STRUCTURAL INSIGHTS INTO RECOGNITION OF ADENOVIRUS BY
IMMUNOLOGIC AND SERUM FACTORS
by
JUSTIN WAYNE FLATT
Submitted in partial fulfillment of the requirements for the degree of Doctor of
Philosophy
Dissertation Adviser: Dr. Phoebe L. Stewart
Department of Pharmacology
CASE WESTERN RESERVE UNIVERSITY
May, 2014
CASE WESTERN RESERVE UNIVERSITY
SCHOOL OF GRADUATE STUDIES
We hereby approve the thesis/dissertation of
4 JUSTIN WAYNE FLATT G 4 candidate for the A DOCTOR OF PHILOSOPHY a degree*.
(signed) a JASON MEARS A (chair of committee)
A PHOEBE STEWART a
A ANDREAS ENGEL A
A SICHUN YANG D
A AMY RUSCHAK A
(date) A DECEMBER 13, 2013 A
*We also certify that written approval has been obtained for any proprietary
material contained therein.
ii
To Victoria and Samson.
iii
TABLE OF CONTENTS
List of Tables…………………………………………………………………………….xii
List of Figures…………………………………………………………………………...xiv
List of Abbreviations…………………………………………………………………….xx
Acknowledgements……………………………………………………………………..xxv
Abstract……………………………………………………………………………………1
Chapter 1 Introduction and background………….…………………………………..3
1.1 Adenoviruses, from pathogens to therapeutics…………………………....3
1.1.1 Adenoviruses as pathogens………………………………………………..3
1.1.2 Adenovirus structure and cell entry……………………………………….5
1.1.3 Adenoviruses as therapeutic agents……………………………………….9
1.2 Adenovirus-host interactions…………………………………………….10
1.2.1 Host immune response to adenovirus……………………………………...11
1.2.2 Recognition of adenovirus by human alpha defensin 5……………….....12
1.2.3 Recognition of adenovirus by serum factors…………………………….15
iv
1.2.4 Recognition of adenovirus capsid-incorporated HIV antigen……….…..18
1.3 A hybrid structural approach to analyzing adenovirus recognition……...20
1.3.1 Preparation of samples for cryo-EM……………………………………..20
1.3.2. Virus imaging and three-dimensional reconstruction…………………....22
1.3.3 Modeling of cryo-EM maps……………………………………………...25
Chapter 2 Recognition of adenovirus by human alpha defensin 5 involves intrinsic
disorder……………………………………………………………….….29
2.1 Abstract……………………………………………………………..……29
2.2 Intro……….…………………………………………………………..….30
2.3 Results…………………………………………………………………....33
Cryo-EM structures of HD5 complexed with neutralization-sensitive and –
resistant HAdVs……………………………………………………….....33
Modeling of HD5 monomers at the HAdV vertex…………………….....36
The vertex of the defensin-sensitive HAdV accommodates HD5
dimers…………………………………………………………………….39
Intrinsic disorder at the HD5 binding site……………………………..…40
v
Stabilization of defensin-sensitive HAdV vertex region by HD5…...... 42
2.4 Discussion………………………………………………………………..43
2.5 Material and methods…………………………………………………....48
Cryo-EM and image processing…………………………….…………...48
Atomic model building…………………………………………….…....49
Molecular dynamics flexible fitting………………………………….....50
2.6 Acknowledgements……………………………………………….…….51
2.7 Author contributions…………………………………………….……....51
Chapter 3 Virus-misplaced humoral factor activates innate immunity….…………67
3.1 Abstract………………………………………………………….….…...67
3.2 Report………………………………………………………………..…..68
Cryo-EM and MDFF analysis of the protein-protein interface between FX
and HAdV5…………………………………………………………..…..69
Validation of the FX-HAdV5 interaction model………………..………..70
FX decorated virus activates an innate immune response………..….…...72
Conclusion…………………………………………………….…..……...74
vi
3.3 Material and methods……………………………………………………75
Animal studies………………………………………………………...... 75
Viruses………………………………………………………………...... 75
Electron microscopy, image processing and modeling………………….76
Proteome profiler antibody arrays…………………………………….....77
Antibody for confocal microscopy…………………………………...... 77
Surface plasmon resonance analyses…………………………………….77
Microarray sample processing…………………………………………...78
Microarray data analysis……………………………………………...... 78
Ingenuity pathway analysis…………………………………………...... 79
Promoter analysis of differentially expressed genes……………………..80
Gene ontology category analysis……………………....……………...... 80
Statistical analysis………………………………………………………..80
3.4 Acknowledgements………………………………………………………81
Chapter 4 FVII dimerization on adenovirus capsid may influence infectivity…….110
vii
4.1 Abstract…………………………………………………………………110
4.2 Introduction……………………………………………………………..111
4.3 Materials and methods………………………………………………….114
Cells and viruses………………………………………………………..114
Cryo-electron microscopy, image processing, and modeling…………..115
AdV infection in vitro…………………………………………………..117
AdV attachment assay……………………………………………...... 118
SPR analyses……………………………………………………………118
Statistical analysis………………………………………………………120
Protein structure assessment number………………………………...... 120
4.4 Results…………………………………………………………………..120
Cryo-EM structural analysis adenovirus-FVII interaction……………...120
Mutation of the HVR7 TET amino acid motif reduces the affinity of FVII
binding to the virus……………………………………………………...121
FVII is inefficient at supporting virus attachment and cell transduction,
despite efficient binding to hexon……………………………….……...123
viii
FVII binds HAdV5 hexon in an altered orientation compared to that of
FX……………………………………………………………………....124
FVII dimerizes via SP domain interactions when bound to HAdV5
hexon…………………………………………………………….……...125
FVII domain dimerization obscure putative receptor interacting residues
within the dimer interface……………………………………….……...127
4.5 Discussion…………………………………………………………...... 128
4.6 Acknowledgements…………………………………………………...... 132
Chapter 5 Visualization of adenovirus capsid-incorporated HIV antigen….……...147
5.1 Abstract………………………………………………………………....147
5.2 Intro……….……………………………………….………………...... 148
5.3 Material and methods…………………………………….…………...... 152
Cryo-EM and image processing………………………….…………...... 152
Cryo-EM guided molecular dynamic simulations to model the MPER
insertions……………………………………………………….…….....154
5.4 Results……………………………………………………….………….155
ix
Cryo-EM structure of AdV vector with capsid incorporated MPER
peptide……………………………………………………………..….155
Alpha helices are observed within the AdV capsid and for one MPER
insertion……………………………………………………………...... 157
MPER conformation is constrained by the AdV capsid at one insertion
site……………………………………………………………………...157
Strong helical interactions between MPER insertions at the 3-mer
sites……………………………………………………………………..158
Transient interactions between MPER insertions at the 2-mer sites…...159
Conformational flexibility of MPER next to penton base……………...160
5.5 Discussion……………………………………………………………....160
5.6 Acknowledgements……………………………………………….….....163
5.7 Author contributions………………………………………….………....164
Chapter 6 Summary of discoveries and future directions……………….………....179
6.1 Insights on immune recognition of adenoviruses……………….……....179
6.1.1 Human alpha defensin opposes capsid disassembly…………….……....179
6.1.2 A role for FX in adenovirus innate immunity………………….….…….182
x
6.1.3 Capsid-incorporated HIV antigen adopts multiple conformations…..….185
6.1.4 Conclusion……………………………………………………….……...187
References…………………………………………………………………………...... 188
xi
LIST OF TABLES
Table Page
2.1 Intermolecular nonbonded energies for HD5 monomers with adenovirus vertex proteins…………………………………………………………………………………...52
2.2 Intermolecular nonbonded energies for HD5 dimers with adenovirus vertex proteins…………………………………………………………………………………...53
2.3 Intermolecular nonbonded energies for three HD5 dimers with each of the subunits of fiber and penton base at one defensin-sensitive adenovirus vertex……………………...54
S3.1 Distances at the FX-hexon interface before and after molecular dynamics flexible fitting runs with different starting FX orientations………………………………………90
S3.2 Thirty-four genes co-activated in the spleens of WT and Illr1-/- mice after challenge with HAdV5……………………………………………………………………………...92
S3.3 Differential expression of 34 gene set in the spleens of WT and Illr1-/- mice after challenge with HAdV5 and TEA mutant viruses……………………………………...... 94
4.1 Binding of FVII to adenovirus vectors and summary of SPR fitting parameters for 1:1 kinetic models…………………………………………………………………………..133
4.2 Intermolecular nonbonded energies between FVII molecules and hexons at the icosahedral 2-fold axis of HAdV5 at the end of three 100-ps MDFF simulations……..134
xii
S5.1 Optimization of the helical interface at a 3-mer site with molecular dynamics flexible fitting…………………………………………………………………………………....173
S5.2 Distances between hexon insertion sites at 2-mer regions……………………...... 174
xiii
LIST OF FIGURES
Figure Page
1.1 Schematic representation of the structural organization of AdV based on cryo-EM and
X-ray crystallography……………………………………………………………...... ….28
2.1 Cryo-EM structures of HD5 bound to neutralization-sensitive (Ad5.F35) and –resistant
(Ad5.PB/GYAR) chimeric HAdVs………………………………………………...…...55
2.2 Modeling and cryo-EM guided molecular dynamics simulations of the interaction between HD5 monomers and vertex proteins of the defensin-sensitive (Ad5.F35) and defensin-resistant (Ad5.PB/GYAR) HAdV chimeras…………………………………...56
2.3 Movement of the RGD-containing loop and HD5 during the molecular dynamics simulations of the defensin-sensitive (Ad5.F35) and defensin-resistant (Ad5.PG/GYAR)
HAdV chimeras………………………………………………………………………….57
2.4 Modeling and cryo-EM guided molecular dynamics simulations of the interaction between HD5 dimers and complete vertex regions of the defensin-sensitive (Ad5.F35) and defensin-resistant (Ad5.PB/GYAR) HAdV chimeras…………………………..…….…59
2.5 Structural malleability of the binding pocket within the defensin-sensitive HAdV chimera (Ad5.F35)………………………………………………………………………61
S2.1 Subnanometer resolution of cryo-EM structures of HD5 bound to neutralization- sensitive (Ad5.F35) and –resistant (Ad5.PB/GYAR) chimeric HAdVs……..………....62
xiv
S2.2 Prediction of intrinsically disordered regions with the HAdV5 and HAdV19c penton base proteins by the PrDOS webserver…………………………………………………..63
S2.3 Space filling representation of the vertex region of Ad5.F35 with three bound HD5 dimers…………………………………………………………………………………….65
3.1 Cryo-EM structure of the FX-HAdV5 complex and simulation of the FX-hexon interface using MDFF……………………………………………………………………82
3.2 A single amino acid substitution (T425A) abrogates FX binding to HAdV5..……...84
3.3 FX binding-ablated virus triggers blunted transcriptional response of NFKB1- dependent genes in vivo………………………………………………………………….86
3.4 HAdV5 binding to FX induces NFKB1-dependent inflammatory cytokines and chemokines downstream of TLR4-TRIF/MyD88-TRAF6 signaling axis in vivo…...….88
S3.1 Robustness of FX-hexon interface after molecular dynamics flexible fitting runs with different starting FX orientations………………………………………………...……...95
S3.2 Comparison of FX bound to hexon in the subnanometer resolution cryo-EM structure to stimulated FX/hexon density…………………………………………………...……..96
S3.3 Infectivity of and response to wild type and hexon-mutated viruses in cells cultured in vitro and mouse hepatocytes in vivo……………………………………………………..98
xv
S3.4 Co-localization of virus particles (red) with splenic marginal zone macrophages
(green) observed 1 h after virus injection for indicated viruses analyzed by confocal microscopy…………………………………………………………………………...…100
S3.5 Confocal microscopy analysis of virus particle localization with CD169+ and
MARCO+ marginal zone macrophages in the spleen of mice 30 minutes after challenge with indicated viruses……………………………………………………………...……101
S3.6 Confocal microscopy analysis of virus particle localization with F4/80+ macrophages in the liver of mice 30 minutes after challenge with indicated viruses……………..…..103
S3.7 Z-score map of transcription factor binding site distribution in the -1000br promoter regions of 34 genes co-activated in the spleens of WT and Illr1-/- mice upon challenge with HAdV…………………………………………………………...……….…...... …105
S3.8 Ingenuity Pathway analysis of networks of transcriptional targets for NFKB1,
CREB1, and SRF transcription factors that respond to HAdV5 infection in the spleen of
WT mice 30 minutes after intravenous virus injection……………………….………...107
S3.9 Ingenuity Pathway analysis of networks of transcriptional targets for NFKB1,
CREB1, and SRF transcription factors that respond to HAdV5 in the spleen of IL-1RI- deficient mice 30 minutes after intravenous virus injection………………….………...108
S3.10 The mRNA expression for IL-1β in spleens of WT and indicated gene-deficient mice 30 minutes after virus injection analyzed by the RNAse protection assay……….109
xvi
4.1 Cryo-EM structure of the HAdV5 complex and simulation of the FVII-hexon interface by molecular dynamics flexible fitting.………………………………………..…….....135
4.2 Integrity, infectivity, and kinetic response data and dissociation constants for FVII binding to wild-type AdV and AdV vectors with mutated hexons………………..…....137
4.3 Transduction of CHO-K1 cells with AdV-WT vector and attachment of AdV-WT and
Ad5S vectors to CHO-K1 cells in the presence of FVII and X……………………..….139
4.4 Cryo-EM and MDFF simulations indicate that FVII and FX adopt distinct binding orientations relative to hexon…………………………………………………………...141
4.5 Modeling of two molecules of FVII at the icosahedral 2-fold axis………………...142
4.6 Localization of positively charged amino acids on the surface of serine protease domains of coagulation factors FX and FVII…………………………………………..143
4.7 Proximal FVII molecules interact and bury potential heparan sulfate proteoglycan binding residues that are exposed on FX…………………………………………….....145
5.1 Cryo-EM structure of the Ad-HVR2-GP41-L15 vector at subnanometer resolution……………………………………………………………………………..…165
5.2 Cryo-EM density showing α-helices for two hexons and two MPER insertions………………………………………………………………………………..166
5.3 MPER insertion within a narrow cavity between hexons at the icosahedral 2-fold axis……………………………………………………………………………………...168
xvii
5.4 MPER forms a stable helical bundle at 3-mer sites………………………………...169
5.5 MPER interactions at 2-mer sites are weak and transient……………………...... 170
5.6 Alternate model conformations for MPER next to penton base…….……………...171
5.7 Proposed vector modifications for optimizing MPER presentation at the AdV hexon
HVR2 site……………………………………………………………………………….172
S5.1 Resolution assessment of the Ad-HVR2-GP41-L15 cryo-EM structure…...…...... 175
S5.2 Secondary structure prediction for the inserted MPER and linker sequence……..176
S5.3 Comparison of density at the icosahedral 3-fold axis with simulated hexon/MPER density……………………..…………………………………………………………....177
xviii
LIST OF ABBREVIATIONS
A Alanine
A549 Adenocarcinomic human alveolar basal epithelial cells
AdV Adenovirus
ARD Acute respiratory disease
C Cysteine
CAR Coxsackie-adenovirus-receptor
CHO Chinese hamster ovary cells
Cryo-EM Cryo-electron microscopy cAMP Cyclic adenosine monophosphate
CCD Charge-coupled device
CREB1 Cyclic adenosine monophosphate response element–binding protein 1
D Aspartic acid
DAMPS Danger-associated molecular patterns
E Glutamic acid
xx
F Phenylalanine
FEG Field emission gun
FMDV Foot and mouth disease virus
FSC Fourier Shell Correlation
FVII Coagulation factor VII
FX Coagulation factor X
G Glycine
GFP Green fluorescent protein
GLA γ-carboxyglutamic acid
H Histidine
HAdV Human adenovirus
HD5 Human alpha defensin 5
HIV Human immunodeficiency virus
HPCC High performance computing cluster
HPV Human papillomavirus
xxi
HSPG Heparan sulfate proteoglycan
HVR Hexon hypervariable region
I Isoleucine
IFN Interferon
IFV Influenza virus
IL Interleukin
IL-1R1 Interleukin-1 receptor 1
K Lysine
Kb Kilobase
L Leucine
M Methionine
MDFF Molecular dynamics flexible fitting
MIP-1α Macrophage inflammatory protein-1α
MOI Multiplicity of infection
MPER Membrane proximal ectodomain region
xxii
MyD88 Myeloid differentiation primary response gene (88)
N Asparagine
NFKB Nuclear factor–κB
NPC Nuclear pore complex
P Proline
PAMPS Pathogen-associated molecular patterns
PMNs Polymorphonuclear leukocytes
PV Poliovirus
Q Glutamine
R Arginine
RNAse Ribonuclease
RGD Arginine-Glycine-Asparagine
S Serine
SNX17 Sorting nexin 17
SP Serine protease
xxiii
SPR Surface plasmon resonance
SRF Serum response factor
T Threonine
TLR Toll-like receptor
TRAF TNF receptor associated factors
TRIF TIR-domain-containing adapter-inducing interferon-β
V Valine
VLP Virus like particle
VMD Visual molecular dynamics
W Tryptophan
WT Wild-type
Y Tyrosine
xxiv
ACKNOWLEDGEMENTS
A dissertation is not a work of individual effort. Though my name alone is listed as an author, there are so many others who have contributed to this work either directly or indirectly and supported me in ways that made this possible. Here, I attempt to acknowledge those who have helped me cross the finish line. Unfortunately, it is impossible to list everyone or to thank those listed enough.
I would like to begin by thanking the people who influenced me to enter the race.
Jimmy Mills started my science addiction. Jimmy is truly one of the best teachers I have ever met. His excitement and passion for science was contagious and because of this I decided that I had to try it myself. Jimmy, thank you for teaching me that science is a privilege and responsibility. Pam Twigg gave me the opportunity to work in her lab as an undergraduate. I still think fondly of my time spent in her lab. Thanks Pam for such a positive undergraduate experience. Also thanks to the graduate students I worked with in
Pam’s lab: Randall, Amy, and Ronny. You guys made the lab a fun place to work.
I would be extremely remiss to forget to mention Hassane Mchaourab. Hassane played a special role in this story because he was my connection to my PhD advisor
Phoebe Stewart. I first met Hassane when he visited my undergraduate university to give a lecture on the application of EPR spectroscopy to protein dynamics. We continued to keep in touch by email and he encouraged me to apply to graduate school at Vanderbilt
University. During my senior year as an undergraduate, I visited Hassane’s lab and it was at this time that I was introduced to my future supervisor Phoebe. Thanks Hassane, for
xxv helping out with the transition to graduate student and for introducing me to a Phoebe
Stewart.
I can still remember the first scientific conversation that I had with Phoebe
Stewart. It was wonderful. She talked about cryo-electron microscopy and showed me a stunning image of adenovirus. I enjoyed our conversation so much that I jumped at the first opportunity to work with her, which was the summer before I started graduate school at Vanderbilt. Phoebe has been an excellent PhD advisor. She has put a tremendous amount of time and effort into developing me as an independent scientist. Phoebe I can’t thank you enough for being a great advisor and friend.
This work would not have been possible without the help and support of our collaborators: Dmitry Shayakhmetov, David Curiel, Glen Nemerow, and Jason Smith.
Thank you all for sharing your materials and skills with us. Also, my research was supported by a training grant. I was appointed to a T32 institutional training grant in the pharmacological sciences, entitled the Molecular Therapeutics Training Program.
I sincerely appreciate the intellectual support and guidance from the members of my dissertation committee: Jason Mears, Andreas Engel, Sichun Yang, and Phoebe
Stewart. Thank you all for being patient, supportive, available and inquisitive throughout this journey. Additionally, I thank Amy Ruschak for her willingness to serve on my committee during the final phase of graduate school. Besides my committee I have also benefited from conversations with Derek Taylor and Vera-Moiseenkova-Bell. Vera and
Derek have always had an open door for my impromptu visits.
xxvi
The members of the Stewart lab, both past and present, have contributed immensely to my personal and professional time during graduate school. I would like to thank, in no particular order, Steffen, Jian, Susan, Tara, Seth, Dewight, Neetu, Mariena, and Rob. Thank you all for your help and friendship.
Heather Holdaway provided support for the electron microscopes and aided with sample preparation during my time at the Cleveland Center for Membrane and Structural
Biology. Kris Palczewski and John Mieyal went above and beyond in helping me transition as a graduate student from Vanderbilt to Case Western. There was never a shortage of assistance from the administrative staff which included, Cami Thompson,
Diane Dowd, Ivona Golczak, Vida Tripodo, and Jenny Yang. Thank you all for your help and support.
Of course, I owe all the thanks in the world to my family! I thank my parents, Ron and Julie, and my two sisters, Jordan and Jantzen. Thank you for providing me with an overwhelming amount of love and support throughout graduate school and life in general.
Looking back, I cannot count the many times that I have heard “we believe in you”s. I guess you were right. I finally did it! I thank my grandparents, Dave and Carolyn King, who have enthusiastically followed me every step of the way during graduate school. I thank my wife, who is also my best friend. After staring at this computer screen for much longer than is advised, I have concluded that there is no way of truly expressing my love and gratitude to Victoria. I am without words. Thank you for all that you do. Finally, I thank God for putting me in the race and giving me the strength to run it to completion.
xxvii
Upon entering graduate school I quickly realized that it would take nothing short of a miracle for me to obtain a PhD.
xxviii
Structural Insights into Recognition of Adenovirus by Immunologic and Serum
Factors
ABSTRACT
by
JUSTIN WAYNE FLATT
Adenoviruses (AdVs) are common pathogens that are a major cause of acute infections of the respiratory and intestinal tracts, as well as the eye. Despite having a distinguished and extensive experimental history, there remain many unanswered questions about how AdVs are recognized and eliminated during infection. In order to advance therapy for infectious and inherited diseases, these challenging questions must be addressed. Here we have examined recognition of AdV by immunologic and serum factors using high-resolution cryo-electron microscopy (cryo-EM) and computational modeling. These factors include human alpha defensin 5 (HD5), human blood coagulation factor X (FX), and factor VII (FVII). We also analyzed the structure of an
AdV-based vaccine that is designed to provide protective immunity against human immunodeficiency virus (HIV). Structural analysis and modeling studies on HD5 recognition of AdV implicated a key role for intrinsic disorder in mediating a stabilizing interaction that blocks viral infection. Cryo-EM and functional examination of serum
factor binding to AdV showed that FX, a noninflammatory humoral factor of the coagulation cascade, binds to the surface of AdV and becomes a pathogen-associated molecular pattern that, upon viral entry into liver cells, triggers activation of innate immunity via the TLR/NF-κB pathway. In contrast, FVII does not support AdV entry into liver cells because it binds in an altered orientation compared to that of FX and dimerizes, which buries potential liver receptor binding residues within the dimer interface. Characterization of the AdV-based HIV vaccine demonstrated how the adenoviral capsid influences epitope structure, flexibility, and accessibility, all of which affect the host immune response.
2
CHAPTER 1: Introduction and background
1.1 Adenoviruses, from pathogens to therapeutic agents
Adenoviruses (AdVs) were first identified as pathogens over 50 years ago using explants of adenoids and tonsils grown in cell culture (Rowe et al, 1953). Since this discovery, research efforts have accelerated knowledge on AdV interactions with host cells and even paved the way for developing strategies to adapt AdVs for therapeutic interventions in humans. This first section of the introduction will cover AdVs both as pathogens and as therapeutic agents. It is designed to be a general introduction for a non- virologist.
1.1.1 Adenoviruses as pathogens
It all began in the winter of 1953, when Wallace Rowe, a post-doctoral fellow at the National Institutes of Health in Robert Huebner’s laboratory, discovered a filterable cytopathogenic agent that causes spontaneous degeneration in cultures of human adenoids (Rowe et al, 1953). The cytopathic changes, which caused rounding and grape like clustering of the affected cells, were shown to be caused by a new virus. At the time of this discovery, there was a major outbreak of acute respiratory disease (ARD) in military recruits at the Fort Leonard Wood military base in the Missouri Ozarks. This outbreak, which was thought to be caused by influenza virus (IFV), was under investigation by Maurice Hilleman and colleague Jacqueline Werner from the Army
Medical Service Graduate School in Washington D.C. Hilleman and Werner failed to isolate IFV, but they succeeded in recovering a new agent from throat swabs of patients
3 suffering from ARD (Hilleman & Werner, 1954). Shortly after, it was established that the agent isolated by Hilleman and Werner was related to the virus recovered by Rowe and co-workers (Huebner et al, 1954) and identical to the virus responsible for large outbreaks of ARD in recruits of the armed forces during World War II (Ginsberg et al,
1955). Several of the early groundbreaking studies on these new respiratory agents were carried out by Harold Ginsberg at Western Reserve University (now Case Western
Reserve University) in Cleveland, OH (Babiss, 2003). In 1956 this group of viruses, affecting primarily the respiratory tract, was named adenoviruses (Enders et al, 1956).
Human AdVs, whose members include >55 types, are classified into 7 species (A-
G) on the basis of common biologic, morphologic, and genetic features. These viruses cause a variety of illnesses including acute respiratory, intestinal, and ocular infections.
Transmission typically occurs from person-to-person via respiratory droplets, but the virus can also spread by the conjunctival and fecal-oral routes. Epidemiological data indicates that the majority of AdV infections occur in the first 5 years of life (Hong et al,
2001), with a peak incidence in the first 2 years (Pacini et al, 1987). Most AdV infections are mild and self-limiting. However, in immunosuppressed individuals, disseminated disease occurs frequently and is associated with a high fatality rate (Pham et al, 2003).
Also, AdVs have emerged as important pathogens within the transplant population, where for example, infection occurs in up to 40% of pediatric stem cell transplant patients and in 5-10% of solid organ transplant recipients (de Mezerville et al, 2006;
Humar et al, 2005; Kampmann et al, 2005; Walls et al, 2005). Although human AdVs have been investigated for several decades, there are currently no anti-viral drugs to treat infection.
4
1.1.2 Adenovirus structure and cell entry
AdVs have a non-enveloped, icosahedrally shaped capsid that protects a double- stranded DNA genome. The viral capsid is comprised of three major proteins (hexon, penton base, and fiber) and four minor proteins (IIIa, VI, VIII, and IX) that surround ~36 kilobase (kb) pairs of DNA. The AdV genome encodes more than 40 different proteins; however, only 13 of these proteins have been identified as constituents of the virus particle (Russell, 2009). The virion has a molecular weight of ~150 MDa and is among the largest non-enveloped viruses. Early structural studies focused on individual capsid proteins. The AdV capsid protein hexon was the first animal virus protein to be crystallized (Franklin et al, 1971). This hexon X-ray diffraction study set the stage for more complex structural analyses of the entire AdV virion by X-ray crystallography and cryo-EM. Two decades after hexon was crystallized, a structure of the intact virus was produced by cryo-EM at 35 Å, which provided detailed information on the hexon packing arrangement within the capsid as well as a first structure of the vertex proteins, including the penton base and its associated fiber (Stewart et al, 1991). Combining the X-ray crystal structure of hexon with the image reconstruction of the intact AdV particle yielded a three-dimensional difference map containing information on the minor proteins that stabilize the capsid (Stewart et al, 1993). Cryo-EM has continued to provide insights into
AdV capsid architecture, including inter-capsomer interactions and assignments for the locations of the AdV minor proteins (Fabry et al, 2005; Liu et al, 2010; Reddy et al,
2010; Saban et al, 2006). The vertex region of the AdV capsid has been examined using both X-ray crystallography and cryo-EM. A partial crystal structure of the AdV homo- trimeric fiber protein revealed a novel triple beta-spiral fibrous fold for the shaft as well
5 as atomic information on the receptor-binding head domain (van Raaij et al, 1999). X-ray crystallographic studies of the AdV penton base protein showed that each subunit contains a basal jellyroll domain similar to hexon and also yielded a structural description of how a segment of fiber binds to form the penton complex (Zubieta et al, 2005). A cryo-EM structure of AdV in complex with a monoclonal antibody revealed the presence of an RGD sequence in each of the five loops located at the top of the penton base protein that serve as a binding site for αv integrins, which mediate virus internalization (Stewart et al, 1997). Steady improvements in cryo-EM and genetically modified virus samples
(for example AdV type 5 modified to contain short AdV type 35 fibers referred to as
Ad35F) have significantly enhanced resolution and in turn, understanding of AdV capsid assembly (Saban et al, 2005). High resolution structural comparisons of an immature adenovirus particle (AdV type 2 temperature-sensitive mutant ts1) with mature AdV showed that virus maturation does not involve large scale conformational changes in the capsid, but rather differences in the inner capsid region below that penton base (Perez-
Berna et al, 2009; Silvestry et al, 2009).
Amazingly, in recent years, atomic resolution structures of the AdV virion have been determined using both X-ray crystallography and cryo-EM (Liu et al, 2010; Reddy et al, 2010). In terms of organization, the AdV capsid is built almost entirely of hexons except at the vertices (Fig 1.1). Each of the 12 vertices contains a penton complex, which is composed of a penton base protein non-covalently associated with a fiber protein.
Capsid diameter is ~920 Å not including the fibers. AdV fibers are highly elongated and vary in size (roughly 120 to 315 Å) depending on virus type. The four minor proteins
IIIa, VI, VIII, and IX form interactions on the inner and outer surface of AdV that are
6 critical for capsid stabilization (Vellinga et al, 2005). Exactly how these minor proteins serve as cement to hold together the AdV virion is not yet fully understood and is currently an active area of research. As structure determination methods improve, it is becoming more feasible to visualize the location, folds, and interactions of the minor capsid proteins. Inside the capsid the viral genome is condensed along with proteins V,
VII, mu, Iva2, protease and terminal protein. There is a lack of structural information on this core region of the AdV virion containing viral DNA and associated proteins because the core lacks icosahedral symmetry, making it extremely challenging for analysis by cryo-EM and X-ray crystallography. Also, information is limited regarding adenovirus interactions with host proteins. Atomic resolution crystal structures of AdV fiber domains with fragments of CD46, CAR, and sialyl lactose have provided detailed knowledge on the molecular interactions necessary for attachment to host cells (Bewley et al, 1999;
Burmeister et al, 2004; Persson et al, 2007). Cryo-EM analysis of AdV penton base-cell integrin interactions has shed light on how the virus is internalized into clathrin coated vehicles (Stewart et al, 1997; Lindert et al, 2009). Additionally, a cryo-EM structure of
AdV complexed with anti-hexon antibody has been used to understand a postentry neutralization mechanism in which the hexon capsid is cross-linked by antibodies, thus blocking infection (Varghese et al, 2004). The research covered in this thesis in chapters
2-5 seeks to understand how immunologic and soluble serum factors impact AdV infection.
AdVs enter a variety of postmitotic eukaryotic cells using a stepwise entry program that ensures that their genome reaches the nucleus for replication (Greber et al,
1993; Puntener et al, 2011). For nuclear import to be effective, upstream steps for AdV
7 entry into cells entails traversing several physical barriers and a stepwise uncoating program of the virus, where cellular factors support or restrict the entry program (Mercer
& Greber, 2013; Suomalainen & Greber, 2013). AdVs do not directly cross the plasma membrane at the cell surface but are instead taken up into vesicles. Most AdV types utilize the coxsackie-adenovirus-receptor (CAR) for attachment to cells and integrins are required for internalization (Roelvink et al, 1998; Stewart & Nemerow, 2007). The AdV penton complex contains all necessary components to negotiate viral attachment and internalization into host cells. Attachment occurs via the fiber protein and the penton base protein interacts with cellular integrins through an Arginine-Glycine-Asparagine (RGD) motif located in a hypervariable surface loop to trigger internalization (Wickham et al,
1993). Some AdV types use alternative receptors for attachment including complement receptor CD46, sialic acid, and heparin sulfate proteoglycans (Arnberg et al, 2002;
Gaggar et al, 2003; Tuve et al, 2008; Zhang & Bergelson, 2005). Internalization results in uptake of AdV into clathrin-coated vesicles, which are then transported to the endosomal network. Although this strategy facilitates entry into cells, it does not provide direct access to their intracellular space for transport of DNA into the nucleus. Thus, AdVs entering the endocytic network must quickly subvert their vacuoles to avoid transport to the degradative lysosomal compartments within host cells. To escape endosomes and deliver DNA into the host nucleus, AdVs undergo stepwise disassembly, which involves removal of the protein capsid from the viral genome. This process begins at the plasma membrane, where receptor binding causes loss of protruding fibers. It continues in the early endosomes, where mild acidification causes release of a few vertex regions leading to exposure of the membrane-lytic protein VI and enhanced viral escape from
8 endosomes. The partially disrupted AdV particles then travel along microtubules until they reach the nucleus. The nucleus is protected by the nuclear envelope and AdVs can only gain access through nuclear pore complexes (NPCs) (Greber et al, 1997). The size restriction of NPCs precludes the AdV capsid from directly invading the nucleus. Upon docking at the NPC, final disassembly of AdV is mediated by the microtubule motor protein kinesin-1, NPC filament protein CAN/Nup214, and nuclear histone H1 allowing the viral genome to enter the nucleus for replication (Meier & Greber, 2004; Strunze et al, 2011; Trotman et al, 2001).
1.1.3 Adenoviruses as therapeutic agents
Over fifty years of intense research has resulted in AdVs being arguably one of the best-characterized of human DNA viruses. Detailed knowledge of the AdV replication live cycle has motivated numerous attempts to engineer AdVs as vectors for gene therapy, vaccination, and oncolysis (Russell, 2000). These viruses have several characteristics that make them well suited for therapeutic interventions in humans. First, they have low pathogenicity in immunocompetent individuals, with symptoms that are mild and often associated with the common cold. Second, they are popular for their ability to accommodate relatively large transgenes. AdV-based gene delivery vectors can tolerate foreign gene insertions of up to 7.5 kb pairs of DNA. This is significantly larger than the size of the average human gene, which is about ~1.4 kb pairs of DNA (Lander et al, 2001). Third, AdVs are able to rapidly infect a broad range of human cells and tend to yield high levels of gene transfer. Fourth, the viral genome doesn’t undergo rearrangement at a high rate, and inserted genes are stable over successive rounds of
9 replication. Finally, AdV vectors are easy to manipulate using standard recombinant
DNA techniques.
Indeed, AdV vectors have been extremely popular in efforts to advance therapy for infectious and inherited diseases, yet there has been limited clinical success. Major obstacles that hinder their use as therapeutic agents include: the innate immune response to AdV challenge, pre-existing immunity to AdV vectors, and poor efficiency at targeting gene transfer to specific cell types. Activation of innate immunity to AdV is associated with a reduction in gene transfer efficiency (Worgall et al, 1997) and an acute inflammatory response that can cause serious damage to healthy tissues and may even lead to death (Raper et al, 2003). The limitation posed by pre-existing immunity is anti-
AdV antibodies, which can prevent AdV vectors from transducing target cells (Zaiss et al, 2009). In terms of targeting specific cell types, a common goal is delivery via the bloodstream; however, upon injection, the majority of virus particles accumulate in the liver. Furthermore, the primary receptor for AdV is virtually ubiquitous and causes non- specific uptake into multiple organs. Even with these present challenges, AdVs are among the most commonly used vectors for therapeutic interventions in humans, second only to retroviruses. Research is currently underway, to design safe vectors that can efficiently transduce target tissues.
1.2 Adenovirus-host interactions
AdVs deliver their genome into the nucleus of human cells for replication. They accomplish this task by co-opting host cell machinery whilst avoiding host immunity. In turn, just as these viruses commandeer cellular factors to replicate, cells have evolved
10 elaborate mechanisms to recognize and eliminate AdVs during infection. This evolutionary struggle between virus and host is poorly understood. There is much to be learned from studying AdV-host interactions and this has been the focus of our research.
This section will introduce our studies.
1.2.1 Host immune response to adenovirus
A major barrier to AdV replication is the host immune response. Immunity to viral infections occurs by multiple specific and non-specific mechanisms that vary with respect to activation, duration, and magnitude (Lowenstein & Castro, 2003; Muruve,
2004). Ultimately the reaction of the immune system is modulated by how AdV interacts with host cells and spreads during infection. Viral antigens will be present in different parts of the body depending on both the route of spread and phase of infection. Local infections, for example at mucosal surfaces, can elicit local cell-mediated and humoral
(IgA) immune responses, but not necessarily systemic immunity (Lemiale et al, 2003).
One type of local response involves defensins, which are naturally occurring immune peptides that block AdV infection (Nguyen et al, 2010; Smith & Nemerow, 2008; Smith et al, 2010a). Chapter 2 of this thesis reveals new details on the defensin anti-viral mechanism against AdV. Virus and/or virus-infected cells can also stimulate B lymphocytes to produce antibodies that are specific for viral antigens, primarily against the AdV capsid proteins penton base, hexon, and fiber (Bradley et al, 2012). IgG, IgM, and IgA all exert anti-viral activity against AdV (Ariyawansa & Tobin, 1976). Antibodies can neutralize AdV by: 1) preventing virus-host cell interactions, 2) blocking post-entry steps of AdV infection, or 3) recognizing antigens on infected cells, which can lead to
11 antibody-dependent cytotoxic cells or complement-mediated lysis (Cichon et al, 2001;
Varghese et al, 2004). IgG antibodies are responsible for most anti-viral activity in serum, whereas IgA activity is prevalent when AdVs infect mucosal surfaces (Parkin & Cohen,
2001). This type of immunity is commonly referred to as humoral. Cell mediated immunity involves leukocytes and their production of cytokines in response to virus and virus-infected cells. Cytotoxic lymphocytes, natural killer cells and antiviral macrophages can recognize and kill AdV-infected cells (Thaci et al, 2011; Zhu et al, 2007). Helper T cells can recognize virus-infected cells and produce a number of important cytokines.
Cytokines produced by monocytes, T cells, and natural killer cells play important roles in developing anti-AdV immune responses (Chen & Lee, 2013). Cell-mediated immunity is principally thought to be regulated by toll-like receptors. Toll-like receptors mediate the anti-viral immune response by recognizing virus infections, activating signaling pathways and inducing the production of cytokines and chemokines (Xagorari & Chlichlia, 2008).
Chapter 3 focuses on a new mechanism by which toll-like receptor 4 triggers an immune response to AdV infection. The early, non-specific responses (for example natural killer cell activity and interferon) serve to limit virus infection during the acute phase. The later specific immune responses (humoral and cell-mediated) function to eliminate viral infections at the end of the acute phase, and to maintain specific resistance to infection.
1.2.2 Recognition of adenovirus by human alpha defensin 5
AdVs gain access to their host cells using a multi-step process that involves clathrin-mediated endocytosis and acid activated penetration of endosomes (see section
1.1.2). Successful entry for these non-enveloped viruses depends on the orchestrated
12 disassembly program of their outer capsids. Human alpha defensin 5 (HD5) has been shown to stabilize the AdV capsid and prevent disassembly of the virus during cell entry, thereby blocking infection (Nguyen et al, 2010; Smith & Nemerow, 2008; Smith et al,
2010a). HD5 is a protein of the innate immune system that is produced in Paneth cells of the small intestine. This protein contains 32 amino acids and shares a common structural fold of the alpha-defensin family characterized by an anti-parallel β-sheet structure stabilized by three intramolecular disulfide bonds. Also it is mainly positively charged, amphipathic, and possesses the ability to multimerize both in solution and upon ligand binding. These properties dictate HD5 anti-viral activity against non-enveloped viruses
(Gounder et al, 2012). Although it is has been clearly demonstrated that HD5 stabilizes the AdV virion and prevents release of viral proteins including the membrane lytic protein VI required for endosomal escape, the structural and functional organization that drives this process is poorly defined. Unraveling the mechanism underlying this AdV- host interaction is important for several reasons. First, this may be a common mechanism for restricting infection of non-enveloped viruses. For instance, HD5 has been shown to block human papillomavirus (HPV) from escaping endocytic vesicles (Parker et al,
2006). Thus, there is clear potential for eventual development of broad-spectrum anti- viral agents that target virus disassembly during cell entry. Second, insights gleaned from the mechanism may be useful for optimizing AdV vectors for gene delivery. For example, a capsid modified AdV vector may increase the efficacy of therapeutic gene delivery by avoiding interaction with HD5. Finally, studies of the AdV-HD5 interaction provide insight into AdV stepwise disassembly during cell entry.
13
What do we currently know about HD5 neutralization of AdV? From an initial investigation of how HD5 antagonizes AdV infection, it was discovered that HD5 is a potent inhibitor and is capable of complete inhibition with an IC50 between 3 and 4 µM
(Smith & Nemerow, 2008). Evidence indicated that this potent anti-viral activity involves
HD5 binding to AdV outside the cell and preventing release of the internalized AdV-
HD5 complex from the endosome. This conclusion was based on failure of AdV to mediate the translocation of ribotoxin α-sarcin from the endosome into the cytoplasm in the presence of inhibitory concentrations of HD5. Consistent with this finding, HD5 was shown to inhibit release of the viral protein VI, which is required for AdV-mediated endosome penetration. Moreover, at late times post-infection, AdV particles colocalized with lysosomes instead of the nucleus. This further demonstrated that HD5 causes AdV to remain trapped in the endosomal/lysosomal pathway rather than trafficking to the nucleus. To gain deeper insight into the defensin-mediated mechanism of neutralization, the specificity of the AdV-HD5 interaction was analyzed. Infectivity studies revealed that sensitivity of AdV to HD5 inhibition is species specific. Sensitivity to HD5 was found for many AdV types from species A, B, C, and E, while types from species D and F were resistant to neutralization (Smith et al, 2010a). It was observed that neutralization is dependent upon the tertiary structure of a correctly folded HD5 molecule. HD5 derivatives in which six cysteines were replaced with L-α-aminobutyric acid, to prevent formation of the three intramolecular disulfide bonds that stabilize a correctly folded
HD5 molecule, failed to inhibit AdV infection. A cryo-EM structure of a sensitive AdV vector in complex with HD5, combined with sequence analysis of capsid proteins from sensitive and resistant AdV types, led to a hypothesis that the critical HD5 neutralization
14 site is located in a region spanning the penton base and fiber proteins (Smith et al,
2010a). A critical binding site at the interface of the non-covalently coupled penton base and fiber was supported by two key observations. First, there was an accumulation of extra density at the top of penton base and around the fiber shaft in the cryo-EM structure of the AdV vector complexed with HD5 that was not present in a cryo-EM structure of the AdV vector alone. Second, this hypothesis was supported by infectivity assays, where
HD5 activity was tested against virus chimeras comprised of capsid proteins from sensitive and resistant AdV types. These experiments confirmed the presence of multiple binding determinants in the penton complex that are critical for neutralization and suggested how HD5 may bridge adjacent capsid proteins fiber and penton base to prevent disassembly during cell entry. The goal of our research has been to gain insight into the structural basis for AdV susceptibility to HD5 anti-viral activity and to better define the critical neutralization site. This research is covered in detail in chapter 2 of the thesis.
1.2.3 Recognition of adenovirus by serum factors
As research in AdV biology progresses, a new model for cell infection is emerging, which appreciates that AdV biodistribution, spread, viral persistence, and replication is less reliant on direct receptor binding and more influenced by other complex interactions between virus and host. A first glimpse into this reality came from directly injecting AdV vectors into the bloodstream of mice. The biodistribution of these vectors revealed no correlation with CAR expression (Akiyama et al, 2003; Alemany &
Curiel, 2001). Instead, it was later discovered that virus particles accumulate in the liver and that virus entry into hepatocytes is mediated by interaction between the AdV capsid
15 protein hexon and vitamin K-dependent blood coagulation factors (Kalyuzhniy et al,
2008; Parker et al, 2006; Shayakhmetov et al, 2005; Waddington et al, 2008). Thus, a novel fiber-independent AdV entry mechanism was identified. In vitro studies of this mechanism have shown that several homologous blood coagulation factors, including factors VII, IX, and X (FVII, FIX, and FX) can support virus infection of susceptible cells (Parker et al, 2006). However, in vivo, specific inactivation of FX alone is sufficient to completely abrogate hepatocyte transduction in mice after intravenous administration.
A detailed description of this AdV-FX interaction is of fundamental importance both for understanding basic AdV biology in vivo and for refinement and optimization of AdV vectors for gene therapy. From a gene therapy perspective, FX-mediated sequestration of
AdV vectors into liver hepatocytes and Kupffer cells induces an acute inflammatory response that may be solely responsible for the morbidity and mortality associated with infection (Ni et al, 2005; Raper et al, 2003), and thus, represents the greatest constraint on safety. Therefore, interference with this process could potentially improve the safety of AdV vectors. Furthermore, the highly efficient interaction between AdV-FX and liver cells is a significant hindrance if gene delivery to extrahepatic cells and tissues is required. Precise knowledge on the AdV-FX interaction may inform strategies to bypass this effect. Moreover, studies on AdV-FX interaction might impact development of effective drugs for treating disseminated AdV infections, which are frequently associated with liver failure and a high virus burden in the blood.
Although extensive in vitro, ex vivo, and in vivo analyses has defined a critical role for coagulation FX in facilitating AdV invasion of liver hepatocytes, the exact mechanism of FX binding to AdV has remained elusive. The earliest attempts to
16 understand AdV liver tropism following intravenous administration focused on modifying penton base and fiber proteins. Mutations to ablate CAR and integrin binding worked as expected in vitro, but did not alter tropism in vivo (Akiyama et al, 2003;
Alemany & Curiel, 2001). Shortly after, the focus shifted to investigating AdV interactions with coagulation factors after a seminal study reported that they play a major role in targeting intravenously injected AdV vectors to hepatic cells (Shayakhmetov et al,
2005). Both in vitro and ex vivo data from mice showed that blood factor mediated uptake of virus into hepatocytes is CAR-independent and instead occurs through cell surface heparan sulfate proteoglycans and low-density lipoprotein receptor-related protein. Later studies demonstrated that FX mediates in vivo delivery of AdV into hepatocytes and revealed an unanticipated function for hexon in defining virus infectivity (Kalyuzhniy et al, 2008; Vigant et al, 2008; Waddington et al, 2008). Moderate resolution (< 20 Å) cryo-
EM structures of an AdV vector bound to FX localized attachment to the viral capsid protein hexon (Kalyuzhniy et al, 2008; Waddington et al, 2008). Furthermore, hexon- mutated virus bearing a large insertion within the top of hexon showed markedly reduced
FX binding in vitro and failed to deliver a transgene to hepatocytes in vivo. In our study, we used high-resolution cryo-EM and guided molecular dynamics simulations to characterize the interaction interface between FX and human AdV type 5 hexon. We also provided a link between this interaction and innate immunity. This work is covered in chapter 3. In a follow-up study we analyzed the interaction between FVII and human
AdV type 5 hexon. Coagulation FVII, like FX, binds hexon with very high affinity, yet has a poor capacity for supporting virus entry into hepatocytes. Our results indicated a structural basis for differential infectivity. An explanation is given in chapter 4.
17
1.2.4 Recognition of an adenovirus capsid-incorporated HIV antigen
AdVs are promising vectors for therapeutic interventions in humans. To date, they have been explored as vaccines against cancer and infectious diseases (Tatsis & Ertl,
2004). They are appealing as vaccine carriers based on their ability to induce high antibody titers and robust cytotoxic T-lymphocyte responses. Additionally, it is thought that the high immunogenicity inherent in these vectors may have an adjuvant effect.
These characteristics have underpinned their utility as carriers to deliver antigens from other infectious agents for vaccination. Major efforts are underway for their use to protect against HIV. HIV destroys CD4+ T cells eventually overwhelming the capacity of the immune system to regenerate or fight off other infections. A vaccine to HIV remains elusive. The requirements for eliciting protective neutralizing antibody and cellular responses remain a significant challenge. However, human clinical trials of candidate
HIV vaccines show modest efficacy, suggesting that it should be possible to generate vaccine-elicited protection against HIV infection (Kwong et al, 2011). For example, AdV type 5 vectors containing HIV gene inserts are notable for their ability to induce strong
HIV-specific cellular immune responses and have advanced to clinical trials. One major obstacle to the development of an effective AdV-based HIV vaccine is pre-existing immunity. Neutralizing antibodies, even present at moderate titers, can drastically restrict
AdV vector uptake by cells, including antigen presenting cells (Fitzgerald et al, 2003). As a result, heterologous antigens that are packaged as transgenes within the viral genome are not efficiently expressed, thereby limiting the induction of an antigen-specific immune response. While neutralizing antibodies can inhibit AdV vectors extracellularly,
AdV-specific CD8+ T cells destroy vector expressing cells (Mittal et al, 1993; Yang et al,
18
1996). Studies have shown these effects in animals (Fitzgerald et al, 2003; Xiang et al,
2002) and in humans (Knowles et al, 1995). The mechanisms underlying the dampening effect of pre-existing immunity are still being investigated. Meanwhile, alternative strategies have been developed for using AdV vectors as vaccines against HIV.
One alternative to bypass acquired immunity has focused on AdV types to which the human population is less exposed. In particular, promising results have been obtained from vaccine vectors based on AdVs that infect chimpanzees (Dicks et al, 2012; O'Hara et al, 2012; Xiang et al, 2006). Another approach involves direct incorporation of antigens into the viral capsid. The viral capsid protein hexon has been utilized for antigen display due to its natural role in generation of an anti-AdV immune response and abundance within the AdV virion (Crompton et al, 1994; Matthews et al, 2010; Matthews et al, 2008). Recently an AdV vector presenting an HIV antigen on the surface of the hexon capsid was shown to elicit an anti-HIV cellular response (Matthews et al, 2010).
The HIV-1 antigen is 24 amino acids of the HIV membrane proximal ectodomain region
(MPER). It is derived from the HIV glycoprotein gp41. Viral protein gp41 is part of a complex, often referred to as a “spike”, which decorates the envelope of HIV and is responsible for attaching to and fusing with host cell membranes. The conserved 24 amino acid MPER sequence is a target of two neutralizing human monoclonal antibodies,
2F5 and 4E10, and is therefore an important lead for vaccine design (Zwick et al, 2005).
Here, it has been genetically incorporated within the AdV hexon hypervariable region 2
(HVR2). HVR2 is a highly immunogenic site located at the top of hexon. A critical aspect of understanding how this immunologically promising AdV vector elicits protection against HIV is knowledge of how MPER is displayed at the surface of the viral
19 capsid. Therefore, we undertook a cryo-EM structural study to visualize heterologous antigen display. We found that the AdV capsid influences epitope structure, flexibility and accessibility, all of which affect the host immune response. The results of this work are shared in chapter 5.
1.3. A hybrid structural approach to analyzing adenovirus recognition
Studies on the structural biology of AdV and its interactions with host proteins have benefited tremendously from the combination of cryo-EM structures of full complexes with available atomic structures of components from X-ray crystallography.
This approach was first utilized to study the structure of AdV in the 1990s (Stewart et al,
1993). We have used this method to extend our knowledge on how AdV is recognized during infection. A new hybrid modeling tool, molecular dynamics flexible fitting
(MDFF), allowed us to merge detail from atomic-scale structures with the overall architecture of AdV-host complexes captured in cryo-EM density maps. This section contains information on our approach.
1.3.1 Preparation of samples for cryo-EM
For observation by cryo-EM, biological samples must be cryogenically frozen prior to being examined in a transmission electron microscope. The technique of vitrifying samples for cryo-EM studies was first introduced over 30 years ago. In particular, Dubochet and colleagues demonstrated that a thin layer of an unfixed and unstained specimen, such as a virus, could be vitrified for electron microscopy imaging
(Adrian et al, 1984; Dubochet et al, 1981). Their new freezing method did not damage the
20 specimen but rather cryo-EM images were produced that revealed an impressive level of structural detail. Since this breakthrough discovery, advances in sample preparation have made it possible to examine a wide range of biological specimens from intact tissue sections and plunge-frozen cells to bacteria, viruses, and proteins. In terms of sample preparation, cryo-EM has a distinct advantage over X-ray crystallography in that crystals are not required for structure determination. The size of AdV (~150 MDa) and complexity of AdV-host interactions makes crystallization extremely difficult, if not impossible. Another advantage is that cryo-EM requires relatively little material. A single grid covered by a 2-6 µL droplet of concentrated sample (1 x 1013 virus particles per ml) should easily contain enough particles for an entire data set. The droplet is frozen in a cryogen such that the virus or virus mixture is preserved in a near native or physiologically relevant state. Plunging is done either manually or automatically.
Automatic plunging devices should increase the reproducibility of the grids. Once the sample is frozen, it must be maintained below the devitrification temperature, which is approximately -137°C (Dubochet et al, 1982), to avoid conversion of the amorphous vitreous ice within the specimen into cubic or hexagonal ice. Crystalline ice formation would cause mechanical stresses on the sample, and those stresses might distort the structures.
Preparing AdV for cryo-EM imaging is a multi-step process. AdV is a biosafety level 2 virus and thus requires the use of adequate containment equipment and practices.
Purified virus must be dialyzed out of cesium chloride and into a low salt solution suitable for a high-resolution cryo-EM study. We used Thermo Scientific Slide-A-Lyzer mini unit dialysis devices (10 kDa cutoff) to facilitate removal of cesium chloride. The
21 low salt solution used for buffer exchange contained 150 mM sodium chloride, 25 mM
Tris and was at a pH of 8 (2 L total – 1 fresh L per hour). Exchange was done at 4°C for two hours under gentle stirring conditions. Afterwards, virus was incubated with protein to form a complex. For the AdV-defensin project, virus (160 µg/mL) was combined with
HD5 (5 µM). For the study involving serum factors, 100 µL of AdV at 1 x 1013 virus particles per mL was mixed with either 1 µL of FVII or FX (2mg/mL). We did not mix any protein with the AdV-based HIV vaccine. It is worth noting that some of the AdVs used in these studies were modified to contain short fibers, making them easier to study by cryo-EM. Before plunging the sample, cryo-EM grids were prepared. Quantifoil holey carbon grids (Cu 400 mesh, R2/4, SPI Supplies, US) were freshly glow discharged under
25 mA for 30 seconds on both sides. Sample was then applied to the grids. A total of 6
µL of sample was applied to each grid. Initially, 3 µL of sample was added and immediately blotted by touching the back of the grid with filter paper. We used Whatman filter paper cut into rectangular strips for blotting. After applying an additional 3 µL, the grid was allowed to sit for 20 seconds and was then blotted. The blot time ranged from
10-20 seconds depending on room temperature and humidity. The sample was then vitrified by plunging into liquid ethane. Frozen grids could then be imaged or were stored in a liquid nitrogen dewar for later use.
1.3.2 Virus imaging and three-dimensional reconstruction
After plunging, the next critical step is grid transfer into the microscope. The main objective here is to place the cryo-EM sample in the microscope without warming it above the devitrification temperature and without collecting too much condensation from
22 water droplets in the air. We used an FEI Polara cryo-workstation for grid transfer. The workstation comes equipped with a loading area, vacuum system and an airlock for docking to an FEI Polara transfer device, making it easy and safe to transfer grids into the microscope. The transfer device can preserve up to 5 cryo-grids under liquid nitrogen for transfer into the electron microscope. The design is such that once the device is mounted onto the microscope, one grid is manually inserted while the others are kept at liquid nitrogen temperature. The electron microscope that we used for imagining is an FEI
Polara field emission gun (FEG) transmission electron microscope with a 300 kV electron beam and a highly stable specimen cryo-stage. The microscope was operated at liquid nitrogen temperature for data acquisition to avoid build-up of excessive contamination on the cold specimen. Also, a low dose of electrons (~20 e-/Å2) was used during imaging due to the fact that cryo-EM samples are extremely radiation sensitive.
Images were collected at a magnification of 397,878X on a Gatan Ultrascan 4000 CCD camera. This magnification corresponds to a pixel size of 0.4 Å on the molecular scale.
The defocus values of particle images ranged from -1 to -4 µm. Datasets for the projects varied in size, with the largest being 5,025 for the AdV-based vaccine. For the AdV-HD5 study, 3,515 and 3,620 particle images were collected of HD5 in the presence of a sensitive and a resistant AdV respectively. The dataset size for AdV with serum factor
VII was 1,503 and for AdV with factor X was 1,101.
Each dataset was handled as described here. Individual particles were extracted from cryo-EM images with in house scripts that call IMAGIC subroutines (van Heel et al,
1996). Particle images were computationally binned for the initial stages of refinement.
Averaging of adjacent pixels during the initial phases of image processing helped to
23 improve the image contrast. This, in turn, aided in determining the initial, relatively crude, orientation parameters for each particle in the stack. The program CTFFIND3
(Mindell & Grigorieff, 2003) was used to determine initial estimates for the microscope defocus and astigmatism parameters. The image reconstruction software FREALIGN was used for determining three-dimensional cryo-EM structures. A 6 Å cryo-EM structure of
AdV (Saban et al, 2006) served as a starting model for refinement. Essential steps in refinement included: 1) determining the center in x, and y for each particle in the image stack, and 2) assigning the three Euler angles that describe the projection view, or three- dimensional orientation, for each particle in the image stack. After these five parameters were estimated for each particle image, the particles could be averaged to generate a preliminary three-dimensional reconstruction. We performed many iterative rounds of refinement to improve the orientational parameters, and hence the resolution of our cryo-
EM structures. In the later rounds of refinement we used particle images with smaller binning factors to make use of the higher resolution information contained in the finer pixel sizes. We also refined defocus on a per particle basis at this stage. A key aspect of our refinement was the viral capsid, which has 60-fold symmetry. Icosahedral symmetry averaging effectively increased the number of asymmetric units that were averaged to produce the final cryo-EM structure. In doing so, any features of the virus that followed icosahedral symmetry were strengthened, whereas any features, such as the protruding fibers, that did not follow this symmetry were weakened or averaged away. This procedure allowed us to determine subnanometer resolution cryo-EM structures for the projects discussed in chapters 2-5. Final resolutions were based upon the Fourier shell
24 correlation 0.5 threshold criterion, which was first used to define the resolution of icosahedral viruses in the late 1990s (Bottcher et al, 1997; Conway et al, 1997).
1.3.3 Modeling of cryo-EM maps
State-of-the-art cryo-EM routinely allows direct visualization of whole viruses in their near native state at subnanometer resolutions. This impressive capability is largely due to rapid developments in microscope design, imaging hardware, automation, and image processing, and has yielded astonishing results, which for instance, allowed a huge leap from a stunning first image of AdV in vitrified ice in 1984 (Adrian et al, 1984) to a
3.6 Å cryo-EM structure in 2010 (Liu et al, 2010). Thus, we used this approach to study
AdV-host interactions at subnanometer resolutions. Detail at this resolution level distinguished viral capsid proteins involved in host interactions and informed protein- protein interfaces for virus complexes. For the AdV vaccine vector, subnanometer resolution revealed how an HIV antigen is displayed at the surface of the capsid.
Interpretation of cryo-EM density involved isolating individual subunits, identifying secondary structures, and accurately fitting atomic models. We used several tools for completing these steps. Cryo-EM density was visualized and segmented for study using the UCSF Chimera software package (Pettersen et al, 2004). This program was also used to initially dock atomic resolution structures or computational atomic models into segmented density. In the absence of crystal structures, computational model building was done using a combination of Chimera, I-TASSER (Zhang, 2008), and Rosetta (Rohl et al, 2004). After components were docked into density we used cryo-EM guided
Molecular Dynamics Flexible Fitting (MDFF) (Chan et al, 2012; Trabuco et al, 2008) to
25 further refine the docked models into cryo-EM density. Setup and analysis of MDFF simulations was carried out using a program called Visual Molecular Dynamics (VMD)
(Humphrey et al, 1996). VMD has several useful plugins for analyzing simulations that do tasks such as measuring the root mean square deviation over time, calculating agreement between atomic model and input cryo-EM density map, and evaluating energies of the system over the course of simulation. This general protocol was used in all of our structural studies.
We used MDFF to generate atomic models based on our density maps and, for this reason, time is spent here to briefly highlight how this approach works. With MDFF, a molecular dynamics simulation is performed using a starting atomic model, often built from crystal structures. The goal is to refine these crystal structures or computationally constructed atomic models of components into cryo-EM density to better understand the interactions between the docked units for insight into how function is orchestrated.
Additionally, this method reveals differences between conformations of structures in crystal form versus conformations captured by cryo-EM. Simulations are carried out such that external forces proportional to the gradient of the density map are applied to all atoms, driving the models to occupy regions of high density. Each of the steps described here were done using VMD. More information on each of these steps as well as how to setup and run a MDFF simulation can be read about in detail at the developer’s website located at http://www.ks.uiuc.edu/Research/vmd/plugins/mdff/. First, atomic models are checked for errors in stereochemistry and peptide bond configuration. Then, structural restraints are applied to preserve secondary structural elements so that over-fitting does not occur during the simulation. Afterwards, an MDFF simulation can be setup. The
26 simulations we performed applied a g-scale factor of 0.3 for guiding secondary structural elements into density. The g-scale dictates how strongly the atomic models are pulled into density. A low g-scale of 0.3 was chosen to help prevent over-fitting. All MDFF simulations were carried out using NAMD 2.8 (Phillips et al, 2005) and the CHARMM27 force field. Simulations were setup to run for 100 picoseconds under implicit solvent conditions. They were performed on the Case Western Reserve University High-
Performance Computing Cluster. Typically, these simulations would take ~6 hours to complete, but this time varies with the number of atoms in the system.
27
Figure 1.1 Schematic representation of the structural organization of AdV based on cryo-
EM and X-ray crystallography. Figure is adapted from Russell 2009 and is published here in agreement with the Creative Commons Attribution License.
28
CHAPTER 2: Recognition of adenovirus by human alpha defensin 5 involves intrinsic disorder
This chapter is published: PLoS ONE 8(4): e61571. doi:10.1371/journal.pone.0061571
Research Article:
An intrinsically disordered region of the adenovirus capsid is implicated in neutralization by human alpha defensin 5.
Justin W. Flatt*, Robert Kim*, Jason G. Smith, Glen R. Nemerow, Phoebe L. Stewart
*Contributed equally
2.1 ABSTRACT:
Human α-defensins are proteins of the innate immune system that suppress viral and bacterial infections by multiple mechanisms including membrane disruption. For viruses that lack envelopes, such as human adenovirus (HAdV), other, less well defined, mechanisms must be involved. A previous structural study on the interaction of an α- defensin, human α-defensin 5 (HD5), with HAdV led to a proposed mechanism in which
HD5 stabilizes the vertex region of the capsid and blocks uncoating steps required for
29 infectivity. Studies with virus chimeras comprised of capsid proteins from sensitive and resistant serotypes supported this model. To further characterize the critical binding site, we determined subnanometer resolution cryo-electron microscopy (cryo-EM) structures of HD5 complexed with both neutralization-sensitive and -resistant HAdV chimeras.
Models were built for the vertex regions of these chimeras with monomeric and dimeric forms of HD5 in various initial orientations. Cryo-EM guided molecular dynamics flexible fitting (MDFF) was used to restrain the majority of the vertex model in well- defined cryo-EM density. The RGD-containing penton base loops of both the sensitive and resistant virus chimeras are predicted to be intrinsically disordered, and little cryo-
EM density is observed for them. In simulations these loops from the sensitive virus chimera, interact with HD5, bridge the penton base and fiber proteins, and provides significant stabilization with a three-fold increase in the intermolecular nonbonded interactions of the vertex complex. In the case of the resistant virus chimera, simulations revealed fewer bridging interactions and reduced stabilization by HD5. This study implicates a key dynamic region in mediating a stabilizing interaction between a viral capsid and a protein of the innate immune system with potent anti-viral activity.
2.3 INTRODUCTION:
Human α-defensins are small (3–5 kDa), positively charged, amphipathic, naturally occurring peptides that are abundant in neutrophils and Paneth cells of the small intestine (Ganz, 2003). Structure determination of these molecules revealed that they have a three-stranded beta-sheet fold stabilized by disulfide bonds and can readily form dimeric complexes (Hill et al, 1991; Szyk et al, 2006; Zhang et al, 2010a). New
30 functional studies showed the importance of dimerization for α-defensin mediated inactivation of both bacteria (Rajabi et al, 2012) and viruses (Gounder et al, 2012), and a structural study of membrane-bound α-defensin supports a dimer pore mechanism for membrane disruption (Zhang et al, 2010b). Humans express six α-defensins (HNP1–4,
HD5, and HD6) and multiple β-defensins that are distinguished by the arrangement of their disulfide bonds and their expression patterns. Of the six α-defensins, HD6 forms an atypical dimer that undergoes further ordered self-assembly to form fibrils that entangle bacteria (Chu et al, 2012).
Currently, there is little structural information on the recognition of microbial agents by defensins. The antibacterial activities of defensins against both Gram-positive and Gram-negative organisms have been characterized with the major bactericidal mechanism involving membrane disruption, although other mechanisms have been recently proposed (Brogden, 2005; Lehrer & Lu, 2012; Wilmes et al, 2011). An understanding of the antiviral properties of defensins is beginning to emerge (Gounder et al, 2012; Smith et al, 2010a). For enveloped viruses, defensins can suppress viral infection by direct inactivation of the virion via membrane disruption, interference with viral membrane fusion (Demirkhanyan et al, 2012) and by modulation of immunity and other biological responses of the host (Steinstraesser et al, 2011). Viruses that lack envelopes including human adenovirus (HAdV), human papillomavirus (HPV), and polyomaviruses are neutralized by α-defensins despite the absence of a lipid membrane target (Bastian & Schafer, 2001; Dugan et al, 2008; Gropp et al, 1999; Harvey et al,
2005; Parker et al, 2006; Smith & Nemerow, 2008).
31
A previous structural and functional characterization of α-defensin neutralization of HAdV showed that the mechanism of inactivation is species specific and dependent upon α-defensin tertiary structure (Smith et al, 2010a). A cryo-EM structural analysis of
HD5 complexed with a neutralization-sensitive HAdV chimera led to a model for neutralization in which HD5 binds at a point of contact between the vertex proteins penton base and fiber, preventing release of the fiber protein and stabilizing the capsid.
Loss of the vertex complex formed by penton base and fiber is presumed to be a required step in HAdV cell entry (Lindert et al, 2009; Nakano et al, 2000). In particular, dissociation of the penton base while the virus particle is in the endosome allows release of the internal viral protein VI, which is membrane lytic and can disrupt the endosomal membrane (Moyer & Nemerow, 2011; Wiethoff et al, 2005). The previous structural analysis of a HAdV/HD5 complex was at moderate (12 Å) resolution and led to the identification of strong HD5 binding proximal to the interface between penton base and fiber. Sequence analysis identified a negatively charged region in fiber that might form favorable interactions with HD5 and that was only present in sensitive HAdV species.
Infectivity studies of virus chimeras with this fiber region swapped between sensitive and resistant HAdV species indicated that this region, together with penton base, is involved in adenoviral neutralization (Smith et al, 2010a).
Here we use cryo-EM and Molecular Dynamics Flexible Fitting (MDFF) simulations to further characterize the HD5 binding site at the interface between the
HAdV penton base and fiber proteins. Subnanometer (8–9 Å) resolution cryo-EM structures of HD5 complexed with both neutralization-sensitive and -resistant HAdV chimeras are presented. The MDFF simulations indicate that for the sensitive HAdV
32 chimera HD5 binding is likely to involve a penton base surface loop that is predicted to be intrinsically disordered. The functions of intrinsically disordered proteins include regulation, signaling and molecular recognition, and they may serve to promote binding to multiple interaction partners (Dunker et al, 2001; Uversky et al, 2008). Their hallmark is the absence of rigid 3D structure under physiological conditions. Cryo-EM density is missing for the intrinsically disordered loops of the penton base. Therefore, MDFF was used to restrain the majority of the vertex model in well-defined cryo-EM density while the conformation of the loops and their interaction with HD5 was dependent mainly on the potential energy function. Analysis of the stabilization effect calculated for HD5 binding at this site of the sensitive HAdV chimera provides a plausible molecular mechanism for HAdV susceptibility to defensin inactivation.
2.4 RESULTS:
CRYO-EM STRUCTURES OF HD5 COMPLEXED WITH NEUTRALIZATION-
SENSITIVE AND –RESISTANT HADVS
In our previous structural study on the interaction of HD5 with a neutralization- sensitive HAdV chimera, Ad5.F35, we observed thousands of binding sites for HD5 over a significant portion of the capsid surface (Smith et al, 2010a). In particular we noted that
HD5 interacts with flexible surface loops of the major capsid proteins, hexon and penton base, as well as with fiber shaft. Visualization of thousands of binding sites on the capsid is consistent with the results of an equilibrium-binding assay that showed approximately
3,000 HD5 molecules bound to each HAdV-5 virus particle (Smith et al, 2010a). HD5 binds at lower levels to resistant HAdVs, suggesting that there may be high and low
33 affinity sites on the capsid and that HD5 binding to a subset of these sites on sensitive virus species results in neutralization. The previous structural study was performed with a relatively high (20 µM), neutralizing concentration of HD5 in order to saturate all of the possible binding sites on the virion (Ganz, 2003; Smith & Nemerow, 2008). The former analysis implicated both a negatively charged region near the N-terminus of the fiber
(i.e. 18-EDES-21 in Ad5.F35) and the penton base in HD5 mediated neutralization. In resistant HAdV types the corresponding region of the fiber is non-polar and positively charged (i.e. 18-GYAR-21 in HAdV-19c).
In the current study we used a lower, but still neutralizing concentration of HD5
(5 µM) with the goal of predominantly populating and visualizing individual high affinity binding sites. To this end, we determined subnanometer (<10 Å) resolution cryo-EM structures of HD5 in complex with two HAdV chimeras comprised of capsid proteins from resistant and sensitive serotypes that were previously examined for susceptibility to
HD5 antiviral activity (Smith et al, 2010a). Ad5.F35 contains the short-shafted HAdV-35 fiber incorporated into the HAdV-5 capsid and is sensitive to HD5 neutralization. The sensitivity of Ad5.F35 to HD5 is comparable to that of HAdV-5 and HAdV-35 (Smith et al, 2010a). Ad5.PB/GYAR, which has the HAdV-19c penton base within the HAdV-5 capsid and a mutated HAdV-5 fiber (with 18-GYAR-21), is completely resistant to HD5 activity.
The neutralization-sensitive and -resistant HAdV chimeras were incubated in the presence of 5 µM HD5 then applied to grids and flash frozen for cryo-EM analysis. This
HD5 concentration is sufficient for neutralizing the sensitive HAdV chimera. Datasets
34 were collected on an FEI Polara microscope (300 kV, FEG) under liquid nitrogen temperature. Image processing was performed as described earlier (Saban et al, 2006), and the final 3D reconstructions for Ad5.F35+HD5 and Ad5.PB/GYAR+HD5 have resolutions of 9.7 Å and 8.1 Å, respectively, by the Fourier Shell Correlation 0.5 threshold criterion (Fig. 2.1 and Fig. S2.1).
The most noticeable differences between these two virus-defensin structures lie within the vertex region, and are largely due to penton base and fiber substitutions. The short-shafted fiber is fully visible in the Ad5.F35+HD5 structure (Fig. 2.1A,B), whereas the longer fiber is only partially reconstructed in the Ad5.PB/GYAR+HD5 structure (Fig.
2.1D,E). Fitting of atomic models for the HAdV-5 and HAdV-19c penton base into our cryo-EM density maps shows several α-helices from the docked atomic models that match strong regions in the cryo-EM maps, indicating that the resolution for these complexes is in the subnanometer range (Fig. 2.1C,F and Fig. S2.1). Compared to the previous cryo-EM study which used 20 µM HD5 (Smith et al, 2010a), much less HD5 density is observed. Previously, we found that HD5 interacts with multiple flexible regions within hexon, penton base, and fiber proteins of the HAdV capsid. The strongest
HD5 density observed on top of the hexons in our prior study is similar to what is observed on the hexons in both new cryo-EM structures.
We expected to observe HD5 density at the critical site of the Ad5.F35+HD5 complex and not at the corresponding site of the Ad5.PB/GYAR+HD5 complex. After docking complete atomic models for the vertex regions into the two cryo-EM density maps (Fig. 2.2), we realized that the density at the predicted critical HD5 binding site was
35 weaker than expected for the defensin-sensitive Ad5.F35+HD5 complex. In the cryo-EM structure of Ad5.F35+HD5 there is density above the fiber EDES sequence and next to the penton base RGD loop but it only partially covers a docked HD5 monomer (Fig.
2.2C). Slightly less density was observed at this site in the resistant complex (Fig. 2.2D).
We speculate that significant disorder and flexibility at these sites would make it difficult to directly visualize HD5 density even in higher resolution structures. High resolution x- ray crystallographic and cryo-EM structures of adenovirus have large regions of the penton base RGD loop missing due to disorder (Liu et al, 2010); (Reddy et al, 2010) and these regions are adjacent to the critical site for HD5 neutralization proposed from the previous cryo-EM and viral chimera study (Smith et al, 2010a). In addition, there is a symmetry mismatch between the pentameric penton base and the trimeric fiber that also complicates reconstruction of density in this area.
MODELING OF HD5 MONOMERS AT THE HADV VERTEX
In our previous study we showed that HD5 binds to HAdV capsid determinants within the vertex region at the point of contact between fiber and penton base and that disruption of these sites leads not only to resistance, but also to enhanced infection (Smith et al, 2010a). To gain insight into the structural basis for HAdV susceptibility to HD5 antiviral activity and to better define the critical neutralization site, we built molecular models for the vertex regions of Ad5.F35+HD5 and
Ad5.PB/GYAR+HD5. These molecular models were refined into our cryo-EM density maps using cryo-EM guided molecular dynamics based flexible fitting with the MDFF software package (Trabuco et al, 2008). The MDFF method allows the application of an
36 external force based on the cryo-EM density map to guide atomic models into agreement with the density while at the same time preserving correct stereochemistry by use of a standard potential energy function during the molecular dynamics simulation. The hybrid cryo-EM/MDFF approach has been applied to several macromolecular complexes (Chan et al, 2012), including an engineered HAdV vector (Flatt et al, 2012) and HAdV in complex with coagulation factor X (FX) (Doronin et al, 2012). In the HAdV-FX study, the hybrid approach led to determination of the molecular interaction interface between hexon and FX.
Complete atomic models were built for the penton base/fiber complexes of the
Ad5.F35 and Ad5.PB/GYAR chimeras (Fig. 2.2A,B). The HAdV-5 penton base was constructed using the crystal structure of the HAdV-2 penton base as a template, which is
98% identical (Zubieta et al, 2005). The long flexible RGD-containing loops (77aa) absent in the HAdV-2 penton base crystal structure were added into the HAdV-5 penton base homology model using the Rosetta de novo structure prediction protocol (Rohl et al,
2004). Five different Rosetta generated loop models were docked into the five sites of the pentameric HAdV-5 penton base. A model for the HAdV-19c penton base was obtained using the automated ab initio I-TASSER protein structure prediction server (Zhang,
2008). Fibers for both chimeras were constructed using the HAdV-2 fiber crystal structure (van Raaij et al, 1999) as a template. The HAdV-2 fiber crystal structure contains atomic coordinates for the trimeric fiber knob and shaft domain, which includes a repeating sequence motif. Assembly of the complete vertex models was facilitated by the crystal structure of the HAdV-2 penton base in complex with the N-terminal portion
37 of the HAdV-2 fiber (Zubieta et al, 2005). The fiber peptide in this crystal structure contains a conserved motif that binds at the interface of adjacent penton base subunits.
The crystal structure of an HD5 monomer (Szyk et al, 2006) was positioned directly above the three occurrences of the fiber sequence (EDES or GYAR) that differs in the vertex models for the defensin-sensitive and defensin-resistant chimeras. Slightly more density was observed at this site in the cryo-EM structure of the defensin-sensitive chimera than in the defensin-resistant chimera (Fig. 2.2C,D). Presumably the HD5 density at this site is weak because of both structural heterogeneity in this region and because of smearing by the imposed icosahedral symmetry during calculation of the cryo-
EM structures. Five-fold symmetry was imposed on the vertex regions, which is correct for the pentameric penton base but incorrect for the trimeric fiber and possibly incorrect for HD5. Thus, we are using MDFF to restrain the majority of the HAdV vertex atoms within the cryo-EM density while relying more heavily on the potential energy function to position HD5 and the flexible penton base RGD-containing loops, which do not have strong cryo-EM density.
During MDFF simulations the HD5 monomers remained near the critical fiber sequence (18-EDES-21) in Ad5.F35 (Fig. 2.3). Strikingly however, HD5 monomers were repelled from the corresponding fiber sequence (18-GYAR-21) of Ad5.PB/GYAR (Fig.
2.3). Twenty-four different starting orientations of HD5 monomers were tested and in all cases the penton base RGD-containing loops of Ad5.F35 enveloped and interacted favorably with the HD5 monomers by the end of MDFF simulations. Significant movement of the Ad5.F35 RGD-containing loops was observed toward the HD5
38 monomers even during relatively short (100-picosecond) MDFF simulations (Fig. 2.3). In contrast, significantly less extensive and fewer favorable interactions were found between
HD5 and the RGD-containing loops of defensin-resistant Ad5.PB/GYAR (Fig. 2.3).
These observations correlate with the intermolecular nonbonded interactions calculated between the HD5 monomers and the fiber and penton base proteins at the end of the
MDFF simulations (Table 2.1). The intermolecular nonbonded energies reported include
Van der Waals and electrostatic interactions between separate polypeptide chains.
THE VERTEX OF THE DEFENSIN-SENSITIVE HADV ACCOMODATES HD5
DIMERS
In crystal structures α-defensins are found as dimers (Fig. 2.4A) (Hill et al, 1991;
Szyk et al, 2006). Recent studies indicate the importance of dimerization for viral neutralization activity, as an obligate monomer of HD5 was able to bind to the HAdV-5 capsid but was non-neutralizing (Gounder et al, 2012). Therefore, we repeated the molecular dynamics simulations with HD5 dimers modeled at the vertex sites within
Ad5.F35 and Ad5.PB/GYAR. In the case of Ad5.F35, the fiber/penton base interface easily accommodated HD5 dimers (Fig. 2.4B). In fact, the presence of HD5 dimers rather than monomers leads to even more favorable intermolecular nonbonded interactions between HD5 and the HAdV vertex proteins than observed for HD5 monomers (Table
2.2). In the case of Ad5.PB/GYAR, the short penton base RGD-containing loops are not able to envelop the HD5 dimers nearly as well as Ad5.F35 (Fig. 2.4C), and significantly less favorable interactions are modeled.
39
INTRINSIC DISORDER AT THE HD5 BINDING SITE
In total twenty-four different starting orientations of HD5 monomers and twenty- four orientations of HD5 dimers were simulated for the defensin-sensitive and defensin- resistant HAdV chimeras. In one simulation run, three different HD5 orientations were modeled with a complete vertex formed by a pentameric penton base and a trimeric fiber.
Since the RGD-containing loops of penton base are known to be highly flexible, and numerous feasible loop models were generated by Rosetta, we built a complete penton base with five different initial RGD-containing loop models. We reasoned that it would be more realistic to use five different loop models than to pick one loop model arbitrarily and use it at all five sites within the pentamer. Several factors led to our decision to perform MDFF simulations with HD5 in multiple starting orientations. These factors included the observation of weak and five-fold symmetrized defensin density in the vicinity of the critical binding site, which precluded initial docking of HD5 coordinates on the basis of HD5 shape within the cryo-EM density (Fig. 2.2C,D). In addition, the nature of HD5 with positive charge on multiple surfaces of the monomer and dimer made it possible to orient the peptide in multiple reasonable starting orientations with respect to the negatively-charged 18-EDES-21 sequence of the Ad5.F35 fiber.
One interesting finding that emerged from the MDFF simulations is that the critical binding site between the fiber N-terminal region and the penton base RGD- containing loops seems to be highly structurally malleable. This site in the defensin- sensitive HAdV chimera can form various favorable binding pockets for HD5. This led us to submit the HAdV5 and HAdV19c penton base protein sequences to the PrDOS
40
ProteinDisOrder System prediction webserver (Ishida & Kinoshita, 2007). The HAdV5 penton base is predicted to have a large disordered region (aa 297–373) (Fig. S2.2), which corresponds well to the flexible RGD loop region (aa 297–374) as defined by sequence alignment to the HAdV2 penton base crystal structure (Zubieta et al, 2005). The
HAdV19c penton base also is predicted to have a disordered region (aa 294–316), but it is significantly shorter (Fig. S2.2). The MDFF simulations show that the disordered regions of both penton bases interact with HD5. Regardless of the starting orientations for the HD5 monomer/dimer, or the conformation of the RGD-containing loops, after the simulation we observed favorable intermolecular nonbonded interactions between HD5 and penton base and also between HD5 and fiber. Simulations with the defensin-resistant
HAdV chimera showed that although favorable interactions could be found, they were less favorable overall than for the defensin-sensitive chimera and often only between
HD5 and either penton base or fiber but not both. This implies that although HD5 may bind to the vertex region of the defensin-resistant chimera it might not bridge the penton base and fiber and stabilize the vertex as it appears to do for the sensitive chimera.
The structural malleability of the binding pocket within the defensin-sensitive
HAdV chimera is illustrated in Figure 2.5 for one vertex region. Three HD5 dimers were started in different orientations and in close proximity to the N-terminal fiber residues 18-
EDES-21. At the end of the MDFF simulation all three HD5 dimers remained in close proximity to the EDES sequence, but they were oriented differently with respect to the fiber shaft and formed diverse interactions with loops of the fiber shaft (Fig. 2.5A). The intermolecular nonbonded energies reported in Table 2.3 indicate that each of these three
HD5 dimers formed favorable interactions with one or more of the fiber subunits at the
41 vertex. The 3-fold β-spiral repeat elements comprising the fiber shaft (van Raaij et al,
1999) facilitate the interaction with multiple HD5 dimers.
Similarly, the three HD5 dimers of this same vertex formed multiple and extensive interactions with the intrinsically disordered RGD-containing loops of the penton base (Fig. 2.5B). Although all of the loops moved toward the HD5 dimers and interacted favorably with HD5 by the end of the MDFF simulations, there did not appear to be a single preferred mode of interaction. Rather, the relatively long length of the loop
(77aa) and its conformational flexibility resulted in different, yet highly favorable, interactions with HD5. This was true regardless of the starting orientation of HD5 in all twenty-four simulations of dimers within defensin-sensitive binding pockets. As indicated by the intermolecular nonbonded energies for one vertex (Table 2.3), each loop interacts with more than one HD5 dimer and each HD5 dimer interacts favorably with one or two loops.
STABILIZATION OF THE DEFENSIN-SENSITIVE HADV VERTEX REGION BY
HD5
A striking result from the MDFF simulations is that HD5 dimers bound to the vertex region substantially stabilize the complex of penton base and fiber. If the working model for HAdV cell entry is correct, stabilizing the non-covalent association between fiber and penton base could block subsequent capsid uncoating steps that are necessary for escape of the virion from the endosome and a productive infection. The total nonbonded interaction energy calculated for three HD5 dimers at the vertex shown in Figure 2.5 corresponds to a stabilization of 1,992 kcal/mol. For comparison, the
42 average nonbonded energy calculated between fiber and penton base indicates a more modest favorable interaction of 670 kcal/mol. To help put these MDDF calculated potential energy values in perspective, we note that the stabilization achieved upon maturation of the bacteriophage HK97 capsid has been measured by differential scanning calorimetry to be on the order of 1,930 kcal/mol (Ross et al, 2005). This enormous effect includes both expansion and crosslinking of the capsid proteins.
The MDFF simulations for the vertex shown in Figure 2.5 indicate that three HD5 dimers can provide an increase in stabilization by a factor of 3. Calculations for seven other vertices, corresponding to twenty-one additional HD5 dimer orientations, show similar stabilization increases on the order of factors of 2 to 3. The MDFF simulations presented here suggest that HD5 dimers stabilize the HAdV vertex by bridging the fiber and penton base components at the critical fiber site. This stabilization effect could potentially lock the capsid in a structure that prevents release of the membrane lytic protein VI and therefore restricts viral escape from the endosome.
2.4 DISCUSSION:
The simulations presented here indicate that an intrinsically disordered loop of the adenovirus capsid can interact with HD5 favorably in multiple different ways. Given the strong nonbonded interaction energies calculated for the Ad5.F35/HD5 interaction, it is possible that once HD5 is bound at the vertex sites it is not released. In fact, it has been noted that a protein with intrinsic disorder can bind permanently with an interaction partner (Dunker et al, 2001). If HD5 bound at the vertex sites of sensitive HAdV types can permanently lock the vertex region, this would presumably block release of the
43 membrane lytic factor in the endosome and consequently block cell entry. Consistent with this, recent studies have shown that HD5 selectively increases the tensile strength of the capsid vertex region (Snijder et al, 2013). We built a space filling representation of one Ad5.F35 vertex with three bound HD5 dimers (Fig. S2.3) and we find that several of the RGD sites are still accessible on the surface. This is consistent with the experimental finding that the presence of HD5 does not block virus internalization (Smith & Nemerow,
2008).
This study presents detailed models for the interaction of a human α-defensin,
HD5, with two HAdV chimeras, one of which is sensitive to defensin antiviral activity and one of which is resistant. Previously, a charged stretch of four residues within the N- terminal region of fiber was implicated as playing a role in the critical binding site for
HD5 (Smith et al, 2010a). This site is at a position of symmetry mismatch within the
HAdV capsid, where the trimeric fiber interacts with the pentameric penton base, and is at least partially flexible (Zubieta et al, 2005). In addition, this site is next to the flexible
RGD-containing loop of the penton base. These factors made it difficult to directly resolve density for HD5 in the current structures, although we noted slightly more density attributable to HD5 in this site for the sensitive chimera than in the resistant chimera. In order to overcome this limitation we used a hybrid approach including MDFF to model the interaction of HD5 with both HAdV capsids. In this case we are relying heavily on the MDFF potential energy functions to model the interactions between HD5 and the intrinsically disordered region of penton base, as strong cryo-EM density is not observed for these regions. The simulations reveal that HD5 bridges the penton base and fiber and thus stabilizes the vertex complex formed by these two proteins within the sensitive
44 chimera. The bridging interactions involve the negatively charged EDES sequence in the
N-terminal fiber region, which is of opposite charge in the resistant chimera, and multiple and varying residues of the penton base RGD loop.
The MDFF simulations show the critical binding pocket to have a large degree of malleability, as expected for an intrinsically disordered region. The RGD-containing penton base loops of the sensitive HAdV chimera are observed in the simulations to envelop HD5 monomers and dimers at the top of the penton base near the critical fiber sequence. We speculate that defensin neutralization of HAdV may be taking advantage of the inherent flexibility of the RGD-containing loops. Flexibility of these loops is thought to be important for their interaction with αv integrins during cell entry (Lindert et al,
2009). The simulations also indicate multiple favorable orientations for HD5 dimers at the interface between penton base and fiber. This variability within the atomic models for the HD5 interaction with the HAdV vertex is consistent with difficulty we had in observing density for HD5 at the vertex in the cryo-EM structure of the defensin- sensitive chimera.
A recent study investigating the critical determinants of HD5 activity against
HAdV found that HD5-mediated neutralization depends upon specific binding interactions to the viral capsid mediated in part by critical arginine residues, R9 and R28, in HD5 (Gounder et al, 2012). Visual inspection of the MDFF results for HD5 dimers bound to HAdV vertex proteins indicates that the critical arginines contribute to the calculated intermolecular nonbonded energies. Gounder et al. (Gounder et al, 2012) also found that stabilization of the HD5 dimer is critical for neutralization of HAdV. The
45
MDFF-based finding of more favorable interaction energies for HD5 dimers than for
HD5 monomers with the defensin-sensitive HAdV chimera is consistent with this experimental result. We speculate that higher order multimerization of HD5 above the vertex region would provide further stabilization. The malleability for the critical binding pocket and the long, flexible RGD-containing penton base loops of the sensitive HAdV chimera should allow for various types of HD5 associations with the penton base/fiber complex. It is likely that additional HD5 interactions at the vertex would stabilize the protein complex and would impede its timely dissociation in the endosome.
The average nonbonded interaction energy calculated at the end of the MDFF simulations for one HD5 dimer with a penton base/fiber complex indicates a stabilization of 582 kcal/mol. This is on the order of the nonbonded interaction energy calculated between penton base and fiber at one vertex, which indicates a similar stabilization of
670 kcal/mol. Therefore, if three HD5 dimers bind at the same vertex, the overall increase in stabilization of the vertex is on the order of a factor of three. Calculation of a significant stabilization effect on the vertex by HD5 is consistent with the experimental results from a thermostability assay showing that HD5 stabilizes the defensin-sensitive
HAdV5 capsid (Smith & Nemerow, 2008). In the presence of HD5, the release of fiber and membrane lytic protein VI was shifted from ~49°C to 61° and 67°C respectively.
This thermostability assay is designed to mimic virus disassembly in the endosome (Wiethoff et al, 2005). Our working model for HD5 neutralization of HAdV is that binding of defensin at critical sites on the HAdV capsid could prevent dissociation of fiber from penton base, which is thought to be an early step in HAdV disassembly during cell entry. Stabilization of the association of fiber with the penton base could in turn
46 block subsequent steps in uncoating that lead to the release of protein VI and disruption of the endosomal membrane (Smith et al, 2010a). The ability of HD5 to block viral uncoating during cell entry has been demonstrated (Nguyen et al, 2010).
The cryo-EM structures presented here were determined with 5 µM HD5, which is close to the IC50 for inhibition of HAdV-5 infection of 3–4 µM (Smith & Nemerow,
2008). The concentration of HD5 within small localized areas of the intestine has been estimated as 100 to 1000 times higher (Ganz, 2003). If the model for defensin neutralization of HAdV is correct, then we assume that HD5 must bind and stabilize all twelve vertices of the HAdV capsid in order to completely block infectivity. Given the variability we observed in the MDFF simulations for the interaction of HD5 with the
HAdV vertex proteins, it is possible that at low µM concentrations of HD5 most but not all of the vertices are highly stabilized by defensin. We speculate that if a few vertices are not stabilized by HD5, this could allow partial disassembly of the capsid and release of protein VI in the endosome. This would be consistent with measured low levels of infectivity for multiple sensitive HAdV types in the presence of 15 µM HD5 (Smith et al,
2010a).
In summary, our goal was to gain more precise structural information of the interaction of HD5 with a defensin-sensitive and a resistant HAdV chimera. Due to the intrinsic disorder of the RGD-containing penton base loops of the sensitive chimera we relied on molecular dynamics simulations to enhance the model of defensin interaction.
The simulations indicate that HD5 dimers can stabilize the interaction between penton base and fiber, but only in the context of the sensitive chimera. A high degree of
47 conformational flexibility was found for the critical binding pocket resulting in a variety of stabilizing interactions between defensin and the capsid vertex proteins. Based on our structural analysis and modeling studies, it is likely that intrinsic disorder is an important aspect of the binding pocket that contributes to the susceptibility of a particular HAdV types to defensin neutralization.
2.5 MATERIAL AND METHODS:
CRYO-EM AND IMAGE PROCESSING
Samples of two HAdV chimeras, Ad5.F35 and Ad5.PB/GYAR, were prepared as previously described (Smith et al, 2010a). Purified virus (160 µg/ml) was combined with
HD5 (5 µM) and incubated on ice for 45 minutes. These samples were then applied to
Quantifoil grids and rapidly frozen in liquid ethane using a homebuilt vitrification device.
For high resolution cryo-EM data acquisition, electron micrographs were collected on an
FEI Polara (300 kV, FEG) operated at liquid nitrogen temperature. A Gatan UltraScan
4000 CCD camera was used for recording images at an absolute magnification of
397,878×. The underfocus values of the micrographs ranged from 0.5 µm to 4 µm.
Datasets were collected for HD5 complexed with Ad5.F35 and Ad5.PB/GYAR that included 3,515 and 3,620 particle images, respectively. Individual particles were manually selected and centered using in-house scripts and stacks were generated at various pixel sizes (4.5 Å, 2.2 Å, and 1.5 Å) suitable for single particle image processing.
Particle images with a coarser image size were used at the beginning of refinement.
Initial defocus and astigmatism estimates were determined using the program CTFFIND3
(Mindell & Grigorieff, 2003). A cryo-EM structure of Ad5.F35 (Saban et al, 2006) was
48 used as the starting model for FREALIGN refinement (Grigorieff, 2007). Intermediate refinement rounds were performed using particle images with a 2.2 Å pixel size and the final refinement rounds with a particle image pixel size of 1.5 Å. For both datasets, the orientational and CTF parameters, as well as absolute magnification were refined using
FREALIGN. The resolutions of the final cryo-EM structures of the Ad5.F35/HD5 complex and Ad5.PB/GYAR/HD5 complex were at 9.7 Å and 8.1 Å, respectively, as measured at the Fourier Shell Correlation 0.5 threshold (Fig. S2.1).
ATOMIC MODEL BUILDING
We constructed atomic models for the vertex regions of Ad5.F35 and
Ad5.PB/GYAR, which include the capsid proteins penton base and fiber. Homology models of HAdV-5 and HAdV-19c penton base proteins were generated using the
HAdV-2 penton base crystal structure (Zubieta et al, 2005). For Ad5.F35, the HAdV-2 penton base crystal structure (98% identity) was mutated to match the HAdV-5 penton base sequence using the amino acid substitution command available within the UCSF
Chimera software package (Pettersen et al, 2004). Additionally, the long flexible RGD containing loops (77aa) that are absent in the crystal structure were incorporated into the
HAdV-5 penton base homology model using the Rosetta de novo structure prediction protocol (Rohl et al, 2004). The five RGD containing loops that were attached to the penton base model have different conformations. A homology model of the Ad19c penton base was created using the I-TASSER server for protein 3D structure prediction (Roy et al, 2010; Zhang, 2008). Atomic models were created for the short- shafted fiber present in Ad5.F35 and the longer HAdV-5 fiber present in Ad5.PBGYAR
49 based on the Ad2 fiber crystal structure (van Raaij et al, 1999). The HAdV-2 fiber crystal structure contains atomic coordinates for the trimeric fiber knob and shaft domain, which includes a repeating sequence motif. Attachment of the fibers to penton base proteins was guided by the crystal structure of HAdV-2 penton base in complex with a portion of the fiber (Zubieta et al, 2005). The HAdV-2 N-terminal fiber peptide includes is a highly conserved fiber motif that binds at the interface of adjacent penton base subunits.
MOLECULAR DYNAMICS AND FLEXIBLE FITTING
MDFF is a cryo-EM guided molecular dynamics method that will flexibly fit atomic models into cryo-EM density maps (Trabuco et al, 2008; Trabuco et al, 2009). In
MDFF simulations, forces proportional to the gradient of the density map are applied to all atoms, driving the models to occupy regions of high density. Structural restraints are applied to preserve secondary structural elements so that over-fitting does not occur during the simulation. The simulations were carried out in such a way that only secondary structural elements were guided into density with a g-scale value of 0.3. We purposely kept the g-scale value rather low so that the coordinates for HD5 and the flexible penton base loops would not be forced into cryo-EM density. All MDFF simulations were carried out using NAMD 2.8 (Phillips et al, 2005) and the CHARMM27 force field. Cryo-EM density of the vertex regions was extracted from both the
Ad5.F35+HD5 (defensin-sensitive) and Ad5.PB/GYAR+HD5 (defensin-resistant) maps and vertex models were docked into the selected densities using rigid body fitting in
UCSF Chimera (Pettersen et al, 2004). The docked coordinates served as input for 100- picosecond MDFF implicit solvent simulations. Monomeric and dimeric forms of HD5
50 were tested with both the defensin-sensitive and defensin-resistant vertex regions.
Nonbonded energies were measured between specific polypeptide chains using the
NAMD Energy plugin in VMD (Humphrey et al, 1996). The MDFF simulations were performed on the Case Western Reserve University High-Performance Computing
Cluster.
2.6 ACKNOWLEDGMENTS:
We would like to thank Dewight Williams for EM support. We also thank the
High Performance Computing Cluster (HPCC) at Case Western Reserve University for computational support. The cryo-EM structures have been deposited in the EM Data
Bank with accession numbers EMD-5538 (sensitive) and EMD-5539 (resistant).
2.7 AUTHOR CONTRIBUTIONS:
Conceived and designed the experiments: JGS GRN PLS. Performed the experiments: JWF RK. Analyzed the data: JWF PLS RK. Contributed reagents/materials/analysis tools: JGS GRN. Wrote the paper: JWF JGS GRN PLS.
51
Table 2.1 Intermolecular nonbonded energies for HD5 monomers with adenovirus vertex proteins.
52
Table 2.2 Intermolecular nonbonded energies for HD5 dimers with adenovirus vertex proteins.
53
Table 2.3 Intermolecular nonbonded energies for three HD5 dimers with each of the subunits of fiber and penton base at one defensin-sensitive adenovirus vertex.a
54
Figure 2.1 Cryo-EM structures of HD5 bound to neutralization-sensitive (Ad5.F35) and - resistant (Ad5.PB/GYAR) chimeric HAdVs. (A,D) Reconstructions viewed along icosahedral 2-fold axes and shown radially color-coded in red for the sensitive
HAdV+HD5 complex and blue for the resistant HAdV+HD5 complex. (B,E) Enlarged views of the penton base and fiber of both HAdV+HD5 complexes. Only a portion of the
Ad5.PB/GYAR fiber is reconstructed due to length (>300 Å). (C,F) Density rods are observed for penton base α-helices within both HAdV+HD5 complexes. Atomic models
(black) for the HAdV-5 penton base in Ad5.F35 and the HAdV-19c penton base in
Ad5.PB/GYAR are shown docked within the cryo-EM density. The isosurface threshold level for the density is chosen to highlight the density rods. The enlarged insets show one density rod aligned with one α-helix confirming the subnanometer (<10 Å) resolution of the structures. Scale bars, 100 Å.
55
Figure 2.2 Modeling and cryo-EM guided molecular dynamics simulations of the interaction between HD5 monomers and vertex proteins of the defensin-sensitive
(Ad5.F35) and defensin-resistant (Ad5.PB/GYAR) HAdV chimeras. (A) Atomic model of the penton base (brown with RGD loops in yellow) and fiber (green) of Ad5.F35 with docked HD5 monomers (red). (B) Similar atomic model of Ad5.PB/GYAR. The smaller model representations in panels A and B show the full length fibers. (C,D) Atomic models of the vertex regions with HD5 monomers shown docked within the cryo-EM density.
56
Figure 2.3 Movement of the RGD-containing loop and HD5 during the molecular dynamics simulations of the defensin-sensitive (Ad5.F35) and defensin-resistant
(Ad5.PB/GYAR) HAdV chimeras. (A,C) Initial and final MDFF coordinates for one simulation of the Ad5.F35+HD5 interaction. The final coordinates show the HD5 peptide in close proximity to the fiber sequence 18-EDES-21 (spheres) and enveloped by the
RGD-containing loop of the penton base. The bars indicate the extent of the movement of the RGD-containing loop toward HD5 during the simulation. (B,D) Initial and final
57
MDFF coordinates for one simulation of the Ad5.PB/GYAR+HD5 interaction. The bars indicate the extent of the movement of HD5 away from the fiber sequence 18-GYAR-21
(spheres) during the simulation.
58
Figure 2.4 Modeling and cryo-EM guided molecular dynamics simulations of the interaction between HD5 dimers and complete vertex regions of the defensin-sensitive
(Ad5.F35) and defensin-resistant (Ad5.PB/GYAR) HAdV chimeras. (A) Comparison of
59 the HD5 monomer (left) and HD5 dimer (right) structures (PDB-ID 1ZMP). (B) Final
MDFF coordinates for one vertex simulation of the Ad5.F35+HD5 interaction shown in top and side views. The three HD5 dimers are each in close proximity to the fiber sequence 18-EDES-21 (spheres) and flanked by the RGD-containing loops of the penton base. (C) Final MDFF coordinates for one vertex simulation of the Ad5.PB/GYAR+HD5 interaction. The HD5 dimers are less closely associated with the fiber sequence 18-
GYAR-21 (spheres) and less well enveloped by the RGD-containing loops compared to the defensin-sensitive vertex. The coloring scheme is as in Figure 2.2.
60
Figure 2.5 Structural malleability of the binding pocket within the defensin-sensitive
HAdV chimera (Ad5.F35). (A) Comparison of three different HD5 dimer orientations each with respect to one fiber subunit containing the closest 18-EDES-21 sequence
(spheres). (B) Comparison of penton base RGD-containing loop interactions with the same three HD5 dimers. The penton base loops are shown without the rest of the penton base. The models in panels A and B are all based on final MDFF coordinates from one vertex with a trimeric fiber and a pentameric penton base. The five penton base loops are shown in purple, cyan, blue, gold, and gray and the rest of the coloring scheme is as in Figure 2.2.
61
Figure S2.1 Subnanometer resolution of cryo-EM structures of HD5 bound to neutralization-sensitive (Ad5.F35) and -resistant (Ad5.PB/GYAR) chimeric
HAdVs. (A,C) Density rods are observed for hexon α-helices within both HAdV+HD5 complexes. Atomic models (black) for the HAdV-5 hexon in Ad5.F35 and the HAdV-19c penton base in Ad5.PB/GYAR are shown docked within the cryo-EM density. The isosurface threshold level for the density is chosen to highlight the density rods. (B,D)
Fourier shell correlation plots indicating 9.7 Å resolution for Ad5.F35+HD5 and 8.1 Å resolution for Ad5.PB/GYAR+HD5 at the FSC 0.5 thresholds.
62
Figure S2.2 Prediction of intrinsically disordered regions within the HAdV5 and
HAdV19c penton base proteins by the PrDOS webserver (Ishida & Kinoshita, 2007). (A)
Prediction for the HAdV5 penton base of the Ad5.F35 virus chimera. A long intrinsically
63 disordered region is found between residues 297 and 373. This corresponds well to the
RGD loop (aa297–374) as assigned based on sequence alignment to the flexible residues in the HAdV2 penton base crystal structure (Zubieta et al, 2005). (B) Prediction for the
HAdV19c penton base of the Ad5.PB/GYAR virus chimera. A short intrinsically disordered region is found between residues 294 and 316, overlapping with the RGD loop
(aa290–318). Residues above the 0.5 threshold line in these plots are predicted to be disordered.
64
Figure S2.3 Space filling representation of the vertex region of Ad5.F35 with three bound
HD5 dimers. The penton base is shown mostly in brown with the RGD loops in yellow and the RGD residues in blue. The fiber is shown in green and HD5 in red. This
65 representation was generated with the UCSF Chimera molmap command with a 5 Å resolution filter applied to the final MDFF coordinates for one vertex.
66
CHAPTER 3: Virus-misplaced humoral factor activates innate immunity.
This chapter is published: Science doi:10.1126/science.1226625
Research Article:
Coagulation factor X activates innate immunity to human species C adenovirus.
* * * Konstantin Doronin , Justin W. Flatt , Nelson C. Di Paolo , Oleksandr Kalyuzhniy,
James W. MacDonald, Theo K. Bammler, Richard P. Beyer, Frederico M. Farin, Phoebe
L. Stewart, and Dmitry M. Shayakhmetov.
*Contributed equally
3.1 ABSTRACT
Self-nonself discrimination and adequate response to infection and tissue damage are fundamental functions of the immune system. The rapid and global spread of known and emerging viruses is a testament that the timely detection of viral pathogens that reproduce within host cells, presents a formidable challenge to the immune system. To gain access to a proper reproductive niche, many pathogens travel via the host vasculature and become exposed to humoral factors of the innate immune system. Although a cascade of coagulation factors plays a fundamental role in host defense for “living fossils” such as horseshoe crabs (Xiphosurida spp)
(Kairies et al, 2001; Young et al, 1972), the role of the coagulation system in the activation of innate responses to pathogens in higher organisms remains unclear. When human type C
67 adenovirus (HAdv) enters the circulation, 240 copies of coagulation factor X (FX) bind to the virus particle with picomolar affinity (Kalyuzhniy et al, 2008; Waddington et al, 2008). Here, using molecular dynamics flexible fitting (MDFF) and cryo-electron microscopy (cryo-EM), we modeled the interface between the HAdv5 hexon protein and FX. Based on this structural data, we introduced a single amino acid substitution, T425A, in the hexon that completely abrogated
FX interaction with the virus. In vivo genome-wide transcriptional profiling revealed that FX- binding-ablated virus failed to activate a distinct network of the early response genes, whose expression depends on transcription factor NFKB1. Deconvolution of the signaling network responsible for early gene activation showed that the FX-HAdv complex triggers
MyD88/TRIF/TRAF6 signaling upon activation of toll-like receptor 4 (TLR4) in vivo. Our study implicates host factor “decoration” of the virus as a mechanism to trigger an innate immune sensor that responds to a misplacement of coagulation FX from the blood into intracellular macrophage compartments upon virus entry into the cell. Our results further the mounting evidence of evolutionary conservation between the coagulation system and innate immunity.
3.2 REPORT
Upon infection with microbial and viral pathogens, specialized innate immune sensors recognize unique pathogen-associated chemical moieties (pathogen-associated molecular patterns or PAMPs) (Janeway & Medzhitov 2002). This triggers the activation of effector mechanisms aimed at restricting spread and facilitating pathogen elimination from the host. In the context of sterile inflammation, the innate immune receptors can also recognize host moieties released from damaged cells as an intrinsic danger signal
(danger-associated molecular patterns or DAMPs) (Matzinger 2002) to initiate tissue
68 repair. Often, innate immune recognition of pathogens leads to a severe inflammatory host response that may be solely responsible for the morbidity and mortality associated with the infection (Kash et al, 2006; Imai et al, 2005; Chen & Subbarao, 2007; Wen et al,
2010).
HAdv induces potent innate immune and inflammatory responses (Thomas et al,
2000). For immunocompromized individuals, HAdv infections can be lethal (Ardehali et al, 2001; Leen & Rooney, 2005; Kim et al, 2007). Disseminated HAdv infections are frequently associated with liver and kidney failure and a high virus burden in the blood
(Lynch et al, 2011) (Supplemental note 1). When HAdv-based vectors are injected intravenously in preclinical and clinical gene therapy trials, they induce an acute inflammatory response that may lead to morbidity and mortality (Morral et al, 2002;
Raper et al, 2003). The molecular basis for innate immune activation by HAdv remains poorly characterized. While in the blood, type C HAdv2 and HAdv5 bind coagulation factor X with high affinity (Kalyuzhniy et al, 2008; Waddington et al, 2008). Because coagulation factor activation may trigger systemic inflammation, we hypothesized that
FX binding to HAdv may trigger virus recognition by the innate immune system and activation of an antiviral inflammatory response. As a first step in testing this hypothesis, we analyzed the interaction interface between FX and HAdv by determining a high resolution cryo-EM structure followed by MDFF simulations (Trabuco et al, 2008).
69
CRYO-EM AND MDFF ANALYSIS OF THE PROTEIN-PROTEIN INTERFACE
BETWEEN FX AND HADV5
In the cryo-EM structure, FX density protrudes from the top of each hexon in the
HAdv5 capsid (Fig. 3.1a-c and Movie S1). FX interacts with the capsid by attaching to three-fold symmetric central depressions in each hexon trimer as seen in earlier moderate resolution cryo-EM structures (Kalyuzhniy et al, 2008; Waddington et al, 2008). Atomic models for hexon and FX were docked into cryo-EM density (9Å resolution for the capsid and 11Å resolution for the capsid plus FX). MDFF simulations revealed an electrostatic interaction between FX-GLA domain residue K10 and HAdv5 hexon residues E424 and T425 (Fig. 3.1d-f, Table S3.1, and Movie S2). To test the robustness of the MDFF simulated protein-protein interface, we performed a series of simulations with the FX starting orientation rotated by +/-10, or +/-20° about the 3-fold molecular axis of hexon (Figs. S3.1-S3.2 and Table S3.1). Despite these adjustments, the FX-
GLA/HAdv5 hexon interface snaps back to form the same packing arrangement.
VALIDATION OF THE FX-HADV5 MODEL
To test whether glutamic acid E424 and threonine T425 are essential for formation of the FX-HAdv5 complex, we introduced single amino acid substitutions in this region of the hexon by substituting the wild type amino acid sequence TET
(T423E424T425) for GAT, TAT, or TEA (Fig. 3.2a). The viruses with these mutations were constructed and analyzed for their capacity to directly bind FX using surface plasmon resonance (Kalyuzhniy et al, 2008). This analysis revealed that the affinity of
GAT and TAT mutants for FX remained in low nanomolar range (Fig. 3.2b). In contrast,
70 a single amino acid substitution T425A completely abrogated FX binding to the virus
(Fig. 3.2b). The in vitro infection of CHO-K1 cells (which can only be infected in the presence of FX) showed that, unlike all other viruses, the Ad-TEA mutant failed to infect these cells in the presence of FX (Fig. 3.2c). However, all the vectors analyzed, including
Ad-TEA, infected susceptible lung carcinoma A549 cells with equal efficacy (Fig. S3.3).
As FX-binding-ablation abrogates HAdv5 infection of hepatocytes in vivo (Alba et al,
2009: Kalyuzhniy et al, 2008; Waddington et al, 2008) and Supplemental note 2), we injected mice intravenously with virus mutants expressing GFP and analyzed GFP expression 24 h later. This analysis confirmed that all the mutants, except for Ad-TEA, transduced hepatocytes efficiently (Fig. 3.2d-f). Hepatocyte transduction by FX binding- ablated virus was significantly reduced. In vivo analysis of virus interaction with tissue macrophages showed that macrophages both in the liver and spleen trapped all of the viruses with comparable efficacy (Fig 3.2g-h and Fig. S3.4-S3.5). As tissue macrophages are the principal inducers of inflammation in response to HAdv (Zhang et al, 2001; Paolo et al, 2009), we next analyzed whether in vivo macrophage sensors stimulated aberrant innate immune response to Ad-TEA mutant virus, which is stripped of FX, compared to control HAdv. The analysis of transcripts for >28,000 genes revealed that expression of
519 genes was changed more than 1.5-fold (P < 0.001) in the spleen of WT mice 30 min after HAdv injection. Gene ontology analysis with CateGOrizer tool (see Methods section) showed that over 60% of genes transcriptionally upregulated in response to
HAdv were divided between four categories: stress response (23.88%), metabolism
(20.90%), death pathways (11.94%), and apoptosis (8.96%) (Fig. 3.3a). Because our earlier study demonstrated that signaling downstream of direct HAdv sensors occurs
71 independently of IL-1RI, and IL-1RI triggers feed-forward amplification loops(21), to define the signaling pathway(s) specifically activated by HAdv sensors, we identified
-/- only those genes that were co-activated after HAdv injection in both WT and Il1r1 mice.
We found that after applying a very strict cut off criteria (P < 0.0002), there were 34
-/- genes co-activated after HAdv injection in both WT and Il1r1 mice (Table S3.2). We
-/- next injected WT and Il1r1 mice with Ad-TEA mutant virus and compared the transcriptional signatures induced by HAdv and Ad-TEA on this 34 co-activated gene set.
This analysis showed that there were numerous genes differentially activated by HAdv and FX binding-ablated Ad-TEA mutant (Fig. 3.3b and Table S3.3) and for Nr4a2 (P =
9.61E-07), ATF3 (P = 6.77E-06), Fosb, Fosl2, and Ptgs2 genes, differential activation ranged between 2.2- and 5.8-fold. The P-scan analysis showed that there is enrichment for CREB1, NFKB1 and SRF transcription factor binding sites within the promoters of a
34 gene set (Fig. S3.6). However, only NFKB1 transcription factor binding sites were overrepresented in proximal promoters of the gene set differentially activated by HAdv and Ad-TEA (Fig. 3.3c). Using Ingenuity Pathway Analysis software, we further confirmed that NFKB1, CREB1, and SRF are at the center of a signaling network that responds to HAdv entry into macrophages in vivo (Figs. S3.7, S3.8).
FX DECORATED VIRUS ACTIVATES AN INNATE IMMUNE RESPONSE
Because NFKB1 is a transcription factor that activates numerous early response genes, including genes encoding for inflammatory cytokines and chemokines, we next analyzed whether HAdv and FX binding-ablated Ad-TEA viruses trigger differential inflammatory responses at the protein level. This analysis revealed that Ad-TEA failed to
72 activate a specific set of cytokines and chemokines, including IL-1β, IL-6, and MIP-1α, whose expression is known to be dependent on NF-κB (Fig. 3.4a). Using an RNAse protection assay and the IL-1β gene as a prototypic marker of NF-κB-dependent genes, we further confirmed that Ad-TEA virus failed to activate IL-1β transcription in vivo in
WT mice (Figs. 3.4b and S3.9). Deconvolution of the signaling pathway that leads to NF-
κB activation in gene-deficient mice revealed that IL-1β transcriptional activation required TRAF6, MyD88, and TRIF (Fig. 3.4c and S3.9). Since MyD88 and TRIF are the signaling adapters for the TLR family of receptors, we analyzed IL-1β expression in response to HAdv in mice deficient in specific TLRs. This analysis showed that
-/- transcriptional activation of IL-1β was deficient only in Tlr4 mice and occurred at a wild type level in mice deficient in TLR2, TLR7/8, or TLR9 (Figs. 3.4d and S3.9). The
-/- -/- protein level analysis of inflammatory cytokines and chemokines in Tlr4 , Myd88 , and
-/- Ticam mice confirmed reduced levels of cytokines and chemokines that depend on NF-
κB (Fig. 3.4e). Collectively, these data demonstrate that a full-scale inflammatory response to HAdv requires TLR4/TRAF6/NF-κB signaling and FX binding ablated virus
-/- failed to activate this pathway in vivo. In vitro infection of primary mouse WT and Tlr4 macrophages with wild type HAdv5 confirmed the dependence of a full scale cytokine and chemokine activation on presence of FX (Figs. S3.10). We next analyzed activation of a panel of cytokines and chemokines in response to a number of wild type HAdv species and found that those viruses that bind FX activated a broader spectrum of cytokines and chemokines, compared to wild type HAdv species that do not bind
FX(Kalyuzhniy et al, 2008) (Fig. 3.4f).
73
CONCLUSION
Deconvolution of pathogen sensing mechanisms and signaling pathways that activate innate immunity is fundamental for the development of drugs that target these pathways to moderate exacerbated inflammation. However, upon natural infection, a prototypical pathogen, such as a gram negative bacterium, presents a multitude of
PAMPs to cellular sensors. These PAMPs include TLR activating lipopolysaccharides
(Poltorak et al, 1998), flagellin and components of the secretion apparatus that activate the NLRC4/Caspase-1 inflammasome (Miao et al, 2006), bacterial RNA that activates the
NLRP3/Caspase-1 inflammasome (Kanneganti et al, 2006), and cyclic di-nucleotides that activate the IFN-I signaling pathway (Burdette et al, 2011), among others. This complex interaction of bacteria with the host makes mechanistic dissection of critical events that drive the outcome of natural infection a daunting task. Using a rather simple viral pathogen that assembles in mammalian cells and presents only a limited number of
PAMPs, we defined a novel mechanism of innate immune activation by the misplacement of a host humoral factor from the blood. We showed that a non- inflammatory humoral factor of the coagulation cascade binds to the surface of the virus and becomes a pathogen-associated molecular pattern that, upon viral entry into the cell, triggers activation of innate immunity via the Toll-/NF-κB pathway (Fig. 3.4g). Direct binding of coagulation factors FVII and FX was recently reported for human herpes virus
HSV-1 (Livingston et al, 2006). The data presented here provide the first evidence for an evolutionary conserved link between the coagulation system and innate immunity in higher organisms, where the coagulation system functions in facilitating direct recognition of a pathogen and activating innate immune defenses.
74
3.3 MATERIAL AND METHODS:
ANIMAL STUDIES
All animal studies were carried out with the approval of the Institutional Animal
Care and Use Committee of the University of Washington, Seattle, WA. C57BL/6 mice were purchased from Charles River, Wilmington, MA. Il1r1-/- mice and mice deficient in
TLR2 and TRIF were purchased from Jackson Laboratory. All mice were on C57BL/6 genetic background, matched by age and housed in specific-pathogen free facilities.
C57BL/6, Il1r1-/-, Myd88-/-, Ticam1-/-, Traf6/-, Tlr2-/-, Tlr4-/-, Tlr7-/-/8-/-, Tlr9-/- and
Md2-/- mice were challenged with wild type HAdV5 or HAdV5-based vectors at a dose of 5x1011 virus particles kg-1 via the tail vein infusion. At indicated times mice, were sacrificed and tissues were harvested for further analyses.
VIRUSES
Wild-type human adenovirus species HAdV2, HAdV4, HAdV5, HAdV16,
HAdV21, and HAdV51 were purchased from ATCC (Manassas, VI). Viral particle titers were determined by OD260 measurement. Each produced virus stock was tested for endotoxin contamination using Limulus amebocyte lysate Pyrotell (Cape Cod Inc,
Falmouth, MA). For in vivo experiments, only virus preparations confirmed to be free of endotoxin contamination were used.
75
ELECTRON MICROSCOPY, IMAGE PROCESSING AND MODELING
Digital micrographs were recorded under low-dose conditions on a Polara 300kV
FEG transmission cryo-electron microscope equipped with a Gatan Ultrascan 4kx4k charged coupled device camera. A total dataset of 1,101 particle images were collected at a nominal magnification of 310,000X. Image processing was performed as described earlier (Lindert et al, 2009) except that an in-house script was used for particle selection.
After the final round of refinement, the resolution of the icosahedral capsid was calculated at the Fourier Shell Correlation (FSC) 0.5 level to be 9.1Å. When the outer radius for resolution assessment was extended to 600Å to include FX and fiber density, the resolution was calculated to be 11Å (FSC 0.5). Atomic resolution information is available for the hexons in the HAdV capsid (Liu et al, 2010; Reddy et al, 2010) and a homology model for the zymogenic form of FX was supplied by Dr. Lee G. Pedersen.
After selecting a region of well-resolved FX density in the cryo-EM structure of the FX-
HAdV5 complex and filtering it to 14Å resolution, we used rigid body docking methods
(UCSF Chimera) to fit atomic coordinates for the hexon trimer and the FX model into the cryo-EM density. These docked coordinates served as the input for 100ps MDFF simulations with implicit solvent. The r.m.s.d between the initial and final FX model after flexible fitting is 2.8Å as calculated for the backbone atoms with the RMSD Calculator in
VMD. When individual FX domains are analyzed separately, the r.m.s.d.’s are 1.5 1.7,
1.9 and 2.6 Å, for the GLA, EGF-1, EGF-2 and serine protease domains respectively.
76
PROTEOME PROFILER ANTIBODY ARRAYS
A “Proteome Profiler antibody array: Mouse Cytokine Array Panel A (#ARY006) was from R&D System and were used, according to the manufacturer’s instructions. Each spleen was homogenized in 2 ml of sample solution, and 1 ml (1/2 spleen) was used to incubate with each membrane on a rocking platform overnight. Membranes were developed with ImmunoStar HRP-sustrate (BioRad, #1705041). Primary mouse bone marrow-derived macrophages were differentiated from total bone marrow cells with M-
CSF as described elsewhere.
ANTIBODIES FOR CONFOCAL MICROSCOPY
Antibodies were purchased from Abcam: biotinylated anti-Ad-Hexon (#ab34374, final dilution 1/100), anti-Ad5 (#ab6982, final dilution 1/50). Antibodies from BMA: anti-Marco (BMA, #T2026, 2 ug/ml), anti-Moma-1 (or CD-169) (BMA, #T2011, 2 ug/ml). Secondary antibodies and reagents were from Jackson Immunoresearch: Cy2 or
Cy3-labeled streptavidin, or donkey anti-rat or rabbit antibodies, Cy2-, Cy3- or HRP- labeled.
SURFACE PLASMON RESONANCE ANALYSES
All analyses were carried out on BIACORE 2000 machine. Research grade CM5 sensor chips, along with the following N-hydroxysuccinimide (NHS), Nethyl-N’-(3- diethylaminopropyl) carbodiimide (EDC), ethanolaminehydrochloride, HBSP running buffer and HBSEP regeneration buffer were purchased from the manufacturer
(BIACORE Inc., Piscataway, NJ). All data were collected at 1Hz using two or three
77 replicate injections for each concentration of analyte. Data processing and kinetic analysis were performed using Scrubber software (version 2.0, BioLogic Software,
Campbell, Australia). Data processing included double referencing. Processed data were globally fit to a simple 1:1 interaction model. Data analysis with conformational change model has been done using CLAMP software.
MICROARRAY SAMPLE PROCESSING
Integrity of RNA samples was assessed with an Agilent 2100 Bioanalyzer. Only samples that passed the Bioanalyzer QC and also had OD260/280 and OD 260/230 ratios ratios > 1.8 were further processed for microarray analysis. Affymetrix Mouse Gene 1.0
ST arrays were used for this study. Processing of the RNA samples was carried out according to the Affymetrix GeneChip Whole Transcript Sense Target labeling protocol.
MICROARRAY DATA ANALYSIS
Image generation and feature extraction were performed using Affymetrix GeneChip
Command Console (AGCC) software. The output files from AGCC were processed with
Bioconductor [http://bioconductor.org] software packages to determine various quality control metrics. Only data that met these QC specifications were included in the further analysis. Affymetrix microarray data were normalized using the Robust Multiarray
Average (RMA) method. A linear model was fit to all the data and then contrasts for the various comparisons were computed. This is quite similar to fitting individual t-tests for each comparison, but increases power by using all samples to compute intra-group variability. In addition, the linear model that was fit was slightly modified to take
78 advantage of the fact that we have many thousands of probesets, from which we can estimate an expected intra-group variability. The observed variability of a given probeset is then adjusted to more closely approximate the expected variability using the
Bioconductor package limma. This has been shown to increase power in both real and simulated data sets. Heatmaps generated using Bioconductor gplots package
(http://cran.r-project.org/web/packages/gplots/gplots.pdf). A heatmap is a false color image of the underlying data, where (in this case) genes have been ordered using a hierarchical clustering algorithm, which groups genes according to their Euclidean distance. All collected raw expression array data were deposited in GEO database.
Accession number is GSE36078.
INGENUITY PATHWAY ANALYSIS
Ingenuity Pathway Analysis (IPA; http://www.ingenuity.com/; Ingenuity Systems
Inc., Redwood City, CA) was used to generate signaling network diagrams (figs. S3.89) in the following manner. The gene symbols CREB1, NFkB1, SRF and NR4A2 or ATF3 were used as seeds and IPA’s grow tool was employed to identify genes/proteins upstream and downstream of these seed molecules. Network diagrams show only those genes that were more than 1.5-fold (P < 0.05) differentially expressed in at least one of the two contrasts [HAdV5-WT versus Saline-WT] or [HAdV5-IL-1RI versus Saline-IL-
1RI]. Red indicates up-regulation, whereas green indicates down-regulation relative to the Saline-WT or Saline-IL-1RI groups.
79
PROMOTER ANALYSIS OF DIFFERENTIALLY EXPRESSED GENES
Genes that were differentially expressed more than 1.5-fold (p<0.001) in the two experimental group comparisons [HAdV5-WT versus Saline-WT] and [HAdV5-IL-1RI versus Saline-IL-1RI], were identified. These two comparisons shared 34 differentially expressed genes. The P-scan analysis tool (http://www.beaconlab.it/Pscan) was used to identify transcriptional response elements in the 1 kb upstream regions of the transcriptional start sites of these 34 genes. The Transfac and Jasper databases were chosen for this analysis. Fig. 3.3C shows a small part of the Z-score map generated by P- scan for NFKB1, CREB1, and SRF transcription factors extracted from 34 gene promoter analysis z-score map in fig. S3.7.
GENE ONTOLOGY CATEGORY ANALYSIS
Genes that were differentially expressed more than 1.5-fold (P < 0.05) in the contrast [HAdV versus Saline] in WT mice were identified. The web based tool
CateGOrizer (http://www.animalgenome.org/tools/catego/) was used to sort these differentially expressed genes into Gene Ontology categories and determine percentages of differentially expressed genes for each category. This information was imported into
Microsoft Excel software and a pie chart figure was generated.
STATISTICAL ANALYSIS
Unless otherwise indicated, statistical analysis in each independent experiment was performed with an unpaired, two-tailed Student’s t-test. Data are reported as mean ± standard deviation. P < 0.05 was considered statistically significant
80
ACKNOWLEDGEMENTS
We thank Dr. Shizuo Akira (Osaka University, Japan) for providing mice deficient in
MyD88, MD2, TLR4, TLR7/8, and TLR9 and Dr. Jun-ichiro Inoue (University of Tokyo, Japan) for TRAF6-deficient mice. We gratefully acknowledge the assistance of Dewight R. Williams during cryo-EM data acquisition. The cryo-EM data was collected while JWF and PLS were at
Vanderbilt University. We thank Morley D. Hollenberg (University of Calgary, Canada) and
Alan Aderem and Daniel Zak (SeattleBioMed, USA) for helpful discussion. We thank Lisa K.
Baldwin for manuscript editing. This study was supported by US NIH grants AI065429 and
CA141439 to D.M.S., the UW NIEHS sponsored Center for Ecogenetics & Environmental Health
(P30ES07033), and CIDB Analytical Biopharmacy 1 Core facility, which is funded by the
Washington State Life Sciences Discovery 2 Fund. 3 Author Contribution: D.S. and P.S. designed the research, K.D., J.F., N.D., O.K., conducted 4 experiments, collected and analyzed the data, J.M., T.B., R.B., F.F., collected and processed all 5 the microarray data, P.S. and D.S. wrote the paper.
81
Figure 3.1 Cryo-EM structure of the FX-HAdv5 complex and simulation of the FX- hexon interface using molecular dynamics flexible fitting. (a) The cryo-EM structure of HAdv5 in complex with FX. The density is shown with the hexon capsid in blue, penton base in gold, fiber in green, and FX in red. The scale bar represents 100Å. (b)
82
An enlarged view of the FX-HAdv5 complex showing the network of the FX density above the hexon capsid. (c) The best rigid body fit orientation of the zymogenic FX model (red ribbon) within FX cryo-EM density (transparent pink). This FX density is selected from above a hexon near the icosahedral 3-fold axis of the capsid. (d)
Coordinates from a frame in the MDFF simulation that show hexon residues E424 and
T425 surround residue K10 of the FX-GLA domain. The side chains of these three residues are shown in space filling representation and colored by element. (e) FX-GLA
+ domain and associated Ca2 ions (green) in the central depression of the hexon trimer.
(f) Residue K10 in the FX-GLA domain is in close enough proximity to E424 and
T425 of hexon to engage in electrostatic interactions.
83
Figure 3.2 A single amino acid substitution T425A completely abrogates FX biding to
HAdv5. (a) The schematic diagram of a region of HAdv5 hexon hyper-variable loop 7
(HVR7) showing T423E424T425 amino acids and amino acid substitutions introduced in this region of individual viruses with mutated hexons (shown in red). (b) Kinetic
84 response data and Kd for human coagulation FX binding to different hexon mutated viruses obtained using surface plasmon resonance analysis (BIACORE).
Experimentally obtained data are shown by black lines. Global fits of these data to
1:1 single-site interaction model are shown in orange. The responses are shown in instrument response units vs. time in seconds. Representative data obtained from four independent experiments are shown.
85
Figure 3.3 FX binding-ablated virus triggers blunted transcriptional response of NFKB1- dependent genes in vivo. (a) Gene Ontology pie chart. Genes that were differentially expressed more than 1.5-fold (P < 0.05) in the spleens of WT mice challenged with HAdv5 or mock (saline) were identified. The CateGOrizer tool was used to sort these genes in Gene Ontology categories and determine percentages of differentially expressed genes for each category. (b) Heatmap
-/- representation of the averaged gene expression levels for 34 gene set when WT and Il1r1 mice were challenged with either HAdv or Ad-TEA mutant virus. N = 3 in each experimental group.
86
The subset of genes differentially activated by HAdv and Ad-TEA is highlighted by the blue bar.
The yellow and blue color legend shows log2-transformed fold changes. Heatmap was generated by using the Bioconductor gplots package. (c) Z-score map of the transcription factor binding site frequencies in proximal promoters of indicated genes was generated by P-scan from the analysis of binding sites for 116 transcription factors (Transfac database). The five gene set represents a subset of genes from (b) that are the most differentially inducted by HAdv and Ad-TEA mutant
-7 -6 virus (1.5-fold cut off). * for Nr4a2 P = 9.61x10 ; * for ATF3 P = 6.77x10 . The green and red color legend shows log2-transformed fold z-score changes.
87
Figure 3.4 HAdv induces NFKB1-dependent inflammatory cytokines and chemokines downstream of TLR4-TRIF/MyD88-TRAF6 signaling and viruses unable to bind FX fail to activate this pathway in vivo. (a) Mouse cytokine array panel showing differences in inflammatory cytokines and chemokines in the spleens of WT mice 1h after challenge with HAdv or Ad-TEA mutant, determined by proteome profiler antibody array.
Representative blot from 4 independent experiments is shown. C – mouse was challenged with saline. (b) mRNA expression of IL-1β in the spleen of WT mice 30 min after
88 challenge with indicated viruses. Graphs show mean ± SD, N = 4, ** P < 0.01. AU – arbitrary units reflecting IL-1β to GAPDH mRNA ratios. (c-d) mRNA expression of IL-
1β in the spleen of WT mice and mice deficient for indicated genes 30 min after challenge with HAdv. Graphs show mean ± SD, N = 4, ** P < 0.01. AU – arbitrary units reflecting
IL-1β to GAPDH mRNA ratios. (e) Mouse cytokine array panel showing differences in
-/- -/ - inflammatory cytokines and chemokines in the spleens of WT and Myd88 , Ticam , Tlr4
/- -/- , and Md2 mice 1 h after challenge with HAdv, determined by proteome profiler antibody array. Representative blot from 4 independent experiments is shown. C – mouse was challenged with saline. (f) Mouse cytokine array panel showing differences in inflammatory cytokines and chemokines in the spleens of WT mice 1h after challenge with wild-type human adenoviruses of indicated species, determined by proteome profiler antibody array. Representative blot from 4 independent experiments is shown. C – mouse was challenged with saline. (g) Schematic representation of coagulation factor
“decoration” model of HAdv that triggers TLR4-dependent activation of inflammatory cytokines and chemokines. Coagulation factors that were experimentally confirmed to
3 bind HAdv5 virions are highlighted by the red rectangles
89
Table S3.1 Distances at the FX-hexon interface before and after molecular dynamics flexible fitting runs with different starting FX orientations. Distances are reported in
Angstroms between oxygen atoms of hexon residues E424 and T425 and the nitrogen atom of the FX-GLA domain K10 residue.
90
91
Continued
Table S3.2 Thirty-four genes co-activated in the spleens of WT and Il1r1-/- mice after challenge with HAdV5.
92
93
Continued
Table S3.3 Differential expression of 34 gene set in the spleens of WT and Il1r1-/- mice after challenge with HAdV5 and TEA mutant viruses. The differentially expressed genes with 1.25-fold cut off and P < 0.01 are highlighted in bold. The differentially expressed genes with 1.5-fold cut off and P < 0.001 are highlighted in red.
94
Figure S3.1 Robustness of FX-hexon interface after molecular dynamics flexible fitting runs with different starting FX orientations. (A) Final coordinates after a 100ps MDFF simulation starting with the best rigid body fit (turn 0) FX orientation. (B) Same as in (a) but starting with an FX orientation rotated by +10 degrees around the molecular 3-fold axis of hexon, or (C) -10 degrees, or (D) -20 degrees.
95
Figure S3.2 Comparison of FX bound to hexon in the subnanometer resolution cryo-EM structure to simulated FX/hexon density. (A) Top view of the cryo-EM density filtered to
14Å resolution and contoured to show FX density protruding from the top of the hexon.
(B) Top view of the simulated density for FX (red) and hexon (blue) filtered to 10 Å resolution and calculated from all frames of the MDFF trajectory. (C and D)
96
Perpendicular views of FX bound to hexon. Note that the distal serine protease domain of
FX appears more diffuse in the cryo-EM structure than in the simulated density. (E) The cryo-EM density is contoured to show the FX-GLA domain density in the central depression at the top of hexon. The trimeric hexon contains three symmetrically related and overlapping FX binding sites, which accounts for the triangular shape of GLA density. (F) Simulated density for hexon and FX-GLA domain. The scale bars represent
25Å.
97
Figure S3.3 Infectivity of and response to wild type and hexon-mutated viruses in cells cultured in vitro and mouse hepatocytes in vivo. (A) The schematic diagram of a region of HAdV5 hexon hypervariable loop 7 (HVR7) showing T423E424T425 amino acids and amino acid substitutions introduced in this region of individual viruses with mutated hexons (shown in red). (B) Infection of CHO-K1 cells with indicated viruses with or without the addition of FX (8μg ml-1). Cells were infected with a multiplicity of infection of 200 v.p. cell-1. Mean fluorescent intensity of virus-encoded GFP reporter protein was analyzed by flow cytometry 24 h after virus infection. N = 6. *P < 0.01. n.s.
98
– not significant. (C) A549 cells were infected with indicated viruses at an MOI of 100 v.p. cell-1. Twenty four hours after virus infection, cells were harvested and the mean fluorescent intensity of virus-encoded GFP transgene was analyzed by flow cytometry. N
= 3. N.S. – not statistically significant. (D-E) Inflammatory cytokines and chemokines produced by A549 cells in vitro with and without the infection with indicated viruses at an MOI of 5x103v.p. cell-1 16 h after infection and determined by proteome profiler antibody array (D) or ELISA (E). N = 3. n.d. – not detectable. Punctate line shows the limit of detection. Pos.C – positive control sample of media spiked with 100 pg of IFN-α and IFN-β. (F) Quantitative representation of GFP protein levels adjusted for the actin loading control band intensity for each individual virus shown in Fig. 3.2C processed with histogram pixel density analysis. N = 4. The relative GFP signal intensity is shown in histogram units (HU). ** P < 0.001.
99
Figure S3.4 Co-localization of virus particles (red) with splenic marginal zone macrophages (green) observed 1 h after virus injection for indicated viruses analyzed by confocal microscopy. DAPI staining (blue) reveals the nuclei of cells and the physical border of the germinal centers are depicted by dotted lines. Representative fields are shown. Mock - mice were injected with saline and spleen samples were sectioned and stained with antibodies in a way identical to virus-injected mice. Scale bar is 10 μm.
100
Figure S3.5 Confocal microscopy analysis of virus particle localization with CD169+ and
MARCO+ marginal zone macrophages in the spleen of mice 30 minutes after challenge with indicated viruses. Mice were injected intravenously with indicated viruses at a dose of 1e10 virus particles per mouse and, 30 minutes later, spleens were harvested and sections were prepared and stained with DAPI (blue) to detect nuclei of splenocytes, as
101 well as Abs specific for CD169 and MARCO (green) or adenovirus hexon (red).
Confocal images were obtained with a Zeiss 510 Meta Confocal microscope. The physical border of splenic germinal centers are indicated by punctate lines.
Representative fields are shown. N = 4. Marginal zone macrophages with accumulated virus particles are indicated by arrows. Scale bar is 10 μm.
102
Figure S3.6 Confocal microscopy analysis of virus particle localization with F4/80+ macrophages in the liver of mice 30 minutes after challenge with indicated viruses. Mice were injected intravenously with indicated viruses at a dose of 1e10 virus particles per mouse and 30 minutes later, livers were harvested and sections were prepared and stained with DAPI (blue)
103 to detect nuclei of liver cells, as well as Abs specific for F4/80 (green) or adenovirus hexon (red). Confocal images were obtained with a Zeiss 510 Meta Confocal microscope.
Representative fields are shown. n = 4. F4/80+ macrophages with accumulated virus particles are indicated by arrows. Scale bar is 10 μm.
104
Figure S3.7 Z-score map of transcription factor binding site distribution in the -1000bp promoter regions of 34 genes co-activated in the spleens of WT and Il1r1-/- mice upon challenge with HAdV. The P-scan analysis tool (http://www.beaconlab.it/Pscan) was used to identify transcriptional response elements in the 1kb upstream regions of the
105 transcriptional start sites of these 34 genes. The binding site information for 116 transcription factors was selected from the Transfac database. The Zscore map was computed using P-scan software. There is a marked increase in the Z-score of CREB1,
SRF, and NFKB1 binding sites in the promoter regions of this gene set.
106
Figure S3.8 Ingenuity Pathway analysis of networks of transcriptional targets for
NFKB1, CREB1, and SRF transcription factors that respond to HAdV5 infection in the spleen of WT mice 30 minutes after intravenous virus injection. Genes that are up- regulated in response to virus infection are coded in red, down-regulated in green.
NFKB1, CREB1, and SRF are highlighted with circles. The cut off for transcriptional modulation is 1.5-fold, compared to the levels of the gene expression in mice injected with saline. N = 3; P < 0.05.
107
Figure S3.9 Ingenuity Pathway analysis of networks of transcriptional targets for
NFKB1, CREB1, and SRF transcription factors that respond to HAdV5 in the spleen of
IL-1RI-deficient mice 30 minutes after intravenous virus injection. Genes that are up- regulated in response to virus infection are coded in red, down-regulated in green.
NFKB1, CREB1, and SRF are highlighted with circles. Cut off for transcriptional modulation is 1.5-fold, compared to the levels of the gene expression in mice injected with saline. N = 3; P < 0.05.
108
Figure S3.10 The mRNA expression for IL-1β in spleens of WT and indicated gene- deficient mice 30 minutes after virus injection analyzed by the RNAse protection assay.
Representative gels are shown in biological duplicates. N = 3. C – mice injected with saline.
109
CHAPTER 4: FVII dimerization on adenovirus capsid may influence infectivity
This chapter is published: J. Virol. doi:10.1128/JVI.01070-13
Research Article:
Coagulation factor binding orientation and dimerization may influence infectivity of adenovirus-coagulation factor complexes.
Eric E. Irons*, Justin W. Flatt*, Konstantin Doronin, Tara L. Fox, Mauro Acchione,
Phoebe L. Stewart, Dmitry M. Shayakhmetov
*Contributed equally
4.1 ABSTRACT:
Adenoviruses (AdVs) are promising vectors for therapeutic interventions in humans. When injected into the bloodstream, Ad vectors can bind several vitamin K- dependent blood coagulation factors, which contributes to virus sequestration in the liver by facilitating transduction of hepatocytes. Although both coagulation factors FVII and
FX bind the hexon protein of human AdV serotype 5 (HAdV5) with a very high affinity, only FX appears to play a role in mediating AdV-hepatocyte transduction in vivo. To understand the discrepancy between efficacy of FVII binding to hexon and its apparently poor capacity for supporting virus cell entry, we analyzed the HAdV5-FVII complex by
110 using high-resolution cryo-electron microscopy (cryo-EM) followed by molecular dynamic flexible fitting (MDFF) simulations. The results indicate that although hexon amino acids T423, E424, and T425, identified earlier as critical for FX binding, are also involved in mediating binding of FVII, the FVII GLA domain sits within the surface- exposed hexon trimer depression in a different orientation from that found for FX.
Furthermore, we found that when bound to hexon, two proximal FVII molecules interact via their serine protease (SP) domains and bury potential heparan sulfate proteoglycan
(HSPG) receptor binding residues within the dimer interface. In contrast, earlier cryo-EM studies of the AdV-FX interaction showed no evidence of dimer formation. Dimerization of FVII bound to AdV may be a contributing mechanistic factor for the differential infectivity of AdV-FX and AdV-FVII complexes, despite high-affinity binding of both these coagulation factors to the virus.
4.1 INTRODUCTION:
Viral vectors based on adenoviruses (AdVs) specific to human and animal species have been adapted widely for gene transfer applications both in vitro and in vivo.
Extensive in vitro analyses have revealed a model of AdV cell infection whereby the initial virus attachment to a plasma membrane-localized receptor via the fiber protein is followed by virus penton interaction with integrins that mediate virion internalization into the cell (Nemerow, 2000; Smith et al, 2010b)
. To date, a number of cellular proteins that serve as functional high-affinity virus attachment receptors have been identified. For the most common vectors, based on species C human adenovirus serotype 5 (HAdV5), as well as for HAdV of species A, D,
111
E, and F, it was found that a tight junction protein designated coxsackie and adenovirus receptor (CAR) can serve as a high-affinity attachment receptor (Bergelson et al, 1997;
Roelvink et al, 1998; Tomko et al, 1997). The species B AdVs may utilize CD46 and/or
DSG2 proteins to gain entry into host cells (Gaggar et al, 2003; Segerman et al, 2003;
Wang et al, 2011), while several species D AdVs may utilize CD46, sialic acid, or GD1a glycan to enter the cell (Arnberg et al, 2002; Nilsson et al, 2011). It was also shown that
HAdV can bind a variety of integrin classes that interact with a penton RGD amino acid motif to trigger internalization of cell-bound virus particles into the cell (Smith et al,
2010b).
Although this canonical pathway of HAdV cell entry operates efficiently in vitro and explains the topology and functional interdependence between viral capsid proteins
(Liu et al, 2010; Reddy et al, 2010), the biodistribution of HAd5-based vectors in mice after intravascular administration revealed no correlation with tissue levels of CAR expression (Akiyama et al, 2003; Alemany & Curiel, 2001). Instead, it was found that
AdV particles accumulate in the liver and that virus entry into hepatocytes is mediated by hexon interaction with vitamin K-dependent blood coagulation factors (Kalyuzhniy et al,
2008; Parker et al, 2006; Shayakhmetov et al, 2005; Waddington et al, 2008). While earlier studies showed that several homologous blood coagulation factors, including
FVII, FIX, and FX, can support virus infection of susceptible cells in vitro via this mechanism (Parker et al, 2006), specific inactivation of FX alone was sufficient to completely abrogate hepatocyte transduction with AdV in mice after intravenous virus administration (Waddington et al, 2008). Furthermore, although analyses of binding affinities of different blood coagulation factors for hexon showed that FIX binds poorly
112 to Ad, both FVII and FX bound the virus hexon protein with very high affinities, in the low-nanomolar range (Kalyuzhniy et al, 2008). Using moderate-resolution cryo-electron microscopy (cryo-EM), we and others previously showed that FX interacts with solvent- exposed hypervariable (HVR) loops of HAdV5 hexon via the GLA domain (Kalyuzhniy et al, 2008; Waddington et al, 2008). Substitution of HVR loops in HAdV5 hexon for
HVR loops from HAdV26 hexon, which does not bind FX, or introduction of mutations into solvent-exposed regions of AdV hexon resulted in a loss or significant reduction of the FX binding affinity for hexon (Alba et al, 2009). However, using this extensive mutagenesis approach, the specific amino acids that form the FX-AdV hexon binding interface were not defined.
Using high-resolution cryo-EM followed by molecular dynamic flexible fitting
(MDFF) simulations, we recently modeled the FX interaction with HAdV5 hexon and identified the T423-E424-T425 amino acid motif in HVR7 as critical for high-affinity FX binding to adenovirus. We further demonstrated that a single amino acid substitution,
T425A, completely abrogated FX binding to AdV (Doronin et al, 2012). To reconcile the discrepancy between high-affinity binding of FVII to the virus and its poor capacity to support virus entry into the cell, in this study we analyzed the HAdV5-FVII interaction interface by using high-resolution cryo-EM followed by MDFF simulations. Our analyses revealed that although hexon amino acids T423, E424, and T425 are also involved in mediating binding of FVII, the FVII GLA domain sits within the surface-exposed hexon trimer depression, in a different orientation from that found for FX. The MDFF simulations indicated that when two proximal FVII molecules are bound to hexon, they interact via their serine protease (SP) domains and bury potential heparan sulfate
113 proteoglycan (HSPG) receptor binding residues within the dimer interface. In contrast, earlier cryo-EM studies showed no indication that FX interaction with hexon leads to the formation of SP domain dimers and indicated that virus attachment to cells is efficient in the presence of FX. Dimerization of FVII bound to AdV may be a contributing mechanistic factor for the differential infectivity of AdV-FX and AdV-FVII complexes despite high-affinity binding of both these coagulation factors to the virus.
4.3 MATERIAL AND METHODS:
CELLS AND VIRUSES
293 cells were obtained from Microbix (Toronto, Canada). CHO-K1 cells
(expressing HSPG; ATCC CCL-61) were obtained from the American Type Culture
Collection. CHO-CAR cells were kindly provided by Jeffery Bergelson and were described earlier (Bergelson et al, 1997). All cell lines were grown on Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum. The E1- and E3-region-deleted replication-defective HAdV5-based vectors AdV-WT, AdV-GAT,
AdV-TAT, AdV-TEA, and Ad5VS were previously constructed and described in detail elsewhere (Doronin et al, 2012; Shayakhmetov & Lieber, 2000; Shayakhmetov et al,
2000). The Ad5S vector possesses a short Ad9-derived fiber shaft domain but the fiber tail and knob domains of HAdV5. This vector was used in lieu of AdV-WT virus to obtain high-contrast cryo-EM images of virus-FVII complexes. All these vectors express the green fluorescent protein (GFP) gene under the control of the human cytomegalovirus
(CMV) early promoter. The GFP gene expression cassette is located in the E3 region of the AdV genome. Per previously published designations, AdV-WT possesses the wild-
114 type HAdV5 capsid without any modifications, AdV5-GAT possesses T423G and E424A mutations in the hexon HVR7 loop, AdV5-TAT possesses the E424A mutation, and
AdV-TEA possesses the T425A mutation in the hexon HVR7 loop (Doronin et al, 2012).
For AdV amplification, 293 cells were infected under conditions that prevented cross- contamination. Viruses were banded in CsCl gradients, dialyzed, and stored in aliquots as described earlier (Shayakhmetov et al, 2000). AdV genome titers were determined by measuring the optical density at 260 nm (OD260). For each AdV used in this study, at least two independently prepared virus stocks were obtained. Each produced virus stock was tested for endotoxin contamination by using Pyrotell Limulusamebocyte lysate (Cape
Cod Inc., Falmouth, MA). For in vivo experiments, only virus preparations confirmed to be free of endotoxin contamination were used.
CRYO-ELECTRON MICROSCOPY, IMAGE PROCESSING, AND MODELING
For cryo-electron microscopy analyses, we utilized a modified HAdV5-based vector, referred to earlier as Ad5S (Doronin et al, 2012), which contains HAdV5 hexon, the HAdV5 penton base, and a modified fiber protein. In this fiber protein, the knob and tail domains are from HAdV5, but the fiber shaft domain is derived from HAdV9, rendering the fiber about 3-fold shorter than that of unmodified HAdV5. This form of
HAdV5 was chosen to produce thinner ice on cryo-EM grids and to improve image contrast in cryo-electron micrographs. Samples were prepared by mixing 50 fmol of modified HAdV5 with a short-shafted fiber with 200 pmol of FVII in 50 mM Tris, pH
7.6, 150 mM NaCl, 2 mM CaCl2, and 2 mM MgCl2 for 15 min at room temperature.
After incubation, 3.5-μl aliquots of virus particles with FVII bound were applied to
115
Quantifoil grids (Quantifoil Micro Tools GmbH) and suspended in thin vitrified ice layers by use of a homebuilt vitrification device. Digital micrographs of the HAdV5-FVII complex were acquired under low-dose conditions at liquid nitrogen temperature on an
FEI Polara 300-kV FEG transmission cryo-electron microscope with a Gatan UltraScan
4000 charge-coupled device (CCD) camera. A total of 1,503 particle images were collected at an absolute magnification of 397,878×, corresponding to a pixel size of 0.4 Å on the molecular scale. The defocus values of the micrographs ranged from −1 to −4 μm.
Individual particles were selected from micrographs by use of in-house scripts calling
IMAGIC subroutines (van Heel et al, 2000) and were binned to generate particle image stacks with pixel sizes of 4.5, 2.2, and 1.5 Å that were used for initial, intermediate, and final refinement rounds, respectively. The program CTFFIND3 (Mindell & Grigorieff,
2003) was used to determine estimates for the microscope defocus and astigmatism parameters. A cryo-EM structure of Ad5.F35 (Saban et al, 2006) was used as the starting model for FREALIGN refinement (Grigorieff, 2007). After the final round of refinement, the resolution of the icosahedral capsid (radii of 325 to 460 Å) at the Fourier shell correlation (FSC) 0.5 threshold was calculated to be 8.7 Å. The final map included 837 particles selected from the total data set.
A composite structural model of the zymogenic form of FVII (residues 1 to 406) was built using existing X-ray crystallographic data (Banner et al, 1996; Eigenbrot et al,
2001) as well as a computational model of the full-length protein (Perera et al, 2002).
Atomic-resolution information is available for HAdV5 hexon (Liu et al, 2010; Reddy et al, 2010). In order to model the interface between HAdV5 and FVII, atomic coordinates for one or two hexon trimers and one or two copies of the FVII model were docked into
116 the HAdV5-FVII cryo-EM density at the 2-fold axis of the icosahedral capsid. Weak density for the EGF1 and EGF2 domains between the GLA and SP domains helped to guide the overall position. Attempts to use the GLA domain position from the previous
HAdV5-FX model (Doronin et al, 2012) to guide positioning of the FVII GLA domain resulted in models that did not match the experimental HAdV5-FVII cryo-EM density.
The docked FVII coordinates served as input for multiple 100-ps MDFF simulations
(Trabuco et al, 2008), with implicit solvent and a g-scale parameter of 0.3. The robustness of the MDFF-refined FVII orientation was tested by rotating the FVII starting orientation by ±10° or ±20° about the 3-fold molecular axis of hexon. Many of the same protein-protein interactions were recovered compared to the 0° starting orientation, lending support to the MDFF-derived fit of FVII within the cryo-EM density. Nonbonded interaction energies were evaluated with the NAMD Energy plug-in in VMD (Humphrey et al, 1996). The MDFF simulations were performed on the Case Western Reserve
University High-Performance Computing Cluster.
ADV INFECTION IN VITRO
Unless noted otherwise, 2.5 × 105 CHO-K1 or CHO-CAR cells were infected at a multiplicity of infection (MOI) of 2,000 virus particles (v.p.)/cell in 400 μl saline, with or without FVII or FX. Coagulation factors were added at concentrations ranging from 0.5
μg/ml to 25 μg/ml (each). The 12-μg/ml concentration of FX corresponds to its physiological concentration in the blood (1 U/ml). The physiological concentration of
FVII in the blood is 0.5 to 0.6 μg/ml, which is about 17-fold lower on a molar basis than that of FX. We used the same amounts of coagulation factors in all comparative analyses.
117
Two hours after the addition of viruses to cells, the virus-containing saline was replaced by fresh growth medium. Reporter gene activity was analyzed 48 h later by flow cytometry. Affinity-purified human blood coagulation factors FX and FVII were purchased from Hematologic Technologies, Inc.
ADV ATTACHMENT ASSAY
AdV attachment studies were performed based on a protocol published elsewhere.
Briefly, 5 × 105 cells were incubated for 1 h on ice with equal amounts of [3H]thymidine- labeled HAdV5 or Ad5S at an MOI of 2,000 v.p./cell, either alone or in the presence of human FVII or FX (at 12 μg/ml) in 150 μl of ice-cold adhesion buffer (DMEM supplemented with 2 mM MgCl2, 1% bovine serum albumin [BSA], and 20 mM
HEPES). Next, the cells were pelleted by centrifugation for 4 min at 1,000 × g and washed two times with 0.5 ml of ice-cold phosphate-buffered saline (PBS). After the last wash, the cells were pelleted at 1,500 × g, the supernatant was removed, and the cell- associated radioactivity was determined by use of a scintillation counter. The number of viral particles bound per cell was calculated using the virion-specific radioactivity and the number of cells as described earlier.
SPR ANALYSES
All surface plasmon resonance (SPR) analyses were carried out on a Biacore
T100 machine (15). HAdV5 and FVII samples were dialyzed against HBS-N buffer (10 mM HEPES, 150 mM NaCl, 1 mM CaCl2, 0.5 mM MgCl2, pH 7.4) twice, using 1 liter buffer and 0.1- to 0.5-ml dialysis cassettes (molecular weight cutoff [MWCO], 10,000)
118
(Pierce Co.). Samples were diluted 1:1 with dialysis buffer to ensure that the volume was greater than 100 μl prior to loading into the cassette, and samples were dialyzed for at least 2 h at 4°C prior to changing the buffer once. Dialyzed samples were clarified in 1.5- ml centrifuge tubes at 14,000 × gprior to SPR analysis.
Dialyzed virus samples were diluted 1:9 in 10 mM sodium acetate, pH 4.0, and immobilized onto research-grade CM5 Biacore chips (Biacore Inc., Piscataway, NJ).
Once immobilized, the chips were allowed to equilibrate in running buffer for 10 min prior to testing binding of FVII and regeneration. The running buffer was 10 mM
HEPES, 150 mM NaCl, 1 mM CaCl2, 0.5 mM MgCl2, 0.1% (wt/vol) BSA, and 0.005%
(vol/vol) polysorbate 20, pH 7.4. All buffers were filtered through a 0.2-μm membrane prior to use. The regeneration buffer was the same as the running buffer, without calcium or magnesium and including 3 mM EDTA. After immobilization, the chips were tested with a manual run to confirm function and to establish the best protocols for kinetic analysis. From this analysis, the operating conditions chosen were a flow rate of 50
μl/min, a 1-min association time, a 5-min dissociation time, and two 2-min regeneration cycles, with 1 min of stabilization. Samples were run in triplicate with buffer injections for double referencing. All data were collected at 1 Hz, using two or three replicate injections for each concentration of analyte. The dissociation time of 5 min was chosen to aid in regeneration rather than being necessary for the kinetic fits, so for data analysis the dissociation curve was truncated at 90 s. Fits to a 1:1 kinetic model were good for the wild-type AdV-WT virus and adequate for two of the mutants, AdV-GAT and AdV-
TAT. The AdV-TEA mutant showed no significant binding to FVII within the range of
119 tested concentrations of FVII (0 to 500 nM). Data processing and kinetic analysis were performed using Origin 8.5 software.
STATISTICAL ANALYSIS
Unless otherwise indicated, statistical analysis in each independent experiment was performed with an unpaired, two-tailed Student's ttest. Data are reported as means ± standard deviations. P values of <0.05 were considered statistically significant.
PROTEIN STRUCTURE ASSESSMENT NUMBER
The cryo-EM structure from this study has been deposited in the EM Data Bank under accession number EMD-5594.
4.4 RESULTS:
CRYO-EM STRUCTURAL ANALYSIS OF ADENOVIRUS-FVII INTERACTION
Previous cryo-EM structural studies indicated that FX binds in the central depression of the 240 hexons within the adenoviral capsid (Doronin et al, 2012;
Kalyuzhniy et al, 2008; Waddington et al, 2008) and led to identification of key interaction residues on hexon (Doronin et al, 2012). Here we applied the same approach to investigate the interaction between FVII and HAdV5. A subnanometer (8.7-Å)- resolution cryo-EM structure was determined for the HAdV5-FVII complex (Fig.
4.1A and B). In the structure, the FVII density is observed to extend outward from each hexon, creating a network of density surrounding the HAdV5 capsid. It is clear from the structure that FVII does not interact with either the penton base or fiber capsid protein.
120
Atomic models of FVII and hexon were docked into a region of cryo-EM density taken from near the icosahedral 2-fold axis of the HAdV5 capsid. A good fit was obtained for the GLA domain, in the density within the hexon central depression, and for the distal SP domain, within the density above the capsid (Fig. 4.1C). Presumably the full-length FVII molecule is somewhat flexible at the domain linker regions. This could explain why not much density is observed for the intervening EGF1 and EGF2 domains and why the SP domain density is weaker than that of the GLA domain. An MDFF analysis revealed a stable interaction between the FVII GLA domain and hexon. In particular, an Arg residue of FVII (R28) formed interactions with a patch of polar and negatively charged residues on hexon, i.e., T423-E424-T425 (Fig. 4.1D to F). While the MDFF simulation indicated additional interactions between the GLA domain and hexon, we noted these three residues in particular because they were also predicted to be involved in the interaction with FX (Doronin et al, 2012; Kalyuzhniy et al, 2008).
MUTATION OF THE HVR7 TET AMINO ACID MOTIF REDUCES THE AFFINITY
OF FVII BINDING TO THE VIRUS
Prior to the analysis of FVII binding to a set of HVR7-mutated AdV vectors, we evaluated the integrity and infectivity of a virus preparation of each individual hexon- mutated vector. Polyacrylamide gel electrophoresis showed that all major proteins for all vectors were present in similar amounts (Fig. 4.2A). The analysis of vector infectivities on susceptible CHO-CAR cells further confirmed that all the hexon-mutated vectors exhibited similar transduction properties (Fig. 4.2B). Because MDFF simulations of the
AdV-FVII hexon interaction interface showed that the T423-E424-T425 amino acids are
121 likely to contribute to the formation and/or stabilization of the complex, we next analyzed the FVII binding affinities of AdV-WT and AdV mutants possessing amino acid substitutions in this region of hexon by using surface plasmon resonance (Biacore,
Piscataway, NJ). To avoid complications with interpretation of the results caused by the multivalent nature of the virus particle, AdV-WT and the previously described AdV mutants AdV-GAT, AdV-TAT, and AdV-TEA (Doronin et al, 2012) were immobilized on a CM5 sensor chip by cross-linking. Next, various concentrations of FVII, ranging from 0 nM to 500 nM, were injected over sensor surfaces with immobilized virus, and binding responses were recorded and analyzed. Figure 4.2 shows a representative data set obtained from the analysis of the AdV-FVII interaction together with global fits to a 1:1 interaction model. The kinetic dissociation constant (KD) for the AdV-WT–FVII complex was determined to be 2.99 nM, in agreement with our previously published measurements (Kalyuzhniy et al, 2008). Using the same experimental approach, we found that mutations of the TET amino acids within the hexon HVR7 region drastically reduced the KD of FVII-AdV complexes (Table 4.1). Specifically, the kinetic KD determined for the FVII–AdV-GAT complex was 32.7 nM, and that for the
FVII–AdV-TAT complex was 28.5 nM, while FVII failed to bind to the AdV-TEA mutant immobilized on the sensor chip. Collectively, these data confirm that the TET amino acid motif in HVR7 predicted by MDFF to interact with the FVII GLA domain is indeed critical for mediating high-affinity binding between FVII and AdV hexon.
122
FVII IS INEFFICIENT AT SUPPORTING VIRUS ATTACHMENT AND CELL
TRANSDUCTION, DESPITE EFFICIENT BINDING TO HEXON
The observation of high-affinity FVII binding to AdV hexon (in the low- nanomolar range) and the lack of an FVII contribution to transduction of hepatocytes in vivo after intravenous AdV administration (Parker et al, 2006; Waddington et al, 2008) are puzzling. These findings suggest that if the AdV-FVII complex is formed in vivo, the cell transduction properties of such a complex are likely to be inferior to those of AdV complexed with FX. To directly compare the infectivities of the AdV-FVII and AdV-FX complexes, we infected CHO-K1 cells with AdV-WT in the presence of human FVII or
FX at concentrations ranging from 0.5 μg/ml to 25 μg/ml. This analysis demonstrated that the addition of FX to AdV-WT prior to infection of this cell line with the virus significantly increased the percentage of virus-transduced cells and the GFP mean fluorescence intensity (Fig. 4.3A and B). In contrast, the addition of FVII to AdV-WT prior to cell infection did not increase the percentage of virus-transduced cells or the GFP mean fluorescence intensity compared to that in experimental settings where cells were infected with the virus alone. We hypothesized that differential virus attachment to cells in the presence of FVII and FX could explain the reduced infectivity of the AdV-WT–
FVII complex compared to the AdV-WT–FX complex. To directly assess this possibility, we incubated [3H]thymidine-labeled AdV-WT vector with CHO-K1 cells in the presence of either FVII or FX and analyzed the efficacy of virus-cell attachment. This analysis confirmed that the AdV-FX complex bound to cells significantly more efficiently than the AdV-FVII complex did (Fig. 4.3C). Because we utilized the Ad5S variant for collection of high-contrast cryo-EM images of virus-FVII complexes, we also analyzed
123 whether Ad5S attachment to cells was differentially affected by FX and FVII, using a
[3H]thymidine-labeled AdV5S vector. This analysis demonstrated, similar to the case with AdV-WT, that only the addition of FX could increase the number of attached
AdV5S particles to cells, while the addition of FVII failed to increase virus attachment
(Fig. 4.3D).
FVII BINDS HADV5 HEXON IN AN ALTERED ORIENTATION COMPARED TO
THAT OF FX
Comparison of the network of factor density in the HAdV5-FVII complex (Fig.
4.1A and B) with the cryo-EM density observed for the HAdV5-FX complex (Doronin et al, 2012) shows a different distribution for each coagulation factor despite interactions with the HAdV5 capsid via homologous GLA domains. The FVII and FX GLA domains have 60% identity. Our MDFF analyses of the two HAdV5 factor complexes indicated that several of the same hexon residues play a role in mediating binding (Fig. 4.1)
(Doronin et al, 2012). In addition, our Biacore data for the two complexes showed that mutation of hexon residue T425 to alanine ablates binding of both factors (Fig. 4.2)
(Doronin et al, 2012). Interestingly, despite these similarities, the cryo-EM and MDFF analyses indicated that the FVII GLA domain sits within the hexon trimer, in a different orientation from that found for FX (Fig. 4.4A and B). This leads to a different angle for the rest of the factor molecule with respect to the viral capsid and to a different presentation of the SP domain (Fig. 4.4C and D). While an Arg residue (R28) in FVII appears to form critical interactions with the hexon TET sequence (Fig. 4.1D to F), the key factor residue in the modeled FX interaction appears to be a Lys residue (K10)
124
(Doronin et al, 2012). These different interaction residues result in the FVII GLA domain sitting deeper within the central hexon depression and in a different orientation, resulting in an altered positioning of FVII compared to FX (Fig. 4.4C and D).
FVII DIMERIZES VIA SP DOMAIN INTERACTIONS WHEN BOUND TO HADV5
HEXON
Visual inspection of the cryo-EM density assigned to FVII near the icosahedral 2- fold axis suggested that two SP domains might interact (Fig. 4.5). The volume of the SP density at this site is larger than expected for one copy of SP and nearly large enough to encompass two copies of the SP domain (Fig. 4.5B). We suspect that flexibility between
FVII domains leads to weaker and more diffuse SP domain density than that of the GLA domains or hexon. In general, we found a much weaker SP domain density in the
HAdV5-FVII cryo-EM structure than in the HAdV5-FX cryo-EM structure (Doronin et al, 2012). This is likely due to flexibility within the FVII molecule combined with formation of different SP domain interactions over the capsid. We assume that each hexon trimer can bind one FVII molecule in three different orientations related by 120° rotations. Therefore, an FVII molecule bound to a specific hexon may have a variety of different neighboring FVII molecules with which it could dimerize, leading to poorly defined SP domain density in the cryo-EM structure.
To simulate the interaction between neighboring FVII molecules, we docked two copies of FVII and two hexon trimers into the cryo-EM density extracted from a region near the 2-fold axis. This region displays the strongest SP domain density in the HAdV5-
FVII cryo-EM structure. The hexon trimers could be docked with a high level of
125 precision because strong density rods were observed for α-helices within hexon. Both
FVII molecules were initially docked using the MDFF-refined orientation found for a single FVII molecule and then subjected to further MDFF simulations (Table 4.2).
Although the two FVII models were positioned in the same way with respect to hexon at the beginning of simulation 1, at the end of this simulation the orientations of the two
FVII molecules diverged. In addition, the two FVII molecules showed a significant interaction via their SP domains, as indicated by a relatively large and favorable calculated nonbonded interaction energy (−367 kcal/mol). Also, at the end of simulation
1, we found that one FVII-hexon pair had a less favorable interaction than the other
(−638 versus −1,027 kcal/mol). After an additional simulation (simulation 2), the two
FVII molecules showed a stronger favorable interaction between them, and the nonbonded interaction energies for the two FVII-hexon pairs diverged even farther.
These results indicate that neighboring FVII molecules interact with and influence each other. In order to remove the influence of one FVII molecule on the other, a third simulation (simulation 3) was performed with only one FVII molecule and one hexon trimer present. This resulted in a strengthening of the calculated nonbonded interaction energy between FVII and hexon. These results for the HAdV5-FVII complex are clearly different from the previous cryo-EM and MDFF findings for the HAdV5-FX complex, in which no evidence for FX dimerization was observed (Doronin et al, 2012).
126
FVII DOMAIN DIMERIZATION OBSCURES PUTATIVE RECEPTOR-
INTERACTING RESIDUES WITHIN THE DIMER INTERFACE
In the FX SP domain, seven basic amino acids (R273, K276, R306, R347, K351,
K420, and R424) have been shown to mediate surface attachment of HAdV-FX complexes to HSPGs on hepatocytes (Duffy et al, 2011). Previous analyses identified that
K420 and R424 are the most critical residues for heparin binding to FXa, and together with the other charged amino acids, they form the so-called heparin-binding exosite on the surface of the FX SP domain (Rezaie, 2000). Efficient heparin binding was also confirmed for coagulation factors FVII (Martinez-Martinez et al, 2011) and FIX (Yang et al, 2002), and the locations of heparin-binding exosites were proposed to be similar on these factors, based on structural homology analyses (Martinez-Martinez et al, 2011;
Rezaie, 2000). Highlighting positively charged amino acids, including those constituting the experimentally confirmed heparin-binding exosite on the available crystal structures for FXa (R306, K420, R424, K427, R429, K433, and K435) and the suggested heparin- binding exosite on FVII (H249, R271, R277, K389, R392, R396, and R402), showed that these amino acids cluster on one and the same side of the SP domain for both FX and
FVII (Fig. 4.6). However, highlighting these amino acids on the three-dimensional surface of the FVII SP domain in complex with HAdV5 revealed that these potential
HSPG-binding amino acids are mostly buried in the middle of the modeled SP domain dimerization interface (Fig. 4.7A). This finding suggests that dimerization between FVII molecules bound to HAdV5 may shield putative receptor-interacting amino acids from
HSPGs on the cell surface and contribute to the inability of FVII to mediate virus attachment via HSPGs. This is consistent with our data demonstrating that the addition of
127
FVII to the virus does not increase its attachment to CHO-K1 cells (Fig. 4.3C and D).
Furthermore, our earlier cryo-EM analysis of AdV-FX revealed that the known HSPG- binding residues of the FX SP domain are solvent accessible on the surface of the
HAdV5-FX complex (Fig. 4.7B). The structural analysis indicates that FX bound to
HAdV5 should be capable of mediating virus attachment, consistent with our experimental data on HSPG-expressing cells (Fig. 4.3). At the icosahedral 2-fold axis, we measured the distance of closest approach between FX SP domains to be ∼40 Å. The absence of FX dimerization when FX is bound to HAdV5, as well as a favorable binding orientation for FX, may therefore explain the efficient cell transduction by Ad in the presence of FX. Our data also indicate that in contrast to the case for FX, the poor efficiency of FVII in supporting AdV cell transduction in vitro can be explained, at least in part, by an unfavorable binding orientation leading to dimerization of SP domains and inaccessibility of HSPG-binding residues at the SP domain dimer interface.
DISCUSSION:
Viral vectors based on HAdV5 are highly efficient at infecting both dividing and nondividing cells and are now being used in clinical trials to deliver therapeutic genes to various diseased cell and tissue types in vivo. However, upon intravascular administration, AdVs are rapidly trapped in the liver and transduce hepatocytes via a mechanism that relies on binding of the virus hexon protein to blood coagulation factors
(Baker et al, 2007; Kalyuzhniy et al, 2008; Waddington et al, 2008). Although several related and structurally homologous vitamin K-dependent blood coagulation factors, including FVII, FIX, and FX, were shown to be able to mediate virus entry into cells of
128 hepatic origin in vitro, with various efficacies (Parker et al, 2006), the selective inactivation of FX alone was fully sufficient to ablate virus-mediated hepatocyte transduction in vivo (Waddington et al, 2008). Using surface plasmon resonance, we demonstrated that FIX binds poorly to HAdV5 particles immobilized on the surface of a sensor chip. However, both FVII and FX bind to immobilized HAdV particles or purified hexon trimers with a very high affinity, in the low-nanomolar range (Kalyuzhniy et al,
2008). Although the physiological molar concentration of FVII in the blood is lower than that of FX, the lack of any contribution of FVII to hepatocyte transduction in vivo is puzzling. This is true even under conditions where FX has selectively been inactivated in vivo (Waddington et al, 2008), indicating that FVII lacks transduction-supporting capability, at least in mouse models of intravascular HAdV5 delivery. We hypothesized that the lack of contribution of FVII to mediating hepatocyte transduction by AdV after intravascular virus administration may be due in part to differences in presentation of
HSPG receptor-interacting amino acids on SP domains of FX and FVII upon their binding to the virus.
To address this hypothesis experimentally, we utilized high-resolution cryo-EM followed by MDFF to model the HAdV5-FVII interaction. We recently obtained a subnanometer-resolution cryo-EM structure of the HAdV5-FX complex by using the same approach (Doronin et al, 2012). Subsequent MDFF simulations revealed a potential
FX-hexon interaction interface with a single dominant orientation of the FX GLA domain within the central depression of the hexon trimer, involving K10 in the FX GLA domain and HVR7 amino acids E424 and T425 in the HAdV5 hexon protein. Indeed, a single amino acid substitution, T425A, completely abrogated FX binding to hexon and resulted
129 in a loss of hepatocyte transduction by a mutated vector possessing this T425A substitution in hexon. The cryo-EM visualization of the HAdV5-FVII complex presented here confirms that FVII interacts with virus hexons via its GLA domain (Fig. 4.1).
Furthermore, our MDFF analysis indicates that HVR7 amino acids T423, E424, and
T425 appear to be critical for the formation of the complex between hexon and the FVII
GLA domain, consistent with experimental data showing that mutation of these individual amino acids significantly reduces FVII's binding affinity for the virus.
Surprisingly, attempts to use the GLA domain position from the previous HAdV5-FX model (Doronin et al, 2012) to guide positioning of the FVII GLA domain resulted in models that did not match the experimental HAdV5-FVII cryo-EM density. The de novo cryo-EM density-guided MDFF simulations revealed that the FVII GLA domain is positioned deeper within the hexon trimer central depression and in a different orientation from that of FX (Fig. 4.4). The most surprising finding, however, was that upon binding to the virus via their GLA domains, two adjacent FVII molecules appear to form stable dimers via their SP domains, a feature that was not observed in the HAdV5-FX complex.
Although MDFF simulations were performed only for FVII molecules bound to the hexons at the icosahedral 2-fold axis, manual docking of FVII molecules at other hexons in the capsid was performed. This indicated that a relatively small amount of flexibility between FVII domains, combined with three possible binding orientations of
FVII at each trimeric hexon, could lead to dimerization of most, if not all, of the bound
FVII molecules on Ad. Averaging of multiple possible dimerization patterns for FVII bound at the hexons other than those at the 2-fold axis would lead to weak SP domain density except at the 2-fold axis. Indeed, a poorly defined SP domain density was
130 observed over most of the AdV capsid in the cryo-EM structure. The hexons at the icosahedral 2-fold axis provide a unique environment with one favored FVII dimerization pair, which leads to stronger SP domain density at this particular site. The structural and functional data on localization of HSPG-interacting residues within FVII (Martinez-
Martinez et al, 2011), combined with our cryo-EM-based model of HAdV5-FVII, indicate that most of the HSPG residues are hidden within the interacting interface of modeled SP domain dimers formed by adjacent FVII molecules bound to the virus (Fig.
4.7). The suboptimal presentation of these basic residues on the surfaces of SP domains of virus-bound FVII may contribute to the poor capacity of FVII in supporting AdV cell infection. Indeed, the direct assessment of attachment of AdV to the cell surface in the presence of FX and FVII showed that FVII failed to increase the number of attached virus particles to the cell.
Collectively, our structural and functional analyses revealed an unexpected divergence of properties for AdV complexes with FX and FVII. Although several highly homologous vitamin K-dependent coagulation factors and protein C were shown to bind to the virus and support cell infection in isolated systems (Kalyuzhniy et al, 2008; Parker et al, 2006), our study reveals unanticipated and potentially mechanistic insight into the highly efficient and selective role of FX in mediating efficient virus entry into HSPG- expressing cells. Although coadministration of HAdV5 and mouse coagulation factor
FVII (at concentrations similar to those of FX) into warfarin-treated mice could reveal the potential role of FVII in supporting virus entry into hepatocytes in a mouse model in vivo, mouse FVII is not commercially available. In addition, administration of human
FVII in this setting would not provide definitive information on the role of mouse FVII in
131 virus entry into hepatocytes. Our structural and computational analyses of the HAdV5-
FX and HAdV5-FVII complexes indicate a favorable SP domain orientation only for FX, which leads to presentation of heparan sulfate proteoglycan receptor-interacting residues on the virus-bound coagulation factor. Our findings may have important implications for strategies of cell type-specific AdV targeting via FX GLA domain-containing heterologous bifunctional ligands (Chen et al, 2010). When bound to virus, these novel targeting ligands can interact with each other in unexpected ways, reducing the efficacy of virus cell infection, as we demonstrated here for AdV-FVII complexes.
ACKNOWLEDGEMENTS:
We thank Lee Pederson for the zymogenic FVII atomic model and Dewight
Williams for EM support. This study was supported by NIH R01 grants CA141439 and AI065429 to D.M.S.
132
Table 4.1 Binding of FVII to adenovirus vectors and summary of SPR fitting parameters for 1:1 kinetic modelsa
133
Table 4.2 Intermolecular nonbonded energies between FVII molecules and hexons at the icosahedral 2-fold axis of HAdV5 at the end of three 100-ps MDFF simulations.
134
Figure 4.1 Cryo-EM structure of the HAdV5-FVII complex and simulation of the FVII- hexon interface by molecular dynamic flexible fitting. (A) Cryo-EM structure of HAdV5 in complex with FVII. The density is filtered to 16 Å and is shown with the hexon capsid in blue, the penton base in gold, the fiber in green, and FVII in purple. (B) Enlarged view of the HAdV5-FVII complex showing the network of FVII density above the hexons of the viral capsid. The inset shows the density near an icosahedral 2-fold axis of the capsid, which is indicated by a white oval. (C) The best MDFF-refined model of zymogenic FVII
(purple ribbon) within the FVII cryo-EM density (transparent light purple). This FVII density was selected from above a hexon near an icosahedral 2-fold axis of the capsid.
(D) Coordinates from the final frame of the best MDFF simulation that show hexon residue E424 forming a salt bridge with residue R28 of the FVII GLA domain. The side chains of these two residues are shown in space-filling representation and colored by
135 element. (E) FVII GLA domain and associated Ca2+ ions (green) in the central depression of the hexon trimer. Residue R28 of the FVII GLA domain is shown, along with the nearby hexon residues T423, E424, and T425. (F) Enlarged view of residue R28 near T423, E424, and T425 as shown in panel E. Bars, 100 Å.
136
Figure 4.2 Integrity, infectivity, and kinetic response data and dissociation constants for
FVII binding to wild-type AdV and AdV vectors with mutated hexons. (A) Twofold dilutions of purified virus preparations for the indicated vectors were loaded onto a polyacrylamide gel, resolved, and stained with Coomassie blue to visualize major virus capsid proteins (indicated on the right). M, molecular size ladder. (B) Infectivity of hexon-mutated vectors on virus-susceptible CHO-CAR cells, analyzed by flow cytometry
137 at 24 h postinfection. Error bars indicate standard deviations of the means (n = 6). (C)
Experimentally obtained data are shown by multicolored lines, with different colors representing different concentrations of FVII used for the analyses. Global fits of these data to a 1:1 single-site interaction model are shown in gray. The responses are shown in instrument response units (RU) versus time in seconds (s). Representative data obtained from three independent experiments are shown.
138
Figure 4.3 Transduction of CHO-K1 cells with AdV-WT vector and attachment of AdV-
WT and Ad5S vectors to CHO-K1 cells in the presence of FVII and FX. (A and B) CHO-
K1 cells were infected with an AdV-WT vector that expresses the GFP reporter gene under the control of the CMV early promoter, with or without the addition of FX or FVII to the infection medium, and were analyzed by flow cytometry 48 h after infection. In control settings, cells were treated with saline instead of the virus (mock). (C and D)
Attachment of [3H]thymidine-labeled AdV-WT (C) and Ad5S (D) to CHO-K1 cells in the presence of FVII or FX. Representative data from two independent experiments done in triplicate are shown. The attachment of virus particles to CHO-K1 cells with or without the addition of coagulation factors was analyzed as described in Materials and
139
Methods and is expressed as the number of virus particles per cell (vp/cell). *, P < 0.05; n.s., not statistically significant.
140
Figure 4.4 Cryo-EM and MDFF simulations indicate that FVII and FX adopt distinct binding orientations relative to hexon. (A) FVII GLA domain (purple) docked into the central depression of hexon, shown with (upper) and without (lower) hexon. Residue R28 of the GLA domain is shown in space-filling representation and colored by element. (B)
The FX GLA domain (red) is shown similarly, with residue K10 of the GLA domain. (C)
Overlay showing the distinct orientations of the zymogenic forms of FVII (purple) and
FX (red) relative to hexon. Basic residues, including those that constitute heparin-binding exosites and are known to contribute to heparan sulfate proteoglycan binding in the FX serine protease domain (R273, K276, R306, R347, K351, K420, and R424) and the FVII serine protease domain (H249, R271, R277, K389, R392, R396, and R402), are shown.
(D) Perpendicular view of panel C.
141
Figure 4.5 Modeling of two molecules of FVII at the icosahedral 2-fold axis. (A) MDFF- refined model of neighboring capsid-bound FVII molecules. These FVII proteins are shown as light and dark purple ribbons, and two hexon trimers are shown as blue ribbons docked within the cryo-EM density. The density is contoured to reveal a small region of density for the FVII GLA domain within the central depression of hexon (purple). (B)
Similar to panel A, but with the density contoured to include the distal FVII serine protease domains. (C and D) Side views perpendicular to panels A and B.
142
Figure 4.6 Localization of positively charged amino acids on the surfaces of serine protease domains of coagulation factors FX (A) and FVII (B). Positively charged amino acids are highlighted in yellow within the crystal structures of FXa (structure 1HCG in the structure database [SDB]) and FVIIa (structure 1KLJ in SDB). Side views of the SP domains of FX and FVII are shown in the left panels. The right panels show positively charged amino acids, highlighted in yellow, in front views of the SP domains of FX and
FVII. For FX, amino acids K420 and R424 contribute to the heparin-binding exosite and were experimentally confirmed as the most critical residues for heparin binding. For
143
FVII, amino acids R392, R396, and R402 were suggested to be part of the heparin- binding exosite, based on biochemical analyses of heparin binding to FVIIa and on structural homology analyses of the exosite location in thrombin and FIXa (Martinez-
Martinez et al, 2011).
144
Figure 4.7 Proximal FVII molecules interact and bury potential heparan sulfate proteoglycan binding residues that are exposed on FX. (A) Top and side views of the
MDFF-modeled FVII interactions with two hexon trimers at the icosahedral 2-fold axis.
HAdV5 hexons are shown in blue, and FVII molecules are shown in dark and light purple space-filling representations. Seven basic residues (H249, R271, R277, K389, R392,
R396, and R402) in the FVII serine protease domain, localized on the side that was proposed to bind heparin and heparan sulfate proteoglycans, are shown in yellow. A side view with one transparent hexon and FVII molecule reveals that these basic residues are mostly buried within the modeled serine protease dimer interface. (B) Similar views for the MDFF-modeled FX interactions, with two hexon trimers at the icosahedral 2-fold axis. FX molecules are shown in red and do not interact. Seven basic residues in the serine protease domain (R273, K276, R306, R347, K351, K420, and R424) that bind
145 heparan sulfate proteoglycans are shown in green. The residues of both FVII and FX are numbered according to the zymogen models. Bar, 100 Å.
146
CHAPTER 5: Visualization of adenovirus capsid-incorporated HIV antigen
This chapter is published: PLoS ONE 7(11): e49607. doi:10.1371/journal.pone.0049607
Research Article:
Cryo-EM visualization of an adenovirus capsid-incorporated HIV antigen.
Justin W. Flatt, Tara L. Fox, Natalia Makarova, Jerry L. Blackwell, Igor P. Dmitriev,
Elena A. Kashentseva, David T. Curiel, Phoebe L. Stewart
5.1 ABSTRACT:
Adenoviral (Ad) vectors show promise as platforms for vaccine applications against infectious diseases including HIV. However, the requirements for eliciting protective neutralizing antibody and cellular immune responses against HIV remain a major challenge. In a novel approach to generate 2F5- and 4E10-like antibodies, we engineered an AdV vector with the HIV membrane proximal ectodomain region (MPER) epitope displayed on the hypervariable region 2 (HVR2) of the viral hexon capsid, instead of expressed as a transgene. The structure and flexibility of MPER epitopes, and the structural context of these epitopes within viral vectors, play important roles in the induced host immune responses. In this regard, understanding the critical factors for epitope presentation would facilitate optimization strategies for developing viral vaccine
147 vectors. Therefore we undertook a cryo-EM structural study of this AdV vector, which was previously shown to elicit MPER-specific humoral immune responses. A subnanometer resolution cryo-EM structure was analyzed with guided molecular dynamics simulations. Due to the arrangement of hexons within the AdV capsid, there are twelve unique environments for the inserted peptide that lead to a variety of conformations for MPER, including individual α-helices, interacting α-helices, and partially extended forms. This finding is consistent with the known conformational flexibility of MPER. The presence of an extended form, or an induced extended form, is supported by interaction of this vector with the human HIV monoclonal antibody 2F5, which recognizes 14 extended amino acids within MPER. These results demonstrate that the AdV capsid influences epitope structure, flexibility and accessibility, all of which affect the host immune response. In summary, this cryo-EM structural study provided a means to visualize an epitope presented on an engineered viral vector and suggested modifications for the next generation of AdV vectors with capsid-incorporated HIV epitopes.
INTRODUCTION:
Viruses and virus-like particles (VLPs) with capsid-incorporated or chemically- attached heterologous epitopes are being explored as vaccine platforms to provide protective immunity against pathogens (Plummer & Manchester, 2010). The potential advantages of viral vaccine vectors include multivalent display of epitopes and the ability of viral particles to stimulate both the adaptive and innate immune systems. Recently it has been found that viral pathogens trigger innate immune sensors that recognize unique
148 pathogen-associated molecular patterns (PAMPs) (Kawai & Akira, 2007). Vectors have been designed that display either antigenic peptides or, in some cases, whole protein domains. A hepatitis B VLP has been engineered to present a peptide from foot and mouth disease virus (FMDV), and this VLP elicits an immune response that is stronger than that from the FMDV peptide alone (Clarke et al, 1987). The hepatitis B VLP system has also been utilized to present GFP, and vectors have been produced that lead to a strong humoral immune response against GFP in rabbits (Kratz et al, 1999). In another example, a recombinant VLP was designed to display domains from the anthrax toxin receptor. A single administration of this VLP in rats led to a potent immune response against a lethal anthrax toxin challenge (Manayani et al, 2007).
Adenoviral (Ad) based vectors have a number of advantages for vaccine applications including the ability to infect a broad range of target cells, the fact that their genome can accept large insertions of up to 8 kb, and their natural immunogenicity. In a recent study, a noninfectious, disrupted AdV vector with a covalently linked cocaine analog was found to evoke high-titer immunity to cocaine in mice (Hicks et al, 2011).
Administration of this vaccine to mice was also found to reverse the hyperlocomotor activity induced by intravenous delivery of cocaine, suggesting that the vaccine- generated immune response blocks cocaine from reaching its target receptor in the brain.
AdV vectors have also been explored for use in HIV vaccine approaches with antigens either expressed as a transgene or displayed on the capsid surface with an “antigen capsid-incorporation” strategy. A combination of these approaches was employed to generate a multivalent vaccine vector presenting an HIV antigen within the major surface exposed AdV capsid protein hexon (Matthews et al, 2010). Vectors were developed both
149 with and without transgene expression of Gag. In this study, the capsid incorporated antigen was a twenty-four amino acid region of the HIV membrane proximal ectodomain region (MPER), derived from HIV glycoprotein gp41. This peptide region was selected for display on the AdV capsid because it is the target of one of the first identified broadly neutralizing anti-HIV-1 antibodies, 2F5 (Purtscher et al, 1994). Vaccinations in mice with
MPER incorporated AdV vectors elicited an HIV epitope-specific humoral immune response as well as an anti-HIV Gag cellular response (Matthews et al, 2010).
The magnitude and quality of the immune responses generated by an engineered viral vector or designed immunogen is influenced by the structure and flexibility of the epitope, as well as the structural context and accessibility of the epitope (Plummer &
Manchester, 2010). Consideration of the structural environment of the epitope is particularly important in the case of viral based vaccines of the “capsid-incorporation” type where insertions can often be introduced at a variety of possible sites. Due to the spatial restrictions imposed by the packing of viral capsid proteins, different insertion sites are likely to tolerate various size insertions and lead to a range of surface exposure levels. Systematic investigations on the effect of the insertion point within the hepatitis B core antigen show that fusion to an immunodominant site in the middle of the nucleocapsid protein produces the strongest immune response for the inserted foreign epitope (Schodel et al, 1992). Furthermore, epitope fusions at the amino terminus of the nucleocapsid protein, which is not surface exposed in the VLP, correlated with the lowest level of immunogenicity.
150
Both the structure and flexibility of the presented epitope can affect the immune response. Neutralizing antibodies have been found for specific conformations, extended or α-helical, for neighboring residues within the MPER region. The broadly neutralizing monoclonal antibody 2F5 binds an extended conformation of MPER (Ofek et al,
2004) while another such antibody, 4E10, targets an α-helical conformation (Cardoso et al, 2007). An elegant paper by Ofek et al. (Ofek et al, 2010) shows that by designing an immunogen with the proper conformation of MPER displayed on a protein scaffold it is possible to elicit antibodies that induce this region of gp41 to assume its 2F5-recognized shape. This paper also demonstrated a correlation between epitope flexibility and immunogenicity. Although all of the designed scaffold proteins presented similar extended conformations of MPER, the scaffolds with more flexibly held epitopes generated greater immune responses in animals. Antigen structure and flexibility has been shown to have an effect on antigen processing and presentation of helper T-cell epitopes (Dai et al, 2001). The presence of flexible flanking regions around immunodominant helper T-cell epitopes against HIV gp120 is implicated in enhancing antigen processing in both mice and humans (Dai et al, 2001) .
Human clinical trials of candidate HIV vaccines have shown some modest efficacy, suggesting that it should be possible to generate vaccine-elicited protection against HIV-1 infection (Kwong et al, 2011). While induction of neutralizing antibodies is a major goal of HIV-1 vaccine development (Kwong et al, 2011), an ideal vaccine would also induce cytotoxic T lymphocyte and helper T cell responses (Mirano-Bascos et al, 2008). The field of rational vaccine design is still in the relatively early phase of
151 figuring out the principles that correlate antigen display with the ultimate, multi-faceted immune response.
Atomic level structural information for the viral capsid is highly valuable during the vector design process. We reasoned that a structure of an engineered vector that produces an immune response would also be useful in the analysis phase and in guiding future vector design. In cases where the antigen is conformationally flexible, such as
MPER of HIV gp41, it is not possible to know a priori which conformation will be presented on an engineered viral vector. We have undertaken a high resolution cryo-EM structural study of the most immunologically promising AdV vector with a capsid- incorporated HIV MPER epitope described by Matthews et al. (Matthews et al, 2010).
Our hypothesis is that an understanding of how the MPER epitope is presented on the viral capsid surface will lead to ideas for rationally modifying the next generation of
AdV-based vectors to potentially produce a stronger immune response. This study illustrates that cryo-EM combined with guided molecular dynamics simulations can be used to describe the structure and flexibility of an HIV epitope incorporated into an engineered AdV capsid.
MATERIALS AND METHODS:
CRYO-EM AND IMAGE PROCESSING
A concentrated sample of the Ad-HVR2-GP41-L15 vector (~0.2 mg/ml) was produced as described previously (Matthews et al, 2010). Cryo-EM samples were prepared with Quantifoil grids (Quantifoil Micro Tools GmbH) and a homebuilt
152 vitrification device. An FEI Polara microscope (300 kV, FEG) operated at liquid nitrogen temperature was used for data acquisition. Digital micrographs were recorded on a Gatan
UltraScan 4000 CCD camera at an absolute magnification of 397,878X, corresponding to a pixel size of 0.4 Å on the molecular scale. The defocus values of the micrographs ranged from −1 to −4 µm. Individual particles were selected from micrographs with in- house scripts that call IMAGIC subroutines (van Heel et al, 1996) and computationally binned to generate matching particle image stacks with pixel sizes of 4.5 Å and 2.2 Å.
The data with the coarser pixel size was used for the initial refinement rounds. The program CTFFIND3 (Mindell & Grigorieff, 2003) was used to determine initial estimates for the microscope defocus and astigmatism parameters. A cryo-EM structure of
Ad5.F35 (Saban et al, 2006) was used as the starting three-dimensional model for
FREALIGN refinement (Grigorieff, 2007). The orientational and CTF parameters, as well as the absolute magnification values, were refined for the total dataset of 5,025 particle images. The resolution of the final reconstruction was estimated to be 8.7 Å by the Fourier Shell Correlation 0.5 threshold for the icosahedral capsid (radii 325–460 Å)
(Fig. S5.1). Comparison of the cryo-EM structure of Ad-HVR2-GP41-L15 with both the crystal structure of an intact AdV virion (Reddy et al, 2010) and the cryo-EM structure of
Ad5.F35(Saban et al, 2006) facilitated identification of density regions corresponding to the MPER insertions.
153
CRYO-EM GUIDED MOLECULAR DYNAMICS SIMULATIONS TO MODEL THE
MPER INSERTIONS
Given the diversity in sequence and structure for the gp41 MPER region in the
PDB, we performed secondary structure prediction on the specific MPER sequence and neighboring linker residues that are incorporated into the Ad-HVR2-GP41-L15 vector
(Fig. S5.2). The Jufo (Meiler & Baker, 2003), SAM (Karplus et al, 1998), and Psi-
Pred (Jones, 1999) secondary structure methods were applied and the α-helical propensity results were averaged. Various initial models for the MPER section of the insertion were built with Swiss-PDB Viewer (Guex & Peitsch, 1997) including a long α-helix, and α- helices with a few extended residues at either end. Manual docking of these models into the cryo-EM density aligned with the crystal structure of the intact AdV virion (Reddy et al, 2010) led to selection of a model with four residues extended at the N-terminal end of the MPER sequence. An initial model for the 3-mer site was constructed using the selected monomeric model and the packing arrangement observed in the crystal structure of a partial MPER sequence stabilized with an isoleucine zipper motif (PDB-ID
3G9R) (Liu et al, 2009). Additional 3-mer models were made with the individual helices rotated by +/−10°, 20°, 90°, and 180° along the helical axis to evaluate the packing interface.
Initial models for the 1-mer, 2-mer, and 3-mer MPER insertions were incorporated into hexon trimer coordinate files (PDB-ID 1VSZ) (Reddy et al, 2010). The molecular dynamics flexible fitting (MDFF) program (Trabuco et al, 2008) was used together with NAMD 2.8 (Phillips et al, 2005) and VMD 1.9 (Humphrey et al, 1996) for
154 cryo-EM guided molecular dynamics simulations. All MDFF simulations were performed with implicit solvent, secondary structural restraints, a g-scale factor of 0.3, and the
CHARMM force field. The 2-mer and 3-mer simulations involved extensive minimization (20,000 steps), followed by a 500 ps molecular dynamics phase, another minimization phase (40,000 steps), followed by a second 100 ps molecular dynamics phase, and final minimization (20,000 steps). The 1-mer models underwent less extensive simulations with initial minimization (20,000 steps), a 100 ps molecular dynamics phase, and final minimization (20,000 steps). Nonbonded and total potential energies were calculated using the NAMD Energy plugin. Simulated density maps were generated with the Molmap command in UCSF Chimera (Pettersen et al, 2004). To simulate the 3-mer density from the final MDFF refined MPER coordinates, Molmap was run with three overlapping copies of the helical bundle related by 0°, 120° and 240°. All molecular graphics figures were made with UCSF Chimera. The MDFF simulations were run on
Case Western Reserve University’s High Performance Computing Cluster.
RESULTS:
CRYO-EM STRUCTURE OF ADV VECTOR WITH CAPSID INCORPORATED
MPER PEPTIDE
A recombinant type 5 adenovirus vector, Ad-HVR2-GP41-L15, was generated with a 24 amino acid MPER insertion within hypervariable region 2 (HVR2) of the hexon capsid protein (Matthews et al, 2010). The rationale for selecting a portion of MPER for incorporation was based on the fact that the gp41 envelope protein ectodomain is a target of three broadly neutralizing anti-HIV-1 antibodies (Ofek et al, 2004). The selection of
155 the HVR2 insertion site within hexon was made after consideration of both the HVR2 and HVR5 sites, both of which were established as technically feasible capsid incorporation sites (Matthews et al, 2008; Wu et al, 2005). HVR2 sites are immunogenic and near the top of the hexon trimer (Roberts et al, 2006). Based on atomic structural information for the intact AdV virion (Liu et al, 2010; Reddy et al, 2010), we reasoned during the design phase that insertions at this site should be exposed on the viral capsid surface. A purified sample of Ad-HVR2-GP41-L15 was preserved on cryo-EM grids and imaged with an FEI Polara microscope (300 kV, FEG). Over 5,000 cryo-EM particle images were processed to generate a structure at 9 Å resolution (FSC 0.5) for the icosahedral capsid (Fig. 5.1A). MPER density is observed between adjacent hexons in the capsid. The size and shape of the MPER density varies as seen in the asymmetric unit of the capsid (Fig. 5.1B). There are four hexon trimers in the asymmetric unit (H1-H4) displaying in total 12 MPER insertions, each with a unique environment within the capsid.
At the icosahedral 3-fold axis, MPER insertions from three neighboring hexons meet and form one merged density region. A similar 3-way interaction between MPER insertions is observed at a position of local 3-fold symmetry between hexons H2, H3 and
H4. Smaller regions of density are also observed between pairs of hexons, where two
MPER insertions are displayed in close proximity (Fig. 5.1B). These 3-mer and 2-mer
MPER interactions account for 10 of the 12 insertion sites within the asymmetric unit. At the remaining two sites, which are near the icosahedral 2-fold axis and near the penton base, density is observed for isolated, non-interacting (1-mer) MPER peptides.
156
ALPHA-HELICES ARE OBSERVED WITHIN THE ADV CAPSID AND FOR ONE
MPER INSERTION
In subnanometer (<10 Å) resolution cryo-EM structures, well-ordered α-helices should be resolved as density rods. Indeed, α-helices within hexon trimers are clearly resolved as rods within the cryo-EM structure of Ad-HVR2-GP41-L15 (Fig. 5.2). The
MPER epitope is known to adopt numerous conformations, some α-helical and some extended, and its structural plasticity may be important in facilitating viral membrane fusion. A crystal structure of the MPER region stabilized with an isoleucine zipper motif reveals a parallel triple-stranded coiled coil (Liu et al, 2009). An NMR structure of an
MPER segment in the presence of DPC micelles shows a kinked conformation with two separate helical segments (Sun et al, 2008). We observe a well resolved density rod, indicative of an α-helix, at the 1-mer site near the icosahedral 2-fold axis of the capsid
(Fig. 5.2). At the other 1-mer site near the penton base, some density for MPER is observed along the side of hexon, but there is not a clearly resolved α-helix. The strength of the cryo-EM density varies among the 2-mer and 3-mer sites, indicating different levels of conformational heterogeneity, but none of these regions have resolved α-helices.
MPER CONFORMATION IS CONSTRAINED BY THE ADV CAPSID AT ONE
INSERTION SITE
The variation in the MPER cryo-EM density at the 12 unique insertion sites is a reflection of diverse local environments created by the hexon and penton base packing arrangement within the AdV capsid. At the icosahedral 2-fold axis, hexon packing restricts the MPER insertion to within a narrow cavity (Fig. 5.3). The cryo-EM structure
157 shows a well resolved density rod, which suggests that the insertion at this site adopts a rigid helical structure with limited flexibility. Interactions with other MPER insertions are precluded at this site. To further interpret the structure, we used cryo-EM guided molecular dynamics simulations. We built initial atomic models based on the available
MPER structures and secondary structure prediction for the inserted sequence (Fig. S5.2).
These models served as input for Molecular Dynamics Flexible Fitting (MDFF) simulations (Trabuco et al, 2008). The MDFF simulation for the 1-mer site near the icosahedral 2-fold axis reveals that an α-helix for the majority of the 24 amino acid
MPER sequence fits within the narrow cavity between hexons.
STRONG HELICAL INTERACTIONS BETWEEN MPER INSERTIONS AT THE 3-
MER SITES
Analysis of the cryo-EM density at the 3-mer positions indicates that a stable interaction between MPER insertions occurs as the observed MPER density is as strong as the hexon density in the AdV capsid. Based on the crystal structure of the MPER region stabilized with an isoleucine zipper motif (Liu et al, 2009), we speculated that 3 nearby MPER insertions might interact as a parallel triple-stranded coiled coil. To test this idea, we built an atomic model for three interacting MPER α-helices based on the parallel coiled coil crystal structure (Liu et al, 2009). MDFF simulations were performed with the atomic model docked into cryo-EM density. Various initial helical interfaces
(with +/−10°, 20°, 90°, and 180° rotations about the helical axis) were tested to optimize the helical packing arrangement. The selected final model for the triple helical bundle has a minimum in both the Nonbonded and Total Potential energy terms calculated between
158 the helical MPER residues (Table S5.1). The MDFF refined model for the MPER interaction at the icosahedral 3-fold axis fits within the cryo-EM density and displays a reasonable sidechain packing arrangement (Fig. 5.4). This model fits equally well into the density at the other unique 3-mer site within the AdV capsid (between hexons H2, H3 and H4). Simulated density based on the MDFF refined model closely resembles the cryo-EM density after 3-fold averaging (Fig. S5.3). Modeling indicates that the length and flexibility of the linker regions of the MPER insertion would allow the triple helical bundle to tilt in various directions. This would result in an average representation in the cryo-EM structure and explain why resolved density rods for α-helices are not observed at the 3-mer sites.
TRANSIENT INTERACTIONS BETWEEN MPER INSERTIONS AT THE 2-MER
SITES
The cryo-EM density at the three 2-mer sites within the asymmetric unit is variable with the strongest density region between hexons H3 and H4 of a neighboring asymmetric unit (Fig. 5.1B). All of the 2-mer density regions are weaker than those at the
3-mer sites. We built an initial atomic model for two interacting MPER α-helices based on the MDFF refined 3-mer model. However, during MDFF simulations the two helices drifted apart and ended up as only weakly interacting (Fig. 5.5). Various starting models docked within the strongest cryo-EM 2-mer density were tested in MDFF, but the final
Nonbonded and Total Potential energy terms were not as favorable as for the 3-mer model. The density differences observed between the three unique 2-mer sites in the cryo-EM structure can be explained by variation in distances between the insertion sites
159 of neighboring hexons (Table S5.2). The closer the insertion sites are, the weaker the cryo-EM density is in that region. This implies that if the insertion sites are too close then favorable interactions between neighboring MPER sequences are less likely to be formed.
Taken together, the cryo-EM and modeling results indicate that the MPER interactions at
2-mer sites are transient and weak in nature.
CONFORMATIONAL FLEXIBILITY OF MPER NEXT TO PENTON BASE
The MPER insertion closest to the penton base has a unique local environment within the capsid. We observe a small amount of MPER density for this insertion along the side of the peripentonal hexon facing the penton base. However, the density is not well defined, indicating a low affinity binding site for MPER on the side of the hexon.
Therefore, we speculate that the MPER insertion at this site is the most conformationally flexible within the AdV capsid.
DISCUSSION:
A hybrid cryo-EM and molecular dynamics approach enabled us to model the conformation of MPER presented on an engineered AdV vaccine vector that shows a promising immunological response in mice (Matthews et al, 2010). There are twelve unique hexon HVR2 insertion sites within the AdV capsid that result in different MPER conformations, including well resolved individual α-helices at the icosahedral 2-fold axis, interacting α-helices, and partially extended forms. At the insertion site closest to the icosahedral 2-fold axis, a well resolved density rod is observed for MPER, strongly suggesting that the MPER sequence is adopting an α-helical conformation at this position
160 in the capsid. This MPER insertion site is constrained by the AdV capsid with the helix observed within a narrow cavity between two hexons. We speculate that the MPER sequence presented at this confined site on the AdV capsid would not be accessible to antibody molecules. The observation of strong density at the 3-mer sites indicates a stable interaction between MPER regions at these capsid locations. MDFF simulations with atomic models based on a crystal structure of a partial MPER sequence that forms a parallel triple helical bundle reveal that a favorable helical packing arrangement can be achieved for the MPER sequence in the Ad-HVR2-GP41-L15 vector. The MDFF refined
3-mer model has the aromatic sidechains buried at the helical interfaces. The cryo-EM density at the 3-mer sites can be reasonably well simulated by 3-fold averaging the final helical bundle model. Given the strength of the cryo-EM density at the 3-mer sites it seems likely that these MPER regions are stably associating and thus are unlikely to be available for interaction with antibodies.
In comparison, the cryo-EM density at the 2-mer sites is more variable. MDFF simulations show a more limited interaction between MPER regions where only two insertions can interact. The 1-mer insertion site next to the penton base leads to little observable cryo-EM density for the MPER region. Although we cannot directly observe flexibility of a peptide region in a cryo-EM structure, the presence of small amino acids
(Gly and Ser) in the linkers, and the weak nature of the density, both suggest flexible
MPER presentation at these sites. The 2-mer insertions could be in equilibrium between interacting with a neighboring insertion and forming an extended and flexible conformation above the capsid. Similarly, the 1-mer sites next to the penton base could be in equilibrium between interacting with hexon and forming an extended conformation
161
(Fig. 5.6). The presence of extended MPER conformations on the Ad-HVR2-GP41-L15 vector is supported by the observation that this vector binds the human HIV monoclonal antibody 2F5 (Matthews et al, 2010), which recognizes an extended 14 amino acid region of MPER.
A major obstacle to the rational design of viruses as vaccine vectors using the capsid-incorporation strategy is the lack of understanding of the rules governing the relationship between epitope presentation and host immune response. We reasoned that a structural analysis of the Ad-HVR2-GP41-L15 vector would lead to ideas on how to modify the next generation of AdV-based vectors and potentially produce stronger immune responses. Consideration of the modeled MPER structure at the most constrained position within the AdV capsid led us to propose several modifications that should be structurally tolerated at the hexon HVR2 site (Fig. 5.7). We were careful to avoid increasing the insertion length so much that it might interfere with hexon-hexon packing interactions during capsid assembly. One possible modification would be to add a maximum of 3aa to the N-terminal linker, while keeping the MPER insertion sequence and the C-terminal linker the same (Fig. 5.7B). Modeling indicates that this longer insertion would still fit at the constrained 1-mer location within the capsid, while enhancing the flexibility of the MPER insertions at the 1-mer site next to the penton base and at the 2-mer sites. A second possibility is to increase the C-terminal linker by up to
4aa (Fig. 5.7C). Alternatively, both the N- and C-terminal linkers could potentially be lengthened by 2aa (Fig. 5.7D). All three of these proposed modifications should enhance the exposure of the heterologous MPER epitope on the capsid surface and potentially enhance the generation of antibodies. In addition, the lengthened, flexible linkers could
162 improve antigen processing and subsequent presentation on helper T-cells (Mirano-
Bascos et al, 2008). The fourth proposed vector modification is a swap of the N- and C- terminal linkers (Fig. 5.7E), which modeling indicates would be structurally permissible at all 12 insertion sites within the asymmetric unit. This would put the longer, and presumably more flexible, linker before the MPER epitope and would subtly alter antigen presentation on the capsid as well as potentially enhance antigen processing (Dai et al,
2001; Mirano-Bascos et al, 2008).
The structural information gained from this study could be applicable to the design of vectors with alternative HIV epitopes within the AdV hexon HVR2 site. More than a dozen broadly neutralizing anti-HIV monoclonal antibodies have been isolated (Kwong et al, 2011). Once they are structurally characterized in complex with their cognate epitopes, this information could be used to design the optimal epitope insertion, similar in length to the MPER α-helix, for display at the AdV hexon HVR2 sites. Moreover, future studies are in development to test which of the predicted modifications will lead to increased immunogenicity. In summary, this cryo-EM study provided a way to visualize and model the presentation of a heterologous epitope incorporated within an AdV capsid and suggested potential ways to optimize MPER presentation on the AdV capsid.
ACKNOWLEDGEMENTS:
The authors gratefully acknowledge Robert Kim and Dewight Williams for their assistance with cryo-EM data acquisition and Qiana Matthews for providing the original sample. We also thank the Advanced Computing Center for Research and Education
163
(ACCRE) at Vanderbilt University and the High Performance Computing Cluster
(HPCC) at Case Western Reserve University for computational support. The cryo-EM structure has been deposited in the EM Data Bank with accession number EMD-5511.
AUTHOR CONTRIBUTIONS:
Conceived and designed the experiments: JLB DTC PLS. Performed the experiments:
JWF TLF. Analyzed the data: JWF TFL PLS. Wrote the paper: JWF TLF JLB DTC PLS.
Helped analyze perspective with respect to virology and immunology: NM IPD EAK.
164
Figure 5.1 Cryo-EM structure of the Ad-HVR2-GP41-L15 vector at subnanometer resolution. (A) Full virion viewed along an icosahedral 3-fold axis. Density assigned to the MPER insertion within the top facet is colored in red and gold, with red representing the MPER density within one asymmetric unit. This AdV vector is based on human AdV type 5, which has long and flexible fibers (>300 Å). Only short portions of the fiber (out to a radius of 463 Å) have been reconstructed (3 fibers are indicated with arrows). (B)
Enlarged view with the 12 MPER density regions within one asymmetric unit numbered
1–12. Interacting MPER density regions from adjacent asymmetric units are numbered in parentheses. The four hexons are labeled H1–H4 and the penton base is labeled P. The icosahedral 2- and 3-fold axes are indicated with oval and triangle symbols respectively.
Scale bars represent 100 Å.
165
Figure 5.2 Cryo-EM density showing α-helices for two hexons and two MPER insertions.
(A) Coordinates from the final frame of an MDFF simulation show α-helices of both hexon and MPER docked within density rods of the cryo-EM structure (gray). (B)
166
Perpendicular view showing one of the two MPER density rods. (C) Additional view showing the second of the two MPER density rods at the icosahedral 2-fold axis with the final MDFF coordinates, which were not icosahedrally constrained. The isosurface threshold levels were chosen to highlight the well resolved density rods within each panel. Ribbon representations are shown for the hexon backbone (purple and blue), the
MPER sequence (red), and the linker regions (green). The icosahedral 2-fold axis is indicated with an oval symbol.
167
Figure 5.3 MPER insertion within a narrow cavity between hexons at the icosahedral 2- fold axis. Density for two hexons (blue and purple) simulated from the final MDFF frame shown surrounding experimentally determined cryo-EM density (gray) for one MPER insertion. A ribbon representation of the final MDFF model of the MPER insertion is overlaid, colored as in Figure 5.2. A tilted view is shown to emphasize the confined nature of the cavity between the hexons.
168
Figure 5.4 MPER forms a stable helical bundle at 3-mer sites. (A) Overlay of MDFF refined model (ribbons) and the cryo-EM density (gray) at the icosahedral 3-fold axis.
The ribbon coloring scheme is the same as in Figure 5.2. (B) Enlarged view of the helical packing interface viewed along the bundle axis with aromatic sidechains displayed
(gold).
169
Figure 5.5 MPER interactions at 2-mer sites are weak and transient. Overlay of MDFF refined model (ribbons) and the cryo-EM density (gray) at the strongest 2-mer site
(between hexons H3 and H4 of a neighboring asymmetric unit). The ribbon coloring scheme is the same as in Figure 5.2.
170
Figure 5.6 Alternate model conformations for MPER next to penton base. (A) Simulated density for hexon (blue) shown with a ribbon representation of an MDFF model of the
MPER in an α-helical conformation or (B) extended conformation. The 14 amino acid
2F5 epitope is shown with a thicker ribbon in the extended model. The ribbons are colored as in Figure 5.2.
171
Figure 5.7 Proposed vector modifications for optimizing MPER presentation at the AdV hexon HVR2 site. (A) Schematic representation of the MPER insertion incorporated within the Ad-HVR2-GP41-L15 vector. The hexon capsid protein (gray) is shown together with the N-terminal 3aa linker (blue), the MPER peptide (green), and the C- terminal 10aa linker (red). The protein-protein interface between hexons in the AdV capsid is represented by a hashed region. (B-E) Based on the structural analysis of the
Ad-HVR2-GP41-L15 vector, four possible modifications are proposed which include (B) extending the N-terminal linker by 3aa, (C) extending the C-terminal linker by 4aa, (D) extending both the N- and C-terminal linkers by 2aa, and (E) swapping the N- and C- terminal linkers.
172
Table S5.1 Optimization of the helical interface at a 3-mer site with molecular dynamics flexible fitting.
173
Table S5.2 Distances between hexon insertion sites at 2-mer sites.
174
Figure S5.1 Resolution assessment of the Ad-HVR2-GP41-L15 cryo-EM structure. The
Fourier shell correlation (FSC) curve is calculated for the icosahedral capsid (radii 325–
460 Å). The resolution as assessed by the FSC = 0.5 criterion is 8.7 Å.
175
Figure S5.2 Secondary structure prediction for the inserted MPER and linker sequence.
Average of the predicted α-helical propensity as a function of the amino acid sequence, including the 24aa MPER and the N- and C-terminal linkers. The prediction is an average of the results from Jufo, SAM, and Psi-Pred. The linkers are shown in green, and the
MPER sequence is shown in purple for the region modeled as extended, and red for the region modeled as α-helical. The MPER and linker sequence was inserted within the hexon HVR2 regions after Val-188 and before Pro-193.
176
Figure S5.3 Comparison of density at the icosahedral 3-fold axis with simulated hexon/MPER density. (A) Top view of the cryo-EM density contoured to show the
MPER insertion density between three hexons. (B) Corresponding view of the simulated density for three hexons (blue) and the MPER residues of the MDFF refined 3-mer model
(red). The simulated hexon density is filtered to 8 Å resolution. The simulated MPER density is filtered to 12 Å resolution and 3-fold averaged to account for the observation that the flexible linkers would presumably allow the helical bundle to tilt in three
177 different directions. Ribbon representations are shown for the hexon backbone (blue), the
MPER sequence (red), and the linker regions (green). (C and D) Perpendicular views.
178
CHAPTER 6: Summary of discoveries and future directions
6.1 Insights on immune recognition of adenoviruses
The work presented in this thesis focused on the structural biology underlying immune recognition of AdV. By using a hybrid structural approach, we investigated the interplay between viral and host factors in shaping the outcome of AdV infection. These host factors included HD5, FX, and FVII. We also characterized how a specific HIV epitope, the gp41 MPER peptide, is presented on the capsid surface of an AdV vector designed as a candidate HIV vaccine. This section contains a summary of our discoveries as well as future directions.
6.1.1 Human alpha defensin 5 opposes capsid disassembly
Studies on human alpha defensins, such as HD5, have demonstrated that they directly act on both host cells and pathogens to prevent infections; however, how they work, particularly against non-enveloped viruses, is not well understood. We have sought to help answer this question by studying how HD5 neutralizes non-enveloped AdV. Our goal has been to describe the HD5 anti-viral mechanism. An earlier study from the
Stewart lab identified the capsid proteins critical for HD5 neutralization. This strongly showed that the HD5 critical binding site is located in a region spanning the fiber and penton base proteins. In our new study, we showed that HD5 binding has opposing effects on capsid disassembly and infectivity. We determined cryo-EM structures of HD5 complexed with both neutralization-sensitive and –resistant AdV chimeras. Models were built for the vertex regions of these virus chimeras with monomeric and dimeric forms of
179
HD5 in various initial orientations. Cryo-EM guided MDFF was used to restrain the majority of these vertex models in well-defined cryo-EM density. The RGD-containing penton base loops of both the neutralization-sensitive and -resistant virus chimeras were predicted to be intrinsically disordered, and little cryo-EM density was observed for them. In simulations these loops from the sensitive virus chimera, interacted with HD5, bridged the penton base and fiber proteins, and provided significant stabilization with a three-fold increase in the intermolecular non-bonded interactions of the vertex complex.
For the resistant virus chimera, simulations revealed fewer bridging interactions and reduced stabilization by HD5. Thus, we demonstrated that HD5 can provide significant stabilization to the AdV vertex region formed by fiber and penton base, and that this stabilization would likely block subsequent steps in disassembly that lead to the release of protein VI and disruption of the endosomal membrane during cell entry. In addition, the HD5 stabilization effect is dependent on the presence of a conformationally flexible binding pocket that provides numerous alternative ways for HD5 monomers and dimers to interact with the vertex region.
In terms of future studies, we plan to investigate how HD5 neutralizes non- enveloped human papillomavirus (HPV). For HPV, time course experiments suggested that an early step in infection was blocked, and microscopic examination of cells infected in the presence of inhibitory concentrations of HD5 showed that the virus appeared to be trapped in endosomes (Parker et al, 2006). These results are strikingly similar to how
HD5 blocks AdV infection, where inhibitory concentrations of HD5 prevent virus escape from endosomes. Therefore, it is possible that defensins use a common mode of neutralization for non-enveloped viruses. HPV has an icosahedrally shaped capsid that is
180 roughly 550 Å in diameter. The capsid is composed of two structural proteins, the major protein L1 and the minor protein L2. HPVs infect cells via a clathrin-dependent pathway.
Endosomal acidification triggers L1 capsid disassembly, releasing L2 which facilitates viral escape from late endosomes via sorting nexin 17 (Bergant Marusic et al, 2012).
Sorting nexin 17 (SNX17) is a key regulator of endosomal recycling and has been identified as a strong interacting partner of HPV L2 (Bergant Marusic et al, 2012). The interaction involves a highly conserved SNX17 consensus binding motif that is present in the majority of HPV L2 proteins. While it is known that the SNX 17 interaction is essential for viral escape from the endosome, details of how this process occurs remain unknown. Our hypothesis is that HD5 binding has a stabilizing effect on the viral capsid and, as a result, blocks disassembly steps required for release of L2 and infection. To test this hypothesis, we will examine how HD5 interacts with HPV using primarily cryo-EM.
Visualization by cryo-EM should indicate whether or not HD5 bridges adjacent capsid subunits and will allow us to build a model for neutralization. We can then test the functional predictions of our model by infectivity studies, site-directed mutagenesis, and molecular dynamics calculations. Finally, we can compare the results of this study with the results from the AdV study to see if this is indeed a general mechanism against non- enveloped viruses. It is worth noting that there are several important questions that remain to be answered, which are not addressed in our future study, but are necessary for fitting virus-defensin interactions into the larger context of the innate immune response.
Do non-enveloped viruses such as AdV and HPV induce the release of appropriate defensins at sites of infection via their direct interaction with cells or does this occur indirectly by stimulating the production of inflammatory mediators? Does stabilizing the
181
AdV capsid and preventing release of viral DNA reduce TLR9 activation and therefore, the overall magnitude of the inflammatory response? If so, is that good for the virus or the host? It will be exciting to learn the answers to these questions and much more about this virus-host interaction from future studies.
6.1.2 A role for FX in adenovirus innate immunity
Host FX, better known for playing an essential role in the clotting process, also works as a determinant of viral tropism. While in the blood, type C adenoviruses, including AdV types 5 and 2, bind FX with high affinity and this interaction contributes to virus sequestration in the liver by facilitating transduction of hepatocytes. In a previous study from the Stewart lab, we demonstrated that the major AdV type 5 capsid protein, hexon, binds FX with an affinity of 229 pM (Kalyuzhniy et al, 2008). Here, we have dissected the AdV-FX interaction and discovered a novel link between this interaction and the innate immune response. A 9 Å cryo-EM structure was determined for AdV type
5 with FX bound. Cryo-EM guided MDFF simulations were performed to characterize the interaction interface between the viral protein hexon and FX. Based on our structural data, a single amino acid substitution was introduced into hexon (T425A) that completely abrogated FX binding. This FX-binding-ablated virus failed to infect hepatocytes when injected in mice. Because hepatocyte transduction contributes to the induction of anti-
AdV inflammatory and immune responses, we examined the effect of FX ablation on immune sensing in mice. Immune responses were assessed for both AdV-FX and FX- binding-ablated virus. This led to the novel observation that FX becomes a pathogen associated molecular pattern upon binding to the surface of the virus. FX decorated virus
182 triggers a TLR4-dependent innate immune response. Thus, our work demonstrated a link between coagulation and innate immunity. In a follow-up study, we performed an in- depth analysis of FVII interaction with the AdV capsid. FVII and FX are structurally homologous and each bind to hexon via their GLA domains. They both contain a serine protease domain for binding to HSPGs on hepatocytes. However, FVII does not contribute to transduction of hepatocytes in vivo. Surface plasmon resonance demonstrated that FVII binding to hexon is inhibited by the same mutation that prevents
FX attachment. Cryo-EM and MDFF modeling revealed that FVII and FX adopt distinct binding orientations relative to hexon and therefore, are presented differently on the capsid surface. Furthermore, we found that neighboring FVII molecules interact via their serine protease domains and bury potential HSPG receptor binding residues within the dimer interface. In contrast, our earlier cryo-EM study of the AdV-FX interaction showed no evidence of FX dimerization. Our study implicates serum factor binding orientation and dimerization as playing major roles in defining the differential infectivity of AdV-
FVII and AdV-FX complexes.
Important questions remain to be answered about the AdV-FX interaction. We demonstrated how FX binding to AdV activates innate immunity. At first glimpse, it appeared that we had discovered a new host defense mechanism. However, a new study provided evidence that FX is being utilized by AdV type 5 to protect against immune attack (Xu et al, 2013). The authors observed FX-independent liver transduction in antibody deficient mice. Injection of purified natural IgM in mice that have T cells but lack B cells restricted liver transduction in cases where AdV was unable to bind FX.
Furthermore, liver transduction was unaffected in warfarinized mice that lacked
183 complement components C1q and C4. In vitro experiments revealed that IgM mediates complement-dependent (C1q and C4 dependent) neutralization of AdV in the absence of
FX. However, FX protected the virus from neutralization. Also, complement did not destroy the virus but rather impaired transduction of cells. These results are rather astonishing considering previous data from immune-competent mice, where warfarin treatment of mice ablated RFP transgene expression in hepatocytes after injection of an
AdV type 5 RFP containing vector (Kalyuzhniy et al, 2008). In this older study, ablation was completely reverted by restoration of physiological levels of FX. In the future we will investigate how IgM, Cq1 and C4 neutralize AdV. Our hypothesis is that coating of
AdV by antibody and complement prevents its ability to attach to host cells, thereby blocking productive infection. This would be consistent with the finding that complement impairs virus transduction of the liver as well as other organs (i.e. lung, kidney, spleen).
For future studies, we will continue to collaborate with Dmitry Shayakhmetov who worked with us on the AdV-FX project. There are several questions that must be addressed. How do IgM and complement proteins Cq1 and C4 neutralize AdV? How does AdV enter hepatocytes in vivo in mice that lack IgM or complement components?
Are AdV species that do not bind FX inactivated by complement in the bloodstream, or are there additional host defense mechanisms? Will IgM-mediated complement- dependent neutralization of FX-binding-ablated virus make it impossible to design AdV vectors for gene delivery to extrahepatic sites? The more we learn about AdV-host interactions the more we realize how complex these interactions are.
184
6.1.3 Capsid-incorporated HIV antigen adopts multiple conformations
AdVs are promising as vectors for vaccination against infectious diseases. They offer a serious advantage over traditional vaccines because in addition to inducing high antibody titers, they also promote cell mediated immunity. This involves the generation of antigen-specific T cells. CD8+ T cells kill infected cells, while CD4+ T cells, in addition to killing infected cells, recruit other branches of the immune system, such as antibodies, to attack target cells. By contrast, traditional vaccines only stimulate antibody responses and lack the ability to induce T cell immunity. For this reason, AdVs have been used for a variety of vaccine applications. We used cryo-EM to perform a structural analysis of a multivalent AdV vector that was used to elicit anti-HIV humoral and cell mediated immunity in mice. The vector was modified to express HIV Gag as a transgene.
Transgene expression led to an anti-HIV Gag cellular response in mice. Also, the vector contained a capsid-incorporated HIV epitope, the gp41 MPER peptide, for presentation on the surface of the virus. The MPER epitope was inserted into the top of the major viral capsid protein hexon. MPER presentation led to an anti-HIV specific humoral response in mice. A high-resolution cryo-EM structure allowed us to visualize how MPER is displayed on the surface of AdV. Cryo-EM guided modeling of MPER presentation revealed that the inserted peptide adopts multiple conformations on the surface of AdV due to the arrangement of hexons within the viral capsid. The different peptide conformations we observed included individual rigid α-helices, interacting bent α-helices, and partially extended forms. This occurred despite the fact that only a single insertion site within hexon was employed for antigen display. Taken altogether, these results demonstrate that the AdV capsid significantly influences epitope structure, flexibility and
185 accessibility, all of which are critical factors in inducing an antigen specific immune response. Our study represents an important early step for rational design of AdV-based vaccines that utilize the virus capsid as an immune-enhancing scaffold for displaying heterologous antigens.
The idea that heterologous epitopes, such as MPER, can be displayed on viral capsids for inducing epitope-specific immunity is of growing interest to the field of rational vaccine design. This approach has been tested several times using the AdV capsid. The first attempt dates back to the mid-1990s when an 8 amino acid sequence of the VP1 capsid protein of poliovirus (PV) type 3 was incorporated into the AdV structural protein hexon. Antiserum raised against the engineered vector containing polio antigen specifically recognized the VP1 capsid of PV (Crompton et al, 1994). Since then, several research groups have attempted the capsid-incorporation strategy, but progress has been slow due to a limited understanding of the principles that correlate antigen display with the ultimate, multifaceted immune response. Using cryo-EM and MDFF modeling, we provided a first look into how the AdV capsid influences the structure and presentation of a hexon-inserted heterologous epitope. This new information is useful for future studies aimed at optimizing AdV vectors for targeted heterologous epitope presentation. For example, based on our structural analysis, we suggested four capsid modifications that may lead to a more robust MPER-specific humoral immune response, while remaining structurally tolerated. These modifications involve adding length to the flanking sequences that attach the heterologous epitope to the viral capsid to increase surface exposure. Our hypothesis is that AdV scaffolds with more flexibly held MPER epitopes will elicit greater immune responses. This idea is supported by recent studies
186 that correlate epitope flexibility with immunogenicity. In one elegant study, epitope scaffolds possessing a range of flexibilities were designed, structurally characterized, and tested in mice (Ofek et al, 2010). In doing so, it was observed that a more flexibly displayed MPER epitope generates a significantly higher immune response than a more rigidly constrained form. Understanding and controlling this process is critical to the success of the capsid-incorporation strategy. We envision that cryo-EM structural characterization of our proposed modifications along with assessment of other capsid- incorporated heterologous epitopes will play a critical role in harnessing the full potential of the AdV capsid as a scaffold for antigen display.
6.1.4 Conclusion
The projects included in this thesis involved primarily structural studies to develop a more detailed description of AdV-host interactions during infection.
Specifically, high resolution cryo-EM combined with atomic resolution structures from
X-ray crystallography and newly developed computational modeling tools greatly enhanced our understanding of how specific regions of the AdV capsid are recognized by immune and coagulation factors. This approach also enabled us to examine how a specific HIV epitope is displayed on the capsid of an AdV vector designed as a candidate
HIV vaccine. Besides providing a better understanding of AdV infection biology, the information gleaned from these hybrid structural studies may inform new strategies for controlling AdVs as both pathogens and therapeutic agents.
187
REFERENCES
Adrian M, Dubochet J, Lepault J, McDowall AW (1984) Cryo-electron microscopy of viruses. Nature 308: 32-36
Akiyama M, Roelvink P, Einfeld D, Kovesdi I, King CR (2003) Effect of ablating CAR and integrin binding on the biodistribution and cellular localization of adenovirus vectors following intravenous or intraperitoneal delivery. Mol Ther 7: S173-S174
Alba R, Bradshaw AC, Parker AL, Bhella D, Waddington SN, Nicklin SA, van Rooijen N, Custers J, Goudsmit J, Barouch DH, Mcvey JH, Baker AH (2009) Identification of coagulation factor (F)X binding sites on the adenovirus serotype 5 hexon: effect of mutagenesis on FX interactions and gene transfer. Blood 114: 965-971
Alemany R, Curiel DT (2001) CAR-binding ablation does not change biodistribution and toxicity of adenoviral vectors. Gene therapy 8: 1347-1353
Ariyawansa JP, Tobin JO (1976) Fluorescent antibody responses to adenoviruses in humans. Journal of clinical pathology 29: 411-416
Arnberg N, Pring-Akerblom P, Wadell G (2002) Adenovirus type 37 uses sialic acid as a cellular receptor on Chang C cells. Journal of virology 76: 8834-8841
Babiss LE (2003) In memoriam: Harold S. Ginsberg (1917-2003). Archives of virology 148: 1655-1657
Baker AH, Mcvey JH, Waddington SN, Di Paolo NC, Shayakhmetov DM (2007) The influence of blood on in vivo adenovirus bio-distribution and transduction. Mol Ther 15: 1410-1416
Banner DW, DArcy A, Chene C, Winkler FK, Guha A, Konigsberg WH, Nemerson Y, Kirchhofer D (1996) The crystal structure of the complex of blood coagulation factor VIIa with soluble tissue factor. Nature 380: 41-46
Bastian A, Schafer H (2001) Human alpha-defensin 1 (HNP-1) inhibits adenoviral infection in vitro. Regulatory peptides 101: 157-161
188
Bergant Marusic M, Ozbun MA, Campos SK, Myers MP, Banks L (2012) Human papillomavirus L2 facilitates viral escape from late endosomes via sorting nexin 17. Traffic 13: 455-467
Bergelson JM, Cunningham JA, Droguett G, Kurt-Jones EA, Krithivas A, Hong JS, Horwitz MS, Crowell RL, Finberg RW (1997) Isolation of a common receptor for Coxsackie B viruses and adenoviruses 2 and 5. Science 275: 1320-1323
Bewley MC, Springer K, Zhang YB, Freimuth P, Flanagan JM (1999) Structural analysis of the mechanism of adenovirus binding to its human cellular receptor, CAR. Science 286: 1579-1583
Bottcher B, Wynne SA, Crowther RA (1997) Determination of the fold of the core protein of hepatitis B virus by electron cryomicroscopy. Nature 386: 88-91
Bradley RR, Lynch DM, Iampietro MJ, Borducchi EN, Barouch DH (2012) Adenovirus serotype 5 neutralizing antibodies target both hexon and fiber following vaccination and natural infection. Journal of virology 86: 625-629
Brogden KA (2005) Antimicrobial peptides: pore formers or metabolic inhibitors in bacteria? Nature reviews Microbiology 3: 238-250
Burmeister WP, Guilligay D, Cusack S, Wadell G, Arnberg N (2004) Crystal structure of species D adenovirus fiber knobs and their sialic acid binding sites. Journal of virology 78: 7727-7736
Cardoso RM, Brunel FM, Ferguson S, Zwick M, Burton DR, Dawson PE, Wilson IA (2007) Structural basis of enhanced binding of extended and helically constrained peptide epitopes of the broadly neutralizing HIV-1 antibody 4E10. Journal of molecular biology 365: 1533-1544
Chan KY, Trabuco LG, Schreiner E, Schulten K (2012) Cryo-electron microscopy modeling by the molecular dynamics flexible fitting method. Biopolymers 97: 678-686
Chen CY, May SM, Barry MA (2010) Targeting Adenoviruses with Factor X-Single- Chain Antibody Fusion Proteins. Human gene therapy 21: 739-749
Chen RF, Lee CY (2013) Adenoviruses Types, Cell Receptors and Local Innate Cytokines in Adenovirus Infection. International reviews of immunology
189
Chu H, Pazgier M, Jung G, Nuccio SP, Castillo PA, de Jong MF, Winter MG, Winter SE, Wehkamp J, Shen B, Salzman NH, Underwood MA, Tsolis RM, Young GM, Lu W, Lehrer RI, Baumler AJ, Bevins CL (2012) Human alpha-defensin 6 promotes mucosal innate immunity through self-assembled peptide nanonets. Science 337: 477-481
Cichon G, Boeckh-Herwig S, Schmidt HH, Wehnes E, Muller T, Pring-Akerblom P, Burger R (2001) Complement activation by recombinant adenoviruses. Gene therapy 8: 1794-1800
Clarke BE, Newton SE, Carroll AR, Francis MJ, Appleyard G, Syred AD, Highfield PE, Rowlands DJ, Brown F (1987) Improved immunogenicity of a peptide epitope after fusion to hepatitis B core protein. Nature 330: 381-384
Conway JF, Cheng N, Zlotnick A, Wingfield PT, Stahl SJ, Steven AC (1997) Visualization of a 4-helix bundle in the hepatitis B virus capsid by cryo-electron microscopy. Nature 386: 91-94
Crompton J, Toogood CI, Wallis N, Hay RT (1994) Expression of a foreign epitope on the surface of the adenovirus hexon. The Journal of general virology 75 ( Pt 1): 133-139
Dai G, Steede NK, Landry SJ (2001) Allocation of helper T-cell epitope immunodominance according to three-dimensional structure in the human immunodeficiency virus type I envelope glycoprotein gp120. The Journal of biological chemistry 276: 41913-41920 de Mezerville MH, Tellier R, Richardson S, Hebert D, Doyle J, Allen U (2006) Adenoviral infections in pediatric transplant recipients: a hospital-based study. The Pediatric infectious disease journal 25: 815-818
Demirkhanyan LH, Marin M, Padilla-Parra S, Zhan C, Miyauchi K, Jean-Baptiste M, Novitskiy G, Lu W, Melikyan GB (2012) Multifaceted mechanisms of HIV-1 entry inhibition by human alpha-defensin. The Journal of biological chemistry 287: 28821- 28838
Dicks MD, Spencer AJ, Edwards NJ, Wadell G, Bojang K, Gilbert SC, Hill AV, Cottingham MG (2012) A novel chimpanzee adenovirus vector with low human seroprevalence: improved systems for vector derivation and comparative immunogenicity. PloS one 7: e40385
190
Doronin K, Flatt JW, Di Paolo NC, Khare R, Kalyuzhniy O, Acchione M, Sumida JP, Ohto U, Shimizu T, Akashi-Takamura S, Miyake K, MacDonald JW, Bammler TK, Beyer RP, Farin FM, Stewart PL, Shayakhmetov DM (2012) Coagulation factor X activates innate immunity to human species C adenovirus. Science 338: 795-798
Dubochet J, Booy FP, Freeman R, Jones AV, Walter CA (1981) Low temperature electron microscopy. Annual review of biophysics and bioengineering 10: 133-149
Dubochet J, Lepault J, Freeman R, Berriman JA, Homo JC (1982) Electron-Microscopy of Frozen Water and Aqueous-Solutions. J Microsc-Oxford 128: 219-237
Duffy MR, Bradshaw AC, Parker AL, Mcvey JH, Baker AH (2011) A Cluster of Basic Amino Acids in the Factor X Serine Protease Mediates Surface Attachment of Adenovirus/FX Complexes. Journal of virology 85: 10914-10919
Dugan AS, Maginnis MS, Jordan JA, Gasparovic ML, Manley K, Page R, Williams G, Porter E, O'Hara BA, Atwood WJ (2008) Human alpha-defensins inhibit BK virus infection by aggregating virions and blocking binding to host cells. The Journal of biological chemistry 283: 31125-31132
Dunker AK, Lawson JD, Brown CJ, Williams RM, Romero P, Oh JS, Oldfield CJ, Campen AM, Ratliff CM, Hipps KW, Ausio J, Nissen MS, Reeves R, Kang C, Kissinger CR, Bailey RW, Griswold MD, Chiu W, Garner EC, Obradovic Z (2001) Intrinsically disordered protein. Journal of molecular graphics & modelling 19: 26-59
Eigenbrot C, Kirchhofer D, Dennis MS, Santell L, Lazarus RA, Stamos J, Ultsch MH (2001) The factor VII zymogen structure reveals reregistration of beta strands during activation. Structure 9: 627-636
Enders JF, Bell JA, Dingle JH, Francis T, Jr., Hilleman MR, Huebner RJ, Payne AM (1956) Adenoviruses: group name proposed for new respiratory-tract viruses. Science 124: 119-120
Fabry CM, Rosa-Calatrava M, Conway JF, Zubieta C, Cusack S, Ruigrok RW, Schoehn G (2005) A quasi-atomic model of human adenovirus type 5 capsid. The EMBO journal 24: 1645-1654
Fitzgerald JC, Gao GP, Reyes-Sandoval A, Pavlakis GN, Xiang ZQ, Wlazlo AP, Giles- Davis W, Wilson JM, Ertl HC (2003) A simian replication-defective adenoviral recombinant vaccine to HIV-1 gag. Journal of immunology 170: 1416-1422
191
Flatt JW, Fox TL, Makarova N, Blackwell JL, Dmitriev IP, Kashentseva EA, Curiel DT, Stewart PL (2012) CryoEM visualization of an adenovirus capsid-incorporated HIV antigen. PloS one 7: e49607
Franklin RM, Pettersson U, Akervall K, Strandberg B, Philipson L (1971) Structural proteins of adenovirus. V. Size and structure of the adenovirus type 2 hexon. Journal of molecular biology 57: 383-395
Gaggar A, Shayakhmetov DM, Lieber A (2003) CD46 is a cellular receptor for group B adenoviruses. Nature medicine 9: 1408-1412
Ganz T (2003) Defensins: antimicrobial peptides of innate immunity. Nature reviews Immunology 3: 710-720
Ginsberg HS, Badger GF, Dingle JH, Jordan WS, Jr., Katz S (1955) Etiologic relationship of the RI-67 agent to acute respiratory disease (ARD). The Journal of clinical investigation 34: 820-831
Gounder AP, Wiens ME, Wilson SS, Lu W, Smith JG (2012) Critical determinants of human alpha-defensin 5 activity against non-enveloped viruses. The Journal of biological chemistry 287: 24554-24562
Greber UF, Suomalainen M, Stidwill RP, Boucke K, Ebersold MW, Helenius A (1997) The role of the nuclear pore complex in adenovirus DNA entry. The EMBO journal 16: 5998-6007
Greber UF, Willetts M, Webster P, Helenius A (1993) Stepwise dismantling of adenovirus 2 during entry into cells. Cell 75: 477-486
Grigorieff N (2007) FREALIGN: high-resolution refinement of single particle structures. Journal of structural biology 157: 117-125
Gropp R, Frye M, Wagner TO, Bargon J (1999) Epithelial defensins impair adenoviral infection: implication for adenovirus-mediated gene therapy. Human gene therapy 10: 957-964
Guex N, Peitsch MC (1997) SWISS-MODEL and the Swiss-PdbViewer: an environment for comparative protein modeling. Electrophoresis 18: 2714-2723
192
Harvey SA, Romanowski EG, Yates KA, Gordon YJ (2005) Adenovirus-directed ocular innate immunity: the role of conjunctival defensin-like chemokines (IP-10, I-TAC) and phagocytic human defensin-alpha. Investigative ophthalmology & visual science 46: 3657-3665
Hicks MJ, De BP, Rosenberg JB, Davidson JT, Moreno AY, Janda KD, Wee S, Koob GF, Hackett NR, Kaminsky SM, Worgall S, Toth M, Mezey JG, Crystal RG (2011) Cocaine analog coupled to disrupted adenovirus: a vaccine strategy to evoke high-titer immunity against addictive drugs. Mol Ther 19: 612-619
Hill CP, Yee J, Selsted ME, Eisenberg D (1991) Crystal structure of defensin HNP-3, an amphiphilic dimer: mechanisms of membrane permeabilization. Science 251: 1481-1485
Hilleman MR, Werner JH (1954) Recovery of new agent from patients with acute respiratory illness. Proceedings of the Society for Experimental Biology and Medicine Society for Experimental Biology and Medicine 85: 183-188
Hong JY, Lee HJ, Piedra PA, Choi EH, Park KH, Koh YY, Kim WS (2001) Lower respiratory tract infections due to adenovirus in hospitalized Korean children: epidemiology, clinical features, and prognosis. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America 32: 1423-1429
Huebner RJ, Rowe WP, Ward TG, Parrott RH, Bell JA (1954) Adenoidal-pharyngeal- conjunctival agents: a newly recognized group of common viruses of the respiratory system. The New England journal of medicine 251: 1077-1086
Humar A, Kumar D, Mazzulli T, Razonable RR, Moussa G, Paya CV, Covington E, Alecock E, Pescovitz MD, Group PVS (2005) A surveillance study of adenovirus infection in adult solid organ transplant recipients. American journal of transplantation : official journal of the American Society of Transplantation and the American Society of Transplant Surgeons 5: 2555-2559
Humphrey W, Dalke A, Schulten K (1996) VMD: visual molecular dynamics. Journal of molecular graphics 14: 33-38, 27-38
Ishida T, Kinoshita K (2007) PrDOS: prediction of disordered protein regions from amino acid sequence. Nucleic acids research 35: W460-464
193
Jones DT (1999) Protein secondary structure prediction based on position-specific scoring matrices. Journal of molecular biology 292: 195-202
Kairies N, Beisel HG, Fuentes-Prior P, Tsuda R, Muta T, Iwanaga S, Bode W, Huber R, Kawabata S (2001) The 2.0-A crystal structure of tachylectin 5A provides evidence for the common origin of the innate immunity and the blood coagulation systems. Proceedings of the National Academy of Sciences of the United States of America 98: 13519-13524
Kalyuzhniy O, Di Paolo NC, Silvestry M, Hofherr SE, Barry MA, Stewart PL, Shayakhmetov DM (2008) Adenovirus serotype 5 hexon is critical for virus infection of hepatocytes in vivo. Proceedings of the National Academy of Sciences of the United States of America 105: 5483-5488
Kampmann B, Cubitt D, Walls T, Naik P, Depala M, Samarasinghe S, Robson D, Hassan A, Rao K, Gaspar H, Davies G, Jones A, Cale C, Gilmour K, Real M, Foo M, Bennett- Rees N, Hewitt A, Amrolia P, Veys P (2005) Improved outcome for children with disseminated adenoviral infection following allogeneic stem cell transplantation. British journal of haematology 130: 595-603
Karplus K, Barrett C, Hughey R (1998) Hidden Markov models for detecting remote protein homologies. Bioinformatics 14: 846-856
Kawai T, Akira S (2007) Antiviral signaling through pattern recognition receptors. Journal of biochemistry 141: 137-145
Knowles MR, Hohneker KW, Zhou Z, Olsen JC, Noah TL, Hu PC, Leigh MW, Engelhardt JF, Edwards LJ, Jones KR, et al. (1995) A controlled study of adenoviral- vector-mediated gene transfer in the nasal epithelium of patients with cystic fibrosis. The New England journal of medicine 333: 823-831
Kratz PA, Bottcher B, Nassal M (1999) Native display of complete foreign protein domains on the surface of hepatitis B virus capsids. Proceedings of the National Academy of Sciences of the United States of America 96: 1915-1920
Kwong PD, Mascola JR, Nabel GJ (2011) Rational design of vaccines to elicit broadly neutralizing antibodies to HIV-1. Cold Spring Harbor perspectives in medicine 1: a007278
194
Lander ES, Linton LM, Birren B, Nusbaum C, Zody MC, Baldwin J, Devon K, Dewar K, Doyle M, FitzHugh W, Funke R, Gage D, Harris K, Heaford A, Howland J, Kann L, Lehoczky J, LeVine R, McEwan P, McKernan K, Meldrim J, Mesirov JP, Miranda C, Morris W, Naylor J, Raymond C, Rosetti M, Santos R, Sheridan A, Sougnez C, Stange- Thomann N, Stojanovic N, Subramanian A, Wyman D, Rogers J, Sulston J, Ainscough R, Beck S, Bentley D, Burton J, Clee C, Carter N, Coulson A, Deadman R, Deloukas P, Dunham A, Dunham I, Durbin R, French L, Grafham D, Gregory S, Hubbard T, Humphray S, Hunt A, Jones M, Lloyd C, McMurray A, Matthews L, Mercer S, Milne S, Mullikin JC, Mungall A, Plumb R, Ross M, Shownkeen R, Sims S, Waterston RH, Wilson RK, Hillier LW, McPherson JD, Marra MA, Mardis ER, Fulton LA, Chinwalla AT, Pepin KH, Gish WR, Chissoe SL, Wendl MC, Delehaunty KD, Miner TL, Delehaunty A, Kramer JB, Cook LL, Fulton RS, Johnson DL, Minx PJ, Clifton SW, Hawkins T, Branscomb E, Predki P, Richardson P, Wenning S, Slezak T, Doggett N, Cheng JF, Olsen A, Lucas S, Elkin C, Uberbacher E, Frazier M, Gibbs RA, Muzny DM, Scherer SE, Bouck JB, Sodergren EJ, Worley KC, Rives CM, Gorrell JH, Metzker ML, Naylor SL, Kucherlapati RS, Nelson DL, Weinstock GM, Sakaki Y, Fujiyama A, Hattori M, Yada T, Toyoda A, Itoh T, Kawagoe C, Watanabe H, Totoki Y, Taylor T, Weissenbach J, Heilig R, Saurin W, Artiguenave F, Brottier P, Bruls T, Pelletier E, Robert C, Wincker P, Smith DR, Doucette-Stamm L, Rubenfield M, Weinstock K, Lee HM, Dubois J, Rosenthal A, Platzer M, Nyakatura G, Taudien S, Rump A, Yang H, Yu J, Wang J, Huang G, Gu J, Hood L, Rowen L, Madan A, Qin S, Davis RW, Federspiel NA, Abola AP, Proctor MJ, Myers RM, Schmutz J, Dickson M, Grimwood J, Cox DR, Olson MV, Kaul R, Raymond C, Shimizu N, Kawasaki K, Minoshima S, Evans GA, Athanasiou M, Schultz R, Roe BA, Chen F, Pan H, Ramser J, Lehrach H, Reinhardt R, McCombie WR, de la Bastide M, Dedhia N, Blocker H, Hornischer K, Nordsiek G, Agarwala R, Aravind L, Bailey JA, Bateman A, Batzoglou S, Birney E, Bork P, Brown DG, Burge CB, Cerutti L, Chen HC, Church D, Clamp M, Copley RR, Doerks T, Eddy SR, Eichler EE, Furey TS, Galagan J, Gilbert JG, Harmon C, Hayashizaki Y, Haussler D, Hermjakob H, Hokamp K, Jang W, Johnson LS, Jones TA, Kasif S, Kaspryzk A, Kennedy S, Kent WJ, Kitts P, Koonin EV, Korf I, Kulp D, Lancet D, Lowe TM, McLysaght A, Mikkelsen T, Moran JV, Mulder N, Pollara VJ, Ponting CP, Schuler G, Schultz J, Slater G, Smit AF, Stupka E, Szustakowski J, Thierry-Mieg D, Thierry-Mieg J, Wagner L, Wallis J, Wheeler R, Williams A, Wolf YI, Wolfe KH, Yang SP, Yeh RF, Collins F, Guyer MS, Peterson J, Felsenfeld A, Wetterstrand KA, Patrinos A, Morgan MJ, de Jong P, Catanese JJ, Osoegawa K, Shizuya H, Choi S, Chen YJ, International Human Genome Sequencing C (2001) Initial sequencing and analysis of the human genome. Nature 409: 860-921
Lehrer RI, Lu W (2012) alpha-Defensins in human innate immunity. Immunological reviews 245: 84-112
Lemiale F, Kong WP, Akyurek LM, Ling X, Huang Y, Chakrabarti BK, Eckhaus M, Nabel GJ (2003) Enhanced mucosal immunoglobulin A response of intranasal adenoviral vector human immunodeficiency virus vaccine and localization in the central nervous system. Journal of virology 77: 10078-10087
195
Lindert S, Silvestry M, Mullen TM, Nemerow GR, Stewart PL (2009) Cryo-electron microscopy structure of an adenovirus-integrin complex indicates conformational changes in both penton base and integrin. Journal of virology 83: 11491-11501
Liu H, Jin L, Koh SB, Atanasov I, Schein S, Wu L, Zhou ZH (2010) Atomic structure of human adenovirus by cryo-EM reveals interactions among protein networks. Science 329: 1038-1043
Liu J, Deng Y, Dey AK, Moore JP, Lu M (2009) Structure of the HIV-1 gp41 membrane- proximal ectodomain region in a putative prefusion conformation. Biochemistry 48: 2915-2923
Lowenstein PR, Castro MG (2003) Inflammation and adaptive immune responses to adenoviral vectors injected into the brain: peculiarities, mechanisms, and consequences. Gene therapy 10: 946-954
Manayani DJ, Thomas D, Dryden KA, Reddy V, Siladi ME, Marlett JM, Rainey GJ, Pique ME, Scobie HM, Yeager M, Young JA, Manchester M, Schneemann A (2007) A viral nanoparticle with dual function as an anthrax antitoxin and vaccine. PLoS pathogens 3: 1422-1431
Martinez-Martinez I, Ordonez A, Pedersen S, de la Morena-Barrio ME, Navarro- Fernandez J, Kristensen SR, Minano A, Padilla J, Vicente V, Corral J (2011) Heparin affinity of factor VIIa: Implications on the physiological inhibition by antithrombin and clearance of recombinant factor VIIa. Thromb Res 127: 154-160
Matthews QL, Fatima A, Tang Y, Perry BA, Tsuruta Y, Komarova S, Timares L, Zhao C, Makarova N, Borovjagin AV, Stewart PL, Wu H, Blackwell JL, Curiel DT (2010) HIV antigen incorporation within adenovirus hexon hypervariable 2 for a novel HIV vaccine approach. PloS one 5: e11815
Matthews QL, Yang P, Wu Q, Belousova N, Rivera AA, Stoff-Khalili MA, Waehler R, Hsu HC, Li Z, Li J, Mountz JD, Wu H, Curiel DT (2008) Optimization of capsid- incorporated antigens for a novel adenovirus vaccine approach. Virology journal 5: 98
Meier O, Greber UF (2004) Adenovirus endocytosis. The journal of gene medicine 6 Suppl 1: S152-163
196
Meiler J, Baker D (2003) Coupled prediction of protein secondary and tertiary structure. Proceedings of the National Academy of Sciences of the United States of America 100: 12105-12110
Mercer J, Greber UF (2013) Virus interactions with endocytic pathways in macrophages and dendritic cells. Trends in microbiology 21: 380-388
Mindell JA, Grigorieff N (2003) Accurate determination of local defocus and specimen tilt in electron microscopy. Journal of structural biology 142: 334-347
Mirano-Bascos D, Tary-Lehmann M, Landry SJ (2008) Antigen structure influences helper T-cell epitope dominance in the human immune response to HIV envelope glycoprotein gp120. European journal of immunology 38: 1231-1237
Mittal SK, McDermott MR, Johnson DC, Prevec L, Graham FL (1993) Monitoring foreign gene expression by a human adenovirus-based vector using the firefly luciferase gene as a reporter. Virus research 28: 67-90
Moyer CL, Nemerow GR (2011) Viral weapons of membrane destruction: variable modes of membrane penetration by non-enveloped viruses. Current opinion in virology 1: 44-49
Muruve DA (2004) The innate immune response to adenovirus vectors. Human gene therapy 15: 1157-1166
Nakano MY, Boucke K, Suomalainen M, Stidwill RP, Greber UF (2000) The first step of adenovirus type 2 disassembly occurs at the cell surface, independently of endocytosis and escape to the cytosol. Journal of virology 74: 7085-7095
Nemerow GR (2000) Cell receptors involved in adenovirus entry. Virology 274: 1-4
Nguyen EK, Nemerow GR, Smith JG (2010) Direct evidence from single-cell analysis that human {alpha}-defensins block adenovirus uncoating to neutralize infection. Journal of virology 84: 4041-4049
Ni S, Bernt K, Gaggar A, Li ZY, Kiem HP, Lieber A (2005) Evaluation of biodistribution and safety of adenovirus vectors containing group B fibers after intravenous injection into baboons. Human gene therapy 16: 664-677
197
Nilsson EC, Storm RJ, Bauer J, Johansson SM, Lookene A, Angstrom J, Hedenstrom M, Eriksson TL, Frangsmyr L, Rinaldi S, Willison HJ, Pedrosa Domellof F, Stehle T, Arnberg N (2011) The GD1a glycan is a cellular receptor for adenoviruses causing epidemic keratoconjunctivitis. Nature medicine 17: 105-109
O'Hara GA, Duncan CJ, Ewer KJ, Collins KA, Elias SC, Halstead FD, Goodman AL, Edwards NJ, Reyes-Sandoval A, Bird P, Rowland R, Sheehy SH, Poulton ID, Hutchings C, Todryk S, Andrews L, Folgori A, Berrie E, Moyle S, Nicosia A, Colloca S, Cortese R, Siani L, Lawrie AM, Gilbert SC, Hill AV (2012) Clinical assessment of a recombinant simian adenovirus ChAd63: a potent new vaccine vector. The Journal of infectious diseases 205: 772-781
Ofek G, Guenaga FJ, Schief WR, Skinner J, Baker D, Wyatt R, Kwong PD (2010) Elicitation of structure-specific antibodies by epitope scaffolds. Proceedings of the National Academy of Sciences of the United States of America 107: 17880-17887
Ofek G, Tang M, Sambor A, Katinger H, Mascola JR, Wyatt R, Kwong PD (2004) Structure and mechanistic analysis of the anti-human immunodeficiency virus type 1 antibody 2F5 in complex with its gp41 epitope. Journal of virology 78: 10724-10737
Pacini DL, Collier AM, Henderson FW (1987) Adenovirus infections and respiratory illnesses in children in group day care. The Journal of infectious diseases 156: 920-927
Parker AL, Waddington SN, Nicol CG, Shayakhmetov DM, Buckley SM, Denby L, Kemball-Cook G, Ni S, Lieber A, McVey JH, Nicklin SA, Baker AH (2006) Multiple vitamin K-dependent coagulation zymogens promote adenovirus-mediated gene delivery to hepatocytes. Blood 108: 2554-2561
Parkin J, Cohen B (2001) An overview of the immune system. Lancet 357: 1777-1789
Perera L, Darden TA, Pedersen LG (2002) Predicted solution structure of zymogen human coagulation FVII. Journal of computational chemistry 23: 35-47
Perez-Berna AJ, Marabini R, Scheres SH, Menendez-Conejero R, Dmitriev IP, Curiel DT, Mangel WF, Flint SJ, San Martin C (2009) Structure and uncoating of immature adenovirus. Journal of molecular biology 392: 547-557
198
Persson BD, Reiter DM, Marttila M, Mei YF, Casasnovas JM, Arnberg N, Stehle T (2007) Adenovirus type 11 binding alters the conformation of its receptor CD46. Nature structural & molecular biology 14: 164-166
Pettersen EF, Goddard TD, Huang CC, Couch GS, Greenblatt DM, Meng EC, Ferrin TE (2004) UCSF Chimera--a visualization system for exploratory research and analysis. Journal of computational chemistry 25: 1605-1612
Pham TT, Burchette JL, Jr., Hale LP (2003) Fatal disseminated adenovirus infections in immunocompromised patients. American journal of clinical pathology 120: 575-583
Phillips JC, Braun R, Wang W, Gumbart J, Tajkhorshid E, Villa E, Chipot C, Skeel RD, Kale L, Schulten K (2005) Scalable molecular dynamics with NAMD. Journal of computational chemistry 26: 1781-1802
Plummer EM, Manchester M (2010) Viral nanoparticles and virus-like particles: platforms for contemporary vaccine design. Wiley interdisciplinary reviews Nanomedicine and nanobiotechnology
Puntener D, Engelke MF, Ruzsics Z, Strunze S, Wilhelm C, Greber UF (2011) Stepwise loss of fluorescent core protein V from human adenovirus during entry into cells. Journal of virology 85: 481-496
Purtscher M, Trkola A, Gruber G, Buchacher A, Predl R, Steindl F, Tauer C, Berger R, Barrett N, Jungbauer A, et al. (1994) A broadly neutralizing human monoclonal antibody against gp41 of human immunodeficiency virus type 1. AIDS research and human retroviruses 10: 1651-1658
Rajabi M, Ericksen B, Wu X, de Leeuw E, Zhao L, Pazgier M, Lu W (2012) Functional determinants of human enteric alpha-defensin HD5: crucial role for hydrophobicity at dimer interface. The Journal of biological chemistry 287: 21615-21627
Raper SE, Chirmule N, Lee FS, Wivel NA, Bagg A, Gao GP, Wilson JM, Batshaw ML (2003) Fatal systemic inflammatory response syndrome in a ornithine transcarbamylase deficient patient following adenoviral gene transfer. Molecular genetics and metabolism 80: 148-158
Reddy VS, Natchiar SK, Stewart PL, Nemerow GR (2010) Crystal structure of human adenovirus at 3.5 A resolution. Science 329: 1071-1075
199
Rezaie AR (2000) Identification of basic residues in the heparin-binding exosite of factor Xa critical for heparin and factor Va binding. Journal of Biological Chemistry 275: 3320- 3327
Roberts DM, Nanda A, Havenga MJ, Abbink P, Lynch DM, Ewald BA, Liu J, Thorner AR, Swanson PE, Gorgone DA, Lifton MA, Lemckert AA, Holterman L, Chen B, Dilraj A, Carville A, Mansfield KG, Goudsmit J, Barouch DH (2006) Hexon-chimaeric adenovirus serotype 5 vectors circumvent pre-existing anti-vector immunity. Nature 441: 239-243
Roelvink PW, Lizonova A, Lee JG, Li Y, Bergelson JM, Finberg RW, Brough DE, Kovesdi I, Wickham TJ (1998) The coxsackievirus-adenovirus receptor protein can function as a cellular attachment protein for adenovirus serotypes from subgroups A, C, D, E, and F. Journal of virology 72: 7909-7915
Rohl CA, Strauss CE, Chivian D, Baker D (2004) Modeling structurally variable regions in homologous proteins with rosetta. Proteins 55: 656-677
Ross PD, Cheng N, Conway JF, Firek BA, Hendrix RW, Duda RL, Steven AC (2005) Crosslinking renders bacteriophage HK97 capsid maturation irreversible and effects an essential stabilization. The EMBO journal 24: 1352-1363
Rowe WP, Huebner RJ, Gilmore LK, Parrott RH, Ward TG (1953) Isolation of a cytopathogenic agent from human adenoids undergoing spontaneous degeneration in tissue culture. Proceedings of the Society for Experimental Biology and Medicine Society for Experimental Biology and Medicine 84: 570-573
Roy A, Kucukural A, Zhang Y (2010) I-TASSER: a unified platform for automated protein structure and function prediction. Nature protocols 5: 725-738
Russell WC (2000) Update on adenovirus and its vectors. The Journal of general virology 81: 2573-2604
Russell WC (2009) Adenoviruses: update on structure and function. The Journal of general virology 90: 1-20
Saban SD, Nepomuceno RR, Gritton LD, Nemerow GR, Stewart PL (2005) CryoEM structure at 9A resolution of an adenovirus vector targeted to hematopoietic cells. Journal of molecular biology 349: 526-537
200
Saban SD, Silvestry M, Nemerow GR, Stewart PL (2006) Visualization of alpha-helices in a 6-angstrom resolution cryoelectron microscopy structure of adenovirus allows refinement of capsid protein assignments. Journal of virology 80: 12049-12059
Schodel F, Moriarty AM, Peterson DL, Zheng JA, Hughes JL, Will H, Leturcq DJ, McGee JS, Milich DR (1992) The position of heterologous epitopes inserted in hepatitis B virus core particles determines their immunogenicity. Journal of virology 66: 106-114
Segerman A, Atkinson JP, Marttila M, Dennerquist V, Wadell G, Arnberg N (2003) Adenovirus type 11 uses CD46 as a cellular receptor. Journal of virology 77: 9183-9191
Shayakhmetov DM, Gaggar A, Ni SH, Li ZY, Lieber A (2005) Adenovirus binding to blood factors results in liver cell infection and hepatotoxicity. Journal of virology 79: 7478-7491
Shayakhmetov DM, Lieber A (2000) Dependence of adenovirus infectivity on length of the fiber shaft domain. Journal of virology 74: 10274-10286
Shayakhmetov DM, Papayannopoulou T, Stamatoyannopoulos G, Lieber A (2000) Efficient gene transfer into human CD34(+) cells by a retargeted adenovirus vector. Journal of virology 74: 2567-2583
Silvestry M, Lindert S, Smith JG, Maier O, Wiethoff CM, Nemerow GR, Stewart PL (2009) Cryo-electron microscopy structure of adenovirus type 2 temperature-sensitive mutant 1 reveals insight into the cell entry defect. Journal of virology 83: 7375-7383
Smith JG, Nemerow GR (2008) Mechanism of adenovirus neutralization by Human alpha-defensins. Cell host & microbe 3: 11-19
Smith JG, Silvestry M, Lindert S, Lu W, Nemerow GR, Stewart PL (2010a) Insight into the mechanisms of adenovirus capsid disassembly from studies of defensin neutralization. PLoS pathogens 6: e1000959
Smith JG, Wiethoff CM, Stewart PL, Nemerow GR (2010b) Adenovirus. Current topics in microbiology and immunology 343: 195-224
201
Snijder J, Reddy VS, May ER, Roos WH, Nemerow GR, Wuite GJ (2013) Integrin and defensin modulate the mechanical properties of adenovirus. Journal of virology 87: 2756- 2766
Steinstraesser L, Kraneburg U, Jacobsen F, Al-Benna S (2011) Host defense peptides and their antimicrobial-immunomodulatory duality. Immunobiology 216: 322-333
Stewart PL, Burnett RM, Cyrklaff M, Fuller SD (1991) Image reconstruction reveals the complex molecular organization of adenovirus. Cell 67: 145-154
Stewart PL, Fuller SD, Burnett RM (1993) Difference imaging of adenovirus: bridging the resolution gap between X-ray crystallography and electron microscopy. The EMBO journal 12: 2589-2599
Stewart PL, Nemerow GR (2007) Cell integrins: commonly used receptors for diverse viral pathogens. Trends in microbiology 15: 500-507
Strunze S, Engelke MF, Wang IH, Puntener D, Boucke K, Schleich S, Way M, Schoenenberger P, Burckhardt CJ, Greber UF (2011) Kinesin-1-mediated capsid disassembly and disruption of the nuclear pore complex promote virus infection. Cell host & microbe 10: 210-223
Sun ZY, Oh KJ, Kim M, Yu J, Brusic V, Song L, Qiao Z, Wang JH, Wagner G, Reinherz EL (2008) HIV-1 broadly neutralizing antibody extracts its epitope from a kinked gp41 ectodomain region on the viral membrane. Immunity 28: 52-63
Suomalainen M, Greber UF (2013) Uncoating of non-enveloped viruses. Current opinion in virology 3: 27-33
Szyk A, Wu Z, Tucker K, Yang D, Lu W, Lubkowski J (2006) Crystal structures of human alpha-defensins HNP4, HD5, and HD6. Protein science : a publication of the Protein Society 15: 2749-2760
Tatsis N, Ertl HC (2004) Adenoviruses as vaccine vectors. Mol Ther 10: 616-629
Thaci B, Ulasov IV, Wainwright DA, Lesniak MS (2011) The challenge for gene therapy: innate immune response to adenoviruses. Oncotarget 2: 113-121
202
Tomko RP, Xu R, Philipson L (1997) HCAR and MCAR: the human and mouse cellular receptors for subgroup C adenoviruses and group B coxsackieviruses. Proceedings of the National Academy of Sciences of the United States of America 94: 3352-3356
Trabuco LG, Villa E, Mitra K, Frank J, Schulten K (2008) Flexible fitting of atomic structures into electron microscopy maps using molecular dynamics. Structure 16: 673- 683
Trabuco LG, Villa E, Schreiner E, Harrison CB, Schulten K (2009) Molecular dynamics flexible fitting: a practical guide to combine cryo-electron microscopy and X-ray crystallography. Methods 49: 174-180
Trotman LC, Mosberger N, Fornerod M, Stidwill RP, Greber UF (2001) Import of adenovirus DNA involves the nuclear pore complex receptor CAN/Nup214 and histone H1. Nature cell biology 3: 1092-1100
Tuve S, Wang H, Jacobs JD, Yumul RC, Smith DF, Lieber A (2008) Role of cellular heparan sulfate proteoglycans in infection of human adenovirus serotype 3 and 35. PLoS pathogens 4: e1000189
Uversky VN, Oldfield CJ, Dunker AK (2008) Intrinsically disordered proteins in human diseases: introducing the D2 concept. Annual review of biophysics 37: 215-246 van Heel M, Gowen B, Matadeen R, Orlova EV, Finn R, Pape T, Cohen D, Stark H, Schmidt R, Schatz M, Patwardhan A (2000) Single-particle electron cryo-microscopy: towards atomic resolution. Q Rev Biophys 33: 307-369 van Heel M, Harauz G, Orlova EV, Schmidt R, Schatz M (1996) A new generation of the IMAGIC image processing system. Journal of structural biology 116: 17-24 van Raaij MJ, Mitraki A, Lavigne G, Cusack S (1999) A triple beta-spiral in the adenovirus fibre shaft reveals a new structural motif for a fibrous protein. Nature 401: 935-938
Varghese R, Mikyas Y, Stewart PL, Ralston R (2004) Postentry neutralization of adenovirus type 5 by an antihexon antibody. Journal of virology 78: 12320-12332
Vellinga J, Van der Heijdt S, Hoeben RC (2005) The adenovirus capsid: major progress in minor proteins. The Journal of general virology 86: 1581-1588
203
Vigant F, Descamps D, Jullienne B, Esselin S, Connault E, Opolon P, Tordjmann T, Vigne E, Perricaudet M, Benihoud K (2008) Substitution of hexon hypervariable region 5 of adenovirus serotype 5 abrogates blood factor binding and limits gene transfer to liver. Mol Ther 16: 1474-1480
Waddington SN, McVey JH, Bhella D, Parker AL, Barker K, Atoda H, Pink R, Buckley SM, Greig JA, Denby L, Custers J, Morita T, Francischetti IM, Monteiro RQ, Barouch DH, van Rooijen N, Napoli C, Havenga MJ, Nicklin SA, Baker AH (2008) Adenovirus serotype 5 hexon mediates liver gene transfer. Cell 132: 397-409
Walls T, Hawrami K, Ushiro-Lumb I, Shingadia D, Saha V, Shankar AG (2005) Adenovirus infection after pediatric bone marrow transplantation: is treatment always necessary? Clinical infectious diseases : an official publication of the Infectious Diseases Society of America 40: 1244-1249
Wang H, Li ZY, Liu Y, Persson J, Beyer I, Moller T, Koyuncu D, Drescher MR, Strauss R, Zhang XB, Wahl JK, 3rd, Urban N, Drescher C, Hemminki A, Fender P, Lieber A (2011) Desmoglein 2 is a receptor for adenovirus serotypes 3, 7, 11 and 14. Nature medicine 17: 96-104
Wickham TJ, Mathias P, Cheresh DA, Nemerow GR (1993) Integrins alpha v beta 3 and alpha v beta 5 promote adenovirus internalization but not virus attachment. Cell 73: 309- 319
Wiethoff CM, Wodrich H, Gerace L, Nemerow GR (2005) Adenovirus protein VI mediates membrane disruption following capsid disassembly. Journal of virology 79: 1992-2000
Wilmes M, Cammue BP, Sahl HG, Thevissen K (2011) Antibiotic activities of host defense peptides: more to it than lipid bilayer perturbation. Natural product reports 28: 1350-1358
Worgall S, Wolff G, Falck-Pedersen E, Crystal RG (1997) Innate immune mechanisms dominate elimination of adenoviral vectors following in vivo administration. Human gene therapy 8: 37-44
Wu H, Han T, Belousova N, Krasnykh V, Kashentseva E, Dmitriev I, Kataram M, Mahasreshti PJ, Curiel DT (2005) Identification of sites in adenovirus hexon for foreign peptide incorporation. Journal of virology 79: 3382-3390
204
Xagorari A, Chlichlia K (2008) Toll-like receptors and viruses: induction of innate antiviral immune responses. The open microbiology journal 2: 49-59
Xiang Z, Gao G, Reyes-Sandoval A, Cohen CJ, Li Y, Bergelson JM, Wilson JM, Ertl HC (2002) Novel, chimpanzee serotype 68-based adenoviral vaccine carrier for induction of antibodies to a transgene product. Journal of virology 76: 2667-2675
Xiang Z, Li Y, Cun A, Yang W, Ellenberg S, Switzer WM, Kalish ML, Ertl HC (2006) Chimpanzee adenovirus antibodies in humans, sub-Saharan Africa. Emerging infectious diseases 12: 1596-1599
Xu ZL, Qiu Q, Tian J, Smith JS, Conenello GM, Morita T, Byrnes AP (2013) Coagulation factor X shields adenovirus type 5 from attack by natural antibodies and complement. Nature medicine 19: 452-+
Yang LK, Manithody C, Rezaie AR (2002) Localization of the heparin binding exosite of factor IXa. Journal of Biological Chemistry 277: 50756-50760
Yang Y, Jooss KU, Su Q, Ertl HC, Wilson JM (1996) Immune responses to viral antigens versus transgene product in the elimination of recombinant adenovirus-infected hepatocytes in vivo. Gene therapy 3: 137-144
Young NS, Levin J, Prendergast RA (1972) An invertebrate coagulation system activated by endotoxin: evidence for enzymatic mediation. The Journal of clinical investigation 51: 1790-1797
Zaiss AK, Machado HB, Herschman HR (2009) The influence of innate and pre-existing immunity on adenovirus therapy. Journal of cellular biochemistry 108: 778-790
Zhang Y (2008) I-TASSER server for protein 3D structure prediction. BMC bioinformatics 9: 40
Zhang Y, Bergelson JM (2005) Adenovirus receptors. Journal of virology 79: 12125- 12131
Zhang Y, Doherty T, Li J, Lu W, Barinka C, Lubkowski J, Hong M (2010a) Resonance assignment and three-dimensional structure determination of a human alpha-defensin, HNP-1, by solid-state NMR. Journal of molecular biology 397: 408-422
205
Zhang Y, Lu W, Hong M (2010b) The membrane-bound structure and topology of a human alpha-defensin indicate a dimer pore mechanism for membrane disruption. Biochemistry 49: 9770-9782
Zhu JG, Huang XP, Yang YP (2007) Innate immune response to adenoviral vectors is mediated by both Toll-like receptor-dependent and -independent pathways. Journal of virology 81: 3170-3180
Zubieta C, Schoehn G, Chroboczek J, Cusack S (2005) The structure of the human adenovirus 2 penton. Molecular cell 17: 121-135
Zwick MB, Jensen R, Church S, Wang M, Stiegler G, Kunert R, Katinger H, Burton DR (2005) Anti-human immunodeficiency virus type 1 (HIV-1) antibodies 2F5 and 4E10 require surprisingly few crucial residues in the membrane-proximal external region of glycoprotein gp41 to neutralize HIV-1. Journal of virology 79: 1252-1261
206