Glycan Receptor Binding Determinants of
Influenza A Virus Hemagglutinin
by
MASSACHUSETTS INSTITUTE
OF TECHNOLOGY
Xiaoying Koh
OCT 42010
Bachelor of Science in Biomedical Engineering
Johns Hopkins University, 2004
LIBRARIES
ARCHIVES
SUBMITTED TO THE DEPARTMENT OF BIOLOGICAL ENGINEERING
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY IN BIOLOGICAL ENGINEERING
AT THE
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
September 2010
© 2010 Massachusetts Institute of Technology. All Rights Reserved.
Signature of author:
Department of Biological Engineering
August 6, 2010
Certified by:
Sasisekharan, Ph.D
Director & Edward Hood Taplin Professor of Health Sciences & Technology
Thesis Supervisor
Certified by:
Jianzhu Chen, Ph.D
Ivan R. Cottrell Professor of Immunology and Professor of Biology
Thesis Supervisor
Accepted by:
Bevin Engelward, Sc.D
Associate Professor of Biological Engineering
Thesis Committee Chair
This doctoral thesis has been examined by a committee of the Biological Engineering
Department as follows:
Bevin Engelward, Sc.D
Associate Professor of Biological Engineering
Thesis Committee Chair
Ram Sasisekharan, Ph.D
Director & Edward Hood Taplin Professor of
Health Sciences & Technology
Thesis Supervisor
Jianzhu Chen, Ph.D
Ivan R. Cottrell Professor of Immunology
and Professor of Biology
Thesis Supervisor
Rahul Raman, Ph.D
Research Sciehtist
Harvard-MIT Division of Health Sciences & Technology
Glycan Receptor Binding Determinants of
Influenza A Virus Hemagglutinin
by
Xiaoying Koh
Submitted to the Department of Biological Engineering on
August 6, 2010 in Partial Fulfillment of the Requirements
for the Degree of Doctor of Philosophy in Biological Engineering
ABSTRACT
An understanding of the factors involved in the human adaptation of influenza A viruses
is critical for various aspects of influenza preparedness, including the development of
appropriate surveillance measures, preventive strategies and effective treatments. A key step in
influenza human adaptation is the acquisition of mutations in the viral coat glycoprotein,
hemagglutinin (HA), which changes its binding specificity towards glycan receptors in the
human upper respiratory epithelia (referred to as human receptors). In this thesis, determinants
that mediate changes in HA-glycan receptor binding specificity are investigated, with focus on
the molecular environments within and surrounding the glycan receptor binding site (RBS) of
HA. The glycan receptor binding properties of HA from different influenza subtypes (H1N1,
H2N2, H3N2 and H5N1) are studied using a combination of approaches including dosedependent glycan binding, human tissue staining and structural modeling. Using these
complementary analyses, it is shown in this thesis that the molecular interactions between amino
acids in and proximal to the RBS (referred to as amino acid interaction networks), including
those between the RBS and glycosylation at sites proximal to the RBS, and interactions between
the RBS and the glycan receptor together govern the high affinity binding of HA to human
receptors.
The thesis is divided into three sections. First, the evolution of glycan receptor binding
specificity of recent human-adapted H3 strains such as A/Fujian/411/02 and A/Panama/2007/99
is investigated, with implications on vaccine production in chicken eggs. Second, the
determinants of glycan receptor binding affinity of potentially pandemic avian viruses is studied
in the context of the recently circulating H2 A/Chicken/Pennsylvania/2004 and the highly
pathogenic H5 A/Vietnam/1203/2004. Here it is shown that mutations which cause human
adaptation of H2 do not increase human receptor binding affinity in H5, and the importance of
amino acid interaction networks is implicated. Third, determinants that govern the high affinity
human receptor binding of pandemic influenza HAs is investigated using the prototypic 1918
H1N1 HA as a model system. The roles of amino acid interaction networks and the molecular
interactions between the RBS and glycosylation at sites proximal to the RBS in contributing to
the high affinity human receptor binding of 1918 H1N
HA are investigated.
The approaches presented in this thesis to systematically investigate molecular
interactions between HA and glycan receptors that impinge on quantitative HA-glycan receptor
binding affinity offer a new angle towards studying determinants of human receptor binding
specificity and affinity of influenza A virus HAs.
Thesis Supervisor: Ram Sasisekharan
Title: Director & Edward Hood Taplin Professor of Health Sciences & Technology
Thesis Supervisor: Jianzhu Chen
Title: Ivan R. Cottrell Professor of Immunology and Professor of Biology
Acknowledgements
This thesis, and my journey as a graduate student, would not be possible without the
support of many to whom I am immensely grateful. First and foremost, I would like to thank my
advisors Professor Jianzhu Chen and Professor Ram Sasisekharan for their belief in me and for
giving me the wonderful opportunity to work with them. My experiences interacting with them
and working in their labs have helped me grow throughout my career as a graduate student and
molded me as a researcher.
I would also like to thank my thesis committee chair Professor Bevin Engelward for her
invaluable support and guidance. The interest she has shown in my work and her contagious
enthusiasm towards science have been an inspiration for me. I have also had a fantastic
experience as a teaching assistant for her and Professor Pete Dedon in the class 20.440 "Analysis
of Biological Networks". That Fall semester working together with these indomitable BE
professors and first year students will stay with me as a memorable part of my experience at MIT.
Dr. Rahul Raman, a member of my committee and a steadfast mentor in the lab, has been
instrumental to my progress as a graduate student. I have gained much from our numerous
exchanges about research, which have prepared me to think more rigorously and learn to
embrace the big picture. He has been an excellent mentor and role model, and I am immensely
grateful to him.
I would also like to express my heartfelt gratitude to an exceptional mentor and friend, Dr.
Karthik Viswanathan. I have had the privilege of sharing a bench space and adjacent desk with
him during his time in the lab, which had resulted in hours of both serious rumination and
delightful conversation. Thank you, Karthik, for your witty sense of humor, your valuable insight
and most of all your constant encouragement and support.
I am grateful to the Agency for Science, Technology and Research (A*STAR, Singapore)
for funding my PhD studies.
My collaborators in Beijing, Professor Taijiao Jiang and Cao Yang, have added to my
PhD experience and I am grateful to them.
Behind every cloud lies a silver lining, and perhaps, behind some silver linings there had
once been a cloud. I am forever grateful to Professor Douglas Lauffenburger, Dalia Fares,
Professor John Essigman, Ram and Jianzhu of MIT for their unquestioning support during a
difficult period of time. I am indebted to former chairman of A*STAR, Mr Philip Yeo, for his
understanding, care and provision. Thank you Mr Keat Chuan Yeoh, Assistant Managing
Director of the Singapore Economic Development Board, and Mr Teck Thai Heng, former
director of the Boston center of the Singapore Economic Development Board, and his wife
Mariam, and Ms. Yena Lim, Managing Director of A*STAR, and Ms. Shannon Quek for their
assistance, friendship, and for providing me an important link to home. Thank you Wind Chay
for assistance, encouragement and support. I am grateful to Dr. Hans R. Agrawal for his
generosity, thoughtfulness, valuable insight and most of all, his understanding over the past five
years. And last but not least in this context, I would like to thank Dr. Alan E. Siegel for his good
judgement which started my travels on this important road.
A special thanks to my dear friend and fellow pursuer of, among other things, domestic
prowess-Karen Sheh. You have helped me discover hobbies which I cherish and love, and
shared many memorable moments with me. Your steadfast friendship has been a gift and a
blessing in my life. Thank you baby Nicholas for your smiles and for just being you.
My time at MIT would not be complete without my labmates from the Chen and
Sasisekharan labs. They have been the best colleagues and friends one could ever ask for, and I
thank them for being a core part of my experience working in the lab. Thank you Akila
Jayaraman and Karunya Srinivasan for your friendship and companionship every day in lab, rain
or shine. You have made every weekday an enjoyable one. Thank you Udayanath Aich for being
a good friend and lunchmate. In the Sasisekharan lab, I have also had the privilege of working
with Luke Robinson, Beam Charlermchai Artpradit, David Eavarone, Troy Rurak, Vidya
Subramanian, Zachary Shriver, Venkataramanan Soundararajan, Alex Keesling, Ken Warnock,
Aarthi Chandrasekaran, Aravind Srinivasan, Shuhua Nong, Ido Bachelet and Ganpan Gao. From
the Chen lab, where I had so many beginnings, I would like to thank Carol McKinley, Jaimie Lee,
Steve Chen, Adam Drake, Zhuyan Guo, Ching Hung Shen, Bettina Iliopoulou, Ailin Bai, Eileen
Higham, Mobolaji Olurinde, Peter Bak, Mimi Fragoso, Camille Jusino, Oezcan Talay, Lars
Heinig, Maroun Khoury, Li Yan and Amanda Souza. From the Koch Institute core facilities, I
would like to thank Kyja Robinson for pleasant early morning conversations and for her
encouragement and support.
A big thank you to Ada Ziolkowski who from the beginning has been a friend and
reliable aid. Thank you Ada for welcoming me to your office every time I needed something, and
with your spiritedness, for transforming those stressful days to ones full of sunshine. Thank you
for understanding even when so little had been said and for simply being you. Thank you dear
Pemmican, my special furry friend, for bringing me happiness, offering me chances to give you
belly rubs and for those afternoons we hung out in the hallway.
I would like to thank my classmates in Biological Engineering for memorable moments
at MIT: Alex Sheh, Nidhi Shrivastav, James Mutamba, Brandon Kwong, Bonnie Huang,
Abhinav Arneja, Ta Hang and many others. I would also like to thank my neighbours Barbara,
"Chubby", Billy and Billy Jr. for their support. I am grateful to my friends from and through the
Cambridgeport Baptist Church for fellowship and friendship: Pastor Dan Szatkowski, Kathy
Szatkowski, Tracy Wemett, Linda Woodbury, Kunle Adeyemo, Gwendalyn King and Yingshu
Jin.
Words cannot express my heartfelt gratitude to my fiance, Gregory Stenta, for his
patience, kindness, understanding and love. You have been my best supporter and lit up my life.
A warm thank you to Greg's parents, Vesta and Albert Stenta, for their welcoming arms,
assistance and encouragement. Geoff Stenta and Cathy Laramie have been wonderful people in
my life and I also thank them for all they have done. Figaro, my other special furry friend, has
been great company and a wonderful stress-reliever.
I am extremely grateful to my family for their love, support, guidance and so much more.
I know that silently or vocally, you have always been cheering me on and that means so much to
me. I look forward to coming home soon. Love and gratitude to my Dad (Baba), Mom (Mama),
sister Bonnie (Qiqi) and Grandma (Popo).
And last but not least, thank you God for leading me to you and for your blessings,
provision and unconditional love.
"ForI know the plans I have for you," declares the LORD,
'plans to prosperyou and not to harm you,
plans to give you hope and afuture."
Jeremiah 29:11
Table of Contents
Abstract...................................................................................................
5
Acknowledgements......................................................................................
7
List of Figures...........................................................................................
11
1. Introduction..........................................................................................
13
13
Sum m ary ..............................................................................................
14
1.1 The Influenza A V irus........................................................................
16
1.2 Host Adaptation and Pandemic Formation................................................
18
A
Virus...................................................
1.3 Hemagglutinin of the Influenza
18
1.3.1 The HA receptor binding site and glycan receptors..............................
20
1.3.2 Methods to investigate HA-receptor interactions.................................
1.3.3 Role of glycan receptor binding properties of HA in viral transmission...... 24
1.3.4 Role of HA in the host immunity against influenza A virus..................... 26
1.4 Challenges Correlating Biochemical Glycan Receptor Binding with Host Adaptation
.....................................................
28
29
1.5 Motivation for this Thesis and Thesis Outline.............................................
31
1.6 Significance........................................................................................
32
R eferences.............................................................................................
37
2. Glycan Receptor Specificity of Recent Human-Adapted H3 Viruses....................
.. 37
Summ ary ...........................................................................................
Introduction ....................................................................................
R esu lts..........................................................................................
2.2.1 A rapid computational method for predicting HA-receptor binding activity.
2.2.2 A virtual water model for evaluating HA-receptor binding energy............
2.2.3 Glycan receptor binding specificity of recent human H3N2 viruses...........
2.2.4 Experimental validation of residues critical for receptor binding..............
2.3 D iscussion .....................................................................................
2.4 M aterials and M ethods.......................................................................
....
R eferences........................................................................................
2.1
2 .2
38
39
39
40
42
46
46
49
51
55
3. Determinants of Glycan Receptor Specificity of H2 and H5 HA.........................
.
55
Sum m ary............................................................................................
3.1
3.2
56
Introduction ....................................................................................
57
Determinants of Glycan Receptor Specificity of H2N2 HA............................
58
Results................................................................................
3.2.1
58
HA.......
of
Alb58
specificity
3.2.1.1 Characterization of glycan receptor binding
3.2.1.2 Mutations in RBS of CkPA04 and their effects on its glycan receptor
binding specificity.................................................................
3.3 Effect of HA Mutations Gln226Leu and Gly228Ser on Avian H5N1 HA Receptor
B inding Specificity...........................................................................
3.3.1
R esults................................................................................
3.3.1.1 Dose-dependent glycan binding assays of H5N1 HA..........................
3.4 D iscussion .....................................................................................
3.5 Materials and Methods.......................................................................
R eferences............................................................................................
62
69
70
70
72
75
78
4. Determinants of Glycan Receptor Specificity of H1 HA....................................
Summ ary ...............................................................................................
4. Amino Acid Networks Govern H1NI HA-Glycan Interactions..........................
4.1.1
Introduction...........................................................................
4 .1.2
Results.................................................................................
4.1.2.1 Regions of molecular interactions between HA and sialylated glycans........
4.1.2.2 H2/H3 human adaptive mutations in Hi HA......................... ...........
4 .1.3
D iscussion ............................................................................
4.2 Effects of HIN1 HA Glycosylation on Receptor Binding.................................
4.2.1
Introduction..........................................................................
4.2.2
Results and Discussion..............................................................
4.2.2.1 Identification of conserved N-glycosylation site proximal to RBS of H1N1
HA ....................................................................................
4.2.2.2 Circular dichroism analysis of mutant HAs......................................
4.2.2.3 Direct binding analysis of mutant and wildtype HAs on the glycan array....
4.2.3
C onclusions...........................................................................
4.3 Materials and Methods........................................................................
References..............................................................................................
93
93
95
10 1
101
102
5. Summary and Significance of Thesis Research.............................................
105
81
81
82
82
82
82
85
90
91
91
93
List of Figures
Chapter 1
Figure 1.1:
Figure 1.2:
Figure 1.3:
Figure 1.4:
Figure 1.5:
Figure 1.6:
Schematic diagram of the influenza A virus and receptor............................... 15
Host-switch events that lead to emergence of pandemic influenza A virus strains... 17
19
The trimeric ectodomain and receptor binding site of HA............................
22
Glycan arrays for analysis of influenza binding to glycan receptors................
25
tract.........................................
respiratory
Tissue staining of the human
Schematic diagram of the HAl subunit of a HA monomer and the five antigenic sites
. . 27
A -E .............................................................................................
Chapter 2
Figure 2.1: Correlation of energy functions with perturbations of crystal structures of H3N2 HA
43
in complex with an avian receptor analog..................................................
Figure 2.2: Evolution of receptor binding specificity of the Panama- and Fujian-like viruses... 45
Figure 2.3: Experimental validation of amino acid residues critical for altered receptor binding
47
specificity of Fujian-like viruses..........................................................
Chapter 3
Figure 3.1: Glycan receptor binding specificity of Alb58 HA......................................
Figure 3.2: Glycan receptor binding specificity of mutant forms of Alb58 HA..................
Figure 3.3: Glycan receptor binding specificity of CkPA04 HA.....................................
Figure 3.4: Homology-based structural model of HA-glycan receptor complexes.................
Figure 3.5: Glycan receptor binding specificity of mutant forms of CkPA04 HA...............
Figure 3.6: Glycan receptor binding affmities of the mutant forms of CkPA04 HA.............
Figure 3.7: Dose-dependent direct glycan binding of wildtype and mutant Viet04 HA..........
Chapter 4
Figure 4.1:
Figure 4.2:
Figure 4.3:
Figure 4.4:
Figure
Figure
Figure
Figure
59
61
63
64
66
68
.71
Structural models of avian HI, H3 and H5 HA interacting with LSTa.............. 83
Structural models of wildtype and mutant HI HA interacting with LSTc........... 87
Dose-dependent direct glycan binding of wildtype and mutant HI HA............. 89
Alignment of avian and human H1NI HA sequences show conserved
94
N-glycosylation sites...........................................................................
signatures
spectral
similar
HAs
show
4.5: Circular dichroism of wildtype and mutant Hi
96
indicative of no general misfolding due to deglycosylation..........................
4.6: Glycan receptor binding affinity and structural model of wildtype SC18 HA and
glycan receptor binding affinity of deglycosylated SC18 HA........................ 98
4.7: Glycan receptor binding affinity and structural model of wildtype NYl8 HA and
99
glycan receptor binding affinity of deglycosylated NYl 8 HA........................
4.8: Glycan receptor binding affinity and structural model of wildtype AV18 HA and
100
glycan receptor binding affinity of deglyosylated AV18 HA........................
Chapter 1
Introduction
Summary
The global threat posed by the highly virulent 'bird flu' H5N1 and the more recent 2009
'swine flu' H1N1 viruses have accelerated efforts to understand the factors involved in the
adaptation of influenza for efficient airborne transmission in humans. A key step in influenza
human adaptation is the acquisition of mutations in the viral coat glycoprotein, hemagglutinin
(HA), which changes its binding specificity towards glycan receptors in the human upper
respiratory epithelia (referred to as human receptors). In this thesis, determinants that mediate
changes in HA-glycan receptor binding specificity are investigated, with focus on the molecular
environments within and surrounding the glycan receptor binding site (RBS) of HA. The glycan
receptor binding properties of HA from different influenza subtypes (HINI, H2N2, H3N2 and
H5N1) are studied using a combination of approaches including dose-dependent glycan binding,
human tissue staining and structural modeling. Using these complementary analyses, it is shown
in this thesis that the molecular interactions between amino acids in and proximal to the RBS
(referred to as amino acid interaction networks), including those between the RBS and
glycosylation at sites proximal to the RBS, and interactions between the RBS and the glycan
receptor together govern the high affinity binding of HA to human receptors.
1.1
The Influenza A Virus
The influenza A virus is a negative-sense, single-stranded RNA virus of the
Orthomyxoviridae family. It affects millions of people worldwide, resulting in 250,000-500,000
deaths
each
year
due
to
respiratory
(http://www.who.int/mediacentre/factsheets/fs211/en/).
illness
and
ensuing
complications
The virus tends to spread locally by
aerosol and globally by international travel and migrating birds, a unique property which makes
it one of the most infectious pathogens known. The genome of the virus is composed of eight
gene segments which encode at least ten known proteins [1]. Hemagglutinin (HA) and
neuraminidase (NA) are the two major viral envelope glycoproteins bearing surface antigens
(Figure 1.1). To date, sixteen HA and nine NA subtypes have been identified, of which three HA
subtypes (Hi, H2, H3) and two NA subtypes (N1, N2) have circulated in humans. Other encoded
proteins include the M2 ion channel, the structural protein M1 and components of the viral RNA
polymerase complex PB2, PB1 and PA [2]. Two additional polymerase proteins, PBl-F2 and
PB 1 N40, have been recently identified, although they are not encoded by all strains [3].
The life cycle of the influenza A virus begins with the binding of HA to sialylated glycan
receptors expressed on host cells. This is a critical step that contributes to infection, transmission
and host tropism of the virus. Wild aquatic birds predominantly express a2-3 sialylated glycans
(referred to as avian receptors or ax2-3) (Figure 1.1, Inset) on their gut epithelium, and are the
natural hosts for influenza viruses. Humans, on the other hand, express a2-6 sialylated glycans
(referred to as human receptors or a2-6) (Figure 1.1, Inset) on their upper respiratory tracts. The
subtle difference in glycan recognition of human versus avian receptors is an important
determinant of the species barrier that prevents avian viruses from efficiently infecting and
transmitting in humans.
o'HA
oo'NA
N-acetytglucosamine
/M2
6
Avian a2-3 receptor
sialicacid
Human a2-6 receptor
Figure 1.1
Schematic diagram of the influenza A virus and receptor. The major surface
glycoproteins HA and NA, and the M2 ion channel protein are shown. Inset: Sialylated glycan receptors
that are a2-3 linked (avian; shown in yellow) or a2-6 linked (human; shown in green).
1.2
Host Adaptation and Pandemic Formation
The influenza A virus has a low-fidelity viral RNA polymerase lacking proofreading
capability, resulting in high error rate and frequent mutations in its genome [4, 5]. The
evolutionary pressure that leads to selection of HA mutations can transform its binding
specificity from avian to human receptors and result in the cross-over and adaptation of the virus
to the human host [6] (Figure 1.2). Another way the virus can adapt to the human host is through
reassortment of avian virus gene segments into existing human-adapted viruses, which is
believed to occur in "mixing vessels" or intermediate hosts such as swine [7] that are susceptible
to both avian and human viruses (Figure 1.2). Both host-switch events can result in viral
acquisition of human receptor specificity and sustained human-to-human transmission, lead to
the introduction of novel influenza antigens into the human population, and ultimately give rise
to occasional pandemics that vary in severity from mild to catastrophic [8].
The current "swine flu" pandemic outbreak is caused by the 2009 H1Nl virus. To date,
this
virus
has
resulted
in
43-89
million
cases
(http://www.cdc.gov/h1nlflu/estimates_2009_hlnl.htm).
In the
of
2 0 th
infection
worldwide
century, other pandemics
were the H1N1 Spanish flu of 1918 (50 million deaths), H2N2 Asian flu of 1957-58 (1 million
deaths) and H3N2 Hong Kong flu of 1967-68 (0.5 million deaths), caused by virus subtypes
which had efficiently adapted to the human host. The 1918 HINI pandemic virus was believed
to have a direct avian origin (Reid 1999), while the 1957 Asian and 1968 Hong Kong pandemics
were caused by reassortant H2N2 and H3N2 viruses respectively. H1N1 and H3N2 viruses
continue to circulate in the human population causing seasonal and epidemic outbreaks. H2N2
viruses stopped circulating in humans in 1968 but continue to circulate in the avian and swine
_:::::
.:.....
.............
...............................
.
...
.....
...
..
.. .. ....
"natural
reservoir"
"mixing vessel"
4)
(
Figure 1.2
Host-switch events that lead to emergence of pandemic influenza A virus strains. HA
mutations can transform its binding specificity from avian to human receptors and result in the
adaptation of the virus to the human host. The avian virus gene segments can also reassort into
existing human-adapted viruses, an event believed to occur in "mixing vessels" or intermediate
hosts such as swine.
populations [9, 10]. In the past decade, much attention has been directed towards avian viruses
including those of the H2, H5, H7 and H9 subtypes which have been responsible for sporadic
outbreaks in humans, and pose a constant threat for new pandemics. Efforts are ongoing to
understand the factors that govern the efficient human adaptation of these avian viruses, among
which H5N1 has been the deadliest with a high mortality rate in humans.
1.3
Hemagglutinin of the Influenza A Virus
HA is a homotrimeric transmembrane protein with an ectodomain consisting of a
globular head and a stem region. Each monomer of HA has a molecular weight of -80 kDa.
Structures of HA of multiple subtypes, either non-complexed or in complex with sialic acid
receptor analogs, have been analyzed using X-ray crystallography and nuclear magnetic
resonance [11-14]. These studies enable the identification of key HA residues involved in
receptor binding, and highlight the important contacts that these residues form with the glycan
receptor.
1.3.1
The HA receptor binding site and glycan receptors
The receptor binding site (RBS) of HA is located on the membrane distal head (Figure
1.3), and is defined by three secondary structure elements: the 190-helix (residues 190-198), the
130-loop (residues 135-138) and the 220-loop (residues 221-228) (Figure 1.3, Inset). The RBS
forms a shallow depression composed of residues conserved in all subtypes of influenza: Y98,
S/T136, W153, H183, L/1194 (numbered based on H3 HA) [15]. Analyses of HA-glycan
cocrystal structures have shown that the orientation of the terminal sialic acid sugar is fixed
relative to the HA glycan binding site [15, 16]. Long chain a2-6 glycans tend to fold and extend
190 he
22 1op
Figure 1.3 The trimeric ectodomain and receptor binding site of HA. The head and stem region of
the trimeric ectodomain of HA is shown in complex with the long-chained a2-6 receptor. Inset: The
receptor binding site is located on the HAI subunit and is defined by the 190-helix (blue), 220-loop (red)
and 130-loop (green). The glycan receptor is shown in yellow. This schematic was constructed based on a
1930 H1 swine virus (PDB:lRVT) and using the visualization software Pymol.
into the RBS (Figure 1.3, Inset) because of the flexible C5-C6 bond of the penultimate galactose.
a2-3 glycans, on the other hand, adopt a more rigid, linear conformation because of the
glycosidic oxygen between the carbon rings of sialic acid and galactose (Figure 1.1, Inset). We
have previously shown that human and avian receptors are characterized by their respective
glycan topologies, and that specific recognition of the human receptor is an important
determinant of virus adaptation to humans [15].
1.3.2
Methods to investigate HA-receptor interactions
In addition to structural studies, array-, cell- and tissue-based binding assays have
enhanced our collective knowledge of influenza virus-host interactions, and have been used to
assess the potential emergence of new human-adapted viruses. These assays can characterize,
either qualitatively or quantitatively, the alteration of HA receptor recognition in response to
specific mutations [17, 18].
Glycan arrays
Array platforms which enable rapid investigation of glycan-binding protein interactions
on a single chip coupled to hundreds of different glycan structures have been widely employed to
study the receptor binding of recombinantly-expressed HAs and whole viruses [15,16,17,18]
(Figure 1.4A). One such glycan array platform, composed of representative avian and human
receptors, has been developed by the Consortium for Functional Glycomics. This microarray has
been used to analyze several HAs from avian and human Hi, H3 and H5 subtypes, including
strains from the 1918 HI pandemic (A/South Carolina/l/1918) and the highly pathogenic H5
avian virus (A/Vietnam/1203/2004) [19, 20]. From glycan microarray studies, it has been found
that factors determining binding properties of the different HAs are not restricted to sialic acid
linkage type-fucosylation and sulphation are other glycan modifications that impact receptor
binding [18].
To capture the effects of multivalent HA-glycan interactions, we have previously
developed a quantitative biochemical framework based on dose-dependent binding of
recombinant HA to glycans representative of different structural topologies [15, 21]. We
recapitulate the properties of multivalent interactions by precomplexing HA with secondary and
primary Abs before addition to each glycan-plated well, such that four HA molecules will be
complexed to each enzyme-linked secondary Ab (Figure 1.4B). By selecting glycans of different
lengths and linkages, we take into account the different spatial topologies that sialylated
receptors can adopt when interacting with the HA receptor binding domain. This selection of
glycans permits us to distinguish the human receptor binding of human-adapted HA from other
glycan binding patterns. This method, which has been used to study several avian and human
HAs, enables the quantitative characterization of HA receptor binding affinity based on the
linearized Hill equation [21]. Importantly, it has been shown that the human receptor-binding
affinity of HlN1 HAs correlate with the efficiency of airborne viral transmission in the ferret
animal model [21, 22], which is an established model to evaluate viral transmissibility in humans
[23-27].
Cell-based assays
For many years, studies on influenza-host interactions have made use of red blood cells
(RBCs) based on their presentation of endogenous surface sialic acid and their ability to
agglutinate [17, 28]. RBC-based assays include the hemagglutination assay, which measures the
agglutination of RBCs by whole viruses, and the hemagglutination inhibition assay, which
detects antibodies and soluble receptor analogues that inhibit RBC agglutination by whole
.............. ..............................
......
..
............
......
............
...
....... .
.
...............
.....
.
B
100- oTrisToric
o 3'-LN
go. oV&XaX-l.
HA J4Mmary Ab
Salyated
gycan
60
* Labeled
1
ry Ab
44
C
~40
20
Sequendal (no precomplex)
HApAb:eAb pr-complexed
HA concentration at 40 pg/mI
Figure 1.4 Glycan arrays for analysis of influenza binding to glycan receptors. (A) Schematic
representation of whole-virus and recombinant HA methods used to assess receptor binding to a glycan
microarray. (B) Receptor binding intensity for recombinant HA is significantly increased with
precomplexation using secondary and primary antibodies because of increased valency.
viruses [7]. Another RBC-based assay is the hemadsorption assay, which utilizes adherent
mammalian cells expressing cell surface HA that bind RBCs, followed by quantification of RBC
heme group absorbance at 540 nm [17]. In the above assays, RBCs can be desialylated by
sialidases and then resialylated using either a2-3 or a2-6 sialyltransferases, enabling the
measurement of HA binding to specific receptors.
The RBC assays using intact, whole viruses have the advantage of potentially reflecting
natural virus-host binding dynamics, such as high valency and avidity binding. However, they
also bear disadvantages and the factors involved are often difficult to control. Agglutination data
is relative and does not allow quantification of apparent binding constants. The quality of
desialylation and resialylation of RBCs depends on many experimental conditions such as
enzyme efficacy and purity. In addition, RBC batches can vary in quality leading to further
discrepancies. Another factor to consider in whole-virus assays is the presence of NA on the
virus surface, which may influence the assay because of the enzymatic removal of sialic acid.
Hence, cell-based assays often reflect the overall affinity of a virus for the human or avian
receptor, but unlike glycan arrays do not reveal other nuances of binding such as the ability of
the virus to discriminate between different receptor structures and glycan modifications.
Tissue staining
The staining of formalin-fixed, paraffin-embedded respiratory tissue sections with whole
virus or recombinant HA have been performed as direct assessments of virus attachment to
physiological host receptors [15, 22, 29]. This tool for studying receptor binding offers a
qualitative view of virus tropism for the respiratory tracts of humans and animal models such as
ferrets and mice. In one such study, van Riel et al. [29] found that seasonal and pandemic human
influenza viruses attach better to human upper respiratory tract epithelium than avian influenza
viruses. The highly pathogenic avian H5N1 virus, on the other hand, attaches better to the lower
respiratory tract of humans [30].
The respiratory epithelium (Figure 1.5A) of the human upper respiratory tract consists of
goblet cells (more common in the upper trachea), ciliated cells (more common in the lower
trachea) and the apical membrane lining, all of which exhibit positive staining by the a2-6specific lectin Sambucus Nigra-I (SNA-I) (Figure 1.5B). Alveolar cells of the human lower
respiratory tract exhibit positive staining by the a2-3-specific Maackia Amurensis Lectin-II
(MAL-II) (Figure 1.5B). a2-6 sialylated cells in the human upper respiratory tract tend to be
long-chained as demonstrated by the predominant presence of long oligosaccharide branches
with multiple lactosamine repeats [15].
As will be shown later in this thesis, the glycan array binding properties of pandemic and
avian HAs correlate with their binding to physiological glycan receptors in the upper and lower
human respiratory tracts.
1.3.3
Role ofglycan receptor binding properties of HA in viral transmission
Mutations that have conferred human receptor binding to HA and led to the avian-tohuman crossover of certain strains of influenza A viruses have been well characterized.
Following the first characterization of a H3 HA crystal structure [31], it was found that H3
viruses bearing Gln226 preferentially bound c2-3 sialylated glycans, while those bearing Leu226
preferentially bound a2-6 sialylated glycans [32]. Subsequent studies of H3 HA evolution also
identified site 228 to be important in glycan recognition-avian H3 HAs bear Gly228 and
human-adapted H3 HAs always bear Ser228 [33]. H2 HA have similar human adaptative
mutations (Gln226Leu and Gly228Ser) as H3 HA [34]. In the case of many H1 viruses,
A#.4O
Figure 1.5 Tissue staining of the human respiratory tract. (A) Hematoxylin and eosin stain of a trachea
section showing the respiratory epithelial layer (RE, left) and goblet cells (GC, right). (B) Image of a
trachea section taken by confocal microscopy showing staining in red by propidium iodide and staining in
green by the a2-6 specific lectin Sambucus Nigra-I (left). A corresponding image of an alveolus from the
lower respiratory tract shows staining in green by the a2-3-specific Maackia Amurensis Lectin-II (MALII) (right).
adaptation to humans is characterized by mutations Glul90Asp and Gly225Asp at the avian
consensus sequence [17]. High affinity binding to human receptors has recently been established
to be a correlate of transmissibility.
1.3.4
Role of HA in the host immunity against influenza A virus
In addition to facilitating host cell recognition through sialic acid binding, HA also
mediates other functions. One of them is the presentation of antigens predominantly on the
membrane distal portion of the HA1 subunit (Figure 1.6). The HA1 subunit consists of 329
amino acids, out of which approximately 130 are thought to lie in or near antigenic sites A-E
[31]. These antigenic sites are involved in antibody-binding and are important for eliciting
neutralizing immune responses by the host-a crucial step in vaccination. However, surface
epitopes on HAl can escape immune recognition by undergoing constant change caused by
antigenic drift and shift, which can either introduce mutations to HA or alter the HA subtype
through gene reassortment. The effectiveness of vaccination hence depends on the accurate
prediction of future circulating strains, followed by annual vaccine reformulation.
Recently, neutralizing antibodies that block viral membrane fusion with the host cell have
been identified [35, 36]. These antibodies recognize an epitope located on the HA2 subunit that
is conserved across different subtypes of HA. This advance brings us closer to the development
of vaccines that can offer broad cross-protective immunity, at the same time, it does face
challenges in application because the conserved epitope is too close to the virus surface to be
readily accessible for B cell activation.
...........
:...
x.......................
x...
x. -......
.......................................................
. . ..
......
. ..
- --w .11,11,11"..
,- TW W" " " " Mw- -
.............
......
................
...........
....
..
....
-A
Antigenic Sites
Figure 1.6 Schematic diagram of the HA1 subunit of a HA monomer and the five antigenic sites AE. Antigenic sites are color-coded as indicated in the legend. These antigenic sites are involved in
antibody-binding and are important for eliciting neutralizing immune responses by the host.
Visualization was performed using the software Pymol and the HA1 monomer of H3-X31 (PDB ID:
1HGD).
In another immunization study, H5N1 HA receptor binding domain mutants were used to
develop vaccines and generate monoclonal antibodies in mice that were able to effectively
neutralize the new variants [37]. These results demonstrate that alterations in HA receptor
binding specificity can facilitate the generation of antibodies with improved neutralizing
efficiency. Evidence from this study also suggests the importance of characterizing, in a
systematic and informed manner, how structure-based modifications of HA can lead to altered
receptor binding specificity. Such a representation will provide insight into the properties of
potential pandemic strains, and pave the way towards effective immune strategies.
1.4
Challenges Correlating Biochemical Glycan Receptor Binding with Host
Adaptation
Binding of recombinant HAs to glycan receptors on some arrays, such as the glycan
microarray, is detected by fluorescent antibodies that recognize histidine tags on the HAs. The
strength of HA binding to the receptors is thus interpreted by the intensity of the fluorescent
signals. Such data is useful for providing qualitative information about HA receptor preference,
however it does not provide quantitative information because experiments are typically
performed at single-dose binding conditions, which do not reflect the kinetics of binding.
Furthermore, as binding activity between HA and glycans presented on the array approaches
equilibrium, higher detection signals tend to reflect high affinity interactions while lower
detection signals usually reflect low affinity interactions. However in cases of extreme HA-toglycan ratios, saturation of glycans by HA may mean that signals are exaggerated even for low
affinity interactions. As a result, signal magnitudes can have different implications depending on
kinetic behavior.
The above challenges in quantifying biochemical glycan receptor binding highlight the
importance of developing dose-dependent assays to assess binding affinities. Moreover, the
factors that lead to pandemic emergence are many and complex. Among the factors influencing
efficient viral transmissibility between humans, it has been shown that the human receptor
binding affinity (quantified using Kd') is especially pertinent. The high Kd' for human receptor
binding by the 1918 pandemic HlNl and the recently declared 2009 pandemic HlN1 viruses
correlates with the transmissibility of the virus via respiratory droplets in ferrets [21, 22].
Recently, this correlation has also been extended to H2N2 viruses [38]. Taken together, the
above observations point to the fact that high affinity binding to human receptors appears to be a
necessary determinant for efficient human adaptation and transmission. Significantly, it has
recently been demonstrated that changes in the interactions between amino acids within and
proximal to the RBS, arising from substitutions due to antigenic drift, have profound effects on
HA-glycan interactions which in turn influence the glycan binding affinity of HA [22, 39].
Hence there is a need to correlate changes in the molecular environment of the HA-glycan
complex with changes in biochemical glycan receptor binding properties, which in turn is
strongly connected with the propensity for human adaptation and transmission.
1.5
Motivation for this Thesis and Thesis Outline
In the context of the challenges mentioned in the previous section-from quantifying
human receptor binding affinity, assessing transmissibility in relation to receptor binding
properties, to correlating changes in the molecular environment with glycan receptor binding
changes-an overall theme in this thesis is influenza receptor binding, focusing on glycan
interactions with the HA receptor binding site. The main goal of this thesis is to understand the
determinants that mediate change in HA receptor binding behavior, such as substitutions that
switch HA receptor preference between avian and human, and to examine the molecular
environment associated with these changes. Here, a range of influenza subtypes (HiNI, H2N2,
H3N2, H5N1) have been studied. The avian and human-adapted forms of these viruses have
been investigated using a combination of approaches including dose-dependent glycan binding,
human tissue staining and structural modeling. From these analyses, it has been found that salient
features in the receptor binding domain are important for receptor binding. For example, high
affinity binding occurs as a result of stable interactions between HA and its receptor, and the
formation of these stable interactions requires a favorable biochemical and structural
environment. Inter-residue contacts of HA play a significant role in providing this environment.
The following part of the thesis is divided into three sections, connected by the principal
concept that networks of amino acids govern HA-glycan receptor interactions. The three sections
are:
1)
Glycan receptor binding specificity of recent human-adapted H3 viruses
The H3N2 virus caused the Hong Kong flu pandemic in 1968. Chapter 1 of this thesis
investigates the receptor binding properties of recent human strains A/Fujian/411/02 and
A/Panama/2007/99, and how receptor binding has evolved over time with implications on
vaccine production. This work has been performed in collaboration with Professor Taijiao
Jiang's group in the Institute of Biophysics, Chinese Academy of Sciences, Beijing, China.
2)
Determinants of glycan receptor binding specificity of H2 and H5 HA
The H2N2 virus caused the Asian flu pandemic in 1957 and stopped circulating in
humans by 1968. In Chapter 2, human adaptation of the H2N2 virus in the context of currently
circulating avian strains is explored. Another avian virus of pandemic potential is the H5N1
virus, which is highly pathogenic in birds and has been known to infect humans, but not transmit
efficiently between humans. Here it will be shown that mutations which cause human adaptation
of H2 do not increase human receptor binding in H5, and the importance of amino acid
interaction networks is implicated.
3)
Determinants ofglycan receptor specificity ofHi HA
As a follow-up to the results of Chapter 2, this part of the thesis investigates determinants
that govern the different behavior of substitutions across subtypes. In Chapter 3, determinants
that govern high affinity human receptor binding of pandemic influenza HAs is investigated
using the prototypic pandemic 1918 H1NI model system. The results suggest that HA residues
can form contacts with one another to provide a stable environment for receptor binding. Using
the same 1918 HIN1 model system, this chapter also investigates the manner in which HA
glycosylation influences receptor binding. Notably, glycans as well as amino acids of HA can
participate in interaction networks and affect receptor binding properties.
1.6
Significance
Development of an understanding of the factors involved in the human adaptation of
influenza virus is critical for various aspects of influenza preparedness, including the
development of appropriate surveillance measures, preventive strategies and effective treatments.
This opening chapter of the thesis has introduced the receptor binding function of the key
influenza virus surface protein HA, described how receptor binding properties relate to host
adaptation and transmission, and the assays commonly employed to study receptor binding. This
chapter has also presented a preview of the next four sections of the thesis, which are linked by
the central theme that networks of amino acid interactions govern HA-glycan interactions.
In the current understanding of influenza receptor binding, most studies have employed a
sequence-centered approach, often applying mutations from one HA to another based on its
effects on the first HA alone. More often than not, influenza receptor binding studies do not
establish a link between quantitative binding and host adaptation. This thesis will address the
need to characterize the way in which structure-based modifications of HA can lead to altered
specificity, keeping in mind their link to transmissibility in humans. Subtle changes in the
molecular environment will be accounted for and their impact on biochemical glycan receptor
binding will be considered. Such a representation will provide insight into the threshold
conditions for acquiring human receptor specificity, and also pave the way towards prophylactic
vaccine and therapeutic monoclonal antibody development.
References
1.
Watanabe, T., Watanabe, S., and Kawaoka, Y. Cellular networks involved in the
influenza virus life cycle. Cell Host Microbe 7,427-439.
2.
Skehel, J.J., and Schild, G.C. (1971). The polypeptide composition of influenza A
viruses. Virology 44, 396-408.
Wise, H.M., Foeglein, A., Sun, J., Dalton, R.M., Patel, S., Howard, W., Anderson, E.C.,
Barclay, W.S., and Digard, P. (2009). A complicated message: Identification of a novel
PB 1-related protein translated from influenza A virus segment 2 mRNA. J Virol 83,
8021-8031.
3.
4.
Doherty, P.C., Turner, S.J., Webby, R.G., and Thomas, P.G. (2006). Influenza and the
challenge for immunology. Nat Immunol 7, 449-455.
5.
Taubenberger, J.K., and Kash, J.C. Influenza virus evolution, host adaptation, and
pandemic formation. Cell Host Microbe 7, 440-451.
6.
Landolt, G.A., and Olsen, C.W. (2007). Up to new tricks - a review of cross-species
transmission of influenza A viruses. Anim Health Res Rev 8, 1-21.
Stevens, J., Blixt, 0., Paulson, J.C., and Wilson, I.A. (2006). Glycan microarray
technologies: tools to survey host specificity of influenza viruses. Nat Rev Microbiol 4,
7.
857-864.
8.
9.
Salomon, R., and Webster, R.G. (2009). The influenza virus enigma. Cell 136, 402-410.
Schafer, J.R., Kawaoka, Y., Bean, W.J., Suss, J., Senne, D., and Webster, R.G. (1993).
Origin of the pandemic 1957 H2 influenza A virus and the persistence of its possible
progenitors in the avian reservoir. Virology 194, 781-788.
32
10.
Ma, W., Vincent, A.L., Gramer, M.R., Brockwell, C.B., Lager, K.M., Janke, B.H.,
Gauger, P.C., Patnayak, D.P., Webby, R.J., and Richt, J.A. (2007). Identification of
H2N3 influenza A viruses from swine in the United States. Proc Natl Acad Sci U S A
104, 20949-20954.
11.
12.
13.
14.
15.
16.
17.
18.
Sauter, N.K., Hanson, J.E., Glick, G.D., Brown, J.H., Crowther, R.L., Park, S.J., Skehel,
J.J., and Wiley, D.C. (1992). Binding of influenza virus hemagglutinin to analogs of its
cell-surface receptor, sialic acid: analysis by proton nuclear magnetic resonance
spectroscopy and X-ray crystallography. Biochemistry 31, 9609-962 1.
Eisen, M.B., Sabesan, S., Skehel, J.J., and Wiley, D.C. (1997). Binding of the influenza
A virus to cell-surface receptors: structures of five hemagglutinin-sialyloligosaccharide
complexes determined by X-ray crystallography. Virology 232, 19-3 1.
Ha, Y., Stevens, D.J., Skehel, J.J., and Wiley, D.C. (2003). X-ray structure of the
hemagglutinin of a potential H3 avian progenitor of the 1968 Hong Kong pandemic
influenza virus. Virology 309, 209-218.
Gamblin, S.J., Haire, L.F., Russell, R.J., Stevens, D.J., Xiao, B., Ha, Y., Vasisht, N.,
Steinhauer, D.A., Daniels, R.S., Elliot, A., Wiley, D.C., and Skehel, J.J. (2004). The
structure and receptor binding properties of the 1918 influenza hemagglutinin. Science
303, 1838-1842.
Chandrasekaran, A., Srinivasan, A., Raman, R., Viswanathan, K., Raguram, S., Tumpey,
T.M., Sasisekharan, V., and Sasisekharan, R. (2008). Glycan topology determines human
adaptation of avian H5N1 virus hemagglutinin. Nat Biotechnol 26, 107-113.
Skehel, J.J., and Wiley, D.C. (2000). Receptor binding and membrane fusion in virus
entry: the influenza hemagglutinin. Annu Rev Biochem 69, 531-569.
Glaser, L., Stevens, J., Zamarin, D., Wilson, I.A., Garcia-Sastre, A., Tumpey, T.M.,
Basler, C.F., Taubenberger, J.K., and Palese, P. (2005). A single amino acid substitution
in 1918 influenza virus hemagglutinin changes receptor binding specificity. J Virol 79,
11533-11536.
Stevens, J., Blixt, 0., Glaser, L., Taubenberger, J.K., Palese, P., Paulson, J.C., and
Wilson, I.A. (2006). Glycan microarray analysis of the hemagglutinins from modem and
pandemic influenza viruses reveals different receptor specificities. J Mol Biol 355, 11431155.
19.
20.
21.
22.
Stevens, J., Blixt, 0., Chen, L.M., Donis, R.O., Paulson, J.C., and Wilson, I.A. (2008).
Recent avian H5N1 viruses exhibit increased propensity for acquiring human receptor
specificity. J Mol Biol 381, 1382-1394.
Stevens, J., Blixt, 0., Tumpey, T.M., Taubenberger, J.K., Paulson, J.C., and Wilson, I.A.
(2006). Structure and receptor specificity of the hemagglutinin from an H5N1 influenza
virus. Science 312, 404-410.
Srinivasan, A., Viswanathan, K., Raman, R., Chandrasekaran, A., Raguram, S., Tumpey,
T.M., Sasisekharan, V., and Sasisekharan, R. (2008). Quantitative biochemical rationale
for differences in transmissibility of 1918 pandemic influenza A viruses. Proc Natl Acad
Sci U S A 105, 2800-2805.
Maines, T.R., Jayaraman, A., Belser, J.A., Wadford, D.A., Pappas, C., Zeng, H., Gustin,
K.M., Pearce, M.B., Viswanathan, K., Shriver, Z.H., Raman, R., Cox, N.J., Sasisekharan,
R., Katz, J.M., and Tumpey, T.M. (2009). Transmission and pathogenesis of swine-origin
2009 A(H1Nl) influenza viruses in ferrets and mice. Science 325, 484-487.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
Van Hoeven, N., Pappas, C., Belser, J.A., Maines, T.R., Zeng, H., Garcia-Sastre, A.,
Sasisekharan, R., Katz, J.M., and Tumpey, T.M. (2009). Human HA and polymerase
subunit PB2 proteins confer transmission of an avian influenza virus through the air. Proc
Natl Acad Sci U S A 106, 3366-3371.
Itoh, Y., Shinya, K., Kiso, M., Watanabe, T., Sakoda, Y., Hatta, M., Muramoto, Y.,
Tamura, D., Sakai-Tagawa, Y., Noda, T., Sakabe, S., Imai, M., Hatta, Y., Watanabe, S.,
Li, C., Yamada, S., Fujii, K., Murakami, S., Imai, H., Kakugawa, S., Ito, M., Takano, R.,
Iwatsuki-Horimoto, K., Shimojima, M., Horimoto, T., Goto, H., Takahashi, K., Makino,
A., Ishigaki, H., Nakayama, M., Okamatsu, M., Warshauer, D., Shult, P.A., Saito, R.,
Suzuki, H., Furuta, Y., Yamashita, M., Mitamura, K., Nakano, K., Nakamura, M.,
Brockman-Schneider, R., Mitamura, H., Yamazaki, M., Sugaya, N., Suresh, M., Ozawa,
M., Neumann, G., Gem, J., Kida, H., Ogasawara, K., and Kawaoka, Y. (2009). In vitro
and in vivo characterization of new swine-origin HiN1 influenza viruses. Nature 460,
1021-1025.
Wan, H., Sorrell, E.M., Song, H., Hossain, M.J., Ramirez-Nieto, G., Monne, I., Stevens,
J., Cattoli, G., Capua, I., Chen, L.M., Donis, R.O., Busch, J., Paulson, J.C., Brockwell,
C., Webby, R., Blanco, J., Al-Natour, M.Q., and Perez, D.R. (2008). Replication and
transmission of H9N2 influenza viruses in ferrets: evaluation of pandemic potential.
PLoS One 3, e2923.
Tumpey, T.M., Maines, T.R., Van Hoeven, N., Glaser, L., Solorzano, A., Pappas, C.,
Cox, N.J., Swayne, D.E., Palese, P., Katz, J.M., and Garcia-Sastre, A. (2007). A twoamino acid change in the hemagglutinin of the 1918 influenza virus abolishes
transmission. Science 315, 655-659.
Maines, T.R., Chen, L.M., Matsuoka, Y., Chen, H., Rowe, T., Ortin, J., Falcon, A.,
Nguyen, T.H., Mai le, Q., Sedyaningsih, E.R., Harun, S., Tumpey, T.M., Donis, R.O.,
Cox, N.J., Subbarao, K., and Katz, J.M. (2006). Lack of transmission of H5N1 avianhuman reassortant influenza viruses in a ferret model. Proc Natl Acad Sci U S A 103,
12121-12126.
Rogers, G.N., and Paulson, J.C. (1983). Receptor determinants of human and animal
influenza virus isolates: differences in receptor specificity of the H3 hemagglutinin based
on species of origin. Virology 127, 361-373.
van Riel, D., den Bakker, M.A., Leijten, L.M., Chutinimitkul, S., Munster, V.J., de Wit,
E., Rimmelzwaan, G.F., Fouchier, R.A., Osterhaus, A.D., and Kuiken, T. Seasonal and
pandemic human influenza viruses attach better to human upper respiratory tract
epithelium than avian influenza viruses. Am J Pathol 176, 1614-1618.
van Riel, D., Munster, V.J., de Wit, E., Rimmelzwaan, G.F., Fouchier, R.A., Osterhaus,
A.D., and Kuiken, T. (2006). H5N1 Virus Attachment to Lower Respiratory Tract.
Science 312, 399.
Wilson, I.A., Skehel, J.J., and Wiley, D.C. (1981). Structure of the haemagglutinin
membrane glycoprotein of influenza virus at 3 A resolution. Nature 289, 366-373.
Rogers, G.N., Paulson, J.C., Daniels, R.S., Skehel, J.J., Wilson, I.A., and Wiley, D.C.
(1983). Single amino acid substitutions in influenza haemagglutinin change receptor
binding specificity. Nature 304, 76-78.
Bean, W.J., Schell, M., Katz, J., Kawaoka, Y., Naeve, C., Gorman, 0., and Webster, R.G.
(1992). Evolution of the H3 influenza virus hemagglutinin from human and nonhuman
hosts. J Virol 66, 1129-1138.
34.
35.
36.
37.
38.
39.
Connor, R.J., Kawaoka, Y., Webster, R.G., and Paulson, J.C. (1994). Receptor specificity
in human, avian, and equine H2 and H3 influenza virus isolates. Virology 205, 17-23.
Sui, J., Hwang, W.C., Perez, S., Wei, G., Aird, D., Chen, L.M., Santelli, E., Stec, B.,
Cadwell, G., Ali, M., Wan, H., Murakami, A., Yammanuru, A., Han, T., Cox, N.J.,
Bankston, L.A., Donis, R.O., Liddington, R.C., and Marasco, W.A. (2009). Structural
and functional bases for broad-spectrum neutralization of avian and human influenza A
viruses. Nat Struct Mol Biol 16, 265-273.
Ekiert, D.C., Bhabha, G., Elsliger, M.A., Friesen, R.H., Jongeneelen, M., Throsby, M.,
Goudsmit, J., and Wilson, I.A. (2009). Antibody recognition of a highly conserved
influenza virus epitope. Science 324, 246-251.
Yang, Z.Y., Wei, C.J., Kong, W.P., Wu, L., Xu, L., Smith, D.F., and Nabel, G.J. (2007).
Immunization by avian H5 influenza hemagglutinin mutants with altered receptor binding
specificity. Science 317, 825-828.
Pappas, C., Viswanathan, K., Chandrasekaran, A., Raman, R., Katz, J.M., Sasisekharan,
R., and Tumpey, T.M. Receptor specificity and transmission of H2N2 subtype viruses
isolated from the pandemic of 1957. PLoS One 5, el 1158.
Hensley, S.E., Das, S.R., Bailey, A.L., Schmidt, L.M., Hickman, H.D., Jayaraman, A.,
Viswanathan, K., Raman, R., Sasisekharan, R., Bennink, J.R., and Yewdell, J.W. (2009).
Hemagglutinin receptor binding avidity drives influenza A virus antigenic drift. Science
326, 734-736.
Chapter 2
Glycan Receptor Specificity of Recent Human-Adapted H3
Viruses
Summary
A critical step for avian influenza viruses to infect human hosts and cause epidemics or
pandemics is acquisition of the ability of the viral hemagglutinin (HA) to bind human receptors.
However, current global influenza surveillance does not monitor HA receptor binding
specificity. Here, we report the results of a collaborative study with Professor Taijiao Jiang of the
Institute of Biophysics, Chinese Academy of Sciences, Bejing, China. Professor Jiang and group
have developed a computational method that uses an effective energy function to quantify HAreceptor binding activities with high accuracy and speed. Application of this method reveals
receptor specificity changes and its temporal relationship with antigenic changes during the
evolution of human H3N2 viruses. The method predicts that two amino acid differences between
HAs of A/Fujian/411/02 and A/Panama/2007/99 viruses account for their differences in binding
to both avian and human receptors; this prediction was verified experimentally. The new
computational method could provide an urgently needed tool for rapid and large-scale analysis of
HA receptor specificities for global influenza surveillance. Furthermore, the two amino acid
positions form potential interactions with each other to influence receptor binding strength,
suggesting a role for interactions between amino acids of the RBS in HA receptor binding.
2.1
Introduction
As described in Chapter 1, a critical step for avian influenza viruses to infect human hosts
and cause epidemics or pandemics is acquisition of the ability of the viral HA to bind human
receptors. However, current global influenza surveillance does not monitor HA receptor binding
specificity due to a lack of rapid and reliable assays. In addition to influenza surveillance, the
receptor specificity monitoring of HA is also critical for the timely production of influenza
vaccines, because the most widely used method for vaccine production requires growing human
influenza viruses in embryonated chicken eggs. If the HA of the human virus does not bind well
to receptors in chicken eggs, vaccine production could be adversely affected.
The World Health Organization (WHO) recommended the use of H3N2 influenza
A/Fujian/411/02 (Fujian) virus for vaccine production for the 2003-2004 flu season. However,
the A/Fujian/411/02 virus failed to replicate well in chicken eggs [1, 2], possibly due to weak
HA binding to sialic acid receptors in the allantoic and amniotic cavities. As a result, an earlier
H3N2 strain of A/Panama/2007/99 (Panama) had to be used in its place for vaccine production.
The antigenic difference between the two strains was so large that immunity induced by the
Panama virus did not protect against infection by the Fujian-like viruses, rendering the vaccine
ineffective during the 2003-2004 flu season [3]. Differences in amino acids at positions 155 and
156 account for the antigenic differences between the Panama and Fujian viruses [2]. Using
reverse genetics technology, it has also been shown that two amino acid changes of either
G186V and V2261, or H183L and V226A are sufficient for the Fujian virus to adapt for growth
in eggs [1]. Despite these progresses, little is known about the molecular basis for the altered
receptor binding specificity in the evolution of Fujian-like viruses from the Panama virus, or the
evolutionary relationship between changes in antigenicity and receptor binding specificity.
Given that the receptor binding specificity of HA directly affects influenza transmission
[4] from avians to humans, it is imperative to develop robust and effective methods for
monitoring changes in the receptor binding specificity of influenza viruses as they evolve. Major
efforts have been devoted to understand the molecular mechanisms governing HA binding
specificities by determining crystal structures of HA-receptor complexes and analyzing HAglycan binding using glycan arrays [5-13]. Computational models have also been developed to
assess HA-receptor binding on a large scale. Several studies have used the ab initio fragment
molecular orbital method, in conjunction with molecular dynamics and molecular mechanics
approaches, to calculate HA-receptor binding energy [4, 14, 15]. However, the utility of these
computational methods is limited because they lack experimental validation and tend to be
computationally expensive, which poses a barrier for practical applications in influenza
surveillance.
Here, we report the results of a collaborative study with Professor Taijiao Jiang of the
Institute of Biophysics, Chinese Academy of Sciences, Bejing, China. Professor Jiang and group
have developed a computational method that uses an effective energy function to quantify HAreceptor binding activities with high accuracy and speed. We demonstrate the utility of this
method, with experimental verification, in identifying molecular events that underlie the receptor
specificity changes during the evolution of human H3N2 Fujian-like viruses.
2.2
Results
2.2.1 A rapid computational method for predicting HA-receptor binding activity
The binding of an influenza virus to a host cell can be approximated by the interaction
between HA and the host glycan receptor [16]. In this study, the a2-3 pentasaccharide (LSTa,
Neu5Aca(2-3)Galp(1-3)GlcNAcp(1-3)Galp(1-4)Glc)
and
a2-6
pentasaccharide
(LSTc,
Neu5Aca(2-6)Galp(1-4)GlcNAcp(1-3)Galp(1-4)Glc) were used for avian and human receptor
analogs, respectively. To model HA-receptor binding activity, a computational method was
designed consisting of four steps. First, the target HA structure was constructed based on
homology modelling using an available HA crystal structure that is highly similar to the target
HA (>70% sequence identity). Second, the HA structure was aligned against known HA-receptor
complex templates and the binding position of the receptor analog in the HA receptor binding
domain was initialized. Third, the conformations of side chains at receptor binding sites were
refined. Fourth, the receptor position was optimized according to an energy function that
calculates the binding energy between HA and the receptor analog, which is described below in
equation 1. The fourth step was iterated until the minimum binding energy between HA and
receptor analog was achieved as an output. The computations took less than an hour on an Intel
Xeon 2.8 GHz processor, much less computationally demanding than traditional molecular
dynamics techniques that usually took weeks for similar tasks. Thus, the approach developed
here can be used in large-scale applications such as the rapid monitoring of evolution of
influenza receptor binding specificities.
2.2.2 A virtual water modelfor evaluating HA-receptor binding energy
The accurate modelling of HA-receptor binding activity requires an effective energy
function that can quantify the binding energy between HA and its receptor. To this end, an
empirical energy function based on a simple virtual water model that takes into account the
effects of solvents on molecular interactions was used. In this virtual water model, water
molecules were added onto the surfaces of target molecules for quantifying their hydrogen
binding and hydrophobic interactions with the target molecules, which constitute a solvation
energy (Gsr,). The hydrogen binding energies between the solvents and ionizable atoms (GwaterhbA)
or other polar atoms (Gwater-hbB) were quantified by calculating probabilities of hydrogen
bond formation. The hydrophobic interaction energy was quantified by the shape
complementarity of the layer of added water molecules on the surface, which is denoted as
Gshape.
In addition to solvation energy, we considered two classical intra- or intermolecular
interaction potentials: van der Waals interaction (Evdw) and Coulombic electrostatics (Eeiec).
Therefore, in the virtual water model, the energy function can be represented as a linear
combination of the energy terms:
AG =
wilEdu + W)Eei,
where wi,
+
W2, W3, W4 and
wvAGw,,r-bA
+
W4 AGaer-hbB
+ 'V5AGhop(,
ws are the relative weights of the different energy terms.
Assuming that AGeompex,
AGHA
and AGreceptor are the energies of the HA-receptor complex,
unbound HA and the receptor respectively, binding energy (AGbind) between HA and receptor
analog, which indicates HA-receptor binding strength, was calculated as follows:
AGb,,d =AG,,O - AG
r-eAGcepr
(2)
In order for the energy function to accurately quantify the binding strength between HA and
receptor, we parameterized the relative weights of the energy terms in equation 1 using proteinprotein interaction data. The availability of high resolution crystal structures of many different
HAs in complex with analogs of both avian and human receptors allowed us to test whether the
energy function that was parameterized on protein-protein interaction data could be used to study
HA-receptor interactions [5, 6, 12]. A good energy function should be able to differentiate
between conformations close to and far from the native conformations, which should adopt the
lowest possible binding energies. We used eight structures of HAs of human HlN1, avian H3N2,
avian H5N1 and swine H9N2 viruses in complex with both avian and human receptor analogs
(PDB ID: 1RVZ, 1RVX, 1MQM, lMQN, 1JSH, JSI, 1JSO, lJSN). For each complex, a set of
structures was generated by displacing receptor positions in all three dimensions at steps of 0.5
A, ranging from 0.5 A to 5 A root mean square deviation (RMSD), and 5' rotation around each
axis, and then their binding energies were calculated using equations 1 and 2. In all the
complexes the native binding positions were located at or near the lowest interaction energy, and
the further they were moved away from the native binding positions, the higher the binding
energy became (Figure 2. lA). The test also allowed assessment of how individual energy terms
such as solvation energy, van der Waals energy and electrostatic energy contributed to HAreceptor interactions. As shown in Figure 2.1B-D, the solvation energy showed a better
correlation with RMSD than the van der Waals and electrostatic energy. This indicates that the
effect of water molecules plays a major role in the HA-receptor binding, which is consistent with
the observation made from the analysis of the co-crystal structures of HA with receptor analog.
Thus, the energy model is able to differentiate native conformations from non-native
conformations at a high resolution, suggesting the validity of using the binding energy to
estimate HA-receptor binding strength.
2.2.3 Glycan receptor binding specificity of recent human H3N2 viruses
We determined the molecular events underlying receptor specificity changes in the
evolution of Fujian-like viruses by combining the binding energy predictions with large-scale
HA sequencing. HA genes from a total of 207 human H3N2 viruses isolated from diverse
regions in China between 2000 and 2002 were sequenced. Based on these sequences,
phylogenetic analyses of these H3N2 HAs revealed two temporally distinct clades that we
designate as Panama and Fujian, after the WHO-recommended vaccine strains of the
10
0
E
-10
-20
.R
-30
-40
10
0
E
-10
-20
-30
-40
B
0
1
2
3
4
5
6
0
1
20
10
0
-10
-20
-30
20 10
0
-10
-20
-30
1
2
3
4
RMSD (A)
3
4
5
6.
5
6.
RMSD (A)
RMSD (A)
0
2
5
6
0
1
2 3 4
RMSD (A)
Figure 2.1 Correlation of energy functions with perturbations of crystal structures of H3N2 HA in
complex with an avian receptor analog. The binding position of receptor analog with HA in the
complex (PDB ID: 1MQM) was altered by translating receptor positions at steps of 0.5 A, ranging from
0.5 A to 5 A RMSD in each direction of Cartesian space and by looping rotations of 5* around Cartesian
axis for each of the translation. The energy of each conformation was minimized by tweaking side chain
torsion angles. Then, the changes of total binding energy AGbi,, (A), solvation term AG,,, (B), van der
Waals term AGydw (C) and electrostatic term AG,, (D), respectively, with degree of perturbation with
RMSD are recorded.
Panama and Fujian viruses (Figure 2.2A). The Fujian clade appeared in China in the 2002-2003
flu season, one season earlier than in the United States. To gain molecular insights into the
evolution of the Fujian clade, we further tracked the detailed amino acid changes over time. As
shown in Figure 2.2B, this method revealed remarkable sequence diversity present in the HA1
subunit, allowing a visual representation of almost all amino acid substitution intermediates
between the Panama-like viruses and the Fujian-like viruses over time.
To investigate the receptor binding changes during the evolution of the Fujian-like
viruses, the binding energies of HA to both the human and avian receptor analogs were
calculated for each of the 207 virus isolates. Figure 2.2C shows the dynamic change of binding
strength of HA to both the human and avian receptor analogs during evolution by tracking the
binding energies for each HA on the phylogenetic tree. It is noticeable that the increase in
binding energy (i.e. decrease in binding strength) for both the avian receptor and human receptor
occurred twice during the evolution of Fujian-like viruses from Panama-like viruses: the first one
occurred at the beginning of the 2001-2002 flu season and the second one at the beginning of the
2002-2003 flu season. A side-by-side, visual comparison of amino acid changes (Figure 2.2B)
and receptor binding changes (Figure 2.2C) allows association of specific amino acid change(s)
with alterations in receptor binding properties. For example, the mutations of W222R and
G225D located at the receptor binding region took place at the beginning of the 2001-2002 flu
season, coinciding with the time when predicted binding strength decreased. Upon modelling the
effect on receptor binding of each mutation at all thirteen sites that differed between
A/Panama/2007/99 and A/Fujian/411/02 viruses, we found that the mutations at residues 222 and
225 are important for changing the receptor binding properties of these human H3N2 HAs.
Interestingly, these two residues which are proximally positioned to each other share the same
....
..
. ....
.
C
B
A
......
...........................
.......
0.6
L
G
A
CO Q156HC
N
L25IH75QH155
A131T
MU
N
-0
0.4
R
'T
H
Q
W
1
E
K
Y
0.2
0.0
-0.2
P
N
CC
x
cc
S186G
E83K|V202
W222RR50G
G225D
thoi U aafa2-3
a2-6
Figure 2.2 Evolution of receptor binding specificity of the Panama- and Fujian-like viruses.
(A) Phylogenetic analysis of human H3N2 viruses isolated in China from 2000 to 2002, covering
flu seasons from 1999-2000 to 2002-2003. The stars denote the positions of Panama and Fujian
viruses on the phylogenetic tree. Key amino acid changes are shown at the indicated positions
during the evolution from Panama-like viruses to Fujian-like viruses. (B) Dynamic changes of
amino acids at 13 sites that differ between Panama-like viruses and Fujian-like viruses. Right:
The color code of amino acids is shown. 'X' and '-' indicate unknown amino acid and gaps,
respectively, in the HA sequence. (C) Comparison of calculated binding energies for HAs to both
avian (a2-3) and human (a2-6) receptor analogs, normalized to those of HA of Panama. The heat
map of binding energies is aligned with the corresponding HAs in the phylogenetic tree to
indicate dynamic change of the binding energies of these viral HAs to both avian and human
receptors during evolution. Right: Scale of normalized binding energy values is shown.
biochemical properties: they are either neutral W and G or polar R and D. This common
characteristic suggests a potential interaction between the amino acids 222 and 225 that forms an
environment conducive for receptor binding.
2.2.4 Experimental validation of residues critical for receptor binding
To experimentally validate if the amino acid changes at residues 222 and 225
significantly changed HA receptor binding specificity, we mutated the residues 222 and 225 in
HA of the Fujian virus (R and D) back to the corresponding residues of HA of the Panama virus
(W and G) and measured binding of the double mutant and two wildtype HAs with either a2-3 or
a2-6 linked sialic acid receptors by a hemadsorption assay (see 2.4 Materials and Methods). In
agreement with computational prediction, the wild type Fujian virus HA exhibited a significantly
weaker binding to both a2-3 and a2-6 receptors than the wildtype Panama virus HA (Figure 2.3).
The Fujian double mutant containing corresponding residues of the Panama virus HA showed a
significant increase in binding activity to both a2-3 and a2-6 receptors. These results confirm the
prediction, showing that the changes in HA residues 222 (W to R) and 225 (G to D) indeed
caused a significantly low receptor binding activity for a2-3 sialic acid receptor during the
evolution of Fujian-like viruses.
2.3
Discussion
In this chapter, a novel computational model for measuring interaction strengths between
influenza HA and its host cell receptors is described. The computational method can predict
binding strengths of a wide variety of influenza HA, and has been tested for accuracy by
correlation with experimental measurements. Application of this computational method has
enabled us to identify how receptor specificities change during the evolution of human H3N2
46
*
0.10-
r-
-*
m
Tm
*
0.080.06-
0.040.020.00
GFP PnWT FjW
a2-3
GFP PnWT FJWT
a2-6
Figure 2.3 Experimental validation of amino acid residues critical for altered receptor binding
specificity of Fujian-like viruses. Comparison of receptor binding activities of wild type Panama and
Fujian HAs and Fujian HA with R222W and D225G mutations (FjR222W/D225G). Briefly, the wild type
and mutant HAs were expressed on the surface of 293T cells. Sialic acid was removed from chicken red
blood cells with neuraminidase and resialylated to express either a2-3 or a2-6 linked sialic acid. The
amount of red blood cells bound to HA expressed on the 293T cell surface was measured by absorbance
at 540 nm. GFP-transfected 293T cells were used as a control for nonspecific binding. *p<0.05.
Fujian-like viruses. We predicted and further validated by experimentation that W222R and
G225D mutations in HA resulted in a change in receptor binding specificity during the evolution
of the Fujian-like viruses.
The residues at 222 and 225 stand out from the other eleven amino acid positions that
differ between Panama and Fujian in that 1) 222 and 225 are both positioned on the 220 loop of
the RBS, an important secondary structural element for receptor binding, 2) 222 and 225 share
the same biochemical characteristics even as they evolve (neutral W and G as opposed to polar R
and D), 3) the 220 loop is shaped such that 222 and 225 can form potential interactions with each
other. We have shown that the simultaneous change in sequences in 222 and 225 is important for
defining receptor binding strength in H3N2 viruses. Furthermore, the potential interactions
between these adjacent residues indicate that amino acid interactions in the RBS do play a role in
determining the evolution of residues and ultimately, receptor binding properties of HA.
Accurate and timely monitoring of the evolution of HA's receptor specificity is critical
for global preparedness for influenza epidemics and pandemics. Since their introduction into
humans, the receptor-binding specificity of human H3N2 viruses has changed constantly [8].
Several groups have investigated the molecular basis underlying the poor replication of the
Fujian-like viruses in embryonated chicken eggs. Using reverse genetics, Lu et al. found that the
unbalanced HA receptor-binding activity and NA enzymatic activity in the Fujian virus
contribute to its poor growth in embryonated chicken eggs [1]. Better virus growth could be
achieved by either increasing the HA receptor-binding activity via G186V and V2261 mutations
in HA or lowering the NA enzymatic activity via El 19Q and Q136K mutations in NA [1]. Three
different pairs of mutations, including G186V and V2261, H183L and V226A, or H183L and
D188Y, have been identified in adaptation of Fujian virus in eggs [1, 17]. However, none of
these mutations have appeared in either Fujian or Panama clade, raising the question of whether
they are likely to mediate the evolution of Fujian-like viruses in nature.
Studies have shown that H155T and Q156H substitutions in HA were sufficient to render
the Panama virus antigenically equivalent to A/Wyoming/03/03, an A/Fujian/411/02-like H3N2
virus [2]. While antigenicity changes at residues 155 and 156 occurred in the 2002-2003 flu
season, changes that impacted on receptor binding occurred over a year earlier at the beginning
of the 2000-2001 flu season. It can be envisioned that such findings relating antigenic change to
receptor binding change are of critical importance for global influenza surveillance as they can
alert us to imminent influenza epidemics or pandemics.
2.4
Materials and Methods
Computational method for predicting HA-receptor binding energy
Structure templates for HA-receptor complexes were obtained from Protein Data Bank (PDB).
The HA moiety of 1RVX of H1N1 virus, 1MQN of H3N2 virus, and 2IBX of H5N1 virus were
used as HA templates for their respective virus subtypes. The avian and human receptor analogs
were LSTa (Neu5Aca(2-3)Galp(1-3)GlcNAc (1-3)Galp(1-4)Glc)
and LSTc (Neu5Aca(2-
6)Galp(1-4)GlcNAcp(1-3)Galp(1-4)Glc), respectively. The computational method to calculate
HA-receptor binding strength consists of four steps. In step 1, the structure of the target HA
sequence is built. The HA sequence is first aligned to the template HA of same subtype virus by
CLUSTALW, is then threaded to the template, and finally the conformations of its side chains
are modeled by a fast side chain construction program, SCWRL4. Any site with a gap or
insertion is automatically ignored. In step 2, the receptor analog is transferred to the newly built
HA in the same position as that relative to template HA in the template. Its coordinates were
manually adjusted to avoid steric clashes with HA templates. In step 3, the amino acid side
chains which contain atoms within a 12 A distance from the receptor analog are repacked using a
heuristic iteration search algorithm to optimize side chain conformations sequentially based on
an empirical scoring function parameterized over known HA structures. This scoring function
contains van der Waals, salt bridge and solvation effects. The side chain conformations use the
Dunbrack rotamer library. In step 4, a docking procedure is performed to search for the best
binding position of HA with the receptor analog guided by an effective empirical energy function
we developed based on a virtual water model. The receptor analog performs a random walk bias
to the direction of van der Waals and electrostatic force in the space at the binding region while
the HA protein is kept static. At each step of random walk, the receptor analog is translated in all
three dimensions and rotated around each axis. The binding energy is evaluated and compared to
the energy of the previous step according to the Metropolis criteria at the new position. The
fourth step is iterated until the minimum binding energy is achieved and outputted as the binding
energy.
Sequencing and sequence analysis
HA sequences used in this analysis were generated at the Chinese Influenza Center as part of an
ongoing routine genetic analysis of HA genes of variant and typical influenza field strains.
Hemadsorption glycan-binding assay
The hemadsorption glycan-binding assay protocol was modified from Glaser et al. [18]. 10%
chicken red blood cells (Charles River Laboratories) were incubated with 50 mU neuraminidase
(Roche) at 37*C for 1 h to remove all of the sialic acid. Complete removal of sialic acid was
tested by the hemagglutination assay using the A/PuertoRico/8/34 virus, and unmodified chicken
red blood cells were used as controls. Resialylation of chicken red blood cells was performed by
incubating 100 pl chicken red blood cells with either 0.5 mU a2-3-(N)-sialyltransferease
(Calbiochem) or 1.0 mU a2-6-(N)-sialyltransferase (a gift from Dr. James Paulson, Scripps), 1.5
mM CMP-sialic acid at 37*C for 1 h. 293T human embryonic kidney cells in 6-well plates were
transfected with 2 pg of HA in pcDNA3.1 vector using Mirus transfection reagent (Mirus Bio,
LLC). 48 hrs later, the transfected 293T cells were treated with 5 mU neuraminidase for 30 min
to remove sialic acids on the expressed HA that may interfere with glycan-binding, and washed
three times in PBS, pH 7.4. 2ml of 0.5% resialylated chicken red blood cells were incubated with
the 293T cells at 4*C for 30 min, and carefully washed three times in PBS to remove unbound
chicken red blood cells. The bound chicken red blood cells were lysed in 375 pl of lysis buffer
(20 mM Tris-HCl, 177mM NH4Cl, pH 7.5) for 30 min at room temperature. The amount of
bound chicken red blood cells was quantified by measuring the absorbance of released heme
groups at 540 nm. The negative control which consists of 293T cells not transfected with HA but
expressed GFP shows consistently low, background signals as compared to cells expressing HA.
It implies that 1) chicken red blood cells have been successfully resialylated, and 2) resialylated
chicken red blood cells discriminate between HA and non-HA expressing 293T cells.
References
Lu, B., Zhou, H., Ye, D., Kemble, G., and Jin, H. (2005). Improvement of influenza
A/Fujian/411/02 (H3N2) virus growth in embryonated chicken eggs by balancing the
hemagglutinin and neuraminidase activities, using reverse genetics. J Virol 79, 67636771.
Jin, H., Zhou, H., Liu, H., Chan, W., Adhikary, L., Mahmood, K., Lee, M.S., and
2.
Kemble, G. (2005). Two residues in the hemagglutinin of A/Fujian/411/02-like influenza
viruses are responsible for antigenic drift from A/Panama/2007/99. Virology 336, 113119.
CDC (2004). Assessment of the effectiveness of the 2003-04 influenza vaccine among
3.
children and adults-Colorado,2003. Morb. Mortal. Wkly. Rep. 53, 707-7 10.
1.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
Das, P., Li, J., Royyuru, A.K., and Zhou, R. (2009). Free energy simulations reveal a
double mutant avian H5N1 virus hemagglutinin with altered receptor binding specificity.
J Comput Chem 30, 1654-1663.
Gamblin, S.J., Haire, L.F., Russell, R.J., Stevens, D.J., Xiao, B., Ha, Y., Vasisht, N.,
Steinhauer, D.A., Daniels, R.S., Elliot, A., Wiley, D.C., and Skehel, J.J. (2004). The
structure and receptor binding properties of the 1918 influenza hemagglutinin. Science
303, 1838-1842.
Ha, Y., Stevens, D.J., Skehel, J.J., and Wiley, D.C. (2003). X-ray structure of the
hemagglutinin of a potential H3 avian progenitor of the 1968 Hong Kong pandemic
influenza virus. Virology 309, 209-218.
Kumari, K., Gulati, S., Smith, D.F., Gulati, U., Cummings, R.D., and Air, G.M. (2007).
Receptor binding specificity of recent human H3N2 influenza viruses. Virol J 4, 42.
Matrosovich, M., Tuzikov, A., Bovin, N., Gambaryan, A., Klimov, A., Castrucci, M.R.,
Donatelli, I., and Kawaoka, Y. (2000). Early alterations of the receptor-binding properties
of H1, H2, and H3 avian influenza virus hemagglutinins after their introduction into
mammals. J Virol 74, 8502-8512.
Srinivasan, A., Viswanathan, K., Raman, R., Chandrasekaran, A., Raguram, S., Tumpey,
T.M., Sasisekharan, V., and Sasisekharan, R. (2008). Quantitative biochemical rationale
for differences in transmissibility of 1918 pandemic influenza A viruses. Proc Natl Acad
Sci U S A 105, 2800-2805.
Stevens, J., Blixt, 0., Chen, L.M., Donis, R.O., Paulson, J.C., and Wilson, I.A. (2008).
Recent avian H5N1 viruses exhibit increased propensity for acquiring human receptor
specificity. J Mol Biol 381, 1382-1394.
Stevens, J., Blixt, 0., Glaser, L., Taubenberger, J.K., Palese, P., Paulson, J.C., and
Wilson, I.A. (2006). Glycan microarray analysis of the hemagglutinins from modem and
pandemic influenza viruses reveals different receptor specificities. J Mol Biol 355, 11431155.
Stevens, J., Blixt, 0., Tumpey, T.M., Taubenberger, J.K., Paulson, J.C., and Wilson, I.A.
(2006). Structure and receptor specificity of the hemagglutinin from an H5N1 influenza
virus. Science 312, 404-410.
Stevens, J., Corper, A.L., Basler, C.F., Taubenberger, J.K., Palese, P., and Wilson, I.A.
(2004). Structure of the uncleaved human HI hemagglutinin from the extinct 1918
influenza virus. Science 303, 1866-1870.
Iwata, T., Fukuzawa, K., Nakajima, K., Aida-Hyugaji, S., Mochizuki, Y., Watanabe, H.,
and Tanaka, S. (2008). Theoretical analysis of binding specificity of influenza viral
hemagglutinin to avian and human receptors based on the fragment molecular orbital
method. Comput Biol Chem 32, 198-211.
Sawada, T., Hashimoto, T., Nakano, H., Suzuki, T., Ishida, H., and Kiso, M. (2006). Why
does avian influenza A virus hemagglutinin bind to avian receptor stronger than to human
receptor? Ab initio fragment molecular orbital studies. Biochem Biophys Res Commun
351, 40-43.
Eisen, M.B., Sabesan, S., Skehel, J.J., and Wiley, D.C. (1997). Binding of the influenza
A virus to cell-surface receptors: structures of five hemagglutinin-sialyloligosaccharide
complexes determined by X-ray crystallography. Virology 232, 19-31.
17.
18.
Widjaja, L., Ilyushina, N., Webster, R.G., and Webby, R.J. (2006). Molecular changes
associated with adaptation of human influenza A virus in embryonated chicken eggs.
Virology 350, 137-145.
Glaser, L., Stevens, J., Zamarin, D., Wilson, I.A., Garcia-Sastre, A., Tumpey, T.M.,
Basler, C.F., Taubenberger, J.K., and Palese, P. (2005). A single amino acid substitution
in 1918 influenza virus hemagglutinin changes receptor binding specificity. J Virol 79,
11533-11536.
Chapter 3
Determinants of Glycan Receptor Specificity of H2 and H5 HA
Summary
The highly pathogenic H5N1 avian influenza virus causes fatalities in 60% of human
infections, and is one of the greatest pandemic influenza threats posed at this time. Other avian
influenza viruses, including those of the H9N2, H7N3 and H7N7 subtypes, have also been
known to infect humans. The H2N2 virus, which caused the Asian flu pandemic of 1957-58, is a
particular cause for concern because even though it had stopped circulating in humans by 1968,
strains of the H2 subtype continue to circulate in birds and have the potential of reentering the
human population. A crucial factor in the cross-over and adaptation of avian influenza viruses to
the human host is associated with the selection of HA mutations that transforms its binding
specificity from avian to human glycan receptors, leading to efficient replication and rapid
transmission of the virus. This chapter investigates the human adaptation of avian H2 and
determinants of glycan receptor specificity of H5 HA. The determinants of glycan receptor
specificity of H2N2 HA are investigated using representative HAs from a 1957-58 pandemic
strain and a recent avian strain. Following that, it is demonstrated that mutations in H5N1 HA
based on key amino acids in human adapted H2N2 and H3N2 HAs do not increase H5 binding
affinity to human glycan receptors.
3.1
Introduction
Wild waterfowl are the natural reservoir for influenza viruses and carry them without
exhibiting disease. Avian influenza viruses are, however, very contagious and often pathogenic
in domesticated birds including chickens, ducks and turkeys. Occasionally, avian influenza
viruses can lead to direct infection of humans, mostly when contact has been made with infected
poultry. The highly pathogenic H5N1 avian influenza virus causes fatalities in 60% of human
infections, and is one of the greatest pandemic influenza threats posed at this time [1]. Although
most infections primarily affect the respiratory system, it has recently been reported that the
H5N1 virus can enter the central nervous system and induce neuroinflammation and
neurodegeneration in mice [2]. Other avian influenza viruses, including those of the H9N2,
H7N3 and H7N7 subtypes, have also been known to infect humans.
Efforts are being made to understand the factors that can govern the human adaptation of
avian viruses, an event that will result in widespread and sustained disease in humans. A crucial
factor in the cross-over and adaptation of avian influenza viruses to the human host is associated
with the selection of HA mutations that transforms its binding specificity from avian to human
glycan receptors, leading to efficient replication and rapid transmission of the virus [3-5]. As
described in Chapter 1, such sustained cross-over events have already taken place for the HIN1,
H2N2 and H3N2 subtypes leading to pandemics during the 2 0 th century. The H2N2 virus, which
caused the Asian flu pandemic of 1957-58, is a particular cause for concern because even though
it had stopped circulating in humans by 1968, strains of the H2 subtype continue to circulate in
birds and have been isolated from pigs on several occasions [6-8]. The circulation of H2 subtype
viruses in domestic birds and pigs increases the risk of human exposure to these viruses and
reintroduction of the viruses to the human population. Such a reintroduction will pose a
significant global health threat given the lack of pre-existing immunity in a huge subset of the
human population born after 1968.
This chapter investigates the human adaptation of avian HA, focusing on the H2 and H5
subtypes given the imminent risk they pose in the event of species cross-over. First, the
determinants of glycan receptor specificity of H2N2 HA will be investigated using representative
HAs from a 1957-58 pandemic strain and a recent avian strain. Following that, we demonstrate
that mutations in H5N1 HA made based on key amino acids in human adapted H2N2 and H3N2
HAs do not increase its binding affinity to the human glycan receptors.
3.2
Determinants of Glycan Receptor Specificity of H2N2 HA
Previous structural and biochemical studies have provided insights into interactions of the
RBS of HA with avian and human receptors for both wild type and mutant forms derived from
the 1957-58 H2N2 pandemic strains [9, 10].
However, it has recently been demonstrated that
changes in the interactions between amino acids within and proximal to the RBS, arising from
substitutions due to antigenic drift, have profound effects on HA-glycan interactions which in
turn influence the glycan binding affinity of HA [11, 12]. This observation is particularly
relevant to HA from recent avian H2 strains that have diverged considerably in sequence
compared to the HA sequences of the pandemic H2N2 strains [13]. Therefore in order to monitor
the evolution of recent avian H2 viruses with possible human-adaptive changes, it is important to
understand the mutations in their HAs that would confer human receptor-binding affmnity similar
to that of human-adapted HAs and specifically to that of the H2N2 HA from the 1957-58
pandemic viruses.
In this study, we have systematically analyzed the effects of mutations in the RBS of
pandemic and recent avian H2N2 HAs on their respective glycan binding specificities. The HA
from a representative 1957-58 pandemic H2N2 strain, A/Albany/6/58 (Alb58), was chosen as a
reference
human-adapted
HA. The HA from a representative
avian H2N2
virus,
A/Chicken/Pennsylvania/2004 (CkPA04), which is among the most recent strains isolated from
birds, was also evaluated [13]. We first characterized the glycan receptor-binding affinity and
human respiratory tissue binding properties of these avian and human-adapted H2N2 HAs. The
glycan receptor-binding affinity of HA was quantitatively defined using an apparent binding
constant Kd' that takes into account the cooperativity and avidity in the multivalent HA-glycan
interactions as described previously [14]. Next, using homology-based structural models of
Alb58 HA-human receptor and CkPA04 HA-avian receptor complexes we analyzed the RBS of
these HAs and designed and evaluated mutations in CkPA04 HA that would make its human
receptor binding affinity comparable to that of Alb58 HA.
3.2.1
Results
3.2.1.1 Characterization of glycan receptor binding specificity of A1b58 HA.
We have previously developed a dose-dependent glycan array binding assay [14, 15] to
quantitatively characterize glycan receptor binding affinity of HA by calculating an apparent
binding constant Kd'. Alb58 HA was recombinantly expressed and analyzed on this assay, which
utilizes a2-6 and a2-3 (referred to as 6' and 3' respectively) sialylated glycan receptors linked to
single or multiple lactosamine (referred to as LN) repeats. Alb58 HA bound with high affinity to
the representative human receptors, 6'SLN (Kd'~ 35 pM) and 6'SLN-LN (Kd'~ 5 pM) (Figure
3.lA). Notably, the binding affinity of Alb58 HA to 6'SLN-LN is in the same range as that of
......................................
.............
20
10
5
2
1
0.5
[HA] pglmI
B
Thachea
Mb58 HAIPI
Amyelu
AIb
HAI PI
Figure 3.1 Glycan receptor binding specificity of Alb58 HA. (A) Dose-dependent direct glycan array
binding of A1b58 HA, which shows high affinity binding to human receptors in comparison with avian
receptors. (B) Extensive staining of apical surface of human tracheal epithelia and observable staining of
alveolar tissue section by Alb58 HA (in green) shown against propidium idodide staining (in red).
the pandemic HINi (A/South Carolina/1/1918 or SC18) HA [14]. However unlike SC18 HA,
Alb58 HA also showed substantial binding to the representative avian receptors 3'SLN-LN (Kd'
~1.5 nM) and 3'SLN-LN-LN (Kd'~ 1 nM) on the glycan array (Figure 3.1A). Staining of Alb58
HA on human upper respiratory tracheal tissue sections revealed extensive binding of the protein
to the apical side (Figure 3.1B) and thus correlated with its high affinity binding to human
receptors. Additionally, the substantial a2-3 sialylated glycan binding of Alb58 observed in the
glycan array assay was also reflected in its binding to the human deep lung alveolar tissue
(Figure 3.1B) that predominantly expresses these glycans [14, 15].
Previous studies have pointed to the roles played by the amino acids in positions 226 and
228 in the RBS of H2N2 HAs in governing glycan receptor binding specificity [9, 10]. The
observation includes the fact that HA from most human H2N2 isolates have Leu226 and Ser228
within its RBS, whereas HA from most avian H2 isolates have Gln226 and Gly228. To
understand the roles of these residues on the quantitative glycan receptor binding affinity of
Alb58 HA, three mutant forms of Alb58 were designed. Two of these mutants possess a single
amino acid change, Leu2264Gln (Alb58-QS mutant) and Ser228+Gly (Alb58-LG). The third
mutant carried two amino acid changes, Leu2264Gln / Ser2284Gly (Alb58-QG).
The Alb58-LG mutant retained the human receptor binding specificity of the wildtype
Alb58 HA but showed a complete loss in the avian receptor binding in the dose-dependent direct
glycan binding assay (Figure 3.2A). On the other hand, Alb58-QG mutant showed a complete
loss in human receptor binding and but displayed a substantial binding to avian receptors in
contrast to Alb58 HA (Figure 3.2B). Surprisingly, Alb58-QS mutant exhibited little to no
binding to either the avian or human glycan receptor (Figure 3.2C). Circular dichroism analysis
of Alb58-QS ruled out the possibility of Alb58-QS being misfolded (Figure 3.2E). A homology-
.....................
...
.
....
....
Ab58-LG mutant HA
A
'SUI4N4N
l
6
1
86
fi
6
2
I*
1
.5
a *
6.1
.2
.'L
666
MINOW
A&b-QG Mutant HA
asr
400
3-15W.*
J-20
tedsdpne
gly
ant
bindig s
i
3ILN4J
A
l58as
1
H2AbS
d p
mutant
0:2-1-10
ityby
.
.20L
R618'S6
Alb8-Q
muat(B5ati
2'
16.6
0.,
200
.'&IOMh
MPOW ato
hwna
nbie
220
wihtehmnrcpo
240
Wavelength (tim)
260
2n0
i4MQshwin()
Figure 3.2 Glycan receptor binding specificity of mutant forms of AIb58 HA. Shown in (A)-(C) are
the dose-dependent glycan array binding of AlbS8-LG, Alb58-QG and A~b58-QS mutants respectively. A
single amino acid change from Ser228->Gly (Alb58-LG mutant) leads to a loss of avian receptor binding
observed in AlbS8 HA. An additional Leu226-*Gln mutation (on Alb58-LG) completely transforms the
binding specificity by making the Alb58-QG mutant bind predominantly to avian receptors. Alb58-QS
mutant shows loss of both avian and human receptor binding. Homology based structural model of
Alb58-QS mutant (RBS part is shown as a cartoon in beige) with the human receptor is shown in (D).
Both the Leu226 and Gln226 side chains are marked. The Gln226 in the mutant is positioned to interact
with Ser228 hence making the 226 position less favorable for contacts with both human and avian
receptors. (E) Circular dichroism analysis of Alb58-QS and Alb58 shows no misfolding of the mutant
protein.
based structural model of the Alb58-QS mutant was constructed to investigate the molecular
basis of the observed biochemical binding property. Analysis of the glycan receptor-binding site
of this mutant in the model showed that Ser228 is positioned to form a hydrogen bond with
Gln226 (Figure 3.2D). The interaction between Gln226 and Ser228 potentially disrupts the
favorable positioning of Gln226 for optimal contact with avian receptor. This observation offers
an explanation for the loss of avian receptor binding in the Alb58-QS mutant. Furthermore, the
absence of contacts between Gln226 and human receptor could explain the loss of human
receptor binding. This data emphasizes how interactions between amino acids can influence the
environment necessary for optimal HA-glycan receptor interactions.
3.2.1.2
Mutations in RBS of CkPA04 and their effects on its glycan receptor binding
specificity.
The dose-dependent glycan array binding of CkPA04 HA showed high affinity binding
to the representative avian receptors 3'SLN, 3'SLN-LN and 3'SLN-LN-LN with minimal
binding to human receptors (Figure 3.3A). Furthermore, the glycan array binding properties of
CkPA04 correlated with its extensive binding to the human alveolar tissues and minimal binding
to the apical side of the tracheal tissues (Figure 3.3B).
To understand the molecular aspects of the H2 HA-glycan receptor interactions, we
constructed homology-based structural models of the CkPAO4-avian (Figure 3.4A) and the
Alb58-human receptor complexes (Figure 3.4B). Based on these structural models of CkPA04
and Alb58 HAs, the amino acids positioned to interact with the glycan receptors were compared
(Table 3.1). In addition to the differences in 226 and 228 positions, there were differences in
other positions including 137 and 193. The amino acids at positions 137 and 193 are oriented to
interact with Neu5Acc2-+6Gal motif as well as sugars beyond this motif in the context of the
...
........
..........
....
............
....
CkPA04 HA
a rtN"
* S-suA
31sut44
*SSLN-M
o.
ErstN-.NM
40.
j20.
II
20
*
1e
2
1
0.5
[HAI pWgmI
0.2
0.1
0*5
Figure 3.3 Glycan receptor binding specificity of CkPA04 HA. (A) Dose-dependent direct glycan
array binding of CkPA04 HA, which shows high affinity binding to avian receptors in comparison with
human receptors. (B) Extensive alveolar staining and minimal staining of apical surface of the human
tracheal epithelia by CkPA04 HA (in green) shown against propidium idodide staining (in red).
Figure 3.4. Homology-based structural model of HA-glycan receptor complexes. (A) Stereo view of
the RBS (shown as cartoon in cyan) of CkPA04 HA - avian receptor structural complex constructed using
co-crystal structure of A/Chicken/NY/91-avian receptor (PRB ID: 2WR2) as a template. The resolved
coordinates of the avian receptor (Neu5Aca2-+3Gal 1-+3GlcNAc) are shown using a stick
representation (in green). (B) Stereo view of RBS (shown as cartoon in gray) of Alb58 HA - human
receptor complex constructed using co-crystal structure of A/Singapore/1/57 - human receptor (PDB ID:
2WR7)
as
the template.
The
resolved
coordinates
of the
human
receptor
(Neu5Aca2-+6Galp l--4GlcNAcp1-+3Gal) are shown using a stick representation (in orange). The side
chains of the key residues involved in interaction with glycan receptor are shown and labeled. The
residues in the RBS that differ between CkPA04 and Alb58 HA are labeled in red.
'..,............
......................
............................
human receptor (and potentially play a role in antigenic variations among current strains of
H2viruses; see 3.2.2 Discussion). These differences potentially impinge on the human receptor
binding of H2N2 HA. Notably, CkPA04 HA differs from earlier avian-adapted H2N2 HAs in the
137 and 193 positions. Therefore, while the Gln2264Leu and Gly228+Ser substitutions would
make the RBS of earlier avian-adapted H2N2 HAs almost identical to that of the pandemic
Alb58 HA, additional amino acid changes are required in the more recent avian-adapted HAs,
including CkPA04.
136
137
153
155
156
183
196
187
189
190
193
194
222
226
CkPA04
S
Q
W
T
K
H
N
D
T
R
A
L
K
Q
AIb58
S
R w T
R T
L
K L
K H N D T
Table 3.1 Comparison of key amino acids in the RBS of CkPA04 and Alb58 HAs.
Based on the above analysis, three sets of mutations were progressively made on CkPA04
to improve its contacts with the human receptor. The first mutant comprised of the two amino
acid changes Gln226+Leu / Gly228+Ser (CkPA04-LS). The second mutant, CkPA04-TLS,
included an additional Ala193+Thr amino acid change in the CkPA04-LS HA. The third
mutant, CkPA04-RTLS, was generated by introducing an additional Gln1374Arg mutation in
the CkPA04-TLS HA. These HA mutants were recombinantly expressed and characterized in
terms of their quantitative glycan receptor binding affinity and human tissue binding properties.
CkPA04-LS showed decreased binding to avian receptors and substantial binding to
human receptors in comparison with CkPA04 (Figure 3.5A). CkPA04-TLS showed substantially
higher binding signals to both human and avian receptors when compared to CkPA04-LS (Figure
3.5C).
CkPA06-LS HA
w
'6LN
o r'"N
M35LN4.N
* 6SLN4.N
aSLN4M.4
0.2
01
0.05
CkPA04-TLS HA
CkPAO4-RTS HA
CAPAO4RTS MAI P1
CkPA044RLS "AI PI
DW POW~m
Figure 3.5 Glycan receptor binding specificity of mutant forms of CkPA04 HA. Shown in the figure
is the dose-dependent glycan receptor binding (A, C, E) and human tissue binding (B, D, F) of CkPA04LS, CkPA04-TLS and CkPA04-RTLS mutants respectively. All the mutants show substantial
improvement in the human receptor binding and reduction in avian receptor binding in comparison to the
wildtype CkPA04 HA as observed in both the glycan array tissue-binding experiments.
CkPA04-RTLS on the other hand showed increased binding signals to human receptor and
similar binding signals to avian receptor as compared to CkPA04-LS (Figure 3.5E). The human
respiratory tissue binding of these mutant H2 HAs was in agreement with their observed glycan
array binding (Figures 3.5B, D, F). The dose-dependent glycan binding data of the described
HAs were used to calculate Kd' and n values (n ~ 1.3 for all the HAs) by fitting the binding data
to the Hill equation (for multivalent binding) and this was then used to generate theoretical
binding curves to clearly distinguish the relative binding affinities of wildtype and mutant H2
HAs to representative avian and human receptors (Figure 3.6). The human receptor binding
affinity of CkPA04-LS (Kd'
-
50 pM) was 10-fold lower than that of the Alb58 HA (Kd' ~ 5
pM). On the other hand the human receptor binding affinity of both CkPA04-TLS (Kd' ~ 3 pM)
and CkPA04-RTLS (Kd' ~ 8 pM) were several fold higher than that of CkPA04-LS and in the
same range as that of Alb58 HA. The avian receptor binding affinity of CkPA04-TLS (Kd' ~ 50
pM) was in the same range as that of the wildtype CkPA04 HA (Kd' ~ 20 pM) and several fold
higher than that of CkPA04-LS (Kd' ~ 220 pM) and CkPA04-RTLS (Kd'
220 pM). Therefore,
among the different mutants, CkPA04-RTLS was the closest to Alb58 HA in terms of its relative
human to avian receptor binding affinity.
Based on our structural understanding, this
observation is consistent with the fact that the RBS of CkPA04-RTLS and Alb58 are very similar
to each other, with extended range contacts with the glycan receptor beyond the Neu5Ac linkage.
....
...................
::v:::v:::
............
.........
3'SLN-LN Binding Curves
6'SLN-LN Binding Curves
KX- 60 pM
-..- chmo-LS
-- kA-OuRs
0
2
4
6
8
10
12
14
16
1
2&
I")l POWm
Figure 3.6 Glycan receptor binding affinities of the mutant forms of CkPA04 HA. (A) The
theoretical binding curves (with the apparent binding constant Kd') that depict the differences in the
binding affinity of the wildtype and mutant H2N2 HAs to the representative avian receptor (3'SLN-LN).
(B) The theoretical binding curves that depict the differences in the binding affinity of the wildtype and
mutant H2N2 HAs to the representative human receptor (6'SLN-LN). The range of Kd' values (3-8 pM) is
shown for CkPA04-TLS, Alb58 and CkPA04-RTLS that is contrasted with the Kd' value of CkPA04-LS.
The binding curves were generated by fitting to the Hill equation (see Methods) and plotting the
theoretically calculated fractional saturation (y-axis) against HA concentration (x-axis). The n value for all
the binding events is around 1.3.
3.3
Effect of HA Mutations Gln226Leu and Gly228Ser on Avian H5N1 HA
Receptor Binding Specificity
The pandemic potential of the H5Nl virus has prompted several investigations of H5
receptor binding properties [16-18], including extensive glycan microarray studies of H5
receptor binding [19, 20]. The glycan microarray data for H5N1 virus and recombinant HA
demonstrate that H5N1 is not human-adapted; it exhibits strong a2-3 binding and minimal a2-6
binding, as characteristic of avian-adapted viruses [15]. Given that the NA and PB2 genes of
circulating strains of H5N1 are already human-adapted, it is important to determine the trajectory
of HA mutations that would render it human-adapted. The results raise the question-which
residue switches in H5N1 HA may lead to receptor binding change towards human adaptation?
In avian and human Hi, H2 and H3 HAs, biochemical, biophysical and X-ray crystal
structure studies [9, 10, 21-23] have led to identification of specific residues that play key roles
in determining the glycan receptor binding specificity. For H2 and H3 HAs, sites 226 and 228
have been identified to be important in glycan recognition-avian viruses have Gln226/Gly228
and human viruses have Leu226/Ser228 [23-26]. For the case of HI HA, avian and human
viruses are characterized by Glul90/Gly225 and Aspl90/Asp225 respectively [24]. These Hi,
H2 and H3 human adaptive HA mutations that govern x2-6 binding specificity have been
presented on H5N1 HA. When screened on glycan microarrays [19, 20], none of these H5
mutants showed dramatic receptor binding changes comparable to those elicited in the original
Hi, H2 and H3 HAs though some moderate increase in binding to a2-6 glycans on the glycan
array was observed for Gln226Leu/Gly228Ser double mutants of recent H5N1 clades in
comparison with the wild-type [19, 20]. Based on these results, such mutants were suggested to
provide possible routes by which the H5N1 virus could enter the human population. In spite of
these advances in understanding glycan receptor-binding properties of H5N1 HAs and their
mutants, it has been challenging to quantitatively link the glycan array results with the efficient
human adaptation of these viruses.
In this section, the relative a2-3 and a2-6 binding affinities of wild-type and mutant
forms (comprising the Gln226Leu/Gly228Ser double mutations) of H5N1 HA are systematically
evaluated using dose-dependent direct glycan binding on an array platform. We demonstrate that
mutations in H5N1 HA made based on these key amino acids in human adapted H2N2 and
H3N2 HAs do not increase its binding affinity to the human glycan receptors. Though the mutant
H5N1 HA shows c2-6 binding signals at high HA concentrations (typically used in primary
screening assays with glycan arrays), its relative a2-6 binding affinity is substantially lower than
that of a typical human adapted HA. This study demonstrates the need for using stringent
quantitative biochemical assays to define a2-6 binding specificity to evaluate the human
adaptation of HA.
3.3.1
Results
3.3.1.1 Dose-dependent glycan binding assays of H5N1 HA
The mutations Gln226Leu/Gly228Ser (LS) have been the hallmark for human adaptation
of H2 and H3 HAs. Previously, researchers have introduced LS mutations in H5N1 HA and
investigated their glycan specificity using glycan arrays, which have emerged as useful tools to
screen pathogen binding to diverse glycan receptor motifs [20]. In these solid-phase binding
assays, detection signals at equilibrium reflect the magnitude of HA-glycan binding affinities.
However these assays have typically been performed at a single 'high' concentration of HA, thus
biasing the assays to exaggerate low affinity interactions. To address the above kinetics
consideration, and to capture the effects of multivalent HA-glycan interactions, use a quantitative
70
....
......................................................................................................
100
100-
~80
80"3'LN
60
" 6'SLN
" 3SLN-LN
C 60-
*6'SLN-LN
LJ C
>40 -
40
.
l 3'SLN-LN-LN
20
* 20
0
40 20
0
10 5 2.5 1.25
[HA] p~g/mi
<
JII Ait-.
40 20 10
5
2
1
0.5 0.2 0.1 0.05 0.02 0.01
[HA] pg/mi
100
S100
80
$ 80
C 60
so
60
C
*40
140
20
> 0 20
0
_
40
_
_
20 10
5
2
1
... .laa
.
11.1 as
0.5 0.2 0.1 0.050.020.01
[HA] pg/mi
.1
0
2
4
6
8
10 12
14 16
18 20
[HA] pg/mi
Figure 3.7 Dose-dependent direct glycan binding of wildtype and mutant Viet04 HA. Dosedependent direct glycan binding of (A) wildtype Viet04, (B) VietO4-LS and (C) VietO4-RLS to selected
glycans of different topologies. (D) Comparison of 3'SLN-LN-LN binding curves for VietO4, Viet04-LS
and VietO4-RLS.
biochemical framework based on dose-dependent binding of recombinant HA to a selected panel
of glycans [14, 15]. This biochemical assay, which has been described earlier in this chapter,
enables us to capture subtle differences in the glycan binding preference of different HAs.
Based on the prototypic H5 wildtype A/Vietnam/1203/2004 (Viet04) virus HA, we
created H5 HA mutants with LS alone or in combination with the H5 clade-specific mutation
Lys193Arg (RLS) by site-directed mutagenesis for use on the biochemical assay. These
constructs were expressed in Sf9 cells using the baculovirus system and the purified HA was
used to perform dose dependent glycan array analyses. Both LS and RLS H5 HA mutants
revealed little propensity of H5 HA to bind 6'SLN-LN characteristic of human receptors (Figure
3.7B and C). Further, both LS and RLS retained the predominant binding to a2-3 sialylated
glycans, a characteristic of avian receptors. Similar to wildtype Viet04 (Figure 3.7A), Viet04-LS
showed high affinity binding to the longer a2-3 sialylated glycans (3'SLN-LN; 3'SLN-LN-LN)
and no significant binding to the a2-6 sialylated glyans (6'SLN; 6'SLN-LN) (Figure 3.7B). On
the other hand, a comparison of binding curves depicts that Viet04-RLS displayed relatively
lower affinity to a2-3 sialylated glycans (Figure 3.7C). Our quantitative examination of binding
affinities for Gln226Leu/Gly228Ser mutants in H5 HA shows that substitutions which cause
human adaptation in H2 and H3 subtypes do not confer human adaptation in H5.
3.4
Discussion
Development of an understanding of the factors involved in the human adaptation of
influenza virus is critical for various aspects of influenza preparedness, including the
development of appropriate surveillance measures, preventive strategies and effective treatments.
The emergence of the 2009 HlN 'swine flu' virus strain, that swiftly gained a foothold in the
human population and was declared a pandemic, demonstrates a practical need for better
predictive capability and knowledge of influenza human adaptation. In this light, the highly
pathogenic avian H5N1 virus and the avian H2N2 virus remain pandemic threats. Despite
multiple investigations into how avian HA may adapt to humans, the evolutionary molecular
pathways that may confer human receptor specificity remain elusive.
This study using H2 and H5 HA highlights the importance of integrating a systematic
sequence and structure analysis of HA-glycan molecular interactions and a quantitative binding
assay to study the effects of these interactions on the biochemical glycan receptor binding
affinity of HA. Previous studies have focused on amino acid substitutions in 226 and 228
positions in the RBS of pandemic H2N2 HAs and avian H5N1 HAs [9, 20]. Recently the glycan
receptor-binding properties of the Alb58 virus and the wildtype and mutant forms (with
substitutions in 226 and 228 positions in HA) of a related pandemic H2N2 virus - A/El
Salvador/2/57 were characterized by analyzing these whole viruses in a dose dependent fashion
on the glycan array platform [27]. The glycan receptor-binding properties of wildtype and mutant
forms of the recombinant Alb58 HA reported in the present study are in good agreement with
those obtained using the whole viruses [27]. Our results further augment these observations by
characterizing the effect of substitutions in the 226 and 228 positions on the quantitative glycan
receptor binding affinity of Alb58 HA.
When introduced to H5N1 HA, the substitutions Leu226 and Ser228 do not give rise to
dramatic human receptor binding changes as observed in H2N2 HA. Our conclusion, drawn after
a systematic quantitative dose-dependent glycan binding study, is in contrast to that made by
others [20] by way of single-dose glycan microarray analysis. Despite overall structural
similarities between H2 and H5 HA, the same mutations give rise to subtle changes in the HA
molecular environment that impact receptor binding in different ways. This highlights the
importance of the molecular environment in and proximal to the RBS in determining glycan
receptor binding specificity. Taken together, this study demonstrates the need for quantitative
biochemical assessments of u.2-6 binding specificity to evaluate the human adaptation of HA.
The Kd' values calculated in our study are used primarily to compare the binding
affinities of different recombinant HAs by taking into account a defined spatial arrangement of
HA (that is fixed for all the HAs) relative to the glycans. Among the various factors that
influence the efficient viral transmissibility in humans we have shown in both the 1918 pandemic
H1N1 and the recently declared 2009 pandemic HlN1 that the binding affinity to the human
receptors (quantified using Kd') correlates with the transmissibility of the virus via respiratory
droplets in ferrets [11, 14]. The human receptor binding affinity of Alb58 HA being in the same
range as that of the SC 18 HA, taken together with the efficient respiratory droplet transmission
of the Alb58 virus [27], extends this correlation to the H2N2 viruses. Furthermore, given that the
Alb58 virus transmits efficiently via respiratory droplets in ferrets, our results underscore the fact
that a complete switch from avian to human receptor binding is not the critical determinant for
human adaptation of influenza A virus HAs. Both the quantitative glycan array and human tissue
binding results of Alb58 HA show substantial avian receptor binding. Instead, it appears that the
high affinity binding to human receptors is a common factor shared by H2 HA with that of other
human-adapted virus subtypes (HI and H3) [14, 15] and therefore this property appears to be a
necessary determinant for efficient human adaptation and transmission. In summary, our study
offers important strategies to monitor the evolution of human-adaptive mutations in the HA of
currently circulating avian H2N2 and highly pathogenic avian H5N1 influenza A viruses.
3.5
Materials and Methods
Homology based modeling of CkPA04 HA- and Alb58 HA-glycan structural complexes -
The co-crystal structures of A/Singapore/1/57 H2N2 HA - human receptor (PDB ID: 2WR7) and
A/ck/NewYork/91 - avian receptor (PDB ID: 2WR2) were used as templates to model the
structural complexes of Alb58 - human receptor and CkPA04 - avian receptor respectively.
Homology modeling was performed using the SWISS-MODEL web-based program
(http://swissmodel.expasy.org/SWISS-MODEL.html).
Cloning, mutagenesis and expression of HA - The Alb58 and CkPA04 plasmids were gifts
from Dr. Terrence Tumpey and Dr. Adolfo Garcia-Sastre respectively. The human and avian
wildtype H2N2 HA genes were subcloned into a pAcGP67A vector to generate pAcGp67Alb58-HA and pAcGp67-CkPA04-HA respectively for baculovirus expression in insect cells.
Using pAcGp67- CkPA04-HA as a template the gene was mutated to yield pAcGp67-LS-HA
[Gln226Leu,
Gly228Ser], pAcGp67-TLS-HA
[Alal93Thr, Gln226Leu, Gly228Ser]
and
pAcGp67-RTLS-HA [Glnl37Arg, Alal93Thr, Gln226Leu, Gly228Ser]. pAcGp67-VietO4LSHA and pAcGp67-VietO4RLS-HA plasmids were generated from pAcGP67-VietO4-HA [14] by
[Gln 226Leu; Gly 2 28Ser] and [Lys193Arg; Gln226Leu; Gly 228Ser ] mutations respectively. The
primers for mutagenesis were designed using PrimerX (http://bioinformatics.org/primerxf) and
synthesized by IDT DNA technologies (Coralville, IA). The mutagenesis reaction was carried
out using the QuikChange Multi Site- Directed Mutagenesis Kit (Stratagene, CA). Alb58,
CkPA04, CkPA04-LS, CkPA04-TLS and CkPA04-RTLS baculoviruses were created from their
respective plasmids, using the Baculogold system (BD Biosciences, CA) as per the
manufacturer's instructions. The baculoviruses were used to infect 300 ml suspension cultures of
Sf9 cells (Invitrogen, Carlsbad, CA) cultured in Sf- 900 II SFM medium (Invitrogen, Carlsbad,
CA). The infected cultures were monitored and harvested 4-5 days post-infection. The soluble
trimeric form of HA was purified from the supernatant of infected cells using modification of the
protocol described previously [28]. In brief, the supernatant was concentrated using Centricon
Plus-70 centrifugal filters (Millipore, Billerica, MA) and the trimeric HA was recovered from the
concentrated cell supernatant using affinity chromatography with columns packed with Ni-NTA
beads (Qiagen, Valencia, CA). The fractions containing HA were pooled together and subjected
to ultrafiltration using Amicon Ultra 100 K NMWL membrane filters (Millipore, Billerica, MA).
The protein was reconstituted in PBS and concentrated. The purified protein concentration was
determined using Bio-Rad's protein assay (Bio-Rad, CA).
Dose dependent direct glycan array-binding assay - To investigate the multivalent HA-glycan
interactions a streptavidin plate array comprising representative biotinylated a2-3 and a2-6
sialylated glycans as described previously [14]. The glycans 3'SLN, 3'SLN-LN, 3'SLN-LN-LN
are representative avian receptors. 6'SLN and 6'SLN-LN are representative human receptors.
LN corresponds to lactosamine (Galpl-4GlcNAc) and 3'SLN and 6'SLN respectively
correspond to Neu5Aca2-3 and Neu5Aca2-6 linked to LN. The biotinylated glycans were
obtained from the Consortium of Functional Glycomics through their resource request program.
Streptavidin-coated High Binding Capacity 384-well plates (Pierce) were loaded to the full
capacity of each well by incubating the well with 50 il of 2.4 pM of biotinylated glycans
overnight at 4'C. Excess glycans were removed through extensive washing with PBS. The
trimeric HA unit comprises of three HA monomers (and hence three RBS, one for each
monomer). The spatial arrangement of the biotinylated glycans in the wells of the streptavidin
plate array favors binding to only one of the three HA monomers in the trimeric HA unit.
Therefore in order to specifically enhance the multivalency in the HA-glycan interactions, the
recombinant HA proteins were pre-complexed with the primary and secondary antibodies in the
ratio of 4:2:1 (HA:primary:secondary). The identical arrangement of 4 trimeric HA units in the
precomplex for all the HAs permits comparison between their glycan binding affinities. A stock
solution containing appropriate amounts of Histidine tagged HA protein, primary antibody
(Mouse anti 6X His tag IgG) and secondary antibody (HRP conjugated goat anti Mouse IgG
(Santacruz Biotechnology) in the ratio 4:2:1 and incubated on ice for 20 min. Appropriate
amounts of precomplexed stock HA were diluted to 250 lI with 1%BSA in PBS. 50 tl of this
precomplexed HA was added to each of the glycan-coated wells and incubated at room
temperature for 2 hrs followed by the above wash steps. The binding signal was determined
based on HRP activity using Amplex Red Peroxidase Assay (Invitrogen, CA) according to the
manufacturer's instructions. The experiments were done in triplicate. Minimal binding signals
were observed in the negative controls including binding of precomplexed unit to wells without
glycans and binding of the antibodies alone to the wells with glycans. The binding parameters,
cooperativity (n) and apparent binding constant (Kd'), for H2 HA-glycan binding were calculated
by fitting the average signal value (from the triplicate analysis) and the HA concentration to the
linearized form of the Hill equation: log0 Y)
=
n * log(HA])-
og(Kd'), where y is the fractional
saturation (average binding signal/maximum observed binding signal). The theoretical y values
calculated using the Hill equation Y=
(for the set of n and Kd' parameters) were
[HA]"
[HA]" +Kd
plotted against the varying concentration of HA to obtain the binding curves for representative
human (6'SLN-LN) and avian receptors (3'SLN-LN) shown in Figure 6.
Human respiratory tissue binding assay - Formalin fixed and paraffin embedded normal
human tracheal and alveolar tissue sections were purchased from US Biological and US Biomax,
respectively. Tissue sections were incubated for 30 minutes in a hybridization oven at 60*C to
melt the paraffin. Excess paraffin was removed by multiple washes in xlyene. Sections were
subsequently rehydrated in a series of ethanol washes. In order to prevent nonspecific binding,
sections were pre-blocked with 1% BSA in PBS for 30 minutes at room temperature (RT). For
the generation of HA-antibody precomplexes, the histidine tagged purified recombinant HAs
(Alb58, CkPA04, LS and TLS) were incubated with primary antibody against his tag (mouse anti
6X His tag, Abcam) and secondary (Alexa Fluor 488 goat anti mouse IgG, Invitrogen) antibody
in a ratio of 4:2:1 respectively for 20 minutes on ice. Tissue sections were incubated with the
HA-antibody precomplexed unit, diluted to different final concentrations in 1% BSA-PBS, for 3
hours at RT. In order to demonstrate that the recombinant HAs specifically bound to the
sialylated glycan receptors on the tissue sections, control experiments were carried out by
preincubating the tissue sections with 0.2 units of Sialidase A (recombinant from Arthrobacter
ureafaciens, Prozyme) for 3 hours at 370 C prior to incubation with HA. This enzyme has been
demonstrated to cleave the terminal Neu5Ac from both Neu5Aca2-3Gal and Neu5Aca2-6Gal
motifs. Sections were then incubated with propidium iodide to counterstain the nuclei
(Invitrogen; 1:100 in TBST) for 20 minutes at RT. After thorough washing, sections were
mounted and analyzed using a Zeiss LSM5 10 laser scanning confocal microscope.
References
1.
2.
3.
Organization, W.H. (2010). Cumulative Number of Confirmed Human Cases of Avian
Influenza A/(H5Nl) Reported to WHO.
Jang, H., Boltz, D., Sturm-Ramirez, K., Shepherd, K.R., Jiao, Y., Webster, R., and
Smeyne, R.J. (2009). Highly pathogenic H5N1 influenza virus can enter the central
nervous system and induce neuroinflammation and neurodegeneration. Proc Natl Acad
Sci U S A 106, 14063-14068.
Russell, C.J., and Webster, R.G. (2005). The genesis of a pandemic influenza virus. Cell
123, 368-371.
4.
5.
Yen, H.L., and Webster, R.G. (2009). Pandemic influenza as a current threat. Curr Top
Microbiol Immunol 333, 3-24.
Tumpey, T.M., Basler, C.F., Aguilar, P.V., Zeng, H., Solorzano, A., Swayne, D.E., Cox,
N.J., Katz, J.M., Taubenberger, J.K., Palese, P., and Garcia-Sastre, A. (2005).
Characterization of the reconstructed 1918 Spanish influenza pandemic virus. Science
310, 77-80.
6.
Makarova, N.V., Kaverin, N.V., Krauss, S., Senne, D., and Webster, R.G. (1999).
Transmission of Eurasian avian H2 influenza virus to shorebirds in North America. J Gen
Virol 80 (Pt 12), 3167-3171.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
Schafer, J.R., Kawaoka, Y., Bean, W.J., Suss, J., Senne, D., and Webster, R.G. (1993).
Origin of the pandemic 1957 H2 influenza A virus and the persistence of its possible
progenitors in the avian reservoir. Virology 194, 781-788.
Ma, W., Vincent, A.L., Gramer, M.R., Brockwell, C.B., Lager, K.M., Janke, B.H.,
Gauger, P.C., Patnayak, D.P., Webby, R.J., and Richt, J.A. (2007). Identification of
H2N3 influenza A viruses from swine in the United States. Proc Natl Acad Sci U S A
104, 20949-20954.
Liu, J., Stevens, D.J., Haire, L.F., Walker, P.A., Coombs, P.J., Russell, R.J., Gamblin,
S.J., and Skehel, J.J. (2009). Structures of receptor complexes formed by hemagglutinins
from the Asian Influenza pandemic of 1957. Proc Natl Acad Sci U S A 106, 1717517180.
Xu, R., McBride, R., Paulson, J.C., Basler, C.F., and Wilson, I.A. (2009). Structure,
receptor binding and antigenicity of influenza virus hemagglutinins from the 1957 H2N2
pandemic. J Virol.
Maines, T.R., Jayaraman, A., Belser, J.A., Wadford, D.A., Pappas, C., Zeng, H., Gustin,
K.M., Pearce, M.B., Viswanathan, K., Shriver, Z.H., Raman, R., Cox, N.J., Sasisekharan,
R., Katz, J.M., and Tumpey, T.M. (2009). Transmission and pathogenesis of swine-origin
2009 A(H1Nl) influenza viruses in ferrets and mice. Science 325, 484-487.
Hensley, S.E., Das, S.R., Bailey, A.L., Schmidt, L.M., Hickman, H.D., Jayaraman, A.,
Viswanathan, K., Raman, R., Sasisekharan, R., Bennink, J.R., and Yewdell, J.W. (2009).
Hemagglutinin receptor binding avidity drives influenza A virus antigenic drift. Science
326, 734-736.
Glaser, L., Zamarin, D., Acland, H.M., Spackman, E., Palese, P., Garcia-Sastre, A., and
Tewari, D. (2006). Sequence analysis and receptor specificity of the hemagglutinin of a
recent influenza H2N2 virus isolated from chicken in North America. Glycoconj J 23, 9399.
Srinivasan, A., Viswanathan, K., Raman, R., Chandrasekaran, A., Raguram, S., Tumpey,
T.M., Sasisekharan, V., and Sasisekharan, R. (2008). Quantitative biochemical rationale
for differences in transmissibility of 1918 pandemic influenza A viruses. Proc Natl Acad
Sci U S A 105, 2800-2805.
Chandrasekaran, A., Srinivasan, A., Raman, R., Viswanathan, K., Raguram, S., Tumpey,
T.M., Sasisekharan, V., and Sasisekharan, R. (2008). Glycan topology determines human
adaptation of avian H5N1 virus hemagglutinin. Nat Biotechnol 26, 107-113.
Gambaryan, A., Tuzikov, A., Pazynina, G., Bovin, N., Balish, A., and Klimov, A. (2006).
Evolution of the receptor binding phenotype of influenza A (H5) viruses. Virology 344,
432-438.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
Yamada, S., Suzuki, Y., Suzuki, T., Le, M.Q., Nidom, C.A., Sakai-Tagawa, Y.,
Muramoto, Y., Ito, M., Kiso, M., Horimoto, T., Shinya, K., Sawada, T., Usui, T., Murata,
T., Lin, Y., Hay, A., Haire, L.F., Stevens, D.J., Russell, R.J., Gamblin, S.J., Skehel, J.J.,
and Kawaoka, Y. (2006). Haemagglutinin mutations responsible for the binding of H5N1
influenza A viruses to human-type receptors. Nature 444, 378-382.
Yen, H.L., Aldridge, J.R., Boon, A.C., Ilyushina, N.A., Salomon, R., Hulse-Post, D.J.,
Marjuki, H., Franks, J., Boltz, D.A., Bush, D., Lipatov, A.S., Webby, R.J., Rehg, J.E.,
and Webster, R.G. (2009). Changes in H5N1 influenza virus hemagglutinin receptor
binding domain affect systemic spread. Proc Natl Acad Sci U S A 106, 286-291.
Stevens, J., Blixt, 0., Tumpey, T.M., Taubenberger, J.K., Paulson, J.C., and Wilson, I.A.
(2006). Structure and receptor specificity of the hemagglutinin from an H5N1 influenza
virus. Science 312, 404-410.
Stevens, J., Blixt, 0., Chen, L.M., Donis, R.O., Paulson, J.C., and Wilson, I.A. (2008).
Recent avian H5N1 viruses exhibit increased propensity for acquiring human receptor
specificity. J Mol Biol 381, 1382-1394.
Matrosovich, M., Tuzikov, A., Bovin, N., Gambaryan, A., Klimov, A., Castrucci, M.R.,
Donatelli, I., and Kawaoka, Y. (2000). Early alterations of the receptor-binding properties
of H1, H2, and H3 avian influenza virus hemagglutinins after their introduction into
mammals. J Virol 74, 8502-8512.
Stevens, J., Blixt, 0., Glaser, L., Taubenberger, J.K., Palese, P., Paulson, J.C., and
Wilson, I.A. (2006). Glycan microarray analysis of the hemagglutinins from modem and
pandemic influenza viruses reveals different receptor specificities. J Mol Biol 355, 11431155.
Tumpey, T.M., Maines, T.R., Van Hoeven, N., Glaser, L., Solorzano, A., Pappas, C.,
Cox, N.J., Swayne, D.E., Palese, P., Katz, J.M., and Garcia-Sastre, A. (2007). A twoamino acid change in the hemagglutinin of the 1918 influenza virus abolishes
transmission. Science 315, 655-659.
Glaser, L., Stevens, J., Zamarin, D., Wilson, I.A., Garcia-Sastre, A., Tumpey, T.M.,
Basler, C.F., Taubenberger, J.K., and Palese, P. (2005). A single amino acid substitution
in 1918 influenza virus hemagglutinin changes receptor binding specificity. J Virol 79,
11533-11536.
Bean, W.J., Schell, M., Katz, J., Kawaoka, Y., Naeve, C., Gorman, 0., and Webster, R.G.
(1992). Evolution of the H3 influenza virus hemagglutinin from human and nonhuman
hosts. J Virol 66, 1129-1138.
Connor, R.J., Kawaoka, Y., Webster, R.G., and Paulson, J.C. (1994). Receptor specificity
in human, avian, and equine H2 and H3 influenza virus isolates. Virology 205, 17-23.
Pappas, C., Viswanathan, K., Chandrasekaran, A., Raman, R., Katz, J.M., Sasisekharan,
R., and Tumpey, T.M. Receptor specificity and transmission of H2N2 subtype viruses
isolated from the pandemic of 1957. PLoS One 5, el 1158.
Stevens, J., Corper, A.L., Basler, C.F., Taubenberger, J.K., Palese, P., and Wilson, I.A.
(2004). Structure of the uncleaved human HI hemagglutinin from the extinct 1918
influenza virus. Science 303, 1866-1870.
Chapter 4
Determinants of Glycan Receptor Specificity of H1 HA
Summary
In this chapter, we further examine the effects of H2/H3 HA human-adaptive mutations
Gln226Leu/Gly228Ser on human glycan receptor specificity using a prototypic human adapted
H1N1 HA as a model system. Our data provides evidence that H1 HA responds very differently
from H2/H3 HA to the same Gln226Leu/Gly228Ser substitutions. In the context of H2/H3, these
substitutions confer a2-6 specific binding to HA; in Hi they actually decrease a2-6 binding
when 225 is Asp (SC18-LS), and obliterate all glycan binding when 225 is Gly (NY18-LS).
We
conclude that it is insufficient to consider the receptor binding effects of individual HA residues
in isolation. Aside from binding to glycan receptors, HA itself is heavily glycosylated. The
effects of H1 HA N-glycosylation on influenza receptor binding are also explored in this chapter.
Here, the specific role of a conserved N-glycosylation site in the receptor binding property of the
same H1N1 HA model system is studied by knocking out glycosylation. This study reveals the
effects of subtle HA conformational changes, resulting from HA deglycosylation, on quantitative
receptor binding affinities. Notably, glycans as well as amino acids of HA can participate in
interaction networks to affect receptor binding properties. These interaction networks play a key
role in providing the appropriate environment for optimal HA-glycan receptor interactions.
4.1
Amino Acid Networks Govern H1N1 HA-Glycan Interactions
4.1.1
Introduction
Previous chapters have described the impact of interactions between amino acids, within
and proximal to the RBS, on the glycan receptor binding properties of H2 and H3 HA. Here, the
role that amino acid interaction networks play in governing HA-glycan interactions is further
assessed using the 1918 pandemic H1 model system. An integrated approach that combines
sequence and structure analysis with biochemical assays as earlier presented in Chapter 3 is used.
The implications of introducing cross-subtype human-adaptive mutations on HA are investigated
by incorporating human-adaptive mutations of H2N2 and H3N2 subtypes into a prototypic
human-adapted 1918 H1N1 HA. Chapter 4 further emphasizes that rather than considering
amino acid substitutions in isolation from their surrounding residues, there is a need to carefully
investigate the effect of introducing mutations in the context of their interactions with other
amino acids in the glycan receptor-binding site of HAs.
4.1.2
Results
4.1.2.1 Regions of molecular interactions between HA and sialylated glycans
The results from Chapter 3.3 (Effect of HA Mutations Gln226Leu and Gly228Ser on
Avian H5N1 HA Receptor Binding Specificity) lead us to pose the question-how is it that
Leu226/Ser228 substitutions lead to a gain of a2-6 recognition in the context of avian H2/H3
HA, but do not cause a similar receptor recognition change in the context of avian H5 HA? An
analysis of the structure of HA from various subtypes reveal that the HA of H3 and H5 subtypes
are distinguished by subtle differences in structure and sequence. Firstly, as reported previously
[1], H5 HA has a narrower receptor binding pocket than H3 HA, and one that is more similar to
that of H1 HA (Figure 4.1). This structural difference results in different sets of amino acids
82
....
...................
.......
---
....................................
..
....
.....
......
f---i
H1 AV18-LSTa
H3 DuckUkraine -LSTa
H5 VietQ4-LSTa
Figure 4.1 Sturctural models of avian H1, H3 and 115 HA interacting with LSTa. The membrane
distal portion of HA are shown for avian H1(PDB ID: lruz), H3 (PDB ID: lmql) and H5 HA (PDB ID:
2fko) . Black arrows indicate the widths of the HA receptor binding pockets.
involved in the interaction of HA and glycan receptors. Specifically, a structure-based sequence
comparison of HA of currently circulating H5 (avian) , H3 (avian and human) and H1 (avian and
human) viruses reveals groups of key amino acids that interact with specific regions of the
glycan; these groups of amino acids tend to carry different biochemical characteristics depending
on subtype (Table 4.1).
Extension Region
Base RegkM
H1 MaNardAerta76
DucidAIetal81
DucdtaM
CaO409
SC18
NY18
SC18-LS
NY18-LS
H3 Duckarare3
DuckJHokkaidoi86
MaldIAlaskal07
Aichil68
Panama(99
Fujnf02
H5 Vietl123/04
Viet/1203A04-LS
Viet/1203104-RLS
136
T
T
T
T
T
T
T
T
S
S
S
S
S
S
S
S
S
Cukter I
137 145
A
S
A
S
A
S
A
K
S
A
S
A
A
S
S
A
N
S
N
S
S
N
N
S
K
S
S
K
S
S
S
S
S
S
Neu6Ao-1
in
K
T
K
V
T
T
T
T
T
T
T
T
H
T
I
I
I
222
K
K
K
K
K
K
K
K
W
W
W
W
W
R
K
K
K
Chr 2
25 226 226
G
G
Q
Q
G
G
Q
G
G
G
D Q
G
D Q
G
Q
G
D
L
S
S
G
L
Q
G
G
G
G
Q
G
Q
G
S
G
L
G
VJS
S
D
V
G
G
Q
L
S
G
L
S
G
Gal-2
16
P
P
P
S
P
P
P
P
S
S
S
S
S
G
N
N
N
Custer 3
187 189
S
S
T
G
T
S
T
A
T
T
T
T
T
T
T
T
T
Q
T
Q
T
Q
T
Q
S
T
T
S
D
A
A
D
D
A
GIcNAc-3
190
E
E
E
D
D
D
D
D
E
E
E
E
D
D
E
E
E
192
Q
0
Q
Q
Q
Q
Q
Q
T
T
T
T
I
I
T
T
T
Cuwer 4
193 196 186
Q
K
S
Q
K
S
Q
K
S
K
S
Q
K
S
Q
S
Q
K
S
Q
K
K
S
Q
K
N
V
V
K
N
N
V
K
K
S
V
Q
S
A
H
S
A
K
K
Q
K
Q
K
Q
K
R
Ga-, GIc,...
1N
S
T
S
N
S
S
S
S
S
S
N
S
Y
Y
S
S
S
Table 4.1 Glycan binding residues of Hl, H3 and H5 HAs. Avian residues are shaded in grey. The
residues are organized according to the amino acid clusters that form interaction networks with respective
glycans, as indicated on the bottom row. Clusters 1 and 2 of the base region interact with Neu5Ac-1 and
Gal-2, and clusters 3 and 4 of the extension region interact with GlcNAc-3 and the glycans beyond.
Residues of, or carried over, from the avian consensus sequence are indicated in red. Examples of amino
acid interaction networks of the 220-loop are highlighted in blue.
The interaction of HA with long-chained a2-6 glycans can be divided into base and
extension regions, defined by amino acid contacts with the Neu5Aca2-6Galpl- and the GlcNAcal-3Galpl-4Glcp1- motifs respectively [2]. The base region, which anchors the sialic
acid in a relatively fixed position, includes contacts with key amino acids positioned at 136, 137,
145, 222, 225, 226 and 228, while the extension region involves contacts with amino acids
positioned near or on the 190-loop, including 186, 187, 189, 190 and 193.
84
Performing a
sequence analysis of Hi, H3 and H5 HA, we grouped amino acids of HA that are important for
making contacts with the glycan receptor according to the structural regions in which they lie
(Table 4.1). This visual representation of HA sequences, combined with biochemical and
structural information, enabled us to identify clusters, or networks, of amino acids that form
potential interactions with one another. Table 4.1 highlights one such network formed by the
group of amino acids 222, 225 and 226. This group of amino acids is positioned on the 220-loop
to form contacts with the terminal sialic acid, the penultimate galactose, as well as within the
network. In human-adapted H3 HA, this network tends to favor nonpolar residues, and the
hydrophobic interactions with the glycan and within the network of amino acids dominate. A
change in the characteristic of one amino acid influences a corresponding change on its
neighboring amino acids -
this is apparent in amino acids 222 and 225 which appear to co-
evolve from the nonpolar residues Trp and Gly prevalent in H3 HA before 2000, to polar
residues Arg and Asp in the epidemic-causing H3 Fujian virus of 2002. The avian H5 HA, on
the other hand, bears the consensus sequences Gly and Gln at residues 225 and 226, and
resembles an avian HI rather than an avian H3 HA at the 220-loop because of the presence of the
polar Lys rather than the nonpolar Trp (Table 4.1).
4.1.2.2 H2/H3 human adaptive mutations in H1 HA
To further study the implication of introduction of heterologous HA adaptive mutations,
we chose the 1918 Hi HA system and introduced the H2/H3 adaptive mutations on it. Unlike
human H2/H3 viruses, human H1 viruses have neither Leu226 nor Ser228. In the context of Hi,
the amino acids on the 220-loop tend to have polar residues (Table 4.1) and form an interaction
network. Specifically, the amino acids Lys222, Asp225 and Gln226 of the prototypic 1918
pandemic virus A/South Carolina/1/1918 (SC 18) can form favorable ionic contacts and hydrogen
bonds with the glycan and within the network. Structurally, H1 HA resembles H5 more than H3
given its narrow receptor binding pocket (Figure 4.1). To understand the implication of the
H2/H3 human-adaptive mutations on a different subtype, we studied two classic human-adapted
H1 viruses SC 18 and A/New York/1918 (NYl 8), which have been well-characterized in terms of
their structure, receptor binding properties and transmissibility [3-5]. NYl 8 differs from SC 18 at
position 225 and has Gly225 instead of Asp225 thus affecting the interaction network on the
220-loop. The effect of introducing a nonpolar residue in this network will provide an indication
of how/why avian H5 signaling is affected by Leu226(non-polar)/Ser228.
We have previously shown that SC 18 is an extremely specific binder of long a2-6
glycans while NYl8
displays significant a2-3 and minimal long a2-6 binding [6].
Correspondingly, SC18 transmits efficiently in ferrets but NYl8 does not [5]. In conjunction
with Leu226/Ser228 mutational studies, we modeled and analyzed the structures of SC18 and
NYl 8 to identify possible bond formation between HA and either LSTa (a2-3-linked lactoseries
tetrasaccharide a) or LSTc (a2-6-linked lactoseries tetrasaccharide c) glycans. For this structural
analysis, we aligned the SC18 monomer (PDB ID: lruz) against the HA structures of other
HIl
viruses in complex with glycans: human PR34 with LSTa (PDB ID:1rvx), and swine
SW30 with LSTc (PDB ID: lrvt). In addition to amino acid-glycan interactions, we identified
possible interactions between amino acids that may stabilize or destabilize the amino acid
interaction networks. The amino acid interaction network consisting of, but not limited to amino
acids 222, 225 and 226 was of particular interest as it includes residues important for receptor
specificity switching. In the interaction of SC18 with long a2-6 glycan, there is potential
hydrogen bond formation between Lys222 and Gal-2, Asp225 and Gal-2, Gln226 and Gal-2, and
Gln226 and Neu5Ac (Figure 4.2A). In addition, a hydrogen bond is also formed between the
. ..................
A
B
'222K
2221(
L..-------.B6T
C
L-
D
16T
-------
5DD
1222K
1222K
L6T
226
Figure 4.2 Structurl models of wildtype and mutant HI HA interacting with LSTc. Structure models
of (A) SC18, (B) NY18, (C) SC18-LS and (D) NY18-LS HA are visualized using Pymol. Dotted lines
show hydrogen bond formations between key HA residues and the glycan (green), and the amino acid
interaction network surrounding the base region is outlined in red.
conserved residue Thr136 and Neu5Ac. These interactions, in addition to possible stabilizing
ionic contacts within the network, provide an environment conducive for long a2-6 binding.
The Asp225Gly of NYl8 is part of this amino acid interaction network. A loss of polar
Asp in this network results in a dramatic effect on binding leading to a decrease in long a2-6
binding and an increase in short and long a2-3 binding [2, 6]. From the homology- model of
NYl8, it is apparent that the hydrogen bond contact between 225 and the glycan that is seen in
SC18 is absent here (Figure 4.2B). Besides, the absence of Asp225 also has an impact on the
NY18 amino acid interaction network- preventing the interaction between Asp225 and Lys222
and potentially destabilizing the entire network of amino acids around it. Mutating Gln226 to
polar Leu226 in NY18-LS further disrupts the interaction network; although there is some
compensation of hydrogen bond formation from Ser228 (Figure 4.2D). Indeed, expression of
NYl8-LS and its glycan array analysis reveals a complete loss of binding to all a2-3 and a2-6
glycans (Figure 4.3B). In contrast to the H3 system which has a nonpolar network and shows
predominant a2-3 binding in the presence of Leu226/Ser228, the dramatic loss of glycan binding
in NYl8-LS suggests that the introduction of the nonpolar Leu226 into a predominantly polar
group of amino acids disrupts the network of interactions. When the Gln226Leu/Gly228Ser
substitutions were incorporated on SC 18, the expressed protein showed a decreased in its binding
affinity toward long a2-6 glycans (Figure 4.3A and C). Significantly, there was no concomitant
increase in a2-3 binding. On SC18-LS, the presence of polar Asp225 was able to rescue the
compelte loss of binding observed in NY18-LS to some extent (Figure 4.3C). Asp225 likely
provides an important stabilizing contact within the polar network environment, especially by
establishing interactions with Lys222. The above receptor binding analysis using the Hl model
...........
..........
....
B
100,
100
8.0
d60
060,
I
40
~20
0
40 20
10
5
2
1
0
0.5 0.2 0.1 0.5 0.020.01
20
d
10
5
2
[HAlp9/ni
1
0.5 0.2 0.1 0.50.620.01
[HA] pg/mi
100
80
60
40
20
0
0
2
4
6
8
10
12
14
16
18
20
[HA] pg/nm
Figure 4.3 Dose-dependent direct glycan binding of wildtype and mutant H1 HA. Dose-dependent
direct glycan binding of (A) SCi 8-LS and (B) NYl 8-LS to selected glycans of different topologies,
shows loss of human receptor binding for NYl 8-LS. (C) Comparison of human receptor binding curves
for SC18 and SC18-LS.
system and a dose-dependent, quantitative framework demonstrates how amino acid network
stability is implicated in HA receptor binding strength. SC18-LS and NYl8-LS, which
incorporate nonpolar Leu226 within previously polar networks, display marked loss of binding
to glycans as compared to the wildtype. Avian H5 Viet04-LS and Viet04-RLS, which also
incorporate a nonpolar Leu226 within a polar network, experience decreased a2-3 binding and
do not display any human-adaptive characteristics in the form of increased a2-6 binding.
4.1.3 Discussion
In this chapter, the H2 and H3 HA human adaptive Gln226Leu/Gly228Ser substitutions
are introduced into H1 HA, as had been explored in the context of avian H5 HA in the previous
chapter. Our data provides evidence that H1 HA responds very differently from H2/H3 HA to the
same Gln226Leu/Gly228Ser substitutions. In the context of H2/H3, these substitutions confer
a2-6 specific binding to HA; in H1 they actually decrease a2-6 binding when 225 is Asp (SC 18LS), and obliterate all glycan binding when 225 is Gly (NY18-LS).
Using a human-adapted H1 HA model system, we have furthered our study on the
implications of introducing cross-subtype adaptive mutations. This has led to the understanding
that the establishment of stable interactions between HA residues and the glycan is dependent on
the biochemical and structural environment in which these contacts take place. A nonpolar
residue, when introduced into a network of predominantly polar amino acids, is likely to cause a
disruptive effect not favorable for bond formation. The stability of such amino acid networks is
only one of the contributing factors towards recognition of different glycan topologies.
Structurally, subtle differences in lengths between residues (Glu vs Asp) can have a significant
impact on the extent to which HA can interact with the glycan.
Drawing from the conclusions of this chapter and going back to H5 HA, our preliminary
analysis shows that Asp190 rather than Glul90 is more likely to form stable contacts with
GlcNAc of a glycan in the umbrella topology characteristic of human receptors. Closer
examination of the environment around residue 190 suggests that a Glul90Asp substitution is
insufficient for increasing human receptor binding: Ala189 of H5 does not provide the stabilizing
contact for 190 in the same way as Thr189 of H1, which may explain why Glul90Asp of Viet04
did not show glycan interactions in a microarray study [7].
4.2
Effects of H1N1 HA Glycosylation on Receptor Binding
4.2.1
Introduction
Influenza HA binds host cell glycans to mediate viral attachment and entry into cells. HA
itself is heavily glycosylated [8], and the role of glycosylation in receptor binding has been an
active area of research. Molecular dynamics simulations predict that HA glycans may form
interactions near the binding pocket to influence receptor binding [9]. In earlier mutational
studies, it has been shown that the deletion of two N-glycans near the binding site of a H7N1
fowl plague virus could enhance virus hemadsorbing activity [10], and that a variant of human
seasonal H1N1 without glycosylation at site 131 exhibited more X2-6 binding than the
glycosylated variant [11]. More recently, the role of N-glycans in receptor binding of H5N1 HA
has been examined [12]. Wang et al. showed that truncation of the N-glycan structures on a
H5N1 HA increases a2-3 glycan receptor binding affmities, supporting the notion that N-glycans
of HA play a role in receptor binding. Moreover the loss of a specific N-glycosylation site at
position 154 in a H5N1 virus has been associated with increased virulence in mice [13]. Such
studies provide insights into the overall role of glycans on HA in receptor binding and virulence
91
in the host. While the mutations responsible for changes in receptor binding consist in the loss of
glycans located close to the receptor binding site, specific interactions between HA glycans and
the receptor binding site have not been demonstrated in these glycan studies.
In addition to receptor binding, HA also mediates other functional properties, many of
which have been shown to be affected by glycosylation. Glycans expressed on HA epitopes
shield HA from neutralizing antibodies and contribute to antigenic drift [14]. Near the cleavage
site and in the stem region, HA glycans modulate proteolytic activation and maintain HA in the
metastable form required for fusion activity [8].
Recently, Wei et al. reported the presence of conserved N-glycosylation sites in the stem
and side of the globular head domain of HIN1 HA in a study of the structural basis for crossneutralization between pandemic viruses [14]. Because some of these N-glycosylation sites,
including those near the RBS, are so steadily conserved, there is potential that they may be
important not only for immune evasion but also receptor binding. In this last chapter of the
thesis, the specific role of a conserved N-glycosylation site in the receptor binding property of a
prototypic pandemic HlN1 A/South Carolina/l/18 (SC18) HA model system is explored. The
receptor binding property of wildtype SC 18, a highly specific binder of the human receptor, has
been described earlier in this chapter. We have also previously characterized the receptor binding
properties of NY18 and AV18 HA [6]. NYl8, which differs from SC18 by an Asp-to-Gly
mutation at 225, has a mixed a2-6/a2-3 receptor binding preference and AV18, which differs
from NYl8 by a Gln-to-Asp mutation at 190, has an avian-like c2-3 glycan receptor binding
preference. This set of three closely-related variants of the 1918 pandemic form a wellcharacterized model system suitable for studying the effects of the loss of a specific
glycosylation site near the RBS of HA. This study reveals the effects of subtle HA
conformational changes, resulting from HA deglycosylation, on quantitative receptor binding
affinities.
4.2.2
Results and Discussion
4.2.2.1 Identification of conserved N-glycosylation site proximal to RBS of H1N1 HA
First, sequence and structural analyses were performed on SC 18 HA to identify the
conserved N-glycosylation site of HA that could potentially impinge its receptor binding.
Structural analysis of SC 18 showed that the conserved N-glycosylation site at position 91 (in H3
numbering) could potentially be involved in stabilization of the 220 loop (Figure 4.4). The Nglycan sequon (defined as a sequence of three consecutive amino acids Asn-X-Ser/Thr, where X
is any amino acid except Pro, in a protein that can serve as the attachment site to an N-glycan) at
this site is shown to be glycosylated in the recent crystal structures of HlN1 HAs recombinantly
expressed in insect cells [15]. As described in chapter 1, the 220 loop forms the base of the
receptor binding pocket and makes contacts with the terminal sialic acid and the penultimate
galactose of the non-reducing end. In order to validate the role of the N-glycosylation site at
position 91 in receptor binding, we generated the T93A mutant (Thr of Asn-X-Thr mutated to
Ala) of SC18, NY18 and AV18 HA by site-directed mutagenesis.
4.2.2.2 Circular dichroism analysis of mutant HAs
N-glycosylation is known to play an important role in folding and maintaining the threedimensional structure of HA [8]. In order to ensure that deletion of the glycosylation site at
position 91 did not affect the folding of the protein, we performed a circular dichroism analysis
of the SC18-T93A, NYl8-T93A and AV18-T93A mutant proteins. The spectral signatures
generated for the mutant proteins were compared with that of the wild-type HAs. The circular
l
A/Dk/Alberta/35/76
A/Dk/NY/15024/96
Al SC/i /18
A/TX/36/91
A/N.ewCaiedonia/2 0/99
A/PR/8/34
A/CA/04/09
.. ..
...
...
MEAKLFVLFCTFTVLKADTICVGYHAI
-----QKQGKI KSTKMEAKLFVLFCTFTVLKADTICVGYRAI
MEARLLVLLCAFAA'INADTICIGYRA
-----MKAKLLVLLCAFTATYADTICIGYHA
-----MKAKLLVLLCTFTA IYADTICIGYHA
-----MKANLLVLLSALAAADADTICIGY
-----MKAILVVLLYTFATANADTLCIGY]
------
rIVDTVLE KVTHSVN 4
rIVDTVLE KEMVTHSVNLL
KTSVNLL
TVDTVLEIVTHSVNjL
riVDTVLE KVHSVNLL
THSVNL1
TVLEt
THSVNLL
~TVTLE
~TVLE
A/NewCaiedonia/2 0/99
A/PR/8/34
A/CA/04/09
EDSHNGKLCSLNGIAP1iQLGKCNVAGWLLGNPECDLLLTANSWSYI IETSNS EYPG
EDSHNGKWCSLNGIAPWLGKCNVAWLLGNPECDLLLTANSWSYI IETSNSECYPG
EDSHNGKCKLKGIAPW4LGKCNIAWLLGNPECDLLLTASSWSYIVETRMS
CYP3
EDSHNGKLCRLKGIAPQLGNCSVWILGNPESKES
WSYIAETPECYP
EDSHNGKLCLLKGIAPIWLGNCSVAGWILGNPECELL IS KESWSYIVETPECYP
EDSHNGKLCRLKGIAP1QLGKCNIAGWLLGNIPECDPLLPVRSWSYIVETFPISENGI CYPG
EDKHNGKLCKLRGVAPHLGKCNIGWILGNPECESLSTASSWSYIVETPS SDCYPG
A/Dk/Alberta/35/76
A/Dk/NY/15024/96
A/Sc/i /18
A/TX/36/91
A/NewCaledonia/ 20/99
A/PR/B /34
A/CA/04/09
EFIDYEELREQLSSISSFEKFEIFPKASSWPNHETTKG
SYSGASSFYRNLLWITK
EFIDYEELREQLSSVSSFEKFEIFPKANSWPNHETTK,
G
SYSGASSFYRNLLWITK
**
.**
****
.**. *
.*
*.*
**..
*
***
**
.**
**I
DFIDYEELREQLSSVSSFEKFEIFPKTSSWPNETTKGC SYAGASSFYRNLLWLTK
YFADYEELREQLSSVSSFERFE IFPESSWPMVTKGCSNGKSSFYRNLLWLTK
YFADYEEREQLSSVSSFERFEIFPKESSWPVT- GCSHNGKSSFYRNLLWL IG
DFIDYEELREQLSSVSSFERFEIFPKESSWPNHIITN-GgSHEGKSSFYRNLLWLTE
DFIDYEELREQLSSVSSFERFE IFPKTSSWPNIIDSNKG
PHAGAKSFYKNLIWLVK
A/Dk/Aiberta/35/76
A/Dk/NY/15024/96
A1 'Sc!1 /18
A/TX/36/91
*
A/Dk/Alberta/35/76
A/Dk/NY/i5024/96
A/SC/i/lB
A/TX/36/91
A/NewCaiedonia/20/ 99
A/PR! 8/34
A/CA/04/09
A/Dk/Alberta/35/76
A/Dk/NY/15024/96
A/SC/i /18
A/TX/36/91
A/NewCaledonia/20/99
A/PR/8/34
A/CA/04/09
***********.****.******
AYV.* *
***.**.*.
KEGSYPKLKNSYVNKKEVLWGI
HHPPNSUAYVSVVTSNYNRRFTPE
KGNSYPKISKSYIDKGEVLVLWGIHHPSTSIII TYVFVGSSRYSKKFKPE
I AR1YNY
TLLDQGTITFEATNLIAPWYAFALNKGSD3I ITSDAPVH
IAA
NY
TLLDQGDTITFEATNLIAPWYAFAINKGSDSGI ITSDAPVH
IAARRMNYLTLLEPGDT
TFEATLIAPWYAFALNRGSGSGI ITSDAPVH
IARINYWTLLEPGDrI IFEANGNLIAPWYAFALSRGFGSGIITSD
IAKRR
INYYWTLEGDIIFEANGNLIAPWYAFALSRGFGSGI ITSNAPMD
IAEUMNYYWTLLKPGYIIFEANGNLIAPMYAFALSRGFGSGI ITSMHM
IAIR
MNYYWTLVEPGDKITFEATGNLVVPRYAFAMERNAGSGII ISDTPVR
NCTR***.**
A/Dk/Aiberta/35/76
A/Dk/NY/i5024/96
A/SC/i/lB
A/TX/36/91
A/NewCaiedonia/20/99
A/PR/8/34
A/CA/04/09
**
*****
KGTSYPKIJKSYTNNKGKEVLVLWGVHPPSV3AYVSVGSSKYNRRFAPE
KGTSYPKSKSYNNKGKEVLVLWGVHHPPT
TDAYVSVGSSKYNRRFTPE
MGSSYISKSYVNNKKCEVVLWGVHHPPTGIDAYVSVSSYRRFTPE
KNGLY iKSYVNUEKEVLVHHGVRPSNIAYVSVVSSHYSRRFTPE
TH *
**.******.
**
*
**
***.
*
****.
.
****
NCDTRCQTPHGALLFQNVHPITIIGECPKWIKSTKLRMATGLRNVPSIQSRGLFGAI
NTRCQTPHGALLPFQNIVr IGECPKYI KSTKLRMATGLRNVPSI QSRGLFGAI
EDNKCQTPGAILP FQNIHPVrIGECPKYVRSTKLRMATGLRNIPSIQSRGLFGAI
ECDAKCQTPQGAIL FQNVHPVIIGE CPKYVRSTKLRMVTGLRNI PSIQSRGLFGAI
ECDAKCQTPGAIL FQNVRPVrIGECPKYVRSAKLRMVTGLRNI PSIQSRGLFGAI
DCNTTCQTPKGAI
FQNIHPITIGKCPKWVKSTKLRLATGLRNI PSIQSRGLFGAI
Figure 4.4 Alignment of avian and human HiNi HA sequences show conserved N-giycosyiation
sites. Green:sites distal to the RBS; grey:sites within the RBS secondary structure elements; red. site 9193 proximal to the RBS.
dichroism spectra of all the mutant proteins were generated between 190nm and 280nm. All the
mutants showed similar circular dichroism specral signatures as that of their wild-type
counterparts (Figure 4.5).
4.2.2.3 Direct binding analysis of mutant and wildtype HAs on the glycan array
To understand the differences in the glycan binding specificities and affmities of T93A
mutant HAs from the wild-type proteins, a systematic dose-dependent glycan array analysis was
performed. SC18-T93A showed a characteristic binding to the 6'SLN-LN glycan similar to that
of the wild-type SC 18 HA. However the binding affinity of the mutant was significantly lower
than that of the wild-type protein (Figure 4.6A). From the HA residue standpoint, the high
affinity human receptor binding by SC 18 can explained by the observation that wildtype SC 18
has all the polar interactions favorable for LSTc binding; potential hydrogen bonds can be
formed between : (1) 222Lys and Gal-2, (2) 225Asp and Gal-2, (3) 226Gln and Neu5Ac, (4)
226Gln and Gal-2. In addition, the long residue 224Arg extends downwards to form a potential
hydrogen bond (2.5A distance) with the glycan at glycosylated residue 9lAsn. This holds the
receptor binding domain, especially the 220-loop of HA, in a rigid conformation favorable for
making contacts with LSTc (Figure 4.6D). In the absence of this glycan due to a T93A mutation,
the 220-loop loses its rigidity and critical contacts between LSTc and 222Lys, 225Asp, 226Gln
may not form effectively. This potential loss of contact could explain the lower human receptor
binding affinity by SC18-T93A. The decreased human receptor binding affmity also follows a
time-dependent trend; freshly prepared SC19-T93A at day 1 (Figure 4.6B) exhibits a higher
affinity than SC18-T93A stored until day 4 (Figure 4.6C), lending further support to the
explanation of a subtle, gradual change in conformation along the 220-loop due to
deglycosylation and destabilization, as opposed to a more dramatic change in folding.
...
.........................
........
.
.
4
2
**.
0
'
-..
**
-2
- SC18
" SC18-T93A
- NY18
+
NY18-T93A
- AV18
. AV18-T93A
190
200
210
220
230
240
250
260
270
280
Wavelength (nm)
Figure 4.5 Circular dichroism of wildtype and mutant H1 HAs show similar spectral signatures
indicative of no general misfolding due to deglycosylation. Circular dichroism spectra for SC18, SC18T93A, NY18, NY18-T93A, AV18 and AV18-T93A are shown as indicated in the legend. Deletion of the
glycan at the N-glycosylation site of amino acid 91 does not cause misfolding of HA.
NY18-T93A has less polar interactions favorable for LSTc binding than SC18-T93A; potential
interactions formed between 222Lys and 225Asp, and 225Asp and Gal-2 are absent. With only
polar lysine at 222 and glutamine at 226 to support hydrogen bond formation on opposing ends
of the 220-loop (due to absence of Asp at 225), and lack of the anchoring interaction formed by
224Arg, NY18-T93A loses human receptor binding completely. Wildtype NY18 is able to form
glycan interactions with LSTa because it does not have any polar side chains on 225 that may
repel carbon atoms of the Gal-2 ring. The long residue 224Arg extends downwards to form a
potential hydrogen bond (2.5A distance) with the glycan at glycosylated residue 9lAsn. This
holds the receptor binding domain, especially 226Gln on the 220-loop of HA, in a rigid
conformation favorable for making contacts with LSTa. In the absence of this glycan due to a
T93A mutation, the 220-loop loses its rigidity and critical contacts between LSTa and 226Gln
may not form effectively. Aspartate at 190 is too far away from LSTa to make contacts to
stabilize the glycan. In addition to loss of human receptor binding, NY18-T93A loses avian
receptor binding completely.
Wildtype AV18 is able to bind the avian receptor (Figure 4.8A) because it does not have
any polar side chains on 225 that may repel carbon atoms of the Gal-2 ring. In addition, it has a
long chain residue 190E that forms critical contacts with Gal-2. The long residue 224R extends
downwards to form a potential hydrogen bond (2.5A distance) with the glycan at glycosylated
residue 91N. This holds the receptor binding domain, especially 226Q on the 220-loop of HA, in
a rigid conformation favorable for making contacts with LSTa (Figure 4.8C). In the absence of
this glycan due to a T93A mutation, the 220-loop loses its rigidity and critical contacts between
LSTa and 226Q may not form effectively. Glutamate at 190 is able to compensate for loss of
SLN
a 3'5LN.LN
f sLN-LN
63SU4N-LND
020
10
5
2
1
05
0.2
0.1
00
HA COncentratiln Q(pI)
B
A CSaLN
'so
*3SLNLN
XSULNLN
M3&NULNN
40
20 to 5 2 1 05 02 01 005
20 0 10
A C2
(p1.0ral
5
2
eO
0
863LN
63rSL4LN
*
60
U20
2D
10
5
2
1
HAConcentralon
06
0,2
0-)
006
tigf
Figure 4.6 Glycan receptor binding affinity and structural model of wildtype SC18 HA and glycan
receptor binding affinity of deglycosylated SC18 HA. (A-C) Glycan receptor binding affinities of
wildtype SC18, SC18-T93A (day 1) and SC18-T93A (day 4) respectively. (D) Structure model of
wildtype SC18 shows 224R extending downwards to interact with the N-glycan (blue) at site 91. The
glycan receptor is shown as LSTc in green.
.......................
.....
~IOD
..................................
MSN
a 7"L.LN
$h0
*VSLNLN
w3VIN1.N-LN
140
120
20 to
6
2
1
HACooceiftlon
0.5
O
01
0.1 060
B00
M6SN
s 3LN-LN
8GVX.-L4
s3LN4-LNLN
sao
14D
0
-
Ihj"&b..-.,.-.LA-
20
10
5
2
1
0.5
,
0.2
-
0.1
.---
0.05
HAConewaon 4Wm
Figure 4.7 Glycan receptor binding affinity and structural model of wildtype NY18 HA and glycan
receptor binding affinity of deglycosylated NY18 HA. (A-B) Glycan receptor binding affinities of
wildtype NY18 and NY18-T93A respectively. (C-D) Structure model of wildtype NY18 shows 224R
extending downwards to interact with the N-glycan (blue) at site 91. The glycan receptors are shown as
LSTc in green (left) and as LSTa in pink (right).
-
0
1
- - -
2
CC
- --
.
2
00
10
-
- - -- -- 1--
-
-
-- . --
--
-- -- -
YS1N
SLN0
3SLN-U4
*3SN4LN
3GSLN.4.N-
-
20
-.
5
2
1
05
0.2
0.1
0.05
HA Concenaon (g/ml)
10
u3'SLN-LN
SSLN-LN-LN
*20
20
10
5
2
1
0.5
0.2
0.1
0.05
HA Concentraton (pg/m)
Figure 4.8 Glycan receptor binding affinity and structural model of wildtype AV18 HA and glycan
receptor binding affinity of deglyosylated AV18 HA. (A-B) Glycan receptor binding affinities of
wildtype AV18 and AV18-T93A respectively. (C) Structure model of wildtype AV18 shows 224R
extending downwards to interact with the N-glycan (blue) at site 91. The glycan receptor is shown as
LSTa in pink (right).
100
,
- , j.
4A
::::::::.:.
any glycan contacts caused by deglycosylation. As a result, AV18-T93A loses avian receptor
binding only slightly (Figure 4.8B).
4.2.3 Conclusions
The data presented in this chapter demonstrate that deglycosylation of HA of the 1918
pandemic H1 model system near the receptor binding site (91-93 sequon) directly impacts
interactions with the glycan receptor and its amino acid network. Specifically, the
deglycosylation decreases human receptor binding progressively over time for SC18-T93A,
eradicates receptor binding for NY1 8-T93A, and slightly decreases avian receptor binding for
AV18-T93A. The explanation provided here is a destabilization of 220-loop of the receptor
binding site in absence of anchoring glycan; deglycosylation of this particular site does not alter
HA secondary structure as shown by circular dichroism. This evidence points to subtle changes
in HA structure that contribute to moderate-to-dramatic changes in receptor binding, depending
on the amino acid network environment. Earlier in the chapter, it was demonstrated that the
establishment of stable interactions between HA residues and the glycan receptor is dependent
on the molecular environment, again implicating amino acid networks. Notably, glycans as well
as amino acids of HA can participate in interaction networks to affect receptor binding
properties.
4.3
Materials and Methods
Cloning, baculovirus synthesis, expression and purification of HA
The soluble form of HA was expressed using the baculovirus expression vector system (BEVS).
pAcGp67-SC 18LS-HA and pAcGp67-NY1 8LS-HA plasmids were generated from pAcGP67-
101
SC18-HA [6] by [Gln226Leu; Gly228Ser] and [Asp225Gly; Gln226Leu; Gly228Ser ] mutations
respectively. Mutagenesis was carried out using the QuikChange multisite-directed mutagenesis
kit (Stratagene). The primers used for mutagenesis were synthesized by IDT DNA Technologies.
Baculoviruses were created from their respective HA constructs using the Baculogold system
(BD Biosciences) according to the manufacturer's instructions. Purification and concentration of
HA were performed in a similar manner as that described in Chapter 3.2.3.
Dose-dependent direct glycan binding of HA - Concentrated mutant and wildtype H1 HA was
applied to a dose-dependent glycan array similar to that described in Chapter 3.2.3.
References
1.
2.
3.
4.
5.
6.
7.
Stevens, J., Corper, A.L., Basler, C.F., Taubenberger, J.K., Palese, P., and Wilson, I.A.
(2004). Structure of the uncleaved human HI hemagglutinin from the extinct 1918
influenza virus. Science 303, 1866-1870.
Chandrasekaran, A., Srinivasan, A., Raman, R., Viswanathan, K., Raguram, S., Tumpey,
T.M., Sasisekharan, V., and Sasisekharan, R. (2008). Glycan topology determines human
adaptation of avian H5N1 virus hemagglutinin. Nat Biotechnol 26, 107-113.
Gamblin, S.J., Haire, L.F., Russell, R.J., Stevens, D.J., Xiao, B., Ha, Y., Vasisht, N.,
Steinhauer, D.A., Daniels, R.S., Elliot, A., Wiley, D.C., and Skehel, J.J. (2004). The
structure and receptor binding properties of the 1918 influenza hemagglutinin. Science
303, 1838-1842.
Glaser, L., Stevens, J., Zamarin, D., Wilson, L.A., Garcia-Sastre, A., Tumpey, T.M.,
Basler, C.F., Taubenberger, J.K., and Palese, P. (2005). A single amino acid substitution
in 1918 influenza virus hemagglutinin changes receptor binding specificity. J Virol 79,
11533-11536.
Tumpey, T.M., Maines, T.R., Van Hoeven, N., Glaser, L., Solorzano, A., Pappas, C.,
Cox, N.J., Swayne, D.E., Palese, P., Katz, J.M., and Garcia-Sastre, A. (2007). A twoamino acid change in the hemagglutinin of the 1918 influenza virus abolishes
transmission. Science 315, 655-659.
Srinivasan, A., Viswanathan, K., Raman, R., Chandrasekaran, A., Raguram, S., Tumpey,
T.M., Sasisekharan, V., and Sasisekharan, R. (2008). Quantitative biochemical rationale
for differences in transmissibility of 1918 pandemic influenza A viruses. Proc Natl Acad
Sci U S A 105, 2800-2805.
Stevens, J., Blixt, 0., Tumpey, T.M., Taubenberger, J.K., Paulson, J.C., and Wilson, I.A.
(2006). Structure and receptor specificity of the hemagglutinin from an H5N1 influenza
virus. Science 312, 404-410.
102
8.
9.
10.
11.
12.
13.
14.
15.
Ohuchi, R., Ohuchi, M., Garten, W., and Klenk, H.D. (1997). Oligosaccharides in the
stem region maintain the influenza virus hemagglutinin in the metastable form required
for fusion activity. J Virol 71, 3719-3725.
Varki, A. (2009). Essentials of Glycobiology, 2 Edition (Cold Spring Harbor Laboratory
Press).
Ohuchi, M., Ohuchi, R., Feldmann, A., and Klenk, H.D. (1997). Regulation of receptor
binding affinity of influenza virus hemagglutinin by its carbohydrate moiety. J Virol 71,
8377-8384.
Gambaryan, A.S., Marinina, V.P., Tuzikov, A.B., Bovin, N.V., Rudneva, I.A., Sinitsyn,
B.V., Shilov, A.A., and Matrosovich, M.N. (1998). Effects of host-dependent
glycosylation of hemagglutinin on receptor-binding properties on H1N1 human influenza
A virus grown in MDCK cells and in embryonated eggs. Virology 247, 170-177.
Wang, C.C., Chen, J.R., Tseng, Y.C., Hsu, C.H., Hung, Y.F., Chen, S.W., Chen, C.M.,
Khoo, K.H., Cheng, T.J., Cheng, Y.S., Jan, J.T., Wu, C.Y., Ma, C., and Wong, C.H.
(2009). Glycans on influenza hemagglutinin affect receptor binding and immune
response. Proc Natl Acad Sci U S A 106, 18137-18142.
Chen, H., Bright, R.A., Subbarao, K., Smith, C., Cox, N.J., Katz, J.M., and Matsuoka, Y.
(2007). Polygenic virulence factors involved in pathogenesis of 1997 Hong Kong H5N1
influenza viruses in mice. Virus Res 128, 159-163.
Wei, C.J., Boyington, J.C., Dai, K., Houser, K.V., Pearce, M.B., Kong, W.P., Yang, Z.Y.,
Tumpey, T.M., and Nabel, G.J. Cross-neutralization of 1918 and 2009 influenza viruses:
role of glycans in viral evolution and vaccine design. Sci Transl Med 2, 24ra2 1.
Xu, R., Ekiert, D.C., Krause, J.C., Hai, R., Crowe, J.E., Jr., and Wilson, I.A. Structural
basis of preexisting immunity to the 2009 H1N1 pandemic influenza virus. Science 328,
357-360.
103
Chapter 5
Summary and Significance of Thesis Research
An understanding of the factors involved in the human adaptation of influenza A viruses
is critical for various aspects of influenza preparedness, including the development of
appropriate surveillance measures, preventive strategies and effective treatments. A key step in
influenza human adaptation is the acquisition of mutations in HA which changes its binding
specificity towards glycan receptors in the human upper respiratory epithelia. In this thesis,
determinants that mediate changes in HA-glycan receptor binding specificity are investigated,
with focus on the molecular environments within and surrounding the RBS. The glycan receptor
binding properties of HA from different influenza subtypes (H1N1, H2N2, H3N2 and H5Nl) are
studied using a combination of approaches including dose-dependent glycan binding, human
tissue staining and structural modeling. Using these complementary analyses, it is shown in this
thesis that the molecular interactions between amino acids in and proximal to the RBS (referred
to as amino acid interaction networks), including those between the RBS and glycosylation at
sites proximal to the RBS, and interactions between the RBS and the glycan receptor together
govern the high affinity binding of HA to human receptors.
The thesis research is divided into three sections. In Chapter 2, the evolution of glycan
receptor binding specificity of recent human-adapted H3 strains such as A/Fujian/411/02 and
A/Panama/2007/99 is investigated, with implications on vaccine production in chicken eggs. The
residues at 222 and 225 of the HA RBS are found to play a role in determining H3N2 receptor
binding strengths. Notably, these adjacent residues share the same biochemical characteristics as
105
they evolve, suggesting potential interactions between them that are important for glycan
receptor binding. In Chapter 3, the determinants of glycan receptor binding affinity of potentially
pandemic avian viruses are studied in the context of the recently circulating H2
A/Chicken/Pennsylvania/2004 and the highly pathogenic H5 A/Vietnam/1203/2004. Here it is
shown that mutations at 226 and 228 which cause human adaptation of H2 do not increase
human receptor binding affinity in H5, and the importance of amino acid interaction networks is
again highlighted. Additionally, mutations at 137 and 193 of the recently circulating avian H2
strain change its receptor binding to resemble that of the pandemic A/Albany/6/58, consistent
with the fact that these residues have extended range contacts with the glycan receptor beyond
the sialic acid linkage. In Chapter 4, determinants that govern the high affinity human receptor
binding of pandemic influenza HAs is investigated using the prototypic 1918 HINI HA as a
model system. The roles of amino acid interaction networks and the molecular interactions
between the RBS and glycosylation at sites proximal to the RBS in contributing to the high
affinity human receptor binding of 1918 H1Nl HA are investigated. This study reveals the
effects of subtle HA conformational changes on quantitative receptor binding affinities.
The advent of the 2009 H1N
'swine flu' pandemic and the threat from the highly
pathogenic avian H5N1 virus demonstrate the need for improved surveillance measures and
understanding of the factors behind influenza host adaptation. Previous approaches for studying
receptor binding often entail the substitution of amino acids in the RBS in isolation of their
surrounding molecular environment. It is shown in this thesis that the amino acid interaction
networks give rise to subtle changes in the molecular environment of HA which are key for
receptor binding affinities. The approaches presented in this thesis to systematically investigate
molecular interactions between HA and glycan receptors that impinge on quantitative HA-glycan
106
receptor binding affinity offer a new angle towards studying determinants of human receptor
binding specificity and affinity of influenza A virus HAs.
107