NMR study of the human NCK2 SH3 domains structure determination, binding diversity, folding and amyloidogenesis 1

1
Chapter 1 Introduction

In multicellular organisms, a multitude of different signal transduction processes are
required for coordinating the behaviour of individual cells to support the function of
the organism as a whole. The organism’s sensing of both the external and internal
environments at the cellular level relies on signal transduction. Eph receptors and
ephrins have captured the interest of the biology community in recent years. The
ephrin signalling pathway plays many important roles in cells: segmentation, axon
guidance, fascination, cell migration, angiogenesis, limb development and
tumorigenesis. Some proteins coordinate with ephrins and Eph receptors in these
pathways. One of them is Nck2. The Nck2 adaptor protein, which directly binds with
the phosphorylation site on the cytoplasmic side of ephrinB/EphB through its SH2
domain, interacts with numerous downstream functional effectors. Realising the
importance of Nck2 in coordinating indirect ephrin/Eph-effector interactions and
eventually a huge assembly to function properly, we chose this particular protein as
our target to investigate bindings between SH3 domains that are independent tandem
segments and effectors from a structural point of view. In this thesis, the SH3 binding
mechanism and its abnormal folding process of mutants were investigated through a
series of biophysical methods.

1.1 Ephrin signalling pathway and the roles of ephrins/Ephs

1.1.1 Ephrin signalling pathway
The Eph receptor and ephrins are membrane-bound proteins that function as a
receptor-ligand pair. There are 13 Eph receptors and 8 ephrins that have been

2
identified in mammals (Tuzi and Gullick, 1994; Orioli and Klein, 1997; Pasquale,
1997). Only one Eph receptor and one ephrin are found in Drosophila melanogaster

and one Eph receptor and 4 ephrins are found in C. elegants (Scully et al., 1999;
Wang et al., 1999; Bossing and Brand, 2002). Eph receptors and ephrins can be
divided into two groups based on a sequence and homology and binding preference.
Recent research has indicated that EphA receptor binds preferentially to ephrin-As
(with the exception of EphA4), and EphB receptors have a preference to bind with
ephrin-Bs. However, cross class interactions have been found in specific contexts
(Himanen et al., 2004). Ephrin-As and ephrin-Bs exhibit different structural features,
because ephrin-As are tethered to the membrane by means of a
glycosylphosphatidylinosito anchor, whereas ephrin-Bs span across a membrane with
a cytoplasmic domain. In spite of the limited interactions between ephrin-As and
ephrin-Bs, they are promiscuous within a class with different Eph receptors binding to
a given ephrin, and vice versa. The promiscuity may guarantee functional
redundancies that have been observed in vivo for both ephrin and Eph receptors
(Reldheim et at., 2000).

Eph receptors can be regarded as one member of a superfamily of receptor tyrosine
kinases. They are autophosphorylated upon binding to their cognate ephrin ligands,
and subsequently activate downstream signalling cascades. The reverse signalling
transduction can also be activated after the ephrin binds to Eph recepors, activating
the downstream effector of ephrins. Transmembrane ephrins can either be activated
through the tyrosine phosphorylation of their cytoplasmic tail or through interaction
with various signalling molecules. Although the GPI-linked ephrin stimulation is still
unclear, they have been shown to activate a member of the Src-family kinases. In

3
addition, oligomerisation and clusting of Eph receptors and ephrins are essential for
their signalling function and might be regulated by localisation in membrane
microdomains (Cowan and Henkemeyer, 2002; Kullanderand Klein, 2002). Eph-
ephrin complexes can progressively aggregate into larger clusters, the size of which
might depend on the densities of Eph receptors and ephrins on the cell surface.

Several weak ephrin-ephrin and receptor-receptor interactions could promote the
association of the complexes into an interconnected network (Elena B. Pasquale,
2005).


















4


Figure 1.1 Ephrin signalling pathway Eph receptor tyrosine kinases and their ephrin
receptors are recognised to regulate several important processes during development
including axon guidance, cell migration, angiogenesis, synaptic plasticity, etc. (Image from
protein lounge signalling pathway database)

5
1.1.2 Roles of ephrins and Eph receptors

1.1.2.1 Axon guidance
The function of Eph receptors and ephrin ligands was firstly studied in axon guidance.
It was shown that cells expressing Eph receptors avoided territories expressing
ephrins, thus providing necessary cues to guide axons to their appropriate target
(O’Leary and Wilkinson, 1999). Recently, it was also found that Eph receptors and
ephrins can also regulate axon pathfinding through attractive interactions (Knoll et al.,
2001; Kullander et al., 2001; Hindges et al., 2002; Mann et al., 2002; Eberhart et al.,
2004) and ephrins can act as receptors on navigating axons. Henkemeyer proposed
that EphB2 acts as a ligand to activate an ephrin-induced reverse signalling and direct
the migration of ephrin-expressing (Henkemeyer et al., 1996). Another Eph receptor,
EphA4 (which can bind ephrin-B2 and ephrin-B3 in addition to ephrin-As), has also
been shown to act as a ligand to control the formation of the anterior commissure tract
(Kullander et al., 2001). Ephrin-B3 is a known ligand for EphA4, however no defects
of the anterior commissure were reported in the ephrin-B3 null mice (Kullander et al.,
2001), raising the possibility that EphA4 is acting as a ligand for one of the ephrin-As.
ephrin-induced reverse signalling has also been implicated in retinal axon pathfinding

(Birgbauer et al., 2000; Birgbauer,et al., 2000; Birgbauer et al., 2000; Hornberger et
al., 1999).

1.1.2.2 Segmentation
Ephs and ephrins have been recognised early for their role in segmentation
(Wilkinson, 2000). Initial expression studies of these proteins have shown that several
members of both the Eph receptor family and the ephrin family are expressed in a

6
segmented pattern in the hindbrain and in somites, suggesting that Ephs and ephrins
could have a role in segmentation during embryogenesis (Gale et al., 1996). Cells
expressing ephrin-B2 are excluded from the Eph receptor-expressing rhombomeres,
presumably after the activation of the receptors (Xu et al., 1999). Zebrafish mutant
studies, in which the disruption of somite formation is associated with the loss of
EphA4 and uniform ephrin-B2 expression in paraxial mesoderm, have shown that
reverse signalling is required for the formation of boundaries during somite
morphogenesis (Barrios et al., 2003). Forward signalling was shown to be responsible
for the epithelialisation of the somite, in an autonomous and non-autonomous manner.
It is still unclear whether the repulsion mechanism is utilised in the formation of the
boundary in the paraxial mesoderm.

1.1.2.3 Cell migration
Eph receptors and ephrins also regulate both cranial and trunk neural crest cell (NCC)
migration (Holder and Klein, 1999; Wilkinson, 2000). Similarly, a cell’s autonomous
forward signalling has been shown to regulate branchial NCCs’ migration in mice
(Adams et al., 2001). It has been suggested that forward signalling in Eph-expressing
NCCs was necessary and sufficient for proper branchial arch development, and
ephrin-B2 expressed in the neural tube ( in r4 and r6 ) might be involved in the
delamination of Eph-expressing NCCs (Adams et al., 2001). Ephrin-B1 is also
required for the proper migration of branchial NCCs because mutant mice display a

cleft palate, consistent with a defect in NCC (Davy et al., 2004). Ephrin-B1 also acts
autonomously in the cell in NCC to regulate their targeted migration, and this reverse
signalling involving the binding of PDZ containing protein is required for this
function. Very recently, forward and reverse signalling have been implicated in the

7
migration and adhesion of cells involved in the separation of the urorectal region
(Dravis et al., 2004). Dravis proposed that the simultaneous activation of forward and
reverse signalling in the same cell leads to adhesion, while the unidirectional
activation of either forward or reverse signalling leads to repulsion.

1.1.2.4 Angiogenesis
Several studies have implicated Eph receptors and ephrins in angiogenesis (Adams,
2002). Studies show that the deletion of ephrin-B2 and ephrin-B4 result in identical
phenotypes, characterised by defective angiogenic remodeling (Wang et al., 1998;
Adams et al., 1999; Gerety et al., 1999). Based on the fact that EphB4 is expressed on
veins, while ephrin-B2 is restricted to arteries, Adams proposed that this
receptor/ligand pair might be involved in setting up the arterial and venous identity of
blood vessels, possibly by means of repulsion between ephrin-B2 and EphB4
expressing endothelial cells. The same authors subsequently provided evidence that
ephrin- induced reverse signalling was required for blood vessel remodelling ,
because the expression of a deleted form of ephrin-B2 lacking the cytoplasmic
domain, was unable to rescue the angiogenesis defects associated with the loss of
ephrin-B2 (Adams et al., 2001). The role of ephrin-induced reverse signalling in
angiogenesis was considered to be quite controversial until recently. However, the
Cowan study demonstrated that angiogenesis proceeds normally in the absence of the
ephrin-B2 cytoplasmic domain, inferring that forward signalling is sufficient for this
process (Cowan et al., 2004).

1.1.2.5 GPI-anchored ephrins

8
GPI-linked ephrins could act as receptors and activate a reverse signalling pathway to
regulate epidermal morphogenesis. One study showed that GPI-linked ephrins act as
receptors in a guidance decision affecting vomeronasal axons (Knoll et al., 2001;
Knoll and Drescher, 2002). In stripe assays, vomeronasal axons prefer to grow on Eph
receptors rather than a control protein, suggesting that ephrin-A5-expressing axons
are attracted towards EphA6 expressing territories (Knoll et al., 2001).

1.2 Nck2 adaptor protein and its SH3 domain

1.2.1 Nck2 adaptor protein
Signal transduction involves the coordinated relay of information from extracellular
cues to intracellular effectors. The formation of multimeric protein complexes is a
critical step in the activation of most intracellular signal transduction cascades. In
many cases, the proteins consisting of src homology 2 and 3 (domains) are very
important in these processes. The “adaptor” term is used to describe the features of
these proteins that lack intrinsic enzymatic functions and consist almost entirely of
SH2 and SH3 domains.

Various biochemical and genetic analyses have identified the SH2/SH3 adaptor
proteins as critical mediators in the activation of diverse signal responses. The
abnormal activation of these proteins resulted in developmental defects and the onset
of various abnormalities (Mayer BJ, 1988; Clark SG, 1992; Simon MA, 1993; Garrity
PA, 1996). There are many excellent papers reporting the detailed signal functions of
these adaptor proteins, including p58, Grb2, Crk and nck2. The human Nck cDNA
was originally cloned by (Lehmann et al., 1990). Using monoclonal antibodies that

9

Figure 1.2 Grb4 (Nck2) in the ephrin signalling pathway (Nature, 413, 13, 2001)











10
recognise the melanoma-specific MUC18 antigen, NCK was identified as a false
positive during the screening of melanoma cDNA expression. In order to identify the
SH2-containing proteins, Margolis and colleagues later cloned the murine homolog of
Nck, termed Grb4. The human and mouse Nck cDNA encode a 47 kDa protein
consisting entirely of SH3 domains and single C-terminal SH2 domains. Studies
shows that the Nck proteins are expressed in many tissues and cell types (Park D,
1992; Li W, 1992), implying its function in diverse cellular events. Nck SH2 Domain
associates with the tyrosine phosphorylated proteins including EGP, PDGF, HGF and
VEGF. Here we should mention that Nck-associated proteins are involved in GTPase
activation. Many cell surface receptors utilise members of the Ras superfamily of low
molecular weight GTP-binding proteins to regulate the activities of multicellular
signalling cascades (Vojtek AB, 1995). Williams and colleagues first reported that
Nck directly interacts with the SOS GTP exchange factor, resulting in enhanced
transcription from a ras-dependant reporter gene. This investigation shows that Nck
and SOS constitutively associate in quiescent or growth factor stimulated cells via a
second Nck SH3 domain and C-terminal proline-rich region of SOS. Following PDGF

stimulation, Nck and hyperphosphorylated SOS translocate to the activated RTK,
leading to SOS-mediate ras activation. However, Tanaka et al. report that exogenous
over-expressions of various dominant negative Nck proteins have little effect on the
EGF activation. One possible explanation of this discrepancy is that both ras-
dependent and ras-independent signalling pathways control ERK activation. In
addition to the reports of association with ras, several groups have reported a
functional link between Nck, the Paks and various Rho family members. Two
members of the Pak family Pak1 and Pak3 have been shown to be constitutively
associated in quiescent or growth factor-stimulated cells with the second Nck domain

11
(Galisteo ML, 1996; Bagrodia S, 1995; Bokoch G, 1996; Lu W, 1997). Associated
Rho-specific GTP exchange factors are most likely to be involved in this process,
although the functional relationship between these factors and Nck or the PAKs
remains to be determined. Together, all data show that Nck proteins have a functional
role in the activation of signal events involving members of the Ras superfamily in
low molecular weight GTP-binding proteins.

Recent evidence from different labs strongly supports the ability of different domains
of the Nck to regulate diverse signal transduction events. Not only does it involve the
cellular proliferation, but also the regulation of cellular migration and actin
cytoskeletal dynamics. Mutation studies in WASP show the reduced mobility of
lymphoid immune cells, reflecting the fact that WASP and its relatives play a critical
role in the organisation of the actin cytoskeleton (Mayer BJ, 2001; Machesky LM,
1999; Mullins RD, 2000).

1.2.1.1 Nck2 interaction with Bcr-Abl
Sunita Coutinho, et al. identified Grb4 as an interacting protein using yeast 2-hybrid
in which the Bcr-Abl was used as the bait. The interaction was tested both in vitro and
in vivo, and grb4 was an excellent substrate of the Bcr-Abl tyrosine kinase. The

association of Grb4 and Bcl-Abl in intact cells was mediated by an SH2-mediated
phosphotyrosine-depentent interaction as well as an SH3-mediated phosphotyrosine-
indepentent interaction. Subcellular localisation studies have shown that Grb4
localises to both the nucleus and cytoplasm. A coexpression of kinase-active Bcr-Abl
with Grb4 results in the translocation of Grb4 from the cytoplasm and nucleus to the
cytoplasm, to colocalise with Bcr-Abl. This coexpression also induced a redistribution

12
of actin-associated Bcr-Abl. Sunita Coutinho et al. finally concluded that Grb4 in
conjunction with Bcr-Abl may be capable of modulating the cytoskeletal structure and
negatively interfering with the signalling of oncogenic Abl kinases. She also
speculated that Grb4 may therefore play a role in the molecular pathogenesis of
chronic myelogenous leukemia (Sunita Coutinho et al., 2000).

1.2.1.2 Nck2 acting as a nuclear repressor of transcription
Co-immunoprecipitation experiments have revealed that the co-expression of Grb4
and v-Abl results in a complex formation between v-Abl and Grb4 independent of the
kinase activity of v-Abl. The interaction was identified as SH3-dependent but
independent of the SH2 domain of Grb4. The results showed that an overexpression
of Grb4 results in the significant inhibition of the v-Abl-induced transcriptional
activation from the promitogenic enhancer elements such as activator protein 1(AP-1)
and the serum-responsive element (SRE). A further mutational analysis revealed that
the first two SH3 domains primarily mediate the inhibitory function. Co-expression
experiments indicated that the inhibitory activity of Grb4 was specific for c-jun/c-fos-
regulated promoter elements. Finally, based on all these results, Thomas Jahn, et al.
suggested a novel role for Grb4 in the inhibition of promitogenic enhancer elements,
such as the 12-O-tetradecanoylphorbol-13-acetate-responsive element and the SRE
(Jahn T, 2001).

1.2.1.3 Nck2 interaction with Dock180 – a signalling protein implicated in

regulation of membrane ruffling and migration

Using a yeast 2-hybrid screen, Yizeng, Tu et al. identified one new signalling protein,
DOCK180, which interacts with Grb4. An SPR analysis showed that the second and
third SH3 domain interact with the C-terminal of DOCK180. In terms of binding

13
affinity, interactions mediated by the individual SH3 were much weaker that that of
the full length Grb4. The main binding region of DOCK180 was located at region
1819-1836. In addition, two of the other weak binding sites were pinpointed at the
region of 1793-1810 and 1835-1852 respectively. Finally, these results showed that a
new interaction between DOCK180 and Grb4 had been identified. More importantly,
the protein complex was mediated by multiple SH3 domains, enhancing the weak
interactions mediated by each individual SH3 domain (Tu Y, 2001).

1.2.1.4 Nck2 interaction with focal adhesion kinase (FAK)
Recently, Silvia M. Goicoechea et al. reported that the Nck-2-FAK complex was
mediated by interactions involving multiple SH2 and SH3 domains of Nck-2. The
SH2 mediated interaction with FAK was dependent on phosphorylation of Tyr397,
which was involved in the regulation of cell mobility. An overexpression of Nck2
modestly decreased cell mobility, whereas an overexpression of a mutant form of
Nck2 containing only the SH2 domain significantly promoted cell mobility
(Goicoechea SM, 2002).

1.2.1.5 Nck2 interaction with PINCH- LIM only protein known to mediate
integrin signal

Recently, Tu et al. have identified the Nck2 adaptor proteins as a potential
convergence point between integrin signalling pathways and growth factor signalling
(Tu Y, 1998). PINCH, which is widely expressed and evolutionally conserved,

interacted with the Nck2. The interaction was mediated by the fourth LIM domain of
PINCH and the third SH3 domain of Nck-2 (Tu Y, 1998). Also, Algirdas Velyvis et al.
showed that the PINCH LIM4 domain, while maintaining the conserved LIM scaffold,
recognised the third SH3 domain of the Nck2 by using NMR spectroscopy. A quite

14
weak binding affinity was obtained by using the NMR titration method, and got a 3.1
mM K
d
value (Algirdas Velyvis, 2003).

1.2.1.6 Nck2 interaction with PDGFR beta regulates actin polymerisation
Min Chen, et al. recently reported a specific role for the Nck2 in platelet-derived
growth factor-induced actin polymerisation in NIH 3T3 cells. Data showed that an
overexpression of Nck2 blocked PDGF-stimulated membrane ruffling and formation
of lamellipoda. The mutation of the SH2 domain or the middle SH3 domain of Nck2
abolished its interfering effects. It was shown that Nck2 bound at Tyr-1009, which
was different from the Nck1 binding site and did not compete with
phosphatidylinositol-3 kinase for binding to PDGFR. Constitutively membrane-bound
Nck2 blocked Rac1-L62-induced membrane ruffling and the formation of
lamellipodia, suggesting that Nck2 acted in parallel to or down-stream of Rac1. It was
the first report of cross talk between receptor tyrosine kinase signalling and the actin
cytoskeleton (Chen M, 2000).

1.2.1.7 Nck2 interaction with Tyrosine-phosphorylated Disabled 1 and
redistribution in Reelin-stimulated neurons

Recently, Nck2 was identified as a phosphorylated-dependent Dab1-interacting
protein. The tyrosine phosphorylation sites of the Disabled1 (Dab1) docking protein
were essential for the transmission of the Reelin signal, which regulated neuronal

placement. The binding sites of Dab1 with Nck2 were identified as Y220 or Y232.
Nck2 was coexpressed with Dab1 in the developing brain and in cultured neurons,
where Reelin stimulation led to the redistribution of Nck2 from the cell soma into
neuronal processes. Albena Pramatarova et al. found that tyrosine-phosphorylated
Dab1 in synergy with Nck2 disrupted the actin cytoskeleton. A morphological

15
experiment indicated that conserved interaction existed between the Disabled and Nck
family members. The author also proposed a model in which Dab1 phosphorylation
led to the recruitment of Nck2 to the membrane, where it acted in remodelling the
actin cytoskeleton (Pramatarova A, 2003).

1.2.1.8 Nck2 interaction with TrkB tyrosine kinase receptor
Shingo Suzuki, et al. recently identified Nck2 as a binding partner with the TrkB
tyrosine kinase receptor by using a yeast 2-hybrid screen. The brain-derived
neurotrophic factor (BDNF) bound to and activated the TrkB tyrosine receptor kinase
to regulate cell differentiation, survival and neural plasticity in the nervous system.
Three tyrosines Y694, Y695 and Y771 were believed to be crucial residues for this
interaction. The data indicated that BDNF stimulation promoted the interaction of
Nck2 with TrkB in cortical neurons (Suzuki S, 2002).

1.2.1.9 Nck2 interacts with WASP-Binding Protein WIRE
Recently, Pontus Aspenstrom found that Nck2 interacted with WASP-Binding Protein
WIRE which had a role in the regulation of the action filament system. WIRE was
localised to action filaments in transiently transfected PAE/PDGFRβcells, and in cells
simultaneously expressing WIRE and WASP, WIRE relocalised WASP to actin
filaments, a relocalisation that required direct interaction between the two proteins. A
PDGF treatment of cells ectopically expressing WIRE resulted in the formation of
peripheral protrusions composed of filopodia and a lamelliodia-like structure. All data
showed that WIRE had a role in the WASP-mediated organisation of the actin

cytoskeleton, and that WIRE was a potential link between the activated PDGF
receptor and the actin polymerisation machinery (Aspenstrom P, 2002).

16

1.2.1.10 Nck2 interaction with wrch1, belonging to Cdc42 subfamily
Jan Saras recently identified that Wrch1 is one of the binding partners of Nck2. The
interaction was shown to be mediated via PxxP motifs in the N-terminal of Wrch1 to
the second and third SH3 domains of Nck2. Wrch1, which belongs to the Cdc42
subfamily, was one of the least characterised family members. The Wrch1 had no
detectable GTPase activity in vitro and its intrinsic nucleotide exchange rate was very
high in comparison with Cdc42. Being tranfected with Wrch1, NIH3T3 cells showed
an up-rounded, retracted phenotype. The serum stimulation of cells expressing Wrch1
induced vigorous membrane blebbing, a phenomenon dependent on the activity of
ROCK (Jan Saras, 2004).

1.2.2 SH3 domain

1.2.2.1 The history of SH3 domain
SH3 domains were first noted by researchers as regions of sequence similarity
between diverging signalling proteins such as the Src family of tyrosine kinase, the
Crk adaptor proteins and phospholipase protein (Mayer, 1988). It was soon apparent
that this ~60-residue region of similarity is present in many proteins. Because the
small size of the module seemed to preclude enzymatic activity, the search for
function naturally focused on potential protein interactions, and a screening of
expression libraries using isolated SH3 domains soon identified seemingly specific
binding partners (Cicchetti et al., 1992). Earlier studies indicated that the region
bound by SH3 domains is in all cases proline-rich and identified PxxP as a core
conserved binding motif (Ren et al., 1993). A host of subsequent studies using

17
alanine-scanning mutagenesis of known binding sites, phage display, combinatorial
chemistry and high-resolution structure determination have revealed the specifics of
binding in great detail (Kay et al., 2000). What is clear is that the surface of the SH3
domain bears a relatively flat, hydrophobic ligand-binding surface, which consists of
three shallow pockets or grooves defined by conserved aromatic residues. The ligand
adopts an extended, left-handed helical conformation, termed the polyproline-2 (or
PPII) helix. The PPII helix has three residues per turn; this means it is roughly
triangular in cross-section, and the base of this triangle sits on the surface of the SH3
domain. Two of the three ligand-binding pockets of the SH3 domain are occupied by
two hydrophobic-proline (φP) dipeptides in register on two adjacent turns of the helix,
whereas the third ‘specificity’ pocket in most cases interacts with a basic residue in
the ligand distal to the φPxφP core. Remarkably, different binding modules such as
WW domains and profilin have converged on very similar modes of interaction with
proline-rich, PPII-helical ligands (Kay et al., 2000; Zarrinpar and Lim, 2000).








18

Figure 1.3 SH3 domain, binding pockets/grooves and binding ligand. (A) The structure
shown is based on PDB accession code 1CKB. The SH3 domain is depicted in ribbons, with
secondary-structural elements shown in different colours and labelled. The bound peptide,
PPPALPPKKR, is shown in blue with side chains. Interface residues on the SH3 domain are
shown in pink, for aromatic residues, and in light blue, for non-aromatic residues. The

locations of the xP grooves and specificity pocket on the SH3 domain are identified by broken
arrows. (B) A schematic representation of the same structure to highlight the characteristics of
the ligand-binding surface on the SH3 domain such as the enrichment of aromatic residues.
The same colouring scheme is used in (A) and (B) for the purposes of comparison.
(Biochemical Journal, 2005, Volume 390, 641–653)



19
1.2.2.2 SH3 binding specificity and affinity

1.2.2.2.1 SH3 ligand consensus sequences
The target specificity of particular SH3 domain is so important issue that we can
predict ligands for particular protein of interest and develop ways to inhibit the
interactions in vivo. The PPII helix of SH3 ligand has the similar overall structure and
ligand can bind in either of two orientations: class I has general consensus +xφPxφP
and class II have the general consensus φPxφPx+ (x stands for any amino acid; +
stands for basic residue) (Feng et al. 1994; Lim et al., 1994; Mayer and Eck, 1995).
The dipeptides occupy different positions on the surface of SH3 domain relative to the
helix axis. The determinants for binding to φP pockets are a continuous surface
presented by the alkylated amide nitrogen of proline residue, and the carbonyl, α
carbon and a side chain of preceding residue (Nguyen et al., 1998). The xP dipeptide
is the only naturally occurring amino acid in which two alkylated backbone atoms are
separated by one single carbon (Mayer, 2001). The recently published Hck
SH3/Ligand complex structure suggests that the class I ligand consensus sequence
+xφPxφP should be expanded by a ligand consensus sequence +xxφPxφP (Holger
Schmidt, 2007).

1.2.2.2.2 Artificial ligands for SH3 domain
Several groups have used phage display or combinatorial chemistry approaches to

identify optimal ligands for particular SH3 domains, and in general have been able to
find consensus sequences showing some specificity for a particular domain (Cestra et
al., 1999; Cheadle et al., 1994; Feng et al., 1994; Rickles et al., 1994; Sparks et al.,
1994; Sparks et al., 1996). For synthetic peptide ligands, KD values for ‘specific’

20
ligands tend to be around 10 mM, although in a few cases submicromolar affinities
have been reported (Posern et al., 1998). Such studies have served as the foundation
for an algorithm based on sequence comparison, contacts identified in crystal
structures, and phage-display binding data, which is designed to predict specific
binding sites for any SH3 domain from its primary sequence (Brannetti et al., 2000).

1.2.2.2.3 Contribution of non-core residues of SH3 ligand to interaction
Because the core φPxφP scaffold restricts the variability that can be used to generate
specificity, efforts have addressed the contribution of a third specificity pocket to the
binding specificity of the SH3 domain. Although most SH3 domains prefer an
arginine residue at this site, the Crk SH3 has high selectivity for class II sites that
instead contain a lysine residue (Wu et al., 1995). The Abl SH3, which ironically was
the first for which specific targets were identified, is atypical in that it prefers class I
ligands in which hydrophobic residues contact the third specificity pocket; this is
because it lacks conserved acidic residues found in other SH3 domains, which make
specific contact with the arginine residues of typical ligands (Ren et al., 1993; Weng
et al., 1995).

The residues outside binding core region of the PPII helix sometimes contribute to the
binding with the RT-loop and n-Src loop of the SH3 domain (Feng et al., 1995). For
example, the binding residues for the second SH3 domain of the adaptor Nck have
been carefully mapped, indicating a strong reliance on a serine residue downstream of
the core class II binding motif (PxxPxRxxS). On the other hand, phosphoserine,
acidic and proline residues are not tolerated in positions immediately next to the C-

terminal in the core motif (Zhao et al., 2000).

21

1.2.2.2.4 Evolutionary balance between affinity and specificity
The binding of the SH3 domain of the Src family kinase Hck to the HIV Nef protein
illustrates how regions of a ligand protein distant from the PPII helix can also
modulate specificity. In this case, a single residue in the RT loop of the Hck SH3
confers high-affinity binding to the native Nef, but not to the isolated PxxP-containing
peptide (Lee et al., 1995). After screening a library of mutant Hck SH3 domains (in
which six RT loop residues were randomised) by phage display for binding to the
native Nef, Saksela and colleagues were able to isolate mutant domains that have up
to a 40-fold higher affinity than the wild-type Hck SH3 (Hiipakka et al., 1999). They
also generated mutant SH3 domains that have a very high affinity and selectivity for
an Nef variant bearing a point mutation in the region predicted to contact the RT loop.
This is an excellent example that it is possible to generate SH3 domains that bind very
selectively to a specific target protein. This possibly allows binding sites to be
blocked while leaving others untouched. There must be some evolutionary advantage
to maintaining a relatively low affinity and selectivity for SH3-mediated interaction
(Mayer, 2001).

1.2.2.2.5 non-PxxP ligands
Accumulating data suggest that the SH3 domain sometimes binds with non-PxxP
ligands. For example, the Pix SH3-binding site in Pak (PPPVIAPRPETKS) (Manser
et al., 1998), the Eps8 SH3-binding consensus (PxxDY) (Mongiovi et al., 1999), the
Hbp SH3-binding sites on UBPY (Px(V/I)(D/N)RxxKP) (Kato et al., 2000). The
ultraweak protein-protein interaction exists between the third

SH3 domain and fourth
LIM domain of the PINCH. The interface was highly electrostatic and contains two

22
salt bridges involving the LIM4 R197/SH3-3 D257 pair and the LIM4 R198/SH3-3
E233 pair, and one hydrogen bond involving LIM 4 R198/SH3-3 and N250. This
interaction was mainly electrostatic; however, aliphatic side chains made some
hydrophobic contacts with SH3-3 (Vaynberg J, 2005).

1.2.2.2.6 N-substituted residues
It was proposed that the requirement for proline in the two φP dipeptdies of the
φPxφP core might be due to the unique structure of proline: it is the only N-
substituted amino acid. Nguyen et al. constructed synthetic peptoid ligands, in which
one or both of the proline residues were replaced with various non-natural N-
substituted residues. Higher binding affinities and selectivities were observed for
these non-natural residues compared with natural peptides (Nguyen et al., 1998;
Nguyen et al., 2000). Numerous examples imply that SH3 domains do not generally
recognise the entire ring of proline, but only a small portion of the ring near the
backbone nitrogen. Each groove accommodates two peptide residues. The minimal
and sufficient requirement at each binding groove is a pair of sequential Ca- and N-
substituted residues. The Cα/N-substituted pair may be required because, in this
arrangement, substituent groups are separated by only a single backbone carbon atom,
forming a relatively continuous ridge that can pack efficiently into the domain
grooves (Fig. 1.4b).

1.2.2.3 SH3, WASP and actin polymerisation
Mayer proposed that SH3-mediated interactions can help recruit the WASP to sites of
tyrosine phosphorylation, directly activate the VASP-Arp2/3 complex in the absence
of the GTP-Cdc42, recruit myosin motor and additional Arp2/3-binding sites to sites

23

Figure 1.4 Structural basis of peptoid recognition (A) Structure of wild-type Sos peptide
(PPPVPPRRR) bound to Crk SH3 domain (X. Wu et al., 1995). Proline-rich core binding
grooves are indicated by dashed boxes. Highly conserved surface residues among the four
SH3 domains studied here (one or two conservative amino acid types) are shown in green.
Variable surface residues (31 amino acid types) are in brown. The ligand PXXP core binds at
the most conserved surface on the protein. (B) Structure of peptide 34 bound to Crk SH3
domain. N-(S)-1- Phenylethyl peptoid side chain (orange) bound at site P2. Close-up view
from the same perspective as above. (C) Structure of peptide 39 bound to the Sem5 SH3
domain. N-Cyclopropylmethyl peptoid side chain (orange) bound at site P
-1
. Close-up view
from the same perspective as above (Science, 1998, 282).








24
of actin nucleation, recruit Pak kinase to sites of actin nucleation and activate myosin,
indirectly recruit the WASP to the Nck, couple myosin to the actin-bundling activity,
and indirectly recruit myosin motors and additional Arp2/3-binding sites to the WASP
(Mayer, 2001). The whole picture shows that the nucleation of a new filament at the
appropriate time and place in a cell is controlled by a multicomponent complex of
proteins assembled through SH3-mediated and other protein-protein interactions. The
central role provided by SH3 domain interaction grants more flexibility in possible
interaction, because of the low affinity and modest specificity.

1.3 Protein NMR

1.3.1 The application of NMR in protein structure assignment
Nuclear magnetic resonance spectroscopy, most commonly known as NMR
spectroscopy, is the name given to the technique that exploits the magnetic properties
of certain nuclei. Protein NMR is one of the most important applications of nuclear
magnetic resonance, which mostly explores proton and carbon-13 NMR spectroscopy
for determining three-dimensional structures, dynamic properties and intermolecular
interactions.

Basically, some of the most useful information for structure determination in one-
dimensional NMR spectrum comes from J-coupling or scalar coupling between NMR
active nuclei. This coupling arises from the interaction of different spin states through
the chemical bonds of molecules and results in the splitting of NMR signals. These
splitting patterns can be complex or simple and, likewise, can be straightforwardly

25
interpretable or deceptive. This coupling provides a detailed insight into the
connectivity of atoms in a molecule.

Later, by the introduction of additional spectral dimensions these spectra are
simplified and some extra information is obtained. The invention of multidimensional
spectra was the major leap in NMR spectroscopy apart from the introduction of FT-
NMR. Correlation spectroscopy is one of several types of two-dimensional nuclear
magnetic resonance spectroscopy. Other types of two-dimensional NMR include J-
spectroscopy, exchange spectroscopy (EXSY) and Nuclear Overhauser effect
spectroscopy (NOESY). The first two-dimensional experiment, COSY, was proposed
by Jean Jeener, a professor at University Libre de Bruxelles in 1971. This experiment
was later implemented by Walter P. Aue, Enrico Bartholdi and Richard R. Ernst, who

published their work in 1976 (R.R. Ernst, 1976).

Practically, in NOESY, the Nuclear Overhauser effect (NOE) between nuclear spins
is used to establish correlations. Hence the cross-peaks in the resulting two-
dimensional spectrum connect resonances form spins that are spatially close. NOESY
spectra from large biomolecules can often be assigned using Sequential Walking. The
NOESY experiment can also be performed in a one-dimensional fashion by pre-
selecting individual resonances. This only reveals which peaks have measurable nOes
to the resonance of interest, but obviously takes much less time than the full 2D
experiment.

Much of the recent innovation within NMR spectroscopy has been within the field of
protein NMR, which has become a very important technique in structural biology.