University of Southern California Dissertations and Theses

Dual effects of transmembrane proline residues on integrin function

(USC Thesis Other)

Dual effects of transmembrane proline residues on integrin function

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content i | P a g e
Dual Effects of
Transmembrane Proline
Residues on Integrin
Function
By
Thomas Schmidt
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the Requirements for the Degree
DOCTOR OF PHILOSOPHY
Genetics, Molecular and Cellular Biology
August 2015

ii | P a g e

Dedicated

Zu Meinem Opa,
Walter Schmidt

iii | P a g e

Acknowledgement

Foremost, I would like to express my sincere thanks and admiration to my advisor, Dr. Tobias S. Ulmer,
for giving me the opportunity to follow my desire for the unknown and express novel discoveries,
which resulted from my passion, in this thesis. He has been instrumental through his support, helpful
guidance, and encouragement to my academic and personal development. It was that him and his
girlfriend, Al, who supported me like no other when my son Otto was born through advice as well as
general care for my son's wellbeing. I am also thankful to my guidance and thesis committee members,
Dr. Ian Haworth, Dr. Ralf Langen, Dr. Robert Chow and Dr. Robert Farley. Especially Dr. Farley, who
invested his personal time to hone me in the art of scientific presentation and was open to give constant
scientific and personal recommendations. I thank the current and past members of the Ulmerlab for
helpful suggestions and amiable acquaintanceship which greatly contributed to a productive and
enjoyable graduate experience, particularly Alan Situ for his talented execution of experiment,
professional suggestions and friendship. I thank further members of the graduate program for technical
help and friendly discussions, especially Joey, Andy, Lindsey, Charles, Ayesha, and others. Special
thanks are due to my family in Germany and Singapore, especially my wife Jenn and my son Otto for
all the fun, encouragement and moral support.

Thank you.

iv | P a g e

Der Zauberlehrling
By Johann Wolfgang von Goethe

Hat der alte Hexenmeister
Sich doch einmal wegbegeben!
Und nun sollen seine Geister
Auch nach meinem Willen leben.
Seine Wort und Werke
Merkt ich und den Brauch,
Und mit Geistesstärke
Tu ich Wunder auch.
Walle! walle
Manche Strecke,
Daß, zum Zwecke,
Wasser fließe
Und mit reichem, vollem Schwalle
Zu dem Bade sich ergieße.
Und nun komm, du alter Besen,
Nimm die schlechten Lumpenhüllen!
Bist schon lange Knecht gewesen:
Nun erfülle meinen Willen!
Auf zwei Beinen stehe,
Oben sei ein Kopf,
Eile nun und gehe
Mit dem Wassertopf!
Walle! walle
Manche Strecke,
Daß, zum Zwecke,
Wasser fließe
Und mit reichem, vollem Schwalle
Zu dem Bade sich ergieße.
Seht, er läuft zum Ufer nieder!
Wahrlich! ist schon an dem Flusse,
Und mit Blitzesschnelle wieder
Ist er hier mit raschem Gusse.
Schon zum zweiten Male!
Wie das Becken schwillt!
Wie sich jede Schale
Voll mit Wasser füllt!
Stehe! stehe!
Denn wir haben
Deiner Gaben
Vollgemessen! -
Ach, ich merk es! Wehe! wehe!
Hab ich doch das Wort vergessen!
Ach, das Wort, worauf am Ende
Er das wird, was er gewesen!
Ach, er läuft und bringt behende!
Wärst du doch der alte Besen!
Immer neue Güsse
Bringt er schnell herein,
Ach, und hundert Flüsse
Stürzen auf mich ein!

Nein, nicht länger
Kann ichs lassen:
Will ihn fassen!
Das ist Tücke!
Ach, nun wird mir immer bänger!
Welche Miene! welche Blicke!
O, du Ausgeburt der Hölle!
Soll das ganze Haus ersaufen?
Seh ich über jede Schwelle
Doch schon Wasserströme laufen.
Ein verruchter Besen,
Der nicht hören will!
Stock, der du gewesen,
Steh doch wieder still!
Willst am Ende
Gar nicht lassen?
Will dich fassen,
Will dich halten
Und das alte Holz behende
Mit dem scharfen Beile spalten!
Seht, da kommt er schleppend wieder!
Wie ich mich nur auf dich werfe,
Gleich, o Kobold, liegst du nieder;
Krachend trifft die glatte Schärfe.
Wahrlich! brav getroffen!
Seht, er ist entzwei!
Und nun kann ich hoffen,
Und ich atme frei!
Wehe! wehe!
Beide Teile
Stehn in Eile
Schon als Knechte
Völlig fertig in die Höhe!
Helft mir, ach! ihr hohen Mächte!
Und sie laufen! Naß und nässer
Wirds im Saal und auf den Stufen:
Welch entsetzliches Gewässer!
Herr und Meister, hör mich rufen! -
Ach, da kommt der Meister!
Herr, die Not ist groß!
Die ich rief, die Geister,
Werd ich nun nicht los.
"In die Ecke,
Besen! Besen!
Seids gewesen!
Denn als Geister
Ruft euch nur, zu seinem Zwecke,
Erst hervor der alte Meister."

v | P a g e

Table of Contents

List of Figures ………………………………………………………………………………….. vi
Abstract………………………………………………………………………………………… vii
Introduction …………………………………………………….................................................. 1
Chapter 1 - “Integrin β
3
Lysine 716 Forms a Regulator of Integrin Signaling via
its ‘Snorkeling Mechanism’.” …………...…………………………………..….. 11
Chapter 2 - “β
3
(Ala711P) increases dimer formation due to increased glycine
packing in the OMC”……………………………………….......................……... 51
Chapter 3 - “N-terminal Proline Influences Structure to Function Relationship
in Integrin α
IIb
Transmembrane Region” …………………………………….… 80
Chapter 4 - Materials and Methods ……...………………….....……………………………. 119
Future Directions …………………………………………….……………………………... 126
References ………………………………………………..………………………………….. 129

vi | P a g e

List of Figures
Figure Page Title
Fig. 1-1 15 Sequence alignment of the β
3
TMD family.
Fig. 1-2 19 Lys716’s amide group interacts with the lipid head group.
Fig. 1-3 20 β
3
TMD changes its tilt angle upon exposure to anionic lipid system.
Fig. 1-4 24 Anionic lipids allow for increased dimer formation.
Fig. 1-5 29 Lysine alkylation introduced topological changes.
Fig. 1-6 31 Lysine alkylation increases α
IIb
β
3
TMD dimerization in POPC and POPG lipid
systems.
Fig. 1-7 35 Mutations of β
3
(Lys 716) disrupt the α
ΙΙb
β
3
TMD interaction.
Fig. 1-8 37 β
3
TMD K716E mutations significantly disturb β
3
TMD topology.
Fig. 1-9 41 β
3
TMD Lys716 mutation depicts topological changes.

Fig. 2-1 56 Ala711P stabilizes the α
ΙΙb
β
3
(K716A) TMD interaction and reduces integrin
activation.
Fig. 2-2 60 β
3
(Ala711P) mutation stabilizes the α
IIb
β
3
heterodimer (ITC).
Fig. 2-3 62 β
3
(Ala711P) mutation stabilizes the α
IIb
β
3
heterodimer (NMR).
Fig. 2-4 65 β
3
(Ala711P) mutation stabilizes the α
IIb
β
3
heterodimer (FRET).
Fig. 2-5 68 Proline introduced in the TMD forms a flexible kink.
Fig. 2-6 71 Structure of the integrin α
IIb
β
3
transmembrane complex.

Fig. 3-1 85 Sequence alignment of selected human integrin α and interleukin βc (IL3RB)
segments.
Fig. 3-2 89 The engineered GB3A-α
IIb
TMD peptide forms a valid system to characterize
structural properties in the linker region.
Fig. 3-3 91 Secondary structure rearrangement and its effect on the backbone dynamics of
the αIIb linker region.
Fig. 3-4 95 N-terminal proline 965 influences chemical surrounding of close by residues.
Fig. 3-5 96 Structural and topological comparison between and αIIb and its Pro965A/E
mutants.
Fig. 3-6 100 α
IIb
TMD Pro965A/E mutation does not significantly disturb α
IIb
/β
3
TMD
heterodimerization.
Fig. 3-7 104 N-terminal proline 441 influences chemical surrounding of close by residues.
Fig. 3-8 105 Structures and topological comparison of interleukin β
c
TMD and its Pro441A
mutation.
Fig. 3-9 109 N-terminal membrane-water interface residue restrains.

Fig. F-1 128 Arginine interacts with lipid headgroup.

vii | P a g e

Abstract

With 50% of all drug targets, membrane proteins present an important subcategory for
structural characterization, even though they code for less than a third of all proteins in the human
genome (Ulmschneider et al., 2005). Due to specific properties, the structure of membrane proteins are
not easily characterized by conventional methods, therefore NMR presents a unique tool for that
purpose. Integrins are a type I heterodimeric receptors that are essential for cell adhesion and
migration. Integrins are large heterodimeric membrane receptors that tie the extracellular matrix to the
cytoskeleton (Harburger and Calderwood, 2009). It is essential to both the arrest of bleeding at sites of
vascular injury and pathological thrombosis culminating in heart attack and stroke. Integrin bi ‐
directional transmembrane (TM) signaling involves the dissociation of a complex formed by the α–β
TM segments, which is accompanied by large rearrangements of its extracellular domains (Anthis and
Campbell, 2011).
TM segment sequences of the 8 β human subunits are well conserved, especially within a
distinct positive charged residue at the inner membrane clasp, β
3
(K716). In our research, we have
shown that charge variation of the lipid system influences K716’s amide-lipid interaction at the C-
terminal membrane interface, therefore defining a regulation mechanism for α
IIb
β
3
heterodimer
formation. Site-directed mutation of K716 caused integrin to achieve its high-affinity state by
stabilizing its monomeric species. Using directed evolution of β
3
(K716A), we identified a substitution,
A711P, to restore the integrin α
IIb
β
3
default low-affinity state. Quantitative dimerization analysis,
utilizing ITC, NMR and FRET, independently verified an increase in TM domain dimer formation
upon proline mutation, therefore recovering the low-affinity state. Structural analysis of the monomeric
β
3
(A711P/K716A) identified a kink of 30 ± 1° at the border of the outer and inner membrane clasps,
thereby decoupling the tilt between these segments. Subsequently, the integrin α
IIb
β
3
(A711P)
heterodimer revealed a strong glycine packing at the OMC compared to the wt, hence its complex
represents the first structure of a heterodimeric TM receptor of its kind and reveals a dimerization
interface of captivating complexity.
Utilizing NMR and MD simulation, we determined proline's extended role at the N-terminal
membrane-water interface of single-pass α-helical TM domains in integrin α
IIb
and interleukin β
c
. The
structural characterization of these mutants revealed proline's function as N-terminal helix-cap and
linker terminator. Therefore, α
IIb
(P965) establishes an equilibrium between the length of the TM helix
and its N-terminal linker domain, regulating signal transmission towards the extracellular domains.
These results describe signaling and activation tendencies of members in the integrin α family. Taken
together, the structural knowledge presented in this thesis provides insights to the disease mechanism
and will form a scientific basis for future therapeutic treatments of stroke and cancer.

1 | P a g e

Introduction

The goal of structural biology presents the elucidation of structure-to-function relationship in
biomolecules. No other field is influenced by such discoveries as protein chemistry, where the
development of protein structures is pushing the frontier for drug development. Success stories include,
but are not limited to, retroviral drugs and cancer medication targeting proteins in the kinase family (e.g.
methotrexate in complex with dihydrofolate reductase) (Baenziger et al, 2000; Merzlyakov et al, 2007).
Over the past decade, the structure development, and subsequently the structure-to-function
relationship, of membrane proteins has proven to be more valuable and difficult to tackle than all other
protein groups. Part of its interest resides in its pivotal role during physiological events, hence
membrane proteins form regulators of a wide variety of cellular processes such as cell signaling,
metabolism and homeostasis. Defects in transmembrane proteins result in some of the most known
diseases reaching from cystic fibrosis to familial hemiplegic migraine. Additionally, protein-membrane
interactions form regulators of important signaling events by embracing different lipid compositions,
therefore an impaired lipid-protein interaction could compromise functional regulation. Due to the
importance of membrane proteins in healthy and diseased physiological systems, they often form the
target of pharmacological substances, constituting ~50% of all drug targets even so the presents only
<30% of the entire protein coding genome (Ulmschneider & Ulmschneider, 2008). Despite its
physiological importance and relevance to the pharmacological industry, membrane protein structure
and function are poorly understood compared to aqueous soluble proteins. Overall, structural insights
of membrane proteins allow to form an understanding and appreciation for its role in physiological and
pathological signaling events. Specifically, it provides insights of initiation and transmission of
molecular signaling across the membrane, environmental modulators, and transmembrane protein-
protein and protein-lipid interactions. The difficulty of forming a molecular understanding of
membrane proteins is underlined by the ratio of overall unique protein structures solved (>20,000)

2 | P a g e

compared to the fraction representing membrane proteins (<500). The low fraction (~5%) emphasizes
the struggle to study membrane proteins by conventional methods, such as X-ray crystallography.
Difficulty of forming high resolution crystals is due to (1) the actual crystallization process and (2) to
form a well ordered crystal lattice. Crystal formation results in a well ordered lattice, however due to
the hydrophobic nature of membrane proteins, lipids/detergents are required for the protein to remain in
solution. The surrounding lipids will compete with the protein packing and such interfere with the
actual crystal formation. The second problem arises due to the purification of the transmembrane
section, as its hydrophobicity frequently results in impurities. The impurities will impair with the
crystal lattice, resulting in crystals with low resolution. Furthermore, of the few structures solved by X-
ray crystallography only a minor fraction include lipids imbedded in the crystal unit’s cell, questioning
their physiological relevance.
NMR forms a valid alternative to X-ray crystallography as it allows to calculate protein
structures at atomic resolution via specific nuclear spin interactions. However, its greatest advantage
remains in its ability to function under near physiological conditions, which includes a membrane
environment. In NMR the hydrophobicity of transmembrane proteins is mostly matched via detergent
micelles in place of lipid membranes, however we have previously shown that the difference in
detergent environment alters the protein conformation (Lau et al, 2008b). Therefore, in our studies
transmembrane proteins were presented in a bicellular environment; bicelles mirror lipid membrane
well due to its lipid bilayer geometry. Furthermore, NMR is able to sample dynamic conformational
and topological changes of the protein, and assess residue motilities under native lipid
conditions(Nietlispach & Gautier, 2011). The studies were implemented with molecular dynamics (MD)
simulation to form a molecular understanding of lipid-protein interactions.
One of the principal tools in the theoretical study of biological molecules is the method of
molecular dynamics simulations (MD). This computational method calculates the time dependent
behavior of a molecular system. MD simulations have provided detailed information on the

3 | P a g e

fluctuations and conformational changes of proteins and nucleic acids. These methods are now
routinely used to investigate the structure, dynamics and thermodynamics of biological molecules and
their complexes. Molecular dynamics simulations generate information at the microscopic level,
including atomic positions and velocities. The conversion of this microscopic information to
macroscopic observables such as pressure, energy, heat capacities, etc., requires statistical mechanics.
Statistical mechanics is fundamental to the study of biological systems by molecular dynamics
simulation, and has a great potential to be implemented in systems targeting membrane proteins.
The presented thesis exploits transmembrane signaling of single pass α-helical transmembrane
domains, with the major interest encircling the signaling mechanism of integrin α
IIb
β
3
. The research
reaches to identify various environmental molecular modulators that influence integrin signaling.
Residues, that regulate the signaling mechanism, were sought after allowing to formulate a working
understanding of integrin regulation. NMR in combination with MD simulation and FRET was utilized
to elucidate structural transitions between the two TM integin states, its monomeric and heterodimeric
state (Merzlyakov et al, 2006b).Therefore, integrin α
IIb
β
3
TM heterodimer forms a unique model
system for structural studies
It is a central dogma that it is necessary for most cells of a multicellular organism to interact
with the extracellular matrix. Cell migration plays a key role in a variety of biological processes,
including growth, development, and wound healing, therefore it commonly utilized such as the red
blood cells during thrombosis or leukocytes during the immune response (Hynes et al, 2002). However,
a strict regulation system must be in place to allow for a dynamic response to either adhere or
disengage from its surrounding when undergoing cell migration. The central position cell adhesion and
migration holds in normal physiology makes it vulnerable to a wide range of pathologies (e.g. cardio
vascular diseases and cancer) (Harburger & Calderwood, 2009). Normal adherent cells respond to
detachment by apoptosis (anoikis), however in cancer this mechanism has been altered to allow
metastasis to proceed. The following consequence of tissue invasion by cancer cells requires additional

4 | P a g e

adhesion ability to the counterpart of the previous detachment (Arnaout et al, 2005). The mediation
between the adhesion and migration in the pathological processes or their physiological counterparts
are regulated by the integrin protein family.
Integrin protein are “bridging” the gap between the interior of the cell and the extracellular
space by interacting with the extracellular matrix. Integrin are transmembrane receptors; the protein
crosses membrane and interacts with the ECM and actin cytoskeleton at the extra- and intracellular side
via adaptor proteins, respectively (Harburger & Calderwood, 2009). The overall anatomy of integrins
follows a common trend for all members of the family; the heterodimer consists of one α and β subunit,
which in turn exposes a several globular extracellular domains, a single transmembrane (TM) domain,
and relatively short cytoplasmic tail at the C-terminal (Arnaout et al, 2005). As consequence of the
versatile variety in integrin functions, the integrin family has evolved to several members, in mammals,
that includes 18 different α subunits and 8 different β subunits in humans (Hughes et al, 1996). Due to
its ability to form non-covalently linked heterdimers, 24 unique integrin proteins occupy their own
overlapping but nonredundant physiological function as shown by KO mice (Hodivala-Dilke et al,
1999; Hynes et al, 1999; Tsakiris et al, 1999).
With 24 integrin having nonredundant physiological function, a large interest is placed on its
function-to-structure relationship. The extracellular portion of the α and β subunit distinguish
themselves by both length and domain content. The β integrin subunit is about 700 residues in length
and consists of a PSI domain, a hybrid domain, four I-EGF domains, and a tail domain(Arnaout et al,
2005; Springer & Wang, 2004; Zhu et al, 2008). The α subunit is about 940 to 1,120 residues in length
and consists of a Propeller, a Thigh, Calf1 and Calf2 domain (Zhu et al, 2008). On the other hand the
cytoplasmic domains are rather small, where the α and β tail encompass a length of 15-78 and 46-70
amino acids (Wegener et al, 2007). The β tail is mostly unstructured, as previously shown, and presents
a high sequence similarity throughout the family; additionally, the general structure has a membrane-
proximal helix and a membrane-distal unstructured region. In comparison to the α tails, which MP

5 | P a g e

portion contains a highly-conserved GFFKR motif, the β MP portion consists of a two NPxY or NPxY-
like motifs (Rocco et al, 2008; Wegener et al, 2007). Interestingly, the β-tail serves as major hub for the
interaction between integrins and cytoplasmic proteins to allow for integrin-related signaling pathways.
The most appealing character of integrin proteins encircles its ability to transmit signals in both
direction of the membrane through allosteric means; the signal is either transmitted via an inside-out or
an outside-in signaling pathway (Arnaout et al, 2005; Harburger & Calderwood, 2009; Hughes et al,
1996). During an outside-in signaling event, an extracellular ligand triggers allosteric changes as well
as integrin clustering which ultimately results in the recruitment and activation of tyrosine kinases and
focal adhesion kinase (Hynes et al, 2002). On the contrary, during an insight-out signaling even,
integrin activation is adhered by the cytoskeletal protein talin (Hughes et al, 1996; Wegener et al,
2007). Talin functions by interrupting the interaction between the cytoplasmic and TM portion of the α
and β subunits resulting structural reorganization of the extracellular domains which will eventually
result in an increased binding to the ECM (Kim et al, 2009a; Kim et al, 2003; Wegener et al, 2007).
Due to its complex strategy, the insight-out signaling process is quite unique and cannot be found in
other physiological systems in nature.
Integrins play a central role in a variety of physiological processes, from growth and
development to haemostasis and leukocyte trafficking (Hynes & Hodivala-Dilke, 1999; Hynes et al,
2002; Tsakiris et al, 1999). As previously mentioned, integrins play an important role through the
interaction with ECM proteins, such as collagen, laminin, fibronectin and vitronectin (Hynes et al,
2002). The main strategy is to connect the ECM to the actin cytskeletion in large transient complexes,
which are itself composed of many integrins (focal adhesion). In a similar matter can epithelial cells
attach to the ECM via hemidesmosomes, which is characterized by integrin α
6
β
4
linking the basement
membrane laminin to keratin intermediate filaments. Others include the leukocyte specific β
2
and β
7
, in
which the β
2
allows for cell-cell adhesion via binding to the ICAM molecules on the cellular surface
(Takagi, 2002; Takagi et al, 2001a; Takagi et al, 2001b). Integrin β
3
is, upon expression, inactive and

6 | P a g e

remains so till activated during an immune response, at which point it can adhere to the vascular wall
and allow invasion of the surrounding tissue. Evolution has also developed crosstalk between integrins,
in leukocytes the binding of α
4
β
7
to extracellular ligands can trigger the activation of α
L
β
2
(Takagi et al,
2001a). Pathological defects in β
2
will cause in patients recurrent infections, impaired wound healing,
and leukocytosis (Hynes et al, 1999). More interesting is the structure and function of integrin α
IIb
β
3
,
which is expressed by RBCs in its low affinity state. Upon activation through the inside-out signaling
pathway, integrin α
IIb
β
3
will bind to fibrinogen and allow for thrombosis (Xiao et al, 2004). Even so the
activation pathway is relatively uncertain, several initiators include GPCRs and collagen. Pathological
interruption in the pathway can lead to Glanzmann's thrombasthenia, characterized by defective
clotting and excessive bleeding (Hynes & Hodivala-Dilke, 1999). Thus, leukocyte integrins are a
potential target for clinical therapies against inflammation and autoimmune diseases and α
IIb
β
3
is
already a target of effective antithrombotic drugs. However, the majority of the integrin proteins exist
in adherent state, therefore they are able to form a variety of adhesions ranging from weak to transient
focal complexes to rather sable adhesions (Bennett, 2005). Integrins participate in particularly strong
semi-permanent adhesions in the case of striated muscle, such as in costameres or the myotendinous
junction (Anthis et al, 2010).
Integrins, being by nature a rather complex structure, has only one representative of which the
entire structure exist at atomic resolution, α
IIb
β
3
. The integrin α
IIb
β
3
is the only of two intgrins for which
the extracellular structure is solved (the other being α
V
β
3
) and still the only one of which the TM
domain heterodimer is known (Zhu et al, 2007; Zhu et al, 2008). Furthermore, extensive studies have
revealed the unstructured cytoplasmic tail. The majority of structures known of the integrin family
treats the extracellular site, as it is readily accessible to X-ray crystallography; the TM domains and
cytosolic tails, however, are rather unraveled by NMR due to their greater flexibility and extended
complications arising due to the membrane environment (Anthis & Campbell, 2011). Additional work

7 | P a g e

on full-length α
IIb
β
3
by electron microscopy (EM) were able to identify large structural rearrangements
in the extracellular domains upon activation (Adair et al, 2005; Rocco et al, 2008).
Due to the hydrophobic environment of the membrane studies of the TM segment are rather
limited. Even so the integrin α/β integrin complex was observed via modeling studies, the only given
physiological relevant structure was solved by Lau et al (2009) (Lau et al, 2009). In 2008, Lau et al.
published structures of the α
IIb
and β
3
TM segments in lipid bicelles (Lau et al, 2008a; Lau et al, 2008b).
The structures of α
IIb
and its counterpart β
3
were unique thought specific characteristics. The 24 residue
long α
IIb
TMD segment forms a helix that sits vertically in the lipid bilayer, constrained by tryptophan
residues at each membrane interface (Lau et al, 2008a; Lau et al, 2008b). Furthermore, the C-terminus
is composed of the GFFKR motif (residues 991-995), which is highly conserved among α integrins.
Different from previous modeling methods, the two C-terminal phenylalanine are packed against the
helix upon association (C* et al, 2011; Lau et al, 2009). The β
3
TMD segment, on the other hand, forms
a longer 29-residue helix. Agreeing with hydrophobic mismatch, the helix was tilted by around 30°
from a vertical orientation within the membrane. As we will explore in the thesis, β
3
TMD surprisingly
has a positive charged lysine residue placed 5 amino acids into the hydrophobic membrane interior (C*
et al, 2011; Lau et al, 2008b). It is believed that the positive charge is able to snorkel through the
membrane to the more polar environment at the interface with the cytoplasm. The discovery of K716
actually be located in the interior of the membrane came as a surprise as it was prior to its discovery,
treated as start of the TM section. Post solving the monomeric structure Lau et al. became invested in
solving the NMR structure of the α
IIb
β
3
TMD complex in lipid bicelles (Lau et al, 2009). The dimer
structure revealed to be an exciting subject as individual segments exhibiting similar structures to what
has been found individually and crossing at an angle of about 30°(Lau et al, 2009). Two main sections
allow for the interaction to take place, the outer and inner membrane clasp. The outer membrane clasp
is stabilized via a glycine-mediated TM helix packing against aliphatic side chains (Lau et al, 2009).
The inner membrane clasp includes the D723/R995 cytoplasmic salt bridge and packing between αIIb

8 | P a g e

F992/F993 between the α and β helix (Lau et al, 2009; Lau et al, 2008b; Ulmer et al, 2001).
Specifically interesting was the cytoplasmic salt bridge, which was originally identified by mutagenesis
(Hughes et al., 1996) and figures centrally into the mechanism of integrin activation. This assembly
places the salt bridge between α
IIb
(R995) and β
3
(D723) within the relatively low dielectric environment
of lipid headgroups, particularly compared with aqueous solution. Thus, integrin α
IIb
β
3
forms a TM
dimer of unique structural complexity. The heterodimeric structure has given insights of the working
mechanism and explains why previous results on the cytosolic site gave rather dubious results.
Additionally, it provided the necessary argument that many modeling structures of coiled-coil are
rather wrong as they are missing several main features. Following studies tried to in-cooperate the
integrin activation via the extracellular domain. Zhu et al. used disulfide crosslinking to generate
restraints to model the structure of the α
IIb
β
3
TMD segment (Zhu et al, 2008). Thus, in the case of the
TM domains—as with the rest of the integrin heterodimer, atomic resolution structural data has greatly
advanced our understanding of integrin function. Taken in context with other studies, this achievement
has given great insight to the inside-out signaling pathway.
The integrin family mainly as attractive to many researchers due to their dual function by
allowing for an inside-out as well as outside-in signaling pathway. In regard to this thesis is, integrin
activation is solely presented via the inside-out signaling pathway (Kim et al, 2009a). Inside-out
signaling, in-vivo, is always activated via the binding of talin to the β tail. The actual binding event will
cause a separation of the α and β TMD and cytoplasmic regions, and these changes are propagated to
the extracellular domains, causing the integrin to adopt a conformation with higher affinity for the
extracellular matrix (Hughes et al, 1996; Partridge et al, 2005). The exact nature of the subsequent
extracellular changes remain unknown, however two prominent models include the switchblade-model
and the deadbolt model. In the switchblade-model large structural rearrangements directly modulate
integrin affinity. However in the deadbolt model only moderate conformational changes are needed for
integrin activation, and the extended conformation observed by EM rather refers to mobility (Adair et

9 | P a g e

al, 2005). The outside-in signaling pathway on the other hand can be activated through integrin
conformational changes, integrin clustering, mechanical force, or even a combination of these.
Additionally, divalent cations, particularly Mn
2+
, cause an increase in integrin affinity for extracellular
ligands. There are also very distinct differences between Mn
2+
-initiated activation and talin-induced
inside-out activation (Ye et al, 2008). The previously mentioned salt bridge between the α
IIb
R995 or β
3

D723 forms a very interesting regulation mechanism. Interestingly, mutation of α
IIb
R995 or β
3
D723 to
an oppositely-charged residue causes constitutive α
IIb
β
3
activation, however upon mutation of residues
to their opposite charges reverts them back to the wt activity (Lau et al, 2009; Ulmer et al, 2001).
Which makes sense in the picture of Lau et al heterodimer structure, as the salt bridge largely
contributes to the dimer stability of the IMC (Lau et al, 2008a; Lau et al, 2009; Lau et al, 2008b). Initial
studies to detect the salt bridge formation by NMR did not succeed as the TMD were missing, which is
necessary for electrostatic interaction. The studies on the the α
IIb
R995/β
3
D723 have advanced our
understanding on the insight-out signaling pathway, we now understand important driving forces for
the dimerization.
Follow up studies on the salt bridge formation has lead the Springer lab to conduct additional
disulfide crosslinking experiments on α
IIb
β
3
(Springer & Wang, 2004; Xiao et al, 2004). The results
proclaim that the cytosolic tail separation, found during the inside-out signaling, is required for the
outside-in signaling. The results showed that disulfide-linked TM domains were deficient in several
outside-in signaling activities. In agreement with the previous reported structure by Lau et al, the turn
at the C-terminus of the α
IIb
helix had placed part of the two phenylalanines of the GFFKR motif back
in the membrane (Lau et al, 2008a; Lau et al, 2009; Springer et al, 2008; Zhu et al, 2008). The point
that the two structures were derived independently places a great emphasis that they indeed present the
in-vivo structure.
Even so it was identified early-on that integrins undergo conformational change in conjunction
with signaling events, the presented evidence heavily relied on indirect methods. Early evidence was

10 | P a g e

based on FRET to identify a conformational change in αIIbβ3 in response to thrombin-induced inside-
out activation, or antibodies that recognize activated conformations therefore the nature of the
conformational change remain unknown to mankind (Hughes et al, 1996). Finally in 1995, upon
solving the structure of the I/A domains of integrin α, conformational changes were seen in response to
cation binding (Ye et al, 2008). Soon it became clear that the release of the interaction of the cytosolic
tails and the TMDs were responsible for inside-out integrin activation; however, a mechanism coupling
tail separation to extracellular ligand affinity remained elusive. Eventually in 2009 Anthis et al.
identified the separation mechanism upon talin binding through an electrostatic competition (Anthis et
al, 2009b).
Overall, although bidirectional integrin signaling is a complex and somewhat heterogeneous
process, good evidence exists for a switchblade mechanism of inside-out integrin activation. In this
mechanism the tail and, subsequent, TM separation induce major structural rearrangements, freeing the
extracellular domains to interact with the EMC (Springer et al, 2008). Studies have shown that the
integrin heterodimer exist in an equilibrium between two states: a closed low affinity state and an open
high affinity state. Interestingly, the high affinity state is stabilized through binding of extracellular
ligands and sequential separation of the TMDs and the cytosolic tail (Anthis et al, 2010; Goult et al,
2009). Inside-out activation, however, appears to strictly follow a switchblade mechanism, and studies
of integrin activation by talin have further clarified the molecular details of this process (Ye et al,
2008).
In our research we aim to establish a more detailed understanding in regulators of the TMD
separation upon insight-out signaling. It has come to our notice that the chemical environment
promotes integrin to uptake either the low- or high-affinity state. We believe that environmental
factors, such lipid composition hold key to fine tune integrin activity. Furthermore, a great interest
occupies the transmission of signal between the TMD and the extracellular domains, therefore we
investigate the linker region between them.

11 | P a g e

Chapter 1 – Integrin β
3
Lysine 716 Forms a Regulator of
Integrin Signaling via its “Snorkeling Mechanism.”

Abstract
Introduction
Results
K716 interaction with the lipid head group
K716-lipid interaction promotes dimer formation
Lys716 alkylation models POPG lipid environment
K716 mutations influence integrin activity via interfering with heterodimerization
Discussion
Conclusion

12 | P a g e

Abstract

Integrins are composed of α and β type I transmembrane subunits, and association between the
α and β TMDs partly regulates bidirectional transmembrane signal transduction. A conserved
lysine/arginine residue is located at the fifth amino acid near the inner TMD boundary that precedes an
additional hydrophobic patch, termed the ‘membrane proximal region.’ In accordance with previous
literature, lysine residues are frequently located at the membrane-water interface of TM proteins as
snorkeling moieties, and as such their positively charged primary amide group forms an electrostatic
interaction with the negatively charged lipid head group. Here we use NMR spectroscopy, MD
simulation, and fluorescence to show that integrin β
3
(Lys 716) regulates β
3
TMD topology. The α
ΙΙb
β
3

TMD heterodimer structure indicates that precise β
3
TMD crossing angles enable the assembly of outer
and inner membrane ‘clasps’ that hold the αβ TMD together to limit transmembrane signaling.
Variations in the lipid environment, anionic POPS, and POPG or net neutral POPC systems influence
both the lysine’s amide location in the membrane and the topology of the β
3
TMD, which eventually
results in a variation of associations between the β
3
and α
IIb
transmembrane domains. Therefore, we
evaluated the molecular and structural basis for this change by comparing an equilibrium constant via
chemical modifications of Lys716 and all atom MD simulations. The results suggest that Lys716 places
its amide group in the phosphate region if presented to POPC, but rather displaces it towards an
aqueous environment upon encountering anionic lipids. Furthermore, mutations in β
3
(Lys 716) caused
a dissociation of α
ΙΙb
β
3
TMDs and subsequent integrin activation by shifting the topology, as explored
by a bimane shift assay. Thus, widely occurring snorkeling residues in TMDs can help maintain TMD
topography and membrane-embedding, thereby regulating transmembrane signaling. Therefore, we
gathered convincing evidence that Lys716 forms a regulation mechanism which describes integrin
function in health and disease.

13 | P a g e

Introduction

Integrins are composed of α and β type I transmembrane subunits, and association of the α and
β TMDs regulates bidirectional transmembrane signal transduction (Arnaout et al, 2005). For most
metazoa, integrin β subunits contain a positively charged Lys or Arg near the inner TMD boundary
which precedes an additional hydrophobic patch, termed the ‘membrane proximal region’ (Figure 1-
1.A) (Ulmer, 2010). Disruption of the interaction between integrin α and β TMDs leads to allosteric
rearrangements that increases ligand-binding affinity of the extracellular domain (integrin activation)
and activation of inside-out signaling pathways (Hughes et al, 1996; Partridge et al, 2005). A stable αβ
TMD association, which is crucial in integrin regulation and physiological functions, requires the
simultaneous formation of two discrete assemblies: an inner and outer membrane clasp (IMC and OMC,
respectively) (Kim et al, 2009b). Because the β
3
TMD forms a continuous α-helix, its crossing angle
appears critical for the simultaneous assembly of these clasps. It is believed that the positively charged
lysine/arginine residue exposes a regulation mechanism by restraining the β
3
topology inside the
transmembrane section (Kim et al, 2009b). Lysine residues are frequently located at membrane-water
interfaces of TM proteins, in which they seem to function as a “snorkeling” moiety. It is surprising that
lysine, a charged residue, can occur in a relative hydrophobic environment as presented by the
membrane interior (Figure 1-1.A) (von Heijne, 1989). The free energy cost is minimized by placing the
long, flexible hydrocarbon part of the side chain within the hydrophobic interior of the membrane,
while the charged amino group is placed in the more polar interface region (van Klompenburg et al,
1997). Snorkeling residue functions have appeared for various membrane proteins, such as peripheral
and integral proteins (Conforti et al, 1990; Klepsch et al, 2011). Peripheral proteins (amphipathic
peptides) interact with the membrane surface by placing their hydrophobic residues towards the
membrane interior, and this wedging process is facilitated by snorkeling moieties, as shown by several
publications (Ambroso et al, 2014; Bugg et al, 2012; Jao et al, 2004; Jao et al, 2008; Varkey et al, 2013;

14 | P a g e

Woodham et al, 2012). More interestingly, transmembrane proteins utilize snorkeling residues by (1)
relieving hydrophobic mismatch and (2) modulating transmembrane topology (van Klompenburg et al,
1997; von Heijne, 1989). Hydrophobic mismatch describes the difference between the hydrophobic
length of the transmembrane segment and hydrophobic hydrocarbon region (Park & Opella, 2005;
Petrache et al, 2000). The decrease in bilayer thickness by disordering their acyl chains or rather
straightening them out, has specific effects on the TMD topology. The tilt and conformational changes,
which allow the side chains to orientate, reduce the energy landscape of the transmembrane section
(Park & Opella, 2005). The latter option is facilitated by the snorkeling mechanism, which effects are
frequently found for lysine and arginine residues; therefore, snorkeling residues function as regulators
of TM protein topology. Here we show that this mechanism additionally functions as a regulator of
single-pass transmembrane protein interactions.
In this chapter, we describe the molecular and structural basis of integrin α
IIb
β
3
activation upon
exposure to an anionic lipid environment, as is the case in the inner leaflet of red blood cells. The
regulating mechanism is based on the “snorkeling” capabilities of K716, which, if presented to
zwitterionic, net neutral lipid POPC, will interact with its negatively charged phosphate lipid head
group (van Klompenburg et al, 1997). To our surprise, upon exposure to anionic lipids, such as POPG
and POPS, lysine’s amide group discriminated against the interaction with the phosphate group, and
rather forms an electrostatic interaction with the carboxyl group placed at the membrane-water
interface. The resulting dimerization rates were able to be mimicked by forming a tertiary amide
(K716-(CH
3
)
2
) via an alkylation reaction. Finally, we were able to describe the stability of dimerization
in integrin α
IIb
β
3
TMD upon mutating Lys716 to Ala, Leu, or Glu by shifting its tilt or rotation angle.
Those findings describe the regulating mechanism on a molecular level exposed by integrin β
3
(K716),
which will hopefully lead to the formation of therapeutics to treat illnesses, such as cancer and stroke,
in which integrins have been implicated.

15 | P a g e

Figure 1-1 Sequence alignment of the β3TMD family. The transmembrane domains span a distance
of roughly 28 amino acids. A conserved positively charged Lysine/Arginine is found five amino acid
into the hydrocarbon interior from the cytosolic side of the membrane. Conserved amino acids are
colored by the JalView multiple alignment editor using the ClustalX color scheme. The N- and C-
terminal membrane-water interface of the monomeric β3TMD subunits are depicted.

16 | P a g e

Results

MD simulation and Solid State Data (Contributed by Dr. Christian V. Grant)
All atom MD simulations were performed for 30ns with 3 different starting conditions in POPC
and POPS environment having slight variation in tilt and displacement (Elmore, 2006; Gumbart et al,
2005). It is expected that the observed lysine’s positive charged amide interacts with a negative charge
placed at the lipid head group (Deol et al, 2006). The simulations observed two different lipid
environments: (1) the net neutral lipid POPC and (2) the anionic lipid POPS. Previous studies
determined that anionic lipids allow an increase in integrin inactivation by promoting
heterodimerization (Bogdanov et al, 2009; Shahidullah & London, 2008; Suk et al, 2012). Lysine 716
establishes itself to function as a regulator of integrin singling in a lipid dependent manor; our research
presents that the variation in lipid properties at the membrane water interface favor diverse low energy
population of Lys716’s amide group. In case of a net neutral lipid environment, such as POPC, lysine
is located in close proximity to the phosphate head group. In the simulations, Lys716’s primary amide
group was found in ~80% of the time at a close distance (≤4Å) to the POPC lipid phosphate head group
(Figure1-2.A). The preferred distance of ~2Å to the phosphate group occurred for ~60% of all frames
(Figure 1-2.B). However, minor distances occurred between 4-6Å, with each being around 10% of the
population. Interestingly, in the anionic POPS system Lys716 preferred to interact with the negative
charged carboxyl moiety of the lipid head group rather than the phosphate group. In fact, the amide
group had almost null interaction with the phosphate group, but occurred in ~70% of the simulation
time in close distance to the carboxyl group. Similar to POPC, the preferred distance for the primary
amide in the lysine side chain to the lipid’s carboxyl group is ~2Å, occurring at ~70% of all the frames.
However, minor populations occur at 6-9Å with <10% occurrence (Figure 1-2.A&B). Therefore,
Lys716 occurs to prefer the interaction with the negative charge placed at lipid’s phosphate head group
if faced with net neutral zwitter ionic lipids. The attraction is weakened in the POPS environment, in

17 | P a g e

which the primary amide migrates towards the lipid’s carboxyl group (Figure 1-2.C).
The two different lipid systems repeatedly appeared stable as no large deviation
occurred in the RMSD value presenting the C
α
carbon of the protein backbone for all three starting
conditions (Figure 1-3.A). The validity of the simulation system is important as we were able to extract
information treating global topological behavior, such as tilt angle. In Figure 1-3.B, the average tilt
angles for the three sets of β
3
TMD in POPC and POPS are compared suggesting a similar tilt in POPC
and POPS bilayer. The tilt angle observed for both membrane systems corresponded to ~40°. The
relatively unaffected angle can be explained via hydrophobic mismatch, in which case the hydrophobic
region of the TM helix matches the hydrocarbon region of the membrane bilayer (Park & Opella, 2005).
This effect builds up a relative large energy barrier which is unaffected by the snorkeling mechanism
(~0.7 kcal/mol). The difference in integrin population between neutral and anionic lipid environment
can attributed to other topological changes, such as rotation angle or lipid dispersion (Figure 1-3.C).
The shift in location of the lysine amide, upon change in lipid environment, causes a force that
influences the embedding environment of the β
3
TMD through various topological changes. The
reported tilt angle was further confirmed by solid state NMR experiment, shown in Figure 1-3.D.
Helical tilt was estimated by comparing the experimental spectrum with calculated helical wheels
generated from an ideal α-helix model using Ramachandran angles of φ = -57, ψ = -47 (Cross & Opella,
1980; Grant et al, 2009; Nevzorov & Opella, 2007). By applying PASIMA based experiments the tilt
angle was estimate to encircle ~40°, which is similar to the previously noted tilt determined from the
MD simulation. Cross Polarization (CP) 1D experiments gave insights to the tilt behavior upon dimer
formation (Figure 1-3.E) (Park et al, 2006b). The tilt angles were determined by comparing the CP
spectrum with chemical shift projections of ideal helical wheels (Nevzorov & Opella, 2007). In this
Figure the two spectra compare the CP 1D spectra of the β
3
TMD alone (lower) or associated with
unlabeled α
IIb
TMD (upper), each fitted with the ideal helical wheel projection. Similar to the 2D
SAMPI4 spectrum, the CP 1D spectrum of β
3
TMD resulted in a tilt of ~40°, however if associated with

18 | P a g e

α
IIb
TMD the tilt angle decreases to ~30° (Park et al, 2006a; Park et al, 2006b). Interestingly, the tilt
change occurring upon integrin association should be related to the topological changes introduced
upon Lys716 migration, therefore Lys716 might lower the energy barrier to form the integrin
heterodimer.

19 | P a g e

Figure 1-2 Lys716’s amide group interacts with the lipid head group. (A) MD simulations of
β
3
TMD in POPC environment concluded that in 80% of the simulation time it forms electrostatic
interactions with the phosphate head group. However, POPS environment shifts the interaction away
from the phosphate group and rather forms an electrostatic interaction with the negatively charged
carboxyl group in the lipid head group region. (B) In POPC, the majority of interactions occur at a
distance of ~2Å with a smaller population at up to ~5Å. Upon performing the simulations in POPS
environment, the distance increases beyond 10Å, surpassing the limit for electrostatic interactions.
However, a new interaction occurs between the negatively charged carboxyl groups in the lipid head
region with a majority at a distance of ~2Å. (C) The illustration depicts the interaction of the lysine’s
amide group with the phosphate head group of POPC. Lysine interacts with POPS lipid (lower figure);
the interactions rather occur with the lipid’s carboxyl group compared to the phosphate group in POPC,
as shown previously.

20 | P a g e

21 | P a g e

Figure 1-3 β
3
TMD changes its tilt angle upon exposure to anionic lipid system. (A) The protein
does not show elaborate deviation in the backbone structure as shown by the RMSD of the C
α
.
Therefore the AA MD simulation in the 2 different lipid system with 3 starting condition appear to
behave similar as observed by NMR. (B) The tilt angle was defined as the deviation of the protein
backbone from the membrane normal, which is orthogonal to the membrane-water interface. The tilt
angle in a POPC and POPS lipid environment was ~40°. (C) The illustration depicts topological
changes occurring upon exposure to either POPC or POPS bicelles. Initially, when exposed to POPC
lipids, Lys716 interacts with the lipid phosphate head group; however, upon exposure to POPS, Lys716
shifts outwards to rather interact with the negative charge placed on the lipid’s carboxyl group. The
difference in lysine location causes a shift in topology to form the heterodimer species. (D) 2D
SAMPI4 spectrum with superimposed 40˚ PISA wheel simulation for the β3TMD subunit which fits to
a ~40˚ tilt angle (E) 1D cross polarization spectra with an 80 µs contact time for (upper spectrum)
15
N
labeled β
3
TMD subunit in the presence of the unlabeled α
IIb
subunit and (lower spectrum) the
15
N
labeled β
3
TMD subunit alone. The superimposed fits are chemical shift projections of ideal PISA
wheel simulations indicating a tilt angle of 30˚ in the presence of the α
IIb
subunit (upper) and 40˚ for the
beta subunit alone (lower).

22 | P a g e

Titration with anionic lipids
In MD simulation, we have shown that the amide group, if presented to anionic lipids, migrates
further into the hydrophilic section of the membrane interface. This amide migration will lead to
specific topological changes which effect the rotation angle and lipid dispersion. The dimer formation
and the POPS/POPG associated lysine migration appear to be associated; therefore, one would see an
increase of heterodimer formation upon exposure to various lipid systems. In Figure 1-4.A, three
different lipid systems were compared. The net neutral, zwitterionic POPC lipid was compared to the
anionic lipids POPS and POPG, which are expected to modify β
3
TMD topology based on the induced
lysine migration. The equilibrium constants were determined by ITC titrations between the β
3
and
α
IIb
TMD in a bicellular environment with a q-factor of 0.39 (Tellinghuisen, 2008; Turnbull & Daranas,
2003). ITC measurements, studying bicelles that incorporate POPG, confirmed the anionic lipid-
mediated complex stabilization of the α
IIb
β
3
heterodimer (Table 1.1, Figure 1-4.A). Interestingly, POPS
was slightly more stabilizing than POPG, showing that the charge distribution of the lipid head group
contributes to protein-lipid interactions. However, the presented method suffers from solely observing
the heat exchange instead of observing the direct population differences of the dimer and monomer
fraction (Chou et al, 2004). Therefore we employed NMR titrations to reconfirm the observed
phenomena.
The NMR peaks, corresponding to TMD monomers and heterodimer of integrin α
IIb
β
3
, are
simultaneously observable at 28° C. Homodimerization of α
IIb
and β
3
is absent at the selected solution
conditions. This exchange behavior permits the direct extraction of the α
IIb
β
3
dimer fraction by
quantifying the peak volume decline of monomeric residues (Bocharov et al, 2012; Suk et al, 2012).
Specifically, we maintained the fixed concentration of
2
H/
15
N-labeled β
3
peptide and varied the
concentration of the unlabeled, partnering α
IIb
peptide to determine K
XY
(Figure 1-4.B&D). Isotope
labeling of either α
IIb
or β
3
yielded analogous results. Surprisingly, a POPS-mediated α
IIb
β
3
complex
stabilization of at least 0.22 ± 0.02 kcal/mol was apparent at a POPC:POPS=2:1 ratio, which

23 | P a g e

approximates the ratio of anionic lipids in the intracellular membrane leaflet of RBC. The dimer
fraction behaves as a linear function of the POPS ratio (Table 2 and Figure 1-3.C), either indicating that
the anionic lipid-binding site on α
IIb
β
3
was not yet saturated or that other residues contribute to the
dimerization mechanism.

24 | P a g e

25 | P a g e

Figure 1-4 Anionic lipids allow for increased dimer formation. (A) ITC measurements in
phospholipid bicelles consisting of 43 mM DHPC and 17 mM of the depicted long-chain lipid
(q
eff
=0.5). A starting β3 concentration of 10 µM was used. (B) Measurement of α
IIb
+ β
3
⇄ α
IIb
β
3

equilibrium constant. NMR measurements in phospholipid bicelles consisting of 400 mM DHPC and
120 mM of the depicted long-chain lipid (q
eff
=0.31). The disappearance of the monomeric αIIb(G972)
and β3(G702) resonances is plotted as 1–V
M
/V
M,0
where V
M,0
denotes the resonance volume of pure
monomer at 100 μM and V
M
is the residual monomer volume at increasing concentrations of the
partnering peptide. Peptide concentrations are expressed as peptide-to-lipid ratios. (C) Free energy
change, ΔG°, of α
IIb
β
3
association as a function of the depicted POPS-to-POPC ratio. (D) Schematics
describing the heterodimer formation starting with the monomer form of β
3
TMD (blue) and α
IIb
TMD
(green) in a bicelle environment.

26 | P a g e

Table 1-1 Thermodynamic Parameter of IIb 3 TMD Association Measured by ITC.
Peptides Lipid K
XY
a

H
o

[kcal/mol]
S
o

[kcal/mol]
G
o

[kcal/mol]

IIb
+ 
3

POPC 3820 ± 70 -16.0 ± 0.1 -11.0 ± 0.1 -4.91 ± 0.01

IIb
+ 
3
POPG 7000 ± 200 -18.9 ± 0.2 -13.6 ± 0.2 -5.09 ± 0.02

IIb
+ 
3
POPS 8800 ± 300 -19.4 ± 0.3 -13.9 ± 0.3 -5.44 ± 0.02
a
Quoted are values for measurements in 43 mM DHPC, 17 mM of the indicated long-chain lipid, 25
mM NaH
2
PO
4
/Na
2
HPO
4
pH 7.4 at 28° C. The presence of 9 mM free, non bicellar DHPC is noted.

Table 1-2 Thermodynamic Parameter of IIb 3 TMD Association Measured by NMR.
Peptides Lipids K
XY
a

G
o

[kcal/mol]

IIb
+
2
H/
15
N- 3 POPC 2070 ± 40 -4.57 ± 0.01

IIb
+
2
H/
15
N- 3
2 POPC:
1 POPS
3000 ± 100 -4.79 ± 0.02

IIb
+
2
H/
15
N- 3
1 POPC:
2 POPS
3400 ± 200 -4.86 ± 0.04

IIb
+
2
H/
15
N- 3 POPS 3800 ± 100 -4.93 ± 0.02
a
Quoted are values for 3(Gly702) in 400 mM DHPC, 120 mM of the indicated long-chain lipids, 25
mM HEPES pH 7.4, 0.02% NaN
3
at 28° C. The presence of 9 mM free, non bicellar DHPC is noted.

27 | P a g e

Lysine 716 alkylation models anionic lipid environment
As shown in the MD simulation, K716 is expected to interact with the lipid phosphate head
groups on the inner membrane interface, enforcing a specific topology on the transmembrane domain.
The strength of this electrostatic interaction is regulated by coulombs law (F=kq
1
q
2
/r
2
), which dictates
that the force attracting the two charged particles is reversely related to the distance between them.
Since the amide group is separated from the negatively charged phosphate oxygen by a proton, the
short distance (~1Å) results in a relatively strong bonding interaction, which is evident in Figure 1-2.B.
The electrostatic interaction was modulated by alkylating the primary amide of the Lys716 two methyl
groups, transforming the primary amide into a tertiary amide. The resulting distance to the opposite
charge increases due to the volume uptake of the methyl group (~3Å), resulting in a decreased
attraction between the two particles (Figure 1-5.D). Furthermore, due to its retained positive charge, the
tertiary amide will be actively excluded from the interior of the hydrophobic polycarbon region,
therefore modeling a similar environment as found under anionic lipid conditions. In the PRE
protection assay, the N-terminal membrane border is shifted by ~1 amino acid in the i-1 direction,
embodying topological changes taking place (Lau et al, 2008b). More interestingly, the C-terminal
membrane-water interface is disturbed upon alkylation, visualized by a sharp spike in the residue post
the alkylated Lys716. The weakened salt bridge and the increase in volume by the methyl group results
in a soft exposure of the positively charged tertiary amide by the hydrophobic lipid environment. This
wedge inside at the membrane interface results in a distortion of the PRE pattern. Furthermore, the
force exerted on the protein results in a shift of residues surrounding Lys716, exposing them to
different chemical environmental conditions. This change can be recorded in the TROSY spectrum, in
which the corresponding peaks are shifted in the amide and proton dimension, whereas peaks
unaffected by the methylation remain stationary. Interestingly, small shift changes for residues located
at the N-terminal membrane interface are observed, in addition to the methylation site. The chemical
shift change was quantified by measuring the difference between the wt and its methylated counterpart.

28 | P a g e

The largest shift change occurred at the methylation site, but minor changes are also seen at the
membrane interfaces. It can be concluded that distortion of the “snorkeling” effect between the
positively charged Lys716 amide group and the negatively charged phosphate lipid head group,
induced upon methylation, has a slight effect on the protein embedding and the conformation of the
β
3
TMD peptide.
Previously we determined that small changes in β
3
TMD’s topology take place upon disturbing
the snorkeling Lys716 via its methylation to a tertiary amide. The expulsion and complimentary force
exerted on the TM section of β
3
by the tertiary amide, reassembles the snorkeling preference of the
primary amide for the carbonyl in the lipid head group upon exposure to anionic lipids (POPS).
Therefore it is suggested that due to this force, an increase in dimer formation will be observed which
contains anionic lipid characteristics. The interaction between the α
IIb
and β
3
subunit was quantitatively
examined by ITC. In Figures 1-6.A, we compare two different lipid systems: the net neutral lipid POPC
and the anionic lipid POPG as previously presented. It appears that similar to the previously determined
POPS, POPG increases heterodimer formation, partly due to Lys716’s snorkeling mechanism. As
expected from the literature, the anionic lipid promotes heterodimerization compared to the net neutral
POPC system. Additionally, we compared the heterodimer formation of α
IIb
with either the β
3
wt
variant or its K716 alkylated counterpart. In the case of the POPC lipid system, the dimerization
increased upon K716 alkylation, resulting in a ΔG of -5.13, translating to an increase of 0.26 kcal/mol
compared to the non-alkylated peptide. However, in the anionic POPG lipid system, the dimerization
further increased to ΔG of -5.41 upon alkylation, which closely models the dimer formation in POPS
environment (ΔG
POPS
=-5.44). Lys716 appears to form a switch, which partly regulates the integrin
heterodimer formation via its snorkeling interplay with its lipid surroundings by discriminating
between specific properties at the membrane-water interface.

29 | P a g e

30 | P a g e

Figure 1-5 Lysine alkylation introduced topological changes. (A) The
15
N
H
and
1
H
N
chemical shift
differences between the methylation states located topological changes on the C- and the N-terminal
water-membrane interface. (B) The PRE protection assay determines a slight shift of ~1 amino acid in
the i+1 direction and a membrane distortion/break in the C- and N-terminal membrane-water interface.
(C) The TROSY spectra of β
3
TMD with and without a dimethyl addition to K716 presents valuable
information regarding structural, topological, and dynamical changes introduced upon methylation. (D)
The original 40° tilted β
3
TMD helix allows its positively charged lysine amide to interact with the
negatively charged phosphate lipid head group. However, upon methylation of K716, the increased
distance weakens this interaction by ~20% and the spatial increase, correlated with the methyl group,
distorts the lipid packing as it actively tries to exclude the positive charge from the hydrocarbon tail.
The excluded amide group therefore causes an uptake of a new conformation (tilt or rotational angle)
effecting both the C- and N-terminal end of the helix.

31 | P a g e

32 | P a g e

Figure 1-6 Lysine alkylation increases αIIbβ3TMD dimerization in POPC and POPG lipid systems.
(A) ITC data allows to quantitate the equilibrium constant by measuring the heat exchange upon
association. The titration curves represent the volume change compared to its molar ratio of the titrant
to substrate. Titration curves compare the β3TMD association with αIIb in its wt to its K716
dimethylated state in both POPC and POPG environment. (B) The resulting equilibrium constants,
entropy and enthalpy, give rise to ΔG. The ΔΔG in the POPC bilayer is ~0.24 kcal/mol, therefore
shifting the equilibrium towards the dimeric species upon methylation. The Lys716 methylation shifts
the dimer equilibrium close to POPG conditions. However, ΔΔG in the POPG bilayer is ~0.26 kcal/mol,
shifting the dimer/monomer equilibrium close to POPS conditions. The data supports lysine’s active
role in regulating heterodimer formation. (C) Schematics describing the heterodimer formation starting
with the monomer form of β3TMD (blue) and αIIbTMD (green) in a bicelle environment.

33 | P a g e

Table 1-3 Thermodynamic Parameter of IIb 3 TMD Association Measured by ITC.
Peptides Lipid K
XY
a

H
o

[kcal/mol]
S
o

[kcal/mol]
G
o

[kcal/mol]

IIb
+ 
3
POPC 3820 ± 70 -16.0 ± 0.1 -11.0 ± 0.1 -4.91 ± 0.01

IIb
+ 
3
POPG 7000 ± 200 -18.9 ± 0.2 -13.6 ± 0.2 -5.09 ± 0.02

IIb
+ 
3
-(CH
3
)
2
POPC 8800 ± 300 -19.4 ± 0.3 -13.9 ± 0.3 -5.15 ± 0.02

IIb
+ 
3
-(CH
3
)
2
POPG 8800 ± 300 -19.4 ± 0.3 -13.9 ± 0.3 -5.41 ± 0.02

IIb
+ 
3
POPS 8800 ± 300 -19.4 ± 0.3 -13.9 ± 0.3 -5.44 ± 0.02
a
Quoted are values for measurements in 43 mM DHPC, 17 mM of the indicated long-chain lipid, 25
mM NaH
2
PO
4
/Na
2
HPO
4
pH 7.4 at 28° C. The presence of 9 mM free, non bicellar DHPC is noted.

34 | P a g e

Lys716 mutation effects Integrin activity (C. Kim and M. Ginsberg collaborators)
For the capture assay, an αΙΙb mini-integrin bait containing the TMD and cytoplasmic tail of α
ΙΙb

(Figure 1-7.A) joined to a C-terminal tandem affinity purification (TAP) tag for rapid efficient
purification was expressed with prey comprised of the extracellular domain of the Tac (IL-2 receptor α)
joined to the TMD and tail of β
3
or β
3
-bearing Lys 716 substitutions (Figure 1-7.A) . When the cells
were lysed and bait were captured using calmodulin beads, we found that the α
ΙΙb
bait captured the β
3

prey, as expected; however, neutral (Ala), polar (Cys, Ser), acidic (Glu), or hydrophobic (Leu)
substitutions at β
3
(Lys 716) blocked the α
ΙΙb
β
3
TMD association. In contrast, a basic amino acid
substitution (Arg) did not disrupt the association, consistent with the idea that a snorkeling residue in
this position is required for the formation of the α
ΙΙb
β
3
TMD complex. To examine the potential effects
of β
3
(Lys 716) mutations on transmembrane signaling, we assayed their effects on the affinity state of
integrin α
ΙΙb
β
3
by measuring binding to an activation-specific α
ΙΙb
β
3
antibody (PAC1) as in Figure 1-7.B.
The results precisely correlated with the effects on αβ TMD interaction: all substitutions, with the
exception of Arg, led to spontaneous integrin activation (Figure 1-7.C). Thus, the loss of a conserved
basic residue in integrin β TMDs leads to disruption of the α–β TMD interaction and spontaneous
transmembrane signaling.

35 | P a g e

Figure 1-7 Mutations of β
3
(Lys 716) disrupt the α
ΙΙb
β
3
TMD interaction. (A) CHO cells were co-
transfected with α
ΙΙb
TMD-tail construct fused with C-terminal TAP tag, α
ΙΙb
TM-TAP, and β
3
TMD-tail
constructs fused with N-terminal Tac extracellular domain, Tac-β
3
TMD, bearing various mutations at
β
3
(K716), as indicated. α
ΙΙb
TM-TAP proteins were isolated and associated Tac-β3TM was detected by
western blotting (upper panel). Expressed Tac-β3TM proteins (middle panel) and captured α
ΙΙb
TM-TAP
proteins (bottom panel) are also shown. (B) β3(K716E) activates integrin α
ΙΙb
β
3
. CHO cells stably
expressing integrin α
ΙΙb
(CHO/α
ΙΙb
) were transiently transfected with wild-type (WT) integrin β
3
or
β
3
(K716E). Eighteen hours later, surface expression (D57 binding) and affinity of α
ΙΙb
β
3
(PAC1 binding)
were analyzed. Geometric means of PAC1 binding in cells expressing different quantities of α
ΙΙb
β
3
were
plotted as larger red dots. (C) Multiple β
3
(K716) mutations activate integrin α
ΙΙb
β
3
. CHO/α
ΙΙb
cells were
transfected with integrin β
3
bearing different mutations in the K716 residue as indicated. The geometric
means of PAC1 binding to those CHO/α
ΙΙb
cells were plotted against D57 binding. [Contributed by Dr.
Ginsberg at UCSD]

36 | P a g e

K716E changes TMD tilt
The effects of the K716X mutation are interesting as it allows for a constant active integrin
population and overwrites the otherwise inactive integrin species. Specific interests surround the
Lys716E mutant as it replaces the positively charged lysine residue with a negatively glutamate residue,
and as such, reversing its polarity at the C-terminal membrane interface. Chemical shift changes of
K716E relative to the wild type, as isolated from its TROSY spectrum in Figure 1-8.B, suggest
topological rearrangement rather than changes in secondary structure. Figure 1-8.C compares
13
C
α

secondary chemical shifts between the TMD region of the wt integrin β
3
and β
3
(K716E) TMD.
Secondary
13
C
α
chemical shifts, defined as the difference between the observed and tabulated random-
coil
13
C
α
shift of a residue, correlate with the backbone conformation. The minor shift differences
between the β
3
variant demonstrate the absence of any significant rearrangements in secondary
structure conformation or content. However, differences are seen in the embedding between β
3
(Lys 716)
and its glutamate mutant. The TROSY cross-peak signal intensity of a residue in the presence and
absence of 1 mM Mn
2+
EDDA
2−
in the aqueous phase, I/I
0
, was measured to quantify accessibility to the
paramagnetic reagent. It appears that the TM section of the helix shortened from its initial 28 amino
acids in the wt to ~23 residues in β
3
(K716E). The predicted interaction of Lys716 side chain ε-NH
3
+

with a lipid’s PO
4
−
group (red) and the interaction of glutamate’s γ-COO
−
with a POPS lipid’s amino
NH
3
+
group (blue) are illustrated in Figure 1-8.E, upon which the glutamate’s negative charge prohibits
it from interacting with the phosphate group in the lipid head region.

37 | P a g e

38 | P a g e

Figure 1-8 β
3
TMD K716E mutations significantly disturb β
3
TMD topology. (A) Chemical shift
changes of K716E relative to the wild type as isolated from its TROSY spectrum in B. (C) Comparison
of
13
C
α
secondary chemical shifts between the TMD region of wild type integrin β
3
and β
3
(K716E)
TMD. Secondary
13
C
α
chemical shifts, defined as the difference between the observed and tabulated
random-coil
13
C
α
shift of a residue, correlate with the backbone conformation. The minor shift
differences between the β3 variant demonstrate the absence of any significant rearrangements in
secondary structure conformation or content. Chemical shifts were measured in bicelles composed of
350 mM DHPC, 70 mM POPC, and 35 mM POPS. (D) Mutation of β
3
(Lys 716) changes TMD
membrane embedding. The TROSY H–N cross-peak signal intensity of a residue in the presence and
absence of 1 mM Mn
2+
EDDA
2−
in the aqueous phase, I/I
0
, was measured to quantify accessibility to the
paramagnetic reagent. Experiments were performed in duplicates using independently prepared
samples and quote the standard error between data sets. (E) The predicted interaction of Lys716 side
chain ε-NH
3
+
with a lipid’s PO
4
−
group (red) and the interaction of glutamate’s γ-COO
−
with a POPS
lipid’s amino NH
3
+
group (blue) are illustrated. (F) NMR structure of integrin β
3
TMD upon K716E
mutation. Side chains of residue K716E interacts with the lipid phosphate head group.

39 | P a g e

K716X mutation causes topological rearrangement
As previously shown, K716X mutations have significant effects on integrin activity, and thus it
appears essential to understand the structural basis underlying the phenomena. The previous section
accredited the activating mutation K716E to a shortening of the TM region to ~23 amino acids;
however, this explanation does not suit K716 mutation to either Ala or Leu. Figure 1-9.A compares
chemical shift changes of the β
3
TMD wt to the K716A/L/E mutants, and it is surprising that both
K716A/L mutants cause a larger shift change than the K716E mutant in both the amide and proton
dimension. This change can be accredited to the shift in chemical environment as expected due to helix
rotation. Figure 1-9.C compares
13
C
α
secondary chemical shifts between the TMD region of wt integrin
β
3
and β
3
(K716A/L/E) TMD. The minor shift differences between the β
3
variant demonstrate the
absence of any significant rearrangements in secondary structure conformation or content. Similar to
the
13
C
α
secondary chemical shifts, no significant differences are seen in the embedding between
β
3
(Lys 716) and its Ala or Leu mutant when compared to the Glu mutant (Figure 1-9.C). In Figure 1-
9.D, the TROSY cross-peak signal intensity of a residue in the presence and absence of 1 mM
Mn
2+
EDDA
2−
in the aqueous phase, I/I
0
, was measured to quantify accessibility to the paramagnetic
reagent. Embedding of the β
3
TMD wt and its mutants inside the lipid bilayer was further monitored via
cysteine constructs at location Leu694, 714C, and 718C and labeled with monobromo-bimane. The
peptides were then inserted into LUVs composed of POPC lipids. Changes in topology can be
identified by monitoring the emission spectrum of the LUVs containing the wt β
3
mutants (K716A/L)
compared to the emission spectrum of β
3
TMD wt. As expected, the bimane emission was shifted to a
higher wave number for the Leu694C-Bimane moiety as compared to the 714C and 718C positions of
the bimane label. Thus, K716A/L causes side chains at both the N- and C-terminal end of the β
3
TMD
to reside in a more hydrophilic environment. Even so, the relative embedding changes of those three
residues are not directly comparable to the change in fluorescence, but are in agreement with the PRE
protection assay in Figure 1-9.D. Thus, K716A/L can induce changes in topology by changing the

40 | P a g e

rotation angle (Figure 1-9.F). Moreover, rotation and/or pistoning could contribute to the changes we
observed in integrin activity upon K716 mutation to either Ala or Leu; however, pistoning would cause
opposing effects on the two TMD termini and can therefore be excluded.

41 | P a g e

42 | P a g e

Figure 1-9 β3TMD Lys716 mutation depicts topological changes. (A) Chemical shift changes of
K716E relative to the wild type as isolated from its TROSY spectrum in B. (C) Comparison of 13Cα
secondary chemical shifts between the TMD region of wild type integrin β3 and β3(K716E) TMD.
Secondary
13
C
α
chemical shifts, defined as the difference between the observed and tabulated random-
coil
13
C
α
shift of a residue, correlate with the backbone conformation. The minor shift differences
between the β3 variant demonstrate the absence of any significant rearrangements in secondary
structure conformation or content. (D) Mutation of β3(Lys 716) changes TMD membrane embedding.
The TROSY H–N cross-peak signal intensity of a residue in the presence and absence of 1 mM
Mn
2+
EDDA
2−
in the aqueous phase, I/I
0
, was measured to quantify accessibility to the paramagnetic
reagent. (E) Emission spectra of POPC LUVs containing β3TMD, bimane labeled at L694C (L694C,
I714C, or L718C) in the presence or absence of K716A/L mutation. Data are representative of three
independent experiments. Deceased emission intensities and shift to larger wavelengths suggest bimane
samples a less hydrophobic environment upon mutation of Lys716 to either Ala or Leu. (F) Cartoon
depicts bimane orientation upon TM rotation. Prior to rotation, bimane moieties were oriented towards
the hydrocarbon core; upon mutation, a rotational change in topology shifted the orientation towards
the membrane water interface.

43 | P a g e

Discussion

Disruption of the interaction between integrin α and β TMDs leads to allosteric rearrangements
that result in increased ligand-binding affinity of the extracellular domain (integrin activation) and
activation of inside-out signaling pathways (Anthis et al, 2009a; Anthis et al, 2009b). A stable αβ TMD
association, which is crucial in the regulation of physiological functions of integrins, requires the
simultaneous formation of two discrete assemblies, an inner and outer membrane clasp (IMC and
OMC, respectively). Because the β
3
TMD forms a continuous α-helix, its crossing angle appears
critical for the simultaneous assembly of these clasps (Ulmer, 2010). Therefore, a sophisticated
mechanisms has evolved to regulate this assembly. Figure 1.1 presents the alignment of the various
integrin β family members, with a special interest in the β
3
subunit. Surprisingly, all members locate a
positively charged residue, preferably a lysine, five amino acids into the membrane proximal region.
Previously it was suggested that the lysine residue functions by interacting with the membrane-water
interface via a specific ‘snorkeling’ mechanism (Lau et al, 2009).
The β
3
TMD segment locates one lysine at the N-terminal membrane proximal region (Figure 1-
1). Given the lack of anionic lipid-mediated α
IIb
β
3
complex destabilization, we investigated whether the
membrane immersion depth of β
3
(K716) allowed for variations in electrostatic interaction with the lipid
head group. Forming an atomistic understanding for β
3
(Lys716)-lipid contacts, all atom molecular
dynamics (MD) simulations were carried out for β
3
immersed in a bilayer of either POPC or POPS. For
Lys716, lipid contacts were observed (Figure 1-2.A); the incidence of electrostatic ε-NH contacts to
lipid head groups was apparent in both POPC and POPS. Remarkably, β
3
(Lys716/ε-NH) shifts its
interaction depending on the lipid surroundings; it interacts with the phosphate head group in POPC
bilayer but locates its ε-NH closer to the carbonyl group in a POPS environment and therefore actively
disregards the interaction with its phosphate group. This expands the snorkeling mechanism to further
interact with its lipid surroundings, as is exemplified by the distance the Lys716/ε-H
N
occupies in

44 | P a g e

relation to its interaction partner. As expected for electrostatic interactions, the preferred distance for
the negatively charged group in POPC or POPS is 2Å, as represented by roughly 70% percent of the
interactions. However, minor interactions occur for both systems which could represent a possible
scouting mechanism of interaction partners. It appears that Lys716 is suitably positioned to interact
with lipid head groups, and further, is able to discriminate between zwitterionic and anionic lipids
(Figure 1-2.C). We conclude that β
3
(Lys716/ε-H
N
) interacts differently with POPC or POPS, possibly
forming a mechanism to induce topological changes upon interaction with anionic lipids.
The α
IIb
β
3
complex formation is tightly regulated. The simulation suggested no significant
difference in tilt between the POPC and POPS system could be observed (Figure 1-3.B); therefore,
other changes in topology, due to the difference in Lys716 snorkeling, allow for the increase in
heterodimer formation. Figure 1-3.C suggests that during the interaction with POPC verses POPS,
β
3
TMD rearranges in the membrane bilayer, adopting a rotation angle and lipid dispersion, thus
preparing for a stable heterodimer formation. Interestingly, the initial tilt angle of ~40° is further
supported by solid state NMR (Figure 1-3.D). Since both independent methods reach the same
conclusion, the tilt angle appears to encircle ~40°, however the exposure to POPS might catalyze a
change in tilt which cannot be accounted for in MD simulation. Further observation by solid state NMR
estimated a tilt change of ~10° upon association of β
3
with its counterpart, α
IIb
. The tilt angles were
determined by comparing the CP spectrum with chemical shift projections of ideal helical wheels. In
Figure 1-3.E the two spectra compare the CP 1D spectra of the β
3
TMD alone (lower) or associated with
unlabeled α
IIb
TMD (upper), each fitted with the ideal helical wheel projection. The CP 1D spectrum of
β
3
TMD resulted in a tilt of ~40°, however if associated with α
IIb
TMD the tilt angle decreased to ~30°.
If correlated, the relation between dimerization and anionic lipid content should be both concentration
and charge dependent.
ITC titration allows for monitoring of the dimerization process and to extract the change of free
energy in dimerization in order to compare the effects of the lipid environment. In Figure 1-4.A, three

45 | P a g e

different lipid environments are compered; POPC, a net neutral lipid, appears to have the lowest K
XY

for dimer formation, whereby the introduction of anionic lipids (e.g. POPG and POPS) increase the
dimer stability (Table 1.1). Furthermore, POPS is more electronegative due to its carboxyl group
compared to POPG’s hydroxyl groups, therefore increases dimer stability. NMR titrations were utilized
to extract information on the POPS concentration dependence. The peak volume ratios relate to the
overall concentration of a peptide species apparent during the titration, allowing for distinction between
the monomer and dimer populations. The decrease in monomer population relates to the increase in
dimeric species, allowing the equilibrium constant to be extracted. In Figure 1-4.B, the long-chain lipid
ratio between POPC and POPS was varied from 0% to 100% POPS. The shift in the equilibrium curve
to the left suggests an increase in the dimer population related to the increase in anionic lipid
concentration. Figure 1-4.C plots the linear relationship between POPS content and dimer formation,
suggesting that the dimer formation is indeed influenced by the anionic lipid concentration found in the
inner membrane leaflet. However, we did notice that the linear behavior indicates that the anionic lipid-
binding site on α
IIb
β
3
was not yet saturated and allows for either (1) further increase in anionic strength
and the Lys716-ε-H
N
/POPS packing mechanism or (2) further lipid binding sites.
The transmembrane domain dimerization process establishes itself as highly sophisticated, as
variation in environmental factors can easily influence the equilibrium constants. Considering the
presence of other possible snorkeling residues, such as Trp715 and Arg724 at the β
3
TMD, and also
Trp967and 968, and Arg995 and 997 at the α
IIb
TMD, sampling the specific impact of Lys716 on
heterodimerization upon exposure to variations in lipid environment is rather challenging (Ulmer,
2010; Ulmer et al, 2001). Additionally, the salt bridge formed between α
IIb
(R995) and β
3
(D723), with
which anionic lipids could compete to form the α
IIb
β
3
TMD complex, may cast doubts on the data
presented previously (Ulmer et al, 2001). Lysine’s unique capability for chemical modifications, such
as acetylation and methylation, allows it to specifically isolate its effect on dimer formation compared
to other amino acids. Lysine’s snorkeling mechanism was further evaluated via its alkylated, tertiary

46 | P a g e

amide state of its side chain. The initial electrostatic interactions between the positive charge of the
lysine amide group and the negatively charged phosphate lipid head group was inhibited by the volume
increase of the methyl groups. Electrostatic interactions between two charged groups prefer a distance
of ~2Å (Figure 1-2.B); however, through the addition of two methyl groups to the lysine amide, the
membrane interface network is disturbed which results in a weaker snorkeling effect. The PRE data,
shown in Figure 1-5.A, supports the notion that the C-terminal membrane-water interface is disturbed
by the tertiary amide, which is accompanied by a chemical shift change in Lys716’s surrounding amino
acid (Figure 1-5.C). Additionally, a rather minor shift change in the N-terminal interface presents an
alteration in the membrane topology, as expected from the membrane mismatch at the C-terminal
helical end. The energy barrier for a snorkeling residue is 0.7 kcal/mol, and upon methylation, the
energy barrier is expected to increase, thus creating a weakened snorkeling effect. However, due to the
positive charge still pending on the lysine side chain, lysine is unlikely to penetrate the hydrophobic
core of the membrane and shift towards the hydrophilic exterior or retain a space devoid of lipids in
close range (Figure 1-5.D). In either case, the result is a disturbance of the snorkeling mechanism and
dissociation from the lipid phosphate group similar to the POPS environment.
To what extent the disturbed snorkeling mechanism affected the dimer formation was probed
via the previously reported ITC method. Figure 1-6.A and Table 1.3 compare the dimerization rate for
β
3
(K716) methylated and unmethylated in POPC bicelles (q=0.31). Interestingly, the change of free
energy decreases to values similar to a POPG environment (Figure 1-6.B). Even more surprising, we
found that a similar experiment in POPG bicelles (q=0.31) models a dimer/monomer equilibrium close
to values observed for the POPS environment. During the experiments no differences occurred, except
in the methylation status of the lysine residue. The results in Figure 1-6.A argue that the weakened
electrostatic interaction between Lys716 and PO
4
-
stabilizes the dimer species. One could jump to the
conclusion that Lys716/ε-H
N
competes for the interaction of the PO
4
-
or COO
-
(in case of POPS). Upon
methylation, K716's interaction with PO
4
-
is weakened, thereby disturbing the membrane location. In

47 | P a g e

POPG, the methylated amide group further distances itself from the interfacial PO
4
-
groups,
subsequently approaching the hydroxyl sites at the lipid head group. In summary, the dimethylation of
lysine's amide group resulted in an increase in dimer stability by discriminating against the interaction
with PO
4
-
, and the resulting equilibrium constants are modeled closely to interactions in a more anionic
bicelle environment.
To examine the effect of β
3
(Lys 716) mutations on the α
ΙΙb
β
3
TMD association in mammalian
cell membranes, our collaborators used a mini-integrin affinity capture assay and chemical cross-
linking. For the capture assay, an α
ΙΙb
mini-integrin bait containing the TMD and cytoplasmic tail of α
ΙΙb

(Figure 1-7.A) joined to a C-terminal tandem affinity purification (TAP) tag for rapid efficient
purification was expressed with prey comprising the extracellular domain of the Tac (IL-2 receptor α)
joined to the TMD and tail of β
3
or β
3
-bearing Lys 716 substitutions (Figure 1-7.A). When the cells
were lysed and bait captured using calmodulin beads, we found that the α
ΙΙb
bait captured the β
3
prey, as
expected; however, neutral (Ala), polar (Cys, Ser), acidic (Glu), or hydrophobic (Leu) substitutions at
β
3
(Lys 716) blocked the α
ΙΙb
β
3
TMD association. In contrast, a basic amino-acid substitution (Arg) did
not disrupt the association, consistent with the idea that a snorkeling residue in this position is required
for the formation of the α
ΙΙb
β
3
TMD complex. To examine the potential effects of β
3
(Lys 716)
mutations on transmembrane signaling, we assayed their effects on the affinity state of integrin α
ΙΙb
β
3

by measuring binding to an activation-specific α
ΙΙb
β
3
antibody (PAC1) as in Figure 1-7.B. The results
precisely correlated with the effects on αβ TMD interaction; all substitutions, with the exception of Arg,
led to spontaneous integrin activation (Figure 1-7.C). Thus, loss of a conserved basic residue in integrin
β TMDs leads to disruption of the α–β TMD interaction and spontaneous transmembrane signaling.
The findings that β
3
(Lys 716) substitutions cause dissociation in the α
ΙΙb
β
3
TMD complex and
activate integrins was surprising as it implicates its significance in a physiological system.
Nevertheless, Rosetta modelling with sparse restraints, provided by cysteine cross-linking, predicted
seven clusters of integrin α
ΙΙb
β
3
TMD structures, some of which resembled the calculated NMR

48 | P a g e

structure. The Rosetta structures suggested that β
3
(Lys 716) can form hydrogen bonds with the α
ΙΙb

backbone carbonyl groups of Phe 992 and Phe 993, thereby stabilizing the α–β interaction; the same
paper reported that mutations at β
3
(Lys 716) resulted in integrin activation, a result we confirmed.
However, NMR-based structural restraints preclude β
3
(Lys 716/ε-NH
3
+
)–α
ΙΙb
(Phe 993/CO) interactions.
Furthermore, embedding of the isolated β
3
TMD was similar to that observed in the α
ΙΙb
β
3
complex and,
as shown, loss of the snorkeling β
3
(Lys 716) alters β
3
TMD embedding. β
3
(Lys 716) mutations on α
ΙΙb
β
3

TMD association and on integrin activation were dominated by changes in the β
3
membrane
topography.
To assess the role of β
3
(Lys 716) in TMD topography, the Lys was mutated to Ala, Leu, and
Glu residues of which the embedding in phospholipid bicelles was assessed by measuring the
protection of the backbone amide protons from the paramagnetic Mn
2+
EDDA
2−
agent. Lipid embedding
on the extracellular side, defined by the protection pattern of Leu 694–Val 696, was unchanged in all
β
3
(K716) mutants (Figure 1-9.D). The same holds true for the intracellular side with regards to the Ala
and Leu mutants. In contrast, β
3
(K716E) reduced protection on the intracellular side by approximately
five residues, shifting the membrane border from residue 721 to 716 and decreasing the membrane
crossing angle. The absence of significant
13
C
α
chemical shift changes between β
3
and β
3
(K716E)
indicated there was no change in secondary structure as a consequence of the mutation (Figure 1-8.C;
1-9.C). At the level of H
N
shifts, which are sensitive to the surrounding chemical environment,
relatively small H
N
chemical changes between β
3
and β
3
(K716E) indicated limited rearrangements in
bicelle structure (Figure 1-8.A), suggesting that Glu716 was still localized within the lipid–water
interface. In analogy to the K716(ε-NH
3
+
)-lipid(PO
4
−
) snorkeling interaction, glutamate’s γ-COO
−

group may engage a POPC lipid’s choline N(CH
3
)
3
+
group or a POPS lipid's amino NH
3
+
group within
the lipid head group region (Figure 1-8.E) (C* et al, 2011). The β3(K716E) TMD-tail was not
aggregated, indicating that displacement of the negatively charged Glu716 from the hydrophobic core
shifted Leu(717)–Ile(721) into a more polar environment. Thus, Lys716 substitutions perturb β3 TMD

49 | P a g e

membrane embedding and crossing angle (Figure 1-8.F)
Interestingly, as previously stated, a mutation of the positively charged Lys716 to a
neutral/hydrophobic residue, such as Ala or Leu, impairs integrin function. Even so, no large changes
in membrane embedding were found, but compared to K716E, the chemical shifts in the NH and HN
dimension showed differences compared to the wt. In comparison to the K716E mutation, the shift
changes are larger for either of the two mutants, suggesting a rearrangement of the TMD in the bicelle
environment. Since limited differences occurred in the PRE assay, the TMD rearrangement is expected
to occur via rotational changes of the TMD. The chemical environment of individual residues (Leu694,
Ile714, and Leu718) up to a distance of around 6Å were sampled via the bimane fluorophore. Bimane’s
fluorescence depends on the chemical surrounding; an aqueous environment causes emission at a lower
intensity but higher wavelength compared to a rather hydrophobic environment in which the intensity
increases but the emission wavelength is shifted downfield. The bimane labeling sites were chosen
based on their topological location compared the membrane; in the β
3
TMD, wt Ile714 is sampling
primarily the inner hydrocarbon space compared to Leu 694 and Leu718, which are located closer to
the membrane-water interface and therefore are more susceptible to changes in hydrophobicity. In the
case of the wt, all the bimane labeling sites are pointing inwards. However, upon Lys716A/L mutation,
the fluorescence intensity and emission wavelength suggest an orientation towards a hydrophilic
environment. These findings support the hypothesis that Lys716 influences the rotation angle of
β
3
TMD.

50 | P a g e

Conclusion

The evolution of membrane proteins in a bilayer allowed for an adaptation to the outer and inner leaflet,
in which the inner leaflet contains a large quantity of anionic lipids. As the hydrocarbon core developed
hydrophobic mismatch, the membrane-water interface has developed a “snorkeling” mechanism. Here
we present evidence that this “snorkeling” mechanism has resulted in the regulation of electrostatic
contacts between anionic lipids and cationic protein residues to establish a specific membrane protein
topology. In summary, we showed that snorkeling can affect transmembrane signaling by altering the
stability of interactions between integrin TMDs. The equilibrium between the monomer and
heterodimer of integrin α
IIb
β
3
heavily depends on the interaction between Lys716 and negative charges
distributed throughout the membrane interface, where anionic lipids strongly favor the dimeric species.
The anionic lipid interaction appears to coincide with changes introduced upon methylation, which
seems to discriminate against the lysine-phosphate interaction. Furthermore, we demonstrated that the
loss of a snorkeling residue (K716) in integrin β TMDs can change membrane embedding, and further,
the membrane-crossing angle, providing direct evidence that snorkeling can specify the topography of
TMDs. Thus, the long-appreciated snorkeling of basic residues in TMDs can have an important role in
their topography and lateral association and consequently in signal transduction. Moreover, the
presence of anionic lipids is of universal importance to the folding and stability of membrane proteins,
while causing limited restrictions on electrostatic helix-helix interactions in the head group region.

51 | P a g e

Chapter 2 – β
3
(Ala711P) increases dimer formation due to
increased glycine packing in the OMC

Abstract
Introduction
Results
Ala711P mutation reduces integrin activation.
Ala711P mutation increases α
IIb
β
3
heterodimer stability
FRET to measure TMD dimerization at low error
Structure of the β
3
(K716A,A711P) TMD
Structure of the α
IIb
β
3
(A711P) heterodimer
Discussion
Conclusion

52 | P a g e

Abstract

Membrane proteins constitute 50% of all drug targets and are an important subcategory for
structural characterization despite accounting for less than a third of all proteins in the human genome
(Ulmschneider & Sansom, 2001). Due to specific properties within the structure, membrane proteins
are not easily characterized by conventional methods, therefore NMR presents a unique tool for that
purpose (Nietlispach & Gautier, 2011). Integrin has a pivotal role in the regulation of cellular migration,
therefore it is important to truly understand the molecular mechanism that controls its activity.
Previously, we have shown that K716 facilitated integrin manipulation via electrostatic interaction with
the lipid head group; mutation of β
3
(Lys 716) resulted in the dissociation of α
ΙΙb
β
3
TMDs and
subsequent integrin activation (Kim et al, 2011). Upon confirming that the altered topography of
β
3
(Lys 716) mutants activated α
ΙΙb
β
3
, we used directed evolution of β
3
(K716A) to identify substitutions
that restored the original state. Introduction of Pro(711) at the midpoint of β
3
TMD (A711P) increased
α
ΙΙb
β
3
TMD association and inactivated integrin α
ΙΙb
β
3
(A711P,K716A). The β
3
(Pro 711) modification
caused a TMD kink of 30 ± 1° at the border of the outer and inner membrane clasps, resulting in a
decoupling of the tilt between these segments. Here we proved the structure-to-function relationship
between the mentioned phenomena through quantifying the dimerization process via three isolated and
independent methods: ITC, NMR and FRET (Bocharov et al, 2012; Merzlyakov et al, 2006a;
Tellinghuisen, 2008). Additionally, significant improvements were made to an established FRET
titration method, which decreased errors caused by colocalization to less than 2% independent of
labeling efficiency (Li et al, 2005a; Merzlyakov et al, 2007; Merzlyakov et al, 2006a; Posokhov et al,
2008; You et al, 2005). Finally, we determined the monomer and heterodimer structure of the
α
ΙΙb
β
3
(A711P) mutant, giving us new insights into the molecular basis of integrin activity. The results
provided herein outline integrin activation during the inside-out signaling event, which may lead to the
development of novel treatments in integrin-associated diseases.

53 | P a g e

Introduction

The transmembrane (TM) segments of integral membrane proteins are embedded in a
phospholipid bilayer, an environment which strongly limits the range of possible structures. Taking this
into account, the effect of proline residues in membrane protein folding should be based on its capacity
to perturb helix packing more than the helicity of the region itself (Bowie, 2011; Partridge et al, 2005).
Therefore introduction of a proline residue may convey beneficial effects of protein interaction in the
hydrophobic bilayer. There are many advantages of transmembrane helix kinks for membrane protein
structure and function. The literature sites the following: weak points that allow movement during
catalytic cycles, precise positioning of key side chains, water recruitment to functional sites, and
prevention of off-pathway folding (Cao & Bowie, 2012; Orzáez et al, 2004a; Rankenberg et al, 2012;
Senes et al, 2004; Solomaha & Palfrey, 2005). It is surprising that distortions are much more common
in transmembrane than soluble protein helices since helices are more stable in non-polar membrane
environments where backbone hydrogen bonds are stronger (Cao & Bowie, 2012). How these
distortions could have possibly been generated by the evolutionary occurrence of random point
mutations remains a mystery. Here we present a model in which random point mutations are able to
increase helix packing and interaction. Furthermore, we’re able to rescue the initial monomeric state
upon proline introduction.
Many structural models of signaling events in the membrane have been proposed. The structural
information of the monomeric, dissociated α
IIb
and β
3
TMD states show an increase in possible
structural transitions between the associated and dissociated states (Lau et al, 2008a; Lau et al, 2009;
Lau et al, 2008b; Ulmer, 2010). In the previous chapter we had focused on Lys716, which has the
potential to regulate TMD heterodimerization. K716, upon mutation, can destabilize the heterodimer
and activate integrin signaling. In this chapter our focus shifts towards counteracting this event through
stabilization of the dimeric species of the protein. To do this we screened random mutations and

54 | P a g e

identified β
3
(A711P) as a mutation that decreases integrin activity via enhanced heterodimer formation
and rescues the original phenotype. The effects were determined to be localized at the TMD region of
the α
IIb
β
3
TMD heterodimer. Due to its potential application in the pharmaceutical industry and the
advancement in membrane protein science we set our goal in identifying the structure-to-function
relationship of this unique phenomena.
The interaction of TM helices can be observed via analytical ultracentrifugation, fluorescence,
ITC and NMR spectroscopy (Bocharov et al, 2012; Merzlyakov et al, 2006b; Tellinghuisen, 2008).
Here we present a titration experiment that utilizes three independent methods to quantify the
monomer/dimer equilibrium change established by the Ala711P mutation. Both ITC and NMR are
relevant methods as both ground themselves in utilizing bicelles to measure the equilibrium constant.
Additionally, we advanced a FRET-based method that is able to determine the equilibrium constant
within a LUV system, adding strength and validity to the measurements obtained in the previous
methods. The results from all three experiments and systems show that β
3
(A711P) mutation increases
the dimer formation.
After the mutation-based experiments were carried out we used NMR to see which structural
components resulted in the enhanced dimer stability. From this data we are able to report the structure
of the covalently associated α
IIb
β
3
(A711P) TMD complex. This structure ‐based site ‐directed
mutagenesis and lipid embedding gives us an estimate of the large number of integrin TM signaling
possibilities. Aside from providing insight into integrin biology, the integrin α
IIb
β
3
(A711P) TMD
complex structure also advances the understanding of dimeric cell receptor TM complexes of which
there have been limited reports of heterodimeric structures.

55 | P a g e

Results

Ala711P mutation reduces integrin activation.
We used random mutagenesis in a window of 5 residues N-terminal and 10 residues C-terminal
to the β
3
(K716) substitution to identify mutations that would complement the activating effects. We
chose β
3
(K716A) as the activating mutation; its effects may be less profound than β
3
(K716E), thus
favoring the discovery of weakly compensating mutations. Lentivirus particles carry two genomes,
making it possible that a single particle might encode two mutants. To test this possibility, we
transfected packaging cells with a mixture of lentiviral plasmids encoding integrin β
3
and β
3
(K716A).
When CHO cells bearing integrin α
IIb
were infected with the viruses we found only populations
containing either fully active or fully inactive α
IIb
β
3
, no intermediate phenotype was seen. Thus, the
packaging cells incorporated two copies of identical genomes into each viral particle. We performed
PCR using a primer that was synthesized with 9% contamination of incorrect nucleotides in the
randomized windows. The contamination level was predicted by computer simulation to cover most
single amino acid changes within the window. We ligated the PCR fragments containing random
mutations into a lentiviral vector encoding full-length integrin β
3
to create a randomized β
3
cDNA
library.
CHO/α
IIb
cells were infected with the mutant β
3
lentiviral particles and the infected cells were
analyzed for integrin expression (D57 antibody binding) and activation (PAC1antibody binding)
through flow cytometry. In contrast to the β
3
(K716A)-infectedcells, cells infected with the mutant β
3

library had a population of cells that expressed inactive α
IIb
β
3
. To identify these mutations, we collected
~7,000 cells in the R3 region and purified genomic DNA, then used PCR to isolate the region of
integrin β3 containing the mutagenized window. Sequencing of the bulk product revealed that a
β
3
(A711P) mutation and stop codons at residue 724 and 725 were found in the mutagenized region. We
performed NheI and BamHI digestion to isolate individual fragments containing the mutations and,

56 | P a g e

after ligation into wild type integrin β3, confirmed the compensating effect of the mutations through
transient transfection into cells expressing wild-type integrin αIIb and measurement of PAC1 binding.
The clones showing a compensating effect were sequenced, resulting in 3 major groups: clones
containing Pro substitutions at Ala711 position (group 1), stop codons at Arg724 or Lys725 positions
(group 2), and a neutral residue substitution at Glu726 position (group 3).

Figure 2-1 Ala711P stabilizes the α
ΙΙb
β
3
(K716A) TMD interaction and reduces integrin activation.
A. β
3
(A711P) complements β
3
(Lys 716) mutations in integrin activation. CHO/α
ΙΙb
cells were
transiently transfected with various integrin β
3
constructs as indicated, binding to PAC1 and D57 was
analyzed as described above. Error bars, s.e.m. (n=3). B. β3(A711P) stabilizes the TMD interactions of
α
ΙΙb
β
3
(K716A). CHO cells were co-transfected with α
ΙΙb
TM-TAP constructs and wild type, or mutant
Tac-β
3
TM constructs as indicated, and the interaction between those integrin TMDs were analyzed as
previously reported

57 | P a g e

Ala711P mutation increases α
IIb
β
3
heterodimer stability
The previous section identified β
3
(Ala711P) as a integrin-deactivating/heterodimer-promoting
mutation that is able to rescue the otherwise constantly active integrin mutants (e.g. β3(Lys716A))
under physiological conditions within CHO cells. Ala711P is located in the TMD section of the β
3

subunit, therefore it was important to consider which effect it would expose on the heterodimer TMD
of α
IIb
β
3
in the hydrophobic environment (Figure 1-1)
.
The dimerization process can be expressed by the
α
IIb
+ β
3
⇄ α
IIb
β
3
equilibrium, which gives rise to the equilibrium constant, K
XY
(Fleming & Engelman,
2001). The interaction of TM helices can be quantified by analytical ultracentrifugation, fluorescence,
ITC and NMR spectroscopy. Here we show the titration experiment utilizing two different,
independent methods to measure the monomer/dimer equilibrium change established by the Ala711P
mutation. In this case, both ITC and NMR form relevant methods, as they utilize two different aspects
during kinetic studies; during ITC titration the heat exchange upon dimer formation will give rise to the
equilibrium constant; while during the NMR titration the quantification of monomer and dimer
population allows gives rise to K
XY
values.
ITC has been used to study the interaction of membrane proteins with ligands in the aqueous
phase, here we applied ITC because both ligand and protein are in the hydrophobic phase. ITC
measurements determine TM helix-helix association and dissociation rate constants based on the heat
exchange upon dimer formation. ITC depends on the direct titration of one peptide to its partnering
peptide in the sample cells to detect relatively small heat changes upon complex formation (Son et al,
2012; Wiseman et al, 1989). Figure 2-2 explores the difference introduced upon various mutations in
β
3
TMD, which are taken from Figure 2-1. Initially the dimer formation for the wt α
IIb
β
3
was observed,
resulting in ΔG=−4.91 kcal/mol. Upon Lys716A mutation, in β
3
TM section, a drastic decrease in dimer
formation was observed when compared to the wt; the Lys716A mutation appears to favor the

58 | P a g e

monomer species with ΔG=−3.60 kcal/mol. Figure 2-1.A suggests that the wt dimer equilibrium can be
reestablished upon a proline mutation at position Ala711. The ITC trace confirms this expectation, the
Ala711P mutation does increase the heterodimer formation even so it does not reach the initial wt
equilibrium. In Chapter 1, we presented convincing evidence that mutations in K716 cause a disruption
in the integrin heterodimer by dissociating the TMD sections, this is further supported by ITC
measurements. Even so the difference is rather small, compared to β
3
K716A (ΔΔG=−0.2 kcal/mol), it
could provide evidence for an inactivation mechanism which allows a certain threshold. More
interestingly, the “rescue” mutation has the ability to further increase the dimerization compared to the
wt, showing a novel characteristic of proline residues that are found in single-pass transmembrane
domains (Table 2.1). Overall, the ITC method observed transmembrane dimerization characteristics of
α
IIb
β
3
upon dimer formation. Quantitative measurements by this method gave insights to rescue
mutation introduced in Figure 2-1.A and further isolated Ala711P mutation to shift the equilibrium
towards the dimer species, in the K716X mutants as well as wt constructs. However, ITC
measurements dismiss the population differences between the monomer and dimer species and rely on
the heat exchange upon association, allowing the usage of NMR as a valid complimentary method to
validate our findings (Bocharov et al, 2012).
NMR depends on the reliable detection of protein resonances, limiting applicable particle sizes
to molecular weights of 30-50 kDa. The dependence of NMR line widths on rotational correlation time
limits NMR measurements of the integrin α
IIb
β
3
TMD complex to the usage of bicelles with q factors
~0.5. NMR was employed to determine α
IIb
+ β
3
⇄ α
IIb
β
3
equilibrium constant on a mole-fractions
scale, termed K
XY
. On the NMR chemical shift timescale the α
IIb
+ β
3
⇄ α
IIb
β
3
equilibrium is in a slow
exchange limit at 28 °C, i.e., both monomer and dimer resonances are observable simultaneously (Suk
et al, 2012). This permits the direct quantification of the fraction of dimer formed as 1-V
M
/V
M,0
where,
under the employed isotope labeling scheme, V
M,0
is a β3 resonance volume in the absence of α
IIb
and
V
M
its volume at a specific α
IIb
concentration (Bocharov et al, 2012). K
XY
can be obtained from one

59 | P a g e

measurement at a non-zero α
IIb
concentration. However, to enhance accuracy, K
XY
was obtained by
non-linear curve fitting of five concentrations of α
IIb
(Figure 2-3.A). Hence NMR based titrations give
insights to the population distribution; if the exchange rate is slow enough it can be used to validate the
titration data gathered via ITC. In Figure 2-3.A the α
IIb
β
3
wt is compared to its α
IIb
β
3
(Ala711P) mutant,
when comparing the dimer fractions it becomes clear that the Ala711P mutation allows for an earlier
saturation and supports the formation of the dimer species. The K
XY
of the α
IIb
β
3
TMD is 2070M
-1
while
the β
3
TMD(A711P)/α
IIb
TMD is 3747M
-1
, corresponding to a ΔG of -4.57 kcal/mol and -4.92 kcal/mol
respectively (Table 2-1). Even though the values deviate from the reported ITC measurements they
hold true for its increase in dimerization upon proline mutation. This equilibrium difference can be
explained by the dissimilarity of the observed variables; during the ITC measurements we disregard the
population difference and need to take in account the lipid variation environment, as presented by ΔS
and ΔH. However in the NMR approach the observation of the monmeric and dimeric species
disregards the kinetic exchange upon bicelle collision, as only population differences are observed
(Chou et al, 2004).

60 | P a g e

61 | P a g e

Figure 2-2 β
3
(Ala711P) mutation stabilizes the α
IIb
β
3
heterodimer. ITC measurements in
phospholipid bicelles consisting of 43 mM DHPC and 17 mM of the depicted long-chain lipid
(q
eff
=0.5). The ITC titration of various β3 mutants establishes an understanding of the kinetic
characteristics. The α
IIb
β
3
wt forms a heterodimer at ΔG=−4.91 kcal/mol, however upon introducing an
alanine at position 716 a loss of heterodimerization is seen (ΔG=− 3.60kcal/mol), this can be rescued
via the Ala711P mutation (ΔG=− 3.79kcal/mol). Furthermore, upon introduction of the Ala711P
mutation in the wt, a shift towards the dimeric species could be observed (ΔG=−5.75 kcal/mol). These
results are in agreement with the data presented in Figure 2-1.

62 | P a g e

Figure 2-3 β
3
(Ala711P) mutation stabilizes the α
IIb
β
3
heterodimer. NMR measurements in
phospholipid bicelles consisting of 400 mM DHPC and 120 mM of the depicted long-chain lipid. (A)
The decrease of the monomeric b3(G702) peak volume is plotted as 1–V
M
/V
M,0
where V
M,0
denotes the
resonance volume of pure monomer at 0.1 mM and V
M
the residual monomer volume at increasing
concentrations of partnering peptide. (B) Shown is a schematic presenting the difference between the
α
IIb
β
3
wt and its Ala711P mutant; the proline mutation stabilizes the dimer form by
ΔΔG=−0.36kcal/mol.

63 | P a g e

FRET to measure TMD dimerization at low error
The previous experiments compared the kinetics of dimerization of the α
IIb
β
3
wt and its Ala711P
mutant. Both approaches yield similar results in that the dimerization is increased in the presence of
A711P mutation. ITC experiments depend on the direct titration of one peptide to its partnering peptide
in the sample cells to detect relatively small heat exchanges upon complex formation, but it does not
observe the direct population of monomeric to dimeric species. For that reason NMR was used to
confirm the ITC based equilibrium constant, which is based on the TROSY peak volumes of the
monomer and dimer resonances that were observed simultaneously. There is a large set back caused by
the requirement of a bicellular environment in both ITC and NMR titrations, therefore it was beneficial
to observe the dimerization condition under larger membrane conditions such as large unilamellar
vesicles (LUVs) (Merzlyakov et al, 2006a).
Previously, Merzlyakov et al (2006) devised a method utilizing FRET to measure TMD
dimerization in larger membrane systems, however his method was flawed due to labeling efficiency as
well as a large error due to colocalization (Merzlyakov et al, 2006a). Förster resonance energy transfer
(FRET) relies on resonance transfer between a donor and acceptor, which requires the tagging of TM
peptides with fluorophores and the correction of the detected signal for the random co-localization of
donor and acceptor (Li et al, 2005b). Therefore we revised the previous method by utilizing
fluorophores with a relative small Förster radius (R
0
), such as naturally occurring tryptophan, in
combination with 5-([2-[(iodoacetyl)amino]ethyl]amino)naphthalene-1-sulfonic acid (IAEDANS)
(Figure 2-4.A) (Klepsch et al, 2011). Figure 2-4.B presents the fluorescence titration of the α
IIb
β
3
TMD
heterodimer, the combination of a set concentration of peptide is titrated with an increasing
concentration of POPC LUVs (Jayasinghe et al, 2001). The increase in POPC LUVs causes the dimer
to dissociate, therefore decreasing the IAEDANS fluorescence intensity. We noticed that the
tryptophan fluorescence does not increase with a decrease in IAEDANS fluorescence as would be
expected in a FRET experiment (Hristova & White, 2005; Jayasinghe et al, 2001; Li et al, 2005b; Li et

64 | P a g e

al, 2006; Merzlyakov et al, 2007; Merzlyakov et al, 2006b; Posokhov et al, 2008; You et al, 2005). The
reason revolves around the presence of multiple Trp residues in the heterodimer (Figure 1-1 and 2-4.A)
that contributed to the fluorescence. However, due to the small R
0
of 22Å for the fluorophore pair, the
other Trp are out of range and do not significantly contribute to the Trp-IAEDANS’ förster resonance
transmission (Figure 2-4.A). The benefit of using innate Trp as acceptor fluorophore is its certainty in
labeling efficiency, allowing the determination of transfer efficiency by the IAEDANS fluorescence
intensity at ~477nm without taking in account its labeling efficiency. The error in the final result is
decreased by: (1) shortening R
0
from ~60Å to 22Å, and (2) eliminating the dependence of the labeling
efficiency. The final result is presented in Figure 2-4.C, which compares the wt heterodimer as well as
its Ala711P mutant, additionally, we included α
IIb
β
3
(D723A). D723A is well establish to disrupt the
heterodimer formation in-vitro and in–vivo experiments and such serves as a negative control.

3
(D723A) disrupts electrostatic contacts to 
IIb
(R995) in the intracellular lipid head group region,
making dimer formation unfavorable.
As expected the D723A mutation did indeed decrease the dimer formation as seen by the
inefficiency to saturate at the given concentrations. The fitting of the rate equation resulted in
ΔG=−2.69 kcal/mol, which clearly shows favoring of the monomer species compared to the wt by
ΔΔG=0.88 kcal/mol. Based on this presented data we can assume the applied modifications to the
previously published method does indeed to determine the change in free energy due to dimerization at
low error and minimal contribution by co-localization (<2%). Similar to previous results, an increase in
heterodimerization is observed upon introduction of a proline residue at position Ala711 (Table 2-1).
The greatest advantage of the FRET method is the ability to measure the association of TMDs that are
readily placed in the bilayer environment, therefore the collisions exchange, as present with bicelles,
can be disregarded. Even though the margins very between the different experiments all three agree
that the Ala711P mutation stabilizes the heterodimer in a lipid environment.

65 | P a g e

66 | P a g e

Figure 2-4 β
3
(Ala711P) mutation stabilizes the α
IIb
β
3
heterodimer. (A) Schematic represents foster
resonance transfer (FRET) between the donor (Trp) and acceptor fluorophore (IAEDANS) with
R0=22Å. FRET solely occurs between α
IIb
’s Trp967 and Trp968 due to the increased distance to C-
terminal Trp. (B) fluorescence titration of the α
IIb
β
3
TMD heterodimer (C) α
IIb
β
3
TMD dimerization in
POPC LUVs. Total dimer fraction is a function of total peptide concentration and gives the ratio of
monomer/dimer distribution. The system was evaluated with the well-known β
3
(D723A) mutation
which will render the heterodimer in its monomeric state (blue). A shift in the equilibration curve upon
Ala711P mutation suggests a rather stable heterodimer compared to its wt (ΔΔG=−0.27 kcal/mol). (D)
The schematic presents the difference of heterodimer formation upon proline introduction as observed
via FRET.

Table 2-1
Construct K
XY
ΔG [kcal/mol] ΔΔG [kcal/mol] Lipid System
ITC

α
IIb
β
3
3629 -4.91 -
Large Bicelle
α
IIb
β
3
(Ala711P) 14938 -5.75 -0.84
NMR

α
IIb
β
3
2070 -4.57 - Small Bicelle
(q≤0.5) α
IIb
β
3
(Ala711P) 3747 -4.92 -0.35
FRET

α
IIb
β
3
291 -3.39 -
LUV
α
IIb
β
3
(Ala711P) 452 -3.66 -0.27

67 | P a g e

Structure of the β
3
(K716A,A711P) TMD
Previous experiments explored the kinetic relationship between the β
3
(Ala711P) mutation and
the sequential decrease in integrin activity. Considering that proline residues frequently found at center
position of single-pass transmembrane proteins (~25%), it is of general interest to establish a structure-
to-function relationship. Elicidation of the TMD structure of the β
3
(A711P, K716A) monomer and the
α
IIb
β
3
(A711P) heterodimer are necessary to form a concise molecular understanding of the proteins. We
chose β
3
(K716A) as an activating mutation since its effects may be less profound than β
3
(K716E)
(Figure 1-8) and may favor rescue by the compensating Ala711P mutation. Initial structural description
was carried out through comparing the TROSY spectrum of the wt and its Ala711P mutant (Figure 2-
5.A&B). As shown in the previous spectra of chapter 1, chemical shift changes occur in the
surrounding environment of the Lys716A mutation, however more significant changes are presented by
the Ala711P mutation. The chemical shift changes suggest a structural reorganization in-between the
inner and outer membrane clasp due to the induced proline mutation. A change in secondary structure
is marked by secondary
13
C
α
shifts and the trace does not appear to alter the length of the α-helix
compared to the wt peptide (Figure 2-5.B). Sudden, minor changes in the C-terminus present the impact
of the Lys716A mutation, as it appears to follow a similar pattern as seen in Figure 1-9.C. The largest
change is observed with residues around P711; these two residues the i-1 direction appear to take up a
random coil conformation, whereas the i+1 direction induces a α-helical conformation. This pattern
presents a helix break caused by P711, the proline is known to break a helix in the i-1 direction due to
its missing amide hydrogen, however it also induces the helix formation in the i+1 direction due to its
dihedral angle. The final structure is seen in Figure 2-5.E. Interestingly, placement in the bicelle is
different from the wt, as it shifts the N-terminal membrane-water interface by three additional residues
in the i-1 direction. Overall the presented structure gives insight to the effects that proline residues have
on the single-pass α-helices, but the heterodimer structure will yield a greater understanding of the
interaction increase.

68 | P a g e

69 | P a g e

Figure 2-5 Proline introduced in the TMD forms a flexible kink. (A) Chemical shift changes of
β
3
(A711P, K716A) relative to the wild type extracted from the TROSY spectrum at (B). (C)
Comparison of the backbone structural parameters of the β3 transmembrane segment between the wt
and β
3
(A711P, K716A) mutant. The α
IIb
TMD sary
13
C
α
chemical shifts obtained from the wt and its
mutant are compared, despite a strong agreement for residues K689 to I704, a minor shift difference
occur at residues L705 to H703. The difference suggests a break of α-helical structure upon Ala711P
mutation. (D) Comparison of the normalized ratios of
1
H-
15
N TROSY signal intensities in the presence
or absence of 1mM Mn
+2
EDDA
-2
, I/I
0
, between β3TMD and its mutant embedded in DHPC/POPC
(q=0.3). The N-terminal membrane border is shifted by 3 residues in the i-1 direction. (E) Structure of
the β
3
(K716A,A711P) TMD. Structure of the bicelle-embedded integrin β
3
(A711P,K716A) TMD
segment compared with the β
3
wt TMD structure. To illustrate the proline-induced kink, the structures
were superimposed on the backbone heavy atom coordinates of Ile 693–Leu 709.

70 | P a g e

Structure of the α
IIb
β
3
(A711P) heterodimer
The formation of a covalently linked α
IIb
β
3
(A711P) heterodimer was previously described . Figure 2-
6.A compares the TROSY spectra of heterodimer vs monomer and shows the disulfide linker indeed
stabilizes and favors heterodimer formation. No significant changes in the secondary structure between
monomeric and heterodimeric states were observed (Figure 2-6.B) (Suk et al, 2012). However minor
changes in the secondary
13
C
α
chemical shifts at the C-terminal end of the β-subunit suggests the newly
formed α
IIb
(R995)–β
3
(D723) salt bridge stabilizes the heterodimer. Furthermore, minor changes post
Pro711 show a dynamic interface that is formed with the α
IIb
counterpart. Further analysis of H
α
,
15
N,
13
C
α
and
13
C′ backbone chemical shifts and HN–HN NOEs confirms that backbone secondary
structures are indeed similar for heterodimeric to monomeric states. NOESY experiments identified
close α
IIb
–β
3
inter-proton distances while suppressing the otherwise dominating intra-subunit NOE
signals by using a unique labeling schematic. Specifically, 35 inter-subunit distance restraints were
identified; in combination with RDC and torsion restrains this resulted in the calculation of 20
simulated annealing structures (Figure 2-6.D). Guided by packing interactions with three distinct
glycine residues, α
IIb
(G972), α
IIb
(G976) and β
3
(G708), the TMD helices cross within their N ‐terminal
halves at an angle of 25±3° (Figure 2-5.B–D). This and the differing length of the α
IIb
β
3
TMD helices
results in a loss of inter-helical contacts C ‐terminal to β
3
(L712) on the intracellular side.

71 | P a g e

72 | P a g e

Figure 2-6 Structure of the integrin α
IIb
β
3
transmembrane complex. (A) Structural properties of the
covalently associated integrin α
IIb
β
3
TMD complex. Comparison of H–N correlation spectra of
covalently and noncovalently associated α
IIb
β
3
TMD domains. (B) Comparison of 13Cα secondary
chemical shifts between monomeric and heterodimeric integrin α
IIb
–β
3
transmembrane states. The
minor differences between monomeric and heterodimeric shifts indicate the absence of significant
backbone rearrangements on heterodimerization and the stabilization of secondary structure at the
intracellular side in the presence of α
IIb
(R995)–β
3
(D723) electrostatic interactions. (C) NOESY spectra
to measure intrahelical distances. (D) Superposition of the ensemble of 20 calculated simulated
annealing structures. α
IIb
(I966 ‐R995) and β
3
(I693 ‐D723) adopt well ‐structured conformations. The
outer membrane clasp: illustration of glycine packing. α
IIb
(G972), α
IIb
(G976) and β
3
(G708) are shown
in green spheres, with their β3 and αIIb packing residues shown in blue and red spheres, respectively.
However The inner membrane clasp: stabilization of the α–β TMD helix arrangement by α
IIb
(F992 ‐
F993) ‐mediated inter-helical packing and α
IIb
(R995)–β
3
(D723) electrostatic interaction. β
3
(W715) and
β
3
(D723) are shown in blue, α
IIb
(R995) in red, and α
IIb
(F992 ‐F993) in green spheres. (E) Cartoon
compares the wt heterodimer with its proline mutant counterpart.

73 | P a g e

Discussion

In the previous chapter we showed that point mutations in K716 destabilize the TMD
heterodimer, which leads to a constantly activated, high-affinity integrin α
IIb
β
3
. However, it was of
great interest to find a rescue mutation that is able to overcome the dimer disruption. The only
compensating mutation consistently observed in the TMD was β
3
(A711P), an integrin TMD point
mutation that inhibits transmembrane signaling and is inaccessible to talin or other cytoplasmic
proteins (Kim et al, 2011). Interestingly, Ala711P is unlikely to provide a direct interaction with α
IIb

since it is not in the α-β TMD binding interface. We confirmed that β
3
(A711P) compensates for both
β
3
(K716A) and β
3
(K716E) mutations with respect to integrin activation (Figure 2-1.A&B). In addition,
we show β
3
(A711P) increases the interaction of β
3
(K716A) TMD with its α
IIb
counterpart (Figure 2-
1.B). Thus, the only consistent TMD mutation that compensated for β
3
(Lys716) substitutions was that
of Pro introduction at position 711.
As shown in-vivo, integrins low affinity state is stabilized by the Ala711P mutation. The result
suggests that the TMD heterodimer is stabilized or a conformational change effects the extracellular
domains, leading to the low affinity characterizing compact structure. Kinetic studies of the TMD
dimerization suggested that the mutations are affecting the transmembrane heterodimer (Bocharov et al,
2012; Chen & Xu, 2006; Fleming, 2002; Li et al, 2006). To understand and utilize the natural folding
design of membrane proteins, next to determining the structures of associated and dissociated TM
helices, we need to assess the thermodynamic and kinetic basis of TM helix-helix interactions.
Additionally, to understand the interaction of more complex membrane proteins analogous
characterization processes are required. The interaction of TM helices in-vitro has been quantified
using approaches based on analytical ultracentrifugation, fluorescence and NMR spectroscopy
(Bocharov et al, 2012; Chen & Xu, 2006; Fleming, 2002; Li et al, 2006). The ITC titration in Figure 2-
2 explores and quantifies the effects Ala711P mutation on dimerization. In case of the wt, it can

74 | P a g e

enhance the dimer formation by ΔΔG=−0.84 kcal/mol, therefore the proline mutation between the IMC
and OMC acts as TMD stabilizing. As previously mentioned, the K716A mutation causes a loss of
dimer formation due to interference with Lys716 snorkeling mechanism, however upon introduction of
the Ala711P mutation the dimer formation can be rescued, thus mirroring the results presented in
Figure 2-1. The efficient, but relatively small difference in change of free energy (ΔΔG=−0.2 kcal/mol)
might present a threshold for integrin activation and correlate ΔΔG with the previously presented
physiological data.
The reaction enthalpy (∆H
o
) and equilibrium constant (K
XY
), but not the a
IIb
b
3
complex
stoichiometry, were calculated from the measured heat changes, δH
i
, as described previously
(Tellinghuisen, 2008). The ITC dimerization is based on the observation of heat exchange in the system
and does not observe the monomer and dimer population, therefore the kinetic study was repeated by
NMR (Turnbull & Daranas, 2003; Wiseman et al, 1989). In the TROSY spectrum, peaks corresponding
to TM monomers and heterodimer of integrin a
IIb
b
3
are simultaneously observable at 28°C.
Homodimerization of α
IIb
and β
3
is absent in the selected solution condition, and this exchange
behavior permits direct detection of the α
IIb
β
3
dimer population by quantifying the peak volume decline
of monomer residues (specifically Gly702). Specifically, we maintained a stable concentration of
2
H/
15
N-labeled β
3
while varying the amount of unlabeled, partnering peptide (α
IIb
) to determine K
XY

(Figure 2-3.A). Isotope labeling of either α
IIb
or β
3
yielded similar results (data not shown). The NMR
titration in Figure 2-3.A explores and quantifies the effects Ala711P mutation on dimerization. In case
of the wt, it can enhance the dimer formation by ΔΔG=−0.36 kcal/mol, therefore stabilizing the
heterodimer (Figure 2-3.B). The NMR and ITC titration yield similar results, suggesting that the
Ala711P mutation indeed targets the transmembrane section of the integrin α
IIb
β
3
subunit by stabilizing
the heterodimer (Table 2.1). The short comings of the experiment are found within the bicellular
conditions, as it only allows monomer/dimer exchange upon bicelle collision.
Fluorescence spectroscopy relies on resonance transfer between a donor and acceptor, which

75 | P a g e

requires the tagging of TM peptides with fluorophores and the correction of the detected signal for the
random co-localization of donor and acceptor. Recent advances in FRET have suffered due to using
fluorophores with a relatively large R
0
(~60Å), therefore we improved the method by utilizing
fluorophores with a relatively small R
0
(~22Å) and minimizing the signal due to co-localization to less
than 2%, as shown by Monte-Carlo simulations (Merzlyakov et al, 2006a). We carried out FRET
measurements of the dimerization free energies of a
IIb
b
3
using two different sets of FRET pairs:
Tryptophan/IAEDANS, and IAEDANS/Dabcyl (which allows for a quenching mechanism). Figure 2-
4.B shows the fluorescence spectrum of a sample containing 2µM α
IIb
β
3
in various concentrations of
POPC. For comparison, we show the emission spectra of samples containing up to 8mM POPC. Using
FRET, in Figure 2-4.B, we observe a decrease in donor fluorescence (around 290 nm) and the
appearance of sensitized acceptor fluorescence (around 477 nm). FRET efficiency was calculated from
the increase in acceptor fluorescence solely and therefore is independent of labeling efficiency. FRET
that arises due to proximity effects was modeled according to the analysis reported by Wolber &
Hudson via Monte Carlo simulation on Mathematica (Hristova & White, 2005). The FRET efficiency
due to sequence-specific dimerization was therefore obtained by subtracting the predicted FRET
efficiency due to proximity from the measured FRET signal. Figure 2-4.C shows the FRET efficiency
due to sequence-specific dimerization for

α
IIb
β
3
, as a function of total peptide concentration. The FRET
efficiency for the wt was measure in both fluorophore pairs (Trp/IAEDANS and IAEDANS/Dabcyl)
with similar results, commenting on the viability of the presented method. The theoretical curves fitted
to the data sets were calculated according to the rate equation with an equilibrium constant (K
XY
). In
this fit, the only varied parameter was K
XY
, or the free energy of dimerization ΔG=−RT lnKxy(Hristova
& White, 2005). Applying the improved method to α
IIb
β
3
wt and (Ala711P) mutant gave rise to ∆∆G of
−3.39 kcal/mol and −3.66 kcal/mol, respectively. The change in the free energy of dimerization due to
the Ala711→Pro mutation is therefore ΔΔG =−0.27 kcal/mol in a POPC LUV system. The measured
ΔΔG=−0.27 kcal/mol is consistent with the previous titration experiments by ITC and NMR (Table

76 | P a g e

2.1).
The structure of the β
3
(A711P,K716A) TMD peptide embedded in phospholipid bicelles was
determined. In accordance with the capacity of Pro to introduce kinks in transmembrane helices,
A711P induced a 30 ± 1° kink between preceding and succeeding helical segments (Figure 2-5.C&E)
caused by its side-chain geometry and the ensuing disruption of helix-stabilizing hydrogen bonds
between Ile 707 and the proline residue and between Gly 708 and Leu 712 (Kim et al, 2011). This kink
precisely separates the α-helical segments of β
3
that constitute the inner and outer clasps. Moreover,
A711P introduced a tilt between these elements that is independent of the overall membrane β
3
TMD
crossing angle. Thus, even if the tilt angle of the inner helix is perturbed, this kink can aid the
formation of both inner and outer membrane clasps to stabilize the αβ dimer (Orzáez et al, 2004b;
Petrache et al, 2000). To verify that both clasps still partake in maintaining the inactive state of the
integrin, mutations that disrupted either clasp were examined. The OMC involves packing interactions
of α
ΙΙb
(Gly 972 and Gly 976) and β
3
(Gly 708). Thus, the kink introduced by β
3
(A711P) allows the
formation of the two membrane clasps required to stabilize integrin in the off-state (Figure 2-6.D).
The structure of the covalently associated α
IIb
β
3
(A711P) TMD complex was determined in the
phospholipid bilayer environment of small bicelles. The bicelle system ensures a physiological,
accurate integrin TM conformations compared to detergent systems. The conformation restrains were
obtained by solution NMR spectroscopy using peptides of various labeling schematics (
2
H/
13
C/
15
N). As
the monomeric α
IIb
and β
3
(A711P, K716A) TMD structures have been previously solved, they serve as
reference for assessing possible changes in backbone conformation upon heterodimerization (Lau et al,
2009). Secondary
13
C
α
chemical shifts exhibit a high correlation with backbone conformation and low
dependence on the chemical environment, which makes them the most sensitive NMR parameter for
assessing changes in the secondary structure. However, despite relatively large changes in chemical
shift of the TROSY spectra (Figure 2-6.A), the secondary
13
C
α
chemical shifts, between monomeric
and heterodimeric states, are less effected (Figure 2-6.B) (Lau et al, 2009).

77 | P a g e

The α
IIb
and β
3
(A711P) TMD helices reside in close proximity at the extracellular membrane
border with close side-chain distances detectable in the hetero NOESY spectrum (Figure 2-6.C)
(Nietlispach & Gautier, 2011). Guided by packing interactions with three distinct glycine residues,
α
IIb
(G972), α
IIb
(G976) and β
3
(G708), the TMD helices align along their N ‐terminal halves (Figure 2-
6.C&D). Due to the proline kink, close interhelical contacts at the OMC do not interfere with the
aromatic contacts at the IMC. Thus, the proline residue allows a crossing angle of ~30° to compensate
for the placement of α
IIb
(F992 ‐F993) between the TMD helices, which results from the distinct α
IIb

backbone reversal C ‐terminal to its TMD helix (Figure 2-6.D) (Lau et al, 2009). The aromatic rings of
α
IIb
(F992 ‐F993) are in proximity to that of β
3
(W715), and the hydrophobicity of this area is augmented
by contacts between β
3
(I719) and α
IIb
(F992‐F993) as well as the hydrophobic moiety of α
IIb
(R995)'s
side-chain. More interestingly, the previously identified electrostatic attractions between α
IIb
(R995) and
β
3
(D723) remains in contact within the relatively low dielectric environment of lipid head groups
(Bogdanov et al, 2009). Thus, integrin α
IIb
β
3
(A711P) forms a TMD dimer complex of unique structural
complexity.
In comparison between the α
IIb
β
3
TMD wt and the proline mutant complex, two association
elements are differentiated: an outer membrane association motif or clasp (OMC) characterized by
packing interactions centered on α
IIb
(G972), α
IIb
(G976) and β
3
(G708) (Figure 2-6.D&E), and an inner
membrane association motif or clasp (IMC) based on the hydrophobic α
IIb
(F992 ‐F993) and electrostatic
α
IIb
(R995)–β3(D723) bridges (Lau et al, 2009). Interestingly, the proline kink allows a separation
between the IMC and OMC. The resulting clasps are more readily able to pack in a low energy
conformation, which mainly presented in the glycine packing of the OMC (Figure 2-6.E) (Senes et al,
2004). Thus, two dominant integrin TMD association motifs are discerned (IMC and OMC), structural
explorations suggest that the α–β TMD affinity of integrins, most notably the OMC, is reassembled
upon proline mutation. In summary, the integrin αIIbβ3(A711P) TMD complex represents the first

78 | P a g e

structure of a heterodimeric TMD receptor of its kind and reveals a dimerization interface of intriguing
complexity.

Conclusion

We have presented convincing evidence that β
3
(A711P) mutation forms a rescue mutation to
counteract the previously dimer disrupting K716X mutation. Data was presented which gives evidence
that, in-vivo as well as in-vitro, the β
3
(A711P) can rescue the K716A mutation, even more interesting it
is actually able to increase the dimer formation in the α
IIb
β
3
wt heterodimer. The α
IIb
β
3
(A711P) complex
formation was characterized, based on ITC, NMR and fluorescence titration experiments. ITC and
NMR, both treated the α
IIb
β
3
complex formation in a bicellular system by observing the heat exchange
or the monomer/dimer population, respectively. The results, even so quantitatively different, argue for
stabilization of the heterodimer upon proline mutation. However, due to the bicelle environment,
monomer/dimer exchange only occur upon bicelle collision, therefore kinetic measurements in POPC
LUV will give further insights to the dimerization process. Therefore, we present a method to
characterize the energetics of heterodimerization in lipid bilayers based on the previous FRET titration
by Merzlyakov et al (2006). The modified method allows to (1) disregard labeling efficiency, as we
observe the acceptor fluorophore and (2) reduce the contribution of random co-localization of donor
and acceptor to less than 2%. We use it to determine the propensity for heterodimer formation between
the α
IIb
β
3
wt TMD and the α
IIb
β
3
(A711P) mutant. Therefore, the presented method to quantify
heterodimer energetics should have a broad utility for studies in induction mechanisms of pathology, as
well as in studies of the general membrane protein folding principles. Ultimately utilizing three
independent methods we were able to characterize the heterodimer stabilization due to A711P mutation
(Table 2-1).

79 | P a g e

Upon kinetic characterization of this phenomena, it became apparent that we needed to
deterimine the structure-to-function relationship. Thus, even if the tilt angle of the inner helix is
perturbed, this kink is able to aid in the formation of both inner and outer membrane clasps, stabilizing
the αβ dimer interface. The monomeric structure, β
3
(A711P/K716A), is characterized by a kink
formation between the IMC and OMC. The uncoupled clasps allow for a better binding interaction and
dismiss the dependencies on either inner or outer membrane clasp. The true nature of the proline
mutation was revealed when addressing the heterodimeric structure of α
IIb
β
3
(A711P), which in addition
to the uncoupling mechanism, revealed a tighter packing of the interfacial glycine (specifically G708),
forming a strong interaction between α
IIb
and β
3
OMC. In summary, we discovered that a Pro-induced
helix-kink was the reason for the stabilization of integrin α and β TMDs, specifically through
facilitating the formation of the inner and outer membrane clasps.

80 | P a g e

Chapter 3 - N-terminal Proline Influences Structure to
Function Relationship in Integrin αIIb Transmembrane
Region

Abstract
Introduction
Results
α
IIb
P965 Influences Linker Region
α
IIb
TMD Pro965A/E Structure and Orientation
α
IIb
TMD (Pro965A/E) Heterodimer Formation
β
c
wt and P441A/E Mutant’s Structure and Orientation
Localized Changes
Discussion
Conclusion

81 | P a g e

Abstract

Prolines are frequently found at borders of transmembrane helices with a likelihood of ~25%,
thus attaching a significant structure-to-function relationship to their occurrence (Ulmschneider &
Sansom, 2001). Utilizing NMR and MD simulation, we determined proline's extended role at the N-
terminal membrane-water interface of single-pass α-helical transmembrane domains in integrin α
IIb
and
interleukin β
c
. Our study exploits proline’s role as a regulatory residue for dynamical and structural
characteristics in the transmembrane and the linker section of α
IIb
. We identified specific mutations in
Pro965 to Ala and Glu which elongate the α-helical structure, leading to a crossing of the N-terminal
membrane-water interface and, consequently, a shortening of the linker region. The shortened linker
region was correlated with restrained dynamics, allowing a tauter coupling between transmembrane
and extracellular domain. The structural characterization of these mutants revealed proline's function
as N-terminal helix-cap and linker terminator. Therefore, Pro965 establishes an equilibrium between
the length of the transmembrane helix and its N-terminal linker domain, regulating the rate a signal
travels between the two domains. Furthermore, similar structural differences were observed for the
common interleukin β receptor and its Pro441A mutant, confirming its significance across the
transmembrane protein family. The structural knowledge provides insights to disease mechanism and
will form a scientific basis for future therapeutic treatments of stroke and cancer.

82 | P a g e

Introduction

Identification of particular structural characteristics of proteins is vital to understand their
molecular function and enable the production of targeted drugs designed to fit to the protein structure
and modulate its function. Membrane proteins comprise 50% of all currently known drug targets, and
therefore present an important subcategory for structural characterization, even though they code for
fewer than a quarter of all proteins in the human genome (Ulmschneider et al, 2010). Single-pass α-
helical membrane proteins present a large family of receptors related to disease, such as stroke (e.g.
integrin receptors) (Harburger & Calderwood, 2009). Thus, the structural characterization of these
membrane proteins will present key insights into cardiovascular disease etiology and drug design.
Previously, we reported that the proline substitution Ala711P in β
3
TM increases dimer
formation due to a 30° kink formation at the substitution site (Kim et al, 2011). Here we report a
conserved proline located at the α
IIb
N-terminal membrane-water interface, which contributes to
α
IIb
TMD's integrity and signaling propensity (Figure 3-1 A). The structural properties assigned to the
N-terminal proline appears to be found in other transmembrane domains with similar structural
characteristics, such as β
c
(Figure 3-1 B). Due to its restraining characteristics, proline is the least likely
amino acid to appear in an α-helix of soluble proteins. However, it occurs frequently in TM helices
with a prevalence of 25% at the middle, and 30% at N- and C-terminal membrane border of an α-helix
(Tieleman et al, 2001; Ulmschneider & Sansom, 2001).
In transmembrane domains local preferences for residues across the membrane spanning
segment allows for signal transfer between two aqueous environments by spanning the hydrophobic
membrane interior. Therefore, a conserved proline could contribute to integrin's integrity (Bowie,
1997). Proline is well known to induce or disturb α-helices (Orzáez et al, 2004a; Rankenberg et al,
2012; Williams & Deber, 1991). Previous studies have identified that the cost of transition due to

83 | P a g e

disturbing in the H-bond network equals to 1 kcal/mol in soluble proteins (Cao & Bowie, 2012).
Therefore, the cost of shifting hydrogen bonds is relatively modest, laying the foundation to introduce
flexibility into transmembrane helices (Bowie, 2011; Cao & Bowie, 2012; Faham et al, 2004). Previous
studies have identified a structural preference for proline residues at the N-terminus of transmembrane
proteins (Cao & Bowie, 2012; Hong et al, 2010; Rankenberg et al, 2012; Tieleman et al, 2001). Proline
substitutions did not cause local effects, but rather cause structural distortions throughout the protein.
Therefore, proline mutations are seldom tolerated, arguing for proline’s functionality upon placement
in membrane proteins. In literature, proline's location was correlated with structural adaptive features
leading to specific protein function(Sansom & Weinstein, 2000; Solomaha & Palfrey, 2005; Tieleman
et al, 2001). Proline induced hinge points are of special interest due to their elaborate influence on
protein dynamics, folding mechanism, and protein interaction. However, proline-linked destabilizing
effects are hard to predict and difficult to attach a functional relationship to, therefore further studies
need to be accomplished. Specifically, significant unresolved questions remain revolving around the
proline-function relationship at the N-terminus, implicating dynamics and orientations of individual
residues on either side of the proline with respect to the membrane-water interface. Here we present a
direct link between N-terminal proline in single-pass α-helical membrane proteins and their
functionality in case of integrin αIIb’s influence on integrin signaling.
In integrin α
IIb
and interleukin β
c,
an N-terminal proline defines the membrane border via a
capping mechanism, in which proline destabilizes the hydrogen-bonding network of the helix’s N-
terminus as well as the physical constrains conveyed by its pyrollidine ring. The latter further
destabilizes the helix at the i-1 position but strongly induces α-helical formation at the i+1 position,
arguing for proline’s N-terminal placement (Tieleman et al, 2001). Drastic changes were observed
upon mutation of the N-terminal proline to alanine or glutamate. The structural basics which underline
these were explored by NMR spectroscopy and molecular dynamics. In either mutation an elongation
of the N-terminal α-helix occurred, which can be prescribed to the addition of the amide hydrogen and

84 | P a g e

removal of the φ angle constrain. Structural analysis of the α
IIb
TM section revealed an elongation of the
α-helical structure in the i-1 direction for ~3 residues. Upon addition of an extracellular domain to the
α
IIb
TMD, the mutation in Pro965 seemed to influence the correlation between them via a mobility
change of the linker region. The elongation of the α-helix has consequence of shortening the linker
region from ~9 to ~6 residues, therefore impeding with signaling necessary linker dynamics. In case of
integrin α
IIb
, no influence of the mutation on the functionality in heterodimer formation was determined.
Therefore, either of the mutations will lead to distress upon a signaling event, due to the regulation in
linker length. Overall, our research gives rise to general understanding in integrin signaling as well as
the differences in between members of the integrin family (Figure 3-1.A). At last, the research forms
the basis for future discoveries in the understanding and treatment of varies diseases (e.g. stroke, heart
attack and several auto immune diseases).

85 | P a g e

A

B

Figure 3-1 Sequence alignment of selected human integrin α and interleukin βc (IL3RB) segments.
Conserved amino acids are colored by the JalView multiple alignment editor using the ClustalX color
scheme. The N- and C-terminal membrane-water interface of the monomeric (A) integrin α and (B)
interleukin TMD subunits are depicted as well as the residue of interest, Pro965 and Pro441. The
expected linker region follows sharply post the N-terminal proline residue, such connecting the
extracellular and TM domain. The alignment of all 18 α subunits shows similar trends.

86 | P a g e

Results

α
IIb
P965 Influences Linker Region
In membrane receptors, crosstalk between extracellular domains and their transmembrane
domain is essential for signal transmission across the membrane. The signal transfer between the two
domains is directly correlated to the mobility of the linker region between them, therefore an
understanding of its linker dynamics leads to insights for integrin activation. GB3A-αIIb is composed
of the GB3A domain attached to αIIb TMD via αIIb's linker region (Figure 3-2 A). In the engineered
protein, the extracellular domain Calf2 is replaced by the smaller GB3A domain in order to increase
signal-to-noise ratio during NMR measurements. GB3A, a structural homolog of Calf2, does not affect
α
IIb
's structural integrity (Figure 3-2 B). The GB3A-α
IIb
domains fold autonomous into their native
structure as indicated by the secondary
13
C
α
shift comparison to the wt GB3 and α
IIb
TMD domains
(Figure 3-2 B). GB3A-α
IIb
assumes native topology upon inserting its TMD into hydrophobic bicelle
and exposes its extracellular domain to the aqueous environment, as shown by the paramagnetic
enhancement protection assay (Figure 3-2 C). In Figure 3-3 A, secondary
13
C
α
chemical shifts identify
secondary structural differences between wt and its Pro965E mutant. The difference between the
observed and tabulated random-coil
13
C
α
,
13
C
β
and
13
C' shifts of a residue define the
13
C
α
secondary
shifts, which correlate with the backbone torsion angles (Delaglio et al, 1995). Besides their agreement
for its GB3A and α
IIb
TMD domains, a shift change at R962-I966 occurs, supporting a structural change
at the N-terminal membrane-water interface upon mutation. The close to random-coil chemical shifts in
the wt, i.e. zero, for N-terminal residues display increased α-helix propensity towards the linker region,
i.e. 1ppm, upon Pro965E mutation.
The Relaxation data, composed of the NOE, T1, and T2 acquisition, serves well to characterize
the effects of Pro965A/E on the inter-domain crosstalk between the TMD and GB3A. NOE results

87 | P a g e

(Figure 3-3 B) confirm the linker's high mobility for residues A958 to I966. A decreased mobility is
seen for E961 to I964 upon P965E mutation, corresponding to increased helicity in the linker region (G
& A, 1982). The combination of Figures 3-3 A & B confirm proline's capping properties regulate
structural preference and mobility in the linker region. GB3A-α
IIb
's total correlation time was calculated
in DHPC/POPC bicelle system for q=0.3 and 0.5. The two domains were treated as separate entities to
evaluate their mobility separately from each other. The τc for the TMD varied between q=0.5 and
q=0.3 due to increased bicelle size, however τc corresponding to GB3A remained fixed as the dynamic
linker uncouples GB3A from its TMD/bicelle system (Table 1) (Chou et al, 2004; Lipari & Szabo,
1982). As shown previously, the linker becomes restrained upon P965E mutation, hence GB3A
displays an increased coupling to its TMD. GB3A's τ
c
for q=0.3 bicelles for both mutation varies as a
function of linker mobility; the dynamic wt and its Pro965E mutant correspond to τ
c
of 8.2(±0.0)ns and
9.3(±0.2)ns, respectively. The extracellular domains mobility decreases by ~20% due to increased
helical propensity in the linker region due to the Pro965E mutation. The results support P965
regulatory mechanism on the inter-domain crosstalk through its coupling/decoupling properties.

88 | P a g e

Construct wt P965E
Q-Factor (POPC/DHPC) 0.3 0.5 0.5
TM 25.2±0.09 28.1±0.47 25.8±0.31
GB3A 8.2±0.00 7.6±0.02 9.3±0.02

Table 3-1 Total correlation time increases upon Pro965E mutation for the extracellular GB3A
domain. The total correlation time (τ
c
) presents a measure of molecular mobility and is based on
measurements of the T1, T2 and NOE relaxation rates. The TM section is placed in the bicellular
environment and therefore tumbles relatively slow due to its size, however its extracellular GB3A
domain is relatively unrestrained due to the flexible linker region. Upon increased bicelle size (q=0.5)
the τ
c
for the TM increases respectively, however GB3A’s τ
c
remains similar. The Pro965E mutation
results in an increase in τ
c
compared its wt counter part due to the restrained linker region.

89 | P a g e

90 | P a g e

Figure 3-2 The engineered GB3A-α
IIb
TMD peptide forms a valid system to characterize structural
properties in the linker region. (A) Conserved amino acids are colored by the JalView multiple
alignment editor using the ClustalX color scheme. The N-terminal TM and the C-terminal extracellular
border of the monomeric α
IIb
subunit are depicted, presenting the linker region between them. Different
constructs were tested under which GB3A-α
IIb
TMD presented itself as most liable. (B) Structural
alignment of the integrin αIIb extracellular domain Calf2 with its structural homolog GB3 shows
similar secondary structure. Furthermore, the location of structural important residues are highlighted
encouraging its substitution. (C) Secondary Carbon shifts compare the observed carbon shifts to the
tabulated carbon shift of a random coil, the difference between them is correlated to the secondary
structure. The similarities between the overlaid GB3A-α
IIb
TMD with the individual GB3 wt and
α
IIb
TMD domains suggest proper folding of the engineered protein. (D) The PRE protection assay
observes the embedding of the GB3A-α
IIb
TMD peptide based on the ratio of TROSY peak intensity
with (I) to without (I
0
) paramagnetic reagent (Mn
2+
EDDA
2-
). The engineered protein does correctly
insert itself into the DHPC/POPC bicelle (q=0.3) based on the comparison to the PRE footprint of
α
IIb
TMD wt. However, the linker region and the extracellular GB3 domain remain exposed to the
aqueous environment.

91 | P a g e

92 | P a g e

Figure 3-3 Secondary structure rearrangement and its effect on the backbone dynamics of the αIIb
linker region. (A) The GB3A-αIIbTMD
13
C
α
chemical shifts were used to identify secondary structural
differences between wt and its Pro965E mutant. Besides their congruence for its GB3A and αIIbTMD
domains, a shift change for R962-I966 was observed, supporting a structural change at the N-terminal
membrane-water interface. Close to random-coil chemical shifts, i.e., zero, were shifted towards α-
helical values, i.e. 1ppm, increasing the α-helix propensity at the linker region. (B) Pico- to nanosecond
dynamics, correlative to backbone order (44), were evaluated in the form of heteronuclear [1H]-15N
NOE values. High values, as seen for the extracellular GB3A or the TMD, correspond to restrained,
structured residues, whereby low values rather present mobile unordered residues. Values for L956-
I964 support a relatively mobile linker in the wt, which becomes more ordered upon P965E mutation.
(C) Modeled structure of the engineered fusion peptide, GB3A-αIIbTMD, in its native lipid
environment was calculated by MODELLER. The structure depicts the graphical presentation of (A),
(B), most importantly the linker region is presented by a highly dynamic random-coil, which allows for
a lose coupling between TMD and GB3A.

93 | P a g e

α
IIb
TM Pro965A/E Structure and Orientation
Mutating Pro965 to either Ala or Glu will manifest structural changes in the protein backbone
of the αIIb peptide. The changes are modulated by the addition of an amide hydrogen and removal of
the proline restrains on the i-1 and i+1 residue (Tieleman et al, 2001). Significant changes occur in the
secondary structure, as indicated by the
13
C
α
secondary shifts. Figure 3-5 A compares the secondary
structure of α
IIb
wt and its P965A/E mutants as calculated from the
13
C
α
,
13
C
β
,
13
C’ shifts. An α-helical
formation identifies with
13
C
α
secondary shift values above 1 ppm, whereby values between 1 to -1
ppm will rather refer to a random coil (Maltsev et al, 2012). The wt and P965A/E mutants are
structurally tantamount between residues W968-R997, but distinct differences arise at the N-terminus,
E960 to W987. In the given residue range, the wt will render to a random-coil but the P965A/E mutant
reflects an increased α-helical propensity. In combination with RDC and relaxation measurements, the
precise helical dimension encompasses residues R962 to R997 in both mutations (Bax, 2003). These
results are in agreement with Conchran et al., describing prolines at N1 position promote helix
formation for residues in the i+1 direction, but disfavors α-helical formation for residues in the i-1
direction.
Membrane protein function is partly regulated by the orientation inside the membrane bilayer
(Conforti et al, 1990). The orientation is composed of membrane embedding, helical tilt, and helical
rotation as compared to the membrane normal. In Figure 3-4 A the amide proton chemical shifts are
presenting changes in the chemical environment. Upon mutations in Pro965 to either Ala or Glu a
relatively large shift change occurs at surrounding residues in the i-1 and i+1 direction, presenting an
uptake of a new conformation. The overlaid TROSY spectra of the wt and its Pro965A/E mutants
reveals the impact the Pro965 mutations on the peak location for individual residues (Figure 3-4 C). In
Figure 3-4 B, similar to the H
N
, the amide nitrogen chemical shifts are additional influenced by the
protein secondary structure. The shift changes in both Pro965A and Pro965E suggest a structural
reorientation at the N-terminal membrane-water interface. PRE protection assay was used to initially

94 | P a g e

describe αIIb's membrane embedding as a function of Pro965 mutation. The paramagnetic reagent,
Mn
+2
EDDA
-2
, causes signal broadening for nuclei in close approximation, as described by the Soloman
equation(Lau et al, 2008a; Lau et al, 2009; Lau et al, 2008b). Mn
+2
EDDA
-2
, a polar reagent, prefers a
hydrophilic aqueous environment over the hydrophobic membrane interior. The strict separation
between those two environments in the membrane bilayer results in a steep Mn
+2
EDDA
-2
gradient,
which is reflected in the sequential
1
H-
15
N TROSY signal broadening. In Figure 3-5 B signal
broadening is quantified by the normalized ratios of
1
H-
15
N TROSY signal intensities in the presence
or absence of 1 mM Mn
+2
EDDA
-2
(I/I
0
). The magnitude change in the amino acid sequence reflects the
membrane borders of the peptide at the N- and C-termini (Killian & von Heijne, 2000); residues
exposed to aqueous environment have a higher probability to come in close contact with Mn
+2
EDDA
-2

compared to residues in the hydrocarbon region. The N- and C-terminal membrane borders for the
α
IIb
TM wt are placed at P965 and K989, respectively; however the Pro965A mutation shifts the
borders by 3 residues in the i-1 direction at the N-terminus, from P965 to R962. The opposite holds true
for α
IIb
(P965E), whereby the membrane border shifts 3 residues towards the interior of the membrane,
from residue P965 to W968. Furthermore, a difference in relaxation pattern is observed in the mutants
compared to the wt, specifically at the N-terminal membrane interface. Interestingly the N-terminal
boarder was shifted by three residues in the T1, T2, and NOE graphs as well (G & A, 1982). Additional
support comes from MD simulation data, which gained similar results in embedding and relaxation
data. Furthermore, upon measuring the exposure of protein residues to water at a radius of 4Å the
protection assay can be confirmed on a molecular basis. Thus, Pro965 exhibits structural characteristics
to regulate membrane embedding of the α-helical conformation.

95 | P a g e

Figure 3-4 N-terminal proline 965 influences chemical surrounding of close by residues. (A) The
amide proton chemical shifts are heavily influenced by changes in the surrounding chemical
environment. Upon mutating Pro965 to either Ala or Glu a relatively large shift change occurs at
surrounding residues in the i-1 and i+1 direction. (B) Similar to the
H
N, the amide nitrogen chemical
shifts are additional influenced by the protein secondary structure. The shift changes in both Pro965A
and Pro965E suggest a structural reorientation at the N-terminal membrane-water interface. (C) The
overlaid TROSY spectra of the wt and its Pro965A/E mutants reveals the impact the Pro965 mutation
have on the peak location.

96 | P a g e

97 | P a g e

Figure 3-5 Structural and topological comparison between and αIIb and its Pro965A/E mutants. (A)
Comparison of αIIbTMD's backbone structural parameters between the wt and its Pro965A/E mutants.
The αIIbTMD 13Cα chemical shifts obtained from the wt and its mutants present indicators for the
secondary structure. A strong congruity is observed for residues I966 to R995 which is presented by an
α-helix; however shift differences of ~1.5 ppm occur for residues E960 to P/A/E965. Changes pinpoint
alterations in secondary structure, shifting it from elongated (~0 ppm) to α-helical (>1ppm)
conformation upon P965A/E mutation. (B) Comparison of the normalized ratios of H-N TROSY signal
intensities in the presence or absence of 1mM Mn
+2
EDDA
-2
, I/I0, between α
IIb
TMD and its P965A/E
mutants embedded in DHPC/POPC (q=0.3). The α
IIb
TM wt , however the N-terminal border is shifted
by 2 residues upon P965A mutation. (C) The structure of α
IIb
TMD Pro965A was calculated as
discussed earlier, the wt and mutant structures are directly compared upon embedding in the membrane.
The topology is based on (B) and MD simulations, in such the α-helix crosses the lipid head group
region into aqueous environment. α
IIb
TMD Pro965E is expected to have a comparable structure to
Pro965A as (A) does not significantly differentiate between them.

98 | P a g e

α
IIb
TMD Heterodimer Formation
The integrin α
IIb
and β
3
transmembrane domains can form a heterodimer, presenting a
regulating mechanism of integrin activity (Harburger & Calderwood, 2009). Quantification of the
equilibrium between the α
IIb
and β
3
transmembrane subunits give insights on regulatory mechanisms
governing the dimer formation (Hughes et al, 1996; Kim et al, 2009b; Ulmer, 2010). In our study, the
equilibrium constants were determined by NMR; specifically, TROSY spectra were recorded at
different αIIb-to-β3 molar ratios. An increase of α
IIb
(or its mutants), while holding β
3
constant,
increases α
IIb
/β
3
heterodimer population. The two molecular populations lead to the formation of a
separate set of heterodimer specific TROSY peaks. The monomeric peak volume decrease, as referred
by residue β3 Gly702 (or the average of L698, G702, L705, L706, A710, and L717) as seen in Figure
2B, is directly proportional to the change in hetero-dimer fraction upon increase in α
IIb
TMD
concentration. The equilibrium constant is extrapolated, allowing to calculate the change in free energy
of dimerization via ΔG=-RTlnK
XY
(Son et al, 2012; Tellinghuisen, 2008; Turnbull & Daranas, 2003).
The difference in change of Gibb's free energy between the wt and either of the mutants determines its
favoritism towards the monomeric or dimeric state. The integrin titration curves of α
IIb
/β
3
TMD,
α
IIb
(P965A)/β
3
TMD, and α
IIb
(P965E)/β
3
TMD are shown in Figure 3-6 C, in addition their individual
equilibrium constants are shown in Table 3.2. The best estimates for the free energies of dimerization,
derived from various optimization schemes, are ΔG= -4.57±0.01 kcal/mol for α
IIb
(wt), ΔG= -
4.31(±0.02) kcal/mol upon Pro965A mutation, and ΔG= -4.76(±0.04) kcal/mol Pro965E mutation.
Therefore, the change in the free energy of dimerization due to the Pro965A mutation corresponds to
ΔΔG= 0.26 kcal/mol and upon Pro965E mutation is ΔΔG= -0.19 kcal/mol. Therefore Pro965A slightly
favors the monomeric species, whereby the Pro965E marginally favors the dimer formation. The
change is minimal and therefore is not expected to significantly contribute to the TMD dimer formation
under physiological conditions.

99 | P a g e

K
XY
ΔG(kcal/mol) ΔΔG(kcal/mol
αIIb/β3TMD 2070 -4.57±0.01 -
αIIb(P965A)/β3TMD 1340 -4.31±0.02 0.26
αIIb(P965E)/β3TMD 2850 -4.76±0.04 -0.19

Table 3-2 αIIbTMD Pro965 mutation to either Ala or Glu have minor effects on the heterodimer
formation. The change in the free energy of dimerization due to the Pro965 mutation to Ala or Glu
corresponds to ΔΔG= 0.26 kcal/mol and ΔΔG= -0.19 kcal/mol, respectively. Therefore Pro965A
slightly favors the monomeric species, whereby the Pro965E marginally favors the dimer formation.
However, the change is minimal and therefore is not expected to significantly contribute to the TM
dimer formation under physiological conditions.

100 | P a g e

101 | P a g e

Figure 3-6 α
IIb
TMD Pro965A/E mutation does not significantly disturb α
IIb
/β
3
TMD
heterodimerization. (A) Schematic view of a lipid bilayer with two TM helical species, α
IIb
(Pro965A)
and β3, associating into heterodimer α
IIb
(Pro965A)/β
3
. The monomeric and dimeric species give rise to
individual TROSY peak sets. (B) The β
3
monomer (0.1 mM) is titrated with α
IIb
allowing to form the
dimer species. The monomeric Trosy peak volumes decrease upon complex formation and
subsequently increase heterodimeric peak volumes. (C) Energetics of α
IIb
TM wt and β
3
TMD
dimerization in POPC compared to its Pro967A/E mutants. The free energy of dimerization was
determined by fitting the theoretical curves derived from the rate equation to the experimental data, as
described in Materials and Methods. The best estimates for the free energies of dimerization, derived
from various optimization schemes (see Table 2 and text), are ΔG= -4.57±0.01 kcal/mol for αIIb(wt),
ΔG= -4.31(±0.02) kcal/mol upon Pro965A mutation, and ΔG= -4.76(±0.04) kcal/mol Pro965E
mutation. Therefore, the change in the free energy of dimerization due to the Pro965A mutation is
ΔΔG= -0.26 kcal/mol and upon Pro965E mutation is ΔΔG= 0.20 kcal/mol. Therefore Pro965A favors
the monomeric species, whereby the Pro965E favors the dimer formation; however we notice the
change is minimal and therefore is not expected to significantly contribute to the TM dimer formation
under physiological conditions.

102 | P a g e

β
c
wt and P441A/E Mutant’s Structure and Orientation
The importance of N-terminal prolines, on the conformation uptake of single-pass
transmembrane domains, was presented on integrin α
IIb
. The transmembrane domain of common
interleukin receptor subunit βc was characterized to ensure that α
IIb
Pro965 does not present an isolated
singular event, but applies to a wider range of α-helical transmembrane peptides. Similar to α
IIb
, β
c

harbors a proline (Pro441) at the N-terminal membrane interface, located strategically between the α-
helical segment and the extracellular random coil (Figure 3-1 B). Localized changes in the chemical
environment upon Pro441A mutation, as presented by the TROSY spectra overlay of the wt and mutant
TM segment (Figure 3-7 C). The amide proton and nitrogen chemical shift changes are plotted in
Figure 3-7 B and A, respectively, and give insights to changes in conformation and chemical
environment. In the wt, an α-helical conformation is observed between residues Pro441 till G464 as
indicated by the secondary shifts. A sudden shift from random coil to α-helix occurs at Pro441,
signifying its function as N-terminal helix cap and helix inducer in the i+1 direction (Figure 3-8 A).
Both the wt and its mutant are tantamount in their backbone conformation between residues W443 till
R472, but drastic differences arise towards the N-terminal membrane interface. Upon Pro441A
mutation the β
c
helix extends in the i-1 direction, elongating the helix by ~5 additional residues at the
N-terminal membrane-water interface. The mutant conformation stretches between T436-G464,
shifting the initial N-terminal helix border at Pro441 to T436. The PRE protection assay further
stretches the importance of P441 on membrane embedding (Figure 3-8 B). The membrane N- and C-
terminal membrane borders were identified at Pro441 and G464, respectively, embedding 23 residues
in the POPC bilayer. Pro441A mutation causes the N-terminal membrane border to shift in the i-1
direction for an additional residue, stabilizing at L440. Even so the backbone conformation is deeply
impacted by the mutation, the membrane orientation remains rather similar to the wt. Taking in account
the PRE protection assay, Δδ, and secondary carbon shifts, P441 exhibits conformational control on
βc's N-terminal membrane interface similar to P965 on α
IIb
. The structure calculation for the wt and its

103 | P a g e

Pro441A were performed as previously mentioned in Materials and Methods, including RDC, χ1, and
NOE measurements. The finalized structures are presented in Figure 3-8 C. The orientation in the
bicelle is based in the previous mentioned PRE protection results.

104 | P a g e

Figure 3-7 N-terminal proline 441 influences chemical surrounding of close by residues. (A) The
amide proton chemical shifts are heavily influenced by changes in the surrounding chemical
environment. Upon mutating Pro441 to either Ala a relatively large shift change occurs at surrounding
residues in the i-1 (~5residues) and i+1 direction. (B) Similar to the H
N
, the amide nitrogen chemical
shifts are additional influenced by the protein secondary structure. The shift changes in both Pro441A
suggest a structural reorientation at the N-terminal membrane-water interface. (C) The overlaid
TROSY spectra of the wt and its Pro441A mutants reveals the impact the Pro441 mutation have on the
peak location.

105 | P a g e

106 | P a g e

Figure 3-8 Structures and topological comparison of interleukin β
c
TMD and its Pro441A mutation.
(A) Sequence alignment of selected human interleukin receptor segments. The N- and C-terminal
membrane-water interface of the monomeric interleukin subunits are depicted as well as the residue of
interest, Pro441. (B) Comparison of the backbone structural parameters of the βc transmembrane
segment between the wt and the Pro441A mutant. The α
IIb
TM sary 13Cα chemical shifts obtained from
the wt and its P965A/E mutants are compared. Despite a strong agreement for residues L445 to R473,
a significant shift difference of ~3.0 ppm occur at residues T436 to L440. The difference suggests a
strong conformational preference for α-helical structure upon P441A mutation. (C) Comparison of the
normalized ratios of 1H-15N TROSY signal intensities in the presence or absence of 1mM
Mn
+2
EDDA
-2
, I/I
0
, between βcTMD and its P441A mutant embedded in DHPC/POPC (q=0.3). Solid
blue line mark the membrane borders of the α
IIb
TMD wt based on the paramagnetic protection
environment of a membrane. (D) The structure of β
c
TMD wt and its Pro441A mutant were calculated
as discussed earlier, the wt and mutant structures are directly compared upon embedding in the
membrane. The topology is based on (B), in such the α-helix crosses the lipid headgroup region into
aqueous environment.

107 | P a g e

Localized Changes
Previous results supported the hypothesis that proline in the N-terminal membrane water
interface regulate linker dynamics as well as helical conformation (Wang et al, 1999). NMR suits the
phenomenon as it is capable of determining high resolution protein structure and measure localized
dynamical changes (Nietlispach & Gautier, 2011). The TROSY spectrum observes local changes in
orientation and mobility at the
1
H
N
or
15
N
H
dimension, due variations in the chemical environment. The
peak intensities are influenced by mobility of specific nuclei, whereby shift changes correlate with
alterations in orientation and chemical environment. Figure 3-9 B compares the membrane normalized
TROSY intensities of individual α
IIb
and β
c
residues in a DHPC/POPC bicelle. Broadened peak
intensities, compared to the norm (dashed line), present a change in dynamics. In both proteins
intensities fell steeply subsequent the N-terminal proline for the following six residues (i+1...6). The
initial 3 residues shared the lowest intensity followed by a slow recovery for the next three residues.
Additionally, both proteins at least one Trp is located in the i+2 position of proline, which are known to
function as membrane anchor (Ridder et al, 2005; van der Wel et al, 2007; Vostrikov et al, 2010).
Figure 3-9 A presents a comparison of the signal intensities for α
IIb
and its Pro965A/E mutants; upon
mutation the signal intensity for I966 and W967 increase to membrane normal, relating the signal
decrease to the occurrence of an N-terminal proline residue. The intensity change is more significant in
the P965A mutant. These results suggest that proline heavily influences molecular dynamic
characteristics at the membrane interfaces. Structural changes can also be accessed via TROSY spectra
by comparing wt and mutants’ chemical shifts in the
15
N and
1
H
N
dimensions. In Figure 3-4 A and B,
significant chemical shift differences are observed in the N-terminal region for both αIIb P965A and
P965E for 6 residues post Pro965 (i+1...6). However, the magnitude of change is larger for P965E
mutant compared to the Pro965A mutation. The difference arises due to P965E negative charge. The
charge is placed in close approximation to the hydrophobic membrane, disturbing the membrane
integrity and causing α
IIb
to uptake a new topology. Similar results present themselves comparing the

108 | P a g e

signal intensities for β
c
wt and its Pro441A mutant. As previously mentioned, the signal decrease for
post Pro441 residues (Pro441-L447) identifies a restrained behavior at the N-terminal membrane
interface. The shift change underlines important structural aspect of proline's capping mechanism,
enhancing its significance at the N-terminal region. The combination of these results indicate that P965
defines the backbone structure and dynamics in the i+1 direction, underlining its function as capping
residue.

109 | P a g e

110 | P a g e

Figure 3-9 N-terminal membrane-water interface residue restrains. (A) Examination of the
normalized
1
H-
15
N TROSY signal intensities between α
IIb
TMD and β
c
suggests a common trend by
broadening signals for 6 residues in the i+1 direction (α
IIb
: [Pro965-V971]; βc[Pro441-L447]). (B)
Normalized
1
H-
15
N TROSY signal intensities between α
IIb
TMD and its P965A and P965E mutants
embedded in DHPC/POPC (q=0.3) bicelle. A strong decrease in signal intensity for residues I966 to
G972 occurs post P965; during the Pro965A/E mutation an increase in signal intensity for the
immediate following residues is observed. (C) Normalized
1
H-
15
N TROSY signal intensities between
β
c
TMD and its P441A mutant embedded in DHPC/POPC (q=0.3) bicelle. A strong decrease in signal
intensity for residues V442 to L447 suggest a similar behavior at the N-terminal membrane-water
interface as α
IIb
; the Pro441A mutation significantly increases all six residues in the i+1 direction to the
membrane normal (dashed line).

111 | P a g e

Discussion:

N-terminal Helic Cap vs Linker Terminator
The αIIbTMD α-helix is characterized by a consecutive, main-chain, i → i – 4 hydrogen
bonding network between each amide hydrogen and the carbonyl oxygen of the adjacent helical turn
(Cao & Bowie, 2012; Yohannan et al, 2004). The N- and C-terminal ends of the TM helix are
characterized by a disturbed H-bond matrix, known as a helix-cap (Cao & Bowie, 2012). Helix-capping
is instrumental in the stability and function of proteins, as such multiple motifs have evolved (Bowie,
2011). Its functional relevance is generally associated with structural stabilization, which may apply to
associated diseases and disorders (Hodivala-Dilke et al, 1999). It is well documented that prolines play
an essential role in the N-cap position at which it serves as a helix breaker due to the absence of its
amide proton. Here we extend the definition from helix breaker to single-pass α-helical transmembrane
domains in which proline, additionally, functions as a linker terminator. The structural and functional
aspects were investigated on α
IIb
and β
c
TMD by mutating the N-terminal proline to either Ala or/and
Glu. In the wt, the N-cap prohibits the formation of H-bonds with the protein backbone but rather
forms bonds with water (Williams & Deber, 1991). Proline’s function as a helix-cap is promoted by the
absence of proline’s amide hydrogen and therefore disabling the H-bond formation necessary for a
stable helix turn. Furthermore, proline’s pyrollidine ring locks the backbone φ angle at approximately
−60 which induces an α-helix in the i+1 direction (φ= -60±15º) (Rankenberg et al, 2012; Sansom &
Weinstein, 2000; Tieleman et al, 2001; Williams & Deber, 1991). These restrains promote proline’s
function as a linker terminator. Upon mutation to Ala/Glu, an amide hydrogen is introduced at proline’s
position in the protein backbone, dismissing its function as a helix-cap. A new N-cap is formed in the i-
4 position via the interaction of charged residues (Glu, Arg) with the amide protons, stabilizing the N-
terminus [49]. Further implementation affects the linker region in αIIb, between the TMD and its
extracellular domain, Calf2. The mutations remove proline’s restraints on the protein backbone,

112 | P a g e

therefore disabling its function as linker terminator (Figure 2C). Our research establishes N-terminal
proline’s function as a regulator of the equilibrium between linker region and α-helix conformation in
single-pass α-helical TM receptors.

α
IIb
Linker Domain
Signaling across the membrane by transmembrane proteins implies that the exterior and interior
domains are inter-communicating through the TM segment. The protein architecture and function
relationship heavily relies on the flexibility of a linker region that interconnects various domains of the
protein. Jesut et al (2004) established a clear relationship between linker peptides and the functional
dynamics they enable. A flexible linker has three functions: (1) to allow for interaction, (2) to increase
separation, and/or (3) to maintain distance between domains. In transmembrane signaling proteins,
linkers can play an essential role in maintaining cooperative inter-domain interaction between the
extracellular, TM and intracellular domains. In case of integrin, its linker dynamics appears to be key
for the signal transfer. Previous literature explored the elongation of the linker region upon inserting of
a poly-glycine stretch resulted in a constantly active integrin α
M
β
8
. The flexible poly-Gly stretch
introduced additional mobility, leading to the uptake of integrin's high affinity state. Therefore, great
interest encircles the linker’s ability to influence the extracellular and TM interactions. Identification of
key residues mediating this synergy forms the basis for our research. Integrin αIIb identifies an N-
terminal linker sequence between L955-A963 (Figure 3-1 A). The linker length and amino acid
composition mirrors its identification as a medium size linker with ~9 residues and high occurrence of
Glu and Arg (Ulmer et al, 2001). The linker database http://mathbio.nimr.mrc.ac.uk identified a similar
linker sequence of known structure in tRNA splicing endonuclease increasing the linker propensity. In
this model, we utilized NMR to record changes in mobility/dynamic tendencies at the linker region
between the extracellular domain and its transmembrane domain (Figure 3-3 B). In membrane
receptors (e.g. integrins), the linker region dictates the transmission across the membrane, behaving

113 | P a g e

either sluggish or rather rapid. Previous literature reported that a mobile linker enables two system to
act independently due to the lose coupling between them [53]. However the opposite is true for a
restrained linker, in which a tighter coupling allows a rapid signal transmission [58]. The mutation
P965E was shown to increase helix propensity in the Linker region (Figures 3-3 A, 3-5 A), hence it is
expected to induce a tighter coupling. Figure 3-3 C presents the model peptide GB3A-α
IIb
TMD
integrated in membrane environment, the α
IIb
TMD section is fully emerged into the bicelle
environment and both domains fold into their native low-energy structure. The bicelle environment
heavily restrains the α
IIb
section of the protein, however the extracellular GB3A domain is rather
mobile (τ
c
=8.2ns), characteristic of a flexible linker region. Table 3 presents the correlation times (τ
c
)
for the TM (restricted) and the extracellular domain (dynamic) for the wt in q-factor 0.3 and 0.5
bicelles. It is noticed that τ
c
for the TMD increases with the increased bicelle size, however due to the
mobile linker region GB3A domain is relatively uncoupled from the bicelle restrains, therefore τ
c

remains relatively constant. The N-terminal residue in the linker region occurs to be a proline, which is
well known for its linker propensity due to its random coil induction in the i-1 direction. Figure 3-1 A
presents the frequent occurrence of Pro965 at the membrane-water interface across the integrin family,
however variations occur (e.g. α
3
, α
D
). In the GB3A-α
IIb
TMD wt, Pro965 acts as a linker terminator/N-
terminal helix-cap due to its restraining properties. If Pro965 is replaced by Glu, as in integrin α3, the
helix propensity increases in the linker region as evident by Figure 3-3 B. Removal of the proline
restrains extends the α
IIb
TMD helix in the i-1 direction, decreasing the linker by 3 residues from ~9 to
~6 amino acids transferring it into the rather small linker domain. Upon P965E mutation the GB3A
domain increases on correlation time (τ
c
=9.2ns) suggesting a rather restrained linker region, compared
to the wt. The decrease in mobility is mirrored by the NOE relaxation (Figure 3-3 B) rate in this region;
the α-helix is able to efficiently retain a conformational restrain at the membrane-water interface
causing the NOE values to differ from the wt. The extended helix shortens the length of the flexible
linker, which in turn decreases its mobility. The data has important implications for the integrin family,

114 | P a g e

as it indicates that the helix length at the N-terminus partially controls integrins inter-domain crosstalk
and therefore is accountable for variations in between the integrin family.

Structure and Orientation Comparison between α
IIb
TMD and its P965A/E Mutant.
The α
IIb
TMD bicelle embedded structure was previously solved by Lau at al (2008) and several
important residues have been determined to present regulators of integrin signaling (Lau et al, 2008a;
Lau et al, 2008b). P965 is a conserved residue located at the N-terminal membrane-water interface,
forming the N-terminal helix-cap for the TMD. We calculated the structure of α
IIb
TMD (Pro965A)
based on various structure restrains as described in the materials and methods. Its Pro965E mutant was
expected to expose similar structural behavior to Pro965A as judged by Figure 3-5.A, therefore
structure modelling redeemed fitting for the structure evaluation. As previously mentioned the absent
H-bond, between the helix-turn, shifts the secondary structure from α-helix to random-coil in the i-1
direction (Figure 3-5 A&C). In both mutations, Ala or Glu, the helical content increased upon addition
of the amide hydrogen. Such increasing the α-helix length by an additional 3 residues. The original N-
cap solely dependent on the proline’s function as helix breaker, however upon its removal a new N-cap
forms. The new N-cap encompasses residues E960 and R962 forming a big-box motif [49]. The energy
comparison of H-bonds located in hydrophilic or hydrophobic environment 1.5 kcal/mol to 5 kcal/mol
(Bowie, 2011). Therefore it appears of great interest to locate the newly formed H-bonds in a relative
hydrophobic environment, reasoning a slight shift in Figure 3B towards the membrane interior. Further
transitions are inhibited by a sequence of highly charged residues (E960, E961, R962), which disfavor
the membrane environment. The same holds true for α
IIb
Pro965E mutant in which a shift occurs in the
opposite direction due to its negative charge, which is being extruded from the membrane surroundings.
Figure 3-5 C depicts the described circumstances, the α
IIb
TMD wt structure is aligned as previously
shown by Lau et. al. in a orthogonal fashion compared to the membrane. The α
IIb
Pro965A structure is
slightly tilted to accommodate the newly formed H-bonds, and its Pro965E mutant assumes a similar

115 | P a g e

conformation as the wt. The TROSY spectrum identifies a signal intensity pattern in which six residues
following the Pro965 have decreased signal intensity (Figure 3-9 A), which might identify their
restraining surroundings. However upon mutation the signal increases pinpointing an increase in
residue dynamics. The dynamics are additionally stabilized by post-proline W967 and W968 located at
the N-terminal membrane-water interface. A change in χ1 angles was observed, in the tryptophan pair
following the proline loosening W967 (Shahidullah & London, 2008; Sparks et al, 2014; van der Wel
et al, 2007; Vostrikov et al, 2010). The tryptophan are well established to function as membrane anchor,
at which point the χ1 angle is important in stabilizing the energetically favorable conformation.
Attaching a significant function to P965 as a regulator of Trp orientation. Summarizing the results, N-
terminal Pro965 has a significant effect on length, dynamics and orientation of α
IIb
TMD segment.
The integrin α
IIb
and β
3
transmembrane subunits associate to form a heterodimer, α
IIb
β
3
(Figure
3-6 C). To present, several α
IIb
TMD regulatory residues of integrin signaling events were identified,
upon its interaction with β
3
TMD. Specifically, snorkeling phenylalanines (F992 and F993), glycine
packing and salt bridge formation have shown to mediate the association with the β3-subunit. Due to its
impact and the α
IIb
TMD helix, Pro965 might convey a similar regulatory mechanism on the TM
heterodimer formation. In our study, the association rates were determined by NMR; specifically,
TROSY spectra were recorded at different α
IIb
-to-β
3
molar ratios (as described earlier) free energy of
dimerization via ΔG=-RTlnK
XY
. The difference in change of Gibb's free energy between the wt and
either of the mutants determines its favoritism towards the monomeric or dimeric state is stabilized by
the specific mutation. The integrin titration curves of α
IIb
/β
3
TMD, α
IIb
(P965A)/β
3
TMD, and
α
IIb
(P965E)/β
3
TMD are shown in Figure 3-6 A. Therefore, the change in the free energy of
dimerization due to the Pro965A mutation is ΔΔG= -0.26 kcal/mol and upon Pro965E mutation is
ΔΔG= 0.19 kcal/mol. Quantification of the equilibrium constants, between the αIIb and β
transmembrane subunits support a minor preference of Pro965A for the monomeric species, whereby
the Pro965E marginally favors the dimer formation; however, the change is minimal and therefore is

116 | P a g e

not expected to significantly contribute to the TM dimer formation under physiological conditions.

Comparison to interleukin β
c

Cytokines in the GM-CSF/IL-3/IL-5 family are responsible for actuating a conformational
change in their receptors, transducing a signal across the membrane that governs hematopoetic cells to
survive or apoptose, proliferate, differentiate, migrate, or reactive oxygen species release (Rourke et al,
1996). During an allergic inflammation of the lung an up regulation of those cytokines occurs causing
an increase in eosinophil number and their activation, which allows for different phases of eosinophil
infiltration and determine a localized versus a generalized eosinopil-mediated inflammation. Structural
evaluation of the interleukin β
c
TMD domain shows similar behavior as seen previously in α
IIb
TMD. It
appears essential to structurally illuminate this interaction, to allow modulation of β
c
and such
counteract diseases related to GM-CSF/IL-3/IL-5 family signaling (Rourke et al, 1996). We were able
to determine both, the β
c
TMD wild type and its mutant β
c
TMD Pro441A structure via various NMR
restrains, as described in materials and methods. Structurally, β
c
TMD mirrors α
IIb
TMD, forming an α-
helix between residues Pro441 to G464, which is terminated at the N-, and C-terminal membrane-water
interface by Pro441 and Cys463/Gly464, respectively (Figure 3-8 A). It can be safely suggested that
Pro441 forms a helix cap, therefore restricting the length of the helix. Proline’s function arises due to
its lacking amide hydrogen, which denies H-bond formation with the residue in the i+4 position.
Similar to αIIb, Pro441 induces helix formation in the i+1 direction and supports random coil
conformation in the i-1 direction, due to its pyrollidine ring side chain. Upon mutating Pro441 to
alanine the α-helix extends by four additional residues towards the N-terminus, compared to the wild
type (Figure 3-8 A). However, the change in secondary structure has a relative minor impact on
membrane water interface as suggested by the paramagnetic protection assay in Figure 3-8 B, similar to
the αIIb Pro965A mutation. Even so H-bonds are more stable in the membrane interior, the relative
high occurrence of electronegative residues at the helix extension makes it unfavorable to be included

117 | P a g e

in the bicelle. Therefore, the Pro441Ala mutation allows the helix to cross the N-terminal memberane-
water interface, restraining the linker region to the extracellular domains. Analyzing the TROSY
spectra, the N-terminal amino acid composition of β
c
modulates the signal intensities, causing a
broadened peaks in post Pro441 residues (i-1 direction); showing a similar dynamical behavior as seen
in α
IIb
TMD (Figure 3-9 B). The signal decrease correlates well to the restraining surroundings of the
effected residues. Interestingly, Pro441 mutation to alanine allows an increase in the otherwise
broadened signal of the N-terminal residues. The change in signal intensity is even more significant as
found in α
IIb
TMD Pro965A. Additionally, the occurrence of W443 as membrane anchor at the N-
terminal membrane-water interface, refers to its restraining properties as suggested by the χ1 angle (S.
Table 3). A change in W441’s χ1 angles was observed upon Pro441A mutation, allowing for an
increase in mobility. In conclusion, the structural features proclaimed towards Pro965 in α
IIb
are not an
isolated event but rather are found across other single-pass TMDs (e.g. β
c
) featuring its significance
throughout the membrane protein family.

118 | P a g e

Conclusion

In conclusion, Pro965’s impact on the linker region, helix formation and its inability to significantly
regulate heterodimer formation underlines its regulation on the linker region dynamics. Therefore
Pro965 has both helix cap and linker terminator properties, and acts by moderating the equilibrium
between them. The transmembrane domain region in integrin α
IIb
β
3
is a key component during the
signaling mechanism, it allows the transfer of signal from the extracellular towards the intracellular
environment, and vice versa. In integrin receptors the signaling mechanism is characterized by a cross-
talk between the extracellular, transmembrane, and intracellular domains.
Here, we characterized the N-terminal Pro965 of integrin α
IIb
TMD as a regulator of the linker dynamics.
Proline has shown to convey structural mobility to its prequel residues by functioning as a helix breaker
in the i-1 direction and a linker terminator in the i+1 direction. Furthermore, two mutations (Pro965E/A)
were isolated which eliminate prolines helical-cap formation and elongate the TM helix in the i-1
direction. The P965A structure shows the extension of the α-helix by an additional 3 residues compared
to the wt, which we show to contribute to a decrease in linker dynamics. At last our observations do not
restrain itself to only integrin, but also were shown to hold true for other transmembrane receptors such
as interleukin β receptor (β
c
). β
c
showed similar structural effects as α
IIb
TMD. It is in our believe that
our observation will contribute to the general appreciation of transmembrane proteins and eventually
give rise to a clearer understanding and treatment of its related diseases, such as stroke or heart.

119 | P a g e

Chapter 4 - Materials and Methods

Integrin α
IIb
Expression and Isolation
The integrin α
IIb
TMD expression construct was previously prepared by Lau et al. (2008) [23]. In short,
the gene of interest (α
IIb
TMD: G955 to P997) was subcloned into pet-44 expression vector (Novagen),
which included the third IgG-binding domain of protein G (GB3) as N-terminal fusion peptide. To
separate the α
IIb
TMD from its fusion peptide during the isolation process, an intervening tobacco etch
virus protease cleavage site was introduced between the C-terminal and N-terminal of GB3 and
α
IIb
TMD, respectively. The α
IIb
(P965A) and (P965E) mutants were prepared via site-specific
mutagenesis using the QuikChange protocol (Novagen). Expression and Isolation preceded as
previously discribed. Isotope labeling was achieved by supplying
15
N-NH
4
Cl and
13
C-D-glucose and
culturing in 99% D
2
O for highly deuteriated peptide.

NMR Sample Preparation
The peptide concentration was measured in a 3-to-1 acetonitrile-water solution by UV spectroscopy
(ε
280
nm[aIIb]=16500 M-1cm-1). Specific peptide fractions at various concentrations were prepared and
lyophilized to dry powder form. The resulting fractions were dissolved in 280-320 μl of aqueous bicelle
solution during a heating (42°)/cooling cycle (-20°). The bicelle solution was comprised of long-chain
lipids, 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-choline (POPC), and short chain lipids, 1,2-
dihexanoyl-sn-glycero-3-phosphocholine (DHPC), at a q-factor of 0.3 (or otherwise mentioned), 25
mM HEPES ∘NaOH, pH 7.4, 6% D2O, 0.02% w/v NaN3.

120 | P a g e

NMR-Spectroscopy
NMR experiments were carried out on a cryoprobe equipped Bruker Avance 700 spectrometer at 40°C,
unless otherwise noted. Experimental data was processed and analyzed with the nmrPipe package and
CARA, respectively. Assignments for the H
N
,N,C
α
,C
β
, and C' nuclei were obtained via HNCA,
HNCACB, HNCO, and HN(CA)CO experiments. Quantitative J-correlation spectroscopy allowed to
determine the
3
J
CC
and
3
J
NC
couplings with a dephasing time of 50 and 100 ms, respectively, for the
aromatic and aliphatic residues. The residual dipolar couplings (RDC) of the peptide-bicelle complex
were determined in stretched, negatively charged polyacrylamide gels that were aligned relative to the
magnetic field(Bax, 2003). A
2
H splitting of 1.2 Hz was gained by polymerizing a 320 μL from a 4.5%
w/v solution of acrylamide, 2-acrylamido-2-methyl-1-propanesulfonate (AMPS), and bisacrylamide
with a monomer-to-cross-linker ratio of 49:1 (w/w) and a molar ratio of 94:6 of acrylamide to AMPS in
150 mM Tris1⁄7HCl, pH 8.0. For efficient buffer exchange, the gels were dialyzed overnight in 50 ml
of 100 mM NaH2PO4/Na2HPO4, pH 6.8, followed by H
2
O dialysis overnight. The gel was dried at
room temperature and soaked in 320 μL of α
IIb
-bicelle for 48 h and transferred into an open-ended
NMR tube (Chou et al, 2004).
1
J
NH
,
1
J
CC
, 1J
CN
and
1
J
NH

1
D
NH
,
1
J
CC

1
D
CC
, and
1
J
CN

1
D
CN
couplings were
determined from
1
J
NH
scaled HNCO experiments and from quantitative J-correlation HNCO
experiments on isotropic and aligned samples, respectively (Chou et al, 2000). [
1
H]-
15
N NOE
measurements were carried out using 5s of presaturation preceded by a recycling delay of 4 s for the
NOE experiment and a 9-s recycle delay for the reference experiment. To measure H
N
-H
N
and H
N
-H
O

NOEs, a 15N-edited NOESY spectrum (150-ms mixing time) was recorded.

Linker Study
The proteins were expressed via the commercially available pET44 vector system (Novagen, Inc.). The
α
IIb
gene encompasses the residues Ala958-Arg997, which includes the linker and the transmembrane
domain, additionally the extracellular domain Calf2 was replaced with the smaller but structurally

121 | P a g e

homolog protein GB3A resulting the in pET44-GB3A-α
IIb
TMD construct. The single-point mutation
P965E was introduced in the construct via the quickchange protocol, resulting in the pET44-GB3A-
α
IIb
TMD(P965E) construct. The isolation protocol followed as previously described.

Titration
To compare the equilibrium constants, K
XY
, of the heterodimer complex α
IIb
/β
3
wt to its
α
IIb
(P965A/E)/β
3
mutants, 0.1 mM [
15
N/
2
H] β
3
wt was titrated with α
IIb
wt, P965A or P965E in 400 mM
DHPC, 120 mM POPC, 6% D2O, 0.02% w/v NaN
3
, and 25 mM HEPES at pH7.4. The concentration
of the αIIb peptide was varied between 0 to 0.8 mM while holding the other conditions constant (Figure
3-6 B). The monomer/dimer relationship is based on the TROSY peak volumes corresponding to the
monomeric β
3
(Gly702) residue. The volume of the monomer peak was used to quantify the dimer
formation due its increased signal to noise ratio compared to the dimer peaks. The dimer fraction was
isolated by subtracting the monomer fraction, the dividend of the observed and initial peak volume,
from the total volume observed (Bocharov et al, 2012). The equilibrium constant, K
XY
, was calculated
by fitting rate equation to the observed dimer fractions for each titration point.

ITC measurements
Employing a Microcal VP-ITC calorimeter, 10 M of 
3
peptide in the 1.425 ml sample cell was
titrated with 
IIb
peptide. We injected 
IIb
in 9 l aliquots over a period of 10 s. Unless otherwise
specified, measurements were carried out at 28 °C in 25 mM NaH
2
PO
4
/Na
2
HPO
4
, pH 7.4, 43 mM
DHPC and 17 mM of long-chain lipid. Prior to data analysis, the measurements were corrected for the
heat of dilutions of the 
IIb
and 
3
peptides. The reaction enthalpy ( H
o
) and equilibrium constant
(K
XY
), but not the 
IIb

3
complex stoichiometry, were calculated from the measured heat changes, δH
i
,
as described previously (Tellinghuisen, 2008).

122 | P a g e

Structure Calculation
The molecular structures of integrin α
IIb
(P965A), interleukin β
c
and β
c
(P441A) were calculated via
XPLOR-NIH applying simulated annealing by starting at 3000K (Schwieters et al, 2003). Secondary
carbon shifts, RDCs, and relaxation experiments determined that the N-terminus and C-terminus occurs
in random-coil conformation (Bax, 2003). The folded region for α
IIb
and β
c
, were calculated by
applying the standard force field terms for covalent geometry (bonds, angles and improper torsions)
and non-bounded contacts (van der Waals repulsion). In addition, a harmonic potential was used to
quantify the difference between the predicted and experimental residual dipolar couplings; to assist the
convergence of the residual dipolar coupling restraints, dihedral angle restrains were implemented via a
quadratic square-well potential. TALOS+ was implemented to statistically determine the backbone
dihedral angle restrains by comparing N, C
α
, C
β
, and C' chemical shifts to a structural database (Shen &
Bax, 2013; Shen et al, 2009). The χ1 angle restrains for the aliphatic and aromatic side-chains were
extracted from the J
CC
and J
NC
coupling constants. The helical conformation was refined utilizing a
backbone-backbone hydrogen-bonding potential implemented with NOE interproton distance restraints
using a quadratic square-well potential, which were referenced to the linear-helical structure. The α
IIb

rotamers were guided to an asserted conformation by a torsion angle potential of mean force to
implement an α-helical formation. The final outcome of the calculations were 20.

MD-Simulation
MD-simulation were employed to characterize protein dynamics. In the model system, Charmm27 was
implemented to prepare a peptide structure with embedded in a POPC bilayer solvated by TIP3P water
and a periodic boundary conditions were applied (Ulmschneider et al, 2007).The system was
neutralized by adding Na
+
and Cl
-
ions at a 0.2 mM salt concentration. For the parameter for the MD

123 | P a g e

simulation, performed on NAMD ver2.9, a time step of 1fs switching distance for smoothing functions
of non-bonded force were employed (Humphrey et al, 1996). The SHAKE algorithm was used to
define bonds involving an H-atom in proteins to rigid, whereby the SETTLE algorithm was used in
waters (Hristova & White, 2005). During the simulation the Particle Mesh Ewald method was used to
calculate full electrostatic interactions every 4 fs. The overall atomic coordinates of the systems were
saved every 5 ps. The system was energy-minimized for three consecutive 10,000 conjugate-gradient
steps: (1) with all protein and lipid atoms fixed, (2) with only the protein atom fixed, and (3) with all
atoms free. In the final equilibration run the pressure was maintained at 1 atm by the Langevin piston
method and the temperature was kept constant at 310 K by Langevin dynamics with a damping
coefficient of 5 ps−1. The system was continuously energy-minimized 100ns.

Analyzes of MD Trajectories
VMD was employed to analyze simulations, render molecular graphics, and generate trajectory videos.
The RMSD over the simulation time was calculated to identify trajectories which present the energy–
minimized state, which were used for further analysis (Humphrey et al, 1996). The autocorrelation
function (C[t]) was calculated for each individual residue’s amide utilizing the Gromacs package
utilizing the 2nd order Legendre polynomial for the HN bond vector [37], which in turn allowed for the
calculation of order parameter (S
2
). Furthermore, the chemical shifts (H
N
,
H
N, C
α
, C
β
, and C`) were
back calculated via Sparta+, a database system for empirical prediction of backbone chemical shifts to
ensure structural stability (Shen & Bax, 2010). The χ1 and χ2 angles were determined in order to
quantify the Trp967/968 orientation and mobility and plotted against simulation time. The tilt angle
between the TM region and the membrane normal was calculated by observing the crossing angle
between the protein vector, defined by the C
α
heavy atoms in the peptide backbone, and a membrane
vector, defined by the phosphates in inner-leaflet of the membrane.

124 | P a g e

Assay of full-length integrin α
IIb
β
3
activity in lipid nanodiscs
The receptor was purified from outdated human platelets based on a protocol modified from Ye et al.
In brief, out-dated platelets were centrifuged at 300g to remove erythrocytes and leucocytes, followed
by centrifugation at 1800g to pellet platelets. The platelets were washed twice with Tris-buffered saline
and membrane proteins extracted by incubating overnight in 20 mM Tris, pH 7.4, 150 mM NaCl, 1%
Triton X-100, 5 mM PMSF, 0.5 mM CaCl
2
, 10 M Leupeptin, 10 M protease inhibitor E64 (Sigma),
2.76 M Calpeptin buffer. Integrin IIb3 was purified on a Con A column and passed through a
Heparin column to remove thrombospondin 1. Subsequently, the receptor was purified by gel filtration
chromatography and the inactive fraction was isolated as the flow-through of an immobilized
KYGRGDS affinity matrix, as described by Steiner and co-workers. Purified integrins were stored in
20 mM Tris, pH 7.4, 150 mM NaCl, 0.1% Triton X-100, 1 mM MgCl
2
and 1 mM CaCl
2
buffer at -80
°C.
Integrin nanodiscs were assembled based on a protocol adapted from a previous report [46,47].
POPC, POPS and POPG were solubilized in chloroform, mixed thoroughly, and dried onto a glass tube
under a stream of nitrogen gas. The homogeneous lipid mixture was then solubilized in 100 mM
cholate, 10 mM Tris, pH 7.4, and 100 mM NaCl, to result in a lipid concentration of 50 mM. 72 l of
the lipid solution was then mixed with 200 μl of 200 μM membrane scaffold protein (MSP) in dH
2
O
and 200 μl of 5 μM purified inactive integrin (see above). This resulted in a final lipids:MSP:protein
ratio of 90:1:0.025 in a total volume of 472 μl. The integrin nanodiscs were assembled by removing the
detergents with SM-2 biobeads overnight at room temperature. Finally, the assembled integrin
nanodiscs were purified by gel filtration using a Superdex 200 column in 20 mM Tris, pH 7.4, 150 mM
NaCl, 0.5 mM CaCl
2
solution. Successful nanodisc assembly was verified by SDS-PAGE and electron
microscopy.

125 | P a g e

To assay the activation state of integrin 
IIb

3
obtained in nanodiscs as a function of lipid
composition, the binding of activity-state dependent antibody PAC1 was quantified by ELISA as
described previously(Kim et al, 2009a; Kim et al, 2003). Briefly, ELISA plates were coated with 5
g/ml AP3 antibody overnight at 4 °C and then blocked with BSA for 1 hr at 37 °C. After washing the
plate, integrin nanodiscs were added and incubated for 2 hr at room temperature. The PAC1 antibody
was used to detect active integrin receptors. PAC1 binding in the presence of activating antibody anti-
LIBS6 was used as control for full activation. As negative control, PAC1 binding in the presence of 20
M eptifibatide was evaluated. After 2 hr incubation with PAC1 antibody, the wells were washed again
and HRP conjugated anti-mouse IgM was added for 1 more hour of incubation. Subsequent to the final
wash, luminescence of the added ECL reagent was read using a VICTOR2 plate reader. An activation
index was calculated as (L-L
0
)/(L
max
-L
0
) where L denotes the luminescence intensity, L
0
the
luminescence in the presence of 20 M eptifibatide, and L
max
the luminescence in the presence of anti-
LIBS6 antibody.

126 | P a g e

Future Directions

Arginine snorkeling
The evolution of membrane proteins in a bilayer, with anionic lipids concentrated in the inner
leaflet, has resulted in the regulation of electrostatic contacts between anionic lipids and cationic
protein residues to establish membrane protein topology (Bogdanov et al, 2009; van Klompenburg et al,
1997; von Heijne, 1989). Elaborating on the mechanism, we examined the influence of anionic lipids
on the stability of the integrin 
IIb

3
TM complex. The heterodimer exhibits a highly conserved

IIb
(R995)- 
3
(D723) electrostatic interaction within the cytosolic lipid headgroup region. Point
mutation of either residue prompts TM complex dissociation and integrin activation. However, it is
unknown whether anionic lipids limit or modulate the strength of this interaction.
Curiously, in apoptotic cells, a breakdown in lipid membrane asymmetry exposes anionic
phosphatidylserine (PS)-based lipids to the outer membrane leaflet, which triggers phagocytosis by
macrophages. However, in some cells, most notably blood platelets, this strategy has been adapted for
physiological function. The activation of the blood clotting enzyme thrombin requires the emergence of
PS lipids in the outer membrane leaflet of platelets is required. Its failure results in the bleeding
disorder known as Scott syndrome, in which the phospholipid scramblase TMEM16F is compromised.
With 80,000 copies per platelet, the activation of the integrin 
IIb

3
adhesion receptor represents the
pivotal step in platelet aggregation and, thus, physiological and pathological thrombosis. In principle,
PS scrambling will reduce competition to 
IIb
(R995)- 
3
(D723), stabilize the TM complex and stimulate
the inactive receptor state. Anionic lipids, next to their static function in setting membrane protein
topography, may therefore dynamically modulate integrin IIb 3 stability and activity.
Our research employs biophysical, biochemical and computational techniques to quantify the

127 | P a g e

influence of anionic versus zwitterionic lipids on the stability of the integrin 
IIb

3
TM complex.
Notwithstanding the competitive nature of anionic lipids reflected in preferential interactions with basic

IIb
and 
3
residues including 
IIb
(R995), anionic lipids stabilize the 
IIb

3
TM complex relative to
electroneutral lipids. Anionic lipids thus deliver a net stabilizing effect that appears of broad
significance to the structure and folding of membrane proteins.We have now shown that, despite of the
weakened key 
IIb
(Arg995)- 
3
(Asp723) electrostatic interaction of the integrin 
IIb

3
TM complex, the
overall complex stability increased in the presence of anionic lipids. Relative to zwitterionic lipids,
anionic lipids stabilize the 
IIb

3
complex by 0.50 ± 0.01 kcal/mol. The “scrambling” of anionic lipids
in activated platelets will therefore weaken the 
IIb

3
TM complex. The electrostatic competition of
anionic lipids for 
IIb
(Arg995) is sensitive to headgroup structure and is compensated by additional
protein-lipid interactions. The cross-linking of TM helices by anionic lipids could contribute to TM
complex stabilization. However, thermodynamic data more acutely highlight possible contributions
from the narrowing of TM helix topography distributions and lipid type-dependent changes in protein-
lipid interactions between associated and dissociated TM helices. Anionic lipids can stabilize a wide
range of membrane proteins and therefore, as extrapolated from the integrin 
IIb

3
TM complex, the
structural basis of this observation is likely a general trend for TM helix-helix associations (Baenziger
et al, 2000; Rourke et al, 1996). Thus, the presence of anionic lipids is of universal importance to the
folding and stability of membrane proteins, while causing limited restrictions on electrostatic helix-
helix interactions in the headgroup region.

128 | P a g e

Figure F-1. Arginine interacts with lipid headgroup. Lipid dependence of arginine side chain-lipid
contacts. The contacts of arginine -H
N
with PO
4
-
or COO
-
of POPC and POPC, and water were
quantified. For each lipid type, three 30 ns MD simulations starting from different initial starting
conditions were performed for a total of 12 simulations

129 | P a g e

References

Adair BD, Xiong JP, Maddock C, Goodman SL, Arnaout MA, Yeager M (2005) Three-dimensional
EM structure of the ectodomain of integrin {alpha}V{beta}3 in a complex with fibronectin. J Cell Biol
168: 1109-1118

Ambroso MR, Hegde BG, Langen R (2014) Endophilin A1 induces different membrane shapes using a
conformational switch that is regulated by phosphorylation. Proc Natl Acad Sci U S A 111: 6982-6987

Anthis NJ, Campbell ID (2011) The tail of integrin activation. Trends Biochem Sci 36: 191-198

Anthis NJ, Haling JR, Oxley CL, Memo M, Wegener KL, Lim CJ, Ginsberg MH, Campbell ID (2009a)
Beta integrin tyrosine phosphorylation is a conserved mechanism for regulating talin-induced integrin
activation. J Biol Chem 284: 36700-36710

Anthis NJ, Wegener KL, Critchley DR, Campbell ID (2010) Structural diversity in integrin/talin
interactions. Structure 18: 1654-1666

Anthis NJ, Wegener KL, Ye F, Kim C, Goult BT, Lowe ED, Vakonakis I, Bate N, Critchley DR,
Ginsberg MH, Campbell ID (2009b) The structure of an integrin/talin complex reveals the basis of
inside-out signal transduction. EMBO J 28: 3623-3632

Arnaout MA, Mahalingam B, Xiong JP (2005) Integrin structure, allostery, and bidirectional signaling.
Annu Rev Cell Dev Biol 21: 381-410

Baenziger JE, Morris ML, Darsaut TE, Ryan SE (2000) Effect of membrane lipid composition on the
conformational equilibria of the nicotinic acetylcholine receptor. J Biol Chem 275: 777-784

Bax A (2003) Weak alignment offers new NMR opportunities to study protein structure and dynamics.
Protein Sci 12: 1-16

Bennett JS (2005) Structure and function of the platelet integrin alphaIIbbeta3. J Clin Invest 115: 3363-
3369

Bocharov EV, Mineev KS, Goncharuk MV, Arseniev AS (2012) Structural and thermodynamic insight
into the process of "weak" dimerization of the ErbB4 transmembrane domain by solution NMR.
Biochim Biophys Acta 1818: 2158-2170

130 | P a g e

Bogdanov M, Xie J, Dowhan W (2009) Lipid-protein interactions drive membrane protein topogenesis
in accordance with the positive inside rule. J Biol Chem 284: 9637-9641

Bowie JU (1997) Helix packing in membrane proteins. J Mol Biol 272: 780-789

Bowie JU (2011) Membrane protein folding: how important are hydrogen bonds? Curr Opin Struct
Biol 21: 42-49

Bugg CW, Isas JM, Fischer T, Patterson PH, Langen R (2012) Structural features and domain
organization of huntingtin fibrils. J Biol Chem 287: 31739-31746

C* K, Schmidt T*, EG C, F Y, TS U, MH G (2011) Basic amino-acid side chains regulate
transmembrane integrin signalling. Nature

Cao Z, Bowie JU (2012) Shifting hydrogen bonds may produce flexible transmembrane helices. Proc
Natl Acad Sci U S A 109: 8121-8126

Chen Z, Xu Y (2006) Energetics and stability of transmembrane helix packing: a replica-exchange
simulation with a knowledge-based membrane potential. Proteins 62: 539-552

Chou JJ, Baber JL, Bax A (2004) Characterization of phospholipid mixed micelles by translational
diffusion. J Biomol NMR 29: 299-308

Chou JJ, Delaglio F, Bax A (2000) Measurement of one-bond 15N-13C' dipolar couplings in medium
sized proteins. J Biomol NMR 18: 101-105

Conforti G, Zanetti A, Pasquali-Ronchetti I, Quaglino D, Neyroz P, Dejana E (1990) Modulation of
vitronectin receptor binding by membrane lipid composition. J Biol Chem 265: 4011-4019

Cross TA, Opella SJ (1980) Structural properties of fd coat protein in sodium dodecyl sulfate micelles.
Biochem Biophys Res Commun 92: 478-484

Delaglio F, Grzesiek S, Vuister GW, Zhu G, Pfeifer J, Bax A (1995) NMRPipe: a multidimensional
spectral processing system based on UNIX pipes. J Biomol NMR 6: 277-293

Deol SS, Domene C, Bond PJ, Sansom MS (2006) Anionic phospholipid interactions with the
potassium channel KcsA: simulation studies. Biophys J 90: 822-830

Elmore DE (2006) Molecular dynamics simulation of a phosphatidylglycerol membrane. FEBS Lett
580: 144-148

131 | P a g e

Faham S, Yang D, Bare E, Yohannan S, Whitelegge JP, Bowie JU (2004) Side-chain contributions to
membrane protein structure and stability. J Mol Biol 335: 297-305

Fleming KG (2002) Standardizing the free energy change of transmembrane helix-helix interactions. J
Mol Biol 323: 563-571

Fleming KG, Engelman DM (2001) Specificity in transmembrane helix-helix interactions can define a
hierarchy of stability for sequence variants. Proc Natl Acad Sci U S A 98: 14340-14344

G L, A S (1982) Model-free approach to the interpretation of nuclear magnetic-resonance relaxation
in macromolecules. 1. Theory and range of validity. Journal of the American Chemical Society 104
(17), pp 4546–4559

Goult BT, Bate N, Anthis NJ, Wegener KL, Gingras AR, Patel B, Barsukov IL, Campbell ID, Roberts
GC, Critchley DR (2009) The structure of an interdomain complex that regulates talin activity. J Biol
Chem 284: 15097-15106

Grant CV, Yang Y, Glibowicka M, Wu CH, Park SH, Deber CM, Opella SJ (2009) A Modified
Alderman-Grant Coil makes possible an efficient cross-coil probe for high field solid-state NMR of
lossy biological samples. J Magn Reson 201: 87-92

Gumbart J, Wang Y, Aksimentiev A, Tajkhorshid E, Schulten K (2005) Molecular dynamics
simulations of proteins in lipid bilayers. Curr Opin Struct Biol 15: 423-431

Harburger DS, Calderwood DA (2009) Integrin signalling at a glance. J Cell Sci 122: 159-163

Hodivala-Dilke KM, McHugh KP, Tsakiris DA, Rayburn H, Crowley D, Ullman-Culleré M, Ross FP,
Coller BS, Teitelbaum S, Hynes RO (1999) Beta3-integrin-deficient mice are a model for Glanzmann
thrombasthenia showing placental defects and reduced survival. J Clin Invest 103: 229-238

Hong H, Blois TM, Cao Z, Bowie JU (2010) Method to measure strong protein-protein interactions in
lipid bilayers using a steric trap. Proc Natl Acad Sci U S A 107: 19802-19807

Hristova K, White SH (2005) An experiment-based algorithm for predicting the partitioning of
unfolded peptides into phosphatidylcholine bilayer interfaces. Biochemistry 44: 12614-12619

Hughes PE, Diaz-Gonzalez F, Leong L, Wu C, McDonald JA, Shattil SJ, Ginsberg MH (1996)
Breaking the integrin hinge. A defined structural constraint regulates integrin signaling. J Biol Chem
271: 6571-6574

Humphrey W, Dalke A, Schulten K (1996) VMD: visual molecular dynamics. J Mol Graph 14: 33-38,
27-38

132 | P a g e

Hynes RO, Bader BL, Hodivala-Dilke K (1999) Integrins in vascular development. Braz J Med Biol
Res 32: 501-510

Hynes RO, Hodivala-Dilke KM (1999) Insights and questions arising from studies of a mouse model of
Glanzmann thrombasthenia. Thromb Haemost 82: 481-485

Hynes RO, Lively JC, McCarty JH, Taverna D, Francis SE, Hodivala-Dilke K, Xiao Q (2002) The
diverse roles of integrins and their ligands in angiogenesis. Cold Spring Harb Symp Quant Biol 67:
143-153

Jao CC, Der-Sarkissian A, Chen J, Langen R (2004) Structure of membrane-bound alpha-synuclein
studied by site-directed spin labeling. Proc Natl Acad Sci U S A 101: 8331-8336

Jao CC, Hegde BG, Chen J, Haworth IS, Langen R (2008) Structure of membrane-bound alpha-
synuclein from site-directed spin labeling and computational refinement. Proc Natl Acad Sci U S A 105:
19666-19671

Jayasinghe S, Hristova K, White SH (2001) Energetics, stability, and prediction of transmembrane
helices. J Mol Biol 312: 927-934

Killian JA, von Heijne G (2000) How proteins adapt to a membrane-water interface. Trends Biochem
Sci 25: 429-434

Kim C, Lau TL, Ulmer TS, Ginsberg MH (2009a) Interactions of platelet integrin alphaIIb and beta3
transmembrane domains in mammalian cell membranes and their role in integrin activation. Blood 113:
4747-4753

Kim C, Lau TL, Ulmer TS, Ginsberg MH (2009b) Interactions of platelet integrin alphaIIb and beta3
transmembrane domains in mammalian cell membranes and their role in integrin activation. Blood 113:
4747-4753

Kim C, Schmidt T, Cho EG, Ye F, Ulmer TS, Ginsberg MH (2011) Basic amino-acid side chains
regulate transmembrane integrin signalling. Nature

Kim M, Carman CV, Springer TA (2003) Bidirectional transmembrane signaling by cytoplasmic
domain separation in integrins. Science 301: 1720-1725

Klepsch MM, Kovermann M, Löw C, Balbach J, Permentier HP, Fusetti F, de Gier JW, Slotboom DJ,
Berntsson RP (2011) Escherichia coli peptide binding protein OppA has a preference for positively
charged peptides. J Mol Biol 414: 75-85

133 | P a g e

Lau TL, Dua V, Ulmer TS (2008a) Structure of the integrin alphaIIb transmembrane segment. J Biol
Chem 283: 16162-16168

Lau TL, Kim C, Ginsberg MH, Ulmer TS (2009) The structure of the integrin alphaIIbbeta3
transmembrane complex explains integrin transmembrane signalling. EMBO J 28: 1351-1361

Lau TL, Partridge AW, Ginsberg MH, Ulmer TS (2008b) Structure of the integrin beta3
transmembrane segment in phospholipid bicelles and detergent micelles. Biochemistry 47: 4008-4016

Li E, You M, Hristova K (2005a) Sodium dodecyl sulfate-polyacrylamide gel electrophoresis and
forster resonance energy transfer suggest weak interactions between fibroblast growth factor receptor 3
(FGFR3) transmembrane domains in the absence of extracellular domains and ligands. Biochemistry 44:
352-360

Li E, You M, Hristova K (2005b) Sodium dodecyl sulfate-polyacrylamide gel electrophoresis and
forster resonance energy transfer suggest weak interactions between fibroblast growth factor receptor 3
(FGFR3) transmembrane domains in the absence of extracellular domains and ligands. Biochemistry 44:
352-360

Li E, You M, Hristova K (2006) FGFR3 dimer stabilization due to a single amino acid pathogenic
mutation. J Mol Biol 356: 600-612

Lipari G, Szabo A (1982) Model-free approach to the interpretation of nuclear magnetic resonance
relaxation in macromolecules. Journal of American Chemical Society

Maltsev AS, Ying J, Bax A (2012) Deuterium isotope shifts for backbone ¹H, ¹ ⁵N and ¹³C nuclei in
intrinsically disordered protein α-synuclein. J Biomol NMR 54: 181-191

Merzlyakov M, Chen L, Hristova K (2007) Studies of receptor tyrosine kinase transmembrane domain
interactions: the EmEx-FRET method. J Membr Biol 215: 93-103

Merzlyakov M, You M, Li E, Hristova K (2006a) Transmembrane helix heterodimerization in lipid
bilayers: probing the energetics behind autosomal dominant growth disorders. J Mol Biol 358: 1-7

Merzlyakov M, You M, Li E, Hristova K (2006b) Transmembrane helix heterodimerization in lipid
bilayers: probing the energetics behind autosomal dominant growth disorders. J Mol Biol 358: 1-7

Nevzorov AA, Opella SJ (2007) Selective averaging for high-resolution solid-state NMR spectroscopy
of aligned samples. J Magn Reson 185: 59-70

Nietlispach D, Gautier A (2011) Solution NMR studies of polytopic α-helical membrane proteins. Curr
Opin Struct Biol 21: 497-508

134 | P a g e

Orzáez M, Salgado J, Giménez-Giner A, Pérez-Payá E, Mingarro I (2004a) Influence of proline
residues in transmembrane helix packing. J Mol Biol 335: 631-640

Orzáez M, Salgado J, Giménez-Giner A, Pérez-Payá E, Mingarro I (2004b) Influence of proline
residues in transmembrane helix packing. J Mol Biol 335: 631-640

Park SH, De Angelis AA, Nevzorov AA, Wu CH, Opella SJ (2006a) Three-dimensional structure of
the transmembrane domain of Vpu from HIV-1 in aligned phospholipid bicelles. Biophys J 91: 3032-
3042

Park SH, Mrse AA, Nevzorov AA, De Angelis AA, Opella SJ (2006b) Rotational diffusion of
membrane proteins in aligned phospholipid bilayers by solid-state NMR spectroscopy. J Magn Reson
178: 162-165

Park SH, Opella SJ (2005) Tilt angle of a trans-membrane helix is determined by hydrophobic
mismatch. J Mol Biol 350: 310-318

Partridge AW, Liu S, Kim S, Bowie JU, Ginsberg MH (2005) Transmembrane domain helix packing
stabilizes integrin alphaIIbbeta3 in the low affinity state. J Biol Chem 280: 7294-7300

Petrache HI, Grossfield A, MacKenzie KR, Engelman DM, Woolf TB (2000) Modulation of
glycophorin A transmembrane helix interactions by lipid bilayers: molecular dynamics calculations. J
Mol Biol 302: 727-746

Posokhov YO, Merzlyakov M, Hristova K, Ladokhin AS (2008) A simple "proximity" correction for
Förster resonance energy transfer efficiency determination in membranes using lifetime measurements.
Anal Biochem 380: 134-136

Rankenberg JM, Vostrikov VV, DuVall CD, Greathouse DV, Koeppe RE, Grant CV, Opella SJ (2012)
Proline kink angle distributions for GWALP23 in lipid bilayers of different thicknesses. Biochemistry
51: 3554-3564

Ridder A, Skupjen P, Unterreitmeier S, Langosch D (2005) Tryptophan supports interaction of
transmembrane helices. J Mol Biol 354: 894-902

Rocco M, Rosano C, Weisel JW, Horita DA, Hantgan RR (2008) Integrin conformational regulation:
uncoupling extension/tail separation from changes in the head region by a multiresolution approach.
Structure 16: 954-964

Rourke AM, Cha Y, Collins D (1996) Stabilization of granulocyte colony-stimulating factor and
structurally analogous growth factors by anionic phospholipids. Biochemistry 35: 11913-11917

135 | P a g e

Sansom MS, Weinstein H (2000) Hinges, swivels and switches: the role of prolines in signalling via
transmembrane alpha-helices. Trends Pharmacol Sci 21: 445-451

Schwieters CD, Kuszewski JJ, Tjandra N, Clore GM (2003) The Xplor-NIH NMR molecular structure
determination package. J Magn Reson 160: 65-73

Senes A, Engel DE, DeGrado WF (2004) Folding of helical membrane proteins: the role of polar,
GxxxG-like and proline motifs. Curr Opin Struct Biol 14: 465-479

Shahidullah K, London E (2008) Effect of lipid composition on the topography of membrane-
associated hydrophobic helices: stabilization of transmembrane topography by anionic lipids. J Mol
Biol 379: 704-718

Shen Y, Bax A (2010) SPARTA+: a modest improvement in empirical NMR chemical shift prediction
by means of an artificial neural network. J Biomol NMR 48: 13-22

Shen Y, Bax A (2013) Protein backbone and sidechain torsion angles predicted from NMR chemical
shifts using artificial neural networks. J Biomol NMR 56: 227-241

Shen Y, Delaglio F, Cornilescu G, Bax A (2009) TALOS+: a hybrid method for predicting protein
backbone torsion angles from NMR chemical shifts. J Biomol NMR 44: 213-223

Solomaha E, Palfrey HC (2005) Conformational changes in dynamin on GTP binding and
oligomerization reported by intrinsic and extrinsic fluorescence. Biochem J 391: 601-611

Son WS, Park SH, Nothnagel HJ, Lu GJ, Wang Y, Zhang H, Cook GA, Howell SC, Opella SJ (2012)
'q-Titration' of long-chain and short-chain lipids differentiates between structured and mobile residues
of membrane proteins studied in bicelles by solution NMR spectroscopy. J Magn Reson 214: 111-118

Sparks KA, Gleason NJ, Gist R, Langston R, Greathouse DV, Koeppe RE (2014) Comparisons of
interfacial Phe, Tyr, and Trp residues as determinants of orientation and dynamics for GWALP
transmembrane peptides. Biochemistry 53: 3637-3645

Springer TA, Wang JH (2004) The three-dimensional structure of integrins and their ligands, and
conformational regulation of cell adhesion. Adv Protein Chem 68: 29-63

Springer TA, Zhu J, Xiao T (2008) Structural basis for distinctive recognition of fibrinogen gammaC
peptide by the platelet integrin alphaIIbbeta3. J Cell Biol 182: 791-800

Suk JE, Situ AJ, Ulmer TS (2012) Construction of covalent membrane protein complexes and high-
throughput selection of membrane mimics. J Am Chem Soc 134: 9030-9033

136 | P a g e

Takagi J (2002) [First atomic view of alpha V beta 3 integrin extracellular domain]. Tanpakushitsu
Kakusan Koso 47: 153-159

Takagi J, Beglova N, Yalamanchili P, Blacklow SC, Springer TA (2001a) Definition of EGF-like,
closely interacting modules that bear activation epitopes in integrin beta subunits. Proc Natl Acad Sci U
S A 98: 11175-11180

Takagi J, Erickson HP, Springer TA (2001b) C-terminal opening mimics 'inside-out' activation of
integrin alpha5beta1. Nat Struct Biol 8: 412-416

Tellinghuisen J (2008) Isothermal titration calorimetry at very low c. Anal Biochem 373: 395-397

Tieleman DP, Shrivastava IH, Ulmschneider MR, Sansom MS (2001) Proline-induced hinges in
transmembrane helices: possible roles in ion channel gating. Proteins 44: 63-72

Tsakiris DA, Scudder L, Hodivala-Dilke K, Hynes RO, Coller BS (1999) Hemostasis in the mouse
(Mus musculus): a review. Thromb Haemost 81: 177-188

Turnbull WB, Daranas AH (2003) On the value of c: can low affinity systems be studied by isothermal
titration calorimetry? J Am Chem Soc 125: 14859-14866

Ulmer TS (2010) Structural basis of transmembrane domain interactions in integrin signaling. Cell Adh
Migr 4: 243-248

Ulmer TS, Yaspan B, Ginsberg MH, Campbell ID (2001) NMR analysis of structure and dynamics of
the cytosolic tails of integrin alpha IIb beta 3 in aqueous solution. Biochemistry 40: 7498-7508

Ulmschneider MB, Sansom MS (2001) Amino acid distributions in integral membrane protein
structures. Biochim Biophys Acta 1512: 1-14

Ulmschneider MB, Smith JC, Ulmschneider JP (2010) Peptide partitioning properties from direct
insertion studies. Biophys J 98: L60-62

Ulmschneider MB, Ulmschneider JP (2008) Membrane adsorption, folding, insertion and translocation
of synthetic trans-membrane peptides. Mol Membr Biol 25: 245-257

Ulmschneider MB, Ulmschneider JP, Sansom MS, Di Nola A (2007) A generalized born implicit-
membrane representation compared to experimental insertion free energies. Biophys J 92: 2338-2349

van der Wel PC, Reed ND, Greathouse DV, Koeppe RE (2007) Orientation and motion of tryptophan

137 | P a g e

interfacial anchors in membrane-spanning peptides. Biochemistry 46: 7514-7524

van Klompenburg W, Nilsson I, von Heijne G, de Kruijff B (1997) Anionic phospholipids are
determinants of membrane protein topology. EMBO J 16: 4261-4266

Varkey J, Mizuno N, Hegde BG, Cheng N, Steven AC, Langen R (2013) α-Synuclein oligomers with
broken helical conformation form lipoprotein nanoparticles. J Biol Chem 288: 17620-17630

von Heijne G (1989) Control of topology and mode of assembly of a polytopic membrane protein by
positively charged residues. Nature 341: 456-458

Vostrikov VV, Daily AE, Greathouse DV, Koeppe RE (2010) Charged or aromatic anchor residue
dependence of transmembrane peptide tilt. J Biol Chem 285: 31723-31730

Wang YX, Jacob J, Cordier F, Wingfield P, Stahl SJ, Lee-Huang S, Torchia D, Grzesiek S, Bax A
(1999) Measurement of 3hJNC' connectivities across hydrogen bonds in a 30 kDa protein. J Biomol
NMR 14: 181-184

Wegener KL, Partridge AW, Han J, Pickford AR, Liddington RC, Ginsberg MH, Campbell ID (2007)
Structural basis of integrin activation by talin. Cell 128: 171-182

Williams KA, Deber CM (1991) Proline residues in transmembrane helices: structural or dynamic role?
Biochemistry 30: 8919-8923

Wiseman T, Williston S, Brandts JF, Lin LN (1989) Rapid measurement of binding constants and heats
of binding using a new titration calorimeter. Anal Biochem 179: 131-137

Woodham AW, Da Silva DM, Skeate JG, Raff AB, Ambroso MR, Brand HE, Isas JM, Langen R, Kast
WM (2012) The S100A10 subunit of the annexin A2 heterotetramer facilitates L2-mediated human
papillomavirus infection. PLoS One 7: e43519

Xiao T, Takagi J, Coller BS, Wang JH, Springer TA (2004) Structural basis for allostery in integrins
and binding to fibrinogen-mimetic therapeutics. Nature 432: 59-67

Ye F, Liu J, Winkler H, Taylor KA (2008) Integrin alpha IIb beta 3 in a membrane environment
remains the same height after Mn2+ activation when observed by cryoelectron tomography. J Mol Biol
378: 976-986

Yohannan S, Faham S, Yang D, Grosfeld D, Chamberlain AK, Bowie JU (2004) A C alpha-H...O
hydrogen bond in a membrane protein is not stabilizing. J Am Chem Soc 126: 2284-2285

138 | P a g e

You M, Li E, Wimley WC, Hristova K (2005) Forster resonance energy transfer in liposomes:
measurements of transmembrane helix dimerization in the native bilayer environment. Anal Biochem
340: 154-164

Zhu J, Carman CV, Kim M, Shimaoka M, Springer TA, Luo BH (2007) Requirement of alpha and beta
subunit transmembrane helix separation for integrin outside-in signaling. Blood 110: 2475-2483

Zhu J, Luo BH, Xiao T, Zhang C, Nishida N, Springer TA (2008) Structure of a complete integrin
ectodomain in a physiologic resting state and activation and deactivation by applied forces. Mol Cell 32:
849-861

Abstract (if available)

Abstract With 50% of all drug targets, membrane proteins present an important subcategory for structural characterization, even though they code for less than a third of all proteins in the human genome. Due to specific properties, the structure of membrane proteins are not easily characterized by conventional methods, therefore NMR presents a unique tool for that purpose. Integrins are a type I heterodimeric receptors that are essential for cell adhesion and migration. Integrins are large heterodimeric membrane receptors that tie the extracellular matrix to the cytoskeleton (Harburger and Calderwood, 2009). It is essential to both the arrest of bleeding at sites of vascular injury and pathological thrombosis culminating in heart attack and stroke. Integrin bi‐directional transmembrane (TM) signaling involves the dissociation of a complex formed by the α–β TM segments, which is accompanied by large rearrangements of its extracellular domains. ❧ TM segment sequences of the 8 β human subunits are well conserved, especially within a distinct positive charged residue at the inner membrane clasp, β₃(K716). In our research, we have shown that charge variation of the lipid system influences K716’s amide-lipid interaction at the C-terminal membrane interface, therefore defining a regulation mechanism for αIIb β₃ heterodimer formation. Site-directed mutation of K716 caused integrin to achieve its high-affinity state by stabilizing its monomeric species. Using directed evolution of β₃(K716A), we identified a substitution, A711P, to restore the integrin α IIb β₃ default low-affinity state. Quantitative dimerization analysis, utilizing ITC, NMR and FRET, independently verified an increase in TM domain dimer formation upon proline mutation, therefore recovering the low-affinity state. Structural analysis of the monomeric β₃(A711P/K716A) identified a kink of 30 ± 1° at the border of the outer and inner membrane clasps, thereby decoupling the tilt between these segments. Subsequently, the integrin αIIb β₃(A711P) heterodimer revealed a strong glycine packing at the OMC compared to the wt, hence its complex represents the first structure of a heterodimeric TM receptor of its kind and reveals a dimerization interface of captivating complexity. ❧ Utilizing NMR and MD simulation, we determined proline's extended role at the N-terminal membrane-water interface of single-pass α-helical TM domains in integrin αIIb and interleukin βc. The structural characterization of these mutants revealed proline's function as N-terminal helix-cap and linker terminator. Therefore, αIIb (P965) establishes an equilibrium between the length of the TM helix and its N-terminal linker domain, regulating signal transmission towards the extracellular domains. These results describe signaling and activation tendencies of members in the integrin α family. Taken together, the structural knowledge presented in this thesis provides insights to the disease mechanism and will form a scientific basis for future therapeutic treatments of stroke and cancer.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Molecular basis of CD33 receptor function, and effects of a destabilizing transmembrane motif on αXβ2 integrin

PDF

Study of integrin αIIbβ3 transmembrane and cytosolic domain interaction and the expression of extracellular domain by Pichia Pastoris

PDF

Integrin β3TM structural investigations using phospholipid nanodiscs

PDF

Structural studies of the IAPP membrane-mediated aggregation pathway

PDF

The folding, misfolding and refolding of membrane proteins and the design of a phosphatidylserine-specific membrane sensor

PDF

The structure and function of membrane curving proteins on different membrane shapes and their regulation by post-translational modifications

PDF

Expression and purification of extracellular domain of human CD33 in Escherichia coli and Pichia pastoris

PDF

Role of CD33 structure and function in Alzheimer’s disease

PDF

Structure and kinetics of the Orb2 functional amyloid

PDF

Before they were amyloid: understanding the toxicity of disease-associated monomers and oligomers prior to their aggregation

PDF

Structural studies on functional amyloids and the mechanism of aggregation and disaggregation of Huntingtin Exon 1

PDF

Oligomer formation of functional amyloid protein - Orb2A

PDF

Regulating functional amyloid formation: the promiscuous behavior of the Orb2A N-terminal amphipathic region

PDF

Protein phosphatase 2A and annexin A5: modulators of cellular functions

PDF

Mechanistic studies of the disintegrin contortrostatin and characterization of the recombinant protein vicrostatin

PDF

Mechanism of secretion and function of heat shock protein-90 (Hsp90) family genes

PDF

Developing peptide and antibody-mimetic ligands for the cell surface receptors β2AR and DC-SIGN

PDF

Structural and biochemical determinants of APOBEC1 substrate recognition and enzymatic function

PDF

Structural characterization of pseudorepeat shuffled alpha-synuclein

PDF

The role of secreted Hsp90α in tissue repair and cancer

Asset Metadata

Creator Schmidt, Thomas (author)

Core Title Dual effects of transmembrane proline residues on integrin function

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Genetic, Molecular and Cellular Biology

Publication Date 06/17/2015

Defense Date 08/11/2014

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag integrin,NMR,OAI-PMH Harvest,protein,structure biology

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Ulmer, Tobias (committee chair), Chow, Robert H. (committee member), Farley, Robert A. (committee member), Haworth, Ian (committee member), Langen, Ralf (committee member)

Creator Email schmidt.nedlitz@hotmail.de

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-571451

Unique identifier UC11301761

Identifier etd-SchmidtTho-3469.pdf (filename),usctheses-c3-571451 (legacy record id)

Legacy Identifier etd-SchmidtTho-3469.pdf

Dmrecord 571451

Document Type Dissertation

Format application/pdf (imt)

Rights Schmidt, Thomas

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA