Submitted Report Final Draft

NMR studies of peptide binding to a Src SH3
domain
Owen James Walton
Project Director: Professor Jennifer Potts
Co-director: Dr Gareth Evans

2
Contents:
Abstract................................................................................................................ 4
1.0 Introduction.................................................................................................... 5
1.1 C-Src and N1-Src ................................................................................. 5
1.2 SH3 Domains and their Ligands........................................................... 6
1.3 PD1 Peptide Discovery......................................................................... 8
1.4 NMR Peptide Titration, ITC and DSC .................................................. 9
1.5 In Vitro and Cell-based Peptide Studies .............................................. 9
1.6 Glutathione S-Transferase protein tag................................................ 10
1.7 Project Aims ....................................................................................... 12
2.0 Materials and Methods................................................................................ 13
2.1 GST-N1SH3 ....................................................................................... 13
2.2 Over-expression of 15
N labelled GST-N1SH3..................................... 13
2.3 Harvesting E. coli................................................................................ 13
2.4 Soluble Lysate Preparation ................................................................ 14
2.5 Purification of 15
N GST-N1SH3 .......................................................... 14
2.6 SDS-PAGE......................................................................................... 15
2.7 Size Exclusion Chromatography ........................................................ 15
2.8 Centrifugal Protein Concentrating and Buffer Exchange .................... 16
2.9 NMR Sample Preparation................................................................... 17
2.10 NMR Spectra Acquisition ............................................................... 17
2.11 Saturation Transfer Difference NMR.............................................. 18
3.0 Results ......................................................................................................... 19
3.1 SDS-PAGE Analysis of 15
N GST-N1SH3 Overexpression ................. 19
N GST-N1SH3 .......................................................... 20
3.3 SDS-PAGE Analysis of GST Affinity Fractions................................... 21
3.4 GST Affinity Chromatography with Protease Inhibitors....................... 22
3.5 Protease Cleavage of 15
N GST-N1SH3.............................................. 24
3.6 Size Exclusion Separation of GST and N1SH3 .................................. 25
3.7 SDS-PAGE of Concentrated NMR Samples ...................................... 27
3.8 STD NMR........................................................................................... 28
3.9 HSQC Sample Preparation ................................................................ 31

3
3.10 Analysis and Assignment of N1SH3 HSQC Spectrum................... 31
3.11 Comparison of GST-N1SH3 and N1SH3 HSQC Spectra............... 33
3.12 Shift Perturbation Assay and Structure Mapping............................ 35
4.0 Discussion......................................................................................................................... 38
4.1 Differential PD1 binding to GST-N1SH3 and N1SH3 ......................... 38
4.2 Structural Alterations around the n-Src loop in GST-N1SH3.............. 39
4.3 Conclusions........................................................................................ 41
4.4 Future Studies .................................................................................... 41
Acknowledgements........................................................................................... 42
5.0 References................................................................................................... 43
6.0 Abbreviations .............................................................................................. 47
7.0 Appendices.................................................................................................. 48
7.1 Appendix 1 – Plasmid Map................................................................. 48
7.2 Appendix 2 – Amino Acid Sequences and Protein Data..................... 49
7.3 Appendix 3 – Media Recipes.............................................................. 50
7.4 Appendix 4 – Buffer Recipes.............................................................. 52
7.5 Appendix 5 – SDS-PAGE Reagents................................................... 53
7.6 Appendix 6 – N1-Src SH3 Assigned Residues................................... 53
7.7 Appendix 7 – Overlay of HSQC spectra of C-Src and N1-Src SH3 .... 55

4
Abstract:
The C-Src splice variant, N1-Src, differs only by a six residue microexon insert in
the n-Src loop of its SH3 domain. This insert significantly alters N1-Src’s binding
specificity. A small linear peptide (PD1) which binds a GST tagged SH3 domain of
N1-Src was generated by phage display. In vitro studies have shown this peptide
has biological activity, however, various thermodynamic biophysical studies have
shown no evidence of binding. We have therefore used several nuclear magnetic
resonance (NMR) spectroscopy techniques to examine the binding of this
potential ligand to the N1-Src SH3 domain. Saturation transfer difference (STD)
NMR produced preliminary evidence that the SH3 domain may bind PD1 in its
GST tagged state but does not appear to bind after tag cleavage. Heteronuclear
single quantum coherence (HSQC) NMR and a shift perturbation assay were then
used to determine the structural differences between the GST tagged and
cleaved SH3 domains. This revealed a cluster of residues proximal to the GST
tag/linker and in close spatial proximity around the n-Src loop which had
significantly shifted. These residues also form part of the peptide binding epitope
of the SH3 domain. It therefore seems likely that the GST tag induces an altered
conformation in N1SH3 which facilitates PD1 binding. In view of these results,
whether or not PD1 still represents a biologically relevant N1-Src SH3 ligand must
now be re-evaluated.
Word Count: 225

5
1.0 Introduction:
1.1 C-Src and N1-Src
C-Src is a 536 amino acid non-receptor tyrosine kinase (1) with a role in a myriad
of different cellular processes including differentiation, cell-cell interactions and
extracellular signalling responses (2). Also known as Proto-oncogene tyrosine-
protein kinase Src, it is a pathologically over-expressed oncogene in many
cancers implicated in promoting metastasis (3). Although expressed in all cell
types, C-Src is upregulated in cell types with highly active secretion systems,
particularly neurons. Two distinct neuronal splice variants of C-Src exist, known
as N1-Src and N2-Src. One key difference between C-Src and its neuronal splice
variants is in their differential binding to several proteins involved in vesicle
trafficking such as dynamin and synapsin (4). Whereas C-Src binds these
proteins, N1-Src does not. Indeed the number of N1-Src binding partners is
significantly reduced compared to C-Src and as such, the mechanism of this
specificity is of great interest.
C-Src consists of six distinct domains, a kinase domain, SH4 domain, unique
domain, negative regulatory domain, SH2 domain and SH3 domain, the last two
of which are involved in the regulation of kinase activity (5). Intramolecular
associations between the SH3 domain and the SH2-kinase linker act to repress
kinase activity as the linker resembles an SH3 ligand. One of several ways this
auto-inhibition is relieved is by SH3 ligand binding which displaces the SH2-
kinase linker, allowing the protein to move from a closed to an open and active
conformation (6). This tertiary structure is the same for N1- and N2-Src, both of
which differ from C-Src only by short inserts in their SH3 domain (figure 1). The
N1-Src insert consists of six residues inserted by the microexon N1 into the n-Src
loop of the SH3 domain (5). Although few N1-Src binding partners are known, it is
hypothesised that this alteration is necessary in conferring on it a separate and
distinct role from C-Src in the signalling events that take place in vertebrate
central nervous system development and early neuronal differentiation (7).

6
Neuronal splice variants of C-Src are currently poorly represented in the scientific
literature and as such any research into this area has the potential to be highly
novel. N1-Src overexpression has also been shown in neuroblastoma cells with
the increased ability to differentiate into cells with a neuronal phenotype. This
ability in patients leads to good prognoses, therefore elucidating its function could
potentially lead to novel treatment strategies (8).
1.2 SH3 Domains and their Ligands
SH3 domains are a vast and diverse family of protein modules of between 60 -70
residues which mediate many different protein to protein interactions. The N1-Src
SH3 domain is slightly longer at 73 residues due to the six residue insert. SH3
domains have a highly conserved tertiary structure, that of a beta-barrel with three
distinct loops and a very short 310-helix (9). There is a great deal of structural and
sequence data concerning the known SH3 binding ligands, and meta-analysis of
these data has revealed a consensus motif in the form PxxP in which P stands for
proline and x for any other amino acid (10). This motif has been further
characterised and subdivided into two separate sequences, class I, containing a
positively charged residue (Arg or Lys) before the PxxP (R/KxxPxxP), and class
II, containing a positively charged residue after the PxxP (PxxPxR/K) (11).
Despite these common motifs, new findings continue to suggest that there also
Figure 1. Domain sequence of C-Src and N1-Src showing SH3 insert. U is the unique
domain and R is the regulatory domain.

7
exists a large number of atypical non-consensus sequence peptides which bind to
SH3 domains via unconventional structural interactions (12). The conventional
structural interactions involve a three pocket model in which the peptide xP
sequence binds two hydrophobic pockets and the rest of the peptide binds a
pocket created by the variable n-Src and RT loop (11). N and C terminal flanking
residues to the core proline motif are extremely important in determining binding
affinity and specificity. Class I and II ligands have been shown to adopt opposite
orientations in SH3 binding, indicating the importance of maintaining the positive
residue in the same position relative to SH3. Structural characterisation of ligand
binding orientation to various SH3 domains has led to a wealth of information
enabling researchers to now predict the relative affinity and orientation of general
SH3 ligand binding (13).
Figure 2. (A) Ribbon diagram of the solution structure of the C-Src SH3 domain (23)
with secondary structures labelled. (B) Ribbon diagram of the solution structure of C-
Src SH3 bound to a class II peptide (PRL1) in red (13). PRL1 sequence –
AFAPPLPPR. (C) Secondary structure sequence of the N1-Src SH3 domain showing
the six residue insert in the n-Src loop. β indicates beta strand, RT is the the RT loop,
n-src is the n-Src loop, DH is the distal hairpin and H is the 310-helix.

8
Numerous, SH3 binding, C-Src ligands are known and the structural basis of their
interaction has been elucidated primarily by NMR studies (14) and a few
crystallographic studies (15). The small size of SH3 domains (~8-9 kDa, well
below the conventional NMR limit of 35 kDa) makes them particularly attractive
for NMR based studies, as evidenced by the 361 papers in PubMed containing
the words ‘SH3’ and ‘NMR’ (as of 26/03/14). It is hoped that NMR can be used, in
the same way as for C-Src, to structurally characterise the N1-Src SH3 domain
and its interactions with binding partners.
1.3 PD1 Peptide Discovery
Since very few ligands of N1-Src SH3 are known, an experimental rationale for
attempting to discover novel binding proteins was developed (5). If a consensus
sequence for N1-Src SH3 binding could be discovered, then a bioinformatic
screen for that sequence in the mammalian proteome could be carried out,
identifying novel potential binding partners. The technique chosen to identify this
consensus sequence was phage display (5) as this had previously been used to
successfully validate the binding of class I and II core motif containing peptides to
C-Src SH3 and identify novel flanking peptide residues (16). A random, non-
biased library of 12-mer linear peptides was used. The peptide eventually
selected by this technique, was named PD1 (Phage Display 1) and had the
sequence WHRMPAYTAKYP. Interestingly, the six highest affinity peptides
generated against N1SH3 all contained the motif +xPxxTx+, where x is any amino
acid and + is a positively charged amino acid. This sequence is not a canonical
SH3 binding motif (PxxP) and therefore represented a novel binding sequence.
The presence of a positively charged residue at either end is also unusual and
means the peptide could not be classed according to the class I/class II
nomenclature. A mutant version of PD1 named P5A was generated as a negative
SH3 binding control with the key proline residue mutated to an alanine. P5A:
WHRMAAYTAKYP.

9
1.4 NMR Peptide Titration, ITC and DSC
To validate binding, an HSQC NMR experiment was performed, in which cleaved
15
N N1-SH3 was titrated with PD1 in order to determine whether there were any
peak shifts associated with binding1
. The results of this suggested that no binding
took place as none of the peaks in the spectrum shifted. A second attempt to
validate binding was tried, using the more quantitative thermodynamic techniques
of Isothermal Titration Calorimetry (ITC) and Differential Scanning Calorimetry
(DSC). The results of the ITC showed no evidence of binding between cleaved
N1-SH3 and PD1 and the DSC showed highly abnormal curves not indicative of
binding2
. This raised the question as to whether, during the phage display, the
PD1 peptide had bound to an altered conformation of the SH3 domain induced by
the bulky GST tag which it then could not bind to once cleaved. Alternatively, the
peptide could have bound to an epitope shared by GST and N1-SH3 or the 13
residue linker joining them. In order to test this, an HSQC comparison of free SH3
and GST-N1SH3 was necessary to determine whether there were any significant
peak differences. If enough peaks were present in the GST-N1SH3 fusion
spectrum, then a PD1 titration with GST-N1SH3 could be performed to test for
peak shifts characteristic of binding.
1.5 In Vitro and Cell-based Peptide Studies
Perplexingly, despite the lack of biophysical validation for PD1 binding free SH3,
in vitro kinase assays showed that the presence of PD1 c-terminal to an ideal Src
substrate significantly increased substrate phosphorylation by N1-Src compared
to C-Src (5). Presumably PD1 enhances substrate docking to N1-Src and so
decreases the Km of binding. Data from cultured cell experiments also showed
apparent PD1 activity (5). COS7 fibroblast-like cells were used as they do not
naturally express neuronal Src isoforms. N1-Src and CFP tagged PD1 were
transfected into the cells and over-expressed. The morphology of heterologous
cells like COS7 which have been transfected with N1-Src is characterised by
neurite-like outgrowths from the cell body. The co-expression of CFP-PD1
1
Prof J. Potts, personal communication.
2
Dr G. Evans, personal communication.

10
however resulted in an inhibition of this neurite-like morphology. Importantly the
P5A mutant form of PD1 did not inhibit this morphology when co-transfected with
N1-Src. This morphological inhibition could be due to other non N1-Src specific
interactions of PD1, however, contrary to this, it was also shown that PD1 and
N1-Src form a complex, as they co-immunoprecipitate from the COS7 lysate. An
in vitro study involving PD1 titration into a reaction mix of Src substrate and N1-
Src was also performed and found phosphorylation decreased. P5A had no
effect. It was therefore hypothesised that PD1 titrates out the natural substrate of
N1-Src via a very specific interaction with the SH3 domain and so inhibits its
downstream signalling effects. It is important to note that in these studies, the full
length N1-Src was used as opposed to the SH3 domain alone as in the
biophysical studies. PD1 was also present with a protein tag in these
experiments; however it was cleaved from the fusion tag in the biophysical
experiments. SH3 and PD1 may well be structurally/functionally altered when
attached to much larger molecules.
1.6 Glutathione S-Transferase protein tag
Glutathione S-Transferase (GST) is a very widely used protein tag for affinity
purification of proteins and domains for biochemical and biophysical studies. It
has the benefit of increasing expression, solubility and stability of the fused
protein (17). The potential drawback of using GST compared to other common
affinity tags such as poly-histidine tags for Ni2+
affinity purification is its large size
of 25 kDa. Very few problems have been reported with GST however and it has
been used as an expression tag for many other studies of SH3 binding peptides
using phage display (16, 18).
GST tags have other NMR specific benefits which depend on its dimerization into
a 50 kDa complex which is then not seen in the NMR spectrum, whilst leaving the
fused protein/domain visible (19). This relies on the fact that the larger a protein
is, the slower it tumbles and therefore the quicker its signal decays. This gives a
weaker signal with a poor signal to noise ratio and broad line-widths. If a smaller,
NMR visible, protein or domain is attached to GST via a sufficiently long and
flexible linker, then it should theoretically tumble fast enough to give a good signal

11
independent of GST. Therefore the GST does not necessarily need to be cleaved
before NMR can be carried out. As apparent in Figure 3, the C terminus of GST,
where the linker-N1SH3 domain is attached, is on the opposite side of the
molecule to the dimerization interface. This implies that any structural variations in
N1SH3 are not dependent on GST dimerization.
A potential disadvantage of GST tags is their reported ability to cause false
positive peptide hits in phage display. K.K. Murthy et al. (20) reported
identification of a peptide which bound to a GST-PDZ domain fusion, of a similar
size to the GST-N1SH3 fusion, but did not bind to the free PDZ domain. They
verified this by an HSQC peptide titration assay similar to the one carried out on
N1SH3 and PD1, however they offered no structural or explanatory analysis of
these results. M. Zhang et al. (21) conducted a similar study involving GST fused
to different domains of a viral protein to examine binding to scFv antibody
fragments and found they had identified several false positive results. Again, they
offered no real explanatory analysis. Importantly however, they observe that
Figure 3. X-ray crystal structure of dimerized GST. Monomers shown in blue and
green. C terminal lysines highlighted in red indicate position of linker and N1SH3 in
relation to dimer interface. Structure generated in PyMOL using Protein Database
(PDB) structure 1HNB (31).

12
these false positives were excluded when using capture phage ELISA as
opposed to indirect phage ELISA to characterise the phage display selected
clones. Interestingly, the PD1 generating phage display did not use ELISA
methods to validate peptide hits and instead used a kinase assay based method.
1.7 Project Aims
The fundamental aim of this study was to ascertain by heteronuclear 2-
dimensional NMR whether there is a conformational change in the SH3 domain
when fused to GST, compared to when free. If possible, the residues whose
peaks have shifted would be identified using assignments of the SH3 domain
based on a previous peak assignment of the C-Src SH3 domain (22). If this
conformational change was confirmed, a peptide titration of PD1 with the two SH3
variants, GST fused and free, would be performed. This would assess which
peaks shifted and therefore may be involved in the binding interface between PD1
and the SH3 domain. If residues which had been altered in the GST fused state
shifted, this would be indicative of a GST dependent structural alteration of the
SH3 domain which facilitates PD1 binding.

13
2.0 Materials & Methods:
2.1 GST-N1SH3
A glycerol stock of BL21 Escherichia coli transformed with the pGEX6P-1 plasmid
(ampicillin resistant) with a GST-N1SH3 insert was used for all experiments
(Appendix 1). The GST- N1SH3 fusion contains a 13 residue linker with a Human
Rhinovirus 3C Protease cleavage site (sequence in Appendix 2) and was in an
inducible lac operon. The post-3C cleaved N1SH3 domain was 78 residues long
as it retained a five residue linker remnant on the N-terminus. The post 3C
cleaved GST was 226 residues long including the eight residue linker remnant on
the C-terminus.
2.2 Over-expression of 15
N labelled GST-N1SH3
50 ml of LB (Appendix 3, Table 1) containing 100 µg/ml ampicillin (Melford Labs)
was inoculated with a glycerol stock of transformed BL21 cells. This was
incubated at 37°C in a shaker at 180 rpm overnight. 10 ml of overnight LB culture
was added to 1 L of 15
NH4Cl supplemented M9 minimal media (Appendix 3, Table
2) in a 2 L baffled conical flask. The M9 culture was incubated at 37°C with
shaking at 200 rpm until reaching a mid log density (OD600 = ~0.6) as measured
using an Eppendorf® BioPhotometer. A 1 ml aliquot of 1 M IPTG (Melford Labs)
in H2O was added to the culture at this point in order to induce 15
N labelled GST-
N1SH3 over-expression. The 1 L M9 culture was then placed in a 20°C shaker,
shaking at 180 rpm overnight for around 18 hours.
2.3 Harvesting E. coli
The following day a final OD600 reading of the M9 culture was taken. The culture
was split into two equal volumes and centrifuged in a Sorvall Evolution
Ultracentrifuge at 5,471 xg at 4°C for 20 minutes. The supernatant was discarded
and each pellet was re-suspended in 15 ml of PBS (Appendix 4, Table 3). Both 15

14
ml suspensions were pooled in a 50 ml Falcon tube and either stored at -20°C or
used immediately for GST affinity chromatography.
1 ml of culture was removed prior to the addition of IPTG (point of induction
sample) and another 1 ml removed in the morning after induction (post-induction
sample). Both samples were centrifuged at 3.5 xg for five minutes in a Fisher
Scientific accuspin Micro17 centrifuge. The supernatant was discarded and the
pellet re-suspended in a volume of PBS 1/20th
of the value of the OD600 for each
sample’s time point. E.g. if the OD600 = 1, then the pellet was resuspended in 0.05
ml. 10 µl of this suspension was added to 10 µl of sample buffer. The remaining
post-induction sample had a volume of 10X BugBuster® (Merck Millipore) 1/10th
of the volume of the sample added to it. This was left rocking at room temperature
for 20 minutes and then centrifuged at 3.5 xg for five minutes. 10 µl of
supernatant was removed and added to 10 µl of sample buffer. The samples were
then prepared according to the SDS-PAGE protocol (chapter 2.6).
2.4 Soluble Lysate Preparation
The 30 ml PBS pellet suspension was sonicated on ice with a Misonix Sonicator
3000. A three minute programme was run with a pulse sequence of three
seconds ON and seven seconds OFF with a power of ~75 Watts. The resulting
lysate was centrifuged at 39,191 xg for 30 minutes at 4°C on a Beckman Avanti
J26 Ultracentrifuge. The supernatant was recovered to use for GST affinity
chromatography and the pellet discarded.
N GST-N1SH3
An Amersham Pharmacia Biotech AKTA Prime or AKTA Purifier FPLC system
was used for GST affinity chromatography. Two 5 ml GE Healthcare GSTrapTM
HP columns were used in series to create a 10 ml column. After the system was
washed in H20 and the 10 ml column equilibrated with binding buffer (PBS), 30 ml
of lysate was flowed through at a flow rate of 1.0 ml/min for 90 minutes, with the
output flow re-directed into the lysate to ensure cyclic flow. Subsequently, a wash

15
step was performed by flowing binding buffer at 3.0 ml/min through the column
and system until the UV280 trace reached zero. The flow-through was collected
and stored. Bound proteins were eluted from the column using a 100% step
gradient from binding buffer to elution buffer (Binding buffer + 10mM Glutathione
Reduced (Fisher Bioreagents)) with a flow rate of 3.0 ml/min collecting 4 ml
fractions. Fractions containing eluted protein were indicated by the UV280 peak on
the chromatogram. The lysate was collected after the 90 minute binding step and
stored to be run on an SDS-PAGE gel with the eluted protein fractions. This same
protocol was used when protease inhibitors (Appendix 4, Table 4) were added to
the buffers.
2.6 SDS-PAGE
Precast NuPAGE® Novex® 12% Bis-Tris, 1.0 mm 12 well gels were used for
every experiment. 5 µl of BioRAD Precision Plus ProteinTM
All Blue molecular
weight marker was loaded into the first well of each gel. Samples were prepared
by adding Sample Buffer (Appendix 5, Table 6) and Elga filtered H20 to dilute if
necessary and then heating at 95°C for five minutes. Gels were run in a gel-rig
with 500 ml of 1X MES buffer (Appendix 5, Table 7) for 45 minutes at 200 V.
Proteins were visualised by staining in Coomassie brilliant blue stain and then
destained with destaining solution (10% Ethanol and 10% acetic acid in dH20).
Gels were visualised on the lower white setting of the Syngene Gene Genius Bio
Imaging system at 40 ms and the images were photographed and saved with the
Syngene GeneSnap program.
2.7 Size Exclusion Chromatography
An Amersham Biosciences AKTA purifier with a HiLoadTM
16/60 SuperdexTM
75
prep grade column was used. The column and system were first washed with 0.2
µm filtered H20. Prior to sample loading, the column was equilibrated with Size
Exclusion Buffer (Appendix 4, Table 5) for around two hours with a flow rate of 1
ml/min. To prepare the protein samples for size exclusion chromatography, they
were first concentrated to around 1 ml using the centrifugal method with Vivaspin

16
tubes (Cha. 2.8). The samples were loaded with a syringe and then run using the
programme ‘Superdex 75 1660 Sec B1’ with a flow rate of 1.0 ml/min, pressure
limit of 0.5 MPa and 100% Size Exclusion Buffer.
2.8 Centrifugal Protein Concentrating and Buffer Exchange
Depending on the molecular weight of the purified protein intended for
concentrating, a Vivaspin column with a Molecular Weight Cut-Off (MWCO) of
around 1/3rd
of the protein MW was used. The available Sartorius Stedium
Biotech Vivaspin tubes had MWCOs of 3 kDa, 5 kDa or 10 kDa. No Buffer
exchange was performed on the samples straight from size exclusion
chromatography; however it was performed on samples straight from GST affinity
chromatography. This was because concentrated NMR samples must be in SE
(Size Exclusion) buffer and because GST affinity elution buffer contains
glutathione which may appear in the 1
H spectrum unless diluted.
To concentrate proteins already in SE buffer, the Vivaspin tube was first washed
with 20 ml of SE buffer in order to remove any glycerol present on the filter. This
was achieved by centrifuging in the Scientific Heraeus Megafuge 16R at 4,696 xg
at 4°C for as long as required for all buffer to flow through into the waste
chamber, typically ~20 minutes. The flow-through liquid was discarded and the
protein loaded and centrifuged under the same conditions until only the desired
volume of protein remained in the top fraction (500 µl – 700 µl for NMR).
The buffer exchange procedure was the same with the exception that once the
protein was concentrated to 1 ml, 19 ml of SE buffer was added and centrifuged
again. This was performed three or four times in order to achieve a glutathione
dilution to 0.125 µM or 6.25 nM, respectively. After the last wash, the protein was
concentrated to 500 µl – 700 µl. All proteins were concentrated to a minimum
concentration of 200 µM within the stated volume. All concentrations were
calculated using the Beer Lambert law, and an Eppendorf® BioPhotometer was
used to obtain the UV280 readings. This spectrophotometer substracts the 320 nm
reading from UV280 to give a more accurate reading which accounts for light
scattering interference.

17
2.9 NMR Sample Preparation
The final NMR samples needed 10% Deuterium Oxide (D20) in a volume between
550-600 µl. This was to provide the lock signal to shim the spectrometer magnet.
60 µl of D20 (ARMAR chemicals) was added to 540 µl of concentrated protein.
The sample was transferred to the 5 mm Thin Wall Precision NMR sample tube
(Wilmad-LabGlass) and back to the eppendorf and then pH adjusted to the
desired pH (6.9) with 1 µl additions of 0.1 M or 1 M solutions of HCl and NaOH.
The samples were then transferred to NMR tubes ready for loading into the
spectrometer. The addition of D20 decreases the protein concentration by 10%
and so this was factored into the final concentration calculations presented in the
HSQC figure legends in the Results chapter.
2.10 NMR Spectra Acquisition
NMR spectra were recorded using a Bruker Avance 700 MHz spectrometer with a
triple resonance TXI probe and two gradient coils at 298 K and Topspin software.
All NMR experiments were carried out by Pedro Aguiar in the centre for magnetic
resonance in the Chemistry department. 1D spectra and 2D Heteronuclear Single
Quantum Coherence (HSQC) spectra were acquired for each protein. HSQC
spectra were converted from Topspin 2.0 to NMRViewJ format for peak analysis.
Pulse Program hsqcetf3gpsi
No. Of Scans 24 – GST-N1SH3, 48 – GST and N1-SH3
1
H 15
N
No. Of Points 2048 256
Sweepwidth (Hz) 11,261.262 2128.558
(ppm) 16.0845 30.00
Offsets (ppm) 4.690 118
Table 1. 2D HSQC spectra acquisition parameters.

18
2.11 Saturation Transfer Difference NMR
The PD1 peptide (1,733 Da) was supplied by alta Bioscience Ltd. This had >90%
purity according to manufacturer.
Sequence: Acetyl-GGGWHRMPAYTAKYP-amide
N-terminal acetyl glycine cap and C-terminal amide protecting groups added to
PD1 sequence in order to remove charge and reactivity of termini and mimic in-
protein like sequence. All solutions of PD1 were prepared in Elga H20.
Concentrations were calculated estimating that PD1 contained 20% H20 in dry
weight. Samples made for GST-N1SH3 and N1SH3 only, not GST. Each sample
was 600 µL with 10% D20, 1 mM PD1 and 10 µM protein, giving a 100:1 ratio of
peptide (ligand) to protein. Both samples were pH matched.
Pulse Program zgesgp
No. Of Scans 32
1
H
No. Of Points 16,384
Sweepwidth (Hz) 11,160.714
(ppm) 15.9408
Offsets (ppm) 4.690
(Hz) 3287.11
Table 2. 1D spectra acquisition parameters for STD experiments.

19
3.0 Results:
3.1 SDS-PAGE Analysis of 15
N GST-N1SH3 Overexpression
The purpose of these experiments was to verify, prior to purifying 15
N GST-
N1SH3, that the protein had been over-expressed and was in the soluble fraction
of the bacterial lysate. If over-expressed and soluble then GST affinity
chromatography was performed on the cell lysate to purify the fusion protein.
Figure 4 is consistent with the assumption that GST-N1SH3 (MW = 35.023 kDa)
was not noticeably expressed prior to induction and was over-expressed 18 hr
post-induction. This is evidenced by the absence of a distinct band directly below
the 37 kDa MW marker band in the Ind lane and the presence of a strong, distinct
band in the 18 lane. Lane 18 (BB) shows that GST-N1SH3 is in the soluble
fraction. After over-expression, the BL21 bacterial cells were harvested (Cha. 2.3)
Figure 4. SDS-PAGE gel showing different level of 15
N GST-N1SH3 expression prior
to induction and 18 hrs post induction. Lane MW is molecular weight marker. Lane Ind
is the point of induction sample, lane 18 is the 18 hr post induction total lysate sample
and lane 18 (BB) is the same sample but only the soluble protein fraction. Equivalent
cell density loadings according to Cha. 2.3.

20
and then sonicated and centrifuged (Cha. 2.4) in order to separate the soluble
from the insoluble components of the cells.
N GST-N1SH3
15
N GST-N1SH3 was purified from the lysate by GST affinity chromatography
(Cha. 2.5).
Figure 5 shows a typical GST affinity chromatogram with a clear peak around
fractions 6 and 7 which tails off in fractions 8 – 12. A selection of these fractions
were then analysed by SDS-PAGE in order to determine whether they contained
pure 15
N GST-N1SH3.
Figure 5. Typical chromatogram from GST affinity chromatography. Y axis showing
UV280 absorbance in milli Absorbance Units (mAU). X axis showing elution volume
(mls) and fraction number in red. Blue trace shows UV280 absorbance reading, brown
shows conductivity, however the units (mS/cm) are not shown and the green trace
shows the % input from line B which in this experiment was in elution buffer (PBS +
10 mM Glutathione). Chromatogram generated by the UNICORN 5.11 program.

21
3.3 SDS-PAGE Analysis of GST Affinity Fractions
5 µl of lysate was removed before and after it was run on the GST column and 5
µl of the flow-through from the wash step was removed. These volumes were
added to individual volumes of 15 µl of H20 and 10 µl of sample buffer and were
prepared and loaded on an SDS-PAGE gel (Cha. 2.6). For each of the fraction
samples, 10 µl of the eluate was added to 10 µl of H20 and 10 µl of sample buffer.
A 10 µl aliquot of each sample was loaded on the gel.
From figure 6 it is apparent that the chromatogram peak fractions 6-10 contain, as
expected, GST-N1SH3. The purification was not perfect however and a large
amount of contaminating GST was present in fractions 6-10 as seen by the
intensity of the bands just above the 25 kDa marker. Small amounts of
free/cleaved N1SH3 are present in fractions 6 and 7. A very small amount of GST
is to be expected as a contaminant after GST affinity chromatography as any
GST present in the lysate will also bind to the column and elute with the fusion.
However, this does not explain the presence of a high concentration of GST in the
fractions and also does not explain the presence of free N1SH3. This would
Figure 6. SDS-PAGE gel showing the eluate containing fractions from the GST affinity
chromatrography as well as the lysate before and after flowing through the GSTrap
columns. Lane MW is molecular weight marker, lane L is the pre-chromatography lysate,
lanes 3-10 are the fraction samples (See Figure 5), lane PFL is the post-chromatography
flow lysate and lane FT is the wash step flow-through sample.

22
indicate that 3C protease cleavage was taking place in the column as the GST
and N1SH3 must have formed after the fusion bound to the column. The
explanation for this observation was that the GSTrap column used in this
purification had recently been used for an on column 3C cleavage.
A second purification was therefore performed using a different GSTrap column
and protease inhibitors were added to the binding and elution buffer to inhibit any
other potential protease contamination. The partially cleaved GST-N1SH3 was
saved for full 3C cleavage in order to separate GST and N1SH3 by Size
Exclusion chromatography ready for NMR.
3.4 GST Affinity Chromatography with Protease Inhibitors
15
N labelled GST-N1SH3 was overexpressed and extracted from BL21 cells. For
this experiment, two purifications of GST-N1SH3 from two lysates of two separate
1 L M9 cultures, A and B respectively, were performed. The first purification, i.e.
A1, was performed on the lysate straight after sonication and centrifugation and
the second purification, i.e. A2, was performed on the lysate after the first round
of chromatography. All fractions containing the elution peak were pooled for each
purification and a UV280 reading was taken to measure the concentration of
protein in the pooled fractions. The concentration was converted into mg/ml and
then the volume containing 2 µg was calculated. This volume was multiplied by
three and added to H20 to make a total volume of 20 µl. 10 µl of sample buffer
was added to give a final sample containing three 2 µg/10 µl gel loads. These
samples were prepared and loaded on a gel.

23
From figure 7 it is apparent that the purity of GST-N1SH3 from these elutions was
particularly high and there appeared to be almost no GST or N1SH3
contamination. Since these samples were all loaded with the same amount of
protein and this was a known amount compared to the gel in figure 6, there
appears to be less protein loaded in the gel in figure 7. This might explain why
there seems to be less contamination, as the concentration of contaminants may
be lower and therefore the bands less intense. Nevertheless, contamination
appeared insignificant for the purposes of NMR and so samples A1, A2, B1 and
B2 were pooled. This pooled GST-N1SH3 could then be concentrated and buffer
exchanged into Size Exclusion buffer for NMR (Cha 2.8). In order to obtain
purified samples of GST and N1SH3 the previously pooled volume of partially
cleaved GST-N1SH3 was fully cleaved.
Figure 7. SDS-PAGE gel showing 2 µg sample loads of pooled eluate fractions from two
consecutive purifications of two separate GST-N1SH3 expression cultures. Lane MW is
molecular weight marker, lanes A1 and B1 are the pooled eluate from the first rounds of
chromatography on each separate culture and lanes A2 and B2 are pooled eluate from
the second rounds of chromatography.

24
3.5 Protease Cleavage of 15
N GST-N1SH3
The partially cleaved GST-N1SH3 sample was concentrated to a volume of ~500
µl and the concentration determined by UV280 absorbance. It was then fully
cleaved by the addition of 3C Protease in a ratio of 1 part protease: 50 parts
protein by mass (mg). This ratio was based on the results of a previously
conducted cleavage trial of GST-N1SH33
. The cleavage reaction was left at 4°C
overnight for approximately 18 hr.
Figure 8 shows that full cleavage of GST-N1SH3 into GST and N1SH3 was
achieved with a 1:50, 3C to protein ratio in 18 hr.
3
J. Hawkhead, personal communication.
Figure 8. SDS-PAGE gel showing partial and full cleavage of GST-N1SH3 before and
after addition of 3C protease. Lane MW is molecular weight marker, lane Pre is the
partially cleaved sample prior to 3C addition and lane Post is the fully cleaved sample
18 hr after 3C addition.

25
3.6 Size Exclusion Separation of GST and N1SH3
3C protease, GST and N1SH3 were separated in order to obtain pure GST and
N1SH3 for concentrating (Cha 2.8).
From the chromatogram alone it was not possible to determine whether
seperation of GST and N1SH3 was clean and complete and so an SDS-PAGE
gel of the fractions was run. A selection of eight fractions covering each of the
four peaks were run on the gel. 5 µl of each fraction was added to 2.5 µl of
sample buffer and 7 µl of each of these samples was loaded on the gel.
Figure 9. Size Exclusion chromatogram showing three distinct peaks and 1 very shallow
peak. Y axis showing UV280 absorbance in milli Absorbance Units (mAU). X axis showing
elution volume (mls) and fraction number in red. Blue trace shows UV280 absorbance
reading, brown shows conductivity however the units (mS/cm) are not shown and the
green trace shows the % input from line B which in this experiment was 100% Size
Exclusion Buffer. Chromatogram generated by the UNICORN 5.11 program.

26
Faint bands in fractions A11 – B15 can be observed running just below the 75
kDa marker. These are most likely 3C protease running at a lower weight than
expected which is typical. There are also faint bands in fractions A11 and A13
which run at around 50 kDa and just under 37 kDa. The 50 kDa bands are most
likely GST dimers and the ~37 kDa bands are most likely contaminating GST-
N1SH3. GST dimers should not appear in denaturing SDS-PAGE gels, however
often trace amounts persist for reasons unknown. Due to the significant 3C
contamination in B15 and the low concentration of GST in B14, only fractions
A11-A14 were pooled for the GST sample. Due to the low concentration of
N1SH3 and the presence of traces of GST in B12, only fractions B7 and B8 were
pooled for the N1SH3 sample. These pooled samples of GST and N1SH3 could
then be concentrated (Cha. 2.8).
Figure 10. SDS-PAGE gel showing the separation of GST and N1SH3 into different
fractions after size exclusion chromatography. Lane MW is molecular weight marker,
the rest of the lanes correspond to the fractions on the chromatogram (figure 9). A11-
A14 is the first peak, B15 – B14 is the second peak, B12 is the third peak and B8-B7
is the fourth peak.

27
3.7 SDS-PAGE of Concentrated NMR Samples
The concentrated GST-N1SH3, GST and N1SH3 samples were all run on an
SDS-PAGE gel in order to assess their purity and whether the fusion protein was
still intact and un-cleaved. Sample concentrations were: GST-N1SH3 = 1.78 mM,
GST = 371 µM and N1SH3 = 643 µM. These concentrations all decreased by
10% in the final NMR samples due to the addition of D20.
From figure 11 it appears that only N1SH3 is completely pure, however the level
of contamination in the other samples seems very low. The GST-N1SH3 sample
appears to have minor 3C protease contamination from the band running just
above 75 kDa and also some GST contamination. GST appears to have
contaminating fusion and a larger contamination of some unknown proteins
running between 50-70 kDa. These contaminants do not run at the same
molecular weight as 3C protease, however no other contaminants apart from
potentially dimerized GST-N1SH3 (MW = ~70 kDa) could be present in the
sample. Despite these contaminations, the desired protein species were deemed
pure enough (>95%) for the purposes of NMR.
Figure 11. SDS-PAGE gel showing each of the concentrated samples in Size
Exclusion Buffer. Lane MW is molecular weight marker, lane F20 is a 20 µg load of
GST-N1SH3, F2 is a 2 µg load of the same protein, GST is a 2 µg load of GST and
SH3 is a 2 µg load of N1SH3.

28
3.8 STD NMR
Although initially not an experiment planned for this project, STD was used to
verify whether PD1 bound to GST-N1SH3 and not to N1SH3 which was the
starting hypothesis of this project. A 100 fold higher concentration of PD1
compared to each protein was used so that only the PD1 1
H peaks were visible.
Four different positions within the 1D spectra of the proteins were irradiated (0.49,
0.76, 5.38 and 8.91ppm) in order to see which achieved optimal saturation.
Figure 12. (A) 1D spectra of 1 mM PD1 + 10 µM N1SH3 on top and N1SH3 (276 µM) on bottom.
(B) 1D spectra of 1mM PD1 + 10µM GST-N1SH3 on top and GST-N1SH3 (294 µM) on bottom.
The four irradiation points on the protein spectra are indicated on the top 1D spectra with black
arrows and the PD1 control irradiation (6.72ppm) is marked with a blue arrow. H20 signal seen
at ~4.7ppm.

29
In order to ensure irradiation was occurring at the correct point, a control
experiment was carried out in which one position (6.72 ppm) on the PD1
spectrum was irradiated. As apparent in figure 13, the difference spectra for these
irradiation points shows only one peak in the region irradiated, indicating a
specific irradiation event. From the four protein irradiation points, 0.49ppm was
selected as the best spectrum.
As the 6.72ppm irradiation showed that irradiation was occurring at the right part
of the spectrum and was effective in the saturation of the 1
H signal, the
experiments where the protein peaks were irradiated can be considered reliable.
Figure 13. 1D spectrum of off-resonance (non-irradiated) PD1 + N1SH3 (red). 1D
difference spectra (off resonance – on resonance spectra) showing specific saturation of
the 6.72ppm PD1 signal (blue). Difference spectra looked the same for the 6.72ppm
irradiation experiment for the GST-N1SH3 sample (not shown).

30
The difference spectrum shown in figure 14 (A) strongly suggest that no binding
takes place between free N1SH3 and PD1 since there are no clear peaks beside
that at 0.49ppm even when scaled by a factor of 64. It is difficult to say whether
the difference spectrum in (B) suggests any binding between PD1 and GST-
Figure 14. 1D off resonance spectrum of PD1 + N1SH3 (A) and PD1 + GST-N1SH3
(B) in red and 0.49ppm irradiation difference spectra for each in blue. (A) Black arrow
indicates irradiation peak at 0.49ppm. (B) No irradiation peak seen, black arrows
denote potential difference peaks indicating PD1 binding. Both difference spectra
scaled up by a factor of 64 compared to off resonance 1D spectra.

31
N1SH3. There appear to be three small peaks which correspond to the PD1 1D
spectrum and therefore could represent a 1
H saturation transfer between GST-
N1SH3 and PD1 indicative of binding. This is the first time any biophysical studies
to test for binding between GST-N1SH3 and PD1 have been carried out. Although
the result is not entirely conclusive, there is some evidence of binding and
therefore the HSQC comparison between GST-N1SH3 and N1SH3 was carried
out to examine any potential structural differences.
3.9 HSQC Sample Preparation
HSQC spectra were acquired at pH 6.9. This was due to the fact that despite the
samples being prepared in Size Exclusion buffer at pH 6.5, the GST-N1SH3
sample was measured at pH 6.9. This was after having added D20, washed the
sample down the inner side of the NMR tube and then transferred it back to the
eppendorf for pH adjusting. Upon the addition of 1 µl of 0.1 M HCl, cloudy white
precipitate began to form. It is likely that this precipitate may have been mostly
contaminating GST (pI = 6.09) and may have precipitated due to a local
concentration effect of HCl addition. The sample was centrifuged and the soluble
fraction removed and the concentration recalculated (1.23 mM). Both GST and
N1SH3 samples were then pH adjusted to 6.9. The GST sample appeared slightly
cloudy, suggesting a small degree of precipitation.
3.10 Analysis and Assignment of N1SH3 HSQC Spectrum
The N1SH3 domain in these studies is a 78 residue version with a five residue N-
terminal section of attached linker. The domain contains three prolines which
have no amide backbone peaks and therefore 74 backbone peaks were
expected. This prediction takes into account that, due to rapid proton exchange
with the solvent, the N terminal residue is never seen in an HSQC and so an extra
peak can be subtracted. The domain also contains two tryptophans (contributing
one side chain peak each), three glutamines and three asparagines (each
contributing two side chain peaks). This results in 14 side chain peaks in total and
88 peaks for the whole spectrum. After counting the spectral peaks for N1SH3

32
and dividing them regionally into side chain and backbone peaks it was concluded
that the spectrum matched the prediction exactly.
Residues were then assigned to the peaks based on the assignments of a C-Src
SH3 spectrum generated within the lab, which in turn was based on assignments
of a C-Src SH3 domain by Yu et al. (22). Residues were assigned according to
two categories, confident and tentative based on how well they overlapped with
the C-Src SH3 HSQC spectrum (C-Src and N1-Src SH3 HSQC overlay in
appendix 7); confident being a high degree of overlap and tentative being
reasonably similar spatial proximity (table of residues and their assigned category
in appendix 6).
Only 55 peaks could be clearly assigned this way, 47 of which were backbone
peaks. The remaining peaks showed no spatial similarity to the C-Src SH3 HSQC
and so could not be confidently assigned. This is most likely due to structural
differences between the two variants caused by the n-Src loop insert. Since there
is no conclusive triple resonance assignment for N1SH3 yet, these assignments
remain partial and to some extent speculative.
Figure 15. HSQC spectrum of N1SH3 (579 µM, pH 6.9, 48 Scans)
with assigned peaks labelled and side chain peaks circled.

33
3.11 Comparison of GST-N1SH3 and N1SH3 HSQC Spectra
As expected, very few peaks (17 peaks out of 226 residues) were seen on the
GST HSQC (figure 16) due to GST dimerization. The fact that the intensities of
these peaks remained high enough to be detected suggests that they are
residues within disordered regions of GST for instance loops or the eight residue
C-terminal linker section. This linker section contains a glutamine and there is a
clear glutamine/asparagine side chain in the spectrum.
Since figure 16 shows that GST produces some peak signals, it was important to
establish which peaks in the GST-N1SH3 HSQC are contributed by GST so that
there was no confusion in matching up peaks with the N1SH3 HSQC for the shift
perturbation assay.
Figure 16. HSQC spectrum of GST (334 µM, pH 6.9, 48 scans) showing only 17 peaks.
The glutamine/asparagine side chain peaks are circled.

34
Figure 17 highlights the overlap of peaks between GST-N1SH3 and GST and it is
clear that only around seven of the peaks are shared by both spectra. This
implies that the rest of the peaks are from the N1SH3 domain which, as expected,
has shown up in the spectrum. Having separated the GST peaks from the N1SH3
peaks, it was possible to compare the GST-N1SH3 and N1SH3 spectra.
Figure 17. Overlay of GST-N1SH3 (Blue) (1.23 mM, pH 6.9, 24 scans) and GST
(Red) HSQC spectra showing peaks in GST-N1SH spectrum which are
contributed by GST. Arrows highlighting seven clear peak overlaps.

35
The fact that the peaks of GST-N1SH3 and N1SH3 overlap so well indicates that
the N1SH3 domain within the fusion is structurally similar to the free domain.
Therefore, any structural difference which may lead to differential binding of PD1
to GST-N1SH3 must be very subtle. A shift perturbation analysis was performed
on the assigned peaks in order to gain a quantitative insight into these structural
differences.
3.12 Shift Perturbation Assay and Structure Mapping
All residue numbers are quoted in terms of the cleaved N1SH3. Parallel residues
on GST-N1SH3 will all be numbered from 227 (residue G1 on N1SH3).
The formula used to calculate the combined chemical shift difference between the
peaks is shown below:
Figure 18. HSQC overlay of GST-N1SH3 (Blue) and N1SH3 (Red).

36
αN is the scaling factor for the difference in the 15
N chemical shift which is 0.2.
This is because the nitrogen chemical shift range in the HSQC is around five
times larger than that of hydrogen. The SH3 domain used for structural mapping
was the C-Src SH3 solution structure solved by the Shreiber lab (23).
Figure 19. (A) Graph of combined 15
N and 1
H chemical shift perturbations for each of the
peaks of the assigned residues including side chain peaks. Dashed blue line shows
calculated average chemical shift differences for assigned residues (0.0211). Red line
shows the summation of the average and standard deviation (0.395) which here
represents the line of significance for any peak shifts. V36 not included as no peak seen
in GST-N1SH3 spectrum. All shifted residues are backbone peaks. (B) Residues with
significant peak shifts mapped on to C-Src SH3 structure (PDB=1SRL) in PyMOL. Left
hand side shows surface representation and right hand side shows ribbon
representation.

37
As shown in figure 19 (A), there were nine significant peak shifts within the
assigned residues and potentially more within the unassigned residues.
Interestingly, these peaks are contained within two distinct clusters in the
sequence with the exception of S65. However, as is seen in the structure
mapping of these shifted residues, the two clusters are in spatial proximity in the
tertiary structure of the domain, as are S65 and V36. V36 has disappeared or
massively shifted in the GST-N1SH3 spectrum and so can be assumed to be
involved in some kind of structural alteration. G6 and G7 have not been mapped
onto the structure as these residues are not present in the truncated C-Src SH3
domain used for this mapping.

38
4.0 Discussion:
4.1 Differential PD1 binding to GST-N1SH3 and N1SH3
STD NMR is a highly sensitive technique which is able to test for binding between
proteins/domains and ligands, and works optimally for complexes with
dissociation constants (Kd) between 10-3
and 10-8
M. However, the lower the Kd
or the slower the koff, the smaller the peaks in the difference spectrum (24). This
technique was suggested for this study as it was hypothesised that PD1:GST-
N1SH3 may have a similar dissociation constant (around µM levels) to most short
linear consensus motif C-Src SH3 peptides (25).
The difference spectra from the STD experiments (Cha. 3.8) showed no evidence
of PD1 binding to N1SH3 but some limited evidence that it binds to GST-N1SH3.
Indeed all previous studies, such as the HSQC peptide titration, ITC and DSC,
have revealed no evidence of binding between N1SH3 and PD14
, further
supporting the results in this report. Although there are some peaks in the
difference spectrum of the GST-N1SH3 experiment, these are very small and only
visible after enlarging by a factor of 64. This suggests that either no binding
occurs and these peaks are artefacts or that PD1 binds with a much lower Kd
than expected (perhaps nM-pM). This could also be explained by an unusually
slow off rate (koff) (26). The slower exchange rate would therefore mean a lower
concentration of PD1 would receive 1
H saturation via Nuclear Overhauser spin
diffusion from the protein, as fewer peptide molecules would have associated and
disassociated with the protein. This would lead to the on-resonance spectrum
appearing much the same as the off-resonance, leading to smaller difference
spectrum peaks. There are other methods to verify high affinity ligand binding by
STD. These rely on competition binding experiments in which the ligand of
interest displaces another lower affinity ligand which binds the same protein
moiety (27). However, these experiments would require another GST-N1SH3
ligand of known low affinity which is currently not possible as no such ligand
exists. Within the current set-up however, the difference spectrum could be
4
Prof J Potts and Dr G Evans, personal communication.

39
improved and the size of the peaks increased by the addition of more PD1 or
performing longer irradiations, to counteract the low kd and koff respectively.
It may be that NMR based techniques are not ideal for these N1-SH3 studies and
other techniques may be worth attempting. Surface Plasmon Resonance (SPR),
with its very high sensitivity has the ability to detect ligand binding of complexes
with Kds in the pM range and so may be more appropriate. Fluorescence
anisotropy, another highly sensitive technique ideal for studying interactions
between molecules with large size differences, for example peptides and
domains, might also be promising. Both techniques have already been used to
measure binding of peptides to SH3 domains (28, 29)
4.2 Structural Alterations around the n-Src loop in GST-N1SH3
The GST-N1SH3 fusion and free N1SH3 appear to have very similar overlapping
HSQC spectra with the exception of a few GST contributed residues and a few
other peaks in the GST-N1SH3 spectrum. These peaks have very strong
intensities and are clustered between 8.5ppm and 7.5ppm, indicating potential
disorder, and are therefore likely to be linker residues. As seen in figure 18,
besides V36, most of the other shifts are extremely small. However, since this
experiment is examining potentially subtle structural alterations between N1SH3
and GST-N1SH3, they may still be significant. It must be remembered that the
structure of C-Src SH3 is different to N1-SH3 and so the predictive structure
mapping in figure 19 can only be used as a speculative guide to any structural
variations between GST-N1SH3 and N1SH3.
Interestingly, the clusters of shifted residues are all on the side of the domain
closest to the N-terminus and therefore the linker and GST. They are also all
clustered around the n-Src loop, particularly the β sheet which follows this loop.
Therefore it may be possible that the N1SH3 domain is in some way interacting
with the linker or GST in a region of SH3 which is likely to be involved in ligand
binding. The n-Src loop insert is the only difference between C-Src and N1-Src
and is enough to dramatically alter ligand binding and so this loop is likely to be in
the ligand binding epitope. Previous studies have shown binding contacts

40
between the 310-Helix, the n-Src loop of Src SH3 and nine residue peptide ligands
(13). Since the shifted S65 is located in the 310-helix and residues in spatial
proximity to the n-Src loop have shifted, this implies a potential disturbance in the
ligand binding epitope between N1SH3 and GST-N1SH3. S65 is the only residue
in the 310-helix assigned and none of the n-Src loop residues are assigned. This is
due to the presence of the six residue insert which is not present in the C-Src
SH3 HSQC spectrum used to assign this N1-Src SH3 spectrum. Residues
flanking the insert will also not be assigned, presumably due to the vast structural,
and therefore peak position, differences of these residues. As the assignments for
N1SH3 are based on comparisons to the C-Src SH3 HSQC, any major
differences will mean those residues cannot be reasonably assigned. This could
mean that more, even greater, peak perturbations exist in these 310-helix and n-
Src loop residues however they are unidentifiable as they have not been
assigned.
This highlights a key limitation in this study which is that only 55 of 88 peaks on
the N1SH3 spectrum have been assigned and only 36 of these ‘confidently’
(Appendix 6). In order to improve the reliability of similar future studies, it would
be worth carrying out triple resonance NMR on the N1SH3 domain in order to
unambiguously and sequentially assign residues. These experiments have been
carried out, however the assignments are not yet complete5
. Since the structural
mapping in figure 18 uses the C-Src SH3 domain and there are likely large
structural differences between this and the N1-Src SH3 domain (See HSQC
overlay in Appendix 7), a solution structure of N1SH3 would be extremely useful
in reliably mapping out shifted residues. It would also mean that a peptide titration
of PD1 with N1SH3 and GST-N1SH3 could be performed and yield information as
to the exact residues which are involved in PD1 binding. This was initially
intended to be carried out as part of this project, however was not possible due to
time constraints.
5

41
4.3 Conclusions
From the evidence available, a tentative conclusion can be drawn that there is
indeed a structural alteration in the peptide binding region of N1SH3 when bound
to GST. As PD1 appears only to bind to GST-N1SH3 and not N1SH3, it can be
assumed that the presence of GST is inducing this differential binding as the
peptide binding region is in close proximity to the linker and GST. This would
explain why PD1 does not bind cleaved N1SH3, but does not explain the
apparent biological activity of PD1 in the in vitro and cell based studies (Cha.1.5).
It may be that GST-N1SH3 represents the biologically relevant form of the SH3
domain in full length N1-Src. This may be due to an interaction between the N-
terminal SH4 domain/linker and the SH3 domain. Future studies could be
designed to test this hypothesis.
4.4 Future Studies
If the interactions of the SH3 domain are dependent on its proximity and/or
interactions with other domains within N1-Src, then it could be worth expressing
and purifying larger fragments of N1-Src for biochemical and biophysical studies,
e.g. the SH3 and SH4 domain. The ability of PD1 to bind this protein fragment
could then be tested. If NMR studies are to be carried out on these proteins and
their larger size begins to decrease the quality of their spectra, then it may be
necessary to produce deuterated proteins to improve the signal to noise ratio.
In addition to the suggested improvements to the specific experiments of this
project, there are numerous other possibilities for developing the future studies of
N1-Src and identifying its ligands and inter-protein interactions. Similar studies to
those already used for N1-Src SH3 involving peptide identification by phage
display could be performed. Larger fragments e.g. SH4-SH3 with smaller fusion
tags such as His-tags could be used. This could prevent any non-biologically
relevant structural perturbations of the protein induced by a bulky tag and so
reduce the likelihood of false positive peptide binding. An SH4 binding control
experiment would have to be performed when screening phage however, in order
to prevent selecting phage which bind to the SH4 domain only. A capture phage

42
ELISA method for validating peptide binding could also be optimised for these
studies to increase the reliability of any identified peptides (21). Other more
biologically relevant techniques for ligand identification such as phototrapping with
Tandem Affinity Purification (TAP) and mass spectrometry could be used to
isolate and identify in cellulo binding partners specific to the SH3 domain of N1-
Src (30). Attempts to use this method are already underway6
.
It is hoped that the results of this study will help inform future efforts to uncover
novel peptide or protein binding partners of the N1-Src SH3 domain. This study
may also represent the first attempt to structurally characterise the phenomenon
of GST induced false positives in phage display and prove a reminder of the need
for cautious analysis of phage display results.
Word Count: 7,999
Acknowledgements
I would like to thank Gemma Harris and all of the Potts team for their patience,
assistance and advice, both practically and analytically.
I would also like to thank Pedro Aguiar in the Department of Chemistry for his
help in acquiring all of my spectra.
6

43
5.0 References:
In the style of The Journal of Biological Chemistry, as formatted by Mendeley
Desktop.
1. Proto-oncogene tyrosine-protein kinase Src - SRC - Homo sapiens
(Human) [online] http://www.uniprot.org/uniprot/P12931 (Accessed March
26, 2014).
2. Superti-Furga, G., and Courtneidge, S. A. (1995) Structure-function
relationships in Src family and related protein tyrosine kinases. Bioessays
17, 321–30 [online] http://www.ncbi.nlm.nih.gov/pubmed/7537961
(Accessed November 11, 2013).
3. Irby, R. B., and Yeatman, T. J. (2000) Role of Src expression and activation
in human cancer. Oncogene 19, 5636–42 [online]
http://www.nature.com/onc/journal/v19/n49/full/1203912a.html (Accessed
November 11, 2013).
4. Foster-Barber, A., and Bishop, J. M. (1998) Src interacts with dynamin and
synapsin in neuronal cells. Proc. Natl. Acad. Sci. U. S. A. 95, 4673–7
[online]
http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=22549&tool=pmc
entrez&rendertype=abstract (Accessed February 16, 2014).
5. Keenan, S. (2012) Structure-function studies of the neuronal Src kinases.
Ph.D. thesis, The University of York.
6. Yadav, S. S., and Miller, W. T. (2007) Cooperative activation of Src family
kinases by SH3 and SH2 ligands. Cancer Lett. 257, 116–123 [online]
http://www.sciencedirect.com/science/article/pii/S0304383507003060
7. Marin, V., Groveman, B. R., Qiao, H., Xu, J., Ali, M. K., Fang, X.-Q., Lin, S.-
X., Rizkallah, R., Hurt, M. H., Bienkiewicz, E. A., and Yu, X.-M. (2010)
Characterization of neuronal Src kinase purified from a bacterial expression
system. Protein Expr. Purif. 74, 289–297 [online]
8. Matsunaga, T., Shirasawa, H., Tanabe, M., Ohnuma, N., Takahashi, H.,
and Simizu, B. (1993) Expression of alternatively spliced src messenger
RNAs related to neuronal differentiation in human neuroblastomas. Cancer
Res. 53, 3179–85 [online] http://www.ncbi.nlm.nih.gov/pubmed/8319227
(Accessed March 18, 2014).

44
9. Alberghina, L., Höfer, T., Vanoni, M., Carducci, M., Perfetto, L., Briganti, L.,
Paoluzi, S., Costa, S., Zerweck, J., Schutkowski, M., Castagnoli, L., and
Cesareni, G. (2012) The protein interaction network mediated by human
SH3 domains. Biotechnol. Adv. 30, 4–15 [online]
(Accessed October 9, 2013).
10. Kaneko, T., Li, L., and Li, S. S.-C. (2008) The SH3 domain--a family of
versatile peptide- and protein-recognition module. Front. Biosci. 13, 4938–
52 [online] http://www.ncbi.nlm.nih.gov/pubmed/18508559 (Accessed
November 11, 2013).
11. Dergai, M., Tsyba, L., Dergai, O., Zlatskii, I., Skrypkina, I., Kovalenko, V.,
and Rynditch, A. (2010) Microexon-based regulation of ITSN1 and Src SH3
domains specificity relies on introduction of charged amino acids into the
interaction interface. Biochem. Biophys. Res. Commun. 399, 307–12
[online] http://www.ncbi.nlm.nih.gov/pubmed/20659428 (Accessed October
8, 2013).
12. Sudol, M., Cesareni, G., Superti-Furga, G., Just, W., Saksela, K., and
Permi, P. (2012) SH3 domain ligand binding: What’s the consensus and
where’s the specificity? FEBS Lett. 586, 2609–2614 [online]
13. Feng, S., Chen, J. K., Yu, H., Simon, J. A., and Schreiber, S. L. (1994) Two
binding orientations for peptides to the Src SH3 domain: development of a
general model for SH3-ligand interactions. Science 266, 1241–7 [online]
http://www.ncbi.nlm.nih.gov/pubmed/7526465 (Accessed March 29, 2014).
14. Gunasekara, N., Sykes, B., and Hugh, J. (2012) Characterization of a novel
weak interaction between MUC1 and Src-SH3 using nuclear magnetic
resonance spectroscopy, [online]
http://www.sciencedirect.com/science/article/pii/S0006291X12007802
15. Xiao, R., Xi, X.-D., Chen, Z., Chen, S.-J., and Meng, G. (2013) Structural
framework of c-Src activation by integrin β3. Blood 121, 700–6 [online]
http://bloodjournal.hematologylibrary.org/content/121/4/700.long (Accessed
November 11, 2013).
16. Rickles, R. J., Botfield, M. C., Zhou, X. M., Henry, P. A., Brugge, J. S., and
Zoller, M. J. (1995) Phage display selection of ligand residues important for
Src homology 3 domain binding specificity. Proc. Natl. Acad. Sci. U. S. A.
92, 10909–13 [online]
http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=40540&tool=pmc
entrez&rendertype=abstract (Accessed February 11, 2014).
17. Zhou, P., and Wagner, G. (2010) Overcoming the solubility limit with
solubility-enhancement tags: successful applications in biomolecular NMR

45
studies. J. Biomol. NMR 46, 23–31 [online]
http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2879018&tool=p
mcentrez&rendertype=abstract (Accessed February 23, 2014).
18. Sparks, A. B., Rider, J. E., and Kay, B. K. (1998) Mapping the specificity of
SH3 domains with phage-displayed random-peptide libraries. Methods Mol.
Biol. 84, 87–103 [online] http://www.ncbi.nlm.nih.gov/pubmed/9666443
(Accessed February 26, 2014).
19. Liew, C. K., Gamsjaeger, R., Mansfield, R. E., and Mackay, J. P. (2008)
NMR spectroscopy as a tool for the rapid assessment of the conformation
of GST-fusion proteins. Protein Sci. 17, 1630–5 [online]
mcentrez&rendertype=abstract (Accessed March 5, 2014).
20. Murthy, K. K., Ekiel, I., Shen, S. H., and Banville, D. (1999) Fusion proteins
could generate false positives in peptide phage display. Biotechniques 26,
142–9 [online] http://www.ncbi.nlm.nih.gov/pubmed/9894603 (Accessed
February 11, 2014).
21. Zhang, M.-Y., Schillberg, S., Zimmermann, S., Liao, Y.-C., Breuer, G., and
Fischer, R. (2001) GST fusion proteins cause false positives during
selection of viral movement protein specific single chain antibodies. J. Virol.
Methods 91, 139–147 [online]
(Accessed October 9, 2013).
22. Yu, H., Rosen, M. K., and Schreiber, S. L. (1993) 1H and 15N assignments
and secondary structure of the Src SH3 domain. FEBS Lett. 324, 87–92
[online] http://www.ncbi.nlm.nih.gov/pubmed/8504863 (Accessed February
11, 2014).
23. Yu, H., Rosen, M., Shin, T., Seidel-Dugan, C., Brugge, J., and Schreiber, S.
(1992) Solution structure of the SH3 domain of Src and identification of its
ligand-binding site. Science (80-. ). 258, 1665–1668 [online]
http://www.sciencemag.org/content/258/5088/1665 (Accessed February 26,
2014).
24. Viegas, A., Manso, J., Nobrega, F. L., and Cabrita, E. J. (2011) Saturation-
Transfer Difference (STD) NMR: A Simple and Fast Method for Ligand
Screening and Characterization of Protein Binding. J. Chem. Educ. 88,
990–994 [online] http://dx.doi.org/10.1021/ed101169t (Accessed March 28,
2014).
25. Yu, H. (1994) Structural basis for the binding of proline-rich peptides to SH3
domains. Cell 76, 933–945 [online]
http://www.sciencedirect.com/science/article/pii/0092867494903670

46
26. Pérez-Victoria, I., Kemper, S., Patel, M. K., Edwards, J. M., Errey, J. C.,
Primavesi, L. F., Paul, M. J., Claridge, T. D. W., and Davis, B. G. (2009)
Saturation transfer difference NMR reveals functionally essential kinetic
differences for a sugar-binding repressor protein. Chem. Commun.
(Camb)., 5862–4 [online] http://www.ncbi.nlm.nih.gov/pubmed/19787122
(Accessed April 7, 2014).
27. Wang, Y.-S., Liu, D., and Wyss, D. F. (2004) Competition STD NMR for the
detection of high-affinity ligands and NMR-based screening. Magn. Reson.
Chem. 42, 485–9 [online] http://www.ncbi.nlm.nih.gov/pubmed/15137040
28. Lemmon, M. A., Ladbury, J. E., Mandiyan, V., Zhou, M., and Schlessinger,
J. (1994) Independent binding of peptide ligands to the SH2 and SH3
domains of Grb2. J. Biol. Chem. 269, 31653–8 [online]
29. Ferguson, M. R., Fan, X., Mukherjee, M., Luo, J., Khan, R., Ferreon, J. C.,
Hilser, V. J., Shope, R. E., and Fox, R. O. (2004) Directed discovery of
bivalent peptide ligands to an SH3 domain. Protein Sci. 13, 626–32 [online]
mcentrez&rendertype=abstract (Accessed March 29, 2014).
30. Okada, H., Uezu, A., Mason, F. M., Soderblom, E. J., Moseley, M. A., and
Soderling, S. H. (2011) SH3 domain-based phototrapping in living cells
reveals Rho family GAP signaling complexes. Sci. Signal. 4, rs13 [online]
mcentrez&rendertype=abstract (Accessed November 11, 2013).
31. Raghunathan, S., Chandross, R. J., Kretsinger, R. H., Allison, T. J.,
Penington, C. J., and Rule, G. S. (1994) Crystal structure of human class
mu glutathione transferase GSTM2-2. Effects of lattice packing on
conformational heterogeneity. J. Mol. Biol. 238, 815–32 [online]

47
6.0 Abbreviations
CFP Cyan fluorescent protein
DSC Differential scanning calorimetry
DTT Dithiothreitol
ELISA Enzyme-linked immunosorbent assay
GST-N1SH3 GST tagged N1SH3 domain
HRV 3C Protease Human rhinovirus 3C protease
HSQC Heteronuclear single quantum coherence
IPTG Isopropyl β-D-1 thiogalactopyranoside
ITC Isothermal titration calorimetry
LB Luria Broth
MES 2-(N-morpholino)ethanesulfonic acid
MW Molecular Weight
N1SH3 N1-Src SH3 domain
NMR Nuclear magnetic resonance spectroscopy
PBS Phosphate buffered saline
ppm Parts per million
ScFv Single-chain variable fragment
SDS-PAGE Sodium dodecyl sulfate – polyacrylamide gel
electrophoresis
SE Size exclusion
SH3 Src homology 3 domain
STD Saturation transfer difference

48
7.0 Appendices
7.1 Appendix 1 – Plasmid map
GST-N1SH3 cloned into pGEX6P-1 using 5’BamHI and 3’SalI restriction sites and
the following primers:
Sense: CCG CGT GGA TCC GGT GGG GTG ACT ACC TTT GTG GCC
Anti-sense: CAC AGC GTC GAC TCA CTC CTC AGC CTG GAT GGA GTC GAA
Appendix 1. Plasmid map of pGEX6P-1 with the GST-N1SH3 insert. Map
generated by PlasMapper 2.0 (http://wishart.biology.ualberta.ca/PlasMapper/).

49
7.2 Appendix 2 - Amino Acid Sequences and Protein Data (Calculated via
ExPASy ProtParam):
GST-N1SH3:
M S P I L G Y W K I K G L V Q P T R L L L E Y L E E K Y E E H L Y E R D E G D
K W R N K K F E L G L E F P N L P Y Y I D G D V K L T Q S M A I I R Y I A D K
H N M L G G C P K E R A E I S M L E G A V L D I R Y G V S R I A Y S K D F E T
L K V D F L S K L P E M L K M F E D R L C H K T Y L N G D H V T H P D F M L
Y D A L D V V L Y M D P M C L D A F P K L V C F K K R I E A I P Q I D K Y L K
S S K Y I A W P L Q G W Q A T F G G G D H P P K S D L E V L F Q /// G P L G
S G G V T T F V A L Y D Y E S R T E T D L S F K K G E R L Q I V N N T R K V
D V R E G D W W L A H S L S T G Q T G Y I P S N Y V A P S D S I Q A E E
Stop
Yellow: Linker.
///: 3C protease cleavage site.
Blue: 6 residue n-Src loop insert.
Number of amino acids: 304
Molecular weight: 35,023.1 Da
Theoretical pI: 5.38
Total number of negatively charged residues (Asp + Glu): 46
Total number of positively charged residues (Arg + Lys): 37
Extinction coefficients (280nm): 60,070 M-1
cm-1
GST:
M S P I L G Y W K I K G L V Q P T R L L L E Y L E E K Y E E H L Y E R D E G D
K W R N K K F E L G L E F P N L P Y Y I D G D V K L T Q S M A I I R Y I A D K
H N M L G G C P K E R A E I S M L E G A V L D I R Y G V S R I A Y S K D F E T
L K V D F L S K L P E M L K M F E D R L C H K T Y L N G D H V T H P D F M L

50
Y D A L D V V L Y M D P M C L D A F P K L V C F K K R I E A I P Q I D K Y L K
S S K Y I A W P L Q G W Q A T F G G G D H P P K S D L E V L F Q
cm-1
N1-SH3 domain:
G P L G S G G V T T F V A L Y D Y E S R T E T D L S F K K G E R L Q I V N N
T R K V D V R E G D W W L A H S L S T G Q T G Y I P S N Y V A P S D S I Q A
E E
cm-1
7.3 Appendix 3 - Media Recipes:
Constituent g/L
Tryptone Granulated (Melford Laboratories Ltd) 10
Yeast Extract Microgranulated (FormediumTM
) 5
NaCl (Fisher Scientific) 10
Table 1. Luria Broth (LB) recipe. Autoclaved and cooled prior to inoculation.

51
Salts (Fisher Scientific)
Na2HPO4 6 g/L
KH2PO4 3 g/L
NaCl 0.5 g/L
Trace Metals [1000x] (made up to 100mls with H20) 1ml in 1L M9
0.1M FeCl3 (in 0.1M HCl) 50ml
1M CaCl2.2H20 2ml
1M MnCl2.4H20 1ml
1M ZnSO4.7H20 1ml
0.2M CoCl2.6H20 1ml
0.1M CuCl2 2ml
0.2M NiCl2.6H20 1ml
0.1M Na2MoO4.2H20 2ml
0.1M NaSeO3.5H20 2ml
0.1M H3BO3 2ml
Vitamins [1000x] (dissolved in 100mls of H20) 1ml in 1L M9
Riboflavin 0.1g
Nicotinamide 0.1g
Pyridoxine 0.1g
Thiamine 0.1g
Other
1M MgSO4 2ml
1M CaCl2 0.2ml
15
NH4Cl (Cambridge Isotope Laboratories Inc) 1g/L
20% (w/v) D-Glucose (in H20) 20ml
100mg/ml Ampicillin (Melford Labs Ltd) 1ml
Table 2. M9 minimal media recipe. Salts added prior to autoclaving. All other
components added just prior to inoculation.

52
7.4 Appendix 4 - Buffer Recipes:
 All buffers and solutions filtered either with Whatman Nylon membrane
filters (0.2µm) using vacuum pump or using syringes (BD Plastipak) and
0.2µm filters (Merck Millipore).
 All dilutions made in and all buffer components dissolved in Elga PURLAB
Ultra filtered H20 unless otherwise stated.
Constituent Concentration (Molarity)
Na2HPO4 10mM
KH2PO4 1.8mM
NaCl 140mM
KCl 2.7mM
Constituent Concentration (mg/ml)
Benzamide HCl 1.6
Leupeptin 1
Pepstatin A
Aprotinin
1
1
Constituent Per litre
1M NaH2PO4.H20/ K2HPO4 (pH 6.5) 50mls
NaCl 5.84g (100mM)
Table 3. PBS/Binding Buffer recipe pH 7.3. All compounds supplied by Fisher
Scientific. GE Healthcare Amersham recipe.
Table 4. 1000X protease inhibitor cocktail recipe. Dissolved in 100% Ethanol.
Table 5. Size Exclusion Buffer recipe (pH 6.5).

53
7.5 Appendix 5 – SDS-PAGE Reagents
7.6 Appendix 6 – N1-Src SH3 Assigned Residues
C-Src Residue No. Amino Acid
N-Src
Residue No. Tentative/Confident
5 G 6 Con
6 G 7 Con
8 T 9 Con
9 T 10 Con
10 F 11 Con
11 V 12 Con
12 A 13 Con
13 L 14 Con
14 Y 15 Tent
15 D 16 Con
16 Y 17 Con
17 E 18 Con
18 S 19 Con
20 T 21 Tent
21 E 22 Tent
Constituent In 14mls
Glycerol 12g
H20 3ml
10% SDS
1M Tris pH 7.2
10ml
1ml
Bromophenol Blue 0.06g
Constituent g/L
MES [2-(N-morpholino)ethanesulfonic acid] (Melford Laboratories) 97.6
Tris [tris(hydroxymethyl)aminomethane] (Sigma) 60.6
SDS [Sodium dodecyl sulphate] (Sigma) 10
EDTA [Ethylenediaminetetraacetic acid] (Sigma) 3
Table 6. Recipe for 4X sample buffer. 700µl of 4X sample buffer added to 300µl of 1M
DTT before use.
Table 7. 20X MES recipe.

54
22 T 23 Con
23 D 24 Tent
24 L 25 Tent
25 S 26 Con
26 F 27 Con
27 K 28 Tent
28 K 29 Tent
29 G 30 Con
30 E 31 Tent
31 R 32 Tent
32 L 33 Tent
33 Q 34 Tent
33 Q-Side Chain 34 Tent
34 I 35 Tent
35 V 36 Con
36 N-Side Chain 37 Con
36 N-Side Chain 37 Con
37 N 38 Con
38 T 39 Con
42 W-Side Chain 49 Con
42 W 49 Tent
43 W-Side Chain 50 Con
43 W 50 Tent
44 L 51 Con
45 A 52 Con
46 H 53 Con
48 L 55 Con
50 T 57 Con
52 Q 59 Con
53 T 60 Con
54 G 61 Con
55 Y 62 Tent
56 I 63 Con
58 S 65 Con
61 V 68 Con
62 A 69 Con
C terminus C-terminus 78 Con

55
7.7 Appendix 7 - Overlay of HSQC spectra of C-Src SH3 (Red) and N1-Src
SH3 (Blue)

Submitted Report Final Draft

Recommended

Recommended

More Related Content

Viewers also liked

Viewers also liked (11)

Similar to Submitted Report Final Draft

Similar to Submitted Report Final Draft (20)

Submitted Report Final Draft