Citation
Quantifying the association between phosphatidic acid and SRC

Material Information

Title:
Quantifying the association between phosphatidic acid and SRC
Creator:
Yang, Pengsue
Publication Date:
Language:
English
Physical Description:
x, 103 leaves : ; 28 cm

Subjects

Subjects / Keywords:
Phosphatidic acids ( lcsh )
Fertilization (Biology) ( lcsh )
Protein-tyrosine kinase ( lcsh )
Proto-oncogenes ( lcsh )
Fertilization (Biology) ( fast )
Phosphatidic acids ( fast )
Protein-tyrosine kinase ( fast )
Proto-oncogenes ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 96-103).
General Note:
Department of Integrative Biology
Statement of Responsibility:
by Pengsue Yang.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
710983166 ( OCLC )
ocn710983166
Classification:
LD1193.L45 2010m Y36 ( lcc )

Full Text
QUANTIFYING THE ASSOCIATION BETWEEN PHOSPHATIDIC ACID AND SRC
By
Pengsue Yang
B.S., University of California San Diego
A thesis submitted to the
University of Colorado Denver
in partial fulfillment
of the requirements of the degree of
Master of Biology
2010


This thesis for the Masters of Biology
degree by
Pengsue Yang
has been approved
by
ScZ
Date


Yang, Pengsue [M.S., Biology)
Quantifying the Association Betweeen Phosphatidic Acid and Src
Thesis Directed by Professor Bradley J. Stith
ABSTRACT
The exact molecular mechanism of fertilization is still unknown, but our
current model suggests that sperm elevate phosphatidic acid (PA) and that this
lipid binds to and activates Src. This would lead to the activation of
phospholipase C-gamma, the release of inositol 1,4,5-triphosphate and
intracellular calcium in the zygote, and subsequent fertilization events. Using
PIP strips, Xenopus Src was found to bind phosphatidic acid with specificity.
This current work estimates the dissociation constant (Kd) of human Src binding
to phosphatidic acid and compares it to the affinity of Src and a control anionic
lipid. We also compare the affinity of Src binding to phosphatidic acid with that
of Bovine Serum Albumin (BSA) binding to phosphatidic acid. We tested two
different vesicle sedimentation assays where we varied protein concentrations
or varied liposomal concentrations. Protein concentrations were recorded using
280nm absorbance or the fluorescent CBQCA Protein Quantification Kit
(Molecular Probes Inc., Carlsbad, CA). The protein quantification kit utilizes the
ATTO-TAG CBQCA reagent that, in the presence of cyanide, reacts with primary
amines to form fluorescent derivatives, which are quantified using a


fluorometer. The tag required the addition of the KCl-Protein Release
modification. We found that the preferred method of affinity analysis was the
vesicle sedimentation assay (McLaughlin, 1998) where the binding of Src
protein to large unilamellar vesicles (LUVs) was quantified at different liposomal
concentrations. The large unilamellar vesicles were comprised of
phosphatidylcholine, phosphatidylcholine plus phosphatidic acid or
phosphatidylcholine with a control anionic lipid. The lipid vesicles were created
using the Avanti Extruder (Avanti Polar Lipids, Inc. Alabaster, AL). Src was
found to bind with a higher affinity to vesicles consisting of phosphatidic acid
and phosphatidylcholine (Kd = 0,009 M) than to vesicles containing just
phosphatidylcholine (Kd = 0.018 M) or phosphatidylcholine plus
lysophosphatidic acid (Kd = 0.023 M). In addition, we were able to show that Src
bound with a higher affinity than BSA in all experimental sets. Our data
supports our hypothesis that the increase in PA observed during fertilization
may be binding Src.
This abstract accurately represents the content of the candidates thesis. I
recommend its publication.
Signed_______________________________
Bradley J. Stith


DEDICATION
I dedicate this thesis to my loving parents, who taught me the value of
patience, perseverance and resolve. I also dedicate this to my siblings, for
their love and support. Finally, I dedicate this to Cassandra, for her
friendship and understanding while I was completing this thesis.


ACKNOWLEDGEMENT
My thanks to my advisor, Dr. Bradley J. Stith, for his contribution to my
research. 1 also wish to thank the members of my committee, Dr. Aimee
Bernard and Dr. Amanda Charlesworth, for their valuable participation and
insights. In addition, 1 would like to thank Dr. Michael Wunder for his help in
analyzing my data. Special thanks to Ryan Bates, Josh Snyder and Colby Fees
for their contribution to my research.


TABLE OF CONTENTS
LIST OF FIGURES....................................................ix
LIST OF TABLES......................................................x
1. INTRODUCTION.....................................................1
1.1 Background...................................................1
1.2 Role of Observed Increase in Calcium Concentrations..............5
1.3 Fertilization Causes the Observed Calcium Wave...................6
1.4 Src is Involved in PLC-y Activation..............................7
1.5 Phosphatidic Acid activates Src..............................8
1.6 Phosphatidic Acid...........................................11
1.7 Src Tyrosine Kinase.........................................13
1.8 Observed Increases in PA and Src............................15
1.9 Association Between PA and Src..............................15
1.10 Removal of N-Term Basic Amino Acids Lower Src-Lipid Affinity....16
1.11 The Affinity Between Src and Phosphatidic Acid..................21
2. MATERIALS AND METHODS...........................................22
2.1 Preparing Sucrose-loaded Liposomes..........................22
2.2 Avanti Mini Extruder Assembly...............................25
2.3 Liposome Protein Binding Protocol: Varying Protein Concentration.27
2.4 Liposome-Protein Binding Protocol: Varying Protein Concentration with
KC1 Protein Release..............................................31
2.5 Liposome-Protein Binding Protocol: Varying Liposome Concentration
with KC1 Protein Release.........................................32
2.6 Testing Protein Quantification Using 280nm Spectrophotometry.....35
2.7 Testing Protein Quantification Using ATTO-Tag CBQCA.........37
2.8 Affinity Constants..........................................38
2.9 Statistical Analysis........................................41
3. RESULTS.........................................................43
3.1 Protein Detection Using the Absorbance Assay (280nm)........43
3.2 Testing the ATTO-TAG CBQCA..................................44
3.3 Preliminary Protein-Liposome Binding Data...................46
3.4 Protein-Liposome Binding Data with KCl-Protein Release......48
3.5 Liposome-BSA Binding........................................57
3.6 Summary of BSA Binding Kds.................................61
3.7 Liposome-Polylysine Binding.................................63
3.8 Liposome-Src Binding........................................66
3.9 Summary of Binding to Various Liposomes.....................69
vii


4. DISCUSSION.....................................................72
4.1 280nm Absorbance Assay......................................72
4.2 ATTO-Tag CBQCA..............................................72
4.3 Effects of KC1 on Releasing Protein From Liposomes.........73
4.4 Comparing Varied Protein Concentration Protocol to Varied Liposomal
Concentration Protocol..........................................75
4.5 Polylysine Binding..........................................76
4.6 Calculating the Kd of Protein Binding.......................79
4.7 Compare Vesicle Sedimentation Results to PIP Strip Results..80
4.8 Physiological Relevance of PA and Src.......................81
4.9 Conclusion..................................................85
Appendix A.................................................... 88
Appendix B........................................................89
Appendix C........................................................91
Appendix D........................................................93
Appendix E........................................................95
Bibliography......................................................96
viii


LIST OF FIGURES
Figure 1.1.1 Xenopus Laevis Fertilization Pathway......................4
Figure 1.5.1 PLDlb Cleavage of Phosphatidylcholine....................10
Figure 1.6.1 Phosphatidic Acid........................................11
Figure 1.7.1 Domain Structure of Src Protein..........................14
Figure 1.10.1 Comparing the N terminus of Xenopus and Human Src..........17
Figure 1.10.2 Mutated Basic Amino Acid Residues in Src Lowers Binding
Affinity to Liposomes Containing PS...................................20
Figure 2.1.1 Protocol for Preparing Sucrose Loaded Liposomes..........24
Figure 2.2.1 Avanti Mini-Extruder Set-up..............................26
Figure 2.2.2 Extrusion Technique......................................27
Figure 2.3.1 Liposome-Protein Binding Protocol: Varying Protein
Concentration.........................................................29
Figure 2.3.2 Beckman Airfuge: Relating Pressure to Centrifugation Speeds.30
Figure 2.4.1 Liposome-Protein Binding Protocol: Varying Protein
Concentration with KC1 Protein Release................................32
Figure 2.5.1 Liposome-Protein Binding Protocol: Varying Liposome
Concentration.........................................................35
Figure 2.7.1 CBQCA Reaction...........................................38
Figure 3.1.1 BSA Concentration vs Calculated BSA Concentration Using 280nm
Absorbance Assay......................................................44
Figure 3.2.1 Testing the ATTO-TAG CBQCA using BSA.....................45
Figure 3.3.1 Preliminary Protein-Liposome Binding Data................48
Figure 3.4.1 BSA Binding to PC/PA Liposomes...........................50
Figure 3.4.2 BSA Binding to PC Liposomes..............................52
Figure 3.4.3 BSA Binding to PC/LPA Liposomes..........................55
Figure 3.4.4 Comparing BSA-Liposome Binding...........................56
Figure 3.5.1 BSA Binding to Liposomes.................................58
Figure 3.6.1 Relating Confidence Intervals with P-values..............95
Figure 3.6.2 Estimated Kds of BSA and Liposomes with 95% Cl..............63
Figure 3.7.1 Polylysine Binding.......................................65
Figure 3.8.1 Src Binding to Liposomes.................................67
Figure 3.8.2 Src Binding to 4:1 Liposomes.............................68
Figure 3.9.1 Estimated Kd's of BSA and Liposomes with 95% Confidence
Interval..............................................................71
Figure E.l Relating Confidence Intervals with P-values................95
IX


LIST OF TABLES
Table 1.10.1 Calculated Kds of c-Src Binding to PC:PS Liposomes..........19
Table 3.4.1 Fluorometer Readouts for BSA Binding Assay Using PC/PS
Liposomes................................................................54
Table 3.6.1 Relative Binding of Control Protein BSA to Various Liposomes.61
Table 3.9.1 Estimated Kds of Src, BSA and Polylysine.....................69
x


1. INTRODUCTION
1.1 Background
At fertilization, the fusion of sperm and egg initiates a number of
cellular and biochemical events that give rise to the development of the
embryo. These events include the acrosome reaction, the acceleration of
metabolic processes, and the cortical reaction (cortical granule exocytosis).
A glycoprotein-rich extracellular layer called the vitelline envelope (Xenopus
Laevis) or the zona pellucida (mammals) perform multiple functions during
fertilization. For example, the vitelline envelope plays a role in the induction
of the sperm acrosome reaction, prevents the penetration of sperm from
other species, and hardens after fertilization to prevent multiple sperm entry
(polyspermy) (Tian et al., 1997).
Sperm normally carry proteolytic enzymes, along with other enzymes,
in a large exocytotic vesicle in the head region of the sperm, called the
acrosome. On contact with the egg jelly of the egg, a process called the
acrosome reaction occurs in which the contents of the vesicle are exocytosed.
The enzymes help degrade the egg jelly, zona pellucida, or vitelline envelope
so sperm can reach the egg plasma membrane (Austin and Bishop, 1958).
Fusion between the sperm and egg membranes activate two major blocks to
1


polyspermy: the fast block, which is due to a change in the plasma membrane
potential (depolarized plasma membrane will not bind sperm), and the slow
block, which is initiated by the intracellular release of calcium (Ca2+). The
slow block involves destruction of sperm receptors. Slow block to
polyspermy is due to the release of cortical granules. The contents of the
cortical granule modifies the vitelline envelope and forms a tough
fertilization membrane, which lifts off the egg surface and is very resistant to
further sperm entry.
To date, the egg of all species studied exhibit a transient rise in
intracellular free calcium at sperm-egg fusion (Jaffe, 1983; Wittingham, 1980;
Whitaker and Steinhardt, 1985). The increase in intracellular calcium has
been observed to play an important role in the induction of the fertilization
potential (Chambers and De Armendi, 1979; Obata and Kuroda, 1987),
exocytosis of cortical granules, a process known as the cortical reaction
(Kline and Kline, 1992) and increasing metabolic activity (Whitaker and
Steinhardt, 1985).
The mechanism that brings about the rise in intracellular free calcium
is not fully understood, but recent studies have shown that this process
requires the production of inositol 1,4,5-triphosphate (IP3), a second
2


messenger for Ca2+ release (Miyazaki et al., 1993; Sato et al., 2006). The
production of IP3 is carefully regulated and involves phospholipase C (PLC)
and the tyrosine kinase, Src. The current working hypothesis in frog models
is that the sperm somehow triggers the activation of Src tyrosine kinase,
which directly activates the gamma (y) isoform of PLC. Activation of PLC-y
produces IP3 and sn-1,2 diacylglycerol (DAG) from phosphatidylinositol 4,5-
bisphosphate (PIP2) (Figure 1.1.1). Much work by Stith and associates
suggest that sperm binding activates phospholipase Dlb (PLDlb), which
hydrolyzes phosphatidylcholine (PC) to form phosphatidic acid (PA) and
choline. PA binds and activates Src tyrosine kinase which goes on to activate
PLC-y. As noted, activated PLC-y leads to IP3 production, which in turn
releases Ca2+ through IP3 receptors in the endoplasmic reticulum (ER) of the
egg-
3


Xenopus laevis Fertilization Pathway
1.Sperm binding
2. ? activates Src
3. Src activatesPLC
4. PLC cleaves PIP2
into DAG and IP3
5.IP3 translocates to
ER IP3 receptor
6. Ca2+ is released into
cytplasm
7. Ca2+ binds and
activates CACCs
8. CI- leaves cell causing
voltage change and
fast block to polyspermy
2-aminoethoxydiphenylborate
2-APB
Figure 1.1.1 Xenopus Laevis Fertilization Pathway
The molecular pathway during fertilization in Xenopus Laevis. It is proposed
that sperm binding to the egg (1) activates phospholipase Dlb, which cleaves
phophatidylcholine to choline and phosphatidic acid. Phosphatidic acid goes
on to activate Src (2), which goes on to activate PLC (3). PLC then cleaves
PIP2 into DAG and IP3 (4). IP3 then translocates to receptors on the
endoplasmic reticulum (5). This binding releases calcium into the cytoplasm
(6). Image created by Shilo Smith and Joshua Snyder (University of Colorado,
Denver].
4


Previous experimental data has led us to hypothesize that
phosphatidic acid, an acidic phospholipid of the cell bilayer, is responsible for
activating Src tyrosine kinase directly. In order to determine whether our
hypothesis is physiologically plausible and whether its association is
significant, we quantified the association between Src and PA.
1.2 Role of Observed Increase in Calcium Concentrations
The increase in calcium concentration during fertilization is known to
play an important role in egg activation (Jaffe, 1983; Wittingham, 1980;
Whitaker and Steinhardt, 1985). The normal concentration of intracellular
calcium in the eggs of most species is usually about 50 200 nM (Taylor et
al., 1993). During egg activation, the concentration of calcium can be seen to
increase approximately 10-fold to 1 2 pM (Taylor et al., 1993; Nuccitelli et
al., 1993). The increase in [Ca] has been observed to occur as a wave that
sweeps across the egg that begins at the site of sperm-egg fusion (Busa and
Nuccitelli, 1985).
There are three main functions of egg activation: block to polyspermy,
activation of egg metabolism, and resumption of the cell cycle. The increase
in calcium ion levels also causes an increase in pH by activating the Na+/H+
5


exchange pump. In sea urchins, the increase in pH may stimulate new
protein synthesis and DNA synthesis (Winkler et al., 1980; Whitaker and
Steinhardt, 1982; Rees et al., 1995). Finally, elevated calcium induces a
degradation of cyclin that leads to resumption of the cell cycle (Whitaker,
2008).
1.3 Fertilization Causes the Observed Calcium Wave
The signaling pathway that brings about the rise in intracellular free
calcium is not fully understood. Studies suggest that inositol phospholipid
turnover and intracellular stores are responsible for egg activation (Miyazaki
et al., 1993). In this pathway, phopholipase C-gamma (PLC-y), a plasma
membrane associated enzyme, is believed to be activated at fertilization to
catalyze the cleavage of phosphatidylinositol-4-5-bisphosphate (PIP2), to
diacylglycerol (DAG) and inositol triphosphate (IPs). The water-soluble IP3
molecule diffuses away from the membrane, and bind and open ligand-gated
calcium channels on the smooth endoplasmic reticulum to mediate calcium
release. DAG remains within the plasma membrane and possibly activates
protein kinase C (PKC), (Sato et al., 1999) and this may lead to other
fertilization events.
There are three main isoforms of PLC: p, y, and 5. A study conducted
6


by Shen et al. (1999), using sea urchin eggs, suggested that the y isoform is
responsible for egg activation in lower organisms (not mammals). Although
it has been shown that PLC-y can be activated by tyrosine phosphorylation
(Nishibe et al., 1990), the regulation of it is not well understood. Using
immunoprecipitation, Sato et al. (2000) showed that Xenopus egg PLC-y is
tyrosine phosphorylated and that this occurs at two minutes after
fertilization.
1.4 Src is Involved in PLC-y Activation
Recent work by Sato et al. (1999) suggests that Src, a tyrosine kinase,
has a major role in Xenopus fertilization. Using mass spectrophotometry and
Western blotting, Sato et al. (1996) were able to show that Src is expressed in
Xenopus eggs and that the addition of sperm to Xenopus eggs promotes Src-
dependent phosphorylation of PLC-y, IP3 production, and calcium release.
Other studies in sea urchin have demonstrated that calcium release can be
blocked by a PLC-specific inhibitor (U-73122), an IP3 receptor antagonist
(heparin) (Stith et al., 1994), and various Src inhibitors (Sato et al., 1999;
Glahn et al., 1999). A fertilization-induced colocalization of PLC-y and a Src
protein-tyrosine kinase was also observed (Sato et al., 2000). When the
7


tyrosine kinase was microinjected into the egg, activation of tyrosine
phosphorylation of PLC-y increased. However, the addition of a Src-specific
protein tyrosine kinase inhibitor, PP1, effectively blocked the PLC-y-Src
association. Blockage of the PLC-y-Src association inhibited sperm-induced
IP3 production and calcium release, as well as all calcium-dependent events
such as cortical contraction and the elevation of the fertilization envelope.
The use of a PLC-y inhibitor, U-73122, produced the same inhibition. The
biochemical steps between sperm binding and Src activation however, has
not been well studied.
1.5 Phosphatidic Acid activates Src
As noted, much data suggest that PLC-y plays an important role in
Xenopus fertilization; however, the regulation of it is not well understood. A
study showed that PLC activity was affected by the amount of acidic lipid
present (Rebecchi et al., 1992). This finding led to the hypothesis that
perhaps an acidic lipid helps regulate this pathway. At fertilization, in
Xenopus, there is an increase for inositol triphosphate as well as a larger, late
increase in DAG (Stith et al., 1993; 1997). From these and other data, they
suggest that activation of phospholipase D (PLD) and the subsequent
8


production and degradation of phosphatidic acid is responsible for the late
DAG increase (Petcoff et al., 2008). It is known that phospholipase D can
hydrolyze phosphatidylcholine (PC) to phosphatidic acid (PA) and choline
(Exton, 1994) (Figure 1.5.1), and that PA can be converted to DAG via the
action of phosphatidic acid phosphohydrolase (Carman and Han, 2006).
Using an enzymatic method, Stith et al. (1997) found that there is an increase
in the mass of choline very early after insemination. In addition, using high
performance liquid chromatography and evaporative light scattering
detection, the mass of phosphatidic acid has also been observed to increase
after insemination (Stith, unpublished data). Phosphatidic acid is believed to
facilitate membrane fusion (Zeniou-Meyer et al., 2007) and directly or
indirectly activate phospholipase C (Jones and Carpenter, 1993), making it a
strong candidate for an activator of intracellular calcium release, sperm-egg
fusion or cortical granule merger with the plasma membrane during
fertilization.
It has been suggested PA has the ability to bind to proteins (Hong et
al., 2009) and Src binds to cell membranes (Linstedt et al., 1992), even
forming lateral membrane domains (Wanaski et al., 2003). Using "fat blots"
(PIP strips), Stith et al. (manuscript in preparation) demonstrated that
9


Xenopus Src bound with a relatively high specificity to PA, but Xenopus PLC-y
did not. Reverse transcriptase PCR and Western blot results suggest that
only the message and gene product of phospholipase Dlb is present in the
egg and that this is the isoform activated by sperm to produce phosphatidic
acid at fertilization. PLDlb catalyzes the cleavage of phosphatidylcholine to
produce phosphatidic acid and choline (Figure 1.5.1]. In addition, when
phosphatidic acid was added to Xenopus eggs, Src and PLC-y was stimulated.
All these studies suggest that sperm activates phospholipase Dlb to produce
phosphatidic acid and that this lipid binds and activates Src.
I
Glycerol
Backbone
Choline
Head
Group
t
o
Figure 1.5.1 PLDlb Cleavage of Phosphatidylcholine
The chemical structure of phosphatidylcholine. Circled is the glycerol
backbone, and the choline head group is boxed. The vertical line shows
where PLDlb cleaves phosphatidylcholine to produce phosphatidic acid and
choline.
10


1.6 Phosphatidic Acid
Phosphatidic acid consists of a glycerol backbone, typically with a
saturated fatty acid bonded to carbon-1, and an unsaturated fatty acid
bonded to carbon-2. Finally, a phosphate group is bonded to carbon-3
(Figure 1.6.1). For our experiments, we used l-Palmitoyl-2-oleoyl-sn-
glycero-3-phosphate (C16:0, C18:l) (see Appendix A). Phosphatidic acid can
be formed in one of four ways: (1) by phospholipase D, via the hydrolysis of
the phosphatidylcholine to produce phosphatidic acid and choline
(Liscovitch et al., 2000), (2) by phosphorylation of DAG by DAG kinase
(Salvador et al., 2005), (3) by the acylation of lysophosphatidic acid by
lysoPA-acyltransferase (Yamashita et al., 2001), or (4) by de novo synthesis
(Alberts et al., 2008).
Figure 1.6.1 Phosphatidic Acid
The picture on the left shows the chemical structure of phosphatidic acid.
The glycerol backbone is circled. The picture on the right shows a molecular
model of phosphatidic acid.
11


Phosphatidic acid has a number of roles in the cell: it is the precursor
for the biosynthesis of many other lipids (Schneider, 1972), and influences
membrane curvature (Kooijman et al., 2003). Phosphatidic acid has been
known to act as a signaling lipid; PA enhances the activity of Fgr, a tyrosine
kinase of the Src tyrosine kinase family (Sergeant et al., 2001). The
observation that PA positively regulates another tyrosine kinase is significant
to our research in that Src is closely related in structure to Fgr (Hanks et al.,
1988). It is plausible to assume that PA is capable of enhancing the activity of
Src as well. In other studies, PA binding domains have been found to exist
that allow certain proteins, such as Raf-1 (sequence of binding domain: Raf
(390-423), FRNEVAVLRKTRHVNILLFMGYMTKDNLA1VTQW) (Ghosh etal.,
2003) and protein phosphatase-1 (sequence of PA binding domain: protein
phosphatase-1 (286 296): DETLMCSFQIL) (Jones et al., 2005) to bind
specifically to PA. Note the basic amino acids present in the Raf-1 amino acid
sequence (underlined) and hydrophobic amino acids present in protein
phosphatase-1 amino acid sequence (underlined). The existence of binding
domains specific to PA in other proteins allow for the possibility that Src has
12


a specific binding domain as well, which when bound to PA, can lead to its
activation.
1.7 Src Tyrosine Kinase
Src is a membrane-associated protein that was first identified as a
viral oncogene and it is a member of the Src family of protein tyrosine
kinases. The Src family consists of nine mammalian members including: Src,
Yes, Fyn, Fgr, Lck, Hck, Blk, Lyn, and Frk. Src consists of a myristoylated tail
at the N-terminus that helps to anchor Src to the plasma membrane, an SH2
domain, which binds phosphotyrosines, an SH3 domain that binds to proline
rich areas of proteins, and a kinase domain (Figure 1.7.1). There are two
critical tyrosine phosphorylation sites on Src: tyrosine 416/418 and tyrosine
529 (numbering systems vary: tyrosine 414 on frog Src, tyrosine 416 on
chicken Src, tyrosine 418 in human and tyrosine 424 on mouse Src).
Tyrosine 416/418 is located in the catalytic domain and
autophosphorylation at this site is required for full catalytic activity. The
tyrosine 529 is located near the carboxyl terminus of Src and acts as a
negative regulator. Phosphorylation of the tyrosine 529 by Csk induces an
intramolecular interaction between the SH2 domain and the carboxyl
terminus, causing the protein to fold back on itself (Roskoski, 2004), thereby
13


holding Src in its inactive form. This conformation blocks phosphorylation of
tyrosine 416/418, thereby preventing Src activation. When tyrosine 529 is
dephosphorylated, tyrosine 416/418 can be maximally phosphorylated and
Src becomes active. Full catalytic activity of Src requires the phosphorylation
of tyrosine 416/418.
Gly-2 myristylation site
Ser-12 : PKC
Ser-17 : cAMP-PK
Ser-48 : PKC
Tyr-192
Tyr-416:
autophosphorylation site
Tyr-527

SH4 located unique SH3,
on the mem- domain negative
brane regulation
SH2.
negative
regulation and
recognition
Figure 1.7.1 Domain Structure of Src Protein
Src has a myristylated N-terminus, a kinase domain, and SH2 and SH3
domains, which cooperate to negatively inhibit Src activity. When tyrosine
527 is phosphorylated, Src is inactive. The dephosphorylation of tyrosine
527 and the autophosphorylation of tyrosine 416/418 activate Src. Sequence
illustration is from Tatosyan and Mizenina (2000).
14


1.8 Observed Increases in PA and Src
PA levels have been observed to increase from 0.93 + 0.04 ng/egg
(n=20) to an average value of 2.51 + 0.13 ng/egg (n=29) (P<0.001) after
fertilization (unpublished data, Holland et al.). Sato and associates have
shown that Src is activated at 2 minutes after insemination during Xenopus
fertilization and that a tyrosine kinase phophorylates and activates PLC-y,
leading to intracellular calcium release (Sato et al., 2001; 2003). Using
Western blotting with phosphospecific antibodies, Stith et al. (ms in prep)
were also able to confirm the early activation of Xenopus Src during
fertilization.
1.9 Association Between PA and Src
Stith et al. (manuscript in preparation) tested whether the addition of
PA could stimulate tyrosine kinase activity as well. Two minutes after the
addition of PA to Xenopus eggs, they observed that there was an elevation of
Src activity. Using a method called Fat blots", Western blotting, and the
Odyssey System, Stith et al. (manuscript in preparation), also observed that
Xenopus Src bound more to PA than to thirteen other lipids.
15


1.10 Removal of N-Term Basic Amino Acids Lower Src-Lipid Affinity
Sigal et al. (1994) showed that Src interacts electro-statically with the
acidic phospholipids of the cell membrane. Even though Src has a
myristoylated (lipid) tail that allows it to insert into the lipid membrane,
Sigal et al. (1994) showed that Src-anionic lipid binding was more important.
The research team focused mainly on the interactions of only one kind of
acidic phospholipid (phosphatidylserine) with the six basic residues of the
amino-terminal of human c-Src. The Src they used for binding studies had a
myristoyl group, and they found that the hydrophilic amino acid sequence
with 6 positively charged residues (Myr-GSSKSKPKDPSQRRR; K = Lysine, R =
Arginine) at the N-terminus were required for PS binding. In Figure 1.10.1 it
is important to note that the six vertical lines highlighting basic amino acids
are conserved in both Xenopus and human Src.
16


30
*.SPC/1-532
hS.RC21-536
20
40
ilivilsirQP- F TSiSMOXPN
E P A|N vi AOOO A CPMHi P S
Cons(vjtior
111 k
Quality
J
Consensus
MO KSKP
ft RSL
II
Figure 1.10.1 Comparing the N terminus of Xenopus and Human Src
Highlighted are the 6 basic residues that Sigal et al. (1994) examined (K =
Lysine, R = Arginine). Amino acid coding for Xenopus Src is on top, and amino
acid coding for Human Src on the bottom. "Conservation (based on a 1 10
scale) measures the conserved physico-chemical properties of the amino
acids. "Quality is inversely proportional to the average cost of all pairs of
mutations observed in a particular column of the alignment. "Consensus"
describes whether the two amino acids are exactly the same. N-terminus is
not fully conserved, however the important basic amino acids are. In
addition, the rest of the amino acids (not shown) are highly conserved.
Image created by Ryan Bates.
Prior studies have suggested that although myristylation of Src is
required for interaction with membrane bound receptors, NH2-terminal
lysines are essential for membrane localization (Silverman and Resh, 1992).
Through mutation analysis it has been shown that six basic residues play a
17


major role in membrane binding (Sigal et al., 1994). In this experment, wild-
type and mutant Src constructs were radiolabeled and added to varying
concentrations of liposomes. Table 1.10.1 shows the different Src constructs
and notes that dissociation constants increased as the number of basic amino
acids was reduced. The samples were allowed to incubate and then the
liposomes and bound Src were pelleted and Src was separated by SDS-PAGE.
The radiolabeled Src was excised from the gels and quantitated using liquid
scintillation counting. The data gathered were plotted as percent of Src
(wild-type or mutant) that bound to the liposomes (radioactivity in Src
bound to liposomes), divided by total Src (normalized to protein available for
binding, times 100) versus the concentration of lipid that was used (Figure
1.10.2). The binding data were fitted using a binding model (Section 2.9) and
the dissociation constants were calculated. Wt Src (Myr-
GSSKSKPKDPSORRR) was mutated at the 5 and 9 position (N5N9: Myr-
GSSNSKPNDPSQRRR), or at the 5, 9 and 14-16 positions (N5N9N14-16: Myr-
GSSNSKPKDPSONNN) to a neutral amino acid (asparagine).
18


Table 1.10.1 Calculated Kds of c-Src Binding to PC:PS Liposomes
As the number of positive (basic) amino acids decreased, the dissociation
constant (Kd) increased, suggesting that the N-terminal basic amino acids aid
in Src binding to anionic phospholipids. The highest Kd for c-Src binding to
liposomes was found with liposomes lacking anionic PS ("PC") (Sigal et al.,
1994).
Kd
C-Src [PC:PS] 6 2L10"7 M
N5N9 [PC:PS] 2&105M
N5N9N14-16 [PC:PS] 1 x 104 M
Number of (+)
6
4
1
C-Src [PC] 2 x 103M 6
The increased dissociation constants suggest that the affinity between
the protein and the acidic lipid used (PS) decreased as the number of basic
amino acids decreased. The dissociation constants of mutants approach the
dissociation constant of wild-type Src binding to liposomes that do not
contain any acidic phospholipids (e.g. the "PC group). The six basic residues
of the Src protein are critical for binding the acidic phospholipids in the cell
membrane, thus they play an important role in Srcs ability to bind to cell
membranes.
19


Lipid, M
Figure 1.10.2 Mutated Basic Amino Acid Residues in Src Lowers Binding Affinity
to Liposomes Containing PS
Plot comparing percent of liposome-bound Src versus lipid concentrations.
Percent bound was calculated as Src that bound to the liposomes divided by
total Src (normalized by fraction available for binding) multiplied by 100. All
PC:PS liposomes used were made in a 2:1 ratio. As more basic amino acids at
the N-terminus were mutated to neutral amino acids, the dissociation
constants increased, suggesting that these basic amino acids present in Src
are important for acid phospholipid (PS) binding (Sigal et al., 1994).
20


1.11 The Affinity Between Src and Phosphatidic Acid
As noted, there is evidence that phosphatidic acid is acting as an
upstream regulator of Src kinase and PLC-y to stimulate calcium release from
intracellular stores during Xenopus fertilization. However, the Src-
phosphatidic acid binding affinity has never been measured. While PIP
strips are useful as a preliminary method in comparing selectivity of binding,
this method does not allow for accurate dissociation constant estimations.
Through comparison of dissociation constants, it is possible to determine
whether the binding between Src and PA is relatively high and specific. A
second main concern with PIP Strips is that the lipids are simply in a
monolayer on a flat paper and the orientation of the lipids is random. This
means that Src may not be interacting with the lipids as it would in situ and
this may result in misleading binding data. Thus, the dissociation constant
between Src and phosphatidic acid is measured with liposomes that
reproduce in vivo conditions.
21


2. MATERIALS AND METHODS
2.1 Preparing Sucrose-loaded Liposomes
Different lipids (PC, PA, and LPA) that were suspended in chloroform,
were mixed to the correct proportions (100% PC, 2:1 and 4:1 PC:PA and
PC:LPA) (Figure 2.1.1). All lipids were purchased from Avanti Polar Lipids
Inc. (Alabaster, AL): l-Palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine
(C16:0, C18,l), l-Palmitoyl-2-oleoyl-sn-glycero-3-phosphate (C16:0, C18:l)
and l-oleoyl-2-hydroxy-sn-glycero-3-phosphate (C18:l) (see Appendix A).
The long chain lipids were chosen to mimic the lipids normally found in the
egg, and to better stabilize the liposomes. Using the Meyer N-Evap Analytical
Evaporator (Organomation and Associates Inc, Berlin, MA) and nitrogen, the
chloroform was evaporated, and the lipid film was rehydrated with lmL of
"Inside Buffer (176mM Sucrose and ImM MOPS) (see Appendix A).
Liposomes were made with lipids and inside buffer using the Mini-Extruder
(Avanti Polar Lipids Inc.; Alabaster, AL). The extruder-heating block (Figure
2.2.2) was placed on a hot plate and allowed to heat up to the phase
transition temperature of the lipids used. The dried lipids were then
hydrated using "Inside Buffer" for one hour. The lipid suspension was kept
above the phase transition temperature of the lipid during hydration as well
22


as extrusion. This increases the entrapment of water-soluble compounds
such as sucrose. When the sample was fully hydrated, it was loaded into one
of the gas-tight syringes and carefully placed into one end of the Mini-
Extruder. An empty gas-tight syringe was then placed into the other end of
the mini-extruder. The fully assembled extruder apparatus was then
inserted into the extruder-heating block, and the syringes were clipped on.
The lipid suspension was allowed to equilibrate with the temperature of the
heating block, and then the lipids were extruded. The plunger of the syringe
containing the lipids was gently pushed until the solution was completely
transferred through a filter to the empty syringe. This step was repeated 11
times to ensure a homogenous lipid solution. Figure 2.2.1 shows the proper
way to set up the extruder and syringes into the heat block. The liposome
solution was divided into two centrifuge tubes and 4mL of "Outside Buffer"
(see Appendix A] was added to each tube. The samples were separated into,
two centrifuge tubes because the centrifuge tubes could not hold larger
volumes. The sucrose-loaded liposomes were pelleted (100,000 x g, 1 hr.)
using an ultracentrifuge (Beckman Coulter, Brea, CA; Model: L5-50; Rotor
Type: 60 Ti). The supernatant was discarded, and lmL of Outside Buffer was
added to the pellet. The result was 2 mL of 100 nanometer-diameter
23


sucrose-loaded liposomes (Avanti Polar Lipids, Alabaster, AL) with a final
concentration that was about 30% less than what was originally calculated
(see Appendix B). This value is based on a study that showed an average of
30% lipid loss during extrusion (Loughrey et al., 2007). The liposomes were
stored at 4C, and were used less than one week later.
Mix the correct
proportion of lipids m a
test-tube (1ml)
i
Evaporate chloroform
from the lipids using the
Organ ova p (underN2B) j
Add lmL of Inside Suffer
(176mM Sucrose and
ImM MOPS) to rehydrate
/
/
i
Dry Film of lipids
/
/
/
1
i
/
lmL ftehydrated lipids
Extrude the solution using j
Avanti lipid Mini-Extruder j
-------------------------------> ^
lmi. Liposomes
f
0.5ml
.iposomes
4.Sml
liposomes in
Outside Suffer
Pelleted
Liposomes
1 Add4mLof ! Centrifuge at
Outside Buffer ; 100,000 x B for 1
(lOQmM KC! and ! hour using
ImM MOPS) j ultracentrifuge
Remove supernatant immediately
and discard. Add 1ml of Outside
Buffer. Store at 4*C
Sucrose-loaded
Jposomes (100
nanometer in diameter}
Figure 2.1.1 Protocol for Preparing Sucrose Loaded liposomes
Protocol for preparing liposomes used for vesicle sedimentation binding
assay. Notice that the end result is 2 mL of 100 nanometer-diameter
sucrose-loaded liposomes, which can be stored at 4C for up to a week.
24


2.2 Avanti Mini Extruder Assembly
The Mini-Extruder was set up according to the instructions provided
by Avanti Polar Lipids Inc. (Alabaster, AL) (Figure 2.2.1). The two internal
membrane supports were set on a flat surface with the O-rings facing up.
Two filter supports were pre-wet with Inside Buffer and placed over the
Teflon orifice inside the O-ring inner diameter of the first internal membrane
support. One Polycarbonate Filter (100 nanometer diameter pores) was
then pre-wet with Inside Buffer and placed on top of the Filter Supports. The
first Internal Membrane Support, containing the two Filter Supports and the
Polycarbonate Filter, was then placed in the Extruder Outer Casing with the
Polycarbonate Filter and Filter Supports facing up. Two more Filter Supports
were pre-wet with Inside Buffer and placed over the Teflon Orifice inside the
second O-ring Diameter of the second Internal Membrane Support. This was
also placed into the Extruder Outer Casing, but with the Filter Supports
facing down. The Flat Washer was then placed on top, and the Retainer Nut
was placed on the threaded end of the Extruder Casing and tightened.
25


On^g
Channel
V.
Bctrudar Chter
Chang
t
Internal
Mmbrane
Suppcrt
Pdycarbcnate
Mmhrane
Teflon
Bearing
Retainer
Nit
Figure 2.2.1 Avanti Mini-Extruder Set-up
This figure shows the extruder set up. Notice the polycarbonate filter in the
middle, surrounded by two filter supports on either side. Two internal
membrane supports surround the membrane and filter supports. This whole
set-up fits into the extruder outer casing, and the Teflon bearing and retainer
nut work to keep it all together. The O-rings help the lipid solution from
seeping out into the extruder outer casing during extrusion. Image from
Avanti Polar Lipids.
26


Figure 2.2.2 Extrusion Technique
The syringes must be injected into the extruder before being placed on the
heating block. Extrusion should be repeated 11 times to ensure homogenous
lipid solution. Image from Avanti Polar Lipids.
2.3 Liposome Protein Binding Protocol: Varying Protein Concentration
We used two different vesicle sedimentation protocols for Kd
determination: varying the protein or varying the liposome (lipid)
concentration. For the former protocol, different concentrations of protein
were prepared in advance using a "Binding Buffer" (20mM HEPES, pH 7.0,
100 mM NaCl, 0.1% Triton X). Boving Serum Albumin (BSA) was chosen to
initially test this protocol. The BSA concentrations used were: 0 nM, 10 nM,
27


20 nM, 40 nM, 70 nM, 80 nM, and 100 nM. Each concentration [135 (J.L) was
mixed with 85 pL of the 8 mM sucrose-loaded liposomes to obtain a total
volume of 220 pL [Figure 2.3.1). This mixture was incubated at room
temperature for 15 minutes to allow BSA binding to the liposomes, and then
the mixture was centrifuged at 126,000 x g [80,000 rpm, 1.5 hrs.) using an
airfuge [Beckman Coulter, Brea, CA; Model: 350624; Rotor Type: 30 A-100).
Approximate centrifugation speeds were calculated using the airfuge manual
(Figure 2.3.2). The centrifugation speed and time were chosen to closely
resemble the speed and time used by Sigal et al. (1994). After centrifugation,
110 pL of the supernatant was withdrawn for protein analysis. The pellet
was then resuspended with 220pL of Binding Buffer. Half (110 pL) of the
resuspended pellet solution was removed in order to quantify the amount of
protein bound to liposomes Bound Protein". Binding Buffer (110 pL), 20
mM KCN (5 pL) and 5 mM ATTO-TAG CBQCA (10 pL) (Molecular Probes Inc.,
Carlsbad, CA) were added. This mixture was allowed to incubate for an hour
and the amount of protein was determined with the TD-700 Fluorometer
(Turner Designs; Sunnyvale, CA). The Filter Cylinder and 13mm Sample
Adapter was used.
28


85uL of ! 135uL Protein
Liposomes | Solution
\

220ul of
Liposome and
Protein Mixture
Incubate for IS j Then, Centrifuge at 126,000
| x g {80,000rpm} on
, | Beckman Airfuge
S/N (a) and Pellet
(a) of Liposome-
Src Solution
Remove 110u^ of
S/N for analysis
I
Add HOuL of Src-
Smdmg Buffer, SuL of
20mM ofSmM ATTO-TAG
CBaCA (Molecular
Probes}
Pellet
220uL of Src
Binding 3uffer (2C
mW HEp£S, 100
mM NaCI, 0.1%
Triton )
Resuspended
Pellet
23SuLofS/N Sample for Incubate for 1 hour and Analyze using TD-/00 235uLofS/N Sample for
*
analysis Laboratory Fluorometer analysis
Figure 2.3.1 Liposome-Protein Binding Protocol: Varying Protein Concentration
Protocol used to quantify the amount of protein bound to liposomes, while
varying protein concentration. S/N is short for supernatant.
29


700
90
_ 80

o 70
| 60
o
£ 50
^ 40
a. 30
CE
20
10
0
69 kPa 138 207
(10 psig) (20) (30)
Pressure at Instrument Gauge
Air Pressure psig (kPa) Nominal Rotor Speed rpm Relative Centrifugal Field <**) atrmax
18 Rotor 30 Rotor
22 (152) 80 000 105 000 126 000
27 (186) 90 000 133 000 160 000
30 (207) 95 000 149 000 1 78 000
Figure 2.3.2 Beckman Airfuge: Relating Pressure to Centrifugation Speeds
Centrifugation speeds were calculated using these charts. Top graph
compares Presure at Instrument Gauge (psig) to centrifugation speed (rpm
in thousands). Bottom chart shows the relative centrifugal field at various
run speeds (Beckman Coulter, Brea, CA).
30


2.4 Liposome-Protein Binding Protocol: Varying Protein Concentration
with KCI Protein Release
The basic steps of this protocol were the same as that noted in Section
2.3 (Figure 2.3.1]. However, after centrifugation, the pellet was resuspended
with 220 pL of 150 mM KCI and allowed to incubate at room temperature for
15 minutes (Figure 2.4.1). The added salt should maximize release of bound
protein from the pelleted liposomes (Schoeffler et al., 2003). The KCl-pellet
mixture was centrifuged again at 55,000 x g (35,000 rpm, 15 min.) using the
Beckman Airfuge to pellet liposomes. Since this centrifugation step was used
to clarify the resuspended pellet, and the removal of all liposomes was not
necessary, the high speed and long centrifugation time was not needed. After
centrifugation, llOpL of the supernatant was removed in order to quantify
the amount of "bound" protein released from the liposomes. KCN and ATTO-
TAG CBQCA (Molecular Probes Inc., Carlsbad, CA) were added as noted
(Section 2.3) to quantify bound and total BSA protein.
31


85uL of
Liposomes
13SuL Protein
Solution
Pellet (a)
220uL of ISOmM
\
22CuL of
Liposome and Src
Protein Mixture
I Incubate for IS
Minutes
i Then, Centrifuge at 126,000
| x g {80,000rpm} on
j Beckman Airfuge
Discard
remaining S/N
/
Pellet (a)
rehydrated in Incubate for IS
Minutes
Then, Centrifuge at
55,000 x g (3S,OOOrpm)
on Beckman Airfuge
S/N (a) and Pellet S/N (b) and 3ellet
(a) of Liposome- (b) of rehydrated I
Src Solution Pellet (a) |
Remove IIOul of
S/N for analysis
235uL of S/N
Sample for
analysis
Add llOulof Src-
3inding Buffer, SuL of
20mM of 5mM ATTO-TAG
CBQCA (Molecular
Probes)
Incubate for 1 hour and
Analyze using TD-700
Laboratory Huorometer
Figure 2.4.1 Liposome-Protein Binding Protocol: Varying Protein Concentration
with KCI Protein Release
Protocol used to quantify the amount of protein bound to liposomes, while
varying protein concentration.
2.5 Liposome-Protein Binding Protocol: Varying Liposome Concentration
with KCI Protein Release
As different laboratories report both methods, we also varied the
liposome concentration instead of the protein concentration (Figure 2.5.1).
The initial liposome concentrations used were: 0 mM, 2 mM, 4 mM, 6 mM,
and 8 mM. Liposomes (lOOpL) of were mixed with 120pL of 150nM BSA. In
32


subsequent experiments, 10 active units of human, recombinant, non-
myristoylated Src (Upstate Biotechnology, Lake Placid, NY) in 135pL, or
150nM poly-D-lysine (MW 30,000 to 70,000) (Sigma-Aldrich, St. Louis, MO)
was used instead of BSA. Using an estimated specific activity of 10
pmole/pg/min for Src, we calculated the concentration of 10 units of Src to
equal 122 nM, which is close to the 150 nM concentrations we used for the
control proteins (Appendix C). Analysis of the amino acid residues suggest
that Xenopus Src and Human Src are highly conserved (Figure 1.10.1). In
addition, the recombinant human Src protein that we used did not have a
myristoylated tail as it was our intention to focus on the interaction between
the protein and the head group of the lipid, rather than the interaction
between the tail and membranes.
The final total volume of the protein and liposome mixture was 220pL
(see Appendix C). This mixture was incubated at room temperature for 15
minutes and then centrifuged at 126,000 x g (80,000 rpm, 1.5 hrs.) in the
Beckman Airfuge. As noted, the centrifugation speed and time were set to
resemble Sigal et al. (1994). After centrifugation, the supernatant was
removed. The pellet was rehydrated with 220pL of 150mM KC1 and allowed
to incubate for 15 minutes (R.T.), and centrifuged again at 55,000 x g (35,000
33


rpm, 15 min., Beckman Airfuge). The lower speed and shorter centrifugation
time were used to clarify the resuspended pellet so that protein
quantification can be achieved. After centrifugation, HOpL of the
supernatant was removed in order to quantify the amount of "bound" protein
present with the ATTO-TAG CBQCA assay (Molecular Probes Inc., Carlsbad,
CA). Control samples containing no lipid were also analyzed to assess
protein aggregation. The percent bound protein was corrected by
subtraction of the percent aggregated and normalized by use of the fraction
of protein available for binding. Therefore, the percent of bound protein is
equal to the amount of protein in the liposome pellet, divided by the
normalized fraction of protein available for binding (x 100 to acquire
percentage).
34


lOOulof
Liposomes
120ul Protein
Solution
/
/
/

220uL of
Liposome and Src
Protein Mixture
Incubate for 15 | Then, Centrifuge at 126.000
| x g (80,0GQrpn) on
\ , ( 3eckman Airfuge
Pellet {a)
220uL oflSOmM
\ *
Pellet (a)
rehydrated in Incubate for 15
Minutes
Then, Centrifuge at
i 55,000 x g (35,OOQrpm)
v 1 on Beckman Airfuge
S/N (a) and Pellet S/N (bj and 3e(iet
(a) of Liposome- (b) of rehydrated
Src Solution Pellet (a)
i
Remove
Supernatant
Add HOuL of Src-
3inding Suffer, SuL of
20mM of 5mM ATTO-TAG
CBQCA (Molecular
Probes)
i
Remove llOuL of
S/N for analysis !
| Incubate for 1 hour and
j Analyze using TD-700
| Laboratory Fluorometer
235uL ofS/N
Sample for
analysis
Figure 2.5.1 Liposome-Protein Binding Protocol: Varying Liposome
Concentration with KCI Protein Release
The protocol used to quantify the amount of protein bound to liposomes,
while varying liposomal concentration. Supernatant was shortened to "S/N"
in the figure.
2.6 Testing Protein Quantification Using 280nm Spectrophotometry
To quantify liposome bound protein (BSA, Src, or PDL), we pellet the
liposomes and bound protein, then quantify the amount of protein in the
liposome pellet. We first tried to measure BSA by absorbance. Various
bovine serum albumin protein concentrations were made by diluting the
35


stock BSA solution with Binding Buffer (Appendix A). The BSA
concentrations used were 0 nM, 75 nM, 150 nM, 300 nM, 450 nM, 600 nM
and 750 nM. The protein concentrations chosen were in the nanomolar
range as this is the range used in the Kd determination and to test whether
this method would be sufficiently sensitive. The spectrophotometer was set
to 280 nm and the various protein concentrations (500 pL) were pipetted
into a cuvette and absorbance was recorded. A 1% (10 mg/mL) of BSA has
an absorbance of ~0.667 in a cuvette that has a 1cm path length (Pierce,
Rockford, IL). Therefore, to calculate the concentration of BSA, we used this
equation:
Spec, absorbance reading = X Eq. 1
6.67 1%
This is equation can be rewritten as:
Spec, absorbance reading = X Eq. 2
0.667 0.1%
However, since 0.1% of BSA is the same as 15.06pM (see Appendix B), we
were able to rewrite equation 1 as:
Spec, absorbance reading = X Eq. 3
0.667 15.06pM
36


Using equation 3, we plugged in the spectrophotometer absorbance reading
and solved for X; where X is the micromolar protein concentration detected
(Pierce Biotechnology, Rockford, IL).
2.7 Testing Protein Quantification Using ATTO-Tag CBQCA
Protocols for testing the CBQCA tag were acquired from Molecular
Probes Inc. (Carlsbad, CA). Various BSA protein concentrations were made
using a stock solution (150 mM). Potassium cyanide (7.5 pL, 20 mM) was
then added to each protein sample. The CBQCA tag was prepared by diluting
a stock solution to a 5mM working solution with 0.1 M sodium borate buffer
(pH 9.3). The working solution of CBQCA (15 pL) was then added to each
protein sample. The samples were allowed to incubate for ~1 hour, and
were protected from light. After incubation, the fluorescence was measured
using a TD-700 fluorometer (Turner Designs; Sunnyvale, CA). The ATTO-Tag
CBQCA reagent is effective at quantifying accessible amines in protein in the
presence of cyanide (Figure 2.7.1). The absorbance wavelength of this tag is
465 nm, and the emission wavelength is 550 nm. Molecular Probes states
that the tag is able to detect protein quantities ranging from 10 ng to 150 pg.
37


Figure 2.7.1 CBQCA Reaction
In the presence of cyanide, CBQ is able to react with primary amines to form
a product that absorbs light at 465nm and emits light at 550nm (Molecular
Probes Inc., Carlsbad, CA],
2.8 Affinity Constants
Affinity constants are numeric representations of the strength with
which two molecules interact. When coupled with biological experiments,
affinity constants can provide valuable insight into the mechanisms of
interactions and aid in predicting if specific interactions take place in cells.
Equilibrium exists when there is no overall change in the
concentrations of the free and bound species in the reaction over time. In
these experiments, "total" protein is compared with liposome "bound
protein to determine the Kd. It is important to note that at equilibrium, a
38


binding reaction is not static. Within a system, complexes are constantly
forming while others are dissociating; however, at equilibrium there is no net
change in bound versus free.
The affinity constants are: Dissociation constants (Kd) and association
constants (Ka). They are quantitative parameters used to describe affinity
that can be experimentally measured. A Kd is expressed in units of M
(moles/liter] while a Ka is expressed in units of M1. Kd is inversely related to
affinity: as affinity increases, Kd decreases. On the other hand, Ka is directly
related to affinity: as affinity increases, Ka increases.
The observed binding interactions between two biomolecules can be
described by the equation:
K.
Protein (P) + Liposome (L) =£ [P-L]
Eq. 4
where P and L represent two different interacting biomolecules. P is free
and is approximately equal to the total protein, whereas P-L is protein that is
bound to liposomes ("bound protein). Affinity constants are related to the
concentrations of the three components present in a bimolecular reaction at
39


equilibrium. The following equations show the relationship between Ka, Kd
and the three components:
Kd = [P] [L] / [P-L] Eq. 5
Ka = [P-L] / [P] [L] Eq. 6.
In these equations, [P], [L] and [P-L] are the molar concentrations of these
reaction components at equilibrium.
The population of P molecules will be split between free P and P that
have been bound in P-L complexes; where
[P]Total = [P] + [P-L] ~ free P or [P] Eq. 7
The same is true of the L molecules, where
[L]Total = [L] + [P-L] Eq. 8
Understanding these concepts, we can use these equations to derive this
equation:
[P-L] / [P]Totai = [L] / (Kd + [L]) Eq. 9
Here, [P-L] / [PJiotai represents the fraction of P that is bound (in the P-L
complex); fraction of protein bound to liposomes. [L] represents the
concentration of the liposomes in the reaction at equilibrium. In order to
measure the Kd using this equation, the concentration of P in the reaction
was varied and the amount of L added to reactions was varied. This is the
40


vesicle sedimentation method where protein concentration was varied, but
one can solve for an equation with varying lipids. At each point of L
(liposome concentration), the concentrations of the bound protein is
measured P-L. This equation can further be derived to obtain the equation:
[P-L] / [P]iotai = Ka[L] / (1 + Ka[L]) or Eq. 10
Bound [P] / Total [P] = Ka[free liposome] / (1 + Ka[L]) Eq. 11
Using this equation, if [L] = Ka and since (Ka Ka) = 1, then (Ka Ka) / (1 + (Ka
* Ka)) = (1 / (1 + 1)) = Vz. Therefore, when estimating Ka's using a graph of
varying lipid concentration (x-axis) versus percent protein binding (y-axis),
the concentration (x-axis point) that correlates with 50% of maximal binding
is considered the Ka (see Figure 1.10.2 as an example).
2.9 Statistical Analysis
The R program was used to calculate Kds and 95% confidence
intervals using our data (R Development Core Team, 2007) (see Appendix C).
We fit the model to the equation y = K*x / (1 + K*x), where y is percentage
bound, x is the lipid concentration and 1/K is the parameter of interest.
When comparing the estimated Kds, if the confidence interval in A
significantly overlaps B to include the estimated Kd, we can say that A and B
41


are not significantly different at the 0.05 level (p > 0.05). If the confidence
interval of B overlaps the confidence intervals of C, but does not overlap the
estimated Kd, we can say that B and C are significantly different at the 0.05
level (p < 0.05). Finally, if the confidence interval of A does not overlap that
of C, then we can say that A and C are significantly different at the 0.05 level
(p 0.05) (Appendix D).
42


3. RESULTS
3.1 Protein Detection Using the Absorbance Assay (280nm)
As noted, one must measure bound protein. The 280nm Absorbance
Assay was the first method we chose for protein detection. Absorbance
assays are fast and convenient since no additional reagents are required. In
addition, absorbance assays do not consume the protein. These positives
made it the perfect choice for our project for we wanted to use protein
concentrations in the nanomolar range. Bovine Serum Albumin (BSA) was
the protein chosen to run our preliminary tests due to its relatively modest
price. We tested various BSA concentrations (0 nM, 75 nM, 150 nM, 300 nM,
450 nM, 600 nM and 750 nM) to see if this method would accurately record
protein concentrations.
The concentrations of BSA used were plotted against the BSA
concentration that was calculated using absorbance (Figure 3.1.1). The slope
and estimated protein concentration from each trial varied considerably,
ranging anywhere from below, to doubling that of the actual concentration.
For example, for 300 nM [BSA], the method determined concentrations from
200 to 600 nM. This method was eventually abandoned due to the high level
of variance.
43


0 200 400 600 800
BSA Concentration (nM)
Figure 3.1.1 BSA Concentration vs Calculated BSA Concentration Using 280nm
Absorbance Assay Produced Variable Results
"BSA Concentration refers to the known concentration of the protein
solution. The "Calculated BSA Concentration" refers to the concentration
that was calculated using equation 3 (Section 2.6), after the absorbance at
280 nm was recorded. Each line was one experiment.
3.2 Testing the ATTO-TAG CBQCA
The fluorescent CBQCA Protein Quantification Kit (Molecular Probes
Inc., Carlsbad, CA) utilizes the ATTO-TAG CBQCA reagent that, in the
presence of cyanide, reacts with primary amines to form fluorescent
44


derivatives that are quantified using a fluorometer. The kit was tested using
various concentrations of BSA, and a standard line was created from the data
collected. The raw fluorescence was compared to the actual BSA amount
used, and a best-fit line was drawn through the points (Figure 3.2.1). The
equation of the best-fit line is: y = 1.0031x 11.732; the R-squared value is
0.99609 (see Appendix B).
BSA (ng)
Figure 3.2.1 Testing the ATTO-TAG CBQCA using BSA
This figure shows the standard line of BSA, using the CBQCA tag. These
values are based on an average of two experiments (n=2).
45


3.3 Preliminary Protein-Liposome Binding Data
The CBQCA kit proved to be accurate and consistent in detecting
various standard protein concentrations in the nanomolar range. However,
we wanted to test whether the CBQCA tag would work with our vesicle
sedimentation-binding assay and we ran preliminary tests using BSA. Sigal
et al. (1994) were able to show that the six basic residues of the Src protein
are critical for binding the acidic phospholipids in the cell membrane (see
introduction, section 1.10). BSA is known to have basic residues that would
allow it to bind acidic lipids. It was expected that if basic amino acids
(positively charged) played an important role in protein binding to
membranes (Sigal et al., 1994), then protein should bind more readily to
liposomes containing acidic phospholipids (negatively charged, such as PA).
In order to determine whether the protein is actually binding to the
liposomes the way we expected, we chose to test our liposome-binding assay
using liposomes containing PC and PA lipids. This data would be compared
with our protein binding to liposomes only with PC. Liposomes containing
2:1 PC/PA were created using the Avanti Mini Extruder. The 2:1 ratio was
chosen based experiments by Sigal et al. (2004).
In our first experiments, various concentrations of BSA were added to
46


the liposomes. The samples were allowed to incubate, and then liposomes
and bound protein were pelleted. The CBQCA tag was added to the
supernatant and bound protein from the liposome pellet (see Methods:
Section 2.3; Figure 2.3.1). Samples were analyzed using a fluorometer. The
BSA concentration detected in the pellet ("bound" BSA) was plotted against
the BSA concentration that was initially added ("total" BSA) (Figure 3.3.1).
It was expected that as the amount of protein added to the liposomes
increased, the amount of protein that bound would increase as well. We
expected to see a shape similar to that of a hyperbolic binding curve (Sigal et
al., 1994). However, our results showed no significant trend. Our data also
revealed a lack of consistency between the different trials, even though the
same method and stock protein solution were used. The most disturbing
aspect was the realization that after subtracting out the background noise
using our blank (0 ng) some points exhibited negative protein levels. These
observations made it unclear as to whether any binding actually occurred.
The inconsistency between trials, the negative protein binding observed, and
the fact that the resuspended liposome solution appeared murky, led us to
suspect that the liposomes may be affecting the fluorometry readings, and
that liposomes may be trapping protein in the pellet.
47


500
400 -
300 -
200 -
CD 100 -
T3
C
-100 -
-200 -
-300 -
0 200 400 600 800 1000 1200
BSA(ng)
Figure 3.3.1 Preliminary Protein-Liposome Binding Data
The x-axis shows the "total protein concentration that was added to the
liposomes, whereas the y-axis shows the protein that bound to the
liposomes. It is unclear as to whether the protein is actually binding to the
liposomes, due to the inconsistency and the negative binding concentrations
(when compared to zero).
3.4 Protein-Liposome Binding Data with KCI-Protein Release
Potassium chloride was added to the pellet containing the liposomes
and bound protein in order to release the protein from the liposomes. The
resuspended pellet was then centrifuged at a low speed, to pellet out the
48


sucrose-loaded liposomes and the released protein was measured in the
supernatant (Figure 2.4.1). This allowed for quantification of the protein
without the interference of the liposomes.
The concentration of BSA in the resulting supernatant (bound
protein) was plotted on the y-axis against the estimated free (or total)
concentration of BSA (x-axis) (Figure 3.4.1). It was observed that as the
amount of protein added increased, the amount of protein binding increased.
This observation not only suggested that protein is binding to our liposomes,
but that the KCl-protein release modification is working to release the
protein from the liposomes, allowing for its quantification. More
importantly, with this method (Figure 2.4.1), we were able to successfully
acquire reproducible data. After averaging three different experiments, our
standard error bars were small.
49


0 20 40 60 80 100
BSA (nM)
Figure 3.4.1 BSA Binding to PC/PA Liposomes
Comparing the concentration of BSA that bound, to the concentration of
"free" BSA. The data points were taken from an average of three different
experiments (n=3) using the same stock BSA solution.
Having acquired the expected dose response curve using the KC1-
protein release modification, we decided to continue the next set of tests. To
further test our protocol, liposomes containing solely PC were used. It was
still unclear as to whether PA was the primary reason for the observed
protein-lipid interaction, or whether the vast amount of PC played a bigger
role in the interaction. To test this, various amounts of protein were added
50


to liposomes containing PC alone. The samples were analyzed using the KC1
release method (Figure 2.4.1) as used in Figure 3.4.1.
The concentration of PC-bound BSA was plotted against the
concentration of "free (or total) BSA (Figure 3.4.2). As the concentration of
protein added increased, the concentration of protein that bound to
liposomes increased, similar to that of PC/PA liposomes. It is possible that
the liposomes are able to trap a small amount of the protein in the pellet even
if there is no protein-lipid binding. Therefore, it was expected that a trace
amount of protein would be detected in the pellet. Furthermore, PC has both
negative and positive charges that could bind to BSA. The percent of BSA
that bound to PC liposomes increased as the concentration of BSA added
increased.
51


Figure 3.4.2 BSA Binding to PC Liposomes
Comparing the concentration of BSA that bound to the liposomes containing
PC lipids [no PA present), to the concentration of "free" (or total) BSA. The
data points were taken from an average of three different experiments (n=3)
using the same stock BSA solution.
The low standard errors in both PC/PA and PC binding experiments
suggested that our protocol was ready to be used to test Src binding. We
wanted to compare our PA data with those of another control acidic
phospholipid, and phosphatidylserine has been used (Sigal et al., 1994).
Phosphatidylserine is a lipid that has a net negative charge: phosphate has a
negative one charge, and the head group has both a positive and negative
52


charge for a net negative one charge. Liposomes containing
phosphatidylcholine and phosphatidylserine were created using our
extrusion protocol. Various protein concentrations were added to the
liposomes, incubated (15 min.), centrifuged, and the pellet was resuspended
in 150mM KC1. The resuspended pellet was centrifuged again for 15 minutes
and the supernatant was removed for protein analysis. The experimental
trials yielded above maximum readings (Table 3.4.1). After consideration of
the phosphatidylserine structure, we suggested that the PS head group
contains an amine that the CBQCA tag can react with. Although the KC1
protein release method seems to be effective in separating the protein from
the liposomes, with the low speed clarification, lipids are still present in
supernatant. It was suspected that PS vesicles or free phosphatidylserine
increased the fluorometer readings and that protein signal was relatively
minor.
53


Table 3.4.1 Fluorometer Readouts for BSA Binding Assay Using PC/PS
Liposomes
The fluorescence of the resuspended pellets were too high to be detected;
even the samples that contained no protein were over-fluorescing.
BSA (ng) Fluorometer Readout PC/PS Pellet Trial 1 Fluorometer Readout PC/PS Pellet Trial 2 Fluorometer Readout PC/PS Pellet Trial 3
0 OVER OVER OVER
100 OVER OVER OVER
200 OVER OVER OVER
400 OVER OVER OVER
700 OVER OVER OVER
Having failed to acquire any data using PS, a different acidic
phospholipid was chosen: Lysophosphatidic acid (LPA). LPA is similar to PA
in every respect, except that LPA has only one fatty acid tail, whereas PA has
two. The identical negative phosphate head groups made this lipid an
appealing choice. The amount of bound BSA was plotted against the amount
of total BSA (Figure 3.4.3). After averaging three experiments, the standard
errors calculated were low enough to suggest that our protocol is working in
a consistent manner.
54


Figure 3.4.3 BSA Binding to PC/LPA Liposomes
A comparison of the concentration of BSA that bound to the 2:1 PC/LPA
liposomes, to the concentration of "free (or total) BSA. The data points were
taken from an average of three different experiments (n=3) using the same
stock BSA solution.
All three graphs (Figure 3.4.1, Figure 3.4.2, Figure 3.4.3) were plotted
together in order to compare the amount of binding observed between the
different liposomes (Figure 3.4.4). The data suggested that liposomes
containing PA bound better to BSA than liposomes that contained only PC or
liposomes that contained PC and LPA.
55


40
Figure 3.4.4 Comparing BSA-Liposome Binding
This figure compares the amount of BSA that bound to liposomes that were
made up of 100% PC (closed circles) versus the amount of BSA that bound to
liposomes that were made up of 2:1 PC:PA (open circles) versus the amount
of BSA that bound to liposomes that were made up of 2:1 PC:LPA (triangles).
The data represent three experiments (n=3).
Although we were able to achieve protein binding and protein release
with our protocol, we realized that using our data to plot percent of bound
lipid versus the concentration of protein would be rather difficult task since
our protocol only allows us to detect the amount of protein. It could be
56


assumed that one binding occurs in a one-to-one ratio, but this may not
always be the case. It could easily be that patches of lipid are responsible for
attracting the protein to the membrane.
3.5 Liposome-BSA Binding
We switched over to the alternate liposome binding assay: varying
lipid protocol, as this method was recommended by Sigal et al. (1994) and
Buser et al. (1998) (see Figure 1.10.2). We measured the affinity of control
protein BSA to liposomes containing 2:1 and 4:1 PC/LPA or PC/PA or PC
alone (Figure 2.5.1).
With our new method, we plotted the percent of bound BSA against
the log of the concentration of lipids that were added (Figure 3.5.1). In
addition, we noted a sigmoidal curve that was similar to those reported by
Sigal et al. (1994) (Figure 1.10.2). Using regression analysis with the R
program (Appendix C), we fit our data to the model fb = Ka[L]/(l + Ka[L]) and
we were able to estimate Kds for BSA and these liposomes. In this equation,
fb is the fraction bound, [L] is the accessible lipid concentration, and Ka is the
apparent association constant. The apparent dissociation constant Kd is
equal to 1/Ka. Our estimated Kd's were a bit higher than the range suggested
57


by Mclaughlin (1995), which was consistent with the idea that BSA would not
show high affinity to liposomes. The estimated Kd for BSA binding to 2:1
PC/LPA liposomes was 0.19 M (95% Cl: 0.15, 0.27). The estimated Kd for
BSA binding to 2:1 PC/PA was 0.021 M (95% Cl: 0.017, 0.027). The
estimated Kd for BSA binding to PC liposomes was 0.058 M (95% Cl: 0.046,
0.074) (Table 3.6.1).
Figure 3.5.1 BSA Binding to Liposomes
Percent binding of BSA to liposomes consisting of 2:1 PC/LPA, 2:1 PC/PA or
just PC. Inner graph is an enlarged image of points between -3 and -2 M lipid.
58


Note the differences in binding were only at higher lipid
concentrations. PA may form highly concentrated patches in the membrane.
We also wanted to lower the percentage of PA and measure binding. We
decided to decrease the 2:1 PC/PA ratio to 4:1 PC/PA and 4:1 PC/LPA
(method as in Figure 2.5.1). With regression analysis of the data (Figure
3.5.2), we estimated the Kd's for these liposomes with lower PA as well. The
estimated Kd for BSA binding to 4:1 PC/LPA liposomes was 0.25 M (95% Cl:
0.19, 0.34), and for 4:1 PC/PA was 0.025 M (95% Cl: 0.020, 0.034) (Table
3.6.1).
59


100
Figure 3.5.2 BSA Binding to 4:1 Liposomes
BSA binding to liposomes consisting of 4:1 PC/LPA, 4:1 PC/PA or PC alone.
To provide a comparison, PC data is from a prior figure (Figure 3.5.1). Inner
graph is an enlarged image of points between -3 and -2.
60


3.6 Summary of BSA Binding Kd's
Table 3.6.1 Relative Binding of Control Protein BSA to Various Liposomes
Estimated dissociation constants between BSA and the varying liposome
composition, along with their corresponding confidence intervals (Cl). BSA
bound best to liposomes containing PA and worst to liposomes containing
LPA.
2:1 PC/PA 0.021 0.017,0.027
2:1 PC/LPA 0.19 0.15,0.27
PC 0.058 0.046, 0.074
4:1 PC/PA 0.025 0.020, 0.034
4:1 PC/LPA 0.25 0.19, 0.34
When the Kd's and the 95% confidence intervals (Table 3.6.1) are
plotted against their respective liposome composition (Figure 3.6.1), one can
61


determine whether the estimated Kd's are significantly different (see
Appendix E).
The 95% confidence intervals of the dissociation constants estimated
using the 2:1 PC/PA liposomes and the 4:1 PC/PA liposomes overlap
significantly, suggesting that although the lipid concentration of PA has
decreased, it did not significantly affect the estimated Kds (p > 0.05).
Similarly 2:1 PC/LPA and 4:1 PC/LPA liposomes show similarly binding to
BSA (p > 0.05). However, when comparing BSA binding to PC/PA liposomes
with PC/LPA liposomes BSA bound approximately ten-fold better to PA than
LPA (p 0.05). Surpisingly, LPA binding was approximately 4 times worse
than binding of BSA to PC (P 0.05).
62


4:1 PC/LPA -

o
cft
o
CL
a)
Q.
2:1 PC/LPA -
PC -
4:1 PC/PA -
2:1 PC/PA m
-------1--------1--------1--------1-------1--------1--------1--------
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40
Estimated Kd (M)
Figure 3.6.1 Estimated Kds of BSA and Liposomes with 95% Ci
Graph of the estimated Kds between BSA and the different type of liposomes.
Kds are measured in molar concentrations (M). The 95% confidence
intervals shown determine how likely it is that the interval contains the true
dissociation constant.
3.7 Liposome-Polylysine Binding
Polylysine binds strongly to liposomes containing acidic
phospholipids (e.g. PS), but does not bind to liposomes containing only PC
(Hartmann and Galla, 1978; Schwieger and Blume, 2007). To evaluate our
63


method with a second binding control, we evaluated binding of the
polylysine (150 nM) to liposomes containing PA or LPA. Similar to the BSA
and Src experiments, we used liposomes containing 2:1 PC/LPA, liposomes
containing 2:1 PC/PA, and liposomes containing PC (see method in Figure
2.5.1). Varying liposomal concentrations were added to 150 nM protein. The
log of the lipid concentration was plotted against the percent of polylysine
binding (Figure 3.7.1).
64


120
Figure 3.7.1 Polylysine Binding
Percent binding of Polylysine (150 nM) to liposomes consisting of 2:1
PC/LPA, 2:1 PC/PA or just PC. These values are an average of 3 experiments
(n = 3).
Polylysine bound exceptionally well to liposomes containing PA and
LPA (Figure 3.7.1). Our data correlates with the other studies that suggest
that polylysine binds well to all negatively charged lipids without specific
binding to any one anionic phospholipid (Hartmann and Galla, 1978;
Schwieger and Blume, 2007). The estimated Kd for poly-D-lysine binding to
65


2:1 PC/LPA liposomes was 0.00012 M (95% Cl: 0.00009, 0.00016). The
estimated Kd for polylysine binding to 2:1 PC/PA was 0.00010 M (95% Cl:
0.00006, 0.00018). The estimated Kd for Src binding to PC liposomes was
1.764 M (95% Cl: 0.87, 27.07). Due to negligible polylysine binding to PC
vesicles, we conclude that pelleting the liposomes are not trapping protein
and that the binding we observed between Src and liposomes reflects specific
interactions that may play an important role in Src in vivo regulation.
3.8 Liposome-Src Binding
Similar to the BSA and polylysine experiments, we used liposomes
containing 2:1 PC/LPA, 2:1 PC/PA, and PC alone. Various liposomal
concentrations were added to human, recombinant, non-myristoylated Src
protein (Figure 2.5.1), and binding it shown in Figure 3.8.1.
The estimated Kd for Src binding to 2:1 PC/LPA liposomes was 0.023
M (95% Cl: 0.017, 0.032). The estimated Kd for Src binding to 2:1 PC/PA was
0.0089 M (95% Cl: 0.0078, 0.0103). The estimated Kd for Src binding to PC
liposomes was 0.018 M (95% Cl: 0.014, 0.024).
66


Figure 3.8.1 Src Binding to Liposomes
Human, recombinant, non-myristoylated Src binding to liposomes consisting
of 2:1 PC/LPA, 2:1 PC/PA or PC alone. Inner graph is an enlarged image of
points between -3 and -2 M lipid.
In a similar manner (Figure 2.5.1), varying concentrations of 4:1
PC/LPA and 4:1 PC/PA liposomes were added to Src protein and binding was
quantified (Figure 3.8.2). The estimated Kd for Src binding to 4:1 PC/LPA
liposomes was 0.029 M (95% Cl: 0.021, 0.043) (Table 3.9.1). The estimated
Kd for Src binding to 4:1 PC/PA was 0.011 M (95% Cl: 0.010, 0.012). The
67


estimated Kd for Src binding to PC liposomes was 0.018 M (95% Cl: 0.014,
0.024).
Figure 3.8.2 Src Binding to 4:1 Liposomes
Percent binding of Src to liposomes consisting of 4:1 PC/LPA, 4:1 PC/PA or
just PC. Inner graph is an enlarged image of points between -3 and -2 M lipid.
68


3.9 Summary of Binding to Various Liposomes
Table 3.9.1 Estimated Kds of Src, BSA and Polylysine
Dissociation constants between Src and the varying liposome composition,
along with their corresponding confidence intervals (Cl).
Src with 2:1 PC/PA 0.009 0.0078,0.0103
Src with 2:1 PC/LPA 0.023 0.017, 0.032
Src with PC alone 0.018 0.014, 0.024
Src with 4:1 PC/PA 0.011 0.010, 0.012
Src with 4:1 PC/LPA 0.029 0.021,0.043
BSA with 2:1 PC/PA 0.021 0.017, 0.027
BSA with 2:1 PC/LPA 0.19 0.15, 0.27
BSA with PC alone 0.058 0.046, 0.074
BSA with 4:1 PC/PA 0.025 0.020, 0.034
BSA with 4:1 PC/LPA 0.25 0.19, 0.34
Polylysine with 2:1 PC/PA 0.00010 0.00006, 0.00018
Polylysine with 2:1 PC/LPA 0.00012 0.00009, 0.00016
Polylysine with PC alone 1.764 0.87,27.07
69


The estimated Kds (Table 3.9.1) have been plotted against the
respective liposomes (Figure 3.9.1). The estimated Kd's and the
corresponding confidence intervals were compared in the same manner as
the BSA data (Figure 3.6.2). The Src binding data suggests that 2:1 PC/LPA
and 4:1 PC/LPA were not significantly different since the 95% confidence
intervals overlap significantly. However, whereas it was observed in the BSA
experiments that 2:1 PC/PA and 4:1 PC/PA were not significantly different, it
is more difficult to determine for Src since the 95% intervals barely overlap.
In addition, for the Src experiments, the data suggest that Src binding to PC
was not significantly different from 2:1 PC/LPA. However, Src binds to
liposomes containing PA much better to vesicles with PC alone or with
PC/LPA (p 0.05).
70


4:1 PC/LPA -

o
cn
o
Cl
0)
Q.
>.
2:1 PC/LPA -
PC -
4:1 PC/PA -
2:1 PC/PA
-------1-------1-------1------1-------1-------1------1-------
0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 0.045
Estimated Kd (M)
Figure 3.9.1 Estimated Kd's of Src and Liposomes with 95% Confidence Interval
Dissociation constants (Kd) for the binding of human, recombinant, non-
myristoylated Src and the different type of liposomes. Kds are measured in
molar concentrations (M). The 95% confidence intervals shown determine
how likely it is that the interval contains the true dissociation constant.
71


4. DISCUSSION
4.1 280nm Absorbance Assay
The 280nm absorbance assay was initially chosen as the method for
protein quantification. It is a relatively simple method, and the fact that it
does not consume the protein during protein detection made this method
appealing. Although the data reflect that the 280nm technique was able to
detect BSA in the nM range, the results were not consistent from one
experiment to the next. Although the same stock BSA solution, the same
cuvette and the same technique were used, the calculated protein
concentration never matched that of the actual protein concentration used.
We suggest that the low protein concentrations were at the lower limit of
sensitivity of the spectrophotometer where there is significant error.
4.2 ATTO-Tag CBQCA
The ATTO-Tag CBQCA kit (Molecular Probes Inc., Carlsbad, CA) was
chosen as the second method for protein detection after the abandonment of
the 280nm Absorbance Assay. The CBQCA kit was originally developed as a
derivitization reagent for amines in proteins being separated by liquid
chromatography.
Similar to the 280nm absorbance assay, the ATTO-Tag CBQCA kit was
72


tested using BSA in the nM range. Whereas the 280nm Absorbance Assay
failed in its ability to accurately and consistently detect protein in the
nanomolar range, the CBQCA kit succeeded in both of those areas. After
averaging data from two trials (Figure 3.2.1), an accurate reproducible
regression line was obtained, with an R-squared value (a measure of variance
from the regression line) of 0.996. As this number is close to 1, the data
gathered correlated well with the best-fit line.
4.3 Effects of KCI on Releasing Protein From Liposomes
We were concerned that bound protein was not released from
pelleted liposomes and that we would record inaccurate, low protein levels
for bound protein. Although the CBQCA tag worked well in quantitating BSA
alone, when coupled with the liposome-binding assay (varying protein
concentrations), the results lacked consistency. In addition, it may be that
the liposomes are absorbing or reflecting some of the light. For example, the
negative fluorescence readings (Figure 3.3.1) suggested that there were less
protein present in the sample than the control that contained no protein. In
order to resolve these issues, another modification was made to the liposome
binding protocol: KCI protein release.
With the KCI modification, the KCI is added to the liposome pellet to
73


fully release liposome-bound protein (Schoeffler et al., 2003). With no fatty
tail allowing the protein to insert itself into the liposomal membrane, it is
expected that the main interaction between the liposome and the protein are
through electrostatic forces. The addition of 150 mM KC1 will disturb the
ionic interactions between the amino acids in the protein and the lipid head
groups. After incubation, the resuspended pellet is centrifuged again at a
slower rate so that the liposomes pellet, and the "bound protein remained in
the supernatant. Since this step was only used to clarify the solution, the
removal of all liposomes was not necessary, and so the high centrifugation
speed was not used. This "bound protein is then compared with "total
protein.
Using 2:1 PC/PA liposomes, the KC1 protein release method produced
superior results, as the bound" protein determinations were higher and
more consistent (Figure 3.4.1). As the concentration of BSA increases, more
protein will bind to the constant PA concentration; this is what we found
after the use of KC1.
The head group of PC has both positive and negative charges so it is
not surprising that BSA and Src bound to vesicles made with PC alone.
However, the affinity of protein for PC vesicles was relatively low as
74


compared with PA.
As the PS head group contains an amine with which the CBQCA tag
can react, LPA was chosen for a control. The only difference between LPA
and PA is that PA has two fatty acid tails whereas LPA only has one. Since
LPA has a negative head group, there was an increase in BSA binding as
protein or liposome concentration increased.
However, BSA bound with higher affinity to liposomes containing PA,
than to liposomes with LPA (Figure 3.4.4). In addition, BSA bound better to
liposomes with PC alone than to liposomes containing 2:1 or 4:1 PC/LPA.
However, PC bound BSA better than LPA containing vesicles. The positive
charged choline or phosphate of PC may form weak ionic bonds with BSA to
explain these results.
4.4 Comparing Varied Protein Concentration Protocol to Varied
Liposomal Concentration Protocol
The protocol using various protein concentrations was initially
chosen for a number of reasons. First, measuring protein concentration is
simple and accurate. As we assumed lipid concentration, the amount of
lipids lost during the extrusion process was not quantifiable and we used an
estimate for the final concentration of the liposomes in the method we
75


subsequently employed. Secondly, varying protein concentrations would
allow us to use up a smaller amount of protein (when we varied liposome
concentration).
Unfortunately, there were some drawbacks to using varying protein
concentrations. Thus, we changed to the use of different lipid concentrations
while keeping the protein concentration constant. One benefit of this change
is that we can more effectively compare our data to a study of Src and Ps
binding (Sigal et al., 1994). The switch to varying liposomal concentrations
produced results similar to those of Sigal et al. (1994).
4.5 Polylysine Binding
We were concerned that the airfuge may be cooling the samples and
causing aggregation of the liposomes. Liposome aggregation could result in
the trapping of protein and no protein-lipid binding actually occured.
Protein trapping would increase the apparent amount of protein binding and
artificially lower the estimated Kd. However, our airfuge was run at room
temperature (compared to most airfuges that are run in a cold box), and a
change in the temperature of the samples or the rotor after centrifugation
was not apparent. However, in order to convincingly show that protein
trapping in the liposome pellet was not an issue, we tested liposome and
76


polylysine binding. Polylysine was chosen due to its positive charges and
that it has been well documented to bind to acidic phospholipids (such as PA
and PS), but not to neutral phospholipids such as PC (Hartmann and Galla,
1978; Schwieger and Blume, 2007). A study suggested that polylysine (MW
up to 30,000) has the ability to penetrate giant liposomes (30 90 pm in
diameter) (Menger et al., 2003). However, if the polycationic polymer is too
large, it cannot pass through the membrane. The polylysine we chose had a
high molecular weight (30,000 to 70,000). The use of a big polymer and
small liposomes (100 nanometers in diameter) should have limited the
ability of polylysine to penetrate our liposomes.
Our data showed that polylysine bound equally well to liposomes
containing PA or LPA, but poorly to liposomes containing only PC (Figure
3.7.1 and Table 3.9.1). We estimated equivalent Kd's of for polylysine
binding to PC/PA and PC/LPA liposomes, but binding to liposomes
containing PC alone was 17,840 fold lower. Polylysine is a highly positively
charged polypeptide (without negative charge, except for the C-terminus), so
the observed binding of polylysine to liposomes containing acidic
phospholipids is not surprising. Polylysine is also known for its ability to
adopt all three common secondary structures, i.e. random coil, a-helix and p-
77


sheet. Carrier and Pezolet (1984) suggested that the ability of polylysine to
undergo a conformational transition toward an ordered an a-helical
structure maximizes the electrostatic interactions. This may account for the
maximal and equivalent protein binding we detected using liposomes
containing PA and LPA. We found that our data correlated well with other
published literature (Hartmann and Galla, 1978; Schwieger and Blume,
2007), which suggest that polylysine binds well with acidic phospholipids,
but not to neutral phospholipids. More importantly, the polylysine binding
data showing lack of binding to PC liposomes supports the belief that Src
binding to PC liposomes were not simply due to protein trapping in the
liposome pellet.
The low affinity between Src or BSA and liposomes containing LPA
suggestes that non-electrostatic forces inhibit protein interaction with LPA.
LPA has the same negative headgroup as PA, but without one fatty acid off
the glycerol backbone. LPA may take on a poor orientation or have too high
of mobility for optimal interaction with protein. Also, Src may interact with
the fatty acids of lipids as LPA (which lacks on fatty acid) appears to have the
lowest affinity for protein as compared to PC alone or PC/PA.
78


4.6 Calculating the Kd of Protein Binding
To calculate the dissociation constant, we used the equation:
fb = Ka[L]/(l + Ka[L]), Eq. 12
where fb is the fraction bound, [L] is the accessible lipid concentration, and Ka
is the partition coefficient or apparent association constant. The apparent
dissociation constant Kd is equal to 1/Ka. The accessible lipid concentration
is equivalent to half the total lipid concentration, since the vesicles are not
permeable to Src (Sigal et al., 1994). A study suggested that an average of
30% of lipids may be lost during extrusion (Loughrey et al., 2007).
Therefore, total lipid concentration was corrected for lipid loss (30%) during
extrusion. As expected, our Kd values (0.018M for Src to PC liposomes, to
0.009M for Src to PC:PA liposomes) are rather high when compared to those
that Sigal et al. (1994) (0.002M for Src to PC liposomes, to 0.000001M for Src
binding to PC:PS vesicles). Based on estimates (Sigal et al., 1994), this
difference could be explained by our use of non-myristoylated Src. Sigal et al.
(1994) used myristoylated Src and the myristoyl tail increases the affinity of
Src to the membrane. Since Src is suggested to bind to acidic phospholipids
largely through ionic interactions (Sigal et al., 1994), non-myristoylated Src
should bind to acidic phospholipids although with a weaker affinity.
79


However, by using a non-myristoylated Src, our data emphasize the
specificity with which Src interacts with certain anionic lipids. The Kd for
PA-containing lipids and Src is 2 fold to 3.2 fold lower than PC or PC:LPA
(respectively) liposomes (Table 3.9.1; Figure 3.9.1). Since we found Kd
values that were 10 to 1000 fold greater for the non-myristoylated Src, and
since we found only a 2 fold increase in affinity of PA over PC vesicles, our
data suggest that the myristoyl tail plays a more important role than anionic
lipid binding.
Overall, our data showed that Src bound better to all liposome
compositions than BSA. These findings may be due to the presence of an SH2
domain in Src, and a lack of it in BSA. A study has shown that an SH2 domain
can play a role in membrane binding, and a mutation in this region may
reduce the ability of the protein to associate with the membrane (Hong et al.,
2009).
4.7 Compare Vesicle Sedimentation Results to PIP Strip Results
Our results showed that BSA bound with the highest affinity to PC/PA
liposomes, and with the lowest affinity to PC/LPA liposomes. Src showed the
highest affinity for PC/PA liposomes, and a nonsignificant trend between
liposomes containing PC/LPA or PC alone. This correlates well with the PIP
80


Strip data (Section 1.9), which suggests that Src has the higher specificity for
PA, and negligible specificity for LPA lipids. In a typical example of PIP Strip
results, PA binding to Src was 6.5 arbitraty units, whereas Src binding to PC
was zero, and LPA (-0.5) was lowest. Though both our data and the PIP strip
data reveal similar results, it should be noted that the large unilamellar
vesicle liposomes used in our protocol is a better representation of the
membrane than the phospholipid monolayer on paper used in the PIP strip
experiments.
4.8 Physiological Relevance of PA and Src
Two different lipid ratios were used to produce the liposomes; for
example, 2:1 PC/PA and 4:1 PC/PA. Comparisons of the Kds between 2:1 and
4:1 liposomes (using the same lipid) reveal that there is only a slight
difference in affinity with protein. Since a higher PA concentration suggests
that there are more sites for protein to bind, we expected to see a difference
in affinity. While it appears that Src binds with a higher affinity to the 2:1
concentrations (when comparing Kds) than the 4:1 concentrations (Figure
3.9.1), the 95% confidence interval ranges overlap, so the two are not
significantly different. A study looking at SNARE syntaxin protein binding to
membranes also reported only minor differences in binding of the syntaxin
81


protein to liposomes containing 20% PA and 30% PA (Mendonsa and
Engebrecht, 2009; Figure 1). It is possible that a higher concentration of PA
may increase the binding, but if PA forms patches, the concentration of PA in
2:1 or 4:1 vesicles may be similar. The presence of PA patches increases the
area of negative charge seen by the protein, thereby increasing the attraction
of the protein to the membrane. In addition, it has been observed that
proteins often tend to enhance the ability of the lipid membrane to form
patches (Hinderliter and May, 2006).
It has also been suggested that protein-induced membrane domains
(PA patches) may affect cell signaling by organizing signal transduction
components within the membrane and change reaction rates (Wanaski et al.,
2003). It is possible that once a protein binds to the membrane, more PA is
drawn to the area forming a patch. The patch is then able to attract more
protein, working in a cooperative manner.
Comparisons of Kd's between PC/PA liposomes and PC/LPA
liposomes find that Src has a higher affinity for PA. Interestingly, the only
difference between PA and LPA is that LPA has one less fatty acid tail. High
affinity phosphatidic acid-protein interaction has been observed for many
proteins in mammals, plants and yeast cells.
82


When comparing the PA binding region of the different proteins, no
consensus sequence is apparent. It is not surprising though, that one general
feature that constantly arises is the presence of basic amino acid residues.
In addition, it appears that hydrophobic residues are important as
well (Stace and Ktistakis, 2006, Jones et al. 2005)). This fact may explain why
LPA has surprisingly low affinity for protein it lacks one very hydrophobic
fatty acid tail. However, Src is thought by some to only interact with the
surface hydrophilic head group (Mclaughlin and Aderem, 1995).
The basic residues and hydrophobic residues present in the proteins
studied, along with the shape of PA may explain why Src has a higher affinity
for PA over other lipids like PS or PC. PA is the only anionic phospholipid
with a pronounced cone shape under physiological conditions. In most
phospholipids, the head group is relatively large compared to the fatty acid
tails. PA has a small head group (phosphate), but the same fatty acid tails as
other glycerophospholipids, giving PA its pronounced inverted cone shape.
Cone shaped lipids facilitate protein penetration into the membrane by
reducing lipid head group packing, forming favorable insertion sites in the
headgroup region of the lipid bilayer (Brink-van der Laan et al., 2004). Thus
there is reason to believe that Src might interact with the hydrophobic tails
83


of PA.
In addition, the single acyl chain of LPA and a higher mobility may
explain why BSA and Src had a higher affinity for liposomes containing only
PC than for liposomes containing LPA. Since LPA only has a single acyl chain,
the molecular area that LPA possesses is smaller than that of PC and PA. A
smaller molecular area results in an increase in charge density of the lipid
(Kooijman and Burger, 2009], which may lead a decrease in the dissociation
of hydrogen ions from the oxygen molecules of the phosphate group. The
inability of the oxygen groups on LPA to dissociate hydrogen may lead to an
inability to bind to Src.
In addition, it is possible that the docking of basic protein domains on
PA may be followed by insertion of hydrophobic protein domains into the
hydrophobic interior of the lipid bilayer. This hydrophobic interaction has
been observed for the GTPase dynamin, where after dynamin binds to
monolayers containing PA, there was considerably more insertion compared
to the binding of dynamin to monolayers containing other negatively charged
phospholipids (Burger et al., 2000]. It is possible that the single acyl chain
and the negative head group of LPA allows for tighter packing of the PC/LPA
liposomes which would prevent BSA or Src interaction with the hydrophobic
84


core. While the double fatty acid chains of most phospholipids may be
sufficient in keeping other lipids from crowding it, the single acyl chain of
LPA may not be as effective, resulting in tighter lipid packing. The tight
packing limits the amount of negative charges seen by the protein. In
addition, the tight packing encourages heavier interaction between the acyl
chains of PC and LPA, thereby limiting hydrophobic interactions as well.
On the other hand, with liposomes that contain just PC, the lipid
interactions are not as tight due to the double acyl chain and the big head
group, therefore, it is possible that the proteins are still able to have limited
binding with the membrane through hydrophobic interactions. Of course,
the positive and negative charges of the PC headgroup may also interact with
BSA or Src. While acidic phospholipids are important in drawing the
proteins to the membrane as observed with PA and PS, our data suggests that
there are other factors that may affect the affinity that lipids have with the
protein.
4.9 Conclusion
Using Sigal et al.s (1994) work as reference, we initially hypothesized
that acidic phospholipids were responsible for binding to and activating Src.
If Src does specifically bind to PA, then the interaction may activate Src.
85


There are data showing PA addition to Xenopus oocyte stimulates Src
(upublished data, Stith et al.). The vesicle sedimentation assay has been used
to quantify the binding affinity between Src and liposomes that contained
either: PC, PC and PA, or PC and LPA (another acidic phospholipid). We have
modified the vesicle sedimentation assay by switching to altering the
liposomal concentration, and using the CBQCA tag for protein quantification.
This method produced similar results to that of Sigal et al. (1994), and
allowed the calculation of Kd's.
The estimated Kd's from this experiment suggests that Src has a 2 3
fold higher affinity to PA than the other lipids tested. The higher affinity
supports the idea that the PA increase at fertilization may be contributing to
Src activation as hypothesized.
Our data showed that Src bound better to liposomes that were made
up of only PC than to liposomes that contained both PC and LPA. Our data
analysis suggests that although the negative charges in acidic phospholipids
do affect the binding affinity of Src to the membrane (as observed in Sigal et
al. (1994)), there are other factors that may affect the binding of Src to the
membrane as well. We suggest that one of the reasons Src is not able to bind
as well to LPA is mainly due to its single acyl chain. The single acyl chain
86


does not allow for sufficient separation from the other lipids, shielding the
negative head group from the protein. In addition, the tight packing does not
allow for as many hydrophobic interactions between membrane lipids and
Src, as that seen with liposomes that contain only PC. In addition, we suggest
that the higher affinity of Src for PA involves more than just the electrostatic
interactions, but that cooperativity plays a role. As mentioned, the amount of
phosphatidic acid in a typical cell is usually very low, and at fertilization, an
increase in PA is observed. When the amount of PA is low, it may be
ineffective in interacting with Src, however, when PA increases, patches of PA
may form and increase the affinity to Src, which would attract Src to the
membrane.
Recent research, using Xenopus frogs, suggest that PLC-y is activated
through Src Tyrosine Kinase (Sato et al., 2003). However, it is still unclear as
to how sperm binding leads to the activation of Src. Stith et al. (unpublished
data) have suggested that the observed increase in PA may activate Src at
fertilization. Our binding experiment has shown that Src has a relatively high
affinity to PA even without the presence of a myristoyl tail. The interaction
between Src and PA may be sufficient in attracting Src to the membrane.
Whether it is PA that directly activates Src or not, is yet to be determine
87


Appendix A
Materials
Chemical Name Supplier Product Number
l,2-dioleoyl-s/?-glycero-3-phospho -
L-serine (DOPS] Avanti 840035C
l-oleoyl-2-hydroxy-sn-glycero- 3-phosphate (LPA) Avanti 857130
l-Palmitoyl-2-oleoyl-sn-glycero- 3-phosphate (POPA) Avanti 840857C
l-Palmitoyl-2-oleoyl-sn-glycero- 3-phosphocholine (POPC) Avanti 850457C
ATTO-Tag CBQCA kit Molecular Probes A-2333
Bovine Serum Albumin Sigma A-0281
HEPES Calbiochem 391338
MOPS Sigma M-8899
Poly-D-Lysine Sigma P-7886
Potassium Chloride Sigma P3911
p60c-src Upstate Biotech 14-117
Sodium Chloride OmniPur 7760
Sucrose Calbiochem 573113
Triton X-100 Calbiochem 648463
Outside Buffer is made up of 100 mM KC1 & 1 mM MOPS (pH 7)
Inside buffer is made up of 176 mM Surcrose & 1 mM MOPS (pH 7)
Binding Buffer is made up of 20 mM HEPES (pH 7], 100 mM NaCl, 0.1% Triton
88


Appendix B
Lipid Concentration Calculations
As noted, we expected a 30% lipid loss during the extrusion process. If we
used a total lipid concentration of 12 mM, we would expect for there to be
about a lipid loss of about 4 mM after extrusion. Therefore, our final lipid
concentration would be stated as 8 mM instead of 12 mM.
280nm Calculations
0.1% (1 mg/ml) of BSA has an absorbance of 0.667 in a cuvette with a 1cm
pathlength
The molar extinction coefficient at 280 nm for BSA is approx. 43,824 M^cnr1
Spec Abs Reading = X = X
0.667 0.1% 15.06 pM
1 mg/ml of BSA = (1 g/L of BSA) x (lmol/66400g) = 15.06uM of BSA
Where X equals the concentration of BSA.
For example, if our protein sample gave us a spectrophotometer absorbance
reading of 0.005, we can plug in this value into the equation and calculate the
estimated concentration.
0.005 = X
0.667 15.06 pM
Solving for X, we get 0.667(X) = 75 xlO-9
X = 113 x 10'9 or 113 nM
89


Fluorometer Calculations
To calculate protein concentrations using the the CBQCA tag and fluormeter,
a standard line was made, following the protocol in Section 2.7, where
various protein concentrations were made, tagged, and its fluorescence was
measured. The fluorescence (y-axis) was plotted against the amount of
protein used (x-axis) (Figure 3.2.1). A best-fit line is created using the data
collected. To quantify the concentration of protein present in a sample, the
fluorescence is measured, and this value is plugged into the equation, and the
amount of protein is calculated. The amount of protein can then be divided
by the volume of the sample, the protein concentration can be calculated.
For example, we calculated a best-fit line of y = 1.0031x 11.732 for BSA. If
we were to measure a fluorescence of 100 fsu for our sample, we would
simply plug 100 in to the equation (y) and solve for x to calculate protein
amount.
100 fsu = 1.0031x- 11.732
Solving for x, we see that the amount of BSA present in the sample is 111.39
ng. This concentration can be calculated by dividing the protein amount by
the volume of the sample.
90