UNIVERSITY OF CALIFORNIA SAN DIEGO

Responses of Germline Targeting Immunogens to Influenza

A Thesis submitted in partial satisfaction of the requirements for the degree Master of Science

in

Biology

by

Justin Romualdo Abadejos

Committee in charge:

Professor Stephen Hedrick, Chair Professor Li-Fan Lu, Co-Chair Professor Elina Zuniga

2018

Copyright

Justin Romualdo Abadejos, 2018

All rights reserved.

The Thesis of Justin Romualdo Abadejos is approved and it is acceptable in quality and form for publication on microfilm and electronically:

Co-Chair

Chair

University of California San Diego

2018

iii DEDICATION

I would like to thank everyone at the old Kirak lab: Oktay Kirak, Manching Ku,

Nathaniel Phillips, Niklas Münchmeier, Shih-En Chang, Julio Hernandez, Jesus

Zaragoza, Amy Nham, and Brett Freeman for taking me in and introducing me to the world of science.

I would like to also thank Dr. Stephen Hedrick, Dr. Li-Fan Lu, and Dr. Zuniga for allowing me to continue my research at TSRI and for acting as my committee.

I would also like to thank everyone at the Nemazee lab: David Nemazee, Amanda Gavin,

Patrick Skog, Tanya Blane, Deli Huang, Sabrina Volpi, Miyo Ota, Terri Thinnes, and

Wenjuan Li for taking me in to a wonderful family and helping me along the way.

A special thank you to David Nemazee, Takayuki Ota, Shanshan Lang for being wonderful mentors and for guiding me throughout my time in science.

I want to take the time to thank Nicole Aguirre, for continuously pushing me to be my best and supporting me in everything I do.

A huge thank you and odes of appreciation to my parents, Efren and Tessie Abadejos for putting up with me and being my foundation, words can’t explain how grateful I am for you two. I’d like to dedicate this thesis to my grandfather and my late grandmother, Felix and Isadora Romualdo.

iv EPIGRAPH

“Never doubt that a small group of thoughtful, committed citizens can change the world; indeed, it's the only thing that ever has.” -Margaret Mead

“If you want to go fast, go alone. If you want to go far, go together.” -African Proverb

v TABLE OF CONTENTS

Signature Page ..……………………………………………………………………. iii

Dedication ...……………………………………………………………………….. iv

Epigraph …………………………………………………………………………… v

Table of Contents ...………………………………………………………………... vi

List of Abbreviations ………………………………………………………………. vii

List of Figures ……………………………………………………………………… ix

List of Tables ………………………………………………………………………. xi

Acknowledgements ………………………………………………………………... xii

Vita…………………………………………………………………………………. xiii

Abstract of the Thesis ...…………………………..……………………………….. xiv

Chapter 1: Introduction…………………………………………………………….. 1 1.1: Figures………………………………………………………………… 9

Chapter 2: Testing the ability of anti-idiotype K1-18 and engineered HA 587H1-X to induce a broadly neutralizing HA stem response in a 9114 gH mouse model…….. 16 2.1: Generation of a 9114 gH mouse model……………………………….. 16 2.2: Testing of anti-idiotypic antibody K1-18 as a germline targeting immunogen…………………………………………………………………. 17 2.3: Testing of engineered HA 587H1-X as a germline targeting immunogen…………………………………………………………………. 20 2.4: Discussion……………………………………………………………… 24 2.5: Methods………………………………………………………………... 30 2.6: Figures and Tables…………………………………………………...… 42

References…………………………………………………………………………... 57

vi LIST OF ABBREVIATIONS

BLI biolayer inferometry bnAb broadly neutralizing antibody

BSA bovine serum albumin

CDR complementarity determining region

CDRH complementarity determining region heavy chain

ELISA Enzyme linked immunosorbent assay

FAB fragment antigen binding

FACS fluorescence-activated cell sorting

Fc fragment crystallizable

FR framework region gH germline heavy chain

GL germline

HA hemagglutinin

IgG immunoglobulin G

Kd dissociation constant

M molar

MHC major histocompatibility complex

NA neuraminidase

Ni-NTA nickel-nitrilotriacetic acid

PBMC peripheral blood mononuclear cells

PBS phosphate buffered saline

vii PDB Protein Data Bank

PEG polyethylene glycol

SSRL Stanford Synchrotron Radiation Lightsource

UCA unmutated common ancestor

viii LIST OF FIGURES

Figure 1.1. The influenza life cycle and potential immune responses……………… 9

Figure 1.2. Structures and subtypes of hemagglutinin……………………………… 10

Figure 1.3. Isolation, identification, and theoretical development of broadly neutralizing antibodies………………………………………………………………. 11

Figure 1.4. Crystal structures revealing HA stem binding contacts of VH1-69 bnAb CR6261……………………………………………………………………….. 12

Figure 1.5. VH1-69 bnAbs utilizing key CDRH2 residues to bind to HA stem…..... 13

Figure 1.6. Binding ability of various influenza bnAbs to the different HA subtypes 14

Figure 1.7. DNA Vaxibody dimer structure………………………………………... 15

Figure 2.1. Design of 9114 gH knock-in mice……………………………………... 42

Figure 2.2. Characterization of the B6 9114 gH Mouse Model……………………. 43

Figure 2.3. Structure and paratope of K1-18……………………….………….…… 44

Figure 2.4. K1-18 activates B6 9114 gH B cells…………………………………… 45

Figure 2.5. K1-18 Immunizations show 587H1-X IFY enhancing mutation-specific response via Octet…………………………………………………………………… 46

Figure 2.6. K1-18 Immunizations show 9114-specific, 587H1-X IFY enhancing mutation-specific response via competitive ELISA……………………………..…. 47

Figure 2.7. Vaxibody DNA immunizations with K1-18 scFv as the antigenic unit show responses to K1-18 and 587H1 via endpoint titre ELISA………………………….. 48

Figure 2.8. Structures outlining mutations in Cali09/H1 that improve binding to germline VH1-69 bnAb CR6261……………………………………………………………… 50

Figure 2.9. In vitro binding and activation of 587X-H1 to B6 9114 gH B cells…… 51

Figure 2.10. 587H1-X Immunizations show IFY motif enhancing mutation-specific response via octet and ELISA assays………………………………………………. 52

Figure 2.11. Immunizations of B6 9114 gH mice with 587H1-X generate mutations in 9114 heavy chain…………………………………………………………………. 54

ix

Figure 2.12. Future directions in testing germline targeting immunogens for development of influenza broadly neutralizing antibodies………………………………………… 56

x LIST OF TABLES

Table 2.1. Mutations in Cali09/H1 that improve binding to germline VH1-69 bnAb CR6261……………………………………………………………………………… 49

xi ACKNOWLEDGEMENTS

I’d like to thank Shanshan Lang (TSRI) for developing our germline targeting immunogens, providing and creating protein for analysis, for structures of K1-18, 587H1-

X, CR6261, CR9114, and 27F3; Takayuki Ota (TSRI) for generating the 9114 gH mouse model, for help characterizing the mouse model, for proving mutation analysis data and in general for all his help and mentorship on all aspects of this project; Stephen M.

Hedrick, Li-Fan Lu, and Elina Zuniga for allowing me to continue my research at TSRI and for gratefully serving on my committee; Patrick Skög for his help with optimizing immunizations; Tanya Blane for her help with mouse techniques and immunizations;

Gunnveig Grødeland for mentorship and providing the vaxibody construct; TSRI’s FACS

++ core for help with flow cytometry and providing resources for FACS analysis and Ca flux; TSRI’s department of animal resources for outstanding services. Lastly, I’d like to thank Dr. David Nemazee for his oversight of the entire project and his mentorship and guidance throughout these past two years.

xii VITA

2016 Bachelor of Science, Human Biology, University of California San

Diego

2016-2017 Teaching Assistant, Department of Biological Sciences

University of California San Diego

2018 Master of Science, Biology, University of California San Diego

PUBLICATIONS

Ku, M., S. E. Chang, J. Hernandez, J. R. Abadejos, M. Sabouri-Ghomi, N. J. Muenchmeier, A. Schwarz, A. M. Valencia and O. Kirak (2016). "Nuclear transfer nTreg model reveals fate-determining TCR-beta and novel peripheral nTreg precursors." Proc Natl Acad Sci U S A 113(16): E2316-2325.

FIELDS OF STUDY

Major Field: Human Biology

Studies in Immunology

Professors Oktay Kirak and David Nemazee

xiii

ABSTRACT OF THE THESIS

Response of Germline Targeting Immunogens to Influenza

by

Justin Romualdo Abadejos

Master of Science in Biology

University of California San Diego, 2018

Professor Stephen Hedrick, Chair

Professor Li-Fan Lu, Co-Chair

Influenza virus is a constant threat to international public health. The current, seasonal vaccine generates a response to epitopes of influenza virus that have a high rate

xiv of mutation leading to antigenic drift and antigenic shift leaving current memory responses useless and potentially resulting in a pandemic. The need for a universal influenza vaccine is evident. High affinity, broadly neutralizing antibodies (bnAbs)

against a conserved epitope of the influenza virion, such as the stem of hemagglutinin, show potential to protect against a wide breadth, if not all influenza viruses. Identifying germline precursors of these bnAbs and generating immunogens that drive the maturation of a heterosubtypic response could provide a universal answer to influenza. In this thesis, I show the development and characterization of a mouse model consisting of B cells containing the germline heavy chain of the influenza bnAb CR9114.

Furthermore, I outline the responses of two germline targeting immunogens to influenza: an anti-idiotypic antibody known as K1-18 and an engineered variant of an H1- hemagglutinin protein known as 587H1-X. Both K1-18 and 587H1-X are shown via calcium flux and activation markers to simulate B cells of our 9114 gH mouse model.

Immunization studies with both immunogens show their ability to generate an IgG response that cross binds to wild type hemagglutinin. 587H1-X in particular shows mutation capabilities that drive these gl9114 antibodies towards their mature stage. This study sheds light on a germline targeting approach to inducing a universal response to influenza.

xv CHAPTER 1: INTRODUCTION

The influenza pandemic of 1918 was known as one of the most dramatic events of medical history and claimed over 50 million lives worldwide (1). Infection with the virus in 1918 led to symptoms similar to other respiratory infections, but also included: sudden onset of three-day fever, muscle pain, and prostration that led ultimately to millions of deaths. Since then, influenza has remained a public health concern.

The Influenza Virion

Influenza is a negative sense, single stranded RNA virus of the family orthomyxoviridae. The influenza genome is separated into eight RNA segments that encode for ten major viral proteins (2). The major components of influenza are: hemagglutinin (HA), neuraminidase (NA), the Matrix 2 proton channel, the Matrix 1 protein, nucleoprotein, and the negative strand RNA genome (3). The influenza life cycle begins with infection by HA binding to sialic acid receptors leading to virion endocytosis.

The viral and endosomal membranes fuse utilizing proton transfer via the M2 proton channel which initiates a low pH conformational change in HA leading to fusion of the two membranes carried out by the stem region of HA allowing for the viral genes to be released into the cytoplasm. Nuclear localization signals allow the viral genes to move into the nucleus followed by protein synthesis, packaging, and budding. The release of new virions is triggered by neuraminidase which enzymatically cleaves the sialic acid receptors from the cell membrane, allowing for further cell infection (Fig. 1.1) (4).

1 The two surface molecules, hemagglutinin and neuraminidase, are trimeric and tetrameric molecules, respectively and are the target of most influenza vaccines and anti- viral drugs. Hemagglutinin comprises over 80% of the surface membrane molecules of the influenza virion and is composed of two structures: the highly variable, globular head group and the more conserved stalk or stem region (Fig. 1.2A)(3). The globular head group poses a major obstacle to influenza vaccine development due to two reasons: its high immunogenicity, poising pre-existing immune responses away from the highly conserved stem region and more prominently, its high rate of mutation at 5.52 variants per HA residue within avian viruses and 4.42 variants per residue for non-avian viruses

(5-7). This high rate of mutation can potentially be attributed to the lack of a functional role of the head domain other than the sialic acid binding site, making the other epitopes highly volatile. On the other hand, the stem domain has a direct function in viral membrane fusion, which requires a major conformational change, a quality that may explain its conservation. The high mutation rate of the head region leads to antigenic drift and ultimately is the primary reason the current vaccination is required on a seasonal basis. In terms of public health, the high rate of mutation in HA is cause for concern because accumulated or specific mutations along with genetic reassortment of various virus strains could lead to antigenic shift and can result in pandemics such as the 2009 swine flu pandemic (8). Previous seasonal influenza viruses have also evolved over time by acquiring glycosylations that may add another obstacle to vaccine development by shielding antigenically relevant regions (9). A vaccine that generates antibodies specific for a conserved epitope on HA such as the HA stem would, in theory, be able to

2 neutralize a breadth of influenza virus strains and eliminate the need for a seasonal flu vaccine.

The Current Seasonal Influenza Vaccine

The current influenza vaccine and the timely process of how it is created year after year explains and highlights the need for a universal vaccine. The current seasonal vaccine is made annually beginning in January/February when a panel of scientists selects three or four influenza strains that they predict will be circulating come flu season.

Once they are selected, the timely process of growing the virus in chicken eggs begins and ends sometime in June when they are inactivated, tested, and mixed into one seasonal, trivalent/quadrivalent vaccine (10). A major complication in the current vaccine system is the time of creation versus dissemination of the vaccine. There is a substantial risk that the hypothesized strains could be incorrect due to antigenic drift and genetic reassortment through the months of development, ultimately leading to an ineffective vaccine come peak flu season (11).

The path towards a universal influenza vaccine

A universal influenza vaccine would theoretically bring about a response that generates a class of broadly neutralizing antibodies able to protect against a wide variety, if not all, strains of influenza. Influenza strains are categorized based on their HA amino acid sequences into two main groups: group A and group B. Group A is further separated into subgroup 1 and 2 (Fig. 1.2B)(3). Prioritization for generating universal influenza vaccines for major circulating strains such as H1, H2, and H3 rather than generating

3 immunogens or vaccine strategies that can protect against all strains would be a massive step up from our current, hypothesized vaccine strategy (12). The concept of creating a vaccine that develops mature broadly neutralizing antibodies is outlined in Figure 1.3, beginning with the isolation and identification of these antibodies in human individuals through blood samples. These mature bnAbs are then reverse engineered: their germline versions as well as their intermediate versions are inferred ultimately leading to the identification of a germline or unmutated common ancestor (UCA). Then, various immunogens are designed specifically to reverse this process from germline towards intermediate ancestors and eventually the mature bnAb (13).

Broadly neutralizing antibodies to influenza

The identification of broadly neutralizing antibodies from individual antibody repertoires has been done by the generation of antibody phage display libraries and single cell cloning (14). Various antibody phage display libraries identified a family of heterosubtypic neutralizing antibodies that utilize specifically the VH1-69 variable heavy chain gene and ultimately found the most broadly neutralizing of these antibodies

CR6261, which was able to neutralize multiple group 1 viruses (15-19). The Crucell group went on to use antibody phage display to identify three more influenza bnAbs, one of them also utilizing the VH1-69 gene and being one of the most broadly binding flu bnAbs identified, known as CR9114 (20). CR9114 was found to neutralize virtually all influenza A subtypes, but also influenza B through targeting of the stem region. Because of its breadth of binding and potential neutralization, learning how to re-elicit CR9114 by germline targeting vaccination is the topic of this thesis.

4 The VH1-69 derived anti-stem broadly neutralizing antibodies utilize a characteristic form of binding. Crystal structures of CR6261 bound to HA of SC1918/H1 and Viet04/H5 reveal only heavy chain contacts and little to no light chain contacts (Fig.

1.4) (19). Paired heavy and light chain analysis using various VH1-69 bnAbs showed maximal binding to H1Cali09 HA when the mature heavy chain was paired with the germline light chain, further showing the importance of the heavy chain in binding (21).

Mutational analysis shows the importance of two elements required for acquiring breadth of binding, phenylalanine at residue 54 and a conserved tyrosine at residue 98 in CDRH3

(21, 22). Crystal structures of four identified VH1-69 anti-stem bnAbs also reveal key contacts with HA stem in CDRH2 with two highly conserved residues, isoleucine-53 and phenylalanine-54 (Fig. 1.5) (23). Although other broadly neutralizing flu antibodies not utilizing the VH1-69 use both heavy and light chain contacts, VH1-69 family antibodies and specifically CR9114 have been shown to have the widest breadth of binding to almost all group A HAs and group B (Fig. 1.6) (20, 24, 25).

The mechanism of neutralization of VH1-69 broadly neutralizing antibodies is carried out by blocking the low pH conformational change of HA, ultimately leading to the prevention of viral and endosomal membrane fusion carried out by the stem domain

(26, 27). CR9114 was shown to neutralize through two different mechanisms, the aforementioned block of the stem region’s low pH conformational change and via a FcγR dependent mechanism (28).

5 Utilizing germline targeting immunogens to generate an anti-stem broadly neutralizing antibody response

To further study the mechanism of neutralization and ultimately to test germline targeting immunogens that can potentially generate a broadly neutralizing response, Dr.

Taka Ota and Dr. David Nemazee developed a mouse model that expressed the germline reverted 9114 heavy chain knocked into the IgH locus (9114 gH). Several immunogens have been designed to generate anti-stem responses by removing the immunodominant head domain (29-32). Our alternative approach involves increasing the affinity of HA or other immunogens to VH1-69 precursors to specifically prime and develop a broadly neutralizing, anti-stem response through targeting of the germline lineage in an otherwise infrequent, stem response.

Shanshan Lang of the Wilson Lab at TSRI solved the crystal structure of an anti- idiotypic antibody known as K1-18 that binds efficiently to CR6261 and other VH1-69 anti-HA stem bnAbs (33). Crystal structures of the K1-18 anti-idiotype with CR6261 show that K1-18 binds a critical motif on VH1-69 antibodies, known as the IFY motif

(Ile53, Phe54, and Tyr98) that are known to interact with the HA stem. A critical and unusual property of K1-18 is that it binds independent, mutated broadly neutralizing anti-

HAs that express VH1-69 as well as monoclonal antibodies with germline VH1-69. Anti- idiotypic vaccines have been shown to trigger both cellular and humoral immune responses in vivo (34) and taken with the VH1-69 binding ability, K1-18 has the potential to elicit a VH1-69 specific, broadly neutralizing response. Shanshan Lang utilized the binding specificity of K1-18 to VH1-69 bnAbs to identify three key residues on HA that when mutated strengthen binding to the aforementioned IFY motif: L38Y, T49Y, and

6 S291H. The engineered HA with the three key mutations is known as H1Cali09-X or

587H1-X and binds to the germline antibody of CR6261 with a Kd of ~31 nM, over

1000-fold higher than the wild type H1Cali09 HA. When tested via calcium flux in vitro activation assay, K1-18 and 587H1-X were shown by me to activate B cells isolated from our 9114 gH mouse model and to upregulate the activation marker CD86. 9114 gH mice immunized with K1-18 showed a 9114 IgG response specific for the IFY motif binding site. 9114 gH mice immunized with 587H1-X again showed a 9114 IgG response specific to the three mutations introduced to increase binding to the IFY motif. Additionally, heavy chain sequencing analysis showed that 587H1-X generates some of the mutations that match to mature CR9114.

Incorporating germline targeting immunogens into the Vaxibody DNA platform

The Nemazee lab collaborated with Gunnveig Grødeland, who is a primary researcher and developer of a vaccine platform known as Vaxibody. Vaxibody constructs are dimers of fusion proteins each carrying the CH3 domain of IgG3 linked to an APC- targeting unit and the “antigen” (Fig. 1.7) (40, 41). The construct DNA is injected intramuscularly followed by introduction of an electric current allowing for the DNA construct to be transfected into local cells where they can express and secrete the vaxibody proteins. The targeting unit then directs these proteins towards antigen presenting cells utilizing units such as an anti-MHCII single chain variable fragment

(scFv) that will direct a more humoral immunity driven response or the MIP-1α chemokine unit that targets antigen to CD8+ dendritic cells to direct a cellular immunity

7 driven response. The antigenic unit is able to be switched out to contain any immunogen of choice, in our case, K1-18. DNA Vaxibody immunization with K1-18 scFv as the antigenic unit showed a significant response at day 21 post-immunization binding to K1-

18 and also show a wild-type 587H1 response. The Vaxibody platform is a promising and time-efficient alternative to immunization using germline targeting antigenic units.

Further immunization and neutralization studies are necessary to optimize these germline targeting immunogens and determine their ability to induce a universal, anti- stem response.

8 1.1: Figures

Figure 1.1. The influenza life cycle and potential immune responses. Influenza infection begins by the attachment of hemagglutinin to sialic acid receptors located on the cell surface (1) followed by endocytosis of the virion (2). Viral and endosomal membrane fusion occurs utilizing the M2 proton channel (3, 4) resulting in low pH conformational change in HA resulting in fusion. The viral genes are released into the cytoplasm and localize to the nucleus followed by viral protein synthesis, packaging, and budding (5). Release of the virions from the host cell is triggered by the cleavage of neuraminidase of the sialic acid receptors from the cell membrane (6). Anti-HA antibodies can neutralize by blocking virus attachment. Anti-M2 antibodies can prevent infection by stopping the release of viral particles into the cytoplasm. Cellular immunity can prevent infection through CD8+ T cells specific for nucleoproteins (NP) leading to cytotoxic apoptosis. Anti-NA antibodies prevent infection by preventing release of new virions. *Figure is adapted from the literature (4).

9 B

Figure 1.2. Structures and subtypes of hemagglutinin. (A) Ribbon diagram of a monomer of the trimeric influenza hemagglutinin surface protein. HA is composed of two domains: the highly variable, immunodominant globular head domain and the highly conserved stalk or stem domain. (B) Phylogenetic tree of all HA subtypes. HA is separated into two major strains, A and B. Strain A is broken up further into two groups, 1 and 2. *Figures are adapted from the literature (3).

10

Figure 1.3. Isolation, identification, and theoretical development of broadly neutralizing antibodies. (1) Blood samples and antibodies are isolated from individuals from populations that potentially have bnAbs, these antibodies are isolated and their lineage is inferred to generate a unmutated common ancestor. (2) Immunogens are engineered that are able to bind to the germline antibody and/or intermediate clones to potentially drive the development to the mature broadly neutralizing antibody. (3) Immunization studies are conducted to show in vivo the efficacy of these immunogens of generating and maturing a response that is broadly neutralizing. *Figure is adapted from the literature (13).

11

Figure 1.4. Crystal structures revealing HA stem binding contacts of VH1-69 bnAb CR6261. (A) CR6261 in complex with SC1918/H1 trimeric hemagglutinin. One monomer is shown colored for clarity: HA1 is shown in purple, HA2 in cyan, CR6261 heavy chain in yellow, and CR6261 light chain in orange. (B) CR6261 in complex with Viet04/H5 trimeric hemagglutinin, colored as in (A). Binding of CR6261 to these two HAs is specific to the stem region located at the bottom and shows a majority of binding contacts coming from the heavy chain. *Figures are adapted from the literature (19).

12

Figure 1.5. VH1-69 bnAbs utilizing key CDRH2 residues to bind to HA stem. Ile53 and F54 of CDRH2 in various VH1-69 antibodies to the HA stem. (A) 27F3 and CR9114 are found to bind to both group 1 and 2 of influenza strain A. They show similar conformations that allow hydrogen bonding with the HA stem. (B) CR6261 and F10 antibodies are found to bind broadly to group 1 HAs and have different conformations compared to 27F3 and CR9114 making hydrogen bonding less favorable. *Figures are adapted from the literature (23).

13

Figure 1.6. Binding ability of various influenza bnAbs to the different HA subtypes. Phylogenetic tree of influenza HA subtypes reveal a majority of identified influenza broadly neutralizing antibodies do not show universal binding ability, but CR9114 shows the most breadth out of all tested. *Figure is adapted from the literature (24).

14 Targeting unit

- scFv (VH+VL) (αMHCII) - chemokine MIP-1α

hinge Dimerization unit

Cγ3

Antigenic unit K1-18 ScFv

Figure 1.7. DNA Vaxibody dimer structure. The general dimerized Vaxibody structure is composed of a targeting unit (aMHC-II or MIP1α), a CH3 dimerization unit from IgG3, and an antigenic unit of choice (K1-18 scFv). *Figure provided by Dr. Gunnveig Grødeland, University of Oslo, Norway

15

CHAPTER 2: TESTING THE ABILITY OF ANTI-IDIOTYPE K1-18 AND

ENGINEERED HA 587H1-X TO INDUCE A BROADLY NEUTRALIZING HA

STEM RESPONSE IN A 9114 gH MOUSE MODEL

2.1: Generation of a 9114 gH mouse model

In order to understand the development of human VH1-69 broadly neutralizing antibodies and specifically the generation of a mature CR9114 bnAb using germline targeting immunogens, Dr. Taka Ota and Dr. David Nemazee designed and characterized a mouse model containing the germline heavy chain of CR9114. Due to the fact that mice lack a gene similar to VH1-69, the creation of this model was necessary to study the development of these influenza bnAbs. The design replaced the mouse region DQ52-Jh with a germline reverted VDJ exon from CR9114 along with its associated promoter and leader elements (Fig. 2.1). The location of the knock-in gene allows for normal regulation, class switching, and somatic hypermutation making these mice an ideal baseline to study and test the development of vaccinations with germline targeting immunogens.

The 9114 gH mouse model was characterized by Dr. David Nemazee and Patrick

Skög and shown in Figure 2.2. When compared to wild-type littermates, 9114 gH knock- in mice exhibited normal B cell numbers in various lymphoid tissues (Fig. 2.2A). To determine the frequency of B cells expressing the 9114 germline reverted heavy chain,

9114 gH IgHb/b and wild-type IgHb/b mice were bred to wild-type IgHa/a mice. Wild-type

16 IgHa/b mice show an equal ratio expressing IgMa and IgMb, whereas in 9114 gH IgHb/b mice crossed with wild-type IgHa/a ~90% of B cells show expression of IgHb and thus, showing suppression of endogenous mouse heavy chains (Fig. 2.2B). FACS staining of B cells from 9114 gH mice with PE-labeled HA from 575H3 and 587H1 revealed the propensity of a few of these precursor CR9114 B cells to bind to a group 1 HA and cross- bind to a group 2 HA (Fig. 2.2C). Characterization of light chain usage in 9114 gH mice revealed an elevated expression of lambda light chain in B cells as compared to a littermate control, ~4% vs. ~25%, showing an increased preference of germline 9114 heavy chains to pair with characteristic lambda light chains (Fig. 2.2D). Although, HA- binding cells mainly used kappa light chains. The 9114 gH mouse model, with an artificially high precursor frequency by design, is an ideal baseline to study a proof of concept that germline targeting immunogens can drive the development of mature influenza bnAbs.

2.2: Testing of anti-idiotypic antibody K1-18 as a germline targeting immunogen

Anti-idiotype immunizations have been shown to trigger cellular and humoral immune responses, showing viability for these antibodies to act as immunogens (34). The anti-idiotype of the flu bnAb CR6261 known as K1-18 has been shown to bind efficiently to antibodies utilizing the VH1-69 heavy chain (Fig. 2.3A) (33). K1-18 has ~40nM affinity to germline reverted CR6261 and utilizes a specific motif for binding known as the IFY motif (Ile53, Phe54, and Tyr98) located on CR6261. These three residues are conserved across most germline and mature VH1-69 bnAbs and specifically bind to the

17 hydrophobic groove between HA1 and HA2 using key residues also found on K1-18 such as Tyr100d and Tyr58 (Fig. 2.3B.).

Activation of germline 9114 B cells by K1-18 anti-idiotype

To determine if K1-18 is able to induce activation of germline 9114 B cells, I isolated splenocytes from 9114 gH mice and tested isolated B cells for stimulation via calcium flux (Fig. 2.4A). Addition of full length K1-18 carrying a human Fc showed B cell stimulation via an increase in calcium flux marker Indo-1 as compared to wild-type

H1Cali09 and human IgG controls. Activation was further enhanced by cross-linked

(CXL) dimerization using a secondary anti-human Fc antibody. Activation was further shown through overnight incubation of 9114 gH B cells with K1-18 hFc and K1-18 hFc

CXL followed by staining for the B cell activation marker CD86. Both immunogens showed an increase in CD86 levels as compared to H1Cali09 wild-type (Fig. 2.4B).

Immunization of 9114 gH mice with K1-18 reveal 587H1-X mutation specific response

Preliminary immunization studies with K1-18 hFc showed little to no response generated potentially due to a lack of T cell help. Thus, we tried to alleviate this by conjugating the T cell peptide ovalbumin (OVA) to the K1-18 light chain and began immunizations with priming of 10μg OVA. After one immunization, a majority of the response generated was specific to 587H1-X and the response was minimal to 587H1-

WT (Fig. 2.5A). When 587H1-X and 587H1-WT solutions were preincubated with anti- stem bnAb 27F3, octet analysis showed minimal to no binding indicating that the response generated was 9114 gH specific. These responses were confirmed via

18 competitive ELISA (Fig. 2.6A). After a second immunization, three of the four mice decreased in 9114 specific response to 587H1-X and 587H1-WT, but the fourth mouse exhibited breadth of binding to 587H1-X, 587H1-WT, and cross binding to 270H5 via octet analysis (Fig. 2.5B) and ELISA (Fig.2.6B).

DNA Vaxibody constructs with K1-18 scFv show H1 crossreactivity

In collaboration with Gunnveig Grødeland, I developed a Vaxibody DNA construct containing K1-18 scFv as the antigenic unit, utilizing either the chemokine

MIP1α and αMHC-II as the targeting unit. Immunization of B6 x BALB/c F1 mice with the two different constructs at different amounts (50μg and 200μg) showed a significant response binding to K1-18 by day 21 post-immunization via endpoint titer ELISA. These

Vaxibody constructs also generated a response able to crossreact to 587H1 at day 21 post-immunization showing viability of the constructs in vivo and their ability to generate a potential heterosubtypic response. My activation studies, showing K1-18’s ability to stimulate VH1-69 germline reverted B cells combined with my results from protein and

DNA immunization studies show a potential for K1-18 to induce a broadly neutralizing,

9114 response.

19 2.3: Testing of engineered HA 587H1-X as a germline targeting immunogen

Artificially engineering an immunogen such as HA to have a higher affinity and thus, propensity to bind and activate precursor anti-stem bnAbs is the approach taken with 587H1-X. Shanshan Lang of the Wilson Lab at TSRI utilized the knowledge of the binding of K1-18 to germline VH1-69 bnAbs to generate point mutations in the HA of the H1Cali09 virus to enhance binding to the conserved IFY motif. Three mutations were generated to increase binding affinity to germline reverted CR6261 to ~31nM (Table 2.1,

Fig. 2.8A). The first mutation, T49Y, improved binding to Ile53 and Phe54 by providing additional hydrophobic interactions. The second mutation, L38Y, enhanced binding to the conserved Tyr98 by mimicking the Tyr100d in the K1-18 heavy chain. The third mutation, S291H, promoted germline reverted VH1-69 binding by establishing positive residues able to contact the negatively charged FR3 surface. Ultimately, the combination of the three mutations achieved nanomolar affinity and showed ability to bind to various

VH1-69 bnAbs such as 27F3 (Fig. 2.8B).

Calcium flux and binding to 9114 gH B cells show viability of mutated 587H1-X

The binding ability of the various 587H1 mutants were tested via staining of 9114 gH mouse splenocytes with tetramerized PE-labeled HAs (Fig. 2.9A). With the three mutations in wild type H1Cali09 HA the number of B cells able to bind the HA increases to a maximum amount for 587H1-X at ~33%. The L38Y mutation was shown to have fewer B cells able to bind than the 587H1-WT, so we tested the affinity and binding ability of a combination of just the other two mutations (T49Y and S291H) and it showed a lower affinity and percent of binding B cells as compared to the version with three

20 mutations, 587H1-X (data not shown). This indicates the L38Y or 587H1-C mutation is significant for germline VH1-69 binding. 587H1-X proved to activate 9114 gH B cells, as measured by calcium flux as compared to 587H1-WT which did not activate (Fig.

2.9B). Oligomerizing 587H1-X via Ni2+-NTA liposomes showed a significant increase in calcium flux, indicating the efficacy of multimerization in enhancing the antigenicity of these immunogens.

587H1-X immunizations show mutation specific 9114 response

Immunizations with 587H1-X, like with K1-18 hFc, showed no response to the immunogen, again potentially due to a lack of T cell help. We mitigated this again by the addition of ovalbumin peptide to the HA mutated and wild-type protein, generating

587H1-X-OVA and 587H1-WT-OVA, and by priming immunized mice with 10μg OVA prior to immunogen vaccination. One immunization with 587H1-X-OVA showed a 9114 specific response via octet in two mice that bound to wild-type 587H1 which was competed out by flu stem bnAb 27F3, while the wild-type 587H1 immunization generated no response (Fig. 2.10A). When these sera were tested via ELISA, there was a significant response to 587H1-X indicating a majority of the response was towards the X mutations (Fig. 2.10B). The two mice able to cross bind to 587H1-WT had minimal response via ELISA. After a third immunization with 587H1-X or 587H1-WT, octet analysis showed the two mice able to bind to 587H1-WT again, but also showed a significant increase in the nonspecific response, as assessed by the minimal competition of 27F3 (Fig. 2.10C). 9114 gH mice immunized with 587H1-WT exhibited a majority of

21 unspecific/non-9114 specific binding to 587H1-WT. ELISA analysis demonstrated the two mice immunized with 587H1-X again able to cross bind to wild-type 587H1 as well as a significant response to the mutated HA (Fig. 2.10D). Mice immunized three times with 587H1-WT also showed a 9114 specific response to both 587H1-X and 587H1-WT.

Sequencing analysis of heavy chains from 9114 gH mice immunized with 587H1-X show mature CR9114 matched mutations

After the immunization study, we took the two mice immunized three times with

587H1-X that showed ability to cross bind to wild-type 587H1, isolated splenocytes, and sorted their B cells to test for their ability to bind to 587H1 and 270H5. RNA from these antigen-sorted cells was isolated, RT-PCR was performed, and 9114 heavy chain PCR was done using a CR9114 specific primer and a degenerate IgG reverse primer. After sequencing of 33 clones that bound to 587H1 and 40 clones that bound to 270H5, we found that these heavy chains achieved a minimum of two mutations to a maximum of

13, with most clones averaging about 5-6 mutations in total (Fig. 2.11A). The locations of these mutations were found throughout the variable heavy chain (CDR1, 2, and 3) but more notably, the 270H5 binding clones had more mutations that matched to bnAb

CR9114 (Fig. 2.11B). Amino acid alignment and comparison of the 270H5 binding clones revealed two key mutations that matched those in CR9114: serine to asparagine located in the middle of CDR1 and alanine to threonine located at the end CDR2 (Fig.

2.11C). Expression of these clones as membrane bound IgM paired to germline 9114 light chain on 293 cells further reinforced the importance of these mutations (Fig. 2.11D).

Clone 3, a weak binder to 587H1 but a moderate binder to 270H5 carried the CDR2

22 mutation Ala to Thr. Clone 4 a moderate binder to 587H1 and a strong binder to 270H5 had both the Ala to Thr mutation in CDR2 and Ser to Asn in CDR1. These two clones reveal the importance of these two mutations for binding to various wild-type HA proteins. Clone 18 contained various other matched mutations in CDR1 and framework region 3 (FR3) that allowed it to strongly bind to 587H1 and 270H5 indicating the aforementioned two key mutations to potentially be redundant as further matched mutations occur. As a whole, these data show the ability of 587H1-X to be an immunogen capable of priming a VH1-69 and specifically a 9114 immune response able to cross bind to different HAs and to accumulate mutations that drive the maturation towards flu stem broadly neutralizing antibodies.

23 2.4: Discussion

The call for a universal influenza vaccine is made annually as seasonal influenza leads to a yearly death rate of about 49,000 and a hospitalization rate at over 200,000 per year (35). Additionally, the likelihood of genetic reassortment and antigenic shift puts public health at risk for a potential pandemic. Albeit the risk is not of the scope of the

1918 Spanish influenza pandemic because of our advances in healthcare and antiviral therapy through drugs such as the neuraminidase inhibitors Zanamivir and Oseltamivir

(36). As influenza virus continues to evade humoral immunity and antiviral drug resistance grows (37), the call for a universal vaccine grows as well. Various critical needs have been identified to address the development of a universal vaccine, some of which I address in this thesis and our approach: research that supports the rational design of immunogens that elicit long-lasting, high-affinity, broadly protective antibodies and animal models that elicit bnAbs and permit studies of their role in protecting against infection (38).

The development of the B6 9114 gH mouse model provides us a foundation for studying the progression of potential flu bnAbs and the immunogens capable of eliciting them. The development and identification of the two germline targeting immunogens by

Dr. Shanshan Lang of the Wilson Lab, 587H1-X and K1-18 prove to be promising immunogens. I have shown that both immunogens exhibit ability to stimulate precursor

9114 B cells via calcium flux and B cell activation markers. Through immunization studies, I have shown the potential of K1-18 vaccinations to elicit a response specific for mutations binding to the characteristic IFY motif of VH1-69 flu bnAbs and a response capable of binding to two wild-type group 1 HAs, 270H5 and 587H1 via ELISA and

24 biolayer inferometry assays. To further study K1-18, we inserted the K1-18 scFv into the

Vaxibody platform provided by Gunnveig Grødeland and showed APC targeting DNA immunizations generated a response specific for K1-18 that was able to crossreact to

587H1. The 587H1-X engineered HA immunogen, based on the binding of K1-18 to germline CR6261, demonstrated an increased ability to bind to germline 9114 B cells relative to wild-type 587H1. Through immunization studies with 587H1-X we have again shown a response elicited to the three introduced mutations and a response able to bind to the wild-type 587H1 in ELISA and octet assays. Further, after three immunizations of

9114 gH mice with 587H1-X, heavy chain sequencing analysis revealed mutations that matched to mature CR9114 and that allowed cross binding to wild type H1 and H5 HAs.

Ultimately, K1-18 and 587H1-X appear to generate a response that can crossreact to group 1 virus HAs and shows development towards the mature bnAb stage. Further binding tests to group 2 and potentially strain B viruses are needed and protection/neutralization studies are on the horizon.

The 9114 gH mouse model provides a precursor standard for a proof of concept that these immunogens are able to initiate maturation of germline bnAb precursors in humans, but the model may also come with limitations. The high (~90%) percentage of

9114 gH B cells raises concern for clonal competition, a significantly high number of B cells may compete for the amount of immunogen and hinder B cell maturation.

Furthermore, the increased lambda usage may potentially be due to a characteristic high autoreactivity for the germline 9114 heavy chain leading to light chain editing. The concern of the paired light chain to the 9114 heavy chain extends to the heavy chain sequencing analysis, the testing of binding to 587H1 and 270H5 when paired to a fixed

25 germline 9114 light chain showed a significant number of clones unable to bind either

HA whilst obtaining CR9114 matched mutations. This could indicate a role of the light chain in binding either affecting the conformation and binding ability of the heavy chain as a whole or simply being short of length thus allowing an increase in heavy chain contacts through the decrease in potential steric hindrance. The need for a specific light chain for binding could also explain why only ~30% of B cells in 9114 gH mice bind to

587H1-X. Further light chain sequencing is necessary to understand the role of light chain in binding and the development of response.

Future Directions

The efficiency of K1-18 and 587H1-X to generate a 9114-specific response that can crossreact to various wild-type HAs provides evidence of binding that can plausibly provide protection in infection and neutralization studies. From our data, K1-18 showed a

25% efficacy rate and 587H1-X showed a 50% efficacy rate in ability to generate a wild- type binding response to at least 587H1. There may be multiple reasons for this suboptimal efficacy, such as the following: a high percentage of 9114 B cells may lead to clonal competition and hinder B cell maturation, a lack of immunogenicity of our germline targeting immunogens and specifically the stem region to name a few. To increase efficiency and immunogenicity, these immunogens could be multimerized.

Multimerizing immunogens on ferritin nanoparticles has been shown to increase immunogenicity and to elicit neutralizing antibodies able to protect against a variety of

H1 viruses (39). Loading our two immunogens on ferritin nanoparticles may also increase our efficiency of response simply through maximizing accessibility to these

26 epitopes, which are normally near the viral membrane. In order to address the potential of clonal competition in a high percentage of 9114 B cells in our mouse model, diluting out germline 9114 B cells by transfer to wild-type recipients followed by immunizations studies may alleviate this problem. Another future mouse model the Nemazee lab is developing is a mouse containing a rearrangeable VH1-69*01 gene allowing for the development of a wide array of anti-stem flu bnAbs (Fig. 2.12A). This mouse model would decrease the number of specific germline reverted bnAbs and focus on the VH1-69 family and its ability to randomly generate potential germline antibodies. Another method to increase wild-type heterosubtypic crossreactivity is simply by priming with one of our two immunogens to increase and stimulate the specific germline reverted B cells we are targeting and follow up with a boost immunization with a wild-type HA, potentially an

H1 HA to increase a humoral response on a similar stem epitope or a group 2/H3 HA to influence group 2 cross binding ability. An approach used to increase efficiency of the anti-stem humoral response is removal of the immunodominant head region of HA and to trimerize the stem region for use as an immunogen (30, 32). This approach would decrease distracting epitopes generating unspecific response, and utilizing these stem immunogens in the aforementioned wild-type boost would increase the stem response we desire. Furthermore, incorporating the three 587H1-X mutations into a stem immunogen would further increase a stem-specific response. Truncating a full length HA removes a majority of immunodominant epitopes but also removes many T cell epitopes, raising a concern of poor T cell help.

The Vaxibody DNA immunization platform is a potential way to ameliorate a lack of immunogenicity through direct APC targeting. Utilizing different targeting units

27 such as MIP1α or αMHC-II allows us to direct the immune response towards cellular or humoral immunity, respectively. Manipulating the targeting units to potentially target

VH1-69 precursor B cells utilizing K1-18 scFv as a targeting unit may be a way to target and increase an influenza bnAb response. Manipulating the antigenic units to contain

587H1-X engineered HA can potentially increase immunization efficiency as well.

Studies to understand the formation and secretion of Vaxibody proteins are necessary to understand the mechanism of the response using DNA vaccination and electroporation.

Furthermore, past immunization studies with Vaxibody utilized only a truncated, monomeric HA protein begging the question whether the constructs can hold and maintain the trimeric conformation of full length HA. The Vaxibody platform proves to be an efficient and flexible alternative to protein and egg-based development of influenza vaccines. Immunization experiments are continuing to study the immune response generated by these K1-18 Vaxibody constructs and ultimately to determine if the response is protective against infection.

The final stage of immunogen testing is assessing the ability of the generated responses to protect against viral challenge and more importantly, heterosubtypic viral challenge. The initial protection assay would be testing mouse serum antibodies by in vitro neutralization of viruses carrying distinct HA subtypes, but a recent concern is the discovery of the FcγR mediated mechanism of neutralization, which will not show neutralization via in vitro assays (28). Another concern in vaccine development is previous exposure and how previous immune responses can prevent or affect these vaccines. This concern being the deciding factor as to when a theoretical vaccination should occur in an individual’s life, if vaccination is necessary early in a child’s life or

28 prior exposure is an issue. Studies show previous exposure of mice with non-lethal doses of PR8/H1 virus or HA subunit exposure via DNA vaccination does not prevent the development of pan-H1 bnAbs (42), thus showing prior exposure does not prevent the development of broadly neutralizing responses. Our data provide evidence that 587H1-X and K1-18 both show promise as immunogens capable of generating a broadly binding response via germline targeting, but ultimately the final question is whether they act as viable vaccine candidates and protect against heterosubtypic viral challenges. As a whole, our observations of K1-18 and 587H1-X show a novel method of vaccination through the concept of germline targeting in mouse models containing genes expressing the precursor heavy chains of influenza broadly neutralizing antibodies.

29 2.5: Methods

Sequence determination of K1-18 antibody. The K1-18 hybridoma cells were obtained from MAB Institute Inc (33). Cellular RNA was extracted and cDNA was prepared by using random primers. The VHDJH sequence was amplified by PCR with a mixture of downward primers that cover all the mouse VH genes and a mixture of upward primers that cover four JH genes. The VKJKCK sequence was amplified by PCR with a mixture of downward primers that cover all the mouse VK genes, and an upward primer that is complementary to the 3' end region of CK gene. The vector expressing Fab form of mouse antibody contains a SfiI-XhoI site to accommodate VHDJH sequence followed by the CH1 gene of mIgG1. It also contains an NcoI-AscI site to accommodate VKJKCK gene. At first, the amplified VKJKCK sequence was inserted into the vector, and then the

VHDJH sequence was further inserted. Several clones were picked and their sequences were confirmed to be identical. K1-18 Fab was prepared using the generated construct and their binding activity to VH1-69 antibodies was determined to be consistent with that of original K1-18 clone.

*Performed by Shanshan Lang of the Wilson Lab, TSRI

Fab and IgG cloning, expression and purification. The germline Fabs were expressed in insect cells. In brief, the germline Fab genes were cloned into the pFastBac Dual vector (Invitrogen) with N-terminal gp67 secretion signal peptide fused to the heavy

30 chain and N-terminal HBM secretion signal fused to the light chain. Recombinant bacmid

DNA was prepared using the Bac-to-Bac system (Invitrogen) and Baculovirus was generated by transfecting purified bacmid DNA in to Sf9 cells. Fabs were expressed by inoculating High Five cells with the recombinant virus, shaking at 110 r.p.m. for 72h at

28ºC. The Fabs were directly collected from the supernatant via Ni-NTA Superflow

(Qiagen) and purified using monoS chromatography (GE Healthcare).

K1-18, CR6261 and their mutant Fabs were expressed in 293F mammalian cells. The heavy and light chains of the Fab were cloned independently into the phCMV3 vector and fused with the N-terminal IgK secretion signal peptide. A His6 tag was added to the

C-terminus of the Fab heavy chain. Recombinant DNAs for both heavy and light chains were purified separately and co-transfected into 293F cells. The cells were cultured for 6-

7 days at 37ºC, while shaking at 110 r.p.m. Secreted Fabs were purified as described above.

To generate the VH1-69 GL IgG, the DNA fragment of the VH domain was fused with the DNA fragment of heavy chain Fc domain of human IgG1 via PCR. The full-length gene was cloned into the phCMV3 vector with the N-terminal IgK secretion signal peptide. IgG was expressed in 293F cells, same as above and purified by Protein G and monoS chromatography (GE Healthcare) and gel filtration. Mutations were introduced into each of the original plasmids by QuikChange II Site-Directed Mutagenesis (Agilent technologies). Recombinant plasmids were sequenced and purified for expression.

*Performed by Shanshan Lang of the Wilson Lab, TSRI

31 Kd determination and binding kinetics analysis by bio-layer interferometry. An

Octet RED instrument (FortéBio, Inc.) was used to determine Kd of the antibody-antigen interactions by bio-layer interferometry. The association and dissociation curves were processed using the Prism GraphPad. To determine the binding affinity of GL-CR6261

IgG to HA, biotinylated-HA or HA mutants were diluted to 10-50 μg/ml in PBS pH 7.4,

0.01% BSA and 0.002% Tween 20. HAs were immobilized onto streptavidin-coated biosensors (FortéBio, Inc.) and incubated with VH1-69 IgG at 2-8 μM. The signals for each binding events were measured in real-time and Kd values determined by fitting to a

1:2 bivalent analyte model.

Anti-human Fab-CH1 biosensors (FortéBio, Inc.) were used to test the binding affinity of

K1-18 to the VH1-69 GL, CR6261 or their related mutants. In brief, the targeted antibodies were immobilized on the anti-human Fab-CH1 sensors and then incubated with K1-18 IgG (12.5-100 nM) or Fab (0.25-1.0 μM). The Kd values are determined by fitting to a 1:2 bivalent analyte model (IgG) and a 1:1 binding model (Fab), respectively.

To determine the binding affinity of immunized mouse sera to HA, biotinylated-HA or

HA mutants were diluted to 10-50 μg/ml in PBS pH 7.4, 0.01% BSA and 0.002% Tween

20. HAs were immobilized onto streptavidin-coated biosensors (FortéBio, Inc.) and incubated with immunized mouse sera diluted 1/20. The signals for each binding events were measured in real-time.

*Performed by Shanshan Lang of the Wilson Lab, TSRI

32

Crystallization and structural determination of K1-18 in complex with VH1-69 germline antibodies. K1-18 Fab was mixed with GL-CR6261 Fab in a 1:1 molar ratio, and the sample was purified by gel-filtration on a Superdex75 column (GE healthcare).

The complex was concentrated to 8.0 mg/ml in Tris-HCl pH 7.0, 50mM NaCl. The complex was initially crystalized by sitting-drop vapor diffusion with 0.1 M MES pH 6.0,

30% PEG6000 as mother liquor (20ºC). However, the initial crystals diffracted poorly.

For further screening, similar crystals were transferred to ~40 μl mother liquor and mechanically homogenized by vortexing with a seed bead (Hampton Research). The seed solution (0.05 μl) and complex sample (0.15 μl) were then mixed using a Douglas Oryx 8 crystallization robot with JCSG Top96 cryo solutions (0.15 μl) by vapor diffusion (43).

The best crystals grew in the well with mother liquor consisting of 5% (w/v) PEG 6000,

0.1 M Citric acid pH 4.0, and 25% PEG200 (4°C). The crystals were directly cryoprotected in liquid nitrogen for data collection on beamline 11-1 at the Stanford

Synchrotron Radiation Lightsource (SSRL). The diffraction data were processed with

HKL2000 and the structure determined to 2.1 Å resolution by molecular replacement

(MR) using Phaser (44) with mouse antibodies (PDB 3VRL and 1FJ1) and GL-CR6261

Fab as the MR starting models. Refinement was carried out in Refmac (45), Phenix (46) and manual adjustments in Coot (47). The final model has 96.5% residues in the favored regions of the Ramachandran plot with only 0.4% outliers (48).

*Performed by Shanshan Lang of the Wilson Lab, TSRI

33 Crystallization and structural determination of K1-18 in complex with CR6261. The complex was prepared as described above and concentrated to 7.5 mg/ml in 50 mM

NaOAc, pH 5.5. The initial crystal hits came from vapor diffusion screening with mother liquor of 0.1M bicine pH 8.16, 1M lithium chloride, 15% PEG2000 (20°C). Initial crystals were mixed with ~40 μl of mother liquor and mechanically homogenized by vortexing, and the suspension used for microseeding. Crystals grew in multiple conditions and the best hits came from 0.2 M calcium acetate, 10% (w/v) PEG 8000, 0.1

M imidazole pH 8.0, 20% ethylene glycol (20oC). Diffraction data were collected at the

GM/CA-CAT 23ID-B beamline. The structure was determined to 3.5 Å by molecular replacement (MR) in Phaser (44) using CR6261 and K1-18 Fabs as MR models and further refined by Refmac (45), Phenix (46) with manual adjustments using Coot (47).

The final model has 95.2% of the residues in the favored regions of the Ramachandran plot with 0.9% outliers (48).

*Performed by Shanshan Lang of the Wilson Lab, TSRI

Crystallization and structure determination of the complex of H1Cali09-X with

VH1-69 Fabs. The H1Cali09-X HA was expressed in High Five Cells and purified as the protocol above. HA is partially digested with trypsin (mass ratio of protein to enzyme was about 1000:1, at 4°C overnight) to cleave off the tags and convert HA0 into HA1 and

HA2. Cleaved HA trimer was purified by gel filtration. The HA complex with either

27F3, GL-CR6261 or UCA9 was formed by mixing the HA trimer with the targeted Fab in a molar ratio of 1: 3.6 and purified by gel filtration on a Superdex200 column (GE

Healthcare) in 20 mM Tris-HCl pH 8.0, 150 mM NaCl. For the complex of H1Cali09-

34 X_VHGL-CR9114, the sample was prepared by directly mixing the HA trimer and

VHGL-CR9114 in a molar ratio of 3:1 without further gel filtration.

All complex samples were concentrated to 8-10 mg/ml for crystallization screening on our high-throughput robotic Rigaku CrystalMation system at TSRI using sitting-drop vapor diffusion. Multiple crystals were observed in diverse conditions for every complex samples. The conditions producing the best diffracting crystals and corresponding cryo- protections are summarized as following

H1Cali09-X_27F3:

0.1 M Tris-HCl, pH 8.5, 1,26 M ammonium sulfate, 0.2M lithium sulfate

cryo-protected by addition of 20% glycerol

H1Cali09-X_GL-CR6261:

0.1 M sodium chloride, 1.6 M ammonium sulfate, 0.1 HEPES 7.5;

cryo-protected by addition of 20% glycerol

X-ray diffraction data were collected at multiple beamline (Table 4.1, 4.2). The diffraction data were processed with HKL2000 and the structure was determined by molecular replacement in Phaser (44) using a 2009 pandemic H1 HA model (PDB

4M4Y)(54) and corresponding Fab models (see in Chapter-2) . Refinement was carried out in Refmac (45), Phenix (46), and model rebuilding was performed manually in Coot

(47) and the model was validated by MolProbity (48).

35 *Performed by Shanshan Lang of the Wilson Lab, TSRI

Structural analysis. Interaction and interface analysis is carried out on online server

PDBePISA at www.ebi.ac.uk/pdbe/pisa/. Structure figures were generated by MacPyMol

(DeLano Scientific LLC).

*Performed by Shanshan Lang of the Wilson Lab, TSRI

HA expression, purification and biotinylation. Baculovirus-expressed HA was prepared for the binding studies by cloning the HA genes into the pFastBac vector with an N-terminal gp67 secretion signal peptide and a C-terminal BirA biotinylation site, thrombin cleavage site, foldon trimerization domain, and His6 tag. Recombinant bacmid

DNA was generated via the Bac-to-Bac system (Invitrogen) and Baculovirus was generated by transfecting purified bacmid DNA in to Sf9 cells. HAs were expressed by infecting the High Five cells with the recombinant virus, shaking at 110 r.p.m. for 72h at

28ºC. The secreted HAs were purified from the supernatant via Ni-NTA Superflow

(Qiagen). Concentrated HAs were biotinylated with BirA enzyme in 100mM Tris-HCl

(pH8.0), 10 mM ATP, 10 mM magnesium acetate, 50 μM biotin at 37ºC for 1-2 hour.

After gel filtration, the biotinylated-HAs were concentrated and stored at -80ºC.

Mutations of HA genes were introduced into each of the original pFastBac plasmids by

QuikChange II Site-Directed Mutagenesis (Agilent technologies).

*Performed by Shanshan Lang of the Wilson Lab, TSRI

36 VH1-69GL B Cell activation

The gl-CR9114 knock-in mice were generated in the same way as previously described.

In brief, splenocytes were suspended at 4 million cells/ml in HBSS, labeled with 0.75 μM

Indo-1 (Invitrogen), for 30 min at 37°C and washed with 2 mM CaCl2 HBSS, followed by another 30 min at 37°C. 300ul cells at 2 million cells/ml were then stimulated at 37°C with the indicated ligands. Ca2+ signals were recorded for 300 s, measuring the 405/485- nm emission ratio of Indo-1 fluorescence upon UV excitation. Calcium flux analysis was performed on an LSR II cytometer (BD Biosciences). Kinetic analysis was performed using FlowJo (Tree Star).

Mice

8-12 week old mice were used in most experiments. C57BL/6J mice were purchased from the Scripps Research Institute breeding colony. 9114 gH mice were generated on the C57BL/6J background. BALB/c mice were crossed to C57BL/6J mice and the F1 mice were used for DNA vaccinations. All mice were bred and maintained in the Scripps

Research Institute Animal Resources facility according to The Scripps Research Institute

Institutional Animal Care and Use Committee guidelines.

Transgenic mouse generation

Knock-in mouse generation of 9114 gH mice was carried out essentially as described (55,

56), by substituting the 9114 gH VDJ exon in the targeting construct by overlap PCR.

The gl9114 H chain gene fragment was synthesized at Initrogen. It included a part of the

37 intron, gl9114 VH, and part of 5’ arm. The fragment was cloned to double digested gl-

VRC01 targeting vector. Briefly, linearized gl9114 H with SalI was introduced into

C57BL/6J-derived embryonic stem cells and selected in media supplemented with G418.

Cells that insert the gene non-specifically carry the DTA gene, which is toxic and cells lacking the neomycin resistance gene were counterselected. Positive clones were identified initially by a PCR strategy and confirmed by southern blot analysis.

Polymerase chain reaction conditions were as follows. Reactions were carried out with

Q5 Hot start Master mix (NEB) according to the instructions of the manufacturer, using primers at a final concentration of 0.25 mM and ~200 ng genomic DNA template.

Reaction conditions were 98°C 45 sec (1 cycle), followed by 98°C 15 sec, 67°C 30 sec,

72°C 1 min (35 cycles) then 72°C 10 min. Primers used were as follows: gv3_13: GTG CAG TCT GGG GCT GAG GTG AAG gv3_3123: AGC ACT CAG AGA

AGC CCA CCC ATC T

The amino acid sequence used for 9114 gH encodes the following predicted germline protein V region:

QVQLVQSGAEVKKPGSSVKVSCKASGGTFSSYAISWVRQAPGQGLEWMGGIIPIF

GTANYAQKFQGRVTITADESTSTAYMELSSLRSEDTAVYYCARHGNYYYYSGM

DVWGQ

38 The upstream promoter, leader and intron elements were from VHJ558.85.191.

Targeting, embryonic stem cell screening and mouse generation were as described (55,

56).

*Performed by Takayuki Ota of the Nemazee Lab, TSRI

Vaccination

In each of the protein immunization studies, B6 9114 gH mice were immunized via intraperitoneal injection with 20μg of protein dissolved in PBS and added to 100μL Ribi, also known as Sigma Adjuvant System (SAS).

In the Vaxibody DNA immunization studies, B6 x BALB/c F1 mice were anesthetized by isoflurane and hind legs were shaved. 50μL DNA construct dissolved in PBS was injected intramuscularly on each hind leg of the mouse, immediately followed by skin electroporation using the TriGrid Delivery System (Ichor).

Flow cytometry analyses

Analyses for surface markers and Ca2+ flux assay were performed using standard protocols. The following mAbs were used at 1:200 in all experiments described and were from eBioscience unless otherwise indicated: CD19 (1D3; PE/Cy7), TCRβ (H57;

PerCPCy5.5), StreptAvidin-PE, B220 (RA3-6B2; FITC; BioLegend), IgG1 (A85-1;

FITC; BD Pharmingen), IgG2a (RMG2a-62; FITC; Biolegend), IgG2b (R12-3; FITC;

BD Pharmingen), IgG3 (R40-82; FITC; BD Pharmingen), CD86 (IT2.2; PE), Igκ (187.1;

FITC or Pacific Blue), Igλ (RLM-42; APC; Biolegend). mAbs against mouse IgM (M41;

39 Alexa Fluor 488) and B220 (Pacific blue) were labeled in-house. Antigen binding B cells were detected with biotinylated 587H1-WT, 587H1-X, 270H5, 575H3 followed by streptavidin-PE. Samples were read on an LSR-II instrument (BD) and analyzed using the FlowJo program (Tree Star, Inc.). Mean fluorescence intensity values given were calculated by geometric means of the indicated fluorophore staining of the gated cells.

Mutation analysis

Splenocytes were enriched with biotinylated-587H1-WT or 270H5 with Anti-Biotin

Microbeads. The enriched fraction were stained with appropriately diluted Anti-B220-

BV421 (BD), Anti-IgG1 (A85-1, BD), 2a, 2b-FITC (eBioscience), Strepavidin-PE

(eBioscience), Anti-TCRb (H57)-PerCP Cy5.5 (eBioscience), Anti-IgM (M41) Alexa-

647, Anti-IgD (JA12.5) Alexa-647. B220+, IgGs+, TCRb-, IgM-, IgD- fraction was purified using using FACS Aria (BD). Cells were directly sorted in Cells-to-cDNA lysis buffer (Thermo Fisher) and RNAase was inactivated with incubation at 75ºC for 10 minutes. RT-PCR was done using OneStep Ahead RT PCR kit (QIAGEN) with Forward primer:

5’-TGCAGTCTGGGGCTGAGGTGAA-3’

And degenerate IgG reverse primers:

IgGR1 5’-CAGATGGGGGTGTCGTTTTGGCTGA-3’

IgGR2a 5’-CTGATGGGGGTGTTGTTTTGGCTGA-3’

IgGR2b 5’CCGATGGGGCTGTTGTTTTGGCTGA-3’

40 Amplified heavy chain product was cloned into PCR cloning kit (NEB) and each clones sequence was confirmed. For heavy chain expression analysis, leader sequence, VH, and

IgM constant region (membrane bound form) were amplified separately and subcloned into pCI-WPRE plasmid with NEBuilder. Each heavy chain plasmid was co-transfected with gl9114 light chain plasmid into 293T cells using Lipofectamine 2000 (Thermo

Fisher). Surface Ig expression and binding to 270H5 or 587H1 was confirmed with FACS analysis.

Competitive 9114 Serum ELISA

Maxisorp plates (Thermo Fisher Scientific) were coated with 10µg/mL streptavidin

(Sigma) overnight at 4ºC, then blocked with 1% BSA in buffered saline (ELISA buffer) for 1 hour. followed by incubation with 2µg/mL of 587H1-X, 587H1-WT, or 270H5 at room temperature for 2 hours. Plates were then blocked with 1% BSA in buffered saline for 30 minutes followed by a competitive epitope block using 10 µg/mL mature CR9114 human IgG for 1 hour followed by incubation with mouse sera diluted in ELISA buffer for 1 hour. Purified mature CR9114 mouse IgG was used as a standard. Horseradish peroxidase-conjugated polyclonal anti-mouse IgG diluted in ELISA buffer was used to report signals using Ultrasensitive TMB solution (Millipore) per manufacturer’s instructions. Signals were recorded at 450nm using a microplate reader (Versa-Max;

MDS Analytical Technologies).

41 2.6: Figures and Tables

Figure 2.1. Design of 9114 gH knock-in mice. 9114 gH mice were generated by replacing the cluster from DQ52-JH with germline reverted VDJ from CR9114 along with associated promoter and leader elements. *Figure provided by Dr. Takayuki Ota of the Wyatt Lab, TSRI and Dr. David Nemazee

42

C D gl9114 WT

105 27.2 7.06

104 1.39

103 : 575

102 H3

0 λ - 11.8 67.6 88.7

0 102 103 104 105

575 : 587 spleen, B220+ 587-H1 κ

Figure 2.2. Characterization of the B6 9114 gH Mouse Model. (A) B6 9114 gH mice show normal B cell numbers in various lymphoid tissues. (B) B6 9114 gH mice show heavy chain allelic exclusion with ~90% oh B cells expressing germline 9114 heavy chain. Left panel shows wild type IgHa/b mice, right shows F1 9114 gH IgHb/b x IgHa/a mice. (C) A population of 9114 gH B cells bind to 575/Victoria2011/H3. (D) 9114 gH B cells express an elevated percentage of λ light chain. *Figures (A, B) provided by Dr. David Nemazee and Patrick Skög of the Nemazee Lab, TSRI

43 A B Contact surface on GL-CR6261 Fab A GL-CR6261 H K1-18H B Contact surface summary on GL Fab Region Surface area (Å2) Percentage GL-CR6261 H K1-18H Region Surface area (Å2) Percentage Total 771 100% Total 771 100% Chain H 673 673 87% 87% 80 10% CDR H1 80 10% CDR H2 268 268 35% 35% FR3 100 100 13% 13% K1-18L CDR H3 224 224 29% 29% K1-18L Chain L 98 98 13% 13%

B Paratope of K1-18 C K1-18 H D E Y58 Y58 K1-18 L L94 K1-18 L I53 L94 CDR H3 L94 GL GL CDR H3 CDR H2 W96 CDR H2 Y98 I53 F54 K1-18 L H95 W96 Y100d CDR H2 Y98 K1-18 H Y100d F54

W96 H95 Y100d

K1-18 H

Figure 2.3. Structure and paratope of K1-18. (A) Crystal structure of K1-18 fab (heavy chain in green, light chain in blue) in complex with germline VH1-69 antibody CR6261 (heavy chain in yellow, light chain in gray). IFY motif (Ile53, Phe54, Tyr98) in red. (B) Interaction overview of K1-18 and CDRH2 of germline CR6261 showing binding of the IFY motif of CR6261 (yellow) to Tyr100d and Tyr58, two residues also found in the hydrophobic groove of HA between HA1 and HA2. *Structures and figures provided by Dr. Shanshan Lang of the Wilson Lab, TSRI

44 A

B

Figure 2.4. K1-18 activates B6 9114 gH B cells. (A) Calcium flux activation assay of B6 9114 gH B cells stimulated with various antigens reveals K1-18 and dimerized K1-18 (yellow and orange, respectively) induces stronger activation as compared to wild-type HA. (B) Overnight incubation with K1-18 and K1-18 CXL (color as in (A)) shows increase in B cell activation marker CD86.

45 A Immunization: OVA > K1-18-OVA To 587H1-X: To 587H1-wt: 1/20 sera dilution binding to 587X 1/20 sera dilution binding to 587wt

4 4 K1-18-OVA 1 K1-18-OVA 1 K1-18-OVA 2 3 3 K1-18-OVA 2 K1-18-OVA 3 K1-18-OVA 3 2 K1-18-OVA 4 2 K1-18-OVA 4 nm Day0 nm Day0 1 1

0 0 100 200 300 100 200 300 -1 Time (s) -1 Time (s)

To 587H1-X blocked by stem Fab (27F3): To 587H1-wt blocked by stem Fab (27F3): 1/20 sera dilution binding to 587X + 27F3 Fab 1/20 sera dilution binding to 587wt + 27F3 4 K1-18-OVA 1 4 K1-18-OVA 1 3 K1-18-OVA 2 3 K1-18-OVA 2 K1-18-OVA 3 K1-18-OVA 3 2 K1-18-OVA 4 2 K1-18-OVA 4

nm Day0 1 nm Day0 1

0 100 200 300 0 100 200 300 -1 Time (s) -1 Time (s)

B Immunization: OVA > K1-18-OVA > K1-18-OVA To 587H1-X: To 587H1-wt:

To 270H5:

Figure 2.5. K1-18 Immunizations show 587H1-X IFY enhancing mutation-specific response via Octet. Binding of mouse sera from four B6 9114 gH mice immunized once (A) and twice (B) with OVA conjugated K1-18 to respective antigens. Same sera was then blocked using a competitive anti-stem flu bnAb, 27F3 ((A), bottom row). Each line corresponds to a different mouse. *Figures provided by Dr. Shanshan Lang of the Wilson Lab, TSRI

46 A Immunization: OVA > K1-18-OVA

p =0.0024 10000 **

1000

100

10 [9114 mIgG] µg/mL mIgG] [9114

1

587-Bio 587X-Bio

B Immunization: OVA > K1-18-OVA > K1-18-OVA

10000

1000

100

10 [9114 mIgG] µg/mL mIgG] [9114

1

587-Bio 587X-Bio 270H5-Bio

Figure 2.6. K1-18 Immunizations show 9114-specific, 587H1-X IFY enhancing mutation-specific response via competitive ELISA. Binding of mouse sera from four B6 9114 gH mice immunized once (A) and twice (B) with OVA conjugated K1-18 to 587H1-X biotin labeled (587X-Bio), 587H1 biotin labeled (587-Bio), and/or 270H5 biotin labeled (270H5-Bio). Each dot corresponds to a different mouse.

47 A

B

Figure 2.7 Vaxibody DNA immunizations with K1-18 scFv as the antigenic unit show responses to K1-18 and 587H1 via endpoint titre ELISA. Binding of mouse sera from BALB/c x B6 9114 gH mice immunized once intramuscularly with respective DNA constructs to K1-18 (A) and 587H1 (B). 50 and 200 indicate immunization with 50μg and 200μg of DNA construct, respectively. n=5 per group, *P < 0.0001

48 Immunogens that target anti-HA stem B cells Dr. Schief chose a similar approach to engineer immunogens for anti-stem gl-bNAbs, using both scaffolds and modified full length HA. In the scaffolded approach, we developed a monomeric engineered HA stem scaffold, or "eStem", by remodeling the stem epitope with a de novo three helical bundle core, different connectivity and loops, and additional stabilizing, resurfacing and glycan-masking mutations to focus immunogenicity to the stem bnAb epitopes (Fig. 4). The eStem is essentially a perfect monomer in solution (Fig. 4A), exhibits extreme thermo-stability, with no significant changes in structure detected by CD or DSC at 100° C (Fig. 4B-C) and binds with extremely high affinity to stem bnAbs (Fig. 4D). We will employ eStem as well as the published headless HA scaffolds from the VRC (20) and Crucell/Wilson (21) as platforms on which to add mutations to confer affinity for germline variants of stem bnAbs. While developing eStem, we worked in parallel to begin discovery of germline-targeting mutations on full-length HA. In this effort, computational analyses were used to design yeast display libraries carrying full-length HAs with combinations of mutations predicted to facilitate germline binding. We sought HA variants that i) bound well to germline-reverted versions of both CR9114 and CR6261, ii) bind at least as well or even better to the original mature CR9114 and CR6261, and iii) failed to bind to irrelevant, non neutralizing antibodies. In addition, initial hits with reasonable affinities were mutated randomly to identify HAs with improved properties. We identified a first generation HA mutant (GT-HA) that bound to 9114 gH when paired with mutated 9114 L (Fig 3). Subsequent selection rounds identified HA-GT5, which has no L-chain dependence (Fig 6, below). This same general approach will be used for the scaffolded antigens, but we will switch to mammalian display as described below. In Aim 1 we will engineer variants of eStem and headless HA (20,21) that bind germline variants of CR9114 and related stem bnAbs, as well as the mature stem bnAbs.

CR9114 CR9114 gH / mutated L Figure 3. Isolation of full-length HA mutant WT HA GT HA WT HA GT HA engineered to bind germline-reverted 1"nM" 1"nM" 100"nM" 100"nM" CR9114, and hence is germline targeted (GT HA). FACS assay shows binding of CR9114 or CR9114 gH/mutated L antibody to yeast displaying GT HA or unmutated, wild type HA IgG binding (WT HA). These data verify that gCR9114 has negligible affinity for WT HA (tested at 100 nM). myc-TAG (Display) Generation of germline targeting HA stem immunogens based on the solution of novel structures. Using a combination structural analysis of VH1-69 flu stem bnAb:HA complexes, the novel complex of gl- CR6261 with anti-idiotypic antibody K1-18 and site-directed mutagenesis, the Wilson lab has made important insights about the design of germline targeting HA that can be applied to the VH1-69 anti-stem class antibody and perhaps be extended to the VH6-1+DH3-3 class. The analysis suggests that the key contacts on the antibody “IFY motif”-- I53, F54, Y98-- encoded within gl-VH1-69 were recognized with higher affinity by K1-18. A series of point mutants of H1Cali04 was generated with the aim to artificially strengthen the contacts by substituting K1-18 contacts. Binding to Y98 was enhanced by the I38Y change. Similarly, T49Y improved binding to I53/F54. Finally, it was noted that several flu stem bnAbs are mutated at position 72 in FW3 to contact S291 on HA. The mutation S291H was found to promote gl-VH1-69 binding. The I38Y+T49Y+S291H combination (denoted here by the suffix "X") could enhance the binding of H1Cali04 to Tablegl-VH1-69 2.1. by 1000Mutations-fold (Table in 2).Cali09/H1 that improve binding to germline VH1-69 bnAb

CR6261.Table 2. Identification of mutations in H1 that improve binding by full length HA to gl-VH1-69 bnAb CR6261. Clone ID Mutations Affinity to V 1-69 IgG (nM) H 587 (H1 Cali04) WT ~54,000 587c L38Y 1145.0±499.0 587g T49Y 6049.0±123.8 587i L38Y T49Y 190.9±7.7 587x L38Y T49Y S291H 31.1±1.2

*TableImportantly, provided these three by Dr.HA mutations Shanshan not Lang only enhance of the Wilsongl-VH1-69 Lab, binding, TSRI but also maintain tight binding of mature bnAbs CR9114 and CR6261. Moreover, these same mutations could be grafted onto H1 SC1918, enhancing affinity from undetectable to sub-micromolar, 82 nM for the IgG form (Table 3). 5

49

D

H1Cali09-X

Fab 27F3

Figure 2.8. Structures outlining mutations in Cali09/H1 that improve binding to germline VH1-69 bnAb CR6261. (A-C) show the mutated residues and their contacts to bnAb 27F3. (A) T49 was mutated to a Tyr in order to provide additional hydrophobic interactions to Ile53. (B) L38 was mutated to Tyr in order to enhance binding to Y98. (C) Positively charged residues were introduced to the S291 site, resulting in H291 leading to the highest affinity. By combining all three mutations, affinity to germline CR6261 rose to ~31nM. (D) Crystal structure of H1Cali09-X (one monomer is colored: HA1 in green, HA2 in blue) in complex with anti-stem flu bnAb 27F3 (heavy chain in pink, light chain in light gray). *Structures and figures provided by Dr. Shanshan Lang of the Wilson Lab, TSRI

50

A 587x 587i 587g 587c 587

33.4 23.6 22.7 7.93 16.2 9114 gH

0.395 0.526 0.61 0.358 0.958 Wild-type B220 Tetramer

B

Figure 2.8. In vitro binding and activation of 587X-H1 to B6 9114 gH B cells. (A) Staining of B220+ B cells from B6 9114 gH mice with tetramerized PE-labeled HAs of the various mutations. Refer to Table 2.1 for mutation details. Top row, 9114 gH spleenocytes. Bottom row, C57BL/6J wild type spleenocytes. (B) Calcium flux B cell activation assay of 9114 gH B cells stimulated with designated antigens. H1Cali09-wild type and H1Cali09-X were oligomerized onto Ni2+-NTA liposomes and used for stimulation along with the soluble H1Cali09-X antigen. *Data developed alongside Dr. Taka Ota of the Wyatt Lab, TSRI

51 A Immunization: OVA > 587H1-X Immunization: OVA > 587H1-WT

Binding587x-ova to : 587H1sera to -587H1wt Binding587wt-ova to : 587H1 sera to-wt 587H1 2.0 2.0 #1 #5 1.5 #2 1.5 #6 #3 #7 1.0 #4 1.0 #8 nm Day0 nm Day0 0.5 0.5

0.0 0.0 100 200 300 100 200 300 -0.5 Time (s) -0.5 Time (s)

Binding587x-ova to: sera 587H1 to 587H1-27F3-wt + 27F3 Binding587wt-ova to: 587H1sera to 587H1-27F3-wt + 27F3

2.0 2.0 #1 #5 1.5 #2 1.5 #6 #3 #7 1.0 #4 1.0 #8

nm Day0 nm Day0 0.5 0.5

0.0 0.0 100 200 300 100 200 300 -0.5 Time (s) -0.5 Time (s)

B Immunization: OVA > 587H1-X Immunization: OVA > 587H1-WT

** * 10000 P = 0.0051 10000 P = 0.0293

1000 1000

100 100

10 10 [9114 mIgG] µg/mL mIgG] [9114 [9114 mIgG] µg/mL mIgG] [9114

1 1

587-Bio 587-Bio 587X-Bio 587X-Bio

Figure 2.10. 587H1-X Immunizations show IFY motif enhancing mutation-specific response via octet and ELISA assays. Binding of mouse sera from four B6 9114 gH mice immunized once with OVA conjugated 587H1-X (left column) or once with OVA conjugated 587H1-WT (right column) to 587H1-X and 587H1-WT via octet/biolayer inferometry (A) and ELISA (B). Same sera was then blocked using a competitive anti- stem flu bnAb, 27F3 ((A), bottom row) to detect 9114 specific binding. Each line/dot corresponds to a different mouse. *Data for (A) provided by Dr. Shanshan Lang of the Wilson Lab, TSRI

52 C Immunization: OVA > 3x 587H1-X Immunization: OVA > 3x 587H1-WT 587x-ova sera to 587H1 587wt-ova sera to 587H1 587x-ova sera to 587H1 587wt-ova sera to 587H1 5 5 Binding587x-ova to : 587H1 sera -towt 587H1 #1 Binding to : 587H1-wt 5 587x-ova sera to 587H1 5 587wt-ova sera to 587H1 #5 4 #1 587wt-ova sera to 587H1 #2 4 #5 #6 4 5 5 5 #2 #3#1 45 #6 3 #1 3 #7#5 4 #3 #5 34 #4#2 3 4 #7 #8#6 2 #2 4 #6

nm #4 2 #3 nm #8 2 3 Day0 3 Day0#7

nm #3 2 3 Day0 nm 3 #7 1 #4 1 Day0 #8 2 #4 #8

1 nm 2 2 Day0 1 nm

nm 2 Day0 0 Day0 nm 0 Day0 0 1 100 200 300 1 100 200 300 1 01 -1 100 200 300 -1 Time (s) 100 Time200 (s) 300 -1 0 -1 0 0 Time100 (s) 200 300 0 Time100 (s) 200 300 -1 100 200 300 -1 100 200 300 -1 Time (s) -1 Time (s) Time (s) Time (s)

Binding587x-ova to: 587H1 sera to -587H1-27F3wt + 27F3 Binding587wt-ova to: sera 587H1 to 587H1-27F3-wt + 27F3 587x-ova sera to 587H1-27F3 587wt-ova sera to 587H1-27F3 5 587x-ova sera to 587H1-27F3 #1 5 5 587x-ova sera to 587H1-27F3 5 587wt-ova sera to 587H1-27F3 #5 4 #1 587wt-ova sera to 587H1-27F3 #2 4 #5 #6 4 5 5 5 #2 #3#1 45 #6 3 #1 3 #7#5 4 #3 #5 34 #4#2 3 4 #7 #8#6 2 #2 4 #6

nm #4 2 #3 nm #8 2 3 Day0 3 Day0#7

nm #3 2 3 Day0 nm 3 #7 1 #4 1 Day0 #8 2 #4 #8

1 nm 2 2 Day0 1 nm

nm 2 Day0 0 Day0 nm 0 Day0 0 1 100 200 300 1 100 200 300 1 01 -1 100 200 300 -1 Time (s) 100 Time200 (s) 300 -1 0 -1 0 0 Time100 (s) 200 300 0 Time100 (s) 200 300 -1 100 200 300 -1 100 200 300 -1 Time (s) -1 Time (s) Time (s) Time (s) D Immunization: OVA > 3x 587H1-X Immunization: OVA > 3x 587H1-WT

10000 10000

1000 1000

100 100

10 10 [9114 mIgG] µg/mL mIgG] [9114 [9114 mIgG] µg/mL mIgG] [9114

1 1

587-Bio 587-Bio 587X-Bio 587X-Bio (Cont.) Figure 2.10. 587H1-X Immunizations show IFY motif enhancing mutation- specific response via octet and ELISA assays. Binding of mouse sera from four B6 9114 gH mice immunized three times with OVA conjugated 587H1-X (left column) or three times with OVA conjugated 587H1-WT (right column) to 587H1-X-biotin labeled and 587H1-WT-biotin labeled via octet/biolayer inferometry (C) and ELISA (D). Same sera was then blocked using a competitive anti-stem flu bnAb, 27F3 ((C), bottom row) to detect 9114 specific binding. Each line/dot corresponds to a different mouse. *Data for (C) provided by Dr. Shanshan Lang of the Wilson Lab, TSRI

53 A B 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 15 FR1

CDR1

10 FR2

CDR2

5 FR3 CDR3

Mutation numnber AA J 0 MATCHED 587 270 Antigen 587 270

C

CDR1 CDR2

CDR3

Figure 2.11. Immunizations of B6 9114 gH mice with 587H1-X generate mutations in 9114 heavy chain. B cells from B6 9114 gH mice immunized three times with 587X- H1 were sorted for binding to 587H1-WT and 270H5-WT and respective heavy chains were sequenced and analyzed. (A) Number of mutations in heavy chain sequences analyzed. n=33 (587) and n=40 (270). (B) Overview of frequency of mutations in various 9114 heavy chain regions of clones in (A). (C) Amino acids and mutations present at major residues in CDR1-3 in the 9114 heavy chain of 270H5-WT binding B cells. Neutral amino acids are green, hydrophobic amino acids are black, and hydrophillic amino acids are blue. *Data developed alongside Dr. Taka Ota of the Wyatt Lab, TSRI

54 D Clone 3 Clone 4 Clone 18

105 105 105

104 104 104

103 103 103 PE- A : 5 8 7 PE- A : 5 8 7 PE- A : 5 8 7 587

102 102 102

0 0 0

0 102 103 104 105 0 102 103 104 105 0 102 103 104 105 APC- A: IgM APC- A: IgM APC- A: IgM

5 10 105 105

4 10 104 104

3 10 103 103 PE- A : 2 7 0 270 PE- A : 2 7 0 PE- A : 2 7 0

2 10 102 102

0 0 0

2 3 4 5 0 10 10 10 10 0 102 103 104 105 0 102 103 104 105 APC- A: IgM APC- A: IgM APC- A: IgM IgM

(Cont.) Figure 2.11. Immunizations of B6 9114 gH mice with 587H1-X generate mutations in 9114 heavy chain. B cells from B6 9114 gH mice immunized three times with 587X-H1 were sorted for binding to 587H1-WT and 270H5-WT and respective heavy chains were sequenced and analyzed. (D) Heavy chain clones paired with germline 9114 light chain expressed as membrane bound IgM on 293 cells and stained with PE- labeled 587H1 HA (top) and 270H5 HA (bottom). *Data developed alongside Dr. Taka Ota of the Wyatt Lab, TSRI

55

CRISPR

Other VHs DH JH1-4 Cm Cd Cg3 VH81x ... wt

targeting construct

VH1-69*01

targeted

Introduced VH1-69*01

Figure 2.12. Future directions in testing germline targeting immunogens for development of influenza broadly neutralizing antibodies. (A) Development of a VH1-69*01 mouse model allows VH1-69 to rearrange randomly to a variety of DJ segments allowing for development of a number of potential influenza bnAbs. *Figure provided by Dr. David Nemazee

56 REFERENCES

1. Potter, C. W. (2001). "A history of influenza." J Appl Microbiol 91(4): 572-579.

2. Subbarao, K. and T. Joseph (2007). "Scientific barriers to developing vaccines against avian influenza viruses." Nat Rev Immunol 7(4): 267-278.

3. Jang, Y. H. and B. L. Seong (2014). "Options and obstacles for designing a universal influenza vaccine." Viruses 6(8): 3159-3180.

4. Subbarao, K., B. R. Murphy and A. S. Fauci (2006). "Development of effective vaccines against pandemic influenza." Immunity 24(1): 5-9.

5. Altman, M. O., J. R. Bennink, J. W. Yewdell and B. R. Herrin (2015). "Lamprey VLRB response to influenza virus supports universal rules of immunogenicity and antigenicity." Elife 4.

6. Das, S. R., S. E. Hensley, W. L. Ince, C. B. Brooke, A. Subba, M. G. Delboy, G. Russ, J. S. Gibbs, J. R. Bennink and J. W. Yewdell (2013). "Defining influenza A virus hemagglutinin antigenic drift by sequential monoclonal antibody selection." Cell Host Microbe 13(3): 314-323.

7. Obenauer, J. C., J. Denson, P. K. Mehta, X. Su, S. Mukatira, D. B. Finkelstein, X. Xu, J. Wang, J. Ma, Y. Fan, K. M. Rakestraw, R. G. Webster, E. Hoffmann, S. Krauss, J. Zheng, Z. Zhang and C. W. Naeve (2006). "Large-scale sequence analysis of avian influenza isolates." Science 311(5767): 1576-1580.

8. Trifonov, V., H. Khiabanian and R. Rabadan (2009). "Geographic dependence, surveillance, and origins of the 2009 influenza A (H1N1) virus." N Engl J Med 361(2): 115-119.

9. Medina, R. A., S. Stertz, B. Manicassamy, P. Zimmermann, X. Sun, R. A. Albrecht, H. Uusi-Kerttula, O. Zagordi, R. B. Belshe, S. E. Frey, T. M. Tumpey and A. Garcia-Sastre (2013). "Glycosylations in the globular head of the hemagglutinin protein modulate the virulence and antigenic properties of the H1N1 influenza viruses." Sci Transl Med 5(187): 187ra170.

10. Lambert, L. C. and A. S. Fauci (2010). "Influenza vaccines for the future." N Engl J Med 363(21): 2036-2044.

11. Salomon, R. and R. G. Webster (2009). "The influenza virus enigma." Cell 136(3): 402-410.

57 12. Nabel, G. J. and A. S. Fauci (2010). "Induction of unnatural immunity: prospects for a broadly protective universal influenza vaccine." Nat Med 16(12): 1389- 1391.

13. Haynes BF, Bradley T. Broadly Neutralizing Antibodies and the Development of Vaccines. JAMA.2015;313(24):2419–2420. doi:10.1001/jama.2015.2427

14. Hammers, C. M. and J. R. Stanley (2014). "Antibody phage display: technique and applications." J Invest Dermatol 134(2): 1-5.

15. Lerner, R. A. (2011). "Rare antibodies from combinatorial libraries suggests an S.O.S. component of the human immunological repertoire." Mol Biosyst 7(4): 1004-1012.

16. Sui, J., W. C. Hwang, S. Perez, G. Wei, D. Aird, L. M. Chen, E. Santelli, B. Stec, G. Cadwell, M. Ali, H. Wan, A. Murakami, A. Yammanuru, T. Han, N. J. Cox, L. A. Bankston, R. O. Donis, R. C. Liddington and W. A. Marasco (2009). "Structural and functional bases for broad-spectrum neutralization of avian and human influenza A viruses." Nat Struct Mol Biol 16(3): 265-273.

17. Kashyap, A. K., J. Steel, A. F. Oner, M. A. Dillon, R. E. Swale, K. M. Wall, K. J. Perry, A. Faynboym, M. Ilhan, M. Horowitz, L. Horowitz, P. Palese, R. R. Bhatt and R. A. Lerner (2008). "Combinatorial antibody libraries from survivors of the Turkish H5N1 avian influenza outbreak reveal virus neutralization strategies." Proc Natl Acad Sci U S A 105(16): 5986-5991.

18. Throsby, M., E. van den Brink, M. Jongeneelen, L. L. Poon, P. Alard, L. Cornelissen, A. Bakker, F. Cox, E. van Deventer, Y. Guan, J. Cinatl, J. ter Meulen, I. Lasters, R. Carsetti, M. Peiris, J. de Kruif and J. Goudsmit (2008). "Heterosubtypic neutralizing monoclonal antibodies cross-protective against H5N1 and H1N1 recovered from human IgM+ memory B cells." PLoS One 3(12): e3942.

19. Ekiert, D. C., G. Bhabha, M. A. Elsliger, R. H. Friesen, M. Jongeneelen, M. Throsby, J. Goudsmit and I. A. Wilson (2009). "Antibody recognition of a highly conserved influenza virus epitope." Science 324(5924): 246-251.

20. Dreyfus, C., N. S. Laursen, T. Kwaks, D. Zuijdgeest, R. Khayat, D. C. Ekiert, J. H. Lee, Z. Metlagel, M. V. Bujny, M. Jongeneelen, R. van der Vlugt, M. Lamrani, H. J. Korse, E. Geelen, O. Sahin, M. Sieuwerts, J. P. Brakenhoff, R. Vogels, O. T. Li, L. L. Poon, M. Peiris, W. Koudstaal, A. B. Ward, I. A. Wilson, J. Goudsmit and R. H. Friesen (2012). "Highly conserved protective epitopes on influenza B viruses." Science 337(6100): 1343-1348.

58 21. Pappas, L., M. Foglierini, L. Piccoli, N. L. Kallewaard, F. Turrini, C. Silacci, B. Fernandez-Rodriguez, G. Agatic, I. Giacchetto-Sasselli, G. Pellicciotta, F. Sallusto, Q. Zhu, E. Vicenzi, D. Corti and A. Lanzavecchia (2014). "Rapid development of broadly influenza neutralizing antibodies through redundant mutations." Nature 516(7531): 418-422.

22. Ekiert, D. C., A. K. Kashyap, J. Steel, A. Rubrum, G. Bhabha, R. Khayat, J. H. Lee, M. A. Dillon, R. E. O'Neil, A. M. Faynboym, M. Horowitz, L. Horowitz, A. B. Ward, P. Palese, R. Webby, R. A. Lerner, R. R. Bhatt and I. A. Wilson (2012). "Cross-neutralization of influenza A viruses mediated by a single antibody loop." Nature 489(7417): 526-532.

23. Lang, S., J. Xie, X. Zhu, N. C. Wu, R. A. Lerner and I. A. Wilson (2017). "Antibody 27F3 Broadly Targets Influenza A Group 1 and 2 Hemagglutinins through a Further Variation in VH1-69 Antibody Orientation on the HA Stem." Cell Rep 20(12): 2935-2943.

24. Dreyfus, C., D. C. Ekiert and I. A. Wilson (2013). "Structure of a classical broadly neutralizing stem antibody in complex with a pandemic H2 influenza virus hemagglutinin." J Virol 87(12): 7149-7154.

25. Joyce, M. G., A. K. Wheatley, P. V. Thomas, G. Y. Chuang, C. Soto, R. T. Bailer, A. Druz, I. S. Georgiev, R. A. Gillespie, M. Kanekiyo, W. P. Kong, K. Leung, S. N. Narpala, M. S. Prabhakaran, E. S. Yang, B. Zhang, Y. Zhang, M. Asokan, J. C. Boyington, T. Bylund, S. Darko, C. R. Lees, A. Ransier, C. H. Shen, L. Wang, J. R. Whittle, X. Wu, H. M. Yassine, C. Santos, Y. Matsuoka, Y. Tsybovsky, U. Baxa, N. C. S. Program, J. C. Mullikin, K. Subbarao, D. C. Douek, B. S. Graham, R. A. Koup, J. E. Ledgerwood, M. Roederer, L. Shapiro, P. D. Kwong, J. R. Mascola and A. B. McDermott (2016). "Vaccine-Induced Antibodies that Neutralize Group 1 and Group 2 Influenza A Viruses." Cell 166(3): 609-623.

26. Ekiert, D. C., R. H. Friesen, G. Bhabha, T. Kwaks, M. Jongeneelen, W. Yu, C. Ophorst, F. Cox, H. J. Korse, B. Brandenburg, R. Vogels, J. P. Brakenhoff, R. Kompier, M. H. Koldijk, L. A. Cornelissen, L. L. Poon, M. Peiris, W. Koudstaal, I. A. Wilson and J. Goudsmit (2011). "A highly conserved neutralizing epitope on group 2 influenza A viruses." Science 333(6044): 843-850.

27. Friesen, R. H., P. S. Lee, E. J. Stoop, R. M. Hoffman, D. C. Ekiert, G. Bhabha, W. Yu, J. Juraszek, W. Koudstaal, M. Jongeneelen, H. J. Korse, C. Ophorst, E. C. Brinkman-van der Linden, M. Throsby, M. J. Kwakkenbos, A. Q. Bakker, T. Beaumont, H. Spits, T. Kwaks, R. Vogels, A. B. Ward, J. Goudsmit and I. A. Wilson (2014). "A common solution to group 2 influenza virus neutralization." Proc Natl Acad Sci U S A 111(1): 445-450.

59 28. Sutton, T. C., E. W. Lamirande, K. W. Bock, I. N. Moore, W. Koudstaal, M. Rehman, G. J. Weverling, J. Goudsmit and K. Subbarao (2017). "In Vitro Neutralization Is Not Predictive of Prophylactic Efficacy of Broadly Neutralizing Monoclonal Antibodies CR6261 and CR9114 against Lethal H2 Influenza Virus Challenge in Mice." J Virol 91(24).

29. Fleishman, S. J., T. A. Whitehead, D. C. Ekiert, C. Dreyfus, J. E. Corn, E. M. Strauch, I. A. Wilson and D. Baker (2011). "Computational design of proteins targeting the conserved stem region of influenza hemagglutinin." Science 332(6031): 816-821.

30. Yassine, H. M., J. C. Boyington, P. M. McTamney, C. J. Wei, M. Kanekiyo, W. P. Kong, J. R. Gallagher, L. Wang, Y. Zhang, M. G. Joyce, D. Lingwood, S. M. Moin, H. Andersen, Y. Okuno, S. S. Rao, A. K. Harris, P. D. Kwong, J. R. Mascola, G. J. Nabel and B. S. Graham (2015). "Hemagglutinin-stem nanoparticles generate heterosubtypic influenza protection." Nat Med 21(9): 1065-1070.

31. Steel, J., A. C. Lowen, T. T. Wang, M. Yondola, Q. Gao, K. Haye, A. Garcia- Sastre and P. Palese (2010). "Influenza virus vaccine based on the conserved hemagglutinin stalk domain." MBio 1(1).

32. Impagliazzo, A., F. Milder, H. Kuipers, M. V. Wagner, X. Zhu, R. M. Hoffman, R. van Meersbergen, J. Huizingh, P. Wanningen, J. Verspuij, M. de Man, Z. Ding, A. Apetri, B. Kukrer, E. Sneekes-Vriese, D. Tomkiewicz, N. S. Laursen, P. S. Lee, A. Zakrzewska, L. Dekking, J. Tolboom, L. Tettero, S. van Meerten, W. Yu, W. Koudstaal, J. Goudsmit, A. B. Ward, W. Meijberg, I. A. Wilson and K. Radosevic (2015). "A stable trimeric influenza hemagglutinin stem as a broadly protective immunogen." Science 349(6254): 1301-1306.

33. Ohshima, N., R. Kubota-Koketsu, Y. Iba, Y. Okuno and Y. Kurosawa (2014). "Two types of antibodies are induced by vaccination with A/California/2009 pdm virus: binding near the sialic acid-binding pocket and neutralizing both H1N1 and H5N1 viruses." PLoS One 9(2): e87305.

34. Fagerberg, J., M. Steinitz, H. Wigzell, P. Askelof and H. Mellstedt (1995). "Human anti-idiotypic antibodies induced a humoral and cellular immune response against a colorectal carcinoma-associated antigen in patients." Proc Natl Acad Sci U S A 92(11): 4773-4777.

35. Yewdell, J. W., D. J. Spiro, H. Golding, H. Quill, A. Mittelman and G. J. Nabel (2013). "Getting to the heart of influenza." Sci Transl Med 5(191): 191ed198.

36. Moscona, A. (2005). "Neuraminidase inhibitors for influenza." N Engl J Med 353(13): 1363-1373.

60

37. Hayden, F. G. and M. D. de Jong (2011). "Emerging influenza antiviral resistance threats." J Infect Dis 203(1): 6-10.

38. Van der Most, R. G., F. P. Roman, B. Innis, E. Hanon, D. W. Vaughn, P. Gillard, K. Walravens and M. Wettendorff (2014). "Seeking help: B cells adapting to flu variability." Sci Transl Med 6(246): 246ps248.

39. Kanekiyo, M., C. J. Wei, H. M. Yassine, P. M. McTamney, J. C. Boyington, J. R. Whittle, S. S. Rao, W. P. Kong, L. Wang and G. J. Nabel (2013). "Self- assembling influenza nanoparticle vaccines elicit broadly neutralizing H1N1 antibodies." Nature 499(7456): 102-106.

40. Grodeland, G., S. Mjaaland, K. H. Roux, A. B. Fredriksen and B. Bogen (2013). "DNA vaccine that targets hemagglutinin to MHC class II molecules rapidly induces antibody-mediated protection against influenza." J Immunol 191(6): 3221-3231.

41. Grodeland, G., A. B. Fredriksen, G. A. Loset, E. Vikse, L. Fugger and B. Bogen (2016). "Antigen Targeting to Human HLA Class II Molecules Increases Efficacy of DNA Vaccination." J Immunol 197(9): 3575-3585.

42. Wei, C. J., H. M. Yassine, P. M. McTamney, J. G. Gall, J. R. Whittle, J. C. Boyington and G. J. Nabel (2012). "Elicitation of broadly neutralizing influenza antibodies in animals with previous influenza exposure." Sci Transl Med 4(147): 147ra114.

43. A. D'Arcy, F. Villard, M. Marsh, An automated microseed matrix-screening method for protein crystallization. Acta Crystallogr. D Biol. Crystallogr. 63, 550- 554 (2007).

44. Mccoy, A. J., R. W. Grosse-Kunstleve, P. D. Adams, M. D. Winn, L. C. Storoni and R. J. Read (2007). "Phaser crystallographic software." Journal of Applied Crystallography 40: 658-674.

45. P. Skubak, G. N. Murshudov, N. S. Pannu, Direct incorporation of experimental phase information in model refinement. Acta Crystallogr. D Biol. Crystallogr. 60, 2196-2201 (2004).

46. Adams, P. D., P. V. Afonine, G. Bunkoczi, V. B. Chen, I. W. Davis, N. Echols, J. J. Headd, L. W. Hung, G. J. Kapral, R. W. Grosse-Kunstleve, A. J. McCoy, N. W. Moriarty, R. Oeffner, R. J. Read, D. C. Richardson, J. S. Richardson, T. C. Terwilliger and P. H. Zwart (2010). "PHENIX: a comprehensive Python-based system for macromolecular structure solution." Acta Crystallographica Section D- Biological Crystallography 66: 213-221.

61

47. P. Emsley, K. Cowtan, Coot: model-building tools for molecular graphics. Acta Crystallogr. D Biol. Crystallogr. 60, 2126-2132 (2004).

48. Chen, V. B., W. B. Arendall, J. J. Headd, D. A. Keedy, R. M. Immormino, G. J. Kapral, L. W. Murray, J. S. Richardson and D. C. Richardson (2010). "MolProbity: all-atom structure validation for macromolecular crystallography." Acta Crystallographica Section D-Biological Crystallography 66: 12-21.

49. S. Sheriff, W. A. Hendrickson, J. L. Smith, Structure of myohemerythrin in the azidomet state at 1.7/1.3 Å resolution. J. Mol. Biol. 197, 273-296 (1987).

50. M. L. Connolly, Analytical molecular-surface calculation. J. Appl. Crystallogr. 16, 548-558 (1983).

51. Dolinsky, T. J., P. Czodrowski, H. Li, J. E. Nielsen, J. H. Jensen, G. Klebe and N. A. Baker (2007). "PDB2PQR: expanding and upgrading automated preparation of biomolecular structures for molecular simulations." Nucleic Acids Research 35: W522-W525.

52. T. J. Dolinsky, J. E. Nielsen, J. A. McCammon, N. A. Baker, PDB2PQR: an automated pipeline for the setup of Poisson-Boltzmann electrostatics calculations. Nucleic Acids Res. 32, W665-667 (2004).

53. N. A. Baker, D. Sept, S. Joseph, M. J. Holst, J. A. McCammon, Electrostatics of nanosystems: application to microtubules and the ribosome. Proc. Natl. Acad. Sci. USA 98, 10037-10041 (2001).

54. Hong, M. S., P. S. Lee, R. M. B. Hoffman, X. Y. Zhu, J. C. Krause, N. S. Laursen, S. I. Yoon, L. Z. Song, L. Tussey, J. E. Crowe, A. B. Ward and I. A. Wilson (2013). "Antibody Recognition of the Pandemic H1N1 Influenza Virus Hemagglutinin Receptor Binding Site." Journal of Virology 87(22): 12471-12480.

55. Jardine, J. G., T. Ota, D. Sok, M. Pauthner, D. W. Kulp, O. Kalyuzhniy, P. D. Skog, T. C. Thinnes, D. Bhullar, B. Briney, S. Menis, M. Jones, M. Kubitz, S. Spencer, Y. Adachi, D. R. Burton, W. R. Schief and D. Nemazee (2015). "HIV-1 VACCINES. Priming a broadly neutralizing antibody response to HIV-1 using a germline-targeting immunogen." Science 349(6244): 156-161.

56. Ota, T., C. Doyle-Cooper, A. B. Cooper, K. J. Doores, M. Aoki-Ota, K. Le, W. R. Schief, R. T. Wyatt, D. R. Burton and D. Nemazee (2013). "B cells from knock-in mice expressing broadly neutralizing HIV antibody b12 carry an innocuous B cell receptor responsive to HIV vaccine candidates." J Immunol 191(6): 3179-3185.

62