SOLUTION STRUCTURE AND CHARACTERIZATION OF LIPID BINDING OF THE NOXO1β PX DOMAIN
By
NICOLE YOLANDA DAVIS
A Dissertation Submitted to the Graduate Faculty of
WAKE FOREST UNIVERSITY GRADUATE SCHOOL OF ARTS AND SCIENCES
In Partial Fulfillment of the Requirements
For the Degree of
DOCTOR OF PHILOSOPHY
In the Biochemistry and Molecular Biology Program
August 2010
Winston-Salem, North Carolina
Approved by:
Dr. David A. Horita, Ph.D., Advisor
Examining Committee:
Dr. Sean Reid, Ph.D., Chair
Dr. Tom Hollis, Ph.D.
Dr. Greg Kucera, Ph.D.
Dr. Linda C. McPhail, Ph.D.
ACKNOWLEDGEMENTS
I would like to thank my advisor, David Horita and my committee members (Sean Reid, Linda McPhail, Tom Hollis and Greg Kucera) for their help and guidance throughout my time here. I know my project (along with myself) was difficult at times and I thank you for your patience. I would also like to thank some current and former members of the Horita lab. Joel, I cannot thank you enough for the work you did early on with NOXO1. You helped push the project along while I worked on other projects. I appreciate all the times you immediately stopped to answer any questions I had in lab. Kai, I want to thank you for the little bits of encouragement you gave me. When times were really stressful and frustrating, remembering times when you congratulated me for what seemed to be the most mundane task helped me push through. Lindsay, I am sad we did not get more time together in lab, but I am glad we got to know each other.
To all of my friends: I would never have made it through graduate school without you. From personal to professional, I cannot thank you enough for your constant support. To H!, you know too much, we have to stay friends forever. To the martini night girls (Keri, Carla, Heather and Karon), I will miss the conversations and $5 martinis we had over the summer, they are times I will always fondly remember. To E, you are a sick man, SICK! but I thank you for being there for me in times of laughter and times of tears. To Joy, thank you for all of the useful and humorous career advice you shared with me. I still believe you need to write a book on life post-grad school. Erin, thank you for simultaneously telling people I am your smartest friend yet treating me like I’m a normal person. You are always someone I look to for advice and you have never let me down. To Valerie and Amanda, thanks for the constant emails, gossip and vacation plans. I am glad we’ve reconnected since high school. I would like to thank Jim Gaffigan for being a constant source of inspiration, humor…and bacon.
To my parents, thank you for paying for the first fourteen years of school while I “paid” for the last eleven. Thank you for letting me be who I am, for supporting me and being my biggest cheerleaders. I hope I have made you proud. To my mom thank you for trying to understand what I am researching; you understand it more than you realize. To my dad, I want to thank you for asking me random questions dealing with any area of science and thinking I should know the answer. To my brother, I’ve enjoyed getting to know you as an adult and learning we share a similar, twisted sense of humor. I also thank you for succinctly stating that all I do is just stick goo in a magnet.
ii TABLE OF CONTENTS
Page LIST OF ILLUSTRATIONS…………….……………………………………………...iv
LIST OF ABBREVIATIONS…………………...... vi
ABSTRACT……………………………………………………………………………...x
CHAPTER I. INTRODUCTION……………………………………………….1
CHAPTER II. NOXO1β PX BINDS TO PI(4,5)P2 IN ADDITION TO NEUTRAL MEMBRANE LIPIDS……………………………..29
CHAPTER III. SOLUTION STRUCTURE OF NOXO1β PX………….………76
CHAPTER IV. DISCUSSION…………………………………………………..131
APPENDIX……………………………………………………………………………..142
CURRICULUM VITAE………………………………………………………………..152
iii LIST OF ILLUSTRATIONS
CHAPTER I Page Figure 1. Components and Domain Structure of the Phagocytic NADPH 5 Oxidase
Figure 2. NADPH Oxidase Isoforms 10
Figure 3. Domain Comparison of NOXO1 and p47phox and Alignment of 15
NOXO1 PX Isoforms
CHAPTER II
Figure 1. Dot Blots for NOXO1β PX, p40phox PX and p47phox PX 43
Figure 2. SDS-PAGE Gel of GST-NOXO1β PX Binding to PA LUVs. 46
Figure 3. 12% SDS-PAGE Gel of NOXO1β PX Binding to PI(5)P LUVs 49
Figure 4. The Magnetic Bead Assay is not a Reliable Method for 51 Measuring Lipid Binding
Figure 5. PA LUVs Do Not Dissociate from a GST-p47phox PX Coated 54 Surface
Figure 6. Increasing the NaCl Concentration Decreased Binding of 58 NOXO1β PX to a POPC/POPE Surface
Figure 7. The Addition of Phosphate Decreases Binding of NOXO1β 60 PX to a POPE/POPC Surface
Figure 8. The Addition of Phosphate Decreases Binding of NOXO1β 62 PX to a 3% PI(4,5)P2 Surface
Figure 9. NOXO1β PX Binds to Background Phospholipids 65
Figure 10. NOXO1β PX Binds to PI(4,5)P2 67
phox phox Figure 11. P40 PX and p47 PX bind to PI(3)P and PI(3,4)P2 Respectively 70
Figure 12. NOXO1β PX Does Not Bind to Low Levels of Anionic Lipids 72
iv CHAPTER III
Figure 1. 15N, 1H HSQC of NOXO1β PX 88
Figure 2. Predicted secondary structure of NOXO1β PX by TALOS+ 90
Figure 3. NOXO1β PX is a monomer in solution 93
Figure 4. Solution Structure of NOXO1β PX 96
Table 1. NOXO1β PX Restraints and Structure Statistics 98
2 Figure 5. S , τe and Rex of NOXO1β PX 102
Figure 6. NOXO1β PX does not have a stable core 105
Figure 7. POPC Nanodiscs are a suitable membrane mimetic for 108 NOXO1β PX
Figure 8. NOXO1β PX exhibits non-specific binding to a POPC nanodisc 110
Figure 9. NOXO1β PX diC8-PI(4,5)P2 titration 113
Figure 10. NOXO1β PX does not undergo many chemical shift changes 116 upon addition of diC4-PI(3,4)P2
Figure 11. NOXO1β PX undergoes chemical shift changes upon addition 119 of diC4-PI(4,5)P2
Figure 12. NOXO1β PX does not contain conserved residues for binding to PIs 122 phosphorylated at the D3 position
Figure 13. NOXO1β PX contains conserved residues for binding to PIs 125 phosphorylated at the D4 position
Figure 14. Candidate residues of NOXO1β PX for binding to PIs 128 phosphorylated at the D5 position
APPENDIX
Figure 1. Domain Organization of RIP1 and Constructs of RIP1 148 Used For Binding Assays
v LIST OF ABBREVIATIONS
AIR, autoinhibitory region
ApoA1, Apolipoprotein A-1
BSA, bovine serum albumin
CGD, chronic granulomatous disease
CHAPS, (3-[(3-cholamidopropyl)-dimethylammonio]-1-propane sulfonate
CISK, cytokine-independent survival kinase
CMC, critical micelle concentration
DCN, 2H13C15N isotopically labeled protein
DCN-ILV, 2H13C15N-1H(Iδ1,L,V) isotopically labeled protein
DCN-ILV-F, 2H13C15N-1H(Iδ1,L,V)-15N Phe isotopically labeled protein
DD, death domain
DHPC, 1,2-dihexanoyl-sn-glycero-3-phosphocholine
DMPC, 1,2-dimyristoyl-sn-glycero-3-phosphocholine
DN, 2H15N isotopically labeled protein
DUOX(1-2), dual oxidase (1-2)
E. coli, Escherichia coli
FAD, flavin adenine dinucleotide gp91phox (NOX2), 91-kDa phagocytic oxidase component
GST, glutathione S transferase
ID, intermediate domain
IP3, inositol-3-phosphate
vi KD, kinase domain
LPA, lysophosphatidic acid
LPC, lysophosphatidyl choline
LUV, large unilamellar vesicles
MAPK, mitogen activated protein kinase
NADPH, nicotinamide adenine dinucleotide phosphate
NMR, Nuclear Magnetic Resonance
NOX(1-5), NADPH oxidase (1-5)
NOXA1, NADPH oxidase activator 1
NOXO1, NADPH oxidase organizer 1
- O2 , superoxide anion
OG, n-octyl-β-D-glucopyranoside p22phox, 22-kDa phagocytic oxidase component
p40phox, 40-kDa phagocytic oxidase component
p47phox, 47-kDa phagocytic oxidase component
p67phox, 67-kDa phagocytic oxidase component
PB1, PHOX Bem1 domain
PBS, phosphate buffered saline
PC, phosphatidylcholine
PE, phosphatidylethanolamine
PG, phosphatidylglycerol
PHOX, phagocytic oxidase
vii PI, phosphatidylinositols
PI(3)P, phosphatidylinositol-3-phosphate
PI(4)P, phosphatidylinositol-4-phosphate
PI(5)P, phosphatidylinositol-5-phosphate
PI(3,4)P2, phosphatidylinositol-3,4-bisphosphate
PI(3,5)P2, phosphatidylinositol-3,5-bisphosphate
PI(4,5)P2, phosphatidylinositol-4,5-bisphosphate
PI(3,4,5)P3, phosphatidylinositol-3,4,5-triphosphate
PI3K, phosphatidylinositol-3-kinase
PIP, phosphatidylinositol phosphate
PLD, phospholipase D
PR, proline-rich-region
PS, phosphatidylserine
PX, phox homology
RHIM, receptor interacting protein homotypic interaction motif
RIP1, receptor interacting protein 1
ROS, reactive oxygen species
RU, response units
S1P, sphingosine-1-phosphate
SDS-PAGE, sodium dodecylsulfate polyacrylamide gel electrophoresis
SH3, Src homology 3 siRNA, small interfering RNA
S/N, signal-to-noise
viii SPR, surface plasmon resonance
TBST, Tris-buffered saline with Tween
Tks4, tyrosine kinase substrate with four SH3 domains
Tks5, tyrosine kinase substrate with five SH3 domains
TNF, tumor necrosis factor
TNFR1, tumor necrosis factor receptor 1
TRADD, tumor necrosis factor receptor 1 associated death domain protein
TRAF2, tumor necrosis factor receptor 1 associated factor 2
TROSY, transverse relaxation-optimized spectroscopy
Trx, thioredoxin
ix ABSTRACT
Nicole Y. Davis
SOLUTION STRUCTURE AND CHARACTERIZATION OF LIPID BINDING OF THE NOXO1β PX DOMAIN
Dissertation under the direction of
David A. Horita, Ph.D., Associate Professor of Biochemistry
The NADPH oxidases are multiprotein enzyme complexes which catalyze the
formation of reactive oxygen species (ROS). The most studied of these is the phagocytic
NADPH oxidase (PHOX), which is found in neutrophils, macrophages and monocytes.
The PHOX enzyme plays a role in the host immune response through production of ROS.
Deficiencies in components of the phagocytic NADPH oxidase lead to weakened host
defense, as seen in chronic granulomatous disease. There are also non-phagocytic
homologues of the phagocytic NADPH oxidase (NOX) found in a broad range of tissues.
These homologues play roles in host defense, signaling, regulation of cell growth and
metabolism. Aberrant NOX function has been implicated in several disease states such
as certain cancers and Alzheimer’s disease. Increased expression and dysregulation of
the non-phagocytic NADPH oxidase has been indicated to have a role in abnormal
growth in cancers.
Three cytosolic components p40phox, p47phox and NOXO1 (the nonphagocytic
homologue of p47phox) contain a PX (Phox homology) domain which is important in lipid binding. We have solved the solution structure of NOXO1β PX and characterized the lipid binding of NOXO1β PX using dot blots and surface plasmon resonance (SPR). We
x found that NOXO1β PX binds to neutral membrane lipids which has not been reported for any other PX domain. NOXO1β PX also has a modest increase in binding to the phosphorylated phosphatidylinositols. NOXO1β PX bound the most to PI(4,5)P2, which does not agree with published literature on NOXO1β PX’s lipid binding.
The structure of NOXO1β PX resembles that of other known PX structures.
NOXO1β PX contains three anti-parallel β-strands surrounded by four α-helicies. Using the structure of NOXO1β PX along with known residue-specific PX-PIP binding data, we have identified candidate residues in NOXO1β PX that can potentially bind to the D4 and D5 phosphate of the inositol headgroup. Additionally, we have identified that structurally, NOXO1β PX should not bind to PIs phosphorylated at the D3 position.
Preliminary lipid experiments using solution-state NMR showed changes specific to
PI(4,5)P2 and not to PI(3,4)P2, again suggesting that NOXO1β PX preferentially binds to
PI(4,5)P2 and does not bind to PIPs phosphorylated at the D3 position.
The goal of this dissertation is to understand the mechanisms of phospholipid targeting and membrane binding by NOXO1β PX. Results of these studies will provide additional insight into the structural basis for regulation of NADPH oxidases by the PX domain.
xi Chapter I
INTRODUCTION
Reactive Oxygen Species and the Phagocytic NADPH Oxidase
Reactive oxygen species (ROS) are a family of small molecules derived from
molecular oxygen. These ROS species are generally derived via a string of reactions
- from superoxide (O2 ). Superoxide itself is not very stable and rapidly dismutates or
reacts with other molecules to form other ROS species such as hydrogen peroxide (H2O2)
and hypochlorous acid (HOCl) (Babior 2004, Borregaard et al. 1993, Hampton et al.
1998, Dallegri et al. 1986, Thomas et al. 1988 and Shepherd 1986). ROS can interact with a wide range of biological molecules including proteins, lipids and nucleic acids.
Although interaction of ROS with these biological molecules can damage the molecules,
one beneficial role for ROS was found in host defense involving neutrophils.
Before the phagocytic oxidase (PHOX) system was identified, it was observed that neutrophils had increased oxygen consumption (Baldridge and Gerard 1933).
Although it was first believed that the source of oxygen consumption was a result of increased mitochondrial respiration, experimental data indicated that the oxygen comsumption was due to another source within the neutrophil (Sbarra and Karnovsky
1959). Over several years, investigators identified the PHOX enzyme complex as the source of this increased oxygen consumption.
Upon infection, neutrophils are recruited to the site of infection. Once the invading pathogen has been engulfed in the phagosome, a series of oxygen-dependent and oxygen-independent events take place. The oxygen-independent response involves
1 chemotaxis (Schall and Bacon 1994), phagocytosis and degranulation and the release of
degradative enzymes into the phagosome (Wiernik 1985). The oxygen-dependent
function is termed the respiratory burst. The respiratory burst catalyzed by the PHOX
enzyme is shown in the following reaction:
+ - + NADPH + 2O2 Æ NADP +2O2 + H
Superoxide is produced by the reduction of molecular oxygen via the transfer of an electron from NADPH (nicotinamide adenine dinucleotide phosphate). As previously mentioned, the superoxide anion is highly reactive and is readily converted into other
ROS to aid in the killing of the pathogen.
Components and Interactions of the Phagocytic NADPH oxidase
The PHOX complex is a multiprotein enzyme complex composed of both
membrane-bound and cytosolic components. The membrane bound components are two
integral membrane proteins, gp91phox and p22phox. The cytosolic proteins are p47phox, p67phox and p40phox, which are present as a trimeric complex (Dusi et al. 1996, Heyworth
et al. 1991), as well as the small GTPase Rac. Functional oxidase is dependent on
translocation and assembly of the cytosolic components at the membrane (Baboir 2004).
The membrane-bound components are a heterodimer of gp91phox (or NOX2,
phox NADPH Oxidase 2) and p22 and are collectively called flavocytochrome b558.
Gp91phox contains seven transmembrane helicies as well as cytosolic binding sites for
FAD (flavin adenine dinucleotide) and NADPH (Vignais 2002, Finegold et al. 1996,
Doussiere et al. 1993). There are also two coordinated heme groups present in gp91phox that are thought to help facilitate single electron transfer from NADPH to molecular
2 oxygen (Cross and Segal 2004, Doussiere et al. 1996). Gp91phox is also heavily
glycosylated (Harper et al. 1985). Gp91phox requires dimerization with p22phox, the other
membrane-bound component, in order to be able to facilitate electron transfer from
NADPH to molecular oxygen. P22phox provides sites of interaction with both the
membrane-components as well as the cytosolic components. P22phox also has two
potential phosphorylation sites present on cytosolic portions of the protein that may act as
potential sites of regulation (Lewis et al. 2010 and Lewis 2007).
The cytosolic components are p67phox, p47phox, p40phox and the small GTPase Rac.
In the resting state, the cytosolic components are sequestered away from the membrane
bound components, with p67phox, p47phox, p40phox in a trimeric complex. Upon activation of the innate immune response, a series of activation events trigger the translocation of the cytosolic components to the membrane where they can interact with both the
flavocytochrome and the membrane.
P47phox contains an N-terminal PHOX homology (PX) domain, two tandem Src- homology 3 domains (SH3), a domain containing multiple serine phosphorylation sites termed the autoinhibitory domain (AIR) and a C-terminal proline-rich region (PR)
(Figure 1). In resting neutrophils, p47phox exists in an autoinhibited form; the C-terminal
tail of the protein folds back upon itself, binding with the SH3 domains (Figure 1)
(Groemping et al. 2003). In its autoinhibited form, p47phox cannot interact with the flavocytochrome or membrane. Upon stimulation and activation of the neutrophil, p47phox is phosphorylated on multiple serine residues in the AIR which in turn causes the
protein to unfold from its inhibited conformation and allows it to interact with the
3
Figure 1. Components of the Phagocytic Oxidase. A schematic of the phagocytic
NADPH oxidase membrane-bound and cytosolic components. The cytosolic units
(p47phox, p67phox and p40phox) are shown with their domain organization. Protein-protein and protein-lipid interactions within the PHOX complex are indicated by arrows. A detailed description of the interactions is included in the text.
4
Flavocytochrome b558
Outside
gp91phox p22phox Membrane
Inside
PI(3,4)P2, PA p47phox PX SH3 SH3 AIR PR
RAC p67phox TPR AD SH3 PB1 PR SH3
PI(3)P p40phox PX SH3 PB1
5 flavocytochrome (Ago et al. 2003). P47phox is necessary for translocation of the cytosolic
components to the flavocytochrome. Studies with p67phox-deficient neutrophils still showed translocation of p47phox to the membrane whereas in p47phox-deficient neutrophils,
p67phox did not translocate to the plasma membrane (Heyworth et al. 1991).
P67phox contains an N-terminal four tetratricopeptide repeat (TPR) followed by an
activation domain (AD) which is shown to bind to gp91phox and possibly aide electron transfer through the oxidase (Han et al. 1998, Han and Lee 2000). P67phox also contains two SH3 domains flanking a PB1 (PHOX and Bem1) domain as well as a central PR region (Figure 1).
P40phox is the shortest of the three cytosolic components. Like p47phox, p40phox contains an N-terminal PX domain. P40phox also contains an SH3 and PB1 domain
(Figure 1). The function of p40phox within the PHOX complex has remained
controversial. Multiple groups have shown that p40phox both activates and inhibits the
PHOX complex (Ellson et al. 2006, Lopes et al. 2004, Matute et al. 2005, Sathyamoorthy
et al. 1997, Suh et al. 2006).
In the activated PHOX complex are multiple protein-protein and protein-lipid
interactions (Figure 1). Interactions between p47phox and p22phox occur through the SH3
domains of p47phox to the C-terminal tail of p22phox. Another protein-protein interaction
occurs between p47phox and p67phox. This interaction occurs through a proline-rich region
at the C-terminal tail of p47phox and a C-terminal SH3 domain of p67phox. P67phox also has
a protein-protein interaction with the flavocytochrome. An activation domain in p67phox interacts with a cytosolic portion of gp91phox. The AD is thought to aid in the transfer of
electrons through the flavocytochrome and is required for activation of the NADPH
6 oxidase (Han et al. 1998, Nisimoto et al. 1999). P67phox also interacts with Rac through
its TPR domain (Koga et al. 1999). The PB1 domain on p67phox interacts with the C-
terminal PB1 domain in p40phox (Nakamura et al. 1998). The protein-lipid interactions
occur through the PX domains of both p40phox and p47phox. The PX domain of p40phox preferentially interacts with phosphatidylinositol-3-phosphate (PI(3)P) (Ellson et al.
2001, Kanai et al. 2001) while the PX domain of p47phox preferentially interacts with
PI(3,4)P2 (Kanai et al. 2001).
Chronic Granulomatous Disease
Chronic Granulomatous Disease (CGD) is an inherited primary
immunodeficiency that is a result of defects in components of the PHOX complex.
Patients with CGD have nonfunctional, reduced amounts or a complete absence of the
NADPH oxidase component (Babior et al. 1991). Patients with this disease are highly
susceptible to chronic infections due to a decrease or loss in their ability to produce ROS
through the PHOX complex.
Mutations in the gene encoding gp91phox are the most common, making up ~65%
of all CGD cases (Jurkowska et al. 2004). Mutations in the p47phox gene result in ~25%
of CGD cases, while mutations in p22phox, p67phox and p40phox make up the remaining
cases (Matute et al. 2009). Most of the identified mutations result in reduced levels or a complete absence of the affected component. Of the cases that result in normal protein levels, most of the mutations are found in gp91phox, with one mutation each in p22phox and
p67phox (Heyworth et al. 2003). Although there are no currently diagnosed CGD patients
with normal levels of p47phox, several CGD patients who have an absence of p47phox are
7 heterozygous, with one allele containing a mutation leading to a premature stop codon
and therefore absence of the p47phox component. The other allele contains a point mutation that predicts the amino acid change R42Q (Noack et al. 2001). R42 is a highly conserved residue among PX domains. The mutation R42Q diminished phosphoinositide
(PI) binding to the p47phox PX domain (Kanai et al. 2001). These cases of CGD highlight
the importance of each component in the PHOX complex as well as the importance of
their interactions in producing a fully functional NADPH oxidase.
Non-Phagocytic Oxidase Homologues
Over the past decade, homologues to several components of the PHOX system
have been identified. Multiple homologues of gp91phox have been identified in several
tissues. These homologues are classified into three groups according to their domain
similarity to gp91phox, which is also referred to as NOX2. In the first group are NOX1
(Suh et al. 1999), NOX3 (Lambeth et al. 2000) and NOX4 (Geiszt et al. 2000). NOX1,
NOX3 and NOX4 all have an overall domain structure similar to that of gp91phox. In the
second group is NOX5 (Banfi et al. 2001), which contains an additional N-terminal EF
hand-containing extension. This calmodulin-like domain of NOX5 contains four calcium
binding sites and has been shown by Banfi et al. (2001) to require calcium for activation.
The last group includes DUOX1 (Lambeth et al. 2000) and DUOX2 (Edens et al. 2001).
DUOX1 and DUOX2 build upon the NOX5 structure, having the N-terminal calmodulin-
like domain with an additional extracellular peroxidase homology domain. The
peroxidase homology domain is homologous to myeloperoxidase although it differs in
residues that are thought to be critical for the activity of myeloperoxidase (Zeng and
8
Figure 2. NADPH Oxidase Isoforms. Schematic representation of the human NADPH oxidase isoforms and the components necessary for an active complex. A. NOX1 requires p22phox, NOXO1 (or p47phox), NOXA1 (or p67phox) and Rac. B. NOX2
(gp91phox) requires p22phox, p47phox (or NOXO1), p67phox (or NOXA1), p40phox and Rac.
C. NOX3 requires p22phox and NOXO1. D. NOX4 requires only p22phox. E. NOX5 does not appear to require additional subunits and is activated by Ca2+ through an additional
N-terminus calmodulin-like domain that is not seen in NOX1-4. F. DUOX1/DUOX2 do not appear to require additional subunits but requires Ca2+ for activation through an N- terminal calmodulin-like domain as well as possessing an additional domain that appears to have a peroxidase-like function.
9
A. B. C.
NOX1 p22 NOX2 p22 NOX3 p22
Phox NOXO1 Rac NOXA1 NOXO1 Rac p67Phox p47 p40Phox
D. E. F. DUOX1 NOX4 p22 NOX5 DUOX2
Ca2+ Ca2+
10 Fenna 2002). This peroxidase domain in DUOX1 and DUOX2 does appear to catalyze
peroxide-dependent peroxidative reactions (Edens et al. 2001). NOX1-4 all appear to
require dimerization with p22phox while NOX5 and DUOX1/2 appear to be functional as
monomers (Figure 2) (Bedard and Krause 2007).
In addition to the non-phagocytic homologues of the gp91phox subunit, homologues to both p47phox and p67phox have been identified (Banfi et al. 2003, Geiszt et al. 2003, Takeya et al. 2003, Cheng and Lambeth 2004). The p47phox homologue is
called NOXO1 (Nox Organizer 1). NOXO1 has been shown to activate NOX1, NOX2
and NOX3 (Figure 2). With NOX3, only NOXO1 was required to obtain high levels of
oxidase activation; addition of NOXA1 (Nox Activator 1), the p67phox homologue, did not significantly increase superoxide generation (Cheng and Lambeth 2005, Ueno et al.
2005, Ueyama et al. 2006). With NOX1, NOXO1 and NOXA1 were required to generate high levels of superoxide in a stimulus-dependent manner (Cheng and Lambeth 2004,
Takeya et al. 2003) (Figure 2). NOXO1 has also been shown to interact with components of both the phagocytic and non-phagocytic NADPH oxidase systems.
NOXO1 interacts with the C-terminal tail of p22phox through its tandem SH3 domains.
NOXO1 can also interact with both NOXA1 and p67phox through its proline-rich region
(Takeya et al. 2003).
Recently, other proteins involved in NADPH oxidase activity that were described
as “p47phox-related” were identified from their involvement of NOX1-derived ROS
generated in a colon cancer cell line (Diaz et al. 2009). These homologues, named Tks4
(tyrosine kinase substrate with four SH3 domains) (Courtneidge 2003 and Courtneidge et
al. 2005) and Tks5 (tyrosine kinase substrate with five SH3 domains) (Lock et al. 1998)
11 contain an N-terminal PX domain with either four (Tks4) or five (Tks5) SH3 domains.
Both Tks4 and Tks5 were able to support NOX1 and NOX3 activity, but not NOX2 or
NOX4 activity (Gianni et al. 2009). Tks4 and Tks5 were also able to interact with
NOXA1 via its SH3 domains (Gianni et al. 2009). The identity of Tks4 and Tks5
provide four distinct proteins (including NOXO1 and p47phox) containing a PX/SH3
domain organization that are involved in functional PHOX/NOX-derived ROS
generation. Tks4 and Tks5 provide further evidence of the importance of the PX domain
and SH3 domains and their interactions within NOX complexes.
Functions of Non-Phagocytic Oxidase
The non-phagocytic oxidases are hypothesized to play a wide variety of roles
ranging from immune response, cell growth and inflammation. As a result, the oxidases
are implicated in a wide variety of diseases. NOX isoforms have been implicated in
vascular disease through NOX-derived H2O2 (Cai 2005, Soccio et al. 2005, Cai et al.
2003, Meyer and Schmitt 2000). NOX1, found primarily in colon tissue, has been
proposed to be involved in inflammatory bowel disease (Szanto et al. 2005). Also,
several NOX isoforms have been proposed to be involved in neuron inflammation,
leading to several types of dementia including Alzheimer’s disease (Zekry et al. 2003).
Consequently, treatment of diseases could potentially be done through specific targeting of members of the NOX family (Cave et al. 2005, Zekry et al. 2003, Meyer and Schmitt
2000, Fukai and Nakamura 2008).
In relation to NOX1 and its involvement in cancer, further studies have indicated
that NOX1-derived ROS play a role in multiple tumorgenic processes (Chamulitrat et al.
12 2003, Shinohara et al. 2010, Komatsu et al. 2008). Shinohara et al. (2010) found that
ROS generated by NOX1 upregulated oncogenic Ras-induced matrix (MMP-9)
production. Chamulitrat et al. (2003) found that NOX1 expression and activity was
upregulated in cells displaying an anchorage-independent growth. The involvement of
ROS by various NOX enzymes in angiogenesis has been well documented (Arbiser et al.
2002, Fukai and Nakamura 2008, Kamata 2009). One specific example of NOX1
involvement in angiogenesis has been in a Ras-induced angiogenesis (Komatsu et al.
2008). These studies suggest the involvement of NOX1-derived ROS in different steps
associated with tumor progression.
Recent studies have shown that NADPH oxidase-generated superoxide is
involved in necrotic signaling events (Kim et al. 2007). NOX1 has been shown to be
involved in TNF (tumor necrosis factor)-induced necrotic death pathways. The NOX1
complex has been shown to interact with the (TNF Receptor 1) TNFR1 complex. The
mechanism by which the NOX1 complex is thought to be recruited to the TNFR1
complex is through a NOXO1-RIP1 (receptor interacting protein 1) mediated binding
(Kim et al. 2007). The role of NOX1 and NOXO1 in necrotic signaling events is
discussed in greater detail in the Appendix.
NOXO1
NOXO1 has domain organization similar to that of p47phox. Both contain an N-
terminal PX domain followed by two tandem SH3 domains and a C-terminal proline-rich
region (PR) (Figure 3). One noticeable difference between the domain organization of p47phox and NOXO1 is that NOXO1 lacks an autoinhibitory domain, including the
13
Figure 3. Domain Comparison of NOXO1 and p47phox and Alignment of NOXO1 PX
Isoforms. A. Domain organization of p47phox and NOXO1. Known interactions with lipids/proteins are noted by arrows. B. Sequence alignment of the PX domain of the four
NOXO1 isoforms. The two sites that differ in sequence are marked with * for the K50 deletion and ** for the GQASL insert.
14
A
p47phox PX SH3 SH3 AIR PR
PI(3,4)P2 p22phox p67phox PI(3,5)P2 NOXA1
NOXO1 PX SH3 SH3 PR
B *
**
15 multiple serine phosphorylation sites, which are present in p47phox. Another feature of
NOXO1 is the identification of four isoforms named NOXO1α, NOXO1β, NOXO1γ, and
NOXO1δ (Cheng and Lambeth 2005, Takeya et al. 2006, Ueyama et al. 2007). These isoforms are naturally occurring splice variants on the 3’ end of exon 3 (for the α/β isoforms) which results in a deletion of K50 or the 5’ end of exon 3 (for the γ/δ isoforms), which results in a five amino acid insert (GQASL) near position 70. The amino acid sequence of these splice variants differs only in the PX domain (Figure 3).
The different NOXO1 isoforms appear to have some variance in their tissue distribution, with the NOXO1β isoform appearing to be the predominant isoform. In testis and certain fetal tissues, NOXO1γ appears to be the predominant isoform. Both
NOXO1α and NOXO1δ do not appear to be major isoforms in any tissue (Cheng and
Lambeth 2005). Additionally, the subcellular localization of the NOXO1 isoforms has been examined. NOXO1β was localized primarily to the plasma membrane and to membranes of intracellular organelles while NOXO1γ was localized primarily in the nucleus and to a lesser extent to the plasma membrane (Ueyama et al. 2007). NOXO1α and NOXO1δ were distributed throughout the cytosol. The role of these different isoforms is not yet known.
Translocation of NOXO1 to the membrane appears to be mediated primarily through the PX domain. Studies by Cheng and Lambeth (2004) showed a similar localization to plasma membrane for full length NOXO1β and the isolated PX domain of
NOXO1β. In a construct of NOXO1β lacking the PX domain, the protein was distributed throughout the cell and did not exhibit the localization pattern seen with constructs containing the PX domain. These results led to the hypothesis that only the PX
16 domain of NOXO1β contributes to membrane binding. NOXO1β has been shown to
preferentially bind phosphatidylinositol-3,5-bisphosphate (PI(3,5)P2) (Cheng and
Lambeth 2004). NOXO1 also binds to PI(3,4)P2, phosphatidylinositol-4-phosphate
(PI(4)P) and phosphatidylinositol-5-phosphate (PI(5)P). No binding was seen to PI(3)P, phosphatidylyinositol-4,5-bisphosphate (PI(4,5)P2) or phosphatidlyinositole-3,4,5-
triphosphate (PI(3,4,5)P3). A mutation of R40 abolished binding of the NOXO1β PX domain to all of the phosphoinositides that were tested.
The PX domain
The PX domain was first identified by Ponting et al. (1996) in both p40phox and p47phox. The first assigned function of the PX domain was as a protein binding domain,
specifically to SH3 domains. Not until 2001 was the PX domain identified as a
phosphoinositide binding domain (Sato et al. 2001, Song et al. 2001, Wishart et al. 2001,
Simonsen and Stenmark 2001, Xu et al. 2001, Kanai et al. 2001, Ellson et al. 2001). The
PX domain is approximately 120 amino acids in length and is found in a variety of
eukaryotic proteins that have a wide range of functions. The PX domain is found in
several cytosolic components of the NOX and PHOX systems, sorting nexins, kinases
and phospholipases.
Most of the known PX domains bind preferentially to PI(3)P. The yeast system
of PX domains were all described as binding specifically to PI(3)P (Yu and Lemmon
2001). Other examples of PX domains that preferentially bind to one PIP include: p40phox
(PI(3)P, Ellson et al. 2001, Kanai et al. 2001), Bem1p (PI(4)P, Schuck 1997) and PI3K-
C2α (PI(4,5)P2, Song et al. 2001). In addition to those PX domains that appear to bind
17 only one PIP, several PX domains appear to be more promiscuous in their PIP-binding
profile, including those PX domains for p47phox and NOXO1β.
Structure of PX Domains and Implications for Lipid Binding
The first published structure of a PX domain was the NMR structure of p47phox
PX. Since the first structure was published in 2001, multiple structures for PX domains have been solved by both solution-state NMR and X-ray crystallography. The PX domain contains a fold comprised of three antiparallel β-strands packed against a helical bundle composed of four helices (Bravo et al. 2001 (p40phox PX), Hiroaki et al. 2001 and
Karathanassis et al. 2002 (p47phox PX), Xing et al. 2004 (CISK PX), Zhou et al. 2003
(Grd19p PX), Lu et al. 2002 (Vam7p PX), Stahelin et al. 2007 (Bem1p), Song et al. 2001
(PI3K C2α)). Chapter 3 will introduce the first solved structure of NOXO1β PX.
The structures for both p40phox PX and Grd19p (a yeast sorting nexin) PX were
phox solved bound to soluble diC4PI(3)P (Bravo et al. 2001 (p40 PX), Zhou et al. 2003
(Grd19p PX)). A comparison of the residues involved in binding to the
phosphatidylinositol showed similarities between the two PX domains. In both
structures, an arginine side chain (R58 of p40phox PX and R81 of Grd19p PX) forms
hydrogen bonds with the D3-phosphate. Mutations of these arginine residues in p40phox
PX led to a decrease in binding (Bravo et al. 2001), suggesting that residues analogous to
R58/R81 in PX domains are important for binding the D3-phosphate of the inositol ring.
The position of this arginine is highly conserved among PX domains that bind PI(3)P
(Bravo et al. 2001). The PI3K (phosphatidylinositol-3-kinase) C2α PX domain does not contain this conserved arginine residue and does not bind to 3-phosphate-containing
18 phosphatidylinositols (Song et al. 2001). The backbone amide groups of Y59/R60 (for
p40phox PX) or S83/R118 (for Grd19p PX) also form hydrogen bonds with the 3- phosphate. Another interaction between the p40phox PX domain and the 3-phosphate on
the inositol ring is a salt bridge formed with R105.
Interactions with the 1-phosphate and the 4- and 5-hydroxyls on the inositol ring
were also seen in both structures. For the 1-phosphate, the side chains of K92 and R60 of
p40phox PX form hydrogen bonds with the non-bridging oxygens. Mutation of either K92
or R60 in p40phox PX decreased the affinity of binding to PI(3)P. The 4- and 5-hydroxyl
moieties on the inositol ring form hydrogen bonds with the NH1 and NH2 of R105 in
p40phox PX and R127 in Grd19p PX. Mutation of R105 in p40phox PX eliminated binding to PI(3)P. Mutation of an analogous residue in other PX domains, including CISK
(cytokine independent survival kinase) PX and p47phox PX eliminated binding of these
domains to phospholipids (Bravo et al. 2001, Xing et al. 2004, Karathanassis et al. 2002).
The Grd19p PX structures are the only case of an apo- and PIP-bound structure
for the same PX domain, which can provide additional information in how PX domains
bind to PIPs. Overall, the apo- and PI(3)P-bound structures were very similar with just
two regions having any major difference. One of the two major differences seen between
the apo- and bound structures was the loop connecting β1 to β2. This loop is near the
PIP binding site and is involved in the dimer formation seen in the apo- but not bound
structure (Zhou et al. 2003). The second difference was in the loop connecting α1 to α2.
This area is also proximally located near the PIP binding pocket and provides one side of
the PI(3)P binding site (Zhou et al. 2003). These two structures suggest that binding of a
19 PIP to the PX domain does not induce a large conformational change and that
perturbances are localized around the PIP binding region.
Several structures of p47phox PX exist, however, none of the structures is bound to lipids. One crystal structure of p47phox PX domain was complexed with sulfate ions which indicated two basic pockets for binding phospholipids (Karathanassis et al., 2002).
One binding pocket is proposed to specifically bind PI(3,4)P2. This was based on the
location of the binding pocket for PI(3)P in the p40phox PX domain and the similarity in
fold between the PX domains. The position of the sulfate in the p47phox PX structure mimics the position of the 3-phosphate of PI(3)P in the p40phox PX structure (Bravo et al.
2001). In the p47phox PX crystal structure, the side chain of R43 mimics the position of
R58 in the p40phox PX structure and presumably interacts with the 3-phosphate. R90 in
the p47phox PX structure mimics the position of R105 in the p40phox PX structure and
would presumably interact with the 4- and 5-positions on the inositol headgroup.
Mutation of R43 to Q substantially decreased binding of the p47phox PX domain to
PI(3,4)P2-containing vesicles (Karathanassis et al. 2002). Mutation of R90 to A also had
phox a reduction in the binding affinity of p47 PX to PI(3,4)P2, although the effect was not
as severe as the R43 mutation (Karathanassis et al. 2002). Modeling PI(3,4)P2 into the
binding pocket of p47phox PX in the same orientation as PI(3)P was in the p40phox crystal
structure introduces several steric clashes. The shape of the binding pocket in the (apo)
p47phox PX crystal structure cannot accommodate the 4-phosphate. Either there is a
conformational change in the binding pocket upon binding PI(3,4)P2 or the inositide
headgroup binds in another orientation. Karathanassis et al. (2002) proposed that the
headgroup of PI(3,4)P2 would be rotated in the binding pocket so that favorable
20 interactions can occur between the 3- and 4-phosphates and arginine side chains within
the pocket.
The second binding pocket in the p47phox PX structure is thought to bind acidic phospholipids, such as phosphatidic acid. This second binding pocket is composed of
several basic residues, including R70, K55 and H51. R70 forms several hydrogen bonds
between the sulfate ion and neighboring residues in the protein. (Karathanassis et al.
2002). Mutation of R70 to Q had a decrease in binding affinity for either PA or
phosphatidyl serine (PS) (Karathanassis et al. 2002). An interesting feature of this
postulated second binding pocket for p47phox PX is that the two basic residues analogous to R70/K55 are not common among PX domains (Karathanassis et al. 2002). These basic residues appear to be present in the PX domains for p47phox, PLD and NOXO1.
The PX domain has shown to be an integral domain in formation of active
NADPH oxidase, both in the phagocytic and non-phagocytic NADPH oxidase systems.
Having a better understanding of the lipid-binding of these PX domains can provide
insight into signaling events necessary for functional complex as well as providing
insight into a lipid signal for recruitment of the NADPH oxidase to other complexes. The
goals of this project are to solve the solution structure of the NOXO1β PX domain and
characterize the lipid binding profile for this PX domain using multiple assays. The data in this thesis will provide more information to help in understanding PX domains and how their lipid binding specificity can be thoroughly addressed. The data also aims to test the hypothesis that the NOXO1β PX domain is a spatial regulator not only within the
NADPH oxidase complex, but also in bringing the NADPH oxidase to other complexes.
21 References
Ago, T., Kuribayashi, F., Hiroaki, H., Takeya, R., Ito, T., Kohda, D. and Sumimoto, H. (2003) Phosphorylation of p47phox directs Phox homology domain from SH3 domain toward phosphoinositides, leading to phagocyte NADPH oxidase activation. Proc. Natl. Acad. Sci. 100, 4474-4479. Arbiser, J.L., Petros, J. Klafter, R. Govindajaran, B., McLaughlin, E.R., Brown, L.F., Cohen, C., Moses, M., Kilroy, S., Arnold, R.S. and Lambeth, J.D.(2002) Reactive oxygen generated by Nox1 triggers the angiogenic switch. Proc. Natl. Acad. Sci. USA 99, 715-720. Babior, B.M. (1991) The respiratory burst oxidase and the molecular basis of chronic granulomatous disease. Am. J. Hematol. 37, 263-266. Babior, B.M. (2004) NADPH oxidase. Curr. Opin. Immunol. 16, 42-47. Baldrige, C.W., Gerad, R. W. (1933) The extra respiration of phagocytes. American Journal of Physiology 103, 235-236. Banfi, B., Molnar, G., Maturana, A., Steger, K., Hegedus, B., Demaurex, N., Krause, K.H. (2001) A Ca2+-activated NADPH oxidase in testis, spleen and lymph nodes. J. Biol. Chem. 276, 37594-37601. Banfi, B., Clark, R.A., Steger, K., Krause, K.H. (2003) Two novel proteins activate superoxide generation by the NADPH oxidase NOX1. J. Biol. Chem. 278, 3510- 3513. Bedard, K., and Krause, K.H. (2007) The NOX family of ROS-generating NADPH oxidases: physiology and pathophysiology. Physiol Rev 87, 245-313. Bokoch, G.M. and Knuas, U.G. (2003) NADPH oxidase: not just for leukocytes anymore! TIBS 28, 502-508. Borregaard, N., Lollike, K., Kjeldsen, L., Sengelov, H., Bastholm, L., Nielsen, M.H., and Bainton, D.F. (1993) Human neutrophil granules and secretory vesicles. Eur. J. Haematol. 51,187-198. Bravo J., Karathanassis, D., Pacold, C.M., Pacold, M.E., Ellson, C.D., Anderson, K.E., Butler, P.J.G., Lavenir, I., Perisic, O., Hawkins, P.T., Stephens, L., Williams, R.L. (2001) The crystal structure of the PX domain from p40phox bound to phosphatidylinositol 3-phosphate. Mol. Cell 8, 829-839. Cai, H. (2005) NAD(P)H oxidase-dependent self-propagation of hydrogen peroxide and vascular disease. Circ. Res. 96, 818-822. Cai, H., Griendling, K.K., Harrison, D.G. (2003) The vascular NAD(P)H oxidases as therapeutic targets in cardiovascular diseases. Trends Pharmacol Sci. 24, 471- 478. Cave, A., Grieve, D., Johar, S., Zhang, M., Shah, A.M. (2005) NADPH oxidase-derived reactive oxygen species in cardiac pathophysiology. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 360, 2327-34. Chamulitrat, W., Schmidt, R., Tomakidi, P., Stremmel, W., Chunglok, W., Kawahara, T. and Rokutan, K. (2003) Association of gp91phox homolog Nox1 with anchorage- independent growth and MAP kinase-activation of transformed human keratinocytes. Oncogene 22, 6045-6053.
22 Cheng, G. and Lambeth, J.D. (2004) NOXO1, regulation of lipid binding, localization and activation of Nox1 by the Phox homology (PX) domain. J. Biol. Chem. 279, 4737-4742. Cheng, G. and Lambeth, J.D. (2005) Alternative mRNA splice forms of NOXO1: Differential tissue expression and regulation of Nox1 and Nox3. Gene 356, 118- 126. Courtneidge, S.A., Azucena, E.F., Pass, I., Seals, D.F. and Tesfay, L. (2005) The SRC substrate Tks5, podosomes (invadopodia), and cancer cell invasion. Cold Spring Harb Symp Quant Biol. 70, 167-71. Courtneidge S.A. (2003) Isolation of novel Src substrates. Biochem Soc Trans. 31, 25-28 Cross, A.R. and Segal, A.W. (2004) The NADPH oxidase of professional phagocytes-- prototype of the NOX electron transport chain systems. Biochim. Biophys. Acta. 1657,1-22. Dallegri, F., Ballestrero, A., Frumento, G., and Patrone, F. (1986) Role of hypochlorous acid and chloramines in the extracellular cytolysis by neutrophil polymorphonuclear leukocytes. J. Clin. Lab. Immunol. 20, 37-41. Diaz, B., Shani, G., Pass, I., Anderson, D., Quintavalle, M. and Courtneidge, S.A. (2009) Tks5-dependent, nox-mediated generation of reactive oxygen species is necessary for invadopodia formation. Sci Signal 2, ra53. Doussiere, J., Brandolin, G., Derrien, V., and Vignais, P.V. (1993) Critical assessment of the presence of an NADPH binding site on neutrophil cytochrome b558 by photoaffinity and immunochemical labeling. Biochemistry 32, 8880-8887. Doussiere, J., Gaillard, J. and Vignais, P.V. (1996) Electron transfer across the O2- generating flavocytochrome b of neutrophils. Evidence for a transition from a low-spin state to a high-spin state of the heme iron component. Biochemistry 35,13400-13410. Dusi, S., Donini, M. and Rossi, F. (1996) Mechanisms of NADPH oxidase activation: translocation of p40phox, Rac1 and Rac2 from the cytosol to the membranes in human neutrophils lacking p47phox or p67phox. Biochem. J. 314, 409-412. Edens, W.A., Sharling, L., Cheng, G., Shapira, R., Kinkade, J.M., Lee, T., Edens, H.A., Tang, X., Sullards, C., Flaherty, D.B., Benian, G.M., Lambeth, J.D. (2001) Tyrosine crosslinking of extracellular matrix is catalyzed by Duox, a multidomain oxidase/peroxidase with homology to the phagocytic oxidase subunit gp91phox. J. Cell Biol. 154, 879-891. Ellson, C.D., Gobert-Gosse, S., Anderson, K.E., Davidson, K., Erdjument-Bromage, H., Tempst, P., Thuring, J.W., Cooper, M.A., Lin, Z-Y., Holmes, A.B., Gaffney, P.R.J., Coadwell, J., Chilvers, E.R., Hawkins, P.T. and Stephens, L.R. (2001) PtdIns(3)P regulates the neutrophil oxidase complex by binding to the PX domain of p40phox. Nat. Cell Bio. 3, 679-682. Ellson, C.D., Andrews, S., Stephens, L.R., Hawkins, P.T. (2002) The PX domain: a new phosphoinositide-binding module. J. Cell Sci. 115, 1099-1105. Ellson, C.D., Davidson, K., Ferguson, G.J., O'Connor, R., Stephens, L.R. and Hawkins, P.T. (2006) Neutrophils from p40phox-/- mice exhibit severe defects in NADPH oxidase regulation and oxidant-dependent bacterial killing. J. Exp. Med. 203,1927-1937.
23 Finegold, A.A., Shatwell, K.P., Segal, A.W., Klausner, R.D. and Dancis, A. (1996) Intramembrane bis-heme motif for transmembrane electron transport conserved in a yeast iron reductase and the human NADPH oxidase. J. Biol. Chem. 271, 31021-31024. Geiszt, M., Kopp, J.B., Varnai, P., Leto, T.L. (2000) Identification of renox, an NAD(P)H oxidase in kidney. Proc. Natl. Acad. Sci. 97, 8010-8014. Geiszt, M., Kapus, A., Ligeti, E. (2001) Chronic granulomatous disease: more than the lack of superoxide? J. Leukoc. Biol. 69, 191-196. Geiszt, M., Lekstrom, K., Witta, J., Leto, T.L. (2003) Proteins homologous to p47phox and p67phox support superoxide production by NAD(P)H oxidase 1 in colon epithelial cells. J. Biol. Chem. 278, 20006-20012. Gianni, D., Diaz, B., Taulet, N., Fowler, B., Courtneidge, S.A. and Bokoch, G.M. (2009) Novel p47(phox)-related organizers regulate localized NADPH oxidase 1 (Nox1) activity. Sci Signal 2, ra54. Groemping, Y., Lapouge, K., Smerdon, S.J. and Rittinger, K. (2003) Molecular basis of phosphorylation-induced activation of the NADPH oxidase. Cell 113, 343-355. Hampton, M.B., Kettle, A.J., and Winterbourn, C.C. (1998) Inside the neutrophil phagosome: oxidants, myeloperoxidase, and bacterial killing. Blood 92, 3007- 3017. Han, C.H., Freeman, J.L.R., Lee, T., Motalebi, S.A. and Lambeth, J.D. (1998) Regulation of the neutrophil respiratory burst oxidase: identification of an activation domain in p67phox. J. Biol. Chem. 273, 16663-16668. Han, C.H. and Lee, M.H. (2000) Activation domain in P67phox regulates the steady state reduction of FAD in gp91phox. J. Vet. Sci. 1, 27-31. Harper, A.M., Chaplin, M.F. and Segal, A.W. (1985) Cytochrome b-245 from human neutrophils is a glycoprotein. Biochem. J. 227, 783-788. Hata K, Ito T, Takeshige K, Sumimoto H. (1998) Anionic amphiphile-independent activation of the phagocyte NADPH oxidase in a cell-free system by p47phox and p67phox, both in C terminally truncated forms. Implication for regulatory Src homology 3 domain-mediated interactions. J. Biol. Chem. 273, 4232-4236. Heyworth, P.G., Curnutte, J.T., Nauseef ,W.M., Volpp, B.D., Pearson, D.W., Rosen, H. and Clark, R.A. (1991) Neutrophil NADPH oxidase assembly. Membrane translocation of p47-phox and p67-phox requires interaction between p47-phox and cytochrome b558. J. Clin. Invest. 87, 352–356. Heyworth, P.G., Cross, A.R., Curnutte, J.T. (2003) Chronic granulomatous disease. Curr. Opin. Immunol. 15, 578-84. Hiroaki, H., Ago, T., Ito, T., Sumimoto, H., Kohda, D. (2001) Solution structure of the PX domain, a target of the SH3 domain. Nat. Struct. Biol. 8, 526-530. Jurkowska, M., Bernatowska, E., Bal, J. (2004) Genetic and biochemical background of chronic granulomatous disease. Arch. Immunol. Ther. Exp. 52, 113-20. Kamata, T. (2009) Roles of Nox1 and other Nox isoforms in cancer development. Cancer Sci. 100, 1382-1388. Kanai, F., Liu, H., Field, S.J., Akbary, H., Matsuo, T., Brown, G.E., Cantley, L.C. and Yaffe, M.B. (2001) The PX domains of p47phox and p40phox bind to lipid products of PI(3)K. Nat. Cell Biol. 3, 675–678.
24 Karathanassis, D., Stahelin, R.V., Bravo, J., Perisic, O., Pacold, C.M., Cho, W., Williams, R.L. (2002) Binding of the PX domain of p47phox to phosphatidylinositol 3,4-bisphosphate and phosphatidic acid is masked by an intramolecular interaction. EMBO J. 21, 5057-5068. Kim, Y-S., Morgan, M.J., Choksi, S. and Liu, Z-G. (2007) TNF-induced activation of the Nox1 NADPH oxidase and its role in the induction of necrotic cell death. Mol. Cell. 26, 675-687. Koga, H., Terasawa, H., Nunoi, H., Takeshige, K., Inagaki, F. and Sumimoto, H. (1999) Tetratricopeptide repeat (TPR) motifs of p67phox participate in interaction with the small GTPase Rac and activation of the phagocyte NADPH oxidase. J. Biol. Chem. 274, 25051–25060. Komatsu, D., Kato, M., Nakayama, J., Miyagawa, S. and Kamata, T. (2008) NADPH oxidase 1 plays a critical mediating role in oncogenic Ras-induced vascular endothelial growth factor expression. Oncogene 27, 4724-4732. Kutateladze, T.G., Capelluto, D.G.S., Ferguson, C.G., Cheever, M.L., Kutateladze, A.G., Prestwich, G.D., Overduin, M. (2003) Multivalent Mechanism of Membrane Insertion by the FYVE Domain. J. Biol. Chem. 279, 3050-3057. Lambeth, J.D., Cheng, G., Arnold, R.S., Edens, W.E. (2000) Novel homologs of gp91phox. Trends Biochem. Sci. 25, 459-461. Lapouge, K., Smith, S.J.M., Groemping, Y., Rittinger, K. (2002) Architecture of the p40- p47-p67phox complex in the resting state of the NADPH oxidase. J. Biol. Chem. 277, 10121–10128. Lewis, E.M. (2007) Regulatory threonines within the NADPH oxidase component p22phox and the diagnosis of a flavocytochrome b558 deficient patient with chronic granulomatous disease. Lewis, E.M., Sergeant, S., Ledford, B., Stull, N., Dinauer, M.C. and McPhail, L.C. (2010) Phosphorylation of p22phox on threonine 147 enhances NADPH oxidase activity by promoting p47phox binding. J. Biol. Chem. 285, 2959-2967. Lock, P., Abram, C.L., Gibson, T. and Courtneidge, S.A. (1998) A new method for isolating tyrosine kinase substrates used to identify fish, an SH3 and PX domain- containing protein, and Src substrate. EMBO J. 17, 4346-4357. Lopes, L.R., Dagher, M.C., Gutierrez, A., Young, B., Bouin, A.P., Fuchs, A. and Babior, B.M. (2004) Phosphorylated p40PHOX as a negative regulator of NADPH oxidase. Biochemistry 43, 3723-3730. Lu, J., Garcia, J., Dulubova, I., Sudhof, T.C., Rizo, J. (2002) Solution structure of the VAM7P PX domain. Biochemistry 41, 5956-5962. Matute, J.D., Arias, A.A., Dinauer, M.C. and Patino, P.J. (2005) p40phox: the last NADPH oxidase subunit. Blood Cells Mol. Dis. 35, 291-302. Matute, J.D., Arias, A.A., Wright, N.A., Wrobel, I., Waterhouse, C.C., Li, X.J., Marchal, C.C., Stull, N.D., Lewis, D.B., Steele, M., Kellner, J.D., Yu, W., Meroueh, S.O., Nauseef, W.M. and Dinauer, M.C. (2009) A new genetic subgroup of chronic granulomatous disease with autosomal recessive mutations in p40 phox and selective defects in neutrophil NADPH oxidase activity. Blood. 114, 3309-3315. Meyer, J.W. and Schmitt, M.E. (2000) A central role for the endothelial NADPH oxidase in atherosclerosis. FEBS Lett. 472, 1-4.
25 Nakamura, R., Sumimoto, H., Mizuki, K., Hata, K., Ago, T., Kitajima, S., Takeshige, K., Sakaki, Y. and Ito, T. (1998) The PC motif: a novel and evolutionarily conserved sequence involved in interaction between p40 Phox and p67phox, SH3 domain- containing cytosolic factors of the phagocyte NADPH oxidase. Eur. J. Biochem. 251, 583–589. Nisimoto, Y., Motalebi, S., Han, C.H. and Lambeth, J.D. (1999) The p67phox activation domain regulates electron transfer flow from NADPH to flavin in flavocytochrome b588. J. Biol. Chem. 274, 22999-23005. Noack, D., Rae, J., Cross, A.R., Ellis, B.A., Newburger, P.E., Curnutte, J.T., Heyworth, P.G. (2001) Autosomal recessive chronic granulomatous disease caused by defects in NCF-1, the gene encoding the phagocyte p47-phox: mutations not arising in the NCF-1 pseudogenes. Blood 97, 305-311. Palicz, A., Foubert, T.R., Jesaitis, A.J., Marodi, L., McPhail, L.C. (2001) Phoshphatidic acid and diacylglycerol directly activate NADPH oxidase by interacting with the enzyme components. J. Biol. Chem. 276, 3090-3097. Park, H.S., Park, J.W. (1998) Conformational changes of the leukocyte NADPH oxidase subunit p47(Phox) during activation studied through its intrinsic fluorescence. Biochim. Biophys. Acta. 1387, 406-414. Ponting, C.P. (1996) Novel domains in NADPH oxidase subunits, sorting nexins and PtdIns 3-kinases: Binding partners of SH3 domains? Protein Sci. 5, 2353-2357. Quinn, M.T., Mullen, M.L., Jesaitis, A.J. (1992) Human neutrophil cytochrome-b contains multiple hemes. Evidence for heme associated with both subunits. J. Biol. Chem. 267, 7303-7309. Quinn, M.T., Gauss, K.A. (2004) Structure and regulation of the neutrophil respiratory burst oxidase: comparison with nonphagocyte oxidases. J. Leukoc. Bio. 76, 1-22. Roos, D., de Boer, M., Kuribayashi, F., Weening, R.S., Segal, A.W., Ahlin, A., Nemet, K., Hassle, J.P., Bernatowska-Matuszkiewicz, E., Middleton-Price, H. (1996) Mutations in the X-linked and autosomal recessive forms of Chronic Granulomatous Disease. Blood 87, 1663-l681. Rotrosen, D., Yeung, C.L., Leto, T.L., Malech, H.L., Kwong, C.H. (1992) Cytochrome b558. The flavin-binding component of the phagocyte NADPH oxidase. Science 256, 1459-1462. Sathyamoorthy, M., de Mendez, I., Adams, A.G. and Leto, T.L. (1997) p40(phox) down- regulates NADPH oxidase activity through interactions with its SH3 domain. J. Biol. Chem. 272, 9141-9146. Sbarra, A.J., and Karnovsky, M.L. (1959) The biochemical basis of phagocytosis. I. Metabolic changes during the ingestion of particles by polymorphonuclear leukocytes. J. Biol. Chem. 234, 1355-1362. Schall, T.J. and Bacon, K.B. (1994) Chemokines, leukocyte trafficking and inflammation. Curr. Opin. Immunol. 6, 865-873. Schuck, P. (1997) Use of surface plasmon resonance to probe the equilibrium and dynamic aspects of interactions between biological macromolecules. Annu. Rev. Biophys. Biomol. Struct. 26, 541-566. Shatwell, K.P. and Segal, A.W. (1996) NADPH Oxidase Int. J. Biochem. Cell Biol. 28, 1191-1195.
26 Shepherd, V.L. (1986) The role of the respiratory burst of phagocytes in host defense. Semin. Respir. Infect. 1, 99-106. Shinohara, M., Adachi, Y., Mitsushita, J., Kuwabara, M., Nagasawa, A., Harada, S., Furuta, S., Zhang, Y., Seheli, K., Miyazaki, H. and Kamata, T. (2010) Reactive oxygen generated by NADPH oxidase 1 (Nox1) contributes to cell invasion by regulating matrix metalloprotease-9 production and cell migration. J. Biol. Chem. 285, 4481-4488. Soccio, M., Toniato, E., Evangelista, V., Carluccio, M., De Caterina, R. (2005) Oxidative stress and cardiovascular risk: the role of vascular NAD(P)H oxidase and its genetic variants. Eur. J. Clin. Invest. 35, 305-314. Song, X., Xu, W., Zhang, A., Huang, G., Liang, X., Virbasius, J.V., Czech, M. P., and Zhou, G. W. (2001) Phox homology domains specifically bind phosphatidylinositol phosphates. Biochemistry 40, 8940–8944. Stahelin, R.V., Burian, A., Bruzik, K.S., Murray, D., Cho, W. (2003) Membrane binding mechanism of the PX domains of NADPH oxidase p40phox and p47phox. J. Biol. Chem. 278, 14469-14479. Stahelin, R.V., Karathanassis, D., Murray, D., Williams, R.L. and Cho, W. (2007) Structural and membrane binding analysis of the phox homology domain of Bem1p. J. Biol. Chem. 282, 25737-25747. Suh, C.I., Stull, N.D., Li, X.J., Tian, W., Price, M.O., Grinstein, S., Yaffe, M.B., Atkinson, S. and Dinauer, M.C. (2006) The phosphoinositide-binding protein p40phox activates the NADPH oxidase during FcgammaIIA receptor-induced phagocytosis. J. Exp. Med. 203,1915-1925. Suh, Y.-A., Arnold, R.S., Lassegue, B., Shi, J., Xu, X., Sorescu, D., Chung, A.B., Griendling, K.K., Lambeth, J.D. (1999) Cell transformation by the superoxide- generating oxidase Mox1. Nature 401, 79-82. Sumimoto, H., Kage, Y., Nunoi, H., Sasaki, H., Nose, T., Fukumaki, Y., Ohno, M., Minakami, S., Takeshige, K. (1994) Role of Src homology 3 domains in assembly and activation of the phagocyte NADPH oxidase. Proc. Natl. Acad. Sci. U.S.A. 91, 5345-5349. Swain, S.D., Helgerson, S.L., Davis, A.R., Nelson, L.K., Quinn, M.T. (1996) Analysis of activation-induced conformational changes in p47phox using tryptophan fluorescence spectroscopy. J. Biol. Chem. 272, 29502-29510. Szanto, I., Rubbia-Brandt, L., Kiss, P., Steger, K., Banfi, B., Kovari, E., Herrmann, F., Hadengue, A., Krause, K.H. (2005) Expression of NOX1, a superoxide- generating NADPH oxidase, in colon cancer and inflammatory bowel disease. J. Pathol. 207, 164-176. Takeya, R., Ueno, N., Kami, K., Taura, M., Kohjima, M., Izaki, T., Nunoi, H., Sumimoto, H. (2003) Novel human homologues of p47phox and p67phox participate in activation of superoxide-producing NADPH oxidase. J. Biol. Chem. 278, 25234-25246. Takeya, R., Taura, M., Yamasaki, T., Naito, S. and Summimoto, H. (2006) Expression and function of Noxo1gamma, an alternative splicing form of the NADPH oxidase organizer 1. Febs J. 273, 3663-3677. Thomas, E.L., Lehrer, R.I., and Rest, R.F. (1988) Human neutrophil antimicrobial activity. Rev. Infect. Dis. 10, S450-S456.
27 Ueno, N., Takeya, R., Miyano, K., Kikuchi, H. and Sumimoto, H. (2005) The NADPH oxidase Nox3 constitutively produces superoxide in a p22phox-dependent manner: its regulation by oxidase organizers and activators. J. Biol. Chem. 280, 23328-23339. Ueyama, T., Geiszt, M. and Leto, T.L. (2006) Involvment of Rac1 in activation of multicomponent Nox1- and Nox3-based NADPH oxidases. Mol. Cell Biol. 26, 2160-2174. Ueyama, T., Lekstrom, K., Tsujibe, S., Saito, N and Leto, T.L. (2007) Subcellular localization and function of alternatively spliced Noxo1 isoforms. Free Radic. Biol. Med. 42, 180-190. Vignais, P.V. (2002) The superoxide-generating NADPH oxidase: structural aspects and activation mechanism. Cell Mol. Life Sci. 59, 1428-1459. Wiernik, P.H. (1985) Neutrophil function and infection. Mediguide to Oncology 1, 1-12. Xing, Y., Liu, D., Zhang, R., Joachimiak, A., Songyang, Z., Xu, W. (2004) Structural basis of membrane targeting by the Phox homology domain of cytokine- independent survival kinase (CISK-PX). J. Biol. Chem. 279, 30662-30669. Yaffe, M.B. (2002) The p47phox PX domain: two heads are better than one! Structure (Camb). 10, 1288-1290. Yu, J.W. and Lemmon, M.A. (2001) All phox homology (PX) domains from Saccharomyces cerevisiae specifically recognize phosphatidylinositol 3- phosphate. J. Biol. Chem. 276, 44179-44184. Zekry, D., Epperson, T.K., Krause, K.H. (2003) A role for NOX NADPH oxidases in Alzheimer's disease and other types of dementia? IUBMB Life. 55, 307-313. Zeng, J. & Fenna, R. E. (1992) X-ray crystal structure of canine myeloperoxidase at 3 Å resolution. J. Mol. Biol. 226, 185–207. Zhan, Y., Virbasius, J.V., Song, X., Pomerleau, D.P., Zhou, G.W. (2002) The p40phox and p47phox PX domains of NADPH oxidase target cell membranes via direct and indirect recruitment by phosphoinositides. J. Biol. Chem. 277, 4512-4518. Zhou, C.-Z., de La Sierra-Gallay, I.L., Quevillon-Cheruel, S., Collinet, B., Minard, P., Blondeau, K., Henckes, G., Aufrère, R. Leulliot, N., Graille, M., Sorel, I., Savarin, P., de la Torre, F., Poupon, A., Janin, J., and van Tilbeurgh, H. (2003) Crystal structure of the yeast Phox homology (PX) domain protein Grd19p complexed to phosphatidylinositol-3-phosphate. J. Biol. Chem. 278, 50371- 50376.
28 CHAPTER II
NOXO1β PX BINDS TO PI(4,5)P2 IN ADDITION TO NEUTRAL MEMBRANE
LIPIDS.
N. Y. Davis and D.A. Horita
The following experiments shown were performed by N. Y. Davis. P47phox PX was purified by Dr. K. Shen. P40phox PX was purified by J. V. Tuttle. Dr. D. A. Horita acted in an advisory capacity.
Introduction
One of the key interactions within the NOX and PHOX complexes is the protein- lipid interaction. The PX domain was first identified in two cytosolic proteins in the
PHOX complex, p40phox and p47phox. Surprisingly, the PX domain was first characterized as a protein-binding domain (Ponting et al. 1996) and it was not until 2001 that it was identified as a lipid-binding domain (Sato et al. 2001, Song et al. 2001, Wishart et al.
2001, Simonsen and Stenmark 2001, Xu et al. 2001, Kanai et al. 2001, Ellson et al.
2001). PX domains are found in a variety of proteins outside of the NOX/PHOX complexes including sorting nexins, phospholipases and kinases. These PX domains target membrane lipids, specifically phosphorylated phosphatidylinositols. The binding specificity of PX domains varies by proteins, with some proteins being very specific for a single phosphoinositide while others are more promiscuous and can bind to several PIPs.
Most PX domains, including the PX domain of p40phox, bind specifically to PI(3)P
(Yu and Lemmon 2001, Kanai et al. 2001, Cheever et al. 2001, Xu et al. 2001, Zhan et al. 2002, Stahelin et al. 2003, Ellson et al. 2004). Conversely, p47phox PX has a broader
phox lipid binding specificity. P47 PX preferentially binds PI(3,4)P2 (Kanai et al. 2001,
Karathanassis et al. 2002, Zhan et al. 2002, Stahelin et al. 2003) but can also bind to other phosphatidylinositols (Kanai et al. 2001, Karathanassis et al. 2002, Ledford, 2004).
In addition to the PIPs, p47phox PX can bind to acidic phospholipids, specifically PA
(Yaffe 2002, Karathanassis et al. 2002, Ledford 2004). The binding site for PA is proposed to be distinct from the binding site for PIPs (Karathanassis et al. 2002). This dual binding allows for two distinct signals (PLD-driven hydrolysis producing PA and
30 production of PI(3,5)P2) that are needed to signal and regulate the assembly of the PHOX
complex.
The only published data for NOXO1β PX lipid binding is using dot blots.
NOXO1β PX is reported to bind preferentially to PI(3,5)P2 while also binding to
PI(3,4)P2, PI(4)P and PI(5)P (Cheng and Lambeth 2004, Cheng and Lambeth 2005,
Takeya et al. 2006, Ueyama et al. 2007). NOXO1β PX also has the potential to bind to
acidic phospholipids in that it has high sequence similarity with the residues in p47phox
PX that are believed to be involved in binding to PA (Karathanassis et al. 2002). This illustrates how little is known and has been tested about the lipid binding of NOXO1β
PX.
There have been several reports where results from using protein-lipid overlay experiments differ from those from other binding assays (Yu et al. 2004, Narayan and
Lemmon 2006). Even Echelon, the manufacturers of several lipid membrane products, recommends “researchers use alternative methods to fully characterize the lipid binding preference of a particular protein” (Echelon Biosciences, 2006). To better characterize the lipid binding profile of NOXO1β PX, we performed multiple lipid binding assays to gain a better characterization of the lipid binding profile of NOXO1β PX. We compared the binding of NOXO1β PX to both p40phox PX and p47phox PX to illustrate the varying
lipid binding properties of the PX domains found in the NOX and PHOX complexes.
31 Experimental Procedures
NOXO1β PX Expression and Purification-The sequence for human NOXO1β PX (1-
144) was cloned into pGEX-6P-1 (Amersham) and transformed into the Escherichia coli
(E. coli) strain BL21(DE3)Codon+ RIPL (Stratagene). Cultures were grown at 32 °C in
minimal media until OD600 reached 0.6 and induced with 0.5mM IPTG at 22 °C
overnight The harvested cell pellet was resuspended in lysis buffer (20mM Tris, 20mM
NaCl, 2mM DTT, 0.1% (v/v) Tween-20, pH 8.0) and lysed by sonication on ice for 8
cycles (4 minutes/cycle, level 8, 30% duty cycle). 10μl Benzonase (Novagen) was added
to lysates and incubated at room temperature for one hour to digest bacterial DNA.
Lysates were clarified by centrifugation for 45 minutes at 30,000 x g at 4 °C. Clarified
lysate containing GST-NOXO1β PX was run over SP Sepharose resin equilibrated in SP-
A buffer (50mM NaPi, 100mM NaCl, 2mM DTT, pH 7.4). The fusion protein was eluted
with a salt gradient (SP-B: 50mM NaPi, 1.5M NaCl, 2mM DTT, pH 7.4) and eluted at
~800mM NaCl (~45mS/cm). The pooled SP Sepharose peak was loaded onto GSH-
Sepharose FF resin equilibrated in GSH-A (50mM NaPi, 800mM NaCl, 2mM DTT, pH
7.4). GST-NOXO1β PX was eluted with 15mM reduced glutathione (GSH-B: 50mM
NaPi, 800mM NaCl, 2mM DTT, 15mM gluathione, pH 7.4). The pooled GSH-Sepharose
FF eluted peak was dialyzed against 2L SP-A with simultaneous digestion with 10μl (80
μg) PreScission Protease (GE Biosciences). The digested fusion protein was loaded onto
Source 15S resin equilibrated in SP-A. Cleaved NOXO1β PX was eluted with a salt gradient (SP-B), with the protein eluting at ~400mM NaCl (~33 mS/cm). The pooled
NOXO1β PX was treated with Pefabloc PSC (Roche, 0.1 mg/ml final concentration) for
32 30 minutes at 4 °C. Final yield of purified NOXO1β PX was ~3 mg/L culture. GST-
NOXO1β PX used in assays was purified as described above with the omission of
PreScission Protease. For dot blots, GST-NOXO1β PX was dialyzed into TBS-T (20mM
Tris, 136mM NaCl, 0.1% (v/v) Tween-20, pH 7.2). For vesicle and SPR experiments,
NOXO1β PX was dialyzed into 20mM HEPES, pH 7.2, 137mM KCl, 1mM MgCl2, 1mM
EGTA.
NOXO1β PX (1-144) was also cloned in pMHPb, with the residual fusion tag mutated to
have the same sequence as the construct using the pGEX-6P-1 plasmid. This fusion
system was also cloned into BL21(DE3)Codon+ RIPL E. coli. Cultures were grown at
37 °C until OD600=0.6 and induced with 0.5mM IPTG overnight at 15 °C. After
harvesting, cells were resuspended into 50mM KPi, 1M NaCl, 0.1% (v/v) Tween-20, pH
8.0 and lysed by sonication on ice for 5 cycles with a 3/8” probe (3 minutes/cycle, level
9, 50% duty). Lysates were cleared by centrifugation for 40 minutes at 40,000 x g at 4
°C. Clarified lysate was loaded onto a Ni-NTA resin equilibrated with Ni-A1 (50mM
KPi, 1M NaCl, pH 8.0). After washing with Ni-A2 (50mM KPi, 50mM NaCl, 5mM
imidazole, pH 7.4), MBP-His6-NOXO1β PX was eluted with Ni-B1 (50mM KPi, 50mM
NaCl, 200mM imidazole, pH 7.4). The pooled Ni-NTA elute was loaded onto SP
Sepharose resin equilibrated with SPA (50mM KPi, 50mM NaCl, 2mM DTT, pH 7.5)
and eluted with a NaCl gradient (SPB, 50mM KPi, 1M NaCl, 2mM DTT, pH 7.5). The
fusion protein eluted at ~55% SPB. The eluted peak was dialyzed against 2L of Q-A
(50mM Tris-HCl, 25mM NaCl, 2mM DTT, pH 8.0) overnight at 4 °C. After dialysis, the
protein was loaded onto Q Sepharose resin equilibrated with Q-A and eluted with a NaCl
33 gradient (Q-B, 50mM Tris-HCl, 1M NaCl, 2mM DTT, pH 8.0) with MBP-His6-NOXO1β
PX eluting at ~32% Q-B. The pooled Q Sepharose peak was dialyzed against 2L SPA
with simulatneous digestion with 10μl PreScission Protease overnight at 4 °C. The
digested fusion protein was loaded onto SP Sepharose resin equilibrated with SPA and
eluted with a NaCl gradient (SPB). NOXO1β PX eluted at ~60% SPB. Final yield for
the pMHPb expression system was 9-10mg NOXO1β PX/L culture. NOXO1β PX for
vesicle and SPR experiments was dialyzed into 20mM HEPES, pH 7.2, 137mM KCl,
1mM MgCl2, 1mM EGTA.
P47phox PX Expression and Purification-The sequence for human p47phox PX (2-134) is
expressed as a GST-fusion (pGEX-6P-1) in E. coli BL21(DE3)tuner. Bacterial cells are
grown at 37 °C to OD600=0.6. Cultures are induced with the addition of IPTG to a final
concentration of 0.5mM for four hours at 37 °C. After harvesting cells by centrifugation,
the cell pellet is resuspended in 1X PBS, 2mM DTT and sonicated on ice with ¾” probe
for 3 cycles of 3 minutes, level 7, 50% duty. The cell lysate is incubated with 10 μl
Benzonase for 30 minutes at room temperature. Cell lysate is clarified by centrifugation at 25,000 x g for 45 minutes at 4 °C. A 50% glutathione-agarose slurry is added to the clarified lysate and is nutated at 4 °C for four hours to allow fusion protein to bind with the resin. The glutathione resin is washed with 1X PBS, 2mM DTT. GST-p47phox PX is eluted with 1X PBS, 2mM DTT, 15mM reduced glutathione. Pooled fusion protein is digested overnight at 4 °C by addition of 10 μl PreScission Protease. For assays using
GST-p47phox PX the PreScission Protease step is omitted. P47phox PX is dialyzed into
5mM KPi, 10mM NaCl, 2mM DTT, pH 7.0 and loaded onto a 40ml SP Sepharose
34 column. P47phox PX is eluted with a NaCl gradient and comes off the column at
~15mS/cm (~150mM NaCl). Pooled p47phox PX is dialyzed into 20mM HEPES, pH 7.2,
phox 137mM KCl, 1mM MgCl2, 1mM EGTA for SPR and lipid vesicle assays. GST-p47
PX is dialyzed into TBS-T for dot blots.
P40phox PX Expression and Purification-The sequence for human p40phox PX domain
(residues 2-149) were cloned into pET49b and transformed into the
BL21(DE3)CodonPlus-RIPL E coli. cell line. The cells were grown at 37 °C until OD600 of 0.6 and induced with 0.5mM IPTG overnight at 15 °C. Harvested cells were resuspended in 50mM KPi, 500mM NaCl, 10mM EDTA, 2mM DTT, pH 7.4 with 0.1%
(v/v) Tween 20. The cell suspension was lysed on ice by sonication (5 cycles at 3
minutes per cycle, level 9, 50% duty). The lysate was clarified by centrifugation for 40
minutes at 40,000 x g. Clarified lysate was loaded onto a GSH-Sepharose FF column
phox that had been equilibrated with the lysis buffer. Bound GST-His6-p40 PX was eluted
with 15mM GSH, 100mM Tris-HCl, 50mM NaCl, 1mM EDTA, 2mM DTT, pH 8.0. The fusion protein was cleaved with PreScission Protease overnight at 4 °C while being dialyzed into 50mM KPi, 150mM NaCl, 1mM EDTA, 2mM DTT, pH 7.3. For assays
phox using GST-His6-p40 PX, the PreScission Protease was omitted. The digested fusion
protein was loaded onto a GSH-Sepharose-FF column equilibrated with 50mM KPi, pH
7.3, 150mM NaCl, 1mM EDTA, 2mM DTT. The flowthrough containing p40phox PX
was dialyzed overnight at 4 °C into 50mM Tris-HCl, pH 8.0, 1mM EDTA, 2mM DTT.
The dialysate was loaded onto a Source 15Q column and eluted with a NaCl gradient.
phox P40 was dialyzed into 20mM HEPES, pH 7.2, 137mM KCl, 1mM MgCl2, 1mM
35 phox EGTA for vesicle binding and SPR experiments. GST-His6-p40 PX was dialyzed into
TBS-T for dot blots.
Dot Blots-PIP Strips (Echelon Biosciences) were blocked for one hour at room
temperature in TBS-T with 3% fatty acid free BSA (Equitech-Bio, Inc.). The blocked
strips were incubated with 100 ng/ml of GST- or GST-His6-fusion protein overnight at 4
°C. Strips were washed three times with TBS-T followed by a one hour incubation with
primary antibody (1:3000 mouse anti-His (GE Biosciences) for His6 fusions and 1:2000
polyclonal rabbit anti-GST (Santa Cruz) for GST fusions). Strips were washed three
times with TBS-T followed by a one hour incubation with alkaline phosphatase- conjugated antibody (1:10,000 goat anti-mouse (Southern Biotech Associates) for GST-
His6 fusions and 1:10,000 goat anti-rabbit (Southern Biotech Associates) for GST
fusions). Color development was carried out by incubation in a BCIP/NBT (5-bromo-4-
chloro-3-indolyl phosphate/nitroblue tetrazolium salt) solution (Zymed Laboratories)
followed by thorough washing in water.
Preparation of Large Unilamellar Vesicles-Large unilamellar vesicles (LUVs) are
made as described in Mayer et al. (1986), Tortorella et al. (1993) and Lehman et al.
(2007). For vesicle assays, lipid mixtures containing egg PE and egg PC at a constant 4:1 molar ratio, 0.7% biotinylated-PE plus the target lipid are mixed as chloroform suspensions. For SPR experiments, lipid mixtures containing POPE:POPC in a 20:80 ratio for reference lane and POPE:POPC:x in a 20:74:6 ratio for sample lanes were mixed as chloroform suspensions. Lipid mixtures are dried under a nitrogen stream to remove
36 organic solvents. Dried lipids are rehydrated at 37 °C with mixing for one hour in buffer
containing 20mM HEPES, pH 7.2, 137mM KCl, 1mM EGTA, 1mM MgCl2. To obtain
uniformly sized vesicles, the lipid mixture is extruded nineteen times through a 100nm
pore polycarbonate membrane in a LipoFast extruder (Avestin). Size of liposomes were
verified by dynamic light scattering and concentration by quantitative lipid phosphorous.
Quantitative Lipid Phosphorous Assay-A quantitative lipid phosphorous assay is done as described in Rouser et al. (1966). Briefly, 5μl of LUV samples were placed in a glass tube along with phosphate standards containing 0, 5, 10, 25 or 50 nmoles of phosphate
(KH2PO4). Samples were dried in a heat block at 180 °C until all liquid was removed.
150μl of 70% perchloric acid was added to each sample and allowed to reflux at 180 °C for thirty minutes. The tubes were removed from the heat source and allowed to cool for
15 minutes. 900μl of water, 167μl of 2.5% ammonium molybdate and 167μl 10%
ascorbic acid were added to each tube (for final concentrations of 0.3% ammonium
molybdate and 1.2 % ascorbic acid) and incubated for fifteen minutes in a 50 °C water
bath. The absorbance of the samples was measured at 820nm paired versus water. A
standard curve was calculated using the phosphate standards and the LUV concentrations
were measured by fitting to the linear equation derived from the standard curve.
Vesicle Sedimentation Assay-5 μl LUVs (for a final concentration of 250μM) were added to solutions containing 1.25μM GST and 1.25μM GST-PX domain and incubated at room temperature for 30 minutes on a shaker plate. Streptavidin was added to each sample for a streptavidin:biotin ratio of 1:4.3 and returned to shaker plate for one minute.
37 Samples were spun at 100,000 x g for 20 minutes at 25 °C. Supernatants were removed
from each sample and the pellets were resuspended in 100μl of buffer. SDS loading
buffer was added to each sample and the samples were boiled for 10 minutes at >90 °C.
Samples were ran on a 12% SDS-PAGE gel and the % bound was determined by
densitometry using the equation: %Bound = pellet x 100/(pellet + supernatant).
Magnetic Bead-Vesicle Assay-25μl of M-280 Streptavidin Dynabeads (Invitrogen
Dynal AS) equilibrated in binding buffer were incubated with 5μl of LUVs for 30 minutes at room temperature. The supernatants were removed to remove any unbound vesicles and solutions containing 5 μM GST and 5 μM PX domain were incubated with the bead-bound LUVs for thirty minutes at room temperature with shaking. The supernatants were removed and the pellets were brought up in an equal volume of binding buffer. SDS running buffer was added to each sample and the samples were heated to >90 °C for 5 minutes. The samples were run on a 12% SDS-PAGE gel and stained with Coomassie Brilliant Blue. Bound protein was determined by densitometry of the bands. Alphaease Image Quant Software was used for densitometry. Boxes corresponding to the specific bands were used to determine the area quantified and these values were subtracted from background staining. To determine the percentage of bound protein the following equation was used: %Bound = pellet x 100/(pellet + supernatant).
Surface Plasmon Resonance-SPR experiments were carried out on a Biacore T100.
For CM3/CM5 experiments, GST fusion proteins were attached via amine-coupled anti-
GST antibody while the binding of LUVs over the protein surface was monitored.
38 Briefly, anti-GST polyclonal antibody (30μg/ml) was covalently attached via amine coupling to the dextran layer on a CM5 or CM3 chip. After blocking remaining sites with ethanolamine to prevent further coupling, the chip binding capacity was tested with free GST. Kinetic experiments were done by flowing GST fusion proteins at 5 μl/min over the chip surface and allowing for a stabilization period to ensure a stable binding surface. LUVs were injected onto the chip surface at 50μl/min for 150 seconds. The dissociation of the LUVs from the GST-fusion protein coated surface was monitored for
10 minutes. Remaining LUVs and fusion protein were removed with a two minute injection of 30mM OG followed by a 2 minute injection of 10mM glycine, pH 2.2 to regenerate the anti-GST surface.
For L1 experiments, LUVs were attached to the chip surface via hydrophobic interactions while the binding of PX domains was measured using a modified equilibrium experiment. Briefly, the surface of an L1 chip was prepared with 3 short pulses (6μl at
100μl/min) of 20mM CHAPS (3-[(3-cholamidopropyl)-dimethylammonio]-1-propane sulfonate). 0.1mM lipid solutions (POPE:POPC (20:80) for reference lane and
POPE:POPC:x (20:74:6) for sample lane) were injected at 5μl/min for a total contact time of 10 minutes resulting in 4200-4500 RUs of lipid on each lane. The lipid surface was then subjected to three short pulses (6μl at 100μl/min) of 50mM NaOH to remove loosely bound lipids. The prepared lipid surfaces were blocked with a five minute injection (at 5μl/min) of 0.1 mg/ml fatty acid-free BSA. PX domains were injected at
5μl/min for 20 minutes to allow equilibrium binding to be established. After each protein injection, the L1 chip surface was regenerated with five minute injections (at 5μl/min) of
39 40mM OG (n-octyl-β-D-glucopyranoside) and 40mM CHAPS. New lipid surfaces were prepared prior to each protein injection. Data was analyzed by using 10 seconds prior to protein injection as t(ime)=0 and subtracting the RU level at t=0 to have injections begin at RU=0. The ratio of the reference (POPE:POPC, 20:80) and sample (POPE:POPC:x,
20:7x:x) was calculated at each time point. Statistics were calculated using SigmaPlot
11. Rank anova was used for NOXO1β PX to the 8 PIPs while a paired t-test was used for p40phox PX , p47phox PX and NOXO1β PX to PA and PS.
40 Results
Dot blots for NOXO1β PX show a wide range of PIP binding (Figure 1A). The strongest signal was seen to the monophosphorylated PIPs (PI(3)P, PI(4)P, and PI(5)P) along with PS. A much lighter signal was seen to the bisphosphorylated and triphosphorylated PIPs as well as PA and PI. No detectable signal was seen to LPA,
LPC, PE, PC or S1P. Additionally, a Membrane Strip™ was tested which contains additional lipids not found on the PIP Strip™ (Figure 1B). Of lipids not on the PIP
Strip™, a strong signal was seen with NOXO1β PX to phosphatidylglycerol (PG), cardiolipin and 3-sulfogalactosylceramide. No detectable signal was seen to triglyceride, diacylglycerol, cholesterol or sphingomyelin. PIP Strips™ for p40phox PX showed a strong signal to PI(3)P and a weak signal to PA, but not to any other lipids (Figure 1C).
The PIP Strip™ for p40phox PX agrees with published data that shows that p40phox PX is
selective to PI(3)P (Ellson et al. 2001). PIP Strips™ for p47phox PX showed a range of
modest binding, with the strongest signal seen for PI(3,4,5)P3, with weak binding to the
bisphosphates (Figure 1D). This pattern is not consistent with the published lipid
binding specificity of p47phox PX (Kanai et al. 2001, Zhan et al. 2002).
The dot blots for NOXO1β PX do not completely agree with those that have been
published by multiple groups (Cheng and Lambeth 2004, Cheng and Lambeth 2005,
Takeya et al. 2006 and Ueyama et al. 2007). Cheng and Lambeth (2004, 2005) and
Takeya et al. (2006) saw binding primarily to PI(3,5)P2, PI(4)P and PI(5)P with some
binding to PI(3,4)P2. Ueyama et al. (2007) saw similar binding patterns with the addition
of binding to PA (neither Takeya et al. or Cheng and Lambeth tested for PA binding).
41
Figure 1. Dot blots of NOXO1β PX, p40phox PX and p47phox PX. A. PIP Strip™ of
GST-NOXO1β PX B. Membrane Strip™ of GST-NOXO1β PX C. PIP Strip™ of GST-
phox phox His6-p40 PX D. PIP Strip™ of GST-p47 PX. All dot blots were carried out in
TBS-T according to manufacturer’s instructions. Spots were visualized via BCIP/NBT
solution. The dot blots for NOXO1β PX show that it binds to a variety of negatively
charged lipids and does not interact with neutral lipids. P40phox PX only binds to PI(3)P
while p47phox PX also exhibits a broader binding profile like that seen in NOXO1β PX.
42
A NOXO1β PX B NOXO1β PX
LPA S1P TG PI LPC PI(3,4)P2 DAG PI(4)P PI PI(3,5)P2 PA PI(3)P PI(4,5)P2 PI(4,5)P2 PS PI(4)P PI(3,4,5)P PI(3,4,5)P3 3 PI(5)P PA PE Cholesterol PE PS PC Sphingomyelin PC Blank PG Sulfatide Cardiolipin Blank
C p40phox PX D p47phox PX
LPA LPA S1P S1P LPC PI(3,4)P LPC 2 PI(3,4)P2
PI PI(3,5)P2 PI PI(3,5)P2 PI(3)P PI(4,5)P2 PI(3)P PI(4,5)P2 PI(4)P PI(3,4,5)P3 PI(4)P PI(3,4,5)P3 PI(5)P PA PI(5)P PA PE PS PE PS PC Blank PC Blank
43 While our dot blots for NOXO1β PX were similar in terms of binding to the
monophosphorylated PIPs, we saw similar intensities to all three of the bisphosphorylated
PIPs (that is, not a stronger signal to PI(3,5)P2 over the other bisphosphorylated PIPs). In
addition, the signal for PS was much stronger than the signal for PA in our dot blots.
With some slight discrepancies in the binding profiles obtained by the dot blots, further
characterization of the lipid binding profile of NOXO1β PX was studied using both
LUV-based assays and SPR.
A vesicle sedimentation assay that was developed and used to characterize the
binding profile of p47phox PX to the PIPs and acidic lipids including PA and PS (Ledford
2004, Lehman et al. 2007) was an attractive starting point to test the lipid binding of
NOXO1β PX. After gels for the assay were run, it was noticed that little to no bands were seen for NOXO1β PX in the soluble fractions, even with the no lipid control sample
(Figure 2). Different fusion constructs (GST- vs MBP-tagged NOXO1β PX) and assay
buffers were tested to see if the missing bands were a result of non-optimal conditions.
After careful testing and troubleshooting, both GST-NOXO1β PX and MBP-NOXO1β
PX are not stable to the centrifugation step needed to efficiently pellet the crosslinked liposomes. Measuring the concentration of the tagged NOXO1β PX before and after one spin revealed that the protein concentration dropped by about half. Although marginal improvements were seen with different fusions and buffer conditions, there was still a significant loss of protein during the centrifugation step.
The bands for the pellet fractions in the LUV assay did increase as the target lipid increased (as well as no band visible in the no lipid control pellet), thus it appeared that the assay was working and a different separation technique was needed. Using
44
Figure 2. 12% SDS-PAGE Gel of GST-NOXO1β PX Binding to PA LUVs. The LUV
samples are labeled at the top of the gel. LUVs used were no lipid (NL) and 0, 10, 20,
40, 60 and 99.3 mol % PA-containing LUVs. The soluble (S) and pellet (P) fractions are labeled on the bottom of the gel. The band under the LUV labeling corresponds to GST-
NOXO1β PX while the bands above the fraction labeling correspond to GST. The amount of GST-NOXO1β PX in the pellets (P) increased as the amount of PA increased.
However, little to no GST-NOXO1β PX was seen in the soluble (S) fractions.
45
0% 10% 20% 40% 60% 99.3% NL PA PA PA PA PA PA
S P S P S P S P S P S P S P
46
streptavidin-linked magnetic beads provided an alternate method to separate bound and
unbound fractions.
The assay developed using the magnetic beads was intended to replace the vesicle
sedimentation assay for NOXO1β PX. However, the magnetic bead assay proved to be
unreliable and problematic. The bands in the soluble and pellet fractions were visible,
but their combined intensities varied between samples (Figure 3). The total protein
between each soluble/pellet pair was not consistent. This is possibly due to problems in
gel loading. The soluble samples were often dried out after heating, despite significantly
shortening the sample heating time, such that the full volume (25μl) could not be loaded
onto the gel. For the pellet samples, the magnetic beads would occasionally clog the gel
loading tip making even loading in the wells difficult.
These problems can be seen in the densitometry on gels for the assay (Figure 4).
The percent bound value did not increase with increasing lipid and the values were often
very different between repeats, creating large errors. This was seen for p40phox PX, p47phox PX and NOXO1β PX for all lipids tested. As the assay currently stands, it is not
reliable to use to characterize lipid binding. It is not certain if scaling the assay volume
up would improve the assay reliability.
To characterize the lipid binding of NOXO1β PX, SPR experiments were used.
SPR can provide ways to directly measure binding constants (Kd) as well as binding rates
phox (kon and koff). A pilot experiment measuring the binding of p47 PX to 0 and 10% mol
PA was performed by Biacore, Inc. The pilot experiments measured LUV binding to
47
Figure 3. 12% SDS-PAGE Gel of NOXO1β PX Binding to PI(5)P LUVs. Magnetic bead assay testing NOXO1β PX binding to PI(5)P LUVs. The lipid samples are labeled at the top of the gel. LUVs used were 0, 0.3, 1, 3 and 10mol % PI(5)P-containing LUVs.
The soluble (S) and pellet (P) fractions are labeled on the bottom of the gel. The band under the lipid sample labeling correspond to GST while the bands above the soluble/pellet fraction labeling correspond to NOXO1β PX. Although bands are visible in both the soluble and pellet fractions, densitometry on those bands do not show and increase in binding of NOXO1β PX as the amount of PI(5)P increases (Figure 4).
48
0% 0.3% 1% 3% 10% PI(5)P PI(5)P PI(5)P PI(5)P PI(5)P
SSSSSPP PP P
49
Figure 4. The Magnetic Bead Assay is not a Reliable Method for Measuring Lipid
Binding. A. magnetic bead assay for NOXO1β PX to PI(5)P B. magnetic bead assay for
p47phox PX to PI(3)P C. magnetic bead assay for p40phox PX to PI(3)P. The LUV binding
assay using magnetic beads to separate the soluble from the bound fractions was
unreliable and problematic. The data did not indicate that either NOXO1β PX, p47phox
PX or p40phox PX increased binding as the amount of the target lipid increased.
Additionally, the percent bound in the no lipid (NL) and 0% PIP was high for all three
PX domains. Data represent mean ± SD of three individual runs.
50 A
70
60
50
40 PX Bound β 30
% NOXO1 20
10
0 NL 0 0.3 1 3 10 mol % PI(5)P
B 100
80
60
PX Bound phox phox 40 % p47
20
0 NL 0 0.3 1 3 10 mol % PI(3)P
C
80
60
PX Bound 40 phox % p40 20
0 NL 0 0.3 1 3 10 mol % PI(3)P
51 a surface of GST-p47phox PX and although binding was seen by the LUVs to the chip
surface, Biacore, Inc. noticed several problems that needed to be addressed. The first
problem was that the Response Units (RU) were above the level that Biacore, Inc. has
confidence in for determining accurate binding rates. The second problem was that no
dissociation was seen with the PA-containing vesicles. The major suggestion was to use
a much lower percentage of PA in the LUVs (<1 mol %).
Initial experiments carried out here at Wake Forest presented a major problem that was not noticed by Biacore, Inc. The LUVs were binding more to the reference lane, which did not contain anti-GST antibody or GST-fusion protein, than the sample lane.
Upon careful review, initial pilot experiments completed by Biacore, Inc. used a CM3 chip. While both the CM3 and CM5 chip contain a dextran matrix, the CM3 chip has a much shorter dextran matrix than the CM5 (~30nm and ~100nm respectively). Using a
CM3 chip for the experiments did eliminate most of the non-specific binding of the
LUVs to the chip surface.
Another problem with the kinetic experiments was that little to no dissociation was seen by the PA LUVs from GST-p47phox PX. Very little dissociation was seen at
lower PA concentrations (<10 mol % PA) and no dissociation was seen at the higher PA
concentrations (>10mol % PA) (Figure 5). It was finally concluded that this
experimental setup was not suitable for studying the PX-lipid interaction. Further
Biacore experiments would use an L1 chip, which contains a hydrophobic surface to
attach lipids. The L1 chip setup is the reverse of the CM3/CM5 experimental setup in
that the lipids are attached to the chip surface and PX domains are injected over the lipid
surface to measure binding.
52
Figure 5. PA LUVs Do Not Dissociate from a GST-p47phox PX Coated Surface.
Sensorgram of varying PA-containing LUVs over a GST-p47phox PX surface. Time (in
seconds) is on the x-axis and arbitrary response units (RU) are on the y-axis. An increase
in binding (higher RUs) is seen with increasing amounts of PA in the LUVs, but the
LUVs also do not dissociate from the GST-p47phox PX surface.
53
10000 No Lipid 0% PA 1% PA 9000 10% PA 20% PA 60% PA 8000
7000
6000 Response Units (RU) Response
5000
4000 600 800 1000 1200 1400 1600 Time (sec)
54 Initial experiments testing lipid binding of NOXO1β PX to PIP-containing membranes using SPR revealed a high level of “background” binding to a PE/PC bilayer
(data not shown). Systematic testing was done to see if the high level of background binding was due to a specific lipid headgroup (PE vs. PC), lipid source (natural vs synthetic) or if changing the buffer condition could minimize this observation.
In the case of lipid source, the lipids form different supported structures on the L1 chip surface depending on their source. For natural lipids (egg sources of PE and PC), the LUVs tend to attach to the L1 chip surface as intact liposomes. This can be seen in the high level of RUs (>10,000 RUs). Anderluh et al. (2005) visualized the lipid-loaded chip surface by fluorescence microscopy and distinct, intact vesicles could be seen on the surface of the dextran matrix. For synthetic lipids (POPE, POPC), the LUVs tend to plate out and form a supported bilayer on the chip surface. This is seen by a much lower level of RUs (4000-6500 RUs) after lipid absorption and was visualized by Erb et al. (2000) using fluoresence microscopy and AFM. It was possible that the non-specific binding that was seen was due to the nature of the chip surface when natural lipids were used.
However, only a minimal difference (< 8%) was seen between using a lipid surface composed of natural versus synthetic lipids (data not shown).
Next, the “background” lipid composition was varied to see if NOXO1β PX was preferentially binding to one headgroup (PE or PC) over another. The ratio of PE:PC was varied from 0:100, 20:80, 50:50 and 80:20. There was < 5% difference in binding level of 1μM NOXO1β PX between the 0:100 and 80:20 PE:PC lipids (data not shown). These results combined with the results of the PIP Strip indicate that NOXO1β PX binds to neutral membranes with no preference towards either a PE or PC headgroup.
55 Finally the components of the buffer were examined to see if raising the NaCl concentration or adding phosphate to the buffer would reduce the nonspecific binding.
The initial NaCl concentration was 100mM. Increasing the NaCl concentration to
150mM did reduce the binding level of 1μM NOXO1β PX by ~50% (Figure 6A and 6B) for both a 0:100 and 80:20 POPE:POPC surface. Phosphate concentrations of 2mM,
25mM and 50mM NaPi were tested to see what effect it had on binding. The decrease in binding of 1μM NOXO1β PX was ~20% for 2mM phosphate, ~80% for 25mM phosphate and >90% for 50mM phosphate (Figure 7A and 7B) for both a 0:100 and
80:20 POPE:POPC surface as compared to buffer containing no phosphate.
Increasing the salt concentration and/or adding a small amount of phosphate to the buffer reduced the binding of NOXO1β PX to a (PO)PE/(PO)PC lipid surface. The next question was whether this buffer condition would also decrease NOXO1β PX binding to a PIP-containing surface. The results indicated that there was an overall decrease in the lipid binding of NOXO1β PX, but the decrease was seen in both the reference
(POPE/POPC) and sample (3% PI(4,5)P2) lanes (Figure 8).
The high background binding seen in the SPR experiments agrees with data published by two independent groups who noted that in mammalian cell lines transfected with wildtype NOXO1β, the protein was localized at the plasma membrane without external agonist (Cheng and Lambeth 2004, Takeya et al. 2006). It was concluded that the binding to a neutral (PE/PC) membrane was a real observation for the NOXO1β PX domain and is a feature that has not been reported for other PX domains.
While the high background binding eliminated the possibility of measuring direct binding constants and rates, SPR experiments were used to determine the ratio of protein
56
Figure 6. Increasing the NaCl Concentration Decreased Binding of NOXO1β PX to
a POPC/POPE Surface. A. 1μM NOXO1β PX binding to a 100% POPC surface in either 100mM NaCl (black) or 150mM NaCl (red) B. 1μM NOXO1β PX binding to a
80:20 POPE:POPC surface in either 100mM NaCl (black) or 150mM NaCl (red).
Increasing the NaCl concentration was enough to decrease the level of NOXO1β PX
binding by approximately half for both lipid surfaces.
57 A
1000
800
100mM NaCl 150mM NaCl 600
RU 400
200
0 0 20 40 60 80 100 120 140 160 180 200 Time (sec)
B
1000
100mM NaCl 150mM NaCl 800
600
RU 400
200
0 0 20 40 60 80 100 120 140 160 180 200
Time (sec)
58
Figure 7. The Addition of Phosphate Decreases Binding of NOXO1β PX to a
POPE/POPC Surface. A. 1μM NOXO1β PX binding to a 100% POPC surface in either
0 (black), 5 (red), 25 (green) or 50mM (yellow) NaPi B. 1μM NOXO1β PX binding to a
POPE:POPC (80:20) surface in either 0 (black), 5 (red), 25 (green) or 50mM (yellow)
NaPi. An incremental decrease in binding of 1μM NOXO1β PX to a POPC or
POPE:POPC surface was seen upon addition of NaPi to the sample buffer. The majority
of binding (>90%) was diminished with the addition of 50mM NaPi.
59 A
1000
800
600 0mM NaPi RU 5mM NaPi 400 25mM NaPi 50mM NaPi
200
0 0 50 100 150 200 Time (sec)
B
1000
800
600 0mM NaPi RU 5mM NaPi 400 25mM NaPi 50mM NaPi
200
0 0 20 40 60 80 100 120 140 160 180 200 Time (sec)
60
Figure 8. The Addition of Phosphate Decreases Binding of NOXO1β PX to a 3%
PI(4,5)P2 Surface. Sensorgram of 0μM or 4μM NOXO1β PX binding to a 3% PI(4,5)P2 surface in either 0mM, 2mM or 15mM NaPi. The addition of 2mM NaPi was sufficient
to decrease binding of 4μM NOXO1β PX by about half. Addition of 15mM NaPi brought binding levels to that of 0μM NOXO1β PX.
61
1200
no Pi, 0μM NOXO1β PX 1000 no Pi, 4μM NOXO1β PX
2mM Pi, 0μM NOXO1β PX
800 2mM Pi, 4μM NOXO1β PX
15mM Pi, 0μM NOXO1β PX
15mM Pi, 4μM NOXO1β PX RUs 600
400
200
0 0 200 400 600 800
Time (sec.)
62 bound to a PIP or anionic lipid-containing lipid surface over that of a background, or
neutral lipid surface. The same experimental protocol was done for p40phox PX and
p47phox PX to provide a comparison of PHOX/NOX-related PX domains under the same
conditions. All PIPs along with PA and PS were tested with NOXO1β PX while p40phox
phox PX and p47 PX were tested with only PI(3)P and PI(3,4)P2.
One initial distinction between the PX domains is that p40phox PX does not bind to
background surfaces while NOXO1β PX does (Figure 9). NOXO1β PX shows a high level of binding to a POPE:POPC background, with a slight increase in binding to the lane including 6% PIP (Figure 9A) which is in contrast to p40phox PX (Figure 9B).
P40phox PX showed low levels of binding to a POPE:POPC background, with a large
increase in binding to PI(3)P.
NOXO1β PX generally did not bind significantly above background to the PIPs
(Figure 10). All of the PIPs showed little to no binding above background up to the 1μM concentration point. At 5μM NOXO1β PX, most of the PIPs reached their maximal binding, with PI(4,5)P2 statistically significant over PI (Figure 10B). NOXO1β PX did
not exhibit any binding above background to PI and only saw binding just above
background for the monophosphates (PI(3)P, PI(4)P and PI(5)P) at the highest
concentration of NOXO1β PX. The binding by two of the bisphosphates (PI(3,4)P2 and
PI(3,5)P2) and of PI(3,4,5)P3 were very similar at the higher concentrations. These data
suggests that although NOXO1β PX does bind to PIPs over background, it is not a very
strong interaction.
63
Figure 9. NOXO1β PX Binds to Background Phospholipids. A. Sensorgram of
NOXO1β PX to a 80:20 POPE:POPC surface (black) and a 6% PI(4,5)P2 surface (red).
B. Sensorgram of p40phox PX to a 80:20 POPE:POPC surface (black) and a 6% PI(3)P surface (red). NOXO1β PX exhibits high binding to a neutral membrane while p40phox
PX binds very little to neutral membranes. Conversely, NOXO1β PX only shows a
phox modest increase over background binding with the addition of PI(4,5)P2 while p40 PX
shows a robust increase over background.
64 A
1200 NOXO1β PX
1000
800
600 RU
400
200 POPE:POPC
6% PI(4,5)P2 0 0 200 400 600 800 1000
Time (sec)
B
1000 phox p40 PX POPE:POPC 6% PI(3)P 800
600
RU 400
200
0 0 200 400 600 800 1000 Time (sec)
65
Figure 10. NOXO1β PX Binds to PI(4,5)P2. A. Ratio of sample lane/reference lane
RUs versus NOXO1β PX concentration for PI and the monophosphate PIPs. B. Ratio of sample lane/reference lane RUs versus NOXO1β PX concentration for PI and the bis- and triphosphate PIPs. NOXO1β PX does not bind above background to PI and only slightly above background to the monophosphates. A larger increase in binding above background is seen for the bis- and triphosphate, with only PI(4,5)P2 being significant over PI (B, marked with asterisk). Data represent mean±stdev of three separate experiments.
66 A
1.4
1.2
1.0
0.8
0.6
0.4 6% PI 0.2 6% PI(3)P
Ratio RU Sample Reference Lane/RU Lane 6% PI(4)P 6% PI(5)P 0.0 024681 0 [NOXO1β PX] μM
B
1.4 *
1.2
1.0
0.8
0.6
0.4 6% PI 6% PI(3,4)P2 6% PI(3,5)P 0.2 2
Ratio RU Sample Lane/RU Reference Lane Reference Lane/RU Sample RU Ratio 6% PI(4,5)P2
6% PI(3,4,5)P3 0.0 0246810 [NOXO1 β PX] μM
67 For p40phox PX, significant binding above background was seen for both 5μM and
10μM p40phox PX to PI(3)P (Figure 11A). No binding above background was seen for
phox PI(3,4)P2 over the entire p40 PX concentration range. These data are in agreement
with published results for p40phox PX; that it specifically binds to PI(3)P only (Kanai et
al. 2001, Ellson et al. 2001). P47phox PX showed no binding above background to PI(3)P
phox and no statistically significant binding to PI(3,4)P2 (Figure 11B). The data set for p47
PX to PI(3,4)P2 was variable in the response at the higher protein concentrations, which
is evident by the large error bars. After multiple attempts of rerunning the p47phox PX
data points, the same variability was seen. The best guess for this observation is that
there was a problem with the protein sample used in the experiments. The p47phox PX
data should be rerun with freshly prepared protein to see if the expected results, that is,
significant binding to PI(3,4)P2 (Kanai et al. 2001 Karathanassis et al. 2002), is seen.
Lastly, binding of NOXO1β PX to anionic lipids was measured using the same
protocol as for the PIPs. This was done due to the results of the dot blots with a strong
signal to PS (Figure 1A), the sequence homology to the region of p47phox PX that is
proposed to interact with anionic lipids (Karathanassis et al. 2002) and the basic nature of
NOXO1β PX (theoretical pI=10.98, calculated by ExPASy ProtParam). The only difference for the PA and PS experiments was that sample surfaces contained only 0.5%
PA or PS (as opposed to the 6% PIP used). Such a low amount was used based on initial pilot experiments performed by Biacore, Inc. (discussed previously in this chapter). At this low of a concentration of PA and PS, no binding above background is seen with
NOXO1β PX (Figure 12). Had a higher concentration of PA and PS been used, it would
68
phox phox Figure 11. P40 PX and p47 PX bind to PI(3)P and PI(3,4)P2 Respectively. A.
Ratio of binding to 6% PI(3)P (closed circles) or 6% PI(3,4)P2 (open circles) to reference lane (POPE:POPC) versus p40phox PX concentration B. Ratio of binding to 6% PI(3)P
(closed circles) or 6% PI(3,4)P2 (open circles) to reference lane (POPE:POPC) versus
p47phox PX concentration. P40phox PX reaches maximal binding to PI(3)P at 5μM. P40phox
phox PX shows no binding to a 6% PI(3,4)P2 lane over background. Data points for p40
PX at 5μM and 10μM for PI(3)P are significant with P-values <0.03 (indicated by *).
P47phox PX shows no binding above background to PI(3)P and shows no significant binding to PI(3,4)P2, despite PI(3,4)P2 being the PIP it binds the most to. Graph and
statistics were made using SigmaPlot 11. Data represent mean ± SD of three separate
experiments.
69
A 20 6% PI(3)P
6% PI(3,4)P2
15 *
10 *
5
Ratio RU Sample Lane/RU Reference Lane Reference Lane/RU Sample RU Ratio
0 0246810 [p40phox PX] μM
B 6 6% PI(3)P 6% PI(3,4)P2 5
4
3
2
1
Ratio RU Sample Lane/RU Reference Lane Reference Ratio Lane/RU RU Sample
0 0246810
phox [p47 PX] μM
70
Figure 12. NOXO1β PX Does Not Bind to Low Levels of Anionic Lipids. Ratio of
binding of NOXO1β PX to 0.5% PS (closed diamonds) or PA (open diamonds) to 80:20
POPE:POPC background. NOXO1β PX does not bind above background to either PA or
PS at the concentrations used in the assay. Graph and statistics were done using
SigmaPlot 11. Data represent mean ± SD of three separate experiments.
71
1.4 0.5% PS 0.5% PA
1.2
1.0
0.8
0.6
0.4
0.2 RatioRU Sample Lane Reference Lane/RU
0.0 0246810
[NOXO1β PX] μM
72 seem that NOXO1β PX would bind more to the more negatively charged surface.
However, from this data set it is not apparent if NOXO1β PX would bind more to PA or
PS or bind to them in equal amounts.
73 References
Anderluh, G., Beseničar, M., Kladnik, A., Lakey, J.H. and Maček, P. (2005) Properties of nonfused liposomes immobilized on an L1 Biacore chip and their permeabilization by a eukaryotic pore-forming toxin. Anal. Biochem. 344, 43-52. Cheever, M.L., Sato, T.K., de Beer, T., Kutateladze, T.G., Emr, S.D. and Overduin, M. (2001) Phox domain interaction with PtdIns(3)P targets the Vam7 t-SNARE to vacuole membranes. Nat. Cell. Bio. 3, 613-618. Cheng, G. and Lambeth, J.D. (2004) NOXO1, regulation of lipid binding, localization and activation of Nox1 by the Phox homology (PX) domain. J. Biol. Chem. 279, 4737-4742. Cheng, G. and Lambeth, J.D. (2005) Alternative mRNA splice forms of NOXO1: Differential tissue expression and regulation of Nox1 and Nox3. Gene 356, 118- 126. Erb, E.-M., Chen, X., Allen, S., Roberts, C.J., Tendler, S.J.B., Davies, M.C and Forsén, S. (2000) Characterization of the surfaces generated by liposome binding to the modified dextran matrix of a surface plasmon resonance sensor chip. Anal. Biochem. 280, 29-35. Echelon Biosciences (2006) Protocol P-6000 Rev: 5 (01/11/2006) Ellson, C.D., Gobert-Gosse, S., Anderson, K.E., Davidson, K., Erdjument-Bromage, H., Tempst, P., Thuring, J.W., Cooper, M.A., Lin, Z-Y., Holmes, A.B., Gaffney, P.R.J., Coadwell, J., Chilvers, E.R., Hawkins, P.T. and Stephens, L.R. (2001) PtdIns(3)P regulates the neutrophil oxidase complex by binding to the PX domain of p40phox. Nat. Cell Bio. 3, 679-682. Kanai, F., Liu, H., Field, S.J., Akbary, H., Matsuo, T., Brown, G.E., Cantley, L.C., and Yaffe, M.B. (2001) The PX domains of p47phox and p40phox bind to lipid products of PI(3)K. Nat. Cell Biol. 3, 675–678. Karathanassis, D., Stahelin, R.V., Bravo, J., Perisic, O., Pacold, C.M., Cho, W., Williams, R.L. (2002) Binding of the PX domain of p47phox to phosphatidylinositol 3,4-bisphosphate and phosphatidic acid is masked by an intramolecular interaction. EMBO J. 21, 5057-5068. Ledford, B. (2004) Characterization of phoshpolipid binding to the NADPH oxidase component p47-phox. Lehman, N., Ledford, B., Di Fulviom, M., Frondorf, K., McPhail, L.C. and Gomez- Cambronero, J. (2007) Phospholipase D2-derived phosphatidic acid binds to and activates ribosomal p70 S6 kinase independently of mTOR. FASEB J. 21, 1075- 1087. Mayer, L.D., Hope, M.J., Cullis, P.R. (1986) Vesicles of variable sizes produced by a rapid extrusion procedure. Biochim. Biophys. Acta. 858, 161-168. Narayan, K. and Lemmon, M.A. (2006) Determining selectivity of phosphoinositide- binding domains. Methods 39, 122-133 Ponting, C.P. (1996) Novel domains in NADPH oxidase subunits, sorting nexins and PtdIns 3-kinases: Binding partners of SH3 domains? Protein Sci. 5, 2353-2357. Rouser, G., Siakotas, A.N. and Fleischer, S. (1966) Quantitative analysis of phospholipids by thin-layer chromatography and phosphorous analysis of spots. Lipids 1, 85-86.
74 Sato, T.K., Overduin, M. and Emr, S.D. (2001) Location, location, location: membrane targeting directed by the PX domain. Science 294, 1881-1885. Simonsen, A. and Stenmark, H. (2001) PX domains: attracted by phosphoinositides. Nat. Cell Bio. 3, E179-E182. Stahelin, R.V., Burian, A., Bruzik, K.S., Murray, D., Cho, W. (2003) Membrane binding mechanism of the PX domains of NADPH oxidase p40phox and p47phox. J. Biol. Chem. 278, 14469-14479. Song, X., Xu, W., Zhang, A., Huang, G., Liang, X., Virbasius, J.V., Czech, M.P. and Zhou, G.W. (2001) Phox homology domains specifically bind phosphatidylinositol phosphates. Biochemistry 40, 8940-8944. Takeya, R., Taura, M., Yamasaki, T., Naito, S. and Summimoto, H. (2006) Expression and function of Noxo1gamma, an alternative splicing form of the NADPH oxidase organizer 1. Febs J. 273, 3663-3677. Tortorella, D., Ulbrandt, N.D., London, E. (1993) Simple centrifugation method for efficient pelleting of both small and large unilamellar vesicles that allows for convenient measurement of protein binding. Biochemistry 32, 9181-9188. Ueyama, T., Lekstrom, K., Tsujibe, S., Saito, N and Leto, T.L. (2007) Subcellular localization and function of alternatively spliced Noxo1 isoforms. Free Radic. Biol. Med. 42, 180-190. Wishart, M.J., Taylor, G.S. and Dixon, J.E. (2001) Phoxy lipids: revealing PX domains as phosphoinositide binding modules. Cell 105, 817-820. Xu, Y., Hortsman, H., Lifong, S., Wong, S.W. and Hong, W. (2001) SNX3 regulates endosomal function through its PX-domain-mediated interaction with PtdIns(3)P. Nat. Cell Bio. 3, 658-666. Yaffe, M.B. (2002) The p47phox PX domain: two heads are better than one! Structure (Camb). 10, 1288-1290. Yu, J.W. and Lemmon, M.A. (2001) All phox homology (PX) domains from Saccharomyces cerevisiae specifically recognize phosphatidylinositol 3- phosphate. J. Biol. Chem. 276, 44179-44184. Yu, J.W., Mendrola, J.M., Audhya, A., Singh, S., Keleti, D., DeWald, D.B., Murray, D., Emr, S.D. and Lemmon, M.A. (2004) Genome-wide analysis of membrane targeting by S. cerevisiae pleckstrin homology domains. Mol. Cell 13, 677-688. Zhan, Y., Virbasius, J.V., Song, X., Pomerleau, D.P., Zhou, G.W. (2002) The p40phox and p47phox PX domains of NADPH oxidase target cell membranes via direct and indirect recruitment by phosphoinositides. J. Biol. Chem. 277, 4512-4518.
75 CHAPTER III
SOLUTION STRUCTURE OF NOXO1β PX
N. Y. Davis and D.A. Horita
The following experiments shown were performed by N. Y. Davis. Dr. K. Shen prepared
charged gels used for RDC experiments. Dr. D. A. Horita prepared nanodiscs, acquired
13 and analyzed C-edited NOESY-HSQC, assigned methyl groups, Gln sidechain NH2 and
Phe resonances and helped with structure calculations.
Introduction
An integral part of the NADPH oxidase complexes are the multiple interactions
between proteins in the complex and the membrane. A protein domain first identified in
two proteins within the PHOX complex, the PX domain, is necessary for several of the
protein-lipid interactions within the NADPH oxidases. In the PHOX complex, the PX
domain of p40phox specifically targets PI(3)P (Kanai et al. 2001, Zhan et al. 2002,
phox Stahelin et al. 2003, Ellson et al. 2004). The PX domain of p47 binds to PI(3,4)P2 in addition to anionic lipids such as PA (Kanai et al. 2001, Karathanassis et al. 2002, Zhan et al. 2002, Stahelin et al. 2003, Ledford 2004). The PX domain of NOXO1β PX was determined to bind to both neutral membranes as well as PI(4,5)P2 (Chapter II). Among these three PX domains found in the NADPH oxidases, there are varying lipids and cellular signals being targeted by this single domain.
Structures for several PX domains have been solved by both solution-state NMR and X-ray crystallography. Although there is typically low sequence similarity between
PX domains, they share a common three-dimensional fold. The PX domain structures have a three-dimensional fold comprising three β-strands followed by four α-helices
(Bravo et al. 2001 (p40phox PX), Hiroaki et al. 2001 and Karathanassis et al. 2002 (p47phox
PX), Xing et al. 2004 (CISK PX), Zhou et al. 2003 (Grd19p PX), Lu et al. 2002 (Vam7p
PX), Stahelin et al. 2006 (Bem1p PX), Stahelin et al. 2007 (PI3K-C2α PX)). The PIP binding pocket in the PX domain is formed by residues in three regions: the β3-α1 loop, the loop connecting α1-α2 and the N-terminal end of α2.
To better understand how PX domains bind to phospholipids, structures of several
PX domains have included either soluble PI(3)P (p40phox PX ( Bravo et al. 2001) and
77 Grd19p PX (Zhou et al. 2003)) or sulfate ions (p47phox PX (Karathanassis et al. 2001) and
PI3K-C2α (Stahelin et al. 2007)). These structures provide additional residue-specific information about the protein-lipid interactions in the PX domain. This structural data coupled with mutagenesis data identify the critical residues and nature of the interactions that are necessary for PX domain-PIP binding.
Although there is residue-specific information about the PX domain-PIP headgroup interaction, little is known about how the PX domain binds to a membrane.
Studies by Stahelin et al. (2003) identified the mechanism by which p40phox PX and
p47phox PX target a membrane. Their studies showed that the membrane binding was
initially characterized by a non-specific electrostatic interaction followed by membrane
insertion of hydrophobic residues. Extensive experiments with both Vam7p PX (Lee et
al. 2006) and p40phox PX (Malkova et al. 2006) provide residue-specific information
about how the PX domain binds to a membrane. Both groups identified residues that
penetrated into the membrane (Y94 and V95 in p40phox PX; V70, L71 and W75 in Vam7p
PX), which when coupled with the PIP headgroup data can begin to shed light on how
these, along with other PX domains, bind to a membrane.
In this chapter, we present the solution structure of NOXO1β PX along with
dynamic and preliminary lipid binding experiments using solution-state NMR. The data
presented in this chapter aim to structurally rationalize the binding data presented in
Chapter II by identifying candidate residues for phospholipid binding of NOXO1β PX.
Additionally, we aim to build upon the known PX-phospholipid binding information for
PX domains by analysis of lipid binding of NOXO1β PX by solution-state NMR.
78 Experimental Procedures
Protein Expression and Purification- The sequence for human NOXO1β PX (1-144)
was cloned into pGEX-6P-1 (Amersham) or pMHPb and transformed into the E. coli
strain BL21(DE3)Codon+ RIPL (Stratagene). Cultures were grown at 32 °C (pGEX-6P-
15 13 1) or 37 °C (pMHPb) in minimal media containing 1g/L NH4Cl and 2g/L C6 glucose
2 13 2 ( H7 C6 glucose for H samples) until OD600 reached 0.6 and induced with 0.5mM IPTG
at 22 °C (pGEX-6P-1) or 15 °C (pMHPb) overnight (for 1H media) or for 72 hours (for
2H media). For DCN-ILV (2H, 13C, 15N, 1H(Iδ1,L,V)) samples, 85mg/L of α-
13 13 ketoisovalerate ( C5, 3-d1) and 50mg/L α-ketobutyric acid ( C4, 3,3-d2) was added to
cultures one hour prior to induction (Goto et al. 1999). For DNC-ILV-F (2H, 13C, 15N,
1 δ1 15 13 H(I ,L,V), N F) 100mg/L of α-ketoisovalerate ( C5, 3-d1), 50mg/L α-ketobutyric acid
13 15 ( C4, 3,3-d2) and 40mg N-phenylalanine were added to cultures one hour prior to
induction. Standard protocols for growth and purification (Chapter II) were followed
after addition of the chemicals for the DCN-ILV/ILV-F labeling. Final yield for the
pGEX-6P-1 expression system was ~3mg NOXO1β PX/L culture and the pMHPB
expression system was 9-10mg NOXO1β PX/L culture.
NMR data collection and analysis-NMR spectra were acquired at 293K on a Bruker
Avance DRX-600 spectrophotometer equipped with a triple resonance Z-gradient
cryoprobe. For NMR experiments, NOXO1β PX was dialyzed into 100mM NaPi,
100mM NaCl, 0.1mM EDTA, pH 6.9. Samples used for NOESY data were dialyzed into
2 100mM NaPi, 100mM NaCl, 0.1mM EDTA, pH 6.5. 5% (v/v) H5 glycerol was added after samples were concentrated. NOXO1β PX was concentrated to ~500μM for NMR
79 experiments used in structure calculations. For bicelle and nanodisc experiments, samples
were in 20mM NaPi, 100mM NaCl, pH 7.2 at ~100μM concentration. For lipid titration experiments, 20mM HEPES, 137mM KCl, 0.1mM EGTA, pH 7.2 was used with
~100μM NOXO1β PX.
Backbone Assignments- Standard pulse sequences were used to acquire 3D HNCA,
HNCO, HN(CA)CB, HN(CO)CA, HN(CA)CO and HN(COCA)CB data. NMR data was
processed using NMRPipe (Delaglio et al. 1995) and analyzed using NMRView (Johnson et al. 1994).
NOESY Data-Three-dimensional 15N- and 13C-edited NOESY-HSQC spectra (Xia et al.
2 15 2 13 15 2003) were collected in H2O at 20 °C on H, N (DN), H C N (DCN), DCN-ILV and
o DCN-ILV-F samples in H2O and D2O at 20 C with 200ms and 300ms mixing times. A
three-dimensional, CT-NOESY-HSQC was collected in D2O with a 300ms mixing time.
Isotopic half-filtered NOESY spectra (Otting and Wüthrich 1989) were also collected on
the DCN-ILV-F labeled sample in H2O.
Assignment Data-Methyl groups were assigned using the HmeCmeCGCBCA sequence
(Tugarinov and Kay 2003), an HmeCmeCBCA sequence which provided higher sensitivity
for Val residues and was derived from the previous sequence by deleting one carbon
transfer step, and the H(C)C(CO)NH / (H)CC(CO)NH TOCSY sequences (Link and
Wagner 1999). Data were collected on both fully DCN-ILV labeled samples and on
samples containing Leu and Val residues that were labeled 1H,13C on one methyl group
2 12 and H, C on the other methyl group (Tugarinov and Kay 2003). All Gln NH2 resonance
80 were assigned using a 2D version of the H2N(CO)CGCB sequence (Farmer et al. 1996).
Phenylalanine resonances were assigned through analysis of NOESY, COSY, and
TOCSY spectra.
Structure Calculation-Structures were calculated using Aria2.2/CNS 1.2 (Brünger et al.
1998 and Rieping et al. 2007). Restraints involving Phe side chains were set at 1.8 Å for the lower limit and either 4.0 or 6.0 Å for the upper limit, depending on cross peak intensity. All other restraints were automatically calibrated by Aria using 1.8 Å as the lower limit and 6.0 Å as the upper limit. Aria calculations used the initial unambiguous and ambiguous distance restraints, 83 χ1 ( Berjanskii et al. 2006) and 230 backbone φ+ψ
(Shen et al. 2009 and Cornilescu et al. 1999) dihedral angle restraints, 53 1H methyl chemical shifts, and 92 1HN-15N residual dipolar coupling restraints. The hydrogen bond restraint list was built iteratively based on the presence of secondary structural elements identified from backbone dihedral angles, cross-strand HN-HN NOEs, and analysis of structures using HB-PLUS (McDonald and Thornton 1994). Ultimately, we incorporated
61 hydrogen bond restraints. We followed standard ARIA protocols with the exception of longer cooling stages (300,000 and 200,000 steps) which greatly increased convergence properties (Fossi et al. 2005). Typically we generated 144 structures per iteration. The final iteration yielded 637 unambiguous NOEs. Fourteen low-energy structures were used for refinement in explicit water molecules.
Residual Dipolar Couplings (RDCs)- RDCs were measured in both uncharged, vertically stretched gels and charged, vertically compressed gels (Tycko et al. 2000). For
81 uncharged, vertically stretched gels (Chou et al. 2001, Jones and Opella 2004), 7%
acrylamide gels were made. The gel solution was poured into a 6mm Teflon funnel (New
Era) and the ends sealed. Gels were allowed to polymerize overnight followed by
extensive washing in deionized water for 72 hours After washing was complete, gels were cut into four equal lengths and allowed to dry at room temperature. Dried gels were
reconstituted by placing the gel into the 6mm funnel and adding 400μL of NOXO1β PX
(~500μM concentration in standard NMR buffer) and allowing to hydrate overnight.
Fully hydrated gels were compressed into a 5mm NMR tube using the New Era Gel
NMR sample kit. 3D transverse relaxation-optimized spectroscopy (TROSY)-HNCO-
based experiments (Yang et al. 1999) were used to obtain backbone RDCs (1D(HN, N),
1D(Co,Ca) and 3D(HN,Ca)).
For charged gels ( Meier et al. 2002, Cierpicki and Bushweller 2004) copolymerization
was done by preparing 7% acrylamide copolymer gels containing 25% 3-
(acrylamidopropyl)-trimethylammonium chloride (APTMAC). The gel mixtures were
polymerized overnight in a 19.2 cm piece of 3.2mm (inner diameter) PVC tubing
(Nalgene). After polymerization, gels were washed extensively in deionized water over
72 hours to remove any unpolymerized acrylamide copolymers and other impurities.
Gels were cut into 12 equal pieces and dried on Teflon plates to prevent the gels from
sticking and distorting during the drying process. To prepare NOXO1β PX in the
charged gels, the dried gel was put into a 5mm Shigemi tube and protein solution (300μL
of 500μM NOXO1β PX in standard NMR buffer) was added and the gel was hydrated
overnight. To achieve vertical compression of the gels, the Shigemi plunger was inserted
82 in the tube to reduce the gel length by 10%. As with the uncharged, horizontally
compressed gels, 3D TROSY-HNCO-based pulse sequences (Yang et al. 1999) were
used to measure 1D(HN, N).
1H-2H Exchange Experiments- 300μL of 500μM NOXO1β PX in standard NMR buffer
2 (100mM NaPi, 100mM NaCl, 0.1mM EDTA, pH 6.9, 5% (v/v) H5 glycerol) was
lyophilized to remove H2O. The lyophilized protein sample was brought back up in an
2 2 1 15 equal volume of cold 99.9% H2O. Immediately following addition of H2O, H, N-
HSQCs were acquired every 12 minutes for one hour and then every 24 minutes for 4.5
hours. The disappearance of 1H-N peaks was monitored over time.
Bicelle Preparation-Bicelle mixtures were prepared according to Vold et al. (1997) at a
2X concentration (long chain/short chain(q)=0.5 cL=30%). Due to the very hygroscopic
nature of the short chain lipids, DHPC (1,2-dihexanoyl-sn-glycero-3-phosphocholine)
was measured using a glove bag under nitrogen. The DHPC was solubilized into bicelle
buffer (20mM NaPi, 100mM NaCl, pH 7.2). The DMPC (1,2-dimyristoyl-sn-glycero-3-
phosphocholine) was mixed with the solubilized short chain lipids and repeatedly
vortexed between incubations at 42 °C and on ice. After mixing, bicelle solutions were
hydrated overnight at 4 °C. The final bicelle solution was clear and free of particulate.
Purified protein is added to the bicelle solution giving a final bicelle concentration of 1X
(q=0.5, cL=15%).
83 Nanodisc Preparation- POPC was prepared by drying 6.1mg of POPC to remove
chloroform and redissolving in 550ml of cholate buffer (30mg/ml NaCholate). The
POPC solution was shaken for 20 minutes at 37 °C followed by a incubation for one hour
at 37 °C. Finally, the solution was vortexed for one hour at room temperature. 2.38ml of
Apolipoprotein A-1 (ApoA1) (1mg/ml), 0.55ml of the POPC solution, 0.30ml buffer
(10mM NH4HCO3, pH 7.7) and 2.32ml of H2O were swirled together and let rest for 10
minutes at room temperature. The nanodisc mixture was dialyzed six times against 2L of
10mM NH4HCO3, pH 7.7 at room temperature with at least two hours between buffer
changes. Discs were purified on size exclusion column. Discs were concentrated to 2X
while being exchanged into 20mM NaPi, 100mM NaCl, pH 7.2. 200μM NOXO1β PX was added to make a final concentration of 100μM NOXO1β PX and 0.2mM nanodiscs
(in terms of ApoA1 concentration).
PIP Titrations-Short chain (diC4 or diC8) PIPs were resuspended in 20mM HEPES,
137mM KCl, 0.1mM EGTA, pH 7.2 for one hour at room temperature to a concentration
of 2.5nmol/μL PIP. ~100μM NOXO1β PX (20nmol protein) in 20mM HEPES, 137mM
KCl, 0.1mM EGTA, pH 7.2 was put into a 4mm Shigemi tube. 5nmol (2μL) PIP (for
protein:lipid (mol:mol) data points of 1:0, 1:0.25, 1:0.5, 1:0.75 and 1:1) or 10nmol (4μL)
PIP added (for protein:lipid data points of 1:1.5, 1:2 and 1:2.5) was added to the protein
1 15 solution followed by acquisition of a H, N HSQC. For the diC8-PI(4,5)P2 sample,
20μL of 1M NaPi, pH 7.2 was added to the 1:1 protein/lipid solution giving a final
1 15 concentration of 100mM NaPi followed by acquisition of a H, N HSQC. Protein
84 concentrations were measured by Bradford assay after the final data acquisition for each
PIP.
85 Results
Backbone Assignments
Backbone assignments for NOXO1β PX are near complete for 132 of 135 non-
proline residues. 95% of NH groups (Figure 1), 95% Cα, 89% Cβ and 95% C′ atoms
have been assigned. Sidechain NH2 for all Q residues have been assigned as well as the
NH for the indole of W30 (Figure 1). Secondary structure predicted by TALOS+ (Shen et al. 2009) (Figure 2) is consistent with observed secondary structure in other PX domains with known structures. Backbone assignments for NOXO1β PX have been deposited in the BMRB with accession code 16749.
Although NOXO1β PX is a small domain (~16.8 kDa MW), deuteration of the protein was necessary to obtain backbone carbon assignments. For 1H15N13C-labeled
protein, only 152 total Cα peaks and 15 total Cβ peaks were observed in HNCA and
HNCACB spectra respectively. The 2H15N13C NOXO1β PX contained 253 Cα peaks
and 182 Cβ peaks in HNCA and HNCACB spectra. We hypothesized that the poor
magnetization transfer in the protonated protein was a result of a longer τc (correlation time) due to the addition of glycerol to the sample buffer and the need for running experiments at a cooler temperature (293K). A calculated τc for a protein with a
theoretical molecular weight of 16,800 g/mol would have a theoretical τc of ~8.5 ns at
293K (calculated according to equations 1.41 and 1.42 from Cavanagh et al. 1996).
NOXO1β PX under NMR conditions has a calculated τc of 21.297 ns (calculated using
FastmodelFree, Cole and Loria 2003). This is equivalent to a theoretical τc for a protein
with a molecular weight of ~63,000 g/mol at 293K.
86
Figure 1. 15N, 1H HSQC of NOXO1β PX. The HSQC spectrum was obtained with
2 ~500μM NOXO1β PX in 100mM NaPi, 100mM NaCl, 0.1mM EDTA, 5% (v/v) H5 glycerol, pH 6.9 with D2O in an outer 5mm tube surrounding the 4mm Shigemi tube.
Assigned peaks, including sidechain amides and W30 indole, are indicated.
87
88
Figure 2. Predicted secondary structure of NOXO1β PX by TALOS+. Residue
number is indicated along the x-axis and the secondary structure is indicated along the y-
axis with a value of +1 for β-sheet and -1 for α-helix. The height of the bars indicate the probability of the residue being either a β-sheet (1 maximum) or α-helix (-1 maximum)
89
90 To ensure that the protein was not forming oligomers, size exclusion experiments
were done and indicated that the protein is not forming stable aggregates or oligomers.
The size exclusion data suggests that NOXO1β PX is close to the calculated size of a
monomer (16.8kDa calculated MW, 20.9kDa from column) and is slightly smaller than a
construct of the p47phox PX domain (1-151) that is seven residues larger (18.1kDa calculated MW, 21.7kDa from column) (Figure 3). In addition, 1H, 15N HSQC
experiments were run on NOXO1β PX over a range of protein concentrations (from
~50μM-500μM) and no changes in chemical shift were seen between the spectra indicating that no stable oligmer of NOXO1β PX is forming under these conditions.
Aggregation has been seen in other PX domains solved by solution state NMR (Vamp7p
PX domain, Lu et al. 2002), so it is possible that there is non-specific and non-stable
aggregation of NOXO1β PX under NMR conditions, which would factor into the large
calculated τc.
Cryogenic probes are able to increase the sensitivity of NMR experiments by increasing signal strength while decreasing noise, resulting in an increase in signal-to- noise (S/N) ratio. However, electrically conductive samples reduce this increase in S/N ratio in cryogenic probes. For NOXO1β PX solubility and stability, it was necessary for
sample buffer to contain high levels of both phosphate and salt. Sodium phosphate and
sodium chloride were chosen over their potassium counterparts due to a slightly lower
conductivity (Kelly et al. 2002). To help increase S/N ratio in experiments for NOXO1β
PX, a 4mm Shigemi tube was used for data acquisition. Voehler et al. (2006) reported that a smaller diameter NMR tube is a viable avenue for increasing S/N of high conductivity samples in a cryogenic probe.
91
Figure 3. NOXO1β PX is a monomer in solution. Size exclusion column (Superdex75) of NOXO1β PX and p47phox PX (1-151) show that although both PX domains run slightly larger than their calculated MW, they are not forming stable oligomers. RNAseA (13.7 kDa), chymotrypsin (25kDa) and ovalbumin (43kDa) were used as standards. NOXO1β
PX (16.5kDa) runs at a predicted 20.9kDa MW and p47phox PX (1-151) (18.1kDa) runs at a predicted 21.7kDa.
92
80
rnasea p47phox PX (1-151) 75 NOXO1β PX (1-149)
70
chymotrypsin
65
60 Elution Volume (ml)
ovalbumin 55
50 4.1 4.2 4.3 4.4 4.5 4.6 4.7 log (MW) 10
93 Using a 4mm vs. 5mm Shigemi tube, the sample volume is decreased. The protein
sample was put into a 4mm Shigemi tube that was placed inside a 5mm flat bottomed
tube containing D2O. Putting the D2O in the outer tube allowed for spectrometer locking
without any dilution effects to the protein sample. Although the sample volume
decreases by ~30%, the gain in S/N of data acquired in a 4mm tube is ~40% higher than
that in a 5mm tube for the same protein concentration. Utilizing a smaller tube also
decreased the 1H 90° pulse from ~15μs to 10.2μs on average.
Structure and Dynamics of NOXO1β PX
Structure calculations for NOXO1β PX produced a structure for residues 2-144
(7-149 of the protein construct) (Figure 4). The first five N-terminal residues, a residual
fusion tag, and the first Met of NOXO1β PX were not included in final structure
calculations. Structural statistics for the 14 lowest energy structures are presented in
Table 1. A total of 696 distance restraints were used in the structure calculations.
Typically, about two times that number of restraints would be wanted for use in structure calculations for a protein of this size. From the structure bundle (Figure 4A), one can see
that the regions that had the fewest restraints (and thus more possible conformations) were in loop regions and at the N- and C-termini. From the RMSD (Table 1), the backbone atoms of residues in secondary structures, that is helix or sheets, are within 1Å.
When the entire molecule is considered, the RMSD for backbone atoms increases to nearly 3Å. The Ramachandran statistics (Table 1) show that the majority of residues are in allowed regions. Of the 14 lowest energy structures, 99.3% of residues were within allowed regions of a Ramachadran plot (80.3% of the residues were in the most favored
94
Figure 4. Solution Structure of NOXO1β PX. A. Bundle of the 14 lowest energy structures as calculated using Aria2.2/CNS1.2 B. Average structure of NOXO1β PX.
Strands are colored yellow, helicies are colored red and loops are colored green. Figures created with PyMOL.
95
A.
B.
96
Table 1. NOXO1β PX Restraints and Structure Statistics. Restraints used in structure
calculations and the statistics generated from the 14 lowest energy structures used in
refinement. Restraints were obtained using Procheck and statistics were calculated from
Aria2.2
97
Table 1 NOXO1β PX Restraints and Structure Statistics Conformational Restraints NOE Restraints 696
Intra-residue 89 Sequential 159 Medium (<4 Å) 142 Long-range (>4 Å) 247
Ambiguous 59
Dihedral Angle restraints chi 83
phi +psi 230 Structure Statistics
Violations Number per RMSD Structure Distance Restraints (>0.3 Å) 0.92 (±0.79) 3.25x10-2 (±7.77x10-3)
Dihedral angle restrains (>5°) 7.07 (±2.22) 1.45 (±0.26) RMSD (Å) All residues (7-149) Secondary Structure Backbone heavy atoms 2.96 (±0.59) 0.97 (±0.24) All heavy atoms 3.39 (±0.57) 1.41 (±0.24)
Ramachandran plot
Most favored regions (%) 80.3 Additionally allowed regions (%) 18.3
Generously allowed regions (%) 0.7 Disallowed regions (%) 0.7
98 region, 18.3% in additional allowed region and 0.8% in generously allowed regions).
Only 0.7% of residues were in disallowed regions.
The PX domain of NOXO1β resembles the overall three dimensional fold seen
for other PX domains. The NOXO1β PX domain contains three anti-parallel β-strands
surrounded by four α-helicies and a 310 helix. The three β-strands correspond to residues
12-22 (β1), 27-35 (β2) and 39-45 (β3). There is a bulge in the β-sheet (V10 and Q11)
that twists the β-sheet towards the PIP binding pocket. This β-bulge is also seen in other
PX domains. The four α-helicies correspond to residues 47-56 (α1), 89-107 (α2), 112-
116 (α3) and 118-124 (α4). The short 310 helix is made by residues 128-130. A long
loop region (residues 57-88), whose length is highly variable amongst PX domains,
connects α1 to α2.
The sequence used to determine the structure of NOXO1β PX is slightly longer
than that used for most of the solved PX domain structures. In addition to the typical
secondary structural elements seen in PX domains, we also noticed a short β-strand
forming at the C-terminus (residues 141-142) (Figure 4B). This short strand was also
seen in the X-ray structure for a longer p47phox PX construct. This strand could
potentially be a secondary structural element connecting the PX domain to the SH3
domains in full length p47phox PX as well as NOXO1β PX.
The program FastmodelFree was used to calcuate the amplitude and time scale for intramolecular motions using R1 (longitudinal relaxation rate constant), R2 (transverse
relaxational rate constant) and heteronuclear NOE data. The program calculates Rex, τe
(effective correlation time for internal motions) and S2 (generalized order parameter).
99 The generalized order parameter, S2, characterizes the amplitude of the
intramolecular motions. The amplitude of S2 ranges between 0≤S2≤1, with lower values
(those nearing 0) having larger amplitudes of internal motion. Both experimental (closed
circles) and simulated (open circles) S2 values were calculated by FastmodelFree (Figure
5A). The experimental and simulated S2 values generally fit well. Of the residues that
had a large difference between experimental and simulated S2, most had a calculated S2 of 1 (that is, no internal motion). V8 had one of the largest differences between the experimental and calculated S2, with the calculated S2 being closer to the experimental
and calculated S2 values for the subsequent residues. The majority of the values below
0.8 were located in loop structures, including L74, which is found in the long loop region
connecting α1-α2 and is located near the PIP binding pocket.
The effective correlation time, τe, was also calculated by FastmodelFree (Figure
5B). Fewer residues fit this model of exchange as evidenced by fewer data points calculated by the program. In general, the difference between simulated (open triangles) versus calculated (closed triangles) was within 10% of the values. For L74, the τe, was much slower than for other residues. This residue also had a much lower S2 value as
mentioned previously. This region of NOXO1β PX appears to have different dynamics
versus the remaining protein. This area was not very well defined in the structures
(Figure 4A), and appears to be due to the dynamics of the protein and not just a lack of
distance restraints for this region.
The exchange term, Rex, did not fit data for most of the residues in NOXO1β PX
(Figure 5C). Although the simulated (open squares) and calculated (closed squares)
100
2 2 Figure 5. S , τe and Rex of NOXO1β PX. A. Calculated and simulated S of NOXO1β
PX versus residue. The peak corresponding to L74 appears to exhibit greater motion than the overall protein. B. Calculated and simulated τe of NOXO1β PX versus residue.
Motions differing from the other residues with a τe component are apparent in residue
L74. C. Calculated and simulated Rex of NOXO1β PX. Rex does not appear to be
contributing to the dynamics of NOXO1β PX. All parameters were calculated by
FastmodelFree and were graphed with SigmaPlot 11.0.
101 A
1.0
0.8
0.6
2 S
0.4
0.2
S2 S2sim 0.0 0 20 40 60 80 100 120 140 Residue B 1200 τe
τe sim 1000
800
(ps) 600 e τ
400
200
0 0 20406080100120140 Residue C 30
Rex
Rex sim 25
20 ) -1
(s 15
ex R
10
5
0 0 20 40 60 80 100 120 140 Residue
102 agreed well, there does not appear to be any region of NOXO1β PX that has different
dynamics due to exchange.
Another method of looking at protein dynamics is by using proton-deuterium
exchange experiments. In these experiments, the exchange of labile protons with
deuterons in the bulk solvent is measured by the disappearance of 1H-N peaks in a 1H,
15N HSQC measured over time. 1H-N peaks that are solvent accessible will exchange
rapidly and disappear from spectra early. 1H-N peaks that are buried, such as those found
in the protein core, will exchange more slowly and remain visible for a longer period of
time. NOXO1β PX does not appear to have a very stable core, as seen in the rapid
disappearance of 1H-N peaks in the HSQC (Figure 6). By 24 minutes after data collection began, most of the peaks had disappeared. This suggests that the protein is experiencing “breathing” motions. This rapid exchange could have contributed to the lower than expected distance restraints measured from the NMR data.
Lipid Binding Experiments
There are multiple methods for looking at protein-lipid interactions using solution-state NMR. One method is using membrane mimetics, which include micelles and bicelles. Bicelles provide a flat bilayer surface for examining protein-lipid interactions in contrast with the curved lipid surface of micelles. Previous data from our lab indicate that the PX domain unfolds when micelles are added to the protein solution.
Our hypothesis is that the PX domain has a large surface that interacts with the lipid bilayer and the high curvature of the micelles causes the PX domain to become disordered. When using isotropic bicelles, the PX domain appears to remain folded.
103
Figure 6. NOXO1β PX does not have a stable core. A. 1H, 15N-HSQC of NOXO1β
2 PX in H2O over time. The disappearance of peaks in the spectrum is monitored over
time and are a result of the 1H-N exchanging with 2H in the bulk solvent. After the first
24 minutes, the majority of peaks are missing, indicating that those 1H-N have exchanged
with 2H. B. Map of NOXO1β PX for 1H-2H exchange. Residues shown in yellow still
have a signal after 24 minutes of data acquisition. These residues were resistant to 1H-2H
exchange.
104
A Black: T=0min. Blue: T=24min. Red: T=48min. Green: T=96min.
1 5 N
p p p m
1H ppm
B
105 The approach using bicelles was to look at the 1H-N chemical shift changes of
NOXO1β PX with a DMPC/DHPC bicelle. The 1H,15N-HSQC of NOXO1β PX with
DMPC/DHPC bicelles appears to be folded due to positions of peaks compared to an
HSQC of NOXO1β PX alone (Figure 7A and 7B). However, the poor resolution of the
spectrum, as evidence by the few number of peaks that appear in the spectrum, indicate
that the isotropic bicelles under these experimental conditions are not a suitable
membrane mimetic for studying the protein-lipid interaction of NOXO1β PX. We
hypothesized that the poor spectra quality is due to the protein interacting with the bicelle
and creating a larger complex which is leading to the poor spectrum resolution. This can
be remedied by running TROSY-based NMR pulse sequences, which are designed for
large macromolecular complexes. In addition to the poor spectrum resolution, the life of
the NOXO1β PX-bicelle sample was very short (< 24 hours). We next switched to POPC
nanodiscs (Nath et al. 2007). The nanodiscs, which contain ApoA1 wrapped around a
bilayer of POPC, are more stable over time. Using a combination of TROSY-based pulse
sequences and POPC nanodiscs should provide the experimental conditions and sample
stability needed to examine NOXO1β PX’s interaction with a POPC lipid surface.
Using POPC nanodiscs, a 1H, 15N TROSY-HSQC appears to be folded (Figure
7C). The spectrums show a large number of chemical shift changes from NOXO1β PX
(black) and NOXO1β PX with POPC nanodiscs (red) (Figure 8A). A map of the shifted residues on the structure of NOXO1β PX (Figure 8B) show that the changes are spread across the PX domain, with concentrated areas on either end of the β-sheet. These data suggest that NOXO1β PX binds to a POPC membrane in a non-specific manner and does not indicate NOXO1β PX binds to neutral membranes on a single face.
106
Figure 7. POPC Nanodiscs are a suitable membrane mimetic for NOXO1β PX. A.
1H,15N-HSQC of NOXO1β PX B. 1H,15N-HSQC of NOXO1β PX with isotropic
DMPC/DHPC bicelles C. 1H,15N-TROSY-HSQC of NOXO1β PX with POPC nanodiscs.
The addition of isotropic bicelles to NOXO1β PX domain does not appear to unfold the protein as evidenced by the positions of visible peaks in the spectrum compared to that of wild type protein. While isotropic bicelles might provide a useful membrane mimetic for studying the PX-lipid surface interaction, the sample stability is extremely short and the spectrum resolution is poor. The use of POPC nanodiscs is a suitable system for analysis of the NOXO1β PX-lipid interaction. As seen in C, the NOXO1β PX domain remains folded in the presence of POPC nanodiscs.
107
A
.
15 N p p m
1H ppm
B .
1 5 N
p p m
1H ppm
C .
1 5 N p p m
1H ppm
108
Figure 8. NOXO1β PX exhibits non-specific binding to a POPC nanodisc. A. 1H,
15N-TROSY-HSQC of NOXO1β PX (black) and NOXO1β PX with POPC nanodiscs
(red). B. Chemical shift map of NOXO1β PX for POPC nanodisc-dependent changes.
NOXO1β PX undergoes chemical shift changes across the entire structure (shown in yellow) upon addition of POPC nanodiscs, suggesting that NOXO1β PX binds to POPC nanodiscs in a non-specific manner.
109
A
N ppm N 15
1H ppm
B
110
Another approach for looking at the NOXO1β PX-lipid interaction is by titrating
in soluble, short chain lipids with the protein and monitoring any changes in 1H-N
chemical shifts. We first titrated diC8-PI(4,5)P2 into NOXO1β PX in a modified binding
assay buffer (20mM HEPES, pH 7.2, 137mM KCl, 0.1mM EGTA) using a 1:0, 1:0.25,
1:0.5, 1:0.75 and 1:1 molar ratio of NOXO1β PX: diC8-PI(4,5)P2. Changes in chemical
shift were seen upon the first addition of PI(4,5)P2 (Figure 9A). However, the spectra
quality dramatically decreased after addition of PI(4,5)P2 to a 1:0.75 mol ratio.
To test if some type of aggregate (whether due to protein aggregation or
formation of micelles) was forming and was creating a large molecular complex and thus
leading to the poor spectra quality, 100mM NaPi was added to the NMR sample to
disrupt any protein-lipid binding. This concentration was chosen to mimic the typical
NMR conditions as well as binding data that indicate that 50mM Pi is sufficient to abolish
binding of NOXO1β to a POPC surface (Chapter II, Figure 7). A similar spectra was
seen after addition of phosphate to the 1:1 NOXO1β PX: diC8-PI(4,5)P2 sample.(Figure
9B). The formation of lipid micelles does not appear to be the cause of the poor spectra
resolution, which is further supported by Campbell et al. (2003) who commented that the
CMC (critical micelle concentration) for a phosphorylated inositol headgroup would
occur in the low millimolar concentrations. In these lipid titration experiments, the lipid
concentration would not exceed 200μM. The protein concentration was taken after the
last PI(4,5)P2 titration point and the protein concentration had decreased by 4-fold from the expected concentration (from ~80μM accounting for sample dilution to <20μM). The
decrease is spectra resolution seen with the diC8 chain length could be due to NOXO1β
111
1 15 Figure 9. NOXO1β PX diC8-PI(4,5)P2 titration. A. H, N HSQCs of NOXO1β PX
with increasing amounts of diC8-PI(4,5)P2 (black:0mol, blue:0.5mol, red:0.75mol,
1 15 green:1mol) B. H, N HSQCs of NOXO1β PX with increasing amounts of diC8-
PI(4,5)P2 plus the addition of 100mM NaPi (black:0mol, blue:0.5mol, green:1mol,
1 cyan:1mol + 100mM NaPi). With increasing amounts of diC8-PI(4,5)P2, H-N peaks in a
1H, 15N-HSQC disappear. The addition of phosphate to the buffer should disrupt all
protein-lipid interaction yet has no effect on the spectra.
112
A 1:0.5 NOXO1β PX:diC -PI(4,5)P 1:0 NOXO1β PX:diC8-PI(4,5)P2 8 2
15 15 N ppm N ppm
1H ppm 1H ppm 1:1 NOXO1β PX:diC -PI(4,5)P 1:0 NOXO1β PX:diC8-PI(4,5)P2 8 2
15 15 N ppm N ppm
1H ppm 1H ppm B 1:0 NOXO1β PX:diC8-PI(4,5)P2 1:0.5 NOXO1β PX:diC8-PI(4,5)P2
15 15 N ppm N ppm
1H ppm 1H ppm 1:1 NOXO1β PX:diC -PI(4,5)P + 100mM NaP 1:1 NOXO1β PX:diC8-PI(4,5)P2 8 2 i
15 15 N ppm N ppm
1H ppm 1H ppm
113
PX aggregating due to the fatty acyl chain or it could be an indication of intermediate exchange occuring.
To test whether the decrease in spectral resolution was due to the lipid chain length, we next titrated NOXO1β PX with diC4-PI(3,4)P2 in a 1:0, 1:0.25, 1:0.5, 1:0.75,
1:1 and 1:2 ratio. Unlike the spectra for diC8-PI(4,5)P2, there were very few changes in
chemical shifts when diC4-PI(3,4)P2 was added, yet the spectra resolution did not decrease. Peaks for Q122, E107 and S86 are weaker as the amount of diC4-PI(3,4)P2
increases, although there is no apparent shift in the peaks (Figure 10). There are two
additional peaks that appear to be weaker as the amount of PI(3,4)P2 increases, but due to
changes in the peak positions in the buffer used for the PIP titration versus that used for
structure assignment and calculations (20mM HEPES, 137mM KCl, 1mM EGTA, 5%
2 2 H5-glycerol, pH 7.2 and 100mM NaPi, 100mM NaCl, 0.1mM EDTA, 5% H5-glycerol,
pH 6.9 respectively), which peaks are moving is not clear (peaks are in a section of the
spectra containing L26, D32, K53, R62 and A106, Figure 10A).
The final concentration of NOXO1β PX after the titration with diC4-PI(3,4)P2 was lower than expected on account of dilution alone (~80μM accounting for sample dilution to 20μM). Although it was slightly higher than the concentration for NOXO1β PX after titrating with diC8-PI(4,5)P2, the concentration difference was not enough to account for
the differences in the spectra. These data indicate that the diC8-PI(4,5)P2 is causing
NOXO1β PX to form aggregates in solution with diC8-PI(4,5)P2. To test if this was a
result of the chain length or the inositol headgroup, diC4-PI(4,5)P2 was next titrated into
NOXO1β PX.
114
Figure 10. NOXO1β PX does not undergo many chemical shift changes upon
1 15 addition of diC4-PI(3,4)P2. A. H, N HSQC of NOXO1β PX with increasing amounts
of diC4-PI(3,4)P2 (black:0mol, blue:0.5mol, red:1mol, green:2mol). B. Chemical shift
map of NOXO1β PX titrated with diC4-PI(3,4)P2. Residues that undergo changes in
chemical shift upon addition of diC4-PI(3,4)P2 are colored in yellow. DiC4-PI(3,4)P2 causes few chemical shift changes which are spread across the structure of NOXO1β PX.
115
A Black: 1:0 Blue: 1:0.5 Red: 1:1 Green 1:2
N ppm N 15
1H ppm
B
116 Titrating diC4-PI(4,5)P2 we saw improved resolution for the spectra with the higher diC4-PI(4,5)P2 amounts compared to that of diC8-PI(4,5)P2 (Figure 11A vs Figure
9A). However, unlike the case of diC4-PI(3,4)P2, there were several changes in chemical shift upon addition of increasing amounts of diC4-PI(4,5)P2 (Figure 11B). Some of the chemical shift changes were for residues near the PIP-binding pocket (T23, D75 and
S86). Several residues, whose sidechains pointed down from α1 (V38, W42, D43,
R46/K49 and K50), also saw changes in chemical shift. The remaining identified chemical shift changes (V10, G61, Q96, L103, R108, R111 and Q122) were scattered away from the PIP-binding pocket.
The changes seen in chemical shift are due to a change in the chemical environment around that residue. Chemical shift changes can be seen due to direct interaction of a residue with part of the PIP headgroup or some other local perturbation in
the structure. In the spectra monitoring the chemical shift changes for the PIP titrations,
we are looking at backbone 1H-N atoms. It might be necessary to do experiments that
would include sidechain perturbations to see if those changes are localized towards the
PIP-binding pocket. Although the data do not reveal any direct evidence of PI(4,5)P2-
specific binding in the PIP binding pocket, there were differences between the titrations
with diC4-PI(4,5)P2 and diC4-PI(3,4)P2. In the diC4-PI(3,4)P2 titration there were very
few peaks that had changes (either chemical shift changes or decrease in peak intensity)
while the diC4-PI(4,5)P2 titration saw many more peaks have a change in chemical shift
with increasing amounts of lipid.
117
Figure 11. NOXO1β PX undergoes chemical shift changes upon addition of diC4-
1 15 PI(4,5)P2. A. H, N HSQC of NOXO1β PX with increasing amounts of diC4-PI(4,5)P2
(black:0mol, blue:0.5mol, red:1mol, green:2mol). B. Chemical shift map of NOXO1β
PX titrated with diC4-PI(4,5)P2. Residues that undergo changes in chemical shift upon
addition of diC4-PI(4,5)P2 are colored in yellow. DiC4-PI(4,5)P2 causes many chemical
shift changes which are spread across the structure of NOXO1β PX. The location of the
chemical shift changes does not indicate specific binding to the short chain lipid in the
PIP binding pocket.
118
A Black: 1:0 Blue: 1:0.5 Red: 1:1 Green 1:2
N ppm N
15
1H ppm
B
119
Structural Analysis of the PIP-Binding Pocket
For PX domains that bind to PIPs phosphorylated at the D3 position, several
interactions are involved. In the case of p40phox PX and Grd19p PX, which both have
phox diC4-PI(3)P-bound structures (Bravo et al. 2001 (p40 PX), Zhou et al. 2003 (Grd19p
PX)), an Arg coordinates with the D3 phosphate (R58 in p40phox PX and R81 in Grd19p
phox PX). P47 PX, which preferentially binds PI(3,4)P2, also contains an Arg analogous to
the R58/R81 position in p40phox PX/Grd19p (Karathanassis et al. 2002).
Although this Arg is highly conserved among PX domains, several PX domains
contain a Tyr, Ser or Thr in this position. These PX domains appear to not bind PIPs
phosphorylated at the D3 position, with the sidechain hydroxyl groups not allowing a
favorable interaction with the phosphate group. This is seen in PI3K-C2α PX domain that contains a Thr (T1462) in this position and selectively binds PI(4,5)P2 (Song et al.
2001, Stahelin et al. 2006) and in the Bem1p PX domain which has a Tyr (T318) and selectively binds PI(4)P (Stahelin et al. 2007).
Like the PX domains of PI3K-C2α and Bem1p, NOXO1β PX also lacks this Arg
and contains a Ser (S41). In agreement with the binding data presented in Chapter II,
NOXO1β PX should not bind to PIPs phosphorylated at the D3 position. A view of this
region of the PX domains of p40phox (PDB ID: 1H6H), PI3K-C2α (PDB ID: 2IWL) and
NOXO1β illustrate the position of the residues involved in binding to a phosphate at the
D3 position (Figure 12). Although these results conflict with the current published binding data for NOXO1β PX (Cheng and Lambeth 2004, Cheng and Lambeth 2005,
120
Figure 12. NOXO1β PX does not contain conserved residues for binding to PIs phosphorylated at the D3 position. A. Structure of NOXO1β PX (red) with S41 highlighted. B. Structure of p40phox PX (PDB:1H6H, green) with R58 highlighted. C.
Structure of PI3K-C2α (PDB:2IWL, blue) with T1462 highlighted. NOXO1β PX lacks a conserved Arg that is present in p40phox PX and confers binding to the D3 phsophate.
PI3K-C2α also lacks this conserved Arg and binds specifically to PI(4,5)P2. The structure of NOXO1β PX suggests that it does not bind to PIPs phosphorylated at the D3 position.
121
A
NOXO1β PX
S41
B
p40phox PX R58
C
PI3K-C2α T1462
122 Ueyama et al. 2007), the structural data are consistent with the binding data for NOXO1β
PX presented in Chapter II.
In addition to the loss of the D3 phosphate coordinating residue, NOXO1β PX also lacks a basic residue to interact with the D1 phosphate. P40phox PX contains a Lys
(K92) that coordinates the D1 phosphate of the inositol headgroup acting as a “positive
selector” for binding PI(3)P (Bravo et al. 2001). The structurally similar region of
NOXO1β PX (present in the long loop connecting α1-α2) does not contain a basic
residue (the sequence in this region is LDAPLL).
For PX domains that bind to PIPs phosphorylated at the D4 position, the structure
of Bem1p PX (Stahelin et al. 2007) provides information for a PI(4)P-binding PX domain
while the PX domain of PI3K-C2α PX structure provides information for a PI(4,5)P2- binding PX domain (Song et al. 2001, Stahelin et al. 2006). In the PI3K-C2α structure, the loop where the D4 and D5 phosphates are proposed to interact was highly disordered so that no structural basis of binding at the D4 and D5 positions could be determined.
However, based on a bound sulfate ion in the PI3K-C2α crystal structure, an Arg
(R1503) is proposed to coordinate the phosphate at the D4 position. R1503 of PI3K-C2α superimposes onto an Arg in p40phox PX (R105) that coordinates the 4- and 5-OH of
PI(3)P (Stahelin et al. 2006). Similarly, the PX domain of Bem1p contains an Arg
(R369) that appears to be a candidate in coordinating the D4 phosphate based on a model
of PI(4)P bound to the Bem1p PX crystal structure (Stahelin et al. 2007). Like R1503 of
PI3K-C2α, R369 of Bem1p PX also is in a position analagous to R105 of p40phox PX. In
NOXO1β PX, R91 is in a position analagous to R369 of Bem1p PX and R1503 of PI3K-
C2α PX (Figure 13) and is a candidate for coordinating the phosphate at the D4 position.
123
Figure 13. NOXO1β PX contains conserved residues for binding to PIs phosphorylated at the D4 position. A. Structure of NOXO1β PX (red) with R91 highlighted. B. Structure of Bem1p PX (PDB:2CZO, teal) with R369 highlighted. C.
Structure of PI3K-C2α (PDB:2IWL, blue) with R1503 highlighted. NOXO1β PX
contains a conserved Arg that is present in Bem1p PX which binds specifically to PI(4)P.
PI3K-C2α also contains this conserved Arg which binds specifically to the D4 position of
PI(4,5)P2 based on the presence of a sulfate ion. The structure of NOXO1β PX suggests
that it can bind to PIs phosphorylated at the D4 position.
124 A
NOXO1β PX R91
B
Bem1p PX R369
C
PI3K-C2α PX
R1503
125 Examining the residue(s) responsible for coordinating the D5 phosphate does not have a
detailed structure-based model. As mentioned previously, a disordered loop in the
PI3K-C2α crystal structure prevented Stahelin and coworkers (2006) from suggesting candidate residues. Several basic residues in the disordered loop in the P13K-C2α PX structure were proposed as being able to coordinate the D5 phosphate. Based on the position of the proposed D4 coordinating arginine (R1503), the authors proposed a series of nearby basic residues, including R1488, R1493 and K1497. Three candidate arginine residues in NOXO1β PX are located near the position of R91 and could be responsible for coordinating the D5 phosphate. R81, R84 and R87 are located proximal to where
R1488, R1493 and K1497 of PI3K-C2α would potentially be in the disordered loop
(Figure 14). There is not a basic residue in NOXO1β PX located close to where position
R1488 would be, which might suggest that it is not a likely candidate in PI3K-C2α to coordinate the D5 phosphate. R81 in NOXO1β PX in located approximately where
R1493 would be in the disordered loop of PI3K-C2α and both R84 and R87 of NOXO1β
PX are close to where position K1497 would be in PI3K-C2α (Figure 14). Interestingly, the dynamics calculated for NOXO1β PX indicated slower motions around residue 74
(Figure 5), which is in the same loop as these candidate Arg in NOXO1β PX. The dynamics present near this loop could indicate flexibility necessary to bind to PIPs phosphorylated at the D5 position.
126
Figure 14. Candidate residues of NOXO1β PX for binding to PIs phosphorylated at
the D5 position. A. Structure of NOXO1β PX (red) with R81, R84 and R87 highlighted.
B. Structure of PI3K-C2α (PDB:2IWL, blue) showing the disordered loop between residues 1487-1498. C. Alignment of NOXO1β PX (red) and PI3K-C2α (blue)
superimposes candidate Arg residues over the missing loop in PI3K-C2α. A series of basic residues (R1488, R1493 and K1497) were identified as candidates to coordinate the
D5 phosphate in PI3K-C2α based on its proximity to R1503 (Figure 13) which coordinates the D4 phosphate. R81 is located near where R1493 would be positioned in
PI3K-C2α. R84 and R87 are located near where K1497 would be located in PI3K-C2α.
There are no basic residues in NOXO1β PX near the position of R1488, suggesting that
R1488 is not a candidate for binding the D5 phosphate in PI3K-C2α.
127
A
R84
R87
R81
B
1498
1487
C
128 References
Berjanskii, M.V., Neal, S. and Wishart, D.S. (2006) PREDITOR: a web server for predicting protein torsion angle restraints. Nucleic Acids Res. 34, W63-W69. Brunger, A.T., Adams, P.D., Clore, G.M., DeLano, W.L., Gros, P., Grosse-Kuntstleve, R.W., Jiang, J.-S., Kuszewski, J., Nilges, M., Pannu, N.S., Read, R.J., Rice, L.M., Simonson, T. and Warren, G.L. (1998) Crystallography and NMR System: a new software suite for macromolecular structure determination. Acta Cryst. D54, 905- 921. Campbell, R.B., Liu, F. and Ross, A.H. (2003) Allosteric activation of PTEN phosphatase by phosphatidylinositol 4,5-bisphosphate. J. Biol. Chem. 278, 33617- 33620. Cavanagh, J., Fairbrother, W.J., Palmer, A.G., III and Skelton, N.J. (1996) Protein NMR Spectroscopy: Principles and Practice. Academic Press, Inc., California. Chen, J., Feng, L. and Prestwich, G.D. (1998) Asymmetric total synthesis of phosphatidylinositol 3-phosphate and 4-phosphate derivatives. J. Org. Chem. 63, 6511-6522. Chou, J.J., Gaemers, S., Howder, B., Louis, J.M. and Bax, A. (2001) A simple apparatus for generating stretched polyacrylamide gels, yielding uniform alignment of proteins and detergent micelles. J. Biomol. NMR 21, 377-382. Cierpicki, T. and Bushweller, J.H. (2004) Charged gels as orienting media for measurement of residual dipolar couplings in soluble and integral membrane proteins. J. Am. Chem. Soc. 126, 16259-16266. Cole, R. and Loria, J.P. (2003) FAST-Modelfree: a program for rapid automated analysis of solution NMR spin-relaxation data. J. Biomol. NMR 26, 203-213. Cornilescu, G., Delaglio, F. and Bax, A. (1999) Protein backbone angle restraints from searching a database for chemical shift and sequence homology. J. Biomol. NMR 13, 289-302. Delaglio, F., Grzesiek, S., Vuister, G. W., Zhu, G., Pfeifer, J., and Bax, A. (1995) NMRPipe: A multidimensional spectral processing system based on UNIX pipes. J. Biomol. NMR 6, 277-293. Ellson, C.D., Gobert-Gosse, S., Anderson, K.E., Davidson, K., Erdjument-Bromage, H., Tempst, P., Thuring, J.W., Cooper, M.A., Lin, Z-Y., Holmes, A.B., Gaffney, P.R.J., Coadwell, J., Chilvers, E.R., Hawkins, P.T. and Stephens, L.R. (2001) PtdIns(3)P regulates the neutrophil oxidase complex by binding to the PX domain of p40phox. Nat. Cell Bio. 3, 679-682. Farmer, B.T., II and Venters, R.A.(1996) Assignment of aliphatic side-chain 1HN/15N resonance in perdeuterated proteins. J. Biomol. NMR 7, 59-71. Fossi, M., Oschkinat, H., Nilges, M. and Ball, L.J. (2005) Quantitative study of the effects of chemical shift tolerances and rates of SA colling on structure calculation from automatically assigned NOE data. J. Magn. Reson. 175, 92-102. Goto, N.K., Gardner, K.H., Mueller, G.A., Willis, R.C. and Kay, L.E. (1999) A robust and cost-effective method for the production of Val, Leu, Ile(δ1) methyl- protonated 15N-, 13C-, 2H-labeled proteins. J. Biomol. NMR 13, 369-374.
129 Johnson, B. A. and Blevins, R.A. (1994) NMR View: A computer program for the visualization and analysis of NMR data. J. Biomol. NMR 4, 603-614. Jones, D.H. and Opella, S.J. (2004) Weak alignment of membrane proteins in stressed polyacrylamide gels. J. Magn. Reson. 171, 258-269. Kanai, F., Liu, H., Field, S.J., Akbary, H., Matsuo, T., Brown, G.E., Cantley, L.C., and Yaffe, M.B. (2001) The PX domains of p47phox and p40phox bind to lipid products of PI(3)K. Nat. Cell Biol. 3, 675–678. Karathanassis, D., Stahelin, R.V., Bravo, J., Perisic, O., Pacold, C.M., Cho, W., Williams, R.L. (2002) Binding of the PX domain of p47phox to phosphatidylinositol 3,4-bisphosphate and phosphatidic acid is masked by an intramolecular interaction. EMBO J. 21, 5057-5068. Kelly, A.E., Ou, H.D., Withers, R. and Dötsch, V. (2002) Low-conductivity buffers for high-sensitivity NMR measurements. J.Am. Chem. Soc. 124, 12013-12019. Kleinschmidt, J.H. and Tamm, L.K. (2002) Structural transitions in short-chain lipid assemblies studied by (31)P-NMR spectroscopy. Biophys. J. 83, 994-1003. Lee, S.A., Kovacs, J., Stahelin, R.V., Cheever, M.L., Overduin, M., Setty, T.G., Burd, C.G., Cho, W. and Kutateladze, T. (2006) Molecular Mechanism of Membrane Docking by the Vam7p PX Domain. J. Biol. Chem. 281, 37091-37101. Ledford, B. (2004) Characterization of phoshpolipid binding to the NADPH oxidase component p47-phox. Link, Y. and Wagner, G. (1999) Efficient side-chain and backbone assignment in large proeins: application to tGCN5. J. Biomol. NMR 15, 227-239. Lu, J., Garcia, J., Dulubova, I. Sudof, T.C. and Rizo, J. (2002) Solution Structure of the Vam7p PX Domain. Biochemistry 41, 5956-5962. Malkova, S., Stahelin, R.V., Pingali, S.V., Cho, W. and Schlossman, M.L. (2006) Orientation and Penetration Depth of Monolayer-Bound p40phox-PX. Biochemistry 45, 13566-13575. McDonald, I.K. and Thornton, J.M. (1994) Satisfying hydrogen bonding potential in proteins. J. Mol. Biol. 238, 777-793. Meier, S., Haussinger, D. and Grzesiek, S. (2002) Charged acrylamide copolymer gels as media for weak alignment. J. Biomol. NMR 24, 351-356. Nath, A., Atkins, W.M. and Sligar, S.G. (2007). .Applications of phospholipids bilayer nanodiscs in the study of membranes and membrane proteins. Biochemistry, 46, 2059-2069. Otting, G. and Wüthrich, K. (1989) Extended heteronuclear editing of 2D 1H NMR spectra of isotope-labeled proteins, using the X(w1, w2) double half filter. J. Magn. Reson. 85, 586-594. Rebecchi, M.J., Eberhardt, R., Delaney, T., Ali, S. and Bittman, R. (1993) Hydrolysis of short acyl chain inositol lipids by phospholipase C-delta 1 R. J. Biol. Chem. 268, 1735-1741. Rieping, W., Habeck, M., Bardiaux, B., Bernard, A., Malliavin, T.E. and Nilges, M. (2007) ARIA2: automated NOE assignment and data integration in NMR structure calculation. Bioinformatics 23, 381-382. Rosen, M. J. (1984) Structure-Performance Relationships in Surfactants, American Cancer Society Symposium Series 23, ACS, Washington, D. C.
130 Shen, Y., Delaglio, F., Cornilescu, G., and Bax, A. (2009) TALOS+: A hybrid method for predicting protein backbone torsion angles from NMR chemical shifts. J. Biomol. NMR 44, 213-223. Stahelin, R.V., Burian, A., Bruzik, K.S., Murray, D., Cho, W. (2003) Membrane binding mechanism of the PX domains of NADPH oxidase p40phox and p47phox. J. Biol. Chem. 278, 14469-14479. Stahelin, R.V., Karathanassis, D., Bruzik, K.S., Waterfield, M.D., Bravo, J., Williams, R.L. and Cho, W. (2006) Structural and membrane binding analysis of the phox homology domain of phosphoinoside 3-kinase-C2a. J. Biol. Chem. 281, 39396- 39406. Stahelin, R.V., Karathanassis, D., Murray, D., Williams, R.L. and Cho, W. (2007) Structural and membrane binding analysis of the phox homology domain of Bem1p. J. Biol. Chem. 282, 25737-25747. Song, X., Xu, W., Zhang, A., Huang, G., Liang, X., Virbasius, J.V., Czech, M.P. and Zhou, G.W. (2001) Phox homology domains specifically bind phosphatidylinositol phosphates. Biochemistry 40, 8940-8944. Tugarinov, V. and Kay, L.E. (2003) Ile, Leu and Val methyl assignments of the 723- residue malate synthase G using a new labeling strategy and novel NMR methods. J. Am. Chem. Soc. 125, 13868-13878. Tycko, R., Blanco, F.J. and Ishii, Y. (2000) Alignment of biopolymers in strained gels: a new way to create detectable dipole-dipole couplings in high-resolution biomolecular NMR. J. Am. Chem. Soc. 122, 9340-9341. Ueyama, T., Lekstrom, K., Tsujibe, S., Saito, N and Leto, T.L. (2007) Subcellular localization and function of alternatively spliced Noxo1 isoforms. Free Radic. Biol. Med. 42, 180-190. Voehler, M. V., Collier, G., Young, J. K., Stone, M. P., and Germann, M. W. (2006) Performance of cryogenic probes as a function of ionic strength and sample tube geometry. J. Magn. Reson. 183, 102-109. Vold, R.R., Prosser, R.S., and Deese, A.J. (1997). Isotropic solutions of phospholipids bicelles: A new membrane mimetic for high-resolution NMR studies of polypeptides. J. Biomol. NMR 9, 329-335. Xia, Y., Yee, A., Arrowsmith, G.H. and Gao, X. (2003) 1HC and 1HN total NOE correlations in a single 3D NMR experiment. 15N and 13C time-sharing in t1 and t2 dimensions for simultaneous data acquisition. J. Biomol. NMR 27, 193-203. Yang, D., Venters, R.A., Mueller, G.A., Choy, W.Y. and Kay, L.E. (1999) TROSY-based HNCO pulse sequences for the measurement of 1HN-15N, 15N-13CO, 1HN- 13CO, 13CO-13Cα and 1HN-13Cα dipolar couplings in 15N, 13C, 2H-labeled proteins. J. Biomol. NMR. 14, 333-343. Zhan, Y., Virbasius, J.V., Song, X., Pomerleau, D.P., Zhou, G.W. (2002) The p40phox and p47phox PX domains of NADPH oxidase target cell membranes via direct and indirect recruitment by phosphoinositides. J. Biol. Chem. 277, 4512-4518.
131 Chapter IV
DISCUSSION
The NADPH Oxidases are a family of enzyme complexes that catalyze the
generation of ROS. These ROS are used in a variety of biological functions ranging from
host defense, cell growth and differentiation to inflammation. The best characterized
NADPH oxidase is the phagocytic oxidase or PHOX complex. The PHOX complex has
both membrane bound and cytosolic components that are sequestered from each other
when the enzyme is inactive. Fully functional and active complex is dependent upon
translocation of the cytosolic components to the membrane-bound components. Two of
the cytosolic components, p47phox and p40phox, both contain an N-terminal PX domain,
which binds to membrane phospholipids. The PX domains of p47phox and p40phox aid in
targeting the cytosolic components to the membrane.
Homologues to several of the PHOX components have been identified recently.
Homologues to gp91phox (NOX2) include NOX1-5 (Suh et al. 1999, Lambeth et al. 2000,
Geiszt et al. 2000, Banfi et al. 2001) as well as DUOX1-2 (Lambeth et al. 2000, Edens et
al. 2001). Homologues to two of the cytosolic components have also been identified.
The p47phox homologue is NOXO1 and the p67phox homologue is NOXA1 (Banfi et al.
2003, Geiszt et al. 2003, Takeya et al. 2003, Cheng and Lambeth 2004). NOXO1, like
p47phox and p40phox, contains a N-terminal PX domain.
Our lab has recently been investigating both p47phox and NOXO1 to probe the
structure-function relationship of the PX domains. In this dissertation, we aimed to examine the structure-function relationship of the PX domain of NOXO1β. To do this
131 we characterized the lipid binding of NOXO1β PX (Chapter II) and determined the solution structure of this domain (Chapter III).
To characterize the lipid binding of NOXO1β PX, we utilized several assays to
varying degrees of success. The first step was to do a dot blot for NOXO1β PX. The
nature of the dot blot only allows for the qualitative observation that NOXO1β PX binds
to negatively charged lipids (Chapter II, Figure 1A). The relative intensity of the bands
may or may not indicate a preference for one lipid over another. Narayan and Lemmon
(2006) noted that in the case of the PIP Strip™, some lipids wash off of the membranes
to a greater extent than other lipids. The authors noted that PI(3,4,5)P3 was the most
soluble followed by the bisphosphates and then the monophosphates. This observation
follows the pattern of intensities that were seen in the PIP Strips™ for NOXO1β PX; that
is, that the strongest signal was seen to the PIPs that were the least soluble (and least
likely to wash off the membrane) and the weakest signal to the PIP that was the most
soluble (and more likely to wash off the membrane over time). Similar to NOXO1β PX,
p47phox PX also displayed a broader range of lipids that it binds to, although the strongest
phox signal was not seen to PI(3,4)P2 (Chapter II, Figure 1D). P40 PX only showed an
interaction with PI(3)P (Chapter II, Figure 1C), which agrees with published data
(Ellson et al. 2001, Kanai et al. 2001) that has shown that p40phox PX is specific to PI(3)P
only. From the PHOX/NOX PX domains, it appears dot blots are only reliable if there is
a very specific and very strong binding of the PX domain to a particular PIP, such as the
case with p40phox PX.
SPR experiments on NOXO1β PX showed a high level of binding to what is
considered “background lipids” for PX domains (Chapter II, Figure 9A). The binding
132 of NOXO1β PX was so high to a POPC:POPE surface that reference subtracting to calculate binding parameters (Kd, kon, koff) was not appropiate based on Biacore, Inc.’s criteria as well as the inherent error in calculating binding parameters where often there was <10% difference between the reference and sample lipid binding. Both the dot blots and the screen of (PO)PE/(PO)PC conditions indicate that the binding is not specific to either the PE or PC headgroup but to a general membrane surface. This binding to a neutral membrane agrees with published data that showed NOXO1β PX colocalized at the plasma membrane without the presence of external stimuli (Cheng and Lambeth
2004, Ueyama et al. 2007).
Contrary to published reports (Cheng and Lambeth 2004, Cheng and Lambeth
2005, Takeya et al. 2006, Ueyama et al. 2007), we have shown that NOXO1β PX
preferentially binds to PI(4,5)P2 using SPR (Chapter II, Figure 10). The binding data show that NOXO1β PX does not bind to the PIPs much above background, which was
phox phox seen in both p40 PX (for PI(3)P) and p47 PX (for PI(3,4)P2) (Chapter II, Figure
11). NOXO1β PX appears to be a PX domain that binds strongly to neutral membranes and has a weaker interaction with PIPs, which has not been reported for any other PX domain. We also tested whether NOXO1β PX would bind to anionic lipids, such as PA and PS, which is the case for the p47phox PX (Karathanassis et al. 2002, Ledford 2004).
At low levels of either PA or PS, no binding above background is seen for NOXO1β PX
(Chapter II, Figure 12). We hypothesize that at a higher amount of PA or PS, an
increase in binding over background would be seen with NOXO1β PX. This is based on
the basic nature of the residue (theoretical pI of 10.98) and the sequence similarity it has
to the residues of p47phox that interacts with PA (Karathanassis et al. 2002).
133 Two vesicle binding assays were attempted to further characterize the lipid
binding of NOXO1β PX. However, both assays had problems that could not be
addressed by the time of this dissertation. Having another assay would help to either
corroborate the SPR binding data or perhaps provide additional data that suggests
different lipid binding. One thing to note is that these binding assays were done on the
isolated PX domain. Although it is hypothesized that only the PX domain of NOXO1β contributes to membrane binding (Cheng and Lambeth 2004), full length NOXO1β might exhibit different lipid binding than the PX domain alone.
The solution structure of NOXO1β PX exhibits the overall fold seen in other PX domains with known structures. The structure contains three anti-parallel β-strands
surrounded by four α-helices and a 310 helix (Chapter III, Figure 4). The solution
structure of NOXO1β PX definitively shows the position of the amino acid differences in the isoform variants. The structure shows that the K50 deletion, seen in NOXO1α and
NOXO1δ, is located within helix α1. We hypothesize that the K50 deletion could cause a
shift in the helical register and cause some degree of disorder or unfolding in the PX
domain. This hypothesis is supported by the observation by Ueyama et al. (2007) where
NOXO1α and NOXO1δ were localized in the cytosol. This observation, coupled with the consistent localization to the plasma membrane of NOXO1β and the dependence of that localization on the PX domain (Cheng and Lambeth 2004), suggests that NOXO1α
and NOXO1δ are not fully functional.
The five amino acid insert in NOXO1γ and NOXOδ is in a large loop region near
the proposed lipid binding surface. It remains unknown if this insert has any structural or
functional effects, but due to its proximity to the PIP binding pocket, this insert has the
134 potential to alter lipid binding of NOXO1γ compared to that of NOXO1β. Dot blots for
NOXO1γ PX were identical to those for NOXO1β PX (Cheng and Lambeth 2005, data
not shown). Additionally, Ueyama et al. (2007) noted that NOXO1γ was localized
primarily in the nucleus. The five amino acid insert (GQASL) in NOXO1γ PX is not a
known nuclear localizing sequence, but it could play some role in targeting NOXO1γ to
the nucleus. Whether this insert merely aids in targeting NOXO1γ to the nucleus or alters
the lipid binding of NOXO1γ still remains to be determined.
The dynamics of NOXO1β PX (Chapter III, Figure 5) suggest some flexibility
in the loop connecting α1-α2 that is involved in the formation of one side of the PIP-
binding pocket. This flexibility could be necessary in NOXO1β PX being able to bind
more to PI(4,5)P2 and to a lesser extent to PI(3,4,5)P3 over the other PIPs, which contain
multiple bulky phosphate groups. This flexibility could also be inherent in PX domains that have a broad lipid binding specificity, allowing multiple inositol headgroups to access the conserved basic residues that line the pocket.
Additional protein-lipid experiments were performed using solution-state NMR.
First we examined the residue-specific interaction of NOXO1β PX with a POPC nanodisc (Chapter III, Figure 8). We saw a large number of chemical shift changes that were spread across the PX domain. This suggests a non-specific interaction between the
POPC surface and NOXO1β PX. We also did several short-chain PIP titrations. We noticed an induction of changes in chemical shifts upon addition of diC4-PI(4,5)P2 but not with diC4-PI(3,4)P2. This suggests that NOXO1β PX is interacting with the diC4-
PI(4,5)P2 and not the diC4-PI(3,4)P2, although the nature of the interaction does not appear to be specific to the PIP binding pocket.
135 The NOX complexes and how they are regulated are not yet well understood.
Many of the interactions found within the PHOX system are thought to also be involved in the NOX1 and NOX3 systems, since they appear to be the most similar to the PHOX system. The role of NOXO1β is thought to be very similar to that of its homologue, p47phox. While there are some similarities between the two proteins, there are also several striking differences.
The PX domain of both p47phox and NOXO1β are more promiscuous in their lipid binding specificity, both being able to bind to varying degrees to most of the PIPs.
However, NOXO1β PX binds strongly to a neutral membrane, which has not been reported for any other PX domain. In contrast to the high level of binding to a neutral membrane, NOXO1β PX does not bind much over background to the PIPs. NOXO1β also has poor sequence similarity to the AIR of p47phox, which contains multiple phosphorylation sites and is involved in the autoinhibition of p47phox in resting neutrophils. Between the general, constitutive membrane binding and lack of the AIR region, NOXO1β is regulated within the NOX complexes differently than p47phox is in the PHOX system.
NOXO1β has been shown to bind to p22phox via its tandem SH3 domains (Dutta and Rittinger 2010) and to a lesser extent via portion of the C-terminal tail (Dutta and
Rittinger 2010, Yamamoto et al. 2007). Like p47phox (Groemping et al. 2003), NOXO1 forms a “superSH3 domain” and binds with high affinity to p22phox (Dutta and Rittinger
2010). However, Cheng and Lambeth (2004) showed that a NOXO1β construct lacking the PX domain does not localize to the plasma membrane. Therefore, the NOXO1- p22phox interaction does not appear to be sufficient for NOXO1’s constitutive localization
136 at the plasma membrane, which further illustrates the importance of the PX domain for
NOXO1β.
The presence of high amounts of PI(4,5)P2 in the plasma membrane could
provide an additional anchorage point for NOXO1β PX’s membrane binding. PI(4,5)P2 is present in high concentrations in mammalian cells. PI(4,5)P2 has been reported to
make up >99% of the bisphosphorylated PIs in mammalian cells (Vanhaesebroeck et al.
2001) and to make up ~1% of the total phospholipid in human erythyrocytes (Ferrell and
Huestis 1984). NOXO1β PX’s high binding capacity to neutral membranes coupled to binding to the highly present PI(4,5)P2 could “lock” NOXO1β PX to the plasma
membrane. This would put NOXO1β PX proximal to p22phox and the NOX enzyme in the
membrane. This would allow other signals, including reversible interactions with other
components in the oxidase (p22phox and NOXA1, Dutta and Rittinger 2010), to regulate
the NOX complex.
The preference of NOXO1β PX to PI(4,5)P2 provides a potential signal to known
NOX functions. As discussed in Chapter I, NOX1 has been shown to be involved in
multiple tumorgenic processes (Chamulitrat et al. 2003, Shinohara et al. 2010, Komatsu et al. 2008). In the study by Shinohara and colleagues (2010), NOX1-generated ROS contributed to cell invasion via regulation of matrix metalloprotease-9 production and cell migration. If ROS derived from NOX1 is one of the players in the signaling events leading to cell invasion, the preference of NOXO1β PX for PI(4,5)P2, which is involved in cytoskeleton rearrangements (Raucher et al. 2000, McLaughlin et al. 2002), could be a signal to target the NOX1 complex. However, it is not clear if NOXO1β is present in this
137 NOX activity, so a direct link between NOXO1β PI(4,5)P2 binding and NOX1-derived
ROS contributing to cell invasion can not be made.
Additionally, there are published reports for the PX domain of both p47phox and
p40phox to bind to moesin in a phosphoinositide-dependent manner (Wientjes et al. 2001,
Zhan et al. 2004) as well as reports of an actin binding site in p47phox (Tamura et al.
2006). It remains unknown if NOXO1β can bind to any cytoskeleton proteins, but it provides another pathway where NOXO1β PX’s binding to PI(4,5)P2 is targeting it to
additional complexes associated with PI(4,5)P2.
The research presented in this dissertation provides a more thorough
characterization of the lipid binding of NOXO1β PX. We have shown that NOXO1β PX
binds to neutral membranes at high levels and that it is a modest increase in binding to
PIP-containing membranes. Of the PIPs tested, NOXO1β PX bound the most to
PI(4,5)P2, and an analysis of the PIP-binding pocket in the NOXO1β PX structure
suggests that NOXO1β PX can bind to PIs phosphorylated at the D4 and D5 positions.
These data suggest that NOXO1β PX specifically targets PI(4,5)P2 in addition to binding
neutral membranes. This dual binding allows NOXO1β to be constitutively at the
membrane yet still be able to target itself, along with the NOX complexes, to signaling
events involving PI(4,5)P2.
138 References
Banfi, B., Molnar, G., Maturana, A., Steger, K., Hegedus, B., Demaurex, N., Krause, K.H. (2001) A Ca2+-activated NADPH oxidase in testis, spleen and lymph nodes. J. Biol. Chem. 276, 37594-37601. Banfi, B., Clark, R.A., Steger, K., Krause, K.H. (2003) Two novel proteins activate superoxide generation by the NADPH oxidase NOX1. J. Biol. Chem. 278, 3510- 3513. Chamulitrat, W., Schmidt, R., Tomakidi, P., Stremmel, W., Chunglok, W., Kawahara, T. and Rokutan, K. (2003) Association of gp91phox homolog Nox1 with anchorage- independent growth and MAP kinase-activation of transformed human keratinocytes. Oncogene 22, 6045-6053. Cheng, G. and Lambeth, J.D. (2004) NOXO1, regulation of lipid binding, localization and activation of Nox1 by the Phox homology (PX) domain. J. Biol. Chem. 279, 4737-4742. Cheng, G. and Lambeth, J.D. (2005) Alternative mRNA splice forms of NOXO1: Differential tissue expression and regulation of Nox1 and Nox3. Gene 356, 118- 126. Dutta, S. and Rittinger, K. (2010) Regulation of NOXO1 Activity through Reversible interactions with p22phox and NOXA1. PLoS ONE 5, e10478. Edens, W.A., Sharling, L., Cheng, G., Shapira, R., Kinkade, J.M., Lee, T., Edens, H.A., Tang, X., Sullards, C., Flaherty, D.B., Benian, G.M., Lambeth, J.D. (2001) Tyrosine crosslinking of extracellular matrix is catalyzed by Duox, a multidomain oxidase/peroxidase with homology to the phagocytic oxidase subunit gp91phox. J. Cell Biol. 154, 879-891. Ellson, C.D., Gobert-Gosse, S., Anderson, K.E., Davidson, K., Erdjument-Bromage, H., Tempst, P., Thuring, J.W., Cooper, M.A., Lin, Z-Y., Holmes, A.B., Gaffney, P.R.J., Coadwell, J., Chilvers, E.R., Hawkins, P.T. and Stephens, L.R. (2001) PtdIns(3)P regulates the neutrophil oxidase complex by binding to the PX domain of p40phox. Nat. Cell Bio. 3, 679-682. Ferrell, J.E. Jr and Huestis, W.H. (1984) Phosphoinositide metabolism and the morphology of human erythrocytes. J. Cell Biol. 98, 1992-1998. Geiszt, M., Kopp, J.B., Varnai, P., Leto, T.L. (2000) Identification of renox, an NAD(P)H oxidase in kidney. Proc. Natl. Acad. Sci. 97, 8010-8014. Geiszt, M., Lekstrom, K., Witta, J., Leto, T.L. (2003) Proteins homologous to p47phox and p67phox support superoxide production by NAD(P)H oxidase 1 in colon epithelial cells. J. Biol. Chem. 278, 20006-20012. Groemping, Y., Lapouge, K., Smerdon, S.J. and Rittinger, K. (2003). Molecular basis of phosphorylation-induced activation of the NADPH oxidase. Cell 113, 343-355. Kamata, T. (2009) Roles of Nox1 and other Nox isoforms in cancer development. Cancer Sci. 100, 1382-1388. Kanai, F., Liu, H., Field, S.J., Akbary, H., Matsuo, T., Brown, G.E., Cantley, L.C. and Yaffe, M.B. (2001) The PX domains of p47phox and p40phox bind to lipid products of PI(3)K. Nat. Cell Biol. 3, 675–678. Karathanassis, D., Stahelin, R.V., Bravo, J., Perisic, O., Pacold, C.M., Cho, W., Williams, R.L. (2002) Binding of the PX domain of p47phox to
139 phosphatidylinositol 3,4-bisphosphate and phosphatidic acid is masked by an intramolecular interaction. EMBO 21, 5057-5068. Komatsu, D., Kato, M., Nakayama, J., Miyagawa, S. and Kamata, T. (2008) NADPH oxidase 1 plays a critical mediating role in oncogenic Ras-induced vascular endothelial growth factor expression. Oncogene 27, 4724-4732. Lambeth, J.D., Cheng, G., Arnold, R.S., Edens, W.E. (2000) Novel homologs of gp91phox. Trends Biochem. Sci. 25, 459-461. Ledford, B. (2004) Characterization of phoshpolipid binding to the NADPH oxidase component p47-phox. McLaughlin, S., Wang, J., Gambhir, A. and Murray, D. (2002) PIP2 and Proteins: Interactions, organization and information flow. Annu. Rev. Biophys. Biomol. Struct. 31, 151-175. Narayan, K. and Lemmon, M.A. (2006) Determining selectivity of phosphoinositide- binding domains. Methods. 39, 122-133. Raucher, D., Stauffer, T., Chen, W., Shen, K., Guo, S., York, J.D., Sheetz, M.P. and Meyer, T. (2000) Phosphatidyinositol 4,5-bisphosphate functions as a second messenger that regulates cytoskeleton-plasma membrane adhesion. Cell. 100, 221-228. Shinohara, M., Adachi, Y., Mitsushita, J., Kuwabara, M., Nagasawa, A., Harada, S., Furuta, S., Zhang, Y., Seheli, K., Miyazaki, H. and Kamata, T. (2010) Reactive oxygen generated by NADPH oxidase 1 (Nox1) contributes to cell invasion by regulating matrix metalloprotease-9 production and cell migration. J. Biol. Chem. 285, 4481-4488. Suh, Y.-A., Arnold, R.S., Lassegue, B., Shi, J., Xu, X., Sorescu, D., Chung, A.B., Griendling, K.K., Lambeth, J.D. (1999) Cell transformation by the superoxide- generating oxidase Nox1. Nature 401, 79-82. Takeya, R., Ueno, N., Kami, K., Taura, M., Kohjima, M., Izaki, T., Nunoi, H., Sumimoto, H. (2003) Novel human homologues of p47phox and p67phox participate in activation of superoxide-producing NADPH oxidase. J. Biol. Chem. 278, 25234-25246. Takeya, R., Taura, M., Yamasaki, T., Naito, S. and Summimoto, H. (2006) Expression and function of Noxo1gamma, an alternative splicing form of the NADPH oxidase organizer 1. Febs J. 273, 3663-3677. Tamura, M., Itoh, K., Akita, H., Takano, K. and Oku, S. (2006) Identification of an actin- binding site in p47phox and organizer protein of NADPH oxidase. FEBS Letters 580, 261-267. Ueyama, T., Lekstrom, K., Tsujibe, S., Saito, N and Leto, T.L. (2007) Subcellular localization and function of alternatively spliced Noxo1 isoforms. Free Radic. Biol. Med. 42, 180-190. Van Meer, G., Voelker, D.R. and Feigenson, G.W. (2008) Membrane lipids: where they are and how they behave. Nat. Rev. Mol Cell Bio. 9, 112-124. Vanhaesebroeck, B., Leevers, S.J., Ahmadi, K., Timms, J. Katso, R., Driscoll, P.C., Woscholski, R., Parker, P.J. and Waterfield, M.D. (2001) Synthesis and function of 3-phosphorylated inositol lipids. Annu. Rev. Biochem. 70, 535-602.
140 Wientjes, F.B., Reeves, E.P., Soskic, V., Furthmayr, H. and Segal, A.W. (2001) The NADPH oxidase components p47phox and p40phox bind to moesin through their PX domain. Biochem. and Biophys. Res. Comm. 289, 382-388. Yamamoto, A., Kami, K. Takeya, R. and Sumimoto, H. (2007) Interaction between the SH3 domains and C-terminal proline-rich region in NADPH oxidase organizer 1 (Noxo1). Biochem. Biophys. Res. Commun. 352, 560-565. Zhan, Y., He, D., Newburger, P.E. and Zhou, G.W. (2004) p47phox PX domain of NADPH oxidase targets cell membrane via moesin-mediated association wih the actin cytoskeleton. J. Cell. Biochem. 92, 795-809.
141 APPENDIX
INVESTIGATION OF THE PROPOSED NOXO1β-RIP1 INTERACTION
Introduction The TNFR1 signaling complex is a member of the death receptor pathways.
Upon binding of TNF to TNFR1, the receptor oligomerizes resulting in oligomerization
of the intracellular death domains (DD). TRADD (TNFR1 associated death domain
protein) is recruited to the complex where it interacts via the death domains. After
receptor oligomerization, TRADD can further recruit additional proteins, such as RIP1
and TRAF2 (TNFR1 associated factor 2), to the complex. A downstream signaling event
in the activation of this death receptor pathway is the production of ROS.
It has been noted that ROS are involved in TNF-induced necrotic cell death.
Early suggestions pointed to the mitochondrial complex I as the source of this TNF- induced ROS seen in necrotic cell death (Fiers et al. 1999). However, in 2007, Kim et al.
showed that superoxide anion is generated under conditions of necrotic cell death and
that this superoxide anion was specific to TNF and not to other inflammatory cytokines.
Antioxidants were able to block both TNF-induced necrotic cell death and superoxide
production, suggesting a link between these two processes (Kim et al. 2007).
Of the seven known oxidases, only NOX2 had been previously shown to be
inducible by TNF (Frey et al. 2002). However, in the cell lines used by Kim et al.
(2007), no NOX2 was detected. This led the authors to investigate if other NOX proteins
were a potential source of the superoxide anion that was produced. NOX1 was detected
in the cell lines used and siRNA (small interfering RNA) confirmed that NOX1 was the
source of the TNF-induced superoxide produced (Kim et al. 2007). Further experiments
142 revealed that NOX1 formed a complex with TRADD, RIP1 and Rac1 following TNF treatment and that this complex was dependent on NOX1.
The next question asked was how NOX1 was recruited to the TRADD complex following TNF stimulation. Since no direct interaction was detected between NOX1 and either TRADD or RIP1, Kim et al. (2007) hypothesized that another component in the
NOX complex was interacting with the TRADD complex. A strong TNF-dependent interaction was detected between RIP1 and NOXO1β as well as RIP1 and NOXA1. A strong TNF-dependent interaction between TRADD and NOXO1β was also detected in the cell lines. The TRADD-NOXO1β interaction was identified between a known SH3- binding motif in the central domain of TRADD to the N-terminal SH3 domain of
NOXO1β. The group was not able to identify the region of interaction between RIP1 and
NOXO1β.
RIP1 is an intriguing binding partner to NOXO1β because of its key involvement in multiple cell death pathways (Festjens et al. 2007). RIP1 is part of the seven member
RIP family of serine/threonine kinases and appears to play a central role in cell death and survival. It can be activated by multiple pathways, including death receptor pathways, toll-like receptors (Cusson-Hermance et al. 2005) and T-cell receptors (Cusson et al.
2002). RIP1 also has different downstream effects including activation of NFkappaB
(Kelliher et al. 1998), activation of MAPK pathways (Lee et al. 2003) as well as activation or inhibition of both apoptosis and necrosis (Holler et el. 2000, Kreuz et al.
2004, Lin et al. 2004, Meylan et al. 2005).
RIP1 contains an N-terminal kinase domain that autophosphorylates when active
(Hsu et al. 1996 and Ting et al. 1996) followed by an intermediate domain (ID) that
143 contains a RIP homotypic interaction motif (RHIM) that can interact with RHIMs in
other RIP family members (Meylan et al. 2005). The role of this intermediate domain is
not well understood. RIP1 also contains a C-terminal DD, which is involved in protein- protein interactions with other DDs.
We aimed to indentify the region of RIP1 that interacts with NOXO1β. We made
fusion proteins of various RIP1 domain constructs and attempted to detect binding to full
length NOXO1β by using a pulldown assay with recombinant proteins.
144 Experimental Procedures
RIP1 and NOXO1β Protein Expression-RIP1 constructs were made by cloning
different fragments of human RIP1 into pET expression vectors (pET47, pET48 and
pET50). The RIP1 constructs used were the intermediate domain and death domain
(residues 294-671) in pET47, pET48 and pET50; the intermediate domain (residues 294-
585) in pET47; the death domain (residues 551-671) in pET47 (Figure 1). Plasmids
containing the RIP1 constructs were transfected into E. coli Bl21(DE3)Codon+RIPL and
grown in LB at 37 °C until OD600=0.6. Cultures were induced with 0.5mM IPTG for 4 hours at 37 °C or 8 hours at 15 °C. Cells were harvested by centrifugation for thirty minutes at 4,000 rpm at 4 °C and resuspended in 20mM Tris, pH 8.0, 20mM NaCl, 2mM
DTT. Cells were lysed by sonicating on ice for two times at four minutes each cycle
(level 5, 50% duty). Lysates were cleared by centrifugation for thirty minutes at 4,000 rpm at 4 °C.
Full length NOXO1β (1-371) was cloned into pGEX-6P-1. E. coli strain
BL21(DE3)Codon+RIPL was transformed with the plasmid and grown in LB at 32 °C until OD600=0.6 and induced overnight at 22 °C. Bacterial cells were harvested by centrifugation (30 minutes, 4,000 rpm at 4 °C) and resuspended in 20mM Tris, ph 8.0,
20mM NaCl, 2mM DTT. Cells were lysed for four cycles at four minutes each cycle
(level 5, 50% duty) and lysates were clarified by centrifugation at 4,000 rpm for 30 minutes at 4 °C.
RIP1-NOXO1β Binding Assay- Initial tests were performed to see if any of the fusion
proteins bound non-specifically to either a Ni-NTA resin or a GSH-Sepharose-FF resin.
145 None of the His6-RIP1 constructs bound non-specifically to the GSH-Sepharose-FF resin,
while GST-NOXO1β did have some background binding to the Ni-NTA resin. For all binding experiments, the NOXO1β-RIP1 binding was tested using immobilized GST-
NOXO1β on GSH-Sepharose-FF resin. Lysates containing GST-NOXO1β were incubated with GSH-Sepharose-FF resin and washed extensively to remove unbound proteins. Bacterial lysate containing the specific RIP1 construct was incubated with the
GST-NOXO1β bound GSH-Sepharose resin for one hour at room temperature. GST-
NOXO1β was eluted with the addition of 15mM GSH to the buffer and the fractions were run on a SDS-PAGE to see if both NOXO1β and RIP1 were present in eluted fractions.
146
Figure 1. Domain Organization of RIP1 and Constructs of RIP1 Used For Binding Assays.
A. Domain organization of RIP1. RIP1 contains a N-terminal kinase domain (KD) followed by the intermediate domain which contains the RIP1 homotypic interaction motif (RHIM) and a C- terminal death domain (DD). B. To test the NOXO1β-RIP1 interaction, three constructs of RIP1 were used. The first construct contained both the ID and DD (294-671). The second construct contained the ID only (294-585) and the third construct contained the DD only (585-671). The
KD was not used in any of the constructs.
147
A.
RIP1 KD ID DD RHIM
B.
RIP1 ID-DD (294-671) ID DD RHIM
RIP1 ID (295-585) ID RHIM
RIP1 DD (551-671) DD
148 Results and Discussion
The first step in examining the potential NOXO1β-RIP1 interaction included expressing recombinant RIP1 proteins. The three RIP1 constructs (Figure 1) were cloned into three pET vectors (pET47: His6, pET48: His6-thioredoxin(Trx), pET50: His6-
NusA) and induced at two different temperatures to test for solubility. The ID-DD RIP1 construct (294-671) was sucessfully cloned into all three vectors. Both the ID (294-585) and DD (585-671) alone were only sucessfully cloned into pET47.
Of the three different fusion tags attached to the ID-DD RIP1 construct (residues
294-671), the pET47 fusion was not soluble at either induction temperature, so only the pET48 and pET50 fusions were used in the binding assay. No binding was seen with the
His6-NusA-ID-DD RIP1 (pET50). The His6-Trx RIP1 ID-DD fusion (pET48) ran in the same position as GST-NOXO1β on the gel so that comassie stained gels were not able to distinguish the presence of one or both proteins. No binding was detected for the pET47
(His6) fusions of either the ID alone (residues 295-585) or the DD alone (551-671).
Although no RIP1-NOXO1β interaction was detected, several possibilities for this arise. In the paper by Kim et al. (2007), the pulldown assay where both RIP1 and
NOXO1β were detected used lysates from cells stimulated with TNF. Cells that were not stimulated with TNF prior to the pulldown experiment did not detect NOXO1β in the pulldown assays. In addition to the TNF-dependent interaction seen by Kim et al. (2007), another reason no interaction between NOXO1β and RIP1 was detected could be due to missing intermediate factors. The published assays used mammalian cells expressing the entire TNF signaling complex in addition to the NOX1 complex. Intermediate factors necessary for RIP1 and NOXO1β to interact could be missing. It is also possible that the
149 pulldown assay was not detecting a direct interaction between RIP1 and NOXO1β and
that, in fact, a larger portion of the TNF signaling complex and/or NADPH oxidase complex is needed to see the RIP1-NOXO1β connection.
Expressing recombinant protein domains bacterially might be lacking a critical post-translational modification so that either NOXO1β or RIP1 (or another component in the complexes critical for the RIP1-NOXO1β interaction) is not in its biologically active
form. The constructs used in the assay also might have not contained the region of RIP1 necessary for the interaction. The N-terminal kinase domain of RIP1 was not used in the pulldown assays. Our initial hypothesis was that the interaction would occur between either the death domain or intermediate domain since these are both known protein- protein interaction domains. If there is a direct interaction between NOXO1β and RIP1 the possibility of the kinase domain being involved cannot be ruled out.
The experimental setup to confirm the RIP1-NOXO1β interaction was not ideal
for identifying the NOXO1β-RIP1 interaction. A system using proteins and/or whole cell
lysates from a mammalian cell line would be better suited for probing this interaction.
Future experiments will need to systematically attempt to address several questions
individually, which the current experiments failed to do. After confirming the pulldown
assay results that Kim et al. (2007) obtained, it needs to be determined what, if any,
protein(s) among the TNFR1 and NOX1 complexes is(are) necessary for the pulldown
results. Finally, it needs to be determined if the pulldown results are in fact due to a
direct NOXO1β-RIP1 interaction before attempting to identify the specific location of the
proposed NOXO1β-RIP1 interaction.
150 References
Cusson, N., Oikemus, S., Kilpatrick, E.D., Cunningham, L. and Kelliher, M. The death domain kinase RIP protects thymocytes from tumor necrosis factor receptor type 2-induced cell death. J. Exp. Med. 196, 15-26. Cusson-Hermance, N. Khurana, S., Lee, T.H., Fitzgerald, K.A. and Kelliher, M.A. (2005) Rip1 mediates the Trif-dependent toll-like receptor 3- and 4-induced NF- (kappa)B activation but does not contribute to interferon regulatory factor 3 activation. J. Biol. Chem. 280, 36560-36566. Festjens, N., Berghe, V.T., Cornelis, S. and Vandenabeele, P. (2007) RIP1, A kinase on the crossroads of a cell’s decsion to live or die. Cell Death and Diff. 14, 400-410. Fiers, W., Beyaert, R., Declercq, W. and Vandenabeele, P. (1999) More than one way to die: apoptosis, necrosis and reactive oxygen damage. Oncogene. 18, 7719-7730. Frey, R.S., Rahman, A., Kefer, J.C., Minshall, R.D. and Malik, A.B. (2002) PKCzeta regulates TNF-alpha-induced activation of NADPH oxidase in endothelial cells. Circ. Res. 90, 1012-1019. Holler, N., Zaru, R., Micheau, O., Thome, M., Attinger, A., Valititti, S., Bodmer, J.L., Schneider, P., Seed, B. and Tschopp, J. (2000) Fas triggers an alternative, caspase-8-independent cell death pathway using the kinase RIP as effector molecule. Nat. Immunol. 1, 489-495. Hsu, H., Huang, J., Shu, H.B., Baichwal, V. and Goeddel, D.V. (1996) TNF-dependent recruitment of the protein kinase RIP to the TNF receptor-1 signaling complex. Immunity. 4, 387-396. Kelliher, M.A., Grimm, S., Ishida, Y., Kuo, F., Stanger, B.Z. and Leder, P. (1998) The death domain kinase RIP mediates the TNF-induced NF-kappaB signal. Immunity. 8, 297-303. Kim, Y-S., Morgan, M.J., Choksi, S. and Liu, Z-G. (2007) TNF-induced activation of the Nox1 NADPH oxidase and its role in the induction of necrotic cell death. Mol. Cell. 26, 675-687. Kreuz, S., Siegmund, D., Rumpf, J.J., Samel, D. Leverkus, M., Janssen, O., Häcker, G., Dittrich-Breiholz, O., Kracht, M., Scheurich, P. and Wajant, H. (2004) NFkappaB acivation by Fas is mediated through FADD, caspase-8, and RIP and is inhibited by FLIP. J. Cell Biol. 166, 369-380. Lee, T.H., Huang, Q., Oikemus, S., Shank, J., Ventura, J.J., Cusson, N., Vaillancourt, R.R., Su, B., Davis, R.J. and Kelliher, M.A. (2003) The death domain kinase RIP1 is essential for tumor necrosis factor alpha signaling to p38 mitogen- activated protein kinase. Mol Cell Biol. 23, 8377-8385. Lin, Y., Choksi, S., Shen, H-M., Yang, Q-F., Hur, G.M., Kim, Y.S., Tran, J.H., Nedospasov, S.A. and Liu, Z-G. (2004) Tumor necrosis factor-induced nonapoptotic cell death requires receptor-interacting protein-mediated cellular reactive oxygen species accumulation. J. Biol. Chem. 279, 10822-10828. Meylan, E. and Tschopp, J. (2005) The RIP kinases: crucial integrators of cellular stress. Trends Biochem. Sci. 30, 151-159. Ting, A.T., Pimentel-Muinos, F.X. and Seed, B. (1996) RIP mediates tumor necrosis factor receptor 1 activation of NF-kappaB but not Fas/APO-1-initiated apoptosis. EMBO J. 15, 6189-6196.
151 Curriculum Vitae Nicole Y. Davis
Contact Information: Department of Biochemistry Wake Forest University School of Medicine Medical Center Boulevard Winston-Salem, NC 27157 (336)-716-5525 (w) [email protected]
Education: 2003-present Ph.D. Candidate, Department of Biochemistry Wake Forest University School of Medicine, Winston-Salem, NC. Thesis Advisor: David A. Horita, Ph.D. 1998-2002 B.S., magna cum laude, Biochemistry (minors: biology and mathematics), Florida State University, Tallahassee, FL. Undergraduate Advisor: Timothy M. Logan, Ph.D.
Related Experience 2001-2002 Undergraduate Research, Florida State University, Department of Chemistry and Biochemistry, Institute for Molecular Biophysics, Tallahassee, FL.
Undergraduate Research: Analyzed kinetic and thermodynamic parameters of protein folding and unfolding of mutants of the FK506 binding protein (FKBP). The research included cloning, expressing and purifying recombinant mutant proteins as well as stopped-flow kinetic measurements of folding and unfolding of the mutant proteins.
Honors and Awards: 2005 Suraj Manrao Award for Student Travel, 46th ENC Conference 2005 Wake Forest University Graduate School of Arts and Sciences Alumni Student Travel Award 2009 Wake Forest University Graduate School of Arts and Sciences Alumni Student Travel Award
Laboratory Skills: • Recombinant protein expression and purification • Chromatography o Size exclusion o Ion Exchange o Affinity (GST-, His-, MBP-tagged proteins) • Protein Electrophoresis
152 o SDS-PAGE • Spectrometry o Mass (MALDI-TOF) o UV-Vis • Multidimensional solution-state NMR • PCR methods o Cloning PCR-generated DNA fragments o site directed mutagenesis using PCR o purification of PCR fragments for cloning • DNA Methods o Restriction digests o DNA ligation and transformation o Agarose gel electrophoresis o Plasmid isolation • Light Scattering • Lyophilization • Western Blot • Surface Plasmon Resonance: o kinetic and affinity experiments o affinity capture and hydrophobic adsorption chip surfaces • Lipid work o Extruded vesicle preparation o Isotropic bicelle preparation • ultracentrifugation • assay development
Related Graduate Courses: Metabolism and Bioenergetics Proteins and Enzymes Molecular Biology I Intracellular Signaling Structural Biology Cell Biology Bioinformatics Enzyme Kinetics Physics of Macromolecules Computational Molecular Biology Lab Computational Analysis in Molecular Biology Principles of Teaching
Teaching Experience: Courses Taught Biochemistry I, Wake Forest University School of Medicine (BCM 705), Fall 2008. Lecture-Glycolysis and the Pentose Phosphate Pathway
153 Lecture-Online Resources Tutorial Biochemistry I, Wake Forest University School of Medicine (BCM 705), Fall 2009 Lecture-Glycolysis and the Pentose Phosphate Pathway Tutoring Biochemistry I, Wake Forest University School of Medicine (BCM 705), Fall 2008 Biochemistry I, Wake Forest University School of Medicine (BCM 705), Fall 2009
Presentations and Seminars 2008 Biochemistry Student Seminar, Wake Forest University “Investigation of Noxo1 Protein-protein and Protein-lipid Interactions” 2007 Biochemistry Student Seminar, Wake Forest University “Structure and Lipid Interactions of the PX Domain” 2006 Biochemistry Student Seminar, Wake Forest University “Structural Studies of Membrane Binding by the PX Domains of p47phox and Noxo1” 2004 Biochemistry Student Seminar, Wake Forest University “Structural Studies of Membrane Binding by the PX Domain” 2004 Biochemistry Student Seminar, Wake Forest University “Exonuclease I: A Processive Exonuclease” 2003 Biochemistry Student Seminar, Wake Forest University “NMR Studies of the 52 kDa TREX2 Homodimer”
Posters and Abstracts: 2005 46th ENC “Mapping the Lipid-Binding Interface of the Phox Homology (PX) Domain by Solution-State NMR.” N.Y. Davis, B.G. Ledford, L.C. McPhail, D.A. Horita 2009 Keystone Symposia: Frontiers of NMR in Biology “Noxo1β PX Domain Binds to Both Neutral and PIP-Containing Membranes.” N.Y. Davis, L.C. McPhail, D.A. Horita
Publications: E.M. Lewis, A.S. Wilkinson, N.Y. Davis, D.A. Horita, J.C. Wilkinson. Non-Degradative Ubiquitination of Apoptosis Inducing Factor by XIAP at a Residue Critical for AIF- Mediated Chromatin Condensation. (submitted)
N.Y. Davis, L.C. McPhail, D.A. Horita. Backbone Resonance Assignments of NOXO1β PX Domain. (in preparation)
References: David A. Horita, Ph.D. Associate Professor of Biochemistry
154 Wake Forest University School of Medicine Medical Center Boulevard Winston-Salem, NC 27157 Tel: 336-713-4194 E-mail: [email protected]
Linda C. McPhail, Ph.D. Associate Dean Graduate School of Arts and Sciences Wake Forest University Tel: 336-716-0221
Professor of Biochemistry Wake Forest University School of Medicine Medical Center Boulevard Winston-Salem, NC 27157 Tel: 336-716-2621 E-mail: [email protected]
Roy R. Hantgan, Ph.D. Associate Professor of Biochemistry Wake Forest University School of Medicine Medical Center Boulevard Winston-Salem, NC 27157 Tel: 336-716-4675 E-mail: [email protected]
155