosshard, H.R. and Pluckthun, A.
(1998) Ribosome display efficiently selects and evolves high-affinity antibodies in
vitro from immune libraries. Proc Natl Acad Sei USA, 95, 14130-14135.
137. Ryabova, L.A., Desplancq, D., Spirin, A.S. and Pluckthun, A. (1997) Functional
antibody production using cell-free translation: effects of protein disulfide isomerase
and chaperones. Nat Biotechnol, 15, 79-84.
138. Jermutus, L., Honegger, A., Schwesinger, F., Hanes, J. and Pluckthun, A. (2001)
Tailoring in vitro evolution for protein affinity or stability. Proc Natl Acad Sei USA,
98, 75-80.
139. Knappik, A., Gc, L., Honegger, A., Pack, P., Fischer, M., Wellnhofer, G., Hocss, A.,
Wolle, J., Pluckthun, A. and Virnekas, B. (2000) Fully synthetic human combinatorial
antibody libraries (HuCAL) based on modular consensus frameworks and CDRs
randomized with trinucleotides. J Mol Biol, 296, 57-86.
140. Hanes, J., Schaffitzel, C, Knappik, A. and Pluckthun, A. (2000) Picomolar affinity
antibodies from a fully synthetic naive library selected and evolved by ribosome
display. Nat Biotechnol, 18, 1287-1292.
141. Zahnd, C, Spinelli, S., Luginbuhl, B.s Amstutz, P., Cambillau, C. and Pluckthun, A.
(2004) Directed in vitro evolution and crystallographic analysis of a peptide-binding
single chain antibody fragment (scFv) with low picomolar affinity. J Biol Chem, 279,
18870-18877.
142. Bertschinger, J. and Neri, D. (2004) Covalent DNA display as a novel tool for
directed evolution of proteins in vitro. Protein Eng Des Sel, 17, 699-707.
143. Tawfik, D.S. and Griffiths, A.D. (1998) Man-made cell-like compartments for
molecular evolution. Nat Biotechnol, 16, 652-656.
127
144. Chen, L., MacMillan, A.M., Chang, W., Ezaz-Nikpay, K., Lane, W.S. and Verdine,
G.L. (1991) Direct identification of the active-site nucleophile in a DNA (cytosine-5)-
methyltransferase. Biochemistry, 30,11018-11025.
145. Zardi, L., Carnemolla, B., Siri, A., Petersen, T.E., Paolella, G., Sebastio, G. and
Baralle, F.E. ( 1987) Transformed human cells produce a new fibronectin isoform by
preferential alternative splicing of a previously unobserved exon. EmboJ, 6, 2337-
2342.
146. Bertschingcr, J., Grabulovski, D. and Neri, D. (2007) Selection of single domain
binding proteins by covalent DNA display. Protein Eng Des Sel, 20, 57-68.
147. Peters, T., Jr. (1985) Serum albumin. Adv Protein Chem, 37, 161-245.
148. Dennis, M.S., Zhang, M., Meng, Y.G., Kadkhodayan, M., Kirchhofer, D., Combs, D.
and Damico, L.A. (2002) Albumin binding as a general strategy for improving the
pharmacokinetics of proteins. J Biol Chem, 111, 35035-35043.
149. Makrides, S.C., Nygren, P.A., Andrews, B., Ford, P.J., Evans, K.S., Hayman, E.G.,
Adari, H., Uhlen, M. and Toth, C.A. (1996) Extended in vivo half-life of human
soluble complement receptor type 1 fused to a serum albumin-binding receptor. J
Pharmacol Exp Ther, 277, 534-542.
150. Markussen, J., Havelund, S., Kurtzhals, P., Andersen, A.S., Halstrom, J., Hasselager,
E„ Larsen, U.D., Ribel, U., Schaffer, L., Vad, K. et al (1996) Soluble, fatty acid
acylated insulins bind to albumin and show protracted action in pigs. Diabetologia,
39, 281-288.
151. Nguyen, A., Reyes, A.E., 2nd, Zhang, M., McDonald, P., Wong, W.L., Damico, L.A.
and Dennis, M.S. (2006) The pharmacokinetics of an albumin-binding Fab (AB.Fab)
can be modulated as a function of affinity for albumin. Protein Eng Des Sel, 19, 291-
297.
128
152. Hu, S., Shively, L., Raubitschek, A., Sherman, M., Williams, L.E., Wong, J.Y.,
Shively, J.E. and Wu, A.M. (1996) Minibody: A novel engineered anti-
carcinoembryonic antigen antibody fragment (single-chain Fv-CH3) which exhibits
rapid, high-level targeting of xenografts. Cancer Res, 56, 3055-3061.
153. Ettorre, A., Rosli, C, Silacci, M., Brack, S., McCombie, G., Knochenmuss, R., Elia,
G. and Neri, D. (2006) Recombinant antibodies for the depletion of abundant proteins
from human serum. Proteomics.
154. Bosslet, K., Straub, R., Blumrich, M., Czech, J., Gerken, M., Sperker, B., Kroemer,
H.K., Gesson, J.P., Koch, M. and Monneret, C. (1998) Elucidation of the mechanism
enabling tumor selective prodrug monotherapy. Cancer Res, 58, 1195-1201.
155. Ramachandran, C. and Melnick, S.J. (1999) Multidrug resistance in human tumors-
molecular diagnosis and clinical significance. Mol Diagn, 4, 81-94.
156. Konerding, M.A., Fait, E. and Gaumann, A. (2001) 3D microvascular architecture of
pre-cancerous lesions and invasive carcinomas of the colon. Br J Cancer, 84, 1354-
1362.
157. Eberhard, A., Kahlert, S., Goede, V., Hemmerlein, B., Plate, K.H. and Augustin, H.G.
(2000) Heterogeneity of angiogenesis and blood vessel maturation in human tumors:
implications for antiangiogenic tumor therapies. Cancer Res, 60, 1388-1393.
158. McDonald, D.M. and Choyke, P.L. (2003) Imaging of angiogenesis: from microscope
to clinic. Nat Med, 9, 713-725.
159. Tozer, G.M., Ameer-Beg, S.M., Baker, J., Barber, P.R., Hill, S.A., Hodgkiss, R.J.,
Locke, R., Prise, V.E., Wilson, I. and Vojnovic, B. (2005) Intravital imaging of
tumour vascular networks using multi-photon fluorescence microscopy. Adv Drug
DelivRev, 57, 135-152.
129
160. Tozer, G.M., Lewis, S., Michalowski, A. and Aber, V. (1990) The relationship
between regional variations in blood flow and histology in a transplanted rat
fibrosarcoma. Br J Cancer, 61, 250-257.
161. Helmlinger, G„ Yuan, F., Dellian, M. and Jain, R.K. ( 1997) Interstitial pH and p02
gradients in solid tumors in vivo: high-resolution measurements reveal a lack of
correlation. Nat Med, 3, 177-182.
162. Denckamp, J. and Hobson, B. (1982) Endothelial-cell proliferation in experimental
tumours. Br J Cancer, 46, 711-720.
163. Brown, N.S. and Bicknell, R. (2001) Hypoxia and oxidative stress in breast cancer.
Oxidative stress: its effects on the growth, metastatic potential and response to
therapy of breast cancer. Breast Cancer Res, 3, 323-327.
164. Toyokuni, S., Okamoto, K., Yodoi, J. and Hiai, H. (1995) Persistent oxidative stress
in cancer. FEBS Lett, 358, 1-3.
165. Carnemolla, B., Balza, E., Siri, A., Zardi, L., Nicotra, M.R., Bigotti, A. andNatali,
P.G. (1989) A tumor-associated fibronectin isoform generated by alternative splicing
of messenger RNA precursors. J Cell Biol, 108, 1139-1148.
166. Fukuda, T., Yoshida, N., Kataoka, Y, Manabe, R., Mizuno-Horikawa, Y., Sato, M.,
Kuriyama, K., Yasui, N. and Sekiguchi, K. (2002) Mice lacking the EDB segment of
fibronectin develop normally but exhibit reduced cell growth and fibronectin matrix
assembly in vitro. Cancer Res, 62, 5603-5610.
167. Winter, G., Griffiths, A.D., Hawkins, R.E. and Hoogenboom, H.R. (1994) Making
antibodies by phage display technology. Annu Rev Immunol, 12, 433-455.
168. Pini, A., Viti, F., Santucci, A., Carnemolla, B., Zardi, L., Neri, P. and Neri, D. (1998)
Design and use of a phage display library. Human antibodies with subnanomolar
affinity against a marker of angiogenesis eluted from a two-dimensional gel. J Biol
Chem, 273, 21769-21776.
130
169. Carnemolla, B., Neri, D., Castellani, P., Leprini, A., Neri, G., Pini, A., Winter, G. and
Zardi, L. (1996) Phage antibodies with pan-species recognition of the oncofoetal
angiogenesis marker fibronectin ED-B domain. Int J Cancer, 68, 397-405.
170. Neri, D., Carnemolla, B., Nissim, A., Leprini, A., Querze, G., Balza, E., Pini, A.,
Tarli, L., Halin, C, Neri, P. et al. (1997) Targeting by affinity-matured recombinant
antibody fragments of an angiogenesis associated fibronectin isoform. Nat Biotechnol,
15, 1271-1275.
171. Demartis, S., Tarli, L., Borsi, L., Zardi, L. and Neri, D. (2001) Selective targeting of
tumour neovasculature by a radiohalogenatcd human antibody fragment specific for
the ED-B domain of fibronectin. EurJNucl Med, 28, 534-539.
172. Tarli, L., Balza, E., Viti, F., Borsi, L., Castellani, P., Berndorff, D., Dinkelborg, L.,
Neri, D. and Zardi, L. (1999) A high-affinity human antibody that targets tumoral
blood vessels. Blood, 94, 192-198.
173. Viti, F., Tarli, L., Giovannoni, L., Zardi, L. and Neri, D. (1999) Increased binding
affinity and valence of recombinant antibody fragments lead to improved targeting of
tumoral angiogenesis. Cancer Res, 59, 347-352.
174. Santimaria, M., Moscatelli, G., Viale, G.L., Giovannoni, L., Neri, G., Viti, F., Leprini,
A., Borsi, L., Castellani, P., Zardi, L. et al. (2003) Immunoscintigraphic detection of
the ED-B domain of fibronectin, a marker of angiogenesis, in patients with cancer.
Clin Cancer Res, 9, 571-579.
175. Birchler, M., Neri, G., Tarli, L., Halin, C, Viti, F. and Neri, D. (1999) Infrared
photodetection for the in vivo localisation of phage-derived antibodies directed
against angiogenic markers. JImmunol Methods, 231, 239-248.
176. Birchler, M., Viti, F., Zardi, L., Spiess, B. and Neri, D. (1999) Selective targeting and
photocoagulation of ocular angiogenesis mediated by a phage-derived human
antibody fragment. Nat Biotechnol, 17, 984-988.
131
177. Borsi, L., Balza, E., Bestagno, M., Castellani, P., Carnemolla, B., Biro, A., Leprini,
A., Sepulveda, J., Burronc, O., Ncri, D. et al. (2002) Selective targeting of tumoral
vasculature: comparison of different formats of an antibody (LI9) to the ED-B
domain of fibronectin. Int J Cancer, 102, 75-85.
178. Bcrndorff, D., Borkowski, S., Sieger, S., Rother, A., Friebe, M., Viti, F., Hilger, CS.,
Cyr, J.E. and Dinkelborg, L.M. (2005) Radioimmunotherapy of solid tumors by
targeting extra domain B fibronectin: identification of the best-suited
radioimmunoconjugate. Clin Cancer Res, 11, 7053s-7063s.
179. Marty, C, Odermatt, B., Schott, H., Neri, D., Ballmer-Hofer, K., Klemenz, R. and
Schwendener, R.A. (2002) Cytotoxic targeting of F9 teratocarcinoma tumours with
anti-ED-B fibronectin scFv antibody modified liposomes. Br J Cancer, 87, 106-112.
180. Nilsson, F., Kosmehl, H., Zardi, L. and Neri, D. (2001) Targeted delivery of tissue
factor to the ED-B domain of fibronectin, a marker of angiogencsis, mediates the
infarction of solid tumors in mice. Cancer Res, 61, 711-716.
181. Carnemolla, B., Borsi, L., Balza, E., Castellani, P., Meazza, R., Berndt, A., Ferrini, S.,
Kosmehl, H., Neri, D. and Zardi, L. (2002) Enhancement of the antitumor properties
of interlcukin-2 by its targeted delivery to the tumor blood vessel extracellular matrix.
Blood,99, 1659-1665.
182. Halin, C, Rondini, S., Nilsson, F., Berndt, A., Kosmehl, H., Zardi, L. and Neri, D.
(2002) Enhancement of the antitumor activity of interleukin-12 by targeted delivery to
neovasculature. Nat Biotechnol, 20, 264-269.
183. Halin, C, Niesncr, U., Villani, M.E., Zardi, L. and Neri, D. (2002) Tumor-targeting
properties of antibody-vascular endothelial growth factor fusion proteins. Int J
Cancer, 102, 109-116.
132
184. Borsi, L., Balza, E., Carnemolla, B., Sassi, F., Castellani, P., Berndt, A., Kosmehl, H.,
Biro, A., Siri, A., Orecchia, P. et al. (2003) Selective targeted delivery of TNFalpha to
tumor blood vessels. Blood, 102, 4384-4392.
185. Ebbinghaus, C, Ronca, R., Kaspar, M., Grabulovski, D., Berndt, A., Kosmehl, H.,
Zardi, L, and Neri, D. (2005) Engineered vascular-targeting antibody-interfcron-
gamma fusion protein for cancer therapy. Int J Cancer, 116, 304-313.
186. Heinis, C, Alessi, P. and Neri, D. (2004) Engineering a thermostable human prolyl
endopeptidase for antibody-directed enzyme prodrug therapy. Biochemistry, 43, 6293-
6303.
187. Melkko, S., Halin, C, Borsi, L., Zardi, L. and Neri, D. (2002) An antibody-
calmodulin fusion protein reveals a functional dependence between macromolecular
isoelectric point and tumor targeting performance. Int JRadiât Oncol Biol Phys, 54,
1485-1490.
188. Niesner, U., Halin, C, Lozzi, L., Gunthert, M., Neri, P., Wunderli-Allcnspach, H.,
Zardi, L. and Neri, D. (2002) Quantitation of the tumor-targeting properties of
antibody fragments conjugated to cell-permeating H1V-1 TAT peptides. Bioconjug
Chem, 13, 729-736.
189. Menrad, A. and Mcnssen, H.D. (2005) ED-B fibronectin as a target for antibody-
based cancer treatments. Expert Opin Ther Targets, 9, 491-500.
190. Siri, A., Carnemolla, B., Saginati, M., Leprini, A., Casari, G., Baralle, F. and Zardi, L.
(1991) Human tenascin: primary structure, pre-mRNA splicing patterns and
localization of the epitopes recognized by two monoclonal antibodies. Nucleic Acids
Res, 19, 525-531.
191. Borsi, L., Carnemolla, B., Nicolo, G., Spina, B., Tanara, G. and Zardi, L. (1992)
Expression of different tenascin isoforms in normal, hyperplastic and neoplastic
human breast tissues. Int J Cancer, 52, 688-692.
133
192. Carnemolla, B., Castellani, P., Ponassi, M., Borsi, L., Urbini, S., Nicolo, G.,
Dorcaratto, A, Vialc, G., Winter, G., Neri, D. et al. (1999) Identification of a
glioblastoma-associatcd tenascin-C isoform by a high affinity recombinant antibody.
Am JPathol, 154, 1345-1352.
193. Silacci, M., Brack, S.S., Spath, N., Buck, A., Hillingcr, S., Ami, S.s Weder, W., Zardi,
L. and Neri, D. (2006) Human monoclonal antibodies to domain C of tcnascin-C
selectively target solid tumors in vivo. Protein Eng Des Sel, 19, 471-478.
194. Bartolomei, M., Mazzetta, C, Handkiewicz-Junak, D., Bodei, L., Rocca, P., Grana,
C, Maira, G., Sturiale, C, Villa, G. and Paganelli, G. (2004) Combined treatment of
glioblastoma patients with locoregional pre-targeted 90Y-biotin radioimmunotherapy
and temozolomide. Q JNucl MedMol Imaging, 48, 220-228.
195. Paganelli, G., Grana, C, Chinol, M., Cremonesi, M., De Cicco, C, De Braud, F.,
Robertson, C, Zurrida, S., Casadio, C, Zoboli, S. et al. (1999) Antibody-guided
three-step therapy for high grade glioma with yttrium-90 biotin. Eur J Nucl Med, 26,
348-357.
196. Riva, P., Franceschi, G., Frattarelli, M., Lazzari, S., Riva, N., Giuliani, G., Casi, M.,
Sarti, G., Guiducci, G., Giorgetti, G. et al (1999) Loco-regional radioimmunotherapy
of high-grade malignant gliomas using specific monoclonal antibodies labeled with
90Y: a phase I study. Clin Cancer Res, 5, 3275s-3280s.
197. Riva, P., Franceschi, G., Frattarelli, M., Riva, N., Guiducci, G., Cremonini, A.M.,
Giuliani, G., Casi, M., Gentile, R., Jekunen, A.A. et al. (1999) 1311 radioconjugated
antibodies for the locoregional radioimmunotherapy of high-grade malignant glioma—
phase I and II study. Acta Oncol, 38, 351-359.
198. Leins, A., Riva, P., Lindstedt, R.s Davidoff, M.S., Mehraein, P. and Weis, S. (2003)
Expression of tenascin-C in various human brain tumors and its relevance for survival
in patients with astrocytoma. Cancer, 98, 2430-2439.
134
199. Brack, S.S., Silacci, M., Birchler, M. and Neri, D. (2006) Tumor-targeting properties
of novel antibodies specific to the large isoform of tenascin-C. Clin Cancer Res, 12,
3200-3208.
200. Ran, S., Downes, A. and Thorpe, P.E. (2002) Increased exposure of anionic
phospholipids on the surface of tumor blood vessels. Cancer Res, 62, 6132-6140.
201. Bucki, R., Janmey, P.A., Vegners, R., Giraud, F. and Sulpice, J.C. (2001)
Involvement ofphosphatidylinositol 4,5-bisphosphate in phosphatidylserine exposure
in platelets: use of a permeant phosphoinositide-binding peptide. Biochemistry, 40,
15752-15761.
202. Ran, S., He, J., Huang, X., Soares, M., Scothorn, D. and Thorpe, P.E. (2005)
Antitumor effects of a monoclonal antibody that binds anionic phospholipids on the
surface of tumor blood vessels in mice. Clin Cancer Res, 11, 1551 -1562.
203. Huang, X., Bennett, M. and Thorpe, P.E. (2005) A monoclonal antibody that binds
anionic phospholipids on tumor blood vessels enhances the antitumor effect of
docetaxel on human breast tumors in mice. Cancer Res, 65, 4408-4416.
204. Oh, P., Li, Y., Yu, J., Durr, E., Krasinska, K.M., Carver, L.A., Testa, J.E. and
Schnitzer, J.E. (2004) Subtractive proteomic mapping of the endothelial surface in
lung and solid tumours for tissue-specific therapy. Nature, 429, 629-635.
205. Hood, i.D., Bednarski, M, Frausto, R., Guccionc, S., Reisfeld, R.A., Xiang, R. and
Chcrcsh, D.A. (2002) Tumor regression by targeted gene delivery to the
neovasculature. Science, 296, 2404-2407.
206. Gutheil, J.C, Campbell, T.N., Pierce, P.R., Watkins, J.D., Huse, W.D., Bodkin, D.J.
and Cheresh, D.A. (2000) Targeted antiangiogenic therapy for cancer using Vitaxin: a
humanized monoclonal antibody to the integrin alphavbeta3. Clin Cancer Res, 6,
3056-3061.
135
207. Posey, J.A., Khazacli, MB., DelGrosso, A., Saleh, M.N., Lin, C.Y., Huse, W. and
LoBuglio, A.F. (2001) A pilot trial of Vitaxin, a humanized anti-vitronectin receptor
(anti alpha v beta 3) antibody in patients with metastatic cancer. Cancer Biother
Radiopharm, 16, 125-132.
208. Baluk, P., Morikawa, S., Haskell, A., Mancuso, M. and McDonald, D.M. (2003)
Abnormalities of basement membrane on blood vessels and endothelial sprouts in
tumors. Am JPathol, 163, 1801-1815.
209. Sipkins, D.A., Cheresh, D.A., Kazemi, M.R., Nevin, L.M., Bednarski, M.D. and Li,
K.C. (1998) Detection of tumor angiogenesis in vivo by alphaVbeta3-targeted
magnetic resonance imaging. Nat Med, 4, 623-626.
210. Brekken, R.A., Huang, X., King, S.W. and Thorpe, P.E. (1998) Vascular endothelial
growth factor as a marker of tumor endothelium. Cancer Res, 58, 1952-1959.
211. Fcrrara, N„ Hillan, K.J., Gerber, H.P. and Novotny, W. (2004) Discovery and
development of bevacizumab, an anti-VEGF antibody for treating cancer. Nat Rev
DrugDiscov, 3, 391-400.
212. Ferrara, N. (2004) Vascular endothelial growth factor: basic science and clinical
progress. EndocrRev, 25, 581-611.
213. Cooke, S.P., Boxer, G.M., Lawrence, L., Pedley, R.B., Spencer, D.I., Begent, R.H.
and Chester, K.A. (2001) A strategy for antitumor vascular therapy by targeting the
vascular endothelial growth factor: receptor complex. Cancer Res, 61, 3653-3659.
214. Kc, L., Qu, H., Nagy, J.A., Eckelhoefer, I.A., Masse, E.M., Dvorak, A.M. and
Dvorak, H.F. (1996) Vascular targeting of solid and ascites tumours with antibodies
to vascular endothelial growth factor. Eur J Cancer, 32A, 2467-2473.
215. Prewett, M., Huber, J., Li, Y., Santiago, A., O'Connor, W., King, K., Overholser, J.,
Hooper, A., Pytowski, B., Witte, L. et al. (1999) Antivascular endothelial growth
136
factor receptor (fetal liver kinase 1) monoclonal antibody inhibits tumor angiogenesis
and growth of several mouse and human tumors. Cancer Res, 59, 5209-5218.
216. Zhu, Z. and Witte, L. ( 1999) Inhibition of tumor growth and metastasis by targeting
tumor-associated angiogenesis with antagonists to the receptors of vascular
endothelial growth factor. Invest New Drugs, 17, 195-212.
217. Brckken, R.A., Overholser, J.P., Stastny, V.A., Waltenberger, J., Minna, J.D. and
Thorpe, P.E. (2000) Selective inhibition of vascular endothelial growth factor (VEGF)
receptor 2 (KDR/Flk-1) activity by a monoclonal anti-VEGF antibody blocks tumor
growth in mice. Cancer Res, 60, 5117-5124.
218. Bostwick, D.G., Grignon, D.J., Hammond, M.E., Amin, M.B., Cohen, M., Crawford,
D., Gospadarowicz, M., Kaplan, R.S., Miller, D.S., Montironi, R. et al. (2000)
Prognostic factors in prostate cancer. College of American Pathologists Consensus
Statement 1999. Arch Pathol Lab Med, 124, 995-1000.
219. Chang, S.S., Reuter, V.E., Heston, W.D., Bander, N.H., Grauer, L.S. and Gaudin,
P.B. (1999) Five different anti-prostate-specific membrane antigen (PSMA)
antibodies confirm PSMA expression in tumor-associated neovasculature. Cancer
Res, 59, 3192-3198.
220. Liu, H., Moy, P., Kim, S., Xia, Y., Rajasckaran, A., Navarro, V., Knudsen, B. and
Bander, N.H. (1997) Monoclonal antibodies to the extracellular domain of prostate-
specific membrane antigen also react with tumor vascular endothelium. Cancer Res,
57, 3629-3634.
221. Li, Y., Tian, Z., Rizvi, S.M., Bander, N.H. and Allen, B.J. (2002) In vitro and
preclinical targeted alpha therapy ofhuman prostate cancer with Bi-213 labeled J591
antibody against the prostate specific membrane antigen. Prostate Cancer Prostatic
Dis, 5, 36-46.
137
222. Bander, N.H., Nanus, D.M., Milowsky, M.I., Kostakoglu, L., Vallabahajosula, S. and
Goldsmith, SJ. (2003) Targeted systemic therapy of prostate cancer with a
monoclonal antibody to prostate-specific membrane antigen. Semin Oncol, 30, 667-
676.
223. Milowsky, M.I., Nanus, D.M., Kostakoglu, L., Vallabhajosula, S., Goldsmith, S.J. and
Bander, N.H. (2004) Phase I trial of yttrium-90-labeled anti-prostate-specific
membrane antigen monoclonal antibody J591 for androgen-independent prostate
cancer. J Clin Oncol, 22, 2522-2531.
224. Bander, N.H., Milowsky, M.I., Nanus, D.M., Kostakoglu, L., Vallabhajosula, S. and
Goldsmith, S.J. (2005) Phase 1 trial of 1771utetium-labeled J591, a monoclonal
antibody to prostate-specific membrane antigen, in patients with androgen-
independent prostate cancer. J Clin Oncol, 23, 4591-4601.
225. Wang, J.M., Kumar, S., Pye, D., van Agthoven, A.J., Krupinski, J. and Hunter, R.D.
(1993) A monoclonal antibody detects heterogeneity in vascular endothelium of
tumours and normal tissues. Int J Cancer, 54, 363-370.
226. Burrows, F.J., Derbyshire, E.J., Tazzari, P.L., Amlot, P., Gazdar, A.F., King, S.W.,
Letarte, M., Vitetta, E.S. and Thorpe, P.E. (1995) Up-regulation of endoglin on
vascular endothelial cells in human solid tumors: implications for diagnosis and
therapy. Clin Cancer Res, 1, 1623-1634.
227. Matsubara, S., Bourdeau, A., terBrugge, K.G., Wallace, C. and Letarte, M. (2000)
Analysis of endoglin expression in normal brain tissue and in cerebral arteriovenous
malformations. Stroke, 31, 2653-2660.
228. Balza, E., Castellani, P., Zijlstra, A., Neri, D., Zardi, L. and Siri, A. (2001) Lack of
specificity of endoglin expression for tumor blood vessels. Int J Cancer, 94, 579-585.
138
229. Bredow, S., Lewin, M., Hofmann, B., Marecos, E. and Weissleder, R. (2000) Imaging
of tumour neovasculature by targeting the TGF-beta binding receptor endoglin. Eur J
Cancer, 36, 675-681.
230. Fonsatti, E., Jekunen, A.P., Kairemo, K.J., Coral, S., Snellman, M., Nicotra, M.R.,
Natali, P.G., Altomonte, M. and Maio, M. (2000) Endoglin is a suitable target for
efficient imaging of solid tumors: in vivo evidence in a canine mammary carcinoma
model. Clin Cancer Res, 6, 2037-2043.
231. Porkka, K., Laakkonen, P., Hoffman, J.A., Bernasconi, M. and Ruoslahti, E. (2002) A
fragment of the HMGN2 protein homes to the nuclei of tumor cells and tumor
endothelial cells in vivo. Proc Natl Acad Sei USA, 99, 7444-7449.
232. Christian, S., Pilch, J., Akerman, M.E., Porkka, K., Laakkonen, P. and Ruoslahti, E.
(2003) Nucleolin expressed at the cell surface is a marker of endothelial cells in
angiogenic blood vessels. J Cell Biol, 163, 871-878.
233. Pasqualini, R., Koivuncn, E. and Ruoslahti, E. (1997) Alpha v intcgrins as receptors
for tumor targeting by circulating ligands. Nat Biotechnol, 15, 542-546.
234. Pasqualini, R., Koivunen, E., Kain, R., Lahdenranta, J., Sakamoto, M., Stryhn, A.,
Ashmun, R.A., Shapiro, L.H., Arap, W. and Ruoslahti, E. (2000) Aminopeptidase N
is a receptor for tumor-homing peptides and a target for inhibiting angiogenesis.
Cancer Res, 60, 722-727.
235. Curnis, F., Sacchi, A., Borgna, L., Magni, F., Gasparri, A. and Corti, A. (2000)
Enhancement of tumor necrosis factor alpha antitumor immunothcrapeutic properties
by targeted delivery to aminopeptidase N (CD13). Nat Biotechnol, 18, 1185-1190.
236. Curnis, F., Sacchi, A. and Corti, A. (2002) Improving chemotherapeutic drug
penetration in tumors by vascular targeting and barrier alteration. J Clin Invest, 110,
475-482.
139
237. Wakai, Y., Matsui, J., Koizumi, K., Tsunoda, S., Makimoto, H., Ohizumi, I.,
Taniguchi, K., Kaiho, S., Saito, H., Utoguchi, N. et ah (2000) Effective cancer
targeting using an anti-tumor tissue vascular endothclium-specific monoclonal
antibody (TES-23). Jpn J Cancer Res, 91, 1319-1325.
238. Huminiecki, L., Gorn, M., Suchting, S„ Poulsom, R. and Bickncll, R. (2002) Magic
roundabout is a new member of the roundabout receptor family that is endothelial
specific and expressed at sites of active angiogenesis. Genomics, 19, 547-552.
239. Aoka, Y., Johnson, F.L., Penta, K., Hirata Ki, K., Hidai, C, Schatzman, R., Varner,
J.A. and Quertermous, T. (2002) The embryonic angiogenic factor Dell accelerates
tumor growth by enhancing vascular formation. Microvasc Res, 64, 148-161.
240. Sullivan, D.C., Huminiecki, L., Moore, J.W., Boyle, J.J., Poulsom, R., Creamer, D.,
Barker, J. and Bicknell, R. (2003) EndoPDI, a novel protein-disulfide isomerasc-like
protein that is preferentially expressed in endothelial cells acts as a stress survival
factor. J Biol Chem, 278, 47079-47088.
241. Fukumoto, M. (2004) Single-photon agents for tumor imaging: 201T1, 99mTc-MIBI,
and 99mTc-tetrofosmin. Ann Nucl Med, 18, 79-95.
242. Vercl, 1., Visser, G.W. and van Dongen, G.A. (2005) The promise of immuno-PET in
radioimmunotherapy. J Nucl Med, 46 Suppl 1, 164S-171S.
243. Dzik-Jurasz, A.S. (2003) Molecular imaging in vivo: an introduction. Br J Radiol, 76
Spec No 2, S98-109.
244. Podoleanu, A.G. (2005) Optical coherence tomography. Br JRadiol, 78, 976-988.
245. Ntziachristos, V., Yodh, A.G., Schnall, M. and Chance, B. (2000) Concurrent MRI
and diffuse optical tomography of breast after indocyanine green enhancement. Proc
Natl Acad Sei USA, 97, 2767-2772.
140
246. Weissleder, R. and Ntziachristos, V. (2003) Shedding light onto live molecular
targets. Nat Med, 9, 123-128.
247. Joseph, S., Olbrich, C, Kirsch, J., Hasbach, M., Briel, A. and Schirner, M. (2004) A
real-time in vitro assay for studying functional characteristics of target-specific
ultrasound contrast agents. Pharm Res, 21, 920-926.
248. Weiler, G.E., Wong, M.K., Modzelewski, R.A., Lu, E., Klibanov, A.L., Wagner,
W.R. and Villanueva, F.S. (2005) Ultrasonic imaging oftumor angiogenesis using
contrast microbubbles targeted via the tumor-binding peptide arginine-arginine-
leucine. Cancer Res, 65, 533-539.
249. Wcissleder, R., Kelly, K., Sun, E.Y., Shtatland, T. and Josephson, L. (2005) Cell-
specific targeting of nanoparticles by multivalent attachment of small molecules. Nat
Biotechnol, 23, 1418-1423.
250. Wan, S., Parrish, J.A., Anderson, R.R. and Madden, M. (1981) Transmittance of
nonionizing radiation in human tissues. Photochem Photobiol, 34, 679-681.
251. Pelegrin, A., Folli, S., Bucheggcr, F., Mach, J.P., Wagnieres, G. and van den Bergh,
H. (1991) Antibody-fluorescein conjugates for photoimmunodiagnosis of human
colon carcinoma in nude mice. Cancer, 67, 2529-2537.
252. Matter, CM., Schüler, P.K., Alessi, P., Meier, P., Ricci, R., Zhang, D., Halin, C,
Castellani, P., Zardi, L., Hofer, CK. et al. (2004) Molecular imaging of
atherosclerotic plaques using a human antibody against the extra-domain B of
fibronectin. Circ Res, 95, 1225-1233.
253. Halin, C, Gafner, V., Villani, M.E., Borsi, L., Berndt, A., Kosmehl, H., Zardi, L. and
Neri, D. (2003) Synergistic therapeutic effects of a tumor targeting antibody
fragment, fused to interleukin 12 and to tumor necrosis factor alpha. Cancer Res, 63,
3202-3210.
141
254. Thorpe, P.E. and Ran, S. (2000) Tumor infarction by targeting tissue factor to tumor
vasculature. Cancer J, 6 Suppl 3, S237-244.
255. Fabbrini, M., Trachsel, E., Soldani, P., Bindi, S., Alessi, P., Bracci, L., Kosmehl, H.,
Zardi, L., Ncri, D. and Neri, P. (2005) Selective occlusion of tumor blood vessels by
targeted delivery of an antibody-photosensitizer conjugate. Int J Cancer.
256. Li, J., Hu, P., Khawli, L.A. and Epstein, A.L. (2003) LEC/chTNT-3 fusion protein for
the immunotherapy of experimental solid tumors. JImmunother, 26, 320-331.
257. Damle, N.K. (2004) Tumour-targeted chemotherapy with immunoconjugates of
calicheamicin. Expert OpinBiol Ther, 4, 1445-1452.
258. Bagshawe, K.D., Sharma, S.K. and Begent, R.H. (2004) Antibody-directed enzyme
prodrug therapy (ADEPT) for cancer. Expert Opin Biol Ther, 4, 1777-1789.
259. Brack, S.S., Dinkelborg, L.M. and Neri, D. (2004) Molecular targeting of
angiogencsis for imaging and therapy. Eur J Nucl Med Mol Imaging, 31,1327-1341.
260. Marty, C. and Schwendener, R.A. (2005) Cytotoxic tumor targeting with scFv
antibody-modified liposomes. Methods Mol Med, 109, 389-402.
261. Hainfcld, J.F. (1992) Uranium-loaded apoferritin with antibodies attached: molecular
design for uranium neutron-capture therapy. Proc Natl Acad Sei USA, 89, 11064-
11068.
262. Hoogcnboom, H.R., Griffiths, A.D., Johnson, K.S., Chiswell, D.J., Hudson, P. and
Winter, G. (1991) Multi-subunit proteins on the surface of filamentous phage:
methodologies for displaying antibody (Fab) heavy and light chains. Nucleic Acids
Res, 19,4133-4137.
263. Carnemolla, B., Leprini, A., Allemanni, G., Saginati, M. and Zardi, L. (1992) The
inclusion of the type III repeat ED-B in the fibronectin molecule generates
142
conformational modifications that unmask a cryptic sequence. J Biol Chem, 267,
24689-24692.
264. Harrison, J.L., Williams, S.C., Winter, G. and Nissim, A. (1996) Screening of phage
antibody libraries. Methods Enzymol, 267, 83-109.
143
Curriculum vitae
Personal details
First Name
Last Name
Date of birth
Place ofbirth
Address
Phone
Nationality
Marital Status
Dragan
Grabulovski
March 9, 1978
Winterthur, Switzerland
Imfeldstrasse 84a
CH-8037 Zürich
+ 4143 343 13 24
Swiss, Macedonian
Unmarried
Education
Nov 2004 - May 2007 PhD thesis: „The SH3 domain of Fyn Kinase as a Scaffold for
the Generation ofNew Binding Proteins", Institute of
Pharmaceutical Sciences, ETH Zürich, Supervisor: Prof. Dr.
Dario Neri
Oct 2000 - Mar 2005 Studies in „Didaktischer Ausweis", ETH Zürich
Graduated: Didaktischer Ausweis ETH in Pharmazie
144
Oct 1998 - Oct 2003 Studies in Pharmacy, ETH Zürich
Graduated: Eidg. Dipl. Apotheker ETH
(average mark: 5.8 out of 6)
Mar 2003 - Jul 2003 Diploma thesis: „Random synthesis: Isolation, structure
elucidation and activity screening of tyrosine kinase inhibitors
produced by y-irradiation", Supervisor: Prof. Dr. Gerd Folkers,
ETH Zürich
Oct 2002 - Feb 2003 Semester thesis: „Cloning and expression of mutants of LI9-
IFNy", Supervisor: Prof. Dr. Dario Neri, ETH Zürich
Aug 1991 - Jan 1998 Swiss Matura B, Kantonsschulc Rychenberg, Winterthur
Practical trainings
Mar 2004 Compulsory practical training as teacher in biology, Kantons¬
schulc Wettingen (Dr. S. Ginsburg)
Aug 2000 - Aug 2001 Compulsory one-year practical training, Adler Apotheke,
Winterthur (Dr. U. Reinhard)
Working experience
As ofMay 2007 Co-founder and CSO ofCovagen AG, a spin-off company of
ETH Zürich
145
Jan 2006 - May 2007 Saturday's Service as pharmacist,
Apotheke im Shoppingcenter, 8957 Spreitenbach
(Dr. R. Gysler)
Nov 2004 - Mar 2006 Sunday's Service as pharmacist, Apotheke
im Bahnhof, 8400 Winterthur (H. Sonanini)
Jan 2004 - Oct 2004 Pharmacist (80%), Apotheke im Bahnhof, 8400 Winterthur
(H. Sonanini)
Oct 2001 - May 2003 Saturday's and Sunday's service as
cand.pharm. (Apotheker-Assistent)
- Apotheke Scuzach, 8472 Seuzach (Dr. P. Steigrad)
- Apotheke im Bahnhof, 8400 Winterthur (H. Sonanini)
Awards
Feb 2006 Venture Competition 2006, ETH Zürich & McKinsey: One of
the ten best business ideas and business plans, in team with Dr.
Julian Bertschinger
Nov 2004 ETH medal for Diploma thesis
Nov 2003>nd2n place Amedis Förderpreis 2003 for the Diploma thesis
Publications & Patents
2007 Grabulovski. D. and Ncri, D. (2007) Vascular tumor targeting,
in Tumor angiogenesis, invited chapter, ed. Marme D &
Fusenig N, Springer Verlag, in press
146
Feb 2007
Feb 2007
Bertschinger, J.*, Grabulovski, P.* and Neri, D. (2007)
Selection of single domain binding proteins by covalent DNA
display. Protein Eng Des Sel, 20, 57-68.
*these authors contributed equally to the work
Grabulovski. P.. Kaspar, M. and Neri, P. (2007) A Novel,
Non-imrnunogenic Fyn SH3-derived Binding Protein with
Tumor Vascular Targeting Properties. JBiol Chem, 282, 3196-
3204.
May 2006 Patent application EP 06017336.6, Grabulovski D. and Neri D.
Mar 2005 Ebbinghaus, C, Ronca, R., Kaspar, M., Grabulovski, P..
Berndt, A., Kosmehl, H., Zardi, L. and Neri, D. (2005)
Engineered vascular-targeting antibody-intcrferon-gamma
fusion protein for cancer therapy. IntJ Cancer, 116, 304-313.
Languages
Macedonian
German
English
French
Serbian / Croatian
Spanish
Bulgarian
Native Speaker
Fluent
Fluent
Good knowledge
Good knowledge
Basic knowledge
Basic knowledge
Hobbies
Sports (Basketball, Snowboard, Swimming)
Zurich, May, 2007
147
7 Acknowledgements
Above all, I would like to express my sincere gratitude to my supervisor, Prof. Dr. Dario
Neri, who offered me an inspiring and exciting scientific environment for my thesis. When I
started with my PhD studies, you gave me a new, challenging and stimulating project being
different from your „corc"-research in the area of antibody phage and ESAC technology.
Despite this difference, you never left me alone with my scientific questions and you always
found some spare time to discuss my experiments and to bring in your ideas and critical
questions. I am very grateful for the constant support on a professional as well as personal
level. Particularly, I want to thank you for your assistance that Julian and 1 needed to start a
spin-off company and for your extremely kind support in any topic during the founding
phase. Especially throughout this time, I have got to know you as an extremely transparent,
fair and professional doctor father.
My special thanks go to Prof. Dr. Gerd Folkers for having kindly accepted to be my co-
examiner.
I am grateful to Julian Bertschinger for discussions, advices, critical inputs and for the co-
founding of our ETH spin-off company. I think that we are a good team and we will make out
of Covagen AG a success story. I further want to thank Simon Brack and Michela Silacci
who introduced me to phage display technology and library design strategies. Special thanks
go to Manuela Kaspar and Frank Bootz for their indispensable help in the animal
experiments. I further want to thank all other members of our research group for the nice
atmosphere in the lab. I also want to thank my former diploma student Jasmin Rauer for her
valuable help in the laboratory.
I thank all people who made the founding of Covagen AG possible, especially Dario Neri,
Julian Bertschinger, Rudolf Gygax, Corina Schutt and Mattias Beck.
I am most grateful to my parents, who helped and supported me all my life to reach my goals,
for their encouragement, trust and love.
148
